In the months since we went live with probably hundreds of millions
of messages passing through our signers/verifiers, this is the only thing
that I've seen with any consistency that breaks the body with simple.
How many of those were signed though? And what was the intent of the
signer wrt these mail-formed mails? And hundreds of millions over many
months sounds like a tiny sample.
And more to the point, do you have any way of tracking what happened to these
messages between the time they were composed to the time they were read? Most
submission agents find it necessary to perform a huge array of fixups because
the stuff that clients emit is often egregiouly bad. Just because the tiny
fraction that underwent signing at some point after initial submission and
verification at some point shortly thereafter did OK doesn't mean that nothing
interesting happened at other points along the path.
I have to say I see a lot of "blind man and the elephant" thinking going on
here. We have have our own set of experiences with email, but the Internet is a
huge and surprisingly disjoint place, and none of us can claim to have a good
understanding of everything out there. In my particular corner I deal with
problems caused by malformed messages all the time. (Last week it was some
voicemail thingie that was submitting what amounted to raw binary material in
plain old SMTP. And the system in question is apparently widely deployed - the
people who developed it are trotting out the time honored "this is the first
time there has been a problem so it must be someone else's fault" canard.) But
I have no problem believing that someone else's experiences in a similar
position with a different development group would be quite different.
Ned
_______________________________________________
NOTE WELL: This list operates according to
http://mipassoc.org/dkim/ietf-list-rules.html