[Mailman-Users] Messages are lost...I have followed FAQs 3.14

[Stephen J. Turnbull]

>>>>> "Mike" == Mike Avery <mavery at mail.otherwhen.com> writes:

    Mike> Looking at the archives to the mailing list, I've seen that
    Mike> different people have had this problem since 2003 in one
    Mike> form or another.

Actually, not.  (See below.)

    Mike> Still, it seems Mailman should be able to handle the illegal
    Mike> data a bit more elegantly.... what's the old system designer
    Mike> mantra, "Never test for an error condition you don't know
    Mike> how to handle"?

The error condition being tested for is "all other errors", and we
know how to handle those ... "preserve the post, log the condition,
and let someone burn real neurons figuring it out."  Although I'm not
a Mailman developer, I do know a little about I18N, and it seems to me
from a brief look at the code that the problem is that there are a
number of different ways that such data can leak out to where it needs
to get processed, and the process has been to fix them in place as
they are identified.  (I've seen two or three of these bugs fixed in
the last year, I suspect your spammers have found Yet Another Path to
the error handler.)

What really needs to be done for Mailman 2.1 is a complete audit of
all the places where headers are accessed, but you know how expensive
that is.  For Mailman 3, what probably should be done is to rip out
all of the current just-in-time I18N processing of headers, and
preprocess every header, tagging them with their charsets.  Binary
headers would be tagged as "bogus" rather than "binary."

Now if I could just beg, borrow, or steal a few "round tuits"....

-- 
School of Systems and Information Engineering http://turnbull.sk.tsukuba.ac.jp
University of Tsukuba                    Tennodai 1-1-1 Tsukuba 305-8573 JAPAN
               Ask not how you can "do" free software business;
              ask what your business can "do for" free software.