1 // © 2016 and later: Unicode, Inc. and others.
2 // License & terms of use: http://www.unicode.org/copyright.html
3 /*
4 *******************************************************************************
5 * Copyright (C) 1997-2016, International Business Machines Corporation and *
6 * others. All Rights Reserved. *
7 *******************************************************************************
8 *
9 * File SMPDTFMT.CPP
10 *
11 * Modification History:
12 *
13 * Date Name Description
14 * 02/19/97 aliu Converted from java.
15 * 03/31/97 aliu Modified extensively to work with 50 locales.
16 * 04/01/97 aliu Added support for centuries.
17 * 07/09/97 helena Made ParsePosition into a class.
18 * 07/21/98 stephen Added initializeDefaultCentury.
19 * Removed getZoneIndex (added in DateFormatSymbols)
20 * Removed subParseLong
21 * Removed chk
22 * 02/22/99 stephen Removed character literals for EBCDIC safety
23 * 10/14/99 aliu Updated 2-digit year parsing so that only "00" thru
24 * "99" are recognized. {j28 4182066}
25 * 11/15/99 weiv Added support for week of year/day of week format
26 ********************************************************************************
27 */
28
29 #define ZID_KEY_MAX 128
30
31 #include "unicode/utypes.h"
32
33 #if !UCONFIG_NO_FORMATTING
34 #include "unicode/smpdtfmt.h"
35 #include "unicode/dtfmtsym.h"
36 #include "unicode/ures.h"
37 #include "unicode/msgfmt.h"
38 #include "unicode/calendar.h"
39 #include "unicode/gregocal.h"
40 #include "unicode/timezone.h"
41 #include "unicode/decimfmt.h"
42 #include "unicode/dcfmtsym.h"
43 #include "unicode/uchar.h"
44 #include "unicode/uniset.h"
45 #include "unicode/ustring.h"
46 #include "unicode/basictz.h"
47 #include "unicode/simpleformatter.h"
48 #include "unicode/simpletz.h"
49 #include "unicode/rbtz.h"
50 #include "unicode/tzfmt.h"
51 #include "unicode/ucasemap.h"
52 #include "unicode/utf16.h"
53 #include "unicode/vtzone.h"
54 #include "unicode/udisplaycontext.h"
55 #include "unicode/brkiter.h"
56 #include "unicode/rbnf.h"
57 #include "uresimp.h"
58 #include "olsontz.h"
59 #include "patternprops.h"
60 #include "fphdlimp.h"
61 #include "hebrwcal.h"
62 #include "cstring.h"
63 #include "uassert.h"
64 #include "cmemory.h"
65 #include "umutex.h"
66 #include <float.h>
67 #include "smpdtfst.h"
68 #include "sharednumberformat.h"
69 #include "ucasemap_imp.h"
70 #include "ustr_imp.h"
71 #include "charstr.h"
72 #include "uvector.h"
73 #include "cstr.h"
74 #include "dayperiodrules.h"
75 #include "tznames_impl.h" // ZONE_NAME_U16_MAX
76 #include "number_utypes.h"
77
78 #if defined( U_DEBUG_CALSVC ) || defined (U_DEBUG_CAL)
79 #include <stdio.h>
80 #endif
81
82 // *****************************************************************************
83 // class SimpleDateFormat
84 // *****************************************************************************
85
86 U_NAMESPACE_BEGIN
87
88 /**
89 * Last-resort string to use for "GMT" when constructing time zone strings.
90 */
91 // For time zones that have no names, use strings GMT+minutes and
92 // GMT-minutes. For instance, in France the time zone is GMT+60.
93 // Also accepted are GMT+H:MM or GMT-H:MM.
94 // Currently not being used
95 //static const UChar gGmt[] = {0x0047, 0x004D, 0x0054, 0x0000}; // "GMT"
96 //static const UChar gGmtPlus[] = {0x0047, 0x004D, 0x0054, 0x002B, 0x0000}; // "GMT+"
97 //static const UChar gGmtMinus[] = {0x0047, 0x004D, 0x0054, 0x002D, 0x0000}; // "GMT-"
98 //static const UChar gDefGmtPat[] = {0x0047, 0x004D, 0x0054, 0x007B, 0x0030, 0x007D, 0x0000}; /* GMT{0} */
99 //static const UChar gDefGmtNegHmsPat[] = {0x002D, 0x0048, 0x0048, 0x003A, 0x006D, 0x006D, 0x003A, 0x0073, 0x0073, 0x0000}; /* -HH:mm:ss */
100 //static const UChar gDefGmtNegHmPat[] = {0x002D, 0x0048, 0x0048, 0x003A, 0x006D, 0x006D, 0x0000}; /* -HH:mm */
101 //static const UChar gDefGmtPosHmsPat[] = {0x002B, 0x0048, 0x0048, 0x003A, 0x006D, 0x006D, 0x003A, 0x0073, 0x0073, 0x0000}; /* +HH:mm:ss */
102 //static const UChar gDefGmtPosHmPat[] = {0x002B, 0x0048, 0x0048, 0x003A, 0x006D, 0x006D, 0x0000}; /* +HH:mm */
103 //static const UChar gUt[] = {0x0055, 0x0054, 0x0000}; // "UT"
104 //static const UChar gUtc[] = {0x0055, 0x0054, 0x0043, 0x0000}; // "UT"
105
106 typedef enum GmtPatSize {
107 kGmtLen = 3,
108 kGmtPatLen = 6,
109 kNegHmsLen = 9,
110 kNegHmLen = 6,
111 kPosHmsLen = 9,
112 kPosHmLen = 6,
113 kUtLen = 2,
114 kUtcLen = 3
115 } GmtPatSize;
116
117 // Stuff needed for numbering system overrides
118
119 typedef enum OvrStrType {
120 kOvrStrDate = 0,
121 kOvrStrTime = 1,
122 kOvrStrBoth = 2
123 } OvrStrType;
124
125 static const UDateFormatField kDateFields[] = {
126 UDAT_YEAR_FIELD,
127 UDAT_MONTH_FIELD,
128 UDAT_DATE_FIELD,
129 UDAT_DAY_OF_YEAR_FIELD,
130 UDAT_DAY_OF_WEEK_IN_MONTH_FIELD,
131 UDAT_WEEK_OF_YEAR_FIELD,
132 UDAT_WEEK_OF_MONTH_FIELD,
133 UDAT_YEAR_WOY_FIELD,
134 UDAT_EXTENDED_YEAR_FIELD,
135 UDAT_JULIAN_DAY_FIELD,
136 UDAT_STANDALONE_DAY_FIELD,
137 UDAT_STANDALONE_MONTH_FIELD,
138 UDAT_QUARTER_FIELD,
139 UDAT_STANDALONE_QUARTER_FIELD,
140 UDAT_YEAR_NAME_FIELD,
141 UDAT_RELATED_YEAR_FIELD };
142 static const int8_t kDateFieldsCount = 16;
143
144 static const UDateFormatField kTimeFields[] = {
145 UDAT_HOUR_OF_DAY1_FIELD,
146 UDAT_HOUR_OF_DAY0_FIELD,
147 UDAT_MINUTE_FIELD,
148 UDAT_SECOND_FIELD,
149 UDAT_FRACTIONAL_SECOND_FIELD,
150 UDAT_HOUR1_FIELD,
151 UDAT_HOUR0_FIELD,
152 UDAT_MILLISECONDS_IN_DAY_FIELD,
153 UDAT_TIMEZONE_RFC_FIELD,
154 UDAT_TIMEZONE_LOCALIZED_GMT_OFFSET_FIELD };
155 static const int8_t kTimeFieldsCount = 10;
156
157
158 // This is a pattern-of-last-resort used when we can't load a usable pattern out
159 // of a resource.
160 static const UChar gDefaultPattern[] =
161 {
162 0x79, 0x79, 0x79, 0x79, 0x4D, 0x4D, 0x64, 0x64, 0x20, 0x68, 0x68, 0x3A, 0x6D, 0x6D, 0x20, 0x61, 0
163 }; /* "yyyyMMdd hh:mm a" */
164
165 // This prefix is designed to NEVER MATCH real text, in order to
166 // suppress the parsing of negative numbers. Adjust as needed (if
167 // this becomes valid Unicode).
168 static const UChar SUPPRESS_NEGATIVE_PREFIX[] = {0xAB00, 0};
169
170 /**
171 * These are the tags we expect to see in normal resource bundle files associated
172 * with a locale.
173 */
174 static const UChar QUOTE = 0x27; // Single quote
175
176 /*
177 * The field range check bias for each UDateFormatField.
178 * The bias is added to the minimum and maximum values
179 * before they are compared to the parsed number.
180 * For example, the calendar stores zero-based month numbers
181 * but the parsed month numbers start at 1, so the bias is 1.
182 *
183 * A value of -1 means that the value is not checked.
184 */
185 static const int32_t gFieldRangeBias[] = {
186 -1, // 'G' - UDAT_ERA_FIELD
187 -1, // 'y' - UDAT_YEAR_FIELD
188 1, // 'M' - UDAT_MONTH_FIELD
189 0, // 'd' - UDAT_DATE_FIELD
190 -1, // 'k' - UDAT_HOUR_OF_DAY1_FIELD
191 -1, // 'H' - UDAT_HOUR_OF_DAY0_FIELD
192 0, // 'm' - UDAT_MINUTE_FIELD
193 0, // 's' - UDAT_SECOND_FIELD
194 -1, // 'S' - UDAT_FRACTIONAL_SECOND_FIELD (0-999?)
195 -1, // 'E' - UDAT_DAY_OF_WEEK_FIELD (1-7?)
196 -1, // 'D' - UDAT_DAY_OF_YEAR_FIELD (1 - 366?)
197 -1, // 'F' - UDAT_DAY_OF_WEEK_IN_MONTH_FIELD (1-5?)
198 -1, // 'w' - UDAT_WEEK_OF_YEAR_FIELD (1-52?)
199 -1, // 'W' - UDAT_WEEK_OF_MONTH_FIELD (1-5?)
200 -1, // 'a' - UDAT_AM_PM_FIELD
201 -1, // 'h' - UDAT_HOUR1_FIELD
202 -1, // 'K' - UDAT_HOUR0_FIELD
203 -1, // 'z' - UDAT_TIMEZONE_FIELD
204 -1, // 'Y' - UDAT_YEAR_WOY_FIELD
205 -1, // 'e' - UDAT_DOW_LOCAL_FIELD
206 -1, // 'u' - UDAT_EXTENDED_YEAR_FIELD
207 -1, // 'g' - UDAT_JULIAN_DAY_FIELD
208 -1, // 'A' - UDAT_MILLISECONDS_IN_DAY_FIELD
209 -1, // 'Z' - UDAT_TIMEZONE_RFC_FIELD
210 -1, // 'v' - UDAT_TIMEZONE_GENERIC_FIELD
211 0, // 'c' - UDAT_STANDALONE_DAY_FIELD
212 1, // 'L' - UDAT_STANDALONE_MONTH_FIELD
213 -1, // 'Q' - UDAT_QUARTER_FIELD (1-4?)
214 -1, // 'q' - UDAT_STANDALONE_QUARTER_FIELD
215 -1, // 'V' - UDAT_TIMEZONE_SPECIAL_FIELD
216 -1, // 'U' - UDAT_YEAR_NAME_FIELD
217 -1, // 'O' - UDAT_TIMEZONE_LOCALIZED_GMT_OFFSET_FIELD
218 -1, // 'X' - UDAT_TIMEZONE_ISO_FIELD
219 -1, // 'x' - UDAT_TIMEZONE_ISO_LOCAL_FIELD
220 -1, // 'r' - UDAT_RELATED_YEAR_FIELD
221 #if UDAT_HAS_PATTERN_CHAR_FOR_TIME_SEPARATOR
222 -1, // ':' - UDAT_TIME_SEPARATOR_FIELD
223 #else
224 -1, // (no pattern character currently) - UDAT_TIME_SEPARATOR_FIELD
225 #endif
226 };
227
228 // When calendar uses hebr numbering (i.e. he@calendar=hebrew),
229 // offset the years within the current millenium down to 1-999
230 static const int32_t HEBREW_CAL_CUR_MILLENIUM_START_YEAR = 5000;
231 static const int32_t HEBREW_CAL_CUR_MILLENIUM_END_YEAR = 6000;
232
233 static UMutex LOCK;
234
UOBJECT_DEFINE_RTTI_IMPLEMENTATION(SimpleDateFormat)235 UOBJECT_DEFINE_RTTI_IMPLEMENTATION(SimpleDateFormat)
236
237 SimpleDateFormat::NSOverride::~NSOverride() {
238 if (snf != NULL) {
239 snf->removeRef();
240 }
241 }
242
243
free()244 void SimpleDateFormat::NSOverride::free() {
245 NSOverride *cur = this;
246 while (cur) {
247 NSOverride *next_temp = cur->next;
248 delete cur;
249 cur = next_temp;
250 }
251 }
252
253 // no matter what the locale's default number format looked like, we want
254 // to modify it so that it doesn't use thousands separators, doesn't always
255 // show the decimal point, and recognizes integers only when parsing
fixNumberFormatForDates(NumberFormat & nf)256 static void fixNumberFormatForDates(NumberFormat &nf) {
257 nf.setGroupingUsed(FALSE);
258 DecimalFormat* decfmt = dynamic_cast<DecimalFormat*>(&nf);
259 if (decfmt != NULL) {
260 decfmt->setDecimalSeparatorAlwaysShown(FALSE);
261 }
262 nf.setParseIntegerOnly(TRUE);
263 nf.setMinimumFractionDigits(0); // To prevent "Jan 1.00, 1997.00"
264 }
265
createSharedNumberFormat(NumberFormat * nfToAdopt)266 static const SharedNumberFormat *createSharedNumberFormat(
267 NumberFormat *nfToAdopt) {
268 fixNumberFormatForDates(*nfToAdopt);
269 const SharedNumberFormat *result = new SharedNumberFormat(nfToAdopt);
270 if (result == NULL) {
271 delete nfToAdopt;
272 }
273 return result;
274 }
275
createSharedNumberFormat(const Locale & loc,UErrorCode & status)276 static const SharedNumberFormat *createSharedNumberFormat(
277 const Locale &loc, UErrorCode &status) {
278 NumberFormat *nf = NumberFormat::createInstance(loc, status);
279 if (U_FAILURE(status)) {
280 return NULL;
281 }
282 const SharedNumberFormat *result = createSharedNumberFormat(nf);
283 if (result == NULL) {
284 status = U_MEMORY_ALLOCATION_ERROR;
285 }
286 return result;
287 }
288
allocSharedNumberFormatters()289 static const SharedNumberFormat **allocSharedNumberFormatters() {
290 const SharedNumberFormat **result = (const SharedNumberFormat**)
291 uprv_malloc(UDAT_FIELD_COUNT * sizeof(const SharedNumberFormat*));
292 if (result == NULL) {
293 return NULL;
294 }
295 for (int32_t i = 0; i < UDAT_FIELD_COUNT; ++i) {
296 result[i] = NULL;
297 }
298 return result;
299 }
300
freeSharedNumberFormatters(const SharedNumberFormat ** list)301 static void freeSharedNumberFormatters(const SharedNumberFormat ** list) {
302 for (int32_t i = 0; i < UDAT_FIELD_COUNT; ++i) {
303 SharedObject::clearPtr(list[i]);
304 }
305 uprv_free(list);
306 }
307
getNumberFormatByIndex(UDateFormatField index) const308 const NumberFormat *SimpleDateFormat::getNumberFormatByIndex(
309 UDateFormatField index) const {
310 if (fSharedNumberFormatters == NULL ||
311 fSharedNumberFormatters[index] == NULL) {
312 return fNumberFormat;
313 }
314 return &(**fSharedNumberFormatters[index]);
315 }
316
317 //----------------------------------------------------------------------
318
~SimpleDateFormat()319 SimpleDateFormat::~SimpleDateFormat()
320 {
321 delete fSymbols;
322 if (fSharedNumberFormatters) {
323 freeSharedNumberFormatters(fSharedNumberFormatters);
324 }
325 if (fTimeZoneFormat) {
326 delete fTimeZoneFormat;
327 }
328 freeFastNumberFormatters();
329
330 #if !UCONFIG_NO_BREAK_ITERATION
331 delete fCapitalizationBrkIter;
332 #endif
333 }
334
335 //----------------------------------------------------------------------
336
SimpleDateFormat(UErrorCode & status)337 SimpleDateFormat::SimpleDateFormat(UErrorCode& status)
338 : fLocale(Locale::getDefault()),
339 fSymbols(NULL),
340 fTimeZoneFormat(NULL),
341 fSharedNumberFormatters(NULL),
342 fCapitalizationBrkIter(NULL)
343 {
344 initializeBooleanAttributes();
345 construct(kShort, (EStyle) (kShort + kDateOffset), fLocale, status);
346 initializeDefaultCentury();
347 }
348
349 //----------------------------------------------------------------------
350
SimpleDateFormat(const UnicodeString & pattern,UErrorCode & status)351 SimpleDateFormat::SimpleDateFormat(const UnicodeString& pattern,
352 UErrorCode &status)
353 : fPattern(pattern),
354 fLocale(Locale::getDefault()),
355 fSymbols(NULL),
356 fTimeZoneFormat(NULL),
357 fSharedNumberFormatters(NULL),
358 fCapitalizationBrkIter(NULL)
359 {
360 fDateOverride.setToBogus();
361 fTimeOverride.setToBogus();
362 initializeBooleanAttributes();
363 initializeCalendar(NULL,fLocale,status);
364 fSymbols = DateFormatSymbols::createForLocale(fLocale, status);
365 initialize(fLocale, status);
366 initializeDefaultCentury();
367
368 }
369 //----------------------------------------------------------------------
370
SimpleDateFormat(const UnicodeString & pattern,const UnicodeString & override,UErrorCode & status)371 SimpleDateFormat::SimpleDateFormat(const UnicodeString& pattern,
372 const UnicodeString& override,
373 UErrorCode &status)
374 : fPattern(pattern),
375 fLocale(Locale::getDefault()),
376 fSymbols(NULL),
377 fTimeZoneFormat(NULL),
378 fSharedNumberFormatters(NULL),
379 fCapitalizationBrkIter(NULL)
380 {
381 fDateOverride.setTo(override);
382 fTimeOverride.setToBogus();
383 initializeBooleanAttributes();
384 initializeCalendar(NULL,fLocale,status);
385 fSymbols = DateFormatSymbols::createForLocale(fLocale, status);
386 initialize(fLocale, status);
387 initializeDefaultCentury();
388
389 processOverrideString(fLocale,override,kOvrStrBoth,status);
390
391 }
392
393 //----------------------------------------------------------------------
394
SimpleDateFormat(const UnicodeString & pattern,const Locale & locale,UErrorCode & status)395 SimpleDateFormat::SimpleDateFormat(const UnicodeString& pattern,
396 const Locale& locale,
397 UErrorCode& status)
398 : fPattern(pattern),
399 fLocale(locale),
400 fTimeZoneFormat(NULL),
401 fSharedNumberFormatters(NULL),
402 fCapitalizationBrkIter(NULL)
403 {
404
405 fDateOverride.setToBogus();
406 fTimeOverride.setToBogus();
407 initializeBooleanAttributes();
408
409 initializeCalendar(NULL,fLocale,status);
410 fSymbols = DateFormatSymbols::createForLocale(fLocale, status);
411 initialize(fLocale, status);
412 initializeDefaultCentury();
413 }
414
415 //----------------------------------------------------------------------
416
SimpleDateFormat(const UnicodeString & pattern,const UnicodeString & override,const Locale & locale,UErrorCode & status)417 SimpleDateFormat::SimpleDateFormat(const UnicodeString& pattern,
418 const UnicodeString& override,
419 const Locale& locale,
420 UErrorCode& status)
421 : fPattern(pattern),
422 fLocale(locale),
423 fTimeZoneFormat(NULL),
424 fSharedNumberFormatters(NULL),
425 fCapitalizationBrkIter(NULL)
426 {
427
428 fDateOverride.setTo(override);
429 fTimeOverride.setToBogus();
430 initializeBooleanAttributes();
431
432 initializeCalendar(NULL,fLocale,status);
433 fSymbols = DateFormatSymbols::createForLocale(fLocale, status);
434 initialize(fLocale, status);
435 initializeDefaultCentury();
436
437 processOverrideString(locale,override,kOvrStrBoth,status);
438
439 }
440
441 //----------------------------------------------------------------------
442
SimpleDateFormat(const UnicodeString & pattern,DateFormatSymbols * symbolsToAdopt,UErrorCode & status)443 SimpleDateFormat::SimpleDateFormat(const UnicodeString& pattern,
444 DateFormatSymbols* symbolsToAdopt,
445 UErrorCode& status)
446 : fPattern(pattern),
447 fLocale(Locale::getDefault()),
448 fSymbols(symbolsToAdopt),
449 fTimeZoneFormat(NULL),
450 fSharedNumberFormatters(NULL),
451 fCapitalizationBrkIter(NULL)
452 {
453
454 fDateOverride.setToBogus();
455 fTimeOverride.setToBogus();
456 initializeBooleanAttributes();
457
458 initializeCalendar(NULL,fLocale,status);
459 initialize(fLocale, status);
460 initializeDefaultCentury();
461 }
462
463 //----------------------------------------------------------------------
464
SimpleDateFormat(const UnicodeString & pattern,const DateFormatSymbols & symbols,UErrorCode & status)465 SimpleDateFormat::SimpleDateFormat(const UnicodeString& pattern,
466 const DateFormatSymbols& symbols,
467 UErrorCode& status)
468 : fPattern(pattern),
469 fLocale(Locale::getDefault()),
470 fSymbols(new DateFormatSymbols(symbols)),
471 fTimeZoneFormat(NULL),
472 fSharedNumberFormatters(NULL),
473 fCapitalizationBrkIter(NULL)
474 {
475
476 fDateOverride.setToBogus();
477 fTimeOverride.setToBogus();
478 initializeBooleanAttributes();
479
480 initializeCalendar(NULL, fLocale, status);
481 initialize(fLocale, status);
482 initializeDefaultCentury();
483 }
484
485 //----------------------------------------------------------------------
486
487 // Not for public consumption; used by DateFormat
SimpleDateFormat(EStyle timeStyle,EStyle dateStyle,const Locale & locale,UErrorCode & status)488 SimpleDateFormat::SimpleDateFormat(EStyle timeStyle,
489 EStyle dateStyle,
490 const Locale& locale,
491 UErrorCode& status)
492 : fLocale(locale),
493 fSymbols(NULL),
494 fTimeZoneFormat(NULL),
495 fSharedNumberFormatters(NULL),
496 fCapitalizationBrkIter(NULL)
497 {
498 initializeBooleanAttributes();
499 construct(timeStyle, dateStyle, fLocale, status);
500 if(U_SUCCESS(status)) {
501 initializeDefaultCentury();
502 }
503 }
504
505 //----------------------------------------------------------------------
506
507 /**
508 * Not for public consumption; used by DateFormat. This constructor
509 * never fails. If the resource data is not available, it uses the
510 * the last resort symbols.
