[Bug 500695] [NEW] Some sendmail bounces not detected

Public bug reported: Some bounces from sendmail are not detected by the current bounce code. A regression was introduced in svn commit r4486: http://mailman.svn.sourceforge.net/viewvc/mailman/trunk/mailman/Mailman/Bouncers/SimpleMatch.py?view=diff&r1=4485&r2=4486 When ?P<addr> was changed to ?P<fulladdr>, the code in the process() function was not updated to match. ** Affects: mailman Importance: Undecided Status: New -- Some sendmail bounces not detected https://bugs.launchpad.net/bugs/500695 You received this bug notification because you are a member of Mailman Coders, which is subscribed to GNU Mailman.

** Attachment added: "Update SimpleMatch bouncer to check both addr and fulladdr regexes" http://launchpadlibrarian.net/37219816/SimpleMatch.patch -- Some sendmail bounces not detected https://bugs.launchpad.net/bugs/500695 You received this bug notification because you are a member of Mailman Coders, which is subscribed to GNU Mailman.

** Attachment added: "Add test to test_bounces.py" http://launchpadlibrarian.net/37219843/test_bounces.patch -- Some sendmail bounces not detected https://bugs.launchpad.net/bugs/500695 You received this bug notification because you are a member of Mailman Coders, which is subscribed to GNU Mailman.

** Attachment added: "sendmail bounce message not grokked by current code" http://launchpadlibrarian.net/37219852/sendmail_02.txt -- Some sendmail bounces not detected https://bugs.launchpad.net/bugs/500695 You received this bug notification because you are a member of Mailman Coders, which is subscribed to GNU Mailman.

** Changed in: mailman Status: New => Invalid -- Some sendmail bounces not detected https://bugs.launchpad.net/bugs/500695 You received this bug notification because you are a member of Mailman Coders, which is subscribed to GNU Mailman.

It is unclear whether or not there is a bug in SimpleMatch or not. The regexp itself is ambiguous. If in fact there is a problem, it should be resolved differently - e.g. by making two tests: (_c('the following addresses had'), _c('transcript of session follows'), _c(r'\(expanded from: <?(?P<addr>[^>)]*)>?\)')), (_c('the following addresses had'), _c('transcript of session follows'), _c(r'<(?P<addr>[^>]*)>')), although, I'm not convinced that wouldn't cause other problems. We never want to match the first address in '<a@example.com> (expanded from: b@example.net)' since that address is not the address we sent to. In any case, what is your problem? Your sample bounce message is recognized correctly by the DSN.py module as it should be (at least since 2.1.11 if not before). -- Some sendmail bounces not detected https://bugs.launchpad.net/bugs/500695 You received this bug notification because you are a member of Mailman Coders, which is subscribed to GNU Mailman.

Hmm, I wonder why these bounces were not caught by my 2.1.12 installation. I'll recheck against uncaught bounces on January 1st. In the meantime, this patch removes the useless fulladdr part of the regex since the process function doesn't reference it. ** Attachment added: "SimpleMatch.py.patch" http://launchpadlibrarian.net/37237387/SimpleMatch.py.patch -- Some sendmail bounces not detected https://bugs.launchpad.net/bugs/500695 You received this bug notification because you are a member of Mailman Coders, which is subscribed to GNU Mailman.

** Changed in: mailman Status: Invalid => Won't Fix -- Some sendmail bounces not detected https://bugs.launchpad.net/bugs/500695 You received this bug notification because you are a member of Mailman Coders, which is subscribed to GNU Mailman.

There are many places where named regexp groups are not referenced. This is not a bug per se. Your reference to Mailman Day (Jan 1) tells me that what you are seeing are not "uncaught" bounces, but rather are "unprocessed" bounces of messages from the site list, i.e., bounced password reminders. For more on this, see the FAQ at <http://wiki.list.org/x/aICB>. -- Some sendmail bounces not detected https://bugs.launchpad.net/bugs/500695 You received this bug notification because you are a member of Mailman Coders, which is subscribed to GNU Mailman.
participants (2)
-
Mark Sapiro
-
Matt Selsky