Re: [Exim] improvments to exim

> >It occurred to me as I walked home on Friday that it would not be so
> >very difficult to add a new precondition to the routers, of the form
> > acl = ....
>
> The immediate risk I can see with this feature is that it adds another
> kind of statically-unbounded looping into Exim, since ACLs call routers
> via verify=sender and verify=recipient, and routers would then be able
> to call ACLs. This is probably not such a problem given that ACL nesting
> is limited to a depth of 20, and routing is limited to 100 redirections.

you could
1) not allow anything in an acl that does verification (or in turn causes
routing)
2) skip the router if hit via an acl
3) use/require use of no_verify on the router.

Personally, I'd choose #1 over #2.

Are there any times other than verification an acl runs through the routers?

--
Lab tests show that use of micro$oft causes cancer in lab animals