qrunner hung by spam

Bug #265907 reported by Greglindahl-users
2
Affects Status Importance Assigned to Milestone
GNU Mailman
New
High
Unassigned

Bug Description

I realized no mail had gone out for 6 days, and saw
that qrunner was barfing on SOMETHING. This is with
Mailman 2.0.6. Even
though I'm behind the times, I'd bet this bug could be
in 2.1.X -- I just don't know enough Python to see
what's going on. I blew away a bunch of files in my
qfiles directory; I didn't save the bad message.

Apr 24 00:39:01 2003 qrunner(2397): Traceback
(innermost last):
Apr 24 00:39:01 2003 qrunner(2397): File
"/home/mailman/cron/qrunner", line 282, in ?
Apr 24 00:39:01 2003 qrunner(2397): kids = main(lock)
Apr 24 00:39:01 2003 qrunner(2397): File
"/home/mailman/cron/qrunner", line 252, in main
Apr 24 00:39:01 2003 qrunner(2397): keepqueued =
dispose_message(mlist, msg, msgdata)
Apr 24 00:39:01 2003 qrunner(2397): File
"/home/mailman/cron/qrunner", line 121, in dispose_message
Apr 24 00:39:01 2003 qrunner(2397): if
BouncerAPI.ScanMessages(mlist, mimemsg):
Apr 24 00:39:01 2003 qrunner(2397): File
"/home/mailman/Mailman/Bouncers/BouncerAPI.py", line
59, in ScanMessages
Apr 24 00:39:01 2003 qrunner(2397): addrs = func(msg)
Apr 24 00:39:01 2003 qrunner(2397): File
"/home/mailman/Mailman/Bouncers/Postfix.py", line 39,
in process
Apr 24 00:39:01 2003 qrunner(2397): more =
mfile.next()
Apr 24 00:39:01 2003 qrunner(2397): File
"/usr/lib/python1.5/multifile.py", line 121, in next
Apr 24 00:39:01 2003 qrunner(2397): while
self.readline(): pass
Apr 24 00:39:01 2003 qrunner(2397): File
"/usr/lib/python1.5/multifile.py", line 93, in readline
Apr 24 00:39:01 2003 qrunner(2397): if marker ==
self.section_divider(sep):
Apr 24 00:39:01 2003 qrunner(2397): File
"/usr/lib/python1.5/multifile.py", line 157, in
section_divider
Apr 24 00:39:01 2003 qrunner(2397): return "--" + str
Apr 24 00:39:01 2003 qrunner(2397): TypeError :
illegal argument type for built-in operation

[http://sourceforge.net/tracker/index.php?func=detail&aid=726736&group_id=103&atid=100103]

Revision history for this message
Ber-users (ber-users) wrote :

I also encountered that bug once at a site of a friend.
Take a look at Mailman/Bouncers/Postfix.py
and especially the line with
       boundary = msg.getparam('boundary')
this can result in boundary being "None"
which is then not treated properly.
So add the following check below this with
the correct indentation:
        if not boundary:
                        return None

If I remember correctly this was handled completely differently
in Mailman 2.1 so I believe that this bug is limited to
Mailman 2.0.x.

Revision history for this message
Ber-users (ber-users) wrote :

I've submitted a patch for this to the patch tracker 783651
https://sourceforge.net/tracker/index.php?func=detail&aid=783651&group_id=103&atid=300103

Revision history for this message
Greglindahl-users (greglindahl-users) wrote :

Thanks -- I see that you fixed all of them in your patch.

This should be added to the 2.0.X series.

Revision history for this message
Ber-users (ber-users) wrote :

Greg, maybe you should close the bug
in addition to send another email to the mailman lists
to ask for an 2.0.14 release.
An alternative would be to make a patch release,
this bug certainly is critical for 2.0.x.
On the other hand it most likely does not affect 2.1.x.

Revision history for this message
Lindahl (lindahl) wrote :

It's not the reporter's place to close a bug report --
thanks for the fix, but the Mailman owners have to close the
bug. I'm afraid that the mailman community doesn't spend
much time keeping their sourceforge bug list clean.

I've given a couple of other people the fix, it's an
extremely common situation today.

To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.