Changeset 34100

Timestamp:

2020-04-09T11:39:25+12:00 (4 years ago)

Author:

ak19

Message:

Related to trac changesets committed for 33793, 34098 and 34099 related to fixing encoding related parsing issues on different (Chinese) locales. Those commits were necessary to support different locales, but also had the side-effect of apparently fixing the issue of UTF-8 like a-macron chars not displaying correctly in the running GS3 server. A fix to resolve the same issue for when GS3 was launched in other ways (not necessary through our own tomcat) was needed, however. This commit carries out the idea Dr Bainbridge described in the email thread with subject 'GS3 encoding problem reproduces immediately', in his email response dated Apr 8, 10:50 AM.

File:

• main/trunk/greenstone3/src/java/org/greenstone/gsdl3/util/XMLConverter.java (modified) (6 diffs)

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

@@ -72 +72 @@
     static Logger logger = Logger.getLogger(org.greenstone.gsdl3.util.XMLConverter.class.getName());
 
+    
+    public static final String XML_FILE_ENCODING="UTF-8";
+
+    static {
+    String fileEncodingProperty = System.getProperty("file.encoding");
+
+    // log a warning if the file.encoding Java property departs from XML_FILE_ENCODING,
+    // but always use the latter for reading in XML files in getDOM(File, ...) methods
+    if(fileEncodingProperty == null) {
+        logger.warn("file.encoding Java property was not set.");
+        System.err.println("Using XMLConverter.XML_FILE_ENCODING="+XML_FILE+ENCODING+"in XMLConverter.getDOM() calls.");
+    }
+    else if(!fileEncodingProperty.equals(XML_FILE_ENCODING)) {
+        logger.warn("file.encoding property passed in with JAVA_TOOLS_OPTIONS: "
+                   + fileEncodingProperty);
+        logger.warn("does not match XMLConverter.XML_FILE_ENCODING: " + XML_FILE_ENCODING);
+        logger.warn("Using XMLConverter.XML_FILE_ENCODING in XMLConverter.getDOM() calls.");
+    }
+    };
+    
+
     /** the no-args constructor */
     public XMLConverter()
@@ -78 +99 @@
     }
 
-    /** returns a DOM Document */
+    /** returns a DOM Document 
+     * Question: why is this not simply calling the getDOM(String in, String encoding) version?
+     * Answer: Because the string, being a Java String, is already fully formed in unicode,
+     * so it doesn't need to call that other version.
+     */
   public static Document getDOM(String in)
     {
@@ -123 +148 @@
     /** returns a DOM Document */
   public static Document getDOM(File in) {
+
+      // now we're always going to explicitly use XML_FILE_ENCODING (UTF-8)
+      // as the encoding to read in a file unless otherwise specified.
+      return getDOM(in, XML_FILE_ENCODING, null);
+      
+      /*
         try
         {
@@ -140 +171 @@
         }
         return null;
+      */
     }
 
@@ -169 +201 @@
 
   public static Document getDOM(File in, EntityResolver er) {
-    
+      // now we're always going to explicitly use XML_FILE_ENCODING (UTF-8)
+      // as the encoding to read in a file unless otherwise specified.
+      return getDOM(in, XML_FILE_ENCODING, er);
+      /*
     try {      
       InputSource xml_source = new InputSource(new FileInputStream(in));
@@ -182 +217 @@
       }
     return null;
+      */
   }