One symbol table to rule them all (#1080)

Author: eggrobin

unicodetools/src/main/java/org/unicode/props/UnicodeProperty.java

@@ -32,6 +32,7 @@
 import java.util.TreeMap;
 import java.util.function.Predicate;
 import java.util.regex.Pattern;
+import org.unicode.cldr.util.Rational.RationalParser;
 import org.unicode.cldr.util.props.UnicodeLabel;
 
 public abstract class UnicodeProperty extends UnicodeLabel {
@@ -198,6 +199,7 @@ public UnicodeProperty setDelimiter(String value) {
             EXTENDED_MASK = 1,
             CORE_MASK = ~EXTENDED_MASK,
             BINARY_MASK = (1 << BINARY) | (1 << EXTENDED_BINARY),
+            NUMERIC_MASK = (1 << NUMERIC) | (1 << EXTENDED_NUMERIC),
             STRING_MASK = (1 << STRING) | (1 << EXTENDED_STRING),
             STRING_OR_MISC_MASK =
                     (1 << STRING) | (1 << EXTENDED_STRING) | (1 << MISC) | (1 << EXTENDED_MISC),
@@ -443,19 +445,25 @@ public final UnicodeSet getSet(String propertyValue, UnicodeSet result) {
         if (isMultivalued && propertyValue != null && propertyValue.contains(delimiter)) {
             throw new IllegalArgumentException(
                     "Multivalued property values can't contain the delimiter.");
+        }
+        if (propertyValue == null) {
+            return getSet(NULL_MATCHER, result);
+        }
+        Comparator<String> comparator;
+        if (isType(NUMERIC_MASK)) {
+            // UAX44-LM1.
+            comparator = RATIONAL_COMPARATOR;
+        } else if (getName().equals("Name") || getName().equals("Name_Alias")) {
+            // UAX44-LM2.
+            comparator = CHARACTER_NAME_COMPARATOR;
+        } else if (isType(BINARY_OR_ENUMERATED_OR_CATALOG_MASK)) {
+            // UAX44-LM3
+            comparator = PROPERTY_COMPARATOR;
         } else {
-            return getSet(
-                    propertyValue == null
-                            ? NULL_MATCHER
-                            : new SimpleMatcher(
-                                    propertyValue,
-                                    getName().equals("Name") || getName().equals("Name_Alias")
-                                            ? CHARACTER_NAME_COMPARATOR
-                                            : isType(STRING_OR_MISC_MASK)
-                                                    ? null
-                                                    : PROPERTY_COMPARATOR),
-                    result);
+            // String-valued or Miscellaneous property.
+            comparator = null;
         }
+        return getSet(new SimpleMatcher(propertyValue, comparator), result);
     }
 
     private UnicodeMap<String> unicodeMap = null;
@@ -725,6 +733,26 @@ public static String toSkeleton(String source) {
         return skeletonBuffer.toString();
     }
 
+    public static final Comparator<String> RATIONAL_COMPARATOR =
+            new Comparator<String>() {
+                @Override
+                public int compare(String x, String y) {
+                    return compareRationals(x, y);
+                }
+            };
+
+    public static int compareRationals(String a, String b) {
+        if (a == b) return 0;
+        if (a == null) return -1;
+        if (b == null) return 1;
+        final boolean aIsNaN = equalNames(a, "NaN");
+        final boolean bIsNaN = equalNames(b, "NaN");
+        if (aIsNaN && bIsNaN) return 0;
+        if (aIsNaN) return -1;
+        if (bIsNaN) return 1;
+        return RationalParser.BASIC.parse(a).compareTo(RationalParser.BASIC.parse(b));
+    }
+
     public static final Comparator<String> CHARACTER_NAME_COMPARATOR =
             new Comparator<String>() {
                 @Override