ICU-13597 Adding support for field positions to the NumberFormatter C API.

X-SVN-Rev: 41159

icu4c/source/i18n/fieldposutil.h

@@ -0,0 +1,50 @@
+// © 2018 and later: Unicode, Inc. and others.
+// License & terms of use: http://www.unicode.org/copyright.html
+
+#include "unicode/utypes.h"
+
+#if !UCONFIG_NO_FORMATTING
+#ifndef __SOURCE_FIELDPOSUTIL_H__
+#define __SOURCE_FIELDPOSUTIL_H__
+
+U_NAMESPACE_BEGIN
+
+/**
+ * Wraps a UFieldPosition and makes it usable as a FieldPosition. Example:
+ *
+ * <pre>
+ * UFieldPositionWrapper wrapper(myUFPos);
+ * u_favorite_function_taking_ufpos(wrapper);
+ * // when destructed, the wrapper saves the data back into myUFPos
+ * </pre>
+ */
+class UFieldPositionWrapper : public UMemory {
+  public:
+    explicit UFieldPositionWrapper(UFieldPosition& ufpos)
+            : _ufpos(ufpos) {
+        _fpos.setField(_ufpos.field);
+        _fpos.setBeginIndex(_ufpos.beginIndex);
+        _fpos.setEndIndex(_ufpos.endIndex);
+    }
+
+    /** When destructed, copies the information from the fpos into the ufpos. */
+    ~UFieldPositionWrapper() {
+        _ufpos.field = _fpos.getField();
+        _ufpos.beginIndex = _fpos.getBeginIndex();
+        _ufpos.endIndex = _fpos.getEndIndex();
+    }
+
+    /** Conversion operator to FieldPosition */
+    operator FieldPosition&() {
+        return _fpos;
+    }
+
+  private:
+    FieldPosition _fpos;
+    UFieldPosition& _ufpos;
+};
+
+U_NAMESPACE_END
+
+#endif //__SOURCE_FIELDPOSUTIL_H__
+#endif /* #if !UCONFIG_NO_FORMATTING */

icu4c/source/i18n/number_capi.cpp

@@ -13,6 +13,7 @@
 #include "number_utypes.h"
 #include "unicode/numberformatter.h"
 #include "unicode/unumberformatter.h"
+#include "fieldposutil.h"
 
 using namespace icu;
 using namespace icu::number;
@@ -153,16 +154,47 @@ unumf_resultToString(const UFormattedNumber* uresult, UChar* buffer, int32_t buf
 }
 
 U_CAPI void U_EXPORT2
-unumf_closeResult(const UFormattedNumber* uresult, UErrorCode* ec) {
-    const UFormattedNumberData* impl = UFormattedNumberData::validate(uresult, *ec);
+unumf_resultGetField(const UFormattedNumber* uresult, UFieldPosition* ufpos, UErrorCode* ec) {
+    if (ufpos == nullptr) {
+        *ec = U_ILLEGAL_ARGUMENT_ERROR;
+        return;
+    }
+
+    const UFormattedNumberData* result = UFormattedNumberData::validate(uresult, *ec);
     if (U_FAILURE(*ec)) { return; }
+
+    UFieldPositionWrapper helper(*ufpos);
+    result->string.populateFieldPosition(helper, 0, *ec);
+}
+
+U_CAPI void U_EXPORT2
+unumf_resultGetAllFields(const UFormattedNumber* uresult, UFieldPositionIterator* ufpositer,
+                         UErrorCode* ec) {
+    if (ufpositer == nullptr) {
+        *ec = U_ILLEGAL_ARGUMENT_ERROR;
+        return;
+    }
+
+    const UFormattedNumberData* result = UFormattedNumberData::validate(uresult, *ec);
+    if (U_FAILURE(*ec)) { return; }
+
+    auto* helper = reinterpret_cast<FieldPositionIterator*>(ufpositer);
+    result->string.populateFieldPositionIterator(*helper, *ec);
+}
+
+U_CAPI void U_EXPORT2
+unumf_closeResult(const UFormattedNumber* uresult) {
+    UErrorCode localStatus = U_ZERO_ERROR;
+    const UFormattedNumberData* impl = UFormattedNumberData::validate(uresult, localStatus);
+    if (U_FAILURE(localStatus)) { return; }
     delete impl;
 }
 
