2003-06-09 Rob Adams <robadams@ucla.edu>
Revamp placement policy for windows that are maximized when they
are mapped, including windows that set a hint to be maximized or
windows that are auto-maximized using our heuristic. See #111902.
* src/window.h: add new flag maximize_after_placement and new
function meta_window_maximize_internal.
* src/window.c (meta_window_new): initialize
maximize_after_placement to FALSE and remove the automaximize
heuristic.
(meta_window_maximize_internal): new function accepts a saved_rect
argument to be used as the new saved_rect for the window, and does
not queue a move_resize.
(meta_window_maximize): re-implement using
meta_window_maximize_internal.
(update_net_wm_state): If a window has a maximize hint set on
startup set maximize_after_placement to TRUE
* src/constraints.c (meta_window_constrain): Update the xinerama
information in the ConstraintInfo after placing the window, and
maximize the window after placement if
window->maximize_after_placement
* src/place.c (find_first_fit): take a natural xinerama list as an
argument instead of generating it here
(constrain_placement): remove function, since it is no longer
needed
(meta_window_place): generate the natural xinerama list here and
pass it into find_first_fit. If find_first_fit fails, use the
list to find empty xineramas where we can place windows that may
be maximized later. This makes maximized windows follow the
correct placement policy. Move the automaximize heuristic here.
2003-04-05 Rob Adams <robadams@ucla.edu>
Update placement policy for screen with multiple xineramas.
Windows will be placed preferentially on the xinerama with the
pointer, and progressively further away as needed to find a place
where the window does not overlap other windows.
* src/place.c (rect_fits_in_work_area): function
fit_rect_in_xinerama greatly simplified to work with new placement
policy.
(find_first_fit): implement new first first placement scheme
* src/screen.c (meta_screen_get_xinerama_neighbor): look for an
xinerama in the xinerama list that is adjacent to the specified
xinerama.
(meta_screen_get_natural_xinerama_list): return a list of
xineramas in the order to be preferred by the placement algorithm
as determined by the current location of the pointer.
* src/screen.h: add function prototypes and an enum used by
meta_screen_get_xinerama_neighbor.
2003-02-23 Havoc Pennington <hp@pobox.com>
Patch from Rob Adams addresses #95014 (placement issues),
makes first fit algorithm "center tile", adds most code
for per-xinerama workspaces (#86682) but disables it for now.
* src/workspace.c (meta_workspace_get_work_area_for_xinerama)
(meta_workspace_get_work_area_all_xineramas): new xinerama
functions, maintain workspace->work_areas with a different
work area for each xinerama. However for now all the work
areas are the same, because haven't quite figured out how
_NET_WM_STRUT is supposed to work
* src/window.c: adapt to new meta_window_* xinerama APIs
(meta_window_get_work_area_current_xinerama): new xinerama
API
(meta_window_get_work_area_for_xinerama): new xinerama API
(constrain_position): be a bit more clever about which xinerama's
work area we choose to use.
* src/stack.c: adapt to new Xinerama API
* src/screen.c (reload_xinerama_infos): invalidate all work areas
(meta_screen_get_xinerama_for_rect): new function
(meta_screen_window_intersects_xinerama): new function
* src/place.c (find_first_fit): change to use
"center tiling" (center a screen full of tiled windows,
rather than aligning them top left). Adapt to new
xinerama functions.
2002-10-01 Havoc Pennington <hp@pobox.com>
* src/place.c (find_next_cascade): try extra cascades alongside
the first, if the first fails; patch from readams@hmc.edu
2002-09-27 Havoc Pennington <hp@pobox.com>
* src/place.c (constrain_placement): constrain placement to try to
keep windows from going offscreen to the right/bottom
* src/stack.c (compute_layer): rearrange the logic here to say
that a window must always be in at least as high a layer as any of
its transient parents or group members, rather than special-casing
fullscreen. Also, group_member_is_fullscreen was leaking the list
of group members every time, a fairly major memory leak.
2002-09-24 Havoc Pennington <hp@pobox.com>
* src/place.c (fit_rect_in_xinerama): update best_overlap as we go
through the loop - doh. Fix from readams@hmc.edu for #90799.
(find_first_fit): try the origin of each xinerama screen
after the first. Also from readams@hmc.edu
2002-09-12 Havoc Pennington <hp@redhat.com>
* src/workspace.c (meta_workspace_screen_index)
(meta_workspace_index): fix compiler warnings
* src/tools/metacity-window-demo.c (menu_items): add a test for
dialogs with no transient parent
* src/place.c (find_first_fit): Try placing window at origin of
first Xinerama, even if there are no windows to place next to;
makes placement work when no other windows are open on the screen.
