Philip Hazel wrote:
>>I like the new acl_*-variables, but maybe it would be better to keep
>>them while routing the mail (at least the acl_m-variables). This would
>>save us from adding headers to hand informations from the ACL to the
>>routers.
>
> I *knew* somebody would ask for that when I implemented what I was asked
> for, namely ACL-time variables. That is why I named them $acl_xxx. One
> reason for not handing them on is that this would mean more data has to
> be kept in the -H file, but I suppose that is a small point really.
Sorry, forgot to ask for it while running the snapshot.
I don't think that the -H files will be a real problem, as we only need
to save variables that were really in use, so for the most cases it adds
a VERY small overhead to the -H parsing.
If you really care about the overhead, maybe this could be optional
[(save|keep)_acl_message_variables, default false]. I don't think many
people need it, but it's nice to know that it's possible. :)
>>The question is, WHAT comes next? When I started to use exim3, I thought
>>WOW!! can this be even better? But Philip proved that it's possible,
>>with exim4 :)
>
> Every program reaches its natural peak at some point, and some new
> program then comes along to replace it. I have no illusions about that,
> having been writing programs for quite a long time now...
I didn't meant that exim will last forever, at some point (far far away)
it will be most probably replaced with something better, but currently
exim is the tip of the MTA evolution ;).
And the flexible configuration allows us to mutate it even further to a
new species ;)
