Optimize VM JSON parser for memory use.

sdk/lib/core/iterable.dart

Index: sdk/lib/core/iterable.dart
diff --git a/sdk/lib/core/iterable.dart b/sdk/lib/core/iterable.dart
index 8ea9c99f825f1a6bb03048fa0d7f0fb54acc7464..8eb91bdea0be59383bc4fa0e3d34aa6f8421572e 100644
--- a/sdk/lib/core/iterable.dart
+++ b/sdk/lib/core/iterable.dart
@@ -44,9 +44,8 @@ abstract class Iterable<E> {
    * As an Iterable, [:new Iterable.generate(n, generator)):] is equivalent to
    * [:const [0, ..., n - 1].map(generator):]
    */
-  factory Iterable.generate(int count, E generator(int index)) {
-    return new _GeneratorIterable<E>(count, generator);
-  }
+  const factory Iterable.generate(int count, E generator(int index)) =
+    _GeneratorIterable<E>;
 
   /**
    * Returns an Iterator that iterates over this Iterable object.
@@ -291,7 +290,7 @@ class _GeneratorIterable<E> extends IterableBase<E>
                             implements EfficientLength {
   final int _count;
   final _Generator<E> _generator;
-  _GeneratorIterable(this._count, this._generator);
+  const _GeneratorIterable(this._count, this._generator);
   Iterator<E> get iterator => new _GeneratorIterator(_count, _generator);
   int get length => _count;
 }