[Mailman-Users] One message in ./qfiles hold up entire show (qrunner)

[Christopher Schulte]

I noticed messages were not being processed by mailman to my MTA.  With MTA 
(postfix) and mailman (2.0.6) logs, I narrowed it down to one bounced 
message from a remote site that caused this error when qrunner tried to run 
and clear out the mailman queue:

Sep 04 15:16:09 2001 qrunner(20389): Traceback (innermost last):
Sep 04 15:16:09 2001 qrunner(20389):   File "/home/mailman/cron/qrunner", 
line 282, in ?
Sep 04 15:16:09 2001 qrunner(20389):      kids = main(lock)
Sep 04 15:16:09 2001 qrunner(20389):   File "/home/mailman/cron/qrunner", 
line 252, in main
Sep 04 15:16:09 2001 qrunner(20389):      keepqueued = 
dispose_message(mlist, msg, msgdata)
Sep 04 15:16:09 2001 qrunner(20389):   File "/home/mailman/cron/qrunner", 
line 121, in dispose_message
Sep 04 15:16:09 2001 qrunner(20389):      if BouncerAPI.ScanMessages(mlist, 
mimemsg):
Sep 04 15:16:09 2001 qrunner(20389):   File 
"/home/mailman/Mailman/Bouncers/BouncerAPI.py", line 59, in ScanMessages
Sep 04 15:16:09 2001 qrunner(20389):      addrs = func(msg)
Sep 04 15:16:09 2001 qrunner(20389):   File 
"/home/mailman/Mailman/Bouncers/DSN.py", line 46, in process
Sep 04 15:16:09 2001 qrunner(20389):      if string.lower(msg.gettype()) <> 
'multipart/report' or \
Sep 04 15:16:09 2001 qrunner(20389): TypeError :  read-only character 
buffer, None

When I moved the offending messages out of the queue, qrunner ran just fine 
and 12,000 queued messages went out.

I'm running Mailman 2.0.6 and python:
Python 1.5.2 (#2, Jul 11 2000, 07:30:30)  [GCC 2.95.2 19991024 (release)] 
on freebsd4.

Because I don't want to spam the list with tons of message text, the 
messages that caused mailman to hang are located here:

http://smtp.schulte.org/smtp/mailman-bad-messages/

I will leave these up for as long as their required to debug this problem.

Is this a problem which I can fix, or perhaps some bug?  I see similar 
message in the list archive from 6-12 months ago, but nothing new.

Thanks,

--chris