searching for reviewer(cpearce)
41ab55fee001f7fa381a77d9afc8395786417e1a: Bug 1498742 - Part 2 - Start the GMP sandbox earlier during process startup r=jya,cpearce
Haik Aftandilian <haftandilian@mozilla.com> - Fri, 14 Jun 2019 22:55:27 +0000 - rev 538622
Push 11522 by ffxbld-merge at Mon, 01 Jul 2019 09:00:55 +0000
Bug 1498742 - Part 2 - Start the GMP sandbox earlier during process startup r=jya,cpearce Change the Mac GMP process launch to include sandboxing params on the command line to allow the sandbox to be started earlier during GMP process launch. Content, extension, and RDD processes have already been changed to start the sandbox earlier. Update GMPProcessParent to override GeckoChildProcessHost methods used to construct sandboxing parameters. Pass the plugin path as a sandbox parameter so that the sandbox rules can whitelist the plugin directory which is now read after the sandbox is enabled in the plugin process. On development builds, pass "testingReadPath" params so directories needed during automated tests can be whitelisted. Update Mac sandboxing code to detect GMP sandbox params on the command line and enable the sandbox with additional arguments needed for early sandbox start. Allow reverting to the old implementation by setting security.sandbox.gmp.mac.earlyinit to false. Differential Revision: https://phabricator.services.mozilla.com/D34085
f6da94d9035047b44f1521bfd2d37960263a2f36: Bug 1498742 - Part 2 - Start the GMP sandbox earlier during process startup r=jya,cpearce
Haik Aftandilian <haftandilian@mozilla.com> - Fri, 14 Jun 2019 20:12:01 +0000 - rev 538571
Push 11522 by ffxbld-merge at Mon, 01 Jul 2019 09:00:55 +0000
Bug 1498742 - Part 2 - Start the GMP sandbox earlier during process startup r=jya,cpearce Change the Mac GMP process launch to include sandboxing params on the command line to allow the sandbox to be started earlier during GMP process launch. Content, extension, and RDD processes have already been changed to start the sandbox earlier. Update GMPProcessParent to override GeckoChildProcessHost methods used to construct sandboxing parameters. Pass the plugin path as a sandbox parameter so that the sandbox rules can whitelist the plugin directory which is now read after the sandbox is enabled in the plugin process. On development builds, pass "testingReadPath" params so directories needed during automated tests can be whitelisted. Update Mac sandboxing code to detect GMP sandbox params on the command line and enable the sandbox with additional arguments needed for early sandbox start. Allow reverting to the old implementation by setting security.sandbox.gmp.mac.earlyinit to false. Differential Revision: https://phabricator.services.mozilla.com/D34085
6729dc168afd4186db7cfb4a3bb5810d8fc6a826: Bug 1498742 - Part 2 - Start the GMP sandbox earlier during process startup r=jya,cpearce
Haik Aftandilian <haftandilian@mozilla.com> - Thu, 13 Jun 2019 06:32:37 +0000 - rev 538276
Push 11522 by ffxbld-merge at Mon, 01 Jul 2019 09:00:55 +0000
Bug 1498742 - Part 2 - Start the GMP sandbox earlier during process startup r=jya,cpearce Change the Mac GMP process launch to include sandboxing params on the command line to allow the sandbox to be started earlier during GMP process launch. Content, extension, and RDD processes have already been changed to start the sandbox earlier. Update GMPProcessParent to override GeckoChildProcessHost methods used to construct sandboxing parameters. Pass the plugin path as a sandbox parameter so that the sandbox rules can whitelist the plugin directory which is now read after the sandbox is enabled in the plugin process. On development builds, pass "testingReadPath" params so directories needed during automated tests can be whitelisted. Update Mac sandboxing code to detect GMP sandbox params on the command line and enable the sandbox with additional arguments needed for early sandbox start. Allow reverting to the old implementation by setting security.sandbox.gmp.mac.earlyinit to false. Differential Revision: https://phabricator.services.mozilla.com/D34085
5f9d32e26c712ca9494a2bcf3b77778a2e4e8119: Bug 1498742 - Part 2 - Start the GMP sandbox earlier during process startup r=jya,cpearce
Haik Aftandilian <haftandilian@mozilla.com> - Wed, 12 Jun 2019 23:41:53 +0000 - rev 538228
Push 11522 by ffxbld-merge at Mon, 01 Jul 2019 09:00:55 +0000
