ICU-20900 Fix createCanonical

See #922

diff --git a/icu4c/source/common/uloc.cpp b/icu4c/source/common/uloc.cpp
index ea81c0ce87fe13b15e568b8e6bc9fd721dd1629b..cfe9ea0061a47e9fd41d5d72bcf402d5bd600bef 100644 (file)
--- a/icu4c/source/common/uloc.cpp
+++ b/icu4c/source/common/uloc.cpp
@@ -466,15 +466,16 @@ typedef struct CanonicalizationMap {
  * different semantic kinds of transformations.
  */
 static const CanonicalizationMap CANONICALIZE_MAP[] = {
-    { "art_LOJBAN",     "jbo" }, /* registered name */
+    { "art__LOJBAN",    "jbo" }, /* registered name */
     { "hy__AREVELA",    "hy" }, /* Registered IANA variant */
     { "hy__AREVMDA",    "hyw" }, /* Registered IANA variant */
+    { "zh__GUOYU",      "zh" }, /* registered name */
+    { "zh__HAKKA",      "hak" }, /* registered name */
+    { "zh__XIANG",      "hsn" }, /* registered name */
+    // subtags with 3 chars won't be treated as variants.
     { "zh_GAN",         "gan" }, /* registered name */
-    { "zh_GUOYU",       "zh" }, /* registered name */
-    { "zh_HAKKA",       "hak" }, /* registered name */
     { "zh_MIN_NAN",     "nan" }, /* registered name */
     { "zh_WUU",         "wuu" }, /* registered name */
-    { "zh_XIANG",       "hsn" }, /* registered name */
     { "zh_YUE",         "yue" }, /* registered name */
 };
 

diff --git a/icu4c/source/test/intltest/loctest.cpp b/icu4c/source/test/intltest/loctest.cpp
index bd51a1e4268597b6be9301ee609960a91887eb04..d7ad6722504393ebb3c31c37d3a3e2ea7a5396a6 100644 (file)
--- a/icu4c/source/test/intltest/loctest.cpp
+++ b/icu4c/source/test/intltest/loctest.cpp
@@ -255,6 +255,7 @@ void LocaleTest::runIndexedTest( int32_t index, UBool exec, const char* &name, c
     TESTCASE_AUTO(TestBug13277);
     TESTCASE_AUTO(TestBug13554);
     TESTCASE_AUTO(TestBug20410);
+    TESTCASE_AUTO(TestBug20900);
     TESTCASE_AUTO(TestConstructorAcceptsBCP47);
     TESTCASE_AUTO(TestForLanguageTag);
     TESTCASE_AUTO(TestToLanguageTag);
@@ -3100,7 +3101,7 @@ void LocaleTest::TestBug20410() {
 
     static const char locid3[] = "art__lojban@x=0";
     Locale result3 = Locale::createCanonical(locid3);
-    static const Locale expected3("art__LOJBAN@x=0");
+    static const Locale expected3("jbo@x=0");
     assertEquals(locid3, expected3.getName(), result3.getName());
 
     static const char locid4[] = "art-lojban-x-0";
@@ -3109,6 +3110,32 @@ void LocaleTest::TestBug20410() {
     assertEquals(locid4, expected4.getName(), result4.getName());
 }
 
+void LocaleTest::TestBug20900() {
+    static const struct {
+        const char *localeID;    /* input */
+        const char *canonicalID; /* expected canonicalize() result */
+    } testCases[] = {
+        { "art-lojban", "jbo" },
+        { "zh-guoyu", "zh" },
+        { "zh-hakka", "hak" },
+        { "zh-xiang", "hsn" },
+        { "zh-min-nan", "nan" },
+        { "zh-gan", "gan" },
+        { "zh-wuu", "wuu" },
+        { "zh-yue", "yue" },
+    };
+
+    IcuTestErrorCode status(*this, "TestBug20900");
+    for (int32_t i=0; i < UPRV_LENGTHOF(testCases); i++) {
+        Locale loc = Locale::createCanonical(testCases[i].localeID);
+        std::string result = loc.toLanguageTag<std::string>(status);
+        const char* tag = loc.isBogus() ? "BOGUS" : result.c_str();
+        status.errIfFailureAndReset("FAIL: createCanonical(%s).toLanguageTag() expected \"%s\"",
+                    testCases[i].localeID, tag);
+        assertEquals("createCanonical", testCases[i].canonicalID, tag);
+    }
+}
+
 void LocaleTest::TestConstructorAcceptsBCP47() {
     IcuTestErrorCode status(*this, "TestConstructorAcceptsBCP47");
 

diff --git a/icu4c/source/test/intltest/loctest.h b/icu4c/source/test/intltest/loctest.h
index 0e8c09a5acfb6592f81df6948314a29b4d2d2cc2..6c41b132bdd21b67df13ff3daa59124e56a0b226 100644 (file)
--- a/icu4c/source/test/intltest/loctest.h
+++ b/icu4c/source/test/intltest/loctest.h
@@ -119,6 +119,7 @@ public:
     void TestBug13277();
     void TestBug13554();
     void TestBug20410();
+    void TestBug20900();
     void TestConstructorAcceptsBCP47();
 
