mythfilldatabase segfault just before finished

[Eric Fontaine <ericrfontaine@…>]

Fresh mythtv backend install on parabola (arch) linux x86-64. I'm able to setup backend and record shows and whatnot, but I'm getting this segfault before exit (although I believe most of mythfilldatabase completes, since I can see the guide data in frontend).

Thread 7 "Qt bearer threa" received signal SIGSEGV, Segmentation fault. [Switching to Thread 0x7fffcaffd700 (LWP 7454)] QDBusMetaType::typeToSignature (type=8) at qdbusmetatype.cpp:437

See attached gdb console output and backtrace.

I should note that my computer is on Qt version 5.6.0.

[Eric Fontaine <ericrfontaine@…>]

gdb console

[Eric Fontaine <ericrfontaine@…>]

backtrace

[Gary Buhrmaster <gary.buhrmaster@…>]

Just a FYI to the devs: I cannot (quickly) find the reference, but I have seen something like this reported for other (non-MythTV) apps with Qt 5.6.0. I suspect it is upstream. As I (vaguely) recall, it may have to do with Qt's final cleanup processing (and possibly QNetworkAccessManager?). It might also be timing related (so it may not always crash, either).

Btw, to the OP, with schedules direct, presuming your system is not under-powered, using the --dd-grab-all option is usually recommended (but I suspect it would not change the crash, unless it changes the timing sufficiently so that you get "lucky").

[falk0069@…]

Looks like I have the same problem and the fix is probably to upgrade Qt from 5.6 to 5.7. I'm running Centos7 and they do not have 5.7 available yet, so I'll probably just wait it out since the job completes. It is just annoying to see the daily core [abrt] report appear in my inbox.

Here are a couple of related threads: ​https://stackoverflow.com/questions/37986256/qt-qtest-segmentation-fault-at-end-of-execution

​https://bugreports.qt.io/browse/QTBUG-52988?focusedCommentId=325416&page=com.atlassian.jira.plugin.system.issuetabpanels%3Acomment-tabpanel#comment-325416

[Gary Buhrmaster]

This ticket should (probably) be closed as upstream bug, as it was acknowledged as Qt 5.6 issue, and was resolved (upstream) quite some time ago (including for the OPs OS).