Re: [OT] Re: [Evolution] strategies for handling lots of mail

[Not Zed <notzed ximian com>]

> > > I don't know about the new features in CVS, but I'm pretty happy
> > with
> > > the present state (all my colleagues as well).
> >
> > Well, "redesign" means ... well, redesign.   i.e. not the same no
> > more.
> >
> > I guess we'll see how it goes once its out there.
>
> I've managed to hack HEAD to compile (the gal_a11y stuff is still
> screwed up, lots of missing dependencies on libevolution-a11y.so), and
> the performance is noticeably snappier.  Your etree performance fixes
> seem to have really helped.  Nice work.

FWIW These have gone into the 2.2 branch too, the etree and mail
changes.

The experimental/rapidly changing/quite unstable 'disksummary' branch is
where the more radical stuff is happening.  Too many changes to list
really, but for example 'vfolders' are ceasing to exist entirely as
'automatically updated searches presented as a folder', although they
are to be replaced with something that can be made to look the same, but
it is based on alternate folder indices instead.

The changes address various major and/or architecturally related issues:
 - overall mailer memory use
 - vfolder startup time
 - help any eventual movement of mail to a desktop-wide CORBA service
 - imap
 - cleaning up some old/redundant apis