Fixed a logic bug in mapTMWtoTM and mapTMtoTMW.

You can now specify which Unicode font to use via 'java
-Dthdl.tmw.to.unicode.font=Ximalaya ...'.

source/org/thdl/tib/input/Jskad.java

@@ -403,7 +403,8 @@ public class Jskad extends JPanel implements DocumentListener {
                     public void theRealActionPerformed(ActionEvent e) {
                         StringBuffer errors = new StringBuffer();
                         boolean errorReturn
-                            = ((TibetanDocument)dp.getDocument()).convertToUnicode(0, -1, errors); // entire document
+                            = ((TibetanDocument)dp.getDocument()).convertToUnicode(0, -1, errors,
+                                                                                   ThdlOptions.getStringOption("thdl.tmw.to.unicode.font").intern()); // entire document
                         if (errorReturn) {
                             JOptionPane.showMessageDialog(Jskad.this,
                                                           "At least one error occurred while converting Tibetan Machine Web\nto Unicode.  Your document is mostly converted,\nexcept for the following glyphs, which you should replace manually\nbefore retrying:\n"

source/org/thdl/tib/input/TibetanConverter.java

@@ -177,7 +177,10 @@ public class TibetanConverter {
                 } else if (convertToUnicodeMode) {
                     StringBuffer errors = new StringBuffer();
                     // Convert to Unicode:
-                    if (((TibetanDocument)dp.getDocument()).convertToUnicode(0, dp.getDocument().getLength(), errors)) {
+                    if (((TibetanDocument)dp.getDocument()).convertToUnicode(0,
+                                                                             dp.getDocument().getLength(),
+                                                                             errors,
+                                                                             ThdlOptions.getStringOption("thdl.tmw.to.unicode.font").intern())) {
                         System.err.println(errors);
                         exitCode = 42;
                     }

source/org/thdl/tib/text/DuffCode.java

@@ -93,7 +93,8 @@ public final class DuffCode {
 	}
 
     private void setFontNum(int font) {
-        ThdlDebug.verify(font >= 1 && font <= 10);
+        if (!(font >= 1 && font <= 10))
+            throw new IllegalArgumentException("DuffCodes work with font numbers in the range [1, 5] or [1, 10].  This isn't in the range [1, 10]: " + font);
         fontNum = font;
     }
 

source/org/thdl/tib/text/TibetanDocument.java

@@ -202,9 +202,10 @@ public class TibetanDocument extends DefaultStyledDocument {
         @see TibetanMachineWeb#getUnicodeAttributeSet()
     */
     private void replaceDuffsWithUnicode(int fontSize, int startOffset,
-                                         int endOffset, String unicode) {
+                                         int endOffset, String unicode,
+                                         String unicodeFont) {
 		MutableAttributeSet mas
-            = TibetanMachineWeb.getUnicodeAttributeSet();
+            = TibetanMachineWeb.getUnicodeAttributeSet(unicodeFont);
         StyleConstants.setFontSize(mas, fontSize);
 		try {
             replace(startOffset, endOffset - startOffset, unicode, mas);
@@ -512,7 +513,7 @@ public class TibetanDocument extends DefaultStyledDocument {
         cases will be appended to this StringBuffer
     */
     public boolean convertToTM(int begin, int end, StringBuffer errors) {
-        return convertHelper(begin, end, true, false, errors);
+        return convertHelper(begin, end, true, false, errors, null);
     }
 
     /** Converts all TibetanMachine glyphs in the document to
@@ -528,7 +529,7 @@ public class TibetanDocument extends DefaultStyledDocument {
         cases will be appended to this StringBuffer
     */
     public boolean convertToTMW(int begin, int end, StringBuffer errors) {
-        return convertHelper(begin, end, false, false, errors);
+        return convertHelper(begin, end, false, false, errors, null);
     }
 
     /** Converts all TibetanMachineWeb glyphs in the document to
@@ -541,9 +542,13 @@ public class TibetanDocument extends DefaultStyledDocument {
         @return false on 100% success, true if any exceptional case
         was encountered
         @param errors if non-null, then notes about all exceptional
-        cases will be appended to this StringBuffer */
+        cases will be appended to this StringBuffer
-    public boolean convertToUnicode(int begin, int end, StringBuffer errors) {
+        @param unicodeFont the name of the Unicode font to use;
-        return convertHelper(begin, end, false, true, errors);
+        defaults to Arial Unicode MS if null
+    */
+    public boolean convertToUnicode(int begin, int end, StringBuffer errors,
+                                    String unicodeFont) {
+        return convertHelper(begin, end, false, true, errors, unicodeFont);
     }
 
