Re: Mail work

[Joe Shaw <joeshaw novell com>]

On Mon, 2004-11-29 at 13:03 -0800, Alex Graveley wrote:
> Regarding relevancy information: what other information will indexing
> the mail directly get us?

That's the bulk of it.  We can also get rid of a lot of wasted CPU
cycles because mbox files are seekable (since every message starts with
"From ") and the Evo summaries are not.  Which means that we don't have
to iterate over 70,000 messages just to index 3 new ones.  We can keep
an offset in the mbox.

> If you are going to avoid using the Evo summary files altogether how do
> you plan to get things like read/unread, flagged, replied to, etc
> information into Lucene?  The e-plugin?

We lose the flags, but AFAICT that's all we lose.  And yeah, those would
be tracked through the e-plugin.

Joe