[spambayes-bugs] [ spambayes-Bugs-747945 ] apparent corruption of Received lines

[SourceForge.net]

Bugs item #747945, was opened at 2003-06-03 18:09
Message generated for change (Comment added) made by anadelonbrin
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=498103&aid=747945&group_id=61702

Category: hammie
Group: None
Status: Open
Resolution: None
Priority: 5
Submitted By: Raphael Manfredi (rmanfredi)
Assigned to: Nobody/Anonymous (nobody)
Summary: apparent corruption of Received lines

Initial Comment:
It looks like the message header is corrupted by
hammiefilter.py.  Here's my evidence:

I'm running the 1.0a2 release.

I use "mailagent" to filter my mail, which is like
procmail (with a much cleaner code base and readable
rules) and I've set up the rules like that:

### Begin SPAM filtering section ###
       
{
    FEED ~/bin/spambayes/hammiefilter.py;
    RESYNC;
    REJECT;
};
         
X-Spambayes-Classification:
    /^spam/     { REJECT SPAM }
    /^ham/      { REJECT INITIAL }
    /^unsure/   { REJECT LOST };
       
# Dispatch to spam/lost, keep going if ham...
<LOST,SPAM>     { VACATION off; BIFF off; REJECT };   
<LOST>          { SAVE lost };
<SPAM>          { SAVE spam };
         
### End of SPAM filtering section ###

The "FEED" command pipes the message to the specified
scripts and gets the output back as the new message.
The RESYNC command reparses the headers, since they
might have been changed by the command (and indeed,
they are since X-Spambayes-Classification is added!).

However, here are the logs I see:

------------------------------
03/06/03 07:44:37 mailagent[277851]: FROM
<tvd_documentation at nai.com>
03/06/03 07:44:37 mailagent[277851]: RELAYED
hbs43-138.hbs.edu, orb.pobox.com, pop.free.fr, lyon.ram.loc
03/06/03 07:44:37 mailagent[277851]: ABOUT Approved
03/06/03 07:44:37 mailagent[277851]: MATCH on rule #68
in mode INITIAL
03/06/03 07:44:39 mailagent[277851]: FED [qm16938]
through '~/bin/spambayes/hammiefilter.py'
03/06/03 07:44:40 mailagent[277851]: WARNING no by in
first Received: line 'from localhost (mail at localhost
[127.0.0.1])h535iTik016935 for <ram at localhost>; Tue, 3
Jun 2003 07:44:30 +0200'
03/06/03 07:44:40 mailagent[277851]: weird Received:
line 'from localhost (mail at localhost
[127.0.0.1])h535iTik016935 for <ram at localhost>; Tue, 3
Jun 2003 07:44:30 +0200'
03/06/03 07:44:40 mailagent[277851]: RESYNCED [qm16938]
03/06/03 07:44:40 mailagent[277851]: REJECTED [qm16938]
in state INITIAL
03/06/03 07:44:40 mailagent[277851]: MATCH on rule #69
in mode INITIAL
03/06/03 07:44:40 mailagent[277851]: entering new state
SPAM
---------------------------------

Look at the WARNING about the absence of "by" in
Received.  The problem is that no such warning is
originally emitted, but it appears during RESYNC when
the headers are reparsed after the message went through
hammiefilter.py.

This indicates a corruption problem, since mailagent
uses the same code to reparse the headers during RESYNC
as it used originally, and no warning were emitted
before FEED was run.

----------------------------------------------------------------------

>Comment By: Tony Meyer (anadelonbrin)
Date: 2003-07-28 14:31

Message:
Logged In: YES 
user_id=552329

Could you try this with alpha4 and let us know if it still occurs?

----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=498103&aid=747945&group_id=61702