[Raw Msg Headers][Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Zmailer shortcomings

> Ok, folks, let's talk turkey.

  Yes, its the season..

> Zmailer has a number of shortcomings, or rather, short sightedness,
> that I was wondering if anyone else found annoying enough to fix...
> 1] From: header lines that are continued on the following line
>    (perfectly legal RFC822) *never* work, especially if it is a
>    comment that is being continued. An example:
> From: mshappe@harper-hall.cit.cornell.edu (By way of 
> 	nobody@nowhere.in.particular.ny.us)
> This format is used by Eudora (a popular Mac POP client in wide use
> here at Cornell) to denote some forms of forwarding. It's perfectly
> legal 822, but Zmailer bounces it saying that it's not complete. It
> ignores the continuation line entirely and wonders where the closing
> paren is.

  I know of this behaviour, altough I haven't checked if it really is legal
format or not.  On the other hand, both POPx and IMAPx -protocols are
only half-way implementation of the thing.  There should be a protocol
for bi-directional traffic between Message-Store/Message-Transport-Agent
and Message-User-Agent.  Now POPx and IMAPx specify SMTP to be used
for transmit...

> 2] X.400 addresses in "/PN=Fishbait/ADMD=Moose Liver/"@sprint.com
> format never seem to work. Ususally the error is something along the
> lines of 'illegal domain name', I believe. 

  That has worked just fine for me for ages -- it is more of MUA->MTA
interaction problem than anything else.   I had to hack seriously on
elm 2.3  to get it right (and to have those features into elm 2.4..)
Basically those quotes MUST bet into mailer...

	/lib/sendmail -bt
	z$ router "/PN=.../ADMD=Moose Liver/"@sprint.com
		  ^                         ^
Guess what happens to those in this example ?  They disappear...

	z$ router "\"/PN=.../ADMD=Moose Liver/\""@sprint.com

Now it works.  Similarly to your mail clients.
POPper feeds via SMTP, and you can see those incoming addresses in "to "-
headers before normal headers.

> These are two off the top of my head that I can think of. I have no
> doubt they're reparable, but I have yet to be able to disentangle
> Rayan's code enough to figure out where the fix should go :-)

  I am hunting now wierd scheduler crashes with outqueue beeing around
1k files..  (NULL pointer dereference..)

> Thoughts?

	/Matti Aarnio <mea@nic.funet.fi>