2006-10-01 18:30:10 -04:00
/* -*- mode: C; c-file-style: "gnu"; indent-tabs-mode: nil; -*- */
2008-12-18 07:41:56 -05:00
/*
2001-08-05 12:04:52 -04:00
* Copyright ( C ) 2001 Havoc Pennington , Anders Carlsson
2003-02-23 12:09:46 -05:00
* Copyright ( C ) 2002 , 2003 Red Hat , Inc .
* Copyright ( C ) 2003 Rob Adams
2006-01-10 14:43:21 -05:00
* Copyright ( C ) 2004 - 2006 Elijah Newren
2008-12-18 07:41:56 -05:00
*
2001-05-30 11:36:31 -04:00
* This program is free software ; you can redistribute it and / or
* modify it under the terms of the GNU General Public License as
* published by the Free Software Foundation ; either version 2 of the
* License , or ( at your option ) any later version .
*
* This program is distributed in the hope that it will be useful , but
* WITHOUT ANY WARRANTY ; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE . See the GNU
* General Public License for more details .
2008-12-18 07:41:56 -05:00
*
2001-05-30 11:36:31 -04:00
* You should have received a copy of the GNU General Public License
2014-01-11 20:42:06 -05:00
* along with this program ; if not , see < http : //www.gnu.org/licenses/>.
2001-05-30 11:36:31 -04:00
*/
2013-03-11 11:52:36 -04:00
/**
* SECTION : window
* @ title : MetaWindow
* @ short_description : Mutter X managed windows
*/
2001-12-09 17:41:12 -05:00
# include <config.h>
2008-05-18 20:00:09 -04:00
# include "window-private.h"
2010-10-18 14:34:14 -04:00
# include "boxes-private.h"
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
# include "edge-resistance.h"
2013-09-11 04:18:53 -04:00
# include "util-private.h"
2011-03-05 19:29:12 -05:00
# include "frame.h"
# include <meta/errors.h>
2008-09-18 11:09:11 -04:00
# include "workspace-private.h"
2001-06-10 14:46:46 -04:00
# include "stack.h"
2009-01-29 09:16:01 -05:00
# include "keybindings-private.h"
2001-06-23 01:49:35 -04:00
# include "ui.h"
2001-06-24 02:47:54 -04:00
# include "place.h"
2011-03-05 19:29:12 -05:00
# include <meta/prefs.h>
2014-03-19 09:11:23 -04:00
# include <meta/group.h>
2002-03-17 12:22:23 -05:00
# include "resizepopup.h"
2003-03-08 18:43:17 -05:00
# include "constraints.h"
2009-08-04 03:06:30 -04:00
# include "mutter-enum-types.h"
2013-02-07 18:20:03 -05:00
# include "core.h"
2001-06-06 00:47:37 -04:00
2001-06-02 00:14:18 -04:00
# include <X11/Xatom.h>
2013-03-04 09:57:01 -05:00
# include <stdlib.h>
2002-10-26 12:40:50 -04:00
# include <string.h>
2012-03-12 12:09:17 -04:00
# include <math.h>
2001-06-02 00:14:18 -04:00
2003-01-05 02:51:02 -05:00
# ifdef HAVE_SHAPE
# include <X11/extensions/shape.h>
# endif
2013-10-03 15:29:44 -04:00
# include "meta/compositor-mutter.h"
2013-08-12 12:04:34 -04:00
2014-03-18 21:56:45 -04:00
# include "x11/window-x11.h"
# include "x11/window-props.h"
# include "x11/xprops.h"
2014-03-18 22:02:47 -04:00
# include "x11/session.h"
2014-03-18 21:56:45 -04:00
2014-03-19 09:26:17 -04:00
# include "wayland/window-wayland.h"
2014-03-18 22:01:31 -04:00
# include "wayland/meta-wayland-private.h"
2012-03-12 12:09:17 -04:00
/* Windows that unmaximize to a size bigger than that fraction of the workarea
2012-03-15 15:42:11 -04:00
* will be scaled down to that size ( while maintaining aspect ratio ) .
* Windows that cover an area greater then this size are automaximized on map .
*/
2012-03-12 12:09:17 -04:00
# define MAX_UNMAXIMIZED_WINDOW_AREA .8
2002-01-18 22:50:03 -05:00
static int destroying_windows_disallowed = 0 ;
2001-06-09 17:58:30 -04:00
2003-02-27 20:24:44 -05:00
static void invalidate_work_areas ( MetaWindow * window ) ;
2014-02-01 19:03:33 -05:00
static void set_wm_state ( MetaWindow * window ) ;
2002-10-21 17:44:35 -04:00
static void set_net_wm_state ( MetaWindow * window ) ;
2011-07-02 07:08:17 -04:00
static void meta_window_set_above ( MetaWindow * window ,
gboolean new_value ) ;
2004-01-10 12:16:07 -05:00
2011-02-02 19:14:27 -05:00
static void meta_window_force_placement ( MetaWindow * window ) ;
2001-06-06 00:47:37 -04:00
static void meta_window_show ( MetaWindow * window ) ;
static void meta_window_hide ( MetaWindow * window ) ;
2002-09-24 17:33:34 -04:00
static void meta_window_save_rect ( MetaWindow * window ) ;
2008-08-15 23:09:40 -04:00
static void save_user_window_placement ( MetaWindow * window ) ;
static void force_save_user_window_placement ( MetaWindow * window ) ;
2002-09-24 17:33:34 -04:00
2001-08-28 23:37:03 -04:00
static void meta_window_move_resize_internal ( MetaWindow * window ,
MetaMoveResizeFlags flags ,
int resize_gravity ,
int root_x_nw ,
int root_y_nw ,
int w ,
int h ) ;
2001-07-28 02:35:19 -04:00
2004-08-05 13:48:39 -04:00
static void ensure_mru_position_after ( MetaWindow * window ,
MetaWindow * after_this_one ) ;
2001-06-02 00:14:18 -04:00
2006-01-20 17:03:56 -05:00
static void meta_window_move_resize_now ( MetaWindow * window ) ;
2001-06-23 14:30:27 -04:00
2007-06-16 22:15:07 -04:00
static void meta_window_unqueue ( MetaWindow * window , guint queuebits ) ;
Refactor thrice-duplicated queue code in window.c. Closes #376760.
2007-06-10 Thomas Thurman <thomas@thurman.org.uk>
Refactor thrice-duplicated queue code in window.c. Closes #376760.
* src/window.c (meta_window_queue, meta_window_unqueue):
New functions.
* src/window.[ch] (meta_window_unqueue_*, meta_window_queue_*):
Removed functions.
* src/window.c (meta_window_new_with_attrs, meta_window_free,
meta_window_flush_calc_showing, queue_calc_showing_func,
meta_window_minimize, meta_window_unminimize, meta_window_maximize,
meta_window_make_fullscreen, meta_window_shade,
meta_window_unshade, meta_window_move_resize_internal,
window_stick_impl, window_unstick_impl,
meta_window_client_message, process_property_notify): Modified to
use new queueing functions.
* src/window.c (idle_move_resize, idle_update_icon,
idle_calc_showing): update to receive queue number from pointer.
* src/window.h (MetaQueueType): new enum.
* src/window.h (MetaWindow): *_queued replaced with is_in_queue
bitfield.
* src/core.c (meta_core_queue_frame_resize):
* src/display.c (event_callback,
meta_display_queue_retheme_all_windows): Using new queueing functions.
* src/frame.c (meta_window_destroy_frame): Using new queueing functions.
* src/screen.c (queue_resize, meta_screen_resize_func,
queue_windows_showing): Using new queueing functions.
* src/window-props.c (reload_mwm_hints, reload_wm_hints,
reload_transient_for): Using new queueing functions.
* src/workspace.c (meta_workspace_add_window,
meta_workspace_remove_window, meta_workspace_queue_calc_showing,
meta_workspace_invalidate_work_area): Using new queueing functions.
svn path=/trunk/; revision=3236
2007-06-10 21:15:33 -04:00
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
static void update_move ( MetaWindow * window ,
gboolean snap ,
int x ,
int y ) ;
static gboolean update_move_timeout ( gpointer data ) ;
static void update_resize ( MetaWindow * window ,
gboolean snap ,
int x ,
int y ,
gboolean force ) ;
static gboolean update_resize_timeout ( gpointer data ) ;
2011-02-28 05:41:23 -05:00
static gboolean should_be_on_all_workspaces ( MetaWindow * window ) ;
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
2006-01-20 17:03:56 -05:00
static void meta_window_flush_calc_showing ( MetaWindow * window ) ;
2001-06-21 23:20:21 -04:00
2005-05-26 13:15:56 -04:00
static gboolean queue_calc_showing_func ( MetaWindow * window ,
void * data ) ;
2001-07-27 00:36:44 -04:00
static void meta_window_apply_session_info ( MetaWindow * window ,
const MetaWindowSessionInfo * info ) ;
2011-05-20 04:56:12 -04:00
static void meta_window_move_between_rects ( MetaWindow * window ,
const MetaRectangle * old_area ,
const MetaRectangle * new_area ) ;
2001-07-27 00:36:44 -04:00
2006-04-16 23:54:49 -04:00
static void unmaximize_window_before_freeing ( MetaWindow * window ) ;
static void unminimize_window_and_all_transient_parents ( MetaWindow * window ) ;
2009-09-25 15:47:22 -04:00
/* Idle handlers for the three queues (run with meta_later_add()). The
* " data " parameter in each case will be a GINT_TO_POINTER of the
* index into the queue arrays to use .
Refactor thrice-duplicated queue code in window.c. Closes #376760.
2007-06-10 Thomas Thurman <thomas@thurman.org.uk>
Refactor thrice-duplicated queue code in window.c. Closes #376760.
* src/window.c (meta_window_queue, meta_window_unqueue):
New functions.
* src/window.[ch] (meta_window_unqueue_*, meta_window_queue_*):
Removed functions.
* src/window.c (meta_window_new_with_attrs, meta_window_free,
meta_window_flush_calc_showing, queue_calc_showing_func,
meta_window_minimize, meta_window_unminimize, meta_window_maximize,
meta_window_make_fullscreen, meta_window_shade,
meta_window_unshade, meta_window_move_resize_internal,
window_stick_impl, window_unstick_impl,
meta_window_client_message, process_property_notify): Modified to
use new queueing functions.
* src/window.c (idle_move_resize, idle_update_icon,
idle_calc_showing): update to receive queue number from pointer.
* src/window.h (MetaQueueType): new enum.
* src/window.h (MetaWindow): *_queued replaced with is_in_queue
bitfield.
* src/core.c (meta_core_queue_frame_resize):
* src/display.c (event_callback,
meta_display_queue_retheme_all_windows): Using new queueing functions.
* src/frame.c (meta_window_destroy_frame): Using new queueing functions.
* src/screen.c (queue_resize, meta_screen_resize_func,
queue_windows_showing): Using new queueing functions.
* src/window-props.c (reload_mwm_hints, reload_wm_hints,
reload_transient_for): Using new queueing functions.
* src/workspace.c (meta_workspace_add_window,
meta_workspace_remove_window, meta_workspace_queue_calc_showing,
meta_workspace_invalidate_work_area): Using new queueing functions.
svn path=/trunk/; revision=3236
2007-06-10 21:15:33 -04:00
*
* TODO : Possibly there is still some code duplication among these , which we
* need to sort out at some point .
*/
static gboolean idle_calc_showing ( gpointer data ) ;
static gboolean idle_move_resize ( gpointer data ) ;
static gboolean idle_update_icon ( gpointer data ) ;
2006-01-20 18:36:23 -05:00
2014-03-19 09:36:36 -04:00
G_DEFINE_ABSTRACT_TYPE ( MetaWindow , meta_window , G_TYPE_OBJECT ) ;
2008-11-07 04:57:00 -05:00
2008-12-05 20:03:27 -05:00
enum {
PROP_0 ,
PROP_TITLE ,
PROP_ICON ,
2008-12-18 07:41:56 -05:00
PROP_MINI_ICON ,
2009-02-11 07:16:41 -05:00
PROP_DECORATED ,
2009-03-17 08:27:13 -04:00
PROP_FULLSCREEN ,
2010-02-12 15:26:52 -05:00
PROP_MAXIMIZED_HORIZONTALLY ,
PROP_MAXIMIZED_VERTICALLY ,
2011-05-31 13:52:04 -04:00
PROP_MINIMIZED ,
2009-08-31 12:06:53 -04:00
PROP_WINDOW_TYPE ,
2009-11-17 05:02:06 -05:00
PROP_USER_TIME ,
2009-11-20 03:51:19 -05:00
PROP_DEMANDS_ATTENTION ,
2010-01-18 05:12:17 -05:00
PROP_URGENT ,
2014-01-30 16:39:54 -05:00
PROP_SKIP_TASKBAR ,
2011-03-24 21:36:47 -04:00
PROP_MUTTER_HINTS ,
2011-03-03 04:01:04 -05:00
PROP_APPEARS_FOCUSED ,
2011-07-02 07:08:17 -04:00
PROP_RESIZEABLE ,
PROP_ABOVE ,
2011-11-25 18:42:41 -05:00
PROP_WM_CLASS ,
2012-01-17 13:27:10 -05:00
PROP_GTK_APPLICATION_ID ,
PROP_GTK_UNIQUE_BUS_NAME ,
PROP_GTK_APPLICATION_OBJECT_PATH ,
PROP_GTK_WINDOW_OBJECT_PATH ,
PROP_GTK_APP_MENU_OBJECT_PATH ,
PROP_GTK_MENUBAR_OBJECT_PATH
2008-12-05 20:03:27 -05:00
} ;
2009-01-20 07:52:06 -05:00
enum
{
WORKSPACE_CHANGED ,
2009-02-02 06:00:13 -05:00
FOCUS ,
2009-02-02 09:08:46 -05:00
RAISED ,
2009-10-13 10:34:34 -04:00
UNMANAGED ,
2013-12-06 15:39:24 -05:00
SIZE_CHANGED ,
POSITION_CHANGED ,
2009-01-20 07:52:06 -05:00
LAST_SIGNAL
} ;
static guint window_signals [ LAST_SIGNAL ] = { 0 } ;
2011-11-25 19:52:24 -05:00
static void
prefs_changed_callback ( MetaPreference pref ,
gpointer data )
{
MetaWindow * window = data ;
2012-07-13 18:22:51 -04:00
if ( pref = = META_PREF_WORKSPACES_ONLY_ON_PRIMARY )
{
meta_window_update_on_all_workspaces ( window ) ;
meta_window_queue ( window , META_QUEUE_CALC_SHOWING ) ;
}
else if ( pref = = META_PREF_ATTACH_MODAL_DIALOGS & &
window - > type = = META_WINDOW_MODAL_DIALOG )
{
window - > attached = meta_window_should_attach_to_parent ( window ) ;
2014-02-01 18:58:58 -05:00
meta_window_recalc_features ( window ) ;
2012-07-13 18:22:51 -04:00
meta_window_queue ( window , META_QUEUE_MOVE_RESIZE ) ;
}
2011-11-25 19:52:24 -05:00
}
2014-03-18 11:20:22 -04:00
static void
meta_window_real_get_default_skip_hints ( MetaWindow * window ,
gboolean * skip_taskbar_out ,
gboolean * skip_pager_out )
{
* skip_taskbar_out = FALSE ;
* skip_pager_out = FALSE ;
}
2008-11-07 04:57:00 -05:00
static void
meta_window_finalize ( GObject * object )
{
MetaWindow * window = META_WINDOW ( object ) ;
2008-12-18 07:41:56 -05:00
2008-11-07 04:57:00 -05:00
if ( window - > icon )
g_object_unref ( G_OBJECT ( window - > icon ) ) ;
if ( window - > mini_icon )
g_object_unref ( G_OBJECT ( window - > mini_icon ) ) ;
2011-03-22 15:36:12 -04:00
if ( window - > frame_bounds )
cairo_region_destroy ( window - > frame_bounds ) ;
2013-09-14 08:19:46 -04:00
if ( window - > shape_region )
cairo_region_destroy ( window - > shape_region ) ;
2013-01-14 20:45:31 -05:00
if ( window - > opaque_region )
cairo_region_destroy ( window - > opaque_region ) ;
2013-09-16 08:44:37 -04:00
if ( window - > transient_for )
g_object_unref ( window - > transient_for ) ;
2008-11-07 04:57:00 -05:00
meta_icon_cache_free ( & window - > icon_cache ) ;
2008-12-18 07:41:56 -05:00
2008-11-07 04:57:00 -05:00
g_free ( window - > sm_client_id ) ;
g_free ( window - > wm_client_machine ) ;
g_free ( window - > startup_id ) ;
g_free ( window - > role ) ;
g_free ( window - > res_class ) ;
g_free ( window - > res_name ) ;
g_free ( window - > title ) ;
g_free ( window - > desc ) ;
2011-03-19 05:31:25 -04:00
g_free ( window - > gtk_theme_variant ) ;
2012-01-17 13:27:10 -05:00
g_free ( window - > gtk_application_id ) ;
g_free ( window - > gtk_unique_bus_name ) ;
g_free ( window - > gtk_application_object_path ) ;
g_free ( window - > gtk_window_object_path ) ;
g_free ( window - > gtk_app_menu_object_path ) ;
g_free ( window - > gtk_menubar_object_path ) ;
2013-04-23 08:47:19 -04:00
G_OBJECT_CLASS ( meta_window_parent_class ) - > finalize ( object ) ;
2008-11-07 04:57:00 -05:00
}
2008-12-05 20:03:27 -05:00
static void
meta_window_get_property ( GObject * object ,
guint prop_id ,
GValue * value ,
GParamSpec * pspec )
{
MetaWindow * win = META_WINDOW ( object ) ;
switch ( prop_id )
{
case PROP_TITLE :
g_value_set_string ( value , win - > title ) ;
break ;
case PROP_ICON :
g_value_set_object ( value , win - > icon ) ;
break ;
case PROP_MINI_ICON :
g_value_set_object ( value , win - > mini_icon ) ;
2008-12-18 07:41:56 -05:00
break ;
2009-02-11 07:16:41 -05:00
case PROP_DECORATED :
g_value_set_boolean ( value , win - > decorated ) ;
break ;
2009-03-17 08:27:13 -04:00
case PROP_FULLSCREEN :
g_value_set_boolean ( value , win - > fullscreen ) ;
break ;
2010-02-12 15:26:52 -05:00
case PROP_MAXIMIZED_HORIZONTALLY :
g_value_set_boolean ( value , win - > maximized_horizontally ) ;
break ;
case PROP_MAXIMIZED_VERTICALLY :
g_value_set_boolean ( value , win - > maximized_vertically ) ;
break ;
2011-05-31 13:52:04 -04:00
case PROP_MINIMIZED :
g_value_set_boolean ( value , win - > minimized ) ;
break ;
2009-08-04 03:06:30 -04:00
case PROP_WINDOW_TYPE :
g_value_set_enum ( value , win - > type ) ;
break ;
2009-08-31 12:06:53 -04:00
case PROP_USER_TIME :
g_value_set_uint ( value , win - > net_wm_user_time ) ;
break ;
2009-11-17 05:02:06 -05:00
case PROP_DEMANDS_ATTENTION :
g_value_set_boolean ( value , win - > wm_state_demands_attention ) ;
break ;
2009-11-20 03:51:19 -05:00
case PROP_URGENT :
2014-03-18 13:53:42 -04:00
g_value_set_boolean ( value , win - > urgent ) ;
2009-11-20 03:51:19 -05:00
break ;
2014-01-30 16:39:54 -05:00
case PROP_SKIP_TASKBAR :
g_value_set_boolean ( value , win - > skip_taskbar ) ;
break ;
2010-01-18 05:12:17 -05:00
case PROP_MUTTER_HINTS :
g_value_set_string ( value , win - > mutter_hints ) ;
break ;
2011-03-24 21:36:47 -04:00
case PROP_APPEARS_FOCUSED :
g_value_set_boolean ( value , meta_window_appears_focused ( win ) ) ;
break ;
2011-03-03 04:01:04 -05:00
case PROP_WM_CLASS :
g_value_set_string ( value , win - > res_class ) ;
break ;
2011-07-02 07:08:17 -04:00
case PROP_RESIZEABLE :
g_value_set_boolean ( value , win - > has_resize_func ) ;
break ;
case PROP_ABOVE :
g_value_set_boolean ( value , win - > wm_state_above ) ;
break ;
2012-01-17 13:27:10 -05:00
case PROP_GTK_APPLICATION_ID :
g_value_set_string ( value , win - > gtk_application_id ) ;
2011-11-25 18:42:41 -05:00
break ;
2012-01-17 13:27:10 -05:00
case PROP_GTK_UNIQUE_BUS_NAME :
g_value_set_string ( value , win - > gtk_unique_bus_name ) ;
2011-12-05 12:52:35 -05:00
break ;
2012-01-17 13:27:10 -05:00
case PROP_GTK_APPLICATION_OBJECT_PATH :
g_value_set_string ( value , win - > gtk_application_object_path ) ;
break ;
case PROP_GTK_WINDOW_OBJECT_PATH :
g_value_set_string ( value , win - > gtk_window_object_path ) ;
break ;
case PROP_GTK_APP_MENU_OBJECT_PATH :
g_value_set_string ( value , win - > gtk_app_menu_object_path ) ;
break ;
case PROP_GTK_MENUBAR_OBJECT_PATH :
g_value_set_string ( value , win - > gtk_menubar_object_path ) ;
2011-12-05 12:52:35 -05:00
break ;
2008-12-05 20:03:27 -05:00
default :
G_OBJECT_WARN_INVALID_PROPERTY_ID ( object , prop_id , pspec ) ;
2008-12-18 07:41:56 -05:00
break ;
2008-12-05 20:03:27 -05:00
}
}
static void
meta_window_set_property ( GObject * object ,
guint prop_id ,
const GValue * value ,
GParamSpec * pspec )
{
switch ( prop_id )
{
default :
G_OBJECT_WARN_INVALID_PROPERTY_ID ( object , prop_id , pspec ) ;
2008-12-18 07:41:56 -05:00
break ;
2008-12-05 20:03:27 -05:00
}
}
2008-11-07 04:57:00 -05:00
static void
meta_window_class_init ( MetaWindowClass * klass )
{
GObjectClass * object_class = G_OBJECT_CLASS ( klass ) ;
object_class - > finalize = meta_window_finalize ;
2008-12-18 07:41:56 -05:00
2008-12-05 20:03:27 -05:00
object_class - > get_property = meta_window_get_property ;
object_class - > set_property = meta_window_set_property ;
2008-12-18 07:41:56 -05:00
2014-03-18 11:20:22 -04:00
klass - > get_default_skip_hints = meta_window_real_get_default_skip_hints ;
2008-12-05 20:03:27 -05:00
g_object_class_install_property ( object_class ,
PROP_TITLE ,
g_param_spec_string ( " title " ,
" Title " ,
" The title of the window " ,
NULL ,
2008-12-18 07:41:56 -05:00
G_PARAM_READABLE ) ) ;
2008-12-05 20:03:27 -05:00
g_object_class_install_property ( object_class ,
PROP_ICON ,
g_param_spec_object ( " icon " ,
" Icon " ,
2012-12-03 14:58:13 -05:00
" 96 pixel sized icon " ,
2008-12-05 20:03:27 -05:00
GDK_TYPE_PIXBUF ,
G_PARAM_READABLE ) ) ;
g_object_class_install_property ( object_class ,
PROP_MINI_ICON ,
g_param_spec_object ( " mini-icon " ,
" Mini Icon " ,
" 16 pixel sized icon " ,
GDK_TYPE_PIXBUF ,
2008-12-18 07:41:56 -05:00
G_PARAM_READABLE ) ) ;
2009-01-20 07:52:06 -05:00
2009-02-11 07:16:41 -05:00
g_object_class_install_property ( object_class ,
PROP_DECORATED ,
g_param_spec_boolean ( " decorated " ,
" Decorated " ,
2010-02-12 15:26:52 -05:00
" Whether window is decorated " ,
2009-02-11 07:16:41 -05:00
TRUE ,
G_PARAM_READABLE ) ) ;
2009-03-17 08:27:13 -04:00
g_object_class_install_property ( object_class ,
PROP_FULLSCREEN ,
g_param_spec_boolean ( " fullscreen " ,
" Fullscreen " ,
2010-02-12 15:26:52 -05:00
" Whether window is fullscreened " ,
FALSE ,
G_PARAM_READABLE ) ) ;
g_object_class_install_property ( object_class ,
PROP_MAXIMIZED_HORIZONTALLY ,
g_param_spec_boolean ( " maximized-horizontally " ,
" Maximized horizontally " ,
" Whether window is maximized horizontally " ,
FALSE ,
G_PARAM_READABLE ) ) ;
g_object_class_install_property ( object_class ,
PROP_MAXIMIZED_VERTICALLY ,
g_param_spec_boolean ( " maximized-vertically " ,
" Maximizing vertically " ,
" Whether window is maximized vertically " ,
2009-03-17 08:27:13 -04:00
FALSE ,
G_PARAM_READABLE ) ) ;
2011-05-31 13:52:04 -04:00
g_object_class_install_property ( object_class ,
PROP_MINIMIZED ,
g_param_spec_boolean ( " minimized " ,
" Minimizing " ,
" Whether window is minimized " ,
FALSE ,
G_PARAM_READABLE ) ) ;
2009-03-17 08:27:13 -04:00
2009-08-04 03:06:30 -04:00
g_object_class_install_property ( object_class ,
PROP_WINDOW_TYPE ,
g_param_spec_enum ( " window-type " ,
" Window Type " ,
" The type of the window " ,
2012-02-23 14:39:58 -05:00
META_TYPE_WINDOW_TYPE ,
2009-08-04 03:06:30 -04:00
META_WINDOW_NORMAL ,
G_PARAM_READABLE ) ) ;
2009-08-31 12:06:53 -04:00
g_object_class_install_property ( object_class ,
PROP_USER_TIME ,
g_param_spec_uint ( " user-time " ,
" User time " ,
" Timestamp of last user interaction " ,
0 ,
G_MAXUINT ,
0 ,
G_PARAM_READABLE ) ) ;
2009-11-17 05:02:06 -05:00
g_object_class_install_property ( object_class ,
PROP_DEMANDS_ATTENTION ,
g_param_spec_boolean ( " demands-attention " ,
" Demands Attention " ,
" Whether the window has _NET_WM_STATE_DEMANDS_ATTENTION set " ,
FALSE ,
G_PARAM_READABLE ) ) ;
2009-11-20 03:51:19 -05:00
g_object_class_install_property ( object_class ,
PROP_URGENT ,
g_param_spec_boolean ( " urgent " ,
" Urgent " ,
" Whether the urgent flag of WM_HINTS is set " ,
FALSE ,
G_PARAM_READABLE ) ) ;
2014-01-30 16:39:54 -05:00
g_object_class_install_property ( object_class ,
PROP_SKIP_TASKBAR ,
g_param_spec_boolean ( " skip-taskbar " ,
" Skip taskbar " ,
" Whether the skip-taskbar flag of WM_HINTS is set " ,
FALSE ,
G_PARAM_READABLE ) ) ;
2010-01-18 05:12:17 -05:00
g_object_class_install_property ( object_class ,
PROP_MUTTER_HINTS ,
g_param_spec_string ( " mutter-hints " ,
" _MUTTER_HINTS " ,
" Contents of the _MUTTER_HINTS property of this window " ,
NULL ,
G_PARAM_READABLE ) ) ;
2011-03-24 21:36:47 -04:00
g_object_class_install_property ( object_class ,
PROP_APPEARS_FOCUSED ,
g_param_spec_boolean ( " appears-focused " ,
" Appears focused " ,
" Whether the window is drawn as being focused " ,
FALSE ,
G_PARAM_READABLE ) ) ;
2011-07-02 07:08:17 -04:00
g_object_class_install_property ( object_class ,
PROP_RESIZEABLE ,
g_param_spec_boolean ( " resizeable " ,
" Resizeable " ,
" Whether the window can be resized " ,
FALSE ,
G_PARAM_READABLE ) ) ;
g_object_class_install_property ( object_class ,
PROP_ABOVE ,
g_param_spec_boolean ( " above " ,
" Above " ,
" Whether the window is shown as always-on-top " ,
FALSE ,
G_PARAM_READABLE ) ) ;
2011-03-03 04:01:04 -05:00
g_object_class_install_property ( object_class ,
PROP_WM_CLASS ,
g_param_spec_string ( " wm-class " ,
" WM_CLASS " ,
" Contents of the WM_CLASS property of this window " ,
NULL ,
G_PARAM_READABLE ) ) ;
2011-11-25 18:42:41 -05:00
g_object_class_install_property ( object_class ,
2012-01-17 13:27:10 -05:00
PROP_GTK_APPLICATION_ID ,
g_param_spec_string ( " gtk-application-id " ,
" _GTK_APPLICATION_ID " ,
" Contents of the _GTK_APPLICATION_ID property of this window " ,
NULL ,
G_PARAM_READABLE ) ) ;
g_object_class_install_property ( object_class ,
PROP_GTK_UNIQUE_BUS_NAME ,
g_param_spec_string ( " gtk-unique-bus-name " ,
" _GTK_UNIQUE_BUS_NAME " ,
" Contents of the _GTK_UNIQUE_BUS_NAME property of this window " ,
NULL ,
G_PARAM_READABLE ) ) ;
g_object_class_install_property ( object_class ,
PROP_GTK_APPLICATION_OBJECT_PATH ,
g_param_spec_string ( " gtk-application-object-path " ,
" _GTK_APPLICATION_OBJECT_PATH " ,
" Contents of the _GTK_APPLICATION_OBJECT_PATH property of this window " ,
2011-11-25 18:42:41 -05:00
NULL ,
G_PARAM_READABLE ) ) ;
2011-12-05 12:52:35 -05:00
g_object_class_install_property ( object_class ,
2012-01-17 13:27:10 -05:00
PROP_GTK_WINDOW_OBJECT_PATH ,
g_param_spec_string ( " gtk-window-object-path " ,
" _GTK_WINDOW_OBJECT_PATH " ,
" Contents of the _GTK_WINDOW_OBJECT_PATH property of this window " ,
2011-12-05 12:52:35 -05:00
NULL ,
G_PARAM_READABLE ) ) ;
g_object_class_install_property ( object_class ,
2012-01-17 13:27:10 -05:00
PROP_GTK_APP_MENU_OBJECT_PATH ,
g_param_spec_string ( " gtk-app-menu-object-path " ,
" _GTK_APP_MENU_OBJECT_PATH " ,
" Contents of the _GTK_APP_MENU_OBJECT_PATH property of this window " ,
NULL ,
G_PARAM_READABLE ) ) ;
g_object_class_install_property ( object_class ,
PROP_GTK_MENUBAR_OBJECT_PATH ,
g_param_spec_string ( " gtk-menubar-object-path " ,
" _GTK_MENUBAR_OBJECT_PATH " ,
" Contents of the _GTK_MENUBAR_OBJECT_PATH property of this window " ,
2011-12-05 12:52:35 -05:00
NULL ,
G_PARAM_READABLE ) ) ;
2009-01-20 07:52:06 -05:00
window_signals [ WORKSPACE_CHANGED ] =
g_signal_new ( " workspace-changed " ,
G_TYPE_FROM_CLASS ( object_class ) ,
G_SIGNAL_RUN_LAST ,
2014-03-18 11:21:02 -04:00
0 ,
2011-10-18 18:06:14 -04:00
NULL , NULL , NULL ,
2009-01-20 07:52:06 -05:00
G_TYPE_NONE , 1 ,
G_TYPE_INT ) ;
2009-02-02 06:00:13 -05:00
window_signals [ FOCUS ] =
g_signal_new ( " focus " ,
G_TYPE_FROM_CLASS ( object_class ) ,
G_SIGNAL_RUN_LAST ,
2014-03-18 11:21:02 -04:00
0 ,
2011-10-18 18:06:14 -04:00
NULL , NULL , NULL ,
2009-02-02 06:00:13 -05:00
G_TYPE_NONE , 0 ) ;
2009-02-02 09:08:46 -05:00
window_signals [ RAISED ] =
g_signal_new ( " raised " ,
G_TYPE_FROM_CLASS ( object_class ) ,
G_SIGNAL_RUN_LAST ,
2014-03-18 11:21:02 -04:00
0 ,
2011-10-18 18:06:14 -04:00
NULL , NULL , NULL ,
2009-02-02 09:08:46 -05:00
G_TYPE_NONE , 0 ) ;
2009-10-13 10:34:34 -04:00
window_signals [ UNMANAGED ] =
g_signal_new ( " unmanaged " ,
G_TYPE_FROM_CLASS ( object_class ) ,
G_SIGNAL_RUN_LAST ,
2014-03-18 11:21:02 -04:00
0 ,
2011-10-18 18:06:14 -04:00
NULL , NULL , NULL ,
2009-10-13 10:34:34 -04:00
G_TYPE_NONE , 0 ) ;
2013-12-06 15:39:24 -05:00
/**
* MetaWindow : : position - changed :
* @ window : a # MetaWindow
*
* This is emitted when the position of a window might
* have changed . Specifically , this is emitted when the
* position of the toplevel window has changed , or when
* the position of the client window has changed .
*/
window_signals [ POSITION_CHANGED ] =
g_signal_new ( " position-changed " ,
G_TYPE_FROM_CLASS ( object_class ) ,
G_SIGNAL_RUN_LAST ,
0 ,
NULL , NULL , NULL ,
G_TYPE_NONE , 0 ) ;
/**
* MetaWindow : : size - changed :
* @ window : a # MetaWindow
*
* This is emitted when the position of a window might
* have changed . Specifically , this is emitted when the
* size of the toplevel window has changed , or when the
* size of the client window has changed .
*/
window_signals [ SIZE_CHANGED ] =
g_signal_new ( " size-changed " ,
G_TYPE_FROM_CLASS ( object_class ) ,
G_SIGNAL_RUN_LAST ,
0 ,
NULL , NULL , NULL ,
G_TYPE_NONE , 0 ) ;
2008-11-07 04:57:00 -05:00
}
static void
meta_window_init ( MetaWindow * self )
{
2011-11-25 19:52:24 -05:00
meta_prefs_add_listener ( prefs_changed_callback , self ) ;
2008-11-07 04:57:00 -05:00
}
2003-05-16 17:59:08 -04:00
static gboolean
is_desktop_or_dock_foreach ( MetaWindow * window ,
void * data )
{
gboolean * result = data ;
* result =
window - > type = = META_WINDOW_DESKTOP | |
window - > type = = META_WINDOW_DOCK ;
if ( * result )
return FALSE ; /* stop as soon as we find one */
else
return TRUE ;
}
/* window is the window that's newly mapped provoking
* the possible change
*/
static void
maybe_leave_show_desktop_mode ( MetaWindow * window )
{
gboolean is_desktop_or_dock ;
Make the "showing desktop" mode be per-workspace instead of per-screen.
2004-10-16 Elijah Newren <newren@math.utah.edu>
Make the "showing desktop" mode be per-workspace instead of
per-screen. (fixes #142198)
* src/keybindings.c (handle_toggle_desktop): access
showing_desktop through the active workspace
* src/screen.c (meta_screen_new): remove initialization of
screen->showing_desktop,
(meta_screen_update_showing_desktop_hint): rename and make not
static and access showing_desktop through the active workspace,
(queue_windows_showing): replace meta_display_list_windows() with
screen->active_workspace->windows,
(meta_screen_minimize_all_on_active_workspace_except): renamed
from meta_screen_minimize_all_except since it now only works on
the active workspace, (meta_screen_show_desktop,
meta_screen_unshow_desktop): access showing_desktop through the
active workspace
* src/screen.h (struct _MetaScreen): remove showing_desktop field,
(meta_screen_minimize_all_on_active_workspace_except): rename from
meta_screen_minimize_all_except,
(meta_screen_update)_showing_desktop_hint): export this function too
* src/window.c (maybe_leave_show_desktop_mode): access
showing_desktop through the active workspace and use new name for
meta_screen_minimize_all_on_active_workspace_except,
(window_should_be_showing): access showing_desktop through the
active workspace
* src/workspace.c (meta_workspace_new): initialize
workspace->showing_desktop, (meta_workspace_activate_with_focus):
add note that old can be NULL, update showing_desktop_hint if
different on this workspace than the previous one
* src/workspace.h (struct _MetaWorkspace): add showing_desktop
field
2004-10-17 00:28:29 -04:00
if ( ! window - > screen - > active_workspace - > showing_desktop )
2003-05-16 17:59:08 -04:00
return ;
/* If the window is a transient for the dock or desktop, don't
* leave show desktop mode when the window opens . That ' s
* so you can e . g . hide all windows , manipulate a file on
* the desktop via a dialog , then unshow windows again .
*/
is_desktop_or_dock = FALSE ;
2003-05-18 13:16:03 -04:00
is_desktop_or_dock_foreach ( window ,
& is_desktop_or_dock ) ;
2003-05-16 17:59:08 -04:00
meta_window_foreach_ancestor ( window , is_desktop_or_dock_foreach ,
& is_desktop_or_dock ) ;
if ( ! is_desktop_or_dock )
{
Make the "showing desktop" mode be per-workspace instead of per-screen.
2004-10-16 Elijah Newren <newren@math.utah.edu>
Make the "showing desktop" mode be per-workspace instead of
per-screen. (fixes #142198)
* src/keybindings.c (handle_toggle_desktop): access
showing_desktop through the active workspace
* src/screen.c (meta_screen_new): remove initialization of
screen->showing_desktop,
(meta_screen_update_showing_desktop_hint): rename and make not
static and access showing_desktop through the active workspace,
(queue_windows_showing): replace meta_display_list_windows() with
screen->active_workspace->windows,
(meta_screen_minimize_all_on_active_workspace_except): renamed
from meta_screen_minimize_all_except since it now only works on
the active workspace, (meta_screen_show_desktop,
meta_screen_unshow_desktop): access showing_desktop through the
active workspace
* src/screen.h (struct _MetaScreen): remove showing_desktop field,
(meta_screen_minimize_all_on_active_workspace_except): rename from
meta_screen_minimize_all_except,
(meta_screen_update)_showing_desktop_hint): export this function too
* src/window.c (maybe_leave_show_desktop_mode): access
showing_desktop through the active workspace and use new name for
meta_screen_minimize_all_on_active_workspace_except,
(window_should_be_showing): access showing_desktop through the
active workspace
* src/workspace.c (meta_workspace_new): initialize
workspace->showing_desktop, (meta_workspace_activate_with_focus):
add note that old can be NULL, update showing_desktop_hint if
different on this workspace than the previous one
* src/workspace.h (struct _MetaWorkspace): add showing_desktop
field
2004-10-17 00:28:29 -04:00
meta_screen_minimize_all_on_active_workspace_except ( window - > screen ,
window ) ;
2008-12-18 07:41:56 -05:00
meta_screen_unshow_desktop ( window - > screen ) ;
2003-05-16 17:59:08 -04:00
}
}
window: make determination of attached dialog windows more consistent
Different bits of code were using slightly different checks to test
whether a window was an attached dialog. Add a new
meta_window_is_attached_dialog(), and use that everywhere.
Also, freeze the is-attached status when the window is first shown,
rather than recomputing it each time the caller asks, since this could
cause problems if a window changes its type after it has already been
attached, etc. However, if an attached window's parent is destroyed,
or an attached window changes its transient-for, then fix things up by
destroying the old MetaWindow and creating a new one (causing
compositor unmap and map events to be fired off, allowing the display
of the window to be fixed up).
Remove some code in display.c that tried to fix existing windows if
the gconf setting changed, but which didn't actually do anything (at
least under gnome-shell). However, if 654643 was fixed then the new
behavior with this patch would be that changing the gconf setting
would affect new dialogs, but not existing ones.
https://bugzilla.gnome.org/show_bug.cgi?id=646761
2011-04-05 10:14:52 -04:00
gboolean
meta_window_should_attach_to_parent ( MetaWindow * window )
{
MetaWindow * parent ;
if ( ! meta_prefs_get_attach_modal_dialogs ( ) | |
window - > type ! = META_WINDOW_MODAL_DIALOG )
return FALSE ;
parent = meta_window_get_transient_for ( window ) ;
if ( ! parent )
return FALSE ;
2011-07-22 12:34:10 -04:00
switch ( parent - > type )
{
case META_WINDOW_NORMAL :
case META_WINDOW_DIALOG :
case META_WINDOW_MODAL_DIALOG :
return TRUE ;
default :
return FALSE ;
}
window: make determination of attached dialog windows more consistent
Different bits of code were using slightly different checks to test
whether a window was an attached dialog. Add a new
meta_window_is_attached_dialog(), and use that everywhere.
Also, freeze the is-attached status when the window is first shown,
rather than recomputing it each time the caller asks, since this could
cause problems if a window changes its type after it has already been
attached, etc. However, if an attached window's parent is destroyed,
or an attached window changes its transient-for, then fix things up by
destroying the old MetaWindow and creating a new one (causing
compositor unmap and map events to be fired off, allowing the display
of the window to be fixed up).
Remove some code in display.c that tried to fix existing windows if
the gconf setting changed, but which didn't actually do anything (at
least under gnome-shell). However, if 654643 was fixed then the new
behavior with this patch would be that changing the gconf setting
would affect new dialogs, but not existing ones.
https://bugzilla.gnome.org/show_bug.cgi?id=646761
2011-04-05 10:14:52 -04:00
}
Always map the client and frame windows
Traditionally, WMs unmap windows when minimizing them, and map them
when restoring them or wanting to show them for other reasons, like
upon creation.
However, as metacity morphed into mutter, we optionally chose to keep
windows mapped for the lifetime of the window under the user option
"live-window-previews", which makes the code keep windows mapped so it
can show window preview for minimized windows in other places, like
Alt-Tab and Expose.
I removed this preference two years ago mechanically, by removing all
the if statements, but never went through and cleaned up the code so
that windows are simply mapped for the lifetime of the window -- the
"architecture" of the old code that maps and unmaps on show/hide was
still there.
Remove this now.
The one case we still need to be careful of is shaded windows, in which
we do still unmap the client window. In the future, we might want to
show previews of shaded windows in the overview and Alt-Tab. In that
we'd also keep shaded windows mapped, and could remove all unmap logic,
but we'd need a more complex method of showing the shaded titlebar, such
as using a different actor.
At the same time, simplify the compositor interface by removing
meta_compositor_window_[un]mapped API, and instead adding/removing the
window on-demand.
https://bugzilla.gnome.org/show_bug.cgi?id=720631
2014-02-01 19:06:51 -05:00
static gboolean
client_window_should_be_mapped ( MetaWindow * window )
{
return ! window - > shaded ;
}
static void
sync_client_window_mapped ( MetaWindow * window )
{
gboolean should_be_mapped = client_window_should_be_mapped ( window ) ;
if ( window - > mapped = = should_be_mapped )
return ;
window - > mapped = should_be_mapped ;
meta_error_trap_push ( window - > display ) ;
if ( should_be_mapped )
{
XMapWindow ( window - > display - > xdisplay , window - > xwindow ) ;
}
else
{
XUnmapWindow ( window - > display - > xdisplay , window - > xwindow ) ;
window - > unmaps_pending + + ;
}
meta_error_trap_pop ( window - > display ) ;
}
2014-03-18 21:19:51 -04:00
static void
meta_window_update_desc ( MetaWindow * window )
{
g_clear_pointer ( & window - > desc , g_free ) ;
if ( window - > title )
window - > desc = g_strdup_printf ( " 0x%lx (%.10s) " , window - > xwindow , window - > title ) ;
else
window - > desc = g_strdup_printf ( " 0x%lx " , window - > xwindow ) ;
}
2014-02-01 19:19:22 -05:00
MetaWindow *
_meta_window_shared_new ( MetaDisplay * display ,
MetaScreen * screen ,
MetaWindowClientType client_type ,
MetaWaylandSurface * surface ,
Window xwindow ,
gulong existing_wm_state ,
MetaCompEffect effect ,
XWindowAttributes * attrs )
2003-11-20 21:32:05 -05:00
{
MetaWindow * window ;
2001-06-12 20:56:08 -04:00
MetaWorkspace * space ;
Patch from Jens Granseuer to fix build with gcc 2.95. #322622.
2006-01-02 Elijah Newren <newren@gmail.com>
Patch from Jens Granseuer to fix build with gcc 2.95. #322622.
* src/boxes.c (meta_rectangle_region_to_string,
meta_rectangle_edge_list_to_string, fix_up_edges):
* src/constraints.c (meta_window_constrain, setup_constraint_info,
place_window_if_needed, constrain_maximization,
constrain_fullscreen, constrain_size_increments,
constrain_size_limits, constrain_aspect_ratio,
do_screen_and_xinerama_relative_constrai,
constrain_to_single_xinerama, constrain_fully_onscreen,
constrain_partially_onscreen):
* src/edge-resistance.c (find_nearest_position,
apply_edge_resistance, apply_edge_resistance_to_each_side):
* src/testboxes.c (test_clamping_to_region,
test_clipping_to_region, test_shoving_into_region):
* src/window.c (meta_window_new_with_attrs,
meta_window_apply_session_info, meta_window_resize,
meta_window_resize_with_gravity, meta_window_configure_request):
Remove C99 style variable initiailization
2006-01-02 13:37:46 -05:00
MetaMoveResizeFlags flags ;
2003-11-20 21:32:05 -05:00
g_assert ( attrs ! = NULL ) ;
2008-12-18 07:41:56 -05:00
2013-11-19 14:42:50 -05:00
meta_verbose ( " attrs->map_state = %d (%s) \n " ,
2003-11-20 21:32:05 -05:00
attrs - > map_state ,
( attrs - > map_state = = IsUnmapped ) ?
2001-08-19 02:23:59 -04:00
" IsUnmapped " :
2003-11-20 21:32:05 -05:00
( attrs - > map_state = = IsViewable ) ?
2001-08-19 02:23:59 -04:00
" IsViewable " :
2003-11-20 21:32:05 -05:00
( attrs - > map_state = = IsUnviewable ) ?
2001-08-19 02:23:59 -04:00
" IsUnviewable " :
" (unknown) " ) ;
2008-12-18 07:41:56 -05:00
2014-03-18 11:20:22 -04:00
if ( client_type = = META_WINDOW_CLIENT_TYPE_X11 )
window = g_object_new ( META_TYPE_WINDOW_X11 , NULL ) ;
else
2014-03-19 09:26:17 -04:00
window = g_object_new ( META_TYPE_WINDOW_WAYLAND , NULL ) ;
2001-05-30 11:36:31 -04:00
deleted and moved into window-props.c (meta_window_new_with_attrs): added
2006-03-25 Thomas Thurman <thomas@thurman.org.uk>
* src/window.c, src/window.h (update_net_wm_state,
update_mwm_hints, update_wm_class, update_transient_for):
deleted and moved into window-props.c
(meta_window_new_with_attrs): added constructing field
and four new initial properties (as above)
(meta_window_recalc_features,
meta_window_recalc_window_type): new functions
* src/window-props.c (init_net_wm_state, reload_net_wm_state
init_mwm_hints, reload_mwm_hints, init_wm_class,
reload_mwm_class, init_transient_for, reload_transient_for):
new functions, moved in from window.c
(meta_display_init_window_prop_hooks): initialise new properties
2006-03-25 17:57:22 -05:00
window - > constructing = TRUE ;
2008-12-18 07:41:56 -05:00
2002-04-13 00:58:45 -04:00
window - > dialog_pid = - 1 ;
2008-12-18 07:41:56 -05:00
2012-01-07 17:21:32 -05:00
window - > client_type = client_type ;
window - > surface = surface ;
2001-05-30 11:36:31 -04:00
window - > xwindow = xwindow ;
2014-02-01 19:16:25 -05:00
window - > surface_mapped = FALSE ;
2008-12-18 07:41:56 -05:00
2001-05-31 02:42:58 -04:00
/* this is in window->screen->display, but that's too annoying to
* type
*/
window - > display = display ;
Partially resolve the conflicting requirements of windows on multiple
2004-12-22 Elijah Newren <newren@gmail.com>
Partially resolve the conflicting requirements of windows on
multiple workspaces and hidden being a global quantity for windows
(fixes bug 156182; the remainder of the work is bug 87531 and is a
libwnck issue)
* src/display.c (event_callback):
* src/window.c (meta_window_visible_on_workspace, meta_window_unstick):
* src/workspace.c (meta_workspace_add_window,
meta_workspace_contains_window,
meta_workspace_queue_calc_showing):
* src/workspace.h:
Remove meta_workspace_contains_window, replace with simple
comparison utilizing window->workspace
* src/place.c (meta_window_place):
* src/window.c (meta_window_shares_some_workspace):
* src/window.h:
Remove meta_window_shares_some_workspace, replace with a simple
comparison utilizing window->workspace
* src/session.c (save_state),
* src/window.c (meta_window_new_with_attrs,
meta_window_apply_session_info, meta_window_free,
window_showing_on_its_workspace,
meta_window_change_workspace_without_transients,
meta_window_unstick, meta_window_set_current_workspace_hint,
meta_window_get_workspaces):
* src/window.h:
* src/workspace.c (meta_workspace_free, meta_workspace_add_window,
meta_workspace_remove_window):
Only one workspace now
2004-12-22 18:52:52 -05:00
window - > workspace = NULL ;
2002-12-09 22:23:04 -05:00
# ifdef HAVE_XSYNC
Fix bug 143333, support for update counter spec, and 109362, schedule
Sat Jun 19 02:21:08 2004 Soeren Sandmann <sandmann@daimi.au.dk>
Fix bug 143333, support for update counter spec, and 109362,
schedule compensation events when events are ignored.
* src/display.c (meta_display_open): Add _NET_WM_SYNC_REQUEST and
_NET_WM_SYNC_REQUEST_COUNTER atoms. Remove the old
METACITY_SYNC_COUNTER stuff.
(meta_display_begin_op): Setup the sync counter
* src/xprops.c, src/xprops.h, src/window-props.c, src/display.h:
Add new atoms.
* src/window.c (send_sync_request): new function.
(meta_window_move_resize_internal): send a sync request before
resizing.
(check_move_resize_frequence): Rework logic to also check the SYNC
case. If an event is ignored return the remaining time.
(update_resize_timeout): Timeout that gets called when a
compensation event is scheduled.
(uddate_resize): schedule compensation events when an event is
ignored.
(meta_window_handle_mouse_grap_op_event): When an alarm is
received and sync was turned off, turn it back on.
* src/window.h (struct MetaWindow) Add some variables
2004-06-18 20:45:24 -04:00
window - > sync_request_counter = None ;
window - > sync_request_serial = 0 ;
2013-03-13 15:31:11 -04:00
window - > sync_request_timeout_id = 0 ;
2011-06-08 11:26:03 -04:00
window - > sync_request_alarm = None ;
2002-12-09 22:23:04 -05:00
# endif
2002-05-27 15:26:56 -04:00
2008-11-23 14:28:40 -05:00
window - > screen = screen ;
2001-06-04 00:58:22 -04:00
2014-03-18 21:19:51 -04:00
meta_window_update_desc ( window ) ;
2003-06-07 13:15:33 -04:00
2008-11-23 14:28:40 -05:00
window - > override_redirect = attrs - > override_redirect ;
2001-06-10 23:24:20 -04:00
/* avoid tons of stack updates */
meta_stack_freeze ( window - > screen - > stack ) ;
2003-01-05 02:51:02 -05:00
2003-11-20 21:32:05 -05:00
window - > rect . x = attrs - > x ;
window - > rect . y = attrs - > y ;
window - > rect . width = attrs - > width ;
window - > rect . height = attrs - > height ;
2001-06-02 21:33:27 -04:00
2001-06-04 00:58:22 -04:00
/* And border width, size_hints are the "request" */
2003-11-20 21:32:05 -05:00
window - > border_width = attrs - > border_width ;
window - > size_hints . x = attrs - > x ;
window - > size_hints . y = attrs - > y ;
window - > size_hints . width = attrs - > width ;
window - > size_hints . height = attrs - > height ;
2003-04-26 13:40:32 -04:00
/* initialize the remaining size_hints as if size_hints.flags were zero */
meta_set_normal_hints ( window , NULL ) ;
2001-06-07 01:18:10 -04:00
/* And this is our unmaximized size */
window - > saved_rect = window - > rect ;
2001-08-28 23:37:03 -04:00
window - > user_rect = window - > rect ;
2008-12-18 07:41:56 -05:00
2003-11-20 21:32:05 -05:00
window - > depth = attrs - > depth ;
window - > xvisual = attrs - > visual ;
window - > colormap = attrs - > colormap ;
2008-12-18 07:41:56 -05:00
2001-06-02 00:14:18 -04:00
window - > title = NULL ;
2001-06-23 22:22:10 -04:00
window - > icon = NULL ;
2001-09-15 20:30:45 -04:00
window - > mini_icon = NULL ;
2002-02-09 18:03:52 -05:00
meta_icon_cache_init ( & window - > icon_cache ) ;
window - > wm_hints_pixmap = None ;
window - > wm_hints_mask = None ;
2001-06-02 21:33:27 -04:00
window - > frame = NULL ;
2001-06-04 00:58:22 -04:00
window - > has_focus = FALSE ;
2011-03-24 21:36:47 -04:00
window - > attached_focus_window = NULL ;
2001-06-04 02:17:52 -04:00
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
window - > maximized_horizontally = FALSE ;
window - > maximized_vertically = FALSE ;
window - > maximize_horizontally_after_placement = FALSE ;
window - > maximize_vertically_after_placement = FALSE ;
2006-10-01 17:59:09 -04:00
window - > minimize_after_placement = FALSE ;
2002-02-23 17:58:02 -05:00
window - > fullscreen = FALSE ;
2010-01-20 10:59:07 -05:00
window - > fullscreen_after_placement = FALSE ;
2008-11-16 21:57:20 -05:00
window - > fullscreen_monitors [ 0 ] = - 1 ;
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
window - > require_fully_onscreen = TRUE ;
2009-08-31 09:52:16 -04:00
window - > require_on_single_monitor = TRUE ;
2006-08-07 19:34:55 -04:00
window - > require_titlebar_visible = TRUE ;
2001-06-09 23:17:15 -04:00
window - > on_all_workspaces = FALSE ;
2011-02-28 05:41:23 -05:00
window - > on_all_workspaces_requested = FALSE ;
2010-06-24 14:41:28 -04:00
window - > tile_mode = META_TILE_NONE ;
2011-10-08 11:00:16 -04:00
window - > tile_monitor_number = - 1 ;
2001-06-07 22:17:48 -04:00
window - > shaded = FALSE ;
2001-06-04 02:17:52 -04:00
window - > initially_iconic = FALSE ;
window - > minimized = FALSE ;
2005-10-03 16:02:31 -04:00
window - > tab_unminimized = FALSE ;
2001-06-04 02:17:52 -04:00
window - > iconic = FALSE ;
2003-11-20 21:32:05 -05:00
window - > mapped = attrs - > map_state ! = IsUnmapped ;
Simplify relationship between mapping and visibility
Previously, changes to the visibility of a window could be indicated
by meta_compositor_map_window(), meta_compositor_unminimize_window(),
meta_compositor_set_window_hidden(), etc, with the exact behavior
depending on the 'live_hidden_windows' preference.
Simplify this so that visibility is controlled by:
meta_compositor_show_window()
meta_compositor_hide_window()
With an 'effect' parameter provided to indicate the appropriate
effect (CREATE/UNMINIMIZE/MINIMIZE/DESTROY/NONE.)
The map state of the window is signalled separately by:
meta_compositor_map_window()
meta_compositor_unmap_window()
And is used only to control resource handling.
Other changes:
* The desired effect on show/hide is explicitly stored in
MetaWindow, avoiding the need for the was_minimized flag.
At idle, once we calculate the window state, we pass the
effect to the compositor if it matches the new window
state, and then clear the effect to start over for future
map state changes.
* meta_compositor_switch_workspace() is called before any windows
are hidden or shown, allowing the compositor to avoid hiding
or showing an effect for windows involved in the switch.
http://bugzilla.gnome.org/show_bug.cgi?id=582341
* Handling of post-effect cleanups for MutterWindow are
simplified - instead of trying to do different things based
on the individual needs of different effects, we just wait until
all effects complete and sync the window state to what it
should be.
* On unmap, once we destroy the pixmap, we tell ClutterX11Pixmap
that we've done so, so it can clean up and unbind. (The
unbinding doesn't seem to be working properly because of
ClutterGLXPixmap or video driver issues.)
http://bugzilla.gnome.org/show_bug.cgi?id=587251
2009-06-28 17:10:40 -04:00
window - > hidden = FALSE ;
window - > visible_to_compositor = FALSE ;
window - > pending_compositor_effect = effect ;
2004-12-28 01:01:42 -05:00
/* if already mapped, no need to worry about focus-on-first-time-showing */
window - > showing_for_first_time = ! window - > mapped ;
2009-06-15 18:25:07 -04:00
/* if already mapped we don't want to do the placement thing;
* override - redirect windows are placed by the app */
window - > placed = ( ( window - > mapped & & ! window - > hidden ) | | window - > override_redirect ) ;
2001-06-24 04:09:10 -04:00
if ( window - > placed )
2002-02-07 22:34:26 -05:00
meta_topic ( META_DEBUG_PLACEMENT ,
" Not placing window 0x%lx since it's already mapped \n " ,
xwindow ) ;
2010-01-20 10:59:07 -05:00
window - > force_save_user_rect = TRUE ;
2005-02-11 14:20:44 -05:00
window - > denied_focus_and_not_transient = FALSE ;
2001-06-21 23:20:21 -04:00
window - > unmanaging = FALSE ;
Refactor thrice-duplicated queue code in window.c. Closes #376760.
2007-06-10 Thomas Thurman <thomas@thurman.org.uk>
Refactor thrice-duplicated queue code in window.c. Closes #376760.
* src/window.c (meta_window_queue, meta_window_unqueue):
New functions.
* src/window.[ch] (meta_window_unqueue_*, meta_window_queue_*):
Removed functions.
* src/window.c (meta_window_new_with_attrs, meta_window_free,
meta_window_flush_calc_showing, queue_calc_showing_func,
meta_window_minimize, meta_window_unminimize, meta_window_maximize,
meta_window_make_fullscreen, meta_window_shade,
meta_window_unshade, meta_window_move_resize_internal,
window_stick_impl, window_unstick_impl,
meta_window_client_message, process_property_notify): Modified to
use new queueing functions.
* src/window.c (idle_move_resize, idle_update_icon,
idle_calc_showing): update to receive queue number from pointer.
* src/window.h (MetaQueueType): new enum.
* src/window.h (MetaWindow): *_queued replaced with is_in_queue
bitfield.
* src/core.c (meta_core_queue_frame_resize):
* src/display.c (event_callback,
meta_display_queue_retheme_all_windows): Using new queueing functions.
* src/frame.c (meta_window_destroy_frame): Using new queueing functions.
* src/screen.c (queue_resize, meta_screen_resize_func,
queue_windows_showing): Using new queueing functions.
* src/window-props.c (reload_mwm_hints, reload_wm_hints,
reload_transient_for): Using new queueing functions.
* src/workspace.c (meta_workspace_add_window,
meta_workspace_remove_window, meta_workspace_queue_calc_showing,
meta_workspace_invalidate_work_area): Using new queueing functions.
svn path=/trunk/; revision=3236
2007-06-10 21:15:33 -04:00
window - > is_in_queues = 0 ;
2001-06-23 01:49:35 -04:00
window - > keys_grabbed = FALSE ;
window - > grab_on_frame = FALSE ;
2001-07-11 02:22:00 -04:00
window - > all_keys_grabbed = FALSE ;
2001-06-23 14:30:27 -04:00
window - > withdrawn = FALSE ;
2001-06-24 02:47:54 -04:00
window - > initial_workspace_set = FALSE ;
Add support for _NET_WM_USER_TIME
2004-06-17 Elijah Newren <newren@math.utah.edu>
Add support for _NET_WM_USER_TIME
* src/display.c:
(meta_display_open): Add _NET_WM_USER_TIME to atom_names[],
(event_callback): Manually set _NET_WM_USER_TIME upon KeyPress
(doesn't work since keyboard isn't grabbed) and ButtonPress (does
work), this is just a fallback for applications that don't update
this themselves.
* src/display.h: (struct _MetaDisplay): Add atom_net_wm_user_time field
* src/screen.c: (meta_screen_apply_startup_properties): Check for
TIMESTAMP provided from startup sequence as well.
* src/stack.c:
s/meta_window_set_stack_position/meta_window_set_stack_position_no_sync/,
(meta_window_set_stack_position): New function which calls the
meta_window_set_stack_position_no_sync function followed immediately
by calling meta_stack_sync_to_server.
* src/window-props.c:
(init_net_wm_user_time), (reload_net_wm_user_time): new functions,
(reload_wm_hints): also load atom_net_wm_user_time
* src/window.c:
new XSERVER_TIME_IS_LATER macro (accounts for timestamp wraparound),
(meta_window_new_with_attrs): add timestamp attributes,
(window_takes_focus_on_map): use TIMESTAMP from startup
notification and _NET_WM_USER_TIME to decide whether to focus new
windows,
(meta_window_show): if app doesn't take focus on map, place it
just below the focused window in the stack
(process_property_notify): check for changes to _NET_WM_USRE_TIME,
(meta_window_stack_just_below): new function
* src/window.h:
(_MetaWindow struct): new fields for initial_timestamp,
initial_timestamp_set, net_wm_user_time_set, and net_wm_user_time,
(meta_window_stack_just_below): new function
2004-06-24 11:47:05 -04:00
window - > initial_timestamp_set = FALSE ;
window - > net_wm_user_time_set = FALSE ;
2007-03-31 01:19:41 -04:00
window - > user_time_window = None ;
2009-11-09 13:47:16 -05:00
window - > take_focus = FALSE ;
window - > delete_window = FALSE ;
2014-02-18 20:28:23 -05:00
window - > can_ping = FALSE ;
2009-11-09 13:47:16 -05:00
window - > input = TRUE ;
2001-06-24 04:09:10 -04:00
window - > calc_placement = FALSE ;
2003-06-12 01:55:06 -04:00
window - > shaken_loose = FALSE ;
2003-10-13 16:15:40 -04:00
window - > have_focus_click_grab = FALSE ;
Fix bug 143333, support for update counter spec, and 109362, schedule
Sat Jun 19 02:21:08 2004 Soeren Sandmann <sandmann@daimi.au.dk>
Fix bug 143333, support for update counter spec, and 109362,
schedule compensation events when events are ignored.
* src/display.c (meta_display_open): Add _NET_WM_SYNC_REQUEST and
_NET_WM_SYNC_REQUEST_COUNTER atoms. Remove the old
METACITY_SYNC_COUNTER stuff.
(meta_display_begin_op): Setup the sync counter
* src/xprops.c, src/xprops.h, src/window-props.c, src/display.h:
Add new atoms.
* src/window.c (send_sync_request): new function.
(meta_window_move_resize_internal): send a sync request before
resizing.
(check_move_resize_frequence): Rework logic to also check the SYNC
case. If an event is ignored return the remaining time.
(update_resize_timeout): Timeout that gets called when a
compensation event is scheduled.
(uddate_resize): schedule compensation events when an event is
ignored.
(meta_window_handle_mouse_grap_op_event): When an alarm is
received and sync was turned off, turn it back on.
* src/window.h (struct MetaWindow) Add some variables
2004-06-18 20:45:24 -04:00
window - > disable_sync = FALSE ;
2008-12-18 07:41:56 -05:00
2001-06-10 23:24:20 -04:00
window - > unmaps_pending = 0 ;
2008-12-18 07:41:56 -05:00
2001-06-23 14:30:27 -04:00
window - > mwm_decorated = TRUE ;
2002-06-21 23:23:02 -04:00
window - > mwm_border_only = FALSE ;
2001-06-23 14:30:27 -04:00
window - > mwm_has_close_func = TRUE ;
window - > mwm_has_minimize_func = TRUE ;
window - > mwm_has_maximize_func = TRUE ;
window - > mwm_has_move_func = TRUE ;
window - > mwm_has_resize_func = TRUE ;
2008-12-18 07:41:56 -05:00
2012-01-07 17:21:32 -05:00
if ( client_type = = META_WINDOW_CLIENT_TYPE_X11 )
window - > decorated = TRUE ;
else
window - > decorated = FALSE ;
2001-06-08 02:39:38 -04:00
window - > has_close_func = TRUE ;
window - > has_minimize_func = TRUE ;
window - > has_maximize_func = TRUE ;
2001-06-23 14:30:27 -04:00
window - > has_move_func = TRUE ;
window - > has_resize_func = TRUE ;
2001-06-23 02:54:28 -04:00
window - > has_shade_func = TRUE ;
2002-02-23 17:58:02 -05:00
window - > has_fullscreen_func = TRUE ;
2003-05-16 17:59:08 -04:00
window - > always_sticky = FALSE ;
2008-12-18 07:41:56 -05:00
2002-07-06 15:05:37 -04:00
window - > skip_taskbar = FALSE ;
window - > skip_pager = FALSE ;
2002-10-01 20:47:29 -04:00
window - > wm_state_above = FALSE ;
window - > wm_state_below = FALSE ;
Fix some support for EWMH hints, and fix USER_TIME support to include the
2004-07-31 Rob Adams <readams@readams.net>
Fix some support for EWMH hints, and fix USER_TIME support to
include the DEMANDS_ATTENTION hint. Also includes some code for
implementing _NET_RESTACK_WINDOW and _NET_MOVERESIZE_WINDOW, but
this is disabled pending feature thaw.
* COMPLIANCE: update with new information
* src/display.c (meta_display_open): add new hints to list
* src/display.h (_MetaDisplay): Add new atoms to struct
* src/screen.c (set_supported_hint): update the list of support
hints.
(set_desktop_viewport_hint): new function sets the viewport hint
to (0,0) as required by the spec for WMs with no viewport support.
(set_desktop_geometry_hint): new function to set the desktop size
hint to the size of the display, since we don't implement large
desktop support, as required by the spec.
(meta_screen_resize): update the geometry hint on screen resize
* src/window.c (meta_window_new_with_attrs): Initialize
demands_attention state
(set_net_wm_state): Set demands_attention hint in the window state
(meta_window_show): If we don't pop up a window because of
USER_TIME, set DEMANDS_ATTENTION on the window.
(meta_window_focus): When a window receives focus, remove
DEMANDS_ATTENTION hint
(meta_window_client_message): Allow other apps to set
DEMANDS_ATTENTION on a window. Also, if the _NET_ACTIVE_WINDOW
hint includes a timestamp, use it.
(update_net_wm_state): Read DEMANDS_ATTENTION state also
* src/window.h (_MetaWindow): add wm_state_demands_attention bit.
2004-07-31 15:56:10 -04:00
window - > wm_state_demands_attention = FALSE ;
2008-12-18 07:41:56 -05:00
2001-06-09 17:58:30 -04:00
window - > res_class = NULL ;
window - > res_name = NULL ;
window - > role = NULL ;
window - > sm_client_id = NULL ;
2002-04-13 00:58:45 -04:00
window - > wm_client_machine = NULL ;
2013-02-20 15:17:21 -05:00
window - > is_remote = FALSE ;
2002-11-30 22:58:04 -05:00
window - > startup_id = NULL ;
2008-12-18 07:41:56 -05:00
2002-04-13 00:58:45 -04:00
window - > net_wm_pid = - 1 ;
2008-12-18 07:41:56 -05:00
2001-06-09 17:58:30 -04:00
window - > xtransient_for = None ;
window - > xclient_leader = None ;
2002-09-27 15:42:59 -04:00
window - > transient_parent_is_root_window = FALSE ;
2008-12-18 07:41:56 -05:00
2001-06-09 23:17:15 -04:00
window - > type = META_WINDOW_NORMAL ;
2001-06-10 03:27:11 -04:00
2003-06-25 23:09:38 -04:00
window - > struts = NULL ;
2002-08-12 17:32:13 -04:00
2002-09-28 12:33:39 -04:00
window - > layer = META_LAYER_LAST ; /* invalid value */
window - > stack_position = - 1 ;
2001-06-24 02:47:54 -04:00
window - > initial_workspace = 0 ; /* not used */
Add support for _NET_WM_USER_TIME
2004-06-17 Elijah Newren <newren@math.utah.edu>
Add support for _NET_WM_USER_TIME
* src/display.c:
(meta_display_open): Add _NET_WM_USER_TIME to atom_names[],
(event_callback): Manually set _NET_WM_USER_TIME upon KeyPress
(doesn't work since keyboard isn't grabbed) and ButtonPress (does
work), this is just a fallback for applications that don't update
this themselves.
* src/display.h: (struct _MetaDisplay): Add atom_net_wm_user_time field
* src/screen.c: (meta_screen_apply_startup_properties): Check for
TIMESTAMP provided from startup sequence as well.
* src/stack.c:
s/meta_window_set_stack_position/meta_window_set_stack_position_no_sync/,
(meta_window_set_stack_position): New function which calls the
meta_window_set_stack_position_no_sync function followed immediately
by calling meta_stack_sync_to_server.
* src/window-props.c:
(init_net_wm_user_time), (reload_net_wm_user_time): new functions,
(reload_wm_hints): also load atom_net_wm_user_time
* src/window.c:
new XSERVER_TIME_IS_LATER macro (accounts for timestamp wraparound),
(meta_window_new_with_attrs): add timestamp attributes,
(window_takes_focus_on_map): use TIMESTAMP from startup
notification and _NET_WM_USER_TIME to decide whether to focus new
windows,
(meta_window_show): if app doesn't take focus on map, place it
just below the focused window in the stack
(process_property_notify): check for changes to _NET_WM_USRE_TIME,
(meta_window_stack_just_below): new function
* src/window.h:
(_MetaWindow struct): new fields for initial_timestamp,
initial_timestamp_set, net_wm_user_time_set, and net_wm_user_time,
(meta_window_stack_just_below): new function
2004-06-24 11:47:05 -04:00
window - > initial_timestamp = 0 ; /* not used */
2008-10-31 08:57:47 -04:00
window - > compositor_private = NULL ;
2008-12-18 07:41:56 -05:00
2011-02-25 10:37:05 -05:00
window - > monitor = meta_screen_get_monitor_for_window ( window - > screen , window ) ;
2012-02-24 11:08:55 -05:00
window - > tile_match = NULL ;
2009-09-18 20:50:11 -04:00
/* Assign this #MetaWindow a sequence number which can be used
* for sorting .
*/
window - > stable_sequence = + + display - > window_sequence_counter ;
2003-11-24 14:09:25 -05:00
2013-12-06 13:44:31 -05:00
window - > opacity = 0xFF ;
2014-03-19 09:36:15 -04:00
META_WINDOW_GET_CLASS ( window ) - > manage ( window ) ;
2008-12-18 07:41:56 -05:00
2013-09-10 09:53:31 -04:00
if ( window - > override_redirect )
{
window - > decorated = FALSE ;
window - > always_sticky = TRUE ;
window - > has_close_func = FALSE ;
window - > has_shade_func = FALSE ;
window - > has_move_func = FALSE ;
window - > has_resize_func = FALSE ;
}
2008-11-23 14:28:40 -05:00
if ( ! window - > override_redirect )
meta_window_update_icon_now ( window ) ;
2003-06-04 15:15:46 -04:00
2001-06-04 02:17:52 -04:00
if ( window - > initially_iconic )
{
2001-06-06 00:47:37 -04:00
/* WM_HINTS said minimized */
window - > minimized = TRUE ;
meta_verbose ( " Window %s asked to start out minimized \n " , window - > desc ) ;
2001-06-04 02:17:52 -04:00
}
2001-06-24 02:47:54 -04:00
if ( existing_wm_state = = IconicState )
{
/* WM_STATE said minimized */
window - > minimized = TRUE ;
meta_verbose ( " Window %s had preexisting WM_STATE = IconicState, minimizing \n " ,
window - > desc ) ;
/* Assume window was previously placed, though perhaps it's
* been iconic its whole life , we have no way of knowing .
*/
window - > placed = TRUE ;
}
2002-11-30 22:58:04 -05:00
/* Apply any window attributes such as initial workspace
* based on startup notification
*/
meta_screen_apply_startup_properties ( window - > screen , window ) ;
2007-11-09 15:41:45 -05:00
/* Try to get a "launch timestamp" for the window. If the window is
* a transient , we ' d like to be able to get a last - usage timestamp
* from the parent window . If the window has no parent , there isn ' t
* much we can do . . . except record the current time so that any children
* can use this time as a fallback .
*/
2009-06-14 08:37:57 -04:00
if ( ! window - > override_redirect & & ! window - > net_wm_user_time_set ) {
2008-02-26 23:39:10 -05:00
/* First, maybe the app was launched with startup notification using an
* obsolete version of the spec ; use that timestamp if it exists .
*/
2007-11-09 15:41:45 -05:00
if ( window - > initial_timestamp_set )
2008-02-26 23:39:10 -05:00
/* NOTE: Do NOT toggle net_wm_user_time_set to true; this is just
* being recorded as a fallback for potential transients
*/
2007-11-09 15:41:45 -05:00
window - > net_wm_user_time = window - > initial_timestamp ;
2013-09-16 08:44:37 -04:00
else if ( window - > transient_for ! = NULL )
meta_window_set_user_time ( window , window - > transient_for - > net_wm_user_time ) ;
2007-11-09 15:41:45 -05:00
else
2008-02-26 23:39:10 -05:00
/* NOTE: Do NOT toggle net_wm_user_time_set to true; this is just
* being recorded as a fallback for potential transients
*/
2007-11-09 15:41:45 -05:00
window - > net_wm_user_time =
meta_display_get_current_time_roundtrip ( window - > display ) ;
}
2008-12-18 07:41:56 -05:00
window: make determination of attached dialog windows more consistent
Different bits of code were using slightly different checks to test
whether a window was an attached dialog. Add a new
meta_window_is_attached_dialog(), and use that everywhere.
Also, freeze the is-attached status when the window is first shown,
rather than recomputing it each time the caller asks, since this could
cause problems if a window changes its type after it has already been
attached, etc. However, if an attached window's parent is destroyed,
or an attached window changes its transient-for, then fix things up by
destroying the old MetaWindow and creating a new one (causing
compositor unmap and map events to be fired off, allowing the display
of the window to be fixed up).
Remove some code in display.c that tried to fix existing windows if
the gconf setting changed, but which didn't actually do anything (at
least under gnome-shell). However, if 654643 was fixed then the new
behavior with this patch would be that changing the gconf setting
would affect new dialogs, but not existing ones.
https://bugzilla.gnome.org/show_bug.cgi?id=646761
2011-04-05 10:14:52 -04:00
window - > attached = meta_window_should_attach_to_parent ( window ) ;
if ( window - > attached )
2014-02-01 18:58:58 -05:00
meta_window_recalc_features ( window ) ;
window: make determination of attached dialog windows more consistent
Different bits of code were using slightly different checks to test
whether a window was an attached dialog. Add a new
meta_window_is_attached_dialog(), and use that everywhere.
Also, freeze the is-attached status when the window is first shown,
rather than recomputing it each time the caller asks, since this could
cause problems if a window changes its type after it has already been
attached, etc. However, if an attached window's parent is destroyed,
or an attached window changes its transient-for, then fix things up by
destroying the old MetaWindow and creating a new one (causing
compositor unmap and map events to be fired off, allowing the display
of the window to be fixed up).
Remove some code in display.c that tried to fix existing windows if
the gconf setting changed, but which didn't actually do anything (at
least under gnome-shell). However, if 654643 was fixed then the new
behavior with this patch would be that changing the gconf setting
would affect new dialogs, but not existing ones.
https://bugzilla.gnome.org/show_bug.cgi?id=646761
2011-04-05 10:14:52 -04:00
2001-06-08 02:39:38 -04:00
if ( window - > decorated )
meta_window_ensure_frame ( window ) ;
2001-06-04 02:17:52 -04:00
2001-06-23 01:49:35 -04:00
meta_window_grab_keys ( window ) ;
2008-11-23 14:28:40 -05:00
if ( window - > type ! = META_WINDOW_DOCK & & ! window - > override_redirect )
2008-10-05 21:34:17 -04:00
{
meta_display_grab_window_buttons ( window - > display , window - > xwindow ) ;
meta_display_grab_focus_window_button ( window - > display , window ) ;
}
2003-12-14 14:19:40 -05:00
if ( window - > type = = META_WINDOW_DESKTOP | |
2011-02-28 05:41:23 -05:00
window - > type = = META_WINDOW_DOCK )
2003-12-14 14:19:40 -05:00
{
/* Change the default, but don't enforce this if the user
* focuses the dock / desktop and unsticks it using key shortcuts .
* Need to set this before adding to the workspaces so the MRU
* lists will be updated .
*/
2011-02-28 05:41:23 -05:00
window - > on_all_workspaces_requested = TRUE ;
2003-12-14 14:19:40 -05:00
}
2008-12-18 07:41:56 -05:00
2011-02-28 05:41:23 -05:00
window - > on_all_workspaces = should_be_on_all_workspaces ( window ) ;
2001-06-24 02:47:54 -04:00
/* For the workspace, first honor hints,
* if that fails put transients with parents ,
* otherwise put window on active space
*/
2008-12-18 07:41:56 -05:00
2001-06-24 02:47:54 -04:00
if ( window - > initial_workspace_set )
{
2002-02-06 22:07:56 -05:00
if ( window - > initial_workspace = = ( int ) 0xFFFFFFFF )
2001-10-24 21:39:36 -04:00
{
2002-03-03 23:54:46 -05:00
meta_topic ( META_DEBUG_PLACEMENT ,
" Window %s is initially on all spaces \n " ,
window - > desc ) ;
2008-12-18 07:41:56 -05:00
2003-12-14 14:19:40 -05:00
/* need to set on_all_workspaces first so that it will be
* added to all the MRU lists
*/
2011-02-28 05:41:23 -05:00
window - > on_all_workspaces_requested = TRUE ;
2003-12-13 18:00:06 -05:00
window - > on_all_workspaces = TRUE ;
2003-12-14 14:19:40 -05:00
meta_workspace_add_window ( window - > screen - > active_workspace , window ) ;
2001-10-24 21:39:36 -04:00
}
else
{
2002-03-03 23:54:46 -05:00
meta_topic ( META_DEBUG_PLACEMENT ,
" Window %s is initially on space %d \n " ,
window - > desc , window - > initial_workspace ) ;
2001-10-24 21:39:36 -04:00
space =
2002-10-16 16:12:24 -04:00
meta_screen_get_workspace_by_index ( window - > screen ,
window - > initial_workspace ) ;
2008-12-18 07:41:56 -05:00
2001-10-24 21:39:36 -04:00
if ( space )
meta_workspace_add_window ( space , window ) ;
}
2001-06-24 02:47:54 -04:00
}
2008-12-18 07:41:56 -05:00
2009-06-15 18:02:32 -04:00
/* override-redirect windows are subtly different from other windows
* with window - > on_all_workspaces = = TRUE . Other windows are part of
* some workspace ( so they can return to that if the flag is turned off ) ,
* but appear on other workspaces . override - redirect windows are part
* of no workspace .
*/
if ( ! window - > override_redirect )
2001-06-24 02:47:54 -04:00
{
2009-06-15 18:02:32 -04:00
if ( window - > workspace = = NULL & &
2013-09-16 08:44:37 -04:00
window - > transient_for ! = NULL )
2009-06-15 18:02:32 -04:00
{
/* Try putting dialog on parent's workspace */
2013-09-16 08:44:37 -04:00
if ( window - > transient_for - > workspace )
2009-06-15 18:02:32 -04:00
{
meta_topic ( META_DEBUG_PLACEMENT ,
" Putting window %s on same workspace as parent %s \n " ,
2013-09-16 08:44:37 -04:00
window - > desc , window - > transient_for - > desc ) ;
2008-12-18 07:41:56 -05:00
2013-09-16 08:44:37 -04:00
if ( window - > transient_for - > on_all_workspaces_requested )
2011-02-28 05:41:23 -05:00
{
window - > on_all_workspaces_requested = TRUE ;
window - > on_all_workspaces = TRUE ;
}
2008-12-18 07:41:56 -05:00
2009-06-15 18:02:32 -04:00
/* this will implicitly add to the appropriate MRU lists
*/
2013-09-16 08:44:37 -04:00
meta_workspace_add_window ( window - > transient_for - > workspace , window ) ;
2009-06-15 18:02:32 -04:00
}
2001-06-24 02:47:54 -04:00
}
2008-12-18 07:41:56 -05:00
2009-06-15 18:02:32 -04:00
if ( window - > workspace = = NULL )
{
meta_topic ( META_DEBUG_PLACEMENT ,
" Putting window %s on active workspace \n " ,
window - > desc ) ;
2008-12-18 07:41:56 -05:00
2009-06-15 18:02:32 -04:00
space = window - > screen - > active_workspace ;
2001-06-24 02:47:54 -04:00
2009-06-15 18:02:32 -04:00
meta_workspace_add_window ( space , window ) ;
}
2008-12-18 07:41:56 -05:00
2009-06-15 18:02:32 -04:00
/* for the various on_all_workspaces = TRUE possible above */
2009-06-14 08:37:57 -04:00
meta_window_set_current_workspace_hint ( window ) ;
2008-12-18 07:41:56 -05:00
2009-06-14 08:37:57 -04:00
meta_window_update_struts ( window ) ;
}
2002-08-01 19:52:47 -04:00
2011-02-28 15:26:11 -05:00
g_signal_emit_by_name ( window - > screen , " window-entered-monitor " , window - > monitor - > number , window ) ;
Fix several bugs with handling of fullscreen windows, causing them to not
2006-08-21 Elijah Newren <newren gmail com>
Fix several bugs with handling of fullscreen windows, causing them
to not actually be fullscreen. #343115 (and also #346927,
#350547, #351643, the recent additional WINE-related issue
mentioned on the mailing list, and probably others...)
* src/constraints.c (setup_constraint_info): if a window tries to
resize to fullscreen-size and it has a fullscreen function but
isn't actually marked as fullscreen then assist it by marking it
as such, (constrain_fully_onscreen, constrain_titlebar_visible):
ignore this constraint for fullscreen windows since such windows
have a separate specialized constraint
* src/stack.c (window_is_fullscreen_size, get_standalone_layer):
remove the old window_is_fullscreen_size() hack for detecting
windows to treat as fullscreen since it doesn't work well with the
new constraints framework (i.e. we needed a slightly different
hack)
* src/window.[ch] (meta_window_new_with_addrs): shuffle the order
of adding the window to the stack and moveresizing the window
since moveresizing can cause stack changes if the window's initial
size is fullscreen size, (meta_window_make_fullscreen,
meta_window_make_fullscreen_internal): split
meta_window_make_fullscreen() similar to meta_window_maximize() so
that constraints can make use of it
2006-08-21 14:08:05 -04:00
/* Must add window to stack before doing move/resize, since the
* window might have fullscreen size ( i . e . should have been
* fullscreen ' d ; acrobat is one such braindead case ; it withdraws
* and remaps its window whenever trying to become fullscreen . . . )
* and thus constraints may try to auto - fullscreen it which also
* means restacking it .
*/
2009-06-15 14:31:04 -04:00
if ( ! window - > override_redirect )
meta_stack_add ( window - > screen - > stack ,
window ) ;
else
window - > layer = META_LAYER_OVERRIDE_REDIRECT ; /* otherwise set by MetaStack */
Fix several bugs with handling of fullscreen windows, causing them to not
2006-08-21 Elijah Newren <newren gmail com>
Fix several bugs with handling of fullscreen windows, causing them
to not actually be fullscreen. #343115 (and also #346927,
#350547, #351643, the recent additional WINE-related issue
mentioned on the mailing list, and probably others...)
* src/constraints.c (setup_constraint_info): if a window tries to
resize to fullscreen-size and it has a fullscreen function but
isn't actually marked as fullscreen then assist it by marking it
as such, (constrain_fully_onscreen, constrain_titlebar_visible):
ignore this constraint for fullscreen windows since such windows
have a separate specialized constraint
* src/stack.c (window_is_fullscreen_size, get_standalone_layer):
remove the old window_is_fullscreen_size() hack for detecting
windows to treat as fullscreen since it doesn't work well with the
new constraints framework (i.e. we needed a slightly different
hack)
* src/window.[ch] (meta_window_new_with_addrs): shuffle the order
of adding the window to the stack and moveresizing the window
since moveresizing can cause stack changes if the window's initial
size is fullscreen size, (meta_window_make_fullscreen,
meta_window_make_fullscreen_internal): split
meta_window_make_fullscreen() similar to meta_window_maximize() so
that constraints can make use of it
2006-08-21 14:08:05 -04:00
2001-06-09 17:58:30 -04:00
/* Put our state back where it should be,
* passing TRUE for is_configure_request , ICCCM says
* initial map is handled same as configure request
*/
Patch from Jens Granseuer to fix build with gcc 2.95. #322622.
2006-01-02 Elijah Newren <newren@gmail.com>
Patch from Jens Granseuer to fix build with gcc 2.95. #322622.
* src/boxes.c (meta_rectangle_region_to_string,
meta_rectangle_edge_list_to_string, fix_up_edges):
* src/constraints.c (meta_window_constrain, setup_constraint_info,
place_window_if_needed, constrain_maximization,
constrain_fullscreen, constrain_size_increments,
constrain_size_limits, constrain_aspect_ratio,
do_screen_and_xinerama_relative_constrai,
constrain_to_single_xinerama, constrain_fully_onscreen,
constrain_partially_onscreen):
* src/edge-resistance.c (find_nearest_position,
apply_edge_resistance, apply_edge_resistance_to_each_side):
* src/testboxes.c (test_clamping_to_region,
test_clipping_to_region, test_shoving_into_region):
* src/window.c (meta_window_new_with_attrs,
meta_window_apply_session_info, meta_window_resize,
meta_window_resize_with_gravity, meta_window_configure_request):
Remove C99 style variable initiailization
2006-01-02 13:37:46 -05:00
flags =
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
META_IS_CONFIGURE_REQUEST | META_IS_MOVE_ACTION | META_IS_RESIZE_ACTION ;
2009-06-15 18:25:07 -04:00
if ( ! window - > override_redirect )
meta_window_move_resize_internal ( window ,
flags ,
window - > size_hints . win_gravity ,
window - > size_hints . x ,
window - > size_hints . y ,
window - > size_hints . width ,
window - > size_hints . height ) ;
2001-06-09 17:58:30 -04:00
2001-07-27 00:36:44 -04:00
/* Now try applying saved stuff from the session */
{
const MetaWindowSessionInfo * info ;
info = meta_window_lookup_saved_state ( window ) ;
if ( info )
{
meta_window_apply_session_info ( window , info ) ;
meta_window_release_saved_state ( info ) ;
}
}
2008-12-18 07:41:56 -05:00
2008-11-23 14:28:40 -05:00
if ( ! window - > override_redirect )
{
/* FIXME we have a tendency to set this then immediately
* change it again .
*/
2014-02-01 19:03:33 -05:00
set_wm_state ( window ) ;
2008-11-23 14:28:40 -05:00
set_net_wm_state ( window ) ;
}
2005-01-25 21:47:49 -05:00
2014-03-18 17:06:42 -04:00
meta_compositor_add_window ( screen - > display - > compositor , window ) ;
2014-02-24 12:56:30 -05:00
2001-06-10 23:24:20 -04:00
/* Sync stack changes */
meta_stack_thaw ( window - > screen - > stack ) ;
2002-05-11 01:47:04 -04:00
2009-06-14 08:04:28 -04:00
/* Usually the we'll have queued a stack sync anyways, because we've
* added a new frame window or restacked . But if an undecorated
* window is mapped , already stacked in the right place , then we
* might need to do this explicitly .
*/
meta_stack_tracker_queue_sync_stack ( window - > screen - > stack_tracker ) ;
2002-05-11 01:47:04 -04:00
/* disable show desktop mode unless we're a desktop component */
2003-05-16 17:59:08 -04:00
maybe_leave_show_desktop_mode ( window ) ;
2008-12-18 07:41:56 -05:00
Always map the client and frame windows
Traditionally, WMs unmap windows when minimizing them, and map them
when restoring them or wanting to show them for other reasons, like
upon creation.
However, as metacity morphed into mutter, we optionally chose to keep
windows mapped for the lifetime of the window under the user option
"live-window-previews", which makes the code keep windows mapped so it
can show window preview for minimized windows in other places, like
Alt-Tab and Expose.
I removed this preference two years ago mechanically, by removing all
the if statements, but never went through and cleaned up the code so
that windows are simply mapped for the lifetime of the window -- the
"architecture" of the old code that maps and unmaps on show/hide was
still there.
Remove this now.
The one case we still need to be careful of is shaded windows, in which
we do still unmap the client window. In the future, we might want to
show previews of shaded windows in the overview and Alt-Tab. In that
we'd also keep shaded windows mapped, and could remove all unmap logic,
but we'd need a more complex method of showing the shaded titlebar, such
as using a different actor.
At the same time, simplify the compositor interface by removing
meta_compositor_window_[un]mapped API, and instead adding/removing the
window on-demand.
https://bugzilla.gnome.org/show_bug.cgi?id=720631
2014-02-01 19:06:51 -05:00
sync_client_window_mapped ( window ) ;
Refactor thrice-duplicated queue code in window.c. Closes #376760.
2007-06-10 Thomas Thurman <thomas@thurman.org.uk>
Refactor thrice-duplicated queue code in window.c. Closes #376760.
* src/window.c (meta_window_queue, meta_window_unqueue):
New functions.
* src/window.[ch] (meta_window_unqueue_*, meta_window_queue_*):
Removed functions.
* src/window.c (meta_window_new_with_attrs, meta_window_free,
meta_window_flush_calc_showing, queue_calc_showing_func,
meta_window_minimize, meta_window_unminimize, meta_window_maximize,
meta_window_make_fullscreen, meta_window_shade,
meta_window_unshade, meta_window_move_resize_internal,
window_stick_impl, window_unstick_impl,
meta_window_client_message, process_property_notify): Modified to
use new queueing functions.
* src/window.c (idle_move_resize, idle_update_icon,
idle_calc_showing): update to receive queue number from pointer.
* src/window.h (MetaQueueType): new enum.
* src/window.h (MetaWindow): *_queued replaced with is_in_queue
bitfield.
* src/core.c (meta_core_queue_frame_resize):
* src/display.c (event_callback,
meta_display_queue_retheme_all_windows): Using new queueing functions.
* src/frame.c (meta_window_destroy_frame): Using new queueing functions.
* src/screen.c (queue_resize, meta_screen_resize_func,
queue_windows_showing): Using new queueing functions.
* src/window-props.c (reload_mwm_hints, reload_wm_hints,
reload_transient_for): Using new queueing functions.
* src/workspace.c (meta_workspace_add_window,
meta_workspace_remove_window, meta_workspace_queue_calc_showing,
meta_workspace_invalidate_work_area): Using new queueing functions.
svn path=/trunk/; revision=3236
2007-06-10 21:15:33 -04:00
meta_window_queue ( window , META_QUEUE_CALC_SHOWING ) ;
2005-05-26 13:15:56 -04:00
/* See bug 303284; a transient of the given window can already exist, in which
* case we think it should probably be shown .
*/
meta_window_foreach_transient ( window ,
queue_calc_showing_func ,
NULL ) ;
2006-10-01 17:59:09 -04:00
/* See bug 334899; the window may have minimized ancestors
* which need to be shown .
*
* However , we shouldn ' t unminimize windows here when opening
* a new display because that breaks passing _NET_WM_STATE_HIDDEN
* between window managers when replacing them ; see bug 358042.
2008-03-28 14:44:59 -04:00
*
* And we shouldn ' t unminimize windows if they were initially
* iconic .
2006-04-16 23:54:49 -04:00
*/
2009-06-14 08:11:29 -04:00
if ( ! window - > override_redirect & &
! display - > display_opening & &
! window - > initially_iconic )
2006-10-01 17:59:09 -04:00
unminimize_window_and_all_transient_parents ( window ) ;
2002-10-21 17:44:35 -04:00
deleted and moved into window-props.c (meta_window_new_with_attrs): added
2006-03-25 Thomas Thurman <thomas@thurman.org.uk>
* src/window.c, src/window.h (update_net_wm_state,
update_mwm_hints, update_wm_class, update_transient_for):
deleted and moved into window-props.c
(meta_window_new_with_attrs): added constructing field
and four new initial properties (as above)
(meta_window_recalc_features,
meta_window_recalc_window_type): new functions
* src/window-props.c (init_net_wm_state, reload_net_wm_state
init_mwm_hints, reload_mwm_hints, init_wm_class,
reload_mwm_class, init_transient_for, reload_transient_for):
new functions, moved in from window.c
(meta_display_init_window_prop_hooks): initialise new properties
2006-03-25 17:57:22 -05:00
window - > constructing = FALSE ;
2009-10-13 10:34:34 -04:00
meta_display_notify_window_created ( display , window ) ;
2009-10-28 13:19:24 -04:00
if ( window - > wm_state_demands_attention )
g_signal_emit_by_name ( window - > display , " window-demands-attention " , window ) ;
2013-12-30 14:01:28 -05:00
return window ;
2012-01-07 17:21:32 -05:00
}
MetaWindow *
2014-02-01 19:19:22 -05:00
meta_window_wayland_new ( MetaDisplay * display ,
MetaWaylandSurface * surface )
2012-01-07 17:21:32 -05:00
{
XWindowAttributes attrs ;
MetaScreen * scr = display - > screens - > data ;
MetaWindow * window ;
attrs . x = 0 ;
attrs . y = 0 ;
2013-11-19 15:32:33 -05:00
attrs . width = 1 ;
attrs . height = 1 ;
2012-01-07 17:21:32 -05:00
attrs . border_width = 0 ;
attrs . depth = 24 ;
attrs . visual = NULL ;
attrs . root = scr - > xroot ;
attrs . class = InputOutput ;
attrs . bit_gravity = NorthWestGravity ;
attrs . win_gravity = NorthWestGravity ;
attrs . backing_store = 0 ;
attrs . backing_planes = ~ 0 ;
attrs . backing_pixel = 0 ;
attrs . save_under = 0 ;
attrs . colormap = 0 ;
attrs . map_installed = 1 ;
attrs . map_state = IsUnmapped ;
attrs . all_event_masks = ~ 0 ;
attrs . your_event_mask = 0 ;
attrs . do_not_propagate_mask = 0 ;
attrs . override_redirect = 0 ;
attrs . screen = scr - > xscreen ;
2013-12-30 14:01:28 -05:00
/* XXX: Note: In the Wayland case we currently still trap X errors while
* creating a MetaWindow because we will still be making various redundant
* X requests ( passing a window xid of None ) until we thoroughly audit all
* the code to make sure it knows about non X based clients . . .
2012-01-07 17:21:32 -05:00
*/
meta_error_trap_push ( display ) ; /* Push a trap over all of window
* creation , to reduce XSync ( ) calls
*/
2014-02-01 19:19:22 -05:00
window = _meta_window_shared_new ( display ,
scr ,
META_WINDOW_CLIENT_TYPE_WAYLAND ,
surface ,
None ,
WithdrawnState ,
META_COMP_EFFECT_CREATE ,
& attrs ) ;
2014-02-18 20:28:23 -05:00
window - > can_ping = TRUE ;
2012-01-07 17:21:32 -05:00
meta_error_trap_pop ( display ) ; /* pop the XSync()-reducing trap */
return window ;
}
2003-11-20 21:32:05 -05:00
/* This function should only be called from the end of meta_window_new_with_attrs () */
2001-07-27 00:36:44 -04:00
static void
meta_window_apply_session_info ( MetaWindow * window ,
const MetaWindowSessionInfo * info )
{
2002-06-22 20:21:20 -04:00
if ( info - > stack_position_set )
{
meta_topic ( META_DEBUG_SM ,
" Restoring stack position %d for window %s \n " ,
info - > stack_position , window - > desc ) ;
/* FIXME well, I'm not sure how to do this. */
}
if ( info - > minimized_set )
{
meta_topic ( META_DEBUG_SM ,
" Restoring minimized state %d for window %s \n " ,
info - > minimized , window - > desc ) ;
if ( window - > has_minimize_func & & info - > minimized )
meta_window_minimize ( window ) ;
}
if ( info - > maximized_set )
{
meta_topic ( META_DEBUG_SM ,
" Restoring maximized state %d for window %s \n " ,
info - > maximized , window - > desc ) ;
2008-12-18 07:41:56 -05:00
2002-06-22 20:21:20 -04:00
if ( window - > has_maximize_func & & info - > maximized )
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
{
2014-03-13 18:32:20 -04:00
meta_window_maximize ( window , META_MAXIMIZE_BOTH ) ;
2002-09-03 22:44:52 -04:00
if ( info - > saved_rect_set )
{
meta_topic ( META_DEBUG_SM ,
" Restoring saved rect %d,%d %dx%d for window %s \n " ,
info - > saved_rect . x ,
info - > saved_rect . y ,
info - > saved_rect . width ,
info - > saved_rect . height ,
window - > desc ) ;
2008-12-18 07:41:56 -05:00
2002-09-03 22:44:52 -04:00
window - > saved_rect . x = info - > saved_rect . x ;
window - > saved_rect . y = info - > saved_rect . y ;
window - > saved_rect . width = info - > saved_rect . width ;
window - > saved_rect . height = info - > saved_rect . height ;
}
}
2002-06-22 20:21:20 -04:00
}
2008-12-18 07:41:56 -05:00
2001-07-27 00:36:44 -04:00
if ( info - > on_all_workspaces_set )
{
2011-02-28 05:41:23 -05:00
window - > on_all_workspaces_requested = info - > on_all_workspaces ;
meta_window_update_on_all_workspaces ( window ) ;
2002-02-07 22:34:26 -05:00
meta_topic ( META_DEBUG_SM ,
" Restoring sticky state %d for window %s \n " ,
2011-02-28 05:41:23 -05:00
window - > on_all_workspaces_requested , window - > desc ) ;
2001-07-27 00:36:44 -04:00
}
2008-12-18 07:41:56 -05:00
2001-07-27 00:36:44 -04:00
if ( info - > workspace_indices )
{
GSList * tmp ;
2008-12-18 07:41:56 -05:00
GSList * spaces ;
2001-07-27 00:36:44 -04:00
spaces = NULL ;
2008-12-18 07:41:56 -05:00
2001-07-27 00:36:44 -04:00
tmp = info - > workspace_indices ;
while ( tmp ! = NULL )
{
2008-12-18 07:41:56 -05:00
MetaWorkspace * space ;
2001-07-27 00:36:44 -04:00
space =
2002-10-16 16:12:24 -04:00
meta_screen_get_workspace_by_index ( window - > screen ,
GPOINTER_TO_INT ( tmp - > data ) ) ;
2008-12-18 07:41:56 -05:00
2001-07-27 00:36:44 -04:00
if ( space )
spaces = g_slist_prepend ( spaces , space ) ;
2008-12-18 07:41:56 -05:00
2001-07-27 00:36:44 -04:00
tmp = tmp - > next ;
}
if ( spaces )
{
/* This briefly breaks the invariant that we are supposed
* to always be on some workspace . But we paranoically
* ensured that one of the workspaces from the session was
* indeed valid , so we know we ' ll go right back to one .
*/
Partially resolve the conflicting requirements of windows on multiple
2004-12-22 Elijah Newren <newren@gmail.com>
Partially resolve the conflicting requirements of windows on
multiple workspaces and hidden being a global quantity for windows
(fixes bug 156182; the remainder of the work is bug 87531 and is a
libwnck issue)
* src/display.c (event_callback):
* src/window.c (meta_window_visible_on_workspace, meta_window_unstick):
* src/workspace.c (meta_workspace_add_window,
meta_workspace_contains_window,
meta_workspace_queue_calc_showing):
* src/workspace.h:
Remove meta_workspace_contains_window, replace with simple
comparison utilizing window->workspace
* src/place.c (meta_window_place):
* src/window.c (meta_window_shares_some_workspace):
* src/window.h:
Remove meta_window_shares_some_workspace, replace with a simple
comparison utilizing window->workspace
* src/session.c (save_state),
* src/window.c (meta_window_new_with_attrs,
meta_window_apply_session_info, meta_window_free,
window_showing_on_its_workspace,
meta_window_change_workspace_without_transients,
meta_window_unstick, meta_window_set_current_workspace_hint,
meta_window_get_workspaces):
* src/window.h:
* src/workspace.c (meta_workspace_free, meta_workspace_add_window,
meta_workspace_remove_window):
Only one workspace now
2004-12-22 18:52:52 -05:00
if ( window - > workspace )
meta_workspace_remove_window ( window - > workspace , window ) ;
2001-07-27 00:36:44 -04:00
Partially resolve the conflicting requirements of windows on multiple
2004-12-22 Elijah Newren <newren@gmail.com>
Partially resolve the conflicting requirements of windows on
multiple workspaces and hidden being a global quantity for windows
(fixes bug 156182; the remainder of the work is bug 87531 and is a
libwnck issue)
* src/display.c (event_callback):
* src/window.c (meta_window_visible_on_workspace, meta_window_unstick):
* src/workspace.c (meta_workspace_add_window,
meta_workspace_contains_window,
meta_workspace_queue_calc_showing):
* src/workspace.h:
Remove meta_workspace_contains_window, replace with simple
comparison utilizing window->workspace
* src/place.c (meta_window_place):
* src/window.c (meta_window_shares_some_workspace):
* src/window.h:
Remove meta_window_shares_some_workspace, replace with a simple
comparison utilizing window->workspace
* src/session.c (save_state),
* src/window.c (meta_window_new_with_attrs,
meta_window_apply_session_info, meta_window_free,
window_showing_on_its_workspace,
meta_window_change_workspace_without_transients,
meta_window_unstick, meta_window_set_current_workspace_hint,
meta_window_get_workspaces):
* src/window.h:
* src/workspace.c (meta_workspace_free, meta_workspace_add_window,
meta_workspace_remove_window):
Only one workspace now
2004-12-22 18:52:52 -05:00
/* Only restore to the first workspace if the window
* happened to be on more than one , since we have replaces
* window - > workspaces with window - > workspace
*/
2008-12-18 07:41:56 -05:00
meta_workspace_add_window ( spaces - > data , window ) ;
2001-07-27 00:36:44 -04:00
Partially resolve the conflicting requirements of windows on multiple
2004-12-22 Elijah Newren <newren@gmail.com>
Partially resolve the conflicting requirements of windows on
multiple workspaces and hidden being a global quantity for windows
(fixes bug 156182; the remainder of the work is bug 87531 and is a
libwnck issue)
* src/display.c (event_callback):
* src/window.c (meta_window_visible_on_workspace, meta_window_unstick):
* src/workspace.c (meta_workspace_add_window,
meta_workspace_contains_window,
meta_workspace_queue_calc_showing):
* src/workspace.h:
Remove meta_workspace_contains_window, replace with simple
comparison utilizing window->workspace
* src/place.c (meta_window_place):
* src/window.c (meta_window_shares_some_workspace):
* src/window.h:
Remove meta_window_shares_some_workspace, replace with a simple
comparison utilizing window->workspace
* src/session.c (save_state),
* src/window.c (meta_window_new_with_attrs,
meta_window_apply_session_info, meta_window_free,
window_showing_on_its_workspace,
meta_window_change_workspace_without_transients,
meta_window_unstick, meta_window_set_current_workspace_hint,
meta_window_get_workspaces):
* src/window.h:
* src/workspace.c (meta_workspace_free, meta_workspace_add_window,
meta_workspace_remove_window):
Only one workspace now
2004-12-22 18:52:52 -05:00
meta_topic ( META_DEBUG_SM ,
" Restoring saved window %s to workspace %d \n " ,
window - > desc ,
meta_workspace_index ( spaces - > data ) ) ;
2008-12-18 07:41:56 -05:00
2001-07-27 00:36:44 -04:00
g_slist_free ( spaces ) ;
}
}
2001-07-28 02:35:19 -04:00
if ( info - > geometry_set )
{
int x , y , w , h ;
Patch from Jens Granseuer to fix build with gcc 2.95. #322622.
2006-01-02 Elijah Newren <newren@gmail.com>
Patch from Jens Granseuer to fix build with gcc 2.95. #322622.