Bug 1498742 - Part 2 - Start the GMP sandbox earlier during process startup r=jya,cpearce Change the Mac GMP process launch to include sandboxing params on the command line to allow the sandbox to be started earlier during GMP process launch. Content, extension, and RDD processes have already been changed to start the sandbox earlier. Update GMPProcessParent to override GeckoChildProcessHost methods used to construct sandboxing parameters. Pass the plugin path as a sandbox parameter so that the sandbox rules can whitelist the plugin directory which is now read after the sandbox is enabled in the plugin process. On development builds, pass "testingReadPath" params so directories needed during automated tests can be whitelisted. Update Mac sandboxing code to detect GMP sandbox params on the command line and enable the sandbox with additional arguments needed for early sandbox start. Allow reverting to the old implementation by setting security.sandbox.gmp.mac.earlyinit to false. Differential Revision: https://phabricator.services.mozilla.com/D34085
7cf9846578e49787c666d7cb436f45692c0f325b: Bug 1536102 - P4: Enabled WPT for encrypted media scheme checking. r=cpearce
Bryce Van Dyk <bvandyk@mozilla.com> - Wed, 12 Jun 2019 05:31:52 +0000 - rev 538028
Push 11522 by ffxbld-merge at Mon, 01 Jul 2019 09:00:55 +0000
Bug 1536102 - P4: Enabled WPT for encrypted media scheme checking. r=cpearce Enable the web platform tests for clearkey scheme checking. Since these tests simply check if the functionality is implemented, and do not check if the browsers actually support different encryption schemes, it's okay to do this even though we don't have cbcs support in clearkey yet. I.e. it's enough that a page can ask "do you support cbcs in clearkey?" to Firefox to pass the test, the answer from Firefox doesn't have to be "yes." Add the pref setting to the DRM scheme checking test, though leave the expectations as they are on this test, as in automation the test will still not pass due to Widevine downloads being blocked. My hope is that we can find a solution to this Widevine download issue in automation, at which point we'd expect that tests to start passing due to the pref added in this patch -- at which point we could toggle the expectations. Differential Revision: https://phabricator.services.mozilla.com/D34303
70b2674c95c5d6de3f83f9e44700535b480337f2: Bug 1536102 - P3: Implement filtering behaviour for encryption scheme if passed on MediaKeySystemMediaCapability. r=cpearce
Bryce Van Dyk <bvandyk@mozilla.com> - Wed, 12 Jun 2019 05:31:38 +0000 - rev 538027
Push 11522 by ffxbld-merge at Mon, 01 Jul 2019 09:00:55 +0000
Bug 1536102 - P3: Implement filtering behaviour for encryption scheme if passed on MediaKeySystemMediaCapability. r=cpearce Drive by reordering of includes per google style and remove several unused headers. Differential Revision: https://phabricator.services.mozilla.com/D34302
be993cc8f451d7c848e26a7136d5d3f5359209da: Bug 1536102 - P1: Add pref to gate encryption scheme check functionality behind. r=cpearce
Bryce Van Dyk <bvandyk@mozilla.com> - Wed, 12 Jun 2019 05:31:00 +0000 - rev 538025
Push 11522 by ffxbld-merge at Mon, 01 Jul 2019 09:00:55 +0000
Bug 1536102 - P1: Add pref to gate encryption scheme check functionality behind. r=cpearce First step in implementing behaviour proposed in https://github.com/WICG/encrypted-media-encryption-scheme/blob/master/explainer.md Differential Revision: https://phabricator.services.mozilla.com/D34300
267d1abce0581f08bdefc977f24e421c161e603f: Bug 1552717 - P3: Set mInitDataType on sample info regardless of if the sample is encrypted. r=cpearce
Bryce Van Dyk <bvandyk@mozilla.com> - Tue, 28 May 2019 12:40:16 +0000 - rev 535467
Push 11522 by ffxbld-merge at Mon, 01 Jul 2019 09:00:55 +0000
Bug 1552717 - P3: Set mInitDataType on sample info regardless of if the sample is encrypted. r=cpearce Unencrypted samples can carry init data and thus we should set the init data type regardless of if the sample itself is encrypted (otherwise the init data info is incomplete for unencrypted samples). Depends on D32751 Differential Revision: https://phabricator.services.mozilla.com/D32752
71edbdac5e129ab4c49b0119d2e43962e1b24c34: Bug 1552717 - P2: Add test media for key rotation with a clear lead. r=cpearce
Bryce Van Dyk <bvandyk@mozilla.com> - Tue, 28 May 2019 08:13:13 +0000 - rev 535466
Push 11522 by ffxbld-merge at Mon, 01 Jul 2019 09:00:55 +0000
