K-user mailing list

[Eric Huber <echuber2 AT illinois.edu>]

Hi, I'm wondering if I understand correctly the differences between Token and Lexer. Lexer is not greedy, so it must be given the follow rule with -/- to show when it may NOT stop collecting characters for one token, correct? Whereas Token is greedy and can not use -/- rules.

Also, this detail from ref-manual.k seems counter-intuitive:
  The construction {<char set> "<separator>"}+ will match 'A', 'A-A',
  'A-A-A', but not 'A-'.

Doesn't the syntax suggest that the separator character must follow each character? I take it that the only reason this works is because in the particular example:

syntax MODNAME ::= Lexer{"#"?{[a-z0-9A-Z\_]+ "-"}+}
syntax MODNAME -/- [a-z0-9A-Z\-\_]

it is using Lexer with the construction, so it is not a greedy selection, and the follow exclusion rule says that - may not appear at the end. But the line in the documentation makes it sound like a special construction to specify ranges with a separator.

Thank you,
Eric Huber