On Aug 16, 2011, at 7:53 PM, Hector Santos wrote:
Keith Moore wrote:
On Aug 16, 2011, at 6:29 PM, Hector Santos wrote:
Agree. I just wanted to show at least the one way found why/how it can
happen, otherwise, with my tendency to give everyone the benefit of the
doubt first, I would still be pulling my hair on this one. :)
There are some other insight to this, that IMV is worthy of implementation
notes, but yes, I agree, the SMTP client MUST escape the dot lines if its
not the true end of data marker.
This is not a new problem; I've occasionally seen clients that botched this
ever since the mid-1980s. (which is to say, ever since I first had anything
to do with SMTP).
I'm not sure how implementation notes would help here. If client
implementors can't be bothered to read the spec, why would they read the
implementation notes?
maybe we need to make the protocol more complicated so that they have to do
so :)
Ok, I can see your position, therefore we should use this principle to
reduce, eliminate all WG, IETF/IETF reviews, conflicts, strong passions that
delay progress when in fact, no one really need to follow anything standard
or put in the sweat to at least try and all software really just need to
follow one principle:
Be 100% liberal in what you receive, and don't worry about being
very conservative in what you send.
what are you smoking?
I'm only saying that all we can do is document how things should work. If
we've done that, and done it clearly,
adding more documentation probably won't help.
Keith