RE: last dev HM (CVS)

Got it - I was stupidly checking to see whether the msgid was a duplicate even if the current message didn't have a msgid.

I've commited a fixed version of struct.c to CVS.

Paul

On Tue, 8 Jun 1999, Otis Gospodnetic wrote:

> I'll try it by hand later (I have to run now).
> The archive is: http://www.infojump.com/lists/archives/zine-talk/old/
> And the first message that has null Message-ID is 0000.html
> (http://www.infojump.com/lists/archives/zine-talk/old/0000.html - third from
> the bottom).
> You will also notice some weird duplicate messages in that archive - I got
> them after I switched from 1.02 to 2.0b3.
>
> Thanks,
>
> Otis
>
> > > It doesn't dump any core (no core in the archive dir, no core
> > in dir where
> > > hypermail is, I'm not sure where else it could be).
> >
> > Sometimes the system won't leave dump core if its not sure who the real
> > user is or if its running in a setuid environment - this may be the case
> > when delivering through an alias. Could you try saving one of your input
> > messages and running hypermail by hand? (but then you were doing the
> > sendmail run as root so it _should_ be OK - try by hand anyway).
> >
> > > > It dies with signal 13, which is (SIG)PIPE...
> > > I'm running this on Red Hat 5.2, kernel 2.0.36, with sendmail 8.8.7.
Received on Tue 08 Jun 1999 04:26:06 PM GMT