1 /*
2 *******************************************************************************
3 * Copyright (C) 2007-2012, International Business Machines Corporation and
4 * others. All Rights Reserved.
5 *******************************************************************************
6 *
7 * File DTPTNGEN.CPP
8 *
9 *******************************************************************************
10 */
11
12 #include "unicode/utypes.h"
13 #if !UCONFIG_NO_FORMATTING
14
15 #include "unicode/datefmt.h"
16 #include "unicode/decimfmt.h"
17 #include "unicode/dtfmtsym.h"
18 #include "unicode/dtptngen.h"
19 #include "unicode/msgfmt.h"
20 #include "unicode/smpdtfmt.h"
21 #include "unicode/udat.h"
22 #include "unicode/udatpg.h"
23 #include "unicode/uniset.h"
24 #include "unicode/uloc.h"
25 #include "unicode/ures.h"
26 #include "unicode/ustring.h"
27 #include "unicode/rep.h"
28 #include "cpputils.h"
29 #include "ucln_in.h"
30 #include "mutex.h"
31 #include "cmemory.h"
32 #include "cstring.h"
33 #include "locbased.h"
34 #include "gregoimp.h"
35 #include "hash.h"
36 #include "uresimp.h"
37 #include "dtptngen_impl.h"
38
39 #define LENGTHOF(array) (int32_t)(sizeof(array)/sizeof((array)[0]))
40
41 #if U_CHARSET_FAMILY==U_EBCDIC_FAMILY
42 /**
43 * If we are on EBCDIC, use an iterator which will
44 * traverse the bundles in ASCII order.
45 */
46 #define U_USE_ASCII_BUNDLE_ITERATOR
47 #define U_SORT_ASCII_BUNDLE_ITERATOR
48 #endif
49
50 #if defined(U_USE_ASCII_BUNDLE_ITERATOR)
51
52 #include "unicode/ustring.h"
53 #include "uarrsort.h"
54
55 struct UResAEntry {
56 UChar *key;
57 UResourceBundle *item;
58 };
59
60 struct UResourceBundleAIterator {
61 UResourceBundle *bund;
62 UResAEntry *entries;
63 int32_t num;
64 int32_t cursor;
65 };
66
67 /* Must be C linkage to pass function pointer to the sort function */
68
69 U_CDECL_BEGIN
70
71 static int32_t U_CALLCONV
ures_a_codepointSort(const void * context,const void * left,const void * right)72 ures_a_codepointSort(const void *context, const void *left, const void *right) {
73 //CompareContext *cmp=(CompareContext *)context;
74 return u_strcmp(((const UResAEntry *)left)->key,
75 ((const UResAEntry *)right)->key);
76 }
77
78 U_CDECL_END
79
ures_a_open(UResourceBundleAIterator * aiter,UResourceBundle * bund,UErrorCode * status)80 static void ures_a_open(UResourceBundleAIterator *aiter, UResourceBundle *bund, UErrorCode *status) {
81 if(U_FAILURE(*status)) {
82 return;
83 }
84 aiter->bund = bund;
85 aiter->num = ures_getSize(aiter->bund);
86 aiter->cursor = 0;
87 #if !defined(U_SORT_ASCII_BUNDLE_ITERATOR)
88 aiter->entries = NULL;
89 #else
90 aiter->entries = (UResAEntry*)uprv_malloc(sizeof(UResAEntry)*aiter->num);
91 for(int i=0;i<aiter->num;i++) {
92 aiter->entries[i].item = ures_getByIndex(aiter->bund, i, NULL, status);
93 const char *akey = ures_getKey(aiter->entries[i].item);
94 int32_t len = uprv_strlen(akey)+1;
95 aiter->entries[i].key = (UChar*)uprv_malloc(len*sizeof(UChar));
96 u_charsToUChars(akey, aiter->entries[i].key, len);
97 }
98 uprv_sortArray(aiter->entries, aiter->num, sizeof(UResAEntry), ures_a_codepointSort, NULL, TRUE, status);
99 #endif
100 }
101
ures_a_close(UResourceBundleAIterator * aiter)102 static void ures_a_close(UResourceBundleAIterator *aiter) {
103 #if defined(U_SORT_ASCII_BUNDLE_ITERATOR)
104 for(int i=0;i<aiter->num;i++) {
105 uprv_free(aiter->entries[i].key);
106 ures_close(aiter->entries[i].item);
107 }
108 #endif
109 }
110
ures_a_getNextString(UResourceBundleAIterator * aiter,int32_t * len,const char ** key,UErrorCode * err)111 static const UChar *ures_a_getNextString(UResourceBundleAIterator *aiter, int32_t *len, const char **key, UErrorCode *err) {
112 #if !defined(U_SORT_ASCII_BUNDLE_ITERATOR)
113 return ures_getNextString(aiter->bund, len, key, err);
114 #else
115 if(U_FAILURE(*err)) return NULL;
116 UResourceBundle *item = aiter->entries[aiter->cursor].item;
117 const UChar* ret = ures_getString(item, len, err);
118 *key = ures_getKey(item);
119 aiter->cursor++;
120 return ret;
121 #endif
122 }
123
124
125 #endif
126
127
128 U_NAMESPACE_BEGIN
129
130
131 // *****************************************************************************
132 // class DateTimePatternGenerator
133 // *****************************************************************************
134 static const UChar Canonical_Items[] = {
135 // GyQMwWEdDFHmsSv
136 CAP_G, LOW_Y, CAP_Q, CAP_M, LOW_W, CAP_W, CAP_E, LOW_D, CAP_D, CAP_F,
137 CAP_H, LOW_M, LOW_S, CAP_S, LOW_V, 0
138 };
139
140 static const dtTypeElem dtTypes[] = {
141 // patternChar, field, type, minLen, weight
142 {CAP_G, UDATPG_ERA_FIELD, DT_SHORT, 1, 3,},
143 {CAP_G, UDATPG_ERA_FIELD, DT_LONG, 4, 0},
144 {LOW_Y, UDATPG_YEAR_FIELD, DT_NUMERIC, 1, 20},
145 {CAP_Y, UDATPG_YEAR_FIELD, DT_NUMERIC + DT_DELTA, 1, 20},
146 {LOW_U, UDATPG_YEAR_FIELD, DT_NUMERIC + 2*DT_DELTA, 1, 20},
147 {CAP_U, UDATPG_YEAR_FIELD, DT_SHORT, 1, 3},
148 {CAP_U, UDATPG_YEAR_FIELD, DT_LONG, 4, 0},
149 {CAP_U, UDATPG_YEAR_FIELD, DT_NARROW, 5, 0},
150 {CAP_Q, UDATPG_QUARTER_FIELD, DT_NUMERIC, 1, 2},
151 {CAP_Q, UDATPG_QUARTER_FIELD, DT_SHORT, 3, 0},
152 {CAP_Q, UDATPG_QUARTER_FIELD, DT_LONG, 4, 0},
153 {LOW_Q, UDATPG_QUARTER_FIELD, DT_NUMERIC + DT_DELTA, 1, 2},
154 {LOW_Q, UDATPG_QUARTER_FIELD, DT_SHORT + DT_DELTA, 3, 0},
155 {LOW_Q, UDATPG_QUARTER_FIELD, DT_LONG + DT_DELTA, 4, 0},
156 {CAP_M, UDATPG_MONTH_FIELD, DT_NUMERIC, 1, 2},
157 {CAP_M, UDATPG_MONTH_FIELD, DT_SHORT, 3, 0},
158 {CAP_M, UDATPG_MONTH_FIELD, DT_LONG, 4, 0},
159 {CAP_M, UDATPG_MONTH_FIELD, DT_NARROW, 5, 0},
160 {CAP_L, UDATPG_MONTH_FIELD, DT_NUMERIC + DT_DELTA, 1, 2},
161 {CAP_L, UDATPG_MONTH_FIELD, DT_SHORT - DT_DELTA, 3, 0},
162 {CAP_L, UDATPG_MONTH_FIELD, DT_LONG - DT_DELTA, 4, 0},
163 {CAP_L, UDATPG_MONTH_FIELD, DT_NARROW - DT_DELTA, 5, 0},
164 {LOW_L, UDATPG_MONTH_FIELD, DT_NUMERIC + DT_DELTA, 1, 1},
165 {LOW_W, UDATPG_WEEK_OF_YEAR_FIELD, DT_NUMERIC, 1, 2},
166 {CAP_W, UDATPG_WEEK_OF_MONTH_FIELD, DT_NUMERIC + DT_DELTA, 1, 0},
167 {CAP_E, UDATPG_WEEKDAY_FIELD, DT_SHORT, 1, 3},
168 {CAP_E, UDATPG_WEEKDAY_FIELD, DT_LONG, 4, 0},
169 {CAP_E, UDATPG_WEEKDAY_FIELD, DT_NARROW, 5, 0},
170 {LOW_C, UDATPG_WEEKDAY_FIELD, DT_NUMERIC + 2*DT_DELTA, 1, 2},
171 {LOW_C, UDATPG_WEEKDAY_FIELD, DT_SHORT - 2*DT_DELTA, 3, 0},
172 {LOW_C, UDATPG_WEEKDAY_FIELD, DT_LONG - 2*DT_DELTA, 4, 0},
173 {LOW_C, UDATPG_WEEKDAY_FIELD, DT_NARROW - 2*DT_DELTA, 5, 0},
174 {LOW_E, UDATPG_WEEKDAY_FIELD, DT_NUMERIC + DT_DELTA, 1, 2}, // LOW_E is currently not used in CLDR data, should not be canonical
175 {LOW_E, UDATPG_WEEKDAY_FIELD, DT_SHORT - DT_DELTA, 3, 0},
176 {LOW_E, UDATPG_WEEKDAY_FIELD, DT_LONG - DT_DELTA, 4, 0},
177 {LOW_E, UDATPG_WEEKDAY_FIELD, DT_NARROW - DT_DELTA, 5, 0},
178 {LOW_D, UDATPG_DAY_FIELD, DT_NUMERIC, 1, 2},
179 {CAP_D, UDATPG_DAY_OF_YEAR_FIELD, DT_NUMERIC + DT_DELTA, 1, 3},
180 {CAP_F, UDATPG_DAY_OF_WEEK_IN_MONTH_FIELD, DT_NUMERIC + 2*DT_DELTA, 1, 0},
181 {LOW_G, UDATPG_DAY_FIELD, DT_NUMERIC + 3*DT_DELTA, 1, 20}, // really internal use, so we don't care
182 {LOW_A, UDATPG_DAYPERIOD_FIELD, DT_SHORT, 1, 0},
183 {CAP_H, UDATPG_HOUR_FIELD, DT_NUMERIC + 10*DT_DELTA, 1, 2}, // 24 hour
184 {LOW_K, UDATPG_HOUR_FIELD, DT_NUMERIC + 11*DT_DELTA, 1, 2},
185 {LOW_H, UDATPG_HOUR_FIELD, DT_NUMERIC, 1, 2}, // 12 hour
186 {LOW_K, UDATPG_HOUR_FIELD, DT_NUMERIC + DT_DELTA, 1, 2},
187 {LOW_M, UDATPG_MINUTE_FIELD, DT_NUMERIC, 1, 2},
188 {LOW_S, UDATPG_SECOND_FIELD, DT_NUMERIC, 1, 2},
189 {CAP_S, UDATPG_FRACTIONAL_SECOND_FIELD, DT_NUMERIC + DT_DELTA, 1, 1000},
190 {CAP_A, UDATPG_SECOND_FIELD, DT_NUMERIC + 2*DT_DELTA, 1, 1000},
191 {LOW_V, UDATPG_ZONE_FIELD, DT_SHORT - 2*DT_DELTA, 1, 0},
192 {LOW_V, UDATPG_ZONE_FIELD, DT_LONG - 2*DT_DELTA, 4, 0},
193 {LOW_Z, UDATPG_ZONE_FIELD, DT_SHORT, 1, 3},
194 {LOW_Z, UDATPG_ZONE_FIELD, DT_LONG, 4, 0},
195 {CAP_Z, UDATPG_ZONE_FIELD, DT_SHORT - DT_DELTA, 1, 3},
196 {CAP_Z, UDATPG_ZONE_FIELD, DT_LONG - DT_DELTA, 4, 0},
197 {CAP_V, UDATPG_ZONE_FIELD, DT_SHORT - DT_DELTA, 1, 3},
198 {CAP_V, UDATPG_ZONE_FIELD, DT_LONG - DT_DELTA, 4, 0},
199 {0, UDATPG_FIELD_COUNT, 0, 0, 0} , // last row of dtTypes[]
200 };
201
202 static const char* const CLDR_FIELD_APPEND[] = {
203 "Era", "Year", "Quarter", "Month", "Week", "*", "Day-Of-Week", "Day", "*", "*", "*",
204 "Hour", "Minute", "Second", "*", "Timezone"
205 };
206
207 static const char* const CLDR_FIELD_NAME[] = {
208 "era", "year", "quarter", "month", "week", "*", "weekday", "day", "*", "*", "dayperiod",
