From 00d924f0db3c74afef370eadca9d7f21cfc3b637 Mon Sep 17 00:00:00 2001
From: James Rutherford <jrutherford@users.sourceforge.net>
Date: Mon, 9 Apr 2007 12:03:14 +0000
Subject: [PATCH] (Stuart Lewis) - SF Patch #1587225 Google and html sitemap
 generator

Note that this patch was modified to work with jdk 1.4. Once we are using jdk
1.5 or higher, the annotations can go back in.


git-svn-id: http://scm.dspace.org/svn/repo/trunk@1778 9c30dcfa-912a-0410-8fc2-9e0234be79fd
---
 dspace/CHANGES                       |  1 +
 dspace/config/dspace.cfg             | 16 +++++++++++++
 dspace/etc/dspace-web.xml            | 35 ++++++++++++++++++++++++----
 dspace/jsp/layout/footer-default.jsp |  1 +
 dspace/jsp/robots.txt                |  1 +
 5 files changed, 50 insertions(+), 4 deletions(-)

diff --git a/dspace/CHANGES b/dspace/CHANGES
index b09d507795..e77d027d3b 100644
--- a/dspace/CHANGES
+++ b/dspace/CHANGES
@@ -40,6 +40,7 @@
 - SF Patch #1628889 Improve file size descriptions in ItemTag
 - SF Patch #1641678 [dspace]/bin scripts for import and export
 - SF Patch #1642336 Restrict domains of self-registered users
+- SF Patch #1587225 Google and html sitemap generator
 
 (Monika Mevenkamp)
 - SF Patch #1654469 stat scripts always return exit code 1
diff --git a/dspace/config/dspace.cfg b/dspace/config/dspace.cfg
index 9820adda8a..dae159649e 100644
--- a/dspace/config/dspace.cfg
+++ b/dspace/config/dspace.cfg
@@ -631,6 +631,22 @@ webui.subjectlinks.enable = false
 #
 # webui.html.max-depth-guess = 3
 
+
+#### Sitemap settings #####
+#
+# Comma-separated list of search engine URLs to 'ping' when a new Sitemap has
+# been created.  Include everything except the Sitemap URL itself (which will
+# be URL-encoded and appended to form the actual URL 'pinged').
+#
+sitemap.engineurls = http://www.google.com/webmasters/sitemaps/ping?sitemap=
+
+# Add this to the above parameter if you have an application ID with Yahoo
+# (Replace REPLACE_ME with your application ID)
+# http://search.yahooapis.com/SiteExplorerService/V1/updateNotification?appid=REPLACE_ME&url=
+#
+# No known Sitemap 'ping' URL for MSN/Live search
+
+
 #### Checksum Checker Settings ####
 # Default dispatcher in case none specified
 plugin.single.org.dspace.checker.BitstreamDispatcher=org.dspace.checker.SimpleDispatcher
diff --git a/dspace/etc/dspace-web.xml b/dspace/etc/dspace-web.xml
index 4e13301bdd..19648595e5 100644
--- a/dspace/etc/dspace-web.xml
+++ b/dspace/etc/dspace-web.xml
@@ -8,7 +8,7 @@
  -
  - Date: $Date$
  -
- - Copyright (c) 2002, Hewlett-Packard Company and Massachusetts
+ - Copyright (c) 2002-2006, Hewlett-Packard Company and Massachusetts
  - Institute of Technology.  All rights reserved.
  -
  - Redistribution and use in source and binary forms, with or without
@@ -274,6 +274,15 @@
     <servlet-class>org.dspace.app.webui.servlet.HTMLServlet</servlet-class>
   </servlet>
 
