On 03/May/10 16:07, Matus UHLAR - fantomas wrote:
>>>  On 01/May/10 00:03, Sam Varshavchik wrote:
>>>>  No, you don't need smtpfilter in that case. Just install rcptfilter that
>>>>  terminates with a 0 exit code.
> [...]
>>>>  Since the rcptfilter never returns 99, the smtpfilter is never checked.
>
>>  Alessandro Vesely writes:
>>>  Fine. What I still don't understand, is how should the other users say
>>>  that they /want/ the global filter to be run. I've always used
>>>  "allfilters", so I'm not familiar with this at all.
>
> On 01.05.10 09:49, Sam Varshavchik wrote:
>>  This is a three-step process.
>>
>>  First, each local recipient's rcptfilter runs, if it exists. A
>>  nonexistent rcptfilter is considered a whitelisted result, for these
>>  purposes. The first recipient's result determines whether the entire
>>  message is considered to be whitelisted or not.
>>
>>  In the second step, courierfilters are run. This stage uses
>>  courierfilters installed in either the filters or the allfilters
>>  directory, depending upon whether the message is whitelisted, or not.
>>
>>  In the third step, if the message is not whitelisted, each local
>>  recipient's smtpfilter runs to selective reject the message for that
>>  recipient only.
>
> now back to my question: do rcptfilters (or smtpfilters? I don't mind much)
> have any chance to know who the other recipients are? or at least pass any
> info from first recipient to others? So when first recipient is non-abuse,
> only other non-abuse recipients would be passed and vice versa?
>
> That way I could apply basically different filtering for abuse and other
> users, which is what I need...

Perhaps the opposite than I initially thought: if you install an 
rcptfilter for abuse that returns 99, mail to abuse would go through 
global non-all filters, if any. Only users who also install an 
rcptfilter that returns 99 may be co-recipient with abuse that way. In 
all probability, maildrop will look for abuse's smtpfilter, and assume 
an exit code of 0 (=accept) if not found.

-- 


------------------------------------------------------------------------------
_______________________________________________
courier-users mailing list
[email protected]
Unsubscribe: https://lists.sourceforge.net/lists/listinfo/courier-users

Reply via email to