Hide Forgot
Description of problem: Problem was initially reported by me to pgmodeler upstream (submitted on review) - https://github.com/pgmodeler/pgmodeler/issues/777 According to our discuss crash in QT: Program received signal SIGSEGV, Segmentation fault. 0x00007fffe259ef08 in QXcbWindow::setParent (this=0x555555e891d0, parent=0x0) at qxcbwindow.cpp:1499 1499 xcb_parent_id = xcbScreen()->root(); Missing separate debuginfos, use: dnf debuginfo-install bzip2-libs-1.0.6-17.fc23.x86_64 flac-libs-1.3.1-5.fc23.x86_64 libgcc-5.1.1-4.fc23.x86_64 libstdc++-5.1.1-4.fc23.x86_64 postgresql95-libs-9.5-beta2_1PGDG.f23.x86_64 (gdb) bt #0 0x00007fffe259ef08 in QXcbWindow::setParent(QPlatformWindow const*) (this=0x555555e891d0, parent=0x0) at qxcbwindow.cpp:1499 #1 0x00007ffff53e78bb in QWindow::setParent(QWindow*) (this=0x555557359750, parent=parent@entry=0x55555702f450) at kernel/qwindow.cpp:614 #2 0x00007ffff5bc25d5 in QWidgetPrivate::setParent_sys(QWidget*, QFlags<Qt::WindowType>) (this=this@entry=0x555557008d60, newparent=newparent@entry=0x555556e5f780, f=...) at kernel/qwidget.cpp:10543 #3 0x00007ffff5bcdfda in QWidget::setParent(QWidget*, QFlags<Qt::WindowType>) (this=this@entry=0x555555c96e30, parent=parent@entry=0x555556e5f780, f=...) at kernel/qwidget.cpp:10398 #4 0x00007ffff5bcf0cc in QWidget::setParent(QWidget*) (this=this@entry=0x555555c96e30, parent=parent@entry=0x555556e5f780) at kernel/qwidget.cpp:10334 #5 0x00007ffff5cb511c in QDialogButtonBox::addButton(QAbstractButton*, QDialogButtonBox::ButtonRole) (this=0x555556e5f780, button=0x555555c96e30, role=role@entry=QDialogButtonBox::AcceptRole) at widgets/qdialogbuttonbox.cpp:761 And by documentation (http://doc.qt.io/qt-5/qfiledialog.html) parent should be optional.
Created attachment 1105603 [details] Test case for bug Added testcase for this bug, it also appears to happen when the parent pointer is set. It appears to be intermittent, it may take several attempts before it crashes. Seems more likely to crash if dialog is opened shortly after the application is launched. The backtrace produced by this test case is similar to the one above. The segault occurs as xcbScreen() returns a null pointer.
*** Bug 1291019 has been marked as a duplicate of this bug. ***
I've seen the same with kate and my own applications.
Another user experienced a similar problem: I was looking for required details to configire WPA2 Enterprise security, reporter: libreport-2.6.3 backtrace_rating: 4 cmdline: /usr/bin/kde5-nm-connection-editor crash_function: QXcbWindow::setParent executable: /usr/bin/kde5-nm-connection-editor global_pid: 28921 kernel: 4.2.7-300.fc23.x86_64 package: plasma-nm-5.5.0-2.fc23 reason: kde5-nm-connection-editor killed by SIGSEGV runlevel: N 5 type: CCpp uid: 1000
Created attachment 1107787 [details] File: backtrace
I *suspect* this may be fixed in qt-5.6 (I get warnings, but no crashes running kate using 5.6-beta), tentatively marking it so.
