Expose allocation tracing over service protocol

runtime/observatory/tests/service/get_cpu_profile_test.dart

+// Copyright (c) 2015, the Dart project authors.  Please see the AUTHORS file
+// for details. All rights reserved. Use of this source code is governed by a
+// BSD-style license that can be found in the LICENSE file.
+// VMOptions=--compile_all --error_on_bad_type --error_on_bad_override
+
+import 'dart:developer';
+import 'package:observatory/service_io.dart';
+import 'package:observatory/cpu_profile.dart';
+import 'package:unittest/unittest.dart';
+
+import 'test_helper.dart';
+
+class Foo {
+  Foo() {
+    print('new Foo');
+  }
+}
+
+void test() {
+  // Initial.
+  debugger();
+  // Toggle on.
+  debugger();
+  // Step.
+  debugger();
+  // Allocation.
+  new Foo();
+  debugger();
+}
+
+var tests = [
+  hasStoppedAtBreakpoint,

[turnidge, 2015/07/13 23:37:48]

Can you add blank lines between tests?  it helps me to parse it visually.

[Cutch, 2015/07/14 15:53:47]

Done.

+  // Initial.
+  (Isolate isolate) async {
+    // Verify initial state of 'Foo'.
+    var fooClass = await getRootClass(isolate, 'Foo');
+    expect(fooClass, isNotNull);
+    expect(fooClass.name, equals('Foo'));
+    print(fooClass.id);
+    expect(fooClass.traceAllocations, isFalse);
+  },
+  resumeIsolate,
+  hasStoppedAtBreakpoint,

[turnidge, 2015/07/13 23:37:48]

It looks to me like this resume/breakpoint is unnecessary -- there is no need to
advance program state to check that setTraceAllocations works.

[Cutch, 2015/07/14 15:53:47]

There is an issue between debugger stepping and allocation tracing- I need to
step twice (or allocate twice) in order for the allocation stub code to switch
over.

+  // Toggle on.
+  (Isolate isolate) async {
+    // Enable allocation tracing for 'Foo'.
+    var fooClass = await getRootClass(isolate, 'Foo');
+    await fooClass.reload();
+    expect(fooClass.traceAllocations, isFalse);
+    await fooClass.setTraceAllocations(true);
+    await fooClass.reload();
+    expect(fooClass.traceAllocations, isTrue);
+  },
+  resumeIsolate,
+  hasStoppedAtBreakpoint,
+  // Step.

[turnidge, 2015/07/13 23:37:48]

Calling this Step is inaccurate, since we aren't stepping.

[Cutch, 2015/07/14 15:53:47]

Done.

+  resumeIsolate,
+  hasStoppedAtBreakpoint,
+  // Allocation profile.
+  (Isolate isolate) async {
+    var fooClass = await getRootClass(isolate, 'Foo');
+    await fooClass.reload();
+    expect(fooClass.traceAllocations, isTrue);
+    var profileResponse = await fooClass.getAllocationProfile();
+    expect(profileResponse, isNotNull);
+    expect(profileResponse['type'], equals('_CpuProfile'));
+    await fooClass.setTraceAllocations(false);
+    await fooClass.reload();
+    expect(fooClass.traceAllocations, isFalse);
+    CpuProfile cpuProfile = new CpuProfile();

[turnidge, 2015/07/13 23:37:48]

I think we should change the name CpuProfile to something else now.  Not sure
what, maybe TreeProfile, CallTreeProfile, SampleProfile, ...

[Cutch, 2015/07/14 15:53:47]

I agree. The CPU profile is slowly becoming a general stack trace sampler. I've
added a TODO (in cpu_profile.dart) because doing the rename right will touch
lots of files (and rename some files).

+    cpuProfile.load(isolate, profileResponse);
+    cpuProfile.buildCodeCallerAndCallees();
+    cpuProfile.buildFunctionCallerAndCallees();
+    var tree = cpuProfile.loadCodeTree('exclusive');
+    var node = tree.root;
+    var expected =
+        ['Root', 'test', 'test', '_FunctionImpl.call', 'runIsolateTests'];
+    for (var i = 0; i < expected.length; i++) {
+      expect(node.profileCode.code.name, equals(expected[i]));
+      // Depth first traversal.
+      if (node.children.length == 0) {
+        node = null;
+      } else {
+        node = node.children[0];
+      }
+      expect(node, isNotNull);
+    }
+  },
+  resumeIsolate,
+];
+
+main(args) async => runIsolateTests(args, tests, testeeConcurrent:test);