From: Bill Moseley (moseley@hank.org)
Date: Fri Dec 28 2001 - 22:30:01 CST
At 07:22 PM 12/28/01 -0500, Jim Lill wrote:
>
>I use procmail to split up incoming stuff from a list into 10 sub-cats and
>then let HM makes 10 archives. I do this by subject. I need 10X hmrc files
>to do this although the bulk of the hmrc files is identical. Is there a
>function I'm missing that would allow include of a basic hmrc into the
>by-subject specific hmrc.
Hi Jim,
This is not really related but it solves a similar problem for me:
Just in the last few hours I wrote a perl script to automate hypermail to
some degree. I've got a bunch of lists that I need to archive. The lists
all have mbox files in gzip and text format (ezmlm generated lists) with
each collection of mbox files in their own directory. I'm also adding a
search feature to the archive using swish-e.
The way it works is you setup one config file where [text] marks a section
that defines a list. e.g.:
[Title of This List One]
list_email = somelist@foo.org
archive_dir = otherdir
strip_subject = [Somelist Discussion]
mbox_dir = /path/to/mbox/files
hypermail_opts = gmtime=On, showhtml=1
header_order = List-Post To Cc
[Title of This List Two]
...
Running the program like:
./mail_archive.pl --create
scans the config file and for each section (list archive) that doesn't
already exist, it creates a directory for the archive, and if mbox_dir
exists, reads in any mbox files and runs hypermail for each one building
the hypermail archive from scratch. Then it indexes the archive for
searching with swish-e.
In .procmailrc, I use:
MAILDIR=$ARCHIVE_DIR
: 0w
| ./mail_archive.pl --mode=update
Then the mail headers are scanned to route it to the correct archive. It
uses the header_order config setting and looks for the address set by
list_email.
Then there's also a crontab entry:
0,10,20,30,40,50 * * * * ./mail_archive.pl --mode=index --chdir $HOME/archives
which reindexes (for swish-e) any archives that have been added to in the
last ten minutes.
For your setup I suppose you could set
header_order = Subject
match_string = foo
to match automatically for you. I suppose it would be helpful to make
match_string a regular expression.
It's probably overkill for your setup unless you want to provide a search
engine for the archive, as well.
I'm currently testing this at:
http://search.apache.org/archives/
But that's only a temporary location.
If interested let me know by private email -- I can always use help testing...
I can also use pointers on how to best configure higher volume lists, and
how best way to purge old messages.
-- Bill Moseley mailto:moseley@hank.org