[antlr-interest] C code generation differences on Windows

[Michael J Iatauro]

Using ANTLR 3.1.3, I'm seeing some odd code generation differences in 
the tree parser between Windows and *nix platforms, specifically having 
to do with token names and IDs, which are causing my tree parser to 
reject all incoming trees.  There are several tokens missing in the 
generated Windows header, and those missing tokens as well as several 
others aren't given strings in the TreeTokenNames array in the C code. 
Even more weirdly, the token that's always expected to be the root is 
given different IDs, even between the parser and the tree parser, which 
is causing the rejection.  Has anybody else seen anything like this, or 
have a sense of what's going on?
Thanks!

~MJI