mutter/src/core/testboxes.c

1415 lines
49 KiB
C
Raw Normal View History

/* -*- mode: C; c-file-style: "gnu"; indent-tabs-mode: nil; -*- */
/* Mutter box operation testing program */
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 14:58:50 +00:00
2014-05-02 13:34:02 +00:00
/*
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 14:58:50 +00:00
* Copyright (C) 2005 Elijah Newren
2014-05-02 13:34:02 +00:00
*
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 14:58:50 +00:00
* 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.
2014-05-02 13:34:02 +00:00
*
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 14:58:50 +00:00
* You should have received a copy of the GNU General Public License
* along with this program; if not, see <http://www.gnu.org/licenses/>.
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 14:58:50 +00:00
*/
#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 14:58:50 +00:00
#include <glib.h>
#include <stdlib.h>
#include <stdio.h>
#include <X11/Xutil.h> /* Just for the definition of the various gravities */
#include <time.h> /* To initialize random seed */
#define NUM_RANDOM_RUNS 10000
static void
init_random_ness ()
{
srand(time(NULL));
}
static void
get_random_rect (MetaRectangle *rect)
{
rect->x = rand () % 1600;
rect->y = rand () % 1200;
rect->width = rand () % 1600 + 1;
rect->height = rand () % 1200 + 1;
}
static MetaRectangle*
new_meta_rect (int x, int y, int width, int height)
{
MetaRectangle* temporary;
temporary = g_new (MetaRectangle, 1);
temporary->x = x;
temporary->y = y;
temporary->width = width;
temporary->height = height;
return temporary;
}
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
2007-04-03 03:41:10 +00:00
static MetaStrut*
new_meta_strut (int x, int y, int width, int height, int side)
{
MetaStrut* temporary;
temporary = g_new (MetaStrut, 1);
temporary->rect = meta_rect(x, y, width, height);
temporary->side = side;
return temporary;
}
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 14:58:50 +00:00
static MetaEdge*
new_screen_edge (int x, int y, int width, int height, int side_type)
{
MetaEdge* temporary;
temporary = g_new (MetaEdge, 1);
temporary->rect.x = x;
temporary->rect.y = y;
temporary->rect.width = width;
temporary->rect.height = height;
temporary->side_type = side_type;
temporary->edge_type = META_EDGE_SCREEN;
return temporary;
}
static MetaEdge*
new_monitor_edge (int x, int y, int width, int height, int side_type)
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 14:58:50 +00:00
{
MetaEdge* temporary;
temporary = g_new (MetaEdge, 1);
temporary->rect.x = x;
temporary->rect.y = y;
temporary->rect.width = width;
temporary->rect.height = height;
temporary->side_type = side_type;
temporary->edge_type = META_EDGE_MONITOR;
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 14:58:50 +00:00
return temporary;
}
static void
test_area ()
{
MetaRectangle temp;
int i;
for (i = 0; i < NUM_RANDOM_RUNS; i++)
{
get_random_rect (&temp);
g_assert (meta_rectangle_area (&temp) == temp.width * temp.height);
}
temp = meta_rect (0, 0, 5, 7);
g_assert (meta_rectangle_area (&temp) == 35);
printf ("%s passed.\n", G_STRFUNC);
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 14:58:50 +00:00
}
static void
test_intersect ()
{
MetaRectangle a = {100, 200, 50, 40};
MetaRectangle b = { 0, 50, 110, 152};
MetaRectangle c = { 0, 0, 10, 10};
MetaRectangle d = {100, 100, 50, 50};
MetaRectangle b_intersect_d = {100, 100, 10, 50};
MetaRectangle temp;
MetaRectangle temp2;
meta_rectangle_intersect (&a, &b, &temp);
temp2 = meta_rect (100, 200, 10, 2);
g_assert (meta_rectangle_equal (&temp, &temp2));
g_assert (meta_rectangle_area (&temp) == 20);
meta_rectangle_intersect (&a, &c, &temp);
g_assert (meta_rectangle_area (&temp) == 0);
meta_rectangle_intersect (&a, &d, &temp);
g_assert (meta_rectangle_area (&temp) == 0);
meta_rectangle_intersect (&b, &d, &b);
g_assert (meta_rectangle_equal (&b, &b_intersect_d));
printf ("%s passed.\n", G_STRFUNC);
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 14:58:50 +00:00
}
static void
test_equal ()
{
MetaRectangle a = {10, 12, 4, 18};
MetaRectangle b = a;
MetaRectangle c = {10, 12, 4, 19};
MetaRectangle d = {10, 12, 7, 18};
MetaRectangle e = {10, 62, 4, 18};
MetaRectangle f = {27, 12, 4, 18};
g_assert ( meta_rectangle_equal (&a, &b));
g_assert (!meta_rectangle_equal (&a, &c));
g_assert (!meta_rectangle_equal (&a, &d));
g_assert (!meta_rectangle_equal (&a, &e));
g_assert (!meta_rectangle_equal (&a, &f));
printf ("%s passed.\n", G_STRFUNC);
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 14:58:50 +00:00
}
static void
test_overlap_funcs ()
{
MetaRectangle temp1, temp2;
int i;
for (i = 0; i < NUM_RANDOM_RUNS; i++)
{
get_random_rect (&temp1);
get_random_rect (&temp2);
g_assert (meta_rectangle_overlap (&temp1, &temp2) ==
(meta_rectangle_horiz_overlap (&temp1, &temp2) &&
meta_rectangle_vert_overlap (&temp1, &temp2)));
}
temp1 = meta_rect ( 0, 0, 10, 10);
temp2 = meta_rect (20, 0, 10, 5);
g_assert (!meta_rectangle_overlap (&temp1, &temp2));
g_assert (!meta_rectangle_horiz_overlap (&temp1, &temp2));
g_assert ( meta_rectangle_vert_overlap (&temp1, &temp2));
printf ("%s passed.\n", G_STRFUNC);
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 14:58:50 +00:00
}
static void
test_basic_fitting ()
{
MetaRectangle temp1, temp2, temp3;
int i;
/* Four cases:
* case temp1 fits temp2 temp1 could fit temp2
* 1 Y Y
* 2 N Y
* 3 Y N
* 4 N N
* Of the four cases, case 3 is impossible. An alternate way of looking
* at this table is that either the middle column must be no, or the last
* column must be yes. So we test that. Also, we can repeat the test
* reversing temp1 and temp2.