Bug 1552717 - P2: Add test media for key rotation with a clear lead. r=cpearce Add another test case for the mp4 key rotation (pssh in fragments) test. In this case, the test file has a clear (unencrypted) lead. This test seeks to ensure we don't regress surfacing of init info even if we encounter it for fragments that start with unencrypted samples. Add a further check to the key rotation test to ensure that the initDataType is being surfaced correctly on the encrypted event. Media files created with shaka packager via: ``` packager-win.exe in=bipbop.mp4,stream=audio,out=bipbop-clearkey-keyrotation-clear-lead-audio.mp4 in=bipbop.mp4,stream=video,out=bipbop-clearkey-keyrotation-clear-lead-video.mp4 --enable_raw_key_encryption --keys label=:key_id=00112233445566778899AABBCCDDEEFF:key=00112233445566778899AABBCCDDEEFF --crypto_period_duration 5 --fragment_duration 5 --clear_lead 3 ``` Note that the way shaka packager handles key rotation in this case is just to left shift the key id and the key. In this case, where crypto_period_duration == fragment_duration, a left shift of 1 will take place each time the keys rotate. This happens once in the test file leaving us with 2 key ids + keys. Depends on D32750 Differential Revision: https://phabricator.services.mozilla.com/D32751
aeb0cd86768f22552e0df6b8e64529f81842be17: Bug 1552717 - P1: Refactor test_eme_pssh_in_moof mochitest so it's easier to add further test media. r=cpearce
Bryce Van Dyk <bvandyk@mozilla.com> - Tue, 28 May 2019 08:11:38 +0000 - rev 535465
Push 11522 by ffxbld-merge at Mon, 01 Jul 2019 09:00:55 +0000
Bug 1552717 - P1: Refactor test_eme_pssh_in_moof mochitest so it's easier to add further test media. r=cpearce test_eme_pssh_in_moof tests if our key rotation works correctly. It currently does so by using a single video with an audio and video track. This patch refactors the test so that it does the same thing (all going well), but in a more extensible way. The changes in this patch seek to lean more heavily on test harness functionality in manifest.js and eme.js where possible. This cuts down on some boilerplate, but means we have to make some concessions in a more verbose expression of our test media list so the eme.js functions work with it. Differential Revision: https://phabricator.services.mozilla.com/D32750
40dda46eb12632aafb5648430772015620656a73: Bug 1550422 - P27. Do not set WebRender preferences as code don't expect them to exists. r=cpearce
Jean-Yves Avenard <jyavenard@mozilla.com> - Sun, 26 May 2019 14:32:42 +0000 - rev 535275
Push 11522 by ffxbld-merge at Mon, 01 Jul 2019 09:00:55 +0000
Bug 1550422 - P27. Do not set WebRender preferences as code don't expect them to exists. r=cpearce Listing the preferences in either all.js or StaticPrefList.h would also make them appear in about:config which is something we don't want. Additionally, rename some pref constants to improve code clarity as we can no longer rely on using the StaticPrefs accessor. Differential Revision: https://phabricator.services.mozilla.com/D32416
0b40296717100fa26c8f7bdea8e365eeb523cfbd: Bug 1550422 - P27. Do not set WebRender preferences as code don't expect them to exists. r=cpearce
Jean-Yves Avenard <jyavenard@mozilla.com> - Fri, 24 May 2019 11:40:04 +0000 - rev 535185
Push 11522 by ffxbld-merge at Mon, 01 Jul 2019 09:00:55 +0000
Bug 1550422 - P27. Do not set WebRender preferences as code don't expect them to exists. r=cpearce Listing the preferences in either all.js or StaticPrefList.h would also make them appear in about:config which is something we don't want. Additionally, rename some pref constants to improve code clarity as we can no longer rely on using the StaticPrefs accessor. Differential Revision: https://phabricator.services.mozilla.com/D32416
c4781d18e0f5d7f276c91a809db4cfc831c7a24d: Bug 1547757 - Use a RefPtr to hold owner element ref when moving media resource load to background. r=cpearce
Bryce Van Dyk <bvandyk@mozilla.com> - Thu, 16 May 2019 07:01:02 +0000 - rev 534295
Push 11522 by ffxbld-merge at Mon, 01 Jul 2019 09:00:55 +0000
Bug 1547757 - Use a RefPtr to hold owner element ref when moving media resource load to background. r=cpearce Differential Revision: https://phabricator.services.mozilla.com/D31019
9f479818871fec638fc21e91ad10a946e4769bfd: Bug 1552717 - P3: Set mInitDataType on sample info regardless of if the sample is encrypted. r=cpearce a=jcristau
Bryce Van Dyk <bvandyk@mozilla.com> - Tue, 28 May 2019 12:40:16 +0000 - rev 533721
Push 11409 by rmaries@mozilla.com at Mon, 10 Jun 2019 14:23:46 +0000
Bug 1552717 - P3: Set mInitDataType on sample info regardless of if the sample is encrypted. r=cpearce a=jcristau Unencrypted samples can carry init data and thus we should set the init data type regardless of if the sample itself is encrypted (otherwise the init data info is incomplete for unencrypted samples). Depends on D32751 Differential Revision: https://phabricator.services.mozilla.com/D32752
4bf11a9d5633083a474caba9e0c5353f3a078882: Bug 1552717 - P2: Add test media for key rotation with a clear lead. r=cpearce a=jcristau
Bryce Van Dyk <bvandyk@mozilla.com> - Tue, 28 May 2019 08:13:13 +0000 - rev 533720
