test/jdk/java/util/Locale/LocaleEnhanceTest.java - master - githubmirror/panama-vector

test/jdk/java/util/Locale/LocaleEnhanceTest.java
1 468 строк58 KB
Justin Lu
8387455: Restrict legacy Locale compatibility handling to exact matches
13 июл 2026, 19:23
13 июл 2026, 19:23d343e6c
/*
 * Copyright (c) 2010, 2026, Oracle and/or its affiliates. All rights reserved.
 * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
 *
 * This code is free software; you can redistribute it and/or modify it
 * under the terms of the GNU General Public License version 2 only, as
 * published by the Free Software Foundation.
 *
 * This code is distributed in the hope that it will be useful, but WITHOUT
 * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
 * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License
 * version 2 for more details (a copy is included in the LICENSE file that
 * accompanied this code).
 *
 * You should have received a copy of the GNU General Public License version
 * 2 along with this work; if not, write to the Free Software Foundation,
 * Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA.
 *
 * Please contact Oracle, 500 Oracle Parkway, Redwood Shores, CA 94065 USA
 * or visit www.oracle.com if you need additional information or have any
 * questions.
 */

import java.io.BufferedReader;
import java.io.ByteArrayInputStream;
import java.io.ByteArrayOutputStream;
import java.io.File;
import java.io.FileInputStream;
import java.io.InputStreamReader;
import java.io.ObjectInputStream;
import java.io.ObjectOutputStream;
import java.net.URISyntaxException;
import java.net.URL;
import java.nio.charset.StandardCharsets;
import java.text.DecimalFormatSymbols;
import java.util.Calendar;
import java.util.IllformedLocaleException;
import java.util.List;
import java.util.Locale;
import java.util.Locale.Builder;
import java.util.Set;

import org.junit.jupiter.api.Test;
import org.junit.jupiter.params.ParameterizedTest;
import org.junit.jupiter.params.provider.EmptySource;
import org.junit.jupiter.params.provider.NullSource;
import org.junit.jupiter.params.provider.ValueSource;

import static org.junit.jupiter.api.Assertions.assertDoesNotThrow;
import static org.junit.jupiter.api.Assertions.assertEquals;
import static org.junit.jupiter.api.Assertions.assertNotNull;
import static org.junit.jupiter.api.Assertions.assertNull;
import static org.junit.jupiter.api.Assertions.assertThrows;
import static org.junit.jupiter.api.Assertions.assertTrue;
import static org.junit.jupiter.api.Assertions.fail;

/*
 * @test
 * @bug 6875847 6992272 7002320 7015500 7023613 7032820 7033504 7004603
 *      7044019 8008577 8176853 8255086 8263202 8287868 8174269 8369452
 *      8369590 8387185 8387253 8387455
 * @summary test API changes to Locale
 * @modules jdk.localedata
 * @run junit/othervm -esa LocaleEnhanceTest
 */
public class LocaleEnhanceTest {

    /** A canonical language code. */
    private static final String l = "en";

    /** A canonical script code.. */
    private static final String s = "Latn";

    /** A canonical region code. */
    private static final String c = "US";

    /** A canonical variant code. */
    private static final String v = "NewYork";

    ///
    /// Generic sanity tests
    ///

    /**
     * Ensure that Builder builds locales that have the expected
     * tag and java6 ID.  Note the odd cases for the ID.
     */
    @Test
    public void testCreateLocaleCanonicalValid() {
        String[] valids = {
            "en-Latn-US-NewYork", "en_US_NewYork_#Latn",
            "en-Latn-US", "en_US_#Latn",
            "en-Latn-NewYork", "en__NewYork_#Latn", // double underscore
            "en-Latn", "en__#Latn", // double underscore
            "en-US-NewYork", "en_US_NewYork",
            "en-US", "en_US",
            "en-NewYork", "en__NewYork", // double underscore
            "en", "en",
            "und-Latn-US-NewYork", "_US_NewYork_#Latn",
            "und-Latn-US", "_US_#Latn",
            "und-Latn-NewYork", "", // variant only not supported
            "und-Latn", "",
            "und-US-NewYork", "_US_NewYork",
            "und-US", "_US",
            "und-NewYork", "", // variant only not supported
            "und", ""
        };

        Builder builder = new Builder();

        for (int i = 0; i < valids.length; i += 2) {
            String tag = valids[i];
            String id = valids[i+1];

            String idl = (i & 16) == 0 ? l : "";
            String ids = (i & 8) == 0 ? s : "";
            String idc = (i & 4) == 0 ? c : "";
            String idv = (i & 2) == 0 ? v : "";

            String msg = String.valueOf(i/2) + ": '" + tag + "' ";

            try {
                Locale l = builder
                    .setLanguage(idl)
                    .setScript(ids)
                    .setRegion(idc)
                    .setVariant(idv)
                    .build();
                assertEquals(idl, l.getLanguage(), msg + "language");
                assertEquals(ids, l.getScript(), msg + "script");
                assertEquals(idc, l.getCountry(), msg + "country");
                assertEquals(idv, l.getVariant(), msg + "variant");
                assertEquals(tag, l.toLanguageTag(), msg + "tag");
                assertEquals(id, l.toString(), msg + "id");
            }
            catch (IllegalArgumentException e) {
                fail(msg + e.getMessage());
            }
        }
    }

    /**
     * Test that locale construction works with 'multiple variants'.
     * <p>
     * The string "Newer__Yorker" is treated as three subtags,
     * "Newer", "", and "Yorker", and concatenated into one
     * subtag by omitting empty subtags and joining the remainer
     * with underscores.  So the resulting variant tag is "Newer_Yorker".
     * Note that 'New' and 'York' are invalid BCP47 variant subtags
     * because they are too short.
     */
    @Test
    public void testCreateLocaleMultipleVariants() {

        String[] valids = {
            "en-Latn-US-Newer-Yorker",  "en_US_Newer_Yorker_#Latn",
            "en-Latn-Newer-Yorker",     "en__Newer_Yorker_#Latn",
            "en-US-Newer-Yorker",       "en_US_Newer_Yorker",
            "en-Newer-Yorker",          "en__Newer_Yorker",
            "und-Latn-US-Newer-Yorker", "_US_Newer_Yorker_#Latn",
            "und-Latn-Newer-Yorker",    "",
            "und-US-Newer-Yorker",      "_US_Newer_Yorker",
            "und-Newer-Yorker",         "",
        };

        Builder builder = new Builder(); // lenient variant

        final String idv = "Newer_Yorker";
        for (int i = 0; i < valids.length; i += 2) {
            String tag = valids[i];
            String id = valids[i+1];

            String idl = (i & 8) == 0 ? l : "";
            String ids = (i & 4) == 0 ? s : "";
            String idc = (i & 2) == 0 ? c : "";

            String msg = String.valueOf(i/2) + ": " + tag + " ";
            try {
                Locale l = builder
                    .setLanguage(idl)
                    .setScript(ids)
                    .setRegion(idc)
                    .setVariant(idv)
                    .build();

                assertEquals(idl, l.getLanguage(), msg + " language");
                assertEquals(ids, l.getScript(), msg + " script");
                assertEquals(idc, l.getCountry(), msg + " country");
                assertEquals(idv, l.getVariant(), msg + " variant");

                assertEquals(tag, l.toLanguageTag(), msg + "tag");
                assertEquals(id, l.toString(), msg + "id");
            }
            catch (IllegalArgumentException e) {
                fail(msg + e.getMessage());
            }
        }
    }