*/
for (i = 0; i < NUM_RANDOM_RUNS; i++)
{
get_random_rect (&temp1);
get_random_rect (&temp2);
g_assert (meta_rectangle_contains_rect (&temp1, &temp2) == FALSE ||
meta_rectangle_could_fit_rect (&temp1, &temp2) == TRUE);
g_assert (meta_rectangle_contains_rect (&temp2, &temp1) == FALSE ||
meta_rectangle_could_fit_rect (&temp2, &temp1) == TRUE);
}
temp1 = meta_rect ( 0, 0, 10, 10);
temp2 = meta_rect ( 5, 5, 5, 5);
temp3 = meta_rect ( 8, 2, 3, 7);
g_assert ( meta_rectangle_contains_rect (&temp1, &temp2));
g_assert (!meta_rectangle_contains_rect (&temp2, &temp1));
g_assert (!meta_rectangle_contains_rect (&temp1, &temp3));
g_assert ( meta_rectangle_could_fit_rect (&temp1, &temp3));
g_assert (!meta_rectangle_could_fit_rect (&temp3, &temp2));
printf ("%s passed.\n", G_STRFUNC);
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 14:58:50 +00:00
}
static void
free_strut_list (GSList *struts)
{
GSList *tmp = struts;
while (tmp)
{
g_free (tmp->data);
tmp = tmp->next;
}
g_slist_free (struts);
}
static GSList*
get_strut_list (int which)
{
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
2007-04-03 03:41:10 +00:00
GSList *ans;
MetaDirection wc = 0; /* wc == who cares? ;-) */
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 14:58:50 +00:00
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
2007-04-03 03:41:10 +00:00
ans = 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 14:58:50 +00:00
g_assert (which >=0 && which <= 6);
switch (which)
{
case 0:
break;
case 1:
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
2007-04-03 03:41:10 +00:00
ans = g_slist_prepend (ans, new_meta_strut ( 0, 0, 1600, 20, wc));
ans = g_slist_prepend (ans, new_meta_strut ( 400, 1160, 1600, 40, wc));
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 14:58:50 +00:00
break;
case 2:
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
2007-04-03 03:41:10 +00:00
ans = g_slist_prepend (ans, new_meta_strut ( 0, 0, 1600, 20, wc));
ans = g_slist_prepend (ans, new_meta_strut ( 800, 1100, 400, 100, wc));
ans = g_slist_prepend (ans, new_meta_strut ( 300, 1150, 150, 50, wc));
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 14:58:50 +00:00
break;
case 3:
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
2007-04-03 03:41:10 +00:00
ans = g_slist_prepend (ans, new_meta_strut ( 0, 0, 1600, 20, wc));
ans = g_slist_prepend (ans, new_meta_strut ( 800, 1100, 400, 100, wc));
ans = g_slist_prepend (ans, new_meta_strut ( 300, 1150, 80, 50, wc));
ans = g_slist_prepend (ans, new_meta_strut ( 700, 525, 200, 150, wc));
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 14:58:50 +00:00
break;
case 4:
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
2007-04-03 03:41:10 +00:00
ans = g_slist_prepend (ans, new_meta_strut ( 0, 0, 800, 1200, wc));
ans = g_slist_prepend (ans, new_meta_strut ( 800, 0, 1600, 20, wc));
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 14:58:50 +00:00
break;
case 5:
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
2007-04-03 03:41:10 +00:00
ans = g_slist_prepend (ans, new_meta_strut ( 800, 0, 1600, 20, wc));
ans = g_slist_prepend (ans, new_meta_strut ( 0, 0, 800, 1200, wc));
ans = g_slist_prepend (ans, new_meta_strut ( 800, 10, 800, 1200, wc));
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 14:58:50 +00:00
break;
case 6:
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
2007-04-03 03:41:10 +00:00
ans = g_slist_prepend (ans, new_meta_strut ( 0, 0, 1600, 40, wc));
ans = g_slist_prepend (ans, new_meta_strut ( 0, 0, 1600, 20, wc));
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 14:58:50 +00:00
break;
}
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
2007-04-03 03:41:10 +00:00
return ans;
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 14:58:50 +00:00
}
static GList*
get_screen_region (int which)
{
GList *ret;
GSList *struts;
MetaRectangle basic_rect;
basic_rect = meta_rect (0, 0, 1600, 1200);
ret = NULL;
struts = get_strut_list (which);
ret = meta_rectangle_get_minimal_spanning_set_for_region (&basic_rect, struts);
free_strut_list (struts);
return ret;
}
static GList*
get_screen_edges (int which)
{
GList *ret;
GSList *struts;
MetaRectangle basic_rect;
basic_rect = meta_rect (0, 0, 1600, 1200);
ret = NULL;
struts = get_strut_list (which);
ret = meta_rectangle_find_onscreen_edges (&basic_rect, struts);
free_strut_list (struts);
return ret;
}
static GList*
get_monitor_edges (int which_monitor_set, int which_strut_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 14:58:50 +00:00
{
GList *ret;
GSList *struts;
GList *xins;
xins = NULL;
g_assert (which_monitor_set >=0 && which_monitor_set <= 3);
switch (which_monitor_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 14:58:50 +00:00
{
case 0:
xins = g_list_prepend (xins, new_meta_rect ( 0, 0, 1600, 1200));
break;
case 1:
xins = g_list_prepend (xins, new_meta_rect ( 0, 0, 800, 1200));
xins = g_list_prepend (xins, new_meta_rect (800, 0, 800, 1200));
break;
case 2:
xins = g_list_prepend (xins, new_meta_rect ( 0, 0, 1600, 600));
xins = g_list_prepend (xins, new_meta_rect ( 0, 600, 1600, 600));
break;
case 3:
xins = g_list_prepend (xins, new_meta_rect ( 0, 0, 1600, 600));
xins = g_list_prepend (xins, new_meta_rect ( 0, 600, 800, 600));
xins = g_list_prepend (xins, new_meta_rect (800, 600, 800, 600));
break;
}
ret = NULL;
struts = get_strut_list (which_strut_set);
ret = meta_rectangle_find_nonintersected_monitor_edges (xins, struts);
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 14:58:50 +00:00
free_strut_list (struts);
meta_rectangle_free_list_and_elements (xins);
return ret;
}
#if 0
static void
test_merge_regions ()
{
/* logarithmically distributed random number of struts (range?)
* logarithmically distributed random size of struts (up to screen size???)
