EME: Implement Session Closed algorithm

third_party/WebKit/LayoutTests/media/encrypted-media/encrypted-media-keystatuses.html

Index: third_party/WebKit/LayoutTests/media/encrypted-media/encrypted-media-keystatuses.html
diff --git a/third_party/WebKit/LayoutTests/media/encrypted-media/encrypted-media-keystatuses.html b/third_party/WebKit/LayoutTests/media/encrypted-media/encrypted-media-keystatuses.html
index fe5f1a939484ba15c898373e347e212c633220b5..4546f6a99e9c76eda3c28882b81074bb6b079957 100644
--- a/third_party/WebKit/LayoutTests/media/encrypted-media/encrypted-media-keystatuses.html
+++ b/third_party/WebKit/LayoutTests/media/encrypted-media/encrypted-media-keystatuses.html
@@ -142,8 +142,10 @@
                     else
                     {
                         // The second keystatuseschange (caused by close())
-                        // should not have any keys.
+                        // should not have any keys, and the expiration time
+                        // should be NaN.
                         assert_equals(mediaKeySession.keyStatuses.size, 0);
+                        assert_true(isNaN(mediaKeySession.expiration));

[ddorwin, 2016/10/11 01:41:07]

We might need to update the spec test.

[jrummell, 2016/10/11 18:47:35]

I wanted to make sure that we verified that this was NaN after the close. But
since clearkey doesn't set expiration, it really doesn't make a difference (it's
always NaN). But in the future we may, so this will catch any regression.

                         test.done();
                     }
                 }