Hi!

-------- Original-Nachricht --------
> Datum: Thu, 17 Apr 2008 10:40:32 +0200
> Von: "Thomas N." <[EMAIL PROTECTED]>
> An: [email protected]
> Betreff: Re: [argouml-dev] Re: ANTLR java.g (was JavaRecognizer)

> Hi Luis!
> 
> > Does ANTLR 3 use a different package? If not I think this will be 
> > difficult, since, AFAIK, the classloader is the same for ArgoUML core 
> > and the remaining modules.
> 
> The Antlr jar file antlrall.jar need to be replaced. Since it belongs to
> the core of ArgoUML and other modules depend on it, it must stay. So, there
> must be found a way for the new Antlr 3 based Java import, to use it's own
> Antlr jar instead of the existing one, both for the build and at runtime.
> (Later, there could be an approach to move Antlr away from the core.)
> I'd definitely need help here!

Fork in CVS/SVN? Otherwise you would have to rename the methods in Antlr. Not so
easy, if you want to update the Antlr stuff regularly...

> > Also, will we get an update of the classfile reveng grammar? It would be
> > nice if that would be updated also to keep consistency amongst the java 
> > reveng functionality.
> 
> I have not planned to do so, as I'm not familiar with that grammar.
> Andreas, you contributed it, would this be difficult? Can you give me a hint 
> so
> that I could deal with it?

Depends on the modifications, you want to do. Antlr 3 shouldn't pose a big 
problem.
Or has the Java bytecode changed? That's a bigger problem, since you have to
study the entire bytecode specs, and back then, when I wrote the grammar, it was
not very well documented (had to read a lot between the lines and study some 
compiled
code).

--<snip>--

> > But, I may give it a try after finishing the work I'm doing in the 
> > profile subsystem.
> 
> The problem here is, that there is no C++ grammar for Antlr 3 so far, I
> think. For Java I found one, so my (nontrivial) task is to put the ArgoUML
> code into it. For the C++ grammar the approach could be different: change the
> current C++ grammar with ArgoUML code in it so that it could be processed
> by Antlr 3. Requires deeper Antlr knowledge, and the result would not be a
> Antlr 3 optimizes grammar.
> 
> My suggestion: For C++, stay with Antlr 2.7.7 (unless someone comes up
> with a C++ grammar newly written for Antlr 3), and let us benefit from it by
> setting up a working enwironment with both Java and C++ import based on two
> different Parser generators! Would be a nice showcase to prove that ArgoUML
> import functionality is independent from a certain parser generator.

Problem is, that we need C++ support. Just yesterday, Jan posted some comments
on the CRRC list, about problems with Argo. And those folks won't be happy with
Argo unless they have a good C++ plus, I guess. Just like some companies etc.

I would like to have more feedback from the users on their needs etc, that's why
I like the idea of the Argo user website so much. It would be cool, if some OSS
projects (that use Argo) would have a contact person, that reports experiences
with Argo etc. and would be willing to answer questions from the Argo folks, 
helps
them with testing etc.

Ciao,
Andreas

-- 
Psst! Geheimtipp: Online Games kostenlos spielen bei den GMX Free Games! 
http://games.entertainment.gmx.net/de/entertainment/games/free

---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Reply via email to