One could however offer a alternate standard set of plugins
implementing the discussed ideas, I think.

One could.  I've been considering it, but mine's not as configurable (in
terms of handling mailer behaviour on a hit) as I'd like.  Probably
wouldn't be hard to rework that bit.

The project already maintains multiple different daemon architectures, and a separate set of plugins for -async... IMO it would be an unnecessary burden to maintain an alternate set of plugins. I agree with Chris that methodology changes done right would make plugins simpler, though, eliminating the need for this.

Of course the semi-related issue of per-recip configuration directives does seem like something that could complicate plugins, unless that complexity were moved to core somehow (e.g. the plugin says DENY and core possibly decides to do something else)

-Jared

Reply via email to