[antlr-interest] How to handle first non whitespace comment characters?

[Johannes Luber]

Jim Loverde wrote:
> Hello,
> 
> I was hoping someone could help me out, as I've been struggling with this for a bit.  Basically, how do you handle something like a COBOL asterisk comment character, where it is only a valid line comment indicator if it is the first non whitespace character on a line (but if it's part of an expression such as 5 * 7 it shouldn't be treated as a line comment character).
> 
> For example:
> 
> * This is a line comment
> IF 5 * 5 = 20 THEN
>     * this is also a line comment
>     NOTHING
> END-IF.
> 
> Ideally I'd like to have the line comments channeled to the hidden channel similar to the "normal" line comment such as:
> 
> LINE_COMMENT
>     :    '!' ~('\n'|'\r')*  '\r'? '\n' {$channel=HIDDEN;}
> 
> Also, while on the topic, does anyone have a COBOL grammar for ANTLR (or even a partial grammar) that they can share?

I have a similar problem with C#'s preprocessor directives. Only at the
beginning of the line or only whitespace may be in front of "#". My
solution is to use a boolean variable, which is reset at the NEW_LINE token:

NEW_LINE
   :    { AllowPPDirective(); } '\u000D' // Carriage return character
   |    { AllowPPDirective(); } '\u000A' // Line feed character
   |    { AllowPPDirective(); } '\u000D\u000A'  // Carriage return
character followed by line feed character
   |    { AllowPPDirective(); } '\u0085' // Next line character
   |    { AllowPPDirective(); } '\u2028' // Line separator character
   |    { AllowPPDirective(); } '\u2029' // Paragraph separator character
   ;

Then at every non-fragment rule I put:

SINGLE_LINE_COMMENT
   :    { DisallowPPDirective(); } '//' INPUT_CHARACTER*
   ;

And I do the testing like:

PP_DEFINE
   :    { isPPDirectiveAllowed }? '#' WHITESPACE? 'define'
   ;

Unfortunately I couldn't get my parser running due to unrelated
problems, so I can't vouch that this solution will actually work. On the
other hand, there isn't anything what would make me doubt my solution's
capabilities.

Best regards,
Johannes Luber