]> git.saurik.com Git - apple/icu.git/blobdiff - icuSources/test/intltest/utxttest.cpp
ICU-8.11.tar.gz
[apple/icu.git] / icuSources / test / intltest / utxttest.cpp
diff --git a/icuSources/test/intltest/utxttest.cpp b/icuSources/test/intltest/utxttest.cpp
new file mode 100644 (file)
index 0000000..da6a4af
--- /dev/null
@@ -0,0 +1,1302 @@
+/********************************************************************
+ * COPYRIGHT: 
+ * Copyright (c) 2005-2006, International Business Machines Corporation and
+ * others. All Rights Reserved.
+ ********************************************************************/
+/************************************************************************
+*   Tests for the UText and UTextIterator text abstraction classses
+*
+************************************************************************/
+
+#include "unicode/utypes.h"
+
+#include <string.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <unicode/utext.h>
+#include <unicode/utf8.h>
+#include <unicode/ustring.h>
+#include <unicode/uchriter.h>
+#include "utxttest.h"
+
+static UBool  gFailed = FALSE;
+static int    gTestNum = 0;
+
+// Forward decl
+UText *openFragmentedUnicodeString(UText *ut, UnicodeString *s, UErrorCode *status);
+
+#define TEST_ASSERT(x) \
+{ if ((x)==FALSE) {errln("Test #%d failure in file %s at line %d\n", gTestNum, __FILE__, __LINE__);\
+                     gFailed = TRUE;\
+   }}
+
+
+#define TEST_SUCCESS(status) \
+{ if (U_FAILURE(status)) {errln("Test #%d failure in file %s at line %d. Error = \"%s\"\n", \
+       gTestNum, __FILE__, __LINE__, u_errorName(status)); \
+       gFailed = TRUE;\
+   }}
+
+UTextTest::UTextTest() {
+}
+
+UTextTest::~UTextTest() {
+}
+
+
+void
+UTextTest::runIndexedTest(int32_t index, UBool exec,
+                          const char* &name, char* /*par*/) {
+    switch (index) {
+        case 0: name = "TextTest";
+            if (exec) TextTest();    break;
+        case 1: name = "ErrorTest";
+            if (exec) ErrorTest();   break;
+        case 2: name = "FreezeTest";
+            if (exec) FreezeTest();  break;
+        default: name = "";          break;
+    }
+}
+
+//
+// Quick and dirty random number generator.
+//   (don't use library so that results are portable.
+static uint32_t m_seed = 1;
+static uint32_t m_rand()
+{
+    m_seed = m_seed * 1103515245 + 12345;
+    return (uint32_t)(m_seed/65536) % 32768;
+}
+
+
+//
+//   TextTest()
+//
+//       Top Level function for UText testing.
+//       Specifies the strings to be tested, with the acutal testing itself
+//       being carried out in another function, TestString().
+//
+void  UTextTest::TextTest() {
+    int32_t i, j;
+
+    TestString("abcd\\U00010001xyz");
+    TestString("");
+
+    // Supplementary chars at start or end
+    TestString("\\U00010001");
+    TestString("abc\\U00010001");
+    TestString("\\U00010001abc");
+
+    // Test simple strings of lengths 1 to 60, looking for glitches at buffer boundaries
+    UnicodeString s;
+    for (i=1; i<60; i++) {
+        s.truncate(0);
+        for (j=0; j<i; j++) {
+            if (j+0x30 == 0x5c) {
+                // backslash.  Needs to be escaped
+                s.append((UChar)0x5c);
+            }
+            s.append(UChar(j+0x30));
+        }
+        TestString(s);
+    }
+
+   // Test strings with odd-aligned supplementary chars,
+   //    looking for glitches at buffer boundaries
+    for (i=1; i<60; i++) {
+        s.truncate(0);
+        s.append((UChar)0x41);
+        for (j=0; j<i; j++) {
+            s.append(UChar32(j+0x11000));
+        }
+        TestString(s);
+    }
+
+    // String of chars of randomly varying size in utf-8 representation.
+    //   Exercise the mapping, and the varying sized buffer.
+    //
+    s.truncate(0);
+    UChar32  c1 = 0;
+    UChar32  c2 = 0x100;
+    UChar32  c3 = 0xa000;
+    UChar32  c4 = 0x11000;
+    for (i=0; i<1000; i++) {
+        int len8 = m_rand()%4 + 1;
+        switch (len8) {
+            case 1: 
+                c1 = (c1+1)%0x80;
+                // don't put 0 into string (0 terminated strings for some tests)
+                // don't put '\', will cause unescape() to fail.
+                if (c1==0x5c || c1==0) {
+                    c1++;
+                }
+                s.append(c1);
+                break;
+            case 2:
+                s.append(c2++);
+                break;
+            case 3:
+                s.append(c3++);
+                break;
+            case 4:
+                s.append(c4++);
+                break;
+        }
+    }
+    TestString(s);
+}
+
+
+//
+//  TestString()     Run a suite of UText tests on a string.
+//                   The test string is unescaped before use.
+//
+void UTextTest::TestString(const UnicodeString &s) {
+    int32_t       i;
+    int32_t       j;
+    UChar32       c;
+    int32_t       cpCount = 0;
+    UErrorCode    status  = U_ZERO_ERROR;
+    UText        *ut      = NULL;
+    int32_t       saLen;
+
+    UnicodeString sa = s.unescape();
+    saLen = sa.length();
+
+    //
+    // Build up a mapping between code points and UTF-16 code unit indexes.
+    //
+    m *cpMap = new m[sa.length() + 1];
+    j = 0;
+    for (i=0; i<sa.length(); i=sa.moveIndex32(i, 1)) {
+        c = sa.char32At(i);
+        cpMap[j].nativeIdx = i;
+        cpMap[j].cp = c;
+        j++;
+        cpCount++;
+    }
+    cpMap[j].nativeIdx = i;   // position following the last char in utf-16 string.    
