1 // Copyright (C) 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/utf16.h"
52 #include "unicode/vtzone.h"
53 #include "unicode/udisplaycontext.h"
54 #include "unicode/brkiter.h"
55 #include "uresimp.h"
56 #include "olsontz.h"
57 #include "patternprops.h"
58 #include "fphdlimp.h"
59 #include "hebrwcal.h"
60 #include "cstring.h"
61 #include "uassert.h"
62 #include "cmemory.h"
63 #include "umutex.h"
64 #include <float.h>
65 #include "smpdtfst.h"
66 #include "sharednumberformat.h"
67 #include "ustr_imp.h"
68 #include "charstr.h"
69 #include "uvector.h"
70 #include "cstr.h"
71 #include "dayperiodrules.h"
72
73 #if defined( U_DEBUG_CALSVC ) || defined (U_DEBUG_CAL)
74 #include <stdio.h>
75 #endif
76
77 // *****************************************************************************
78 // class SimpleDateFormat
79 // *****************************************************************************
80
81 U_NAMESPACE_BEGIN
82
83 /**
84 * Last-resort string to use for "GMT" when constructing time zone strings.
85 */
86 // For time zones that have no names, use strings GMT+minutes and
87 // GMT-minutes. For instance, in France the time zone is GMT+60.
88 // Also accepted are GMT+H:MM or GMT-H:MM.
89 // Currently not being used
90 //static const UChar gGmt[] = {0x0047, 0x004D, 0x0054, 0x0000}; // "GMT"
91 //static const UChar gGmtPlus[] = {0x0047, 0x004D, 0x0054, 0x002B, 0x0000}; // "GMT+"
92 //static const UChar gGmtMinus[] = {0x0047, 0x004D, 0x0054, 0x002D, 0x0000}; // "GMT-"
93 //static const UChar gDefGmtPat[] = {0x0047, 0x004D, 0x0054, 0x007B, 0x0030, 0x007D, 0x0000}; /* GMT{0} */
94 //static const UChar gDefGmtNegHmsPat[] = {0x002D, 0x0048, 0x0048, 0x003A, 0x006D, 0x006D, 0x003A, 0x0073, 0x0073, 0x0000}; /* -HH:mm:ss */
95 //static const UChar gDefGmtNegHmPat[] = {0x002D, 0x0048, 0x0048, 0x003A, 0x006D, 0x006D, 0x0000}; /* -HH:mm */
96 //static const UChar gDefGmtPosHmsPat[] = {0x002B, 0x0048, 0x0048, 0x003A, 0x006D, 0x006D, 0x003A, 0x0073, 0x0073, 0x0000}; /* +HH:mm:ss */
97 //static const UChar gDefGmtPosHmPat[] = {0x002B, 0x0048, 0x0048, 0x003A, 0x006D, 0x006D, 0x0000}; /* +HH:mm */
98 //static const UChar gUt[] = {0x0055, 0x0054, 0x0000}; // "UT"
99 //static const UChar gUtc[] = {0x0055, 0x0054, 0x0043, 0x0000}; // "UT"
100
101 typedef enum GmtPatSize {
102 kGmtLen = 3,
103 kGmtPatLen = 6,
104 kNegHmsLen = 9,
105 kNegHmLen = 6,
106 kPosHmsLen = 9,
107 kPosHmLen = 6,
108 kUtLen = 2,
109 kUtcLen = 3
110 } GmtPatSize;
111
112 // Stuff needed for numbering system overrides
113
114 typedef enum OvrStrType {
115 kOvrStrDate = 0,
116 kOvrStrTime = 1,
117 kOvrStrBoth = 2
118 } OvrStrType;
119
120 static const UDateFormatField kDateFields[] = {
121 UDAT_YEAR_FIELD,
122 UDAT_MONTH_FIELD,
123 UDAT_DATE_FIELD,
124 UDAT_DAY_OF_YEAR_FIELD,
125 UDAT_DAY_OF_WEEK_IN_MONTH_FIELD,
126 UDAT_WEEK_OF_YEAR_FIELD,
127 UDAT_WEEK_OF_MONTH_FIELD,
128 UDAT_YEAR_WOY_FIELD,
129 UDAT_EXTENDED_YEAR_FIELD,
130 UDAT_JULIAN_DAY_FIELD,
131 UDAT_STANDALONE_DAY_FIELD,
132 UDAT_STANDALONE_MONTH_FIELD,
133 UDAT_QUARTER_FIELD,
134 UDAT_STANDALONE_QUARTER_FIELD,
135 UDAT_YEAR_NAME_FIELD,
136 UDAT_RELATED_YEAR_FIELD };
137 static const int8_t kDateFieldsCount = 16;
138
139 static const UDateFormatField kTimeFields[] = {
140 UDAT_HOUR_OF_DAY1_FIELD,
141 UDAT_HOUR_OF_DAY0_FIELD,
142 UDAT_MINUTE_FIELD,
143 UDAT_SECOND_FIELD,
144 UDAT_FRACTIONAL_SECOND_FIELD,
145 UDAT_HOUR1_FIELD,
146 UDAT_HOUR0_FIELD,
147 UDAT_MILLISECONDS_IN_DAY_FIELD,
148 UDAT_TIMEZONE_RFC_FIELD,
149 UDAT_TIMEZONE_LOCALIZED_GMT_OFFSET_FIELD };
150 static const int8_t kTimeFieldsCount = 10;
151
152
153 // This is a pattern-of-last-resort used when we can't load a usable pattern out
154 // of a resource.
155 static const UChar gDefaultPattern[] =
156 {
157 0x79, 0x79, 0x79, 0x79, 0x4D, 0x4D, 0x64, 0x64, 0x20, 0x68, 0x68, 0x3A, 0x6D, 0x6D, 0x20, 0x61, 0
158 }; /* "yyyyMMdd hh:mm a" */
159
160 // This prefix is designed to NEVER MATCH real text, in order to
161 // suppress the parsing of negative numbers. Adjust as needed (if
162 // this becomes valid Unicode).
163 static const UChar SUPPRESS_NEGATIVE_PREFIX[] = {0xAB00, 0};
164
165 /**
166 * These are the tags we expect to see in normal resource bundle files associated
167 * with a locale.
168 */
169 static const UChar QUOTE = 0x27; // Single quote
170
171 /*
172 * The field range check bias for each UDateFormatField.
173 * The bias is added to the minimum and maximum values
174 * before they are compared to the parsed number.
175 * For example, the calendar stores zero-based month numbers
176 * but the parsed month numbers start at 1, so the bias is 1.
177 *
178 * A value of -1 means that the value is not checked.
179 */
180 static const int32_t gFieldRangeBias[] = {
181 -1, // 'G' - UDAT_ERA_FIELD
182 -1, // 'y' - UDAT_YEAR_FIELD
183 1, // 'M' - UDAT_MONTH_FIELD
184 0, // 'd' - UDAT_DATE_FIELD
185 -1, // 'k' - UDAT_HOUR_OF_DAY1_FIELD
186 -1, // 'H' - UDAT_HOUR_OF_DAY0_FIELD
187 0, // 'm' - UDAT_MINUTE_FIELD
188 0, // 's' - UDAT_SECOND_FIELD
189 -1, // 'S' - UDAT_FRACTIONAL_SECOND_FIELD (0-999?)
190 -1, // 'E' - UDAT_DAY_OF_WEEK_FIELD (1-7?)
191 -1, // 'D' - UDAT_DAY_OF_YEAR_FIELD (1 - 366?)
192 -1, // 'F' - UDAT_DAY_OF_WEEK_IN_MONTH_FIELD (1-5?)
193 -1, // 'w' - UDAT_WEEK_OF_YEAR_FIELD (1-52?)
194 -1, // 'W' - UDAT_WEEK_OF_MONTH_FIELD (1-5?)
195 -1, // 'a' - UDAT_AM_PM_FIELD
196 -1, // 'h' - UDAT_HOUR1_FIELD
197 -1, // 'K' - UDAT_HOUR0_FIELD
198 -1, // 'z' - UDAT_TIMEZONE_FIELD
199 -1, // 'Y' - UDAT_YEAR_WOY_FIELD
200 -1, // 'e' - UDAT_DOW_LOCAL_FIELD
201 -1, // 'u' - UDAT_EXTENDED_YEAR_FIELD
202 -1, // 'g' - UDAT_JULIAN_DAY_FIELD
203 -1, // 'A' - UDAT_MILLISECONDS_IN_DAY_FIELD
204 -1, // 'Z' - UDAT_TIMEZONE_RFC_FIELD
205 -1, // 'v' - UDAT_TIMEZONE_GENERIC_FIELD
206 0, // 'c' - UDAT_STANDALONE_DAY_FIELD
207 1, // 'L' - UDAT_STANDALONE_MONTH_FIELD
208 -1, // 'Q' - UDAT_QUARTER_FIELD (1-4?)
209 -1, // 'q' - UDAT_STANDALONE_QUARTER_FIELD
210 -1, // 'V' - UDAT_TIMEZONE_SPECIAL_FIELD
211 -1, // 'U' - UDAT_YEAR_NAME_FIELD
212 -1, // 'O' - UDAT_TIMEZONE_LOCALIZED_GMT_OFFSET_FIELD
213 -1, // 'X' - UDAT_TIMEZONE_ISO_FIELD
214 -1, // 'x' - UDAT_TIMEZONE_ISO_LOCAL_FIELD
215 -1, // 'r' - UDAT_RELATED_YEAR_FIELD
216 #if UDAT_HAS_PATTERN_CHAR_FOR_TIME_SEPARATOR
217 -1, // ':' - UDAT_TIME_SEPARATOR_FIELD
218 #else
219 -1, // (no pattern character currently) - UDAT_TIME_SEPARATOR_FIELD
220 #endif
221 };
222
223 // When calendar uses hebr numbering (i.e. he@calendar=hebrew),
224 // offset the years within the current millenium down to 1-999
225 static const int32_t HEBREW_CAL_CUR_MILLENIUM_START_YEAR = 5000;
226 static const int32_t HEBREW_CAL_CUR_MILLENIUM_END_YEAR = 6000;
227
228 static UMutex LOCK = U_MUTEX_INITIALIZER;
229
UOBJECT_DEFINE_RTTI_IMPLEMENTATION(SimpleDateFormat)230 UOBJECT_DEFINE_RTTI_IMPLEMENTATION(SimpleDateFormat)
231
232 SimpleDateFormat::NSOverride::~NSOverride() {
233 if (snf != NULL) {
234 snf->removeRef();
235 }
236 }
237
238
free()239 void SimpleDateFormat::NSOverride::free() {
240 NSOverride *cur = this;
241 while (cur) {
242 NSOverride *next = cur->next;
243 delete cur;
244 cur = next;
245 }
246 }
247
248 // no matter what the locale's default number format looked like, we want
249 // to modify it so that it doesn't use thousands separators, doesn't always
250 // show the decimal point, and recognizes integers only when parsing
fixNumberFormatForDates(NumberFormat & nf)251 static void fixNumberFormatForDates(NumberFormat &nf) {
252 nf.setGroupingUsed(FALSE);
253 DecimalFormat* decfmt = dynamic_cast<DecimalFormat*>(&nf);
254 if (decfmt != NULL) {
255 decfmt->setDecimalSeparatorAlwaysShown(FALSE);
256 }
257 nf.setParseIntegerOnly(TRUE);
258 nf.setMinimumFractionDigits(0); // To prevent "Jan 1.00, 1997.00"
259 }
260
createSharedNumberFormat(NumberFormat * nfToAdopt)261 static const SharedNumberFormat *createSharedNumberFormat(
262 NumberFormat *nfToAdopt) {
263 fixNumberFormatForDates(*nfToAdopt);
264 const SharedNumberFormat *result = new SharedNumberFormat(nfToAdopt);
265 if (result == NULL) {
266 delete nfToAdopt;
267 }
268 return result;
269 }
270
createSharedNumberFormat(const Locale & loc,UErrorCode & status)271 static const SharedNumberFormat *createSharedNumberFormat(
272 const Locale &loc, UErrorCode &status) {
273 NumberFormat *nf = NumberFormat::createInstance(loc, status);
274 if (U_FAILURE(status)) {
275 return NULL;
276 }
277 const SharedNumberFormat *result = createSharedNumberFormat(nf);
278 if (result == NULL) {
279 status = U_MEMORY_ALLOCATION_ERROR;
280 }
281 return result;
282 }
283
allocSharedNumberFormatters()284 static const SharedNumberFormat **allocSharedNumberFormatters() {
285 const SharedNumberFormat **result = (const SharedNumberFormat**)
286 uprv_malloc(UDAT_FIELD_COUNT * sizeof(const SharedNumberFormat*));
287 if (result == NULL) {
288 return NULL;
289 }
290 for (int32_t i = 0; i < UDAT_FIELD_COUNT; ++i) {
291 result[i] = NULL;
292 }
293 return result;
294 }
295
freeSharedNumberFormatters(const SharedNumberFormat ** list)296 static void freeSharedNumberFormatters(const SharedNumberFormat ** list) {
297 for (int32_t i = 0; i < UDAT_FIELD_COUNT; ++i) {
298 SharedObject::clearPtr(list[i]);
299 }
300 uprv_free(list);
301 }
302
getNumberFormatByIndex(UDateFormatField index) const303 const NumberFormat *SimpleDateFormat::getNumberFormatByIndex(
304 UDateFormatField index) const {
305 if (fSharedNumberFormatters == NULL ||
306 fSharedNumberFormatters[index] == NULL) {
307 return fNumberFormat;
308 }
309 return &(**fSharedNumberFormatters[index]);
310 }
311
312 class SimpleDateFormatMutableNFNode {
313 public:
314 const NumberFormat *key;
315 NumberFormat *value;
SimpleDateFormatMutableNFNode()316 SimpleDateFormatMutableNFNode()
317 : key(NULL), value(NULL) { }
~SimpleDateFormatMutableNFNode()318 ~SimpleDateFormatMutableNFNode() {
319 delete value;
320 }
321 private:
322 SimpleDateFormatMutableNFNode(const SimpleDateFormatMutableNFNode &);
323 SimpleDateFormatMutableNFNode &operator=(const SimpleDateFormatMutableNFNode &);
324 };
325
326 // Single threaded cache of non const NumberFormats. Designed to be stack
327 // allocated and used for a single format call.
328 class SimpleDateFormatMutableNFs : public UMemory {
329 public:
SimpleDateFormatMutableNFs()330 SimpleDateFormatMutableNFs() {
331 }
332
333 // Returns a non-const clone of nf which can be safely modified.
334 // Subsequent calls with same nf will return the same non-const clone.
335 // This object maintains ownership of all returned non-const
336 // NumberFormat objects. On memory allocation error returns NULL.
337 // Caller must check for NULL return value.
get(const NumberFormat * nf)338 NumberFormat *get(const NumberFormat *nf) {
339 if (nf == NULL) {
340 return NULL;
341 }
342 int32_t idx = 0;
343 while (nodes[idx].value) {
344 if (nf == nodes[idx].key) {
345 return nodes[idx].value;
346 }
347 ++idx;
348 }
349 U_ASSERT(idx < UDAT_FIELD_COUNT);
350 nodes[idx].key = nf;
351 nodes[idx].value = (NumberFormat *) nf->clone();
352 return nodes[idx].value;
353 }
354 private:
355 // +1 extra for sentinel. If each field had its own NumberFormat, this
356 // cache would have to allocate UDAT_FIELD_COUNT mutable versions worst
357 // case.
358 SimpleDateFormatMutableNFNode nodes[UDAT_FIELD_COUNT + 1];
359 SimpleDateFormatMutableNFs(const SimpleDateFormatMutableNFs &);
360 SimpleDateFormatMutableNFs &operator=(const SimpleDateFormatMutableNFs &);
361 };
362
363 //----------------------------------------------------------------------
364
~SimpleDateFormat()365 SimpleDateFormat::~SimpleDateFormat()
366 {
367 delete fSymbols;
368 if (fSharedNumberFormatters) {
369 freeSharedNumberFormatters(fSharedNumberFormatters);
370 }
371 if (fTimeZoneFormat) {
372 delete fTimeZoneFormat;
373 }
374
375 #if !UCONFIG_NO_BREAK_ITERATION
376 delete fCapitalizationBrkIter;
377 #endif
378 }
379
380 //----------------------------------------------------------------------
381
SimpleDateFormat(UErrorCode & status)382 SimpleDateFormat::SimpleDateFormat(UErrorCode& status)
383 : fLocale(Locale::getDefault()),
384 fSymbols(NULL),
385 fTimeZoneFormat(NULL),
386 fSharedNumberFormatters(NULL),
387 fCapitalizationBrkIter(NULL)
388 {
389 initializeBooleanAttributes();
390 construct(kShort, (EStyle) (kShort + kDateOffset), fLocale, status);
391 initializeDefaultCentury();
392 }
393
394 //----------------------------------------------------------------------
395
SimpleDateFormat(const UnicodeString & pattern,UErrorCode & status)396 SimpleDateFormat::SimpleDateFormat(const UnicodeString& pattern,
397 UErrorCode &status)
398 : fPattern(pattern),
399 fLocale(Locale::getDefault()),
400 fSymbols(NULL),
401 fTimeZoneFormat(NULL),
402 fSharedNumberFormatters(NULL),
403 fCapitalizationBrkIter(NULL)
404 {
405 fDateOverride.setToBogus();
406 fTimeOverride.setToBogus();
407 initializeBooleanAttributes();
408 initializeCalendar(NULL,fLocale,status);
409 fSymbols = DateFormatSymbols::createForLocale(fLocale, status);
410 initialize(fLocale, status);
411 initializeDefaultCentury();
412
413 }
414 //----------------------------------------------------------------------
415
SimpleDateFormat(const UnicodeString & pattern,const UnicodeString & override,UErrorCode & status)416 SimpleDateFormat::SimpleDateFormat(const UnicodeString& pattern,
417 const UnicodeString& override,
418 UErrorCode &status)
419 : fPattern(pattern),
420 fLocale(Locale::getDefault()),
421 fSymbols(NULL),
422 fTimeZoneFormat(NULL),
423 fSharedNumberFormatters(NULL),
424 fCapitalizationBrkIter(NULL)
425 {
426 fDateOverride.setTo(override);
427 fTimeOverride.setToBogus();
428 initializeBooleanAttributes();
429 initializeCalendar(NULL,fLocale,status);
430 fSymbols = DateFormatSymbols::createForLocale(fLocale, status);
431 initialize(fLocale, status);
432 initializeDefaultCentury();
433
434 processOverrideString(fLocale,override,kOvrStrBoth,status);
435
436 }
437
438 //----------------------------------------------------------------------
439
SimpleDateFormat(const UnicodeString & pattern,const Locale & locale,UErrorCode & status)440 SimpleDateFormat::SimpleDateFormat(const UnicodeString& pattern,
441 const Locale& locale,
442 UErrorCode& status)
443 : fPattern(pattern),
444 fLocale(locale),
445 fTimeZoneFormat(NULL),
446 fSharedNumberFormatters(NULL),
447 fCapitalizationBrkIter(NULL)
448 {
449
450 fDateOverride.setToBogus();
451 fTimeOverride.setToBogus();
452 initializeBooleanAttributes();
453
454 initializeCalendar(NULL,fLocale,status);
455 fSymbols = DateFormatSymbols::createForLocale(fLocale, status);
456 initialize(fLocale, status);
457 initializeDefaultCentury();
458 }
459
460 //----------------------------------------------------------------------
461
SimpleDateFormat(const UnicodeString & pattern,const UnicodeString & override,const Locale & locale,UErrorCode & status)462 SimpleDateFormat::SimpleDateFormat(const UnicodeString& pattern,
463 const UnicodeString& override,
464 const Locale& locale,
465 UErrorCode& status)
466 : fPattern(pattern),
467 fLocale(locale),
468 fTimeZoneFormat(NULL),
469 fSharedNumberFormatters(NULL),
470 fCapitalizationBrkIter(NULL)
471 {
472
473 fDateOverride.setTo(override);
474 fTimeOverride.setToBogus();
475 initializeBooleanAttributes();
476
477 initializeCalendar(NULL,fLocale,status);
478 fSymbols = DateFormatSymbols::createForLocale(fLocale, status);
479 initialize(fLocale, status);
480 initializeDefaultCentury();
481
482 processOverrideString(locale,override,kOvrStrBoth,status);
483
484 }
485
486 //----------------------------------------------------------------------
487
SimpleDateFormat(const UnicodeString & pattern,DateFormatSymbols * symbolsToAdopt,UErrorCode & status)488 SimpleDateFormat::SimpleDateFormat(const UnicodeString& pattern,
489 DateFormatSymbols* symbolsToAdopt,
490 UErrorCode& status)
491 : fPattern(pattern),
492 fLocale(Locale::getDefault()),
493 fSymbols(symbolsToAdopt),
494 fTimeZoneFormat(NULL),
495 fSharedNumberFormatters(NULL),
496 fCapitalizationBrkIter(NULL)
497 {
498
499 fDateOverride.setToBogus();
500 fTimeOverride.setToBogus();
501 initializeBooleanAttributes();
502
503 initializeCalendar(NULL,fLocale,status);
504 initialize(fLocale, status);
505 initializeDefaultCentury();
506 }
507
508 //----------------------------------------------------------------------
509
SimpleDateFormat(const UnicodeString & pattern,const DateFormatSymbols & symbols,UErrorCode & status)510 SimpleDateFormat::SimpleDateFormat(const UnicodeString& pattern,
511 const DateFormatSymbols& symbols,
512 UErrorCode& status)
513 : fPattern(pattern),
514 fLocale(Locale::getDefault()),
515 fSymbols(new DateFormatSymbols(symbols)),
516 fTimeZoneFormat(NULL),
517 fSharedNumberFormatters(NULL),
518 fCapitalizationBrkIter(NULL)
519 {
520
521 fDateOverride.setToBogus();
522 fTimeOverride.setToBogus();
523 initializeBooleanAttributes();
524
525 initializeCalendar(NULL, fLocale, status);
526 initialize(fLocale, status);
527 initializeDefaultCentury();
528 }
529
530 //----------------------------------------------------------------------
531
532 // Not for public consumption; used by DateFormat
SimpleDateFormat(EStyle timeStyle,EStyle dateStyle,const Locale & locale,UErrorCode & status)533 SimpleDateFormat::SimpleDateFormat(EStyle timeStyle,
534 EStyle dateStyle,
535 const Locale& locale,
536 UErrorCode& status)
537 : fLocale(locale),
538 fSymbols(NULL),
539 fTimeZoneFormat(NULL),
540 fSharedNumberFormatters(NULL),
541 fCapitalizationBrkIter(NULL)
542 {
543 initializeBooleanAttributes();
544 construct(timeStyle, dateStyle, fLocale, status);
545 if(U_SUCCESS(status)) {
546 initializeDefaultCentury();
547 }
548 }
549
550 //----------------------------------------------------------------------
551
552 /**
553 * Not for public consumption; used by DateFormat. This constructor
554 * never fails. If the resource data is not available, it uses the
555 * the last resort symbols.