+  <servlet>
+    <servlet-name>htmlmap</servlet-name>
+    <servlet-class>org.dspace.app.webui.servlet.SitemapServlet</servlet-class>
+    <init-param>
+      <param-name>type</param-name>
+      <param-value>html</param-value>
+    </init-param>
+  </servlet>
+
   <servlet>
     <servlet-name>internal-error</servlet-name>
     <servlet-class>org.dspace.app.webui.servlet.InternalErrorServlet</servlet-class>
@@ -347,6 +356,15 @@
     <servlet-name>simple-search</servlet-name>
     <servlet-class>org.dspace.app.webui.servlet.SimpleSearchServlet</servlet-class>
   </servlet>
+  
+  <servlet>
+    <servlet-name>sitemap</servlet-name>
+    <servlet-class>org.dspace.app.webui.servlet.SitemapServlet</servlet-class>
+    <init-param>
+      <param-name>type</param-name>
+      <param-value>sitemaps.org</param-value>
+    </init-param>
+  </servlet>
 
   <servlet>
     <servlet-name>statistics</servlet-name>
@@ -387,8 +405,7 @@
     <servlet-name>workspace</servlet-name>
     <servlet-class>org.dspace.app.webui.servlet.WorkspaceServlet</servlet-class>
   </servlet>
-  
-    
+
   <servlet>
     <servlet-name>controlledvocabulary</servlet-name>
     <servlet-class>org.dspace.app.webui.servlet.ControlledVocabularyServlet</servlet-class>
@@ -527,6 +544,11 @@
     <url-pattern>/html/*</url-pattern>
   </servlet-mapping>
 
+  <servlet-mapping>
+    <servlet-name>htmlmap</servlet-name>
+    <url-pattern>/htmlmap</url-pattern>
+  </servlet-mapping>
+
   <servlet-mapping>
     <servlet-name>internal-error</servlet-name>
     <url-pattern>/internal-error</url-pattern>
@@ -596,6 +618,11 @@
     <servlet-name>simple-search</servlet-name>
     <url-pattern>/simple-search</url-pattern>
   </servlet-mapping>
+  
+  <servlet-mapping>
+    <servlet-name>sitemap</servlet-name>
+    <url-pattern>/sitemap</url-pattern>
+  </servlet-mapping>
 
   <servlet-mapping>
     <servlet-name>statistics</servlet-name>
@@ -637,7 +664,6 @@
     <url-pattern>/workspace</url-pattern>
   </servlet-mapping>
   
-    
   <servlet-mapping>
     <servlet-name>controlledvocabulary</servlet-name>
     <url-pattern>/controlledvocabulary</url-pattern>
@@ -648,6 +674,7 @@
     <url-pattern>/subject-search</url-pattern>
   </servlet-mapping> 
 
+
   <!-- Icon MIME type -->
   <mime-mapping>
     <extension>ico</extension>
diff --git a/dspace/jsp/layout/footer-default.jsp b/dspace/jsp/layout/footer-default.jsp
index f0e49e58eb..5dbaee3eb0 100644
--- a/dspace/jsp/layout/footer-default.jsp
+++ b/dspace/jsp/layout/footer-default.jsp
@@ -87,6 +87,7 @@
                             <td class="pageFootnote">
                                 <fmt:message key="jsp.layout.footer-default.text"/>&nbsp;-
                                 <a target="_blank" href="<%= request.getContextPath() %>/feedback"><fmt:message key="jsp.layout.footer-default.feedback"/></a>
+                                <a href="<%= request.getContextPath() %>/htmlmap"></a>
                             </td>
                             <td nowrap="nowrap" valign="middle"> <%-- nowrap, valign for broken NS 4.x --%>
                             </td>
diff --git a/dspace/jsp/robots.txt b/dspace/jsp/robots.txt
index 23a9ed92db..e0bfd82fa2 100644
--- a/dspace/jsp/robots.txt
+++ b/dspace/jsp/robots.txt
@@ -4,3 +4,4 @@ Disallow: /browse-author
 Disallow: /items-by-author
 Disallow: /browse-date
 Disallow: /browse-subject
+Disallow: /browse-title