Push 11409 by rmaries@mozilla.com at Mon, 10 Jun 2019 14:23:46 +0000
Bug 1552717 - P2: Add test media for key rotation with a clear lead. r=cpearce a=jcristau Add another test case for the mp4 key rotation (pssh in fragments) test. In this case, the test file has a clear (unencrypted) lead. This test seeks to ensure we don't regress surfacing of init info even if we encounter it for fragments that start with unencrypted samples. Add a further check to the key rotation test to ensure that the initDataType is being surfaced correctly on the encrypted event. Media files created with shaka packager via: ``` packager-win.exe in=bipbop.mp4,stream=audio,out=bipbop-clearkey-keyrotation-clear-lead-audio.mp4 in=bipbop.mp4,stream=video,out=bipbop-clearkey-keyrotation-clear-lead-video.mp4 --enable_raw_key_encryption --keys label=:key_id=00112233445566778899AABBCCDDEEFF:key=00112233445566778899AABBCCDDEEFF --crypto_period_duration 5 --fragment_duration 5 --clear_lead 3 ``` Note that the way shaka packager handles key rotation in this case is just to left shift the key id and the key. In this case, where crypto_period_duration == fragment_duration, a left shift of 1 will take place each time the keys rotate. This happens once in the test file leaving us with 2 key ids + keys. Depends on D32750 Differential Revision: https://phabricator.services.mozilla.com/D32751
a6f7233484a044e74ebb6ddef80ea03957d689f4: Bug 1552717 - P1: Refactor test_eme_pssh_in_moof mochitest so it's easier to add further test media. r=cpearce a=jcristau
Bryce Van Dyk <bvandyk@mozilla.com> - Tue, 28 May 2019 08:11:38 +0000 - rev 533719
Push 11409 by rmaries@mozilla.com at Mon, 10 Jun 2019 14:23:46 +0000
Bug 1552717 - P1: Refactor test_eme_pssh_in_moof mochitest so it's easier to add further test media. r=cpearce a=jcristau test_eme_pssh_in_moof tests if our key rotation works correctly. It currently does so by using a single video with an audio and video track. This patch refactors the test so that it does the same thing (all going well), but in a more extensible way. The changes in this patch seek to lean more heavily on test harness functionality in manifest.js and eme.js where possible. This cuts down on some boilerplate, but means we have to make some concessions in a more verbose expression of our test media list so the eme.js functions work with it. Differential Revision: https://phabricator.services.mozilla.com/D32750
326fdc435656a1f34b13591b038017eed68814fb: Bug 1547757 - Use a RefPtr to hold owner element ref when moving media resource load to background. r=cpearce a=jcristau
Bryce Van Dyk <bvandyk@mozilla.com> - Thu, 16 May 2019 07:01:02 +0000 - rev 533319
Push 11293 by archaeopteryx@coole-files.de at Wed, 22 May 2019 14:51:55 +0000
Bug 1547757 - Use a RefPtr to hold owner element ref when moving media resource load to background. r=cpearce a=jcristau Differential Revision: https://phabricator.services.mozilla.com/D31019
4b635d928b2bd42890ffc97c55622072092351e9: Bug 1549544 - Use the 'cenc' init data type when encountering cbcs media. r=cpearce
Bryce Van Dyk <bvandyk@mozilla.com> - Tue, 14 May 2019 02:31:47 +0000 - rev 532860
Push 11272 by apavel@mozilla.com at Thu, 16 May 2019 15:28:22 +0000
Bug 1549544 - Use the 'cenc' init data type when encountering cbcs media. r=cpearce When encountering media using the common encryption format (mp4s containing pssh boxes), the encrypted event we fire should specify the cenc init data type. Fixes a bug where for cbcs the 'cbcs' init data type was used -- which is a bogus value. Differential Revision: https://phabricator.services.mozilla.com/D30095
a14a4e0b60ebc2ca635bbc777474cf2fa8cbc62d: Bug 1544602 - Make media.gmp.insecure.allow an atomic. r=cpearce
Bryce Van Dyk <bvandyk@mozilla.com> - Tue, 16 Apr 2019 22:20:10 +0000 - rev 528699
Push 11265 by ffxbld-merge at Mon, 13 May 2019 10:53:39 +0000
Bug 1544602 - Make media.gmp.insecure.allow an atomic. r=cpearce Since this pref can be accessed from the GMP thread (per the bug), it should be an atomic. See bug 1465852 for the code the added the assertion hit in this bug, as well as further context as to why this should be an atomic. Differential Revision: https://phabricator.services.mozilla.com/D27831
46235152b24e33c652c4e94430897369bbb2090b: Bug 1525209 - Revert spec-incompliant pause event from bug 1461877. r=cpearce
Andreas Pehrson <apehrson@mozilla.com> - Wed, 17 Apr 2019 09:57:59 +0000 - rev 528670
Push 11265 by ffxbld-merge at Mon, 13 May 2019 10:53:39 +0000
Bug 1525209 - Revert spec-incompliant pause event from bug 1461877. r=cpearce Differential Revision: https://phabricator.services.mozilla.com/D27864
