Description Victor Noël 2010-07-26 11:23:33 EDT

Hi,

I noticed a bug when using such a grammar (using a clean new Xtext project):
Model:
	('y' Parent1) | ('z' Parent2) | Test;

Test:
	l=ID;

Parent1:
	Sub1 | Sub2;

Parent2:
	Sub2 | Sub3;

Sub1:
	'a' x=ID;

Sub2:
	'b' x=ID;

Sub3:
	'c' x=ID;

The problem here is that Parent1 and Parent2 get the x property and for a reason I don't really understand, when the ecore implementation are generated, the package for the language contains duplicate features for Sub2:

  int SUB2__X = PARENT1__X;
  int SUB2__X = PARENT1_FEATURE_COUNT + 0;