Merge pull request yasserg#55 from MadEgg/bug-html-metarefresh

Fix bug where metaRefresh was used instead of metaLocation
hellojvm · Jul 19, 2015 · 373f09d · 373f09d
2 parents 7ed6385 + 4f6b7db
commit 373f09d
Showing 1 changed file with 3 additions and 3 deletions.
diff --git a/src/main/java/edu/uci/ics/crawler4j/parser/HtmlContentHandler.java b/src/main/java/edu/uci/ics/crawler4j/parser/HtmlContentHandler.java
@@ -124,14 +124,14 @@ public void startElement(String uri, String localName, String qName, Attributes
           int pos = content.toLowerCase().indexOf("url=");
           if (pos != -1) {
             metaRefresh = content.substring(pos + 4);
+            addToOutgoingUrls(metaRefresh, localName);
           }
-          addToOutgoingUrls(metaRefresh, localName);
         }
 
         // http-equiv="location" content="http://foo.bar/..."
         if ("location".equals(equiv) && (metaLocation == null)) {
           metaLocation = content;
-          addToOutgoingUrls(metaRefresh, localName);
+          addToOutgoingUrls(metaLocation, localName);
         }
       }
     } else if (element == Element.BODY) {
@@ -194,4 +194,4 @@ public String getBaseUrl() {
   public Map<String, String> getMetaTags() {
     return metaTags;
   }
-}
+}