GNOME Bugzilla – Bug 510236
rhythmbox crashed with SIGSEGV in rhythmdb_entry_allocate()
Last modified: 2009-04-27 08:28:44 UTC
The report has been filled here: https://bugs.launchpad.net/ubuntu/+source/rhythmbox/+bug/183797 "Rhythmbox was reading my Music folder and importing the metadata in the database for the first time when it crashed. I was also playing an mp3 file at the time. I was NOT using the crossfading engine by the way," rhythmbox version 0.11.4 "#0 rhythmdb_entry_allocate (db=0x828e000, type=0x0) at rhythmdb.c:1206 size = <value optimized out>
+ Trace 185729
Probably fixed in svn: 2008-01-27 Jonathan Matthew <jonathan@d14n.org> * rhythmdb/rhythmdb.c: (rhythmdb_add_import_error_entry), (rhythmdb_missing_plugins_cb), (default_sync_metadata): Don't pass in NULL for error/ignore entry types, and if we don't have a valid entry type for error entries, don't create one. Probably fixes #510236.
*** Bug 536336 has been marked as a duplicate of this bug. ***
*** Bug 535261 has been marked as a duplicate of this bug. ***
*** Bug 540102 has been marked as a duplicate of this bug. ***
*** Bug 542467 has been marked as a duplicate of this bug. ***
*** Bug 540647 has been marked as a duplicate of this bug. ***
*** Bug 542936 has been marked as a duplicate of this bug. ***
This is NOT fixed in version 0.11.5, see the duplicates listed. Reopening.
*** Bug 543343 has been marked as a duplicate of this bug. ***
*** Bug 548510 has been marked as a duplicate of this bug. ***
*** Bug 548521 has been marked as a duplicate of this bug. ***
The following is to confirm that in Debian the bug still exists in version 0.11.6 (my former report was about version 0.11.5). Distribution: Debian lenny/sid Gnome Release: 2.22.3 2008-06-30 (Debian) BugBuddy Version: 2.22.0 System: Linux 2.6.25-2-amd64 #1 SMP Fri Jun 27 04:49:22 UTC 2008 x86_64 X Vendor: The X.Org Foundation X Vendor Release: 10402000 Selinux: No Accessibility: Disabled GTK+ Theme: Yattacier3 Icon Theme: NoiaWarm Memory status: size: 171429888 vsize: 171429888 resident: 40894464 share: 18538496 rss: 40894464 rss_rlim: 18446744073709551615 CPU usage: start_time: 1221064673 rtime: 254 utime: 220 stime: 34 cutime:4 cstime: 3 timeout: 0 it_real_value: 0 frequency: 100 Backtrace was generated from '/usr/bin/rhythmbox' [Thread debugging using libthread_db enabled] [New Thread 0xf6595740 (LWP 10977)] [New Thread 0xf0dfbb90 (LWP 11050)] [New Thread 0xf1dfdb90 (LWP 11049)] [New Thread 0xf15fcb90 (LWP 11048)] [New Thread 0xf25feb90 (LWP 11047)] [New Thread 0xeedf7b90 (LWP 11032)] [New Thread 0xef5f8b90 (LWP 11030)] [New Thread 0xefdf9b90 (LWP 11029)] [New Thread 0xf05fab90 (LWP 11028)] [New Thread 0xf4dd3b90 (LWP 11025)] [New Thread 0xf2dffb90 (LWP 11009)] [New Thread 0xf37ffb90 (LWP 11008)] [New Thread 0xf5ecdb90 (LWP 11007)] [New Thread 0xf41ffb90 (LWP 10992)] 0xf7f4f430 in __kernel_vsyscall ()
+ Trace 206599
Thread 1 (Thread 0xf6595740 (LWP 10977))
*** Bug 553127 has been marked as a duplicate of this bug. ***
*** Bug 549220 has been marked as a duplicate of this bug. ***
*** Bug 552795 has been marked as a duplicate of this bug. ***
*** Bug 558972 has been marked as a duplicate of this bug. ***
*** Bug 563641 has been marked as a duplicate of this bug. ***
*** Bug 563644 has been marked as a duplicate of this bug. ***
Created attachment 125089 [details] [review] no allocate error entry for RHYTHMDB_EVENT_METADATA_LOAD events event with type RHYTHMDB_EVENT_METADATA_LOAD have NULL error_type and ignore_type fields (maybe not always, I don't know). rhythmdb_entry_new:1499 crash because of try to a get size of entry depends on error_type.
No, that won't work at all. We *only* create import error entries in response to metadata load events, so this is not the right solution.
Created attachment 125550 [details] [review] set error_type and ignore_type for all errors
This appears to set the error and ignore types to 'invalid' for all metadata load operations, so it completely disables all processing of import errors and files that should be ignored. To fix this, we need to understand how a metadata event arrives with one or both of those fields set to NULL. Blindly hacking things won't help.
*** Bug 570460 has been marked as a duplicate of this bug. ***
*** Bug 570463 has been marked as a duplicate of this bug. ***
*** Bug 570464 has been marked as a duplicate of this bug. ***
Finally figured it out. 2009-03-26 Jonathan Matthew <jonathan@d14n.org> * plugins/generic-player/rb-nokia770-source.c: (rb_nokia770_source_new): * plugins/generic-player/rb-psp-source.c: (rb_psp_source_new): Set the ignore and error entry types for these to 'invalid' rather than leaving them NULL. Fixes #510236.
*** Bug 580106 has been marked as a duplicate of this bug. ***
*** Bug 580397 has been marked as a duplicate of this bug. ***