209 "hour", "minute", "second", "*", "zone"
210 };
211
212 static const char* const Resource_Fields[] = {
213 "day", "dayperiod", "era", "hour", "minute", "month", "second", "week",
214 "weekday", "year", "zone", "quarter" };
215
216 // For appendItems
217 static const UChar UDATPG_ItemFormat[]= {0x7B, 0x30, 0x7D, 0x20, 0x251C, 0x7B, 0x32, 0x7D, 0x3A,
218 0x20, 0x7B, 0x31, 0x7D, 0x2524, 0}; // {0} \u251C{2}: {1}\u2524
219
220 //static const UChar repeatedPatterns[6]={CAP_G, CAP_E, LOW_Z, LOW_V, CAP_Q, 0}; // "GEzvQ"
221
222 static const char DT_DateTimePatternsTag[]="DateTimePatterns";
223 static const char DT_DateTimeCalendarTag[]="calendar";
224 static const char DT_DateTimeGregorianTag[]="gregorian";
225 static const char DT_DateTimeAppendItemsTag[]="appendItems";
226 static const char DT_DateTimeFieldsTag[]="fields";
227 static const char DT_DateTimeAvailableFormatsTag[]="availableFormats";
228 //static const UnicodeString repeatedPattern=UnicodeString(repeatedPatterns);
229
230 UOBJECT_DEFINE_RTTI_IMPLEMENTATION(DateTimePatternGenerator)
UOBJECT_DEFINE_RTTI_IMPLEMENTATION(DTSkeletonEnumeration)231 UOBJECT_DEFINE_RTTI_IMPLEMENTATION(DTSkeletonEnumeration)
232 UOBJECT_DEFINE_RTTI_IMPLEMENTATION(DTRedundantEnumeration)
233
234 DateTimePatternGenerator* U_EXPORT2
235 DateTimePatternGenerator::createInstance(UErrorCode& status) {
236 return createInstance(Locale::getDefault(), status);
237 }
238
239 DateTimePatternGenerator* U_EXPORT2
createInstance(const Locale & locale,UErrorCode & status)240 DateTimePatternGenerator::createInstance(const Locale& locale, UErrorCode& status) {
241 DateTimePatternGenerator *result = new DateTimePatternGenerator(locale, status);
242 if (result == NULL) {
243 status = U_MEMORY_ALLOCATION_ERROR;
244 }
245 if (U_FAILURE(status)) {
246 delete result;
247 result = NULL;
248 }
249 return result;
250 }
251
252 DateTimePatternGenerator* U_EXPORT2
createEmptyInstance(UErrorCode & status)253 DateTimePatternGenerator::createEmptyInstance(UErrorCode& status) {
254 DateTimePatternGenerator *result = new DateTimePatternGenerator(status);
255 if (result == NULL) {
256 status = U_MEMORY_ALLOCATION_ERROR;
257 }
258 if (U_FAILURE(status)) {
259 delete result;
260 result = NULL;
261 }
262 return result;
263 }
264
DateTimePatternGenerator(UErrorCode & status)265 DateTimePatternGenerator::DateTimePatternGenerator(UErrorCode &status) :
266 skipMatcher(NULL),
267 fAvailableFormatKeyHash(NULL)
268 {
269 fp = new FormatParser();
270 dtMatcher = new DateTimeMatcher();
271 distanceInfo = new DistanceInfo();
272 patternMap = new PatternMap();
273 if (fp == NULL || dtMatcher == NULL || distanceInfo == NULL || patternMap == NULL) {
274 status = U_MEMORY_ALLOCATION_ERROR;
275 }
276 }
277
DateTimePatternGenerator(const Locale & locale,UErrorCode & status)278 DateTimePatternGenerator::DateTimePatternGenerator(const Locale& locale, UErrorCode &status) :
279 skipMatcher(NULL),
280 fAvailableFormatKeyHash(NULL)
281 {
282 fp = new FormatParser();
283 dtMatcher = new DateTimeMatcher();
284 distanceInfo = new DistanceInfo();
285 patternMap = new PatternMap();
286 if (fp == NULL || dtMatcher == NULL || distanceInfo == NULL || patternMap == NULL) {
287 status = U_MEMORY_ALLOCATION_ERROR;
288 }
289 else {
290 initData(locale, status);
291 }
292 }
293
DateTimePatternGenerator(const DateTimePatternGenerator & other)294 DateTimePatternGenerator::DateTimePatternGenerator(const DateTimePatternGenerator& other) :
295 UObject(),
296 skipMatcher(NULL),
297 fAvailableFormatKeyHash(NULL)
298 {
299 fp = new FormatParser();
300 dtMatcher = new DateTimeMatcher();
301 distanceInfo = new DistanceInfo();
302 patternMap = new PatternMap();
303 *this=other;
304 }
305
306 DateTimePatternGenerator&
operator =(const DateTimePatternGenerator & other)307 DateTimePatternGenerator::operator=(const DateTimePatternGenerator& other) {
308 pLocale = other.pLocale;
309 fDefaultHourFormatChar = other.fDefaultHourFormatChar;
310 *fp = *(other.fp);
311 dtMatcher->copyFrom(other.dtMatcher->skeleton);
312 *distanceInfo = *(other.distanceInfo);
313 dateTimeFormat = other.dateTimeFormat;
314 decimal = other.decimal;
315 // NUL-terminate for the C API.
316 dateTimeFormat.getTerminatedBuffer();
317 decimal.getTerminatedBuffer();
318 delete skipMatcher;
319 if ( other.skipMatcher == NULL ) {
320 skipMatcher = NULL;
321 }
322 else {
323 skipMatcher = new DateTimeMatcher(*other.skipMatcher);
324 }
325 for (int32_t i=0; i< UDATPG_FIELD_COUNT; ++i ) {
326 appendItemFormats[i] = other.appendItemFormats[i];
327 appendItemNames[i] = other.appendItemNames[i];
328 // NUL-terminate for the C API.
329 appendItemFormats[i].getTerminatedBuffer();
330 appendItemNames[i].getTerminatedBuffer();
331 }
332 UErrorCode status = U_ZERO_ERROR;
333 patternMap->copyFrom(*other.patternMap, status);
334 copyHashtable(other.fAvailableFormatKeyHash, status);
335 return *this;
336 }
337
338
339 UBool
operator ==(const DateTimePatternGenerator & other) const340 DateTimePatternGenerator::operator==(const DateTimePatternGenerator& other) const {
341 if (this == &other) {
342 return TRUE;
343 }
344 if ((pLocale==other.pLocale) && (patternMap->equals(*other.patternMap)) &&
345 (dateTimeFormat==other.dateTimeFormat) && (decimal==other.decimal)) {
346 for ( int32_t i=0 ; i<UDATPG_FIELD_COUNT; ++i ) {
347 if ((appendItemFormats[i] != other.appendItemFormats[i]) ||
348 (appendItemNames[i] != other.appendItemNames[i]) ) {
349 return FALSE;
350 }
351 }
352 return TRUE;
353 }
354 else {
355 return FALSE;
356 }
357 }
358
359 UBool
operator !=(const DateTimePatternGenerator & other) const360 DateTimePatternGenerator::operator!=(const DateTimePatternGenerator& other) const {
361 return !operator==(other);
362 }
363
~DateTimePatternGenerator()364 DateTimePatternGenerator::~DateTimePatternGenerator() {
365 if (fAvailableFormatKeyHash!=NULL) {
366 delete fAvailableFormatKeyHash;
367 }
368
369 if (fp != NULL) delete fp;
370 if (dtMatcher != NULL) delete dtMatcher;
371 if (distanceInfo != NULL) delete distanceInfo;
372 if (patternMap != NULL) delete patternMap;
373 if (skipMatcher != NULL) delete skipMatcher;
374 }
375
376 void
initData(const Locale & locale,UErrorCode & status)377 DateTimePatternGenerator::initData(const Locale& locale, UErrorCode &status) {
378 //const char *baseLangName = locale.getBaseName(); // unused
379
380 skipMatcher = NULL;
381 fAvailableFormatKeyHash=NULL;
382 addCanonicalItems();
383 addICUPatterns(locale, status);
384 if (U_FAILURE(status)) {
385 return;
386 }
387 addCLDRData(locale, status);
388 setDateTimeFromCalendar(locale, status);
389 setDecimalSymbols(locale, status);
390 } // DateTimePatternGenerator::initData
391
392 UnicodeString
getSkeleton(const UnicodeString & pattern,UErrorCode &)393 DateTimePatternGenerator::getSkeleton(const UnicodeString& pattern, UErrorCode&
394 /*status*/) {
395 dtMatcher->set(pattern, fp);
396 return dtMatcher->getSkeletonPtr()->getSkeleton();
397 }
398
399 UnicodeString
getBaseSkeleton(const UnicodeString & pattern,UErrorCode &)400 DateTimePatternGenerator::getBaseSkeleton(const UnicodeString& pattern, UErrorCode& /*status*/) {
401 dtMatcher->set(pattern, fp);
402 return dtMatcher->getSkeletonPtr()->getBaseSkeleton();
403 }
404
405 void
addICUPatterns(const Locale & locale,UErrorCode & status)406 DateTimePatternGenerator::addICUPatterns(const Locale& locale, UErrorCode& status) {
407 UnicodeString dfPattern;
408 UnicodeString conflictingString;
409 DateFormat* df;
410
411 if (U_FAILURE(status)) {
412 return;
413 }
414
415 // Load with ICU patterns
416 for (int32_t i=DateFormat::kFull; i<=DateFormat::kShort; i++) {
417 DateFormat::EStyle style = (DateFormat::EStyle)i;
418 df = DateFormat::createDateInstance(style, locale);
419 SimpleDateFormat* sdf;
420 if (df != NULL && (sdf = dynamic_cast<SimpleDateFormat*>(df)) != NULL) {
421 addPattern(sdf->toPattern(dfPattern), FALSE, conflictingString, status);
422 }
423 // TODO Maybe we should return an error when the date format isn't simple.
424 delete df;
425 if (U_FAILURE(status)) {
426 return;
427 }
428
429 df = DateFormat::createTimeInstance(style, locale);
430 if (df != NULL && (sdf = dynamic_cast<SimpleDateFormat*>(df)) != NULL) {
431 addPattern(sdf->toPattern(dfPattern), FALSE, conflictingString, status);
432 // HACK for hh:ss
433 if ( i==DateFormat::kMedium ) {
434 hackPattern = dfPattern;
435 }
436 }
437 // TODO Maybe we should return an error when the date format isn't simple.
