NUTCH-2404 Fix for Failed Jenkin build #1588 after merging pull request

apache#192 (NUTCH-2389).
kaidul · Jul 31, 2017 · a6870de · a6870de
1 parent 5f6c383
commit a6870de
Show file tree

Hide file tree

Showing 2 changed files with 6 additions and 11 deletions.
diff --git a/src/plugin/jsoup-extractor/build.xml b/src/plugin/jsoup-extractor/build.xml
@@ -20,9 +20,4 @@
 
   <import file="../build-plugin.xml"/>
 
-  <!-- Deploy Unit test dependencies -->
-  <target name="deploy-test">
-    <ant target="deploy" inheritall="false" dir="../nutch-extensionpoints"/>
-  </target>
-
 </project>
diff --git a/.../jsoup-extractor/src/test/org/apache/nutch/parse/jsoup/extractor/TestJsoupHtmlParser.java b/.../jsoup-extractor/src/test/org/apache/nutch/parse/jsoup/extractor/TestJsoupHtmlParser.java
@@ -27,11 +27,11 @@
 import java.net.MalformedURLException;
 import java.net.URL;
 import java.nio.ByteBuffer;
+import java.nio.charset.StandardCharsets;
 import java.util.Map.Entry;
 
 import org.apache.avro.util.Utf8;
 import org.apache.hadoop.conf.Configuration;
-import org.apache.hadoop.hbase.util.Bytes;
 import org.apache.nutch.core.jsoup.extractor.JsoupExtractorConstants;
 import org.apache.nutch.parse.ParseException;
 import org.apache.nutch.parse.ParseUtil;
@@ -43,8 +43,8 @@ public class TestJsoupHtmlParser {
 
   private static final String SAMPLE_CONF_FILE = "jsoup-extractor-example.xml";
   private static final String SAMPLE_URL = "https://www.youtube.com/watch?v=pzMpwW4ppRM";
-  private static final String TITLE = "Large scale crawling with Apache Nutch\t";
-  private static final String PUBLISHER = "LuceneSolrRevolution\t";
+  private static final String TITLE = "Large scale crawling with Apache Nutch";
+  private static final String PUBLISHER = "LuceneSolrRevolution";
 
   @Test
   public void testJsoupHtmlParser() {
@@ -70,11 +70,11 @@ public void testJsoupHtmlParser() {
       parser.parse(SAMPLE_URL, page);
 
       for(Entry<CharSequence, ByteBuffer> entry: page.getMetadata().entrySet()) {
-        System.out.println(entry.getKey().toString() + " => " + Bytes.toString(entry.getValue().array()));
+         System.out.println(entry.getKey().toString() + " => " + new String(entry.getValue().array(), StandardCharsets.UTF_8));
       }
 
-      assertEquals(Bytes.toString(page.getMetadata().get(new Utf8("title")).array()), TITLE);
-      assertEquals(Bytes.toString(page.getMetadata().get(new Utf8("publisherName")).array()), PUBLISHER);
+      assertEquals(new String(page.getMetadata().get(new Utf8("title")).array(), StandardCharsets.UTF_8).replaceAll("\\t", ""), TITLE);
+      assertEquals(new String(page.getMetadata().get(new Utf8("publisherName")).array(), StandardCharsets.UTF_8).replaceAll("\\t", ""), PUBLISHER);
 
     } catch (MalformedURLException ex) {
       ex.printStackTrace();