+
+
+    // UChar * test, null terminated
+    status = U_ZERO_ERROR;
+    UChar *buf = new UChar[saLen+1];
+    sa.extract(buf, saLen+1, status);
+    TEST_SUCCESS(status);
+    ut = utext_openUChars(NULL, buf, -1, &status);
+    TEST_SUCCESS(status);
+    TestAccess(sa, ut, cpCount, cpMap);
+    utext_close(ut);
+    delete [] buf;
+
+    // UChar * test, with length
+    status = U_ZERO_ERROR;
+    buf = new UChar[saLen+1];
+    sa.extract(buf, saLen+1, status);
+    TEST_SUCCESS(status);
+    ut = utext_openUChars(NULL, buf, saLen, &status);
+    TEST_SUCCESS(status);
+    TestAccess(sa, ut, cpCount, cpMap);
+    utext_close(ut);
+    delete [] buf;
+
+
+    // UnicodeString test
+    status = U_ZERO_ERROR;
+    ut = utext_openUnicodeString(NULL, &sa, &status);
+    TEST_SUCCESS(status);
+    TestAccess(sa, ut, cpCount, cpMap);
+    TestCMR(sa, ut, cpCount, cpMap, cpMap);
+    utext_close(ut);
+
+
+    // Const UnicodeString test
+    status = U_ZERO_ERROR;
+    ut = utext_openConstUnicodeString(NULL, &sa, &status);
+    TEST_SUCCESS(status);
+    TestAccess(sa, ut, cpCount, cpMap);
+    utext_close(ut);
+
+
+    // Replaceable test.  (UnicodeString inherits Replaceable)
+    status = U_ZERO_ERROR;
+    ut = utext_openReplaceable(NULL, &sa, &status);
+    TEST_SUCCESS(status);
+    TestAccess(sa, ut, cpCount, cpMap);
+    TestCMR(sa, ut, cpCount, cpMap, cpMap);
+    utext_close(ut);
+
+    // Character Iterator Tests
+    status = U_ZERO_ERROR;
+    const UChar *cbuf = sa.getBuffer();
+    CharacterIterator *ci = new UCharCharacterIterator(cbuf, saLen, status);
+    TEST_SUCCESS(status);
+    ut = utext_openCharacterIterator(NULL, ci, &status);
+    TEST_SUCCESS(status);
+    TestAccess(sa, ut, cpCount, cpMap);
+    utext_close(ut);
+    delete ci;
+    
+
+    // Fragmented UnicodeString  (Chunk size of one)
+    //
+    status = U_ZERO_ERROR;
+    ut = openFragmentedUnicodeString(NULL, &sa, &status);
+    TEST_SUCCESS(status);
+    TestAccess(sa, ut, cpCount, cpMap);
+    utext_close(ut);
+
+    //
+    // UTF-8 test
+    //
+
+    // Convert the test string from UnicodeString to (char *) in utf-8 format
+    int32_t u8Len = sa.extract(0, sa.length(), NULL, 0, "utf-8");
+    char *u8String = new char[u8Len + 1];
+    sa.extract(0, sa.length(), u8String, u8Len+1, "utf-8");
+
+    // Build up the map of code point indices in the utf-8 string
+    m * u8Map = new m[sa.length() + 1];
+    i = 0;   // native utf-8 index
+    for (j=0; j<cpCount ; j++) {  // code point number
+        u8Map[j].nativeIdx = i;
+        U8_NEXT(u8String, i, u8Len, c)
+        u8Map[j].cp = c;
+    }
+    u8Map[cpCount].nativeIdx = u8Len;   // position following the last char in utf-8 string.
+
+    // Do the test itself
+    status = U_ZERO_ERROR;
+    ut = utext_openUTF8(NULL, u8String, -1, &status);
+    TEST_SUCCESS(status);
+    TestAccess(sa, ut, cpCount, u8Map);
+    utext_close(ut);
+
+
+
+       delete []cpMap;
+       delete []u8Map;
+       delete []u8String;
+}
+
+//  TestCMR   test Copy, Move and Replace operations.
+//              us         UnicodeString containing the test text.
+//              ut         UText containing the same test text.
+//              cpCount    number of code points in the test text.
+//              nativeMap  Mapping from code points to native indexes for the UText.
+//              u16Map     Mapping from code points to UTF-16 indexes, for use with the UnicodeString.
+//
+//     This function runs a whole series of opertions on each incoming UText.
+//     The UText is deep-cloned prior to each operation, so that the original UText remains unchanged.
+//     
+void UTextTest::TestCMR(const UnicodeString &us, UText *ut, int cpCount, m *nativeMap, m *u16Map) {
+    TEST_ASSERT(utext_isWritable(ut) == TRUE);
+
+    int  srcLengthType;       // Loop variables for selecting the postion and length
+    int  srcPosType;          //   of the block to operate on within the source text.
+    int  destPosType; 
+
+    int  srcIndex  = 0;       // Code Point indexes of the block to operate on for
+    int  srcLength = 0;       //   a specific test.
+
+    int  destIndex = 0;       // Code point index of the destination for a copy/move test.
+
+    int32_t  nativeStart = 0; // Native unit indexes for a test.
+    int32_t  nativeLimit = 0;
+    int32_t  nativeDest  = 0;
+
+    int32_t  u16Start    = 0; // UTF-16 indexes for a test.
+    int32_t  u16Limit    = 0; //   used when performing the same operation in a Unicode String
+    int32_t  u16Dest     = 0;
+
+    // Iterate over a whole series of source index, length and a target indexes.
+    // This is done with code point indexes; these will be later translated to native
+    //   indexes using the cpMap.
+    for (srcLengthType=1; srcLengthType<=3; srcLengthType++) {
+        switch (srcLengthType) {
+            case 1: srcLength = 1; break;
+            case 2: srcLength = 5; break;
+            case 3: srcLength = cpCount / 3;
+        }
+        for (srcPosType=1; srcPosType<=5; srcPosType++) {
+            switch (srcPosType) {
+                case 1: srcIndex = 0; break;
+                case 2: srcIndex = 1; break;
+                case 3: srcIndex = cpCount - srcLength; break;
+                case 4: srcIndex = cpCount - srcLength - 1; break;
+                case 5: srcIndex = cpCount / 2; break;
+            }
+            if (srcIndex < 0 || srcIndex + srcLength > cpCount) {
+                // filter out bogus test cases - 
+                //   those with a source range that falls of an edge of the string.
+                continue;
+            }
+
+            //
+            // Copy and move tests.
+            //   iterate over a variety of destination positions.
+            //
+            for (destPosType=1; destPosType<=4; destPosType++) {
+                switch (destPosType) {
+                    case 1: destIndex = 0; break;
+                    case 2: destIndex = 1; break;
+                    case 3: destIndex = srcIndex - 1; break;
+                    case 4: destIndex = srcIndex + srcLength + 1; break;
+                    case 5: destIndex = cpCount-1; break;
+                    case 6: destIndex = cpCount; break;
+                }
+                if (destIndex<0 || destIndex>cpCount) {
+                    // filter out bogus test cases.