    /**
     * Ensure that all these invalid formats are not recognized by
     * forLanguageTag.
     */
    @Test
    public void testCreateLocaleCanonicalInvalidSeparator() {
        String[] invalids = {
            // trailing separator
            "en_Latn_US_NewYork_",
            "en_Latn_US_",
            "en_Latn_",
            "en_",
            "_",

            // double separator
            "en_Latn_US__NewYork",
            "_Latn_US__NewYork",
            "en_US__NewYork",
            "_US__NewYork",

            // are these OK?
            // "en_Latn__US_NewYork", // variant is 'US_NewYork'
            // "_Latn__US_NewYork", // variant is 'US_NewYork'
            // "en__Latn_US_NewYork", // variant is 'Latn_US_NewYork'
            // "en__US_NewYork", // variant is 'US_NewYork'

            // double separator without language or script
            "__US",
            "__NewYork",

            // triple separator anywhere except within variant
            "en___NewYork",
            "en_Latn___NewYork",
            "_Latn___NewYork",
            "___NewYork",
        };

        for (int i = 0; i < invalids.length; ++i) {
            String id = invalids[i];
            Locale l = Locale.forLanguageTag(id);
            assertEquals("und", l.toLanguageTag(), id);
        }
    }

    /**
     * Ensure that all current locale ids parse.  Use DateFormat as a proxy
     * for all current locale ids.
     */
    @Test
    public void testCurrentLocales() {
        Locale[] locales = java.text.DateFormat.getAvailableLocales();
        Builder builder = new Builder();

        for (Locale target : locales) {
            String tag = target.toLanguageTag();

            // the tag recreates the original locale,
            // except no_NO_NY
            Locale tagResult = Locale.forLanguageTag(tag);
            if (!target.getVariant().equals("NY")) {
                assertEquals(target, tagResult, "tagResult");
            }

            // the builder also recreates the original locale,
            // except ja_JP_JP, th_TH_TH and no_NO_NY
            Locale builderResult = builder.setLocale(target).build();
            if (target.getVariant().length() != 2) {
                assertEquals(target, builderResult, "builderResult");
            }
        }
    }

    /**
     * Ensure that all icu locale ids parse.
     */
    @Test
    public void testIcuLocales() throws Exception {
        BufferedReader br = new BufferedReader(
            new InputStreamReader(
                LocaleEnhanceTest.class.getResourceAsStream("icuLocales.txt"),
                    StandardCharsets.UTF_8));
        String id = null;
        while (null != (id = br.readLine())) {
            Locale result = Locale.forLanguageTag(id);
            assertEquals(id, result.toLanguageTag(), "ulocale");
        }
    }

    ///
    /// Compatibility tests
    ///

    @Test
    public void testConstructor() {
        // all the old weirdness still holds, no new weirdness
        String[][] tests = {
                // language to lower case, region to upper, variant unchanged
                // short
                {"X", "y", "z", "x", "Y"},
                // long
                {"xXxXxXxXxXxX", "yYyYyYyYyYyYyYyY", "zZzZzZzZzZzZzZzZ",
                        "xxxxxxxxxxxx", "YYYYYYYYYYYYYYYY"},
                // mapped language ids
                {"he", "IL", "", "he"},
                {"iw", "IL", "", "he"},
                {"yi", "DE", "", "yi"},
                {"ji", "DE", "", "yi"},
                {"id", "ID", "", "id"},
                {"in", "ID", "", "id"},
                // special variants
                {"ja", "JP", "JP"},
                {"th", "TH", "TH"},
                {"no", "NO", "NY"},
                {"no", "NO", "NY"},
                // no canonicalization of 3-letter language codes
                {"eng", "US", ""}
        };
        for (int i = 0; i < tests.length; ++i) {
            String[] test = tests[i];
            String id = String.valueOf(i);
            Locale locale = Locale.of(test[0], test[1], test[2]);
            assertEquals(test.length > 3 ? test[3] : test[0], locale.getLanguage(), id + " lang");
            assertEquals(test.length > 4 ? test[4] : test[1], locale.getCountry(), id + " region");
            assertEquals(test.length > 5 ? test[5] : test[2], locale.getVariant(), id + " variant");
        }
    }

    ///
    /// Locale API Tests
    ///

    @Test
    public void testGetScript() {
        // forLanguageTag normalizes case
        Locale locale = Locale.forLanguageTag("und-latn");
        assertEquals("Latn", locale.getScript(), "forLanguageTag");

        // Builder normalizes case
        locale = new Builder().setScript("LATN").build();
        assertEquals("Latn", locale.getScript(), "builder");

        // empty string is returned, not null, if there is no script
        locale = Locale.forLanguageTag("und");
        assertEquals("", locale.getScript(), "script is empty string");
    }

    @Test
    public void testGetExtension() {
        // forLanguageTag does NOT normalize to hyphen
        Locale locale = Locale.forLanguageTag("und-a-some_ex-tension");
        assertNull(locale.getExtension('a'), "some_ex-tension");

        // regular extension
        locale = new Builder().setExtension('a', "some-ex-tension").build();
        assertEquals("some-ex-tension", locale.getExtension('a'), "builder");

        // returns null if extension is not present
        assertNull(locale.getExtension('b'), "empty b");

        // throws exception if extension tag is illegal
        assertThrows(IllegalArgumentException.class, () -> Locale.forLanguageTag("").getExtension('\uD800'));

        // 'x' is not an extension, it's a private use tag, but it's accessed through this API
        locale = Locale.forLanguageTag("x-y-z-blork");
        assertEquals("y-z-blork", locale.getExtension('x'), "x");
    }

    @Test
    public void testGetExtensionKeys() {
        Locale locale = Locale.forLanguageTag("und-a-xx-yy-b-zz-ww");
        Set<Character> result = locale.getExtensionKeys();
        assertEquals(2, result.size(), "result size");
        assertTrue(result.contains('a') && result.contains('b'), "'a','b'");

        // result is not mutable
        assertThrows(UnsupportedOperationException.class, () -> result.add('x'));

        // returns empty set if no extensions
        locale = Locale.forLanguageTag("und");
        assertTrue(locale.getExtensionKeys().isEmpty(), "empty result");
    }

    @Test
    public void testGetUnicodeLocaleAttributes() {
        Locale locale = Locale.forLanguageTag("en-US-u-abc-def");
        Set<String> attributes = locale.getUnicodeLocaleAttributes();
        assertEquals(2, attributes.size(), "number of attributes");
        assertTrue(attributes.contains("abc"), "attribute abc");
        assertTrue(attributes.contains("def"), "attribute def");

        locale = Locale.forLanguageTag("en-US-u-ca-gregory");
        attributes = locale.getUnicodeLocaleAttributes();
        assertTrue(attributes.isEmpty(), "empty attributes");
    }

