Source code

001/*
002 * Licensed to the Apache Software Foundation (ASF) under one or more
003 * contributor license agreements.  See the NOTICE file distributed with
004 * this work for additional information regarding copyright ownership.
005 * The ASF licenses this file to You under the Apache License, Version 2.0
006 * (the "License"); you may not use this file except in compliance with
007 * the License.  You may obtain a copy of the License at
008 *
009 *      http://www.apache.org/licenses/LICENSE-2.0
010 *
011 * Unless required by applicable law or agreed to in writing, software
012 * distributed under the License is distributed on an "AS IS" BASIS,
013 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
014 * See the License for the specific language governing permissions and
015 * limitations under the License.
016 */
017package org.apache.commons.lang3;
018
019import java.util.ArrayList;
020import java.util.Arrays;
021import java.util.Collections;
022import java.util.HashSet;
023import java.util.List;
024import java.util.Locale;
025import java.util.Set;
026import java.util.concurrent.ConcurrentHashMap;
027import java.util.concurrent.ConcurrentMap;
028import java.util.function.Predicate;
029import java.util.stream.Collectors;
030
031/**
032 * Operations to assist when working with a {@link Locale}.
033 *
034 * <p>This class tries to handle {@code null} input gracefully.
035 * An exception will not be thrown for a {@code null} input.
036 * Each method documents its behavior in more detail.</p>
037 *
038 * @since 2.2
039 */
040public class LocaleUtils {
041
042    // class to avoid synchronization (Init on demand)
043    static class SyncAvoid {
044        /** Unmodifiable list of available locales. */
045        private static final List<Locale> AVAILABLE_LOCALE_LIST;
046        /** Unmodifiable set of available locales. */
047        private static final Set<Locale> AVAILABLE_LOCALE_SET;
048
049        static {
050            final List<Locale> list = new ArrayList<>(Arrays.asList(Locale.getAvailableLocales()));  // extra safe
051            AVAILABLE_LOCALE_LIST = Collections.unmodifiableList(list);
052            AVAILABLE_LOCALE_SET = Collections.unmodifiableSet(new HashSet<>(list));
053        }
054    }
055    private static final char UNDERSCORE = '_';
056    private static final String UNDETERMINED = "und";
057
058    private static final char DASH = '-';
059
060    /** Concurrent map of language locales by country. */
061    private static final ConcurrentMap<String, List<Locale>> cLanguagesByCountry =
062        new ConcurrentHashMap<>();
063
064    /** Concurrent map of country locales by language. */
065    private static final ConcurrentMap<String, List<Locale>> cCountriesByLanguage =
066        new ConcurrentHashMap<>();
067
068    /**
069     * Obtains an unmodifiable list of installed locales.
070     *
071     * <p>This method is a wrapper around {@link Locale#getAvailableLocales()}.
072     * It is more efficient, as the JDK method must create a new array each
073     * time it is called.</p>
074     *
075     * @return the unmodifiable list of available locales
076     */
077    public static List<Locale> availableLocaleList() {
078        return SyncAvoid.AVAILABLE_LOCALE_LIST;
079    }
080
081    private static List<Locale> availableLocaleList(final Predicate<Locale> predicate) {
082        return availableLocaleList().stream().filter(predicate).collect(Collectors.toList());
083    }
084
085    /**
086     * Obtains an unmodifiable set of installed locales.
087     *
088     * <p>This method is a wrapper around {@link Locale#getAvailableLocales()}.
089     * It is more efficient, as the JDK method must create a new array each
090     * time it is called.</p>
091     *
092     * @return the unmodifiable set of available locales
093     */
094    public static Set<Locale> availableLocaleSet() {
095        return SyncAvoid.AVAILABLE_LOCALE_SET;
096    }
097
098    /**
099     * Obtains the list of countries supported for a given language.
100     *
101     * <p>This method takes a language code and searches to find the
102     * countries available for that language. Variant locales are removed.</p>
103     *
104     * @param languageCode  the 2 letter language code, null returns empty
105     * @return an unmodifiable List of Locale objects, not null
106     */
107    public static List<Locale> countriesByLanguage(final String languageCode) {
108        if (languageCode == null) {
109            return Collections.emptyList();
110        }
111        return cCountriesByLanguage.computeIfAbsent(languageCode, lc -> Collections.unmodifiableList(
112            availableLocaleList(locale -> languageCode.equals(locale.getLanguage()) && !locale.getCountry().isEmpty() && locale.getVariant().isEmpty())));
113    }
114
115    /**
116     * Checks if the locale specified is in the set of available locales.
117     *
118     * @param locale the Locale object to check if it is available
119     * @return true if the locale is a known locale
120     */
121    public static boolean isAvailableLocale(final Locale locale) {
122        return availableLocaleSet().contains(locale);
123    }
124
125    /**
126     * Checks whether the given String is a ISO 3166 alpha-2 country code.
