• Home
  • Line#
  • Scopes#
  • Navigate#
  • Raw
  • Download
1 // © 2016 and later: Unicode, Inc. and others.
2 // License & terms of use: http://www.unicode.org/copyright.html
3 /********************************************************************
4  * COPYRIGHT:
5  * Copyright (c) 2005-2016, International Business Machines Corporation and
6  * others. All Rights Reserved.
7  ********************************************************************/
8 /************************************************************************
9 *   Tests for the UText and UTextIterator text abstraction classses
10 *
11 ************************************************************************/
12 
13 #include <string.h>
14 #include <stdio.h>
15 #include <stdlib.h>
16 #include "unicode/utypes.h"
17 #include "unicode/utext.h"
18 #include "unicode/utf8.h"
19 #include "unicode/utf16.h"
20 #include "unicode/ustring.h"
21 #include "unicode/uchriter.h"
22 #include "cmemory.h"
23 #include "cstr.h"
24 #include "utxttest.h"
25 
26 static UBool  gFailed = FALSE;
27 static int    gTestNum = 0;
28 
29 // Forward decl
30 UText *openFragmentedUnicodeString(UText *ut, UnicodeString *s, UErrorCode *status);
31 
32 #define TEST_ASSERT(x) UPRV_BLOCK_MACRO_BEGIN { \
33     if ((x)==FALSE) { \
34         errln("Test #%d failure in file %s at line %d\n", gTestNum, __FILE__, __LINE__); \
35         gFailed = TRUE; \
36     } \
37 } UPRV_BLOCK_MACRO_END
38 
39 
40 #define TEST_SUCCESS(status) UPRV_BLOCK_MACRO_BEGIN { \
41     if (U_FAILURE(status)) { \
42         errln("Test #%d failure in file %s at line %d. Error = \"%s\"\n", \
43               gTestNum, __FILE__, __LINE__, u_errorName(status)); \
44         gFailed = TRUE; \
45     } \
46 } UPRV_BLOCK_MACRO_END
47 
UTextTest()48 UTextTest::UTextTest() {
49 }
50 
~UTextTest()51 UTextTest::~UTextTest() {
52 }
53 
54 
55 void
runIndexedTest(int32_t index,UBool exec,const char * & name,char *)56 UTextTest::runIndexedTest(int32_t index, UBool exec,
57                           const char* &name, char* /*par*/) {
58     TESTCASE_AUTO_BEGIN;
59     TESTCASE_AUTO(TextTest);
60     TESTCASE_AUTO(ErrorTest);
61     TESTCASE_AUTO(FreezeTest);
62     TESTCASE_AUTO(Ticket5560);
63     TESTCASE_AUTO(Ticket6847);
64     TESTCASE_AUTO(Ticket10562);
65     TESTCASE_AUTO(Ticket10983);
66     TESTCASE_AUTO(Ticket12130);
67     TESTCASE_AUTO(Ticket13344);
68     TESTCASE_AUTO_END;
69 }
70 
71 //
72 // Quick and dirty random number generator.
73 //   (don't use library so that results are portable.
74 static uint32_t m_seed = 1;
m_rand()75 static uint32_t m_rand()
76 {
77     m_seed = m_seed * 1103515245 + 12345;
78     return (uint32_t)(m_seed/65536) % 32768;
79 }
80 
81 
82 //
83 //   TextTest()
84 //
85 //       Top Level function for UText testing.
86 //       Specifies the strings to be tested, with the acutal testing itself
87 //       being carried out in another function, TestString().
88 //
TextTest()89 void  UTextTest::TextTest() {
90     int32_t i, j;
91 
92     TestString("abcd\\U00010001xyz");
93     TestString("");
94 
95     // Supplementary chars at start or end
96     TestString("\\U00010001");
97     TestString("abc\\U00010001");
98     TestString("\\U00010001abc");
99 
100     // Test simple strings of lengths 1 to 60, looking for glitches at buffer boundaries
101     UnicodeString s;
102     for (i=1; i<60; i++) {
103         s.truncate(0);
104         for (j=0; j<i; j++) {
105             if (j+0x30 == 0x5c) {
106                 // backslash.  Needs to be escaped
107                 s.append((UChar)0x5c);
108             }
109             s.append(UChar(j+0x30));
110         }
111         TestString(s);
112     }
113 
114    // Test strings with odd-aligned supplementary chars,
115    //    looking for glitches at buffer boundaries
116     for (i=1; i<60; i++) {
117         s.truncate(0);
118         s.append((UChar)0x41);
119         for (j=0; j<i; j++) {
120             s.append(UChar32(j+0x11000));
121         }
122         TestString(s);
123     }
124 
125     // String of chars of randomly varying size in utf-8 representation.
126     //   Exercise the mapping, and the varying sized buffer.
127     //
128     s.truncate(0);
129     UChar32  c1 = 0;
130     UChar32  c2 = 0x100;
131     UChar32  c3 = 0xa000;
132     UChar32  c4 = 0x11000;
133     for (i=0; i<1000; i++) {
134         int len8 = m_rand()%4 + 1;
135         switch (len8) {
136             case 1:
137                 c1 = (c1+1)%0x80;
138                 // don't put 0 into string (0 terminated strings for some tests)
139                 // don't put '\', will cause unescape() to fail.
140                 if (c1==0x5c || c1==0) {
141                     c1++;
142                 }
143                 s.append(c1);
144                 break;
145             case 2:
146                 s.append(c2++);
147                 break;
148             case 3:
149                 s.append(c3++);
150                 break;
151             case 4:
152                 s.append(c4++);
153                 break;
154         }
155     }
156     TestString(s);
157 }
158 
159 
160 //
161 //  TestString()     Run a suite of UText tests on a string.
162 //                   The test string is unescaped before use.
163 //
TestString(const UnicodeString & s)164 void UTextTest::TestString(const UnicodeString &s) {
165     int32_t       i;
166     int32_t       j;
167     UChar32       c;
168     int32_t       cpCount = 0;
169     UErrorCode    status  = U_ZERO_ERROR;
170     UText        *ut      = NULL;
171     int32_t       saLen;
172 
173     UnicodeString sa = s.unescape();
174     saLen = sa.length();
175 
176     //
177     // Build up a mapping between code points and UTF-16 code unit indexes.
178     //
179     m *cpMap = new m[sa.length() + 1];
180     j = 0;
181     for (i=0; i<sa.length(); i=sa.moveIndex32(i, 1)) {
182         c = sa.char32At(i);
183         cpMap[j].nativeIdx = i;
184         cpMap[j].cp = c;
185         j++;
186         cpCount++;
187     }
188     cpMap[j].nativeIdx = i;   // position following the last char in utf-16 string.
189 
190 
191     // UChar * test, null terminated
192     status = U_ZERO_ERROR;
193     UChar *buf = new UChar[saLen+1];
194     sa.extract(buf, saLen+1, status);
195     TEST_SUCCESS(status);
196     ut = utext_openUChars(NULL, buf, -1, &status);
197     TEST_SUCCESS(status);
198     TestAccess(sa, ut, cpCount, cpMap);
199     utext_close(ut);
200     delete [] buf;
201 
202     // UChar * test, with length
203     status = U_ZERO_ERROR;
204     buf = new UChar[saLen+1];
205     sa.extract(buf, saLen+1, status);
206     TEST_SUCCESS(status);
207     ut = utext_openUChars(NULL, buf, saLen, &status);
208     TEST_SUCCESS(status);
209     TestAccess(sa, ut, cpCount, cpMap);
210     utext_close(ut);
211     delete [] buf;
212 
213 
214     // UnicodeString test
215     status = U_ZERO_ERROR;
216     ut = utext_openUnicodeString(NULL, &sa, &status);
217     TEST_SUCCESS(status);
218     TestAccess(sa, ut, cpCount, cpMap);
219     TestCMR(sa, ut, cpCount, cpMap, cpMap);
220     utext_close(ut);
221 
222 
223     // Const UnicodeString test
224     status = U_ZERO_ERROR;
225     ut = utext_openConstUnicodeString(NULL, &sa, &status);
226     TEST_SUCCESS(status);
227     TestAccess(sa, ut, cpCount, cpMap);
228     utext_close(ut);
229 
230 
231     // Replaceable test.  (UnicodeString inherits Replaceable)
232     status = U_ZERO_ERROR;
233     ut = utext_openReplaceable(NULL, &sa, &status);
234     TEST_SUCCESS(status);
235     TestAccess(sa, ut, cpCount, cpMap);
236     TestCMR(sa, ut, cpCount, cpMap, cpMap);
237     utext_close(ut);
238 
239     // Character Iterator Tests
240     status = U_ZERO_ERROR;
241     const UChar *cbuf = sa.getBuffer();
242     CharacterIterator *ci = new UCharCharacterIterator(cbuf, saLen, status);
243     TEST_SUCCESS(status);
244     ut = utext_openCharacterIterator(NULL, ci, &status);
245     TEST_SUCCESS(status);
246     TestAccess(sa, ut, cpCount, cpMap);
247     utext_close(ut);
248     delete ci;
249 
250 
251     // Fragmented UnicodeString  (Chunk size of one)
252     //
253     status = U_ZERO_ERROR;
254     ut = openFragmentedUnicodeString(NULL, &sa, &status);
255     TEST_SUCCESS(status);
256     TestAccess(sa, ut, cpCount, cpMap);
257     utext_close(ut);
258 
259     //
260     // UTF-8 test
261     //
262 
263     // Convert the test string from UnicodeString to (char *) in utf-8 format
264     int32_t u8Len = sa.extract(0, sa.length(), NULL, 0, "utf-8");
265     char *u8String = new char[u8Len + 1];
266     sa.extract(0, sa.length(), u8String, u8Len+1, "utf-8");
267 
268     // Build up the map of code point indices in the utf-8 string
269     m * u8Map = new m[sa.length() + 1];
270     i = 0;   // native utf-8 index
271     for (j=0; j<cpCount ; j++) {  // code point number
272         u8Map[j].nativeIdx = i;
273         U8_NEXT(u8String, i, u8Len, c);
274         u8Map[j].cp = c;
275     }
276     u8Map[cpCount].nativeIdx = u8Len;   // position following the last char in utf-8 string.
