Changeset 5832

Timestamp:

2003-11-13T13:55:46+13:00 (20 years ago)

Author:

kjdon

Message:

now have changed the format parsing so that it allows format over mulitple lines too

File:

• trunk/gli/src/org/greenstone/gatherer/cdm/CollectionConfiguration.java (modified) (6 diffs)

trunk/gli/src/org/greenstone/gatherer/cdm/CollectionConfiguration.java

@@ -758 +758 @@
         String language_str = command_element.getAttribute(LANGUAGE_ATTRIBUTE);
         // If this is element is in english, and it is the first one found, we don't need to write the language argument.
-        // we should always write the language string
         //if(!language_str.equals(ENGLISH_LANGUAGE_STR) || known_metadata == null || known_metadata.contains(name_str)) {
+        // changed so that we always write the language string
         text.append(LBRACKET_CHARACTER);
         text.append(LANGUAGE_ARGUMENT);
@@ -765 +765 @@
         text.append(RBRACKET_CHARACTER);
         text.append(SPACE_CHARACTER);
-            //}
+        //}
         if(known_metadata != null) {
             known_metadata.add(name_str);
@@ -840 +840 @@
             }
             if(command_element == null && command_type.equals(FORMAT_STR)) {
-            command_element = parseFormat(command_str);
+            command_element = parseFormat(command_str, in);
             }
             if(command_element == null && (command_type.equals(INDEX_STR)  || command_type.equals(COMMENTED_INDEXES_STR))) {
@@ -958 +958 @@
     }
 
-    private Element parseFormat(String command_str) {
+    private Element parseFormat(String command_str, BufferedReader in) {
     Element command_element = null;
     try {
@@ -967 +967 @@
         tokenizer.nextToken();
         command_element.setAttribute(NAME_ATTRIBUTE, tokenizer.nextToken());
-        String format_value = tokenizer.nextToken();
+        String value_str = tokenizer.nextToken();
         // If the value is true or false we add it as an attribute
-        if(format_value.equalsIgnoreCase(TRUE_STR) || format_value.equalsIgnoreCase(FALSE_STR)) {
-            command_element.setAttribute(VALUE_ATTRIBUTE, format_value.toLowerCase());
+        if(value_str.equalsIgnoreCase(TRUE_STR) || value_str.equalsIgnoreCase(FALSE_STR)) {
+            command_element.setAttribute(VALUE_ATTRIBUTE, value_str.toLowerCase());
         }
         // Otherwise it gets added as a text node
         else {
-            // Strip any speech marks
-            if(format_value.startsWith(SPEECH_CHARACTER) && format_value.endsWith(SPEECH_CHARACTER)) {
-            format_value = format_value.substring(1, format_value.length() - 1);
-            }
-            // Decode from Greenstone back into text
-            format_value = Codec.transform(format_value, Codec.GREENSTONE_TO_DOM);
-            // And add to DOM
-            MSMUtils.setValue(command_element, format_value);
-        }
-        format_value = null;
+
+            // now we need to handle the case where the value is enclosed in quotes (single or double) and may extend across multiple lines
+            String start_string = value_str.substring(0,1);
+            if (start_string.equals("\"") || start_string.equals("\'")) {
+            if (value_str.endsWith(start_string) && value_str.length()!=1) {
+                // we remove the quotes from the ends
+                value_str = value_str.substring(1, value_str.length() - 1);
+            } else {
+                
+                // remove the first quote
+                StringBuffer value_raw = new StringBuffer(value_str.substring(1));
+                // add the new line back in
+                value_raw.append(StaticStrings.NEW_LINE_CHAR);
+                int pos = value_raw.indexOf(start_string);
+                int old_pos = 0;
+                while (pos != -1 && value_raw.charAt(pos-1)=='\\') {
+                old_pos = pos+1;
+                pos = value_raw.indexOf(start_string, old_pos);
+                }
+                while(pos == -1) {
+                String next_line = in.readLine();
+                if(next_line != null) {
+                    value_raw.append(next_line);
+                    value_raw.append(StaticStrings.NEW_LINE_CHAR);
+                }
+                next_line = null;
+                pos = value_raw.indexOf(start_string, old_pos);
+                while (pos != -1 && value_raw.charAt(pos-1)=='\\') {
+                    old_pos = pos+1;
+                    pos = value_raw.indexOf(start_string, old_pos);
+                }
+                }
+                
+                value_str = value_raw.substring(0, value_raw.lastIndexOf(start_string));
+                value_raw = null;
+                
+            } // else
+            } // if starts with a quote
+                    
+            if(value_str != null) {
+            // Ready the value str (which can contain all sorts of funky characters) for writing as a DOM value
+            value_str = Codec.transform(value_str, Codec.GREENSTONE_TO_DOM);
+            MSMUtils.setValue(command_element, value_str);
+            }
+            else {
+            command_element = null;
+            }
+            start_string = null;
+        }
+        value_str = null;
+        
         }
         tokenizer = null;
     }
-    catch(Exception exception) {
+    catch (Exception exception) {
     }
     return command_element;
@@ -1165 +1206 @@
             } else {
             
-            Gatherer.println("cdm.CollectionConfiguration.parseMEtadata: we have found a starting quote but not an ending quote, so now read lines until we get to an end quote");
+            // remove the first quote
             StringBuffer value_raw = new StringBuffer(value_str.substring(1));
+            // add the new line back in
+            value_raw.append(StaticStrings.NEW_LINE_CHAR);
+
             int pos = value_raw.indexOf(start_string);
             int old_pos = 0;