438 delete df;
439 if (U_FAILURE(status)) {
440 return;
441 }
442 }
443 }
444
445 void
hackTimes(const UnicodeString & hackPattern,UErrorCode & status)446 DateTimePatternGenerator::hackTimes(const UnicodeString& hackPattern, UErrorCode& status) {
447 UnicodeString conflictingString;
448
449 fp->set(hackPattern);
450 UnicodeString mmss;
451 UBool gotMm=FALSE;
452 for (int32_t i=0; i<fp->itemNumber; ++i) {
453 UnicodeString field = fp->items[i];
454 if ( fp->isQuoteLiteral(field) ) {
455 if ( gotMm ) {
456 UnicodeString quoteLiteral;
457 fp->getQuoteLiteral(quoteLiteral, &i);
458 mmss += quoteLiteral;
459 }
460 }
461 else {
462 if (fp->isPatternSeparator(field) && gotMm) {
463 mmss+=field;
464 }
465 else {
466 UChar ch=field.charAt(0);
467 if (ch==LOW_M) {
468 gotMm=TRUE;
469 mmss+=field;
470 }
471 else {
472 if (ch==LOW_S) {
473 if (!gotMm) {
474 break;
475 }
476 mmss+= field;
477 addPattern(mmss, FALSE, conflictingString, status);
478 break;
479 }
480 else {
481 if (gotMm || ch==LOW_Z || ch==CAP_Z || ch==LOW_V || ch==CAP_V) {
482 break;
483 }
484 }
485 }
486 }
487 }
488 }
489 }
490
491 #define ULOC_LOCALE_IDENTIFIER_CAPACITY (ULOC_FULLNAME_CAPACITY + 1 + ULOC_KEYWORD_AND_VALUES_CAPACITY)
492
493 static const UChar hourFormatChars[] = { CAP_H, LOW_H, CAP_K, LOW_K, 0 }; // HhKk, the hour format characters
494
495 void
addCLDRData(const Locale & locale,UErrorCode & err)496 DateTimePatternGenerator::addCLDRData(const Locale& locale, UErrorCode& err) {
497 UResourceBundle *rb, *calTypeBundle, *calBundle;
498 UResourceBundle *patBundle, *fieldBundle, *fBundle;
499 UnicodeString rbPattern, value, field;
500 UnicodeString conflictingPattern;
501 const char *key=NULL;
502 int32_t i;
503
504 UnicodeString defaultItemFormat(TRUE, UDATPG_ItemFormat, LENGTHOF(UDATPG_ItemFormat)-1); // Read-only alias.
505
506 err = U_ZERO_ERROR;
507
508 fDefaultHourFormatChar = 0;
509 for (i=0; i<UDATPG_FIELD_COUNT; ++i ) {
510 appendItemNames[i]=CAP_F;
511 if (i<10) {
512 appendItemNames[i]+=(UChar)(i+0x30);
513 }
514 else {
515 appendItemNames[i]+=(UChar)0x31;
516 appendItemNames[i]+=(UChar)(i-10 + 0x30);
517 }
518 // NUL-terminate for the C API.
519 appendItemNames[i].getTerminatedBuffer();
520 }
521
522 rb = ures_open(NULL, locale.getName(), &err);
523 if (rb == NULL || U_FAILURE(err)) {
524 return;
525 }
526 const char *curLocaleName=ures_getLocaleByType(rb, ULOC_ACTUAL_LOCALE, &err);
527 const char * calendarTypeToUse = DT_DateTimeGregorianTag; // initial default
528 char calendarType[ULOC_KEYWORDS_CAPACITY]; // to be filled in with the type to use, if all goes well
529 if ( U_SUCCESS(err) ) {
530 char localeWithCalendarKey[ULOC_LOCALE_IDENTIFIER_CAPACITY];
531 // obtain a locale that always has the calendar key value that should be used
532 (void)ures_getFunctionalEquivalent(localeWithCalendarKey, ULOC_LOCALE_IDENTIFIER_CAPACITY, NULL,
533 "calendar", "calendar", locale.getName(), NULL, FALSE, &err);
534 localeWithCalendarKey[ULOC_LOCALE_IDENTIFIER_CAPACITY-1] = 0; // ensure null termination
535 // now get the calendar key value from that locale
536 int32_t calendarTypeLen = uloc_getKeywordValue(localeWithCalendarKey, "calendar", calendarType, ULOC_KEYWORDS_CAPACITY, &err);
537 if (U_SUCCESS(err) && calendarTypeLen < ULOC_KEYWORDS_CAPACITY) {
538 calendarTypeToUse = calendarType;
539 }
540 err = U_ZERO_ERROR;
541 }
542 calBundle = ures_getByKeyWithFallback(rb, DT_DateTimeCalendarTag, NULL, &err);
543 calTypeBundle = ures_getByKeyWithFallback(calBundle, calendarTypeToUse, NULL, &err);
544
545 key=NULL;
546 int32_t dtCount=0;
547 patBundle = ures_getByKeyWithFallback(calTypeBundle, DT_DateTimePatternsTag, NULL, &err);
548 while (U_SUCCESS(err)) {
549 rbPattern = ures_getNextUnicodeString(patBundle, &key, &err);
550 dtCount++;
551 if (rbPattern.length()==0 ) {
552 break; // no more pattern
553 }
554 else {
555 if (dtCount==9) {
556 setDateTimeFormat(rbPattern);
557 } else if (dtCount==4) { // short time format
558 // set fDefaultHourFormatChar to the hour format character from this pattern
559 int32_t tfIdx, tfLen = rbPattern.length();
560 UBool ignoreChars = FALSE;
561 for (tfIdx = 0; tfIdx < tfLen; tfIdx++) {
562 UChar tfChar = rbPattern.charAt(tfIdx);
563 if ( tfChar == SINGLE_QUOTE ) {
564 ignoreChars = !ignoreChars; // toggle (handle quoted literals & '' for single quote)
565 } else if ( !ignoreChars && u_strchr(hourFormatChars, tfChar) != NULL ) {
566 fDefaultHourFormatChar = tfChar;
567 break;
568 }
569 }
570 }
571 }
572 }
573 ures_close(patBundle);
574
575 err = U_ZERO_ERROR;
576 patBundle = ures_getByKeyWithFallback(calTypeBundle, DT_DateTimeAppendItemsTag, NULL, &err);
577 key=NULL;
578 UnicodeString itemKey;
579 while (U_SUCCESS(err)) {
580 rbPattern = ures_getNextUnicodeString(patBundle, &key, &err);
581 if (rbPattern.length()==0 ) {
582 break; // no more pattern
583 }
584 else {
585 setAppendItemFormat(getAppendFormatNumber(key), rbPattern);
586 }
587 }
588 ures_close(patBundle);
589
590 key=NULL;
591 err = U_ZERO_ERROR;
592 fBundle = ures_getByKeyWithFallback(calTypeBundle, DT_DateTimeFieldsTag, NULL, &err);
593 for (i=0; i<MAX_RESOURCE_FIELD; ++i) {
594 err = U_ZERO_ERROR;
595 patBundle = ures_getByKeyWithFallback(fBundle, Resource_Fields[i], NULL, &err);
596 fieldBundle = ures_getByKeyWithFallback(patBundle, "dn", NULL, &err);
597 rbPattern = ures_getNextUnicodeString(fieldBundle, &key, &err);
598 ures_close(fieldBundle);
599 ures_close(patBundle);
600 if (rbPattern.length()==0 ) {
601 continue;
602 }
603 else {
604 setAppendItemName(getAppendNameNumber(Resource_Fields[i]), rbPattern);
605 }
606 }
607 ures_close(fBundle);
608
609 // add available formats
610 UBool firstTimeThrough = TRUE;
611 err = U_ZERO_ERROR;
612 initHashtable(err);
613 while (TRUE) {
614 // At the start of the loop:
615 // - rb is the open resource bundle for the current locale being processed,
616 // whose actual name is in curLocaleName.
617 // - if U_SUCCESS(err), then calBundle and calTypeBundle are open;
618 // process contents of calTypeBundle, then close calBundle and calTypeBundle.
619 if (U_SUCCESS(err)) {
620 // process contents of calTypeBundle
621 patBundle = ures_getByKeyWithFallback(calTypeBundle, DT_DateTimeAvailableFormatsTag, NULL, &err);
622 if (U_SUCCESS(err)) {
623 int32_t numberKeys = ures_getSize(patBundle);
624 int32_t len;
625 const UChar *retPattern;
626 key=NULL;
627 #if defined(U_USE_ASCII_BUNDLE_ITERATOR)
628 UResourceBundleAIterator aiter;
629 ures_a_open(&aiter, patBundle, &err);
630 #endif
631 for(i=0; i<numberKeys; ++i) {
632 #if defined(U_USE_ASCII_BUNDLE_ITERATOR)
633 retPattern=ures_a_getNextString(&aiter, &len, &key, &err);
634 #else
635 retPattern=ures_getNextString(patBundle, &len, &key, &err);
636 #endif
637 UnicodeString format=UnicodeString(retPattern);
638 UnicodeString retKey=UnicodeString(key, -1, US_INV);
639 if ( firstTimeThrough || !isAvailableFormatSet(retKey) ) {
640 setAvailableFormat(retKey, err);
641 // Add pattern with its associated skeleton. Override any duplicate derived from std patterns,
642 // but not a previous availableFormats entry:
643 addPatternWithSkeleton(format, &retKey, TRUE, conflictingPattern, err);
644 }
645 }
646 #if defined(U_USE_ASCII_BUNDLE_ITERATOR)
647 ures_a_close(&aiter);
648 #endif
649 ures_close(patBundle);
650 }
651 firstTimeThrough = FALSE;
652 // close calBundle and calTypeBundle
653 ures_close(calTypeBundle);
654 ures_close(calBundle);
655 }
656 if (uprv_strcmp(curLocaleName,"root")==0 || uprv_strlen(curLocaleName)==0) {
657 // we just finished handling root, nothing more to check
658 ures_close(rb);
659 break;
660 }
661 // Find the name of the appropriate parent locale (from %%Parent if present, else
662 // uloc_getParent on the actual locale name)
663 // (It would be nice to have a ures function that did this...)
664 err = U_ZERO_ERROR;
665 char parentLocale[ULOC_FULLNAME_CAPACITY];
666 int32_t locNameLen;
667 const UChar * parentUName = ures_getStringByKey(rb, "%%Parent", &locNameLen, &err);
668 if (U_SUCCESS(err) && err != U_USING_FALLBACK_WARNING && locNameLen < ULOC_FULLNAME_CAPACITY) {
669 u_UCharsToChars(parentUName, parentLocale, locNameLen + 1);
670 } else {
671 err = U_ZERO_ERROR;
672 uloc_getParent(curLocaleName, parentLocale, ULOC_FULLNAME_CAPACITY, &err);
673 if (U_FAILURE(err) || err == U_STRING_NOT_TERMINATED_WARNING) {
674 // just fallback to root, since we are not already there
675 parentLocale[0] = 0;
676 err = U_ZERO_ERROR;
677 }
678 }
679 // Close current locale bundle
680 ures_close(rb);
681 // And open its parent, which becomes the new current locale being processed
682 rb = ures_open(NULL, parentLocale, &err);
683 if ( U_FAILURE(err) ) {
684 err = U_ZERO_ERROR;
685 break;
686 }
687 // Get the name of the parent / new current locale
688 curLocaleName=ures_getLocaleByType(rb, ULOC_ACTUAL_LOCALE, &err);
689 if ( U_FAILURE(err) ) {
690 curLocaleName = parentLocale;
691 err = U_ZERO_ERROR;
692 }
693 // Open calBundle and calTypeBundle
694 calBundle = ures_getByKeyWithFallback(rb, DT_DateTimeCalendarTag, NULL, &err);
695 if (U_SUCCESS(err)) {
696 calTypeBundle = ures_getByKeyWithFallback(calBundle, calendarTypeToUse, NULL, &err);
697 if ( U_FAILURE(err) ) {
698 ures_close(calBundle);
699 }
700 }
701 // Go to the top of the loop to process contents of calTypeBundle
702 }
703
704 if (hackPattern.length()>0) {
705 hackTimes(hackPattern, err);
706 }
707 }
708
709 void
initHashtable(UErrorCode & err)710 DateTimePatternGenerator::initHashtable(UErrorCode& err) {
711 if (fAvailableFormatKeyHash!=NULL) {
712 return;
713 }
714 if ((fAvailableFormatKeyHash = new Hashtable(FALSE, err))==NULL) {
715 err=U_MEMORY_ALLOCATION_ERROR;
716 return;
717 }
718 }
719
720
721 void
setAppendItemFormat(UDateTimePatternField field,const UnicodeString & value)722 DateTimePatternGenerator::setAppendItemFormat(UDateTimePatternField field, const UnicodeString& value) {
723 appendItemFormats[field] = value;
724 // NUL-terminate for the C API.
