On Tue, Jun 11, 2002 at 09:42:50PM -0000, mutt-users-digest wrote:
> Date: Tue, 11 Jun 2002 14:04:03 -0500
> From: David T-G <[EMAIL PROTECTED]>
> Subject: Re: More send-hook (+folder-hook) questions (long)
>
> Your description is a little convoluted. Please allow me to attempt to
> clearly restate your goal and confirm or deny the presentation.
>
> In =3Dlists/mutt
> sending in general (to the list, to me, to your mom)
> From: mutt@yoursite
>
> In =3Dlists/loganalysis
> sending in general (to the list, to me, to your mom)
> From: loganalysis@yoursite
>
> Anywhere
> sending to mutt list
> From: mutt@yoursite
> sending to loganalysis list
> From: loganalysis@yoursite
I'd say that that's accurate, except I'd change the
phrases "sending in general" to "sending in general except to another
list", which I think is what you meant anyways.
David continues:
> If you want all three, then it does, indeed, get tricky. I think that
> the most straightforward approach is to pull all of the send-hooks (both
> for lists and for exceptional_persons) into a muttrc that you source
> from within a folder-hook, and then for each list folder construct a
> folder-hook about like
>
> folder-hook =3Dlists/mutt \
> 'send-hook . my_hdr From: mutt@yoursite ; \
> source $HOME/.mutt/muttrc.sendhooks'
, which Peter Abplanalp and Rocco Rutte also suggested in
one form or another. I had simplified things greatly in my rc.testing
file to track down the problem and then make it clear for my post; the
actual configs are generated by a perl script parsing XML config files
for ...
$ find ~/mail -type f -print | wc -l
994
... 994 different mailboxes and ...
find ~/.mutt/addresses -type f -print | \
> while read LINE ; do wc -l $LINE ; done | \
> awk '{LINE+=$1};END{print LINE}'
812
... 812 different entries in my address book. If I need
to, I _could_ have the script generate every possible combination of
mailboxes and recipients, but even given that not all of them have hooks
defined, it's not a pretty picture; if I can avoid doing a cartesian
join, I'd prefer to do so--I'd estimate that that could use up as much
as 20 megs of RAM for the hooks (assuming they are stored in text format
and not some more compact representation). For that matter, I have no
way of knowing until I figure out what's going on with the simple
two-folder scenario of =lists/mutt and =inbox (from the rc.testing file),
then I could well run into the exact same problem for my cartesian join.
So, does anybody know why I'm seeing the behaviour I'm
seeing? Is it a bug, or a "feature"? David -- I can't find anything
in the manpage about a debug log; is it a compile-time option?
TIA,
Sweth.
--
Sweth Chandramouli Idiopathic Systems Consulting
[EMAIL PROTECTED] http://www.idiopathic.net/