* uniformly distributed location of center of struts (within screen)
* merge all regions that are possible
* print stats on problem setup
2014-05-02 13:34:02 +00:00
* number of (non-completely-occluded?) struts
* percentage of screen covered
* length of resulting non-minimal spanning set
* length of resulting minimal spanning set
* print stats on merged regions:
* number boxes merged
* number of those merges that were of the form A contains B
* number of those merges that were of the form A partially contains B
* number of those merges that were of the form A is adjacent to B
*/
GList* region;
GList* compare;
int num_contains, num_merged, num_part_contains, num_adjacent;
num_contains = num_merged = num_part_contains = num_adjacent = 0;
compare = region = get_screen_region (2);
g_assert (region);
printf ("Merging stats:\n");
printf (" Length of initial list: %d\n", g_list_length (region));
#ifdef PRINT_DEBUG
char rect1[RECT_LENGTH], rect2[RECT_LENGTH];
char region_list[(RECT_LENGTH + 2) * g_list_length (region)];
meta_rectangle_region_to_string (region, ", ", region_list);
printf (" Initial rectangles: %s\n", region_list);
#endif
while (compare && compare->next)
{
MetaRectangle *a = compare->data;
GList *other = compare->next;
g_assert (a->width > 0 && a->height > 0);
while (other)
{
MetaRectangle *b = other->data;
GList *delete_me = NULL;
g_assert (b->width > 0 && b->height > 0);
#ifdef PRINT_DEBUG
printf (" -- Comparing %s to %s --\n",
meta_rectangle_to_string (a, rect1),
meta_rectangle_to_string (b, rect2));
#endif
/* If a contains b, just remove b */
if (meta_rectangle_contains_rect (a, b))
{
delete_me = other;
num_contains++;
num_merged++;
}
/* If b contains a, just remove a */
else if (meta_rectangle_contains_rect (a, b))
{
delete_me = compare;
num_contains++;
num_merged++;
}
/* If a and b might be mergeable horizontally */
else if (a->y == b->y && a->height == b->height)
{
/* If a and b overlap */
if (meta_rectangle_overlap (a, b))
{
int new_x = MIN (a->x, b->x);
a->width = MAX (a->x + a->width, b->x + b->width) - new_x;
a->x = new_x;
delete_me = other;
num_part_contains++;
num_merged++;
}
/* If a and b are adjacent */
else if (a->x + a->width == b->x || a->x == b->x + b->width)
{
int new_x = MIN (a->x, b->x);
a->width = MAX (a->x + a->width, b->x + b->width) - new_x;
a->x = new_x;
delete_me = other;
num_adjacent++;
num_merged++;
}
}
/* If a and b might be mergeable vertically */
else if (a->x == b->x && a->width == b->width)
{
/* If a and b overlap */
if (meta_rectangle_overlap (a, b))
{
int new_y = MIN (a->y, b->y);
a->height = MAX (a->y + a->height, b->y + b->height) - new_y;
a->y = new_y;
delete_me = other;
num_part_contains++;
num_merged++;
}
/* If a and b are adjacent */
else if (a->y + a->height == b->y || a->y == b->y + b->height)
{
int new_y = MIN (a->y, b->y);
a->height = MAX (a->y + a->height, b->y + b->height) - new_y;
a->y = new_y;
delete_me = other;
num_adjacent++;
num_merged++;
}
}
other = other->next;
/* Delete any rectangle in the list that is no longer wanted */
if (delete_me != NULL)
{
#ifdef PRINT_DEBUG
MetaRectangle *bla = delete_me->data;
printf (" Deleting rect %s\n",
meta_rectangle_to_string (bla, rect1));
#endif
/* Deleting the rect we're compare others to is a little tricker */
if (compare == delete_me)
{
compare = compare->next;
other = compare->next;
a = compare->data;
}
/* Okay, we can free it now */
g_free (delete_me->data);
region = g_list_delete_link (region, delete_me);
}
#ifdef PRINT_DEBUG
char region_list[(RECT_LENGTH + 2) * g_list_length (region)];
meta_rectangle_region_to_string (region, ", ", region_list);
printf (" After comparison, new list is: %s\n", region_list);
#endif
}
compare = compare->next;
}
2014-05-02 13:34:02 +00:00
printf (" Num rectangles contained in others : %d\n",
num_contains);
2014-05-02 13:34:02 +00:00
printf (" Num rectangles partially contained in others: %d\n",
num_part_contains);
2014-05-02 13:34:02 +00:00
printf (" Num rectangles adjacent to others : %d\n",
num_adjacent);
printf (" Num rectangles merged with others : %d\n",
num_merged);
#ifdef PRINT_DEBUG
char region_list2[(RECT_LENGTH + 2) * g_list_length (region)];
meta_rectangle_region_to_string (region, ", ", region_list2);
printf (" Final rectangles: %s\n", region_list2);
#endif
meta_rectangle_free_spanning_set (region);
region = NULL;
printf ("%s passed.\n", G_STRFUNC);
}
#endif
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 14:58:50 +00:00
static void
verify_lists_are_equal (GList *code, GList *answer)
{
int which = 0;
while (code && answer)
{
MetaRectangle *a = code->data;
MetaRectangle *b = answer->data;
if (a->x != b->x ||
a->y != b->y ||
a->width != b->width ||
a->height != b->height)
{
g_error ("%dth item in code answer answer lists do not match; "
"code rect: %d,%d + %d,%d; answer rect: %d,%d + %d,%d\n",
which,
a->x, a->y, a->width, a->height,
b->x, b->y, b->width, b->height);
}
code = code->next;
answer = answer->next;
which++;
}
/* Ought to be at the end of both lists; check if we aren't */
if (code)
{
MetaRectangle *tmp = code->data;
g_error ("code list longer than answer list by %d items; "
"first extra item: %d,%d +%d,%d\n",
g_list_length (code),
tmp->x, tmp->y, tmp->width, tmp->height);
}
if (answer)
{
MetaRectangle *tmp = answer->data;
g_error ("answer list longer than code list by %d items; "
"first extra item: %d,%d +%d,%d\n",
g_list_length (answer),
tmp->x, tmp->y, tmp->width, tmp->height);
}
}
static void
test_regions_okay ()
{
GList* region;
GList* tmp;
2014-05-02 13:34:02 +00:00
/*************************************************************/
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 14:58:50 +00:00
/* Make sure test region 0 has the right spanning rectangles */
2014-05-02 13:34:02 +00:00
/*************************************************************/
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 14:58:50 +00:00
region = get_screen_region (0);
tmp = NULL;
tmp = g_list_prepend (tmp, new_meta_rect (0, 0, 1600, 1200));
verify_lists_are_equal (region, tmp);
meta_rectangle_free_list_and_elements (tmp);
meta_rectangle_free_list_and_elements (region);
2014-05-02 13:34:02 +00:00