511 */
SimpleDateFormat(const Locale & locale,UErrorCode & status)512 SimpleDateFormat::SimpleDateFormat(const Locale& locale,
513 UErrorCode& status)
514 : fPattern(gDefaultPattern),
515 fLocale(locale),
516 fSymbols(NULL),
517 fTimeZoneFormat(NULL),
518 fSharedNumberFormatters(NULL),
519 fCapitalizationBrkIter(NULL)
520 {
521 if (U_FAILURE(status)) return;
522 initializeBooleanAttributes();
523 initializeCalendar(NULL, fLocale, status);
524 fSymbols = DateFormatSymbols::createForLocale(fLocale, status);
525 if (U_FAILURE(status))
526 {
527 status = U_ZERO_ERROR;
528 delete fSymbols;
529 // This constructor doesn't fail; it uses last resort data
530 fSymbols = new DateFormatSymbols(status);
531 /* test for NULL */
532 if (fSymbols == 0) {
533 status = U_MEMORY_ALLOCATION_ERROR;
534 return;
535 }
536 }
537
538 fDateOverride.setToBogus();
539 fTimeOverride.setToBogus();
540
541 initialize(fLocale, status);
542 if(U_SUCCESS(status)) {
543 initializeDefaultCentury();
544 }
545 }
546
547 //----------------------------------------------------------------------
548
SimpleDateFormat(const SimpleDateFormat & other)549 SimpleDateFormat::SimpleDateFormat(const SimpleDateFormat& other)
550 : DateFormat(other),
551 fLocale(other.fLocale),
552 fSymbols(NULL),
553 fTimeZoneFormat(NULL),
554 fSharedNumberFormatters(NULL),
555 fCapitalizationBrkIter(NULL)
556 {
557 initializeBooleanAttributes();
558 *this = other;
559 }
560
561 //----------------------------------------------------------------------
562
operator =(const SimpleDateFormat & other)563 SimpleDateFormat& SimpleDateFormat::operator=(const SimpleDateFormat& other)
564 {
565 if (this == &other) {
566 return *this;
567 }
568 DateFormat::operator=(other);
569 fDateOverride = other.fDateOverride;
570 fTimeOverride = other.fTimeOverride;
571
572 delete fSymbols;
573 fSymbols = NULL;
574
575 if (other.fSymbols)
576 fSymbols = new DateFormatSymbols(*other.fSymbols);
577
578 fDefaultCenturyStart = other.fDefaultCenturyStart;
579 fDefaultCenturyStartYear = other.fDefaultCenturyStartYear;
580 fHaveDefaultCentury = other.fHaveDefaultCentury;
581
582 fPattern = other.fPattern;
583 fHasMinute = other.fHasMinute;
584 fHasSecond = other.fHasSecond;
585
586 // TimeZoneFormat in ICU4C only depends on a locale for now
587 if (fLocale != other.fLocale) {
588 delete fTimeZoneFormat;
589 fTimeZoneFormat = NULL; // forces lazy instantiation with the other locale
590 fLocale = other.fLocale;
591 }
592
593 #if !UCONFIG_NO_BREAK_ITERATION
594 if (other.fCapitalizationBrkIter != NULL) {
595 fCapitalizationBrkIter = (other.fCapitalizationBrkIter)->clone();
596 }
597 #endif
598
599 if (fSharedNumberFormatters != NULL) {
600 freeSharedNumberFormatters(fSharedNumberFormatters);
601 fSharedNumberFormatters = NULL;
602 }
603 if (other.fSharedNumberFormatters != NULL) {
604 fSharedNumberFormatters = allocSharedNumberFormatters();
605 if (fSharedNumberFormatters) {
606 for (int32_t i = 0; i < UDAT_FIELD_COUNT; ++i) {
607 SharedObject::copyPtr(
608 other.fSharedNumberFormatters[i],
609 fSharedNumberFormatters[i]);
610 }
611 }
612 }
613
614 UErrorCode localStatus = U_ZERO_ERROR;
615 freeFastNumberFormatters();
616 initFastNumberFormatters(localStatus);
617
618 return *this;
619 }
620
621 //----------------------------------------------------------------------
622
623 SimpleDateFormat*
clone() const624 SimpleDateFormat::clone() const
625 {
626 return new SimpleDateFormat(*this);
627 }
628
629 //----------------------------------------------------------------------
630
631 UBool
operator ==(const Format & other) const632 SimpleDateFormat::operator==(const Format& other) const
633 {
634 if (DateFormat::operator==(other)) {
635 // The DateFormat::operator== check for fCapitalizationContext equality above
636 // is sufficient to check equality of all derived context-related data.
637 // DateFormat::operator== guarantees following cast is safe
638 SimpleDateFormat* that = (SimpleDateFormat*)&other;
639 return (fPattern == that->fPattern &&
640 fSymbols != NULL && // Check for pathological object
641 that->fSymbols != NULL && // Check for pathological object
642 *fSymbols == *that->fSymbols &&
643 fHaveDefaultCentury == that->fHaveDefaultCentury &&
644 fDefaultCenturyStart == that->fDefaultCenturyStart);
645 }
646 return FALSE;
647 }
648
649 //----------------------------------------------------------------------
650
construct(EStyle timeStyle,EStyle dateStyle,const Locale & locale,UErrorCode & status)651 void SimpleDateFormat::construct(EStyle timeStyle,
652 EStyle dateStyle,
653 const Locale& locale,
654 UErrorCode& status)
655 {
656 // called by several constructors to load pattern data from the resources
657 if (U_FAILURE(status)) return;
658
659 // We will need the calendar to know what type of symbols to load.
660 initializeCalendar(NULL, locale, status);
661 if (U_FAILURE(status)) return;
662
663 // Load date time patterns directly from resources.
664 const char* cType = fCalendar ? fCalendar->getType() : NULL;
665 LocalUResourceBundlePointer bundle(ures_open(NULL, locale.getBaseName(), &status));
666 if (U_FAILURE(status)) return;
667
668 UBool cTypeIsGregorian = TRUE;
669 LocalUResourceBundlePointer dateTimePatterns;
670 if (cType != NULL && uprv_strcmp(cType, "gregorian") != 0) {
671 CharString resourcePath("calendar/", status);
672 resourcePath.append(cType, status).append("/DateTimePatterns", status);
673 dateTimePatterns.adoptInstead(
674 ures_getByKeyWithFallback(bundle.getAlias(), resourcePath.data(),
675 (UResourceBundle*)NULL, &status));
676 cTypeIsGregorian = FALSE;
677 }
678
679 // Check for "gregorian" fallback.
680 if (cTypeIsGregorian || status == U_MISSING_RESOURCE_ERROR) {
681 status = U_ZERO_ERROR;
682 dateTimePatterns.adoptInstead(
683 ures_getByKeyWithFallback(bundle.getAlias(),
684 "calendar/gregorian/DateTimePatterns",
685 (UResourceBundle*)NULL, &status));
686 }
687 if (U_FAILURE(status)) return;
688
689 LocalUResourceBundlePointer currentBundle;
690
691 if (ures_getSize(dateTimePatterns.getAlias()) <= kDateTime)
692 {
693 status = U_INVALID_FORMAT_ERROR;
694 return;
695 }
696
697 setLocaleIDs(ures_getLocaleByType(dateTimePatterns.getAlias(), ULOC_VALID_LOCALE, &status),
698 ures_getLocaleByType(dateTimePatterns.getAlias(), ULOC_ACTUAL_LOCALE, &status));
699
700 // create a symbols object from the locale
701 fSymbols = DateFormatSymbols::createForLocale(locale, status);
702 if (U_FAILURE(status)) return;
703 /* test for NULL */
704 if (fSymbols == 0) {
705 status = U_MEMORY_ALLOCATION_ERROR;
706 return;
707 }
708
709 const UChar *resStr,*ovrStr;
710 int32_t resStrLen,ovrStrLen = 0;
711 fDateOverride.setToBogus();
712 fTimeOverride.setToBogus();
713
714 // if the pattern should include both date and time information, use the date/time
715 // pattern string as a guide to tell use how to glue together the appropriate date
716 // and time pattern strings.
717 if ((timeStyle != kNone) && (dateStyle != kNone))
718 {
719 currentBundle.adoptInstead(
720 ures_getByIndex(dateTimePatterns.getAlias(), (int32_t)timeStyle, NULL, &status));
721 if (U_FAILURE(status)) {
722 status = U_INVALID_FORMAT_ERROR;
723 return;
724 }
725 switch (ures_getType(currentBundle.getAlias())) {
726 case URES_STRING: {
727 resStr = ures_getString(currentBundle.getAlias(), &resStrLen, &status);
728 break;
729 }
730 case URES_ARRAY: {
731 resStr = ures_getStringByIndex(currentBundle.getAlias(), 0, &resStrLen, &status);
732 ovrStr = ures_getStringByIndex(currentBundle.getAlias(), 1, &ovrStrLen, &status);
733 fTimeOverride.setTo(TRUE, ovrStr, ovrStrLen);
734 break;
735 }
736 default: {
737 status = U_INVALID_FORMAT_ERROR;
738 return;
739 }
740 }
741
742 UnicodeString tempus1(TRUE, resStr, resStrLen);
743
744 currentBundle.adoptInstead(
745 ures_getByIndex(dateTimePatterns.getAlias(), (int32_t)dateStyle, NULL, &status));
746 if (U_FAILURE(status)) {
747 status = U_INVALID_FORMAT_ERROR;
748 return;
749 }
750 switch (ures_getType(currentBundle.getAlias())) {
751 case URES_STRING: {
752 resStr = ures_getString(currentBundle.getAlias(), &resStrLen, &status);
753 break;
754 }
755 case URES_ARRAY: {
756 resStr = ures_getStringByIndex(currentBundle.getAlias(), 0, &resStrLen, &status);
757 ovrStr = ures_getStringByIndex(currentBundle.getAlias(), 1, &ovrStrLen, &status);
758 fDateOverride.setTo(TRUE, ovrStr, ovrStrLen);
759 break;
760 }
761 default: {
762 status = U_INVALID_FORMAT_ERROR;
763 return;
764 }
765 }
766
767 UnicodeString tempus2(TRUE, resStr, resStrLen);
768
769 int32_t glueIndex = kDateTime;
770 int32_t patternsSize = ures_getSize(dateTimePatterns.getAlias());
771 if (patternsSize >= (kDateTimeOffset + kShort + 1)) {
772 // Get proper date time format
773 glueIndex = (int32_t)(kDateTimeOffset + (dateStyle - kDateOffset));
774 }
775
776 resStr = ures_getStringByIndex(dateTimePatterns.getAlias(), glueIndex, &resStrLen, &status);
777 SimpleFormatter(UnicodeString(TRUE, resStr, resStrLen), 2, 2, status).
778 format(tempus1, tempus2, fPattern, status);
779 }
780 // if the pattern includes just time data or just date date, load the appropriate
781 // pattern string from the resources
782 // setTo() - see DateFormatSymbols::assignArray comments
783 else if (timeStyle != kNone) {
784 currentBundle.adoptInstead(
785 ures_getByIndex(dateTimePatterns.getAlias(), (int32_t)timeStyle, NULL, &status));
786 if (U_FAILURE(status)) {
787 status = U_INVALID_FORMAT_ERROR;
788 return;
789 }
790 switch (ures_getType(currentBundle.getAlias())) {
791 case URES_STRING: {
792 resStr = ures_getString(currentBundle.getAlias(), &resStrLen, &status);
793 break;
794 }
795 case URES_ARRAY: {
796 resStr = ures_getStringByIndex(currentBundle.getAlias(), 0, &resStrLen, &status);
797 ovrStr = ures_getStringByIndex(currentBundle.getAlias(), 1, &ovrStrLen, &status);
798 fDateOverride.setTo(TRUE, ovrStr, ovrStrLen);
799 break;
800 }
801 default: {
802 status = U_INVALID_FORMAT_ERROR;
803 return;
804 }
805 }
806 fPattern.setTo(TRUE, resStr, resStrLen);
807 }
808 else if (dateStyle != kNone) {
809 currentBundle.adoptInstead(
810 ures_getByIndex(dateTimePatterns.getAlias(), (int32_t)dateStyle, NULL, &status));
811 if (U_FAILURE(status)) {
812 status = U_INVALID_FORMAT_ERROR;
813 return;
814 }
815 switch (ures_getType(currentBundle.getAlias())) {
816 case URES_STRING: {
817 resStr = ures_getString(currentBundle.getAlias(), &resStrLen, &status);
818 break;
819 }
820 case URES_ARRAY: {
821 resStr = ures_getStringByIndex(currentBundle.getAlias(), 0, &resStrLen, &status);
822 ovrStr = ures_getStringByIndex(currentBundle.getAlias(), 1, &ovrStrLen, &status);
823 fDateOverride.setTo(TRUE, ovrStr, ovrStrLen);
824 break;
825 }
826 default: {
827 status = U_INVALID_FORMAT_ERROR;
828 return;
829 }
830 }
831 fPattern.setTo(TRUE, resStr, resStrLen);
832 }
833
834 // and if it includes _neither_, that's an error
835 else
836 status = U_INVALID_FORMAT_ERROR;
837
838 // finally, finish initializing by creating a Calendar and a NumberFormat
839 initialize(locale, status);
840 }
841
842 //----------------------------------------------------------------------
843
844 Calendar*
initializeCalendar(TimeZone * adoptZone,const Locale & locale,UErrorCode & status)845 SimpleDateFormat::initializeCalendar(TimeZone* adoptZone, const Locale& locale, UErrorCode& status)
846 {
847 if(!U_FAILURE(status)) {
848 fCalendar = Calendar::createInstance(adoptZone?adoptZone:TimeZone::createDefault(), locale, status);
849 }
850 return fCalendar;
851 }
852
853 void
initialize(const Locale & locale,UErrorCode & status)854 SimpleDateFormat::initialize(const Locale& locale,
855 UErrorCode& status)
856 {
857 if (U_FAILURE(status)) return;
858
859 parsePattern(); // Need this before initNumberFormatters(), to set fHasHanYearChar
860
861 // Simple-minded hack to force Gannen year numbering for ja@calendar=japanese
862 // if format is non-numeric (includes 年) and fDateOverride is not already specified.
863 // Now this does get updated if applyPattern subsequently changes the pattern type.
864 if (fDateOverride.isBogus() && fHasHanYearChar &&
865 fCalendar != nullptr && uprv_strcmp(fCalendar->getType(),"japanese") == 0 &&
866 uprv_strcmp(fLocale.getLanguage(),"ja") == 0) {
867 fDateOverride.setTo(u"y=jpanyear", -1);
868 }
869
870 // We don't need to check that the row count is >= 1, since all 2d arrays have at
871 // least one row
872 fNumberFormat = NumberFormat::createInstance(locale, status);
873 if (fNumberFormat != NULL && U_SUCCESS(status))
874 {
875 fixNumberFormatForDates(*fNumberFormat);
876 //fNumberFormat->setLenient(TRUE); // Java uses a custom DateNumberFormat to format/parse
877
878 initNumberFormatters(locale, status);
879 initFastNumberFormatters(status);
880
881 }
882 else if (U_SUCCESS(status))
883 {
884 status = U_MISSING_RESOURCE_ERROR;
885 }
886 }
887
888 /* Initialize the fields we use to disambiguate ambiguous years. Separate
889 * so we can call it from readObject().
890 */
initializeDefaultCentury()891 void SimpleDateFormat::initializeDefaultCentury()
892 {
893 if(fCalendar) {
894 fHaveDefaultCentury = fCalendar->haveDefaultCentury();
895 if(fHaveDefaultCentury) {
896 fDefaultCenturyStart = fCalendar->defaultCenturyStart();
897 fDefaultCenturyStartYear = fCalendar->defaultCenturyStartYear();
898 } else {
899 fDefaultCenturyStart = DBL_MIN;
900 fDefaultCenturyStartYear = -1;
901 }
902 }
903 }
904
905 /*
906 * Initialize the boolean attributes. Separate so we can call it from all constructors.
907 */
initializeBooleanAttributes()908 void SimpleDateFormat::initializeBooleanAttributes()
909 {
910 UErrorCode status = U_ZERO_ERROR;
911
912 setBooleanAttribute(UDAT_PARSE_ALLOW_WHITESPACE, true, status);
913 setBooleanAttribute(UDAT_PARSE_ALLOW_NUMERIC, true, status);
914 setBooleanAttribute(UDAT_PARSE_PARTIAL_LITERAL_MATCH, true, status);
915 setBooleanAttribute(UDAT_PARSE_MULTIPLE_PATTERNS_FOR_MATCH, true, status);
916 }
917
918 /* Define one-century window into which to disambiguate dates using
919 * two-digit years. Make public in JDK 1.2.
920 */
parseAmbiguousDatesAsAfter(UDate startDate,UErrorCode & status)921 void SimpleDateFormat::parseAmbiguousDatesAsAfter(UDate startDate, UErrorCode& status)
922 {
923 if(U_FAILURE(status)) {
924 return;
925 }
926 if(!fCalendar) {
927 status = U_ILLEGAL_ARGUMENT_ERROR;
928 return;
929 }
930
931 fCalendar->setTime(startDate, status);
932 if(U_SUCCESS(status)) {
933 fHaveDefaultCentury = TRUE;
934 fDefaultCenturyStart = startDate;
935 fDefaultCenturyStartYear = fCalendar->get(UCAL_YEAR, status);
936 }
937 }
938
939 //----------------------------------------------------------------------
940
941 UnicodeString&
format(Calendar & cal,UnicodeString & appendTo,FieldPosition & pos) const942 SimpleDateFormat::format(Calendar& cal, UnicodeString& appendTo, FieldPosition& pos) const
943 {
944 UErrorCode status = U_ZERO_ERROR;
945 FieldPositionOnlyHandler handler(pos);
946 return _format(cal, appendTo, handler, status);
947 }
948
949 //----------------------------------------------------------------------
950
951 UnicodeString&
format(Calendar & cal,UnicodeString & appendTo,FieldPositionIterator * posIter,UErrorCode & status) const952 SimpleDateFormat::format(Calendar& cal, UnicodeString& appendTo,
953 FieldPositionIterator* posIter, UErrorCode& status) const
954 {
955 FieldPositionIteratorHandler handler(posIter, status);
956 return _format(cal, appendTo, handler, status);
957 }
958
959 //----------------------------------------------------------------------
960
961 UnicodeString&
_format(Calendar & cal,UnicodeString & appendTo,FieldPositionHandler & handler,UErrorCode & status) const962 SimpleDateFormat::_format(Calendar& cal, UnicodeString& appendTo,
963 FieldPositionHandler& handler, UErrorCode& status) const
964 {
965 if ( U_FAILURE(status) ) {
966 return appendTo;
967 }
968 Calendar* workCal = &cal;
969 Calendar* calClone = NULL;
970 if (&cal != fCalendar && uprv_strcmp(cal.getType(), fCalendar->getType()) != 0) {
971 // Different calendar type
972 // We use the time and time zone from the input calendar, but
973 // do not use the input calendar for field calculation.
974 calClone = fCalendar->clone();
975 if (calClone != NULL) {
976 UDate t = cal.getTime(status);
977 calClone->setTime(t, status);
978 calClone->setTimeZone(cal.getTimeZone());
979 workCal = calClone;
980 } else {
981 status = U_MEMORY_ALLOCATION_ERROR;
982 return appendTo;
983 }
984 }
985
986 UBool inQuote = FALSE;
987 UChar prevCh = 0;
988 int32_t count = 0;
989 int32_t fieldNum = 0;
990 UDisplayContext capitalizationContext = getContext(UDISPCTX_TYPE_CAPITALIZATION, status);
991
992 // loop through the pattern string character by character
993 for (int32_t i = 0; i < fPattern.length() && U_SUCCESS(status); ++i) {
994 UChar ch = fPattern[i];
995
996 // Use subFormat() to format a repeated pattern character
997 // when a different pattern or non-pattern character is seen
998 if (ch != prevCh && count > 0) {
999 subFormat(appendTo, prevCh, count, capitalizationContext, fieldNum++, handler, *workCal, status);
1000 count = 0;
1001 }
1002 if (ch == QUOTE) {
1003 // Consecutive single quotes are a single quote literal,
1004 // either outside of quotes or between quotes
1005 if ((i+1) < fPattern.length() && fPattern[i+1] == QUOTE) {
1006 appendTo += (UChar)QUOTE;
1007 ++i;
1008 } else {
1009 inQuote = ! inQuote;
1010 }
1011 }
1012 else if (!inQuote && isSyntaxChar(ch)) {
1013 // ch is a date-time pattern character to be interpreted
1014 // by subFormat(); count the number of times it is repeated
1015 prevCh = ch;
1016 ++count;
1017 }
1018 else {
1019 // Append quoted characters and unquoted non-pattern characters
1020 appendTo += ch;
1021 }
1022 }
1023
1024 // Format the last item in the pattern, if any
1025 if (count > 0) {
1026 subFormat(appendTo, prevCh, count, capitalizationContext, fieldNum++, handler, *workCal, status);
1027 }
1028
1029 if (calClone != NULL) {
1030 delete calClone;
1031 }
1032
1033 return appendTo;
1034 }
1035
1036 //----------------------------------------------------------------------
1037
1038 /* Map calendar field into calendar field level.
1039 * the larger the level, the smaller the field unit.
1040 * For example, UCAL_ERA level is 0, UCAL_YEAR level is 10,
1041 * UCAL_MONTH level is 20.
1042 * NOTE: if new fields adds in, the table needs to update.
1043 */
1044 const int32_t
1045 SimpleDateFormat::fgCalendarFieldToLevel[] =
1046 {
1047 /*GyM*/ 0, 10, 20,
1048 /*wW*/ 20, 30,
1049 /*dDEF*/ 30, 20, 30, 30,
1050 /*ahHm*/ 40, 50, 50, 60,
1051 /*sS*/ 70, 80,
1052 /*z?Y*/ 0, 0, 10,
1053 /*eug*/ 30, 10, 0,
1054 /*A?.*/ 40, 0, 0
1055 };
1056
getLevelFromChar(UChar ch)1057 int32_t SimpleDateFormat::getLevelFromChar(UChar ch) {
1058 // Map date field LETTER into calendar field level.
1059 // the larger the level, the smaller the field unit.
1060 // NOTE: if new fields adds in, the table needs to update.
1061 static const int32_t mapCharToLevel[] = {
1062 -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
1063 //
1064 -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
1065 // ! " # $ % & ' ( ) * + , - . /
1066 -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
1067 #if UDAT_HAS_PATTERN_CHAR_FOR_TIME_SEPARATOR
1068 // 0 1 2 3 4 5 6 7 8 9 : ; < = > ?
1069 -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, 0, -1, -1, -1, -1, -1,
1070 #else
1071 // 0 1 2 3 4 5 6 7 8 9 : ; < = > ?
1072 -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
1073 #endif
1074 // @ A B C D E F G H I J K L M N O
1075 -1, 40, -1, -1, 20, 30, 30, 0, 50, -1, -1, 50, 20, 20, -1, 0,
1076 // P Q R S T U V W X Y Z [ \ ] ^ _
1077 -1, 20, -1, 80, -1, 10, 0, 30, 0, 10, 0, -1, -1, -1, -1, -1,
1078 // ` a b c d e f g h i j k l m n o
1079 -1, 40, -1, 30, 30, 30, -1, 0, 50, -1, -1, 50, 0, 60, -1, -1,
1080 // p q r s t u v w x y z { | } ~
1081 -1, 20, 10, 70, -1, 10, 0, 20, 0, 10, 0, -1, -1, -1, -1, -1
1082 };
1083
1084 return ch < UPRV_LENGTHOF(mapCharToLevel) ? mapCharToLevel[ch] : -1;
1085 }
1086
isSyntaxChar(UChar ch)1087 UBool SimpleDateFormat::isSyntaxChar(UChar ch) {
1088 static const UBool mapCharToIsSyntax[] = {
1089 //
1090 FALSE, FALSE, FALSE, FALSE, FALSE, FALSE, FALSE, FALSE,
1091 //
1092 FALSE, FALSE, FALSE, FALSE, FALSE, FALSE, FALSE, FALSE,
1093 //
1094 FALSE, FALSE, FALSE, FALSE, FALSE, FALSE, FALSE, FALSE,
1095 //
1096 FALSE, FALSE, FALSE, FALSE, FALSE, FALSE, FALSE, FALSE,
1097 // ! " # $ % & '
1098 FALSE, FALSE, FALSE, FALSE, FALSE, FALSE, FALSE, FALSE,
1099 // ( ) * + , - . /
1100 FALSE, FALSE, FALSE, FALSE, FALSE, FALSE, FALSE, FALSE,
1101 // 0 1 2 3 4 5 6 7
1102 FALSE, FALSE, FALSE, FALSE, FALSE, FALSE, FALSE, FALSE,
1103 #if UDAT_HAS_PATTERN_CHAR_FOR_TIME_SEPARATOR
1104 // 8 9 : ; < = > ?
1105 FALSE, FALSE, TRUE, FALSE, FALSE, FALSE, FALSE, FALSE,
1106 #else
1107 // 8 9 : ; < = > ?
1108 FALSE, FALSE, FALSE, FALSE, FALSE, FALSE, FALSE, FALSE,
1109 #endif
1110 // @ A B C D E F G
1111 FALSE, TRUE, TRUE, TRUE, TRUE, TRUE, TRUE, TRUE,
1112 // H I J K L M N O
1113 TRUE, TRUE, TRUE, TRUE, TRUE, TRUE, TRUE, TRUE,
1114 // P Q R S T U V W
1115 TRUE, TRUE, TRUE, TRUE, TRUE, TRUE, TRUE, TRUE,
1116 // X Y Z [ \ ] ^ _
1117 TRUE, TRUE, TRUE, FALSE, FALSE, FALSE, FALSE, FALSE,
1118 // ` a b c d e f g
1119 FALSE, TRUE, TRUE, TRUE, TRUE, TRUE, TRUE, TRUE,
1120 // h i j k l m n o
1121 TRUE, TRUE, TRUE, TRUE, TRUE, TRUE, TRUE, TRUE,
1122 // p q r s t u v w
1123 TRUE, TRUE, TRUE, TRUE, TRUE, TRUE, TRUE, TRUE,
1124 // x y z { | } ~
1125 TRUE, TRUE, TRUE, FALSE, FALSE, FALSE, FALSE, FALSE
1126 };
1127
1128 return ch < UPRV_LENGTHOF(mapCharToIsSyntax) ? mapCharToIsSyntax[ch] : FALSE;
1129 }
1130
1131 // Map index into pattern character string to Calendar field number.
