[ietf-dkim] 822/2822 or just 2822
MarkD+dkim at yahoo-inc.com
Thu Jul 20 19:52:48 PDT 2006
> >Maybe, but if they're going through the usual MTAs, they're breaking
> >already so I can't get too worried about it.
> That's not true. Sendmail does not change these strings.
Is that true on the SUBMIT front? or the MTA front?
> We really have
> no clue as to whether this is going to be a problem.
I guess I'm a little confused about the problem here. For non-signed
mail we don't care, yes?
For signed mail, the spec is asking signers to ensure compliance with
2822. If, as you say, sendmail passes through whatever bits it gets,
then as long as signers ensure compliant messages, what problem are we
trying to solve?
Certainly gratuitous CRs and LFs are common, and while mindful of
Ned's observation about anyone truly *knowing*, I will nonetheless
suggest that such spurious characters are mostly introduced through
careless SUBMIT agents, not transfer agents. Assuming SUBMIT is most
likely prior to signing the conflict is for the signer to resolve with
Put another way, my question is this: what evidence is there to
suggest that stray characters are likely to be introduced *after*
signing of an rfc2822 compliant mail?
More information about the ietf-dkim