| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
| |
The user may change the client’s signals priority setting while it is stopped, which would not restart it.
We may as well also always change the actual priority.
|
|
|
|
| |
This is a per-client setting: activating it in the menu will have the process being SIGSTOPped/SIGCONTed instead of having its priority changed on focus change.
|
|
|
|
|
|
| |
In the very last commit I got the visibility check wrong.
ref 0680250d720649a95fd3d9631c7baf6829a1a665
|
|
|
|
|
|
|
|
| |
We kept an extra reference for wayland clients and dropped it on their final
render to stop crashes. However, if they're not on screen at the time of
their deletion this reference still needed to be dropped.
Fixes clients remaining on the deskmirror after dying on another desktop.
|
|
|
|
|
|
|
|
|
| |
so getting top object was broken. it didnt account for repeat event
objects that would be included. so get the full l,ist and walk them
top to bottom for the first one thats a client. THAT is the correct
thing to do. this would affect both x11 and wayland.
@fix
|
|
|
|
|
|
|
| |
it seems on openbsd sigterm doesnt kill things... sigkill does so move
to that. fixes T4121
@fix
|
|
|
|
| |
the coords passed in this method are relative to the window geometry
|
|
|
|
| |
Signed-off-by: Chris Michael <cp.michael@samsung.com>
|
|
|
|
|
|
|
|
| |
in the case where there is no smart parent, this fixes the object tree
and resolves some stacking issues. if there is a parent or a parent is added
later, nothing is changed
fixes eg. screenshot preview visibility in wayland
|
|
|
|
|
|
| |
hard drive.
@fix T4119
|
|
|
|
|
|
|
| |
i found this didn't emit signals correctly and multiple times when not
needed. this fixes that.
@fix
|
|
|
|
|
|
|
|
|
| |
Quite a hack overall, but if you don't need those parameter, just put
nothing and it will work.
@fix T4112
Signed-off-by: Cedric Bail <cedric@osg.samsung.com>
|
|
|
|
|
|
| |
Moves and resizes tripped up the new xdg focus logic, so test if the
focus is leaving the client entirely before trying to determine if it's
leaving into a parent surface.
|
|
|
|
|
| |
This stops recent xdg_popup focus tracking changes from causing multiple
key events to be sent to the focused client.
|
|
|
|
|
| |
Otherwise the pointer is removed when the next drag starts, so this
makes debugging dnd problems easier.
|
| |
|
|
|
|
| |
ref b5576dbb8c98934ab0e8c5ff59bed106c02d9874
|
|
|
|
| |
global sizing policy++
|
|
|
|
| |
fixes usage in non-composited x11 environments (eg. openbox)
|
|
|
|
| |
fixes autodetection for x11 wayland output without setting E_WL_FORCE
|
|
|
|
| |
Signed-off-by: Chris Michael <cp.michael@samsung.com>
|
|
|
|
|
|
| |
Oops, a client can have multiple top level windows.
ref f391a0fb6724d9efe9aa0c89fe47e2372af8d35d
|
| |
|
|
|
|
|
|
| |
These were inadvertently removed by e1369a24bd7dc464df0b1eb568adf59e9cc5f1c2.
@fix
|
|
|
|
|
|
|
|
| |
Reviewers: zmike!
Subscribers: cedric
Differential Revision: https://phab.enlightenment.org/D4129
|
| |
|
|
|
|
|
| |
For now the setting of a new layout does not mean the complete file
needs to be recompiled.
|
|
|
|
|
|
|
|
|
|
|
| |
this commit introduces the setting of the index. Setting the index here
means that the layout with the id 0..n, out of the compiled keymap file
will be used. After a new index is set the modifiers are updated, that
the client are aware of the new resulting group.
If the api is called before the compositor is inited (this can happen
in e_xkb, so the drm can use the keymap at startup) then the index is
saved in between and will be flushed once the compositor does the init.
|
|
|
|
| |
feature freeze is over, have fun
|
| |
|
| |
|
| |
|
|
|
|
|
| |
On xdg_shell we should only ever send enter/leave to top level surfaces
or GTK becomes sad.
|
|
|
|
|
|
|
| |
There are actually toolkits that create surfaces, do nothing with them,
and destroy them. Sending keyboard leave events for this causes problems.
Fixes a bug in handling of some GTK popups.
|
| |
|
|
|
|
| |
This fixes Qt sub-windows.
|
|
|
|
|
| |
Otherwise, if a wayland client does its first commit without a buffer
attached it will confuse our focus logic.
|
|
|
|
| |
fixes dismissal during comp input regrabs
|
|
|
|
|
|
|
|
|
|
| |
Mirrors can be rendered independently of what they're mirroring,
which (at least under wayland) can result in a situation where the
mirror is rendered before the parent sets up their image pointers
properly.
We give mirrors their own callback to prevent that from causing a
crash.
|
|
|
|
|
|
|
| |
This reverts commit 326d638ba274ba4623c4112a68250b47f7b250e1.
This, for some reason, resulted in client surfaces being blacked out,
despite only ever changing mirror surfaces.
|
|
|
|
|
|
|
|
| |
Wayland pixmap ids are a different data type for internal and
external windows. cast them both to 64-bits so they're the same
size regardless of arch.
ref d3ba524a62713ff58c1537524030c155d0713ca2
|
|
|
|
| |
this fixes none changing sinks in the app.
|
|
|
|
| |
as said on the ML
|
|
|
|
| |
after talking to zmike, this is by intend.
|
|
|
|
|
|
|
| |
otherwise we are going to set 0 as default group.
This commit refactors out _eval_cur_group from the init timer, since wl
does not need a init timer.
|
| |
|
|
|
|
|
|
| |
Both the X and WL code try to set up the output pointer for the
zone. Make the WL output take precedence, since it's actually
used for events.
|
|
|
|
|
|
|
|
| |
in a choice between fixing a corner case popup behavior and breaking dnd
or having functional dnd and adding hacks to fix corner case popup behavior,
adding more hacks was the obvious correct solution
ref 03a4ecbdb0972f8267c07ea14d752fcee69bd9fa
|
| |
|
|
|
|
| |
Avoid that NULL pointer deref.
|