Here is the result of the test. The callout was cached, so I suppressed
the cache before sending the message. I also removed greylisting because
it would have been unpractical for debugging.
I included the full file just in case.
On 2/13/2006, "Philip Hazel" <ph10@???> wrote:
>On Sat, 11 Feb 2006, Renaud Allard wrote:
>
>> I can do any tests you wish, but if you could tell me what would be
>> useful to you, it would be better.
>
>I have been checking the Exim test suite, and it already contains a test
>for callouts that fail at RCPT, and (needless to say) the message that
>is given is the correct one. So it is not an "obvious" bug. It must
>depend on some particular circumstances that are different on your
>system. The test that is needed is a debugging test that forces Exim to
>do the callout. You can do this by running
>
>exim -d -bhc 1.2.3.4 2>/tmp/eximstderr
>ehlo a.domain
>mail from:<mail.llorien.org-1139481506-testing@???>
>rcpt to:<yyrutebjkuwpbh@???>
>quit
>
>This will simulate incoming mail from 1.2.3.4 (unimportant), but will do
>the callout as it would normally do. The debugging output on
>/tmp/eximstderr should show everything that is going on.
>
>There is one possible problem: the callout cache. If this callout check
>has happened recently, Exim will get the answer from the cache and not
>do the callout, but let's see if that happens before trying anything
>else.
>
>--
>Philip Hazel University of Cambridge Computing Service
>Get the Exim 4 book: http://www.uit.co.uk/exim-boo
