Re: [Evolution-hackers] Re: Create/Modify/Remove in the Calendar

[JP Rosevear <jpr ximian com>]

On Fri, 2003-09-12 at 18:44, Rodrigo Moya wrote:
> On Fri, 2003-09-12 at 09:05 -0400, JP Rosevear wrote:
> > On Fri, 2003-09-12 at 06:35, Rodrigo Moya wrote:
> > > On Fri, 2003-09-12 at 01:32 -0400, JP Rosevear wrote:
> > > > So, using create/modify/remove object methods in the calendar (matching
> > > > the addressbook api) is starting to shape up nicely, however I
> > > > encountered one problem.  Create has the nice feature of returning the
> > > > uid of the object created (addressbook does this because it can't
> > > > dictate to the ldap server the id I believe).  This is nice since it
> > > > allows us to match the server ids (esp. in the groupwise case where we
> > > > can't set any old uid).
> > > > 
> > > > Now, the problem I've encountered is that itip/imip needs to keep the
> > > > exact uid for 3rd party objects arriving via email.
> > > > 
> > > hmm, right
> > > 
> > > > Aside from dropping this system entirely, I've thought about a couple of
> > > > things:
> > > > 
> > > > 1. Add a receiveObject call to match the sendObject call - both will be
> > > > itip/imip specific (I think this still causes some work in the case of
> > > > groupwise since this part will force us to retain a mapping of 3rd party
> > > > uids to groupwise uids).
> > > > 
> > > retain a mapping? Can't you create objects in the GW server with any
> > > UID?
> > 
> > Nope.  You put it in the system and it gives you an id.  This id is only
> > guaranteed to be unique to the server as well.
> > 
> hmm, so what does it do then when getting iTip components from other
> servers/clients?

I don't know.  Scott?

We'll end up keeping a mapping anyhow likely.

-JP
-- 
JP Rosevear <jpr ximian com>
Ximian, Inc.