Path: utzoo!attcan!utgpu!jarvis.csri.toronto.edu!rutgers!tut.cis.ohio-state.edu!unmvax!indri!polyslo!csun!srhqla!Upacbell
From: comp-mail-mush@srhqla.uucp
Newsgroups: comp.mail.mush
Subject: Re: My domain in From: (mush 6.5)
Message-ID: <652@srhqla.UUCP>
Date: 17 May 89 13:44:12 GMT
Sender: Upacbell@srhqla.UUCP
Lines: 30

From: Marc Rouleau <mer6g@virginia.acc.virginia.edu>

On May 16, 16:03, Dan Heller wrote:
> 
> There's no question that it is desireable to allow it.  the strongest
> arguments against it is the simple fact that malformed headers have to
> be dealt with -- error recovery is complicated even when it's detected.
> If undetected, there could be unpredictable results.

I haven't been following this discussion too closely, so please forgive
me if I'm treading ground already covered.

What's so hard about error _detection_?  Just hand the message off to the
MTA -- if you get a bad return (!= MTA_EXIT), dump it in dead.letter and
forget about it.  If you get a good return, it's the MTA's responsibility
to deal with it.

The worst case is that the user puts an incorrect or even syntactically
illegal string in the From: line and the MTA accepts the message only to
find later that it (or some MTA farther down the line) cannot return it.
In that case it'll drop it in the local "orphanage" or somesuch for the
local administrator to deal with.  Really, is this, the _worst_ case, so
bad?

I think there oughta be a variable (not an #ifdef) allow_from_edit or
whatever so that the "experts" can do what they want.  Occasionally one
of them will send a piece of mail that is neither delivered nor returned,
but that's his/her problem, not mush's.

	-- Marc Rouleau