Overlap type arguments of a type with the type arguments of its super type

Overlap type arguments of a type with the type arguments of its super type
when possible. Overlapping may be partial.
Add flag --overlap-type-arguments controlling this optimization for debugging
and performance measurement (default is true).

[hausner, 2013-10-11 18:00:58]

LGTM but I suggest waiting for a second pair of eyes to take a look as well.

https://codereview.chromium.org/26955002/diff/1/runtime/vm/object.cc
File runtime/vm/object.cc (right):

https://codereview.chromium.org/26955002/diff/1/runtime/vm/object.cc#newcode1710
runtime/vm/object.cc:1710: intptr_t Class::NumOwnTypeArguments() const {
As discussed offline, it may be worth it to cache this computed value in the
class object. (Separate change.)

https://codereview.chromium.org/26955002/diff/1/runtime/vm/object.cc#newcode1...
runtime/vm/object.cc:11714: return
String::Handle(name()).Equals(String::Handle(other_type_param.name()));
Can't you just compare raw pointers here? The names are canonicalized strings.

[regis, 2013-10-11 18:36:09]

Thanks!

If other reviewers have more comments, I'll address them in the follow up change
caching the number of type arguments in Class objects.

https://codereview.chromium.org/26955002/diff/1/runtime/vm/object.cc
File runtime/vm/object.cc (right):

https://codereview.chromium.org/26955002/diff/1/runtime/vm/object.cc#newcode1710
runtime/vm/object.cc:1710: intptr_t Class::NumOwnTypeArguments() const {
On 2013/10/11 18:00:58, hausner wrote:
> As discussed offline, it may be worth it to cache this computed value in the
> class object. (Separate change.)

Yes, it is worth doing (in another change).

https://codereview.chromium.org/26955002/diff/1/runtime/vm/object.cc#newcode1...
runtime/vm/object.cc:11714: return
String::Handle(name()).Equals(String::Handle(other_type_param.name()));
On 2013/10/11 18:00:58, hausner wrote:
> Can't you just compare raw pointers here? The names are canonicalized strings.

Done.

[regis, 2013-10-11 18:50:55]

I'm closing this cl manually:

svn: Commit failed unexpectedly --- please try again later (1)

... it was committed anyway as r28539.