You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@nutch.apache.org by ma...@apache.org on 2011/04/12 14:51:16 UTC
svn commit: r1091390 - in /nutch/branches/branch-1.3: CHANGES.txt
src/plugin/subcollection/src/java/org/apache/nutch/collection/Subcollection.java
Author: markus
Date: Tue Apr 12 12:51:16 2011
New Revision: 1091390
URL: http://svn.apache.org/viewvc?rev=1091390&view=rev
Log:
NUTCH-891 Subcollection plugin won't require blacklist any more (markus)
Modified:
nutch/branches/branch-1.3/CHANGES.txt
nutch/branches/branch-1.3/src/plugin/subcollection/src/java/org/apache/nutch/collection/Subcollection.java
Modified: nutch/branches/branch-1.3/CHANGES.txt
URL: http://svn.apache.org/viewvc/nutch/branches/branch-1.3/CHANGES.txt?rev=1091390&r1=1091389&r2=1091390&view=diff
==============================================================================
--- nutch/branches/branch-1.3/CHANGES.txt (original)
+++ nutch/branches/branch-1.3/CHANGES.txt Tue Apr 12 12:51:16 2011
@@ -2,6 +2,8 @@ Nutch Change Log
Release 1.3 - Current Development
+* NUTCH-891 Subcollection plugin won't require blacklist any more (markus)
+
* NUTCH-972 CrawlDbMerger doesn't break on non-existent input (Gabriele Kahlout via jnioche)
* NUTCH-967 Upgrade to Tika 0.9 (jnioche)
Modified: nutch/branches/branch-1.3/src/plugin/subcollection/src/java/org/apache/nutch/collection/Subcollection.java
URL: http://svn.apache.org/viewvc/nutch/branches/branch-1.3/src/plugin/subcollection/src/java/org/apache/nutch/collection/Subcollection.java?rev=1091390&r1=1091389&r2=1091390&view=diff
==============================================================================
--- nutch/branches/branch-1.3/src/plugin/subcollection/src/java/org/apache/nutch/collection/Subcollection.java (original)
+++ nutch/branches/branch-1.3/src/plugin/subcollection/src/java/org/apache/nutch/collection/Subcollection.java Tue Apr 12 12:51:16 2011
@@ -25,6 +25,7 @@ import org.apache.hadoop.conf.Configured
import org.apache.nutch.net.URLFilter;
import org.apache.xerces.util.DOMUtil;
import org.w3c.dom.Element;
+import org.w3c.dom.NodeList;
/**
* SubCollection represents a subset of index, you can define url patterns that
@@ -170,11 +171,15 @@ public class Subcollection extends Confi
collection.getElementsByTagName(TAG_NAME).item(0)).trim();
this.wlString = DOMUtil.getChildText(
collection.getElementsByTagName(TAG_WHITELIST).item(0)).trim();
- this.blString = DOMUtil.getChildText(
- collection.getElementsByTagName(TAG_BLACKLIST).item(0)).trim();
parseList(this.whiteList, wlString);
- parseList(this.blackList, blString);
+
+ // Check if there's a blacklist we need to parse
+ NodeList nodeList = collection.getElementsByTagName(TAG_BLACKLIST);
+ if (nodeList.getLength() > 0) {
+ this.blString = DOMUtil.getChildText(nodeList.item(0)).trim();
+ parseList(this.blackList, blString);
+ }
}
/**