Summary: | Crashes in Scene::Window::buildQuads() | ||
---|---|---|---|
Product: | [Plasma] kwin | Reporter: | David Martini <Damar1980> |
Component: | compositing | Assignee: | KWin default assignee <kwin-bugs-null> |
Status: | RESOLVED UNMAINTAINED | ||
Severity: | crash | CC: | amantia, andresbajotierra, arvidjaar, bigras.bruno, jbrsubscribe, john.bartlett, jonas.vejlin, markotahal, null, thomas.bigot, vide80, vo.zaeb, wittemar, zahl |
Priority: | VHI | ||
Version: | unspecified | ||
Target Milestone: | --- | ||
Platform: | Unlisted Binaries | ||
OS: | Linux | ||
Latest Commit: | Version Fixed In: | ||
Attachments: |
New crash information added by DrKonqi
New crash information added by DrKonqi |
Description
David Martini
2009-06-03 15:23:44 UTC
Backtrace looks related to bug 192172 which mentions using the Snow effect. Were you using Desktop compositing / effects ? Any special effect? Thanks Kwin with compositing activated, no extra activated special effects besides, maybe it's an issue due to the the open driver for the ati readeon graphic card or to slow hardware? radeon X1250 onboard-graphic-card i don't think it's about the acid test at all and likely not about the driver. but i wonder about the static_cast's<Client*>(toplevel) in Scene::Window::buildQuads (there's one in the if/else query relying on the toplevel's geometry and another one in effects->buildQuads() sure that's ok? (i.e. it's called from simple screen here, but can't it ever be a Deleted : Toplevel?) *** Bug 195304 has been marked as a duplicate of this bug. *** Sorry, that dup was a lie. :} ...and yeah, I can't see that having anything to do with acid3, nor can I duplicate it. Changing title to reflect. *** Bug 196152 has been marked as a duplicate of this bug. *** *** Bug 216609 has been marked as a duplicate of this bug. *** *** Bug 216813 has been marked as a duplicate of this bug. *** I see two crashes here, although both in the buildQuads. With yast I can 100% reproduce the crash (see bug 216813), I just need to search for something in it. *** Bug 217461 has been marked as a duplicate of this bug. *** *** Bug 217552 has been marked as a duplicate of this bug. *** *** Bug 217671 has been marked as a duplicate of this bug. *** *** Bug 217761 has been marked as a duplicate of this bug. *** Created attachment 38907 [details]
New crash information added by DrKonqi
there is no need to add crash reports to this bug. We are quite aware of this bug and there are lot's of duplicates which each contains a crash report. *** Bug 217817 has been marked as a duplicate of this bug. *** *** Bug 217831 has been marked as a duplicate of this bug. *** The crashes starting from comment #7 are NOT a duplicate of this bug, but of bug 216908, which is fixed. By that I remove the blocker and will reshuffle all the duplicates. Please set new duplicates to bug 216908. May I /once more/ (see comment #3 ;-) point the static_cast then, esp. now as an untested dynamic_cast from Toplevel -> Client has apparently caused bug #21608? (The backtrace should possibly end in Client::effectWindow(), then but it seems wrong anyway) So i suggest to move WindowQuadList ret; Client *client = dynamic_cast<Client*>( toplevel ); // <- one context up (I did never encounter this bug nor would know to reproduce it, so i can't test if this is it) and test if (client) effects->buildQuads( client->effectWindow(), ret ); (In reply to comment #19) > May I /once more/ (see comment #3 ;-) point the static_cast then, esp. now as > an untested dynamic_cast from Toplevel -> Client has apparently caused bug > #21608? it wasn't an untested dynamic_cast it was a typo ;-) A client was used where a toplevel should have been used. > (I did never encounter this bug nor would know to reproduce it, so i can't test > if this is it) me neither > > and test > if (client) > effects->buildQuads( client->effectWindow(), ret ); why at all the cast to Client? Why not toplevel->effectWindow()? (In reply to comment #20) > it wasn't an untested dynamic_cast it was a typo ;-) Rather a "Typo" ;-P - I hadn't looked at the diff but just the bugreport. > why at all the cast to Client? Why not toplevel->effectWindow()? Indeed, it's not virtual. May this actually cause trouble with some compilers? The function offsets should not be affected (as it's a base function) but it's also an inliner...? Created attachment 39097 [details]
New crash information added by DrKonqi
(In reply to comment #22) > Created an attachment (id=39097) [details] > New crash information added by DrKonqi Your crash has nothing to do with this crash. Your crash is reported in bug 216908 and already fixed. (In reply to comment #23) > Your crash has nothing to do with this crash. Your crash is reported in bug > 216908 and already fixed. Sorry, too much trust in the new DrKonqui functionalities :P Forget it This crash had been reported against 4.3. This is a very long time ago and the stack trace does no longer match our code at all. Most likely the crash has been resolved, e.g. the issue pointed out in comment #3 is no longer present. I assume that this crash had been resolved, as I cannot be completely sure about it, I change the state of this crash report to unmaintained, which is in fact true about 4.3. In case you are still experiencing this crash with a recent version (at the time of this writing 4.7.4 or 4.8.1) please provide a new crash trace and reopen the bug. |