725 appendItemFormats[field].getTerminatedBuffer();
726 }
727
728 const UnicodeString&
getAppendItemFormat(UDateTimePatternField field) const729 DateTimePatternGenerator::getAppendItemFormat(UDateTimePatternField field) const {
730 return appendItemFormats[field];
731 }
732
733 void
setAppendItemName(UDateTimePatternField field,const UnicodeString & value)734 DateTimePatternGenerator::setAppendItemName(UDateTimePatternField field, const UnicodeString& value) {
735 appendItemNames[field] = value;
736 // NUL-terminate for the C API.
737 appendItemNames[field].getTerminatedBuffer();
738 }
739
740 const UnicodeString&
getAppendItemName(UDateTimePatternField field) const741 DateTimePatternGenerator:: getAppendItemName(UDateTimePatternField field) const {
742 return appendItemNames[field];
743 }
744
745 void
getAppendName(UDateTimePatternField field,UnicodeString & value)746 DateTimePatternGenerator::getAppendName(UDateTimePatternField field, UnicodeString& value) {
747 value = SINGLE_QUOTE;
748 value += appendItemNames[field];
749 value += SINGLE_QUOTE;
750 }
751
752 UnicodeString
getBestPattern(const UnicodeString & patternForm,UErrorCode & status)753 DateTimePatternGenerator::getBestPattern(const UnicodeString& patternForm, UErrorCode& status) {
754 return getBestPattern(patternForm, UDATPG_MATCH_NO_OPTIONS, status);
755 }
756
757 UnicodeString
getBestPattern(const UnicodeString & patternForm,UDateTimePatternMatchOptions options,UErrorCode & status)758 DateTimePatternGenerator::getBestPattern(const UnicodeString& patternForm, UDateTimePatternMatchOptions options, UErrorCode& status) {
759 const UnicodeString *bestPattern=NULL;
760 UnicodeString dtFormat;
761 UnicodeString resultPattern;
762
763 int32_t dateMask=(1<<UDATPG_DAYPERIOD_FIELD) - 1;
764 int32_t timeMask=(1<<UDATPG_FIELD_COUNT) - 1 - dateMask;
765
766 UnicodeString patternFormCopy = UnicodeString(patternForm);
767 patternFormCopy.findAndReplace(UnicodeString(LOW_J), UnicodeString(fDefaultHourFormatChar));
768
769 resultPattern.remove();
770 dtMatcher->set(patternFormCopy, fp);
771 const PtnSkeleton* specifiedSkeleton=NULL;
772 bestPattern=getBestRaw(*dtMatcher, -1, distanceInfo, &specifiedSkeleton);
773 if ( distanceInfo->missingFieldMask==0 && distanceInfo->extraFieldMask==0 ) {
774 resultPattern = adjustFieldTypes(*bestPattern, specifiedSkeleton, FALSE, options);
775
776 return resultPattern;
777 }
778 int32_t neededFields = dtMatcher->getFieldMask();
779 UnicodeString datePattern=getBestAppending(neededFields & dateMask, options);
780 UnicodeString timePattern=getBestAppending(neededFields & timeMask, options);
781 if (datePattern.length()==0) {
782 if (timePattern.length()==0) {
783 resultPattern.remove();
784 }
785 else {
786 return timePattern;
787 }
788 }
789 if (timePattern.length()==0) {
790 return datePattern;
791 }
792 resultPattern.remove();
793 status = U_ZERO_ERROR;
794 dtFormat=getDateTimeFormat();
795 Formattable dateTimeObject[] = { timePattern, datePattern };
796 resultPattern = MessageFormat::format(dtFormat, dateTimeObject, 2, resultPattern, status );
797 return resultPattern;
798 }
799
800 UnicodeString
replaceFieldTypes(const UnicodeString & pattern,const UnicodeString & skeleton,UErrorCode & status)801 DateTimePatternGenerator::replaceFieldTypes(const UnicodeString& pattern,
802 const UnicodeString& skeleton,
803 UErrorCode& status) {
804 return replaceFieldTypes(pattern, skeleton, UDATPG_MATCH_NO_OPTIONS, status);
805 }
806
807 UnicodeString
replaceFieldTypes(const UnicodeString & pattern,const UnicodeString & skeleton,UDateTimePatternMatchOptions options,UErrorCode &)808 DateTimePatternGenerator::replaceFieldTypes(const UnicodeString& pattern,
809 const UnicodeString& skeleton,
810 UDateTimePatternMatchOptions options,
811 UErrorCode& /*status*/) {
812 dtMatcher->set(skeleton, fp);
813 UnicodeString result = adjustFieldTypes(pattern, NULL, FALSE, options);
814 return result;
815 }
816
817 void
setDecimal(const UnicodeString & newDecimal)818 DateTimePatternGenerator::setDecimal(const UnicodeString& newDecimal) {
819 this->decimal = newDecimal;
820 // NUL-terminate for the C API.
821 this->decimal.getTerminatedBuffer();
822 }
823
824 const UnicodeString&
getDecimal() const825 DateTimePatternGenerator::getDecimal() const {
826 return decimal;
827 }
828
829 void
addCanonicalItems()830 DateTimePatternGenerator::addCanonicalItems() {
831 UnicodeString conflictingPattern;
832 UErrorCode status = U_ZERO_ERROR;
833
834 for (int32_t i=0; i<UDATPG_FIELD_COUNT; i++) {
835 addPattern(UnicodeString(Canonical_Items[i]), FALSE, conflictingPattern, status);
836 }
837 }
838
839 void
setDateTimeFormat(const UnicodeString & dtFormat)840 DateTimePatternGenerator::setDateTimeFormat(const UnicodeString& dtFormat) {
841 dateTimeFormat = dtFormat;
842 // NUL-terminate for the C API.
843 dateTimeFormat.getTerminatedBuffer();
844 }
845
846 const UnicodeString&
getDateTimeFormat() const847 DateTimePatternGenerator::getDateTimeFormat() const {
848 return dateTimeFormat;
849 }
850
851 void
setDateTimeFromCalendar(const Locale & locale,UErrorCode & status)852 DateTimePatternGenerator::setDateTimeFromCalendar(const Locale& locale, UErrorCode& status) {
853 const UChar *resStr;
854 int32_t resStrLen = 0;
855
856 Calendar* fCalendar = Calendar::createInstance(locale, status);
857 CalendarData calData(locale, fCalendar?fCalendar->getType():NULL, status);
858 UResourceBundle *dateTimePatterns = calData.getByKey(DT_DateTimePatternsTag, status);
859 if (U_FAILURE(status)) return;
860
861 if (ures_getSize(dateTimePatterns) <= DateFormat::kDateTime)
862 {
863 status = U_INVALID_FORMAT_ERROR;
864 return;
865 }
866 resStr = ures_getStringByIndex(dateTimePatterns, (int32_t)DateFormat::kDateTime, &resStrLen, &status);
867 setDateTimeFormat(UnicodeString(TRUE, resStr, resStrLen));
868
869 delete fCalendar;
870 }
871
872 void
setDecimalSymbols(const Locale & locale,UErrorCode & status)873 DateTimePatternGenerator::setDecimalSymbols(const Locale& locale, UErrorCode& status) {
874 DecimalFormatSymbols dfs = DecimalFormatSymbols(locale, status);
875 if(U_SUCCESS(status)) {
876 decimal = dfs.getSymbol(DecimalFormatSymbols::kDecimalSeparatorSymbol);
877 // NUL-terminate for the C API.
878 decimal.getTerminatedBuffer();
879 }
880 }
881
882 UDateTimePatternConflict
addPattern(const UnicodeString & pattern,UBool override,UnicodeString & conflictingPattern,UErrorCode & status)883 DateTimePatternGenerator::addPattern(
884 const UnicodeString& pattern,
885 UBool override,
886 UnicodeString &conflictingPattern,
887 UErrorCode& status)
888 {
889 return addPatternWithSkeleton(pattern, NULL, override, conflictingPattern, status);
890 }
891
892 // For DateTimePatternGenerator::addPatternWithSkeleton -
893 // If skeletonToUse is specified, then an availableFormats entry is being added. In this case:
894 // 1. We pass that skeleton to matcher.set instead of having it derive a skeleton from the pattern.
895 // 2. If the new entry's skeleton or basePattern does match an existing entry but that entry also had a skeleton specified
896 // (i.e. it was also from availableFormats), then the new entry does not override it regardless of the value of the override
897 // parameter. This prevents later availableFormats entries from a parent locale overriding earlier ones from the actual
898 // specified locale. However, availableFormats entries *should* override entries with matching skeleton whose skeleton was
899 // derived (i.e. entries derived from the standard date/time patters for the specified locale).
900 // 3. When adding the pattern (patternMap->add), we set a new boolean to indicate that the added entry had a
901 // specified skeleton (which sets a new field in the PtnElem in the PatternMap).
902 UDateTimePatternConflict
addPatternWithSkeleton(const UnicodeString & pattern,const UnicodeString * skeletonToUse,UBool override,UnicodeString & conflictingPattern,UErrorCode & status)903 DateTimePatternGenerator::addPatternWithSkeleton(
904 const UnicodeString& pattern,
905 const UnicodeString* skeletonToUse,
906 UBool override,
907 UnicodeString& conflictingPattern,
908 UErrorCode& status)
909 {
910
911 UnicodeString basePattern;
912 PtnSkeleton skeleton;
913 UDateTimePatternConflict conflictingStatus = UDATPG_NO_CONFLICT;
914
915 DateTimeMatcher matcher;
916 if ( skeletonToUse == NULL ) {
917 matcher.set(pattern, fp, skeleton);
918 matcher.getBasePattern(basePattern);
919 } else {
920 matcher.set(*skeletonToUse, fp, skeleton); // no longer trims skeleton fields to max len 3, per #7930
921 matcher.getBasePattern(basePattern); // or perhaps instead: basePattern = *skeletonToUse;
922 }
923 UBool entryHadSpecifiedSkeleton;
924 const UnicodeString *duplicatePattern = patternMap->getPatternFromBasePattern(basePattern, entryHadSpecifiedSkeleton);
925 if (duplicatePattern != NULL ) {
926 conflictingStatus = UDATPG_BASE_CONFLICT;
927 conflictingPattern = *duplicatePattern;
928 if (!override || (skeletonToUse != NULL && entryHadSpecifiedSkeleton)) {
929 return conflictingStatus;
930 }
931 }
932 const PtnSkeleton* entrySpecifiedSkeleton = NULL;
933 duplicatePattern = patternMap->getPatternFromSkeleton(skeleton, &entrySpecifiedSkeleton);
934 if (duplicatePattern != NULL ) {
935 conflictingStatus = UDATPG_CONFLICT;
936 conflictingPattern = *duplicatePattern;
937 if (!override || (skeletonToUse != NULL && entrySpecifiedSkeleton != NULL)) {
938 return conflictingStatus;
939 }
940 }
941 patternMap->add(basePattern, skeleton, pattern, skeletonToUse != NULL, status);
942 if(U_FAILURE(status)) {
943 return conflictingStatus;
944 }
945
946 return UDATPG_NO_CONFLICT;
947 }
948
949
950 UDateTimePatternField
getAppendFormatNumber(const char * field) const951 DateTimePatternGenerator::getAppendFormatNumber(const char* field) const {
952 for (int32_t i=0; i<UDATPG_FIELD_COUNT; ++i ) {
953 if (uprv_strcmp(CLDR_FIELD_APPEND[i], field)==0) {
954 return (UDateTimePatternField)i;
955 }
956 }
957 return UDATPG_FIELD_COUNT;
958 }
959
960 UDateTimePatternField
getAppendNameNumber(const char * field) const961 DateTimePatternGenerator::getAppendNameNumber(const char* field) const {
962 for (int32_t i=0; i<UDATPG_FIELD_COUNT; ++i ) {
963 if (uprv_strcmp(CLDR_FIELD_NAME[i],field)==0) {
964 return (UDateTimePatternField)i;
965 }
966 }
967 return UDATPG_FIELD_COUNT;
968 }
969
970 const UnicodeString*
getBestRaw(DateTimeMatcher & source,int32_t includeMask,DistanceInfo * missingFields,const PtnSkeleton ** specifiedSkeletonPtr)971 DateTimePatternGenerator::getBestRaw(DateTimeMatcher& source,
972 int32_t includeMask,
973 DistanceInfo* missingFields,
974 const PtnSkeleton** specifiedSkeletonPtr) {
975 int32_t bestDistance = 0x7fffffff;
976 DistanceInfo tempInfo;
977 const UnicodeString *bestPattern=NULL;
978 const PtnSkeleton* specifiedSkeleton=NULL;
979
980 PatternMapIterator it;
981 for (it.set(*patternMap); it.hasNext(); ) {
982 DateTimeMatcher trial = it.next();
983 if (trial.equals(skipMatcher)) {
984 continue;
985 }
986 int32_t distance=source.getDistance(trial, includeMask, tempInfo);
987 if (distance<bestDistance) {
988 bestDistance=distance;
989 bestPattern=patternMap->getPatternFromSkeleton(*trial.getSkeletonPtr(), &specifiedSkeleton);
990 missingFields->setTo(tempInfo);
991 if (distance==0) {
992 break;
993 }
994 }
995 }
996
997 // If the best raw match had a specified skeleton and that skeleton was requested by the caller,
998 // then return it too. This generally happens when the caller needs to pass that skeleton
999 // through to adjustFieldTypes so the latter can do a better job.
