Jason L Tibbitts III <tibbs@math.uh.edu> replied to me:
> n> This report would be generated as soon as at least one attempt has been
> n> made for each subscribed email address to transmit the post to the MX
> n> host for that address.
>
> That is so MTA dependent it's not funny. How do you plan to know when the
> MTA has done this?
>
> n> The current strategy is that this is done by sendmail or qmail, running
> n> either at the remote host or at a smarthost, and Majordomo talks to this
> n> MTA via SMTP (which makes it neigh impossible to get at the desired
> n> information.)
>
> That is the only reasonable strategy. Unless of course you want to write a
> full MTA in Perl, which you are welcome to do but which I think is an
> incredible waste of time.
Sure, but one doesn't need to write a full MTA to make this work. It
would suffice to have something that understands the small subset of
SMTP that Mj2 actually uses and then runs ($sendmail,'-v',@addresses)
and then appends the output of sendmail to a file named after the
message-ID of the message. Each of these processes would have a
non-exclusive lock on an additional per-message lockfile. As
soon as Majordomo has finished pumping batches to these pseudo-MTA
processes it would wait until they're all finished (by requiring an
exclusive lock on that lockfile) and then analyze the output from
sendmail to generate the confirmation message.
Caveat: This will work only with sendmail on the local host, not with
other MTAs or with smarthosts.
-- NB.
>
> - J<
>
Follow-Ups:
References:
|
|