277 
278     // Do the test itself
279     status = U_ZERO_ERROR;
280     ut = utext_openUTF8(NULL, u8String, -1, &status);
281     TEST_SUCCESS(status);
282     TestAccess(sa, ut, cpCount, u8Map);
283     utext_close(ut);
284 
285 
286 
287     delete []cpMap;
288     delete []u8Map;
289     delete []u8String;
290 }
291 
292 //  TestCMR   test Copy, Move and Replace operations.
293 //              us         UnicodeString containing the test text.
294 //              ut         UText containing the same test text.
295 //              cpCount    number of code points in the test text.
296 //              nativeMap  Mapping from code points to native indexes for the UText.
297 //              u16Map     Mapping from code points to UTF-16 indexes, for use with the UnicodeString.
298 //
299 //     This function runs a whole series of opertions on each incoming UText.
300 //     The UText is deep-cloned prior to each operation, so that the original UText remains unchanged.
301 //
TestCMR(const UnicodeString & us,UText * ut,int cpCount,m * nativeMap,m * u16Map)302 void UTextTest::TestCMR(const UnicodeString &us, UText *ut, int cpCount, m *nativeMap, m *u16Map) {
303     TEST_ASSERT(utext_isWritable(ut) == TRUE);
304 
305     int  srcLengthType;       // Loop variables for selecting the postion and length
306     int  srcPosType;          //   of the block to operate on within the source text.
307     int  destPosType;
308 
309     int  srcIndex  = 0;       // Code Point indexes of the block to operate on for
310     int  srcLength = 0;       //   a specific test.
311 
312     int  destIndex = 0;       // Code point index of the destination for a copy/move test.
313 
314     int32_t  nativeStart = 0; // Native unit indexes for a test.
315     int32_t  nativeLimit = 0;
316     int32_t  nativeDest  = 0;
317 
318     int32_t  u16Start    = 0; // UTF-16 indexes for a test.
319     int32_t  u16Limit    = 0; //   used when performing the same operation in a Unicode String
320     int32_t  u16Dest     = 0;
321 
322     // Iterate over a whole series of source index, length and a target indexes.
323     // This is done with code point indexes; these will be later translated to native
324     //   indexes using the cpMap.
325     for (srcLengthType=1; srcLengthType<=3; srcLengthType++) {
326         switch (srcLengthType) {
327             case 1: srcLength = 1; break;
328             case 2: srcLength = 5; break;
329             case 3: srcLength = cpCount / 3;
330         }
331         for (srcPosType=1; srcPosType<=5; srcPosType++) {
332             switch (srcPosType) {
333                 case 1: srcIndex = 0; break;
334                 case 2: srcIndex = 1; break;
335                 case 3: srcIndex = cpCount - srcLength; break;
336                 case 4: srcIndex = cpCount - srcLength - 1; break;
337                 case 5: srcIndex = cpCount / 2; break;
338             }
339             if (srcIndex < 0 || srcIndex + srcLength > cpCount) {
340                 // filter out bogus test cases -
341                 //   those with a source range that falls of an edge of the string.
342                 continue;
343             }
344 
345             //
346             // Copy and move tests.
347             //   iterate over a variety of destination positions.
348             //
349             for (destPosType=1; destPosType<=4; destPosType++) {
350                 switch (destPosType) {
351                     case 1: destIndex = 0; break;
352                     case 2: destIndex = 1; break;
353                     case 3: destIndex = srcIndex - 1; break;
354                     case 4: destIndex = srcIndex + srcLength + 1; break;
355                     case 5: destIndex = cpCount-1; break;
356                     case 6: destIndex = cpCount; break;
357                 }
358                 if (destIndex<0 || destIndex>cpCount) {
359                     // filter out bogus test cases.
360                     continue;
361                 }
362 
363                 nativeStart = nativeMap[srcIndex].nativeIdx;
364                 nativeLimit = nativeMap[srcIndex+srcLength].nativeIdx;
365                 nativeDest  = nativeMap[destIndex].nativeIdx;
366 
367                 u16Start    = u16Map[srcIndex].nativeIdx;
368                 u16Limit    = u16Map[srcIndex+srcLength].nativeIdx;
369                 u16Dest     = u16Map[destIndex].nativeIdx;
370 
371                 gFailed = FALSE;
372                 TestCopyMove(us, ut, FALSE,
373                     nativeStart, nativeLimit, nativeDest,
374                     u16Start, u16Limit, u16Dest);
375 
376                 TestCopyMove(us, ut, TRUE,
377                     nativeStart, nativeLimit, nativeDest,
378                     u16Start, u16Limit, u16Dest);
379 
380                 if (gFailed) {
381                     return;
382                 }
383             }
384 
385             //
386             //  Replace tests.
387             //
388             UnicodeString fullRepString("This is an arbitrary string that will be used as replacement text");
389             for (int32_t replStrLen=0; replStrLen<20; replStrLen++) {
390                 UnicodeString repStr(fullRepString, 0, replStrLen);
391                 TestReplace(us, ut,
392                     nativeStart, nativeLimit,
393                     u16Start, u16Limit,
394                     repStr);
395                 if (gFailed) {
396                     return;
397                 }
398             }
399 
400         }
401     }
402 
403 }
404 
405 //
406 //   TestCopyMove    run a single test case for utext_copy.
407 //                   Test cases are created in TestCMR and dispatched here for execution.
408 //
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)409 void UTextTest::TestCopyMove(const UnicodeString &us, UText *ut, UBool move,
410                     int32_t nativeStart, int32_t nativeLimit, int32_t nativeDest,
411                     int32_t u16Start, int32_t u16Limit, int32_t u16Dest)
412 {
413     UErrorCode      status   = U_ZERO_ERROR;
414     UText          *targetUT = NULL;
415     gTestNum++;
416     gFailed = FALSE;
417 
418     //
419     //  clone the UText.  The test will be run in the cloned copy
420     //  so that we don't alter the original.
421     //
422     targetUT = utext_clone(NULL, ut, TRUE, FALSE, &status);
423     TEST_SUCCESS(status);
424     UnicodeString targetUS(us);    // And copy the reference string.
425 
426     // do the test operation first in the reference
427     targetUS.copy(u16Start, u16Limit, u16Dest);
428     if (move) {
429         // delete out the source range.
430         if (u16Limit < u16Dest) {
431             targetUS.removeBetween(u16Start, u16Limit);
432         } else {
433             int32_t amtCopied = u16Limit - u16Start;
434             targetUS.removeBetween(u16Start+amtCopied, u16Limit+amtCopied);
435         }
436     }
437 
438     // Do the same operation in the UText under test
439     utext_copy(targetUT, nativeStart, nativeLimit, nativeDest, move, &status);
440     if (nativeDest > nativeStart && nativeDest < nativeLimit) {
441         TEST_ASSERT(status == U_INDEX_OUTOFBOUNDS_ERROR);
442     } else {
443         TEST_SUCCESS(status);
444 
445         // Compare the results of the two parallel tests
446         int32_t  usi = 0;    // UnicodeString postion, utf-16 index.
447         int64_t  uti = 0;    // UText position, native index.
