Todd,


Actually, what I'm trying to do is something simple: remove large attachments from an email if the message can't be delivered (incorrect/nonexistent destination address). This will save users in my case from having to grab a huge attachment that they sent just to know that it didn't get delivered (still have POP3 users to consider).

The easiest way I can think to do it is to have qmail run something when it can't deliver a message from the queue just before it returns the mail to the user.

There is a qmail patch that will let you truncate the bounced message if over a certain size, but that can get nasty if it truncates the across a MIME boundary (attachment can get truncated and cause general problems). I want to do it cleaner such that any non text attachment gets the axe before the message is returned to the sender.

Hopefully that gets into the details a bit more to let you know exactly what I want to do. qmail provides hooks for scripts into incoming messages (.qmail), but I want a hook into a bounced message as well (just before it is returned to the sender).

-Scott

Todd A. Jacobs wrote:
On Mon, 28 Jul 2003, Scott Fritzinger wrote:


Now, the only question is how to have the bounce account send the
filtered, bounced email back to the original sender?


I'm not sure who's bouncing what in your scenario. If you want to bounce a message, use bouncesaying. If qmail itself is bouncing the message because of a filter (e.g. rblsmtpd) then you'll need to see what qmail is doing with the exit status and hack accordingly.

In the normal case, messages get bounced hard when a user is over quota or the account doesn't exist. Anything else is probably a filter applied before the .qmail files are accessed, or user-specific where bouncesaying or |preline filters can be brought into play in the dot-qmail files.

Anything more esoteric should probably be brought up to the qmail support
list, although most likely you'll get a "don't do that!" sort of response
since qmail is designed around KISS, and you're probably violating the
KISS principle rather thoroughly with whatever it is you're really trying
to do here.


_______________________________________________ RLUG mailing list [EMAIL PROTECTED] http://www.rlug.org/mailman/listinfo/rlug

Reply via email to