Distributions do package each backend plugin in a dedicated package,
which means just because KWin got compiled with a specific backend,
doesn't mean it is also available at runtime.
In order to better support this reality this change introduces a
list-backends command line option. Each of the plugins got the json
metadata extended by the KPlugin syntax, so that we have a name and
description to print.
As we already locate all plugins anyway, the additional findPlugins for
the selected backend is changed to search the list of all plugin meta
data.
AbstractClient now also supports mainClients, so we can do the
fullscreen window check on setActive in a general way. This ensures
that we do get proper stacking changes for activating fullscreen
shell clients.
Similar to what we have in Client::manage we should not always
invoke the Placement algorithm. E.g. an initially fullscreened window
should not get placed.
This needs to be extended for more checks similar to Client::manage
and might indicate that we also need a ShellClient::manage.
REVIEW: 125469
Whenever we set a transient we must register it with the parent and
remove it again. Also if a parent gets destroyed we must inform the
transients.
This fixes a crash in Deleted::copyToDeleted when connecting the main
clients. The crash condition was hit if the parent got destroyed
before the transient.
Reviewed-By: Marco Martin
Main purpose is to reproduce a crash if the parent window gets
destroyed before the transient. For this the following key combo
can be used:
t: show transient
k: close parent window
q: quit application
This currently causes reliably a crash in Deleted::copyToDeleted for
the connect of the mainClients.
Following the approach how it's done for X11: only create the interface
if the shader succeeded to compile and remove support again if it failed
to compile after a reconfigure.
REVIEW: 125444
Discovered by ASAN on the CI system causing the
TestClientMachine::hostName(ostname) to fail since it's enabled.
What happened is that the QByteArray returned by getHostName() gets
destroyed in the main thread but accessed in the resolve thread. This
is changed by calling getHostName in the resolve thread.
REVIEW: 125458
With Qt 5.5 the physicalSize of a screen is broken if the X system does
not provide the XRandR extension. This caused our screen edge test to
fail on the CI system (Xvfb) due to approach window being based on the
dpi.
The problem itself is addressed for Qt in:
https://codereview.qt-project.org/126808
This change just worksaround till the change has made it into our CI
system (at which point the expected fails will break).
Reviewed-By: David Edmundson
When pressing right control key the input gets grabbed/released.
The grab is for both keyboard and pointer and pointer is also confined
to the window.
How to grab/ungrab is added to the window title.
in a way so that each entry has its own
drawback: present mnemonics changed.
they however also change with i18n.
REVIEW: 125434
BUG: 319695
FIXED-IN: 5.5
the pseudo maximization sanitizes huge windows and
was/is the way kdelibs KMainWindow indicates maximization
It must however build up on the hinted maximization
and not MaximizeRestore, or it may falsely destroy
the maximization hint from the client (which now works
in KF5 ;-)
BUG: 352185
REVIEW: 125037
FIXED-IN: 5.5
* skip all overhead on unrestricted moveResizes
* the titlebar is not guaranteed to be the top border (anymore)
* don't block vertical resize when resizing on a top corner
* reduce QRegion operations by knowledge about the degenerated titlebarregion (rect)
* rename int (sic!) "titleArea" to "requiredPixels"
REVIEW: 125428
commit c6bd68d78a2e00ce094a4087c106aaf4067f9f57 fixed a bug
that allowed early invocation when pushback was forced down.
However, it also turned activation and re-activation time
additive again (ie. invocation could only happen after
r+a ms instead of after r ms)
see also REVIEW 124888
REVIEW: 125143
a) if the move is left after the delay is triggered
but before the timer hit, do not cause a stale QT
indication
b) only delay QT if there's actually another screen
on the other side
BUG: 352744
REVIEW: 125250
Holds AbstractClients that is both X11 and Wayland clients. Allows
to easily change code which needs to operate on all clients to get
to them without needing special handling for Wayland clients. At the
same time we are still able to get to the windowing system specific
clients through the old clientList() and waylandServer()->clients().
This fixes yet another regression from the transient refactoring.
The passed in client might be null, so we need a nullptr check. There
are several already in that code.
Was tricky given that I removed a cast there.
For Wayland transients are popups relative to a parent surface. This
means for a dock window we obviously want them above the dock, otherwise
the context menu would be below the dock.
To not break the existing functionality (which makes sense) we bind
it to whether the transient has a positioning hint - that's only set
for ShellClients.
Merges together the code from ShellClient and Client and removes the
starting differences. Long term it's better to have only one
implementation to prevent diversions in the implementation.
As it doesn't match exactly protected virtual methods are called
which allow more specific implementations for a certain aspect of the
layer resolving.
