LayoutTests/media/encrypted-media/encrypted-media-lifetime-mediakeys.html - Issue 180203002: Add EME content test that forces garbage collection

Side by Side Diff: LayoutTests/media/encrypted-media/encrypted-media-lifetime-mediakeys.html

Issue 180203002: Add EME content test that forces garbage collection (Closed) Base URL: https://chromium.googlesource.com/chromium/blink.git@master

Patch Set: count ActiveDOMObjects Created 6 years, 9 months ago

Use n/p to move between diff chunks; N/P to move between comments. Draft comments are only viewable by you.

Jump to:

View unified diff | Download patch

OLD	NEW
(Empty)
	1 <!DOCTYPE html>

	2 <html>

	3 <head>

	4 <title>Test MediaKeys lifetime</title>

	5 <script src="../../resources/testharness.js"></script>

	6 <script src="../../resources/testharnessreport.js"></script>

	7 <script src="../../resources/gc.js"></script>

	8 <script src="encrypted-media-utils.js"></script>

	9 </head>

	10 <body>

	11 <div id="log"></div>

	12 <script>

	13 // Since MediaKeys are not ActiveDOMObjects, it is hard to

	14 // determine when they are garbage collected. For this test, we add

	15 // a MediaKeySession (which are ActiveDOMObjects) to each one so
	ddorwin 2014/02/27 22:28:17 You should explain why this works and what we assu You should explain why this works and what we assume: MKS objects will not be GC'd as long as their MK is around. jrummell 2014/02/28 00:08:38 Done. Show quoted text On 2014/02/27 22:28:17, ddorwin wrote: > You should explain why this works and what we assume: MKS objects will not be > GC'd as long as their MK is around. Done.
	16 // we can count.

	17 test(function()

	18 {

	19 var mediaKeys;

	20 var initData = stringToUint8Array("mock");

	21 var startingObjects = window.internals.activeDOMObjectCount(docu ment);

	22

	23 function objectsCreated()
	ddorwin 2014/02/27 22:28:17 This sounds like an event handler. :) Maybe numObj This sounds like an event handler. :) Maybe numObjectsCreated. jrummell 2014/02/28 00:08:38 Done. Show quoted text On 2014/02/27 22:28:17, ddorwin wrote: > This sounds like an event handler. :) Maybe numObjectsCreated. Done.
	24 {

	25 return window.internals.activeDOMObjectCount(document) - sta rtingObjects;

	26 }

	27

	28 // Create a MediaKeys object.
	ddorwin 2014/02/27 22:28:17 We could probably have a simple test that releases We could probably have a simple test that releases and gc()'s the empty MK immediately. jrummell 2014/02/28 00:08:38 Done. Show quoted text On 2014/02/27 22:28:17, ddorwin wrote: > We could probably have a simple test that releases and gc()'s the empty MK > immediately. Done.
	29 mediaKeys = new MediaKeys("org.w3.clearkey");

	30 mediaKeys.createSession("video/webm", initData);
	ddorwin 2014/02/27 22:28:17 Maybe this is just testing the rest of Blink, but Maybe this is just testing the rest of Blink, but we could gc() before here to ensure the MK does not go away and mediaKeys remains valid. WDYT? jrummell 2014/02/28 00:08:38 Done. Show quoted text On 2014/02/27 22:28:17, ddorwin wrote: > Maybe this is just testing the rest of Blink, but we could gc() before here to > ensure the MK does not go away and mediaKeys remains valid. WDYT? Done.
	31 assert_not_equals(mediaKeys, null);
	ddorwin 2014/02/27 22:28:17 not really necessary. not really necessary. jrummell 2014/02/28 00:08:38 Done. Show quoted text On 2014/02/27 22:28:17, ddorwin wrote: > not really necessary. Done.
	32 assert_equals(mediaKeys.keySystem, 'org.w3.clearkey');
	ddorwin 2014/02/27 22:28:17 This is probably checking that the mediaKeys is va This is probably checking that the mediaKeys is valid (otherwise, it's just redundant with other spec tests), but that doesn't make much sense until after a gc(). jrummell 2014/02/28 00:08:38 Done. Show quoted text On 2014/02/27 22:28:17, ddorwin wrote: > This is probably checking that the mediaKeys is valid (otherwise, it's just > redundant with other spec tests), but that doesn't make much sense until after a > gc(). Done.
	33 assert_equals(objectsCreated(), 1);

