On Sun, 2008-04-20 at 11:50 -0400, D Bera wrote:
> >  connect(15, {sa_family=AF_FILE, path="/home/meebey/.beagle/socket"}, 110) 
> > = -1 EAGAIN (Resource temporarily unavailable)
> 
> It is strange that connect() is returning EAGAIN. Does the socket
> physically exist ?

Yes the socket exist but is dead because of a crashed beagle (forgot the
evolution# issue?).

[EMAIL PROTECTED]:~$ stat .beagle/socket 
  File: `.beagle/socket'
  Size: 0               Blocks: 0          IO Block: 4096   socket
Device: 806h/2054d      Inode: 328351      Links: 1
Access: (0755/srwxr-xr-x)  Uid: ( 1000/  meebey)   Gid: ( 1000/  meebey)
Access: 2008-04-20 16:28:33.000000000 +0200
Modify: 2008-04-20 16:28:33.000000000 +0200
Change: 2008-04-20 16:28:33.000000000 +0200
[EMAIL PROTECTED]:~$ 

Beagle doesn't crash completly btw, the process is still there but not
even responding to kill, only kill -9 works in that stage, see #475640
for the crash trigger / cause.

After I killed beagled using kill -9, gnome-color-chooser and other GTK+
apps "work" again (not hanging with EAGAIN).

Beagle or libbeagle should handle this case gracefully though (when
beagle crashes / hangs for some reason)

> It is even stranger that gnome-color-chooser is using libbeagle.

probably some gtk widget they use does it...

> I will try to reproduce this on ubuntu and mandriva. Wish me luck.
-- 
Regards,

Mirco 'meebey' Bauer

PGP-Key ID: 0xEEF946C8

FOSS Developer    [EMAIL PROTECTED]  http://www.meebey.net/
PEAR Developer    [EMAIL PROTECTED]     http://pear.php.net/
Debian Developer  [EMAIL PROTECTED]  http://www.debian.org/

Attachment: signature.asc
Description: This is a digitally signed message part

Reply via email to