Menu
▾
▴
[Archive-access-cvs] SF.net SVN: archive-access: [1623] trunk/archive-access/projects/nutchwax
|
From: <sta...@us...> - 2007-03-21 19:14:49
|
Revision: 1623
http://archive-access.svn.sourceforge.net/archive-access/?rev=1623&view=rev
Author: stack-sf
Date: 2007-03-21 12:13:23 -0700 (Wed, 21 Mar 2007)
Log Message:
-----------
M nutchwax/src/main/assembly/distribution.xml
Change where we fetch job and war jars from.
M nutchwax/nutchwax-thirdparty/pom.xml
D nutchwax/build.xml
Removed build.xml. Remove misimpression that
you can use ant to build this project.
Call third.party build.xml directly instead.
Modified Paths:
--------------
trunk/archive-access/projects/nutchwax/nutchwax-thirdparty/pom.xml
trunk/archive-access/projects/nutchwax/src/main/assembly/distribution.xml
Removed Paths:
-------------
trunk/archive-access/projects/nutchwax/build.xml
Deleted: trunk/archive-access/projects/nutchwax/build.xml
===================================================================
--- trunk/archive-access/projects/nutchwax/build.xml 2007-03-21 17:06:07 UTC (rev 1622)
+++ trunk/archive-access/projects/nutchwax/build.xml 2007-03-21 19:13:23 UTC (rev 1623)
@@ -1,342 +0,0 @@
-<?xml version="1.0"?>
-
-<project name="nutchwax" default="all">
- <property name="name" value="${ant.project.name}"/>
- <property name="root" value="${basedir}"/>
-
- <!--'nutch.root' is pointer at core nutch. Expect to find it in
- '${basedir}/third-party' named 'nutch'.
- -->
- <!--Keep this aligned with whats in maven2 pom-->
- <property name="nutchwax.version" value="-0.11.0-SNAPSHOT"/>
- <property name="nutch.root" location="${root}/third-party/nutch"/>
-
- <property file="${user.home}/.$(name}.build.properties" />
-
- <property name="src.dir" location="${root}/src/java"/>
- <property name="src.test" location="${root}/src/test"/>
-
- <available file="${src.test}" type="dir" property="test.available"/>
-
- <property name="conf.dir" location="${root}/conf"/>
-
- <property name="build.dir" location="${root}/target"/>
- <property name="build.classes" location="${build.dir}/classes"/>
- <property name="build.test" location="${build.dir}/test"/>
-
- <property name="build.plugins" location="${nutch.root}/build/plugins"/>
- <property name="deploy.dir" location="${build.plugins}/${name}"/>
-
- <property name="this.web" location="${root}/src/web"/>
- <property name="nutch.web" location="${nutch.root}/src/web"/>
-
- <property name="javac.deprecation" value="off"/>
- <property name="javac.debug" value="on"/>
-
- <property name="javadoc.link"
- value="http://java.sun.com/j2se/1.5/docs/api/"/>
-
- <property name="build.encoding" value="ISO-8859-1"/>
-
- <!-- the normal classpath -->
- <path id="classpath">
- <pathelement location="${build.classes}"/>
- <pathelement location="${nutch.root}/build/classes"/>
- <fileset dir="${nutch.root}/lib">
- <include name="*.jar" />
- </fileset>
- </path>
-
- <!-- the unit test classpath -->
- <path id="test.classpath">
- <pathelement location="${build.test}" />
- <pathelement location="${conf.dir}"/>
- <pathelement location="${nutch.root}/conf"/>
- <pathelement location="${nutch.root}/build"/>
- <path refid="classpath"/>
- </path>
-
- <target name="third.party.plugins" description="Build third-party plugins">
- <echo message="Building nutch third-party dependency (plugins)" />
- <ant dir="third-party/nutch" target="compile-plugins" inheritAll="false" >
- <property name="build.compiler" value="extJavac" />
- </ant>
- </target>
- <target name="third.party.compile" description="Compile third-party src">
- <echo message="Building nutch third-party dependency (compile)" />
- <ant dir="third-party/nutch" target="compile" inheritAll="false" >
- <property name="build.compiler" value="extJavac" />
- </ant>
- </target>
- <target name="third.party.jar" description="Build third-party jars" >
- <echo message="Building nutch third-party dependency (jar)" />
- <ant dir="third-party/nutch" target="jar" inheritAll="false" >
- <property name="build.compiler" value="extJavac" />
