1 /*
2 *******************************************************************************
3 * Copyright (C) 2007-2010, 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 #if !defined (OS390) && !defined (OS400)
70 extern "C"
71 #endif
72 static int32_t U_CALLCONV
ures_a_codepointSort(const void * context,const void * left,const void * right)73 ures_a_codepointSort(const void *context, const void *left, const void *right) {
74 //CompareContext *cmp=(CompareContext *)context;
75 return u_strcmp(((const UResAEntry *)left)->key,
76 ((const UResAEntry *)right)->key);
77 }
78
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, LOW_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_Q, UDATPG_QUARTER_FIELD, DT_NUMERIC, 1, 2},
148 {CAP_Q, UDATPG_QUARTER_FIELD, DT_SHORT, 3, 0},
149 {CAP_Q, UDATPG_QUARTER_FIELD, DT_LONG, 4, 0},
150 {LOW_Q, UDATPG_QUARTER_FIELD, DT_NUMERIC + DT_DELTA, 1, 2},
151 {LOW_Q, UDATPG_QUARTER_FIELD, DT_SHORT + DT_DELTA, 3, 0},
152 {LOW_Q, UDATPG_QUARTER_FIELD, DT_LONG + DT_DELTA, 4, 0},
153 {CAP_M, UDATPG_MONTH_FIELD, DT_NUMERIC, 1, 2},
154 {CAP_M, UDATPG_MONTH_FIELD, DT_SHORT, 3, 0},
155 {CAP_M, UDATPG_MONTH_FIELD, DT_LONG, 4, 0},
156 {CAP_M, UDATPG_MONTH_FIELD, DT_NARROW, 5, 0},
157 {CAP_L, UDATPG_MONTH_FIELD, DT_NUMERIC + DT_DELTA, 1, 2},
158 {CAP_L, UDATPG_MONTH_FIELD, DT_SHORT - DT_DELTA, 3, 0},
159 {CAP_L, UDATPG_MONTH_FIELD, DT_LONG - DT_DELTA, 4, 0},
160 {CAP_L, UDATPG_MONTH_FIELD, DT_NARROW - DT_DELTA, 5, 0},
161 {LOW_W, UDATPG_WEEK_OF_YEAR_FIELD, DT_NUMERIC, 1, 2},
162 {CAP_W, UDATPG_WEEK_OF_MONTH_FIELD, DT_NUMERIC + DT_DELTA, 1, 0},
163 {LOW_E, UDATPG_WEEKDAY_FIELD, DT_NUMERIC + DT_DELTA, 1, 2},
164 {LOW_E, UDATPG_WEEKDAY_FIELD, DT_SHORT - DT_DELTA, 3, 0},
165 {LOW_E, UDATPG_WEEKDAY_FIELD, DT_LONG - DT_DELTA, 4, 0},
166 {LOW_E, UDATPG_WEEKDAY_FIELD, DT_NARROW - DT_DELTA, 5, 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_D, UDATPG_DAY_FIELD, DT_NUMERIC, 1, 2},
175 {CAP_D, UDATPG_DAY_OF_YEAR_FIELD, DT_NUMERIC + DT_DELTA, 1, 3},
176 {CAP_F, UDATPG_DAY_OF_WEEK_IN_MONTH_FIELD, DT_NUMERIC + 2*DT_DELTA, 1, 0},
177 {LOW_G, UDATPG_DAY_FIELD, DT_NUMERIC + 3*DT_DELTA, 1, 20}, // really internal use, so we don't care
178 {LOW_A, UDATPG_DAYPERIOD_FIELD, DT_SHORT, 1, 0},
179 {CAP_H, UDATPG_HOUR_FIELD, DT_NUMERIC + 10*DT_DELTA, 1, 2}, // 24 hour
180 {LOW_K, UDATPG_HOUR_FIELD, DT_NUMERIC + 11*DT_DELTA, 1, 2},
181 {LOW_H, UDATPG_HOUR_FIELD, DT_NUMERIC, 1, 2}, // 12 hour
182 {LOW_K, UDATPG_HOUR_FIELD, DT_NUMERIC + DT_DELTA, 1, 2},
183 {LOW_M, UDATPG_MINUTE_FIELD, DT_NUMERIC, 1, 2},
184 {LOW_S, UDATPG_SECOND_FIELD, DT_NUMERIC, 1, 2},
185 {CAP_S, UDATPG_FRACTIONAL_SECOND_FIELD, DT_NUMERIC + DT_DELTA, 1, 1000},
186 {CAP_A, UDATPG_SECOND_FIELD, DT_NUMERIC + 2*DT_DELTA, 1, 1000},
187 {LOW_V, UDATPG_ZONE_FIELD, DT_SHORT - 2*DT_DELTA, 1, 0},
188 {LOW_V, UDATPG_ZONE_FIELD, DT_LONG - 2*DT_DELTA, 4, 0},
189 {LOW_Z, UDATPG_ZONE_FIELD, DT_SHORT, 1, 3},
190 {LOW_Z, UDATPG_ZONE_FIELD, DT_LONG, 4, 0},
191 {CAP_Z, UDATPG_ZONE_FIELD, DT_SHORT - DT_DELTA, 1, 3},
192 {CAP_Z, UDATPG_ZONE_FIELD, DT_LONG - DT_DELTA, 4, 0},
193 {CAP_V, UDATPG_ZONE_FIELD, DT_SHORT - DT_DELTA, 1, 3},
194 {CAP_V, UDATPG_ZONE_FIELD, DT_LONG - DT_DELTA, 4, 0},
195 {0, UDATPG_FIELD_COUNT, 0, 0, 0} , // last row of dtTypes[]
196 };
197
198 static const char* const CLDR_FIELD_APPEND[] = {
199 "Era", "Year", "Quarter", "Month", "Week", "*", "Day-Of-Week", "Day", "*", "*", "*",
200 "Hour", "Minute", "Second", "*", "Timezone"
201 };
202
203 static const char* const CLDR_FIELD_NAME[] = {
204 "era", "year", "quarter", "month", "week", "*", "weekday", "day", "*", "*", "dayperiod",
205 "hour", "minute", "second", "*", "zone"
206 };
207
208 static const char* const Resource_Fields[] = {
209 "day", "dayperiod", "era", "hour", "minute", "month", "second", "week",
210 "weekday", "year", "zone", "quarter" };
211
212 // For appendItems
213 static const UChar UDATPG_ItemFormat[]= {0x7B, 0x30, 0x7D, 0x20, 0x251C, 0x7B, 0x32, 0x7D, 0x3A,
214 0x20, 0x7B, 0x31, 0x7D, 0x2524, 0}; // {0} \u251C{2}: {1}\u2524
215
216 static const UChar repeatedPatterns[6]={CAP_G, CAP_E, LOW_Z, LOW_V, CAP_Q, 0}; // "GEzvQ"
217
218 static const char DT_DateTimePatternsTag[]="DateTimePatterns";
219 static const char DT_DateTimeCalendarTag[]="calendar";
220 static const char DT_DateTimeGregorianTag[]="gregorian";
221 static const char DT_DateTimeAppendItemsTag[]="appendItems";
222 static const char DT_DateTimeFieldsTag[]="fields";
223 static const char DT_DateTimeAvailableFormatsTag[]="availableFormats";
224 //static const UnicodeString repeatedPattern=UnicodeString(repeatedPatterns);
225
226 UOBJECT_DEFINE_RTTI_IMPLEMENTATION(DateTimePatternGenerator)
UOBJECT_DEFINE_RTTI_IMPLEMENTATION(DTSkeletonEnumeration)227 UOBJECT_DEFINE_RTTI_IMPLEMENTATION(DTSkeletonEnumeration)
228 UOBJECT_DEFINE_RTTI_IMPLEMENTATION(DTRedundantEnumeration)
229
230 DateTimePatternGenerator* U_EXPORT2
231 DateTimePatternGenerator::createInstance(UErrorCode& status) {
232 return createInstance(Locale::getDefault(), status);
233 }
234
235 DateTimePatternGenerator* U_EXPORT2
createInstance(const Locale & locale,UErrorCode & status)236 DateTimePatternGenerator::createInstance(const Locale& locale, UErrorCode& status) {
237 DateTimePatternGenerator *result = new DateTimePatternGenerator(locale, status);
238 if (result == NULL) {
239 status = U_MEMORY_ALLOCATION_ERROR;
240 }
241 if (U_FAILURE(status)) {
242 delete result;
243 result = NULL;
244 }
245 return result;
246 }
247
248 DateTimePatternGenerator* U_EXPORT2
createEmptyInstance(UErrorCode & status)249 DateTimePatternGenerator::createEmptyInstance(UErrorCode& status) {
250 DateTimePatternGenerator *result = new DateTimePatternGenerator(status);
251 if (result == NULL) {
252 status = U_MEMORY_ALLOCATION_ERROR;
253 }
254 if (U_FAILURE(status)) {
255 delete result;
256 result = NULL;
257 }
258 return result;
259 }
260
DateTimePatternGenerator(UErrorCode & status)261 DateTimePatternGenerator::DateTimePatternGenerator(UErrorCode &status) :
262 skipMatcher(NULL),
263 fAvailableFormatKeyHash(NULL)
264 {
265 fp = new FormatParser();
266 dtMatcher = new DateTimeMatcher();
267 distanceInfo = new DistanceInfo();
268 patternMap = new PatternMap();
269 if (fp == NULL || dtMatcher == NULL || distanceInfo == NULL || patternMap == NULL) {
270 status = U_MEMORY_ALLOCATION_ERROR;
271 }
272 }
273
DateTimePatternGenerator(const Locale & locale,UErrorCode & status)274 DateTimePatternGenerator::DateTimePatternGenerator(const Locale& locale, UErrorCode &status) :
275 skipMatcher(NULL),
276 fAvailableFormatKeyHash(NULL)
277 {
278 fp = new FormatParser();
279 dtMatcher = new DateTimeMatcher();
280 distanceInfo = new DistanceInfo();
281 patternMap = new PatternMap();
282 if (fp == NULL || dtMatcher == NULL || distanceInfo == NULL || patternMap == NULL) {
283 status = U_MEMORY_ALLOCATION_ERROR;
284 }
285 else {
286 initData(locale, status);
287 }
288 }
289
DateTimePatternGenerator(const DateTimePatternGenerator & other)290 DateTimePatternGenerator::DateTimePatternGenerator(const DateTimePatternGenerator& other) :
291 UObject(),
292 skipMatcher(NULL),
293 fAvailableFormatKeyHash(NULL)
294 {
295 fp = new FormatParser();
296 dtMatcher = new DateTimeMatcher();
297 distanceInfo = new DistanceInfo();
298 patternMap = new PatternMap();
299 *this=other;
300 }
301
302 DateTimePatternGenerator&
operator =(const DateTimePatternGenerator & other)303 DateTimePatternGenerator::operator=(const DateTimePatternGenerator& other) {
304 pLocale = other.pLocale;
305 fDefaultHourFormatChar = other.fDefaultHourFormatChar;
306 *fp = *(other.fp);
307 dtMatcher->copyFrom(other.dtMatcher->skeleton);
308 *distanceInfo = *(other.distanceInfo);
309 dateTimeFormat = other.dateTimeFormat;
310 decimal = other.decimal;
311 // NUL-terminate for the C API.
