Author: jerome
Date: Thu Feb 16 02:41:49 2006
New Revision: 378219
URL: http://svn.apache.org/viewcvs?rev=378219&view=rev
Log:
NUTCH-196 : Add neko html library (lib-nekohtml)
Added:
lucene/nutch/trunk/src/plugin/lib-nekohtml/
lucene/nutch/trunk/src/plugin/lib-nekohtml/build.xml (with props)
lucene/nutch/trunk/src/plugin/lib-nekohtml/lib/
lucene/nutch/trunk/src/plugin/lib-nekohtml/lib/nekohtml-0.9.4.jar (with
props)
lucene/nutch/trunk/src/plugin/lib-nekohtml/plugin.xml (with props)
Removed:
lucene/nutch/trunk/src/plugin/clustering-carrot2/lib/nekohtml-0.9.2.jar
lucene/nutch/trunk/src/plugin/parse-html/lib/nekohtml-0.9.4.LICENSE.txt
lucene/nutch/trunk/src/plugin/parse-html/lib/nekohtml-0.9.4.jar
Modified:
lucene/nutch/trunk/src/plugin/clustering-carrot2/build.xml
lucene/nutch/trunk/src/plugin/clustering-carrot2/plugin.xml
lucene/nutch/trunk/src/plugin/parse-html/build.xml
lucene/nutch/trunk/src/plugin/parse-html/plugin.xml
Modified: lucene/nutch/trunk/src/plugin/clustering-carrot2/build.xml
URL:
http://svn.apache.org/viewcvs/lucene/nutch/trunk/src/plugin/clustering-carrot2/build.xml?rev=378219&r1=378218&r2=378219&view=diff
==============================================================================
--- lucene/nutch/trunk/src/plugin/clustering-carrot2/build.xml (original)
+++ lucene/nutch/trunk/src/plugin/clustering-carrot2/build.xml Thu Feb 16
02:41:49 2006
@@ -7,7 +7,7 @@
<path id="plugin.deps">
<fileset dir="${nutch.root}/build">
<include name="**/lib-log4j/*.jar" />
- <include name="**/lib-commons-httpclient/*.jar" />
+ <include name="**/lib-nekohtml/*.jar" />
</fileset>
</path>
Modified: lucene/nutch/trunk/src/plugin/clustering-carrot2/plugin.xml
URL:
http://svn.apache.org/viewcvs/lucene/nutch/trunk/src/plugin/clustering-carrot2/plugin.xml?rev=378219&r1=378218&r2=378219&view=diff
==============================================================================
--- lucene/nutch/trunk/src/plugin/clustering-carrot2/plugin.xml (original)
+++ lucene/nutch/trunk/src/plugin/clustering-carrot2/plugin.xml Thu Feb 16
02:41:49 2006
@@ -22,15 +22,12 @@
<library name="commons-pool-1.1.jar"/>
<library name="FSA.jar"/>
<library name="Jama-1.0.1-patched.jar"/>
- <library name="log4j-1.2.8.jar"/>
-
- <library name="nekohtml-0.9.2.jar"/>
</runtime>
<requires>
<import plugin="nutch-extensionpoints"/>
<import plugin="lib-log4j"/>
- <import plugin="lib-commons-httpclient"/>
+ <import plugin="lib-nekohtml"/>
</requires>
<extension id="org.apache.nutch.clustering.carrot2"
Added: lucene/nutch/trunk/src/plugin/lib-nekohtml/build.xml
URL:
http://svn.apache.org/viewcvs/lucene/nutch/trunk/src/plugin/lib-nekohtml/build.xml?rev=378219&view=auto
==============================================================================
--- lucene/nutch/trunk/src/plugin/lib-nekohtml/build.xml (added)
+++ lucene/nutch/trunk/src/plugin/lib-nekohtml/build.xml Thu Feb 16 02:41:49
2006
@@ -0,0 +1,21 @@
+<?xml version="1.0"?>
+
+<project name="lib-nekohtml" default="jar">
+
+ <import file="../build-plugin.xml"/>
+
+ <!--
+ ! Override the compile and jar targets,
+ ! since there is nothing to compile here.
