]>
Commit | Line | Data |
---|---|---|
46f4442e A |
1 | /* |
2 | ********************************************************************** | |
51004dcb | 3 | * Copyright (C) 2005-2013, International Business Machines |
46f4442e A |
4 | * Corporation and others. All Rights Reserved. |
5 | ********************************************************************** | |
6 | */ | |
7 | ||
46f4442e A |
8 | #include "unicode/utypes.h" |
9 | ||
10 | #if !UCONFIG_NO_COLLATION | |
11 | ||
46f4442e | 12 | #include "cmemory.h" |
51004dcb A |
13 | #include "cstring.h" |
14 | #include "ucol_imp.h" | |
15 | ||
46f4442e A |
16 | #include "unicode/coll.h" |
17 | #include "unicode/tblcoll.h" | |
51004dcb | 18 | #include "unicode/usearch.h" |
46f4442e A |
19 | #include "unicode/uset.h" |
20 | #include "unicode/ustring.h" | |
46f4442e | 21 | |
51004dcb A |
22 | #include "unicode/coleitr.h" |
23 | #include "unicode/regex.h" // TODO: make conditional on regexp being built. | |
729e4ab9 | 24 | |
51004dcb A |
25 | #include "colldata.h" |
26 | #include "ssearch.h" | |
46f4442e A |
27 | #include "xmlparser.h" |
28 | ||
51004dcb | 29 | #include <stdio.h> // for sprintf |
46f4442e A |
30 | |
31 | char testId[100]; | |
32 | ||
33 | #define TEST_ASSERT(x) {if (!(x)) { \ | |
34 | errln("Failure in file %s, line %d, test ID = \"%s\"", __FILE__, __LINE__, testId);}} | |
35 | ||
36 | #define TEST_ASSERT_M(x, m) {if (!(x)) { \ | |
51004dcb | 37 | dataerrln("Failure in file %s, line %d. \"%s\"", __FILE__, __LINE__, m);return;}} |
46f4442e A |
38 | |
39 | #define TEST_ASSERT_SUCCESS(errcode) {if (U_FAILURE(errcode)) { \ | |
729e4ab9 | 40 | dataerrln("Failure in file %s, line %d, test ID \"%s\", status = \"%s\"", \ |
46f4442e A |
41 | __FILE__, __LINE__, testId, u_errorName(errcode));}} |
42 | ||
43 | #define ARRAY_SIZE(array) (sizeof array / sizeof array[0]) | |
729e4ab9 A |
44 | #define NEW_ARRAY(type, count) (type *) uprv_malloc((count) * sizeof(type)) |
45 | #define DELETE_ARRAY(array) uprv_free((void *) (array)) | |
46f4442e A |
46 | |
47 | //--------------------------------------------------------------------------- | |
48 | // | |
49 | // Test class boilerplate | |
50 | // | |
51 | //--------------------------------------------------------------------------- | |
52 | SSearchTest::SSearchTest() | |
53 | { | |
54 | } | |
55 | ||
56 | SSearchTest::~SSearchTest() | |
57 | { | |
58 | } | |
59 | ||
60 | void SSearchTest::runIndexedTest( int32_t index, UBool exec, const char* &name, char *params ) | |
61 | { | |
62 | if (exec) logln("TestSuite SSearchTest: "); | |
63 | switch (index) { | |
64 | #if !UCONFIG_NO_BREAK_ITERATION | |
65 | case 0: name = "searchTest"; | |
66 | if (exec) searchTest(); | |
67 | break; | |
68 | ||
69 | case 1: name = "offsetTest"; | |
70 | if (exec) offsetTest(); | |
71 | break; | |
72 | ||
73 | case 2: name = "monkeyTest"; | |
74 | if (exec) monkeyTest(params); | |
75 | break; | |
729e4ab9 | 76 | |
51004dcb A |
77 | case 3: name = "sharpSTest"; |
78 | if (exec) sharpSTest(); | |
729e4ab9 A |
79 | break; |
80 | ||
51004dcb | 81 | case 4: name = "goodSuffixTest"; |
729e4ab9 A |
82 | if (exec) goodSuffixTest(); |
83 | break; | |
84 | ||
51004dcb | 85 | case 5: name = "searchTime"; |
729e4ab9 A |
86 | if (exec) searchTime(); |
87 | break; | |
46f4442e A |
88 | #endif |
89 | default: name = ""; | |
90 | break; //needed to end loop | |
91 | } | |
92 | } | |
93 | ||
94 | ||
95 | #if !UCONFIG_NO_BREAK_ITERATION | |
96 | ||
97 | #define PATH_BUFFER_SIZE 2048 | |
98 | const char *SSearchTest::getPath(char buffer[2048], const char *filename) { | |
99 | UErrorCode status = U_ZERO_ERROR; | |
100 | const char *testDataDirectory = IntlTest::getSourceTestData(status); | |
101 | ||
102 | if (U_FAILURE(status) || strlen(testDataDirectory) + strlen(filename) + 1 >= PATH_BUFFER_SIZE) { | |
103 | errln("ERROR: getPath() failed - %s", u_errorName(status)); | |
104 | return NULL; | |
105 | } | |
106 | ||
107 | strcpy(buffer, testDataDirectory); | |
108 | strcat(buffer, filename); | |
109 | return buffer; | |
110 | } | |
111 | ||
112 | ||
113 | void SSearchTest::searchTest() | |
114 | { | |
729e4ab9 | 115 | #if !UCONFIG_NO_REGULAR_EXPRESSIONS && !UCONFIG_NO_FILE_IO |
46f4442e A |
116 | UErrorCode status = U_ZERO_ERROR; |
117 | char path[PATH_BUFFER_SIZE]; | |
118 | const char *testFilePath = getPath(path, "ssearch.xml"); | |
119 | ||
120 | if (testFilePath == NULL) { | |
121 | return; /* Couldn't get path: error message already output. */ | |
122 | } | |
123 | ||
729e4ab9 | 124 | LocalPointer<UXMLParser> parser(UXMLParser::createParser(status)); |
46f4442e | 125 | TEST_ASSERT_SUCCESS(status); |
729e4ab9 | 126 | LocalPointer<UXMLElement> root(parser->parseFile(testFilePath, status)); |
46f4442e A |
127 | TEST_ASSERT_SUCCESS(status); |
128 | if (U_FAILURE(status)) { | |
129 | return; | |
130 | } | |
131 | ||
132 | const UnicodeString *debugTestCase = root->getAttribute("debug"); | |
133 | if (debugTestCase != NULL) { | |
134 | // setenv("USEARCH_DEBUG", "1", 1); | |
135 | } | |
136 | ||
137 | ||
138 | const UXMLElement *testCase; | |
139 | int32_t tc = 0; | |
140 | ||
141 | while((testCase = root->nextChildElement(tc)) != NULL) { | |
142 | ||
143 | if (testCase->getTagName().compare("test-case") != 0) { | |
144 | errln("ssearch, unrecognized XML Element in test file"); | |
145 | continue; | |
146 | } | |
147 | const UnicodeString *id = testCase->getAttribute("id"); | |
148 | *testId = 0; | |
149 | if (id != NULL) { | |
150 | id->extract(0, id->length(), testId, sizeof(testId), US_INV); | |
151 | } | |
152 | ||
153 | // If debugging test case has been specified and this is not it, skip to next. | |
154 | if (id!=NULL && debugTestCase!=NULL && *id != *debugTestCase) { | |
155 | continue; | |
156 | } | |
157 | // | |
158 | // Get the requested collation strength. | |
159 | // Default is tertiary if the XML attribute is missing from the test case. | |
160 | // | |
161 | const UnicodeString *strength = testCase->getAttribute("strength"); | |
729e4ab9 | 162 | UColAttributeValue collatorStrength = UCOL_PRIMARY; |
46f4442e A |
163 | if (strength==NULL) { collatorStrength = UCOL_TERTIARY;} |
164 | else if (*strength=="PRIMARY") { collatorStrength = UCOL_PRIMARY;} | |
165 | else if (*strength=="SECONDARY") { collatorStrength = UCOL_SECONDARY;} | |
166 | else if (*strength=="TERTIARY") { collatorStrength = UCOL_TERTIARY;} | |
167 | else if (*strength=="QUATERNARY") { collatorStrength = UCOL_QUATERNARY;} | |
168 | else if (*strength=="IDENTICAL") { collatorStrength = UCOL_IDENTICAL;} | |
169 | else { | |
170 | // Bogus value supplied for strength. Shouldn't happen, even from | |
171 | // typos, if the XML source has been validated. | |
172 | // This assert is a little deceiving in that strength can be | |
173 | // any of the allowed values, not just TERTIARY, but it will | |
174 | // do the job of getting the error output. | |
175 | TEST_ASSERT(*strength=="TERTIARY") | |
176 | } | |
177 | ||
178 | // | |
179 | // Get the collator normalization flag. Default is UCOL_OFF. | |
180 | // | |
181 | UColAttributeValue normalize = UCOL_OFF; | |
182 | const UnicodeString *norm = testCase->getAttribute("norm"); | |
183 | TEST_ASSERT (norm==NULL || *norm=="ON" || *norm=="OFF"); | |
184 | if (norm!=NULL && *norm=="ON") { | |
185 | normalize = UCOL_ON; | |
186 | } | |
187 | ||
