>>>>> "BS" == Bill Silvert <email@example.com> writes:
BS> This morning I discovered that one of our lists had been truncated from
BS> 375 members to 1 (one), and all I can identify was a routine
BS> subscription process. This is with 1.94b5, perl 5.003.
I've given code to torture Majordomo endlessly, but I don't think anyone
has the fortitude to run it except me. I could not duplicate any list
BS> although the erasures may have occured during the period when our
BS> system was hung because of too many processes, as I described in an
BS> earlier message.
Well, all bets are off when your system hangs.
BS> It seems to me that MJ should be able to test for this sort of
BS> unbalanced brackets as part of the subscription process, given the
BS> power of perl as a pattern-matching language.
I wrote this code too late for inclusion in 1.94. I think Evan blasted it
with several addresses of this type when I put the test code up on a
server, but I took it down to work on the betas.
BS> I have run into a lot pf problems which turned out to be the result of
BS> sendmail (Berkeley 8.25) changing all of the aliases to lower case.
Yep. It's a pain. It doesn't always do it, either. Some versions smash
case only when queuing.
BS> but it should be mentioned in the documentation, and I hope that MJ and
BS> associated programs (like digest) can learn to stop using upper-case
cough FAQ cough.
We don't recommend using any arguments to resend except for the required
ones (which are in lower case). That was the whole point behind the config
files. I'm not familiar with how digest works, but soon it won't matter
because 1.95 will send the message to archive and digest internally
(because of the problems with headers and footers).