312 dateTimeFormat.getTerminatedBuffer();
313 decimal.getTerminatedBuffer();
314 delete skipMatcher;
315 if ( other.skipMatcher == NULL ) {
316 skipMatcher = NULL;
317 }
318 else {
319 skipMatcher = new DateTimeMatcher(*other.skipMatcher);
320 }
321 for (int32_t i=0; i< UDATPG_FIELD_COUNT; ++i ) {
322 appendItemFormats[i] = other.appendItemFormats[i];
323 appendItemNames[i] = other.appendItemNames[i];
324 // NUL-terminate for the C API.
325 appendItemFormats[i].getTerminatedBuffer();
326 appendItemNames[i].getTerminatedBuffer();
327 }
328 UErrorCode status = U_ZERO_ERROR;
329 patternMap->copyFrom(*other.patternMap, status);
330 copyHashtable(other.fAvailableFormatKeyHash, status);
331 return *this;
332 }
333
334
335 UBool
operator ==(const DateTimePatternGenerator & other) const336 DateTimePatternGenerator::operator==(const DateTimePatternGenerator& other) const {
337 if (this == &other) {
338 return TRUE;
339 }
340 if ((pLocale==other.pLocale) && (patternMap->equals(*other.patternMap)) &&
341 (dateTimeFormat==other.dateTimeFormat) && (decimal==other.decimal)) {
342 for ( int32_t i=0 ; i<UDATPG_FIELD_COUNT; ++i ) {
343 if ((appendItemFormats[i] != other.appendItemFormats[i]) ||
344 (appendItemNames[i] != other.appendItemNames[i]) ) {
345 return FALSE;
346 }
347 }
348 return TRUE;
349 }
350 else {
351 return FALSE;
352 }
353 }
354
355 UBool
operator !=(const DateTimePatternGenerator & other) const356 DateTimePatternGenerator::operator!=(const DateTimePatternGenerator& other) const {
357 return !operator==(other);
358 }
359
~DateTimePatternGenerator()360 DateTimePatternGenerator::~DateTimePatternGenerator() {
361 if (fAvailableFormatKeyHash!=NULL) {
362 delete fAvailableFormatKeyHash;
363 }
364
365 if (fp != NULL) delete fp;
366 if (dtMatcher != NULL) delete dtMatcher;
367 if (distanceInfo != NULL) delete distanceInfo;
368 if (patternMap != NULL) delete patternMap;
369 if (skipMatcher != NULL) delete skipMatcher;
370 }
371
372 void
initData(const Locale & locale,UErrorCode & status)373 DateTimePatternGenerator::initData(const Locale& locale, UErrorCode &status) {
374 //const char *baseLangName = locale.getBaseName(); // unused
375
376 skipMatcher = NULL;
377 fAvailableFormatKeyHash=NULL;
378 addCanonicalItems();
379 addICUPatterns(locale, status);
380 if (U_FAILURE(status)) {
381 return;
382 }
383 addCLDRData(locale, status);
384 setDateTimeFromCalendar(locale, status);
385 setDecimalSymbols(locale, status);
386 } // DateTimePatternGenerator::initData
387
388 UnicodeString
getSkeleton(const UnicodeString & pattern,UErrorCode &)389 DateTimePatternGenerator::getSkeleton(const UnicodeString& pattern, UErrorCode&
390 /*status*/) {
391 dtMatcher->set(pattern, fp);
392 return dtMatcher->getSkeletonPtr()->getSkeleton();
393 }
394
395 UnicodeString
getBaseSkeleton(const UnicodeString & pattern,UErrorCode &)396 DateTimePatternGenerator::getBaseSkeleton(const UnicodeString& pattern, UErrorCode& /*status*/) {
397 dtMatcher->set(pattern, fp);
398 return dtMatcher->getSkeletonPtr()->getBaseSkeleton();
399 }
400
401 void
addICUPatterns(const Locale & locale,UErrorCode & status)402 DateTimePatternGenerator::addICUPatterns(const Locale& locale, UErrorCode& status) {
403 UnicodeString dfPattern;
404 UnicodeString conflictingString;
405 UDateTimePatternConflict conflictingStatus;
406 DateFormat* df;
407
408 if (U_FAILURE(status)) {
409 return;
410 }
411
412 // Load with ICU patterns
413 for (int32_t i=DateFormat::kFull; i<=DateFormat::kShort; i++) {
414 DateFormat::EStyle style = (DateFormat::EStyle)i;
415 df = DateFormat::createDateInstance(style, locale);
416 if (df != NULL && df->getDynamicClassID() == SimpleDateFormat::getStaticClassID()) {
417 SimpleDateFormat* sdf = (SimpleDateFormat*)df;
418 conflictingStatus = addPattern(sdf->toPattern(dfPattern), FALSE, conflictingString, status);
419 }
420 // TODO Maybe we should return an error when the date format isn't simple.
421 delete df;
422 if (U_FAILURE(status)) {
423 return;
424 }
425
426 df = DateFormat::createTimeInstance(style, locale);
427 if (df != NULL && df->getDynamicClassID() == SimpleDateFormat::getStaticClassID()) {
428 SimpleDateFormat* sdf = (SimpleDateFormat*)df;
429 conflictingStatus = addPattern(sdf->toPattern(dfPattern), FALSE, conflictingString, status);
430 // HACK for hh:ss
431 if ( i==DateFormat::kMedium ) {
432 hackPattern = dfPattern;
433 }
434 }
435 // TODO Maybe we should return an error when the date format isn't simple.
436 delete df;
437 if (U_FAILURE(status)) {
438 return;
439 }
440 }
441 }
442
443 void
hackTimes(const UnicodeString & hackPattern,UErrorCode & status)444 DateTimePatternGenerator::hackTimes(const UnicodeString& hackPattern, UErrorCode& status) {
445 UDateTimePatternConflict conflictingStatus;
446 UnicodeString conflictingString;
447
448 fp->set(hackPattern);
449 UnicodeString mmss;
450 UBool gotMm=FALSE;
451 for (int32_t i=0; i<fp->itemNumber; ++i) {
452 UnicodeString field = fp->items[i];
453 if ( fp->isQuoteLiteral(field) ) {
454 if ( gotMm ) {
455 UnicodeString quoteLiteral;
456 fp->getQuoteLiteral(quoteLiteral, &i);
457 mmss += quoteLiteral;
458 }
459 }
460 else {
461 if (fp->isPatternSeparator(field) && gotMm) {
462 mmss+=field;
463 }
464 else {
465 UChar ch=field.charAt(0);
466 if (ch==LOW_M) {
467 gotMm=TRUE;
468 mmss+=field;
469 }
470 else {
471 if (ch==LOW_S) {
472 if (!gotMm) {
473 break;
474 }
475 mmss+= field;
476 conflictingStatus = addPattern(mmss, FALSE, conflictingString, status);
477 break;
478 }
479 else {
480 if (gotMm || ch==LOW_Z || ch==CAP_Z || ch==LOW_V || ch==CAP_V) {
481 break;
482 }
483 }
484 }
485 }
486 }
487 }
488 }
489
490 #define ULOC_LOCALE_IDENTIFIER_CAPACITY (ULOC_FULLNAME_CAPACITY + 1 + ULOC_KEYWORD_AND_VALUES_CAPACITY)
491
492 static const UChar hourFormatChars[] = { CAP_H, LOW_H, CAP_K, LOW_K, 0 }; // HhKk, the hour format characters
493
494 void
addCLDRData(const Locale & locale,UErrorCode & err)495 DateTimePatternGenerator::addCLDRData(const Locale& locale, UErrorCode& err) {
496 UResourceBundle *rb, *calTypeBundle, *calBundle;
497 UResourceBundle *patBundle, *fieldBundle, *fBundle;
498 UnicodeString rbPattern, value, field;
499 UnicodeString conflictingPattern;
500 UDateTimePatternConflict conflictingStatus;
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 err = U_ZERO_ERROR;
611 initHashtable(err);
612 patBundle = ures_getByKeyWithFallback(calTypeBundle, DT_DateTimeAvailableFormatsTag, NULL, &err);
613 if (U_SUCCESS(err)) {
614 int32_t numberKeys = ures_getSize(patBundle);
615 int32_t len;
616 const UChar *retPattern;
617 key=NULL;
618 #if defined(U_USE_ASCII_BUNDLE_ITERATOR)
619 UResourceBundleAIterator aiter;
620 ures_a_open(&aiter, patBundle, &err);
621 #endif
622 for(i=0; i<numberKeys; ++i) {
623 #if defined(U_USE_ASCII_BUNDLE_ITERATOR)
624 retPattern=ures_a_getNextString(&aiter, &len, &key, &err);
625 #else
626 retPattern=ures_getNextString(patBundle, &len, &key, &err);
627 #endif
628 UnicodeString format=UnicodeString(retPattern);
629 UnicodeString retKey=UnicodeString(key, -1, US_INV);
630 setAvailableFormat(retKey, err);
631 // Add pattern with its associated skeleton. Override any duplicate derived from std patterns,
632 // but not a previous availableFormats entry:
633 conflictingStatus = addPatternWithSkeleton(format, &retKey, TRUE, conflictingPattern, err);
634 }
635 #if defined(U_USE_ASCII_BUNDLE_ITERATOR)
636 ures_a_close(&aiter);
637 #endif
638 }
639 ures_close(patBundle);
640 ures_close(calTypeBundle);
641 ures_close(calBundle);
642 ures_close(rb);
643
644 err = U_ZERO_ERROR;
645 char parentLocale[50];
646 int32_t localeNameLen=0;
647 uprv_strcpy(parentLocale, curLocaleName);
648 while((localeNameLen=uloc_getParent(parentLocale, parentLocale, 50, &err))>=0 ) {
649 rb = ures_open(NULL, parentLocale, &err);
650 curLocaleName=ures_getLocaleByType(rb, ULOC_ACTUAL_LOCALE, &err);
651 uprv_strcpy(parentLocale, curLocaleName);
652 calBundle = ures_getByKey(rb, DT_DateTimeCalendarTag, NULL, &err);
653 calTypeBundle = ures_getByKey(calBundle, calendarTypeToUse, NULL, &err);
654 patBundle = ures_getByKeyWithFallback(calTypeBundle, DT_DateTimeAvailableFormatsTag, NULL, &err);
655 if (U_SUCCESS(err)) {
656 int32_t numberKeys = ures_getSize(patBundle);
657 int32_t len;
658 const UChar *retPattern;
659 key=NULL;
660 #if defined(U_USE_ASCII_BUNDLE_ITERATOR)
661 UResourceBundleAIterator aiter;
662 ures_a_open(&aiter, patBundle, &err);
663 #endif
664 for(i=0; i<numberKeys; ++i) {
665 #if defined(U_USE_ASCII_BUNDLE_ITERATOR)
666 retPattern=ures_a_getNextString(&aiter, &len, &key, &err);
667 #else
668 retPattern=ures_getNextString(patBundle, &len, &key, &err);
669 #endif
670 UnicodeString format=UnicodeString(retPattern);
671 UnicodeString retKey=UnicodeString(key, -1, US_INV);
672 if ( !isAvailableFormatSet(retKey) ) {
673 setAvailableFormat(retKey, err);
674 // Add pattern with its associated skeleton. Override any duplicate derived from std patterns,
675 // but not a previous availableFormats entry:
676 conflictingStatus = addPatternWithSkeleton(format, &retKey, TRUE, conflictingPattern, err);
677 }
678 }
679 #if defined(U_USE_ASCII_BUNDLE_ITERATOR)
680 ures_a_close(&aiter);
681 #endif
682 }
683 err = U_ZERO_ERROR; // reset; if this locale lacks the necessary data, need to keep checking up to root.