729e4ab9 A |
188 | // |
189 | // Get the alternate_handling flag. Default is UCOL_NON_IGNORABLE. | |
190 | // | |
191 | UColAttributeValue alternateHandling = UCOL_NON_IGNORABLE; | |
192 | const UnicodeString *alt = testCase->getAttribute("alternate_handling"); | |
193 | TEST_ASSERT (alt == NULL || *alt == "SHIFTED" || *alt == "NON_IGNORABLE"); | |
194 | if (alt != NULL && *alt == "SHIFTED") { | |
195 | alternateHandling = UCOL_SHIFTED; | |
196 | } | |
197 | ||
46f4442e A |
198 | const UnicodeString defLocale("en"); |
199 | char clocale[100]; | |
200 | const UnicodeString *locale = testCase->getAttribute("locale"); | |
201 | if (locale == NULL || locale->length()==0) { | |
202 | locale = &defLocale; | |
203 | }; | |
204 | locale->extract(0, locale->length(), clocale, sizeof(clocale), NULL); | |
205 | ||
206 | ||
207 | UnicodeString text; | |
208 | UnicodeString target; | |
209 | UnicodeString pattern; | |
210 | int32_t expectedMatchStart = -1; | |
211 | int32_t expectedMatchLimit = -1; | |
212 | const UXMLElement *n; | |
729e4ab9 | 213 | int32_t nodeCount = 0; |
46f4442e A |
214 | |
215 | n = testCase->getChildElement("pattern"); | |
216 | TEST_ASSERT(n != NULL); | |
217 | if (n==NULL) { | |
218 | continue; | |
219 | } | |
220 | text = n->getText(FALSE); | |
221 | text = text.unescape(); | |
222 | pattern.append(text); | |
223 | nodeCount++; | |
224 | ||
225 | n = testCase->getChildElement("pre"); | |
226 | if (n!=NULL) { | |
227 | text = n->getText(FALSE); | |
228 | text = text.unescape(); | |
229 | target.append(text); | |
230 | nodeCount++; | |
231 | } | |
729e4ab9 | 232 | |
46f4442e A |
233 | n = testCase->getChildElement("m"); |
234 | if (n!=NULL) { | |
235 | expectedMatchStart = target.length(); | |
236 | text = n->getText(FALSE); | |
237 | text = text.unescape(); | |
238 | target.append(text); | |
239 | expectedMatchLimit = target.length(); | |
240 | nodeCount++; | |
241 | } | |
242 | ||
243 | n = testCase->getChildElement("post"); | |
244 | if (n!=NULL) { | |
245 | text = n->getText(FALSE); | |
246 | text = text.unescape(); | |
247 | target.append(text); | |
248 | nodeCount++; | |
249 | } | |
250 | ||
251 | // Check that there weren't extra things in the XML | |
252 | TEST_ASSERT(nodeCount == testCase->countChildren()); | |
253 | ||
729e4ab9 | 254 | // Open a collator and StringSearch based on the parameters |
46f4442e A |
255 | // obtained from the XML. |
256 | // | |
257 | status = U_ZERO_ERROR; | |
729e4ab9 A |
258 | LocalUCollatorPointer collator(ucol_open(clocale, &status)); |
259 | ucol_setStrength(collator.getAlias(), collatorStrength); | |
260 | ucol_setAttribute(collator.getAlias(), UCOL_NORMALIZATION_MODE, normalize, &status); | |
261 | ucol_setAttribute(collator.getAlias(), UCOL_ALTERNATE_HANDLING, alternateHandling, &status); | |
262 | LocalUStringSearchPointer uss(usearch_openFromCollator(pattern.getBuffer(), pattern.length(), | |
263 | target.getBuffer(), target.length(), | |
264 | collator.getAlias(), | |
265 | NULL, // the break iterator | |
266 | &status)); | |
267 | ||
46f4442e A |
268 | TEST_ASSERT_SUCCESS(status); |
269 | if (U_FAILURE(status)) { | |
46f4442e A |
270 | continue; |
271 | } | |
272 | ||
273 | int32_t foundStart = 0; | |
274 | int32_t foundLimit = 0; | |
275 | UBool foundMatch; | |
276 | ||
277 | // | |
278 | // Do the search, check the match result against the expected results. | |
279 | // | |
729e4ab9 | 280 | foundMatch= usearch_search(uss.getAlias(), 0, &foundStart, &foundLimit, &status); |
46f4442e | 281 | TEST_ASSERT_SUCCESS(status); |
729e4ab9 A |
282 | if ((foundMatch && expectedMatchStart<0) || |
283 | (foundStart != expectedMatchStart) || | |
284 | (foundLimit != expectedMatchLimit)) { | |
46f4442e A |
285 | TEST_ASSERT(FALSE); // ouput generic error position |
286 | infoln("Found, expected match start = %d, %d \n" | |
287 | "Found, expected match limit = %d, %d", | |
288 | foundStart, expectedMatchStart, foundLimit, expectedMatchLimit); | |
289 | } | |
290 | ||
291 | // In case there are other matches... | |
292 | // (should we only do this if the test case passed?) | |
293 | while (foundMatch) { | |
294 | expectedMatchStart = foundStart; | |
295 | expectedMatchLimit = foundLimit; | |
296 | ||
729e4ab9 | 297 | foundMatch = usearch_search(uss.getAlias(), foundLimit, &foundStart, &foundLimit, &status); |
46f4442e A |
298 | } |
299 | ||
729e4ab9 | 300 | uss.adoptInstead(usearch_openFromCollator(pattern.getBuffer(), pattern.length(), |
46f4442e | 301 | target.getBuffer(), target.length(), |
729e4ab9 | 302 | collator.getAlias(), |
46f4442e | 303 | NULL, |
729e4ab9 | 304 | &status)); |
46f4442e A |
305 | |
306 | // | |
307 | // Do the backwards search, check the match result against the expected results. | |
308 | // | |
729e4ab9 | 309 | foundMatch= usearch_searchBackwards(uss.getAlias(), target.length(), &foundStart, &foundLimit, &status); |
46f4442e | 310 | TEST_ASSERT_SUCCESS(status); |
729e4ab9 A |
311 | if ((foundMatch && expectedMatchStart<0) || |
312 | (foundStart != expectedMatchStart) || | |
313 | (foundLimit != expectedMatchLimit)) { | |
46f4442e A |
314 | TEST_ASSERT(FALSE); // ouput generic error position |
315 | infoln("Found, expected backwards match start = %d, %d \n" | |
316 | "Found, expected backwards match limit = %d, %d", | |
317 | foundStart, expectedMatchStart, foundLimit, expectedMatchLimit); | |
318 | } | |
729e4ab9 A |
319 | } |
320 | #endif | |
321 | } | |
322 | ||
46f4442e A |
323 | struct Order |
324 | { | |
325 | int32_t order; | |
326 | int32_t lowOffset; | |
327 | int32_t highOffset; | |
328 | }; | |
329 | ||
330 | class OrderList | |
331 | { | |
332 | public: | |
333 | OrderList(); | |
334 | OrderList(UCollator *coll, const UnicodeString &string, int32_t stringOffset = 0); | |
335 | ~OrderList(); | |
336 | ||
337 | int32_t size(void) const; | |
338 | void add(int32_t order, int32_t low, int32_t high); | |
339 | const Order *get(int32_t index) const; | |
340 | int32_t getLowOffset(int32_t index) const; | |
341 | int32_t getHighOffset(int32_t index) const; | |
342 | int32_t getOrder(int32_t index) const; | |
343 | void reverse(void); | |
344 | UBool compare(const OrderList &other) const; | |
345 | UBool matchesAt(int32_t offset, const OrderList &other) const; | |
346 | ||
347 | private: | |
348 | Order *list; | |
349 | int32_t listMax; | |
350 | int32_t listSize; | |
351 | }; | |
352 | ||
353 | OrderList::OrderList() | |
729e4ab9 | 354 | : list(NULL), listMax(16), listSize(0) |
46f4442e A |
355 | { |
356 | list = new Order[listMax]; | |
357 | } | |
358 | ||
359 | OrderList::OrderList(UCollator *coll, const UnicodeString &string, int32_t stringOffset) | |
360 | : list(NULL), listMax(16), listSize(0) | |
361 | { | |
362 | UErrorCode status = U_ZERO_ERROR; | |
363 | UCollationElements *elems = ucol_openElements(coll, string.getBuffer(), string.length(), &status); | |
364 | uint32_t strengthMask = 0; | |
365 | int32_t order, low, high; | |
366 | ||
729e4ab9 | 367 | switch (ucol_getStrength(coll)) |
46f4442e A |
368 | { |
369 | default: | |
370 | strengthMask |= UCOL_TERTIARYORDERMASK; | |
371 | /* fall through */ | |
372 | ||
373 | case UCOL_SECONDARY: | |
374 | strengthMask |= UCOL_SECONDARYORDERMASK; | |
375 | /* fall through */ | |
376 | ||
377 | case UCOL_PRIMARY: | |
378 | strengthMask |= UCOL_PRIMARYORDERMASK; | |
379 | } | |
380 | ||
