CLDR-7428 Freeze collators; new class CollatorHelper

-Public CollatorHelper replaces old private versions in CLDRConfig and MapComparator -New CollatorHelper.ROOT_ORDER, CollatorHelper.ROOT_NUMERIC, CollatorHelper.EMOJI_COLLATOR -Use those collators everywhere possible instead of recreating -Fix some mistakes in previous commits where freeze was called too soon -Remove dead code
unicode-org · Nov 19, 2024 · 5db775f · 5db775f
1 parent f03028c
commit 5db775f
Show file tree

Hide file tree

Showing 18 changed files with 58 additions and 176 deletions.
diff --git a/tools/cldr-apps/src/test/java/org/unicode/cldr/unittest/web/TestAll.java b/tools/cldr-apps/src/test/java/org/unicode/cldr/unittest/web/TestAll.java
@@ -7,22 +7,15 @@
 
 import com.ibm.icu.dev.test.TestFmwk.TestGroup;
 import com.ibm.icu.dev.test.TestLog;
-import com.ibm.icu.text.Collator;
-import com.ibm.icu.text.RuleBasedCollator;
-import com.ibm.icu.util.ULocale;
 import java.io.File;
 import java.io.PrintWriter;
 import java.sql.SQLException;
 import java.util.logging.Logger;
 import org.unicode.cldr.test.CheckCLDR;
 import org.unicode.cldr.util.CLDRConfig;
 import org.unicode.cldr.util.CLDRConfig.Environment;
-import org.unicode.cldr.util.CLDRFile;
 import org.unicode.cldr.util.CLDRPaths;
 import org.unicode.cldr.util.CldrUtility;
-import org.unicode.cldr.util.Factory;
-import org.unicode.cldr.util.StandardCodes;
-import org.unicode.cldr.util.SupplementalDataInfo;
 import org.unicode.cldr.web.CLDRProgressIndicator;
 import org.unicode.cldr.web.DBUtils;
 import org.unicode.cldr.web.SurveyLog;
@@ -129,13 +122,6 @@ public TestAll() {
     public static class WebTestInfo {
         private static WebTestInfo INSTANCE = null;
 
-        private SupplementalDataInfo supplementalDataInfo;
-        private StandardCodes sc;
-        private Factory cldrFactory;
-        private CLDRFile english;
-        private CLDRFile root;
-        private RuleBasedCollator col;
-
         public static WebTestInfo getInstance() {
             synchronized (WebTestInfo.class) {
                 if (INSTANCE == null) {
@@ -146,62 +132,6 @@ public static WebTestInfo getInstance() {
         }
 
         private WebTestInfo() {}
-
-        public SupplementalDataInfo getSupplementalDataInfo() {
-            synchronized (this) {
-                if (supplementalDataInfo == null) {
-                    supplementalDataInfo =
-                            SupplementalDataInfo.getInstance(CLDRPaths.SUPPLEMENTAL_DIRECTORY);
-                }
-            }
-            return supplementalDataInfo;
-        }
-
-        public StandardCodes getStandardCodes() {
-            synchronized (this) {
-                if (sc == null) {
-                    sc = StandardCodes.make();
-                }
-            }
-            return sc;
-        }
-
-        public Factory getCldrFactory() {
-            synchronized (this) {
-                if (cldrFactory == null) {
-                    cldrFactory = Factory.make(CLDRPaths.MAIN_DIRECTORY, ".*");
-                }
-            }
-            return cldrFactory;
-        }
-
-        public CLDRFile getEnglish() {
-            synchronized (this) {
-                if (english == null) {
-                    english = getCldrFactory().make("en", true);
-                }
-            }
-            return english;
-        }
-
-        public CLDRFile getRoot() {
-            synchronized (this) {
-                if (root == null) {
-                    root = getCldrFactory().make("root", true);
-                }
-            }
-            return root;
-        }
-
-        public Collator getCollator() {
-            synchronized (this) {
-                if (col == null) {
-                    col = (RuleBasedCollator) Collator.getInstance(ULocale.ROOT);
-                    col.setNumericCollation(true);
-                }
-            }
-            return col;
-        }
     }
 
