[Discuss] DMARC issue, Yahoo and beyond

[richard.pieri at gmail.com (Richard Pieri)]

Derek Martin wrote:
> I admit I'd forgotten this; for the longest time I had a patch which I
> wrote to fix this applied to my mutt; Mutt dev being what it is
> (basically dead) the maintainers didn't have any interest in applying
> it.  I have no use to maintain patches forever so I stopped bothering.

There you go. :)


> Or you can pick the "right" option: make reply-to work sanely in all
> mail programs...  As implemented currently, reply-to is next to useless,
> and as you say, detrimental... even when used as intended.  It should
> not be thus.

Reply-To is not useless (or next to it) when used to direct individual
correspondence. For example, I compose a note to a friend from my work
mailbox and set the Reply-To field to my not-work mailbox to continue
the discussion that way. Or, I was mistakenly included on some bit of
work-related conversation and I want to make sure that the right
correspondents are involved and set the Reply-To accordingly.

All the MUAs that I've used behave consistently in cases like these.
It's when Reply-To is used with mass correspondence such as mailing
lists and large Cc lists that it gets weird or disruptive. And mailing
list software that munges Reply-To fields just makes it worse.


Regarding that Emacs-Lisp code, upon reflection I think that I didn't
write a hack reply handler. I think I used a filter that stripped the
Reply-To header if the mailbox matched the mailbox in the To field.
Writing code to strip invalid headers is something I'm more likely to do
than writing code to work around them.

-- 
Rich P.