Changeset 29730

Timestamp:

2015-02-11T21:09:47+13:00 (9 years ago)

Author:

ak19

Message:

The second and final part of the commits to getting GLI running again and parsing web.xml, after the changes to commit r29687, where web.xml was split into two and included server.xml. In this commit: 1. GLI uses an EntityResolver to resolve entities in web.xml that are defined in the included servlets.xml file. In order to keep XMLTools.java tidy and hopefully make the GLI entity resolver more reusable, the new GLIEntityResolver.java class checks default search paths first when asked to resolve entities. web/WEB-INF, where web.xml and servlets.xml live, has been added to the default search paths, as also the gli user dir where the web.xml and server.xml will be in a client-gli situation. 2. Small tidy up to Greenstone runtime's GSEntityResolver. 3. Remote Greenstone gliserver.pl needs to also transfer the new server.xml file when zipping up web.xml. 4. Minor touchups to the new README on apache.jar.

Location:

main/trunk

Files:

• gli/lib/README.txt (modified) (3 diffs)
• gli/src/org/greenstone/gatherer/util/GLIEntityResolver.java (added)
• gli/src/org/greenstone/gatherer/util/XMLTools.java (modified) (1 diff)
• greenstone2/common-src/cgi-bin/gliserver.pl (modified) (3 diffs)
• greenstone3/src/java/org/greenstone/gsdl3/util/GSEntityResolver.java (modified) (1 diff)

main/trunk/gli/lib/README.txt

@@ +1 @@
+---------------------------
+THE apache.jar FILE
+---------------------------
+
 GLI's apache.jar is a custom jar file that collects together apache libraries necessary to compile, build and run GLI. There's no documentation on the various jar files that it contains, but I've guessed by inspecting the contents. 
 
-Some of the contents are from xercesImpl.jar and xml-apis.jar, which may need to be kept up to date together, since they are related and interdependent jar files.
+Some of the contents are from xercesImpl.jar and xml-apis.jar, which may need to be kept up-to-date in tandem, since they are related and interdependent jar files as updates to xalan.jar for instance come with correspondingly updated xerces, xml-apis and other related jars.
 
 
-Commit 29729 added the contents of xml-apis.jar (sax) and xercesImpl.jar (both found in GS3/web/WEB-INF/lib else packages/tomcat/lib) into apache.jar, replacing the older xerces and xml-apis contents of apache.jar
+Commit r29729 adds the contents of xml-apis.jar (sax) and xercesImpl.jar (both found in GS3/web/WEB-INF/lib else packages/tomcat/lib) into apache.jar, replacing the older xerces and xml-apis contained of apache.jar.
+This update was necessary to handle commit r29687 where web.xml was split into web.xml and servlets.xml, with the former including the latter as an entity. GLI needs to use an EntityResolver to work with this, but the older xerces and xml-apis in apache.jar would still result in GLI failing to start, unable to parse web.xml because of the way in which servlets.xml was included. The xercesImpl and xml-apis jars in GS3/web/WEB-INF/lib (copied into packages/tomcat/lib) contain versions of these jars that do make the EntityResolver in GLI work.
 
+
+---------------------------
+CONTENTS OF apache.jar
+---------------------------
 
 Overview of the contents of apache.jar and whether they got updated in the upcoming commit 29729.
@@ -27 +36 @@
             - codec: appears to be from commons-codec, 
             - logging: appears to be from commons-logging.
-            - not sure where the 4 other subfolders of commons (beanutils, collections, digester, httpclient) are from.
+            - not sure where the 4 other subfolders of commons (beanutils, collections, digester, httpclient) are from. See commit message of http://trac.greenstone.org/log/gli/trunk/lib/apache.jar?rev=14319 for ideas.
         
 
@@ -33 +42 @@
 
 
-After unzipping apache.jar, the contents were updated as described above.
+---------------------------
+UPDATING apache.jar
+---------------------------
+
+After unzipping apache.jar, its contents were updated to use the later xerces and xml-apis as described above.
 
 How to recreate the apache.jar:

main/trunk/gli/src/org/greenstone/gatherer/util/XMLTools.java

@@ -619 +619 @@
             parser.setFeature("http://apache.org/xml/features/dom/defer-node-expansion", true);
             parser.setFeature("http://apache.org/xml/features/dom/include-ignorable-whitespace", false);
+            parser.setEntityResolver(new GLIEntityResolver());
             parser.parse(isc);
             document = parser.getDocument();

main/trunk/greenstone2/common-src/cgi-bin/gliserver.pl

@@ -746 +746 @@
     my $java_classpath = &util::filename_cat($ENV{'GSDLHOME'}, "bin", "java", "GLIServer.jar");
     my $zip_file_path = &util::filename_cat($collection_directory, $collection_tail_name . "-file-" . $timestamp . ".zip");
-    my $java_args = "\"$zip_file_path\" \"$collection_directory\" \"$file\"";
+    my $java_args = "\"$zip_file_path\" \"$collection_directory\" \"$file\" servlets.xml";
     $ENV{'LANG'} = $lang_env;
     my $java_command = "\"$java\" -classpath \"$java_classpath\" org.greenstone.gatherer.remote.ZipFiles $java_args"; 
@@ -787 +787 @@
     }
 
-    my $web_inf_directory = &util::filename_cat($ENV{'GSDL3SRCHOME'}, "web", "WEB-INF");
+    my $web_inf_directory = &util::filename_cat($ENV{'GSDL3HOME'}, "WEB-INF");
     $gsdl_cgi->checked_chdir($web_inf_directory);
 
@@ -799 +799 @@
     my $java_classpath = &util::filename_cat($ENV{'GSDLHOME'}, "bin", "java", "GLIServer.jar");
     my $zip_file_path = &util::filename_cat($web_inf_directory, "webxml" . $timestamp . ".zip");
-    my $java_args = "\"$zip_file_path\" \"$web_inf_directory\" \"$file\"";
+    my $java_args = "\"$zip_file_path\" \"$web_inf_directory\" \"$file\" servlets.xml";
     $ENV{'LANG'} = $lang_env;
     my $java_command = "\"$java\" -classpath \"$java_classpath\" org.greenstone.gatherer.remote.ZipFiles $java_args"; 

main/trunk/greenstone3/src/java/org/greenstone/gsdl3/util/GSEntityResolver.java

@@ -61 +61 @@
         return new InputSource(system_id);
         } else {
-        temp_id = temp_id.substring(temp_id.lastIndexOf("/")+1);
+        temp_id = f.getName(); //temp_id = temp_id.substring(temp_id.lastIndexOf("/")+1);
         }
-    } else {
+    } else { // not a file
         if (temp_id.indexOf("/")!= -1) {
         temp_id = temp_id.substring(temp_id.lastIndexOf("/")+1);