     static boolean dbSetup = false;

diff --git a/tools/cldr-code/src/main/java/org/unicode/cldr/icu/ExtractICUData.java b/tools/cldr-code/src/main/java/org/unicode/cldr/icu/ExtractICUData.java
@@ -12,9 +12,7 @@
 import com.ibm.icu.lang.UCharacter;
 import com.ibm.icu.lang.UProperty;
 import com.ibm.icu.text.Collator;
-import com.ibm.icu.text.RuleBasedCollator;
 import com.ibm.icu.text.Transliterator;
-import com.ibm.icu.util.ULocale;
 import com.ibm.icu.util.UResourceBundle;
 import java.io.BufferedReader;
 import java.io.File;
@@ -35,6 +33,7 @@
 import org.unicode.cldr.util.CLDRFile;
 import org.unicode.cldr.util.CLDRPaths;
 import org.unicode.cldr.util.CldrUtility;
+import org.unicode.cldr.util.CollatorHelper;
 import org.unicode.cldr.util.PathUtilities;
 import org.unicode.cldr.util.PatternCache;
 import org.unicode.cldr.util.SimpleFactory;
@@ -411,9 +410,7 @@ static void testProps() {
             {UProperty.DOUBLE_START, UProperty.DOUBLE_START},
             {UProperty.STRING_START, UProperty.STRING_LIMIT},
         };
-        Collator col = Collator.getInstance(ULocale.ROOT); // freeze below
-        ((RuleBasedCollator) col).setNumericCollation(true);
-        col = col.freeze();
+        Collator col = CollatorHelper.ROOT_NUMERIC;
         Map<String, Set<String>> alpha = new TreeMap<>(col);
 
         for (int range = 0; range < ranges.length; ++range) {
@@ -466,13 +463,6 @@ static void testProps() {
             }
             out.println("</table></td></tr>");
         }
-        Collator c = Collator.getInstance(ULocale.ROOT); // freeze below
-        ((RuleBasedCollator) c).setNumericCollation(true);
-        c = c.freeze();
-
-        // int enumValue = UCharacter.getIntPropertyValue(codePoint, propEnum);
-        // return UCharacter.getPropertyValueName(propEnum,enumValue, (int)nameChoice);
-
     }
 
     private static String getName(int index, String valueName, String shortValueName) {

diff --git a/tools/cldr-code/src/main/java/org/unicode/cldr/tool/CLDRModify.java b/tools/cldr-code/src/main/java/org/unicode/cldr/tool/CLDRModify.java
@@ -3309,8 +3309,9 @@ private static int stepsFromRoot(String origLoc) {
 
     /** Internal */
     public static void testJavaSemantics() {
-        Collator caseInsensitive = Collator.getInstance(ULocale.ROOT).freeze();
+        Collator caseInsensitive = Collator.getInstance(ULocale.ROOT); // freeze below
         caseInsensitive.setStrength(Collator.SECONDARY);
+        caseInsensitive = caseInsensitive.freeze();
         Set<String> setWithCaseInsensitive = new TreeSet<>(caseInsensitive);
         setWithCaseInsensitive.addAll(Arrays.asList(new String[] {"a", "b", "c"}));
         Set<String> plainSet = new TreeSet<>();

diff --git a/tools/cldr-code/src/main/java/org/unicode/cldr/tool/ChartCollation.java b/tools/cldr-code/src/main/java/org/unicode/cldr/tool/ChartCollation.java
@@ -258,8 +258,6 @@ private void addCollator(Map<String, Data> data, String type, RuleBasedCollator
         dataItem.collator = col;
     }
 
