[antlr-interest] Proper way to catch parse errors

[Martin Olsson]

Hi,

I'm developing an Eclipse plugin and I'm using ANTLR for refactoring etc.
Currently parse errors are correctly propagated into the Eclipse problems
view, but I would like to clean up the code a little bit.

In particular I don't understand why start() throws RecognitionException
but at the same time my method override reportError() is also called? Why
is there two different ways to catch the same error? Currently, I ignore
all exceptions thrown by start() which I feel is a little bit ugly.


Sincerly,
Martin

(below is the first part of my parser .g code)
------------------------------------------------------------------


class DUPParser extends Parser;
options {
	k = 4;		// The lookahead depth
}
{
	protected MyEditor editor = null;

	public MyParser(TokenStream lexer, MyEditor editor) {
	  this(lexer, 4);
	  this.editor = editor;
	}

	public void reportError(RecognitionException e) {
		editor.reportError(e.getMessage(),
                                   e.getLine(),
                                   e.getColumn());
	}

	public final void parseIntoModule(MyModule module) {
		try {
			start(module);
		} catch (Exception e) {
                     /* not used?! */
                }
	}
}