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

[Steve Atkins <[e-mail]>]

> On Jan 17, 2019, at 9:24 AM, Paul Smith <[e-mail]> wrote:
> 
> On 17/01/2019 03:31, Gene Hightower wrote:
>> Microsoft, for their hotmail.com service anyway, does support
>> BINARYMIME for received mail. And, as far as I can tell, it works.
> 
>> Strangely, when Microsoft sends mail they don't seem to use BINARYMIME
>> even though I offer the extension.
>> 
> To be honest, I'd think BINARYMIME would be easy to implement on a final delivery SMTP server (as you KNOW what's downstream from you, and that they won't need to recode the message).
> 
> However, on the outgoing part of a submission server (or on an intermediate MTA), it's a lot more tricky, as you don't know what you may be sending to and whether it'll need recoding.
> 
> So, that's quite possibly why Microsoft accept it on incoming, but don't use it on outgoing. Seems sensible to me.

The time I've been bitten by this in the wild was when Inbound mail was delivered to Exchange, which then attempted to forward it to the final recipient. The final recipient had didn't speak binarymime, Exchange had no idea what to do in that case so just dropped the message on the floor.

I fixed it by modifying the final recipient to accept binarymime, but I'm not sure what I'd have done if I hadn't been able to do that and it really showed how that part of the protocol is kind of broken.

Cheers,
  Steve