[I'm new to ANTLR, so please let me know if this is a known issue, and if
not, if I should attempt to file a bug for it.]

I'm trying to add ANTLR support to
Boost.Build<http://www.boost.org/doc/html/bbv2.html>,
based on the ANTLR -depend feature, but -depend seems to output incorrect
target filenames for language=C. For example, for a lexer grammar X, it
incorrect adds the suffix "Lexer" to the header filename:

X.c : X.g
.\X.tokens : X.g
XLexer.h : X.g

For grammars using "import", the problem is a bit more severe. For composite
grammar B, which imports lexer grammar L and parser grammar A, -depend
outputs:

B.g: L.g, A.g
BParser.c : B.g
.\B.tokens : B.g
BParser.h : B.g
BLexer.c : B.g
BLexer.h : B.g
B_L.c : B.g
B_A.c : B.g

However, the actual files generated by ANTLR, with just B.g on the command
line, are:

A.tokens
B.tokens
BParser.c
BParser.h
B_A.c
B_A.h

Note that the lexer isn't generated at all. Is this to be expected? Also
putting L.g on the command line addresses this, but then of course the
output files are L.h and L.c, which don't match the -depend output for B.g.

Thanks for any guidance,
Trevor
_______________________________________________
antlr-dev mailing list
[email protected]
http://www.antlr.org/mailman/listinfo/antlr-dev

Reply via email to