556 */
SimpleDateFormat(const Locale & locale,UErrorCode & status)557 SimpleDateFormat::SimpleDateFormat(const Locale& locale,
558 UErrorCode& status)
559 : fPattern(gDefaultPattern),
560 fLocale(locale),
561 fSymbols(NULL),
562 fTimeZoneFormat(NULL),
563 fSharedNumberFormatters(NULL),
564 fCapitalizationBrkIter(NULL)
565 {
566 if (U_FAILURE(status)) return;
567 initializeBooleanAttributes();
568 initializeCalendar(NULL, fLocale, status);
569 fSymbols = DateFormatSymbols::createForLocale(fLocale, status);
570 if (U_FAILURE(status))
571 {
572 status = U_ZERO_ERROR;
573 delete fSymbols;
574 // This constructor doesn't fail; it uses last resort data
575 fSymbols = new DateFormatSymbols(status);
576 /* test for NULL */
577 if (fSymbols == 0) {
578 status = U_MEMORY_ALLOCATION_ERROR;
579 return;
580 }
581 }
582
583 fDateOverride.setToBogus();
584 fTimeOverride.setToBogus();
585
586 initialize(fLocale, status);
587 if(U_SUCCESS(status)) {
588 initializeDefaultCentury();
589 }
590 }
591
592 //----------------------------------------------------------------------
593
SimpleDateFormat(const SimpleDateFormat & other)594 SimpleDateFormat::SimpleDateFormat(const SimpleDateFormat& other)
595 : DateFormat(other),
596 fLocale(other.fLocale),
597 fSymbols(NULL),
598 fTimeZoneFormat(NULL),
599 fSharedNumberFormatters(NULL),
600 fCapitalizationBrkIter(NULL)
601 {
602 initializeBooleanAttributes();
603 *this = other;
604 }
605
606 //----------------------------------------------------------------------
607
operator =(const SimpleDateFormat & other)608 SimpleDateFormat& SimpleDateFormat::operator=(const SimpleDateFormat& other)
609 {
610 if (this == &other) {
611 return *this;
612 }
613 DateFormat::operator=(other);
614 fDateOverride = other.fDateOverride;
615 fTimeOverride = other.fTimeOverride;
616
617 delete fSymbols;
618 fSymbols = NULL;
619
620 if (other.fSymbols)
621 fSymbols = new DateFormatSymbols(*other.fSymbols);
622
623 fDefaultCenturyStart = other.fDefaultCenturyStart;
624 fDefaultCenturyStartYear = other.fDefaultCenturyStartYear;
625 fHaveDefaultCentury = other.fHaveDefaultCentury;
626
627 fPattern = other.fPattern;
628 fHasMinute = other.fHasMinute;
629 fHasSecond = other.fHasSecond;
630
631 // TimeZoneFormat in ICU4C only depends on a locale for now
632 if (fLocale != other.fLocale) {
633 delete fTimeZoneFormat;
634 fTimeZoneFormat = NULL; // forces lazy instantiation with the other locale
635 fLocale = other.fLocale;
636 }
637
638 #if !UCONFIG_NO_BREAK_ITERATION
639 if (other.fCapitalizationBrkIter != NULL) {
640 fCapitalizationBrkIter = (other.fCapitalizationBrkIter)->clone();
641 }
642 #endif
643
644 if (fSharedNumberFormatters != NULL) {
645 freeSharedNumberFormatters(fSharedNumberFormatters);
646 fSharedNumberFormatters = NULL;
647 }
648 if (other.fSharedNumberFormatters != NULL) {
649 fSharedNumberFormatters = allocSharedNumberFormatters();
650 if (fSharedNumberFormatters) {
651 for (int32_t i = 0; i < UDAT_FIELD_COUNT; ++i) {
652 SharedObject::copyPtr(
653 other.fSharedNumberFormatters[i],
654 fSharedNumberFormatters[i]);
655 }
656 }
657 }
658
659 return *this;
660 }
661
662 //----------------------------------------------------------------------
663
664 Format*
clone() const665 SimpleDateFormat::clone() const
666 {
667 return new SimpleDateFormat(*this);
668 }
669
670 //----------------------------------------------------------------------
671
672 UBool
operator ==(const Format & other) const673 SimpleDateFormat::operator==(const Format& other) const
674 {
675 if (DateFormat::operator==(other)) {
676 // The DateFormat::operator== check for fCapitalizationContext equality above
677 // is sufficient to check equality of all derived context-related data.
678 // DateFormat::operator== guarantees following cast is safe
679 SimpleDateFormat* that = (SimpleDateFormat*)&other;
680 return (fPattern == that->fPattern &&
681 fSymbols != NULL && // Check for pathological object
682 that->fSymbols != NULL && // Check for pathological object
683 *fSymbols == *that->fSymbols &&
684 fHaveDefaultCentury == that->fHaveDefaultCentury &&
685 fDefaultCenturyStart == that->fDefaultCenturyStart);
686 }
687 return FALSE;
688 }
689
690 //----------------------------------------------------------------------
691
construct(EStyle timeStyle,EStyle dateStyle,const Locale & locale,UErrorCode & status)692 void SimpleDateFormat::construct(EStyle timeStyle,
693 EStyle dateStyle,
694 const Locale& locale,
695 UErrorCode& status)
696 {
697 // called by several constructors to load pattern data from the resources
698 if (U_FAILURE(status)) return;
699
700 // We will need the calendar to know what type of symbols to load.
701 initializeCalendar(NULL, locale, status);
702 if (U_FAILURE(status)) return;
703
704 // Load date time patterns directly from resources.
705 const char* cType = fCalendar ? fCalendar->getType() : NULL;
706 LocalUResourceBundlePointer bundle(ures_open(NULL, locale.getBaseName(), &status));
707 if (U_FAILURE(status)) return;
708
709 UBool cTypeIsGregorian = TRUE;
710 LocalUResourceBundlePointer dateTimePatterns;
711 if (cType != NULL && uprv_strcmp(cType, "gregorian") != 0) {
712 CharString resourcePath("calendar/", status);
713 resourcePath.append(cType, status).append("/DateTimePatterns", status);
714 dateTimePatterns.adoptInstead(
715 ures_getByKeyWithFallback(bundle.getAlias(), resourcePath.data(),
716 (UResourceBundle*)NULL, &status));
717 cTypeIsGregorian = FALSE;
718 }
719
720 // Check for "gregorian" fallback.
721 if (cTypeIsGregorian || status == U_MISSING_RESOURCE_ERROR) {
722 status = U_ZERO_ERROR;
723 dateTimePatterns.adoptInstead(
724 ures_getByKeyWithFallback(bundle.getAlias(),
725 "calendar/gregorian/DateTimePatterns",
726 (UResourceBundle*)NULL, &status));
727 }
728 if (U_FAILURE(status)) return;
729
730 LocalUResourceBundlePointer currentBundle;
731
732 if (ures_getSize(dateTimePatterns.getAlias()) <= kDateTime)
733 {
734 status = U_INVALID_FORMAT_ERROR;
735 return;
736 }
737
738 setLocaleIDs(ures_getLocaleByType(dateTimePatterns.getAlias(), ULOC_VALID_LOCALE, &status),
739 ures_getLocaleByType(dateTimePatterns.getAlias(), ULOC_ACTUAL_LOCALE, &status));
740
741 // create a symbols object from the locale
742 fSymbols = DateFormatSymbols::createForLocale(locale, status);
743 if (U_FAILURE(status)) return;
744 /* test for NULL */
745 if (fSymbols == 0) {
746 status = U_MEMORY_ALLOCATION_ERROR;
747 return;
748 }
749
750 const UChar *resStr,*ovrStr;
751 int32_t resStrLen,ovrStrLen = 0;
752 fDateOverride.setToBogus();
753 fTimeOverride.setToBogus();
754
755 // if the pattern should include both date and time information, use the date/time
756 // pattern string as a guide to tell use how to glue together the appropriate date
757 // and time pattern strings.
758 if ((timeStyle != kNone) && (dateStyle != kNone))
759 {
760 currentBundle.adoptInstead(
761 ures_getByIndex(dateTimePatterns.getAlias(), (int32_t)timeStyle, NULL, &status));
762 if (U_FAILURE(status)) {
763 status = U_INVALID_FORMAT_ERROR;
764 return;
765 }
766 switch (ures_getType(currentBundle.getAlias())) {
767 case URES_STRING: {
768 resStr = ures_getString(currentBundle.getAlias(), &resStrLen, &status);
769 break;
770 }
771 case URES_ARRAY: {
772 resStr = ures_getStringByIndex(currentBundle.getAlias(), 0, &resStrLen, &status);
773 ovrStr = ures_getStringByIndex(currentBundle.getAlias(), 1, &ovrStrLen, &status);
774 fTimeOverride.setTo(TRUE, ovrStr, ovrStrLen);
775 break;
776 }
777 default: {
778 status = U_INVALID_FORMAT_ERROR;
779 return;
780 }
781 }
782
783 UnicodeString tempus1(TRUE, resStr, resStrLen);
784
785 currentBundle.adoptInstead(
786 ures_getByIndex(dateTimePatterns.getAlias(), (int32_t)dateStyle, NULL, &status));
787 if (U_FAILURE(status)) {
788 status = U_INVALID_FORMAT_ERROR;
789 return;
790 }
791 switch (ures_getType(currentBundle.getAlias())) {
792 case URES_STRING: {
793 resStr = ures_getString(currentBundle.getAlias(), &resStrLen, &status);
794 break;
795 }
796 case URES_ARRAY: {
797 resStr = ures_getStringByIndex(currentBundle.getAlias(), 0, &resStrLen, &status);
798 ovrStr = ures_getStringByIndex(currentBundle.getAlias(), 1, &ovrStrLen, &status);
799 fDateOverride.setTo(TRUE, ovrStr, ovrStrLen);
800 break;
801 }
802 default: {
803 status = U_INVALID_FORMAT_ERROR;
804 return;
805 }
806 }
807
808 UnicodeString tempus2(TRUE, resStr, resStrLen);
809
810 int32_t glueIndex = kDateTime;
811 int32_t patternsSize = ures_getSize(dateTimePatterns.getAlias());
812 if (patternsSize >= (kDateTimeOffset + kShort + 1)) {
813 // Get proper date time format
814 glueIndex = (int32_t)(kDateTimeOffset + (dateStyle - kDateOffset));
815 }
816
817 resStr = ures_getStringByIndex(dateTimePatterns.getAlias(), glueIndex, &resStrLen, &status);
818 SimpleFormatter(UnicodeString(TRUE, resStr, resStrLen), 2, 2, status).
819 format(tempus1, tempus2, fPattern, status);
820 }
821 // if the pattern includes just time data or just date date, load the appropriate
822 // pattern string from the resources
823 // setTo() - see DateFormatSymbols::assignArray comments
824 else if (timeStyle != kNone) {
825 currentBundle.adoptInstead(
826 ures_getByIndex(dateTimePatterns.getAlias(), (int32_t)timeStyle, NULL, &status));
827 if (U_FAILURE(status)) {
828 status = U_INVALID_FORMAT_ERROR;
829 return;
830 }
831 switch (ures_getType(currentBundle.getAlias())) {
832 case URES_STRING: {
833 resStr = ures_getString(currentBundle.getAlias(), &resStrLen, &status);
834 break;
835 }
836 case URES_ARRAY: {
837 resStr = ures_getStringByIndex(currentBundle.getAlias(), 0, &resStrLen, &status);
838 ovrStr = ures_getStringByIndex(currentBundle.getAlias(), 1, &ovrStrLen, &status);
839 fDateOverride.setTo(TRUE, ovrStr, ovrStrLen);
840 break;
841 }
842 default: {
843 status = U_INVALID_FORMAT_ERROR;
844 return;
845 }
846 }
847 fPattern.setTo(TRUE, resStr, resStrLen);
848 }
849 else if (dateStyle != kNone) {
850 currentBundle.adoptInstead(
851 ures_getByIndex(dateTimePatterns.getAlias(), (int32_t)dateStyle, NULL, &status));
852 if (U_FAILURE(status)) {
853 status = U_INVALID_FORMAT_ERROR;
854 return;
855 }
856 switch (ures_getType(currentBundle.getAlias())) {
857 case URES_STRING: {
858 resStr = ures_getString(currentBundle.getAlias(), &resStrLen, &status);
859 break;
860 }
861 case URES_ARRAY: {
862 resStr = ures_getStringByIndex(currentBundle.getAlias(), 0, &resStrLen, &status);
863 ovrStr = ures_getStringByIndex(currentBundle.getAlias(), 1, &ovrStrLen, &status);
864 fDateOverride.setTo(TRUE, ovrStr, ovrStrLen);
865 break;
866 }
867 default: {
868 status = U_INVALID_FORMAT_ERROR;
869 return;
870 }
871 }
872 fPattern.setTo(TRUE, resStr, resStrLen);
873 }
874
875 // and if it includes _neither_, that's an error
876 else
877 status = U_INVALID_FORMAT_ERROR;
878
879 // finally, finish initializing by creating a Calendar and a NumberFormat
880 initialize(locale, status);
881 }
882
883 //----------------------------------------------------------------------
884
885 Calendar*
initializeCalendar(TimeZone * adoptZone,const Locale & locale,UErrorCode & status)886 SimpleDateFormat::initializeCalendar(TimeZone* adoptZone, const Locale& locale, UErrorCode& status)
887 {
888 if(!U_FAILURE(status)) {
889 fCalendar = Calendar::createInstance(adoptZone?adoptZone:TimeZone::createDefault(), locale, status);
890 }
891 return fCalendar;
892 }
893
894 void
initialize(const Locale & locale,UErrorCode & status)895 SimpleDateFormat::initialize(const Locale& locale,
896 UErrorCode& status)
897 {
898 if (U_FAILURE(status)) return;
899
900 // We don't need to check that the row count is >= 1, since all 2d arrays have at
901 // least one row
902 fNumberFormat = NumberFormat::createInstance(locale, status);
903 if (fNumberFormat != NULL && U_SUCCESS(status))
904 {
905 fixNumberFormatForDates(*fNumberFormat);
906 //fNumberFormat->setLenient(TRUE); // Java uses a custom DateNumberFormat to format/parse
907
908 initNumberFormatters(locale,status);
909
910 }
911 else if (U_SUCCESS(status))
912 {
913 status = U_MISSING_RESOURCE_ERROR;
914 }
915
916 parsePattern();
917 }
918
919 /* Initialize the fields we use to disambiguate ambiguous years. Separate
920 * so we can call it from readObject().
921 */
initializeDefaultCentury()922 void SimpleDateFormat::initializeDefaultCentury()
923 {
924 if(fCalendar) {
925 fHaveDefaultCentury = fCalendar->haveDefaultCentury();
926 if(fHaveDefaultCentury) {
927 fDefaultCenturyStart = fCalendar->defaultCenturyStart();
928 fDefaultCenturyStartYear = fCalendar->defaultCenturyStartYear();
929 } else {
930 fDefaultCenturyStart = DBL_MIN;
931 fDefaultCenturyStartYear = -1;
932 }
933 }
934 }
935
936 /*
937 * Initialize the boolean attributes. Separate so we can call it from all constructors.
938 */
initializeBooleanAttributes()939 void SimpleDateFormat::initializeBooleanAttributes()
940 {
941 UErrorCode status = U_ZERO_ERROR;
942
943 setBooleanAttribute(UDAT_PARSE_ALLOW_WHITESPACE, true, status);
944 setBooleanAttribute(UDAT_PARSE_ALLOW_NUMERIC, true, status);
945 setBooleanAttribute(UDAT_PARSE_PARTIAL_LITERAL_MATCH, true, status);
946 setBooleanAttribute(UDAT_PARSE_MULTIPLE_PATTERNS_FOR_MATCH, true, status);
947 }
948
949 /* Define one-century window into which to disambiguate dates using
950 * two-digit years. Make public in JDK 1.2.
951 */
parseAmbiguousDatesAsAfter(UDate startDate,UErrorCode & status)952 void SimpleDateFormat::parseAmbiguousDatesAsAfter(UDate startDate, UErrorCode& status)
953 {
954 if(U_FAILURE(status)) {
955 return;
956 }
957 if(!fCalendar) {
958 status = U_ILLEGAL_ARGUMENT_ERROR;
959 return;
960 }
961
962 fCalendar->setTime(startDate, status);
963 if(U_SUCCESS(status)) {
964 fHaveDefaultCentury = TRUE;
965 fDefaultCenturyStart = startDate;
966 fDefaultCenturyStartYear = fCalendar->get(UCAL_YEAR, status);
967 }
968 }
969
970 //----------------------------------------------------------------------
971
972 UnicodeString&
format(Calendar & cal,UnicodeString & appendTo,FieldPosition & pos) const973 SimpleDateFormat::format(Calendar& cal, UnicodeString& appendTo, FieldPosition& pos) const
974 {
975 UErrorCode status = U_ZERO_ERROR;
976 FieldPositionOnlyHandler handler(pos);
977 return _format(cal, appendTo, handler, status);
978 }
979
980 //----------------------------------------------------------------------
981
982 UnicodeString&
format(Calendar & cal,UnicodeString & appendTo,FieldPositionIterator * posIter,UErrorCode & status) const983 SimpleDateFormat::format(Calendar& cal, UnicodeString& appendTo,
984 FieldPositionIterator* posIter, UErrorCode& status) const
985 {
986 FieldPositionIteratorHandler handler(posIter, status);
987 return _format(cal, appendTo, handler, status);
988 }
989
990 //----------------------------------------------------------------------
991
992 UnicodeString&
_format(Calendar & cal,UnicodeString & appendTo,FieldPositionHandler & handler,UErrorCode & status) const993 SimpleDateFormat::_format(Calendar& cal, UnicodeString& appendTo,
994 FieldPositionHandler& handler, UErrorCode& status) const
995 {
996 if ( U_FAILURE(status) ) {
997 return appendTo;
998 }
999 Calendar* workCal = &cal;
1000 Calendar* calClone = NULL;
1001 if (&cal != fCalendar && uprv_strcmp(cal.getType(), fCalendar->getType()) != 0) {
1002 // Different calendar type
1003 // We use the time and time zone from the input calendar, but
1004 // do not use the input calendar for field calculation.