381 | list = new Order[listMax]; | |
382 | ||
383 | ucol_setOffset(elems, stringOffset, &status); | |
384 | ||
385 | do { | |
386 | low = ucol_getOffset(elems); | |
387 | order = ucol_next(elems, &status); | |
388 | high = ucol_getOffset(elems); | |
389 | ||
390 | if (order != UCOL_NULLORDER) { | |
391 | order &= strengthMask; | |
392 | } | |
393 | ||
394 | if (order != UCOL_IGNORABLE) { | |
395 | add(order, low, high); | |
396 | } | |
397 | } while (order != UCOL_NULLORDER); | |
398 | ||
399 | ucol_closeElements(elems); | |
400 | } | |
401 | ||
402 | OrderList::~OrderList() | |
403 | { | |
404 | delete[] list; | |
405 | } | |
406 | ||
407 | void OrderList::add(int32_t order, int32_t low, int32_t high) | |
408 | { | |
409 | if (listSize >= listMax) { | |
410 | listMax *= 2; | |
411 | ||
412 | Order *newList = new Order[listMax]; | |
413 | ||
414 | uprv_memcpy(newList, list, listSize * sizeof(Order)); | |
415 | delete[] list; | |
416 | list = newList; | |
417 | } | |
418 | ||
419 | list[listSize].order = order; | |
420 | list[listSize].lowOffset = low; | |
421 | list[listSize].highOffset = high; | |
422 | ||
423 | listSize += 1; | |
424 | } | |
425 | ||
426 | const Order *OrderList::get(int32_t index) const | |
427 | { | |
428 | if (index >= listSize) { | |
429 | return NULL; | |
430 | } | |
431 | ||
432 | return &list[index]; | |
433 | } | |
434 | ||
435 | int32_t OrderList::getLowOffset(int32_t index) const | |
436 | { | |
437 | const Order *order = get(index); | |
438 | ||
439 | if (order != NULL) { | |
440 | return order->lowOffset; | |
441 | } | |
442 | ||
443 | return -1; | |
444 | } | |
445 | ||
446 | int32_t OrderList::getHighOffset(int32_t index) const | |
447 | { | |
448 | const Order *order = get(index); | |
449 | ||
450 | if (order != NULL) { | |
451 | return order->highOffset; | |
452 | } | |
453 | ||
454 | return -1; | |
455 | } | |
456 | ||
457 | int32_t OrderList::getOrder(int32_t index) const | |
458 | { | |
459 | const Order *order = get(index); | |
460 | ||
461 | if (order != NULL) { | |
462 | return order->order; | |
463 | } | |
464 | ||
465 | return UCOL_NULLORDER; | |
466 | } | |
467 | ||
468 | int32_t OrderList::size() const | |
469 | { | |
470 | return listSize; | |
471 | } | |
472 | ||
473 | void OrderList::reverse() | |
474 | { | |
475 | for(int32_t f = 0, b = listSize - 1; f < b; f += 1, b -= 1) { | |
476 | Order swap = list[b]; | |
477 | ||
478 | list[b] = list[f]; | |
479 | list[f] = swap; | |
480 | } | |
481 | } | |
482 | ||
483 | UBool OrderList::compare(const OrderList &other) const | |
484 | { | |
485 | if (listSize != other.listSize) { | |
486 | return FALSE; | |
487 | } | |
488 | ||
489 | for(int32_t i = 0; i < listSize; i += 1) { | |
490 | if (list[i].order != other.list[i].order || | |
491 | list[i].lowOffset != other.list[i].lowOffset || | |
492 | list[i].highOffset != other.list[i].highOffset) { | |
493 | return FALSE; | |
494 | } | |
495 | } | |
496 | ||
497 | return TRUE; | |
498 | } | |
499 | ||
500 | UBool OrderList::matchesAt(int32_t offset, const OrderList &other) const | |
501 | { | |
502 | // NOTE: sizes include the NULLORDER, which we don't want to compare. | |
503 | int32_t otherSize = other.size() - 1; | |
504 | ||
505 | if (listSize - 1 - offset < otherSize) { | |
506 | return FALSE; | |
507 | } | |
508 | ||
509 | for (int32_t i = offset, j = 0; j < otherSize; i += 1, j += 1) { | |
510 | if (getOrder(i) != other.getOrder(j)) { | |
511 | return FALSE; | |
512 | } | |
513 | } | |
514 | ||
515 | return TRUE; | |
516 | } | |
517 | ||
518 | static char *printOffsets(char *buffer, OrderList &list) | |
519 | { | |
520 | int32_t size = list.size(); | |
521 | char *s = buffer; | |
522 | ||
523 | for(int32_t i = 0; i < size; i += 1) { | |
524 | const Order *order = list.get(i); | |
525 | ||
526 | if (i != 0) { | |
527 | s += sprintf(s, ", "); | |
528 | } | |
529 | ||
530 | s += sprintf(s, "(%d, %d)", order->lowOffset, order->highOffset); | |
531 | } | |
532 | ||
533 | return buffer; | |
534 | } | |
535 | ||
536 | static char *printOrders(char *buffer, OrderList &list) | |
537 | { | |
538 | int32_t size = list.size(); | |
539 | char *s = buffer; | |
540 | ||
541 | for(int32_t i = 0; i < size; i += 1) { | |
542 | const Order *order = list.get(i); | |
543 | ||
544 | if (i != 0) { | |
545 | s += sprintf(s, ", "); | |
546 | } | |
547 | ||
548 | s += sprintf(s, "%8.8X", order->order); | |
549 | } | |
550 | ||
551 | return buffer; | |
552 | } | |
553 | ||
554 | void SSearchTest::offsetTest() | |
555 | { | |
556 | const char *test[] = { | |
729e4ab9 A |
557 | // The sequence \u0FB3\u0F71\u0F71\u0F80 contains a discontiguous |
558 | // contraction (\u0FB3\u0F71\u0F80) logically followed by \u0F71. | |
559 | "\\u1E33\\u0FB3\\u0F71\\u0F71\\u0F80\\uD835\\uDF6C\\u01B0", | |
560 | ||
46f4442e A |
561 | "\\ua191\\u16ef\\u2036\\u017a", |
562 | ||
563 | #if 0 | |
564 | // This results in a complex interaction between contraction, | |
565 | // expansion and normalization that confuses the backwards offset fixups. | |
566 | "\\u0F7F\\u0F80\\u0F81\\u0F82\\u0F83\\u0F84\\u0F85", | |
567 | #endif | |
568 | ||
569 | "\\u0F80\\u0F81\\u0F82\\u0F83\\u0F84\\u0F85", | |
570 | "\\u07E9\\u07EA\\u07F1\\u07F2\\u07F3", | |
571 | ||
572 | "\\u02FE\\u02FF" | |
573 | "\\u0300\\u0301\\u0302\\u0303\\u0304\\u0305\\u0306\\u0307\\u0308\\u0309\\u030A\\u030B\\u030C\\u030D\\u030E\\u030F" | |
574 | "\\u0310\\u0311\\u0312\\u0313\\u0314\\u0315\\u0316\\u0317\\u0318\\u0319\\u031A\\u031B\\u031C\\u031D\\u031E\\u031F" | |
575 | "\\u0320\\u0321\\u0322\\u0323\\u0324\\u0325\\u0326\\u0327\\u0328\\u0329\\u032A\\u032B\\u032C\\u032D\\u032E\\u032F" | |
576 | "\\u0330\\u0331\\u0332\\u0333\\u0334\\u0335\\u0336\\u0337\\u0338\\u0339\\u033A\\u033B\\u033C\\u033D\\u033E\\u033F" | |
729e4ab9 | 577 | "\\u0340\\u0341\\u0342\\u0343\\u0344\\u0345\\u0346\\u0347\\u0348\\u0349\\u034A\\u034B\\u034C\\u034D\\u034E", // currently not working, see #8081 |
46f4442e | 578 | |
729e4ab9 A |
579 | "\\u02FE\\u02FF\\u0300\\u0301\\u0302\\u0303\\u0316\\u0317\\u0318", // currently not working, see #8081 |
580 | "a\\u02FF\\u0301\\u0316", // currently not working, see #8081 | |
581 | "a\\u02FF\\u0316\\u0301", | |
582 | "a\\u0430\\u0301\\u0316", | |
583 | "a\\u0430\\u0316\\u0301", | |
46f4442e | 584 | "abc\\u0E41\\u0301\\u0316", |
729e4ab9 A |
585 | "abc\\u0E41\\u0316\\u0301", |
586 | "\\u0E41\\u0301\\u0316", | |
587 | "\\u0E41\\u0316\\u0301", | |
588 | "a\\u0301\\u0316", | |
589 | "a\\u0316\\u0301", | |
590 | "\\uAC52\\uAC53", | |
591 | "\\u34CA\\u34CB", | |
592 | "\\u11ED\\u11EE", | |
593 | "\\u30C3\\u30D0", | |
594 | "p\\u00E9ch\\u00E9", | |
46f4442e A |
595 | "a\\u0301\\u0325", |
596 | "a\\u0300\\u0325", | |
597 | "a\\u0325\\u0300", | |
598 | "A\\u0323\\u0300B", | |
599 | "A\\u0300\\u0323B", | |
600 | "A\\u0301\\u0323B", | |
601 | "A\\u0302\\u0301\\u0323B", | |
602 | "abc", | |
603 | "ab\\u0300c", | |
604 | "ab\\u0300\\u0323c", | |
605 | " \\uD800\\uDC00\\uDC00", | |
606 | "a\\uD800\\uDC00\\uDC00", | |
607 | "A\\u0301\\u0301", | |
608 | "A\\u0301\\u0323", | |
609 | "A\\u0301\\u0323B", | |
610 | "B\\u0301\\u0323C", | |
611 | "A\\u0300\\u0323B", | |
612 | "\\u0301A\\u0301\\u0301", | |
613 | "abcd\\r\\u0301", | |
614 | "p\\u00EAche", | |
615 | "pe\\u0302che", | |
616 | }; | |
617 | ||
618 | int32_t testCount = ARRAY_SIZE(test); | |
619 | UErrorCode status = U_ZERO_ERROR; | |
