On Sat, 21 Nov 2009 23:17:02 +0100, Stefan Schmidt <stefan@datenfreihafen.org> wrote:
> 
> While talloc is great we need to free the g_error by hand.
> 
> Tested-by: Stefan Schmidt <stefan@datenfreihafen.org>
> Signed-off-by: Holger Freyther <zecke@selfish.org>

Thanks to both of you! This is pushed.

As for talloc and glib, it's become clear to me that I'd really like a
talloc-friendly hash-table implementation, (and maybe a nice growing
array). I really haven't seen any other data structures in glib that I'd
need as universally as I need those, (particularly since talloc already
provides a bunch of string functions).

-Carl