ICU-20321 Fix ultag_isUnicodeLocaleKey

Returns false when passingin alphanum digit.
Sync with UTS35
https://www.unicode.org/reports/tr35/#Unicode_locale_identifier

address review feedback add unit tests.

Fix Java too

add test cases to c++

Fix format

change test case

diff --git a/icu4c/source/common/uloc_tag.cpp b/icu4c/source/common/uloc_tag.cpp
index ce5ab9de7a80db9fbcadc83c3c73d5acb157be6e..bdc0652df0aff30da27aa6b85ac65961c55bc555 100644 (file)
--- a/icu4c/source/common/uloc_tag.cpp
+++ b/icu4c/source/common/uloc_tag.cpp
@@ -608,7 +608,7 @@ ultag_isUnicodeLocaleKey(const char* s, int32_t len) {
     if (len < 0) {
         len = (int32_t)uprv_strlen(s);
     }
-    if (len == 2 && _isAlphaNumericString(s, len)) {
+    if (len == 2 && (ISALPHA(*s) || ISNUMERIC(*s)) && ISALPHA(s[1])) {
         return TRUE;
     }
     return FALSE;

diff --git a/icu4c/source/test/cintltst/cloctst.c b/icu4c/source/test/cintltst/cloctst.c
index 2268486a9f16d2f9098d2f8ddb9e6a9817a55e71..99f9970a82b0bb19ea620478da767dcf34445a31 100644 (file)
--- a/icu4c/source/test/cintltst/cloctst.c
+++ b/icu4c/source/test/cintltst/cloctst.c
@@ -53,6 +53,7 @@ static void TestUnicodeDefines(void);
 static void TestIsRightToLeft(void);
 static void TestBadLocaleIDs(void);
 static void TestBug20370(void);
+static void TestBug20321UnicodeLocaleKey(void);
 
 void PrintDataTable();
 
@@ -268,6 +269,7 @@ void addLocaleTest(TestNode** root)
     TESTCASE(TestToLegacyType);
     TESTCASE(TestBadLocaleIDs);
     TESTCASE(TestBug20370);
+    TESTCASE(TestBug20321UnicodeLocaleKey);
 }
 
 
@@ -6287,6 +6289,39 @@ static void TestToUnicodeLocaleKey(void)
     }
 }
 
+static void TestBug20321UnicodeLocaleKey(void)
+{
+    // key = alphanum alpha ;
+    static const char* invalid[] = {
+        "a0",
+        "00",
+        "a@",
+        "0@",
+        "@a",
+        "@a",
+        "abc",
+        "0bc",
+    };
+    for (int i = 0; i < UPRV_LENGTHOF(invalid); i++) {
+        const char* bcpKey = NULL;
+        bcpKey = uloc_toUnicodeLocaleKey(invalid[i]);
+        if (bcpKey != NULL) {
+            log_err("toUnicodeLocaleKey: keyword=%s => %s, expected=NULL\n", invalid[i], bcpKey);
+        }
+    }
+    static const char* valid[] = {
+        "aa",
+        "0a",
+    };
+    for (int i = 0; i < UPRV_LENGTHOF(valid); i++) {
+        const char* bcpKey = NULL;
+        bcpKey = uloc_toUnicodeLocaleKey(valid[i]);
+        if (bcpKey == NULL) {
+            log_err("toUnicodeLocaleKey: keyword=%s => NULL, expected!=NULL\n", valid[i]);
+        }
+    }
+}
+
 static void TestToLegacyKey(void)
 {
     /* $IN specifies the result should be the input pointer itself */

diff --git a/icu4j/main/classes/core/src/com/ibm/icu/impl/locale/UnicodeLocaleExtension.java b/icu4j/main/classes/core/src/com/ibm/icu/impl/locale/UnicodeLocaleExtension.java
index bc27a490303310374f9b98ddf517cedb8f54defb..c0e820f99dc57847ef3fd0d7c3c576223d7baa87 100644 (file)
--- a/icu4j/main/classes/core/src/com/ibm/icu/impl/locale/UnicodeLocaleExtension.java
+++ b/icu4j/main/classes/core/src/com/ibm/icu/impl/locale/UnicodeLocaleExtension.java
@@ -93,8 +93,9 @@ public class UnicodeLocaleExtension extends Extension {
     }
 