     /** For debugging only.  Start with an empty document, and call
@@ -690,7 +695,8 @@ public class TibetanDocument extends DefaultStyledDocument {
         @see convertToTMW(int,int) 
         @see convertToTM(int,int) */
     private boolean convertHelper(int begin, int end, boolean toTM,
-                                  boolean toUnicode, StringBuffer errors) {
+                                  boolean toUnicode, StringBuffer errors,
+                                  String unicodeFont) {
         // To preserve formatting, we go paragraph by paragraph.
 
         // Use positions, not offsets, because our work on paragraph K
@@ -726,7 +732,8 @@ public class TibetanDocument extends DefaultStyledDocument {
                                 ((finalEndPos.getOffset() < p_end)
                                  ? finalEndPos.getOffset()
                                  : p_end),
-                                toTM, toUnicode, errors, ceh);
+                                toTM, toUnicode, errors, ceh,
+                                unicodeFont);
         }
         if (!ceh.errorReturn
             && pl != getParagraphs(begin, finalEndPos.getOffset()).length) {
@@ -742,7 +749,8 @@ public class TibetanDocument extends DefaultStyledDocument {
     /** See the sole caller, convertHelper. */
     private void convertHelperHelper(int begin, int end, boolean toTM,
                                      boolean toUnicode, StringBuffer errors,
-                                     ConversionErrorHelper ceh) {
+                                     ConversionErrorHelper ceh,
+                                     String unicodeFont) {
         final boolean debug = false;
         if (debug)
             System.err.println("cHH: [" + begin + ", " + end + ")");
@@ -835,7 +843,8 @@ public class TibetanDocument extends DefaultStyledDocument {
                             replaceDuffsWithUnicode(replacementFontSize,
                                                     replacementStartIndex,
                                                     endIndex,
-                                                    replacementQueue.toString());
+                                                    replacementQueue.toString(),
+                                                    unicodeFont);
                         } else {
                             replaceDuffs(replacementFontSize,
                                          replacementStartIndex,
@@ -937,7 +946,8 @@ public class TibetanDocument extends DefaultStyledDocument {
                     replaceDuffsWithUnicode(replacementFontSize,
                                             replacementStartIndex,
                                             endIndex,
-                                            replacementQueue.toString());
+                                            replacementQueue.toString(),
+                                            unicodeFont);
                 } else {
                     replaceDuffs(replacementFontSize,
                                  replacementStartIndex,

source/org/thdl/tib/text/TibetanMachineWeb.java

@@ -76,7 +76,7 @@ public class TibetanMachineWeb implements THDLWylieConstants {
 	private static final String DELIMITER = "~";
 	private static Set top_vowels;
     /** the font we use when we convert TMW->Unicode: */
-	private static SimpleAttributeSet unicodeFontAttributeSet = null;
+	private static SimpleAttributeSet defaultUnicodeFontAttributeSet = null;
     /** a way of encoding the choice of TibetanMachineWeb font from
         that family of 10 fonts: */
 	private static SimpleAttributeSet[] webFontAttributeSet = new SimpleAttributeSet[11];
@@ -283,10 +283,9 @@ public class TibetanMachineWeb implements THDLWylieConstants {
             readInTMFontFiles();
         }
 
-        unicodeFontAttributeSet = new SimpleAttributeSet();
+        defaultUnicodeFontAttributeSet = new SimpleAttributeSet();
-        StyleConstants.setFontFamily(unicodeFontAttributeSet,
+        StyleConstants.setFontFamily(defaultUnicodeFontAttributeSet,
-                                     ThdlOptions.getStringOption("thdl.tmw.to.unicode.font",
+                                     "Arial Unicode MS");
-                                                                 "Arial Unicode MS"));
 
