Hi,
has anyone got an idea how to go at indexing email?
What I've seen from looking around, you can't do that via cron - it
has do be done on-the-fly.
For this purpose, there seem to be mostly two contenders: Lucene and
Xapian.
From reading the Xapian docs that come with it, the big gotcha is to
pass an email to it the moment it enters the system.
This seems to imply that somehow vdelivermail and courier-imap (and
qmail-pop3d or courier-pop3d) would have to be patched.
The imap-server, too, because messages can also enter via IMAP (sent-
folder, drafts).
One would have to delete a document from the index, once it's fetched
via POP3, too.
Has anyone given this some thought?
It would be very cool to have a good search in one's webmail that
actually works for big mailboxes....
cheers,
Rainer