[26137] in Athena Bugs

home help back first fref pref prev next nref lref last post

Re: sun4 [9.3.14]: mitmaildel

daemon@ATHENA.MIT.EDU (Robert A Basch)
Mon Oct 18 16:18:48 2004

Date: Mon, 18 Oct 2004 16:17:27 -0400
Message-Id: <200410182017.i9IKHRa8009171@anhedonia.mit.edu>
From: Robert A Basch <rbasch@mit.edu>
To: Jonathon Weiss <jweiss@mit.edu>
In-reply-to: "[26120] in Athena Bugs"
cc: Jacob Morzinski <jmorzins@mit.edu>
cc: bugs@mit.edu
Errors-To: bugs-bounces@mit.edu

As jmorzins pointed out, mitmailscan does not parse the date field in
a message itself, instead using a search query which tells the server
to parse the RFC-822 header, so the failure to recognize the dates of
certain messages is almost certainly a problem with the server.  There
are other search queries which have the server operate on its internal
received date for a message, but a design decision was not to provide
simple options for these, on the theory that this would prove more
confusing than useful.  However, you can query based on this internal
date via the --search-key option, e.g. --search-key="BEFORE 01-Sep-2004".
(The --before option uses the "SENTBEFORE" search query).

I cannot reproduce mitmaildel's failure to delete messages when trying
similar commands against the MIT IMAP servers.  The debug output you
provided suggests that the server does not recognize the given UIDs
as valid, for whatever reason.  (Unfortunately, the server returns
an "OK" response even for an unknown UID; an additional "FETCH"
response with the updated FLAGS value indicates that the STORE
command actually did something to a given UID).  Can you provide
--debug output for the mitmailscan command whose output was piped
into mitmaildel (without actually running mitmaildel)?  Also, do you
get the same results if you specify the UIDs manually to mitmaildel,
i.e. instead of using the "mitmailscan ... | xargs mitmaildel ..."
pipeline?

Bob

home help back first fref pref prev next nref lref last post