Instance equality

* Rename misleading Instance::Equals to CanonicalizeEquals and document it.
* Keep 'Equals' alias for String and Integer.
* Add Instance::OperatorEquals that calls or mimics Object.operator==.

This is partly in preparation for internalizing HashMap/Set.

R=hausner@google.com, rmacnak@google.com

Committed: https://code.google.com/p/dart/source/detail?r=36066

[hausner, 2014-05-12 17:09:29]

LGTM w/Q

https://codereview.chromium.org/274333002/diff/60001/runtime/vm/object.h
File runtime/vm/object.h (right):

https://codereview.chromium.org/274333002/diff/60001/runtime/vm/object.h#newc...
runtime/vm/object.h:4136: // 3. CanonicalizeEquals: used to canonicalize
compile-time constants, e.g.,
Would it make sense to rename this as BitwiseEquals throughout?

[rmacnak, 2014-05-12 17:15:17]

lgtm

https://codereview.chromium.org/274333002/diff/60001/runtime/vm/object.h
File runtime/vm/object.h (right):

https://codereview.chromium.org/274333002/diff/60001/runtime/vm/object.h#newc...
runtime/vm/object.h:4134: // 1. OperatorEquals: true iff 'this == other' is true
in Dart code.
Reminds me of Self's equivalent (=), indistinguishable (==), and identical
(===).

[koda, 2014-05-12 17:41:43]

https://codereview.chromium.org/274333002/diff/60001/runtime/vm/object.h
File runtime/vm/object.h (right):

https://codereview.chromium.org/274333002/diff/60001/runtime/vm/object.h#newc...
runtime/vm/object.h:4134: // 1. OperatorEquals: true iff 'this == other' is true
in Dart code.
On 2014/05/12 17:15:17, rmacnak wrote:
> Reminds me of Self's equivalent (=), indistinguishable (==), and identical
> (===).

Maybe we should add SameSameButDifferent :)

https://codereview.chromium.org/274333002/diff/60001/runtime/vm/object.h#newc...
runtime/vm/object.h:4136: // 3. CanonicalizeEquals: used to canonicalize
compile-time constants, e.g.,
On 2014/05/12 17:09:30, hausner wrote:
> Would it make sense to rename this as BitwiseEquals throughout?

That is almost true, but not technically true for GrowableObjectArray, where it
actually looks inside its single Array field. Also, there still seems to be some
dispute about whether we really should canonicalize NaNs using bitwise equality
(issue 18738).

I mostly want to minimize the risk that people make false assumptions.

[koda, 2014-05-12 20:53:27]

Committed patchset #4 manually as r36066 (presubmit successful).