Add support for IE/Edge/PlayReady #176

jonoward · 2015-09-09T14:42:25Z

Added new EME polyfill for v20140218 which is the flavour of EME implemented by IE11/Edge. No support for peristent licenses, not 100% sure if this is possible in IE11/Edge. Not tested on live streams, don't have suitable content.

Also had to make some changes to work-around browser specific issues:

Had to increase shaka.player.Player.UNDERFLOW_THRESHOLD_ from 0.2 to 0.5; because IE11 + Edge would automatically halt video before playhead reached the underflow theshold, a state it couldn't exit out of.
Had to tweak shaka.util.FakeEventTarget.prototype.dispatchEvent to only define properties once as this caused an issue in IE11 (Edge OK). Also, IE11 didn't like the properties being defined with value/writable, only got it to work properly with get/set funcs. Issue was that event.target was always null, even if it was being set on the preceding line. I did experiment with using Object.defineProperty for each, instead of Object.defineProperties, as was discussed a little while back, but I had the same issue.
Not sure if this is a browser bug in IE11, but when I handled PSSH data in the msneedkey event (within the new EME polyfill), I was occasionally getting the PSSH data duplicated back-to-back (e.g. if media contained PlayReady PSSH "ABC", sometime event.initData in msneedkey would have "ABC", and sometimes it would have "ABCABC"). Never found a reason why this was happening. This eventually caused an exception in IE11. So I tweaked the PSSH parser to keep track of the boundaries for each distinct PSSH box within a given array, which is used by the EME polyfil to remove duplicates within the msneedkey handler.
EME manager uses shaka.util.Uint8ArrayUtils.key() to generate a unique key for a PSSH data array by converting to a comma separated string, which was later used as a property on an object. There must be a limit to the length of property names in IE11, as large values generated by the key() method (when PSSH data is large) caused the value of object.key to always be undefined. Modified this method to keep string representations of the given array internally, and assign it a randomly generated ID if it's not already present in the cache.

Tests:
Unit tests are passing, but am having some flakiness in the integration tests, in both this branch as well as latest master:

PR branch failing tests:
Player end-of-stream behavior permits looping

Master failing tests:
Player setPlaybackRate plays in reverse for negative rates
Player end-of-stream behavior permits looping

Any idea if there are known issues with the integration tests?

joeyparrish · 2015-09-09T16:03:55Z

We are having issues with test flakiness in general. It has gotten better and continues to improve, but we still have some issues with the integration tests. I'll run this PR on our bot and see how it goes.

Thank you so much for putting this PR together. I'll review it as soon as possible.

shaka-bot · 2015-09-09T16:04:23Z

Testing in progress...

joeyparrish · 2015-09-09T16:06:55Z

externs/v20140218_eme.js

@@ -0,0 +1,108 @@
+/**
+ * Copyright 2014 Google Inc.


2015
Also, add a license annotation just before the copyright header, as in 4cc4e96?

shaka-bot · 2015-09-09T16:08:16Z

All tests passed!

joeyparrish · 2015-09-09T16:08:41Z

externs/v20140218_eme.js

@@ -0,0 +1,108 @@
+/**


To keep with the existing convention (externs/mediakeys.js), can you rename this to externs/msmediakeys.js?

joeyparrish · 2015-09-09T17:22:39Z

lib/polyfill/patchedmediakeys_20140218.js

+  // This is only a guess, since we don't really know from the prefixed API.
+  var allowPersistentState = true;
+
+  if (keySystem == 'org.w3.clearkey') {


From what I can tell, IE11 doesn't seem to support clearkey.

Actually you're right, it doesn't, I'll remove this (and any other refs in the polyfill). I can't find definitive documentation about Edge, so I'll presume it doesn't support it either, as most of the EME implementation so far seems to be pretty much identical between them

joeyparrish · 2015-10-14T18:34:14Z

lib/polyfill/patchedmediakeys_20140218.js

+  delete this['mediaKeys'];  // in case there is an existing getter
+  this['mediaKeys'] = mediaKeys;  // work around read-only declaration
+
+  return mediaKeys.setMedia(this);


Since you aren't calling setMedia(null) on the old value of mediaKeys, there seems to be no way for this polyfill to remove listeners created by setMedia. This could lead to a memory leak by holding references to the media element.

joeyparrish · 2015-10-14T18:48:47Z

I think we're nearly there.

The last version I tested (not current) had issues with non-zero start times, which also breaks all live content. The GPAC clip in the demo app uses a non-zero start time, so you can reproduce with that. The issue seems to be similar to one we had in #101 getting Firefox to work. We'll have to change how/when we set currentTime during startup, but that can wait for another patch.

I'll also need to add some PR DRM settings to the demo app. There's one clip ("multi-DRM") that should work but doesn't yet.

jonoward · 2015-10-15T17:21:47Z

Thanks Joey, I'll aim to incorporate these changes in the next few days.

joeyparrish · 2015-10-15T20:12:14Z

Thanks, Jono. Ping me when the next version is ready.