2002-07-12 Havoc Pennington <hp@redhat.com>
* src/window.c (meta_window_new): try to maximize windows that
are too big for the work area
* src/place.c (find_next_cascade): don't let the cascade algorithm
place windows off the screen, and fix it to always exhaustively
search the window list for cascade parents.
2002-06-17 Havoc Pennington <hp@pobox.com>
* src/place.c (find_next_cascade): increase the cascade threshold
a bit.
(find_first_fit): implement a somewhat lame first fit algorithm
2002-06-09 Havoc Pennington <hp@pobox.com>
* src/place.c (meta_window_place): don't run constrain_placement
on windows we allow to go anywhere (docks, etc.). Fixes
positioning of panel windows in certain cases.
2002-06-08 Havoc Pennington <hp@pobox.com>
Yeah I know maximization is broken, I'm too tired to fix it.
Probably because of the change to update_struts() that was
supposed to fix the 100% CPU bug.
* src/place.c (meta_window_place): don't run docks and things
through the placement algorithm. Thought it might fix
metacity-window-demo but it didn't.
* src/window.c (constrain_size): only get work area when needed
(meta_window_new): init the do_not_cover field
2002-06-08 Havoc Pennington <hp@pobox.com>
* src/screen.c (meta_screen_get_xinerama_for_window):
short-circuit the "only one xinerama" case, and use outer rect of
window instead of window->rect, so we get root window coords.
* src/theme.c (meta_frame_layout_get_borders): if fullscreen all
frame edges are zero-width.
* src/frame.c (meta_frame_get_flags): init fullscreen flag.
* src/common.h (enum): add META_FRAME_FULLSCREEN frame flag
* src/place.c: fix up calls to meta_window_get_work_area
* src/window.c (meta_window_get_work_area): add an arg for whether
the work area is for the screen or the xinerama subscreen.
(constrain_position): fix up calls to meta_window_get_work_area
(constrain_size): ditto
* src/screen.c (meta_screen_new): add METACITY_DEBUG_XINERAMA
environment variable which simulates xinerama on a single head.
2002-06-06 Havoc Pennington <hp@pobox.com>
* src/screen.c (meta_screen_get_current_xinerama): implement
* src/place.c (meta_window_place): cascade windows on the active
Xinerama screen
* src/window.c (meta_window_move_resize_internal): strip out the
#if 0 cruft about guessing fullscreen mode
(constrain_position, constrain_size): fullscreen/maximize to the
Xinerama head, not the whole screen
(meta_window_get_work_area): autocreate struts at the Xinerama
physical screen edges for the screen the window is on.
* src/screen.c (meta_screen_get_xinerama_for_window): someone
snuck in a for loop, fix it. ;-)
2002-05-25 Anders Carlsson <andersca@gnu.org>
* src/place.c: (get_vertical_edges), (get_horizontal_edges):
Take Xinerama screen edges into consideration.
* src/screen.c: (meta_rectangle_intersect),
(meta_screen_get_xinerama_for_window):
* src/screen.h:
Add a new function that returns the xinerama monitor that
a window is on.
2002-05-12 Havoc Pennington <hp@pobox.com>
* src/place.c (meta_window_place): move pposition/usposition
honoring code into here, instead of putting it in window.c.
Makes focusing new windows work, and cleans things up a bit.
#81585
2002-05-05 Havoc Pennington <hp@pobox.com>
* src/window.c (meta_window_unminimize): on unminimize, queue
calc_showing on all transients
(meta_window_activate): on activate, unminimize all a window's
ancestors, not just the window itself.
* src/workspace.c (set_work_area_hint): don't increment "tmp" by
16 unsigned long, increment by 4
* src/window.c (meta_window_free): if a window isn't minimized,
restore its WM_STATE to NormalState instead of IconicState,
since IconicState on initial window map means that the window
should be minimized.
* src/workspace.c (meta_workspace_invalidate_work_area): queue an
idle to recompute the work area hint.
(set_work_area_hint): we need 4*num_workspaces ints, not just
num_workspaces.
* src/screen.c (meta_screen_new): add work_area_idle field,
handle it on screen shutdown
* src/common.h (META_PRIORITY_PREFS_NOTIFY,
META_PRIORITY_WORK_AREA_HINT): define some idle priorities
* src/window.c (meta_window_calc_showing): hide windows if
their parent window is minimized
(meta_window_minimize): also queue_calc_showing on all
transients of the window being minimized
* src/place.c (constrain_placement): function to apply
placement-time-only constraints, such as "not off the left of the
screen"
(meta_window_place): put dialogs down a bit over their parent,
not right at the top.