1000 if (bestPattern && specifiedSkeletonPtr) {
1001 *specifiedSkeletonPtr = specifiedSkeleton;
1002 }
1003 return bestPattern;
1004 }
1005
1006 UnicodeString
adjustFieldTypes(const UnicodeString & pattern,const PtnSkeleton * specifiedSkeleton,UBool fixFractionalSeconds,UDateTimePatternMatchOptions options)1007 DateTimePatternGenerator::adjustFieldTypes(const UnicodeString& pattern,
1008 const PtnSkeleton* specifiedSkeleton,
1009 UBool fixFractionalSeconds,
1010 UDateTimePatternMatchOptions options) {
1011 UnicodeString newPattern;
1012 fp->set(pattern);
1013 for (int32_t i=0; i < fp->itemNumber; i++) {
1014 UnicodeString field = fp->items[i];
1015 if ( fp->isQuoteLiteral(field) ) {
1016
1017 UnicodeString quoteLiteral;
1018 fp->getQuoteLiteral(quoteLiteral, &i);
1019 newPattern += quoteLiteral;
1020 }
1021 else {
1022 if (fp->isPatternSeparator(field)) {
1023 newPattern+=field;
1024 continue;
1025 }
1026 int32_t canonicalIndex = fp->getCanonicalIndex(field);
1027 if (canonicalIndex < 0) {
1028 newPattern+=field;
1029 continue; // don't adjust
1030 }
1031 const dtTypeElem *row = &dtTypes[canonicalIndex];
1032 int32_t typeValue = row->field;
1033 if (fixFractionalSeconds && typeValue == UDATPG_SECOND_FIELD) {
1034 UnicodeString newField=dtMatcher->skeleton.original[UDATPG_FRACTIONAL_SECOND_FIELD];
1035 field = field + decimal + newField;
1036 } else if (dtMatcher->skeleton.type[typeValue]!=0) {
1037 // Here:
1038 // - "reqField" is the field from the originally requested skeleton, with length
1039 // "reqFieldLen".
1040 // - "field" is the field from the found pattern.
1041 //
1042 // The adjusted field should consist of characters from the originally requested
1043 // skeleton, except in the case of UDATPG_HOUR_FIELD or UDATPG_MONTH_FIELD or
1044 // UDATPG_WEEKDAY_FIELD or UDATPG_YEAR_FIELD, in which case it should consist
1045 // of characters from the found pattern.
1046 //
1047 // The length of the adjusted field (adjFieldLen) should match that in the originally
1048 // requested skeleton, except that in the following cases the length of the adjusted field
1049 // should match that in the found pattern (i.e. the length of this pattern field should
1050 // not be adjusted):
1051 // 1. typeValue is UDATPG_HOUR_FIELD/MINUTE/SECOND and the corresponding bit in options is
1052 // not set (ticket #7180). Note, we may want to implement a similar change for other
1053 // numeric fields (MM, dd, etc.) so the default behavior is to get locale preference for
1054 // field length, but options bits can be used to override this.
1055 // 2. There is a specified skeleton for the found pattern and one of the following is true:
1056 // a) The length of the field in the skeleton (skelFieldLen) is equal to reqFieldLen.
1057 // b) The pattern field is numeric and the skeleton field is not, or vice versa.
1058
1059 UnicodeString reqField = dtMatcher->skeleton.original[typeValue];
1060 int32_t reqFieldLen = reqField.length();
1061 if (reqField.charAt(0) == CAP_E && reqFieldLen < 3)
1062 reqFieldLen = 3; // 1-3 for E are equivalent to 3 for c,e
1063 int32_t adjFieldLen = reqFieldLen;
1064 if ( (typeValue==UDATPG_HOUR_FIELD && (options & UDATPG_MATCH_HOUR_FIELD_LENGTH)==0) ||
1065 (typeValue==UDATPG_MINUTE_FIELD && (options & UDATPG_MATCH_MINUTE_FIELD_LENGTH)==0) ||
1066 (typeValue==UDATPG_SECOND_FIELD && (options & UDATPG_MATCH_SECOND_FIELD_LENGTH)==0) ) {
1067 adjFieldLen = field.length();
1068 } else if (specifiedSkeleton) {
1069 UnicodeString skelField = specifiedSkeleton->original[typeValue];
1070 int32_t skelFieldLen = skelField.length();
1071 UBool patFieldIsNumeric = (row->type > 0);
1072 UBool skelFieldIsNumeric = (specifiedSkeleton->type[typeValue] > 0);
1073 if (skelFieldLen == reqFieldLen || (patFieldIsNumeric && !skelFieldIsNumeric) || (skelFieldIsNumeric && !patFieldIsNumeric)) {
1074 // don't adjust the field length in the found pattern
1075 adjFieldLen = field.length();
1076 }
1077 }
1078 UChar c = (typeValue!= UDATPG_HOUR_FIELD && typeValue!= UDATPG_MONTH_FIELD &&
1079 typeValue!= UDATPG_WEEKDAY_FIELD && typeValue!= UDATPG_YEAR_FIELD)?
1080 reqField.charAt(0): field.charAt(0);
1081 field.remove();
1082 for (int32_t i=adjFieldLen; i>0; --i) {
1083 field+=c;
1084 }
1085 }
1086 newPattern+=field;
1087 }
1088 }
1089 return newPattern;
1090 }
1091
1092 UnicodeString
getBestAppending(int32_t missingFields,UDateTimePatternMatchOptions options)1093 DateTimePatternGenerator::getBestAppending(int32_t missingFields, UDateTimePatternMatchOptions options) {
1094 UnicodeString resultPattern, tempPattern;
1095 UErrorCode err=U_ZERO_ERROR;
1096 int32_t lastMissingFieldMask=0;
1097 if (missingFields!=0) {
1098 resultPattern=UnicodeString();
1099 const PtnSkeleton* specifiedSkeleton=NULL;
1100 tempPattern = *getBestRaw(*dtMatcher, missingFields, distanceInfo, &specifiedSkeleton);
1101 resultPattern = adjustFieldTypes(tempPattern, specifiedSkeleton, FALSE, options);
1102 if ( distanceInfo->missingFieldMask==0 ) {
1103 return resultPattern;
1104 }
1105 while (distanceInfo->missingFieldMask!=0) { // precondition: EVERY single field must work!
1106 if ( lastMissingFieldMask == distanceInfo->missingFieldMask ) {
1107 break; // cannot find the proper missing field
1108 }
1109 if (((distanceInfo->missingFieldMask & UDATPG_SECOND_AND_FRACTIONAL_MASK)==UDATPG_FRACTIONAL_MASK) &&
1110 ((missingFields & UDATPG_SECOND_AND_FRACTIONAL_MASK) == UDATPG_SECOND_AND_FRACTIONAL_MASK)) {
1111 resultPattern = adjustFieldTypes(resultPattern, specifiedSkeleton, TRUE, options);
1112 distanceInfo->missingFieldMask &= ~UDATPG_FRACTIONAL_MASK;
1113 continue;
1114 }
1115 int32_t startingMask = distanceInfo->missingFieldMask;
1116 tempPattern = *getBestRaw(*dtMatcher, distanceInfo->missingFieldMask, distanceInfo, &specifiedSkeleton);
1117 tempPattern = adjustFieldTypes(tempPattern, specifiedSkeleton, FALSE, options);
1118 int32_t foundMask=startingMask& ~distanceInfo->missingFieldMask;
1119 int32_t topField=getTopBitNumber(foundMask);
1120 UnicodeString appendName;
1121 getAppendName((UDateTimePatternField)topField, appendName);
1122 const Formattable formatPattern[] = {
1123 resultPattern,
1124 tempPattern,
1125 appendName
1126 };
1127 UnicodeString emptyStr;
1128 resultPattern = MessageFormat::format(appendItemFormats[topField], formatPattern, 3, emptyStr, err);
1129 lastMissingFieldMask = distanceInfo->missingFieldMask;
1130 }
1131 }
1132 return resultPattern;
1133 }
1134
1135 int32_t
getTopBitNumber(int32_t foundMask)1136 DateTimePatternGenerator::getTopBitNumber(int32_t foundMask) {
1137 if ( foundMask==0 ) {
1138 return 0;
1139 }
1140 int32_t i=0;
1141 while (foundMask!=0) {
1142 foundMask >>=1;
1143 ++i;
1144 }
1145 if (i-1 >UDATPG_ZONE_FIELD) {
1146 return UDATPG_ZONE_FIELD;
1147 }
1148 else
1149 return i-1;
1150 }
1151
1152 void
setAvailableFormat(const UnicodeString & key,UErrorCode & err)1153 DateTimePatternGenerator::setAvailableFormat(const UnicodeString &key, UErrorCode& err)
1154 {
1155 fAvailableFormatKeyHash->puti(key, 1, err);
1156 }
1157
1158 UBool
isAvailableFormatSet(const UnicodeString & key) const1159 DateTimePatternGenerator::isAvailableFormatSet(const UnicodeString &key) const {
1160 return (UBool)(fAvailableFormatKeyHash->geti(key) == 1);
1161 }
1162
1163 void
copyHashtable(Hashtable * other,UErrorCode & status)1164 DateTimePatternGenerator::copyHashtable(Hashtable *other, UErrorCode &status) {
1165
1166 if (other == NULL) {
1167 return;
1168 }
1169 if (fAvailableFormatKeyHash != NULL) {
1170 delete fAvailableFormatKeyHash;
1171 fAvailableFormatKeyHash = NULL;
1172 }
1173 initHashtable(status);
1174 if(U_FAILURE(status)){
1175 return;
1176 }
1177 int32_t pos = -1;
1178 const UHashElement* elem = NULL;
1179 // walk through the hash table and create a deep clone
1180 while((elem = other->nextElement(pos))!= NULL){
1181 const UHashTok otherKeyTok = elem->key;
1182 UnicodeString* otherKey = (UnicodeString*)otherKeyTok.pointer;
1183 fAvailableFormatKeyHash->puti(*otherKey, 1, status);
1184 if(U_FAILURE(status)){
1185 return;
1186 }
1187 }
1188 }
1189
1190 StringEnumeration*
getSkeletons(UErrorCode & status) const1191 DateTimePatternGenerator::getSkeletons(UErrorCode& status) const {
1192 StringEnumeration* skeletonEnumerator = new DTSkeletonEnumeration(*patternMap, DT_SKELETON, status);
1193 return skeletonEnumerator;
1194 }
1195
1196 const UnicodeString&
getPatternForSkeleton(const UnicodeString & skeleton) const1197 DateTimePatternGenerator::getPatternForSkeleton(const UnicodeString& skeleton) const {
1198 PtnElem *curElem;
1199
1200 if (skeleton.length() ==0) {
1201 return emptyString;
1202 }
1203 curElem = patternMap->getHeader(skeleton.charAt(0));
1204 while ( curElem != NULL ) {
1205 if ( curElem->skeleton->getSkeleton()==skeleton ) {
1206 return curElem->pattern;
1207 }
1208 curElem=curElem->next;
1209 }
1210 return emptyString;
1211 }
1212
1213 StringEnumeration*
getBaseSkeletons(UErrorCode & status) const1214 DateTimePatternGenerator::getBaseSkeletons(UErrorCode& status) const {
1215 StringEnumeration* baseSkeletonEnumerator = new DTSkeletonEnumeration(*patternMap, DT_BASESKELETON, status);
1216 return baseSkeletonEnumerator;
1217 }
1218
