GNOME Bugzilla – Bug 691240
frei0r-filter-opencvfacedetect: crash when adding filter to a clip in pitivi
Last modified: 2013-03-16 12:37:22 UTC
This is what happens with 0.10.x (not tested in 1.x due to the brokenness of it at the moment): (gdb) set pagination 0 (gdb) run /usr/bin/pitivi Starting program: /usr/bin/python /usr/bin/pitivi [Thread debugging using libthread_db enabled] Using host libthread_db library "/lib/libthread_db.so.1". ** Message: pygobject_register_sinkfunc is deprecated (GstObject) [New Thread 0xb6427b40 (LWP 12967)] [Thread 0xb6427b40 (LWP 12967) exited] [New Thread 0xb6427b40 (LWP 12968)] [Thread 0xb6427b40 (LWP 12968) exited] Detaching after fork from child process 12969. /usr/share/pitivi/python/pitivi/application.py:255: Warning: g_object_set_qdata: assertion `G_IS_OBJECT (object)' failed self.mainloop.run() [New Thread 0xb6427b40 (LWP 12971)] /usr/share/pitivi/python/pitivi/ui/filechooserpreview.py:72: Warning: g_object_set_qdata: assertion `G_IS_OBJECT (object)' failed self.__videosink = self.player.get_property("video-sink") [New Thread 0xaff3cb40 (LWP 12972)] [New Thread 0xaf5ffb40 (LWP 12973)] [New Thread 0xaebffb40 (LWP 12974)] [New Thread 0xae3feb40 (LWP 12975)] [New Thread 0xadbfdb40 (LWP 12976)] [New Thread 0xad3fcb40 (LWP 12977)] [Thread 0xb6427b40 (LWP 12971) exited] [Thread 0xae3feb40 (LWP 12975) exited] [Thread 0xaf5ffb40 (LWP 12973) exited] [Thread 0xadbfdb40 (LWP 12976) exited] [Thread 0xad3fcb40 (LWP 12977) exited] [Thread 0xaebffb40 (LWP 12974) exited] [New Thread 0xaebffb40 (LWP 12978)] [New Thread 0xad3fcb40 (LWP 12979)] [New Thread 0xadbfdb40 (LWP 12980)] [New Thread 0xaf5ffb40 (LWP 12981)] [New Thread 0xb6427b40 (LWP 12982)] [New Thread 0xae3feb40 (LWP 12983)] [Thread 0xb6427b40 (LWP 12982) exited] [Thread 0xae3feb40 (LWP 12983) exited] [Thread 0xaebffb40 (LWP 12978) exited] [Thread 0xaf5ffb40 (LWP 12981) exited] [Thread 0xaff3cb40 (LWP 12972) exited] [Thread 0xad3fcb40 (LWP 12979) exited] [New Thread 0xad3fcb40 (LWP 12985)] [New Thread 0xaf5ffb40 (LWP 12986)] [New Thread 0xaff3cb40 (LWP 12987)] [New Thread 0xae3feb40 (LWP 12988)] [New Thread 0xac2ffb40 (LWP 12989)] [New Thread 0xabafeb40 (LWP 12990)] [New Thread 0xa6509b40 (LWP 12996)] [Thread 0xac2ffb40 (LWP 12989) exited] [Thread 0xabafeb40 (LWP 12990) exited] [Thread 0xae3feb40 (LWP 12988) exited] [Thread 0xaff3cb40 (LWP 12987) exited] [Thread 0xaf5ffb40 (LWP 12986) exited] [Thread 0xad3fcb40 (LWP 12985) exited] [Thread 0xadbfdb40 (LWP 12980) exited] [New Thread 0xadbfdb40 (LWP 12998)] [New Thread 0xad3fcb40 (LWP 12999)] [New Thread 0xaf5ffb40 (LWP 13000)] [New Thread 0xaff3cb40 (LWP 13001)] [New Thread 0xac2ffb40 (LWP 13002)] [New Thread 0xabafeb40 (LWP 13003)] [New Thread 0xab2fdb40 (LWP 13004)] [New Thread 0xaa9aab40 (LWP 13005)] [New Thread 0xa9aadb40 (LWP 13006)] [New Thread 0xa8f27b40 (LWP 13007)] [New Thread 0xa8726b40 (LWP 13008)] [New Thread 0xa7f25b40 (LWP 13009)] [Thread 0xaff3cb40 (LWP 13001) exited] [New Thread 0xaff3cb40 (LWP 13010)] [Thread 0xa7f25b40 (LWP 13009) exited] Program received signal SIGSEGV, Segmentation fault.
+ Trace 231340
Thread 2915031872 (LWP 12998)
Thread 24 (Thread 0xadbfdb40 (LWP 12998))
1) Please try to come up with a gst-launch pipeline to reproduce the issue 2) If you must file bugs against 0.10, please at least try with 0.10 from git.
As for using 0.10 git, the 0.10 series is not exactly the most changing thing out there these days, with 1.x around... Full repro instructions with pitivi are: - start pitivi 0.15 provided by your friendly distro - drag a video file (ex: big buck bunny) to the media library, then to the timeline - click the clip on the timeline - in the effects library, drag "Opencvfacedetect" into the clip's effects (the "Clip properties" tab) Boom, segfault. But then I understand it's not as easy as just executing gst launch and that's annoying, but that's the best I can provide for the time being. I'm sorry about not being able to provide better quality info, as gst-launch is black magic to me.
Sorry, but no one is going to look into 0.10 bugs any more. I have tried a static pipeline in 1.0 such as gst-launch-1.0 v4l2src ! videoconvert ! frei0r-filter-opencvfacedetect ! videoconvert ! ximagesink sync=false and I have also added the frei0r-filter-opencvfacedetect element to the gst-plugins-base/tests/icles/test-effect-switch test to see how it behaves with dynamically switching stuff, and didn't notice any problems there. Dynamic switching is something that has changed in 1.0. So let's just close this for now, it's not really useful to keep around. The stack trace also looks a bit wrong. If it's still an issue with 1.0, someone will run into it sooner or later and file a new bug.