(meta_window_place): when centering a dialog, center it
on the current xinerama screen, rather than the entire
screen.
* src/screen.c (meta_screen_get_current_xinerama): new function,
but not implemented
2002-04-30 Havoc Pennington <hp@pobox.com>
* src/window.c (recalc_window_features): don't try to decorate
toolbars.
* src/tools/metacity-window-demo.c: add menu and toolbar tests
* src/place.c (meta_window_place): only dialogs should be centered
over parent, not anything with transient for set.
* src/window.c (meta_window_configure_request): become more
fascist about window positioning if workarounds are disabled, and
less fascist if they are enabled.
* src/metacity.schemas: add a "disable_workarounds" option. Kind
of crack-smoking. But we just can't get all applications
fixed. And I need no-workarounds mode to monitor which apps are
broken and what needs fixing in specs.
* src/window.c (meta_window_configure_request): always allow
windows to resize themselves
* src/keybindings.c (reload_modmap): don't filter out Mode_switch,
apparently some people bind window manager shortcuts to that.
2002-02-09 Havoc Pennington <hp@pobox.com>
* src/themes/Atlanta/metacity-theme-1.xml: put in some kind of
distinctive frame for UTILITY, though it's ugly. Also put in the
borderless look for maximized windows.
* src/stack.c (compute_layer): put splash screen in the splash
layer
* src/stack.h (enum): create a splash screen layer
* src/place.c (meta_window_place): center splashscreen, and fix a
typo in the centering code
* src/window.c (recalc_window_features): disable most features on
splash screens
* src/screen.c (set_supported_hint): add UTILITY and SPLASHSCREEN
hints
* src/window.c: add UTILITY, SPLASHSCREEN implementation
* src/window.h (enum): add UTILITY, SPLASHSCREEN types
* src/theme-parser.c (parse_toplevel_element): parser support
for has_title attribute
* src/theme.c (meta_frame_layout_get_borders): handle a has_title
field in the layout, for utility windows that don't display a
title (would be better to be able to shrink the title text,
but that's kind of tricky to implement :-/)
2002-02-07 Havoc Pennington <hp@pobox.com>
Throughout: move to meta_topic rather than meta_verbose so
metacity.log can start being more useful
* src/util.h (enum): add more debug topics
* src/frames.c: clean up some cruft that caused warnings
2002-01-09 Havoc Pennington <hp@pobox.com>
* src/window.c, src/window.h: store strut information,
update it on property changes, etc. etc. so we avoid panel
on maximize.
* src/workspace.c (meta_workspace_get_work_area): add accessor for
work area so we can compute it lazily
* src/display.h, src/display.c: add _NET_WM_STRUT atom
and _WIN_HINTS atom
2001-10-29 Havoc Pennington <hp@pobox.com>
* src/window.c (idle_calc_showing): handle queue/unqueue of
calc showings as we are iterating over the pending list
(meta_window_show): focus placed transients in here instead
of in meta_window_place - now it should actually work, yay
* src/place.c (meta_window_place): remove focusing of transient
child from here; this was really broken
2001-10-25 Havoc Pennington <hp@pobox.com>
* src/window.c (meta_window_visible_on_workspace):
I was using meta_workspace_contains_window() in a number of
places where on_all_workspaces should also have been considered,
thus this new function. Fixes bugs such as pinned windows
not appearing in the tab order.
(meta_window_client_message): use meta_window_visible_on_workspace
* src/stack.c (find_tab_forward): ditto
(find_tab_backward): ditto
(meta_stack_get_tab_next): ditto
(meta_stack_get_tab_list): ditto
* src/place.c (get_windows_on_same_workspace): ditto
* src/keybindings.c (handle_focus_previous): ditto
(handle_focus_previous): ditto
2001-09-08 Havoc Pennington <hp@pobox.com>
* src/ui.c (meta_ui_init): don't use gdk_display_name
* src/frame.c (meta_window_ensure_frame): create frame
with screen default visual, rather than client window visual;
for DRI games, the client window visual was not allowed to be
a child of another window with the same visual, apparently.
Anyhow now we copy twm, etc. so it must be correct.
* src/place.c (meta_window_place): if a transient is placed and
its parent has focus, focus the transient.
