Extract image URLs in srcset as well

java/org/chromium/distiller/DomUtil.java

 // Copyright 2014 The Chromium Authors. All rights reserved.
 // Use of this source code is governed by a BSD-style license that can be
 // found in the LICENSE file.
 
 package org.chromium.distiller;
 
 import com.google.gwt.core.client.JsArray;
 import com.google.gwt.core.client.JsArrayString;
 import com.google.gwt.dom.client.AnchorElement;
 import com.google.gwt.dom.client.Document;
@@ skipping 171 matching lines @@
     public static native String javascriptTextContent(Node node) /*-{
         return node.textContent;
     }-*/;
 
     /**
      * Get a list of all the parents of this node starting with the node itself.
      * @param n The node to get the parents of.
      * @return A list of the provided node's parents.
      */
     public static List<Node> getParentNodes(Node n) {
-        ArrayList<Node> result = new ArrayList<Node>();
+        ArrayList<Node> result = new ArrayList<>();
         Node curr = n;
         while (curr != null) {
             result.add(curr);
             curr = curr.getParentNode();
         }
         return result;
     }
 
     /**
      * Get the depth of the given node in the DOM tree (only counting elements).
@@ skipping 119 matching lines @@
             makeSrcSetAbsolute(ImageElement.as(root));
         }
         NodeList<Element> imgs = DomUtil.querySelectorAll(root, "IMG[SRCSET]");
         for (int i = 0; i < imgs.getLength(); i++) {
             makeSrcSetAbsolute(ImageElement.as(imgs.getItem(i)));
         }
     }
 
     public static void makeSrcSetAbsolute(ImageElement ie) {
         String srcset = ie.getAttribute("srcset");
-        if (srcset == "") {
+        if (srcset.isEmpty()) {
             ie.removeAttribute("srcset");
             return;
         }
 
         String oldsrc = ie.getSrc();
         String[] sizes = StringUtil.jsSplit(srcset, ",");
         for(int i = 0; i < sizes.length; i++) {
             String size = StringUtil.jsTrim(sizes[i]);
             if (size.isEmpty()) continue;
             String[] comp = size.split(" ");
             ie.setSrc(comp[0]);
             comp[0] = ie.getSrc();
             sizes[i] = StringUtil.join(comp, " ");
         }
         ie.setAttribute("srcset", StringUtil.join(sizes, ", "));
         ie.setSrc(oldsrc);
     }
 
+    public static List<String> getSrcSetUrls(ImageElement ie) {
+        List<String> list = new ArrayList<>();
+        String srcset = ie.getAttribute("srcset");
+        if (srcset.isEmpty()) {
+            return list;
+        }
+
+        String[] sizes = StringUtil.jsSplit(srcset, ",");
+        for(int i = 0; i < sizes.length; i++) {
+            String size = StringUtil.jsTrim(sizes[i]);
+            if (size.isEmpty()) continue;
+            String[] comp = size.split(" ");
+            list.add(comp[0]);
+        }
+        return list;
+    }
+
     public static void stripImageElements(Node root) {
         if (root.getNodeType() == Node.ELEMENT_NODE) {
             Element element = Element.as(root);
             if (element.getTagName().equals("IMG")) {
                 stripImageElement(ImageElement.as(element));
             }
         }
         NodeList<Element> imgs = DomUtil.querySelectorAll(root, "IMG");
         for (int i = 0; i < imgs.getLength(); i++) {
             stripImageElement(ImageElement.as(imgs.getItem(i)));
@@ skipping 163 matching lines @@
     }-*/;
 
     public static native Document createHTMLDocument(Document doc) /*-{
         return doc.implementation.createHTMLDocument();
     }-*/;
 
     public static native Element getFirstElementChild(Document document) /*-{
         return document.firstElementChild;
     }-*/;
 }