From 3e2e688bd097727f457f1aa882c74a128f0a53da Mon Sep 17 00:00:00 2001
From: Markus Jelsma <markus@apache.org>
Date: Thu, 12 Feb 2015 08:38:56 +0000
Subject: [PATCH] NUTCH 1925 Upgrade Tika to version 1.7

git-svn-id: https://svn.apache.org/repos/asf/nutch/trunk@1659168 13f79535-47bb-0310-9956-ffa450edef68
---
 CHANGES.txt                                  |  2 ++
 ivy/ivy.xml                                  |  2 +-
 src/plugin/parse-tika/howto_upgrade_tika.txt |  2 +-
 src/plugin/parse-tika/ivy.xml                |  2 +-
 src/plugin/parse-tika/plugin.xml             | 19 +++++++++----------
 5 files changed, 14 insertions(+), 13 deletions(-)

diff --git a/CHANGES.txt b/CHANGES.txt
index 7b1229b551..57436c347c 100644
--- a/CHANGES.txt
+++ b/CHANGES.txt
@@ -2,6 +2,8 @@ Nutch Change Log
 
 Nutch Current Development 1.10-SNAPSHOT
 
+* NUTCH-1925 Upgrade to Apache Tika 1.7 (Tyler Palsulich via markus)
+
 * NUTCH-1323 AjaxNormalizer (markus)
 
 * NUTCH-1918 TikaParser specifies a default namespace when generating DOM (jnioche)
diff --git a/ivy/ivy.xml b/ivy/ivy.xml
index 8f420913e0..e0ee11d3e3 100644
--- a/ivy/ivy.xml
+++ b/ivy/ivy.xml
@@ -60,7 +60,7 @@
 			<exclude org="ant" name="ant" />
 		</dependency>
 
-		<dependency org="org.apache.tika" name="tika-core" rev="1.6" />
+		<dependency org="org.apache.tika" name="tika-core" rev="1.7" />
 		<dependency org="com.ibm.icu" name="icu4j" rev="4.0.1" />
 
 		<dependency org="xerces" name="xercesImpl" rev="2.9.1" />
diff --git a/src/plugin/parse-tika/howto_upgrade_tika.txt b/src/plugin/parse-tika/howto_upgrade_tika.txt
index 96e9f02e7a..63a05a4bf7 100644
--- a/src/plugin/parse-tika/howto_upgrade_tika.txt
+++ b/src/plugin/parse-tika/howto_upgrade_tika.txt
@@ -5,4 +5,4 @@
 3. Upgrade Tika's own dependencies in src/plugin/parse-tika/plugin.xml
    To get the list of dependencies and their versions execute:
    $ ant -f ./build-ivy.xml
-   $ ls lib/
+   $ ls lib | sed 's/^/      <library name="/g' | sed 's/$/"\/>/g'
diff --git a/src/plugin/parse-tika/ivy.xml b/src/plugin/parse-tika/ivy.xml
index 7cad603ac0..377497b6e9 100644
--- a/src/plugin/parse-tika/ivy.xml
+++ b/src/plugin/parse-tika/ivy.xml
@@ -36,7 +36,7 @@
   </publications>
 
   <dependencies>
-    <dependency org="org.apache.tika" name="tika-parsers" rev="1.6" conf="*->default">
+    <dependency org="org.apache.tika" name="tika-parsers" rev="1.7" conf="*->default">
      <exclude org="org.apache.tika" name="tika-core" />
     </dependency>
     <override module="rome" rev="0.9"/>
diff --git a/src/plugin/parse-tika/plugin.xml b/src/plugin/parse-tika/plugin.xml
index 49f29b312b..e822fd462e 100644
--- a/src/plugin/parse-tika/plugin.xml
+++ b/src/plugin/parse-tika/plugin.xml
@@ -25,7 +25,6 @@
       <library name="parse-tika.jar">
          <export name="*"/>
       </library>
-      
       <library name="apache-mime4j-core-0.7.2.jar"/>
       <library name="apache-mime4j-dom-0.7.2.jar"/>
       <library name="asm-debug-all-4.1.jar"/>
@@ -37,26 +36,26 @@
       <library name="commons-compress-1.8.1.jar"/>
       <library name="commons-httpclient-3.1.jar"/>
       <library name="commons-logging-1.1.1.jar"/>
-      <library name="fontbox-1.8.6.jar"/>
+      <library name="fontbox-1.8.8.jar"/>
       <library name="isoparser-1.0.2.jar"/>
       <library name="java-libpst-0.8.1.jar"/>
       <library name="jcip-annotations-1.0.jar"/>
       <library name="jdom-1.0.jar"/>
-      <library name="jempbox-1.8.6.jar"/>
+      <library name="jempbox-1.8.8.jar"/>
       <library name="jhighlight-1.0.jar"/>
       <library name="jmatio-1.0.jar"/>
       <library name="juniversalchardet-1.0.3.jar"/>
       <library name="metadata-extractor-2.6.2.jar"/>
       <library name="netcdf-4.2.20.jar"/>
-      <library name="pdfbox-1.8.6.jar"/>
-      <library name="poi-3.11-beta2.jar"/>
-      <library name="poi-ooxml-3.11-beta2.jar"/>
-      <library name="poi-ooxml-schemas-3.11-beta2.jar"/>
-      <library name="poi-scratchpad-3.11-beta2.jar"/>
+      <library name="pdfbox-1.8.8.jar"/>
+      <library name="poi-3.11.jar"/>
+      <library name="poi-ooxml-3.11.jar"/>
+      <library name="poi-ooxml-schemas-3.11.jar"/>
+      <library name="poi-scratchpad-3.11.jar"/>
       <library name="rome-0.9.jar"/>
       <library name="slf4j-api-1.6.1.jar"/>
       <library name="tagsoup-1.2.1.jar"/>
-      <library name="tika-parsers-1.6.jar"/>
+      <library name="tika-parsers-1.7.jar"/>
       <library name="unidataCommon-4.2.20.jar"/>
       <library name="vorbis-java-core-0.6.jar"/>
       <library name="vorbis-java-tika-0.6.jar"/>
@@ -64,7 +63,7 @@
       <library name="xml-apis-1.3.03.jar"/>
       <library name="xmlbeans-2.6.0.jar"/>
       <library name="xmpcore-5.1.2.jar"/>
-      <library name="xz-1.5.jar"/>
+      <library name="xz-1.5.jar"/>   
    </runtime>
 
    <requires>