On Sun, Dec 21, 2003 at 09:58:09PM -0500, Eli said:
> No I don't believe that is it, because in the output of running exim with
> -d, it shows it evaluating the condition but it just says that the condition
> failed to trigger the warn statement (thus, verification has passed):
>
> 28805 using ACL "acl_check_helo"
> 28805 processing "warn"
> 28805 check verify = helo
> 28805 warn: condition test failed
> 28805 processing "accept"
>
> See, it processed the warn statement, did the verification (which was
> actually done earlier):
>
> Verification failed, thus meaning that "verify = helo" should return a false
> value. So I don't think the verification stuff is working incorrectly, I
> think it's just the final result code that's stored that is either incorrect
> (or contrary to what documentation says it should be), or that "verify =
> helo" is working backwards or only one-wayish (to always return true).
Don't you want:
!verify = helo
? It seems to me that the statement reads:
if (verify = helo) ; then warn "...."; fi
(rough logic). So when verify = helo fails, you get no warn statement.
Shot in the dark and all,
--
--------------------------------------------------------------------------
| Stephen Gran | I know not how I came into this, shall |
| steve@??? | I call it a dying life or a living |
| http://www.lobefin.net/~steve | death? -- St. Augustine |
--------------------------------------------------------------------------