1293098, 1293494, 1293494, 1290993, 1291426, 1293758, 1292477, 1292355, 1291942, 1285615, 1290988 are all duplicates of this one The bug itself is caused by the assumption, that a window always has a screen - which turned out to be false, IIRC, thus why upstream fixed this with a gap-stopper in the 5.5 branch: 37b7c5164c830458ea833d3757ca94cb7bebcb3a The 5.6 branch has a proper fix that reverts all the gap-stoppers: a094af001795c9651b299d700a992150d1aba33a This bug is pretty annoying and will affect every application that uses QDialogButtonBox, so it would be nice to have it fixed before 5.6 is merged (i.e. before February or whenever afterwards it will happen)
Created attachment 1110768 [details] gap-stopper from the 5.5 branch 37b7c5164c830458ea833d3757ca94cb7bebcb3a
Created attachment 1110769 [details] spec for gap-stopper
I have been using the gap-stopper for several days and it seems to do the trick for now. I posted the required files, so anyone interested can fix their system until the maintainer(|s) ha(s|ve) time to decide how to proceed. The relevant qt commit is http://code.qt.io/cgit/qt/qtbase.git/commit/?h=5.5&id=37b7c5164c830458ea833d3757ca94cb7bebcb3a
Hrm, that commit was supposed to be included in Qt 5.5.1 release, I'll dig to try to find out why/how that didn't get included
OK, I've found that the experimental patch as part of work from bug #1083664 which is a backport from https://codereview.qt-project.org/#/c/138201/ here, http://pkgs.fedoraproject.org/cgit/qt5-qtbase.git/tree/138201.patch?h=f23 removes that check here, among others. I suppose we could consider adjusting it as done in comment #10
*** Bug 1295086 has been marked as a duplicate of this bug. ***
*** Bug 1295074 has been marked as a duplicate of this bug. ***
*** Bug 1294985 has been marked as a duplicate of this bug. ***
*** Bug 1290988 has been marked as a duplicate of this bug. ***
*** ConfidentialBug 1294351 has been marked as a duplicate of this bug. ***
*** Bug 1291712 has been marked as a duplicate of this bug. ***
*** Bug 1293494 has been marked as a duplicate of this bug. ***
*** Bug 1293758 has been marked as a duplicate of this bug. ***
*** Bug 1293751 has been marked as a duplicate of this bug. ***
*** Bug 1292355 has been marked as a duplicate of this bug. ***
*** Bug 1293098 has been marked as a duplicate of this bug. ***
*** Bug 1290993 has been marked as a duplicate of this bug. ***
*** Bug 1292477 has been marked as a duplicate of this bug. ***
*** Bug 1291942 has been marked as a duplicate of this bug. ***
*** Bug 1285615 has been marked as a duplicate of this bug. ***
Another user experienced a similar problem: pasted text into KWrite, clicked "Save As..." button. problem is repeateble on my machine issue also occurs when siply opening kwrite and immediately ckicking "Save As..." or using Ctrl+Shift+s reporter: libreport-2.6.3 backtrace_rating: 4 cmdline: /usr/bin/kwrite crash_function: QXcbWindow::setParent executable: /usr/bin/kwrite global_pid: 4726 kernel: 4.2.8-200.fc22.x86_64 package: kwrite-15.04.3-1.fc22 reason: kwrite killed by SIGSEGV runlevel: N 5 type: CCpp uid: 1000
Rex if I understand correctly fix incorporated. Could you please make update of package?
That's not a fix (at best an only slightly tested workaround). Upstream is working on a different approach, see https://bugreports.qt.io/browse/QTBUG-50081 and https://codereview.qt-project.org/#/c/144896/
Another user experienced a similar problem: was trying to save a file I had just created by pressing Ctrl+S reporter: libreport-2.6.3 backtrace_rating: 4 cmdline: /usr/bin/kate /tmp/fz3temp-1/party_crystal.c crash_function: QXcbWindow::setParent executable: /usr/bin/kate global_pid: 3844 kernel: 4.2.8-200.fc22.i686+PAE package: kate-15.04.3-1.fc22 reason: kate killed by SIGSEGV runlevel: N 5 type: CCpp uid: 1000
Another user experienced a similar problem: using codeblocks in debugger, I stopped debugger without closing files and Cpp11 string objects. I may have contaminated the heap. then opened kwrite which crashed. reporter: libreport-2.6.3 backtrace_rating: 4 cmdline: /usr/bin/kwrite crash_function: QXcbWindow::setParent executable: /usr/bin/kwrite global_pid: 2306 kernel: 4.2.8-200.fc22.x86_64 package: kwrite-15.04.3-1.fc22 reason: kwrite killed by SIGSEGV runlevel: N 5 type: CCpp uid: 1000
Another user experienced a similar problem: Update on last post about codeblocks and heap damage: 1) Rebooted system, launched codeblocks, coppied about 20-lines of code from codeblocks editor, saved to disk, kwrite crashed. There is no possibillity of system heap damage from this action. 2) while problem reporting system was reporting, opened an unrelated file using dolphin into kwrite. saved this unrelated file with a different file name. kwrite worked as expected (no crash). 3) conclusion, something to do with unprintable characters in codeblocks. reporter: libreport-2.6.3 backtrace_rating: 4 cmdline: /usr/bin/kwrite crash_function: QXcbWindow::setParent executable: /usr/bin/kwrite global_pid: 4732 kernel: 4.2.8-200.fc22.x86_64 package: kwrite-15.04.3-1.fc22 reason: kwrite killed by SIGSEGV runlevel: N 5 type: CCpp uid: 1000
While a fix is searched what should I downgrade (in my fedora 22) to go back in a state where the open/save dialogs works?
