Nonsense. I'm talking about searching for a pair of adjacent tokens in a
string. The parsing risks are false positives - unlikely given the tokens
involved - and false negatives - also unlikely given that anything generating
these tokens is going to be new code written to conform to these
specifications.
And as Arnt has pointed out, you also have control over the agent generating
the line you care about most of the time.
Someone else this time, I think, although I've said the same thing in
the past. (My two cents: Parsing Received is simple and easy, making
sense of the parsed information often is neither.)
There is another aspect of this that I feel hasn't been thoroughly
discussed, namely the likelihood that signalling 6532 compliance in a
different way would be more (or less) reliable than the current way.
What I mean is that it certainly would be possible to say e..g "
"Mime-Version: 1.1 means 6532". But lots of people ignore
Mime-Version, including Gmail. So is there an argument that some other
way to signal 6532 compliance would be more reliable, more likely to
be set and used correctly, than the current way?
Arnt
_______________________________________________
ietf-822 mailing list
ietf-822(_at_)ietf(_dot_)org
https://www.ietf.org/mailman/listinfo/ietf-822