I suggest you talk to the author..
On 2011-10-22, Aaron Jackson <jack...@msrce.howard.edu> wrote:
> Sent this to misc about a month ago, but really didn't get a response. Ports
> is listed as the maintainer for smtp-vilter. So, if anybody is interested?
>
> Aaron
>
> Begin forwarded message:
>
>> From: Aaron Jackson <jack...@msrce.howard.edu>
>> Subject: smtp-vilter bug/feature?
>> Date: September 6, 2011 2:41:30 PM EDT
>> To: m...@openbsd.org
>>
>> Irene killed my firewall/web server/mail sever, so I'm in the process of
>> recreating its setup with the current 4.9 release. I was running into a
>> problem with making smtp-vilter (installed from a package) work the way I
>> expected it to work. Specifically, the virus backend via clamav and the spam
>> backend via spam assassin worked fine but I could never get the attachment
>> backend to work. I kept getting the following message in maillog whenever I
>> sent an unwanted attachment:
>>
>> Sep 2 12:54:52 mushmouth smtp-vilter[32388]: failed to replace message body
>>
>> After banging my head for a couple of days (I did search google and the
>> mailing list without luck) I was able to trace the error message to line
>> 1817 of engine.c:
>>
>> if ((virus_strategy == STRATEGY_NOTIFY_RECIPIENT)
>> || (error_strategy == STRATEGY_NOTIFY_RECIPIENT)
>> || (spam_strategy == STRATEGY_NOTIFY_RECIPIENT)
>> || (unwanted_strategy == STRATEGY_NOTIFY_RECIPIENT))
>> desc.xxfi_flags |= SMFIF_CHGBODY;
>>
>> It turns out that for unwanted content, when smtp-vilter registers with
>> sendmail, it never sets the change body flag because
>> STRATEGY_NOTIFY_RECIPIENT is not an allowed strategy for unwanted content. I
>> made the following change then rebuilt and re-installed, and things seem to
>> work as expected.
>>
>> if ((virus_strategy == STRATEGY_NOTIFY_RECIPIENT)
>> || (error_strategy == STRATEGY_NOTIFY_RECIPIENT)
>> || (spam_strategy == STRATEGY_NOTIFY_RECIPIENT)
>> || (unwanted_strategy == STRATEGY_MARK))
>> desc.xxfi_flags |= SMFIF_CHGBODY;
>>
>> It seems like a bug to me, but then again the code is a bit complex and I
>> don't fully understand it. I was just wondering if anybody had any thoughts
>> about this "fix." I don't know if this will effect anything. Anyway, reading
>> code is very educational and I did learn a few things in the process.
>>
>> Aaron
>
>