     void TestAddLikelySubtags();

diff --git a/icu4j/main/classes/core/src/com/ibm/icu/util/ULocale.java b/icu4j/main/classes/core/src/com/ibm/icu/util/ULocale.java
index 06de947bb09b30f21f1c4c89c35b5dde7b360075..f37eade55a5ecd273999c19fb821952352e578a7 100644 (file)
--- a/icu4j/main/classes/core/src/com/ibm/icu/util/ULocale.java
+++ b/icu4j/main/classes/core/src/com/ibm/icu/util/ULocale.java
@@ -389,23 +389,24 @@ public final class ULocale implements Serializable, Comparable<ULocale> {
      * canonicalized id.
      */
     private static String[][] CANONICALIZE_MAP = {
-        { "art_LOJBAN",     "jbo" }, /* registered name */
-        { "cel_GAULISH",    "cel__GAULISH" }, /* registered name */
-        { "de_1901",        "de__1901" }, /* registered name */
-        { "de_1906",        "de__1906" }, /* registered name */
-        { "en_BOONT",       "en__BOONT" }, /* registered name */
-        { "en_SCOUSE",      "en__SCOUSE" }, /* registered name */
+        { "art__LOJBAN",    "jbo" }, /* registered name */
+        { "cel__GAULISH",   "cel__GAULISH" }, /* registered name */
+        { "de__1901",       "de__1901" }, /* registered name */
+        { "de__1906",       "de__1906" }, /* registered name */
+        { "en__BOONT",      "en__BOONT" }, /* registered name */
+        { "en__SCOUSE",     "en__SCOUSE" }, /* registered name */
         { "hy__AREVELA",    "hy", null, null }, /* Registered IANA variant */
         { "hy__AREVMDA",    "hyw", null, null }, /* Registered IANA variant */
-        { "sl_ROZAJ",       "sl__ROZAJ" }, /* registered name */
-        { "zh_GAN",         "zh__GAN" }, /* registered name */
-        { "zh_GUOYU",       "zh" }, /* registered name */
-        { "zh_HAKKA",       "zh__HAKKA" }, /* registered name */
+        { "sl__ROZAJ",      "sl__ROZAJ" }, /* registered name */
+        { "zh__GUOYU",      "zh" }, /* registered name */
+        { "zh__HAKKA",      "hak" }, /* registered name */
+        { "zh__XIANG",      "hsn" }, /* registered name */
+        // Three letter subtags won't be treated as variants.
+        { "zh_GAN",         "gan" }, /* registered name */
         { "zh_MIN",         "zh__MIN" }, /* registered name */
-        { "zh_MIN_NAN",     "zh__MINNAN" }, /* registered name */
-        { "zh_WUU",         "zh__WUU" }, /* registered name */
-        { "zh_XIANG",       "zh__XIANG" }, /* registered name */
-        { "zh_YUE",         "zh__YUE" } /* registered name */
+        { "zh_MIN_NAN",     "nan" }, /* registered name */
+        { "zh_WUU",         "wuu" }, /* registered name */
+        { "zh_YUE",         "yue" } /* registered name */
     };
 
     /**

diff --git a/icu4j/main/tests/core/src/com/ibm/icu/dev/test/util/ULocaleTest.java b/icu4j/main/tests/core/src/com/ibm/icu/dev/test/util/ULocaleTest.java
index d36097e95f9fbe205d3a03857cfd14172004028a..c35deceb0503b30970f02f656799f4bfa822355f 100644 (file)
--- a/icu4j/main/tests/core/src/com/ibm/icu/dev/test/util/ULocaleTest.java
+++ b/icu4j/main/tests/core/src/com/ibm/icu/dev/test/util/ULocaleTest.java
@@ -5137,4 +5137,21 @@ public class ULocaleTest extends TestFmwk {
         Assert.assertEquals(displayName, locale_tag.getDisplayName(displayLocale));
         Assert.assertEquals(displayName, locale_build.getDisplayName(displayLocale));
     }
+
+    @Test
+    public void Test20900() {
+        final String [][] testData = new String[][]{
+            {"art-lojban", "jbo"},
+            {"zh-guoyu", "zh"},
+            {"zh-hakka", "hak"},
+            {"zh-xiang", "hsn"},
+            {"zh-min-nan", "nan"},
+            {"zh-gan", "gan"},
+            {"zh-yue", "yue"},
+        };
+        for (int row=0;row<testData.length;row++) {
+            ULocale loc = ULocale.createCanonical(testData[row][0]);
+            Assert.assertEquals(testData[row][1], loc.toLanguageTag());
+        }
+    }
 }