684 ures_close(patBundle);
685 ures_close(calTypeBundle);
686 ures_close(calBundle);
687 ures_close(rb);
688 if (localeNameLen==0) {
689 break;
690 }
691 }
692
693 if (hackPattern.length()>0) {
694 hackTimes(hackPattern, err);
695 }
696 }
697
698 void
initHashtable(UErrorCode & err)699 DateTimePatternGenerator::initHashtable(UErrorCode& err) {
700 if (fAvailableFormatKeyHash!=NULL) {
701 return;
702 }
703 if ((fAvailableFormatKeyHash = new Hashtable(FALSE, err))==NULL) {
704 err=U_MEMORY_ALLOCATION_ERROR;
705 return;
706 }
707 }
708
709
710 void
setAppendItemFormat(UDateTimePatternField field,const UnicodeString & value)711 DateTimePatternGenerator::setAppendItemFormat(UDateTimePatternField field, const UnicodeString& value) {
712 appendItemFormats[field] = value;
713 // NUL-terminate for the C API.
714 appendItemFormats[field].getTerminatedBuffer();
715 }
716
717 const UnicodeString&
getAppendItemFormat(UDateTimePatternField field) const718 DateTimePatternGenerator::getAppendItemFormat(UDateTimePatternField field) const {
719 return appendItemFormats[field];
720 }
721
722 void
setAppendItemName(UDateTimePatternField field,const UnicodeString & value)723 DateTimePatternGenerator::setAppendItemName(UDateTimePatternField field, const UnicodeString& value) {
724 appendItemNames[field] = value;
725 // NUL-terminate for the C API.
726 appendItemNames[field].getTerminatedBuffer();
727 }
728
729 const UnicodeString&
getAppendItemName(UDateTimePatternField field) const730 DateTimePatternGenerator:: getAppendItemName(UDateTimePatternField field) const {
731 return appendItemNames[field];
732 }
733
734 void
getAppendName(UDateTimePatternField field,UnicodeString & value)735 DateTimePatternGenerator::getAppendName(UDateTimePatternField field, UnicodeString& value) {
736 value = SINGLE_QUOTE;
737 value += appendItemNames[field];
738 value += SINGLE_QUOTE;
739 }
740
741 UnicodeString
getBestPattern(const UnicodeString & patternForm,UErrorCode & status)742 DateTimePatternGenerator::getBestPattern(const UnicodeString& patternForm, UErrorCode& status) {
743 return getBestPattern(patternForm, UDATPG_MATCH_NO_OPTIONS, status);
744 }
745
746 UnicodeString
getBestPattern(const UnicodeString & patternForm,UDateTimePatternMatchOptions options,UErrorCode & status)747 DateTimePatternGenerator::getBestPattern(const UnicodeString& patternForm, UDateTimePatternMatchOptions options, UErrorCode& status) {
748 const UnicodeString *bestPattern=NULL;
749 UnicodeString dtFormat;
750 UnicodeString resultPattern;
751
752 int32_t dateMask=(1<<UDATPG_DAYPERIOD_FIELD) - 1;
753 int32_t timeMask=(1<<UDATPG_FIELD_COUNT) - 1 - dateMask;
754
755 UnicodeString patternFormCopy = UnicodeString(patternForm);
756 patternFormCopy.findAndReplace(UnicodeString(LOW_J), UnicodeString(fDefaultHourFormatChar));
757
758 resultPattern.remove();
759 dtMatcher->set(patternFormCopy, fp);
760 const PtnSkeleton* specifiedSkeleton=NULL;
761 bestPattern=getBestRaw(*dtMatcher, -1, distanceInfo, &specifiedSkeleton);
762 if ( distanceInfo->missingFieldMask==0 && distanceInfo->extraFieldMask==0 ) {
763 resultPattern = adjustFieldTypes(*bestPattern, specifiedSkeleton, FALSE, options);
764
765 return resultPattern;
766 }
767 int32_t neededFields = dtMatcher->getFieldMask();
768 UnicodeString datePattern=getBestAppending(neededFields & dateMask, options);
769 UnicodeString timePattern=getBestAppending(neededFields & timeMask, options);
770 if (datePattern.length()==0) {
771 if (timePattern.length()==0) {
772 resultPattern.remove();
773 }
774 else {
775 return timePattern;
776 }
777 }
778 if (timePattern.length()==0) {
779 return datePattern;
780 }
781 resultPattern.remove();
782 status = U_ZERO_ERROR;
783 dtFormat=getDateTimeFormat();
784 Formattable dateTimeObject[] = { timePattern, datePattern };
785 resultPattern = MessageFormat::format(dtFormat, dateTimeObject, 2, resultPattern, status );
786 return resultPattern;
787 }
788
789 UnicodeString
replaceFieldTypes(const UnicodeString & pattern,const UnicodeString & skeleton,UErrorCode & status)790 DateTimePatternGenerator::replaceFieldTypes(const UnicodeString& pattern,
791 const UnicodeString& skeleton,
792 UErrorCode& status) {
793 return replaceFieldTypes(pattern, skeleton, UDATPG_MATCH_NO_OPTIONS, status);
794 }
795
796 UnicodeString
replaceFieldTypes(const UnicodeString & pattern,const UnicodeString & skeleton,UDateTimePatternMatchOptions options,UErrorCode &)797 DateTimePatternGenerator::replaceFieldTypes(const UnicodeString& pattern,
798 const UnicodeString& skeleton,
799 UDateTimePatternMatchOptions options,
800 UErrorCode& /*status*/) {
801 dtMatcher->set(skeleton, fp);
802 UnicodeString result = adjustFieldTypes(pattern, NULL, FALSE, options);
803 return result;
804 }
805
806 void
setDecimal(const UnicodeString & newDecimal)807 DateTimePatternGenerator::setDecimal(const UnicodeString& newDecimal) {
808 this->decimal = newDecimal;
809 // NUL-terminate for the C API.
810 this->decimal.getTerminatedBuffer();
811 }
812
813 const UnicodeString&
getDecimal() const814 DateTimePatternGenerator::getDecimal() const {
815 return decimal;
816 }
817
818 void
addCanonicalItems()819 DateTimePatternGenerator::addCanonicalItems() {
820 UnicodeString conflictingPattern;
821 UDateTimePatternConflict conflictingStatus;
822 UErrorCode status = U_ZERO_ERROR;
823
824 for (int32_t i=0; i<UDATPG_FIELD_COUNT; i++) {
825 conflictingStatus = addPattern(UnicodeString(Canonical_Items[i]), FALSE, conflictingPattern, status);
826 }
827 }
828
829 void
setDateTimeFormat(const UnicodeString & dtFormat)830 DateTimePatternGenerator::setDateTimeFormat(const UnicodeString& dtFormat) {
831 dateTimeFormat = dtFormat;
832 // NUL-terminate for the C API.
833 dateTimeFormat.getTerminatedBuffer();
834 }
835
836 const UnicodeString&
getDateTimeFormat() const837 DateTimePatternGenerator::getDateTimeFormat() const {
838 return dateTimeFormat;
839 }
840
841 void
setDateTimeFromCalendar(const Locale & locale,UErrorCode & status)842 DateTimePatternGenerator::setDateTimeFromCalendar(const Locale& locale, UErrorCode& status) {
843 const UChar *resStr;
844 int32_t resStrLen = 0;
845
846 Calendar* fCalendar = Calendar::createInstance(locale, status);
847 CalendarData calData(locale, fCalendar?fCalendar->getType():NULL, status);
848 UResourceBundle *dateTimePatterns = calData.getByKey(DT_DateTimePatternsTag, status);
849 if (U_FAILURE(status)) return;
850
851 if (ures_getSize(dateTimePatterns) <= DateFormat::kDateTime)
852 {
853 status = U_INVALID_FORMAT_ERROR;
854 return;
855 }
856 resStr = ures_getStringByIndex(dateTimePatterns, (int32_t)DateFormat::kDateTime, &resStrLen, &status);
857 setDateTimeFormat(UnicodeString(TRUE, resStr, resStrLen));
858
859 delete fCalendar;
860 }
861
862 void
setDecimalSymbols(const Locale & locale,UErrorCode & status)863 DateTimePatternGenerator::setDecimalSymbols(const Locale& locale, UErrorCode& status) {
864 DecimalFormatSymbols dfs = DecimalFormatSymbols(locale, status);
865 if(U_SUCCESS(status)) {
866 decimal = dfs.getSymbol(DecimalFormatSymbols::kDecimalSeparatorSymbol);
867 // NUL-terminate for the C API.
