> There's really no way I can think of to prevent Exim from at least
> creating a spool file at least for a very short period of time.
>
> However, one alternative would be to add a condition on your router
> that says:
>
> condition = ${if first_delivery}
>
> The second time the message is processed, regardless of the retry time,
> the router will pass and will presumably hit the end of your list of
> routers and fail.
You might be able to mix that in with a 2nd router that would catch the
failures from the router before and do a local delivery to /dev/null, which
would (I presume) defeat any local queuing for retries or the like since as
far as Exim would be concerned, it was delivered successfully.
Though, I suppose that if there were more than just this one reason to kill
a message before it's ever completed it's lifecycle in Exim, perhaps
allowing routers/transports the ability to return a value that denotes the
message should be stopped in it's tracks and all corresponding files removed
might be a slightly nicer solution.
Eli.