- </ant>
- </target>
- <target name="third.party.war" description="Build third-party wars" >
- <echo message="Building nutch third-party dependency (war)" />
- <ant dir="third-party/nutch" target="war" inheritAll="false" >
- <property name="build.compiler" value="extJavac" />
- </ant>
- </target>
- <target name="third.party.clean" description="Clean third-party software">
- <echo message="Cleaning nutch third-party dependency" />
- <ant dir="third-party/nutch" target="clean" inheritAll="false" >
- <property name="build.compiler" value="extJavac" />
- </ant>
- </target>
-
- <!-- ====================================================== -->
- <!-- Stuff needed by all targets -->
- <!-- ====================================================== -->
- <target name="init">
- <mkdir dir="${build.dir}"/>
- <mkdir dir="${build.classes}"/>
- <mkdir dir="${build.test}"/>
- </target>
-
- <!-- ====================================================== -->
- <!-- Compile the Java files -->
- <!-- ====================================================== -->
- <target name="compile" depends="init"
- description="Compile nutchwax classes">
- <property name="build.compiler" value="extJavac" />
- <javac
- encoding="${build.encoding}"
- srcdir="${src.dir}"
- includes="**/*.java"
- destdir="${build.classes}"
- debug="${javac.debug}"
- target="1.5"
- source="1.5"
- deprecation="${javac.deprecation}">
- <classpath refid="classpath"/>
- </javac>
- </target>
-
- <!-- ====================================================== -->
- <!-- Compile plugins -->
- <!-- ====================================================== -->
- <target name="compile-plugins"
- description="Compile all nutchwax plugins">
- <ant dir="src/plugin" target="deploy" inheritAll="false">
- <property name="build.compiler" value="extJavac" />
- </ant>
- </target>
-
- <!-- ================================================================== -->
- <!-- Make job jar -->
- <!-- ================================================================== -->
- <!-- -->
- <!-- ================================================================== -->
- <target name="jar" depends="compile, compile-plugins"
- description="Builds nutchwax jobs jar of all tasks to do import, etc." >
- <zip destfile="${build.dir}/${name}-job${nutchwax.version}.jar">
- <zipfileset prefix="META-INF" file="${conf.dir}/MANIFEST.MF"/>
- <zipfileset file="${conf.dir}/log4j.properties"/>
- <zipfileset file="${conf.dir}/wax-parse-plugins.xml"/>
- <zipfileset file="${conf.dir}/wax-default.xml"/>
- <zipfileset file="${conf.dir}/regex-normalize.xml"/>
- <zipfileset file="${conf.dir}/regex-urlfilter.txt"/>
- <zipfileset file="${nutch.root}/conf/mime-types.xml"/>
- <zipfileset file="${nutch.root}/conf/nutch-default.xml"/>
- <zipfileset file="${nutch.root}/conf/common-terms.utf8"/>
- <zipfileset prefix="bin" file="${basedir}/src/plugin/parse-waxext/bin/parse-pdf.sh" filemode="555"/>
- <!--<zipfileset refid="lib.jars"/>
- -->
-
- <!--Include all class files both nutch and nutchwax at top level
- so all needed to launch a job using the 'hadoop jar nutchwax.jobs'
- is on the classpath (Only classes that are at top-level in a jar can
- be found on CLASSPATH. Jars inside jars or classes under 'classes'
- directory cannot be found or added to CLASSPATH, not without custom
- classloader: See
- http://java.sun.com/docs/books/tutorial/deployment/jar/downman.html).
- -->
- <zipfileset dir="${build.dir}/classes" />
- <zipfileset dir="${nutch.root}/build/classes" />
- <!-- Be selective about which plugins to copy over. Otherwise
- the jar gets massive (16Megs with all plugins. 10Megs not
- including plugins used at other than indexing time).
-
- Include query-time filters for case where we're running in
- distributed mode.
-
- Note, we EXCLUDE parse-js. Otherwise, its run as part of
- html parse. We don't want this because the parse-js currently
- adds base url as anchor text polluting the linkdb and its kinda
- messy regards URLs it finds in javascript. It needs some work.