+                    continue;
+                }
+
+                nativeStart = nativeMap[srcIndex].nativeIdx;
+                nativeLimit = nativeMap[srcIndex+srcLength].nativeIdx;
+                nativeDest  = nativeMap[destIndex].nativeIdx;
+
+                u16Start    = u16Map[srcIndex].nativeIdx;
+                u16Limit    = u16Map[srcIndex+srcLength].nativeIdx;
+                u16Dest     = u16Map[destIndex].nativeIdx;
+
+                gFailed = FALSE;
+                TestCopyMove(us, ut, FALSE,
+                    nativeStart, nativeLimit, nativeDest,
+                    u16Start, u16Limit, u16Dest);
+
+                TestCopyMove(us, ut, TRUE,
+                    nativeStart, nativeLimit, nativeDest,
+                    u16Start, u16Limit, u16Dest);
+
+                if (gFailed) {
+                    return;
+                }
+            }
+
+            //
+            //  Replace tests.
+            //
+            UnicodeString fullRepString("This is an arbitrary string that will be used as replacement text");
+            for (int32_t replStrLen=0; replStrLen<20; replStrLen++) {
+                UnicodeString repStr(fullRepString, 0, replStrLen);
+                TestReplace(us, ut,
+                    nativeStart, nativeLimit,
+                    u16Start, u16Limit,
+                    repStr);
+                if (gFailed) {
+                    return;
+                }
+            }
+
+        }
+    }
+
+}
+
+//
+//   TestCopyMove    run a single test case for utext_copy.
+//                   Test cases are created in TestCMR and dispatched here for execution.
+//
+void UTextTest::TestCopyMove(const UnicodeString &us, UText *ut, UBool move,
+                    int32_t nativeStart, int32_t nativeLimit, int32_t nativeDest,
+                    int32_t u16Start, int32_t u16Limit, int32_t u16Dest) 
+{
+    UErrorCode      status   = U_ZERO_ERROR;
+    UText          *targetUT = NULL;
+    gTestNum++;
+    gFailed = FALSE;
+
+    //
+    //  clone the UText.  The test will be run in the cloned copy
+    //  so that we don't alter the original.
+    //
+    targetUT = utext_clone(NULL, ut, TRUE, FALSE, &status);
+    TEST_SUCCESS(status);
+    UnicodeString targetUS(us);    // And copy the reference string.
+
+    // do the test operation first in the reference
+    targetUS.copy(u16Start, u16Limit, u16Dest);
+    if (move) {
+        // delete out the source range.
+        if (u16Limit < u16Dest) {
+            targetUS.removeBetween(u16Start, u16Limit);
+        } else {
+            int32_t amtCopied = u16Limit - u16Start;
+            targetUS.removeBetween(u16Start+amtCopied, u16Limit+amtCopied);
+        }
+    }
+
+    // Do the same operation in the UText under test
+    utext_copy(targetUT, nativeStart, nativeLimit, nativeDest, move, &status);
+    if (nativeDest > nativeStart && nativeDest < nativeLimit) {
+        TEST_ASSERT(status == U_INDEX_OUTOFBOUNDS_ERROR);
+    } else {
+        TEST_SUCCESS(status);
+
+        // Compare the results of the two parallel tests
+        int32_t  usi = 0;    // UnicodeString postion, utf-16 index.
+        int64_t  uti = 0;    // UText position, native index.
+        int32_t  cpi;        // char32 position (code point index) 
+        UChar32  usc;        // code point from Unicode String
+        UChar32  utc;        // code point from UText
+        utext_setNativeIndex(targetUT, 0);
+        for (cpi=0; ; cpi++) {
+            usc = targetUS.char32At(usi);
+            utc = utext_next32(targetUT);
+            if (utc < 0) {
+                break;
+            }
+            TEST_ASSERT(uti == usi);
+            TEST_ASSERT(utc == usc);
+            usi = targetUS.moveIndex32(usi, 1);
+            uti = utext_getNativeIndex(targetUT);
+            if (gFailed) {
+                goto cleanupAndReturn;
+            }
+        }
+        int64_t expectedNativeLength = utext_nativeLength(ut);
+        if (move == FALSE) {
+            expectedNativeLength += nativeLimit - nativeStart;
+        }
+        uti = utext_getNativeIndex(targetUT);
+        TEST_ASSERT(uti == expectedNativeLength);
+    }
+
+cleanupAndReturn:
+    utext_close(targetUT);
+}
+    
+
+//
+//  TestReplace   Test a single Replace operation.
+//
+void UTextTest::TestReplace(
+            const UnicodeString &us,     // reference UnicodeString in which to do the replace 
+            UText         *ut,                // UnicodeText object under test.
+            int32_t       nativeStart,        // Range to be replaced, in UText native units. 
+            int32_t       nativeLimit,
+            int32_t       u16Start,           // Range to be replaced, in UTF-16 units
+            int32_t       u16Limit,           //    for use in the reference UnicodeString.
+            const UnicodeString &repStr)      // The replacement string
+{
+    UErrorCode      status   = U_ZERO_ERROR;
+    UText          *targetUT = NULL;
+    gTestNum++;
+    gFailed = FALSE;
+
+    //
+    //  clone the target UText.  The test will be run in the cloned copy
+    //  so that we don't alter the original.
+    //
+    targetUT = utext_clone(NULL, ut, TRUE, FALSE, &status);
+    TEST_SUCCESS(status);
+    UnicodeString targetUS(us);    // And copy the reference string.
+
+    //
+    // Do the replace operation in the Unicode String, to 
+    //   produce a reference result.
+    //
+    targetUS.replace(u16Start, u16Limit-u16Start, repStr);
+
+    //
+    // Do the replace on the UText under test
+    //
+    const UChar *rs = repStr.getBuffer();
+    int32_t  rsLen = repStr.length();
+    int32_t actualDelta = utext_replace(targetUT, nativeStart, nativeLimit, rs, rsLen, &status);
+    int32_t expectedDelta = repStr.length() - (nativeLimit - nativeStart);
+    TEST_ASSERT(actualDelta == expectedDelta);
+
+    //
+    // Compare the results
+    //
+    int32_t  usi = 0;    // UnicodeString postion, utf-16 index.
+    int64_t  uti = 0;    // UText position, native index.