127     *
128     * @param str the String to check
129     * @return true, is the given String is a ISO 3166 compliant country code.
130     */
131    private static boolean isISO3166CountryCode(final String str) {
132        return StringUtils.isAllUpperCase(str) && str.length() == 2;
133    }
134
135    /**
136     * Checks whether the given String is a ISO 639 compliant language code.
137     *
138     * @param str the String to check.
139     * @return true, if the given String is a ISO 639 compliant language code.
140     */
141    private static boolean isISO639LanguageCode(final String str) {
142        return StringUtils.isAllLowerCase(str) && (str.length() == 2 || str.length() == 3);
143    }
144
145    /**
146     * Tests whether a Locale's language is undetermined.
147     * <p>
148     * A Locale's language tag is undetermined if it's value is {@code "und"}. If a language is empty, or not well-formed (for example, "a" or"e2"), it will be
149     * equal to {@code "und"}.
150     * </p>
151     *
152     * @param locale the locale to test.
153     * @return whether a Locale's language is undetermined.
154     * @see Locale#toLanguageTag()
155     * @since 3.14.0
156     */
157    public static boolean isLanguageUndetermined(final Locale locale) {
158        return locale == null || UNDETERMINED.equals(locale.toLanguageTag());
159    }
160
161    /**
162     * Checks whether the given String is a UN M.49 numeric area code.
163     *
164     * @param str the String to check
165     * @return true, is the given String is a UN M.49 numeric area code.
166     */
167    private static boolean isNumericAreaCode(final String str) {
168        return StringUtils.isNumeric(str) && str.length() == 3;
169    }
170
171    /**
172     * Obtains the list of languages supported for a given country.
173     *
174     * <p>This method takes a country code and searches to find the
175     * languages available for that country. Variant locales are removed.</p>
176     *
177     * @param countryCode  the 2-letter country code, null returns empty
178     * @return an unmodifiable List of Locale objects, not null
179     */
180    public static List<Locale> languagesByCountry(final String countryCode) {
181        if (countryCode == null) {
182            return Collections.emptyList();
183        }
184        return cLanguagesByCountry.computeIfAbsent(countryCode,
185            k -> Collections.unmodifiableList(availableLocaleList(locale -> countryCode.equals(locale.getCountry()) && locale.getVariant().isEmpty())));
186    }
187
188    /**
189     * Obtains the list of locales to search through when performing
190     * a locale search.
191     *
192     * <pre>
193     * localeLookupList(Locale("fr", "CA", "xxx"))
194     *   = [Locale("fr", "CA", "xxx"), Locale("fr", "CA"), Locale("fr")]
195     * </pre>
196     *
197     * @param locale  the locale to start from
198     * @return the unmodifiable list of Locale objects, 0 being locale, not null
199     */
200    public static List<Locale> localeLookupList(final Locale locale) {
201        return localeLookupList(locale, locale);
202    }
203
204    /**
205     * Obtains the list of locales to search through when performing
206     * a locale search.
207     *
208     * <pre>
209     * localeLookupList(Locale("fr", "CA", "xxx"), Locale("en"))
210     *   = [Locale("fr", "CA", "xxx"), Locale("fr", "CA"), Locale("fr"), Locale("en"]
211     * </pre>
212     *
213     * <p>The result list begins with the most specific locale, then the
214     * next more general and so on, finishing with the default locale.
215     * The list will never contain the same locale twice.</p>
216     *
217     * @param locale  the locale to start from, null returns empty list
218     * @param defaultLocale  the default locale to use if no other is found
219     * @return the unmodifiable list of Locale objects, 0 being locale, not null
220     */
221    public static List<Locale> localeLookupList(final Locale locale, final Locale defaultLocale) {
222        final List<Locale> list = new ArrayList<>(4);
223        if (locale != null) {
224            list.add(locale);
225            if (!locale.getVariant().isEmpty()) {
226                list.add(new Locale(locale.getLanguage(), locale.getCountry()));
227            }
228            if (!locale.getCountry().isEmpty()) {
229                list.add(new Locale(locale.getLanguage(), StringUtils.EMPTY));
230            }
231            if (!list.contains(defaultLocale)) {
232                list.add(defaultLocale);
233            }
234        }
235        return Collections.unmodifiableList(list);
236    }
237
238    /**
239     * Tries to parse a locale from the given String.
240     *
241     * @param str the String to parse a locale from.
242     * @return a Locale instance parsed from the given String.
243     * @throws IllegalArgumentException if the given String can not be parsed.