/*************************************************************/
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 14:58:50 +00:00
/* Make sure test region 1 has the right spanning rectangles */
2014-05-02 13:34:02 +00:00
/*************************************************************/
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 14:58:50 +00:00
region = get_screen_region (1);
tmp = NULL;
tmp = g_list_prepend (tmp, new_meta_rect (0, 20, 400, 1180));
tmp = g_list_prepend (tmp, new_meta_rect (0, 20, 1600, 1140));
verify_lists_are_equal (region, tmp);
meta_rectangle_free_list_and_elements (tmp);
meta_rectangle_free_list_and_elements (region);
/*************************************************************/
/* Make sure test region 2 has the right spanning rectangles */
2014-05-02 13:34:02 +00:00
/*************************************************************/
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 14:58:50 +00:00
region = get_screen_region (2);
tmp = NULL;
tmp = g_list_prepend (tmp, new_meta_rect ( 0, 20, 300, 1180));
tmp = g_list_prepend (tmp, new_meta_rect ( 450, 20, 350, 1180));
tmp = g_list_prepend (tmp, new_meta_rect (1200, 20, 400, 1180));
tmp = g_list_prepend (tmp, new_meta_rect ( 0, 20, 800, 1130));
tmp = g_list_prepend (tmp, new_meta_rect ( 0, 20, 1600, 1080));
verify_lists_are_equal (region, tmp);
meta_rectangle_free_list_and_elements (tmp);
meta_rectangle_free_list_and_elements (region);
/*************************************************************/
/* Make sure test region 3 has the right spanning rectangles */
2014-05-02 13:34:02 +00:00
/*************************************************************/
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 14:58:50 +00:00
region = get_screen_region (3);
tmp = NULL;
tmp = g_list_prepend (tmp, new_meta_rect ( 380, 675, 420, 525)); /* 220500 */
tmp = g_list_prepend (tmp, new_meta_rect ( 0, 20, 300, 1180)); /* 354000 */
tmp = g_list_prepend (tmp, new_meta_rect ( 380, 20, 320, 1180)); /* 377600 */
tmp = g_list_prepend (tmp, new_meta_rect ( 0, 675, 800, 475)); /* 380000 */
tmp = g_list_prepend (tmp, new_meta_rect (1200, 20, 400, 1180)); /* 472000 */
tmp = g_list_prepend (tmp, new_meta_rect ( 0, 675, 1600, 425)); /* 680000 */
tmp = g_list_prepend (tmp, new_meta_rect ( 900, 20, 700, 1080)); /* 756000 */
tmp = g_list_prepend (tmp, new_meta_rect ( 0, 20, 700, 1130)); /* 791000 */
tmp = g_list_prepend (tmp, new_meta_rect ( 0, 20, 1600, 505)); /* 808000 */
#if 0
printf ("Got to here...\n");
char region_list[(RECT_LENGTH+2) * g_list_length (region)];
char tmp_list[ (RECT_LENGTH+2) * g_list_length (tmp)];
meta_rectangle_region_to_string (region, ", ", region_list);
meta_rectangle_region_to_string (region, ", ", tmp_list);
printf ("%s vs. %s\n", region_list, tmp_list);
#endif
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 14:58:50 +00:00
verify_lists_are_equal (region, tmp);
meta_rectangle_free_list_and_elements (tmp);
meta_rectangle_free_list_and_elements (region);
/*************************************************************/
/* Make sure test region 4 has the right spanning rectangles */
2014-05-02 13:34:02 +00:00
/*************************************************************/
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 14:58:50 +00:00
region = get_screen_region (4);
tmp = NULL;
tmp = g_list_prepend (tmp, new_meta_rect ( 800, 20, 800, 1180));
verify_lists_are_equal (region, tmp);
meta_rectangle_free_list_and_elements (tmp);
meta_rectangle_free_list_and_elements (region);
/*************************************************************/
/* Make sure test region 5 has the right spanning rectangles */
2014-05-02 13:34:02 +00:00
/*************************************************************/
printf ("The next test intentionally causes a warning, "
"but it can be ignored.\n");
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 14:58:50 +00:00
region = get_screen_region (5);
verify_lists_are_equal (region, NULL);
/* FIXME: Still to do:
* - Create random struts and check the regions somehow
*/
printf ("%s passed.\n", G_STRFUNC);
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 14:58:50 +00:00
}
static void
test_region_fitting ()
{
GList* region;
MetaRectangle rect;
/* See test_basic_fitting() for how/why these automated random tests work */
int i;
region = get_screen_region (3);
for (i = 0; i < NUM_RANDOM_RUNS; i++)
{
get_random_rect (&rect);
g_assert (meta_rectangle_contained_in_region (region, &rect) == FALSE ||
meta_rectangle_could_fit_in_region (region, &rect) == TRUE);
}
meta_rectangle_free_list_and_elements (region);
/* Do some manual tests too */
region = get_screen_region (1);
rect = meta_rect (50, 50, 400, 400);
g_assert (meta_rectangle_could_fit_in_region (region, &rect));
g_assert (meta_rectangle_contained_in_region (region, &rect));
rect = meta_rect (250, 0, 500, 1150);
g_assert (!meta_rectangle_could_fit_in_region (region, &rect));
g_assert (!meta_rectangle_contained_in_region (region, &rect));
rect = meta_rect (250, 0, 400, 400);
g_assert (meta_rectangle_could_fit_in_region (region, &rect));
g_assert (!meta_rectangle_contained_in_region (region, &rect));
meta_rectangle_free_list_and_elements (region);
region = get_screen_region (2);
rect = meta_rect (1000, 50, 600, 1100);
g_assert (meta_rectangle_could_fit_in_region (region, &rect));
g_assert (!meta_rectangle_contained_in_region (region, &rect));
meta_rectangle_free_list_and_elements (region);
printf ("%s passed.\n", G_STRFUNC);
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 14:58:50 +00:00
}
static void
test_clamping_to_region ()
{
GList* region;
MetaRectangle rect;
MetaRectangle min_size;
FixedDirections fixed_directions;
int i;
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 14:58:50 +00:00
min_size.height = min_size.width = 1;
fixed_directions = 0;
region = get_screen_region (3);
for (i = 0; i < NUM_RANDOM_RUNS; i++)
{
MetaRectangle temp;
get_random_rect (&rect);
temp = rect;
meta_rectangle_clamp_to_fit_into_region (region,
fixed_directions,
&rect,
&min_size);
g_assert (meta_rectangle_could_fit_in_region (region, &rect) == TRUE);