    @Test
    public void testGetUnicodeLocaleType() {
        Locale locale = Locale.forLanguageTag("und-u-co-japanese-nu-thai");
        assertEquals("japanese", locale.getUnicodeLocaleType("co"), "collation");
        assertEquals("thai", locale.getUnicodeLocaleType("nu"), "numbers");

        // Unicode locale extension key is case insensitive
        assertEquals("japanese", locale.getUnicodeLocaleType("Co"), "key case");

        // if keyword is not present, returns null
        assertNull(locale.getUnicodeLocaleType("xx"), "locale keyword not present");

        // if no locale extension is set, returns null
        locale = Locale.forLanguageTag("und");
        assertNull(locale.getUnicodeLocaleType("co"), "locale extension not present");

        // typeless keyword
        locale = Locale.forLanguageTag("und-u-kn");
        assertEquals("", locale.getUnicodeLocaleType("kn"), "typeless keyword");

        // invalid keys throw exception
        assertThrows(IllegalArgumentException.class, () -> Locale.forLanguageTag("").getUnicodeLocaleType("q"));
        assertThrows(IllegalArgumentException.class, () -> Locale.forLanguageTag("").getUnicodeLocaleType("abcdefghi"));

        // null argument throws exception
        assertThrows(NullPointerException.class, () -> Locale.forLanguageTag("").getUnicodeLocaleType(null));
    }

    @Test
    public void testGetUnicodeLocaleKeys() {
        Locale locale = Locale.forLanguageTag("und-u-co-japanese-nu-thai");
        Set<String> result = locale.getUnicodeLocaleKeys();
        assertEquals(2, result.size(), "two keys");
        assertTrue(result.contains("co") && result.contains("nu"), "co and nu");

        // result is not modifiable
        assertThrows(UnsupportedOperationException.class, () -> result.add("frobozz"));
    }

    @Test
    public void testPrivateUseExtension() {
        Locale locale = Locale.forLanguageTag("x-y-x-blork-");
        assertEquals("y-x-blork", locale.getExtension(Locale.PRIVATE_USE_EXTENSION), "blork");

        locale = Locale.forLanguageTag("und");
        assertNull(locale.getExtension(Locale.PRIVATE_USE_EXTENSION), "no privateuse");
    }

    @Test
    public void testToLanguageTag() {
        // lots of normalization to test here
        // test locales created using the constructor
        String[][] tests = {
                // empty locale canonicalizes to 'und'
                {"", "", "", "und"},
                // variant alone is not a valid Locale, but has a valid language tag
                {"", "", "NewYork", "und-NewYork"},
                // standard valid locales
                {"", "Us", "", "und-US"},
                {"", "US", "NewYork", "und-US-NewYork"},
                {"EN", "", "", "en"},
                {"EN", "", "NewYork", "en-NewYork"},
                {"EN", "US", "", "en-US"},
                {"EN", "US", "NewYork", "en-US-NewYork"},
                // underscore in variant will be emitted as multiple variant subtags
                {"en", "US", "Newer_Yorker", "en-US-Newer-Yorker"},
                // invalid variant subtags are appended as private use
                {"en", "US", "new_yorker", "en-US-x-lvariant-new-yorker"},
                // the first invalid variant subtags and following variant subtags are appended as private use
                {"en", "US", "Windows_XP_Home", "en-US-Windows-x-lvariant-XP-Home"},
                // too long variant and following variant subtags disappear
                {"en", "US", "WindowsVista_SP2", "en-US"},
                // invalid region subtag disappears
                {"en", "USA", "", "en"},
                // invalid language tag disappears
                {"e", "US", "", "und-US"},
                // three-letter language tags are not canonicalized
                {"Eng", "", "", "eng"},
                // legacy languages canonicalize to modern equivalents
                {"he", "IL", "", "he-IL"},
                {"iw", "IL", "", "he-IL"},
                {"yi", "DE", "", "yi-DE"},
                {"ji", "DE", "", "yi-DE"},
                {"id", "ID", "", "id-ID"},
                {"in", "ID", "", "id-ID"},
                // special values are converted on output
                {"ja", "JP", "JP", "ja-JP-u-ca-japanese-x-lvariant-JP"},
                {"th", "TH", "TH", "th-TH-u-nu-thai-x-lvariant-TH"},
                {"no", "NO", "NY", "nn-NO"}
        };
        for (int i = 0; i < tests.length; ++i) {
            String[] test = tests[i];
            Locale locale = Locale.of(test[0], test[1], test[2]);
            assertEquals(test[3], locale.toLanguageTag(), "case " + i);
        }

        // test locales created from forLanguageTag
        String[][] tests1 = {
                // case is normalized during the round trip
                {"EN-us", "en-US"},
                {"en-Latn-US", "en-Latn-US"},
                // reordering Unicode locale extensions
                {"de-u-co-phonebk-ca-gregory", "de-u-ca-gregory-co-phonebk"},
                // private use only language tag is preserved (no extra "und")
                {"x-elmer", "x-elmer"},
                {"x-lvariant-JP", "x-lvariant-JP"},
                // Legacy locale cases
                // no/NO/NY case is normalized during `toLanguageTag`
                // ja/JP/JP & th/TH/TH case is normalized during `forLanguageTag`
                // Script prevents the legacy conversions
                {"no-Latn-NO-x-lvariant-NY",
                        "no-Latn-NO-x-lvariant-NY"},
                {"ja-Jpan-JP-x-lvariant-JP",
                        "ja-Jpan-JP-x-lvariant-JP"},
                {"th-Thai-TH-x-lvariant-TH",
                        "th-Thai-TH-x-lvariant-TH"},
                // Unexpected extensions prevent the legacy conversions
                {"no-NO-a-foo-x-lvariant-NY",
                        "no-NO-a-foo-x-lvariant-NY"},
                {"ja-JP-a-foo-x-lvariant-JP",
                        "ja-JP-a-foo-x-lvariant-JP"},
                {"th-TH-a-foo-x-lvariant-TH",
                        "th-TH-a-foo-x-lvariant-TH"},
        };
        for (String[] test : tests1) {
            Locale locale = Locale.forLanguageTag(test[0]);
            assertEquals(test[1], locale.toLanguageTag(), "case " + test[0]);
        }

    }

    @Test
    public void testForLanguageTag() {
        // forLanguageTag implements the 'Language-Tag' production of
        // BCP47, so it handles private use and legacy language tags,
        // unlike locale builder.  Tags listed below (except for the
        // sample private use tags) come from 4646bis Feb 29, 2009.