-    // RuleBasedCollator ROOT = (RuleBasedCollator) Collator.getInstance(ULocale.ROOT).freeze();
-
     private class Subchart extends Chart {
         private static final String HIGH_COLLATION_PRIMARY = "\uFFFF";
         String title;

diff --git a/tools/cldr-code/src/main/java/org/unicode/cldr/tool/ChartLanguageGroups.java b/tools/cldr-code/src/main/java/org/unicode/cldr/tool/ChartLanguageGroups.java
@@ -3,8 +3,6 @@
 import com.google.common.collect.ImmutableSet;
 import com.google.common.collect.ImmutableSet.Builder;
 import com.google.common.collect.Multimap;
-import com.ibm.icu.text.Collator;
-import com.ibm.icu.util.ULocale;
 import java.io.IOException;
 import java.util.Collection;
 import java.util.Comparator;
@@ -76,8 +74,6 @@ public String getExplanation() {
                 + "The data doesn't completely match wikipedia’s; there are some patches for CLDR languages.</p>\n";
     }
 
-    Collator ENGLISH_ORDER = Collator.getInstance(ULocale.ROOT).freeze();
-
     @Override
     public void writeContents(FormattedFileWriter pw) throws IOException {
 
@@ -112,7 +108,9 @@ private void show(Multimap<String, String> lg, String parent, TablePrinter table
                         new Comparator<Pair<String, String>>() {
                             @Override
                             public int compare(Pair<String, String> o1, Pair<String, String> o2) {
-                                int diff = ENGLISH_ORDER.compare(o1.getFirst(), o2.getFirst());
+                                int diff =
+                                        CollatorHelper.ROOT_ORDER.compare(
+                                                o1.getFirst(), o2.getFirst());
                                 if (diff != 0) {
                                     return diff;
                                 }

diff --git a/tools/cldr-code/src/main/java/org/unicode/cldr/tool/CompareData.java b/tools/cldr-code/src/main/java/org/unicode/cldr/tool/CompareData.java
@@ -1,9 +1,6 @@
 package org.unicode.cldr.tool;
 
 import com.ibm.icu.dev.util.UOption;
-import com.ibm.icu.text.Collator;
-import com.ibm.icu.text.RuleBasedCollator;
-import com.ibm.icu.util.ULocale;
 import java.util.HashSet;
 import java.util.Iterator;
 import java.util.Set;
@@ -30,14 +27,6 @@ public class CompareData {
 
     String[] directoryList = {"main", "collation", "segmentations"};
 
-    static RuleBasedCollator uca =
-            (RuleBasedCollator) Collator.getInstance(ULocale.ROOT); // freeze below
-
-    {
-        uca.setNumericCollation(true);
-        uca = (RuleBasedCollator) uca.freeze();
-    }
-
     static PrettyPath prettyPathMaker = new PrettyPath();
     static CLDRFile english;
     static Set<String> locales;

diff --git a/tools/cldr-code/src/main/java/org/unicode/cldr/tool/ConvertLanguageData.java b/tools/cldr-code/src/main/java/org/unicode/cldr/tool/ConvertLanguageData.java
@@ -6,9 +6,7 @@
 import com.ibm.icu.impl.Relation;
 import com.ibm.icu.impl.Row;
 import com.ibm.icu.impl.Row.R2;
-import com.ibm.icu.text.Collator;
 import com.ibm.icu.text.NumberFormat;
-import com.ibm.icu.text.RuleBasedCollator;
 import com.ibm.icu.text.UTF16;
 import com.ibm.icu.util.ULocale;
 import java.io.BufferedReader;
@@ -42,6 +40,7 @@
 import org.unicode.cldr.util.CLDRFile;
 import org.unicode.cldr.util.CLDRPaths;
 import org.unicode.cldr.util.CldrUtility;
+import org.unicode.cldr.util.CollatorHelper;
 import org.unicode.cldr.util.Factory;
 import org.unicode.cldr.util.Iso639Data;
 import org.unicode.cldr.util.Iso639Data.Scope;
@@ -1924,12 +1923,6 @@ public String toString() {
 
     public static class GeneralCollator implements Comparator<String> {
         static UTF16.StringComparator cpCompare = new UTF16.StringComparator(true, false, 0);
-        static RuleBasedCollator UCA =
-                (RuleBasedCollator) Collator.getInstance(ULocale.ROOT).freeze();
-
-        static {
-            UCA.setNumericCollation(true);
-        }
 