868 decimal.getTerminatedBuffer();
869 }
870 }
871
872 UDateTimePatternConflict
addPattern(const UnicodeString & pattern,UBool override,UnicodeString & conflictingPattern,UErrorCode & status)873 DateTimePatternGenerator::addPattern(
874 const UnicodeString& pattern,
875 UBool override,
876 UnicodeString &conflictingPattern,
877 UErrorCode& status)
878 {
879 return addPatternWithSkeleton(pattern, NULL, override, conflictingPattern, status);
880 }
881
882 // For DateTimePatternGenerator::addPatternWithSkeleton -
883 // If skeletonToUse is specified, then an availableFormats entry is being added. In this case:
884 // 1. We pass that skeleton to matcher.set instead of having it derive a skeleton from the pattern.
885 // 2. If the new entry's skeleton or basePattern does match an existing entry but that entry also had a skeleton specified
886 // (i.e. it was also from availableFormats), then the new entry does not override it regardless of the value of the override
887 // parameter. This prevents later availableFormats entries from a parent locale overriding earlier ones from the actual
888 // specified locale. However, availableFormats entries *should* override entries with matching skeleton whose skeleton was
889 // derived (i.e. entries derived from the standard date/time patters for the specified locale).
890 // 3. When adding the pattern (patternMap->add), we set a new boolean to indicate that the added entry had a
891 // specified skeleton (which sets a new field in the PtnElem in the PatternMap).
892 UDateTimePatternConflict
addPatternWithSkeleton(const UnicodeString & pattern,const UnicodeString * skeletonToUse,UBool override,UnicodeString & conflictingPattern,UErrorCode & status)893 DateTimePatternGenerator::addPatternWithSkeleton(
894 const UnicodeString& pattern,
895 const UnicodeString* skeletonToUse,
896 UBool override,
897 UnicodeString& conflictingPattern,
898 UErrorCode& status)
899 {
900
901 UnicodeString basePattern;
902 PtnSkeleton skeleton;
903 UDateTimePatternConflict conflictingStatus = UDATPG_NO_CONFLICT;
904
905 DateTimeMatcher matcher;
906 if ( skeletonToUse == NULL ) {
907 matcher.set(pattern, fp, skeleton);
908 matcher.getBasePattern(basePattern);
909 } else {
910 matcher.set(*skeletonToUse, fp, skeleton); // this still trims skeleton fields to max len 3, may need to change it.
911 matcher.getBasePattern(basePattern); // or perhaps instead: basePattern = *skeletonToUse;
912 }
913 UBool entryHadSpecifiedSkeleton;
914 const UnicodeString *duplicatePattern = patternMap->getPatternFromBasePattern(basePattern, entryHadSpecifiedSkeleton);
915 if (duplicatePattern != NULL ) {
916 conflictingStatus = UDATPG_BASE_CONFLICT;
917 conflictingPattern = *duplicatePattern;
918 if (!override || (skeletonToUse != NULL && entryHadSpecifiedSkeleton)) {
919 return conflictingStatus;
920 }
921 }
922 const PtnSkeleton* entrySpecifiedSkeleton = NULL;
923 duplicatePattern = patternMap->getPatternFromSkeleton(skeleton, &entrySpecifiedSkeleton);
924 if (duplicatePattern != NULL ) {
925 conflictingStatus = UDATPG_CONFLICT;
926 conflictingPattern = *duplicatePattern;
927 if (!override || (skeletonToUse != NULL && entrySpecifiedSkeleton != NULL)) {
928 return conflictingStatus;
929 }
930 }
931 patternMap->add(basePattern, skeleton, pattern, skeletonToUse != NULL, status);
932 if(U_FAILURE(status)) {
933 return conflictingStatus;
934 }
935
936 return UDATPG_NO_CONFLICT;
937 }
938
939
940 UDateTimePatternField
getAppendFormatNumber(const char * field) const941 DateTimePatternGenerator::getAppendFormatNumber(const char* field) const {
942 for (int32_t i=0; i<UDATPG_FIELD_COUNT; ++i ) {
943 if (uprv_strcmp(CLDR_FIELD_APPEND[i], field)==0) {
944 return (UDateTimePatternField)i;
945 }
946 }
947 return UDATPG_FIELD_COUNT;
948 }
949
950 UDateTimePatternField
getAppendNameNumber(const char * field) const951 DateTimePatternGenerator::getAppendNameNumber(const char* field) const {
952 for (int32_t i=0; i<UDATPG_FIELD_COUNT; ++i ) {
953 if (uprv_strcmp(CLDR_FIELD_NAME[i],field)==0) {
954 return (UDateTimePatternField)i;
955 }
956 }
957 return UDATPG_FIELD_COUNT;
958 }
959
960 const UnicodeString*
getBestRaw(DateTimeMatcher & source,int32_t includeMask,DistanceInfo * missingFields,const PtnSkeleton ** specifiedSkeletonPtr)961 DateTimePatternGenerator::getBestRaw(DateTimeMatcher& source,
962 int32_t includeMask,
963 DistanceInfo* missingFields,
964 const PtnSkeleton** specifiedSkeletonPtr) {
965 int32_t bestDistance = 0x7fffffff;
966 DistanceInfo tempInfo;
967 const UnicodeString *bestPattern=NULL;
968 const PtnSkeleton* specifiedSkeleton=NULL;
969
970 PatternMapIterator it;
971 for (it.set(*patternMap); it.hasNext(); ) {
972 DateTimeMatcher trial = it.next();
973 if (trial.equals(skipMatcher)) {
974 continue;
975 }
976 int32_t distance=source.getDistance(trial, includeMask, tempInfo);
977 if (distance<bestDistance) {
978 bestDistance=distance;
979 bestPattern=patternMap->getPatternFromSkeleton(*trial.getSkeletonPtr(), &specifiedSkeleton);
980 missingFields->setTo(tempInfo);
981 if (distance==0) {
982 break;
983 }
984 }
985 }
986
987 // If the best raw match had a specified skeleton and that skeleton was requested by the caller,
988 // then return it too. This generally happens when the caller needs to pass that skeleton
989 // through to adjustFieldTypes so the latter can do a better job.
990 if (bestPattern && specifiedSkeletonPtr) {
991 *specifiedSkeletonPtr = specifiedSkeleton;
992 }
993 return bestPattern;
994 }
995
996 UnicodeString
adjustFieldTypes(const UnicodeString & pattern,const PtnSkeleton * specifiedSkeleton,UBool fixFractionalSeconds,UDateTimePatternMatchOptions options)997 DateTimePatternGenerator::adjustFieldTypes(const UnicodeString& pattern,
998 const PtnSkeleton* specifiedSkeleton,
999 UBool fixFractionalSeconds,
1000 UDateTimePatternMatchOptions options) {
1001 UnicodeString newPattern;
1002 fp->set(pattern);
1003 for (int32_t i=0; i < fp->itemNumber; i++) {
1004 UnicodeString field = fp->items[i];
1005 if ( fp->isQuoteLiteral(field) ) {
1006
1007 UnicodeString quoteLiteral;
1008 fp->getQuoteLiteral(quoteLiteral, &i);
1009 newPattern += quoteLiteral;
1010 }
1011 else {
1012 if (fp->isPatternSeparator(field)) {
1013 newPattern+=field;
1014 continue;
1015 }
1016 int32_t canonicalIndex = fp->getCanonicalIndex(field);
1017 if (canonicalIndex < 0) {
1018 newPattern+=field;
1019 continue; // don't adjust
1020 }
1021 const dtTypeElem *row = &dtTypes[canonicalIndex];
1022 int32_t typeValue = row->field;
1023 if (fixFractionalSeconds && typeValue == UDATPG_SECOND_FIELD) {
1024 UnicodeString newField=dtMatcher->skeleton.original[UDATPG_FRACTIONAL_SECOND_FIELD];
1025 field = field + decimal + newField;
1026 }
1027 else {
1028 if (dtMatcher->skeleton.type[typeValue]!=0) {
1029 // Here:
1030 // - "reqField" is the field from the originally requested skeleton, with length
1031 // "reqFieldLen".
1032 // - "field" is the field from the found pattern.
1033 //
1034 // The adjusted field should consist of characters from the originally requested
1035 // skeleton, except in the case of UDATPG_HOUR_FIELD or UDATPG_MONTH_FIELD, in
1036 // which case it should consist of characters from the found pattern.
1037 //
1038 // The length of the adjusted field (adjFieldLen) should match that in the originally
1039 // requested skeleton, except that in the following cases the length of the adjusted field
1040 // should match that in the found pattern (i.e. the length of this pattern field should
1041 // not be adjusted):
1042 // 1. typeValue is UDATPG_HOUR_FIELD/MINUTE/SECOND and the corresponding bit in options is
1043 // not set (ticket #7180). Note, we may want to implement a similar change for other
1044 // numeric fields (MM, dd, etc.) so the default behavior is to get locale preference for
1045 // field length, but options bits can be used to override this.
1046 // 2. There is a specified skeleton for the found pattern and one of the following is true:
1047 // a) The length of the field in the skeleton (skelFieldLen) is equal to reqFieldLen.
1048 // b) The pattern field is numeric and the skeleton field is not, or vice versa.
1049
1050 UnicodeString reqField = dtMatcher->skeleton.original[typeValue];
1051 int32_t reqFieldLen = reqField.length();
1052 int32_t adjFieldLen = reqFieldLen;
1053 if ( (typeValue==UDATPG_HOUR_FIELD && (options & UDATPG_MATCH_HOUR_FIELD_LENGTH)==0) ||
1054 (typeValue==UDATPG_MINUTE_FIELD && (options & UDATPG_MATCH_MINUTE_FIELD_LENGTH)==0) ||
1055 (typeValue==UDATPG_SECOND_FIELD && (options & UDATPG_MATCH_SECOND_FIELD_LENGTH)==0) ) {
1056 adjFieldLen = field.length();
1057 } else if (specifiedSkeleton) {
1058 UnicodeString skelField = specifiedSkeleton->original[typeValue];
1059 int32_t skelFieldLen = skelField.length();
1060 UBool patFieldIsNumeric = (row->type > 0);
1061 UBool skelFieldIsNumeric = (specifiedSkeleton->type[typeValue] > 0);
1062 if (skelFieldLen == reqFieldLen || (patFieldIsNumeric && !skelFieldIsNumeric) || (skelFieldIsNumeric && !patFieldIsNumeric)) {
1063 // don't adjust the field length in the found pattern
1064 adjFieldLen = field.length();
1065 }
1066 }
1067 UChar c = (typeValue!= UDATPG_HOUR_FIELD && typeValue!= UDATPG_MONTH_FIELD)? reqField.charAt(0): field.charAt(0);
1068 field.remove();
1069 for (int32_t i=adjFieldLen; i>0; --i) {
1070 field+=c;
1071 }
1072 }
1073 newPattern+=field;
1074 }
1075 }
1076 }
1077 return newPattern;
1078 }
1079
1080 UnicodeString
getBestAppending(int32_t missingFields,UDateTimePatternMatchOptions options)1081 DateTimePatternGenerator::getBestAppending(int32_t missingFields, UDateTimePatternMatchOptions options) {
1082 UnicodeString resultPattern, tempPattern, formattedPattern;
1083 UErrorCode err=U_ZERO_ERROR;
1084 int32_t lastMissingFieldMask=0;
1085 if (missingFields!=0) {
1086 resultPattern=UnicodeString();
1087 const PtnSkeleton* specifiedSkeleton=NULL;
1088 tempPattern = *getBestRaw(*dtMatcher, missingFields, distanceInfo, &specifiedSkeleton);
1089 resultPattern = adjustFieldTypes(tempPattern, specifiedSkeleton, FALSE, options);
1090 if ( distanceInfo->missingFieldMask==0 ) {
1091 return resultPattern;
1092 }
1093 while (distanceInfo->missingFieldMask!=0) { // precondition: EVERY single field must work!