        String[][] tests = {
                // private use tags only
                {"x-abc", "x-abc"},
                {"x-a-b-c", "x-a-b-c"},
                {"x-a-12345678", "x-a-12345678"},

                // legacy language tags with preferred mappings
                {"i-ami", "ami"},
                {"i-bnn", "bnn"},
                {"i-hak", "hak"},
                {"i-klingon", "tlh"},
                {"i-lux", "lb"}, // two-letter tag
                {"i-navajo", "nv"}, // two-letter tag
                {"i-pwn", "pwn"},
                {"i-tao", "tao"},
                {"i-tay", "tay"},
                {"i-tsu", "tsu"},
                {"art-lojban", "jbo"},
                {"no-bok", "nb"},
                {"no-nyn", "nn"},
                {"sgn-BE-FR", "sfb"},
                {"sgn-BE-NL", "vgt"},
                {"sgn-CH-DE", "sgg"},
                {"zh-guoyu", "cmn"},
                {"zh-hakka", "hak"},
                {"zh-min-nan", "nan"},
                {"zh-xiang", "hsn"},

                // irregular legacy language tags, no preferred mappings, drop illegal fields
                // from end.  If no subtag is mappable, fallback to 'und'
                {"i-default", "en-x-i-default"},
                {"i-enochian", "x-i-enochian"},
                {"i-mingo", "see-x-i-mingo"},
                {"en-GB-oed", "en-GB-x-oed"},
                {"zh-min", "nan-x-zh-min"},
                {"cel-gaulish", "xtg-x-cel-gaulish"},
        };
        for (int i = 0; i < tests.length; ++i) {
            String[] test = tests[i];
            Locale locale = Locale.forLanguageTag(test[0]);
            assertEquals(test[1], locale.toLanguageTag(), "legacy language tag case " + i);
        }

        // forLanguageTag ignores everything past the first place it encounters
        // a syntax error
        tests = new String[][]{
                {"valid",
                        "en-US-Newer-Yorker-a-bb-cc-dd-u-aa-abc-bb-def-x-y-12345678-z",
                        "en-US-Newer-Yorker-a-bb-cc-dd-u-aa-abc-bb-def-x-y-12345678-z"},
                {"segment of private use tag too long",
                        "en-US-Newer-Yorker-a-bb-cc-dd-u-aa-abc-bb-def-x-y-123456789-z",
                        "en-US-Newer-Yorker-a-bb-cc-dd-u-aa-abc-bb-def-x-y"},
                {"segment of private use tag is empty",
                        "en-US-Newer-Yorker-a-bb-cc-dd-u-aa-abc-bb-def-x-y--12345678-z",
                        "en-US-Newer-Yorker-a-bb-cc-dd-u-aa-abc-bb-def-x-y"},
                {"first segment of private use tag is empty",
                        "en-US-Newer-Yorker-a-bb-cc-dd-u-aa-abc-bb-def-x--y-12345678-z",
                        "en-US-Newer-Yorker-a-bb-cc-dd-u-aa-abc-bb-def"},
                {"illegal extension tag",
                        "en-US-Newer-Yorker-a-bb-cc-dd-u-aa-abc-bb-def-\uD800-y-12345678-z",
                        "en-US-Newer-Yorker-a-bb-cc-dd-u-aa-abc-bb-def"},
                {"locale subtag with no value",
                        "en-US-Newer-Yorker-a-bb-cc-dd-u-aa-abc-bb-x-y-12345678-z",
                        "en-US-Newer-Yorker-a-bb-cc-dd-u-aa-abc-bb-x-y-12345678-z"},
                {"locale key subtag invalid",
                        "en-US-Newer-Yorker-a-bb-cc-dd-u-aa-abc-123456789-def-x-y-12345678-z",
                        "en-US-Newer-Yorker-a-bb-cc-dd-u-aa-abc"},
                // locale key subtag invalid in earlier position, all following subtags
                // dropped (and so the locale extension dropped as well)
                {"locale key subtag invalid in earlier position",
                        "en-US-Newer-Yorker-a-bb-cc-dd-u-123456789-abc-bb-def-x-y-12345678-z",
                        "en-US-Newer-Yorker-a-bb-cc-dd"},
        };
        for (int i = 0; i < tests.length; ++i) {
            String[] test = tests[i];
            String msg = "syntax error case " + i + " " + test[0];
            try {
                Locale locale = Locale.forLanguageTag(test[1]);
                assertEquals(test[2], locale.toLanguageTag(), msg);
            } catch (IllegalArgumentException e) {
                fail(msg + " caught exception: " + e);
            }
        }

        // duplicated extension are just ignored
        Locale locale = Locale.forLanguageTag("und-d-aa-00-bb-01-D-AA-10-cc-11-c-1234");
        assertEquals("aa-00-bb-01", locale.getExtension('d'), "extension");
        assertEquals("1234", locale.getExtension('c'), "extension c");

        locale = Locale.forLanguageTag("und-U-ca-gregory-u-ca-japanese");
        assertEquals("ca-gregory", locale.getExtension(Locale.UNICODE_LOCALE_EXTENSION), "Unicode extension");

        // redundant Unicode locale keys in an extension are ignored
        locale = Locale.forLanguageTag("und-u-aa-000-bb-001-bB-002-cc-003-c-1234");
        assertEquals("aa-000-bb-001-cc-003", locale.getExtension(Locale.UNICODE_LOCALE_EXTENSION), "Unicode keywords");
        assertEquals("1234", locale.getExtension('c'), "Duplicated Unicode locake key followed by an extension");
    }

    @Test
    public void testGetDisplayScript() {
        Locale latnLocale = Locale.forLanguageTag("und-latn");
        Locale hansLocale = Locale.forLanguageTag("und-hans");

        Locale oldLocale = Locale.getDefault();

        Locale.setDefault(Locale.US);
        assertEquals("Latin", latnLocale.getDisplayScript(), "latn US");
        assertEquals("Simplified", hansLocale.getDisplayScript(), "hans US");

        Locale.setDefault(Locale.GERMANY);
        assertEquals("Lateinisch", latnLocale.getDisplayScript(), "latn DE");
        assertEquals("Vereinfacht", hansLocale.getDisplayScript(), "hans DE");

        Locale.setDefault(oldLocale);
    }

    @Test
    public void testGetDisplayScriptWithLocale() {
        Locale latnLocale = Locale.forLanguageTag("und-latn");
        Locale hansLocale = Locale.forLanguageTag("und-hans");

        assertEquals("Latin", latnLocale.getDisplayScript(Locale.US), "latn US");
        assertEquals("Simplified", hansLocale.getDisplayScript(Locale.US), "hans US");

        assertEquals("Lateinisch", latnLocale.getDisplayScript(Locale.GERMANY), "latn DE");
        assertEquals("Vereinfacht", hansLocale.getDisplayScript(Locale.GERMANY), "hans DE");
    }

    @Test
    public void testGetDisplayName() {
        final Locale[] testLocales = {
                Locale.ROOT,
                Locale.ENGLISH,
                Locale.US,
                Locale.of("", "US"),
                Locale.of("no", "NO", "NY"),
                Locale.of("", "", "NY"),
                Locale.forLanguageTag("zh-Hans"),
                Locale.forLanguageTag("zh-Hant"),
                Locale.forLanguageTag("zh-Hans-CN"),
                Locale.forLanguageTag("und-Hans"),
        };

        final String[] displayNameEnglish = {
                "",
                "English",
                "English (United States)",
                "United States",
                "Norwegian (Norway, Nynorsk)",
                "Nynorsk",
                "Chinese (Simplified)",
                "Chinese (Traditional)",
                "Chinese (Simplified, China)",
                "Simplified",
        };

        final String[] displayNameSimplifiedChinese = {
                "",
                "\u82f1\u8bed",
                "\u82f1\u8bed (\u7f8e\u56fd)",
                "\u7f8e\u56fd",
                "\u632a\u5a01\u8bed (\u632a\u5a01\uff0cNynorsk)",
                "Nynorsk",
                "\u4e2d\u6587 (\u7b80\u4f53)",
                "\u4e2d\u6587 (\u7e41\u4f53)",
                "\u4e2d\u6587 (\u7b80\u4f53\uff0c\u4e2d\u56fd)",
                "\u7b80\u4f53",
        };

        for (int i = 0; i < testLocales.length; i++) {
            Locale loc = testLocales[i];
            assertEquals(displayNameEnglish[i], loc.getDisplayName(Locale.ENGLISH),
                    "English display name for " + loc.toLanguageTag());
            assertEquals(displayNameSimplifiedChinese[i], loc.getDisplayName(Locale.CHINA),
                    "Simplified Chinese display name for " + loc.toLanguageTag());
        }
    }

