Summary:
If you switch virtual desktops while krunner is sliding in, then
depending on whether your distro strips assert statements away,
KWin can crash.
The reason why it crashes is the sliding popups effect tries to unref
deleted windows that it hasn't referenced before (if there is an active
full screen effect, then popups won't be slided out, which in its turn
means that we won't reference deleted windows). So, in the end, the
refcount of those windows can be -1. That triggers an assert statement
in the destructor of the Deleted class, which checks whether the
refcount is equal to 0.
Popups are not slided while there is an active full screen effect because
we don't know what the full screen effect does.
This patch adjusts the sliding popups effect so it stops all active
animations when user switches virtual desktops or when a full screen
effect kicks in. We need to do that so the effect won't try to
unreference windows in postPaintWindow.
Visually, it doesn't look quite nice, but for now that's good enough.
A proper fix would be more complex: we would need to make sure that
full screen effects ignore sliding popups (and also maybe docks) and
perform some input redirection.
BUG: 400170
FIXED-IN: 5.14.4
Test Plan: I'm not able anymore to reproduce bug 400170.
Reviewers: #kwin, graesslin
Reviewed By: #kwin, graesslin
Subscribers: davidedmundson, graesslin, kwin
Tags: #kwin
Differential Revision: https://phabricator.kde.org/D16731
Summary:
If you activate the Present Windows effect and then the Desktop Grid,
you'll be able to see the close button from PW. The reason for that is
PW doesn't destroy the close button and DG doesn't filter it out.
This patch addesses this problem by syncing DesktopGridEffect::isRelevantWithPresentWindows
with PresentWindowsEffect::isSelectableWindow.
On X11, the close button is filtered by the isSpecialWindow check. On
Wayland, the close button is filtered by the acceptsFocus check.
The proposed solution is kinda hack-ish, but on the other hand, we have
to keep those two methods in sync anyway.
In addition to the close button, notifications won't be displayed too.
BUG: 364710
FIXED-IN: 5.14.3
Test Plan:
* Activate the Present Windows effect;
* Activate the Desktop Grid effect;
* (the close button is no longer visible)
Reviewers: #kwin, graesslin
Reviewed By: #kwin, graesslin
Subscribers: kwin
Tags: #kwin
Differential Revision: https://phabricator.kde.org/D16513
Summary:
When using the natural layout algorithm with the fill-gaps option, a small
error (less than one) is introduced in windows' aspect ratio each time they are
enlarged due to floating-point roundoff.
Currently, the algorithm computes the width and height enlargement factors and
then attempts to enlarge in each of the four possible directions, repeating
until it can't enlarge any windows any further. Hence, this aspect ratio error
can be multiplied by up to four. Especially for small, long, and narrow
windows, this can result in a total error of greater than one by the end of
that loop iteration. If this occurs, on subsequent iterations the height
enlargement factor might then be computed as negative violating some of the
core assumptions of the algorithm and resulting in the loop iterating endlessly
until one of the window dimensions overflows, freezing the program for up to
several minutes.
To fix this, the height enlargement factor should be re-computed based on the
new width each time the window is enlarged, ensuring the error introduced in
the aspect ratio never exceeds one.
BUG: 364709
BUG: 380865
BUG: 368811
FIXED-IN: 5.15.0
Test Plan:
The most reliable way to reproduce the freeze seems to be to activate the
desktop-grid effect while a tool-tip window is fading in.
Ensure desktop-grid is configured to use present windows, and that present
windows is configured to use the natural layout algorithm with the fill gaps
option selected.
The freeze is still intermittent, but using this method should be able to be
triggered within about 10 tries without this fix.
After applying the fix, the freeze has never been observed.
Reviewers: #kwin, zzag
Reviewed By: #kwin, zzag
Subscribers: graesslin, kwin, zzag
Tags: #kwin
Differential Revision: https://phabricator.kde.org/D16278
Summary:
If user selects a shortcut to trigger the effect, that shortcut won't
work after the KCM is closed because we don't save it.
BUG: 400351
FIXED-IN: 5.14.3
Test Plan:
* Specified custom shortcut to trigger the effect;
* Closed the KCM;
* Pressed the shortcut, a thumbnail appeared.
Reviewers: #kwin, graesslin
Reviewed By: #kwin, graesslin
Subscribers: kwin
Tags: #kwin
Differential Revision: https://phabricator.kde.org/D16456
Summary:
Currently, when the user is moving a window in the desktop grid it is drawn from
the paintScreen function to ensure it appears above all other windows. However,
when drawing the window, the screenProjectionMatrix for the WindowPaintData structure
is never specified. Because of this, the window will not be visible if OpenGL
compositing along with Lanczos scaling are in use since its coordinates will never
be projected into screen-space (unless the window is maximized, in which case the
scaling code is bypassed).