+    int32_t  cpi;        // char32 position (code point index) 
+    UChar32  usc;        // code point from Unicode String
+    UChar32  utc;        // code point from UText
+    int64_t  expectedNativeLength = 0;
+    utext_setNativeIndex(targetUT, 0);
+    for (cpi=0; ; cpi++) {
+        usc = targetUS.char32At(usi);
+        utc = utext_next32(targetUT);
+        if (utc < 0) {
+            break;
+        }
+        TEST_ASSERT(uti == usi);
+        TEST_ASSERT(utc == usc);
+        usi = targetUS.moveIndex32(usi, 1);
+        uti = utext_getNativeIndex(targetUT);
+        if (gFailed) {
+            goto cleanupAndReturn;
+        }
+    }
+    expectedNativeLength = utext_nativeLength(ut) + expectedDelta;
+    uti = utext_getNativeIndex(targetUT);
+    TEST_ASSERT(uti == expectedNativeLength);
+
+cleanupAndReturn:
+    utext_close(targetUT);
+}
+
+//
+//  TestAccess()    Test the read only access functions on a UText.
+//                  The text is accessed in a variety of ways, and compared with
+//                  the reference UnicodeString.
+//
+void UTextTest::TestAccess(const UnicodeString &us, UText *ut, int cpCount, m *cpMap) {
+    UErrorCode  status = U_ZERO_ERROR;
+    gTestNum++;
+
+    //
+    //  Check the length from the UText
+    //
+    int64_t expectedLen = cpMap[cpCount].nativeIdx;
+    int64_t utlen = utext_nativeLength(ut);
+    TEST_ASSERT(expectedLen == utlen);
+
+    //
+    //  Iterate forwards, verify that we get the correct code points
+    //   at the correct native offsets.
+    //
+    int         i = 0;
+    int64_t     index;
+    int64_t     expectedIndex = 0;
+    int64_t     foundIndex = 0;
+    UChar32     expectedC;
+    UChar32     foundC;
+    int64_t     len;
+
+    for (i=0; i<cpCount; i++) {
+        expectedIndex = cpMap[i].nativeIdx;
+        foundIndex    = utext_getNativeIndex(ut);
+        TEST_ASSERT(expectedIndex == foundIndex);
+        expectedC     = cpMap[i].cp;
+        foundC        = utext_next32(ut);    
+        TEST_ASSERT(expectedC == foundC);
+        foundIndex    = utext_getPreviousNativeIndex(ut);
+        TEST_ASSERT(expectedIndex == foundIndex);
+        if (gFailed) {
+            return;
+        }
+    }
+    foundC = utext_next32(ut);
+    TEST_ASSERT(foundC == U_SENTINEL);
+    
+    // Repeat above, using macros
+    utext_setNativeIndex(ut, 0);
+    for (i=0; i<cpCount; i++) {
+        expectedIndex = cpMap[i].nativeIdx;
+        foundIndex    = UTEXT_GETNATIVEINDEX(ut);
+        TEST_ASSERT(expectedIndex == foundIndex);
+        expectedC     = cpMap[i].cp;
+        foundC        = UTEXT_NEXT32(ut);    
+        TEST_ASSERT(expectedC == foundC);
+        if (gFailed) {
+            return;
+        }
+    }
+    foundC = UTEXT_NEXT32(ut);
+    TEST_ASSERT(foundC == U_SENTINEL);
+
+    //
+    //  Forward iteration (above) should have left index at the
+    //   end of the input, which should == length().
+    //
+    len = utext_nativeLength(ut);
+    foundIndex  = utext_getNativeIndex(ut);
+    TEST_ASSERT(len == foundIndex);
+
+    //
+    // Iterate backwards over entire test string
+    //
+    len = utext_getNativeIndex(ut);
+    utext_setNativeIndex(ut, len);
+    for (i=cpCount-1; i>=0; i--) {
+        expectedC     = cpMap[i].cp;
+        expectedIndex = cpMap[i].nativeIdx;
+        int64_t prevIndex = utext_getPreviousNativeIndex(ut);
+        foundC        = utext_previous32(ut);
+        foundIndex    = utext_getNativeIndex(ut);
+        TEST_ASSERT(expectedIndex == foundIndex);
+        TEST_ASSERT(expectedC == foundC);
+        TEST_ASSERT(prevIndex == foundIndex);
+        if (gFailed) {
+            return;
+        }
+    }
+
+    //
+    //  Backwards iteration, above, should have left our iterator
+    //   position at zero, and continued backwards iterationshould fail.
+    //
+    foundIndex = utext_getNativeIndex(ut);
+    TEST_ASSERT(foundIndex == 0);
+    foundIndex = utext_getPreviousNativeIndex(ut);
+    TEST_ASSERT(foundIndex == 0);
+
+
+    foundC = utext_previous32(ut);
+    TEST_ASSERT(foundC == U_SENTINEL);
+    foundIndex = utext_getNativeIndex(ut);
+    TEST_ASSERT(foundIndex == 0);
+    foundIndex = utext_getPreviousNativeIndex(ut);
+    TEST_ASSERT(foundIndex == 0);
+
+
+    // And again, with the macros
+    utext_setNativeIndex(ut, len);
+    for (i=cpCount-1; i>=0; i--) {
+        expectedC     = cpMap[i].cp;
+        expectedIndex = cpMap[i].nativeIdx;
+        foundC        = UTEXT_PREVIOUS32(ut);
+        foundIndex    = UTEXT_GETNATIVEINDEX(ut);
+        TEST_ASSERT(expectedIndex == foundIndex);
+        TEST_ASSERT(expectedC == foundC);
+        if (gFailed) {
+            return;
+        }
+    }
+
+    //
+    //  Backwards iteration, above, should have left our iterator
+    //   position at zero, and continued backwards iterationshould fail.
+    //
+    foundIndex = UTEXT_GETNATIVEINDEX(ut);
+    TEST_ASSERT(foundIndex == 0);
+
+    foundC = UTEXT_PREVIOUS32(ut);
+    TEST_ASSERT(foundC == U_SENTINEL);
+    foundIndex = UTEXT_GETNATIVEINDEX(ut);
+    TEST_ASSERT(foundIndex == 0);
+    if (gFailed) {
+        return;
+    }
+
+    //
+    //  next32From(), prevous32From(), Iterate in a somewhat random order.
