[Extension ServiceWorkers] Blink: Passing v8::context to extensions dispatcher

Source/core/workers/WorkerThread.cpp

Index: Source/core/workers/WorkerThread.cpp
diff --git a/Source/core/workers/WorkerThread.cpp b/Source/core/workers/WorkerThread.cpp
index 5c0b410876bd36448b622d837a969523647b7de1..afb588103ee8a920f770f8e732a243ec76a7f982 100644
--- a/Source/core/workers/WorkerThread.cpp
+++ b/Source/core/workers/WorkerThread.cpp
@@ -255,6 +255,9 @@ void WorkerThread::initialize(PassOwnPtr<WorkerThreadStartupData> startupData)
     m_workerGlobalScope->didEvaluateWorkerScript();
     m_workerReportingProxy.didEvaluateWorkerScript(success);
 
+    ScriptState::Scope scope(script->scriptState());
+    m_workerReportingProxy.didInitializeWorkerContext(script->context());

[falken, 2015/08/03 01:55:14]

It seems to make sense to call didInitializeWorkerContext where
workerContextInitialized() is called in line 251.

As written this waits for evaluation to complete. So if the listener wants to
set some flag on the context as the the chromium patch seems to do, the initial
evaluation of the script wouldn't be under the flag (only events that fired
after would have it).

Also consider the while(true) {} case. If the script never terminates,
didInitializeWorkerContext() is never called in the current patch. In the case
of SW there's a browser-side timeout that sends a message to the renderer to
terminate execution and shutdown the thread. We should verify the sequence of
what happens in that case: I'm not sure whether script->context() is valid after
termination.

+
     postInitialize();
 
     m_webScheduler->postIdleTaskAfterWakeup(FROM_HERE, WTF::bind<double>(&WorkerThread::performIdleWork, this));