         @Override
         public int compare(String s1, String s2) {
@@ -1938,7 +1931,7 @@ public int compare(String s1, String s2) {
             } else if (s2 == null) {
                 return 1;
             }
-            int result = UCA.compare(s1, s2);
+            int result = CollatorHelper.ROOT_NUMERIC.compare(s1, s2);
             if (result != 0) return result;
             return cpCompare.compare(s1, s2);
         }

diff --git a/tools/cldr-code/src/main/java/org/unicode/cldr/tool/CountItems.java b/tools/cldr-code/src/main/java/org/unicode/cldr/tool/CountItems.java
@@ -45,6 +45,7 @@
 import org.unicode.cldr.util.CLDRFile;
 import org.unicode.cldr.util.CLDRPaths;
 import org.unicode.cldr.util.CldrUtility;
+import org.unicode.cldr.util.CollatorHelper;
 import org.unicode.cldr.util.Factory;
 import org.unicode.cldr.util.Iso639Data;
 import org.unicode.cldr.util.IsoCurrencyParser;
@@ -318,8 +319,7 @@ public static void genSupplementalZoneData() throws IOException {
     }
 
     public static void genSupplementalZoneData(boolean skipUnaliased) throws IOException {
-        RuleBasedCollator col = (RuleBasedCollator) Collator.getInstance(ULocale.ROOT).freeze();
-        col.setNumericCollation(true);
+        RuleBasedCollator col = (RuleBasedCollator) CollatorHelper.ROOT_NUMERIC;
         StandardCodes sc = StandardCodes.make();
         Map<String, String> zone_country = sc.getZoneToCounty();
         Map<String, Set<String>> country_zone = sc.getCountryToZoneSet();

diff --git a/tools/cldr-code/src/main/java/org/unicode/cldr/tool/GenerateComparison.java b/tools/cldr-code/src/main/java/org/unicode/cldr/tool/GenerateComparison.java
@@ -5,7 +5,6 @@
 import com.ibm.icu.text.Collator;
 import com.ibm.icu.text.NumberFormat;
 import com.ibm.icu.text.UTF16;
-import com.ibm.icu.util.ULocale;
 import java.io.File;
 import java.io.IOException;
 import java.io.PrintWriter;
@@ -18,6 +17,7 @@
 import org.unicode.cldr.util.CLDRFile.Status;
 import org.unicode.cldr.util.CLDRPaths;
 import org.unicode.cldr.util.CldrUtility;
+import org.unicode.cldr.util.CollatorHelper;
 import org.unicode.cldr.util.Counter;
 import org.unicode.cldr.util.EscapingUtilities;
 import org.unicode.cldr.util.Factory;
@@ -30,7 +30,7 @@ public class GenerateComparison {
 
     private static PrettyPath prettyPathMaker;
 
-    private static Collator collator = Collator.getInstance(ULocale.ROOT).freeze();
+    private static Collator collator = CollatorHelper.ROOT_ORDER;
 
     static class EnglishRowComparator implements Comparator<R2<String, String>> {
         private static Comparator<String> unicode = new UTF16.StringComparator(true, false, 0);

diff --git a/tools/cldr-code/src/main/java/org/unicode/cldr/tool/GenerateStatistics.java b/tools/cldr-code/src/main/java/org/unicode/cldr/tool/GenerateStatistics.java
@@ -26,6 +26,7 @@
 import org.unicode.cldr.util.ArrayComparator;
 import org.unicode.cldr.util.CLDRFile;
 import org.unicode.cldr.util.CldrUtility;
+import org.unicode.cldr.util.CollatorHelper;
 import org.unicode.cldr.util.Factory;
 import org.unicode.cldr.util.LanguageTagParser;
 import org.unicode.cldr.util.Log;
@@ -40,7 +41,7 @@ class GenerateStatistics {
     static CLDRFile english;
     static Factory factory;
     static LanguageTagParser ltp = new LanguageTagParser();
-    static Collator col = Collator.getInstance(ULocale.ROOT).freeze();
+    static Collator col = CollatorHelper.ROOT_ORDER;
     static boolean notitlecase = true;
 
