Extract image URLs in srcset as well

java/org/chromium/distiller/DomUtil.java

Index: java/org/chromium/distiller/DomUtil.java
diff --git a/java/org/chromium/distiller/DomUtil.java b/java/org/chromium/distiller/DomUtil.java
index 3cca24bc50de03876d1c477ac2675ff5cb4ff06f..1ff099b4ed6320ca5ebf80de1a191bdb329a97a2 100644
--- a/java/org/chromium/distiller/DomUtil.java
+++ b/java/org/chromium/distiller/DomUtil.java
@@ -189,7 +189,7 @@ public class DomUtil {
      * @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);
@@ -329,7 +329,7 @@ public class DomUtil {
 
     public static void makeSrcSetAbsolute(ImageElement ie) {
         String srcset = ie.getAttribute("srcset");
-        if (srcset == "") {
+        if (srcset.isEmpty()) {
             ie.removeAttribute("srcset");
             return;
         }
@@ -348,6 +348,23 @@ public class DomUtil {
         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);