* src/boxes.c (meta_rectangle_region_to_string,
meta_rectangle_edge_list_to_string, fix_up_edges):
* src/constraints.c (meta_window_constrain, setup_constraint_info,
place_window_if_needed, constrain_maximization,
constrain_fullscreen, constrain_size_increments,
constrain_size_limits, constrain_aspect_ratio,
do_screen_and_xinerama_relative_constrai,
constrain_to_single_xinerama, constrain_fully_onscreen,
constrain_partially_onscreen):
* src/edge-resistance.c (find_nearest_position,
apply_edge_resistance, apply_edge_resistance_to_each_side):
* src/testboxes.c (test_clamping_to_region,
test_clipping_to_region, test_shoving_into_region):
* src/window.c (meta_window_new_with_attrs,
meta_window_apply_session_info, meta_window_resize,
meta_window_resize_with_gravity, meta_window_configure_request):
Remove C99 style variable initiailization
2006-01-02 13:37:46 -05:00
MetaMoveResizeFlags flags ;
2008-12-18 07:41:56 -05:00
2001-07-28 02:35:19 -04:00
window - > placed = TRUE ; /* don't do placement algorithms later */
x = info - > rect . x ;
y = info - > rect . y ;
w = window - > size_hints . base_width +
info - > rect . width * window - > size_hints . width_inc ;
h = window - > size_hints . base_height +
info - > rect . height * window - > size_hints . height_inc ;
/* Force old gravity, ignoring anything now set */
window - > size_hints . win_gravity = info - > gravity ;
2008-12-18 07:41:56 -05:00
2002-02-07 22:34:26 -05:00
meta_topic ( META_DEBUG_SM ,
" Restoring pos %d,%d size %d x %d for %s \n " ,
x , y , w , h , window - > desc ) ;
2008-12-18 07:41:56 -05:00
Patch from Jens Granseuer to fix build with gcc 2.95. #322622.
2006-01-02 Elijah Newren <newren@gmail.com>
Patch from Jens Granseuer to fix build with gcc 2.95. #322622.
* src/boxes.c (meta_rectangle_region_to_string,
meta_rectangle_edge_list_to_string, fix_up_edges):
* src/constraints.c (meta_window_constrain, setup_constraint_info,
place_window_if_needed, constrain_maximization,
constrain_fullscreen, constrain_size_increments,
constrain_size_limits, constrain_aspect_ratio,
do_screen_and_xinerama_relative_constrai,
constrain_to_single_xinerama, constrain_fully_onscreen,
constrain_partially_onscreen):
* src/edge-resistance.c (find_nearest_position,
apply_edge_resistance, apply_edge_resistance_to_each_side):
* src/testboxes.c (test_clamping_to_region,
test_clipping_to_region, test_shoving_into_region):
* src/window.c (meta_window_new_with_attrs,
meta_window_apply_session_info, meta_window_resize,
meta_window_resize_with_gravity, meta_window_configure_request):
Remove C99 style variable initiailization
2006-01-02 13:37:46 -05:00
flags = META_DO_GRAVITY_ADJUST | META_IS_MOVE_ACTION | META_IS_RESIZE_ACTION ;
2001-07-28 02:35:19 -04:00
meta_window_move_resize_internal ( window ,
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
flags ,
window - > size_hints . win_gravity ,
2001-07-28 02:35:19 -04:00
x , y , w , h ) ;
}
2001-07-27 00:36:44 -04:00
}
window: make determination of attached dialog windows more consistent
Different bits of code were using slightly different checks to test
whether a window was an attached dialog. Add a new
meta_window_is_attached_dialog(), and use that everywhere.
Also, freeze the is-attached status when the window is first shown,
rather than recomputing it each time the caller asks, since this could
cause problems if a window changes its type after it has already been
attached, etc. However, if an attached window's parent is destroyed,
or an attached window changes its transient-for, then fix things up by
destroying the old MetaWindow and creating a new one (causing
compositor unmap and map events to be fired off, allowing the display
of the window to be fixed up).
Remove some code in display.c that tried to fix existing windows if
the gconf setting changed, but which didn't actually do anything (at
least under gnome-shell). However, if 654643 was fixed then the new
behavior with this patch would be that changing the gconf setting
would affect new dialogs, but not existing ones.
https://bugzilla.gnome.org/show_bug.cgi?id=646761
2011-04-05 10:14:52 -04:00
static gboolean
detach_foreach_func ( MetaWindow * window ,
void * data )
{
GList * * children = data ;
MetaWindow * parent ;
if ( window - > attached )
{
/* Only return the immediate children of the window being unmanaged */
parent = meta_window_get_transient_for ( window ) ;
if ( parent - > unmanaging )
* children = g_list_prepend ( * children , window ) ;
}
return TRUE ;
}
2001-05-30 11:36:31 -04:00
void
2008-11-07 04:57:00 -05:00
meta_window_unmanage ( MetaWindow * window ,
guint32 timestamp )
2001-05-30 11:36:31 -04:00
{
2001-06-06 00:47:37 -04:00
GList * tmp ;
2008-12-18 07:41:56 -05:00
2001-05-31 02:42:58 -04:00
meta_verbose ( " Unmanaging 0x%lx \n " , window - > xwindow ) ;
2001-06-06 00:47:37 -04:00
2014-03-19 09:36:15 -04:00
/* This needs to happen for both Wayland and XWayland clients,
* so it can ' t be in MetaWindowWayland . */
2014-02-18 16:39:23 -05:00
if ( window - > surface )
meta_wayland_surface_window_unmanaged ( window - > surface ) ;
2014-03-18 17:06:42 -04:00
if ( window - > visible_to_compositor )
meta_compositor_hide_window ( window - > display - > compositor , window ,
META_COMP_EFFECT_DESTROY ) ;
Simplify relationship between mapping and visibility
Previously, changes to the visibility of a window could be indicated
by meta_compositor_map_window(), meta_compositor_unminimize_window(),
meta_compositor_set_window_hidden(), etc, with the exact behavior
depending on the 'live_hidden_windows' preference.
Simplify this so that visibility is controlled by:
meta_compositor_show_window()
meta_compositor_hide_window()
With an 'effect' parameter provided to indicate the appropriate
effect (CREATE/UNMINIMIZE/MINIMIZE/DESTROY/NONE.)
The map state of the window is signalled separately by:
meta_compositor_map_window()
meta_compositor_unmap_window()
And is used only to control resource handling.
Other changes:
* The desired effect on show/hide is explicitly stored in
MetaWindow, avoiding the need for the was_minimized flag.
At idle, once we calculate the window state, we pass the
effect to the compositor if it matches the new window
state, and then clear the effect to start over for future
map state changes.
* meta_compositor_switch_workspace() is called before any windows
are hidden or shown, allowing the compositor to avoid hiding
or showing an effect for windows involved in the switch.
http://bugzilla.gnome.org/show_bug.cgi?id=582341
* Handling of post-effect cleanups for MutterWindow are
simplified - instead of trying to do different things based
on the individual needs of different effects, we just wait until
all effects complete and sync the window state to what it
should be.
* On unmap, once we destroy the pixmap, we tell ClutterX11Pixmap
that we've done so, so it can clean up and unbind. (The
unbinding doesn't seem to be working properly because of
ClutterGLXPixmap or video driver issues.)
http://bugzilla.gnome.org/show_bug.cgi?id=587251
2009-06-28 17:10:40 -04:00
2014-03-18 17:06:42 -04:00
meta_compositor_remove_window ( window - > display - > compositor , window ) ;
2008-12-18 07:41:56 -05:00
2002-10-28 00:35:30 -05:00
if ( window - > display - > window_with_menu = = window )
{
meta_ui_window_menu_free ( window - > display - > window_menu ) ;
window - > display - > window_menu = NULL ;
window - > display - > window_with_menu = NULL ;
}
2008-12-18 07:41:56 -05:00
2002-01-18 22:50:03 -05:00
if ( destroying_windows_disallowed > 0 )
meta_bug ( " Tried to destroy window %s while destruction was not allowed \n " ,
window - > desc ) ;
2008-12-18 07:41:56 -05:00
2001-06-21 23:20:21 -04:00
window - > unmanaging = TRUE ;
2002-08-25 16:59:09 -04:00
window: make determination of attached dialog windows more consistent
Different bits of code were using slightly different checks to test
whether a window was an attached dialog. Add a new
meta_window_is_attached_dialog(), and use that everywhere.
Also, freeze the is-attached status when the window is first shown,
rather than recomputing it each time the caller asks, since this could
cause problems if a window changes its type after it has already been
attached, etc. However, if an attached window's parent is destroyed,
or an attached window changes its transient-for, then fix things up by
destroying the old MetaWindow and creating a new one (causing
compositor unmap and map events to be fired off, allowing the display
of the window to be fixed up).
Remove some code in display.c that tried to fix existing windows if
the gconf setting changed, but which didn't actually do anything (at
least under gnome-shell). However, if 654643 was fixed then the new
behavior with this patch would be that changing the gconf setting
would affect new dialogs, but not existing ones.
https://bugzilla.gnome.org/show_bug.cgi?id=646761
2011-04-05 10:14:52 -04:00
if ( meta_prefs_get_attach_modal_dialogs ( ) )
{
GList * attached_children = NULL , * iter ;
/* Detach any attached dialogs by unmapping and letting them
* be remapped after @ window is destroyed .
*/
meta_window_foreach_transient ( window ,
detach_foreach_func ,
& attached_children ) ;
for ( iter = attached_children ; iter ; iter = iter - > next )
meta_window_unmanage ( iter - > data , timestamp ) ;
g_list_free ( attached_children ) ;
}
2002-08-25 16:59:09 -04:00
if ( window - > fullscreen )
{
MetaGroup * group ;
/* If the window is fullscreen, it may be forcing
* other windows in its group to a higher layer
*/
meta_stack_freeze ( window - > screen - > stack ) ;
group = meta_window_get_group ( window ) ;
if ( group )
meta_group_update_layers ( group ) ;
meta_stack_thaw ( window - > screen - > stack ) ;
}
2002-10-27 09:45:29 -05:00
meta_window_shutdown_group ( window ) ; /* safe to do this early as
* group . c won ' t re - add to the
* group if window - > unmanaging
*/
2008-12-18 07:41:56 -05:00
2002-01-03 18:28:19 -05:00
/* If we have the focus, focus some other window.
* This is done first , so that if the unmap causes
* an EnterNotify the EnterNotify will have final say
* on what gets focused , maintaining sloppy focus
* invariants .
*/
2011-04-13 17:08:12 -04:00
if ( meta_window_appears_focused ( window ) )
meta_window_propagate_focus_appearance ( window , FALSE ) ;
2002-01-03 18:28:19 -05:00
if ( window - > has_focus )
{
meta_topic ( META_DEBUG_FOCUS ,
2004-06-24 16:02:46 -04:00
" Focusing default window since we're unmanaging %s \n " ,
2002-01-03 18:28:19 -05:00
window - > desc ) ;
Partial audit to fix timestamp usage. One step towards fixing #355180; see
2006-09-18 Elijah Newren <newren gmail com>
Partial audit to fix timestamp usage. One step towards fixing
#355180; see important comments in that bug.
* src/core.[ch] (meta_core_unshade, meta_core_shade):
* src/delete.c (meta_window_present_delete_dialog,
delete_ping_timeout_func):
* src/display.[ch] (meta_display_open, meta_display_close,
event_callback, meta_display_begin_grab_op,
process_selection_clear, meta_display_unmanage_screen,
meta_display_unmanage_windows_for_screen):
* src/frames.c (meta_frames_button_press_event):
* src/keybindings.c (handle_toggle_shade):
* src/main.c (main):
* src/screen.[ch] (update_num_workspaces, meta_screen_new,
meta_screen_free, prefs_changed_callback):
* src/window.[ch] (meta_window_free, finish_minimize,
implement_showing, meta_window_show, meta_window_maximize,
meta_window_make_fullscreen_internal,
meta_window_unmake_fullscreen, meta_window_shade,
meta_window_unshade, window_activate, send_sync_request,
meta_window_client_message, menu_callback,
meta_window_update_keyboard_resize):
Remove usage of CurrentTime, meta_display_get_current_time() and
meta_display_get_current_time_roundtrip() where possible, or
document why it isn't possible, or at very least add a FIXME with
some explanation of my laziness and what needs to be done.
2006-09-18 13:27:24 -04:00
meta_workspace_focus_default_window ( window - > screen - > active_workspace ,
window ,
timestamp ) ;
2002-01-03 18:28:19 -05:00
}
else
{
meta_topic ( META_DEBUG_FOCUS ,
" Unmanaging window %s which doesn't currently have focus \n " ,
window - > desc ) ;
}
2002-01-10 22:31:14 -05:00
2011-04-26 08:12:04 -04:00
g_assert ( window - > display - > focus_window ! = window ) ;
2003-06-25 23:09:38 -04:00
if ( window - > struts )
2002-02-07 22:34:26 -05:00
{
2007-04-07 17:42:05 -04:00
meta_free_gslist_and_elements ( window - > struts ) ;
2003-06-25 23:09:38 -04:00
window - > struts = NULL ;
2002-02-07 22:34:26 -05:00
meta_topic ( META_DEBUG_WORKAREA ,
" Unmanaging window %s which has struts, so invalidating work areas \n " ,
window - > desc ) ;
invalidate_work_areas ( window ) ;
}
2008-12-18 07:41:56 -05:00
2013-03-13 15:31:11 -04:00
if ( window - > sync_request_timeout_id )
{
g_source_remove ( window - > sync_request_timeout_id ) ;
window - > sync_request_timeout_id = 0 ;
}
2001-07-11 02:22:00 -04:00
if ( window - > display - > grab_window = = window )
Partial audit to fix timestamp usage. One step towards fixing #355180; see
2006-09-18 Elijah Newren <newren gmail com>
Partial audit to fix timestamp usage. One step towards fixing
#355180; see important comments in that bug.
* src/core.[ch] (meta_core_unshade, meta_core_shade):
* src/delete.c (meta_window_present_delete_dialog,
delete_ping_timeout_func):
* src/display.[ch] (meta_display_open, meta_display_close,
event_callback, meta_display_begin_grab_op,
process_selection_clear, meta_display_unmanage_screen,
meta_display_unmanage_windows_for_screen):
* src/frames.c (meta_frames_button_press_event):
* src/keybindings.c (handle_toggle_shade):
* src/main.c (main):
* src/screen.[ch] (update_num_workspaces, meta_screen_new,
meta_screen_free, prefs_changed_callback):
* src/window.[ch] (meta_window_free, finish_minimize,
implement_showing, meta_window_show, meta_window_maximize,
meta_window_make_fullscreen_internal,
meta_window_unmake_fullscreen, meta_window_shade,
meta_window_unshade, window_activate, send_sync_request,
meta_window_client_message, menu_callback,
meta_window_update_keyboard_resize):
Remove usage of CurrentTime, meta_display_get_current_time() and
meta_display_get_current_time_roundtrip() where possible, or
document why it isn't possible, or at very least add a FIXME with
some explanation of my laziness and what needs to be done.
2006-09-18 13:27:24 -04:00
meta_display_end_grab_op ( window - > display , timestamp ) ;
2002-03-06 17:27:24 -05:00
g_assert ( window - > display - > grab_window ! = window ) ;
2008-12-18 07:41:56 -05:00
2006-03-06 11:33:33 -05:00
if ( window - > maximized_horizontally | | window - > maximized_vertically )
unmaximize_window_before_freeing ( window ) ;
2007-04-12 00:02:57 -04:00
Refactor thrice-duplicated queue code in window.c. Closes #376760.
2007-06-10 Thomas Thurman <thomas@thurman.org.uk>
Refactor thrice-duplicated queue code in window.c. Closes #376760.
* src/window.c (meta_window_queue, meta_window_unqueue):
New functions.
* src/window.[ch] (meta_window_unqueue_*, meta_window_queue_*):
Removed functions.
* src/window.c (meta_window_new_with_attrs, meta_window_free,
meta_window_flush_calc_showing, queue_calc_showing_func,
meta_window_minimize, meta_window_unminimize, meta_window_maximize,
meta_window_make_fullscreen, meta_window_shade,
meta_window_unshade, meta_window_move_resize_internal,
window_stick_impl, window_unstick_impl,
meta_window_client_message, process_property_notify): Modified to
use new queueing functions.
* src/window.c (idle_move_resize, idle_update_icon,
idle_calc_showing): update to receive queue number from pointer.
* src/window.h (MetaQueueType): new enum.
* src/window.h (MetaWindow): *_queued replaced with is_in_queue
bitfield.
* src/core.c (meta_core_queue_frame_resize):
* src/display.c (event_callback,
meta_display_queue_retheme_all_windows): Using new queueing functions.
* src/frame.c (meta_window_destroy_frame): Using new queueing functions.
* src/screen.c (queue_resize, meta_screen_resize_func,
queue_windows_showing): Using new queueing functions.
* src/window-props.c (reload_mwm_hints, reload_wm_hints,
reload_transient_for): Using new queueing functions.
* src/workspace.c (meta_workspace_add_window,
meta_workspace_remove_window, meta_workspace_queue_calc_showing,
meta_workspace_invalidate_work_area): Using new queueing functions.
svn path=/trunk/; revision=3236
2007-06-10 21:15:33 -04:00
meta_window_unqueue ( window , META_QUEUE_CALC_SHOWING |
META_QUEUE_MOVE_RESIZE |
META_QUEUE_UPDATE_ICON ) ;
2002-04-13 00:58:45 -04:00
meta_window_free_delete_dialog ( window ) ;
2008-12-18 07:41:56 -05:00
Partially resolve the conflicting requirements of windows on multiple
2004-12-22 Elijah Newren <newren@gmail.com>
Partially resolve the conflicting requirements of windows on
multiple workspaces and hidden being a global quantity for windows
(fixes bug 156182; the remainder of the work is bug 87531 and is a
libwnck issue)
* src/display.c (event_callback):
* src/window.c (meta_window_visible_on_workspace, meta_window_unstick):
* src/workspace.c (meta_workspace_add_window,
meta_workspace_contains_window,
meta_workspace_queue_calc_showing):
* src/workspace.h:
Remove meta_workspace_contains_window, replace with simple
comparison utilizing window->workspace
* src/place.c (meta_window_place):
* src/window.c (meta_window_shares_some_workspace):
* src/window.h:
Remove meta_window_shares_some_workspace, replace with a simple
comparison utilizing window->workspace
* src/session.c (save_state),
* src/window.c (meta_window_new_with_attrs,
meta_window_apply_session_info, meta_window_free,
window_showing_on_its_workspace,
meta_window_change_workspace_without_transients,
meta_window_unstick, meta_window_set_current_workspace_hint,
meta_window_get_workspaces):
* src/window.h:
* src/workspace.c (meta_workspace_free, meta_workspace_add_window,
meta_workspace_remove_window):
Only one workspace now
2004-12-22 18:52:52 -05:00
if ( window - > workspace )
meta_workspace_remove_window ( window - > workspace , window ) ;
2001-06-06 00:47:37 -04:00
Partially resolve the conflicting requirements of windows on multiple
2004-12-22 Elijah Newren <newren@gmail.com>
Partially resolve the conflicting requirements of windows on
multiple workspaces and hidden being a global quantity for windows
(fixes bug 156182; the remainder of the work is bug 87531 and is a
libwnck issue)
* src/display.c (event_callback):
* src/window.c (meta_window_visible_on_workspace, meta_window_unstick):
* src/workspace.c (meta_workspace_add_window,
meta_workspace_contains_window,
meta_workspace_queue_calc_showing):
* src/workspace.h:
Remove meta_workspace_contains_window, replace with simple
comparison utilizing window->workspace
* src/place.c (meta_window_place):
* src/window.c (meta_window_shares_some_workspace):
* src/window.h:
Remove meta_window_shares_some_workspace, replace with a simple
comparison utilizing window->workspace
* src/session.c (save_state),
* src/window.c (meta_window_new_with_attrs,
meta_window_apply_session_info, meta_window_free,
window_showing_on_its_workspace,
meta_window_change_workspace_without_transients,
meta_window_unstick, meta_window_set_current_workspace_hint,
meta_window_get_workspaces):
* src/window.h:
* src/workspace.c (meta_workspace_free, meta_workspace_add_window,
meta_workspace_remove_window):
Only one workspace now
2004-12-22 18:52:52 -05:00
g_assert ( window - > workspace = = NULL ) ;
2001-06-07 01:18:10 -04:00
2003-12-12 22:58:55 -05:00
# ifndef G_DISABLE_CHECKS
tmp = window - > screen - > workspaces ;
while ( tmp ! = NULL )
{
MetaWorkspace * workspace = tmp - > data ;
g_assert ( g_list_find ( workspace - > windows , window ) = = NULL ) ;
g_assert ( g_list_find ( workspace - > mru_list , window ) = = NULL ) ;
tmp = tmp - > next ;
}
# endif
2008-12-18 07:41:56 -05:00
2011-02-28 15:26:11 -05:00
if ( window - > monitor )
{
g_signal_emit_by_name ( window - > screen , " window-left-monitor " ,
window - > monitor - > number , window ) ;
window - > monitor = NULL ;
}
2009-06-15 14:31:04 -04:00
if ( ! window - > override_redirect )
meta_stack_remove ( window - > screen - > stack , window ) ;
2008-12-18 07:41:56 -05:00
2011-06-08 11:26:03 -04:00
meta_window_destroy_sync_request_alarm ( window ) ;
2009-06-14 08:04:28 -04:00
/* If an undecorated window is being withdrawn, that will change the
* stack as presented to the compositing manager , without actually
* changing the stacking order of X windows .
*/
meta_stack_tracker_queue_sync_stack ( window - > screen - > stack_tracker ) ;
2001-06-23 01:49:35 -04:00
meta_window_ungrab_keys ( window ) ;
2001-07-25 22:55:35 -04:00
meta_display_ungrab_window_buttons ( window - > display , window - > xwindow ) ;
2003-10-13 16:15:40 -04:00
meta_display_ungrab_focus_window_button ( window - > display , window ) ;
2013-10-16 00:19:33 -04:00
if ( window - > display - > autoraise_window = = window )
meta_display_remove_autoraise_callback ( window - > display ) ;
2008-12-18 07:41:56 -05:00
2014-03-19 09:36:15 -04:00
META_WINDOW_GET_CLASS ( window ) - > unmanage ( window ) ;
2002-02-09 18:03:52 -05:00
2014-03-19 10:30:12 -04:00
if ( window - > frame )
meta_window_destroy_frame ( window ) ;
2012-02-29 16:31:30 -05:00
meta_prefs_remove_listener ( prefs_changed_callback , window ) ;
2013-03-14 16:55:49 -04:00
meta_screen_queue_check_fullscreen ( window - > screen ) ;
2009-10-13 10:34:34 -04:00
g_signal_emit ( window , window_signals [ UNMANAGED ] , 0 ) ;
2008-11-07 04:57:00 -05:00
g_object_unref ( window ) ;
2001-05-30 11:36:31 -04:00
}
2011-02-28 05:41:23 -05:00
static gboolean
should_be_on_all_workspaces ( MetaWindow * window )
{
return
window - > on_all_workspaces_requested | |
2011-02-28 07:48:28 -05:00
window - > override_redirect | |
( meta_prefs_get_workspaces_only_on_primary ( ) & &
! meta_window_is_on_primary_monitor ( window ) ) ;
2011-02-28 05:41:23 -05:00
}
void
meta_window_update_on_all_workspaces ( MetaWindow * window )
{
gboolean old_value ;
old_value = window - > on_all_workspaces ;
window - > on_all_workspaces = should_be_on_all_workspaces ( window ) ;
if ( window - > on_all_workspaces ! = old_value & &
! window - > override_redirect )
{
if ( window - > on_all_workspaces )
{
GList * tmp = window - > screen - > workspaces ;
/* Add to all MRU lists */
while ( tmp )
{
MetaWorkspace * work = ( MetaWorkspace * ) tmp - > data ;
if ( ! g_list_find ( work - > mru_list , window ) )
work - > mru_list = g_list_prepend ( work - > mru_list , window ) ;
tmp = tmp - > next ;
}
}
else
{
GList * tmp = window - > screen - > workspaces ;
/* Remove from MRU lists except the window's workspace */
while ( tmp )
{
MetaWorkspace * work = ( MetaWorkspace * ) tmp - > data ;
if ( work ! = window - > workspace )
work - > mru_list = g_list_remove ( work - > mru_list , window ) ;
tmp = tmp - > next ;
}
}
meta_window_set_current_workspace_hint ( window ) ;
}
}
2011-03-08 17:15:56 -05:00
static void
2014-02-01 19:03:33 -05:00
set_wm_state ( MetaWindow * window )
2011-03-08 17:15:56 -05:00
{
2014-02-01 19:19:22 -05:00
if ( window - > client_type = = META_WINDOW_CLIENT_TYPE_X11 )
meta_window_x11_set_wm_state ( window ) ;
2001-06-10 23:24:20 -04:00
}
2002-10-21 17:44:35 -04:00
static void
2001-06-10 23:24:20 -04:00
set_net_wm_state ( MetaWindow * window )
{
2014-01-23 16:33:53 -05:00
if ( window - > client_type = = META_WINDOW_CLIENT_TYPE_X11 )
meta_window_x11_set_net_wm_state ( window ) ;
2001-06-06 00:47:37 -04:00
}
2013-01-14 12:34:14 -05:00
/**
* meta_window_located_on_workspace :
* @ window : a # MetaWindow
* @ workspace : a # MetaWorkspace
*
2013-02-15 13:42:08 -05:00
* Returns : whether @ window is displayed on @ workspace , or whether it
2013-01-14 12:34:14 -05:00
* will be displayed on all workspaces .
*/
2001-10-25 00:51:25 -04:00
gboolean
2004-12-22 19:20:33 -05:00
meta_window_located_on_workspace ( MetaWindow * window ,
2001-10-25 00:51:25 -04:00
MetaWorkspace * workspace )
{
2002-06-19 22:29:19 -04:00
return ( window - > on_all_workspaces & & window - > screen = = workspace - > screen ) | |
Partially resolve the conflicting requirements of windows on multiple
2004-12-22 Elijah Newren <newren@gmail.com>
Partially resolve the conflicting requirements of windows on
multiple workspaces and hidden being a global quantity for windows
(fixes bug 156182; the remainder of the work is bug 87531 and is a
libwnck issue)
* src/display.c (event_callback):
* src/window.c (meta_window_visible_on_workspace, meta_window_unstick):
* src/workspace.c (meta_workspace_add_window,
meta_workspace_contains_window,
meta_workspace_queue_calc_showing):
* src/workspace.h:
Remove meta_workspace_contains_window, replace with simple
comparison utilizing window->workspace
* src/place.c (meta_window_place):
* src/window.c (meta_window_shares_some_workspace):
* src/window.h:
Remove meta_window_shares_some_workspace, replace with a simple
comparison utilizing window->workspace
* src/session.c (save_state),
* src/window.c (meta_window_new_with_attrs,
meta_window_apply_session_info, meta_window_free,
window_showing_on_its_workspace,
meta_window_change_workspace_without_transients,
meta_window_unstick, meta_window_set_current_workspace_hint,
meta_window_get_workspaces):
* src/window.h:
* src/workspace.c (meta_workspace_free, meta_workspace_add_window,
meta_workspace_remove_window):
Only one workspace now
2004-12-22 18:52:52 -05:00
( window - > workspace = = workspace ) ;
2001-10-25 00:51:25 -04:00
}
2003-03-17 01:36:39 -05:00
static gboolean
is_minimized_foreach ( MetaWindow * window ,
void * data )
{
gboolean * result = data ;
* result = window - > minimized ;
if ( * result )
return FALSE ; /* stop as soon as we find one */
else
return TRUE ;
}
static gboolean
ancestor_is_minimized ( MetaWindow * window )
{
gboolean is_minimized ;
is_minimized = FALSE ;
meta_window_foreach_ancestor ( window , is_minimized_foreach , & is_minimized ) ;
return is_minimized ;
}
2011-03-09 10:01:44 -05:00
/**
* meta_window_showing_on_its_workspace :
* @ window : A # MetaWindow
*
* Returns : % TRUE if window would be visible , if its workspace was current
*/
2005-01-28 09:48:47 -05:00
gboolean
meta_window_showing_on_its_workspace ( MetaWindow * window )
2001-06-06 00:47:37 -04:00
{
2004-10-22 16:28:36 -04:00
gboolean showing ;
2003-11-08 00:29:23 -05:00
gboolean is_desktop_or_dock ;
2004-10-22 16:28:36 -04:00
MetaWorkspace * workspace_of_window ;
2001-07-03 22:10:54 -04:00
2004-10-22 16:28:36 -04:00
showing = TRUE ;
2001-10-14 23:39:41 -04:00
2004-10-22 16:28:36 -04:00
/* 1. See if we're minimized */
on unminimize, queue calc_showing on all transients
2002-05-05 Havoc Pennington <hp@pobox.com>
* src/window.c (meta_window_unminimize): on unminimize, queue
calc_showing on all transients
(meta_window_activate): on activate, unminimize all a window's
ancestors, not just the window itself.
* src/workspace.c (set_work_area_hint): don't increment "tmp" by
16 unsigned long, increment by 4
* src/window.c (meta_window_free): if a window isn't minimized,
restore its WM_STATE to NormalState instead of IconicState,
since IconicState on initial window map means that the window
should be minimized.
* src/workspace.c (meta_workspace_invalidate_work_area): queue an
idle to recompute the work area hint.
(set_work_area_hint): we need 4*num_workspaces ints, not just
num_workspaces.
* src/screen.c (meta_screen_new): add work_area_idle field,
handle it on screen shutdown
* src/common.h (META_PRIORITY_PREFS_NOTIFY,
META_PRIORITY_WORK_AREA_HINT): define some idle priorities
* src/window.c (meta_window_calc_showing): hide windows if
their parent window is minimized
(meta_window_minimize): also queue_calc_showing on all
transients of the window being minimized
* src/place.c (constrain_placement): function to apply
placement-time-only constraints, such as "not off the left of the
screen"
(meta_window_place): put dialogs down a bit over their parent,
not right at the top.
(meta_window_place): when centering a dialog, center it
on the current xinerama screen, rather than the entire
screen.
* src/screen.c (meta_screen_get_current_xinerama): new function,
but not implemented
2002-05-05 01:41:13 -04:00
if ( window - > minimized )
showing = FALSE ;
2008-12-18 07:41:56 -05:00
2004-10-22 16:28:36 -04:00
/* 2. See if we're in "show desktop" mode */
2003-11-08 00:29:23 -05:00
is_desktop_or_dock = FALSE ;
is_desktop_or_dock_foreach ( window ,
& is_desktop_or_dock ) ;
meta_window_foreach_ancestor ( window , is_desktop_or_dock_foreach ,
& is_desktop_or_dock ) ;
2004-10-22 16:28:36 -04:00
if ( window - > on_all_workspaces )
workspace_of_window = window - > screen - > active_workspace ;
Partially resolve the conflicting requirements of windows on multiple
2004-12-22 Elijah Newren <newren@gmail.com>
Partially resolve the conflicting requirements of windows on
multiple workspaces and hidden being a global quantity for windows
(fixes bug 156182; the remainder of the work is bug 87531 and is a
libwnck issue)
* src/display.c (event_callback):
* src/window.c (meta_window_visible_on_workspace, meta_window_unstick):
* src/workspace.c (meta_workspace_add_window,
meta_workspace_contains_window,
meta_workspace_queue_calc_showing):
* src/workspace.h:
Remove meta_workspace_contains_window, replace with simple
comparison utilizing window->workspace
* src/place.c (meta_window_place):
* src/window.c (meta_window_shares_some_workspace):
* src/window.h:
Remove meta_window_shares_some_workspace, replace with a simple
comparison utilizing window->workspace
* src/session.c (save_state),
* src/window.c (meta_window_new_with_attrs,
meta_window_apply_session_info, meta_window_free,
window_showing_on_its_workspace,
meta_window_change_workspace_without_transients,
meta_window_unstick, meta_window_set_current_workspace_hint,
meta_window_get_workspaces):
* src/window.h:
* src/workspace.c (meta_workspace_free, meta_workspace_add_window,
meta_workspace_remove_window):
Only one workspace now
2004-12-22 18:52:52 -05:00
else if ( window - > workspace )
workspace_of_window = window - > workspace ;
2004-10-22 16:28:36 -04:00
else /* This only seems to be needed for startup */
workspace_of_window = NULL ;
2008-12-18 07:41:56 -05:00
if ( showing & &
2004-10-22 16:28:36 -04:00
workspace_of_window & & workspace_of_window - > showing_desktop & &
2003-11-08 00:29:23 -05:00
! is_desktop_or_dock )
2001-10-14 23:39:41 -04:00
{
2004-10-22 16:28:36 -04:00
meta_verbose ( " We're showing the desktop on the workspace(s) that window %s is on \n " ,
2001-10-14 23:39:41 -04:00
window - > desc ) ;
on unminimize, queue calc_showing on all transients
2002-05-05 Havoc Pennington <hp@pobox.com>
* src/window.c (meta_window_unminimize): on unminimize, queue
calc_showing on all transients
(meta_window_activate): on activate, unminimize all a window's
ancestors, not just the window itself.
* src/workspace.c (set_work_area_hint): don't increment "tmp" by
16 unsigned long, increment by 4
* src/window.c (meta_window_free): if a window isn't minimized,
restore its WM_STATE to NormalState instead of IconicState,
since IconicState on initial window map means that the window
should be minimized.
* src/workspace.c (meta_workspace_invalidate_work_area): queue an
idle to recompute the work area hint.
(set_work_area_hint): we need 4*num_workspaces ints, not just
num_workspaces.
* src/screen.c (meta_screen_new): add work_area_idle field,
handle it on screen shutdown
* src/common.h (META_PRIORITY_PREFS_NOTIFY,
META_PRIORITY_WORK_AREA_HINT): define some idle priorities
* src/window.c (meta_window_calc_showing): hide windows if
their parent window is minimized
(meta_window_minimize): also queue_calc_showing on all
transients of the window being minimized
* src/place.c (constrain_placement): function to apply
placement-time-only constraints, such as "not off the left of the
screen"
(meta_window_place): put dialogs down a bit over their parent,
not right at the top.
(meta_window_place): when centering a dialog, center it
on the current xinerama screen, rather than the entire
screen.
* src/screen.c (meta_screen_get_current_xinerama): new function,
but not implemented
2002-05-05 01:41:13 -04:00
showing = FALSE ;
2001-10-14 23:39:41 -04:00
}
on unminimize, queue calc_showing on all transients
2002-05-05 Havoc Pennington <hp@pobox.com>
* src/window.c (meta_window_unminimize): on unminimize, queue
calc_showing on all transients
(meta_window_activate): on activate, unminimize all a window's
ancestors, not just the window itself.
* src/workspace.c (set_work_area_hint): don't increment "tmp" by
16 unsigned long, increment by 4
* src/window.c (meta_window_free): if a window isn't minimized,
restore its WM_STATE to NormalState instead of IconicState,
since IconicState on initial window map means that the window
should be minimized.
* src/workspace.c (meta_workspace_invalidate_work_area): queue an
idle to recompute the work area hint.
(set_work_area_hint): we need 4*num_workspaces ints, not just
num_workspaces.
* src/screen.c (meta_screen_new): add work_area_idle field,
handle it on screen shutdown
* src/common.h (META_PRIORITY_PREFS_NOTIFY,
META_PRIORITY_WORK_AREA_HINT): define some idle priorities
* src/window.c (meta_window_calc_showing): hide windows if
their parent window is minimized
(meta_window_minimize): also queue_calc_showing on all
transients of the window being minimized
* src/place.c (constrain_placement): function to apply
placement-time-only constraints, such as "not off the left of the
screen"
(meta_window_place): put dialogs down a bit over their parent,
not right at the top.
(meta_window_place): when centering a dialog, center it
on the current xinerama screen, rather than the entire
screen.
* src/screen.c (meta_screen_get_current_xinerama): new function,
but not implemented
2002-05-05 01:41:13 -04:00
2004-10-22 16:28:36 -04:00
/* 3. See if an ancestor is minimized (note that
on unminimize, queue calc_showing on all transients
2002-05-05 Havoc Pennington <hp@pobox.com>
* src/window.c (meta_window_unminimize): on unminimize, queue
calc_showing on all transients
(meta_window_activate): on activate, unminimize all a window's
ancestors, not just the window itself.
* src/workspace.c (set_work_area_hint): don't increment "tmp" by
16 unsigned long, increment by 4
* src/window.c (meta_window_free): if a window isn't minimized,
restore its WM_STATE to NormalState instead of IconicState,
since IconicState on initial window map means that the window
should be minimized.
* src/workspace.c (meta_workspace_invalidate_work_area): queue an
idle to recompute the work area hint.
(set_work_area_hint): we need 4*num_workspaces ints, not just
num_workspaces.
* src/screen.c (meta_screen_new): add work_area_idle field,
handle it on screen shutdown
* src/common.h (META_PRIORITY_PREFS_NOTIFY,
META_PRIORITY_WORK_AREA_HINT): define some idle priorities
* src/window.c (meta_window_calc_showing): hide windows if
their parent window is minimized
(meta_window_minimize): also queue_calc_showing on all
transients of the window being minimized
* src/place.c (constrain_placement): function to apply
placement-time-only constraints, such as "not off the left of the
screen"
(meta_window_place): put dialogs down a bit over their parent,
not right at the top.
(meta_window_place): when centering a dialog, center it
on the current xinerama screen, rather than the entire
screen.
* src/screen.c (meta_screen_get_current_xinerama): new function,
but not implemented
2002-05-05 01:41:13 -04:00
* ancestor ' s " mapped " field may not be up to date
* since it ' s being computed in this same idle queue )
*/
2008-12-18 07:41:56 -05:00
on unminimize, queue calc_showing on all transients
2002-05-05 Havoc Pennington <hp@pobox.com>
* src/window.c (meta_window_unminimize): on unminimize, queue
calc_showing on all transients
(meta_window_activate): on activate, unminimize all a window's
ancestors, not just the window itself.
* src/workspace.c (set_work_area_hint): don't increment "tmp" by
16 unsigned long, increment by 4
* src/window.c (meta_window_free): if a window isn't minimized,
restore its WM_STATE to NormalState instead of IconicState,
since IconicState on initial window map means that the window
should be minimized.
* src/workspace.c (meta_workspace_invalidate_work_area): queue an
idle to recompute the work area hint.
(set_work_area_hint): we need 4*num_workspaces ints, not just
num_workspaces.
* src/screen.c (meta_screen_new): add work_area_idle field,
handle it on screen shutdown
* src/common.h (META_PRIORITY_PREFS_NOTIFY,
META_PRIORITY_WORK_AREA_HINT): define some idle priorities
* src/window.c (meta_window_calc_showing): hide windows if
their parent window is minimized
(meta_window_minimize): also queue_calc_showing on all
transients of the window being minimized
* src/place.c (constrain_placement): function to apply
placement-time-only constraints, such as "not off the left of the
screen"
(meta_window_place): put dialogs down a bit over their parent,
not right at the top.
(meta_window_place): when centering a dialog, center it
on the current xinerama screen, rather than the entire
screen.
* src/screen.c (meta_screen_get_current_xinerama): new function,
but not implemented
2002-05-05 01:41:13 -04:00
if ( showing )
{
2003-03-17 01:36:39 -05:00
if ( ancestor_is_minimized ( window ) )
showing = FALSE ;
on unminimize, queue calc_showing on all transients
2002-05-05 Havoc Pennington <hp@pobox.com>
* src/window.c (meta_window_unminimize): on unminimize, queue
calc_showing on all transients
(meta_window_activate): on activate, unminimize all a window's
ancestors, not just the window itself.
* src/workspace.c (set_work_area_hint): don't increment "tmp" by
16 unsigned long, increment by 4
* src/window.c (meta_window_free): if a window isn't minimized,
restore its WM_STATE to NormalState instead of IconicState,
since IconicState on initial window map means that the window
should be minimized.
* src/workspace.c (meta_workspace_invalidate_work_area): queue an
idle to recompute the work area hint.
(set_work_area_hint): we need 4*num_workspaces ints, not just
num_workspaces.
* src/screen.c (meta_screen_new): add work_area_idle field,
handle it on screen shutdown
* src/common.h (META_PRIORITY_PREFS_NOTIFY,
META_PRIORITY_WORK_AREA_HINT): define some idle priorities
* src/window.c (meta_window_calc_showing): hide windows if
their parent window is minimized
(meta_window_minimize): also queue_calc_showing on all
transients of the window being minimized
* src/place.c (constrain_placement): function to apply
placement-time-only constraints, such as "not off the left of the
screen"
(meta_window_place): put dialogs down a bit over their parent,
not right at the top.
(meta_window_place): when centering a dialog, center it
on the current xinerama screen, rather than the entire
screen.
* src/screen.c (meta_screen_get_current_xinerama): new function,
but not implemented
2002-05-05 01:41:13 -04:00
}
2002-11-03 15:30:45 -05:00
return showing ;
}
2004-12-19 21:53:08 -05:00
gboolean
meta_window_should_be_showing ( MetaWindow * window )
2004-10-22 16:28:36 -04:00
{
gboolean on_workspace ;
2013-11-19 17:33:57 -05:00
if ( ! window - > surface_mapped )
return FALSE ;
2004-10-22 16:28:36 -04:00
meta_verbose ( " Should be showing for window %s \n " , window - > desc ) ;
/* See if we're on the workspace */
2008-12-18 07:41:56 -05:00
on_workspace = meta_window_located_on_workspace ( window ,
2004-10-22 16:28:36 -04:00
window - > screen - > active_workspace ) ;
if ( ! on_workspace )
meta_verbose ( " Window %s is not on workspace %d \n " ,
window - > desc ,
meta_workspace_index ( window - > screen - > active_workspace ) ) ;
else
meta_verbose ( " Window %s is on the active workspace %d \n " ,
window - > desc ,
meta_workspace_index ( window - > screen - > active_workspace ) ) ;
if ( window - > on_all_workspaces )
meta_verbose ( " Window %s is on all workspaces \n " , window - > desc ) ;
2005-01-28 09:48:47 -05:00
return on_workspace & & meta_window_showing_on_its_workspace ( window ) ;
2004-10-22 16:28:36 -04:00
}
2002-11-03 15:30:45 -05:00
static void
implement_showing ( MetaWindow * window ,
gboolean showing )
{
on unminimize, queue calc_showing on all transients
2002-05-05 Havoc Pennington <hp@pobox.com>
* src/window.c (meta_window_unminimize): on unminimize, queue
calc_showing on all transients
(meta_window_activate): on activate, unminimize all a window's
ancestors, not just the window itself.
* src/workspace.c (set_work_area_hint): don't increment "tmp" by
16 unsigned long, increment by 4
* src/window.c (meta_window_free): if a window isn't minimized,
restore its WM_STATE to NormalState instead of IconicState,
since IconicState on initial window map means that the window
should be minimized.
* src/workspace.c (meta_workspace_invalidate_work_area): queue an
idle to recompute the work area hint.
(set_work_area_hint): we need 4*num_workspaces ints, not just
num_workspaces.
* src/screen.c (meta_screen_new): add work_area_idle field,
handle it on screen shutdown
* src/common.h (META_PRIORITY_PREFS_NOTIFY,
META_PRIORITY_WORK_AREA_HINT): define some idle priorities
* src/window.c (meta_window_calc_showing): hide windows if
their parent window is minimized
(meta_window_minimize): also queue_calc_showing on all
transients of the window being minimized
* src/place.c (constrain_placement): function to apply
placement-time-only constraints, such as "not off the left of the
screen"
(meta_window_place): put dialogs down a bit over their parent,
not right at the top.
(meta_window_place): when centering a dialog, center it
on the current xinerama screen, rather than the entire
screen.
* src/screen.c (meta_screen_get_current_xinerama): new function,
but not implemented
2002-05-05 01:41:13 -04:00
/* Actually show/hide the window */
2002-11-03 18:42:21 -05:00
meta_verbose ( " Implement showing = %d for window %s \n " ,
showing , window - > desc ) ;
2008-12-18 07:41:56 -05:00
Always map the client and frame windows
Traditionally, WMs unmap windows when minimizing them, and map them
when restoring them or wanting to show them for other reasons, like
upon creation.
However, as metacity morphed into mutter, we optionally chose to keep
windows mapped for the lifetime of the window under the user option
"live-window-previews", which makes the code keep windows mapped so it
can show window preview for minimized windows in other places, like
Alt-Tab and Expose.
I removed this preference two years ago mechanically, by removing all
the if statements, but never went through and cleaned up the code so
that windows are simply mapped for the lifetime of the window -- the
"architecture" of the old code that maps and unmaps on show/hide was
still there.
Remove this now.
The one case we still need to be careful of is shaded windows, in which
we do still unmap the client window. In the future, we might want to
show previews of shaded windows in the overview and Alt-Tab. In that
we'd also keep shaded windows mapped, and could remove all unmap logic,
but we'd need a more complex method of showing the shaded titlebar, such
as using a different actor.
At the same time, simplify the compositor interface by removing
meta_compositor_window_[un]mapped API, and instead adding/removing the
window on-demand.
https://bugzilla.gnome.org/show_bug.cgi?id=720631
2014-02-01 19:06:51 -05:00
sync_client_window_mapped ( window ) ;
on unminimize, queue calc_showing on all transients
2002-05-05 Havoc Pennington <hp@pobox.com>
* src/window.c (meta_window_unminimize): on unminimize, queue
calc_showing on all transients
(meta_window_activate): on activate, unminimize all a window's
ancestors, not just the window itself.
* src/workspace.c (set_work_area_hint): don't increment "tmp" by
16 unsigned long, increment by 4
* src/window.c (meta_window_free): if a window isn't minimized,
restore its WM_STATE to NormalState instead of IconicState,
since IconicState on initial window map means that the window
should be minimized.
* src/workspace.c (meta_workspace_invalidate_work_area): queue an
idle to recompute the work area hint.
(set_work_area_hint): we need 4*num_workspaces ints, not just
num_workspaces.
* src/screen.c (meta_screen_new): add work_area_idle field,
handle it on screen shutdown
* src/common.h (META_PRIORITY_PREFS_NOTIFY,
META_PRIORITY_WORK_AREA_HINT): define some idle priorities
* src/window.c (meta_window_calc_showing): hide windows if
their parent window is minimized
(meta_window_minimize): also queue_calc_showing on all
transients of the window being minimized
* src/place.c (constrain_placement): function to apply
placement-time-only constraints, such as "not off the left of the
screen"
(meta_window_place): put dialogs down a bit over their parent,
not right at the top.
(meta_window_place): when centering a dialog, center it
on the current xinerama screen, rather than the entire
screen.
* src/screen.c (meta_screen_get_current_xinerama): new function,
but not implemented
2002-05-05 01:41:13 -04:00
if ( ! showing )
2011-02-02 19:14:27 -05:00
{
/* When we manage a new window, we normally delay placing it
* until it is is first shown , but if we ' re previewing hidden
* windows we might want to know where they are on the screen ,
* so we should place the window even if we ' re hiding it rather
* than showing it .
*/
2012-02-29 22:15:12 -05:00
if ( ! window - > placed )
2011-02-02 19:14:27 -05:00
meta_window_force_placement ( window ) ;
meta_window_hide ( window ) ;
}
2001-06-06 00:47:37 -04:00
else
Simplify relationship between mapping and visibility
Previously, changes to the visibility of a window could be indicated
by meta_compositor_map_window(), meta_compositor_unminimize_window(),
meta_compositor_set_window_hidden(), etc, with the exact behavior
depending on the 'live_hidden_windows' preference.
Simplify this so that visibility is controlled by:
meta_compositor_show_window()
meta_compositor_hide_window()
With an 'effect' parameter provided to indicate the appropriate
effect (CREATE/UNMINIMIZE/MINIMIZE/DESTROY/NONE.)
The map state of the window is signalled separately by:
meta_compositor_map_window()
meta_compositor_unmap_window()
And is used only to control resource handling.
Other changes:
* The desired effect on show/hide is explicitly stored in
MetaWindow, avoiding the need for the was_minimized flag.
At idle, once we calculate the window state, we pass the
effect to the compositor if it matches the new window
state, and then clear the effect to start over for future
map state changes.
* meta_compositor_switch_workspace() is called before any windows
are hidden or shown, allowing the compositor to avoid hiding
or showing an effect for windows involved in the switch.
http://bugzilla.gnome.org/show_bug.cgi?id=582341
* Handling of post-effect cleanups for MutterWindow are
simplified - instead of trying to do different things based
on the individual needs of different effects, we just wait until
all effects complete and sync the window state to what it
should be.
* On unmap, once we destroy the pixmap, we tell ClutterX11Pixmap
that we've done so, so it can clean up and unbind. (The
unbinding doesn't seem to be working properly because of
ClutterGLXPixmap or video driver issues.)
http://bugzilla.gnome.org/show_bug.cgi?id=587251
2009-06-28 17:10:40 -04:00
meta_window_show ( window ) ;
window - > pending_compositor_effect = META_COMP_EFFECT_NONE ;
2001-06-06 00:47:37 -04:00
}
2001-06-21 23:20:21 -04:00
2002-11-03 15:30:45 -05:00
void
meta_window_calc_showing ( MetaWindow * window )
{
2004-12-19 21:53:08 -05:00
implement_showing ( window , meta_window_should_be_showing ( window ) ) ;
2002-11-03 15:30:45 -05:00
}
2001-06-21 23:20:21 -04:00
2009-09-25 15:47:22 -04:00
static guint queue_later [ NUMBER_OF_QUEUES ] = { 0 , 0 , 0 } ;
Refactor thrice-duplicated queue code in window.c. Closes #376760.
2007-06-10 Thomas Thurman <thomas@thurman.org.uk>
Refactor thrice-duplicated queue code in window.c. Closes #376760.
* src/window.c (meta_window_queue, meta_window_unqueue):
New functions.
* src/window.[ch] (meta_window_unqueue_*, meta_window_queue_*):
Removed functions.
* src/window.c (meta_window_new_with_attrs, meta_window_free,
meta_window_flush_calc_showing, queue_calc_showing_func,
meta_window_minimize, meta_window_unminimize, meta_window_maximize,
meta_window_make_fullscreen, meta_window_shade,
meta_window_unshade, meta_window_move_resize_internal,
window_stick_impl, window_unstick_impl,
meta_window_client_message, process_property_notify): Modified to
use new queueing functions.
* src/window.c (idle_move_resize, idle_update_icon,
idle_calc_showing): update to receive queue number from pointer.
* src/window.h (MetaQueueType): new enum.
* src/window.h (MetaWindow): *_queued replaced with is_in_queue
bitfield.
* src/core.c (meta_core_queue_frame_resize):
* src/display.c (event_callback,
meta_display_queue_retheme_all_windows): Using new queueing functions.
* src/frame.c (meta_window_destroy_frame): Using new queueing functions.
* src/screen.c (queue_resize, meta_screen_resize_func,
queue_windows_showing): Using new queueing functions.
* src/window-props.c (reload_mwm_hints, reload_wm_hints,
reload_transient_for): Using new queueing functions.
* src/workspace.c (meta_workspace_add_window,
meta_workspace_remove_window, meta_workspace_queue_calc_showing,
meta_workspace_invalidate_work_area): Using new queueing functions.
svn path=/trunk/; revision=3236
2007-06-10 21:15:33 -04:00
static GSList * queue_pending [ NUMBER_OF_QUEUES ] = { NULL , NULL , NULL } ;
2001-06-21 23:20:21 -04:00
2001-06-24 04:09:10 -04:00
static int
stackcmp ( gconstpointer a , gconstpointer b )
{
MetaWindow * aw = ( gpointer ) a ;
MetaWindow * bw = ( gpointer ) b ;
if ( aw - > screen ! = bw - > screen )
return 0 ; /* don't care how they sort with respect to each other */
else
return meta_stack_windows_cmp ( aw - > screen - > stack ,
aw , bw ) ;
}
2001-06-21 23:20:21 -04:00
static gboolean
idle_calc_showing ( gpointer data )
{
GSList * tmp ;
2001-10-29 21:00:53 -05:00
GSList * copy ;
2002-11-03 15:30:45 -05:00
GSList * should_show ;
GSList * should_hide ;
GSList * unplaced ;
2003-05-30 16:24:00 -04:00
GSList * displays ;
Refactor thrice-duplicated queue code in window.c. Closes #376760.
2007-06-10 Thomas Thurman <thomas@thurman.org.uk>
Refactor thrice-duplicated queue code in window.c. Closes #376760.
* src/window.c (meta_window_queue, meta_window_unqueue):
New functions.
* src/window.[ch] (meta_window_unqueue_*, meta_window_queue_*):
Removed functions.
* src/window.c (meta_window_new_with_attrs, meta_window_free,
meta_window_flush_calc_showing, queue_calc_showing_func,
meta_window_minimize, meta_window_unminimize, meta_window_maximize,
meta_window_make_fullscreen, meta_window_shade,
meta_window_unshade, meta_window_move_resize_internal,
window_stick_impl, window_unstick_impl,
meta_window_client_message, process_property_notify): Modified to
use new queueing functions.
* src/window.c (idle_move_resize, idle_update_icon,
idle_calc_showing): update to receive queue number from pointer.
* src/window.h (MetaQueueType): new enum.
* src/window.h (MetaWindow): *_queued replaced with is_in_queue
bitfield.
* src/core.c (meta_core_queue_frame_resize):
* src/display.c (event_callback,
meta_display_queue_retheme_all_windows): Using new queueing functions.
* src/frame.c (meta_window_destroy_frame): Using new queueing functions.
* src/screen.c (queue_resize, meta_screen_resize_func,
queue_windows_showing): Using new queueing functions.
* src/window-props.c (reload_mwm_hints, reload_wm_hints,
reload_transient_for): Using new queueing functions.
* src/workspace.c (meta_workspace_add_window,
meta_workspace_remove_window, meta_workspace_queue_calc_showing,
meta_workspace_invalidate_work_area): Using new queueing functions.
svn path=/trunk/; revision=3236
2007-06-10 21:15:33 -04:00
guint queue_index = GPOINTER_TO_INT ( data ) ;
2004-06-24 16:02:46 -04:00
2011-03-29 15:16:23 -04:00
g_return_val_if_fail ( queue_pending [ queue_index ] ! = NULL , FALSE ) ;
2002-02-07 22:34:26 -05:00
meta_topic ( META_DEBUG_WINDOW_STATE ,
" Clearing the calc_showing queue \n " ) ;
2001-10-29 21:00:53 -05:00
/* Work with a copy, for reentrancy. The allowed reentrancy isn't
* complete ; destroying a window while we ' re in here would result in
* badness . But it ' s OK to queue / unqueue calc_showings .
*/
Refactor thrice-duplicated queue code in window.c. Closes #376760.
2007-06-10 Thomas Thurman <thomas@thurman.org.uk>
Refactor thrice-duplicated queue code in window.c. Closes #376760.
* src/window.c (meta_window_queue, meta_window_unqueue):
New functions.
* src/window.[ch] (meta_window_unqueue_*, meta_window_queue_*):
Removed functions.
* src/window.c (meta_window_new_with_attrs, meta_window_free,
meta_window_flush_calc_showing, queue_calc_showing_func,
meta_window_minimize, meta_window_unminimize, meta_window_maximize,
meta_window_make_fullscreen, meta_window_shade,
meta_window_unshade, meta_window_move_resize_internal,
window_stick_impl, window_unstick_impl,
meta_window_client_message, process_property_notify): Modified to
use new queueing functions.
* src/window.c (idle_move_resize, idle_update_icon,
idle_calc_showing): update to receive queue number from pointer.
* src/window.h (MetaQueueType): new enum.
* src/window.h (MetaWindow): *_queued replaced with is_in_queue
bitfield.
* src/core.c (meta_core_queue_frame_resize):
* src/display.c (event_callback,
meta_display_queue_retheme_all_windows): Using new queueing functions.
* src/frame.c (meta_window_destroy_frame): Using new queueing functions.
* src/screen.c (queue_resize, meta_screen_resize_func,
queue_windows_showing): Using new queueing functions.
* src/window-props.c (reload_mwm_hints, reload_wm_hints,
reload_transient_for): Using new queueing functions.
* src/workspace.c (meta_workspace_add_window,
meta_workspace_remove_window, meta_workspace_queue_calc_showing,
meta_workspace_invalidate_work_area): Using new queueing functions.
svn path=/trunk/; revision=3236
2007-06-10 21:15:33 -04:00
copy = g_slist_copy ( queue_pending [ queue_index ] ) ;
g_slist_free ( queue_pending [ queue_index ] ) ;
queue_pending [ queue_index ] = NULL ;
2009-09-25 15:47:22 -04:00
queue_later [ queue_index ] = 0 ;
2002-01-18 22:50:03 -05:00
destroying_windows_disallowed + = 1 ;
2008-12-18 07:41:56 -05:00
2002-11-03 15:30:45 -05:00
/* We map windows from top to bottom and unmap from bottom to
* top , to avoid extra expose events . The exception is
* for unplaced windows , which have to be mapped from bottom to
* top so placement works .
2001-06-24 04:09:10 -04:00
*/
2002-11-03 15:30:45 -05:00
should_show = NULL ;
should_hide = NULL ;
unplaced = NULL ;
2003-05-30 16:24:00 -04:00
displays = NULL ;
2001-10-29 21:00:53 -05:00
tmp = copy ;
2001-06-21 23:20:21 -04:00
while ( tmp ! = NULL )
{
MetaWindow * window ;
window = tmp - > data ;
2002-11-03 15:30:45 -05:00
if ( ! window - > placed )
unplaced = g_slist_prepend ( unplaced , window ) ;
2004-12-19 21:53:08 -05:00
else if ( meta_window_should_be_showing ( window ) )
2002-11-03 15:30:45 -05:00
should_show = g_slist_prepend ( should_show , window ) ;
else
should_hide = g_slist_prepend ( should_hide , window ) ;
2008-12-18 07:41:56 -05:00
2002-11-03 15:30:45 -05:00
tmp = tmp - > next ;
}
/* bottom to top */
unplaced = g_slist_sort ( unplaced , stackcmp ) ;
should_hide = g_slist_sort ( should_hide , stackcmp ) ;
/* top to bottom */
should_show = g_slist_sort ( should_show , stackcmp ) ;
should_show = g_slist_reverse ( should_show ) ;
2006-12-03 15:05:23 -05:00
2002-11-03 15:30:45 -05:00
tmp = unplaced ;
while ( tmp ! = NULL )
{
MetaWindow * window ;
window = tmp - > data ;
2001-06-21 23:20:21 -04:00
meta_window_calc_showing ( window ) ;
2008-12-18 07:41:56 -05:00
2002-11-03 15:30:45 -05:00
tmp = tmp - > next ;
}
2006-12-03 15:05:23 -05:00
tmp = should_show ;
2002-11-03 15:30:45 -05:00
while ( tmp ! = NULL )
{
MetaWindow * window ;
window = tmp - > data ;
2006-12-03 15:05:23 -05:00
implement_showing ( window , TRUE ) ;
2008-12-18 07:41:56 -05:00
2002-11-03 15:30:45 -05:00
tmp = tmp - > next ;
}
2008-12-18 07:41:56 -05:00
2006-12-03 15:05:23 -05:00
tmp = should_hide ;
2002-11-03 15:30:45 -05:00
while ( tmp ! = NULL )
{
MetaWindow * window ;
window = tmp - > data ;
2006-12-03 15:05:23 -05:00
implement_showing ( window , FALSE ) ;
2008-12-18 07:41:56 -05:00
2002-11-03 15:30:45 -05:00
tmp = tmp - > next ;
}
2008-12-18 07:41:56 -05:00
2002-11-03 15:30:45 -05:00
tmp = copy ;
while ( tmp ! = NULL )
{
MetaWindow * window ;
window = tmp - > data ;
2002-01-18 22:50:03 -05:00
/* important to set this here for reentrancy -
* if we queue a window again while it ' s in " copy " ,
* then queue_calc_showing will just return since
Refactor thrice-duplicated queue code in window.c. Closes #376760.
2007-06-10 Thomas Thurman <thomas@thurman.org.uk>
Refactor thrice-duplicated queue code in window.c. Closes #376760.
* src/window.c (meta_window_queue, meta_window_unqueue):
New functions.
* src/window.[ch] (meta_window_unqueue_*, meta_window_queue_*):
Removed functions.
* src/window.c (meta_window_new_with_attrs, meta_window_free,
meta_window_flush_calc_showing, queue_calc_showing_func,
meta_window_minimize, meta_window_unminimize, meta_window_maximize,
meta_window_make_fullscreen, meta_window_shade,
meta_window_unshade, meta_window_move_resize_internal,
window_stick_impl, window_unstick_impl,
meta_window_client_message, process_property_notify): Modified to
use new queueing functions.
* src/window.c (idle_move_resize, idle_update_icon,
idle_calc_showing): update to receive queue number from pointer.
* src/window.h (MetaQueueType): new enum.
* src/window.h (MetaWindow): *_queued replaced with is_in_queue
bitfield.
* src/core.c (meta_core_queue_frame_resize):
* src/display.c (event_callback,
meta_display_queue_retheme_all_windows): Using new queueing functions.
* src/frame.c (meta_window_destroy_frame): Using new queueing functions.
* src/screen.c (queue_resize, meta_screen_resize_func,
queue_windows_showing): Using new queueing functions.
* src/window-props.c (reload_mwm_hints, reload_wm_hints,
reload_transient_for): Using new queueing functions.
* src/workspace.c (meta_workspace_add_window,
meta_workspace_remove_window, meta_workspace_queue_calc_showing,
meta_workspace_invalidate_work_area): Using new queueing functions.
svn path=/trunk/; revision=3236
2007-06-10 21:15:33 -04:00
* we are still in the calc_showing queue
2002-01-18 22:50:03 -05:00
*/
Refactor thrice-duplicated queue code in window.c. Closes #376760.
2007-06-10 Thomas Thurman <thomas@thurman.org.uk>
Refactor thrice-duplicated queue code in window.c. Closes #376760.
* src/window.c (meta_window_queue, meta_window_unqueue):
New functions.
* src/window.[ch] (meta_window_unqueue_*, meta_window_queue_*):
Removed functions.
* src/window.c (meta_window_new_with_attrs, meta_window_free,
meta_window_flush_calc_showing, queue_calc_showing_func,
meta_window_minimize, meta_window_unminimize, meta_window_maximize,
meta_window_make_fullscreen, meta_window_shade,
meta_window_unshade, meta_window_move_resize_internal,
window_stick_impl, window_unstick_impl,
meta_window_client_message, process_property_notify): Modified to
use new queueing functions.
* src/window.c (idle_move_resize, idle_update_icon,
idle_calc_showing): update to receive queue number from pointer.
* src/window.h (MetaQueueType): new enum.
* src/window.h (MetaWindow): *_queued replaced with is_in_queue
bitfield.
* src/core.c (meta_core_queue_frame_resize):
* src/display.c (event_callback,
meta_display_queue_retheme_all_windows): Using new queueing functions.
* src/frame.c (meta_window_destroy_frame): Using new queueing functions.
* src/screen.c (queue_resize, meta_screen_resize_func,
queue_windows_showing): Using new queueing functions.
* src/window-props.c (reload_mwm_hints, reload_wm_hints,
reload_transient_for): Using new queueing functions.
* src/workspace.c (meta_workspace_add_window,
meta_workspace_remove_window, meta_workspace_queue_calc_showing,
meta_workspace_invalidate_work_area): Using new queueing functions.
svn path=/trunk/; revision=3236
2007-06-10 21:15:33 -04:00
window - > is_in_queues & = ~ META_QUEUE_CALC_SHOWING ;
2008-12-18 07:41:56 -05:00
2001-06-21 23:20:21 -04:00
tmp = tmp - > next ;
}
2003-05-30 16:24:00 -04:00
2011-06-24 12:14:15 -04:00
if ( meta_prefs_get_focus_mode ( ) ! = G_DESKTOP_FOCUS_MODE_CLICK )
2005-02-21 21:11:25 -05:00
{
/* When display->mouse_mode is false, we want to ignore
* EnterNotify events unless they come from mouse motion . To do
* that , we set a sentinel property on the root window if we ' re
* not in mouse_mode .
*/
tmp = should_show ;
while ( tmp ! = NULL )
{
MetaWindow * window = tmp - > data ;
2008-12-18 07:41:56 -05:00
2005-02-21 21:11:25 -05:00
if ( ! window - > display - > mouse_mode )
meta_display_increment_focus_sentinel ( window - > display ) ;
tmp = tmp - > next ;
}
}
2001-10-29 21:00:53 -05:00
g_slist_free ( copy ) ;
2002-01-18 22:50:03 -05:00
2002-11-03 15:30:45 -05:00
g_slist_free ( unplaced ) ;
g_slist_free ( should_show ) ;
g_slist_free ( should_hide ) ;
2003-05-30 16:24:00 -04:00
g_slist_free ( displays ) ;
2008-12-18 07:41:56 -05:00
2002-01-18 22:50:03 -05:00
destroying_windows_disallowed - = 1 ;
2008-12-18 07:41:56 -05:00
2001-06-21 23:20:21 -04:00
return FALSE ;
}
2009-02-10 23:34:17 -05:00
# ifdef WITH_VERBOSE_MODE
Refactor thrice-duplicated queue code in window.c. Closes #376760.
2007-06-10 Thomas Thurman <thomas@thurman.org.uk>
Refactor thrice-duplicated queue code in window.c. Closes #376760.
* src/window.c (meta_window_queue, meta_window_unqueue):
New functions.
* src/window.[ch] (meta_window_unqueue_*, meta_window_queue_*):
Removed functions.
* src/window.c (meta_window_new_with_attrs, meta_window_free,
meta_window_flush_calc_showing, queue_calc_showing_func,
meta_window_minimize, meta_window_unminimize, meta_window_maximize,
meta_window_make_fullscreen, meta_window_shade,
meta_window_unshade, meta_window_move_resize_internal,
window_stick_impl, window_unstick_impl,
meta_window_client_message, process_property_notify): Modified to
use new queueing functions.
* src/window.c (idle_move_resize, idle_update_icon,
idle_calc_showing): update to receive queue number from pointer.
* src/window.h (MetaQueueType): new enum.
* src/window.h (MetaWindow): *_queued replaced with is_in_queue
bitfield.
* src/core.c (meta_core_queue_frame_resize):
* src/display.c (event_callback,
meta_display_queue_retheme_all_windows): Using new queueing functions.
* src/frame.c (meta_window_destroy_frame): Using new queueing functions.
* src/screen.c (queue_resize, meta_screen_resize_func,
queue_windows_showing): Using new queueing functions.
* src/window-props.c (reload_mwm_hints, reload_wm_hints,
reload_transient_for): Using new queueing functions.
* src/workspace.c (meta_workspace_add_window,
meta_workspace_remove_window, meta_workspace_queue_calc_showing,
meta_workspace_invalidate_work_area): Using new queueing functions.
svn path=/trunk/; revision=3236
2007-06-10 21:15:33 -04:00
static const gchar * meta_window_queue_names [ NUMBER_OF_QUEUES ] =
{ " calc_showing " , " move_resize " , " update_icon " } ;
2009-02-10 23:34:17 -05:00
# endif
Refactor thrice-duplicated queue code in window.c. Closes #376760.
2007-06-10 Thomas Thurman <thomas@thurman.org.uk>
Refactor thrice-duplicated queue code in window.c. Closes #376760.
* src/window.c (meta_window_queue, meta_window_unqueue):
New functions.
* src/window.[ch] (meta_window_unqueue_*, meta_window_queue_*):
Removed functions.
* src/window.c (meta_window_new_with_attrs, meta_window_free,
meta_window_flush_calc_showing, queue_calc_showing_func,
meta_window_minimize, meta_window_unminimize, meta_window_maximize,
meta_window_make_fullscreen, meta_window_shade,
meta_window_unshade, meta_window_move_resize_internal,
window_stick_impl, window_unstick_impl,
meta_window_client_message, process_property_notify): Modified to
use new queueing functions.
* src/window.c (idle_move_resize, idle_update_icon,
idle_calc_showing): update to receive queue number from pointer.
* src/window.h (MetaQueueType): new enum.
* src/window.h (MetaWindow): *_queued replaced with is_in_queue
bitfield.
* src/core.c (meta_core_queue_frame_resize):
* src/display.c (event_callback,
meta_display_queue_retheme_all_windows): Using new queueing functions.
* src/frame.c (meta_window_destroy_frame): Using new queueing functions.
* src/screen.c (queue_resize, meta_screen_resize_func,
queue_windows_showing): Using new queueing functions.
* src/window-props.c (reload_mwm_hints, reload_wm_hints,
reload_transient_for): Using new queueing functions.
* src/workspace.c (meta_workspace_add_window,
meta_workspace_remove_window, meta_workspace_queue_calc_showing,
meta_workspace_invalidate_work_area): Using new queueing functions.
svn path=/trunk/; revision=3236
2007-06-10 21:15:33 -04:00
2006-01-20 17:03:56 -05:00
static void
Refactor thrice-duplicated queue code in window.c. Closes #376760.
2007-06-10 Thomas Thurman <thomas@thurman.org.uk>
Refactor thrice-duplicated queue code in window.c. Closes #376760.
* src/window.c (meta_window_queue, meta_window_unqueue):
New functions.
* src/window.[ch] (meta_window_unqueue_*, meta_window_queue_*):
Removed functions.
* src/window.c (meta_window_new_with_attrs, meta_window_free,
meta_window_flush_calc_showing, queue_calc_showing_func,
meta_window_minimize, meta_window_unminimize, meta_window_maximize,
meta_window_make_fullscreen, meta_window_shade,
meta_window_unshade, meta_window_move_resize_internal,
window_stick_impl, window_unstick_impl,
meta_window_client_message, process_property_notify): Modified to
use new queueing functions.
* src/window.c (idle_move_resize, idle_update_icon,
idle_calc_showing): update to receive queue number from pointer.
* src/window.h (MetaQueueType): new enum.
* src/window.h (MetaWindow): *_queued replaced with is_in_queue
bitfield.
* src/core.c (meta_core_queue_frame_resize):
* src/display.c (event_callback,
meta_display_queue_retheme_all_windows): Using new queueing functions.
* src/frame.c (meta_window_destroy_frame): Using new queueing functions.
* src/screen.c (queue_resize, meta_screen_resize_func,
queue_windows_showing): Using new queueing functions.
* src/window-props.c (reload_mwm_hints, reload_wm_hints,
reload_transient_for): Using new queueing functions.
* src/workspace.c (meta_workspace_add_window,
meta_workspace_remove_window, meta_workspace_queue_calc_showing,
meta_workspace_invalidate_work_area): Using new queueing functions.
svn path=/trunk/; revision=3236
2007-06-10 21:15:33 -04:00
meta_window_unqueue ( MetaWindow * window , guint queuebits )
2001-06-21 23:20:21 -04:00
{
Refactor thrice-duplicated queue code in window.c. Closes #376760.
2007-06-10 Thomas Thurman <thomas@thurman.org.uk>
Refactor thrice-duplicated queue code in window.c. Closes #376760.
* src/window.c (meta_window_queue, meta_window_unqueue):
New functions.
* src/window.[ch] (meta_window_unqueue_*, meta_window_queue_*):
Removed functions.
* src/window.c (meta_window_new_with_attrs, meta_window_free,
meta_window_flush_calc_showing, queue_calc_showing_func,
meta_window_minimize, meta_window_unminimize, meta_window_maximize,
meta_window_make_fullscreen, meta_window_shade,
meta_window_unshade, meta_window_move_resize_internal,
window_stick_impl, window_unstick_impl,
meta_window_client_message, process_property_notify): Modified to
use new queueing functions.
* src/window.c (idle_move_resize, idle_update_icon,
idle_calc_showing): update to receive queue number from pointer.
* src/window.h (MetaQueueType): new enum.
* src/window.h (MetaWindow): *_queued replaced with is_in_queue
bitfield.
* src/core.c (meta_core_queue_frame_resize):
* src/display.c (event_callback,
meta_display_queue_retheme_all_windows): Using new queueing functions.
* src/frame.c (meta_window_destroy_frame): Using new queueing functions.
* src/screen.c (queue_resize, meta_screen_resize_func,
queue_windows_showing): Using new queueing functions.
* src/window-props.c (reload_mwm_hints, reload_wm_hints,
reload_transient_for): Using new queueing functions.
* src/workspace.c (meta_workspace_add_window,
meta_workspace_remove_window, meta_workspace_queue_calc_showing,
meta_workspace_invalidate_work_area): Using new queueing functions.
svn path=/trunk/; revision=3236
2007-06-10 21:15:33 -04:00
gint queuenum ;
2001-10-29 21:00:53 -05:00
Refactor thrice-duplicated queue code in window.c. Closes #376760.
2007-06-10 Thomas Thurman <thomas@thurman.org.uk>
Refactor thrice-duplicated queue code in window.c. Closes #376760.
* src/window.c (meta_window_queue, meta_window_unqueue):
New functions.
* src/window.[ch] (meta_window_unqueue_*, meta_window_queue_*):
Removed functions.
* src/window.c (meta_window_new_with_attrs, meta_window_free,
meta_window_flush_calc_showing, queue_calc_showing_func,
meta_window_minimize, meta_window_unminimize, meta_window_maximize,
meta_window_make_fullscreen, meta_window_shade,
meta_window_unshade, meta_window_move_resize_internal,
window_stick_impl, window_unstick_impl,
meta_window_client_message, process_property_notify): Modified to
use new queueing functions.
* src/window.c (idle_move_resize, idle_update_icon,
idle_calc_showing): update to receive queue number from pointer.
* src/window.h (MetaQueueType): new enum.
* src/window.h (MetaWindow): *_queued replaced with is_in_queue
bitfield.
* src/core.c (meta_core_queue_frame_resize):
* src/display.c (event_callback,
meta_display_queue_retheme_all_windows): Using new queueing functions.
* src/frame.c (meta_window_destroy_frame): Using new queueing functions.
* src/screen.c (queue_resize, meta_screen_resize_func,
queue_windows_showing): Using new queueing functions.
* src/window-props.c (reload_mwm_hints, reload_wm_hints,
reload_transient_for): Using new queueing functions.
* src/workspace.c (meta_workspace_add_window,
meta_workspace_remove_window, meta_workspace_queue_calc_showing,
meta_workspace_invalidate_work_area): Using new queueing functions.
svn path=/trunk/; revision=3236
2007-06-10 21:15:33 -04:00
for ( queuenum = 0 ; queuenum < NUMBER_OF_QUEUES ; queuenum + + )
2001-06-21 23:20:21 -04:00
{
Refactor thrice-duplicated queue code in window.c. Closes #376760.
2007-06-10 Thomas Thurman <thomas@thurman.org.uk>
Refactor thrice-duplicated queue code in window.c. Closes #376760.
* src/window.c (meta_window_queue, meta_window_unqueue):
New functions.
* src/window.[ch] (meta_window_unqueue_*, meta_window_queue_*):
Removed functions.
* src/window.c (meta_window_new_with_attrs, meta_window_free,
meta_window_flush_calc_showing, queue_calc_showing_func,
meta_window_minimize, meta_window_unminimize, meta_window_maximize,
meta_window_make_fullscreen, meta_window_shade,
meta_window_unshade, meta_window_move_resize_internal,
window_stick_impl, window_unstick_impl,
meta_window_client_message, process_property_notify): Modified to
use new queueing functions.
* src/window.c (idle_move_resize, idle_update_icon,
idle_calc_showing): update to receive queue number from pointer.
* src/window.h (MetaQueueType): new enum.
* src/window.h (MetaWindow): *_queued replaced with is_in_queue
bitfield.
* src/core.c (meta_core_queue_frame_resize):
* src/display.c (event_callback,
meta_display_queue_retheme_all_windows): Using new queueing functions.
* src/frame.c (meta_window_destroy_frame): Using new queueing functions.
* src/screen.c (queue_resize, meta_screen_resize_func,
queue_windows_showing): Using new queueing functions.
* src/window-props.c (reload_mwm_hints, reload_wm_hints,
reload_transient_for): Using new queueing functions.
* src/workspace.c (meta_workspace_add_window,
meta_workspace_remove_window, meta_workspace_queue_calc_showing,
meta_workspace_invalidate_work_area): Using new queueing functions.
svn path=/trunk/; revision=3236
2007-06-10 21:15:33 -04:00
if ( ( queuebits & 1 < < queuenum ) /* they have asked to unqueue */
& &
( window - > is_in_queues & 1 < < queuenum ) ) /* it's in the queue */
{
meta_topic ( META_DEBUG_WINDOW_STATE ,
" Removing %s from the %s queue \n " ,
window - > desc ,
meta_window_queue_names [ queuenum ] ) ;
/* Note that window may not actually be in the queue
* because it may have been in " copy " inside the idle handler
2008-12-18 07:41:56 -05:00
*/
Refactor thrice-duplicated queue code in window.c. Closes #376760.
2007-06-10 Thomas Thurman <thomas@thurman.org.uk>
Refactor thrice-duplicated queue code in window.c. Closes #376760.
* src/window.c (meta_window_queue, meta_window_unqueue):
New functions.
* src/window.[ch] (meta_window_unqueue_*, meta_window_queue_*):
Removed functions.
* src/window.c (meta_window_new_with_attrs, meta_window_free,
meta_window_flush_calc_showing, queue_calc_showing_func,
meta_window_minimize, meta_window_unminimize, meta_window_maximize,
meta_window_make_fullscreen, meta_window_shade,
meta_window_unshade, meta_window_move_resize_internal,
window_stick_impl, window_unstick_impl,
meta_window_client_message, process_property_notify): Modified to
use new queueing functions.
* src/window.c (idle_move_resize, idle_update_icon,
idle_calc_showing): update to receive queue number from pointer.
* src/window.h (MetaQueueType): new enum.
* src/window.h (MetaWindow): *_queued replaced with is_in_queue
bitfield.
* src/core.c (meta_core_queue_frame_resize):
* src/display.c (event_callback,
meta_display_queue_retheme_all_windows): Using new queueing functions.
* src/frame.c (meta_window_destroy_frame): Using new queueing functions.
* src/screen.c (queue_resize, meta_screen_resize_func,
queue_windows_showing): Using new queueing functions.
* src/window-props.c (reload_mwm_hints, reload_wm_hints,
reload_transient_for): Using new queueing functions.
* src/workspace.c (meta_workspace_add_window,
meta_workspace_remove_window, meta_workspace_queue_calc_showing,
meta_workspace_invalidate_work_area): Using new queueing functions.
svn path=/trunk/; revision=3236
2007-06-10 21:15:33 -04:00
queue_pending [ queuenum ] = g_slist_remove ( queue_pending [ queuenum ] , window ) ;
window - > is_in_queues & = ~ ( 1 < < queuenum ) ;
/* Okay, so maybe we've used up all the entries in the queue.
* In that case , we should kill the function that deals with
* the queue , because there ' s nothing left for it to do .
*/
2009-09-25 15:47:22 -04:00
if ( queue_pending [ queuenum ] = = NULL & & queue_later [ queuenum ] ! = 0 )
Refactor thrice-duplicated queue code in window.c. Closes #376760.
2007-06-10 Thomas Thurman <thomas@thurman.org.uk>
Refactor thrice-duplicated queue code in window.c. Closes #376760.
* src/window.c (meta_window_queue, meta_window_unqueue):
New functions.
* src/window.[ch] (meta_window_unqueue_*, meta_window_queue_*):
Removed functions.
* src/window.c (meta_window_new_with_attrs, meta_window_free,
meta_window_flush_calc_showing, queue_calc_showing_func,
meta_window_minimize, meta_window_unminimize, meta_window_maximize,
meta_window_make_fullscreen, meta_window_shade,
meta_window_unshade, meta_window_move_resize_internal,
window_stick_impl, window_unstick_impl,
meta_window_client_message, process_property_notify): Modified to
use new queueing functions.
* src/window.c (idle_move_resize, idle_update_icon,
idle_calc_showing): update to receive queue number from pointer.
* src/window.h (MetaQueueType): new enum.
* src/window.h (MetaWindow): *_queued replaced with is_in_queue
bitfield.
* src/core.c (meta_core_queue_frame_resize):
* src/display.c (event_callback,
meta_display_queue_retheme_all_windows): Using new queueing functions.
* src/frame.c (meta_window_destroy_frame): Using new queueing functions.
* src/screen.c (queue_resize, meta_screen_resize_func,
queue_windows_showing): Using new queueing functions.
* src/window-props.c (reload_mwm_hints, reload_wm_hints,
reload_transient_for): Using new queueing functions.
* src/workspace.c (meta_workspace_add_window,
meta_workspace_remove_window, meta_workspace_queue_calc_showing,
meta_workspace_invalidate_work_area): Using new queueing functions.
svn path=/trunk/; revision=3236
2007-06-10 21:15:33 -04:00
{
2009-09-25 15:47:22 -04:00
meta_later_remove ( queue_later [ queuenum ] ) ;
queue_later [ queuenum ] = 0 ;
Refactor thrice-duplicated queue code in window.c. Closes #376760.
2007-06-10 Thomas Thurman <thomas@thurman.org.uk>
Refactor thrice-duplicated queue code in window.c. Closes #376760.
* src/window.c (meta_window_queue, meta_window_unqueue):
New functions.
* src/window.[ch] (meta_window_unqueue_*, meta_window_queue_*):
Removed functions.
* src/window.c (meta_window_new_with_attrs, meta_window_free,
meta_window_flush_calc_showing, queue_calc_showing_func,
meta_window_minimize, meta_window_unminimize, meta_window_maximize,
meta_window_make_fullscreen, meta_window_shade,
meta_window_unshade, meta_window_move_resize_internal,
window_stick_impl, window_unstick_impl,
meta_window_client_message, process_property_notify): Modified to
use new queueing functions.
* src/window.c (idle_move_resize, idle_update_icon,
idle_calc_showing): update to receive queue number from pointer.
* src/window.h (MetaQueueType): new enum.
* src/window.h (MetaWindow): *_queued replaced with is_in_queue
bitfield.
* src/core.c (meta_core_queue_frame_resize):
* src/display.c (event_callback,
meta_display_queue_retheme_all_windows): Using new queueing functions.
* src/frame.c (meta_window_destroy_frame): Using new queueing functions.
* src/screen.c (queue_resize, meta_screen_resize_func,
queue_windows_showing): Using new queueing functions.
* src/window-props.c (reload_mwm_hints, reload_wm_hints,
reload_transient_for): Using new queueing functions.
* src/workspace.c (meta_workspace_add_window,
meta_workspace_remove_window, meta_workspace_queue_calc_showing,
meta_workspace_invalidate_work_area): Using new queueing functions.
svn path=/trunk/; revision=3236
2007-06-10 21:15:33 -04:00
}
}
2001-06-21 23:20:21 -04:00
}
}
2006-01-20 17:03:56 -05:00
static void
2001-10-29 02:00:46 -05:00
meta_window_flush_calc_showing ( MetaWindow * window )
{
Refactor thrice-duplicated queue code in window.c. Closes #376760.
2007-06-10 Thomas Thurman <thomas@thurman.org.uk>
Refactor thrice-duplicated queue code in window.c. Closes #376760.
* src/window.c (meta_window_queue, meta_window_unqueue):
New functions.
* src/window.[ch] (meta_window_unqueue_*, meta_window_queue_*):
Removed functions.
* src/window.c (meta_window_new_with_attrs, meta_window_free,
meta_window_flush_calc_showing, queue_calc_showing_func,
meta_window_minimize, meta_window_unminimize, meta_window_maximize,
meta_window_make_fullscreen, meta_window_shade,
meta_window_unshade, meta_window_move_resize_internal,
window_stick_impl, window_unstick_impl,
meta_window_client_message, process_property_notify): Modified to
use new queueing functions.
* src/window.c (idle_move_resize, idle_update_icon,
idle_calc_showing): update to receive queue number from pointer.
* src/window.h (MetaQueueType): new enum.
* src/window.h (MetaWindow): *_queued replaced with is_in_queue
bitfield.
* src/core.c (meta_core_queue_frame_resize):
* src/display.c (event_callback,
meta_display_queue_retheme_all_windows): Using new queueing functions.
* src/frame.c (meta_window_destroy_frame): Using new queueing functions.
* src/screen.c (queue_resize, meta_screen_resize_func,
queue_windows_showing): Using new queueing functions.
* src/window-props.c (reload_mwm_hints, reload_wm_hints,
reload_transient_for): Using new queueing functions.
* src/workspace.c (meta_workspace_add_window,
meta_workspace_remove_window, meta_workspace_queue_calc_showing,
meta_workspace_invalidate_work_area): Using new queueing functions.
svn path=/trunk/; revision=3236
2007-06-10 21:15:33 -04:00
if ( window - > is_in_queues & META_QUEUE_CALC_SHOWING )
2001-10-29 02:00:46 -05:00
{
Refactor thrice-duplicated queue code in window.c. Closes #376760.
2007-06-10 Thomas Thurman <thomas@thurman.org.uk>
Refactor thrice-duplicated queue code in window.c. Closes #376760.
* src/window.c (meta_window_queue, meta_window_unqueue):
New functions.
* src/window.[ch] (meta_window_unqueue_*, meta_window_queue_*):
Removed functions.
* src/window.c (meta_window_new_with_attrs, meta_window_free,
meta_window_flush_calc_showing, queue_calc_showing_func,
meta_window_minimize, meta_window_unminimize, meta_window_maximize,
meta_window_make_fullscreen, meta_window_shade,
meta_window_unshade, meta_window_move_resize_internal,
window_stick_impl, window_unstick_impl,
meta_window_client_message, process_property_notify): Modified to
use new queueing functions.
* src/window.c (idle_move_resize, idle_update_icon,
idle_calc_showing): update to receive queue number from pointer.
* src/window.h (MetaQueueType): new enum.
* src/window.h (MetaWindow): *_queued replaced with is_in_queue
bitfield.
* src/core.c (meta_core_queue_frame_resize):
* src/display.c (event_callback,
meta_display_queue_retheme_all_windows): Using new queueing functions.
* src/frame.c (meta_window_destroy_frame): Using new queueing functions.
* src/screen.c (queue_resize, meta_screen_resize_func,
queue_windows_showing): Using new queueing functions.
* src/window-props.c (reload_mwm_hints, reload_wm_hints,
reload_transient_for): Using new queueing functions.
* src/workspace.c (meta_workspace_add_window,
meta_workspace_remove_window, meta_workspace_queue_calc_showing,
meta_workspace_invalidate_work_area): Using new queueing functions.
svn path=/trunk/; revision=3236
2007-06-10 21:15:33 -04:00
meta_window_unqueue ( window , META_QUEUE_CALC_SHOWING ) ;
2001-10-29 02:00:46 -05:00
meta_window_calc_showing ( window ) ;
}
}
2001-06-06 00:47:37 -04:00
void
Refactor thrice-duplicated queue code in window.c. Closes #376760.
2007-06-10 Thomas Thurman <thomas@thurman.org.uk>
Refactor thrice-duplicated queue code in window.c. Closes #376760.
* src/window.c (meta_window_queue, meta_window_unqueue):
New functions.
* src/window.[ch] (meta_window_unqueue_*, meta_window_queue_*):
Removed functions.
* src/window.c (meta_window_new_with_attrs, meta_window_free,
meta_window_flush_calc_showing, queue_calc_showing_func,
meta_window_minimize, meta_window_unminimize, meta_window_maximize,
meta_window_make_fullscreen, meta_window_shade,
meta_window_unshade, meta_window_move_resize_internal,
window_stick_impl, window_unstick_impl,
meta_window_client_message, process_property_notify): Modified to
use new queueing functions.
* src/window.c (idle_move_resize, idle_update_icon,
idle_calc_showing): update to receive queue number from pointer.
* src/window.h (MetaQueueType): new enum.
* src/window.h (MetaWindow): *_queued replaced with is_in_queue
bitfield.
* src/core.c (meta_core_queue_frame_resize):
* src/display.c (event_callback,
meta_display_queue_retheme_all_windows): Using new queueing functions.
* src/frame.c (meta_window_destroy_frame): Using new queueing functions.
* src/screen.c (queue_resize, meta_screen_resize_func,
queue_windows_showing): Using new queueing functions.
* src/window-props.c (reload_mwm_hints, reload_wm_hints,
reload_transient_for): Using new queueing functions.
* src/workspace.c (meta_workspace_add_window,
meta_workspace_remove_window, meta_workspace_queue_calc_showing,
meta_workspace_invalidate_work_area): Using new queueing functions.
svn path=/trunk/; revision=3236
2007-06-10 21:15:33 -04:00
meta_window_queue ( MetaWindow * window , guint queuebits )
2001-06-06 00:47:37 -04:00
{
Refactor thrice-duplicated queue code in window.c. Closes #376760.
2007-06-10 Thomas Thurman <thomas@thurman.org.uk>
Refactor thrice-duplicated queue code in window.c. Closes #376760.
* src/window.c (meta_window_queue, meta_window_unqueue):
New functions.
* src/window.[ch] (meta_window_unqueue_*, meta_window_queue_*):
Removed functions.
* src/window.c (meta_window_new_with_attrs, meta_window_free,
meta_window_flush_calc_showing, queue_calc_showing_func,
meta_window_minimize, meta_window_unminimize, meta_window_maximize,
meta_window_make_fullscreen, meta_window_shade,
meta_window_unshade, meta_window_move_resize_internal,
window_stick_impl, window_unstick_impl,
meta_window_client_message, process_property_notify): Modified to
use new queueing functions.
* src/window.c (idle_move_resize, idle_update_icon,
idle_calc_showing): update to receive queue number from pointer.
* src/window.h (MetaQueueType): new enum.
* src/window.h (MetaWindow): *_queued replaced with is_in_queue
bitfield.
* src/core.c (meta_core_queue_frame_resize):
* src/display.c (event_callback,
meta_display_queue_retheme_all_windows): Using new queueing functions.
* src/frame.c (meta_window_destroy_frame): Using new queueing functions.
* src/screen.c (queue_resize, meta_screen_resize_func,
queue_windows_showing): Using new queueing functions.
* src/window-props.c (reload_mwm_hints, reload_wm_hints,
reload_transient_for): Using new queueing functions.
* src/workspace.c (meta_workspace_add_window,
meta_workspace_remove_window, meta_workspace_queue_calc_showing,
meta_workspace_invalidate_work_area): Using new queueing functions.
svn path=/trunk/; revision=3236
2007-06-10 21:15:33 -04:00
guint queuenum ;
2001-06-21 23:20:21 -04:00
2009-06-15 18:25:07 -04:00
/* Easier to debug by checking here rather than in the idle */
g_return_if_fail ( ! window - > override_redirect | | ( queuebits & META_QUEUE_MOVE_RESIZE ) = = 0 ) ;
Refactor thrice-duplicated queue code in window.c. Closes #376760.
2007-06-10 Thomas Thurman <thomas@thurman.org.uk>
Refactor thrice-duplicated queue code in window.c. Closes #376760.
* src/window.c (meta_window_queue, meta_window_unqueue):
New functions.
* src/window.[ch] (meta_window_unqueue_*, meta_window_queue_*):
Removed functions.
* src/window.c (meta_window_new_with_attrs, meta_window_free,
meta_window_flush_calc_showing, queue_calc_showing_func,
meta_window_minimize, meta_window_unminimize, meta_window_maximize,
meta_window_make_fullscreen, meta_window_shade,
meta_window_unshade, meta_window_move_resize_internal,
window_stick_impl, window_unstick_impl,
meta_window_client_message, process_property_notify): Modified to
use new queueing functions.
* src/window.c (idle_move_resize, idle_update_icon,
idle_calc_showing): update to receive queue number from pointer.
* src/window.h (MetaQueueType): new enum.
* src/window.h (MetaWindow): *_queued replaced with is_in_queue
bitfield.
* src/core.c (meta_core_queue_frame_resize):
* src/display.c (event_callback,
meta_display_queue_retheme_all_windows): Using new queueing functions.
* src/frame.c (meta_window_destroy_frame): Using new queueing functions.
* src/screen.c (queue_resize, meta_screen_resize_func,
queue_windows_showing): Using new queueing functions.
* src/window-props.c (reload_mwm_hints, reload_wm_hints,
reload_transient_for): Using new queueing functions.
* src/workspace.c (meta_workspace_add_window,
meta_workspace_remove_window, meta_workspace_queue_calc_showing,
meta_workspace_invalidate_work_area): Using new queueing functions.
svn path=/trunk/; revision=3236
2007-06-10 21:15:33 -04:00
for ( queuenum = 0 ; queuenum < NUMBER_OF_QUEUES ; queuenum + + )
{
if ( queuebits & 1 < < queuenum )
{
/* Data which varies between queues.
* Yes , these do look a lot like associative arrays :
* I seem to be turning into a Perl programmer .
*/
2001-06-21 23:20:21 -04:00
2009-09-25 15:47:22 -04:00
const MetaLaterType window_queue_later_when [ NUMBER_OF_QUEUES ] =
Refactor thrice-duplicated queue code in window.c. Closes #376760.
2007-06-10 Thomas Thurman <thomas@thurman.org.uk>
Refactor thrice-duplicated queue code in window.c. Closes #376760.
* src/window.c (meta_window_queue, meta_window_unqueue):
New functions.
* src/window.[ch] (meta_window_unqueue_*, meta_window_queue_*):
Removed functions.
* src/window.c (meta_window_new_with_attrs, meta_window_free,
meta_window_flush_calc_showing, queue_calc_showing_func,
meta_window_minimize, meta_window_unminimize, meta_window_maximize,
meta_window_make_fullscreen, meta_window_shade,
meta_window_unshade, meta_window_move_resize_internal,
window_stick_impl, window_unstick_impl,
meta_window_client_message, process_property_notify): Modified to
use new queueing functions.
* src/window.c (idle_move_resize, idle_update_icon,
idle_calc_showing): update to receive queue number from pointer.
* src/window.h (MetaQueueType): new enum.
* src/window.h (MetaWindow): *_queued replaced with is_in_queue
bitfield.
* src/core.c (meta_core_queue_frame_resize):
* src/display.c (event_callback,
meta_display_queue_retheme_all_windows): Using new queueing functions.
* src/frame.c (meta_window_destroy_frame): Using new queueing functions.
* src/screen.c (queue_resize, meta_screen_resize_func,
queue_windows_showing): Using new queueing functions.
* src/window-props.c (reload_mwm_hints, reload_wm_hints,
reload_transient_for): Using new queueing functions.
* src/workspace.c (meta_workspace_add_window,
meta_workspace_remove_window, meta_workspace_queue_calc_showing,
meta_workspace_invalidate_work_area): Using new queueing functions.
svn path=/trunk/; revision=3236
2007-06-10 21:15:33 -04:00
{
2013-03-14 16:55:49 -04:00
META_LATER_CALC_SHOWING , /* CALC_SHOWING */
2009-09-25 15:47:22 -04:00
META_LATER_RESIZE , /* MOVE_RESIZE */
META_LATER_BEFORE_REDRAW /* UPDATE_ICON */
Refactor thrice-duplicated queue code in window.c. Closes #376760.
2007-06-10 Thomas Thurman <thomas@thurman.org.uk>
Refactor thrice-duplicated queue code in window.c. Closes #376760.
* src/window.c (meta_window_queue, meta_window_unqueue):
New functions.
* src/window.[ch] (meta_window_unqueue_*, meta_window_queue_*):
Removed functions.
* src/window.c (meta_window_new_with_attrs, meta_window_free,
meta_window_flush_calc_showing, queue_calc_showing_func,
meta_window_minimize, meta_window_unminimize, meta_window_maximize,
meta_window_make_fullscreen, meta_window_shade,
meta_window_unshade, meta_window_move_resize_internal,
window_stick_impl, window_unstick_impl,
meta_window_client_message, process_property_notify): Modified to
use new queueing functions.
* src/window.c (idle_move_resize, idle_update_icon,
idle_calc_showing): update to receive queue number from pointer.
* src/window.h (MetaQueueType): new enum.
* src/window.h (MetaWindow): *_queued replaced with is_in_queue
bitfield.
* src/core.c (meta_core_queue_frame_resize):
* src/display.c (event_callback,
meta_display_queue_retheme_all_windows): Using new queueing functions.
* src/frame.c (meta_window_destroy_frame): Using new queueing functions.
* src/screen.c (queue_resize, meta_screen_resize_func,
queue_windows_showing): Using new queueing functions.
* src/window-props.c (reload_mwm_hints, reload_wm_hints,
reload_transient_for): Using new queueing functions.
* src/workspace.c (meta_workspace_add_window,
meta_workspace_remove_window, meta_workspace_queue_calc_showing,
meta_workspace_invalidate_work_area): Using new queueing functions.
svn path=/trunk/; revision=3236
2007-06-10 21:15:33 -04:00
} ;
2009-09-25 15:47:22 -04:00
const GSourceFunc window_queue_later_handler [ NUMBER_OF_QUEUES ] =
Refactor thrice-duplicated queue code in window.c. Closes #376760.
2007-06-10 Thomas Thurman <thomas@thurman.org.uk>
Refactor thrice-duplicated queue code in window.c. Closes #376760.
* src/window.c (meta_window_queue, meta_window_unqueue):
New functions.
* src/window.[ch] (meta_window_unqueue_*, meta_window_queue_*):
Removed functions.
* src/window.c (meta_window_new_with_attrs, meta_window_free,
meta_window_flush_calc_showing, queue_calc_showing_func,
meta_window_minimize, meta_window_unminimize, meta_window_maximize,
meta_window_make_fullscreen, meta_window_shade,
meta_window_unshade, meta_window_move_resize_internal,
window_stick_impl, window_unstick_impl,
meta_window_client_message, process_property_notify): Modified to
use new queueing functions.
* src/window.c (idle_move_resize, idle_update_icon,
idle_calc_showing): update to receive queue number from pointer.
* src/window.h (MetaQueueType): new enum.
* src/window.h (MetaWindow): *_queued replaced with is_in_queue
bitfield.
* src/core.c (meta_core_queue_frame_resize):
* src/display.c (event_callback,
meta_display_queue_retheme_all_windows): Using new queueing functions.
* src/frame.c (meta_window_destroy_frame): Using new queueing functions.
* src/screen.c (queue_resize, meta_screen_resize_func,
queue_windows_showing): Using new queueing functions.
* src/window-props.c (reload_mwm_hints, reload_wm_hints,
reload_transient_for): Using new queueing functions.
* src/workspace.c (meta_workspace_add_window,
meta_workspace_remove_window, meta_workspace_queue_calc_showing,
meta_workspace_invalidate_work_area): Using new queueing functions.
svn path=/trunk/; revision=3236
2007-06-10 21:15:33 -04:00
{
idle_calc_showing ,
idle_move_resize ,
idle_update_icon ,
} ;
/* If we're about to drop the window, there's no point in putting
* it on a queue .
*/
if ( window - > unmanaging )
break ;
/* If the window already claims to be in that queue, there's no
* point putting it in the queue .
*/
if ( window - > is_in_queues & 1 < < queuenum )
break ;
meta_topic ( META_DEBUG_WINDOW_STATE ,
" Putting %s in the %s queue \n " ,
window - > desc ,
meta_window_queue_names [ queuenum ] ) ;
/* So, mark it as being in this queue. */
window - > is_in_queues | = 1 < < queuenum ;
/* There's not a lot of point putting things into a queue if
* nobody ' s on the other end pulling them out . Therefore ,
* let ' s check to see whether an idle handler exists to do
* that . If not , we ' ll create one .
*/
2001-06-21 23:20:21 -04:00
2009-09-25 15:47:22 -04:00
if ( queue_later [ queuenum ] = = 0 )
queue_later [ queuenum ] = meta_later_add
Refactor thrice-duplicated queue code in window.c. Closes #376760.
2007-06-10 Thomas Thurman <thomas@thurman.org.uk>
Refactor thrice-duplicated queue code in window.c. Closes #376760.
* src/window.c (meta_window_queue, meta_window_unqueue):
New functions.
* src/window.[ch] (meta_window_unqueue_*, meta_window_queue_*):
Removed functions.
* src/window.c (meta_window_new_with_attrs, meta_window_free,
meta_window_flush_calc_showing, queue_calc_showing_func,
meta_window_minimize, meta_window_unminimize, meta_window_maximize,
meta_window_make_fullscreen, meta_window_shade,
meta_window_unshade, meta_window_move_resize_internal,
window_stick_impl, window_unstick_impl,
meta_window_client_message, process_property_notify): Modified to
use new queueing functions.
* src/window.c (idle_move_resize, idle_update_icon,
idle_calc_showing): update to receive queue number from pointer.
* src/window.h (MetaQueueType): new enum.
* src/window.h (MetaWindow): *_queued replaced with is_in_queue
bitfield.
* src/core.c (meta_core_queue_frame_resize):
* src/display.c (event_callback,
meta_display_queue_retheme_all_windows): Using new queueing functions.
* src/frame.c (meta_window_destroy_frame): Using new queueing functions.
* src/screen.c (queue_resize, meta_screen_resize_func,
queue_windows_showing): Using new queueing functions.
* src/window-props.c (reload_mwm_hints, reload_wm_hints,
reload_transient_for): Using new queueing functions.
* src/workspace.c (meta_workspace_add_window,
meta_workspace_remove_window, meta_workspace_queue_calc_showing,
meta_workspace_invalidate_work_area): Using new queueing functions.
svn path=/trunk/; revision=3236
2007-06-10 21:15:33 -04:00
(
2009-09-25 15:47:22 -04:00
window_queue_later_when [ queuenum ] ,
window_queue_later_handler [ queuenum ] ,
Refactor thrice-duplicated queue code in window.c. Closes #376760.
2007-06-10 Thomas Thurman <thomas@thurman.org.uk>
Refactor thrice-duplicated queue code in window.c. Closes #376760.
* src/window.c (meta_window_queue, meta_window_unqueue):
New functions.
* src/window.[ch] (meta_window_unqueue_*, meta_window_queue_*):
Removed functions.
* src/window.c (meta_window_new_with_attrs, meta_window_free,
meta_window_flush_calc_showing, queue_calc_showing_func,
meta_window_minimize, meta_window_unminimize, meta_window_maximize,
meta_window_make_fullscreen, meta_window_shade,
meta_window_unshade, meta_window_move_resize_internal,
window_stick_impl, window_unstick_impl,
meta_window_client_message, process_property_notify): Modified to
use new queueing functions.
* src/window.c (idle_move_resize, idle_update_icon,
idle_calc_showing): update to receive queue number from pointer.
* src/window.h (MetaQueueType): new enum.
* src/window.h (MetaWindow): *_queued replaced with is_in_queue
bitfield.
* src/core.c (meta_core_queue_frame_resize):
* src/display.c (event_callback,
meta_display_queue_retheme_all_windows): Using new queueing functions.
* src/frame.c (meta_window_destroy_frame): Using new queueing functions.
* src/screen.c (queue_resize, meta_screen_resize_func,
queue_windows_showing): Using new queueing functions.
* src/window-props.c (reload_mwm_hints, reload_wm_hints,
reload_transient_for): Using new queueing functions.