244     */
245    private static Locale parseLocale(final String str) {
246        if (isISO639LanguageCode(str)) {
247            return new Locale(str);
248        }
249
250        final String[] segments = str.indexOf(UNDERSCORE) != -1
251            ? str.split(String.valueOf(UNDERSCORE), -1)
252            : str.split(String.valueOf(DASH), -1);
253        final String language = segments[0];
254        if (segments.length == 2) {
255            final String country = segments[1];
256            if (isISO639LanguageCode(language) && isISO3166CountryCode(country) ||
257                    isNumericAreaCode(country)) {
258                return new Locale(language, country);
259            }
260        } else if (segments.length == 3) {
261            final String country = segments[1];
262            final String variant = segments[2];
263            if (isISO639LanguageCode(language) &&
264                    (country.isEmpty() || isISO3166CountryCode(country) || isNumericAreaCode(country)) &&
265                    !variant.isEmpty()) {
266                return new Locale(language, country, variant);
267            }
268        }
269        throw new IllegalArgumentException("Invalid locale format: " + str);
270    }
271
272    /**
273     * Returns the given locale if non-{@code null}, otherwise {@link Locale#getDefault()}.
274     *
275     * @param locale a locale or {@code null}.
276     * @return the given locale if non-{@code null}, otherwise {@link Locale#getDefault()}.
277     * @since 3.12.0
278     */
279    public static Locale toLocale(final Locale locale) {
280        return locale != null ? locale : Locale.getDefault();
281    }
282
283    /**
284     * Converts a String to a Locale.
285     *
286     * <p>This method takes the string format of a locale and creates the
287     * locale object from it.</p>
288     *
289     * <pre>
290     *   LocaleUtils.toLocale("")           = new Locale("", "")
291     *   LocaleUtils.toLocale("en")         = new Locale("en", "")
292     *   LocaleUtils.toLocale("en_GB")      = new Locale("en", "GB")
293     *   LocaleUtils.toLocale("en-GB")      = new Locale("en", "GB")
294     *   LocaleUtils.toLocale("en_001")     = new Locale("en", "001")
295     *   LocaleUtils.toLocale("en_GB_xxx")  = new Locale("en", "GB", "xxx")   (#)
296     * </pre>
297     *
298     * <p>(#) The behavior of the JDK variant constructor changed between JDK1.3 and JDK1.4.
299     * In JDK1.3, the constructor upper cases the variant, in JDK1.4, it doesn't.
300     * Thus, the result from getVariant() may vary depending on your JDK.</p>
301     *
302     * <p>This method validates the input strictly.
303     * The language code must be lowercase.
304     * The country code must be uppercase.
305     * The separator must be an underscore or a dash.
306     * The length must be correct.
307     * </p>
308     *
309     * @param str  the locale String to convert, null returns null
310     * @return a Locale, null if null input
311     * @throws IllegalArgumentException if the string is an invalid format
312     * @see Locale#forLanguageTag(String)
313     */
314    public static Locale toLocale(final String str) {
315        if (str == null) {
316            // TODO Should this return the default locale?
317            return null;
318        }
319        if (str.isEmpty()) { // LANG-941 - JDK 8 introduced an empty locale where all fields are blank
320            return new Locale(StringUtils.EMPTY, StringUtils.EMPTY);
321        }
322        if (str.contains("#")) { // LANG-879 - Cannot handle Java 7 script & extensions
323            throw new IllegalArgumentException("Invalid locale format: " + str);
324        }
325        final int len = str.length();
326        if (len < 2) {
327            throw new IllegalArgumentException("Invalid locale format: " + str);
328        }
329        final char ch0 = str.charAt(0);
330        if (ch0 == UNDERSCORE || ch0 == DASH) {
331            if (len < 3) {
332                throw new IllegalArgumentException("Invalid locale format: " + str);
333            }
334            final char ch1 = str.charAt(1);
335            final char ch2 = str.charAt(2);
336            if (!Character.isUpperCase(ch1) || !Character.isUpperCase(ch2)) {
337                throw new IllegalArgumentException("Invalid locale format: " + str);
338            }
339            if (len == 3) {
340                return new Locale(StringUtils.EMPTY, str.substring(1, 3));
341            }
342            if (len < 5) {
343                throw new IllegalArgumentException("Invalid locale format: " + str);
344            }
345            if (str.charAt(3) != ch0) {
346                throw new IllegalArgumentException("Invalid locale format: " + str);
347            }
348            return new Locale(StringUtils.EMPTY, str.substring(1, 3), str.substring(4));
349        }
350
351        return parseLocale(str);
352    }
353
354    /**
355     * {@link LocaleUtils} instances should NOT be constructed in standard programming.
356     * Instead, the class should be used as {@code LocaleUtils.toLocale("en_GB");}.
357     *
358     * <p>This constructor is public to permit tools that require a JavaBean instance
359     * to operate.</p>
360     */
361    public LocaleUtils() {
362    }
363
364}