move the check down into the function itself, 39df1f4 began syncing on
session change, but isnt checking for the setting itself, so instead of
duplicating checks all over. do it here.
add INCR protocol support for large transfers
fix write handling for partial transfers
fix an issue where wayland windows could die from a paste from an
xwayland window
for xdg-shell, we can ping the wm_base, and thus render an ANR dialog if an app dies
for XWayland, there probably is a similar method, but I don't know about it and don't care.
* xwayland: configure the window on a configure request
* xwayland: move coordinate conversion handling to their own functions
* xwayland: rename configure to configureRequest
timeline check only returns nullopt on ETIME_ERR , meaning the if check
later on returns true if drmSyncobjTimelineWait returns anything else
like EINVAL/EPERM/EACCESS etc, so actually check the returned .value()
of the std::optional. also move the fd to rvalue references.
The explicit settings ignore the aquamarine output.supportsExplicit
attribute, which creates glitches on drivers not supporting explicit
sync (example: freedreno).
If the output has been set as not supporting explicit, disable the
explicit settings.
* fixes#9322, take `active_area_size` into account when sending tip event
* check if `relative_input` is set
As suggested by @y47s5s68tq870r7tc1xpp755pabopg
* refactoring active area in own function to keep it DRY
* coding style
* making transformation static
---------
Co-authored-by: clamydo <clamydo@users.noreply.github.com>
ensure framebuffer textures are detached and deleted, avoid leaving framebuffers bound when not needed
* render: avoid calling glDeleteProgram on no program
its safe to do so but it adds a bunch of unnecessery lines in apitrace
when tracing. if guard it and return early.
* opengl: ensure texture and buffers are unbound
ensure bound buffers are unbound after use, also detach textures from
framebuffer before deleting it otherwise it will become dangling and
essentially leak.
* window: fix resizes with an update callback
* window: fixup sendWindowSize
Remove the size argument from sendWindowSize, since it is now a member of the Window class
and we don't want any mismatches between m_vRealSize and what we report.
Remove sendWindowSize from mapWindow, since we shouldn't need it.
* window: sendWindowSize on animation begin
* window: move most calls to sendWindowSize to the animation begin
callback
* window: remove sendWindowSize in unmanaged if not fullscreen
