[antlr-interest] "code too large" / can't split AST output into smaller files

[Red Rackham]

I'm having same issue.  Have a big grammar (>900 tokens, >4K lines of parser 
grammar) it's translated from an older source and very messy & interconnected 
since it was generated by conversion script.  Was able to get ANTLR to create 
.java files, but the java files don't compile ("code too large").  With 
python output, the lexer works up to a point but mis-identifies a token and I 
haven't figured out why... I wanted to debug in ANTLRWorks but I can't since I 
can't get java to compile.  Catch 22?