* src/workspace.c (meta_workspace_add_window,
meta_workspace_remove_window, meta_workspace_queue_calc_showing,
meta_workspace_invalidate_work_area): Using new queueing functions.
svn path=/trunk/; revision=3236
2007-06-10 21:15:33 -04:00
GUINT_TO_POINTER ( queuenum ) ,
NULL
) ;
/* And now we actually put it on the queue. */
queue_pending [ queuenum ] = g_slist_prepend ( queue_pending [ queuenum ] ,
window ) ;
}
}
2001-06-06 00:47:37 -04:00
}
2002-06-14 21:46:13 -04:00
static gboolean
2005-02-12 02:19:41 -05:00
intervening_user_event_occurred ( MetaWindow * window )
2002-06-14 21:46:13 -04:00
{
Fix issues on 64-bit machines with timestamps by using guint32 (like gtk+
2006-09-13 Elijah Newren <newren gmail com>
* src/common.h (MetaWindowMenuFunc):
* src/core.[ch] (meta_core_user_lower_and_unfocus,
meta_core_user_focus, meta_core_show_window_menu,
meta_core_begin_grab_op, meta_core_end_grab_op):
* src/delete.c (delete_ping_reply_func, delete_ping_timeout_func,
meta_window_delete):
* src/display.[ch] (struct MetaDisplay, struct MetaPingData,
sanity_check_timestamps, meta_display_open, event_callback,
meta_spew_event, meta_display_set_grab_op_cursor,
meta_display_begin_grab_op, meta_display_end_grab_op,
meta_display_ping_timeout, meta_display_ping_window,
process_pong_message, timestamp_too_old,
meta_display_set_input_focus_window):
* src/keybindings.[ch] (grab_keyboard, ungrab_keyboard,
meta_screen_grab_all_keys, meta_window_grab_all_keys,
meta_window_ungrab_all_keys, error_on_generic_command,
error_on_command, error_on_terminal_command):
* src/metacity-dialog.c (on_realize, warn_about_no_sm_support,
error_about_command, main):
* src/screen.[ch] (struct _MetaScreen, meta_screen_new,
meta_screen_show_desktop, meta_screen_apply_startup_properties):
* src/session.c (warn_about_lame_clients_and_finish_interact):
* src/window.[ch] (struct _MetaWindow,
intervening_user_event_occurred, window_activate,
meta_window_delete, meta_window_focus,
meta_window_send_icccm_message, meta_window_client_message,
menu_callback, meta_window_show_menu, struct EventScannerData,
check_use_this_motion_notify, meta_window_begin_grab_op,
meta_window_set_user_time):
* src/workspace.[ch] (focus_ancestor_or_mru_window,
meta_workspace_activate_with_focus, meta_workspace_activate,
meta_workspace_focus_default_window,
focus_ancestor_or_mru_window):
Fix issues on 64-bit machines with timestamps by using guint32
(like gtk+ does) instead of Time. #348305
2006-09-13 12:32:33 -04:00
guint32 compare ;
2005-02-12 02:19:41 -05:00
MetaWindow * focus_window ;
focus_window = window - > display - > focus_window ;
meta_topic ( META_DEBUG_STARTUP ,
" COMPARISON: \n "
" net_wm_user_time_set : %d \n "
Fix issues on 64-bit machines with timestamps by using guint32 (like gtk+
2006-09-13 Elijah Newren <newren gmail com>
* src/common.h (MetaWindowMenuFunc):
* src/core.[ch] (meta_core_user_lower_and_unfocus,
meta_core_user_focus, meta_core_show_window_menu,
meta_core_begin_grab_op, meta_core_end_grab_op):
* src/delete.c (delete_ping_reply_func, delete_ping_timeout_func,
meta_window_delete):
* src/display.[ch] (struct MetaDisplay, struct MetaPingData,
sanity_check_timestamps, meta_display_open, event_callback,
meta_spew_event, meta_display_set_grab_op_cursor,
meta_display_begin_grab_op, meta_display_end_grab_op,
meta_display_ping_timeout, meta_display_ping_window,
process_pong_message, timestamp_too_old,
meta_display_set_input_focus_window):
* src/keybindings.[ch] (grab_keyboard, ungrab_keyboard,
meta_screen_grab_all_keys, meta_window_grab_all_keys,
meta_window_ungrab_all_keys, error_on_generic_command,
error_on_command, error_on_terminal_command):
* src/metacity-dialog.c (on_realize, warn_about_no_sm_support,
error_about_command, main):
* src/screen.[ch] (struct _MetaScreen, meta_screen_new,
meta_screen_show_desktop, meta_screen_apply_startup_properties):
* src/session.c (warn_about_lame_clients_and_finish_interact):
* src/window.[ch] (struct _MetaWindow,
intervening_user_event_occurred, window_activate,
meta_window_delete, meta_window_focus,
meta_window_send_icccm_message, meta_window_client_message,
menu_callback, meta_window_show_menu, struct EventScannerData,
check_use_this_motion_notify, meta_window_begin_grab_op,
meta_window_set_user_time):
* src/workspace.[ch] (focus_ancestor_or_mru_window,
meta_workspace_activate_with_focus, meta_workspace_activate,
meta_workspace_focus_default_window,
focus_ancestor_or_mru_window):
Fix issues on 64-bit machines with timestamps by using guint32
(like gtk+ does) instead of Time. #348305
2006-09-13 12:32:33 -04:00
" net_wm_user_time : %u \n "
2005-02-12 02:19:41 -05:00
" initial_timestamp_set: %d \n "
Fix issues on 64-bit machines with timestamps by using guint32 (like gtk+
2006-09-13 Elijah Newren <newren gmail com>
* src/common.h (MetaWindowMenuFunc):
* src/core.[ch] (meta_core_user_lower_and_unfocus,
meta_core_user_focus, meta_core_show_window_menu,
meta_core_begin_grab_op, meta_core_end_grab_op):
* src/delete.c (delete_ping_reply_func, delete_ping_timeout_func,
meta_window_delete):
* src/display.[ch] (struct MetaDisplay, struct MetaPingData,
sanity_check_timestamps, meta_display_open, event_callback,
meta_spew_event, meta_display_set_grab_op_cursor,
meta_display_begin_grab_op, meta_display_end_grab_op,
meta_display_ping_timeout, meta_display_ping_window,
process_pong_message, timestamp_too_old,
meta_display_set_input_focus_window):
* src/keybindings.[ch] (grab_keyboard, ungrab_keyboard,
meta_screen_grab_all_keys, meta_window_grab_all_keys,
meta_window_ungrab_all_keys, error_on_generic_command,
error_on_command, error_on_terminal_command):
* src/metacity-dialog.c (on_realize, warn_about_no_sm_support,
error_about_command, main):
* src/screen.[ch] (struct _MetaScreen, meta_screen_new,
meta_screen_show_desktop, meta_screen_apply_startup_properties):
* src/session.c (warn_about_lame_clients_and_finish_interact):
* src/window.[ch] (struct _MetaWindow,
intervening_user_event_occurred, window_activate,
meta_window_delete, meta_window_focus,
meta_window_send_icccm_message, meta_window_client_message,
menu_callback, meta_window_show_menu, struct EventScannerData,
check_use_this_motion_notify, meta_window_begin_grab_op,
meta_window_set_user_time):
* src/workspace.[ch] (focus_ancestor_or_mru_window,
meta_workspace_activate_with_focus, meta_workspace_activate,
meta_workspace_focus_default_window,
focus_ancestor_or_mru_window):
Fix issues on 64-bit machines with timestamps by using guint32
(like gtk+ does) instead of Time. #348305
2006-09-13 12:32:33 -04:00
" initial_timestamp : %u \n " ,
2005-02-12 02:19:41 -05:00
window - > net_wm_user_time_set ,
window - > net_wm_user_time ,
window - > initial_timestamp_set ,
window - > initial_timestamp ) ;
if ( focus_window ! = NULL )
{
meta_topic ( META_DEBUG_STARTUP ,
" COMPARISON (continued): \n "
2006-08-09 14:32:30 -04:00
" focus_window : %s \n "
" fw->net_wm_user_time_set : %d \n "
Fix issues on 64-bit machines with timestamps by using guint32 (like gtk+
2006-09-13 Elijah Newren <newren gmail com>
* src/common.h (MetaWindowMenuFunc):
* src/core.[ch] (meta_core_user_lower_and_unfocus,
meta_core_user_focus, meta_core_show_window_menu,
meta_core_begin_grab_op, meta_core_end_grab_op):
* src/delete.c (delete_ping_reply_func, delete_ping_timeout_func,
meta_window_delete):
* src/display.[ch] (struct MetaDisplay, struct MetaPingData,
sanity_check_timestamps, meta_display_open, event_callback,
meta_spew_event, meta_display_set_grab_op_cursor,
meta_display_begin_grab_op, meta_display_end_grab_op,
meta_display_ping_timeout, meta_display_ping_window,
process_pong_message, timestamp_too_old,
meta_display_set_input_focus_window):
* src/keybindings.[ch] (grab_keyboard, ungrab_keyboard,
meta_screen_grab_all_keys, meta_window_grab_all_keys,
meta_window_ungrab_all_keys, error_on_generic_command,
error_on_command, error_on_terminal_command):
* src/metacity-dialog.c (on_realize, warn_about_no_sm_support,
error_about_command, main):
* src/screen.[ch] (struct _MetaScreen, meta_screen_new,
meta_screen_show_desktop, meta_screen_apply_startup_properties):
* src/session.c (warn_about_lame_clients_and_finish_interact):
* src/window.[ch] (struct _MetaWindow,
intervening_user_event_occurred, window_activate,
meta_window_delete, meta_window_focus,
meta_window_send_icccm_message, meta_window_client_message,
menu_callback, meta_window_show_menu, struct EventScannerData,
check_use_this_motion_notify, meta_window_begin_grab_op,
meta_window_set_user_time):
* src/workspace.[ch] (focus_ancestor_or_mru_window,
meta_workspace_activate_with_focus, meta_workspace_activate,
meta_workspace_focus_default_window,
focus_ancestor_or_mru_window):
Fix issues on 64-bit machines with timestamps by using guint32
(like gtk+ does) instead of Time. #348305
2006-09-13 12:32:33 -04:00
" fw->net_wm_user_time : %u \n " ,
2005-02-12 02:19:41 -05:00
focus_window - > desc ,
2006-08-09 14:32:30 -04:00
focus_window - > net_wm_user_time_set ,
2005-02-12 02:19:41 -05:00
focus_window - > net_wm_user_time ) ;
}
/* We expect the most common case for not focusing a new window
* to be when a hint to not focus it has been set . Since we can
* deal with that case rapidly , we use special case it - - this is
* merely a preliminary optimization . : )
*/
if ( ( ( window - > net_wm_user_time_set = = TRUE ) & &
( window - > net_wm_user_time = = 0 ) )
| |
( ( window - > initial_timestamp_set = = TRUE ) & &
( window - > initial_timestamp = = 0 ) ) )
{
meta_topic ( META_DEBUG_STARTUP ,
" window %s explicitly requested no focus \n " ,
window - > desc ) ;
return TRUE ;
}
if ( ! ( window - > net_wm_user_time_set ) & & ! ( window - > initial_timestamp_set ) )
{
meta_topic ( META_DEBUG_STARTUP ,
" no information about window %s found \n " ,
window - > desc ) ;
2007-11-09 14:01:31 -05:00
return FALSE ;
2005-02-12 02:19:41 -05:00
}
2006-08-09 14:32:30 -04:00
if ( focus_window ! = NULL & &
! focus_window - > net_wm_user_time_set )
{
meta_topic ( META_DEBUG_STARTUP ,
" focus window, %s, doesn't have a user time set yet! \n " ,
window - > desc ) ;
return FALSE ;
}
2005-02-12 02:19:41 -05:00
/* To determine the "launch" time of an application,
* startup - notification can set the TIMESTAMP and the
* application ( usually via its toolkit such as gtk or qt ) can
2009-08-08 12:59:01 -04:00
* set the _NET_WM_USER_TIME . If both are set , we need to be
* using the newer of the two values .
*
* See http : //bugzilla.gnome.org/show_bug.cgi?id=573922
2005-02-12 02:19:41 -05:00
*/
2009-08-08 12:59:01 -04:00
compare = 0 ;
if ( window - > net_wm_user_time_set & &
window - > initial_timestamp_set )
compare =
XSERVER_TIME_IS_BEFORE ( window - > net_wm_user_time ,
window - > initial_timestamp ) ?
window - > initial_timestamp : window - > net_wm_user_time ;
else if ( window - > net_wm_user_time_set )
compare = window - > net_wm_user_time ;
else if ( window - > initial_timestamp_set )
compare = window - > initial_timestamp ;
2005-02-12 02:19:41 -05:00
if ( ( focus_window ! = NULL ) & &
XSERVER_TIME_IS_BEFORE ( compare , focus_window - > net_wm_user_time ) )
{
meta_topic ( META_DEBUG_STARTUP ,
Fix issues on 64-bit machines with timestamps by using guint32 (like gtk+
2006-09-13 Elijah Newren <newren gmail com>
* src/common.h (MetaWindowMenuFunc):
* src/core.[ch] (meta_core_user_lower_and_unfocus,
meta_core_user_focus, meta_core_show_window_menu,
meta_core_begin_grab_op, meta_core_end_grab_op):
* src/delete.c (delete_ping_reply_func, delete_ping_timeout_func,
meta_window_delete):
* src/display.[ch] (struct MetaDisplay, struct MetaPingData,
sanity_check_timestamps, meta_display_open, event_callback,
meta_spew_event, meta_display_set_grab_op_cursor,
meta_display_begin_grab_op, meta_display_end_grab_op,
meta_display_ping_timeout, meta_display_ping_window,
process_pong_message, timestamp_too_old,
meta_display_set_input_focus_window):
* src/keybindings.[ch] (grab_keyboard, ungrab_keyboard,
meta_screen_grab_all_keys, meta_window_grab_all_keys,
meta_window_ungrab_all_keys, error_on_generic_command,
error_on_command, error_on_terminal_command):
* src/metacity-dialog.c (on_realize, warn_about_no_sm_support,
error_about_command, main):
* src/screen.[ch] (struct _MetaScreen, meta_screen_new,
meta_screen_show_desktop, meta_screen_apply_startup_properties):
* src/session.c (warn_about_lame_clients_and_finish_interact):
* src/window.[ch] (struct _MetaWindow,
intervening_user_event_occurred, window_activate,
meta_window_delete, meta_window_focus,
meta_window_send_icccm_message, meta_window_client_message,
menu_callback, meta_window_show_menu, struct EventScannerData,
check_use_this_motion_notify, meta_window_begin_grab_op,
meta_window_set_user_time):
* src/workspace.[ch] (focus_ancestor_or_mru_window,
meta_workspace_activate_with_focus, meta_workspace_activate,
meta_workspace_focus_default_window,
focus_ancestor_or_mru_window):
Fix issues on 64-bit machines with timestamps by using guint32
(like gtk+ does) instead of Time. #348305
2006-09-13 12:32:33 -04:00
" window %s focus prevented by other activity; %u < %u \n " ,
window - > desc ,
2008-12-18 07:41:56 -05:00
compare ,
2005-02-12 02:19:41 -05:00
focus_window - > net_wm_user_time ) ;
return TRUE ;
}
else
{
meta_topic ( META_DEBUG_STARTUP ,
" new window %s with no intervening events \n " ,
window - > desc ) ;
return FALSE ;
}
}
2006-01-10 13:33:53 -05:00
/* This function is an ugly hack. It's experimental in nature and ought to be
* replaced by a real hint from the app to the WM if we decide the experimental
* behavior is worthwhile . The basic idea is to get more feedback about how
* usage scenarios of " strict " focus users and what they expect . See # 326159.
*/
2014-03-18 10:08:28 -04:00
static gboolean
window_is_terminal ( MetaWindow * window )
2006-01-10 13:33:53 -05:00
{
2006-10-29 15:30:18 -05:00
if ( window = = NULL | | window - > res_class = = NULL )
2006-01-10 13:33:53 -05:00
return FALSE ;
2006-10-29 15:30:18 -05:00
/*
* Compare res_class , which is not user - settable , and thus theoretically
* a more - reliable indication of term - ness .
*/
2006-01-10 13:33:53 -05:00
/* gnome-terminal -- if you couldn't guess */
2006-10-29 15:30:18 -05:00
if ( strcmp ( window - > res_class , " Gnome-terminal " ) = = 0 )
2006-01-10 13:33:53 -05:00
return TRUE ;
/* xterm, rxvt, aterm */
2006-10-29 15:30:18 -05:00
else if ( strcmp ( window - > res_class , " XTerm " ) = = 0 )
2006-01-10 13:33:53 -05:00
return TRUE ;
/* konsole, KDE's terminal program */
2006-10-29 15:30:18 -05:00
else if ( strcmp ( window - > res_class , " Konsole " ) = = 0 )
2006-01-10 13:33:53 -05:00
return TRUE ;
/* rxvt-unicode */
2006-10-29 15:30:18 -05:00
else if ( strcmp ( window - > res_class , " URxvt " ) = = 0 )
2006-01-10 13:33:53 -05:00
return TRUE ;
/* eterm */
2006-10-29 15:30:18 -05:00
else if ( strcmp ( window - > res_class , " Eterm " ) = = 0 )
2006-01-10 13:33:53 -05:00
return TRUE ;
/* KTerm -- some terminal not KDE based; so not like Konsole */
2006-10-29 15:30:18 -05:00
else if ( strcmp ( window - > res_class , " KTerm " ) = = 0 )
2006-01-10 13:33:53 -05:00
return TRUE ;
/* Multi-gnome-terminal */
2006-10-29 15:30:18 -05:00
else if ( strcmp ( window - > res_class , " Multi-gnome-terminal " ) = = 0 )
2006-01-10 13:33:53 -05:00
return TRUE ;
/* mlterm ("multi lingual terminal emulator on X") */
2006-10-29 15:30:18 -05:00
else if ( strcmp ( window - > res_class , " mlterm " ) = = 0 )
2006-01-10 13:33:53 -05:00
return TRUE ;
2009-10-21 21:11:46 -04:00
/* Terminal -- XFCE Terminal */
else if ( strcmp ( window - > res_class , " Terminal " ) = = 0 )
return TRUE ;
2006-01-10 13:33:53 -05:00
return FALSE ;
}
2005-02-12 02:19:41 -05:00
/* This function determines what state the window should have assuming that it
* and the focus_window have no relation
*/
static void
2008-12-18 07:41:56 -05:00
window_state_on_map ( MetaWindow * window ,
2005-02-12 02:19:41 -05:00
gboolean * takes_focus ,
gboolean * places_on_top )
{
gboolean intervening_events ;
intervening_events = intervening_user_event_occurred ( window ) ;
* takes_focus = ! intervening_events ;
* places_on_top = * takes_focus ;
Add support for _NET_WM_USER_TIME
2004-06-17 Elijah Newren <newren@math.utah.edu>
Add support for _NET_WM_USER_TIME
* src/display.c:
(meta_display_open): Add _NET_WM_USER_TIME to atom_names[],
(event_callback): Manually set _NET_WM_USER_TIME upon KeyPress
(doesn't work since keyboard isn't grabbed) and ButtonPress (does
work), this is just a fallback for applications that don't update
this themselves.
* src/display.h: (struct _MetaDisplay): Add atom_net_wm_user_time field
* src/screen.c: (meta_screen_apply_startup_properties): Check for
TIMESTAMP provided from startup sequence as well.
* src/stack.c:
s/meta_window_set_stack_position/meta_window_set_stack_position_no_sync/,
(meta_window_set_stack_position): New function which calls the
meta_window_set_stack_position_no_sync function followed immediately
by calling meta_stack_sync_to_server.
* src/window-props.c:
(init_net_wm_user_time), (reload_net_wm_user_time): new functions,
(reload_wm_hints): also load atom_net_wm_user_time
* src/window.c:
new XSERVER_TIME_IS_LATER macro (accounts for timestamp wraparound),
(meta_window_new_with_attrs): add timestamp attributes,
(window_takes_focus_on_map): use TIMESTAMP from startup
notification and _NET_WM_USER_TIME to decide whether to focus new
windows,
(meta_window_show): if app doesn't take focus on map, place it
just below the focused window in the stack
(process_property_notify): check for changes to _NET_WM_USRE_TIME,
(meta_window_stack_just_below): new function
* src/window.h:
(_MetaWindow struct): new fields for initial_timestamp,
initial_timestamp_set, net_wm_user_time_set, and net_wm_user_time,
(meta_window_stack_just_below): new function
2004-06-24 11:47:05 -04:00
2002-06-14 21:46:13 -04:00
/* don't initially focus windows that are intended to not accept
* focus
*/
if ( ! ( window - > input | | window - > take_focus ) )
2005-02-12 02:19:41 -05:00
{
* takes_focus = FALSE ;
return ;
}
2002-06-14 21:46:13 -04:00
2006-04-12 14:01:20 -04:00
/* Terminal usage may be different; some users intend to launch
2006-01-10 13:33:53 -05:00
* many apps in quick succession or to just view things in the new
2006-04-12 14:01:20 -04:00
* window while still interacting with the terminal . In that case ,
2006-01-10 13:33:53 -05:00
* apps launched from the terminal should not take focus . This
* isn ' t quite the same as not allowing focus to transfer from
* terminals due to new window map , but the latter is a much easier
* approximation to enforce so we do that .
*/
if ( * takes_focus & &
2011-06-24 12:14:15 -04:00
meta_prefs_get_focus_new_windows ( ) = = G_DESKTOP_FOCUS_NEW_WINDOWS_STRICT & &
2006-01-15 12:03:57 -05:00
! window - > display - > allow_terminal_deactivation & &
2014-03-18 10:08:28 -04:00
window_is_terminal ( window - > display - > focus_window ) & &
2006-01-10 13:33:53 -05:00
! meta_window_is_ancestor_of_transient ( window - > display - > focus_window ,
window ) )
{
meta_topic ( META_DEBUG_FOCUS ,
" focus_window is terminal; not focusing new window. \n " ) ;
* takes_focus = FALSE ;
* places_on_top = FALSE ;
}
2002-06-14 21:46:13 -04:00
switch ( window - > type )
{
2005-02-12 02:19:41 -05:00
case META_WINDOW_UTILITY :
case META_WINDOW_TOOLBAR :
* takes_focus = FALSE ;
* places_on_top = FALSE ;
break ;
2002-06-14 21:46:13 -04:00
case META_WINDOW_DOCK :
case META_WINDOW_DESKTOP :
case META_WINDOW_SPLASHSCREEN :
case META_WINDOW_MENU :
2008-11-23 14:28:40 -05:00
/* override redirect types: */
case META_WINDOW_DROPDOWN_MENU :
case META_WINDOW_POPUP_MENU :
case META_WINDOW_TOOLTIP :
case META_WINDOW_NOTIFICATION :
case META_WINDOW_COMBO :
case META_WINDOW_DND :
case META_WINDOW_OVERRIDE_OTHER :
2005-02-12 02:19:41 -05:00
/* don't focus any of these; places_on_top may be irrelevant for some of
* these ( e . g . dock ) - - but you never know - - the focus window might also be
* of the same type in some weird situation . . .
*/
* takes_focus = FALSE ;
2002-06-14 21:46:13 -04:00
break ;
case META_WINDOW_NORMAL :
Add support for _NET_WM_USER_TIME
2004-06-17 Elijah Newren <newren@math.utah.edu>
Add support for _NET_WM_USER_TIME
* src/display.c:
(meta_display_open): Add _NET_WM_USER_TIME to atom_names[],
(event_callback): Manually set _NET_WM_USER_TIME upon KeyPress
(doesn't work since keyboard isn't grabbed) and ButtonPress (does
work), this is just a fallback for applications that don't update
this themselves.
* src/display.h: (struct _MetaDisplay): Add atom_net_wm_user_time field
* src/screen.c: (meta_screen_apply_startup_properties): Check for
TIMESTAMP provided from startup sequence as well.
* src/stack.c:
s/meta_window_set_stack_position/meta_window_set_stack_position_no_sync/,
(meta_window_set_stack_position): New function which calls the
meta_window_set_stack_position_no_sync function followed immediately
by calling meta_stack_sync_to_server.
* src/window-props.c:
(init_net_wm_user_time), (reload_net_wm_user_time): new functions,
(reload_wm_hints): also load atom_net_wm_user_time
* src/window.c:
new XSERVER_TIME_IS_LATER macro (accounts for timestamp wraparound),
(meta_window_new_with_attrs): add timestamp attributes,
(window_takes_focus_on_map): use TIMESTAMP from startup
notification and _NET_WM_USER_TIME to decide whether to focus new
windows,
(meta_window_show): if app doesn't take focus on map, place it
just below the focused window in the stack
(process_property_notify): check for changes to _NET_WM_USRE_TIME,
(meta_window_stack_just_below): new function
* src/window.h:
(_MetaWindow struct): new fields for initial_timestamp,
initial_timestamp_set, net_wm_user_time_set, and net_wm_user_time,
(meta_window_stack_just_below): new function
2004-06-24 11:47:05 -04:00
case META_WINDOW_DIALOG :
case META_WINDOW_MODAL_DIALOG :
2005-02-12 02:19:41 -05:00
/* The default is correct for these */
2002-06-14 21:46:13 -04:00
break ;
}
}
2006-01-10 00:14:48 -05:00
static gboolean
windows_overlap ( const MetaWindow * w1 , const MetaWindow * w2 )
{
MetaRectangle w1rect , w2rect ;
2013-11-19 12:24:45 -05:00
meta_window_get_frame_rect ( w1 , & w1rect ) ;
meta_window_get_frame_rect ( w2 , & w2rect ) ;
2006-01-10 00:14:48 -05:00
return meta_rectangle_overlap ( & w1rect , & w2rect ) ;
}
2007-10-16 18:29:59 -04:00
/* Returns whether a new window would be covered by any
* existing window on the same workspace that is set
* to be " above " ( " always on top " ) . A window that is not
* set " above " would be underneath the new window anyway .
*
* We take " covered " to mean even partially covered , but
* some people might prefer entirely covered . I think it
* is more useful to behave this way if any part of the
* window is covered , because a partial coverage could be
* ( say ) ninety per cent and almost indistinguishable from total .
*/
static gboolean
window_would_be_covered ( const MetaWindow * newbie )
{
MetaWorkspace * workspace = newbie - > workspace ;
GList * tmp , * windows ;
2008-12-18 07:41:56 -05:00
2007-10-16 18:29:59 -04:00
windows = meta_workspace_list_windows ( workspace ) ;
tmp = windows ;
while ( tmp ! = NULL )
2008-12-18 07:41:56 -05:00
{
2007-10-16 18:29:59 -04:00
MetaWindow * w = tmp - > data ;
2008-03-18 08:22:25 -04:00
if ( w - > wm_state_above & & w ! = newbie )
2007-10-16 18:29:59 -04:00
{
/* We have found a window that is "above". Perhaps it overlaps. */
if ( windows_overlap ( w , newbie ) )
{
g_list_free ( windows ) ; /* clean up... */
return TRUE ; /* yes, it does */
}
}
tmp = tmp - > next ;
}
g_list_free ( windows ) ;
return FALSE ; /* none found */
}
2011-02-02 19:14:27 -05:00
static void
meta_window_force_placement ( MetaWindow * window )
{
if ( window - > placed )
return ;
/* We have to recalc the placement here since other windows may
* have been mapped / placed since we last did constrain_position
*/
/* calc_placement is an efficiency hack to avoid
* multiple placement calculations before we finally
* show the window .
*/
window - > calc_placement = TRUE ;
meta_window_move_resize_now ( window ) ;
window - > calc_placement = FALSE ;
/* don't ever do the initial position constraint thing again.
* This is toggled here so that initially - iconified windows
* still get placed when they are ultimately shown .
*/
window - > placed = TRUE ;
/* Don't want to accidentally reuse the fact that we had been denied
* focus in any future constraints unless we ' re denied focus again .
*/
window - > denied_focus_and_not_transient = FALSE ;
}
2009-06-14 07:59:59 -04:00
static void
2001-06-02 00:14:18 -04:00
meta_window_show ( MetaWindow * window )
{
2002-02-08 01:01:58 -05:00
gboolean did_show ;
Add support for _NET_WM_USER_TIME
2004-06-17 Elijah Newren <newren@math.utah.edu>
Add support for _NET_WM_USER_TIME
* src/display.c:
(meta_display_open): Add _NET_WM_USER_TIME to atom_names[],
(event_callback): Manually set _NET_WM_USER_TIME upon KeyPress
(doesn't work since keyboard isn't grabbed) and ButtonPress (does
work), this is just a fallback for applications that don't update
this themselves.
* src/display.h: (struct _MetaDisplay): Add atom_net_wm_user_time field
* src/screen.c: (meta_screen_apply_startup_properties): Check for
TIMESTAMP provided from startup sequence as well.
* src/stack.c:
s/meta_window_set_stack_position/meta_window_set_stack_position_no_sync/,
(meta_window_set_stack_position): New function which calls the
meta_window_set_stack_position_no_sync function followed immediately
by calling meta_stack_sync_to_server.
* src/window-props.c:
(init_net_wm_user_time), (reload_net_wm_user_time): new functions,
(reload_wm_hints): also load atom_net_wm_user_time
* src/window.c:
new XSERVER_TIME_IS_LATER macro (accounts for timestamp wraparound),
(meta_window_new_with_attrs): add timestamp attributes,
(window_takes_focus_on_map): use TIMESTAMP from startup
notification and _NET_WM_USER_TIME to decide whether to focus new
windows,
(meta_window_show): if app doesn't take focus on map, place it
just below the focused window in the stack
(process_property_notify): check for changes to _NET_WM_USRE_TIME,
(meta_window_stack_just_below): new function
* src/window.h:
(_MetaWindow struct): new fields for initial_timestamp,
initial_timestamp_set, net_wm_user_time_set, and net_wm_user_time,
(meta_window_stack_just_below): new function
2004-06-24 11:47:05 -04:00
gboolean takes_focus_on_map ;
2005-02-12 02:19:41 -05:00
gboolean place_on_top_on_map ;
2006-01-10 00:14:48 -05:00
gboolean needs_stacking_adjustment ;
MetaWindow * focus_window ;
2009-11-17 05:02:06 -05:00
gboolean notify_demands_attention = FALSE ;
Add support for _NET_WM_USER_TIME
2004-06-17 Elijah Newren <newren@math.utah.edu>
Add support for _NET_WM_USER_TIME
* src/display.c:
(meta_display_open): Add _NET_WM_USER_TIME to atom_names[],
(event_callback): Manually set _NET_WM_USER_TIME upon KeyPress
(doesn't work since keyboard isn't grabbed) and ButtonPress (does
work), this is just a fallback for applications that don't update
this themselves.
* src/display.h: (struct _MetaDisplay): Add atom_net_wm_user_time field
* src/screen.c: (meta_screen_apply_startup_properties): Check for
TIMESTAMP provided from startup sequence as well.
* src/stack.c:
s/meta_window_set_stack_position/meta_window_set_stack_position_no_sync/,
(meta_window_set_stack_position): New function which calls the
meta_window_set_stack_position_no_sync function followed immediately
by calling meta_stack_sync_to_server.
* src/window-props.c:
(init_net_wm_user_time), (reload_net_wm_user_time): new functions,
(reload_wm_hints): also load atom_net_wm_user_time
* src/window.c:
new XSERVER_TIME_IS_LATER macro (accounts for timestamp wraparound),
(meta_window_new_with_attrs): add timestamp attributes,
(window_takes_focus_on_map): use TIMESTAMP from startup
notification and _NET_WM_USER_TIME to decide whether to focus new
windows,
(meta_window_show): if app doesn't take focus on map, place it
just below the focused window in the stack
(process_property_notify): check for changes to _NET_WM_USRE_TIME,
(meta_window_stack_just_below): new function
* src/window.h:
(_MetaWindow struct): new fields for initial_timestamp,
initial_timestamp_set, net_wm_user_time_set, and net_wm_user_time,
(meta_window_stack_just_below): new function
2004-06-24 11:47:05 -04:00
2002-02-07 22:34:26 -05:00
meta_topic ( META_DEBUG_WINDOW_STATE ,
" Showing window %s, shaded: %d iconic: %d placed: %d \n " ,
window - > desc , window - > shaded , window - > iconic , window - > placed ) ;
2001-06-07 22:17:48 -04:00
2006-01-10 00:14:48 -05:00
focus_window = window - > display - > focus_window ; /* May be NULL! */
2002-02-08 01:01:58 -05:00
did_show = FALSE ;
2005-02-12 02:19:41 -05:00
window_state_on_map ( window , & takes_focus_on_map , & place_on_top_on_map ) ;
2006-01-10 00:14:48 -05:00
needs_stacking_adjustment = FALSE ;
Add support for _NET_WM_USER_TIME
2004-06-17 Elijah Newren <newren@math.utah.edu>
Add support for _NET_WM_USER_TIME
* src/display.c:
(meta_display_open): Add _NET_WM_USER_TIME to atom_names[],
(event_callback): Manually set _NET_WM_USER_TIME upon KeyPress
(doesn't work since keyboard isn't grabbed) and ButtonPress (does
work), this is just a fallback for applications that don't update
this themselves.
* src/display.h: (struct _MetaDisplay): Add atom_net_wm_user_time field
* src/screen.c: (meta_screen_apply_startup_properties): Check for
TIMESTAMP provided from startup sequence as well.
* src/stack.c:
s/meta_window_set_stack_position/meta_window_set_stack_position_no_sync/,
(meta_window_set_stack_position): New function which calls the
meta_window_set_stack_position_no_sync function followed immediately
by calling meta_stack_sync_to_server.
* src/window-props.c:
(init_net_wm_user_time), (reload_net_wm_user_time): new functions,
(reload_wm_hints): also load atom_net_wm_user_time
* src/window.c:
new XSERVER_TIME_IS_LATER macro (accounts for timestamp wraparound),
(meta_window_new_with_attrs): add timestamp attributes,
(window_takes_focus_on_map): use TIMESTAMP from startup
notification and _NET_WM_USER_TIME to decide whether to focus new
windows,
(meta_window_show): if app doesn't take focus on map, place it
just below the focused window in the stack
(process_property_notify): check for changes to _NET_WM_USRE_TIME,
(meta_window_stack_just_below): new function
* src/window.h:
(_MetaWindow struct): new fields for initial_timestamp,
initial_timestamp_set, net_wm_user_time_set, and net_wm_user_time,
(meta_window_stack_just_below): new function
2004-06-24 11:47:05 -04:00
2005-02-12 02:19:41 -05:00
meta_topic ( META_DEBUG_WINDOW_STATE ,
" Window %s %s focus on map, and %s place on top on map. \n " ,
window - > desc ,
takes_focus_on_map ? " does " : " does not " ,
place_on_top_on_map ? " does " : " does not " ) ;
2007-10-16 18:29:59 -04:00
/* Now, in some rare cases we should *not* put a new window on top.
2007-11-11 19:45:35 -05:00
* These cases include certain types of windows showing for the first
2007-10-16 18:29:59 -04:00
* time , and any window which would be covered because of another window
* being set " above " ( " always on top " ) .
*
* FIXME : Although " place_on_top_on_map " and " takes_focus_on_map " are
* generally based on the window type , there is a special case when the
* focus window is a terminal for them both to be false ; this should
* probably rather be a term in the " if " condition below .
*/
2007-11-11 19:45:35 -05:00
if ( focus_window ! = NULL & & window - > showing_for_first_time & &
( ( ! place_on_top_on_map & & ! takes_focus_on_map ) | |
window_would_be_covered ( window ) )
) {
2006-01-10 00:14:48 -05:00
if ( meta_window_is_ancestor_of_transient ( focus_window , window ) )
2004-10-25 12:28:57 -04:00
{
Simplify relationship between mapping and visibility
Previously, changes to the visibility of a window could be indicated
by meta_compositor_map_window(), meta_compositor_unminimize_window(),
meta_compositor_set_window_hidden(), etc, with the exact behavior
depending on the 'live_hidden_windows' preference.
Simplify this so that visibility is controlled by:
meta_compositor_show_window()
meta_compositor_hide_window()
With an 'effect' parameter provided to indicate the appropriate
effect (CREATE/UNMINIMIZE/MINIMIZE/DESTROY/NONE.)
The map state of the window is signalled separately by:
meta_compositor_map_window()
meta_compositor_unmap_window()
And is used only to control resource handling.
Other changes:
* The desired effect on show/hide is explicitly stored in
MetaWindow, avoiding the need for the was_minimized flag.
At idle, once we calculate the window state, we pass the
effect to the compositor if it matches the new window
state, and then clear the effect to start over for future
map state changes.
* meta_compositor_switch_workspace() is called before any windows
are hidden or shown, allowing the compositor to avoid hiding
or showing an effect for windows involved in the switch.
http://bugzilla.gnome.org/show_bug.cgi?id=582341
* Handling of post-effect cleanups for MutterWindow are
simplified - instead of trying to do different things based
on the individual needs of different effects, we just wait until
all effects complete and sync the window state to what it
should be.
* On unmap, once we destroy the pixmap, we tell ClutterX11Pixmap
that we've done so, so it can clean up and unbind. (The
unbinding doesn't seem to be working properly because of
ClutterGLXPixmap or video driver issues.)
http://bugzilla.gnome.org/show_bug.cgi?id=587251
2009-06-28 17:10:40 -04:00
guint32 timestamp ;
timestamp = meta_display_get_current_time_roundtrip ( window - > display ) ;
2004-10-25 12:28:57 -04:00
/* This happens for error dialogs or alerts; these need to remain on
* top , but it would be confusing to have its ancestor remain
* focused .
*/
meta_topic ( META_DEBUG_STARTUP ,
" The focus window %s is an ancestor of the newly mapped "
" window %s which isn't being focused. Unfocusing the "
" ancestor. \n " ,
2006-01-10 00:14:48 -05:00
focus_window - > desc , window - > desc ) ;
2004-10-25 12:28:57 -04:00
Partial audit to fix timestamp usage. One step towards fixing #355180; see
2006-09-18 Elijah Newren <newren gmail com>
Partial audit to fix timestamp usage. One step towards fixing
#355180; see important comments in that bug.
* src/core.[ch] (meta_core_unshade, meta_core_shade):
* src/delete.c (meta_window_present_delete_dialog,
delete_ping_timeout_func):
* src/display.[ch] (meta_display_open, meta_display_close,
event_callback, meta_display_begin_grab_op,
process_selection_clear, meta_display_unmanage_screen,
meta_display_unmanage_windows_for_screen):
* src/frames.c (meta_frames_button_press_event):
* src/keybindings.c (handle_toggle_shade):
* src/main.c (main):
* src/screen.[ch] (update_num_workspaces, meta_screen_new,
meta_screen_free, prefs_changed_callback):
* src/window.[ch] (meta_window_free, finish_minimize,
implement_showing, meta_window_show, meta_window_maximize,
meta_window_make_fullscreen_internal,
meta_window_unmake_fullscreen, meta_window_shade,
meta_window_unshade, window_activate, send_sync_request,
meta_window_client_message, menu_callback,
meta_window_update_keyboard_resize):
Remove usage of CurrentTime, meta_display_get_current_time() and
meta_display_get_current_time_roundtrip() where possible, or
document why it isn't possible, or at very least add a FIXME with
some explanation of my laziness and what needs to be done.
2006-09-18 13:27:24 -04:00
meta_display_focus_the_no_focus_window ( window - > display ,
window - > screen ,
timestamp ) ;
2004-10-25 12:28:57 -04:00
}
else
{
2006-01-10 00:14:48 -05:00
needs_stacking_adjustment = TRUE ;
2005-02-11 14:20:44 -05:00
if ( ! window - > placed )
window - > denied_focus_and_not_transient = TRUE ;
2004-10-25 12:28:57 -04:00
}
2004-08-05 13:48:39 -04:00
}
Add support for _NET_WM_USER_TIME
2004-06-17 Elijah Newren <newren@math.utah.edu>
Add support for _NET_WM_USER_TIME
* src/display.c:
(meta_display_open): Add _NET_WM_USER_TIME to atom_names[],
(event_callback): Manually set _NET_WM_USER_TIME upon KeyPress
(doesn't work since keyboard isn't grabbed) and ButtonPress (does
work), this is just a fallback for applications that don't update
this themselves.
* src/display.h: (struct _MetaDisplay): Add atom_net_wm_user_time field
* src/screen.c: (meta_screen_apply_startup_properties): Check for
TIMESTAMP provided from startup sequence as well.
* src/stack.c:
s/meta_window_set_stack_position/meta_window_set_stack_position_no_sync/,
(meta_window_set_stack_position): New function which calls the
meta_window_set_stack_position_no_sync function followed immediately
by calling meta_stack_sync_to_server.
* src/window-props.c:
(init_net_wm_user_time), (reload_net_wm_user_time): new functions,
(reload_wm_hints): also load atom_net_wm_user_time
* src/window.c:
new XSERVER_TIME_IS_LATER macro (accounts for timestamp wraparound),
(meta_window_new_with_attrs): add timestamp attributes,
(window_takes_focus_on_map): use TIMESTAMP from startup
notification and _NET_WM_USER_TIME to decide whether to focus new
windows,
(meta_window_show): if app doesn't take focus on map, place it
just below the focused window in the stack
(process_property_notify): check for changes to _NET_WM_USRE_TIME,
(meta_window_stack_just_below): new function
* src/window.h:
(_MetaWindow struct): new fields for initial_timestamp,
initial_timestamp_set, net_wm_user_time_set, and net_wm_user_time,
(meta_window_stack_just_below): new function
2004-06-24 11:47:05 -04:00
2001-06-24 04:09:10 -04:00
if ( ! window - > placed )
2012-03-15 15:42:11 -04:00
{
2012-09-08 05:34:27 -04:00
if ( meta_prefs_get_auto_maximize ( ) & & window - > showing_for_first_time & & window - > has_maximize_func )
2012-03-15 15:42:11 -04:00
{
MetaRectangle work_area ;
meta_window_get_work_area_for_monitor ( window , window - > monitor - > number , & work_area ) ;
/* Automaximize windows that map with a size > MAX_UNMAXIMIZED_WINDOW_AREA of the work area */
if ( window - > rect . width * window - > rect . height > work_area . width * work_area . height * MAX_UNMAXIMIZED_WINDOW_AREA )
{
window - > maximize_horizontally_after_placement = TRUE ;
window - > maximize_vertically_after_placement = TRUE ;
}
}
meta_window_force_placement ( window ) ;
}
2008-12-18 07:41:56 -05:00
2006-09-13 18:21:19 -04:00
if ( needs_stacking_adjustment )
{
2006-09-18 18:00:57 -04:00
gboolean overlap ;
2006-09-13 18:21:19 -04:00
/* This window isn't getting focus on map. We may need to do some
* special handing with it in regards to
* - the stacking of the window
* - the MRU position of the window
* - the demands attention setting of the window
2007-10-16 18:29:59 -04:00
*
* Firstly , set the flag so we don ' t give the window focus anyway
* and confuse people .
2006-09-13 18:21:19 -04:00
*/
2008-12-18 07:41:56 -05:00
2007-10-16 18:29:59 -04:00
takes_focus_on_map = FALSE ;
2006-09-13 18:21:19 -04:00
overlap = windows_overlap ( window , focus_window ) ;
/* We want alt tab to go to the denied-focus window */
ensure_mru_position_after ( window , focus_window ) ;
/* We don't want the denied-focus window to obscure the focus
* window , and if we ' re in both click - to - focus mode and
* raise - on - click mode then we want to maintain the invariant
* that MRU order = = stacking order . The need for this if
* comes from the fact that in sloppy / mouse focus the focus
* window may not overlap other windows and also can be
* considered " below " them ; this combination means that
* placing the denied - focus window " below " the focus window
* in the stack when it doesn ' t overlap it confusingly places
* that new window below a lot of other windows .
*/
2008-12-18 07:41:56 -05:00
if ( overlap | |
2011-06-24 12:14:15 -04:00
( meta_prefs_get_focus_mode ( ) = = G_DESKTOP_FOCUS_MODE_CLICK & &
2006-09-13 18:21:19 -04:00
meta_prefs_get_raise_on_click ( ) ) )
meta_window_stack_just_below ( window , focus_window ) ;
/* If the window will be obscured by the focus window, then the
* user might not notice the window appearing so set the
* demands attention hint .
*
2008-12-18 07:41:56 -05:00
* We set the hint ourselves rather than calling
2006-09-13 18:21:19 -04:00
* meta_window_set_demands_attention ( ) because that would cause
* a recalculation of overlap , and a call to set_net_wm_state ( )
* which we are going to call ourselves here a few lines down .
*/
if ( overlap )
2009-11-17 05:02:06 -05:00
{
if ( ! window - > wm_state_demands_attention )
{
window - > wm_state_demands_attention = TRUE ;
notify_demands_attention = TRUE ;
}
}
2008-12-18 07:41:56 -05:00
}
2006-09-13 18:21:19 -04:00
Always map the client and frame windows
Traditionally, WMs unmap windows when minimizing them, and map them
when restoring them or wanting to show them for other reasons, like
upon creation.
However, as metacity morphed into mutter, we optionally chose to keep
windows mapped for the lifetime of the window under the user option
"live-window-previews", which makes the code keep windows mapped so it
can show window preview for minimized windows in other places, like
Alt-Tab and Expose.
I removed this preference two years ago mechanically, by removing all
the if statements, but never went through and cleaned up the code so
that windows are simply mapped for the lifetime of the window -- the
"architecture" of the old code that maps and unmaps on show/hide was
still there.
Remove this now.
The one case we still need to be careful of is shaded windows, in which
we do still unmap the client window. In the future, we might want to
show previews of shaded windows in the overview and Alt-Tab. In that
we'd also keep shaded windows mapped, and could remove all unmap logic,
but we'd need a more complex method of showing the shaded titlebar, such
as using a different actor.
At the same time, simplify the compositor interface by removing
meta_compositor_window_[un]mapped API, and instead adding/removing the
window on-demand.
https://bugzilla.gnome.org/show_bug.cgi?id=720631
2014-02-01 19:06:51 -05:00
if ( window - > hidden )
2001-06-07 22:17:48 -04:00
{
Always map the client and frame windows
Traditionally, WMs unmap windows when minimizing them, and map them
when restoring them or wanting to show them for other reasons, like
upon creation.
However, as metacity morphed into mutter, we optionally chose to keep
windows mapped for the lifetime of the window under the user option
"live-window-previews", which makes the code keep windows mapped so it
can show window preview for minimized windows in other places, like
Alt-Tab and Expose.
I removed this preference two years ago mechanically, by removing all
the if statements, but never went through and cleaned up the code so
that windows are simply mapped for the lifetime of the window -- the
"architecture" of the old code that maps and unmaps on show/hide was
still there.
Remove this now.
The one case we still need to be careful of is shaded windows, in which
we do still unmap the client window. In the future, we might want to
show previews of shaded windows in the overview and Alt-Tab. In that
we'd also keep shaded windows mapped, and could remove all unmap logic,
but we'd need a more complex method of showing the shaded titlebar, such
as using a different actor.
At the same time, simplify the compositor interface by removing
meta_compositor_window_[un]mapped API, and instead adding/removing the
window on-demand.
https://bugzilla.gnome.org/show_bug.cgi?id=720631
2014-02-01 19:06:51 -05:00
meta_stack_freeze ( window - > screen - > stack ) ;
window - > hidden = FALSE ;
meta_stack_thaw ( window - > screen - > stack ) ;
did_show = TRUE ;
2014-01-22 09:16:56 -05:00
}
Always map the client and frame windows
Traditionally, WMs unmap windows when minimizing them, and map them
when restoring them or wanting to show them for other reasons, like
upon creation.
However, as metacity morphed into mutter, we optionally chose to keep
windows mapped for the lifetime of the window under the user option
"live-window-previews", which makes the code keep windows mapped so it
can show window preview for minimized windows in other places, like
Alt-Tab and Expose.
I removed this preference two years ago mechanically, by removing all
the if statements, but never went through and cleaned up the code so
that windows are simply mapped for the lifetime of the window -- the
"architecture" of the old code that maps and unmaps on show/hide was
still there.
Remove this now.
The one case we still need to be careful of is shaded windows, in which
we do still unmap the client window. In the future, we might want to
show previews of shaded windows in the overview and Alt-Tab. In that
we'd also keep shaded windows mapped, and could remove all unmap logic,
but we'd need a more complex method of showing the shaded titlebar, such
as using a different actor.
At the same time, simplify the compositor interface by removing
meta_compositor_window_[un]mapped API, and instead adding/removing the
window on-demand.
https://bugzilla.gnome.org/show_bug.cgi?id=720631
2014-02-01 19:06:51 -05:00
if ( window - > iconic )
Simplify relationship between mapping and visibility
Previously, changes to the visibility of a window could be indicated
by meta_compositor_map_window(), meta_compositor_unminimize_window(),
meta_compositor_set_window_hidden(), etc, with the exact behavior
depending on the 'live_hidden_windows' preference.
Simplify this so that visibility is controlled by:
meta_compositor_show_window()
meta_compositor_hide_window()
With an 'effect' parameter provided to indicate the appropriate
effect (CREATE/UNMINIMIZE/MINIMIZE/DESTROY/NONE.)
The map state of the window is signalled separately by:
meta_compositor_map_window()
meta_compositor_unmap_window()
And is used only to control resource handling.
Other changes:
* The desired effect on show/hide is explicitly stored in
MetaWindow, avoiding the need for the was_minimized flag.
At idle, once we calculate the window state, we pass the
effect to the compositor if it matches the new window
state, and then clear the effect to start over for future
map state changes.
* meta_compositor_switch_workspace() is called before any windows
are hidden or shown, allowing the compositor to avoid hiding
or showing an effect for windows involved in the switch.
http://bugzilla.gnome.org/show_bug.cgi?id=582341
* Handling of post-effect cleanups for MutterWindow are
simplified - instead of trying to do different things based
on the individual needs of different effects, we just wait until
all effects complete and sync the window state to what it
should be.
* On unmap, once we destroy the pixmap, we tell ClutterX11Pixmap
that we've done so, so it can clean up and unbind. (The
unbinding doesn't seem to be working properly because of
ClutterGLXPixmap or video driver issues.)
http://bugzilla.gnome.org/show_bug.cgi?id=587251
2009-06-28 17:10:40 -04:00
{
Always map the client and frame windows
Traditionally, WMs unmap windows when minimizing them, and map them
when restoring them or wanting to show them for other reasons, like
upon creation.
However, as metacity morphed into mutter, we optionally chose to keep
windows mapped for the lifetime of the window under the user option
"live-window-previews", which makes the code keep windows mapped so it
can show window preview for minimized windows in other places, like
Alt-Tab and Expose.
I removed this preference two years ago mechanically, by removing all
the if statements, but never went through and cleaned up the code so
that windows are simply mapped for the lifetime of the window -- the
"architecture" of the old code that maps and unmaps on show/hide was
still there.
Remove this now.
The one case we still need to be careful of is shaded windows, in which
we do still unmap the client window. In the future, we might want to
show previews of shaded windows in the overview and Alt-Tab. In that
we'd also keep shaded windows mapped, and could remove all unmap logic,
but we'd need a more complex method of showing the shaded titlebar, such
as using a different actor.
At the same time, simplify the compositor interface by removing
meta_compositor_window_[un]mapped API, and instead adding/removing the
window on-demand.
https://bugzilla.gnome.org/show_bug.cgi?id=720631
2014-02-01 19:06:51 -05:00
window - > iconic = FALSE ;
set_wm_state ( window ) ;
Simplify relationship between mapping and visibility
Previously, changes to the visibility of a window could be indicated
by meta_compositor_map_window(), meta_compositor_unminimize_window(),
meta_compositor_set_window_hidden(), etc, with the exact behavior
depending on the 'live_hidden_windows' preference.
Simplify this so that visibility is controlled by:
meta_compositor_show_window()
meta_compositor_hide_window()
With an 'effect' parameter provided to indicate the appropriate
effect (CREATE/UNMINIMIZE/MINIMIZE/DESTROY/NONE.)
The map state of the window is signalled separately by:
meta_compositor_map_window()
meta_compositor_unmap_window()
And is used only to control resource handling.
Other changes:
* The desired effect on show/hide is explicitly stored in
MetaWindow, avoiding the need for the was_minimized flag.
At idle, once we calculate the window state, we pass the
effect to the compositor if it matches the new window
state, and then clear the effect to start over for future
map state changes.
* meta_compositor_switch_workspace() is called before any windows
are hidden or shown, allowing the compositor to avoid hiding
or showing an effect for windows involved in the switch.
http://bugzilla.gnome.org/show_bug.cgi?id=582341
* Handling of post-effect cleanups for MutterWindow are
simplified - instead of trying to do different things based
on the individual needs of different effects, we just wait until
all effects complete and sync the window state to what it
should be.
* On unmap, once we destroy the pixmap, we tell ClutterX11Pixmap
that we've done so, so it can clean up and unbind. (The
unbinding doesn't seem to be working properly because of
ClutterGLXPixmap or video driver issues.)
http://bugzilla.gnome.org/show_bug.cgi?id=587251
2009-06-28 17:10:40 -04:00
}
2008-12-18 07:41:56 -05:00
Simplify relationship between mapping and visibility
Previously, changes to the visibility of a window could be indicated
by meta_compositor_map_window(), meta_compositor_unminimize_window(),
meta_compositor_set_window_hidden(), etc, with the exact behavior
depending on the 'live_hidden_windows' preference.
Simplify this so that visibility is controlled by:
meta_compositor_show_window()
meta_compositor_hide_window()
With an 'effect' parameter provided to indicate the appropriate
effect (CREATE/UNMINIMIZE/MINIMIZE/DESTROY/NONE.)
The map state of the window is signalled separately by:
meta_compositor_map_window()
meta_compositor_unmap_window()
And is used only to control resource handling.
Other changes:
* The desired effect on show/hide is explicitly stored in
MetaWindow, avoiding the need for the was_minimized flag.
At idle, once we calculate the window state, we pass the
effect to the compositor if it matches the new window
state, and then clear the effect to start over for future
map state changes.
* meta_compositor_switch_workspace() is called before any windows
are hidden or shown, allowing the compositor to avoid hiding
or showing an effect for windows involved in the switch.
http://bugzilla.gnome.org/show_bug.cgi?id=582341
* Handling of post-effect cleanups for MutterWindow are
simplified - instead of trying to do different things based
on the individual needs of different effects, we just wait until
all effects complete and sync the window state to what it
should be.
* On unmap, once we destroy the pixmap, we tell ClutterX11Pixmap
that we've done so, so it can clean up and unbind. (The
unbinding doesn't seem to be working properly because of
ClutterGLXPixmap or video driver issues.)
http://bugzilla.gnome.org/show_bug.cgi?id=587251
2009-06-28 17:10:40 -04:00
if ( ! window - > visible_to_compositor )
{
2014-03-18 17:06:42 -04:00
MetaCompEffect effect = META_COMP_EFFECT_NONE ;
2011-01-04 11:43:02 -05:00
window - > visible_to_compositor = TRUE ;
2014-03-18 17:06:42 -04:00
switch ( window - > pending_compositor_effect )
Simplify relationship between mapping and visibility
Previously, changes to the visibility of a window could be indicated
by meta_compositor_map_window(), meta_compositor_unminimize_window(),
meta_compositor_set_window_hidden(), etc, with the exact behavior
depending on the 'live_hidden_windows' preference.
Simplify this so that visibility is controlled by:
meta_compositor_show_window()
meta_compositor_hide_window()
With an 'effect' parameter provided to indicate the appropriate
effect (CREATE/UNMINIMIZE/MINIMIZE/DESTROY/NONE.)
The map state of the window is signalled separately by:
meta_compositor_map_window()
meta_compositor_unmap_window()
And is used only to control resource handling.
Other changes:
* The desired effect on show/hide is explicitly stored in
MetaWindow, avoiding the need for the was_minimized flag.
At idle, once we calculate the window state, we pass the
effect to the compositor if it matches the new window
state, and then clear the effect to start over for future
map state changes.
* meta_compositor_switch_workspace() is called before any windows
are hidden or shown, allowing the compositor to avoid hiding
or showing an effect for windows involved in the switch.
http://bugzilla.gnome.org/show_bug.cgi?id=582341
* Handling of post-effect cleanups for MutterWindow are
simplified - instead of trying to do different things based
on the individual needs of different effects, we just wait until
all effects complete and sync the window state to what it
should be.
* On unmap, once we destroy the pixmap, we tell ClutterX11Pixmap
that we've done so, so it can clean up and unbind. (The
unbinding doesn't seem to be working properly because of
ClutterGLXPixmap or video driver issues.)
http://bugzilla.gnome.org/show_bug.cgi?id=587251
2009-06-28 17:10:40 -04:00
{
2014-03-18 17:06:42 -04:00
case META_COMP_EFFECT_CREATE :
case META_COMP_EFFECT_UNMINIMIZE :
effect = window - > pending_compositor_effect ;
break ;
case META_COMP_EFFECT_NONE :
case META_COMP_EFFECT_DESTROY :
case META_COMP_EFFECT_MINIMIZE :
break ;
2009-01-12 11:41:06 -05:00
}
2014-03-18 17:06:42 -04:00
meta_compositor_show_window ( window - > display - > compositor ,
window , effect ) ;
2001-06-06 00:47:37 -04:00
}
2001-10-29 21:00:53 -05:00
2004-12-28 01:01:42 -05:00
/* We don't want to worry about all cases from inside
* implement_showing ( ) ; we only want to worry about focus if this
* window has not been shown before .
*/
if ( window - > showing_for_first_time )
2001-10-29 21:00:53 -05:00
{
2004-12-28 01:01:42 -05:00
window - > showing_for_first_time = FALSE ;
Add support for _NET_WM_USER_TIME
2004-06-17 Elijah Newren <newren@math.utah.edu>
Add support for _NET_WM_USER_TIME
* src/display.c:
(meta_display_open): Add _NET_WM_USER_TIME to atom_names[],
(event_callback): Manually set _NET_WM_USER_TIME upon KeyPress
(doesn't work since keyboard isn't grabbed) and ButtonPress (does
work), this is just a fallback for applications that don't update
this themselves.
* src/display.h: (struct _MetaDisplay): Add atom_net_wm_user_time field
* src/screen.c: (meta_screen_apply_startup_properties): Check for
TIMESTAMP provided from startup sequence as well.
* src/stack.c:
s/meta_window_set_stack_position/meta_window_set_stack_position_no_sync/,
(meta_window_set_stack_position): New function which calls the
meta_window_set_stack_position_no_sync function followed immediately
by calling meta_stack_sync_to_server.
* src/window-props.c:
(init_net_wm_user_time), (reload_net_wm_user_time): new functions,
(reload_wm_hints): also load atom_net_wm_user_time
* src/window.c:
new XSERVER_TIME_IS_LATER macro (accounts for timestamp wraparound),
(meta_window_new_with_attrs): add timestamp attributes,
(window_takes_focus_on_map): use TIMESTAMP from startup
notification and _NET_WM_USER_TIME to decide whether to focus new
windows,
(meta_window_show): if app doesn't take focus on map, place it
just below the focused window in the stack
(process_property_notify): check for changes to _NET_WM_USRE_TIME,
(meta_window_stack_just_below): new function
* src/window.h:
(_MetaWindow struct): new fields for initial_timestamp,
initial_timestamp_set, net_wm_user_time_set, and net_wm_user_time,
(meta_window_stack_just_below): new function
2004-06-24 11:47:05 -04:00
if ( takes_focus_on_map )
2008-12-18 07:41:56 -05:00
{
Simplify relationship between mapping and visibility
Previously, changes to the visibility of a window could be indicated
by meta_compositor_map_window(), meta_compositor_unminimize_window(),
meta_compositor_set_window_hidden(), etc, with the exact behavior
depending on the 'live_hidden_windows' preference.
Simplify this so that visibility is controlled by:
meta_compositor_show_window()
meta_compositor_hide_window()
With an 'effect' parameter provided to indicate the appropriate
effect (CREATE/UNMINIMIZE/MINIMIZE/DESTROY/NONE.)
The map state of the window is signalled separately by:
meta_compositor_map_window()
meta_compositor_unmap_window()
And is used only to control resource handling.
Other changes:
* The desired effect on show/hide is explicitly stored in
MetaWindow, avoiding the need for the was_minimized flag.
At idle, once we calculate the window state, we pass the
effect to the compositor if it matches the new window
state, and then clear the effect to start over for future
map state changes.
* meta_compositor_switch_workspace() is called before any windows
are hidden or shown, allowing the compositor to avoid hiding
or showing an effect for windows involved in the switch.
http://bugzilla.gnome.org/show_bug.cgi?id=582341
* Handling of post-effect cleanups for MutterWindow are
simplified - instead of trying to do different things based
on the individual needs of different effects, we just wait until
all effects complete and sync the window state to what it
should be.
* On unmap, once we destroy the pixmap, we tell ClutterX11Pixmap
that we've done so, so it can clean up and unbind. (The
unbinding doesn't seem to be working properly because of
ClutterGLXPixmap or video driver issues.)
http://bugzilla.gnome.org/show_bug.cgi?id=587251
2009-06-28 17:10:40 -04:00
guint32 timestamp ;
timestamp = meta_display_get_current_time_roundtrip ( window - > display ) ;
Partial audit to fix timestamp usage. One step towards fixing #355180; see
2006-09-18 Elijah Newren <newren gmail com>
Partial audit to fix timestamp usage. One step towards fixing
#355180; see important comments in that bug.
* src/core.[ch] (meta_core_unshade, meta_core_shade):
* src/delete.c (meta_window_present_delete_dialog,
delete_ping_timeout_func):
* src/display.[ch] (meta_display_open, meta_display_close,
event_callback, meta_display_begin_grab_op,
process_selection_clear, meta_display_unmanage_screen,
meta_display_unmanage_windows_for_screen):
* src/frames.c (meta_frames_button_press_event):
* src/keybindings.c (handle_toggle_shade):
* src/main.c (main):
* src/screen.[ch] (update_num_workspaces, meta_screen_new,
meta_screen_free, prefs_changed_callback):
* src/window.[ch] (meta_window_free, finish_minimize,
implement_showing, meta_window_show, meta_window_maximize,
meta_window_make_fullscreen_internal,
meta_window_unmake_fullscreen, meta_window_shade,
meta_window_unshade, window_activate, send_sync_request,
meta_window_client_message, menu_callback,
meta_window_update_keyboard_resize):
Remove usage of CurrentTime, meta_display_get_current_time() and
meta_display_get_current_time_roundtrip() where possible, or
document why it isn't possible, or at very least add a FIXME with
some explanation of my laziness and what needs to be done.
2006-09-18 13:27:24 -04:00
meta_window_focus ( window , timestamp ) ;
2002-05-16 15:08:20 -04:00
}
Fix some support for EWMH hints, and fix USER_TIME support to include the
2004-07-31 Rob Adams <readams@readams.net>
Fix some support for EWMH hints, and fix USER_TIME support to
include the DEMANDS_ATTENTION hint. Also includes some code for
implementing _NET_RESTACK_WINDOW and _NET_MOVERESIZE_WINDOW, but
this is disabled pending feature thaw.
* COMPLIANCE: update with new information
* src/display.c (meta_display_open): add new hints to list
* src/display.h (_MetaDisplay): Add new atoms to struct
* src/screen.c (set_supported_hint): update the list of support
hints.
(set_desktop_viewport_hint): new function sets the viewport hint
to (0,0) as required by the spec for WMs with no viewport support.
(set_desktop_geometry_hint): new function to set the desktop size
hint to the size of the display, since we don't implement large
desktop support, as required by the spec.
(meta_screen_resize): update the geometry hint on screen resize
* src/window.c (meta_window_new_with_attrs): Initialize
demands_attention state
(set_net_wm_state): Set demands_attention hint in the window state
(meta_window_show): If we don't pop up a window because of
USER_TIME, set DEMANDS_ATTENTION on the window.
(meta_window_focus): When a window receives focus, remove
DEMANDS_ATTENTION hint
(meta_window_client_message): Allow other apps to set
DEMANDS_ATTENTION on a window. Also, if the _NET_ACTIVE_WINDOW
hint includes a timestamp, use it.
(update_net_wm_state): Read DEMANDS_ATTENTION state also
* src/window.h (_MetaWindow): add wm_state_demands_attention bit.
2004-07-31 15:56:10 -04:00
else
2004-10-25 12:28:57 -04:00
{
/* Prevent EnterNotify events in sloppy/mouse focus from
* erroneously focusing the window that had been denied
* focus . FIXME : This introduces a race ; I have a couple
* ideas for a better way to accomplish the same thing , but
* they ' re more involved so do it this way for now .
*/
meta_display_increment_focus_sentinel ( window - > display ) ;
}
2001-10-29 21:00:53 -05:00
}
2002-02-08 01:01:58 -05:00
2006-03-16 22:26:25 -05:00
set_net_wm_state ( window ) ;
Fix some support for EWMH hints, and fix USER_TIME support to include the
2004-07-31 Rob Adams <readams@readams.net>
Fix some support for EWMH hints, and fix USER_TIME support to
include the DEMANDS_ATTENTION hint. Also includes some code for
implementing _NET_RESTACK_WINDOW and _NET_MOVERESIZE_WINDOW, but
this is disabled pending feature thaw.
* COMPLIANCE: update with new information
* src/display.c (meta_display_open): add new hints to list
* src/display.h (_MetaDisplay): Add new atoms to struct
* src/screen.c (set_supported_hint): update the list of support
hints.
(set_desktop_viewport_hint): new function sets the viewport hint
to (0,0) as required by the spec for WMs with no viewport support.
(set_desktop_geometry_hint): new function to set the desktop size
hint to the size of the display, since we don't implement large
desktop support, as required by the spec.
(meta_screen_resize): update the geometry hint on screen resize
* src/window.c (meta_window_new_with_attrs): Initialize
demands_attention state
(set_net_wm_state): Set demands_attention hint in the window state
(meta_window_show): If we don't pop up a window because of
USER_TIME, set DEMANDS_ATTENTION on the window.
(meta_window_focus): When a window receives focus, remove
DEMANDS_ATTENTION hint
(meta_window_client_message): Allow other apps to set
DEMANDS_ATTENTION on a window. Also, if the _NET_ACTIVE_WINDOW
hint includes a timestamp, use it.
(update_net_wm_state): Read DEMANDS_ATTENTION state also
* src/window.h (_MetaWindow): add wm_state_demands_attention bit.
2004-07-31 15:56:10 -04:00
2006-03-16 22:26:25 -05:00
if ( did_show & & window - > struts )
{
meta_topic ( META_DEBUG_WORKAREA ,
" Mapped window %s with struts, so invalidating work areas \n " ,
window - > desc ) ;
invalidate_work_areas ( window ) ;
2002-02-08 01:01:58 -05:00
}
2009-08-08 12:59:01 -04:00
2013-03-14 16:55:49 -04:00
if ( did_show )
meta_screen_queue_check_fullscreen ( window - > screen ) ;
2013-12-03 09:47:10 -05:00
if ( did_show & & window - > client_type = = META_WINDOW_CLIENT_TYPE_WAYLAND )
meta_wayland_compositor_repick ( meta_wayland_compositor_get_default ( ) ) ;
2009-08-08 12:59:01 -04:00
/*
* Now that we have shown the window , we no longer want to consider the
* initial timestamp in any subsequent deliberations whether to focus this
* window or not , so clear the flag .
*
* See http : //bugzilla.gnome.org/show_bug.cgi?id=573922
*/
window - > initial_timestamp_set = FALSE ;
2009-11-17 05:02:06 -05:00
if ( notify_demands_attention )
2009-10-28 13:19:24 -04:00
{
g_object_notify ( G_OBJECT ( window ) , " demands-attention " ) ;
g_signal_emit_by_name ( window - > display , " window-demands-attention " ,
window ) ;
}
2001-06-02 00:14:18 -04:00
}
2006-01-13 14:41:01 -05:00
static void
2001-06-02 00:14:18 -04:00
meta_window_hide ( MetaWindow * window )
{
2002-01-10 22:31:14 -05:00
gboolean did_hide ;
2008-10-24 05:07:24 -04:00
2002-02-07 22:34:26 -05:00
meta_topic ( META_DEBUG_WINDOW_STATE ,
" Hiding window %s \n " , window - > desc ) ;
2002-01-10 22:31:14 -05:00
Simplify relationship between mapping and visibility
Previously, changes to the visibility of a window could be indicated
by meta_compositor_map_window(), meta_compositor_unminimize_window(),
meta_compositor_set_window_hidden(), etc, with the exact behavior
depending on the 'live_hidden_windows' preference.
Simplify this so that visibility is controlled by:
meta_compositor_show_window()
meta_compositor_hide_window()
With an 'effect' parameter provided to indicate the appropriate
effect (CREATE/UNMINIMIZE/MINIMIZE/DESTROY/NONE.)
The map state of the window is signalled separately by:
meta_compositor_map_window()
meta_compositor_unmap_window()
And is used only to control resource handling.
Other changes:
* The desired effect on show/hide is explicitly stored in
MetaWindow, avoiding the need for the was_minimized flag.
At idle, once we calculate the window state, we pass the
effect to the compositor if it matches the new window
state, and then clear the effect to start over for future
map state changes.
* meta_compositor_switch_workspace() is called before any windows
are hidden or shown, allowing the compositor to avoid hiding
or showing an effect for windows involved in the switch.
http://bugzilla.gnome.org/show_bug.cgi?id=582341
* Handling of post-effect cleanups for MutterWindow are
simplified - instead of trying to do different things based
on the individual needs of different effects, we just wait until
all effects complete and sync the window state to what it
should be.
* On unmap, once we destroy the pixmap, we tell ClutterX11Pixmap
that we've done so, so it can clean up and unbind. (The
unbinding doesn't seem to be working properly because of
ClutterGLXPixmap or video driver issues.)
http://bugzilla.gnome.org/show_bug.cgi?id=587251
2009-06-28 17:10:40 -04:00
if ( window - > visible_to_compositor )
{
2014-03-18 17:06:42 -04:00
MetaCompEffect effect = META_COMP_EFFECT_NONE ;
2011-01-04 11:43:02 -05:00
window - > visible_to_compositor = FALSE ;
2014-03-18 17:06:42 -04:00
switch ( window - > pending_compositor_effect )
Simplify relationship between mapping and visibility
Previously, changes to the visibility of a window could be indicated
by meta_compositor_map_window(), meta_compositor_unminimize_window(),
meta_compositor_set_window_hidden(), etc, with the exact behavior
depending on the 'live_hidden_windows' preference.
Simplify this so that visibility is controlled by:
meta_compositor_show_window()
meta_compositor_hide_window()
With an 'effect' parameter provided to indicate the appropriate
effect (CREATE/UNMINIMIZE/MINIMIZE/DESTROY/NONE.)
The map state of the window is signalled separately by:
meta_compositor_map_window()
meta_compositor_unmap_window()
And is used only to control resource handling.
Other changes:
* The desired effect on show/hide is explicitly stored in
MetaWindow, avoiding the need for the was_minimized flag.
At idle, once we calculate the window state, we pass the
effect to the compositor if it matches the new window
state, and then clear the effect to start over for future
map state changes.
* meta_compositor_switch_workspace() is called before any windows
are hidden or shown, allowing the compositor to avoid hiding
or showing an effect for windows involved in the switch.
http://bugzilla.gnome.org/show_bug.cgi?id=582341
* Handling of post-effect cleanups for MutterWindow are
simplified - instead of trying to do different things based
on the individual needs of different effects, we just wait until
all effects complete and sync the window state to what it
should be.
* On unmap, once we destroy the pixmap, we tell ClutterX11Pixmap
that we've done so, so it can clean up and unbind. (The
unbinding doesn't seem to be working properly because of
ClutterGLXPixmap or video driver issues.)
http://bugzilla.gnome.org/show_bug.cgi?id=587251
2009-06-28 17:10:40 -04:00
{
2014-03-18 17:06:42 -04:00
case META_COMP_EFFECT_CREATE :
case META_COMP_EFFECT_UNMINIMIZE :
case META_COMP_EFFECT_NONE :
break ;
case META_COMP_EFFECT_DESTROY :
case META_COMP_EFFECT_MINIMIZE :
effect = window - > pending_compositor_effect ;
break ;
Simplify relationship between mapping and visibility
Previously, changes to the visibility of a window could be indicated
by meta_compositor_map_window(), meta_compositor_unminimize_window(),
meta_compositor_set_window_hidden(), etc, with the exact behavior
depending on the 'live_hidden_windows' preference.
Simplify this so that visibility is controlled by:
meta_compositor_show_window()
meta_compositor_hide_window()
With an 'effect' parameter provided to indicate the appropriate
effect (CREATE/UNMINIMIZE/MINIMIZE/DESTROY/NONE.)
The map state of the window is signalled separately by:
meta_compositor_map_window()
meta_compositor_unmap_window()
And is used only to control resource handling.
Other changes:
* The desired effect on show/hide is explicitly stored in
MetaWindow, avoiding the need for the was_minimized flag.
At idle, once we calculate the window state, we pass the
effect to the compositor if it matches the new window
state, and then clear the effect to start over for future
map state changes.
* meta_compositor_switch_workspace() is called before any windows
are hidden or shown, allowing the compositor to avoid hiding
or showing an effect for windows involved in the switch.
http://bugzilla.gnome.org/show_bug.cgi?id=582341
* Handling of post-effect cleanups for MutterWindow are
simplified - instead of trying to do different things based
on the individual needs of different effects, we just wait until
all effects complete and sync the window state to what it
should be.
* On unmap, once we destroy the pixmap, we tell ClutterX11Pixmap
that we've done so, so it can clean up and unbind. (The
unbinding doesn't seem to be working properly because of
ClutterGLXPixmap or video driver issues.)
http://bugzilla.gnome.org/show_bug.cgi?id=587251
2009-06-28 17:10:40 -04:00
}
2014-03-18 17:06:42 -04:00
meta_compositor_hide_window ( window - > display - > compositor , window , effect ) ;
Simplify relationship between mapping and visibility
Previously, changes to the visibility of a window could be indicated
by meta_compositor_map_window(), meta_compositor_unminimize_window(),
meta_compositor_set_window_hidden(), etc, with the exact behavior
depending on the 'live_hidden_windows' preference.
Simplify this so that visibility is controlled by:
meta_compositor_show_window()
meta_compositor_hide_window()
With an 'effect' parameter provided to indicate the appropriate
effect (CREATE/UNMINIMIZE/MINIMIZE/DESTROY/NONE.)
The map state of the window is signalled separately by:
meta_compositor_map_window()
meta_compositor_unmap_window()
And is used only to control resource handling.
Other changes:
* The desired effect on show/hide is explicitly stored in
MetaWindow, avoiding the need for the was_minimized flag.
At idle, once we calculate the window state, we pass the
effect to the compositor if it matches the new window
state, and then clear the effect to start over for future
map state changes.
* meta_compositor_switch_workspace() is called before any windows
are hidden or shown, allowing the compositor to avoid hiding
or showing an effect for windows involved in the switch.
http://bugzilla.gnome.org/show_bug.cgi?id=582341
* Handling of post-effect cleanups for MutterWindow are
simplified - instead of trying to do different things based
on the individual needs of different effects, we just wait until
all effects complete and sync the window state to what it
should be.
* On unmap, once we destroy the pixmap, we tell ClutterX11Pixmap
that we've done so, so it can clean up and unbind. (The
unbinding doesn't seem to be working properly because of
ClutterGLXPixmap or video driver issues.)
http://bugzilla.gnome.org/show_bug.cgi?id=587251
2009-06-28 17:10:40 -04:00
}
2002-01-10 22:31:14 -05:00
did_hide = FALSE ;
2008-10-24 05:07:24 -04:00
2012-02-29 22:15:12 -05:00
if ( ! window - > hidden )
2001-06-09 23:17:15 -04:00
{
2012-02-29 22:15:12 -05:00
meta_stack_freeze ( window - > screen - > stack ) ;
window - > hidden = TRUE ;
meta_stack_thaw ( window - > screen - > stack ) ;
did_hide = TRUE ;
2001-06-09 23:17:15 -04:00
}
2001-06-04 02:17:52 -04:00
2001-06-06 00:47:37 -04:00
if ( ! window - > iconic )
{
window - > iconic = TRUE ;
2014-02-01 19:03:33 -05:00
set_wm_state ( window ) ;
2001-06-06 00:47:37 -04:00
}
2008-12-18 07:41:56 -05:00
2006-03-16 22:26:25 -05:00
set_net_wm_state ( window ) ;
2008-12-18 07:41:56 -05:00
2006-03-16 22:26:25 -05:00
if ( did_hide & & window - > struts )
{
meta_topic ( META_DEBUG_WORKAREA ,
" Unmapped window %s with struts, so invalidating work areas \n " ,
window - > desc ) ;
invalidate_work_areas ( window ) ;
2002-01-10 22:31:14 -05:00
}
Simplify relationship between mapping and visibility
Previously, changes to the visibility of a window could be indicated
by meta_compositor_map_window(), meta_compositor_unminimize_window(),
meta_compositor_set_window_hidden(), etc, with the exact behavior
depending on the 'live_hidden_windows' preference.
Simplify this so that visibility is controlled by:
meta_compositor_show_window()
meta_compositor_hide_window()
With an 'effect' parameter provided to indicate the appropriate
effect (CREATE/UNMINIMIZE/MINIMIZE/DESTROY/NONE.)
The map state of the window is signalled separately by:
meta_compositor_map_window()
meta_compositor_unmap_window()
And is used only to control resource handling.
Other changes:
* The desired effect on show/hide is explicitly stored in
MetaWindow, avoiding the need for the was_minimized flag.
At idle, once we calculate the window state, we pass the
effect to the compositor if it matches the new window
state, and then clear the effect to start over for future
map state changes.
* meta_compositor_switch_workspace() is called before any windows
are hidden or shown, allowing the compositor to avoid hiding
or showing an effect for windows involved in the switch.
http://bugzilla.gnome.org/show_bug.cgi?id=582341
* Handling of post-effect cleanups for MutterWindow are
simplified - instead of trying to do different things based
on the individual needs of different effects, we just wait until
all effects complete and sync the window state to what it
should be.
* On unmap, once we destroy the pixmap, we tell ClutterX11Pixmap
that we've done so, so it can clean up and unbind. (The
unbinding doesn't seem to be working properly because of
ClutterGLXPixmap or video driver issues.)
http://bugzilla.gnome.org/show_bug.cgi?id=587251
2009-06-28 17:10:40 -04:00
2011-04-26 08:12:04 -04:00
if ( window - > has_focus )
Simplify relationship between mapping and visibility
Previously, changes to the visibility of a window could be indicated
by meta_compositor_map_window(), meta_compositor_unminimize_window(),
meta_compositor_set_window_hidden(), etc, with the exact behavior
depending on the 'live_hidden_windows' preference.
Simplify this so that visibility is controlled by:
meta_compositor_show_window()
meta_compositor_hide_window()
With an 'effect' parameter provided to indicate the appropriate
effect (CREATE/UNMINIMIZE/MINIMIZE/DESTROY/NONE.)
The map state of the window is signalled separately by:
meta_compositor_map_window()
meta_compositor_unmap_window()
And is used only to control resource handling.
Other changes:
* The desired effect on show/hide is explicitly stored in
MetaWindow, avoiding the need for the was_minimized flag.
At idle, once we calculate the window state, we pass the
effect to the compositor if it matches the new window
state, and then clear the effect to start over for future
map state changes.
* meta_compositor_switch_workspace() is called before any windows
are hidden or shown, allowing the compositor to avoid hiding
or showing an effect for windows involved in the switch.
http://bugzilla.gnome.org/show_bug.cgi?id=582341
* Handling of post-effect cleanups for MutterWindow are
simplified - instead of trying to do different things based
on the individual needs of different effects, we just wait until
all effects complete and sync the window state to what it
should be.
* On unmap, once we destroy the pixmap, we tell ClutterX11Pixmap
that we've done so, so it can clean up and unbind. (The
unbinding doesn't seem to be working properly because of
ClutterGLXPixmap or video driver issues.)
http://bugzilla.gnome.org/show_bug.cgi?id=587251
2009-06-28 17:10:40 -04:00
{
MetaWindow * not_this_one = NULL ;
MetaWorkspace * my_workspace = meta_window_get_workspace ( window ) ;
guint32 timestamp = meta_display_get_current_time_roundtrip ( window - > display ) ;
/*
* If this window is modal , passing the not_this_one window to
* _focus_default_window ( ) makes the focus to be given to this window ' s
* ancestor . This can only be the case if the window is on the currently
* active workspace ; when it is not , we need to pass in NULL , so as to
* focus the default window for the active workspace ( this scenario
* arises when we are switching workspaces ) .
2012-10-26 08:39:21 -04:00
* We also pass in NULL if we are in the process of hiding all non - desktop
* windows to avoid unexpected changes to the stacking order .
Simplify relationship between mapping and visibility
Previously, changes to the visibility of a window could be indicated
by meta_compositor_map_window(), meta_compositor_unminimize_window(),
meta_compositor_set_window_hidden(), etc, with the exact behavior
depending on the 'live_hidden_windows' preference.
Simplify this so that visibility is controlled by:
meta_compositor_show_window()
meta_compositor_hide_window()
With an 'effect' parameter provided to indicate the appropriate
effect (CREATE/UNMINIMIZE/MINIMIZE/DESTROY/NONE.)
The map state of the window is signalled separately by:
meta_compositor_map_window()
meta_compositor_unmap_window()
And is used only to control resource handling.
Other changes:
* The desired effect on show/hide is explicitly stored in
MetaWindow, avoiding the need for the was_minimized flag.
At idle, once we calculate the window state, we pass the
effect to the compositor if it matches the new window
state, and then clear the effect to start over for future
map state changes.
* meta_compositor_switch_workspace() is called before any windows
are hidden or shown, allowing the compositor to avoid hiding
or showing an effect for windows involved in the switch.
http://bugzilla.gnome.org/show_bug.cgi?id=582341
* Handling of post-effect cleanups for MutterWindow are
simplified - instead of trying to do different things based
on the individual needs of different effects, we just wait until
all effects complete and sync the window state to what it
should be.
* On unmap, once we destroy the pixmap, we tell ClutterX11Pixmap
that we've done so, so it can clean up and unbind. (The
unbinding doesn't seem to be working properly because of
ClutterGLXPixmap or video driver issues.)
http://bugzilla.gnome.org/show_bug.cgi?id=587251
2009-06-28 17:10:40 -04:00
*/
2012-10-26 08:39:21 -04:00
if ( my_workspace = = window - > screen - > active_workspace & &
! my_workspace - > showing_desktop )
Simplify relationship between mapping and visibility
Previously, changes to the visibility of a window could be indicated
by meta_compositor_map_window(), meta_compositor_unminimize_window(),
meta_compositor_set_window_hidden(), etc, with the exact behavior
depending on the 'live_hidden_windows' preference.
Simplify this so that visibility is controlled by:
meta_compositor_show_window()
meta_compositor_hide_window()
With an 'effect' parameter provided to indicate the appropriate
effect (CREATE/UNMINIMIZE/MINIMIZE/DESTROY/NONE.)
The map state of the window is signalled separately by:
meta_compositor_map_window()
meta_compositor_unmap_window()
And is used only to control resource handling.
Other changes:
* The desired effect on show/hide is explicitly stored in
MetaWindow, avoiding the need for the was_minimized flag.
At idle, once we calculate the window state, we pass the
effect to the compositor if it matches the new window
state, and then clear the effect to start over for future
map state changes.
* meta_compositor_switch_workspace() is called before any windows
are hidden or shown, allowing the compositor to avoid hiding
or showing an effect for windows involved in the switch.
http://bugzilla.gnome.org/show_bug.cgi?id=582341
* Handling of post-effect cleanups for MutterWindow are
simplified - instead of trying to do different things based
on the individual needs of different effects, we just wait until
all effects complete and sync the window state to what it
should be.
* On unmap, once we destroy the pixmap, we tell ClutterX11Pixmap
that we've done so, so it can clean up and unbind. (The
unbinding doesn't seem to be working properly because of
ClutterGLXPixmap or video driver issues.)
http://bugzilla.gnome.org/show_bug.cgi?id=587251
2009-06-28 17:10:40 -04:00
not_this_one = window ;
meta_workspace_focus_default_window ( window - > screen - > active_workspace ,
not_this_one ,
timestamp ) ;
}
2013-03-14 16:55:49 -04:00
if ( did_hide )
meta_screen_queue_check_fullscreen ( window - > screen ) ;
2001-06-04 02:17:52 -04:00
}
2003-03-17 01:36:39 -05:00
static gboolean
on unminimize, queue calc_showing on all transients
2002-05-05 Havoc Pennington <hp@pobox.com>
* src/window.c (meta_window_unminimize): on unminimize, queue
calc_showing on all transients
(meta_window_activate): on activate, unminimize all a window's
ancestors, not just the window itself.
* src/workspace.c (set_work_area_hint): don't increment "tmp" by
16 unsigned long, increment by 4
* src/window.c (meta_window_free): if a window isn't minimized,
restore its WM_STATE to NormalState instead of IconicState,
since IconicState on initial window map means that the window
should be minimized.
* src/workspace.c (meta_workspace_invalidate_work_area): queue an
idle to recompute the work area hint.
(set_work_area_hint): we need 4*num_workspaces ints, not just
num_workspaces.
* src/screen.c (meta_screen_new): add work_area_idle field,
handle it on screen shutdown
* src/common.h (META_PRIORITY_PREFS_NOTIFY,
META_PRIORITY_WORK_AREA_HINT): define some idle priorities
* src/window.c (meta_window_calc_showing): hide windows if
their parent window is minimized
(meta_window_minimize): also queue_calc_showing on all
transients of the window being minimized
* src/place.c (constrain_placement): function to apply
placement-time-only constraints, such as "not off the left of the
screen"
(meta_window_place): put dialogs down a bit over their parent,
not right at the top.
(meta_window_place): when centering a dialog, center it
on the current xinerama screen, rather than the entire
screen.
* src/screen.c (meta_screen_get_current_xinerama): new function,
but not implemented
2002-05-05 01:41:13 -04:00
queue_calc_showing_func ( MetaWindow * window ,
void * data )
{
Refactor thrice-duplicated queue code in window.c. Closes #376760.
2007-06-10 Thomas Thurman <thomas@thurman.org.uk>
Refactor thrice-duplicated queue code in window.c. Closes #376760.
* src/window.c (meta_window_queue, meta_window_unqueue):
New functions.
* src/window.[ch] (meta_window_unqueue_*, meta_window_queue_*):
Removed functions.
* src/window.c (meta_window_new_with_attrs, meta_window_free,
meta_window_flush_calc_showing, queue_calc_showing_func,
meta_window_minimize, meta_window_unminimize, meta_window_maximize,
meta_window_make_fullscreen, meta_window_shade,
meta_window_unshade, meta_window_move_resize_internal,
window_stick_impl, window_unstick_impl,
meta_window_client_message, process_property_notify): Modified to
use new queueing functions.
* src/window.c (idle_move_resize, idle_update_icon,
idle_calc_showing): update to receive queue number from pointer.
* src/window.h (MetaQueueType): new enum.
* src/window.h (MetaWindow): *_queued replaced with is_in_queue
bitfield.
* src/core.c (meta_core_queue_frame_resize):
* src/display.c (event_callback,
meta_display_queue_retheme_all_windows): Using new queueing functions.
* src/frame.c (meta_window_destroy_frame): Using new queueing functions.
* src/screen.c (queue_resize, meta_screen_resize_func,
queue_windows_showing): Using new queueing functions.
* src/window-props.c (reload_mwm_hints, reload_wm_hints,
reload_transient_for): Using new queueing functions.
* src/workspace.c (meta_workspace_add_window,
meta_workspace_remove_window, meta_workspace_queue_calc_showing,
meta_workspace_invalidate_work_area): Using new queueing functions.
svn path=/trunk/; revision=3236
2007-06-10 21:15:33 -04:00
meta_window_queue ( window , META_QUEUE_CALC_SHOWING ) ;
2003-03-17 01:36:39 -05:00
return TRUE ;
on unminimize, queue calc_showing on all transients
2002-05-05 Havoc Pennington <hp@pobox.com>
* src/window.c (meta_window_unminimize): on unminimize, queue
calc_showing on all transients
(meta_window_activate): on activate, unminimize all a window's
ancestors, not just the window itself.
* src/workspace.c (set_work_area_hint): don't increment "tmp" by
16 unsigned long, increment by 4
* src/window.c (meta_window_free): if a window isn't minimized,
restore its WM_STATE to NormalState instead of IconicState,
since IconicState on initial window map means that the window
should be minimized.
* src/workspace.c (meta_workspace_invalidate_work_area): queue an
idle to recompute the work area hint.
(set_work_area_hint): we need 4*num_workspaces ints, not just
num_workspaces.
* src/screen.c (meta_screen_new): add work_area_idle field,
handle it on screen shutdown
* src/common.h (META_PRIORITY_PREFS_NOTIFY,
META_PRIORITY_WORK_AREA_HINT): define some idle priorities
* src/window.c (meta_window_calc_showing): hide windows if
their parent window is minimized
(meta_window_minimize): also queue_calc_showing on all
transients of the window being minimized
* src/place.c (constrain_placement): function to apply
placement-time-only constraints, such as "not off the left of the
screen"
(meta_window_place): put dialogs down a bit over their parent,
not right at the top.
(meta_window_place): when centering a dialog, center it
on the current xinerama screen, rather than the entire
screen.
* src/screen.c (meta_screen_get_current_xinerama): new function,
but not implemented
2002-05-05 01:41:13 -04:00
}
2001-06-04 02:17:52 -04:00
void
meta_window_minimize ( MetaWindow * window )
{
2009-06-14 08:11:29 -04:00
g_return_if_fail ( ! window - > override_redirect ) ;
2001-06-04 02:17:52 -04:00
if ( ! window - > minimized )
{
window - > minimized = TRUE ;
Simplify relationship between mapping and visibility
Previously, changes to the visibility of a window could be indicated
by meta_compositor_map_window(), meta_compositor_unminimize_window(),
meta_compositor_set_window_hidden(), etc, with the exact behavior
depending on the 'live_hidden_windows' preference.
Simplify this so that visibility is controlled by:
meta_compositor_show_window()
meta_compositor_hide_window()
With an 'effect' parameter provided to indicate the appropriate
effect (CREATE/UNMINIMIZE/MINIMIZE/DESTROY/NONE.)
The map state of the window is signalled separately by:
meta_compositor_map_window()
meta_compositor_unmap_window()
And is used only to control resource handling.
Other changes:
* The desired effect on show/hide is explicitly stored in
MetaWindow, avoiding the need for the was_minimized flag.
At idle, once we calculate the window state, we pass the
effect to the compositor if it matches the new window
state, and then clear the effect to start over for future
map state changes.
* meta_compositor_switch_workspace() is called before any windows
are hidden or shown, allowing the compositor to avoid hiding
or showing an effect for windows involved in the switch.
http://bugzilla.gnome.org/show_bug.cgi?id=582341
* Handling of post-effect cleanups for MutterWindow are
simplified - instead of trying to do different things based
on the individual needs of different effects, we just wait until
all effects complete and sync the window state to what it
should be.
* On unmap, once we destroy the pixmap, we tell ClutterX11Pixmap
that we've done so, so it can clean up and unbind. (The
unbinding doesn't seem to be working properly because of
ClutterGLXPixmap or video driver issues.)
http://bugzilla.gnome.org/show_bug.cgi?id=587251
2009-06-28 17:10:40 -04:00
window - > pending_compositor_effect = META_COMP_EFFECT_MINIMIZE ;
Refactor thrice-duplicated queue code in window.c. Closes #376760.
2007-06-10 Thomas Thurman <thomas@thurman.org.uk>
Refactor thrice-duplicated queue code in window.c. Closes #376760.
* src/window.c (meta_window_queue, meta_window_unqueue):
New functions.
* src/window.[ch] (meta_window_unqueue_*, meta_window_queue_*):
Removed functions.
* src/window.c (meta_window_new_with_attrs, meta_window_free,
meta_window_flush_calc_showing, queue_calc_showing_func,
meta_window_minimize, meta_window_unminimize, meta_window_maximize,
meta_window_make_fullscreen, meta_window_shade,
meta_window_unshade, meta_window_move_resize_internal,
window_stick_impl, window_unstick_impl,
meta_window_client_message, process_property_notify): Modified to
use new queueing functions.
* src/window.c (idle_move_resize, idle_update_icon,
idle_calc_showing): update to receive queue number from pointer.
* src/window.h (MetaQueueType): new enum.
* src/window.h (MetaWindow): *_queued replaced with is_in_queue
bitfield.
* src/core.c (meta_core_queue_frame_resize):
* src/display.c (event_callback,
meta_display_queue_retheme_all_windows): Using new queueing functions.
* src/frame.c (meta_window_destroy_frame): Using new queueing functions.
* src/screen.c (queue_resize, meta_screen_resize_func,
queue_windows_showing): Using new queueing functions.
* src/window-props.c (reload_mwm_hints, reload_wm_hints,
reload_transient_for): Using new queueing functions.
* src/workspace.c (meta_workspace_add_window,
meta_workspace_remove_window, meta_workspace_queue_calc_showing,
meta_workspace_invalidate_work_area): Using new queueing functions.
svn path=/trunk/; revision=3236
2007-06-10 21:15:33 -04:00
meta_window_queue ( window , META_QUEUE_CALC_SHOWING ) ;
on unminimize, queue calc_showing on all transients
2002-05-05 Havoc Pennington <hp@pobox.com>
* src/window.c (meta_window_unminimize): on unminimize, queue
calc_showing on all transients
(meta_window_activate): on activate, unminimize all a window's
ancestors, not just the window itself.
* src/workspace.c (set_work_area_hint): don't increment "tmp" by
16 unsigned long, increment by 4
* src/window.c (meta_window_free): if a window isn't minimized,
restore its WM_STATE to NormalState instead of IconicState,
since IconicState on initial window map means that the window
should be minimized.
* src/workspace.c (meta_workspace_invalidate_work_area): queue an
idle to recompute the work area hint.
(set_work_area_hint): we need 4*num_workspaces ints, not just
num_workspaces.
* src/screen.c (meta_screen_new): add work_area_idle field,
handle it on screen shutdown
* src/common.h (META_PRIORITY_PREFS_NOTIFY,
META_PRIORITY_WORK_AREA_HINT): define some idle priorities
* src/window.c (meta_window_calc_showing): hide windows if
their parent window is minimized
(meta_window_minimize): also queue_calc_showing on all
transients of the window being minimized
* src/place.c (constrain_placement): function to apply
placement-time-only constraints, such as "not off the left of the
screen"
(meta_window_place): put dialogs down a bit over their parent,
not right at the top.
(meta_window_place): when centering a dialog, center it
on the current xinerama screen, rather than the entire
screen.
* src/screen.c (meta_screen_get_current_xinerama): new function,
but not implemented
2002-05-05 01:41:13 -04:00
meta_window_foreach_transient ( window ,
queue_calc_showing_func ,
NULL ) ;
2008-12-18 07:41:56 -05:00
2002-01-03 18:28:19 -05:00
if ( window - > has_focus )
{
meta_topic ( META_DEBUG_FOCUS ,
2004-06-24 16:02:46 -04:00
" Focusing default window due to minimization of focus window %s \n " ,
2002-01-03 18:28:19 -05:00
window - > desc ) ;
}
else
{
meta_topic ( META_DEBUG_FOCUS ,
" Minimizing window %s which doesn't have the focus \n " ,
window - > desc ) ;
}
2011-05-31 13:52:04 -04:00
g_object_notify ( G_OBJECT ( window ) , " minimized " ) ;
2001-06-04 02:17:52 -04:00
}
}
void
meta_window_unminimize ( MetaWindow * window )
{
2009-06-14 08:11:29 -04:00
g_return_if_fail ( ! window - > override_redirect ) ;
2001-06-04 02:17:52 -04:00
if ( window - > minimized )
{
window - > minimized = FALSE ;
Simplify relationship between mapping and visibility
Previously, changes to the visibility of a window could be indicated
by meta_compositor_map_window(), meta_compositor_unminimize_window(),
meta_compositor_set_window_hidden(), etc, with the exact behavior
depending on the 'live_hidden_windows' preference.
Simplify this so that visibility is controlled by:
meta_compositor_show_window()
meta_compositor_hide_window()
With an 'effect' parameter provided to indicate the appropriate
effect (CREATE/UNMINIMIZE/MINIMIZE/DESTROY/NONE.)
The map state of the window is signalled separately by:
meta_compositor_map_window()
meta_compositor_unmap_window()
And is used only to control resource handling.
Other changes:
* The desired effect on show/hide is explicitly stored in
MetaWindow, avoiding the need for the was_minimized flag.
At idle, once we calculate the window state, we pass the
effect to the compositor if it matches the new window
state, and then clear the effect to start over for future
map state changes.
* meta_compositor_switch_workspace() is called before any windows
are hidden or shown, allowing the compositor to avoid hiding
or showing an effect for windows involved in the switch.
http://bugzilla.gnome.org/show_bug.cgi?id=582341
* Handling of post-effect cleanups for MutterWindow are
simplified - instead of trying to do different things based
on the individual needs of different effects, we just wait until
all effects complete and sync the window state to what it
should be.
* On unmap, once we destroy the pixmap, we tell ClutterX11Pixmap
that we've done so, so it can clean up and unbind. (The
unbinding doesn't seem to be working properly because of
ClutterGLXPixmap or video driver issues.)
http://bugzilla.gnome.org/show_bug.cgi?id=587251
2009-06-28 17:10:40 -04:00
window - > pending_compositor_effect = META_COMP_EFFECT_UNMINIMIZE ;
Refactor thrice-duplicated queue code in window.c. Closes #376760.
2007-06-10 Thomas Thurman <thomas@thurman.org.uk>
Refactor thrice-duplicated queue code in window.c. Closes #376760.
* src/window.c (meta_window_queue, meta_window_unqueue):
New functions.
* src/window.[ch] (meta_window_unqueue_*, meta_window_queue_*):
Removed functions.
* src/window.c (meta_window_new_with_attrs, meta_window_free,
meta_window_flush_calc_showing, queue_calc_showing_func,
meta_window_minimize, meta_window_unminimize, meta_window_maximize,
meta_window_make_fullscreen, meta_window_shade,
meta_window_unshade, meta_window_move_resize_internal,
window_stick_impl, window_unstick_impl,
meta_window_client_message, process_property_notify): Modified to
use new queueing functions.
* src/window.c (idle_move_resize, idle_update_icon,
idle_calc_showing): update to receive queue number from pointer.
* src/window.h (MetaQueueType): new enum.
* src/window.h (MetaWindow): *_queued replaced with is_in_queue
bitfield.
* src/core.c (meta_core_queue_frame_resize):
* src/display.c (event_callback,
meta_display_queue_retheme_all_windows): Using new queueing functions.
* src/frame.c (meta_window_destroy_frame): Using new queueing functions.
* src/screen.c (queue_resize, meta_screen_resize_func,
queue_windows_showing): Using new queueing functions.
* src/window-props.c (reload_mwm_hints, reload_wm_hints,
reload_transient_for): Using new queueing functions.
* src/workspace.c (meta_workspace_add_window,
meta_workspace_remove_window, meta_workspace_queue_calc_showing,
meta_workspace_invalidate_work_area): Using new queueing functions.
svn path=/trunk/; revision=3236
2007-06-10 21:15:33 -04:00
meta_window_queue ( window , META_QUEUE_CALC_SHOWING ) ;
on unminimize, queue calc_showing on all transients
2002-05-05 Havoc Pennington <hp@pobox.com>
* src/window.c (meta_window_unminimize): on unminimize, queue
calc_showing on all transients
(meta_window_activate): on activate, unminimize all a window's
ancestors, not just the window itself.
* src/workspace.c (set_work_area_hint): don't increment "tmp" by
16 unsigned long, increment by 4
* src/window.c (meta_window_free): if a window isn't minimized,
restore its WM_STATE to NormalState instead of IconicState,
since IconicState on initial window map means that the window
should be minimized.
* src/workspace.c (meta_workspace_invalidate_work_area): queue an
idle to recompute the work area hint.
(set_work_area_hint): we need 4*num_workspaces ints, not just
num_workspaces.
* src/screen.c (meta_screen_new): add work_area_idle field,
handle it on screen shutdown
* src/common.h (META_PRIORITY_PREFS_NOTIFY,
META_PRIORITY_WORK_AREA_HINT): define some idle priorities
* src/window.c (meta_window_calc_showing): hide windows if
their parent window is minimized
(meta_window_minimize): also queue_calc_showing on all
transients of the window being minimized
* src/place.c (constrain_placement): function to apply
placement-time-only constraints, such as "not off the left of the
screen"
(meta_window_place): put dialogs down a bit over their parent,
not right at the top.
(meta_window_place): when centering a dialog, center it
on the current xinerama screen, rather than the entire
screen.
* src/screen.c (meta_screen_get_current_xinerama): new function,
but not implemented
2002-05-05 01:41:13 -04:00
meta_window_foreach_transient ( window ,
queue_calc_showing_func ,
NULL ) ;
2011-05-31 13:52:04 -04:00
g_object_notify ( G_OBJECT ( window ) , " minimized " ) ;
2001-06-07 01:18:10 -04:00
}
}
2007-04-04 17:54:56 -04:00
static void
ensure_size_hints_satisfied ( MetaRectangle * rect ,
const XSizeHints * size_hints )
{
int minw , minh , maxw , maxh ; /* min/max width/height */
int basew , baseh , winc , hinc ; /* base width/height, width/height increment */
int extra_width , extra_height ;
minw = size_hints - > min_width ; minh = size_hints - > min_height ;
maxw = size_hints - > max_width ; maxh = size_hints - > max_height ;
basew = size_hints - > base_width ; baseh = size_hints - > base_height ;
winc = size_hints - > width_inc ; hinc = size_hints - > height_inc ;
2008-12-18 07:41:56 -05:00
2007-04-04 17:54:56 -04:00
/* First, enforce min/max size constraints */
rect - > width = CLAMP ( rect - > width , minw , maxw ) ;
rect - > height = CLAMP ( rect - > height , minh , maxh ) ;
/* Now, verify size increment constraints are satisfied, or make them be */
extra_width = ( rect - > width - basew ) % winc ;
extra_height = ( rect - > height - baseh ) % hinc ;
rect - > width - = extra_width ;
rect - > height - = extra_height ;
/* Adjusting width/height down, as done above, may violate minimum size
* constraints , so one last fix .
*/
if ( rect - > width < minw )
rect - > width + = ( ( minw - rect - > width ) / winc + 1 ) * winc ;
if ( rect - > height < minh )
rect - > height + = ( ( minh - rect - > height ) / hinc + 1 ) * hinc ;
}
2002-09-24 17:33:34 -04:00
static void
meta_window_save_rect ( MetaWindow * window )
{
2010-09-24 14:31:36 -04:00
if ( ! ( META_WINDOW_MAXIMIZED ( window ) | | META_WINDOW_TILED_SIDE_BY_SIDE ( window ) | | window - > fullscreen ) )
2002-09-24 17:33:34 -04:00
{
/* save size/pos as appropriate args for move_resize */
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
if ( ! window - > maximized_horizontally )
2002-09-24 17:33:34 -04:00
{
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
window - > saved_rect . x = window - > rect . x ;
window - > saved_rect . width = window - > rect . width ;
if ( window - > frame )
window - > saved_rect . x + = window - > frame - > rect . x ;
}
if ( ! window - > maximized_vertically )
{
window - > saved_rect . y = window - > rect . y ;
window - > saved_rect . height = window - > rect . height ;
if ( window - > frame )
window - > saved_rect . y + = window - > frame - > rect . y ;
2002-09-24 17:33:34 -04:00
}
}
}
2008-08-15 23:09:40 -04:00
/**
2011-11-02 11:34:45 -04:00
* force_save_user_window_placement :
* @ window : Store current position of this window for future reference
*
2008-08-15 23:09:40 -04:00
* Save the user_rect regardless of whether the window is maximized or
* fullscreen . See save_user_window_placement ( ) for most uses .
*/
static void
force_save_user_window_placement ( MetaWindow * window )
{
meta_window_get_client_root_coords ( window , & window - > user_rect ) ;
}
/**
2011-11-02 11:34:45 -04:00
* save_user_window_placement :
* @ window : Store current position of this window for future reference
*
2008-08-15 23:09:40 -04:00
* Save the user_rect , but only if the window is neither maximized nor
* fullscreen , otherwise the window may snap back to those dimensions
* ( bug # 461927 ) .
*/
2007-11-13 08:31:14 -05:00
static void
2008-08-15 23:09:40 -04:00
save_user_window_placement ( MetaWindow * window )
2007-11-13 08:31:14 -05:00
{
2010-09-24 14:31:36 -04:00
if ( ! ( META_WINDOW_MAXIMIZED ( window ) | | META_WINDOW_TILED_SIDE_BY_SIDE ( window ) | | window - > fullscreen ) )
2007-11-13 08:31:14 -05:00
{
MetaRectangle user_rect ;
meta_window_get_client_root_coords ( window , & user_rect ) ;
if ( ! window - > maximized_horizontally )
{
window - > user_rect . x = user_rect . x ;
window - > user_rect . width = user_rect . width ;
}
if ( ! window - > maximized_vertically )
{
window - > user_rect . y = user_rect . y ;
window - > user_rect . height = user_rect . height ;
}
}
}
2003-06-09 19:49:02 -04:00
void
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
meta_window_maximize_internal ( MetaWindow * window ,
MetaMaximizeFlags directions ,
MetaRectangle * saved_rect )
2003-06-09 19:49:02 -04:00
{
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
/* At least one of the two directions ought to be set */
gboolean maximize_horizontally , maximize_vertically ;
maximize_horizontally = directions & META_MAXIMIZE_HORIZONTAL ;
maximize_vertically = directions & META_MAXIMIZE_VERTICAL ;
g_assert ( maximize_horizontally | | maximize_vertically ) ;
2003-06-09 19:49:02 -04:00
meta_topic ( META_DEBUG_WINDOW_OPS ,
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
" Maximizing %s%s \n " ,
window - > desc ,
maximize_horizontally & & maximize_vertically ? " " :
maximize_horizontally ? " horizontally " :
maximize_vertically ? " vertically " : " BUGGGGG " ) ;
2008-12-18 07:41:56 -05:00
2003-06-09 19:49:02 -04:00
if ( saved_rect ! = NULL )
window - > saved_rect = * saved_rect ;
else
meta_window_save_rect ( window ) ;
2008-12-18 07:41:56 -05:00
2011-12-05 15:52:46 -05:00
if ( maximize_horizontally & & maximize_vertically )
window - > saved_maximize = TRUE ;
2008-12-18 07:41:56 -05:00
window - > maximized_horizontally =
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
window - > maximized_horizontally | | maximize_horizontally ;
2008-12-18 07:41:56 -05:00
window - > maximized_vertically =
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
window - > maximized_vertically | | maximize_vertically ;
2010-01-20 10:59:07 -05:00
if ( maximize_horizontally | | maximize_vertically )
window - > force_save_user_rect = FALSE ;
2006-07-31 15:56:56 -04:00
2014-02-01 18:58:58 -05:00
meta_window_recalc_features ( window ) ;
2003-06-09 19:49:02 -04:00
set_net_wm_state ( window ) ;
2010-02-12 15:26:52 -05:00
2014-02-09 11:45:09 -05:00
if ( window - > surface & & window - > maximized_horizontally & & window - > maximized_vertically )
meta_wayland_surface_send_maximized ( window - > surface ) ;
2010-02-12 15:26:52 -05:00
g_object_freeze_notify ( G_OBJECT ( window ) ) ;
g_object_notify ( G_OBJECT ( window ) , " maximized-horizontally " ) ;
g_object_notify ( G_OBJECT ( window ) , " maximized-vertically " ) ;
g_object_thaw_notify ( G_OBJECT ( window ) ) ;
2003-06-09 19:49:02 -04:00
}
2001-06-07 01:18:10 -04:00
void
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
meta_window_maximize ( MetaWindow * window ,
MetaMaximizeFlags directions )
2001-06-07 01:18:10 -04:00
{
2010-06-24 14:41:28 -04:00
MetaRectangle * saved_rect = NULL ;
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
gboolean maximize_horizontally , maximize_vertically ;
2014-03-18 17:06:42 -04:00
MetaRectangle old_rect ;
MetaRectangle new_rect ;
2009-06-14 08:11:29 -04:00
g_return_if_fail ( ! window - > override_redirect ) ;
/* At least one of the two directions ought to be set */
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
maximize_horizontally = directions & META_MAXIMIZE_HORIZONTAL ;
maximize_vertically = directions & META_MAXIMIZE_VERTICAL ;
g_assert ( maximize_horizontally | | maximize_vertically ) ;
/* Only do something if the window isn't already maximized in the
* given direction ( s ) .
