Rule matching oddness with new CSS parser

[David Hyatt]

Ok, this behavior exactly matches the old parser (and all my test cases 
finally pass, yay!).

The old parser chains the sub-selectors inside simple selectors from 
left to right.  It then hands back the leftmost subselector.  This 
production matches that behavior:

specifier_list specifier {
	$$ = $1;
         CSSSelector *end = $1;
         while( end->tagHistory )
             end = end->tagHistory;
         end->relation = CSSSelector::SubSelector;
         end->tagHistory = $2;
     }

For complex selectors the old parser chained the simple selectors from 
right to left.  The rightmost link in a simple selector would point to 
the leftmost link in the next simple selector.  This production matches 
that behavior.

selector combinator simple_selector {
     	$$ = $3;
         CSSSelector *end = $3;
         while( end->tagHistory )
             end = end->tagHistory;
	end->relation = $2;
	end->tagHistory = $1;
     }

dave
(hyatt@apple.com)