1132 const UCalendarDateFields
1133 SimpleDateFormat::fgPatternIndexToCalendarField[] =
1134 {
1135 /*GyM*/ UCAL_ERA, UCAL_YEAR, UCAL_MONTH,
1136 /*dkH*/ UCAL_DATE, UCAL_HOUR_OF_DAY, UCAL_HOUR_OF_DAY,
1137 /*msS*/ UCAL_MINUTE, UCAL_SECOND, UCAL_MILLISECOND,
1138 /*EDF*/ UCAL_DAY_OF_WEEK, UCAL_DAY_OF_YEAR, UCAL_DAY_OF_WEEK_IN_MONTH,
1139 /*wWa*/ UCAL_WEEK_OF_YEAR, UCAL_WEEK_OF_MONTH, UCAL_AM_PM,
1140 /*hKz*/ UCAL_HOUR, UCAL_HOUR, UCAL_ZONE_OFFSET,
1141 /*Yeu*/ UCAL_YEAR_WOY, UCAL_DOW_LOCAL, UCAL_EXTENDED_YEAR,
1142 /*gAZ*/ UCAL_JULIAN_DAY, UCAL_MILLISECONDS_IN_DAY, UCAL_ZONE_OFFSET,
1143 /*v*/ UCAL_ZONE_OFFSET,
1144 /*c*/ UCAL_DOW_LOCAL,
1145 /*L*/ UCAL_MONTH,
1146 /*Q*/ UCAL_MONTH,
1147 /*q*/ UCAL_MONTH,
1148 /*V*/ UCAL_ZONE_OFFSET,
1149 /*U*/ UCAL_YEAR,
1150 /*O*/ UCAL_ZONE_OFFSET,
1151 /*Xx*/ UCAL_ZONE_OFFSET, UCAL_ZONE_OFFSET,
1152 /*r*/ UCAL_EXTENDED_YEAR,
1153 /*bB*/ UCAL_FIELD_COUNT, UCAL_FIELD_COUNT, // no mappings to calendar fields
1154 #if UDAT_HAS_PATTERN_CHAR_FOR_TIME_SEPARATOR
1155 /*:*/ UCAL_FIELD_COUNT, /* => no useful mapping to any calendar field */
1156 #else
1157 /*no pattern char for UDAT_TIME_SEPARATOR_FIELD*/ UCAL_FIELD_COUNT, /* => no useful mapping to any calendar field */
1158 #endif
1159 };
1160
1161 // Map index into pattern character string to DateFormat field number
1162 const UDateFormatField
1163 SimpleDateFormat::fgPatternIndexToDateFormatField[] = {
1164 /*GyM*/ UDAT_ERA_FIELD, UDAT_YEAR_FIELD, UDAT_MONTH_FIELD,
1165 /*dkH*/ UDAT_DATE_FIELD, UDAT_HOUR_OF_DAY1_FIELD, UDAT_HOUR_OF_DAY0_FIELD,
1166 /*msS*/ UDAT_MINUTE_FIELD, UDAT_SECOND_FIELD, UDAT_FRACTIONAL_SECOND_FIELD,
1167 /*EDF*/ UDAT_DAY_OF_WEEK_FIELD, UDAT_DAY_OF_YEAR_FIELD, UDAT_DAY_OF_WEEK_IN_MONTH_FIELD,
1168 /*wWa*/ UDAT_WEEK_OF_YEAR_FIELD, UDAT_WEEK_OF_MONTH_FIELD, UDAT_AM_PM_FIELD,
1169 /*hKz*/ UDAT_HOUR1_FIELD, UDAT_HOUR0_FIELD, UDAT_TIMEZONE_FIELD,
1170 /*Yeu*/ UDAT_YEAR_WOY_FIELD, UDAT_DOW_LOCAL_FIELD, UDAT_EXTENDED_YEAR_FIELD,
1171 /*gAZ*/ UDAT_JULIAN_DAY_FIELD, UDAT_MILLISECONDS_IN_DAY_FIELD, UDAT_TIMEZONE_RFC_FIELD,
1172 /*v*/ UDAT_TIMEZONE_GENERIC_FIELD,
1173 /*c*/ UDAT_STANDALONE_DAY_FIELD,
1174 /*L*/ UDAT_STANDALONE_MONTH_FIELD,
1175 /*Q*/ UDAT_QUARTER_FIELD,
1176 /*q*/ UDAT_STANDALONE_QUARTER_FIELD,
1177 /*V*/ UDAT_TIMEZONE_SPECIAL_FIELD,
1178 /*U*/ UDAT_YEAR_NAME_FIELD,
1179 /*O*/ UDAT_TIMEZONE_LOCALIZED_GMT_OFFSET_FIELD,
1180 /*Xx*/ UDAT_TIMEZONE_ISO_FIELD, UDAT_TIMEZONE_ISO_LOCAL_FIELD,
1181 /*r*/ UDAT_RELATED_YEAR_FIELD,
1182 /*bB*/ UDAT_AM_PM_MIDNIGHT_NOON_FIELD, UDAT_FLEXIBLE_DAY_PERIOD_FIELD,
1183 #if UDAT_HAS_PATTERN_CHAR_FOR_TIME_SEPARATOR
1184 /*:*/ UDAT_TIME_SEPARATOR_FIELD,
1185 #else
1186 /*no pattern char for UDAT_TIME_SEPARATOR_FIELD*/ UDAT_TIME_SEPARATOR_FIELD,
1187 #endif
1188 };
1189
1190 //----------------------------------------------------------------------
1191
1192 /**
1193 * Append symbols[value] to dst. Make sure the array index is not out
1194 * of bounds.
1195 */
1196 static inline void
_appendSymbol(UnicodeString & dst,int32_t value,const UnicodeString * symbols,int32_t symbolsCount)1197 _appendSymbol(UnicodeString& dst,
1198 int32_t value,
1199 const UnicodeString* symbols,
1200 int32_t symbolsCount) {
1201 U_ASSERT(0 <= value && value < symbolsCount);
1202 if (0 <= value && value < symbolsCount) {
1203 dst += symbols[value];
1204 }
1205 }
1206
1207 static inline void
_appendSymbolWithMonthPattern(UnicodeString & dst,int32_t value,const UnicodeString * symbols,int32_t symbolsCount,const UnicodeString * monthPattern,UErrorCode & status)1208 _appendSymbolWithMonthPattern(UnicodeString& dst, int32_t value, const UnicodeString* symbols, int32_t symbolsCount,
1209 const UnicodeString* monthPattern, UErrorCode& status) {
1210 U_ASSERT(0 <= value && value < symbolsCount);
1211 if (0 <= value && value < symbolsCount) {
1212 if (monthPattern == NULL) {
1213 dst += symbols[value];
1214 } else {
1215 SimpleFormatter(*monthPattern, 1, 1, status).format(symbols[value], dst, status);
1216 }
1217 }
1218 }
1219
1220 //----------------------------------------------------------------------
1221
1222 static number::LocalizedNumberFormatter*
createFastFormatter(const DecimalFormat * df,int32_t minInt,int32_t maxInt,UErrorCode & status)1223 createFastFormatter(const DecimalFormat* df, int32_t minInt, int32_t maxInt, UErrorCode& status) {
1224 const number::LocalizedNumberFormatter* lnfBase = df->toNumberFormatter(status);
1225 if (U_FAILURE(status)) {
1226 return nullptr;
1227 }
1228 return lnfBase->integerWidth(
1229 number::IntegerWidth::zeroFillTo(minInt).truncateAt(maxInt)
1230 ).clone().orphan();
1231 }
1232
initFastNumberFormatters(UErrorCode & status)1233 void SimpleDateFormat::initFastNumberFormatters(UErrorCode& status) {
1234 if (U_FAILURE(status)) {
1235 return;
1236 }
1237 auto* df = dynamic_cast<const DecimalFormat*>(fNumberFormat);
1238 if (df == nullptr) {
1239 return;
1240 }
1241 fFastNumberFormatters[SMPDTFMT_NF_1x10] = createFastFormatter(df, 1, 10, status);
1242 fFastNumberFormatters[SMPDTFMT_NF_2x10] = createFastFormatter(df, 2, 10, status);
1243 fFastNumberFormatters[SMPDTFMT_NF_3x10] = createFastFormatter(df, 3, 10, status);
1244 fFastNumberFormatters[SMPDTFMT_NF_4x10] = createFastFormatter(df, 4, 10, status);
1245 fFastNumberFormatters[SMPDTFMT_NF_2x2] = createFastFormatter(df, 2, 2, status);
1246 }
1247
freeFastNumberFormatters()1248 void SimpleDateFormat::freeFastNumberFormatters() {
1249 delete fFastNumberFormatters[SMPDTFMT_NF_1x10];
1250 delete fFastNumberFormatters[SMPDTFMT_NF_2x10];
1251 delete fFastNumberFormatters[SMPDTFMT_NF_3x10];
1252 delete fFastNumberFormatters[SMPDTFMT_NF_4x10];
1253 delete fFastNumberFormatters[SMPDTFMT_NF_2x2];
1254 fFastNumberFormatters[SMPDTFMT_NF_1x10] = nullptr;
1255 fFastNumberFormatters[SMPDTFMT_NF_2x10] = nullptr;
1256 fFastNumberFormatters[SMPDTFMT_NF_3x10] = nullptr;
1257 fFastNumberFormatters[SMPDTFMT_NF_4x10] = nullptr;
1258 fFastNumberFormatters[SMPDTFMT_NF_2x2] = nullptr;
1259 }
1260
1261
1262 void
initNumberFormatters(const Locale & locale,UErrorCode & status)1263 SimpleDateFormat::initNumberFormatters(const Locale &locale,UErrorCode &status) {
1264 if (U_FAILURE(status)) {
1265 return;
1266 }
1267 if ( fDateOverride.isBogus() && fTimeOverride.isBogus() ) {
1268 return;
1269 }
1270 umtx_lock(&LOCK);
1271 if (fSharedNumberFormatters == NULL) {
1272 fSharedNumberFormatters = allocSharedNumberFormatters();
1273 if (fSharedNumberFormatters == NULL) {
1274 status = U_MEMORY_ALLOCATION_ERROR;
1275 }
1276 }
1277 umtx_unlock(&LOCK);
1278
1279 if (U_FAILURE(status)) {
1280 return;
1281 }
1282
1283 processOverrideString(locale,fDateOverride,kOvrStrDate,status);
1284 processOverrideString(locale,fTimeOverride,kOvrStrTime,status);
1285 }
1286
1287 void
processOverrideString(const Locale & locale,const UnicodeString & str,int8_t type,UErrorCode & status)1288 SimpleDateFormat::processOverrideString(const Locale &locale, const UnicodeString &str, int8_t type, UErrorCode &status) {
1289 if (str.isBogus() || U_FAILURE(status)) {
1290 return;
1291 }
1292
1293 int32_t start = 0;
1294 int32_t len;
1295 UnicodeString nsName;
1296 UnicodeString ovrField;
1297 UBool moreToProcess = TRUE;
1298 NSOverride *overrideList = NULL;
1299
1300 while (moreToProcess) {
1301 int32_t delimiterPosition = str.indexOf((UChar)ULOC_KEYWORD_ITEM_SEPARATOR_UNICODE,start);
1302 if (delimiterPosition == -1) {
1303 moreToProcess = FALSE;
1304 len = str.length() - start;
1305 } else {
1306 len = delimiterPosition - start;
1307 }
1308 UnicodeString currentString(str,start,len);
1309 int32_t equalSignPosition = currentString.indexOf((UChar)ULOC_KEYWORD_ASSIGN_UNICODE,0);
1310 if (equalSignPosition == -1) { // Simple override string such as "hebrew"
1311 nsName.setTo(currentString);
1312 ovrField.setToBogus();
1313 } else { // Field specific override string such as "y=hebrew"
1314 nsName.setTo(currentString,equalSignPosition+1);
1315 ovrField.setTo(currentString,0,1); // We just need the first character.
1316 }
1317
1318 int32_t nsNameHash = nsName.hashCode();
1319 // See if the numbering system is in the override list, if not, then add it.
1320 NSOverride *curr = overrideList;
1321 const SharedNumberFormat *snf = NULL;
1322 UBool found = FALSE;
1323 while ( curr && !found ) {
1324 if ( curr->hash == nsNameHash ) {
1325 snf = curr->snf;
1326 found = TRUE;
1327 }
1328 curr = curr->next;
1329 }
1330
1331 if (!found) {
1332 LocalPointer<NSOverride> cur(new NSOverride);
1333 if (!cur.isNull()) {
1334 char kw[ULOC_KEYWORD_AND_VALUES_CAPACITY];
1335 uprv_strcpy(kw,"numbers=");
1336 nsName.extract(0,len,kw+8,ULOC_KEYWORD_AND_VALUES_CAPACITY-8,US_INV);
1337
1338 Locale ovrLoc(locale.getLanguage(),locale.getCountry(),locale.getVariant(),kw);
1339 cur->hash = nsNameHash;
1340 cur->next = overrideList;
1341 SharedObject::copyPtr(
1342 createSharedNumberFormat(ovrLoc, status), cur->snf);
1343 if (U_FAILURE(status)) {
1344 if (overrideList) {
1345 overrideList->free();
1346 }
1347 return;
1348 }
1349 snf = cur->snf;
1350 overrideList = cur.orphan();
1351 } else {
1352 status = U_MEMORY_ALLOCATION_ERROR;
1353 if (overrideList) {
1354 overrideList->free();
1355 }
1356 return;
1357 }
1358 }
1359
1360 // Now that we have an appropriate number formatter, fill in the appropriate spaces in the
1361 // number formatters table.
1362 if (ovrField.isBogus()) {
1363 switch (type) {
1364 case kOvrStrDate:
1365 case kOvrStrBoth: {
1366 for ( int8_t i=0 ; i<kDateFieldsCount; i++ ) {
1367 SharedObject::copyPtr(snf, fSharedNumberFormatters[kDateFields[i]]);
1368 }
1369 if (type==kOvrStrDate) {
1370 break;
1371 }
1372 U_FALLTHROUGH;
1373 }
1374 case kOvrStrTime : {
1375 for ( int8_t i=0 ; i<kTimeFieldsCount; i++ ) {
1376 SharedObject::copyPtr(snf, fSharedNumberFormatters[kTimeFields[i]]);
1377 }
1378 break;
1379 }
1380 }
1381 } else {
1382 // if the pattern character is unrecognized, signal an error and bail out
1383 UDateFormatField patternCharIndex =
1384 DateFormatSymbols::getPatternCharIndex(ovrField.charAt(0));
1385 if (patternCharIndex == UDAT_FIELD_COUNT) {
1386 status = U_INVALID_FORMAT_ERROR;
1387 if (overrideList) {
1388 overrideList->free();
1389 }
1390 return;
1391 }
1392 SharedObject::copyPtr(snf, fSharedNumberFormatters[patternCharIndex]);
1393 }
1394
1395 start = delimiterPosition + 1;
1396 }
1397 if (overrideList) {
1398 overrideList->free();
1399 }
1400 }
1401
1402 //---------------------------------------------------------------------
1403 void
subFormat(UnicodeString & appendTo,UChar ch,int32_t count,UDisplayContext capitalizationContext,int32_t fieldNum,FieldPositionHandler & handler,Calendar & cal,UErrorCode & status) const1404 SimpleDateFormat::subFormat(UnicodeString &appendTo,
1405 UChar ch,
1406 int32_t count,
1407 UDisplayContext capitalizationContext,
1408 int32_t fieldNum,
1409 FieldPositionHandler& handler,
1410 Calendar& cal,
1411 UErrorCode& status) const
1412 {
1413 if (U_FAILURE(status)) {
1414 return;
1415 }
1416
1417 // this function gets called by format() to produce the appropriate substitution
1418 // text for an individual pattern symbol (e.g., "HH" or "yyyy")
1419
1420 UDateFormatField patternCharIndex = DateFormatSymbols::getPatternCharIndex(ch);
1421 const int32_t maxIntCount = 10;
1422 int32_t beginOffset = appendTo.length();
1423 const NumberFormat *currentNumberFormat;
1424 DateFormatSymbols::ECapitalizationContextUsageType capContextUsageType = DateFormatSymbols::kCapContextUsageOther;
1425
1426 UBool isHebrewCalendar = (uprv_strcmp(cal.getType(),"hebrew") == 0);
1427 UBool isChineseCalendar = (uprv_strcmp(cal.getType(),"chinese") == 0 || uprv_strcmp(cal.getType(),"dangi") == 0);
1428
1429 // if the pattern character is unrecognized, signal an error and dump out
1430 if (patternCharIndex == UDAT_FIELD_COUNT)
1431 {
1432 if (ch != 0x6C) { // pattern char 'l' (SMALL LETTER L) just gets ignored
1433 status = U_INVALID_FORMAT_ERROR;
1434 }
1435 return;
1436 }
1437
1438 UCalendarDateFields field = fgPatternIndexToCalendarField[patternCharIndex];
1439 int32_t value = 0;
1440 // Don't get value unless it is useful
1441 if (field < UCAL_FIELD_COUNT) {
1442 value = (patternCharIndex != UDAT_RELATED_YEAR_FIELD)? cal.get(field, status): cal.getRelatedYear(status);
1443 }
1444 if (U_FAILURE(status)) {
1445 return;
1446 }
1447
1448 currentNumberFormat = getNumberFormatByIndex(patternCharIndex);
1449 if (currentNumberFormat == NULL) {
1450 status = U_INTERNAL_PROGRAM_ERROR;
1451 return;
1452 }
1453 UnicodeString hebr("hebr", 4, US_INV);
1454
1455 switch (patternCharIndex) {
1456
1457 // for any "G" symbol, write out the appropriate era string
1458 // "GGGG" is wide era name, "GGGGG" is narrow era name, anything else is abbreviated name
1459 case UDAT_ERA_FIELD:
1460 if (isChineseCalendar) {
1461 zeroPaddingNumber(currentNumberFormat,appendTo, value, 1, 9); // as in ICU4J
1462 } else {
1463 if (count == 5) {
1464 _appendSymbol(appendTo, value, fSymbols->fNarrowEras, fSymbols->fNarrowErasCount);
1465 capContextUsageType = DateFormatSymbols::kCapContextUsageEraNarrow;
1466 } else if (count == 4) {
1467 _appendSymbol(appendTo, value, fSymbols->fEraNames, fSymbols->fEraNamesCount);
1468 capContextUsageType = DateFormatSymbols::kCapContextUsageEraWide;
1469 } else {
1470 _appendSymbol(appendTo, value, fSymbols->fEras, fSymbols->fErasCount);
1471 capContextUsageType = DateFormatSymbols::kCapContextUsageEraAbbrev;
1472 }
1473 }
1474 break;
1475
1476 case UDAT_YEAR_NAME_FIELD:
1477 if (fSymbols->fShortYearNames != NULL && value <= fSymbols->fShortYearNamesCount) {
1478 // the Calendar YEAR field runs 1 through 60 for cyclic years
1479 _appendSymbol(appendTo, value - 1, fSymbols->fShortYearNames, fSymbols->fShortYearNamesCount);
1480 break;
1481 }
1482 // else fall through to numeric year handling, do not break here
1483 U_FALLTHROUGH;
1484
1485 // OLD: for "yyyy", write out the whole year; for "yy", write out the last 2 digits
1486 // NEW: UTS#35:
1487 //Year y yy yyy yyyy yyyyy
1488 //AD 1 1 01 001 0001 00001
1489 //AD 12 12 12 012 0012 00012
1490 //AD 123 123 23 123 0123 00123
1491 //AD 1234 1234 34 1234 1234 01234
1492 //AD 12345 12345 45 12345 12345 12345
1493 case UDAT_YEAR_FIELD:
1494 case UDAT_YEAR_WOY_FIELD:
1495 if (fDateOverride.compare(hebr)==0 && value>HEBREW_CAL_CUR_MILLENIUM_START_YEAR && value<HEBREW_CAL_CUR_MILLENIUM_END_YEAR) {
1496 value-=HEBREW_CAL_CUR_MILLENIUM_START_YEAR;
1497 }
1498 if(count == 2)
1499 zeroPaddingNumber(currentNumberFormat, appendTo, value, 2, 2);
1500 else
1501 zeroPaddingNumber(currentNumberFormat, appendTo, value, count, maxIntCount);
1502 break;
1503
1504 // for "MMMM"/"LLLL", write out the whole month name, for "MMM"/"LLL", write out the month
1505 // abbreviation, for "M"/"L" or "MM"/"LL", write out the month as a number with the
1506 // appropriate number of digits
1507 // for "MMMMM"/"LLLLL", use the narrow form
1508 case UDAT_MONTH_FIELD:
1509 case UDAT_STANDALONE_MONTH_FIELD:
1510 if ( isHebrewCalendar ) {
1511 HebrewCalendar *hc = (HebrewCalendar*)&cal;
1512 if (hc->isLeapYear(hc->get(UCAL_YEAR,status)) && value == 6 && count >= 3 )
1513 value = 13; // Show alternate form for Adar II in leap years in Hebrew calendar.
1514 if (!hc->isLeapYear(hc->get(UCAL_YEAR,status)) && value >= 6 && count < 3 )
1515 value--; // Adjust the month number down 1 in Hebrew non-leap years, i.e. Adar is 6, not 7.
1516 }
1517 {
1518 int32_t isLeapMonth = (fSymbols->fLeapMonthPatterns != NULL && fSymbols->fLeapMonthPatternsCount >= DateFormatSymbols::kMonthPatternsCount)?
1519 cal.get(UCAL_IS_LEAP_MONTH, status): 0;
1520 // should consolidate the next section by using arrays of pointers & counts for the right symbols...
1521 if (count == 5) {
1522 if (patternCharIndex == UDAT_MONTH_FIELD) {
1523 _appendSymbolWithMonthPattern(appendTo, value, fSymbols->fNarrowMonths, fSymbols->fNarrowMonthsCount,
1524 (isLeapMonth!=0)? &(fSymbols->fLeapMonthPatterns[DateFormatSymbols::kLeapMonthPatternFormatNarrow]): NULL, status);
1525 } else {
1526 _appendSymbolWithMonthPattern(appendTo, value, fSymbols->fStandaloneNarrowMonths, fSymbols->fStandaloneNarrowMonthsCount,
1527 (isLeapMonth!=0)? &(fSymbols->fLeapMonthPatterns[DateFormatSymbols::kLeapMonthPatternStandaloneNarrow]): NULL, status);
1528 }
1529 capContextUsageType = DateFormatSymbols::kCapContextUsageMonthNarrow;
1530 } else if (count == 4) {
1531 if (patternCharIndex == UDAT_MONTH_FIELD) {
1532 _appendSymbolWithMonthPattern(appendTo, value, fSymbols->fMonths, fSymbols->fMonthsCount,
1533 (isLeapMonth!=0)? &(fSymbols->fLeapMonthPatterns[DateFormatSymbols::kLeapMonthPatternFormatWide]): NULL, status);
1534 capContextUsageType = DateFormatSymbols::kCapContextUsageMonthFormat;
1535 } else {
1536 _appendSymbolWithMonthPattern(appendTo, value, fSymbols->fStandaloneMonths, fSymbols->fStandaloneMonthsCount,
1537 (isLeapMonth!=0)? &(fSymbols->fLeapMonthPatterns[DateFormatSymbols::kLeapMonthPatternStandaloneWide]): NULL, status);
1538 capContextUsageType = DateFormatSymbols::kCapContextUsageMonthStandalone;
1539 }
1540 } else if (count == 3) {
1541 if (patternCharIndex == UDAT_MONTH_FIELD) {
1542 _appendSymbolWithMonthPattern(appendTo, value, fSymbols->fShortMonths, fSymbols->fShortMonthsCount,
1543 (isLeapMonth!=0)? &(fSymbols->fLeapMonthPatterns[DateFormatSymbols::kLeapMonthPatternFormatAbbrev]): NULL, status);
1544 capContextUsageType = DateFormatSymbols::kCapContextUsageMonthFormat;
1545 } else {
1546 _appendSymbolWithMonthPattern(appendTo, value, fSymbols->fStandaloneShortMonths, fSymbols->fStandaloneShortMonthsCount,
1547 (isLeapMonth!=0)? &(fSymbols->fLeapMonthPatterns[DateFormatSymbols::kLeapMonthPatternStandaloneAbbrev]): NULL, status);
1548 capContextUsageType = DateFormatSymbols::kCapContextUsageMonthStandalone;
1549 }
1550 } else {
1551 UnicodeString monthNumber;
1552 zeroPaddingNumber(currentNumberFormat,monthNumber, value + 1, count, maxIntCount);
1553 _appendSymbolWithMonthPattern(appendTo, 0, &monthNumber, 1,
1554 (isLeapMonth!=0)? &(fSymbols->fLeapMonthPatterns[DateFormatSymbols::kLeapMonthPatternNumeric]): NULL, status);
1555 }
1556 }
1557 break;
1558
1559 // for "k" and "kk", write out the hour, adjusting midnight to appear as "24"
1560 case UDAT_HOUR_OF_DAY1_FIELD:
1561 if (value == 0)
1562 zeroPaddingNumber(currentNumberFormat,appendTo, cal.getMaximum(UCAL_HOUR_OF_DAY) + 1, count, maxIntCount);
1563 else
1564 zeroPaddingNumber(currentNumberFormat,appendTo, value, count, maxIntCount);
1565 break;
1566
1567 case UDAT_FRACTIONAL_SECOND_FIELD:
1568 // Fractional seconds left-justify
1569 {
1570 int32_t minDigits = (count > 3) ? 3 : count;
1571 if (count == 1) {
1572 value /= 100;
1573 } else if (count == 2) {
1574 value /= 10;
1575 }
1576 zeroPaddingNumber(currentNumberFormat, appendTo, value, minDigits, maxIntCount);
1577 if (count > 3) {
1578 zeroPaddingNumber(currentNumberFormat, appendTo, 0, count - 3, maxIntCount);
1579 }
1580 }
1581 break;
1582
1583 // for "ee" or "e", use local numeric day-of-the-week
1584 // for "EEEEEE" or "eeeeee", write out the short day-of-the-week name
1585 // for "EEEEE" or "eeeee", write out the narrow day-of-the-week name
1586 // for "EEEE" or "eeee", write out the wide day-of-the-week name
1587 // for "EEE" or "EE" or "E" or "eee", write out the abbreviated day-of-the-week name
1588 case UDAT_DOW_LOCAL_FIELD:
1589 if ( count < 3 ) {
1590 zeroPaddingNumber(currentNumberFormat,appendTo, value, count, maxIntCount);
1591 break;
1592 }
1593 // fall through to EEEEE-EEE handling, but for that we don't want local day-of-week,
1594 // we want standard day-of-week, so first fix value to work for EEEEE-EEE.