*/
if ( ( maximize_horizontally & & ! window - > maximized_horizontally ) | |
( maximize_vertically & & ! window - > maximized_vertically ) )
2001-06-07 01:18:10 -04:00
{
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
if ( window - > shaded & & maximize_vertically )
Partial audit to fix timestamp usage. One step towards fixing #355180; see
2006-09-18 Elijah Newren <newren gmail com>
Partial audit to fix timestamp usage. One step towards fixing
#355180; see important comments in that bug.
* src/core.[ch] (meta_core_unshade, meta_core_shade):
* src/delete.c (meta_window_present_delete_dialog,
delete_ping_timeout_func):
* src/display.[ch] (meta_display_open, meta_display_close,
event_callback, meta_display_begin_grab_op,
process_selection_clear, meta_display_unmanage_screen,
meta_display_unmanage_windows_for_screen):
* src/frames.c (meta_frames_button_press_event):
* src/keybindings.c (handle_toggle_shade):
* src/main.c (main):
* src/screen.[ch] (update_num_workspaces, meta_screen_new,
meta_screen_free, prefs_changed_callback):
* src/window.[ch] (meta_window_free, finish_minimize,
implement_showing, meta_window_show, meta_window_maximize,
meta_window_make_fullscreen_internal,
meta_window_unmake_fullscreen, meta_window_shade,
meta_window_unshade, window_activate, send_sync_request,
meta_window_client_message, menu_callback,
meta_window_update_keyboard_resize):
Remove usage of CurrentTime, meta_display_get_current_time() and
meta_display_get_current_time_roundtrip() where possible, or
document why it isn't possible, or at very least add a FIXME with
some explanation of my laziness and what needs to be done.
2006-09-18 13:27:24 -04:00
{
/* Shading sucks anyway; I'm not adding a timestamp argument
* to this function just for this niche usage & corner case .
*/
2008-12-18 07:41:56 -05:00
guint32 timestamp =
Partial audit to fix timestamp usage. One step towards fixing #355180; see
2006-09-18 Elijah Newren <newren gmail com>
Partial audit to fix timestamp usage. One step towards fixing
#355180; see important comments in that bug.
* src/core.[ch] (meta_core_unshade, meta_core_shade):
* src/delete.c (meta_window_present_delete_dialog,
delete_ping_timeout_func):
* src/display.[ch] (meta_display_open, meta_display_close,
event_callback, meta_display_begin_grab_op,
process_selection_clear, meta_display_unmanage_screen,
meta_display_unmanage_windows_for_screen):
* src/frames.c (meta_frames_button_press_event):
* src/keybindings.c (handle_toggle_shade):
* src/main.c (main):
* src/screen.[ch] (update_num_workspaces, meta_screen_new,
meta_screen_free, prefs_changed_callback):
* src/window.[ch] (meta_window_free, finish_minimize,
implement_showing, meta_window_show, meta_window_maximize,
meta_window_make_fullscreen_internal,
meta_window_unmake_fullscreen, meta_window_shade,
meta_window_unshade, window_activate, send_sync_request,
meta_window_client_message, menu_callback,
meta_window_update_keyboard_resize):
Remove usage of CurrentTime, meta_display_get_current_time() and
meta_display_get_current_time_roundtrip() where possible, or
document why it isn't possible, or at very least add a FIXME with
some explanation of my laziness and what needs to be done.
2006-09-18 13:27:24 -04:00
meta_display_get_current_time_roundtrip ( window - > display ) ;
meta_window_unshade ( window , timestamp ) ;
}
2008-12-18 07:41:56 -05:00
2003-06-09 19:49:02 -04:00
/* if the window hasn't been placed yet, we'll maximize it then
*/
if ( ! window - > placed )
{
2008-12-18 07:41:56 -05:00
window - > maximize_horizontally_after_placement =
window - > maximize_horizontally_after_placement | |
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
maximize_horizontally ;
2008-12-18 07:41:56 -05:00
window - > maximize_vertically_after_placement =
window - > maximize_vertically_after_placement | |
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
maximize_vertically ;
2003-06-09 19:49:02 -04:00
return ;
}
2010-06-24 14:41:28 -04:00
if ( window - > tile_mode ! = META_TILE_NONE )
{
saved_rect = & window - > saved_rect ;
window - > maximized_vertically = FALSE ;
2013-03-08 15:37:53 -05:00
window - > tile_mode = META_TILE_NONE ;
2010-06-24 14:41:28 -04:00
}
2008-12-18 07:41:56 -05:00
meta_window_maximize_internal ( window ,
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
directions ,
2010-06-24 14:41:28 -04:00
saved_rect ) ;
2001-10-14 18:13:53 -04:00
2014-03-18 17:06:42 -04:00
meta_window_get_frame_rect ( window , & old_rect ) ;
2008-11-23 14:28:40 -05:00
2014-03-18 17:06:42 -04:00
meta_window_move_resize_now ( window ) ;
2008-09-18 11:09:11 -04:00
2014-03-18 17:06:42 -04:00
meta_window_get_frame_rect ( window , & new_rect ) ;
meta_compositor_maximize_window ( window - > display - > compositor ,
window ,
& old_rect ,
& new_rect ) ;
2001-06-07 01:18:10 -04:00
}
}
2010-11-11 18:42:44 -05:00
/**
* meta_window_get_maximized :
2013-02-15 13:42:08 -05:00
* @ window : a # MetaWindow
2010-11-11 18:42:44 -05:00
*
* Gets the current maximization state of the window , as combination
* of the % META_MAXIMIZE_HORIZONTAL and % META_MAXIMIZE_VERTICAL flags ;
*
* Return value : current maximization state
*/
MetaMaximizeFlags
meta_window_get_maximized ( MetaWindow * window )
{
return ( ( window - > maximized_horizontally ? META_MAXIMIZE_HORIZONTAL : 0 ) |
( window - > maximized_vertically ? META_MAXIMIZE_VERTICAL : 0 ) ) ;
}
/**
* meta_window_is_fullscreen :
2013-02-15 13:42:08 -05:00
* @ window : a # MetaWindow
2010-11-11 18:42:44 -05:00
*
* Return value : % TRUE if the window is currently fullscreen
*/
gboolean
meta_window_is_fullscreen ( MetaWindow * window )
{
return window - > fullscreen ;
}
2013-02-23 15:56:17 -05:00
/**
* meta_window_get_all_monitors :
* @ window : The # MetaWindow
2013-02-27 11:50:48 -05:00
* @ length : ( out ) : gint holding the length , may be % NULL to ignore
2013-02-23 15:56:17 -05:00
*
* Returns : ( array length = length ) ( element - type gint ) ( transfer container ) :
* List of the monitor indices the window is on .
*/
gint *
meta_window_get_all_monitors ( MetaWindow * window , gsize * length )
{
GArray * monitors ;
MetaRectangle window_rect ;
int i ;
monitors = g_array_new ( FALSE , FALSE , sizeof ( int ) ) ;
2013-11-19 12:24:45 -05:00
meta_window_get_frame_rect ( window , & window_rect ) ;
2013-02-23 15:56:17 -05:00
for ( i = 0 ; i < window - > screen - > n_monitor_infos ; i + + )
{
MetaRectangle * monitor_rect = & window - > screen - > monitor_infos [ i ] . rect ;
if ( meta_rectangle_overlap ( & window_rect , monitor_rect ) )
g_array_append_val ( monitors , i ) ;
}
if ( length )
* length = monitors - > len ;
i = - 1 ;
g_array_append_val ( monitors , i ) ;
return ( gint * ) g_array_free ( monitors , FALSE ) ;
}
2013-02-18 07:30:37 -05:00
/**
* meta_window_is_screen_sized :
2013-02-15 13:42:08 -05:00
* @ window : A # MetaWindow
2013-02-18 07:30:37 -05:00
*
* Return value : % TRUE if the window is occupies the
* the whole screen ( all monitors ) .
*/
gboolean
meta_window_is_screen_sized ( MetaWindow * window )
{
MetaRectangle window_rect ;
int screen_width , screen_height ;
meta_screen_get_size ( window - > screen , & screen_width , & screen_height ) ;
2013-11-19 12:24:45 -05:00
meta_window_get_frame_rect ( window , & window_rect ) ;
2013-02-18 07:30:37 -05:00
if ( window_rect . x = = 0 & & window_rect . y = = 0 & &
window_rect . width = = screen_width & & window_rect . height = = screen_height )
return TRUE ;
return FALSE ;
}
2013-02-17 15:12:45 -05:00
/**
* meta_window_is_monitor_sized :
2013-02-18 08:19:28 -05:00
* @ window : a # MetaWindow
2013-02-17 15:12:45 -05:00
*
* Return value : % TRUE if the window is occupies an entire monitor or
* the whole screen .
*/
gboolean
meta_window_is_monitor_sized ( MetaWindow * window )
{
if ( window - > fullscreen )
return TRUE ;
2013-02-18 07:30:37 -05:00
if ( meta_window_is_screen_sized ( window ) )
return TRUE ;
2013-02-17 15:12:45 -05:00
if ( window - > override_redirect )
{
MetaRectangle window_rect , monitor_rect ;
2013-11-19 12:24:45 -05:00
meta_window_get_frame_rect ( window , & window_rect ) ;
2013-02-17 15:12:45 -05:00
meta_screen_get_monitor_geometry ( window - > screen , window - > monitor - > number , & monitor_rect ) ;
if ( meta_rectangle_equal ( & window_rect , & monitor_rect ) )
return TRUE ;
}
return FALSE ;
}
2011-03-09 07:04:18 -05:00
/**
* meta_window_is_on_primary_monitor :
2013-02-15 13:42:08 -05:00
* @ window : a # MetaWindow
2011-03-09 07:04:18 -05:00
*
* Return value : % TRUE if the window is on the primary monitor
*/
gboolean
meta_window_is_on_primary_monitor ( MetaWindow * window )
{
2011-05-20 04:56:12 -04:00
return window - > monitor - > is_primary ;
2011-03-09 07:04:18 -05:00
}
2012-08-30 02:40:18 -04:00
/**
* meta_window_requested_bypass_compositor :
2013-02-15 13:42:08 -05:00
* @ window : a # MetaWindow
2012-08-30 02:40:18 -04:00
*
* Return value : % TRUE if the window requested to bypass the compositor
*/
gboolean
meta_window_requested_bypass_compositor ( MetaWindow * window )
{
2013-02-02 14:05:43 -05:00
return window - > bypass_compositor = = _NET_WM_BYPASS_COMPOSITOR_HINT_ON ;
2012-08-30 02:40:18 -04:00
}
/**
* meta_window_requested_dont_bypass_compositor :
2013-02-15 13:42:08 -05:00
* @ window : a # MetaWindow
2012-08-30 02:40:18 -04:00
*
* Return value : % TRUE if the window requested to opt out of unredirecting
*/
gboolean
meta_window_requested_dont_bypass_compositor ( MetaWindow * window )
{
2013-02-02 14:05:43 -05:00
return window - > bypass_compositor = = _NET_WM_BYPASS_COMPOSITOR_HINT_OFF ;
2012-08-30 02:40:18 -04:00
}
2011-01-19 10:41:16 -05:00
void
2010-06-24 14:41:28 -04:00
meta_window_tile ( MetaWindow * window )
{
2010-09-24 15:35:00 -04:00
MetaMaximizeFlags directions ;
2014-03-18 17:06:42 -04:00
MetaRectangle old_rect ;
MetaRectangle new_rect ;
2010-09-24 15:35:00 -04:00
2010-06-24 14:41:28 -04:00
/* Don't do anything if no tiling is requested */
if ( window - > tile_mode = = META_TILE_NONE )
return ;
2010-09-24 15:35:00 -04:00
if ( window - > tile_mode = = META_TILE_MAXIMIZED )
2014-03-13 18:32:20 -04:00
directions = META_MAXIMIZE_BOTH ;
2010-09-24 15:35:00 -04:00
else
directions = META_MAXIMIZE_VERTICAL ;
meta_window_maximize_internal ( window , directions , NULL ) ;
2012-03-24 06:29:39 -04:00
meta_screen_update_tile_preview ( window - > screen , FALSE ) ;
2010-06-24 14:41:28 -04:00
2014-03-18 17:06:42 -04:00
meta_window_get_frame_rect ( window , & old_rect ) ;
2010-06-24 14:41:28 -04:00
2014-03-18 17:06:42 -04:00
meta_window_move_resize_now ( window ) ;
2010-06-24 14:41:28 -04:00
2014-03-18 17:06:42 -04:00
meta_window_get_frame_rect ( window , & new_rect ) ;
meta_compositor_maximize_window ( window - > display - > compositor ,
window ,
& old_rect ,
& new_rect ) ;
2011-12-05 15:45:26 -05:00
2014-03-18 17:06:42 -04:00
if ( window - > frame )
meta_ui_queue_frame_draw ( window - > screen - > ui ,
window - > frame - > xwindow ) ;
2010-06-24 14:41:28 -04:00
}
2011-04-18 12:27:03 -04:00
static gboolean
meta_window_can_tile_maximized ( MetaWindow * window )
{
2011-12-05 18:52:09 -05:00
return window - > has_maximize_func ;
2011-04-18 12:27:03 -04:00
}
2011-12-05 15:52:54 -05:00
gboolean
2010-09-24 14:31:36 -04:00
meta_window_can_tile_side_by_side ( MetaWindow * window )
2010-06-24 14:41:28 -04:00
{
2012-06-11 10:12:39 -04:00
int monitor ;
2010-06-24 14:41:28 -04:00
MetaRectangle tile_area ;
2013-11-15 16:11:20 -05:00
MetaRectangle client_rect ;
2010-06-24 14:41:28 -04:00
2011-04-18 12:27:03 -04:00
if ( ! meta_window_can_tile_maximized ( window ) )
2010-06-24 14:41:28 -04:00
return FALSE ;
monitor = meta_screen_get_current_monitor ( window - > screen ) ;
2012-06-08 04:21:10 -04:00
meta_window_get_work_area_for_monitor ( window , monitor , & tile_area ) ;
2010-06-24 14:41:28 -04:00
2010-09-24 07:23:18 -04:00
/* Do not allow tiling in portrait orientation */
if ( tile_area . height > tile_area . width )
return FALSE ;
2010-06-24 14:41:28 -04:00
tile_area . width / = 2 ;
2013-11-15 16:11:20 -05:00
meta_window_frame_rect_to_client_rect ( window , & tile_area , & client_rect ) ;
2010-06-24 14:41:28 -04:00
2013-11-15 16:11:20 -05:00
return client_rect . width > = window - > size_hints . min_width & &
client_rect . height > = window - > size_hints . min_height ;
2010-06-24 14:41:28 -04:00
}
2006-01-20 18:36:23 -05:00
static void
unmaximize_window_before_freeing ( MetaWindow * window )
{
meta_topic ( META_DEBUG_WINDOW_OPS ,
" Unmaximizing %s just before freeing \n " ,
window - > desc ) ;
window - > maximized_horizontally = FALSE ;
window - > maximized_vertically = FALSE ;
2006-10-01 17:59:09 -04:00
if ( window - > withdrawn ) /* See bug #137185 */
{
window - > rect = window - > saved_rect ;
set_net_wm_state ( window ) ;
}
else if ( window - > screen - > closing ) /* See bug #358042 */
{
/* Do NOT update net_wm_state: this screen is closing,
* it likely will be managed by another window manager
* that will need the current _NET_WM_STATE atoms .
* Moreover , it will need to know the unmaximized geometry ,
* therefore move_resize the window to saved_rect here
* before closing it . */
meta_window_move_resize ( window ,
FALSE ,
window - > saved_rect . x ,
window - > saved_rect . y ,
window - > saved_rect . width ,
window - > saved_rect . height ) ;
}
2006-01-20 18:36:23 -05:00
}
2010-09-18 16:28:30 -04:00
static void
meta_window_unmaximize_internal ( MetaWindow * window ,
MetaMaximizeFlags directions ,
MetaRectangle * desired_rect ,
int gravity )
2001-06-07 01:18:10 -04:00
{
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
gboolean unmaximize_horizontally , unmaximize_vertically ;
2014-03-18 17:06:42 -04:00
MetaRectangle new_rect ;
2009-06-14 08:11:29 -04:00
g_return_if_fail ( ! window - > override_redirect ) ;
/* At least one of the two directions ought to be set */
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
unmaximize_horizontally = directions & META_MAXIMIZE_HORIZONTAL ;
unmaximize_vertically = directions & META_MAXIMIZE_VERTICAL ;
g_assert ( unmaximize_horizontally | | unmaximize_vertically ) ;
2011-12-05 15:52:46 -05:00
if ( unmaximize_horizontally & & unmaximize_vertically )
window - > saved_maximize = FALSE ;
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
/* Only do something if the window isn't already maximized in the
* given direction ( s ) .
*/
if ( ( unmaximize_horizontally & & window - > maximized_horizontally ) | |
( unmaximize_vertically & & window - > maximized_vertically ) )
2001-06-07 01:18:10 -04:00
{
2007-04-03 23:07:26 -04:00
MetaRectangle target_rect ;
2012-03-12 12:09:17 -04:00
MetaRectangle work_area ;
2013-12-05 08:27:28 -05:00
MetaRectangle old_rect ;
2012-03-12 12:09:17 -04:00
meta_window_get_work_area_for_monitor ( window , window - > monitor - > number , & work_area ) ;
2013-12-05 08:27:28 -05:00
meta_window_get_frame_rect ( window , & old_rect ) ;
2007-04-03 23:07:26 -04:00
2002-02-07 22:34:26 -05:00
meta_topic ( META_DEBUG_WINDOW_OPS ,
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
" Unmaximizing %s%s \n " ,
window - > desc ,
unmaximize_horizontally & & unmaximize_vertically ? " " :
unmaximize_horizontally ? " horizontally " :
unmaximize_vertically ? " vertically " : " BUGGGGG " ) ;
2008-12-18 07:41:56 -05:00
window - > maximized_horizontally =
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
window - > maximized_horizontally & & ! unmaximize_horizontally ;
2008-12-18 07:41:56 -05:00
window - > maximized_vertically =
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
window - > maximized_vertically & & ! unmaximize_vertically ;
2001-06-07 01:18:10 -04:00
2014-02-01 18:58:58 -05:00
/* recalc_features() will eventually clear the cached frame
2013-12-05 08:27:28 -05:00
* extents , but we need the correct frame extents in the code below ,
* so invalidate the old frame extents manually up front .
*/
meta_window_frame_size_changed ( window ) ;
2007-04-03 23:07:26 -04:00
/* Unmaximize to the saved_rect position in the direction(s)
* being unmaximized .
*/
2007-04-08 23:43:55 -04:00
meta_window_get_client_root_coords ( window , & target_rect ) ;
2012-03-12 12:09:17 -04:00
/* Avoid unmaximizing to "almost maximized" size when the previous size
* is greater then 80 % of the work area use MAX_UNMAXIMIZED_WINDOW_AREA of the work area as upper limit
* while maintaining the aspect ratio .
*/
if ( unmaximize_horizontally & & unmaximize_vertically & &
desired_rect - > width * desired_rect - > height > work_area . width * work_area . height * MAX_UNMAXIMIZED_WINDOW_AREA )
{
if ( desired_rect - > width > desired_rect - > height )
{
float aspect = ( float ) desired_rect - > height / ( float ) desired_rect - > width ;
desired_rect - > width = MAX ( work_area . width * sqrt ( MAX_UNMAXIMIZED_WINDOW_AREA ) , window - > size_hints . min_width ) ;
desired_rect - > height = MAX ( desired_rect - > width * aspect , window - > size_hints . min_height ) ;
}
else
{
float aspect = ( float ) desired_rect - > width / ( float ) desired_rect - > height ;
desired_rect - > height = MAX ( work_area . height * sqrt ( MAX_UNMAXIMIZED_WINDOW_AREA ) , window - > size_hints . min_height ) ;
desired_rect - > width = MAX ( desired_rect - > height * aspect , window - > size_hints . min_width ) ;
}
}
2007-04-03 23:07:26 -04:00
if ( unmaximize_horizontally )
{
2010-09-18 16:28:30 -04:00
target_rect . x = desired_rect - > x ;
target_rect . width = desired_rect - > width ;
2007-04-03 23:07:26 -04:00
}
if ( unmaximize_vertically )
{
2010-09-18 16:28:30 -04:00
target_rect . y = desired_rect - > y ;
target_rect . height = desired_rect - > height ;
2007-04-03 23:07:26 -04:00
}
2007-04-04 17:54:56 -04:00
/* Window's size hints may have changed while maximized, making
* saved_rect invalid . # 329152
*/
ensure_size_hints_satisfied ( & target_rect , & window - > size_hints ) ;
2014-03-18 17:06:42 -04:00
meta_window_move_resize_internal ( window ,
META_IS_MOVE_ACTION | META_IS_RESIZE_ACTION ,
gravity ,
target_rect . x ,
target_rect . y ,
target_rect . width ,
target_rect . height ) ;
meta_window_get_frame_rect ( window , & new_rect ) ;
meta_compositor_unmaximize_window ( window - > display - > compositor ,
window ,
& old_rect ,
& new_rect ) ;
2001-06-10 23:24:20 -04:00
2010-06-14 16:40:38 -04:00
/* Make sure user_rect is current.
*/
force_save_user_window_placement ( window ) ;
2010-06-24 14:41:28 -04:00
/* When we unmaximize, if we're doing a mouse move also we could
* get the window suddenly jumping to the upper left corner of
* the workspace , since that ' s where it was when the grab op
* started . So we need to update the grab state . We have to do
* it after the actual operation , as the window may have been moved
* by constraints .
*/
if ( meta_grab_op_is_moving ( window - > display - > grab_op ) & &
window - > display - > grab_window = = window )
{
window - > display - > grab_anchor_window_pos = window - > user_rect ;
}
2014-02-01 18:58:58 -05:00
meta_window_recalc_features ( window ) ;
2001-06-10 23:24:20 -04:00
set_net_wm_state ( window ) ;
2001-06-04 02:17:52 -04:00
}
2010-02-12 15:26:52 -05:00
2014-02-09 11:45:09 -05:00
if ( window - > surface & & ! window - > maximized_horizontally & & ! window - > maximized_vertically )
meta_wayland_surface_send_unmaximized ( window - > surface ) ;
g_object_freeze_notify ( G_OBJECT ( window ) ) ;
g_object_notify ( G_OBJECT ( window ) , " maximized-horizontally " ) ;
g_object_notify ( G_OBJECT ( window ) , " maximized-vertically " ) ;
g_object_thaw_notify ( G_OBJECT ( window ) ) ;
2001-06-02 00:14:18 -04:00
}
2010-09-18 16:28:30 -04:00
void
meta_window_unmaximize ( MetaWindow * window ,
MetaMaximizeFlags directions )
{
meta_window_unmaximize_internal ( window , directions , & window - > saved_rect ,
NorthWestGravity ) ;
}
/* Like meta_window_unmaximize(), but instead of unmaximizing to the
* saved position , we give the new desired size , and the gravity that
* determines the positioning relationship between the area occupied
* maximized and the new are . The arguments are similar to
* meta_window_resize_with_gravity ( ) .
2010-09-19 10:54:08 -04:00
* Unlike meta_window_unmaximize ( ) , tiling is not restored for windows
* with a tile mode other than META_TILE_NONE .
2010-09-18 16:28:30 -04:00
*/
void
meta_window_unmaximize_with_gravity ( MetaWindow * window ,
MetaMaximizeFlags directions ,
int new_width ,
int new_height ,
int gravity )
{
MetaRectangle desired_rect ;
meta_window_get_position ( window , & desired_rect . x , & desired_rect . y ) ;
desired_rect . width = new_width ;
desired_rect . height = new_height ;
meta_window_unmaximize_internal ( window , directions , & desired_rect , gravity ) ;
}
2003-06-27 10:49:56 -04:00
void
meta_window_make_above ( MetaWindow * window )
{
2009-06-14 08:11:29 -04:00
g_return_if_fail ( ! window - > override_redirect ) ;
2011-07-02 07:08:17 -04:00
meta_window_set_above ( window , TRUE ) ;
2003-06-27 10:49:56 -04:00
meta_window_raise ( window ) ;
}
void
meta_window_unmake_above ( MetaWindow * window )
{
2009-06-14 08:11:29 -04:00
g_return_if_fail ( ! window - > override_redirect ) ;
2011-07-02 07:08:17 -04:00
meta_window_set_above ( window , FALSE ) ;
2003-06-27 10:49:56 -04:00
meta_window_raise ( window ) ;
2011-07-02 07:08:17 -04:00
}
static void
meta_window_set_above ( MetaWindow * window ,
gboolean new_value )
{
new_value = new_value ! = FALSE ;
if ( new_value = = window - > wm_state_above )
return ;
window - > wm_state_above = new_value ;
2003-06-27 10:49:56 -04:00
meta_window_update_layer ( window ) ;
set_net_wm_state ( window ) ;
2013-11-15 17:37:50 -05:00
meta_window_frame_size_changed ( window ) ;
2011-07-02 07:08:17 -04:00
g_object_notify ( G_OBJECT ( window ) , " above " ) ;
2003-06-27 10:49:56 -04:00
}
2002-02-23 17:58:02 -05:00
void
Fix several bugs with handling of fullscreen windows, causing them to not
2006-08-21 Elijah Newren <newren gmail com>
Fix several bugs with handling of fullscreen windows, causing them
to not actually be fullscreen. #343115 (and also #346927,
#350547, #351643, the recent additional WINE-related issue
mentioned on the mailing list, and probably others...)
* src/constraints.c (setup_constraint_info): if a window tries to
resize to fullscreen-size and it has a fullscreen function but
isn't actually marked as fullscreen then assist it by marking it
as such, (constrain_fully_onscreen, constrain_titlebar_visible):
ignore this constraint for fullscreen windows since such windows
have a separate specialized constraint
* src/stack.c (window_is_fullscreen_size, get_standalone_layer):
remove the old window_is_fullscreen_size() hack for detecting
windows to treat as fullscreen since it doesn't work well with the
new constraints framework (i.e. we needed a slightly different
hack)
* src/window.[ch] (meta_window_new_with_addrs): shuffle the order
of adding the window to the stack and moveresizing the window
since moveresizing can cause stack changes if the window's initial
size is fullscreen size, (meta_window_make_fullscreen,
meta_window_make_fullscreen_internal): split
meta_window_make_fullscreen() similar to meta_window_maximize() so
that constraints can make use of it
2006-08-21 14:08:05 -04:00
meta_window_make_fullscreen_internal ( MetaWindow * window )
2002-02-23 17:58:02 -05:00
{
if ( ! window - > fullscreen )
{
meta_topic ( META_DEBUG_WINDOW_OPS ,
" Fullscreening %s \n " , window - > desc ) ;
2002-08-10 15:04:55 -04:00
if ( window - > shaded )
Partial audit to fix timestamp usage. One step towards fixing #355180; see
2006-09-18 Elijah Newren <newren gmail com>
Partial audit to fix timestamp usage. One step towards fixing
#355180; see important comments in that bug.
* src/core.[ch] (meta_core_unshade, meta_core_shade):
* src/delete.c (meta_window_present_delete_dialog,
delete_ping_timeout_func):
* src/display.[ch] (meta_display_open, meta_display_close,
event_callback, meta_display_begin_grab_op,
process_selection_clear, meta_display_unmanage_screen,
meta_display_unmanage_windows_for_screen):
* src/frames.c (meta_frames_button_press_event):
* src/keybindings.c (handle_toggle_shade):
* src/main.c (main):
* src/screen.[ch] (update_num_workspaces, meta_screen_new,
meta_screen_free, prefs_changed_callback):
* src/window.[ch] (meta_window_free, finish_minimize,
implement_showing, meta_window_show, meta_window_maximize,
meta_window_make_fullscreen_internal,
meta_window_unmake_fullscreen, meta_window_shade,
meta_window_unshade, window_activate, send_sync_request,
meta_window_client_message, menu_callback,
meta_window_update_keyboard_resize):
Remove usage of CurrentTime, meta_display_get_current_time() and
meta_display_get_current_time_roundtrip() where possible, or
document why it isn't possible, or at very least add a FIXME with
some explanation of my laziness and what needs to be done.
2006-09-18 13:27:24 -04:00
{
/* Shading sucks anyway; I'm not adding a timestamp argument
* to this function just for this niche usage & corner case .
*/
2008-12-18 07:41:56 -05:00
guint32 timestamp =
Partial audit to fix timestamp usage. One step towards fixing #355180; see
2006-09-18 Elijah Newren <newren gmail com>
Partial audit to fix timestamp usage. One step towards fixing
#355180; see important comments in that bug.
* src/core.[ch] (meta_core_unshade, meta_core_shade):
* src/delete.c (meta_window_present_delete_dialog,
delete_ping_timeout_func):
* src/display.[ch] (meta_display_open, meta_display_close,
event_callback, meta_display_begin_grab_op,
process_selection_clear, meta_display_unmanage_screen,
meta_display_unmanage_windows_for_screen):
* src/frames.c (meta_frames_button_press_event):
* src/keybindings.c (handle_toggle_shade):
* src/main.c (main):
* src/screen.[ch] (update_num_workspaces, meta_screen_new,
meta_screen_free, prefs_changed_callback):
* src/window.[ch] (meta_window_free, finish_minimize,
implement_showing, meta_window_show, meta_window_maximize,
meta_window_make_fullscreen_internal,
meta_window_unmake_fullscreen, meta_window_shade,
meta_window_unshade, window_activate, send_sync_request,
meta_window_client_message, menu_callback,
meta_window_update_keyboard_resize):
Remove usage of CurrentTime, meta_display_get_current_time() and
meta_display_get_current_time_roundtrip() where possible, or
document why it isn't possible, or at very least add a FIXME with
some explanation of my laziness and what needs to be done.
2006-09-18 13:27:24 -04:00
meta_display_get_current_time_roundtrip ( window - > display ) ;
meta_window_unshade ( window , timestamp ) ;
}
2002-09-24 17:33:34 -04:00
meta_window_save_rect ( window ) ;
2008-12-18 07:41:56 -05:00
2002-02-23 17:58:02 -05:00
window - > fullscreen = TRUE ;
2010-01-20 10:59:07 -05:00
window - > force_save_user_rect = FALSE ;
2002-02-23 17:58:02 -05:00
2002-08-12 17:32:13 -04:00
meta_stack_freeze ( window - > screen - > stack ) ;
2002-09-24 17:22:06 -04:00
meta_window_update_layer ( window ) ;
2008-12-18 07:41:56 -05:00
2002-02-23 17:58:02 -05:00
meta_window_raise ( window ) ;
2002-08-12 17:32:13 -04:00
meta_stack_thaw ( window - > screen - > stack ) ;
2008-12-18 07:41:56 -05:00
2014-02-01 18:58:58 -05:00
meta_window_recalc_features ( window ) ;
Fix several bugs with handling of fullscreen windows, causing them to not
2006-08-21 Elijah Newren <newren gmail com>
Fix several bugs with handling of fullscreen windows, causing them
to not actually be fullscreen. #343115 (and also #346927,
#350547, #351643, the recent additional WINE-related issue
mentioned on the mailing list, and probably others...)
* src/constraints.c (setup_constraint_info): if a window tries to
resize to fullscreen-size and it has a fullscreen function but
isn't actually marked as fullscreen then assist it by marking it
as such, (constrain_fully_onscreen, constrain_titlebar_visible):
ignore this constraint for fullscreen windows since such windows
have a separate specialized constraint
* src/stack.c (window_is_fullscreen_size, get_standalone_layer):
remove the old window_is_fullscreen_size() hack for detecting
windows to treat as fullscreen since it doesn't work well with the
new constraints framework (i.e. we needed a slightly different
hack)
* src/window.[ch] (meta_window_new_with_addrs): shuffle the order
of adding the window to the stack and moveresizing the window
since moveresizing can cause stack changes if the window's initial
size is fullscreen size, (meta_window_make_fullscreen,
meta_window_make_fullscreen_internal): split
meta_window_make_fullscreen() similar to meta_window_maximize() so
that constraints can make use of it
2006-08-21 14:08:05 -04:00
set_net_wm_state ( window ) ;
2009-03-17 08:27:13 -04:00
2013-03-14 16:55:49 -04:00
/* For the auto-minimize feature, if we fail to get focus */
meta_screen_queue_check_fullscreen ( window - > screen ) ;
2014-02-09 11:45:09 -05:00
if ( window - > surface )
meta_wayland_surface_send_fullscreened ( window - > surface ) ;
2009-03-17 08:27:13 -04:00
g_object_notify ( G_OBJECT ( window ) , " fullscreen " ) ;
Fix several bugs with handling of fullscreen windows, causing them to not
2006-08-21 Elijah Newren <newren gmail com>
Fix several bugs with handling of fullscreen windows, causing them
to not actually be fullscreen. #343115 (and also #346927,
#350547, #351643, the recent additional WINE-related issue
mentioned on the mailing list, and probably others...)
* src/constraints.c (setup_constraint_info): if a window tries to
resize to fullscreen-size and it has a fullscreen function but
isn't actually marked as fullscreen then assist it by marking it
as such, (constrain_fully_onscreen, constrain_titlebar_visible):
ignore this constraint for fullscreen windows since such windows
have a separate specialized constraint
* src/stack.c (window_is_fullscreen_size, get_standalone_layer):
remove the old window_is_fullscreen_size() hack for detecting
windows to treat as fullscreen since it doesn't work well with the
new constraints framework (i.e. we needed a slightly different
hack)
* src/window.[ch] (meta_window_new_with_addrs): shuffle the order
of adding the window to the stack and moveresizing the window
since moveresizing can cause stack changes if the window's initial
size is fullscreen size, (meta_window_make_fullscreen,
meta_window_make_fullscreen_internal): split
meta_window_make_fullscreen() similar to meta_window_maximize() so
that constraints can make use of it
2006-08-21 14:08:05 -04:00
}
}
void
meta_window_make_fullscreen ( MetaWindow * window )
{
2009-06-14 08:11:29 -04:00
g_return_if_fail ( ! window - > override_redirect ) ;
Fix several bugs with handling of fullscreen windows, causing them to not
2006-08-21 Elijah Newren <newren gmail com>
Fix several bugs with handling of fullscreen windows, causing them
to not actually be fullscreen. #343115 (and also #346927,
#350547, #351643, the recent additional WINE-related issue
mentioned on the mailing list, and probably others...)
* src/constraints.c (setup_constraint_info): if a window tries to
resize to fullscreen-size and it has a fullscreen function but
isn't actually marked as fullscreen then assist it by marking it
as such, (constrain_fully_onscreen, constrain_titlebar_visible):
ignore this constraint for fullscreen windows since such windows
have a separate specialized constraint
* src/stack.c (window_is_fullscreen_size, get_standalone_layer):
remove the old window_is_fullscreen_size() hack for detecting
windows to treat as fullscreen since it doesn't work well with the
new constraints framework (i.e. we needed a slightly different
hack)
* src/window.[ch] (meta_window_new_with_addrs): shuffle the order
of adding the window to the stack and moveresizing the window
since moveresizing can cause stack changes if the window's initial
size is fullscreen size, (meta_window_make_fullscreen,
meta_window_make_fullscreen_internal): split
meta_window_make_fullscreen() similar to meta_window_maximize() so
that constraints can make use of it
2006-08-21 14:08:05 -04:00
if ( ! window - > fullscreen )
{
meta_window_make_fullscreen_internal ( window ) ;
2002-02-23 17:58:02 -05:00
/* move_resize with new constraints
*/
Refactor thrice-duplicated queue code in window.c. Closes #376760.
2007-06-10 Thomas Thurman <thomas@thurman.org.uk>
Refactor thrice-duplicated queue code in window.c. Closes #376760.
* src/window.c (meta_window_queue, meta_window_unqueue):
New functions.
* src/window.[ch] (meta_window_unqueue_*, meta_window_queue_*):
Removed functions.
* src/window.c (meta_window_new_with_attrs, meta_window_free,
meta_window_flush_calc_showing, queue_calc_showing_func,
meta_window_minimize, meta_window_unminimize, meta_window_maximize,
meta_window_make_fullscreen, meta_window_shade,
meta_window_unshade, meta_window_move_resize_internal,
window_stick_impl, window_unstick_impl,
meta_window_client_message, process_property_notify): Modified to
use new queueing functions.
* src/window.c (idle_move_resize, idle_update_icon,
idle_calc_showing): update to receive queue number from pointer.
* src/window.h (MetaQueueType): new enum.
* src/window.h (MetaWindow): *_queued replaced with is_in_queue
bitfield.
* src/core.c (meta_core_queue_frame_resize):
* src/display.c (event_callback,
meta_display_queue_retheme_all_windows): Using new queueing functions.
* src/frame.c (meta_window_destroy_frame): Using new queueing functions.
* src/screen.c (queue_resize, meta_screen_resize_func,
queue_windows_showing): Using new queueing functions.
* src/window-props.c (reload_mwm_hints, reload_wm_hints,
reload_transient_for): Using new queueing functions.
* src/workspace.c (meta_workspace_add_window,
meta_workspace_remove_window, meta_workspace_queue_calc_showing,
meta_workspace_invalidate_work_area): Using new queueing functions.
svn path=/trunk/; revision=3236
2007-06-10 21:15:33 -04:00
meta_window_queue ( window , META_QUEUE_MOVE_RESIZE ) ;
2002-02-23 17:58:02 -05:00
}
}
void
meta_window_unmake_fullscreen ( MetaWindow * window )
{
2009-06-14 08:11:29 -04:00
g_return_if_fail ( ! window - > override_redirect ) ;
2002-02-23 17:58:02 -05:00
if ( window - > fullscreen )
{
2007-04-04 17:54:56 -04:00
MetaRectangle target_rect ;
2002-02-23 17:58:02 -05:00
meta_topic ( META_DEBUG_WINDOW_OPS ,
" Unfullscreening %s \n " , window - > desc ) ;
2008-12-18 07:41:56 -05:00
2002-02-23 17:58:02 -05:00
window - > fullscreen = FALSE ;
2007-04-04 17:54:56 -04:00
target_rect = window - > saved_rect ;
/* Window's size hints may have changed while maximized, making
* saved_rect invalid . # 329152
*/
ensure_size_hints_satisfied ( & target_rect , & window - > size_hints ) ;
2002-02-23 17:58:02 -05:00
2011-10-17 14:56:56 -04:00
/* Need to update window->has_resize_func before we move_resize()
*/
2014-02-01 18:58:58 -05:00
meta_window_recalc_features ( window ) ;
2011-10-17 14:56:56 -04:00
set_net_wm_state ( window ) ;
2002-02-23 17:58:02 -05:00
meta_window_move_resize ( window ,
2005-11-22 10:48:30 -05:00
FALSE ,
2007-04-04 17:54:56 -04:00
target_rect . x ,
target_rect . y ,
target_rect . width ,
target_rect . height ) ;
2002-02-23 17:58:02 -05:00
2010-01-20 10:59:07 -05:00
/* Make sure user_rect is current.
*/
force_save_user_window_placement ( window ) ;
2005-01-31 11:49:23 -05:00
meta_window_update_layer ( window ) ;
2008-12-18 07:41:56 -05:00
2014-02-09 11:45:09 -05:00
if ( window - > surface )
meta_wayland_surface_send_unfullscreened ( window - > surface ) ;
2009-03-17 08:27:13 -04:00
g_object_notify ( G_OBJECT ( window ) , " fullscreen " ) ;
2002-02-23 17:58:02 -05:00
}
}
2008-11-16 21:57:20 -05:00
void
meta_window_update_fullscreen_monitors ( MetaWindow * window ,
unsigned long top ,
unsigned long bottom ,
unsigned long left ,
unsigned long right )
{
2009-08-31 09:52:16 -04:00
if ( ( int ) top < window - > screen - > n_monitor_infos & &
( int ) bottom < window - > screen - > n_monitor_infos & &
( int ) left < window - > screen - > n_monitor_infos & &
( int ) right < window - > screen - > n_monitor_infos )
2008-11-16 21:57:20 -05:00
{
window - > fullscreen_monitors [ 0 ] = top ;
window - > fullscreen_monitors [ 1 ] = bottom ;
window - > fullscreen_monitors [ 2 ] = left ;
window - > fullscreen_monitors [ 3 ] = right ;
}
else
{
window - > fullscreen_monitors [ 0 ] = - 1 ;
}
if ( window - > fullscreen )
{
meta_window_queue ( window , META_QUEUE_MOVE_RESIZE ) ;
}
}
2001-06-02 21:33:27 -04:00
void
Partial audit to fix timestamp usage. One step towards fixing #355180; see
2006-09-18 Elijah Newren <newren gmail com>
Partial audit to fix timestamp usage. One step towards fixing
#355180; see important comments in that bug.
* src/core.[ch] (meta_core_unshade, meta_core_shade):
* src/delete.c (meta_window_present_delete_dialog,
delete_ping_timeout_func):
* src/display.[ch] (meta_display_open, meta_display_close,
event_callback, meta_display_begin_grab_op,
process_selection_clear, meta_display_unmanage_screen,
meta_display_unmanage_windows_for_screen):
* src/frames.c (meta_frames_button_press_event):
* src/keybindings.c (handle_toggle_shade):
* src/main.c (main):
* src/screen.[ch] (update_num_workspaces, meta_screen_new,
meta_screen_free, prefs_changed_callback):
* src/window.[ch] (meta_window_free, finish_minimize,
implement_showing, meta_window_show, meta_window_maximize,
meta_window_make_fullscreen_internal,
meta_window_unmake_fullscreen, meta_window_shade,
meta_window_unshade, window_activate, send_sync_request,
meta_window_client_message, menu_callback,
meta_window_update_keyboard_resize):
Remove usage of CurrentTime, meta_display_get_current_time() and
meta_display_get_current_time_roundtrip() where possible, or
document why it isn't possible, or at very least add a FIXME with
some explanation of my laziness and what needs to be done.
2006-09-18 13:27:24 -04:00
meta_window_shade ( MetaWindow * window ,
guint32 timestamp )
2001-06-02 21:33:27 -04:00
{
2009-06-14 08:11:29 -04:00
g_return_if_fail ( ! window - > override_redirect ) ;
2002-02-07 22:34:26 -05:00
meta_topic ( META_DEBUG_WINDOW_OPS ,
" Shading %s \n " , window - > desc ) ;
2001-06-07 22:17:48 -04:00
if ( ! window - > shaded )
2008-12-18 07:41:56 -05:00
{
2001-08-06 00:43:31 -04:00
window - > shaded = TRUE ;
Refactor thrice-duplicated queue code in window.c. Closes #376760.
2007-06-10 Thomas Thurman <thomas@thurman.org.uk>
Refactor thrice-duplicated queue code in window.c. Closes #376760.
* src/window.c (meta_window_queue, meta_window_unqueue):
New functions.
* src/window.[ch] (meta_window_unqueue_*, meta_window_queue_*):
Removed functions.
* src/window.c (meta_window_new_with_attrs, meta_window_free,
meta_window_flush_calc_showing, queue_calc_showing_func,
meta_window_minimize, meta_window_unminimize, meta_window_maximize,
meta_window_make_fullscreen, meta_window_shade,
meta_window_unshade, meta_window_move_resize_internal,
window_stick_impl, window_unstick_impl,
meta_window_client_message, process_property_notify): Modified to
use new queueing functions.
* src/window.c (idle_move_resize, idle_update_icon,
idle_calc_showing): update to receive queue number from pointer.
* src/window.h (MetaQueueType): new enum.
* src/window.h (MetaWindow): *_queued replaced with is_in_queue
bitfield.
* src/core.c (meta_core_queue_frame_resize):
* src/display.c (event_callback,
meta_display_queue_retheme_all_windows): Using new queueing functions.
* src/frame.c (meta_window_destroy_frame): Using new queueing functions.
* src/screen.c (queue_resize, meta_screen_resize_func,
queue_windows_showing): Using new queueing functions.
* src/window-props.c (reload_mwm_hints, reload_wm_hints,
reload_transient_for): Using new queueing functions.
* src/workspace.c (meta_workspace_add_window,
meta_workspace_remove_window, meta_workspace_queue_calc_showing,
meta_workspace_invalidate_work_area): Using new queueing functions.
svn path=/trunk/; revision=3236
2007-06-10 21:15:33 -04:00
meta_window_queue ( window , META_QUEUE_MOVE_RESIZE | META_QUEUE_CALC_SHOWING ) ;
2013-11-15 17:37:50 -05:00
meta_window_frame_size_changed ( window ) ;
2001-06-10 23:24:20 -04:00
2002-01-03 18:28:19 -05:00
/* After queuing the calc showing, since _focus flushes it,
* and we need to focus the frame
*/
meta_topic ( META_DEBUG_FOCUS ,
" Re-focusing window %s after shading it \n " ,
window - > desc ) ;
Partial audit to fix timestamp usage. One step towards fixing #355180; see
2006-09-18 Elijah Newren <newren gmail com>
Partial audit to fix timestamp usage. One step towards fixing
#355180; see important comments in that bug.
* src/core.[ch] (meta_core_unshade, meta_core_shade):
* src/delete.c (meta_window_present_delete_dialog,
delete_ping_timeout_func):
* src/display.[ch] (meta_display_open, meta_display_close,
event_callback, meta_display_begin_grab_op,
process_selection_clear, meta_display_unmanage_screen,
meta_display_unmanage_windows_for_screen):
* src/frames.c (meta_frames_button_press_event):
* src/keybindings.c (handle_toggle_shade):
* src/main.c (main):
* src/screen.[ch] (update_num_workspaces, meta_screen_new,
meta_screen_free, prefs_changed_callback):
* src/window.[ch] (meta_window_free, finish_minimize,
implement_showing, meta_window_show, meta_window_maximize,
meta_window_make_fullscreen_internal,
meta_window_unmake_fullscreen, meta_window_shade,
meta_window_unshade, window_activate, send_sync_request,
meta_window_client_message, menu_callback,
meta_window_update_keyboard_resize):
Remove usage of CurrentTime, meta_display_get_current_time() and
meta_display_get_current_time_roundtrip() where possible, or
document why it isn't possible, or at very least add a FIXME with
some explanation of my laziness and what needs to be done.
2006-09-18 13:27:24 -04:00
meta_window_focus ( window , timestamp ) ;
2008-12-18 07:41:56 -05:00
2001-06-10 23:24:20 -04:00
set_net_wm_state ( window ) ;
2001-06-02 21:33:27 -04:00
}
}
2001-06-06 00:47:37 -04:00
void
Partial audit to fix timestamp usage. One step towards fixing #355180; see
2006-09-18 Elijah Newren <newren gmail com>
Partial audit to fix timestamp usage. One step towards fixing
#355180; see important comments in that bug.
* src/core.[ch] (meta_core_unshade, meta_core_shade):
* src/delete.c (meta_window_present_delete_dialog,
delete_ping_timeout_func):
* src/display.[ch] (meta_display_open, meta_display_close,
event_callback, meta_display_begin_grab_op,
process_selection_clear, meta_display_unmanage_screen,
meta_display_unmanage_windows_for_screen):
* src/frames.c (meta_frames_button_press_event):
* src/keybindings.c (handle_toggle_shade):
* src/main.c (main):
* src/screen.[ch] (update_num_workspaces, meta_screen_new,
meta_screen_free, prefs_changed_callback):
* src/window.[ch] (meta_window_free, finish_minimize,
implement_showing, meta_window_show, meta_window_maximize,
meta_window_make_fullscreen_internal,
meta_window_unmake_fullscreen, meta_window_shade,
meta_window_unshade, window_activate, send_sync_request,
meta_window_client_message, menu_callback,
meta_window_update_keyboard_resize):
Remove usage of CurrentTime, meta_display_get_current_time() and
meta_display_get_current_time_roundtrip() where possible, or
document why it isn't possible, or at very least add a FIXME with
some explanation of my laziness and what needs to be done.
2006-09-18 13:27:24 -04:00
meta_window_unshade ( MetaWindow * window ,
guint32 timestamp )
2001-06-06 00:47:37 -04:00
{
2009-06-14 08:11:29 -04:00
g_return_if_fail ( ! window - > override_redirect ) ;
2002-02-07 22:34:26 -05:00
meta_topic ( META_DEBUG_WINDOW_OPS ,
" Unshading %s \n " , window - > desc ) ;
2001-06-07 22:17:48 -04:00
if ( window - > shaded )
{
window - > shaded = FALSE ;
Refactor thrice-duplicated queue code in window.c. Closes #376760.
2007-06-10 Thomas Thurman <thomas@thurman.org.uk>
Refactor thrice-duplicated queue code in window.c. Closes #376760.
* src/window.c (meta_window_queue, meta_window_unqueue):
New functions.
* src/window.[ch] (meta_window_unqueue_*, meta_window_queue_*):
Removed functions.
* src/window.c (meta_window_new_with_attrs, meta_window_free,
meta_window_flush_calc_showing, queue_calc_showing_func,
meta_window_minimize, meta_window_unminimize, meta_window_maximize,
meta_window_make_fullscreen, meta_window_shade,
meta_window_unshade, meta_window_move_resize_internal,
window_stick_impl, window_unstick_impl,
meta_window_client_message, process_property_notify): Modified to
use new queueing functions.
* src/window.c (idle_move_resize, idle_update_icon,
idle_calc_showing): update to receive queue number from pointer.
* src/window.h (MetaQueueType): new enum.
* src/window.h (MetaWindow): *_queued replaced with is_in_queue
bitfield.
* src/core.c (meta_core_queue_frame_resize):
* src/display.c (event_callback,
meta_display_queue_retheme_all_windows): Using new queueing functions.
* src/frame.c (meta_window_destroy_frame): Using new queueing functions.
* src/screen.c (queue_resize, meta_screen_resize_func,
queue_windows_showing): Using new queueing functions.
* src/window-props.c (reload_mwm_hints, reload_wm_hints,
reload_transient_for): Using new queueing functions.
* src/workspace.c (meta_workspace_add_window,
meta_workspace_remove_window, meta_workspace_queue_calc_showing,
meta_workspace_invalidate_work_area): Using new queueing functions.
svn path=/trunk/; revision=3236
2007-06-10 21:15:33 -04:00
meta_window_queue ( window , META_QUEUE_MOVE_RESIZE | META_QUEUE_CALC_SHOWING ) ;
2013-11-15 17:37:50 -05:00
meta_window_frame_size_changed ( window ) ;
2002-01-03 18:28:19 -05:00
2001-06-09 17:58:30 -04:00
/* focus the window */
2002-01-03 18:28:19 -05:00
meta_topic ( META_DEBUG_FOCUS ,
" Focusing window %s after unshading it \n " ,
window - > desc ) ;
Partial audit to fix timestamp usage. One step towards fixing #355180; see
2006-09-18 Elijah Newren <newren gmail com>
Partial audit to fix timestamp usage. One step towards fixing
#355180; see important comments in that bug.
* src/core.[ch] (meta_core_unshade, meta_core_shade):
* src/delete.c (meta_window_present_delete_dialog,
delete_ping_timeout_func):
* src/display.[ch] (meta_display_open, meta_display_close,
event_callback, meta_display_begin_grab_op,
process_selection_clear, meta_display_unmanage_screen,
meta_display_unmanage_windows_for_screen):
* src/frames.c (meta_frames_button_press_event):
* src/keybindings.c (handle_toggle_shade):
* src/main.c (main):
* src/screen.[ch] (update_num_workspaces, meta_screen_new,
meta_screen_free, prefs_changed_callback):
* src/window.[ch] (meta_window_free, finish_minimize,
implement_showing, meta_window_show, meta_window_maximize,
meta_window_make_fullscreen_internal,
meta_window_unmake_fullscreen, meta_window_shade,
meta_window_unshade, window_activate, send_sync_request,
meta_window_client_message, menu_callback,
meta_window_update_keyboard_resize):
Remove usage of CurrentTime, meta_display_get_current_time() and
meta_display_get_current_time_roundtrip() where possible, or
document why it isn't possible, or at very least add a FIXME with
some explanation of my laziness and what needs to be done.
2006-09-18 13:27:24 -04:00
meta_window_focus ( window , timestamp ) ;
2001-06-10 23:24:20 -04:00
set_net_wm_state ( window ) ;
}
}
2003-03-17 01:36:39 -05:00
static gboolean
unminimize_func ( MetaWindow * window ,
void * data )
{
meta_window_unminimize ( window ) ;
return TRUE ;
}
on unminimize, queue calc_showing on all transients
2002-05-05 Havoc Pennington <hp@pobox.com>
* src/window.c (meta_window_unminimize): on unminimize, queue
calc_showing on all transients
(meta_window_activate): on activate, unminimize all a window's
ancestors, not just the window itself.
* src/workspace.c (set_work_area_hint): don't increment "tmp" by
16 unsigned long, increment by 4
* src/window.c (meta_window_free): if a window isn't minimized,
restore its WM_STATE to NormalState instead of IconicState,
since IconicState on initial window map means that the window
should be minimized.
* src/workspace.c (meta_workspace_invalidate_work_area): queue an
idle to recompute the work area hint.
(set_work_area_hint): we need 4*num_workspaces ints, not just
num_workspaces.
* src/screen.c (meta_screen_new): add work_area_idle field,
handle it on screen shutdown
* src/common.h (META_PRIORITY_PREFS_NOTIFY,
META_PRIORITY_WORK_AREA_HINT): define some idle priorities
* src/window.c (meta_window_calc_showing): hide windows if
their parent window is minimized
(meta_window_minimize): also queue_calc_showing on all
transients of the window being minimized
* src/place.c (constrain_placement): function to apply
placement-time-only constraints, such as "not off the left of the
screen"
(meta_window_place): put dialogs down a bit over their parent,
not right at the top.
(meta_window_place): when centering a dialog, center it
on the current xinerama screen, rather than the entire
screen.
* src/screen.c (meta_screen_get_current_xinerama): new function,
but not implemented
2002-05-05 01:41:13 -04:00
static void
unminimize_window_and_all_transient_parents ( MetaWindow * window )
{
2003-05-18 13:16:03 -04:00
meta_window_unminimize ( window ) ;
2003-03-17 01:36:39 -05:00
meta_window_foreach_ancestor ( window , unminimize_func , NULL ) ;
on unminimize, queue calc_showing on all transients
2002-05-05 Havoc Pennington <hp@pobox.com>
* src/window.c (meta_window_unminimize): on unminimize, queue
calc_showing on all transients
(meta_window_activate): on activate, unminimize all a window's
ancestors, not just the window itself.
* src/workspace.c (set_work_area_hint): don't increment "tmp" by
16 unsigned long, increment by 4
* src/window.c (meta_window_free): if a window isn't minimized,
restore its WM_STATE to NormalState instead of IconicState,
since IconicState on initial window map means that the window
should be minimized.
* src/workspace.c (meta_workspace_invalidate_work_area): queue an
idle to recompute the work area hint.
(set_work_area_hint): we need 4*num_workspaces ints, not just
num_workspaces.
* src/screen.c (meta_screen_new): add work_area_idle field,
handle it on screen shutdown
* src/common.h (META_PRIORITY_PREFS_NOTIFY,
META_PRIORITY_WORK_AREA_HINT): define some idle priorities
* src/window.c (meta_window_calc_showing): hide windows if
their parent window is minimized
(meta_window_minimize): also queue_calc_showing on all
transients of the window being minimized
* src/place.c (constrain_placement): function to apply
placement-time-only constraints, such as "not off the left of the
screen"
(meta_window_place): put dialogs down a bit over their parent,
not right at the top.
(meta_window_place): when centering a dialog, center it
on the current xinerama screen, rather than the entire
screen.
* src/screen.c (meta_screen_get_current_xinerama): new function,
but not implemented
2002-05-05 01:41:13 -04:00
}
2014-01-23 16:33:53 -05:00
void
meta_window_activate_full ( MetaWindow * window ,
guint32 timestamp ,
MetaClientType source_indication ,
MetaWorkspace * workspace )
2001-10-29 02:00:46 -05:00
{
2006-01-10 00:44:40 -05:00
gboolean can_ignore_outdated_timestamps ;
Big patch to cover about 6 different issues in order to correct rare
2005-02-20 Elijah Newren <newren@gmail.com>
Big patch to cover about 6 different issues in order to correct
rare problems with timestamps (make sure window selected in
tasklist actually gets focus, sanity check timestamps to avoid
rogue apps hosing the system, correct the updating of
net_wm_user_time, correctly handle timestamps of 0 when comparing
xserver timestamps for those who have had their systems up for
over 25 days or so, add some debugging information to verbose
logs, some code cleanups). Fixes all issues listed in #167358.
* src/display.h: (struct _MetaDisplay): clarify comment on
last_focus_time, introduce a new variable--last_user_time,
(XSERVER_TIME_IS_BEFORE macro): put this functionality into a
separate macro and then introduce a new macro with this name that
uses the old one but adds additional special-case checks for
timestamps that are 0, (comment to
meta_display_set_input_focus_window): add information about how
last_user_time should be used in this function
* src/display.c (santiy_check_timestamps): new function,
(meta_display_open): intialize display->last_user_time,
(meta_display_get_current_time_roundtrip): use the timestamp,
which is known to be good, in order to sanity_check_timestamps,
(event_callback): use the new meta_window_ste_user_time() function
in order to correct problems, use the timestamp of KeyPress and
ButtonPress events, which are known to be good, in order to
sanity_check_timestamps, (timestamp_too_old): new function for
common behavior of meta_display_focus_the_no_focus_window and
meta_display_set_input_focus_window, with added checking for
display->last_user_time in addition to display->last_focus_time,
(meta_display_set_input_focus_window): replace some of the code
with a call to timestamp_too_old(),
(meta_display_focus_the_no_focus_window): replace some of th ecode
with a call to timestamp_too_old()
* src/window.h: (meta_window_set_user_time): new function to
abstract the many things that need to be done when updating the
net_wm_user_time of any window
* src/window.c: (meta_window_activate): add debugging spew, make
sure the comparison is made with last_user_time NOT
last_focus_time, use meta_window_set_user_time() function in order
to correct problems, (meta_window_client_message): add a newline
to a debugging message to make them easier to read,
(meta_window_set_user_time): new function
* src/window-props.c (reload_net_wm_user_time): use the new
meta_window_ste_user_time() function in order to correct problems
2005-02-20 12:14:16 -05:00
meta_topic ( META_DEBUG_FOCUS ,
Fix issues on 64-bit machines with timestamps by using guint32 (like gtk+
2006-09-13 Elijah Newren <newren gmail com>
* src/common.h (MetaWindowMenuFunc):
* src/core.[ch] (meta_core_user_lower_and_unfocus,
meta_core_user_focus, meta_core_show_window_menu,
meta_core_begin_grab_op, meta_core_end_grab_op):
* src/delete.c (delete_ping_reply_func, delete_ping_timeout_func,
meta_window_delete):
* src/display.[ch] (struct MetaDisplay, struct MetaPingData,
sanity_check_timestamps, meta_display_open, event_callback,
meta_spew_event, meta_display_set_grab_op_cursor,
meta_display_begin_grab_op, meta_display_end_grab_op,
meta_display_ping_timeout, meta_display_ping_window,
process_pong_message, timestamp_too_old,
meta_display_set_input_focus_window):
* src/keybindings.[ch] (grab_keyboard, ungrab_keyboard,
meta_screen_grab_all_keys, meta_window_grab_all_keys,
meta_window_ungrab_all_keys, error_on_generic_command,
error_on_command, error_on_terminal_command):
* src/metacity-dialog.c (on_realize, warn_about_no_sm_support,
error_about_command, main):
* src/screen.[ch] (struct _MetaScreen, meta_screen_new,
meta_screen_show_desktop, meta_screen_apply_startup_properties):
* src/session.c (warn_about_lame_clients_and_finish_interact):
* src/window.[ch] (struct _MetaWindow,
intervening_user_event_occurred, window_activate,
meta_window_delete, meta_window_focus,
meta_window_send_icccm_message, meta_window_client_message,
menu_callback, meta_window_show_menu, struct EventScannerData,
check_use_this_motion_notify, meta_window_begin_grab_op,
meta_window_set_user_time):
* src/workspace.[ch] (focus_ancestor_or_mru_window,
meta_workspace_activate_with_focus, meta_workspace_activate,
meta_workspace_focus_default_window,
focus_ancestor_or_mru_window):
Fix issues on 64-bit machines with timestamps by using guint32
(like gtk+ does) instead of Time. #348305
2006-09-13 12:32:33 -04:00
" _NET_ACTIVE_WINDOW message sent for %s at time %u "
2006-01-10 00:44:40 -05:00
" by client type %u. \n " ,
Fix issues on 64-bit machines with timestamps by using guint32 (like gtk+
2006-09-13 Elijah Newren <newren gmail com>
* src/common.h (MetaWindowMenuFunc):
* src/core.[ch] (meta_core_user_lower_and_unfocus,
meta_core_user_focus, meta_core_show_window_menu,
meta_core_begin_grab_op, meta_core_end_grab_op):
* src/delete.c (delete_ping_reply_func, delete_ping_timeout_func,
meta_window_delete):
* src/display.[ch] (struct MetaDisplay, struct MetaPingData,
sanity_check_timestamps, meta_display_open, event_callback,
meta_spew_event, meta_display_set_grab_op_cursor,
meta_display_begin_grab_op, meta_display_end_grab_op,
meta_display_ping_timeout, meta_display_ping_window,
process_pong_message, timestamp_too_old,
meta_display_set_input_focus_window):
* src/keybindings.[ch] (grab_keyboard, ungrab_keyboard,
meta_screen_grab_all_keys, meta_window_grab_all_keys,
meta_window_ungrab_all_keys, error_on_generic_command,
error_on_command, error_on_terminal_command):
* src/metacity-dialog.c (on_realize, warn_about_no_sm_support,
error_about_command, main):
* src/screen.[ch] (struct _MetaScreen, meta_screen_new,
meta_screen_show_desktop, meta_screen_apply_startup_properties):
* src/session.c (warn_about_lame_clients_and_finish_interact):
* src/window.[ch] (struct _MetaWindow,
intervening_user_event_occurred, window_activate,
meta_window_delete, meta_window_focus,
meta_window_send_icccm_message, meta_window_client_message,
menu_callback, meta_window_show_menu, struct EventScannerData,
check_use_this_motion_notify, meta_window_begin_grab_op,
meta_window_set_user_time):
* src/workspace.[ch] (focus_ancestor_or_mru_window,
meta_workspace_activate_with_focus, meta_workspace_activate,
meta_workspace_focus_default_window,
focus_ancestor_or_mru_window):
Fix issues on 64-bit machines with timestamps by using guint32
(like gtk+ does) instead of Time. #348305
2006-09-13 12:32:33 -04:00
window - > desc , timestamp , source_indication ) ;
Big patch to cover about 6 different issues in order to correct rare
2005-02-20 Elijah Newren <newren@gmail.com>
Big patch to cover about 6 different issues in order to correct
rare problems with timestamps (make sure window selected in
tasklist actually gets focus, sanity check timestamps to avoid
rogue apps hosing the system, correct the updating of
net_wm_user_time, correctly handle timestamps of 0 when comparing
xserver timestamps for those who have had their systems up for
over 25 days or so, add some debugging information to verbose
logs, some code cleanups). Fixes all issues listed in #167358.
* src/display.h: (struct _MetaDisplay): clarify comment on
last_focus_time, introduce a new variable--last_user_time,
(XSERVER_TIME_IS_BEFORE macro): put this functionality into a
separate macro and then introduce a new macro with this name that
uses the old one but adds additional special-case checks for
timestamps that are 0, (comment to
meta_display_set_input_focus_window): add information about how
last_user_time should be used in this function
* src/display.c (santiy_check_timestamps): new function,
(meta_display_open): intialize display->last_user_time,
(meta_display_get_current_time_roundtrip): use the timestamp,
which is known to be good, in order to sanity_check_timestamps,
(event_callback): use the new meta_window_ste_user_time() function
in order to correct problems, use the timestamp of KeyPress and
ButtonPress events, which are known to be good, in order to
sanity_check_timestamps, (timestamp_too_old): new function for
common behavior of meta_display_focus_the_no_focus_window and
meta_display_set_input_focus_window, with added checking for
display->last_user_time in addition to display->last_focus_time,
(meta_display_set_input_focus_window): replace some of the code
with a call to timestamp_too_old(),
(meta_display_focus_the_no_focus_window): replace some of th ecode
with a call to timestamp_too_old()
* src/window.h: (meta_window_set_user_time): new function to
abstract the many things that need to be done when updating the
net_wm_user_time of any window
* src/window.c: (meta_window_activate): add debugging spew, make
sure the comparison is made with last_user_time NOT
last_focus_time, use meta_window_set_user_time() function in order
to correct problems, (meta_window_client_message): add a newline
to a debugging message to make them easier to read,
(meta_window_set_user_time): new function
* src/window-props.c (reload_net_wm_user_time): use the new
meta_window_ste_user_time() function in order to correct problems
2005-02-20 12:14:16 -05:00
2006-01-10 00:44:40 -05:00
/* Older EWMH spec didn't specify a timestamp; we decide to honor these only
* if the app specifies that it is a pager .
*
* Update : Unconditionally honor 0 timestamps for now ; we ' ll fight
* that battle later . Just remove the " FALSE && " in order to only
* honor 0 timestamps for pagers .
2005-02-06 11:52:26 -05:00
*/
2006-01-10 00:44:40 -05:00
can_ignore_outdated_timestamps =
( timestamp ! = 0 | | ( FALSE & & source_indication ! = META_CLIENT_TYPE_PAGER ) ) ;
Big patch to cover about 6 different issues in order to correct rare
2005-02-20 Elijah Newren <newren@gmail.com>
Big patch to cover about 6 different issues in order to correct
rare problems with timestamps (make sure window selected in
tasklist actually gets focus, sanity check timestamps to avoid
rogue apps hosing the system, correct the updating of
net_wm_user_time, correctly handle timestamps of 0 when comparing
xserver timestamps for those who have had their systems up for
over 25 days or so, add some debugging information to verbose
logs, some code cleanups). Fixes all issues listed in #167358.
* src/display.h: (struct _MetaDisplay): clarify comment on
last_focus_time, introduce a new variable--last_user_time,
(XSERVER_TIME_IS_BEFORE macro): put this functionality into a
separate macro and then introduce a new macro with this name that
uses the old one but adds additional special-case checks for
timestamps that are 0, (comment to
meta_display_set_input_focus_window): add information about how
last_user_time should be used in this function
* src/display.c (santiy_check_timestamps): new function,
(meta_display_open): intialize display->last_user_time,
(meta_display_get_current_time_roundtrip): use the timestamp,
which is known to be good, in order to sanity_check_timestamps,
(event_callback): use the new meta_window_ste_user_time() function
in order to correct problems, use the timestamp of KeyPress and
ButtonPress events, which are known to be good, in order to
sanity_check_timestamps, (timestamp_too_old): new function for
common behavior of meta_display_focus_the_no_focus_window and
meta_display_set_input_focus_window, with added checking for
display->last_user_time in addition to display->last_focus_time,
(meta_display_set_input_focus_window): replace some of the code
with a call to timestamp_too_old(),
(meta_display_focus_the_no_focus_window): replace some of th ecode
with a call to timestamp_too_old()
* src/window.h: (meta_window_set_user_time): new function to
abstract the many things that need to be done when updating the
net_wm_user_time of any window
* src/window.c: (meta_window_activate): add debugging spew, make
sure the comparison is made with last_user_time NOT
last_focus_time, use meta_window_set_user_time() function in order
to correct problems, (meta_window_client_message): add a newline
to a debugging message to make them easier to read,
(meta_window_set_user_time): new function
* src/window-props.c (reload_net_wm_user_time): use the new
meta_window_ste_user_time() function in order to correct problems
2005-02-20 12:14:16 -05:00
if ( XSERVER_TIME_IS_BEFORE ( timestamp , window - > display - > last_user_time ) & &
2006-01-10 00:44:40 -05:00
can_ignore_outdated_timestamps )
2005-02-06 11:52:26 -05:00
{
meta_topic ( META_DEBUG_FOCUS ,
Fix issues on 64-bit machines with timestamps by using guint32 (like gtk+
2006-09-13 Elijah Newren <newren gmail com>
* src/common.h (MetaWindowMenuFunc):
* src/core.[ch] (meta_core_user_lower_and_unfocus,
meta_core_user_focus, meta_core_show_window_menu,
meta_core_begin_grab_op, meta_core_end_grab_op):
* src/delete.c (delete_ping_reply_func, delete_ping_timeout_func,
meta_window_delete):
* src/display.[ch] (struct MetaDisplay, struct MetaPingData,
sanity_check_timestamps, meta_display_open, event_callback,
meta_spew_event, meta_display_set_grab_op_cursor,
meta_display_begin_grab_op, meta_display_end_grab_op,
meta_display_ping_timeout, meta_display_ping_window,
process_pong_message, timestamp_too_old,
meta_display_set_input_focus_window):
* src/keybindings.[ch] (grab_keyboard, ungrab_keyboard,
meta_screen_grab_all_keys, meta_window_grab_all_keys,
meta_window_ungrab_all_keys, error_on_generic_command,
error_on_command, error_on_terminal_command):
* src/metacity-dialog.c (on_realize, warn_about_no_sm_support,
error_about_command, main):
* src/screen.[ch] (struct _MetaScreen, meta_screen_new,
meta_screen_show_desktop, meta_screen_apply_startup_properties):
* src/session.c (warn_about_lame_clients_and_finish_interact):
* src/window.[ch] (struct _MetaWindow,
intervening_user_event_occurred, window_activate,
meta_window_delete, meta_window_focus,
meta_window_send_icccm_message, meta_window_client_message,
menu_callback, meta_window_show_menu, struct EventScannerData,
check_use_this_motion_notify, meta_window_begin_grab_op,
meta_window_set_user_time):
* src/workspace.[ch] (focus_ancestor_or_mru_window,
meta_workspace_activate_with_focus, meta_workspace_activate,
meta_workspace_focus_default_window,
focus_ancestor_or_mru_window):
Fix issues on 64-bit machines with timestamps by using guint32
(like gtk+ does) instead of Time. #348305
2006-09-13 12:32:33 -04:00
" last_user_time (%u) is more recent; ignoring "
Big patch to cover about 6 different issues in order to correct rare
2005-02-20 Elijah Newren <newren@gmail.com>
Big patch to cover about 6 different issues in order to correct
rare problems with timestamps (make sure window selected in
tasklist actually gets focus, sanity check timestamps to avoid
rogue apps hosing the system, correct the updating of
net_wm_user_time, correctly handle timestamps of 0 when comparing
xserver timestamps for those who have had their systems up for
over 25 days or so, add some debugging information to verbose
logs, some code cleanups). Fixes all issues listed in #167358.
* src/display.h: (struct _MetaDisplay): clarify comment on
last_focus_time, introduce a new variable--last_user_time,
(XSERVER_TIME_IS_BEFORE macro): put this functionality into a
separate macro and then introduce a new macro with this name that
uses the old one but adds additional special-case checks for
timestamps that are 0, (comment to
meta_display_set_input_focus_window): add information about how
last_user_time should be used in this function
* src/display.c (santiy_check_timestamps): new function,
(meta_display_open): intialize display->last_user_time,
(meta_display_get_current_time_roundtrip): use the timestamp,
which is known to be good, in order to sanity_check_timestamps,
(event_callback): use the new meta_window_ste_user_time() function
in order to correct problems, use the timestamp of KeyPress and
ButtonPress events, which are known to be good, in order to
sanity_check_timestamps, (timestamp_too_old): new function for
common behavior of meta_display_focus_the_no_focus_window and
meta_display_set_input_focus_window, with added checking for
display->last_user_time in addition to display->last_focus_time,
(meta_display_set_input_focus_window): replace some of the code
with a call to timestamp_too_old(),
(meta_display_focus_the_no_focus_window): replace some of th ecode
with a call to timestamp_too_old()
* src/window.h: (meta_window_set_user_time): new function to
abstract the many things that need to be done when updating the
net_wm_user_time of any window
* src/window.c: (meta_window_activate): add debugging spew, make
sure the comparison is made with last_user_time NOT
last_focus_time, use meta_window_set_user_time() function in order
to correct problems, (meta_window_client_message): add a newline
to a debugging message to make them easier to read,
(meta_window_set_user_time): new function
* src/window-props.c (reload_net_wm_user_time): use the new
meta_window_ste_user_time() function in order to correct problems
2005-02-20 12:14:16 -05:00
" _NET_ACTIVE_WINDOW message. \n " ,
window - > display - > last_user_time ) ;
2006-03-29 13:43:22 -05:00
meta_window_set_demands_attention ( window ) ;
2005-02-06 11:52:26 -05:00
return ;
}
2006-01-10 00:44:40 -05:00
2008-12-18 07:41:56 -05:00
/* For those stupid pagers, get a valid timestamp and show a warning */
Partial audit to fix timestamp usage. One step towards fixing #355180; see
2006-09-18 Elijah Newren <newren gmail com>
Partial audit to fix timestamp usage. One step towards fixing
#355180; see important comments in that bug.
* src/core.[ch] (meta_core_unshade, meta_core_shade):
* src/delete.c (meta_window_present_delete_dialog,
delete_ping_timeout_func):
* src/display.[ch] (meta_display_open, meta_display_close,
event_callback, meta_display_begin_grab_op,
process_selection_clear, meta_display_unmanage_screen,
meta_display_unmanage_windows_for_screen):
* src/frames.c (meta_frames_button_press_event):
* src/keybindings.c (handle_toggle_shade):
* src/main.c (main):
* src/screen.[ch] (update_num_workspaces, meta_screen_new,
meta_screen_free, prefs_changed_callback):
* src/window.[ch] (meta_window_free, finish_minimize,
implement_showing, meta_window_show, meta_window_maximize,
meta_window_make_fullscreen_internal,
meta_window_unmake_fullscreen, meta_window_shade,
meta_window_unshade, window_activate, send_sync_request,
meta_window_client_message, menu_callback,
meta_window_update_keyboard_resize):
Remove usage of CurrentTime, meta_display_get_current_time() and
meta_display_get_current_time_roundtrip() where possible, or
document why it isn't possible, or at very least add a FIXME with
some explanation of my laziness and what needs to be done.
2006-09-18 13:27:24 -04:00
if ( timestamp = = 0 )
{
meta_warning ( " meta_window_activate called by a pager with a 0 timestamp; "
" the pager needs to be fixed. \n " ) ;
timestamp = meta_display_get_current_time_roundtrip ( window - > display ) ;
}
2005-02-08 22:15:11 -05:00
Big patch to cover about 6 different issues in order to correct rare
2005-02-20 Elijah Newren <newren@gmail.com>
Big patch to cover about 6 different issues in order to correct
rare problems with timestamps (make sure window selected in
tasklist actually gets focus, sanity check timestamps to avoid
rogue apps hosing the system, correct the updating of
net_wm_user_time, correctly handle timestamps of 0 when comparing
xserver timestamps for those who have had their systems up for
over 25 days or so, add some debugging information to verbose
logs, some code cleanups). Fixes all issues listed in #167358.
* src/display.h: (struct _MetaDisplay): clarify comment on
last_focus_time, introduce a new variable--last_user_time,
(XSERVER_TIME_IS_BEFORE macro): put this functionality into a
separate macro and then introduce a new macro with this name that
uses the old one but adds additional special-case checks for
timestamps that are 0, (comment to
meta_display_set_input_focus_window): add information about how
last_user_time should be used in this function
* src/display.c (santiy_check_timestamps): new function,
(meta_display_open): intialize display->last_user_time,
(meta_display_get_current_time_roundtrip): use the timestamp,
which is known to be good, in order to sanity_check_timestamps,
(event_callback): use the new meta_window_ste_user_time() function
in order to correct problems, use the timestamp of KeyPress and
ButtonPress events, which are known to be good, in order to
sanity_check_timestamps, (timestamp_too_old): new function for
common behavior of meta_display_focus_the_no_focus_window and
meta_display_set_input_focus_window, with added checking for
display->last_user_time in addition to display->last_focus_time,
(meta_display_set_input_focus_window): replace some of the code
with a call to timestamp_too_old(),
(meta_display_focus_the_no_focus_window): replace some of th ecode
with a call to timestamp_too_old()
* src/window.h: (meta_window_set_user_time): new function to
abstract the many things that need to be done when updating the
net_wm_user_time of any window
* src/window.c: (meta_window_activate): add debugging spew, make
sure the comparison is made with last_user_time NOT
last_focus_time, use meta_window_set_user_time() function in order
to correct problems, (meta_window_client_message): add a newline
to a debugging message to make them easier to read,
(meta_window_set_user_time): new function
* src/window-props.c (reload_net_wm_user_time): use the new
meta_window_ste_user_time() function in order to correct problems
2005-02-20 12:14:16 -05:00
meta_window_set_user_time ( window , timestamp ) ;
2004-08-07 20:02:02 -04:00
2002-04-28 00:52:26 -04:00
/* disable show desktop mode unless we're a desktop component */
2003-05-16 17:59:08 -04:00
maybe_leave_show_desktop_mode ( window ) ;
2008-12-18 07:41:56 -05:00
2006-08-07 14:08:27 -04:00
/* Get window on current or given workspace */
if ( workspace = = NULL )
workspace = window - > screen - > active_workspace ;
2008-12-21 00:35:30 -05:00
2009-03-17 08:27:13 -04:00
/* For non-transient windows, we just set up a pulsing indicator,
2008-12-21 00:35:30 -05:00
rather than move windows or workspaces .
See http : //bugzilla.gnome.org/show_bug.cgi?id=482354 */
2013-09-16 08:44:37 -04:00
if ( window - > transient_for = = NULL & &
2008-12-21 00:35:30 -05:00
! meta_window_located_on_workspace ( window , workspace ) )
{
meta_window_set_demands_attention ( window ) ;
/* We've marked it as demanding, don't need to do anything else. */
return ;
}
2013-09-16 08:44:37 -04:00
else if ( window - > transient_for ! = NULL )
2008-12-21 00:35:30 -05:00
{
2009-03-17 08:27:13 -04:00
/* Move transients to current workspace - preference dialogs should appear over
2008-12-21 00:35:30 -05:00
the source window . */
2006-08-07 14:08:27 -04:00
meta_window_change_workspace ( window , workspace ) ;
2008-12-21 00:35:30 -05:00
}
2008-12-18 07:41:56 -05:00
2002-02-08 01:50:09 -05:00
if ( window - > shaded )
Partial audit to fix timestamp usage. One step towards fixing #355180; see
2006-09-18 Elijah Newren <newren gmail com>
Partial audit to fix timestamp usage. One step towards fixing
#355180; see important comments in that bug.
* src/core.[ch] (meta_core_unshade, meta_core_shade):
* src/delete.c (meta_window_present_delete_dialog,
delete_ping_timeout_func):
* src/display.[ch] (meta_display_open, meta_display_close,
event_callback, meta_display_begin_grab_op,
process_selection_clear, meta_display_unmanage_screen,
meta_display_unmanage_windows_for_screen):
* src/frames.c (meta_frames_button_press_event):
* src/keybindings.c (handle_toggle_shade):
* src/main.c (main):
* src/screen.[ch] (update_num_workspaces, meta_screen_new,
meta_screen_free, prefs_changed_callback):
* src/window.[ch] (meta_window_free, finish_minimize,
implement_showing, meta_window_show, meta_window_maximize,
meta_window_make_fullscreen_internal,
meta_window_unmake_fullscreen, meta_window_shade,
meta_window_unshade, window_activate, send_sync_request,
meta_window_client_message, menu_callback,
meta_window_update_keyboard_resize):
Remove usage of CurrentTime, meta_display_get_current_time() and
meta_display_get_current_time_roundtrip() where possible, or
document why it isn't possible, or at very least add a FIXME with
some explanation of my laziness and what needs to be done.
2006-09-18 13:27:24 -04:00
meta_window_unshade ( window , timestamp ) ;
2001-10-29 02:00:46 -05:00
on unminimize, queue calc_showing on all transients
2002-05-05 Havoc Pennington <hp@pobox.com>
* src/window.c (meta_window_unminimize): on unminimize, queue
calc_showing on all transients
(meta_window_activate): on activate, unminimize all a window's
ancestors, not just the window itself.
* src/workspace.c (set_work_area_hint): don't increment "tmp" by
16 unsigned long, increment by 4
* src/window.c (meta_window_free): if a window isn't minimized,
restore its WM_STATE to NormalState instead of IconicState,
since IconicState on initial window map means that the window
should be minimized.
* src/workspace.c (meta_workspace_invalidate_work_area): queue an
idle to recompute the work area hint.
(set_work_area_hint): we need 4*num_workspaces ints, not just
num_workspaces.
* src/screen.c (meta_screen_new): add work_area_idle field,
handle it on screen shutdown
* src/common.h (META_PRIORITY_PREFS_NOTIFY,
META_PRIORITY_WORK_AREA_HINT): define some idle priorities
* src/window.c (meta_window_calc_showing): hide windows if
their parent window is minimized
(meta_window_minimize): also queue_calc_showing on all
transients of the window being minimized
* src/place.c (constrain_placement): function to apply
placement-time-only constraints, such as "not off the left of the
screen"
(meta_window_place): put dialogs down a bit over their parent,
not right at the top.
(meta_window_place): when centering a dialog, center it
on the current xinerama screen, rather than the entire
screen.
* src/screen.c (meta_screen_get_current_xinerama): new function,
but not implemented
2002-05-05 01:41:13 -04:00
unminimize_window_and_all_transient_parents ( window ) ;
Add a raise on click option, basically only because all the major distros
2006-01-10 Elijah Newren <newren@gmail.com>
Add a raise on click option, basically only because all the major
distros are patching it in anyway. See #326156.
* src/metacity.schemas.in: add the new gconf key and explanation
* src/prefs.[ch] (#define KEY_RAISE_ON_CLICK, static gboolean
raise_on_click, update_raise_on_click, meta_prefs_init,
change_notify, meta_prefs_get_raise_on_click,
meta_preference_to_string):
Add all the normal preference handling stuff for this new
raise-on-click option.
* src/core.c (meta_core_show_window_menu):
* src/display.c (event_callback, meta_display_begin_grab_op):
* src/window.c (window_activate, meta_window_configure_request, ):
Only raise the window if in raise_on_click mode.
* src/display.c (meta_display_begin_grab_op,
meta_display_end_grab_op, meta_display_check_threshold_reached):
* src/display.h (struct MetaDisplay):
* src/window.c (meta_window_handle_mouse_grab_op_event):
if not in raise-on-click mode only raise on button release if the
click didn't start a move or resize operation; needs a few extra
MetaDisplay fields to handle this
* src/core.c (meta_core_user_lower_and_unfocus):
no need to do the MRU shuffling if not maintaining the stacking
order == MRU order invariant
* src/frames.c (meta_frames_button_press_event):
* src/window.c (meta_window_begin_grab_op):
remove an unneeded window raising that is already handled elsewhere
2006-01-10 14:35:03 -05:00
if ( meta_prefs_get_raise_on_click ( ) | |
source_indication = = META_CLIENT_TYPE_PAGER )
meta_window_raise ( window ) ;
2005-07-23 21:13:41 -04:00
meta_topic ( META_DEBUG_FOCUS ,
" Focusing window %s due to activation \n " ,
window - > desc ) ;
meta_window_focus ( window , timestamp ) ;
2012-12-24 12:19:26 -05:00
meta_window_check_alive ( window , timestamp ) ;
2001-10-29 02:00:46 -05:00
}
2001-06-10 23:24:20 -04:00
2006-01-10 00:44:40 -05:00
/* This function exists since most of the functionality in window_activate
Comprehensively rename to Mutter
Code:
All references in the code not related to themes, keybindings, or
GConf were changed from 'metacity' to 'mutter'. This includes, among other
things, strings, comments, the atoms used in the message protocol, and
the envvars used for debugging. The GConf schema file was reduced to
the 3 settings new to mutter.
The overall version was brought up to 2.27 to match current gnome.
Structure:
All files named '*metacity*' were renamed '*mutter*' with appropriate
changes in the automake system. Files removed are
doc/creating_themes, src/themes, doc/metacity-theme.dtd,
metacity.doap. These files will eventually end up in an external
gnome-wm-data module.
Installation location:
On the filesystem the mutter-plugindir was change from
$(libdir)/metacity/plugins/clutter to just $(libdir)/mutter/plugins.
The mutter-plugins.pc.in reflects these changes.
Note:
mutter.desktop.in and mutter-wm.desktop both continue to have
X-GNOME-WMSettingsModule=metacity set. This allows
gnome-control-center to continue using libmetacity.so for
configuration. This is fine since most the general keybindings and wm
settings are being read from /apps/metacity/* in gconf.
2009-06-10 06:29:20 -04:00
* is useful for Mutter , but Mutter shouldn ' t need to specify a client
2006-01-10 00:44:40 -05:00
* type for itself . ; - )
*/
void
meta_window_activate ( MetaWindow * window ,
guint32 timestamp )
{
2009-06-14 08:11:29 -04:00
g_return_if_fail ( ! window - > override_redirect ) ;
2006-01-10 00:44:40 -05:00
/* We're not really a pager, but the behavior we want is the same as if
* we were such . If we change the pager behavior later , we could revisit
* this and just add extra flags to window_activate .
*/
2014-01-23 16:33:53 -05:00
meta_window_activate_full ( window , timestamp , META_CLIENT_TYPE_PAGER , NULL ) ;
2006-08-07 14:08:27 -04:00
}
void
meta_window_activate_with_workspace ( MetaWindow * window ,
guint32 timestamp ,
MetaWorkspace * workspace )
{
2009-06-14 08:11:29 -04:00
g_return_if_fail ( ! window - > override_redirect ) ;
2014-01-23 16:33:53 -05:00
meta_window_activate_full ( window , timestamp , META_CLIENT_TYPE_APPLICATION , workspace ) ;
2006-01-10 00:44:40 -05:00
}
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
/* Manually fix all the weirdness explained in the big comment at the
* beginning of meta_window_move_resize_internal ( ) giving positions
* expected by meta_window_constrain ( i . e . positions & sizes of the
* internal or client window ) .
2003-03-08 18:43:17 -05:00
*/
2001-07-28 02:35:19 -04:00
static void
2001-06-10 23:24:20 -04:00
adjust_for_gravity ( MetaWindow * window ,
2011-07-12 00:37:41 -04:00
MetaFrameBorders * borders ,
2001-06-10 23:24:20 -04:00
gboolean coords_assume_border ,
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
int gravity ,
MetaRectangle * rect )
2001-06-10 23:24:20 -04:00
{
int ref_x , ref_y ;
int bw ;
int child_x , child_y ;
int frame_width , frame_height ;
2008-12-18 07:41:56 -05:00
2001-06-10 23:24:20 -04:00
if ( coords_assume_border )
bw = window - > border_width ;
else
bw = 0 ;
2011-07-12 00:37:41 -04:00
if ( borders )
2001-06-10 23:24:20 -04:00
{
2011-09-22 14:17:01 -04:00
child_x = borders - > visible . left ;
child_y = borders - > visible . top ;
frame_width = child_x + rect - > width + borders - > visible . right ;
frame_height = child_y + rect - > height + borders - > visible . bottom ;
2001-06-10 23:24:20 -04:00
}
else
{
child_x = 0 ;
child_y = 0 ;
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
frame_width = rect - > width ;
frame_height = rect - > height ;
2001-06-10 23:24:20 -04:00
}
2008-12-18 07:41:56 -05:00
2001-06-10 23:24:20 -04:00
/* We're computing position to pass to window_move, which is
* the position of the client window ( StaticGravity basically )
*
* ( see WM spec description of gravity computation , but note that
* their formulas assume we ' re honoring the border width , rather
* than compensating for having turned it off )
*/
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
switch ( gravity )
2001-06-10 23:24:20 -04:00
{
case NorthWestGravity :
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
ref_x = rect - > x ;
ref_y = rect - > y ;
2001-06-10 23:24:20 -04:00
break ;
case NorthGravity :
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
ref_x = rect - > x + rect - > width / 2 + bw ;
ref_y = rect - > y ;
2001-06-10 23:24:20 -04:00
break ;
case NorthEastGravity :
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
ref_x = rect - > x + rect - > width + bw * 2 ;
ref_y = rect - > y ;
2001-06-10 23:24:20 -04:00
break ;
case WestGravity :
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
ref_x = rect - > x ;
ref_y = rect - > y + rect - > height / 2 + bw ;
2001-06-10 23:24:20 -04:00
break ;
case CenterGravity :
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
ref_x = rect - > x + rect - > width / 2 + bw ;
ref_y = rect - > y + rect - > height / 2 + bw ;
2001-06-10 23:24:20 -04:00
break ;
case EastGravity :
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
ref_x = rect - > x + rect - > width + bw * 2 ;
ref_y = rect - > y + rect - > height / 2 + bw ;
2001-06-10 23:24:20 -04:00
break ;
case SouthWestGravity :
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
ref_x = rect - > x ;
ref_y = rect - > y + rect - > height + bw * 2 ;
2001-06-10 23:24:20 -04:00
break ;
case SouthGravity :
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
ref_x = rect - > x + rect - > width / 2 + bw ;
ref_y = rect - > y + rect - > height + bw * 2 ;
2001-06-10 23:24:20 -04:00
break ;
case SouthEastGravity :
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
ref_x = rect - > x + rect - > width + bw * 2 ;
ref_y = rect - > y + rect - > height + bw * 2 ;
2001-06-10 23:24:20 -04:00
break ;
case StaticGravity :
default :
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
ref_x = rect - > x ;
ref_y = rect - > y ;
2001-06-10 23:24:20 -04:00
break ;
}
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
switch ( gravity )
2001-06-10 23:24:20 -04:00
{
case NorthWestGravity :
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
rect - > x = ref_x + child_x ;
rect - > y = ref_y + child_y ;
2001-06-10 23:24:20 -04:00
break ;
case NorthGravity :
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
rect - > x = ref_x - frame_width / 2 + child_x ;
rect - > y = ref_y + child_y ;
2001-06-10 23:24:20 -04:00
break ;
case NorthEastGravity :
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
rect - > x = ref_x - frame_width + child_x ;
rect - > y = ref_y + child_y ;
2001-06-10 23:24:20 -04:00
break ;
case WestGravity :
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
rect - > x = ref_x + child_x ;
rect - > y = ref_y - frame_height / 2 + child_y ;
2001-06-10 23:24:20 -04:00
break ;
case CenterGravity :
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
rect - > x = ref_x - frame_width / 2 + child_x ;
rect - > y = ref_y - frame_height / 2 + child_y ;
2001-06-10 23:24:20 -04:00
break ;
case EastGravity :
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
rect - > x = ref_x - frame_width + child_x ;
rect - > y = ref_y - frame_height / 2 + child_y ;
2001-06-10 23:24:20 -04:00
break ;
case SouthWestGravity :
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
rect - > x = ref_x + child_x ;
rect - > y = ref_y - frame_height + child_y ;
2001-06-10 23:24:20 -04:00
break ;
case SouthGravity :
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
rect - > x = ref_x - frame_width / 2 + child_x ;
rect - > y = ref_y - frame_height + child_y ;
2001-06-10 23:24:20 -04:00
break ;
case SouthEastGravity :
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
rect - > x = ref_x - frame_width + child_x ;
rect - > y = ref_y - frame_height + child_y ;
2001-06-10 23:24:20 -04:00
break ;
case StaticGravity :
default :
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
rect - > x = ref_x ;
rect - > y = ref_y ;
2001-06-10 23:24:20 -04:00
break ;
2001-06-07 22:17:48 -04:00
}
}
2011-06-08 11:26:03 -04:00
void
meta_window_create_sync_request_alarm ( MetaWindow * window )
{
# ifdef HAVE_XSYNC
XSyncAlarmAttributes values ;
XSyncValue init ;
if ( window - > sync_request_counter = = None | |
window - > sync_request_alarm ! = None )
return ;
meta_error_trap_push_with_return ( window - > display ) ;
2011-06-08 13:58:54 -04:00
/* In the new (extended style), the counter value is initialized by
* the client before mapping the window . In the old style , we ' re
* responsible for setting the initial value of the counter .
2011-06-08 11:26:03 -04:00
*/
2011-06-08 13:58:54 -04:00
if ( window - > extended_sync_request_counter )
{
if ( ! XSyncQueryCounter ( window - > display - > xdisplay ,
window - > sync_request_counter ,
& init ) )
{
meta_error_trap_pop_with_return ( window - > display ) ;
window - > sync_request_counter = None ;
return ;
}
window - > sync_request_serial =
XSyncValueLow32 ( init ) + ( ( gint64 ) XSyncValueHigh32 ( init ) < < 32 ) ;
}
else
{
XSyncIntToValue ( & init , 0 ) ;
XSyncSetCounter ( window - > display - > xdisplay ,
window - > sync_request_counter , init ) ;
window - > sync_request_serial = 0 ;
}
2011-06-08 11:26:03 -04:00
values . trigger . counter = window - > sync_request_counter ;
2011-06-08 13:58:54 -04:00
values . trigger . test_type = XSyncPositiveComparison ;
2011-06-08 11:26:03 -04:00
/* Initialize to one greater than the current value */
values . trigger . value_type = XSyncRelative ;
XSyncIntToValue ( & values . trigger . wait_value , 1 ) ;
/* After triggering, increment test_value by this until
* until the test condition is false */
XSyncIntToValue ( & values . delta , 1 ) ;
/* we want events (on by default anyway) */
values . events = True ;
window - > sync_request_alarm = XSyncCreateAlarm ( window - > display - > xdisplay ,
XSyncCACounter |
XSyncCAValueType |
XSyncCAValue |
XSyncCATestType |
XSyncCADelta |
XSyncCAEvents ,
& values ) ;
if ( meta_error_trap_pop_with_return ( window - > display ) = = Success )
meta_display_register_sync_alarm ( window - > display , & window - > sync_request_alarm , window ) ;
else
{
window - > sync_request_alarm = None ;
window - > sync_request_counter = None ;
}
# endif
}
void
meta_window_destroy_sync_request_alarm ( MetaWindow * window )
{
# ifdef HAVE_XSYNC
if ( window - > sync_request_alarm ! = None )
{
/* Has to be unregistered _before_ clearing the structure field */
meta_display_unregister_sync_alarm ( window - > display , window - > sync_request_alarm ) ;
XSyncDestroyAlarm ( window - > display - > xdisplay ,
window - > sync_request_alarm ) ;
window - > sync_request_alarm = None ;
}
# endif /* HAVE_XSYNC */
}
2011-06-13 17:53:23 -04:00
/**
* meta_window_updates_are_frozen :
* @ window : a # MetaWindow
*
* Gets whether the compositor should be updating the window contents ;
* window content updates may be frozen at client request by setting
* an odd value in the extended _NET_WM_SYNC_REQUEST_COUNTER counter r
* by the window manager during a resize operation while waiting for
* the client to redraw .
*
* Return value : % TRUE if updates are currently frozen
*/
gboolean
meta_window_updates_are_frozen ( MetaWindow * window )
{
2011-06-08 13:58:54 -04:00
# ifdef HAVE_XSYNC
if ( window - > extended_sync_request_counter & &
window - > sync_request_serial % 2 = = 1 )
return TRUE ;
2013-03-13 13:09:33 -04:00
if ( window - > sync_request_serial < window - > sync_request_wait_serial )
return TRUE ;
# endif
2011-06-13 17:53:23 -04:00
2013-03-13 13:09:33 -04:00
return FALSE ;
2011-06-13 17:53:23 -04:00
}
2010-09-10 05:55:20 -04:00
static gboolean
window: make determination of attached dialog windows more consistent
Different bits of code were using slightly different checks to test
whether a window was an attached dialog. Add a new
meta_window_is_attached_dialog(), and use that everywhere.
Also, freeze the is-attached status when the window is first shown,
rather than recomputing it each time the caller asks, since this could
cause problems if a window changes its type after it has already been
attached, etc. However, if an attached window's parent is destroyed,
or an attached window changes its transient-for, then fix things up by
destroying the old MetaWindow and creating a new one (causing
compositor unmap and map events to be fired off, allowing the display
of the window to be fixed up).
Remove some code in display.c that tried to fix existing windows if
the gconf setting changed, but which didn't actually do anything (at
least under gnome-shell). However, if 654643 was fixed then the new
behavior with this patch would be that changing the gconf setting
would affect new dialogs, but not existing ones.
https://bugzilla.gnome.org/show_bug.cgi?id=646761
2011-04-05 10:14:52 -04:00
maybe_move_attached_dialog ( MetaWindow * window ,
void * data )
2010-09-10 05:55:20 -04:00
{
window: make determination of attached dialog windows more consistent
Different bits of code were using slightly different checks to test
whether a window was an attached dialog. Add a new
meta_window_is_attached_dialog(), and use that everywhere.
Also, freeze the is-attached status when the window is first shown,
rather than recomputing it each time the caller asks, since this could
cause problems if a window changes its type after it has already been
attached, etc. However, if an attached window's parent is destroyed,
or an attached window changes its transient-for, then fix things up by
destroying the old MetaWindow and creating a new one (causing
compositor unmap and map events to be fired off, allowing the display
of the window to be fixed up).
Remove some code in display.c that tried to fix existing windows if
the gconf setting changed, but which didn't actually do anything (at
least under gnome-shell). However, if 654643 was fixed then the new
behavior with this patch would be that changing the gconf setting
would affect new dialogs, but not existing ones.
https://bugzilla.gnome.org/show_bug.cgi?id=646761
2011-04-05 10:14:52 -04:00
if ( meta_window_is_attached_dialog ( window ) )
2010-09-10 05:55:20 -04:00
/* It ignores x,y for such a dialog */
meta_window_move ( window , FALSE , 0 , 0 ) ;
return FALSE ;
}
2011-02-28 09:21:35 -05:00
/**
* meta_window_get_monitor :
* @ window : a # MetaWindow
*
* Gets index of the monitor that this window is on .
*
* Return Value : The index of the monitor in the screens monitor list
*/
int
meta_window_get_monitor ( MetaWindow * window )
{
return window - > monitor - > number ;
}
2011-05-20 04:56:12 -04:00
/* This is called when the monitor setup has changed. The window->monitor
* reference is still " valid " , but refer to the previous monitor setup */
2011-02-25 10:37:05 -05:00
void
2011-05-20 04:56:12 -04:00
meta_window_update_for_monitors_changed ( MetaWindow * window )
{
const MetaMonitorInfo * old , * new ;
int i ;
2012-10-04 13:57:49 -04:00
if ( window - > type = = META_WINDOW_DESKTOP )
return ;
2013-06-24 03:49:44 -04:00
if ( window - > override_redirect )
{
meta_window_update_monitor ( window ) ;
return ;
}
2011-05-20 04:56:12 -04:00
old = window - > monitor ;
/* Start on primary */
new = & window - > screen - > monitor_infos [ window - > screen - > primary_monitor_index ] ;
/* But, if we can find the old output on a new monitor, use that */
for ( i = 0 ; i < window - > screen - > n_monitor_infos ; i + + )
{
MetaMonitorInfo * info = & window - > screen - > monitor_infos [ i ] ;
2013-07-18 07:09:16 -04:00
if ( info - > output_id ! = 0 & &
info - > output_id = = old - > output_id )
2011-05-20 04:56:12 -04:00
{
new = info ;
break ;
}
}
2011-10-08 11:00:16 -04:00
if ( window - > tile_mode ! = META_TILE_NONE )
window - > tile_monitor_number = new - > number ;
2011-05-20 04:56:12 -04:00
/* This will eventually reach meta_window_update_monitor that
* will send leave / enter - monitor events . The old ! = new monitor
* check will always fail ( due to the new monitor_infos set ) so
* we will always send the events , even if the new and old monitor
* index is the same . That is right , since the enumeration of the
* monitors changed and the same index could be refereing
* to a different monitor . */
meta_window_move_between_rects ( window ,
& old - > rect ,
& new - > rect ) ;
}
2014-03-18 10:36:00 -04:00
void
2011-02-25 10:37:05 -05:00
meta_window_update_monitor ( MetaWindow * window )
{
2011-02-28 07:48:28 -05:00
const MetaMonitorInfo * old ;
old = window - > monitor ;
2011-02-25 10:37:05 -05:00
window - > monitor = meta_screen_get_monitor_for_window ( window - > screen , window ) ;
2011-02-28 07:48:28 -05:00
if ( old ! = window - > monitor )
{
meta_window_update_on_all_workspaces ( window ) ;
/* If workspaces only on primary and we moved back to primary, ensure that the
* window is now in that workspace . We do this because while the window is on a
* non - primary monitor it is always visible , so it would be very jarring if it
* disappeared when it crossed the monitor border .
* The one time we want it to both change to the primary monitor and a non - active
* workspace is when dropping the window on some other workspace thumbnail directly .
* That should be handled by explicitly moving the window before changing the
* workspace
2011-05-20 04:56:12 -04:00
* Don ' t do this if old = = NULL , because thats what happens when starting up , and
* we don ' t want to move all windows around from a previous WM instance . Nor do
* we want it when moving from one primary monitor to another ( can happen during
* screen reconfiguration .
2011-02-28 07:48:28 -05:00
*/
if ( meta_prefs_get_workspaces_only_on_primary ( ) & &
meta_window_is_on_primary_monitor ( window ) & &
2011-05-20 04:56:12 -04:00
old ! = NULL & & ! old - > is_primary & &
2011-02-28 07:48:28 -05:00
window - > screen - > active_workspace ! = window - > workspace )
meta_window_change_workspace ( window , window - > screen - > active_workspace ) ;
2011-02-28 15:26:11 -05:00
if ( old )
g_signal_emit_by_name ( window - > screen , " window-left-monitor " , old - > number , window ) ;
g_signal_emit_by_name ( window - > screen , " window-entered-monitor " , window - > monitor - > number , window ) ;
2012-03-12 23:18:03 -04:00
/* If we're changing monitors, we need to update the has_maximize_func flag,
* as the working area has changed . */
2014-02-01 18:58:58 -05:00
meta_window_recalc_features ( window ) ;
2011-02-28 07:48:28 -05:00
}
2011-02-25 10:37:05 -05:00
}
2001-06-07 22:17:48 -04:00
static void
2007-04-09 19:00:55 -04:00
meta_window_move_resize_internal ( MetaWindow * window ,
MetaMoveResizeFlags flags ,
int gravity ,
int root_x_nw ,
int root_y_nw ,
int w ,
int h )
2001-06-09 17:58:30 -04:00
{
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
/* meta_window_move_resize_internal gets called with very different
2007-04-09 19:00:55 -04:00
* meanings for root_x_nw and root_y_nw . w & h are always the area
* of the inner or client window ( i . e . excluding the frame ) and
* gravity is the relevant gravity associated with the request ( note
* that gravity is ignored for move - only operations unless its
* e . g . a configure request ) . The location is different for
* different cases because of how this function gets called ; note
* that in all cases what we want to find out is the upper left
* corner of the position of the inner window :
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
*
2007-04-09 19:00:55 -04:00
* Case | Called from ( flags ; gravity )
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
* - - - - - + - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
* 1 | A resize only ConfigureRequest
* 1 | meta_window_resize
* 1 | meta_window_resize_with_gravity
* 2 | New window
* 2 | Session restore
2007-04-09 19:00:55 -04:00
* 2 | A not - resize - only ConfigureRequest / net_moveresize_window request
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
* 3 | meta_window_move
* 3 | meta_window_move_resize
2013-09-03 06:00:29 -04:00
* 4 | meta_window_move_resize_wayland
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
*
* For each of the cases , root_x_nw and root_y_nw must be treated as follows :
*
* ( 1 ) They should be entirely ignored ; instead the previous position
* and size of the window should be resized according to the given
* gravity in order to determine the new position of the window .