620 | RuleBasedCollator *col = (RuleBasedCollator *) Collator::createInstance(Locale::getEnglish(), status); | |
621 | if (U_FAILURE(status)) { | |
729e4ab9 | 622 | errcheckln(status, "Failed to create collator in offsetTest! - %s", u_errorName(status)); |
46f4442e A |
623 | return; |
624 | } | |
625 | char buffer[4096]; // A bit of a hack... just happens to be long enough for all the test cases... | |
626 | // We could allocate one that's the right size by (CE_count * 10) + 2 | |
627 | // 10 chars is enough room for 8 hex digits plus ", ". 2 extra chars for "[" and "]" | |
628 | ||
629 | col->setAttribute(UCOL_NORMALIZATION_MODE, UCOL_ON, status); | |
630 | ||
631 | for(int32_t i = 0; i < testCount; i += 1) { | |
51004dcb A |
632 | if (!isICUVersionAtLeast(52, 0, 1) && i>=4 && i<=6) { |
633 | continue; // timebomb until ticket #9156 (was #8081) is resolved | |
729e4ab9 | 634 | } |
46f4442e A |
635 | UnicodeString ts = CharsToUnicodeString(test[i]); |
636 | CollationElementIterator *iter = col->createCollationElementIterator(ts); | |
637 | OrderList forwardList; | |
638 | OrderList backwardList; | |
639 | int32_t order, low, high; | |
640 | ||
641 | do { | |
642 | low = iter->getOffset(); | |
643 | order = iter->next(status); | |
644 | high = iter->getOffset(); | |
645 | ||
646 | forwardList.add(order, low, high); | |
647 | } while (order != CollationElementIterator::NULLORDER); | |
648 | ||
649 | iter->reset(); | |
650 | iter->setOffset(ts.length(), status); | |
651 | ||
652 | backwardList.add(CollationElementIterator::NULLORDER, iter->getOffset(), iter->getOffset()); | |
653 | ||
654 | do { | |
655 | high = iter->getOffset(); | |
656 | order = iter->previous(status); | |
657 | low = iter->getOffset(); | |
658 | ||
659 | if (order == CollationElementIterator::NULLORDER) { | |
660 | break; | |
661 | } | |
662 | ||
663 | backwardList.add(order, low, high); | |
664 | } while (TRUE); | |
665 | ||
666 | backwardList.reverse(); | |
667 | ||
668 | if (forwardList.compare(backwardList)) { | |
669 | logln("Works with \"%s\"", test[i]); | |
670 | logln("Forward offsets: [%s]", printOffsets(buffer, forwardList)); | |
671 | // logln("Backward offsets: [%s]", printOffsets(buffer, backwardList)); | |
672 | ||
673 | logln("Forward CEs: [%s]", printOrders(buffer, forwardList)); | |
674 | // logln("Backward CEs: [%s]", printOrders(buffer, backwardList)); | |
675 | ||
676 | logln(); | |
677 | } else { | |
678 | errln("Fails with \"%s\"", test[i]); | |
679 | infoln("Forward offsets: [%s]", printOffsets(buffer, forwardList)); | |
680 | infoln("Backward offsets: [%s]", printOffsets(buffer, backwardList)); | |
681 | ||
682 | infoln("Forward CEs: [%s]", printOrders(buffer, forwardList)); | |
683 | infoln("Backward CEs: [%s]", printOrders(buffer, backwardList)); | |
684 | ||
685 | infoln(); | |
686 | } | |
687 | delete iter; | |
688 | } | |
689 | delete col; | |
690 | } | |
691 | ||
729e4ab9 A |
692 | #if 0 |
693 | static UnicodeString &escape(const UnicodeString &string, UnicodeString &buffer) | |
46f4442e | 694 | { |
729e4ab9 A |
695 | for(int32_t i = 0; i < string.length(); i += 1) { |
696 | UChar32 ch = string.char32At(i); | |
46f4442e | 697 | |
729e4ab9 A |
698 | if (ch >= 0x0020 && ch <= 0x007F) { |
699 | if (ch == 0x005C) { | |
700 | buffer.append("\\\\"); | |
701 | } else { | |
702 | buffer.append(ch); | |
703 | } | |
704 | } else { | |
705 | char cbuffer[12]; | |
46f4442e | 706 | |
729e4ab9 A |
707 | if (ch <= 0xFFFFL) { |
708 | sprintf(cbuffer, "\\u%4.4X", ch); | |
709 | } else { | |
710 | sprintf(cbuffer, "\\U%8.8X", ch); | |
711 | } | |
46f4442e | 712 | |
729e4ab9 | 713 | buffer.append(cbuffer); |
46f4442e A |
714 | } |
715 | ||
729e4ab9 A |
716 | if (ch >= 0x10000L) { |
717 | i += 1; | |
718 | } | |
46f4442e A |
719 | } |
720 | ||
729e4ab9 | 721 | return buffer; |
46f4442e | 722 | } |
729e4ab9 | 723 | #endif |
46f4442e | 724 | |
51004dcb | 725 | void SSearchTest::sharpSTest() |
46f4442e A |
726 | { |
727 | UErrorCode status = U_ZERO_ERROR; | |
729e4ab9 | 728 | UCollator *coll = NULL; |
729e4ab9 A |
729 | UnicodeString lp = "fuss"; |
730 | UnicodeString sp = "fu\\u00DF"; | |
729e4ab9 A |
731 | UnicodeString targets[] = {"fu\\u00DF", "fu\\u00DFball", "1fu\\u00DFball", "12fu\\u00DFball", "123fu\\u00DFball", "1234fu\\u00DFball", |
732 | "ffu\\u00DF", "fufu\\u00DF", "fusfu\\u00DF", | |
733 | "fuss", "ffuss", "fufuss", "fusfuss", "1fuss", "12fuss", "123fuss", "1234fuss", "fu\\u00DF", "1fu\\u00DF", "12fu\\u00DF", "123fu\\u00DF", "1234fu\\u00DF"}; | |
734 | int32_t start = -1, end = -1; | |
735 | ||
736 | coll = ucol_openFromShortString("LEN_S1", FALSE, NULL, &status); | |
51004dcb | 737 | TEST_ASSERT_SUCCESS(status); |
46f4442e | 738 | |
51004dcb A |
739 | UnicodeString lpUnescaped = lp.unescape(); |
740 | UnicodeString spUnescaped = sp.unescape(); | |
729e4ab9 | 741 | |
51004dcb A |
742 | LocalUStringSearchPointer ussLong(usearch_openFromCollator(lpUnescaped.getBuffer(), lpUnescaped.length(), |
743 | lpUnescaped.getBuffer(), lpUnescaped.length(), // actual test data will be set later | |
744 | coll, | |
745 | NULL, // the break iterator | |
746 | &status)); | |
46f4442e | 747 | |
51004dcb A |
748 | LocalUStringSearchPointer ussShort(usearch_openFromCollator(spUnescaped.getBuffer(), spUnescaped.length(), |
749 | spUnescaped.getBuffer(), spUnescaped.length(), // actual test data will be set later | |
750 | coll, | |
751 | NULL, // the break iterator | |
752 | &status)); | |
753 | TEST_ASSERT_SUCCESS(status); | |
46f4442e | 754 | |
729e4ab9 | 755 | for (uint32_t t = 0; t < (sizeof(targets)/sizeof(targets[0])); t += 1) { |
51004dcb | 756 | UBool bFound; |
729e4ab9 | 757 | UnicodeString target = targets[t].unescape(); |
46f4442e | 758 | |
51004dcb A |
759 | start = end = -1; |
760 | usearch_setText(ussLong.getAlias(), target.getBuffer(), target.length(), &status); | |
761 | bFound = usearch_search(ussLong.getAlias(), 0, &start, &end, &status); | |
762 | TEST_ASSERT_SUCCESS(status); | |
763 | if (bFound) { | |
729e4ab9 | 764 | logln("Test %d: found long pattern at [%d, %d].", t, start, end); |
46f4442e | 765 | } else { |
51004dcb | 766 | dataerrln("Test %d: did not find long pattern.", t); |
46f4442e | 767 | } |
46f4442e | 768 | |
51004dcb A |
769 | usearch_setText(ussShort.getAlias(), target.getBuffer(), target.length(), &status); |
770 | bFound = usearch_search(ussShort.getAlias(), 0, &start, &end, &status); | |
771 | TEST_ASSERT_SUCCESS(status); | |
772 | if (bFound) { | |
773 | logln("Test %d: found long pattern at [%d, %d].", t, start, end); | |
46f4442e | 774 | } else { |
51004dcb | 775 | dataerrln("Test %d: did not find long pattern.", t); |
729e4ab9 A |
776 | } |
777 | } | |
46f4442e | 778 | |
729e4ab9 | 779 | ucol_close(coll); |
46f4442e A |
780 | } |
781 | ||
729e4ab9 | 782 | void SSearchTest::goodSuffixTest() |
46f4442e | 783 | { |
729e4ab9 A |
784 | UErrorCode status = U_ZERO_ERROR; |
785 | UCollator *coll = NULL; | |
729e4ab9 A |
786 | UnicodeString pat = /*"gcagagag"*/ "fxeld"; |
787 | UnicodeString target = /*"gcatcgcagagagtatacagtacg"*/ "cloveldfxeld"; | |
729e4ab9 | 788 | int32_t start = -1, end = -1; |
51004dcb | 789 | UBool bFound; |
729e4ab9 A |
790 | |
791 | coll = ucol_open(NULL, &status); | |
51004dcb | 792 | TEST_ASSERT_SUCCESS(status); |
729e4ab9 | 793 | |
51004dcb A |
