Interesting, I just tried the same test you ran and got he following 
(abbreviated) results:



...

Content-Disposition: inline

Content-Type: text/plain; charset=ISO-8859-1; format=flowed

**

test mail

**

test mail

**

EOF



Is it possible that one of your other plugins is causing this problem?  The 
only plugins i'm running are this one, check_loop, and rcpt_ok.  I guess it's 
possible that there could be a bug in the $transaction->set_body_start() method 
that could cause this problem, but it looks correct to me.  I'm running version 
0.32-6 as packaged by Debian.



Craig



On Fri, 25 May 2007 14:45:40 -0500, Tim Tsai <[EMAIL PROTECTED]> wrote:

> I am using swaks to do the testing, and the swaks output follows (I had

> examined the raw before message previously and it's fine - about to

> leave town so want to get this message out before I do).  Note that it

> always does it to the first line of the body.  I think Transaction.pm is

> suspect but I must be the only one that's seeing this, so it could be

> something related to my environment (Fedora Core 6).

> 

> Thanks!!

> 

> Tim

> 

> [EMAIL PROTECTED] swaks --to [EMAIL PROTECTED] --port 2525

> === Trying xxxx:2525...

> === Connected to xxxx.

> <-  220 xxxx ESMTP qpsmtpd 0.33-dev ready; send us your mail, but not

> your spam.

>  -> EHLO xxxx

> <-  250-xxxx Hi xxxx [xxx.xxx.xxx.xxx]

> <-  250-PIPELINING

> <-  250-8BITMIME

> <-  250 SIZE 50000000

>  -> MAIL FROM:<[EMAIL PROTECTED]>

> <-  250 <[EMAIL PROTECTED]>, sender OK - how exciting to get mail from you!

>  -> RCPT TO:<[EMAIL PROTECTED]>

> <-  250 <[EMAIL PROTECTED]>, recipient ok

>  -> DATA

> <-  354 go ahead

>  -> Date: Fri, 25 May 2007 14:36:57 -0500

>  -> To: [EMAIL PROTECTED]

>  -> From: [EMAIL PROTECTED]

>  -> Subject: test Fri, 25 May 2007 14:36:57 -0500

>  -> X-Mailer: swaks v20060621.0 jetmore.org/john/code/#swaks

>  ->

>  -> This is a test mailing

>  ->

>  -> .

> <-  250 Queued!

>  -> QUIT

> <-  221 xxxx closing connection. Have a wonderful day.

> 

> I inserted this code into spamassassin_filter:

> 

>   # Replace this transaction's message contents with the message returned

>   # from spamc.

>   $transaction->header($mail->head());

>   $transaction->set_body_start();

>   $transaction->body_write(join("", @{$mail->body()}));

> 

>   open MSG, ">/tmp/msg2";

>   print MSG $transaction->header()->as_string();

>   $transaction->body_resetpos();

>   print MSG "**\n";

>   print MSG join("", @{$mail->body()});

>   print MSG "**\n";

>   print MSG $transaction->body_as_string();

>   print MSG "**\n";

>   close MSG;

> 

> and the result in /tmp/msg2:

> 

> [EMAIL PROTECTED] more /tmp/msg2

> X-Spam-Checker-Version: SpamAssassin 3.1.8 (2007-02-13) on xxxx

> X-Spam-Level:

> X-Spam-Status: No, score=-100.4 required=5.0

> tests=ALL_TRUSTED,NO_REAL_NAME,

>         RAZOR2_CHECK,USER_IN_WHITELIST autolearn=unavailable version=3.1.8

> Received: from xxxx (HELO xxxx) (xxx.xxx.xxx.xxx) by

>  xxxx (qpsmtpd/0.33-dev) with ESMTP; Fri, 25 May 2007 14:36:58

>  -0500

> Date: Fri, 25 May 2007 14:36:57 -0500

> To: [EMAIL PROTECTED]

> From: [EMAIL PROTECTED]

> Subject: test Fri, 25 May 2007 14:36:57 -0500

> X-Mailer: swaks v20060621.0 jetmore.org/john/code/#swaks

> **

> This is a test mailing

> 

> **

> 

> **

> 

> Craig Gallek wrote:

>> Could you send an example of the problem (a before and after raw message

> maybe)?  The parsing of the spamc response is done by the perl

> Mail::Internet module, so as long as the response is well formed, you

> should get a valid message out of the plugin.

>>

>> Craig

>>

>> On Fri, 25 May 2007 11:24:06 -0500, Tim Tsai <[EMAIL PROTECTED]> wrote:

>>

>>> This is for Craig - I just joined the list and the web archive doesn't

>>> show e-mail addresses.

>>>

>>> I am trying out the spamassassin_filter plugin and it's exactly what I

>>> have been looking for.  We have a very specific spamassassin setup and

>>> moving the configuration off the plugin itself is a great idea.

>>>

>>> I have noticed that the version as posted online munges the first line

>>> of the body though.  Any ideas?  I think it starts when the result is

>>> read back from spamc.  I see the same problem with both .32 and Trunk

>>>

>>> Thanks!

>>>

>>> Tim

Reply via email to