    ///
    /// Builder tests
    ///

    @Test
    public void testBuilderSetLocale() {
        Builder builder = new Builder();
        Builder lenientBuilder = new Builder();

        String languageTag = "en-Latn-US-NewYork-a-bb-ccc-u-co-japanese-x-y-z";
        String target = "en-Latn-US-NewYork-a-bb-ccc-u-co-japanese-x-y-z";

        Locale locale = Locale.forLanguageTag(languageTag);
        Locale result = lenientBuilder
                .setLocale(locale)
                .build();
        assertEquals(target, result.toLanguageTag(), "long tag");
        assertEquals(locale, result, "long tag");

        // null is illegal
        assertThrows(NullPointerException.class, () -> builder.setLocale(null),
                "Setting null locale should throw NPE");

        // builder canonicalizes the three legacy locales:
        // ja_JP_JP, th_TH_TH, no_NY_NO.
        locale = builder.setLocale(Locale.of("ja", "JP", "JP")).build();
        assertEquals("ja-JP-u-ca-japanese", locale.toLanguageTag(), "ja_JP_JP languagetag");
        assertEquals("", locale.getVariant(), "ja_JP_JP variant");

        locale = builder.setLocale(Locale.of("th", "TH", "TH")).build();
        assertEquals("th-TH-u-nu-thai", locale.toLanguageTag(), "th_TH_TH languagetag");
        assertEquals("", locale.getVariant(), "th_TH_TH variant");

        locale = builder.setLocale(Locale.of("no", "NO", "NY")).build();
        assertEquals("nn-NO", locale.toLanguageTag(), "no_NO_NY languagetag");
        assertEquals("nn", locale.getLanguage(), "no_NO_NY language");
        assertEquals("", locale.getVariant(), "no_NO_NY variant");

        // Legacy locales that stripped their compatibility extensions are invalid
        assertThrows(IllformedLocaleException.class,
                () -> new Builder().setLocale(Locale.of("ja", "JP", "JP").stripExtensions()));
        assertThrows(IllformedLocaleException.class,
                () -> new Builder().setLocale(Locale.of("th", "TH", "TH").stripExtensions()));

        // Legacy locales without the correct Unicode locale extension value are invalid
        assertThrows(IllformedLocaleException.class,
                () -> new Builder().setLocale(Locale.forLanguageTag("ja-JP-u-ca-foobar-x-lvariant-JP")));
        assertThrows(IllformedLocaleException.class,
                () -> new Builder().setLocale(Locale.forLanguageTag("th-TH-u-nu-foobar-x-lvariant-TH")));

        // Legacy locales with additional extensions are invalid
        assertThrows(IllformedLocaleException.class,
                () -> new Builder().setLocale(Locale.forLanguageTag("no-NO-a-foo-x-lvariant-NY")));
        assertThrows(IllformedLocaleException.class,
                () -> new Builder().setLocale(Locale.forLanguageTag("ja-JP-a-foo-x-lvariant-JP")));
        assertThrows(IllformedLocaleException.class,
                () -> new Builder().setLocale(Locale.forLanguageTag("th-TH-a-foo-x-lvariant-TH")));

        // Legacy locales with non-empty script are invalid
        assertThrows(IllformedLocaleException.class,
                () -> new Builder().setLocale(Locale.forLanguageTag("ja-Jpan-JP-u-ca-japanese-x-lvariant-JP")));
        assertThrows(IllformedLocaleException.class,
                () -> new Builder().setLocale(Locale.forLanguageTag("th-Thai-TH-u-nu-thai-x-lvariant-TH")));
        assertThrows(IllformedLocaleException.class,
                () -> new Builder().setLocale(Locale.forLanguageTag("no-Latn-NO-x-lvariant-NY")));

        // non-canonical, non-legacy locales are invalid
        assertThrows(IllformedLocaleException.class,
                () -> new Builder().setLocale(Locale.of("123", "4567", "89")), "123_4567_89");
    }

    @Test
    public void testBuilderSetLanguageTag() {
        String source = "eN-LaTn-Us-NewYork-A-Xx-B-Yy-X-1-2-3";
        String target = "en-Latn-US-NewYork-a-xx-b-yy-x-1-2-3";
        Builder builder = new Builder();
        String result = builder
                .setLanguageTag(source)
                .build()
                .toLanguageTag();
        assertEquals(target, result, "language");

        // redundant extensions are ignored
        assertEquals("und-a-xx-yy-b-ww-c-vv",
                new Builder().setLanguageTag("und-a-xx-yy-b-ww-A-00-11-c-vv").build().toLanguageTag());
        // redundant Unicode locale extension keys are ignored
        assertEquals("und-u-cu-usd-nu-thai-xx-1234",
                new Builder().setLanguageTag("und-u-nu-thai-cu-usd-NU-chinese-xx-1234").build().toLanguageTag());
        // redundant Unicode locale extension attributes are ignored
        assertEquals("und-u-bar-foo",
                new Builder().setLanguageTag("und-u-foo-bar-FOO").build().toLanguageTag());
    }

    // Test the values that should clear the builder
    @ParameterizedTest
    @NullSource
    @EmptySource
    public void testBuilderSetLanguageTagClear(String tag) {
        var empty = new Builder();
        var bldr = new Builder();
        bldr.setLanguageTag("en-US");
        assertDoesNotThrow(() -> bldr.setLanguageTag(tag));
        assertEquals(empty.build(), bldr.build(),
                "Setting a %s language tag did not clear the builder"
                .formatted(tag == null ? "null" : "empty"));
    }

    @Test
    public void testBuilderSetLanguage() {
        // language is normalized to lower case
        String source = "eN";
        String target = "en";
        String defaulted = "";
        Builder builder = new Builder();
        String result = builder
                .setLanguage(source)
                .build()
                .getLanguage();
        assertEquals(target, result, "en");

        // setting with empty resets
        result = builder
                .setLanguage(target)
                .setLanguage("")
                .build()
                .getLanguage();
        assertEquals(defaulted, result, "empty");

        // setting with null resets too
        result = builder
                .setLanguage(target)
                .setLanguage(null)
                .build()
                .getLanguage();
        assertEquals(defaulted, result, "null");

        // language codes must be 2-8 alpha
        // for forwards compatibility, 4-alpha and 5-8 alpha (registered)
        // languages are accepted syntax
        for (String arg : List.of("q", "abcdefghi", "13")) {
            assertThrows(IllformedLocaleException.class, () -> new Builder().setLanguage(arg));
        }

