David Woodhouse wrote: > On Sun, 2005-08-21 at 21:10 -0700, Steve Brown wrote:
>
>>So why all the warnings in the docs about not using an address in the DSN
>>and talk about callback loops?
>
>
> That warning is appropriate for normal sender verification callouts. Not
> for postmaster or header_sender callouts, which should have a non-empty
> local sender (which doesn't trigger callouts of its own) -- or for
> recipient callouts, which should use the same reverse-path as the mail
> causing the callout.
So, I'm confused. Surely the avoidance of loops always requires callout
to use a null sender, and postmaster and header-sender callouts are
in contravention of that? I'd say that the warning is *always*
appropriate. Your "doesn't trigger callouts of its own" is not
under your control; it's the remote end which might be deciding to
make yet another callout.
