Bug 700340 – Crash: tag 'track count' should hold value of type 'guint', but value of type '(null)' passed

After an evaluation, GNOME has moved from Bugzilla to GitLab. Learn more about GitLab.
No new issues can be reported in GNOME Bugzilla anymore.
To report an issue in a GNOME project, go to GNOME GitLab.
Do not go to GNOME Gitlab for: Bluefish, Doxygen, GnuCash, GStreamer, java-gnome, LDTP, NetworkManager, Tomboy.

Bug 700340 - Crash: tag 'track count' should hold value of type 'guint', but value of type '(null)' passed


Summary:	Crash: tag 'track count' should hold value of type 'guint', but value of type...


Status:	RESOLVED OBSOLETE

Product:	GStreamer
Classification:	Platform
Component:	gstreamer (core)
Version:	1.0.7
Hardware:	Other Linux

Importance:	Normal major
Target Milestone:	1.5.x
Assigned To:	GStreamer Maintainers
QA Contact:	GStreamer Maintainers

URL:
Whiteboard:

Depends on:
Blocks:

Reported:	2013-05-14 20:28 UTC by Christoph Reiter (lazka)
Modified:	2015-09-16 16:20 UTC

See Also:
GNOME target:	---
GNOME version:	---

Description Christoph Reiter (lazka) 2013-05-14 20:28:40 UTC

1.0.7 on debian unstable, using playbin.

GStreamer-WARNING **: tag 'track count' should hold value of type 'guint', but value of type '(null)' passed

Program received signal SIGSEGV, Segmentation fault.

+ Trace 231946

Thread 140736615933696 (LWP 3063)

#0 g_str_hash
at /tmp/buildd/glib2.0-2.36.1/./glib/ghash.c line 1732
#1 g_hash_table_lookup_node
at /tmp/buildd/glib2.0-2.36.1/./glib/ghash.c line 365
#2 g_hash_table_lookup
at /tmp/buildd/glib2.0-2.36.1/./glib/ghash.c line 1076
#3 gst_tag_lookup
at gsttaglist.c line 444
#4 gst_tag_list_add_value_internal
at gsttaglist.c line 1007
#5 gst_tag_list_copy_foreach
at gsttaglist.c line 1079
#6 gst_structure_foreach
at gststructure.c line 1122
#7 gst_tag_list_insert
at gsttaglist.c line 1108
#8 gst_tag_list_merge
at gsttaglist.c line 1145
#9 gst_tag_demux_send_tag_event
at gsttagdemux.c line 1576
#10 gst_tag_demux_src_getrange
at gsttagdemux.c line 1468
#11 gst_pad_get_range_unchecked
at gstpad.c line 4058
#12 gst_pad_pull_range
at gstpad.c line 4288
#13 gst_base_parse_pull_range
at gstbaseparse.c line 2730
#14 gst_base_parse_scan_frame
at gstbaseparse.c line 2866
#15 gst_base_parse_loop
at gstbaseparse.c line 2970
#16 gst_task_func
at gsttask.c line 316
#17 g_thread_pool_thread_proxy
at /tmp/buildd/glib2.0-2.36.1/./glib/gthreadpool.c line 309
#18 g_thread_proxy
at /tmp/buildd/glib2.0-2.36.1/./glib/gthread.c line 798
#19 start_thread
at pthread_create.c line 311
#20 clone
at ../sysdeps/unix/sysv/linux/x86_64/clone.S line 113

Comment 1 Tim-Philipp Müller 2013-05-14 20:43:17 UTC

Ouch, any chance you could make the file available or e-mail it to me (if smaller than 5MB) or attach the beginning of the file (or just the tag plus a few bytes; needs to be smaller than 900k or so) ?

Comment 2 Christoph Reiter (lazka) 2013-05-14 21:33:09 UTC

It seems it's timing related, happens with random files and not always... this is with PyGObject 3.8.1 btw.

Also got some deadlocks during testing: https://bugzilla.gnome.org/show_bug.cgi?id=700342


Some more warnings I got before it crashed:

GStreamer-WARNING **: tag 'artist ID' should hold value of type 'gchararray', but value of type '(null)' passed

sys:1: Warning: g_value_type_compatible: assertion `G_TYPE_IS_VALUE (src_type)' failed
sys:1: Warning: g_value_copy: assertion `g_value_type_compatible (G_VALUE_TYPE (src_value), G_VALUE_TYPE (dest_value))' failed

(quodlibet:17355): GStreamer-WARNING **: tag 'replaygain album peak' should hold value of type 'gdouble', but value of type '(null)' passed

Comment 3 Tim-Philipp Müller 2013-05-14 23:02:49 UTC

Could you make a new stack trace with the environment variable G_DEBUG=fatal_warnings set?

Comment 4 Sebastian Dröge (slomo) 2013-05-15 16:54:31 UTC

Could also be random memory corruption caused by bug #700342 and everything that goes wrong there.

Comment 5 Christoph Reiter (lazka) 2015-09-13 05:54:33 UTC

Haven't had crashes for some time now -> closing.

Comment 6 GstBlub 2015-09-16 16:20:36 UTC

I run into this issue quite frequently whenever I tear down a pipeline at the "right" time, especially while typefind is in progress.  I submitted a patch in bug #755123 that fixes it for me.