On Wed, 18 Oct 2000, Ian Jackson wrote:
> 1. Is it safe to send an Exim delivery process a SIGTERM ?
I would hope so. The only risk you run is if you kill a delivery between
the time it has actually delivered a message and the time it written
this fact to the journal file - you would then get a duplicate delivery
later.
> it would be nice to have a documented way
> to stop an in-progress delivery attempt.
Just kill the process. Exim should ignore SIGTERM during local
deliveries, incidentally (so as, hopefully, not to leave half-written
files).
> 2. If it is safe, would it be a good idea to patch deliver.c not to
> freeze messages under these conditions ?
I'm slightly confused because I didn't think Exim did freeze messages in
this situation. At least, not in general.
--
Philip Hazel University of Cambridge Computing Service,
ph10@??? Cambridge, England. Phone: +44 1223 334714.