- Meantime, we'll do w/o the URLs it finds in linkdb. See
- NUTCH-425 and
- http://sourceforge.net/tracker/index.php?func=detail&aid=1591709&group_id=118427&atid=681137
- -->
- <zipfileset prefix="plugins" dir="${nutch.root}/build/plugins">
- <!-- See above why we exclude parse-js-->
- <exclude name="parse-js/**" />
- <include name="analysis-*/**" />
- <include name="index-*/**" />
- <include name="language-*/**" />
- <include name="lib-*/**" />
- <include name="nutch-*/**" />
- <include name="scoring-*/**" />
- <include name="query-*/**" />
- <include name="summary-*/**" />
- <include name="urlfilter-*/**" />
- <include name="urlnormalizer-*/**" />
- <include name="parse-*/**" />
- </zipfileset>
- <!--Add wax plugins-->
- <zipfileset prefix="wax-plugins" dir="${build.dir}/wax-plugins">
- <include name="*/**" />
- </zipfileset>
- <!--Include nutch dependencies in job jar. -->
- <zipfileset prefix="lib" file="${nutch.root}/lib/commons-lang*jar"/>
- <zipfileset prefix="lib" file="${nutch.root}/lib/lucene*jar"/>
- <zipfileset prefix="lib" file="${nutch.root}/lib/jakarta-oro*jar"/>
- <zipfileset prefix="lib" file="${nutch.root}/lib/xerces*jar"/>
- <zipfileset prefix="lib" file="${nutch.root}/lib//concurrent-1.3.4.jar"/>
- <!--Finally, include the README.txt file so can tell what
- hadoop and nutch this was built against-->
- <zipfileset file="${root}/README.txt"/>
- </zip>
- </target>
-
- <!-- ================================================================== -->
- <!-- Build all including third-party dependencies (i.e. nutch) -->
- <!-- ================================================================== -->
- <!-- -->
- <!-- ================================================================== -->
- <target name="all" depends="third.party.jar,third.party.war,jar,compile,war" />
-
- <!-- ================================================================== -->
- <!-- Compile test code -->
- <!-- ================================================================== -->
- <target name="compile-test" depends="compile" if="test.available">
- <javac
- encoding="${build.encoding}"
- srcdir="${src.test}"
- includes="**/*.java"
- destdir="${build.test}"
- debug="${debug}">
- <classpath refid="test.classpath"/>
- </javac>
- </target>
-
- <!-- ================================================================== -->
- <!-- Run unit tests -->
- <!-- ================================================================== -->
- <target name="test" depends="compile-test" if="test.available"
- description="Run tests">
-
- <junit printsummary="yes" haltonfailure="no" fork="yes"
- errorProperty="tests.failed" failureProperty="tests.failed">
- <sysproperty key="test.data" value="${build.test}/data"/>
- <sysproperty key="test.input" value="${root}/data"/>
- <classpath refid="test.classpath"/>
- <formatter type="plain" />
- <batchtest todir="${build.test}" unless="testcase">
- <fileset dir="${src.test}"
- includes="**/Test*.java" excludes="**/${test.exclude}.java" />
- </batchtest>
- <batchtest todir="${build.test}" if="testcase">
- <fileset dir="${src.test}" includes="**/${testcase}.java"/>
- </batchtest>
- </junit>
-
- <fail if="tests.failed">Tests failed!</fail>
-
- </target>
-
- <!-- ================================================================== -->
- <!-- build war file -->
- <!-- ================================================================== -->
- <target name="war" depends="compile, compile-plugins"
- description="Builds nutchwax war" >
- <!--Copy our nutchwax nutch-site.xml template into the build dir as
- nutch-site.xml. Then in the below, add it into the WEB-INF/classes dir.
- -->
- <war destfile="${build.dir}/${name}-webapp${nutchwax.version}.war" webxml="${this.web}/web.xml">
- <fileset dir="${nutch.web}/jsp">
- <exclude name="**/search.jsp"/>
- <exclude name="**/web.xml"/>
- <exclude name="**/refine*.xml"/>
- <!--Don't copy these over until they jsp compile.-->
- <exclude name="**/cluster.jsp"/>
- <exclude name="**/refine-query*"/>
- </fileset>
- <fileset dir="${this.web}">
- <exclude name="**/web.xml"/>
- </fileset>
- <classes dir="${nutch.root}/conf" >
- <exclude name="**/*.template"/>
- </classes>
- <classes dir="${root}/conf">
- <exclude name="**/*.template"/>
- </classes>
- <classes dir="${nutch.web}/locale"/>
- <classes file="${this.web}/log4j.properties"/>
- <lib dir="${root}/lib">
- <include name="archive-commons-*.jar" />
- </lib>
- <lib dir="${nutch.root}/build">
- <include name="nutch*.jar"/>
- </lib>
- <lib dir="${nutch.root}/lib">
- <include name="lucene*.jar"/>
- <include name="hadoop*.jar"/>
- <include name="taglibs-*.jar"/>
- <include name="dom4j-*.jar"/>
- <include name="xerces-*.jar"/>
- <include name="log4j-*.jar"/>
- <include name="commons-lang-*.jar"/>
- <include name="commons-cli-*.jar"/>
- <include name="commons-logging-*.jar"/>
- </lib>
- <!--Copy into place the nutchwax classes.-->
- <zipfileset prefix="WEB-INF/classes"
- dir="${build.dir}/classes/" />
-
- <!--Be selective about plugins to copy. Shrinks size of webapp.