 U_CAPI void U_EXPORT2
-unumf_close(UNumberFormatter* f, UErrorCode* ec) {
-    const UNumberFormatterData* impl = UNumberFormatterData::validate(f, *ec);
-    if (U_FAILURE(*ec)) { return; }
+unumf_close(UNumberFormatter* f) {
+    UErrorCode localStatus = U_ZERO_ERROR;
+    const UNumberFormatterData* impl = UNumberFormatterData::validate(f, localStatus);
+    if (U_FAILURE(localStatus)) { return; }
     delete impl;
 }
 

icu4c/source/i18n/unicode/unumberformatter.h

@@ -7,6 +7,9 @@
 #ifndef __UNUMBERFORMATTER_H__
 #define __UNUMBERFORMATTER_H__
 
+#include "unicode/ufieldpositer.h"
+#include "unicode/umisc.h"
+
 
 /**
  * \file
@@ -441,25 +444,62 @@ unumf_resultToString(const UFormattedNumber* uresult, UChar* buffer, int32_t buf
 
 
 /**
- * Releases the UFormattedNumber returned by unumf_formatDouble and friends.
+ * Determines the start and end indices of the first occurrence of the given field in the output string.
+ * This allows you to determine the locations of the integer part, fraction part, and sign.
  *
- * NOTE: This is a C-compatible API; C++ users should build against numberformatter.h instead.
+ * If a field occurs multiple times in an output string, such as a grouping separator, this method will
+ * only ever return the first occurrence. Use unumf_resultGetAllFields() to access all occurrences of an
+ * attribute.
  *
- * @draft ICU 62
+ * @param fpos
+ *         A pointer to a UFieldPosition. On input, position->field is read. On output,
+ *         position->beginIndex and position->endIndex indicate the beginning and ending indices of field
+ *         number position->field, if such a field exists.
  */
 U_DRAFT void U_EXPORT2
-unumf_closeResult(const UFormattedNumber* uresult, UErrorCode* ec);
+unumf_resultGetField(const UFormattedNumber* uresult, UFieldPosition* ufpos, UErrorCode* ec);
 
 
 /**
- * Releases the UNumberFormatter created by unumf_openFromSkeletonAndLocale.
+ * Populates the given iterator with all fields in the formatted output string. This allows you to
+ * determine the locations of the integer part, fraction part, and sign.
+ *
+ * If you need information on only one field, consider using unumf_resultGetField().
+ *
+ * @param fpositer
+ *         A pointer to a UFieldPositionIterator created by {@link #ufieldpositer_open}. Iteration
+ *         information already present in the UFieldPositionIterator is deleted, and the iterator is reset
+ *         to apply to the fields in the formatted string created by this function call. The field values
+ *         and indexes returned by {@link #ufieldpositer_next} represent fields denoted by
+ *         the UNumberFormatFields enum. Fields are not returned in a guaranteed order. Fields cannot
+ *         overlap, but they may nest. For example, 1234 could format as "1,234" which might consist of a
+ *         grouping separator field for ',' and an integer field encompassing the entire string.
+ */
+U_DRAFT void U_EXPORT2
+unumf_resultGetAllFields(const UFormattedNumber* uresult, UFieldPositionIterator* ufpositer,
+                         UErrorCode* ec);
+
+
+/**
+ * Releases the UNumberFormatter created by unumf_openFromSkeletonAndLocale().
  *
  * NOTE: This is a C-compatible API; C++ users should build against numberformatter.h instead.
  *
  * @draft ICU 62
  */
 U_DRAFT void U_EXPORT2
-unumf_close(UNumberFormatter* uformatter, UErrorCode* ec);
+unumf_close(UNumberFormatter* uformatter);
+
+
+/**
+ * Releases the UFormattedNumber created by unumf_openResult().
+ *
+ * NOTE: This is a C-compatible API; C++ users should build against numberformatter.h instead.
+ *
+ * @draft ICU 62
+ */
+U_DRAFT void U_EXPORT2
+unumf_closeResult(const UFormattedNumber* uresult);
 
