Re: bug (?) in libbonobo that is breaking evolution



Hi Jon et al,

On Wed, 2001-09-12 at 19:59, Jon Trowbridge wrote:
> OK, I've done more research, I've figured out what is going on (I
> think).  What is actually going on is pretty obvious... I feel fairly
> silly for missing it before.
> 
> The idle function that starts this all isn't allowed to re-enter, so the
> first lines of the code are
> 
>   if (foo->locked)
>     return TRUE;
>   foo->locked = TRUE;
> 
> to explicitly block re-entry.  When processing events, bonobo keeps
> calling that idle function, which refuses to go away.
> 

Hmm... FWIW, I managed to see this behavior (only once, sigh) in the
mailer once. I had exited from the shell and it was pegging the CPU in
the same gtk main loop code as in the addressbook. I tried to step
through in gdb, but nexting out of gtk_main_iteration caused GDB to
crash. Doh. I don't think the mailer has code like what Jon mentioned...
and the gtk main loop wasn't in an idle call, either :-(

So I guess I'm saying, maybe the solution to this problem is something
else.

Peter

-- 
Peter Williams     peter newton cx / peterw ximian com

"Why should I have to change my name? He's the one who 
sucks!"                              -- Michael Bolton





[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]