1219 StringEnumeration*
getRedundants(UErrorCode & status)1220 DateTimePatternGenerator::getRedundants(UErrorCode& status) {
1221 StringEnumeration* output = new DTRedundantEnumeration();
1222 const UnicodeString *pattern;
1223 PatternMapIterator it;
1224 for (it.set(*patternMap); it.hasNext(); ) {
1225 DateTimeMatcher current = it.next();
1226 pattern = patternMap->getPatternFromSkeleton(*(it.getSkeleton()));
1227 if ( isCanonicalItem(*pattern) ) {
1228 continue;
1229 }
1230 if ( skipMatcher == NULL ) {
1231 skipMatcher = new DateTimeMatcher(current);
1232 }
1233 else {
1234 *skipMatcher = current;
1235 }
1236 UnicodeString trial = getBestPattern(current.getPattern(), status);
1237 if (trial == *pattern) {
1238 ((DTRedundantEnumeration *)output)->add(*pattern, status);
1239 }
1240 if (current.equals(skipMatcher)) {
1241 continue;
1242 }
1243 }
1244 return output;
1245 }
1246
1247 UBool
isCanonicalItem(const UnicodeString & item) const1248 DateTimePatternGenerator::isCanonicalItem(const UnicodeString& item) const {
1249 if ( item.length() != 1 ) {
1250 return FALSE;
1251 }
1252 for (int32_t i=0; i<UDATPG_FIELD_COUNT; ++i) {
1253 if (item.charAt(0)==Canonical_Items[i]) {
1254 return TRUE;
1255 }
1256 }
1257 return FALSE;
1258 }
1259
1260
1261 DateTimePatternGenerator*
clone() const1262 DateTimePatternGenerator::clone() const {
1263 return new DateTimePatternGenerator(*this);
1264 }
1265
PatternMap()1266 PatternMap::PatternMap() {
1267 for (int32_t i=0; i < MAX_PATTERN_ENTRIES; ++i ) {
1268 boot[i]=NULL;
1269 }
1270 isDupAllowed = TRUE;
1271 }
1272
1273 void
copyFrom(const PatternMap & other,UErrorCode & status)1274 PatternMap::copyFrom(const PatternMap& other, UErrorCode& status) {
1275 this->isDupAllowed = other.isDupAllowed;
1276 for (int32_t bootIndex=0; bootIndex<MAX_PATTERN_ENTRIES; ++bootIndex ) {
1277 PtnElem *curElem, *otherElem, *prevElem=NULL;
1278 otherElem = other.boot[bootIndex];
1279 while (otherElem!=NULL) {
1280 if ((curElem = new PtnElem(otherElem->basePattern, otherElem->pattern))==NULL) {
1281 // out of memory
1282 status = U_MEMORY_ALLOCATION_ERROR;
1283 return;
1284 }
1285 if ( this->boot[bootIndex]== NULL ) {
1286 this->boot[bootIndex] = curElem;
1287 }
1288 if ((curElem->skeleton=new PtnSkeleton(*(otherElem->skeleton))) == NULL ) {
1289 // out of memory
1290 status = U_MEMORY_ALLOCATION_ERROR;
1291 return;
1292 }
1293
1294 if (prevElem!=NULL) {
1295 prevElem->next=curElem;
1296 }
1297 curElem->next=NULL;
1298 prevElem = curElem;
1299 otherElem = otherElem->next;
1300 }
1301
1302 }
1303 }
1304
1305 PtnElem*
getHeader(UChar baseChar)1306 PatternMap::getHeader(UChar baseChar) {
1307 PtnElem* curElem;
1308
1309 if ( (baseChar >= CAP_A) && (baseChar <= CAP_Z) ) {
1310 curElem = boot[baseChar-CAP_A];
1311 }
1312 else {
1313 if ( (baseChar >=LOW_A) && (baseChar <= LOW_Z) ) {
1314 curElem = boot[26+baseChar-LOW_A];
1315 }
1316 else {
1317 return NULL;
1318 }
1319 }
1320 return curElem;
1321 }
1322
~PatternMap()1323 PatternMap::~PatternMap() {
1324 for (int32_t i=0; i < MAX_PATTERN_ENTRIES; ++i ) {
1325 if (boot[i]!=NULL ) {
1326 delete boot[i];
1327 boot[i]=NULL;
1328 }
1329 }
1330 } // PatternMap destructor
1331
1332 void
add(const UnicodeString & basePattern,const PtnSkeleton & skeleton,const UnicodeString & value,UBool skeletonWasSpecified,UErrorCode & status)1333 PatternMap::add(const UnicodeString& basePattern,
1334 const PtnSkeleton& skeleton,
1335 const UnicodeString& value,// mapped pattern value
1336 UBool skeletonWasSpecified,
1337 UErrorCode &status) {
1338 UChar baseChar = basePattern.charAt(0);
1339 PtnElem *curElem, *baseElem;
1340 status = U_ZERO_ERROR;
1341
1342 // the baseChar must be A-Z or a-z
1343 if ((baseChar >= CAP_A) && (baseChar <= CAP_Z)) {
1344 baseElem = boot[baseChar-CAP_A];
1345 }
1346 else {
1347 if ((baseChar >=LOW_A) && (baseChar <= LOW_Z)) {
1348 baseElem = boot[26+baseChar-LOW_A];
1349 }
1350 else {
1351 status = U_ILLEGAL_CHARACTER;
1352 return;
1353 }
1354 }
1355
1356 if (baseElem == NULL) {
1357 if ((curElem = new PtnElem(basePattern, value)) == NULL ) {
1358 // out of memory
1359 status = U_MEMORY_ALLOCATION_ERROR;
1360 return;
1361 }
1362 if (baseChar >= LOW_A) {
1363 boot[26 + (baseChar-LOW_A)] = curElem;
1364 }
1365 else {
1366 boot[baseChar-CAP_A] = curElem;
1367 }
1368 curElem->skeleton = new PtnSkeleton(skeleton);
1369 curElem->skeletonWasSpecified = skeletonWasSpecified;
1370 }
1371 if ( baseElem != NULL ) {
1372 curElem = getDuplicateElem(basePattern, skeleton, baseElem);
1373
1374 if (curElem == NULL) {
1375 // add new element to the list.
1376 curElem = baseElem;
1377 while( curElem -> next != NULL )
1378 {
1379 curElem = curElem->next;
1380 }
1381 if ((curElem->next = new PtnElem(basePattern, value)) == NULL ) {
1382 // out of memory
1383 status = U_MEMORY_ALLOCATION_ERROR;
1384 return;
1385 }
1386 curElem=curElem->next;
1387 curElem->skeleton = new PtnSkeleton(skeleton);
1388 curElem->skeletonWasSpecified = skeletonWasSpecified;
1389 }
1390 else {
1391 // Pattern exists in the list already.
1392 if ( !isDupAllowed ) {
1393 return;
1394 }
1395 // Overwrite the value.
1396 curElem->pattern = value;
1397 }
1398 }
1399 } // PatternMap::add
1400
1401 // Find the pattern from the given basePattern string.
1402 const UnicodeString *
getPatternFromBasePattern(UnicodeString & basePattern,UBool & skeletonWasSpecified)1403 PatternMap::getPatternFromBasePattern(UnicodeString& basePattern, UBool& skeletonWasSpecified) { // key to search for
1404 PtnElem *curElem;
1405
1406 if ((curElem=getHeader(basePattern.charAt(0)))==NULL) {
1407 return NULL; // no match
1408 }
1409
1410 do {
1411 if ( basePattern.compare(curElem->basePattern)==0 ) {
1412 skeletonWasSpecified = curElem->skeletonWasSpecified;
1413 return &(curElem->pattern);
1414 }
1415 curElem=curElem->next;
1416 }while (curElem != NULL);
1417
1418 return NULL;
1419 } // PatternMap::getFromBasePattern
1420
1421
1422 // Find the pattern from the given skeleton.
1423 // At least when this is called from getBestRaw & addPattern (in which case specifiedSkeletonPtr is non-NULL),
1424 // the comparison should be based on skeleton.original (which is unique and tied to the distance measurement in bestRaw)
1425 // and not skeleton.baseOriginal (which is not unique); otherwise we may pick a different skeleton than the one with the
1426 // optimum distance value in getBestRaw. When this is called from public getRedundants (specifiedSkeletonPtr is NULL),
1427 // for now it will continue to compare based on baseOriginal so as not to change the behavior unnecessarily.
1428 const UnicodeString *
getPatternFromSkeleton(PtnSkeleton & skeleton,const PtnSkeleton ** specifiedSkeletonPtr)1429 PatternMap::getPatternFromSkeleton(PtnSkeleton& skeleton, const PtnSkeleton** specifiedSkeletonPtr) { // key to search for
1430 PtnElem *curElem;
1431
1432 if (specifiedSkeletonPtr) {
1433 *specifiedSkeletonPtr = NULL;
1434 }
1435
1436 // find boot entry
1437 UChar baseChar='\0';
1438 for (int32_t i=0; i<UDATPG_FIELD_COUNT; ++i) {
1439 if (skeleton.baseOriginal[i].length() !=0 ) {
1440 baseChar = skeleton.baseOriginal[i].charAt(0);
1441 break;
1442 }
1443 }
1444
1445 if ((curElem=getHeader(baseChar))==NULL) {
1446 return NULL; // no match
1447 }
1448
1449 do {
1450 int32_t i=0;
1451 if (specifiedSkeletonPtr != NULL) { // called from DateTimePatternGenerator::getBestRaw or addPattern, use original
1452 for (i=0; i<UDATPG_FIELD_COUNT; ++i) {
1453 if (curElem->skeleton->original[i].compare(skeleton.original[i]) != 0 )
1454 {
1455 break;
1456 }
1457 }
1458 } else { // called from DateTimePatternGenerator::getRedundants, use baseOriginal
1459 for (i=0; i<UDATPG_FIELD_COUNT; ++i) {
1460 if (curElem->skeleton->baseOriginal[i].compare(skeleton.baseOriginal[i]) != 0 )
1461 {
1462 break;
1463 }
1464 }
1465 }
1466 if (i == UDATPG_FIELD_COUNT) {
1467 if (specifiedSkeletonPtr && curElem->skeletonWasSpecified) {
1468 *specifiedSkeletonPtr = curElem->skeleton;
1469 }
1470 return &(curElem->pattern);
1471 }
1472 curElem=curElem->next;
1473 }while (curElem != NULL);
1474
1475 return NULL;
1476 }
1477
1478 UBool
equals(const PatternMap & other)1479 PatternMap::equals(const PatternMap& other) {
1480 if ( this==&other ) {
1481 return TRUE;
1482 }
1483 for (int32_t bootIndex=0; bootIndex<MAX_PATTERN_ENTRIES; ++bootIndex ) {
1484 if ( boot[bootIndex]==other.boot[bootIndex] ) {
1485 continue;
1486 }
1487 if ( (boot[bootIndex]==NULL)||(other.boot[bootIndex]==NULL) ) {
1488 return FALSE;
1489 }
1490 PtnElem *otherElem = other.boot[bootIndex];
1491 PtnElem *myElem = boot[bootIndex];
1492 while ((otherElem!=NULL) || (myElem!=NULL)) {
1493 if ( myElem == otherElem ) {
1494 break;
1495 }
1496 if ((otherElem==NULL) || (myElem==NULL)) {
1497 return FALSE;
1498 }
1499 if ( (myElem->basePattern != otherElem->basePattern) ||
1500 (myElem->pattern != otherElem->pattern) ) {
1501 return FALSE;
1502 }
1503 if ((myElem->skeleton!=otherElem->skeleton)&&
1504 !myElem->skeleton->equals(*(otherElem->skeleton))) {
1505 return FALSE;
1506 }
1507 myElem = myElem->next;
1508 otherElem=otherElem->next;
1509 }
1510 }
1511 return TRUE;
1512 }
1513
1514 // find any key existing in the mapping table already.