1094 if ( lastMissingFieldMask == distanceInfo->missingFieldMask ) {
1095 break; // cannot find the proper missing field
1096 }
1097 if (((distanceInfo->missingFieldMask & UDATPG_SECOND_AND_FRACTIONAL_MASK)==UDATPG_FRACTIONAL_MASK) &&
1098 ((missingFields & UDATPG_SECOND_AND_FRACTIONAL_MASK) == UDATPG_SECOND_AND_FRACTIONAL_MASK)) {
1099 resultPattern = adjustFieldTypes(resultPattern, specifiedSkeleton, FALSE, options);
1100 //resultPattern = tempPattern;
1101 distanceInfo->missingFieldMask &= ~UDATPG_FRACTIONAL_MASK;
1102 continue;
1103 }
1104 int32_t startingMask = distanceInfo->missingFieldMask;
1105 tempPattern = *getBestRaw(*dtMatcher, distanceInfo->missingFieldMask, distanceInfo, &specifiedSkeleton);
1106 tempPattern = adjustFieldTypes(tempPattern, specifiedSkeleton, FALSE, options);
1107 int32_t foundMask=startingMask& ~distanceInfo->missingFieldMask;
1108 int32_t topField=getTopBitNumber(foundMask);
1109 UnicodeString appendName;
1110 getAppendName((UDateTimePatternField)topField, appendName);
1111 const Formattable formatPattern[] = {
1112 resultPattern,
1113 tempPattern,
1114 appendName
1115 };
1116 UnicodeString emptyStr;
1117 formattedPattern = MessageFormat::format(appendItemFormats[topField], formatPattern, 3, emptyStr, err);
1118 lastMissingFieldMask = distanceInfo->missingFieldMask;
1119 }
1120 }
1121 return formattedPattern;
1122 }
1123
1124 int32_t
getTopBitNumber(int32_t foundMask)1125 DateTimePatternGenerator::getTopBitNumber(int32_t foundMask) {
1126 if ( foundMask==0 ) {
1127 return 0;
1128 }
1129 int32_t i=0;
1130 while (foundMask!=0) {
1131 foundMask >>=1;
1132 ++i;
1133 }
1134 if (i-1 >UDATPG_ZONE_FIELD) {
1135 return UDATPG_ZONE_FIELD;
1136 }
1137 else
1138 return i-1;
1139 }
1140
1141 void
setAvailableFormat(const UnicodeString & key,UErrorCode & err)1142 DateTimePatternGenerator::setAvailableFormat(const UnicodeString &key, UErrorCode& err)
1143 {
1144 fAvailableFormatKeyHash->puti(key, 1, err);
1145 }
1146
1147 UBool
isAvailableFormatSet(const UnicodeString & key) const1148 DateTimePatternGenerator::isAvailableFormatSet(const UnicodeString &key) const {
1149 return (UBool)(fAvailableFormatKeyHash->geti(key) == 1);
1150 }
1151
1152 void
copyHashtable(Hashtable * other,UErrorCode & status)1153 DateTimePatternGenerator::copyHashtable(Hashtable *other, UErrorCode &status) {
1154
1155 if (other == NULL) {
1156 return;
1157 }
1158 if (fAvailableFormatKeyHash != NULL) {
1159 delete fAvailableFormatKeyHash;
1160 fAvailableFormatKeyHash = NULL;
1161 }
1162 initHashtable(status);
1163 if(U_FAILURE(status)){
1164 return;
1165 }
1166 int32_t pos = -1;
1167 const UHashElement* elem = NULL;
1168 // walk through the hash table and create a deep clone
1169 while((elem = other->nextElement(pos))!= NULL){
1170 const UHashTok otherKeyTok = elem->key;
1171 UnicodeString* otherKey = (UnicodeString*)otherKeyTok.pointer;
1172 fAvailableFormatKeyHash->puti(*otherKey, 1, status);
1173 if(U_FAILURE(status)){
1174 return;
1175 }
1176 }
1177 }
1178
1179 StringEnumeration*
getSkeletons(UErrorCode & status) const1180 DateTimePatternGenerator::getSkeletons(UErrorCode& status) const {
1181 StringEnumeration* skeletonEnumerator = new DTSkeletonEnumeration(*patternMap, DT_SKELETON, status);
1182 return skeletonEnumerator;
1183 }
1184
1185 const UnicodeString&
getPatternForSkeleton(const UnicodeString & skeleton) const1186 DateTimePatternGenerator::getPatternForSkeleton(const UnicodeString& skeleton) const {
1187 PtnElem *curElem;
1188
1189 if (skeleton.length() ==0) {
1190 return emptyString;
1191 }
1192 curElem = patternMap->getHeader(skeleton.charAt(0));
1193 while ( curElem != NULL ) {
1194 if ( curElem->skeleton->getSkeleton()==skeleton ) {
1195 return curElem->pattern;
1196 }
1197 curElem=curElem->next;
1198 }
1199 return emptyString;
1200 }
1201
1202 StringEnumeration*
getBaseSkeletons(UErrorCode & status) const1203 DateTimePatternGenerator::getBaseSkeletons(UErrorCode& status) const {
1204 StringEnumeration* baseSkeletonEnumerator = new DTSkeletonEnumeration(*patternMap, DT_BASESKELETON, status);
1205 return baseSkeletonEnumerator;
1206 }
1207
1208 StringEnumeration*
getRedundants(UErrorCode & status)1209 DateTimePatternGenerator::getRedundants(UErrorCode& status) {
1210 StringEnumeration* output = new DTRedundantEnumeration();
1211 const UnicodeString *pattern;
1212 PatternMapIterator it;
1213 for (it.set(*patternMap); it.hasNext(); ) {
1214 DateTimeMatcher current = it.next();
1215 pattern = patternMap->getPatternFromSkeleton(*(it.getSkeleton()));
1216 if ( isCanonicalItem(*pattern) ) {
1217 continue;
1218 }
1219 if ( skipMatcher == NULL ) {
1220 skipMatcher = new DateTimeMatcher(current);
1221 }
1222 else {
1223 *skipMatcher = current;
1224 }
1225 UnicodeString trial = getBestPattern(current.getPattern(), status);
1226 if (trial == *pattern) {
1227 ((DTRedundantEnumeration *)output)->add(*pattern, status);
1228 }
1229 if (current.equals(skipMatcher)) {
1230 continue;
1231 }
1232 }
1233 return output;
1234 }
1235
1236 UBool
isCanonicalItem(const UnicodeString & item) const1237 DateTimePatternGenerator::isCanonicalItem(const UnicodeString& item) const {
1238 if ( item.length() != 1 ) {
1239 return FALSE;
1240 }
1241 for (int32_t i=0; i<UDATPG_FIELD_COUNT; ++i) {
1242 if (item.charAt(0)==Canonical_Items[i]) {
1243 return TRUE;
1244 }
1245 }
1246 return FALSE;
1247 }
1248
1249
1250 DateTimePatternGenerator*
clone() const1251 DateTimePatternGenerator::clone() const {
1252 return new DateTimePatternGenerator(*this);
1253 }
1254
PatternMap()1255 PatternMap::PatternMap() {
1256 for (int32_t i=0; i < MAX_PATTERN_ENTRIES; ++i ) {
1257 boot[i]=NULL;
1258 }
1259 isDupAllowed = TRUE;
1260 }
1261
1262 void
copyFrom(const PatternMap & other,UErrorCode & status)1263 PatternMap::copyFrom(const PatternMap& other, UErrorCode& status) {
1264 this->isDupAllowed = other.isDupAllowed;
1265 for (int32_t bootIndex=0; bootIndex<MAX_PATTERN_ENTRIES; ++bootIndex ) {
1266 PtnElem *curElem, *otherElem, *prevElem=NULL;
1267 otherElem = other.boot[bootIndex];
1268 while (otherElem!=NULL) {
1269 if ((curElem = new PtnElem(otherElem->basePattern, otherElem->pattern))==NULL) {
1270 // out of memory
1271 status = U_MEMORY_ALLOCATION_ERROR;
1272 return;
1273 }
1274 if ( this->boot[bootIndex]== NULL ) {
1275 this->boot[bootIndex] = curElem;
1276 }
1277 if ((curElem->skeleton=new PtnSkeleton(*(otherElem->skeleton))) == NULL ) {
1278 // out of memory
1279 status = U_MEMORY_ALLOCATION_ERROR;
1280 return;
1281 }
1282
1283 if (prevElem!=NULL) {
1284 prevElem->next=curElem;
1285 }
1286 curElem->next=NULL;
1287 prevElem = curElem;
1288 otherElem = otherElem->next;
1289 }
1290
1291 }
1292 }
1293
1294 PtnElem*
getHeader(UChar baseChar)1295 PatternMap::getHeader(UChar baseChar) {
1296 PtnElem* curElem;
1297
1298 if ( (baseChar >= CAP_A) && (baseChar <= CAP_Z) ) {
1299 curElem = boot[baseChar-CAP_A];
1300 }
1301 else {
1302 if ( (baseChar >=LOW_A) && (baseChar <= LOW_Z) ) {
1303 curElem = boot[26+baseChar-LOW_A];
1304 }
1305 else {
1306 return NULL;
1307 }
1308 }
1309 return curElem;
1310 }
1311
~PatternMap()1312 PatternMap::~PatternMap() {
1313 for (int32_t i=0; i < MAX_PATTERN_ENTRIES; ++i ) {
1314 if (boot[i]!=NULL ) {
1315 delete boot[i];
1316 boot[i]=NULL;
1317 }
1318 }
1319 } // PatternMap destructor
1320
1321 void
add(const UnicodeString & basePattern,const PtnSkeleton & skeleton,const UnicodeString & value,UBool skeletonWasSpecified,UErrorCode & status)1322 PatternMap::add(const UnicodeString& basePattern,
1323 const PtnSkeleton& skeleton,
1324 const UnicodeString& value,// mapped pattern value
1325 UBool skeletonWasSpecified,
1326 UErrorCode &status) {
1327 UChar baseChar = basePattern.charAt(0);
1328 PtnElem *curElem, *baseElem;
1329 status = U_ZERO_ERROR;
1330
1331 // the baseChar must be A-Z or a-z
1332 if ((baseChar >= CAP_A) && (baseChar <= CAP_Z)) {
1333 baseElem = boot[baseChar-CAP_A];
1334 }
1335 else {
1336 if ((baseChar >=LOW_A) && (baseChar <= LOW_Z)) {
1337 baseElem = boot[26+baseChar-LOW_A];
1338 }
1339 else {
1340 status = U_ILLEGAL_CHARACTER;
1341 return;
1342 }
1343 }
1344
1345 if (baseElem == NULL) {
1346 if ((curElem = new PtnElem(basePattern, value)) == NULL ) {
1347 // out of memory
1348 status = U_MEMORY_ALLOCATION_ERROR;
1349 return;
1350 }
1351 if (baseChar >= LOW_A) {
1352 boot[26 + (baseChar-LOW_A)] = curElem;
1353 }
1354 else {
1355 boot[baseChar-CAP_A] = curElem;
1356 }
1357 curElem->skeleton = new PtnSkeleton(skeleton);
1358 curElem->skeletonWasSpecified = skeletonWasSpecified;
1359 }
1360 if ( baseElem != NULL ) {
1361 curElem = getDuplicateElem(basePattern, skeleton, baseElem);
1362
1363 if (curElem == NULL) {
1364 // add new element to the list.
