Red Hat Bugzilla – Bug 143342
Last modified: 2007-11-30 17:10:57 EST
Every now and then I find a core file produced by gam_server in my home dir.
Haven't noticed what causes it. Backtrace attached in case it's useful anyway.
This is gamin-0.0.19-1.FC3 from updates-testing, but I think I've seen it with
earlier versions too.
Created attachment 108858 [details]
gam_server gdb backtrace
Okay, thanks. In gam_tree_get_children if tree and root are NULL then the
code would segfault. Fixed in CVS,
Nice, thanks. Would it be possible to get a FC3 erratum for this?
By the way, I believe the app that I've seen most of these gamin segfaults
resulting from is kid3 from (pre-)Extras, using KDE on FC3.
Created attachment 111233 [details]
Another gam_server gdb backtrace
I'm not sure if this is the same problem (the backtrace looks different), but
gamin-0.0.24-1.FC3 is even worse than the earlier versions.
Now, launching seemingly *any* KDE app remotely causes a gam_server crash.
$ ssh somewhere
[exit app, wait a while]
$ Mutex destroy failure: Device or resource busy
ICE default IO error handler doing an exit(), pid = 5750, errno = 0
** ERROR **: file gam_tree.c: line 146 (gam_tree_remove): assertion failed:
(g_node_n_children(node->node) == 0)
Backtrace is not especially usful, get a debug trace from the gam_server
with kill -USR2 c.f. instructions at
Created attachment 111238 [details]
gam_server debug trace
Here you go.
gamin-0.0.25-1.FC3 still has the problem.
problem on #4 is a duplicate of #150471, which is fixed in rawhide
Confirmed, 0.0.26-1 rebuilt on FC3 seems to be ok. May I suggest a
FC3 errata for this?
I will give it a few days in rawhide and then push an errata, yes.
It's being pushed to Mandrake Cooker and Ubuntu too so I think within
a week I will get any regression feedback.
Well it's been a month and it hasn't been pushed to FC3 yet. Could an 0.0.26
errata be pushed to FC3 please ? This thing lays core files on all my FC3
systems almos on a daily basis.