[antlr-interest] Bug? "new (new" code generated.

[Steve Bennett]

On 2/2/08, Mark Volkmann <r.mark.volkmann at gmail.com> wrote:
> I think I ran into this same problem. I believe the following is a
> true statement.
>
> If your lexer grammar and parser grammar are in separate files then in
> order to debug the parser grammar using ANTLRWorks, you have to use
> remote debugging.

If so, then the following is also a true statement:

The benefits of having a lexer grammar and parser grammar in separate
files are outweighed by the inconvenience of not being able to use
ANTLRworks.

:)

I think one of the difficulties I'm having is that I don't *have* a
"build script" to modify. I guess there's an unstated, fairly
reasonable assumption that one is attempting to build a parser that
does something: hence one sooner or later has to write a program that
does something with it. Hence, one has a build infrastructure with a
script of some kind.

Unfortunately, I don't have any of that, because I"m not attempting to
build a working program here; my goal is purely to produce a grammar
that describes a certain language. I'm using ANTLR primarily to verify
that the grammar is in fact correct. I don't really expect to use the
generated Java file for anything.

All of this explains why I panic whenever the ANTLRworks IDE fails for
some reason - I don't have much way of investigating the problem. But
I guess I'll have to look at setting up a more serious build
infrastructure than I have so far.

(in the short term, I've remerged my two grammars)

Steve