Retain image sizes

javatests/org/chromium/distiller/ContentExtractorTest.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.dom.client.Document;
 import com.google.gwt.dom.client.Element;
 
 public class ContentExtractorTest extends DomDistillerJsTestCase {
@@ skipping 86 matching lines @@
             "<img id=\"b\" style=\"align: left\" alt=\"b\" data-dummy=\"c\" src=\"image2\">" +
             "<table role=\"grid\"><tbody><tr><td>" +
                 "<img id=\"c\" style=\"a\" alt=\"b\" src=\"/image\" srcset=\"https://example.com/image2x 2x, /image4x 4x,\">" +
                 "<img id=\"d\" style=\"a\" align=\"left\" src=\"/image2\">" +
                 "</td></tr></tbody></table>" +
             "<p>" + CONTENT_TEXT + "</p>";
 
         final String expected =
             "<h1>" + CONTENT_TEXT + "</h1>" +
             "<img src=\"http://example.com/path/image\" " +
-                 "srcset=\"http://example.com/path/image200 200w, http://example.org/image400 400w\">" +
-            "<img alt=\"b\" src=\"http://example.com/path/image2\">" +
+                 "srcset=\"http://example.com/path/image200 200w, http://example.org/image400 400w\" " +
+                    "width=\"0\" height=\"0\">" +

[wychen, 2016/03/11 08:25:35]

If the dimension is not specified in the source, we might want to keep it that
way.

+            "<img alt=\"b\" src=\"http://example.com/path/image2\" " +
+                    "width=\"0\" height=\"0\">" +
             "<table role=\"grid\"><tbody><tr><td>" +
                 "<img alt=\"b\" src=\"http://example.com/image\" " +
                      "srcset=\"https://example.com/image2x 2x, http://example.com/image4x 4x, \">" +
                 "<img src=\"http://example.com/image2\">" +
             "</td></tr></tbody></table>" +
             "<p>" + CONTENT_TEXT + "</p>";
 
         mHead.setInnerHTML("<base href=\"http://example.com/path/\">");
         mBody.setInnerHTML(html);
 
@@ skipping 53 matching lines @@
                 "</tbody>" +
             "</table>";
 
         final String expected =
             "<h1>" +
                 CONTENT_TEXT +
             "</h1>" +
             "<p>" +
                 CONTENT_TEXT +
             "</p>" +
-            "<img src=\"http://example.com/test.png\">" +
+            "<img src=\"http://example.com/test.png\" " +
+                    "width=\"0\" height=\"0\">" +
             "<table>" +
                 "<tbody>" +
                     "<tr>" +
                         "<th>" + CONTENT_TEXT +
                             "<img src=\"http://example.com/test.png\">" +
                         "</th>" +
                         "<th>" + CONTENT_TEXT + "</th>" +
                     "</tr><tr>" +
                         "<td>" + CONTENT_TEXT + "</td>" +
                         "<td>" + CONTENT_TEXT + "</td>" +
                     "</tr>" +
                 "</tbody>" +
             "</table>";
 
         mHead.setInnerHTML("<base href=\"http://example.com/\">");
         mBody.setInnerHTML(html);
 
         ContentExtractor extractor = new ContentExtractor(mRoot);
         String extractedContent = extractor.extractContent();
         assertEquals(expected,
                 TestUtil.removeAllDirAttributes(extractedContent));
     }
 
+    public void testKeepingWidthAndHeightAttributes() {
+        String html =
+            "<h1 style=\"font-weight: folder\">" +

[wychen, 2016/03/11 08:25:35]

The padding text can be simplified. We don't need the style here. It is to test
style removal. You can refer to testImage().

[dalmirsilva, 2016/03/14 18:28:06]

Done.

+                CONTENT_TEXT +
+            "</h1>" +
+            "<p style=\"\">" +
+                CONTENT_TEXT +
+            "</p>" +
+            "<img style=\"align: left\" src=\"/test.png\" " +
+                    "width=\"200\" height=\"300\">";
+
+        final String expected =
+            "<h1>" +
+                CONTENT_TEXT +
+            "</h1>" +
+            "<p>" +
+                CONTENT_TEXT +
+            "</p>" +
+            "<img src=\"http://example.com/test.png\" " +
+                    "width=\"200\" height=\"300\">";

[wychen, 2016/03/11 08:25:35]

Maybe another img with only width, and one with neither width and height?

[dalmirsilva, 2016/03/14 18:28:06]

Done.

+
+        mHead.setInnerHTML("<base href=\"http://example.com/\">");
+        mBody.setInnerHTML(html);
+
+        ContentExtractor extractor = new ContentExtractor(mRoot);
+        String extractedContent = extractor.extractContent();
+        assertEquals(expected,
+                TestUtil.removeAllDirAttributes(extractedContent));
+    }
+
     public void testPreserveOrderedList() {
         Element outerListTag = Document.get().createElement("OL");
         mBody.appendChild(outerListTag);
 
         outerListTag.appendChild(TestUtil.createListItem(CONTENT_TEXT));
         outerListTag.appendChild(TestUtil.createListItem(CONTENT_TEXT));
         outerListTag.appendChild(TestUtil.createListItem(CONTENT_TEXT));
         outerListTag.appendChild(TestUtil.createListItem(CONTENT_TEXT));
 
         ContentExtractor extractor = new ContentExtractor(mRoot);
@@ skipping 382 matching lines @@
 
         final String htmlArticle =
             "<h1>" + CONTENT_TEXT + "</h1>" +
             "<div itemscope itemtype=\"http://schema.org/Movie\">" + article + "</div>";
         final String expected = "<h1>" + CONTENT_TEXT + "</h1>" + article;
 
         // Non-article schema.org types should not use the fast path.
         assertExtractor(expected, htmlArticle);
     }
 }