nautilus crashes when remote systems are mounted

[Tom McLaughlin]

On Sat, 2004-05-15 at 16:31, Joe Marcus Clarke wrote:
> On Thu, 2004-05-13 at 23:48, Tom McLaughlin wrote:
> > On Wed, 2004-05-12 at 15:56, Joe Marcus Clarke wrote: 
> > > On Wed, 2004-05-12 at 00:35, Tom McLaughlin wrote:
> > > > Hi all, I've just started noticing that when I mount a remote system, be
> > > > it smb or sftp, through the Connect to Server option in nautilus,
> > > > nautilus crashes when I go to the computer:/// location.  I cvsuped
> > > > yesterday and rebuilt so I have the latest gnomevfs2 package but that
> > > > still does not help the problem.  The following is what I received from
> > > > gdb after attaching to the /usr/X11R6/libexec/gnome-vfs-daemon process
> > > > and recreating the crash:
> > > 
> > > This looks like a libpthread problem.  It's also one I cannot
> > > reproduce.  I'm running:
> > > 
> > > FreeBSD gyros.marcuscom.com 5.2-CURRENT FreeBSD 5.2-CURRENT #23: Mon May
> > > 10 17:49:51 EDT 2004    
> > > marcus at gyros.marcuscom.com:/usr/obj/usr/src/sys/GYROS  i386
> > > 
> > > Try updating world and kernel, and see if that helps.
> > > 
> > > Joe
> > 
> > 
> > Joe, I cvsuped -current and rebuilt/reinstalled the kernel and world.  I
> > am still receiving the same crash and the backtrace is exactly the
> > same.  I also found something else.  I attached to the gnome-vfs-daemon
> > and decided to fire up rhythmbox so I could get my groove on while
> > looking at this.  The mp3s in my playlist happen to be on an smb share
> > which I mount through nautilus.  Again gnome-vfs-daemon crashes and
> > spits out the same backtrace.  Once I close down gdb I can play my music
> > fine.  
> > 
> > I take it nautilus is essentially fine and only crashing because
> > gnome-vfs-daemon is crashing.  But nautilus only crashes when I go to
> > "computer:///".  I can still browse my remote mounts by clicking their
> > icon on the desktop.  This isn't a problem caused by the smb share since
> > nautilus crashes if I only have an sftp share mounted as well.  Not sure
> > what exactly is going on.
> 
> Not sure, either, but frame 2 from the stack trace below looks very
> suspect.  Maybe you have a symbol conflict somewhere.  Rebuilding glib20
> with debugging symbols may help track down the offender.
> 
> Joe
> 

I updated gnome to 2.6.1 last night and recompiled glib just now.  Here
is the backtrace again.

Tom


(gdb) attach 45811
0x2822041f in pthread_testcancel () from /usr/lib/libpthread.so.1
(gdb) bt
#0  0x2822041f in pthread_testcancel () from /usr/lib/libpthread.so.1
#1  0x282193fd in pthread_mutexattr_init () from
/usr/lib/libpthread.so.1
#2  0x08060000 in ?? ()
#3  0x2824b067 in g_io_channel_write_chars (channel=0x8058f80, buf=0xbb8
<Error reading address 0xbb8: Bad address>, count=2147483647,
bytes_written=0x8073180, error=0x4) at giochannel.c:2181
#4  0x2824a68b in g_io_channel_write_chars (channel=0x8058f80, buf=0x1
<Error reading address 0x1: Bad address>, count=1,
bytes_written=0x8057020, error=0x4c0) at giochannel.c:1934
#5  0x2824acde in g_io_channel_write_chars (channel=0x8059a50,
buf=0x8059a50 "\200\217\005\b\001", count=-1077942264,
bytes_written=0x281399ae, error=0x3) at giochannel.c:2075
#6  0x281399f1 in bonobo_main () from /usr/local/lib/libbonobo-2.so.0
#7  0x0804dbdf in main (argc=3, argv=0xbfbfe860) at
gnome-vfs-daemon.c:605
#8  0x0804c956 in _start ()
(gdb)