Re: [evolution-patches] timezone support in ECalBackendCache

[JP Rosevear <jpr ximian com>]

On Tue, 2004-03-16 at 05:53, Rodrigo Moya wrote:
> On Mon, 2004-03-15 at 23:53 -0500, JP Rosevear wrote:
> 
> > On Sat, 2004-03-13 at 10:04, Rodrigo Moya wrote:
> > > On Fri, 2004-03-12 at 08:40 -0500, JP Rosevear wrote:
> > > 
> > > > On Thu, 2004-03-11 at 16:08, Rodrigo Moya wrote:
> > > > > On Thu, 2004-03-11 at 14:40 -0500, JP Rosevear wrote:
> > > > > 
> > > > > > On Thu, 2004-03-11 at 11:33, Rodrigo Moya wrote:
> > > > > > > This patch adds support for storing / retrieving timezones in the
> > > > > > > cal-backend cache. This is needed for an upcoming patch I'll send
> > > > > > > separately to add timezone support to the GW backend.
> > > > > > 
> > > > > > It would appear that you are looking up the timezone objects in the
> > > > > > cache based on tzid.  Aren't they stored based on UID? 
> > > > > > 
> > > > > the timezones are stored based on their TZID, the events and tasks,
> > > > > based on their UID RID
> > > > 
> > > > Using the same cache?  It seems strange to have two different keys.
> > > > 
> > > well, the IDs should always be different, and the ECalBackendCache class
> > > deals with the differences between the 2 kind of objects. Also, are
> > > VTIMEZONE's guaranteed to have always an UID?
> > 
> > Well, in praticality they should be different, it just seems improper to
> > use two different keys for data access (come on, you're a database
> > guy).
> >
> :-)
> 
> >   Timezones are defined to not have a UID.
> > 
> right, that's why I'm using the TZID. If they've got no UID, we can't
> use the same key as we use for events/tasks. The other solution would be
> to store the timezones in a different file, which seems a bit
> overwhelming to me, or use the same key for all cache objects. But what
> to use?

There is no common key I guess.  I would note this discrepancy as a
"FIXME" I guess.

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