Version-Release number of selected component: trojita-0.5-7.fc23 Additional info: reporter: libreport-2.6.2 backtrace_rating: 4 cmdline: /bin/trojita crash_function: QXcbConnection::internAtom executable: /usr/bin/trojita global_pid: 3296 kernel: 4.2.0-0.rc6.git0.2.fc23.x86_64 runlevel: N 5 type: CCpp uid: 1000 Truncated backtrace: Thread no. 1 (10 frames) #0 QXcbConnection::internAtom at qxcbconnection.cpp:1763 #1 QXcbSystemTrayTracker::create at qxcbsystemtraytracker.cpp:63 #2 QXcbConnection::systemTrayTracker at qxcbconnection.cpp:2048 #3 systemTrayTracker at qxcbnativeinterface.cpp:110 #4 QXcbNativeInterface::nativeResourceForScreen at qxcbnativeinterface.cpp:288 #5 locateSystemTray at util/qsystemtrayicon_x11.cpp:60 #6 QSystemTrayIconPrivate::isSystemTrayAvailable_sys at util/qsystemtrayicon_x11.cpp:352 #7 QSystemTrayIcon::isSystemTrayAvailable at util/qsystemtrayicon.cpp:346 #8 KStatusNotifierItemPrivate::setLegacySystemTrayEnabled at ../../src/kstatusnotifieritem.cpp:801 #9 KStatusNotifierItemPrivate::setLegacyMode at ../../src/kstatusnotifieritem.cpp:773
Created attachment 1066526 [details] File: backtrace
Created attachment 1066527 [details] File: cgroup
Created attachment 1066528 [details] File: core_backtrace
Created attachment 1066529 [details] File: dso_list
Created attachment 1066530 [details] File: environ
Created attachment 1066531 [details] File: limits
Created attachment 1066532 [details] File: maps
Created attachment 1066533 [details] File: mountinfo
Created attachment 1066534 [details] File: namespaces
Created attachment 1066535 [details] File: open_fds
Created attachment 1066536 [details] File: proc_pid_status
Created attachment 1066537 [details] File: var_log_messages
Backtrace indicates some bug in Qt5, so reassigning.
Possible duplication of bug #1255902.
*** Bug 1259470 has been marked as a duplicate of this bug. ***
FTBFS in rawhide, I've to fix that first.
*** Bug 1392511 has been marked as a duplicate of this bug. ***
Maybe it get's better with any upcoming Qt5.7+ …
*** Bug 1256478 has been marked as a duplicate of this bug. ***
It seems the offending method is not thread safe. What about checking for null pointers and return some invalid atom object to let applications not crash? Also, we should ask the maintainers of XCB for some help. xcb_atom_t QXcbConnection::internAtom(const char *name) { if (!name || *name == 0) return XCB_NONE; xcb_intern_atom_cookie_t cookie = xcb_intern_atom(xcb_connection(), false, strlen(name), name); xcb_intern_atom_reply_t *reply = xcb_intern_atom_reply(xcb_connection(), cookie, 0); int atom = reply->atom; free(reply); return atom; }
Similar problem has been detected: it got stuck after trying to login at http://zbranekvalitne.cz/forum/ucp.php?mode=login reporter: libreport-2.8.0 backtrace_rating: 4 cmdline: /usr/bin/qupzilla executable: /usr/bin/qupzilla global_pid: 30095 kernel: 4.8.6-300.fc25.x86_64 package: qupzilla-2.0.2-1.fc25 pkg_fingerprint: 4089 D8F2 FDB1 9C98 pkg_vendor: Fedora Project reason: qupzilla killed by SIGABRT runlevel: N 5 type: CCpp uid: 1000
um, it's nice to get redirected here after a crash of a different application two Fedora versions later :-)
Because it's (nearly) always the same backtrace you report about a bug in the core of Qt5. See explanation in comment #20.
Re-assigning back to Qt, this is almost certainly not a libxcb issue
i tested the trojita on machine with f24 and cannot reproduce this issue. Could someone please add an reproducer for that so i can check? Thanks
I'm closing that bugzilla INSUFFICIENT_DATA because there's no testcase to reproduce this issue. Feel free to reopen it once you find some way how to reproduce it. thanks
(In reply to Ngo Than from comment #25) > Could someone please add an reproducer for that so i can check? sorry, I just don't know what triggered the problem the page mentioned in comment#21 doesn't seem to reproduce the Qupzilla problem now