a38b2423065cdf70dd8db23a468c23de1d3b88dd: Bug 1538206 - Update comments in MediaKeySystemAccess to reflect current numbering in spec. r=cpearce
Bryce Van Dyk <bvandyk@mozilla.com> - Tue, 16 Apr 2019 19:38:46 +0000 - rev 528607
Push 11265 by ffxbld-merge at Mon, 13 May 2019 10:53:39 +0000
Bug 1538206 - Update comments in MediaKeySystemAccess to reflect current numbering in spec. r=cpearce The spec referenced in the comments in code has been updated since the code was originally written. As such, some of the sections referenced in code comments now have different numbering. Update the comments to reflect this. Differential Revision: https://phabricator.services.mozilla.com/D27726
8e1aac075975dd109f48d5ad766c713c6f304af5: Bug 1542921 - turn on block autoplay by default. r=cpearce
alwu <alwu@mozilla.com> - Tue, 09 Apr 2019 23:43:33 +0000 - rev 527499
Push 11265 by ffxbld-merge at Mon, 13 May 2019 10:53:39 +0000
Bug 1542921 - turn on block autoplay by default. r=cpearce As we have successfully rolled out the feature via Normandy in 66, we should turn the feature on in >= 67 to not have to rely on Normandy going forward any more. Differential Revision: https://phabricator.services.mozilla.com/D26713
e217cf252bb06d2c4f4e61b49ccd1e77614a54df: Bug 1540136 - P5: Handle if ChromiumCDMProxy is shutdown in the middle of init. r=cpearce
Bryce Van Dyk <bvandyk@mozilla.com> - Tue, 09 Apr 2019 15:07:49 +0000 - rev 527398
Push 11265 by ffxbld-merge at Mon, 13 May 2019 10:53:39 +0000
Bug 1540136 - P5: Handle if ChromiumCDMProxy is shutdown in the middle of init. r=cpearce - Watch for if a proxy shuts down during init and if so, shutdown the CDM parent that is being initialized. - Make ChromiumCDMParent only store a pointer to a ChromiumCDMProxy when it has successfully initialized. This avoid the lopsided relationship where a if a ChromiumCDMParent fails to initialize it may keep a pointer to a proxy, but the proxy will never have a reference to that CDM parent. Differential Revision: https://phabricator.services.mozilla.com/D26208
f09291096c4f9589a14f43468a04970202ae2468: Bug 1540136 - P4: Remove ChromiumCDMParent mMainThread as it is unused. r=cpearce
Bryce Van Dyk <bvandyk@mozilla.com> - Tue, 09 Apr 2019 15:07:19 +0000 - rev 527397
Push 11265 by ffxbld-merge at Mon, 13 May 2019 10:53:39 +0000
Bug 1540136 - P4: Remove ChromiumCDMParent mMainThread as it is unused. r=cpearce Differential Revision: https://phabricator.services.mozilla.com/D26207
7178347182f50db3fce7af4910e97274d11cb7fa: Bug 1540136 - P3: ChromiumCDMProxy runs CDM init thennable code on the main thread. r=cpearce
Bryce Van Dyk <bvandyk@mozilla.com> - Tue, 09 Apr 2019 15:06:59 +0000 - rev 527396
Push 11265 by ffxbld-merge at Mon, 13 May 2019 10:53:39 +0000
Bug 1540136 - P3: ChromiumCDMProxy runs CDM init thennable code on the main thread. r=cpearce This code is calling other code that expects to be on the main thread, and having this on the main thread (now that the main thread is a serial event target) makes it easier to reason about this and other main thread code. I.e. this cannot be running during other main thread code. Differential Revision: https://phabricator.services.mozilla.com/D26206
bc7e92510e95a39fef9f9fc32224c416e9d6f11f: Bug 1540136 - P2: CDMProxy stores main thread as nsISerialEventTarget, rather than nsIEventTarget. r=cpearce
Bryce Van Dyk <bvandyk@mozilla.com> - Tue, 09 Apr 2019 15:06:38 +0000 - rev 527395
Push 11265 by ffxbld-merge at Mon, 13 May 2019 10:53:39 +0000
Bug 1540136 - P2: CDMProxy stores main thread as nsISerialEventTarget, rather than nsIEventTarget. r=cpearce This gives us greater flexibility in using the main thread member to run promises. The site where we obtain the main thread returns a serial event target, so we're not doing much more work here, we're just keeping the serial event target interface, rather than converting to an event target interface. Differential Revision: https://phabricator.services.mozilla.com/D26205
7c83e08566d34fe194069a5219a16755d1cbe757: Bug 1540136 - P1: Log more info in ChromiumCDMProxy dtor and Shutdown. r=cpearce
Bryce Van Dyk <bvandyk@mozilla.com> - Tue, 09 Apr 2019 15:05:01 +0000 - rev 527394
Push 11265 by ffxbld-merge at Mon, 13 May 2019 10:53:39 +0000
Bug 1540136 - P1: Log more info in ChromiumCDMProxy dtor and Shutdown. r=cpearce Also remove unneeded MOZ_COUNT_[CTOR|DTOR] macros. We already get similar functionality from NS_INLINE_DECL_THREADSAFE_REFCOUNTING. Differential Revision: https://phabricator.services.mozilla.com/D26204
72c34a51659693fb9106e9df3488929d9a8a2117: Bug 1535010 - avoid unnecessary gmp directory creations during startup, r=cpearce.
