Richard P. Groenewegen wrote...
> My questions are:
>
> - what did I do wrong technically? As I said, I just added the line
> delay_warning = 24:1000h at the end of the main configuration
> section.
You're missing the units of time for the first item in the list. You need
to specify "24h" rather than just "24". Exim seems to have assumed that
you were defining a director or router.
> - what did I do wrong procedurally? Maybe a few test messages
> couldn't hurt, but why didn't I get any response that something
> was wrong when I did `eximconfig -i'? What precautions should I
> take when editing exim.conf?
I've never heard of eximconfig, sorry. When I edit exim.cf I HUP or
restart Exim, and I check to be sure that the process is present.
> - why does such a small error (in my mind) have such serious
> consequences? Why does exim not accept the messages and freeze
> them or something? An entry in the paniclog and otherwise not a
> shred of evidence as to what host was connecting is not very
> elegant.
It was a fundamental error in the configuration - Exim wouldn't know how
to decide (among other things) whether the message's recipient could be
verified. Not starting at all is the right thing to do, IMHO. The mail
will be deferred on other hosts until such time as yours becomes available
again.
--
Pete Naylor