1515 // return TRUE if there is an existing key, otherwise return FALSE.
1516 PtnElem*
getDuplicateElem(const UnicodeString & basePattern,const PtnSkeleton & skeleton,PtnElem * baseElem)1517 PatternMap::getDuplicateElem(
1518 const UnicodeString &basePattern,
1519 const PtnSkeleton &skeleton,
1520 PtnElem *baseElem) {
1521 PtnElem *curElem;
1522
1523 if ( baseElem == (PtnElem *)NULL ) {
1524 return (PtnElem*)NULL;
1525 }
1526 else {
1527 curElem = baseElem;
1528 }
1529 do {
1530 if ( basePattern.compare(curElem->basePattern)==0 ) {
1531 UBool isEqual=TRUE;
1532 for (int32_t i=0; i<UDATPG_FIELD_COUNT; ++i) {
1533 if (curElem->skeleton->type[i] != skeleton.type[i] ) {
1534 isEqual=FALSE;
1535 break;
1536 }
1537 }
1538 if (isEqual) {
1539 return curElem;
1540 }
1541 }
1542 curElem = curElem->next;
1543 } while( curElem != (PtnElem *)NULL );
1544
1545 // end of the list
1546 return (PtnElem*)NULL;
1547
1548 } // PatternMap::getDuplicateElem
1549
DateTimeMatcher(void)1550 DateTimeMatcher::DateTimeMatcher(void) {
1551 }
1552
~DateTimeMatcher()1553 DateTimeMatcher::~DateTimeMatcher() {}
1554
DateTimeMatcher(const DateTimeMatcher & other)1555 DateTimeMatcher::DateTimeMatcher(const DateTimeMatcher& other) {
1556 copyFrom(other.skeleton);
1557 }
1558
1559
1560 void
set(const UnicodeString & pattern,FormatParser * fp)1561 DateTimeMatcher::set(const UnicodeString& pattern, FormatParser* fp) {
1562 PtnSkeleton localSkeleton;
1563 return set(pattern, fp, localSkeleton);
1564 }
1565
1566 void
set(const UnicodeString & pattern,FormatParser * fp,PtnSkeleton & skeletonResult)1567 DateTimeMatcher::set(const UnicodeString& pattern, FormatParser* fp, PtnSkeleton& skeletonResult) {
1568 int32_t i;
1569 for (i=0; i<UDATPG_FIELD_COUNT; ++i) {
1570 skeletonResult.type[i]=NONE;
1571 }
1572 fp->set(pattern);
1573 for (i=0; i < fp->itemNumber; i++) {
1574 UnicodeString field = fp->items[i];
1575 if ( field.charAt(0) == LOW_A ) {
1576 continue; // skip 'a'
1577 }
1578
1579 if ( fp->isQuoteLiteral(field) ) {
1580 UnicodeString quoteLiteral;
1581 fp->getQuoteLiteral(quoteLiteral, &i);
1582 continue;
1583 }
1584 int32_t canonicalIndex = fp->getCanonicalIndex(field);
1585 if (canonicalIndex < 0 ) {
1586 continue;
1587 }
1588 const dtTypeElem *row = &dtTypes[canonicalIndex];
1589 int32_t typeValue = row->field;
1590 skeletonResult.original[typeValue]=field;
1591 UChar repeatChar = row->patternChar;
1592 int32_t repeatCount = row->minLen; // #7930 removes cap at 3
1593 while (repeatCount-- > 0) {
1594 skeletonResult.baseOriginal[typeValue] += repeatChar;
1595 }
1596 int16_t subTypeValue = row->type;
1597 if ( row->type > 0) {
1598 subTypeValue += field.length();
1599 }
1600 skeletonResult.type[typeValue] = subTypeValue;
1601 }
1602 copyFrom(skeletonResult);
1603 }
1604
1605 void
getBasePattern(UnicodeString & result)1606 DateTimeMatcher::getBasePattern(UnicodeString &result ) {
1607 result.remove(); // Reset the result first.
1608 for (int32_t i=0; i<UDATPG_FIELD_COUNT; ++i ) {
1609 if (skeleton.baseOriginal[i].length()!=0) {
1610 result += skeleton.baseOriginal[i];
1611 }
1612 }
1613 }
1614
1615 UnicodeString
getPattern()1616 DateTimeMatcher::getPattern() {
1617 UnicodeString result;
1618
1619 for (int32_t i=0; i<UDATPG_FIELD_COUNT; ++i ) {
1620 if (skeleton.original[i].length()!=0) {
1621 result += skeleton.original[i];
1622 }
1623 }
1624 return result;
1625 }
1626
1627 int32_t
getDistance(const DateTimeMatcher & other,int32_t includeMask,DistanceInfo & distanceInfo)1628 DateTimeMatcher::getDistance(const DateTimeMatcher& other, int32_t includeMask, DistanceInfo& distanceInfo) {
1629 int32_t result=0;
1630 distanceInfo.clear();
1631 for (int32_t i=0; i<UDATPG_FIELD_COUNT; ++i ) {
1632 int32_t myType = (includeMask&(1<<i))==0 ? 0 : skeleton.type[i];
1633 int32_t otherType = other.skeleton.type[i];
1634 if (myType==otherType) {
1635 continue;
1636 }
1637 if (myType==0) {// and other is not
1638 result += EXTRA_FIELD;
1639 distanceInfo.addExtra(i);
1640 }
1641 else {
1642 if (otherType==0) {
1643 result += MISSING_FIELD;
1644 distanceInfo.addMissing(i);
1645 }
1646 else {
1647 result += abs(myType - otherType);
1648 }
1649 }
1650
1651 }
1652 return result;
1653 }
1654
1655 void
copyFrom(const PtnSkeleton & newSkeleton)1656 DateTimeMatcher::copyFrom(const PtnSkeleton& newSkeleton) {
1657 for (int32_t i=0; i<UDATPG_FIELD_COUNT; ++i) {
1658 this->skeleton.type[i]=newSkeleton.type[i];
1659 this->skeleton.original[i]=newSkeleton.original[i];
1660 this->skeleton.baseOriginal[i]=newSkeleton.baseOriginal[i];
1661 }
1662 }
1663
1664 void
copyFrom()1665 DateTimeMatcher::copyFrom() {
1666 // same as clear
1667 for (int32_t i=0; i<UDATPG_FIELD_COUNT; ++i) {
1668 this->skeleton.type[i]=0;
1669 this->skeleton.original[i].remove();
1670 this->skeleton.baseOriginal[i].remove();
1671 }
1672 }
1673
1674 UBool
equals(const DateTimeMatcher * other) const1675 DateTimeMatcher::equals(const DateTimeMatcher* other) const {
1676 if (other==NULL) {
1677 return FALSE;
1678 }
1679 for (int32_t i=0; i<UDATPG_FIELD_COUNT; ++i) {
1680 if (this->skeleton.original[i]!=other->skeleton.original[i] ) {
1681 return FALSE;
1682 }
1683 }
1684 return TRUE;
1685 }
1686
1687 int32_t
getFieldMask()1688 DateTimeMatcher::getFieldMask() {
1689 int32_t result=0;
1690
1691 for (int32_t i=0; i<UDATPG_FIELD_COUNT; ++i) {
1692 if (skeleton.type[i]!=0) {
1693 result |= (1<<i);
1694 }
1695 }
1696 return result;
1697 }
1698
1699 PtnSkeleton*
getSkeletonPtr()1700 DateTimeMatcher::getSkeletonPtr() {
1701 return &skeleton;
1702 }
1703
FormatParser()1704 FormatParser::FormatParser () {
1705 status = START;
1706 itemNumber=0;
1707 }
1708
1709
~FormatParser()1710 FormatParser::~FormatParser () {
1711 }
1712
1713
1714 // Find the next token with the starting position and length
1715 // Note: the startPos may
1716 FormatParser::TokenStatus
setTokens(const UnicodeString & pattern,int32_t startPos,int32_t * len)1717 FormatParser::setTokens(const UnicodeString& pattern, int32_t startPos, int32_t *len) {
1718 int32_t curLoc = startPos;
1719 if ( curLoc >= pattern.length()) {
1720 return DONE;
1721 }
1722 // check the current char is between A-Z or a-z
1723 do {
1724 UChar c=pattern.charAt(curLoc);
1725 if ( (c>=CAP_A && c<=CAP_Z) || (c>=LOW_A && c<=LOW_Z) ) {
1726 curLoc++;
1727 }
1728 else {
1729 startPos = curLoc;
1730 *len=1;
1731 return ADD_TOKEN;
1732 }
1733
1734 if ( pattern.charAt(curLoc)!= pattern.charAt(startPos) ) {
1735 break; // not the same token
1736 }
1737 } while(curLoc <= pattern.length());
1738 *len = curLoc-startPos;
1739 return ADD_TOKEN;
1740 }
1741
1742 void
set(const UnicodeString & pattern)1743 FormatParser::set(const UnicodeString& pattern) {
1744 int32_t startPos=0;
1745 TokenStatus result=START;
1746 int32_t len=0;
1747 itemNumber =0;
1748
1749 do {
1750 result = setTokens( pattern, startPos, &len );
1751 if ( result == ADD_TOKEN )
1752 {
1753 items[itemNumber++] = UnicodeString(pattern, startPos, len );
1754 startPos += len;
1755 }
1756 else {
1757 break;
1758 }
1759 } while (result==ADD_TOKEN && itemNumber < MAX_DT_TOKEN);
1760 }
1761
1762 int32_t
getCanonicalIndex(const UnicodeString & s,UBool strict)1763 FormatParser::getCanonicalIndex(const UnicodeString& s, UBool strict) {
1764 int32_t len = s.length();
1765 if (len == 0) {
1766 return -1;
1767 }
1768 UChar ch = s.charAt(0);
1769
1770 // Verify that all are the same character.
1771 for (int32_t l = 1; l < len; l++) {
1772 if (ch != s.charAt(l)) {
1773 return -1;
1774 }
1775 }
1776 int32_t i = 0;
1777 int32_t bestRow = -1;
1778 while (dtTypes[i].patternChar != '\0') {
1779 if ( dtTypes[i].patternChar != ch ) {
1780 ++i;
1781 continue;
1782 }
1783 bestRow = i;
1784 if (dtTypes[i].patternChar != dtTypes[i+1].patternChar) {
1785 return i;
1786 }
1787 if (dtTypes[i+1].minLen <= len) {
1788 ++i;
1789 continue;
1790 }
1791 return i;
1792 }
1793 return strict ? -1 : bestRow;
1794 }
1795
1796 UBool
isQuoteLiteral(const UnicodeString & s) const1797 FormatParser::isQuoteLiteral(const UnicodeString& s) const {
1798 return (UBool)(s.charAt(0)==SINGLE_QUOTE);
1799 }
1800
1801 // This function aussumes the current itemIndex points to the quote literal.
1802 // Please call isQuoteLiteral prior to this function.