+    //
+    int  cpIndex = 0;
+    for (i=0; i<cpCount; i++) {
+        cpIndex = (cpIndex + 9973) % cpCount;
+        index         = cpMap[cpIndex].nativeIdx;
+        expectedC     = cpMap[cpIndex].cp;
+        foundC        = utext_next32From(ut, index);
+        TEST_ASSERT(expectedC == foundC);
+        if (gFailed) {
+            return;
+        }
+    }
+
+    cpIndex = 0;
+    for (i=0; i<cpCount; i++) {
+        cpIndex = (cpIndex + 9973) % cpCount;
+        index         = cpMap[cpIndex+1].nativeIdx;
+        expectedC     = cpMap[cpIndex].cp;
+        foundC        = utext_previous32From(ut, index);
+        TEST_ASSERT(expectedC == foundC);
+        if (gFailed) {
+            return;
+        }
+    }
+
+
+    //
+    // moveIndex(int32_t delta);
+    //
+
+    // Walk through frontwards, incrementing by one
+    utext_setNativeIndex(ut, 0);
+    for (i=1; i<=cpCount; i++) {
+        utext_moveIndex32(ut, 1);
+        index = utext_getNativeIndex(ut);
+        expectedIndex = cpMap[i].nativeIdx;
+        TEST_ASSERT(expectedIndex == index);
+        index = UTEXT_GETNATIVEINDEX(ut);
+        TEST_ASSERT(expectedIndex == index);
+    }
+
+    // Walk through frontwards, incrementing by two
+    utext_setNativeIndex(ut, 0);
+    for (i=2; i<cpCount; i+=2) {
+        utext_moveIndex32(ut, 2);
+        index = utext_getNativeIndex(ut);
+        expectedIndex = cpMap[i].nativeIdx;
+        TEST_ASSERT(expectedIndex == index);
+        index = UTEXT_GETNATIVEINDEX(ut);
+        TEST_ASSERT(expectedIndex == index);
+    }
+
+    // walk through the string backwards, decrementing by one.
+    i = cpMap[cpCount].nativeIdx;
+    utext_setNativeIndex(ut, i);
+    for (i=cpCount; i>=0; i--) {
+        expectedIndex = cpMap[i].nativeIdx;
+        index = utext_getNativeIndex(ut);
+        TEST_ASSERT(expectedIndex == index);
+        index = UTEXT_GETNATIVEINDEX(ut);
+        TEST_ASSERT(expectedIndex == index);
+        utext_moveIndex32(ut, -1);
+    }
+
+
+    // walk through backwards, decrementing by three
+    i = cpMap[cpCount].nativeIdx;
+    utext_setNativeIndex(ut, i);
+    for (i=cpCount; i>=0; i-=3) {
+        expectedIndex = cpMap[i].nativeIdx;
+        index = utext_getNativeIndex(ut);
+        TEST_ASSERT(expectedIndex == index);
+        index = UTEXT_GETNATIVEINDEX(ut);
+        TEST_ASSERT(expectedIndex == index);
+        utext_moveIndex32(ut, -3);
+    }
+
+
+    //
+    // Extract
+    //
+    int bufSize = us.length() + 10;
+    UChar *buf = new UChar[bufSize];
+    status = U_ZERO_ERROR;
+    expectedLen = us.length();
+    len = utext_extract(ut, 0, utlen, buf, bufSize, &status);
+    TEST_SUCCESS(status);
+    TEST_ASSERT(len == expectedLen);
+    int compareResult = us.compare(buf, -1);
+    TEST_ASSERT(compareResult == 0);
+
+    status = U_ZERO_ERROR;
+    len = utext_extract(ut, 0, utlen, NULL, 0, &status);
+    if (utlen == 0) {
+        TEST_ASSERT(status == U_STRING_NOT_TERMINATED_WARNING);
+    } else {
+        TEST_ASSERT(status == U_BUFFER_OVERFLOW_ERROR);
+    }
+    TEST_ASSERT(len == expectedLen);
+
+    status = U_ZERO_ERROR;
+    u_memset(buf, 0x5555, bufSize);
+    len = utext_extract(ut, 0, utlen, buf, 1, &status);
+    if (us.length() == 0) {
+        TEST_SUCCESS(status);
+        TEST_ASSERT(buf[0] == 0);
+    } else {
+        // Buf len == 1, extracting a single 16 bit value.
+        // If the data char is supplementary, it doesn't matter whether the buffer remains unchanged,
+        //   or whether the lead surrogate of the pair is extracted.
+        //   It's a buffer overflow error in either case.
+        TEST_ASSERT(buf[0] == us.charAt(0) ||
+                    buf[0] == 0x5555 && U_IS_SUPPLEMENTARY(us.char32At(0)));
+        TEST_ASSERT(buf[1] == 0x5555);
+        if (us.length() == 1) {
+            TEST_ASSERT(status == U_STRING_NOT_TERMINATED_WARNING);
+        } else {
+            TEST_ASSERT(status == U_BUFFER_OVERFLOW_ERROR);
+        }
+    }
+
+    delete []buf;
+}
+
+
+
+//
+//  ErrorTest()    Check various error and edge cases.
+//
+void UTextTest::ErrorTest() 
+{
+    // Close of an unitialized UText.  Shouldn't blow up.
+    {
+        UText  ut;  
+        memset(&ut, 0, sizeof(UText));
+        utext_close(&ut);
+        utext_close(NULL);
+    }
+
+    // Double-close of a UText.  Shouldn't blow up.  UText should still be usable.
+    {
+        UErrorCode status = U_ZERO_ERROR;
+        UText ut = UTEXT_INITIALIZER;
+        UnicodeString s("Hello, World");
+        UText *ut2 = utext_openUnicodeString(&ut, &s, &status);
+        TEST_SUCCESS(status);
+        TEST_ASSERT(ut2 == &ut);
+
+        UText *ut3 = utext_close(&ut);
+        TEST_ASSERT(ut3 == &ut);
+
+        UText *ut4 = utext_close(&ut);
+        TEST_ASSERT(ut4 == &ut);
+
+        utext_openUnicodeString(&ut, &s, &status);
+        TEST_SUCCESS(status);
+        utext_close(&ut);
+    }
+
+    // Re-use of a UText, chaining through each of the types of UText
+    //   (If it doesn't blow up, and doesn't leak, it's probably working fine)
+    {
+        UErrorCode status = U_ZERO_ERROR;
+        UText ut = UTEXT_INITIALIZER;
+        UText  *utp;
+        UnicodeString s1("Hello, World");
+        UChar s2[] = {(UChar)0x41, (UChar)0x42, (UChar)0};
+        const char  *s3 = "\x66\x67\x68";
+
+        utp = utext_openUnicodeString(&ut, &s1, &status);
+        TEST_SUCCESS(status);
+        TEST_ASSERT(utp == &ut);
+
+        utp = utext_openConstUnicodeString(&ut, &s1, &status);
+        TEST_SUCCESS(status);
+        TEST_ASSERT(utp == &ut);
+
+        utp = utext_openUTF8(&ut, s3, -1, &status);
+        TEST_SUCCESS(status);
+        TEST_ASSERT(utp == &ut);
+
+        utp = utext_openUChars(&ut, s2, -1, &status);
+        TEST_SUCCESS(status);
+        TEST_ASSERT(utp == &ut);
+
+        utp = utext_close(&ut);
+        TEST_ASSERT(utp == &ut);
+
+        utp = utext_openUnicodeString(&ut, &s1, &status);
+        TEST_SUCCESS(status);
+        TEST_ASSERT(utp == &ut);
+    }
+
+    //
+    //  UTF-8 with malformed sequences.
