GNOME Bugzilla – Bug 351704
Crash on gtk_recent_info_get_short_name
Last modified: 2007-01-10 11:23:37 UTC
What were you doing when the application crashed? Tried to view source from epiphany Distribution: Ubuntu 6.10 (edgy) Gnome Release: 2.15.91 2006-08-08 (Ubuntu) BugBuddy Version: 2.15.90 Memory status: size: 33763328 vsize: 0 resident: 33763328 share: 0 rss: 8634368 rss_rlim: 0 CPU usage: start_time: 1155771164 rtime: 0 utime: 21 stime: 0 cutime:18 cstime: 0 timeout: 3 it_real_value: 0 frequency: 0 Backtrace was generated from '/usr/bin/gedit' (no debugging symbols found) Using host libthread_db library "/lib/tls/i686/cmov/libthread_db.so.1". (no debugging symbols found) [Thread debugging using libthread_db enabled] [New Thread -1208186688 (LWP 25379)] (no debugging symbols found) 0xffffe410 in __kernel_vsyscall ()
+ Trace 70571
Thread 1 (Thread -1208186688 (LWP 25379))
> What were you doing when the application crashed? > Tried to view source from epiphany > > #9 0x4da37480 in g_assert_warning () from /usr/lib/libglib-2.0.so.0 > #10 0x4dd7814d in gtk_recent_info_get_short_name () Emmanuele, is this another case of GtkRecent not dealing properly with remote file uris? If yes, can we get it fixed in gtk? It's silly to add workarounds all over the place...
the stack trace is not useful to narrow it down. gtk_recent_info_get_short_name() calls get_uri_shortname_for_display() which is basically the same code used by gnome-vfs (and egg-recent) to get the last fraction of the URI and make sure that it's UTF-8 encoded; if the URI is not local (file:/) it'll try to get the URI scheme and add it to the string. I need the full backtrace, otherwise I don't know where it hit an assert().
apart from the backtrace I useful info to know it's the url of the page of which you were trying to view the source, so that we can try to repro.
(In reply to comment #1) > > Tried to view source from epiphany > > Emmanuele, is this another case of GtkRecent not dealing properly with remote > file uris? If yes, can we get it fixed in gtk? AFAIK Epiphany opens the local cached file in gedit, not the remote one.
Can we move this to gtk+?
<paolo> Can we move bug #351704 to gtk+? <bugsbot> paolo: Bug http://bugzilla.gnome.org/show_bug.cgi?id=351704 cri, High, ---, gedit-maint@gnome.bugs, NEED, Crash on gtk_recent_info_get_short_name <pbor> paolo: mmm... not sure <pbor> paolo: nud's comment is true <pbor> so I am not sure anymore that it's the same issue
if someone experiencing this crasher can get me a complete stack trace I'll more than gladly move this to gtk+; until then, I'd say it can stay as NEEDINFO because I can't see whether this is a bug in GtkRecentInfo creating or being passed wrong data.
*** Bug 360674 has been marked as a duplicate of this bug. ***
I'm not sure that the bug 360674 is directly a gedit bug. It seems to me that this is a gnome-panel bug at first. See the backtrace below
Created attachment 74308 [details] gnome-panel bug
It seems this crash also affects gnome-panel (see comment #10). Moving to gtk+.
I'm experiencing this one as well with Gnome, Gedit and gnome-panel all at version 2.16.1, and it is especially nasty. It started the same way mentioned above: I viewed source on a file in Epiphany, and then gedit and gnome-panel both crashed. On all subsequent startups, gnome-panel and gedit will crash. Here is the important update: if I remove "~/.recently-used.xbel", gnome-panel and gedit will start just fine and a fresh recently-used.xbel will be generated. I've attached the offending file, and a hopefully enlightening trace. I checked it quickly for anything incriminating, but luckily I do my dirty work on my desktop =) Thanks!
Created attachment 75953 [details] gnome-panel startup log and backtrace, and .recently-used.xbel
this feels like a duplicate of bug #363437. I'll dupe this, as the other bug has a meaningful stack trace. *** This bug has been marked as a duplicate of 363437 ***
*** Bug 375011 has been marked as a duplicate of this bug. ***