1365 curElem = baseElem;
1366 while( curElem -> next != NULL )
1367 {
1368 curElem = curElem->next;
1369 }
1370 if ((curElem->next = new PtnElem(basePattern, value)) == NULL ) {
1371 // out of memory
1372 status = U_MEMORY_ALLOCATION_ERROR;
1373 return;
1374 }
1375 curElem=curElem->next;
1376 curElem->skeleton = new PtnSkeleton(skeleton);
1377 curElem->skeletonWasSpecified = skeletonWasSpecified;
1378 }
1379 else {
1380 // Pattern exists in the list already.
1381 if ( !isDupAllowed ) {
1382 return;
1383 }
1384 // Overwrite the value.
1385 curElem->pattern = value;
1386 }
1387 }
1388 } // PatternMap::add
1389
1390 // Find the pattern from the given basePattern string.
1391 const UnicodeString *
getPatternFromBasePattern(UnicodeString & basePattern,UBool & skeletonWasSpecified)1392 PatternMap::getPatternFromBasePattern(UnicodeString& basePattern, UBool& skeletonWasSpecified) { // key to search for
1393 PtnElem *curElem;
1394
1395 if ((curElem=getHeader(basePattern.charAt(0)))==NULL) {
1396 return NULL; // no match
1397 }
1398
1399 do {
1400 if ( basePattern.compare(curElem->basePattern)==0 ) {
1401 skeletonWasSpecified = curElem->skeletonWasSpecified;
1402 return &(curElem->pattern);
1403 }
1404 curElem=curElem->next;
1405 }while (curElem != NULL);
1406
1407 return NULL;
1408 } // PatternMap::getFromBasePattern
1409
1410
1411 // Find the pattern from the given skeleton.
1412 // At least when this is called from getBestRaw & addPattern (in which case specifiedSkeletonPtr is non-NULL),
1413 // the comparison should be based on skeleton.original (which is unique and tied to the distance measurement in bestRaw)
1414 // and not skeleton.baseOriginal (which is not unique); otherwise we may pick a different skeleton than the one with the
1415 // optimum distance value in getBestRaw. When this is called from public getRedundants (specifiedSkeletonPtr is NULL),
1416 // for now it will continue to compare based on baseOriginal so as not to change the behavior unnecessarily.
1417 const UnicodeString *
getPatternFromSkeleton(PtnSkeleton & skeleton,const PtnSkeleton ** specifiedSkeletonPtr)1418 PatternMap::getPatternFromSkeleton(PtnSkeleton& skeleton, const PtnSkeleton** specifiedSkeletonPtr) { // key to search for
1419 PtnElem *curElem;
1420
1421 if (specifiedSkeletonPtr) {
1422 *specifiedSkeletonPtr = NULL;
1423 }
1424
1425 // find boot entry
1426 UChar baseChar='\0';
1427 for (int32_t i=0; i<UDATPG_FIELD_COUNT; ++i) {
1428 if (skeleton.baseOriginal[i].length() !=0 ) {
1429 baseChar = skeleton.baseOriginal[i].charAt(0);
1430 break;
1431 }
1432 }
1433
1434 if ((curElem=getHeader(baseChar))==NULL) {
1435 return NULL; // no match
1436 }
1437
1438 do {
1439 int32_t i=0;
1440 if (specifiedSkeletonPtr != NULL) { // called from DateTimePatternGenerator::getBestRaw or addPattern, use original
1441 for (i=0; i<UDATPG_FIELD_COUNT; ++i) {
1442 if (curElem->skeleton->original[i].compare(skeleton.original[i]) != 0 )
1443 {
1444 break;
1445 }
1446 }
1447 } else { // called from DateTimePatternGenerator::getRedundants, use baseOriginal
1448 for (i=0; i<UDATPG_FIELD_COUNT; ++i) {
1449 if (curElem->skeleton->baseOriginal[i].compare(skeleton.baseOriginal[i]) != 0 )
1450 {
1451 break;
1452 }
1453 }
1454 }
1455 if (i == UDATPG_FIELD_COUNT) {
1456 if (specifiedSkeletonPtr && curElem->skeletonWasSpecified) {
1457 *specifiedSkeletonPtr = curElem->skeleton;
1458 }
1459 return &(curElem->pattern);
1460 }
1461 curElem=curElem->next;
1462 }while (curElem != NULL);
1463
1464 return NULL;
1465 }
1466
1467 UBool
equals(const PatternMap & other)1468 PatternMap::equals(const PatternMap& other) {
1469 if ( this==&other ) {
1470 return TRUE;
1471 }
1472 for (int32_t bootIndex=0; bootIndex<MAX_PATTERN_ENTRIES; ++bootIndex ) {
1473 if ( boot[bootIndex]==other.boot[bootIndex] ) {
1474 continue;
1475 }
1476 if ( (boot[bootIndex]==NULL)||(other.boot[bootIndex]==NULL) ) {
1477 return FALSE;
1478 }
1479 PtnElem *otherElem = other.boot[bootIndex];
1480 PtnElem *myElem = boot[bootIndex];
1481 while ((otherElem!=NULL) || (myElem!=NULL)) {
1482 if ( myElem == otherElem ) {
1483 break;
1484 }
1485 if ((otherElem==NULL) || (myElem==NULL)) {
1486 return FALSE;
1487 }
1488 if ( (myElem->basePattern != otherElem->basePattern) ||
1489 (myElem->pattern != otherElem->pattern) ) {
1490 return FALSE;
1491 }
1492 if ((myElem->skeleton!=otherElem->skeleton)&&
1493 !myElem->skeleton->equals(*(otherElem->skeleton))) {
1494 return FALSE;
1495 }
1496 myElem = myElem->next;
1497 otherElem=otherElem->next;
1498 }
1499 }
1500 return TRUE;
1501 }
1502
1503 // find any key existing in the mapping table already.
1504 // return TRUE if there is an existing key, otherwise return FALSE.
1505 PtnElem*
getDuplicateElem(const UnicodeString & basePattern,const PtnSkeleton & skeleton,PtnElem * baseElem)1506 PatternMap::getDuplicateElem(
1507 const UnicodeString &basePattern,
1508 const PtnSkeleton &skeleton,
1509 PtnElem *baseElem) {
1510 PtnElem *curElem;
1511
1512 if ( baseElem == (PtnElem *)NULL ) {
1513 return (PtnElem*)NULL;
1514 }
1515 else {
1516 curElem = baseElem;
1517 }
1518 do {
1519 if ( basePattern.compare(curElem->basePattern)==0 ) {
1520 UBool isEqual=TRUE;
1521 for (int32_t i=0; i<UDATPG_FIELD_COUNT; ++i) {
1522 if (curElem->skeleton->type[i] != skeleton.type[i] ) {
1523 isEqual=FALSE;
1524 break;
1525 }
1526 }
1527 if (isEqual) {
1528 return curElem;
1529 }
1530 }
1531 curElem = curElem->next;
1532 } while( curElem != (PtnElem *)NULL );
1533
1534 // end of the list
1535 return (PtnElem*)NULL;
1536
1537 } // PatternMap::getDuplicateElem
1538
DateTimeMatcher(void)1539 DateTimeMatcher::DateTimeMatcher(void) {
1540 }
1541
DateTimeMatcher(const DateTimeMatcher & other)1542 DateTimeMatcher::DateTimeMatcher(const DateTimeMatcher& other) {
1543 copyFrom(other.skeleton);
1544 }
1545
1546
1547 void
set(const UnicodeString & pattern,FormatParser * fp)1548 DateTimeMatcher::set(const UnicodeString& pattern, FormatParser* fp) {
1549 PtnSkeleton localSkeleton;
1550 return set(pattern, fp, localSkeleton);
1551 }
1552
1553 void
set(const UnicodeString & pattern,FormatParser * fp,PtnSkeleton & skeletonResult)1554 DateTimeMatcher::set(const UnicodeString& pattern, FormatParser* fp, PtnSkeleton& skeletonResult) {
1555 int32_t i;
1556 for (i=0; i<UDATPG_FIELD_COUNT; ++i) {
1557 skeletonResult.type[i]=NONE;
1558 }
1559 fp->set(pattern);
1560 for (i=0; i < fp->itemNumber; i++) {
1561 UnicodeString field = fp->items[i];
1562 if ( field.charAt(0) == LOW_A ) {
1563 continue; // skip 'a'
1564 }
1565
1566 if ( fp->isQuoteLiteral(field) ) {
1567 UnicodeString quoteLiteral;
1568 fp->getQuoteLiteral(quoteLiteral, &i);
1569 continue;
1570 }
1571 int32_t canonicalIndex = fp->getCanonicalIndex(field);
1572 if (canonicalIndex < 0 ) {
1573 continue;
1574 }
1575 const dtTypeElem *row = &dtTypes[canonicalIndex];
1576 int32_t typeValue = row->field;
1577 skeletonResult.original[typeValue]=field;
1578 UChar repeatChar = row->patternChar;
1579 int32_t repeatCount = row->minLen > 3 ? 3: row->minLen;
1580 while (repeatCount-- > 0) {
1581 skeletonResult.baseOriginal[typeValue] += repeatChar;
1582 }
1583 int16_t subTypeValue = row->type;
1584 if ( row->type > 0) {
1585 subTypeValue += field.length();
1586 }
1587 skeletonResult.type[typeValue] = subTypeValue;
1588 }
1589 copyFrom(skeletonResult);
1590 }
1591
1592 void
getBasePattern(UnicodeString & result)1593 DateTimeMatcher::getBasePattern(UnicodeString &result ) {
1594 result.remove(); // Reset the result first.
