Skip to main content.
home | support | download

Back to List Archive

Re: indexing files with multiple MIME parts

From: Andy Jacobson <andyj(at)>
Date: Tue Aug 17 2004 - 18:23:29 GMT
>>>>> "Bill" == Bill Moseley <> writes:

    Bill> So do you index in a way where you can search like

    Bill>    nnir and (author=andy or email=andyj)

No, whether by choice or by path of least resistance, the interface is
simpler than that. nnir.el prompts me for a search phrase, then finds
the proper messages.  It then creates a new, temporary mail group
(~folder) that contains all the hits.  One can then treat this dynamic
group and all its contents as if it were a standard mail group.

I've peeked under the hood enough to see that there are clever ways to
use different swish-e index files for different purposes, so you could
parse the emails for certain headers and save special index files for
each header, I suppose.  I wouldn't count on that working that way out
of the box, however, not without some work.  (Which unfortunately,
would be in emacs-lisp.)

    Bill> Have you also used mutt?

Nope.  The emacs keybindings have their own reserved place in my
cranium and so I always look for solutions in that universe first.

    Bill> I read my mail over imap from a number of different places,
    Bill> so really want a solution that works anywhere.  That's why
    Bill> I'm thinking about a web-based search tool that can write
    Bill> selected results to a new Maildir -- which then would be
    Bill> available from anyware via imap.

Maildir means you're using Courier-IMAP, right?  We're using UW IMAP
server, which iirc doesn't play nicely with multiple backend formats.
This limitation stopped my explorations in this direction.  

I guess you will also need to be able to run arbitrary processes on
the IMAP and web server.  You need (perl?) code to be able to read,
process, and write the back-end format, and for the IMAP server to
recognize the creation of new groups from external processes.  Sounds
like an interesting project, and possibly of quite wide interest.  I'm
not so sure of the stability of my computing situation--one day I may
need to operate in a more restrictive environment that doesn't give me
those permissions.  If you go forward with this, may I recommend that
you do offer the ability to process all the MIME attachments.  CPAN
has the stuff you need, and I could pass you the (crap) code that I'm

There's also the IMAP search command.  I found this to be unusably
slow for the volume of mail I have archived, and not flexible enough
to do what I wanted anyway.

I went through a phase of "I must be able to do all of this from
anywhere, including Windows machines in random public libraries," but
then I got over it.  A good portion of public machines have some sort
of ssh, and I've always got a laptop with me anyway. The possibility
of keystroke loggers on public machines has seriously lessened the
attractiveness of those computers anyway.  I find that the
pervasiveness of wifi is sufficient to get by with my own machine now
while travelling.



Andy Jacobson

Program in Atmospheric and Oceanic Sciences
Sayre Hall, Forrestal Campus
Princeton University
PO Box CN710 Princeton, NJ 08544-0710 USA

Tel: 609/258-5260  Fax: 609/258-2850
Received on Tue Aug 17 11:24:02 2004