GNU bug report logs - #4358
23.1.50; Arithmetic overflow errors break imap

Previous Next

Package: emacs;

Reported by: James Cloos <cloos <at> jhcloos.com>

Date: Sun, 6 Sep 2009 13:55:05 UTC

Severity: normal

Done: Chong Yidong <cyd <at> stupidchicken.com>

Bug is archived. No further changes may be made.

Full log


Message #20 received at 4358 <at> emacsbugs.donarmstrong.com (full text, mbox):

From: James Cloos <cloos <at> jhcloos.com>
To: 4358 <at> debbugs.gnu.org
Cc: Stefan Monnier <monnier <at> IRO.UMontreal.CA>
Subject: Re: bug#4358: 23.1.50; Arithmetic overflow errors break imap
Date: Wed, 09 Sep 2009 10:34:43 -0400
Having patched the bug in the imapd¹ is use, imap.el no longer errors
out for me.

As such, fixing the imap.el bug is no longer critical for me.

I hope, however, that the bug can remain open, unless there is already
a big integer bug open for gnus with which this can merge.


1] The imapd uses a sql db to store the mail; it implements the imap
   status command by doing a union query to get the number of messages
   in the specified folder, how many are seen, etc.  It then does an
   unseen = exists - seen, using unsigned integers, to get the unseen
   count.  It also made an invalid assumption about how the query
   result set would be ordered; it assumed that, in the absence of an
   explicit order by clause, the results would be implicitly ordered
   by the first column, as if order by 1 had been specified.

   That happened to work with mysql and earlier version of postgres,
   but does not happen with postgresql 8.4.  Adding an explicit
   order by 1 to the query fixed that imapd bug.

-JimC
-- 
James Cloos <cloos <at> jhcloos.com>         OpenPGP: 1024D/ED7DAEA6



This bug report was last modified 14 years and 331 days ago.

Previous Next


GNU bug tracking system
Copyright (C) 1999 Darren O. Benham, 1997,2003 nCipher Corporation Ltd, 1994-97 Ian Jackson.