Re: [Evolution-hackers] ESource shortcomings

On Tue, 2003-11-11 at 19:08, Hans Petter Jansson wrote:
> On Tue, 2003-11-11 at 15:53, Ettore Perazzoli wrote:
> > On Tue, 2003-11-11 at 14:16, Hans Petter Jansson wrote:
> > > I think subclassing would be nicer, since it would specify which
> > > properties are used, and provide easy type checking.
> > Nod.  So, who writes these subclasses?  :-)
> I'll do it for webcal and LDAP. Shouldn't take much.

So I discussed this with ettore and hans offline and we think these
should actually be named properties on the backend.  In this scenario,
the backends should probably be given the source uid rather than the
uri, because they will need to grab the named properties for connecting.

The main issue here is the GUI for some individual backends.  For
instance, where will the LDAP backend gui come from?  and will it be
readily available to outside developers?  I'm reluctant to put the gui
bits into e-d-s but it may logically be the only place to put them.  I
suppose we could have some sort of generated ui, but those tend to
degrade over time.

> > > So basically, I don't think we need the "move" folder op at all, and the
> > > "copy" folder op only for local sources. Is that correct?
> > We need "copy", "rename" and "delete".  These ops are not just for local
> > stores, in some stores you'd want that to still work even if the folders
> > are not local (e.g. think Connector).

The delete method is already on the backend, I think probably the rename
method should be too.  Then the backend can handle the actual rename,
which is useful for groupwise and exchange.

For copy we can probably create a new calendar then fetch all the items
and save them.

JP Rosevear <jpr ximian com>
Ximian, Inc.

[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]