+    //    These should come through as the Unicode replacement char, \ufffd
+    //
+    {
+        UErrorCode status = U_ZERO_ERROR;
+        UText *ut = NULL;
+        const char *badUTF8 = "\x41\x81\x42\xf0\x81\x81\x43";   
+        UChar32  c;
+
+        ut = utext_openUTF8(NULL, badUTF8, -1, &status);
+        TEST_SUCCESS(status);
+        c = utext_char32At(ut, 1);
+        TEST_ASSERT(c == 0xfffd);
+        c = utext_char32At(ut, 3);
+        TEST_ASSERT(c == 0xfffd);
+        c = utext_char32At(ut, 5);
+        TEST_ASSERT(c == 0xfffd);
+        c = utext_char32At(ut, 6);
+        TEST_ASSERT(c == 0x43);
+
+        UChar buf[10];
+        int n = utext_extract(ut, 0, 9, buf, 10, &status);
+        TEST_SUCCESS(status);
+        TEST_ASSERT(n==5);
+        TEST_ASSERT(buf[1] == 0xfffd);
+        TEST_ASSERT(buf[3] == 0xfffd);
+        TEST_ASSERT(buf[2] == 0x42);
+        utext_close(ut);
+    }
+
+
+    //
+    //  isLengthExpensive - does it make the exptected transitions after
+    //                      getting the length of a nul terminated string?
+    //
+    {
+        UErrorCode status = U_ZERO_ERROR;
+        UnicodeString sa("Hello, this is a string");
+        UBool  isExpensive;
+
+        UChar sb[100];
+        memset(sb, 0x20, sizeof(sb));
+        sb[99] = 0;
+
+        UText *uta = utext_openUnicodeString(NULL, &sa, &status);
+        TEST_SUCCESS(status);
+        isExpensive = utext_isLengthExpensive(uta);
+        TEST_ASSERT(isExpensive == FALSE);
+        utext_close(uta);
+
+        UText *utb = utext_openUChars(NULL, sb, -1, &status);
+        TEST_SUCCESS(status);
+        isExpensive = utext_isLengthExpensive(utb);
+        TEST_ASSERT(isExpensive == TRUE);
+        int64_t  len = utext_nativeLength(utb);
+        TEST_ASSERT(len == 99);
+        isExpensive = utext_isLengthExpensive(utb);
+        TEST_ASSERT(isExpensive == FALSE);
+        utext_close(utb);
+    }
+
+    //
+    // Index to positions not on code point boundaries.
+    //
+    {
+        const char *u8str =         "\xc8\x81\xe1\x82\x83\xf1\x84\x85\x86";
+        int32_t startMap[] =        {   0,  0,  2,  2,  2,  5,  5,  5,  5,  9,  9};
+        int32_t nextMap[]  =        {   2,  2,  5,  5,  5,  9,  9,  9,  9,  9,  9};
+        int32_t prevMap[]  =        {   0,  0,  0,  0,  0,  2,  2,  2,  2,  5,  5};
+        UChar32  c32Map[] =    {0x201, 0x201, 0x1083, 0x1083, 0x1083, 0x044146, 0x044146, 0x044146, 0x044146, -1, -1}; 
+        UChar32  pr32Map[] =   {    -1,   -1,  0x201,  0x201,  0x201,   0x1083,   0x1083,   0x1083,   0x1083, 0x044146, 0x044146}; 
+
+        // extractLen is the size, in UChars, of what will be extracted between index and index+1.
+        //  is zero when both index positions lie within the same code point.
+        int32_t  exLen[] =          {   0,  1,   0,  0,  1,  0,  0,  0,  2,  0,  0};
+
+
+        UErrorCode status = U_ZERO_ERROR;
+        UText *ut = utext_openUTF8(NULL, u8str, -1, &status);
+        TEST_SUCCESS(status);
+
+        // Check setIndex
+        int32_t i;
+        int32_t startMapLimit = sizeof(startMap) / sizeof(int32_t);
+        for (i=0; i<startMapLimit; i++) {
+            utext_setNativeIndex(ut, i);
+            int64_t cpIndex = utext_getNativeIndex(ut);
+            TEST_ASSERT(cpIndex == startMap[i]);
+            cpIndex = UTEXT_GETNATIVEINDEX(ut);
+            TEST_ASSERT(cpIndex == startMap[i]);
+        }
+
+        // Check char32At
+        for (i=0; i<startMapLimit; i++) {
+            UChar32 c32 = utext_char32At(ut, i);
+            TEST_ASSERT(c32 == c32Map[i]);
+            int64_t cpIndex = utext_getNativeIndex(ut);
+            TEST_ASSERT(cpIndex == startMap[i]);
+        }
+
+        // Check utext_next32From
+        for (i=0; i<startMapLimit; i++) {
+            UChar32 c32 = utext_next32From(ut, i);
+            TEST_ASSERT(c32 == c32Map[i]);
+            int64_t cpIndex = utext_getNativeIndex(ut);
+            TEST_ASSERT(cpIndex == nextMap[i]);
+        }
+        
+        // check utext_previous32From
+        for (i=0; i<startMapLimit; i++) {
+            gTestNum++;
+            UChar32 c32 = utext_previous32From(ut, i);
+            TEST_ASSERT(c32 == pr32Map[i]);
+            int64_t cpIndex = utext_getNativeIndex(ut);
+            TEST_ASSERT(cpIndex == prevMap[i]);
+        }
+
+        // check Extract
+        //   Extract from i to i+1, which may be zero or one code points,
+        //     depending on whether the indices straddle a cp boundary.
+        for (i=0; i<startMapLimit; i++) {
+            UChar buf[3];
+            status = U_ZERO_ERROR;
+            int32_t  extractedLen = utext_extract(ut, i, i+1, buf, 3, &status);
+            TEST_SUCCESS(status);
+            TEST_ASSERT(extractedLen == exLen[i]);
+            if (extractedLen > 0) {
+                UChar32  c32;
+                U16_GET(buf, 0, 0, extractedLen, c32);
+                TEST_ASSERT(c32 == c32Map[i]);
+            }
+        }
+
+        utext_close(ut);
+    }
+
+
+    {    //  Similar test, with utf16 instead of utf8
+         //  TODO:  merge the common parts of these tests.
