[Mailman-Users] importing archived Maildir email lists into Mailman lists archives

[[IDIS Technical Secretariat] Ricardo Rodríguez]

Thanks for your reply! Please, read below!

On Tue, Sep 1, 2015 at 3:19 AM, Stephen J. Turnbull <stephen at xemacs.org>
wrote:

> [IDIS Technical Secretariat] Ricardo Rodríguez writes:
>  > Thanks Mark, all,
>  >
>  > On Sun, Aug 30, 2015 at 4:09 AM, Mark Sapiro <mark at msapiro.net> wrote:
>  >
>  > > On 08/28/2015 12:52 AM, [IDIS Technical Secretariat] Ricardo Rodríguez
>  > > wrote:
>  > > >
>  > > > I have a number of Maildir format mailing lists archives
>
> No, you don't have Maildir, at least not "Maildir" as most of the
> Internet understands it.  Here's what Dan Bernstein (the inventor or
> at least popularizer of Maildir) says:
>
>     Can a maildir contain more than tmp, new, cur?
>
>     Yes:
>     .qmail: used to do direct deliveries with qmail-local.
>     bulletintime: empty file, used by system-wide bulletin programs.
>     bulletinlock: empty file, used by system-wide bulletin programs.
>     seriallock: empty file, used to serialize AutoTURN.
>
>  > Ricardo-Rodriguezs-Mac-Pro:r.users rrodriguez$ ls
>  > Log bounce digissue headerremove lock mod outlocal remote
>
> "Lock" -- no, this isn't Maildir.  The whole point of Maildir is that
> you don't need locks because reading and writing are done in different
> directories, and changes happen atomically.  (This can even work with
> editing.)
>

That's now far clear for me! As stated in a previous message replying to a
Mark's post, I make a mess interpreting wrongly several messages from our
services provider and some googled information. Ezmlm was behind the scene.
Sorry for the misinformation!


>
>  > allow bouncer dignum indexed lockbounce modsub owner subscribers
>  > archive config editor inlocal mailinglist num prefix text
>  > archived digest headeradd key manager outhost public tstdig
>  >
>  > Within /archive, there are two folders, 0 and 1, with a number of files,
>  > each of them containing one message, and an index file.
>  >
>  > Please, does this made sense for you?
>
> I don't recall anything like that.  Please try to find an explanation
> of the structure in the system documentation, or ask the vendor.
> However, since you think they're "Maildir", probably what is meant is
> that they have a structure that is one message per file rather than
> many messages per file.  You probably just need to figure out how to
> get the order of messages right, then concatenate the messages.
>
>
That's correct: messages are stored in separated files ordered in
subfolders named with a series from 0 onward. 0, 1... each subfolder holds
one hundred files, each of them with a complete message. I think the global
order could be provided by the subfolder order plus the name of each file
within a subfolder.

Mark's reply contents a simple line to get some concatenation. I'll play
with this idea and report back!


> Most likely, all you need for each list are the archive folders and
> the single messages, and maybe the index file will be of some use
> depending on what it contains.  If your documentation and/or the old
> vendor are of no help, see if you can find a whole message file you
> can send to us *as a file attachment* -- we want to see what headers
> are included (it probably doesn't really matter, though, except for
> the "Unix From_" which they probably don't have).  If for privacy
> reasons you don't want to broadcast any message on a public mailing
> list, you can send it to Mark and me personally.  Also it may be
> helpful to figure out the rule for the folders whose names are
> numbers: are they the leading digits where files are named 000 to 999?
> Are they months?  Years?  etc.
>

Most lists are public, so privacy is not a concern! You can find here the
complete folders' structure for one of the lists...

http://datasource.idisantiago.es/r.users/


>
>  > Am I completely lost?
>
> No, of course not.  Just don't delete anything until you're sure the
> new system is working.  I did qualify everything with "probably", it
> may take a couple of guesses to get it right. :-)
>
>
Thanks for your help!

Ricardo

-- 
Ricardo Rodríguez
Research Management and Promotion Technician
Technical Secretariat
Health Research Institute of Santiago de Compostela (IDIS)
http://www.idisantiago.es