[antlr-interest] Grammars with too many keywords?

John Green greenj at ix.netcom.com
Thu Mar 1 16:07:15 PST 2007


Vadim Tropashko wrote:
> BTW, what is the biggest antlr showcase?  

Using Antlr 2.7.6 to generate C++ code, I have to split the parser into 8 compile units for it to compile. I've no idea what I'd have to do in Java.

I do use Java output from Antlr as well, but just for treewalkers, not for the parser. Even at that, the treewalkers are pretty huge, and pushing the limits. I have to trim the data from the static initialization of _tokenNames, because that generates too large of a static initializer (> 64k).

My grammar has over 1100 token types. It's for a database language called "ABL", from Progress Software.

Here's the base grammar for the tree parser, in case you are curious about the size:
  http://www.joanju.com/dist/docs/tree_spec.html

Cheers
-- 
John Green, Partner, Joanju Software Canada
408-33478 Roberts Avenue, Abbotsford, BC, V2S 8E1, Canada
Bus/Mob: 1-604-767-8587, Fax: 1-425-609-1284
http://www.joanju.com/


More information about the antlr-interest mailing list