448         int32_t  cpi;        // char32 position (code point index)
449         UChar32  usc;        // code point from Unicode String
450         UChar32  utc;        // code point from UText
451         utext_setNativeIndex(targetUT, 0);
452         for (cpi=0; ; cpi++) {
453             usc = targetUS.char32At(usi);
454             utc = utext_next32(targetUT);
455             if (utc < 0) {
456                 break;
457             }
458             TEST_ASSERT(uti == usi);
459             TEST_ASSERT(utc == usc);
460             usi = targetUS.moveIndex32(usi, 1);
461             uti = utext_getNativeIndex(targetUT);
462             if (gFailed) {
463                 goto cleanupAndReturn;
464             }
465         }
466         int64_t expectedNativeLength = utext_nativeLength(ut);
467         if (move == FALSE) {
468             expectedNativeLength += nativeLimit - nativeStart;
469         }
470         uti = utext_getNativeIndex(targetUT);
471         TEST_ASSERT(uti == expectedNativeLength);
472     }
473 
474 cleanupAndReturn:
475     utext_close(targetUT);
476 }
477 
478 
479 //
480 //  TestReplace   Test a single Replace operation.
481 //
TestReplace(const UnicodeString & us,UText * ut,int32_t nativeStart,int32_t nativeLimit,int32_t u16Start,int32_t u16Limit,const UnicodeString & repStr)482 void UTextTest::TestReplace(
483             const UnicodeString &us,     // reference UnicodeString in which to do the replace
484             UText         *ut,                // UnicodeText object under test.
485             int32_t       nativeStart,        // Range to be replaced, in UText native units.
486             int32_t       nativeLimit,
487             int32_t       u16Start,           // Range to be replaced, in UTF-16 units
488             int32_t       u16Limit,           //    for use in the reference UnicodeString.
489             const UnicodeString &repStr)      // The replacement string
490 {
491     UErrorCode      status   = U_ZERO_ERROR;
492     UText          *targetUT = NULL;
493     gTestNum++;
494     gFailed = FALSE;
495 
496     //
497     //  clone the target UText.  The test will be run in the cloned copy
498     //  so that we don't alter the original.
499     //
500     targetUT = utext_clone(NULL, ut, TRUE, FALSE, &status);
501     TEST_SUCCESS(status);
502     UnicodeString targetUS(us);    // And copy the reference string.
503 
504     //
505     // Do the replace operation in the Unicode String, to
506     //   produce a reference result.
507     //
508     targetUS.replace(u16Start, u16Limit-u16Start, repStr);
509 
510     //
511     // Do the replace on the UText under test
512     //
513     const UChar *rs = repStr.getBuffer();
514     int32_t  rsLen = repStr.length();
515     int32_t actualDelta = utext_replace(targetUT, nativeStart, nativeLimit, rs, rsLen, &status);
516     int32_t expectedDelta = repStr.length() - (nativeLimit - nativeStart);
517     TEST_ASSERT(actualDelta == expectedDelta);
518 
519     //
520     // Compare the results
521     //
522     int32_t  usi = 0;    // UnicodeString postion, utf-16 index.
523     int64_t  uti = 0;    // UText position, native index.
524     int32_t  cpi;        // char32 position (code point index)
525     UChar32  usc;        // code point from Unicode String
526     UChar32  utc;        // code point from UText
527     int64_t  expectedNativeLength = 0;
528     utext_setNativeIndex(targetUT, 0);
529     for (cpi=0; ; cpi++) {
530         usc = targetUS.char32At(usi);
531         utc = utext_next32(targetUT);
532         if (utc < 0) {
533             break;
534         }
535         TEST_ASSERT(uti == usi);
536         TEST_ASSERT(utc == usc);
537         usi = targetUS.moveIndex32(usi, 1);
538         uti = utext_getNativeIndex(targetUT);
539         if (gFailed) {
540             goto cleanupAndReturn;
541         }
542     }
543     expectedNativeLength = utext_nativeLength(ut) + expectedDelta;
544     uti = utext_getNativeIndex(targetUT);
545     TEST_ASSERT(uti == expectedNativeLength);
546 
547 cleanupAndReturn:
548     utext_close(targetUT);
549 }
550 
551 //
552 //  TestAccess      Test the read only access functions on a UText, including cloning.
553 //                  The text is accessed in a variety of ways, and compared with
554 //                  the reference UnicodeString.
555 //
TestAccess(const UnicodeString & us,UText * ut,int cpCount,m * cpMap)556 void UTextTest::TestAccess(const UnicodeString &us, UText *ut, int cpCount, m *cpMap) {
557     // Run the standard tests on the caller-supplied UText.
558     TestAccessNoClone(us, ut, cpCount, cpMap);
559 
560     // Re-run tests on a shallow clone.
561     utext_setNativeIndex(ut, 0);
562     UErrorCode status = U_ZERO_ERROR;
563     UText *shallowClone = utext_clone(NULL, ut, FALSE /*deep*/, FALSE /*readOnly*/, &status);
564     TEST_SUCCESS(status);
565     TestAccessNoClone(us, shallowClone, cpCount, cpMap);
566 
567     //
568     // Rerun again on a deep clone.
569     // Note that text providers are not required to provide deep cloning,
570     //   so unsupported errors are ignored.
571     //
572     status = U_ZERO_ERROR;
573     utext_setNativeIndex(shallowClone, 0);
574     UText *deepClone = utext_clone(NULL, shallowClone, TRUE, FALSE, &status);
575     utext_close(shallowClone);
576     if (status != U_UNSUPPORTED_ERROR) {
577         TEST_SUCCESS(status);
578         TestAccessNoClone(us, deepClone, cpCount, cpMap);
579     }
580     utext_close(deepClone);
581 }
582 
583 
584 //
585 //  TestAccessNoClone()    Test the read only access functions on a UText.
586 //                         The text is accessed in a variety of ways, and compared with
587 //                         the reference UnicodeString.
588 //
TestAccessNoClone(const UnicodeString & us,UText * ut,int cpCount,m * cpMap)589 void UTextTest::TestAccessNoClone(const UnicodeString &us, UText *ut, int cpCount, m *cpMap) {
590     UErrorCode  status = U_ZERO_ERROR;
591     gTestNum++;
592 
593     //
594     //  Check the length from the UText
595     //
596     int64_t expectedLen = cpMap[cpCount].nativeIdx;
597     int64_t utlen = utext_nativeLength(ut);
598     TEST_ASSERT(expectedLen == utlen);
599 
600     //
601     //  Iterate forwards, verify that we get the correct code points
602     //   at the correct native offsets.
603     //
604     int         i = 0;
605     int64_t     index;
606     int64_t     expectedIndex = 0;
607     int64_t     foundIndex = 0;
608     UChar32     expectedC;
609     UChar32     foundC;
610     int64_t     len;
611 
612     for (i=0; i<cpCount; i++) {
613         expectedIndex = cpMap[i].nativeIdx;
614         foundIndex    = utext_getNativeIndex(ut);
615         TEST_ASSERT(expectedIndex == foundIndex);
616         expectedC     = cpMap[i].cp;
617         foundC        = utext_next32(ut);
618         TEST_ASSERT(expectedC == foundC);
619         foundIndex    = utext_getPreviousNativeIndex(ut);
620         TEST_ASSERT(expectedIndex == foundIndex);
621         if (gFailed) {
622             return;
623         }
624     }
625     foundC = utext_next32(ut);
626     TEST_ASSERT(foundC == U_SENTINEL);
627 
628     // Repeat above, using macros
629     utext_setNativeIndex(ut, 0);
630     for (i=0; i<cpCount; i++) {
631         expectedIndex = cpMap[i].nativeIdx;
632         foundIndex    = UTEXT_GETNATIVEINDEX(ut);
633         TEST_ASSERT(expectedIndex == foundIndex);
634         expectedC     = cpMap[i].cp;
635         foundC        = UTEXT_NEXT32(ut);
636         TEST_ASSERT(expectedC == foundC);
637         if (gFailed) {
638             return;
639         }
640     }
641     foundC = UTEXT_NEXT32(ut);
642     TEST_ASSERT(foundC == U_SENTINEL);
643 
644     //
645     //  Forward iteration (above) should have left index at the
646     //   end of the input, which should == length().
647     //
648     len = utext_nativeLength(ut);
649     foundIndex  = utext_getNativeIndex(ut);
650     TEST_ASSERT(len == foundIndex);
651 
652     //
653     // Iterate backwards over entire test string
654     //
655     len = utext_getNativeIndex(ut);
656     utext_setNativeIndex(ut, len);
657     for (i=cpCount-1; i>=0; i--) {
658         expectedC     = cpMap[i].cp;
659         expectedIndex = cpMap[i].nativeIdx;
660         int64_t prevIndex = utext_getPreviousNativeIndex(ut);
661         foundC        = utext_previous32(ut);
662         foundIndex    = utext_getNativeIndex(ut);
663         TEST_ASSERT(expectedIndex == foundIndex);
664         TEST_ASSERT(expectedC == foundC);
665         TEST_ASSERT(prevIndex == foundIndex);
666         if (gFailed) {
667             return;
668         }
669     }
670 
671     //
672     //  Backwards iteration, above, should have left our iterator
673     //   position at zero, and continued backwards iterationshould fail.
