[Mailman-Users] Thanks, and some archive questions now.

[Mark Sapiro]

Gail wrote:

>My question is, is it possible to somehow get old digests into the
>system?  And if so, is there a way to write a script to do that in some
>automated way?  


It is trivially easy to import messages into a Mailman archive with the
bin/arch command line tool IF and ONLY IF the messages are in unix
mbox format - i.e., a flat file with each message preceded by a line
of the form

>From sender at example.com Mon Mar 13 17:01:06 2006

followed immediately by the message headers, a blank line, the message
body and a blank line.

If the messages aren't in mbox format to begin with, they need somehoy
to be put in that format. There is some info at
<http://www.python.org/cgi-bin/faqw-mm.py?req=show&file=faq03.036.htp>
that may be helpful.


>Also, would I get better, searchable archives out of somehow posting each of
>the messages separately and letting the system create new digests?


What do you have in your existing digests?

Mailman's pipermail archives are indexed by month (by default) and each
month has indices by date, subject, author and thread. The thread
index depends on messages having a Message-Id: header and replies
having a References: or a In-Reply-To: header. If you don't have this
information in the mbox file, the thread index will be the same as the
date index.

You still want individual messages rather than digests because
otherwise, the archive indices will be full of messages with subject
"... digest issue # ...".

Also, Mailman archives are not searchable within Mailman. There are
patches for this, and if the archives are public, they can be searched
with google if google indexes them. Go to
<http://www.python.org/cgi-bin/faqw-mm.py> and search for archive.
Some of the returned articles are relevant.


>And, if that's the case, could I do this update in a second list with only
>one or two dummy members, then somehow merge these archives with the 'real'
>list we'd be running concurrently while all these old messages were being
>processed into the system as I'm describing?


If you need to create archives by posting messages one at a time, then
yes, you can do it on a list with no members, and the archiving
process will build the mbox file which can later be put together with
the one for the live list. See
<http://www.python.org/cgi-bin/faqw-mm.py?req=show&file=faq03.003.htp>
which discusses removing posts from the archive, but the same
information applies to adding to an archive.

Basically, you just put the two mbox files together onto one and
rebuild the archive. There are some issues with message numbering
which are discussed in the FAQ, but they shouldn't be to significant
with a fairly new archive.


>I also wonder if some or all of this rather labor intensive work I'm talking
>about could be handled by some sort of scripts?


Again, depending on what information you have, you can script the
creation of a mbox file, and you can script a list posting process
too, but this shouldn't be necessary because posting a message to a
list is not going to add anything that couldn't be programmatically
added to create a mbox file.

-- 
Mark Sapiro <msapiro at value.net>       The highway is for gamblers,
San Francisco Bay Area, California    better use your sense - B. Dylan