Add WebKit API for doing a hit-test that mimics GestureTap

Source/web/WebViewImpl.cpp

 /*
  * Copyright (C) 2011, 2012 Google Inc. All rights reserved.
  *
  * Redistribution and use in source and binary forms, with or without
  * modification, are permitted provided that the following conditions are
  * met:
  *
  *     * Redistributions of source code must retain the above copyright
  * notice, this list of conditions and the following disclaimer.
  *     * Redistributions in binary form must reproduce the above
@@ skipping 4048 matching lines @@
 HitTestResult WebViewImpl::hitTestResultForWindowPos(const IntPoint& pos)
 {
     if (!m_page->mainFrame()->isLocalFrame())
         return HitTestResult();
     IntPoint docPoint(m_page->deprecatedLocalMainFrame()->view()->windowToContents(pos));
     HitTestResult result = m_page->deprecatedLocalMainFrame()->eventHandler().hitTestResultAtPoint(docPoint, HitTestRequest::ReadOnly | HitTestRequest::Active);
     result.setToShadowHostIfInUserAgentShadowRoot();
     return result;
 }
 
+WebHitTestResult WebViewImpl::hitTestResultForTap(const WebPoint& tapPointWindowPos, const WebSize& tapArea)
+{
+    // EventHandler does not care about the global position.

[hush (inactive), 2014/10/28 22:52:25]

What should be the correct global position here for gestureEvent?

[Rick Byers, 2014/11/04 06:09:19]

In the general tap case, the screen position can be passed through to JavaScript
(eg. as the MouseEvent.screen* properties).  In this case since you're doing
only a hit test, it should be completely unused.  I'd set it to 0,0 instead of
something that looks real.

Also strictly EventHandler does care (in the sense that it'll pass them
through), but the hit-testing phase doesn't.  So replace "EventHandler" with
"hit testing".

[hush (inactive), 2014/11/04 22:42:39]

Done

+    PlatformGestureEvent gestureEvent(PlatformEvent::GestureTap, IntPoint(tapPointWindowPos), IntPoint(tapPointWindowPos), IntSize(tapArea), 0, false, false, false, false, 0, 0, 0, 0);
+    if (!m_page->mainFrame()->isLocalFrame())
+        return HitTestResult();
+    HitTestResult result = m_page->deprecatedLocalMainFrame()->eventHandler().hitTestResultForGestureEvent(gestureEvent, HitTestRequest::ReadOnly | HitTestRequest::Active).hitTestResult();
+
+    result.setToShadowHostIfInUserAgentShadowRoot();

[hush (inactive), 2014/10/28 22:52:25]

Not sure about what this line does, but hitTestResultForWindowPos does this, so
I copied it here.

[Rick Byers, 2014/11/04 06:09:19]

You do want this.  It means, for example, that if a user taps on a form element,
you'll get the <input> element returned, not one of the hidden elements inside
of it used to implement <input> under the covers.

+    return result;
+}
+
 void WebViewImpl::setTabsToLinks(bool enable)
 {
     m_tabsToLinks = enable;
 }
 
 bool WebViewImpl::tabsToLinks() const
 {
     return m_tabsToLinks;
 }
 
@@ skipping 394 matching lines @@
     const PageScaleConstraints& constraints = m_pageScaleConstraintsSet.pageDefinedConstraints();
 
     if (!mainFrameImpl() || !mainFrameImpl()->frameView())
         return false;
 
     return mainFrameImpl()->frameView()->layoutSize().width() == m_size.width
         || (constraints.minimumScale == constraints.maximumScale && constraints.minimumScale != -1);
 }
 
 } // namespace blink