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 693019 - tracker-extract crashed with SIGSEGV in type_check_is_value_type_U()
tracker-extract crashed with SIGSEGV in type_check_is_value_type_U()
Status: RESOLVED INCOMPLETE
Product: tracker
Classification: Core
Component: Extractor
0.14.x
Other Linux
: Normal major
: ---
Assigned To: tracker-extractor
Depends on:
Blocks:
 
 
Reported: 2013-02-01 16:12 UTC by melchiaros
Modified: 2013-06-09 11:40 UTC
See Also:
GNOME target: ---
GNOME version: ---



Description melchiaros 2013-02-01 16:12:36 UTC
This is an upstream report of tracker-extract crash, which occures the first time with
package tracker-extract 0.14.1 in Ubuntu12.10 and is still actual up to package tracker-extract 0.14.4
in Ubuntu13.10 development branch. During time launchpad has counted 2 who are
affected by stacktrace identification.

The original report can be found at:

https://bugs.launchpad.net/ubuntu/+source/tracker/+bug/1052348

and list all dublicates over different versions.

A stactrace is available only for the 0.14.1 version and will be attached here.

Unfortunally I am by myself not able to hit the crash on any version, so a
stacktrace from myself is not available(I only bring this upstream). Anyway,
the stacktrace top is listed as still the same for any affected version.

Also no one of the affected people is able to give a valid reproducing
procedure as can be seen by the comments in the dublicates(crash randomly).


The header of 

https://bugs.launchpad.net/ubuntu/+source/tracker/+bug/1052348

will be copied here and the available stacktrace will be attached.

If there is need to test any further please contact the 2 affected people
automatically by simple commenting on the launchpad report.

Thanks

---------------------------------------------(header of launchpad bug):

tracker-extract crashed while indexing my homedir.

ProblemType: Crash
DistroRelease: Ubuntu 12.10
Package: tracker-extract 0.14.1-1ubuntu5
ProcVersionSignature: Ubuntu 3.5.0-15.20-generic 3.5.4
Uname: Linux 3.5.0-15-generic i686
ApportVersion: 2.5.2-0ubuntu1
Architecture: i386
Date: Tue Sep 18 09:39:29 2012
EcryptfsInUse: Yes
ExecutablePath: /usr/lib/tracker/tracker-extract
ProcCmdline: /usr/lib/tracker/tracker-extract
ProcEnviron:
 SHELL=/bin/bash
 LC_PAPER=fr_FR.UTF-8
 PATH=(custom, no user)
 LANG=en_US.UTF-8
 LC_COLLATE=C
SegvAnalysis:
 Segfault happened at: 0xb7464150 <g_type_check_value+64>: testb $0x8,0xe(%esi)
 PC (0xb7464150) ok
 source "$0x8" ok
 destination "0xe(%esi)" (0x75757582) not located in a known VMA region (needed writable region)!
SegvReason: writing unknown VMA
Signal: 11
SourcePackage: tracker
StacktraceTop:
 type_check_is_value_type_U (type=1970632053) at /build/buildd/glib2.0-2.33.12/./gobject/gtype.c:4096
 g_type_check_value (value=value@entry=0xadde1144) at /build/buildd/glib2.0-2.33.12/./gobject/gtype.c:4138
 gst_value_init_and_copy (dest=dest@entry=0xbffc42ac, src=0xadde1144) at gstvalue.c:4007
 gst_structure_copy (structure=structure@entry=0x99a3f00) at gststructure.c:312
 gst_tag_list_copy (list=0x99a3f00) at gsttaglist.c:985
Title: tracker-extract crashed with SIGSEGV in type_check_is_value_type_U()
UpgradeStatus: Upgraded to quantal on 2011-09-06 (377 days ago)
UserGroups: adm admin audio cdrom dialout dip floppy fuse kvm libvirtd lpadmin plugdev pulse pulse-access scanner staff video
Comment 1 melchiaros 2013-02-01 16:13:17 UTC


