Re: [Exim] Cyrus/Exim incompatibility

トップ ページ
このメッセージを削除
このメッセージに返信
著者: Ken Murchison
日付:  
To: John Holman
CC: info-cyrus, exim-users, Lawrence Greenfield
題目: Re: [Exim] Cyrus/Exim incompatibility

John Holman wrote:
>
> At 15:25 24/01/02, Ken Murchison wrote:
>
> >Philip Hazel wrote:
> > >
> > > On Wed, 23 Jan 2002, John Holman wrote:
> > >
> > > > Messages generated by Cyrus's lmtpd (e.g. as the result of a sieve
> > vacation
> > > > or reject rule) are created with CRLF as line terminators and piped
> > to the
> > > > mail program (by default sendmail, which in our case is actually
> > > > exim). Messages presented to sendmail in this way should, I think,
> > > > conform to the Unix conventions for line termination rather than
> > those for
> > > > SMTP, and therefore not contain CR characters.
> > >
> > > This is also my view.
> > >
> > > > This suggests that exim's author, at least, would consider Cyrus to be
> > > > broken in this respect!
> > >
> > > I am Exim's author.
> >
> >Is there a spec or reference which states that when sending a message to
> >an MTA via stdin that only LF should be used as the line terminator? Or
> >are people saying that this is best common practice?
>
> My guess is that there is no real standard, but I think it makes sense when
> passing a message to an MTA via stdin to comply with the conventions of the
> local operating system. That would mean LF under Unix, CRLF under Windows etc.
>
> Actually, if it's true that there are no standards to be upheld, I also
> think it would be better for Exim (which claims to be a Sendmail clone
> insofar as the command line interface is concerned) to accept either CRLF
> or LF as a line terminator when mail is presented on standard input, if
> that is what Sendmail does.


Agreed. But on the flip side, if there is a best common practice
(whether officially documented or not), we should probably follow it.

> As Lawrence has said, Cyrus could avoid the issue entirely by connecting to
> an SMTP server when it sends mail.


Yeah, I mentioned this to Larry a while back, but he didn't really like
the idea. I don't remember his exact reasoning, but I think it had
something to do with not knowing whether the reject/vacation message
ever got sent or something. Larry?

I would be nice if there was some kind of LMTP/ESMTP backchannel for
sending out a responses to a message just delivered, via the already
open connection to the MTA. I think SMTP TURN may do this, but its use
is deprecated. Hmm.

Ken
--
Kenneth Murchison     Oceana Matrix Ltd.
Software Engineer     21 Princeton Place
716-662-8973 x26      Orchard Park, NY 14127
--PGP Public Key--    http://www.oceana.com/~ken/ksm.pgp