Sounds very promising, was you be able to run a swf built thru this new 
workflow already or is it still too much in early stage at the moment ?


Envoyé depuis un mobile Samsung.

<div>-------- Message d'origine --------</div><div>De : Christofer Dutz 
<christofer.d...@c-ware.de> </div><div>Date :02/08/2014  7:16 PM  (GMT+01:00) 
</div><div>A : "'dev@flex.apache.org'" <dev@flex.apache.org> </div><div>Objet : 
Falcon Antlr4 refactoring </div><div>
</div>
Hi guys, I just wanted to give you an update on my Antlr4-Falcon-Simplification 
I have been working on:

I have actually come quite far with my work on falcon. Unfortunately I had to 
put quite a lot of work into JBurg first.

I created a maven plugin for JBurg and ajdusted JBurgs build tob e able to 
publish Maven artifacts. The plugin allows us to let JBurg use whatever libs it 
needs without interfering with the rest. Then I refactored JBurg to use 
stringtemplate 4. Then I had to migrate our Falcon JBurg rules to JBurg 2.0 
syntax. After that I updated Falcon to work with JBurgs DefaultAdapter instead 
of our IASNodeAdapter and could hereby completely remove any dependency on the 
JBurg libs. Now I still have to fine tune the templates as the CmcEmitter and 
CSSEmitter still have some small problems.

I already have converted all grammars to Antlr4 and the generated parsers seem 
to be parsing their input niceley :)

As soon as the BURM generation works, I'll continue directly building the 
Falcon AST in the first Antlr4 parser run.

And as a result my version builds nicely with Maven ;-)

Chris


PS: All my comitted changes are in Falcons "falcon-antlr4" branch ... but you 
need my updates to JBurg which tom will hopefully apply after I finished my 
work and sent him all of my changes. Perhaps I'll have to help setup a sonatype 
account to which the artifacts can be published and help publishing the 2.0 
release as soon as he has finished it.

Reply via email to