  • #0 type_check_is_value_type_U
    at /build/buildd/glib2.0-2.33.12/./gobject/gtype.c line 4096
  • #1 g_type_check_value
    at /build/buildd/glib2.0-2.33.12/./gobject/gtype.c line 4138
  • #2 gst_value_init_and_copy
    at gstvalue.c line 4007
  • #3 gst_structure_copy
    at gststructure.c line 312
  • #4 gst_tag_list_copy
    at gsttaglist.c line 985
  • #5 _g_type_boxed_copy
    at /build/buildd/glib2.0-2.33.12/./gobject/gtype.c line 4237
  • #6 boxed_proxy_collect_value
    at /build/buildd/glib2.0-2.33.12/./gobject/gboxed.c line 229
  • #7 gst_structure_id_set_valist_internal
    at gststructure.c line 642
  • #8 gst_structure_id_set
    at gststructure.c line 676
  • #9 collect_stream_information
    at gstdiscoverer.c line 605
  • #10 find_stream_for_node
    at gstdiscoverer.c line 845
  • #11 parse_stream_topology
    at gstdiscoverer.c line 894
  • #12 parse_stream_topology
    at gstdiscoverer.c line 931
  • #13 parse_stream_topology
    at gstdiscoverer.c line 931
  • #14 parse_stream_topology
    at gstdiscoverer.c line 991
  • #15 discoverer_collect
    at gstdiscoverer.c line 1071
  • #16 gst_discoverer_discover_uri
    at gstdiscoverer.c line 1579
  • #17 discoverer_init_and_run
    at tracker-extract-gstreamer.c line 1367
  • #18 tracker_extract_gstreamer
    at tracker-extract-gstreamer.c line 1987
  • #19 tracker_extract_get_metadata
    at tracker-extract-gstreamer.c line 2083
  • #20 get_file_metadata
    at tracker-extract.c line 375
  • #21 get_metadata
    at tracker-extract.c line 588
  • #22 dispatch_task_cb
    at tracker-extract.c line 721
  • #23 g_idle_dispatch
    at /build/buildd/glib2.0-2.33.12/./glib/gmain.c line 4806
  • #24 g_main_dispatch
    at /build/buildd/glib2.0-2.33.12/./glib/gmain.c line 2715
  • #25 g_main_context_dispatch
    at /build/buildd/glib2.0-2.33.12/./glib/gmain.c line 3219
  • #26 g_main_context_iterate
    at /build/buildd/glib2.0-2.33.12/./glib/gmain.c line 3290
  • #27 g_main_loop_run
    at /build/buildd/glib2.0-2.33.12/./glib/gmain.c line 3484
  • #28 main
    at tracker-main.c line 451

Comment 2 Martyn Russell 2013-02-01 17:19:43 UTC
Hello, thanks for the details here. I've commented on the original Launchpad bug. Any chance master or a 0.15.x release could be tried here? We use the new GStreamer 1.x API there.

It certainly looks like it could be a GStreamer bug as far as I can see. But I've not tried reproducing it yet.
Comment 3 melchiaros 2013-02-01 17:38:00 UTC
That looks bad. Feature freez on Ubuntu13.04 has already happen, and a search on 

http://packages.debian.org/search?keywords=tracker&searchon=names&suite=all&section=all

give that there are also not appropriate packages from debian which could be incorporated temorary for testing(gdebi).

Compiling from source and installation is what I do not want to do.(To much reinstallations in the last time).

I have just seen your comment on launchpad.

When you can get the affected file out of this people and the gnome project is providing a live cd for your development branches(GNOME3.7.x right?) for sure I could try to reproduce in a VM, but well, that is so quick that you can do it your own I guess.
Comment 4 melchiaros 2013-02-01 17:39:20 UTC
Anyway, thank you for your quick reaction so far.
Comment 5 Martyn Russell 2013-02-02 11:15:30 UTC
The alternative is to upload the file causing the problem so I can test it - is that possible?
Comment 6 melchiaros 2013-02-02 11:19:25 UTC
I am not sure if I understand correct.

Well, if you mean that I should upload it here when one of the affected people provide:  Yes, sure I can do.(Hopefully-> if I see the mail). I can also ask them to subscibe here directly.
Comment 7 melchiaros 2013-02-02 11:23:54 UTC
You know that you can subscibe yourself to the downstream ticket and get a automatically generated message of all notifications:

(right upper corner -> a bit down)

May be that work directly:

https://bugs.launchpad.net/ubuntu/+source/tracker/+bug/1052348/+subscribe
Comment 8 Martyn Russell 2013-02-02 12:59:20 UTC
The stack trace mentions this file:

file:///home/bonnaud/theses/V.Girondel/soutenance/P_HH_seg_CV0.mpeg

If it's consistently breaking with this file, then making it available to test with would help me try to fix this bug without the need for you to build the latest master/etc branch.
Comment 9 melchiaros 2013-02-02 13:22:47 UTC
Aha,

have not seen that the stacktrace gives info of the location of the file.

I have ask both people(have given the second a description how to find his/her specific one with apport .crash file).

May be they react.
Comment 10 melchiaros 2013-02-09 09:10:03 UTC
Hi Martyn,

Laurent Bonnoud has attached the affected file.

You know it is a .mpeg as expected not small.

The 13MB will not be accepted by bugzilla.

A tar.bz2 copression gives also only a result of 11MB.

Could you download it from downstream?

greetings 

melchiaros
Comment 11 melchiaros 2013-02-09 09:11:22 UTC
BTW

a download of the file to my system gives no crash(also compressing and doing arround a bit). -> tracker 0.14.1
Comment 12 Martyn Russell 2013-02-12 08:42:44 UTC
(In reply to comment #10)
> Hi Martyn,

Hello,
 
> Laurent Bonnoud has attached the affected file.
> 
> You know it is a .mpeg as expected not small.
> 
> The 13MB will not be accepted by bugzilla.
> 
> A tar.bz2 copression gives also only a result of 11MB.
> 
> Could you download it from downstream?