 
 #endif //__UNUMBERFORMATTER_H__

icu4c/source/test/cintltst/cintltst.c

@@ -729,4 +729,18 @@ U_CFUNC UBool assertUEquals(const char* message, const UChar* expected,
     return TRUE;
 }
 
+U_CFUNC UBool assertIntEquals(const char* message, int64_t expected, int64_t actual) {
+    if (expected != actual) {
+        log_err("FAIL: %s; got \"%d\"; expected \"%d\"\n",
+                message, actual, expected);
+        return FALSE;
+    }
+#ifdef VERBOSE_ASSERTIONS
+    else {
+        log_verbose("Ok: %s; got \"%d\"\n", message, actual);
+    }
+#endif
+    return TRUE;
+}
+
 #endif

icu4c/source/test/cintltst/cintltst.h

@@ -142,6 +142,11 @@ U_CFUNC UBool assertEquals(const char* msg, const char* expectedString,
 U_CFUNC UBool assertUEquals(const char* msg, const UChar* expectedString,
                             const UChar* actualString);
 
+/**
+ * Assert that two 64-bit integers are equal, returning TRUE if they do.
+ */
+U_CFUNC UBool assertIntEquals(const char* msg, int64_t expected, int64_t actual);
+
 /*
  * note - isICUVersionBefore and isICUVersionAtLeast have been removed.
  * use log_knownIssue() instead.

icu4c/source/test/cintltst/unumberformattertst.c

@@ -10,14 +10,19 @@
 #define UNISTR_FROM_STRING_EXPLICIT
 
 #include "unicode/unumberformatter.h"
+#include "unicode/umisc.h"
+#include "unicode/unum.h"
 #include "cintltst.h"
 
 static void TestSkeletonFormatToString();
 
+static void TestSkeletonFormatToFields();
+
 void addUNumberFormatterTest(TestNode** root);
 
 void addUNumberFormatterTest(TestNode** root) {
     addTest(root, &TestSkeletonFormatToString, "unumberformatter/TestSkeletonFormatToString");
+    addTest(root, &TestSkeletonFormatToFields, "unumberformatter/TestSkeletonFormatToFields");
 }
 
 
@@ -26,39 +31,102 @@ static void TestSkeletonFormatToString() {
     static const int32_t CAPACITY = 30;
     UChar buffer[CAPACITY];
 
-    // SETUP:
+    // setup:
     UNumberFormatter* f = unumf_openFromSkeletonAndLocale(
             u"round-integer currency/USD sign-accounting", -1, "en", &ec);
     assertSuccess("Should create without error", &ec);
     UFormattedNumber* result = unumf_openResult(&ec);
     assertSuccess("Should create result without error", &ec);
 
-    // INT TEST:
+    // int64 test:
     unumf_formatInt(f, -444444, result, &ec);
     assertSuccess("Should format integer without error", &ec);
     unumf_resultToString(result, buffer, CAPACITY, &ec);
     assertSuccess("Should print string to buffer without error", &ec);
     assertUEquals("Should produce expected string result", u"($444,444)", buffer);
 
-    // DOUBLE TEST:
+    // double test:
     unumf_formatDouble(f, -5142.3, result, &ec);
     assertSuccess("Should format double without error", &ec);
     unumf_resultToString(result, buffer, CAPACITY, &ec);
     assertSuccess("Should print string to buffer without error", &ec);
     assertUEquals("Should produce expected string result", u"($5,142)", buffer);
 