g_assert (rect.x == temp.x && rect.y == temp.y);
}
meta_rectangle_free_list_and_elements (region);
/* Do some manual tests too */
region = get_screen_region (1);
rect = meta_rect (50, 50, 10000, 10000);
meta_rectangle_clamp_to_fit_into_region (region,
fixed_directions,
&rect,
&min_size);
g_assert (rect.width == 1600 && rect.height == 1140);
rect = meta_rect (275, -50, 410, 10000);
meta_rectangle_clamp_to_fit_into_region (region,
fixed_directions,
&rect,
&min_size);
g_assert (rect.width == 400 && rect.height == 1180);
rect = meta_rect (50, 50, 10000, 10000);
min_size.height = 1170;
meta_rectangle_clamp_to_fit_into_region (region,
fixed_directions,
&rect,
&min_size);
g_assert (rect.width == 400 && rect.height == 1180);
printf ("The next test intentionally causes a warning, "
"but it can be ignored.\n");
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 14:58:50 +00:00
rect = meta_rect (50, 50, 10000, 10000);
min_size.width = 600; min_size.height = 1170;
meta_rectangle_clamp_to_fit_into_region (region,
fixed_directions,
&rect,
&min_size);
g_assert (rect.width == 600 && rect.height == 1170);
rect = meta_rect (350, 50, 100, 1100);
min_size.width = 1; min_size.height = 1;
fixed_directions = FIXED_DIRECTION_X;
meta_rectangle_clamp_to_fit_into_region (region,
fixed_directions,
&rect,
&min_size);
g_assert (rect.width == 100 && rect.height == 1100);
rect = meta_rect (300, 70, 500, 1100);
min_size.width = 1; min_size.height = 1;
fixed_directions = FIXED_DIRECTION_Y;
meta_rectangle_clamp_to_fit_into_region (region,
fixed_directions,
&rect,
&min_size);
g_assert (rect.width == 400 && rect.height == 1100);
printf ("The next test intentionally causes a warning, "
"but it can be ignored.\n");
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 14:58:50 +00:00
rect = meta_rect (300, 70, 999999, 999999);
min_size.width = 100; min_size.height = 200;
fixed_directions = FIXED_DIRECTION_Y;
meta_rectangle_clamp_to_fit_into_region (region,
fixed_directions,
&rect,
&min_size);
g_assert (rect.width == 100 && rect.height == 999999);
meta_rectangle_free_list_and_elements (region);
printf ("%s passed.\n", G_STRFUNC);
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 14:58:50 +00:00
}
static gboolean
rect_overlaps_region (const GList *spanning_rects,
const MetaRectangle *rect)
{
/* FIXME: Should I move this to boxes.[ch]? */
const GList *temp;
gboolean overlaps;
temp = spanning_rects;
overlaps = FALSE;
while (!overlaps && temp != NULL)
{
overlaps = overlaps || meta_rectangle_overlap (temp->data, rect);
temp = temp->next;
}
return overlaps;
}
gboolean time_to_print = FALSE;
static void
test_clipping_to_region ()
{
GList* region;
MetaRectangle rect, temp;
FixedDirections fixed_directions = 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 14:58:50 +00:00
int i;
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 14:58:50 +00:00
region = get_screen_region (3);
for (i = 0; i < NUM_RANDOM_RUNS; i++)
{
get_random_rect (&rect);
if (rect_overlaps_region (region, &rect))
{
meta_rectangle_clip_to_region (region, 0, &rect);
g_assert (meta_rectangle_contained_in_region (region, &rect) == TRUE);
}
}
meta_rectangle_free_list_and_elements (region);
/* Do some manual tests too */
region = get_screen_region (2);
rect = meta_rect (-50, -10, 10000, 10000);
meta_rectangle_clip_to_region (region,
fixed_directions,
&rect);
g_assert (meta_rectangle_equal (region->data, &rect));
rect = meta_rect (300, 1000, 400, 200);
temp = meta_rect (300, 1000, 400, 150);
meta_rectangle_clip_to_region (region,
fixed_directions,
&rect);
g_assert (meta_rectangle_equal (&rect, &temp));
rect = meta_rect (400, 1000, 300, 200);
temp = meta_rect (450, 1000, 250, 200);
meta_rectangle_clip_to_region (region,
fixed_directions,
&rect);
g_assert (meta_rectangle_equal (&rect, &temp));
rect = meta_rect (400, 1000, 300, 200);
temp = meta_rect (400, 1000, 300, 150);
meta_rectangle_clip_to_region (region,
FIXED_DIRECTION_X,
&rect);
g_assert (meta_rectangle_equal (&rect, &temp));
rect = meta_rect (400, 1000, 300, 200);
temp = meta_rect (400, 1000, 300, 150);
meta_rectangle_clip_to_region (region,
FIXED_DIRECTION_X,
&rect);
g_assert (meta_rectangle_equal (&rect, &temp));
meta_rectangle_free_list_and_elements (region);
printf ("%s passed.\n", G_STRFUNC);
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 14:58:50 +00:00
}
static void
test_shoving_into_region ()
{
GList* region;
MetaRectangle rect, temp;
FixedDirections fixed_directions = 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 14:58:50 +00:00
int i;
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 14:58:50 +00:00
region = get_screen_region (3);
for (i = 0; i < NUM_RANDOM_RUNS; i++)
{
get_random_rect (&rect);
if (meta_rectangle_could_fit_in_region (region, &rect))
{
meta_rectangle_shove_into_region (region, 0, &rect);
g_assert (meta_rectangle_contained_in_region (region, &rect));
}
}
meta_rectangle_free_list_and_elements (region);
/* Do some manual tests too */
region = get_screen_region (2);
rect = meta_rect (300, 1000, 400, 200);
temp = meta_rect (300, 950, 400, 200);
meta_rectangle_shove_into_region (region,
fixed_directions,
&rect);
g_assert (meta_rectangle_equal (&rect, &temp));
rect = meta_rect (425, 1000, 300, 200);
temp = meta_rect (450, 1000, 300, 200);
meta_rectangle_shove_into_region (region,
fixed_directions,
&rect);
g_assert (meta_rectangle_equal (&rect, &temp));
rect = meta_rect (425, 1000, 300, 200);
temp = meta_rect (425, 950, 300, 200);
meta_rectangle_shove_into_region (region,
FIXED_DIRECTION_X,
&rect);
g_assert (meta_rectangle_equal (&rect, &temp));
rect = meta_rect ( 300, 1000, 400, 200);
temp = meta_rect (1200, 1000, 400, 200);
meta_rectangle_shove_into_region (region,
FIXED_DIRECTION_Y,
&rect);
g_assert (meta_rectangle_equal (&rect, &temp));
rect = meta_rect ( 800, 1150, 400, 50); /* Completely "offscreen" :) */
temp = meta_rect ( 800, 1050, 400, 50);
meta_rectangle_shove_into_region (region,
0,
&rect);
g_assert (meta_rectangle_equal (&rect, &temp));
rect = meta_rect (-1000, 0, 400, 150); /* Offscreen in 2 directions */
temp = meta_rect ( 0, 20, 400, 150);
meta_rectangle_shove_into_region (region,
0,
&rect);
