vgritsenko 02/02/28 06:11:42
Modified: src/java/org/apache/cocoon/components/crawler
SimpleCocoonCrawlerImpl.java
src/java/org/apache/cocoon/components/search
SimpleLuceneXMLIndexerImpl.java
Log:
Don't use jdk1.3 methods, use jdk1.2
Revision Changes Path
1.9 +15 -19
xml-cocoon2/src/java/org/apache/cocoon/components/crawler/SimpleCocoonCrawlerImpl.java
Index: SimpleCocoonCrawlerImpl.java
===================================================================
RCS file:
/home/cvs/xml-cocoon2/src/java/org/apache/cocoon/components/crawler/SimpleCocoonCrawlerImpl.java,v
retrieving revision 1.8
retrieving revision 1.9
diff -u -r1.8 -r1.9
--- SimpleCocoonCrawlerImpl.java 27 Feb 2002 05:28:38 -0000 1.8
+++ SimpleCocoonCrawlerImpl.java 28 Feb 2002 14:11:42 -0000 1.9
@@ -82,7 +82,7 @@
* A simple cocoon crawler.
*
* @author <a href="mailto:[EMAIL PROTECTED]";>Bernhard Huber</a>
- * @version CVS $Id: SimpleCocoonCrawlerImpl.java,v 1.8 2002/02/27 05:28:38
vgritsenko Exp $
+ * @version CVS $Id: SimpleCocoonCrawlerImpl.java,v 1.9 2002/02/28 14:11:42
vgritsenko Exp $
*/
public class SimpleCocoonCrawlerImpl extends AbstractLoggable
implements CocoonCrawler, Configurable, Disposable, Recyclable
@@ -118,6 +118,7 @@
* @since
*/
public final static String LINK_VIEW_QUERY_CONFIG = "link-view-query";
+
/**
* Default value of <code>link-view-query</code> configuration value.
* <p>
@@ -157,6 +158,7 @@
* @since
*/
public final static String USER_AGENT_CONFIG = "user-agent";
+
/**
* Default value of <code>user-agent</code> configuration value.
* <p>
@@ -176,6 +178,7 @@
* @since
*/
public final static String ACCEPT_CONFIG = "accept";
+
/**
* Default value of <code>accept</code> configuration value.
* <p>
@@ -445,37 +448,30 @@
*/
private List getLinks(URL url) {
ArrayList url_links = null;
+ String sURL = url.toString();
- if (getLogger().isDebugEnabled()) {
- getLogger().debug("getLinks URL " + url);
- }
-
- if (!isIncludedURL(url.toString())) {
- return null;
- }
- // don't try to get links for url which is excluded
- if (isExcludedURL(url.toString())) {
+ if (!isIncludedURL(sURL) || isExcludedURL(sURL)) {
return null;
}
// don't try to get links for url which has been crawled already
- if (crawled.contains(url.toString())) {
+ if (crawled.contains(sURL)) {
return null;
}
// mark it as crawled
- crawled.add(url.toString());
+ crawled.add(sURL);
+ // get links of url
if (getLogger().isDebugEnabled()) {
- getLogger().debug("Get links of URL: " + url.toString());
+ getLogger().debug("Getting links of URL " + sURL);
}
-
- // get links of url
try {
- URL links_url = new URL(url, url.getPath()
- + ((url.getPath().indexOf("?") == -1) ? "?" : "&")
+ sURL = url.getFile();
+ URL links = new URL(url, sURL
+ + ((sURL.indexOf("?") == -1) ? "?" : "&")
+ linkViewQuery);
- URLConnection links_url_connection = links_url.openConnection();
+ URLConnection links_url_connection = links.openConnection();
InputStream is = links_url_connection.getInputStream();
BufferedReader br = new BufferedReader(new InputStreamReader(is));
@@ -602,7 +598,7 @@
* </p>
*
* @author <a href="mailto:[EMAIL PROTECTED]>Bernhard Huber</a>
- * @version $Id: SimpleCocoonCrawlerImpl.java,v 1.8 2002/02/27 05:28:38
vgritsenko Exp $
+ * @version $Id: SimpleCocoonCrawlerImpl.java,v 1.9 2002/02/28 14:11:42
vgritsenko Exp $
*/
public static class CocoonCrawlerIterator implements Iterator
{
1.9 +3 -3
xml-cocoon2/src/java/org/apache/cocoon/components/search/SimpleLuceneXMLIndexerImpl.java
Index: SimpleLuceneXMLIndexerImpl.java
===================================================================
RCS file:
/home/cvs/xml-cocoon2/src/java/org/apache/cocoon/components/search/SimpleLuceneXMLIndexerImpl.java,v
retrieving revision 1.8
retrieving revision 1.9
diff -u -r1.8 -r1.9
--- SimpleLuceneXMLIndexerImpl.java 27 Feb 2002 05:28:38 -0000 1.8
+++ SimpleLuceneXMLIndexerImpl.java 28 Feb 2002 14:11:42 -0000 1.9
@@ -97,7 +97,7 @@
* A simple class building lucene documents from xml content.
*
* @author <a href="mailto:[EMAIL PROTECTED]";>Bernhard Huber</a>
- * @version CVS $Id: SimpleLuceneXMLIndexerImpl.java,v 1.8 2002/02/27 05:28:38
vgritsenko Exp $
+ * @version CVS $Id: SimpleLuceneXMLIndexerImpl.java,v 1.9 2002/02/28 14:11:42
vgritsenko Exp $
*/
public class SimpleLuceneXMLIndexerImpl extends AbstractLoggable
implements LuceneXMLIndexer, Configurable, Composable, ThreadSafe
@@ -170,8 +170,8 @@
throws ProcessingException {
try {
- URL contentURL = new URL(url, url.getPath()
- + ((url.getPath().indexOf("?") == -1) ? "?" : "&")
+ URL contentURL = new URL(url, url.getFile()
+ + ((url.getFile().indexOf("?") == -1) ? "?" : "&")
+ CONTENT_QUERY);
URLConnection contentURLConnection = contentURL.openConnection();
String contentType = contentURLConnection.getContentType();
----------------------------------------------------------------------
In case of troubles, e-mail: [EMAIL PROTECTED]
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
