Re: undigesting and filtering by list-id

[Ken Irving <fnkci@uaf.edu>]

On Mon, Sep 24, 2007 at 01:27:13PM -0400, Celejar wrote:
> On Mon, 24 Sep 2007 09:15:16 -0700
> John L Fjellstad <john-debian@fjellstad.org> wrote:
> 
> > Celejar <celejar@gmail.com> writes:
> > 
> > > The best practice when filtering list traffic is apparently to use the
> > > 'list-id' header (to catch 'to', 'cc', etc).  But the best
> > > practice is also apparently to undigest digest emails with formail
> > > (greatly easing reading, threading and replying), which doesn't
> > > preserve the 'list-id' header.  Is there no choice, then, but to filter
> > > by 'to' (and 'cc' etc.)?
> > 
> > Couldn't you reinsert the list-id header?  That is, pull out the list-id
> > header before undigesting, store it, undigest, insert list-id.
> 
> Thanks.  That seems to be the consensus solution, to either do that, or
> quickly and dirtily to just insert a hard-coded list header.

You didn't mention how you're doing the filtering, but procmail uses a 
special pattern to match stuff, e.g., maybe something like:

   * \/List_id: .*

then later,

   formail -a "$MATCH" ...

I'm no procmail expert, but have several entries doing this sort of thing.
It generally takes me much trial & error to work out the procmail code,
but then it just runs behind the scenes and I forget about it.

OT, but one concern I have is that stuff like this gets added to
.procmailrc but rarely if ever gets taken out.  Sometime it might be nice
to move to a more transparent, less arcane filtering system, and maybe
such a system could keep track of how often patterns actually get hit.

Ken

-- 
Ken Irving, fnkci@uaf.edu