Age | Commit message (Collapse) | Author | Files | Lines |
|
After resizing a tiled window, tile_resized gets set to true. Since it
never got set back to false when fullscreening, it lead to weird
behavior when unfullscreening the window.
|
|
|
|
|
|
fixes https://github.com/mate-desktop/marco/issues/367
|
|
|
|
|
|
|
|
|
|
|
|
This commit reverts:
https://github.com/mate-desktop/marco/commit/f0541e3dfda29c26fe14c9c9117f95c49006c75c
https://github.com/mate-desktop/marco/commit/d18c2fb4acb58c408c01700682b9922de86e8b2f
And it applies an alternative to fix the deprecated functions:
gdk_screen_get_width
gdk_screen_width
gdk_screen_height
|
|
|
|
META_COMP_WINDOW_DESKTOP is available (#329)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Use that flag to retain the correct window size when performing other
operations. Reset when re-tiling.
|
|
Determine whether the user is initiating a resize action on a tiled
window. If the user is trying to grab the window for resizing
horizontally from the edge that's farther away from the screen edge,
allow the resize to occur. Otherwise maintain the current tile geometry.
Also modified the window hints to allow resizing from the window menu.
Fixes #250
|
|
avoid deprecated:
gdk_screen_get_monitor_geometry
gdk_screen_get_monitor_at_point
|
|
|
|
|
|
use GtkSeat for > gtk+-3.20
|
|
Those were never used in MATE and have a lot of deprecated functions.
Already droped in metacity.
https://git.gnome.org/browse/metacity/commit/?id=b580e3e
|
|
* Determine focused window when processing synthetic events
When a client is passively grabbing keybindings that it does not need,
it sends them up for other clients to process.
Often in this situation, the event contains the wrong window (either
root, for global keybindings, or the original client itself). This means
that Marco will attempt to process the event for the wrong window.
This is not an issue for global keybindings within Marco, as the focused
window does not matter. However, for shortcuts that operate directly on
specific windows, the event gets lost.
This change addresses this by determining what the currently-focused
window is, regardless of which client forwarded the event.
* Determine window under pointer when processing synthetic events
When a client is passively grabbing mouse clicks that it does not need,
it sends them up for other clients to process.
Often in this situation, the event contains the wrong window (either
root, for global keybindings, or the original client itself). This means
that Marco will attempt to process the event for the wrong window.
This change addresses this by determining what the current window under
the mouse pointer is, regardless of which client forwarded the event.
* Remove unused development data
|
|
default or user configured scale. Closes #335 (#336)
|
|
|
|
|
|
|
|
|
|
|
|
This legacy construct was carried over from the gtk2/gtk3 port, and
attempts to initialise MetaFrames with a GtkWindow `type` property.
This property doesn't actually exist, and will cause marco to completely
crash when running with `MARCO_DEBUG=1` for local build testing.
Signed-off-by: Ikey Doherty <[email protected]>
|
|
thanks to James Ward for spotting it
|
|
|
|
|
|
|
|
for consistency with all the other bindings
|
|
|
|
|
|
|
|
|
|
This reverts commit 768fdd8d3852e67555a585da28b6404a97853cbd.
fixes probably "strange window switching foreground/background behavior"
https://github.com/mate-desktop/marco/issues/251
|
|
taken from:
https://git.gnome.org/browse/metacity/commit/?id=7ec9a6f
|
|
_SVID_SOURCE has been deprecated in newer versions of glibc breaking -WError;
the recommended replacement of _DEFAULT_SOURCE is fairly new, so switch to
_XOPEN_SOURCE instead.
Based on mutter commits b159d6a5e17219557574defb6c3788d9b0c39757
and b159d6a5e17219557574defb6c3788d9b0c39757.
https://bugzilla.gnome.org/show_bug.cgi?id=763782
taken from:
https://git.gnome.org/browse/metacity/commit/?id=56bcf350fb599ec2caf68af4f0f92b1c3999027c
|
|
Taken from:
https://git.gnome.org/browse/metacity/commit/?id=3380502
|
|
|