[ietf-dkim] Possible problem with "simple" body canonicalization -- trailing CRLFs

[John Levine]

>Exactly. If you're using something other than SMTP DATA, such as SMTP
>BDAT, it is conceivable that the body could wind up without an ending
>CRLF....

BDAT is part of the CHUNKING extension defined in experimental RFC
1830 in 1995.  It's been 11 years.  Does anyone use it?  Has anyone
even proposed making it standards track?  I don't ever recall running
into it, so I don't see any point in worrying about it.

We may well end up using the DKIM key infrastructure for signatures of
other things, but it's going to take a lot more than deciding which
line endings to ignore to make that happen.

R's,
John

PS: If anyone ever updates that RFC, it's now called Chongqing.