1803 void
getQuoteLiteral(UnicodeString & quote,int32_t * itemIndex)1804 FormatParser::getQuoteLiteral(UnicodeString& quote, int32_t *itemIndex) {
1805 int32_t i=*itemIndex;
1806
1807 quote.remove();
1808 if (items[i].charAt(0)==SINGLE_QUOTE) {
1809 quote += items[i];
1810 ++i;
1811 }
1812 while ( i < itemNumber ) {
1813 if ( items[i].charAt(0)==SINGLE_QUOTE ) {
1814 if ( (i+1<itemNumber) && (items[i+1].charAt(0)==SINGLE_QUOTE)) {
1815 // two single quotes e.g. 'o''clock'
1816 quote += items[i++];
1817 quote += items[i++];
1818 continue;
1819 }
1820 else {
1821 quote += items[i];
1822 break;
1823 }
1824 }
1825 else {
1826 quote += items[i];
1827 }
1828 ++i;
1829 }
1830 *itemIndex=i;
1831 }
1832
1833 UBool
isPatternSeparator(UnicodeString & field)1834 FormatParser::isPatternSeparator(UnicodeString& field) {
1835 for (int32_t i=0; i<field.length(); ++i ) {
1836 UChar c= field.charAt(i);
1837 if ( (c==SINGLE_QUOTE) || (c==BACKSLASH) || (c==SPACE) || (c==COLON) ||
1838 (c==QUOTATION_MARK) || (c==COMMA) || (c==HYPHEN) ||(items[i].charAt(0)==DOT) ) {
1839 continue;
1840 }
1841 else {
1842 return FALSE;
1843 }
1844 }
1845 return TRUE;
1846 }
1847
~DistanceInfo()1848 DistanceInfo::~DistanceInfo() {}
1849
1850 void
setTo(DistanceInfo & other)1851 DistanceInfo::setTo(DistanceInfo &other) {
1852 missingFieldMask = other.missingFieldMask;
1853 extraFieldMask= other.extraFieldMask;
1854 }
1855
PatternMapIterator()1856 PatternMapIterator::PatternMapIterator() {
1857 bootIndex = 0;
1858 nodePtr = NULL;
1859 patternMap=NULL;
1860 matcher= new DateTimeMatcher();
1861 }
1862
1863
~PatternMapIterator()1864 PatternMapIterator::~PatternMapIterator() {
1865 delete matcher;
1866 }
1867
1868 void
set(PatternMap & newPatternMap)1869 PatternMapIterator::set(PatternMap& newPatternMap) {
1870 this->patternMap=&newPatternMap;
1871 }
1872
1873 PtnSkeleton*
getSkeleton()1874 PatternMapIterator::getSkeleton() {
1875 if ( nodePtr == NULL ) {
1876 return NULL;
1877 }
1878 else {
1879 return nodePtr->skeleton;
1880 }
1881 }
1882
1883 UBool
hasNext()1884 PatternMapIterator::hasNext() {
1885 int32_t headIndex=bootIndex;
1886 PtnElem *curPtr=nodePtr;
1887
1888 if (patternMap==NULL) {
1889 return FALSE;
1890 }
1891 while ( headIndex < MAX_PATTERN_ENTRIES ) {
1892 if ( curPtr != NULL ) {
1893 if ( curPtr->next != NULL ) {
1894 return TRUE;
1895 }
1896 else {
1897 headIndex++;
1898 curPtr=NULL;
1899 continue;
1900 }
1901 }
1902 else {
1903 if ( patternMap->boot[headIndex] != NULL ) {
1904 return TRUE;
1905 }
1906 else {
1907 headIndex++;
1908 continue;
1909 }
1910 }
1911
1912 }
1913 return FALSE;
1914 }
1915
1916 DateTimeMatcher&
next()1917 PatternMapIterator::next() {
1918 while ( bootIndex < MAX_PATTERN_ENTRIES ) {
1919 if ( nodePtr != NULL ) {
1920 if ( nodePtr->next != NULL ) {
1921 nodePtr = nodePtr->next;
1922 break;
1923 }
1924 else {
1925 bootIndex++;
1926 nodePtr=NULL;
1927 continue;
1928 }
1929 }
1930 else {
1931 if ( patternMap->boot[bootIndex] != NULL ) {
1932 nodePtr = patternMap->boot[bootIndex];
1933 break;
1934 }
1935 else {
1936 bootIndex++;
1937 continue;
1938 }
1939 }
1940 }
1941 if (nodePtr!=NULL) {
1942 matcher->copyFrom(*nodePtr->skeleton);
1943 }
1944 else {
1945 matcher->copyFrom();
1946 }
1947 return *matcher;
1948 }
1949
PtnSkeleton()1950 PtnSkeleton::PtnSkeleton() {
1951 }
1952
1953
PtnSkeleton(const PtnSkeleton & other)1954 PtnSkeleton::PtnSkeleton(const PtnSkeleton& other) {
1955 for (int32_t i=0; i<UDATPG_FIELD_COUNT; ++i) {
1956 this->type[i]=other.type[i];
1957 this->original[i]=other.original[i];
1958 this->baseOriginal[i]=other.baseOriginal[i];
1959 }
1960 }
1961
1962 UBool
equals(const PtnSkeleton & other)1963 PtnSkeleton::equals(const PtnSkeleton& other) {
1964 for (int32_t i=0; i<UDATPG_FIELD_COUNT; ++i) {
1965 if ( (type[i]!= other.type[i]) ||
1966 (original[i]!=other.original[i]) ||
1967 (baseOriginal[i]!=other.baseOriginal[i]) ) {
1968 return FALSE;
1969 }
1970 }
1971 return TRUE;
1972 }
1973
1974 UnicodeString
getSkeleton()1975 PtnSkeleton::getSkeleton() {
1976 UnicodeString result;
1977
1978 for(int32_t i=0; i< UDATPG_FIELD_COUNT; ++i) {
1979 if (original[i].length()!=0) {
1980 result += original[i];
1981 }
1982 }
1983 return result;
1984 }
1985
1986 UnicodeString
getBaseSkeleton()1987 PtnSkeleton::getBaseSkeleton() {
1988 UnicodeString result;
1989
1990 for(int32_t i=0; i< UDATPG_FIELD_COUNT; ++i) {
1991 if (baseOriginal[i].length()!=0) {
1992 result += baseOriginal[i];
1993 }
1994 }
1995 return result;
1996 }
1997
~PtnSkeleton()1998 PtnSkeleton::~PtnSkeleton() {
1999 }
2000
PtnElem(const UnicodeString & basePat,const UnicodeString & pat)2001 PtnElem::PtnElem(const UnicodeString &basePat, const UnicodeString &pat) :
2002 basePattern(basePat),
2003 skeleton(NULL),
2004 pattern(pat),
2005 next(NULL)
2006 {
2007 }
2008
~PtnElem()2009 PtnElem::~PtnElem() {
2010
2011 if (next!=NULL) {
2012 delete next;
2013 }
2014 delete skeleton;
2015 }
2016
DTSkeletonEnumeration(PatternMap & patternMap,dtStrEnum type,UErrorCode & status)2017 DTSkeletonEnumeration::DTSkeletonEnumeration(PatternMap &patternMap, dtStrEnum type, UErrorCode& status) {
2018 PtnElem *curElem;
2019 PtnSkeleton *curSkeleton;
2020 UnicodeString s;
2021 int32_t bootIndex;
2022
2023 pos=0;
2024 fSkeletons = new UVector(status);
2025 if (U_FAILURE(status)) {
2026 delete fSkeletons;
2027 return;
2028 }
2029 for (bootIndex=0; bootIndex<MAX_PATTERN_ENTRIES; ++bootIndex ) {
2030 curElem = patternMap.boot[bootIndex];
2031 while (curElem!=NULL) {
2032 switch(type) {
2033 case DT_BASESKELETON:
2034 s=curElem->basePattern;
2035 break;
2036 case DT_PATTERN:
2037 s=curElem->pattern;
2038 break;
2039 case DT_SKELETON:
2040 curSkeleton=curElem->skeleton;
2041 s=curSkeleton->getSkeleton();
2042 break;
2043 }
2044 if ( !isCanonicalItem(s) ) {
2045 fSkeletons->addElement(new UnicodeString(s), status);
2046 if (U_FAILURE(status)) {
2047 delete fSkeletons;
2048 fSkeletons = NULL;
2049 return;
2050 }
2051 }
2052 curElem = curElem->next;
2053 }
2054 }
2055 if ((bootIndex==MAX_PATTERN_ENTRIES) && (curElem!=NULL) ) {
2056 status = U_BUFFER_OVERFLOW_ERROR;
2057 }
2058 }
2059
2060 const UnicodeString*
snext(UErrorCode & status)2061 DTSkeletonEnumeration::snext(UErrorCode& status) {
2062 if (U_SUCCESS(status) && pos < fSkeletons->size()) {
2063 return (const UnicodeString*)fSkeletons->elementAt(pos++);
2064 }
2065 return NULL;
2066 }
2067
2068 void
reset(UErrorCode &)2069 DTSkeletonEnumeration::reset(UErrorCode& /*status*/) {
2070 pos=0;
2071 }
2072
2073 int32_t
count(UErrorCode &) const2074 DTSkeletonEnumeration::count(UErrorCode& /*status*/) const {
2075 return (fSkeletons==NULL) ? 0 : fSkeletons->size();
2076 }
2077
2078 UBool
isCanonicalItem(const UnicodeString & item)2079 DTSkeletonEnumeration::isCanonicalItem(const UnicodeString& item) {
2080 if ( item.length() != 1 ) {
2081 return FALSE;
2082 }
2083 for (int32_t i=0; i<UDATPG_FIELD_COUNT; ++i) {
2084 if (item.charAt(0)==Canonical_Items[i]) {
2085 return TRUE;
2086 }
2087 }
2088 return FALSE;
2089 }
2090
~DTSkeletonEnumeration()2091 DTSkeletonEnumeration::~DTSkeletonEnumeration() {
2092 UnicodeString *s;
2093 for (int32_t i=0; i<fSkeletons->size(); ++i) {
2094 if ((s=(UnicodeString *)fSkeletons->elementAt(i))!=NULL) {
2095 delete s;
2096 }
2097 }
2098 delete fSkeletons;
2099 }
2100
DTRedundantEnumeration()2101 DTRedundantEnumeration::DTRedundantEnumeration() {
2102 pos=0;
2103 fPatterns = NULL;
2104 }
2105
2106 void
add(const UnicodeString & pattern,UErrorCode & status)2107 DTRedundantEnumeration::add(const UnicodeString& pattern, UErrorCode& status) {
2108 if (U_FAILURE(status)) return;
2109 if (fPatterns == NULL) {
2110 fPatterns = new UVector(status);
2111 if (U_FAILURE(status)) {
2112 delete fPatterns;
2113 fPatterns = NULL;
2114 return;
2115 }
2116 }
2117 fPatterns->addElement(new UnicodeString(pattern), status);
2118 if (U_FAILURE(status)) {
2119 delete fPatterns;
2120 fPatterns = NULL;
2121 return;
2122 }
2123 }
2124
2125 const UnicodeString*
snext(UErrorCode & status)2126 DTRedundantEnumeration::snext(UErrorCode& status) {
2127 if (U_SUCCESS(status) && pos < fPatterns->size()) {
2128 return (const UnicodeString*)fPatterns->elementAt(pos++);
2129 }
2130 return NULL;
2131 }
2132
2133 void
reset(UErrorCode &)2134 DTRedundantEnumeration::reset(UErrorCode& /*status*/) {
2135 pos=0;
2136 }
2137
2138 int32_t
count(UErrorCode &) const2139 DTRedundantEnumeration::count(UErrorCode& /*status*/) const {
2140 return (fPatterns==NULL) ? 0 : fPatterns->size();
2141 }
2142
2143 UBool
isCanonicalItem(const UnicodeString & item)2144 DTRedundantEnumeration::isCanonicalItem(const UnicodeString& item) {
2145 if ( item.length() != 1 ) {
2146 return FALSE;
2147 }
2148 for (int32_t i=0; i<UDATPG_FIELD_COUNT; ++i) {
2149 if (item.charAt(0)==Canonical_Items[i]) {
2150 return TRUE;
2151 }
2152 }
2153 return FALSE;
2154 }
2155
~DTRedundantEnumeration()2156 DTRedundantEnumeration::~DTRedundantEnumeration() {
2157 UnicodeString *s;
2158 for (int32_t i=0; i<fPatterns->size(); ++i) {
2159 if ((s=(UnicodeString *)fPatterns->elementAt(i))!=NULL) {
2160 delete s;
2161 }
2162 }
2163 delete fPatterns;
2164 }
2165
2166 U_NAMESPACE_END
2167
2168
2169 #endif /* #if !UCONFIG_NO_FORMATTING */
2170
2171 //eof
2172