-    // DECIMAL TEST:
+    // decnumber test:
     unumf_formatDecimal(f, "9.876E2", -1, result, &ec);
     assertSuccess("Should format decimal without error", &ec);
     unumf_resultToString(result, buffer, CAPACITY, &ec);
     assertSuccess("Should print string to buffer without error", &ec);
     assertUEquals("Should produce expected string result", u"$988", buffer);
 
-    // CLEANUP:
-    unumf_closeResult(result, &ec);
-    assertSuccess("Should close without error", &ec);
-    unumf_close(f, &ec);
-    assertSuccess("Should close without error", &ec);
+    // cleanup:
+    unumf_closeResult(result);
+    unumf_close(f);
+}
+
+
+static void TestSkeletonFormatToFields() {
+    UErrorCode ec = U_ZERO_ERROR;
+
+    // setup:
+    UNumberFormatter* uformatter = unumf_openFromSkeletonAndLocale(
+            u".00 measure-unit/length-meter sign-always", -1, "en", &ec);
+    assertSuccess("Should create without error", &ec);
+    UFormattedNumber* uresult = unumf_openResult(&ec);
+    assertSuccess("Should create result without error", &ec);
+    unumf_formatInt(uformatter, 9876543210L, uresult, &ec); // "+9,876,543,210.00 m"
+
+    // field position test:
+    UFieldPosition ufpos = {UNUM_DECIMAL_SEPARATOR_FIELD};
+    unumf_resultGetField(uresult, &ufpos, &ec);
+    assertIntEquals("Field position should be correct", 14, ufpos.beginIndex);
+    assertIntEquals("Field position should be correct", 15, ufpos.endIndex);
+
+    // field position iterator test:
+    UFieldPositionIterator* ufpositer = ufieldpositer_open(&ec);
+    assertSuccess("Should create iterator without error", &ec);
+    unumf_resultGetAllFields(uresult, ufpositer, &ec);
+    static const UFieldPosition expectedFields[] = {
+            // Field, begin index, end index
+            {UNUM_SIGN_FIELD, 0, 1},
+            {UNUM_GROUPING_SEPARATOR_FIELD, 2, 3},
+            {UNUM_GROUPING_SEPARATOR_FIELD, 6, 7},
+            {UNUM_GROUPING_SEPARATOR_FIELD, 10, 11},
+            {UNUM_INTEGER_FIELD, 1, 14},
+            {UNUM_DECIMAL_SEPARATOR_FIELD, 14, 15},
+            {UNUM_FRACTION_FIELD, 15, 17}};
+    UFieldPosition actual;
+    for (int32_t i = 0; i < sizeof(expectedFields) / sizeof(*expectedFields); i++) {
+        // Iterate using the UFieldPosition to hold state...
+        UFieldPosition expected = expectedFields[i];
+        actual.field = ufieldpositer_next(ufpositer, &actual.beginIndex, &actual.endIndex);
+        assertTrue("Should not return a negative index yet", actual.field >= 0);
+        if (expected.field != actual.field) {
+            log_err(
+                    "FAIL: iteration %d; expected field %d; got %d\n", i, expected.field, actual.field);
+        }
+        if (expected.beginIndex != actual.beginIndex) {
+            log_err(
+                    "FAIL: iteration %d; expected beginIndex %d; got %d\n",
+                    i,
+                    expected.beginIndex,
+                    actual.beginIndex);
+        }
+        if (expected.endIndex != actual.endIndex) {
+            log_err(
+                    "FAIL: iteration %d; expected endIndex %d; got %d\n",
+                    i,
+                    expected.endIndex,
+                    actual.endIndex);
+        }
+    }
+    actual.field = ufieldpositer_next(ufpositer, &actual.beginIndex, &actual.endIndex);
+    assertTrue("No more fields; should return a negative index", actual.field < 0);
+
+    // cleanup:
+    unumf_closeResult(uresult);
+    unumf_close(uformatter);
+    ufieldpositer_close(ufpositer);
 }