674     //
675     foundIndex = utext_getNativeIndex(ut);
676     TEST_ASSERT(foundIndex == 0);
677     foundIndex = utext_getPreviousNativeIndex(ut);
678     TEST_ASSERT(foundIndex == 0);
679 
680 
681     foundC = utext_previous32(ut);
682     TEST_ASSERT(foundC == U_SENTINEL);
683     foundIndex = utext_getNativeIndex(ut);
684     TEST_ASSERT(foundIndex == 0);
685     foundIndex = utext_getPreviousNativeIndex(ut);
686     TEST_ASSERT(foundIndex == 0);
687 
688 
689     // And again, with the macros
690     utext_setNativeIndex(ut, len);
691     for (i=cpCount-1; i>=0; i--) {
692         expectedC     = cpMap[i].cp;
693         expectedIndex = cpMap[i].nativeIdx;
694         foundC        = UTEXT_PREVIOUS32(ut);
695         foundIndex    = UTEXT_GETNATIVEINDEX(ut);
696         TEST_ASSERT(expectedIndex == foundIndex);
697         TEST_ASSERT(expectedC == foundC);
698         if (gFailed) {
699             return;
700         }
701     }
702 
703     //
704     //  Backwards iteration, above, should have left our iterator
705     //   position at zero, and continued backwards iterationshould fail.
706     //
707     foundIndex = UTEXT_GETNATIVEINDEX(ut);
708     TEST_ASSERT(foundIndex == 0);
709 
710     foundC = UTEXT_PREVIOUS32(ut);
711     TEST_ASSERT(foundC == U_SENTINEL);
712     foundIndex = UTEXT_GETNATIVEINDEX(ut);
713     TEST_ASSERT(foundIndex == 0);
714     if (gFailed) {
715         return;
716     }
717 
718     //
719     //  next32From(), prevous32From(), Iterate in a somewhat random order.
720     //
721     int  cpIndex = 0;
722     for (i=0; i<cpCount; i++) {
723         cpIndex = (cpIndex + 9973) % cpCount;
724         index         = cpMap[cpIndex].nativeIdx;
725         expectedC     = cpMap[cpIndex].cp;
726         foundC        = utext_next32From(ut, index);
727         TEST_ASSERT(expectedC == foundC);
728         if (gFailed) {
729             return;
730         }
731     }
732 
733     cpIndex = 0;
734     for (i=0; i<cpCount; i++) {
735         cpIndex = (cpIndex + 9973) % cpCount;
736         index         = cpMap[cpIndex+1].nativeIdx;
737         expectedC     = cpMap[cpIndex].cp;
738         foundC        = utext_previous32From(ut, index);
739         TEST_ASSERT(expectedC == foundC);
740         if (gFailed) {
741             return;
742         }
743     }
744 
745 
746     //
747     // moveIndex(int32_t delta);
748     //
749 
750     // Walk through frontwards, incrementing by one
751     utext_setNativeIndex(ut, 0);
752     for (i=1; i<=cpCount; i++) {
753         utext_moveIndex32(ut, 1);
754         index = utext_getNativeIndex(ut);
755         expectedIndex = cpMap[i].nativeIdx;
756         TEST_ASSERT(expectedIndex == index);
757         index = UTEXT_GETNATIVEINDEX(ut);
758         TEST_ASSERT(expectedIndex == index);
759     }
760 
761     // Walk through frontwards, incrementing by two
762     utext_setNativeIndex(ut, 0);
763     for (i=2; i<cpCount; i+=2) {
764         utext_moveIndex32(ut, 2);
765         index = utext_getNativeIndex(ut);
766         expectedIndex = cpMap[i].nativeIdx;
767         TEST_ASSERT(expectedIndex == index);
768         index = UTEXT_GETNATIVEINDEX(ut);
769         TEST_ASSERT(expectedIndex == index);
770     }
771 
772     // walk through the string backwards, decrementing by one.
773     i = cpMap[cpCount].nativeIdx;
774     utext_setNativeIndex(ut, i);
775     for (i=cpCount; i>=0; i--) {
776         expectedIndex = cpMap[i].nativeIdx;
777         index = utext_getNativeIndex(ut);
778         TEST_ASSERT(expectedIndex == index);
779         index = UTEXT_GETNATIVEINDEX(ut);
780         TEST_ASSERT(expectedIndex == index);
781         utext_moveIndex32(ut, -1);
782     }
783 
784 
785     // walk through backwards, decrementing by three
786     i = cpMap[cpCount].nativeIdx;
787     utext_setNativeIndex(ut, i);
788     for (i=cpCount; i>=0; i-=3) {
789         expectedIndex = cpMap[i].nativeIdx;
790         index = utext_getNativeIndex(ut);
791         TEST_ASSERT(expectedIndex == index);
792         index = UTEXT_GETNATIVEINDEX(ut);
793         TEST_ASSERT(expectedIndex == index);
794         utext_moveIndex32(ut, -3);
795     }
796 
797 
798     //
799     // Extract
800     //
801     int bufSize = us.length() + 10;
802     UChar *buf = new UChar[bufSize];
803     status = U_ZERO_ERROR;
804     expectedLen = us.length();
805     len = utext_extract(ut, 0, utlen, buf, bufSize, &status);
806     TEST_SUCCESS(status);
807     TEST_ASSERT(len == expectedLen);
808     int compareResult = us.compare(buf, -1);
809     TEST_ASSERT(compareResult == 0);
810 
811     status = U_ZERO_ERROR;
812     len = utext_extract(ut, 0, utlen, NULL, 0, &status);
813     if (utlen == 0) {
814         TEST_ASSERT(status == U_STRING_NOT_TERMINATED_WARNING);
815     } else {
816         TEST_ASSERT(status == U_BUFFER_OVERFLOW_ERROR);
817     }
818     TEST_ASSERT(len == expectedLen);
819 
820     status = U_ZERO_ERROR;
821     u_memset(buf, 0x5555, bufSize);
822     len = utext_extract(ut, 0, utlen, buf, 1, &status);
823     if (us.length() == 0) {
824         TEST_SUCCESS(status);
825         TEST_ASSERT(buf[0] == 0);
826     } else {
827         // Buf len == 1, extracting a single 16 bit value.
828         // If the data char is supplementary, it doesn't matter whether the buffer remains unchanged,
829         //   or whether the lead surrogate of the pair is extracted.
830         //   It's a buffer overflow error in either case.
831         TEST_ASSERT(buf[0] == us.charAt(0) ||
832                     (buf[0] == 0x5555 && U_IS_SUPPLEMENTARY(us.char32At(0))));
833         TEST_ASSERT(buf[1] == 0x5555);
834         if (us.length() == 1) {
835             TEST_ASSERT(status == U_STRING_NOT_TERMINATED_WARNING);
836         } else {
837             TEST_ASSERT(status == U_BUFFER_OVERFLOW_ERROR);
838         }
839     }
840 
841     delete []buf;
842 }
843 
844 //
845 //  ErrorTest()    Check various error and edge cases.
846 //
ErrorTest()847 void UTextTest::ErrorTest()
848 {
849     // Close of an unitialized UText.  Shouldn't blow up.
850     {
851         UText  ut;
852         memset(&ut, 0, sizeof(UText));
853         utext_close(&ut);
854         utext_close(NULL);
855     }
856 
857     // Double-close of a UText.  Shouldn't blow up.  UText should still be usable.
