Just one more...
I've found it useful to have remote_host available in my pre-conection
plugin (specifically for enforcing concurrency limits specified in a
rule based on host patterns). Attached is the patch I have applied to
make this possible.
Now I'm off to bed!
-Jared
Index: qpsmtpd-f
I noticed the previous iteration of this patch would allow cached global
config values to override per-recipient config values. The attached
patch adds the following change to avoid this:
--- lib/Qpsmtpd.pm (revision 967)
+++ lib/Qpsmtpd.pm (working copy)
@@ -142,7 +142,7 @@
my ($self, $c
On Fri, 09 Jan 2009 22:00:36 -0600, Jared Johnson wrote:
> Most notably for Address:host(), user(),
> notes() and Connection::notes(), Transaction::notes(): it changes the
> 'set' version of these get/set methods to return $self. This makes it
> possible to do things like:
>
> $rcpt->user('jaredj
I can't say I don't hope you're overruled, as I personally enjoy using
that syntax all over the place and would be happy to submit patches to
have the rest of QP act that way :) Nevertheless, here's another
variant that retains the previous behavior, but with the more
standard/one-line-shorter
Matt Sergeant wrote:
>
> On Fri, 09 Jan 2009 22:00:36 -0600, Jared Johnson wrote:
> > Most notably for Address:host(), user(),
> > notes() and Connection::notes(), Transaction::notes(): it changes the
> > 'set' version of these get/set methods to return $self. This makes it
> > possible to do t
What's the point of this? If you need the hostname in the pre-connect
hook you can look it up there from the IP address which you already
have.
Also, what about async?
-R
Jared Johnson wrote:
>
> Just one more...
>
> I've found it useful to have remote_host available in my pre-conection
>
What's the goal of this? It also needs some documentation or examples.
Jared Johnson wrote:
>
> Adding a config method for Qpsmtpd::Address objects turned out to be
> relatively straightforward. I was reluctant to subclass Qpsmtpd, but
> this actually seems like it could be useful in the futu
While I agree that 'DECLINED' is confusing, 'NEXT' isn't terribly
clear either.
I think I'd just stick with the status quo. You get used to it
pretty quick.
-R
Jared Johnson wrote:
>
> A number of my associates who have had occasion to review my QP plugin
> code have been perplexed by our
On Jan 10, 2009, at 21:55, Robert Spier wrote:
While I agree that 'DECLINED' is confusing, 'NEXT' isn't terribly
clear either.
Blame the Apache 1.x API -- that was the API I knew best in ~2001, so
it got emulated ...
I think I'd just stick with the status quo. You get used to it
pretty