794 | LocalUStringSearchPointer ss(usearch_openFromCollator(pat.getBuffer(), pat.length(), |
795 | target.getBuffer(), target.length(), | |
796 | coll, | |
797 | NULL, // the break iterator | |
798 | &status)); | |
799 | TEST_ASSERT_SUCCESS(status); | |
46f4442e | 800 | |
51004dcb A |
801 | bFound = usearch_search(ss.getAlias(), 0, &start, &end, &status); |
802 | TEST_ASSERT_SUCCESS(status); | |
803 | if (bFound) { | |
729e4ab9 A |
804 | logln("Found pattern at [%d, %d].", start, end); |
805 | } else { | |
51004dcb | 806 | dataerrln("Did not find pattern."); |
729e4ab9 A |
807 | } |
808 | ||
729e4ab9 | 809 | ucol_close(coll); |
46f4442e A |
810 | } |
811 | ||
729e4ab9 A |
812 | // |
813 | // searchTime() A quick and dirty performance test for string search. | |
814 | // Probably doesn't really belong as part of intltest, but it | |
815 | // does check that the search succeeds, and gets the right result, | |
816 | // so it serves as a functionality test also. | |
817 | // | |
818 | // To run as a perf test, up the loop count, select by commenting | |
819 | // and uncommenting in the code the operation to be measured, | |
820 | // rebuild, and measure the running time of this test alone. | |
821 | // | |
822 | // time LD_LIBRARY_PATH=whatever ./intltest collate/SSearchTest/searchTime | |
823 | // | |
824 | void SSearchTest::searchTime() { | |
825 | static const char *longishText = | |
826 | "Whylom, as olde stories tellen us,\n" | |
827 | "Ther was a duk that highte Theseus:\n" | |
828 | "Of Athenes he was lord and governour,\n" | |
829 | "And in his tyme swich a conquerour,\n" | |
830 | "That gretter was ther noon under the sonne.\n" | |
831 | "Ful many a riche contree hadde he wonne;\n" | |
832 | "What with his wisdom and his chivalrye,\n" | |
833 | "He conquered al the regne of Femenye,\n" | |
834 | "That whylom was y-cleped Scithia;\n" | |
835 | "And weddede the quene Ipolita,\n" | |
836 | "And broghte hir hoom with him in his contree\n" | |
837 | "With muchel glorie and greet solempnitee,\n" | |
838 | "And eek hir yonge suster Emelye.\n" | |
839 | "And thus with victorie and with melodye\n" | |
840 | "Lete I this noble duk to Athenes ryde,\n" | |
841 | "And al his hoost, in armes, him bisyde.\n" | |
842 | "And certes, if it nere to long to here,\n" | |
843 | "I wolde han told yow fully the manere,\n" | |
844 | "How wonnen was the regne of Femenye\n" | |
845 | "By Theseus, and by his chivalrye;\n" | |
846 | "And of the grete bataille for the nones\n" | |
847 | "Bitwixen Athen's and Amazones;\n" | |
848 | "And how asseged was Ipolita,\n" | |
849 | "The faire hardy quene of Scithia;\n" | |
850 | "And of the feste that was at hir weddinge,\n" | |
851 | "And of the tempest at hir hoom-cominge;\n" | |
852 | "But al that thing I moot as now forbere.\n" | |
853 | "I have, God woot, a large feeld to ere,\n" | |
854 | "And wayke been the oxen in my plough.\n" | |
855 | "The remenant of the tale is long y-nough.\n" | |
856 | "I wol nat letten eek noon of this route;\n" | |
857 | "Lat every felawe telle his tale aboute,\n" | |
858 | "And lat see now who shal the soper winne;\n" | |
859 | "And ther I lefte, I wol ageyn biginne.\n" | |
860 | "This duk, of whom I make mencioun,\n" | |
861 | "When he was come almost unto the toun,\n" | |
862 | "In al his wele and in his moste pryde,\n" | |
863 | "He was war, as he caste his eye asyde,\n" | |
864 | "Wher that ther kneled in the hye weye\n" | |
865 | "A companye of ladies, tweye and tweye,\n" | |
866 | "Ech after other, clad in clothes blake; \n" | |
867 | "But swich a cry and swich a wo they make,\n" | |
868 | "That in this world nis creature livinge,\n" | |
869 | "That herde swich another weymentinge;\n" | |
870 | "And of this cry they nolde never stenten,\n" | |
871 | "Til they the reynes of his brydel henten.\n" | |
872 | "'What folk ben ye, that at myn hoomcominge\n" | |
873 | "Perturben so my feste with cryinge'?\n" | |
874 | "Quod Theseus, 'have ye so greet envye\n" | |
875 | "Of myn honour, that thus compleyne and crye? \n" | |
876 | "Or who hath yow misboden, or offended?\n" | |
877 | "And telleth me if it may been amended;\n" | |
878 | "And why that ye ben clothed thus in blak'?\n" | |
879 | "The eldest lady of hem alle spak,\n" | |
880 | "When she hadde swowned with a deedly chere,\n" | |
881 | "That it was routhe for to seen and here,\n" | |
882 | "And seyde: 'Lord, to whom Fortune hath yiven\n" | |
883 | "Victorie, and as a conquerour to liven,\n" | |
884 | "Noght greveth us your glorie and your honour;\n" | |
885 | "But we biseken mercy and socour.\n" | |
886 | "Have mercy on our wo and our distresse.\n" | |
887 | "Som drope of pitee, thurgh thy gentilesse,\n" | |
888 | "Up-on us wrecched wommen lat thou falle.\n" | |
889 | "For certes, lord, ther nis noon of us alle,\n" | |
890 | "That she nath been a duchesse or a quene;\n" | |
891 | "Now be we caitifs, as it is wel sene:\n" | |
892 | "Thanked be Fortune, and hir false wheel,\n" | |
893 | "That noon estat assureth to be weel.\n" | |
894 | "And certes, lord, t'abyden your presence,\n" | |
895 | "Here in the temple of the goddesse Clemence\n" | |
896 | "We han ben waytinge al this fourtenight;\n" | |
897 | "Now help us, lord, sith it is in thy might.\n" | |
898 | "I wrecche, which that wepe and waille thus,\n" | |
899 | "Was whylom wyf to king Capaneus,\n" | |
900 | "That starf at Thebes, cursed be that day!\n" | |
901 | "And alle we, that been in this array,\n" | |
902 | "And maken al this lamentacioun,\n" | |
903 | "We losten alle our housbondes at that toun,\n" | |
904 | "Whyl that the sege ther-aboute lay.\n" | |
905 | "And yet now th'olde Creon, weylaway!\n" | |
906 | "The lord is now of Thebes the citee, \n" | |
907 | "Fulfild of ire and of iniquitee,\n" | |
908 | "He, for despyt, and for his tirannye,\n" | |
909 | "To do the dede bodyes vileinye,\n" | |
910 | "Of alle our lordes, whiche that ben slawe,\n" | |
911 | "Hath alle the bodyes on an heep y-drawe,\n" | |
912 | "And wol nat suffren hem, by noon assent,\n" | |
913 | "Neither to been y-buried nor y-brent,\n" | |
914 | "But maketh houndes ete hem in despyt. zet'\n"; | |
915 | ||
729e4ab9 A |
916 | const char *cPattern = "maketh houndes ete hem"; |
917 | //const char *cPattern = "Whylom"; | |
918 | //const char *cPattern = "zet"; | |
919 | const char *testId = "searchTime()"; // for error macros. | |
920 | UnicodeString target = longishText; | |
921 | UErrorCode status = U_ZERO_ERROR; | |
46f4442e | 922 | |
46f4442e | 923 | |
729e4ab9 | 924 | LocalUCollatorPointer collator(ucol_open("en", &status)); |
729e4ab9 A |
925 | //ucol_setStrength(collator.getAlias(), collatorStrength); |
926 | //ucol_setAttribute(collator.getAlias(), UCOL_NORMALIZATION_MODE, normalize, &status); | |
927 | UnicodeString uPattern = cPattern; | |
729e4ab9 A |
928 | LocalUStringSearchPointer uss(usearch_openFromCollator(uPattern.getBuffer(), uPattern.length(), |
929 | target.getBuffer(), target.length(), | |
930 | collator.getAlias(), | |
931 | NULL, // the break iterator | |
932 | &status)); | |
933 | TEST_ASSERT_SUCCESS(status); | |
46f4442e | 934 | |
729e4ab9 A |
935 | // int32_t foundStart; |
936 | // int32_t foundEnd; | |
937 | UBool found; | |
938 | ||
939 | // Find the match position usgin strstr | |
940 | const char *pm = strstr(longishText, cPattern); | |
