[Archive-access-cvs] archive-access/projects/nutch .classpath,1.11,1.12 build.xml,1.13,1.14 project.properties,1.16,1.17

[Michael S. <sta...@us...>]

Update of /cvsroot/archive-access/archive-access/projects/nutch
In directory sc8-pr-cvs1.sourceforge.net:/tmp/cvs-serv30555

Modified Files:
	.classpath build.xml project.properties 
Log Message:
Merge 'mapred' branch into HEAD.
* .classpath
* project.properties
    Update to point at new 0.8 nutch.
* build.xml
    Merge in 'mapred'.  Add job target.  
* conf/nutch-site.xml
    Cleanup. Removed unused properties or properties that have same values
    as nutch-default.xml (Except 'searcher.dir' -- keeping that here because
    we'll usually want to change it). Reordered so archive properties are
    towards the end. Brought forward descriptions from nutch-default where
    missing.
* conf/nutch-site.xml.template
    Copy of nutch-site.xml but with the nutchwax defaults turned on.
* src/plugin/build.xml
    Commented out parse-default.
* src/plugin/parse-ext/plugin.xml
    Changed path to parse-pdf.sh.
* src/web/search.jsp
    'mapred' update.
* bin/indexArcs.sh
* conf/ia-parse-plugins.xml 
* lib/commons-codec-1.3.jar 
* src/java/org/archive/access/nutch/ImportArcs.java 
* src/java/org/archive/access/nutch/IndexArcs.java 
    Added.
* bin/arc2seg.sh 
* src/java/org/archive/access/nutch/Arc2Segment.java 
    Removed.


Index: .classpath
===================================================================
RCS file: /cvsroot/archive-access/archive-access/projects/nutch/.classpath,v
retrieving revision 1.11
retrieving revision 1.12
diff -C2 -d -r1.11 -r1.12
*** .classpath	2 Sep 2005 01:08:18 -0000	1.11
--- .classpath	29 Nov 2005 21:43:42 -0000	1.12
***************
*** 7,11 ****
  	<classpathentry kind="con" path="org.eclipse.jdt.launching.JRE_CONTAINER"/>
  	<classpathentry kind="lib" path="nutch/lib/lucene-1.9-rc1-dev.jar"/>
! 	<classpathentry kind="lib" path="nutch/build/nutch-0.7.jar"/>
  	<classpathentry kind="lib" path="lib/arc-1.5.1-200508191341.jar"/>
  	<classpathentry kind="lib" path="lib/commons-httpclient-3.0-alpha2.jar"/>
--- 7,11 ----
  	<classpathentry kind="con" path="org.eclipse.jdt.launching.JRE_CONTAINER"/>
  	<classpathentry kind="lib" path="nutch/lib/lucene-1.9-rc1-dev.jar"/>
! 	<classpathentry kind="lib" path="nutch/build/nutch-0.8-dev.jar"/>
  	<classpathentry kind="lib" path="lib/arc-1.5.1-200508191341.jar"/>
  	<classpathentry kind="lib" path="lib/commons-httpclient-3.0-alpha2.jar"/>

Index: project.properties
===================================================================
RCS file: /cvsroot/archive-access/archive-access/projects/nutch/project.properties,v
retrieving revision 1.16
retrieving revision 1.17
diff -C2 -d -r1.16 -r1.17
*** project.properties	22 Oct 2005 02:43:23 -0000	1.16
--- project.properties	29 Nov 2005 21:43:42 -0000	1.17
***************
*** 18,23 ****
  # Local jars to add to classpath.
  maven.jar.override = on
- maven.jar.corenutch = ${basedir}/nutch/build/nutch-0.7.jar
  maven.jar.lucene = ${basedir}/nutch/lib/lucene-1.9-rc1-dev.jar
  maven.jar.arc = ${basedir}/lib/arc-1.5.1-200508191341.jar
  maven.jar.servlet-api = ${basedir}/nutch/lib/servlet-api.jar
--- 18,23 ----
  # Local jars to add to classpath.
  maven.jar.override = on
  maven.jar.lucene = ${basedir}/nutch/lib/lucene-1.9-rc1-dev.jar
+ maven.jar.corenutch = ${basedir}/nutch/build/nutch-0.8-dev.jar
  maven.jar.arc = ${basedir}/lib/arc-1.5.1-200508191341.jar
  maven.jar.servlet-api = ${basedir}/nutch/lib/servlet-api.jar

Index: build.xml
===================================================================
RCS file: /cvsroot/archive-access/archive-access/projects/nutch/build.xml,v
retrieving revision 1.13
retrieving revision 1.14
diff -C2 -d -r1.13 -r1.14
*** build.xml	28 Nov 2005 22:12:31 -0000	1.13
--- build.xml	29 Nov 2005 21:43:42 -0000	1.14
***************
*** 66,74 ****
      <copy file="${root}/src/plugin/parse-ext/plugin.xml"
        todir="${nutch.root}/build/plugins/parse-ext/"
!       overwrite="true"
!       filtering="true" >
!         <filterset>
!             <filter token="PWD" value="${root}"/>
!         </filterset>
      </copy>
  
--- 66,70 ----
      <copy file="${root}/src/plugin/parse-ext/plugin.xml"
        todir="${nutch.root}/build/plugins/parse-ext/"
!       overwrite="true">
      </copy>
  
***************
*** 113,116 ****
--- 109,127 ----
  
    <!-- ================================================================== -->
+   <!-- Make job jar                                                       -->
+   <!-- ================================================================== -->
+   <!--                                                                    -->
+   <!-- ================================================================== -->
+   <target name="job" depends="compile">
+     <zip destfile="${build.dir}/${name}.job.jar">
+       <zipfileset prefix="classes" file="${conf.dir}/ia-parse-plugins.xml"/>
+       <zipfileset prefix="bin" file="bin/parse-pdf.sh" filemode="755"/>
+       <zipfileset prefix="classes" dir="${build.classes}"/>
+       <zipfileset refid="lib.jars"/>
+     </zip>
+   </target>
+ 
+ 
+   <!-- ================================================================== -->
    <!-- Compile test code                                                  --> 
    <!-- ================================================================== -->