Florian Quèze <florian@queze.net> - Thu, 28 Mar 2019 00:06:29 +0000 - rev 525367
Push 11265 by ffxbld-merge at Mon, 13 May 2019 10:53:39 +0000
Bug 1535010 - avoid unnecessary gmp directory creations during startup, r=cpearce. Differential Revision: https://phabricator.services.mozilla.com/D23320
1694fdeb8a6ad706eb77001076bb6f0ba95a577a: Bug 1525209 - Revert spec-incompliant pause event from bug 1461877. r=cpearce a=pascalc
Andreas Pehrson <apehrson@mozilla.com> - Wed, 17 Apr 2019 09:57:59 +0000 - rev 523246
Push 11115 by aiakab@mozilla.com at Fri, 19 Apr 2019 16:14:22 +0000
Bug 1525209 - Revert spec-incompliant pause event from bug 1461877. r=cpearce a=pascalc Differential Revision: https://phabricator.services.mozilla.com/D27864
5710827ee9cf5011b6698590a7d31e922f636cf2: Bug 1540136 - P5: Handle if ChromiumCDMProxy is shutdown in the middle of init. r=cpearce a=lizzard
Bryce Van Dyk <bvandyk@mozilla.com> - Tue, 09 Apr 2019 15:07:49 +0000 - rev 523226
Push 11107 by archaeopteryx@coole-files.de at Thu, 18 Apr 2019 15:56:49 +0000
Bug 1540136 - P5: Handle if ChromiumCDMProxy is shutdown in the middle of init. r=cpearce a=lizzard - Watch for if a proxy shuts down during init and if so, shutdown the CDM parent that is being initialized. - Make ChromiumCDMParent only store a pointer to a ChromiumCDMProxy when it has successfully initialized. This avoid the lopsided relationship where a if a ChromiumCDMParent fails to initialize it may keep a pointer to a proxy, but the proxy will never have a reference to that CDM parent. Differential Revision: https://phabricator.services.mozilla.com/D26208
ecb3ee5969c19fe8c52cedb2bd0291edf11dc83c: Bug 1540136 - P4: Remove ChromiumCDMParent mMainThread as it is unused. r=cpearce a=lizzard
Bryce Van Dyk <bvandyk@mozilla.com> - Tue, 09 Apr 2019 15:07:19 +0000 - rev 523225
Push 11107 by archaeopteryx@coole-files.de at Thu, 18 Apr 2019 15:56:49 +0000
Bug 1540136 - P4: Remove ChromiumCDMParent mMainThread as it is unused. r=cpearce a=lizzard Differential Revision: https://phabricator.services.mozilla.com/D26207
385fde9e64e20795ca20d6cb7db114b0c0bcc8fe: Bug 1540136 - P3: ChromiumCDMProxy runs CDM init thennable code on the main thread. r=cpearce a=lizzard
Bryce Van Dyk <bvandyk@mozilla.com> - Tue, 09 Apr 2019 15:06:59 +0000 - rev 523224
Push 11107 by archaeopteryx@coole-files.de at Thu, 18 Apr 2019 15:56:49 +0000
Bug 1540136 - P3: ChromiumCDMProxy runs CDM init thennable code on the main thread. r=cpearce a=lizzard This code is calling other code that expects to be on the main thread, and having this on the main thread (now that the main thread is a serial event target) makes it easier to reason about this and other main thread code. I.e. this cannot be running during other main thread code. Differential Revision: https://phabricator.services.mozilla.com/D26206
48e7d2342ffa2f1bc57853dcd0c3a8e4ca838c42: Bug 1540136 - P2: CDMProxy stores main thread as nsISerialEventTarget, rather than nsIEventTarget. r=cpearce a=lizzard
Bryce Van Dyk <bvandyk@mozilla.com> - Tue, 09 Apr 2019 15:06:38 +0000 - rev 523223
Push 11107 by archaeopteryx@coole-files.de at Thu, 18 Apr 2019 15:56:49 +0000
Bug 1540136 - P2: CDMProxy stores main thread as nsISerialEventTarget, rather than nsIEventTarget. r=cpearce a=lizzard This gives us greater flexibility in using the main thread member to run promises. The site where we obtain the main thread returns a serial event target, so we're not doing much more work here, we're just keeping the serial event target interface, rather than converting to an event target interface. Differential Revision: https://phabricator.services.mozilla.com/D26205
40fb1ced2598f5f2372a29819baaab5c5aab8208: Bug 1540136 - P1: Log more info in ChromiumCDMProxy dtor and Shutdown. r=cpearce a=lizzard
Bryce Van Dyk <bvandyk@mozilla.com> - Tue, 09 Apr 2019 15:05:01 +0000 - rev 523222
Push 11107 by archaeopteryx@coole-files.de at Thu, 18 Apr 2019 15:56:49 +0000
Bug 1540136 - P1: Log more info in ChromiumCDMProxy dtor and Shutdown. r=cpearce a=lizzard Also remove unneeded MOZ_COUNT_[CTOR|DTOR] macros. We already get similar functionality from NS_INLINE_DECL_THREADSAFE_REFCOUNTING. Differential Revision: https://phabricator.services.mozilla.com/D26204
778c3da2f7ae901e6cb55115217f8fb32a798bae: Bug 1542921 - turn on block autoplay by default. r=cpearce a=pascalc
alwu <alwu@mozilla.com> - Tue, 09 Apr 2019 23:43:33 +0000 - rev 523127
Push 11064 by apavel@mozilla.com at Thu, 11 Apr 2019 08:46:03 +0000
Bug 1542921 - turn on block autoplay by default. r=cpearce a=pascalc As we have successfully rolled out the feature via Normandy in 66, we should turn the feature on in >= 67 to not have to rely on Normandy going forward any more. Differential Revision: https://phabricator.services.mozilla.com/D26713
062f54735111d741f9189f2c15e2b4da740783fa: Bug 1522547 - Have MediaKeys observe if the owning document becomes inactive so it can shutdown. r=cpearce
Bryce Van Dyk <bvandyk@mozilla.com> - Tue, 05 Mar 2019 16:21:40 +0000 - rev 520333
Push 10862 by ffxbld-merge at Mon, 11 Mar 2019 13:01:11 +0000