Sure, is there a link to download from ?
Comment 13 melchiaros 2013-02-14 10:40:54 UTC
Yes, here it is:

https://launchpadlibrarian.net/130338741/P_HH_seg_CV0.mpeg
Comment 14 Martyn Russell 2013-02-26 10:11:15 UTC
Thanks for this.

I just tested here with master and it seems fine:

<snip>
$ /usr/libexec/tracker-extract -v 3 -m gstreamer -f ~/Downloads/P_HH_seg_CV0.mpegInitializing tracker-extract...
Tracker-Message: Setting up monitor for changes to config file:'/home/martyn/.config/tracker/tracker-extract.cfg'
Locale 'TRACKER_LOCALE_LANGUAGE' was set to 'en_GB.UTF-8'
Locale 'TRACKER_LOCALE_TIME' was set to 'en_GB.UTF-8'
Locale 'TRACKER_LOCALE_COLLATE' was set to 'en_GB.UTF-8'
Locale 'TRACKER_LOCALE_NUMERIC' was set to 'en_GB.UTF-8'
Locale 'TRACKER_LOCALE_MONETARY' was set to 'en_GB.UTF-8'
Initializing Storage...
Mount monitors set up for to watch for added, removed and pre-unmounts...
No mounts found to iterate
Setting scheduler policy to SCHED_IDLE
Setting priority nice level to 19
Loading extractor rules... (/usr/share/tracker/extract-rules)
  Loaded rule '10-abw.rule'
  Loaded rule '10-dvi.rule'
  Loaded rule '10-epub.rule'
  Loaded rule '10-flac.rule'
  Loaded rule '10-gif.rule'
  Loaded rule '10-html.rule'
  Loaded rule '10-ico.rule'
  Loaded rule '10-jpeg.rule'
  Loaded rule '10-mp3.rule'
  Loaded rule '10-msoffice.rule'
  Loaded rule '10-oasis.rule'
  Loaded rule '10-pdf.rule'
  Loaded rule '10-png.rule'
  Loaded rule '10-ps.rule'
  Loaded rule '10-svg.rule'
  Loaded rule '10-tiff.rule'
  Loaded rule '10-vorbis.rule'
  Loaded rule '10-xmp.rule'
  Loaded rule '11-msoffice-xml.rule'
  Loaded rule '15-gstreamer-guess.rule'
  Loaded rule '15-playlist.rule'
  Loaded rule '90-gstreamer-generic.rule'
  Loaded rule '90-text-generic.rule'
  Loaded rule '93-mplayer-generic.rule'
  Loaded rule '93-totem-generic.rule'
Extractor rules loaded
Setting memory limitations: total is 8.4 GB, minimum is 256 MB, recommended is ~1 GB
  Virtual/Heap set to 4.2 GB (50% of total or MAXLONG)
Guessing mime type as '(null)'
Module used 'libextract-gstreamer.so' (due to --force-module='libextract-gstreamer.so')
Extracting...
  Using /usr/lib/tracker-0.16/extract-modules/libextract-gstreamer.so...
GStreamer backend in use:
  Discoverer/GUPnP-DLNA
Retrieving geolocation metadata...
Processing media art: artist:'', title:'', type:'video', uri:'file:///home/martyn/Downloads/P_HH_seg_CV0.mpeg'. Buffer is 0 bytes, mime:'(null)'
Album art path could not be obtained, not processing any further
Done (6 items)

SPARQL pre-update:
--
--

SPARQL item:
--
 a nmm:Video ;
	 nfo:duration 21 ;
	 nfo:frameRate 30 ;
	 nfo:width 640 ;
	 nfo:height 480 ;
	 nfo:aspectRatio 1 .
--

SPARQL where clause:
--
--

SPARQL post-update:
--
--
</snip>

So it looks like it may have been fixed. Are you able to check with master too in case I am missing something? Thanks.
Comment 15 melchiaros 2013-02-28 13:23:31 UTC
Not really.

The earliest timepoint is when the development of Ubuntu13.10 will start, which will import the latest stable tracker automatically.

So, May 2013
Comment 16 Martyn Russell 2013-02-28 14:35:13 UTC
(In reply to comment #15)
> Not really.
> 
> The earliest timepoint is when the development of Ubuntu13.10 will start, which
> will import the latest stable tracker automatically.
> 
> So, May 2013

OK, but I can't reproduce this, so I am keen to know if it was fixed in the mean time or I am missing something. Which 0.14.1 is quite old now too.
Comment 17 Tobias Mueller 2013-06-09 11:40:07 UTC
Closing this bug report as no further information has been provided. Please feel free to reopen this bug if you can provide the information asked for.
Thanks!