task switchers below appear on top left corner under Wayland. breeze dark only text small icons informative compact thumbnails opensuse krypton, qt 5.10
Which exact version are you using?
version 5.11.90git.20171028T043813~145bfd168-93.1 "large icons" switcher is affected too.
Just tried and can reproduce.
So tried some more. When using multi screen it works sometimes.
*** Bug 387456 has been marked as a duplicate of this bug. ***
plasma 5.12 beta is affected.
Also: sometimes the task switcher or just its background blur momentarily appears in the screen centre. Observed on plasma 5.12.
It centers for me on the first invocation sometimes, but then the entries are not highlighted. On the next invocation, it appears at the top left.
(In reply to Michael D from comment #8) > It centers for me on the first invocation sometimes, but then the entries > are not highlighted. On the next invocation, it appears at the top left. Same on my system.
Sometimes after changing of visualization in Task Switcher - it appears in the center (looks like more likely after changing into 'Grid' visualization - using it - and changing back into 'Breeze Dark' for example) - and also after some time of changing of various visualizations - they all starting to show up in the center Plasma: 5.12.4 Apps: 17.12.4 Frameworks: 5.44.0 Qt: 5.10.1 Kernel: 4.14.32-1-MANJARO OS: Netrunner Rolling Video: Intel 4400 Driver: xf86-video-intel 1:2.99.917+823+gd9bf46e4-1 Screen: 1600x900 wayland-protocols 1.13-1 wayland 1.14.0-1
*** Bug 394153 has been marked as a duplicate of this bug. ***
I investigated this problem a bit further. At this late position https://phabricator.kde.org/source/kwin/browse/master/tabbox/tabboxhandler.cpp;184b490c78a3a3e5ae25105c08b19fc375e8c64c$395 the correct d->window()->x() value is still set (in my case to 498). Some time later it gets reset to 0 for some reason. For example in the delayed call to initHighlightWindows it is already 0.
If you fix this, I'll be really close to being able to make the switch to Wayland full time! :)
I have created a Phab task for in-depth technical discussion of the problem(s): https://phabricator.kde.org/T8771
Potential fix: https://phabricator.kde.org/D13084
*** Bug 397032 has been marked as a duplicate of this bug. ***
*** Bug 397586 has been marked as a duplicate of this bug. ***
Git commit 1fb2eace3fbfc6645e024aa0f5003306ac60a417 by Roman Gilg. Committed on 03/09/2018 at 09:25. Pushed by romangilg into branch 'master'. Avoid invalid geometry of internal clients through plasma surface interface Summary: Internal KWin windows might be not in sync with their PlasmaShellSurface. This could be a problem in general, but for now atleast guard against invalid setPosition requests. Test Plan: Manually Reviewers: #kwin, davidedmundson Subscribers: davidedmundson, ngraham, graesslin, kwin Tags: #kwin Maniphest Tasks: T8771 Differential Revision: https://phabricator.kde.org/D13084 M +6 -1 shell_client.cpp https://commits.kde.org/kwin/1fb2eace3fbfc6645e024aa0f5003306ac60a417
This bug is back on neon dev unstable. The following task switchers are not centered on screen: compact thumbnails grid large icons small icons text only informative breeze dark
I'd say that was intentional, see 9b922f88332fc470dcb50e9b55c7d69ab7da4d4c
We need to fix that then, because these task switchers are supposed to be centered.
> We need to fix that then As any other bug ;-) There were definitely some problems with 1fb2eac (if you read the review comments), so we have to find a better solution.
I think there's a solution coming. Once we kill the wl_surface from the internal windows all the races go away and it will "just work". AFAIK that's blocked mostly on shadows that I started.
Also blocked on me having less time currently. Hopefully next week...
Since there is now again enough time on the next release, maybe a proper fix can land before that. If we don't find one in time, the if-conditional can be added back shortly before release. Should still work as a workaround. Of course not so convenient for people on Dev Unstable until then, but it's not a fatal bug. Thanks for reporting back and reopening Patrick.
This hasn't been fixed for 5.16, right? The workaround should then be added back again, which got removed in 9b922f88332fc470dcb50e9b55c7d69ab7da4d4c.