* ( 2 ) Needs to be fixed up by adjust_for_gravity ( ) as these
* coordinates are relative to some corner or side of the outer
* window ( except for the case of StaticGravity ) and we want to
* know the location of the upper left corner of the inner window .
2013-09-03 06:00:29 -04:00
* ( 3 ) These values are already the desired position of the NW corner
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
* of the inner window
2013-09-03 06:00:29 -04:00
* ( 4 ) These values are already the desired position of the NW corner
* of the inner window ( which is also the outer window , because
* we don ' t decorate wayland clients ) , and the client has acknowledged
* the window size change .
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
*/
2011-03-22 15:36:12 -04:00
gboolean frame_shape_changed = FALSE ;
2014-03-19 10:30:12 -04:00
2001-08-28 23:37:03 -04:00
gboolean is_configure_request ;
gboolean do_gravity_adjust ;
2001-09-10 22:57:05 -04:00
gboolean is_user_action ;
2013-09-03 06:00:29 -04:00
gboolean is_wayland_resize ;
2013-02-15 21:27:00 -05:00
gboolean did_placement ;
2002-01-19 18:59:11 -05:00
/* used for the configure request, but may not be final
* destination due to StaticGravity etc .
*/
2003-03-08 18:43:17 -05:00
MetaRectangle new_rect ;
MetaRectangle old_rect ;
2014-03-19 10:30:12 -04:00
MetaRectangle requested_rect ;
MetaMoveResizeResultFlags result ;
MetaFrameBorders borders ;
2008-12-18 07:41:56 -05:00
2009-06-15 18:25:07 -04:00
g_return_if_fail ( ! window - > override_redirect ) ;
2001-08-28 23:37:03 -04:00
is_configure_request = ( flags & META_IS_CONFIGURE_REQUEST ) ! = 0 ;
do_gravity_adjust = ( flags & META_DO_GRAVITY_ADJUST ) ! = 0 ;
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
is_user_action = ( flags & META_IS_USER_ACTION ) ! = 0 ;
2013-09-03 06:00:29 -04:00
is_wayland_resize = ( flags & META_IS_WAYLAND_RESIZE ) ! = 0 ;
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
2013-09-03 06:00:29 -04:00
/* The action has to be a move, a resize or the wayland client
* acking our choice of size .
*/
g_assert ( flags & ( META_IS_MOVE_ACTION | META_IS_RESIZE_ACTION | META_IS_WAYLAND_RESIZE ) ) ;
2008-12-18 07:41:56 -05:00
2002-05-29 00:14:15 -04:00
/* We don't need it in the idle queue anymore. */
Refactor thrice-duplicated queue code in window.c. Closes #376760.
2007-06-10 Thomas Thurman <thomas@thurman.org.uk>
Refactor thrice-duplicated queue code in window.c. Closes #376760.
* src/window.c (meta_window_queue, meta_window_unqueue):
New functions.
* src/window.[ch] (meta_window_unqueue_*, meta_window_queue_*):
Removed functions.
* src/window.c (meta_window_new_with_attrs, meta_window_free,
meta_window_flush_calc_showing, queue_calc_showing_func,
meta_window_minimize, meta_window_unminimize, meta_window_maximize,
meta_window_make_fullscreen, meta_window_shade,
meta_window_unshade, meta_window_move_resize_internal,
window_stick_impl, window_unstick_impl,
meta_window_client_message, process_property_notify): Modified to
use new queueing functions.
* src/window.c (idle_move_resize, idle_update_icon,
idle_calc_showing): update to receive queue number from pointer.
* src/window.h (MetaQueueType): new enum.
* src/window.h (MetaWindow): *_queued replaced with is_in_queue
bitfield.
* src/core.c (meta_core_queue_frame_resize):
* src/display.c (event_callback,
meta_display_queue_retheme_all_windows): Using new queueing functions.
* src/frame.c (meta_window_destroy_frame): Using new queueing functions.
* src/screen.c (queue_resize, meta_screen_resize_func,
queue_windows_showing): Using new queueing functions.
* src/window-props.c (reload_mwm_hints, reload_wm_hints,
reload_transient_for): Using new queueing functions.
* src/workspace.c (meta_workspace_add_window,
meta_workspace_remove_window, meta_workspace_queue_calc_showing,
meta_workspace_invalidate_work_area): Using new queueing functions.
svn path=/trunk/; revision=3236
2007-06-10 21:15:33 -04:00
meta_window_unqueue ( window , META_QUEUE_MOVE_RESIZE ) ;
2003-02-23 12:09:46 -05:00
2007-04-08 23:43:55 -04:00
meta_window_get_client_root_coords ( window , & old_rect ) ;
2008-12-18 07:41:56 -05:00
2003-03-08 18:43:17 -05:00
meta_topic ( META_DEBUG_GEOMETRY ,
" Move/resize %s to %d,%d %dx%d%s%s from %d,%d %dx%d \n " ,
window - > desc , root_x_nw , root_y_nw , w , h ,
is_configure_request ? " (configure request) " : " " ,
is_user_action ? " (user move/resize) " : " " ,
old_rect . x , old_rect . y , old_rect . width , old_rect . height ) ;
2008-12-18 07:41:56 -05:00
2014-03-19 10:30:12 -04:00
meta_frame_calc_borders ( window - > frame , & borders ) ;
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
2014-03-19 10:30:12 -04:00
requested_rect . x = root_x_nw ;
requested_rect . y = root_y_nw ;
requested_rect . width = w ;
requested_rect . height = h ;
new_rect = requested_rect ;
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
/* If this is a resize only, the position should be ignored and
* instead obtained by resizing the old rectangle according to the
* relevant gravity .
*/
2008-12-18 07:41:56 -05:00
if ( ( flags & ( META_IS_MOVE_ACTION | META_IS_RESIZE_ACTION ) ) = =
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
META_IS_RESIZE_ACTION )
2008-12-18 07:41:56 -05:00
{
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
meta_rectangle_resize_with_gravity ( & old_rect ,
& new_rect ,
2007-04-09 19:00:55 -04:00
gravity ,
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
new_rect . width ,
new_rect . height ) ;
meta_topic ( META_DEBUG_GEOMETRY ,
" Compensated for gravity in resize action; new pos %d,%d \n " ,
new_rect . x , new_rect . y ) ;
}
else if ( is_configure_request | | do_gravity_adjust )
2008-12-18 07:41:56 -05:00
{
2003-03-08 18:43:17 -05:00
adjust_for_gravity ( window ,
2011-07-12 00:37:41 -04:00
window - > frame ? & borders : NULL ,
2003-03-08 18:43:17 -05:00
/* configure request coords assume
* the border width existed
*/
is_configure_request ,
2007-04-09 19:00:55 -04:00
gravity ,
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
& new_rect ) ;
2003-03-08 18:43:17 -05:00
meta_topic ( META_DEBUG_GEOMETRY ,
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
" Compensated for configure_request/do_gravity_adjust needing "
" weird positioning; new pos %d,%d \n " ,
new_rect . x , new_rect . y ) ;
2003-03-08 18:43:17 -05:00
}
2001-06-09 23:17:15 -04:00
2013-02-15 21:27:00 -05:00
did_placement = ! window - > placed & & window - > calc_placement ;
2013-09-03 06:00:29 -04:00
if ( flags & ( META_IS_MOVE_ACTION | META_IS_RESIZE_ACTION ) )
{
meta_window_constrain ( window ,
flags ,
gravity ,
& old_rect ,
& new_rect ) ;
}
2003-06-25 23:09:38 -04:00
2014-03-19 10:30:12 -04:00
/* Do the protocol-specific move/resize logic */
META_WINDOW_GET_CLASS ( window ) - > move_resize_internal ( window , gravity , requested_rect , new_rect , flags , & result ) ;
2001-06-09 23:17:15 -04:00
2010-01-20 10:59:07 -05:00
if ( ! window - > placed & & window - > force_save_user_rect & & ! window - > fullscreen )
2008-08-15 23:09:40 -04:00
force_save_user_window_placement ( window ) ;
else if ( is_user_action )
save_user_window_placement ( window ) ;
2014-03-19 10:30:12 -04:00
if ( result & META_MOVE_RESIZE_RESULT_MOVED )
2013-12-06 15:39:24 -05:00
g_signal_emit ( window , window_signals [ POSITION_CHANGED ] , 0 ) ;
2014-03-19 10:30:12 -04:00
if ( result & META_MOVE_RESIZE_RESULT_RESIZED )
2013-12-06 15:39:24 -05:00
g_signal_emit ( window , window_signals [ SIZE_CHANGED ] , 0 ) ;
2014-03-19 10:30:12 -04:00
if ( ( result & ( META_MOVE_RESIZE_RESULT_MOVED | META_MOVE_RESIZE_RESULT_RESIZED ) ) ! = 0 | |
2013-09-03 06:00:29 -04:00
did_placement | | is_wayland_resize )
2001-09-10 22:57:05 -04:00
{
int newx , newy ;
meta_window_get_position ( window , & newx , & newy ) ;
2002-02-07 22:34:26 -05:00
meta_topic ( META_DEBUG_GEOMETRY ,
" New size/position %d,%d %dx%d (user %d,%d %dx%d) \n " ,
newx , newy , window - > rect . width , window - > rect . height ,
window - > user_rect . x , window - > user_rect . y ,
window - > user_rect . width , window - > user_rect . height ) ;
2014-03-18 18:55:32 -04:00
if ( window - > visible_to_compositor )
meta_compositor_sync_window_geometry ( window - > display - > compositor ,
window ,
did_placement ) ;
2001-09-10 22:57:05 -04:00
}
else
{
2002-02-07 22:34:26 -05:00
meta_topic ( META_DEBUG_GEOMETRY , " Size/position not modified \n " ) ;
2001-09-10 22:57:05 -04:00
}
2008-12-18 07:41:56 -05:00
2009-05-07 13:53:47 -04:00
meta_window_refresh_resize_popup ( window ) ;
2008-12-18 07:41:56 -05:00
2011-02-25 10:37:05 -05:00
meta_window_update_monitor ( window ) ;
2001-06-09 23:17:15 -04:00
/* Invariants leaving this function are:
* a ) window - > rect and frame - > rect reflect the actual
* server - side size / pos of window - > xwindow and frame - > xwindow
* b ) all constraints are obeyed by window - > rect and frame - > rect
*/
2010-09-10 05:55:20 -04:00
2011-03-22 15:36:12 -04:00
if ( frame_shape_changed & & window - > frame_bounds )
{
cairo_region_destroy ( window - > frame_bounds ) ;
window - > frame_bounds = NULL ;
}
window: make determination of attached dialog windows more consistent
Different bits of code were using slightly different checks to test
whether a window was an attached dialog. Add a new
meta_window_is_attached_dialog(), and use that everywhere.
Also, freeze the is-attached status when the window is first shown,
rather than recomputing it each time the caller asks, since this could
cause problems if a window changes its type after it has already been
attached, etc. However, if an attached window's parent is destroyed,
or an attached window changes its transient-for, then fix things up by
destroying the old MetaWindow and creating a new one (causing
compositor unmap and map events to be fired off, allowing the display
of the window to be fixed up).
Remove some code in display.c that tried to fix existing windows if
the gconf setting changed, but which didn't actually do anything (at
least under gnome-shell). However, if 654643 was fixed then the new
behavior with this patch would be that changing the gconf setting
would affect new dialogs, but not existing ones.
https://bugzilla.gnome.org/show_bug.cgi?id=646761
2011-04-05 10:14:52 -04:00
meta_window_foreach_transient ( window , maybe_move_attached_dialog , NULL ) ;
2012-02-24 11:08:55 -05:00
meta_stack_update_window_tile_matches ( window - > screen - > stack ,
window - > screen - > active_workspace ) ;
2001-06-07 22:17:48 -04:00
}
2011-03-09 10:01:44 -05:00
/**
* meta_window_resize :
* @ window : a # MetaWindow
* @ user_op : bool to indicate whether or not this is a user operation
* @ w : desired width
* @ h : desired height
*
* Resize the window to the desired size .
*/
2001-06-07 22:17:48 -04:00
void
meta_window_resize ( MetaWindow * window ,
2001-08-28 23:37:03 -04:00
gboolean user_op ,
2001-06-07 22:17:48 -04:00
int w ,
int h )
{
2001-06-09 23:17:15 -04:00
int x , y ;
Patch from Jens Granseuer to fix build with gcc 2.95. #322622.
2006-01-02 Elijah Newren <newren@gmail.com>
Patch from Jens Granseuer to fix build with gcc 2.95. #322622.
* src/boxes.c (meta_rectangle_region_to_string,
meta_rectangle_edge_list_to_string, fix_up_edges):
* src/constraints.c (meta_window_constrain, setup_constraint_info,
place_window_if_needed, constrain_maximization,
constrain_fullscreen, constrain_size_increments,
constrain_size_limits, constrain_aspect_ratio,
do_screen_and_xinerama_relative_constrai,
constrain_to_single_xinerama, constrain_fully_onscreen,
constrain_partially_onscreen):
* src/edge-resistance.c (find_nearest_position,
apply_edge_resistance, apply_edge_resistance_to_each_side):
* src/testboxes.c (test_clamping_to_region,
test_clipping_to_region, test_shoving_into_region):
* src/window.c (meta_window_new_with_attrs,
meta_window_apply_session_info, meta_window_resize,
meta_window_resize_with_gravity, meta_window_configure_request):
Remove C99 style variable initiailization
2006-01-02 13:37:46 -05:00
MetaMoveResizeFlags flags ;
2001-06-09 23:17:15 -04:00
2009-06-15 18:25:07 -04:00
g_return_if_fail ( ! window - > override_redirect ) ;
2001-06-09 23:17:15 -04:00
meta_window_get_position ( window , & x , & y ) ;
2008-12-18 07:41:56 -05:00
Patch from Jens Granseuer to fix build with gcc 2.95. #322622.
2006-01-02 Elijah Newren <newren@gmail.com>
Patch from Jens Granseuer to fix build with gcc 2.95. #322622.
* src/boxes.c (meta_rectangle_region_to_string,
meta_rectangle_edge_list_to_string, fix_up_edges):
* src/constraints.c (meta_window_constrain, setup_constraint_info,
place_window_if_needed, constrain_maximization,
constrain_fullscreen, constrain_size_increments,
constrain_size_limits, constrain_aspect_ratio,
do_screen_and_xinerama_relative_constrai,
constrain_to_single_xinerama, constrain_fully_onscreen,
constrain_partially_onscreen):
* src/edge-resistance.c (find_nearest_position,
apply_edge_resistance, apply_edge_resistance_to_each_side):
* src/testboxes.c (test_clamping_to_region,
test_clipping_to_region, test_shoving_into_region):
* src/window.c (meta_window_new_with_attrs,
meta_window_apply_session_info, meta_window_resize,
meta_window_resize_with_gravity, meta_window_configure_request):
Remove C99 style variable initiailization
2006-01-02 13:37:46 -05:00
flags = ( user_op ? META_IS_USER_ACTION : 0 ) | META_IS_RESIZE_ACTION ;
2001-08-28 23:37:03 -04:00
meta_window_move_resize_internal ( window ,
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
flags ,
2001-06-30 19:17:52 -04:00
NorthWestGravity ,
x , y , w , h ) ;
2001-06-07 22:17:48 -04:00
}
2013-09-03 06:00:29 -04:00
/*
* meta_window_move_resize_wayland :
*
* Complete a resize operation from a wayland client .
*
*/
void
meta_window_move_resize_wayland ( MetaWindow * window ,
int width ,
int height ,
int dx ,
int dy )
{
int x , y ;
MetaMoveResizeFlags flags ;
flags = META_IS_WAYLAND_RESIZE ;
meta_window_get_position ( window , & x , & y ) ;
x + = dx ; y + = dy ;
if ( x ! = window - > expected_rect . x | | y ! = window - > expected_rect . y )
flags | = META_IS_MOVE_ACTION ;
if ( width ! = window - > expected_rect . width | |
height ! = window - > expected_rect . height )
flags | = META_IS_RESIZE_ACTION ;
meta_window_move_resize_internal ( window , flags , NorthWestGravity ,
x , y , width , height ) ;
2014-03-11 12:27:44 -04:00
save_user_window_placement ( window ) ;
2013-09-03 06:00:29 -04:00
}
2011-03-09 10:01:44 -05:00
/**
* meta_window_move :
* @ window : a # MetaWindow
* @ user_op : bool to indicate whether or not this is a user operation
* @ root_x_nw : desired x pos
* @ root_y_nw : desired y pos
*
* Moves the window to the desired location on window ' s assigned workspace .
* NOTE : does NOT place according to the origin of the enclosing
* frame / window - decoration , but according to the origin of the window ,
* itself .
*/
2001-06-07 22:17:48 -04:00
void
meta_window_move ( MetaWindow * window ,
2001-08-28 23:37:03 -04:00
gboolean user_op ,
2001-06-07 22:17:48 -04:00
int root_x_nw ,
int root_y_nw )
{
2009-06-15 18:25:07 -04:00
MetaMoveResizeFlags flags ;
g_return_if_fail ( ! window - > override_redirect ) ;
flags = ( user_op ? META_IS_USER_ACTION : 0 ) | META_IS_MOVE_ACTION ;
2001-08-28 23:37:03 -04:00
meta_window_move_resize_internal ( window ,
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
flags ,
2001-06-30 19:17:52 -04:00
NorthWestGravity ,
2001-06-09 23:17:15 -04:00
root_x_nw , root_y_nw ,
window - > rect . width ,
window - > rect . height ) ;
2001-06-07 22:17:48 -04:00
}
2011-03-09 10:01:44 -05:00
/**
* meta_window_move_frame :
* @ window : a # MetaWindow
* @ user_op : bool to indicate whether or not this is a user operation
* @ root_x_nw : desired x pos
* @ root_y_nw : desired y pos
*
* Moves the window to the desired location on window ' s assigned
* workspace , using the northwest edge of the frame as the reference ,
* instead of the actual window ' s origin , but only if a frame is present .
* Otherwise , acts identically to meta_window_move ( ) .
*/
void
meta_window_move_frame ( MetaWindow * window ,
gboolean user_op ,
int root_x_nw ,
int root_y_nw )
{
2013-11-15 16:11:20 -05:00
MetaRectangle rect = { root_x_nw , root_y_nw , 0 , 0 } ;
2012-03-12 23:11:07 -04:00
2013-11-15 16:11:20 -05:00
meta_window_frame_rect_to_client_rect ( window , & rect , & rect ) ;
meta_window_move ( window , user_op , rect . x , rect . y ) ;
2011-03-09 10:01:44 -05:00
}
2001-06-07 22:17:48 -04:00
2011-05-20 04:56:12 -04:00
static void
meta_window_move_between_rects ( MetaWindow * window ,
const MetaRectangle * old_area ,
const MetaRectangle * new_area )
{
int rel_x , rel_y ;
double scale_x , scale_y ;
rel_x = window - > user_rect . x - old_area - > x ;
rel_y = window - > user_rect . y - old_area - > y ;
scale_x = ( double ) new_area - > width / old_area - > width ;
scale_y = ( double ) new_area - > height / old_area - > height ;
window - > user_rect . x = new_area - > x + rel_x * scale_x ;
window - > user_rect . y = new_area - > y + rel_y * scale_y ;
window - > saved_rect . x = window - > user_rect . x ;
window - > saved_rect . y = window - > user_rect . y ;
meta_window_move_resize ( window , FALSE ,
window - > user_rect . x ,
window - > user_rect . y ,
window - > user_rect . width ,
window - > user_rect . height ) ;
}
2011-06-04 21:02:44 -04:00
/**
* meta_window_move_resize_frame :
* @ window : a # MetaWindow
* @ user_op : bool to indicate whether or not this is a user operation
* @ root_x_nw : new x
* @ root_y_nw : new y
* @ w : desired width
* @ h : desired height
*
* Resizes the window so that its outer bounds ( including frame )
* fit within the given rect
*/
void
meta_window_move_resize_frame ( MetaWindow * window ,
gboolean user_op ,
int root_x_nw ,
int root_y_nw ,
int w ,
int h )
{
2013-11-15 16:11:20 -05:00
MetaRectangle rect = { root_x_nw , root_y_nw , w , h } ;
meta_window_frame_rect_to_client_rect ( window , & rect , & rect ) ;
2012-03-12 23:11:07 -04:00
2013-11-15 16:11:20 -05:00
meta_window_move_resize ( window , user_op , rect . x , rect . y , rect . width , rect . height ) ;
2011-06-04 21:02:44 -04:00
}
2011-03-21 07:36:01 -04:00
/**
* meta_window_move_to_monitor :
* @ window : a # MetaWindow
* @ monitor : desired monitor index
*
* Moves the window to the monitor with index @ monitor , keeping
* the relative position of the window ' s top left corner .
*/
void
meta_window_move_to_monitor ( MetaWindow * window ,
int monitor )
{
MetaRectangle old_area , new_area ;
if ( monitor = = window - > monitor - > number )
return ;
meta_window_get_work_area_for_monitor ( window ,
window - > monitor - > number ,
& old_area ) ;
meta_window_get_work_area_for_monitor ( window ,
monitor ,
& new_area ) ;
2011-10-08 11:00:16 -04:00
if ( window - > tile_mode ! = META_TILE_NONE )
window - > tile_monitor_number = monitor ;
2011-05-20 04:56:12 -04:00
meta_window_move_between_rects ( window , & old_area , & new_area ) ;
2011-03-21 07:36:01 -04:00
}
2001-06-07 22:17:48 -04:00
void
meta_window_move_resize ( MetaWindow * window ,
2001-08-28 23:37:03 -04:00
gboolean user_op ,
2001-06-07 22:17:48 -04:00
int root_x_nw ,
int root_y_nw ,
int w ,
int h )
{
2009-06-15 18:25:07 -04:00
MetaMoveResizeFlags flags ;
g_return_if_fail ( ! window - > override_redirect ) ;
flags = ( user_op ? META_IS_USER_ACTION : 0 ) |
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
META_IS_MOVE_ACTION | META_IS_RESIZE_ACTION ;
2001-08-28 23:37:03 -04:00
meta_window_move_resize_internal ( window ,
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
flags ,
2001-06-30 19:17:52 -04:00
NorthWestGravity ,
2001-06-07 22:17:48 -04:00
root_x_nw , root_y_nw ,
w , h ) ;
2001-06-06 00:47:37 -04:00
}
2001-06-30 19:17:52 -04:00
void
meta_window_resize_with_gravity ( MetaWindow * window ,
2001-08-28 23:37:03 -04:00
gboolean user_op ,
2001-06-30 19:17:52 -04:00
int w ,
int h ,
int gravity )
{
int x , y ;
Patch from Jens Granseuer to fix build with gcc 2.95. #322622.
2006-01-02 Elijah Newren <newren@gmail.com>
Patch from Jens Granseuer to fix build with gcc 2.95. #322622.
* src/boxes.c (meta_rectangle_region_to_string,
meta_rectangle_edge_list_to_string, fix_up_edges):
* src/constraints.c (meta_window_constrain, setup_constraint_info,
place_window_if_needed, constrain_maximization,
constrain_fullscreen, constrain_size_increments,
constrain_size_limits, constrain_aspect_ratio,
do_screen_and_xinerama_relative_constrai,
constrain_to_single_xinerama, constrain_fully_onscreen,
constrain_partially_onscreen):
* src/edge-resistance.c (find_nearest_position,
apply_edge_resistance, apply_edge_resistance_to_each_side):
* src/testboxes.c (test_clamping_to_region,
test_clipping_to_region, test_shoving_into_region):
* src/window.c (meta_window_new_with_attrs,
meta_window_apply_session_info, meta_window_resize,
meta_window_resize_with_gravity, meta_window_configure_request):
Remove C99 style variable initiailization
2006-01-02 13:37:46 -05:00
MetaMoveResizeFlags flags ;
2001-06-30 19:17:52 -04:00
meta_window_get_position ( window , & x , & y ) ;
2008-12-18 07:41:56 -05:00
Patch from Jens Granseuer to fix build with gcc 2.95. #322622.
2006-01-02 Elijah Newren <newren@gmail.com>
Patch from Jens Granseuer to fix build with gcc 2.95. #322622.
* src/boxes.c (meta_rectangle_region_to_string,
meta_rectangle_edge_list_to_string, fix_up_edges):
* src/constraints.c (meta_window_constrain, setup_constraint_info,
place_window_if_needed, constrain_maximization,
constrain_fullscreen, constrain_size_increments,
constrain_size_limits, constrain_aspect_ratio,
do_screen_and_xinerama_relative_constrai,
constrain_to_single_xinerama, constrain_fully_onscreen,
constrain_partially_onscreen):
* src/edge-resistance.c (find_nearest_position,
apply_edge_resistance, apply_edge_resistance_to_each_side):
* src/testboxes.c (test_clamping_to_region,
test_clipping_to_region, test_shoving_into_region):
* src/window.c (meta_window_new_with_attrs,
meta_window_apply_session_info, meta_window_resize,
meta_window_resize_with_gravity, meta_window_configure_request):
Remove C99 style variable initiailization
2006-01-02 13:37:46 -05:00
flags = ( user_op ? META_IS_USER_ACTION : 0 ) | META_IS_RESIZE_ACTION ;
2001-08-28 23:37:03 -04:00
meta_window_move_resize_internal ( window ,
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
flags ,
2001-06-30 19:17:52 -04:00
gravity ,
x , y , w , h ) ;
}
2006-01-20 17:03:56 -05:00
static void
2001-06-24 04:09:10 -04:00
meta_window_move_resize_now ( MetaWindow * window )
2001-06-09 17:58:30 -04:00
{
2007-04-08 23:43:55 -04:00
/* If constraints have changed then we want to snap back to wherever
* the user had the window . We use user_rect for this reason . See
* also bug 426519 comment 3.
2002-04-21 23:10:53 -04:00
*/
2007-04-08 23:43:55 -04:00
meta_window_move_resize ( window , FALSE ,
window - > user_rect . x ,
window - > user_rect . y ,
window - > user_rect . width ,
window - > user_rect . height ) ;
2001-06-09 17:58:30 -04:00
}
2002-01-18 22:50:03 -05:00
static gboolean
idle_move_resize ( gpointer data )
{
GSList * tmp ;
GSList * copy ;
Refactor thrice-duplicated queue code in window.c. Closes #376760.
2007-06-10 Thomas Thurman <thomas@thurman.org.uk>
Refactor thrice-duplicated queue code in window.c. Closes #376760.
* src/window.c (meta_window_queue, meta_window_unqueue):
New functions.
* src/window.[ch] (meta_window_unqueue_*, meta_window_queue_*):
Removed functions.
* src/window.c (meta_window_new_with_attrs, meta_window_free,
meta_window_flush_calc_showing, queue_calc_showing_func,
meta_window_minimize, meta_window_unminimize, meta_window_maximize,
meta_window_make_fullscreen, meta_window_shade,
meta_window_unshade, meta_window_move_resize_internal,
window_stick_impl, window_unstick_impl,
meta_window_client_message, process_property_notify): Modified to
use new queueing functions.
* src/window.c (idle_move_resize, idle_update_icon,
idle_calc_showing): update to receive queue number from pointer.
* src/window.h (MetaQueueType): new enum.
* src/window.h (MetaWindow): *_queued replaced with is_in_queue
bitfield.
* src/core.c (meta_core_queue_frame_resize):
* src/display.c (event_callback,
meta_display_queue_retheme_all_windows): Using new queueing functions.
* src/frame.c (meta_window_destroy_frame): Using new queueing functions.
* src/screen.c (queue_resize, meta_screen_resize_func,
queue_windows_showing): Using new queueing functions.
* src/window-props.c (reload_mwm_hints, reload_wm_hints,
reload_transient_for): Using new queueing functions.
* src/workspace.c (meta_workspace_add_window,
meta_workspace_remove_window, meta_workspace_queue_calc_showing,
meta_workspace_invalidate_work_area): Using new queueing functions.
svn path=/trunk/; revision=3236
2007-06-10 21:15:33 -04:00
guint queue_index = GPOINTER_TO_INT ( data ) ;
2002-01-18 22:50:03 -05:00
2002-02-07 22:34:26 -05:00
meta_topic ( META_DEBUG_GEOMETRY , " Clearing the move_resize queue \n " ) ;
2002-01-18 22:50:03 -05:00
/* Work with a copy, for reentrancy. The allowed reentrancy isn't
* complete ; destroying a window while we ' re in here would result in
* badness . But it ' s OK to queue / unqueue move_resizes .
*/
Refactor thrice-duplicated queue code in window.c. Closes #376760.
2007-06-10 Thomas Thurman <thomas@thurman.org.uk>
Refactor thrice-duplicated queue code in window.c. Closes #376760.
* src/window.c (meta_window_queue, meta_window_unqueue):
New functions.
* src/window.[ch] (meta_window_unqueue_*, meta_window_queue_*):
Removed functions.
* src/window.c (meta_window_new_with_attrs, meta_window_free,
meta_window_flush_calc_showing, queue_calc_showing_func,
meta_window_minimize, meta_window_unminimize, meta_window_maximize,
meta_window_make_fullscreen, meta_window_shade,
meta_window_unshade, meta_window_move_resize_internal,
window_stick_impl, window_unstick_impl,
meta_window_client_message, process_property_notify): Modified to
use new queueing functions.
* src/window.c (idle_move_resize, idle_update_icon,
idle_calc_showing): update to receive queue number from pointer.
* src/window.h (MetaQueueType): new enum.
* src/window.h (MetaWindow): *_queued replaced with is_in_queue
bitfield.
* src/core.c (meta_core_queue_frame_resize):
* src/display.c (event_callback,
meta_display_queue_retheme_all_windows): Using new queueing functions.
* src/frame.c (meta_window_destroy_frame): Using new queueing functions.
* src/screen.c (queue_resize, meta_screen_resize_func,
queue_windows_showing): Using new queueing functions.
* src/window-props.c (reload_mwm_hints, reload_wm_hints,
reload_transient_for): Using new queueing functions.
* src/workspace.c (meta_workspace_add_window,
meta_workspace_remove_window, meta_workspace_queue_calc_showing,
meta_workspace_invalidate_work_area): Using new queueing functions.
svn path=/trunk/; revision=3236
2007-06-10 21:15:33 -04:00
copy = g_slist_copy ( queue_pending [ queue_index ] ) ;
g_slist_free ( queue_pending [ queue_index ] ) ;
queue_pending [ queue_index ] = NULL ;
2009-09-25 15:47:22 -04:00
queue_later [ queue_index ] = 0 ;
2002-01-18 22:50:03 -05:00
destroying_windows_disallowed + = 1 ;
2008-12-18 07:41:56 -05:00
2002-01-18 22:50:03 -05:00
tmp = copy ;
while ( tmp ! = NULL )
{
MetaWindow * window ;
window = tmp - > data ;
/* As a side effect, sets window->move_resize_queued = FALSE */
2008-12-18 07:41:56 -05:00
meta_window_move_resize_now ( window ) ;
2002-01-18 22:50:03 -05:00
tmp = tmp - > next ;
}
g_slist_free ( copy ) ;
destroying_windows_disallowed - = 1 ;
2008-12-18 07:41:56 -05:00
2002-01-18 22:50:03 -05:00
return FALSE ;
}
2001-06-09 17:58:30 -04:00
void
meta_window_get_position ( MetaWindow * window ,
int * x ,
int * y )
{
if ( window - > frame )
{
2001-07-12 01:53:56 -04:00
if ( x )
* x = window - > frame - > rect . x + window - > frame - > child_x ;
if ( y )
* y = window - > frame - > rect . y + window - > frame - > child_y ;
2001-06-09 17:58:30 -04:00
}
else
{
2001-07-12 01:53:56 -04:00
if ( x )
* x = window - > rect . x ;
if ( y )
* y = window - > rect . y ;
2001-06-09 17:58:30 -04:00
}
}
2007-04-08 23:43:55 -04:00
void
meta_window_get_client_root_coords ( MetaWindow * window ,
MetaRectangle * rect )
{
meta_window_get_position ( window , & rect - > x , & rect - > y ) ;
rect - > width = window - > rect . width ;
rect - > height = window - > rect . height ;
}
2001-07-07 14:36:50 -04:00
void
meta_window_get_gravity_position ( MetaWindow * window ,
2007-04-09 19:00:55 -04:00
int gravity ,
2001-07-07 14:36:50 -04:00
int * root_x ,
int * root_y )
{
MetaRectangle frame_extents ;
int w , h ;
int x , y ;
2008-12-18 07:41:56 -05:00
2001-07-07 14:36:50 -04:00
w = window - > rect . width ;
h = window - > rect . height ;
2007-04-09 19:00:55 -04:00
if ( gravity = = StaticGravity )
2001-08-18 22:29:38 -04:00
{
frame_extents = window - > rect ;
if ( window - > frame )
{
frame_extents . x = window - > frame - > rect . x + window - > frame - > child_x ;
frame_extents . y = window - > frame - > rect . y + window - > frame - > child_y ;
}
}
2001-07-07 14:36:50 -04:00
else
2001-08-18 22:29:38 -04:00
{
if ( window - > frame = = NULL )
frame_extents = window - > rect ;
else
frame_extents = window - > frame - > rect ;
}
2001-07-07 14:36:50 -04:00
x = frame_extents . x ;
y = frame_extents . y ;
2008-12-18 07:41:56 -05:00
2007-04-09 19:00:55 -04:00
switch ( gravity )
2001-07-07 14:36:50 -04:00
{
case NorthGravity :
case CenterGravity :
case SouthGravity :
/* Find center of frame. */
x + = frame_extents . width / 2 ;
/* Center client window on that point. */
x - = w / 2 ;
break ;
2008-12-18 07:41:56 -05:00
2001-07-07 14:36:50 -04:00
case SouthEastGravity :
case EastGravity :
case NorthEastGravity :
/* Find right edge of frame */
x + = frame_extents . width ;
/* Align left edge of client at that point. */
x - = w ;
break ;
default :
break ;
}
2008-12-18 07:41:56 -05:00
2007-04-09 19:00:55 -04:00
switch ( gravity )
2001-07-07 14:36:50 -04:00
{
case WestGravity :
case CenterGravity :
case EastGravity :
/* Find center of frame. */
y + = frame_extents . height / 2 ;
/* Center client window there. */
y - = h / 2 ;
break ;
case SouthWestGravity :
case SouthGravity :
case SouthEastGravity :
/* Find south edge of frame */
y + = frame_extents . height ;
/* Place bottom edge of client there */
y - = h ;
break ;
default :
break ;
}
2008-12-18 07:41:56 -05:00
2001-07-07 14:36:50 -04:00
if ( root_x )
* root_x = x ;
if ( root_y )
* root_y = y ;
}
2001-07-28 02:35:19 -04:00
void
meta_window_get_geometry ( MetaWindow * window ,
int * x ,
int * y ,
int * width ,
int * height )
{
2007-04-09 19:00:55 -04:00
meta_window_get_gravity_position ( window ,
window - > size_hints . win_gravity ,
x , y ) ;
2001-07-28 02:35:19 -04:00
* width = ( window - > rect . width - window - > size_hints . base_width ) /
window - > size_hints . width_inc ;
* height = ( window - > rect . height - window - > size_hints . base_height ) /
window - > size_hints . height_inc ;
}
2011-06-29 13:37:04 -04:00
/**
* meta_window_get_input_rect :
* @ window : a # MetaWindow
* @ rect : ( out ) : pointer to an allocated # MetaRectangle
*
* Gets the rectangle that bounds @ window that is responsive to mouse events .
* This includes decorations - the visible portion of its border - and ( if
* present ) any invisible area that we make make responsive to mouse clicks in
* order to allow convenient border dragging .
*/
void
meta_window_get_input_rect ( const MetaWindow * window ,
MetaRectangle * rect )
{
if ( window - > frame )
* rect = window - > frame - > rect ;
else
* rect = window - > rect ;
}
2013-11-15 16:11:20 -05:00
/**
* meta_window_client_rect_to_frame_rect :
* @ window : a # MetaWindow
2013-11-29 07:52:18 -05:00
* @ client_rect : client rectangle in root coordinates
* @ frame_rect : ( out ) : location to store the computed corresponding frame bounds .
2013-11-15 16:11:20 -05:00
*
* Converts a desired bounds of the client window - what is passed to meta_window_move_resize ( ) -
* into the corresponding bounds of the window frame ( excluding invisible borders
* and client side shadows . )
*/
void
meta_window_client_rect_to_frame_rect ( MetaWindow * window ,
2013-11-29 07:52:18 -05:00
MetaRectangle * client_rect ,
MetaRectangle * frame_rect )
2013-11-15 16:11:20 -05:00
{
2013-11-29 07:52:18 -05:00
if ( ! frame_rect )
2013-11-15 16:11:20 -05:00
return ;
2013-11-29 07:52:18 -05:00
* frame_rect = * client_rect ;
2013-11-15 16:11:20 -05:00
2013-11-15 18:03:50 -05:00
/* The support for G_MAXINT here to mean infinity is a convenience for
* constraints . c : get_size_limits ( ) and not something that we provide
* in other locations or document .
*/
2013-11-15 16:11:20 -05:00
if ( window - > frame )
{
MetaFrameBorders borders ;
meta_frame_calc_borders ( window - > frame , & borders ) ;
2013-11-29 07:52:18 -05:00
frame_rect - > x - = borders . visible . left ;
frame_rect - > y - = borders . visible . top ;
if ( frame_rect - > width ! = G_MAXINT )
frame_rect - > width + = borders . visible . left + borders . visible . right ;
if ( frame_rect - > height ! = G_MAXINT )
frame_rect - > height + = borders . visible . top + borders . visible . bottom ;
2013-11-15 16:11:20 -05:00
}
else
{
if ( window - > has_custom_frame_extents )
{
const GtkBorder * extents = & window - > custom_frame_extents ;
2013-11-29 07:52:18 -05:00
frame_rect - > x + = extents - > left ;
frame_rect - > y + = extents - > top ;
if ( frame_rect - > width ! = G_MAXINT )
frame_rect - > width - = extents - > left + extents - > right ;
if ( frame_rect - > height ! = G_MAXINT )
frame_rect - > height - = extents - > top + extents - > bottom ;
2013-11-15 16:11:20 -05:00
}
}
}
/**
* meta_window_frame_rect_to_client_rect :
* @ window : a # MetaWindow
* @ frame_rect : desired frame bounds for the window
* @ client_rect : ( out ) : location to store the computed corresponding client rectangle .
*
* Converts a desired frame bounds for a window into the bounds of the client
* window - what is passed to meta_window_move_resize ( ) .
*/
void
meta_window_frame_rect_to_client_rect ( MetaWindow * window ,
MetaRectangle * frame_rect ,
MetaRectangle * client_rect )
{
if ( ! client_rect )
return ;
* client_rect = * frame_rect ;
if ( window - > frame )
{
MetaFrameBorders borders ;
meta_frame_calc_borders ( window - > frame , & borders ) ;
client_rect - > x + = borders . visible . left ;
client_rect - > y + = borders . visible . top ;
client_rect - > width - = borders . visible . left + borders . visible . right ;
client_rect - > height - = borders . visible . top + borders . visible . bottom ;
}
else
{
if ( window - > has_custom_frame_extents )
{
const GtkBorder * extents = & window - > custom_frame_extents ;
client_rect - > x - = extents - > left ;
client_rect - > y - = extents - > top ;
client_rect - > width + = extents - > left + extents - > right ;
client_rect - > height + = extents - > top + extents - > bottom ;
}
}
}
2010-11-22 16:26:37 -05:00
/**
2013-11-19 12:24:45 -05:00
* meta_window_get_frame_rect :
2010-11-22 16:26:37 -05:00
* @ window : a # MetaWindow
* @ rect : ( out ) : pointer to an allocated # MetaRectangle
*
2013-11-19 12:24:45 -05:00
* Gets the rectangle that bounds @ window that is what the user thinks of
* as the edge of the window . This doesn ' t include any extra reactive
* area that we or the client adds to the window , or any area that the
* client adds to draw a client - side shadow .
2010-11-22 16:26:37 -05:00
*/
2001-07-12 01:53:56 -04:00
void
2013-11-19 12:24:45 -05:00
meta_window_get_frame_rect ( const MetaWindow * window ,
2006-01-10 00:14:48 -05:00
MetaRectangle * rect )
2001-07-12 01:53:56 -04:00
{
if ( window - > frame )
2011-06-29 13:37:04 -04:00
{
MetaFrameBorders borders ;
* rect = window - > frame - > rect ;
meta_frame_calc_borders ( window - > frame , & borders ) ;
rect - > x + = borders . invisible . left ;
rect - > y + = borders . invisible . top ;
rect - > width - = borders . invisible . left + borders . invisible . right ;
rect - > height - = borders . invisible . top + borders . invisible . bottom ;
}
2001-07-12 01:53:56 -04:00
else
2013-08-08 16:58:20 -04:00
{
* rect = window - > rect ;
if ( window - > has_custom_frame_extents )
{
2013-08-20 18:32:23 -04:00
const GtkBorder * extents = & window - > custom_frame_extents ;
2013-08-08 16:58:20 -04:00
rect - > x + = extents - > left ;
rect - > y + = extents - > top ;
rect - > width - = extents - > left + extents - > right ;
rect - > height - = extents - > top + extents - > bottom ;
}
}
2001-07-12 01:53:56 -04:00
}
2013-11-19 12:24:45 -05:00
/**
* meta_window_get_outer_rect :
* @ window : a # MetaWindow
* @ rect : ( out ) : pointer to an allocated # MetaRectangle
*
* Gets the rectangle that bounds @ window that is what the user thinks of
* as the edge of the window . This doesn ' t include any extra reactive
* area that we or the client adds to the window , or any area that the
* client adds to draw a client - side shadow .
*
2013-11-29 03:12:16 -05:00
* Deprecated : 3.12 : Use meta_window_get_frame_rect ( ) instead .
2013-11-19 12:24:45 -05:00
*/
void
meta_window_get_outer_rect ( const MetaWindow * window ,
MetaRectangle * rect )
{
meta_window_get_frame_rect ( window , rect ) ;
}
2014-02-26 19:33:43 -05:00
/**
* meta_window_get_client_area_rect :
* @ window : a # MetaWindow
* @ rect : ( out ) : pointer to a cairo rectangle
*
* Gets the rectangle for the boundaries of the client area , relative
* to the frame . If the window is shaded , the height of the rectangle
* is 0.
*/
void
meta_window_get_client_area_rect ( const MetaWindow * window ,
cairo_rectangle_int_t * rect )
{
if ( window - > frame )
{
rect - > x = window - > frame - > child_x ;
rect - > y = window - > frame - > child_y ;
}
else
{
rect - > x = 0 ;
rect - > y = 0 ;
}
rect - > width = window - > rect . width ;
if ( window - > shaded )
rect - > height = 0 ;
2014-02-27 10:32:19 -05:00
else
rect - > height = window - > rect . height ;
2014-02-26 19:33:43 -05:00
}
2002-11-30 22:58:04 -05:00
const char *
meta_window_get_startup_id ( MetaWindow * window )
{
if ( window - > startup_id = = NULL )
{
MetaGroup * group ;
group = meta_window_get_group ( window ) ;
if ( group ! = NULL )
return meta_group_get_startup_id ( group ) ;
}
return window - > startup_id ;
}
2005-01-25 12:26:06 -05:00
static MetaWindow *
get_modal_transient ( MetaWindow * window )
{
GSList * windows ;
GSList * tmp ;
MetaWindow * modal_transient ;
/* A window can't be the transient of itself, but this is just for
* convenience in the loop below ; we manually fix things up at the
* end if no real modal transient was found .
*/
modal_transient = window ;
2009-06-15 15:32:23 -04:00
windows = meta_display_list_windows ( window - > display , META_LIST_DEFAULT ) ;
2005-01-25 12:26:06 -05:00
tmp = windows ;
while ( tmp ! = NULL )
{
MetaWindow * transient = tmp - > data ;
2008-12-18 07:41:56 -05:00
2013-09-16 08:44:37 -04:00
if ( transient - > transient_for = = modal_transient & &
2014-03-18 10:58:26 -04:00
transient - > type = = META_WINDOW_MODAL_DIALOG )
2005-01-25 12:26:06 -05:00
{
modal_transient = transient ;
tmp = windows ;
continue ;
}
2008-12-18 07:41:56 -05:00
2005-01-25 12:26:06 -05:00
tmp = tmp - > next ;
}
g_slist_free ( windows ) ;
if ( window = = modal_transient )
modal_transient = NULL ;
return modal_transient ;
}
2006-05-23 17:42:34 -04:00
/* XXX META_EFFECT_FOCUS */
2001-06-04 00:58:22 -04:00
void
meta_window_focus ( MetaWindow * window ,
Fix issues on 64-bit machines with timestamps by using guint32 (like gtk+
2006-09-13 Elijah Newren <newren gmail com>
* src/common.h (MetaWindowMenuFunc):
* src/core.[ch] (meta_core_user_lower_and_unfocus,
meta_core_user_focus, meta_core_show_window_menu,
meta_core_begin_grab_op, meta_core_end_grab_op):
* src/delete.c (delete_ping_reply_func, delete_ping_timeout_func,
meta_window_delete):
* src/display.[ch] (struct MetaDisplay, struct MetaPingData,
sanity_check_timestamps, meta_display_open, event_callback,
meta_spew_event, meta_display_set_grab_op_cursor,
meta_display_begin_grab_op, meta_display_end_grab_op,
meta_display_ping_timeout, meta_display_ping_window,
process_pong_message, timestamp_too_old,
meta_display_set_input_focus_window):
* src/keybindings.[ch] (grab_keyboard, ungrab_keyboard,
meta_screen_grab_all_keys, meta_window_grab_all_keys,
meta_window_ungrab_all_keys, error_on_generic_command,
error_on_command, error_on_terminal_command):
* src/metacity-dialog.c (on_realize, warn_about_no_sm_support,
error_about_command, main):
* src/screen.[ch] (struct _MetaScreen, meta_screen_new,
meta_screen_show_desktop, meta_screen_apply_startup_properties):
* src/session.c (warn_about_lame_clients_and_finish_interact):
* src/window.[ch] (struct _MetaWindow,
intervening_user_event_occurred, window_activate,
meta_window_delete, meta_window_focus,
meta_window_send_icccm_message, meta_window_client_message,
menu_callback, meta_window_show_menu, struct EventScannerData,
check_use_this_motion_notify, meta_window_begin_grab_op,
meta_window_set_user_time):
* src/workspace.[ch] (focus_ancestor_or_mru_window,
meta_workspace_activate_with_focus, meta_workspace_activate,
meta_workspace_focus_default_window,
focus_ancestor_or_mru_window):
Fix issues on 64-bit machines with timestamps by using guint32
(like gtk+ does) instead of Time. #348305
2006-09-13 12:32:33 -04:00
guint32 timestamp )
2008-12-18 07:41:56 -05:00
{
2005-01-25 12:26:06 -05:00
MetaWindow * modal_transient ;
2009-06-14 08:11:29 -04:00
g_return_if_fail ( ! window - > override_redirect ) ;
2002-01-03 18:28:19 -05:00
meta_topic ( META_DEBUG_FOCUS ,
" Setting input focus to window %s, input: %d take_focus: %d \n " ,
window - > desc , window - > input , window - > take_focus ) ;
2008-12-18 07:41:56 -05:00
2001-07-11 02:22:00 -04:00
if ( window - > display - > grab_window & &
2013-12-18 09:23:08 -05:00
window - > display - > grab_window - > all_keys_grabbed & &
! window - > display - > grab_window - > unmanaging )
2001-07-11 02:22:00 -04:00
{
2002-01-05 22:15:49 -05:00
meta_topic ( META_DEBUG_FOCUS ,
" Current focus window %s has global keygrab, not focusing window %s after all \n " ,
window - > display - > grab_window - > desc , window - > desc ) ;
2001-07-11 02:22:00 -04:00
return ;
}
2001-10-29 02:00:46 -05:00
2005-01-25 12:26:06 -05:00
modal_transient = get_modal_transient ( window ) ;
2005-06-02 11:24:40 -04:00
if ( modal_transient ! = NULL & &
! modal_transient - > unmanaging )
2005-01-25 12:26:06 -05:00
{
meta_topic ( META_DEBUG_FOCUS ,
" %s has %s as a modal transient, so focusing it instead. \n " ,
window - > desc , modal_transient - > desc ) ;
if ( ! modal_transient - > on_all_workspaces & &
modal_transient - > workspace ! = window - > screen - > active_workspace )
2008-12-18 07:41:56 -05:00
meta_window_change_workspace ( modal_transient ,
2005-01-25 12:26:06 -05:00
window - > screen - > active_workspace ) ;
window = modal_transient ;
}
2001-10-29 02:00:46 -05:00
meta_window_flush_calc_showing ( window ) ;
2002-05-10 22:25:11 -04:00
2008-11-03 07:36:55 -05:00
if ( ( ! window - > mapped | | window - > hidden ) & & ! window - > shaded )
2002-05-10 22:25:11 -04:00
{
meta_topic ( META_DEBUG_FOCUS ,
" Window %s is not showing, not focusing after all \n " ,
window - > desc ) ;
return ;
}
Fix some support for EWMH hints, and fix USER_TIME support to include the
2004-07-31 Rob Adams <readams@readams.net>
Fix some support for EWMH hints, and fix USER_TIME support to
include the DEMANDS_ATTENTION hint. Also includes some code for
implementing _NET_RESTACK_WINDOW and _NET_MOVERESIZE_WINDOW, but
this is disabled pending feature thaw.
* COMPLIANCE: update with new information
* src/display.c (meta_display_open): add new hints to list
* src/display.h (_MetaDisplay): Add new atoms to struct
* src/screen.c (set_supported_hint): update the list of support
hints.
(set_desktop_viewport_hint): new function sets the viewport hint
to (0,0) as required by the spec for WMs with no viewport support.
(set_desktop_geometry_hint): new function to set the desktop size
hint to the size of the display, since we don't implement large
desktop support, as required by the spec.
(meta_screen_resize): update the geometry hint on screen resize
* src/window.c (meta_window_new_with_attrs): Initialize
demands_attention state
(set_net_wm_state): Set demands_attention hint in the window state
(meta_window_show): If we don't pop up a window because of
USER_TIME, set DEMANDS_ATTENTION on the window.
(meta_window_focus): When a window receives focus, remove
DEMANDS_ATTENTION hint
(meta_window_client_message): Allow other apps to set
DEMANDS_ATTENTION on a window. Also, if the _NET_ACTIVE_WINDOW
hint includes a timestamp, use it.
(update_net_wm_state): Read DEMANDS_ATTENTION state also
* src/window.h (_MetaWindow): add wm_state_demands_attention bit.
2004-07-31 15:56:10 -04:00
2014-03-20 15:16:57 -04:00
META_WINDOW_GET_CLASS ( window ) - > focus ( window , timestamp ) ;
Fix some support for EWMH hints, and fix USER_TIME support to include the
2004-07-31 Rob Adams <readams@readams.net>
Fix some support for EWMH hints, and fix USER_TIME support to
include the DEMANDS_ATTENTION hint. Also includes some code for
implementing _NET_RESTACK_WINDOW and _NET_MOVERESIZE_WINDOW, but
this is disabled pending feature thaw.
* COMPLIANCE: update with new information
* src/display.c (meta_display_open): add new hints to list
* src/display.h (_MetaDisplay): Add new atoms to struct
* src/screen.c (set_supported_hint): update the list of support
hints.
(set_desktop_viewport_hint): new function sets the viewport hint
to (0,0) as required by the spec for WMs with no viewport support.
(set_desktop_geometry_hint): new function to set the desktop size
hint to the size of the display, since we don't implement large
desktop support, as required by the spec.
(meta_screen_resize): update the geometry hint on screen resize
* src/window.c (meta_window_new_with_attrs): Initialize
demands_attention state
(set_net_wm_state): Set demands_attention hint in the window state
(meta_window_show): If we don't pop up a window because of
USER_TIME, set DEMANDS_ATTENTION on the window.
(meta_window_focus): When a window receives focus, remove
DEMANDS_ATTENTION hint
(meta_window_client_message): Allow other apps to set
DEMANDS_ATTENTION on a window. Also, if the _NET_ACTIVE_WINDOW
hint includes a timestamp, use it.
(update_net_wm_state): Read DEMANDS_ATTENTION state also
* src/window.h (_MetaWindow): add wm_state_demands_attention bit.
2004-07-31 15:56:10 -04:00
if ( window - > wm_state_demands_attention )
2006-03-29 13:43:22 -05:00
meta_window_unset_demands_attention ( window ) ;
2006-05-25 10:40:50 -04:00
2009-05-07 13:53:47 -04:00
/* meta_effect_run_focus(window, NULL, NULL); */
2001-06-04 00:58:22 -04:00
}
2002-11-22 01:00:18 -05:00
static void
meta_window_change_workspace_without_transients ( MetaWindow * window ,
MetaWorkspace * workspace )
2002-08-06 00:11:23 -04:00
{
2009-01-20 07:52:06 -05:00
int old_workspace = - 1 ;
2001-06-09 01:14:43 -04:00
meta_verbose ( " Changing window %s to workspace %d \n " ,
window - > desc , meta_workspace_index ( workspace ) ) ;
2008-12-18 07:41:56 -05:00
2011-02-28 05:41:23 -05:00
if ( ! window - > on_all_workspaces_requested )
2009-01-20 07:52:06 -05:00
{
old_workspace = meta_workspace_index ( window - > workspace ) ;
}
2008-12-18 07:41:56 -05:00
/* unstick if stuck. meta_window_unstick would call
2002-10-04 10:49:00 -04:00
* meta_window_change_workspace recursively if the window
* is not in the active workspace .
*/
2011-02-28 05:41:23 -05:00
if ( window - > on_all_workspaces_requested )
2002-10-04 10:49:00 -04:00
meta_window_unstick ( window ) ;
2002-06-17 21:41:08 -04:00
/* See if we're already on this space. If not, make sure we are */
Partially resolve the conflicting requirements of windows on multiple
2004-12-22 Elijah Newren <newren@gmail.com>
Partially resolve the conflicting requirements of windows on
multiple workspaces and hidden being a global quantity for windows
(fixes bug 156182; the remainder of the work is bug 87531 and is a
libwnck issue)
* src/display.c (event_callback):
* src/window.c (meta_window_visible_on_workspace, meta_window_unstick):
* src/workspace.c (meta_workspace_add_window,
meta_workspace_contains_window,
meta_workspace_queue_calc_showing):
* src/workspace.h:
Remove meta_workspace_contains_window, replace with simple
comparison utilizing window->workspace
* src/place.c (meta_window_place):
* src/window.c (meta_window_shares_some_workspace):
* src/window.h:
Remove meta_window_shares_some_workspace, replace with a simple
comparison utilizing window->workspace
* src/session.c (save_state),
* src/window.c (meta_window_new_with_attrs,
meta_window_apply_session_info, meta_window_free,
window_showing_on_its_workspace,
meta_window_change_workspace_without_transients,
meta_window_unstick, meta_window_set_current_workspace_hint,
meta_window_get_workspaces):
* src/window.h:
* src/workspace.c (meta_workspace_free, meta_workspace_add_window,
meta_workspace_remove_window):
Only one workspace now
2004-12-22 18:52:52 -05:00
if ( window - > workspace ! = workspace )
2002-08-06 00:11:23 -04:00
{
Partially resolve the conflicting requirements of windows on multiple
2004-12-22 Elijah Newren <newren@gmail.com>
Partially resolve the conflicting requirements of windows on
multiple workspaces and hidden being a global quantity for windows
(fixes bug 156182; the remainder of the work is bug 87531 and is a
libwnck issue)
* src/display.c (event_callback):
* src/window.c (meta_window_visible_on_workspace, meta_window_unstick):
* src/workspace.c (meta_workspace_add_window,
meta_workspace_contains_window,
meta_workspace_queue_calc_showing):
* src/workspace.h:
Remove meta_workspace_contains_window, replace with simple
comparison utilizing window->workspace
* src/place.c (meta_window_place):
* src/window.c (meta_window_shares_some_workspace):
* src/window.h:
Remove meta_window_shares_some_workspace, replace with a simple
comparison utilizing window->workspace
* src/session.c (save_state),
* src/window.c (meta_window_new_with_attrs,
meta_window_apply_session_info, meta_window_free,
window_showing_on_its_workspace,
meta_window_change_workspace_without_transients,
meta_window_unstick, meta_window_set_current_workspace_hint,
meta_window_get_workspaces):
* src/window.h:
* src/workspace.c (meta_workspace_free, meta_workspace_add_window,
meta_workspace_remove_window):
Only one workspace now
2004-12-22 18:52:52 -05:00
meta_workspace_remove_window ( window - > workspace , window ) ;
meta_workspace_add_window ( workspace , window ) ;
2009-01-20 07:52:06 -05:00
g_signal_emit ( window , window_signals [ WORKSPACE_CHANGED ] , 0 ,
old_workspace ) ;
2002-08-06 00:11:23 -04:00
}
2001-06-09 01:14:43 -04:00
}
2003-03-17 01:36:39 -05:00
static gboolean
2002-11-22 01:00:18 -05:00
change_workspace_foreach ( MetaWindow * window ,
void * data )
{
meta_window_change_workspace_without_transients ( window , data ) ;
2003-03-17 01:36:39 -05:00
return TRUE ;
2002-11-22 01:00:18 -05:00
}
void
meta_window_change_workspace ( MetaWindow * window ,
MetaWorkspace * workspace )
{
2009-06-14 08:11:29 -04:00
g_return_if_fail ( ! window - > override_redirect ) ;
2012-04-14 09:12:48 -04:00
if ( window - > always_sticky )
return ;
2002-11-22 01:00:18 -05:00
meta_window_change_workspace_without_transients ( window , workspace ) ;
meta_window_foreach_transient ( window , change_workspace_foreach ,
workspace ) ;
2005-10-03 15:00:01 -04:00
meta_window_foreach_ancestor ( window , change_workspace_foreach ,
workspace ) ;
2002-11-22 01:00:18 -05:00
}
2005-01-26 18:25:05 -05:00
static void
window_stick_impl ( MetaWindow * window )
2001-06-09 23:17:15 -04:00
{
2002-11-19 23:37:55 -05:00
meta_verbose ( " Sticking window %s current on_all_workspaces = %d \n " ,
window - > desc , window - > on_all_workspaces ) ;
2008-12-18 07:41:56 -05:00
2011-02-28 05:41:23 -05:00
if ( window - > on_all_workspaces_requested )
2001-06-09 23:17:15 -04:00
return ;
/* We don't change window->workspaces, because we revert
* to that original workspace list if on_all_workspaces is
* toggled back off .
*/
2011-02-28 05:41:23 -05:00
window - > on_all_workspaces_requested = TRUE ;
2013-11-15 17:37:50 -05:00
meta_window_frame_size_changed ( window ) ;
2011-02-28 05:41:23 -05:00
meta_window_update_on_all_workspaces ( window ) ;
2008-12-18 07:41:56 -05:00
Refactor thrice-duplicated queue code in window.c. Closes #376760.
2007-06-10 Thomas Thurman <thomas@thurman.org.uk>
Refactor thrice-duplicated queue code in window.c. Closes #376760.
* src/window.c (meta_window_queue, meta_window_unqueue):
New functions.
* src/window.[ch] (meta_window_unqueue_*, meta_window_queue_*):
Removed functions.
* src/window.c (meta_window_new_with_attrs, meta_window_free,
meta_window_flush_calc_showing, queue_calc_showing_func,
meta_window_minimize, meta_window_unminimize, meta_window_maximize,
meta_window_make_fullscreen, meta_window_shade,
meta_window_unshade, meta_window_move_resize_internal,
window_stick_impl, window_unstick_impl,
meta_window_client_message, process_property_notify): Modified to
use new queueing functions.
* src/window.c (idle_move_resize, idle_update_icon,
idle_calc_showing): update to receive queue number from pointer.
* src/window.h (MetaQueueType): new enum.
* src/window.h (MetaWindow): *_queued replaced with is_in_queue
bitfield.
* src/core.c (meta_core_queue_frame_resize):
* src/display.c (event_callback,
meta_display_queue_retheme_all_windows): Using new queueing functions.
* src/frame.c (meta_window_destroy_frame): Using new queueing functions.
* src/screen.c (queue_resize, meta_screen_resize_func,
queue_windows_showing): Using new queueing functions.
* src/window-props.c (reload_mwm_hints, reload_wm_hints,
reload_transient_for): Using new queueing functions.
* src/workspace.c (meta_workspace_add_window,
meta_workspace_remove_window, meta_workspace_queue_calc_showing,
meta_workspace_invalidate_work_area): Using new queueing functions.
svn path=/trunk/; revision=3236
2007-06-10 21:15:33 -04:00
meta_window_queue ( window , META_QUEUE_CALC_SHOWING ) ;
2001-06-09 23:17:15 -04:00
}
2005-01-26 18:25:05 -05:00
static void
window_unstick_impl ( MetaWindow * window )
2001-06-09 23:17:15 -04:00
{
2011-02-28 05:41:23 -05:00
if ( ! window - > on_all_workspaces_requested )
2001-06-09 23:17:15 -04:00
return ;
/* Revert to window->workspaces */
2011-02-28 05:41:23 -05:00
window - > on_all_workspaces_requested = FALSE ;
2013-11-15 17:37:50 -05:00
meta_window_frame_size_changed ( window ) ;
2011-02-28 05:41:23 -05:00
meta_window_update_on_all_workspaces ( window ) ;
2003-08-15 18:09:55 -04:00
2001-06-09 23:17:15 -04:00
/* We change ourselves to the active workspace, since otherwise you'd get
* a weird window - vaporization effect . Once we have UI for being
* on more than one workspace this should probably be add_workspace
* not change_workspace .
*/
Partially resolve the conflicting requirements of windows on multiple
2004-12-22 Elijah Newren <newren@gmail.com>
Partially resolve the conflicting requirements of windows on
multiple workspaces and hidden being a global quantity for windows
(fixes bug 156182; the remainder of the work is bug 87531 and is a
libwnck issue)
* src/display.c (event_callback):
* src/window.c (meta_window_visible_on_workspace, meta_window_unstick):
* src/workspace.c (meta_workspace_add_window,
meta_workspace_contains_window,
meta_workspace_queue_calc_showing):
* src/workspace.h:
Remove meta_workspace_contains_window, replace with simple
comparison utilizing window->workspace
* src/place.c (meta_window_place):
* src/window.c (meta_window_shares_some_workspace):
* src/window.h:
Remove meta_window_shares_some_workspace, replace with a simple
comparison utilizing window->workspace
* src/session.c (save_state),
* src/window.c (meta_window_new_with_attrs,
meta_window_apply_session_info, meta_window_free,
window_showing_on_its_workspace,
meta_window_change_workspace_without_transients,
meta_window_unstick, meta_window_set_current_workspace_hint,
meta_window_get_workspaces):
* src/window.h:
* src/workspace.c (meta_workspace_free, meta_workspace_add_window,
meta_workspace_remove_window):
Only one workspace now
2004-12-22 18:52:52 -05:00
if ( window - > screen - > active_workspace ! = window - > workspace )
2001-06-09 23:17:15 -04:00
meta_window_change_workspace ( window , window - > screen - > active_workspace ) ;
2008-12-18 07:41:56 -05:00
Refactor thrice-duplicated queue code in window.c. Closes #376760.
2007-06-10 Thomas Thurman <thomas@thurman.org.uk>
Refactor thrice-duplicated queue code in window.c. Closes #376760.
* src/window.c (meta_window_queue, meta_window_unqueue):
New functions.
* src/window.[ch] (meta_window_unqueue_*, meta_window_queue_*):
Removed functions.
* src/window.c (meta_window_new_with_attrs, meta_window_free,
meta_window_flush_calc_showing, queue_calc_showing_func,
meta_window_minimize, meta_window_unminimize, meta_window_maximize,
meta_window_make_fullscreen, meta_window_shade,
meta_window_unshade, meta_window_move_resize_internal,
window_stick_impl, window_unstick_impl,
meta_window_client_message, process_property_notify): Modified to
use new queueing functions.
* src/window.c (idle_move_resize, idle_update_icon,
idle_calc_showing): update to receive queue number from pointer.
* src/window.h (MetaQueueType): new enum.
* src/window.h (MetaWindow): *_queued replaced with is_in_queue
bitfield.
* src/core.c (meta_core_queue_frame_resize):
* src/display.c (event_callback,
meta_display_queue_retheme_all_windows): Using new queueing functions.
* src/frame.c (meta_window_destroy_frame): Using new queueing functions.
* src/screen.c (queue_resize, meta_screen_resize_func,
queue_windows_showing): Using new queueing functions.
* src/window-props.c (reload_mwm_hints, reload_wm_hints,
reload_transient_for): Using new queueing functions.
* src/workspace.c (meta_workspace_add_window,
meta_workspace_remove_window, meta_workspace_queue_calc_showing,
meta_workspace_invalidate_work_area): Using new queueing functions.
svn path=/trunk/; revision=3236
2007-06-10 21:15:33 -04:00
meta_window_queue ( window , META_QUEUE_CALC_SHOWING ) ;
2001-06-09 23:17:15 -04:00
}
2005-01-26 18:25:05 -05:00
static gboolean
stick_foreach_func ( MetaWindow * window ,
void * data )
{
gboolean stick ;
stick = * ( gboolean * ) data ;
if ( stick )
window_stick_impl ( window ) ;
else
window_unstick_impl ( window ) ;
return TRUE ;
}
void
meta_window_stick ( MetaWindow * window )
{
gboolean stick = TRUE ;
2009-06-14 08:11:29 -04:00
g_return_if_fail ( ! window - > override_redirect ) ;
2005-01-26 18:25:05 -05:00
window_stick_impl ( window ) ;
meta_window_foreach_transient ( window ,
stick_foreach_func ,
& stick ) ;
}
void
meta_window_unstick ( MetaWindow * window )
{
gboolean stick = FALSE ;
2009-06-14 08:11:29 -04:00
g_return_if_fail ( ! window - > override_redirect ) ;
2005-01-26 18:25:05 -05:00
window_unstick_impl ( window ) ;
meta_window_foreach_transient ( window ,
stick_foreach_func ,
& stick ) ;
}
2001-06-21 01:23:24 -04:00
unsigned long
meta_window_get_net_wm_desktop ( MetaWindow * window )
{
Partially resolve the conflicting requirements of windows on multiple
2004-12-22 Elijah Newren <newren@gmail.com>
Partially resolve the conflicting requirements of windows on
multiple workspaces and hidden being a global quantity for windows
(fixes bug 156182; the remainder of the work is bug 87531 and is a
libwnck issue)
* src/display.c (event_callback):
* src/window.c (meta_window_visible_on_workspace, meta_window_unstick):
* src/workspace.c (meta_workspace_add_window,
meta_workspace_contains_window,
meta_workspace_queue_calc_showing):
* src/workspace.h:
Remove meta_workspace_contains_window, replace with simple
comparison utilizing window->workspace
* src/place.c (meta_window_place):
* src/window.c (meta_window_shares_some_workspace):
* src/window.h:
Remove meta_window_shares_some_workspace, replace with a simple
comparison utilizing window->workspace
* src/session.c (save_state),
* src/window.c (meta_window_new_with_attrs,
meta_window_apply_session_info, meta_window_free,
window_showing_on_its_workspace,
meta_window_change_workspace_without_transients,
meta_window_unstick, meta_window_set_current_workspace_hint,
meta_window_get_workspaces):
* src/window.h:
* src/workspace.c (meta_workspace_free, meta_workspace_add_window,
meta_workspace_remove_window):
Only one workspace now
2004-12-22 18:52:52 -05:00
if ( window - > on_all_workspaces )
2001-06-21 01:23:24 -04:00
return 0xFFFFFFFF ;
else
Partially resolve the conflicting requirements of windows on multiple
2004-12-22 Elijah Newren <newren@gmail.com>
Partially resolve the conflicting requirements of windows on
multiple workspaces and hidden being a global quantity for windows
(fixes bug 156182; the remainder of the work is bug 87531 and is a
libwnck issue)
* src/display.c (event_callback):
* src/window.c (meta_window_visible_on_workspace, meta_window_unstick):
* src/workspace.c (meta_workspace_add_window,
meta_workspace_contains_window,
meta_workspace_queue_calc_showing):
* src/workspace.h:
Remove meta_workspace_contains_window, replace with simple
comparison utilizing window->workspace
* src/place.c (meta_window_place):
* src/window.c (meta_window_shares_some_workspace):
* src/window.h:
Remove meta_window_shares_some_workspace, replace with a simple
comparison utilizing window->workspace
* src/session.c (save_state),
* src/window.c (meta_window_new_with_attrs,
meta_window_apply_session_info, meta_window_free,
window_showing_on_its_workspace,
meta_window_change_workspace_without_transients,
meta_window_unstick, meta_window_set_current_workspace_hint,
meta_window_get_workspaces):
* src/window.h:
* src/workspace.c (meta_workspace_free, meta_workspace_add_window,
meta_workspace_remove_window):
Only one workspace now
2004-12-22 18:52:52 -05:00
return meta_workspace_index ( window - > workspace ) ;
2001-06-21 01:23:24 -04:00
}
2002-10-21 17:44:35 -04:00
void
2001-06-09 23:17:15 -04:00
meta_window_set_current_workspace_hint ( MetaWindow * window )
{
/* FIXME if on more than one workspace, we claim to be "sticky",
* the WM spec doesn ' t say what to do here .
*/
unsigned long data [ 1 ] ;
Partially resolve the conflicting requirements of windows on multiple
2004-12-22 Elijah Newren <newren@gmail.com>
Partially resolve the conflicting requirements of windows on
multiple workspaces and hidden being a global quantity for windows
(fixes bug 156182; the remainder of the work is bug 87531 and is a
libwnck issue)
* src/display.c (event_callback):
* src/window.c (meta_window_visible_on_workspace, meta_window_unstick):
* src/workspace.c (meta_workspace_add_window,
meta_workspace_contains_window,
meta_workspace_queue_calc_showing):
* src/workspace.h:
Remove meta_workspace_contains_window, replace with simple
comparison utilizing window->workspace
* src/place.c (meta_window_place):
* src/window.c (meta_window_shares_some_workspace):
* src/window.h:
Remove meta_window_shares_some_workspace, replace with a simple
comparison utilizing window->workspace
* src/session.c (save_state),
* src/window.c (meta_window_new_with_attrs,
meta_window_apply_session_info, meta_window_free,
window_showing_on_its_workspace,
meta_window_change_workspace_without_transients,
meta_window_unstick, meta_window_set_current_workspace_hint,
meta_window_get_workspaces):
* src/window.h:
* src/workspace.c (meta_workspace_free, meta_workspace_add_window,
meta_workspace_remove_window):
Only one workspace now
2004-12-22 18:52:52 -05:00
if ( window - > workspace = = NULL )
2002-03-10 10:48:38 -05:00
{
2008-12-18 07:41:56 -05:00
/* this happens when unmanaging windows */
2002-10-21 17:44:35 -04:00
return ;
2002-03-10 10:48:38 -05:00
}
2008-12-18 07:41:56 -05:00
2001-06-21 01:23:24 -04:00
data [ 0 ] = meta_window_get_net_wm_desktop ( window ) ;
2001-06-09 23:17:15 -04:00
2006-01-20 17:03:56 -05:00
meta_verbose ( " Setting _NET_WM_DESKTOP of %s to %lu \n " ,
2001-06-09 23:17:15 -04:00
window - > desc , data [ 0 ] ) ;
2008-12-18 07:41:56 -05:00
2001-06-09 23:17:15 -04:00
meta_error_trap_push ( window - > display ) ;
XChangeProperty ( window - > display - > xdisplay , window - > xwindow ,
2008-05-02 14:49:01 -04:00
window - > display - > atom__NET_WM_DESKTOP ,
2001-06-09 23:17:15 -04:00
XA_CARDINAL ,
32 , PropModeReplace , ( guchar * ) data , 1 ) ;
2010-10-25 14:44:30 -04:00
meta_error_trap_pop ( window - > display ) ;
2001-06-09 23:17:15 -04:00
}
2005-02-12 02:34:30 -05:00
static gboolean
find_root_ancestor ( MetaWindow * window ,
void * data )
{
MetaWindow * * ancestor = data ;
/* Overwrite the previously "most-root" ancestor with the new one found */
* ancestor = window ;
/* We want this to continue until meta_window_foreach_ancestor quits because
* there are no more valid ancestors .
*/
return TRUE ;
}
2010-09-01 15:39:53 -04:00
/**
* meta_window_find_root_ancestor :
* @ window : a # MetaWindow
*
* Follow the chain of parents of @ window , skipping transient windows ,
* and return the " root " window which has no non - transient parent .
*
* Returns : ( transfer none ) : The root ancestor window
*/
2006-03-30 14:22:38 -05:00
MetaWindow *
meta_window_find_root_ancestor ( MetaWindow * window )
2001-06-06 00:47:37 -04:00
{
2005-02-12 02:34:30 -05:00
MetaWindow * ancestor ;
ancestor = window ;
meta_window_foreach_ancestor ( window , find_root_ancestor , & ancestor ) ;
2006-03-30 14:22:38 -05:00
return ancestor ;
}
void
meta_window_raise ( MetaWindow * window )
{
MetaWindow * ancestor ;
2009-06-14 08:11:29 -04:00
g_return_if_fail ( ! window - > override_redirect ) ;
2006-03-30 14:22:38 -05:00
ancestor = meta_window_find_root_ancestor ( window ) ;
2005-02-12 02:34:30 -05:00
2002-02-07 22:34:26 -05:00
meta_topic ( META_DEBUG_WINDOW_OPS ,
2008-12-18 07:41:56 -05:00
" Raising window %s, ancestor of %s \n " ,
2005-02-12 02:34:30 -05:00
ancestor - > desc , window - > desc ) ;
2001-06-06 00:47:37 -04:00
2005-07-24 21:52:50 -04:00
/* Raise the ancestor of the window (if the window has no ancestor,
* then ancestor will be set to the window itself ) ; do this because
* it ' s weird to see windows from other apps stacked between a child
* and parent window of the currently active app . The stacking
* constraints in stack . c then magically take care of raising all
* the child windows appropriately .
*/
if ( window - > screen - > stack = = ancestor - > screen - > stack )
meta_stack_raise ( window - > screen - > stack , ancestor ) ;
else
2005-02-12 02:34:30 -05:00
{
meta_warning (
" Either stacks aren't per screen or some window has a weird "
" transient_for hint; window->screen->stack != "
" ancestor->screen->stack. window = %s, ancestor = %s. \n " ,
window - > desc , ancestor - > desc ) ;
2005-07-24 21:52:50 -04:00
/* We could raise the window here, but don't want to do that twice and
* so we let the case below handle that .
*/
2005-02-12 02:34:30 -05:00
}
2005-07-24 21:52:50 -04:00
/* Okay, so stacking constraints misses one case: If a window has
* two children and we want to raise one of those children , then
* raising the ancestor isn ' t enough ; we need to also raise the
* correct child . See bug 307875.
*/
if ( window ! = ancestor )
meta_stack_raise ( window - > screen - > stack , window ) ;
2009-02-02 09:08:46 -05:00
g_signal_emit ( window , window_signals [ RAISED ] , 0 ) ;
2001-06-06 00:47:37 -04:00
}
2001-09-16 17:50:27 -04:00
void
meta_window_lower ( MetaWindow * window )
{
2009-06-14 08:11:29 -04:00
g_return_if_fail ( ! window - > override_redirect ) ;
2002-02-07 22:34:26 -05:00
meta_topic ( META_DEBUG_WINDOW_OPS ,
" Lowering window %s \n " , window - > desc ) ;
2001-09-16 17:50:27 -04:00
meta_stack_lower ( window - > screen - > stack , window ) ;
}
2001-06-04 00:58:22 -04:00
void
meta_window_send_icccm_message ( MetaWindow * window ,
Atom atom ,
Fix issues on 64-bit machines with timestamps by using guint32 (like gtk+
2006-09-13 Elijah Newren <newren gmail com>
* src/common.h (MetaWindowMenuFunc):
* src/core.[ch] (meta_core_user_lower_and_unfocus,
meta_core_user_focus, meta_core_show_window_menu,
meta_core_begin_grab_op, meta_core_end_grab_op):
* src/delete.c (delete_ping_reply_func, delete_ping_timeout_func,
meta_window_delete):
* src/display.[ch] (struct MetaDisplay, struct MetaPingData,
sanity_check_timestamps, meta_display_open, event_callback,
meta_spew_event, meta_display_set_grab_op_cursor,
meta_display_begin_grab_op, meta_display_end_grab_op,
meta_display_ping_timeout, meta_display_ping_window,
process_pong_message, timestamp_too_old,
meta_display_set_input_focus_window):
* src/keybindings.[ch] (grab_keyboard, ungrab_keyboard,
meta_screen_grab_all_keys, meta_window_grab_all_keys,
meta_window_ungrab_all_keys, error_on_generic_command,
error_on_command, error_on_terminal_command):
* src/metacity-dialog.c (on_realize, warn_about_no_sm_support,
error_about_command, main):
* src/screen.[ch] (struct _MetaScreen, meta_screen_new,
meta_screen_show_desktop, meta_screen_apply_startup_properties):
* src/session.c (warn_about_lame_clients_and_finish_interact):
* src/window.[ch] (struct _MetaWindow,
intervening_user_event_occurred, window_activate,
meta_window_delete, meta_window_focus,
meta_window_send_icccm_message, meta_window_client_message,
menu_callback, meta_window_show_menu, struct EventScannerData,
check_use_this_motion_notify, meta_window_begin_grab_op,
meta_window_set_user_time):
* src/workspace.[ch] (focus_ancestor_or_mru_window,
meta_workspace_activate_with_focus, meta_workspace_activate,
meta_workspace_focus_default_window,
focus_ancestor_or_mru_window):
Fix issues on 64-bit machines with timestamps by using guint32
(like gtk+ does) instead of Time. #348305
2006-09-13 12:32:33 -04:00
guint32 timestamp )
2001-06-04 00:58:22 -04:00
{
/* This comment and code are from twm, copyright
* Open Group , Evans & Sutherland , etc .
*/
2008-12-18 07:41:56 -05:00
2001-06-04 00:58:22 -04:00
/*
* ICCCM Client Messages - Section 4.2 .8 of the ICCCM dictates that all
* client messages will have the following form :
*
* event type ClientMessage
* message type _XA_WM_PROTOCOLS
* window tmp - > w
* format 32
* data [ 0 ] message atom
* data [ 1 ] time stamp
*/
2008-12-18 07:41:56 -05:00
2001-06-04 00:58:22 -04:00
XClientMessageEvent ev ;
2008-12-18 07:41:56 -05:00
2001-06-04 00:58:22 -04:00
ev . type = ClientMessage ;
ev . window = window - > xwindow ;
2008-05-02 14:49:01 -04:00
ev . message_type = window - > display - > atom_WM_PROTOCOLS ;
2001-06-04 00:58:22 -04:00
ev . format = 32 ;
ev . data . l [ 0 ] = atom ;
ev . data . l [ 1 ] = timestamp ;
2002-04-21 23:23:29 -04:00
meta_error_trap_push ( window - > display ) ;
2001-06-04 00:58:22 -04:00
XSendEvent ( window - > display - > xdisplay ,
window - > xwindow , False , 0 , ( XEvent * ) & ev ) ;
2010-10-25 14:44:30 -04:00
meta_error_trap_pop ( window - > display ) ;
2001-06-04 00:58:22 -04:00
}
2007-04-09 19:00:55 -04:00
void
meta_window_move_resize_request ( MetaWindow * window ,
guint value_mask ,
int gravity ,
int new_x ,
int new_y ,
int new_width ,
int new_height )
2001-06-02 21:33:27 -04:00
{
2001-07-07 14:36:50 -04:00
int x , y , width , height ;
2002-04-30 23:23:46 -04:00
gboolean allow_position_change ;
2002-07-24 11:41:44 -04:00
gboolean in_grab_op ;
Patch from Jens Granseuer to fix build with gcc 2.95. #322622.
2006-01-02 Elijah Newren <newren@gmail.com>
Patch from Jens Granseuer to fix build with gcc 2.95. #322622.
* src/boxes.c (meta_rectangle_region_to_string,
meta_rectangle_edge_list_to_string, fix_up_edges):
* src/constraints.c (meta_window_constrain, setup_constraint_info,
place_window_if_needed, constrain_maximization,
constrain_fullscreen, constrain_size_increments,
constrain_size_limits, constrain_aspect_ratio,
do_screen_and_xinerama_relative_constrai,
constrain_to_single_xinerama, constrain_fully_onscreen,
constrain_partially_onscreen):
* src/edge-resistance.c (find_nearest_position,
apply_edge_resistance, apply_edge_resistance_to_each_side):
* src/testboxes.c (test_clamping_to_region,
test_clipping_to_region, test_shoving_into_region):
* src/window.c (meta_window_new_with_attrs,
meta_window_apply_session_info, meta_window_resize,
meta_window_resize_with_gravity, meta_window_configure_request):
Remove C99 style variable initiailization
2006-01-02 13:37:46 -05:00
MetaMoveResizeFlags flags ;
2002-07-24 11:41:44 -04:00
/* We ignore configure requests while the user is moving/resizing
* the window , since these represent the app sucking and fighting
* the user , most likely due to a bug in the app ( e . g . pfaedit
* seemed to do this )
*
* Still have to do the ConfigureNotify and all , but pretend the
* app asked for the current size / position instead of the new one .
2008-12-18 07:41:56 -05:00
*/
2002-07-24 11:41:44 -04:00
in_grab_op = FALSE ;
if ( window - > display - > grab_op ! = META_GRAB_OP_NONE & &
window = = window - > display - > grab_window )
{
switch ( window - > display - > grab_op )
{
case META_GRAB_OP_MOVING :
case META_GRAB_OP_RESIZING_SE :
case META_GRAB_OP_RESIZING_S :
case META_GRAB_OP_RESIZING_SW :
case META_GRAB_OP_RESIZING_N :
case META_GRAB_OP_RESIZING_NE :
case META_GRAB_OP_RESIZING_NW :
case META_GRAB_OP_RESIZING_W :
case META_GRAB_OP_RESIZING_E :
in_grab_op = TRUE ;
break ;
default :
break ;
}
}
2008-12-18 07:41:56 -05:00
2001-07-07 12:50:01 -04:00
/* it's essential to use only the explicitly-set fields,
* and otherwise use our current up - to - date position .
*
* Otherwise you get spurious position changes when the app changes
* size , for example , if window - > rect is not in sync with the
* server - side position in effect when the configure request was
* generated .
*/
2007-04-09 19:00:55 -04:00
meta_window_get_gravity_position ( window ,
gravity ,
& x , & y ) ;
2001-07-07 12:50:01 -04:00
2002-04-30 23:23:46 -04:00
allow_position_change = FALSE ;
2008-12-18 07:41:56 -05:00
2002-04-30 23:23:46 -04:00
if ( meta_prefs_get_disable_workarounds ( ) )
{
if ( window - > type = = META_WINDOW_DIALOG | |
window - > type = = META_WINDOW_MODAL_DIALOG | |
window - > type = = META_WINDOW_SPLASHSCREEN )
; /* No position change for these */
else if ( ( window - > size_hints . flags & PPosition ) | |
/* USPosition is just stale if window is placed;
* no - - geometry involved here .
*/
( ( window - > size_hints . flags & USPosition ) & &
! window - > placed ) )
allow_position_change = TRUE ;
}
else
{
2002-05-24 22:03:09 -04:00
allow_position_change = TRUE ;
2002-04-30 23:23:46 -04:00
}
2002-07-24 11:41:44 -04:00
if ( in_grab_op )
allow_position_change = FALSE ;
2008-12-18 07:41:56 -05:00
2002-04-30 23:23:46 -04:00
if ( allow_position_change )
2001-08-17 18:44:36 -04:00
{
2007-04-09 19:00:55 -04:00
if ( value_mask & CWX )
x = new_x ;
if ( value_mask & CWY )
y = new_y ;
if ( value_mask & ( CWX | CWY ) )
2002-04-30 23:23:46 -04:00
{
/* Once manually positioned, windows shouldn't be placed
* by the window manager .
*/
window - > placed = TRUE ;
}
2001-07-07 12:50:01 -04:00
}
2002-05-24 22:03:09 -04:00
else
{
meta_topic ( META_DEBUG_GEOMETRY ,
2008-12-18 07:41:56 -05:00
" Not allowing position change for window %s PPosition 0x%lx USPosition 0x%lx type %u \n " ,
window - > desc , window - > size_hints . flags & PPosition ,
2002-05-24 22:03:09 -04:00
window - > size_hints . flags & USPosition ,
2008-12-18 07:41:56 -05:00
window - > type ) ;
2002-05-24 22:03:09 -04:00
}
2001-07-07 12:50:01 -04:00
width = window - > rect . width ;
height = window - > rect . height ;
2002-07-24 11:41:44 -04:00
if ( ! in_grab_op )
{
2007-04-09 19:00:55 -04:00
if ( value_mask & CWWidth )
width = new_width ;
2008-12-18 07:41:56 -05:00
2007-04-09 19:00:55 -04:00
if ( value_mask & CWHeight )
height = new_height ;
2002-07-24 11:41:44 -04:00
}
2001-07-07 12:50:01 -04:00
/* ICCCM 4.1.5 */
2008-12-18 07:41:56 -05:00
2001-07-07 12:50:01 -04:00
/* We're ignoring the value_mask here, since sizes
* not in the mask will be the current window geometry .
*/
window - > size_hints . x = x ;
window - > size_hints . y = y ;
window - > size_hints . width = width ;
window - > size_hints . height = height ;
2007-04-08 23:43:55 -04:00
/* NOTE: We consider ConfigureRequests to be "user" actions in one
* way , but not in another . Explanation of the two cases are in the
* next two big comments .
*/
/* The constraints code allows user actions to move windows
* offscreen , etc . , and configure request actions would often send
* windows offscreen when users don ' t want it if not constrained
* ( e . g . hitting a dropdown triangle in a fileselector to show more
* options , which makes the window bigger ) . Thus we do not set
* META_IS_USER_ACTION in flags to the
* meta_window_move_resize_internal ( ) call .
2001-08-18 22:45:55 -04:00
*/
Patch from Jens Granseuer to fix build with gcc 2.95. #322622.
2006-01-02 Elijah Newren <newren@gmail.com>
Patch from Jens Granseuer to fix build with gcc 2.95. #322622.
* src/boxes.c (meta_rectangle_region_to_string,
meta_rectangle_edge_list_to_string, fix_up_edges):
* src/constraints.c (meta_window_constrain, setup_constraint_info,
place_window_if_needed, constrain_maximization,
constrain_fullscreen, constrain_size_increments,
constrain_size_limits, constrain_aspect_ratio,
do_screen_and_xinerama_relative_constrai,
constrain_to_single_xinerama, constrain_fully_onscreen,
constrain_partially_onscreen):
* src/edge-resistance.c (find_nearest_position,
apply_edge_resistance, apply_edge_resistance_to_each_side):
* src/testboxes.c (test_clamping_to_region,
test_clipping_to_region, test_shoving_into_region):
* src/window.c (meta_window_new_with_attrs,
meta_window_apply_session_info, meta_window_resize,
meta_window_resize_with_gravity, meta_window_configure_request):
Remove C99 style variable initiailization
2006-01-02 13:37:46 -05:00
flags = META_IS_CONFIGURE_REQUEST ;
2007-04-09 19:00:55 -04:00
if ( value_mask & ( CWX | CWY ) )
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
flags | = META_IS_MOVE_ACTION ;
2007-04-09 19:00:55 -04:00
if ( value_mask & ( CWWidth | CWHeight ) )
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
flags | = META_IS_RESIZE_ACTION ;
if ( flags & ( META_IS_MOVE_ACTION | META_IS_RESIZE_ACTION ) )
2008-12-18 07:41:56 -05:00
meta_window_move_resize_internal ( window ,
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
flags ,
2007-04-09 19:00:55 -04:00
gravity ,
x ,
y ,
width ,
height ) ;
2001-07-07 13:08:42 -04:00
2007-04-08 23:43:55 -04:00
/* window->user_rect exists to allow "snapping-back" the window if a
* new strut is set ( causing the window to move ) and then the strut
* is later removed without the user moving the window in the
* interim . We ' d like to " snap-back " to the position specified by
* ConfigureRequest events ( at least the constrained version of the
* ConfigureRequest , since that is guaranteed to be onscreen ) so we
* set user_rect here .
*
* See also bug 426519.
*/
2008-08-15 23:09:40 -04:00
save_user_window_placement ( window ) ;
2007-04-09 19:00:55 -04:00
}
2008-11-20 04:14:58 -05:00
/*
* Move window to the requested workspace ; append controls whether new WS
* should be created if one does not exist .
*/
void
meta_window_change_workspace_by_index ( MetaWindow * window ,
gint space_index ,
2013-10-09 15:36:51 -04:00
gboolean append )
2008-11-20 04:14:58 -05:00
{
MetaWorkspace * workspace ;
MetaScreen * screen ;
2009-06-14 08:11:29 -04:00
g_return_if_fail ( ! window - > override_redirect ) ;
2008-11-20 04:14:58 -05:00
if ( space_index = = - 1 )
{
meta_window_stick ( window ) ;
return ;
}
screen = window - > screen ;
workspace =
meta_screen_get_workspace_by_index ( screen , space_index ) ;
if ( ! workspace & & append )
2013-10-09 15:36:51 -04:00
workspace = meta_screen_append_new_workspace ( screen , FALSE , CurrentTime ) ;
2008-11-20 04:14:58 -05:00
if ( workspace )
{
2011-02-28 05:41:23 -05:00
if ( window - > on_all_workspaces_requested )
2008-11-20 04:14:58 -05:00
meta_window_unstick ( window ) ;
meta_window_change_workspace ( window , workspace ) ;
}
}
2001-06-02 21:33:27 -04:00
2011-10-26 14:00:18 -04:00
static void
meta_window_appears_focused_changed ( MetaWindow * window )
{
set_net_wm_state ( window ) ;
2013-11-15 17:37:50 -05:00
meta_window_frame_size_changed ( window ) ;
2011-10-26 14:00:18 -04:00
g_object_notify ( G_OBJECT ( window ) , " appears-focused " ) ;
if ( window - > frame )
meta_frame_queue_draw ( window - > frame ) ;
2014-02-18 23:03:56 -05:00
if ( window - > surface )
{
if ( meta_window_appears_focused ( window ) )
meta_wayland_surface_activated ( window - > surface ) ;
else
meta_wayland_surface_deactivated ( window - > surface ) ;
}
2011-10-26 14:00:18 -04:00
}
2011-04-13 17:08:12 -04:00
/**
* meta_window_propagate_focus_appearance :
* @ window : the window to start propagating from
* @ focused : % TRUE if @ window ' s ancestors should appear focused ,
* % FALSE if they should not .
*
* Adjusts the value of # MetaWindow : appears - focused on @ window ' s
* ancestors ( but not @ window itself ) . If @ focused is % TRUE , each of
* @ window ' s ancestors will have its % attached_focus_window field set
* to the current % focus_window . If @ focused if % FALSE , each of
* @ window ' s ancestors will have its % attached_focus_window field
* cleared if it is currently % focus_window .
*/
2011-03-24 21:36:47 -04:00
void
meta_window_propagate_focus_appearance ( MetaWindow * window ,
gboolean focused )
2010-09-10 05:55:20 -04:00
{
2011-04-13 17:08:12 -04:00
MetaWindow * child , * parent , * focus_window ;
2010-09-10 05:55:20 -04:00
2011-04-13 17:08:12 -04:00
focus_window = window - > display - > focus_window ;
2011-03-24 21:36:47 -04:00
child = window ;
parent = meta_window_get_transient_for ( child ) ;
window: make determination of attached dialog windows more consistent
Different bits of code were using slightly different checks to test
whether a window was an attached dialog. Add a new
meta_window_is_attached_dialog(), and use that everywhere.
Also, freeze the is-attached status when the window is first shown,
rather than recomputing it each time the caller asks, since this could
cause problems if a window changes its type after it has already been
attached, etc. However, if an attached window's parent is destroyed,
or an attached window changes its transient-for, then fix things up by
destroying the old MetaWindow and creating a new one (causing
compositor unmap and map events to be fired off, allowing the display
of the window to be fixed up).
Remove some code in display.c that tried to fix existing windows if
the gconf setting changed, but which didn't actually do anything (at
least under gnome-shell). However, if 654643 was fixed then the new
behavior with this patch would be that changing the gconf setting
would affect new dialogs, but not existing ones.
https://bugzilla.gnome.org/show_bug.cgi?id=646761
2011-04-05 10:14:52 -04:00
while ( parent & & ( ! focused | | meta_window_is_attached_dialog ( child ) ) )
2011-03-24 21:36:47 -04:00
{
gboolean child_focus_state_changed ;
2010-09-10 05:55:20 -04:00
2011-03-24 21:36:47 -04:00
if ( focused )
{
2011-04-13 17:08:12 -04:00
if ( parent - > attached_focus_window = = focus_window )
2011-03-24 21:36:47 -04:00
break ;
child_focus_state_changed = ( parent - > attached_focus_window = = NULL ) ;
2011-04-13 17:08:12 -04:00
parent - > attached_focus_window = focus_window ;
2011-03-24 21:36:47 -04:00
}
else
{
2011-04-13 17:08:12 -04:00
if ( parent - > attached_focus_window ! = focus_window )
2011-03-24 21:36:47 -04:00
break ;
child_focus_state_changed = ( parent - > attached_focus_window ! = NULL ) ;
parent - > attached_focus_window = NULL ;
}
2011-04-26 08:12:04 -04:00
if ( child_focus_state_changed & & ! parent - > has_focus )
2011-03-24 21:36:47 -04:00
{
2011-10-26 14:00:18 -04:00
meta_window_appears_focused_changed ( parent ) ;
2011-03-24 21:36:47 -04:00
}
child = parent ;
parent = meta_window_get_transient_for ( child ) ;
}
2010-09-10 05:55:20 -04:00
}
2011-04-26 08:12:04 -04:00
void
2012-12-16 21:24:32 -05:00
meta_window_set_focused_internal ( MetaWindow * window ,
gboolean focused )
2012-12-16 21:22:05 -05:00
{
2012-12-16 21:24:32 -05:00
if ( focused )
2012-12-16 21:22:05 -05:00
{
2012-12-16 22:24:15 -05:00
window - > has_focus = TRUE ;
2013-05-01 17:37:50 -04:00
if ( window - > override_redirect )
return ;
2012-12-16 22:24:15 -05:00
/* Move to the front of the focusing workspace's MRU list.
* We should only be " removing " it from the MRU list if it ' s
* not already there . Note that it ' s possible that we might
* be processing this FocusIn after we ' ve changed to a
* different workspace ; we should therefore update the MRU
* list only if the window is actually on the active
* workspace .
*/
if ( window - > screen - > active_workspace & &
meta_window_located_on_workspace ( window ,
window - > screen - > active_workspace ) )
2012-12-16 21:22:05 -05:00
{
2012-12-16 22:24:15 -05:00
GList * link ;
link = g_list_find ( window - > screen - > active_workspace - > mru_list ,
window ) ;
g_assert ( link ) ;
window - > screen - > active_workspace - > mru_list =
g_list_remove_link ( window - > screen - > active_workspace - > mru_list ,
link ) ;
g_list_free ( link ) ;
window - > screen - > active_workspace - > mru_list =
g_list_prepend ( window - > screen - > active_workspace - > mru_list ,
window ) ;
}
2012-12-16 21:22:05 -05:00
2012-12-16 22:24:15 -05:00
if ( window - > frame )
meta_frame_queue_draw ( window - > frame ) ;
2012-12-16 21:22:05 -05:00
2012-12-16 22:24:15 -05:00
meta_error_trap_push ( window - > display ) ;
XInstallColormap ( window - > display - > xdisplay ,
window - > colormap ) ;
meta_error_trap_pop ( window - > display ) ;
2012-12-16 21:22:05 -05:00
2012-12-16 22:24:15 -05:00
/* move into FOCUSED_WINDOW layer */
meta_window_update_layer ( window ) ;
2012-12-16 21:22:05 -05:00
2012-12-16 22:24:15 -05:00
/* Ungrab click to focus button since the sync grab can interfere
* with some things you might do inside the focused window , by
* causing the client to get funky enter / leave events .
*
* The reason we usually have a passive grab on the window is
* so that we can intercept clicks and raise the window in
* response . For click - to - focus we don ' t need that since the
* focused window is already raised . When raise_on_click is
* FALSE we also don ' t need that since we don ' t do anything
* when the window is clicked .
*
* There is dicussion in bugs 102209 , 115072 , and 461577
*/
if ( meta_prefs_get_focus_mode ( ) = = G_DESKTOP_FOCUS_MODE_CLICK | |
! meta_prefs_get_raise_on_click ( ) )
meta_display_ungrab_focus_window_button ( window - > display , window ) ;
2012-12-16 21:22:05 -05:00
2012-12-16 22:24:15 -05:00
g_signal_emit ( window , window_signals [ FOCUS ] , 0 ) ;
2012-12-16 21:22:05 -05:00
2012-12-16 22:24:15 -05:00
if ( ! window - > attached_focus_window )
meta_window_appears_focused_changed ( window ) ;
2012-12-16 21:22:05 -05:00
2012-12-16 22:24:15 -05:00
meta_window_propagate_focus_appearance ( window , TRUE ) ;
2012-12-16 21:22:05 -05:00
}
2012-12-16 21:24:32 -05:00
else
2012-11-15 16:35:42 -05:00
{
2012-12-16 22:24:15 -05:00
window - > has_focus = FALSE ;
2013-05-01 17:37:50 -04:00
if ( window - > override_redirect )
return ;
meta_window_propagate_focus_appearance ( window , FALSE ) ;
2012-11-15 16:35:42 -05:00
2012-12-16 22:24:15 -05:00
if ( ! window - > attached_focus_window )
meta_window_appears_focused_changed ( window ) ;
2012-11-15 16:35:42 -05:00
2012-12-16 22:24:15 -05:00
meta_error_trap_push ( window - > display ) ;
XUninstallColormap ( window - > display - > xdisplay ,
window - > colormap ) ;
meta_error_trap_pop ( window - > display ) ;
2012-11-15 16:35:42 -05:00
2012-12-16 22:24:15 -05:00
/* move out of FOCUSED_WINDOW layer */
meta_window_update_layer ( window ) ;
2012-11-15 16:35:42 -05:00
2012-12-16 22:24:15 -05:00
/* Re-grab for click to focus and raise-on-click, if necessary */
if ( meta_prefs_get_focus_mode ( ) = = G_DESKTOP_FOCUS_MODE_CLICK | |
! meta_prefs_get_raise_on_click ( ) )
meta_display_grab_focus_window_button ( window - > display , window ) ;
2012-11-15 16:35:42 -05:00
}
}
2009-03-31 17:29:48 -04:00
/**
* meta_window_get_icon_geometry :
* @ window : a # MetaWindow
2013-01-20 15:35:32 -05:00
* @ rect : ( out ) : rectangle into which to store the returned geometry .
2009-03-31 17:29:48 -04:00
*
* Gets the location of the icon corresponding to the window . The location
* will be provided set by the task bar or other user interface element
2013-02-01 05:32:53 -05:00
* displaying the icon , and is relative to the root window .
2009-03-31 17:29:48 -04:00
*
* Return value : % TRUE if the icon geometry was succesfully retrieved .
*/
2002-10-24 01:21:51 -04:00
gboolean
2001-08-05 12:04:52 -04:00
meta_window_get_icon_geometry ( MetaWindow * window ,
MetaRectangle * rect )
{
2009-06-14 08:37:57 -04:00
g_return_val_if_fail ( ! window - > override_redirect , FALSE ) ;
2013-02-01 05:32:53 -05:00
if ( window - > icon_geometry_set )
2001-08-19 21:42:44 -04:00
{
2002-01-03 19:58:39 -05:00
if ( rect )
2013-02-01 05:32:53 -05:00
* rect = window - > icon_geometry ;
2001-08-05 12:04:52 -04:00
2002-01-03 19:58:39 -05:00
return TRUE ;
}
return FALSE ;
2001-08-05 12:04:52 -04:00
}
2013-01-22 19:11:23 -05:00
/**
* meta_window_set_icon_geometry :
* @ window : a # MetaWindow
* @ rect : ( allow - none ) : rectangle with the desired geometry or % NULL .
*
* Sets or unsets the location of the icon corresponding to the window . If
* set , the location should correspond to a dock , task bar or other user
* interface element displaying the icon , and is relative to the root window .
*/
void
meta_window_set_icon_geometry ( MetaWindow * window ,
MetaRectangle * rect )
{
if ( rect )
{
window - > icon_geometry = * rect ;
window - > icon_geometry_set = TRUE ;
}
else
{
window - > icon_geometry_set = FALSE ;
}
}
2002-10-21 17:44:35 -04:00
static void
2003-02-27 20:24:44 -05:00
redraw_icon ( MetaWindow * window )
{
/* We could probably be smart and just redraw the icon here,
* instead of the whole frame .
*/
Always map the client and frame windows
Traditionally, WMs unmap windows when minimizing them, and map them
when restoring them or wanting to show them for other reasons, like
upon creation.
However, as metacity morphed into mutter, we optionally chose to keep
windows mapped for the lifetime of the window under the user option
"live-window-previews", which makes the code keep windows mapped so it
can show window preview for minimized windows in other places, like
Alt-Tab and Expose.
I removed this preference two years ago mechanically, by removing all
the if statements, but never went through and cleaned up the code so
that windows are simply mapped for the lifetime of the window -- the
"architecture" of the old code that maps and unmaps on show/hide was
still there.
Remove this now.
The one case we still need to be careful of is shaded windows, in which
we do still unmap the client window. In the future, we might want to
show previews of shaded windows in the overview and Alt-Tab. In that
we'd also keep shaded windows mapped, and could remove all unmap logic,
but we'd need a more complex method of showing the shaded titlebar, such
as using a different actor.
At the same time, simplify the compositor interface by removing
meta_compositor_window_[un]mapped API, and instead adding/removing the
window on-demand.
https://bugzilla.gnome.org/show_bug.cgi?id=720631
2014-02-01 19:06:51 -05:00
if ( window - > frame )
2003-02-27 20:24:44 -05:00
meta_ui_queue_frame_draw ( window - > screen - > ui , window - > frame - > xwindow ) ;
}
Added "above" to the list of flags a frame can have, so that we know when
* common.h: Added "above" to the list of flags a frame can have, so
that we know when to mark it as always on top. Added six grab ops,
one to do and one to undo each of the three new titlebar buttons
(shade, above, stick). Added six new button functions, similarly.
(#96229)
* frame.c (meta_frame_get_flags): If a frame has the WM_STATE_ABOVE X
attribute, set META_FRAME_ABOVE in its flags.
* frames.c (meta_frames_apply_shapes): Allow variable amounts of
rounding. (#113162)
* frames.c (show_tip_now, meta_frames_paint_to_drawable, control_rect,
get_control): extend handling of existing buttons to the
3*2 new kinds of button. (#96229)
* frames.c (meta_frames_button_press_event): translate clicks on the 3*2
new kinds of button to the new grab ops. (#96229)
* frames.c (meta_frames_button_release_event): implement the various
actions for the 3*2 new kinds of button. (#96229)
* frames.c (meta_frames_update_prelit_control,
meta_frames_motion_notify_event): extend existing motion
notifications for buttons to the 3*2 new kinds of button. (#96229)
* frames.c (meta_frames_set_window_background): handle specified
background colours and alpha transparency. (#151261)
* frames.h (MetaFrameControl): New control types for the 3*2 new kinds
of button. (#96229)
* iconcache.[ch] (meta_read_icons): use theme's fallback icons if a
window has no icon; use metacity's fallback icons only if the theme
does not provide any. (#11363)
* iconcache.[ch] (meta_invalidate_default_icons (new function)): clear
icon cache on windows using default icons, and update them. (#11363)
* main.c (main): added \n to error message.
