[antlr-interest] Non-greedy lexer rule in ANTLR4

[Bart Kiers]

> Is it possible to write a non-greedy lexer rule that does not contain a
> dot in ANTLR4? For example:
>
> ANCHOR
>    : '<<' ( ~( '>>' | '\r' | '\n' ) )* '>>'
>    ;
>

Hi Michel,

You can't negate two (or more) characters: negation only works for
char-sets that match a single character.



> I'm using 4.0b2 and the above seems to be greedy.
>
> Alternatively, would it be possible to rewrite this rule so it includes
> everything from '<<' to the first '>>' but not more?
>

Try something like this:

ANCHOR
   : '<<' ( ~( '>' | '\r' | '\n' ) | '>' ~'>' )* '>>'
   ;

regards,

Bart.