Summary: | lot of QXcbConnection: XCB error: 3 (BadWindow) | ||
---|---|---|---|
Product: | [Plasma] kwin | Reporter: | luigide |
Component: | compositing | Assignee: | KWin default assignee <kwin-bugs-null> |
Status: | RESOLVED WORKSFORME | ||
Severity: | normal | CC: | abone27, jsamyth, steffen.klee |
Priority: | NOR | ||
Version: | 5.1.0 | ||
Target Milestone: | --- | ||
Platform: | Arch Linux | ||
OS: | Linux | ||
Latest Commit: | Version Fixed In: | ||
Sentry Crash Report: |
Description
luigide
2014-10-24 20:05:08 UTC
The warning about the decoration says that you don't have the Oxygen decoration installed. Otherwise none of them are errors and just debug output. fyi, the xcb errors happen because "someone" (kwin in this case) attempts to alter a property on a window that has already been destroyed. X11 is an async protocol, so this can happen. The only way around would be to - grab the server - check whether the window is still there - alter the property and then - release the server You would not like that behavior, though ;-) there is another way, though I haven't tested it yet. At XDC I had been told by the designer of XCB that if we used the checked variants and discard the reply we wouldn't get the errors any more. because xcb sucks the error instead of putting it into the event queue =) probably required for kwindowsystem because of the "gdk: i kill myself on X11 errors" attitude. Wow. Two years later and this stuff is still flooding everybody's logs. People have no pride in workmanship these days. I've fired guys for less. @John: this is a volunteer project. This is totally not acceptable behavior. Please have a look at https://www.kde.org/code-of-conduct/ and reflect what you wrote. |