| Summary: | KWin Crashes on Start-Up | ||
|---|---|---|---|
| Product: | [Plasma] kwin | Reporter: | Wedge009 <wedge009> |
| Component: | general | Assignee: | KWin default assignee <kwin-bugs-null> |
| Status: | RESOLVED DUPLICATE | ||
| Severity: | crash | ||
| Priority: | NOR | ||
| Version First Reported In: | unspecified | ||
| Target Milestone: | --- | ||
| Platform: | Ubuntu | ||
| OS: | Linux | ||
| Latest Commit: | Version Fixed/Implemented In: | ||
| Sentry Crash Report: | |||
|
Description
Wedge009
2011-08-08 08:36:39 UTC
yes, is dupe. no. is not threading but (apparently) a bug in the mesa vertex buffer object code - just "gg bind_inputs segfault" The virtualbox issue is known. Compositing shouldn't be enabled in virtualbox since the software renderer doesn't work - this "blacklisting" was turned off by accident. No idea about the N455 - but there're several known bugs & issues with the intel GL implementation. You'll have to file a bug for such crash (alongside the backtrace of course) for further investigation. Most annoying bug here is bug #252817 Have a look, check whether the workaround (not suspending fulsscreen windows) works for you. But, good god: do not post a "me too" to that report ;-) *** This bug has been marked as a duplicate of bug 278838 *** Thanks again for the information. Does that mean the bug is in the Mesa libraries? I understand that a virtual machine is probably not the best place to be playing around with compositing, but VirtualBox is not the same as VMWare, as far as I know, and as I said, it was working (at least the basic features) quite nicely in 4.6.x. To have it disabled completely in 4.7.0 is a bit of a let-down. I'm not too fussed about the Intel implementation, but if it becomes too annoying, I may try obtaining a stack trace on the N455. When it does crash, it's only at the start-up and there doesn't seem to be further issues after manually re-enabling compositing. I'm certainly not going to add to that bug report you mentioned. |