+        
+        UnicodeString u16str("\\u1000\\U00011000\\u2000\\U00022000");
+        int32_t startMap[]  ={ 0,     1,   1,    3,     4,  4,     6,  6};
+        int32_t nextMap[]  = { 1,     3,   3,    4,     6,  6,     6,  6};
+        int32_t prevMap[]  = { 0,     0,   0,    1,     3,  3,     4,  4};
+        UChar32  c32Map[] =  {0x1000, 0x11000, 0x11000, 0x2000,  0x22000, 0x22000, -1, -1}; 
+        UChar32  pr32Map[] = {    -1, 0x1000,  0x1000,  0x11000, 0x2000,  0x2000,   0x22000,   0x22000}; 
+        int32_t  exLen[] =   {   1,  0,   2,  1,  0,  2,  0,  0,};
+
+        u16str = u16str.unescape();
+        UErrorCode status = U_ZERO_ERROR;
+        UText *ut = utext_openUnicodeString(NULL, &u16str, &status);
+        TEST_SUCCESS(status);
+
+        int32_t startMapLimit = sizeof(startMap) / sizeof(int32_t);
+        int i;
+        for (i=0; i<startMapLimit; i++) {
+            utext_setNativeIndex(ut, i);
+            int64_t cpIndex = utext_getNativeIndex(ut);
+            TEST_ASSERT(cpIndex == startMap[i]);
+        }
+
+        // Check char32At
+        for (i=0; i<startMapLimit; i++) {
+            UChar32 c32 = utext_char32At(ut, i);
+            TEST_ASSERT(c32 == c32Map[i]);
+            int64_t cpIndex = utext_getNativeIndex(ut);
+            TEST_ASSERT(cpIndex == startMap[i]);
+        }
+
+        // Check utext_next32From
+        for (i=0; i<startMapLimit; i++) {
+            UChar32 c32 = utext_next32From(ut, i);
+            TEST_ASSERT(c32 == c32Map[i]);
+            int64_t cpIndex = utext_getNativeIndex(ut);
+            TEST_ASSERT(cpIndex == nextMap[i]);
+        }
+        
+        // check utext_previous32From
+        for (i=0; i<startMapLimit; i++) {
+            UChar32 c32 = utext_previous32From(ut, i);
+            TEST_ASSERT(c32 == pr32Map[i]);
+            int64_t cpIndex = utext_getNativeIndex(ut);
+            TEST_ASSERT(cpIndex == prevMap[i]);
+        }
+
+        // check Extract
+        //   Extract from i to i+1, which may be zero or one code points,
+        //     depending on whether the indices straddle a cp boundary.
+        for (i=0; i<startMapLimit; i++) {
+            UChar buf[3];
+            status = U_ZERO_ERROR;
+            int32_t  extractedLen = utext_extract(ut, i, i+1, buf, 3, &status);
+            TEST_SUCCESS(status);
+            TEST_ASSERT(extractedLen == exLen[i]);
+            if (extractedLen > 0) {
+                UChar32  c32;
+                U16_GET(buf, 0, 0, extractedLen, c32);
+                TEST_ASSERT(c32 == c32Map[i]);
+            }
+        }
+
+        utext_close(ut);
+    }
+
+    {    //  Similar test, with UText over Replaceable
+         //  TODO:  merge the common parts of these tests.
+        
+        UnicodeString u16str("\\u1000\\U00011000\\u2000\\U00022000");
+        int32_t startMap[]  ={ 0,     1,   1,    3,     4,  4,     6,  6};
+        int32_t nextMap[]  = { 1,     3,   3,    4,     6,  6,     6,  6};
+        int32_t prevMap[]  = { 0,     0,   0,    1,     3,  3,     4,  4};
+        UChar32  c32Map[] =  {0x1000, 0x11000, 0x11000, 0x2000,  0x22000, 0x22000, -1, -1}; 
+        UChar32  pr32Map[] = {    -1, 0x1000,  0x1000,  0x11000, 0x2000,  0x2000,   0x22000,   0x22000}; 
+        int32_t  exLen[] =   {   1,  0,   2,  1,  0,  2,  0,  0,};
+
+        u16str = u16str.unescape();
+        UErrorCode status = U_ZERO_ERROR;
+        UText *ut = utext_openReplaceable(NULL, &u16str, &status);
+        TEST_SUCCESS(status);
+
+        int32_t startMapLimit = sizeof(startMap) / sizeof(int32_t);
+        int i;
+        for (i=0; i<startMapLimit; i++) {
+            utext_setNativeIndex(ut, i);
+            int64_t cpIndex = utext_getNativeIndex(ut);
+            TEST_ASSERT(cpIndex == startMap[i]);
+        }
+
+        // Check char32At
+        for (i=0; i<startMapLimit; i++) {
+            UChar32 c32 = utext_char32At(ut, i);
+            TEST_ASSERT(c32 == c32Map[i]);
+            int64_t cpIndex = utext_getNativeIndex(ut);
+            TEST_ASSERT(cpIndex == startMap[i]);
+        }
+
+        // Check utext_next32From
+        for (i=0; i<startMapLimit; i++) {
+            UChar32 c32 = utext_next32From(ut, i);
+            TEST_ASSERT(c32 == c32Map[i]);
+            int64_t cpIndex = utext_getNativeIndex(ut);
+            TEST_ASSERT(cpIndex == nextMap[i]);
+        }
+        
+        // check utext_previous32From
+        for (i=0; i<startMapLimit; i++) {
+            UChar32 c32 = utext_previous32From(ut, i);
+            TEST_ASSERT(c32 == pr32Map[i]);
+            int64_t cpIndex = utext_getNativeIndex(ut);
+            TEST_ASSERT(cpIndex == prevMap[i]);
+        }
+
+        // check Extract
+        //   Extract from i to i+1, which may be zero or one code points,
+        //     depending on whether the indices straddle a cp boundary.
+        for (i=0; i<startMapLimit; i++) {
+            UChar buf[3];
+            status = U_ZERO_ERROR;
+            int32_t  extractedLen = utext_extract(ut, i, i+1, buf, 3, &status);
+            TEST_SUCCESS(status);
+            TEST_ASSERT(extractedLen == exLen[i]);
+            if (extractedLen > 0) {
+                UChar32  c32;
+                U16_GET(buf, 0, 0, extractedLen, c32);
+                TEST_ASSERT(c32 == c32Map[i]);
+            }
+        }
+
+        utext_close(ut);
+    }
+}
+
+
+void UTextTest::FreezeTest() {
+    // Check isWritable() and freeze() behavior.