* prefs.c (button_function_from_string): extend for 3 new button
types. (#96229)
* prefs.c (button_opposite_function (new function)): return a button
function's inverse (shade -> unshade, etc) (#96229)
* prefs.c (update_button_layout): allocate space for a button's
inverse, if it has one. (#96229)
* theme-parser.c (ParseState): add state for fallback icons (#11363)
* theme-parser.c (ParseInfo): add format_version; remove
menu_icon_* (#114305)
* theme-parser.c (parse_positive_integer): add lookup for integer
constants (#331356)
* theme-parser.c (parse_rounding (new function)): parse window
rounding amount (#113162)
* theme-parser.c (parse_alpha): don't set error if the number can't
be parsed since it'll already be set; change tolerance in comparison
from 1e6 to 1e-6
* theme-parser.c (parse_color (new function)): parse colour, including
possible constant lookup.
* theme-parser.c (parse_toplevel_element): allow defining of various
new kinds of constant; allow
hide_buttons (#121639) and more detailed rounding attributes on
<frame_geometry> (#113162); allow background and alpha attributes on
<frame_style>; (#151261) remove support for <menu_icon> except as
stub; (#114305) add support for loading stock images (#113465); add
support for <fallback>. (#11363))
* theme-parser.c (parse_draw_op_element): add from and to attribute
for arcs. (#121603)
* theme-parser.c (parse_style_element): add check for theme version
supporting a button function. (#96229)
* theme-parser.c (parse_style_set_element): add ability for shaded
windows to be resizable (#114304)
* theme-parser.c (meta_theme_load): add theme versioning routine.
* theme.c ( meta_frame_layout_get_borders): return rectangles for
the new 3*2 kinds of button, except where they're
inapplicable. (#96229)
* theme.c (meta_frame_layout_calc_geometry): don't format buttons on
windows with no buttons (#121639); strip the 3*2 new kinds of button
correctly (#96229); allow variable amounts of rounding (#113162).
* theme.c (meta_frame_style_new): set alpha to 255 by
default. (#151261)
* theme.c (meta_frame_style_unref): free colour spec if
allocated. (#151261)
* theme.c (meta_frame_style_validate): it's only an error not to
include a button if that button is valid in the current
theme. (#96229)
* theme.c (button_rect): return rectangles for the new 3*2 kinds
of button. (#96229)
* theme.c (meta_frame_style_set_unref): free differently resizable
shaded styles. (#114304)
* theme.c (get_style): look up differently resizable styles
for shaded windows. (#114304)
* theme.c (free_menu_ops (removed function), get_menu_icon
(removed function), meta_theme_draw_menu_icon (removed function),
meta_menu_icon_type_from_string (removed function),
meta_menu_icon_type_to_string (removed function),
meta_theme_free, meta_theme_validate): removed menu icon code. (#114305)
* theme.c (meta_theme_load_image): add size_of_theme_icons
parameter. (#113465)
* theme.c (meta_theme_define_color_constant (new function),
meta_theme_lookup_color_constant (new function)): allow
definition of colour constants. (#129747)
* theme.c (meta_button_type_from_string, meta_button_type_to_string):
add the 3*2 new kinds of button. (#96229)
* theme.c (meta_theme_earliest_version_with_button (new function)):
return the theme version each button was introduced in. (#96229)
* theme.h ( MetaFrameLayout): add "hide_buttons" flag (#121639) and
corner radiuses. (#113162)
* theme.h (MetaFrameGeometry): add rectangles for the 3*2 new
buttons. (#96229)
* theme.h (MetaButtonType): the 3*2 new buttons. (#96229)
* theme.h (MetaFrameStyle): add window_background_color and
window_background_alpha so that we can specify background on a
<frame_style>. (#151261)
* theme.h (MetaFrameStyleSet): shaded_styles gets resize
dimension. (#114304)
* theme.h (MetaTheme): added format_version, color_constants
hash, (#129747) fallback_icon and fallback_mini_icon, (#11363)
and removed menu_icons. (#114305)
* theme.h (META_THEME_ALLOWS (new macro)): return whether a theme
supports a given feature. Also, several macros representing
new features in v2.
* ui.c (meta_ui_set_current_theme)): also invalidate default
icons. (#11363)
* window.[ch] (meta_window_update_icon_now)): became
non-static. (#11363)
2006-10-07 12:56:47 -04:00
void
2003-02-27 20:24:44 -05:00
meta_window_update_icon_now ( MetaWindow * window )
2001-08-19 21:42:44 -04:00
{
2002-02-09 18:03:52 -05:00
GdkPixbuf * icon ;
GdkPixbuf * mini_icon ;
2001-08-19 21:42:44 -04:00
2009-06-14 08:37:57 -04:00
g_return_if_fail ( ! window - > override_redirect ) ;
2002-02-09 18:03:52 -05:00
icon = NULL ;
mini_icon = NULL ;
2008-12-18 07:41:56 -05:00
2002-02-09 18:03:52 -05:00
if ( meta_read_icons ( window - > screen ,
window - > xwindow ,
& window - > icon_cache ,
window - > wm_hints_pixmap ,
window - > wm_hints_mask ,
& icon ,
META_ICON_WIDTH , META_ICON_HEIGHT ,
& mini_icon ,
META_MINI_ICON_WIDTH ,
META_MINI_ICON_HEIGHT ) )
{
if ( window - > icon )
g_object_unref ( G_OBJECT ( window - > icon ) ) ;
2008-12-18 07:41:56 -05:00
2002-02-09 18:03:52 -05:00
if ( window - > mini_icon )
g_object_unref ( G_OBJECT ( window - > mini_icon ) ) ;
2008-12-18 07:41:56 -05:00
2002-02-09 18:03:52 -05:00
window - > icon = icon ;
window - > mini_icon = mini_icon ;
2003-02-27 20:24:44 -05:00
2008-12-05 20:03:27 -05:00
g_object_freeze_notify ( G_OBJECT ( window ) ) ;
g_object_notify ( G_OBJECT ( window ) , " icon " ) ;
g_object_notify ( G_OBJECT ( window ) , " mini-icon " ) ;
g_object_thaw_notify ( G_OBJECT ( window ) ) ;
2003-02-27 20:24:44 -05:00
redraw_icon ( window ) ;
2001-08-23 20:32:17 -04:00
}
2008-12-18 07:41:56 -05:00
2001-09-17 01:50:02 -04:00
g_assert ( window - > icon ) ;
g_assert ( window - > mini_icon ) ;
2001-06-23 22:22:10 -04:00
}
2003-02-27 20:24:44 -05:00
static gboolean
idle_update_icon ( gpointer data )
2002-05-24 04:16:59 -04:00
{
2003-02-27 20:24:44 -05:00
GSList * tmp ;
GSList * copy ;
Refactor thrice-duplicated queue code in window.c. Closes #376760.
2007-06-10 Thomas Thurman <thomas@thurman.org.uk>
Refactor thrice-duplicated queue code in window.c. Closes #376760.
* src/window.c (meta_window_queue, meta_window_unqueue):
New functions.
* src/window.[ch] (meta_window_unqueue_*, meta_window_queue_*):
Removed functions.
* src/window.c (meta_window_new_with_attrs, meta_window_free,
meta_window_flush_calc_showing, queue_calc_showing_func,
meta_window_minimize, meta_window_unminimize, meta_window_maximize,
meta_window_make_fullscreen, meta_window_shade,
meta_window_unshade, meta_window_move_resize_internal,
window_stick_impl, window_unstick_impl,
meta_window_client_message, process_property_notify): Modified to
use new queueing functions.
* src/window.c (idle_move_resize, idle_update_icon,
idle_calc_showing): update to receive queue number from pointer.
* src/window.h (MetaQueueType): new enum.
* src/window.h (MetaWindow): *_queued replaced with is_in_queue
bitfield.
* src/core.c (meta_core_queue_frame_resize):
* src/display.c (event_callback,
meta_display_queue_retheme_all_windows): Using new queueing functions.
* src/frame.c (meta_window_destroy_frame): Using new queueing functions.
* src/screen.c (queue_resize, meta_screen_resize_func,
queue_windows_showing): Using new queueing functions.
* src/window-props.c (reload_mwm_hints, reload_wm_hints,
reload_transient_for): Using new queueing functions.
* src/workspace.c (meta_workspace_add_window,
meta_workspace_remove_window, meta_workspace_queue_calc_showing,
meta_workspace_invalidate_work_area): Using new queueing functions.
svn path=/trunk/; revision=3236
2007-06-10 21:15:33 -04:00
guint queue_index = GPOINTER_TO_INT ( data ) ;
2003-02-27 20:24:44 -05:00
meta_topic ( META_DEBUG_GEOMETRY , " Clearing the update_icon queue \n " ) ;
/* Work with a copy, for reentrancy. The allowed reentrancy isn't
* complete ; destroying a window while we ' re in here would result in
* badness . But it ' s OK to queue / unqueue update_icons .
*/
Refactor thrice-duplicated queue code in window.c. Closes #376760.
2007-06-10 Thomas Thurman <thomas@thurman.org.uk>
Refactor thrice-duplicated queue code in window.c. Closes #376760.
* src/window.c (meta_window_queue, meta_window_unqueue):
New functions.
* src/window.[ch] (meta_window_unqueue_*, meta_window_queue_*):
Removed functions.
* src/window.c (meta_window_new_with_attrs, meta_window_free,
meta_window_flush_calc_showing, queue_calc_showing_func,
meta_window_minimize, meta_window_unminimize, meta_window_maximize,
meta_window_make_fullscreen, meta_window_shade,
meta_window_unshade, meta_window_move_resize_internal,
window_stick_impl, window_unstick_impl,
meta_window_client_message, process_property_notify): Modified to
use new queueing functions.
* src/window.c (idle_move_resize, idle_update_icon,
idle_calc_showing): update to receive queue number from pointer.
* src/window.h (MetaQueueType): new enum.
* src/window.h (MetaWindow): *_queued replaced with is_in_queue
bitfield.
* src/core.c (meta_core_queue_frame_resize):
* src/display.c (event_callback,
meta_display_queue_retheme_all_windows): Using new queueing functions.
* src/frame.c (meta_window_destroy_frame): Using new queueing functions.
* src/screen.c (queue_resize, meta_screen_resize_func,
queue_windows_showing): Using new queueing functions.
* src/window-props.c (reload_mwm_hints, reload_wm_hints,
reload_transient_for): Using new queueing functions.
* src/workspace.c (meta_workspace_add_window,
meta_workspace_remove_window, meta_workspace_queue_calc_showing,
meta_workspace_invalidate_work_area): Using new queueing functions.
svn path=/trunk/; revision=3236
2007-06-10 21:15:33 -04:00
copy = g_slist_copy ( queue_pending [ queue_index ] ) ;
g_slist_free ( queue_pending [ queue_index ] ) ;
queue_pending [ queue_index ] = NULL ;
2009-09-25 15:47:22 -04:00
queue_later [ queue_index ] = 0 ;
2003-02-27 20:24:44 -05:00
destroying_windows_disallowed + = 1 ;
2008-12-18 07:41:56 -05:00
2003-02-27 20:24:44 -05:00
tmp = copy ;
while ( tmp ! = NULL )
{
MetaWindow * window ;
window = tmp - > data ;
2008-12-18 07:41:56 -05:00
meta_window_update_icon_now ( window ) ;
Refactor thrice-duplicated queue code in window.c. Closes #376760.
2007-06-10 Thomas Thurman <thomas@thurman.org.uk>
Refactor thrice-duplicated queue code in window.c. Closes #376760.
* src/window.c (meta_window_queue, meta_window_unqueue):
New functions.
* src/window.[ch] (meta_window_unqueue_*, meta_window_queue_*):
Removed functions.
* src/window.c (meta_window_new_with_attrs, meta_window_free,
meta_window_flush_calc_showing, queue_calc_showing_func,
meta_window_minimize, meta_window_unminimize, meta_window_maximize,
meta_window_make_fullscreen, meta_window_shade,
meta_window_unshade, meta_window_move_resize_internal,
window_stick_impl, window_unstick_impl,
meta_window_client_message, process_property_notify): Modified to
use new queueing functions.
* src/window.c (idle_move_resize, idle_update_icon,
idle_calc_showing): update to receive queue number from pointer.
* src/window.h (MetaQueueType): new enum.
* src/window.h (MetaWindow): *_queued replaced with is_in_queue
bitfield.
* src/core.c (meta_core_queue_frame_resize):
* src/display.c (event_callback,
meta_display_queue_retheme_all_windows): Using new queueing functions.
* src/frame.c (meta_window_destroy_frame): Using new queueing functions.
* src/screen.c (queue_resize, meta_screen_resize_func,
queue_windows_showing): Using new queueing functions.
* src/window-props.c (reload_mwm_hints, reload_wm_hints,
reload_transient_for): Using new queueing functions.
* src/workspace.c (meta_workspace_add_window,
meta_workspace_remove_window, meta_workspace_queue_calc_showing,
meta_workspace_invalidate_work_area): Using new queueing functions.
svn path=/trunk/; revision=3236
2007-06-10 21:15:33 -04:00
window - > is_in_queues & = ~ META_QUEUE_UPDATE_ICON ;
2008-12-18 07:41:56 -05:00
2003-02-27 20:24:44 -05:00
tmp = tmp - > next ;
}
g_slist_free ( copy ) ;
destroying_windows_disallowed - = 1 ;
2008-12-18 07:41:56 -05:00
2003-02-27 20:24:44 -05:00
return FALSE ;
}
2003-06-25 23:09:38 -04:00
GList *
2002-01-10 01:31:31 -05:00
meta_window_get_workspaces ( MetaWindow * window )
{
if ( window - > on_all_workspaces )
2002-10-16 16:12:24 -04:00
return window - > screen - > workspaces ;
2012-03-13 00:37:31 -04:00
else if ( window - > workspace ! = NULL )
2006-02-11 00:06:08 -05:00
return window - > workspace - > list_containing_self ;
2012-03-13 00:37:31 -04:00
else
return NULL ;
2002-01-10 01:31:31 -05:00
}
static void
invalidate_work_areas ( MetaWindow * window )
{
GList * tmp ;
tmp = meta_window_get_workspaces ( window ) ;
2008-12-18 07:41:56 -05:00
2002-01-10 01:31:31 -05:00
while ( tmp ! = NULL )
{
meta_workspace_invalidate_work_area ( tmp - > data ) ;
tmp = tmp - > next ;
}
}
2003-06-25 23:09:38 -04:00
void
meta_window_update_struts ( MetaWindow * window )
2002-01-10 01:31:31 -05:00
{
Patch from Carlo Wood to fix handling of unidirectional maximization and
2007-04-02 Elijah Newren <newren gmail com>
Patch from Carlo Wood to fix handling of unidirectional
maximization and partial struts. #358311.
* src/constraints.c (constrain_maximization):
determine target size for unidirectionally maximized windows by
determining how far they can be maximized without hitting
orthogonal struts. Avoids weird "empty spaces".
* src/boxes.[ch] (meta_rectangle_expand_to_avoiding_struts):
new function
2007-04-02 Elijah Newren <newren gmail com>
Make the strut lists (stored in workspaces) record both the
rectangle and the side that the strut is on. Lots of code
cleanups relating to struts.
* src/boxes.h (struct MetaStrut):
new struct for struts
* src/window.[ch] (struct MetaStruts, struct MetaWindow,
meta_window_update_struts):
overhaul to make window's struts remember their side as well as
their rectangular location, and just use a list instead of several
copies of near-identical code for left/right/top/bottom (allowing
us to nuke MetaStruts struct as well)
* src/testboxes.c (new_meta_strut, get_strut_list):
* src/workspace.c (ensure_work_areas_validated):
* src/boxes.c (meta_rectangle_get_minimal_spanning_set_for_region,
meta_rectangle_expand_to_avoiding_struts,
get_disjoint_strut_rect_list_in_region, fix_up_edges,
meta_rectangle_find_onscreen_edges,
meta_rectangle_find_nonintersected_xinerama_edges):
modify to handle struts being rectangle + side instead of just rectangle
* src/workspace.c (ensure_work_areas_validated):
simplify strut list creation considerably given MetaWindow change,
modify work_area computations to take advantage of region
computations being done (makes the code shorter as well as more
robust against pathological cases).
* src/util.[ch] (meta_free_gslist_and_elements):
new convenience function
* src/common.h (enum MetaDirection):
* src/edge-resistance.c (movement_towards_edge):
* src/boxes.c (meta_rectangle_edge_aligns,
rectangle_and_edge_intersection, split_edge):
Add more MetaDirection fields for convenience
* src/boxes.h (enum FixedDirections):
* src/constraints.c (setup_constraint_info, place_window_if_needed):
add a FIXED_DIRECTION_NONE to the FixedDirections enum to make
code more clear
svn path=/trunk/; revision=3144
2007-04-02 23:41:10 -04:00
GSList * old_struts ;
GSList * new_struts ;
GSList * old_iter , * new_iter ;
2002-01-10 01:31:31 -05:00
gulong * struts = NULL ;
int nitems ;
Patch from Carlo Wood to fix handling of unidirectional maximization and
2007-04-02 Elijah Newren <newren gmail com>
Patch from Carlo Wood to fix handling of unidirectional
maximization and partial struts. #358311.
* src/constraints.c (constrain_maximization):
determine target size for unidirectionally maximized windows by
determining how far they can be maximized without hitting
orthogonal struts. Avoids weird "empty spaces".
* src/boxes.[ch] (meta_rectangle_expand_to_avoiding_struts):
new function
2007-04-02 Elijah Newren <newren gmail com>
Make the strut lists (stored in workspaces) record both the
rectangle and the side that the strut is on. Lots of code
cleanups relating to struts.
* src/boxes.h (struct MetaStrut):
new struct for struts
* src/window.[ch] (struct MetaStruts, struct MetaWindow,
meta_window_update_struts):
overhaul to make window's struts remember their side as well as
their rectangular location, and just use a list instead of several
copies of near-identical code for left/right/top/bottom (allowing
us to nuke MetaStruts struct as well)
* src/testboxes.c (new_meta_strut, get_strut_list):
* src/workspace.c (ensure_work_areas_validated):
* src/boxes.c (meta_rectangle_get_minimal_spanning_set_for_region,
meta_rectangle_expand_to_avoiding_struts,
get_disjoint_strut_rect_list_in_region, fix_up_edges,
meta_rectangle_find_onscreen_edges,
meta_rectangle_find_nonintersected_xinerama_edges):
modify to handle struts being rectangle + side instead of just rectangle
* src/workspace.c (ensure_work_areas_validated):
simplify strut list creation considerably given MetaWindow change,
modify work_area computations to take advantage of region
computations being done (makes the code shorter as well as more
robust against pathological cases).
* src/util.[ch] (meta_free_gslist_and_elements):
new convenience function
* src/common.h (enum MetaDirection):
* src/edge-resistance.c (movement_towards_edge):
* src/boxes.c (meta_rectangle_edge_aligns,
rectangle_and_edge_intersection, split_edge):
Add more MetaDirection fields for convenience
* src/boxes.h (enum FixedDirections):
* src/constraints.c (setup_constraint_info, place_window_if_needed):
add a FIXED_DIRECTION_NONE to the FixedDirections enum to make
code more clear
svn path=/trunk/; revision=3144
2007-04-02 23:41:10 -04:00
gboolean changed ;
2003-06-28 12:12:32 -04:00
2009-06-14 08:37:57 -04:00
g_return_if_fail ( ! window - > override_redirect ) ;
2002-01-10 01:31:31 -05:00
meta_verbose ( " Updating struts for %s \n " , window - > desc ) ;
2003-06-25 23:09:38 -04:00
Patch from Carlo Wood to fix handling of unidirectional maximization and
2007-04-02 Elijah Newren <newren gmail com>
Patch from Carlo Wood to fix handling of unidirectional
maximization and partial struts. #358311.
* src/constraints.c (constrain_maximization):
determine target size for unidirectionally maximized windows by
determining how far they can be maximized without hitting
orthogonal struts. Avoids weird "empty spaces".
* src/boxes.[ch] (meta_rectangle_expand_to_avoiding_struts):
new function
2007-04-02 Elijah Newren <newren gmail com>
Make the strut lists (stored in workspaces) record both the
rectangle and the side that the strut is on. Lots of code
cleanups relating to struts.
* src/boxes.h (struct MetaStrut):
new struct for struts
* src/window.[ch] (struct MetaStruts, struct MetaWindow,
meta_window_update_struts):
overhaul to make window's struts remember their side as well as
their rectangular location, and just use a list instead of several
copies of near-identical code for left/right/top/bottom (allowing
us to nuke MetaStruts struct as well)
* src/testboxes.c (new_meta_strut, get_strut_list):
* src/workspace.c (ensure_work_areas_validated):
* src/boxes.c (meta_rectangle_get_minimal_spanning_set_for_region,
meta_rectangle_expand_to_avoiding_struts,
get_disjoint_strut_rect_list_in_region, fix_up_edges,
meta_rectangle_find_onscreen_edges,
meta_rectangle_find_nonintersected_xinerama_edges):
modify to handle struts being rectangle + side instead of just rectangle
* src/workspace.c (ensure_work_areas_validated):
simplify strut list creation considerably given MetaWindow change,
modify work_area computations to take advantage of region
computations being done (makes the code shorter as well as more
robust against pathological cases).
* src/util.[ch] (meta_free_gslist_and_elements):
new convenience function
* src/common.h (enum MetaDirection):
* src/edge-resistance.c (movement_towards_edge):
* src/boxes.c (meta_rectangle_edge_aligns,
rectangle_and_edge_intersection, split_edge):
Add more MetaDirection fields for convenience
* src/boxes.h (enum FixedDirections):
* src/constraints.c (setup_constraint_info, place_window_if_needed):
add a FIXED_DIRECTION_NONE to the FixedDirections enum to make
code more clear
svn path=/trunk/; revision=3144
2007-04-02 23:41:10 -04:00
old_struts = window - > struts ;
2008-12-18 07:41:56 -05:00
new_struts = NULL ;
2003-06-25 23:09:38 -04:00
2002-01-10 01:31:31 -05:00
if ( meta_prop_get_cardinal_list ( window - > display ,
window - > xwindow ,
2008-05-02 14:49:01 -04:00
window - > display - > atom__NET_WM_STRUT_PARTIAL ,
2002-01-10 01:31:31 -05:00
& struts , & nitems ) )
{
2003-06-25 23:09:38 -04:00
if ( nitems ! = 12 )
Patch from Carlo Wood to fix handling of unidirectional maximization and
2007-04-02 Elijah Newren <newren gmail com>
Patch from Carlo Wood to fix handling of unidirectional
maximization and partial struts. #358311.
* src/constraints.c (constrain_maximization):
determine target size for unidirectionally maximized windows by
determining how far they can be maximized without hitting
orthogonal struts. Avoids weird "empty spaces".
* src/boxes.[ch] (meta_rectangle_expand_to_avoiding_struts):
new function
2007-04-02 Elijah Newren <newren gmail com>
Make the strut lists (stored in workspaces) record both the
rectangle and the side that the strut is on. Lots of code
cleanups relating to struts.
* src/boxes.h (struct MetaStrut):
new struct for struts
* src/window.[ch] (struct MetaStruts, struct MetaWindow,
meta_window_update_struts):
overhaul to make window's struts remember their side as well as
their rectangular location, and just use a list instead of several
copies of near-identical code for left/right/top/bottom (allowing
us to nuke MetaStruts struct as well)
* src/testboxes.c (new_meta_strut, get_strut_list):
* src/workspace.c (ensure_work_areas_validated):
* src/boxes.c (meta_rectangle_get_minimal_spanning_set_for_region,
meta_rectangle_expand_to_avoiding_struts,
get_disjoint_strut_rect_list_in_region, fix_up_edges,
meta_rectangle_find_onscreen_edges,
meta_rectangle_find_nonintersected_xinerama_edges):
modify to handle struts being rectangle + side instead of just rectangle
* src/workspace.c (ensure_work_areas_validated):
simplify strut list creation considerably given MetaWindow change,
modify work_area computations to take advantage of region
computations being done (makes the code shorter as well as more
robust against pathological cases).
* src/util.[ch] (meta_free_gslist_and_elements):
new convenience function
* src/common.h (enum MetaDirection):
* src/edge-resistance.c (movement_towards_edge):
* src/boxes.c (meta_rectangle_edge_aligns,
rectangle_and_edge_intersection, split_edge):
Add more MetaDirection fields for convenience
* src/boxes.h (enum FixedDirections):
* src/constraints.c (setup_constraint_info, place_window_if_needed):
add a FIXED_DIRECTION_NONE to the FixedDirections enum to make
code more clear
svn path=/trunk/; revision=3144
2007-04-02 23:41:10 -04:00
meta_verbose ( " _NET_WM_STRUT_PARTIAL on %s has %d values instead "
" of 12 \n " ,
window - > desc , nitems ) ;
2003-06-25 23:09:38 -04:00
else
{
Patch from Carlo Wood to fix handling of unidirectional maximization and
2007-04-02 Elijah Newren <newren gmail com>
Patch from Carlo Wood to fix handling of unidirectional
maximization and partial struts. #358311.
* src/constraints.c (constrain_maximization):
determine target size for unidirectionally maximized windows by
determining how far they can be maximized without hitting
orthogonal struts. Avoids weird "empty spaces".
* src/boxes.[ch] (meta_rectangle_expand_to_avoiding_struts):
new function
2007-04-02 Elijah Newren <newren gmail com>
Make the strut lists (stored in workspaces) record both the
rectangle and the side that the strut is on. Lots of code
cleanups relating to struts.
* src/boxes.h (struct MetaStrut):
new struct for struts
* src/window.[ch] (struct MetaStruts, struct MetaWindow,
meta_window_update_struts):
overhaul to make window's struts remember their side as well as
their rectangular location, and just use a list instead of several
copies of near-identical code for left/right/top/bottom (allowing
us to nuke MetaStruts struct as well)
* src/testboxes.c (new_meta_strut, get_strut_list):
* src/workspace.c (ensure_work_areas_validated):
* src/boxes.c (meta_rectangle_get_minimal_spanning_set_for_region,
meta_rectangle_expand_to_avoiding_struts,
get_disjoint_strut_rect_list_in_region, fix_up_edges,
meta_rectangle_find_onscreen_edges,
meta_rectangle_find_nonintersected_xinerama_edges):
modify to handle struts being rectangle + side instead of just rectangle
* src/workspace.c (ensure_work_areas_validated):
simplify strut list creation considerably given MetaWindow change,
modify work_area computations to take advantage of region
computations being done (makes the code shorter as well as more
robust against pathological cases).
* src/util.[ch] (meta_free_gslist_and_elements):
new convenience function
* src/common.h (enum MetaDirection):
* src/edge-resistance.c (movement_towards_edge):
* src/boxes.c (meta_rectangle_edge_aligns,
rectangle_and_edge_intersection, split_edge):
Add more MetaDirection fields for convenience
* src/boxes.h (enum FixedDirections):
* src/constraints.c (setup_constraint_info, place_window_if_needed):
add a FIXED_DIRECTION_NONE to the FixedDirections enum to make
code more clear
svn path=/trunk/; revision=3144
2007-04-02 23:41:10 -04:00
/* Pull out the strut info for each side in the hint */
int i ;
for ( i = 0 ; i < 4 ; i + + )
{
MetaStrut * temp ;
int thickness , strut_begin , strut_end ;
thickness = struts [ i ] ;
if ( thickness = = 0 )
continue ;
strut_begin = struts [ 4 + ( i * 2 ) ] ;
strut_end = struts [ 4 + ( i * 2 ) + 1 ] ;
temp = g_new ( MetaStrut , 1 ) ;
2009-03-17 16:53:01 -04:00
temp - > side = 1 < < i ; /* See MetaSide def. Matches nicely, eh? */
Patch from Carlo Wood to fix handling of unidirectional maximization and
2007-04-02 Elijah Newren <newren gmail com>
Patch from Carlo Wood to fix handling of unidirectional
maximization and partial struts. #358311.
* src/constraints.c (constrain_maximization):
determine target size for unidirectionally maximized windows by
determining how far they can be maximized without hitting
orthogonal struts. Avoids weird "empty spaces".
* src/boxes.[ch] (meta_rectangle_expand_to_avoiding_struts):
new function
2007-04-02 Elijah Newren <newren gmail com>
Make the strut lists (stored in workspaces) record both the
rectangle and the side that the strut is on. Lots of code
cleanups relating to struts.
* src/boxes.h (struct MetaStrut):
new struct for struts
* src/window.[ch] (struct MetaStruts, struct MetaWindow,
meta_window_update_struts):
overhaul to make window's struts remember their side as well as
their rectangular location, and just use a list instead of several
copies of near-identical code for left/right/top/bottom (allowing
us to nuke MetaStruts struct as well)
* src/testboxes.c (new_meta_strut, get_strut_list):
* src/workspace.c (ensure_work_areas_validated):
* src/boxes.c (meta_rectangle_get_minimal_spanning_set_for_region,
meta_rectangle_expand_to_avoiding_struts,
get_disjoint_strut_rect_list_in_region, fix_up_edges,
meta_rectangle_find_onscreen_edges,
meta_rectangle_find_nonintersected_xinerama_edges):
modify to handle struts being rectangle + side instead of just rectangle
* src/workspace.c (ensure_work_areas_validated):
simplify strut list creation considerably given MetaWindow change,
modify work_area computations to take advantage of region
computations being done (makes the code shorter as well as more
robust against pathological cases).
* src/util.[ch] (meta_free_gslist_and_elements):
new convenience function
* src/common.h (enum MetaDirection):
* src/edge-resistance.c (movement_towards_edge):
* src/boxes.c (meta_rectangle_edge_aligns,
rectangle_and_edge_intersection, split_edge):
Add more MetaDirection fields for convenience
* src/boxes.h (enum FixedDirections):
* src/constraints.c (setup_constraint_info, place_window_if_needed):
add a FIXED_DIRECTION_NONE to the FixedDirections enum to make
code more clear
svn path=/trunk/; revision=3144
2007-04-02 23:41:10 -04:00
temp - > rect = window - > screen - > rect ;
switch ( temp - > side )
{
case META_SIDE_RIGHT :
temp - > rect . x = BOX_RIGHT ( temp - > rect ) - thickness ;
/* Intentionally fall through without breaking */
case META_SIDE_LEFT :
temp - > rect . width = thickness ;
temp - > rect . y = strut_begin ;
temp - > rect . height = strut_end - strut_begin + 1 ;
break ;
case META_SIDE_BOTTOM :
temp - > rect . y = BOX_BOTTOM ( temp - > rect ) - thickness ;
/* Intentionally fall through without breaking */
case META_SIDE_TOP :
temp - > rect . height = thickness ;
temp - > rect . x = strut_begin ;
temp - > rect . width = strut_end - strut_begin + 1 ;
break ;
default :
g_assert_not_reached ( ) ;
}
new_struts = g_slist_prepend ( new_struts , temp ) ;
}
meta_verbose ( " _NET_WM_STRUT_PARTIAL struts %lu %lu %lu %lu for "
" window %s \n " ,
2008-12-18 07:41:56 -05:00
struts [ 0 ] , struts [ 1 ] , struts [ 2 ] , struts [ 3 ] ,
2003-06-25 23:09:38 -04:00
window - > desc ) ;
}
2002-01-10 01:31:31 -05:00
meta_XFree ( struts ) ;
}
else
{
meta_verbose ( " No _NET_WM_STRUT property for %s \n " ,
window - > desc ) ;
}
Patch from Carlo Wood to fix handling of unidirectional maximization and
2007-04-02 Elijah Newren <newren gmail com>
Patch from Carlo Wood to fix handling of unidirectional
maximization and partial struts. #358311.
* src/constraints.c (constrain_maximization):
determine target size for unidirectionally maximized windows by
determining how far they can be maximized without hitting
orthogonal struts. Avoids weird "empty spaces".
* src/boxes.[ch] (meta_rectangle_expand_to_avoiding_struts):
new function
2007-04-02 Elijah Newren <newren gmail com>
Make the strut lists (stored in workspaces) record both the
rectangle and the side that the strut is on. Lots of code
cleanups relating to struts.
* src/boxes.h (struct MetaStrut):
new struct for struts
* src/window.[ch] (struct MetaStruts, struct MetaWindow,
meta_window_update_struts):
overhaul to make window's struts remember their side as well as
their rectangular location, and just use a list instead of several
copies of near-identical code for left/right/top/bottom (allowing
us to nuke MetaStruts struct as well)
* src/testboxes.c (new_meta_strut, get_strut_list):
* src/workspace.c (ensure_work_areas_validated):
* src/boxes.c (meta_rectangle_get_minimal_spanning_set_for_region,
meta_rectangle_expand_to_avoiding_struts,
get_disjoint_strut_rect_list_in_region, fix_up_edges,
meta_rectangle_find_onscreen_edges,
meta_rectangle_find_nonintersected_xinerama_edges):
modify to handle struts being rectangle + side instead of just rectangle
* src/workspace.c (ensure_work_areas_validated):
simplify strut list creation considerably given MetaWindow change,
modify work_area computations to take advantage of region
computations being done (makes the code shorter as well as more
robust against pathological cases).
* src/util.[ch] (meta_free_gslist_and_elements):
new convenience function
* src/common.h (enum MetaDirection):
* src/edge-resistance.c (movement_towards_edge):
* src/boxes.c (meta_rectangle_edge_aligns,
rectangle_and_edge_intersection, split_edge):
Add more MetaDirection fields for convenience
* src/boxes.h (enum FixedDirections):
* src/constraints.c (setup_constraint_info, place_window_if_needed):
add a FIXED_DIRECTION_NONE to the FixedDirections enum to make
code more clear
svn path=/trunk/; revision=3144
2007-04-02 23:41:10 -04:00
if ( ! new_struts & &
meta_prop_get_cardinal_list ( window - > display ,
window - > xwindow ,
2008-05-02 14:49:01 -04:00
window - > display - > atom__NET_WM_STRUT ,
Patch from Carlo Wood to fix handling of unidirectional maximization and
2007-04-02 Elijah Newren <newren gmail com>
Patch from Carlo Wood to fix handling of unidirectional
maximization and partial struts. #358311.
* src/constraints.c (constrain_maximization):
determine target size for unidirectionally maximized windows by
determining how far they can be maximized without hitting
orthogonal struts. Avoids weird "empty spaces".
* src/boxes.[ch] (meta_rectangle_expand_to_avoiding_struts):
new function
2007-04-02 Elijah Newren <newren gmail com>
Make the strut lists (stored in workspaces) record both the
rectangle and the side that the strut is on. Lots of code
cleanups relating to struts.
* src/boxes.h (struct MetaStrut):
new struct for struts
* src/window.[ch] (struct MetaStruts, struct MetaWindow,
meta_window_update_struts):
overhaul to make window's struts remember their side as well as
their rectangular location, and just use a list instead of several
copies of near-identical code for left/right/top/bottom (allowing
us to nuke MetaStruts struct as well)
* src/testboxes.c (new_meta_strut, get_strut_list):
* src/workspace.c (ensure_work_areas_validated):
* src/boxes.c (meta_rectangle_get_minimal_spanning_set_for_region,
meta_rectangle_expand_to_avoiding_struts,
get_disjoint_strut_rect_list_in_region, fix_up_edges,
meta_rectangle_find_onscreen_edges,
meta_rectangle_find_nonintersected_xinerama_edges):
modify to handle struts being rectangle + side instead of just rectangle
* src/workspace.c (ensure_work_areas_validated):
simplify strut list creation considerably given MetaWindow change,
modify work_area computations to take advantage of region
computations being done (makes the code shorter as well as more
robust against pathological cases).
* src/util.[ch] (meta_free_gslist_and_elements):
new convenience function
* src/common.h (enum MetaDirection):
* src/edge-resistance.c (movement_towards_edge):
* src/boxes.c (meta_rectangle_edge_aligns,
rectangle_and_edge_intersection, split_edge):
Add more MetaDirection fields for convenience
* src/boxes.h (enum FixedDirections):
* src/constraints.c (setup_constraint_info, place_window_if_needed):
add a FIXED_DIRECTION_NONE to the FixedDirections enum to make
code more clear
svn path=/trunk/; revision=3144
2007-04-02 23:41:10 -04:00
& struts , & nitems ) )
2003-06-25 23:09:38 -04:00
{
Patch from Carlo Wood to fix handling of unidirectional maximization and
2007-04-02 Elijah Newren <newren gmail com>
Patch from Carlo Wood to fix handling of unidirectional
maximization and partial struts. #358311.
* src/constraints.c (constrain_maximization):
determine target size for unidirectionally maximized windows by
determining how far they can be maximized without hitting
orthogonal struts. Avoids weird "empty spaces".
* src/boxes.[ch] (meta_rectangle_expand_to_avoiding_struts):
new function
2007-04-02 Elijah Newren <newren gmail com>
Make the strut lists (stored in workspaces) record both the
rectangle and the side that the strut is on. Lots of code
cleanups relating to struts.
* src/boxes.h (struct MetaStrut):
new struct for struts
* src/window.[ch] (struct MetaStruts, struct MetaWindow,
meta_window_update_struts):
overhaul to make window's struts remember their side as well as
their rectangular location, and just use a list instead of several
copies of near-identical code for left/right/top/bottom (allowing
us to nuke MetaStruts struct as well)
* src/testboxes.c (new_meta_strut, get_strut_list):
* src/workspace.c (ensure_work_areas_validated):
* src/boxes.c (meta_rectangle_get_minimal_spanning_set_for_region,
meta_rectangle_expand_to_avoiding_struts,
get_disjoint_strut_rect_list_in_region, fix_up_edges,
meta_rectangle_find_onscreen_edges,
meta_rectangle_find_nonintersected_xinerama_edges):
modify to handle struts being rectangle + side instead of just rectangle
* src/workspace.c (ensure_work_areas_validated):
simplify strut list creation considerably given MetaWindow change,
modify work_area computations to take advantage of region
computations being done (makes the code shorter as well as more
robust against pathological cases).
* src/util.[ch] (meta_free_gslist_and_elements):
new convenience function
* src/common.h (enum MetaDirection):
* src/edge-resistance.c (movement_towards_edge):
* src/boxes.c (meta_rectangle_edge_aligns,
rectangle_and_edge_intersection, split_edge):
Add more MetaDirection fields for convenience
* src/boxes.h (enum FixedDirections):
* src/constraints.c (setup_constraint_info, place_window_if_needed):
add a FIXED_DIRECTION_NONE to the FixedDirections enum to make
code more clear
svn path=/trunk/; revision=3144
2007-04-02 23:41:10 -04:00
if ( nitems ! = 4 )
meta_verbose ( " _NET_WM_STRUT on %s has %d values instead of 4 \n " ,
window - > desc , nitems ) ;
else
2003-06-25 23:09:38 -04:00
{
Patch from Carlo Wood to fix handling of unidirectional maximization and
2007-04-02 Elijah Newren <newren gmail com>
Patch from Carlo Wood to fix handling of unidirectional
maximization and partial struts. #358311.
* src/constraints.c (constrain_maximization):
determine target size for unidirectionally maximized windows by
determining how far they can be maximized without hitting
orthogonal struts. Avoids weird "empty spaces".
* src/boxes.[ch] (meta_rectangle_expand_to_avoiding_struts):
new function
2007-04-02 Elijah Newren <newren gmail com>
Make the strut lists (stored in workspaces) record both the
rectangle and the side that the strut is on. Lots of code
cleanups relating to struts.
* src/boxes.h (struct MetaStrut):
new struct for struts
* src/window.[ch] (struct MetaStruts, struct MetaWindow,
meta_window_update_struts):
overhaul to make window's struts remember their side as well as
their rectangular location, and just use a list instead of several
copies of near-identical code for left/right/top/bottom (allowing
us to nuke MetaStruts struct as well)
* src/testboxes.c (new_meta_strut, get_strut_list):
* src/workspace.c (ensure_work_areas_validated):
* src/boxes.c (meta_rectangle_get_minimal_spanning_set_for_region,
meta_rectangle_expand_to_avoiding_struts,
get_disjoint_strut_rect_list_in_region, fix_up_edges,
meta_rectangle_find_onscreen_edges,
meta_rectangle_find_nonintersected_xinerama_edges):
modify to handle struts being rectangle + side instead of just rectangle
* src/workspace.c (ensure_work_areas_validated):
simplify strut list creation considerably given MetaWindow change,
modify work_area computations to take advantage of region
computations being done (makes the code shorter as well as more
robust against pathological cases).
* src/util.[ch] (meta_free_gslist_and_elements):
new convenience function
* src/common.h (enum MetaDirection):
* src/edge-resistance.c (movement_towards_edge):
* src/boxes.c (meta_rectangle_edge_aligns,
rectangle_and_edge_intersection, split_edge):
Add more MetaDirection fields for convenience
* src/boxes.h (enum FixedDirections):
* src/constraints.c (setup_constraint_info, place_window_if_needed):
add a FIXED_DIRECTION_NONE to the FixedDirections enum to make
code more clear
svn path=/trunk/; revision=3144
2007-04-02 23:41:10 -04:00
/* Pull out the strut info for each side in the hint */
int i ;
for ( i = 0 ; i < 4 ; i + + )
2003-06-25 23:09:38 -04:00
{
Patch from Carlo Wood to fix handling of unidirectional maximization and
2007-04-02 Elijah Newren <newren gmail com>
Patch from Carlo Wood to fix handling of unidirectional
maximization and partial struts. #358311.
* src/constraints.c (constrain_maximization):
determine target size for unidirectionally maximized windows by
determining how far they can be maximized without hitting
orthogonal struts. Avoids weird "empty spaces".
* src/boxes.[ch] (meta_rectangle_expand_to_avoiding_struts):
new function
2007-04-02 Elijah Newren <newren gmail com>
Make the strut lists (stored in workspaces) record both the
rectangle and the side that the strut is on. Lots of code
cleanups relating to struts.
* src/boxes.h (struct MetaStrut):
new struct for struts
* src/window.[ch] (struct MetaStruts, struct MetaWindow,
meta_window_update_struts):
overhaul to make window's struts remember their side as well as
their rectangular location, and just use a list instead of several
copies of near-identical code for left/right/top/bottom (allowing
us to nuke MetaStruts struct as well)
* src/testboxes.c (new_meta_strut, get_strut_list):
* src/workspace.c (ensure_work_areas_validated):
* src/boxes.c (meta_rectangle_get_minimal_spanning_set_for_region,
meta_rectangle_expand_to_avoiding_struts,
get_disjoint_strut_rect_list_in_region, fix_up_edges,
meta_rectangle_find_onscreen_edges,
meta_rectangle_find_nonintersected_xinerama_edges):
modify to handle struts being rectangle + side instead of just rectangle
* src/workspace.c (ensure_work_areas_validated):
simplify strut list creation considerably given MetaWindow change,
modify work_area computations to take advantage of region
computations being done (makes the code shorter as well as more
robust against pathological cases).
* src/util.[ch] (meta_free_gslist_and_elements):
new convenience function
* src/common.h (enum MetaDirection):
* src/edge-resistance.c (movement_towards_edge):
* src/boxes.c (meta_rectangle_edge_aligns,
rectangle_and_edge_intersection, split_edge):
Add more MetaDirection fields for convenience
* src/boxes.h (enum FixedDirections):
* src/constraints.c (setup_constraint_info, place_window_if_needed):
add a FIXED_DIRECTION_NONE to the FixedDirections enum to make
code more clear
svn path=/trunk/; revision=3144
2007-04-02 23:41:10 -04:00
MetaStrut * temp ;
int thickness ;
thickness = struts [ i ] ;
if ( thickness = = 0 )
continue ;
temp = g_new ( MetaStrut , 1 ) ;
temp - > side = 1 < < i ;
temp - > rect = window - > screen - > rect ;
switch ( temp - > side )
{
case META_SIDE_RIGHT :
temp - > rect . x = BOX_RIGHT ( temp - > rect ) - thickness ;
/* Intentionally fall through without breaking */
case META_SIDE_LEFT :
temp - > rect . width = thickness ;
break ;
case META_SIDE_BOTTOM :
temp - > rect . y = BOX_BOTTOM ( temp - > rect ) - thickness ;
/* Intentionally fall through without breaking */
case META_SIDE_TOP :
temp - > rect . height = thickness ;
break ;
default :
g_assert_not_reached ( ) ;
}
new_struts = g_slist_prepend ( new_struts , temp ) ;
2003-06-25 23:09:38 -04:00
}
2008-12-18 07:41:56 -05:00
Patch from Carlo Wood to fix handling of unidirectional maximization and
2007-04-02 Elijah Newren <newren gmail com>
Patch from Carlo Wood to fix handling of unidirectional
maximization and partial struts. #358311.
* src/constraints.c (constrain_maximization):
determine target size for unidirectionally maximized windows by
determining how far they can be maximized without hitting
orthogonal struts. Avoids weird "empty spaces".
* src/boxes.[ch] (meta_rectangle_expand_to_avoiding_struts):
new function
2007-04-02 Elijah Newren <newren gmail com>
Make the strut lists (stored in workspaces) record both the
rectangle and the side that the strut is on. Lots of code
cleanups relating to struts.
* src/boxes.h (struct MetaStrut):
new struct for struts
* src/window.[ch] (struct MetaStruts, struct MetaWindow,
meta_window_update_struts):
overhaul to make window's struts remember their side as well as
their rectangular location, and just use a list instead of several
copies of near-identical code for left/right/top/bottom (allowing
us to nuke MetaStruts struct as well)
* src/testboxes.c (new_meta_strut, get_strut_list):
* src/workspace.c (ensure_work_areas_validated):
* src/boxes.c (meta_rectangle_get_minimal_spanning_set_for_region,
meta_rectangle_expand_to_avoiding_struts,
get_disjoint_strut_rect_list_in_region, fix_up_edges,
meta_rectangle_find_onscreen_edges,
meta_rectangle_find_nonintersected_xinerama_edges):
modify to handle struts being rectangle + side instead of just rectangle
* src/workspace.c (ensure_work_areas_validated):
simplify strut list creation considerably given MetaWindow change,
modify work_area computations to take advantage of region
computations being done (makes the code shorter as well as more
robust against pathological cases).
* src/util.[ch] (meta_free_gslist_and_elements):
new convenience function
* src/common.h (enum MetaDirection):
* src/edge-resistance.c (movement_towards_edge):
* src/boxes.c (meta_rectangle_edge_aligns,
rectangle_and_edge_intersection, split_edge):
Add more MetaDirection fields for convenience
* src/boxes.h (enum FixedDirections):
* src/constraints.c (setup_constraint_info, place_window_if_needed):
add a FIXED_DIRECTION_NONE to the FixedDirections enum to make
code more clear
svn path=/trunk/; revision=3144
2007-04-02 23:41:10 -04:00
meta_verbose ( " _NET_WM_STRUT struts %lu %lu %lu %lu for window %s \n " ,
2008-12-18 07:41:56 -05:00
struts [ 0 ] , struts [ 1 ] , struts [ 2 ] , struts [ 3 ] ,
2003-06-25 23:09:38 -04:00
window - > desc ) ;
}
Patch from Carlo Wood to fix handling of unidirectional maximization and
2007-04-02 Elijah Newren <newren gmail com>
Patch from Carlo Wood to fix handling of unidirectional
maximization and partial struts. #358311.
* src/constraints.c (constrain_maximization):
determine target size for unidirectionally maximized windows by
determining how far they can be maximized without hitting
orthogonal struts. Avoids weird "empty spaces".
* src/boxes.[ch] (meta_rectangle_expand_to_avoiding_struts):
new function
2007-04-02 Elijah Newren <newren gmail com>
Make the strut lists (stored in workspaces) record both the
rectangle and the side that the strut is on. Lots of code
cleanups relating to struts.
* src/boxes.h (struct MetaStrut):
new struct for struts
* src/window.[ch] (struct MetaStruts, struct MetaWindow,
meta_window_update_struts):
overhaul to make window's struts remember their side as well as
their rectangular location, and just use a list instead of several
copies of near-identical code for left/right/top/bottom (allowing
us to nuke MetaStruts struct as well)
* src/testboxes.c (new_meta_strut, get_strut_list):
* src/workspace.c (ensure_work_areas_validated):
* src/boxes.c (meta_rectangle_get_minimal_spanning_set_for_region,
meta_rectangle_expand_to_avoiding_struts,
get_disjoint_strut_rect_list_in_region, fix_up_edges,
meta_rectangle_find_onscreen_edges,
meta_rectangle_find_nonintersected_xinerama_edges):
modify to handle struts being rectangle + side instead of just rectangle
* src/workspace.c (ensure_work_areas_validated):
simplify strut list creation considerably given MetaWindow change,
modify work_area computations to take advantage of region
computations being done (makes the code shorter as well as more
robust against pathological cases).
* src/util.[ch] (meta_free_gslist_and_elements):
new convenience function
* src/common.h (enum MetaDirection):
* src/edge-resistance.c (movement_towards_edge):
* src/boxes.c (meta_rectangle_edge_aligns,
rectangle_and_edge_intersection, split_edge):
Add more MetaDirection fields for convenience
* src/boxes.h (enum FixedDirections):
* src/constraints.c (setup_constraint_info, place_window_if_needed):
add a FIXED_DIRECTION_NONE to the FixedDirections enum to make
code more clear
svn path=/trunk/; revision=3144
2007-04-02 23:41:10 -04:00
meta_XFree ( struts ) ;
}
else if ( ! new_struts )
{
meta_verbose ( " No _NET_WM_STRUT property for %s \n " ,
window - > desc ) ;
2003-06-25 23:09:38 -04:00
}
2008-12-18 07:41:56 -05:00
Patch from Carlo Wood to fix handling of unidirectional maximization and
2007-04-02 Elijah Newren <newren gmail com>
Patch from Carlo Wood to fix handling of unidirectional
maximization and partial struts. #358311.
* src/constraints.c (constrain_maximization):
determine target size for unidirectionally maximized windows by
determining how far they can be maximized without hitting
orthogonal struts. Avoids weird "empty spaces".
* src/boxes.[ch] (meta_rectangle_expand_to_avoiding_struts):
new function
2007-04-02 Elijah Newren <newren gmail com>
Make the strut lists (stored in workspaces) record both the
rectangle and the side that the strut is on. Lots of code
cleanups relating to struts.
* src/boxes.h (struct MetaStrut):
new struct for struts
* src/window.[ch] (struct MetaStruts, struct MetaWindow,
meta_window_update_struts):
overhaul to make window's struts remember their side as well as
their rectangular location, and just use a list instead of several
copies of near-identical code for left/right/top/bottom (allowing
us to nuke MetaStruts struct as well)
* src/testboxes.c (new_meta_strut, get_strut_list):
* src/workspace.c (ensure_work_areas_validated):
* src/boxes.c (meta_rectangle_get_minimal_spanning_set_for_region,
meta_rectangle_expand_to_avoiding_struts,
get_disjoint_strut_rect_list_in_region, fix_up_edges,
meta_rectangle_find_onscreen_edges,
meta_rectangle_find_nonintersected_xinerama_edges):
modify to handle struts being rectangle + side instead of just rectangle
* src/workspace.c (ensure_work_areas_validated):
simplify strut list creation considerably given MetaWindow change,
modify work_area computations to take advantage of region
computations being done (makes the code shorter as well as more
robust against pathological cases).
* src/util.[ch] (meta_free_gslist_and_elements):
new convenience function
* src/common.h (enum MetaDirection):
* src/edge-resistance.c (movement_towards_edge):
* src/boxes.c (meta_rectangle_edge_aligns,
rectangle_and_edge_intersection, split_edge):
Add more MetaDirection fields for convenience
* src/boxes.h (enum FixedDirections):
* src/constraints.c (setup_constraint_info, place_window_if_needed):
add a FIXED_DIRECTION_NONE to the FixedDirections enum to make
code more clear
svn path=/trunk/; revision=3144
2007-04-02 23:41:10 -04:00
/* Determine whether old_struts and new_struts are the same */
old_iter = old_struts ;
new_iter = new_struts ;
while ( old_iter & & new_iter )
{
MetaStrut * old_strut = ( MetaStrut * ) old_iter - > data ;
MetaStrut * new_strut = ( MetaStrut * ) new_iter - > data ;
if ( old_strut - > side ! = new_strut - > side | |
! meta_rectangle_equal ( & old_strut - > rect , & new_strut - > rect ) )
break ;
old_iter = old_iter - > next ;
new_iter = new_iter - > next ;
}
changed = ( old_iter ! = NULL | | new_iter ! = NULL ) ;
/* Update appropriately */
meta_free_gslist_and_elements ( old_struts ) ;
window - > struts = new_struts ;
if ( changed )
{
2002-06-08 01:13:55 -04:00
meta_topic ( META_DEBUG_WORKAREA ,
" Invalidating work areas of window %s due to struts update \n " ,
window - > desc ) ;
invalidate_work_areas ( window ) ;
}
else
{
meta_topic ( META_DEBUG_WORKAREA ,
" Struts on %s were unchanged \n " , window - > desc ) ;
}
2002-01-10 01:31:31 -05:00
}
2014-03-13 16:52:31 -04:00
static void
2013-09-10 09:53:31 -04:00
meta_window_type_changed ( MetaWindow * window )
{
gboolean old_decorated = window - > decorated ;
GObject * object = G_OBJECT ( window ) ;
2008-12-18 07:41:56 -05:00
2013-09-10 09:53:31 -04:00
window - > attached = meta_window_should_attach_to_parent ( window ) ;
2014-02-01 18:58:58 -05:00
meta_window_recalc_features ( window ) ;
2008-12-18 07:41:56 -05:00
2013-09-10 09:53:31 -04:00
if ( ! window - > override_redirect )
set_net_wm_state ( window ) ;
2008-12-18 07:41:56 -05:00
2013-09-10 09:53:31 -04:00
/* Update frame */
if ( window - > decorated )
meta_window_ensure_frame ( window ) ;
else
meta_window_destroy_frame ( window ) ;
2003-02-27 17:44:36 -05:00
2013-09-10 09:53:31 -04:00
/* update stacking constraints */
meta_window_update_layer ( window ) ;
meta_window_grab_keys ( window ) ;
2009-02-11 07:16:41 -05:00
2013-09-10 09:53:31 -04:00
g_object_freeze_notify ( object ) ;
2009-08-04 03:06:30 -04:00
2013-09-10 09:53:31 -04:00
if ( old_decorated ! = window - > decorated )
g_object_notify ( object , " decorated " ) ;
2009-08-04 03:06:30 -04:00
2013-09-10 09:53:31 -04:00
g_object_notify ( object , " window-type " ) ;
2009-08-04 03:06:30 -04:00
2013-09-10 09:53:31 -04:00
g_object_thaw_notify ( object ) ;
2001-06-09 23:17:15 -04:00
}
2014-03-13 16:52:31 -04:00
void
meta_window_set_type ( MetaWindow * window ,
MetaWindowType type )
{
if ( window - > type = = type )
return ;
window - > type = type ;
meta_window_type_changed ( window ) ;
}
2013-11-15 17:37:50 -05:00
void
meta_window_frame_size_changed ( MetaWindow * window )
{
if ( window - > frame )
meta_frame_clear_cached_borders ( window - > frame ) ;
}
2002-10-21 17:44:35 -04:00
static void
2002-08-01 19:52:47 -04:00
set_allowed_actions_hint ( MetaWindow * window )
{
2007-04-11 13:15:54 -04:00
# define MAX_N_ACTIONS 12
2002-08-01 19:52:47 -04:00
unsigned long data [ MAX_N_ACTIONS ] ;
int i ;
i = 0 ;
if ( window - > has_move_func )
{
2008-05-02 14:49:01 -04:00
data [ i ] = window - > display - > atom__NET_WM_ACTION_MOVE ;
2002-08-01 19:52:47 -04:00
+ + i ;
}
if ( window - > has_resize_func )
{
2008-05-02 14:49:01 -04:00
data [ i ] = window - > display - > atom__NET_WM_ACTION_RESIZE ;
2003-11-17 11:46:01 -05:00
+ + i ;
2007-07-22 20:46:03 -04:00
}
if ( window - > has_fullscreen_func )
{
2008-05-02 14:49:01 -04:00
data [ i ] = window - > display - > atom__NET_WM_ACTION_FULLSCREEN ;
2003-11-16 19:06:58 -05:00
+ + i ;
}
if ( window - > has_minimize_func )
{
2008-05-02 14:49:01 -04:00
data [ i ] = window - > display - > atom__NET_WM_ACTION_MINIMIZE ;
2002-08-01 19:52:47 -04:00
+ + i ;
}
if ( window - > has_shade_func )
{
2008-05-02 14:49:01 -04:00
data [ i ] = window - > display - > atom__NET_WM_ACTION_SHADE ;
2002-08-01 19:52:47 -04:00
+ + i ;
}
Comprehensively rename to Mutter
Code:
All references in the code not related to themes, keybindings, or
GConf were changed from 'metacity' to 'mutter'. This includes, among other
things, strings, comments, the atoms used in the message protocol, and
the envvars used for debugging. The GConf schema file was reduced to
the 3 settings new to mutter.
The overall version was brought up to 2.27 to match current gnome.
Structure:
All files named '*metacity*' were renamed '*mutter*' with appropriate
changes in the automake system. Files removed are
doc/creating_themes, src/themes, doc/metacity-theme.dtd,
metacity.doap. These files will eventually end up in an external
gnome-wm-data module.
Installation location:
On the filesystem the mutter-plugindir was change from
$(libdir)/metacity/plugins/clutter to just $(libdir)/mutter/plugins.
The mutter-plugins.pc.in reflects these changes.
Note:
mutter.desktop.in and mutter-wm.desktop both continue to have
X-GNOME-WMSettingsModule=metacity set. This allows
gnome-control-center to continue using libmetacity.so for
configuration. This is fine since most the general keybindings and wm
settings are being read from /apps/metacity/* in gconf.
2009-06-10 06:29:20 -04:00
/* sticky according to EWMH is different from mutter's sticky;
* mutter doesn ' t support EWMH sticky
2003-11-16 19:06:58 -05:00
*/
if ( window - > has_maximize_func )
2002-08-01 19:52:47 -04:00
{
2008-05-02 14:49:01 -04:00
data [ i ] = window - > display - > atom__NET_WM_ACTION_MAXIMIZE_HORZ ;
2003-11-16 19:06:58 -05:00
+ + i ;
2008-05-02 14:49:01 -04:00
data [ i ] = window - > display - > atom__NET_WM_ACTION_MAXIMIZE_VERT ;
2002-08-01 19:52:47 -04:00
+ + i ;
}
/* We always allow this */
2008-05-02 14:49:01 -04:00
data [ i ] = window - > display - > atom__NET_WM_ACTION_CHANGE_DESKTOP ;
2002-08-01 19:52:47 -04:00
+ + i ;
2003-11-16 19:06:58 -05:00
if ( window - > has_close_func )
{
2008-05-02 14:49:01 -04:00
data [ i ] = window - > display - > atom__NET_WM_ACTION_CLOSE ;
2003-11-16 19:06:58 -05:00
+ + i ;
}
2002-08-01 19:52:47 -04:00
2007-04-11 13:15:54 -04:00
/* I guess we always allow above/below operations */
2008-05-02 14:49:01 -04:00
data [ i ] = window - > display - > atom__NET_WM_ACTION_ABOVE ;
2007-04-11 13:15:54 -04:00
+ + i ;
2008-05-02 14:49:01 -04:00
data [ i ] = window - > display - > atom__NET_WM_ACTION_BELOW ;
2007-04-11 13:15:54 -04:00
+ + i ;
2002-08-01 19:52:47 -04:00
g_assert ( i < = MAX_N_ACTIONS ) ;
meta_verbose ( " Setting _NET_WM_ALLOWED_ACTIONS with %d atoms \n " , i ) ;
2008-12-18 07:41:56 -05:00
2002-08-01 19:52:47 -04:00
meta_error_trap_push ( window - > display ) ;
XChangeProperty ( window - > display - > xdisplay , window - > xwindow ,
2008-05-02 14:49:01 -04:00
window - > display - > atom__NET_WM_ALLOWED_ACTIONS ,
2002-08-01 19:52:47 -04:00
XA_ATOM ,
32 , PropModeReplace , ( guchar * ) data , i ) ;
2010-10-25 14:44:30 -04:00
meta_error_trap_pop ( window - > display ) ;
2002-08-01 19:52:47 -04:00
# undef MAX_N_ACTIONS
}
2014-03-18 11:20:22 -04:00
static void
meta_window_get_default_skip_hints ( MetaWindow * window ,
gboolean * skip_taskbar_out ,
gboolean * skip_pager_out )
{
META_WINDOW_GET_CLASS ( window ) - > get_default_skip_hints ( window , skip_taskbar_out , skip_pager_out ) ;
}
2014-03-18 11:10:22 -04:00
static void
meta_window_recalc_skip_features ( MetaWindow * window )
{
switch ( window - > type )
{
/* Force skip taskbar/pager on these window types */
case META_WINDOW_DESKTOP :
case META_WINDOW_DOCK :
case META_WINDOW_TOOLBAR :
case META_WINDOW_MENU :
case META_WINDOW_UTILITY :
case META_WINDOW_SPLASHSCREEN :
case META_WINDOW_DROPDOWN_MENU :
case META_WINDOW_POPUP_MENU :
case META_WINDOW_TOOLTIP :
case META_WINDOW_NOTIFICATION :
case META_WINDOW_COMBO :
case META_WINDOW_DND :
case META_WINDOW_OVERRIDE_OTHER :
window - > skip_taskbar = TRUE ;
window - > skip_pager = TRUE ;
break ;
case META_WINDOW_DIALOG :
case META_WINDOW_MODAL_DIALOG :
/* only skip taskbar if we have a real transient parent
( and ignore the application hints ) */
if ( window - > transient_for ! = NULL )
window - > skip_taskbar = TRUE ;
else
window - > skip_taskbar = FALSE ;
break ;
case META_WINDOW_NORMAL :
2014-03-18 11:20:22 -04:00
{
gboolean skip_taskbar_hint , skip_pager_hint ;
meta_window_get_default_skip_hints ( window , & skip_taskbar_hint , & skip_pager_hint ) ;
window - > skip_taskbar = skip_taskbar_hint ;
window - > skip_pager = skip_pager_hint ;
}
2014-03-18 11:10:22 -04:00
break ;
}
}
2003-02-27 20:24:44 -05:00
void
meta_window_recalc_features ( MetaWindow * window )
2001-06-23 14:30:27 -04:00
{
2002-08-01 19:52:47 -04:00
gboolean old_has_close_func ;
gboolean old_has_minimize_func ;
gboolean old_has_move_func ;
gboolean old_has_resize_func ;
gboolean old_has_shade_func ;
gboolean old_always_sticky ;
2014-01-30 16:39:54 -05:00
gboolean old_skip_taskbar ;
2002-08-01 19:52:47 -04:00
old_has_close_func = window - > has_close_func ;
old_has_minimize_func = window - > has_minimize_func ;
old_has_move_func = window - > has_move_func ;
old_has_resize_func = window - > has_resize_func ;
old_has_shade_func = window - > has_shade_func ;
old_always_sticky = window - > always_sticky ;
2014-01-30 16:39:54 -05:00
old_skip_taskbar = window - > skip_taskbar ;
2002-08-01 19:52:47 -04:00
2001-06-23 14:30:27 -04:00
/* Use MWM hints initially */
2013-09-10 09:53:31 -04:00
if ( window - > client_type = = META_WINDOW_CLIENT_TYPE_X11 )
window - > decorated = window - > mwm_decorated ;
else
window - > decorated = FALSE ;
2002-06-21 23:23:02 -04:00
window - > border_only = window - > mwm_border_only ;
2001-06-23 14:30:27 -04:00
window - > has_close_func = window - > mwm_has_close_func ;
window - > has_minimize_func = window - > mwm_has_minimize_func ;
window - > has_maximize_func = window - > mwm_has_maximize_func ;
window - > has_move_func = window - > mwm_has_move_func ;
2008-12-18 07:41:56 -05:00
window - > has_resize_func = TRUE ;
2002-08-24 16:54:23 -04:00
/* If min_size == max_size, then don't allow resize */
if ( window - > size_hints . min_width = = window - > size_hints . max_width & &
window - > size_hints . min_height = = window - > size_hints . max_height )
window - > has_resize_func = FALSE ;
else if ( ! window - > mwm_has_resize_func )
{
/* We ignore mwm_has_resize_func because WM_NORMAL_HINTS is the
* authoritative source for that info . Some apps such as mplayer or
* xine disable resize via MWM but not WM_NORMAL_HINTS , but that
* leads to e . g . us not fullscreening their windows . Apps that set
* MWM but not WM_NORMAL_HINTS are basically broken . We complain
* about these apps but make them work .
*/
2008-12-18 07:41:56 -05:00
2013-09-11 04:37:36 -04:00
meta_warning ( " Window %s sets an MWM hint indicating it isn't resizable, but sets min size %d x %d and max size %d x %d; this doesn't make much sense. \n " ,
2002-08-24 16:54:23 -04:00
window - > desc ,
window - > size_hints . min_width ,
window - > size_hints . min_height ,
window - > size_hints . max_width ,
window - > size_hints . max_height ) ;
}
2001-06-23 14:30:27 -04:00
window - > has_shade_func = TRUE ;
2002-02-23 17:58:02 -05:00
window - > has_fullscreen_func = TRUE ;
2002-05-26 11:54:38 -04:00
window - > always_sticky = FALSE ;
2008-12-18 07:41:56 -05:00
2001-06-23 14:30:27 -04:00
/* Semantic category overrides the MWM hints */
2002-04-30 23:23:46 -04:00
if ( window - > type = = META_WINDOW_TOOLBAR )
window - > decorated = FALSE ;
2002-05-26 11:54:38 -04:00
if ( window - > type = = META_WINDOW_DESKTOP | |
2008-11-23 14:28:40 -05:00
window - > type = = META_WINDOW_DOCK | |
window - > override_redirect )
2002-05-26 11:54:38 -04:00
window - > always_sticky = TRUE ;
2008-12-18 07:41:56 -05:00
2011-10-17 11:42:35 -04:00
if ( window - > override_redirect | |
meta_window_get_frame_type ( window ) = = META_FRAME_TYPE_LAST )
2001-06-23 14:30:27 -04:00
{
2001-07-30 23:59:55 -04:00
window - > decorated = FALSE ;
2001-06-23 14:30:27 -04:00
window - > has_close_func = FALSE ;
window - > has_shade_func = FALSE ;
2002-04-30 23:23:46 -04:00
/* FIXME this keeps panels and things from using
* NET_WM_MOVERESIZE ; the problem is that some
* panels ( edge panels ) have fixed possible locations ,
* and others ( " floating panels " ) do not .
*
* Perhaps we should require edge panels to explicitly
* disable movement ?
*/
2001-06-23 14:30:27 -04:00
window - > has_move_func = FALSE ;
window - > has_resize_func = FALSE ;
}
2008-12-18 07:41:56 -05:00
2001-06-23 14:30:27 -04:00
if ( window - > type ! = META_WINDOW_NORMAL )
{
window - > has_minimize_func = FALSE ;
window - > has_maximize_func = FALSE ;
2002-02-23 17:58:02 -05:00
window - > has_fullscreen_func = FALSE ;
2001-06-23 14:30:27 -04:00
}
2001-08-03 00:20:27 -04:00
if ( ! window - > has_resize_func )
2008-12-18 07:41:56 -05:00
{
2002-02-23 17:58:02 -05:00
window - > has_maximize_func = FALSE ;
2008-12-18 07:41:56 -05:00
2002-08-24 16:54:23 -04:00
/* don't allow fullscreen if we can't resize, unless the size
* is entire screen size ( kind of broken , because we
2009-08-31 09:52:16 -04:00
* actually fullscreen to monitor size not screen size )
2002-08-24 16:54:23 -04:00
*/
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
if ( window - > size_hints . min_width = = window - > screen - > rect . width & &
2006-08-04 18:33:20 -04:00
window - > size_hints . min_height = = window - > screen - > rect . height )
2002-08-15 01:14:49 -04:00
; /* leave fullscreen available */
else
window - > has_fullscreen_func = FALSE ;
2002-09-22 21:43:20 -04:00
}
2002-08-24 16:54:23 -04:00
2002-09-22 21:43:20 -04:00
/* We leave fullscreen windows decorated, just push the frame outside
* the screen . This avoids flickering to unparent them .
*
* Note that setting has_resize_func = FALSE here must come after the
* above code that may disable fullscreen , because if the window
* is not resizable purely due to fullscreen , we don ' t want to
* disable fullscreen mode .
*/
if ( window - > fullscreen )
{
window - > has_shade_func = FALSE ;
window - > has_move_func = FALSE ;
window - > has_resize_func = FALSE ;
window - > has_maximize_func = FALSE ;
2002-02-23 17:58:02 -05:00
}
2001-08-03 00:20:27 -04:00
2012-03-12 23:18:03 -04:00
if ( window - > has_maximize_func )
{
2013-11-15 16:11:20 -05:00
MetaRectangle work_area , client_rect ;
2012-03-12 23:18:03 -04:00
meta_window_get_work_area_current_monitor ( window , & work_area ) ;
2013-11-15 16:11:20 -05:00
meta_window_frame_rect_to_client_rect ( window , & work_area , & client_rect ) ;
2012-03-12 23:18:03 -04:00
2013-11-15 16:11:20 -05:00
if ( window - > size_hints . min_width > = client_rect . width | |
window - > size_hints . min_height > = client_rect . height )
2012-03-12 23:18:03 -04:00
window - > has_maximize_func = FALSE ;
}
2002-09-22 21:43:20 -04:00
meta_topic ( META_DEBUG_WINDOW_OPS ,
" Window %s fullscreen = %d not resizable, maximizable = %d fullscreenable = %d min size %dx%d max size %dx%d \n " ,
window - > desc ,
window - > fullscreen ,
window - > has_maximize_func , window - > has_fullscreen_func ,
window - > size_hints . min_width ,
window - > size_hints . min_height ,
window - > size_hints . max_width ,
window - > size_hints . max_height ) ;
2008-12-18 07:41:56 -05:00
2001-08-03 00:20:27 -04:00
/* no shading if not decorated */
2002-07-24 22:45:38 -04:00
if ( ! window - > decorated | | window - > border_only )
2001-08-03 00:20:27 -04:00
window - > has_shade_func = FALSE ;
2002-07-06 15:05:37 -04:00
2014-03-18 11:10:22 -04:00
meta_window_recalc_skip_features ( window ) ;
2002-08-01 19:52:47 -04:00
2002-09-22 21:43:20 -04:00
meta_topic ( META_DEBUG_WINDOW_OPS ,
" Window %s decorated = %d border_only = %d has_close = %d has_minimize = %d has_maximize = %d has_move = %d has_shade = %d skip_taskbar = %d skip_pager = %d \n " ,
window - > desc ,
window - > decorated ,
window - > border_only ,
window - > has_close_func ,
window - > has_minimize_func ,
window - > has_maximize_func ,
window - > has_move_func ,
window - > has_shade_func ,
window - > skip_taskbar ,
window - > skip_pager ) ;
2008-12-18 07:41:56 -05:00
2014-01-30 16:39:54 -05:00
if ( old_skip_taskbar ! = window - > skip_taskbar )
g_object_notify ( G_OBJECT ( window ) , " skip-taskbar " ) ;
2002-08-01 19:52:47 -04:00
/* FIXME:
2014-02-01 18:58:58 -05:00
* Lame workaround for recalc_features being used overzealously .
* The fix is to only recalc_features when something has
* actually changed .
2002-08-01 19:52:47 -04:00
*/
2007-04-10 22:37:30 -04:00
if ( window - > constructing | |
old_has_close_func ! = window - > has_close_func | |
2002-08-01 19:52:47 -04:00
old_has_minimize_func ! = window - > has_minimize_func | |
old_has_move_func ! = window - > has_move_func | |
old_has_resize_func ! = window - > has_resize_func | |
old_has_shade_func ! = window - > has_shade_func | |
old_always_sticky ! = window - > always_sticky )
set_allowed_actions_hint ( window ) ;
2008-12-18 07:41:56 -05:00
2011-07-02 07:08:17 -04:00
if ( window - > has_resize_func ! = old_has_resize_func )
g_object_notify ( G_OBJECT ( window ) , " resizeable " ) ;
2013-11-15 17:37:50 -05:00
meta_window_frame_size_changed ( window ) ;
2001-06-23 14:30:27 -04:00
/* FIXME perhaps should ensure if we don't have a shade func,
* we aren ' t shaded , etc .
*/
}
2001-06-23 01:49:35 -04:00
static void
menu_callback ( MetaWindowMenu * menu ,
Display * xdisplay ,
Window client_xwindow ,
Fix issues on 64-bit machines with timestamps by using guint32 (like gtk+
2006-09-13 Elijah Newren <newren gmail com>
* src/common.h (MetaWindowMenuFunc):
* src/core.[ch] (meta_core_user_lower_and_unfocus,
meta_core_user_focus, meta_core_show_window_menu,
meta_core_begin_grab_op, meta_core_end_grab_op):
* src/delete.c (delete_ping_reply_func, delete_ping_timeout_func,
meta_window_delete):
* src/display.[ch] (struct MetaDisplay, struct MetaPingData,
sanity_check_timestamps, meta_display_open, event_callback,
meta_spew_event, meta_display_set_grab_op_cursor,
meta_display_begin_grab_op, meta_display_end_grab_op,
meta_display_ping_timeout, meta_display_ping_window,
process_pong_message, timestamp_too_old,
meta_display_set_input_focus_window):
* src/keybindings.[ch] (grab_keyboard, ungrab_keyboard,
meta_screen_grab_all_keys, meta_window_grab_all_keys,
meta_window_ungrab_all_keys, error_on_generic_command,
error_on_command, error_on_terminal_command):
* src/metacity-dialog.c (on_realize, warn_about_no_sm_support,
error_about_command, main):
* src/screen.[ch] (struct _MetaScreen, meta_screen_new,
meta_screen_show_desktop, meta_screen_apply_startup_properties):
* src/session.c (warn_about_lame_clients_and_finish_interact):
* src/window.[ch] (struct _MetaWindow,
intervening_user_event_occurred, window_activate,
meta_window_delete, meta_window_focus,
meta_window_send_icccm_message, meta_window_client_message,
menu_callback, meta_window_show_menu, struct EventScannerData,
check_use_this_motion_notify, meta_window_begin_grab_op,
meta_window_set_user_time):
* src/workspace.[ch] (focus_ancestor_or_mru_window,
meta_workspace_activate_with_focus, meta_workspace_activate,
meta_workspace_focus_default_window,
focus_ancestor_or_mru_window):
Fix issues on 64-bit machines with timestamps by using guint32
(like gtk+ does) instead of Time. #348305
2006-09-13 12:32:33 -04:00
guint32 timestamp ,
2001-06-23 01:49:35 -04:00
MetaMenuOp op ,
int workspace_index ,
gpointer data )
{
MetaDisplay * display ;
MetaWindow * window ;
2003-12-21 01:31:40 -05:00
MetaWorkspace * workspace ;
2008-12-18 07:41:56 -05:00
2001-06-23 01:49:35 -04:00
display = meta_display_for_x_display ( xdisplay ) ;
window = meta_display_lookup_x_window ( display , client_xwindow ) ;
2003-12-21 01:31:40 -05:00
workspace = NULL ;
2008-12-18 07:41:56 -05:00
2001-06-23 01:49:35 -04:00
if ( window ! = NULL ) /* window can be NULL */
{
2006-01-20 17:03:56 -05:00
meta_verbose ( " Menu op %u on %s \n " , op , window - > desc ) ;
2008-12-18 07:41:56 -05:00
2001-06-23 01:49:35 -04:00
switch ( op )
{
2010-03-31 16:34:02 -04:00
case META_MENU_OP_NONE :
/* nothing */
break ;
2001-06-23 01:49:35 -04:00
case META_MENU_OP_DELETE :
2004-06-21 12:47:27 -04:00
meta_window_delete ( window , timestamp ) ;
2001-06-23 01:49:35 -04:00
break ;
case META_MENU_OP_MINIMIZE :
meta_window_minimize ( window ) ;
break ;
case META_MENU_OP_UNMAXIMIZE :
2014-03-13 18:32:20 -04:00
meta_window_unmaximize ( window , META_MAXIMIZE_BOTH ) ;
2001-06-23 01:49:35 -04:00
break ;
2008-12-18 07:41:56 -05:00
2001-06-23 01:49:35 -04:00
case META_MENU_OP_MAXIMIZE :
2014-03-13 18:32:20 -04:00
meta_window_maximize ( window , META_MAXIMIZE_BOTH ) ;
2001-06-23 01:49:35 -04:00
break ;
case META_MENU_OP_UNSHADE :
Partial audit to fix timestamp usage. One step towards fixing #355180; see
2006-09-18 Elijah Newren <newren gmail com>
Partial audit to fix timestamp usage. One step towards fixing
#355180; see important comments in that bug.
* src/core.[ch] (meta_core_unshade, meta_core_shade):
* src/delete.c (meta_window_present_delete_dialog,
delete_ping_timeout_func):
* src/display.[ch] (meta_display_open, meta_display_close,
event_callback, meta_display_begin_grab_op,
process_selection_clear, meta_display_unmanage_screen,
meta_display_unmanage_windows_for_screen):
* src/frames.c (meta_frames_button_press_event):
* src/keybindings.c (handle_toggle_shade):
* src/main.c (main):
* src/screen.[ch] (update_num_workspaces, meta_screen_new,
meta_screen_free, prefs_changed_callback):
* src/window.[ch] (meta_window_free, finish_minimize,
implement_showing, meta_window_show, meta_window_maximize,
meta_window_make_fullscreen_internal,
meta_window_unmake_fullscreen, meta_window_shade,
meta_window_unshade, window_activate, send_sync_request,
meta_window_client_message, menu_callback,
meta_window_update_keyboard_resize):
Remove usage of CurrentTime, meta_display_get_current_time() and
meta_display_get_current_time_roundtrip() where possible, or
document why it isn't possible, or at very least add a FIXME with
some explanation of my laziness and what needs to be done.
2006-09-18 13:27:24 -04:00
meta_window_unshade ( window , timestamp ) ;
2001-06-23 01:49:35 -04:00
break ;
2008-12-18 07:41:56 -05:00
2001-06-23 01:49:35 -04:00
case META_MENU_OP_SHADE :
Partial audit to fix timestamp usage. One step towards fixing #355180; see
2006-09-18 Elijah Newren <newren gmail com>
Partial audit to fix timestamp usage. One step towards fixing
#355180; see important comments in that bug.
* src/core.[ch] (meta_core_unshade, meta_core_shade):
* src/delete.c (meta_window_present_delete_dialog,
delete_ping_timeout_func):
* src/display.[ch] (meta_display_open, meta_display_close,
event_callback, meta_display_begin_grab_op,
process_selection_clear, meta_display_unmanage_screen,
meta_display_unmanage_windows_for_screen):
* src/frames.c (meta_frames_button_press_event):
* src/keybindings.c (handle_toggle_shade):
* src/main.c (main):
* src/screen.[ch] (update_num_workspaces, meta_screen_new,
meta_screen_free, prefs_changed_callback):
* src/window.[ch] (meta_window_free, finish_minimize,
implement_showing, meta_window_show, meta_window_maximize,
meta_window_make_fullscreen_internal,
meta_window_unmake_fullscreen, meta_window_shade,
meta_window_unshade, window_activate, send_sync_request,
meta_window_client_message, menu_callback,
meta_window_update_keyboard_resize):
Remove usage of CurrentTime, meta_display_get_current_time() and
meta_display_get_current_time_roundtrip() where possible, or
document why it isn't possible, or at very least add a FIXME with
some explanation of my laziness and what needs to be done.
2006-09-18 13:27:24 -04:00
meta_window_shade ( window , timestamp ) ;
2001-06-23 01:49:35 -04:00
break ;
2008-12-18 07:41:56 -05:00
2003-12-21 01:31:40 -05:00
case META_MENU_OP_MOVE_LEFT :
workspace = meta_workspace_get_neighbor ( window - > screen - > active_workspace ,
META_MOTION_LEFT ) ;
break ;
2001-06-23 01:49:35 -04:00
2003-12-21 01:31:40 -05:00
case META_MENU_OP_MOVE_RIGHT :
workspace = meta_workspace_get_neighbor ( window - > screen - > active_workspace ,
META_MOTION_RIGHT ) ;
break ;
2001-06-23 01:49:35 -04:00
2003-12-21 01:31:40 -05:00
case META_MENU_OP_MOVE_UP :
workspace = meta_workspace_get_neighbor ( window - > screen - > active_workspace ,
META_MOTION_UP ) ;
break ;
case META_MENU_OP_MOVE_DOWN :
workspace = meta_workspace_get_neighbor ( window - > screen - > active_workspace ,
META_MOTION_DOWN ) ;
break ;
case META_MENU_OP_WORKSPACES :
workspace = meta_screen_get_workspace_by_index ( window - > screen ,
workspace_index ) ;
2001-06-23 01:49:35 -04:00
break ;
case META_MENU_OP_STICK :
meta_window_stick ( window ) ;
break ;
case META_MENU_OP_UNSTICK :
meta_window_unstick ( window ) ;
break ;
2003-12-21 01:31:40 -05:00
case META_MENU_OP_ABOVE :
2007-04-16 00:48:14 -04:00
case META_MENU_OP_UNABOVE :
2006-12-27 10:51:49 -05:00
if ( window - > wm_state_above = = FALSE )
meta_window_make_above ( window ) ;
else
meta_window_unmake_above ( window ) ;
2003-12-21 01:31:40 -05:00
break ;
2001-07-11 02:22:00 -04:00
case META_MENU_OP_MOVE :
Patch to provide extra cues to the user when using window menu move and
2002-08-08 Craig Black <blackc@speakeasy.net>
Patch to provide extra cues to the user when using
window menu move and resize, #85724.
* src/common.h: add new cursors
* src/display.c: (grab_op_is_mouse)
(meta_display_create_x_cursor), (xcursor_for_op),
(meta_display_set_grab_op_cursor),
(meta_display_begin_grab_op):
The keyboard move and resize grab ops now also use the mouse.
Allow the grab cursor to be changed during the grab op.
Hold onto the initial grab position in case of reset.
* src/display.h: save the initial grab position
* src/keybindings.c: (process_keyboard_move_grab),
(process_keyboard_resize_grab), (handle_begin_move),
(handle_begin_resize):
The keyboard move and resize grab ops now also use the mouse.
* src/window.c: (meta_window_client_message), (menu_callback),
(update_move), (update_resize),
(meta_window_handle_mouse_grab_op_event), (warp_pointer),
(meta_window_warp_pointer), (meta_window_begin_grab_op),
(meta_window_update_resize_grab_op):
When moving or resizing a window use the last grab position
in computing change increment.
Provide support for warping the mouse pointer.
* src/window.h: new warp pointer and grab op helper functions
2002-08-09 00:27:23 -04:00
meta_window_begin_grab_op ( window ,
META_GRAB_OP_KEYBOARD_MOVING ,
2006-10-12 20:28:34 -04:00
TRUE ,
Partial audit to fix timestamp usage. One step towards fixing #355180; see
2006-09-18 Elijah Newren <newren gmail com>
Partial audit to fix timestamp usage. One step towards fixing
#355180; see important comments in that bug.
* src/core.[ch] (meta_core_unshade, meta_core_shade):
* src/delete.c (meta_window_present_delete_dialog,
delete_ping_timeout_func):
* src/display.[ch] (meta_display_open, meta_display_close,
event_callback, meta_display_begin_grab_op,
process_selection_clear, meta_display_unmanage_screen,
meta_display_unmanage_windows_for_screen):
* src/frames.c (meta_frames_button_press_event):
* src/keybindings.c (handle_toggle_shade):
* src/main.c (main):
* src/screen.[ch] (update_num_workspaces, meta_screen_new,
meta_screen_free, prefs_changed_callback):
* src/window.[ch] (meta_window_free, finish_minimize,
implement_showing, meta_window_show, meta_window_maximize,
meta_window_make_fullscreen_internal,
meta_window_unmake_fullscreen, meta_window_shade,
meta_window_unshade, window_activate, send_sync_request,
meta_window_client_message, menu_callback,
meta_window_update_keyboard_resize):
Remove usage of CurrentTime, meta_display_get_current_time() and
meta_display_get_current_time_roundtrip() where possible, or
document why it isn't possible, or at very least add a FIXME with
some explanation of my laziness and what needs to be done.
2006-09-18 13:27:24 -04:00
timestamp ) ;
2001-07-11 02:22:00 -04:00
break ;
case META_MENU_OP_RESIZE :
Patch to provide extra cues to the user when using window menu move and
2002-08-08 Craig Black <blackc@speakeasy.net>
Patch to provide extra cues to the user when using
window menu move and resize, #85724.
* src/common.h: add new cursors
* src/display.c: (grab_op_is_mouse)
(meta_display_create_x_cursor), (xcursor_for_op),
(meta_display_set_grab_op_cursor),
(meta_display_begin_grab_op):
The keyboard move and resize grab ops now also use the mouse.
Allow the grab cursor to be changed during the grab op.
Hold onto the initial grab position in case of reset.
* src/display.h: save the initial grab position
* src/keybindings.c: (process_keyboard_move_grab),
(process_keyboard_resize_grab), (handle_begin_move),
(handle_begin_resize):
The keyboard move and resize grab ops now also use the mouse.
* src/window.c: (meta_window_client_message), (menu_callback),
(update_move), (update_resize),
(meta_window_handle_mouse_grab_op_event), (warp_pointer),
(meta_window_warp_pointer), (meta_window_begin_grab_op),
(meta_window_update_resize_grab_op):
When moving or resizing a window use the last grab position
in computing change increment.
Provide support for warping the mouse pointer.
* src/window.h: new warp pointer and grab op helper functions
2002-08-09 00:27:23 -04:00
meta_window_begin_grab_op ( window ,
META_GRAB_OP_KEYBOARD_RESIZING_UNKNOWN ,
2006-10-12 20:28:34 -04:00
TRUE ,
Partial audit to fix timestamp usage. One step towards fixing #355180; see
2006-09-18 Elijah Newren <newren gmail com>
Partial audit to fix timestamp usage. One step towards fixing
#355180; see important comments in that bug.
* src/core.[ch] (meta_core_unshade, meta_core_shade):
* src/delete.c (meta_window_present_delete_dialog,
delete_ping_timeout_func):
* src/display.[ch] (meta_display_open, meta_display_close,
event_callback, meta_display_begin_grab_op,
process_selection_clear, meta_display_unmanage_screen,
meta_display_unmanage_windows_for_screen):
* src/frames.c (meta_frames_button_press_event):
* src/keybindings.c (handle_toggle_shade):
* src/main.c (main):
* src/screen.[ch] (update_num_workspaces, meta_screen_new,
meta_screen_free, prefs_changed_callback):
* src/window.[ch] (meta_window_free, finish_minimize,
implement_showing, meta_window_show, meta_window_maximize,
meta_window_make_fullscreen_internal,
meta_window_unmake_fullscreen, meta_window_shade,
meta_window_unshade, window_activate, send_sync_request,
meta_window_client_message, menu_callback,
meta_window_update_keyboard_resize):
Remove usage of CurrentTime, meta_display_get_current_time() and
meta_display_get_current_time_roundtrip() where possible, or
document why it isn't possible, or at very least add a FIXME with
some explanation of my laziness and what needs to be done.
2006-09-18 13:27:24 -04:00
timestamp ) ;
2001-07-11 02:22:00 -04:00
break ;
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
case META_MENU_OP_RECOVER :
meta_window_shove_titlebar_onscreen ( window ) ;
break ;
2008-12-18 07:41:56 -05:00
2001-06-23 01:49:35 -04:00
default :
meta_warning ( G_STRLOC " : Unknown window op \n " ) ;
break ;
}
2003-12-21 01:31:40 -05:00
if ( workspace )
{
meta_window_change_workspace ( window ,
workspace ) ;
}
2001-06-23 01:49:35 -04:00
}
else
{
meta_verbose ( " Menu callback on nonexistent window \n " ) ;
}
2002-10-28 00:35:30 -05:00
if ( display - > window_menu = = menu )
{
display - > window_menu = NULL ;
display - > window_with_menu = NULL ;
}
2008-12-18 07:41:56 -05:00
2001-06-23 01:49:35 -04:00
meta_ui_window_menu_free ( menu ) ;
}
void
meta_window_show_menu ( MetaWindow * window ,
int root_x ,
int root_y ,
int button ,
Fix issues on 64-bit machines with timestamps by using guint32 (like gtk+
2006-09-13 Elijah Newren <newren gmail com>
* src/common.h (MetaWindowMenuFunc):
* src/core.[ch] (meta_core_user_lower_and_unfocus,
meta_core_user_focus, meta_core_show_window_menu,
meta_core_begin_grab_op, meta_core_end_grab_op):
* src/delete.c (delete_ping_reply_func, delete_ping_timeout_func,
meta_window_delete):
* src/display.[ch] (struct MetaDisplay, struct MetaPingData,
sanity_check_timestamps, meta_display_open, event_callback,
meta_spew_event, meta_display_set_grab_op_cursor,
meta_display_begin_grab_op, meta_display_end_grab_op,
meta_display_ping_timeout, meta_display_ping_window,
process_pong_message, timestamp_too_old,
meta_display_set_input_focus_window):
* src/keybindings.[ch] (grab_keyboard, ungrab_keyboard,
meta_screen_grab_all_keys, meta_window_grab_all_keys,
meta_window_ungrab_all_keys, error_on_generic_command,
error_on_command, error_on_terminal_command):
* src/metacity-dialog.c (on_realize, warn_about_no_sm_support,
error_about_command, main):
* src/screen.[ch] (struct _MetaScreen, meta_screen_new,
meta_screen_show_desktop, meta_screen_apply_startup_properties):
* src/session.c (warn_about_lame_clients_and_finish_interact):
* src/window.[ch] (struct _MetaWindow,
intervening_user_event_occurred, window_activate,
meta_window_delete, meta_window_focus,
meta_window_send_icccm_message, meta_window_client_message,
menu_callback, meta_window_show_menu, struct EventScannerData,
check_use_this_motion_notify, meta_window_begin_grab_op,
meta_window_set_user_time):
* src/workspace.[ch] (focus_ancestor_or_mru_window,
meta_workspace_activate_with_focus, meta_workspace_activate,
meta_workspace_focus_default_window,
focus_ancestor_or_mru_window):
Fix issues on 64-bit machines with timestamps by using guint32
(like gtk+ does) instead of Time. #348305
2006-09-13 12:32:33 -04:00
guint32 timestamp )
2001-06-23 01:49:35 -04:00
{
MetaMenuOp ops ;
MetaMenuOp insensitive ;
MetaWindowMenu * menu ;
2003-12-21 01:31:40 -05:00
MetaWorkspaceLayout layout ;
2004-08-07 13:50:40 -04:00
int n_workspaces ;
2007-05-25 10:36:12 -04:00
gboolean ltr ;
2008-12-18 07:41:56 -05:00
2009-06-14 08:11:29 -04:00
g_return_if_fail ( ! window - > override_redirect ) ;
2003-03-16 20:19:07 -05:00
if ( window - > display - > window_menu )
{
meta_ui_window_menu_free ( window - > display - > window_menu ) ;
window - > display - > window_menu = NULL ;
window - > display - > window_with_menu = NULL ;
}
2010-03-31 16:34:02 -04:00
ops = META_MENU_OP_NONE ;
insensitive = META_MENU_OP_NONE ;
2001-06-23 01:49:35 -04:00
2004-08-07 13:50:40 -04:00
ops | = ( META_MENU_OP_DELETE | META_MENU_OP_MINIMIZE | META_MENU_OP_MOVE | META_MENU_OP_RESIZE ) ;
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
if ( ! meta_window_titlebar_is_onscreen ( window ) & &
window - > type ! = META_WINDOW_DOCK & &
window - > type ! = META_WINDOW_DESKTOP )
ops | = META_MENU_OP_RECOVER ;
2011-03-09 07:06:21 -05:00
if ( ! meta_prefs_get_workspaces_only_on_primary ( ) | |
meta_window_is_on_primary_monitor ( window ) )
{
n_workspaces = meta_screen_get_n_workspaces ( window - > screen ) ;
2004-08-07 13:50:40 -04:00
2011-03-09 07:06:21 -05:00
if ( n_workspaces > 1 )
ops | = META_MENU_OP_WORKSPACES ;
2003-12-21 01:31:40 -05:00
2011-03-09 07:06:21 -05:00
meta_screen_calc_workspace_layout ( window - > screen ,
n_workspaces ,
meta_workspace_index ( window - > screen - > active_workspace ) ,
& layout ) ;
2003-12-21 01:31:40 -05:00
2011-03-09 07:06:21 -05:00
if ( ! window - > on_all_workspaces )
{
ltr = meta_ui_get_direction ( ) = = META_UI_DIRECTION_LTR ;
if ( layout . current_col > 0 )
ops | = ltr ? META_MENU_OP_MOVE_LEFT : META_MENU_OP_MOVE_RIGHT ;
if ( ( layout . current_col < layout . cols - 1 ) & &
( layout . current_row * layout . cols + ( layout . current_col + 1 ) < n_workspaces ) )
ops | = ltr ? META_MENU_OP_MOVE_RIGHT : META_MENU_OP_MOVE_LEFT ;
if ( layout . current_row > 0 )
ops | = META_MENU_OP_MOVE_UP ;
if ( ( layout . current_row < layout . rows - 1 ) & &
( ( layout . current_row + 1 ) * layout . cols + layout . current_col < n_workspaces ) )
ops | = META_MENU_OP_MOVE_DOWN ;
}
2004-08-07 13:50:40 -04:00
2011-03-09 07:06:21 -05:00
meta_screen_free_workspace_layout ( & layout ) ;
2004-01-17 13:57:26 -05:00
2011-03-09 07:06:21 -05:00
ops | = META_MENU_OP_UNSTICK ;
ops | = META_MENU_OP_STICK ;
}
2005-10-03 15:22:40 -04:00
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
if ( META_WINDOW_MAXIMIZED ( window ) )
2001-06-23 01:49:35 -04:00
ops | = META_MENU_OP_UNMAXIMIZE ;
else
ops | = META_MENU_OP_MAXIMIZE ;
2008-12-18 07:41:56 -05:00
2007-04-16 00:48:14 -04:00
if ( window - > wm_state_above )
ops | = META_MENU_OP_UNABOVE ;
else
ops | = META_MENU_OP_ABOVE ;
2008-12-18 07:41:56 -05:00
2001-06-23 02:54:28 -04:00
if ( ! window - > has_maximize_func )
insensitive | = META_MENU_OP_UNMAXIMIZE | META_MENU_OP_MAXIMIZE ;
2008-12-18 07:41:56 -05:00
2001-06-23 01:49:35 -04:00
if ( ! window - > has_minimize_func )
insensitive | = META_MENU_OP_MINIMIZE ;
2008-12-18 07:41:56 -05:00
2001-06-23 01:49:35 -04:00
if ( ! window - > has_close_func )
insensitive | = META_MENU_OP_DELETE ;
2001-06-23 02:54:28 -04:00
if ( ! window - > has_shade_func )
insensitive | = META_MENU_OP_SHADE | META_MENU_OP_UNSHADE ;
2001-07-11 02:22:00 -04:00
2002-07-02 22:32:40 -04:00
if ( ! META_WINDOW_ALLOWS_MOVE ( window ) )
2001-07-11 02:22:00 -04:00
insensitive | = META_MENU_OP_MOVE ;
2002-07-02 22:32:40 -04:00
if ( ! META_WINDOW_ALLOWS_RESIZE ( window ) )
2001-07-11 02:22:00 -04:00
insensitive | = META_MENU_OP_RESIZE ;
2002-05-26 11:54:38 -04:00
if ( window - > always_sticky )
2007-07-31 08:20:14 -04:00
insensitive | = META_MENU_OP_STICK | META_MENU_OP_UNSTICK | META_MENU_OP_WORKSPACES ;
2004-01-17 13:57:26 -05:00
if ( ( window - > type = = META_WINDOW_DESKTOP ) | |
( window - > type = = META_WINDOW_DOCK ) | |
2012-05-30 07:57:04 -04:00
( window - > type = = META_WINDOW_SPLASHSCREEN | |
META_WINDOW_MAXIMIZED ( window ) ) )
2007-04-16 00:48:14 -04:00
insensitive | = META_MENU_OP_ABOVE | META_MENU_OP_UNABOVE ;
2004-12-24 14:40:09 -05:00
/* If all operations are disabled, just quit without showing the menu.
* This is the case , for example , with META_WINDOW_DESKTOP windows .
*/
if ( ( ops & ~ insensitive ) = = 0 )
return ;
2008-12-18 07:41:56 -05:00
2001-06-23 01:49:35 -04:00
menu =
meta_ui_window_menu_new ( window - > screen - > ui ,
window - > xwindow ,
ops ,
insensitive ,
meta_window_get_net_wm_desktop ( window ) ,
meta_screen_get_n_workspaces ( window - > screen ) ,
menu_callback ,
2008-12-18 07:41:56 -05:00
NULL ) ;
2001-06-23 01:49:35 -04:00
2002-10-28 00:35:30 -05:00
window - > display - > window_menu = menu ;
window - > display - > window_with_menu = window ;
2008-12-18 07:41:56 -05:00
2001-06-23 01:49:35 -04:00
meta_verbose ( " Popping up window menu for %s \n " , window - > desc ) ;
2008-12-18 07:41:56 -05:00
2001-06-23 01:49:35 -04:00
meta_ui_window_menu_popup ( menu , root_x , root_y , button , timestamp ) ;
}
2001-06-24 04:09:10 -04:00
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
void
meta_window_shove_titlebar_onscreen ( MetaWindow * window )
{
2013-11-19 12:24:45 -05:00
MetaRectangle frame_rect ;
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
GList * onscreen_region ;
int horiz_amount , vert_amount ;
int newx , newy ;
2009-06-15 18:25:07 -04:00
g_return_if_fail ( ! window - > override_redirect ) ;
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
/* If there's no titlebar, don't bother */
if ( ! window - > frame )
return ;
/* Get the basic info we need */
2013-11-19 12:24:45 -05:00
meta_window_get_frame_rect ( window , & frame_rect ) ;
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
onscreen_region = window - > screen - > active_workspace - > screen_region ;
/* Extend the region (just in case the window is too big to fit on the
* screen ) , then shove the window on screen , then return the region to
* normal .
*/
2013-11-19 12:24:45 -05:00
horiz_amount = frame_rect . width ;
vert_amount = frame_rect . height ;
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
meta_rectangle_expand_region ( onscreen_region ,
horiz_amount ,
2008-12-18 07:41:56 -05:00
horiz_amount ,
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
0 ,
vert_amount ) ;
meta_rectangle_shove_into_region ( onscreen_region ,
FIXED_DIRECTION_X ,
2013-11-19 12:24:45 -05:00
& frame_rect ) ;
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
meta_rectangle_expand_region ( onscreen_region ,
- horiz_amount ,
2008-12-18 07:41:56 -05:00
- horiz_amount ,
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
0 ,
- vert_amount ) ;
2013-11-19 12:24:45 -05:00
newx = frame_rect . x + window - > frame - > child_x ;
newy = frame_rect . y + window - > frame - > child_y ;
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
meta_window_move_resize ( window ,
2005-11-22 10:48:30 -05:00
FALSE ,
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
newx ,
newy ,
2008-12-18 07:41:56 -05:00
window - > rect . width ,
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
window - > rect . height ) ;
}
gboolean
meta_window_titlebar_is_onscreen ( MetaWindow * window )
{
MetaRectangle titlebar_rect ;
GList * onscreen_region ;
gboolean is_onscreen ;
const int min_height_needed = 8 ;
2013-12-19 10:56:31 -05:00
const float min_width_percent = 0.5 ;
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
const int min_width_absolute = 50 ;
/* Titlebar can't be offscreen if there is no titlebar... */
if ( ! window - > frame )
2014-03-05 19:40:09 -05:00
return TRUE ;
2008-12-18 07:41:56 -05:00
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
/* Get the rectangle corresponding to the titlebar */
2013-11-19 12:24:45 -05:00
meta_window_get_frame_rect ( window , & titlebar_rect ) ;
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
titlebar_rect . height = window - > frame - > child_y ;
/* Run through the spanning rectangles for the screen and see if one of
* them overlaps with the titlebar sufficiently to consider it onscreen .
*/
is_onscreen = FALSE ;
onscreen_region = window - > screen - > active_workspace - > screen_region ;
while ( onscreen_region )
{
MetaRectangle * spanning_rect = onscreen_region - > data ;
MetaRectangle overlap ;
2008-12-18 07:41:56 -05:00
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
meta_rectangle_intersect ( & titlebar_rect , spanning_rect , & overlap ) ;
if ( overlap . height > MIN ( titlebar_rect . height , min_height_needed ) & &
2008-12-18 07:41:56 -05:00
overlap . width > MIN ( titlebar_rect . width * min_width_percent ,
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
min_width_absolute ) )
{
is_onscreen = TRUE ;
break ;
}
2008-12-18 07:41:56 -05:00
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
onscreen_region = onscreen_region - > next ;
}
return is_onscreen ;
}
Fix bug 143333, support for update counter spec, and 109362, schedule
Sat Jun 19 02:21:08 2004 Soeren Sandmann <sandmann@daimi.au.dk>
Fix bug 143333, support for update counter spec, and 109362,
schedule compensation events when events are ignored.
* src/display.c (meta_display_open): Add _NET_WM_SYNC_REQUEST and
_NET_WM_SYNC_REQUEST_COUNTER atoms. Remove the old
METACITY_SYNC_COUNTER stuff.
(meta_display_begin_op): Setup the sync counter
* src/xprops.c, src/xprops.h, src/window-props.c, src/display.h:
Add new atoms.
* src/window.c (send_sync_request): new function.
(meta_window_move_resize_internal): send a sync request before
resizing.
(check_move_resize_frequence): Rework logic to also check the SYNC
case. If an event is ignored return the remaining time.
(update_resize_timeout): Timeout that gets called when a
compensation event is scheduled.
(uddate_resize): schedule compensation events when an event is
ignored.
(meta_window_handle_mouse_grap_op_event): When an alarm is
received and sync was turned off, turn it back on.
* src/window.h (struct MetaWindow) Add some variables
2004-06-18 20:45:24 -04:00
static double
timeval_to_ms ( const GTimeVal * timeval )
2002-10-03 17:49:13 -04:00
{
Fix bug 143333, support for update counter spec, and 109362, schedule
Sat Jun 19 02:21:08 2004 Soeren Sandmann <sandmann@daimi.au.dk>
Fix bug 143333, support for update counter spec, and 109362,
schedule compensation events when events are ignored.
* src/display.c (meta_display_open): Add _NET_WM_SYNC_REQUEST and
_NET_WM_SYNC_REQUEST_COUNTER atoms. Remove the old
METACITY_SYNC_COUNTER stuff.
(meta_display_begin_op): Setup the sync counter
* src/xprops.c, src/xprops.h, src/window-props.c, src/display.h:
Add new atoms.
* src/window.c (send_sync_request): new function.
(meta_window_move_resize_internal): send a sync request before
resizing.
(check_move_resize_frequence): Rework logic to also check the SYNC
case. If an event is ignored return the remaining time.
(update_resize_timeout): Timeout that gets called when a
compensation event is scheduled.
(uddate_resize): schedule compensation events when an event is
ignored.
(meta_window_handle_mouse_grap_op_event): When an alarm is
received and sync was turned off, turn it back on.
* src/window.h (struct MetaWindow) Add some variables
2004-06-18 20:45:24 -04:00
return ( timeval - > tv_sec * G_USEC_PER_SEC + timeval - > tv_usec ) / 1000.0 ;
}
static double
time_diff ( const GTimeVal * first ,
const GTimeVal * second )
{
double first_ms = timeval_to_ms ( first ) ;
double second_ms = timeval_to_ms ( second ) ;
return first_ms - second_ms ;
2002-10-03 17:49:13 -04:00
}
2002-08-10 01:27:17 -04:00
static gboolean
2008-12-18 07:41:56 -05:00
check_moveresize_frequency ( MetaWindow * window ,
Fix bug 143333, support for update counter spec, and 109362, schedule
Sat Jun 19 02:21:08 2004 Soeren Sandmann <sandmann@daimi.au.dk>
Fix bug 143333, support for update counter spec, and 109362,
schedule compensation events when events are ignored.
* src/display.c (meta_display_open): Add _NET_WM_SYNC_REQUEST and
_NET_WM_SYNC_REQUEST_COUNTER atoms. Remove the old
METACITY_SYNC_COUNTER stuff.
(meta_display_begin_op): Setup the sync counter
* src/xprops.c, src/xprops.h, src/window-props.c, src/display.h:
Add new atoms.
* src/window.c (send_sync_request): new function.
(meta_window_move_resize_internal): send a sync request before
resizing.
(check_move_resize_frequence): Rework logic to also check the SYNC
case. If an event is ignored return the remaining time.
