On Tue 23/Apr/2024 17:03:38 +0200 Murray S. Kucherawy wrote:
On Mon, Apr 22, 2024 at 4:31 PM John R. Levine <[email protected]> wrote:
[...] Ale has been proposing ways to undo list modifications for ages,
again no interest. >
I still think the modification reversal idea is worth investigating, and
I've posted my own drafts about it in the past, but I can't really advocate
for this when I don't have the bandwidth right now to produce an
implementation.
My implementation[*] followed more or less Murray's draft[†], except for MLs
involvement. It works unreliably. Wei too proposed a draft[‡] on that.
My opinion is that a generic transformation tracking, à la git, besides
implying an unduly overhead, cannot be trusted to respect the original content
and intent of the message, unless the sender itself is trusted. And if the
sender is trusted ARC is enough.
Having MLs affix a declaration of which transformations they applied, chosen
from a small set of allowed transformation types is not going to be much more
reliable than a "blind" transformation reversal. OTOH, codifying all the
fields that DKIM can sign and all the ways a mailing list can alter the
content, e.g. by using quoted-printable rather than base64, can get so detailed
as to result in generic transformation tracking, as it has to be done at the
content layer which is not the layer DKIM signs.
Best
Ale
--
[*] https://datatracker.ietf.org/doc/html/draft-vesely-dmarc-mlm-transform
[†] https://datatracker.ietf.org/doc/draft-kucherawy-dkim-transform/
[‡] https://datatracker.ietf.org/doc/draft-chuang-mailing-list-modifications/
_______________________________________________
dmarc mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/dmarc
