Consuming a touch move prevents only the next scroll update.

content/browser/renderer_host/input/gesture_event_packet.h

Index: content/browser/renderer_host/input/gesture_event_packet.h
diff --git a/content/browser/renderer_host/input/gesture_event_packet.h b/content/browser/renderer_host/input/gesture_event_packet.h
index 1227417ee86302c6fc6daee6ef39fb76adccd45c..4d5f89ab6473058c387d27d8c507fc48708f1f5d 100644
--- a/content/browser/renderer_host/input/gesture_event_packet.h
+++ b/content/browser/renderer_host/input/gesture_event_packet.h
@@ -15,10 +15,12 @@ namespace content {
 class CONTENT_EXPORT GestureEventPacket {
  public:
   enum GestureSource {
-    INVALID = -1,  // Used only for a default-constructed packet..
-    TOUCH_BEGIN,   // The start of a new gesture sequence.
-    TOUCH,         // Continuation of an existing gesture sequence.
-    TOUCH_TIMEOUT, // Timeout from an existing gesture sequence.
+    INVALID = -1,      // Used only for a default-constructed packet.
+    TOUCH_SEQUENCE_BEGIN,       // The start of a new gesture sequence.
+    TOUCH_BEGIN, // A non-primary pointer was put down. TODO - reword these.
+    TOUCH_MOVE,        // A touch point moved.
+    TOUCH_END,   // A touch point lifted.

[jdduke (slow), 2014/02/11 16:02:25]

Now that we're adding a few types, do you think it's worth having a
TOUCH_SEQUENCE_END that would pop a sequence as soon as it's ack'ed?  Rather
than just waiting until we get the ack of a TOUCH_SEQUENCE_BEGIN?   

Also, are we taking into consideration that a TouchCancel cannot be
preventDefault'ed? Looking at the code, I think this gets handled properly
(e.g., GestureScrollEnd/GestureTapCancel on a TouchCancel), but we should
probably add some test for this.

[tdresser, 2014/02/11 21:04:32]

Done.

+    TOUCH_TIMEOUT,     // Timeout from an existing gesture sequence.
   };
 
   GestureEventPacket();
@@ -42,6 +44,7 @@ private:
   blink::WebGestureEvent gestures_[kMaxGesturesPerTouch];
   size_t gesture_count_;
   GestureSource gesture_source_;
+  bool is_independent_;

[jdduke (slow), 2014/02/11 16:02:25]

Do we still need this flag?

[tdresser, 2014/02/11 21:04:33]

Nope, removed.

 };
 
 }  // namespace content