Remove duplicate resolveMember, fix type substitution for subtype of generic class

Remove duplicate resolveMember, fix type substitution for subtype of generic class. This should get our type checks closer to working for generics.

Committed: https://code.google.com/p/dart/source/detail?r=1810

[Jennifer Messerly, 2011-11-23 19:37:02]

tbr

[jimhug, 2011-11-23 19:43:30]

LGTM

More correct, less code - always nice to see.

However, note my comment below about what I fear is a nasty lingering issue.

http://codereview.chromium.org/8679014/diff/2001/frog/type.dart
File frog/type.dart (right):

http://codereview.chromium.org/8679014/diff/2001/frog/type.dart#newcode594
frog/type.dart:594: _subtypes.add(s.resolveTypeParams(this));
Does this always work?  It's clearly better than what I had (which was to just
use the raw types).  However, it seems that this would get confused when the
names of type params get changed around i.e.
class C<T> {}
class D<X, Y> extends C<Y> {}

Should this work, or do we need a TODO to figure this one out?

[Jennifer Messerly, 2011-11-23 23:39:36]

http://codereview.chromium.org/8679014/diff/2001/frog/type.dart
File frog/type.dart (right):

http://codereview.chromium.org/8679014/diff/2001/frog/type.dart#newcode594
frog/type.dart:594: _subtypes.add(s.resolveTypeParams(this));
On 2011/11/23 19:43:31, jimhug wrote:
> Does this always work?  It's clearly better than what I had (which was to just
> use the raw types).  However, it seems that this would get confused when the
> names of type params get changed around i.e.
> class C<T> {}
> class D<X, Y> extends C<Y> {}
> 
> Should this work, or do we need a TODO to figure this one out?

Doh. Great catch. That's going to be nasty. I'll see if I can come up with a
clean fix.