On Mar 14, 4:23 pm, Glenn Boysko <[EMAIL PROTECTED]> wrote:
> Here's what I've done (and members of the group may add more insights/
> better ways if they know of them).
>
> First, here's a method I use to parse a JavaScript file into a
> ScriptOrFn node (which represents the root of the file contents):
>
> private ScriptOrFnNode parseJavascript(File file) throws
> IOException {
> // Try to open a reader to the file supplied...
> Reader reader = new FileReader(file);
>
> // Setup the compiler environment, error reporter...
> CompilerEnvirons compilerEnv = new CompilerEnvirons();
> ErrorReporter errorReporter = compilerEnv.getErrorReporter();
>
> // Create an instance of the parser...
> Parser parser = new Parser(compilerEnv, errorReporter);
>
> String sourceURI;
>
> try {
> sourceURI = file.getCanonicalPath();
> } catch (IOException e) {
> sourceURI = file.toString();
> }
>
> // Try to parse the reader...
> ScriptOrFnNode scriptOrFnNode = parser.parse(reader,
> sourceURI, 1);
>
> return scriptOrFnNode;
> }
>
> This top-level object represents the root "node" in this file. It
> extends the org.mozilla.javascript.Node class.
>
> From there, you have a rough syntax tree that you can use for your
> analysis of the JavaScript.
>
> To get a textual representation of this syntax tree, you would call
> the toStringTree method (and pass in the scriptOrFnNode as its
> parameter).
>
> Note that in the standard distribution, this returns null. To change
> that, I had to change a constant in Token.java:
>
> public static final boolean printTrees = true; // was false
>
> Not sure if there are easier ways to do what I have done.
>
> Regards,
> Glenn
>
> On Mar 14, 2:40 pm, [EMAIL PROTECTED] wrote:> On Mar 13, 2:17 pm, Glenn
> Boysko <[EMAIL PROTECTED]> wrote:
>
> > > Hello:
>
> > > I've been using Rhino for some time as a parser for JavaScript. The
> > > lack of a complete grammar for ANTLR makes Rhino a good option for me.
> > > I'm using it for things like specialized documentation tools (JavaDoc-
> > > style docs for JavaScript classes) or other code analysis.
>
> > > I'm sure that you get a lot of these types of questions. What do you
> > > recommend?
>
> > > For example, I've been using 1.5 and 1.6 in the past and found no node
> > > visitor class. Does one exist in 1.7? There doesn't seem to be a way
> > > to search for a node (by type) in 1.6 either.
>
> > > If you have any suggestions for writing JavaScript parsers based on
> > > Rhino, I'd really appreciate it.
>
> > > Thanks,
> > > Glenn
>
> > Hello Glenn,
>
> > Just yesterday I realized I need to parse JS in a project -- to
> > generate GraphViz files for example. Since you are doing it already,
> > could you point to any howto or documentation that the compiler-
> > challenged among us can use? Anything to avoid? Any of you experience
> > would be really useful.
>
> > Thanks!
> > Jaime
"printTrees" is intended only for debugging, but is a good way to see
what the IR is like.
--N
_______________________________________________
dev-tech-js-engine-rhino mailing list
[email protected]
https://lists.mozilla.org/listinfo/dev-tech-js-engine-rhino