        // language code validation is NOT performed, any 2-8-alpha passes
        assertNotNull(builder.setLanguage("zz").build(), "2alpha");
        assertNotNull(builder.setLanguage("abcdefgh").build(), "8alpha");

        // three-letter language codes are NOT canonicalized to two-letter
        result = builder
                .setLanguage("eng")
                .build()
                .getLanguage();
        assertEquals("eng", result, "eng");
    }

    @Test
    public void testBuilderSetScript() {
        // script is normalized to title case
        String source = "lAtN";
        String target = "Latn";
        String defaulted = "";
        Builder builder = new Builder();
        String result = builder
                .setScript(source)
                .build()
                .getScript();
        assertEquals(target, result, "script");

        // setting with empty resets
        result = builder
                .setScript(target)
                .setScript("")
                .build()
                .getScript();
        assertEquals(defaulted, result, "empty");

        // settting with null also resets
        result = builder
                .setScript(target)
                .setScript(null)
                .build()
                .getScript();
        assertEquals(defaulted, result, "null");

        // ill-formed script codes throw IAE
        // must be 4alpha
        for (String arg : List.of("abc", "abcde", "l3tn")) {
            assertThrows(IllformedLocaleException.class, () -> new Builder().setScript(arg));
        }


        // script code validation is NOT performed, any 4-alpha passes
        assertEquals("Wxyz", builder.setScript("wxyz").build().getScript(), "4alpha");
    }

    @Test
    public void testBuilderSetRegion() {
        // region is normalized to upper case
        String source = "uS";
        String target = "US";
        String defaulted = "";
        Builder builder = new Builder();
        String result = builder
                .setRegion(source)
                .build()
                .getCountry();
        assertEquals(target, result, "us");

        // setting with empty resets
        result = builder
                .setRegion(target)
                .setRegion("")
                .build()
                .getCountry();
        assertEquals(defaulted, result, "empty");

        // setting with null also resets
        result = builder
                .setRegion(target)
                .setRegion(null)
                .build()
                .getCountry();
        assertEquals(defaulted, result, "null");

        // ill-formed region codes throw IAE
        // 2 alpha or 3 numeric
        for (String arg : List.of("q", "abc", "12", "1234", "a3", "12a")) {
            assertThrows(IllformedLocaleException.class, () -> new Builder().setRegion(arg));
        }

        // region code validation is NOT performed, any 2-alpha or 3-digit passes
        assertEquals("ZZ", builder.setRegion("ZZ").build().getCountry(), "2alpha");
        assertEquals("000", builder.setRegion("000").build().getCountry(), "3digit");
    }

    @Test
    public void testBuilderSetVariant() {
        // Variant case is not normalized in lenient variant mode
        String source = "NewYork";
        String target = source;
        String defaulted = "";
        Builder builder = new Builder();
        String result = builder
                .setVariant(source)
                .build()
                .getVariant();
        assertEquals(target, result, "NewYork");

        result = builder
                .setVariant("NeWeR_YoRkEr")
                .build()
                .toLanguageTag();
        assertEquals("und-NeWeR-YoRkEr", result, "newer yorker");

        // subtags of variant are NOT reordered
        result = builder
                .setVariant("zzzzz_yyyyy_xxxxx")
                .build()
                .getVariant();
        assertEquals("zzzzz_yyyyy_xxxxx", result, "zyx");

        // setting to empty resets
        result = builder
                .setVariant(target)
                .setVariant("")
                .build()
                .getVariant();
        assertEquals(defaulted, result, "empty");

        // setting to null also resets
        result = builder
                .setVariant(target)
                .setVariant(null)
                .build()
                .getVariant();
        assertEquals(defaulted, result, "null");

        // ill-formed variants throw IAE
        // digit followed by 3-7 characters, or alpha followed by 4-8 characters.
        for (String arg : List.of("abcd", "abcdefghi", "1ab", "1abcdefgh")) {
            assertThrows(IllformedLocaleException.class, () -> new Builder().setVariant(arg));
        }


        // 4 characters is ok as long as the first is a digit
        assertEquals("1abc", builder.setVariant("1abc").build().getVariant(), "digit+3alpha");

        // all subfields must conform
        assertThrows(IllformedLocaleException.class, () -> new Builder().setVariant("abcde-fg"));

    }

    @Test
    public void testBuilderSetExtension() {
        // upper case characters are normalized to lower case
        final char sourceKey = 'a';
        final String sourceValue = "aB-aBcdefgh-12-12345678";
        String target = "ab-abcdefgh-12-12345678";
        Builder builder = new Builder();
        String result = builder
                .setExtension(sourceKey, sourceValue)
                .build()
                .getExtension(sourceKey);
        assertEquals(target, result, "extension");

        // setting with empty resets
        result = builder
                .setExtension(sourceKey, sourceValue)
                .setExtension(sourceKey, "")
                .build()
                .getExtension(sourceKey);
        assertNull(result, "empty");

        // setting with null also resets
        result = builder
                .setExtension(sourceKey, sourceValue)
                .setExtension(sourceKey, null)
                .build()
                .getExtension(sourceKey);
        assertNull(result, "null");

        // ill-formed extension keys throw IAE
        // must be in [0-9a-ZA-Z]
        assertThrows(IllformedLocaleException.class,
                () -> new Builder().setExtension('$', sourceValue));


        // each segment of value must be 2-8 alphanum
        assertThrows(IllformedLocaleException.class,
                () -> new Builder().setExtension(sourceKey, "ab-cd-123456789"));


        // no multiple hyphens.
        assertThrows(IllformedLocaleException.class,
                () -> new Builder().setExtension(sourceKey, "ab--cd"));


        // locale extension key has special handling
        Locale locale = builder
                .setExtension('u', "co-japanese")
                .build();
        assertEquals("japanese", locale.getUnicodeLocaleType("co"), "locale extension");

        // locale extension has same behavior with set locale keyword
        Locale locale2 = builder
                .setUnicodeLocaleKeyword("co", "japanese")
                .build();
        assertEquals(locale, locale2, "locales with extension");

        // setting locale extension overrides all previous calls to setLocaleKeyword
        Locale locale3 = builder
                .setExtension('u', "xxx-nu-thai")
                .build();
        assertNull(locale3.getUnicodeLocaleType("co"), "remove co");
        assertEquals("thai", locale3.getUnicodeLocaleType("nu"), "override thai");
        assertEquals(1, locale3.getUnicodeLocaleAttributes().size(), "override attribute");

        // setting locale keyword extends values already set by the locale extension
        Locale locale4 = builder
                .setUnicodeLocaleKeyword("co", "japanese")
                .build();
        assertEquals("japanese", locale4.getUnicodeLocaleType("co"), "extend");
        assertEquals("thai", locale4.getUnicodeLocaleType("nu"), "extend");

        // locale extension subtags are reordered
        result = builder
                .clear()
                .setExtension('u', "456-123-zz-123-yy-456-xx-789")
                .build()
                .toLanguageTag();
        assertEquals("und-u-123-456-xx-789-yy-456-zz-123", result, "reorder");

