defining the new idmap subsystem

[Jeremy Allison]

On Mon, Oct 02, 2006 at 11:00:17PM -0400, simo wrote:
> 
> The problem is that this will duplicate all query calls, it's doable,
> but I wonder if it is really worth it.

Yes - it's not only worth it, it's essential.

> So you would like to see something like splitting
> idmap_uid_to_sid() into idmap_uid_to_sid_query() and
> idmap_uid_to_sid_allocate() ?

Yes. We need to know *explicitly* when an allocate happens.

> I think the code will be much uglier but if most want that, so be it.
> But I'd like some more people complaining about that, before changing.

I would vote -1 on you putting it in as-is - I think this
is one case where I'm going to pull rank on you in design,
sorry. I will not accept a flag-based replacement for this.

Jeremy.