Access to imports in the VM's runtime mirrors. Extend test coverage of the source mirrors.

Access to imports in the VM's runtime mirrors. Extend test coverage of the source mirrors.

BUG=http://dartbug.com/10360
R=gbracha@google.com, hausner@google.com

Committed: https://code.google.com/p/dart/source/detail?r=33397

[rmacnak, 2014-02-04 22:47:19]

RFC regarding runtime versus source mirror differences.

https://codereview.chromium.org/154733003/diff/60001/sdk/lib/mirrors/mirrors....
File sdk/lib/mirrors/mirrors.dart (right):

https://codereview.chromium.org/154733003/diff/60001/sdk/lib/mirrors/mirrors....
sdk/lib/mirrors/mirrors.dart:628: 
This is the same API as the source mirrors, with String substituted by Symbol.

https://codereview.chromium.org/154733003/diff/60001/tests/compiler/dart2js/d...
File tests/compiler/dart2js/dart2js.status (right):

https://codereview.chromium.org/154733003/diff/60001/tests/compiler/dart2js/d...
tests/compiler/dart2js/dart2js.status:25: mirrors/library_exports_hidden_test:
Fail
The VM and the source mirrors disagree about whether to include the implicit
import of dart:core (VM: yes, source mirrors: no). The VM does not remember
whether the import is implicit, nor the groupings of the combinators.

[gbracha, 2014-02-25 01:56:44]

The API and tests lgtm as far as they go.  I note that you never test the
dependencies combinators to see if they are shows or hides,or who their library
of origin is.

[rmacnak, 2014-02-25 01:59:01]

+Matthias re implementation

[hausner, 2014-03-06 20:50:15]

Curious: why are you attaching the metadata to the Namespace object rather than
the library, where all other metadata is?

[rmacnak, 2014-03-06 20:58:54]

On 2014/03/06 20:50:15, hausner wrote:
> Curious: why are you attaching the metadata to the Namespace object rather
than
> the library, where all other metadata is?

The other metadata is associated with declarations that have reliable names (or
the library itself of which there is only one per library). The imports and
exports can have conflicting names (prefixes) or no name.

[hausner, 2014-03-06 21:26:00]

LGTM.

[rmacnak, 2014-03-06 22:02:47]

Committed patchset #5 manually as r33397 (presubmit successful).