1595 for (int32_t i=0; i<UDATPG_FIELD_COUNT; ++i ) {
1596 if (skeleton.baseOriginal[i].length()!=0) {
1597 result += skeleton.baseOriginal[i];
1598 }
1599 }
1600 }
1601
1602 UnicodeString
getPattern()1603 DateTimeMatcher::getPattern() {
1604 UnicodeString result;
1605
1606 for (int32_t i=0; i<UDATPG_FIELD_COUNT; ++i ) {
1607 if (skeleton.original[i].length()!=0) {
1608 result += skeleton.original[i];
1609 }
1610 }
1611 return result;
1612 }
1613
1614 int32_t
getDistance(const DateTimeMatcher & other,int32_t includeMask,DistanceInfo & distanceInfo)1615 DateTimeMatcher::getDistance(const DateTimeMatcher& other, int32_t includeMask, DistanceInfo& distanceInfo) {
1616 int32_t result=0;
1617 distanceInfo.clear();
1618 for (int32_t i=0; i<UDATPG_FIELD_COUNT; ++i ) {
1619 int32_t myType = (includeMask&(1<<i))==0 ? 0 : skeleton.type[i];
1620 int32_t otherType = other.skeleton.type[i];
1621 if (myType==otherType) {
1622 continue;
1623 }
1624 if (myType==0) {// and other is not
1625 result += EXTRA_FIELD;
1626 distanceInfo.addExtra(i);
1627 }
1628 else {
1629 if (otherType==0) {
1630 result += MISSING_FIELD;
1631 distanceInfo.addMissing(i);
1632 }
1633 else {
1634 result += abs(myType - otherType);
1635 }
1636 }
1637
1638 }
1639 return result;
1640 }
1641
1642 void
copyFrom(const PtnSkeleton & newSkeleton)1643 DateTimeMatcher::copyFrom(const PtnSkeleton& newSkeleton) {
1644 for (int32_t i=0; i<UDATPG_FIELD_COUNT; ++i) {
1645 this->skeleton.type[i]=newSkeleton.type[i];
1646 this->skeleton.original[i]=newSkeleton.original[i];
1647 this->skeleton.baseOriginal[i]=newSkeleton.baseOriginal[i];
1648 }
1649 }
1650
1651 void
copyFrom()1652 DateTimeMatcher::copyFrom() {
1653 // same as clear
1654 for (int32_t i=0; i<UDATPG_FIELD_COUNT; ++i) {
1655 this->skeleton.type[i]=0;
1656 this->skeleton.original[i].remove();
1657 this->skeleton.baseOriginal[i].remove();
1658 }
1659 }
1660
1661 UBool
equals(const DateTimeMatcher * other) const1662 DateTimeMatcher::equals(const DateTimeMatcher* other) const {
1663 if (other==NULL) {
1664 return FALSE;
1665 }
1666 for (int32_t i=0; i<UDATPG_FIELD_COUNT; ++i) {
1667 if (this->skeleton.original[i]!=other->skeleton.original[i] ) {
1668 return FALSE;
1669 }
1670 }
1671 return TRUE;
1672 }
1673
1674 int32_t
getFieldMask()1675 DateTimeMatcher::getFieldMask() {
1676 int32_t result=0;
1677
1678 for (int32_t i=0; i<UDATPG_FIELD_COUNT; ++i) {
1679 if (skeleton.type[i]!=0) {
1680 result |= (1<<i);
1681 }
1682 }
1683 return result;
1684 }
1685
1686 PtnSkeleton*
getSkeletonPtr()1687 DateTimeMatcher::getSkeletonPtr() {
1688 return &skeleton;
1689 }
1690
FormatParser()1691 FormatParser::FormatParser () {
1692 status = START;
1693 itemNumber=0;
1694 }
1695
1696
~FormatParser()1697 FormatParser::~FormatParser () {
1698 }
1699
1700
1701 // Find the next token with the starting position and length
1702 // Note: the startPos may
1703 FormatParser::TokenStatus
setTokens(const UnicodeString & pattern,int32_t startPos,int32_t * len)1704 FormatParser::setTokens(const UnicodeString& pattern, int32_t startPos, int32_t *len) {
1705 int32_t curLoc = startPos;
1706 if ( curLoc >= pattern.length()) {
1707 return DONE;
1708 }
1709 // check the current char is between A-Z or a-z
1710 do {
1711 UChar c=pattern.charAt(curLoc);
1712 if ( (c>=CAP_A && c<=CAP_Z) || (c>=LOW_A && c<=LOW_Z) ) {
1713 curLoc++;
1714 }
1715 else {
1716 startPos = curLoc;
1717 *len=1;
1718 return ADD_TOKEN;
1719 }
1720
1721 if ( pattern.charAt(curLoc)!= pattern.charAt(startPos) ) {
1722 break; // not the same token
1723 }
1724 } while(curLoc <= pattern.length());
1725 *len = curLoc-startPos;
1726 return ADD_TOKEN;
1727 }
1728
1729 void
set(const UnicodeString & pattern)1730 FormatParser::set(const UnicodeString& pattern) {
1731 int32_t startPos=0;
1732 TokenStatus result=START;
1733 int32_t len=0;
1734 itemNumber =0;
1735
1736 do {
1737 result = setTokens( pattern, startPos, &len );
1738 if ( result == ADD_TOKEN )
1739 {
1740 items[itemNumber++] = UnicodeString(pattern, startPos, len );
1741 startPos += len;
1742 }
1743 else {
1744 break;
1745 }
1746 } while (result==ADD_TOKEN && itemNumber < MAX_DT_TOKEN);
1747 }
1748
1749 int32_t
getCanonicalIndex(const UnicodeString & s,UBool strict)1750 FormatParser::getCanonicalIndex(const UnicodeString& s, UBool strict) {
1751 int32_t len = s.length();
1752 if (len == 0) {
1753 return -1;
1754 }
1755 UChar ch = s.charAt(0);
1756
1757 // Verify that all are the same character.
1758 for (int32_t l = 1; l < len; l++) {
1759 if (ch != s.charAt(l)) {
1760 return -1;
1761 }
1762 }
1763 int32_t i = 0;
1764 int32_t bestRow = -1;
1765 while (dtTypes[i].patternChar != '\0') {
1766 if ( dtTypes[i].patternChar != ch ) {
1767 ++i;
1768 continue;
1769 }
1770 bestRow = i;
1771 if (dtTypes[i].patternChar != dtTypes[i+1].patternChar) {
1772 return i;
1773 }
1774 if (dtTypes[i+1].minLen <= len) {
1775 ++i;
1776 continue;
1777 }
1778 return i;
1779 }
1780 return strict ? -1 : bestRow;
1781 }
1782
1783 UBool
isQuoteLiteral(const UnicodeString & s) const1784 FormatParser::isQuoteLiteral(const UnicodeString& s) const {
1785 return (UBool)(s.charAt(0)==SINGLE_QUOTE);
1786 }
1787
1788 // This function aussumes the current itemIndex points to the quote literal.
1789 // Please call isQuoteLiteral prior to this function.