1595 value = cal.get(UCAL_DAY_OF_WEEK, status);
1596 if (U_FAILURE(status)) {
1597 return;
1598 }
1599 // fall through, do not break here
1600 U_FALLTHROUGH;
1601 case UDAT_DAY_OF_WEEK_FIELD:
1602 if (count == 5) {
1603 _appendSymbol(appendTo, value, fSymbols->fNarrowWeekdays,
1604 fSymbols->fNarrowWeekdaysCount);
1605 capContextUsageType = DateFormatSymbols::kCapContextUsageDayNarrow;
1606 } else if (count == 4) {
1607 _appendSymbol(appendTo, value, fSymbols->fWeekdays,
1608 fSymbols->fWeekdaysCount);
1609 capContextUsageType = DateFormatSymbols::kCapContextUsageDayFormat;
1610 } else if (count == 6) {
1611 _appendSymbol(appendTo, value, fSymbols->fShorterWeekdays,
1612 fSymbols->fShorterWeekdaysCount);
1613 capContextUsageType = DateFormatSymbols::kCapContextUsageDayFormat;
1614 } else {
1615 _appendSymbol(appendTo, value, fSymbols->fShortWeekdays,
1616 fSymbols->fShortWeekdaysCount);
1617 capContextUsageType = DateFormatSymbols::kCapContextUsageDayFormat;
1618 }
1619 break;
1620
1621 // for "ccc", write out the abbreviated day-of-the-week name
1622 // for "cccc", write out the wide day-of-the-week name
1623 // for "ccccc", use the narrow day-of-the-week name
1624 // for "ccccc", use the short day-of-the-week name
1625 case UDAT_STANDALONE_DAY_FIELD:
1626 if ( count < 3 ) {
1627 zeroPaddingNumber(currentNumberFormat,appendTo, value, 1, maxIntCount);
1628 break;
1629 }
1630 // fall through to alpha DOW handling, but for that we don't want local day-of-week,
1631 // we want standard day-of-week, so first fix value.
1632 value = cal.get(UCAL_DAY_OF_WEEK, status);
1633 if (U_FAILURE(status)) {
1634 return;
1635 }
1636 if (count == 5) {
1637 _appendSymbol(appendTo, value, fSymbols->fStandaloneNarrowWeekdays,
1638 fSymbols->fStandaloneNarrowWeekdaysCount);
1639 capContextUsageType = DateFormatSymbols::kCapContextUsageDayNarrow;
1640 } else if (count == 4) {
1641 _appendSymbol(appendTo, value, fSymbols->fStandaloneWeekdays,
1642 fSymbols->fStandaloneWeekdaysCount);
1643 capContextUsageType = DateFormatSymbols::kCapContextUsageDayStandalone;
1644 } else if (count == 6) {
1645 _appendSymbol(appendTo, value, fSymbols->fStandaloneShorterWeekdays,
1646 fSymbols->fStandaloneShorterWeekdaysCount);
1647 capContextUsageType = DateFormatSymbols::kCapContextUsageDayStandalone;
1648 } else { // count == 3
1649 _appendSymbol(appendTo, value, fSymbols->fStandaloneShortWeekdays,
1650 fSymbols->fStandaloneShortWeekdaysCount);
1651 capContextUsageType = DateFormatSymbols::kCapContextUsageDayStandalone;
1652 }
1653 break;
1654
1655 // for "a" symbol, write out the whole AM/PM string
1656 case UDAT_AM_PM_FIELD:
1657 if (count < 5) {
1658 _appendSymbol(appendTo, value, fSymbols->fAmPms,
1659 fSymbols->fAmPmsCount);
1660 } else {
1661 _appendSymbol(appendTo, value, fSymbols->fNarrowAmPms,
1662 fSymbols->fNarrowAmPmsCount);
1663 }
1664 break;
1665
1666 // if we see pattern character for UDAT_TIME_SEPARATOR_FIELD (none currently defined),
1667 // write out the time separator string. Leave support in for future definition.
1668 case UDAT_TIME_SEPARATOR_FIELD:
1669 {
1670 UnicodeString separator;
1671 appendTo += fSymbols->getTimeSeparatorString(separator);
1672 }
1673 break;
1674
1675 // for "h" and "hh", write out the hour, adjusting noon and midnight to show up
1676 // as "12"
1677 case UDAT_HOUR1_FIELD:
1678 if (value == 0)
1679 zeroPaddingNumber(currentNumberFormat,appendTo, cal.getLeastMaximum(UCAL_HOUR) + 1, count, maxIntCount);
1680 else
1681 zeroPaddingNumber(currentNumberFormat,appendTo, value, count, maxIntCount);
1682 break;
1683
1684 case UDAT_TIMEZONE_FIELD: // 'z'
1685 case UDAT_TIMEZONE_RFC_FIELD: // 'Z'
1686 case UDAT_TIMEZONE_GENERIC_FIELD: // 'v'
1687 case UDAT_TIMEZONE_SPECIAL_FIELD: // 'V'
1688 case UDAT_TIMEZONE_LOCALIZED_GMT_OFFSET_FIELD: // 'O'
1689 case UDAT_TIMEZONE_ISO_FIELD: // 'X'
1690 case UDAT_TIMEZONE_ISO_LOCAL_FIELD: // 'x'
1691 {
1692 UChar zsbuf[ZONE_NAME_U16_MAX];
1693 UnicodeString zoneString(zsbuf, 0, UPRV_LENGTHOF(zsbuf));
1694 const TimeZone& tz = cal.getTimeZone();
1695 UDate date = cal.getTime(status);
1696 const TimeZoneFormat *tzfmt = tzFormat(status);
1697 if (U_SUCCESS(status)) {
1698 if (patternCharIndex == UDAT_TIMEZONE_FIELD) {
1699 if (count < 4) {
1700 // "z", "zz", "zzz"
1701 tzfmt->format(UTZFMT_STYLE_SPECIFIC_SHORT, tz, date, zoneString);
1702 capContextUsageType = DateFormatSymbols::kCapContextUsageMetazoneShort;
1703 } else {
1704 // "zzzz" or longer
1705 tzfmt->format(UTZFMT_STYLE_SPECIFIC_LONG, tz, date, zoneString);
1706 capContextUsageType = DateFormatSymbols::kCapContextUsageMetazoneLong;
1707 }
1708 }
1709 else if (patternCharIndex == UDAT_TIMEZONE_RFC_FIELD) {
1710 if (count < 4) {
1711 // "Z"
1712 tzfmt->format(UTZFMT_STYLE_ISO_BASIC_LOCAL_FULL, tz, date, zoneString);
1713 } else if (count == 5) {
1714 // "ZZZZZ"
1715 tzfmt->format(UTZFMT_STYLE_ISO_EXTENDED_FULL, tz, date, zoneString);
1716 } else {
1717 // "ZZ", "ZZZ", "ZZZZ"
1718 tzfmt->format(UTZFMT_STYLE_LOCALIZED_GMT, tz, date, zoneString);
1719 }
1720 }
1721 else if (patternCharIndex == UDAT_TIMEZONE_GENERIC_FIELD) {
1722 if (count == 1) {
1723 // "v"
1724 tzfmt->format(UTZFMT_STYLE_GENERIC_SHORT, tz, date, zoneString);
1725 capContextUsageType = DateFormatSymbols::kCapContextUsageMetazoneShort;
1726 } else if (count == 4) {
1727 // "vvvv"
1728 tzfmt->format(UTZFMT_STYLE_GENERIC_LONG, tz, date, zoneString);
1729 capContextUsageType = DateFormatSymbols::kCapContextUsageMetazoneLong;
1730 }
1731 }
1732 else if (patternCharIndex == UDAT_TIMEZONE_SPECIAL_FIELD) {
1733 if (count == 1) {
1734 // "V"
1735 tzfmt->format(UTZFMT_STYLE_ZONE_ID_SHORT, tz, date, zoneString);
1736 } else if (count == 2) {
1737 // "VV"
1738 tzfmt->format(UTZFMT_STYLE_ZONE_ID, tz, date, zoneString);
1739 } else if (count == 3) {
1740 // "VVV"
1741 tzfmt->format(UTZFMT_STYLE_EXEMPLAR_LOCATION, tz, date, zoneString);
1742 } else if (count == 4) {
1743 // "VVVV"
1744 tzfmt->format(UTZFMT_STYLE_GENERIC_LOCATION, tz, date, zoneString);
1745 capContextUsageType = DateFormatSymbols::kCapContextUsageZoneLong;
1746 }
1747 }
1748 else if (patternCharIndex == UDAT_TIMEZONE_LOCALIZED_GMT_OFFSET_FIELD) {
1749 if (count == 1) {
1750 // "O"
1751 tzfmt->format(UTZFMT_STYLE_LOCALIZED_GMT_SHORT, tz, date, zoneString);
1752 } else if (count == 4) {
1753 // "OOOO"
1754 tzfmt->format(UTZFMT_STYLE_LOCALIZED_GMT, tz, date, zoneString);
1755 }
1756 }
1757 else if (patternCharIndex == UDAT_TIMEZONE_ISO_FIELD) {
1758 if (count == 1) {
1759 // "X"
1760 tzfmt->format(UTZFMT_STYLE_ISO_BASIC_SHORT, tz, date, zoneString);
1761 } else if (count == 2) {
1762 // "XX"
1763 tzfmt->format(UTZFMT_STYLE_ISO_BASIC_FIXED, tz, date, zoneString);
1764 } else if (count == 3) {
1765 // "XXX"
1766 tzfmt->format(UTZFMT_STYLE_ISO_EXTENDED_FIXED, tz, date, zoneString);
1767 } else if (count == 4) {
1768 // "XXXX"
1769 tzfmt->format(UTZFMT_STYLE_ISO_BASIC_FULL, tz, date, zoneString);
1770 } else if (count == 5) {
1771 // "XXXXX"
1772 tzfmt->format(UTZFMT_STYLE_ISO_EXTENDED_FULL, tz, date, zoneString);
1773 }
1774 }
1775 else if (patternCharIndex == UDAT_TIMEZONE_ISO_LOCAL_FIELD) {
1776 if (count == 1) {
1777 // "x"
1778 tzfmt->format(UTZFMT_STYLE_ISO_BASIC_LOCAL_SHORT, tz, date, zoneString);
1779 } else if (count == 2) {
1780 // "xx"
1781 tzfmt->format(UTZFMT_STYLE_ISO_BASIC_LOCAL_FIXED, tz, date, zoneString);
1782 } else if (count == 3) {
1783 // "xxx"
1784 tzfmt->format(UTZFMT_STYLE_ISO_EXTENDED_LOCAL_FIXED, tz, date, zoneString);
1785 } else if (count == 4) {
1786 // "xxxx"
1787 tzfmt->format(UTZFMT_STYLE_ISO_BASIC_LOCAL_FULL, tz, date, zoneString);
1788 } else if (count == 5) {
1789 // "xxxxx"
1790 tzfmt->format(UTZFMT_STYLE_ISO_EXTENDED_LOCAL_FULL, tz, date, zoneString);
1791 }
1792 }
1793 else {
1794 UPRV_UNREACHABLE;
1795 }
1796 }
1797 appendTo += zoneString;
1798 }
1799 break;
1800
1801 case UDAT_QUARTER_FIELD:
1802 if (count >= 4)
1803 _appendSymbol(appendTo, value/3, fSymbols->fQuarters,
1804 fSymbols->fQuartersCount);
1805 else if (count == 3)
1806 _appendSymbol(appendTo, value/3, fSymbols->fShortQuarters,
1807 fSymbols->fShortQuartersCount);
1808 else
1809 zeroPaddingNumber(currentNumberFormat,appendTo, (value/3) + 1, count, maxIntCount);
1810 break;
1811
1812 case UDAT_STANDALONE_QUARTER_FIELD:
1813 if (count >= 4)
1814 _appendSymbol(appendTo, value/3, fSymbols->fStandaloneQuarters,
1815 fSymbols->fStandaloneQuartersCount);
1816 else if (count == 3)
1817 _appendSymbol(appendTo, value/3, fSymbols->fStandaloneShortQuarters,
1818 fSymbols->fStandaloneShortQuartersCount);
1819 else
1820 zeroPaddingNumber(currentNumberFormat,appendTo, (value/3) + 1, count, maxIntCount);
1821 break;
1822
1823 case UDAT_AM_PM_MIDNIGHT_NOON_FIELD:
1824 {
1825 const UnicodeString *toAppend = NULL;
1826 int32_t hour = cal.get(UCAL_HOUR_OF_DAY, status);
1827
1828 // Note: "midnight" can be ambiguous as to whether it refers to beginning of day or end of day.
1829 // For ICU 57 output of "midnight" is temporarily suppressed.
1830
1831 // For "midnight" and "noon":
1832 // Time, as displayed, must be exactly noon or midnight.
1833 // This means minutes and seconds, if present, must be zero.
1834 if ((/*hour == 0 ||*/ hour == 12) &&
1835 (!fHasMinute || cal.get(UCAL_MINUTE, status) == 0) &&
1836 (!fHasSecond || cal.get(UCAL_SECOND, status) == 0)) {
1837 // Stealing am/pm value to use as our array index.
1838 // It works out: am/midnight are both 0, pm/noon are both 1,
1839 // 12 am is 12 midnight, and 12 pm is 12 noon.
1840 int32_t val = cal.get(UCAL_AM_PM, status);
1841
1842 if (count <= 3) {
1843 toAppend = &fSymbols->fAbbreviatedDayPeriods[val];
1844 } else if (count == 4 || count > 5) {
1845 toAppend = &fSymbols->fWideDayPeriods[val];
1846 } else { // count == 5
1847 toAppend = &fSymbols->fNarrowDayPeriods[val];
1848 }
1849 }
1850
1851 // toAppend is NULL if time isn't exactly midnight or noon (as displayed).
1852 // toAppend is bogus if time is midnight or noon, but no localized string exists.
1853 // In either case, fall back to am/pm.
1854 if (toAppend == NULL || toAppend->isBogus()) {
1855 // Reformat with identical arguments except ch, now changed to 'a'.
1856 subFormat(appendTo, 0x61, count, capitalizationContext, fieldNum,
1857 handler, cal, status);
1858 } else {
1859 appendTo += *toAppend;
1860 }
1861
1862 break;
1863 }
1864
1865 case UDAT_FLEXIBLE_DAY_PERIOD_FIELD:
1866 {
1867 // TODO: Maybe fetch the DayperiodRules during initialization (instead of at the first
1868 // loading of an instance) if a relevant pattern character (b or B) is used.
1869 const DayPeriodRules *ruleSet = DayPeriodRules::getInstance(this->getSmpFmtLocale(), status);
1870 if (U_FAILURE(status)) {
1871 // Data doesn't conform to spec, therefore loading failed.
1872 break;
1873 }
1874 if (ruleSet == NULL) {
1875 // Data doesn't exist for the locale we're looking for.
1876 // Falling back to am/pm.
1877 subFormat(appendTo, 0x61, count, capitalizationContext, fieldNum,
1878 handler, cal, status);
1879 break;
1880 }
1881
1882 // Get current display time.
1883 int32_t hour = cal.get(UCAL_HOUR_OF_DAY, status);
1884 int32_t minute = 0;
1885 if (fHasMinute) {
1886 minute = cal.get(UCAL_MINUTE, status);
1887 }
1888 int32_t second = 0;
1889 if (fHasSecond) {
1890 second = cal.get(UCAL_SECOND, status);
1891 }
1892
1893 // Determine day period.
1894 DayPeriodRules::DayPeriod periodType;
1895 if (hour == 0 && minute == 0 && second == 0 && ruleSet->hasMidnight()) {
1896 periodType = DayPeriodRules::DAYPERIOD_MIDNIGHT;
1897 } else if (hour == 12 && minute == 0 && second == 0 && ruleSet->hasNoon()) {
1898 periodType = DayPeriodRules::DAYPERIOD_NOON;
1899 } else {
1900 periodType = ruleSet->getDayPeriodForHour(hour);
1901 }
1902
1903 // Rule set exists, therefore periodType can't be UNKNOWN.
1904 // Get localized string.
1905 U_ASSERT(periodType != DayPeriodRules::DAYPERIOD_UNKNOWN);
1906 UnicodeString *toAppend = NULL;
1907 int32_t index;
1908
1909 // Note: "midnight" can be ambiguous as to whether it refers to beginning of day or end of day.
1910 // For ICU 57 output of "midnight" is temporarily suppressed.
1911
1912 if (periodType != DayPeriodRules::DAYPERIOD_AM &&
1913 periodType != DayPeriodRules::DAYPERIOD_PM &&
1914 periodType != DayPeriodRules::DAYPERIOD_MIDNIGHT) {
1915 index = (int32_t)periodType;
1916 if (count <= 3) {
1917 toAppend = &fSymbols->fAbbreviatedDayPeriods[index]; // i.e. short
1918 } else if (count == 4 || count > 5) {
1919 toAppend = &fSymbols->fWideDayPeriods[index];
1920 } else { // count == 5
1921 toAppend = &fSymbols->fNarrowDayPeriods[index];
1922 }
1923 }
1924
1925 // Fallback schedule:
1926 // Midnight/Noon -> General Periods -> AM/PM.
1927
1928 // Midnight/Noon -> General Periods.
1929 if ((toAppend == NULL || toAppend->isBogus()) &&
1930 (periodType == DayPeriodRules::DAYPERIOD_MIDNIGHT ||
1931 periodType == DayPeriodRules::DAYPERIOD_NOON)) {
1932 periodType = ruleSet->getDayPeriodForHour(hour);
1933 index = (int32_t)periodType;
1934
1935 if (count <= 3) {
1936 toAppend = &fSymbols->fAbbreviatedDayPeriods[index]; // i.e. short
1937 } else if (count == 4 || count > 5) {
1938 toAppend = &fSymbols->fWideDayPeriods[index];
1939 } else { // count == 5
1940 toAppend = &fSymbols->fNarrowDayPeriods[index];
1941 }
1942 }
1943
1944 // General Periods -> AM/PM.
1945 if (periodType == DayPeriodRules::DAYPERIOD_AM ||
1946 periodType == DayPeriodRules::DAYPERIOD_PM ||
1947 toAppend->isBogus()) {
1948 subFormat(appendTo, 0x61, count, capitalizationContext, fieldNum,
1949 handler, cal, status);
1950 }
1951 else {
1952 appendTo += *toAppend;
1953 }
1954
1955 break;
1956 }
1957
1958 // all of the other pattern symbols can be formatted as simple numbers with
1959 // appropriate zero padding
1960 default:
1961 zeroPaddingNumber(currentNumberFormat,appendTo, value, count, maxIntCount);
1962 break;
1963 }
1964 #if !UCONFIG_NO_BREAK_ITERATION
1965 // if first field, check to see whether we need to and are able to titlecase it
1966 if (fieldNum == 0 && fCapitalizationBrkIter != NULL && appendTo.length() > beginOffset &&
1967 u_islower(appendTo.char32At(beginOffset))) {
1968 UBool titlecase = FALSE;
1969 switch (capitalizationContext) {
1970 case UDISPCTX_CAPITALIZATION_FOR_BEGINNING_OF_SENTENCE:
1971 titlecase = TRUE;
1972 break;
1973 case UDISPCTX_CAPITALIZATION_FOR_UI_LIST_OR_MENU:
1974 titlecase = fSymbols->fCapitalization[capContextUsageType][0];
1975 break;
1976 case UDISPCTX_CAPITALIZATION_FOR_STANDALONE:
1977 titlecase = fSymbols->fCapitalization[capContextUsageType][1];
1978 break;
1979 default:
1980 // titlecase = FALSE;
1981 break;
1982 }
1983 if (titlecase) {
1984 BreakIterator* const mutableCapitalizationBrkIter = fCapitalizationBrkIter->clone();
1985 UnicodeString firstField(appendTo, beginOffset);
1986 firstField.toTitle(mutableCapitalizationBrkIter, fLocale, U_TITLECASE_NO_LOWERCASE | U_TITLECASE_NO_BREAK_ADJUSTMENT);
1987 appendTo.replaceBetween(beginOffset, appendTo.length(), firstField);
1988 delete mutableCapitalizationBrkIter;
1989 }
1990 }
1991 #endif
1992
1993 handler.addAttribute(fgPatternIndexToDateFormatField[patternCharIndex], beginOffset, appendTo.length());
1994 }
1995
1996 //----------------------------------------------------------------------
1997
adoptNumberFormat(NumberFormat * formatToAdopt)1998 void SimpleDateFormat::adoptNumberFormat(NumberFormat *formatToAdopt) {
1999 fixNumberFormatForDates(*formatToAdopt);
2000 delete fNumberFormat;
2001 fNumberFormat = formatToAdopt;
2002
2003 // We successfully set the default number format. Now delete the overrides
2004 // (can't fail).
2005 if (fSharedNumberFormatters) {
2006 freeSharedNumberFormatters(fSharedNumberFormatters);
2007 fSharedNumberFormatters = NULL;
2008 }
2009
2010 // Also re-compute the fast formatters.
2011 UErrorCode localStatus = U_ZERO_ERROR;
2012 freeFastNumberFormatters();
2013 initFastNumberFormatters(localStatus);
2014 }
2015
adoptNumberFormat(const UnicodeString & fields,NumberFormat * formatToAdopt,UErrorCode & status)2016 void SimpleDateFormat::adoptNumberFormat(const UnicodeString& fields, NumberFormat *formatToAdopt, UErrorCode &status){
2017 fixNumberFormatForDates(*formatToAdopt);
2018 LocalPointer<NumberFormat> fmt(formatToAdopt);
2019 if (U_FAILURE(status)) {
2020 return;
2021 }
2022
2023 // We must ensure fSharedNumberFormatters is allocated.
