RE: [exim] Possible bug -- def:acl_m0 operator misbehavior.

Philip wrote:
> You are correct. The docs do say something about this,
> though. From section 39.19:

Yeah, I had read that as well. I was adding the header lines in the DATA
phase (not predata) at the beginning of it, and then trying to test for the
headers I was adding at the very end of the DATA phase - so it was all
within the same ACL (well, I use require lines to include other ACLs, but it
was the DATA phase executing them all).

It was something like ${if def:h_X-Spam-Scan ... } and it wasn't working for
me. I might be able to reproduce it, but in the end I actually decided not
to use it since a) it didn't work :) and b) the header could have been
injected when I received it so I opted to use variables instead.

> There is a Wish List item to re-design and re-implement the
> way header additions and substractions are handled in all the
> places in Exim that this happens. But don't hold your breath.

Yep :) Waiting patiently :)

Eli.