Another user experienced a similar problem: just click on add source files and app crash with this backtrace reporter: libreport-2.6.3 backtrace_rating: 4 cmdline: mkvtoolnix-gui crash_function: QXcbWindow::setParent executable: /usr/bin/mkvtoolnix-gui global_pid: 2734 kernel: 4.2.8-300.fc23.x86_64 package: mkvtoolnix-gui-8.6.1-1.fc23 reason: mkvtoolnix-gui killed by SIGSEGV runlevel: N 5 type: CCpp uid: 500
*** Bug 1295966 has been marked as a duplicate of this bug. ***
qt5-qtbase-5.5.1-11.fc23 has been submitted as an update to Fedora 23. https://bodhi.fedoraproject.org/updates/FEDORA-2016-4899ebd424
qt5-qtbase-5.5.1-11.fc22 has been submitted as an update to Fedora 22. https://bodhi.fedoraproject.org/updates/FEDORA-2016-e81f52ffe7
Another user experienced a similar problem: Baloo file is disabled via console. Went to its systemsettings module and casually decided to remove a lonely entry in the exceptions list. The Enable File Search box was automatically checked and then the program crashed. reporter: libreport-2.6.3 backtrace_rating: 4 cmdline: /usr/bin/systemsettings5 crash_function: QXcbWindow::setParent executable: /usr/bin/systemsettings5 global_pid: 6924 kernel: 4.2.8-300.fc23.x86_64 package: plasma-systemsettings-5.5.1-1.fc23 reason: systemsettings5 killed by SIGSEGV runlevel: N 5 type: CCpp uid: 1000
qt5-qtbase-5.5.1-11.fc23 has been pushed to the Fedora 23 testing repository. If problems still persist, please make note of it in this bug report. See https://fedoraproject.org/wiki/QA:Updates_Testing for instructions on how to install test updates. You can provide feedback for this update here: https://bodhi.fedoraproject.org/updates/FEDORA-2016-4899ebd424
qt5-qtbase-5.5.1-11.fc22 has been pushed to the Fedora 22 testing repository. If problems still persist, please make note of it in this bug report. See https://fedoraproject.org/wiki/QA:Updates_Testing for instructions on how to install test updates. You can provide feedback for this update here: https://bodhi.fedoraproject.org/updates/FEDORA-2016-e81f52ffe7
the update seems to work for me (f22)
No I see a lots of 0x55a9cc007880 void QWindowPrivate::setTopLevelScreen(QScreen*, bool) ( QScreen(0x55a9cb464530) ): Attempt to set a screen on a child window. 0x55a9cc007ce0 void QWindowPrivate::setTopLevelScreen(QScreen*, bool) ( QScreen(0x55a9cb464530) ): Attempt to set a screen on a child window. 0x55a9cc008140 void QWindowPrivate::setTopLevelScreen(QScreen*, bool) ( QScreen(0x55a9cb464530) ): Attempt to set a screen on a child window. 0x55a9cc008500 void QWindowPrivate::setTopLevelScreen(QScreen*, bool) ( QScreen(0x55a9cb464530) ): Attempt to set a screen on a child window. but it does not segfaults. Thank you.
qt5-qtbase-5.5.1-11.fc23 has been pushed to the Fedora 23 stable repository. If problems still persist, please make note of it in this bug report.
*** Bug 1296957 has been marked as a duplicate of this bug. ***
Another user experienced a similar problem: 1) Start kwrite 2) Do not even type anything 3) Immediately proceed to save 4) Crash (It is the save that does it) reporter: libreport-2.6.3 backtrace_rating: 4 cmdline: /usr/bin/kwrite crash_function: QXcbWindow::setParent executable: /usr/bin/kwrite global_pid: 15580 kernel: 4.2.8-200.fc22.x86_64 package: kwrite-15.04.3-1.fc22 reason: kwrite killed by SIGSEGV runlevel: N 5 type: CCpp uid: 1000
qt5-qtbase-5.5.1-11.fc22 has been pushed to the Fedora 22 stable repository. If problems still persist, please make note of it in this bug report.