2024 if (fSharedNumberFormatters == NULL) {
2025 fSharedNumberFormatters = allocSharedNumberFormatters();
2026 if (fSharedNumberFormatters == NULL) {
2027 status = U_MEMORY_ALLOCATION_ERROR;
2028 return;
2029 }
2030 }
2031 const SharedNumberFormat *newFormat = createSharedNumberFormat(fmt.orphan());
2032 if (newFormat == NULL) {
2033 status = U_MEMORY_ALLOCATION_ERROR;
2034 return;
2035 }
2036 for (int i=0; i<fields.length(); i++) {
2037 UChar field = fields.charAt(i);
2038 // if the pattern character is unrecognized, signal an error and bail out
2039 UDateFormatField patternCharIndex = DateFormatSymbols::getPatternCharIndex(field);
2040 if (patternCharIndex == UDAT_FIELD_COUNT) {
2041 status = U_INVALID_FORMAT_ERROR;
2042 newFormat->deleteIfZeroRefCount();
2043 return;
2044 }
2045
2046 // Set the number formatter in the table
2047 SharedObject::copyPtr(
2048 newFormat, fSharedNumberFormatters[patternCharIndex]);
2049 }
2050 newFormat->deleteIfZeroRefCount();
2051 }
2052
2053 const NumberFormat *
getNumberFormatForField(UChar field) const2054 SimpleDateFormat::getNumberFormatForField(UChar field) const {
2055 UDateFormatField index = DateFormatSymbols::getPatternCharIndex(field);
2056 if (index == UDAT_FIELD_COUNT) {
2057 return NULL;
2058 }
2059 return getNumberFormatByIndex(index);
2060 }
2061
2062 //----------------------------------------------------------------------
2063 void
zeroPaddingNumber(const NumberFormat * currentNumberFormat,UnicodeString & appendTo,int32_t value,int32_t minDigits,int32_t maxDigits) const2064 SimpleDateFormat::zeroPaddingNumber(
2065 const NumberFormat *currentNumberFormat,
2066 UnicodeString &appendTo,
2067 int32_t value, int32_t minDigits, int32_t maxDigits) const
2068 {
2069 const number::LocalizedNumberFormatter* fastFormatter = nullptr;
2070 // NOTE: This uses the heuristic that these five min/max int settings account for the vast majority
2071 // of SimpleDateFormat number formatting cases at the time of writing (ICU 62).
2072 if (currentNumberFormat == fNumberFormat) {
2073 if (maxDigits == 10) {
2074 if (minDigits == 1) {
2075 fastFormatter = fFastNumberFormatters[SMPDTFMT_NF_1x10];
2076 } else if (minDigits == 2) {
2077 fastFormatter = fFastNumberFormatters[SMPDTFMT_NF_2x10];
2078 } else if (minDigits == 3) {
2079 fastFormatter = fFastNumberFormatters[SMPDTFMT_NF_3x10];
2080 } else if (minDigits == 4) {
2081 fastFormatter = fFastNumberFormatters[SMPDTFMT_NF_4x10];
2082 }
2083 } else if (maxDigits == 2) {
2084 if (minDigits == 2) {
2085 fastFormatter = fFastNumberFormatters[SMPDTFMT_NF_2x2];
2086 }
2087 }
2088 }
2089 if (fastFormatter != nullptr) {
2090 // Can use fast path
2091 number::impl::UFormattedNumberData result;
2092 result.quantity.setToInt(value);
2093 UErrorCode localStatus = U_ZERO_ERROR;
2094 fastFormatter->formatImpl(&result, localStatus);
2095 if (U_FAILURE(localStatus)) {
2096 return;
2097 }
2098 appendTo.append(result.getStringRef().toTempUnicodeString());
2099 return;
2100 }
2101
2102 // Check for RBNF (no clone necessary)
2103 auto* rbnf = dynamic_cast<const RuleBasedNumberFormat*>(currentNumberFormat);
2104 if (rbnf != nullptr) {
2105 FieldPosition pos(FieldPosition::DONT_CARE);
2106 rbnf->format(value, appendTo, pos); // 3rd arg is there to speed up processing
2107 return;
2108 }
2109
2110 // Fall back to slow path (clone and mutate the NumberFormat)
2111 if (currentNumberFormat != nullptr) {
2112 FieldPosition pos(FieldPosition::DONT_CARE);
2113 LocalPointer<NumberFormat> nf(currentNumberFormat->clone());
2114 nf->setMinimumIntegerDigits(minDigits);
2115 nf->setMaximumIntegerDigits(maxDigits);
2116 nf->format(value, appendTo, pos); // 3rd arg is there to speed up processing
2117 }
2118 }
2119
2120 //----------------------------------------------------------------------
2121
2122 /**
2123 * Return true if the given format character, occuring count
2124 * times, represents a numeric field.
2125 */
isNumeric(UChar formatChar,int32_t count)2126 UBool SimpleDateFormat::isNumeric(UChar formatChar, int32_t count) {
2127 return DateFormatSymbols::isNumericPatternChar(formatChar, count);
2128 }
2129
2130 UBool
isAtNumericField(const UnicodeString & pattern,int32_t patternOffset)2131 SimpleDateFormat::isAtNumericField(const UnicodeString &pattern, int32_t patternOffset) {
2132 if (patternOffset >= pattern.length()) {
2133 // not at any field
2134 return FALSE;
2135 }
2136 UChar ch = pattern.charAt(patternOffset);
2137 UDateFormatField f = DateFormatSymbols::getPatternCharIndex(ch);
2138 if (f == UDAT_FIELD_COUNT) {
2139 // not at any field
2140 return FALSE;
2141 }
2142 int32_t i = patternOffset;
2143 while (pattern.charAt(++i) == ch) {}
2144 return DateFormatSymbols::isNumericField(f, i - patternOffset);
2145 }
2146
2147 UBool
isAfterNonNumericField(const UnicodeString & pattern,int32_t patternOffset)2148 SimpleDateFormat::isAfterNonNumericField(const UnicodeString &pattern, int32_t patternOffset) {
2149 if (patternOffset <= 0) {
2150 // not after any field
2151 return FALSE;
2152 }
2153 UChar ch = pattern.charAt(--patternOffset);
2154 UDateFormatField f = DateFormatSymbols::getPatternCharIndex(ch);
2155 if (f == UDAT_FIELD_COUNT) {
2156 // not after any field
2157 return FALSE;
2158 }
2159 int32_t i = patternOffset;
2160 while (pattern.charAt(--i) == ch) {}
2161 return !DateFormatSymbols::isNumericField(f, patternOffset - i);
2162 }
2163
2164 void
parse(const UnicodeString & text,Calendar & cal,ParsePosition & parsePos) const2165 SimpleDateFormat::parse(const UnicodeString& text, Calendar& cal, ParsePosition& parsePos) const
2166 {
2167 UErrorCode status = U_ZERO_ERROR;
2168 int32_t pos = parsePos.getIndex();
2169 if(parsePos.getIndex() < 0) {
2170 parsePos.setErrorIndex(0);
2171 return;
2172 }
2173 int32_t start = pos;
2174
2175 // Hold the day period until everything else is parsed, because we need
2176 // the hour to interpret time correctly.
2177 int32_t dayPeriodInt = -1;
2178
2179 UBool ambiguousYear[] = { FALSE };
2180 int32_t saveHebrewMonth = -1;
2181 int32_t count = 0;
2182 UTimeZoneFormatTimeType tzTimeType = UTZFMT_TIME_TYPE_UNKNOWN;
2183
2184 // For parsing abutting numeric fields. 'abutPat' is the
2185 // offset into 'pattern' of the first of 2 or more abutting
2186 // numeric fields. 'abutStart' is the offset into 'text'
2187 // where parsing the fields begins. 'abutPass' starts off as 0
2188 // and increments each time we try to parse the fields.
2189 int32_t abutPat = -1; // If >=0, we are in a run of abutting numeric fields
2190 int32_t abutStart = 0;
2191 int32_t abutPass = 0;
2192 UBool inQuote = FALSE;
2193
2194 MessageFormat * numericLeapMonthFormatter = NULL;
2195
2196 Calendar* calClone = NULL;
2197 Calendar *workCal = &cal;
2198 if (&cal != fCalendar && uprv_strcmp(cal.getType(), fCalendar->getType()) != 0) {
2199 // Different calendar type
2200 // We use the time/zone from the input calendar, but
2201 // do not use the input calendar for field calculation.
2202 calClone = fCalendar->clone();
2203 if (calClone != NULL) {
2204 calClone->setTime(cal.getTime(status),status);
2205 if (U_FAILURE(status)) {
2206 goto ExitParse;
2207 }
2208 calClone->setTimeZone(cal.getTimeZone());
2209 workCal = calClone;
2210 } else {
2211 status = U_MEMORY_ALLOCATION_ERROR;
2212 goto ExitParse;
2213 }
2214 }
2215
2216 if (fSymbols->fLeapMonthPatterns != NULL && fSymbols->fLeapMonthPatternsCount >= DateFormatSymbols::kMonthPatternsCount) {
2217 numericLeapMonthFormatter = new MessageFormat(fSymbols->fLeapMonthPatterns[DateFormatSymbols::kLeapMonthPatternNumeric], fLocale, status);
2218 if (numericLeapMonthFormatter == NULL) {
2219 status = U_MEMORY_ALLOCATION_ERROR;
2220 goto ExitParse;
2221 } else if (U_FAILURE(status)) {
2222 goto ExitParse; // this will delete numericLeapMonthFormatter
2223 }
2224 }
2225
2226 for (int32_t i=0; i<fPattern.length(); ++i) {
2227 UChar ch = fPattern.charAt(i);
2228
2229 // Handle alphabetic field characters.
2230 if (!inQuote && isSyntaxChar(ch)) {
2231 int32_t fieldPat = i;
2232
2233 // Count the length of this field specifier
2234 count = 1;
2235 while ((i+1)<fPattern.length() &&
2236 fPattern.charAt(i+1) == ch) {
2237 ++count;
2238 ++i;
2239 }
2240
2241 if (isNumeric(ch, count)) {
2242 if (abutPat < 0) {
2243 // Determine if there is an abutting numeric field.
2244 // Record the start of a set of abutting numeric fields.
2245 if (isAtNumericField(fPattern, i + 1)) {
2246 abutPat = fieldPat;
2247 abutStart = pos;
2248 abutPass = 0;
2249 }
2250 }
2251 } else {
2252 abutPat = -1; // End of any abutting fields
2253 }
2254
2255 // Handle fields within a run of abutting numeric fields. Take
2256 // the pattern "HHmmss" as an example. We will try to parse
2257 // 2/2/2 characters of the input text, then if that fails,
2258 // 1/2/2. We only adjust the width of the leftmost field; the
2259 // others remain fixed. This allows "123456" => 12:34:56, but
2260 // "12345" => 1:23:45. Likewise, for the pattern "yyyyMMdd" we
2261 // try 4/2/2, 3/2/2, 2/2/2, and finally 1/2/2.
2262 if (abutPat >= 0) {
2263 // If we are at the start of a run of abutting fields, then
2264 // shorten this field in each pass. If we can't shorten
2265 // this field any more, then the parse of this set of
2266 // abutting numeric fields has failed.
2267 if (fieldPat == abutPat) {
2268 count -= abutPass++;
2269 if (count == 0) {
2270 status = U_PARSE_ERROR;
2271 goto ExitParse;
2272 }
2273 }
2274
2275 pos = subParse(text, pos, ch, count,
2276 TRUE, FALSE, ambiguousYear, saveHebrewMonth, *workCal, i, numericLeapMonthFormatter, &tzTimeType);
2277
2278 // If the parse fails anywhere in the run, back up to the
2279 // start of the run and retry.
2280 if (pos < 0) {
2281 i = abutPat - 1;
2282 pos = abutStart;
2283 continue;
2284 }
2285 }
2286
2287 // Handle non-numeric fields and non-abutting numeric
2288 // fields.
2289 else if (ch != 0x6C) { // pattern char 'l' (SMALL LETTER L) just gets ignored
2290 int32_t s = subParse(text, pos, ch, count,
2291 FALSE, TRUE, ambiguousYear, saveHebrewMonth, *workCal, i, numericLeapMonthFormatter, &tzTimeType, &dayPeriodInt);
2292
2293 if (s == -pos-1) {
2294 // era not present, in special cases allow this to continue
2295 // from the position where the era was expected
2296 s = pos;
2297
2298 if (i+1 < fPattern.length()) {
2299 // move to next pattern character
2300 UChar c = fPattern.charAt(i+1);
2301
2302 // check for whitespace
2303 if (PatternProps::isWhiteSpace(c)) {
2304 i++;
2305 // Advance over run in pattern
2306 while ((i+1)<fPattern.length() &&
2307 PatternProps::isWhiteSpace(fPattern.charAt(i+1))) {
2308 ++i;
2309 }
2310 }
2311 }
2312 }
2313 else if (s <= 0) {
2314 status = U_PARSE_ERROR;
2315 goto ExitParse;
2316 }
2317 pos = s;
2318 }
2319 }
2320
2321 // Handle literal pattern characters. These are any
2322 // quoted characters and non-alphabetic unquoted
2323 // characters.
2324 else {
2325
2326 abutPat = -1; // End of any abutting fields
2327
2328 if (! matchLiterals(fPattern, i, text, pos, getBooleanAttribute(UDAT_PARSE_ALLOW_WHITESPACE, status), getBooleanAttribute(UDAT_PARSE_PARTIAL_LITERAL_MATCH, status), isLenient())) {
2329 status = U_PARSE_ERROR;
2330 goto ExitParse;
2331 }
2332 }
2333 }
2334
2335 // Special hack for trailing "." after non-numeric field.
2336 if (text.charAt(pos) == 0x2e && getBooleanAttribute(UDAT_PARSE_ALLOW_WHITESPACE, status)) {
2337 // only do if the last field is not numeric
2338 if (isAfterNonNumericField(fPattern, fPattern.length())) {
2339 pos++; // skip the extra "."
2340 }
2341 }
2342
2343 // If dayPeriod is set, use it in conjunction with hour-of-day to determine am/pm.
2344 if (dayPeriodInt >= 0) {
2345 DayPeriodRules::DayPeriod dayPeriod = (DayPeriodRules::DayPeriod)dayPeriodInt;
2346 const DayPeriodRules *ruleSet = DayPeriodRules::getInstance(this->getSmpFmtLocale(), status);
2347
2348 if (!cal.isSet(UCAL_HOUR) && !cal.isSet(UCAL_HOUR_OF_DAY)) {
2349 // If hour is not set, set time to the midpoint of current day period, overwriting
2350 // minutes if it's set.
2351 double midPoint = ruleSet->getMidPointForDayPeriod(dayPeriod, status);
2352
2353 // If we can't get midPoint we do nothing.
2354 if (U_SUCCESS(status)) {
2355 // Truncate midPoint toward zero to get the hour.
2356 // Any leftover means it was a half-hour.
2357 int32_t midPointHour = (int32_t) midPoint;
2358 int32_t midPointMinute = (midPoint - midPointHour) > 0 ? 30 : 0;
2359
2360 // No need to set am/pm because hour-of-day is set last therefore takes precedence.
2361 cal.set(UCAL_HOUR_OF_DAY, midPointHour);
2362 cal.set(UCAL_MINUTE, midPointMinute);
2363 }
2364 } else {
2365 int hourOfDay;
2366
2367 if (cal.isSet(UCAL_HOUR_OF_DAY)) { // Hour is parsed in 24-hour format.
2368 hourOfDay = cal.get(UCAL_HOUR_OF_DAY, status);
2369 } else { // Hour is parsed in 12-hour format.
2370 hourOfDay = cal.get(UCAL_HOUR, status);
2371 // cal.get() turns 12 to 0 for 12-hour time; change 0 to 12
2372 // so 0 unambiguously means a 24-hour time from above.
2373 if (hourOfDay == 0) { hourOfDay = 12; }
2374 }
2375 U_ASSERT(0 <= hourOfDay && hourOfDay <= 23);
2376
2377
2378 // If hour-of-day is 0 or 13 thru 23 then input time in unambiguously in 24-hour format.
2379 if (hourOfDay == 0 || (13 <= hourOfDay && hourOfDay <= 23)) {
2380 // Make hour-of-day take precedence over (hour + am/pm) by setting it again.
2381 cal.set(UCAL_HOUR_OF_DAY, hourOfDay);
2382 } else {
2383 // We have a 12-hour time and need to choose between am and pm.
2384 // Behave as if dayPeriod spanned 6 hours each way from its center point.
2385 // This will parse correctly for consistent time + period (e.g. 10 at night) as
2386 // well as provide a reasonable recovery for inconsistent time + period (e.g.
2387 // 9 in the afternoon).
2388
2389 // Assume current time is in the AM.
2390 // - Change 12 back to 0 for easier handling of 12am.
2391 // - Append minutes as fractional hours because e.g. 8:15 and 8:45 could be parsed
2392 // into different half-days if center of dayPeriod is at 14:30.
2393 // - cal.get(MINUTE) will return 0 if MINUTE is unset, which works.
2394 if (hourOfDay == 12) { hourOfDay = 0; }
2395 double currentHour = hourOfDay + (cal.get(UCAL_MINUTE, status)) / 60.0;
2396 double midPointHour = ruleSet->getMidPointForDayPeriod(dayPeriod, status);
2397
2398 if (U_SUCCESS(status)) {
2399 double hoursAheadMidPoint = currentHour - midPointHour;
2400
2401 // Assume current time is in the AM.
2402 if (-6 <= hoursAheadMidPoint && hoursAheadMidPoint < 6) {
2403 // Assumption holds; set time as such.
2404 cal.set(UCAL_AM_PM, 0);
2405 } else {
2406 cal.set(UCAL_AM_PM, 1);
2407 }
2408 }
2409 }
2410 }
2411 }
2412
2413 // At this point the fields of Calendar have been set. Calendar
2414 // will fill in default values for missing fields when the time
2415 // is computed.
2416
2417 parsePos.setIndex(pos);
2418
2419 // This part is a problem: When we call parsedDate.after, we compute the time.
2420 // Take the date April 3 2004 at 2:30 am. When this is first set up, the year
2421 // will be wrong if we're parsing a 2-digit year pattern. It will be 1904.
2422 // April 3 1904 is a Sunday (unlike 2004) so it is the DST onset day. 2:30 am
2423 // is therefore an "impossible" time, since the time goes from 1:59 to 3:00 am
2424 // on that day. It is therefore parsed out to fields as 3:30 am. Then we
2425 // add 100 years, and get April 3 2004 at 3:30 am. Note that April 3 2004 is
2426 // a Saturday, so it can have a 2:30 am -- and it should. [LIU]
2427 /*
2428 UDate parsedDate = calendar.getTime();
2429 if( ambiguousYear[0] && !parsedDate.after(fDefaultCenturyStart) ) {
2430 calendar.add(Calendar.YEAR, 100);
2431 parsedDate = calendar.getTime();
2432 }
2433 */
2434 // Because of the above condition, save off the fields in case we need to readjust.
2435 // The procedure we use here is not particularly efficient, but there is no other
2436 // way to do this given the API restrictions present in Calendar. We minimize
2437 // inefficiency by only performing this computation when it might apply, that is,
2438 // when the two-digit year is equal to the start year, and thus might fall at the
2439 // front or the back of the default century. This only works because we adjust
2440 // the year correctly to start with in other cases -- see subParse().
2441 if (ambiguousYear[0] || tzTimeType != UTZFMT_TIME_TYPE_UNKNOWN) // If this is true then the two-digit year == the default start year
2442 {
2443 // We need a copy of the fields, and we need to avoid triggering a call to
2444 // complete(), which will recalculate the fields. Since we can't access
2445 // the fields[] array in Calendar, we clone the entire object. This will
2446 // stop working if Calendar.clone() is ever rewritten to call complete().
2447 Calendar *copy;
2448 if (ambiguousYear[0]) {
2449 copy = cal.clone();
2450 // Check for failed cloning.
2451 if (copy == NULL) {
2452 status = U_MEMORY_ALLOCATION_ERROR;
2453 goto ExitParse;
2454 }
2455 UDate parsedDate = copy->getTime(status);
2456 // {sfb} check internalGetDefaultCenturyStart
2457 if (fHaveDefaultCentury && (parsedDate < fDefaultCenturyStart)) {
2458 // We can't use add here because that does a complete() first.
2459 cal.set(UCAL_YEAR, fDefaultCenturyStartYear + 100);
2460 }
2461 delete copy;
2462 }
2463
2464 if (tzTimeType != UTZFMT_TIME_TYPE_UNKNOWN) {
2465 copy = cal.clone();
2466 // Check for failed cloning.
2467 if (copy == NULL) {
2468 status = U_MEMORY_ALLOCATION_ERROR;
2469 goto ExitParse;
2470 }
2471 const TimeZone & tz = cal.getTimeZone();
2472 BasicTimeZone *btz = NULL;
2473
2474 if (dynamic_cast<const OlsonTimeZone *>(&tz) != NULL
2475 || dynamic_cast<const SimpleTimeZone *>(&tz) != NULL
2476 || dynamic_cast<const RuleBasedTimeZone *>(&tz) != NULL
2477 || dynamic_cast<const VTimeZone *>(&tz) != NULL) {
2478 btz = (BasicTimeZone*)&tz;
2479 }
2480
2481 // Get local millis
2482 copy->set(UCAL_ZONE_OFFSET, 0);
2483 copy->set(UCAL_DST_OFFSET, 0);
2484 UDate localMillis = copy->getTime(status);
2485
2486 // Make sure parsed time zone type (Standard or Daylight)
2487 // matches the rule used by the parsed time zone.
2488 int32_t raw, dst;
2489 if (btz != NULL) {
2490 if (tzTimeType == UTZFMT_TIME_TYPE_STANDARD) {
2491 btz->getOffsetFromLocal(localMillis,
2492 BasicTimeZone::kStandard, BasicTimeZone::kStandard, raw, dst, status);
2493 } else {
2494 btz->getOffsetFromLocal(localMillis,
2495 BasicTimeZone::kDaylight, BasicTimeZone::kDaylight, raw, dst, status);
2496 }
2497 } else {
2498 // No good way to resolve ambiguous time at transition,
2499 // but following code work in most case.
2500 tz.getOffset(localMillis, TRUE, raw, dst, status);
2501 }
2502
2503 // Now, compare the results with parsed type, either standard or daylight saving time
2504 int32_t resolvedSavings = dst;
2505 if (tzTimeType == UTZFMT_TIME_TYPE_STANDARD) {
2506 if (dst != 0) {
2507 // Override DST_OFFSET = 0 in the result calendar
2508 resolvedSavings = 0;
2509 }
2510 } else { // tztype == TZTYPE_DST
2511 if (dst == 0) {
2512 if (btz != NULL) {
2513 UDate time = localMillis + raw;
2514 // We use the nearest daylight saving time rule.
2515 TimeZoneTransition beforeTrs, afterTrs;
2516 UDate beforeT = time, afterT = time;
2517 int32_t beforeSav = 0, afterSav = 0;
2518 UBool beforeTrsAvail, afterTrsAvail;
2519
2520 // Search for DST rule before or on the time
2521 while (TRUE) {
2522 beforeTrsAvail = btz->getPreviousTransition(beforeT, TRUE, beforeTrs);
2523 if (!beforeTrsAvail) {
2524 break;
2525 }
2526 beforeT = beforeTrs.getTime() - 1;
2527 beforeSav = beforeTrs.getFrom()->getDSTSavings();
2528 if (beforeSav != 0) {
2529 break;
2530 }
2531 }
2532
2533 // Search for DST rule after the time
2534 while (TRUE) {
2535 afterTrsAvail = btz->getNextTransition(afterT, FALSE, afterTrs);
2536 if (!afterTrsAvail) {
2537 break;
2538 }
2539 afterT = afterTrs.getTime();
2540 afterSav = afterTrs.getTo()->getDSTSavings();
2541 if (afterSav != 0) {
2542 break;
2543 }
2544 }
2545
2546 if (beforeTrsAvail && afterTrsAvail) {
2547 if (time - beforeT > afterT - time) {
2548 resolvedSavings = afterSav;
2549 } else {
2550 resolvedSavings = beforeSav;
2551 }
2552 } else if (beforeTrsAvail && beforeSav != 0) {
2553 resolvedSavings = beforeSav;
2554 } else if (afterTrsAvail && afterSav != 0) {
2555 resolvedSavings = afterSav;
2556 } else {
2557 resolvedSavings = btz->getDSTSavings();
2558 }
2559 } else {
2560 resolvedSavings = tz.getDSTSavings();
2561 }
2562 if (resolvedSavings == 0) {
2563 // final fallback
2564 resolvedSavings = U_MILLIS_PER_HOUR;
2565 }
2566 }
2567 }
2568 cal.set(UCAL_ZONE_OFFSET, raw);
2569 cal.set(UCAL_DST_OFFSET, resolvedSavings);
2570 delete copy;
2571 }
2572 }
2573 ExitParse:
2574 // Set the parsed result if local calendar is used
2575 // instead of the input calendar
2576 if (U_SUCCESS(status) && workCal != &cal) {
2577 cal.setTimeZone(workCal->getTimeZone());
2578 cal.setTime(workCal->getTime(status), status);
2579 }
2580
2581 if (numericLeapMonthFormatter != NULL) {
2582 delete numericLeapMonthFormatter;
2583 }
2584 if (calClone != NULL) {
2585 delete calClone;
2586 }
2587
2588 // If any Calendar calls failed, we pretend that we
2589 // couldn't parse the string, when in reality this isn't quite accurate--
2590 // we did parse it; the Calendar calls just failed.
