Changeset 33730 for main

Timestamp:

2019-11-29T23:40:14+13:00 (4 years ago)

Author:

ak19

Message:

Finally, got the code back to achieving the same thing as the partial bugfix from yesterday to the issue of non-ascii meta filenames not retaining any meta assigned at file level. The difference in today's commit is that the reworking does it Dr Bainbridge's way: instead of hex encoding one value to compare to a hex encoded value, am now ensuring both are decoded hex values when comparing. Adding a bugfix to recently introduced FilenameEncoding.decodeStringContainingHexEntities() since I forgot to test for the most basic case of all: when a string is not hex encoded, it should return the same string (the bug caused it to return nothing and which caused my rewrite of yesterday's work to look like it wasn't working for so long).

Location:

main/trunk/gli/src/org/greenstone/gatherer/metadata

Files:

• FilenameEncoding.java (modified) (3 diffs)
• MetadataXMLFile.java (modified) (5 diffs)

main/trunk/gli/src/org/greenstone/gatherer/metadata/FilenameEncoding.java

@@ -311 +311 @@
     public static String decodeStringContainingHexEntities(String str) {
         String result = "";
-        boolean done = false;       
         Pattern hexPattern = Pattern.compile("(&#x[0-9a-zA-Z]{1,4}+;)");
         Matcher matcher = hexPattern.matcher(str);
@@ -345 +344 @@
         }
         
-        if(endMatchIndex != -1) {
+        if(endMatchIndex != -1) { // attach any suffix once we finished processing all the hex codes
             result += str.substring(endMatchIndex);
             //System.err.println("suffix: " + str.substring(endMatchIndex));
         }
+        else { // there were no hex codes to decode, return string as is
+            result = str;
+        }
         
         return result;
@@ -355 +357 @@
     /** Attempting to produce the equivalent method fileToURLEncoding() above, but taking a String as input parameter */
      public static String fileNameToHex(String filename) {
-        /*String filename_url_encoded = "";
-        try {
-            URI filename_uri = new URI(filename);
-            String filename_ascii = filename_uri.toASCIIString();
-            String filename_raw_bytes = URLDecoder.decode(filename_ascii,"ISO-8859-1");
-            filename_url_encoded = iso_8859_1_filename_to_url_encoded(filename_raw_bytes); 
-            return filename_url_encoded;
-        } catch (Exception e) {
-            e.printStackTrace();
-            // Give up trying to convert
-            filename_url_encoded = filename; 
-        }
-        return filename_url_encoded;
-        */
-        
+
         String hexFilename = "";
         for(int i = 0; i < filename.length(); i++) {

main/trunk/gli/src/org/greenstone/gatherer/metadata/MetadataXMLFile.java

@@ -35 +35 @@
 import org.w3c.dom.*;
 
+import org.greenstone.gatherer.util.Utility;
 
 /** This class represents one metadata.xml file */
@@ -262 +263 @@
         String current_filename_element_value = XMLTools.getElementTextValue(current_filename_element);
 
-        String regexed_file_relative_path = MetadataTools.getRegularExpressionThatMatchesFilePath(file_relative_path);
-        //System.err.println("Looking in meta.xml for regexed version of filename: " + regexed_file_relative_path);
+        //System.err.println("\n  Original TAIL filename was: " + Utility.debugUnicodeString(file.getName()));
+        String hexdecoded_regexed_file_relative_path = FilenameEncoding.decodeStringContainingHexEntities(file_relative_path);
+        //System.err.println("Looking in meta.xml for hexdecoded_regexed_file_RELATIVE_path: " + hexdecoded_regexed_file_relative_path
+            //+ " - debug version: " + Utility.debugUnicodeString(hexdecoded_regexed_file_relative_path));      
     
         // Does this fileset specify metadata for one file only?
@@ -272 +275 @@
         }
 
-        String current_filename_element_value_hex = FilenameEncoding.fileNameToHex(current_filename_element_value);
+        String hexdecoded_current_filename_element_value = FilenameEncoding.decodeStringContainingHexEntities(current_filename_element_value);
+        //System.err.println("   Checking to see if it matches " + hexdecoded_current_filename_element_value + " - debug: " + Utility.debugUnicodeString(hexdecoded_current_filename_element_value));
+            //System.err.println("   Checking to see if it matches " + current_filename_element_value + " - debug: " + Utility.debugUnicodeString(current_filename_element_value));
         
         // This fileset specifies metadata for the file
-        // MetadataXMLFile.addMetadata(CollectionTreeNode, ArrayList) stored filename in uppercase hex, so need to compare with the same
-        if (file_relative_path.matches(current_filename_element_value_hex)) { //if (file_relative_path.matches(current_filename_element_value)) {
-            //System.err.println("Found a match in meta.xml for file name: " + regexed_file_relative_path);
+        // MetadataXMLFile.addMetadata(CollectionTreeNode, ArrayList) stored filename in uppercase hex
+        // so need to make sure everything hex has been decoded (no more hex) to compare apples with apples     
+        if (hexdecoded_regexed_file_relative_path.matches(hexdecoded_current_filename_element_value)) { //if (file_relative_path.matches(current_filename_element_value)) {
+            //System.err.println("   @@@ Found a match in meta.xml for hexdecoded_regexed_file_relative_path: " + hexdecoded_regexed_file_relative_path);
             current_fileset_matches = true;
             if (!file_relative_path.equals("") && current_filename_element_value.equals(DIRECTORY_FILENAME)) {
@@ -283 +289 @@
             }
             break;
-        }       
+        } //else {
+            //System.err.println( hexdecoded_regexed_file_relative_path + " does not match " + hexdecoded_current_filename_element_value);
+            //System.err.println( Utility.debugUnicodeString(hexdecoded_regexed_file_relative_path) + " does not match " + Utility.debugUnicodeString(hexdecoded_current_filename_element_value));
+        //}
         
         // This fileset specifies metadata for the folder the file is in
-        if (regexed_file_relative_path.startsWith(current_filename_element_value + FilenameEncoding.URL_FILE_SEPARATOR)) {
+        if (hexdecoded_regexed_file_relative_path.startsWith(hexdecoded_current_filename_element_value + FilenameEncoding.URL_FILE_SEPARATOR)) {
             current_fileset_matches = true;
             folder_metadata_inherited_from = new File(metadata_xml_file_directory, current_filename_element_value);
@@ -616 +625 @@
     // If we have a file loaded into memory and it has been modified, save it now
     if (loaded_file != null && loaded_file_changed == true) {
+        //System.err.println("START saveLoadedFile(), loaded_file_document:\n" + XMLTools.elementToString(loaded_file_document.getDocumentElement(), true));
+        
         XMLTools.writeXMLFile(loaded_file, loaded_file_document, nonEscapingElements);
     
+        Document doc = XMLTools.parseXMLFile(loaded_file);
+        //System.err.println("AT END saveLoadedFile(), PARSED loaded_file contains:\n" +  XMLTools.elementToString(doc.getDocumentElement(), true));
         
         loaded_file_changed = false;