Add saveImageFromDataURL() and use it in saveImageAt().

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 3273 matching lines @@
         // There isn't actually an image at these coordinates.  Might be because
         // the window scrolled while the context menu was open or because the page
         // changed itself between when we thought there was an image here and when
         // we actually tried to retreive the image.
         //
         // FIXME: implement a cache of the most recent HitTestResult to avoid having
         //        to do two hit tests.
         return;
     }
 
     m_page->deprecatedLocalMainFrame()->editor().copyImage(result);

[Justin Novosad, 2014/09/05 17:22:46]

Don't we have the same bug here too?

[zino, 2014/09/05 18:00:25]

On 2014/09/05 17:22:46, junov wrote:

Internally, the API copies a wrapper object for bitmap image and its url to
shared memory and then renderer sends a pointer of the shared memory to browser.
So, it is okay because GURL marshalling doesn't happen.

 }
 
 void WebViewImpl::saveImageAt(const WebPoint& point)
 {
-    if (!m_page)
+    if (!m_client)
         return;
 
-    KURL url = hitTestResultForWindowPos(point).absoluteImageURLIncludingCanvasDataURL();
-
-    if (url.isEmpty())
+    Node* node = hitTestResultForWindowPos(point).innerNonSharedNode();
+    if (!node || !(isHTMLCanvasElement(*node) || isHTMLImageElement(*node)))
         return;
 
-    ResourceRequest request(url);
-    request.setRequestContext(WebURLRequest::RequestContextDownload);
-    m_page->deprecatedLocalMainFrame()->loader().client()->loadURLExternally(
-        request, NavigationPolicyDownloadTo, WebString());
+    String url = toElement(*node).imageSourceURL();
+    if (!KURL(KURL(), url).protocolIsData())
+        return;

[Justin Novosad, 2014/09/05 17:22:46]

If the user attempts to save an image with an http URL, we just return? Isn't
that a regression?  I think what kbr was suggestion is that in the case we
should do things the old way.

[zino, 2014/09/05 18:00:26]

On 2014/09/05 17:22:46, junov wrote:

If the user attempts to save an image with an http URL, we don't use this code
path. In that case, browser process can immediately start to download the image
without this API. because the process already knows the URL.

<canvas> : use saveImageAt()
<img> with a large data url : use saveImageAt()
<img> with http, https, and so on : use WebContents::saveFrame()

[Ken Russell (switch to Gerrit), 2014/09/11 00:43:06]

Thanks for clarifying this.

+
+    m_client->saveImageFromDataURL(url);
 }
 
 void WebViewImpl::dragSourceEndedAt(
     const WebPoint& clientPoint,
     const WebPoint& screenPoint,
     WebDragOperation operation)
 {
     PlatformMouseEvent pme(clientPoint, screenPoint, LeftButton, PlatformEvent::MouseMoved,
         0, false, false, false, false, PlatformMouseEvent::RealOrIndistinguishable, 0);
     m_page->deprecatedLocalMainFrame()->eventHandler().dragSourceEndedAt(pme,
@@ skipping 957 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