941 | TEST_ASSERT_M(pm!=NULL, "No pattern match with strstr"); | |
942 | int32_t refMatchPos = (int32_t)(pm - longishText); | |
943 | int32_t icuMatchPos; | |
944 | int32_t icuMatchEnd; | |
729e4ab9 A |
945 | usearch_search(uss.getAlias(), 0, &icuMatchPos, &icuMatchEnd, &status); |
946 | TEST_ASSERT_SUCCESS(status); | |
729e4ab9 | 947 | TEST_ASSERT_M(refMatchPos == icuMatchPos, "strstr and icu give different match positions."); |
46f4442e | 948 | |
729e4ab9 | 949 | int32_t i; |
4388f060 | 950 | // int32_t j=0; |
729e4ab9 A |
951 | |
952 | // Try loopcounts around 100000 to some millions, depending on the operation, | |
953 | // to get runtimes of at least several seconds. | |
954 | for (i=0; i<10000; i++) { | |
729e4ab9 | 955 | found = usearch_search(uss.getAlias(), 0, &icuMatchPos, &icuMatchEnd, &status); |
729e4ab9 A |
956 | //TEST_ASSERT_SUCCESS(status); |
957 | //TEST_ASSERT(found); | |
958 | ||
959 | // usearch_setOffset(uss.getAlias(), 0, &status); | |
960 | // icuMatchPos = usearch_next(uss.getAlias(), &status); | |
46f4442e | 961 | |
729e4ab9 A |
962 | // The i+j stuff is to confuse the optimizer and get it to actually leave the |
963 | // call to strstr in place. | |
964 | //pm = strstr(longishText+j, cPattern); | |
965 | //j = (j + i)%5; | |
46f4442e | 966 | } |
729e4ab9 | 967 | |
4388f060 | 968 | //printf("%ld, %d\n", pm-longishText, j); |
46f4442e A |
969 | } |
970 | ||
971 | //---------------------------------------------------------------------------------------- | |
972 | // | |
973 | // Random Numbers. Similar to standard lib rand() and srand() | |
974 | // Not using library to | |
975 | // 1. Get same results on all platforms. | |
976 | // 2. Get access to current seed, to more easily reproduce failures. | |
977 | // | |
978 | //--------------------------------------------------------------------------------------- | |
979 | static uint32_t m_seed = 1; | |
980 | ||
981 | static uint32_t m_rand() | |
982 | { | |
983 | m_seed = m_seed * 1103515245 + 12345; | |
984 | return (uint32_t)(m_seed/65536) % 32768; | |
985 | } | |
986 | ||
987 | class Monkey | |
988 | { | |
989 | public: | |
990 | virtual void append(UnicodeString &test, UnicodeString &alternate) = 0; | |
991 | ||
992 | protected: | |
993 | Monkey(); | |
994 | virtual ~Monkey(); | |
995 | }; | |
996 | ||
997 | Monkey::Monkey() | |
998 | { | |
999 | // ook? | |
1000 | } | |
1001 | ||
1002 | Monkey::~Monkey() | |
1003 | { | |
1004 | // ook? | |
1005 | } | |
1006 | ||
1007 | class SetMonkey : public Monkey | |
1008 | { | |
1009 | public: | |
1010 | SetMonkey(const USet *theSet); | |
1011 | ~SetMonkey(); | |
1012 | ||
1013 | virtual void append(UnicodeString &test, UnicodeString &alternate); | |
1014 | ||
1015 | private: | |
1016 | const USet *set; | |
1017 | }; | |
1018 | ||
1019 | SetMonkey::SetMonkey(const USet *theSet) | |
1020 | : Monkey(), set(theSet) | |
1021 | { | |
1022 | // ook? | |
1023 | } | |
1024 | ||
1025 | SetMonkey::~SetMonkey() | |
1026 | { | |
1027 | //ook... | |
1028 | } | |
1029 | ||
1030 | void SetMonkey::append(UnicodeString &test, UnicodeString &alternate) | |
1031 | { | |
1032 | int32_t size = uset_size(set); | |
1033 | int32_t index = m_rand() % size; | |
1034 | UChar32 ch = uset_charAt(set, index); | |
1035 | UnicodeString str(ch); | |
1036 | ||
1037 | test.append(str); | |
1038 | alternate.append(str); // flip case, or some junk? | |
1039 | } | |
1040 | ||
1041 | class StringSetMonkey : public Monkey | |
1042 | { | |
1043 | public: | |
729e4ab9 | 1044 | StringSetMonkey(const USet *theSet, UCollator *theCollator, CollData *theCollData); |
46f4442e A |
1045 | ~StringSetMonkey(); |
1046 | ||
1047 | void append(UnicodeString &testCase, UnicodeString &alternate); | |
1048 | ||
1049 | private: | |
1050 | UnicodeString &generateAlternative(const UnicodeString &testCase, UnicodeString &alternate); | |
1051 | ||
1052 | const USet *set; | |
729e4ab9 A |
1053 | UCollator *coll; |
1054 | CollData *collData; | |
46f4442e A |
1055 | }; |
1056 | ||
729e4ab9 A |
1057 | StringSetMonkey::StringSetMonkey(const USet *theSet, UCollator *theCollator, CollData *theCollData) |
1058 | : Monkey(), set(theSet), coll(theCollator), collData(theCollData) | |
46f4442e A |
1059 | { |
1060 | // ook. | |
1061 | } | |
1062 | ||
1063 | StringSetMonkey::~StringSetMonkey() | |
1064 | { | |
1065 | // ook? | |
1066 | } | |
1067 | ||
1068 | void StringSetMonkey::append(UnicodeString &testCase, UnicodeString &alternate) | |
1069 | { | |
1070 | int32_t itemCount = uset_getItemCount(set), len = 0; | |
1071 | int32_t index = m_rand() % itemCount; | |
1072 | UChar32 rangeStart = 0, rangeEnd = 0; | |
1073 | UChar buffer[16]; | |
1074 | UErrorCode err = U_ZERO_ERROR; | |
1075 | ||
1076 | len = uset_getItem(set, index, &rangeStart, &rangeEnd, buffer, 16, &err); | |
1077 | ||
1078 | if (len == 0) { | |
1079 | int32_t offset = m_rand() % (rangeEnd - rangeStart + 1); | |
1080 | UChar32 ch = rangeStart + offset; | |
1081 | UnicodeString str(ch); | |
1082 | ||
1083 | testCase.append(str); | |
1084 | generateAlternative(str, alternate); | |
1085 | } else if (len > 0) { | |
1086 | // should check that len < 16... | |
1087 | UnicodeString str(buffer, len); | |
1088 | ||
1089 | testCase.append(str); | |
1090 | generateAlternative(str, alternate); | |
1091 | } else { | |
1092 | // shouldn't happen... | |
1093 | } | |
1094 | } | |
1095 | ||
1096 | UnicodeString &StringSetMonkey::generateAlternative(const UnicodeString &testCase, UnicodeString &alternate) | |
1097 | { | |
1098 | // find out shortest string for the longest sequence of ces. | |
1099 | // needs to be refined to use dynamic programming, but will be roughly right | |
729e4ab9 A |
1100 | UErrorCode status = U_ZERO_ERROR; |
1101 | CEList ceList(coll, testCase, status); | |
46f4442e A |
1102 | UnicodeString alt; |
1103 | int32_t offset = 0; | |
1104 | ||
1105 | if (ceList.size() == 0) { | |
1106 | return alternate.append(testCase); | |
1107 | } | |
1108 | ||
1109 | while (offset < ceList.size()) { | |
1110 | int32_t ce = ceList.get(offset); | |
729e4ab9 | 1111 | const StringList *strings = collData->getStringList(ce); |
46f4442e A |
1112 | |
1113 | if (strings == NULL) { | |
1114 | return alternate.append(testCase); | |
1115 | } | |
1116 | ||
1117 | int32_t stringCount = strings->size(); | |
1118 | int32_t tries = 0; | |
729e4ab9 | 1119 | |
46f4442e | 1120 | // find random string that generates the same CEList |
729e4ab9 A |
1121 | const CEList *ceList2 = NULL; |
1122 | const UnicodeString *string = NULL; | |
1123 | UBool matches = FALSE; | |
46f4442e A |
1124 | |
1125 | do { | |
1126 | int32_t s = m_rand() % stringCount; | |
1127 | ||
1128 | if (tries++ > stringCount) { | |
1129 | alternate.append(testCase); | |
1130 | return alternate; | |
1131 | } | |
1132 | ||
1133 | string = strings->get(s); | |
729e4ab9 A |
1134 | ceList2 = collData->getCEList(string); |
1135 | matches = ceList.matchesAt(offset, ceList2); | |
1136 | ||
1137 | if (! matches) { | |
1138 | collData->freeCEList((CEList *) ceList2); | |
1139 | } | |
1140 | } while (! matches); | |
46f4442e A |
1141 | |
1142 | alt.append(*string); | |
1143 | offset += ceList2->size(); | |
729e4ab9 | 1144 | collData->freeCEList(ceList2); |
46f4442e A |
1145 | } |
1146 | ||
729e4ab9 | 1147 | const CEList altCEs(coll, alt, status); |
46f4442e A |
1148 | |
1149 | if (ceList.matchesAt(0, &altCEs)) { | |
