You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@lenya.apache.org by mi...@apache.org on 2005/04/27 17:12:59 UTC
svn commit: r164991 -
/lenya/branches/BRANCH_1_2_X/src/java/org/apache/lenya/search/crawler/CrawlerConfiguration.java
Author: michi
Date: Wed Apr 27 08:12:58 2005
New Revision: 164991
URL: http://svn.apache.org/viewcvs?rev=164991&view=rev
Log:
show configuration of crawler fixed
Modified:
lenya/branches/BRANCH_1_2_X/src/java/org/apache/lenya/search/crawler/CrawlerConfiguration.java
Modified: lenya/branches/BRANCH_1_2_X/src/java/org/apache/lenya/search/crawler/CrawlerConfiguration.java
URL: http://svn.apache.org/viewcvs/lenya/branches/BRANCH_1_2_X/src/java/org/apache/lenya/search/crawler/CrawlerConfiguration.java?rev=164991&r1=164990&r2=164991&view=diff
==============================================================================
--- lenya/branches/BRANCH_1_2_X/src/java/org/apache/lenya/search/crawler/CrawlerConfiguration.java (original)
+++ lenya/branches/BRANCH_1_2_X/src/java/org/apache/lenya/search/crawler/CrawlerConfiguration.java Wed Apr 27 08:12:58 2005
@@ -15,7 +15,7 @@
*
*/
-/* $Id: CrawlerConfiguration.java,v 1.9 2004/03/05 11:00:06 michi Exp $ */
+/* $Id$ */
package org.apache.lenya.search.crawler;
@@ -31,7 +31,7 @@
/**
- * DOCUMENT ME!
+ * Web-Crawler (it might make sense to replace this by Nutch)
*/
public class CrawlerConfiguration {
static Category log = Category.getInstance(CrawlerConfiguration.class);
@@ -97,31 +97,28 @@
}
} else {
parameter = ce.getBaseURL();
- System.out.println(parameter);
+ System.out.println("Crawler Config: Base URL: " + parameter);
parameter = ce.getScopeURL();
- System.out.println(parameter);
+ System.out.println("Crawler Config: Scope URL: " + parameter);
parameter = ce.getUserAgent();
- System.out.println(parameter);
+ System.out.println("Crawler Config: User Agent: " + parameter);
parameter = ce.getURIList();
- System.out.println(parameter);
- System.out.println(ce.resolvePath(parameter));
+ System.out.println("Crawler Config: URI List: " + ce.resolvePath(parameter) + " (" + parameter + ")");
parameter = ce.getHTDocsDumpDir();
- System.out.println("htdocs-dump-dir/@src: " + parameter);
- System.out.println(ce.resolvePath(parameter));
+ System.out.println("Crawler Config: HTDocs Dump Dir: " + ce.resolvePath(parameter) + " (" + parameter + ")");
parameter = ce.getRobotsFile();
if (parameter != null) {
- System.out.println("robots/@src: " + parameter);
- System.out.println(ce.resolvePath(parameter));
+ System.out.println("Crawler Config: Robots File: " + ce.resolvePath(parameter + " (" + parameter + ")"));
}
parameter = ce.getRobotsDomain();
if (parameter != null) {
- System.out.println("robots/@domain: " + parameter);
+ System.out.println("Crawler Config: Robots Domain: " + parameter);
}
}
}
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@lenya.apache.org
For additional commands, e-mail: commits-help@lenya.apache.org