1005 calClone = fCalendar->clone();
1006 if (calClone != NULL) {
1007 UDate t = cal.getTime(status);
1008 calClone->setTime(t, status);
1009 calClone->setTimeZone(cal.getTimeZone());
1010 workCal = calClone;
1011 } else {
1012 status = U_MEMORY_ALLOCATION_ERROR;
1013 return appendTo;
1014 }
1015 }
1016
1017 UBool inQuote = FALSE;
1018 UChar prevCh = 0;
1019 int32_t count = 0;
1020 int32_t fieldNum = 0;
1021 UDisplayContext capitalizationContext = getContext(UDISPCTX_TYPE_CAPITALIZATION, status);
1022
1023 // Create temporary cache of mutable number format objects. This way
1024 // subFormat won't have to clone the const NumberFormat for each field.
1025 // if several fields share the same NumberFormat, which will almost
1026 // always be the case, this is a big save.
1027 SimpleDateFormatMutableNFs mutableNFs;
1028 // loop through the pattern string character by character
1029 for (int32_t i = 0; i < fPattern.length() && U_SUCCESS(status); ++i) {
1030 UChar ch = fPattern[i];
1031
1032 // Use subFormat() to format a repeated pattern character
1033 // when a different pattern or non-pattern character is seen
1034 if (ch != prevCh && count > 0) {
1035 subFormat(appendTo, prevCh, count, capitalizationContext, fieldNum++, handler, *workCal, mutableNFs, status);
1036 count = 0;
1037 }
1038 if (ch == QUOTE) {
1039 // Consecutive single quotes are a single quote literal,
1040 // either outside of quotes or between quotes
1041 if ((i+1) < fPattern.length() && fPattern[i+1] == QUOTE) {
1042 appendTo += (UChar)QUOTE;
1043 ++i;
1044 } else {
1045 inQuote = ! inQuote;
1046 }
1047 }
1048 else if (!inQuote && isSyntaxChar(ch)) {
1049 // ch is a date-time pattern character to be interpreted
1050 // by subFormat(); count the number of times it is repeated
1051 prevCh = ch;
1052 ++count;
1053 }
1054 else {
1055 // Append quoted characters and unquoted non-pattern characters
1056 appendTo += ch;
1057 }
1058 }
1059
1060 // Format the last item in the pattern, if any
1061 if (count > 0) {
1062 subFormat(appendTo, prevCh, count, capitalizationContext, fieldNum++, handler, *workCal, mutableNFs, status);
1063 }
1064
1065 if (calClone != NULL) {
1066 delete calClone;
1067 }
1068
1069 return appendTo;
1070 }
1071
1072 //----------------------------------------------------------------------
1073
1074 /* Map calendar field into calendar field level.
1075 * the larger the level, the smaller the field unit.
1076 * For example, UCAL_ERA level is 0, UCAL_YEAR level is 10,
1077 * UCAL_MONTH level is 20.
1078 * NOTE: if new fields adds in, the table needs to update.
1079 */
1080 const int32_t
1081 SimpleDateFormat::fgCalendarFieldToLevel[] =
1082 {
1083 /*GyM*/ 0, 10, 20,
1084 /*wW*/ 20, 30,
1085 /*dDEF*/ 30, 20, 30, 30,
1086 /*ahHm*/ 40, 50, 50, 60,
1087 /*sS*/ 70, 80,
1088 /*z?Y*/ 0, 0, 10,
1089 /*eug*/ 30, 10, 0,
1090 /*A?.*/ 40, 0, 0
1091 };
1092
getLevelFromChar(UChar ch)1093 int32_t SimpleDateFormat::getLevelFromChar(UChar ch) {
1094 // Map date field LETTER into calendar field level.
1095 // the larger the level, the smaller the field unit.
1096 // NOTE: if new fields adds in, the table needs to update.
1097 static const int32_t mapCharToLevel[] = {
1098 -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
1099 //
1100 -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
1101 // ! " # $ % & ' ( ) * + , - . /
1102 -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
1103 #if UDAT_HAS_PATTERN_CHAR_FOR_TIME_SEPARATOR
1104 // 0 1 2 3 4 5 6 7 8 9 : ; < = > ?
1105 -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, 0, -1, -1, -1, -1, -1,
1106 #else
1107 // 0 1 2 3 4 5 6 7 8 9 : ; < = > ?
1108 -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
1109 #endif
1110 // @ A B C D E F G H I J K L M N O
1111 -1, 40, -1, -1, 20, 30, 30, 0, 50, -1, -1, 50, 20, 20, -1, 0,
1112 // P Q R S T U V W X Y Z [ \ ] ^ _
1113 -1, 20, -1, 80, -1, 10, 0, 30, 0, 10, 0, -1, -1, -1, -1, -1,
1114 // ` a b c d e f g h i j k l m n o
1115 -1, 40, -1, 30, 30, 30, -1, 0, 50, -1, -1, 50, 0, 60, -1, -1,
1116 // p q r s t u v w x y z { | } ~
1117 -1, 20, 10, 70, -1, 10, 0, 20, 0, 10, 0, -1, -1, -1, -1, -1
1118 };
1119
1120 return ch < UPRV_LENGTHOF(mapCharToLevel) ? mapCharToLevel[ch] : -1;
1121 }
1122
isSyntaxChar(UChar ch)1123 UBool SimpleDateFormat::isSyntaxChar(UChar ch) {
1124 static const UBool mapCharToIsSyntax[] = {
1125 //
1126 FALSE, FALSE, FALSE, FALSE, FALSE, FALSE, FALSE, FALSE,
1127 //
1128 FALSE, FALSE, FALSE, FALSE, FALSE, FALSE, FALSE, FALSE,
1129 //
1130 FALSE, FALSE, FALSE, FALSE, FALSE, FALSE, FALSE, FALSE,
1131 //
1132 FALSE, FALSE, FALSE, FALSE, FALSE, FALSE, FALSE, FALSE,
1133 // ! " # $ % & '
1134 FALSE, FALSE, FALSE, FALSE, FALSE, FALSE, FALSE, FALSE,
1135 // ( ) * + , - . /
1136 FALSE, FALSE, FALSE, FALSE, FALSE, FALSE, FALSE, FALSE,
1137 // 0 1 2 3 4 5 6 7
1138 FALSE, FALSE, FALSE, FALSE, FALSE, FALSE, FALSE, FALSE,
1139 #if UDAT_HAS_PATTERN_CHAR_FOR_TIME_SEPARATOR
1140 // 8 9 : ; < = > ?
1141 FALSE, FALSE, TRUE, FALSE, FALSE, FALSE, FALSE, FALSE,
1142 #else
1143 // 8 9 : ; < = > ?
1144 FALSE, FALSE, FALSE, FALSE, FALSE, FALSE, FALSE, FALSE,
1145 #endif
1146 // @ A B C D E F G
1147 FALSE, TRUE, TRUE, TRUE, TRUE, TRUE, TRUE, TRUE,
1148 // H I J K L M N O
1149 TRUE, TRUE, TRUE, TRUE, TRUE, TRUE, TRUE, TRUE,
1150 // P Q R S T U V W
1151 TRUE, TRUE, TRUE, TRUE, TRUE, TRUE, TRUE, TRUE,
1152 // X Y Z [ \ ] ^ _
1153 TRUE, TRUE, TRUE, FALSE, FALSE, FALSE, FALSE, FALSE,
1154 // ` a b c d e f g
1155 FALSE, TRUE, TRUE, TRUE, TRUE, TRUE, TRUE, TRUE,
1156 // h i j k l m n o
1157 TRUE, TRUE, TRUE, TRUE, TRUE, TRUE, TRUE, TRUE,
1158 // p q r s t u v w
1159 TRUE, TRUE, TRUE, TRUE, TRUE, TRUE, TRUE, TRUE,
1160 // x y z { | } ~
1161 TRUE, TRUE, TRUE, FALSE, FALSE, FALSE, FALSE, FALSE
1162 };
1163
1164 return ch < UPRV_LENGTHOF(mapCharToIsSyntax) ? mapCharToIsSyntax[ch] : FALSE;
1165 }
1166
1167 // Map index into pattern character string to Calendar field number.
1168 const UCalendarDateFields
1169 SimpleDateFormat::fgPatternIndexToCalendarField[] =
1170 {
1171 /*GyM*/ UCAL_ERA, UCAL_YEAR, UCAL_MONTH,
1172 /*dkH*/ UCAL_DATE, UCAL_HOUR_OF_DAY, UCAL_HOUR_OF_DAY,
1173 /*msS*/ UCAL_MINUTE, UCAL_SECOND, UCAL_MILLISECOND,
1174 /*EDF*/ UCAL_DAY_OF_WEEK, UCAL_DAY_OF_YEAR, UCAL_DAY_OF_WEEK_IN_MONTH,
1175 /*wWa*/ UCAL_WEEK_OF_YEAR, UCAL_WEEK_OF_MONTH, UCAL_AM_PM,
1176 /*hKz*/ UCAL_HOUR, UCAL_HOUR, UCAL_ZONE_OFFSET,
1177 /*Yeu*/ UCAL_YEAR_WOY, UCAL_DOW_LOCAL, UCAL_EXTENDED_YEAR,
1178 /*gAZ*/ UCAL_JULIAN_DAY, UCAL_MILLISECONDS_IN_DAY, UCAL_ZONE_OFFSET,
1179 /*v*/ UCAL_ZONE_OFFSET,
1180 /*c*/ UCAL_DOW_LOCAL,
1181 /*L*/ UCAL_MONTH,
1182 /*Q*/ UCAL_MONTH,
1183 /*q*/ UCAL_MONTH,
1184 /*V*/ UCAL_ZONE_OFFSET,
1185 /*U*/ UCAL_YEAR,
1186 /*O*/ UCAL_ZONE_OFFSET,
1187 /*Xx*/ UCAL_ZONE_OFFSET, UCAL_ZONE_OFFSET,
1188 /*r*/ UCAL_EXTENDED_YEAR,
1189 /*bB*/ UCAL_FIELD_COUNT, UCAL_FIELD_COUNT, // no mappings to calendar fields
1190 #if UDAT_HAS_PATTERN_CHAR_FOR_TIME_SEPARATOR
1191 /*:*/ UCAL_FIELD_COUNT, /* => no useful mapping to any calendar field */
1192 #else
1193 /*no pattern char for UDAT_TIME_SEPARATOR_FIELD*/ UCAL_FIELD_COUNT, /* => no useful mapping to any calendar field */
1194 #endif
1195 };
1196
1197 // Map index into pattern character string to DateFormat field number
1198 const UDateFormatField
1199 SimpleDateFormat::fgPatternIndexToDateFormatField[] = {
1200 /*GyM*/ UDAT_ERA_FIELD, UDAT_YEAR_FIELD, UDAT_MONTH_FIELD,
1201 /*dkH*/ UDAT_DATE_FIELD, UDAT_HOUR_OF_DAY1_FIELD, UDAT_HOUR_OF_DAY0_FIELD,
1202 /*msS*/ UDAT_MINUTE_FIELD, UDAT_SECOND_FIELD, UDAT_FRACTIONAL_SECOND_FIELD,
1203 /*EDF*/ UDAT_DAY_OF_WEEK_FIELD, UDAT_DAY_OF_YEAR_FIELD, UDAT_DAY_OF_WEEK_IN_MONTH_FIELD,
1204 /*wWa*/ UDAT_WEEK_OF_YEAR_FIELD, UDAT_WEEK_OF_MONTH_FIELD, UDAT_AM_PM_FIELD,
1205 /*hKz*/ UDAT_HOUR1_FIELD, UDAT_HOUR0_FIELD, UDAT_TIMEZONE_FIELD,
1206 /*Yeu*/ UDAT_YEAR_WOY_FIELD, UDAT_DOW_LOCAL_FIELD, UDAT_EXTENDED_YEAR_FIELD,
1207 /*gAZ*/ UDAT_JULIAN_DAY_FIELD, UDAT_MILLISECONDS_IN_DAY_FIELD, UDAT_TIMEZONE_RFC_FIELD,
1208 /*v*/ UDAT_TIMEZONE_GENERIC_FIELD,
1209 /*c*/ UDAT_STANDALONE_DAY_FIELD,
1210 /*L*/ UDAT_STANDALONE_MONTH_FIELD,
1211 /*Q*/ UDAT_QUARTER_FIELD,
1212 /*q*/ UDAT_STANDALONE_QUARTER_FIELD,
1213 /*V*/ UDAT_TIMEZONE_SPECIAL_FIELD,
1214 /*U*/ UDAT_YEAR_NAME_FIELD,
1215 /*O*/ UDAT_TIMEZONE_LOCALIZED_GMT_OFFSET_FIELD,
1216 /*Xx*/ UDAT_TIMEZONE_ISO_FIELD, UDAT_TIMEZONE_ISO_LOCAL_FIELD,
1217 /*r*/ UDAT_RELATED_YEAR_FIELD,
1218 /*bB*/ UDAT_AM_PM_MIDNIGHT_NOON_FIELD, UDAT_FLEXIBLE_DAY_PERIOD_FIELD,
1219 #if UDAT_HAS_PATTERN_CHAR_FOR_TIME_SEPARATOR
1220 /*:*/ UDAT_TIME_SEPARATOR_FIELD,
1221 #else
1222 /*no pattern char for UDAT_TIME_SEPARATOR_FIELD*/ UDAT_TIME_SEPARATOR_FIELD,
1223 #endif
1224 };
1225
1226 //----------------------------------------------------------------------
1227
1228 /**
1229 * Append symbols[value] to dst. Make sure the array index is not out
1230 * of bounds.
1231 */
1232 static inline void
_appendSymbol(UnicodeString & dst,int32_t value,const UnicodeString * symbols,int32_t symbolsCount)1233 _appendSymbol(UnicodeString& dst,
1234 int32_t value,
1235 const UnicodeString* symbols,
1236 int32_t symbolsCount) {
1237 U_ASSERT(0 <= value && value < symbolsCount);
1238 if (0 <= value && value < symbolsCount) {
1239 dst += symbols[value];
1240 }
1241 }
1242
1243 static inline void
_appendSymbolWithMonthPattern(UnicodeString & dst,int32_t value,const UnicodeString * symbols,int32_t symbolsCount,const UnicodeString * monthPattern,UErrorCode & status)1244 _appendSymbolWithMonthPattern(UnicodeString& dst, int32_t value, const UnicodeString* symbols, int32_t symbolsCount,
1245 const UnicodeString* monthPattern, UErrorCode& status) {
1246 U_ASSERT(0 <= value && value < symbolsCount);
1247 if (0 <= value && value < symbolsCount) {
1248 if (monthPattern == NULL) {
1249 dst += symbols[value];
1250 } else {
1251 SimpleFormatter(*monthPattern, 1, 1, status).format(symbols[value], dst, status);
1252 }
1253 }
1254 }
1255
1256 //----------------------------------------------------------------------
1257 void
initNumberFormatters(const Locale & locale,UErrorCode & status)1258 SimpleDateFormat::initNumberFormatters(const Locale &locale,UErrorCode &status) {
1259 if (U_FAILURE(status)) {
1260 return;
1261 }
1262 if ( fDateOverride.isBogus() && fTimeOverride.isBogus() ) {
1263 return;
1264 }
1265 umtx_lock(&LOCK);
1266 if (fSharedNumberFormatters == NULL) {
1267 fSharedNumberFormatters = allocSharedNumberFormatters();
1268 if (fSharedNumberFormatters == NULL) {
1269 status = U_MEMORY_ALLOCATION_ERROR;
1270 }
1271 }
1272 umtx_unlock(&LOCK);
1273
1274 if (U_FAILURE(status)) {
1275 return;
1276 }
1277
1278 processOverrideString(locale,fDateOverride,kOvrStrDate,status);
1279 processOverrideString(locale,fTimeOverride,kOvrStrTime,status);
1280 }
1281
1282 void
processOverrideString(const Locale & locale,const UnicodeString & str,int8_t type,UErrorCode & status)1283 SimpleDateFormat::processOverrideString(const Locale &locale, const UnicodeString &str, int8_t type, UErrorCode &status) {
1284 if (str.isBogus() || U_FAILURE(status)) {
1285 return;
1286 }
1287
1288 int32_t start = 0;
1289 int32_t len;
1290 UnicodeString nsName;
1291 UnicodeString ovrField;
1292 UBool moreToProcess = TRUE;
1293 NSOverride *overrideList = NULL;
1294
1295 while (moreToProcess) {
1296 int32_t delimiterPosition = str.indexOf((UChar)ULOC_KEYWORD_ITEM_SEPARATOR_UNICODE,start);
1297 if (delimiterPosition == -1) {
1298 moreToProcess = FALSE;
1299 len = str.length() - start;
1300 } else {
1301 len = delimiterPosition - start;
1302 }
1303 UnicodeString currentString(str,start,len);
1304 int32_t equalSignPosition = currentString.indexOf((UChar)ULOC_KEYWORD_ASSIGN_UNICODE,0);
1305 if (equalSignPosition == -1) { // Simple override string such as "hebrew"
1306 nsName.setTo(currentString);
1307 ovrField.setToBogus();
1308 } else { // Field specific override string such as "y=hebrew"
1309 nsName.setTo(currentString,equalSignPosition+1);
1310 ovrField.setTo(currentString,0,1); // We just need the first character.
1311 }
1312
1313 int32_t nsNameHash = nsName.hashCode();
1314 // See if the numbering system is in the override list, if not, then add it.
1315 NSOverride *cur = overrideList;
1316 const SharedNumberFormat *snf = NULL;
1317 UBool found = FALSE;
1318 while ( cur && !found ) {
1319 if ( cur->hash == nsNameHash ) {
1320 snf = cur->snf;
1321 found = TRUE;
1322 }
1323 cur = cur->next;
1324 }
1325
1326 if (!found) {
1327 LocalPointer<NSOverride> cur(new NSOverride);
1328 if (!cur.isNull()) {
1329 char kw[ULOC_KEYWORD_AND_VALUES_CAPACITY];
1330 uprv_strcpy(kw,"numbers=");
1331 nsName.extract(0,len,kw+8,ULOC_KEYWORD_AND_VALUES_CAPACITY-8,US_INV);
1332
1333 Locale ovrLoc(locale.getLanguage(),locale.getCountry(),locale.getVariant(),kw);
1334 cur->hash = nsNameHash;
1335 cur->next = overrideList;
1336 SharedObject::copyPtr(
1337 createSharedNumberFormat(ovrLoc, status), cur->snf);
1338 if (U_FAILURE(status)) {
1339 if (overrideList) {
1340 overrideList->free();
1341 }
1342 return;
1343 }
1344 snf = cur->snf;
1345 overrideList = cur.orphan();
1346 } else {
1347 status = U_MEMORY_ALLOCATION_ERROR;
1348 if (overrideList) {
1349 overrideList->free();
1350 }
1351 return;
1352 }
1353 }
1354
1355 // Now that we have an appropriate number formatter, fill in the appropriate spaces in the
1356 // number formatters table.