1150 | return alternate.append(alt); | |
1151 | } | |
1152 | ||
1153 | return alternate.append(testCase); | |
1154 | } | |
1155 | ||
1156 | static void generateTestCase(UCollator *coll, Monkey *monkeys[], int32_t monkeyCount, UnicodeString &testCase, UnicodeString &alternate) | |
1157 | { | |
1158 | int32_t pieces = (m_rand() % 4) + 1; | |
729e4ab9 | 1159 | UErrorCode status = U_ZERO_ERROR; |
46f4442e A |
1160 | UBool matches; |
1161 | ||
1162 | do { | |
1163 | testCase.remove(); | |
1164 | alternate.remove(); | |
1165 | monkeys[0]->append(testCase, alternate); | |
1166 | ||
1167 | for(int32_t piece = 0; piece < pieces; piece += 1) { | |
1168 | int32_t monkey = m_rand() % monkeyCount; | |
1169 | ||
1170 | monkeys[monkey]->append(testCase, alternate); | |
1171 | } | |
1172 | ||
729e4ab9 A |
1173 | const CEList ceTest(coll, testCase, status); |
1174 | const CEList ceAlt(coll, alternate, status); | |
46f4442e A |
1175 | |
1176 | matches = ceTest.matchesAt(0, &ceAlt); | |
1177 | } while (! matches); | |
1178 | } | |
1179 | ||
46f4442e A |
1180 | static UBool simpleSearch(UCollator *coll, const UnicodeString &target, int32_t offset, const UnicodeString &pattern, int32_t &matchStart, int32_t &matchEnd) |
1181 | { | |
1182 | UErrorCode status = U_ZERO_ERROR; | |
1183 | OrderList targetOrders(coll, target, offset); | |
1184 | OrderList patternOrders(coll, pattern); | |
1185 | int32_t targetSize = targetOrders.size() - 1; | |
1186 | int32_t patternSize = patternOrders.size() - 1; | |
729e4ab9 A |
1187 | UBreakIterator *charBreakIterator = ubrk_open(UBRK_CHARACTER, ucol_getLocaleByType(coll, ULOC_VALID_LOCALE, &status), |
1188 | target.getBuffer(), target.length(), &status); | |
46f4442e A |
1189 | |
1190 | if (patternSize == 0) { | |
729e4ab9 A |
1191 | // Searching for an empty pattern always fails |
1192 | matchStart = matchEnd = -1; | |
1193 | ubrk_close(charBreakIterator); | |
46f4442e A |
1194 | return FALSE; |
1195 | } | |
1196 | ||
1197 | matchStart = matchEnd = -1; | |
1198 | ||
1199 | for(int32_t i = 0; i < targetSize; i += 1) { | |
1200 | if (targetOrders.matchesAt(i, patternOrders)) { | |
1201 | int32_t start = targetOrders.getLowOffset(i); | |
1202 | int32_t maxLimit = targetOrders.getLowOffset(i + patternSize); | |
1203 | int32_t minLimit = targetOrders.getLowOffset(i + patternSize - 1); | |
1204 | ||
1205 | // if the low and high offsets of the first CE in | |
1206 | // the match are the same, it means that the match | |
1207 | // starts in the middle of an expansion - all but | |
1208 | // the first CE of the expansion will have the offset | |
1209 | // of the following character. | |
1210 | if (start == targetOrders.getHighOffset(i)) { | |
1211 | continue; | |
1212 | } | |
1213 | ||
1214 | // Make sure match starts on a grapheme boundary | |
1215 | if (! ubrk_isBoundary(charBreakIterator, start)) { | |
1216 | continue; | |
1217 | } | |
1218 | ||
1219 | // If the low and high offsets of the CE after the match | |
1220 | // are the same, it means that the match ends in the middle | |
1221 | // of an expansion sequence. | |
1222 | if (maxLimit == targetOrders.getHighOffset(i + patternSize) && | |
1223 | targetOrders.getOrder(i + patternSize) != UCOL_NULLORDER) { | |
1224 | continue; | |
1225 | } | |
1226 | ||
1227 | int32_t mend = maxLimit; | |
1228 | ||
1229 | // Find the first grapheme break after the character index | |
1230 | // of the last CE in the match. If it's after character index | |
1231 | // that's after the last CE in the match, use that index | |
1232 | // as the end of the match. | |
1233 | if (minLimit < maxLimit) { | |
4388f060 A |
1234 | // When the last CE's low index is same with its high index, the CE is likely |
1235 | // a part of expansion. In this case, the index is located just after the | |
1236 | // character corresponding to the CEs compared above. If the index is right | |
1237 | // at the break boundary, move the position to the next boundary will result | |
1238 | // incorrect match length when there are ignorable characters exist between | |
1239 | // the position and the next character produces CE(s). See ticket#8482. | |
1240 | if (minLimit == targetOrders.getHighOffset(i + patternSize - 1) && ubrk_isBoundary(charBreakIterator, minLimit)) { | |
1241 | mend = minLimit; | |
1242 | } else { | |
1243 | int32_t nba = ubrk_following(charBreakIterator, minLimit); | |
1244 | ||
1245 | if (nba >= targetOrders.getHighOffset(i + patternSize - 1)) { | |
1246 | mend = nba; | |
1247 | } | |
46f4442e A |
1248 | } |
1249 | } | |
1250 | ||
1251 | if (mend > maxLimit) { | |
1252 | continue; | |
1253 | } | |
1254 | ||
1255 | if (! ubrk_isBoundary(charBreakIterator, mend)) { | |
1256 | continue; | |
1257 | } | |
1258 | ||
1259 | matchStart = start; | |
1260 | matchEnd = mend; | |
1261 | ||
1262 | ubrk_close(charBreakIterator); | |
1263 | return TRUE; | |
1264 | } | |
1265 | } | |
1266 | ||
1267 | ubrk_close(charBreakIterator); | |
1268 | return FALSE; | |
1269 | } | |
1270 | ||
1271 | #if !UCONFIG_NO_REGULAR_EXPRESSIONS | |
1272 | static int32_t getIntParam(UnicodeString name, UnicodeString ¶ms, int32_t defaultVal) { | |
1273 | int32_t val = defaultVal; | |
1274 | ||
1275 | name.append(" *= *(-?\\d+)"); | |
1276 | ||
1277 | UErrorCode status = U_ZERO_ERROR; | |
1278 | RegexMatcher m(name, params, 0, status); | |
1279 | ||
1280 | if (m.find()) { | |
1281 | // The param exists. Convert the string to an int. | |
1282 | char valString[100]; | |
1283 | int32_t paramLength = m.end(1, status) - m.start(1, status); | |
1284 | ||
1285 | if (paramLength >= (int32_t)(sizeof(valString)-1)) { | |
1286 | paramLength = (int32_t)(sizeof(valString)-2); | |
1287 | } | |
1288 | ||
1289 | params.extract(m.start(1, status), paramLength, valString, sizeof(valString)); | |
51004dcb | 1290 | val = uprv_strtol(valString, NULL, 10); |
46f4442e A |
1291 | |
1292 | // Delete this parameter from the params string. | |
1293 | m.reset(); | |
1294 | params = m.replaceFirst("", status); | |
1295 | } | |
1296 | ||
1297 | //U_ASSERT(U_SUCCESS(status)); | |
1298 | if (! U_SUCCESS(status)) { | |
1299 | val = defaultVal; | |
1300 | } | |
1301 | ||
1302 | return val; | |
1303 | } | |
1304 | #endif | |
1305 | ||
1306 | #if !UCONFIG_NO_COLLATION | |
1307 | int32_t SSearchTest::monkeyTestCase(UCollator *coll, const UnicodeString &testCase, const UnicodeString &pattern, const UnicodeString &altPattern, | |
1308 | const char *name, const char *strength, uint32_t seed) | |
1309 | { | |
1310 | UErrorCode status = U_ZERO_ERROR; | |
1311 | int32_t actualStart = -1, actualEnd = -1; | |
1312 | //int32_t expectedStart = prefix.length(), expectedEnd = prefix.length() + altPattern.length(); | |
1313 | int32_t expectedStart = -1, expectedEnd = -1; | |
1314 | int32_t notFoundCount = 0; | |
729e4ab9 A |
1315 | LocalUStringSearchPointer uss(usearch_openFromCollator(pattern.getBuffer(), pattern.length(), |
1316 | testCase.getBuffer(), testCase.length(), | |
1317 | coll, | |
1318 | NULL, // the break iterator | |
1319 | &status)); | |
46f4442e A |
1320 | |
1321 | // **** TODO: find *all* matches, not just first one **** | |
1322 | simpleSearch(coll, testCase, 0, pattern, expectedStart, expectedEnd); | |
1323 | ||
729e4ab9 | 1324 | usearch_search(uss.getAlias(), 0, &actualStart, &actualEnd, &status); |
46f4442e | 1325 | |
