Bug 4652 - [review] __OIMO_MSGID header pattern is not correct
Summary: [review] __OIMO_MSGID header pattern is not correct
Status: RESOLVED FIXED
Alias: None
Product: Spamassassin
Classification: Unclassified
Component: Rules (show other bugs)
Version: 3.0.0
Hardware: All All
: P3 normal
Target Milestone: 3.1.1
Assignee: Daryl C. W. O'Shea
URL:
Whiteboard: ready
Keywords: triage
Depends on:
Blocks:
 
Reported: 2005-10-29 05:08 UTC by Mark Noyes
Modified: 2006-01-24 17:10 UTC (History)
1 user (show)


Attachment Type Modified Status Actions Submitter/CLA Status
Sample email in mbx format that reproduces the problem text/plain None Bob Menschel [HasCLA]
3.1 patch patch None Daryl C. W. O'Shea [HasCLA]
3.1 patch patch None Daryl C. W. O'Shea [HasCLA]
Show Obsolete (1) 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 Mark Noyes 2005-10-29 05:08:04 UTC 
The rules file rules/20_ratware.cf contains a pattern
  header __OIMO_MSGID  MESSAGEID =~ /^[A-P]{26}A[AB]\.[-_\w]+\S+>$/m

The version of Outlook 2000 installed on my machine generates Message-id 
headers such as the following:
Message-id: <LOBBJIAKABGDMAEIEABEMEOJAPAC.mvnoyes@verizon.net>
The "[AB]" in the message id pattern needs to be "[ABC]" in order to match 
correctly.  As a result, email from my system frequently gets incorrectly 
categorized as spam.
Comment 1 Loren Wilton 2005-10-29 09:41:39 UTC 
Hum.  Looks like it might be possible to optimize that pattern too.  The ending 
[stuff]+\S+ is a pattern that could easily have N different boundaries and all 
be valid.  Since the username probably doesn't start with - or _, maybe 
something like \w\S+ would be sufficient?  Or maybe just \S?
Comment 2 Bob Menschel 2005-10-30 17:46:40 UTC 
Mark, You flag this as "all" hardware and O/s, but version 2.0 of SA? 

Can you please let us know which version of MS Outlook 2000 this is? 

Can you send an email to me at RMSA@Menschel.net, from that Outlook 2k system,
so I can attach a sample email with headers as an example we can test against? 

Thanks.
Comment 3 Bob Menschel 2005-10-31 05:01:38 UTC 
Received requested email.  SA version 3.0.x, Outlook version 9.0.0.2711. Will
attach the email
Comment 4 Bob Menschel 2005-10-31 06:04:08 UTC 
Created attachment 3220 [details]
Sample email in mbx format that reproduces the problem
Comment 5 Daryl C. W. O'Shea 2005-11-18 02:49:26 UTC 
I've tested some modifications to the rule (the originally posted rule was
mis-copied, BTW):

# original rule
header T_OIMO_MSGID_0   MESSAGEID =~ /^<[A-P]{26}A[AB]\.[-_\w.]+\@\S+>$/m

# original rule with what I assume was meant to be an escaped period
header T_OIMO_MSGID_1   MESSAGEID =~ /^<[A-P]{26}A[AB]\.[-\w\.]+\@\S+>$/m

# modified rule without escaped period
header T_OIMO_MSGID_2   MESSAGEID =~ /^<[A-P]{26}A[ABC]\.[-\w.]+\@\S+>$/m

# modified rule with escaped period
header T_OIMO_MSGID_3   MESSAGEID =~ /^<[A-P]{26}A[ABC]\.[-\w\.]+\@\S+>$/m


Mass-check results:
http://buildbot.spamassassin.org/ruleqa/ruleqa?s_defcorpus=on&daterev=20051117-r345134-n&rule=%2F%5ET_OIMO&s_zero=on&s_detail=checked+&g=Change

T_OIMO_MSGID_3 has a better S/O, hitting lots more ham with a negligible
increase in spam hits.
Comment 6 Daryl C. W. O'Shea 2005-11-18 03:13:42 UTC 
Created attachment 3257 [details]
3.1 patch

trunk r345404
Comment 7 Justin Mason 2005-11-18 06:56:16 UTC 
+1
Comment 8 Daryl C. W. O'Shea 2005-11-24 02:49:58 UTC 
Created attachment 3272 [details]
3.1 patch

duh! dots aren't character class metacharacters

This patch removes the extraneous backslash.  It doesn't affect mass-check
results.

trunk r348610
Comment 9 Justin Mason 2005-12-13 20:24:09 UTC 
+1
Comment 10 Duncan Findlay 2006-01-25 02:01:30 UTC 
+1
Comment 11 Daryl C. W. O'Shea 2006-01-25 02:10:00 UTC 
Thanks Duncan!

Sending        rules/20_ratware.cf
Transmitting file data .
Committed revision 372069.