Re: [exim] How to disable spamassassin scans for outbound me…

Página Principal
Apagar esta mensagem
Responder a esta mensagem
Autor: Steve Garcia
Data:  
Para: Renaud Allard
CC: exim-users
Assunto: Re: [exim] How to disable spamassassin scans for outbound messages?
Renaud Allard wrote:
>
>
> On 1/9/09 8:07 PM, Steve Garcia wrote:
>> Renaud Allard wrote:
>>> On Thu, 08 Jan 2009 18:03:26 -0800 Steve Garcia<sgarcia@???>
>>> wrote
>>>
>>>> Eli Sand wrote:
>>>>> Steve wrote:
>>>>>
>>>>>> How would I go about disabling spamassassin scans for outbound
>>>>>> messages?
>>>>> Combine it with the "authenticated" acl test to see if your users have
>>>>> authenticated via SMTP (hopefully you use SMTP auth...).
>>>> No, they don't. They authenticate with PAM. They are all local to the
>>>> machine (via ssh), using alpine.
>>>>
>>> In this case, this is even easier, just put
>>> !hosts = localhost :
>>>
>>> Before your spamassassin ACL
>>
>> Aha! This sounds promising.
>>
>> I'm still wrapping my mind around how the ACLs work. I'm using a very
>> slightly modified Debian Etch exim4 config, and the spamassassin stuff
>> is all in "acl_check_data".
>>
>> As near as I can tell, all that I'm doing in acl_check_data is examining
>> the results of the spamassassin scan and making determinations of
>> message disposition based on score. Would I place the statement at the
>> *top* of that ACL?
>>
>> acl_check_data:
>>     !hosts = localhost :

>>
>> or before the acl?
>>
>> !hosts = localhost :
>> acl_check_data:
>>
>> or inside each clause that looks at the spamassassin results?
>>
>> warn
>>     message = X-Spam-Flag: YES
>>     !hosts = localhost :
>>     spam = Debian-exim

>>
>> The last looks like the most likely, am I correct?
>>
>
> In fact I did not explain it very well, but you are correct, you have to
> put it in every ACL that calls spamassassin. Like the one shown at the
> end of your post.
>


Perfect! I've set it up that way and it working great. Incoming emails
are correctly scanned, but outgoing messages are sent instantly, no
delay perceptible by the user at all.

Thanks!

--
Steve Garcia
sgarcia@???
Ignorance killed the cat; Curiosity was framed.