	34

	35 // Run gc(), should not affect MediaKeys object since we have
	ddorwin 2014/02/27 22:28:17 This is basically the test in my comment at line 3 This is basically the test in my comment at line 30. BUT, we shouldn't have created the session yet. jrummell 2014/02/28 00:08:38 Done. Show quoted text On 2014/02/27 22:28:17, ddorwin wrote: > This is basically the test in my comment at line 30. BUT, we shouldn't have > created the session yet. Done.
	36 // a reference to it.

	37 gc();

	38 assert_not_equals(mediaKeys, null);

	39 assert_equals(mediaKeys.keySystem, 'org.w3.clearkey');

	40 assert_equals(objectsCreated(), 1);

	41

	42 // Drop reference to the MediaKeys object and run gc again.

	43 // Object should be collected this time. Running gc() twice
	ddorwin 2014/02/27 22:28:17 This is basically my comment at line 28. BUT, we s This is basically my comment at line 28. BUT, we shouldn't have created the session yet. jrummell 2014/02/28 00:08:38 Done. Show quoted text On 2014/02/27 22:28:17, ddorwin wrote: > This is basically my comment at line 28. BUT, we shouldn't have > created the session yet. Done.
	44 // as the first pass may not actually get the created session.

	45 mediaKeys = null;

	46 gc();

	47 gc();
	ddorwin 2014/02/27 22:28:17 explain explain jrummell 2014/02/28 00:08:38 Expanded comment. Show quoted text On 2014/02/27 22:28:17, ddorwin wrote: > explain Expanded comment.
	48 assert_equals(objectsCreated(), 0);

	49
	ddorwin 2014/02/27 22:28:17 We should probably rerun the above tests here but We should probably rerun the above tests here but with the session. Really, this would be better as a separate test(), though. jrummell 2014/02/28 00:08:38 Not sure the benefit of a separate test. Show quoted text On 2014/02/27 22:28:17, ddorwin wrote: > We should probably rerun the above tests here but with the session. > > Really, this would be better as a separate test(), though. Not sure the benefit of a separate test.
	50 // Create a large number of MediaKeys objects.
	ddorwin 2014/02/27 22:28:17 I'm not sure what this does. If we keep it, it sho I'm not sure what this does. If we keep it, it should be a separate test(). jrummell 2014/02/28 00:08:38 ditto. Show quoted text On 2014/02/27 22:28:17, ddorwin wrote: > I'm not sure what this does. > If we keep it, it should be a separate test(). ditto.
	51 for (var i = 0; i < 100; ++i) {

	52 mediaKeys = new MediaKeys("org.w3.clearkey");

	53 mediaKeys.createSession("video/webm", initData);

	54 }

	55 assert_equals(objectsCreated(), 100);

	56

	57 // All but the last one created should be garbage collected.

	58 gc();

	59 gc();

	60 assert_equals(objectsCreated(), 1);

	61

	62 // Last MediaKeys object created should still be referenced.

	63 assert_not_equals(mediaKeys, null);

	64 assert_equals(mediaKeys.keySystem, 'org.w3.clearkey');

	65

	66 // Release the last MediaKeys object created.

	67 mediaKeys = null;

	68 gc();

	69 gc();

	70 assert_equals(objectsCreated(), 0);

	71 }, "MediaKeys lifetime");

	72 </script>

	73 </body>

	74 </html>

OLD	NEW

« no previous file with comments | « no previous file | LayoutTests/media/encrypted-media/encrypted-media-lifetime-mediakeys-expected.txt » ('j') | no next file with comments »