Re: [Emailcore] Ticket #14: G.7.8. Review different size limits

[Ned Freed <[e-mail]>]

> --On Saturday, July 17, 2021 19:06 -0700 Ned Freed
> <[e-mail]> wrote:

> >> Actually, given an 8BITMIME message and a nexthop SMTP server
> >> that does not support 8BITMIME, Postfix will in fact (by
> >> default) do on the fly MIME quotable-printable encoding.
> >
> >>     http://www.postfix.org/postconf.5.html#disable_mime_outpu
> >>     t_conversion
> >
> > Our MTA defaults to this as well. But it's trivial to switch
> > off, and I have no way of knowing how many sites do so.

> Completely out of curiosity, as the amount of traffic that does
> not use Latin script increases (IMO, quoted-printable is far
> less useful when most of the characters are outside the ASCII
> repertoire), do either of you do the fairly minimal analysis
> needed to decide when Base64 would be more compact and no less
> useful?

In our case it depends on when the downgrade occurs. If we're doing it
on the fly with an open SMTP connection we just go with a default; we don't
do the analysis. If we're doing it earlier when we have the data we do
the calculation and pick the better encoding.

> Or provide a switch s.t. if encoding is to be done,
> Base64 might be preferred by the configuration instead of
> quoted-printable?   It seems like that might be worthwhile if
> one were carrying a significant amount of non-Latin-script,
> traffic.

I've had storing the analysis results on the to-do list literally for decades.
But it's never been a sufficient issue that it's ever budged from its position
near the bottom of the list. At this point it's unlikely it ever will.

> of course, none of this has anything to do with changes that
> might be needed in rfc5321bis, AFAICT.

No, it doesn't.

				Ned