BUG: 361371
BUG: 364509
FIXED-IN: 5.14.3
Test Plan:
Ensure OpenGL compositing is enabled and the scaling method is set to "Accurate" on
hardware supporting this option. Additionally, ensure the desktop grid effect is
active and is configured to use Present Windows. Open a window and ensure is it not
maximized. Click on the window and drag it around - its contents should remain visible
the entire time and follow the mouse.
Reviewers: #kwin, zzag, graesslin, davidedmundson
Reviewed By: #kwin, zzag, davidedmundson
Subscribers: broulik, davidedmundson, kwin, #kwin
Tags: #kwin
Differential Revision: https://phabricator.kde.org/D16430
Summary:
Currently, when the user is moving a window in the desktop grid it is drawn from
the paintScreen function to ensure it appears above all other windows. However,
when drawing the window, the screenProjectionMatrix for the WindowPaintData structure
is never specified. Because of this, the window will not be visible if OpenGL
compositing along with Lanczos scaling are in use since its coordinates will never
be projected into screen-space (unless the window is maximized, in which case the
scaling code is bypassed).
BUG: 361371
BUG: 364509
FIXED-IN: 5.14.3
Test Plan:
Ensure OpenGL compositing is enabled and the scaling method is set to "Accurate" on
hardware supporting this option. Additionally, ensure the desktop grid effect is
active and is configured to use Present Windows. Open a window and ensure is it not
maximized. Click on the window and drag it around - its contents should remain visible
the entire time and follow the mouse.
Reviewers: #kwin, zzag, graesslin, davidedmundson
Reviewed By: #kwin, zzag, davidedmundson
Subscribers: broulik, davidedmundson, kwin, #kwin
Tags: #kwin
Differential Revision: https://phabricator.kde.org/D16430
Summary:
Background behind some windows is not blurred because the invert effect
sets PAINT_WINDOW_TRANSFORMED. This essentially "turns off" the blur and
the background contrast effect unless we set the force roles.
Because the invert effect is a "post-processing"(not really) effect we
don't have to set that flag.
BUG: 279076
BUG: 359583
FIXED-IN: 5.14.3
Test Plan: {F6341916}
Reviewers: #kwin, davidedmundson
Reviewed By: #kwin, davidedmundson
Subscribers: kwin
Tags: #kwin
Differential Revision: https://phabricator.kde.org/D16358
Avoids triggering an assert and communicates that the timeout is running.
Ideally, there was a countdown label of some sort but this is better than crashing.
BUG: 399644
FIXED-IN: 5.12.8
Differential Revision: https://phabricator.kde.org/D16124
Summary:
The Dim Inactive effect expects that the windowClosed signal always
proceeds the windowDeleted signal. But in some cases that's not the case.
If a window gets destroyed before becoming ready for painting, only
the windowDeleted signal will be emitted. In addition to that, KWin will
activate that window, which means we'll probably start a transition for
it.
Because this effect cannot terminate active transitions for such
windows, KWin can crash in postPaintScreen.
This change addresses the crash in postPaintScreen by adding extra clean
up stuff in the windowDeleted slot to make sure that there are no
transitions for deleted windows.
The proper fix would be to not emit windowActivated signal for windows
that are not ready for painting.
BUG: 399612
FIXED-IN: 5.14.1
Test Plan:
Ran
```
x <- seq(5, 15, length=1000)
y <- dnorm(x, mean=10, sd=3)
plot(x, y, type="l", lwd=1)
```
in RKWard multiple times.
Reviewers: #kwin
Subscribers: kwin
Tags: #kwin
Differential Revision: https://phabricator.kde.org/D16130
Summary:
There's an attempt to block updates to activities whilst the popup menu
is showing.
In one of the two code paths for positioning the menu the block is set
but instead of releasing the block at the end of the exec locks it
again.
CCBUG: 335725
Test Plan:
On X11 changed activities with the popup menu near the bottom.
It still flickered like crazy whilst the menu is open (see bug report)
But now at least fixes itself after
Reviewers: #kwin, zzag
Reviewed By: #kwin, zzag
Subscribers: kwin
Tags: #kwin
Differential Revision: https://phabricator.kde.org/D16101
Summary:
Qt does not like having no screens. Both Qt XCB and Wayland QPAs have
systems pretend there's always at least 1 screen
present.