        // multiple keyword types
        result = builder
                .clear()
                .setExtension('u', "nu-thai-foobar")
                .build()
                .getUnicodeLocaleType("nu");
        assertEquals("thai-foobar", result, "multiple types");

        // redundant locale extensions are ignored
        result = builder
                .clear()
                .setExtension('u', "nu-thai-NU-chinese-xx-1234")
                .build()
                .toLanguageTag();
        assertEquals("und-u-nu-thai-xx-1234", result, "duplicate keys");

        // redundant locale attributes are ignored
        result = builder
                .clear()
                .setExtension('u', "posix-posix")
                .build()
                .toLanguageTag();
        assertEquals("und-u-posix", result, "duplicate attributes");
    }

    @Test
    public void testBuilderAddUnicodeLocaleAttribute() {
        Builder builder = new Builder();
        Locale locale = builder
                .addUnicodeLocaleAttribute("def")
                .addUnicodeLocaleAttribute("abc")
                .build();

        Set<String> uattrs = locale.getUnicodeLocaleAttributes();
        assertEquals(2, uattrs.size(), "number of attributes");
        assertTrue(uattrs.contains("abc"), "attribute abc");
        assertTrue(uattrs.contains("def"), "attribute def");

        // remove attribute
        locale = builder.removeUnicodeLocaleAttribute("xxx")
                .build();

        uattrs = locale.getUnicodeLocaleAttributes();
        assertEquals(2, uattrs.size(), "remove bogus");

        // add duplicate
        locale = builder.addUnicodeLocaleAttribute("abc")
                .build();
        uattrs = locale.getUnicodeLocaleAttributes();
        assertEquals(2, uattrs.size(), "add duplicate");

        // null attribute throws NPE
        assertThrows(NullPointerException.class,
                () ->  new Builder().addUnicodeLocaleAttribute(null), "null attribute");

        assertThrows(NullPointerException.class,
                () -> new Builder().removeUnicodeLocaleAttribute(null), "null attribute removal");

        // illformed attribute throws IllformedLocaleException
        assertThrows(IllformedLocaleException.class,
                     () -> new Builder().addUnicodeLocaleAttribute("ca"), "invalid attribute");
    }

    @Test
    public void testBuildersetUnicodeLocaleKeyword() {
        // Note: most behavior is tested in testBuilderSetExtension
        Builder builder = new Builder();
        Locale locale = builder
                .setUnicodeLocaleKeyword("co", "japanese")
                .setUnicodeLocaleKeyword("nu", "thai")
                .build();
        assertEquals("japanese", locale.getUnicodeLocaleType("co"), "co");
        assertEquals("thai", locale.getUnicodeLocaleType("nu"), "nu");
        assertEquals(2, locale.getUnicodeLocaleKeys().size(), "keys");

        // can clear a keyword by setting to null, others remain
        String result = builder
                .setUnicodeLocaleKeyword("co", null)
                .build()
                .toLanguageTag();
        assertEquals("und-u-nu-thai", result, "empty co");

        // locale keyword extension goes when all keywords are gone
        result = builder
                .setUnicodeLocaleKeyword("nu", null)
                .build()
                .toLanguageTag();
        assertEquals("und", result, "empty nu");

        // locale keywords are ordered independent of order of addition
        result = builder
                .setUnicodeLocaleKeyword("zz", "012")
                .setUnicodeLocaleKeyword("aa", "345")
                .build()
                .toLanguageTag();
        assertEquals("und-u-aa-345-zz-012", result, "reordered");

        // null keyword throws NPE
        assertThrows(NullPointerException.class,
                () -> new Builder().setUnicodeLocaleKeyword(null, "thai"), "keyword");


        // well-formed keywords are two alphanum
        for (String arg : List.of("a", "abc")) {
            assertThrows(IllformedLocaleException.class,
                    () -> new Builder().setUnicodeLocaleKeyword(arg, "value"));
        }

        // well-formed values are 3-8 alphanum
        for (String arg : List.of("ab", "abcdefghi")) {
            assertThrows(IllformedLocaleException.class,
                    () -> new Builder().setUnicodeLocaleKeyword("ab", arg));
        }
    }

    @Test
    public void testBuilderPrivateUseExtension() {
        // normalizes hyphens to underscore, case to lower
        String source = "c-B-a";
        String target = "c-b-a";
        Builder builder = new Builder();
        String result = builder
                .setExtension(Locale.PRIVATE_USE_EXTENSION, source)
                .build()
                .getExtension(Locale.PRIVATE_USE_EXTENSION);
        assertEquals(target, result, "abc");

        // multiple hyphens are ill-formed
        assertThrows(IllformedLocaleException.class,
                () -> new Builder().setExtension(Locale.PRIVATE_USE_EXTENSION, "a--b"),
                "multiple-hyphens should throw IAE");
    }

    @Test
    public void testBuilderClear() {
        String monster = "en-latn-US-NewYork-a-bb-cc-u-co-japanese-x-z-y-x-x";
        Builder builder = new Builder();
        Locale locale = Locale.forLanguageTag(monster);
        String result = builder
                .setLocale(locale)
                .clear()
                .build()
                .toLanguageTag();
        assertEquals("und", result, "clear");
    }

    @Test
    public void testBuilderRemoveUnicodeAttribute() {
        // tested in testBuilderAddUnicodeAttribute
    }

    @Test
    public void testBuilderBuild() {
        // tested in other test methods
    }

    @Test
    public void testSerialize() {
        final Locale[] testLocales = {
                Locale.ROOT,
                Locale.ENGLISH,
                Locale.US,
                Locale.of("en", "US", "Win"),
                Locale.of("en", "US", "Win_XP"),
                Locale.JAPAN,
                Locale.of("ja", "JP", "JP"),
                Locale.of("th", "TH"),
                Locale.of("th", "TH", "TH"),
                Locale.of("no", "NO"),
                Locale.of("nb", "NO"),
                Locale.of("nn", "NO"),
                Locale.of("no", "NO", "NY"),
                Locale.of("nn", "NO", "NY"),
                Locale.of("he", "IL"),
                Locale.of("he", "IL", "var"),
                Locale.of("Language", "Country", "Variant"),
                Locale.of("", "US"),
                Locale.of("", "", "Java"),
                Locale.forLanguageTag("en-Latn-US"),
                Locale.forLanguageTag("zh-Hans"),
                Locale.forLanguageTag("zh-Hant-TW"),
                Locale.forLanguageTag("ja-JP-u-ca-japanese"),
                Locale.forLanguageTag("und-Hant"),
                Locale.forLanguageTag("und-a-123-456"),
                Locale.forLanguageTag("en-x-java"),
                Locale.forLanguageTag("th-TH-u-ca-buddist-nu-thai-x-lvariant-TH"),
        };

        for (Locale locale : testLocales) {
            try {
                // write
                ByteArrayOutputStream bos = new ByteArrayOutputStream();
                ObjectOutputStream oos = new ObjectOutputStream(bos);
                oos.writeObject(locale);

                // read
                ByteArrayInputStream bis = new ByteArrayInputStream(bos.toByteArray());
                ObjectInputStream ois = new ObjectInputStream(bis);
                Object o = ois.readObject();

                assertEquals(locale, o, "roundtrip " + locale);
            } catch (Exception e) {
                fail(locale + " encountered exception:" + e.getLocalizedMessage());
            }
        }
    }