g_assert (meta_rectangle_equal (&rect, &temp));
meta_rectangle_free_list_and_elements (region);
printf ("%s passed.\n", G_STRFUNC);
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 14:58:50 +00:00
}
static void
verify_edge_lists_are_equal (GList *code, GList *answer)
{
int which = 0;
while (code && answer)
{
MetaEdge *a = code->data;
MetaEdge *b = answer->data;
if (!meta_rectangle_equal (&a->rect, &b->rect) ||
a->side_type != b->side_type ||
a->edge_type != b->edge_type)
{
g_error ("%dth item in code answer answer lists do not match; "
"code rect: %d,%d + %d,%d; answer rect: %d,%d + %d,%d\n",
which,
a->rect.x, a->rect.y, a->rect.width, a->rect.height,
b->rect.x, b->rect.y, b->rect.width, b->rect.height);
}
code = code->next;
answer = answer->next;
which++;
}
/* Ought to be at the end of both lists; check if we aren't */
if (code)
{
MetaEdge *tmp = code->data;
g_error ("code list longer than answer list by %d items; "
"first extra item rect: %d,%d +%d,%d\n",
g_list_length (code),
tmp->rect.x, tmp->rect.y, tmp->rect.width, tmp->rect.height);
}
if (answer)
{
MetaEdge *tmp = answer->data;
g_error ("answer list longer than code list by %d items; "
"first extra item rect: %d,%d +%d,%d\n",
g_list_length (answer),
tmp->rect.x, tmp->rect.y, tmp->rect.width, tmp->rect.height);
}
}
static void
test_find_onscreen_edges ()
{
GList* edges;
GList* tmp;
int left = META_DIRECTION_LEFT;
int right = META_DIRECTION_RIGHT;
int top = META_DIRECTION_TOP;
int bottom = META_DIRECTION_BOTTOM;
2014-05-02 13:34:02 +00:00
/*************************************************/
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 14:58:50 +00:00
/* Make sure test region 0 has the correct edges */
2014-05-02 13:34:02 +00:00
/*************************************************/
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 14:58:50 +00:00
edges = get_screen_edges (0);
tmp = NULL;
tmp = g_list_prepend (tmp, new_screen_edge ( 0, 1200, 1600, 0, bottom));
tmp = g_list_prepend (tmp, new_screen_edge ( 0, 0, 1600, 0, top));
tmp = g_list_prepend (tmp, new_screen_edge (1600, 0, 0, 1200, right));
tmp = g_list_prepend (tmp, new_screen_edge ( 0, 0, 0, 1200, left));
verify_edge_lists_are_equal (edges, tmp);
meta_rectangle_free_list_and_elements (tmp);
meta_rectangle_free_list_and_elements (edges);
2014-05-02 13:34:02 +00:00
/*************************************************/
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 14:58:50 +00:00
/* Make sure test region 1 has the correct edges */
2014-05-02 13:34:02 +00:00
/*************************************************/
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 14:58:50 +00:00
edges = get_screen_edges (1);
tmp = NULL;
tmp = g_list_prepend (tmp, new_screen_edge ( 0, 1200, 400, 0, bottom));
tmp = g_list_prepend (tmp, new_screen_edge ( 400, 1160, 1200, 0, bottom));
tmp = g_list_prepend (tmp, new_screen_edge ( 0, 20, 1600, 0, top));
tmp = g_list_prepend (tmp, new_screen_edge (1600, 20, 0, 1140, right));
tmp = g_list_prepend (tmp, new_screen_edge ( 400, 1160, 0, 40, right));
tmp = g_list_prepend (tmp, new_screen_edge ( 0, 20, 0, 1180, left));
verify_edge_lists_are_equal (edges, tmp);
meta_rectangle_free_list_and_elements (tmp);
meta_rectangle_free_list_and_elements (edges);
2014-05-02 13:34:02 +00:00
/*************************************************/
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 14:58:50 +00:00
/* Make sure test region 2 has the correct edges */
2014-05-02 13:34:02 +00:00
/*************************************************/
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 14:58:50 +00:00
edges = get_screen_edges (2);
tmp = NULL;
tmp = g_list_prepend (tmp, new_screen_edge (1200, 1200, 400, 0, bottom));
tmp = g_list_prepend (tmp, new_screen_edge ( 450, 1200, 350, 0, bottom));
tmp = g_list_prepend (tmp, new_screen_edge ( 0, 1200, 300, 0, bottom));
tmp = g_list_prepend (tmp, new_screen_edge ( 300, 1150, 150, 0, bottom));
tmp = g_list_prepend (tmp, new_screen_edge ( 800, 1100, 400, 0, bottom));
tmp = g_list_prepend (tmp, new_screen_edge ( 0, 20, 1600, 0, top));
tmp = g_list_prepend (tmp, new_screen_edge (1600, 20, 0, 1180, right));
tmp = g_list_prepend (tmp, new_screen_edge ( 800, 1100, 0, 100, right));
tmp = g_list_prepend (tmp, new_screen_edge ( 300, 1150, 0, 50, right));
tmp = g_list_prepend (tmp, new_screen_edge (1200, 1100, 0, 100, left));
tmp = g_list_prepend (tmp, new_screen_edge ( 450, 1150, 0, 50, left));
tmp = g_list_prepend (tmp, new_screen_edge ( 0, 20, 0, 1180, left));
verify_edge_lists_are_equal (edges, tmp);
meta_rectangle_free_list_and_elements (tmp);
meta_rectangle_free_list_and_elements (edges);
2014-05-02 13:34:02 +00:00
/*************************************************/
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 14:58:50 +00:00
/* Make sure test region 3 has the correct edges */
2014-05-02 13:34:02 +00:00
/*************************************************/
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 14:58:50 +00:00
edges = get_screen_edges (3);
tmp = NULL;
tmp = g_list_prepend (tmp, new_screen_edge (1200, 1200, 400, 0, bottom));
tmp = g_list_prepend (tmp, new_screen_edge ( 380, 1200, 420, 0, bottom));
tmp = g_list_prepend (tmp, new_screen_edge ( 0, 1200, 300, 0, bottom));
tmp = g_list_prepend (tmp, new_screen_edge ( 300, 1150, 80, 0, bottom));
tmp = g_list_prepend (tmp, new_screen_edge ( 800, 1100, 400, 0, bottom));
tmp = g_list_prepend (tmp, new_screen_edge ( 700, 525, 200, 0, bottom));
tmp = g_list_prepend (tmp, new_screen_edge ( 700, 675, 200, 0, top));
tmp = g_list_prepend (tmp, new_screen_edge ( 0, 20, 1600, 0, top));
tmp = g_list_prepend (tmp, new_screen_edge (1600, 20, 0, 1180, right));
tmp = g_list_prepend (tmp, new_screen_edge ( 800, 1100, 0, 100, right));
tmp = g_list_prepend (tmp, new_screen_edge ( 700, 525, 0, 150, right));
tmp = g_list_prepend (tmp, new_screen_edge ( 300, 1150, 0, 50, right));
tmp = g_list_prepend (tmp, new_screen_edge (1200, 1100, 0, 100, left));
tmp = g_list_prepend (tmp, new_screen_edge ( 900, 525, 0, 150, left));
tmp = g_list_prepend (tmp, new_screen_edge ( 380, 1150, 0, 50, left));
tmp = g_list_prepend (tmp, new_screen_edge ( 0, 20, 0, 1180, left));
