[1241] in NetBSD-Development
Re: Masquerading dupe suppresion loses on FQ and non-FQ together
daemon@ATHENA.MIT.EDU (Eric Allman)
Mon Feb 5 14:40:42 1996
To: John Hawkinson <jhawk@MIT.EDU>
From: Eric Allman <eric@sendmail.org>
Cc: "Kari E. Hurtta" <Kari.Hurtta@dionysos.fmi.fi>, sendmail-bugs@sendmail.org,
netbsd-dev@MIT.EDU
In-Reply-To: Mail from John Hawkinson <jhawk@mit.edu>
dated Mon, 05 Feb 1996 03:56:21 EST
<199602050856.DAA03716@lola-granola.MIT.EDU>
Date: Mon, 05 Feb 1996 11:35:14 -0800
There is no ESMTP negotiation for things like line length. Also,
the relay mailer assumes you are staying in your domain, hence you
can pass headers that may not be completely up to snuff as far as
the specs are concerned.
The difference between your two addresses were that one had a dot on
the end of "mit.edu.", one didn't.
This stuff can probably be improved, but these days I don't have much
time to work on anything that isn't pretty serious. I don't think
this counts.
eric
============= In Reply To: ===========================================
: From: John Hawkinson <jhawk@mit.edu>
: Subject: Re: Masquerading dupe suppresion loses on FQ and non-FQ together
: Date: Mon, 5 Feb 1996 03:56:21 -0500
:
: Kari writes:
: > relay mailer assumes, that relay is running sendmail and thereofe uses
: > different kind limits for line length and so on.
: >
: > (So 'relay' mailer is 'esmtp' mailer + 'smtp8' mailer + bigger line length.
)
:
: Hmm, ok; it would seem that it would make sense to just junk it and
: let ESMTP negotiation take care of it, but perhaps the efficieny is
: worth it (??). I guess the relay mailer is much more useful for
: non-smtp configurations? [note that somewhere we moved from a
: discussion of Mrelay to confRELAY_MAILER, and the two aren't the same
: concepts...]
:
: > > Unfortunately, it's still not quite right:
: >
: > Use also: define(`confSMTP_MAILER', `esmtp')
:
: Well, sure :-) Again, wonders why that's not the default...
: (actually, in my case, the MX of MIT.EDU is not currently an
: esmtp-capable mailer, so this is not necessarily advantageous, not
: that it matters much).
:
: Crux of the easy-to-fix bug: It certainly seems to be the case that
: when using FEATURE(allmasquerade), one should expect confRELAY_MAILER
: and confSMTP_MAILER to resolve to the same thing. Either that or
: sameaddr() should consider some mailers equivalent (smtp, esmtp,
: etc.).
:
: [ various quotes on what the relay mailer can be used for ]
:
: Right, I know what it can be used for, but it doesn't seem clear why
: it's necessary at all...
:
: No ideas on the next problem (sameaddr(), etc.)?
:
: --jhawk