(update_resize_timeout): Timeout that gets called when a
compensation event is scheduled.
(uddate_resize): schedule compensation events when an event is
ignored.
(meta_window_handle_mouse_grap_op_event): When an alarm is
received and sync was turned off, turn it back on.
* src/window.h (struct MetaWindow) Add some variables
2004-06-18 20:45:24 -04:00
gdouble * remaining )
2002-08-10 01:27:17 -04:00
{
GTimeVal current_time ;
2013-03-13 15:31:11 -04:00
const double max_resizes_per_second = 25.0 ;
const double ms_between_resizes = 1000.0 / max_resizes_per_second ;
double elapsed ;
2008-12-18 07:41:56 -05:00
2002-08-10 01:27:17 -04:00
g_get_current_time ( & current_time ) ;
2002-12-09 22:23:04 -05:00
# ifdef HAVE_XSYNC
2013-03-13 15:31:11 -04:00
/* If we are throttling via _NET_WM_SYNC_REQUEST, we don't need
* an artificial timeout - based throttled */
Fix bug 143333, support for update counter spec, and 109362, schedule
Sat Jun 19 02:21:08 2004 Soeren Sandmann <sandmann@daimi.au.dk>
Fix bug 143333, support for update counter spec, and 109362,
schedule compensation events when events are ignored.
* src/display.c (meta_display_open): Add _NET_WM_SYNC_REQUEST and
_NET_WM_SYNC_REQUEST_COUNTER atoms. Remove the old
METACITY_SYNC_COUNTER stuff.
(meta_display_begin_op): Setup the sync counter
* src/xprops.c, src/xprops.h, src/window-props.c, src/display.h:
Add new atoms.
* src/window.c (send_sync_request): new function.
(meta_window_move_resize_internal): send a sync request before
resizing.
(check_move_resize_frequence): Rework logic to also check the SYNC
case. If an event is ignored return the remaining time.
(update_resize_timeout): Timeout that gets called when a
compensation event is scheduled.
(uddate_resize): schedule compensation events when an event is
ignored.
(meta_window_handle_mouse_grap_op_event): When an alarm is
received and sync was turned off, turn it back on.
* src/window.h (struct MetaWindow) Add some variables
2004-06-18 20:45:24 -04:00
if ( ! window - > disable_sync & &
2011-06-08 11:26:03 -04:00
window - > sync_request_alarm ! = None )
2013-03-13 15:31:11 -04:00
return TRUE ;
2002-12-09 22:23:04 -05:00
# endif /* HAVE_XSYNC */
Fix bug 143333, support for update counter spec, and 109362, schedule
Sat Jun 19 02:21:08 2004 Soeren Sandmann <sandmann@daimi.au.dk>
Fix bug 143333, support for update counter spec, and 109362,
schedule compensation events when events are ignored.
* src/display.c (meta_display_open): Add _NET_WM_SYNC_REQUEST and
_NET_WM_SYNC_REQUEST_COUNTER atoms. Remove the old
METACITY_SYNC_COUNTER stuff.
(meta_display_begin_op): Setup the sync counter
* src/xprops.c, src/xprops.h, src/window-props.c, src/display.h:
Add new atoms.
* src/window.c (send_sync_request): new function.
(meta_window_move_resize_internal): send a sync request before
resizing.
(check_move_resize_frequence): Rework logic to also check the SYNC
case. If an event is ignored return the remaining time.
(update_resize_timeout): Timeout that gets called when a
compensation event is scheduled.
(uddate_resize): schedule compensation events when an event is
ignored.
(meta_window_handle_mouse_grap_op_event): When an alarm is
received and sync was turned off, turn it back on.
* src/window.h (struct MetaWindow) Add some variables
2004-06-18 20:45:24 -04:00
2013-03-13 15:31:11 -04:00
elapsed = time_diff ( & current_time , & window - > display - > grab_last_moveresize_time ) ;
Fix bug 143333, support for update counter spec, and 109362, schedule
Sat Jun 19 02:21:08 2004 Soeren Sandmann <sandmann@daimi.au.dk>
Fix bug 143333, support for update counter spec, and 109362,
schedule compensation events when events are ignored.
* src/display.c (meta_display_open): Add _NET_WM_SYNC_REQUEST and
_NET_WM_SYNC_REQUEST_COUNTER atoms. Remove the old
METACITY_SYNC_COUNTER stuff.
(meta_display_begin_op): Setup the sync counter
* src/xprops.c, src/xprops.h, src/window-props.c, src/display.h:
Add new atoms.
* src/window.c (send_sync_request): new function.
(meta_window_move_resize_internal): send a sync request before
resizing.
(check_move_resize_frequence): Rework logic to also check the SYNC
case. If an event is ignored return the remaining time.
(update_resize_timeout): Timeout that gets called when a
compensation event is scheduled.
(uddate_resize): schedule compensation events when an event is
ignored.
(meta_window_handle_mouse_grap_op_event): When an alarm is
received and sync was turned off, turn it back on.
* src/window.h (struct MetaWindow) Add some variables
2004-06-18 20:45:24 -04:00
2013-03-13 15:31:11 -04:00
if ( elapsed > = 0.0 & & elapsed < ms_between_resizes )
{
meta_topic ( META_DEBUG_RESIZING ,
" Delaying move/resize as only %g of %g ms elapsed \n " ,
elapsed , ms_between_resizes ) ;
2008-12-18 07:41:56 -05:00
2013-03-13 15:31:11 -04:00
if ( remaining )
* remaining = ( ms_between_resizes - elapsed ) ;
Fix bug 143333, support for update counter spec, and 109362, schedule
Sat Jun 19 02:21:08 2004 Soeren Sandmann <sandmann@daimi.au.dk>
Fix bug 143333, support for update counter spec, and 109362,
schedule compensation events when events are ignored.
* src/display.c (meta_display_open): Add _NET_WM_SYNC_REQUEST and
_NET_WM_SYNC_REQUEST_COUNTER atoms. Remove the old
METACITY_SYNC_COUNTER stuff.
(meta_display_begin_op): Setup the sync counter
* src/xprops.c, src/xprops.h, src/window-props.c, src/display.h:
Add new atoms.
* src/window.c (send_sync_request): new function.
(meta_window_move_resize_internal): send a sync request before
resizing.
(check_move_resize_frequence): Rework logic to also check the SYNC
case. If an event is ignored return the remaining time.
(update_resize_timeout): Timeout that gets called when a
compensation event is scheduled.
(uddate_resize): schedule compensation events when an event is
ignored.
(meta_window_handle_mouse_grap_op_event): When an alarm is
received and sync was turned off, turn it back on.
* src/window.h (struct MetaWindow) Add some variables
2004-06-18 20:45:24 -04:00
2013-03-13 15:31:11 -04:00
return FALSE ;
}
2008-12-18 07:41:56 -05:00
2013-03-13 15:31:11 -04:00
meta_topic ( META_DEBUG_RESIZING ,
" Checked moveresize freq, allowing move/resize now (%g of %g seconds elapsed) \n " ,
elapsed / 1000.0 , 1.0 / max_resizes_per_second ) ;
2008-12-18 07:41:56 -05:00
2013-03-13 15:31:11 -04:00
return TRUE ;
2002-08-10 01:27:17 -04:00
}
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
static gboolean
update_move_timeout ( gpointer data )
{
MetaWindow * window = data ;
2008-12-18 07:41:56 -05:00
update_move ( window ,
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
window - > display - > grab_last_user_action_was_snap ,
window - > display - > grab_latest_motion_x ,
window - > display - > grab_latest_motion_y ) ;
return FALSE ;
}
2002-08-10 01:27:17 -04:00
static void
2001-07-25 22:55:35 -04:00
update_move ( MetaWindow * window ,
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
gboolean snap ,
2001-07-25 22:55:35 -04:00
int x ,
int y )
{
int dx , dy ;
2001-07-25 23:58:24 -04:00
int new_x , new_y ;
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
MetaRectangle old ;
2003-06-12 01:55:06 -04:00
int shake_threshold ;
2006-02-25 11:54:39 -05:00
MetaDisplay * display = window - > display ;
2008-12-18 07:41:56 -05:00
2006-02-25 11:54:39 -05:00
display - > grab_latest_motion_x = x ;
display - > grab_latest_motion_y = y ;
2008-12-18 07:41:56 -05:00
2006-02-25 11:54:39 -05:00
dx = x - display - > grab_anchor_root_x ;
dy = y - display - > grab_anchor_root_y ;
2003-06-12 01:55:06 -04:00
2006-02-25 11:54:39 -05:00
new_x = display - > grab_anchor_window_pos . x + dx ;
new_y = display - > grab_anchor_window_pos . y + dy ;
2003-06-12 01:55:06 -04:00
2003-10-12 02:25:38 -04:00
meta_verbose ( " x,y = %d,%d anchor ptr %d,%d anchor pos %d,%d dx,dy %d,%d \n " ,
x , y ,
2006-02-25 11:54:39 -05:00
display - > grab_anchor_root_x ,
display - > grab_anchor_root_y ,
display - > grab_anchor_window_pos . x ,
display - > grab_anchor_window_pos . y ,
2003-10-12 02:25:38 -04:00
dx , dy ) ;
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
/* Don't bother doing anything if no move has been specified. (This
* happens often , even in keyboard moving , due to the warping of the
* pointer .
*/
if ( dx = = 0 & & dy = = 0 )
return ;
2010-06-24 14:41:28 -04:00
/* Originally for detaching maximized windows, but we use this
* for the zones at the sides of the monitor where trigger tiling
* because it ' s about the right size
2003-06-12 01:55:06 -04:00
*/
# define DRAG_THRESHOLD_TO_SHAKE_THRESHOLD_FACTOR 6
shake_threshold = meta_ui_get_drag_threshold ( window - > screen - > ui ) *
DRAG_THRESHOLD_TO_SHAKE_THRESHOLD_FACTOR ;
2008-12-18 07:41:56 -05:00
2011-10-20 10:41:52 -04:00
if ( snap )
{
/* We don't want to tile while snapping. Also, clear any previous tile
request . */
window - > tile_mode = META_TILE_NONE ;
window - > tile_monitor_number = - 1 ;
}
else if ( meta_prefs_get_edge_tiling ( ) & &
! META_WINDOW_MAXIMIZED ( window ) & &
! META_WINDOW_TILED_SIDE_BY_SIDE ( window ) )
2010-06-24 14:41:28 -04:00
{
const MetaMonitorInfo * monitor ;
MetaRectangle work_area ;
2010-09-24 15:35:00 -04:00
/* For side-by-side tiling we are interested in the inside vertical
2011-03-29 14:11:07 -04:00
* edges of the work area of the monitor where the pointer is located ,
* and in the outside top edge for maximized tiling .
*
* For maximized tiling we use the outside edge instead of the
* inside edge , because we don ' t want to force users to maximize
* windows they are placing near the top of their screens .
2011-10-08 11:00:16 -04:00
*
* The " current " idea of meta_window_get_work_area_current_monitor ( ) and
* meta_screen_get_current_monitor ( ) is slightly different : the former
* refers to the monitor which contains the largest part of the window ,
* the latter to the one where the pointer is located .
2010-06-24 14:41:28 -04:00
*/
2013-06-23 15:18:21 -04:00
monitor = meta_screen_get_current_monitor_info_for_pos ( window - > screen , x , y ) ;
2010-06-24 14:41:28 -04:00
meta_window_get_work_area_for_monitor ( window ,
monitor - > number ,
& work_area ) ;
2011-03-29 14:11:07 -04:00
/* Check if the cursor is in a position which triggers tiling
* and set tile_mode accordingly .
2010-09-24 15:35:00 -04:00
*/
2011-03-29 14:11:07 -04:00
if ( meta_window_can_tile_side_by_side ( window ) & &
x > = monitor - > rect . x & & x < ( work_area . x + shake_threshold ) )
window - > tile_mode = META_TILE_LEFT ;
else if ( meta_window_can_tile_side_by_side ( window ) & &
x > = work_area . x + work_area . width - shake_threshold & &
x < ( monitor - > rect . x + monitor - > rect . width ) )
window - > tile_mode = META_TILE_RIGHT ;
else if ( meta_window_can_tile_maximized ( window ) & &
y > = monitor - > rect . y & & y < = work_area . y )
window - > tile_mode = META_TILE_MAXIMIZED ;
else
window - > tile_mode = META_TILE_NONE ;
2011-10-08 11:00:16 -04:00
if ( window - > tile_mode ! = META_TILE_NONE )
window - > tile_monitor_number = monitor - > number ;
2010-06-24 14:41:28 -04:00
}
/* shake loose (unmaximize) maximized or tiled window if dragged beyond
* the threshold in the Y direction . Tiled windows can also be pulled
* loose via X motion .
*/
if ( ( META_WINDOW_MAXIMIZED ( window ) & & ABS ( dy ) > = shake_threshold ) | |
2010-09-24 14:31:36 -04:00
( META_WINDOW_TILED_SIDE_BY_SIDE ( window ) & & ( MAX ( ABS ( dx ) , ABS ( dy ) ) > = shake_threshold ) ) )
2003-06-12 01:55:06 -04:00
{
2003-07-27 22:09:20 -04:00
double prop ;
2010-12-02 16:43:21 -05:00
/* Shake loose, so that the window snaps back to maximized
* when dragged near the top ; do not snap back if tiling
* is enabled , as top edge tiling can be used in that case
*/
window - > shaken_loose = ! meta_prefs_get_edge_tiling ( ) ;
2010-06-24 14:41:28 -04:00
window - > tile_mode = META_TILE_NONE ;
2008-12-18 07:41:56 -05:00
2003-07-27 22:09:20 -04:00
/* move the unmaximized window to the cursor */
2008-12-18 07:41:56 -05:00
prop =
( ( double ) ( x - display - > grab_initial_window_pos . x ) ) /
2006-02-25 11:54:39 -05:00
( ( double ) display - > grab_initial_window_pos . width ) ;
2003-06-12 01:55:06 -04:00
2008-12-18 07:41:56 -05:00
display - > grab_initial_window_pos . x =
2004-02-16 14:01:25 -05:00
x - window - > saved_rect . width * prop ;
2006-02-25 11:54:39 -05:00
display - > grab_initial_window_pos . y = y ;
2003-07-27 22:09:20 -04:00
if ( window - > frame )
2004-02-16 14:01:25 -05:00
{
2006-02-25 11:54:39 -05:00
display - > grab_initial_window_pos . y + = window - > frame - > child_y / 2 ;
2004-02-16 14:01:25 -05:00
}
2006-02-25 11:54:39 -05:00
window - > saved_rect . x = display - > grab_initial_window_pos . x ;
window - > saved_rect . y = display - > grab_initial_window_pos . y ;
display - > grab_anchor_root_x = x ;
display - > grab_anchor_root_y = y ;
2003-07-27 22:09:20 -04:00
2014-03-13 18:32:20 -04:00
meta_window_unmaximize ( window , META_MAXIMIZE_BOTH ) ;
2003-06-12 01:55:06 -04:00
return ;
}
2010-06-24 14:41:28 -04:00
2009-08-31 09:52:16 -04:00
/* remaximize window on another monitor if window has been shaken
* loose or it is still maximized ( then move straight )
2003-06-12 01:55:06 -04:00
*/
2010-06-24 14:41:28 -04:00
else if ( ( window - > shaken_loose | | META_WINDOW_MAXIMIZED ( window ) ) & &
2010-09-24 15:35:00 -04:00
window - > tile_mode ! = META_TILE_LEFT & & window - > tile_mode ! = META_TILE_RIGHT )
2003-06-12 01:55:06 -04:00
{
2009-08-31 09:52:16 -04:00
const MetaMonitorInfo * wmonitor ;
2003-07-27 22:09:20 -04:00
MetaRectangle work_area ;
2003-06-12 01:55:06 -04:00
int monitor ;
2010-09-24 15:35:00 -04:00
window - > tile_mode = META_TILE_NONE ;
2009-08-31 09:52:16 -04:00
wmonitor = meta_screen_get_monitor_for_window ( window - > screen , window ) ;
2003-06-12 01:55:06 -04:00
2009-08-31 09:52:16 -04:00
for ( monitor = 0 ; monitor < window - > screen - > n_monitor_infos ; monitor + + )
2003-06-12 01:55:06 -04:00
{
2009-08-31 09:52:16 -04:00
meta_window_get_work_area_for_monitor ( window , monitor , & work_area ) ;
2003-07-27 22:09:20 -04:00
2009-08-31 09:52:16 -04:00
/* check if cursor is near the top of a monitor work area */
2003-07-27 22:09:20 -04:00
if ( x > = work_area . x & &
x < ( work_area . x + work_area . width ) & &
2004-02-16 14:01:25 -05:00
y > = work_area . y & &
y < ( work_area . y + shake_threshold ) )
2003-06-12 01:55:06 -04:00
{
/* move the saved rect if window will become maximized on an
* other monitor so user isn ' t surprised on a later unmaximize
*/
2009-08-31 09:52:16 -04:00
if ( wmonitor - > number ! = monitor )
2003-06-12 01:55:06 -04:00
{
2003-07-27 22:09:20 -04:00
window - > saved_rect . x = work_area . x ;
2004-02-16 14:01:25 -05:00
window - > saved_rect . y = work_area . y ;
2008-12-18 07:41:56 -05:00
if ( window - > frame )
2003-06-12 01:55:06 -04:00
{
window - > saved_rect . x + = window - > frame - > child_x ;
window - > saved_rect . y + = window - > frame - > child_y ;
}
2006-12-05 20:26:30 -05:00
window - > user_rect . x = window - > saved_rect . x ;
window - > user_rect . y = window - > saved_rect . y ;
2014-03-13 18:32:20 -04:00
meta_window_unmaximize ( window , META_MAXIMIZE_BOTH ) ;
2003-06-12 01:55:06 -04:00
}
2001-07-25 22:55:35 -04:00
2006-02-25 11:54:39 -05:00
display - > grab_initial_window_pos = work_area ;
display - > grab_anchor_root_x = x ;
display - > grab_anchor_root_y = y ;
2004-02-16 14:01:25 -05:00
window - > shaken_loose = FALSE ;
2008-12-18 07:41:56 -05:00
2014-03-13 18:32:20 -04:00
meta_window_maximize ( window , META_MAXIMIZE_BOTH ) ;
2004-02-16 14:01:25 -05:00
return ;
2003-06-12 01:55:06 -04:00
}
}
}
2010-06-24 14:41:28 -04:00
/* Delay showing the tile preview slightly to make it more unlikely to
* trigger it unwittingly , e . g . when shaking loose the window or moving
* it to another monitor .
*/
2012-03-24 06:29:39 -04:00
meta_screen_update_tile_preview ( window - > screen ,
2010-06-24 14:41:28 -04:00
window - > tile_mode ! = META_TILE_NONE ) ;
2009-05-07 13:53:47 -04:00
meta_window_get_client_root_coords ( window , & old ) ;
2005-08-02 22:22:00 -04:00
2010-06-24 14:41:28 -04:00
/* Don't allow movement in the maximized directions or while tiled */
2010-09-24 14:31:36 -04:00
if ( window - > maximized_horizontally | | META_WINDOW_TILED_SIDE_BY_SIDE ( window ) )
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
new_x = old . x ;
if ( window - > maximized_vertically )
new_y = old . y ;
/* Do any edge resistance/snapping */
2008-12-18 07:41:56 -05:00
meta_window_edge_resistance_for_move ( window ,
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
old . x ,
old . y ,
& new_x ,
& new_y ,
update_move_timeout ,
snap ,
FALSE ) ;
2009-05-07 13:53:47 -04:00
meta_window_move ( window , TRUE , new_x , new_y ) ;
2001-07-25 22:55:35 -04:00
}
2010-09-18 16:28:30 -04:00
/* When resizing a maximized window by using alt-middle-drag (resizing
* with the grips or the menu for a maximized window is not enabled ) ,
* the user can " break " out of the maximized state . This checks for
* that possibility . During such a break - out resize the user can also
* return to the previous maximization state by resizing back to near
* the original size .
*/
static MetaMaximizeFlags
check_resize_unmaximize ( MetaWindow * window ,
int dx ,
int dy )
{
int threshold ;
MetaMaximizeFlags new_unmaximize ;
# define DRAG_THRESHOLD_TO_RESIZE_THRESHOLD_FACTOR 3
threshold = meta_ui_get_drag_threshold ( window - > screen - > ui ) *
DRAG_THRESHOLD_TO_RESIZE_THRESHOLD_FACTOR ;
new_unmaximize = 0 ;
if ( window - > maximized_horizontally | |
2010-09-19 10:54:08 -04:00
window - > tile_mode ! = META_TILE_NONE | |
2010-09-18 16:28:30 -04:00
( window - > display - > grab_resize_unmaximize & META_MAXIMIZE_HORIZONTAL ) ! = 0 )
{
int x_amount ;
/* We allow breaking out of maximization in either direction, to make
* the window larger than the monitor as well as smaller than the
* monitor . If we wanted to only allow resizing smaller than the
* monitor , we ' d use - dx for NE / E / SE and dx for SW / W / NW .
*/
switch ( window - > display - > grab_op )
{
case META_GRAB_OP_RESIZING_NE :
case META_GRAB_OP_KEYBOARD_RESIZING_NE :
case META_GRAB_OP_RESIZING_E :
case META_GRAB_OP_KEYBOARD_RESIZING_E :
case META_GRAB_OP_RESIZING_SE :
case META_GRAB_OP_KEYBOARD_RESIZING_SE :
case META_GRAB_OP_RESIZING_SW :
case META_GRAB_OP_KEYBOARD_RESIZING_SW :
case META_GRAB_OP_RESIZING_W :
case META_GRAB_OP_KEYBOARD_RESIZING_W :
case META_GRAB_OP_RESIZING_NW :
case META_GRAB_OP_KEYBOARD_RESIZING_NW :
x_amount = dx < 0 ? - dx : dx ;
break ;
default :
x_amount = 0 ;
break ;
}
if ( x_amount > threshold )
new_unmaximize | = META_MAXIMIZE_HORIZONTAL ;
}
if ( window - > maximized_vertically | |
( window - > display - > grab_resize_unmaximize & META_MAXIMIZE_VERTICAL ) ! = 0 )
{
int y_amount ;
switch ( window - > display - > grab_op )
{
case META_GRAB_OP_RESIZING_N :
case META_GRAB_OP_KEYBOARD_RESIZING_N :
case META_GRAB_OP_RESIZING_NE :
case META_GRAB_OP_KEYBOARD_RESIZING_NE :
case META_GRAB_OP_RESIZING_NW :
case META_GRAB_OP_KEYBOARD_RESIZING_NW :
case META_GRAB_OP_RESIZING_SE :
case META_GRAB_OP_KEYBOARD_RESIZING_SE :
case META_GRAB_OP_RESIZING_S :
case META_GRAB_OP_KEYBOARD_RESIZING_S :
case META_GRAB_OP_RESIZING_SW :
case META_GRAB_OP_KEYBOARD_RESIZING_SW :
y_amount = dy < 0 ? - dy : dy ;
break ;
default :
y_amount = 0 ;
break ;
}
if ( y_amount > threshold )
new_unmaximize | = META_MAXIMIZE_VERTICAL ;
}
/* Metacity doesn't have a full user interface for only horizontally or
* vertically maximized , so while only unmaximizing in the direction drags
* has some advantages , it will also confuse the user . So , we always
* unmaximize both ways if possible .
*/
if ( new_unmaximize ! = 0 )
{
new_unmaximize = 0 ;
if ( window - > maximized_horizontally | |
( window - > display - > grab_resize_unmaximize & META_MAXIMIZE_HORIZONTAL ) ! = 0 )
new_unmaximize | = META_MAXIMIZE_HORIZONTAL ;
if ( window - > maximized_vertically | |
( window - > display - > grab_resize_unmaximize & META_MAXIMIZE_VERTICAL ) ! = 0 )
new_unmaximize | = META_MAXIMIZE_VERTICAL ;
}
return new_unmaximize ;
}
Fix bug 143333, support for update counter spec, and 109362, schedule
Sat Jun 19 02:21:08 2004 Soeren Sandmann <sandmann@daimi.au.dk>
Fix bug 143333, support for update counter spec, and 109362,
schedule compensation events when events are ignored.
* src/display.c (meta_display_open): Add _NET_WM_SYNC_REQUEST and
_NET_WM_SYNC_REQUEST_COUNTER atoms. Remove the old
METACITY_SYNC_COUNTER stuff.
(meta_display_begin_op): Setup the sync counter
* src/xprops.c, src/xprops.h, src/window-props.c, src/display.h:
Add new atoms.
* src/window.c (send_sync_request): new function.
(meta_window_move_resize_internal): send a sync request before
resizing.
(check_move_resize_frequence): Rework logic to also check the SYNC
case. If an event is ignored return the remaining time.
(update_resize_timeout): Timeout that gets called when a
compensation event is scheduled.
(uddate_resize): schedule compensation events when an event is
ignored.
(meta_window_handle_mouse_grap_op_event): When an alarm is
received and sync was turned off, turn it back on.
* src/window.h (struct MetaWindow) Add some variables
2004-06-18 20:45:24 -04:00
static gboolean
update_resize_timeout ( gpointer data )
{
MetaWindow * window = data ;
2008-12-18 07:41:56 -05:00
update_resize ( window ,
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
window - > display - > grab_last_user_action_was_snap ,
window - > display - > grab_latest_motion_x ,
window - > display - > grab_latest_motion_y ,
TRUE ) ;
Fix bug 143333, support for update counter spec, and 109362, schedule
Sat Jun 19 02:21:08 2004 Soeren Sandmann <sandmann@daimi.au.dk>
Fix bug 143333, support for update counter spec, and 109362,
schedule compensation events when events are ignored.
* src/display.c (meta_display_open): Add _NET_WM_SYNC_REQUEST and
_NET_WM_SYNC_REQUEST_COUNTER atoms. Remove the old
METACITY_SYNC_COUNTER stuff.
(meta_display_begin_op): Setup the sync counter
* src/xprops.c, src/xprops.h, src/window-props.c, src/display.h:
Add new atoms.
* src/window.c (send_sync_request): new function.
(meta_window_move_resize_internal): send a sync request before
resizing.
(check_move_resize_frequence): Rework logic to also check the SYNC
case. If an event is ignored return the remaining time.
(update_resize_timeout): Timeout that gets called when a
compensation event is scheduled.
(uddate_resize): schedule compensation events when an event is
ignored.
(meta_window_handle_mouse_grap_op_event): When an alarm is
received and sync was turned off, turn it back on.
* src/window.h (struct MetaWindow) Add some variables
2004-06-18 20:45:24 -04:00
return FALSE ;
}
2001-07-25 22:55:35 -04:00
static void
update_resize ( MetaWindow * window ,
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
gboolean snap ,
Fix bug 143333, support for update counter spec, and 109362, schedule
Sat Jun 19 02:21:08 2004 Soeren Sandmann <sandmann@daimi.au.dk>
Fix bug 143333, support for update counter spec, and 109362,
schedule compensation events when events are ignored.
* src/display.c (meta_display_open): Add _NET_WM_SYNC_REQUEST and
_NET_WM_SYNC_REQUEST_COUNTER atoms. Remove the old
METACITY_SYNC_COUNTER stuff.
(meta_display_begin_op): Setup the sync counter
* src/xprops.c, src/xprops.h, src/window-props.c, src/display.h:
Add new atoms.
* src/window.c (send_sync_request): new function.
(meta_window_move_resize_internal): send a sync request before
resizing.
(check_move_resize_frequence): Rework logic to also check the SYNC
case. If an event is ignored return the remaining time.
(update_resize_timeout): Timeout that gets called when a
compensation event is scheduled.
(uddate_resize): schedule compensation events when an event is
ignored.
(meta_window_handle_mouse_grap_op_event): When an alarm is
received and sync was turned off, turn it back on.
* src/window.h (struct MetaWindow) Add some variables
2004-06-18 20:45:24 -04:00
int x , int y ,
Partial audit to fix timestamp usage. One step towards fixing #355180; see
2006-09-18 Elijah Newren <newren gmail com>
Partial audit to fix timestamp usage. One step towards fixing
#355180; see important comments in that bug.
* src/core.[ch] (meta_core_unshade, meta_core_shade):
* src/delete.c (meta_window_present_delete_dialog,
delete_ping_timeout_func):
* src/display.[ch] (meta_display_open, meta_display_close,
event_callback, meta_display_begin_grab_op,
process_selection_clear, meta_display_unmanage_screen,
meta_display_unmanage_windows_for_screen):
* src/frames.c (meta_frames_button_press_event):
* src/keybindings.c (handle_toggle_shade):
* src/main.c (main):
* src/screen.[ch] (update_num_workspaces, meta_screen_new,
meta_screen_free, prefs_changed_callback):
* src/window.[ch] (meta_window_free, finish_minimize,
implement_showing, meta_window_show, meta_window_maximize,
meta_window_make_fullscreen_internal,
meta_window_unmake_fullscreen, meta_window_shade,
meta_window_unshade, window_activate, send_sync_request,
meta_window_client_message, menu_callback,
meta_window_update_keyboard_resize):
Remove usage of CurrentTime, meta_display_get_current_time() and
meta_display_get_current_time_roundtrip() where possible, or
document why it isn't possible, or at very least add a FIXME with
some explanation of my laziness and what needs to be done.
2006-09-18 13:27:24 -04:00
gboolean force )
2001-07-25 22:55:35 -04:00
{
int dx , dy ;
int new_w , new_h ;
int gravity ;
2002-12-09 22:23:04 -05:00
MetaRectangle old ;
Fix bug 143333, support for update counter spec, and 109362, schedule
Sat Jun 19 02:21:08 2004 Soeren Sandmann <sandmann@daimi.au.dk>
Fix bug 143333, support for update counter spec, and 109362,
schedule compensation events when events are ignored.
* src/display.c (meta_display_open): Add _NET_WM_SYNC_REQUEST and
_NET_WM_SYNC_REQUEST_COUNTER atoms. Remove the old
METACITY_SYNC_COUNTER stuff.
(meta_display_begin_op): Setup the sync counter
* src/xprops.c, src/xprops.h, src/window-props.c, src/display.h:
Add new atoms.
* src/window.c (send_sync_request): new function.
(meta_window_move_resize_internal): send a sync request before
resizing.
(check_move_resize_frequence): Rework logic to also check the SYNC
case. If an event is ignored return the remaining time.
(update_resize_timeout): Timeout that gets called when a
compensation event is scheduled.
(uddate_resize): schedule compensation events when an event is
ignored.
(meta_window_handle_mouse_grap_op_event): When an alarm is
received and sync was turned off, turn it back on.
* src/window.h (struct MetaWindow) Add some variables
2004-06-18 20:45:24 -04:00
double remaining ;
2010-09-18 16:28:30 -04:00
MetaMaximizeFlags new_unmaximize ;
2008-12-18 07:41:56 -05:00
2002-12-09 22:23:04 -05:00
window - > display - > grab_latest_motion_x = x ;
window - > display - > grab_latest_motion_y = y ;
2008-12-18 07:41:56 -05:00
2003-10-12 02:25:38 -04:00
dx = x - window - > display - > grab_anchor_root_x ;
dy = y - window - > display - > grab_anchor_root_y ;
2001-07-25 22:55:35 -04:00
2012-07-03 16:30:02 -04:00
/* Attached modal dialogs are special in that size
* changes apply to both sides , so that the dialog
2011-03-21 13:23:53 -04:00
* remains centered to the parent .
*/
window: make determination of attached dialog windows more consistent
Different bits of code were using slightly different checks to test
whether a window was an attached dialog. Add a new
meta_window_is_attached_dialog(), and use that everywhere.
Also, freeze the is-attached status when the window is first shown,
rather than recomputing it each time the caller asks, since this could
cause problems if a window changes its type after it has already been
attached, etc. However, if an attached window's parent is destroyed,
or an attached window changes its transient-for, then fix things up by
destroying the old MetaWindow and creating a new one (causing
compositor unmap and map events to be fired off, allowing the display
of the window to be fixed up).
Remove some code in display.c that tried to fix existing windows if
the gconf setting changed, but which didn't actually do anything (at
least under gnome-shell). However, if 654643 was fixed then the new
behavior with this patch would be that changing the gconf setting
would affect new dialogs, but not existing ones.
https://bugzilla.gnome.org/show_bug.cgi?id=646761
2011-04-05 10:14:52 -04:00
if ( meta_window_is_attached_dialog ( window ) )
2012-07-03 16:30:02 -04:00
{
dx * = 2 ;
dy * = 2 ;
}
2011-03-21 13:23:53 -04:00
2003-10-12 02:25:38 -04:00
new_w = window - > display - > grab_anchor_window_pos . width ;
new_h = window - > display - > grab_anchor_window_pos . height ;
2001-07-25 22:55:35 -04:00
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
/* Don't bother doing anything if no move has been specified. (This
* happens often , even in keyboard resizing , due to the warping of the
* pointer .
*/
if ( dx = = 0 & & dy = = 0 )
return ;
2008-12-18 07:41:56 -05:00
2005-08-02 22:22:00 -04:00
if ( window - > display - > grab_op = = META_GRAB_OP_KEYBOARD_RESIZING_UNKNOWN )
{
if ( ( dx > 0 ) & & ( dy > 0 ) )
{
window - > display - > grab_op = META_GRAB_OP_KEYBOARD_RESIZING_SE ;
meta_window_update_keyboard_resize ( window , TRUE ) ;
}
else if ( ( dx < 0 ) & & ( dy > 0 ) )
{
window - > display - > grab_op = META_GRAB_OP_KEYBOARD_RESIZING_SW ;
meta_window_update_keyboard_resize ( window , TRUE ) ;
}
else if ( ( dx > 0 ) & & ( dy < 0 ) )
{
window - > display - > grab_op = META_GRAB_OP_KEYBOARD_RESIZING_NE ;
meta_window_update_keyboard_resize ( window , TRUE ) ;
}
else if ( ( dx < 0 ) & & ( dy < 0 ) )
{
window - > display - > grab_op = META_GRAB_OP_KEYBOARD_RESIZING_NW ;
meta_window_update_keyboard_resize ( window , TRUE ) ;
}
else if ( dx < 0 )
{
window - > display - > grab_op = META_GRAB_OP_KEYBOARD_RESIZING_W ;
meta_window_update_keyboard_resize ( window , TRUE ) ;
}
else if ( dx > 0 )
{
window - > display - > grab_op = META_GRAB_OP_KEYBOARD_RESIZING_E ;
meta_window_update_keyboard_resize ( window , TRUE ) ;
}
else if ( dy > 0 )
{
window - > display - > grab_op = META_GRAB_OP_KEYBOARD_RESIZING_S ;
meta_window_update_keyboard_resize ( window , TRUE ) ;
}
else if ( dy < 0 )
{
window - > display - > grab_op = META_GRAB_OP_KEYBOARD_RESIZING_N ;
meta_window_update_keyboard_resize ( window , TRUE ) ;
}
}
2008-12-18 07:41:56 -05:00
2010-09-18 16:28:30 -04:00
new_unmaximize = check_resize_unmaximize ( window , dx , dy ) ;
2001-07-25 22:55:35 -04:00
switch ( window - > display - > grab_op )
{
case META_GRAB_OP_RESIZING_SE :
case META_GRAB_OP_RESIZING_NE :
case META_GRAB_OP_RESIZING_E :
Patch to provide extra cues to the user when using window menu move and
2002-08-08 Craig Black <blackc@speakeasy.net>
Patch to provide extra cues to the user when using
window menu move and resize, #85724.
* src/common.h: add new cursors
* src/display.c: (grab_op_is_mouse)
(meta_display_create_x_cursor), (xcursor_for_op),
(meta_display_set_grab_op_cursor),
(meta_display_begin_grab_op):
The keyboard move and resize grab ops now also use the mouse.
Allow the grab cursor to be changed during the grab op.
Hold onto the initial grab position in case of reset.
* src/display.h: save the initial grab position
* src/keybindings.c: (process_keyboard_move_grab),
(process_keyboard_resize_grab), (handle_begin_move),
(handle_begin_resize):
The keyboard move and resize grab ops now also use the mouse.
* src/window.c: (meta_window_client_message), (menu_callback),
(update_move), (update_resize),
(meta_window_handle_mouse_grab_op_event), (warp_pointer),
(meta_window_warp_pointer), (meta_window_begin_grab_op),
(meta_window_update_resize_grab_op):
When moving or resizing a window use the last grab position
in computing change increment.
Provide support for warping the mouse pointer.
* src/window.h: new warp pointer and grab op helper functions
2002-08-09 00:27:23 -04:00
case META_GRAB_OP_KEYBOARD_RESIZING_SE :
case META_GRAB_OP_KEYBOARD_RESIZING_NE :
case META_GRAB_OP_KEYBOARD_RESIZING_E :
2001-07-25 22:55:35 -04:00
new_w + = dx ;
break ;
case META_GRAB_OP_RESIZING_NW :
case META_GRAB_OP_RESIZING_SW :
case META_GRAB_OP_RESIZING_W :
Patch to provide extra cues to the user when using window menu move and
2002-08-08 Craig Black <blackc@speakeasy.net>
Patch to provide extra cues to the user when using
window menu move and resize, #85724.
* src/common.h: add new cursors
* src/display.c: (grab_op_is_mouse)
(meta_display_create_x_cursor), (xcursor_for_op),
(meta_display_set_grab_op_cursor),
(meta_display_begin_grab_op):
The keyboard move and resize grab ops now also use the mouse.
Allow the grab cursor to be changed during the grab op.
Hold onto the initial grab position in case of reset.
* src/display.h: save the initial grab position
* src/keybindings.c: (process_keyboard_move_grab),
(process_keyboard_resize_grab), (handle_begin_move),
(handle_begin_resize):
The keyboard move and resize grab ops now also use the mouse.
* src/window.c: (meta_window_client_message), (menu_callback),
(update_move), (update_resize),
(meta_window_handle_mouse_grab_op_event), (warp_pointer),
(meta_window_warp_pointer), (meta_window_begin_grab_op),
(meta_window_update_resize_grab_op):
When moving or resizing a window use the last grab position
in computing change increment.
Provide support for warping the mouse pointer.
* src/window.h: new warp pointer and grab op helper functions
2002-08-09 00:27:23 -04:00
case META_GRAB_OP_KEYBOARD_RESIZING_NW :
case META_GRAB_OP_KEYBOARD_RESIZING_SW :
case META_GRAB_OP_KEYBOARD_RESIZING_W :
2001-07-25 22:55:35 -04:00
new_w - = dx ;
break ;
2009-05-07 13:53:47 -04:00
default :
break ;
}
2008-12-18 07:41:56 -05:00
2001-07-25 22:55:35 -04:00
switch ( window - > display - > grab_op )
{
case META_GRAB_OP_RESIZING_SE :
case META_GRAB_OP_RESIZING_S :
case META_GRAB_OP_RESIZING_SW :
Patch to provide extra cues to the user when using window menu move and
2002-08-08 Craig Black <blackc@speakeasy.net>
Patch to provide extra cues to the user when using
window menu move and resize, #85724.
* src/common.h: add new cursors
* src/display.c: (grab_op_is_mouse)
(meta_display_create_x_cursor), (xcursor_for_op),
(meta_display_set_grab_op_cursor),
(meta_display_begin_grab_op):
The keyboard move and resize grab ops now also use the mouse.
Allow the grab cursor to be changed during the grab op.
Hold onto the initial grab position in case of reset.
* src/display.h: save the initial grab position
* src/keybindings.c: (process_keyboard_move_grab),
(process_keyboard_resize_grab), (handle_begin_move),
(handle_begin_resize):
The keyboard move and resize grab ops now also use the mouse.
* src/window.c: (meta_window_client_message), (menu_callback),
(update_move), (update_resize),
(meta_window_handle_mouse_grab_op_event), (warp_pointer),
(meta_window_warp_pointer), (meta_window_begin_grab_op),
(meta_window_update_resize_grab_op):
When moving or resizing a window use the last grab position
in computing change increment.
Provide support for warping the mouse pointer.
* src/window.h: new warp pointer and grab op helper functions
2002-08-09 00:27:23 -04:00
case META_GRAB_OP_KEYBOARD_RESIZING_SE :
case META_GRAB_OP_KEYBOARD_RESIZING_S :
case META_GRAB_OP_KEYBOARD_RESIZING_SW :
2001-07-25 22:55:35 -04:00
new_h + = dy ;
break ;
2008-12-18 07:41:56 -05:00
2001-07-25 22:55:35 -04:00
case META_GRAB_OP_RESIZING_N :
case META_GRAB_OP_RESIZING_NE :
case META_GRAB_OP_RESIZING_NW :
Patch to provide extra cues to the user when using window menu move and
2002-08-08 Craig Black <blackc@speakeasy.net>
Patch to provide extra cues to the user when using
window menu move and resize, #85724.
* src/common.h: add new cursors
* src/display.c: (grab_op_is_mouse)
(meta_display_create_x_cursor), (xcursor_for_op),
(meta_display_set_grab_op_cursor),
(meta_display_begin_grab_op):
The keyboard move and resize grab ops now also use the mouse.
Allow the grab cursor to be changed during the grab op.
Hold onto the initial grab position in case of reset.
* src/display.h: save the initial grab position
* src/keybindings.c: (process_keyboard_move_grab),
(process_keyboard_resize_grab), (handle_begin_move),
(handle_begin_resize):
The keyboard move and resize grab ops now also use the mouse.
* src/window.c: (meta_window_client_message), (menu_callback),
(update_move), (update_resize),
(meta_window_handle_mouse_grab_op_event), (warp_pointer),
(meta_window_warp_pointer), (meta_window_begin_grab_op),
(meta_window_update_resize_grab_op):
When moving or resizing a window use the last grab position
in computing change increment.
Provide support for warping the mouse pointer.
* src/window.h: new warp pointer and grab op helper functions
2002-08-09 00:27:23 -04:00
case META_GRAB_OP_KEYBOARD_RESIZING_N :
case META_GRAB_OP_KEYBOARD_RESIZING_NE :
case META_GRAB_OP_KEYBOARD_RESIZING_NW :
2001-07-25 22:55:35 -04:00
new_h - = dy ;
break ;
default :
break ;
}
2013-03-18 14:16:41 -04:00
# ifdef HAVE_XSYNC
/* If we're waiting for a request for _NET_WM_SYNC_REQUEST, we'll
* resize the window when the window responds , or when we time
* the response out .
*/
if ( window - > sync_request_timeout_id ! = 0 )
return ;
# endif
Fix bug 143333, support for update counter spec, and 109362, schedule
Sat Jun 19 02:21:08 2004 Soeren Sandmann <sandmann@daimi.au.dk>
Fix bug 143333, support for update counter spec, and 109362,
schedule compensation events when events are ignored.
* src/display.c (meta_display_open): Add _NET_WM_SYNC_REQUEST and
_NET_WM_SYNC_REQUEST_COUNTER atoms. Remove the old
METACITY_SYNC_COUNTER stuff.
(meta_display_begin_op): Setup the sync counter
* src/xprops.c, src/xprops.h, src/window-props.c, src/display.h:
Add new atoms.
* src/window.c (send_sync_request): new function.
(meta_window_move_resize_internal): send a sync request before
resizing.
(check_move_resize_frequence): Rework logic to also check the SYNC
case. If an event is ignored return the remaining time.
(update_resize_timeout): Timeout that gets called when a
compensation event is scheduled.
(uddate_resize): schedule compensation events when an event is
ignored.
(meta_window_handle_mouse_grap_op_event): When an alarm is
received and sync was turned off, turn it back on.
* src/window.h (struct MetaWindow) Add some variables
2004-06-18 20:45:24 -04:00
if ( ! check_moveresize_frequency ( window , & remaining ) & & ! force )
{
/* we are ignoring an event here, so we schedule a
* compensation event when we would otherwise not ignore
* an event . Otherwise we can become stuck if the user never
* generates another event .
*/
if ( ! window - > display - > grab_resize_timeout_id )
{
2008-12-18 07:41:56 -05:00
window - > display - > grab_resize_timeout_id =
Fix bug 143333, support for update counter spec, and 109362, schedule
Sat Jun 19 02:21:08 2004 Soeren Sandmann <sandmann@daimi.au.dk>
Fix bug 143333, support for update counter spec, and 109362,
schedule compensation events when events are ignored.
* src/display.c (meta_display_open): Add _NET_WM_SYNC_REQUEST and
_NET_WM_SYNC_REQUEST_COUNTER atoms. Remove the old
METACITY_SYNC_COUNTER stuff.
(meta_display_begin_op): Setup the sync counter
* src/xprops.c, src/xprops.h, src/window-props.c, src/display.h:
Add new atoms.
* src/window.c (send_sync_request): new function.
(meta_window_move_resize_internal): send a sync request before
resizing.
(check_move_resize_frequence): Rework logic to also check the SYNC
case. If an event is ignored return the remaining time.
(update_resize_timeout): Timeout that gets called when a
compensation event is scheduled.
(uddate_resize): schedule compensation events when an event is
ignored.
(meta_window_handle_mouse_grap_op_event): When an alarm is
received and sync was turned off, turn it back on.
* src/window.h (struct MetaWindow) Add some variables
2004-06-18 20:45:24 -04:00
g_timeout_add ( ( int ) remaining , update_resize_timeout , window ) ;
}
return ;
}
/* Remove any scheduled compensation events */
if ( window - > display - > grab_resize_timeout_id )
{
g_source_remove ( window - > display - > grab_resize_timeout_id ) ;
window - > display - > grab_resize_timeout_id = 0 ;
}
2008-12-18 07:41:56 -05:00
2009-05-07 13:53:47 -04:00
old = window - > rect ; /* Don't actually care about x,y */
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
/* One sided resizing ought to actually be one-sided, despite the fact that
* aspect ratio windows don ' t interact nicely with the above stuff . So ,
* to avoid some nasty flicker , we enforce that .
*/
switch ( window - > display - > grab_op )
{
case META_GRAB_OP_RESIZING_S :
case META_GRAB_OP_RESIZING_N :
new_w = old . width ;
break ;
2008-12-18 07:41:56 -05:00
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
case META_GRAB_OP_RESIZING_E :
case META_GRAB_OP_RESIZING_W :
new_h = old . height ;
break ;
default :
break ;
}
2003-10-12 02:25:38 -04:00
2001-07-25 22:55:35 -04:00
/* compute gravity of client during operation */
2002-03-17 12:22:23 -05:00
gravity = meta_resize_gravity_from_grab_op ( window - > display - > grab_op ) ;
g_assert ( gravity > = 0 ) ;
2008-12-18 07:41:56 -05:00
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
/* Do any edge resistance/snapping */
meta_window_edge_resistance_for_resize ( window ,
old . width ,
old . height ,
& new_w ,
& new_h ,
gravity ,
update_resize_timeout ,
snap ,
FALSE ) ;
2010-09-18 16:28:30 -04:00
if ( new_unmaximize = = window - > display - > grab_resize_unmaximize )
{
/* We don't need to update unless the specified width and height
* are actually different from what we had before .
*/
if ( old . width ! = new_w | | old . height ! = new_h )
{
2011-07-26 20:13:53 -04:00
meta_window_resize_with_gravity ( window , TRUE , new_w , new_h , gravity ) ;
2010-09-18 16:28:30 -04:00
}
}
else
{
if ( ( new_unmaximize & ~ window - > display - > grab_resize_unmaximize ) ! = 0 )
{
meta_window_unmaximize_with_gravity ( window ,
( new_unmaximize & ~ window - > display - > grab_resize_unmaximize ) ,
new_w , new_h , gravity ) ;
}
if ( ( window - > display - > grab_resize_unmaximize & ~ new_unmaximize ) )
{
MetaRectangle saved_rect = window - > saved_rect ;
meta_window_maximize ( window ,
( window - > display - > grab_resize_unmaximize & ~ new_unmaximize ) ) ;
window - > saved_rect = saved_rect ;
}
}
window - > display - > grab_resize_unmaximize = new_unmaximize ;
2002-12-09 22:23:04 -05:00
Fix bug 143333, support for update counter spec, and 109362, schedule
Sat Jun 19 02:21:08 2004 Soeren Sandmann <sandmann@daimi.au.dk>
Fix bug 143333, support for update counter spec, and 109362,
schedule compensation events when events are ignored.
* src/display.c (meta_display_open): Add _NET_WM_SYNC_REQUEST and
_NET_WM_SYNC_REQUEST_COUNTER atoms. Remove the old
METACITY_SYNC_COUNTER stuff.
(meta_display_begin_op): Setup the sync counter
* src/xprops.c, src/xprops.h, src/window-props.c, src/display.h:
Add new atoms.
* src/window.c (send_sync_request): new function.
(meta_window_move_resize_internal): send a sync request before
resizing.
(check_move_resize_frequence): Rework logic to also check the SYNC
case. If an event is ignored return the remaining time.
(update_resize_timeout): Timeout that gets called when a
compensation event is scheduled.
(uddate_resize): schedule compensation events when an event is
ignored.
(meta_window_handle_mouse_grap_op_event): When an alarm is
received and sync was turned off, turn it back on.
* src/window.h (struct MetaWindow) Add some variables
2004-06-18 20:45:24 -04:00
/* Store the latest resize time, if we actually resized. */
2006-01-02 13:48:30 -05:00
if ( window - > rect . width ! = old . width | | window - > rect . height ! = old . height )
g_get_current_time ( & window - > display - > grab_last_moveresize_time ) ;
2001-07-25 22:55:35 -04:00
}
2010-09-24 15:35:00 -04:00
static void
update_tile_mode ( MetaWindow * window )
{
switch ( window - > tile_mode )
{
case META_TILE_LEFT :
case META_TILE_RIGHT :
if ( ! META_WINDOW_TILED_SIDE_BY_SIDE ( window ) )
window - > tile_mode = META_TILE_NONE ;
2010-12-04 15:20:11 -05:00
break ;
2010-09-24 15:35:00 -04:00
case META_TILE_MAXIMIZED :
if ( ! META_WINDOW_MAXIMIZED ( window ) )
window - > tile_mode = META_TILE_NONE ;
break ;
}
}
2014-03-19 10:30:12 -04:00
void
meta_window_update_resize ( MetaWindow * window ,
gboolean snap ,
int x , int y ,
gboolean force )
{
update_resize ( window , snap , x , y , force ) ;
}
2012-11-15 17:18:06 -05:00
# ifdef HAVE_XSYNC
2001-07-25 22:55:35 -04:00
void
2012-11-15 17:18:06 -05:00
meta_window_update_sync_request_counter ( MetaWindow * window ,
2011-06-08 13:58:54 -04:00
gint64 new_counter_value )
2001-07-25 22:55:35 -04:00
{
2013-02-14 13:40:55 -05:00
gboolean needs_frame_drawn = FALSE ;
gboolean no_delay_frame = FALSE ;
if ( window - > extended_sync_request_counter & & new_counter_value % 2 = = 0 )
2012-11-15 17:49:57 -05:00
{
2013-02-14 13:40:55 -05:00
needs_frame_drawn = TRUE ;
no_delay_frame = new_counter_value = = window - > sync_request_serial + 1 ;
2012-11-15 17:49:57 -05:00
}
2011-06-13 18:09:59 -04:00
2011-06-08 13:58:54 -04:00
window - > sync_request_serial = new_counter_value ;
meta_compositor_set_updates_frozen ( window - > display - > compositor , window ,
meta_window_updates_are_frozen ( window ) ) ;
2013-01-30 14:49:31 -05:00
if ( window = = window - > display - > grab_window & &
meta_grab_op_is_resizing ( window - > display - > grab_op ) & &
2013-03-13 13:09:33 -04:00
new_counter_value > = window - > sync_request_wait_serial & &
2013-03-15 16:16:28 -04:00
( ! window - > extended_sync_request_counter | | new_counter_value % 2 = = 0 ) & &
window - > sync_request_timeout_id )
2001-07-25 22:55:35 -04:00
{
2002-12-09 22:23:04 -05:00
meta_topic ( META_DEBUG_RESIZING ,
" Alarm event received last motion x = %d y = %d \n " ,
window - > display - > grab_latest_motion_x ,
window - > display - > grab_latest_motion_y ) ;
Fix bug 143333, support for update counter spec, and 109362, schedule
Sat Jun 19 02:21:08 2004 Soeren Sandmann <sandmann@daimi.au.dk>
Fix bug 143333, support for update counter spec, and 109362,
schedule compensation events when events are ignored.
* src/display.c (meta_display_open): Add _NET_WM_SYNC_REQUEST and
_NET_WM_SYNC_REQUEST_COUNTER atoms. Remove the old
METACITY_SYNC_COUNTER stuff.
(meta_display_begin_op): Setup the sync counter
* src/xprops.c, src/xprops.h, src/window-props.c, src/display.h:
Add new atoms.
* src/window.c (send_sync_request): new function.
(meta_window_move_resize_internal): send a sync request before
resizing.
(check_move_resize_frequence): Rework logic to also check the SYNC
case. If an event is ignored return the remaining time.
(update_resize_timeout): Timeout that gets called when a
compensation event is scheduled.
(uddate_resize): schedule compensation events when an event is
ignored.
(meta_window_handle_mouse_grap_op_event): When an alarm is
received and sync was turned off, turn it back on.
* src/window.h (struct MetaWindow) Add some variables
2004-06-18 20:45:24 -04:00
2013-03-13 15:31:11 -04:00
g_source_remove ( window - > sync_request_timeout_id ) ;
window - > sync_request_timeout_id = 0 ;
2008-12-18 07:41:56 -05:00
2013-01-30 14:49:31 -05:00
/* This means we are ready for another configure;
* no pointer round trip here , to keep in sync */
update_resize ( window ,
window - > display - > grab_last_user_action_was_snap ,
window - > display - > grab_latest_motion_x ,
window - > display - > grab_latest_motion_y ,
TRUE ) ;
2002-12-09 22:23:04 -05:00
}
2013-02-14 13:40:55 -05:00
2013-03-15 16:16:28 -04:00
/* If sync was previously disabled, turn it back on and hope
* the application has come to its senses ( maybe it was just
* busy with a pagefault or a long computation ) .
*/
window - > disable_sync = FALSE ;
2013-02-14 13:40:55 -05:00
if ( needs_frame_drawn )
meta_compositor_queue_frame_drawn ( window - > display - > compositor , window ,
no_delay_frame ) ;
2012-11-15 17:18:06 -05:00
}
2002-12-09 22:23:04 -05:00
# endif /* HAVE_XSYNC */
2008-12-18 07:41:56 -05:00
2013-12-05 16:13:03 -05:00
void
meta_window_handle_mouse_grab_op_event ( MetaWindow * window ,
const ClutterEvent * event )
2012-11-15 17:18:06 -05:00
{
2013-10-03 15:29:44 -04:00
gboolean is_window_root = ( event - > any . stage ! = NULL & &
window & &
window - > screen & &
CLUTTER_ACTOR ( event - > any . stage ) = =
meta_get_stage_for_screen ( window - > screen ) ) ;
switch ( event - > type )
2002-12-09 22:23:04 -05:00
{
2013-10-03 15:29:44 -04:00
case CLUTTER_BUTTON_RELEASE :
2013-10-16 16:29:54 -04:00
if ( event - > button . button = = 1 | |
2013-11-19 14:03:09 -05:00
event - > button . button = = ( unsigned int ) meta_prefs_get_mouse_button_resize ( ) )
2013-07-23 12:42:22 -04:00
{
meta_display_check_threshold_reached ( window - > display ,
2013-10-03 15:29:44 -04:00
event - > button . x ,
event - > button . y ) ;
2013-07-23 12:42:22 -04:00
/* If the user was snap moving then ignore the button
* release because they may have let go of shift before
* releasing the mouse button and they almost certainly do
* not want a non - snapped movement to occur from the button
* release .
*/
if ( ! window - > display - > grab_last_user_action_was_snap )
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
{
2013-07-23 12:42:22 -04:00
if ( meta_grab_op_is_moving ( window - > display - > grab_op ) )
{
if ( window - > tile_mode ! = META_TILE_NONE )
meta_window_tile ( window ) ;
2013-10-03 15:29:44 -04:00
else if ( is_window_root )
2013-07-23 12:42:22 -04:00
update_move ( window ,
2013-10-03 15:29:44 -04:00
event - > button . modifier_state & CLUTTER_SHIFT_MASK ,
event - > button . x ,
event - > button . y ) ;
2013-07-23 12:42:22 -04:00
}
else if ( meta_grab_op_is_resizing ( window - > display - > grab_op ) )
{
2013-10-03 15:29:44 -04:00
if ( is_window_root )
2013-07-23 12:42:22 -04:00
update_resize ( window ,
2013-10-03 15:29:44 -04:00
event - > button . modifier_state & CLUTTER_SHIFT_MASK ,
event - > button . x ,
event - > button . y ,
2013-07-23 12:42:22 -04:00
TRUE ) ;
/* If a tiled window has been dragged free with a
* mouse resize without snapping back to the tiled
* state , it will end up with an inconsistent tile
* mode on mouse release ; cleaning the mode earlier
* would break the ability to snap back to the tiled
* state , so we wait until mouse release .
*/
update_tile_mode ( window ) ;
}
2013-10-03 15:29:44 -04:00
meta_display_end_grab_op ( window - > display , event - > any . time ) ;
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
}
2002-12-09 22:23:04 -05:00
}
2008-12-18 07:41:56 -05:00
break ;
2002-12-09 22:23:04 -05:00
2013-10-03 15:29:44 -04:00
case CLUTTER_MOTION :
Add a raise on click option, basically only because all the major distros
2006-01-10 Elijah Newren <newren@gmail.com>
Add a raise on click option, basically only because all the major
distros are patching it in anyway. See #326156.
* src/metacity.schemas.in: add the new gconf key and explanation
* src/prefs.[ch] (#define KEY_RAISE_ON_CLICK, static gboolean
raise_on_click, update_raise_on_click, meta_prefs_init,
change_notify, meta_prefs_get_raise_on_click,
meta_preference_to_string):
Add all the normal preference handling stuff for this new
raise-on-click option.
* src/core.c (meta_core_show_window_menu):
* src/display.c (event_callback, meta_display_begin_grab_op):
* src/window.c (window_activate, meta_window_configure_request, ):
Only raise the window if in raise_on_click mode.
* src/display.c (meta_display_begin_grab_op,
meta_display_end_grab_op, meta_display_check_threshold_reached):
* src/display.h (struct MetaDisplay):
* src/window.c (meta_window_handle_mouse_grab_op_event):
if not in raise-on-click mode only raise on button release if the
click didn't start a move or resize operation; needs a few extra
MetaDisplay fields to handle this
* src/core.c (meta_core_user_lower_and_unfocus):
no need to do the MRU shuffling if not maintaining the stacking
order == MRU order invariant
* src/frames.c (meta_frames_button_press_event):
* src/window.c (meta_window_begin_grab_op):
remove an unneeded window raising that is already handled elsewhere
2006-01-10 14:35:03 -05:00
meta_display_check_threshold_reached ( window - > display ,
2013-10-03 15:29:44 -04:00
event - > motion . x ,
event - > motion . y ) ;
2002-12-09 22:23:04 -05:00
if ( meta_grab_op_is_moving ( window - > display - > grab_op ) )
{
2013-10-03 15:29:44 -04:00
if ( is_window_root )
2003-01-11 14:32:12 -05:00
{
2013-12-05 16:13:03 -05:00
update_move ( window ,
event - > button . modifier_state & CLUTTER_SHIFT_MASK ,
event - > motion . x ,
event - > motion . y ) ;
2003-01-11 14:32:12 -05:00
}
2002-12-09 22:23:04 -05:00
}
else if ( meta_grab_op_is_resizing ( window - > display - > grab_op ) )
{
2013-10-03 15:29:44 -04:00
if ( is_window_root )
2003-01-11 14:32:12 -05:00
{
2013-12-05 16:13:03 -05:00
update_resize ( window ,
event - > button . modifier_state & CLUTTER_SHIFT_MASK ,
event - > motion . x ,
event - > motion . y ,
FALSE ) ;
2003-01-11 14:32:12 -05:00
}
2002-12-09 22:23:04 -05:00
}
2001-07-25 22:55:35 -04:00
break ;
2002-12-09 22:23:04 -05:00
2001-07-25 22:55:35 -04:00
default :
break ;
}
}
2003-02-23 12:09:46 -05:00
static void
2009-08-31 09:52:16 -04:00
get_work_area_monitor ( MetaWindow * window ,
MetaRectangle * area ,
int which_monitor )
2002-01-10 01:31:31 -05:00
{
2008-12-18 07:41:56 -05:00
GList * tmp ;
2009-08-31 09:52:16 -04:00
g_assert ( which_monitor > = 0 ) ;
2002-01-10 01:31:31 -05:00
2009-08-31 09:52:16 -04:00
/* Initialize to the whole monitor */
* area = window - > screen - > monitor_infos [ which_monitor ] . rect ;
2008-12-18 07:41:56 -05:00
tmp = meta_window_get_workspaces ( window ) ;
2002-01-10 01:31:31 -05:00
while ( tmp ! = NULL )
{
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
MetaRectangle workspace_work_area ;
2009-08-31 09:52:16 -04:00
meta_workspace_get_work_area_for_monitor ( tmp - > data ,
which_monitor ,
& workspace_work_area ) ;
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
meta_rectangle_intersect ( area ,
& workspace_work_area ,
area ) ;
2002-01-10 01:31:31 -05:00
tmp = tmp - > next ;
}
2008-12-18 07:41:56 -05:00
2002-02-07 22:34:26 -05:00
meta_topic ( META_DEBUG_WORKAREA ,
2009-08-31 09:52:16 -04:00
" Window %s monitor %d has work area %d,%d %d x %d \n " ,
window - > desc , which_monitor ,
2003-03-08 18:43:17 -05:00
area - > x , area - > y , area - > width , area - > height ) ;
2002-01-10 01:31:31 -05:00
}
2002-01-27 00:50:34 -05:00
2013-01-27 23:08:17 -05:00
/**
* meta_window_get_work_area_current_monitor :
* @ window : a # MetaWindow
* @ area : ( out ) : a location to store the work area
*
* Get the work area for the monitor @ window is currently on .
*/
2003-02-23 12:09:46 -05:00
void
2009-08-31 09:52:16 -04:00
meta_window_get_work_area_current_monitor ( MetaWindow * window ,
MetaRectangle * area )
2003-02-23 12:09:46 -05:00
{
2009-08-31 09:52:16 -04:00
const MetaMonitorInfo * monitor = NULL ;
monitor = meta_screen_get_monitor_for_window ( window - > screen ,
window ) ;
2003-02-23 12:09:46 -05:00
2009-08-31 09:52:16 -04:00
meta_window_get_work_area_for_monitor ( window ,
monitor - > number ,
area ) ;
2003-02-23 12:09:46 -05:00
}
2013-01-27 23:08:17 -05:00
/**
* meta_window_get_work_area_for_monitor :
* @ window : a # MetaWindow
* @ which_monitor : a moniotr to get the work area for
* @ area : ( out ) : a location to store the work area
*
* Get the work area for @ window , given the monitor index
* @ which_monitor .
*/
2003-02-23 12:09:46 -05:00
void
2009-08-31 09:52:16 -04:00
meta_window_get_work_area_for_monitor ( MetaWindow * window ,
int which_monitor ,
MetaRectangle * area )
2003-02-23 12:09:46 -05:00
{
2009-08-31 09:52:16 -04:00
g_return_if_fail ( which_monitor > = 0 ) ;
2008-12-18 07:41:56 -05:00
2009-08-31 09:52:16 -04:00
get_work_area_monitor ( window ,
area ,
which_monitor ) ;
2003-03-08 18:43:17 -05:00
}
2013-01-27 23:08:17 -05:00
/**
* meta_window_get_work_area_all_monitors :
* @ window : a # MetaWindow
* @ area : ( out ) : a location to store the work area
*
* Get the work area for all monitors for @ window .
*/
2003-03-08 18:43:17 -05:00
void
2009-08-31 09:52:16 -04:00
meta_window_get_work_area_all_monitors ( MetaWindow * window ,
MetaRectangle * area )
2003-03-08 18:43:17 -05:00
{
2008-12-18 07:41:56 -05:00
GList * tmp ;
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
/* Initialize to the whole screen */
* area = window - > screen - > rect ;
2008-12-18 07:41:56 -05:00
tmp = meta_window_get_workspaces ( window ) ;
2003-03-08 18:43:17 -05:00
while ( tmp ! = NULL )
{
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
MetaRectangle workspace_work_area ;
2009-08-31 09:52:16 -04:00
meta_workspace_get_work_area_all_monitors ( tmp - > data ,
& workspace_work_area ) ;
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
meta_rectangle_intersect ( area ,
& workspace_work_area ,
area ) ;
2003-03-08 18:43:17 -05:00
tmp = tmp - > next ;
}
meta_topic ( META_DEBUG_WORKAREA ,
" Window %s has whole-screen work area %d,%d %d x %d \n " ,
window - > desc , area - > x , area - > y , area - > width , area - > height ) ;
2003-02-23 12:09:46 -05:00
}
2012-03-24 06:29:39 -04:00
int
meta_window_get_current_tile_monitor_number ( MetaWindow * window )
{
int tile_monitor_number = window - > tile_monitor_number ;
if ( tile_monitor_number < 0 )
{
meta_warning ( " %s called with an invalid monitor number; using 0 instead \n " , G_STRFUNC ) ;
tile_monitor_number = 0 ;
}
return tile_monitor_number ;
}
2010-06-24 14:41:28 -04:00
void
meta_window_get_current_tile_area ( MetaWindow * window ,
MetaRectangle * tile_area )
{
2011-10-08 11:00:16 -04:00
int tile_monitor_number ;
2010-06-24 14:41:28 -04:00
g_return_if_fail ( window - > tile_mode ! = META_TILE_NONE ) ;
2012-03-24 06:29:39 -04:00
tile_monitor_number = meta_window_get_current_tile_monitor_number ( window ) ;
2011-10-08 11:00:16 -04:00
meta_window_get_work_area_for_monitor ( window , tile_monitor_number , tile_area ) ;
2010-06-24 14:41:28 -04:00
if ( window - > tile_mode = = META_TILE_LEFT | |
window - > tile_mode = = META_TILE_RIGHT )
tile_area - > width / = 2 ;
if ( window - > tile_mode = = META_TILE_RIGHT )
tile_area - > x + = tile_area - > width ;
}
2003-03-08 18:43:17 -05:00
2002-01-27 00:50:34 -05:00
gboolean
meta_window_same_application ( MetaWindow * window ,
MetaWindow * other_window )
{
2008-07-12 23:26:39 -04:00
MetaGroup * group = meta_window_get_group ( window ) ;
MetaGroup * other_group = meta_window_get_group ( other_window ) ;
2002-11-30 22:58:04 -05:00
return
2008-07-12 23:26:39 -04:00
group ! = NULL & &
other_group ! = NULL & &
group = = other_group ;
2002-01-27 00:50:34 -05:00
}
2002-03-17 12:22:23 -05:00
2014-03-05 19:01:11 -05:00
/**
* meta_window_is_client_decorated :
*
* Check if if the window has decorations drawn by the client .
* ( window - > decorated refers only to whether we should add decorations )
*/
gboolean
meta_window_is_client_decorated ( MetaWindow * window )
{
if ( window - > client_type = = META_WINDOW_CLIENT_TYPE_WAYLAND )
{
/* Assume all Wayland clients draw decorations - not strictly
* true but good enough for current purposes .
*/
return TRUE ;
}
else
{
/* Currently the implementation here is hackish -
* has_custom_frame_extents ( ) is set if _GTK_FRAME_EXTENTS is set
* to any value even 0. GTK + always sets _GTK_FRAME_EXTENTS for
* client - side - decorated window , even if the value is 0 because
* the window is maxized and has no invisible borders or shadows .
*/
return window - > has_custom_frame_extents ;
}
}
2002-03-17 12:22:23 -05:00
void
meta_window_refresh_resize_popup ( MetaWindow * window )
{
if ( window - > display - > grab_op = = META_GRAB_OP_NONE )
return ;
if ( window - > display - > grab_window ! = window )
return ;
2010-08-25 16:23:50 -04:00
switch ( window - > display - > grab_op )
{
case META_GRAB_OP_RESIZING_SE :
case META_GRAB_OP_RESIZING_S :
case META_GRAB_OP_RESIZING_SW :
case META_GRAB_OP_RESIZING_N :
case META_GRAB_OP_RESIZING_NE :
case META_GRAB_OP_RESIZING_NW :
case META_GRAB_OP_RESIZING_W :
case META_GRAB_OP_RESIZING_E :
case META_GRAB_OP_KEYBOARD_RESIZING_UNKNOWN :
case META_GRAB_OP_KEYBOARD_RESIZING_S :
case META_GRAB_OP_KEYBOARD_RESIZING_N :
case META_GRAB_OP_KEYBOARD_RESIZING_W :
case META_GRAB_OP_KEYBOARD_RESIZING_E :
case META_GRAB_OP_KEYBOARD_RESIZING_SE :
case META_GRAB_OP_KEYBOARD_RESIZING_NE :
case META_GRAB_OP_KEYBOARD_RESIZING_SW :
case META_GRAB_OP_KEYBOARD_RESIZING_NW :
break ;
default :
/* Not resizing */
return ;
}
2002-03-17 12:22:23 -05:00
if ( window - > display - > grab_resize_popup = = NULL )
{
2002-10-18 18:37:01 -04:00
if ( window - > size_hints . width_inc > 1 | |
window - > size_hints . height_inc > 1 )
window - > display - > grab_resize_popup =
meta_ui_resize_popup_new ( window - > display - > xdisplay ,
window - > screen - > number ) ;
2002-03-17 12:22:23 -05:00
}
2008-12-18 07:41:56 -05:00
2002-03-17 12:22:23 -05:00
if ( window - > display - > grab_resize_popup ! = NULL )
{
2006-03-29 11:05:23 -05:00
MetaRectangle rect ;
2008-12-18 07:41:56 -05:00
2009-05-07 13:53:47 -04:00
meta_window_get_client_root_coords ( window , & rect ) ;
2008-12-18 07:41:56 -05:00
2002-03-17 12:22:23 -05:00
meta_ui_resize_popup_set ( window - > display - > grab_resize_popup ,
2006-03-29 11:05:23 -05:00
rect ,
2002-03-17 12:22:23 -05:00
window - > size_hints . base_width ,
window - > size_hints . base_height ,
window - > size_hints . width_inc ,
Remove the unused attributes resize_gravity, width_inc, height_inc,
2006-05-18 Björn Lindqvist <bjourne@gmail.com>
* resizepopup.c: Remove the unused attributes resize_gravity,
width_inc, height_inc, min_width, min_height, frame_left,
frame_right, frame_top, frame_bottom, tick_origin_x, tick_origin_y
from the MetaResizePopup struct. Delete all code that references
those attributes.
2006-05-18 04:32:16 -04:00
window - > size_hints . height_inc ) ;
2002-03-17 12:22:23 -05:00
meta_ui_resize_popup_set_showing ( window - > display - > grab_resize_popup ,
TRUE ) ;
}
}
on unminimize, queue calc_showing on all transients
2002-05-05 Havoc Pennington <hp@pobox.com>
* src/window.c (meta_window_unminimize): on unminimize, queue
calc_showing on all transients
(meta_window_activate): on activate, unminimize all a window's
ancestors, not just the window itself.
* src/workspace.c (set_work_area_hint): don't increment "tmp" by
16 unsigned long, increment by 4
* src/window.c (meta_window_free): if a window isn't minimized,
restore its WM_STATE to NormalState instead of IconicState,
since IconicState on initial window map means that the window
should be minimized.
* src/workspace.c (meta_workspace_invalidate_work_area): queue an
idle to recompute the work area hint.
(set_work_area_hint): we need 4*num_workspaces ints, not just
num_workspaces.
* src/screen.c (meta_screen_new): add work_area_idle field,
handle it on screen shutdown
* src/common.h (META_PRIORITY_PREFS_NOTIFY,
META_PRIORITY_WORK_AREA_HINT): define some idle priorities
* src/window.c (meta_window_calc_showing): hide windows if
their parent window is minimized
(meta_window_minimize): also queue_calc_showing on all
transients of the window being minimized
* src/place.c (constrain_placement): function to apply
placement-time-only constraints, such as "not off the left of the
screen"
(meta_window_place): put dialogs down a bit over their parent,
not right at the top.
(meta_window_place): when centering a dialog, center it
on the current xinerama screen, rather than the entire
screen.
* src/screen.c (meta_screen_get_current_xinerama): new function,
but not implemented
2002-05-05 01:41:13 -04:00
2010-04-24 17:19:25 -04:00
/**
* meta_window_foreach_transient :
* @ window : a # MetaWindow
2010-09-24 11:26:36 -04:00
* @ func : ( scope call ) ( closure user_data ) : Called for each window which is a transient of @ window ( transitively )
* @ user_data : User data
2010-04-24 17:19:25 -04:00
*
* Call @ func for every window which is either transient for @ window , or is
* a transient of a window which is in turn transient for @ window .
* The order of window enumeration is not defined .
*
* Iteration will stop if @ func at any point returns % FALSE .
*/
on unminimize, queue calc_showing on all transients
2002-05-05 Havoc Pennington <hp@pobox.com>
* src/window.c (meta_window_unminimize): on unminimize, queue
calc_showing on all transients
(meta_window_activate): on activate, unminimize all a window's
ancestors, not just the window itself.
* src/workspace.c (set_work_area_hint): don't increment "tmp" by
16 unsigned long, increment by 4
* src/window.c (meta_window_free): if a window isn't minimized,
restore its WM_STATE to NormalState instead of IconicState,
since IconicState on initial window map means that the window
should be minimized.
* src/workspace.c (meta_workspace_invalidate_work_area): queue an
idle to recompute the work area hint.
(set_work_area_hint): we need 4*num_workspaces ints, not just
num_workspaces.
* src/screen.c (meta_screen_new): add work_area_idle field,
handle it on screen shutdown
* src/common.h (META_PRIORITY_PREFS_NOTIFY,
META_PRIORITY_WORK_AREA_HINT): define some idle priorities
* src/window.c (meta_window_calc_showing): hide windows if
their parent window is minimized
(meta_window_minimize): also queue_calc_showing on all
transients of the window being minimized
* src/place.c (constrain_placement): function to apply
placement-time-only constraints, such as "not off the left of the
screen"
(meta_window_place): put dialogs down a bit over their parent,
not right at the top.
(meta_window_place): when centering a dialog, center it
on the current xinerama screen, rather than the entire
screen.
* src/screen.c (meta_screen_get_current_xinerama): new function,
but not implemented
2002-05-05 01:41:13 -04:00
void
meta_window_foreach_transient ( MetaWindow * window ,
MetaWindowForeachFunc func ,
2010-04-24 17:19:25 -04:00
void * user_data )
on unminimize, queue calc_showing on all transients
2002-05-05 Havoc Pennington <hp@pobox.com>
* src/window.c (meta_window_unminimize): on unminimize, queue
calc_showing on all transients
(meta_window_activate): on activate, unminimize all a window's
ancestors, not just the window itself.
* src/workspace.c (set_work_area_hint): don't increment "tmp" by
16 unsigned long, increment by 4
* src/window.c (meta_window_free): if a window isn't minimized,
restore its WM_STATE to NormalState instead of IconicState,
since IconicState on initial window map means that the window
should be minimized.
* src/workspace.c (meta_workspace_invalidate_work_area): queue an
idle to recompute the work area hint.
(set_work_area_hint): we need 4*num_workspaces ints, not just
num_workspaces.
* src/screen.c (meta_screen_new): add work_area_idle field,
handle it on screen shutdown
* src/common.h (META_PRIORITY_PREFS_NOTIFY,
META_PRIORITY_WORK_AREA_HINT): define some idle priorities
* src/window.c (meta_window_calc_showing): hide windows if
their parent window is minimized
(meta_window_minimize): also queue_calc_showing on all
transients of the window being minimized
* src/place.c (constrain_placement): function to apply
placement-time-only constraints, such as "not off the left of the
screen"
(meta_window_place): put dialogs down a bit over their parent,
not right at the top.
(meta_window_place): when centering a dialog, center it
on the current xinerama screen, rather than the entire
screen.
* src/screen.c (meta_screen_get_current_xinerama): new function,
but not implemented
2002-05-05 01:41:13 -04:00
{
GSList * windows ;
GSList * tmp ;
2009-06-15 15:32:23 -04:00
windows = meta_display_list_windows ( window - > display , META_LIST_DEFAULT ) ;
on unminimize, queue calc_showing on all transients
2002-05-05 Havoc Pennington <hp@pobox.com>
* src/window.c (meta_window_unminimize): on unminimize, queue
calc_showing on all transients
(meta_window_activate): on activate, unminimize all a window's
ancestors, not just the window itself.
* src/workspace.c (set_work_area_hint): don't increment "tmp" by
16 unsigned long, increment by 4
* src/window.c (meta_window_free): if a window isn't minimized,
restore its WM_STATE to NormalState instead of IconicState,
since IconicState on initial window map means that the window
should be minimized.
* src/workspace.c (meta_workspace_invalidate_work_area): queue an
idle to recompute the work area hint.
(set_work_area_hint): we need 4*num_workspaces ints, not just
num_workspaces.
* src/screen.c (meta_screen_new): add work_area_idle field,
handle it on screen shutdown
* src/common.h (META_PRIORITY_PREFS_NOTIFY,
META_PRIORITY_WORK_AREA_HINT): define some idle priorities
* src/window.c (meta_window_calc_showing): hide windows if
their parent window is minimized
(meta_window_minimize): also queue_calc_showing on all
transients of the window being minimized
* src/place.c (constrain_placement): function to apply
placement-time-only constraints, such as "not off the left of the
screen"
(meta_window_place): put dialogs down a bit over their parent,
not right at the top.
(meta_window_place): when centering a dialog, center it
on the current xinerama screen, rather than the entire
screen.
* src/screen.c (meta_screen_get_current_xinerama): new function,
but not implemented
2002-05-05 01:41:13 -04:00
tmp = windows ;
while ( tmp ! = NULL )
{
MetaWindow * transient = tmp - > data ;
2008-12-18 07:41:56 -05:00
on unminimize, queue calc_showing on all transients
2002-05-05 Havoc Pennington <hp@pobox.com>
* src/window.c (meta_window_unminimize): on unminimize, queue
calc_showing on all transients
(meta_window_activate): on activate, unminimize all a window's
ancestors, not just the window itself.
* src/workspace.c (set_work_area_hint): don't increment "tmp" by
16 unsigned long, increment by 4
* src/window.c (meta_window_free): if a window isn't minimized,
restore its WM_STATE to NormalState instead of IconicState,
since IconicState on initial window map means that the window
should be minimized.
* src/workspace.c (meta_workspace_invalidate_work_area): queue an
idle to recompute the work area hint.
(set_work_area_hint): we need 4*num_workspaces ints, not just
num_workspaces.
* src/screen.c (meta_screen_new): add work_area_idle field,
handle it on screen shutdown
* src/common.h (META_PRIORITY_PREFS_NOTIFY,
META_PRIORITY_WORK_AREA_HINT): define some idle priorities
* src/window.c (meta_window_calc_showing): hide windows if
their parent window is minimized
(meta_window_minimize): also queue_calc_showing on all
transients of the window being minimized
* src/place.c (constrain_placement): function to apply
placement-time-only constraints, such as "not off the left of the
screen"
(meta_window_place): put dialogs down a bit over their parent,
not right at the top.
(meta_window_place): when centering a dialog, center it
on the current xinerama screen, rather than the entire
screen.
* src/screen.c (meta_screen_get_current_xinerama): new function,
but not implemented
2002-05-05 01:41:13 -04:00
if ( meta_window_is_ancestor_of_transient ( window , transient ) )
2003-03-17 01:36:39 -05:00
{
2010-04-24 17:19:25 -04:00
if ( ! ( * func ) ( transient , user_data ) )
2003-03-17 01:36:39 -05:00
break ;
}
2008-12-18 07:41:56 -05:00
on unminimize, queue calc_showing on all transients
2002-05-05 Havoc Pennington <hp@pobox.com>
* src/window.c (meta_window_unminimize): on unminimize, queue
calc_showing on all transients
(meta_window_activate): on activate, unminimize all a window's
ancestors, not just the window itself.
* src/workspace.c (set_work_area_hint): don't increment "tmp" by
16 unsigned long, increment by 4
* src/window.c (meta_window_free): if a window isn't minimized,
restore its WM_STATE to NormalState instead of IconicState,
since IconicState on initial window map means that the window
should be minimized.
* src/workspace.c (meta_workspace_invalidate_work_area): queue an
idle to recompute the work area hint.
(set_work_area_hint): we need 4*num_workspaces ints, not just
num_workspaces.
* src/screen.c (meta_screen_new): add work_area_idle field,
handle it on screen shutdown
* src/common.h (META_PRIORITY_PREFS_NOTIFY,
META_PRIORITY_WORK_AREA_HINT): define some idle priorities
* src/window.c (meta_window_calc_showing): hide windows if
their parent window is minimized
(meta_window_minimize): also queue_calc_showing on all
transients of the window being minimized
* src/place.c (constrain_placement): function to apply
placement-time-only constraints, such as "not off the left of the
screen"
(meta_window_place): put dialogs down a bit over their parent,
not right at the top.
(meta_window_place): when centering a dialog, center it
on the current xinerama screen, rather than the entire
screen.
* src/screen.c (meta_screen_get_current_xinerama): new function,
but not implemented
2002-05-05 01:41:13 -04:00
tmp = tmp - > next ;
}
g_slist_free ( windows ) ;
}
2010-04-24 17:19:25 -04:00
/**
* meta_window_foreach_ancestor :
* @ window : a # MetaWindow
2010-09-24 11:26:36 -04:00
* @ func : ( scope call ) ( closure user_data ) : Called for each window which is a transient parent of @ window
* @ user_data : User data
2010-04-24 17:19:25 -04:00
*
* If @ window is transient , call @ func with the window for which it ' s transient ,
* repeatedly until either we find a non - transient window , or @ func returns % FALSE .
*/
2003-03-17 01:36:39 -05:00
void
meta_window_foreach_ancestor ( MetaWindow * window ,
MetaWindowForeachFunc func ,
2010-04-24 17:19:25 -04:00
void * user_data )
on unminimize, queue calc_showing on all transients
2002-05-05 Havoc Pennington <hp@pobox.com>
* src/window.c (meta_window_unminimize): on unminimize, queue
calc_showing on all transients
(meta_window_activate): on activate, unminimize all a window's
ancestors, not just the window itself.
* src/workspace.c (set_work_area_hint): don't increment "tmp" by
16 unsigned long, increment by 4
* src/window.c (meta_window_free): if a window isn't minimized,
restore its WM_STATE to NormalState instead of IconicState,
since IconicState on initial window map means that the window
should be minimized.
* src/workspace.c (meta_workspace_invalidate_work_area): queue an
idle to recompute the work area hint.
(set_work_area_hint): we need 4*num_workspaces ints, not just
num_workspaces.
* src/screen.c (meta_screen_new): add work_area_idle field,
handle it on screen shutdown
* src/common.h (META_PRIORITY_PREFS_NOTIFY,
META_PRIORITY_WORK_AREA_HINT): define some idle priorities
* src/window.c (meta_window_calc_showing): hide windows if
their parent window is minimized
(meta_window_minimize): also queue_calc_showing on all
transients of the window being minimized
* src/place.c (constrain_placement): function to apply
placement-time-only constraints, such as "not off the left of the
screen"
(meta_window_place): put dialogs down a bit over their parent,
not right at the top.
(meta_window_place): when centering a dialog, center it
on the current xinerama screen, rather than the entire
screen.
* src/screen.c (meta_screen_get_current_xinerama): new function,
but not implemented
2002-05-05 01:41:13 -04:00
{
MetaWindow * w ;
2008-12-18 07:41:56 -05:00
2003-03-17 01:36:39 -05:00
w = window ;
2011-05-06 05:09:04 -04:00
do
2003-03-17 01:36:39 -05:00
{
2013-09-16 08:44:37 -04:00
if ( w - > transient_for = = NULL )
2003-03-17 01:36:39 -05:00
break ;
2008-12-18 07:41:56 -05:00
2013-09-16 08:44:37 -04:00
w = w - > transient_for ;
on unminimize, queue calc_showing on all transients
2002-05-05 Havoc Pennington <hp@pobox.com>
* src/window.c (meta_window_unminimize): on unminimize, queue
calc_showing on all transients
(meta_window_activate): on activate, unminimize all a window's
ancestors, not just the window itself.
* src/workspace.c (set_work_area_hint): don't increment "tmp" by
16 unsigned long, increment by 4
* src/window.c (meta_window_free): if a window isn't minimized,
restore its WM_STATE to NormalState instead of IconicState,
since IconicState on initial window map means that the window
should be minimized.
* src/workspace.c (meta_workspace_invalidate_work_area): queue an
idle to recompute the work area hint.
(set_work_area_hint): we need 4*num_workspaces ints, not just
num_workspaces.
* src/screen.c (meta_screen_new): add work_area_idle field,
handle it on screen shutdown
* src/common.h (META_PRIORITY_PREFS_NOTIFY,
META_PRIORITY_WORK_AREA_HINT): define some idle priorities
* src/window.c (meta_window_calc_showing): hide windows if
their parent window is minimized
(meta_window_minimize): also queue_calc_showing on all
transients of the window being minimized
* src/place.c (constrain_placement): function to apply
placement-time-only constraints, such as "not off the left of the
screen"
(meta_window_place): put dialogs down a bit over their parent,
not right at the top.
(meta_window_place): when centering a dialog, center it
on the current xinerama screen, rather than the entire
screen.
* src/screen.c (meta_screen_get_current_xinerama): new function,
but not implemented
2002-05-05 01:41:13 -04:00
}
2011-05-06 05:09:04 -04:00
while ( w & & ( * func ) ( w , user_data ) ) ;
2003-03-17 01:36:39 -05:00
}
on unminimize, queue calc_showing on all transients
2002-05-05 Havoc Pennington <hp@pobox.com>
* src/window.c (meta_window_unminimize): on unminimize, queue
calc_showing on all transients
(meta_window_activate): on activate, unminimize all a window's
ancestors, not just the window itself.
* src/workspace.c (set_work_area_hint): don't increment "tmp" by
16 unsigned long, increment by 4
* src/window.c (meta_window_free): if a window isn't minimized,
restore its WM_STATE to NormalState instead of IconicState,
since IconicState on initial window map means that the window
should be minimized.
* src/workspace.c (meta_workspace_invalidate_work_area): queue an
idle to recompute the work area hint.
(set_work_area_hint): we need 4*num_workspaces ints, not just
num_workspaces.
* src/screen.c (meta_screen_new): add work_area_idle field,
handle it on screen shutdown
* src/common.h (META_PRIORITY_PREFS_NOTIFY,
META_PRIORITY_WORK_AREA_HINT): define some idle priorities
* src/window.c (meta_window_calc_showing): hide windows if
their parent window is minimized
(meta_window_minimize): also queue_calc_showing on all
transients of the window being minimized
* src/place.c (constrain_placement): function to apply
placement-time-only constraints, such as "not off the left of the
screen"
(meta_window_place): put dialogs down a bit over their parent,
not right at the top.
(meta_window_place): when centering a dialog, center it
on the current xinerama screen, rather than the entire
screen.
* src/screen.c (meta_screen_get_current_xinerama): new function,
but not implemented
2002-05-05 01:41:13 -04:00
2003-03-17 01:36:39 -05:00
typedef struct
{
MetaWindow * ancestor ;
gboolean found ;
} FindAncestorData ;
static gboolean
find_ancestor_func ( MetaWindow * window ,
void * data )
{
FindAncestorData * d = data ;
if ( window = = d - > ancestor )
{
d - > found = TRUE ;
return FALSE ;
}
return TRUE ;
}
2009-08-04 03:32:36 -04:00
/**
* meta_window_is_ancestor_of_transient :
* @ window : a # MetaWindow
* @ transient : a # MetaWindow
*
* The function determines whether @ window is an ancestor of @ transient ; it does
* so by traversing the @ transient ' s ancestors until it either locates @ window
* or reaches an ancestor that is not transient .
*
* Return Value : ( transfer none ) : % TRUE if window is an ancestor of transient .
*/
2003-03-17 01:36:39 -05:00
gboolean
meta_window_is_ancestor_of_transient ( MetaWindow * window ,
MetaWindow * transient )
{
FindAncestorData d ;
d . ancestor = window ;
d . found = FALSE ;
meta_window_foreach_ancestor ( transient , find_ancestor_func , & d ) ;
return d . found ;
on unminimize, queue calc_showing on all transients
2002-05-05 Havoc Pennington <hp@pobox.com>
* src/window.c (meta_window_unminimize): on unminimize, queue
calc_showing on all transients
(meta_window_activate): on activate, unminimize all a window's
ancestors, not just the window itself.
* src/workspace.c (set_work_area_hint): don't increment "tmp" by
16 unsigned long, increment by 4
* src/window.c (meta_window_free): if a window isn't minimized,
restore its WM_STATE to NormalState instead of IconicState,
since IconicState on initial window map means that the window
should be minimized.
* src/workspace.c (meta_workspace_invalidate_work_area): queue an
idle to recompute the work area hint.
(set_work_area_hint): we need 4*num_workspaces ints, not just
num_workspaces.
* src/screen.c (meta_screen_new): add work_area_idle field,
handle it on screen shutdown
* src/common.h (META_PRIORITY_PREFS_NOTIFY,
META_PRIORITY_WORK_AREA_HINT): define some idle priorities
* src/window.c (meta_window_calc_showing): hide windows if
their parent window is minimized
(meta_window_minimize): also queue_calc_showing on all
transients of the window being minimized
* src/place.c (constrain_placement): function to apply
placement-time-only constraints, such as "not off the left of the
screen"
(meta_window_place): put dialogs down a bit over their parent,
not right at the top.
(meta_window_place): when centering a dialog, center it
on the current xinerama screen, rather than the entire
screen.
* src/screen.c (meta_screen_get_current_xinerama): new function,
but not implemented
2002-05-05 01:41:13 -04:00
}
Patch to provide extra cues to the user when using window menu move and
2002-08-08 Craig Black <blackc@speakeasy.net>
Patch to provide extra cues to the user when using
window menu move and resize, #85724.
* src/common.h: add new cursors
* src/display.c: (grab_op_is_mouse)
(meta_display_create_x_cursor), (xcursor_for_op),
(meta_display_set_grab_op_cursor),
(meta_display_begin_grab_op):
The keyboard move and resize grab ops now also use the mouse.
Allow the grab cursor to be changed during the grab op.
Hold onto the initial grab position in case of reset.
* src/display.h: save the initial grab position
* src/keybindings.c: (process_keyboard_move_grab),
(process_keyboard_resize_grab), (handle_begin_move),
(handle_begin_resize):
The keyboard move and resize grab ops now also use the mouse.
* src/window.c: (meta_window_client_message), (menu_callback),
(update_move), (update_resize),
(meta_window_handle_mouse_grab_op_event), (warp_pointer),
(meta_window_warp_pointer), (meta_window_begin_grab_op),
(meta_window_update_resize_grab_op):
When moving or resizing a window use the last grab position
in computing change increment.
Provide support for warping the mouse pointer.
* src/window.h: new warp pointer and grab op helper functions
2002-08-09 00:27:23 -04:00
2003-10-12 02:25:38 -04:00
/* Warp pointer to location appropriate for grab,
* return root coordinates where pointer ended up .
*/
Patch to provide extra cues to the user when using window menu move and
2002-08-08 Craig Black <blackc@speakeasy.net>
Patch to provide extra cues to the user when using
window menu move and resize, #85724.
* src/common.h: add new cursors
* src/display.c: (grab_op_is_mouse)
(meta_display_create_x_cursor), (xcursor_for_op),
(meta_display_set_grab_op_cursor),
(meta_display_begin_grab_op):
The keyboard move and resize grab ops now also use the mouse.
Allow the grab cursor to be changed during the grab op.
Hold onto the initial grab position in case of reset.
* src/display.h: save the initial grab position
* src/keybindings.c: (process_keyboard_move_grab),
(process_keyboard_resize_grab), (handle_begin_move),
(handle_begin_resize):
The keyboard move and resize grab ops now also use the mouse.
* src/window.c: (meta_window_client_message), (menu_callback),
(update_move), (update_resize),
(meta_window_handle_mouse_grab_op_event), (warp_pointer),
(meta_window_warp_pointer), (meta_window_begin_grab_op),
(meta_window_update_resize_grab_op):
When moving or resizing a window use the last grab position
in computing change increment.
Provide support for warping the mouse pointer.
* src/window.h: new warp pointer and grab op helper functions
2002-08-09 00:27:23 -04:00
static gboolean
2003-10-12 02:25:38 -04:00
warp_grab_pointer ( MetaWindow * window ,
MetaGrabOp grab_op ,
int * x ,
int * y )
Patch to provide extra cues to the user when using window menu move and
2002-08-08 Craig Black <blackc@speakeasy.net>
Patch to provide extra cues to the user when using
window menu move and resize, #85724.
* src/common.h: add new cursors
* src/display.c: (grab_op_is_mouse)
(meta_display_create_x_cursor), (xcursor_for_op),
(meta_display_set_grab_op_cursor),
(meta_display_begin_grab_op):
The keyboard move and resize grab ops now also use the mouse.
Allow the grab cursor to be changed during the grab op.
Hold onto the initial grab position in case of reset.
* src/display.h: save the initial grab position
* src/keybindings.c: (process_keyboard_move_grab),
(process_keyboard_resize_grab), (handle_begin_move),
(handle_begin_resize):
The keyboard move and resize grab ops now also use the mouse.
* src/window.c: (meta_window_client_message), (menu_callback),
(update_move), (update_resize),
(meta_window_handle_mouse_grab_op_event), (warp_pointer),
(meta_window_warp_pointer), (meta_window_begin_grab_op),
(meta_window_update_resize_grab_op):
When moving or resizing a window use the last grab position
in computing change increment.
Provide support for warping the mouse pointer.
* src/window.h: new warp pointer and grab op helper functions
2002-08-09 00:27:23 -04:00
{
2007-04-08 23:43:55 -04:00
MetaRectangle rect ;
MetaDisplay * display ;
display = window - > display ;
2003-10-12 02:25:38 -04:00
/* We may not have done begin_grab_op yet, i.e. may not be in a grab
*/
2008-12-18 07:41:56 -05:00
2013-11-19 12:24:45 -05:00
meta_window_get_frame_rect ( window , & rect ) ;
2008-12-18 07:41:56 -05:00
Patch to provide extra cues to the user when using window menu move and
2002-08-08 Craig Black <blackc@speakeasy.net>
Patch to provide extra cues to the user when using
window menu move and resize, #85724.
* src/common.h: add new cursors
* src/display.c: (grab_op_is_mouse)
(meta_display_create_x_cursor), (xcursor_for_op),
(meta_display_set_grab_op_cursor),
(meta_display_begin_grab_op):
The keyboard move and resize grab ops now also use the mouse.
Allow the grab cursor to be changed during the grab op.
Hold onto the initial grab position in case of reset.
* src/display.h: save the initial grab position
* src/keybindings.c: (process_keyboard_move_grab),
(process_keyboard_resize_grab), (handle_begin_move),
(handle_begin_resize):
The keyboard move and resize grab ops now also use the mouse.
* src/window.c: (meta_window_client_message), (menu_callback),
(update_move), (update_resize),
(meta_window_handle_mouse_grab_op_event), (warp_pointer),
(meta_window_warp_pointer), (meta_window_begin_grab_op),
(meta_window_update_resize_grab_op):
When moving or resizing a window use the last grab position
in computing change increment.
Provide support for warping the mouse pointer.
* src/window.h: new warp pointer and grab op helper functions
2002-08-09 00:27:23 -04:00
switch ( grab_op )
{
case META_GRAB_OP_KEYBOARD_MOVING :
case META_GRAB_OP_KEYBOARD_RESIZING_UNKNOWN :
2003-10-12 02:25:38 -04:00
* x = rect . width / 2 ;
* y = rect . height / 2 ;
Patch to provide extra cues to the user when using window menu move and
2002-08-08 Craig Black <blackc@speakeasy.net>
Patch to provide extra cues to the user when using
window menu move and resize, #85724.
* src/common.h: add new cursors
* src/display.c: (grab_op_is_mouse)
(meta_display_create_x_cursor), (xcursor_for_op),
(meta_display_set_grab_op_cursor),
(meta_display_begin_grab_op):
The keyboard move and resize grab ops now also use the mouse.
Allow the grab cursor to be changed during the grab op.
Hold onto the initial grab position in case of reset.
* src/display.h: save the initial grab position
* src/keybindings.c: (process_keyboard_move_grab),
(process_keyboard_resize_grab), (handle_begin_move),
(handle_begin_resize):
The keyboard move and resize grab ops now also use the mouse.
* src/window.c: (meta_window_client_message), (menu_callback),
(update_move), (update_resize),
(meta_window_handle_mouse_grab_op_event), (warp_pointer),
(meta_window_warp_pointer), (meta_window_begin_grab_op),
(meta_window_update_resize_grab_op):
When moving or resizing a window use the last grab position
in computing change increment.
Provide support for warping the mouse pointer.
* src/window.h: new warp pointer and grab op helper functions
2002-08-09 00:27:23 -04:00
break ;
case META_GRAB_OP_KEYBOARD_RESIZING_S :
2003-10-12 02:25:38 -04:00
* x = rect . width / 2 ;
2008-02-17 19:21:50 -05:00
* y = rect . height - 1 ;
Patch to provide extra cues to the user when using window menu move and
2002-08-08 Craig Black <blackc@speakeasy.net>
Patch to provide extra cues to the user when using
window menu move and resize, #85724.
* src/common.h: add new cursors
* src/display.c: (grab_op_is_mouse)
(meta_display_create_x_cursor), (xcursor_for_op),
(meta_display_set_grab_op_cursor),
(meta_display_begin_grab_op):
The keyboard move and resize grab ops now also use the mouse.
Allow the grab cursor to be changed during the grab op.
Hold onto the initial grab position in case of reset.
* src/display.h: save the initial grab position
* src/keybindings.c: (process_keyboard_move_grab),
(process_keyboard_resize_grab), (handle_begin_move),
(handle_begin_resize):
The keyboard move and resize grab ops now also use the mouse.
* src/window.c: (meta_window_client_message), (menu_callback),
(update_move), (update_resize),
(meta_window_handle_mouse_grab_op_event), (warp_pointer),
(meta_window_warp_pointer), (meta_window_begin_grab_op),
(meta_window_update_resize_grab_op):
When moving or resizing a window use the last grab position
in computing change increment.
Provide support for warping the mouse pointer.
* src/window.h: new warp pointer and grab op helper functions
2002-08-09 00:27:23 -04:00
break ;
case META_GRAB_OP_KEYBOARD_RESIZING_N :
2003-10-12 02:25:38 -04:00
* x = rect . width / 2 ;
Patch to provide extra cues to the user when using window menu move and
2002-08-08 Craig Black <blackc@speakeasy.net>
Patch to provide extra cues to the user when using
window menu move and resize, #85724.
* src/common.h: add new cursors
* src/display.c: (grab_op_is_mouse)
(meta_display_create_x_cursor), (xcursor_for_op),
(meta_display_set_grab_op_cursor),
(meta_display_begin_grab_op):
The keyboard move and resize grab ops now also use the mouse.
Allow the grab cursor to be changed during the grab op.
Hold onto the initial grab position in case of reset.
* src/display.h: save the initial grab position
* src/keybindings.c: (process_keyboard_move_grab),
(process_keyboard_resize_grab), (handle_begin_move),
(handle_begin_resize):
The keyboard move and resize grab ops now also use the mouse.
* src/window.c: (meta_window_client_message), (menu_callback),
(update_move), (update_resize),
(meta_window_handle_mouse_grab_op_event), (warp_pointer),
(meta_window_warp_pointer), (meta_window_begin_grab_op),
(meta_window_update_resize_grab_op):
When moving or resizing a window use the last grab position
in computing change increment.
Provide support for warping the mouse pointer.
* src/window.h: new warp pointer and grab op helper functions
2002-08-09 00:27:23 -04:00
* y = 0 ;
break ;
case META_GRAB_OP_KEYBOARD_RESIZING_W :
* x = 0 ;
2003-10-12 02:25:38 -04:00
* y = rect . height / 2 ;
Patch to provide extra cues to the user when using window menu move and
2002-08-08 Craig Black <blackc@speakeasy.net>
Patch to provide extra cues to the user when using
window menu move and resize, #85724.
* src/common.h: add new cursors
* src/display.c: (grab_op_is_mouse)
(meta_display_create_x_cursor), (xcursor_for_op),
(meta_display_set_grab_op_cursor),
(meta_display_begin_grab_op):
The keyboard move and resize grab ops now also use the mouse.
Allow the grab cursor to be changed during the grab op.
Hold onto the initial grab position in case of reset.
* src/display.h: save the initial grab position
* src/keybindings.c: (process_keyboard_move_grab),
(process_keyboard_resize_grab), (handle_begin_move),
(handle_begin_resize):
The keyboard move and resize grab ops now also use the mouse.
* src/window.c: (meta_window_client_message), (menu_callback),
(update_move), (update_resize),
(meta_window_handle_mouse_grab_op_event), (warp_pointer),
(meta_window_warp_pointer), (meta_window_begin_grab_op),
(meta_window_update_resize_grab_op):
When moving or resizing a window use the last grab position
in computing change increment.
Provide support for warping the mouse pointer.
* src/window.h: new warp pointer and grab op helper functions
2002-08-09 00:27:23 -04:00
break ;
case META_GRAB_OP_KEYBOARD_RESIZING_E :
2008-02-17 19:21:50 -05:00
* x = rect . width - 1 ;
2003-10-12 02:25:38 -04:00
* y = rect . height / 2 ;
Patch to provide extra cues to the user when using window menu move and
2002-08-08 Craig Black <blackc@speakeasy.net>
Patch to provide extra cues to the user when using
window menu move and resize, #85724.
* src/common.h: add new cursors
* src/display.c: (grab_op_is_mouse)
(meta_display_create_x_cursor), (xcursor_for_op),
(meta_display_set_grab_op_cursor),
(meta_display_begin_grab_op):
The keyboard move and resize grab ops now also use the mouse.
Allow the grab cursor to be changed during the grab op.
Hold onto the initial grab position in case of reset.
* src/display.h: save the initial grab position
* src/keybindings.c: (process_keyboard_move_grab),
(process_keyboard_resize_grab), (handle_begin_move),
(handle_begin_resize):
The keyboard move and resize grab ops now also use the mouse.
* src/window.c: (meta_window_client_message), (menu_callback),
(update_move), (update_resize),
(meta_window_handle_mouse_grab_op_event), (warp_pointer),
(meta_window_warp_pointer), (meta_window_begin_grab_op),
(meta_window_update_resize_grab_op):
When moving or resizing a window use the last grab position
in computing change increment.
Provide support for warping the mouse pointer.
* src/window.h: new warp pointer and grab op helper functions
2002-08-09 00:27:23 -04:00
break ;
case META_GRAB_OP_KEYBOARD_RESIZING_SE :
2008-02-17 19:21:50 -05:00
* x = rect . width - 1 ;
* y = rect . height - 1 ;
Patch to provide extra cues to the user when using window menu move and
2002-08-08 Craig Black <blackc@speakeasy.net>
Patch to provide extra cues to the user when using
window menu move and resize, #85724.
* src/common.h: add new cursors
* src/display.c: (grab_op_is_mouse)
(meta_display_create_x_cursor), (xcursor_for_op),
(meta_display_set_grab_op_cursor),
(meta_display_begin_grab_op):
The keyboard move and resize grab ops now also use the mouse.
Allow the grab cursor to be changed during the grab op.
Hold onto the initial grab position in case of reset.
* src/display.h: save the initial grab position
* src/keybindings.c: (process_keyboard_move_grab),
(process_keyboard_resize_grab), (handle_begin_move),
(handle_begin_resize):
The keyboard move and resize grab ops now also use the mouse.
* src/window.c: (meta_window_client_message), (menu_callback),
(update_move), (update_resize),
(meta_window_handle_mouse_grab_op_event), (warp_pointer),
(meta_window_warp_pointer), (meta_window_begin_grab_op),
(meta_window_update_resize_grab_op):
When moving or resizing a window use the last grab position
in computing change increment.
Provide support for warping the mouse pointer.
* src/window.h: new warp pointer and grab op helper functions
2002-08-09 00:27:23 -04:00
break ;
case META_GRAB_OP_KEYBOARD_RESIZING_NE :
2008-02-17 19:21:50 -05:00
* x = rect . width - 1 ;
Patch to provide extra cues to the user when using window menu move and
2002-08-08 Craig Black <blackc@speakeasy.net>
Patch to provide extra cues to the user when using
window menu move and resize, #85724.
