From: Wietse Venema (wietseporcupine.org)
Date: Wed Sep 19 2007 - 14:08:33 CDT

• Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]

Victor Duchovni:
> On Wed, Sep 19, 2007 at 02:14:12PM -0400, Wietse Venema wrote:
>
> > > On the other hand, most receiving systems accept 8bits, so conversion
> > > at the edge is not that frequent. Having an ISP smarthost that does not
> > > announce 8bit support is a bit unfortunate though, in that case one
> > > really has no choice but to down-convert first and then sign.
> >
> > They can safely specify "disable_mime_output_conversion=yes". In
> > this day and age, all ISPs run MTAs that are 8BIT clean.
>
> What happens when the ISP (in this case using Exim) fails to down-convert
> content when sending to 7bit-pedantic systems? Part of the reason MIME
> went into Postfix was reports of failures when sending to 7bit-pedantic
> sites. Are these presumed to be far less common now that ~5 years ago?

Speaking of multi-hop settings, when mis-labeled mail is sent
through Exim, you can't win. When Exim forwards the mail to an
MTA that understands MIME, that downstream MTA will see the
Content-Transfer-Encoding: 8bit header, assume that the upstream
MTA didn't understand MIME, and we're back at square one.

> Are we ready to recommend "disable_mime_output_coversion=yes" as
> a default?

No, I would prefer to put some pressure on the maintainers of MTAs
that force Postfix and other RFC-compliant MTAs to break signatures.

> Is there anyone who feels they will benefit in practice (rather than in
> theory) from a feature that allows 7bit mail mislabeled as 8bit to go
> through unconverted?

How would you solve the multi-hop problem?

        Wietse

• Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]