On Mon, 26 Feb 2001, Peter Galbavy wrote:
> > Reasonable - until you make an error and the expansion fails.
>
> Gary Palmer pointed out another problem in an earlier reply, but on general
> philosophy, isn't the chance of a expansion failing one that can happen
> anywhere ? Or are you saying that a "fail" condition would react
> particularly badly with rbl_domains ?
I was merely noting that you hadn't written any code to cope with the
failure case. The expansion would return NULL, and your code would
(IIRC) behave as if rbl_domains were not set.
> Now that becomes moot anyhow, due to
> the problem that Gary pointed out with the RBL check being done at the TCP
> accept bit.
It's all going to be different in Exim 4, for exactly this sort of
reason.
--
Philip Hazel University of Cambridge Computing Service,
ph10@??? Cambridge, England. Phone: +44 1223 334714.
