"Jason R. Mastaler" <[EMAIL PROTECTED]> wrote in message
news:[EMAIL PROTECTED]
> FWIW, this patch won't be something I'll be adding to the
> distribution. First, there isn't enough general interest for
it, and
> second I disagree with it philosophically.
Hypothetical question:
What about adding one more config var: CONFIRM_APPEND_ALWAYS?
By default it is set to false. When it is set to true, valid
confirm-accept tagged emails' envelope senders are ALWAYS
appended to CONFIRM_APPEND, even if the original pending message
has already been deleted.
Thus processing would go as follows:
1 - Mail comes in with a confirm-accept tagged address
2 - TMDA checks if the tag is valid (HMAC is okay)
3 - TMDA checks if the message exists.
If it does, process as usual, appending to CONFIRM_APPEND if
defined, and releasing the message.
If it does not:
4 - check for (new) variable CONFIRM_APPEND_ALWAYS. If set
to TRUE, append the sender to CONFIRM_APPEND
5 - send BOUNCE_TEXT_NONEXISTENT_PENDING
This way if someone is late confirming their message (or if the
server admin has set the PENDING_LIFETIME to an ungodly low
number) they still count as having confirmed their message and
are un-hassled when they send from then on.
--
Jim
_____________________________________________
tmda-users mailing list ([EMAIL PROTECTED])
http://tmda.net/lists/listinfo/tmda-users