1357 if (ovrField.isBogus()) {
1358 switch (type) {
1359 case kOvrStrDate:
1360 case kOvrStrBoth: {
1361 for ( int8_t i=0 ; i<kDateFieldsCount; i++ ) {
1362 SharedObject::copyPtr(snf, fSharedNumberFormatters[kDateFields[i]]);
1363 }
1364 if (type==kOvrStrDate) {
1365 break;
1366 }
1367 U_FALLTHROUGH;
1368 }
1369 case kOvrStrTime : {
1370 for ( int8_t i=0 ; i<kTimeFieldsCount; i++ ) {
1371 SharedObject::copyPtr(snf, fSharedNumberFormatters[kTimeFields[i]]);
1372 }
1373 break;
1374 }
1375 }
1376 } else {
1377 // if the pattern character is unrecognized, signal an error and bail out
1378 UDateFormatField patternCharIndex =
1379 DateFormatSymbols::getPatternCharIndex(ovrField.charAt(0));
1380 if (patternCharIndex == UDAT_FIELD_COUNT) {
1381 status = U_INVALID_FORMAT_ERROR;
1382 if (overrideList) {
1383 overrideList->free();
1384 }
1385 return;
1386 }
1387 SharedObject::copyPtr(snf, fSharedNumberFormatters[patternCharIndex]);
1388 }
1389
1390 start = delimiterPosition + 1;
1391 }
1392 if (overrideList) {
1393 overrideList->free();
1394 }
1395 }
1396
1397 //---------------------------------------------------------------------
1398 void
subFormat(UnicodeString & appendTo,UChar ch,int32_t count,UDisplayContext capitalizationContext,int32_t fieldNum,FieldPositionHandler & handler,Calendar & cal,SimpleDateFormatMutableNFs & mutableNFs,UErrorCode & status) const1399 SimpleDateFormat::subFormat(UnicodeString &appendTo,
1400 UChar ch,
1401 int32_t count,
1402 UDisplayContext capitalizationContext,
1403 int32_t fieldNum,
1404 FieldPositionHandler& handler,
1405 Calendar& cal,
1406 SimpleDateFormatMutableNFs &mutableNFs,
1407 UErrorCode& status) const
1408 {
1409 if (U_FAILURE(status)) {
1410 return;
1411 }
1412
1413 // this function gets called by format() to produce the appropriate substitution
1414 // text for an individual pattern symbol (e.g., "HH" or "yyyy")
1415
1416 UDateFormatField patternCharIndex = DateFormatSymbols::getPatternCharIndex(ch);
1417 const int32_t maxIntCount = 10;
1418 int32_t beginOffset = appendTo.length();
1419 NumberFormat *currentNumberFormat;
1420 DateFormatSymbols::ECapitalizationContextUsageType capContextUsageType = DateFormatSymbols::kCapContextUsageOther;
1421
1422 UBool isHebrewCalendar = (uprv_strcmp(cal.getType(),"hebrew") == 0);
1423 UBool isChineseCalendar = (uprv_strcmp(cal.getType(),"chinese") == 0 || uprv_strcmp(cal.getType(),"dangi") == 0);
1424
1425 // if the pattern character is unrecognized, signal an error and dump out
1426 if (patternCharIndex == UDAT_FIELD_COUNT)
1427 {
1428 if (ch != 0x6C) { // pattern char 'l' (SMALL LETTER L) just gets ignored
1429 status = U_INVALID_FORMAT_ERROR;
1430 }
1431 return;
1432 }
1433
1434 UCalendarDateFields field = fgPatternIndexToCalendarField[patternCharIndex];
1435 int32_t value = 0;
1436 // Don't get value unless it is useful
1437 if (field < UCAL_FIELD_COUNT) {
1438 value = (patternCharIndex != UDAT_RELATED_YEAR_FIELD)? cal.get(field, status): cal.getRelatedYear(status);
1439 }
1440 if (U_FAILURE(status)) {
1441 return;
1442 }
1443
1444 currentNumberFormat = mutableNFs.get(getNumberFormatByIndex(patternCharIndex));
1445 if (currentNumberFormat == NULL) {
1446 status = U_MEMORY_ALLOCATION_ERROR;
1447 return;
1448 }
1449 UnicodeString hebr("hebr", 4, US_INV);
1450
1451 switch (patternCharIndex) {
1452
1453 // for any "G" symbol, write out the appropriate era string
1454 // "GGGG" is wide era name, "GGGGG" is narrow era name, anything else is abbreviated name
1455 case UDAT_ERA_FIELD:
1456 if (isChineseCalendar) {
1457 zeroPaddingNumber(currentNumberFormat,appendTo, value, 1, 9); // as in ICU4J
1458 } else {
1459 if (count == 5) {
1460 _appendSymbol(appendTo, value, fSymbols->fNarrowEras, fSymbols->fNarrowErasCount);
1461 capContextUsageType = DateFormatSymbols::kCapContextUsageEraNarrow;
1462 } else if (count == 4) {
1463 _appendSymbol(appendTo, value, fSymbols->fEraNames, fSymbols->fEraNamesCount);
1464 capContextUsageType = DateFormatSymbols::kCapContextUsageEraWide;
1465 } else {
1466 _appendSymbol(appendTo, value, fSymbols->fEras, fSymbols->fErasCount);
1467 capContextUsageType = DateFormatSymbols::kCapContextUsageEraAbbrev;
1468 }
1469 }
1470 break;
1471
1472 case UDAT_YEAR_NAME_FIELD:
1473 if (fSymbols->fShortYearNames != NULL && value <= fSymbols->fShortYearNamesCount) {
1474 // the Calendar YEAR field runs 1 through 60 for cyclic years
1475 _appendSymbol(appendTo, value - 1, fSymbols->fShortYearNames, fSymbols->fShortYearNamesCount);
1476 break;
1477 }
1478 // else fall through to numeric year handling, do not break here
1479 U_FALLTHROUGH;
1480
1481 // OLD: for "yyyy", write out the whole year; for "yy", write out the last 2 digits
1482 // NEW: UTS#35:
1483 //Year y yy yyy yyyy yyyyy
1484 //AD 1 1 01 001 0001 00001
1485 //AD 12 12 12 012 0012 00012
1486 //AD 123 123 23 123 0123 00123
1487 //AD 1234 1234 34 1234 1234 01234
1488 //AD 12345 12345 45 12345 12345 12345
1489 case UDAT_YEAR_FIELD:
1490 case UDAT_YEAR_WOY_FIELD:
1491 if (fDateOverride.compare(hebr)==0 && value>HEBREW_CAL_CUR_MILLENIUM_START_YEAR && value<HEBREW_CAL_CUR_MILLENIUM_END_YEAR) {
1492 value-=HEBREW_CAL_CUR_MILLENIUM_START_YEAR;
1493 }
1494 if(count == 2)
1495 zeroPaddingNumber(currentNumberFormat, appendTo, value, 2, 2);
1496 else
1497 zeroPaddingNumber(currentNumberFormat, appendTo, value, count, maxIntCount);
1498 break;
1499
1500 // for "MMMM"/"LLLL", write out the whole month name, for "MMM"/"LLL", write out the month
1501 // abbreviation, for "M"/"L" or "MM"/"LL", write out the month as a number with the
1502 // appropriate number of digits
1503 // for "MMMMM"/"LLLLL", use the narrow form
1504 case UDAT_MONTH_FIELD:
1505 case UDAT_STANDALONE_MONTH_FIELD:
1506 if ( isHebrewCalendar ) {
1507 HebrewCalendar *hc = (HebrewCalendar*)&cal;
1508 if (hc->isLeapYear(hc->get(UCAL_YEAR,status)) && value == 6 && count >= 3 )
1509 value = 13; // Show alternate form for Adar II in leap years in Hebrew calendar.
1510 if (!hc->isLeapYear(hc->get(UCAL_YEAR,status)) && value >= 6 && count < 3 )
1511 value--; // Adjust the month number down 1 in Hebrew non-leap years, i.e. Adar is 6, not 7.
1512 }
1513 {
1514 int32_t isLeapMonth = (fSymbols->fLeapMonthPatterns != NULL && fSymbols->fLeapMonthPatternsCount >= DateFormatSymbols::kMonthPatternsCount)?
1515 cal.get(UCAL_IS_LEAP_MONTH, status): 0;
1516 // should consolidate the next section by using arrays of pointers & counts for the right symbols...
1517 if (count == 5) {
1518 if (patternCharIndex == UDAT_MONTH_FIELD) {
1519 _appendSymbolWithMonthPattern(appendTo, value, fSymbols->fNarrowMonths, fSymbols->fNarrowMonthsCount,
1520 (isLeapMonth!=0)? &(fSymbols->fLeapMonthPatterns[DateFormatSymbols::kLeapMonthPatternFormatNarrow]): NULL, status);
1521 } else {
1522 _appendSymbolWithMonthPattern(appendTo, value, fSymbols->fStandaloneNarrowMonths, fSymbols->fStandaloneNarrowMonthsCount,
1523 (isLeapMonth!=0)? &(fSymbols->fLeapMonthPatterns[DateFormatSymbols::kLeapMonthPatternStandaloneNarrow]): NULL, status);
1524 }
1525 capContextUsageType = DateFormatSymbols::kCapContextUsageMonthNarrow;
1526 } else if (count == 4) {
1527 if (patternCharIndex == UDAT_MONTH_FIELD) {
1528 _appendSymbolWithMonthPattern(appendTo, value, fSymbols->fMonths, fSymbols->fMonthsCount,
1529 (isLeapMonth!=0)? &(fSymbols->fLeapMonthPatterns[DateFormatSymbols::kLeapMonthPatternFormatWide]): NULL, status);
1530 capContextUsageType = DateFormatSymbols::kCapContextUsageMonthFormat;
1531 } else {
1532 _appendSymbolWithMonthPattern(appendTo, value, fSymbols->fStandaloneMonths, fSymbols->fStandaloneMonthsCount,
1533 (isLeapMonth!=0)? &(fSymbols->fLeapMonthPatterns[DateFormatSymbols::kLeapMonthPatternStandaloneWide]): NULL, status);
1534 capContextUsageType = DateFormatSymbols::kCapContextUsageMonthStandalone;
1535 }
1536 } else if (count == 3) {
1537 if (patternCharIndex == UDAT_MONTH_FIELD) {
1538 _appendSymbolWithMonthPattern(appendTo, value, fSymbols->fShortMonths, fSymbols->fShortMonthsCount,
1539 (isLeapMonth!=0)? &(fSymbols->fLeapMonthPatterns[DateFormatSymbols::kLeapMonthPatternFormatAbbrev]): NULL, status);
1540 capContextUsageType = DateFormatSymbols::kCapContextUsageMonthFormat;
1541 } else {
1542 _appendSymbolWithMonthPattern(appendTo, value, fSymbols->fStandaloneShortMonths, fSymbols->fStandaloneShortMonthsCount,
1543 (isLeapMonth!=0)? &(fSymbols->fLeapMonthPatterns[DateFormatSymbols::kLeapMonthPatternStandaloneAbbrev]): NULL, status);
1544 capContextUsageType = DateFormatSymbols::kCapContextUsageMonthStandalone;
1545 }
1546 } else {
1547 UnicodeString monthNumber;
1548 zeroPaddingNumber(currentNumberFormat,monthNumber, value + 1, count, maxIntCount);
1549 _appendSymbolWithMonthPattern(appendTo, 0, &monthNumber, 1,
1550 (isLeapMonth!=0)? &(fSymbols->fLeapMonthPatterns[DateFormatSymbols::kLeapMonthPatternNumeric]): NULL, status);
1551 }
1552 }
1553 break;
1554
1555 // for "k" and "kk", write out the hour, adjusting midnight to appear as "24"
1556 case UDAT_HOUR_OF_DAY1_FIELD:
1557 if (value == 0)
1558 zeroPaddingNumber(currentNumberFormat,appendTo, cal.getMaximum(UCAL_HOUR_OF_DAY) + 1, count, maxIntCount);
1559 else
1560 zeroPaddingNumber(currentNumberFormat,appendTo, value, count, maxIntCount);
1561 break;
1562
1563 case UDAT_FRACTIONAL_SECOND_FIELD:
1564 // Fractional seconds left-justify
1565 {
1566 currentNumberFormat->setMinimumIntegerDigits((count > 3) ? 3 : count);
1567 currentNumberFormat->setMaximumIntegerDigits(maxIntCount);
1568 if (count == 1) {
1569 value /= 100;
1570 } else if (count == 2) {
1571 value /= 10;
1572 }
1573 FieldPosition p(FieldPosition::DONT_CARE);
1574 currentNumberFormat->format(value, appendTo, p);
1575 if (count > 3) {
1576 currentNumberFormat->setMinimumIntegerDigits(count - 3);
1577 currentNumberFormat->format((int32_t)0, appendTo, p);
1578 }
1579 }
1580 break;
1581
1582 // for "ee" or "e", use local numeric day-of-the-week
1583 // for "EEEEEE" or "eeeeee", write out the short day-of-the-week name
1584 // for "EEEEE" or "eeeee", write out the narrow day-of-the-week name
1585 // for "EEEE" or "eeee", write out the wide day-of-the-week name
1586 // for "EEE" or "EE" or "E" or "eee", write out the abbreviated day-of-the-week name
1587 case UDAT_DOW_LOCAL_FIELD:
1588 if ( count < 3 ) {
1589 zeroPaddingNumber(currentNumberFormat,appendTo, value, count, maxIntCount);
1590 break;
1591 }
1592 // fall through to EEEEE-EEE handling, but for that we don't want local day-of-week,
1593 // we want standard day-of-week, so first fix value to work for EEEEE-EEE.
1594 value = cal.get(UCAL_DAY_OF_WEEK, status);
1595 if (U_FAILURE(status)) {
1596 return;
1597 }
1598 // fall through, do not break here
1599 U_FALLTHROUGH;
1600 case UDAT_DAY_OF_WEEK_FIELD:
1601 if (count == 5) {
1602 _appendSymbol(appendTo, value, fSymbols->fNarrowWeekdays,
1603 fSymbols->fNarrowWeekdaysCount);
1604 capContextUsageType = DateFormatSymbols::kCapContextUsageDayNarrow;
1605 } else if (count == 4) {
1606 _appendSymbol(appendTo, value, fSymbols->fWeekdays,
1607 fSymbols->fWeekdaysCount);
1608 capContextUsageType = DateFormatSymbols::kCapContextUsageDayFormat;
1609 } else if (count == 6) {
1610 _appendSymbol(appendTo, value, fSymbols->fShorterWeekdays,
1611 fSymbols->fShorterWeekdaysCount);
1612 capContextUsageType = DateFormatSymbols::kCapContextUsageDayFormat;
1613 } else {
1614 _appendSymbol(appendTo, value, fSymbols->fShortWeekdays,
1615 fSymbols->fShortWeekdaysCount);
1616 capContextUsageType = DateFormatSymbols::kCapContextUsageDayFormat;
1617 }
1618 break;
1619
1620 // for "ccc", write out the abbreviated day-of-the-week name
1621 // for "cccc", write out the wide day-of-the-week name
1622 // for "ccccc", use the narrow day-of-the-week name
1623 // for "ccccc", use the short day-of-the-week name
1624 case UDAT_STANDALONE_DAY_FIELD:
1625 if ( count < 3 ) {
1626 zeroPaddingNumber(currentNumberFormat,appendTo, value, 1, maxIntCount);
1627 break;
1628 }
1629 // fall through to alpha DOW handling, but for that we don't want local day-of-week,
1630 // we want standard day-of-week, so first fix value.
1631 value = cal.get(UCAL_DAY_OF_WEEK, status);
1632 if (U_FAILURE(status)) {
1633 return;
1634 }
1635 if (count == 5) {
1636 _appendSymbol(appendTo, value, fSymbols->fStandaloneNarrowWeekdays,
1637 fSymbols->fStandaloneNarrowWeekdaysCount);
1638 capContextUsageType = DateFormatSymbols::kCapContextUsageDayNarrow;
1639 } else if (count == 4) {
1640 _appendSymbol(appendTo, value, fSymbols->fStandaloneWeekdays,
1641 fSymbols->fStandaloneWeekdaysCount);
1642 capContextUsageType = DateFormatSymbols::kCapContextUsageDayStandalone;
1643 } else if (count == 6) {
1644 _appendSymbol(appendTo, value, fSymbols->fStandaloneShorterWeekdays,
1645 fSymbols->fStandaloneShorterWeekdaysCount);
1646 capContextUsageType = DateFormatSymbols::kCapContextUsageDayStandalone;
1647 } else { // count == 3
1648 _appendSymbol(appendTo, value, fSymbols->fStandaloneShortWeekdays,
1649 fSymbols->fStandaloneShortWeekdaysCount);
1650 capContextUsageType = DateFormatSymbols::kCapContextUsageDayStandalone;
1651 }
1652 break;
1653
1654 // for "a" symbol, write out the whole AM/PM string
1655 case UDAT_AM_PM_FIELD:
1656 if (count < 5) {
1657 _appendSymbol(appendTo, value, fSymbols->fAmPms,
1658 fSymbols->fAmPmsCount);
1659 } else {
1660 _appendSymbol(appendTo, value, fSymbols->fNarrowAmPms,
1661 fSymbols->fNarrowAmPmsCount);
1662 }
1663 break;
1664
1665 // if we see pattern character for UDAT_TIME_SEPARATOR_FIELD (none currently defined),
1666 // write out the time separator string. Leave support in for future definition.