* src/common.h: add new cursors
* src/display.c: (grab_op_is_mouse)
(meta_display_create_x_cursor), (xcursor_for_op),
(meta_display_set_grab_op_cursor),
(meta_display_begin_grab_op):
The keyboard move and resize grab ops now also use the mouse.
Allow the grab cursor to be changed during the grab op.
Hold onto the initial grab position in case of reset.
* src/display.h: save the initial grab position
* src/keybindings.c: (process_keyboard_move_grab),
(process_keyboard_resize_grab), (handle_begin_move),
(handle_begin_resize):
The keyboard move and resize grab ops now also use the mouse.
* src/window.c: (meta_window_client_message), (menu_callback),
(update_move), (update_resize),
(meta_window_handle_mouse_grab_op_event), (warp_pointer),
(meta_window_warp_pointer), (meta_window_begin_grab_op),
(meta_window_update_resize_grab_op):
When moving or resizing a window use the last grab position
in computing change increment.
Provide support for warping the mouse pointer.
* src/window.h: new warp pointer and grab op helper functions
2002-08-09 00:27:23 -04:00
* y = 0 ;
break ;
case META_GRAB_OP_KEYBOARD_RESIZING_SW :
* x = 0 ;
2008-02-17 19:21:50 -05:00
* y = rect . height - 1 ;
Patch to provide extra cues to the user when using window menu move and
2002-08-08 Craig Black <blackc@speakeasy.net>
Patch to provide extra cues to the user when using
window menu move and resize, #85724.
* src/common.h: add new cursors
* src/display.c: (grab_op_is_mouse)
(meta_display_create_x_cursor), (xcursor_for_op),
(meta_display_set_grab_op_cursor),
(meta_display_begin_grab_op):
The keyboard move and resize grab ops now also use the mouse.
Allow the grab cursor to be changed during the grab op.
Hold onto the initial grab position in case of reset.
* src/display.h: save the initial grab position
* src/keybindings.c: (process_keyboard_move_grab),
(process_keyboard_resize_grab), (handle_begin_move),
(handle_begin_resize):
The keyboard move and resize grab ops now also use the mouse.
* src/window.c: (meta_window_client_message), (menu_callback),
(update_move), (update_resize),
(meta_window_handle_mouse_grab_op_event), (warp_pointer),
(meta_window_warp_pointer), (meta_window_begin_grab_op),
(meta_window_update_resize_grab_op):
When moving or resizing a window use the last grab position
in computing change increment.
Provide support for warping the mouse pointer.
* src/window.h: new warp pointer and grab op helper functions
2002-08-09 00:27:23 -04:00
break ;
case META_GRAB_OP_KEYBOARD_RESIZING_NW :
* x = 0 ;
* y = 0 ;
break ;
default :
return FALSE ;
}
2003-10-12 02:25:38 -04:00
* x + = rect . x ;
* y + = rect . y ;
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
/* Avoid weird bouncing at the screen edge; see bug 154706 */
* x = CLAMP ( * x , 0 , window - > screen - > rect . width - 1 ) ;
* y = CLAMP ( * y , 0 , window - > screen - > rect . height - 1 ) ;
2008-12-18 07:41:56 -05:00
2007-04-08 23:43:55 -04:00
meta_error_trap_push_with_return ( display ) ;
2003-10-12 02:25:38 -04:00
meta_topic ( META_DEBUG_WINDOW_OPS ,
" Warping pointer to %d,%d with window at %d,%d \n " ,
* x , * y , rect . x , rect . y ) ;
Merge of all the changes on the constraints_experiments branch. This is
2005-11-18 Elijah Newren <newren@gmail.com>
Merge of all the changes on the constraints_experiments branch.
This is just a summary, to get the full ChangeLog of those
changes (approx. 2000 lines):
cvs -q -z3 update -Pd -r constraints_experiments
cvs -q -z3 diff -pu -r CONSTRAINTS_EXPERIMENTS_BRANCHPOINT ChangeLog
Bugs fixed:
unfiled - constraints.c is overly complicated[1]
unfiled - constraints.c is not robust when all constraints
cannot simultaneously be met (constraints need to be
prioritized)
unfiled - keep-titlebar-onscreen constraint is decoration
unaware (since get_outermost_onscreen_positions()
forgets to include decorations)
unfiled - keyboard snap-moving and snap-resizing snap to hidden
edges
109553 - gravity w/ simultaneous move & resize doesn't work
113601 - maximize vertical and horizontal should toggle and be
constrained
122196 - windows show up under vertical panels
122670 - jerky/random resizing of window via keyboard[2]
124582 - keyboard and mouse snap-resizing and snap-moving
erroneously moves the window multidimensionally
136307 - don't allow apps to resize themselves off the screen
(*cough* filechooser *cough*)
142016, 143784 - windows should not span multiple xineramas
unless placed there by the user
143145 - clamp new windows to screensize and force them
onscreen, if they'll fit
144126 - Handle pathological strut lists sanely[3]
149867 - fixed aspect ratio windows are difficult to resize[4]
152898 - make screen edges consistent; allow easy slamming of
windows into the left, right, and bottom edges of the
screen too.
154706 - bouncing weirdness at screen edge with keyboard moving
or resizing
156699 - avoid struts when placing windows, if possible (nasty
a11y blocker)
302456 - dragging offscreen too restrictive
304857 - wireframe moving off the top of the screen is misleading
308521 - make uni-directional resizing easier with
alt-middle-drag and prevent the occasional super
annoying resize-the-wrong-side(s) behavior
312007 - snap-resize moves windows with a minimum size
constraint
312104 - resizing the top of a window can cause the bottom to
grow
319351 - don't instantly snap on mouse-move-snapping, remove
braindeadedness of having order of releasing shift and
releasing button press matter so much
[1] fixed in my opinion, anyway.
[2] Actually, it's not totally fixed--it's just annoying
instead of almost completely unusable. Matthias had a
suggestion that may fix the remainder of the problems (see
http://tinyurl.com/bwzuu).
[3] This bug was originally about not-quite-so-pathological
cases but was left open for the worse cases. The code from
the branch handles the remainder of the cases mentioned in
this bug.
[4] Actually, although it's far better there's still some minor
issues left: a slight drift that's only noticeable after
lots of resizing, and potential problems with partially
onscreen constraints due to not clearing any
fixed_directions flags (aspect ratio windows get resized in
both directions and thus aren't fixed in one of them)
New feature:
81704 - edge resistance for user move and resize operations;
in particular 3 different kinds of resistance are
implemented:
Pixel-Distance: window movement is resisted when it
aligns with an edge unless the movement is greater than
a threshold number of pixels
Timeout: window movement past an edge is prevented until
a certain amount of time has elapsed during the
operation since the first request to move it past that
edge
Keyboard-Buildup: when moving or resizing with the
keyboard, once a window is aligned with a certain edge
it cannot move past until the correct direction has
been pressed enough times (e.g. 2 or 3 times)
Major changes:
- constraints.c has been rewritten; very few lines of code from
the old version remain. There is a comment near the top of
the function explaining the basics of how the new framework
works. A more detailed explanation can be found in
doc/how-constraints-works.txt
- edge-resistance.[ch] are new files implementing edge-resistance.
- boxes.[ch] are new files containing low-level error-prone
functions used heavily in constraints.c and edge-resistance.c,
among various places throughout the code. testboxes.c
contains a thorough testsuite for the boxes.[ch] functions
compiled into a program, testboxes.
- meta_window_move_resize_internal() *must* be told the gravity
of the associated operation (if it's just a move operation,
the gravity will be ignored, but for resize and move+resize
the correct value is needed)
- the craziness of different values that
meta_window_move_resize_internal() accepts has been documented
in a large comment at the beginning of the function. It may
be possible to clean this up some, but until then things will
remain as they were before--caller beware.
- screen and xinerama usable areas (i.e. places not covered by
e.g. panels) are cached in the workspace now, as are the
screen and xinerama edges. These get updated with the
workarea in src/workspace.c:ensure_work_areas_validated()
2005-11-19 09:58:50 -05:00
/* Need to update the grab positions so that the MotionNotify and other
* events generated by the XWarpPointer ( ) call below don ' t cause complete
* funkiness . See bug 124582 and bug 122670.
*/
2007-04-08 23:43:55 -04:00
display - > grab_anchor_root_x = * x ;
display - > grab_anchor_root_y = * y ;
display - > grab_latest_motion_x = * x ;
display - > grab_latest_motion_y = * y ;
2009-05-07 13:53:47 -04:00
meta_window_get_client_root_coords ( window ,
& display - > grab_anchor_window_pos ) ;
2008-12-18 07:41:56 -05:00
2012-12-16 21:47:11 -05:00
XIWarpPointer ( display - > xdisplay ,
META_VIRTUAL_CORE_POINTER_ID ,
None ,
window - > screen - > xroot ,
0 , 0 , 0 , 0 ,
* x , * y ) ;
Patch to provide extra cues to the user when using window menu move and
2002-08-08 Craig Black <blackc@speakeasy.net>
Patch to provide extra cues to the user when using
window menu move and resize, #85724.
* src/common.h: add new cursors
* src/display.c: (grab_op_is_mouse)
(meta_display_create_x_cursor), (xcursor_for_op),
(meta_display_set_grab_op_cursor),
(meta_display_begin_grab_op):
The keyboard move and resize grab ops now also use the mouse.
Allow the grab cursor to be changed during the grab op.
Hold onto the initial grab position in case of reset.
* src/display.h: save the initial grab position
* src/keybindings.c: (process_keyboard_move_grab),
(process_keyboard_resize_grab), (handle_begin_move),
(handle_begin_resize):
The keyboard move and resize grab ops now also use the mouse.
* src/window.c: (meta_window_client_message), (menu_callback),
(update_move), (update_resize),
(meta_window_handle_mouse_grab_op_event), (warp_pointer),
(meta_window_warp_pointer), (meta_window_begin_grab_op),
(meta_window_update_resize_grab_op):
When moving or resizing a window use the last grab position
in computing change increment.
Provide support for warping the mouse pointer.
* src/window.h: new warp pointer and grab op helper functions
2002-08-09 00:27:23 -04:00
2010-10-25 14:44:30 -04:00
if ( meta_error_trap_pop_with_return ( display ) ! = Success )
Patch to provide extra cues to the user when using window menu move and
2002-08-08 Craig Black <blackc@speakeasy.net>
Patch to provide extra cues to the user when using
window menu move and resize, #85724.
* src/common.h: add new cursors
* src/display.c: (grab_op_is_mouse)
(meta_display_create_x_cursor), (xcursor_for_op),
(meta_display_set_grab_op_cursor),
(meta_display_begin_grab_op):
The keyboard move and resize grab ops now also use the mouse.
Allow the grab cursor to be changed during the grab op.
Hold onto the initial grab position in case of reset.
* src/display.h: save the initial grab position
* src/keybindings.c: (process_keyboard_move_grab),
(process_keyboard_resize_grab), (handle_begin_move),
(handle_begin_resize):
The keyboard move and resize grab ops now also use the mouse.
* src/window.c: (meta_window_client_message), (menu_callback),
(update_move), (update_resize),
(meta_window_handle_mouse_grab_op_event), (warp_pointer),
(meta_window_warp_pointer), (meta_window_begin_grab_op),
(meta_window_update_resize_grab_op):
When moving or resizing a window use the last grab position
in computing change increment.
Provide support for warping the mouse pointer.
* src/window.h: new warp pointer and grab op helper functions
2002-08-09 00:27:23 -04:00
{
2003-10-12 02:25:38 -04:00
meta_verbose ( " Failed to warp pointer for window %s \n " ,
window - > desc ) ;
Patch to provide extra cues to the user when using window menu move and
2002-08-08 Craig Black <blackc@speakeasy.net>
Patch to provide extra cues to the user when using
window menu move and resize, #85724.
* src/common.h: add new cursors
* src/display.c: (grab_op_is_mouse)
(meta_display_create_x_cursor), (xcursor_for_op),
(meta_display_set_grab_op_cursor),
(meta_display_begin_grab_op):
The keyboard move and resize grab ops now also use the mouse.
Allow the grab cursor to be changed during the grab op.
Hold onto the initial grab position in case of reset.
* src/display.h: save the initial grab position
* src/keybindings.c: (process_keyboard_move_grab),
(process_keyboard_resize_grab), (handle_begin_move),
(handle_begin_resize):
The keyboard move and resize grab ops now also use the mouse.
* src/window.c: (meta_window_client_message), (menu_callback),
(update_move), (update_resize),
(meta_window_handle_mouse_grab_op_event), (warp_pointer),
(meta_window_warp_pointer), (meta_window_begin_grab_op),
(meta_window_update_resize_grab_op):
When moving or resizing a window use the last grab position
in computing change increment.
Provide support for warping the mouse pointer.
* src/window.h: new warp pointer and grab op helper functions
2002-08-09 00:27:23 -04:00
return FALSE ;
}
2008-12-18 07:41:56 -05:00
Patch to provide extra cues to the user when using window menu move and
2002-08-08 Craig Black <blackc@speakeasy.net>
Patch to provide extra cues to the user when using
window menu move and resize, #85724.
* src/common.h: add new cursors
* src/display.c: (grab_op_is_mouse)
(meta_display_create_x_cursor), (xcursor_for_op),
(meta_display_set_grab_op_cursor),
(meta_display_begin_grab_op):
The keyboard move and resize grab ops now also use the mouse.
Allow the grab cursor to be changed during the grab op.
Hold onto the initial grab position in case of reset.
* src/display.h: save the initial grab position
* src/keybindings.c: (process_keyboard_move_grab),
(process_keyboard_resize_grab), (handle_begin_move),
(handle_begin_resize):
The keyboard move and resize grab ops now also use the mouse.
* src/window.c: (meta_window_client_message), (menu_callback),
(update_move), (update_resize),
(meta_window_handle_mouse_grab_op_event), (warp_pointer),
(meta_window_warp_pointer), (meta_window_begin_grab_op),
(meta_window_update_resize_grab_op):
When moving or resizing a window use the last grab position
in computing change increment.
Provide support for warping the mouse pointer.
* src/window.h: new warp pointer and grab op helper functions
2002-08-09 00:27:23 -04:00
return TRUE ;
}
void
meta_window_begin_grab_op ( MetaWindow * window ,
MetaGrabOp op ,
2006-10-12 20:28:34 -04:00
gboolean frame_action ,
Fix issues on 64-bit machines with timestamps by using guint32 (like gtk+
2006-09-13 Elijah Newren <newren gmail com>
* src/common.h (MetaWindowMenuFunc):
* src/core.[ch] (meta_core_user_lower_and_unfocus,
meta_core_user_focus, meta_core_show_window_menu,
meta_core_begin_grab_op, meta_core_end_grab_op):
* src/delete.c (delete_ping_reply_func, delete_ping_timeout_func,
meta_window_delete):
* src/display.[ch] (struct MetaDisplay, struct MetaPingData,
sanity_check_timestamps, meta_display_open, event_callback,
meta_spew_event, meta_display_set_grab_op_cursor,
meta_display_begin_grab_op, meta_display_end_grab_op,
meta_display_ping_timeout, meta_display_ping_window,
process_pong_message, timestamp_too_old,
meta_display_set_input_focus_window):
* src/keybindings.[ch] (grab_keyboard, ungrab_keyboard,
meta_screen_grab_all_keys, meta_window_grab_all_keys,
meta_window_ungrab_all_keys, error_on_generic_command,
error_on_command, error_on_terminal_command):
* src/metacity-dialog.c (on_realize, warn_about_no_sm_support,
error_about_command, main):
* src/screen.[ch] (struct _MetaScreen, meta_screen_new,
meta_screen_show_desktop, meta_screen_apply_startup_properties):
* src/session.c (warn_about_lame_clients_and_finish_interact):
* src/window.[ch] (struct _MetaWindow,
intervening_user_event_occurred, window_activate,
meta_window_delete, meta_window_focus,
meta_window_send_icccm_message, meta_window_client_message,
menu_callback, meta_window_show_menu, struct EventScannerData,
check_use_this_motion_notify, meta_window_begin_grab_op,
meta_window_set_user_time):
* src/workspace.[ch] (focus_ancestor_or_mru_window,
meta_workspace_activate_with_focus, meta_workspace_activate,
meta_workspace_focus_default_window,
focus_ancestor_or_mru_window):
Fix issues on 64-bit machines with timestamps by using guint32
(like gtk+ does) instead of Time. #348305
2006-09-13 12:32:33 -04:00
guint32 timestamp )
Patch to provide extra cues to the user when using window menu move and
2002-08-08 Craig Black <blackc@speakeasy.net>
Patch to provide extra cues to the user when using
window menu move and resize, #85724.
* src/common.h: add new cursors
* src/display.c: (grab_op_is_mouse)
(meta_display_create_x_cursor), (xcursor_for_op),
(meta_display_set_grab_op_cursor),
(meta_display_begin_grab_op):
The keyboard move and resize grab ops now also use the mouse.
Allow the grab cursor to be changed during the grab op.
Hold onto the initial grab position in case of reset.
* src/display.h: save the initial grab position
* src/keybindings.c: (process_keyboard_move_grab),
(process_keyboard_resize_grab), (handle_begin_move),
(handle_begin_resize):
The keyboard move and resize grab ops now also use the mouse.
* src/window.c: (meta_window_client_message), (menu_callback),
(update_move), (update_resize),
(meta_window_handle_mouse_grab_op_event), (warp_pointer),
(meta_window_warp_pointer), (meta_window_begin_grab_op),
(meta_window_update_resize_grab_op):
When moving or resizing a window use the last grab position
in computing change increment.
Provide support for warping the mouse pointer.
* src/window.h: new warp pointer and grab op helper functions
2002-08-09 00:27:23 -04:00
{
2003-10-12 02:25:38 -04:00
int x , y ;
Patch to provide extra cues to the user when using window menu move and
2002-08-08 Craig Black <blackc@speakeasy.net>
Patch to provide extra cues to the user when using
window menu move and resize, #85724.
* src/common.h: add new cursors
* src/display.c: (grab_op_is_mouse)
(meta_display_create_x_cursor), (xcursor_for_op),
(meta_display_set_grab_op_cursor),
(meta_display_begin_grab_op):
The keyboard move and resize grab ops now also use the mouse.
Allow the grab cursor to be changed during the grab op.
Hold onto the initial grab position in case of reset.
* src/display.h: save the initial grab position
* src/keybindings.c: (process_keyboard_move_grab),
(process_keyboard_resize_grab), (handle_begin_move),
(handle_begin_resize):
The keyboard move and resize grab ops now also use the mouse.
* src/window.c: (meta_window_client_message), (menu_callback),
(update_move), (update_resize),
(meta_window_handle_mouse_grab_op_event), (warp_pointer),
(meta_window_warp_pointer), (meta_window_begin_grab_op),
(meta_window_update_resize_grab_op):
When moving or resizing a window use the last grab position
in computing change increment.
Provide support for warping the mouse pointer.
* src/window.h: new warp pointer and grab op helper functions
2002-08-09 00:27:23 -04:00
2003-10-12 02:25:38 -04:00
warp_grab_pointer ( window ,
op , & x , & y ) ;
Patch to provide extra cues to the user when using window menu move and
2002-08-08 Craig Black <blackc@speakeasy.net>
Patch to provide extra cues to the user when using
window menu move and resize, #85724.
* src/common.h: add new cursors
* src/display.c: (grab_op_is_mouse)
(meta_display_create_x_cursor), (xcursor_for_op),
(meta_display_set_grab_op_cursor),
(meta_display_begin_grab_op):
The keyboard move and resize grab ops now also use the mouse.
Allow the grab cursor to be changed during the grab op.
Hold onto the initial grab position in case of reset.
* src/display.h: save the initial grab position
* src/keybindings.c: (process_keyboard_move_grab),
(process_keyboard_resize_grab), (handle_begin_move),
(handle_begin_resize):
The keyboard move and resize grab ops now also use the mouse.
* src/window.c: (meta_window_client_message), (menu_callback),
(update_move), (update_resize),
(meta_window_handle_mouse_grab_op_event), (warp_pointer),
(meta_window_warp_pointer), (meta_window_begin_grab_op),
(meta_window_update_resize_grab_op):
When moving or resizing a window use the last grab position
in computing change increment.
Provide support for warping the mouse pointer.
* src/window.h: new warp pointer and grab op helper functions
2002-08-09 00:27:23 -04:00
meta_display_begin_grab_op ( window - > display ,
window - > screen ,
window ,
op ,
2003-11-24 13:09:47 -05:00
FALSE ,
2006-10-12 20:28:34 -04:00
frame_action ,
2003-11-24 13:09:47 -05:00
0 /* button */ ,
0 ,
Patch to provide extra cues to the user when using window menu move and
2002-08-08 Craig Black <blackc@speakeasy.net>
Patch to provide extra cues to the user when using
window menu move and resize, #85724.
* src/common.h: add new cursors
* src/display.c: (grab_op_is_mouse)
(meta_display_create_x_cursor), (xcursor_for_op),
(meta_display_set_grab_op_cursor),
(meta_display_begin_grab_op):
The keyboard move and resize grab ops now also use the mouse.
Allow the grab cursor to be changed during the grab op.
Hold onto the initial grab position in case of reset.
* src/display.h: save the initial grab position
* src/keybindings.c: (process_keyboard_move_grab),
(process_keyboard_resize_grab), (handle_begin_move),
(handle_begin_resize):
The keyboard move and resize grab ops now also use the mouse.
* src/window.c: (meta_window_client_message), (menu_callback),
(update_move), (update_resize),
(meta_window_handle_mouse_grab_op_event), (warp_pointer),
(meta_window_warp_pointer), (meta_window_begin_grab_op),
(meta_window_update_resize_grab_op):
When moving or resizing a window use the last grab position
in computing change increment.
Provide support for warping the mouse pointer.
* src/window.h: new warp pointer and grab op helper functions
2002-08-09 00:27:23 -04:00
timestamp ,
2003-10-12 02:25:38 -04:00
x , y ) ;
Patch to provide extra cues to the user when using window menu move and
2002-08-08 Craig Black <blackc@speakeasy.net>
Patch to provide extra cues to the user when using
window menu move and resize, #85724.
* src/common.h: add new cursors
* src/display.c: (grab_op_is_mouse)
(meta_display_create_x_cursor), (xcursor_for_op),
(meta_display_set_grab_op_cursor),
(meta_display_begin_grab_op):
The keyboard move and resize grab ops now also use the mouse.
Allow the grab cursor to be changed during the grab op.
Hold onto the initial grab position in case of reset.
* src/display.h: save the initial grab position
* src/keybindings.c: (process_keyboard_move_grab),
(process_keyboard_resize_grab), (handle_begin_move),
(handle_begin_resize):
The keyboard move and resize grab ops now also use the mouse.
* src/window.c: (meta_window_client_message), (menu_callback),
(update_move), (update_resize),
(meta_window_handle_mouse_grab_op_event), (warp_pointer),
(meta_window_warp_pointer), (meta_window_begin_grab_op),
(meta_window_update_resize_grab_op):
When moving or resizing a window use the last grab position
in computing change increment.
Provide support for warping the mouse pointer.
* src/window.h: new warp pointer and grab op helper functions
2002-08-09 00:27:23 -04:00
}
2002-09-24 17:22:06 -04:00
void
2003-10-12 02:25:38 -04:00
meta_window_update_keyboard_resize ( MetaWindow * window ,
gboolean update_cursor )
Patch to provide extra cues to the user when using window menu move and
2002-08-08 Craig Black <blackc@speakeasy.net>
Patch to provide extra cues to the user when using
window menu move and resize, #85724.
* src/common.h: add new cursors
* src/display.c: (grab_op_is_mouse)
(meta_display_create_x_cursor), (xcursor_for_op),
(meta_display_set_grab_op_cursor),
(meta_display_begin_grab_op):
The keyboard move and resize grab ops now also use the mouse.
Allow the grab cursor to be changed during the grab op.
Hold onto the initial grab position in case of reset.
* src/display.h: save the initial grab position
* src/keybindings.c: (process_keyboard_move_grab),
(process_keyboard_resize_grab), (handle_begin_move),
(handle_begin_resize):
The keyboard move and resize grab ops now also use the mouse.
* src/window.c: (meta_window_client_message), (menu_callback),
(update_move), (update_resize),
(meta_window_handle_mouse_grab_op_event), (warp_pointer),
(meta_window_warp_pointer), (meta_window_begin_grab_op),
(meta_window_update_resize_grab_op):
When moving or resizing a window use the last grab position
in computing change increment.
Provide support for warping the mouse pointer.
* src/window.h: new warp pointer and grab op helper functions
2002-08-09 00:27:23 -04:00
{
2003-10-12 02:25:38 -04:00
int x , y ;
2008-12-18 07:41:56 -05:00
2003-10-12 02:25:38 -04:00
warp_grab_pointer ( window ,
window - > display - > grab_op ,
& x , & y ) ;
Patch to provide extra cues to the user when using window menu move and
2002-08-08 Craig Black <blackc@speakeasy.net>
Patch to provide extra cues to the user when using
window menu move and resize, #85724.
* src/common.h: add new cursors
* src/display.c: (grab_op_is_mouse)
(meta_display_create_x_cursor), (xcursor_for_op),
(meta_display_set_grab_op_cursor),
(meta_display_begin_grab_op):
The keyboard move and resize grab ops now also use the mouse.
Allow the grab cursor to be changed during the grab op.
Hold onto the initial grab position in case of reset.
* src/display.h: save the initial grab position
* src/keybindings.c: (process_keyboard_move_grab),
(process_keyboard_resize_grab), (handle_begin_move),
(handle_begin_resize):
The keyboard move and resize grab ops now also use the mouse.
* src/window.c: (meta_window_client_message), (menu_callback),
(update_move), (update_resize),
(meta_window_handle_mouse_grab_op_event), (warp_pointer),
(meta_window_warp_pointer), (meta_window_begin_grab_op),
(meta_window_update_resize_grab_op):
When moving or resizing a window use the last grab position
in computing change increment.
Provide support for warping the mouse pointer.
* src/window.h: new warp pointer and grab op helper functions
2002-08-09 00:27:23 -04:00
if ( update_cursor )
{
meta_display_set_grab_op_cursor ( window - > display ,
2012-11-20 17:08:31 -05:00
window - > screen ,
Patch to provide extra cues to the user when using window menu move and
2002-08-08 Craig Black <blackc@speakeasy.net>
Patch to provide extra cues to the user when using
window menu move and resize, #85724.
* src/common.h: add new cursors
* src/display.c: (grab_op_is_mouse)
(meta_display_create_x_cursor), (xcursor_for_op),
(meta_display_set_grab_op_cursor),
(meta_display_begin_grab_op):
The keyboard move and resize grab ops now also use the mouse.
Allow the grab cursor to be changed during the grab op.
Hold onto the initial grab position in case of reset.
* src/display.h: save the initial grab position
* src/keybindings.c: (process_keyboard_move_grab),
(process_keyboard_resize_grab), (handle_begin_move),
(handle_begin_resize):
The keyboard move and resize grab ops now also use the mouse.
* src/window.c: (meta_window_client_message), (menu_callback),
(update_move), (update_resize),
(meta_window_handle_mouse_grab_op_event), (warp_pointer),
(meta_window_warp_pointer), (meta_window_begin_grab_op),
(meta_window_update_resize_grab_op):
When moving or resizing a window use the last grab position
in computing change increment.
Provide support for warping the mouse pointer.
* src/window.h: new warp pointer and grab op helper functions
2002-08-09 00:27:23 -04:00
window - > display - > grab_op ,
window - > display - > grab_xwindow ,
2012-12-13 17:06:00 -05:00
window - > display - > grab_timestamp ) ;
Patch to provide extra cues to the user when using window menu move and
2002-08-08 Craig Black <blackc@speakeasy.net>
Patch to provide extra cues to the user when using
window menu move and resize, #85724.
* src/common.h: add new cursors
* src/display.c: (grab_op_is_mouse)
(meta_display_create_x_cursor), (xcursor_for_op),
(meta_display_set_grab_op_cursor),
(meta_display_begin_grab_op):
The keyboard move and resize grab ops now also use the mouse.
Allow the grab cursor to be changed during the grab op.
Hold onto the initial grab position in case of reset.
* src/display.h: save the initial grab position
* src/keybindings.c: (process_keyboard_move_grab),
(process_keyboard_resize_grab), (handle_begin_move),
(handle_begin_resize):
The keyboard move and resize grab ops now also use the mouse.
* src/window.c: (meta_window_client_message), (menu_callback),
(update_move), (update_resize),
(meta_window_handle_mouse_grab_op_event), (warp_pointer),
(meta_window_warp_pointer), (meta_window_begin_grab_op),
(meta_window_update_resize_grab_op):
When moving or resizing a window use the last grab position
in computing change increment.
Provide support for warping the mouse pointer.
* src/window.h: new warp pointer and grab op helper functions
2002-08-09 00:27:23 -04:00
}
}
2003-10-12 02:25:38 -04:00
void
meta_window_update_keyboard_move ( MetaWindow * window )
{
int x , y ;
2008-12-18 07:41:56 -05:00
2003-10-12 02:25:38 -04:00
warp_grab_pointer ( window ,
window - > display - > grab_op ,
& x , & y ) ;
}
2002-09-24 17:22:06 -04:00
void
meta_window_update_layer ( MetaWindow * window )
{
MetaGroup * group ;
2008-12-18 07:41:56 -05:00
2002-09-24 17:22:06 -04:00
meta_stack_freeze ( window - > screen - > stack ) ;
group = meta_window_get_group ( window ) ;
2008-11-03 09:50:22 -05:00
if ( group )
2002-09-24 17:22:06 -04:00
meta_group_update_layers ( group ) ;
else
meta_stack_update_layer ( window - > screen - > stack , window ) ;
meta_stack_thaw ( window - > screen - > stack ) ;
}
Add support for _NET_WM_USER_TIME
2004-06-17 Elijah Newren <newren@math.utah.edu>
Add support for _NET_WM_USER_TIME
* src/display.c:
(meta_display_open): Add _NET_WM_USER_TIME to atom_names[],
(event_callback): Manually set _NET_WM_USER_TIME upon KeyPress
(doesn't work since keyboard isn't grabbed) and ButtonPress (does
work), this is just a fallback for applications that don't update
this themselves.
* src/display.h: (struct _MetaDisplay): Add atom_net_wm_user_time field
* src/screen.c: (meta_screen_apply_startup_properties): Check for
TIMESTAMP provided from startup sequence as well.
* src/stack.c:
s/meta_window_set_stack_position/meta_window_set_stack_position_no_sync/,
(meta_window_set_stack_position): New function which calls the
meta_window_set_stack_position_no_sync function followed immediately
by calling meta_stack_sync_to_server.
* src/window-props.c:
(init_net_wm_user_time), (reload_net_wm_user_time): new functions,
(reload_wm_hints): also load atom_net_wm_user_time
* src/window.c:
new XSERVER_TIME_IS_LATER macro (accounts for timestamp wraparound),
(meta_window_new_with_attrs): add timestamp attributes,
(window_takes_focus_on_map): use TIMESTAMP from startup
notification and _NET_WM_USER_TIME to decide whether to focus new
windows,
(meta_window_show): if app doesn't take focus on map, place it
just below the focused window in the stack
(process_property_notify): check for changes to _NET_WM_USRE_TIME,
(meta_window_stack_just_below): new function
* src/window.h:
(_MetaWindow struct): new fields for initial_timestamp,
initial_timestamp_set, net_wm_user_time_set, and net_wm_user_time,
(meta_window_stack_just_below): new function
2004-06-24 11:47:05 -04:00
2004-08-05 13:48:39 -04:00
/* ensure_mru_position_after ensures that window appears after
* below_this_one in the active_workspace ' s mru_list ( i . e . it treats
* window as having been less recently used than below_this_one )
*/
static void
ensure_mru_position_after ( MetaWindow * window ,
MetaWindow * after_this_one )
{
/* This is sort of slow since it runs through the entire list more
* than once ( especially considering the fact that we expect the
* windows of interest to be the first two elements in the list ) ,
* but it doesn ' t matter while we ' re only using it on new window
* map .
*/
GList * active_mru_list ;
GList * window_position ;
GList * after_this_one_position ;
active_mru_list = window - > screen - > active_workspace - > mru_list ;
window_position = g_list_find ( active_mru_list , window ) ;
after_this_one_position = g_list_find ( active_mru_list , after_this_one ) ;
/* after_this_one_position is NULL when we switch workspaces, but in
* that case we don ' t need to do any MRU shuffling so we can simply
* return .
*/
if ( after_this_one_position = = NULL )
return ;
if ( g_list_length ( window_position ) > g_list_length ( after_this_one_position ) )
{
window - > screen - > active_workspace - > mru_list =
g_list_delete_link ( window - > screen - > active_workspace - > mru_list ,
window_position ) ;
window - > screen - > active_workspace - > mru_list =
g_list_insert_before ( window - > screen - > active_workspace - > mru_list ,
after_this_one_position - > next ,
window ) ;
}
}
Add support for _NET_WM_USER_TIME
2004-06-17 Elijah Newren <newren@math.utah.edu>
Add support for _NET_WM_USER_TIME
* src/display.c:
(meta_display_open): Add _NET_WM_USER_TIME to atom_names[],
(event_callback): Manually set _NET_WM_USER_TIME upon KeyPress
(doesn't work since keyboard isn't grabbed) and ButtonPress (does
work), this is just a fallback for applications that don't update
this themselves.
* src/display.h: (struct _MetaDisplay): Add atom_net_wm_user_time field
* src/screen.c: (meta_screen_apply_startup_properties): Check for
TIMESTAMP provided from startup sequence as well.
* src/stack.c:
s/meta_window_set_stack_position/meta_window_set_stack_position_no_sync/,
(meta_window_set_stack_position): New function which calls the
meta_window_set_stack_position_no_sync function followed immediately
by calling meta_stack_sync_to_server.
* src/window-props.c:
(init_net_wm_user_time), (reload_net_wm_user_time): new functions,
(reload_wm_hints): also load atom_net_wm_user_time
* src/window.c:
new XSERVER_TIME_IS_LATER macro (accounts for timestamp wraparound),
(meta_window_new_with_attrs): add timestamp attributes,
(window_takes_focus_on_map): use TIMESTAMP from startup
notification and _NET_WM_USER_TIME to decide whether to focus new
windows,
(meta_window_show): if app doesn't take focus on map, place it
just below the focused window in the stack
(process_property_notify): check for changes to _NET_WM_USRE_TIME,
(meta_window_stack_just_below): new function
* src/window.h:
(_MetaWindow struct): new fields for initial_timestamp,
initial_timestamp_set, net_wm_user_time_set, and net_wm_user_time,
(meta_window_stack_just_below): new function
2004-06-24 11:47:05 -04:00
void
meta_window_stack_just_below ( MetaWindow * window ,
MetaWindow * below_this_one )
{
g_return_if_fail ( window ! = NULL ) ;
g_return_if_fail ( below_this_one ! = NULL ) ;
if ( window - > stack_position > below_this_one - > stack_position )
{
meta_topic ( META_DEBUG_STACK ,
" Setting stack position of window %s to %d (making it below window %s). \n " ,
2008-12-18 07:41:56 -05:00
window - > desc ,
below_this_one - > stack_position ,
Add support for _NET_WM_USER_TIME
2004-06-17 Elijah Newren <newren@math.utah.edu>
Add support for _NET_WM_USER_TIME
* src/display.c:
(meta_display_open): Add _NET_WM_USER_TIME to atom_names[],
(event_callback): Manually set _NET_WM_USER_TIME upon KeyPress
(doesn't work since keyboard isn't grabbed) and ButtonPress (does
work), this is just a fallback for applications that don't update
this themselves.
* src/display.h: (struct _MetaDisplay): Add atom_net_wm_user_time field
* src/screen.c: (meta_screen_apply_startup_properties): Check for
TIMESTAMP provided from startup sequence as well.
* src/stack.c:
s/meta_window_set_stack_position/meta_window_set_stack_position_no_sync/,
(meta_window_set_stack_position): New function which calls the
meta_window_set_stack_position_no_sync function followed immediately
by calling meta_stack_sync_to_server.
* src/window-props.c:
(init_net_wm_user_time), (reload_net_wm_user_time): new functions,
(reload_wm_hints): also load atom_net_wm_user_time
* src/window.c:
new XSERVER_TIME_IS_LATER macro (accounts for timestamp wraparound),
(meta_window_new_with_attrs): add timestamp attributes,
(window_takes_focus_on_map): use TIMESTAMP from startup
notification and _NET_WM_USER_TIME to decide whether to focus new
windows,
(meta_window_show): if app doesn't take focus on map, place it
just below the focused window in the stack
(process_property_notify): check for changes to _NET_WM_USRE_TIME,
(meta_window_stack_just_below): new function
* src/window.h:
(_MetaWindow struct): new fields for initial_timestamp,
initial_timestamp_set, net_wm_user_time_set, and net_wm_user_time,
(meta_window_stack_just_below): new function
2004-06-24 11:47:05 -04:00
below_this_one - > desc ) ;
2004-08-27 13:22:54 -04:00
meta_window_set_stack_position ( window , below_this_one - > stack_position ) ;
Add support for _NET_WM_USER_TIME
2004-06-17 Elijah Newren <newren@math.utah.edu>
Add support for _NET_WM_USER_TIME
* src/display.c:
(meta_display_open): Add _NET_WM_USER_TIME to atom_names[],
(event_callback): Manually set _NET_WM_USER_TIME upon KeyPress
(doesn't work since keyboard isn't grabbed) and ButtonPress (does
work), this is just a fallback for applications that don't update
this themselves.
* src/display.h: (struct _MetaDisplay): Add atom_net_wm_user_time field
* src/screen.c: (meta_screen_apply_startup_properties): Check for
TIMESTAMP provided from startup sequence as well.
* src/stack.c:
s/meta_window_set_stack_position/meta_window_set_stack_position_no_sync/,
(meta_window_set_stack_position): New function which calls the
meta_window_set_stack_position_no_sync function followed immediately
by calling meta_stack_sync_to_server.
* src/window-props.c:
(init_net_wm_user_time), (reload_net_wm_user_time): new functions,
(reload_wm_hints): also load atom_net_wm_user_time
* src/window.c:
new XSERVER_TIME_IS_LATER macro (accounts for timestamp wraparound),
(meta_window_new_with_attrs): add timestamp attributes,
(window_takes_focus_on_map): use TIMESTAMP from startup
notification and _NET_WM_USER_TIME to decide whether to focus new
windows,
(meta_window_show): if app doesn't take focus on map, place it
just below the focused window in the stack
(process_property_notify): check for changes to _NET_WM_USRE_TIME,
(meta_window_stack_just_below): new function
* src/window.h:
(_MetaWindow struct): new fields for initial_timestamp,
initial_timestamp_set, net_wm_user_time_set, and net_wm_user_time,
(meta_window_stack_just_below): new function
2004-06-24 11:47:05 -04:00
}
else
{
meta_topic ( META_DEBUG_STACK ,
" Window %s was already below window %s. \n " ,
window - > desc , below_this_one - > desc ) ;
}
}
Big patch to cover about 6 different issues in order to correct rare
2005-02-20 Elijah Newren <newren@gmail.com>
Big patch to cover about 6 different issues in order to correct
rare problems with timestamps (make sure window selected in
tasklist actually gets focus, sanity check timestamps to avoid
rogue apps hosing the system, correct the updating of
net_wm_user_time, correctly handle timestamps of 0 when comparing
xserver timestamps for those who have had their systems up for
over 25 days or so, add some debugging information to verbose
logs, some code cleanups). Fixes all issues listed in #167358.
* src/display.h: (struct _MetaDisplay): clarify comment on
last_focus_time, introduce a new variable--last_user_time,
(XSERVER_TIME_IS_BEFORE macro): put this functionality into a
separate macro and then introduce a new macro with this name that
uses the old one but adds additional special-case checks for
timestamps that are 0, (comment to
meta_display_set_input_focus_window): add information about how
last_user_time should be used in this function
* src/display.c (santiy_check_timestamps): new function,
(meta_display_open): intialize display->last_user_time,
(meta_display_get_current_time_roundtrip): use the timestamp,
which is known to be good, in order to sanity_check_timestamps,
(event_callback): use the new meta_window_ste_user_time() function
in order to correct problems, use the timestamp of KeyPress and
ButtonPress events, which are known to be good, in order to
sanity_check_timestamps, (timestamp_too_old): new function for
common behavior of meta_display_focus_the_no_focus_window and
meta_display_set_input_focus_window, with added checking for
display->last_user_time in addition to display->last_focus_time,
(meta_display_set_input_focus_window): replace some of the code
with a call to timestamp_too_old(),
(meta_display_focus_the_no_focus_window): replace some of th ecode
with a call to timestamp_too_old()
* src/window.h: (meta_window_set_user_time): new function to
abstract the many things that need to be done when updating the
net_wm_user_time of any window
* src/window.c: (meta_window_activate): add debugging spew, make
sure the comparison is made with last_user_time NOT
last_focus_time, use meta_window_set_user_time() function in order
to correct problems, (meta_window_client_message): add a newline
to a debugging message to make them easier to read,
(meta_window_set_user_time): new function
* src/window-props.c (reload_net_wm_user_time): use the new
meta_window_ste_user_time() function in order to correct problems
2005-02-20 12:14:16 -05:00
2009-08-30 20:04:48 -04:00
/**
* meta_window_get_user_time :
2013-02-15 13:42:08 -05:00
* @ window : a # MetaWindow
2009-08-30 20:04:48 -04:00
*
* The user time represents a timestamp for the last time the user
* interacted with this window . Note this property is only available
* for non - override - redirect windows .
*
* The property is set by Mutter initially upon window creation ,
* and updated thereafter on input events ( key and button presses ) seen by Mutter ,
* client updates to the _NET_WM_USER_TIME property ( if later than the current time )
* and when focusing the window .
*
* Returns : The last time the user interacted with this window .
*/
guint32
meta_window_get_user_time ( MetaWindow * window )
{
return window - > net_wm_user_time ;
}
Big patch to cover about 6 different issues in order to correct rare
2005-02-20 Elijah Newren <newren@gmail.com>
Big patch to cover about 6 different issues in order to correct
rare problems with timestamps (make sure window selected in
tasklist actually gets focus, sanity check timestamps to avoid
rogue apps hosing the system, correct the updating of
net_wm_user_time, correctly handle timestamps of 0 when comparing
xserver timestamps for those who have had their systems up for
over 25 days or so, add some debugging information to verbose
logs, some code cleanups). Fixes all issues listed in #167358.
* src/display.h: (struct _MetaDisplay): clarify comment on
last_focus_time, introduce a new variable--last_user_time,
(XSERVER_TIME_IS_BEFORE macro): put this functionality into a
separate macro and then introduce a new macro with this name that
uses the old one but adds additional special-case checks for
timestamps that are 0, (comment to
meta_display_set_input_focus_window): add information about how
last_user_time should be used in this function
* src/display.c (santiy_check_timestamps): new function,
(meta_display_open): intialize display->last_user_time,
(meta_display_get_current_time_roundtrip): use the timestamp,
which is known to be good, in order to sanity_check_timestamps,
(event_callback): use the new meta_window_ste_user_time() function
in order to correct problems, use the timestamp of KeyPress and
ButtonPress events, which are known to be good, in order to
sanity_check_timestamps, (timestamp_too_old): new function for
common behavior of meta_display_focus_the_no_focus_window and
meta_display_set_input_focus_window, with added checking for
display->last_user_time in addition to display->last_focus_time,
(meta_display_set_input_focus_window): replace some of the code
with a call to timestamp_too_old(),
(meta_display_focus_the_no_focus_window): replace some of th ecode
with a call to timestamp_too_old()
* src/window.h: (meta_window_set_user_time): new function to
abstract the many things that need to be done when updating the
net_wm_user_time of any window
* src/window.c: (meta_window_activate): add debugging spew, make
sure the comparison is made with last_user_time NOT
last_focus_time, use meta_window_set_user_time() function in order
to correct problems, (meta_window_client_message): add a newline
to a debugging message to make them easier to read,
(meta_window_set_user_time): new function
* src/window-props.c (reload_net_wm_user_time): use the new
meta_window_ste_user_time() function in order to correct problems
2005-02-20 12:14:16 -05:00
void
meta_window_set_user_time ( MetaWindow * window ,
Fix issues on 64-bit machines with timestamps by using guint32 (like gtk+
2006-09-13 Elijah Newren <newren gmail com>
* src/common.h (MetaWindowMenuFunc):
* src/core.[ch] (meta_core_user_lower_and_unfocus,
meta_core_user_focus, meta_core_show_window_menu,
meta_core_begin_grab_op, meta_core_end_grab_op):
* src/delete.c (delete_ping_reply_func, delete_ping_timeout_func,
meta_window_delete):
* src/display.[ch] (struct MetaDisplay, struct MetaPingData,
sanity_check_timestamps, meta_display_open, event_callback,
meta_spew_event, meta_display_set_grab_op_cursor,
meta_display_begin_grab_op, meta_display_end_grab_op,
meta_display_ping_timeout, meta_display_ping_window,
process_pong_message, timestamp_too_old,
meta_display_set_input_focus_window):
* src/keybindings.[ch] (grab_keyboard, ungrab_keyboard,
meta_screen_grab_all_keys, meta_window_grab_all_keys,
meta_window_ungrab_all_keys, error_on_generic_command,
error_on_command, error_on_terminal_command):
* src/metacity-dialog.c (on_realize, warn_about_no_sm_support,
error_about_command, main):
* src/screen.[ch] (struct _MetaScreen, meta_screen_new,
meta_screen_show_desktop, meta_screen_apply_startup_properties):
* src/session.c (warn_about_lame_clients_and_finish_interact):
* src/window.[ch] (struct _MetaWindow,
intervening_user_event_occurred, window_activate,
meta_window_delete, meta_window_focus,
meta_window_send_icccm_message, meta_window_client_message,
menu_callback, meta_window_show_menu, struct EventScannerData,
check_use_this_motion_notify, meta_window_begin_grab_op,
meta_window_set_user_time):
* src/workspace.[ch] (focus_ancestor_or_mru_window,
meta_workspace_activate_with_focus, meta_workspace_activate,
meta_workspace_focus_default_window,
focus_ancestor_or_mru_window):
Fix issues on 64-bit machines with timestamps by using guint32
(like gtk+ does) instead of Time. #348305
2006-09-13 12:32:33 -04:00
guint32 timestamp )
Big patch to cover about 6 different issues in order to correct rare
2005-02-20 Elijah Newren <newren@gmail.com>
Big patch to cover about 6 different issues in order to correct
rare problems with timestamps (make sure window selected in
tasklist actually gets focus, sanity check timestamps to avoid
rogue apps hosing the system, correct the updating of
net_wm_user_time, correctly handle timestamps of 0 when comparing
xserver timestamps for those who have had their systems up for
over 25 days or so, add some debugging information to verbose
logs, some code cleanups). Fixes all issues listed in #167358.
* src/display.h: (struct _MetaDisplay): clarify comment on
last_focus_time, introduce a new variable--last_user_time,
(XSERVER_TIME_IS_BEFORE macro): put this functionality into a
separate macro and then introduce a new macro with this name that
uses the old one but adds additional special-case checks for
timestamps that are 0, (comment to
meta_display_set_input_focus_window): add information about how
last_user_time should be used in this function
* src/display.c (santiy_check_timestamps): new function,
(meta_display_open): intialize display->last_user_time,
(meta_display_get_current_time_roundtrip): use the timestamp,
which is known to be good, in order to sanity_check_timestamps,
(event_callback): use the new meta_window_ste_user_time() function
in order to correct problems, use the timestamp of KeyPress and
ButtonPress events, which are known to be good, in order to
sanity_check_timestamps, (timestamp_too_old): new function for
common behavior of meta_display_focus_the_no_focus_window and
meta_display_set_input_focus_window, with added checking for
display->last_user_time in addition to display->last_focus_time,
(meta_display_set_input_focus_window): replace some of the code
with a call to timestamp_too_old(),
(meta_display_focus_the_no_focus_window): replace some of th ecode
with a call to timestamp_too_old()
* src/window.h: (meta_window_set_user_time): new function to
abstract the many things that need to be done when updating the
net_wm_user_time of any window
* src/window.c: (meta_window_activate): add debugging spew, make
sure the comparison is made with last_user_time NOT
last_focus_time, use meta_window_set_user_time() function in order
to correct problems, (meta_window_client_message): add a newline
to a debugging message to make them easier to read,
(meta_window_set_user_time): new function
* src/window-props.c (reload_net_wm_user_time): use the new
meta_window_ste_user_time() function in order to correct problems
2005-02-20 12:14:16 -05:00
{
/* FIXME: If Soeren's suggestion in bug 151984 is implemented, it will allow
* us to sanity check the timestamp here and ensure it doesn ' t correspond to
* a future time .
*/
2009-06-14 08:37:57 -04:00
g_return_if_fail ( ! window - > override_redirect ) ;
Big patch to cover about 6 different issues in order to correct rare
2005-02-20 Elijah Newren <newren@gmail.com>
Big patch to cover about 6 different issues in order to correct
rare problems with timestamps (make sure window selected in
tasklist actually gets focus, sanity check timestamps to avoid
rogue apps hosing the system, correct the updating of
net_wm_user_time, correctly handle timestamps of 0 when comparing
xserver timestamps for those who have had their systems up for
over 25 days or so, add some debugging information to verbose
logs, some code cleanups). Fixes all issues listed in #167358.
* src/display.h: (struct _MetaDisplay): clarify comment on
last_focus_time, introduce a new variable--last_user_time,
(XSERVER_TIME_IS_BEFORE macro): put this functionality into a
separate macro and then introduce a new macro with this name that
uses the old one but adds additional special-case checks for
timestamps that are 0, (comment to
meta_display_set_input_focus_window): add information about how
last_user_time should be used in this function
* src/display.c (santiy_check_timestamps): new function,
(meta_display_open): intialize display->last_user_time,
(meta_display_get_current_time_roundtrip): use the timestamp,
which is known to be good, in order to sanity_check_timestamps,
(event_callback): use the new meta_window_ste_user_time() function
in order to correct problems, use the timestamp of KeyPress and
ButtonPress events, which are known to be good, in order to
sanity_check_timestamps, (timestamp_too_old): new function for
common behavior of meta_display_focus_the_no_focus_window and
meta_display_set_input_focus_window, with added checking for
display->last_user_time in addition to display->last_focus_time,
(meta_display_set_input_focus_window): replace some of the code
with a call to timestamp_too_old(),
(meta_display_focus_the_no_focus_window): replace some of th ecode
with a call to timestamp_too_old()
* src/window.h: (meta_window_set_user_time): new function to
abstract the many things that need to be done when updating the
net_wm_user_time of any window
* src/window.c: (meta_window_activate): add debugging spew, make
sure the comparison is made with last_user_time NOT
last_focus_time, use meta_window_set_user_time() function in order
to correct problems, (meta_window_client_message): add a newline
to a debugging message to make them easier to read,
(meta_window_set_user_time): new function
* src/window-props.c (reload_net_wm_user_time): use the new
meta_window_ste_user_time() function in order to correct problems
2005-02-20 12:14:16 -05:00
/* Only update the time if this timestamp is newer... */
if ( window - > net_wm_user_time_set & &
XSERVER_TIME_IS_BEFORE ( timestamp , window - > net_wm_user_time ) )
{
meta_topic ( META_DEBUG_STARTUP ,
Fix issues on 64-bit machines with timestamps by using guint32 (like gtk+
2006-09-13 Elijah Newren <newren gmail com>
* src/common.h (MetaWindowMenuFunc):
* src/core.[ch] (meta_core_user_lower_and_unfocus,
meta_core_user_focus, meta_core_show_window_menu,
meta_core_begin_grab_op, meta_core_end_grab_op):
* src/delete.c (delete_ping_reply_func, delete_ping_timeout_func,
meta_window_delete):
* src/display.[ch] (struct MetaDisplay, struct MetaPingData,
sanity_check_timestamps, meta_display_open, event_callback,
meta_spew_event, meta_display_set_grab_op_cursor,
meta_display_begin_grab_op, meta_display_end_grab_op,
meta_display_ping_timeout, meta_display_ping_window,
process_pong_message, timestamp_too_old,
meta_display_set_input_focus_window):
* src/keybindings.[ch] (grab_keyboard, ungrab_keyboard,
meta_screen_grab_all_keys, meta_window_grab_all_keys,
meta_window_ungrab_all_keys, error_on_generic_command,
error_on_command, error_on_terminal_command):
* src/metacity-dialog.c (on_realize, warn_about_no_sm_support,
error_about_command, main):
* src/screen.[ch] (struct _MetaScreen, meta_screen_new,
meta_screen_show_desktop, meta_screen_apply_startup_properties):
* src/session.c (warn_about_lame_clients_and_finish_interact):
* src/window.[ch] (struct _MetaWindow,
intervening_user_event_occurred, window_activate,
meta_window_delete, meta_window_focus,
meta_window_send_icccm_message, meta_window_client_message,
menu_callback, meta_window_show_menu, struct EventScannerData,
check_use_this_motion_notify, meta_window_begin_grab_op,
meta_window_set_user_time):
* src/workspace.[ch] (focus_ancestor_or_mru_window,
meta_workspace_activate_with_focus, meta_workspace_activate,
meta_workspace_focus_default_window,
focus_ancestor_or_mru_window):
Fix issues on 64-bit machines with timestamps by using guint32
(like gtk+ does) instead of Time. #348305
2006-09-13 12:32:33 -04:00
" Window %s _NET_WM_USER_TIME not updated to %u, because it "
" is less than %u \n " ,
Big patch to cover about 6 different issues in order to correct rare
2005-02-20 Elijah Newren <newren@gmail.com>
Big patch to cover about 6 different issues in order to correct
rare problems with timestamps (make sure window selected in
tasklist actually gets focus, sanity check timestamps to avoid
rogue apps hosing the system, correct the updating of
net_wm_user_time, correctly handle timestamps of 0 when comparing
xserver timestamps for those who have had their systems up for
over 25 days or so, add some debugging information to verbose
logs, some code cleanups). Fixes all issues listed in #167358.
* src/display.h: (struct _MetaDisplay): clarify comment on
last_focus_time, introduce a new variable--last_user_time,
(XSERVER_TIME_IS_BEFORE macro): put this functionality into a
separate macro and then introduce a new macro with this name that
uses the old one but adds additional special-case checks for
timestamps that are 0, (comment to
meta_display_set_input_focus_window): add information about how
last_user_time should be used in this function
* src/display.c (santiy_check_timestamps): new function,
(meta_display_open): intialize display->last_user_time,
(meta_display_get_current_time_roundtrip): use the timestamp,
which is known to be good, in order to sanity_check_timestamps,
(event_callback): use the new meta_window_ste_user_time() function
in order to correct problems, use the timestamp of KeyPress and
ButtonPress events, which are known to be good, in order to
sanity_check_timestamps, (timestamp_too_old): new function for
common behavior of meta_display_focus_the_no_focus_window and
meta_display_set_input_focus_window, with added checking for
display->last_user_time in addition to display->last_focus_time,
(meta_display_set_input_focus_window): replace some of the code
with a call to timestamp_too_old(),
(meta_display_focus_the_no_focus_window): replace some of th ecode
with a call to timestamp_too_old()
* src/window.h: (meta_window_set_user_time): new function to
abstract the many things that need to be done when updating the
net_wm_user_time of any window
* src/window.c: (meta_window_activate): add debugging spew, make
sure the comparison is made with last_user_time NOT
last_focus_time, use meta_window_set_user_time() function in order
to correct problems, (meta_window_client_message): add a newline
to a debugging message to make them easier to read,
(meta_window_set_user_time): new function
* src/window-props.c (reload_net_wm_user_time): use the new
meta_window_ste_user_time() function in order to correct problems
2005-02-20 12:14:16 -05:00
window - > desc , timestamp , window - > net_wm_user_time ) ;
}
else
{
meta_topic ( META_DEBUG_STARTUP ,
Fix issues on 64-bit machines with timestamps by using guint32 (like gtk+
2006-09-13 Elijah Newren <newren gmail com>
* src/common.h (MetaWindowMenuFunc):
* src/core.[ch] (meta_core_user_lower_and_unfocus,
meta_core_user_focus, meta_core_show_window_menu,
meta_core_begin_grab_op, meta_core_end_grab_op):
* src/delete.c (delete_ping_reply_func, delete_ping_timeout_func,
meta_window_delete):
* src/display.[ch] (struct MetaDisplay, struct MetaPingData,
sanity_check_timestamps, meta_display_open, event_callback,
meta_spew_event, meta_display_set_grab_op_cursor,
meta_display_begin_grab_op, meta_display_end_grab_op,
meta_display_ping_timeout, meta_display_ping_window,
process_pong_message, timestamp_too_old,
meta_display_set_input_focus_window):
* src/keybindings.[ch] (grab_keyboard, ungrab_keyboard,
meta_screen_grab_all_keys, meta_window_grab_all_keys,
meta_window_ungrab_all_keys, error_on_generic_command,
error_on_command, error_on_terminal_command):
* src/metacity-dialog.c (on_realize, warn_about_no_sm_support,
error_about_command, main):
* src/screen.[ch] (struct _MetaScreen, meta_screen_new,
meta_screen_show_desktop, meta_screen_apply_startup_properties):
* src/session.c (warn_about_lame_clients_and_finish_interact):
* src/window.[ch] (struct _MetaWindow,
intervening_user_event_occurred, window_activate,
meta_window_delete, meta_window_focus,
meta_window_send_icccm_message, meta_window_client_message,
menu_callback, meta_window_show_menu, struct EventScannerData,
check_use_this_motion_notify, meta_window_begin_grab_op,
meta_window_set_user_time):
* src/workspace.[ch] (focus_ancestor_or_mru_window,
meta_workspace_activate_with_focus, meta_workspace_activate,
meta_workspace_focus_default_window,
focus_ancestor_or_mru_window):
Fix issues on 64-bit machines with timestamps by using guint32
(like gtk+ does) instead of Time. #348305
2006-09-13 12:32:33 -04:00
" Window %s has _NET_WM_USER_TIME of %u \n " ,
Big patch to cover about 6 different issues in order to correct rare
2005-02-20 Elijah Newren <newren@gmail.com>
Big patch to cover about 6 different issues in order to correct
rare problems with timestamps (make sure window selected in
tasklist actually gets focus, sanity check timestamps to avoid
rogue apps hosing the system, correct the updating of
net_wm_user_time, correctly handle timestamps of 0 when comparing
xserver timestamps for those who have had their systems up for
over 25 days or so, add some debugging information to verbose
logs, some code cleanups). Fixes all issues listed in #167358.
* src/display.h: (struct _MetaDisplay): clarify comment on
last_focus_time, introduce a new variable--last_user_time,
(XSERVER_TIME_IS_BEFORE macro): put this functionality into a
separate macro and then introduce a new macro with this name that
uses the old one but adds additional special-case checks for
timestamps that are 0, (comment to
meta_display_set_input_focus_window): add information about how
last_user_time should be used in this function
* src/display.c (santiy_check_timestamps): new function,
(meta_display_open): intialize display->last_user_time,
(meta_display_get_current_time_roundtrip): use the timestamp,
which is known to be good, in order to sanity_check_timestamps,
(event_callback): use the new meta_window_ste_user_time() function
in order to correct problems, use the timestamp of KeyPress and
ButtonPress events, which are known to be good, in order to
sanity_check_timestamps, (timestamp_too_old): new function for
common behavior of meta_display_focus_the_no_focus_window and
meta_display_set_input_focus_window, with added checking for
display->last_user_time in addition to display->last_focus_time,
(meta_display_set_input_focus_window): replace some of the code
with a call to timestamp_too_old(),
(meta_display_focus_the_no_focus_window): replace some of th ecode
with a call to timestamp_too_old()
* src/window.h: (meta_window_set_user_time): new function to
abstract the many things that need to be done when updating the
net_wm_user_time of any window
* src/window.c: (meta_window_activate): add debugging spew, make
sure the comparison is made with last_user_time NOT
last_focus_time, use meta_window_set_user_time() function in order
to correct problems, (meta_window_client_message): add a newline
to a debugging message to make them easier to read,
(meta_window_set_user_time): new function
* src/window-props.c (reload_net_wm_user_time): use the new
meta_window_ste_user_time() function in order to correct problems
2005-02-20 12:14:16 -05:00
window - > desc , timestamp ) ;
window - > net_wm_user_time_set = TRUE ;
window - > net_wm_user_time = timestamp ;
if ( XSERVER_TIME_IS_BEFORE ( window - > display - > last_user_time , timestamp ) )
window - > display - > last_user_time = timestamp ;
2006-01-15 12:03:57 -05:00
/* If this is a terminal, user interaction with it means the user likely
* doesn ' t want to have focus transferred for now due to new windows .
*/
2014-03-18 10:08:28 -04:00
if ( meta_prefs_get_focus_new_windows ( ) = = G_DESKTOP_FOCUS_NEW_WINDOWS_STRICT & &
window_is_terminal ( window ) )
2006-01-15 12:03:57 -05:00
window - > display - > allow_terminal_deactivation = FALSE ;
Big patch to cover about 6 different issues in order to correct rare
2005-02-20 Elijah Newren <newren@gmail.com>
Big patch to cover about 6 different issues in order to correct
rare problems with timestamps (make sure window selected in
tasklist actually gets focus, sanity check timestamps to avoid
rogue apps hosing the system, correct the updating of
net_wm_user_time, correctly handle timestamps of 0 when comparing
xserver timestamps for those who have had their systems up for
over 25 days or so, add some debugging information to verbose
logs, some code cleanups). Fixes all issues listed in #167358.
* src/display.h: (struct _MetaDisplay): clarify comment on
last_focus_time, introduce a new variable--last_user_time,
(XSERVER_TIME_IS_BEFORE macro): put this functionality into a
separate macro and then introduce a new macro with this name that
uses the old one but adds additional special-case checks for
timestamps that are 0, (comment to
meta_display_set_input_focus_window): add information about how
last_user_time should be used in this function
* src/display.c (santiy_check_timestamps): new function,
(meta_display_open): intialize display->last_user_time,
(meta_display_get_current_time_roundtrip): use the timestamp,
which is known to be good, in order to sanity_check_timestamps,
(event_callback): use the new meta_window_ste_user_time() function
in order to correct problems, use the timestamp of KeyPress and
ButtonPress events, which are known to be good, in order to
sanity_check_timestamps, (timestamp_too_old): new function for
common behavior of meta_display_focus_the_no_focus_window and
meta_display_set_input_focus_window, with added checking for
display->last_user_time in addition to display->last_focus_time,
(meta_display_set_input_focus_window): replace some of the code
with a call to timestamp_too_old(),
(meta_display_focus_the_no_focus_window): replace some of th ecode
with a call to timestamp_too_old()
* src/window.h: (meta_window_set_user_time): new function to
abstract the many things that need to be done when updating the
net_wm_user_time of any window
* src/window.c: (meta_window_activate): add debugging spew, make
sure the comparison is made with last_user_time NOT
last_focus_time, use meta_window_set_user_time() function in order
to correct problems, (meta_window_client_message): add a newline
to a debugging message to make them easier to read,
(meta_window_set_user_time): new function
* src/window-props.c (reload_net_wm_user_time): use the new
meta_window_ste_user_time() function in order to correct problems
2005-02-20 12:14:16 -05:00
}
2009-08-31 12:06:53 -04:00
g_object_notify ( G_OBJECT ( window ) , " user-time " ) ;
Big patch to cover about 6 different issues in order to correct rare
2005-02-20 Elijah Newren <newren@gmail.com>
Big patch to cover about 6 different issues in order to correct
rare problems with timestamps (make sure window selected in
tasklist actually gets focus, sanity check timestamps to avoid
rogue apps hosing the system, correct the updating of
net_wm_user_time, correctly handle timestamps of 0 when comparing
xserver timestamps for those who have had their systems up for
over 25 days or so, add some debugging information to verbose
logs, some code cleanups). Fixes all issues listed in #167358.
* src/display.h: (struct _MetaDisplay): clarify comment on
last_focus_time, introduce a new variable--last_user_time,
(XSERVER_TIME_IS_BEFORE macro): put this functionality into a
separate macro and then introduce a new macro with this name that
uses the old one but adds additional special-case checks for
timestamps that are 0, (comment to
meta_display_set_input_focus_window): add information about how
last_user_time should be used in this function
* src/display.c (santiy_check_timestamps): new function,
(meta_display_open): intialize display->last_user_time,
(meta_display_get_current_time_roundtrip): use the timestamp,
which is known to be good, in order to sanity_check_timestamps,
(event_callback): use the new meta_window_ste_user_time() function
in order to correct problems, use the timestamp of KeyPress and
ButtonPress events, which are known to be good, in order to
sanity_check_timestamps, (timestamp_too_old): new function for
common behavior of meta_display_focus_the_no_focus_window and
meta_display_set_input_focus_window, with added checking for
display->last_user_time in addition to display->last_focus_time,
(meta_display_set_input_focus_window): replace some of the code
with a call to timestamp_too_old(),
(meta_display_focus_the_no_focus_window): replace some of th ecode
with a call to timestamp_too_old()
* src/window.h: (meta_window_set_user_time): new function to
abstract the many things that need to be done when updating the
net_wm_user_time of any window
* src/window.c: (meta_window_activate): add debugging spew, make
sure the comparison is made with last_user_time NOT
last_focus_time, use meta_window_set_user_time() function in order
to correct problems, (meta_window_client_message): add a newline
to a debugging message to make them easier to read,
(meta_window_set_user_time): new function
* src/window-props.c (reload_net_wm_user_time): use the new
meta_window_ste_user_time() function in order to correct problems
2005-02-20 12:14:16 -05:00
}
2006-03-29 13:43:22 -05:00
2009-09-18 20:50:11 -04:00
/**
* meta_window_get_stable_sequence :
* @ window : A # MetaWindow
*
* The stable sequence number is a monotonicially increasing
* unique integer assigned to each # MetaWindow upon creation .
*
* This number can be useful for sorting windows in a stable
* fashion .
*
* Returns : Internal sequence number for this window
*/
guint32
meta_window_get_stable_sequence ( MetaWindow * window )
{
g_return_val_if_fail ( META_IS_WINDOW ( window ) , 0 ) ;
return window - > stable_sequence ;
}
2006-03-29 13:43:22 -05:00
/* Sets the demands_attention hint on a window, but only
* if it ' s at least partially obscured ( see # 305882 ) .
*/
void
meta_window_set_demands_attention ( MetaWindow * window )
{
MetaRectangle candidate_rect , other_rect ;
GList * stack = window - > screen - > stack - > sorted ;
MetaWindow * other_window ;
gboolean obscured = FALSE ;
2008-12-18 07:41:56 -05:00
2008-02-01 23:05:16 -05:00
MetaWorkspace * workspace = window - > screen - > active_workspace ;
2009-11-17 05:02:06 -05:00
if ( window - > wm_state_demands_attention )
return ;
2008-02-01 23:05:16 -05:00
if ( workspace ! = window - > workspace )
2006-03-29 13:43:22 -05:00
{
2008-02-01 23:05:16 -05:00
/* windows on other workspaces are necessarily obscured */
obscured = TRUE ;
}
2008-10-21 23:21:23 -04:00
else if ( window - > minimized )
{
obscured = TRUE ;
}
2008-02-01 23:05:16 -05:00
else
{
2013-11-19 12:24:45 -05:00
meta_window_get_frame_rect ( window , & candidate_rect ) ;
2006-03-29 13:43:22 -05:00
2008-02-01 23:05:16 -05:00
/* The stack is sorted with the top windows first. */
2008-12-18 07:41:56 -05:00
2008-02-01 23:05:16 -05:00
while ( stack ! = NULL & & stack - > data ! = window )
{
other_window = stack - > data ;
stack = stack - > next ;
2008-12-18 07:41:56 -05:00
2013-01-18 14:55:23 -05:00
if ( meta_window_located_on_workspace ( other_window , window - > workspace ) )
2006-03-29 13:43:22 -05:00
{
2013-11-19 12:24:45 -05:00
meta_window_get_frame_rect ( other_window , & other_rect ) ;
2008-02-01 23:05:16 -05:00
if ( meta_rectangle_overlap ( & candidate_rect , & other_rect ) )
{
obscured = TRUE ;
break ;
}
2006-03-29 13:43:22 -05:00
}
}
2008-10-21 23:21:23 -04:00
}
if ( obscured )
{
2006-03-29 13:43:22 -05:00
meta_topic ( META_DEBUG_WINDOW_OPS ,
2008-10-21 23:21:23 -04:00
" Marking %s as needing attention \n " ,
window - > desc ) ;
window - > wm_state_demands_attention = TRUE ;
set_net_wm_state ( window ) ;
2009-11-17 05:02:06 -05:00
g_object_notify ( G_OBJECT ( window ) , " demands-attention " ) ;
2009-10-28 13:19:24 -04:00
g_signal_emit_by_name ( window - > display , " window-demands-attention " ,
window ) ;
2006-03-29 13:43:22 -05:00
}
2008-10-21 23:21:23 -04:00
else
{
/* If the window's in full view, there's no point setting the flag. */
2006-03-29 13:43:22 -05:00
2008-10-21 23:21:23 -04:00
meta_topic ( META_DEBUG_WINDOW_OPS ,
" Not marking %s as needing attention because "
" it's in full view \n " ,
window - > desc ) ;
}
2006-03-29 13:43:22 -05:00
}
void
meta_window_unset_demands_attention ( MetaWindow * window )
{
meta_topic ( META_DEBUG_WINDOW_OPS ,
" Marking %s as not needing attention \n " , window - > desc ) ;
2008-12-18 07:41:56 -05:00
2009-11-17 05:02:06 -05:00
if ( window - > wm_state_demands_attention )
{
window - > wm_state_demands_attention = FALSE ;
set_net_wm_state ( window ) ;
g_object_notify ( G_OBJECT ( window ) , " demands-attention " ) ;
}
2006-03-29 13:43:22 -05:00
}
2008-05-18 20:00:09 -04:00
2010-09-01 15:39:53 -04:00
/**
* meta_window_get_frame : ( skip )
2013-02-15 13:42:08 -05:00
* @ window : a # MetaWindow
2010-09-01 15:39:53 -04:00
*
*/
2008-05-18 20:00:09 -04:00
MetaFrame *
meta_window_get_frame ( MetaWindow * window )
{
return window - > frame ;
}
2010-11-11 16:04:31 -05:00
/**
* meta_window_appears_focused :
* @ window : a # MetaWindow
*
* Determines if the window should be drawn with a focused appearance . This is
* true for focused windows but also true for windows with a focused modal
* dialog attached .
*
* Return value : % TRUE if the window should be drawn with a focused frame
*/
2010-09-10 05:55:20 -04:00
gboolean
meta_window_appears_focused ( MetaWindow * window )
{
2011-03-24 21:36:47 -04:00
return window - > has_focus | | ( window - > attached_focus_window ! = NULL ) ;
2010-09-10 05:55:20 -04:00
}
2008-05-18 20:00:09 -04:00
gboolean
meta_window_has_focus ( MetaWindow * window )
{
return window - > has_focus ;
}
gboolean
meta_window_is_shaded ( MetaWindow * window )
{
return window - > shaded ;
}
2009-08-14 05:42:56 -04:00
/**
* meta_window_is_override_redirect :
* @ window : A # MetaWindow
*
2013-02-15 13:42:08 -05:00
* Returns : % TRUE if this window isn ' t managed by mutter ; it will
2009-08-14 05:42:56 -04:00
* control its own positioning and mutter won ' t draw decorations
* among other things . In X terminology this is " override redirect " .
*/
gboolean
meta_window_is_override_redirect ( MetaWindow * window )
{
return window - > override_redirect ;
}
2009-08-28 12:39:55 -04:00
/**
* meta_window_is_skip_taskbar :
* @ window : A # MetaWindow
*
* Gets whether this window should be ignored by task lists .
*
* Return value : % TRUE if the skip bar hint is set .
*/
gboolean
meta_window_is_skip_taskbar ( MetaWindow * window )
{
g_return_val_if_fail ( META_IS_WINDOW ( window ) , FALSE ) ;
return window - > skip_taskbar ;
}
2011-02-26 12:39:57 -05:00
/**
* meta_window_get_rect :
* @ window : a # MetaWindow
*
* Gets the rectangle that bounds @ window , ignoring any window decorations .
*
* Return value : ( transfer none ) : the # MetaRectangle for the window
*/
2008-05-18 20:00:09 -04:00
MetaRectangle *
meta_window_get_rect ( MetaWindow * window )
{
return & window - > rect ;
}
2008-11-17 16:34:28 -05:00
/**
* meta_window_get_screen :
* @ window : a # MetaWindow
*
* Gets the # MetaScreen that the window is on .
*
* Return value : ( transfer none ) : the # MetaScreen for the window
*/
2008-05-18 20:00:09 -04:00
MetaScreen *
meta_window_get_screen ( MetaWindow * window )
{
return window - > screen ;
}
2010-09-01 15:39:53 -04:00
/**
* meta_window_get_display :
* @ window : A # MetaWindow
*
* Returns : ( transfer none ) : The display for @ window
*/
2008-05-18 20:00:09 -04:00
MetaDisplay *
meta_window_get_display ( MetaWindow * window )
{
return window - > display ;
}
2010-09-01 15:39:53 -04:00
/**
* meta_window_get_xwindow : ( skip )
2013-02-15 13:42:08 -05:00
* @ window : a # MetaWindow
2010-09-01 15:39:53 -04:00
*
*/
2008-05-18 20:00:09 -04:00
Window
meta_window_get_xwindow ( MetaWindow * window )
{
return window - > xwindow ;
}
2008-08-18 10:42:38 -04:00
MetaWindowType
2008-11-07 04:57:00 -05:00
meta_window_get_window_type ( MetaWindow * window )
2008-08-18 10:42:38 -04:00
{
return window - > type ;
}
2008-11-17 16:34:28 -05:00
/**
* meta_window_get_workspace :
* @ window : a # MetaWindow
*
* Gets the # MetaWorkspace that the window is currently displayed on .
* If the window is on all workspaces , returns the currently active
* workspace .
*
* Return value : ( transfer none ) : the # MetaWorkspace for the window
*/
2008-09-18 11:09:11 -04:00
MetaWorkspace *
meta_window_get_workspace ( MetaWindow * window )
{
if ( window - > on_all_workspaces )
return window - > screen - > active_workspace ;
return window - > workspace ;
}
gboolean
meta_window_is_on_all_workspaces ( MetaWindow * window )
{
return window - > on_all_workspaces ;
}
2008-10-24 05:07:24 -04:00
gboolean
meta_window_is_hidden ( MetaWindow * window )
{
return window - > hidden ;
}
2008-10-30 04:30:42 -04:00
const char *
meta_window_get_description ( MetaWindow * window )
{
if ( ! window )
return NULL ;
return window - > desc ;
}
2008-11-23 14:28:40 -05:00
2009-05-29 13:26:18 -04:00
/**
* meta_window_get_wm_class :
* @ window : a # MetaWindow
2010-03-17 05:20:39 -04:00
*
* Return the current value of the name part of WM_CLASS X property .
2009-05-29 13:26:18 -04:00
*/
const char *
meta_window_get_wm_class ( MetaWindow * window )
{
if ( ! window )
return NULL ;
return window - > res_class ;
}
2010-03-17 05:20:39 -04:00
/**
* meta_window_get_wm_class_instance :
* @ window : a # MetaWindow
*
* Return the current value of the instance part of WM_CLASS X property .
*/
const char *
meta_window_get_wm_class_instance ( MetaWindow * window )
{
if ( ! window )
return NULL ;
return window - > res_name ;
}
2011-11-25 18:42:41 -05:00
/**
2012-01-17 13:27:10 -05:00
* meta_window_get_gtk_application_id :
2011-11-25 18:42:41 -05:00
* @ window : a # MetaWindow
*
* Return value : ( transfer none ) : the application ID
* */
const char *
2012-01-17 13:27:10 -05:00
meta_window_get_gtk_application_id ( MetaWindow * window )
2011-11-25 18:42:41 -05:00
{
2012-01-17 13:27:10 -05:00
return window - > gtk_application_id ;
2011-11-25 18:42:41 -05:00
}
2011-12-05 12:52:35 -05:00
/**
2012-01-17 13:27:10 -05:00
* meta_window_get_gtk_unique_bus_name :
2011-12-05 12:52:35 -05:00
* @ window : a # MetaWindow
*
* Return value : ( transfer none ) : the unique name
* */
const char *
2012-01-17 13:27:10 -05:00
meta_window_get_gtk_unique_bus_name ( MetaWindow * window )
{
return window - > gtk_unique_bus_name ;
}
/**
* meta_window_get_gtk_application_object_path :
* @ window : a # MetaWindow
*
* Return value : ( transfer none ) : the object path
* */
const char *
meta_window_get_gtk_application_object_path ( MetaWindow * window )
{
return window - > gtk_application_object_path ;
}
/**
* meta_window_get_gtk_window_object_path :
* @ window : a # MetaWindow
*
* Return value : ( transfer none ) : the object path
* */
const char *
meta_window_get_gtk_window_object_path ( MetaWindow * window )
{
return window - > gtk_window_object_path ;
}
/**
* meta_window_get_gtk_app_menu_object_path :
* @ window : a # MetaWindow
*
* Return value : ( transfer none ) : the object path
* */
const char *
meta_window_get_gtk_app_menu_object_path ( MetaWindow * window )
2011-12-05 12:52:35 -05:00
{
2012-01-17 13:27:10 -05:00
return window - > gtk_app_menu_object_path ;
2011-12-05 12:52:35 -05:00
}
/**
2012-01-17 13:27:10 -05:00
* meta_window_get_gtk_menubar_object_path :
2011-12-05 12:52:35 -05:00
* @ window : a # MetaWindow
*
* Return value : ( transfer none ) : the object path
* */
const char *
2012-01-17 13:27:10 -05:00
meta_window_get_gtk_menubar_object_path ( MetaWindow * window )
2011-12-05 12:52:35 -05:00
{
2012-01-17 13:27:10 -05:00
return window - > gtk_menubar_object_path ;
2011-12-05 12:52:35 -05:00
}
2009-02-02 15:16:05 -05:00
/**
* meta_window_get_compositor_private :
* @ window : a # MetaWindow
*
* Gets the compositor ' s wrapper object for @ window .
*
* Return value : ( transfer none ) : the wrapper object .
* */
GObject *
2008-11-23 14:28:40 -05:00
meta_window_get_compositor_private ( MetaWindow * window )
{
if ( ! window )
return NULL ;
return window - > compositor_private ;
}
void
2009-02-02 15:16:05 -05:00
meta_window_set_compositor_private ( MetaWindow * window , GObject * priv )
2008-11-23 14:28:40 -05:00
{
if ( ! window )
return ;
window - > compositor_private = priv ;
}
2009-01-12 07:30:31 -05:00
const char *
meta_window_get_role ( MetaWindow * window )
{
if ( ! window )
return NULL ;
return window - > role ;
}
2009-06-09 23:02:14 -04:00
/**
* meta_window_get_title :
* @ window : a # MetaWindow
*
2013-02-15 13:42:08 -05:00
* Returns : the current title of the window .
2009-06-09 23:02:14 -04:00
*/
const char *
meta_window_get_title ( MetaWindow * window )
{
g_return_val_if_fail ( META_IS_WINDOW ( window ) , NULL ) ;
return window - > title ;
}
2009-03-12 17:07:27 -04:00
MetaStackLayer
meta_window_get_layer ( MetaWindow * window )
{
return window - > layer ;
}
2009-06-09 23:28:21 -04:00
/**
* meta_window_get_transient_for :
* @ window : a # MetaWindow
*
* Returns the # MetaWindow for the window that is pointed to by the
* WM_TRANSIENT_FOR hint on this window ( see XGetTransientForHint ( )
* or XSetTransientForHint ( ) ) . Metacity keeps transient windows above their
* parents . A typical usage of this hint is for a dialog that wants to stay
* above its associated window .
*
* Return value : ( transfer none ) : the window this window is transient for , or
* % NULL if the WM_TRANSIENT_FOR hint is unset or does not point to a toplevel
* window that Metacity knows about .
*/
MetaWindow *
meta_window_get_transient_for ( MetaWindow * window )
{
g_return_val_if_fail ( META_IS_WINDOW ( window ) , NULL ) ;
2013-09-16 08:44:37 -04:00
if ( window - > transient_for )
return window - > transient_for ;
else if ( window - > xtransient_for )
2009-06-09 23:28:21 -04:00
return meta_display_lookup_x_window ( window - > display ,
window - > xtransient_for ) ;
else
return NULL ;
}
2009-07-31 13:52:25 -04:00
/**
* meta_window_get_pid :
* @ window : a # MetaWindow
*
* Returns pid of the process that created this window , if known ( obtained from
* the _NET_WM_PID property ) .
*
* Return value : ( transfer none ) : the pid , or - 1 if not known .
*/
int
meta_window_get_pid ( MetaWindow * window )
{
g_return_val_if_fail ( META_IS_WINDOW ( window ) , - 1 ) ;
return window - > net_wm_pid ;
}
/**
* meta_window_get_client_machine :
* @ window : a # MetaWindow
*
* Returns name of the client machine from which this windows was created ,
* if known ( obtained from the WM_CLIENT_MACHINE property ) .
*
* Return value : ( transfer none ) : the machine name , or NULL ; the string is
* owned by the window manager and should not be freed or modified by the
* caller .
*/
const char *
meta_window_get_client_machine ( MetaWindow * window )
{
g_return_val_if_fail ( META_IS_WINDOW ( window ) , NULL ) ;
return window - > wm_client_machine ;
}
2010-06-04 13:57:41 -04:00
/**
* meta_window_is_remote :
* @ window : a # MetaWindow
*
* Returns : % TRUE if this window originates from a host
* different from the one running mutter .
*/
gboolean
meta_window_is_remote ( MetaWindow * window )
{
2013-02-20 15:17:21 -05:00
return window - > is_remote ;
2010-06-04 13:57:41 -04:00
}
2010-01-18 05:12:17 -05:00
/**
* meta_window_get_mutter_hints :
* @ window : a # MetaWindow
*
* Gets the current value of the _MUTTER_HINTS property .
*
* The purpose of the hints is to allow fine - tuning of the Window Manager and
* Compositor behaviour on per - window basis , and is intended primarily for
* hints that are plugin - specific .
*
* The property is a list of colon - separated key = value pairs . The key names for
* any plugin - specific hints must be suitably namespaced to allow for shared
* use ; ' mutter - ' key prefix is reserved for internal use , and must not be used
* by plugins .
*
* Return value : ( transfer none ) : the _MUTTER_HINTS string , or % NULL if no hints
* are set .
*/
const char *
meta_window_get_mutter_hints ( MetaWindow * window )
{
g_return_val_if_fail ( META_IS_WINDOW ( window ) , NULL ) ;
return window - > mutter_hints ;
}
2010-11-11 16:12:11 -05:00
/**
* meta_window_get_frame_type :
* @ window : a # MetaWindow
*
* Gets the type of window decorations that should be used for this window .
*
* Return value : the frame type
*/
MetaFrameType
meta_window_get_frame_type ( MetaWindow * window )
{
MetaFrameType base_type = META_FRAME_TYPE_LAST ;
switch ( window - > type )
{
case META_WINDOW_NORMAL :
base_type = META_FRAME_TYPE_NORMAL ;
break ;
case META_WINDOW_DIALOG :
base_type = META_FRAME_TYPE_DIALOG ;
break ;
case META_WINDOW_MODAL_DIALOG :
window: make determination of attached dialog windows more consistent
Different bits of code were using slightly different checks to test
whether a window was an attached dialog. Add a new
meta_window_is_attached_dialog(), and use that everywhere.
Also, freeze the is-attached status when the window is first shown,
rather than recomputing it each time the caller asks, since this could
cause problems if a window changes its type after it has already been
attached, etc. However, if an attached window's parent is destroyed,
or an attached window changes its transient-for, then fix things up by
destroying the old MetaWindow and creating a new one (causing
compositor unmap and map events to be fired off, allowing the display
of the window to be fixed up).
Remove some code in display.c that tried to fix existing windows if
the gconf setting changed, but which didn't actually do anything (at
least under gnome-shell). However, if 654643 was fixed then the new
behavior with this patch would be that changing the gconf setting
would affect new dialogs, but not existing ones.
https://bugzilla.gnome.org/show_bug.cgi?id=646761
2011-04-05 10:14:52 -04:00
if ( meta_window_is_attached_dialog ( window ) )
2010-11-11 16:12:11 -05:00
base_type = META_FRAME_TYPE_ATTACHED ;
else
base_type = META_FRAME_TYPE_MODAL_DIALOG ;
break ;
case META_WINDOW_MENU :
base_type = META_FRAME_TYPE_MENU ;
break ;
case META_WINDOW_UTILITY :
base_type = META_FRAME_TYPE_UTILITY ;
break ;
case META_WINDOW_DESKTOP :
case META_WINDOW_DOCK :
case META_WINDOW_TOOLBAR :
case META_WINDOW_SPLASHSCREEN :
case META_WINDOW_DROPDOWN_MENU :
case META_WINDOW_POPUP_MENU :
case META_WINDOW_TOOLTIP :
case META_WINDOW_NOTIFICATION :
case META_WINDOW_COMBO :
case META_WINDOW_DND :
case META_WINDOW_OVERRIDE_OTHER :
/* No frame */
base_type = META_FRAME_TYPE_LAST ;
break ;
}
if ( base_type = = META_FRAME_TYPE_LAST )
{
/* can't add border if undecorated */
return META_FRAME_TYPE_LAST ;
}
2013-06-24 14:12:07 -04:00
else if ( window - > border_only | |
2012-07-02 20:19:37 -04:00
( window - > hide_titlebar_when_maximized & & META_WINDOW_MAXIMIZED ( window ) ) | |
( window - > hide_titlebar_when_maximized & & META_WINDOW_TILED_SIDE_BY_SIDE ( window ) ) )
2010-11-11 16:12:11 -05:00
{
/* override base frame type */
return META_FRAME_TYPE_BORDER ;
}
else
{
return base_type ;
}
}
2011-03-22 15:36:12 -04:00
/**
* meta_window_get_frame_bounds :
2013-02-15 13:42:08 -05:00
* @ window : a # MetaWindow
2011-03-22 15:36:12 -04:00
*
* Gets a region representing the outer bounds of the window ' s frame .
*
* Return value : ( transfer none ) ( allow - none ) : a # cairo_region_t
* holding the outer bounds of the window , or % NULL if the window
* doesn ' t have a frame .
*/
cairo_region_t *
meta_window_get_frame_bounds ( MetaWindow * window )
{
if ( ! window - > frame_bounds )
{
if ( window - > frame )
window - > frame_bounds = meta_frame_get_frame_bounds ( window - > frame ) ;
}
return window - > frame_bounds ;
}
window: make determination of attached dialog windows more consistent
Different bits of code were using slightly different checks to test
whether a window was an attached dialog. Add a new
meta_window_is_attached_dialog(), and use that everywhere.
Also, freeze the is-attached status when the window is first shown,
rather than recomputing it each time the caller asks, since this could
cause problems if a window changes its type after it has already been
attached, etc. However, if an attached window's parent is destroyed,
or an attached window changes its transient-for, then fix things up by
destroying the old MetaWindow and creating a new one (causing
compositor unmap and map events to be fired off, allowing the display
of the window to be fixed up).
Remove some code in display.c that tried to fix existing windows if
the gconf setting changed, but which didn't actually do anything (at
least under gnome-shell). However, if 654643 was fixed then the new
behavior with this patch would be that changing the gconf setting
would affect new dialogs, but not existing ones.
https://bugzilla.gnome.org/show_bug.cgi?id=646761
2011-04-05 10:14:52 -04:00
/**
* meta_window_is_attached_dialog :
* @ window : a # MetaWindow
*
* Tests if @ window is should be attached to its parent window .
* ( If the " attach_modal_dialogs " option is not enabled , this will
* always return % FALSE . )
*
* Return value : whether @ window should be attached to its parent
*/
gboolean
meta_window_is_attached_dialog ( MetaWindow * window )
{
return window - > attached ;
}
2012-02-24 11:08:55 -05:00
/**
* meta_window_get_tile_match :
2013-02-15 13:42:08 -05:00
* @ window : a # MetaWindow
2012-02-24 11:08:55 -05:00
*
* Returns the matching tiled window on the same monitor as @ window . This is
* the topmost tiled window in a complementary tile mode that is :
*
* - on the same monitor ;
* - on the same workspace ;
* - spanning the remaining monitor width ;
* - there is no 3 rd window stacked between both tiled windows that ' s
* partially visible in the common edge .
*
* Return value : ( transfer none ) ( allow - none ) : the matching tiled window or
* % NULL if it doesn ' t exist .
*/
MetaWindow *
meta_window_get_tile_match ( MetaWindow * window )
{
return window - > tile_match ;
}
void
meta_window_compute_tile_match ( MetaWindow * window )
{
MetaWindow * match ;
MetaStack * stack ;
2012-03-18 14:53:31 -04:00
MetaTileMode match_tile_mode = META_TILE_NONE ;
2012-02-24 11:08:55 -05:00
window - > tile_match = NULL ;
2012-03-18 14:53:31 -04:00
if ( window - > shaded | | window - > minimized )
2012-02-24 11:08:55 -05:00
return ;
2012-03-18 14:53:31 -04:00
if ( META_WINDOW_TILED_LEFT ( window ) )
match_tile_mode = META_TILE_RIGHT ;
else if ( META_WINDOW_TILED_RIGHT ( window ) )
match_tile_mode = META_TILE_LEFT ;
else
return ;
2012-02-24 11:08:55 -05:00
stack = window - > screen - > stack ;
for ( match = meta_stack_get_top ( stack ) ;
match ;
match = meta_stack_get_below ( stack , match , FALSE ) )
{
if ( ! match - > shaded & &
! match - > minimized & &
match - > tile_mode = = match_tile_mode & &
match - > monitor = = window - > monitor & &
meta_window_get_workspace ( match ) = = meta_window_get_workspace ( window ) )
break ;
}
if ( match )
{
MetaWindow * above , * bottommost , * topmost ;
MetaRectangle above_rect , bottommost_rect , topmost_rect ;
if ( meta_stack_windows_cmp ( window - > screen - > stack , match , window ) > 0 )
{
topmost = match ;
bottommost = window ;
}
else
{
topmost = window ;
bottommost = match ;
}
2013-11-19 12:24:45 -05:00
meta_window_get_frame_rect ( bottommost , & bottommost_rect ) ;
meta_window_get_frame_rect ( topmost , & topmost_rect ) ;
2012-02-24 11:08:55 -05:00
/*
* If there ' s a window stacked in between which is partially visible
* behind the topmost tile we don ' t consider the tiles to match .
*/
for ( above = meta_stack_get_above ( stack , bottommost , FALSE ) ;
above & & above ! = topmost ;
above = meta_stack_get_above ( stack , above , FALSE ) )
{
if ( above - > minimized | |
above - > monitor ! = window - > monitor | |
meta_window_get_workspace ( above ) ! = meta_window_get_workspace ( window ) )
continue ;
2013-11-19 12:24:45 -05:00
meta_window_get_frame_rect ( above , & above_rect ) ;
2012-02-24 11:08:55 -05:00
if ( meta_rectangle_overlap ( & above_rect , & bottommost_rect ) & &
meta_rectangle_overlap ( & above_rect , & topmost_rect ) )
return ;
}
window - > tile_match = match ;
}
}
2013-04-26 11:33:11 -04:00
2013-08-30 03:40:36 -04:00
void
meta_window_set_title ( MetaWindow * window ,
const char * title )
{
g_free ( window - > title ) ;
window - > title = g_strdup ( title ) ;
if ( window - > frame )
meta_ui_set_frame_title ( window - > screen - > ui ,
window - > frame - > xwindow ,
window - > title ) ;
2014-03-18 21:19:51 -04:00
meta_window_update_desc ( window ) ;
2013-08-30 03:40:36 -04:00
g_object_notify ( G_OBJECT ( window ) , " title " ) ;
}
void
meta_window_set_wm_class ( MetaWindow * window ,
const char * wm_class ,
const char * wm_instance )
{
g_free ( window - > res_class ) ;
g_free ( window - > res_name ) ;
window - > res_name = g_strdup ( wm_instance ) ;
window - > res_class = g_strdup ( wm_class ) ;
g_object_notify ( G_OBJECT ( window ) , " wm-class " ) ;
}
void
meta_window_set_gtk_dbus_properties ( MetaWindow * window ,
const char * application_id ,
const char * unique_bus_name ,
const char * appmenu_path ,
const char * menubar_path ,
const char * application_object_path ,
const char * window_object_path )
{
g_object_freeze_notify ( G_OBJECT ( window ) ) ;
g_free ( window - > gtk_application_id ) ;
window - > gtk_application_id = g_strdup ( application_id ) ;
g_object_notify ( G_OBJECT ( window ) , " gtk-application-id " ) ;
g_free ( window - > gtk_unique_bus_name ) ;
window - > gtk_unique_bus_name = g_strdup ( unique_bus_name ) ;
g_object_notify ( G_OBJECT ( window ) , " gtk-unique-bus-name " ) ;
g_free ( window - > gtk_app_menu_object_path ) ;
window - > gtk_app_menu_object_path = g_strdup ( appmenu_path ) ;
g_object_notify ( G_OBJECT ( window ) , " gtk-app-menu-object-path " ) ;
g_free ( window - > gtk_menubar_object_path ) ;
window - > gtk_menubar_object_path = g_strdup ( menubar_path ) ;
g_object_notify ( G_OBJECT ( window ) , " gtk-menubar-object-path " ) ;
g_free ( window - > gtk_application_object_path ) ;
window - > gtk_application_object_path = g_strdup ( application_object_path ) ;
g_object_notify ( G_OBJECT ( window ) , " gtk-application-object-path " ) ;
g_free ( window - > gtk_window_object_path ) ;
window - > gtk_window_object_path = g_strdup ( window_object_path ) ;
g_object_notify ( G_OBJECT ( window ) , " gtk-window-object-path " ) ;
g_object_thaw_notify ( G_OBJECT ( window ) ) ;
}
2013-09-16 08:44:37 -04:00
void
meta_window_set_transient_for ( MetaWindow * window ,
MetaWindow * parent )
{
if ( meta_window_appears_focused ( window ) & & window - > transient_for ! = None )
meta_window_propagate_focus_appearance ( window , FALSE ) ;
/* may now be a dialog */
2014-03-13 16:49:17 -04:00
if ( window - > client_type = = META_WINDOW_CLIENT_TYPE_X11 )
meta_window_x11_recalc_window_type ( window ) ;
2013-09-16 08:44:37 -04:00
if ( ! window - > constructing )
{
/* If the window attaches, detaches, or changes attached
* parents , we need to destroy the MetaWindow and let a new one
* be created ( which happens as a side effect of
* meta_window_unmanage ( ) ) . The condition below is correct
* because we know window - > transient_for has changed .
*/
if ( window - > attached | | meta_window_should_attach_to_parent ( window ) )
{
guint32 timestamp ;
timestamp = meta_display_get_current_time_roundtrip ( window - > display ) ;
meta_window_unmanage ( window , timestamp ) ;
return ;
}
}
/* update stacking constraints */
if ( ! window - > override_redirect )
meta_stack_update_transient ( window - > screen - > stack , window ) ;
/* We know this won't create a reference cycle because we check for loops */
g_clear_object ( & window - > transient_for ) ;
window - > transient_for = parent ? g_object_ref ( parent ) : NULL ;
/* possibly change its group. We treat being a window's transient as
* equivalent to making it your group leader , to work around shortcomings
* in programs such as xmms - - see # 328211.
*/
if ( window - > xtransient_for ! = None & &
window - > xgroup_leader ! = None & &
window - > xtransient_for ! = window - > xgroup_leader )
meta_window_group_leader_changed ( window ) ;
if ( ! window - > constructing & & ! window - > override_redirect )
meta_window_queue ( window , META_QUEUE_MOVE_RESIZE ) ;
if ( meta_window_appears_focused ( window ) & & window - > transient_for ! = None )
meta_window_propagate_focus_appearance ( window , TRUE ) ;
}
2013-10-16 00:06:10 -04:00
2013-12-06 13:44:31 -05:00
void
meta_window_set_opacity ( MetaWindow * window ,
guint opacity )
{
window - > opacity = opacity ;
2014-03-18 17:06:42 -04:00
meta_compositor_window_opacity_changed ( window - > display - > compositor , window ) ;
2013-12-06 13:44:31 -05:00
}
2013-10-16 00:06:10 -04:00
static void
reset_ignored_crossing_serials ( MetaDisplay * display )
{
int i ;
i = 0 ;
while ( i < N_IGNORED_CROSSING_SERIALS )
{
display - > ignored_crossing_serials [ i ] = 0 ;
+ + i ;
}
display - > ungrab_should_not_cause_focus_window = None ;
}
typedef struct
{
MetaWindow * window ;
int pointer_x ;
int pointer_y ;
} MetaFocusData ;
static void
mouse_mode_focus ( MetaWindow * window ,
guint32 timestamp )
{
MetaDisplay * display = window - > display ;
if ( window - > type ! = META_WINDOW_DESKTOP )
{
meta_topic ( META_DEBUG_FOCUS ,
" Focusing %s at time %u. \n " , window - > desc , timestamp ) ;
meta_window_focus ( window , timestamp ) ;
if ( meta_prefs_get_auto_raise ( ) )
meta_display_queue_autoraise_callback ( display , window ) ;
else
meta_topic ( META_DEBUG_FOCUS , " Auto raise is disabled \n " ) ;
}
else
{
/* In mouse focus mode, we defocus when the mouse *enters*
* the DESKTOP window , instead of defocusing on LeaveNotify .
* This is because having the mouse enter override - redirect
* child windows unfortunately causes LeaveNotify events that
* we can ' t distinguish from the mouse actually leaving the
* toplevel window as we expect . But , since we filter out
* EnterNotify events on override - redirect windows , this
* alternative mechanism works great .
*/
if ( meta_prefs_get_focus_mode ( ) = = G_DESKTOP_FOCUS_MODE_MOUSE & &
display - > focus_window ! = NULL )
{
meta_topic ( META_DEBUG_FOCUS ,
" Unsetting focus from %s due to mouse entering "
" the DESKTOP window \n " ,
display - > focus_window - > desc ) ;
meta_display_focus_the_no_focus_window ( display ,
window - > screen ,
timestamp ) ;
}
}
}
static gboolean
window_focus_on_pointer_rest_callback ( gpointer data )
{
MetaFocusData * focus_data = data ;
MetaWindow * window = focus_data - > window ;
MetaDisplay * display = window - > display ;
MetaScreen * screen = window - > screen ;
2014-03-02 17:27:50 -05:00
int root_x , root_y ;
2013-10-16 00:06:10 -04:00
guint32 timestamp ;
2014-03-02 17:27:50 -05:00
ClutterActor * child ;
2013-10-16 00:06:10 -04:00
if ( meta_prefs_get_focus_mode ( ) = = G_DESKTOP_FOCUS_MODE_CLICK )
goto out ;
2014-03-02 17:27:50 -05:00
meta_cursor_tracker_get_pointer ( screen - > cursor_tracker ,
& root_x , & root_y , NULL ) ;
2013-10-16 00:06:10 -04:00
if ( root_x ! = focus_data - > pointer_x | |
root_y ! = focus_data - > pointer_y )
{
focus_data - > pointer_x = root_x ;
focus_data - > pointer_y = root_y ;
return TRUE ;
}
2014-03-02 17:27:50 -05:00
child = clutter_stage_get_actor_at_pos ( CLUTTER_STAGE ( clutter_stage_get_default ( ) ) ,
CLUTTER_PICK_REACTIVE , root_x , root_y ) ;
if ( ! META_IS_SURFACE_ACTOR ( child ) )
2013-10-16 00:06:10 -04:00
goto out ;
window =
meta_stack_get_default_focus_window_at_point ( screen - > stack ,
screen - > active_workspace ,
2014-03-02 17:27:50 -05:00
NULL , root_x , root_y ) ;
2013-10-16 00:06:10 -04:00
if ( window = = NULL )
goto out ;
timestamp = meta_display_get_current_time_roundtrip ( display ) ;
mouse_mode_focus ( window , timestamp ) ;
out :
display - > focus_timeout_id = 0 ;
return FALSE ;
}
/* The interval, in milliseconds, we use in focus-follows-mouse
* mode to check whether the pointer has stopped moving after a
* crossing event .
*/
# define FOCUS_TIMEOUT_DELAY 25
static void
queue_focus_callback ( MetaDisplay * display ,
MetaWindow * window ,
int pointer_x ,
int pointer_y )
{
MetaFocusData * focus_data ;
focus_data = g_new ( MetaFocusData , 1 ) ;
focus_data - > window = window ;
focus_data - > pointer_x = pointer_x ;
focus_data - > pointer_y = pointer_y ;
if ( display - > focus_timeout_id ! = 0 )
g_source_remove ( display - > focus_timeout_id ) ;
display - > focus_timeout_id =
g_timeout_add_full ( G_PRIORITY_DEFAULT ,
FOCUS_TIMEOUT_DELAY ,
window_focus_on_pointer_rest_callback ,
focus_data ,
g_free ) ;
}
void
meta_window_handle_enter ( MetaWindow * window ,
guint32 timestamp ,
guint root_x ,
guint root_y )
{
MetaDisplay * display = window - > display ;
switch ( meta_prefs_get_focus_mode ( ) )
{
case G_DESKTOP_FOCUS_MODE_SLOPPY :
case G_DESKTOP_FOCUS_MODE_MOUSE :
display - > mouse_mode = TRUE ;
if ( window - > type ! = META_WINDOW_DOCK )
{
if ( meta_prefs_get_focus_change_on_pointer_rest ( ) )
queue_focus_callback ( display , window , root_x , root_y ) ;
else
mouse_mode_focus ( window , timestamp ) ;
/* stop ignoring stuff */
reset_ignored_crossing_serials ( display ) ;
}
break ;
case G_DESKTOP_FOCUS_MODE_CLICK :
break ;
}
}
2013-11-19 17:33:57 -05:00
void
meta_window_set_surface_mapped ( MetaWindow * window ,
gboolean surface_mapped )
{
if ( window - > surface_mapped = = ( guint ) surface_mapped )
return ;
window - > surface_mapped = surface_mapped ;
meta_window_queue ( window , META_QUEUE_CALC_SHOWING ) ;
}
2013-11-21 14:01:40 -05:00
2013-12-06 00:57:11 -05:00
Window
meta_window_get_toplevel_xwindow ( MetaWindow * window )
{
return window - > frame ? window - > frame - > xwindow : window - > xwindow ;
}
2014-02-07 17:29:35 -05:00
void
meta_window_set_custom_frame_extents ( MetaWindow * window ,
GtkBorder * extents )
{
if ( extents )
{
window - > has_custom_frame_extents = TRUE ;
window - > custom_frame_extents = * extents ;
}
else
{
window - > has_custom_frame_extents = FALSE ;
}
meta_window_queue ( window , META_QUEUE_MOVE_RESIZE ) ;
}
2014-03-13 18:15:26 -04:00
gboolean
meta_window_can_maximize ( MetaWindow * window )
{
return window - > has_maximize_func ;
}
gboolean
meta_window_can_minimize ( MetaWindow * window )
{
return window - > has_minimize_func ;
}
gboolean
meta_window_can_shade ( MetaWindow * window )
{
return window - > has_shade_func ;
}
gboolean
meta_window_can_close ( MetaWindow * window )
{
return window - > has_close_func ;
}
gboolean
meta_window_is_always_on_all_workspaces ( MetaWindow * window )
{
return window - > always_sticky ;
}
gboolean
meta_window_is_above ( MetaWindow * window )
{
return window - > wm_state_above ;
}
gboolean
meta_window_allows_move ( MetaWindow * window )
{
return META_WINDOW_ALLOWS_MOVE ( window ) ;
}
gboolean
meta_window_allows_resize ( MetaWindow * window )
{
return META_WINDOW_ALLOWS_RESIZE ( window ) ;
}
2014-03-18 13:51:36 -04:00
void
meta_window_set_urgent ( MetaWindow * window ,
gboolean urgent )
{
2014-03-18 13:53:42 -04:00
if ( window - > urgent = = urgent )
2014-03-18 13:51:36 -04:00
return ;
2014-03-18 13:53:42 -04:00
window - > urgent = urgent ;
2014-03-18 16:45:53 -04:00
g_object_notify ( G_OBJECT ( window ) , " urgent " ) ;
2014-03-18 13:51:36 -04:00
if ( urgent )
g_signal_emit_by_name ( window - > display , " window-marked-urgent " , window ) ;
}