     public static boolean isKey(String s) {
-        // 2alphanum
-        return (s.length() == 2) && AsciiUtil.isAlphaNumericString(s);
+        // key = alphanum alpha ;
+        return (s.length() == 2) && AsciiUtil.isAlphaNumeric(s.charAt(0)) &&
+            AsciiUtil.isAlpha(s.charAt(1));
     }
 
     public static boolean isTypeSubtag(String s) {

diff --git a/icu4j/main/tests/core/src/com/ibm/icu/dev/test/util/LocaleBuilderTest.java b/icu4j/main/tests/core/src/com/ibm/icu/dev/test/util/LocaleBuilderTest.java
index 772d0eab26b20e3001e08da5783a4504cc4d6e96..71002bac2aefe9ff6109f3711dbe722d6be0bbdc 100644 (file)
--- a/icu4j/main/tests/core/src/com/ibm/icu/dev/test/util/LocaleBuilderTest.java
+++ b/icu4j/main/tests/core/src/com/ibm/icu/dev/test/util/LocaleBuilderTest.java
@@ -82,12 +82,12 @@ public class LocaleBuilderTest extends TestFmwk {
                 {"E", "z", "ExtZ", "L", "en", "E", "z", null, "T", "en", "en"},
                 {"E", "a", "x", "X"},
                 {"E", "a", "abc_def", "T", "und-a-abc-def", "@a=abc-def"},
-                // Design limitation - typeless u extension keyword 00 below is interpreted as a boolean value true/yes.
+                // Design limitation - typeless u extension keyword 0a below is interpreted as a boolean value true/yes.
                 // With the legacy keyword syntax, "yes" is used for such boolean value instead of "true".
-                // However, once the legacy keyword is translated back to BCP 47 u extension, key "00" is unknown,
+                // However, once the legacy keyword is translated back to BCP 47 u extension, key "0a" is unknown,
                 // so "yes" is preserved - not mapped to "true". We could change the code to automatically transform
-                // "yes" to "true", but it will break roundtrip conversion if BCP 47 u extension has "00-yes".
-                {"L", "en", "E", "u", "bbb-aaa-00", "T", "en-u-aaa-bbb-00-yes", "en@00=yes;attribute=aaa-bbb"},
+                // "yes" to "true", but it will break roundtrip conversion if BCP 47 u extension has "0a-yes".
+                {"L", "en", "E", "u", "bbb-aaa-0a", "T", "en-u-aaa-bbb-0a-yes", "en@0a=yes;attribute=aaa-bbb"},
                 {"L", "fr", "R", "FR", "P", "Yoshito-ICU", "T", "fr-FR-x-yoshito-icu", "fr_FR@x=yoshito-icu"},
                 {"L", "ja", "R", "jp", "K", "ca", "japanese", "T", "ja-JP-u-ca-japanese", "ja_JP@calendar=japanese"},
                 {"K", "co", "PHONEBK", "K", "ca", "gregory", "L", "De", "T", "de-u-ca-gregory-co-phonebk", "de@calendar=gregorian;collation=phonebook"},

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 ba772a1353441f9e9dc673d8da0c54c5e6ca07c9..8ffc478c7ba77bd41f2d8df56176fc85eac53a59 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
@@ -4696,6 +4696,36 @@ public class ULocaleTest extends TestFmwk {
         }
     }
 
+    @Test
+    public void TestBug20321UnicodeLocaleKey() {
+        // key = alphanum alpha ;
+        String[] INVALID = {
+            "a0",
+            "00",
+            "a@",
+            "0@",
+            "@a",
+            "@a",
+            "abc",
+            "0bc",
+        };
+
+        for (String invalid : INVALID) {
+            String bcpKey = ULocale.toUnicodeLocaleKey(invalid);
+            assertNull("keyword=" + invalid, bcpKey);
+        }
+
+        String[] VALID = {
+            "aa",
+            "0a",
+        };
+
+        for (String valid : VALID) {
+            String bcpKey = ULocale.toUnicodeLocaleKey(valid);
+            assertEquals("keyword=" + valid, valid, bcpKey);
+        };
+    }
+
     @Test
     public void TestToLegacyKey() {
         String[][] DATA = {