1667 case UDAT_TIME_SEPARATOR_FIELD:
1668 {
1669 UnicodeString separator;
1670 appendTo += fSymbols->getTimeSeparatorString(separator);
1671 }
1672 break;
1673
1674 // for "h" and "hh", write out the hour, adjusting noon and midnight to show up
1675 // as "12"
1676 case UDAT_HOUR1_FIELD:
1677 if (value == 0)
1678 zeroPaddingNumber(currentNumberFormat,appendTo, cal.getLeastMaximum(UCAL_HOUR) + 1, count, maxIntCount);
1679 else
1680 zeroPaddingNumber(currentNumberFormat,appendTo, value, count, maxIntCount);
1681 break;
1682
1683 case UDAT_TIMEZONE_FIELD: // 'z'
1684 case UDAT_TIMEZONE_RFC_FIELD: // 'Z'
1685 case UDAT_TIMEZONE_GENERIC_FIELD: // 'v'
1686 case UDAT_TIMEZONE_SPECIAL_FIELD: // 'V'
1687 case UDAT_TIMEZONE_LOCALIZED_GMT_OFFSET_FIELD: // 'O'
1688 case UDAT_TIMEZONE_ISO_FIELD: // 'X'
1689 case UDAT_TIMEZONE_ISO_LOCAL_FIELD: // 'x'
1690 {
1691 UChar zsbuf[64];
1692 UnicodeString zoneString(zsbuf, 0, UPRV_LENGTHOF(zsbuf));
1693 const TimeZone& tz = cal.getTimeZone();
1694 UDate date = cal.getTime(status);
1695 if (U_SUCCESS(status)) {
1696 if (patternCharIndex == UDAT_TIMEZONE_FIELD) {
1697 if (count < 4) {
1698 // "z", "zz", "zzz"
1699 tzFormat()->format(UTZFMT_STYLE_SPECIFIC_SHORT, tz, date, zoneString);
1700 capContextUsageType = DateFormatSymbols::kCapContextUsageMetazoneShort;
1701 } else {
1702 // "zzzz" or longer
1703 tzFormat()->format(UTZFMT_STYLE_SPECIFIC_LONG, tz, date, zoneString);
1704 capContextUsageType = DateFormatSymbols::kCapContextUsageMetazoneLong;
1705 }
1706 }
1707 else if (patternCharIndex == UDAT_TIMEZONE_RFC_FIELD) {
1708 if (count < 4) {
1709 // "Z"
1710 tzFormat()->format(UTZFMT_STYLE_ISO_BASIC_LOCAL_FULL, tz, date, zoneString);
1711 } else if (count == 5) {
1712 // "ZZZZZ"
1713 tzFormat()->format(UTZFMT_STYLE_ISO_EXTENDED_FULL, tz, date, zoneString);
1714 } else {
1715 // "ZZ", "ZZZ", "ZZZZ"
1716 tzFormat()->format(UTZFMT_STYLE_LOCALIZED_GMT, tz, date, zoneString);
1717 }
1718 }
1719 else if (patternCharIndex == UDAT_TIMEZONE_GENERIC_FIELD) {
1720 if (count == 1) {
1721 // "v"
1722 tzFormat()->format(UTZFMT_STYLE_GENERIC_SHORT, tz, date, zoneString);
1723 capContextUsageType = DateFormatSymbols::kCapContextUsageMetazoneShort;
1724 } else if (count == 4) {
1725 // "vvvv"
1726 tzFormat()->format(UTZFMT_STYLE_GENERIC_LONG, tz, date, zoneString);
1727 capContextUsageType = DateFormatSymbols::kCapContextUsageMetazoneLong;
1728 }
1729 }
1730 else if (patternCharIndex == UDAT_TIMEZONE_SPECIAL_FIELD) {
1731 if (count == 1) {
1732 // "V"
1733 tzFormat()->format(UTZFMT_STYLE_ZONE_ID_SHORT, tz, date, zoneString);
1734 } else if (count == 2) {
1735 // "VV"
1736 tzFormat()->format(UTZFMT_STYLE_ZONE_ID, tz, date, zoneString);
1737 } else if (count == 3) {
1738 // "VVV"
1739 tzFormat()->format(UTZFMT_STYLE_EXEMPLAR_LOCATION, tz, date, zoneString);
1740 } else if (count == 4) {
1741 // "VVVV"
1742 tzFormat()->format(UTZFMT_STYLE_GENERIC_LOCATION, tz, date, zoneString);
1743 capContextUsageType = DateFormatSymbols::kCapContextUsageZoneLong;
1744 }
1745 }
1746 else if (patternCharIndex == UDAT_TIMEZONE_LOCALIZED_GMT_OFFSET_FIELD) {
1747 if (count == 1) {
1748 // "O"
1749 tzFormat()->format(UTZFMT_STYLE_LOCALIZED_GMT_SHORT, tz, date, zoneString);
1750 } else if (count == 4) {
1751 // "OOOO"
1752 tzFormat()->format(UTZFMT_STYLE_LOCALIZED_GMT, tz, date, zoneString);
1753 }
1754 }
1755 else if (patternCharIndex == UDAT_TIMEZONE_ISO_FIELD) {
1756 if (count == 1) {
1757 // "X"
1758 tzFormat()->format(UTZFMT_STYLE_ISO_BASIC_SHORT, tz, date, zoneString);
1759 } else if (count == 2) {
1760 // "XX"
1761 tzFormat()->format(UTZFMT_STYLE_ISO_BASIC_FIXED, tz, date, zoneString);
1762 } else if (count == 3) {
1763 // "XXX"
1764 tzFormat()->format(UTZFMT_STYLE_ISO_EXTENDED_FIXED, tz, date, zoneString);
1765 } else if (count == 4) {
1766 // "XXXX"
1767 tzFormat()->format(UTZFMT_STYLE_ISO_BASIC_FULL, tz, date, zoneString);
1768 } else if (count == 5) {
1769 // "XXXXX"
1770 tzFormat()->format(UTZFMT_STYLE_ISO_EXTENDED_FULL, tz, date, zoneString);
1771 }
1772 }
1773 else if (patternCharIndex == UDAT_TIMEZONE_ISO_LOCAL_FIELD) {
1774 if (count == 1) {
1775 // "x"
1776 tzFormat()->format(UTZFMT_STYLE_ISO_BASIC_LOCAL_SHORT, tz, date, zoneString);
1777 } else if (count == 2) {
1778 // "xx"
1779 tzFormat()->format(UTZFMT_STYLE_ISO_BASIC_LOCAL_FIXED, tz, date, zoneString);
1780 } else if (count == 3) {
1781 // "xxx"
1782 tzFormat()->format(UTZFMT_STYLE_ISO_EXTENDED_LOCAL_FIXED, tz, date, zoneString);
1783 } else if (count == 4) {
1784 // "xxxx"
1785 tzFormat()->format(UTZFMT_STYLE_ISO_BASIC_LOCAL_FULL, tz, date, zoneString);
1786 } else if (count == 5) {
1787 // "xxxxx"
1788 tzFormat()->format(UTZFMT_STYLE_ISO_EXTENDED_LOCAL_FULL, tz, date, zoneString);
1789 }
1790 }
1791 else {
1792 U_ASSERT(FALSE);
1793 }
1794 }
1795 appendTo += zoneString;
1796 }
1797 break;
1798
1799 case UDAT_QUARTER_FIELD:
1800 if (count >= 4)
1801 _appendSymbol(appendTo, value/3, fSymbols->fQuarters,
1802 fSymbols->fQuartersCount);
1803 else if (count == 3)
1804 _appendSymbol(appendTo, value/3, fSymbols->fShortQuarters,
1805 fSymbols->fShortQuartersCount);
1806 else
1807 zeroPaddingNumber(currentNumberFormat,appendTo, (value/3) + 1, count, maxIntCount);
1808 break;
1809
1810 case UDAT_STANDALONE_QUARTER_FIELD:
1811 if (count >= 4)
1812 _appendSymbol(appendTo, value/3, fSymbols->fStandaloneQuarters,
1813 fSymbols->fStandaloneQuartersCount);
1814 else if (count == 3)
1815 _appendSymbol(appendTo, value/3, fSymbols->fStandaloneShortQuarters,
1816 fSymbols->fStandaloneShortQuartersCount);
1817 else
1818 zeroPaddingNumber(currentNumberFormat,appendTo, (value/3) + 1, count, maxIntCount);
1819 break;
1820
1821 case UDAT_AM_PM_MIDNIGHT_NOON_FIELD:
1822 {
1823 const UnicodeString *toAppend = NULL;
1824 int32_t hour = cal.get(UCAL_HOUR_OF_DAY, status);
1825
1826 // Note: "midnight" can be ambiguous as to whether it refers to beginning of day or end of day.
1827 // For ICU 57 output of "midnight" is temporarily suppressed.
1828
1829 // For "midnight" and "noon":
1830 // Time, as displayed, must be exactly noon or midnight.
1831 // This means minutes and seconds, if present, must be zero.
1832 if ((/*hour == 0 ||*/ hour == 12) &&
1833 (!fHasMinute || cal.get(UCAL_MINUTE, status) == 0) &&
1834 (!fHasSecond || cal.get(UCAL_SECOND, status) == 0)) {
1835 // Stealing am/pm value to use as our array index.
1836 // It works out: am/midnight are both 0, pm/noon are both 1,
1837 // 12 am is 12 midnight, and 12 pm is 12 noon.
1838 int32_t value = cal.get(UCAL_AM_PM, status);
1839
1840 if (count <= 3) {
1841 toAppend = &fSymbols->fAbbreviatedDayPeriods[value];
1842 } else if (count == 4 || count > 5) {
1843 toAppend = &fSymbols->fWideDayPeriods[value];
1844 } else { // count == 5
1845 toAppend = &fSymbols->fNarrowDayPeriods[value];
1846 }
1847 }
1848
1849 // toAppend is NULL if time isn't exactly midnight or noon (as displayed).
1850 // toAppend is bogus if time is midnight or noon, but no localized string exists.
1851 // In either case, fall back to am/pm.
1852 if (toAppend == NULL || toAppend->isBogus()) {
1853 // Reformat with identical arguments except ch, now changed to 'a'.
1854 subFormat(appendTo, 0x61, count, capitalizationContext, fieldNum,
1855 handler, cal, mutableNFs, status);
1856 } else {
1857 appendTo += *toAppend;
1858 }
1859
1860 break;
1861 }
1862
1863 case UDAT_FLEXIBLE_DAY_PERIOD_FIELD:
1864 {
1865 // TODO: Maybe fetch the DayperiodRules during initialization (instead of at the first
1866 // loading of an instance) if a relevant pattern character (b or B) is used.
1867 const DayPeriodRules *ruleSet = DayPeriodRules::getInstance(this->getSmpFmtLocale(), status);
1868 if (U_FAILURE(status)) {
1869 // Data doesn't conform to spec, therefore loading failed.
1870 break;
1871 }
1872 if (ruleSet == NULL) {
1873 // Data doesn't exist for the locale we're looking for.
1874 // Falling back to am/pm.
1875 subFormat(appendTo, 0x61, count, capitalizationContext, fieldNum,
1876 handler, cal, mutableNFs, status);
1877 break;
1878 }
1879
1880 // Get current display time.
1881 int32_t hour = cal.get(UCAL_HOUR_OF_DAY, status);
1882 int32_t minute = 0;
1883 if (fHasMinute) {
1884 minute = cal.get(UCAL_MINUTE, status);
1885 }
1886 int32_t second = 0;
1887 if (fHasSecond) {
1888 second = cal.get(UCAL_SECOND, status);
1889 }
1890
1891 // Determine day period.
1892 DayPeriodRules::DayPeriod periodType;
1893 if (hour == 0 && minute == 0 && second == 0 && ruleSet->hasMidnight()) {
1894 periodType = DayPeriodRules::DAYPERIOD_MIDNIGHT;
1895 } else if (hour == 12 && minute == 0 && second == 0 && ruleSet->hasNoon()) {
1896 periodType = DayPeriodRules::DAYPERIOD_NOON;
1897 } else {
1898 periodType = ruleSet->getDayPeriodForHour(hour);
1899 }
1900
1901 // Rule set exists, therefore periodType can't be UNKNOWN.
1902 // Get localized string.
1903 U_ASSERT(periodType != DayPeriodRules::DAYPERIOD_UNKNOWN);
1904 UnicodeString *toAppend = NULL;
1905 int32_t index;
1906
1907 // Note: "midnight" can be ambiguous as to whether it refers to beginning of day or end of day.
1908 // For ICU 57 output of "midnight" is temporarily suppressed.
1909
1910 if (periodType != DayPeriodRules::DAYPERIOD_AM &&
1911 periodType != DayPeriodRules::DAYPERIOD_PM &&
1912 periodType != DayPeriodRules::DAYPERIOD_MIDNIGHT) {
1913 index = (int32_t)periodType;
1914 if (count <= 3) {
1915 toAppend = &fSymbols->fAbbreviatedDayPeriods[index]; // i.e. short
1916 } else if (count == 4 || count > 5) {
1917 toAppend = &fSymbols->fWideDayPeriods[index];
1918 } else { // count == 5
1919 toAppend = &fSymbols->fNarrowDayPeriods[index];
1920 }
1921 }
1922
1923 // Fallback schedule:
1924 // Midnight/Noon -> General Periods -> AM/PM.
1925
1926 // Midnight/Noon -> General Periods.
1927 if ((toAppend == NULL || toAppend->isBogus()) &&
1928 (periodType == DayPeriodRules::DAYPERIOD_MIDNIGHT ||
1929 periodType == DayPeriodRules::DAYPERIOD_NOON)) {
1930 periodType = ruleSet->getDayPeriodForHour(hour);
1931 index = (int32_t)periodType;
1932
1933 if (count <= 3) {
1934 toAppend = &fSymbols->fAbbreviatedDayPeriods[index]; // i.e. short
1935 } else if (count == 4 || count > 5) {
1936 toAppend = &fSymbols->fWideDayPeriods[index];
1937 } else { // count == 5
1938 toAppend = &fSymbols->fNarrowDayPeriods[index];
1939 }
1940 }
1941
1942 // General Periods -> AM/PM.
1943 if (periodType == DayPeriodRules::DAYPERIOD_AM ||
1944 periodType == DayPeriodRules::DAYPERIOD_PM ||
1945 toAppend->isBogus()) {
1946 subFormat(appendTo, 0x61, count, capitalizationContext, fieldNum,
1947 handler, cal, mutableNFs, status);
1948 }
1949 else {
1950 appendTo += *toAppend;
1951 }
1952
1953 break;
1954 }
1955
1956 // all of the other pattern symbols can be formatted as simple numbers with
1957 // appropriate zero padding
1958 default:
1959 zeroPaddingNumber(currentNumberFormat,appendTo, value, count, maxIntCount);
1960 break;
1961 }
1962 #if !UCONFIG_NO_BREAK_ITERATION
1963 // if first field, check to see whether we need to and are able to titlecase it
1964 if (fieldNum == 0 && u_islower(appendTo.char32At(beginOffset)) && fCapitalizationBrkIter != NULL) {
1965 UBool titlecase = FALSE;
1966 switch (capitalizationContext) {
1967 case UDISPCTX_CAPITALIZATION_FOR_BEGINNING_OF_SENTENCE:
1968 titlecase = TRUE;
1969 break;
1970 case UDISPCTX_CAPITALIZATION_FOR_UI_LIST_OR_MENU:
1971 titlecase = fSymbols->fCapitalization[capContextUsageType][0];
1972 break;
1973 case UDISPCTX_CAPITALIZATION_FOR_STANDALONE:
1974 titlecase = fSymbols->fCapitalization[capContextUsageType][1];
1975 break;
1976 default:
1977 // titlecase = FALSE;
1978 break;
1979 }
1980 if (titlecase) {
1981 UnicodeString firstField(appendTo, beginOffset);
1982 firstField.toTitle(fCapitalizationBrkIter, fLocale, U_TITLECASE_NO_LOWERCASE | U_TITLECASE_NO_BREAK_ADJUSTMENT);
1983 appendTo.replaceBetween(beginOffset, appendTo.length(), firstField);
1984 }
1985 }
1986 #endif
1987
1988 handler.addAttribute(fgPatternIndexToDateFormatField[patternCharIndex], beginOffset, appendTo.length());
1989 }
1990
1991 //----------------------------------------------------------------------
1992
adoptNumberFormat(NumberFormat * formatToAdopt)1993 void SimpleDateFormat::adoptNumberFormat(NumberFormat *formatToAdopt) {
1994 fixNumberFormatForDates(*formatToAdopt);
1995 delete fNumberFormat;
1996 fNumberFormat = formatToAdopt;
1997
1998 // We successfully set the default number format. Now delete the overrides
1999 // (can't fail).
2000 if (fSharedNumberFormatters) {
2001 freeSharedNumberFormatters(fSharedNumberFormatters);
2002 fSharedNumberFormatters = NULL;
2003 }
2004 }
2005
adoptNumberFormat(const UnicodeString & fields,NumberFormat * formatToAdopt,UErrorCode & status)2006 void SimpleDateFormat::adoptNumberFormat(const UnicodeString& fields, NumberFormat *formatToAdopt, UErrorCode &status){
2007 fixNumberFormatForDates(*formatToAdopt);
2008 LocalPointer<NumberFormat> fmt(formatToAdopt);
2009 if (U_FAILURE(status)) {
2010 return;
2011 }
2012
2013 // We must ensure fSharedNumberFormatters is allocated.
2014 if (fSharedNumberFormatters == NULL) {
2015 fSharedNumberFormatters = allocSharedNumberFormatters();
2016 if (fSharedNumberFormatters == NULL) {
2017 status = U_MEMORY_ALLOCATION_ERROR;
2018 return;
2019 }
2020 }
2021 const SharedNumberFormat *newFormat = createSharedNumberFormat(fmt.orphan());
2022 if (newFormat == NULL) {
2023 status = U_MEMORY_ALLOCATION_ERROR;
2024 return;
2025 }
2026 for (int i=0; i<fields.length(); i++) {
2027 UChar field = fields.charAt(i);
2028 // if the pattern character is unrecognized, signal an error and bail out
2029 UDateFormatField patternCharIndex = DateFormatSymbols::getPatternCharIndex(field);
2030 if (patternCharIndex == UDAT_FIELD_COUNT) {
2031 status = U_INVALID_FORMAT_ERROR;
2032 newFormat->deleteIfZeroRefCount();
2033 return;
2034 }
2035
2036 // Set the number formatter in the table
2037 SharedObject::copyPtr(
2038 newFormat, fSharedNumberFormatters[patternCharIndex]);
2039 }
2040 newFormat->deleteIfZeroRefCount();
2041 }
2042
2043 const NumberFormat *
getNumberFormatForField(UChar field) const2044 SimpleDateFormat::getNumberFormatForField(UChar field) const {
2045 UDateFormatField index = DateFormatSymbols::getPatternCharIndex(field);
2046 if (index == UDAT_FIELD_COUNT) {
2047 return NULL;
2048 }
2049 return getNumberFormatByIndex(index);
2050 }
2051
2052 //----------------------------------------------------------------------
2053 void
zeroPaddingNumber(NumberFormat * currentNumberFormat,UnicodeString & appendTo,int32_t value,int32_t minDigits,int32_t maxDigits) const2054 SimpleDateFormat::zeroPaddingNumber(
2055 NumberFormat *currentNumberFormat,
2056 UnicodeString &appendTo,
2057 int32_t value, int32_t minDigits, int32_t maxDigits) const
2058 {
2059 if (currentNumberFormat!=NULL) {
2060 FieldPosition pos(FieldPosition::DONT_CARE);
2061
2062 currentNumberFormat->setMinimumIntegerDigits(minDigits);
2063 currentNumberFormat->setMaximumIntegerDigits(maxDigits);
2064 currentNumberFormat->format(value, appendTo, pos); // 3rd arg is there to speed up processing
2065 }
2066 }
2067
2068 //----------------------------------------------------------------------
2069
2070 /**
2071 * Return true if the given format character, occuring count
2072 * times, represents a numeric field.
2073 */
isNumeric(UChar formatChar,int32_t count)2074 UBool SimpleDateFormat::isNumeric(UChar formatChar, int32_t count) {
2075 return DateFormatSymbols::isNumericPatternChar(formatChar, count);
2076 }
2077
2078 UBool
isAtNumericField(const UnicodeString & pattern,int32_t patternOffset)2079 SimpleDateFormat::isAtNumericField(const UnicodeString &pattern, int32_t patternOffset) {
2080 if (patternOffset >= pattern.length()) {
2081 // not at any field
2082 return FALSE;
2083 }
2084 UChar ch = pattern.charAt(patternOffset);
2085 UDateFormatField f = DateFormatSymbols::getPatternCharIndex(ch);
2086 if (f == UDAT_FIELD_COUNT) {
2087 // not at any field
2088 return FALSE;
2089 }
2090 int32_t i = patternOffset;
2091 while (pattern.charAt(++i) == ch) {}
2092 return DateFormatSymbols::isNumericField(f, i - patternOffset);
2093 }
2094
2095 UBool
isAfterNonNumericField(const UnicodeString & pattern,int32_t patternOffset)2096 SimpleDateFormat::isAfterNonNumericField(const UnicodeString &pattern, int32_t patternOffset) {
2097 if (patternOffset <= 0) {
2098 // not after any field
2099 return FALSE;
2100 }
2101 UChar ch = pattern.charAt(--patternOffset);
2102 UDateFormatField f = DateFormatSymbols::getPatternCharIndex(ch);
2103 if (f == UDAT_FIELD_COUNT) {
2104 // not after any field
2105 return FALSE;
2106 }
2107 int32_t i = patternOffset;
2108 while (pattern.charAt(--i) == ch) {}
2109 return !DateFormatSymbols::isNumericField(f, patternOffset - i);
2110 }
2111
2112 void
parse(const UnicodeString & text,Calendar & cal,ParsePosition & parsePos) const2113 SimpleDateFormat::parse(const UnicodeString& text, Calendar& cal, ParsePosition& parsePos) const
2114 {
2115 UErrorCode status = U_ZERO_ERROR;
2116 int32_t pos = parsePos.getIndex();
2117 if(parsePos.getIndex() < 0) {
2118 parsePos.setErrorIndex(0);
2119 return;
2120 }
2121 int32_t start = pos;
2122
2123 // Hold the day period until everything else is parsed, because we need
2124 // the hour to interpret time correctly.
2125 int32_t dayPeriodInt = -1;
2126
2127 UBool ambiguousYear[] = { FALSE };
2128 int32_t saveHebrewMonth = -1;
2129 int32_t count = 0;
2130 UTimeZoneFormatTimeType tzTimeType = UTZFMT_TIME_TYPE_UNKNOWN;
2131 SimpleDateFormatMutableNFs mutableNFs;
2132
2133 // For parsing abutting numeric fields. 'abutPat' is the
2134 // offset into 'pattern' of the first of 2 or more abutting
2135 // numeric fields. 'abutStart' is the offset into 'text'
2136 // where parsing the fields begins. 'abutPass' starts off as 0
2137 // and increments each time we try to parse the fields.
2138 int32_t abutPat = -1; // If >=0, we are in a run of abutting numeric fields
2139 int32_t abutStart = 0;
2140 int32_t abutPass = 0;
2141 UBool inQuote = FALSE;
2142
2143 MessageFormat * numericLeapMonthFormatter = NULL;
2144
2145 Calendar* calClone = NULL;
2146 Calendar *workCal = &cal;
2147 if (&cal != fCalendar && uprv_strcmp(cal.getType(), fCalendar->getType()) != 0) {
2148 // Different calendar type
2149 // We use the time/zone from the input calendar, but
2150 // do not use the input calendar for field calculation.
2151 calClone = fCalendar->clone();
2152 if (calClone != NULL) {
2153 calClone->setTime(cal.getTime(status),status);
2154 if (U_FAILURE(status)) {
2155 goto ExitParse;
2156 }
2157 calClone->setTimeZone(cal.getTimeZone());
2158 workCal = calClone;
2159 } else {
2160 status = U_MEMORY_ALLOCATION_ERROR;
2161 goto ExitParse;
2162 }
2163 }
2164
2165 if (fSymbols->fLeapMonthPatterns != NULL && fSymbols->fLeapMonthPatternsCount >= DateFormatSymbols::kMonthPatternsCount) {
2166 numericLeapMonthFormatter = new MessageFormat(fSymbols->fLeapMonthPatterns[DateFormatSymbols::kLeapMonthPatternNumeric], fLocale, status);
2167 if (numericLeapMonthFormatter == NULL) {
2168 status = U_MEMORY_ALLOCATION_ERROR;
2169 goto ExitParse;
2170 } else if (U_FAILURE(status)) {
2171 goto ExitParse; // this will delete numericLeapMonthFormatter
2172 }
2173 }
2174
2175 for (int32_t i=0; i<fPattern.length(); ++i) {
2176 UChar ch = fPattern.charAt(i);
2177
2178 // Handle alphabetic field characters.