Bug 1522547 - Have MediaKeys observe if the owning document becomes inactive so it can shutdown. r=cpearce MediaKeys objects are typically created and associated with an HTMLMediaElement, however it is possible to create a MediaKeys object and not associate it with an HTMLMediaElement. This resulted in an issue where these MediaKeys would keep alive other components that would assert during bowrser shutdown (see bug 1522547). We anticipated that MediaKeys associated with an HTMLMediaElement would need to be shutdown if their owning document became inactive, but were not handling the case where the keys never became associated with an element. This patch has the MediaKeys listen directly to their owning document for activity change. The MediaKeys will shutdown if their document becomes inactive. This avoids MediaKeys not associated with HTMLMediaElements keeping other objects alive during browser shutdown. Differential Revision: https://phabricator.services.mozilla.com/D21983
153245ec86c26f88beba43d7858cf5c517e22d84: Bug 1522547 - Log in MediaKeySession dtor. r=cpearce
Bryce Van Dyk <bvandyk@mozilla.com> - Tue, 05 Mar 2019 16:21:30 +0000 - rev 520332
Push 10862 by ffxbld-merge at Mon, 11 Mar 2019 13:01:11 +0000
Bug 1522547 - Log in MediaKeySession dtor. r=cpearce Differential Revision: https://phabricator.services.mozilla.com/D21982
c273b07bc5f06aa68c6d8d2ce7074941a9075cad: Bug 1522547 - Add more logs to MediaKeys, use PRIu32 for printing where appropriate. r=cpearce
Bryce Van Dyk <bvandyk@mozilla.com> - Tue, 05 Mar 2019 16:21:28 +0000 - rev 520331
Push 10862 by ffxbld-merge at Mon, 11 Mar 2019 13:01:11 +0000
Bug 1522547 - Add more logs to MediaKeys, use PRIu32 for printing where appropriate. r=cpearce - Add more logging to MediaKeys. - Replace the instances of the %d formatter with PRIu32 when printing the PromiseId type (which has the underlying uint32_t type). Differential Revision: https://phabricator.services.mozilla.com/D21981
c092112f933effac6fbedd252a99bc0fdf0f9411: Bug 1522547 - Log `this` in ChromiumCDMProxy logs, use PRIu32 where appropriate. r=cpearce
Bryce Van Dyk <bvandyk@mozilla.com> - Tue, 05 Mar 2019 16:21:24 +0000 - rev 520330
Push 10862 by ffxbld-merge at Mon, 11 Mar 2019 13:01:11 +0000
Bug 1522547 - Log `this` in ChromiumCDMProxy logs, use PRIu32 where appropriate. r=cpearce Because multiple ChromiumCDMProxy object may exist during a browser lifetime, logging the value of `this` in their logs is useful for disambiguating log statements, as well as matching ChromiumCDMProxy objects to pointers held by other objects. Update formatter usage in logs to prefer PRIu32 over %u when the underlying type is uint32_t. Differential Revision: https://phabricator.services.mozilla.com/D21980
7e199486530b42d3b366883770197f03f143f924: Bug 1522547 - Differentiate GMPServiceParent and GMPServiceChild __CLASS__ macro. r=cpearce
Bryce Van Dyk <bvandyk@mozilla.com> - Tue, 05 Mar 2019 16:21:20 +0000 - rev 520329
Push 10862 by ffxbld-merge at Mon, 11 Mar 2019 13:01:11 +0000
Bug 1522547 - Differentiate GMPServiceParent and GMPServiceChild __CLASS__ macro. r=cpearce The macros in these classes are used to output class names in logs. Differentiating them helps make logs clearer. Differential Revision: https://phabricator.services.mozilla.com/D21979
6f4b04b69f8073c7c7a9cdfb3bfc511dec887105: Bug 1522547 - Add logging to GMPContentParent. r=cpearce
Bryce Van Dyk <bvandyk@mozilla.com> - Tue, 05 Mar 2019 16:19:43 +0000 - rev 520328
Push 10862 by ffxbld-merge at Mon, 11 Mar 2019 13:01:11 +0000
Bug 1522547 - Add logging to GMPContentParent. r=cpearce Use GMPLog.h in GMPContentPareant. Add logging to most functions. This logging was added to aid in diagnosing a shutdown crash, but should be generally useful to have. Driveby touchup of arg name to ChromiumCDMDestroyed to match header. Differential Revision: https://phabricator.services.mozilla.com/D21978
0bd2a637efa3b877d7145fe9d186afa6550e3ef2: Bug 1532354 - Remove virtual keyword on GMPServiceChild|GMPServiceParent::GetContentParent. r=cpearce
Bryce Van Dyk <bvandyk@mozilla.com> - Tue, 05 Mar 2019 00:32:07 +0000 - rev 520248
Push 10862 by ffxbld-merge at Mon, 11 Mar 2019 13:01:11 +0000
Bug 1532354 - Remove virtual keyword on GMPServiceChild|GMPServiceParent::GetContentParent. r=cpearce These functions already specify the override keyword. This change also makes them more consistent with the overloaded versions of the same functions in these classes. Depends on D21968 Differential Revision: https://phabricator.services.mozilla.com/D21975
65fc1cf8774f32b733ec0235adb1f61677fdff5f: Bug 1532354 - Remove unused DeleteGMPServiceParent class. r=cpearce
Bryce Van Dyk <bvandyk@mozilla.com> - Tue, 05 Mar 2019 00:31:27 +0000 - rev 520247
Push 10862 by ffxbld-merge at Mon, 11 Mar 2019 13:01:11 +0000
Bug 1532354 - Remove unused DeleteGMPServiceParent class. r=cpearce Bug 1352924 removed the usage of this class, so we can safely remove the dead code. Differential Revision: https://phabricator.services.mozilla.com/D21968
c1e80167030701714306b8e18a4b8d3ac49187aa: Bug 1532354 - Remove GMPServiceParent::ProcessPossiblePlugin decl as it is unimplemented. r=cpearce