Inspired by code in pull #176 Change-Id: I2e29310c8a3583ed208d7bd1ae2e747a92ddf480

This will help with porting to IE11. Chrome, Firefox, Safari, and Edge all have native Promises. This polyfill does not support thenables because Shaka does not use them. Other than tests related to thenables, this polyfill passes the A+ test suite. It is also worth noting that this polyfill is incompatible with native Promises, so it should not be used to replace a native implementation or mixed with browser APIs that may use a native implementation internally. To safely test in Chrome, force prefixed EME (to avoid native Promises), set window.Promise to null, then load some content in the test app. If using a verison of Chrome after prefixed EME was dropped, use unencrypted content. To run the A+ test suite, compile the library, install nodejs and the module 'promises-aplus-tests', then run ./test_promise_polyfill.js. Inspired by pull #176 Change-Id: I0d25049f162ff7f3b57bbc795403fcdedf927262

joeyparrish · 2015-10-22T17:03:53Z

We've written our own Promise polyfill to more fully support IE11 after your pull request is merged. Hope that helps.

joeyparrish · 2015-11-01T05:03:45Z

@jonoward @jonoblinkbox

Any idea when you'll be able to make revisions? I'm hoping to wrap up v1.6.0 soon.

jonoward · 2015-11-02T08:29:16Z

@joeyparrish Apologies for the delay, I'm working on this now, so I should have something within a 1-2 days

…and annotation improvements

…nt handlers, removing some redundant code and some code-style tweaks

jonoward · 2015-11-02T13:04:47Z

@joeyparrish Right, that should hopefully be all the discussed changes pushed up. I've rebased on latest master and have removed our usage of the shaka.util.Uint8ArrayUtils.key from the polyfill.

Regarding non-zero start times, yes this still seems to be an issue. In our fork, we made some additional changes to pass playbackStartTime down to each individual stream, to avoid relying on currentTime being set. It's a little messy though, have a look at this branch if you're interested - https://github.com/blinkbox/shaka-player/tree/release/1.4.0. Quite like the idea of waiting for the buffer to contain data before setting currentTime (that you mentioned in #101), rather than waiting on loadedmetadata.

joeyparrish · 2015-11-02T20:53:33Z

@jonoward Thanks for the revisions. Looks good! I'll run this PR through our build bot and then we can merge.

shaka-bot · 2015-11-02T20:54:26Z

Testing in progress...

shaka-bot · 2015-11-02T20:58:53Z

All tests passed!

…support Add support for IE/Edge/PlayReady

joeyparrish · 2015-11-02T21:07:50Z

Woohoo! IE support landed! 1,000,000 thanks to @jonoward!

jonoward · 2015-11-03T08:59:52Z

Awesome, thanks

joeyparrish reviewed Sep 9, 2015
View reviewed changes

joeyparrish self-assigned this Sep 9, 2015

joeyparrish added the type: enhancement New feature or request label Sep 9, 2015

joeyparrish reviewed Sep 9, 2015
View reviewed changes

joeyparrish reviewed Oct 14, 2015
View reviewed changes

joeyparrish added a commit that referenced this pull request Oct 20, 2015

Refine MediaKeys polyfill error type

ff1dcc2

Inspired by code in pull #176 Change-Id: I2e29310c8a3583ed208d7bd1ae2e747a92ddf480

jonoblinkbox and others added 3 commits November 2, 2015 12:26

Add support for IE/Edge/PlayReady

de36492

Implementing review comments; code cleanup, mainly code style tweaks …

c9892f9

…and annotation improvements

Implement suggested changes from PR review; improving clean-up of eve…

6db78ba

…nt handlers, removing some redundant code and some code-style tweaks

jonoward force-pushed the pull-request/ie-edge-playready-support branch from 5325658 to 6db78ba Compare November 2, 2015 12:37

joeyparrish added a commit that referenced this pull request Nov 2, 2015

Merge pull request #176 from blinkbox/pull-request/ie-edge-playready-…

80f48ea

…support Add support for IE/Edge/PlayReady

joeyparrish merged commit 80f48ea into shaka-project:master Nov 2, 2015

joeyparrish mentioned this pull request Jan 13, 2020

Ending support for IE 11 after v3.0 #2339

Closed

github-actions bot added the status: archived Archived and locked; will not be updated label Jul 25, 2023

github-actions bot locked as resolved and limited conversation to collaborators Jul 25, 2023

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Add support for IE/Edge/PlayReady #176

Add support for IE/Edge/PlayReady #176

jonoward commented Sep 9, 2015

joeyparrish commented Sep 9, 2015

shaka-bot commented Sep 9, 2015

joeyparrish Sep 9, 2015

shaka-bot commented Sep 9, 2015

joeyparrish Sep 9, 2015

joeyparrish Sep 9, 2015

jonoward Sep 10, 2015

joeyparrish Oct 14, 2015

joeyparrish commented Oct 14, 2015

jonoward commented Oct 15, 2015

joeyparrish commented Oct 15, 2015

joeyparrish commented Oct 22, 2015

joeyparrish commented Nov 1, 2015

jonoward commented Nov 2, 2015

jonoward commented Nov 2, 2015

joeyparrish commented Nov 2, 2015

shaka-bot commented Nov 2, 2015

shaka-bot commented Nov 2, 2015

joeyparrish commented Nov 2, 2015

jonoward commented Nov 3, 2015

Add support for IE/Edge/PlayReady #176

Add support for IE/Edge/PlayReady #176

Conversation

jonoward commented Sep 9, 2015

joeyparrish commented Sep 9, 2015

shaka-bot commented Sep 9, 2015

joeyparrish Sep 9, 2015

Choose a reason for hiding this comment

shaka-bot commented Sep 9, 2015

joeyparrish Sep 9, 2015

Choose a reason for hiding this comment

joeyparrish Sep 9, 2015

Choose a reason for hiding this comment

jonoward Sep 10, 2015

Choose a reason for hiding this comment

joeyparrish Oct 14, 2015

Choose a reason for hiding this comment

joeyparrish commented Oct 14, 2015

jonoward commented Oct 15, 2015

joeyparrish commented Oct 15, 2015

joeyparrish commented Oct 22, 2015

joeyparrish commented Nov 1, 2015

jonoward commented Nov 2, 2015

jonoward commented Nov 2, 2015

joeyparrish commented Nov 2, 2015

shaka-bot commented Nov 2, 2015

shaka-bot commented Nov 2, 2015

joeyparrish commented Nov 2, 2015

jonoward commented Nov 3, 2015