RFC: Fix for issue 1827 (v2)

[Benoit Boissinot]

On Tue, Feb 16, 2010 at 12:37:44AM +0100, Sune Foldager wrote:
> On 16-02-2010 00:34, Benoit Boissinot wrote:
> > On Mon, Feb 15, 2010 at 11:30:23PM +0100, Sune Foldager wrote:
> 
> > I don't think we should move the hook up one level, every other commit
> > related hook is fired on commitctx, why should it be different for this
> > one?
> 
> Because that's how the bug can get fixed (if we want to fix it). The
> commit hook can't fire until wlock has been released, or commands
> running in it won't see the correct view of the world. Moving the wlock
> stuff around is a much bigger deal.

I see. So we assume nobody who calls commitctx() directly assumes hooks
will be fired?

Do we want to keep some sort of symmetry and move precommit up too?

regards,

Benoit

-- 
:wq