+    //
+
+    UnicodeString  ustr("Hello, World.");
+    const char u8str[] = {char(0x31), (char)0x32, (char)0x33, 0};  
+    const UChar u16str[] = {(UChar)0x31, (UChar)0x32, (UChar)0x44, 0};
+
+    UErrorCode status = U_ZERO_ERROR;
+    UText  *ut        = NULL;
+    UText  *ut2       = NULL;
+
+    ut = utext_openUTF8(ut, u8str, -1, &status);
+    TEST_SUCCESS(status);
+    UBool writable = utext_isWritable(ut);
+    TEST_ASSERT(writable == FALSE);
+    utext_copy(ut, 1, 2, 0, TRUE, &status);
+    TEST_ASSERT(status == U_NO_WRITE_PERMISSION);
+
+    status = U_ZERO_ERROR;
+    ut = utext_openUChars(ut, u16str, -1, &status);
+    TEST_SUCCESS(status);
+    writable = utext_isWritable(ut);
+    TEST_ASSERT(writable == FALSE);
+    utext_copy(ut, 1, 2, 0, TRUE, &status);
+    TEST_ASSERT(status == U_NO_WRITE_PERMISSION);
+
+    status = U_ZERO_ERROR;
+    ut = utext_openUnicodeString(ut, &ustr, &status);
+    TEST_SUCCESS(status);
+    writable = utext_isWritable(ut);
+    TEST_ASSERT(writable == TRUE);
+    utext_freeze(ut);
+    writable = utext_isWritable(ut);
+    TEST_ASSERT(writable == FALSE);
+    utext_copy(ut, 1, 2, 0, TRUE, &status);
+    TEST_ASSERT(status == U_NO_WRITE_PERMISSION);
+    
+    status = U_ZERO_ERROR;
+    ut = utext_openUnicodeString(ut, &ustr, &status);
+    TEST_SUCCESS(status);
+    ut2 = utext_clone(ut2, ut, FALSE, FALSE, &status);  // clone with readonly = false
+    TEST_SUCCESS(status);
+    writable = utext_isWritable(ut2);
+    TEST_ASSERT(writable == TRUE);
+    ut2 = utext_clone(ut2, ut, FALSE, TRUE, &status);  // clone with readonly = true
+    TEST_SUCCESS(status);
+    writable = utext_isWritable(ut2);
+    TEST_ASSERT(writable == FALSE);
+    utext_copy(ut2, 1, 2, 0, TRUE, &status);
+    TEST_ASSERT(status == U_NO_WRITE_PERMISSION);
+
+    status = U_ZERO_ERROR;
+    ut = utext_openConstUnicodeString(ut, (const UnicodeString *)&ustr, &status);
+    TEST_SUCCESS(status);
+    writable = utext_isWritable(ut);
+    TEST_ASSERT(writable == FALSE);
+    utext_copy(ut, 1, 2, 0, TRUE, &status);
+    TEST_ASSERT(status == U_NO_WRITE_PERMISSION);
+
+    // Deep Clone of a frozen UText should re-enable writing in the copy.
+    status = U_ZERO_ERROR;
+    ut = utext_openUnicodeString(ut, &ustr, &status);
+    TEST_SUCCESS(status);
+    utext_freeze(ut);
+    ut2 = utext_clone(ut2, ut, TRUE, FALSE, &status);   // deep clone
+    TEST_SUCCESS(status);
+    writable = utext_isWritable(ut2);
+    TEST_ASSERT(writable == TRUE);
+
+
+    // Deep clone of a frozen UText, where the base type is intrinsically non-writable,
+    //  should NOT enable writing in the copy.
+    status = U_ZERO_ERROR;
+    ut = utext_openUChars(ut, u16str, -1, &status);
+    TEST_SUCCESS(status);
+    utext_freeze(ut);
+    ut2 = utext_clone(ut2, ut, TRUE, FALSE, &status);   // deep clone
+    TEST_SUCCESS(status);
+    writable = utext_isWritable(ut2);
+    TEST_ASSERT(writable == FALSE);
+
+    // cleanup
+    utext_close(ut);
+    utext_close(ut2);
+}
+
+
+//
+//  Fragmented UText
+//      A UText type that works with a chunk size of 1.
+//      Intended to test for edge cases.
+//      Input comes from a UnicodeString.
+//
+//       ut.b    the character.  Put into both halves.
+//
+
+U_CDECL_BEGIN
+static UBool U_CALLCONV
+fragTextAccess(UText *ut, int64_t index, UBool forward) {
+    const UnicodeString *us = (const UnicodeString *)ut->context;
+    UChar  c;
+    int32_t length = us->length();
+    if (forward && index>=0 && index<length) {
+        c = us->charAt((int32_t)index);
+        ut->b = c | c<<16;
+        ut->chunkOffset = 0;
+        ut->chunkLength = 1;
+        ut->chunkNativeStart = index;
+        ut->chunkNativeLimit = index+1;
+        return true;
+    }
+    if (!forward && index>0 && index <=length) {
+        c = us->charAt((int32_t)index-1);
+        ut->b = c | c<<16;
+        ut->chunkOffset = 1;
+        ut->chunkLength = 1;
+        ut->chunkNativeStart = index-1;
+        ut->chunkNativeLimit = index;
+        return true;
+    } 
+    ut->b = 0;
+    ut->chunkOffset = 0;
+    ut->chunkLength = 0;
+    if (index <= 0) {
+        ut->chunkNativeStart = 0;
+        ut->chunkNativeLimit = 0;
+    } else {
+        ut->chunkNativeStart = length;
+        ut->chunkNativeLimit = length;
+    }
+    return false;
+}
+U_CDECL_END
+
+// Function table to be used with this fragmented text provider.
+//   Initialized in the open function.
+UTextFuncs  fragmentFuncs;
+
+// Open function for the fragmented text provider.
+UText *
+openFragmentedUnicodeString(UText *ut, UnicodeString *s, UErrorCode *status) {
+    ut = utext_openUnicodeString(ut, s, status);
+    if (U_FAILURE(*status)) {
+        return ut;
+    }
+
+    // Copy of the function table from the stock UnicodeString UText,
+    //   and replace the entry for the access function.
+    memcpy(&fragmentFuncs, ut->pFuncs, sizeof(fragmentFuncs));
+    fragmentFuncs.access = fragTextAccess;
+    ut->pFuncs = &fragmentFuncs;
+
+    ut->chunkContents = (UChar *)&ut->b;
+    ut->pFuncs->access(ut, 0, TRUE);
+    return ut;
+}
+
+