LayoutTests/media/encrypted-media/encrypted-media-playback-multiple-sessions.html - Issue 203323006: Add layout test for EME WD using multiple sessions

Side by Side Diff: LayoutTests/media/encrypted-media/encrypted-media-playback-multiple-sessions.html

Issue 203323006: Add layout test for EME WD using multiple sessions (Closed) Base URL: https://chromium.googlesource.com/chromium/blink.git@master

Patch Set: New test file 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

« no previous file with comments | « LayoutTests/media/content/test-encrypted-2-keys.webm ('k') | LayoutTests/media/encrypted-media/encrypted-media-playback-multiple-sessions-expected.txt » ('j') | no next file with comments »
Toggle Intra-line Diffs ('i') | Expand Comments ('e') | Collapse Comments ('c') | Hide Comments ('s')

OLD	NEW
(Empty)
	1 <!DOCTYPE html>

	2 <html>

	3 <head>

	4 <title>Clear Key Playback with Multiple Sessions</title>

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

	6 <script src="../w3c-media-utils.js"></script>

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

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

	9 </head>

	10 <body>

	11 <video id="testVideo"></video>

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

	13 <p>Test playback of encrypted media using Clear Key key system with mult iple sessions.</p>

	14 <script>

	15 // Timeout for all tests to run.
	ddorwin 2014/03/21 06:00:20 It would be nice to have a single line for this ti It would be nice to have a single line for this timeout. The comment should fit on the line fine, though I'm not sure what exactly it means or that it's necessary. jrummell 2014/03/21 17:23:01 Done. Show quoted text On 2014/03/21 06:00:20, ddorwin wrote: > It would be nice to have a single line for this timeout. The comment should fit > on the line fine, though I'm not sure what exactly it means or that it's > necessary. Done.
	16 setup({ timeout: 60000 });
	ddorwin 2014/03/21 06:00:20 Oh, maybe this is from other files that ran multip Oh, maybe this is from other files that ran multiple tests. I don't think we need it here. (We already have line 112.) jrummell 2014/03/21 17:23:01 The testharness has 2 timeouts -- 1 for the timeou Show quoted text On 2014/03/21 06:00:20, ddorwin wrote: > Oh, maybe this is from other files that ran multiple tests. I don't think we > need it here. (We already have line 112.) The testharness has 2 timeouts -- 1 for the timeout of each test (default 2s), and 1 for the total time of all tests in the file (default 5s). We typically only have 1 test per file, although encrypted-media-v2-syntax.html has several. Without this line this test would timeout after 5 seconds, regardless of the bigger timeout value specified on line 112. ddorwin 2014/03/21 17:36:49 :( Thanks for the explanation. Show quoted text On 2014/03/21 17:23:01, jrummell wrote: > On 2014/03/21 06:00:20, ddorwin wrote: > > Oh, maybe this is from other files that ran multiple tests. I don't think we > > need it here. (We already have line 112.) > > The testharness has 2 timeouts -- 1 for the timeout of each test (default 2s), > and 1 for the total time of all tests in the file (default 5s). We typically > only have 1 test per file, although encrypted-media-v2-syntax.html has several. > Without this line this test would timeout after 5 seconds, regardless of the > bigger timeout value specified on line 112. :( Thanks for the explanation.
	17

	18 async_test(function(test)

	19 {

	20 var video = document.getElementById("testVideo");

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

	22 var audioMediaKeySession = null;

	23 var videoMediaKeySession = null;

	24 var audioSessionReadyReceived = false;

	25 var videoSessionReadyReceived = false;

	26

	27 // The 2 streams use different key ids and different keys.

	28 var audioKeyId = "1234567890123456";

	29 var audioKey = new Uint8Array([0x30, 0x30, 0x62, 0xF1, 0x68, 0x1 4, 0xD2, 0x7B,

	30 0x68, 0xEF, 0x12, 0x2A, 0xFC, 0xE 4, 0xAE, 0x0A]);

	31 var videoKeyId = "0123456789012345";

	32 var videoKey = new Uint8Array([0x7A, 0x7A, 0x62, 0xF1, 0x68, 0x1 4, 0xD2, 0x7B,

	33 0x68, 0xEF, 0x12, 0x2A, 0xFC, 0xE 4, 0xAE, 0x0A]);

	34

	35 function onNeedKey(event)

	36 {

	37 assert_equals(event.target, video);
	ddorwin 2014/03/21 06:00:20 We're not specifically testing the events, so we p We're not specifically testing the events, so we probably don't need these. WDYT? jrummell 2014/03/21 17:23:01 Sure. I just copied the regular playback test and Show quoted text On 2014/03/21 06:00:20, ddorwin wrote: > We're not specifically testing the events, so we probably don't need these. > WDYT? Sure. I just copied the regular playback test and added support for 2 keys. I didn't do any other cleanup.
	38 assert_true(event instanceof window.MediaKeyNeededEvent);

