gwt getInnerText -> javascript innerText or textContent

src/com/dom_distiller/client/PagingLinksFinder.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.
 
 /*
  * Parts of this file are adapted from Readability.
  *
  * Readability is Copyright (c) 2010 Src90 Inc
  * and licenced under the Apache License, Version 2.0.
  */
@@ skipping 115 matching lines @@
             }
 
             // If it's on a different domain, skip it.
             String[] urlSlashes = StringUtil.split(linkHref, "\\/+");
             if (urlSlashes.length < 3 ||  // Expect at least the protocol, domain, and path.
                     !Window.Location.getHost().equalsIgnoreCase(urlSlashes[1])) {
                 appendDbgStrForLink(link, "ignored: different domain");
                 continue;
             }
 
-            String linkText = link.getInnerText();
+            // Use javascript innerText (instead of javascript textContent) to only get visible
+            // text.
+            String linkText = DomUtil.getInnerText(link);
 
             // If the linkText looks like it's not the next or previous page, skip it.
             if (StringUtil.match(linkText, EXTRANEOUS_REGEX) || linkText.length() > 25) {
                 appendDbgStrForLink(link, "ignored: one of extra");
                 continue;
             }
 
             // For next page link, if the initial part of the URL is identical to the base URL, but
             // the rest of it doesn't contain any digits, it's certainly not a next page link.
             // However, this doesn't apply to previous page link, because most sites will just have
@@ skipping 262 matching lines @@
         // (TODO)kuan): investigate how to get logging when running "ant test.prod" - currently,
         // nothing appears.  In the meantime, throwing an exception with a log message at suspicious
         // codepoints can produce a call stack and help debugging, albeit tediously.
         LogUtil.logToConsole("numLinks=" + allLinks.getLength() + ", found " +
                 (pageLink == PageLink.NEXT ? "next: " : "prev: ") +
                 (pagingHref != null ? pagingHref : "null"));
 
         for (int i = 0; i < allLinks.getLength(); i++) {
             AnchorElement link = AnchorElement.as(allLinks.getItem(i));
 
-            String text = link.getInnerText();
+            // Use javascript innerText (instead of javascript textContent) to get only visible
+            // text.
+            String text = DomUtil.getInnerText(link);
             // Trim unnecessary whitespaces from text.
             String[] words = StringUtil.split(text, "\\s+");
             text = "";
             for (int w = 0; w < words.length; w++) {
                 text += words[w];
                 if (w < words.length - 1) text += " ";
             }
 
             LogUtil.logToConsole(i + ")" + link.getHref() + ", txt=[" + text + "], dbg=[" +
                     mLinkDebugInfo.get(link) + "]");
@@ skipping 15 matching lines @@
     }
 
     private enum PageLink {
         NEXT,
         PREV,
     }
 
     private static final Map<Element, String> mLinkDebugInfo = new HashMap<Element, String>();
 
 }