Document calls to the "rules for updating the text track rendering"

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 ae9acd56ba267b4f882581e388d8ac110b184501..b2b989cecc0fee9bdc3b7db6d7e07eed17e11eaa 100644
--- a/Source/core/html/HTMLMediaElement.cpp
+++ b/Source/core/html/HTMLMediaElement.cpp
@@ -3210,6 +3210,9 @@ void HTMLMediaElement::setClosedCaptionsVisible(bool closedCaptionVisible)
     honorUserPreferencesForAutomaticTextTrackSelection();
     m_processingPreferenceChange = false;
 
+    // As track visibility changed while m_processingPreferenceChange was set,
+    // there was no call to updateTextTrackDisplay(). This call is not in the
+    // spec, see the note in configureTextTrackDisplay().
     updateTextTrackDisplay();
 }
 
@@ -3348,7 +3351,15 @@ void HTMLMediaElement::configureTextTrackDisplay(VisibilityChangeAssumption assu
     mediaControls()->changedClosedCaptionsVisibility();
 
     cueTimeline().updateActiveCues(currentTime());
-    updateTextTrackDisplay();
+
+    // Note: The "time marches on" algorithm (updateActiveCues) runs the "rules
+    // for updating the text track rendering" (updateTextTrackDisplay) only for
+    // "affected tracks", i.e. tracks where the the active cues have changed.
+    // This misses cues in tracks that have gone from hidden to showing. This
+    // appears to be a spec bug, which we work around here:
+    // https://www.w3.org/Bugs/Public/show_bug.cgi?id=28236
+    if (assumption == AssumeVisibleChange)

[fs, 2015/03/19 09:50:33]

I'm not really 100% convinced that this call should be predicated like this -
but I'm not quite sure what the |assumption| is expected to achieve either (I
think I'd rather drop it.)

[philipj_slow, 2015/03/19 10:04:54]

To play it safe I can make it unconditional again, making this CL documentation
only.

I'd also be happy to see VisibilityChangeAssumption (and RemovalNotification and
many flags of this sort) gone, but it would take some history digging to
understand what the point is.

+        updateTextTrackDisplay();
 }
 
 void* HTMLMediaElement::preDispatchEventHandler(Event* event)