Re: mailing list archiver chewing patches

Aidan Van Dyk <aidan(at)highrise(dot)ca> writes:

> I'll note that the whole idea of a "email archive" interface might be a
> very good "advocacy" project as well. AOX might not be a perfect fit,
> but it could be a good learning experience... Really, all the PG mail
> archives need is:
>
> 1) A nice normalized DB schema representing mail messages and their
> relations to other message and "recipients" (or "folders")

We're now hoping that this one will fit:

http://www.archiveopteryx.org/schema

> 2) A "injector" that can parse an email message, and de-compose it into
> the various parts/tables of the DB schema, and insert it

aox has that either as a bulk importer or as a MDA.

> 3) A nice set of SQL queries to return message, parts, threads,
> folders based on $criteria (search, id, folder, etc)

I guess Matteo's working on that…

> 4) A web interface to view the messages/thread/parts #3 returns

And that too.

> The largest part of this is #1, but a good schema would be a very good
> candidate to show of some of PG's more powerful features in a way that
> "others" could see (like the movie store sample somewhere) , such as:
> 1) full text search
> 2) text vs bytea handling (thinking of all the mime parts, and encoding,
> etc)
> 3) CTEs, ltree, recursion, etc, for threading/searching
> 4) Triggers for "materialized views" (for quick threading/folder queries)
> 5) expression indexes

And Tsearch, too, maybe. Oh and pg_trgm might be quite good at providing
suggestion as you type or "Did you mean?" stuff.

Regards,
--
dim