Is anyone out there interested in doing some targeted testing?
You will need to
a) have a production-sized system on which you are prepared to
run bleeding-edge exim code
b) be prepared to build this code from git
c) have an operational load of many messages which, by some means,
appear in a short time in your spool and need delivery externally.
The destination could be all the same (the smarthost scenario)
or all over the show (internet targets, meaning that probably half
of it goes to Google... - or department servers or whatever).
Ideally we'll get takers for both of those scenarios.
The OP prompting this development didn't say how messages arrived
in his queue, but he did mention a queue size of 7 x 10^5 messages.
d) be prepared to comment on performance and limitations, and
compare with your previous configuration and procedures.
Anyone wanting to help, please contact me.
--
Cheers,
Jeremy
