[Mailman-Developers] Mail-to-news gatewaying?
Tokio Kikuchi
tkikuchi at is.kochi-u.ac.jp
Tue Feb 8 04:11:12 CET 2005
Hi,
Brad Knowles wrote:
> The latest crisis that has me wondering is the subject prefix
> stuff. Looking in NewsRunner.py, starting at line 104, we have:
>
> # Should we restore the original, non-prefixed subject for gatewayed
> # messages?
> origsubj = msgdata.get('origsubj')
> if not mlist.news_prefix_subject_too and origsubj is not None:
> del msg['subject']
> msg['subject'] = origsubj
>
>
> However, we still have a problem. If mail clients reply to the
> message to the list (which will include the subject prefix), then this
> will result in a new message with a subject prefix for the list already
> present. Now, Mailman is smart and avoids adding a second subject
> prefix, but it doesn't strip out the subject prefix if one already
> exists. For the mailing list, this subject prefix stripping isn't
> necessary. But for gatewaying to the newsgroup, it might be desirable.
Sorry but I have no newsgroup forwarding environment here. Our system
administrator (myself) determined long ago that newsgroups (usenet and
fj) have extremely low S/N ratio and waste of storage and bandwidth. ;-)
>
> This would also resolve the issue of restoring the original subject
> when gatewaying -- don't bother, and just let the subject prefix
> stripper deal with it.
>
>
> If this were sed or Perl, I could quickly put together a one-line
> hack to implement this feature. For example, just doing something off
> the top of my head, you might be able to use something like:
>
> subject=y/\w*$prefix\w*/ /;
Things are not easy because there are MIME subjects and we also start
numbering prefix in mailman 2.1.6. May be we can calculate
'stripped_subject' in CookHeaders.py and store in msgdata for later use.
Or, strip in a magical ;-) way which I use in CookHeaders.py and in
HyperArch.py, in NewsRunner.py again.
Anyway, I will try.
--
Tokio Kikuchi, tkikuchi@ is.kochi-u.ac.jp
http://weather.is.kochi-u.ac.jp/
More information about the Mailman-Developers
mailing list