Re: [exim] Recipient verify only for non-authenticated users

Graeme Fowler wrote:
> Bill, all
>
> On Fri, 2011-10-28 at 09:07 +0000, W B Hacker wrote:
>> *heavy sigh*
>
> Quite. In order to achieve some clarity, please let me say:
>
>> A triggered 'accept' is not 'permanent' until end of DATA. Period.
>
> This assertion is wrong.
>
> An "accept" verb in an ACL simply means that the ACL in question has
> accepted *whatever it is processing* at that time.
> An "accept" in any other ACL than the final one in the flow simply
> passes control to the NEXT ACL.
>
> In the case of DATA (which is usually the final one), it is a permanent
> acceptance of a message. In *any other ACL*, control passes to the next
> ACL.
>
>> A triggered 'deny' class verb is 'permanent' AT ONCE. WHEREVER it is.
>
> Indeed. For whichever part of the ACL flow you're in, a "deny" is
> immediate.
>
> Please don't muddy the waters.
>
> Graeme
>
>

Given that - absent a link fail, time-out, or other earlier bail-out -
(which by definition cannot be an 'accept') DATA is ALWAYS the 'final
one' in the phases of an smtp session - whether you have any
customization to the default handling or not...

How can you say that the statement that:

>> A triggered 'accept' is not 'permanent' until end of DATA. Period.

...is 'wrong'..

.. and THEN go on to say the same thing!

> In the case of DATA (which is usually the final one), it is a
> permanent acceptance of a message.

The side note:

> An "accept" in any other ACL than the final one in the flow simply
> passes control to the NEXT ACL.
>

Is stipulated. Understood. Inherent. Nature of the beast.
Near-as-dammit unavoidable, even.

Ergo inserting it is irrelevant. Red Herring.

THAT is 'muddying the waters'.

Surely you did not interpret any part of this to have implied that an
earlier-phase 'accept' was somehow infectious, parasitical, viral, or
....wot? child-bearing? ... immortal, even?

Bill
--
韓家標