	39 assert_equals(event.type, "needkey");

	40

	41 // Ideally we could just look at event.contentType, but it
	ddorwin 2014/03/21 06:00:20 This comment is unnecessary - the spec says you ca This comment is unnecessary - the spec says you can't depend on contentType. It seems pretty straightforward that we want to figure out which key ID this is. jrummell 2014/03/21 17:23:01 Done. Show quoted text On 2014/03/21 06:00:20, ddorwin wrote: > This comment is unnecessary - the spec says you can't depend on contentType. It > seems pretty straightforward that we want to figure out which key ID this is. Done.
	42 // is used to specify the encoding format, not the stream

	43 // (and both streams are webm).

	44 var keyId = String.fromCharCode.apply(null, event.initData);

	45 if (keyId == videoKeyId) {

	46 assert_equals(videoMediaKeySession, null);

	47 videoMediaKeySession = mediaKeys.createSession(event.con tentType, event.initData);
	ddorwin 2014/03/21 06:00:20 nit: We could share a bit more code by creating th nit: We could share a bit more code by creating the session above with a local var then assigning it here and waiting for an event after the else. WDYT? jrummell 2014/03/21 17:23:01 Done. Show quoted text On 2014/03/21 06:00:20, ddorwin wrote: > nit: We could share a bit more code by creating the session above with a local > var then assigning it here and waiting for an event after the else. WDYT? Done.
	48 waitForEventAndRunStep("message", videoMediaKeySession, onMessage, test);

	49 } else {

	50 assert_equals(keyId, audioKeyId);

	51 assert_equals(audioMediaKeySession, null);

	52 audioMediaKeySession = mediaKeys.createSession(event.con tentType, event.initData);

	53 waitForEventAndRunStep("message", audioMediaKeySession, onMessage, test);

	54 }

	55 }

	56

	57 function onMessage(event)

	58 {

	59 assert_true(event instanceof window.MediaKeyMessageEvent);

	60 assert_equals(event.type, "message");

	61 var keyId = event.message;

	62

	63 if (event.target == videoMediaKeySession) {

	64 assert_equals(String.fromCharCode.apply(null, keyId), vi deoKeyId);

	65 var jwkSet = stringToUint8Array(createJWKSet(createJWK(k eyId, videoKey)));

	66 videoMediaKeySession.update(jwkSet);

	67 waitForEventAndRunStep("ready", videoMediaKeySession, on Ready, test);

	68 } else {

	69 assert_equals(event.target, audioMediaKeySession);

	70 assert_equals(String.fromCharCode.apply(null, keyId), au dioKeyId);

	71 var jwkSet = stringToUint8Array(createJWKSet(createJWK(k eyId, audioKey)));

	72 audioMediaKeySession.update(jwkSet);

	73 waitForEventAndRunStep("ready", audioMediaKeySession, on Ready, test);

	74 }

	75 }

	76

	77 function onReady(event)

	78 {

	79 assert_true(event instanceof window.Event);

	80 assert_equals(event.type, "ready");

	81 if (event.target == videoMediaKeySession) {

	82 videoSessionReadyReceived = true;

	83 } else {

	84 assert_equals(event.target, audioMediaKeySession);

	85 audioSessionReadyReceived = true;

	86 }

	87 }

	88

	89 function onPlaying(event)
	ddorwin 2014/03/21 06:00:20 OOC, could we just register the timeupdate event d OOC, could we just register the timeupdate event directly? jrummell 2014/03/21 17:23:01 I'm not sure it makes any difference. We still wan Show quoted text On 2014/03/21 06:00:20, ddorwin wrote: > OOC, could we just register the timeupdate event directly? I'm not sure it makes any difference. We still want to see the playing event so we verify that playing doesn't start until after both keys are provided. We won't get timeupdate events until playing starts. Leaving it as is to continue to match what is in the other encrypted-media-playback-* tests. ddorwin 2014/03/21 17:36:49 Do we really not get a playing event until the key Show quoted text On 2014/03/21 17:23:01, jrummell wrote: > On 2014/03/21 06:00:20, ddorwin wrote: > > OOC, could we just register the timeupdate event directly? > > I'm not sure it makes any difference. We still want to see the playing event so > we verify that playing doesn't start until after both keys are provided. We > won't get timeupdate events until playing starts. Leaving it as is to continue > to match what is in the other encrypted-media-playback-* tests. Do we really not get a playing event until the keys are provided? If so, we should verify that by asserting that update() has been called on videoMediaKeySession and audioMediaKeySession. jrummell 2014/03/21 18:03:59 Test that Ready received for both audio and video Show quoted text On 2014/03/21 17:36:49, ddorwin wrote: > On 2014/03/21 17:23:01, jrummell wrote: > > On 2014/03/21 06:00:20, ddorwin wrote: > > > OOC, could we just register the timeupdate event directly? > > > > I'm not sure it makes any difference. We still want to see the playing event > so > > we verify that playing doesn't start until after both keys are provided. We > > won't get timeupdate events until playing starts. Leaving it as is to continue > > to match what is in the other encrypted-media-playback-* tests. > > Do we really not get a playing event until the keys are provided? If so, we > should verify that by asserting that update() has been called on > videoMediaKeySession and audioMediaKeySession. Test that Ready received for both audio and video done in onTimeUpdate(). Converted to asserts here.
	90 {

