Xwayland starts listening to -listenfd file descriptors after the WM_S0
selection is claimed. At the moment, it is claimed asynchronously by
kwin. First, we create a dummy window and change one of its properties
to get the timestamp. After the timestamp has been received, we actually
call xcb_set_selection_owner().
This provides kwin greater control over how X11 sockets are created for
Xwayland. For example, it can be used to ensure that the DISPLAY remains
the same across Xwayland server restarts or launching Xwayland on
demand.
Even though -listen <fd> option is deprecated, we still pass it because
older versions of Xwayland may not have the -listenfd option.
This renames updateXauthorityFile to writeXauthorityEntries as it doesn't
actually update (i.e. change) anything, it just writes new ones.
Error handling is introduced, to avoid that it continues silently without
entries, which would cause all connections to fail.
Once in a while, we receive complaints from other fellow KDE developers
about the file organization of kwin. This change addresses some of those
complaints by moving all of source code in a separate directory, src/,
thus making the project structure more traditional. Things such as tests
are kept in their own toplevel directories.
This change may wreak havoc on merge requests that add new files to kwin,
but if a patch modifies an already existing file, git should be smart
enough to figure out that the file has been relocated.
We may potentially split the src/ directory further to make navigating
the source code easier, but hopefully this is good enough already.
Occasionally, I see complaints about the file organization of kwin,
which is fair enough.
This change makes the source code more relocatable by removing relative
paths from includes.
CMAKE_CURRENT_SOURCE_DIR was added to the interface include directories
of kwin library. This means that as long as you link against kwin target,
the real location of the source code of the library doesn't matter.
With autotests, things are not as convenient as with kwin target. Some
tests use cpp files from kwin core. If we move all source code in a src/
directory, they will need to be adjusted, but mostly only in build
scripts.
Seat correctly emits dragEnded when a DataSourceInterface is destroyed,
and stops the drag.
However WlToXSource lives slightly longer to finish handling any other
visits. See the guarded delete statement in Dnd::endDrag.
With recent fixes use is mostly a hypothetical, but we still don't want
dangling pointers in our code.
XWL only clears up when a "visit" finishes.
If we end a WlToXDrag through clicking we call "WlToXDrag::end" but the
visit object stays alive, which obviously will never finish.
Both the visit and the drag object then effectively leak and are used on
the next drag.
Transfer emits "finished" mid way through it's own methods. If we delete
here transfer is left in the middle of methods like Transfer::timeout
with a dangling "this".
This allows running Xwayland apps as root. Xwayland started with an
empty Xauthority file. After kwin has received the display number, the
file is updated with an actual authority entry.
BUG: 432625
The selection owner must live as long as the X11 connection is valid;
otherwise Xwayland won't catch the corresponding selection ownership
change event, and the initialization sequence will be incomplete.
The major difference between this version and the previous is that kwin
no longer forwards the opacity to the app window if it runs as a
window window manager and a compositing manager. As Keith Packard said [1]
> The window manager support is only needed to forward the property from the
> application window to the frame; with the Composite extension, a
> compositing manager can then take that value into account when
> constructing the desktop image. Any X server can implement the Composite
> extension; the one in the X server project at freedesktop.org is just one
> such. It would be trivial to implement this extension in a direct FB
> based X server, or any other X server which places windows in off-screen
> images.
There are a couple of reasons to do so: (a) simplifies code, (b) we don't
temper with opacities on override-redirect windows.
[1] https://listman.redhat.com/archives/xdg-list/2003-December/msg00153.html
A recent patch made it possible to restart KWin when it crashes, this
opens the possibility to do it for development purposes. So far it could
be done using `killall -9 kwin_wayland`. This patch embraces this
functionality under --replace so it's a bit more documented and under
its own semantics.
It is done by returning the process with 133 exit code and picking that
code specifically as a wish to restart.
Some users reported performances issues after recent compositing
scheduling changes, in particular animations being laggy.
The issue is that sometimes it may take more than a vblank interval for
a buffer swap to complete. Previously, compositing was driven by a
timer and it wasn't synchronized to vblanks. If some frame is running
late, the compositor will just start painting a new frame, i.e. the
screen will be rendered triple buffered.
This change disables the swap events code path on Intel to restore the
previous behavior. Unfortunately, that may add an extra frame of latency.
The swap events code path can be enabled explicitly on Intel by setting
the KWIN_USE_INTEL_SWAP_EVENT environment variable to 1.
On Wayland, the highlight window effect is not guaranteed to work all
the time because it uses the opacity to determine if the animation has
completed.
This change rewrites the highlight window effect using AnimationEffect
API to make it work both on X11 and Wayland. The implementation is based
on how the translucency effect works.
