- User Since
- Apr 18 2015, 8:19 AM (127 w, 6 h)
Fix crash on startup - cursor is created before Workspace, so the filter creation needs to be delayed
Out from experience with the KWin Compositing kcm: hide it well, add more warnings and hide it even more. No matter how much you warn against the settings users will change them. Even if it tells them it's the wrong setting for their system and it's going to break there system: we got bug reports for that.
Thu, Sep 21
Wed, Sep 20
Feel free to push either of the two revisions.
-1 from my side. This is a clear bug in whatever is used in the login process. If XDG_RUNTIME_DIR is not created we might have a thousand more things which don't work as expected. I rather have KWin bail out early on that one than have to hunt down all the things which might not work because the system is weird.
Tue, Sep 19
Mon, Sep 18
Do you have commit rights or should one of us push for you?
What surprises me is that BlurManager recreates for Output changes. That sounds like a bug in KWin (or an area which could be improved).
Sun, Sep 17
Just had a look.
If you want you could add another testcase in autotests/libinput/touch_event_test.cpp to cover the special condition.
Sat, Sep 16
Fri, Sep 15
Honestly I'm not sure whether this makes sense to include into KWin directly. While it's a nice gimmick I don't think it's something we should ship with default offerings. Is that really something every user needs in a default setup? This might be handy for development setup, but not in general to ship.
How do we know that this doesn't break other code? This code is for example used in KWin's Alt+Tab handling which is it's own fair beast. Any changes here might break the code as it might be bug-to-bug compatible.
Thu, Sep 14
Why is this needed? Wasn't the idea behind ecm_mark_as_test that they are not built when BUILD_TESTING is off?
Wed, Sep 13
This seems to be a big problem. Kglobalaccell got more new bugs reported this week, than normally in a year.
Tue, Sep 12
I had no idea that this requires a newer CMake. Thanks for fixing!
Now it looks good to me!
Mon, Sep 11
Sun, Sep 10
Addressed David's comments
Check whether we have SCHED_RESET_ON_FORK and ifdef accordingly