RESOLVED FIXED 208423
[GPUP] Plumb through more MediaPlayer methods
https://bugs.webkit.org/show_bug.cgi?id=208423
Summary [GPUP] Plumb through more MediaPlayer methods
Eric Carlson
Reported 2020-02-29 18:13:51 PST
Plumb through more MediaPlayer methods when media is loaded in the GPU process.
Attachments
Patch (64.11 KB, patch)
2020-02-29 18:22 PST, Eric Carlson
no flags
Patch (61.80 KB, patch)
2020-02-29 19:53 PST, Eric Carlson
no flags
Patch (62.68 KB, patch)
2020-02-29 22:12 PST, Eric Carlson
no flags
Patch (64.68 KB, patch)
2020-03-02 05:42 PST, Eric Carlson
no flags
Patch (64.68 KB, patch)
2020-03-02 06:36 PST, Eric Carlson
no flags
Patch (1.47 KB, patch)
2020-03-02 11:17 PST, Eric Carlson
no flags
Patch (1.42 KB, patch)
2020-03-02 11:18 PST, Eric Carlson
no flags
Radar WebKit Bug Importer
Comment 1 2020-02-29 18:14:00 PST
Eric Carlson
Comment 2 2020-02-29 18:22:35 PST
Eric Carlson
Comment 3 2020-02-29 19:53:46 PST
Eric Carlson
Comment 4 2020-02-29 22:12:19 PST
youenn fablet
Comment 5 2020-03-02 05:11:33 PST
Comment on attachment 392093 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=392093&action=review > Source/WebKit/GPUProcess/media/RemoteMediaPlayerProxy.cpp:279 > + m_cachedState.didPassCORSAccessCheck = m_player->didPassCORSAccessCheck(); Should some of these state updates being added to updateCachedState(), for instance wirelessVideoPlaybackDisabled()? Should we always resync all values for extra safety? > Source/WebKit/WebProcess/GPU/media/MediaPlayerPrivateRemote.cpp:675 > + if (!connection().sendSync(Messages::RemoteMediaPlayerProxy::AccessLog(m_id), Messages::RemoteMediaPlayerProxy::AccessLog::Reply(log), m_id)) Seems to make this fail on iOS since AccessLog is not defined for RemoteMediaPlayerProxy.
Eric Carlson
Comment 6 2020-03-02 05:33:39 PST
Comment on attachment 392093 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=392093&action=review >> Source/WebKit/GPUProcess/media/RemoteMediaPlayerProxy.cpp:279 >> + m_cachedState.didPassCORSAccessCheck = m_player->didPassCORSAccessCheck(); > > Should some of these state updates being added to updateCachedState(), for instance wirelessVideoPlaybackDisabled()? > Should we always resync all values for extra safety? I don't think we want to update all of these in updateCachedState because it is called so frequently I tried to make sure every cached state is updated in the appropriate media player callback. For example, m_cachedState.wirelessVideoPlaybackDisabled and m_cachedState.wirelessPlaybackTargetName are updated in RemoteMediaPlayerProxy::mediaPlayerCurrentPlaybackTargetIsWirelessChanged if the framework changes @externalPlaybackActive or @allowsExternalPlayback.
Eric Carlson
Comment 7 2020-03-02 05:42:37 PST
Eric Carlson
Comment 8 2020-03-02 06:36:12 PST
WebKit Commit Bot
Comment 9 2020-03-02 07:54:35 PST
Comment on attachment 392138 [details] Patch Clearing flags on attachment: 392138 Committed r257711: <https://trac.webkit.org/changeset/257711>
WebKit Commit Bot
Comment 10 2020-03-02 07:54:37 PST
All reviewed patches have been landed. Closing bug.
Eric Carlson
Comment 11 2020-03-02 11:17:03 PST
Reopening to attach new patch.
Eric Carlson
Comment 12 2020-03-02 11:17:04 PST
Eric Carlson
Comment 13 2020-03-02 11:18:37 PST
WebKit Commit Bot
Comment 14 2020-03-02 12:07:48 PST
Comment on attachment 392160 [details] Patch Clearing flags on attachment: 392160 Committed r257724: <https://trac.webkit.org/changeset/257724>
WebKit Commit Bot
Comment 15 2020-03-02 12:07:50 PST
All reviewed patches have been landed. Closing bug.
Note You need to log in before you can comment on or make changes to this bug.