On Wed May 11 2005 22:40, Dave Crocker wrote:
It's a little worse than that, I think. The idea of correlating a comment
with a particular bit of ABNF strikes me as clever, like so many ideas.
However, like so many ideas, the real question is whether the community
requires it, thereby justifying the effort to invent the solution and rev the
specification.
You have missed the point(s):
1. As written, parsing ABNF requires arbitrarily-large lookahead
2. Because of the ambiguity, a parser which appears to be parsing ABNF
(as that is specified) may in fact be parsing a different grammar (one
without that ambiguity), and would not be suitable as a fully-conforming
implementation for the purpose of advancement of the specification to
Draft status.