Clean up attributes of image elements

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 198 matching lines @@
         NodeTree expanded = NodeListExpander.expand(outputNodes);
         Node clonedSubtree = expanded.cloneSubtreeRetainDirection();
 
         if (clonedSubtree.getNodeType() != Node.ELEMENT_NODE) return "";
 
         stripIds(clonedSubtree);
         makeAllLinksAbsolute(clonedSubtree);
         stripFontColorAttributes(clonedSubtree);
         stripTableBackgroundColorAttributes(clonedSubtree);
         stripStyleAttributes(clonedSubtree);
+        stripImageElements(clonedSubtree);
 
         if (textOnly) {
             return DomUtil.getTextFromTree(clonedSubtree);
         }
         return Element.as(clonedSubtree).getString();
     }
 
     /**
      * Makes all anchors and video posters absolute. This calls "makeAllSrcAttributesAbsolute".
      * @param rootNode The root Node to look through.
@@ skipping 24 matching lines @@
     }
 
     private static void makeSrcSetAbsolute(Element 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 == "") {
+            ie.removeAttribute("srcset");
+            return;
+        }
+
         String oldsrc = ie.getSrc();
-        String[] sizes = StringUtil.jsSplit(ie.getAttribute("srcset"), ",");
+        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 void stripImageElements(Node root) {
+        NodeList<Element> imgs = DomUtil.querySelectorAll(root, "IMG");
+        for (int i = 0; i < imgs.getLength(); i++) {
+            stripImageElement(ImageElement.as(imgs.getItem(i)));
+        }
+    }
+
+    /**
+     * Only keep some attributes for image elements.
+     * @param ie The image element to strip in-place.
+     */
+    public static void stripImageElement(ImageElement imgElement) {
+        JsArray<Node> attrs = getAttributes(imgElement);
+        for (int i = 0; i < attrs.length(); ) {
+            String name = attrs.get(i).getNodeName();
+            if (!"src".equals(name) &&
+                !"alt".equals(name) &&
+                !"srcset".equals(name) &&
+                !"dir".equals(name) &&
+                !"title".equals(name)) {
+                imgElement.removeAttribute(name);
+            } else {
+                i++;
+            }
+        }
+    }
+
     /**
      * Makes all "img", "source", "track", and "video" tags have an absolute "src" attribute.
      * @param root The root element to look through.
      */
     public static native void makeAllSrcAttributesAbsolute(Element root) /*-{
         var elementsWithSrc = root.querySelectorAll('img,source,track,video');
         for (var key in elementsWithSrc) {
             if (elementsWithSrc[key].src) {
                 elementsWithSrc[key].src = elementsWithSrc[key].src;
             }
@@ skipping 107 matching lines @@
     }-*/;
 
     public static native Document createHTMLDocument(Document doc) /*-{
         return doc.implementation.createHTMLDocument();
     }-*/;
 
     public static native Element getFirstElementChild(Document document) /*-{
         return document.firstElementChild;
     }-*/;
 }