Kwin already uses a dummy screen on startup, this patch reinserts the
dummy screen if the platform states that all real screens are removed.
BUG: 399564
Test Plan:
Logged in, everything still worked
Did not reproduce the original bug
Reviewers: #kwin, zzag
Reviewed By: #kwin, zzag
Subscribers: kwin
Tags: #kwin
Differential Revision: https://phabricator.kde.org/D16096
Summary:
If one wants to drag a maximized window while he or she has enabled both
the Wobbly Windows and the Maximize effect, visual artifacts can be
shown.
When the Maximize effect animates transition from maximize to restore,
some parts of the window can overshoot repaint regions(because of
wobbliness), leaving us with the visual artifacts.
To fix that, we have to take into account current scale and translation
when calculating dirty region in the Wobbly Windows effect.
BUG: 370612
FIXED-IN: 5.14.1
Test Plan: When dragging maximized windows, there are no visual artifacts anymore.
Reviewers: #kwin, davidedmundson
Reviewed By: #kwin, davidedmundson
Subscribers: abetts, kwin
Tags: #kwin
Differential Revision: https://phabricator.kde.org/D15961
Summary:
QOpenGLContext was changed to call platform->makeCurrent before it
called setCurrentContext.
Because we bind window FBO and ultimately that calls
QOpenGlContext::format we need Qt to know which context is current so it
can tell us the format.
This matches the QtWayland EGL code.
BUG: 399392
Test Plan:
Switched virtual desktop with the OSD (previously reliably crashed)
Now it doesn't
Reviewers: #kwin, zzag
Reviewed By: #kwin, zzag
Subscribers: zzag, kwin
Tags: #kwin
Differential Revision: https://phabricator.kde.org/D15957
Summary:
With the Oxygen cursor theme, a visual glitch can appear in the
bottom-left corner of the cursor when the screen is zoomed in.
The reason for that is we're using GL_REPEAT for GL_TEXTURE_WRAP_S
and GL_TEXTURE_WRAP_T. So, when the linear filter samples neighboring
texels in the bottom, it can sample some texels from the top (because of
GL_REPEAT).
To fix that, we have to set the wrap mode to GL_CLAMP_TO_EDGE.
BUG: 338254
FIXED-IN: 5.14.0
Test Plan:
Before:
{F6303912}
After:
{F6303913}
Tried other cursor themes(Breeze and Adwaita), everything looks okay.
Reviewers: #kwin, davidedmundson
Reviewed By: #kwin, davidedmundson
Subscribers: davidedmundson, kwin
Tags: #kwin
Differential Revision: https://phabricator.kde.org/D15949
Summary:
kwin_wayland now automatically chooses an appropriate backend, such as
DRM, nested wayland or nested X. If nested it will automatically go into
windowed mode regardless of whether --windowed is set and works fine.
Backend choosing logic existed duplicated in older code for kwin_wayland
--windowed, with the subtle unexpected difference that it preferred
running as an X client over running as a wayland cient if both are
present. This simplifies codes and syncs automatic resolution behaviour.
kwin_wayland --windowed with --x11-display or --wayland-display will
remain the same.
Test Plan:
Ran kwin_wayland with and without --windowed inside another wayland.
Got the same backend chosen
Tested that kwin_wayland (without --windowed) on an X machine worked just fine.
Reviewers: #kwin, mart
Reviewed By: #kwin, mart
Subscribers: mart, kwin
Tags: #kwin
Differential Revision: https://phabricator.kde.org/D13658
Summary:
Qt render loops behave quite differently to each other.
KWin scripting as a workaround for another situation cleans
(17553e5a1f) handles tracking of script
windows by deleting the underlying window handle on hide.
This currently happens before the window gets the hideEvent.
Arguably this is a quirk with Qt, but in the current state:
- QSGThreadedRenderLoop deletes the platform window and cleans up
- We then get the hide() event. This no-ops because there's no window.
(else branch of case WM_TryRelease in qsgthreadedrenderloop.cpp)
- We carry on rendering animations despite having no platform
- undefined behaviour
Normally this isn't a problem as typically destruction of the platform window
happens only when a window is being deleted, we're messing with Qt
internals here.
If we make sure the QHideEvent is processed by the render loop first,
things seem fine.
BUG: 397767
Test Plan:
Ran QSG_RENDER_LOOP=threaded
Read output with scenegraph logging rules on
Reviewers: #kwin, broulik
Reviewed By: broulik
Subscribers: kwin
Tags: #kwin
Differential Revision: https://phabricator.kde.org/D15025
