ICU-23265 Fixed parseMeasureUnitOption() so that it correctly handles measurement unit alises.

Author: richgillam

icu4c/source/i18n/measunit.cpp

@@ -2767,16 +2767,21 @@ StringEnumeration* MeasureUnit::getAvailableTypes(UErrorCode &errorCode) {
 }
 
 bool MeasureUnit::validateAndGet(StringPiece type, StringPiece subtype, MeasureUnit &result) {
-    // Find the type index using binary search
+    // Find the type and subtype indices using binary search
     int32_t typeIdx = binarySearch(gTypes, 0, UPRV_LENGTHOF(gTypes), type);
-    if (typeIdx == -1) {
-        return false;  // Type not found
-    }
+    int32_t subtypeIdx = typeIdx >= 0 ? binarySearch(typeIdx, subtype) : -1;
 
-    // Find the subtype within the type's range using binary search
-    int32_t subtypeIdx = binarySearch(typeIdx, subtype);
-    if (subtypeIdx == -1) {
-        return false;  // Subtype not found
+    if (typeIdx >= 0 && subtypeIdx < 0) {
+        // if we did find the type, but didn't find the subtype, that might be because the sybtype name
+        // is an alias-- try using MeasureUnit::forIdentifier(), which will resolve aliases
+        UErrorCode localStatus = U_ZERO_ERROR;
+        MeasureUnit tempUnit = MeasureUnit::forIdentifier(subtype, localStatus);
+        if (U_SUCCESS(localStatus) && uprv_strcmp(type.data(), tempUnit.getType()) == 0) {
+            subtypeIdx = tempUnit.fSubTypeId + gOffsets[tempUnit.fTypeId];
+        }
+    }
+    if (typeIdx < 0 || subtypeIdx < 0) {
+        return false;
     }
 
     // Create the MeasureUnit and return it

icu4c/source/i18n/number_skeletons.cpp

@@ -1076,7 +1076,7 @@ void blueprint_helpers::parseMeasureUnitOption(const StringSegment& segment, Mac
     if (MeasureUnit::validateAndGet(type.toStringPiece(), subType.toStringPiece(), unit)) {
         macros.unit = unit;
         return;
-    } 
+    }
 
     status = U_NUMBER_SKELETON_SYNTAX_ERROR;
 }

icu4c/source/test/depstest/dependencies.txt

@@ -892,7 +892,7 @@ library: i18n
     formatting formattable_cnv regex regex_cnv translit
     double_conversion number_representation number_output numberformatter
     number_skeletons number_usageprefs numberparser
-    units_extra unitsformatter
+    units unitsformatter
     universal_time_scale
     uclean_i18n
     display_options
@@ -1057,7 +1057,7 @@ group: number_skeletons
     number_skeletons.o number_capi.o number_asformat.o numrange_capi.o
   deps
     numberformatter
-    units_extra
+    units
 
 group: number_symbolswrapper
     number_symbolswrapper.o
@@ -1128,21 +1128,16 @@ group: sharedbreakiterator
   deps
     breakiterator
 
-group: units_extra
-    measunit_extra.o
-  deps
-    units bytestriebuilder bytestrie resourcebundle uclean_i18n
-    double_conversion
-
 group: units
-    measunit.o currunit.o
+    measunit.o currunit.o measunit_extra.o
   deps
-    stringenumeration errorcode
+    stringenumeration errorcode bytestriebuilder bytestrie resourcebundle uclean_i18n
+    double_conversion
 
 group: unitsformatter
     units_data.o units_converter.o units_complexconverter.o units_router.o
   deps
-    resourcebundle units_extra double_conversion number_representation formattable sort
+    resourcebundle units double_conversion number_representation formattable sort
     number_rounding
 
 group: decnumber

icu4c/source/test/intltest/numbertest.h

@@ -306,6 +306,7 @@ class NumberSkeletonTest : public IntlTest {
     void perUnitToSkeleton();
     void measurementSystemOverride();
     void longSkeletonCrash();
+    void unitAliases();
 
     void runIndexedTest(int32_t index, UBool exec, const char*& name, char* par = nullptr) override;
 

icu4c/source/test/intltest/numbertest_skeletons.cpp

@@ -36,6 +36,7 @@ void NumberSkeletonTest::runIndexedTest(int32_t index, UBool exec, const char*&
         TESTCASE_AUTO(perUnitToSkeleton);
         TESTCASE_AUTO(measurementSystemOverride);
         TESTCASE_AUTO(longSkeletonCrash);
+        TESTCASE_AUTO(unitAliases);
     TESTCASE_AUTO_END;
 }
 
@@ -581,4 +582,31 @@ void NumberSkeletonTest::longSkeletonCrash() {
     UnlocalizedNumberFormatter nf = NumberFormatter::forSkeleton(skeleton, err);
 }
 
