On Tue, Oct 6, 2020 at 8:09 AM John Levine <johnl(_at_)taugh(_dot_)com> wrote:
In article <75046.1601954458@turing-police> you write:
Only if the entire path is sufficiently 8-bit clean. Somewhere between
ietf.org's
mail server, google, and fetchmail to my laptop the emoji got smashed into
a lower case 'p' followed by a 0x1f181f, almost certainly due to something
not being totally 8-bit-clean.
I would blame fetchmail which is not famous for high quality standards
compliant code. Postfix and Gmail do not screw up MIME encodings,
particularly since all they need to do to comply is nothing. I see you
use exmh but you presumably know what your code does.
Anything along the line of "if you don't have an 8-bit clean path, you
need to use rfc2047 to encode the utf-8" is fine by me.
Well, yes, that's been the rule for about three decades.
We could say that the field isn't encoded and only available for SMTPUTF8
as a forcing function... you want your emojis, step up.
I know, doesn't really work that way, we'd probably just get 5 different
non-standard ways of encoding it instead.
Brandon
_______________________________________________
ietf-822 mailing list
ietf-822(_at_)ietf(_dot_)org
https://www.ietf.org/mailman/listinfo/ietf-822