#if 0
#define FUDGE 50 /* number of edges */
char big_buffer1[(EDGE_LENGTH+2)*FUDGE], big_buffer2[(EDGE_LENGTH+2)*FUDGE];
meta_rectangle_edge_list_to_string (edges, "\n ", big_buffer1);
meta_rectangle_edge_list_to_string (tmp, "\n ", big_buffer2);
2014-05-02 13:34:02 +00:00
printf("Generated edge list:\n %s\nComparison edges list:\n %s\n",
big_buffer1, big_buffer2);
#endif
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 14:58:50 +00:00
verify_edge_lists_are_equal (edges, tmp);
meta_rectangle_free_list_and_elements (tmp);
meta_rectangle_free_list_and_elements (edges);
2014-05-02 13:34:02 +00:00
/*************************************************/
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 14:58:50 +00:00
/* Make sure test region 4 has the correct edges */
2014-05-02 13:34:02 +00:00
/*************************************************/
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 14:58:50 +00:00
edges = get_screen_edges (4);
tmp = NULL;
tmp = g_list_prepend (tmp, new_screen_edge ( 800, 1200, 800, 0, bottom));
tmp = g_list_prepend (tmp, new_screen_edge ( 800, 20, 800, 0, top));
tmp = g_list_prepend (tmp, new_screen_edge (1600, 20, 0, 1180, right));
tmp = g_list_prepend (tmp, new_screen_edge ( 800, 20, 0, 1180, left));
verify_edge_lists_are_equal (edges, tmp);
meta_rectangle_free_list_and_elements (tmp);
meta_rectangle_free_list_and_elements (edges);
2014-05-02 13:34:02 +00:00
/*************************************************/
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 14:58:50 +00:00
/* Make sure test region 5 has the correct edges */
2014-05-02 13:34:02 +00:00
/*************************************************/
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 14:58:50 +00:00
edges = get_screen_edges (5);
tmp = NULL;
verify_edge_lists_are_equal (edges, tmp);
meta_rectangle_free_list_and_elements (tmp);
meta_rectangle_free_list_and_elements (edges);
2014-05-02 13:34:02 +00:00
/*************************************************/
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 14:58:50 +00:00
/* Make sure test region 6 has the correct edges */
2014-05-02 13:34:02 +00:00
/*************************************************/
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 14:58:50 +00:00
edges = get_screen_edges (6);
tmp = NULL;
tmp = g_list_prepend (tmp, new_screen_edge ( 0, 1200, 1600, 0, bottom));
tmp = g_list_prepend (tmp, new_screen_edge ( 0, 40, 1600, 0, top));
tmp = g_list_prepend (tmp, new_screen_edge (1600, 40, 0, 1160, right));
tmp = g_list_prepend (tmp, new_screen_edge ( 0, 40, 0, 1160, left));
verify_edge_lists_are_equal (edges, tmp);
meta_rectangle_free_list_and_elements (tmp);
meta_rectangle_free_list_and_elements (edges);
printf ("%s passed.\n", G_STRFUNC);
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 14:58:50 +00:00
}
static void
test_find_nonintersected_monitor_edges ()
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 14:58:50 +00:00
{
GList* edges;
GList* tmp;
int left = META_DIRECTION_LEFT;
int right = META_DIRECTION_RIGHT;
int top = META_DIRECTION_TOP;
int bottom = META_DIRECTION_BOTTOM;
2014-05-02 13:34:02 +00:00
/*************************************************************************/
/* Make sure test monitor set 0 for with region 0 has the correct edges */
2014-05-02 13:34:02 +00:00
/*************************************************************************/
edges = get_monitor_edges (0, 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 14:58:50 +00:00
tmp = NULL;
verify_edge_lists_are_equal (edges, tmp);
meta_rectangle_free_list_and_elements (tmp);
meta_rectangle_free_list_and_elements (edges);
2014-05-02 13:34:02 +00:00
/*************************************************************************/
/* Make sure test monitor set 2 for with region 1 has the correct edges */
2014-05-02 13:34:02 +00:00
/*************************************************************************/
edges = get_monitor_edges (2, 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 14:58:50 +00:00
tmp = NULL;
tmp = g_list_prepend (tmp, new_monitor_edge ( 0, 600, 1600, 0, bottom));
tmp = g_list_prepend (tmp, new_monitor_edge ( 0, 600, 1600, 0, top));
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 14:58:50 +00:00
verify_edge_lists_are_equal (edges, tmp);
meta_rectangle_free_list_and_elements (tmp);
meta_rectangle_free_list_and_elements (edges);
2014-05-02 13:34:02 +00:00
/*************************************************************************/
/* Make sure test monitor set 1 for with region 2 has the correct edges */
2014-05-02 13:34:02 +00:00
/*************************************************************************/
edges = get_monitor_edges (1, 2);
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 14:58:50 +00:00
tmp = NULL;
tmp = g_list_prepend (tmp, new_monitor_edge ( 800, 20, 0, 1080, right));
tmp = g_list_prepend (tmp, new_monitor_edge ( 800, 20, 0, 1180, left));
#if 0
#define FUDGE 50
char big_buffer1[(EDGE_LENGTH+2)*FUDGE], big_buffer2[(EDGE_LENGTH+2)*FUDGE];
meta_rectangle_edge_list_to_string (edges, "\n ", big_buffer1);
meta_rectangle_edge_list_to_string (tmp, "\n ", big_buffer2);
2014-05-02 13:34:02 +00:00
printf("Generated edge list:\n %s\nComparison edges list:\n %s\n",
big_buffer1, big_buffer2);
#endif
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 14:58:50 +00:00
verify_edge_lists_are_equal (edges, tmp);
meta_rectangle_free_list_and_elements (tmp);
meta_rectangle_free_list_and_elements (edges);
2014-05-02 13:34:02 +00:00
/*************************************************************************/
/* Make sure test monitor set 3 for with region 3 has the correct edges */
2014-05-02 13:34:02 +00:00
/*************************************************************************/
edges = get_monitor_edges (3, 3);
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 14:58:50 +00:00
tmp = NULL;
tmp = g_list_prepend (tmp, new_monitor_edge ( 900, 600, 700, 0, bottom));
tmp = g_list_prepend (tmp, new_monitor_edge ( 0, 600, 700, 0, bottom));
tmp = g_list_prepend (tmp, new_monitor_edge ( 900, 600, 700, 0, top));
