[antlr-interest] 3.0 feature discussion: Metadata

[Richard Clark]

One of the things that's become apparent in gluing ANTLR to eclipse is 
that interactive tools would like to know the structure of the language 
they're working on (e.g. to offer better error recovery, code 
completion, etc.) ANTLR has all of that information, as reflected in 
the code it generates, but the information is not actually available to 
the generated lexer/parser at runtime.

So, I would like to ask if ANTLR could have an option to re-emit the 
language definition in a compact machine-readable form when building 
the lexer and parser. (Perhaps a canonical XML representation, though 
calling that "compact" strains credibility ;) ) In that way, a smart 
editor could find a definition in the language (e.g. "new expects a 
className after it") and use that to offer advice on-screen.

I realize it's possible to take the ANTLR language definition itself 
and alter it to make my own ANTLR meta-parser, but it seems silly to 
duplicate the effort that ANTLR itself would be doing.

  ...Richard



 
Yahoo! Groups Links

<*> To visit your group on the web, go to:
    http://groups.yahoo.com/group/antlr-interest/

<*> To unsubscribe from this group, send an email to:
    antlr-interest-unsubscribe at yahoogroups.com

<*> Your use of Yahoo! Groups is subject to:
    http://docs.yahoo.com/info/terms/