[Mailman-Developers] Serious bug shunting messages

[Barry A. Warsaw]

>>>>> "DB" == Daniel Buchmann <Daniel.Buchmann at bibsys.no> writes:

    DB> Thanks a lot Barry! The recently committed change to i18n.py
    DB> solved the problem! :D My qfiles/shunt/ dir is now empty. I
    DB> also had a problem with the daily cron/senddigests, but it
    DB> went away after patching i18n.py. ;)

Excellent!

    DB> Btw: where is the message pipeline in Mailman? What I want to
    DB> do is change the order of the tests that cause messages to be
    DB> held etc.  (See my post
    DB> http://mail.python.org/pipermail/mailman-developers/2003-February/014847.html)

This part's definitely kludgy, and I have ideas for how to fix it.
But for now, there are two places to look.  First is GLOBAL_PIPELINE
in Defaults.py.in.  There you'll see that it calls out to the Moderate
module and then to the Hold module.  There's been a bit of thrash on
the order of these. ;)

Then, inside Hold.py, you'll see that process() hard codes the order
in which conditions are tried.  Changing this requires hacking the
code.

My idea (for 2.2 or 3.0 -- I'm still conflicted) is that /all/ the
hold tests would be applied to a message and there'd be some way
<wink> for admins to either see all the conditions that apply or
select the priority in which they cause a message to be held or
rejected.  Not much more detail than that, but it's definitely
something on my radar.  Suggestions are welcome!

    DB> Thanks, hope to see you around on #mailman again sometime
    DB> soon. ;)

I'm trying to hang out there occasionally these days!
-Barry