Eliminate MediaPlayer & MediaPlayerClient abstractions(seek, duration, poster APIs)

Source/core/html/HTMLMediaElement.cpp

Index: Source/core/html/HTMLMediaElement.cpp
diff --git a/Source/core/html/HTMLMediaElement.cpp b/Source/core/html/HTMLMediaElement.cpp
index 91f87f6d7aef089312dd1dd65cbea2124e420c4d..900164c3b2b4789a598058c3731ffd91ce2e8cf1 100644
--- a/Source/core/html/HTMLMediaElement.cpp
+++ b/Source/core/html/HTMLMediaElement.cpp
@@ -1841,7 +1841,8 @@ void HTMLMediaElement::seek(double time, ExceptionState& exceptionState)
     scheduleEvent(EventTypeNames::seeking);
 
     // 9 - Set the current playback position to the given new playback position
-    m_player->seek(time);
+    if (webMediaPlayer())

[acolwell GONE FROM CHROMIUM, 2014/06/02 22:34:17]

You should not need this check. You shouldn't be able to get to this point w/o
having a webMediaPlayer(). The only way to get m_readyState > HAVE_NOTHING is
for webMediaPlayer() to actually exist and triggered a transition to something
else.

[Srirama, 2014/06/03 07:06:02]

Done.

+        webMediaPlayer()->seek(time);
 
     // 10-14 are handled, if necessary, when the engine signals a readystate change or otherwise
     // satisfies seek completion and signals a time change.
@@ -1881,7 +1882,7 @@ bool HTMLMediaElement::seeking() const
 
 void HTMLMediaElement::refreshCachedTime() const
 {
-    m_cachedTime = m_player->currentTime();
+    m_cachedTime = webMediaPlayer() ? webMediaPlayer()->currentTime() : 0.0;

[acolwell GONE FROM CHROMIUM, 2014/06/02 22:34:17]

You shouldn't need this check. This method shouldn't be called if we don't have
a webMediaPlayer().

[Srirama, 2014/06/03 07:06:02]

It is null for preload:none case, when it is called from updatePlayState. So
should we have this check here, or should i add it in updatePlayState ?

     m_cachedTimeWallClockUpdateTime = WTF::currentTime();
 }
 
@@ -1915,7 +1916,8 @@ double HTMLMediaElement::currentTime() const
 
     if (m_cachedTime != MediaPlayer::invalidTime() && m_paused) {
 #if LOG_CACHED_TIME_WARNINGS
-        double delta = m_cachedTime - m_player->currentTime();
+        double playerCurrentTime = webMediaPlayer() ? webMediaPlayer()->currentTime() : 0.0;

[acolwell GONE FROM CHROMIUM, 2014/06/02 22:34:17]

This check should be moved to the top with the !m_player check since it will
eventually replace it.

[Srirama, 2014/06/03 07:06:02]

Done.

+        double delta = m_cachedTime - playerCurrentTime;
         if (delta > minCachedDeltaForWarning)
             WTF_LOG(Media, "HTMLMediaElement::currentTime - WARNING, cached time is %f seconds off of media time when paused", delta);
 #endif
@@ -1954,7 +1956,7 @@ double HTMLMediaElement::duration() const
     if (m_mediaSource)
         return m_mediaSource->duration();
 
-    return m_player->duration();
+    return webMediaPlayer() ? webMediaPlayer()->duration() : 0.0;

[acolwell GONE FROM CHROMIUM, 2014/06/02 22:34:17]

This check should not be needed. webMediaPlayer() must be valid for m_readyState
>= HAVE_METADATA.

[Srirama, 2014/06/03 07:06:02]

Done.

 }
 
 bool HTMLMediaElement::paused() const
@@ -2770,7 +2772,8 @@ void HTMLMediaElement::mediaPlayerTimeChanged()
     invalidateCachedTime();
 
     // 4.8.10.9 steps 12-14. Needed if no ReadyState change is associated with the seek.
-    if (m_seeking && m_readyState >= HAVE_CURRENT_DATA && !m_player->seeking())
+    if (m_seeking && m_readyState >= HAVE_CURRENT_DATA
+        && !(webMediaPlayer() && webMediaPlayer()->seeking()))

[acolwell GONE FROM CHROMIUM, 2014/06/02 22:34:17]

the pointer check should not be necessary since webMediaPlayer() must be valid
if m_readyState >= HAVE_CURRENT_DATA. Also this method is actually being called
from the webMediaPlayer() instance so it better be non-null at this point.

[Srirama, 2014/06/03 07:06:02]

Done.

         finishSeek();
 
     // Always call scheduleTimeupdateEvent when the media engine reports a time discontinuity,