Bryce Van Dyk <bvandyk@mozilla.com> - Tue, 05 Mar 2019 00:30:13 +0000 - rev 520246
Push 10862 by ffxbld-merge at Mon, 11 Mar 2019 13:01:11 +0000
Bug 1532354 - Remove GMPServiceParent::ProcessPossiblePlugin decl as it is unimplemented. r=cpearce Differential Revision: https://phabricator.services.mozilla.com/D21967
07c49703b70c4cd01985771942e84aa37a66663c: Bug 1530220 - part4 : separate tesing web audio with gum to another test. r=cpearce
Alastor Wu <alwu@mozilla.com> - Tue, 05 Mar 2019 02:13:55 +0000 - rev 520209
Push 10862 by ffxbld-merge at Mon, 11 Mar 2019 13:01:11 +0000
Bug 1530220 - part4 : separate tesing web audio with gum to another test. r=cpearce Testing web audio with GUM is not related with the original purpose of this test. In order to reduce the complexity of this test, separate this part as another new test. Also modify the naming of functions, because Mozilla coding style is to use the camelCase for the name. Differential Revision: https://phabricator.services.mozilla.com/D21257
f706441a86b530d30be4fba47bff43f96ade97cb: Bug 1530220 - part2 : modify test in order to test different kinds of keycode. r=cpearce
Alastor Wu <alwu@mozilla.com> - Tue, 05 Mar 2019 02:09:50 +0000 - rev 520207
Push 10862 by ffxbld-merge at Mon, 11 Mar 2019 13:01:11 +0000
Bug 1530220 - part2 : modify test in order to test different kinds of keycode. r=cpearce Except printable keys, we would treat `carriage return` and `space` as supported user gesture inputs for activating document. Other keys won't activate document. Differential Revision: https://phabricator.services.mozilla.com/D21255
7c1b053e161b0795c83fc39289a8d4335b146127: Bug 1530220 - part1 : allow some non-printalble keys as supported user gesture inputs to activate document. r=masayuki,cpearce
Alastor Wu <alwu@mozilla.com> - Tue, 05 Mar 2019 02:06:22 +0000 - rev 520206
Push 10862 by ffxbld-merge at Mon, 11 Mar 2019 13:01:11 +0000
Bug 1530220 - part1 : allow some non-printalble keys as supported user gesture inputs to activate document. r=masayuki,cpearce `carriage return` and `space` are common keys which user might use to start media, so we should take account them as supported user gesture inputs. As their pseudo char code are zero, we have to check their key code in order to distinguish them from other controls keys such as shift, alt... Differential Revision: https://phabricator.services.mozilla.com/D21253
6e3eb5df50ff8607f0f60e8f7f8b9ed573f74fa4: Bug 1526355 - part1 : notify the front end side to show the blocking icon if the site is still being blocked after it's back from the bfcache. r=cpearce
Alastor Wu <alwu@mozilla.com> - Mon, 04 Mar 2019 17:25:52 +0000 - rev 520123
Push 10862 by ffxbld-merge at Mon, 11 Mar 2019 13:01:11 +0000
Bug 1526355 - part1 : notify the front end side to show the blocking icon if the site is still being blocked after it's back from the bfcache. r=cpearce In order to display blocking icon when the document comes back from the bfcache, we have to notify front end what's the current blocking status. As the front end side would clear blocking autoplay information when nagivation occurs, and the media might not invoke the play again when they comes back from the bfcache. Therefore, we should notify front end side that the site is still being blocked, and we should show blocking icon for it. Differential Revision: https://phabricator.services.mozilla.com/D21582
62d4e16fbb2b655a82418553929c9883279273a5: Bug 1531500 - Fix HE-AAC with WMF decoder. r=cpearce
Jean-Yves Avenard <jyavenard@mozilla.com> - Fri, 01 Mar 2019 00:54:32 +0000 - rev 519723
Push 10862 by ffxbld-merge at Mon, 11 Mar 2019 13:01:11 +0000
Bug 1531500 - Fix HE-AAC with WMF decoder. r=cpearce The rate changes after decoding the first sample to what was indicated in the container. The code to handle that case was incorrectly removed in bug 1530234 Differential Revision: https://phabricator.services.mozilla.com/D21618
0a003a54ff4554cb8b68f3f0d54a6bd3ef166c16: Bug 1525156 - part2 : Add Telemetry to know how many autoplay media, which has been resumed from blocked, played exactly 7 secocnds or more, or less than 7 seconds. r=cpearce,janerik
Alastor Wu <alwu@mozilla.com> - Tue, 26 Feb 2019 04:44:19 +0000 - rev 519109
Push 10862 by ffxbld-merge at Mon, 11 Mar 2019 13:01:11 +0000
Bug 1525156 - part2 : Add Telemetry to know how many autoplay media, which has been resumed from blocked, played exactly 7 secocnds or more, or less than 7 seconds. r=cpearce,janerik By adding the Telemetry to measure the number of video/audio which played exactly 7 seconds or more, or less than 7 seconds, after those media has been resumed from blocked state, we can know how many media would meet the Chrome's MEI condition, which could help us to know more about the whole landscape of autoplay media. In addition, it could help us know how many media are played 'by users intention' because we assume that users are more likely to stop the media if autoplay media is unblocked by accident. Differential Revision: https://phabricator.services.mozilla.com/D18628