858     {
859         UErrorCode status = U_ZERO_ERROR;
860         UText ut = UTEXT_INITIALIZER;
861         UnicodeString s("Hello, World");
862         UText *ut2 = utext_openUnicodeString(&ut, &s, &status);
863         TEST_SUCCESS(status);
864         TEST_ASSERT(ut2 == &ut);
865 
866         UText *ut3 = utext_close(&ut);
867         TEST_ASSERT(ut3 == &ut);
868 
869         UText *ut4 = utext_close(&ut);
870         TEST_ASSERT(ut4 == &ut);
871 
872         utext_openUnicodeString(&ut, &s, &status);
873         TEST_SUCCESS(status);
874         utext_close(&ut);
875     }
876 
877     // Re-use of a UText, chaining through each of the types of UText
878     //   (If it doesn't blow up, and doesn't leak, it's probably working fine)
879     {
880         UErrorCode status = U_ZERO_ERROR;
881         UText ut = UTEXT_INITIALIZER;
882         UText  *utp;
883         UnicodeString s1("Hello, World");
884         UChar s2[] = {(UChar)0x41, (UChar)0x42, (UChar)0};
885         const char  *s3 = "\x66\x67\x68";
886 
887         utp = utext_openUnicodeString(&ut, &s1, &status);
888         TEST_SUCCESS(status);
889         TEST_ASSERT(utp == &ut);
890 
891         utp = utext_openConstUnicodeString(&ut, &s1, &status);
892         TEST_SUCCESS(status);
893         TEST_ASSERT(utp == &ut);
894 
895         utp = utext_openUTF8(&ut, s3, -1, &status);
896         TEST_SUCCESS(status);
897         TEST_ASSERT(utp == &ut);
898 
899         utp = utext_openUChars(&ut, s2, -1, &status);
900         TEST_SUCCESS(status);
901         TEST_ASSERT(utp == &ut);
902 
903         utp = utext_close(&ut);
904         TEST_ASSERT(utp == &ut);
905 
906         utp = utext_openUnicodeString(&ut, &s1, &status);
907         TEST_SUCCESS(status);
908         TEST_ASSERT(utp == &ut);
909     }
910 
911     // Invalid parameters on open
912     //
913     {
914         UErrorCode status = U_ZERO_ERROR;
915         UText ut = UTEXT_INITIALIZER;
916 
917         utext_openUChars(&ut, NULL, 5, &status);
918         TEST_ASSERT(status == U_ILLEGAL_ARGUMENT_ERROR);
919 
920         status = U_ZERO_ERROR;
921         utext_openUChars(&ut, NULL, -1, &status);
922         TEST_ASSERT(status == U_ILLEGAL_ARGUMENT_ERROR);
923 
924         status = U_ZERO_ERROR;
925         utext_openUTF8(&ut, NULL, 4, &status);
926         TEST_ASSERT(status == U_ILLEGAL_ARGUMENT_ERROR);
927 
928         status = U_ZERO_ERROR;
929         utext_openUTF8(&ut, NULL, -1, &status);
930         TEST_ASSERT(status == U_ILLEGAL_ARGUMENT_ERROR);
931     }
932 
933     //
934     //  UTF-8 with malformed sequences.
935     //    These should come through as the Unicode replacement char, \ufffd
936     //
937     {
938         UErrorCode status = U_ZERO_ERROR;
939         UText *ut = NULL;
940         const char *badUTF8 = "\x41\x81\x42\xf0\x81\x81\x43";
941         UChar32  c;
942 
943         ut = utext_openUTF8(NULL, badUTF8, -1, &status);
944         TEST_SUCCESS(status);
945         c = utext_char32At(ut, 1);
946         TEST_ASSERT(c == 0xfffd);
947         c = utext_char32At(ut, 3);
948         TEST_ASSERT(c == 0xfffd);
949         c = utext_char32At(ut, 5);
950         TEST_ASSERT(c == 0xfffd);
951         c = utext_char32At(ut, 6);
952         TEST_ASSERT(c == 0x43);
953 
954         UChar buf[10];
955         int n = utext_extract(ut, 0, 9, buf, 10, &status);
956         TEST_SUCCESS(status);
957         TEST_ASSERT(n==7);
958         TEST_ASSERT(buf[0] == 0x41);
959         TEST_ASSERT(buf[1] == 0xfffd);
960         TEST_ASSERT(buf[2] == 0x42);
961         TEST_ASSERT(buf[3] == 0xfffd);
962         TEST_ASSERT(buf[4] == 0xfffd);
963         TEST_ASSERT(buf[5] == 0xfffd);
964         TEST_ASSERT(buf[6] == 0x43);
965         utext_close(ut);
966     }
967 
968 
969     //
970     //  isLengthExpensive - does it make the exptected transitions after
971     //                      getting the length of a nul terminated string?
972     //
973     {
974         UErrorCode status = U_ZERO_ERROR;
975         UnicodeString sa("Hello, this is a string");
976         UBool  isExpensive;
977 
978         UChar sb[100];
979         memset(sb, 0x20, sizeof(sb));
980         sb[99] = 0;
981 
982         UText *uta = utext_openUnicodeString(NULL, &sa, &status);
983         TEST_SUCCESS(status);
984         isExpensive = utext_isLengthExpensive(uta);
985         TEST_ASSERT(isExpensive == FALSE);
986         utext_close(uta);
987 
988         UText *utb = utext_openUChars(NULL, sb, -1, &status);
989         TEST_SUCCESS(status);
990         isExpensive = utext_isLengthExpensive(utb);
991         TEST_ASSERT(isExpensive == TRUE);
992         int64_t  len = utext_nativeLength(utb);
993         TEST_ASSERT(len == 99);
994         isExpensive = utext_isLengthExpensive(utb);
995         TEST_ASSERT(isExpensive == FALSE);
996         utext_close(utb);
997     }
998 
999     //
1000     // Index to positions not on code point boundaries.
1001     //
1002     {
1003         const char *u8str =         "\xc8\x81\xe1\x82\x83\xf1\x84\x85\x86";
1004         int32_t startMap[] =        {   0,  0,  2,  2,  2,  5,  5,  5,  5,  9,  9};
1005         int32_t nextMap[]  =        {   2,  2,  5,  5,  5,  9,  9,  9,  9,  9,  9};
1006         int32_t prevMap[]  =        {   0,  0,  0,  0,  0,  2,  2,  2,  2,  5,  5};
1007         UChar32  c32Map[] =    {0x201, 0x201, 0x1083, 0x1083, 0x1083, 0x044146, 0x044146, 0x044146, 0x044146, -1, -1};
1008         UChar32  pr32Map[] =   {    -1,   -1,  0x201,  0x201,  0x201,   0x1083,   0x1083,   0x1083,   0x1083, 0x044146, 0x044146};
1009 
1010         // extractLen is the size, in UChars, of what will be extracted between index and index+1.
1011         //  is zero when both index positions lie within the same code point.
1012         int32_t  exLen[] =          {   0,  1,   0,  0,  1,  0,  0,  0,  2,  0,  0};
1013 
1014 
1015         UErrorCode status = U_ZERO_ERROR;
1016         UText *ut = utext_openUTF8(NULL, u8str, -1, &status);
1017         TEST_SUCCESS(status);
1018 
1019         // Check setIndex
1020         int32_t i;
1021         int32_t startMapLimit = UPRV_LENGTHOF(startMap);
1022         for (i=0; i<startMapLimit; i++) {
1023             utext_setNativeIndex(ut, i);
1024             int64_t cpIndex = utext_getNativeIndex(ut);
1025             TEST_ASSERT(cpIndex == startMap[i]);
1026             cpIndex = UTEXT_GETNATIVEINDEX(ut);
1027             TEST_ASSERT(cpIndex == startMap[i]);
1028         }
1029 
1030         // Check char32At
1031         for (i=0; i<startMapLimit; i++) {
1032             UChar32 c32 = utext_char32At(ut, i);
1033             TEST_ASSERT(c32 == c32Map[i]);
1034             int64_t cpIndex = utext_getNativeIndex(ut);
1035             TEST_ASSERT(cpIndex == startMap[i]);
1036         }
1037 
1038         // Check utext_next32From
1039         for (i=0; i<startMapLimit; i++) {
1040             UChar32 c32 = utext_next32From(ut, i);
1041             TEST_ASSERT(c32 == c32Map[i]);
1042             int64_t cpIndex = utext_getNativeIndex(ut);
1043             TEST_ASSERT(cpIndex == nextMap[i]);
1044         }
1045 
1046         // check utext_previous32From
1047         for (i=0; i<startMapLimit; i++) {
1048             gTestNum++;
1049             UChar32 c32 = utext_previous32From(ut, i);
1050             TEST_ASSERT(c32 == pr32Map[i]);
1051             int64_t cpIndex = utext_getNativeIndex(ut);
1052             TEST_ASSERT(cpIndex == prevMap[i]);
1053         }
1054 
1055         // check Extract
1056         //   Extract from i to i+1, which may be zero or one code points,
1057         //     depending on whether the indices straddle a cp boundary.
1058         for (i=0; i<startMapLimit; i++) {
1059             UChar buf[3];
1060             status = U_ZERO_ERROR;
1061             int32_t  extractedLen = utext_extract(ut, i, i+1, buf, 3, &status);
1062             TEST_SUCCESS(status);
1063             TEST_ASSERT(extractedLen == exLen[i]);
1064             if (extractedLen > 0) {
1065                 UChar32  c32;
1066                 /* extractedLen-extractedLen == 0 is used to get around a compiler warning. */
1067                 U16_GET(buf, 0, extractedLen-extractedLen, extractedLen, c32);
1068                 TEST_ASSERT(c32 == c32Map[i]);
1069             }
1070         }
1071 
1072         utext_close(ut);
1073     }
1074 
1075 
1076     {    //  Similar test, with utf16 instead of utf8
1077          //  TODO:  merge the common parts of these tests.
