| Summary: | crash (assert) on desktop change, apparently in KWinInternal::Client::plainResize | ||
|---|---|---|---|
| Product: | [Plasma] kwin | Reporter: | Melchior Franz <mfranz> |
| Component: | general | Assignee: | KWin default assignee <kwin-bugs-null> |
| Status: | RESOLVED DUPLICATE | ||
| Severity: | crash | ||
| Priority: | NOR | ||
| Version First Reported In: | unspecified | ||
| Target Milestone: | --- | ||
| Platform: | Compiled Sources | ||
| OS: | Linux | ||
| Latest Commit: | Version Fixed/Implemented In: | ||
| Sentry Crash Report: | |||
Can you reproduce the problem? * Lubos Lunak -- Monday 16 February 2004 18:36:
> Can you reproduce the problem?
No. But it's not the first time I saw it. Happened a few times already,
but very seldom (maybe once every few months).
|
Version: (using KDE Devel) Installed from: Compiled sources Compiler: gcc (GCC) 3.3.1 (SuSE Linux) OS: Linux When switching from desktop 1 to desktop 4 (running kuickshow in fullscreen mode and on all desktops amarok), kwin aborted with the below bt. After restarting kwin, there were only two (instead of my normal 7) desktops available, the window style was switched to keramik (normally ModernSystem). I restarted the X server (maybe kdesktop would have been enough?!) and now the selection text color was suddenly black (normally white). All very annoying, but this very report is about the kwin crash. :-) From gdb: Program terminated with signal 6, Aborted. (gdb) bt #0 0x4162eb71 in kill () from /lib/i686/libc.so.6 #1 0x414b0cf1 in pthread_kill () from /lib/i686/libpthread.so.0 #2 0x414b100b in raise () from /lib/i686/libpthread.so.0 #3 0x4162e904 in raise () from /lib/i686/libc.so.6 #4 0x4162fe8c in abort () from /lib/i686/libc.so.6 #5 0x41627e84 in __assert_fail () from /lib/i686/libc.so.6 #6 0x41d86a9d in KWinInternal::Client::plainResize(int, int, KWinInternal::ForceGeometry_t) (this=0x4, w=-1073749280, h=164, force=NormalGeometrySet) at geometry.cpp:1174 #7 0x41d858fa in KWinInternal::Client::configureRequest(int, int, int, int, int, int) (this=0x81b0498, value_mask=15, rx=0, ry=0, rw=311, rh=140, gravity=1) at qsize.h:132 #8 0x41d6776d in KWinInternal::Client::configureRequestEvent(XConfigureRequestEvent*) (this=0x81b0498, e=0xbfffe960) at events.cpp:774 #9 0x41d66f69 in KWinInternal::Client::windowEvent(_XEvent*) (this=0x81b0498, e=0xbfffe960) at events.cpp:521 #10 0x41d66b05 in KWinInternal::Workspace::workspaceEvent(_XEvent*) (this=0x81acb70, e=0xbfffe960) at events.cpp:250 #11 0x41d5bd32 in KWinInternal::Application::x11EventFilter(_XEvent*) (this=0x0, e=0xbfffe960) at workspace.h:71 #12 0x40c04efb in qt_x11EventFilter (ev=0xbfffe960) at kernel/qapplication_x11.cpp:382 #13 0x40c0df20 in QApplication::x11ProcessEvent(_XEvent*) (this=0xbfffeaf0, event=0xbfffe960) at kernel/qapplication_x11.cpp:3253 #14 0x40c28fdc in QEventLoop::processEvents(unsigned) (this=0x812d580, flags=4) at kernel/qeventloop_x11.cpp:192 #15 0x40c958be in QEventLoop::enterLoop() (this=0x812d580) at kernel/qeventloop.cpp:198 #16 0x40c957da in QEventLoop::exec() (this=0x812d580) at kernel/qeventloop.cpp:145 #17 0x40c8189d in QApplication::exec() (this=0xbfffeaf0) at kernel/qapplication.cpp:2737 #18 0x41d5c1cc in kdemain (argc=1, argv=0x8060ca0) at main.cpp:268 #19 0x4096b93c in kdeinitmain (argc=0, argv=0x0) at kwin_dummy.cpp:2 #20 0x0804d687 in launch (argc=1, _name=0x805e84c "kwin", args=0xa <Address 0xa out of bounds>, cwd=0x0, envc=0, envs=0x1 <Address 0x1 out of bounds>, reset_env=false, tty=0x0, avoid_loops=false, startup_id_str=0x8050cf4 "0") at kinit.cpp:604 #21 0x0804ff76 in handle_launcher_request (sock=8) at kinit.cpp:1167 #22 0x0804ec5e in handle_requests (waitForPid=0) at kinit.cpp:1360 #23 0x0804c918 in main (argc=3, argv=0xbffff2c4, envp=0x0) at kinit.cpp:1797 this == 0x4? Coool ... ;-) (gdb) up 6 #6 0x41d86a9d in KWinInternal::Client::plainResize(int, int, KWinInternal::ForceGeometry_t) (this=0x4, w=-1073749280, h=164, force=NormalGeometrySet) at geometry.cpp:1174 1174 assert( false ); (gdb) list 1169 { 1170 // check that the frame is not resized to full size when it should be shaded 1171 if( !shade_geometry_change && h != border_top + border_bottom ) 1172 { 1173 kdDebug() << "h:" << h << ":t:" << border_top << ":b:" << border_bottom << endl; 1174 assert( false ); 1175 } 1176 client_size = QSize( w - border_left - border_right, client_size.height()); 1177 } 1178 updateWorkareaDiffs(); (gdb) I have one tiny local change in geometry.cpp that doesn't seem to be related. So feel free to dismiss the bug report. :-/