2179 if (!inQuote && isSyntaxChar(ch)) {
2180 int32_t fieldPat = i;
2181
2182 // Count the length of this field specifier
2183 count = 1;
2184 while ((i+1)<fPattern.length() &&
2185 fPattern.charAt(i+1) == ch) {
2186 ++count;
2187 ++i;
2188 }
2189
2190 if (isNumeric(ch, count)) {
2191 if (abutPat < 0) {
2192 // Determine if there is an abutting numeric field.
2193 // Record the start of a set of abutting numeric fields.
2194 if (isAtNumericField(fPattern, i + 1)) {
2195 abutPat = fieldPat;
2196 abutStart = pos;
2197 abutPass = 0;
2198 }
2199 }
2200 } else {
2201 abutPat = -1; // End of any abutting fields
2202 }
2203
2204 // Handle fields within a run of abutting numeric fields. Take
2205 // the pattern "HHmmss" as an example. We will try to parse
2206 // 2/2/2 characters of the input text, then if that fails,
2207 // 1/2/2. We only adjust the width of the leftmost field; the
2208 // others remain fixed. This allows "123456" => 12:34:56, but
2209 // "12345" => 1:23:45. Likewise, for the pattern "yyyyMMdd" we
2210 // try 4/2/2, 3/2/2, 2/2/2, and finally 1/2/2.
2211 if (abutPat >= 0) {
2212 // If we are at the start of a run of abutting fields, then
2213 // shorten this field in each pass. If we can't shorten
2214 // this field any more, then the parse of this set of
2215 // abutting numeric fields has failed.
2216 if (fieldPat == abutPat) {
2217 count -= abutPass++;
2218 if (count == 0) {
2219 status = U_PARSE_ERROR;
2220 goto ExitParse;
2221 }
2222 }
2223
2224 pos = subParse(text, pos, ch, count,
2225 TRUE, FALSE, ambiguousYear, saveHebrewMonth, *workCal, i, numericLeapMonthFormatter, &tzTimeType, mutableNFs);
2226
2227 // If the parse fails anywhere in the run, back up to the
2228 // start of the run and retry.
2229 if (pos < 0) {
2230 i = abutPat - 1;
2231 pos = abutStart;
2232 continue;
2233 }
2234 }
2235
2236 // Handle non-numeric fields and non-abutting numeric
2237 // fields.
2238 else if (ch != 0x6C) { // pattern char 'l' (SMALL LETTER L) just gets ignored
2239 int32_t s = subParse(text, pos, ch, count,
2240 FALSE, TRUE, ambiguousYear, saveHebrewMonth, *workCal, i, numericLeapMonthFormatter, &tzTimeType, mutableNFs, &dayPeriodInt);
2241
2242 if (s == -pos-1) {
2243 // era not present, in special cases allow this to continue
2244 // from the position where the era was expected
2245 s = pos;
2246
2247 if (i+1 < fPattern.length()) {
2248 // move to next pattern character
2249 UChar ch = fPattern.charAt(i+1);
2250
2251 // check for whitespace
2252 if (PatternProps::isWhiteSpace(ch)) {
2253 i++;
2254 // Advance over run in pattern
2255 while ((i+1)<fPattern.length() &&
2256 PatternProps::isWhiteSpace(fPattern.charAt(i+1))) {
2257 ++i;
2258 }
2259 }
2260 }
2261 }
2262 else if (s <= 0) {
2263 status = U_PARSE_ERROR;
2264 goto ExitParse;
2265 }
2266 pos = s;
2267 }
2268 }
2269
2270 // Handle literal pattern characters. These are any
2271 // quoted characters and non-alphabetic unquoted
2272 // characters.
2273 else {
2274
2275 abutPat = -1; // End of any abutting fields
2276
2277 if (! matchLiterals(fPattern, i, text, pos, getBooleanAttribute(UDAT_PARSE_ALLOW_WHITESPACE, status), getBooleanAttribute(UDAT_PARSE_PARTIAL_LITERAL_MATCH, status), isLenient())) {
2278 status = U_PARSE_ERROR;
2279 goto ExitParse;
2280 }
2281 }
2282 }
2283
2284 // Special hack for trailing "." after non-numeric field.
2285 if (text.charAt(pos) == 0x2e && getBooleanAttribute(UDAT_PARSE_ALLOW_WHITESPACE, status)) {
2286 // only do if the last field is not numeric
2287 if (isAfterNonNumericField(fPattern, fPattern.length())) {
2288 pos++; // skip the extra "."
2289 }
2290 }
2291
2292 // If dayPeriod is set, use it in conjunction with hour-of-day to determine am/pm.
2293 if (dayPeriodInt >= 0) {
2294 DayPeriodRules::DayPeriod dayPeriod = (DayPeriodRules::DayPeriod)dayPeriodInt;
2295 const DayPeriodRules *ruleSet = DayPeriodRules::getInstance(this->getSmpFmtLocale(), status);
2296
2297 if (!cal.isSet(UCAL_HOUR) && !cal.isSet(UCAL_HOUR_OF_DAY)) {
2298 // If hour is not set, set time to the midpoint of current day period, overwriting
2299 // minutes if it's set.
2300 double midPoint = ruleSet->getMidPointForDayPeriod(dayPeriod, status);
2301
2302 // If we can't get midPoint we do nothing.
2303 if (U_SUCCESS(status)) {
2304 // Truncate midPoint toward zero to get the hour.
2305 // Any leftover means it was a half-hour.
2306 int32_t midPointHour = (int32_t) midPoint;
2307 int32_t midPointMinute = (midPoint - midPointHour) > 0 ? 30 : 0;
2308
2309 // No need to set am/pm because hour-of-day is set last therefore takes precedence.
2310 cal.set(UCAL_HOUR_OF_DAY, midPointHour);
2311 cal.set(UCAL_MINUTE, midPointMinute);
2312 }
2313 } else {
2314 int hourOfDay;
2315
2316 if (cal.isSet(UCAL_HOUR_OF_DAY)) { // Hour is parsed in 24-hour format.
2317 hourOfDay = cal.get(UCAL_HOUR_OF_DAY, status);
2318 } else { // Hour is parsed in 12-hour format.
2319 hourOfDay = cal.get(UCAL_HOUR, status);
2320 // cal.get() turns 12 to 0 for 12-hour time; change 0 to 12
2321 // so 0 unambiguously means a 24-hour time from above.
2322 if (hourOfDay == 0) { hourOfDay = 12; }
2323 }
2324 U_ASSERT(0 <= hourOfDay && hourOfDay <= 23);
2325
2326
2327 // If hour-of-day is 0 or 13 thru 23 then input time in unambiguously in 24-hour format.
2328 if (hourOfDay == 0 || (13 <= hourOfDay && hourOfDay <= 23)) {
2329 // Make hour-of-day take precedence over (hour + am/pm) by setting it again.
2330 cal.set(UCAL_HOUR_OF_DAY, hourOfDay);
2331 } else {
2332 // We have a 12-hour time and need to choose between am and pm.
2333 // Behave as if dayPeriod spanned 6 hours each way from its center point.
2334 // This will parse correctly for consistent time + period (e.g. 10 at night) as
2335 // well as provide a reasonable recovery for inconsistent time + period (e.g.
2336 // 9 in the afternoon).
2337
2338 // Assume current time is in the AM.
2339 // - Change 12 back to 0 for easier handling of 12am.
2340 // - Append minutes as fractional hours because e.g. 8:15 and 8:45 could be parsed
2341 // into different half-days if center of dayPeriod is at 14:30.
2342 // - cal.get(MINUTE) will return 0 if MINUTE is unset, which works.
2343 if (hourOfDay == 12) { hourOfDay = 0; }
2344 double currentHour = hourOfDay + (cal.get(UCAL_MINUTE, status)) / 60.0;
2345 double midPointHour = ruleSet->getMidPointForDayPeriod(dayPeriod, status);
2346
2347 if (U_SUCCESS(status)) {
2348 double hoursAheadMidPoint = currentHour - midPointHour;
2349
2350 // Assume current time is in the AM.
2351 if (-6 <= hoursAheadMidPoint && hoursAheadMidPoint < 6) {
2352 // Assumption holds; set time as such.
2353 cal.set(UCAL_AM_PM, 0);
2354 } else {
2355 cal.set(UCAL_AM_PM, 1);
2356 }
2357 }
2358 }
2359 }
2360 }
2361
2362 // At this point the fields of Calendar have been set. Calendar
2363 // will fill in default values for missing fields when the time
2364 // is computed.
2365
2366 parsePos.setIndex(pos);
2367
2368 // This part is a problem: When we call parsedDate.after, we compute the time.
2369 // Take the date April 3 2004 at 2:30 am. When this is first set up, the year
2370 // will be wrong if we're parsing a 2-digit year pattern. It will be 1904.
2371 // April 3 1904 is a Sunday (unlike 2004) so it is the DST onset day. 2:30 am
2372 // is therefore an "impossible" time, since the time goes from 1:59 to 3:00 am
2373 // on that day. It is therefore parsed out to fields as 3:30 am. Then we
2374 // add 100 years, and get April 3 2004 at 3:30 am. Note that April 3 2004 is
2375 // a Saturday, so it can have a 2:30 am -- and it should. [LIU]
2376 /*
2377 UDate parsedDate = calendar.getTime();
2378 if( ambiguousYear[0] && !parsedDate.after(fDefaultCenturyStart) ) {
2379 calendar.add(Calendar.YEAR, 100);
2380 parsedDate = calendar.getTime();
2381 }
2382 */
2383 // Because of the above condition, save off the fields in case we need to readjust.
2384 // The procedure we use here is not particularly efficient, but there is no other
2385 // way to do this given the API restrictions present in Calendar. We minimize
2386 // inefficiency by only performing this computation when it might apply, that is,
2387 // when the two-digit year is equal to the start year, and thus might fall at the
2388 // front or the back of the default century. This only works because we adjust
2389 // the year correctly to start with in other cases -- see subParse().
2390 if (ambiguousYear[0] || tzTimeType != UTZFMT_TIME_TYPE_UNKNOWN) // If this is true then the two-digit year == the default start year
2391 {
2392 // We need a copy of the fields, and we need to avoid triggering a call to
2393 // complete(), which will recalculate the fields. Since we can't access
2394 // the fields[] array in Calendar, we clone the entire object. This will
2395 // stop working if Calendar.clone() is ever rewritten to call complete().
2396 Calendar *copy;
2397 if (ambiguousYear[0]) {
2398 copy = cal.clone();
2399 // Check for failed cloning.
2400 if (copy == NULL) {
2401 status = U_MEMORY_ALLOCATION_ERROR;
2402 goto ExitParse;
2403 }
2404 UDate parsedDate = copy->getTime(status);
2405 // {sfb} check internalGetDefaultCenturyStart
2406 if (fHaveDefaultCentury && (parsedDate < fDefaultCenturyStart)) {
2407 // We can't use add here because that does a complete() first.
2408 cal.set(UCAL_YEAR, fDefaultCenturyStartYear + 100);
2409 }
2410 delete copy;
2411 }
2412
2413 if (tzTimeType != UTZFMT_TIME_TYPE_UNKNOWN) {
2414 copy = cal.clone();
2415 // Check for failed cloning.
2416 if (copy == NULL) {
2417 status = U_MEMORY_ALLOCATION_ERROR;
2418 goto ExitParse;
2419 }
2420 const TimeZone & tz = cal.getTimeZone();
2421 BasicTimeZone *btz = NULL;
2422
2423 if (dynamic_cast<const OlsonTimeZone *>(&tz) != NULL
2424 || dynamic_cast<const SimpleTimeZone *>(&tz) != NULL
2425 || dynamic_cast<const RuleBasedTimeZone *>(&tz) != NULL
2426 || dynamic_cast<const VTimeZone *>(&tz) != NULL) {
2427 btz = (BasicTimeZone*)&tz;
2428 }
2429
2430 // Get local millis
2431 copy->set(UCAL_ZONE_OFFSET, 0);
2432 copy->set(UCAL_DST_OFFSET, 0);
2433 UDate localMillis = copy->getTime(status);
2434
2435 // Make sure parsed time zone type (Standard or Daylight)
2436 // matches the rule used by the parsed time zone.
2437 int32_t raw, dst;
2438 if (btz != NULL) {
2439 if (tzTimeType == UTZFMT_TIME_TYPE_STANDARD) {
2440 btz->getOffsetFromLocal(localMillis,
2441 BasicTimeZone::kStandard, BasicTimeZone::kStandard, raw, dst, status);
2442 } else {
2443 btz->getOffsetFromLocal(localMillis,
2444 BasicTimeZone::kDaylight, BasicTimeZone::kDaylight, raw, dst, status);
2445 }
2446 } else {
2447 // No good way to resolve ambiguous time at transition,
2448 // but following code work in most case.
2449 tz.getOffset(localMillis, TRUE, raw, dst, status);
2450 }
2451
2452 // Now, compare the results with parsed type, either standard or daylight saving time
2453 int32_t resolvedSavings = dst;
2454 if (tzTimeType == UTZFMT_TIME_TYPE_STANDARD) {
2455 if (dst != 0) {
2456 // Override DST_OFFSET = 0 in the result calendar
2457 resolvedSavings = 0;
2458 }
2459 } else { // tztype == TZTYPE_DST
2460 if (dst == 0) {
2461 if (btz != NULL) {
2462 UDate time = localMillis + raw;
2463 // We use the nearest daylight saving time rule.
2464 TimeZoneTransition beforeTrs, afterTrs;
2465 UDate beforeT = time, afterT = time;
2466 int32_t beforeSav = 0, afterSav = 0;
2467 UBool beforeTrsAvail, afterTrsAvail;
2468
2469 // Search for DST rule before or on the time
2470 while (TRUE) {
2471 beforeTrsAvail = btz->getPreviousTransition(beforeT, TRUE, beforeTrs);
2472 if (!beforeTrsAvail) {
2473 break;
2474 }
2475 beforeT = beforeTrs.getTime() - 1;
2476 beforeSav = beforeTrs.getFrom()->getDSTSavings();
2477 if (beforeSav != 0) {
2478 break;
2479 }
2480 }
2481
2482 // Search for DST rule after the time
2483 while (TRUE) {
2484 afterTrsAvail = btz->getNextTransition(afterT, FALSE, afterTrs);
2485 if (!afterTrsAvail) {
2486 break;
2487 }
2488 afterT = afterTrs.getTime();
2489 afterSav = afterTrs.getTo()->getDSTSavings();
2490 if (afterSav != 0) {
2491 break;
2492 }
2493 }
2494
2495 if (beforeTrsAvail && afterTrsAvail) {
2496 if (time - beforeT > afterT - time) {
2497 resolvedSavings = afterSav;
2498 } else {
2499 resolvedSavings = beforeSav;
2500 }
2501 } else if (beforeTrsAvail && beforeSav != 0) {
2502 resolvedSavings = beforeSav;
2503 } else if (afterTrsAvail && afterSav != 0) {
2504 resolvedSavings = afterSav;
2505 } else {
2506 resolvedSavings = btz->getDSTSavings();
2507 }
2508 } else {
2509 resolvedSavings = tz.getDSTSavings();
2510 }
2511 if (resolvedSavings == 0) {
2512 // final fallback
2513 resolvedSavings = U_MILLIS_PER_HOUR;
2514 }
2515 }
2516 }
2517 cal.set(UCAL_ZONE_OFFSET, raw);
2518 cal.set(UCAL_DST_OFFSET, resolvedSavings);
2519 delete copy;
2520 }
2521 }
2522 ExitParse:
2523 // Set the parsed result if local calendar is used
2524 // instead of the input calendar
2525 if (U_SUCCESS(status) && workCal != &cal) {
2526 cal.setTimeZone(workCal->getTimeZone());
2527 cal.setTime(workCal->getTime(status), status);
2528 }
2529
2530 if (numericLeapMonthFormatter != NULL) {
2531 delete numericLeapMonthFormatter;
2532 }
2533 if (calClone != NULL) {
2534 delete calClone;
2535 }
2536
2537 // If any Calendar calls failed, we pretend that we
2538 // couldn't parse the string, when in reality this isn't quite accurate--
2539 // we did parse it; the Calendar calls just failed.
2540 if (U_FAILURE(status)) {
2541 parsePos.setErrorIndex(pos);
2542 parsePos.setIndex(start);
2543 }
2544 }
2545
2546 //----------------------------------------------------------------------
2547
2548 static int32_t
2549 matchStringWithOptionalDot(const UnicodeString &text,
2550 int32_t index,
2551 const UnicodeString &data);
2552
matchQuarterString(const UnicodeString & text,int32_t start,UCalendarDateFields field,const UnicodeString * data,int32_t dataCount,Calendar & cal) const2553 int32_t SimpleDateFormat::matchQuarterString(const UnicodeString& text,
2554 int32_t start,
2555 UCalendarDateFields field,
2556 const UnicodeString* data,
2557 int32_t dataCount,
2558 Calendar& cal) const
2559 {
2560 int32_t i = 0;
2561 int32_t count = dataCount;
2562
2563 // There may be multiple strings in the data[] array which begin with
2564 // the same prefix (e.g., Cerven and Cervenec (June and July) in Czech).
2565 // We keep track of the longest match, and return that. Note that this
2566 // unfortunately requires us to test all array elements.
2567 int32_t bestMatchLength = 0, bestMatch = -1;
2568 UnicodeString bestMatchName;
2569
2570 for (; i < count; ++i) {
2571 int32_t matchLength = 0;
2572 if ((matchLength = matchStringWithOptionalDot(text, start, data[i])) > bestMatchLength) {
2573 bestMatchLength = matchLength;
2574 bestMatch = i;
2575 }
2576 }
2577
2578 if (bestMatch >= 0) {
2579 cal.set(field, bestMatch * 3);
2580 return start + bestMatchLength;
2581 }
2582
2583 return -start;
2584 }
2585
matchDayPeriodStrings(const UnicodeString & text,int32_t start,const UnicodeString * data,int32_t dataCount,int32_t & dayPeriod) const2586 int32_t SimpleDateFormat::matchDayPeriodStrings(const UnicodeString& text, int32_t start,
2587 const UnicodeString* data, int32_t dataCount,
2588 int32_t &dayPeriod) const
2589 {
2590
2591 int32_t bestMatchLength = 0, bestMatch = -1;
2592
2593 for (int32_t i = 0; i < dataCount; ++i) {
2594 int32_t matchLength = 0;
2595 if ((matchLength = matchStringWithOptionalDot(text, start, data[i])) > bestMatchLength) {
2596 bestMatchLength = matchLength;
2597 bestMatch = i;
2598 }
2599 }
2600
2601 if (bestMatch >= 0) {
2602 dayPeriod = bestMatch;
2603 return start + bestMatchLength;
2604 }
2605
2606 return -start;
2607 }
2608
2609 //----------------------------------------------------------------------
matchLiterals(const UnicodeString & pattern,int32_t & patternOffset,const UnicodeString & text,int32_t & textOffset,UBool whitespaceLenient,UBool partialMatchLenient,UBool oldLeniency)2610 UBool SimpleDateFormat::matchLiterals(const UnicodeString &pattern,
2611 int32_t &patternOffset,
2612 const UnicodeString &text,
2613 int32_t &textOffset,
2614 UBool whitespaceLenient,
2615 UBool partialMatchLenient,
2616 UBool oldLeniency)
2617 {
2618 UBool inQuote = FALSE;
2619 UnicodeString literal;
2620 int32_t i = patternOffset;
2621
2622 // scan pattern looking for contiguous literal characters
2623 for ( ; i < pattern.length(); i += 1) {
2624 UChar ch = pattern.charAt(i);
2625
2626 if (!inQuote && isSyntaxChar(ch)) {
2627 break;
2628 }
2629
2630 if (ch == QUOTE) {
2631 // Match a quote literal ('') inside OR outside of quotes
2632 if ((i + 1) < pattern.length() && pattern.charAt(i + 1) == QUOTE) {
2633 i += 1;
2634 } else {
2635 inQuote = !inQuote;
2636 continue;
2637 }
2638 }
2639
2640 literal += ch;
2641 }
2642
2643 // at this point, literal contains the literal text
2644 // and i is the index of the next non-literal pattern character.