1078 
1079         UnicodeString u16str("\\u1000\\U00011000\\u2000\\U00022000", -1, US_INV);
1080         int32_t startMap[]  ={ 0,     1,   1,    3,     4,  4,     6,  6};
1081         int32_t nextMap[]  = { 1,     3,   3,    4,     6,  6,     6,  6};
1082         int32_t prevMap[]  = { 0,     0,   0,    1,     3,  3,     4,  4};
1083         UChar32  c32Map[] =  {0x1000, 0x11000, 0x11000, 0x2000,  0x22000, 0x22000, -1, -1};
1084         UChar32  pr32Map[] = {    -1, 0x1000,  0x1000,  0x11000, 0x2000,  0x2000,   0x22000,   0x22000};
1085         int32_t  exLen[] =   {   1,  0,   2,  1,  0,  2,  0,  0,};
1086 
1087         u16str = u16str.unescape();
1088         UErrorCode status = U_ZERO_ERROR;
1089         UText *ut = utext_openUnicodeString(NULL, &u16str, &status);
1090         TEST_SUCCESS(status);
1091 
1092         int32_t startMapLimit = UPRV_LENGTHOF(startMap);
1093         int i;
1094         for (i=0; i<startMapLimit; i++) {
1095             utext_setNativeIndex(ut, i);
1096             int64_t cpIndex = utext_getNativeIndex(ut);
1097             TEST_ASSERT(cpIndex == startMap[i]);
1098         }
1099 
1100         // Check char32At
1101         for (i=0; i<startMapLimit; i++) {
1102             UChar32 c32 = utext_char32At(ut, i);
1103             TEST_ASSERT(c32 == c32Map[i]);
1104             int64_t cpIndex = utext_getNativeIndex(ut);
1105             TEST_ASSERT(cpIndex == startMap[i]);
1106         }
1107 
1108         // Check utext_next32From
1109         for (i=0; i<startMapLimit; i++) {
1110             UChar32 c32 = utext_next32From(ut, i);
1111             TEST_ASSERT(c32 == c32Map[i]);
1112             int64_t cpIndex = utext_getNativeIndex(ut);
1113             TEST_ASSERT(cpIndex == nextMap[i]);
1114         }
1115 
1116         // check utext_previous32From
1117         for (i=0; i<startMapLimit; i++) {
1118             UChar32 c32 = utext_previous32From(ut, i);
1119             TEST_ASSERT(c32 == pr32Map[i]);
1120             int64_t cpIndex = utext_getNativeIndex(ut);
1121             TEST_ASSERT(cpIndex == prevMap[i]);
1122         }
1123 
1124         // check Extract
1125         //   Extract from i to i+1, which may be zero or one code points,
1126         //     depending on whether the indices straddle a cp boundary.
1127         for (i=0; i<startMapLimit; i++) {
1128             UChar buf[3];
1129             status = U_ZERO_ERROR;
1130             int32_t  extractedLen = utext_extract(ut, i, i+1, buf, 3, &status);
1131             TEST_SUCCESS(status);
1132             TEST_ASSERT(extractedLen == exLen[i]);
1133             if (extractedLen > 0) {
1134                 UChar32  c32;
1135                 /* extractedLen-extractedLen == 0 is used to get around a compiler warning. */
1136                 U16_GET(buf, 0, extractedLen-extractedLen, extractedLen, c32);
1137                 TEST_ASSERT(c32 == c32Map[i]);
1138             }
1139         }
1140 
1141         utext_close(ut);
1142     }
1143 
1144     {    //  Similar test, with UText over Replaceable
1145          //  TODO:  merge the common parts of these tests.
1146 
1147         UnicodeString u16str("\\u1000\\U00011000\\u2000\\U00022000", -1, US_INV);
1148         int32_t startMap[]  ={ 0,     1,   1,    3,     4,  4,     6,  6};
1149         int32_t nextMap[]  = { 1,     3,   3,    4,     6,  6,     6,  6};
1150         int32_t prevMap[]  = { 0,     0,   0,    1,     3,  3,     4,  4};
1151         UChar32  c32Map[] =  {0x1000, 0x11000, 0x11000, 0x2000,  0x22000, 0x22000, -1, -1};
1152         UChar32  pr32Map[] = {    -1, 0x1000,  0x1000,  0x11000, 0x2000,  0x2000,   0x22000,   0x22000};
1153         int32_t  exLen[] =   {   1,  0,   2,  1,  0,  2,  0,  0,};
1154 
1155         u16str = u16str.unescape();
1156         UErrorCode status = U_ZERO_ERROR;
1157         UText *ut = utext_openReplaceable(NULL, &u16str, &status);
1158         TEST_SUCCESS(status);
1159 
1160         int32_t startMapLimit = UPRV_LENGTHOF(startMap);
1161         int i;
1162         for (i=0; i<startMapLimit; i++) {
1163             utext_setNativeIndex(ut, i);
1164             int64_t cpIndex = utext_getNativeIndex(ut);
1165             TEST_ASSERT(cpIndex == startMap[i]);
1166         }
1167 
1168         // Check char32At
1169         for (i=0; i<startMapLimit; i++) {
1170             UChar32 c32 = utext_char32At(ut, i);
1171             TEST_ASSERT(c32 == c32Map[i]);
1172             int64_t cpIndex = utext_getNativeIndex(ut);
1173             TEST_ASSERT(cpIndex == startMap[i]);
1174         }
1175 
1176         // Check utext_next32From
1177         for (i=0; i<startMapLimit; i++) {
1178             UChar32 c32 = utext_next32From(ut, i);
1179             TEST_ASSERT(c32 == c32Map[i]);
1180             int64_t cpIndex = utext_getNativeIndex(ut);
1181             TEST_ASSERT(cpIndex == nextMap[i]);
1182         }
1183 
1184         // check utext_previous32From
1185         for (i=0; i<startMapLimit; i++) {
1186             UChar32 c32 = utext_previous32From(ut, i);
1187             TEST_ASSERT(c32 == pr32Map[i]);
1188             int64_t cpIndex = utext_getNativeIndex(ut);
1189             TEST_ASSERT(cpIndex == prevMap[i]);
1190         }
1191 
1192         // check Extract
1193         //   Extract from i to i+1, which may be zero or one code points,
1194         //     depending on whether the indices straddle a cp boundary.
1195         for (i=0; i<startMapLimit; i++) {
1196             UChar buf[3];
1197             status = U_ZERO_ERROR;
1198             int32_t  extractedLen = utext_extract(ut, i, i+1, buf, 3, &status);
1199             TEST_SUCCESS(status);
1200             TEST_ASSERT(extractedLen == exLen[i]);
1201             if (extractedLen > 0) {
1202                 UChar32  c32;
1203                 /* extractedLen-extractedLen == 0 is used to get around a compiler warning. */
1204                 U16_GET(buf, 0, extractedLen-extractedLen, extractedLen, c32);
1205                 TEST_ASSERT(c32 == c32Map[i]);
1206             }
1207         }
1208 
1209         utext_close(ut);
1210     }
1211 }
1212 
1213 
FreezeTest()1214 void UTextTest::FreezeTest() {
1215     // Check isWritable() and freeze() behavior.
1216     //
1217 
1218     UnicodeString  ustr("Hello, World.");
1219     const char u8str[] = {char(0x31), (char)0x32, (char)0x33, 0};
1220     const UChar u16str[] = {(UChar)0x31, (UChar)0x32, (UChar)0x44, 0};
1221 
1222     UErrorCode status = U_ZERO_ERROR;
1223     UText  *ut        = NULL;
1224     UText  *ut2       = NULL;
1225 
1226     ut = utext_openUTF8(ut, u8str, -1, &status);
1227     TEST_SUCCESS(status);
1228     UBool writable = utext_isWritable(ut);
1229     TEST_ASSERT(writable == FALSE);
1230     utext_copy(ut, 1, 2, 0, TRUE, &status);
1231     TEST_ASSERT(status == U_NO_WRITE_PERMISSION);
1232 
1233     status = U_ZERO_ERROR;
1234     ut = utext_openUChars(ut, u16str, -1, &status);
1235     TEST_SUCCESS(status);
1236     writable = utext_isWritable(ut);
1237     TEST_ASSERT(writable == FALSE);
1238     utext_copy(ut, 1, 2, 0, TRUE, &status);
1239     TEST_ASSERT(status == U_NO_WRITE_PERMISSION);
1240 
1241     status = U_ZERO_ERROR;
1242     ut = utext_openUnicodeString(ut, &ustr, &status);
1243     TEST_SUCCESS(status);
1244     writable = utext_isWritable(ut);
1245     TEST_ASSERT(writable == TRUE);
1246     utext_freeze(ut);
1247     writable = utext_isWritable(ut);
1248     TEST_ASSERT(writable == FALSE);
1249     utext_copy(ut, 1, 2, 0, TRUE, &status);
1250     TEST_ASSERT(status == U_NO_WRITE_PERMISSION);
1251 
1252     status = U_ZERO_ERROR;
1253     ut = utext_openUnicodeString(ut, &ustr, &status);
1254     TEST_SUCCESS(status);
1255     ut2 = utext_clone(ut2, ut, FALSE, FALSE, &status);  // clone with readonly = false
1256     TEST_SUCCESS(status);
1257     writable = utext_isWritable(ut2);
1258     TEST_ASSERT(writable == TRUE);
1259     ut2 = utext_clone(ut2, ut, FALSE, TRUE, &status);  // clone with readonly = true
1260     TEST_SUCCESS(status);
1261     writable = utext_isWritable(ut2);
1262     TEST_ASSERT(writable == FALSE);
1263     utext_copy(ut2, 1, 2, 0, TRUE, &status);
1264     TEST_ASSERT(status == U_NO_WRITE_PERMISSION);
1265 
1266     status = U_ZERO_ERROR;
1267     ut = utext_openConstUnicodeString(ut, (const UnicodeString *)&ustr, &status);
1268     TEST_SUCCESS(status);
1269     writable = utext_isWritable(ut);
1270     TEST_ASSERT(writable == FALSE);
1271     utext_copy(ut, 1, 2, 0, TRUE, &status);
1272     TEST_ASSERT(status == U_NO_WRITE_PERMISSION);
1273 
1274     // Deep Clone of a frozen UText should re-enable writing in the copy.
