On Tue, 23 Sep 2003, Andreas Metzler wrote:
> If I put "testit,blah@???" in ~testit/.forward while foo.org
> cannot be delivered immidiately, I'll get dupes in /var/mail/testit.
>
> While the mail is undelivered to foo.org any queue-runner processing
> the mail will throw an additional copy into /var/mail/testit (the same
> applies if doing the same thing with exim filter).
This should NOT happen! (I am not quite that stupid. :-)
> I know that this is documented (see one_time option)
If you interpreted the documentation to mean that, then I need to
rewrite the documentation.
> but as I am no old hand with exim I did not realize how this limits
> the usability of .forward, i.e. that it cannot be _reliably_ used to
> make copies of messages unless both copies are always delivered in the
> same queue-run (both local destinations or both remote destinations
> routed to the same host).
This should not be the case. There's a lot of messy logic in Exim to
avoid duplicate deliveries; it gets complicated when .forward files
forward to the same local part, but it is all supposed to work, and
there are tests that are meant to check it. But, of course, bugs are
always present...
The only way I am going to be able to debug this is if you can provide
output from -d for an example. This will need two runs - the first and
the second delivery attempts.
--
Philip Hazel University of Cambridge Computing Service,
ph10@??? Cambridge, England. Phone: +44 1223 334714.
Get the Exim 4 book: http://www.uit.co.uk/exim-book