2645 int32_t p;
2646 int32_t t = textOffset;
2647
2648 if (whitespaceLenient) {
2649 // trim leading, trailing whitespace from
2650 // the literal text
2651 literal.trim();
2652
2653 // ignore any leading whitespace in the text
2654 while (t < text.length() && u_isWhitespace(text.charAt(t))) {
2655 t += 1;
2656 }
2657 }
2658
2659 for (p = 0; p < literal.length() && t < text.length();) {
2660 UBool needWhitespace = FALSE;
2661
2662 while (p < literal.length() && PatternProps::isWhiteSpace(literal.charAt(p))) {
2663 needWhitespace = TRUE;
2664 p += 1;
2665 }
2666
2667 if (needWhitespace) {
2668 int32_t tStart = t;
2669
2670 while (t < text.length()) {
2671 UChar tch = text.charAt(t);
2672
2673 if (!u_isUWhiteSpace(tch) && !PatternProps::isWhiteSpace(tch)) {
2674 break;
2675 }
2676
2677 t += 1;
2678 }
2679
2680 // TODO: should we require internal spaces
2681 // in lenient mode? (There won't be any
2682 // leading or trailing spaces)
2683 if (!whitespaceLenient && t == tStart) {
2684 // didn't find matching whitespace:
2685 // an error in strict mode
2686 return FALSE;
2687 }
2688
2689 // In strict mode, this run of whitespace
2690 // may have been at the end.
2691 if (p >= literal.length()) {
2692 break;
2693 }
2694 }
2695 if (t >= text.length() || literal.charAt(p) != text.charAt(t)) {
2696 // Ran out of text, or found a non-matching character:
2697 // OK in lenient mode, an error in strict mode.
2698 if (whitespaceLenient) {
2699 if (t == textOffset && text.charAt(t) == 0x2e &&
2700 isAfterNonNumericField(pattern, patternOffset)) {
2701 // Lenient mode and the literal input text begins with a "." and
2702 // we are after a non-numeric field: We skip the "."
2703 ++t;
2704 continue; // Do not update p.
2705 }
2706 // if it is actual whitespace and we're whitespace lenient it's OK
2707
2708 UChar wsc = text.charAt(t);
2709 if(PatternProps::isWhiteSpace(wsc)) {
2710 // Lenient mode and it's just whitespace we skip it
2711 ++t;
2712 continue; // Do not update p.
2713 }
2714 }
2715 // hack around oldleniency being a bit of a catch-all bucket and we're just adding support specifically for paritial matches
2716 if(partialMatchLenient && oldLeniency) {
2717 break;
2718 }
2719
2720 return FALSE;
2721 }
2722 ++p;
2723 ++t;
2724 }
2725
2726 // At this point if we're in strict mode we have a complete match.
2727 // If we're in lenient mode we may have a partial match, or no
2728 // match at all.
2729 if (p <= 0) {
2730 // no match. Pretend it matched a run of whitespace
2731 // and ignorables in the text.
2732 const UnicodeSet *ignorables = NULL;
2733 UDateFormatField patternCharIndex = DateFormatSymbols::getPatternCharIndex(pattern.charAt(i));
2734 if (patternCharIndex != UDAT_FIELD_COUNT) {
2735 ignorables = SimpleDateFormatStaticSets::getIgnorables(patternCharIndex);
2736 }
2737
2738 for (t = textOffset; t < text.length(); t += 1) {
2739 UChar ch = text.charAt(t);
2740
2741 if (ignorables == NULL || !ignorables->contains(ch)) {
2742 break;
2743 }
2744 }
2745 }
2746
2747 // if we get here, we've got a complete match.
2748 patternOffset = i - 1;
2749 textOffset = t;
2750
2751 return TRUE;
2752 }
2753
2754 //----------------------------------------------------------------------
2755
matchString(const UnicodeString & text,int32_t start,UCalendarDateFields field,const UnicodeString * data,int32_t dataCount,const UnicodeString * monthPattern,Calendar & cal) const2756 int32_t SimpleDateFormat::matchString(const UnicodeString& text,
2757 int32_t start,
2758 UCalendarDateFields field,
2759 const UnicodeString* data,
2760 int32_t dataCount,
2761 const UnicodeString* monthPattern,
2762 Calendar& cal) const
2763 {
2764 int32_t i = 0;
2765 int32_t count = dataCount;
2766
2767 if (field == UCAL_DAY_OF_WEEK) i = 1;
2768
2769 // There may be multiple strings in the data[] array which begin with
2770 // the same prefix (e.g., Cerven and Cervenec (June and July) in Czech).
2771 // We keep track of the longest match, and return that. Note that this
2772 // unfortunately requires us to test all array elements.
2773 int32_t bestMatchLength = 0, bestMatch = -1;
2774 UnicodeString bestMatchName;
2775 int32_t isLeapMonth = 0;
2776
2777 for (; i < count; ++i) {
2778 int32_t matchLen = 0;
2779 if ((matchLen = matchStringWithOptionalDot(text, start, data[i])) > bestMatchLength) {
2780 bestMatch = i;
2781 bestMatchLength = matchLen;
2782 }
2783
2784 if (monthPattern != NULL) {
2785 UErrorCode status = U_ZERO_ERROR;
2786 UnicodeString leapMonthName;
2787 SimpleFormatter(*monthPattern, 1, 1, status).format(data[i], leapMonthName, status);
2788 if (U_SUCCESS(status)) {
2789 if ((matchLen = matchStringWithOptionalDot(text, start, leapMonthName)) > bestMatchLength) {
2790 bestMatch = i;
2791 bestMatchLength = matchLen;
2792 isLeapMonth = 1;
2793 }
2794 }
2795 }
2796 }
2797
2798 if (bestMatch >= 0) {
2799 if (field < UCAL_FIELD_COUNT) {
2800 // Adjustment for Hebrew Calendar month Adar II
2801 if (!strcmp(cal.getType(),"hebrew") && field==UCAL_MONTH && bestMatch==13) {
2802 cal.set(field,6);
2803 } else {
2804 if (field == UCAL_YEAR) {
2805 bestMatch++; // only get here for cyclic year names, which match 1-based years 1-60
2806 }
2807 cal.set(field, bestMatch);
2808 }
2809 if (monthPattern != NULL) {
2810 cal.set(UCAL_IS_LEAP_MONTH, isLeapMonth);
2811 }
2812 }
2813
2814 return start + bestMatchLength;
2815 }
2816
2817 return -start;
2818 }
2819
2820 static int32_t
matchStringWithOptionalDot(const UnicodeString & text,int32_t index,const UnicodeString & data)2821 matchStringWithOptionalDot(const UnicodeString &text,
2822 int32_t index,
2823 const UnicodeString &data) {
2824 UErrorCode sts = U_ZERO_ERROR;
2825 int32_t matchLenText = 0;
2826 int32_t matchLenData = 0;
2827
2828 u_caseInsensitivePrefixMatch(text.getBuffer() + index, text.length() - index,
2829 data.getBuffer(), data.length(),
2830 0 /* default case option */,
2831 &matchLenText, &matchLenData,
2832 &sts);
2833 U_ASSERT (U_SUCCESS(sts));
2834
2835 if (matchLenData == data.length() /* normal match */
2836 || (data.charAt(data.length() - 1) == 0x2e
2837 && matchLenData == data.length() - 1 /* match without trailing dot */)) {
2838 return matchLenText;
2839 }
2840
2841 return 0;
2842 }
2843
2844 //----------------------------------------------------------------------
2845
2846 void
set2DigitYearStart(UDate d,UErrorCode & status)2847 SimpleDateFormat::set2DigitYearStart(UDate d, UErrorCode& status)
2848 {
2849 parseAmbiguousDatesAsAfter(d, status);
2850 }
2851
2852 /**
2853 * Private member function that converts the parsed date strings into
2854 * timeFields. Returns -start (for ParsePosition) if failed.
2855 */
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,SimpleDateFormatMutableNFs & mutableNFs,int32_t * dayPeriod) const2856 int32_t SimpleDateFormat::subParse(const UnicodeString& text, int32_t& start, UChar ch, int32_t count,
2857 UBool obeyCount, UBool allowNegative, UBool ambiguousYear[], int32_t& saveHebrewMonth, Calendar& cal,
2858 int32_t patLoc, MessageFormat * numericLeapMonthFormatter, UTimeZoneFormatTimeType *tzTimeType, SimpleDateFormatMutableNFs &mutableNFs,
2859 int32_t *dayPeriod) const
2860 {
2861 Formattable number;
2862 int32_t value = 0;
2863 int32_t i;
2864 int32_t ps = 0;
2865 UErrorCode status = U_ZERO_ERROR;
2866 ParsePosition pos(0);
2867 UDateFormatField patternCharIndex = DateFormatSymbols::getPatternCharIndex(ch);
2868 NumberFormat *currentNumberFormat;
2869 UnicodeString temp;
2870 UBool gotNumber = FALSE;
2871
2872 #if defined (U_DEBUG_CAL)
2873 //fprintf(stderr, "%s:%d - [%c] st=%d \n", __FILE__, __LINE__, (char) ch, start);
2874 #endif
2875
2876 if (patternCharIndex == UDAT_FIELD_COUNT) {
2877 return -start;
2878 }
2879
2880 currentNumberFormat = mutableNFs.get(getNumberFormatByIndex(patternCharIndex));
2881 if (currentNumberFormat == NULL) {
2882 return -start;
2883 }
2884 UCalendarDateFields field = fgPatternIndexToCalendarField[patternCharIndex]; // UCAL_FIELD_COUNT if irrelevant
2885 UnicodeString hebr("hebr", 4, US_INV);
2886
2887 if (numericLeapMonthFormatter != NULL) {
2888 numericLeapMonthFormatter->setFormats((const Format **)¤tNumberFormat, 1);
2889 }
2890 UBool isChineseCalendar = (uprv_strcmp(cal.getType(),"chinese") == 0 || uprv_strcmp(cal.getType(),"dangi") == 0);
2891
2892 // If there are any spaces here, skip over them. If we hit the end
2893 // of the string, then fail.
2894 for (;;) {
2895 if (start >= text.length()) {
2896 return -start;
2897 }
2898 UChar32 c = text.char32At(start);
2899 if (!u_isUWhiteSpace(c) /*||*/ && !PatternProps::isWhiteSpace(c)) {
2900 break;
2901 }
2902 start += U16_LENGTH(c);
2903 }
2904 pos.setIndex(start);
2905
2906 // We handle a few special cases here where we need to parse
2907 // a number value. We handle further, more generic cases below. We need
2908 // to handle some of them here because some fields require extra processing on
2909 // the parsed value.
2910 if (patternCharIndex == UDAT_HOUR_OF_DAY1_FIELD || // k
2911 patternCharIndex == UDAT_HOUR_OF_DAY0_FIELD || // H
2912 patternCharIndex == UDAT_HOUR1_FIELD || // h
2913 patternCharIndex == UDAT_HOUR0_FIELD || // K
2914 (patternCharIndex == UDAT_DOW_LOCAL_FIELD && count <= 2) || // e
2915 (patternCharIndex == UDAT_STANDALONE_DAY_FIELD && count <= 2) || // c
2916 (patternCharIndex == UDAT_MONTH_FIELD && count <= 2) || // M
2917 (patternCharIndex == UDAT_STANDALONE_MONTH_FIELD && count <= 2) || // L
2918 (patternCharIndex == UDAT_QUARTER_FIELD && count <= 2) || // Q
2919 (patternCharIndex == UDAT_STANDALONE_QUARTER_FIELD && count <= 2) || // q
2920 patternCharIndex == UDAT_YEAR_FIELD || // y
2921 patternCharIndex == UDAT_YEAR_WOY_FIELD || // Y
2922 patternCharIndex == UDAT_YEAR_NAME_FIELD || // U (falls back to numeric)
2923 (patternCharIndex == UDAT_ERA_FIELD && isChineseCalendar) || // G
2924 patternCharIndex == UDAT_FRACTIONAL_SECOND_FIELD) // S
2925 {
2926 int32_t parseStart = pos.getIndex();
2927 // It would be good to unify this with the obeyCount logic below,
2928 // but that's going to be difficult.
2929 const UnicodeString* src;
2930
2931 UBool parsedNumericLeapMonth = FALSE;
2932 if (numericLeapMonthFormatter != NULL && (patternCharIndex == UDAT_MONTH_FIELD || patternCharIndex == UDAT_STANDALONE_MONTH_FIELD)) {
2933 int32_t argCount;
2934 Formattable * args = numericLeapMonthFormatter->parse(text, pos, argCount);
2935 if (args != NULL && argCount == 1 && pos.getIndex() > parseStart && args[0].isNumeric()) {
2936 parsedNumericLeapMonth = TRUE;
2937 number.setLong(args[0].getLong());
2938 cal.set(UCAL_IS_LEAP_MONTH, 1);
2939 delete[] args;
2940 } else {
2941 pos.setIndex(parseStart);
2942 cal.set(UCAL_IS_LEAP_MONTH, 0);
2943 }
2944 }
2945
2946 if (!parsedNumericLeapMonth) {
2947 if (obeyCount) {
2948 if ((start+count) > text.length()) {
2949 return -start;
2950 }
2951
2952 text.extractBetween(0, start + count, temp);
2953 src = &temp;
2954 } else {
2955 src = &text;
2956 }
2957
2958 parseInt(*src, number, pos, allowNegative,currentNumberFormat);
2959 }
2960
2961 int32_t txtLoc = pos.getIndex();
2962
2963 if (txtLoc > parseStart) {
2964 value = number.getLong();
2965 gotNumber = TRUE;
2966
2967 // suffix processing
2968 if (value < 0 ) {
2969 txtLoc = checkIntSuffix(text, txtLoc, patLoc+1, TRUE);
2970 if (txtLoc != pos.getIndex()) {
2971 value *= -1;
2972 }
2973 }
2974 else {
2975 txtLoc = checkIntSuffix(text, txtLoc, patLoc+1, FALSE);
2976 }
2977
2978 if (!getBooleanAttribute(UDAT_PARSE_ALLOW_WHITESPACE, status)) {
2979 // Check the range of the value
2980 int32_t bias = gFieldRangeBias[patternCharIndex];
2981 if (bias >= 0 && (value > cal.getMaximum(field) + bias || value < cal.getMinimum(field) + bias)) {
2982 return -start;
2983 }
2984 }
2985
2986 pos.setIndex(txtLoc);
2987 }
2988 }
2989
2990 // Make sure that we got a number if
2991 // we want one, and didn't get one
2992 // if we don't want one.
2993 switch (patternCharIndex) {
2994 case UDAT_HOUR_OF_DAY1_FIELD:
2995 case UDAT_HOUR_OF_DAY0_FIELD:
2996 case UDAT_HOUR1_FIELD:
2997 case UDAT_HOUR0_FIELD:
2998 // special range check for hours:
2999 if (value < 0 || value > 24) {
3000 return -start;
3001 }
3002
3003 // fall through to gotNumber check
3004 U_FALLTHROUGH;
3005 case UDAT_YEAR_FIELD:
3006 case UDAT_YEAR_WOY_FIELD:
3007 case UDAT_FRACTIONAL_SECOND_FIELD:
3008 // these must be a number
3009 if (! gotNumber) {
3010 return -start;
3011 }
3012
3013 break;
3014
3015 default:
3016 // we check the rest of the fields below.
3017 break;
3018 }
3019
3020 switch (patternCharIndex) {
3021 case UDAT_ERA_FIELD:
3022 if (isChineseCalendar) {
3023 if (!gotNumber) {
3024 return -start;
3025 }
3026 cal.set(UCAL_ERA, value);
3027 return pos.getIndex();
3028 }
3029 if (count == 5) {
3030 ps = matchString(text, start, UCAL_ERA, fSymbols->fNarrowEras, fSymbols->fNarrowErasCount, NULL, cal);
3031 } else if (count == 4) {
3032 ps = matchString(text, start, UCAL_ERA, fSymbols->fEraNames, fSymbols->fEraNamesCount, NULL, cal);
3033 } else {
3034 ps = matchString(text, start, UCAL_ERA, fSymbols->fEras, fSymbols->fErasCount, NULL, cal);
3035 }
3036
3037 // check return position, if it equals -start, then matchString error
3038 // special case the return code so we don't necessarily fail out until we
3039 // verify no year information also
3040 if (ps == -start)
3041 ps--;
3042
3043 return ps;
3044
3045 case UDAT_YEAR_FIELD:
3046 // If there are 3 or more YEAR pattern characters, this indicates
3047 // that the year value is to be treated literally, without any
3048 // two-digit year adjustments (e.g., from "01" to 2001). Otherwise
3049 // we made adjustments to place the 2-digit year in the proper
3050 // century, for parsed strings from "00" to "99". Any other string
3051 // is treated literally: "2250", "-1", "1", "002".
3052 if (fDateOverride.compare(hebr)==0 && value < 1000) {
3053 value += HEBREW_CAL_CUR_MILLENIUM_START_YEAR;
3054 } else if ((pos.getIndex() - start) == 2 && !isChineseCalendar
3055 && u_isdigit(text.charAt(start))
3056 && u_isdigit(text.charAt(start+1)))
3057 {
3058 // only adjust year for patterns less than 3.
3059 if(count < 3) {
3060 // Assume for example that the defaultCenturyStart is 6/18/1903.
3061 // This means that two-digit years will be forced into the range
3062 // 6/18/1903 to 6/17/2003. As a result, years 00, 01, and 02
3063 // correspond to 2000, 2001, and 2002. Years 04, 05, etc. correspond
3064 // to 1904, 1905, etc. If the year is 03, then it is 2003 if the
3065 // other fields specify a date before 6/18, or 1903 if they specify a
3066 // date afterwards. As a result, 03 is an ambiguous year. All other
3067 // two-digit years are unambiguous.
3068 if(fHaveDefaultCentury) { // check if this formatter even has a pivot year
3069 int32_t ambiguousTwoDigitYear = fDefaultCenturyStartYear % 100;
3070 ambiguousYear[0] = (value == ambiguousTwoDigitYear);
3071 value += (fDefaultCenturyStartYear/100)*100 +
3072 (value < ambiguousTwoDigitYear ? 100 : 0);
3073 }
3074 }
3075 }
3076 cal.set(UCAL_YEAR, value);
3077
3078 // Delayed checking for adjustment of Hebrew month numbers in non-leap years.
3079 if (saveHebrewMonth >= 0) {
3080 HebrewCalendar *hc = (HebrewCalendar*)&cal;
3081 if (!hc->isLeapYear(value) && saveHebrewMonth >= 6) {
3082 cal.set(UCAL_MONTH,saveHebrewMonth);
3083 } else {
3084 cal.set(UCAL_MONTH,saveHebrewMonth-1);
3085 }
3086 saveHebrewMonth = -1;
3087 }
3088 return pos.getIndex();
3089
3090 case UDAT_YEAR_WOY_FIELD:
3091 // Comment is the same as for UDAT_Year_FIELDs - look above
3092 if (fDateOverride.compare(hebr)==0 && value < 1000) {
3093 value += HEBREW_CAL_CUR_MILLENIUM_START_YEAR;
3094 } else if ((pos.getIndex() - start) == 2
3095 && u_isdigit(text.charAt(start))
3096 && u_isdigit(text.charAt(start+1))
3097 && fHaveDefaultCentury )
3098 {
3099 int32_t ambiguousTwoDigitYear = fDefaultCenturyStartYear % 100;
3100 ambiguousYear[0] = (value == ambiguousTwoDigitYear);
3101 value += (fDefaultCenturyStartYear/100)*100 +
3102 (value < ambiguousTwoDigitYear ? 100 : 0);
3103 }
3104 cal.set(UCAL_YEAR_WOY, value);
3105 return pos.getIndex();
3106
3107 case UDAT_YEAR_NAME_FIELD:
3108 if (fSymbols->fShortYearNames != NULL) {
3109 int32_t newStart = matchString(text, start, UCAL_YEAR, fSymbols->fShortYearNames, fSymbols->fShortYearNamesCount, NULL, cal);
3110 if (newStart > 0) {
3111 return newStart;
3112 }
3113 }
3114 if (gotNumber && (getBooleanAttribute(UDAT_PARSE_ALLOW_NUMERIC,status) || value > fSymbols->fShortYearNamesCount)) {
3115 cal.set(UCAL_YEAR, value);
3116 return pos.getIndex();
3117 }
3118 return -start;
3119
3120 case UDAT_MONTH_FIELD:
3121 case UDAT_STANDALONE_MONTH_FIELD:
3122 if (gotNumber) // i.e., M or MM.
