Re: [Tools-discuss] Step backwards: <[e-mail]>

[Henrik Levkowetz <[e-mail]>]

Hi Michael,

On 2019-03-14 00:47, Michael Richardson wrote:
> 
> Henrik Levkowetz <[e-mail]> wrote:
>     > On 2019-03-13 00:25, Patrik Fältström wrote:
>     >> On 13 Mar 2019, at 8:19, Henrik Levkowetz wrote:
>     >>
>     >>> I believe this will make Reply and Reply All work as before, while
>     >>> still avoiding the troubles we've seen from using a From: address
>     >>> that may look like a forgery.
>     >>
>     >> May I suggest stopping changing the body of the email messages (by
>     >> adding a footer). That way DKIM signatures will work, and the email
>     >> will indeed not be defined being forgery which in fact they are as
>     >> long as the body is changed.
> 
>     > That is a different issue (that will hopefully be solved by ARC, which
>     > we hope to deploy soon).
> 
>     > Brian is talking about a recent change away from a practice where
>     > when a reviewer would enter his review into the datatracker, it would
>     > originate a mail purporting to be from <[e-mail]>,
>     > and send that out from the server mail.ietf.org, which is not mentioned
>     > in the SPF record of reviewers.domain.
> 
>     > Since this was becoming increasingly troublesome, as more and more
>     > domains were adopting stricter SPF settings, we've stopped sending
>     > mail purporting to be sent by <[e-mail]>, and are
>     > now sending from <[e-mail]>, but we missed adding a Reply-To:
>     > header field that would make replies work as before.
> 
> Wouldn't the @dmarc.ietf.org mangling have also been an option?

Yees, but much more complex, and with a forced return through the
ietf server, which all seems against KISS.

	Henrik