 		webFontAttributeSet[0] = null;
 		for (int i=1; i<webFontAttributeSet.length; i++) {
@@ -665,11 +664,26 @@ public static SimpleAttributeSet getAttributeSet(int font) {
 * in our TMW->Unicode conversion.  This information is required in
 * order to be able to put styled text into {@link TibetanDocument
 * TibetanDocument}.
+* @param unicodeFont the interned name of the Unicode font to use;
+* defaults to Arial Unicode MS if null
 * @return a SimpleAttributeSet for the Unicode font - that is, a way
 * of encoding the font itself */
-public static SimpleAttributeSet getUnicodeAttributeSet() {
+public static SimpleAttributeSet getUnicodeAttributeSet(String unicodeFont) {
-    return unicodeFontAttributeSet;
+    if (null == unicodeFont
+        || "Arial Unicode MS" == unicodeFont)
+        return defaultUnicodeFontAttributeSet;
+    else {
+        SimpleAttributeSet cached
+            = (SimpleAttributeSet)unicodeAttributeSets.get(unicodeFont);
+        if (null == cached) {
+            cached = new SimpleAttributeSet();
+            StyleConstants.setFontFamily(cached, unicodeFont);
+            unicodeAttributeSets.put(unicodeFont, cached);
+        }
+        return cached;
+    }
 }
+private static HashMap unicodeAttributeSets = new HashMap();
 
 /**
 * Gets the AttributeSet for the given TibetanMachine font.
@@ -1008,25 +1022,25 @@ public static DuffCode mapTMtoTMW(int font, int ordinal, int suggestedFont) {
             if (0 == suggestedFont)
                 return TMW_cr;
             else
-                return new DuffCode(suggestedFont, (char)ordinal);
+                return new DuffCode(suggestedFont, (char)ordinal); // FIXME: don't create a new one each time; it wastes heap
         } else if (ordinal == (int)'\n') {
             if (0 == suggestedFont)
                 return TMW_lf;
             else
-                return new DuffCode(suggestedFont, (char)ordinal);
+                return new DuffCode(suggestedFont, (char)ordinal); // FIXME: don't create a new one each time; it wastes heap
         } else if (ordinal == (int)'\t') {
             if (0 == suggestedFont)
                 return TMW_tab;
             else
-                return new DuffCode(suggestedFont, (char)ordinal);
+                return new DuffCode(suggestedFont, (char)ordinal); // FIXME: don't create a new one each time; it wastes heap
         } else {
             // for robustness, just return font 1, char ordinal.
             ThdlDebug.noteIffyCode();
             return null;
         }
     }
-    if (0 != suggestedFont && 32 == ordinal || 45 == ordinal) {
+    if ((0 != suggestedFont) && (32 == ordinal || 45 == ordinal)) {
-        return new DuffCode(suggestedFont, (char)ordinal);
+        return new DuffCode(suggestedFont, (char)ordinal); // FIXME: don't create a new one each time; it wastes heap
     }
 	return TMtoTMW[font][ordinal-32];
 }
@@ -1061,25 +1075,25 @@ public static DuffCode mapTMWtoTM(int font, int ordinal, int suggestedFont) {
             if (0 == suggestedFont)
                 return TM_cr;
             else
-                return new DuffCode(suggestedFont, (char)ordinal);
+                return new DuffCode(suggestedFont, (char)ordinal); // FIXME: don't create a new one each time; it wastes heap
         } else if (ordinal == (int)'\n') {
             if (0 == suggestedFont)
                 return TM_lf;
             else
-                return new DuffCode(suggestedFont, (char)ordinal);
+                return new DuffCode(suggestedFont, (char)ordinal); // FIXME: don't create a new one each time; it wastes heap
         } else if (ordinal == (int)'\t') {
             if (0 == suggestedFont)
                 return TM_tab;
             else
-                return new DuffCode(suggestedFont, (char)ordinal);
+                return new DuffCode(suggestedFont, (char)ordinal); // FIXME: don't create a new one each time; it wastes heap
         } else {
             // for robustness, just return font 1, char ordinal.
             ThdlDebug.noteIffyCode();
             return null;
         }
     }
-    if (0 != suggestedFont && 32 == ordinal || 45 == ordinal) {
+    if ((0 != suggestedFont) && (32 == ordinal || 45 == ordinal)) {
-        return new DuffCode(suggestedFont, (char)ordinal);
+        return new DuffCode(suggestedFont, (char)ordinal); // FIXME: don't create a new one each time; it wastes heap
     }
     DuffCode ans = TMWtoTM[font][ordinal-32];
 	return ans;