Re: gdu-volume-monitor segfault

[David Zeuthen <david fubar dk>]

On Tue, 2009-12-22 at 12:00 +0100, Martin Koegler wrote:
> If gdu-volume-monitor can not connect to the system dbus, it will
> segfault, while calling gdu_pool_get_presentables, because
> monitor->pool is NULL. 

This should never really be the case - my original thought when writing
this is that if you want to use GIO without a system bus present - then
you are probably doing it wrong.... however, someone pointed out that
this can happen in weird chroot cases - so we should probably do
something better than just crash.

> The following patch corrects this.

I don't like this patch much... Instead, we should create the GduPool
object when the volume monitor process starts and store it in a global
variable. If we can't get a GduPool object, we should return FALSE in
the is_supported() vfunc. This way, GIO processes can fall back to e.g.
GUnixVolumeMonitor or others.

Any chance you can file a bug at b.g.o about this? It's easier to handle
it there than on the mailing list.

Thanks,
David