postfix message_reject_characters

[Gunnar Wrobel]

Zitat von "Andrew J. Kopciuch" <akopciuch at bddf.ca>:

> hi :
>
> I had an automated vcs commit email today that never came through.  When
> another developer told me about it, and I looked into it I found it was
> rejected : 5.7.1 disallowed character
>
> I understand the postfix configuration in main.cf
>
> message_reject_characters = \0
>
> I read through the comments, and also kolab/issue3594.  It was suggested
> originally to use message_strip_characters = \0, but was decided on reject
> instead.
>
> I am wondering if there is anything wrong with using strip, instead of reject
> here?  If the null byte is removed, there should be no problems for cyrus
> with it right?
>
> I couldn't find any argument for reject over strip, other than it would not
> changing the message content.  But what content would actually require those
> null bytes?  Does anyone know of any reason to not just strip those out, and
> let the messages through?

This simply depends on your specific situation.

The RFC clearly says that the message is invalid so the problem is  
caused by the sender. So the long-term and the correct solution will  
be: "Fix the sender!".

For the Kolab server it doesn't make much sense to allow it to deal  
with all types of crappy input. At least not in case there are just a  
few senders that misbehave. Of course we sometimes do add hacks in  
case the misbehaving client is named "Outlook" :) But other than that  
the long term strategy must be to adhere to the standards.

Of course you still have the choice for your local installation to  
decide the stripping hurts you much less than fixing the one sender  
that causes problems.

Cheers,

Gunnar
>
>
> Andy
>



--
Gunnar Wrobel
Developer, Kolab Systems AG

e: wrobel at kolabsys.com
t: +49 700 6245 0000
w: http://www.kolabsys.com

pgp: 9703 43BE

----------------------------------------------------------------
This message was sent using IMP, the Internet Messaging Program.