[antlr-interest] Newbie: Confusion building C-Target project

[Jorge Fuente-Alba]

Hi.

I have a game project, currently being developed under MinGW.

Thanks to Jim I tried Cygwin and managed to build LibAntlr3c.
Then noticed that I was having the same issues building a simple test parser because MinGW doesn't seem to like libantlr3c headers at all.
Tried building a test parser under Cygwin and got dizzy.

I've talked with the guys about building the parser using cygwin as a library and then using said library under mingw. This puts a few bumps in our build process that I'd rather not have. Cygwin forces a unix-like environment that we'd rather not use. We're happily using CodeBlocks and DevCpp before that, both rely on MinGW and we've never had trouble before.

My questions are two:
1.- Is there a reason why MinGW doesn't like libantlr3c? I read something about differences between MinGW-make and Cygwin-make. Is there a way to get libantlr3c to build under mingw?
2.- if the only way to go is to build from Cygwin and then wrap for MinGW, can anyone explain why i'm getting undefined references to '_antlr3AsciiFileStreamNew', '_inifileLexerNew', '_antlr3CommonTokenStreamSourceNew' and '_inifileParserNew' ? Can anyone help me with a proper makefile? (I suspect include and lib paths, as I just did gcc main.c)

Included: Example parser main.c and grammar
main.c
-------
#include "inifileLexer.h"
#include "inifileParser.h"

int main(int argc, char* argv[])
{
    pANTLR3_INPUT_STREAM pIStream = antlr3AsciiFileStreamNew((pANTLR3_UINT8)"MyInputFile.txt");
    pinifileLexer pLexer = inifileLexerNew(pIStream);

    /* Not obvious: */
    pANTLR3_COMMON_TOKEN_STREAM pTokens = antlr3CommonTokenStreamSourceNew(0, pLexer->pLexer->tokSource);

    pinifileParser pParser = inifileParserNew(pTokens);

    pParser->start(pParser);

    return 0;
}

grammar: inifile.g
--------------------
grammar inifile;
options {language=C;}


start	:	header (rule|empty)*;
rule	:	STR '=' data;
data	:	(INTEGER | FLOAT | STR) NL;
header	:	'[' STR ']' NL;
empty	:	WS* NL;

STR	:	('a'..'z' | 'A'..'Z' | '0'..'9' | ' ')*;
INTEGER	:	('0'..'9')*;
FLOAT	:	('0'..'9')  '.' ('0'..'9') ;
NL	:	('\n' | '\r');
WS	:	(' ' | '\t' );

--------
Thanks for any help provided, I'd really hate to end up hacking a brittle custom parser as I know the source of bugs this tends to be. Plus I'm trying to get my teachers to use ANTLR for the classes and students use mostly DevCpp around here :)


Texto: ¡Concursa y gánate un boleto doble para el show de The Police en México! http://mx.starmedia.com/thepolice/