CORS-RFC1918: Introduce the 'addressSpace' IDL attributes.

third_party/WebKit/LayoutTests/http/tests/security/cors-rfc1918/resources/post-addressspace-from-worker.html

Index: third_party/WebKit/LayoutTests/http/tests/security/cors-rfc1918/resources/post-addressspace-from-worker.html
diff --git a/third_party/WebKit/LayoutTests/http/tests/security/cors-rfc1918/resources/post-addressspace-from-worker.html b/third_party/WebKit/LayoutTests/http/tests/security/cors-rfc1918/resources/post-addressspace-from-worker.html
new file mode 100644
index 0000000000000000000000000000000000000000..43681d2a217b5e259dcc00a042aeca27a27259e0
--- /dev/null
+++ b/third_party/WebKit/LayoutTests/http/tests/security/cors-rfc1918/resources/post-addressspace-from-worker.html
@@ -0,0 +1,7 @@
+<script>
+    var w = new Worker("/security/cors-rfc1918/resources/post-addressspace-to-owner.js");
+    w.onmessage = function (e) {
+        window.parent.postMessage(e.data, "*");
+    }
+    w.postMessage("go");

[philipj_slow, 2016/03/04 04:35:08]

Why this little indirection? Timing problem?

[Mike West, 2016/03/04 05:53:54]

I wasn't sure if the handler would catch a `postMessage` from the Worker if its
startup was racing this code. I didn't even try it without this little poke.
Would it have worked?

[philipj_slow, 2016/03/04 10:42:12]

Oh, I misread this. I thought the "go" message would end up at the handler right
above here, but it's for the worker script. But the race condition cannot happen
I think, because even if the worker starts and posts its message before the
message event handler is added on the next line, that message can only be
processed and turned into an event dispatch after this script has finished and
returned to the message loop. Right?

+</script>