write FractionalUCA with blanked weights

Author: markusicu

docs/uca/index.md

@@ -115,10 +115,19 @@ Judgment call. See Cherokee, Deseret, Osage, Vithkuqi for examples.
 
 After running the tool, diff the main mapping file and look for bad changes
 (for example, more bytes per weight for common characters).
+
+A good way for checking changes to both
+the sort order and the number of bytes in collation elements
+is to "blank" most of the non-zero weights in FractionalUCA.txt.
+There is a Python script for that (see below).
+Starting in 2025-may, the output of the Java tool directly includes
+FractionalUCA_blanked.txt which is nearly identical to running the Python script
+over FractionalUCA.txt.
+
 ```
-~/unitools/mine/src$ sed -r -f ~/cldr/uni/src/tools/scripts/uca/blankweights.sed ~/cldr/uni/src/common/uca/FractionalUCA.txt > ../frac-14.0.txt
-~/unitools/mine/src$ sed -r -f ~/cldr/uni/src/tools/scripts/uca/blankweights.sed ../Generated/UCA/15.0.0/CollationAuxiliary/FractionalUCA.txt > ../frac-15.0.txt
-~/unitools/mine/src$ meld ../frac-14.0.txt ../frac-15.0.txt
+~/unitools/mine/src$ sed -r -f ~/cldr/uni/src/tools/scripts/uca/blankweights.sed ~/cldr/uni/src/common/uca/FractionalUCA.txt > ../frac-16.0.txt
+~/unitools/mine/src$ sed -r -f ~/cldr/uni/src/tools/scripts/uca/blankweights.sed ../Generated/UCA/17.0.0/CollationAuxiliary/FractionalUCA.txt > ../frac-17.0.txt
+~/unitools/mine/src$ meld ../frac-16.0.txt ../frac-17.0.txt
 ```
 
 CLDR root data files are checked into $CLDR_SRC/common/uca/

unicodetools/src/main/java/org/unicode/text/UCA/CEList.java

@@ -302,18 +302,30 @@ public static byte remap(int ch, byte type, int t) {
         return (byte) t;
     }
 
+    enum ToString {
+        DEFAULT,
+        NO_SPACES,
+        BLANKED
+    }
+
     @Override
     public String toString() {
+        return toString(ToString.DEFAULT);
+    }
+
+    String toString(ToString option) {
         if (isEmpty()) {
             return toString(0);
         }
 
         final StringBuilder result = new StringBuilder();
         for (int i = startOffset; i < endOffset; ++i) {
-            if (i != startOffset) {
+            if (i != startOffset && option == ToString.DEFAULT) {
                 result.append(' ');
             }
-            result.append(toString(contents[i]));
+            int ce = contents[i];
+            String s = option == ToString.BLANKED ? toBlankedString(ce) : toString(ce);
+            result.append(s);
         }
         return result.toString();
     }
@@ -355,45 +367,21 @@ public static String toString(int ce) {
                 + Utility.hex(getSecondary(ce))
                 + "."
                 + Utility.hex(getTertiary(ce))
-                + "]"
-        // + "(" + NAME3[getTertiary(ce)] + ")"
-        ;
-    }
-
-    static final String[] NAME3 = {
-        "IGNORE", // 0
-        "BLK", // Unused?
-        "MIN",
-        "WIDE",
-        "COMPAT",
-        "FONT",
-        "CIRCLE",
-        "RES-2",
-        "CAP",
-        "WIDECAP",
-        "COMPATCAP",
-        "FONTCAP",
-        "CIRCLECAP",
-        "HIRA-SMALL",
-        "HIRA",
-        "SMALL",
-        "SMALL-NARROW",
-        "KATA",
-        "NARROW",
-        "CIRCLE-KATA",
-        "SUP-MNN",
-        "SUB-MNS",
-        "VERT", // Missing??
-        "AINI",
-        "AMED",
-        "AFIN",
-        "AISO",
-        "NOBREAK", // Missing?
-        "SQUARED",
-        "SQUAREDCAP",
-        "FRACTION",
-        "MAX"
-    };
+                + "]";
+    }
+
+    static String toBlankedString(int ce) {
+        char p = getPrimary(ce);
+        char s = getSecondary(ce);
+        // Tertiary allkeys weights are fixed; do not blank them.
+        return "["
+                + (p == 0 ? "0000" : "pppp")
+                + "."
+                + (s == 0 ? "0000" : "ssss")
+                + "."
+                + Utility.hex(getTertiary(ce))
+                + "]";
+    }
 
     public boolean containsZero() {
         for (int i = startOffset; i < endOffset; ++i) {

unicodetools/src/main/java/org/unicode/text/UCA/Fractional.java

@@ -238,26 +238,39 @@ public static WeightIterator assignTertiaryWeightsForTertiaryCEs(int numWeights)
                 FIRST_IGNORABLE_TER_ASSIGNED, FIRST_IGNORABLE_TER_RESERVED, 0x40, numWeights, 20);
     }
 
-    public static void hexBytes(long x, StringBuffer result) {
+    static StringBuilder hexBytes(long x, StringBuilder result) {
         final int oldLength = result.length();
-        // byte lastb = 1;
         for (int shift = 24; shift >= 0; shift -= 8) {
             final byte b = (byte) (x >>> shift);
             if (b != 0) {
                 if (result.length() != oldLength) {
                     result.append(" ");
                 }
                 result.append(Utility.hex(b));
-                // if (lastb == 0) System.err.println(" bad zero byte: " + result);
             }
-            // lastb = b;
         }
+        return result;
     }
 
-    public static String hexBytes(long x) {
-        final StringBuffer temp = new StringBuffer();
-        hexBytes(x, temp);
-        return temp.toString();
+    static StringBuilder blankedHexBytes(long x, String blanked, StringBuilder result) {
+        if (result == null) {
+            result = new StringBuilder();
+        }
+        final int oldLength = result.length();
+        for (int shift = 24; shift >= 0; shift -= 8) {
+            final byte b = (byte) (x >>> shift);
+            if (b != 0) {
+                if (result.length() != oldLength) {
+                    result.append(" ");
+                }
+                result.append(blanked);
+            }
+        }
+        return result;
+    }
+
+    static String hexBytes(long x) {
+        return hexBytes(x, new StringBuilder()).toString();
     }
 
     /* package */ static short getFixedScript(int ch) {