1275     status = U_ZERO_ERROR;
1276     ut = utext_openUnicodeString(ut, &ustr, &status);
1277     TEST_SUCCESS(status);
1278     utext_freeze(ut);
1279     ut2 = utext_clone(ut2, ut, TRUE, FALSE, &status);   // deep clone
1280     TEST_SUCCESS(status);
1281     writable = utext_isWritable(ut2);
1282     TEST_ASSERT(writable == TRUE);
1283 
1284 
1285     // Deep clone of a frozen UText, where the base type is intrinsically non-writable,
1286     //  should NOT enable writing in the copy.
1287     status = U_ZERO_ERROR;
1288     ut = utext_openUChars(ut, u16str, -1, &status);
1289     TEST_SUCCESS(status);
1290     utext_freeze(ut);
1291     ut2 = utext_clone(ut2, ut, TRUE, FALSE, &status);   // deep clone
1292     TEST_SUCCESS(status);
1293     writable = utext_isWritable(ut2);
1294     TEST_ASSERT(writable == FALSE);
1295 
1296     // cleanup
1297     utext_close(ut);
1298     utext_close(ut2);
1299 }
1300 
1301 
1302 //
1303 //  Fragmented UText
1304 //      A UText type that works with a chunk size of 1.
1305 //      Intended to test for edge cases.
1306 //      Input comes from a UnicodeString.
1307 //
1308 //       ut.b    the character.  Put into both halves.
1309 //
1310 
1311 U_CDECL_BEGIN
1312 static UBool U_CALLCONV
fragTextAccess(UText * ut,int64_t index,UBool forward)1313 fragTextAccess(UText *ut, int64_t index, UBool forward) {
1314     const UnicodeString *us = (const UnicodeString *)ut->context;
1315     UChar  c;
1316     int32_t length = us->length();
1317     if (forward && index>=0 && index<length) {
1318         c = us->charAt((int32_t)index);
1319         ut->b = c | c<<16;
1320         ut->chunkOffset = 0;
1321         ut->chunkLength = 1;
1322         ut->chunkNativeStart = index;
1323         ut->chunkNativeLimit = index+1;
1324         return true;
1325     }
1326     if (!forward && index>0 && index <=length) {
1327         c = us->charAt((int32_t)index-1);
1328         ut->b = c | c<<16;
1329         ut->chunkOffset = 1;
1330         ut->chunkLength = 1;
1331         ut->chunkNativeStart = index-1;
1332         ut->chunkNativeLimit = index;
1333         return true;
1334     }
1335     ut->b = 0;
1336     ut->chunkOffset = 0;
1337     ut->chunkLength = 0;
1338     if (index <= 0) {
1339         ut->chunkNativeStart = 0;
1340         ut->chunkNativeLimit = 0;
1341     } else {
1342         ut->chunkNativeStart = length;
1343         ut->chunkNativeLimit = length;
1344     }
1345     return false;
1346 }
1347 
1348 // Function table to be used with this fragmented text provider.
1349 //   Initialized in the open function.
1350 static UTextFuncs  fragmentFuncs;
1351 
1352 // Clone function for fragmented text provider.
1353 //   Didn't really want to provide this, but it's easier to provide it than to keep it
1354 //   out of the tests.
1355 //
1356 UText *
cloneFragmentedUnicodeString(UText * dest,const UText * src,UBool deep,UErrorCode * status)1357 cloneFragmentedUnicodeString(UText *dest, const UText *src, UBool deep, UErrorCode *status) {
1358     if (U_FAILURE(*status)) {
1359         return NULL;
1360     }
1361     if (deep) {
1362         *status = U_UNSUPPORTED_ERROR;
1363         return NULL;
1364     }
1365     dest = utext_openUnicodeString(dest, (UnicodeString *)src->context, status);
1366     utext_setNativeIndex(dest, utext_getNativeIndex(src));
1367     return dest;
1368 }
1369 
1370 U_CDECL_END
1371 
1372 // Open function for the fragmented text provider.
1373 UText *
openFragmentedUnicodeString(UText * ut,UnicodeString * s,UErrorCode * status)1374 openFragmentedUnicodeString(UText *ut, UnicodeString *s, UErrorCode *status) {
1375     ut = utext_openUnicodeString(ut, s, status);
1376     if (U_FAILURE(*status)) {
1377         return ut;
1378     }
1379 
1380     // Copy of the function table from the stock UnicodeString UText,
1381     //   and replace the entry for the access function.
1382     memcpy(&fragmentFuncs, ut->pFuncs, sizeof(fragmentFuncs));
1383     fragmentFuncs.access = fragTextAccess;
1384     fragmentFuncs.clone  = cloneFragmentedUnicodeString;
1385     ut->pFuncs = &fragmentFuncs;
1386 
1387     ut->chunkContents = (UChar *)&ut->b;
1388     ut->pFuncs->access(ut, 0, TRUE);
1389     return ut;
1390 }
1391 
1392 // Regression test for Ticket 5560
1393 //   Clone fails to update chunkContentPointer in the cloned copy.
1394 //   This is only an issue for UText types that work in a local buffer,
1395 //      (UTF-8 wrapper, for example)
1396 //
1397 //   The test:
1398 //     1.  Create an inital UText
1399 //     2.  Deep clone it.  Contents should match original.
1400 //     3.  Reset original to something different.
1401 //     4.  Check that clone contents did not change.
1402 //
Ticket5560()1403 void UTextTest::Ticket5560() {
1404     /* The following two strings are in UTF-8 even on EBCDIC platforms. */
1405     static const char s1[] = {0x41,0x42,0x43,0x44,0x45,0x46,0}; /* "ABCDEF" */
1406     static const char s2[] = {0x31,0x32,0x33,0x34,0x35,0x36,0}; /* "123456" */
1407 	UErrorCode status = U_ZERO_ERROR;
1408 
1409 	UText ut1 = UTEXT_INITIALIZER;
1410 	UText ut2 = UTEXT_INITIALIZER;
1411 
1412 	utext_openUTF8(&ut1, s1, -1, &status);
1413 	UChar c = utext_next32(&ut1);
1414 	TEST_ASSERT(c == 0x41);  // c == 'A'
1415 
1416 	utext_clone(&ut2, &ut1, TRUE, FALSE, &status);
1417 	TEST_SUCCESS(status);
1418     c = utext_next32(&ut2);
1419 	TEST_ASSERT(c == 0x42);  // c == 'B'
1420     c = utext_next32(&ut1);
1421 	TEST_ASSERT(c == 0x42);  // c == 'B'
1422 
1423 	utext_openUTF8(&ut1, s2, -1, &status);
1424 	c = utext_next32(&ut1);
1425 	TEST_ASSERT(c == 0x31);  // c == '1'
1426     c = utext_next32(&ut2);
1427 	TEST_ASSERT(c == 0x43);  // c == 'C'
1428 
1429     utext_close(&ut1);
1430     utext_close(&ut2);
1431 }
1432 
1433 
1434 // Test for Ticket 6847
1435 //
Ticket6847()1436 void UTextTest::Ticket6847() {
1437     const int STRLEN = 90;
1438     UChar s[STRLEN+1];
1439     u_memset(s, 0x41, STRLEN);
1440     s[STRLEN] = 0;
1441 
1442     UErrorCode status = U_ZERO_ERROR;
1443     UText *ut = utext_openUChars(NULL, s, -1, &status);
1444 
1445     utext_setNativeIndex(ut, 0);
1446     int32_t count = 0;
1447     UChar32 c = 0;
1448     int64_t nativeIndex = UTEXT_GETNATIVEINDEX(ut);
1449     TEST_ASSERT(nativeIndex == 0);
1450     while ((c = utext_next32(ut)) != U_SENTINEL) {
1451         TEST_ASSERT(c == 0x41);
1452         TEST_ASSERT(count < STRLEN);
1453         if (count >= STRLEN) {
1454             break;
1455         }
1456         count++;
1457         nativeIndex = UTEXT_GETNATIVEINDEX(ut);
1458         TEST_ASSERT(nativeIndex == count);
1459     }
1460     TEST_ASSERT(count == STRLEN);
1461     nativeIndex = UTEXT_GETNATIVEINDEX(ut);
1462     TEST_ASSERT(nativeIndex == STRLEN);
1463     utext_close(ut);
1464 }
1465 
1466 
Ticket10562()1467 void UTextTest::Ticket10562() {
1468     // Note: failures show as a heap error when the test is run under valgrind.
