[Mailman-Developers] Hmm... More on header mangling...
Ron Jarrell
jarrell@vt.edu
Sat, 18 May 2002 01:29:11 -0400
At 10:50 PM 5/17/02 -0400, Barry A. Warsaw wrote:
> >>>>> "RJ" == Ron Jarrell <jarrell@vt.edu> writes:
>
> RJ> Mailman did the "mangle the header line prematurely ending the
> RJ> headers" trick to it. But for once I can see what the header
> RJ> looked like originally!
>
>Confirmed (and sorry about that hold warning you probably got Ron ;).
Which had me confused as hell for about 30 seconds, trying to figure out
how one of my tests ended up routed through wooz.org! :-)
>Now I just need to figure out why!
Note that my "three line header" idea got shot to hell when I chopped the
third line of and it still acquired the extra new line. It's showing up
that way in the pickle, so it's either email's parseheaders (I'm thinking
there might be some subtle error with a "too long" line and how it's
interacting with "NL.join") or with the pickling process. Since the extra
nl is in the pickle, everyone downstream in the pipeline just inhereits the
problem.
(Why do I seem to always find the truly strange problems?)