+ ! -->
+ <target name="compile" depends="init">
+ <echo message="Compiling plugin: ${name}"/>
+ </target>
+
+ <target name="jar" depends="compile">
+ <copy todir="${build.dir}" verbose="true">
+ <fileset dir="./lib" includes="**/*.jar"/>
+ </copy>
+ </target>
+
+</project>
Propchange: lucene/nutch/trunk/src/plugin/lib-nekohtml/build.xml
------------------------------------------------------------------------------
svn:eol-style = native
Added: lucene/nutch/trunk/src/plugin/lib-nekohtml/lib/nekohtml-0.9.4.jar
URL:
http://svn.apache.org/viewcvs/lucene/nutch/trunk/src/plugin/lib-nekohtml/lib/nekohtml-0.9.4.jar?rev=378219&view=auto
==============================================================================
Binary file - no diff available.
Propchange: lucene/nutch/trunk/src/plugin/lib-nekohtml/lib/nekohtml-0.9.4.jar
------------------------------------------------------------------------------
svn:mime-type = application/octet-stream
Added: lucene/nutch/trunk/src/plugin/lib-nekohtml/plugin.xml
URL:
http://svn.apache.org/viewcvs/lucene/nutch/trunk/src/plugin/lib-nekohtml/plugin.xml?rev=378219&view=auto
==============================================================================
--- lucene/nutch/trunk/src/plugin/lib-nekohtml/plugin.xml (added)
+++ lucene/nutch/trunk/src/plugin/lib-nekohtml/plugin.xml Thu Feb 16 02:41:49
2006
@@ -0,0 +1,21 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<!--
+ ! NekoHTML is a simple HTML scanner and tag balancer.
+ ! (http://people.apache.org/~andyc/neko/doc/html/index.html)
+ !
+ ! Dowload : http://people.apache.org/~andyc/neko/doc/html/index.html
+ ! License : http://people.apache.org/~andyc/neko/LICENSE
+ !-->
+<plugin
+ id="lib-nekohtml"
+ name="CyberNeko HTML Parser"
+ version="0.9.4"
+ provider-name="org.cyberneko">
+
+ <runtime>
+ <library name="nekohtml-0.9.4.jar">
+ <export name="*"/>
+ </library>
+ </runtime>
+
+</plugin>
Propchange: lucene/nutch/trunk/src/plugin/lib-nekohtml/plugin.xml
------------------------------------------------------------------------------
svn:eol-style = native
Modified: lucene/nutch/trunk/src/plugin/parse-html/build.xml
URL:
http://svn.apache.org/viewcvs/lucene/nutch/trunk/src/plugin/parse-html/build.xml?rev=378219&r1=378218&r2=378219&view=diff
==============================================================================
--- lucene/nutch/trunk/src/plugin/parse-html/build.xml (original)
+++ lucene/nutch/trunk/src/plugin/parse-html/build.xml Thu Feb 16 02:41:49 2006
@@ -4,4 +4,10 @@
<import file="../build-plugin.xml"/>
+ <path id="plugin.deps">
+ <fileset dir="${nutch.root}/build">
+ <include name="**/lib-nekohtml/*.jar" />
+ </fileset>
+ </path>
+
</project>
Modified: lucene/nutch/trunk/src/plugin/parse-html/plugin.xml
URL:
http://svn.apache.org/viewcvs/lucene/nutch/trunk/src/plugin/parse-html/plugin.xml?rev=378219&r1=378218&r2=378219&view=diff
==============================================================================
--- lucene/nutch/trunk/src/plugin/parse-html/plugin.xml (original)
+++ lucene/nutch/trunk/src/plugin/parse-html/plugin.xml Thu Feb 16 02:41:49 2006
@@ -9,12 +9,12 @@
<library name="parse-html.jar">
<export name="*"/>
</library>
- <library name="nekohtml-0.9.4.jar"/>
<library name="tagsoup-1.0rc3.jar"/>
</runtime>
<requires>
<import plugin="nutch-extensionpoints"/>
+ <import plugin="lib-nekohtml"/>
</requires>
<extension id="org.apache.nutch.parse.html"
-------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc. Do you grep through log files
for problems? Stop! Download the new AJAX search engine that makes
searching your log files as easy as surfing the web. DOWNLOAD SPLUNK!
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=103432&bid=230486&dat=121642
_______________________________________________
Nutch-cvs mailing list
Nutch-cvs@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/nutch-cvs