729e4ab9 | 1326 | if (expectedStart >= 0 && (actualStart != expectedStart || actualEnd != expectedEnd)) { |
46f4442e A |
1327 | errln("Search for <pattern> in <%s> failed: expected [%d, %d], got [%d, %d]\n" |
1328 | " strength=%s seed=%d", | |
1329 | name, expectedStart, expectedEnd, actualStart, actualEnd, strength, seed); | |
1330 | } | |
1331 | ||
1332 | if (expectedStart == -1 && actualStart == -1) { | |
1333 | notFoundCount += 1; | |
1334 | } | |
1335 | ||
1336 | // **** TODO: find *all* matches, not just first one **** | |
1337 | simpleSearch(coll, testCase, 0, altPattern, expectedStart, expectedEnd); | |
1338 | ||
729e4ab9 | 1339 | usearch_setPattern(uss.getAlias(), altPattern.getBuffer(), altPattern.length(), &status); |
46f4442e | 1340 | |
729e4ab9 | 1341 | usearch_search(uss.getAlias(), 0, &actualStart, &actualEnd, &status); |
46f4442e | 1342 | |
729e4ab9 | 1343 | if (expectedStart >= 0 && (actualStart != expectedStart || actualEnd != expectedEnd)) { |
46f4442e A |
1344 | errln("Search for <alt_pattern> in <%s> failed: expected [%d, %d], got [%d, %d]\n" |
1345 | " strength=%s seed=%d", | |
1346 | name, expectedStart, expectedEnd, actualStart, actualEnd, strength, seed); | |
1347 | } | |
1348 | ||
1349 | if (expectedStart == -1 && actualStart == -1) { | |
1350 | notFoundCount += 1; | |
1351 | } | |
1352 | ||
46f4442e A |
1353 | return notFoundCount; |
1354 | } | |
1355 | #endif | |
1356 | ||
1357 | void SSearchTest::monkeyTest(char *params) | |
1358 | { | |
1359 | // ook! | |
1360 | UErrorCode status = U_ZERO_ERROR; | |
729e4ab9 A |
1361 | //UCollator *coll = ucol_open(NULL, &status); |
1362 | UCollator *coll = ucol_openFromShortString("S1", FALSE, NULL, &status); | |
1363 | ||
46f4442e | 1364 | if (U_FAILURE(status)) { |
729e4ab9 | 1365 | errcheckln(status, "Failed to create collator in MonkeyTest! - %s", u_errorName(status)); |
46f4442e A |
1366 | return; |
1367 | } | |
729e4ab9 | 1368 | |
51004dcb | 1369 | CollData *monkeyData = new CollData(coll, status); |
729e4ab9 | 1370 | |
46f4442e A |
1371 | USet *expansions = uset_openEmpty(); |
1372 | USet *contractions = uset_openEmpty(); | |
46f4442e A |
1373 | |
1374 | ucol_getContractionsAndExpansions(coll, contractions, expansions, FALSE, &status); | |
1375 | ||
46f4442e A |
1376 | U_STRING_DECL(letter_pattern, "[[:letter:]-[:ideographic:]-[:hangul:]]", 39); |
1377 | U_STRING_INIT(letter_pattern, "[[:letter:]-[:ideographic:]-[:hangul:]]", 39); | |
1378 | USet *letters = uset_openPattern(letter_pattern, 39, &status); | |
1379 | SetMonkey letterMonkey(letters); | |
729e4ab9 A |
1380 | StringSetMonkey contractionMonkey(contractions, coll, monkeyData); |
1381 | StringSetMonkey expansionMonkey(expansions, coll, monkeyData); | |
46f4442e A |
1382 | UnicodeString testCase; |
1383 | UnicodeString alternate; | |
1384 | UnicodeString pattern, altPattern; | |
1385 | UnicodeString prefix, altPrefix; | |
1386 | UnicodeString suffix, altSuffix; | |
1387 | ||
1388 | Monkey *monkeys[] = { | |
1389 | &letterMonkey, | |
1390 | &contractionMonkey, | |
1391 | &expansionMonkey, | |
1392 | &contractionMonkey, | |
1393 | &expansionMonkey, | |
1394 | &contractionMonkey, | |
1395 | &expansionMonkey, | |
1396 | &contractionMonkey, | |
1397 | &expansionMonkey}; | |
1398 | int32_t monkeyCount = sizeof(monkeys) / sizeof(monkeys[0]); | |
729e4ab9 | 1399 | // int32_t nonMatchCount = 0; |
46f4442e A |
1400 | |
1401 | UCollationStrength strengths[] = {UCOL_PRIMARY, UCOL_SECONDARY, UCOL_TERTIARY}; | |
1402 | const char *strengthNames[] = {"primary", "secondary", "tertiary"}; | |
1403 | int32_t strengthCount = sizeof(strengths) / sizeof(strengths[0]); | |
1404 | int32_t loopCount = quick? 1000 : 10000; | |
1405 | int32_t firstStrength = 0; | |
729e4ab9 | 1406 | int32_t lastStrength = strengthCount - 1; //*/ 0; |
46f4442e A |
1407 | |
1408 | if (params != NULL) { | |
1409 | #if !UCONFIG_NO_REGULAR_EXPRESSIONS | |
1410 | UnicodeString p(params); | |
1411 | ||
1412 | loopCount = getIntParam("loop", p, loopCount); | |
1413 | m_seed = getIntParam("seed", p, m_seed); | |
1414 | ||
1415 | RegexMatcher m(" *strength *= *(primary|secondary|tertiary) *", p, 0, status); | |
1416 | if (m.find()) { | |
1417 | UnicodeString breakType = m.group(1, status); | |
1418 | ||
1419 | for (int32_t s = 0; s < strengthCount; s += 1) { | |
1420 | if (breakType == strengthNames[s]) { | |
1421 | firstStrength = lastStrength = s; | |
1422 | break; | |
1423 | } | |
1424 | } | |
1425 | ||
1426 | m.reset(); | |
1427 | p = m.replaceFirst("", status); | |
1428 | } | |
1429 | ||
1430 | if (RegexMatcher("\\S", p, 0, status).find()) { | |
1431 | // Each option is stripped out of the option string as it is processed. | |
1432 | // All options have been checked. The option string should have been completely emptied.. | |
1433 | char buf[100]; | |
1434 | p.extract(buf, sizeof(buf), NULL, status); | |
1435 | buf[sizeof(buf)-1] = 0; | |
1436 | errln("Unrecognized or extra parameter: %s\n", buf); | |
1437 | return; | |
1438 | } | |
1439 | #else | |
1440 | infoln("SSearchTest built with UCONFIG_NO_REGULAR_EXPRESSIONS: ignoring parameters."); | |
1441 | #endif | |
1442 | } | |
1443 | ||
1444 | for(int32_t s = firstStrength; s <= lastStrength; s += 1) { | |
1445 | int32_t notFoundCount = 0; | |
1446 | ||
729e4ab9 | 1447 | logln("Setting strength to %s.", strengthNames[s]); |
46f4442e A |
1448 | ucol_setStrength(coll, strengths[s]); |
1449 | ||
1450 | // TODO: try alternate prefix and suffix too? | |
1451 | // TODO: alterntaes are only equal at primary strength. Is this OK? | |
729e4ab9 | 1452 | for(int32_t t = 0; t < loopCount; t += 1) { |
46f4442e | 1453 | uint32_t seed = m_seed; |
729e4ab9 | 1454 | // int32_t nmc = 0; |
46f4442e A |
1455 | |
1456 | generateTestCase(coll, monkeys, monkeyCount, pattern, altPattern); | |
1457 | generateTestCase(coll, monkeys, monkeyCount, prefix, altPrefix); | |
1458 | generateTestCase(coll, monkeys, monkeyCount, suffix, altSuffix); | |
1459 | ||
1460 | // pattern | |
1461 | notFoundCount += monkeyTestCase(coll, pattern, pattern, altPattern, "pattern", strengthNames[s], seed); | |
1462 | ||
1463 | testCase.remove(); | |
1464 | testCase.append(prefix); | |
1465 | testCase.append(/*alt*/pattern); | |
1466 | ||
1467 | // prefix + pattern | |
1468 | notFoundCount += monkeyTestCase(coll, testCase, pattern, altPattern, "prefix + pattern", strengthNames[s], seed); | |
1469 | ||
1470 | testCase.append(suffix); | |
1471 | ||
1472 | // prefix + pattern + suffix | |
1473 | notFoundCount += monkeyTestCase(coll, testCase, pattern, altPattern, "prefix + pattern + suffix", strengthNames[s], seed); | |
1474 | ||
1475 | testCase.remove(); | |
1476 | testCase.append(pattern); | |
1477 | testCase.append(suffix); | |
729e4ab9 | 1478 | |
46f4442e A |
1479 | // pattern + suffix |
1480 | notFoundCount += monkeyTestCase(coll, testCase, pattern, altPattern, "pattern + suffix", strengthNames[s], seed); | |
1481 | } | |
1482 | ||
729e4ab9 A |
1483 | logln("For strength %s the not found count is %d.", strengthNames[s], notFoundCount); |
1484 | } | |
1485 | ||
1486 | uset_close(contractions); | |
1487 | uset_close(expansions); | |
1488 | uset_close(letters); | |
51004dcb | 1489 | delete monkeyData; |
729e4ab9 | 1490 | |
46f4442e A |
1491 | ucol_close(coll); |
1492 | } | |
1493 | ||
729e4ab9 A |
1494 | #endif |
1495 | ||
46f4442e | 1496 | #endif |