    @Test
    public void testDeserialize6() {
        final String TESTFILEPREFIX = "java6locale_";

        File dataDir = null;
        String dataDirName = System.getProperty("serialized.data.dir");
        if (dataDirName == null) {
            URL resdirUrl = getClass().getClassLoader().getResource("serialized");
            if (resdirUrl != null) {
                try {
                    dataDir = new File(resdirUrl.toURI());
                } catch (URISyntaxException urie) {
                }
            }
        } else {
            dataDir = new File(dataDirName);
        }

        if (dataDir == null) {
            fail("'dataDir' is null. serialized.data.dir Property value is " + dataDirName);
        } else if (!dataDir.isDirectory()) {
            fail("'dataDir' is not a directory. dataDir: " + dataDir.toString());
        }

        File[] files = dataDir.listFiles();
        for (File testfile : files) {
            if (testfile.isDirectory()) {
                continue;
            }
            String name = testfile.getName();
            if (!name.startsWith(TESTFILEPREFIX)) {
                continue;
            }
            Locale locale;
            String locStr = name.substring(TESTFILEPREFIX.length());
            if (locStr.equals("ROOT")) {
                locale = Locale.ROOT;
            } else {
                String[] fields = locStr.split("_", 3);
                String lang = fields[0];
                String country = (fields.length >= 2) ? fields[1] : "";
                String variant = (fields.length == 3) ? fields[2] : "";
                locale = Locale.of(lang, country, variant);
            }

            // deserialize
            try (FileInputStream fis = new FileInputStream(testfile);
                 ObjectInputStream ois = new ObjectInputStream(fis)) {
                Object o = ois.readObject();
                assertEquals(o, locale, "Deserialize Java 6 Locale " + locale);
            } catch (Exception e) {
                fail("Exception while reading " + testfile.getAbsolutePath() + " - " + e.getMessage());
            }
        }
    }

    @Test
    public void testBug7002320() {
        // forLanguageTag() and Builder.setLanguageTag(String)
        // should add a location extension for following two cases.
        //
        // 1. language/country are "ja"/"JP" and the resolved variant (x-lvariant-*)
        //    is exactly "JP" and no BCP 47 extensions are available, then add
        //    a Unicode locale extension "ca-japanese".
        // 2. language/country are "th"/"TH" and the resolved variant is exactly
        //    "TH" and no BCP 47 extensions are available, then add a Unicode locale
        //    extension "nu-thai".
        //
        String[][] testdata = {
                {"ja-JP-x-lvariant-JP", "ja-JP-u-ca-japanese-x-lvariant-JP"},   // special case 1
                {"ja-JP-x-lvariant-JP-XXX"},
                {"ja-JP-u-ca-japanese-x-lvariant-JP"},
                {"ja-JP-u-ca-gregory-x-lvariant-JP"},
                {"ja-JP-u-cu-jpy-x-lvariant-JP"},
                {"ja-x-lvariant-JP"},
                {"th-TH-x-lvariant-TH", "th-TH-u-nu-thai-x-lvariant-TH"},   // special case 2
                {"th-TH-u-nu-thai-x-lvariant-TH"},
                {"en-US-x-lvariant-JP"},
        };

        Builder bldr = new Builder();

        for (String[] data : testdata) {
            String in = data[0];
            String expected = (data.length == 1) ? data[0] : data[1];

            // forLanguageTag
            Locale loc = Locale.forLanguageTag(in);
            String out = loc.toLanguageTag();
            assertEquals(expected, out, "Language tag roundtrip by forLanguageTag with input: " + in);

            // setLanguageTag
            bldr.clear();
            bldr.setLanguageTag(in);
            loc = bldr.build();
            out = loc.toLanguageTag();
            assertEquals(expected, out, "Language tag roundtrip by Builder.setLanguageTag with input: " + in);
        }
    }

    @Test
    public void testBug7023613() {
        String[][] testdata = {
                {"en-Latn", "en__#Latn"},
                {"en-u-ca-japanese", "en__#u-ca-japanese"},
        };

        for (String[] data : testdata) {
            String in = data[0];
            String expected = (data.length == 1) ? data[0] : data[1];

            Locale loc = Locale.forLanguageTag(in);
            String out = loc.toString();
            assertEquals(expected, out, "Empty country field with non-empty script/extension with input: " + in);
        }
    }

    /*
     * 7033504: (lc) incompatible behavior change for ja_JP_JP and th_TH_TH locales
     */
    @Test
    public void testBug7033504() {
        checkCalendar(Locale.of("ja", "JP", "jp"), "java.util.GregorianCalendar");
        checkCalendar(Locale.of("ja", "jp", "jp"), "java.util.GregorianCalendar");
        checkCalendar(Locale.of("ja", "JP", "JP"), "java.util.JapaneseImperialCalendar");
        checkCalendar(Locale.of("ja", "jp", "JP"), "java.util.JapaneseImperialCalendar");
        checkCalendar(Locale.forLanguageTag("en-u-ca-japanese"),
                "java.util.JapaneseImperialCalendar");

        checkDigit(Locale.of("th", "TH", "th"), '0');
        checkDigit(Locale.of("th", "th", "th"), '0');
        checkDigit(Locale.of("th", "TH", "TH"), '\u0e50');
        checkDigit(Locale.of("th", "TH", "TH"), '\u0e50');
        checkDigit(Locale.forLanguageTag("en-u-nu-thai"), '\u0e50');
    }

    // Test that numeric singletons are supported
    @Test
    public void numericSingletonRoundTripTest() {
        var tag = "en-0-foo";
        var value = "foo";
        var singleton = '0';
        // test `forLanguageTag`
        var locale = Locale.forLanguageTag(tag);
        assertEquals(value, locale.getExtension(singleton));
        assertEquals(tag, locale.toLanguageTag());
        // test `Locale.Builder`
        locale = new Builder()
                .setLanguage("en")
                .setExtension(singleton, value)
                .build();
        assertEquals(value, locale.getExtension(singleton));
        assertEquals(tag, locale.toLanguageTag());
    }

    // Ensure that extlang is only accepted after a 2*3ALPHA language subtag
    // That is, the 4 ALPHA and 5*8 ALPHA language subtags should not accept extlangs
    @ParameterizedTest
    @ValueSource(strings = {"quux", "foobar"})
    public void testExtlangAfterReservedLanguage(String lang) {
        String tag = lang + "-baz";
        // Locale.forLanguageTag is lenient and truncates the extlang
        assertEquals(lang, Locale.forLanguageTag(tag).toLanguageTag());
        // Locale.Builder is strict and should throw
        assertThrows(IllformedLocaleException.class,
                () -> new Builder().setLanguageTag(tag));
    }

    private void checkCalendar(Locale loc, String expected) {
        Calendar cal = Calendar.getInstance(loc);
        assertEquals(expected, cal.getClass().getName(), "Wrong calendar");
    }

    private void checkDigit(Locale loc, Character expected) {
        DecimalFormatSymbols dfs = DecimalFormatSymbols.getInstance(loc);
        Character zero = dfs.getZeroDigit();
        assertEquals(expected, zero, "Wrong digit zero char");
    }
}