2591 if (U_FAILURE(status)) {
2592 parsePos.setErrorIndex(pos);
2593 parsePos.setIndex(start);
2594 }
2595 }
2596
2597 //----------------------------------------------------------------------
2598
2599 static int32_t
2600 matchStringWithOptionalDot(const UnicodeString &text,
2601 int32_t index,
2602 const UnicodeString &data);
2603
matchQuarterString(const UnicodeString & text,int32_t start,UCalendarDateFields field,const UnicodeString * data,int32_t dataCount,Calendar & cal) const2604 int32_t SimpleDateFormat::matchQuarterString(const UnicodeString& text,
2605 int32_t start,
2606 UCalendarDateFields field,
2607 const UnicodeString* data,
2608 int32_t dataCount,
2609 Calendar& cal) const
2610 {
2611 int32_t i = 0;
2612 int32_t count = dataCount;
2613
2614 // There may be multiple strings in the data[] array which begin with
2615 // the same prefix (e.g., Cerven and Cervenec (June and July) in Czech).
2616 // We keep track of the longest match, and return that. Note that this
2617 // unfortunately requires us to test all array elements.
2618 int32_t bestMatchLength = 0, bestMatch = -1;
2619 UnicodeString bestMatchName;
2620
2621 for (; i < count; ++i) {
2622 int32_t matchLength = 0;
2623 if ((matchLength = matchStringWithOptionalDot(text, start, data[i])) > bestMatchLength) {
2624 bestMatchLength = matchLength;
2625 bestMatch = i;
2626 }
2627 }
2628
2629 if (bestMatch >= 0) {
2630 cal.set(field, bestMatch * 3);
2631 return start + bestMatchLength;
2632 }
2633
2634 return -start;
2635 }
2636
matchDayPeriodStrings(const UnicodeString & text,int32_t start,const UnicodeString * data,int32_t dataCount,int32_t & dayPeriod) const2637 int32_t SimpleDateFormat::matchDayPeriodStrings(const UnicodeString& text, int32_t start,
2638 const UnicodeString* data, int32_t dataCount,
2639 int32_t &dayPeriod) const
2640 {
2641
2642 int32_t bestMatchLength = 0, bestMatch = -1;
2643
2644 for (int32_t i = 0; i < dataCount; ++i) {
2645 int32_t matchLength = 0;
2646 if ((matchLength = matchStringWithOptionalDot(text, start, data[i])) > bestMatchLength) {
2647 bestMatchLength = matchLength;
2648 bestMatch = i;
2649 }
2650 }
2651
2652 if (bestMatch >= 0) {
2653 dayPeriod = bestMatch;
2654 return start + bestMatchLength;
2655 }
2656
2657 return -start;
2658 }
2659
2660 //----------------------------------------------------------------------
matchLiterals(const UnicodeString & pattern,int32_t & patternOffset,const UnicodeString & text,int32_t & textOffset,UBool whitespaceLenient,UBool partialMatchLenient,UBool oldLeniency)2661 UBool SimpleDateFormat::matchLiterals(const UnicodeString &pattern,
2662 int32_t &patternOffset,
2663 const UnicodeString &text,
2664 int32_t &textOffset,
2665 UBool whitespaceLenient,
2666 UBool partialMatchLenient,
2667 UBool oldLeniency)
2668 {
2669 UBool inQuote = FALSE;
2670 UnicodeString literal;
2671 int32_t i = patternOffset;
2672
2673 // scan pattern looking for contiguous literal characters
2674 for ( ; i < pattern.length(); i += 1) {
2675 UChar ch = pattern.charAt(i);
2676
2677 if (!inQuote && isSyntaxChar(ch)) {
2678 break;
2679 }
2680
2681 if (ch == QUOTE) {
2682 // Match a quote literal ('') inside OR outside of quotes
2683 if ((i + 1) < pattern.length() && pattern.charAt(i + 1) == QUOTE) {
2684 i += 1;
2685 } else {
2686 inQuote = !inQuote;
2687 continue;
2688 }
2689 }
2690
2691 literal += ch;
2692 }
2693
2694 // at this point, literal contains the literal text
2695 // and i is the index of the next non-literal pattern character.
2696 int32_t p;
2697 int32_t t = textOffset;
2698
2699 if (whitespaceLenient) {
2700 // trim leading, trailing whitespace from
2701 // the literal text
2702 literal.trim();
2703
2704 // ignore any leading whitespace in the text
2705 while (t < text.length() && u_isWhitespace(text.charAt(t))) {
2706 t += 1;
2707 }
2708 }
2709
2710 for (p = 0; p < literal.length() && t < text.length();) {
2711 UBool needWhitespace = FALSE;
2712
2713 while (p < literal.length() && PatternProps::isWhiteSpace(literal.charAt(p))) {
2714 needWhitespace = TRUE;
2715 p += 1;
2716 }
2717
2718 if (needWhitespace) {
2719 int32_t tStart = t;
2720
2721 while (t < text.length()) {
2722 UChar tch = text.charAt(t);
2723
2724 if (!u_isUWhiteSpace(tch) && !PatternProps::isWhiteSpace(tch)) {
2725 break;
2726 }
2727
2728 t += 1;
2729 }
2730
2731 // TODO: should we require internal spaces
2732 // in lenient mode? (There won't be any
2733 // leading or trailing spaces)
2734 if (!whitespaceLenient && t == tStart) {
2735 // didn't find matching whitespace:
2736 // an error in strict mode
2737 return FALSE;
2738 }
2739
2740 // In strict mode, this run of whitespace
2741 // may have been at the end.
2742 if (p >= literal.length()) {
2743 break;
2744 }
2745 }
2746 if (t >= text.length() || literal.charAt(p) != text.charAt(t)) {
2747 // Ran out of text, or found a non-matching character:
2748 // OK in lenient mode, an error in strict mode.
2749 if (whitespaceLenient) {
2750 if (t == textOffset && text.charAt(t) == 0x2e &&
2751 isAfterNonNumericField(pattern, patternOffset)) {
2752 // Lenient mode and the literal input text begins with a "." and
2753 // we are after a non-numeric field: We skip the "."
2754 ++t;
2755 continue; // Do not update p.
2756 }
2757 // if it is actual whitespace and we're whitespace lenient it's OK
2758
2759 UChar wsc = text.charAt(t);
2760 if(PatternProps::isWhiteSpace(wsc)) {
2761 // Lenient mode and it's just whitespace we skip it
2762 ++t;
2763 continue; // Do not update p.
2764 }
2765 }
2766 // hack around oldleniency being a bit of a catch-all bucket and we're just adding support specifically for paritial matches
2767 if(partialMatchLenient && oldLeniency) {
2768 break;
2769 }
2770
2771 return FALSE;
2772 }
2773 ++p;
2774 ++t;
2775 }
2776
2777 // At this point if we're in strict mode we have a complete match.
2778 // If we're in lenient mode we may have a partial match, or no
2779 // match at all.
2780 if (p <= 0) {
2781 // no match. Pretend it matched a run of whitespace
2782 // and ignorables in the text.
2783 const UnicodeSet *ignorables = NULL;
2784 UDateFormatField patternCharIndex = DateFormatSymbols::getPatternCharIndex(pattern.charAt(i));
2785 if (patternCharIndex != UDAT_FIELD_COUNT) {
2786 ignorables = SimpleDateFormatStaticSets::getIgnorables(patternCharIndex);
2787 }
2788
2789 for (t = textOffset; t < text.length(); t += 1) {
2790 UChar ch = text.charAt(t);
2791
2792 if (ignorables == NULL || !ignorables->contains(ch)) {
2793 break;
2794 }
2795 }
2796 }
2797
2798 // if we get here, we've got a complete match.
2799 patternOffset = i - 1;
2800 textOffset = t;
2801
2802 return TRUE;
2803 }
2804
2805 //----------------------------------------------------------------------
2806
matchString(const UnicodeString & text,int32_t start,UCalendarDateFields field,const UnicodeString * data,int32_t dataCount,const UnicodeString * monthPattern,Calendar & cal) const2807 int32_t SimpleDateFormat::matchString(const UnicodeString& text,
2808 int32_t start,
2809 UCalendarDateFields field,
2810 const UnicodeString* data,
2811 int32_t dataCount,
2812 const UnicodeString* monthPattern,
2813 Calendar& cal) const
2814 {
2815 int32_t i = 0;
2816 int32_t count = dataCount;
2817
2818 if (field == UCAL_DAY_OF_WEEK) i = 1;
2819
2820 // There may be multiple strings in the data[] array which begin with
2821 // the same prefix (e.g., Cerven and Cervenec (June and July) in Czech).
2822 // We keep track of the longest match, and return that. Note that this
2823 // unfortunately requires us to test all array elements.
2824 int32_t bestMatchLength = 0, bestMatch = -1;
2825 UnicodeString bestMatchName;
2826 int32_t isLeapMonth = 0;
2827
2828 for (; i < count; ++i) {
2829 int32_t matchLen = 0;
2830 if ((matchLen = matchStringWithOptionalDot(text, start, data[i])) > bestMatchLength) {
2831 bestMatch = i;
2832 bestMatchLength = matchLen;
2833 }
2834
2835 if (monthPattern != NULL) {
2836 UErrorCode status = U_ZERO_ERROR;
2837 UnicodeString leapMonthName;
2838 SimpleFormatter(*monthPattern, 1, 1, status).format(data[i], leapMonthName, status);
2839 if (U_SUCCESS(status)) {
2840 if ((matchLen = matchStringWithOptionalDot(text, start, leapMonthName)) > bestMatchLength) {
2841 bestMatch = i;
2842 bestMatchLength = matchLen;
2843 isLeapMonth = 1;
2844 }
2845 }
2846 }
2847 }
2848
2849 if (bestMatch >= 0) {
2850 if (field < UCAL_FIELD_COUNT) {
2851 // Adjustment for Hebrew Calendar month Adar II
2852 if (!strcmp(cal.getType(),"hebrew") && field==UCAL_MONTH && bestMatch==13) {
2853 cal.set(field,6);
2854 } else {
2855 if (field == UCAL_YEAR) {
2856 bestMatch++; // only get here for cyclic year names, which match 1-based years 1-60
2857 }
2858 cal.set(field, bestMatch);
2859 }
2860 if (monthPattern != NULL) {
2861 cal.set(UCAL_IS_LEAP_MONTH, isLeapMonth);
2862 }
2863 }
2864
2865 return start + bestMatchLength;
2866 }
2867
2868 return -start;
2869 }
2870
2871 static int32_t
matchStringWithOptionalDot(const UnicodeString & text,int32_t index,const UnicodeString & data)2872 matchStringWithOptionalDot(const UnicodeString &text,
2873 int32_t index,
2874 const UnicodeString &data) {
2875 UErrorCode sts = U_ZERO_ERROR;
2876 int32_t matchLenText = 0;
2877 int32_t matchLenData = 0;
2878
2879 u_caseInsensitivePrefixMatch(text.getBuffer() + index, text.length() - index,
2880 data.getBuffer(), data.length(),
2881 0 /* default case option */,
2882 &matchLenText, &matchLenData,
2883 &sts);
2884 U_ASSERT (U_SUCCESS(sts));
2885
2886 if (matchLenData == data.length() /* normal match */
2887 || (data.charAt(data.length() - 1) == 0x2e
2888 && matchLenData == data.length() - 1 /* match without trailing dot */)) {
2889 return matchLenText;
2890 }
2891
2892 return 0;
2893 }
2894
2895 //----------------------------------------------------------------------
2896
2897 void
set2DigitYearStart(UDate d,UErrorCode & status)2898 SimpleDateFormat::set2DigitYearStart(UDate d, UErrorCode& status)
2899 {
2900 parseAmbiguousDatesAsAfter(d, status);
2901 }
2902
2903 /**
2904 * Private member function that converts the parsed date strings into
2905 * timeFields. Returns -start (for ParsePosition) if failed.
2906 */
subParse(const UnicodeString & text,int32_t & start,UChar ch,int32_t count,UBool obeyCount,UBool allowNegative,UBool ambiguousYear[],int32_t & saveHebrewMonth,Calendar & cal,int32_t patLoc,MessageFormat * numericLeapMonthFormatter,UTimeZoneFormatTimeType * tzTimeType,int32_t * dayPeriod) const2907 int32_t SimpleDateFormat::subParse(const UnicodeString& text, int32_t& start, UChar ch, int32_t count,
2908 UBool obeyCount, UBool allowNegative, UBool ambiguousYear[], int32_t& saveHebrewMonth, Calendar& cal,
2909 int32_t patLoc, MessageFormat * numericLeapMonthFormatter, UTimeZoneFormatTimeType *tzTimeType,
2910 int32_t *dayPeriod) const
2911 {
2912 Formattable number;
2913 int32_t value = 0;
2914 int32_t i;
2915 int32_t ps = 0;
2916 UErrorCode status = U_ZERO_ERROR;
2917 ParsePosition pos(0);
2918 UDateFormatField patternCharIndex = DateFormatSymbols::getPatternCharIndex(ch);
2919 const NumberFormat *currentNumberFormat;
2920 UnicodeString temp;
2921 UBool gotNumber = FALSE;
2922
2923 #if defined (U_DEBUG_CAL)
2924 //fprintf(stderr, "%s:%d - [%c] st=%d \n", __FILE__, __LINE__, (char) ch, start);
2925 #endif
2926
2927 if (patternCharIndex == UDAT_FIELD_COUNT) {
2928 return -start;
2929 }
2930
2931 currentNumberFormat = getNumberFormatByIndex(patternCharIndex);
2932 if (currentNumberFormat == NULL) {
2933 return -start;
2934 }
2935 UCalendarDateFields field = fgPatternIndexToCalendarField[patternCharIndex]; // UCAL_FIELD_COUNT if irrelevant
2936 UnicodeString hebr("hebr", 4, US_INV);
2937
2938 if (numericLeapMonthFormatter != NULL) {
2939 numericLeapMonthFormatter->setFormats((const Format **)¤tNumberFormat, 1);
2940 }
2941 UBool isChineseCalendar = (uprv_strcmp(cal.getType(),"chinese") == 0 || uprv_strcmp(cal.getType(),"dangi") == 0);
2942
2943 // If there are any spaces here, skip over them. If we hit the end
2944 // of the string, then fail.
2945 for (;;) {
2946 if (start >= text.length()) {
2947 return -start;
2948 }
2949 UChar32 c = text.char32At(start);
2950 if (!u_isUWhiteSpace(c) /*||*/ && !PatternProps::isWhiteSpace(c)) {
2951 break;
2952 }
2953 start += U16_LENGTH(c);
2954 }
2955 pos.setIndex(start);
2956
2957 // We handle a few special cases here where we need to parse
2958 // a number value. We handle further, more generic cases below. We need
2959 // to handle some of them here because some fields require extra processing on
2960 // the parsed value.
2961 if (patternCharIndex == UDAT_HOUR_OF_DAY1_FIELD || // k
2962 patternCharIndex == UDAT_HOUR_OF_DAY0_FIELD || // H
2963 patternCharIndex == UDAT_HOUR1_FIELD || // h
2964 patternCharIndex == UDAT_HOUR0_FIELD || // K
2965 (patternCharIndex == UDAT_DOW_LOCAL_FIELD && count <= 2) || // e
2966 (patternCharIndex == UDAT_STANDALONE_DAY_FIELD && count <= 2) || // c
2967 (patternCharIndex == UDAT_MONTH_FIELD && count <= 2) || // M
2968 (patternCharIndex == UDAT_STANDALONE_MONTH_FIELD && count <= 2) || // L
2969 (patternCharIndex == UDAT_QUARTER_FIELD && count <= 2) || // Q
2970 (patternCharIndex == UDAT_STANDALONE_QUARTER_FIELD && count <= 2) || // q
2971 patternCharIndex == UDAT_YEAR_FIELD || // y
2972 patternCharIndex == UDAT_YEAR_WOY_FIELD || // Y
2973 patternCharIndex == UDAT_YEAR_NAME_FIELD || // U (falls back to numeric)
2974 (patternCharIndex == UDAT_ERA_FIELD && isChineseCalendar) || // G
2975 patternCharIndex == UDAT_FRACTIONAL_SECOND_FIELD) // S
2976 {
2977 int32_t parseStart = pos.getIndex();
2978 // It would be good to unify this with the obeyCount logic below,
2979 // but that's going to be difficult.
2980 const UnicodeString* src;
2981
2982 UBool parsedNumericLeapMonth = FALSE;
2983 if (numericLeapMonthFormatter != NULL && (patternCharIndex == UDAT_MONTH_FIELD || patternCharIndex == UDAT_STANDALONE_MONTH_FIELD)) {
2984 int32_t argCount;
2985 Formattable * args = numericLeapMonthFormatter->parse(text, pos, argCount);
2986 if (args != NULL && argCount == 1 && pos.getIndex() > parseStart && args[0].isNumeric()) {
2987 parsedNumericLeapMonth = TRUE;
2988 number.setLong(args[0].getLong());
2989 cal.set(UCAL_IS_LEAP_MONTH, 1);
2990 delete[] args;
2991 } else {
2992 pos.setIndex(parseStart);
2993 cal.set(UCAL_IS_LEAP_MONTH, 0);
2994 }
2995 }
2996
2997 if (!parsedNumericLeapMonth) {
2998 if (obeyCount) {
2999 if ((start+count) > text.length()) {
3000 return -start;
3001 }
3002
3003 text.extractBetween(0, start + count, temp);
3004 src = &temp;
3005 } else {
3006 src = &text;
3007 }
3008
3009 parseInt(*src, number, pos, allowNegative,currentNumberFormat);
3010 }
3011
3012 int32_t txtLoc = pos.getIndex();
3013
3014 if (txtLoc > parseStart) {
3015 value = number.getLong();
3016 gotNumber = TRUE;
3017
3018 // suffix processing
3019 if (value < 0 ) {
3020 txtLoc = checkIntSuffix(text, txtLoc, patLoc+1, TRUE);
3021 if (txtLoc != pos.getIndex()) {
3022 value *= -1;
3023 }
3024 }
3025 else {
3026 txtLoc = checkIntSuffix(text, txtLoc, patLoc+1, FALSE);
3027 }
3028
3029 if (!getBooleanAttribute(UDAT_PARSE_ALLOW_WHITESPACE, status)) {
3030 // Check the range of the value
3031 int32_t bias = gFieldRangeBias[patternCharIndex];
3032 if (bias >= 0 && (value > cal.getMaximum(field) + bias || value < cal.getMinimum(field) + bias)) {
3033 return -start;
3034 }
3035 }
3036
3037 pos.setIndex(txtLoc);
3038 }
3039 }
3040
3041 // Make sure that we got a number if
3042 // we want one, and didn't get one
3043 // if we don't want one.
3044 switch (patternCharIndex) {
3045 case UDAT_HOUR_OF_DAY1_FIELD:
3046 case UDAT_HOUR_OF_DAY0_FIELD:
3047 case UDAT_HOUR1_FIELD:
3048 case UDAT_HOUR0_FIELD:
3049 // special range check for hours:
3050 if (value < 0 || value > 24) {
3051 return -start;
3052 }
3053
3054 // fall through to gotNumber check
3055 U_FALLTHROUGH;
3056 case UDAT_YEAR_FIELD:
3057 case UDAT_YEAR_WOY_FIELD:
3058 case UDAT_FRACTIONAL_SECOND_FIELD:
3059 // these must be a number
3060 if (! gotNumber) {
3061 return -start;
3062 }
3063
3064 break;
3065
3066 default:
3067 // we check the rest of the fields below.
3068 break;
3069 }
3070
3071 switch (patternCharIndex) {
3072 case UDAT_ERA_FIELD:
3073 if (isChineseCalendar) {
3074 if (!gotNumber) {
3075 return -start;
3076 }
3077 cal.set(UCAL_ERA, value);
3078 return pos.getIndex();
3079 }
3080 if (count == 5) {
3081 ps = matchString(text, start, UCAL_ERA, fSymbols->fNarrowEras, fSymbols->fNarrowErasCount, NULL, cal);
3082 } else if (count == 4) {
3083 ps = matchString(text, start, UCAL_ERA, fSymbols->fEraNames, fSymbols->fEraNamesCount, NULL, cal);
3084 } else {
3085 ps = matchString(text, start, UCAL_ERA, fSymbols->fEras, fSymbols->fErasCount, NULL, cal);
3086 }
3087
3088 // check return position, if it equals -start, then matchString error
3089 // special case the return code so we don't necessarily fail out until we
3090 // verify no year information also
3091 if (ps == -start)
3092 ps--;
3093
3094 return ps;
3095
3096 case UDAT_YEAR_FIELD:
3097 // If there are 3 or more YEAR pattern characters, this indicates
3098 // that the year value is to be treated literally, without any
3099 // two-digit year adjustments (e.g., from "01" to 2001). Otherwise
3100 // we made adjustments to place the 2-digit year in the proper
3101 // century, for parsed strings from "00" to "99". Any other string
3102 // is treated literally: "2250", "-1", "1", "002".
3103 if (fDateOverride.compare(hebr)==0 && value < 1000) {
3104 value += HEBREW_CAL_CUR_MILLENIUM_START_YEAR;
3105 } else if (text.moveIndex32(start, 2) == pos.getIndex() && !isChineseCalendar
3106 && u_isdigit(text.char32At(start))
3107 && u_isdigit(text.char32At(text.moveIndex32(start, 1))))
3108 {
3109 // only adjust year for patterns less than 3.
3110 if(count < 3) {
3111 // Assume for example that the defaultCenturyStart is 6/18/1903.
3112 // This means that two-digit years will be forced into the range
3113 // 6/18/1903 to 6/17/2003. As a result, years 00, 01, and 02
3114 // correspond to 2000, 2001, and 2002. Years 04, 05, etc. correspond
3115 // to 1904, 1905, etc. If the year is 03, then it is 2003 if the
3116 // other fields specify a date before 6/18, or 1903 if they specify a
3117 // date afterwards. As a result, 03 is an ambiguous year. All other
3118 // two-digit years are unambiguous.
3119 if(fHaveDefaultCentury) { // check if this formatter even has a pivot year
3120 int32_t ambiguousTwoDigitYear = fDefaultCenturyStartYear % 100;
3121 ambiguousYear[0] = (value == ambiguousTwoDigitYear);
3122 value += (fDefaultCenturyStartYear/100)*100 +
3123 (value < ambiguousTwoDigitYear ? 100 : 0);
3124 }
3125 }
3126 }
3127 cal.set(UCAL_YEAR, value);
3128
3129 // Delayed checking for adjustment of Hebrew month numbers in non-leap years.
3130 if (saveHebrewMonth >= 0) {
3131 HebrewCalendar *hc = (HebrewCalendar*)&cal;
3132 if (!hc->isLeapYear(value) && saveHebrewMonth >= 6) {
3133 cal.set(UCAL_MONTH,saveHebrewMonth);
3134 } else {
3135 cal.set(UCAL_MONTH,saveHebrewMonth-1);
3136 }
3137 saveHebrewMonth = -1;
3138 }
3139 return pos.getIndex();
3140
3141 case UDAT_YEAR_WOY_FIELD:
3142 // Comment is the same as for UDAT_Year_FIELDs - look above
3143 if (fDateOverride.compare(hebr)==0 && value < 1000) {
3144 value += HEBREW_CAL_CUR_MILLENIUM_START_YEAR;
3145 } else if (text.moveIndex32(start, 2) == pos.getIndex()
3146 && u_isdigit(text.char32At(start))
3147 && u_isdigit(text.char32At(text.moveIndex32(start, 1)))
3148 && fHaveDefaultCentury )
3149 {
3150 int32_t ambiguousTwoDigitYear = fDefaultCenturyStartYear % 100;
3151 ambiguousYear[0] = (value == ambiguousTwoDigitYear);
3152 value += (fDefaultCenturyStartYear/100)*100 +
3153 (value < ambiguousTwoDigitYear ? 100 : 0);
3154 }
3155 cal.set(UCAL_YEAR_WOY, value);
3156 return pos.getIndex();
3157
3158 case UDAT_YEAR_NAME_FIELD:
3159 if (fSymbols->fShortYearNames != NULL) {
3160 int32_t newStart = matchString(text, start, UCAL_YEAR, fSymbols->fShortYearNames, fSymbols->fShortYearNamesCount, NULL, cal);
3161 if (newStart > 0) {
3162 return newStart;
3163 }
3164 }
3165 if (gotNumber && (getBooleanAttribute(UDAT_PARSE_ALLOW_NUMERIC,status) || value > fSymbols->fShortYearNamesCount)) {
3166 cal.set(UCAL_YEAR, value);
3167 return pos.getIndex();
3168 }
3169 return -start;
3170
3171 case UDAT_MONTH_FIELD:
3172 case UDAT_STANDALONE_MONTH_FIELD:
3173 if (gotNumber) // i.e., M or MM.
