Philip Hazel wrote:
>On Thu, 18 Mar 2004, Toralf Lund wrote:
>
>
>
>>I actually tried to draw a somewhat simplified picture with the above
>>example; what we *really* have is
>>
>> 1.
>>
>>scu-archive: "|/usr/bin/mhonarc <and so on>"
>>in /etc/aliases
>>
>> 2.
>>
>>~ks/.forward containing
>># Exim filter
>>unseen noerror pipe /usr/bin/mhonarc <and so on>
>>
>>and similar for other users, i.e. most pipes for real users are set up
>>via .forward so it could be OK to run the ones from aliases under a
>>special id. Maybe I would be defining non-standard behaviour for
>>/etc/aliases that way, though, and I'm not 100% happy with that...
>>
>>
>
>What you have now described is the most common way of running things,
>and yes, the answer is just to run the ones from alias files under a
>special id via a different transport.
> OK.
> In fact, your previous description
>would not have worked the way you said it did. Putting a real login name
>in /etc/aliases does not automatically set up that user for the
>transport. It is the check_local_user option on your .forward router
>that does that. (I didn't point this out before because I suspected you
>were not telling the whole truth. :-)
>
> I actually new it already, or at least, I started to suspect this after
I sent the description. Like I said, I tried to present a
compressed/simplified version of the setup, but I wasn't really thinking
when I did it...
BTW, how will pipes in aliases will work with the default config file?
Or will they work at all? What user is expected to be active at the
point of alias expansion?
>It would be even better if the user could be looked up using
>> getpwnam() like (as far as I understand) check_local_user does,
>> though.
>
>
>
>Check out the "passwd" lookup type.
>
> Ah, yes. I didn't notice it earlier because I was reading the wrong
version of the spec ;-( (Version 4.00 instead of 4.30.)