1790 void
getQuoteLiteral(UnicodeString & quote,int32_t * itemIndex)1791 FormatParser::getQuoteLiteral(UnicodeString& quote, int32_t *itemIndex) {
1792 int32_t i=*itemIndex;
1793
1794 quote.remove();
1795 if (items[i].charAt(0)==SINGLE_QUOTE) {
1796 quote += items[i];
1797 ++i;
1798 }
1799 while ( i < itemNumber ) {
1800 if ( items[i].charAt(0)==SINGLE_QUOTE ) {
1801 if ( (i+1<itemNumber) && (items[i+1].charAt(0)==SINGLE_QUOTE)) {
1802 // two single quotes e.g. 'o''clock'
1803 quote += items[i++];
1804 quote += items[i++];
1805 continue;
1806 }
1807 else {
1808 quote += items[i];
1809 break;
1810 }
1811 }
1812 else {
1813 quote += items[i];
1814 }
1815 ++i;
1816 }
1817 *itemIndex=i;
1818 }
1819
1820 UBool
isPatternSeparator(UnicodeString & field)1821 FormatParser::isPatternSeparator(UnicodeString& field) {
1822 for (int32_t i=0; i<field.length(); ++i ) {
1823 UChar c= field.charAt(i);
1824 if ( (c==SINGLE_QUOTE) || (c==BACKSLASH) || (c==SPACE) || (c==COLON) ||
1825 (c==QUOTATION_MARK) || (c==COMMA) || (c==HYPHEN) ||(items[i].charAt(0)==DOT) ) {
1826 continue;
1827 }
1828 else {
1829 return FALSE;
1830 }
1831 }
1832 return TRUE;
1833 }
1834
1835 void
setTo(DistanceInfo & other)1836 DistanceInfo::setTo(DistanceInfo &other) {
1837 missingFieldMask = other.missingFieldMask;
1838 extraFieldMask= other.extraFieldMask;
1839 }
1840
PatternMapIterator()1841 PatternMapIterator::PatternMapIterator() {
1842 bootIndex = 0;
1843 nodePtr = NULL;
1844 patternMap=NULL;
1845 matcher= new DateTimeMatcher();
1846 }
1847
1848
~PatternMapIterator()1849 PatternMapIterator::~PatternMapIterator() {
1850 delete matcher;
1851 }
1852
1853 void
set(PatternMap & newPatternMap)1854 PatternMapIterator::set(PatternMap& newPatternMap) {
1855 this->patternMap=&newPatternMap;
1856 }
1857
1858 PtnSkeleton*
getSkeleton()1859 PatternMapIterator::getSkeleton() {
1860 if ( nodePtr == NULL ) {
1861 return NULL;
1862 }
1863 else {
1864 return nodePtr->skeleton;
1865 }
1866 }
1867
1868 UBool
hasNext()1869 PatternMapIterator::hasNext() {
1870 int32_t headIndex=bootIndex;
1871 PtnElem *curPtr=nodePtr;
1872
1873 if (patternMap==NULL) {
1874 return FALSE;
1875 }
1876 while ( headIndex < MAX_PATTERN_ENTRIES ) {
1877 if ( curPtr != NULL ) {
1878 if ( curPtr->next != NULL ) {
1879 return TRUE;
1880 }
1881 else {
1882 headIndex++;
1883 curPtr=NULL;
1884 continue;
1885 }
1886 }
1887 else {
1888 if ( patternMap->boot[headIndex] != NULL ) {
1889 return TRUE;
1890 }
1891 else {
1892 headIndex++;
1893 continue;
1894 }
1895 }
1896
1897 }
1898 return FALSE;
1899 }
1900
1901 DateTimeMatcher&
next()1902 PatternMapIterator::next() {
1903 while ( bootIndex < MAX_PATTERN_ENTRIES ) {
1904 if ( nodePtr != NULL ) {
1905 if ( nodePtr->next != NULL ) {
1906 nodePtr = nodePtr->next;
1907 break;
1908 }
1909 else {
1910 bootIndex++;
1911 nodePtr=NULL;
1912 continue;
1913 }
1914 }
1915 else {
1916 if ( patternMap->boot[bootIndex] != NULL ) {
1917 nodePtr = patternMap->boot[bootIndex];
1918 break;
1919 }
1920 else {
1921 bootIndex++;
1922 continue;
1923 }
1924 }
1925 }
1926 if (nodePtr!=NULL) {
1927 matcher->copyFrom(*nodePtr->skeleton);
1928 }
1929 else {
1930 matcher->copyFrom();
1931 }
1932 return *matcher;
1933 }
1934
PtnSkeleton()1935 PtnSkeleton::PtnSkeleton() {
1936 }
1937
1938
PtnSkeleton(const PtnSkeleton & other)1939 PtnSkeleton::PtnSkeleton(const PtnSkeleton& other) {
1940 for (int32_t i=0; i<UDATPG_FIELD_COUNT; ++i) {
1941 this->type[i]=other.type[i];
1942 this->original[i]=other.original[i];
1943 this->baseOriginal[i]=other.baseOriginal[i];
1944 }
1945 }
1946
1947 UBool
equals(const PtnSkeleton & other)1948 PtnSkeleton::equals(const PtnSkeleton& other) {
1949 for (int32_t i=0; i<UDATPG_FIELD_COUNT; ++i) {
1950 if ( (type[i]!= other.type[i]) ||
1951 (original[i]!=other.original[i]) ||
1952 (baseOriginal[i]!=other.baseOriginal[i]) ) {
1953 return FALSE;
1954 }
1955 }
1956 return TRUE;
1957 }
1958
1959 UnicodeString
getSkeleton()1960 PtnSkeleton::getSkeleton() {
1961 UnicodeString result;
1962
1963 for(int32_t i=0; i< UDATPG_FIELD_COUNT; ++i) {
1964 if (original[i].length()!=0) {
1965 result += original[i];
1966 }
1967 }
1968 return result;
1969 }
1970
1971 UnicodeString
getBaseSkeleton()1972 PtnSkeleton::getBaseSkeleton() {
1973 UnicodeString result;
1974
1975 for(int32_t i=0; i< UDATPG_FIELD_COUNT; ++i) {
1976 if (baseOriginal[i].length()!=0) {
1977 result += baseOriginal[i];
1978 }
1979 }
1980 return result;
1981 }
1982
~PtnSkeleton()1983 PtnSkeleton::~PtnSkeleton() {
1984 }
1985
PtnElem(const UnicodeString & basePat,const UnicodeString & pat)1986 PtnElem::PtnElem(const UnicodeString &basePat, const UnicodeString &pat) :
1987 basePattern(basePat),
1988 skeleton(NULL),
1989 pattern(pat),
1990 next(NULL)
1991 {
1992 }
1993
~PtnElem()1994 PtnElem::~PtnElem() {
1995
1996 if (next!=NULL) {
1997 delete next;
1998 }
1999 delete skeleton;
2000 }
2001
DTSkeletonEnumeration(PatternMap & patternMap,dtStrEnum type,UErrorCode & status)2002 DTSkeletonEnumeration::DTSkeletonEnumeration(PatternMap &patternMap, dtStrEnum type, UErrorCode& status) {
2003 PtnElem *curElem;
2004 PtnSkeleton *curSkeleton;
2005 UnicodeString s;
2006 int32_t bootIndex;
2007
2008 pos=0;
2009 fSkeletons = new UVector(status);
2010 if (U_FAILURE(status)) {
2011 delete fSkeletons;
2012 return;
2013 }
2014 for (bootIndex=0; bootIndex<MAX_PATTERN_ENTRIES; ++bootIndex ) {
2015 curElem = patternMap.boot[bootIndex];
2016 while (curElem!=NULL) {
2017 switch(type) {
2018 case DT_BASESKELETON:
2019 s=curElem->basePattern;
2020 break;
2021 case DT_PATTERN:
2022 s=curElem->pattern;
2023 break;
2024 case DT_SKELETON:
2025 curSkeleton=curElem->skeleton;
2026 s=curSkeleton->getSkeleton();
2027 break;
2028 }
2029 if ( !isCanonicalItem(s) ) {
2030 fSkeletons->addElement(new UnicodeString(s), status);
2031 if (U_FAILURE(status)) {
2032 delete fSkeletons;
2033 fSkeletons = NULL;
2034 return;
2035 }
2036 }
2037 curElem = curElem->next;
2038 }
2039 }
2040 if ((bootIndex==MAX_PATTERN_ENTRIES) && (curElem!=NULL) ) {
2041 status = U_BUFFER_OVERFLOW_ERROR;
2042 }
2043 }
2044
2045 const UnicodeString*
snext(UErrorCode & status)2046 DTSkeletonEnumeration::snext(UErrorCode& status) {
2047 if (U_SUCCESS(status) && pos < fSkeletons->size()) {
2048 return (const UnicodeString*)fSkeletons->elementAt(pos++);
2049 }
2050 return NULL;
2051 }
2052
2053 void
reset(UErrorCode &)2054 DTSkeletonEnumeration::reset(UErrorCode& /*status*/) {
2055 pos=0;
2056 }
2057
2058 int32_t
count(UErrorCode &) const2059 DTSkeletonEnumeration::count(UErrorCode& /*status*/) const {
2060 return (fSkeletons==NULL) ? 0 : fSkeletons->size();
2061 }
2062
2063 UBool
isCanonicalItem(const UnicodeString & item)2064 DTSkeletonEnumeration::isCanonicalItem(const UnicodeString& item) {
2065 if ( item.length() != 1 ) {
2066 return FALSE;
2067 }
2068 for (int32_t i=0; i<UDATPG_FIELD_COUNT; ++i) {
2069 if (item.charAt(0)==Canonical_Items[i]) {
2070 return TRUE;
2071 }
2072 }
2073 return FALSE;
2074 }
2075
~DTSkeletonEnumeration()2076 DTSkeletonEnumeration::~DTSkeletonEnumeration() {
2077 UnicodeString *s;
2078 for (int32_t i=0; i<fSkeletons->size(); ++i) {
2079 if ((s=(UnicodeString *)fSkeletons->elementAt(i))!=NULL) {
2080 delete s;
2081 }
2082 }
2083 delete fSkeletons;
2084 }
2085
DTRedundantEnumeration()2086 DTRedundantEnumeration::DTRedundantEnumeration() {
2087 pos=0;
2088 fPatterns = NULL;
2089 }
2090
2091 void
add(const UnicodeString & pattern,UErrorCode & status)2092 DTRedundantEnumeration::add(const UnicodeString& pattern, UErrorCode& status) {
2093 if (U_FAILURE(status)) return;
2094 if (fPatterns == NULL) {
2095 fPatterns = new UVector(status);
2096 if (U_FAILURE(status)) {
2097 delete fPatterns;
2098 fPatterns = NULL;
2099 return;
2100 }
2101 }
2102 fPatterns->addElement(new UnicodeString(pattern), status);
2103 if (U_FAILURE(status)) {
2104 delete fPatterns;
2105 fPatterns = NULL;
2106 return;
2107 }
2108 }
2109
2110 const UnicodeString*
snext(UErrorCode & status)2111 DTRedundantEnumeration::snext(UErrorCode& status) {
2112 if (U_SUCCESS(status) && pos < fPatterns->size()) {
2113 return (const UnicodeString*)fPatterns->elementAt(pos++);
2114 }
2115 return NULL;
2116 }
2117
2118 void
reset(UErrorCode &)2119 DTRedundantEnumeration::reset(UErrorCode& /*status*/) {
2120 pos=0;
2121 }
2122
2123 int32_t
count(UErrorCode &) const2124 DTRedundantEnumeration::count(UErrorCode& /*status*/) const {
2125 return (fPatterns==NULL) ? 0 : fPatterns->size();
2126 }
2127
2128 UBool
isCanonicalItem(const UnicodeString & item)2129 DTRedundantEnumeration::isCanonicalItem(const UnicodeString& item) {
2130 if ( item.length() != 1 ) {
2131 return FALSE;
2132 }
2133 for (int32_t i=0; i<UDATPG_FIELD_COUNT; ++i) {
2134 if (item.charAt(0)==Canonical_Items[i]) {
2135 return TRUE;
2136 }
2137 }
2138 return FALSE;
2139 }
2140
~DTRedundantEnumeration()2141 DTRedundantEnumeration::~DTRedundantEnumeration() {
2142 UnicodeString *s;
2143 for (int32_t i=0; i<fPatterns->size(); ++i) {
2144 if ((s=(UnicodeString *)fPatterns->elementAt(i))!=NULL) {
2145 delete s;
2146 }
2147 }
2148 delete fPatterns;
2149 }
2150
2151 U_NAMESPACE_END
2152
2153
2154 #endif /* #if !UCONFIG_NO_FORMATTING */
2155
2156 //eof
2157