3174 {
3175 // When parsing month numbers from the Hebrew Calendar, we might need to adjust the month depending on whether
3176 // or not it was a leap year. We may or may not yet know what year it is, so might have to delay checking until
3177 // the year is parsed.
3178 if (!strcmp(cal.getType(),"hebrew")) {
3179 HebrewCalendar *hc = (HebrewCalendar*)&cal;
3180 if (cal.isSet(UCAL_YEAR)) {
3181 UErrorCode monthStatus = U_ZERO_ERROR;
3182 if (!hc->isLeapYear(hc->get(UCAL_YEAR, monthStatus)) && value >= 6) {
3183 cal.set(UCAL_MONTH, value);
3184 } else {
3185 cal.set(UCAL_MONTH, value - 1);
3186 }
3187 } else {
3188 saveHebrewMonth = value;
3189 }
3190 } else {
3191 // Don't want to parse the month if it is a string
3192 // while pattern uses numeric style: M/MM, L/LL
3193 // [We computed 'value' above.]
3194 cal.set(UCAL_MONTH, value - 1);
3195 }
3196 return pos.getIndex();
3197 } else {
3198 // count >= 3 // i.e., MMM/MMMM, LLL/LLLL
3199 // Want to be able to parse both short and long forms.
3200 // Try count == 4 first:
3201 UnicodeString * wideMonthPat = NULL;
3202 UnicodeString * shortMonthPat = NULL;
3203 if (fSymbols->fLeapMonthPatterns != NULL && fSymbols->fLeapMonthPatternsCount >= DateFormatSymbols::kMonthPatternsCount) {
3204 if (patternCharIndex==UDAT_MONTH_FIELD) {
3205 wideMonthPat = &fSymbols->fLeapMonthPatterns[DateFormatSymbols::kLeapMonthPatternFormatWide];
3206 shortMonthPat = &fSymbols->fLeapMonthPatterns[DateFormatSymbols::kLeapMonthPatternFormatAbbrev];
3207 } else {
3208 wideMonthPat = &fSymbols->fLeapMonthPatterns[DateFormatSymbols::kLeapMonthPatternStandaloneWide];
3209 shortMonthPat = &fSymbols->fLeapMonthPatterns[DateFormatSymbols::kLeapMonthPatternStandaloneAbbrev];
3210 }
3211 }
3212 int32_t newStart = 0;
3213 if (patternCharIndex==UDAT_MONTH_FIELD) {
3214 if(getBooleanAttribute(UDAT_PARSE_MULTIPLE_PATTERNS_FOR_MATCH, status) || count == 4) {
3215 newStart = matchString(text, start, UCAL_MONTH, fSymbols->fMonths, fSymbols->fMonthsCount, wideMonthPat, cal); // try MMMM
3216 if (newStart > 0) {
3217 return newStart;
3218 }
3219 }
3220 if(getBooleanAttribute(UDAT_PARSE_MULTIPLE_PATTERNS_FOR_MATCH, status) || count == 3) {
3221 newStart = matchString(text, start, UCAL_MONTH, fSymbols->fShortMonths, fSymbols->fShortMonthsCount, shortMonthPat, cal); // try MMM
3222 }
3223 } else {
3224 if(getBooleanAttribute(UDAT_PARSE_MULTIPLE_PATTERNS_FOR_MATCH, status) || count == 4) {
3225 newStart = matchString(text, start, UCAL_MONTH, fSymbols->fStandaloneMonths, fSymbols->fStandaloneMonthsCount, wideMonthPat, cal); // try LLLL
3226 if (newStart > 0) {
3227 return newStart;
3228 }
3229 }
3230 if(getBooleanAttribute(UDAT_PARSE_MULTIPLE_PATTERNS_FOR_MATCH, status) || count == 3) {
3231 newStart = matchString(text, start, UCAL_MONTH, fSymbols->fStandaloneShortMonths, fSymbols->fStandaloneShortMonthsCount, shortMonthPat, cal); // try LLL
3232 }
3233 }
3234 if (newStart > 0 || !getBooleanAttribute(UDAT_PARSE_ALLOW_NUMERIC, status)) // currently we do not try to parse MMMMM/LLLLL: #8860
3235 return newStart;
3236 // else we allowing parsing as number, below
3237 }
3238 break;
3239
3240 case UDAT_HOUR_OF_DAY1_FIELD:
3241 // [We computed 'value' above.]
3242 if (value == cal.getMaximum(UCAL_HOUR_OF_DAY) + 1)
3243 value = 0;
3244
3245 // fall through to set field
3246 U_FALLTHROUGH;
3247 case UDAT_HOUR_OF_DAY0_FIELD:
3248 cal.set(UCAL_HOUR_OF_DAY, value);
3249 return pos.getIndex();
3250
3251 case UDAT_FRACTIONAL_SECOND_FIELD:
3252 // Fractional seconds left-justify
3253 i = countDigits(text, start, pos.getIndex());
3254 if (i < 3) {
3255 while (i < 3) {
3256 value *= 10;
3257 i++;
3258 }
3259 } else {
3260 int32_t a = 1;
3261 while (i > 3) {
3262 a *= 10;
3263 i--;
3264 }
3265 value /= a;
3266 }
3267 cal.set(UCAL_MILLISECOND, value);
3268 return pos.getIndex();
3269
3270 case UDAT_DOW_LOCAL_FIELD:
3271 if (gotNumber) // i.e., e or ee
3272 {
3273 // [We computed 'value' above.]
3274 cal.set(UCAL_DOW_LOCAL, value);
3275 return pos.getIndex();
3276 }
3277 // else for eee-eeeee fall through to handling of EEE-EEEEE
3278 // fall through, do not break here
3279 U_FALLTHROUGH;
3280 case UDAT_DAY_OF_WEEK_FIELD:
3281 {
3282 // Want to be able to parse both short and long forms.
3283 // Try count == 4 (EEEE) wide first:
3284 int32_t newStart = 0;
3285 if(getBooleanAttribute(UDAT_PARSE_MULTIPLE_PATTERNS_FOR_MATCH, status) || count == 4) {
3286 if ((newStart = matchString(text, start, UCAL_DAY_OF_WEEK,
3287 fSymbols->fWeekdays, fSymbols->fWeekdaysCount, NULL, cal)) > 0)
3288 return newStart;
3289 }
3290 // EEEE wide failed, now try EEE abbreviated
3291 if(getBooleanAttribute(UDAT_PARSE_MULTIPLE_PATTERNS_FOR_MATCH, status) || count == 3) {
3292 if ((newStart = matchString(text, start, UCAL_DAY_OF_WEEK,
3293 fSymbols->fShortWeekdays, fSymbols->fShortWeekdaysCount, NULL, cal)) > 0)
3294 return newStart;
3295 }
3296 // EEE abbreviated failed, now try EEEEEE short
3297 if(getBooleanAttribute(UDAT_PARSE_MULTIPLE_PATTERNS_FOR_MATCH, status) || count == 6) {
3298 if ((newStart = matchString(text, start, UCAL_DAY_OF_WEEK,
3299 fSymbols->fShorterWeekdays, fSymbols->fShorterWeekdaysCount, NULL, cal)) > 0)
3300 return newStart;
3301 }
3302 // EEEEEE short failed, now try EEEEE narrow
3303 if(getBooleanAttribute(UDAT_PARSE_MULTIPLE_PATTERNS_FOR_MATCH, status) || count == 5) {
3304 if ((newStart = matchString(text, start, UCAL_DAY_OF_WEEK,
3305 fSymbols->fNarrowWeekdays, fSymbols->fNarrowWeekdaysCount, NULL, cal)) > 0)
3306 return newStart;
3307 }
3308 if (!getBooleanAttribute(UDAT_PARSE_ALLOW_NUMERIC, status) || patternCharIndex == UDAT_DAY_OF_WEEK_FIELD)
3309 return newStart;
3310 // else we allowing parsing as number, below
3311 }
3312 break;
3313
3314 case UDAT_STANDALONE_DAY_FIELD:
3315 {
3316 if (gotNumber) // c or cc
3317 {
3318 // [We computed 'value' above.]
3319 cal.set(UCAL_DOW_LOCAL, value);
3320 return pos.getIndex();
3321 }
3322 // Want to be able to parse both short and long forms.
3323 // Try count == 4 (cccc) first:
3324 int32_t newStart = 0;
3325 if(getBooleanAttribute(UDAT_PARSE_MULTIPLE_PATTERNS_FOR_MATCH, status) || count == 4) {
3326 if ((newStart = matchString(text, start, UCAL_DAY_OF_WEEK,
3327 fSymbols->fStandaloneWeekdays, fSymbols->fStandaloneWeekdaysCount, NULL, cal)) > 0)
3328 return newStart;
3329 }
3330 if(getBooleanAttribute(UDAT_PARSE_MULTIPLE_PATTERNS_FOR_MATCH, status) || count == 3) {
3331 if ((newStart = matchString(text, start, UCAL_DAY_OF_WEEK,
3332 fSymbols->fStandaloneShortWeekdays, fSymbols->fStandaloneShortWeekdaysCount, NULL, cal)) > 0)
3333 return newStart;
3334 }
3335 if(getBooleanAttribute(UDAT_PARSE_MULTIPLE_PATTERNS_FOR_MATCH, status) || count == 6) {
3336 if ((newStart = matchString(text, start, UCAL_DAY_OF_WEEK,
3337 fSymbols->fStandaloneShorterWeekdays, fSymbols->fStandaloneShorterWeekdaysCount, NULL, cal)) > 0)
3338 return newStart;
3339 }
3340 if (!getBooleanAttribute(UDAT_PARSE_ALLOW_NUMERIC, status))
3341 return newStart;
3342 // else we allowing parsing as number, below
3343 }
3344 break;
3345
3346 case UDAT_AM_PM_FIELD:
3347 {
3348 // optionally try both wide/abbrev and narrow forms
3349 int32_t newStart = 0;
3350 // try wide/abbrev
3351 if( getBooleanAttribute(UDAT_PARSE_MULTIPLE_PATTERNS_FOR_MATCH, status) || count < 5 ) {
3352 if ((newStart = matchString(text, start, UCAL_AM_PM, fSymbols->fAmPms, fSymbols->fAmPmsCount, NULL, cal)) > 0) {
3353 return newStart;
3354 }
3355 }
3356 // try narrow
3357 if( getBooleanAttribute(UDAT_PARSE_MULTIPLE_PATTERNS_FOR_MATCH, status) || count >= 5 ) {
3358 if ((newStart = matchString(text, start, UCAL_AM_PM, fSymbols->fNarrowAmPms, fSymbols->fNarrowAmPmsCount, NULL, cal)) > 0) {
3359 return newStart;
3360 }
3361 }
3362 // no matches for given options
3363 return -start;
3364 }
3365
3366 case UDAT_HOUR1_FIELD:
3367 // [We computed 'value' above.]
3368 if (value == cal.getLeastMaximum(UCAL_HOUR)+1)
3369 value = 0;
3370
3371 // fall through to set field
3372 U_FALLTHROUGH;
3373 case UDAT_HOUR0_FIELD:
3374 cal.set(UCAL_HOUR, value);
3375 return pos.getIndex();
3376
3377 case UDAT_QUARTER_FIELD:
3378 if (gotNumber) // i.e., Q or QQ.
3379 {
3380 // Don't want to parse the month if it is a string
3381 // while pattern uses numeric style: Q or QQ.
3382 // [We computed 'value' above.]
3383 cal.set(UCAL_MONTH, (value - 1) * 3);
3384 return pos.getIndex();
3385 } else {
3386 // count >= 3 // i.e., QQQ or QQQQ
3387 // Want to be able to parse both short and long forms.
3388 // Try count == 4 first:
3389 int32_t newStart = 0;
3390
3391 if(getBooleanAttribute(UDAT_PARSE_MULTIPLE_PATTERNS_FOR_MATCH, status) || count == 4) {
3392 if ((newStart = matchQuarterString(text, start, UCAL_MONTH,
3393 fSymbols->fQuarters, fSymbols->fQuartersCount, cal)) > 0)
3394 return newStart;
3395 }
3396 if(getBooleanAttribute(UDAT_PARSE_MULTIPLE_PATTERNS_FOR_MATCH, status) || count == 3) {
3397 if ((newStart = matchQuarterString(text, start, UCAL_MONTH,
3398 fSymbols->fShortQuarters, fSymbols->fShortQuartersCount, cal)) > 0)
3399 return newStart;
3400 }
3401 if (!getBooleanAttribute(UDAT_PARSE_ALLOW_NUMERIC, status))
3402 return newStart;
3403 // else we allowing parsing as number, below
3404 if(!getBooleanAttribute(UDAT_PARSE_MULTIPLE_PATTERNS_FOR_MATCH, status))
3405 return -start;
3406 }
3407 break;
3408
3409 case UDAT_STANDALONE_QUARTER_FIELD:
3410 if (gotNumber) // i.e., q or qq.
3411 {
3412 // Don't want to parse the month if it is a string
3413 // while pattern uses numeric style: q or q.
3414 // [We computed 'value' above.]
3415 cal.set(UCAL_MONTH, (value - 1) * 3);
3416 return pos.getIndex();
3417 } else {
3418 // count >= 3 // i.e., qqq or qqqq
3419 // Want to be able to parse both short and long forms.
3420 // Try count == 4 first:
3421 int32_t newStart = 0;
3422
3423 if(getBooleanAttribute(UDAT_PARSE_MULTIPLE_PATTERNS_FOR_MATCH, status) || count == 4) {
3424 if ((newStart = matchQuarterString(text, start, UCAL_MONTH,
3425 fSymbols->fStandaloneQuarters, fSymbols->fStandaloneQuartersCount, cal)) > 0)
3426 return newStart;
3427 }
3428 if(getBooleanAttribute(UDAT_PARSE_MULTIPLE_PATTERNS_FOR_MATCH, status) || count == 3) {
3429 if ((newStart = matchQuarterString(text, start, UCAL_MONTH,
3430 fSymbols->fStandaloneShortQuarters, fSymbols->fStandaloneShortQuartersCount, cal)) > 0)
3431 return newStart;
3432 }
3433 if (!getBooleanAttribute(UDAT_PARSE_ALLOW_NUMERIC, status))
3434 return newStart;
3435 // else we allowing parsing as number, below
3436 if(!getBooleanAttribute(UDAT_PARSE_MULTIPLE_PATTERNS_FOR_MATCH, status))
3437 return -start;
3438 }
3439 break;
3440
3441 case UDAT_TIMEZONE_FIELD: // 'z'
3442 {
3443 UTimeZoneFormatStyle style = (count < 4) ? UTZFMT_STYLE_SPECIFIC_SHORT : UTZFMT_STYLE_SPECIFIC_LONG;
3444 const TimeZoneFormat *tzfmt = tzFormat(status);
3445 if (U_SUCCESS(status)) {
3446 TimeZone *tz = tzfmt->parse(style, text, pos, tzTimeType);
3447 if (tz != NULL) {
3448 cal.adoptTimeZone(tz);
3449 return pos.getIndex();
3450 }
3451 }
3452 return -start;
3453 }
3454 break;
3455 case UDAT_TIMEZONE_RFC_FIELD: // 'Z'
3456 {
3457 UTimeZoneFormatStyle style = (count < 4) ?
3458 UTZFMT_STYLE_ISO_BASIC_LOCAL_FULL : ((count == 5) ? UTZFMT_STYLE_ISO_EXTENDED_FULL: UTZFMT_STYLE_LOCALIZED_GMT);
3459 const TimeZoneFormat *tzfmt = tzFormat(status);
3460 if (U_SUCCESS(status)) {
3461 TimeZone *tz = tzfmt->parse(style, text, pos, tzTimeType);
3462 if (tz != NULL) {
3463 cal.adoptTimeZone(tz);
3464 return pos.getIndex();
3465 }
3466 }
3467 return -start;
3468 }
3469 case UDAT_TIMEZONE_GENERIC_FIELD: // 'v'
3470 {
3471 UTimeZoneFormatStyle style = (count < 4) ? UTZFMT_STYLE_GENERIC_SHORT : UTZFMT_STYLE_GENERIC_LONG;
3472 const TimeZoneFormat *tzfmt = tzFormat(status);
3473 if (U_SUCCESS(status)) {
3474 TimeZone *tz = tzfmt->parse(style, text, pos, tzTimeType);
3475 if (tz != NULL) {
3476 cal.adoptTimeZone(tz);
3477 return pos.getIndex();
3478 }
3479 }
3480 return -start;
3481 }
3482 case UDAT_TIMEZONE_SPECIAL_FIELD: // 'V'
3483 {
3484 UTimeZoneFormatStyle style;
3485 switch (count) {
3486 case 1:
3487 style = UTZFMT_STYLE_ZONE_ID_SHORT;
3488 break;
3489 case 2:
3490 style = UTZFMT_STYLE_ZONE_ID;
3491 break;
3492 case 3:
3493 style = UTZFMT_STYLE_EXEMPLAR_LOCATION;
3494 break;
3495 default:
3496 style = UTZFMT_STYLE_GENERIC_LOCATION;
3497 break;
3498 }
3499 const TimeZoneFormat *tzfmt = tzFormat(status);
3500 if (U_SUCCESS(status)) {
3501 TimeZone *tz = tzfmt->parse(style, text, pos, tzTimeType);
3502 if (tz != NULL) {
3503 cal.adoptTimeZone(tz);
3504 return pos.getIndex();
3505 }
3506 }
3507 return -start;
3508 }
3509 case UDAT_TIMEZONE_LOCALIZED_GMT_OFFSET_FIELD: // 'O'
3510 {
3511 UTimeZoneFormatStyle style = (count < 4) ? UTZFMT_STYLE_LOCALIZED_GMT_SHORT : UTZFMT_STYLE_LOCALIZED_GMT;
3512 const TimeZoneFormat *tzfmt = tzFormat(status);
3513 if (U_SUCCESS(status)) {
3514 TimeZone *tz = tzfmt->parse(style, text, pos, tzTimeType);
3515 if (tz != NULL) {
3516 cal.adoptTimeZone(tz);
3517 return pos.getIndex();
3518 }
3519 }
3520 return -start;
3521 }
3522 case UDAT_TIMEZONE_ISO_FIELD: // 'X'
3523 {
3524 UTimeZoneFormatStyle style;
3525 switch (count) {
3526 case 1:
3527 style = UTZFMT_STYLE_ISO_BASIC_SHORT;
3528 break;
3529 case 2:
3530 style = UTZFMT_STYLE_ISO_BASIC_FIXED;
3531 break;
3532 case 3:
3533 style = UTZFMT_STYLE_ISO_EXTENDED_FIXED;
3534 break;
3535 case 4:
3536 style = UTZFMT_STYLE_ISO_BASIC_FULL;
3537 break;
3538 default:
3539 style = UTZFMT_STYLE_ISO_EXTENDED_FULL;
3540 break;
3541 }
3542 const TimeZoneFormat *tzfmt = tzFormat(status);
3543 if (U_SUCCESS(status)) {
3544 TimeZone *tz = tzfmt->parse(style, text, pos, tzTimeType);
3545 if (tz != NULL) {
3546 cal.adoptTimeZone(tz);
3547 return pos.getIndex();
3548 }
3549 }
3550 return -start;
3551 }
3552 case UDAT_TIMEZONE_ISO_LOCAL_FIELD: // 'x'
3553 {
3554 UTimeZoneFormatStyle style;
3555 switch (count) {
3556 case 1:
3557 style = UTZFMT_STYLE_ISO_BASIC_LOCAL_SHORT;
3558 break;
3559 case 2:
3560 style = UTZFMT_STYLE_ISO_BASIC_LOCAL_FIXED;
3561 break;
3562 case 3:
3563 style = UTZFMT_STYLE_ISO_EXTENDED_LOCAL_FIXED;
3564 break;
3565 case 4:
3566 style = UTZFMT_STYLE_ISO_BASIC_LOCAL_FULL;
3567 break;
3568 default:
3569 style = UTZFMT_STYLE_ISO_EXTENDED_LOCAL_FULL;
3570 break;
3571 }
3572 const TimeZoneFormat *tzfmt = tzFormat(status);
3573 if (U_SUCCESS(status)) {
3574 TimeZone *tz = tzfmt->parse(style, text, pos, tzTimeType);
3575 if (tz != NULL) {
3576 cal.adoptTimeZone(tz);
3577 return pos.getIndex();
3578 }
3579 }
3580 return -start;
3581 }
3582 // currently no pattern character is defined for UDAT_TIME_SEPARATOR_FIELD
3583 // so we should not get here. Leave support in for future definition.
3584 case UDAT_TIME_SEPARATOR_FIELD:
3585 {
3586 static const UChar def_sep = DateFormatSymbols::DEFAULT_TIME_SEPARATOR;
3587 static const UChar alt_sep = DateFormatSymbols::ALTERNATE_TIME_SEPARATOR;
3588
3589 // Try matching a time separator.
3590 int32_t count_sep = 1;
3591 UnicodeString data[3];
3592 fSymbols->getTimeSeparatorString(data[0]);
3593
3594 // Add the default, if different from the locale.
3595 if (data[0].compare(&def_sep, 1) != 0) {
3596 data[count_sep++].setTo(def_sep);
3597 }
3598
3599 // If lenient, add also the alternate, if different from the locale.
3600 if (isLenient() && data[0].compare(&alt_sep, 1) != 0) {
3601 data[count_sep++].setTo(alt_sep);
3602 }
3603
3604 return matchString(text, start, UCAL_FIELD_COUNT /* => nothing to set */, data, count_sep, NULL, cal);
3605 }
3606
3607 case UDAT_AM_PM_MIDNIGHT_NOON_FIELD:
3608 {
3609 U_ASSERT(dayPeriod != NULL);
3610 int32_t ampmStart = subParse(text, start, 0x61, count,
3611 obeyCount, allowNegative, ambiguousYear, saveHebrewMonth, cal,
3612 patLoc, numericLeapMonthFormatter, tzTimeType);
3613
3614 if (ampmStart > 0) {
3615 return ampmStart;
3616 } else {
3617 int32_t newStart = 0;
3618
3619 // Only match the first two strings from the day period strings array.
3620 if (getBooleanAttribute(UDAT_PARSE_MULTIPLE_PATTERNS_FOR_MATCH, status) || count == 3) {
3621 if ((newStart = matchDayPeriodStrings(text, start, fSymbols->fAbbreviatedDayPeriods,
3622 2, *dayPeriod)) > 0) {
3623 return newStart;
3624 }
3625 }
3626 if (getBooleanAttribute(UDAT_PARSE_MULTIPLE_PATTERNS_FOR_MATCH, status) || count == 5) {
3627 if ((newStart = matchDayPeriodStrings(text, start, fSymbols->fNarrowDayPeriods,
3628 2, *dayPeriod)) > 0) {
3629 return newStart;
3630 }
3631 }
3632 // count == 4, but allow other counts
3633 if (getBooleanAttribute(UDAT_PARSE_MULTIPLE_PATTERNS_FOR_MATCH, status)) {
3634 if ((newStart = matchDayPeriodStrings(text, start, fSymbols->fWideDayPeriods,
3635 2, *dayPeriod)) > 0) {
3636 return newStart;
3637 }
3638 }
3639
3640 return -start;
3641 }
3642 }
3643
3644 case UDAT_FLEXIBLE_DAY_PERIOD_FIELD:
3645 {
3646 U_ASSERT(dayPeriod != NULL);
3647 int32_t newStart = 0;
3648
3649 if (getBooleanAttribute(UDAT_PARSE_MULTIPLE_PATTERNS_FOR_MATCH, status) || count == 3) {
3650 if ((newStart = matchDayPeriodStrings(text, start, fSymbols->fAbbreviatedDayPeriods,
3651 fSymbols->fAbbreviatedDayPeriodsCount, *dayPeriod)) > 0) {
3652 return newStart;
3653 }
3654 }
3655 if (getBooleanAttribute(UDAT_PARSE_MULTIPLE_PATTERNS_FOR_MATCH, status) || count == 5) {
3656 if ((newStart = matchDayPeriodStrings(text, start, fSymbols->fNarrowDayPeriods,
3657 fSymbols->fNarrowDayPeriodsCount, *dayPeriod)) > 0) {
3658 return newStart;
3659 }
3660 }
3661 if (getBooleanAttribute(UDAT_PARSE_MULTIPLE_PATTERNS_FOR_MATCH, status) || count == 4) {
3662 if ((newStart = matchDayPeriodStrings(text, start, fSymbols->fWideDayPeriods,
3663 fSymbols->fWideDayPeriodsCount, *dayPeriod)) > 0) {
3664 return newStart;
3665 }
3666 }
3667
3668 return -start;
3669 }
3670
3671 default:
3672 // Handle "generic" fields
3673 // this is now handled below, outside the switch block
3674 break;
3675 }
3676 // Handle "generic" fields:
3677 // switch default case now handled here (outside switch block) to allow
3678 // parsing of some string fields as digits for lenient case
3679
3680 int32_t parseStart = pos.getIndex();
3681 const UnicodeString* src;
3682 if (obeyCount) {
3683 if ((start+count) > text.length()) {
3684 return -start;
3685 }
3686 text.extractBetween(0, start + count, temp);
3687 src = &temp;
3688 } else {
3689 src = &text;
3690 }
3691 parseInt(*src, number, pos, allowNegative,currentNumberFormat);
3692 if (pos.getIndex() != parseStart) {
3693 int32_t val = number.getLong();
3694
3695 // Don't need suffix processing here (as in number processing at the beginning of the function);
3696 // the new fields being handled as numeric values (month, weekdays, quarters) should not have suffixes.
