Bug 4966 - [review] BSMTP broken - can lose mail
Summary: [review] BSMTP broken - can lose mail
Status: RESOLVED FIXED
Alias: None
Product: Spamassassin
Classification: Unclassified
Component: spamc/spamd (show other bugs)
Version: SVN Trunk (Latest Devel Version)
Hardware: All Linux
: P1 critical
Target Milestone: 3.1.4
Assignee: SpamAssassin Developer Mailing List
URL:
Whiteboard: can be committed
Keywords:
Depends on:
Blocks:
 
Reported: 2006-06-24 22:35 UTC by Brad "anomie" Jorsch
Modified: 2006-07-18 02:18 UTC (History)
0 users


Attachment Type Modified Status Actions Submitter/CLA Status
Patch to fix this bug. patch None Brad "anomie" Jorsch [HasCLA]
Add an attachment (proposed patch, testcase, etc.)

Note You need to log in before you can comment on or make changes to this bug.
Description Brad "anomie" Jorsch 2006-06-24 22:35:55 UTC 
Severity set to critical since mail will be bounced with common exim configurations.

The patch in revision 349405 breaks BSMTP message parsing, such that "\n.\n" at
EOF is no longer recognized as terminating the DATA section of the BSMTP message
("\n.\r\n" will still work, as will anything with more lines after the DATA
block). As this is what is output by exim4, all my mail has been being bounced
with a 421 error since I upgraded to 3.1.3.

I will attach a patch momentarily that seems to fix the problem.
Comment 1 Brad "anomie" Jorsch 2006-06-24 22:37:22 UTC 
Created attachment 3557 [details]
Patch to fix this bug.

Here's the patch.
Comment 2 Brad "anomie" Jorsch 2006-06-24 22:43:16 UTC 
Hrm, my "CLA Status" icon is incorrect. See bug 2667, comment #2.
Comment 3 Justin Mason 2006-07-12 15:15:45 UTC 
yeah, this needs to be fixed.  thanks for the patch!
also, updated yr CLA status.

applied to trunk:
Sending        spamc/libspamc.c
Transmitting file data .
Committed revision 421289.

committers, votes for 3.1.x please...
Comment 4 Theo Van Dinter 2006-07-18 03:44:43 UTC 
Hrm.  This section of code always seems to cause us problems.  If I follow
things, I think it looks good.

+1

We should have some tests to check for all these corner case things (see bug
4593,   etc.)
Comment 5 Daryl C. W. O'Shea 2006-07-18 03:50:16 UTC 
Just looking at this, too.  It seems to work.  I hate C.

+1
Comment 6 Justin Mason 2006-07-18 09:16:33 UTC 
thanks guys.

I agree we should have tests for this.  we really really should ;)
Comment 7 Justin Mason 2006-07-18 09:18:32 UTC 
oops!  forgot to commit it!!  on b3_1_0 as r423021.