- -->
- <zipfileset prefix="WEB-INF/classes/plugins"
- dir="${nutch.root}/build/plugins">
- <include name="analysis-*/**" />
- <include name="clustering-*/**" />
- <include name="language-*/**" />
- <include name="lib-lucene-*/**" />
- <include name="lib-log4j-*/**" />
- <include name="lib-regex-*/**" />
- <include name="microformats-*/**" />
- <include name="nutch-*/**" />
- <include name="query-*/**" />
- <include name="urlfilter-*/**" />
- <include name="urlnormalizer-*/**" />
- <include name="summary-*/**" />
- <include name="urlfilter-*/**" />
- </zipfileset>
- <zipfileset prefix="WEB-INF/classes/plugins"
- dir="${build.dir}/wax-plugins"/>
- <webinf dir="${nutch.root}/lib">
- <include name="taglibs-*.tld"/>
- </webinf>
- </war>
- </target>
-
-
- <!-- ================================================================== -->
- <!-- Clean. Delete the build files, and their directories -->
- <!-- ================================================================== -->
- <target name="clean" description="Clean up all built">
- <delete dir="${build.dir}"/>
- </target>
-
- <!-- ================================================================== -->
- <!-- Clean all. Delete the build files including third-party builds -->
- <!-- and their directories -->
- <!-- ================================================================== -->
- <target name="clean-all"
- depends="clean,third.party.clean"
- description="Clean up all built including third-party dependencies" />
-
-</project>
Modified: trunk/archive-access/projects/nutchwax/nutchwax-thirdparty/pom.xml
===================================================================
--- trunk/archive-access/projects/nutchwax/nutchwax-thirdparty/pom.xml 2007-03-21 17:06:07 UTC (rev 1622)
+++ trunk/archive-access/projects/nutchwax/nutchwax-thirdparty/pom.xml 2007-03-21 19:13:23 UTC (rev 1623)
@@ -37,8 +37,10 @@
Done as part of the generate-sources step so that
we can invoke it from eclipse.
-->
- <echo>Compiling third.party dependencies as part of generate-sources</echo>
- <ant dir=".." target="third.party.jar"/>
+ <echo>Building nutch third-party dependency (jar)</echo>
+ <ant dir="../third-party/nutch" target="jar" inheritAll="false" >
+ <property name="build.compiler" value="extJavac" />
+ </ant>
<!--Copy over the nutch classes to target/classes so they
can be found by later modules (target/classes is what maven
has on its classpath when it goes to build subsequent modules).
@@ -46,7 +48,10 @@
<copy todir="target/classes" overwrite="true">
<fileset dir="../third-party/nutch/build/classes" />
</copy>
- <ant dir=".." target="third.party.plugins"/>
+ <echo>Building nutch third-party dependency (plugins)</echo>
+ <ant dir="../third-party/nutch" target="compile-plugins" inheritAll="false" >
+ <property name="build.compiler" value="extJavac" />
+ </ant>
</tasks>
</configuration>
<goals>
@@ -58,7 +63,10 @@
<phase>clean</phase>
<configuration>
<tasks>
- <ant dir=".." target="clean-all"/>
+ <echo>Cleaning nutch third-party dependency</echo>
+ <ant dir="../third-party/nutch" target="clean" inheritAll="false" >
+ <property name="build.compiler" value="extJavac" />
+ </ant>
</tasks>
</configuration>
<goals>
Modified: trunk/archive-access/projects/nutchwax/src/main/assembly/distribution.xml
===================================================================
--- trunk/archive-access/projects/nutchwax/src/main/assembly/distribution.xml 2007-03-21 17:06:07 UTC (rev 1622)
+++ trunk/archive-access/projects/nutchwax/src/main/assembly/distribution.xml 2007-03-21 19:13:23 UTC (rev 1623)
@@ -17,8 +17,8 @@
<directory>target</directory>
<outputDirectory />
<includes>
- <include>nutchwax*.jar</include>
- <include>nutchwax*.war</include>
+ <include>nutchwax-job/target/nutchwax-job*.jar</include>
+ <include>nutchwax-webapp/nutchwax-webapp*.war</include>
</includes>
</fileSet>
<fileSet>
This was sent by the SourceForge.net collaborative development platform, the world's largest Open Source development site.
|
