today's kwin/kwindowsystem build crashes in kstartupinfo processing Application: (kwin), signal: Segmentation fault Using host libthread_db library "/lib/x86_64-linux-gnu/libthread_db.so.1". [KCrash Handler] #6 __memcmp_sse4_1 () at ../sysdeps/x86_64/multiarch/memcmp-sse4.S:908 #7 0x00007fb75c5ee7c0 in qstrcmp (str1=..., str2=...) at tools/qbytearray.cpp:347 #8 0x00007fb7615b1a3d in operator<(QByteArray const&, QByteArray const&) () from /opt/project-neon5/lib/x86_64-linux-gnu/libKF5WindowSystem.so.5 #9 0x00007fb7615ae643 in KStartupInfoId::operator<(KStartupInfoId const&) const () from /opt/project-neon5/lib/x86_64-linux-gnu/libKF5WindowSystem.so.5 #10 0x00007fb7615b3db3 in bool qMapLessThanKey<KStartupInfoId>(KStartupInfoId const&, KStartupInfoId const&) () from /opt/project-neon5/lib/x86_64-linux-gnu/libKF5WindowSystem.so.5 #11 0x00007fb7615b4693 in QMapNode<KStartupInfoId, KStartupInfo::Data>::lowerBound(KStartupInfoId const&) () from /opt/project-neon5/lib/x86_64-linux-gnu/libKF5WindowSystem.so.5 #12 0x00007fb7615b3cb2 in QMapData<KStartupInfoId, KStartupInfo::Data>::findNode(KStartupInfoId const&) const () from /opt/project-neon5/lib/x86_64-linux-gnu/libKF5WindowSystem.so.5 #13 0x00007fb7615b2e53 in QMap<KStartupInfoId, KStartupInfo::Data>::remove(KStartupInfoId const&) () from /opt/project-neon5/lib/x86_64-linux-gnu/libKF5WindowSystem.so.5 #14 0x00007fb7615ab622 in KStartupInfo::Private::remove_startup_info_internal(KStartupInfoId const&) () from /opt/project-neon5/lib/x86_64-linux-gnu/libKF5WindowSystem.so.5 #15 0x00007fb7615ad3e8 in KStartupInfo::Private::find_pid(int, QByteArray const&, KStartupInfoId*, KStartupInfoData*) () from /opt/project-neon5/lib/x86_64-linux-gnu/libKF5WindowSystem.so.5 #16 0x00007fb7615acfbc in KStartupInfo::Private::check_startup_internal(unsigned long long, KStartupInfoId*, KStartupInfoData*) () from /opt/project-neon5/lib/x86_64-linux-gnu/libKF5WindowSystem.so.5 #17 0x00007fb7615accca in KStartupInfo::checkStartup(unsigned long long, KStartupInfoId&, KStartupInfoData&) () from /opt/project-neon5/lib/x86_64-linux-gnu/libKF5WindowSystem.so.5 #18 0x00007fb764710e81 in KWin::Workspace::checkStartupNotification(unsigned int, KStartupInfoId&, KStartupInfoData&) () from /opt/project-neon5/lib/x86_64-linux-gnu/libkdeinit4_kwin.so #19 0x00007fb764769ea7 in KWin::Client::manage(unsigned int, bool) () from /opt/project-neon5/lib/x86_64-linux-gnu/libkdeinit4_kwin.so #20 0x00007fb76470e0de in KWin::Workspace::createClient(unsigned int, bool) () from /opt/project-neon5/lib/x86_64-linux-gnu/libkdeinit4_kwin.so #21 0x00007fb76475321b in KWin::Workspace::workspaceEvent(xcb_generic_event_t*) () from /opt/project-neon5/lib/x86_64-linux-gnu/libkdeinit4_kwin.so #22 0x00007fb764746d76 in KWin::XcbEventFilter::nativeEventFilter(QByteArray const&, void*, long*) () from /opt/project-neon5/lib/x86_64-linux-gnu/libkdeinit4_kwin.so #23 0x00007fb75c84f4af in QAbstractEventDispatcher::filterNativeEvent (this=0x754c50, eventType=..., message=0x7fb748009f30, result=0x7ffff2b69d00) at kernel/qabstracteventdispatcher.cpp:466 #24 0x00007fb751845dad in QXcbConnection::handleXcbEvent (this=0x735ac0, event=0x7fb748009f30) at qxcbconnection.cpp:806 #25 0x00007fb751847b7b in QXcbConnection::processXcbEvents (this=0x735ac0) at qxcbconnection.cpp:1213 #26 0x00007fb7518a00d7 in QXcbConnection::qt_static_metacall (_o=0x735ac0, _c=QMetaObject::InvokeMetaMethod, _id=0, _a=0x7fb7480018e0) at .moc/moc_qxcbconnection.cpp:179 #27 0x00007fb75c88e985 in QMetaCallEvent::placeMetaCall (this=0x7fb748004640, object=0x735ac0) at kernel/qobject.cpp:487 #28 0x00007fb75c88f9da in QObject::event (this=0x735ac0, e=0x7fb748004640) at kernel/qobject.cpp:1147 #29 0x00007fb75d6706b6 in QApplicationPrivate::notify_helper (this=0x725f90, receiver=0x735ac0, e=0x7fb748004640) at kernel/qapplication.cpp:3482 #30 0x00007fb75d66db19 in QApplication::notify (this=0x7ffff2b6aac0, receiver=0x735ac0, e=0x7fb748004640) at kernel/qapplication.cpp:2903 #31 0x00007fb764746bb8 in KWin::Application::notify(QObject*, QEvent*) () from /opt/project-neon5/lib/x86_64-linux-gnu/libkdeinit4_kwin.so #32 0x00007fb75c853b7e in QCoreApplication::notifyInternal (this=0x7ffff2b6aac0, receiver=0x735ac0, event=0x7fb748004640) at kernel/qcoreapplication.cpp:881 #33 0x00007fb75c8577eb in QCoreApplication::sendEvent (receiver=0x735ac0, event=0x7fb748004640) at ../../include/QtCore/../../src/corelib/kernel/qcoreapplication.h:232 #34 0x00007fb75c854e7e in QCoreApplicationPrivate::sendPostedEvents (receiver=0x0, event_type=0, data=0x701400) at kernel/qcoreapplication.cpp:1485 #35 0x00007fb75c8c7b8d in QEventDispatcherUNIX::processEvents (this=0x754c50, flags=...) at kernel/qeventdispatcher_unix.cpp:587 #36 0x00007fb7518decee in QUnixEventDispatcherQPA::processEvents (this=0x754c50, flags=...) at eventdispatchers/qunixeventdispatcher.cpp:70 #37 0x00007fb75c850902 in QEventLoop::processEvents (this=0x7ffff2b6a8e0, flags=...) at kernel/qeventloop.cpp:136 #38 0x00007fb75c850bcd in QEventLoop::exec (this=0x7ffff2b6a8e0, flags=...) at kernel/qeventloop.cpp:212 #39 0x00007fb75c854257 in QCoreApplication::exec () at kernel/qcoreapplication.cpp:1134 #40 0x00007fb75ce4fcb0 in QGuiApplication::exec () at kernel/qguiapplication.cpp:1332 #41 0x00007fb75d66d3bf in QApplication::exec () at kernel/qapplication.cpp:2707 #42 0x00007fb7647482e6 in kdemain () from /opt/project-neon5/lib/x86_64-linux-gnu/libkdeinit4_kwin.so #43 0x00000000004009af in main () Reproducible: Always
The backtrace is missing debug information for the relevant part. Please provide a backtrace with file and line numbers.
Can't reproduce it anymore -.- More interestingly I have no clue why dbg symbols were broken to begin with... will have a look.
(In reply to comment #2) > Can't reproduce it anymore -.- There was such an issue which I fixed months ago (with a very similar backtrace). Maybe something old got picked?
Nope that was definitely the expected daily build running on a newly created user. If it pops back up I'll get a new backtrace. The debug symbols will be fixed in tomorrow's build. There may have been display plug'n'play involved though. I'll try to recreate it tomorrow with debug symbols available. Perhaps the crash is dependent on the time of day :S At any rate I wasn't doing anything particularly out of the ordinary so I expect it to pop up again.
The crash I fixed was depending on timeouts of kstartup infos. If two timeouts happened in the same minute one got a crash. Extremely random, so time of date might matter ;-)