[hypermail] Incremental updates (was: Problems with 2.1.8)

From: Bill Moseley <moseley_at_hank.org_at_hypermail-project.org>
Date: Wed, 21 Jan 2004 10:52:34 -0800
Message-ID: <20040121185234.GA7446_at_hank.org>

On Tue, Jan 20, 2004 at 03:07:05PM -0800, Peter C.McCluskey wrote:
> moseley_at_hank.org (Bill Moseley) writes:
> >So, this creates an initial archive
> >
> > $ rm -rf archive
> > $ cat lots-of-message | hypermail -i
> >
> >So at this point what's the difference between these (not lack of -u):
> >
> > $ cat new_message | hypermail -i
> > $ cat new_message | hypermail -i -x
>
> If you are doing an incremental update (i.e. have an .hmrc with increment = 1
> or increment = -1 and new_message has no old messages), the -x option makes no
> difference (I guess I ought to add that to the documentation).

And increment = 1 is the same as -u on the command line, correct?

> If you are not doing an incremental update (i.e. have an .hmrc with increment
> = 0 and new_message contains lots-of-message plus something new), then
> if you changed your config file from showhtml = 1 to showhtml = 2 between
> the lots-of-message command and the new_message command, the -x will cause
> the entire archive to have the showhtml = 2 format, and avoiding -x will
> produce and archive with mixed format.

In other words, if you feed hypermail a bunch of messages it normally doesn't process messages it already has seen and just works on new messages in the input stream. The -x option says to treat every message as new -- rewriting existing \d+.html files.

How does hypermail keep track of messages it has already seen?

Why would you send hypermail messages it's already processed? To add to the archive I would think that you would just pipe the new message to hypermail as each one comes in.

Another question. Say I've got an existing archive and some old message needs to be removed or modified. What is the procedure?

-- 
Bill Moseley
moseley_at_hank.org

Received on Thu 22 Jan 2004 03:37:49 AM GMT