This is how i would translate 'nestedConstraint()' to a regular expression ID? ( (DOT | HASH) ID)* DOT LEFT_PARENT constraints RIGHT_PARENT And this is the description in his header "( ID ( DOT | HASH ) )* ID DOT LEFT_PAREN constraints RIGHT_PAREN" The difference is that 'nestedConstraint()' never checks if really starts with ID, so can starts with DOT or HASH, which contradicts his header. El 2014-01-30 22:18, Edson Tirelli escribió: understand your question? although I would implement it in a different way myself, the current implementation looks correct to me... am I missing something? Edson Casanueva <claudio.sanchez(a)andanta.es&gt; wrote: drools code and i have seen strange the method 'getNestedConstraintPrefixLenght()' in DRL6Parser.java: If i translate that function to a regular expression in antlr lexer, it would be "ID? ( (DOT | HASH) ID)* DOT LEFT_PARENT", which have few sense. header of 'nestedConstraint()' describes the expected the expected result as "( ID ( DOT | HASH ) )* ID DOT LEFT_PAREN constraints RIGHT_PAREN" : it internally calls to 'getNestedConstraintPrefixLenght()', consumes the numbers of items retuned by this, calls to 'constrains()' and cosumes a RIGHT_TOKEN. For me the expected results described in the header of 'nestedConstraint()' do not correspond with the real results. >> _______________________________________________ list >> rules-dev(a)lists.jboss.org https://lists.jboss.org/mailman/listinfo/rules-dev [1] > > -- Edson Tirelli and Intelligence Group _______________________________________________ list https://lists.jboss.org/mailman/listinfo/rules-dev [1] Links: ------ [1] https://lists.jboss.org/mailman/listinfo/rules-dev