     public static void generateSize(
@@ -344,7 +345,7 @@ private static void addCounts(
 
     private static class LanguageList implements Comparable<Object> {
         Object[] contents;
-        static Collator col = Collator.getInstance(ULocale.ROOT).freeze();
+        static Collator col = CollatorHelper.ROOT_ORDER;
         static Comparator<Object[]> comp = new ArrayComparator(new Collator[] {col, col, null});
 
         LanguageList(String locale, String englishName, String localName) {

diff --git a/tools/cldr-code/src/main/java/org/unicode/cldr/tool/GenerateTransformCharts.java b/tools/cldr-code/src/main/java/org/unicode/cldr/tool/GenerateTransformCharts.java
@@ -515,9 +515,11 @@ private static boolean isLatin(String target) {
     static Comparator<String> UCA;
 
     static {
-        RuleBasedCollator UCA2 = (RuleBasedCollator) Collator.getInstance(ULocale.ROOT).freeze();
+        RuleBasedCollator UCA2 =
+                (RuleBasedCollator) Collator.getInstance(ULocale.ROOT); // freeze below
         UCA2.setNumericCollation(true);
         UCA2.setStrength(Collator.IDENTICAL);
+        UCA2 = (RuleBasedCollator) UCA2.freeze();
         UCA =
                 new org.unicode.cldr.util.MultiComparator(
                         UCA2, new UTF16.StringComparator(true, false, 0));

diff --git a/tools/cldr-code/src/main/java/org/unicode/cldr/tool/MakeTransliterator.java b/tools/cldr-code/src/main/java/org/unicode/cldr/tool/MakeTransliterator.java
@@ -23,6 +23,7 @@
 import java.util.TreeSet;
 import org.unicode.cldr.draft.FileUtilities;
 import org.unicode.cldr.util.CldrUtility;
+import org.unicode.cldr.util.CollatorHelper;
 import org.unicode.cldr.util.Pair;
 
 /**
@@ -49,7 +50,7 @@ public class MakeTransliterator {
 
     static NumberFormat nf = NumberFormat.getInstance(ULocale.ENGLISH);
 
-    static Collator col = Collator.getInstance(ULocale.ROOT).freeze();
+    static Collator col = CollatorHelper.ROOT_ORDER;
 
     static String cldrDataDir =
             "C:\\cvsdata\\unicode\\cldr\\tools\\java\\org\\unicode\\cldr\\util\\data\\transforms\\";

diff --git a/tools/cldr-code/src/main/java/org/unicode/cldr/tool/Misc.java b/tools/cldr-code/src/main/java/org/unicode/cldr/tool/Misc.java
@@ -39,6 +39,7 @@
 import org.unicode.cldr.util.CLDRFile;
 import org.unicode.cldr.util.CLDRPaths;
 import org.unicode.cldr.util.CldrUtility;
+import org.unicode.cldr.util.CollatorHelper;
 import org.unicode.cldr.util.Factory;
 import org.unicode.cldr.util.Iso3166Data;
 import org.unicode.cldr.util.LanguageTagParser;
@@ -466,8 +467,7 @@ private static void printCurrentTimezoneLocalizations(Set<String> languages)
     }
 
     static void printZoneAliases() {
-        RuleBasedCollator col = (RuleBasedCollator) Collator.getInstance(ULocale.ROOT).freeze();
-        col.setNumericCollation(true);
+        RuleBasedCollator col = (RuleBasedCollator) CollatorHelper.ROOT_NUMERIC;
         StandardCodes sc = StandardCodes.make();
         Map<String, String> zone_countries = sc.getZoneToCounty();
         Map<String, String> old_new = sc.getZoneLinkold_new();