1469     UErrorCode status = U_ZERO_ERROR;
1470 
1471     const char *utf8_string = "\x41\x41\x41\x41\x41\x41\x41\x41\x41\x41\x41\x41\x41\x41\x41";
1472     UText *utf8Text = utext_openUTF8(NULL, utf8_string, -1, &status);
1473     TEST_SUCCESS(status);
1474     UText *deepClone = utext_clone(NULL, utf8Text, TRUE, FALSE, &status);
1475     TEST_SUCCESS(status);
1476     UText *shallowClone = utext_clone(NULL, deepClone, FALSE, FALSE, &status);
1477     TEST_SUCCESS(status);
1478     utext_close(shallowClone);
1479     utext_close(deepClone);
1480     utext_close(utf8Text);
1481 
1482     status = U_ZERO_ERROR;
1483     UnicodeString usString("Hello, World.");
1484     UText *usText = utext_openUnicodeString(NULL, &usString, &status);
1485     TEST_SUCCESS(status);
1486     UText *usDeepClone = utext_clone(NULL, usText, TRUE, FALSE, &status);
1487     TEST_SUCCESS(status);
1488     UText *usShallowClone = utext_clone(NULL, usDeepClone, FALSE, FALSE, &status);
1489     TEST_SUCCESS(status);
1490     utext_close(usShallowClone);
1491     utext_close(usDeepClone);
1492     utext_close(usText);
1493 }
1494 
1495 
Ticket10983()1496 void UTextTest::Ticket10983() {
1497     // Note: failure shows as a seg fault when the defect is present.
1498 
1499     UErrorCode status = U_ZERO_ERROR;
1500     UnicodeString s("Hello, World");
1501     UText *ut = utext_openConstUnicodeString(NULL, &s, &status);
1502     TEST_SUCCESS(status);
1503 
1504     status = U_INVALID_STATE_ERROR;
1505     UText *cloned = utext_clone(NULL, ut, TRUE, TRUE, &status);
1506     TEST_ASSERT(cloned == NULL);
1507     TEST_ASSERT(status == U_INVALID_STATE_ERROR);
1508 
1509     utext_close(ut);
1510 }
1511 
1512 // Ticket 12130 - extract on a UText wrapping a null terminated UChar * string
1513 //                leaves the iteration position set incorrectly when the
1514 //                actual string length is not yet known.
1515 //
1516 //                The test text needs to be long enough that UText defers getting the length.
1517 
Ticket12130()1518 void UTextTest::Ticket12130() {
1519     UErrorCode status = U_ZERO_ERROR;
1520 
1521     const char *text8 =
1522         "Fundamentally, computers just deal with numbers. They store letters and other characters "
1523         "by assigning a number for each one. Before Unicode was invented, there were hundreds "
1524         "of different encoding systems for assigning these numbers. No single encoding could "
1525         "contain enough characters: for example, the European Union alone requires several "
1526         "different encodings to cover all its languages. Even for a single language like "
1527         "English no single encoding was adequate for all the letters, punctuation, and technical "
1528         "symbols in common use.";
1529 
1530     UnicodeString str(text8);
1531     const UChar *ustr = str.getTerminatedBuffer();
1532     UText ut = UTEXT_INITIALIZER;
1533     utext_openUChars(&ut, ustr, -1, &status);
1534     UChar extractBuffer[50];
1535 
1536     for (int32_t startIdx = 0; startIdx<str.length(); ++startIdx) {
1537         int32_t endIdx = startIdx + 20;
1538 
1539         u_memset(extractBuffer, 0, UPRV_LENGTHOF(extractBuffer));
1540         utext_extract(&ut, startIdx, endIdx, extractBuffer, UPRV_LENGTHOF(extractBuffer), &status);
1541         if (U_FAILURE(status)) {
1542             errln("%s:%d %s", __FILE__, __LINE__, u_errorName(status));
1543             return;
1544         }
1545         int64_t ni  = utext_getNativeIndex(&ut);
1546         int64_t expectedni = startIdx + 20;
1547         if (expectedni > str.length()) {
1548             expectedni = str.length();
1549         }
1550         if (expectedni != ni) {
1551             errln("%s:%d utext_getNativeIndex() expected %d, got %d", __FILE__, __LINE__, expectedni, ni);
1552         }
1553         if (0 != str.tempSubString(startIdx, 20).compare(extractBuffer)) {
1554             errln("%s:%d utext_extract() failed. expected \"%s\", got \"%s\"",
1555                     __FILE__, __LINE__, CStr(str.tempSubString(startIdx, 20))(), CStr(UnicodeString(extractBuffer))());
1556         }
1557     }
1558     utext_close(&ut);
1559 
1560     // Similar utext extract, this time with the string length provided to the UText in advance,
1561     // and a buffer of larger than required capacity.
1562 
1563     utext_openUChars(&ut, ustr, str.length(), &status);
1564     for (int32_t startIdx = 0; startIdx<str.length(); ++startIdx) {
1565         int32_t endIdx = startIdx + 20;
1566         u_memset(extractBuffer, 0, UPRV_LENGTHOF(extractBuffer));
1567         utext_extract(&ut, startIdx, endIdx, extractBuffer, UPRV_LENGTHOF(extractBuffer), &status);
1568         if (U_FAILURE(status)) {
1569             errln("%s:%d %s", __FILE__, __LINE__, u_errorName(status));
1570             return;
1571         }
1572         int64_t ni  = utext_getNativeIndex(&ut);
1573         int64_t expectedni = startIdx + 20;
1574         if (expectedni > str.length()) {
1575             expectedni = str.length();
1576         }
1577         if (expectedni != ni) {
1578             errln("%s:%d utext_getNativeIndex() expected %d, got %d", __FILE__, __LINE__, expectedni, ni);
1579         }
1580         if (0 != str.tempSubString(startIdx, 20).compare(extractBuffer)) {
1581             errln("%s:%d utext_extract() failed. expected \"%s\", got \"%s\"",
1582                     __FILE__, __LINE__, CStr(str.tempSubString(startIdx, 20))(), CStr(UnicodeString(extractBuffer))());
1583         }
1584     }
1585     utext_close(&ut);
1586 }
1587 
1588 // Ticket 13344 The macro form of UTEXT_SETNATIVEINDEX failed when target was a trail surrogate
1589 //              of a supplementary character.
1590 
Ticket13344()1591 void UTextTest::Ticket13344() {
1592     UErrorCode status = U_ZERO_ERROR;
1593     const char16_t *str = u"abc\U0010abcd xyz";
1594     LocalUTextPointer ut(utext_openUChars(NULL, str, -1, &status));
1595 
1596     assertSuccess("UTextTest::Ticket13344-status", status);
1597     UTEXT_SETNATIVEINDEX(ut.getAlias(), 3);
1598     assertEquals("UTextTest::Ticket13344-lead", (int64_t)3, utext_getNativeIndex(ut.getAlias()));
1599     UTEXT_SETNATIVEINDEX(ut.getAlias(), 4);
1600     assertEquals("UTextTest::Ticket13344-trail", (int64_t)3, utext_getNativeIndex(ut.getAlias()));
1601     UTEXT_SETNATIVEINDEX(ut.getAlias(), 5);
1602     assertEquals("UTextTest::Ticket13344-bmp", (int64_t)5, utext_getNativeIndex(ut.getAlias()));
1603 
1604     utext_setNativeIndex(ut.getAlias(), 3);
1605     assertEquals("UTextTest::Ticket13344-lead-2", (int64_t)3, utext_getNativeIndex(ut.getAlias()));
1606     utext_setNativeIndex(ut.getAlias(), 4);
1607     assertEquals("UTextTest::Ticket13344-trail-2", (int64_t)3, utext_getNativeIndex(ut.getAlias()));
1608     utext_setNativeIndex(ut.getAlias(), 5);
1609     assertEquals("UTextTest::Ticket13344-bmp-2", (int64_t)5, utext_getNativeIndex(ut.getAlias()));
1610 }
1611 
1612