tmp = g_list_prepend (tmp, new_monitor_edge ( 0, 600, 700, 0, top));
tmp = g_list_prepend (tmp, new_monitor_edge ( 800, 675, 0, 425, right));
tmp = g_list_prepend (tmp, new_monitor_edge ( 800, 675, 0, 525, left));
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 14:58:50 +00:00
verify_edge_lists_are_equal (edges, tmp);
meta_rectangle_free_list_and_elements (tmp);
meta_rectangle_free_list_and_elements (edges);
2014-05-02 13:34:02 +00:00
/*************************************************************************/
/* Make sure test monitor set 3 for with region 4 has the correct edges */
2014-05-02 13:34:02 +00:00
/*************************************************************************/
edges = get_monitor_edges (3, 4);
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 14:58:50 +00:00
tmp = NULL;
tmp = g_list_prepend (tmp, new_monitor_edge ( 800, 600, 800, 0, bottom));
tmp = g_list_prepend (tmp, new_monitor_edge ( 800, 600, 800, 0, top));
tmp = g_list_prepend (tmp, new_monitor_edge ( 800, 600, 0, 600, right));
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 14:58:50 +00:00
verify_edge_lists_are_equal (edges, tmp);
meta_rectangle_free_list_and_elements (tmp);
meta_rectangle_free_list_and_elements (edges);
2014-05-02 13:34:02 +00:00
/*************************************************************************/
/* Make sure test monitor set 3 for with region 5has the correct edges */
2014-05-02 13:34:02 +00:00
/*************************************************************************/
edges = get_monitor_edges (3, 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 14:58:50 +00:00
tmp = NULL;
verify_edge_lists_are_equal (edges, tmp);
meta_rectangle_free_list_and_elements (tmp);
meta_rectangle_free_list_and_elements (edges);
printf ("%s passed.\n", G_STRFUNC);
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 14:58:50 +00:00
}
static void
test_gravity_resize ()
{
MetaRectangle oldrect, rect, temp;
rect.x = -500; /* Some random amount not equal to oldrect.x to ensure that
* the resize is done with respect to oldrect instead of rect
*/
oldrect = meta_rect ( 50, 300, 250, 400);
temp = meta_rect ( 50, 300, 20, 5);
meta_rectangle_resize_with_gravity (&oldrect,
&rect,
NorthWestGravity,
20,
5);
g_assert (meta_rectangle_equal (&rect, &temp));
rect = meta_rect ( 50, 300, 250, 400);
temp = meta_rect (165, 300, 20, 5);
meta_rectangle_resize_with_gravity (&rect,
&rect,
NorthGravity,
20,
5);
g_assert (meta_rectangle_equal (&rect, &temp));
rect = meta_rect ( 50, 300, 250, 400);
temp = meta_rect (280, 300, 20, 5);
meta_rectangle_resize_with_gravity (&rect,
&rect,
NorthEastGravity,
20,
5);
g_assert (meta_rectangle_equal (&rect, &temp));
rect = meta_rect ( 50, 300, 250, 400);
temp = meta_rect ( 50, 695, 50, 5);
meta_rectangle_resize_with_gravity (&rect,
&rect,
SouthWestGravity,
50,
5);
g_assert (meta_rectangle_equal (&rect, &temp));
rect = meta_rect ( 50, 300, 250, 400);
temp = meta_rect (150, 695, 50, 5);
meta_rectangle_resize_with_gravity (&rect,
&rect,
SouthGravity,
50,
5);
g_assert (meta_rectangle_equal (&rect, &temp));
rect = meta_rect ( 50, 300, 250, 400);
temp = meta_rect (250, 695, 50, 5);
meta_rectangle_resize_with_gravity (&rect,
&rect,
SouthEastGravity,
50,
5);
g_assert (meta_rectangle_equal (&rect, &temp));
rect = meta_rect (167, 738, 237, 843);
temp = meta_rect (167, 1113, 832, 93);
meta_rectangle_resize_with_gravity (&rect,
&rect,
WestGravity,
832,
93);
g_assert (meta_rectangle_equal (&rect, &temp));
rect = meta_rect ( 167, 738, 237, 843);
temp = meta_rect (-131, 1113, 833, 93);
meta_rectangle_resize_with_gravity (&rect,
&rect,
CenterGravity,
832,
93);
g_assert (meta_rectangle_equal (&rect, &temp));
rect = meta_rect (300, 1000, 400, 200);
temp = meta_rect (270, 994, 430, 212);
meta_rectangle_resize_with_gravity (&rect,
&rect,
EastGravity,
430,
211);
g_assert (meta_rectangle_equal (&rect, &temp));
rect = meta_rect (300, 1000, 400, 200);
temp = meta_rect (300, 1000, 430, 211);
meta_rectangle_resize_with_gravity (&rect,
&rect,
StaticGravity,
430,
211);
g_assert (meta_rectangle_equal (&rect, &temp));
printf ("%s passed.\n", G_STRFUNC);
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 14:58:50 +00:00
}
static void
test_find_closest_point_to_line ()
{
double x1, y1, x2, y2, px, py, rx, ry;
double answer_x, answer_y;
x1 = 3.0; y1 = 49.0;
x2 = 2.0; y2 = - 1.0;
px = -2.6; py = 19.1;
answer_x = 2.4; answer_y = 19;
meta_rectangle_find_linepoint_closest_to_point (x1, y1,
x2, y2,
px, py,
&rx, &ry);
g_assert (rx == answer_x && ry == answer_y);
/* Special test for x1 == x2, so that slop of line is infinite */
x1 = 3.0; y1 = 49.0;
x2 = 3.0; y2 = - 1.0;
px = -2.6; py = 19.1;
answer_x = 3.0; answer_y = 19.1;
meta_rectangle_find_linepoint_closest_to_point (x1, y1,
x2, y2,
px, py,
&rx, &ry);
g_assert (rx == answer_x && ry == answer_y);
/* Special test for y1 == y2, so perp line has slope of infinity */
x1 = 3.14; y1 = 7.0;
x2 = 2.718; y2 = 7.0;
px = -2.6; py = 19.1;
answer_x = -2.6; answer_y = 7;
meta_rectangle_find_linepoint_closest_to_point (x1, y1,
x2, y2,
px, py,
&rx, &ry);
g_assert (rx == answer_x && ry == answer_y);
/* Test when we the point we want to be closest to is actually on the line */
x1 = 3.0; y1 = 49.0;
x2 = 2.0; y2 = - 1.0;
px = 2.4; py = 19.0;
answer_x = 2.4; answer_y = 19;
meta_rectangle_find_linepoint_closest_to_point (x1, y1,
x2, y2,
px, py,
&rx, &ry);
g_assert (rx == answer_x && ry == answer_y);
printf ("%s passed.\n", G_STRFUNC);
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 14:58:50 +00:00
}
int
main()
{
init_random_ness ();
test_area ();
test_intersect ();
test_equal ();
test_overlap_funcs ();
test_basic_fitting ();
test_regions_okay ();
test_region_fitting ();
test_clamping_to_region ();
test_clipping_to_region ();
test_shoving_into_region ();
/* And now the functions dealing with edges more than boxes */
test_find_onscreen_edges ();
test_find_nonintersected_monitor_edges ();
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 14:58:50 +00:00
/* And now the misfit functions that don't quite fit in anywhere else... */
test_gravity_resize ();
test_find_closest_point_to_line ();
printf ("All tests passed.\n");
return 0;
}