On Fri, 18 Aug 2006, Richard Hoyle wrote:
> I hadn't realised that exim actually exits, refusing to run if it
> receives a debugging request as an unprivileged user.
... and I had forgotten that too. It is treated as "you are trying to
request something that only privileged users may request - go away".
> It probably would be good if the autoreply transport only updated the
> 'once' db AFTER it had successfully submitted the message, but I guess
> it's fair to assume that local deliveries don't fail under normal
> circumstances.
The comment in the code says this:
/* Update the "sent to" log whatever the yield. This errs on the side of
missing out a message rather than risking sending more than one. We either have
cache_fd set to a fixed size, circular buffer file, or dbm_file set to an open
DBM file (or neither, if "once" is not set). */
The code has been untouched for a very long time. I can no longer
remember the detailed considerations that led to its being implemented
in this order.
--
Philip Hazel University of Cambridge Computing Service
Get the Exim 4 book: http://www.uit.co.uk/exim-book
