On Tue, 6 Feb 2001, John Horne wrote:
> A complicated one this one! With exim 3.22 it seems that if we send a
> message via a director which uses the 'headers_add' option, then it works if
> the message is subsequently 'forced' with the -q option, but not -qq.
It will take some time to check this completely, and I can't do that
just at the moment, but ...
> Unfortunately the dreaded 'one_time' is in there as well. Sending a test
> message with the '-d9 -qqRff' option showed it all going well, and then:
I'm fairly sure it is the "one_time" option that is doing this. Have you
tried without? The fact that one_time is set means that it won't be
doing the directing the second time round. Hmm. I suppose that means
that "one_time" should be disabled for the first pass of a -qq delivery.
I'd overlooked that - a typical example of a new feature interacting
badly with another feature.
If you find it works with one_time disabled, you might like to see if
the patch below solves the original problem. Please let me know if it does.
--
Philip Hazel University of Cambridge Computing Service,
ph10@??? Cambridge, England. Phone: +44 1223 334714.
*** directors/df_generated.c Fri Jan 5 14:22:17 2001
--- directors/df_generated.c Fri Feb 9 10:26:36 2001
***************
*** 66,72 ****
next->next = *addr_new;
*addr_new = next;
! if (ob->one_time)
{
for (parent = addr; parent->parent != NULL; parent = parent->parent);
next->onetime_parent = parent->orig;
--- 66,72 ----
next->next = *addr_new;
*addr_new = next;
! if (ob->one_time && !queue_2stage)
{
for (parent = addr; parent->parent != NULL; parent = parent->parent);
next->onetime_parent = parent->orig;