6 ms·

Yeah. Similarly, for the range transformations, instead of `[a:A-z:Z]`, I would suggest `[a-z:A-Z]`; and instead of `[a:b-y:zz:a]`, something like `[a-y:b-z;z:a

• kazinator 2y ago
  I would suggest simply [a-z]:[A-Z], inspired by tr. Then there is no syntactic special case. This is just EXPR:EXPR; the special case is that both EXPR are character class syntax, and so the tr-like range mapping applies.
  • c0nstantine 2y ago
    [a-z] is equivalent to 'a|b|...|z' in the normal regex language. So if we do [a-z]:[A-Z] it should be expanded to: (a|b|...|z):(A|B|...|Z) which is pretty legal in trre but has different meaning of mapping any a-z to ALL the A-Z (generating A-Z on each occurrence of lowercase letter).
    • kazinator 2y ago
      [a-z] is a semantically equivalent regex to a|b|..|z, but the two are not equivalent syntactic forms. Distinct syntactic forms can be given distinct semantics, as long as there is rhyme and reason. Moreover, the right side of the colon is not the normal regex language, it only borrows its syntax. So there, we may be justified in denying that the usual equivalence holds between character class syntax and a disjunction of the symbols denoted by the class.
      • c0nstantine 2y ago
        The right side is a normal regex language syntactically. Semantically it is a generator instead of a parser (consumer). But I got your point. Maybe there could be some ways to do it in consistent way. Just straight tr-like syntax won't work, e.g I really want it something like this to be valid: [a-b]:(x|y) (pairs a:x, b:x, a:y, b:y) and I prefer not handle these in some ad-hoc way.
        • kazinator 2y ago
          I also go your point. The right side is a regular expression because it denotes a regular set.