Re: Request for gnome-session freeze break

[Mark McLoughlin <markmc redhat com>]

Hi,
	I have to admit I got confused[1] when reviewing the patch, I got the
impression it fixed the "splash screen doesn't disappear" bug:

  http://bugzilla.gnome.org/show_bug.cgi?id=116814

	The patch is fairly straightforward and obvious, but unless it fixes an
actual reproducible crasher I'd think I'd prefer delaying this one until
2.8.1. Patches for gnome-session's startup sequence a week before the
release is enough to make anyone nervous I think :-)

Cheers,
Mark.

[1] - Entirely my fault, I should have looked more closely

On Thu, 2004-09-09 at 08:32, Murray Cumming wrote:
> So, assuming that it's actually a crash that's being fixed here, and that
> you've tested this for a couple of days, here's 1 of 2 approvals.
> 
> > Just to make sure no fun is missed, I'd like to request permission to
> > commit an obvious fix to gnome-session.  Please see this bug:
> >
> > http://bugzilla.gnome.org/show_bug.cgi?id=151664
> >
> > The patch attached to it should make the bug really obvious.  Whenever
> > gnome-session adds a client, it installs a timeout handler to purge the
> > client if it doesn't register itself before the timeout expires.
> > However, this timeout handler is never removed.  Thus, if the client
> > gets reincarnated while the timer is running, the old client structure
> > will be free()d and the timeout handler will access freed memory.
> >
> >   Federico