(In reply to Roman Gilg from comment #26) > This hasn't been fixed for 5.16, right? The workaround should then be added > back again, which got removed in 9b922f88332fc470dcb50e9b55c7d69ab7da4d4c. ack, do you want to submit the patch?
Git commit 7c7fb758ad66033c63354c0d71780eb47dc218ce by Roman Gilg. Committed on 05/06/2019 at 14:40. Pushed by romangilg into branch 'Plasma/5.16'. Avoid again invalid geometry of internal clients Summary: This adds again the workaround from 1fb2eace3fbf, which got temporarily removed by 9b922f88332f. Test Plan: Manually Reviewers: #kwin Subscribers: kwin Tags: #kwin Differential Revision: https://phabricator.kde.org/D21601 M +6 -1 shell_client.cpp https://commits.kde.org/kwin/7c7fb758ad66033c63354c0d71780eb47dc218ce
Git commit bebe81209cccc4737a357ad825c6b850ce9302bf by Vlad Zahorodnii, on behalf of Vlad Zagorodniy. Committed on 23/09/2019 at 14:28. Pushed by vladz into branch 'master'. Port QPA away from Wayland Summary: So far wayland was used by internal clients to submit raster buffers and position themselves on the screen. While we didn't have issues with submitting raster buffers, there were some problems with positioning task switchers. Mostly, because we had effectively two paths that may alter geometry. A better approach to deal with internal clients is to let our QPA use kwin core api directly. This way we can eliminate unnecessary roundtrips as well make geometry handling much easier and comprehensible. The last missing piece is shadows. Both Plasma::Dialog and Breeze widget style use platform-specific APIs to set and unset shadows. We need to add shadows API to KWindowSystem. Even though some internal clients lack drop-shadows at the moment, I don't consider it to be a blocker. We can add shadows back later on. Reviewers: #kwin, davidedmundson, romangg Reviewed By: #kwin, romangg Subscribers: romangg, kwin Tags: #kwin Maniphest Tasks: T9600 Differential Revision: https://phabricator.kde.org/D22810 M +6 -0 abstract_client.cpp M +9 -0 abstract_client.h M +9 -7 autotests/integration/debug_console_test.cpp M +8 -6 autotests/integration/decoration_input_test.cpp M +5 -3 autotests/integration/effects/popup_open_close_animation_test.cpp M +6 -4 autotests/integration/globalshortcuts_test.cpp M +44 -54 autotests/integration/internal_window.cpp M +18 -18 composite.cpp M +32 -46 debug_console.cpp M +3 -2 debug_console.h M +7 -1 deleted.cpp M +2 -0 deleted.h M +14 -9 effects.cpp M +1 -4 geometry.cpp M +4 -0 idle_inhibition.cpp M +6 -5 input.cpp M +469 -120 internal_client.cpp M +62 -15 internal_client.h M +6 -6 layers.cpp M +113 -5 platformsupport/scenes/opengl/abstract_egl_backend.cpp M +2 -0 platformsupport/scenes/opengl/abstract_egl_backend.h M +0 -2 plugins/qpa/CMakeLists.txt M +45 -70 plugins/qpa/backingstore.cpp M +3 -15 plugins/qpa/backingstore.h M +3 -61 plugins/qpa/integration.cpp M +2 -18 plugins/qpa/integration.h D +0 -106 plugins/qpa/nativeinterface.cpp D +0 -47 plugins/qpa/nativeinterface.h M +7 -5 plugins/qpa/sharingplatformcontext.cpp M +40 -49 plugins/qpa/window.cpp M +10 -34 plugins/qpa/window.h M +22 -2 plugins/scenes/opengl/scene_opengl.cpp M +10 -0 plugins/scenes/qpainter/scene_qpainter.cpp M +9 -10 pointer_input.cpp M +7 -11 scene.cpp M +8 -0 scene.h M +7 -42 shell_client.cpp M +5 -20 shell_client.h M +1 -1 tabbox/tabbox.cpp M +5 -9 toplevel.cpp M +18 -5 toplevel.h M +2 -35 wayland_server.cpp M +0 -5 wayland_server.h M +47 -25 workspace.cpp M +40 -5 workspace.h https://commits.kde.org/kwin/bebe81209cccc4737a357ad825c6b850ce9302bf