[ On Saturday, March 29, 2003 at 19:51:31 (+0100), David Saez wrote: ]
> Subject: Re: [Exim] sender callout failing
>
> Maybe it will be nice to have an option for callout to specify the
> envelope sender to use when doing a callout verification as now this
> control does not really test if the address exists or not (as the
> fact that the server is not prepared to accept bounce messages does
> not mean that the testing address exists or not). Testing the envelope
> sender with callout using a envelope sender is ok,
To say the least! :-)
To the extent that active sender address testing is worthwhile it must
only be done with "<>" as the test's sender address. Think about it.
Why do you want to verify the validity of a sender address? When is the
sender address used? What sender address "MUST" be used when the
sender address under test becomes a recipient address?
A server that won't accept bounce messages _is_ broken and _all_ sender
addresses routed to that server are, _by_definition_, invalid!
> but maybe it will
> be better to relax it when testing addresses from the header.
Perhaps -- it all depends on why you're doing the testing, i.e. what
policy you are trying to implement.
Note that actively testing addresses from headers is very unlikely to
ever have any valid purpose and _will_ break many other things in many
very unexpected ways. Please don't even consider actively testing
addresses from the message body (including the RFC-[2]822 headers).
--
Greg A. Woods
+1 416 218-0098; <g.a.woods@???>; <woods@???>
Planix, Inc. <woods@???>; VE3TCP; Secrets of the Weird <woods@???>
