Question

I’ve got a rule like,

charGroup
    : '[' .+ ']';

But I’m guessing that’ll match something like [abc\]. Assuming I want it to match only unescaped ]s, how do I do that? In a regular expression I’d use a negative look-behind.

Edit: I’d also like it to be ungreedy/lazy if possible. So as to match only [a] in [a][b].

Answer 1

You probably wanted to do something like:

charGroup
  :  '[' ('\\' . | ~('\\' | ']'))+ ']'
  ;

where ~('\\' | ']') matches a single character other than \ and ]. Note that you can only negate single characters! There’s no such thing as ~('ab'). Another mistake often made is that negating inside parser rules does not negate a character, but a token instead. An example might be in order:

foo : ~(A | D);

A : 'a';
B : 'b';
C : 'c';
D : ~A;

Now parser rule foo matches either token B or token C (so only the characters 'b' and 'c') while lexer rule D matches any character other than 'a'.