Implement for-in.

Implement for-in.

Direct implementation with no optimizations.

Committed: https://code.google.com/p/dart/source/detail?r=3180

[ngeoffray, 2012-01-11 10:09:22]

LGTM!

http://codereview.chromium.org/9148021/diff/1/frog/leg/lib/core.dart
File frog/leg/lib/core.dart (right):

http://codereview.chromium.org/9148021/diff/1/frog/leg/lib/core.dart#newcode341
frog/leg/lib/core.dart:341: class ListIterator<T> /* implements Iterator<T> */ {
I'd prefer calling it ForeignListIterator.

http://codereview.chromium.org/9148021/diff/1/frog/leg/resolver.dart
File frog/leg/resolver.dart (right):

http://codereview.chromium.org/9148021/diff/1/frog/leg/resolver.dart#newcode501
frog/leg/resolver.dart:501: // TODO(lrn): A non-terminating error report would
be fine here.
TODO(lrn): Must change cancel to error.

Is error failing right now?

http://codereview.chromium.org/9148021/diff/1/frog/leg/resolver.dart#newcode502
frog/leg/resolver.dart:502: print(declaration);
Remove print.

http://codereview.chromium.org/9148021/diff/1/frog/leg/ssa/builder.dart
File frog/leg/ssa/builder.dart (right):

http://codereview.chromium.org/9148021/diff/1/frog/leg/ssa/builder.dart#newco...
frog/leg/ssa/builder.dart:1296: //     E <declaredIdentifier> = $iter.next();
Maybe mark E as optional in this statement? Something like:
E? <declaredIdentifier> ...

http://codereview.chromium.org/9148021/diff/1/frog/leg/ssa/builder.dart#newco...
frog/leg/ssa/builder.dart:1315: compiler.namer.instanceName(new
SourceString("hasNext"));
new SourceString -> const SourceString

http://codereview.chromium.org/9148021/diff/1/frog/leg/ssa/builder.dart#newco...
frog/leg/ssa/builder.dart:1331: compiler.namer.instanceName(new
SourceString("next"));
new -> const

http://codereview.chromium.org/9148021/diff/1/frog/leg/ssa/builder.dart#newco...
frog/leg/ssa/builder.dart:1337: compiler.unimplemented("SsaBuilder for loop with
aborting body");
add:
node: node

http://codereview.chromium.org/9148021/diff/1/frog/leg/ssa/builder.dart#newco...
frog/leg/ssa/builder.dart:1342: // We create an update block, even when we are
in a while loop. There the
when -> if.

http://codereview.chromium.org/9148021/diff/1/frog/leg/ssa/builder.dart#newco...
frog/leg/ssa/builder.dart:1342: // We create an update block, even when we are
in a while loop. There the
Remove 'There'

http://codereview.chromium.org/9148021/diff/1/frog/tests/leg_only/src/ForInTe...
File frog/tests/leg_only/src/ForInTest.dart (right):

http://codereview.chromium.org/9148021/diff/1/frog/tests/leg_only/src/ForInTe...
frog/tests/leg_only/src/ForInTest.dart:36: // Enable when we donno longer
generate code that bail out on
As discussed, please re-enable them now :)

[Lasse Reichstein, 2012-01-11 10:26:00]

http://codereview.chromium.org/9148021/diff/1/frog/leg/resolver.dart
File frog/leg/resolver.dart (right):

http://codereview.chromium.org/9148021/diff/1/frog/leg/resolver.dart#newcode501
frog/leg/resolver.dart:501: // TODO(lrn): A non-terminating error report would
be fine here.
No, I was guessing that cancel was the right thing.
Using error now.

http://codereview.chromium.org/9148021/diff/1/frog/leg/resolver.dart#newcode502
frog/leg/resolver.dart:502: print(declaration);
On 2012/01/11 10:09:22, ngeoffray wrote:
> Remove print.

Done.

http://codereview.chromium.org/9148021/diff/1/frog/leg/ssa/builder.dart
File frog/leg/ssa/builder.dart (right):

http://codereview.chromium.org/9148021/diff/1/frog/leg/ssa/builder.dart#newco...
frog/leg/ssa/builder.dart:1296: //     E <declaredIdentifier> = $iter.next();
I should make it optional in Iterator<E> too then.
I prefer this way, even though E might implicitly be "Dynamic".

http://codereview.chromium.org/9148021/diff/1/frog/tests/leg_only/src/ForInTe...
File frog/tests/leg_only/src/ForInTest.dart (right):

http://codereview.chromium.org/9148021/diff/1/frog/tests/leg_only/src/ForInTe...
frog/tests/leg_only/src/ForInTest.dart:36: // Enable when we donno longer
generate code that bail out on
Done.