3697
3698 if (!getBooleanAttribute(UDAT_PARSE_ALLOW_NUMERIC, status)) {
3699 // Check the range of the value
3700 int32_t bias = gFieldRangeBias[patternCharIndex];
3701 if (bias >= 0 && (val > cal.getMaximum(field) + bias || val < cal.getMinimum(field) + bias)) {
3702 return -start;
3703 }
3704 }
3705
3706 // For the following, need to repeat some of the "if (gotNumber)" code above:
3707 // UDAT_[STANDALONE_]MONTH_FIELD, UDAT_DOW_LOCAL_FIELD, UDAT_STANDALONE_DAY_FIELD,
3708 // UDAT_[STANDALONE_]QUARTER_FIELD
3709 switch (patternCharIndex) {
3710 case UDAT_MONTH_FIELD:
3711 // See notes under UDAT_MONTH_FIELD case above
3712 if (!strcmp(cal.getType(),"hebrew")) {
3713 HebrewCalendar *hc = (HebrewCalendar*)&cal;
3714 if (cal.isSet(UCAL_YEAR)) {
3715 UErrorCode monthStatus = U_ZERO_ERROR;
3716 if (!hc->isLeapYear(hc->get(UCAL_YEAR, monthStatus)) && val >= 6) {
3717 cal.set(UCAL_MONTH, val);
3718 } else {
3719 cal.set(UCAL_MONTH, val - 1);
3720 }
3721 } else {
3722 saveHebrewMonth = val;
3723 }
3724 } else {
3725 cal.set(UCAL_MONTH, val - 1);
3726 }
3727 break;
3728 case UDAT_STANDALONE_MONTH_FIELD:
3729 cal.set(UCAL_MONTH, val - 1);
3730 break;
3731 case UDAT_DOW_LOCAL_FIELD:
3732 case UDAT_STANDALONE_DAY_FIELD:
3733 cal.set(UCAL_DOW_LOCAL, val);
3734 break;
3735 case UDAT_QUARTER_FIELD:
3736 case UDAT_STANDALONE_QUARTER_FIELD:
3737 cal.set(UCAL_MONTH, (val - 1) * 3);
3738 break;
3739 case UDAT_RELATED_YEAR_FIELD:
3740 cal.setRelatedYear(val);
3741 break;
3742 default:
3743 cal.set(field, val);
3744 break;
3745 }
3746 return pos.getIndex();
3747 }
3748 return -start;
3749 }
3750
3751 /**
3752 * Parse an integer using fNumberFormat. This method is semantically
3753 * const, but actually may modify fNumberFormat.
3754 */
parseInt(const UnicodeString & text,Formattable & number,ParsePosition & pos,UBool allowNegative,const NumberFormat * fmt) const3755 void SimpleDateFormat::parseInt(const UnicodeString& text,
3756 Formattable& number,
3757 ParsePosition& pos,
3758 UBool allowNegative,
3759 const NumberFormat *fmt) const {
3760 parseInt(text, number, -1, pos, allowNegative,fmt);
3761 }
3762
3763 /**
3764 * Parse an integer using fNumberFormat up to maxDigits.
3765 */
parseInt(const UnicodeString & text,Formattable & number,int32_t maxDigits,ParsePosition & pos,UBool allowNegative,const NumberFormat * fmt) const3766 void SimpleDateFormat::parseInt(const UnicodeString& text,
3767 Formattable& number,
3768 int32_t maxDigits,
3769 ParsePosition& pos,
3770 UBool allowNegative,
3771 const NumberFormat *fmt) const {
3772 UnicodeString oldPrefix;
3773 auto* fmtAsDF = dynamic_cast<const DecimalFormat*>(fmt);
3774 LocalPointer<DecimalFormat> df;
3775 if (!allowNegative && fmtAsDF != nullptr) {
3776 df.adoptInstead(fmtAsDF->clone());
3777 if (df.isNull()) {
3778 // Memory allocation error
3779 return;
3780 }
3781 df->setNegativePrefix(UnicodeString(TRUE, SUPPRESS_NEGATIVE_PREFIX, -1));
3782 fmt = df.getAlias();
3783 }
3784 int32_t oldPos = pos.getIndex();
3785 fmt->parse(text, number, pos);
3786
3787 if (maxDigits > 0) {
3788 // adjust the result to fit into
3789 // the maxDigits and move the position back
3790 int32_t nDigits = pos.getIndex() - oldPos;
3791 if (nDigits > maxDigits) {
3792 int32_t val = number.getLong();
3793 nDigits -= maxDigits;
3794 while (nDigits > 0) {
3795 val /= 10;
3796 nDigits--;
3797 }
3798 pos.setIndex(oldPos + maxDigits);
3799 number.setLong(val);
3800 }
3801 }
3802 }
3803
countDigits(const UnicodeString & text,int32_t start,int32_t end) const3804 int32_t SimpleDateFormat::countDigits(const UnicodeString& text, int32_t start, int32_t end) const {
3805 int32_t numDigits = 0;
3806 int32_t idx = start;
3807 while (idx < end) {
3808 UChar32 cp = text.char32At(idx);
3809 if (u_isdigit(cp)) {
3810 numDigits++;
3811 }
3812 idx += U16_LENGTH(cp);
3813 }
3814 return numDigits;
3815 }
3816
3817 //----------------------------------------------------------------------
3818
translatePattern(const UnicodeString & originalPattern,UnicodeString & translatedPattern,const UnicodeString & from,const UnicodeString & to,UErrorCode & status)3819 void SimpleDateFormat::translatePattern(const UnicodeString& originalPattern,
3820 UnicodeString& translatedPattern,
3821 const UnicodeString& from,
3822 const UnicodeString& to,
3823 UErrorCode& status)
3824 {
3825 // run through the pattern and convert any pattern symbols from the version
3826 // in "from" to the corresponding character in "to". This code takes
3827 // quoted strings into account (it doesn't try to translate them), and it signals
3828 // an error if a particular "pattern character" doesn't appear in "from".
3829 // Depending on the values of "from" and "to" this can convert from generic
3830 // to localized patterns or localized to generic.
3831 if (U_FAILURE(status)) {
3832 return;
3833 }
3834
3835 translatedPattern.remove();
3836 UBool inQuote = FALSE;
3837 for (int32_t i = 0; i < originalPattern.length(); ++i) {
3838 UChar c = originalPattern[i];
3839 if (inQuote) {
3840 if (c == QUOTE) {
3841 inQuote = FALSE;
3842 }
3843 } else {
3844 if (c == QUOTE) {
3845 inQuote = TRUE;
3846 } else if (isSyntaxChar(c)) {
3847 int32_t ci = from.indexOf(c);
3848 if (ci == -1) {
3849 status = U_INVALID_FORMAT_ERROR;
3850 return;
3851 }
3852 c = to[ci];
3853 }
3854 }
3855 translatedPattern += c;
3856 }
3857 if (inQuote) {
3858 status = U_INVALID_FORMAT_ERROR;
3859 return;
3860 }
3861 }
3862
3863 //----------------------------------------------------------------------
3864
3865 UnicodeString&
toPattern(UnicodeString & result) const3866 SimpleDateFormat::toPattern(UnicodeString& result) const
3867 {
3868 result = fPattern;
3869 return result;
3870 }
3871
3872 //----------------------------------------------------------------------
3873
3874 UnicodeString&
toLocalizedPattern(UnicodeString & result,UErrorCode & status) const3875 SimpleDateFormat::toLocalizedPattern(UnicodeString& result,
3876 UErrorCode& status) const
3877 {
3878 translatePattern(fPattern, result,
3879 UnicodeString(DateFormatSymbols::getPatternUChars()),
3880 fSymbols->fLocalPatternChars, status);
3881 return result;
3882 }
3883
3884 //----------------------------------------------------------------------
3885
3886 void
applyPattern(const UnicodeString & pattern)3887 SimpleDateFormat::applyPattern(const UnicodeString& pattern)
3888 {
3889 fPattern = pattern;
3890 parsePattern();
3891
3892 // Hack to update use of Gannen year numbering for ja@calendar=japanese -
3893 // use only if format is non-numeric (includes 年) and no other fDateOverride.
3894 if (fCalendar != nullptr && uprv_strcmp(fCalendar->getType(),"japanese") == 0 &&
3895 uprv_strcmp(fLocale.getLanguage(),"ja") == 0) {
3896 if (fDateOverride==UnicodeString(u"y=jpanyear") && !fHasHanYearChar) {
3897 // Gannen numbering is set but new pattern should not use it, unset;
3898 // use procedure from adoptNumberFormat to clear overrides
3899 if (fSharedNumberFormatters) {
3900 freeSharedNumberFormatters(fSharedNumberFormatters);
3901 fSharedNumberFormatters = NULL;
3902 }
3903 fDateOverride.setToBogus(); // record status
3904 } else if (fDateOverride.isBogus() && fHasHanYearChar) {
3905 // No current override (=> no Gannen numbering) but new pattern needs it;
3906 // use procedures from initNUmberFormatters / adoptNumberFormat
3907 umtx_lock(&LOCK);
3908 if (fSharedNumberFormatters == NULL) {
3909 fSharedNumberFormatters = allocSharedNumberFormatters();
3910 }
3911 umtx_unlock(&LOCK);
3912 if (fSharedNumberFormatters != NULL) {
3913 Locale ovrLoc(fLocale.getLanguage(),fLocale.getCountry(),fLocale.getVariant(),"numbers=jpanyear");
3914 UErrorCode status = U_ZERO_ERROR;
3915 const SharedNumberFormat *snf = createSharedNumberFormat(ovrLoc, status);
3916 if (U_SUCCESS(status)) {
3917 // Now that we have an appropriate number formatter, fill in the
3918 // appropriate slot in the number formatters table.
3919 UDateFormatField patternCharIndex = DateFormatSymbols::getPatternCharIndex(u'y');
3920 SharedObject::copyPtr(snf, fSharedNumberFormatters[patternCharIndex]);
3921 snf->deleteIfZeroRefCount();
3922 fDateOverride.setTo(u"y=jpanyear", -1); // record status
3923 }
3924 }
3925 }
3926 }
3927 }
3928
3929 //----------------------------------------------------------------------
3930
3931 void
applyLocalizedPattern(const UnicodeString & pattern,UErrorCode & status)3932 SimpleDateFormat::applyLocalizedPattern(const UnicodeString& pattern,
3933 UErrorCode &status)
3934 {
3935 translatePattern(pattern, fPattern,
3936 fSymbols->fLocalPatternChars,
3937 UnicodeString(DateFormatSymbols::getPatternUChars()), status);
3938 }
3939
3940 //----------------------------------------------------------------------
3941
3942 const DateFormatSymbols*
getDateFormatSymbols() const3943 SimpleDateFormat::getDateFormatSymbols() const
3944 {
3945 return fSymbols;
3946 }
3947
3948 //----------------------------------------------------------------------
3949
3950 void
adoptDateFormatSymbols(DateFormatSymbols * newFormatSymbols)3951 SimpleDateFormat::adoptDateFormatSymbols(DateFormatSymbols* newFormatSymbols)
3952 {
3953 delete fSymbols;
3954 fSymbols = newFormatSymbols;
3955 }
3956
3957 //----------------------------------------------------------------------
3958 void
setDateFormatSymbols(const DateFormatSymbols & newFormatSymbols)3959 SimpleDateFormat::setDateFormatSymbols(const DateFormatSymbols& newFormatSymbols)
3960 {
3961 delete fSymbols;
3962 fSymbols = new DateFormatSymbols(newFormatSymbols);
3963 }
3964
3965 //----------------------------------------------------------------------
3966 const TimeZoneFormat*
getTimeZoneFormat(void) const3967 SimpleDateFormat::getTimeZoneFormat(void) const {
3968 // TimeZoneFormat initialization might fail when out of memory.
3969 // If we always initialize TimeZoneFormat instance, we can return
3970 // such status there. For now, this implementation lazily instantiates
3971 // a TimeZoneFormat for performance optimization reasons, but cannot
3972 // propagate such error (probably just out of memory case) to the caller.
3973 UErrorCode status = U_ZERO_ERROR;
3974 return (const TimeZoneFormat*)tzFormat(status);
3975 }
3976
3977 //----------------------------------------------------------------------
3978 void
adoptTimeZoneFormat(TimeZoneFormat * timeZoneFormatToAdopt)3979 SimpleDateFormat::adoptTimeZoneFormat(TimeZoneFormat* timeZoneFormatToAdopt)
3980 {
3981 delete fTimeZoneFormat;
3982 fTimeZoneFormat = timeZoneFormatToAdopt;
3983 }
3984
3985 //----------------------------------------------------------------------
3986 void
setTimeZoneFormat(const TimeZoneFormat & newTimeZoneFormat)3987 SimpleDateFormat::setTimeZoneFormat(const TimeZoneFormat& newTimeZoneFormat)
3988 {
3989 delete fTimeZoneFormat;
3990 fTimeZoneFormat = new TimeZoneFormat(newTimeZoneFormat);
3991 }
3992
3993 //----------------------------------------------------------------------
3994
3995
adoptCalendar(Calendar * calendarToAdopt)3996 void SimpleDateFormat::adoptCalendar(Calendar* calendarToAdopt)
3997 {
3998 UErrorCode status = U_ZERO_ERROR;
3999 Locale calLocale(fLocale);
4000 calLocale.setKeywordValue("calendar", calendarToAdopt->getType(), status);
4001 DateFormatSymbols *newSymbols =
4002 DateFormatSymbols::createForLocale(calLocale, status);
4003 if (U_FAILURE(status)) {
4004 delete calendarToAdopt;
4005 return;
4006 }
4007 DateFormat::adoptCalendar(calendarToAdopt);
4008 delete fSymbols;
4009 fSymbols = newSymbols;
4010 initializeDefaultCentury(); // we need a new century (possibly)
4011 }
4012
4013
4014 //----------------------------------------------------------------------
4015
4016
4017 // override the DateFormat implementation in order to
4018 // lazily initialize fCapitalizationBrkIter
4019 void
setContext(UDisplayContext value,UErrorCode & status)4020 SimpleDateFormat::setContext(UDisplayContext value, UErrorCode& status)
4021 {
4022 DateFormat::setContext(value, status);
4023 #if !UCONFIG_NO_BREAK_ITERATION
4024 if (U_SUCCESS(status)) {
4025 if ( fCapitalizationBrkIter == NULL && (value==UDISPCTX_CAPITALIZATION_FOR_BEGINNING_OF_SENTENCE ||
4026 value==UDISPCTX_CAPITALIZATION_FOR_UI_LIST_OR_MENU || value==UDISPCTX_CAPITALIZATION_FOR_STANDALONE) ) {
4027 status = U_ZERO_ERROR;
4028 fCapitalizationBrkIter = BreakIterator::createSentenceInstance(fLocale, status);
4029 if (U_FAILURE(status)) {
4030 delete fCapitalizationBrkIter;
4031 fCapitalizationBrkIter = NULL;
4032 }
4033 }
4034 }
4035 #endif
4036 }
4037
4038
4039 //----------------------------------------------------------------------
4040
4041
4042 UBool
isFieldUnitIgnored(UCalendarDateFields field) const4043 SimpleDateFormat::isFieldUnitIgnored(UCalendarDateFields field) const {
4044 return isFieldUnitIgnored(fPattern, field);
4045 }
4046
4047
4048 UBool
isFieldUnitIgnored(const UnicodeString & pattern,UCalendarDateFields field)4049 SimpleDateFormat::isFieldUnitIgnored(const UnicodeString& pattern,
4050 UCalendarDateFields field) {
4051 int32_t fieldLevel = fgCalendarFieldToLevel[field];
4052 int32_t level;
4053 UChar ch;
4054 UBool inQuote = FALSE;
4055 UChar prevCh = 0;
4056 int32_t count = 0;
4057
4058 for (int32_t i = 0; i < pattern.length(); ++i) {
4059 ch = pattern[i];
4060 if (ch != prevCh && count > 0) {
4061 level = getLevelFromChar(prevCh);
4062 // the larger the level, the smaller the field unit.
4063 if (fieldLevel <= level) {
4064 return FALSE;
4065 }
4066 count = 0;
4067 }
4068 if (ch == QUOTE) {
4069 if ((i+1) < pattern.length() && pattern[i+1] == QUOTE) {
4070 ++i;
4071 } else {
4072 inQuote = ! inQuote;
4073 }
4074 }
4075 else if (!inQuote && isSyntaxChar(ch)) {
4076 prevCh = ch;
4077 ++count;
4078 }
4079 }
4080 if (count > 0) {
4081 // last item
4082 level = getLevelFromChar(prevCh);
4083 if (fieldLevel <= level) {
4084 return FALSE;
4085 }
4086 }
4087 return TRUE;
4088 }
4089
4090 //----------------------------------------------------------------------
4091
4092 const Locale&
getSmpFmtLocale(void) const4093 SimpleDateFormat::getSmpFmtLocale(void) const {
4094 return fLocale;
4095 }
4096
4097 //----------------------------------------------------------------------
4098
4099 int32_t
checkIntSuffix(const UnicodeString & text,int32_t start,int32_t patLoc,UBool isNegative) const4100 SimpleDateFormat::checkIntSuffix(const UnicodeString& text, int32_t start,
4101 int32_t patLoc, UBool isNegative) const {
4102 // local variables
4103 UnicodeString suf;
4104 int32_t patternMatch;
4105 int32_t textPreMatch;
4106 int32_t textPostMatch;
4107
4108 // check that we are still in range
4109 if ( (start > text.length()) ||
4110 (start < 0) ||
4111 (patLoc < 0) ||
4112 (patLoc > fPattern.length())) {
4113 // out of range, don't advance location in text
4114 return start;
4115 }
4116
4117 // get the suffix
4118 DecimalFormat* decfmt = dynamic_cast<DecimalFormat*>(fNumberFormat);
4119 if (decfmt != NULL) {
4120 if (isNegative) {
4121 suf = decfmt->getNegativeSuffix(suf);
4122 }
4123 else {
4124 suf = decfmt->getPositiveSuffix(suf);
4125 }
4126 }
4127
4128 // check for suffix
4129 if (suf.length() <= 0) {
4130 return start;
4131 }
4132
4133 // check suffix will be encountered in the pattern
4134 patternMatch = compareSimpleAffix(suf,fPattern,patLoc);
4135
4136 // check if a suffix will be encountered in the text
4137 textPreMatch = compareSimpleAffix(suf,text,start);
4138
4139 // check if a suffix was encountered in the text
4140 textPostMatch = compareSimpleAffix(suf,text,start-suf.length());
4141
4142 // check for suffix match
4143 if ((textPreMatch >= 0) && (patternMatch >= 0) && (textPreMatch == patternMatch)) {
4144 return start;
4145 }
4146 else if ((textPostMatch >= 0) && (patternMatch >= 0) && (textPostMatch == patternMatch)) {
4147 return start - suf.length();
4148 }
4149
4150 // should not get here
4151 return start;
4152 }
4153
4154 //----------------------------------------------------------------------
4155
4156 int32_t
compareSimpleAffix(const UnicodeString & affix,const UnicodeString & input,int32_t pos) const4157 SimpleDateFormat::compareSimpleAffix(const UnicodeString& affix,
4158 const UnicodeString& input,
4159 int32_t pos) const {
4160 int32_t start = pos;
4161 for (int32_t i=0; i<affix.length(); ) {
4162 UChar32 c = affix.char32At(i);
4163 int32_t len = U16_LENGTH(c);
4164 if (PatternProps::isWhiteSpace(c)) {
4165 // We may have a pattern like: \u200F \u0020
4166 // and input text like: \u200F \u0020
4167 // Note that U+200F and U+0020 are Pattern_White_Space but only
4168 // U+0020 is UWhiteSpace. So we have to first do a direct
4169 // match of the run of Pattern_White_Space in the pattern,
4170 // then match any extra characters.
4171 UBool literalMatch = FALSE;
4172 while (pos < input.length() &&
4173 input.char32At(pos) == c) {
4174 literalMatch = TRUE;
4175 i += len;
4176 pos += len;
4177 if (i == affix.length()) {
4178 break;
4179 }
4180 c = affix.char32At(i);
4181 len = U16_LENGTH(c);
4182 if (!PatternProps::isWhiteSpace(c)) {
4183 break;
4184 }
4185 }
4186
4187 // Advance over run in pattern
4188 i = skipPatternWhiteSpace(affix, i);
4189
4190 // Advance over run in input text
4191 // Must see at least one white space char in input,
4192 // unless we've already matched some characters literally.
4193 int32_t s = pos;
4194 pos = skipUWhiteSpace(input, pos);
4195 if (pos == s && !literalMatch) {
4196 return -1;
4197 }
4198
4199 // If we skip UWhiteSpace in the input text, we need to skip it in the pattern.
4200 // Otherwise, the previous lines may have skipped over text (such as U+00A0) that
4201 // is also in the affix.
4202 i = skipUWhiteSpace(affix, i);
4203 } else {
4204 if (pos < input.length() &&
4205 input.char32At(pos) == c) {
4206 i += len;
4207 pos += len;
4208 } else {
4209 return -1;
4210 }
4211 }
4212 }
4213 return pos - start;
4214 }
4215
4216 //----------------------------------------------------------------------
4217
4218 int32_t
skipPatternWhiteSpace(const UnicodeString & text,int32_t pos) const4219 SimpleDateFormat::skipPatternWhiteSpace(const UnicodeString& text, int32_t pos) const {
4220 const UChar* s = text.getBuffer();
4221 return (int32_t)(PatternProps::skipWhiteSpace(s + pos, text.length() - pos) - s);
4222 }
4223
4224 //----------------------------------------------------------------------
4225
4226 int32_t
skipUWhiteSpace(const UnicodeString & text,int32_t pos) const4227 SimpleDateFormat::skipUWhiteSpace(const UnicodeString& text, int32_t pos) const {
4228 while (pos < text.length()) {
4229 UChar32 c = text.char32At(pos);
4230 if (!u_isUWhiteSpace(c)) {
4231 break;
4232 }
4233 pos += U16_LENGTH(c);
4234 }
4235 return pos;
4236 }
4237
4238 //----------------------------------------------------------------------
4239
4240 // Lazy TimeZoneFormat instantiation, semantically const.
4241 TimeZoneFormat *
tzFormat(UErrorCode & status) const4242 SimpleDateFormat::tzFormat(UErrorCode &status) const {
4243 if (fTimeZoneFormat == NULL) {
4244 umtx_lock(&LOCK);
4245 {
4246 if (fTimeZoneFormat == NULL) {
4247 TimeZoneFormat *tzfmt = TimeZoneFormat::createInstance(fLocale, status);
4248 if (U_FAILURE(status)) {
4249 return NULL;
4250 }
4251
4252 const_cast<SimpleDateFormat *>(this)->fTimeZoneFormat = tzfmt;
4253 }
4254 }
4255 umtx_unlock(&LOCK);
4256 }
4257 return fTimeZoneFormat;
4258 }
4259
parsePattern()4260 void SimpleDateFormat::parsePattern() {
4261 fHasMinute = FALSE;
4262 fHasSecond = FALSE;
4263 fHasHanYearChar = FALSE;
4264
4265 int len = fPattern.length();
4266 UBool inQuote = FALSE;
4267 for (int32_t i = 0; i < len; ++i) {
4268 UChar ch = fPattern[i];
4269 if (ch == QUOTE) {
4270 inQuote = !inQuote;
4271 }
4272 if (ch == 0x5E74) { // don't care whether this is inside quotes
4273 fHasHanYearChar = TRUE;
4274 }
4275 if (!inQuote) {
4276 if (ch == 0x6D) { // 0x6D == 'm'
4277 fHasMinute = TRUE;
4278 }
4279 if (ch == 0x73) { // 0x73 == 's'
4280 fHasSecond = TRUE;
4281 }
4282 }
4283 }
4284 }
4285
4286 U_NAMESPACE_END
4287
4288 #endif /* #if !UCONFIG_NO_FORMATTING */
4289
4290 //eof
4291