Segmentation fault in gtkpod if ipod touch connected; triggered by g_type_check_value_holds
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
libgpod (Ubuntu) |
New
|
Undecided
|
Unassigned |
Bug Description
Today I tried to sync my iPod Touch using gtkpod, and discovered that the program segfaults when the device is plugged in -- whether I start gtkpod while it is plugged in or start it first and plug the device in afterwards.
Here's the relevant output from valgrind:
==6295== Process terminating with default action of signal 11 (SIGSEGV)
==6295== Access not within mapped region at address 0x1F00000036
==6295== at 0x69D6C3D: g_type_
==6295== by 0x75017A7: ??? (in /usr/lib/
==6295== by 0x75064D2: itdb_parse (in /usr/lib/
==6295== by 0x4E57679: gp_import_itdb (in /usr/lib/
==6295== by 0x4E57E49: gp_load_ipod (in /usr/lib/
==6295== by 0x4E6FCF1: ad_timeout_cb (in /usr/lib/
==6295== by 0x5E9DEC7: ??? (in /usr/lib/
==6295== by 0x6C3E702: ??? (in /lib/x86_
==6295== by 0x6C3DCE4: g_main_
==6295== by 0x6C3E047: ??? (in /lib/x86_
==6295== by 0x6C3E309: g_main_loop_run (in /lib/x86_
==6295== by 0x590FFE4: gtk_main (in /usr/lib/
This looks very much like this bug, which was recently fixed upstream:
http://
Could we get an updated package which incorporates this patch?
The bug seems to be related to parsing play counts written by iTunes on the device. I have verified that the workaround described here (deleting the XML file with play counts from the mounted iPod Touch) fixes the problem: