Version: (using KDE 4.2.0) Compiler: GCC 4.3.2 Target: x86_64-unknown-linux-gnu Configured with: ../gcc-4.3.2/configure --prefix=/usr --libexecdir=/usr/lib --enable-shared --enable-threads=posix --enable-__cxa_atexit --enable-clocale=gnu --enable-languages=c,c++,fortran,objc,treelang --disable-multilib --enable-c99 --enable-long-long Thread model: posix OS: Linux Installed from: Compiled From Sources Sometimes plasmoids start to drop over time during long work (about 4-5 hours). Not all at one moment but one by one in some time. But in 20-30 mins you have all the plasmoids disappeared. Plasmoids on desktop become transparent and on panel disappear at all (see the following screenshot). If I click on transparent plasmoid Plasma crashes, but I think it's a separate bug, so I'll file it in another report.
Created attachment 30908 [details] Screenshot with dropped plasmoids
The bug report concerning crash in such a situation is bug #182999
Persists for Qt 4.5 final and KDE 4.2.1.
Sorry, is there any attention to this bug? Actually, it's so annoying because in fact all the plasmoid disappear as if they crashed and you have no control over anything. The only way to get plasmoids working fine again is to kill and start plasma, but it's for another 20-30 mins. Also, not all the apps are restored in tray (for example Stardict floats after killing plasma once) and notifications paints only frame without message (I didn't post a bug report for this because it happens only after killing plasma once). So it's _very_ harmful. I would provide any information if it's needed.
this is almost certianly a graphics driver problem, and certainly not due to plasma.
Aaron, thank you for the response and reference. I really had no ideas about the cause of such a bag. Sorry for blaming Plasma :)
no problem; it happens. plasma exercises a lot of code paths in drivers that few if any other apps are these days. it's getting slowly worked out in the drivers, though there were some annoying regressions in the current intel driver recently.
As to me, this bug is on ATI card with radeonhd-1.2.4. Aaron, could you suggest me, is this problem more likely in driver exactly or it could be something inside X? I'm thinking of recompiling whole X now and it's interesting whether there're some sense doing it. In any case, if I manage to fix it, I'll post the result here.
I've switched to xf86-video-ati-6.12.1 and the bug is reproducible. Is it really not a Qt or KDE bug? KWin and all the other graphics stuff work fine.