+void NumberSkeletonTest::unitAliases() {
+    IcuTestErrorCode status(*this, "unitAliases");
+
+    struct TestCase {
+        const char16_t* skeleton;
+        const char16_t* expectedResult;
+    } testCases[] = {
+        { u"measure-unit/concentr-part-per-1e6", u"3.14 ppm" },
+        { u"measure-unit/concentr-part-per-million", u"3.14 ppm" },
+        { u"measure-unit/concentr-permillion", u"3.14 ppm" },
+        { u"measure-unit/concentr-milligram-ofglucose-per-deciliter", u"3.14 mg/dL" },
+        { u"measure-unit/concentr-milligram-per-deciliter", u"3.14 mg/dL" },
+        { u"measure-unit/mass-tonne", u"3.14 t" },
+        { u"measure-unit/mass-metric-ton", u"3.14 t" },
+    };
+    
+    for (const auto& testCase : testCases) {
+        LocalizedNumberFormatter nf = NumberFormatter::forSkeleton(testCase.skeleton, status).locale(Locale::getUS());
+        UnicodeString actualResult = nf.formatDouble(3.14, status).toString(status);
+        
+        status.setScope(testCase.skeleton);
+        if (!status.errIfFailureAndReset()) {
+            assertEquals(testCase.skeleton, testCase.expectedResult, actualResult);
+        }
+    }
+}
+
 #endif /* #if !UCONFIG_NO_FORMATTING */

icu4j/main/core/src/main/java/com/ibm/icu/number/NumberSkeletonImpl.java

@@ -1094,11 +1094,12 @@ private static void parseMeasureUnitOption(StringSegment segment, MacroProps mac
             }
             String type = segment.subSequence(0, firstHyphen).toString();
             String subType = segment.subSequence(firstHyphen + 1, segment.length()).toString();
-            MeasureUnit unit = MeasureUnit.getUnit(type, subType);
+            MeasureUnit unit = MeasureUnit.validateAndGet(type, subType);
             if (unit != null) {
                 macros.unit = unit;
                 return;
             }
+
             throw new SkeletonSyntaxException("Unknown measure unit", segment);
         }
 

icu4j/main/core/src/main/java/com/ibm/icu/util/MeasureUnit.java

@@ -890,6 +890,23 @@ public static MeasureUnit getUnit(String type, String subtype) {
         return units != null ? units.get(subtype) : null;
     }
 
+    @Deprecated
+    public static MeasureUnit validateAndGet(String type, String subtype) {
+        MeasureUnit result = MeasureUnit.getUnit(type, subtype);
+
+        if (result == null) {
+            try {
+                result = MeasureUnit.forIdentifier(subtype);
+                if (!result.getType().equals(type)) {
+                    result = null;
+                }
+            } catch (IllegalArgumentException e) {
+                // leave result as null
+            }
+        }
+        return result;
+    }
+
     static final UnicodeSet ASCII = new UnicodeSet('a', 'z').freeze();
     static final UnicodeSet ASCII_HYPHEN_DIGITS =
             new UnicodeSet('-', '-', '0', '9', 'a', 'z').freeze();

icu4j/main/core/src/test/java/com/ibm/icu/dev/test/number/NumberSkeletonTest.java

@@ -549,4 +549,45 @@ public void measurementSystemOverride() {
                     "Wrong result: " + languageTag + ":" + skeleton, expectedResult, actualResult);
         }
     }
+
+    @Test
+    public void unitAliases() {
+        class TestCase {
+            String skeleton;
+            String expectedResult;
+
+            TestCase(String skeleton, String expectedResult) {
+                this.skeleton = skeleton;
+                this.expectedResult = expectedResult;
+            }
+        }
+
+        TestCase[] testCases =
+                new TestCase[] {
+                    new TestCase("measure-unit/concentr-part-per-1e6", "3.14 ppm"),
+                    new TestCase("measure-unit/concentr-part-per-million", "3.14 ppm"),
+                    new TestCase("measure-unit/concentr-permillion", "3.14 ppm"),
+                    new TestCase(
+                            "measure-unit/concentr-milligram-ofglucose-per-deciliter",
+                            "3.14 mg/dL"),
+                    new TestCase("measure-unit/concentr-milligram-per-deciliter", "3.14 mg/dL"),
+                    new TestCase("measure-unit/mass-tonne", "3.14 t"),
+                    new TestCase("measure-unit/mass-metric-ton", "3.14 t"),
+                };
+
+        for (TestCase testCase : testCases) {
+            try {
+                LocalizedNumberFormatter nf =
+                        NumberFormatter.forSkeleton(testCase.skeleton).locale(Locale.US);
+                String actualResult = nf.format(3.14).toString();
+
+                assertEquals(
+                        "Wrong result for " + testCase.skeleton + ":",
+                        testCase.expectedResult,
+                        actualResult);
+            } catch (SkeletonSyntaxException e) {
+                fail(testCase.skeleton);
+            }
+        }
+    }
 }