------- You are receiving this mail because: -------
You are on the CC list for the bug.
http://bugs.exim.org/show_bug.cgi?id=722
--- Comment #7 from Roman Odaisky <to.roma.from.eximbug@???> 2008-06-21 09:21:16 ---
Good point. But of the 20 forks in the Exim source code only one can happen
between string expansions (the one that processes user filters), all others
seem to occur either too early or too late.
I’m not concerned about safety in the cryptographic sense. I just want random
numbers that differ. For example, nhash of ${md5:$message_headers$message_body}
that was proposed above is the same when routing an incoming message that has
multiple recipients. I need to improve performance of a large mailing list
served by many low-end machines, so such situations are important.
It would be very convenient to write just $rand or ${eval10: $rand % 3} or
something like that, not having to guess which variables are available at this
point and which variables give the most entropy.
--
Configure bugmail:
http://bugs.exim.org/userprefs.cgi?tab=email
