[Top] [All Lists]

Re: [ietf-smtp] [Proposal] confusing parts of the mail system, was 250-MARKDOWN

2019-01-16 20:48:24

On Fri, Jan 11, 2019 at 1:58 PM Gene Hightower 
<gene(_at_)digilicious(_dot_)com> wrote:

Again: can we start with CHUNKING?  Does anyone *object* to CHUNKING?

CHUNKING works today. Hotmail/Exchange/Office365 was the only thing that I
was aware of that advertised it when I implemented it for Gmail ~2013. IIRC
Exim added it a few years ago. Postfix added it in 2018, not sure if it's
enabled by default.

We (Sun/Oracle) implemented CHUNKING support in 2006. We have not
noticed a significant performance improvement, but our DATA pipeline
is pretty heavily optimized.

A quick check of my home logs over the past year or so shows 37% of outgoing
messages were sent with chunking and 49% were received with it. Not exactly

The usual problem with having CHUNKING enabled by default is when someone
installs a broken SMTP security proxy that doesn't understand CHUNKING and
doesn't know how to filter the CHUNKING extension out of the EHLO response. And
then complains about things not working.

Such garbage is getting rarer over time but is still all too common.

As for BINARYMIME, we implemented that for SUBMIT in 2012 due to a couple of
customer requests. As I explained previously, our view is that BINARYMIME makes
sense for SUBMIT because it is far more common for it to be
bandwidth-constained than SMTP transfers, and transcoding is acceptable at
submit time.

We don't have any real way of knowing to what extent it's used, but given the
relative paucity of problem reports my suspiciion is "not much". I'm not
entirely sure why - a SUBMIT client support should not be that difficult.


ietf-smtp mailing list

<Prev in Thread] Current Thread [Next in Thread>