	91 // Not using waitForEventAndRunStep() to avoid too many EVEN T(onTimeUpdate) logs.

	92 video.addEventListener("timeupdate", onTimeUpdate, true);

	93 }

	94

	95 function onTimeUpdate(event)

	96 {

	97 if (!audioSessionReadyReceived \|\| !videoSessionReadyReceived )

	98 return;

	99 if (event.target.currentTime < 0.2)
	ddorwin 2014/03/21 06:00:20 FYI: I think this only works as a test if the play FYI: I think this only works as a test if the player stops when either audio or video are blocked. I don't think Chrome currently has that behavior. We should have tests that verify the Chrome behavior is fixed, including EME waiting tests. Any thoughts on how to verify both keys are being used in this test? A better test would be a file that uses one key for .1s then another key for the rest of the file. We can address that later, though. jrummell 2014/03/21 17:23:01 If I remove either of the update() calls in onMess Show quoted text On 2014/03/21 06:00:20, ddorwin wrote: > FYI: I think this only works as a test if the player stops when either audio or > video are blocked. I don't think Chrome currently has that behavior. > We should have tests that verify the Chrome behavior is fixed, including EME > waiting tests. > Any thoughts on how to verify both keys are being used in this test? > A better test would be a file that uses one key for .1s then another key for the > rest of the file. We can address that later, though. If I remove either of the update() calls in onMessage(), the playing event is never received and the test times out. ddorwin 2014/03/21 17:36:49 Okay. We should have a test like this when we star Show quoted text On 2014/03/21 17:23:01, jrummell wrote: > On 2014/03/21 06:00:20, ddorwin wrote: > > FYI: I think this only works as a test if the player stops when either audio > or > > video are blocked. I don't think Chrome currently has that behavior. > > We should have tests that verify the Chrome behavior is fixed, including EME > > waiting tests. > > Any thoughts on how to verify both keys are being used in this test? > > A better test would be a file that uses one key for .1s then another key for > the > > rest of the file. We can address that later, though. > > If I remove either of the update() calls in onMessage(), the playing event is > never received and the test times out. Okay. We should have a test like this when we start testing waiting. jrummell 2014/03/21 18:03:59 Done. Show quoted text On 2014/03/21 17:36:49, ddorwin wrote: > On 2014/03/21 17:23:01, jrummell wrote: > > On 2014/03/21 06:00:20, ddorwin wrote: > > > FYI: I think this only works as a test if the player stops when either audio > > or > > > video are blocked. I don't think Chrome currently has that behavior. > > > We should have tests that verify the Chrome behavior is fixed, including EME > > > waiting tests. > > > Any thoughts on how to verify both keys are being used in this test? > > > A better test would be a file that uses one key for .1s then another key for > > the > > > rest of the file. We can address that later, though. > > > > If I remove either of the update() calls in onMessage(), the playing event is > > never received and the test times out. > > Okay. We should have a test like this when we start testing waiting. Done.
	100 return;

	101

	102 test.done();

	103 }

	104

	105 waitForEventAndRunStep("needkey", video, onNeedKey, test);

	106 waitForEventAndRunStep("playing", video, onPlaying, test);

	107

	108 video.src = "../content/test-encrypted-2-keys.webm";
	ddorwin 2014/03/21 06:00:20 nit: We should name it different-av-keys or someth nit: We should name it different-av-keys or something like that that indicates that the a and v have different keys. jrummell 2014/03/21 17:23:01 Done. Show quoted text On 2014/03/21 06:00:20, ddorwin wrote: > nit: We should name it different-av-keys or something like that that indicates > that the a and v have different keys. Done.
	109 video.setMediaKeys(mediaKeys);

	110 video.play();

	111 }, "Playback using Clear Key key system with multiple sessions.",

	112 { timeout: 60000 });

	113 </script>

	114 </body>

	115 </html>

OLD	NEW