3123 {
3124 // When parsing month numbers from the Hebrew Calendar, we might need to adjust the month depending on whether
3125 // 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
3126 // the year is parsed.
3127 if (!strcmp(cal.getType(),"hebrew")) {
3128 HebrewCalendar *hc = (HebrewCalendar*)&cal;
3129 if (cal.isSet(UCAL_YEAR)) {
3130 UErrorCode status = U_ZERO_ERROR;
3131 if (!hc->isLeapYear(hc->get(UCAL_YEAR,status)) && value >= 6) {
3132 cal.set(UCAL_MONTH, value);
3133 } else {
3134 cal.set(UCAL_MONTH, value - 1);
3135 }
3136 } else {
3137 saveHebrewMonth = value;
3138 }
3139 } else {
3140 // Don't want to parse the month if it is a string
3141 // while pattern uses numeric style: M/MM, L/LL
3142 // [We computed 'value' above.]
3143 cal.set(UCAL_MONTH, value - 1);
3144 }
3145 return pos.getIndex();
3146 } else {
3147 // count >= 3 // i.e., MMM/MMMM, LLL/LLLL
3148 // Want to be able to parse both short and long forms.
3149 // Try count == 4 first:
3150 UnicodeString * wideMonthPat = NULL;
3151 UnicodeString * shortMonthPat = NULL;
3152 if (fSymbols->fLeapMonthPatterns != NULL && fSymbols->fLeapMonthPatternsCount >= DateFormatSymbols::kMonthPatternsCount) {
3153 if (patternCharIndex==UDAT_MONTH_FIELD) {
3154 wideMonthPat = &fSymbols->fLeapMonthPatterns[DateFormatSymbols::kLeapMonthPatternFormatWide];
3155 shortMonthPat = &fSymbols->fLeapMonthPatterns[DateFormatSymbols::kLeapMonthPatternFormatAbbrev];
3156 } else {
3157 wideMonthPat = &fSymbols->fLeapMonthPatterns[DateFormatSymbols::kLeapMonthPatternStandaloneWide];
3158 shortMonthPat = &fSymbols->fLeapMonthPatterns[DateFormatSymbols::kLeapMonthPatternStandaloneAbbrev];
3159 }
3160 }
3161 int32_t newStart = 0;
3162 if (patternCharIndex==UDAT_MONTH_FIELD) {
3163 if(getBooleanAttribute(UDAT_PARSE_MULTIPLE_PATTERNS_FOR_MATCH, status) || count == 4) {
3164 newStart = matchString(text, start, UCAL_MONTH, fSymbols->fMonths, fSymbols->fMonthsCount, wideMonthPat, cal); // try MMMM
3165 if (newStart > 0) {
3166 return newStart;
3167 }
3168 }
3169 if(getBooleanAttribute(UDAT_PARSE_MULTIPLE_PATTERNS_FOR_MATCH, status) || count == 3) {
3170 newStart = matchString(text, start, UCAL_MONTH, fSymbols->fShortMonths, fSymbols->fShortMonthsCount, shortMonthPat, cal); // try MMM
3171 }
3172 } else {
3173 if(getBooleanAttribute(UDAT_PARSE_MULTIPLE_PATTERNS_FOR_MATCH, status) || count == 4) {
3174 newStart = matchString(text, start, UCAL_MONTH, fSymbols->fStandaloneMonths, fSymbols->fStandaloneMonthsCount, wideMonthPat, cal); // try LLLL
3175 if (newStart > 0) {
3176 return newStart;
3177 }
3178 }
3179 if(getBooleanAttribute(UDAT_PARSE_MULTIPLE_PATTERNS_FOR_MATCH, status) || count == 3) {
3180 newStart = matchString(text, start, UCAL_MONTH, fSymbols->fStandaloneShortMonths, fSymbols->fStandaloneShortMonthsCount, shortMonthPat, cal); // try LLL
3181 }
3182 }
3183 if (newStart > 0 || !getBooleanAttribute(UDAT_PARSE_ALLOW_NUMERIC, status)) // currently we do not try to parse MMMMM/LLLLL: #8860
3184 return newStart;
3185 // else we allowing parsing as number, below
3186 }
3187 break;
3188
3189 case UDAT_HOUR_OF_DAY1_FIELD:
3190 // [We computed 'value' above.]
3191 if (value == cal.getMaximum(UCAL_HOUR_OF_DAY) + 1)
3192 value = 0;
3193
3194 // fall through to set field
3195 U_FALLTHROUGH;
3196 case UDAT_HOUR_OF_DAY0_FIELD:
3197 cal.set(UCAL_HOUR_OF_DAY, value);
3198 return pos.getIndex();
3199
3200 case UDAT_FRACTIONAL_SECOND_FIELD:
3201 // Fractional seconds left-justify
3202 i = pos.getIndex() - start;
3203 if (i < 3) {
3204 while (i < 3) {
3205 value *= 10;
3206 i++;
3207 }
3208 } else {
3209 int32_t a = 1;
3210 while (i > 3) {
3211 a *= 10;
3212 i--;
3213 }
3214 value /= a;
3215 }
3216 cal.set(UCAL_MILLISECOND, value);
3217 return pos.getIndex();
3218
3219 case UDAT_DOW_LOCAL_FIELD:
3220 if (gotNumber) // i.e., e or ee
3221 {
3222 // [We computed 'value' above.]
3223 cal.set(UCAL_DOW_LOCAL, value);
3224 return pos.getIndex();
3225 }
3226 // else for eee-eeeee fall through to handling of EEE-EEEEE
3227 // fall through, do not break here
3228 U_FALLTHROUGH;
3229 case UDAT_DAY_OF_WEEK_FIELD:
3230 {
3231 // Want to be able to parse both short and long forms.
3232 // Try count == 4 (EEEE) wide first:
3233 int32_t newStart = 0;
3234 if(getBooleanAttribute(UDAT_PARSE_MULTIPLE_PATTERNS_FOR_MATCH, status) || count == 4) {
3235 if ((newStart = matchString(text, start, UCAL_DAY_OF_WEEK,
3236 fSymbols->fWeekdays, fSymbols->fWeekdaysCount, NULL, cal)) > 0)
3237 return newStart;
3238 }
3239 // EEEE wide failed, now try EEE abbreviated
3240 if(getBooleanAttribute(UDAT_PARSE_MULTIPLE_PATTERNS_FOR_MATCH, status) || count == 3) {
3241 if ((newStart = matchString(text, start, UCAL_DAY_OF_WEEK,
3242 fSymbols->fShortWeekdays, fSymbols->fShortWeekdaysCount, NULL, cal)) > 0)
3243 return newStart;
3244 }
3245 // EEE abbreviated failed, now try EEEEEE short
3246 if(getBooleanAttribute(UDAT_PARSE_MULTIPLE_PATTERNS_FOR_MATCH, status) || count == 6) {
3247 if ((newStart = matchString(text, start, UCAL_DAY_OF_WEEK,
3248 fSymbols->fShorterWeekdays, fSymbols->fShorterWeekdaysCount, NULL, cal)) > 0)
3249 return newStart;
3250 }
3251 // EEEEEE short failed, now try EEEEE narrow
3252 if(getBooleanAttribute(UDAT_PARSE_MULTIPLE_PATTERNS_FOR_MATCH, status) || count == 5) {
3253 if ((newStart = matchString(text, start, UCAL_DAY_OF_WEEK,
3254 fSymbols->fNarrowWeekdays, fSymbols->fNarrowWeekdaysCount, NULL, cal)) > 0)
3255 return newStart;
3256 }
3257 if (!getBooleanAttribute(UDAT_PARSE_ALLOW_NUMERIC, status) || patternCharIndex == UDAT_DAY_OF_WEEK_FIELD)
3258 return newStart;
3259 // else we allowing parsing as number, below
3260 }
3261 break;
3262
3263 case UDAT_STANDALONE_DAY_FIELD:
3264 {
3265 if (gotNumber) // c or cc
3266 {
3267 // [We computed 'value' above.]
3268 cal.set(UCAL_DOW_LOCAL, value);
3269 return pos.getIndex();
3270 }
3271 // Want to be able to parse both short and long forms.
3272 // Try count == 4 (cccc) first:
3273 int32_t newStart = 0;
3274 if(getBooleanAttribute(UDAT_PARSE_MULTIPLE_PATTERNS_FOR_MATCH, status) || count == 4) {
3275 if ((newStart = matchString(text, start, UCAL_DAY_OF_WEEK,
3276 fSymbols->fStandaloneWeekdays, fSymbols->fStandaloneWeekdaysCount, NULL, cal)) > 0)
3277 return newStart;
3278 }
3279 if(getBooleanAttribute(UDAT_PARSE_MULTIPLE_PATTERNS_FOR_MATCH, status) || count == 3) {
3280 if ((newStart = matchString(text, start, UCAL_DAY_OF_WEEK,
3281 fSymbols->fStandaloneShortWeekdays, fSymbols->fStandaloneShortWeekdaysCount, NULL, cal)) > 0)
3282 return newStart;
3283 }
3284 if(getBooleanAttribute(UDAT_PARSE_MULTIPLE_PATTERNS_FOR_MATCH, status) || count == 6) {
3285 if ((newStart = matchString(text, start, UCAL_DAY_OF_WEEK,
3286 fSymbols->fStandaloneShorterWeekdays, fSymbols->fStandaloneShorterWeekdaysCount, NULL, cal)) > 0)
3287 return newStart;
3288 }
3289 if (!getBooleanAttribute(UDAT_PARSE_ALLOW_NUMERIC, status))
3290 return newStart;
3291 // else we allowing parsing as number, below
3292 }
3293 break;
3294
3295 case UDAT_AM_PM_FIELD:
3296 {
3297 // optionally try both wide/abbrev and narrow forms
3298 int32_t newStart = 0;
3299 // try wide/abbrev
3300 if( getBooleanAttribute(UDAT_PARSE_MULTIPLE_PATTERNS_FOR_MATCH, status) || count < 5 ) {
3301 if ((newStart = matchString(text, start, UCAL_AM_PM, fSymbols->fAmPms, fSymbols->fAmPmsCount, NULL, cal)) > 0) {
3302 return newStart;
3303 }
3304 }
3305 // try narrow
3306 if( getBooleanAttribute(UDAT_PARSE_MULTIPLE_PATTERNS_FOR_MATCH, status) || count >= 5 ) {
3307 if ((newStart = matchString(text, start, UCAL_AM_PM, fSymbols->fNarrowAmPms, fSymbols->fNarrowAmPmsCount, NULL, cal)) > 0) {
3308 return newStart;
3309 }
3310 }
3311 // no matches for given options
3312 return -start;
3313 }
3314
3315 case UDAT_HOUR1_FIELD:
3316 // [We computed 'value' above.]
3317 if (value == cal.getLeastMaximum(UCAL_HOUR)+1)
3318 value = 0;
3319
3320 // fall through to set field
3321 U_FALLTHROUGH;
3322 case UDAT_HOUR0_FIELD:
3323 cal.set(UCAL_HOUR, value);
3324 return pos.getIndex();
3325
3326 case UDAT_QUARTER_FIELD:
3327 if (gotNumber) // i.e., Q or QQ.
3328 {
3329 // Don't want to parse the month if it is a string
3330 // while pattern uses numeric style: Q or QQ.
3331 // [We computed 'value' above.]
3332 cal.set(UCAL_MONTH, (value - 1) * 3);
3333 return pos.getIndex();
3334 } else {
3335 // count >= 3 // i.e., QQQ or QQQQ
3336 // Want to be able to parse both short and long forms.
3337 // Try count == 4 first:
3338 int32_t newStart = 0;
3339
3340 if(getBooleanAttribute(UDAT_PARSE_MULTIPLE_PATTERNS_FOR_MATCH, status) || count == 4) {
3341 if ((newStart = matchQuarterString(text, start, UCAL_MONTH,
3342 fSymbols->fQuarters, fSymbols->fQuartersCount, cal)) > 0)
3343 return newStart;
3344 }
3345 if(getBooleanAttribute(UDAT_PARSE_MULTIPLE_PATTERNS_FOR_MATCH, status) || count == 3) {
3346 if ((newStart = matchQuarterString(text, start, UCAL_MONTH,
3347 fSymbols->fShortQuarters, fSymbols->fShortQuartersCount, cal)) > 0)
3348 return newStart;
3349 }
3350 if (!getBooleanAttribute(UDAT_PARSE_ALLOW_NUMERIC, status))
3351 return newStart;
3352 // else we allowing parsing as number, below
3353 if(!getBooleanAttribute(UDAT_PARSE_MULTIPLE_PATTERNS_FOR_MATCH, status))
3354 return -start;
3355 }
3356 break;
3357
3358 case UDAT_STANDALONE_QUARTER_FIELD:
3359 if (gotNumber) // i.e., q or qq.
3360 {
3361 // Don't want to parse the month if it is a string
3362 // while pattern uses numeric style: q or q.
3363 // [We computed 'value' above.]
3364 cal.set(UCAL_MONTH, (value - 1) * 3);
3365 return pos.getIndex();
3366 } else {
3367 // count >= 3 // i.e., qqq or qqqq
3368 // Want to be able to parse both short and long forms.
3369 // Try count == 4 first:
3370 int32_t newStart = 0;
3371
3372 if(getBooleanAttribute(UDAT_PARSE_MULTIPLE_PATTERNS_FOR_MATCH, status) || count == 4) {
3373 if ((newStart = matchQuarterString(text, start, UCAL_MONTH,
3374 fSymbols->fStandaloneQuarters, fSymbols->fStandaloneQuartersCount, cal)) > 0)
3375 return newStart;
3376 }
3377 if(getBooleanAttribute(UDAT_PARSE_MULTIPLE_PATTERNS_FOR_MATCH, status) || count == 3) {
3378 if ((newStart = matchQuarterString(text, start, UCAL_MONTH,
3379 fSymbols->fStandaloneShortQuarters, fSymbols->fStandaloneShortQuartersCount, cal)) > 0)
3380 return newStart;
3381 }
3382 if (!getBooleanAttribute(UDAT_PARSE_ALLOW_NUMERIC, status))
3383 return newStart;
3384 // else we allowing parsing as number, below
3385 if(!getBooleanAttribute(UDAT_PARSE_MULTIPLE_PATTERNS_FOR_MATCH, status))
3386 return -start;
3387 }
3388 break;
3389
3390 case UDAT_TIMEZONE_FIELD: // 'z'
3391 {
3392 UTimeZoneFormatStyle style = (count < 4) ? UTZFMT_STYLE_SPECIFIC_SHORT : UTZFMT_STYLE_SPECIFIC_LONG;
3393 TimeZone *tz = tzFormat()->parse(style, text, pos, tzTimeType);
3394 if (tz != NULL) {
3395 cal.adoptTimeZone(tz);
3396 return pos.getIndex();
3397 }
3398 }
3399 break;
3400 case UDAT_TIMEZONE_RFC_FIELD: // 'Z'
3401 {
3402 UTimeZoneFormatStyle style = (count < 4) ?
3403 UTZFMT_STYLE_ISO_BASIC_LOCAL_FULL : ((count == 5) ? UTZFMT_STYLE_ISO_EXTENDED_FULL: UTZFMT_STYLE_LOCALIZED_GMT);
3404 TimeZone *tz = tzFormat()->parse(style, text, pos, tzTimeType);
3405 if (tz != NULL) {
3406 cal.adoptTimeZone(tz);
3407 return pos.getIndex();
3408 }
3409 return -start;
3410 }
3411 case UDAT_TIMEZONE_GENERIC_FIELD: // 'v'
3412 {
3413 UTimeZoneFormatStyle style = (count < 4) ? UTZFMT_STYLE_GENERIC_SHORT : UTZFMT_STYLE_GENERIC_LONG;
3414 TimeZone *tz = tzFormat()->parse(style, text, pos, tzTimeType);
3415 if (tz != NULL) {
3416 cal.adoptTimeZone(tz);
3417 return pos.getIndex();
3418 }
3419 return -start;
3420 }
3421 case UDAT_TIMEZONE_SPECIAL_FIELD: // 'V'
3422 {
3423 UTimeZoneFormatStyle style;
3424 switch (count) {
3425 case 1:
3426 style = UTZFMT_STYLE_ZONE_ID_SHORT;
3427 break;
3428 case 2:
3429 style = UTZFMT_STYLE_ZONE_ID;
3430 break;
3431 case 3:
3432 style = UTZFMT_STYLE_EXEMPLAR_LOCATION;
3433 break;
3434 default:
3435 style = UTZFMT_STYLE_GENERIC_LOCATION;
3436 break;
3437 }
3438 TimeZone *tz = tzFormat()->parse(style, text, pos, tzTimeType);
3439 if (tz != NULL) {
3440 cal.adoptTimeZone(tz);
3441 return pos.getIndex();
3442 }
3443 return -start;
3444 }
3445 case UDAT_TIMEZONE_LOCALIZED_GMT_OFFSET_FIELD: // 'O'
3446 {
3447 UTimeZoneFormatStyle style = (count < 4) ? UTZFMT_STYLE_LOCALIZED_GMT_SHORT : UTZFMT_STYLE_LOCALIZED_GMT;
3448 TimeZone *tz = tzFormat()->parse(style, text, pos, tzTimeType);
3449 if (tz != NULL) {
3450 cal.adoptTimeZone(tz);
3451 return pos.getIndex();
3452 }
3453 return -start;
3454 }
3455 case UDAT_TIMEZONE_ISO_FIELD: // 'X'
3456 {
3457 UTimeZoneFormatStyle style;
3458 switch (count) {
3459 case 1:
3460 style = UTZFMT_STYLE_ISO_BASIC_SHORT;
3461 break;
3462 case 2:
3463 style = UTZFMT_STYLE_ISO_BASIC_FIXED;
3464 break;
3465 case 3:
3466 style = UTZFMT_STYLE_ISO_EXTENDED_FIXED;
3467 break;
3468 case 4:
3469 style = UTZFMT_STYLE_ISO_BASIC_FULL;
3470 break;
3471 default:
3472 style = UTZFMT_STYLE_ISO_EXTENDED_FULL;
3473 break;
3474 }
3475 TimeZone *tz = tzFormat()->parse(style, text, pos, tzTimeType);
3476 if (tz != NULL) {
3477 cal.adoptTimeZone(tz);
3478 return pos.getIndex();
3479 }
3480 return -start;
3481 }
3482 case UDAT_TIMEZONE_ISO_LOCAL_FIELD: // 'x'
3483 {
3484 UTimeZoneFormatStyle style;
3485 switch (count) {
3486 case 1:
3487 style = UTZFMT_STYLE_ISO_BASIC_LOCAL_SHORT;
3488 break;
3489 case 2:
3490 style = UTZFMT_STYLE_ISO_BASIC_LOCAL_FIXED;
3491 break;
3492 case 3:
3493 style = UTZFMT_STYLE_ISO_EXTENDED_LOCAL_FIXED;
3494 break;
3495 case 4:
3496 style = UTZFMT_STYLE_ISO_BASIC_LOCAL_FULL;
3497 break;
3498 default:
3499 style = UTZFMT_STYLE_ISO_EXTENDED_LOCAL_FULL;
3500 break;
3501 }
3502 TimeZone *tz = tzFormat()->parse(style, text, pos, tzTimeType);
3503 if (tz != NULL) {
3504 cal.adoptTimeZone(tz);
3505 return pos.getIndex();
3506 }
3507 return -start;
3508 }
3509 // currently no pattern character is defined for UDAT_TIME_SEPARATOR_FIELD
3510 // so we should not get here. Leave support in for future definition.
3511 case UDAT_TIME_SEPARATOR_FIELD:
3512 {
3513 static const UChar def_sep = DateFormatSymbols::DEFAULT_TIME_SEPARATOR;
3514 static const UChar alt_sep = DateFormatSymbols::ALTERNATE_TIME_SEPARATOR;
3515
3516 // Try matching a time separator.
3517 int32_t count = 1;
3518 UnicodeString data[3];
3519 fSymbols->getTimeSeparatorString(data[0]);
3520
3521 // Add the default, if different from the locale.
3522 if (data[0].compare(&def_sep, 1) != 0) {
3523 data[count++].setTo(def_sep);
3524 }
3525
3526 // If lenient, add also the alternate, if different from the locale.
3527 if (isLenient() && data[0].compare(&alt_sep, 1) != 0) {
3528 data[count++].setTo(alt_sep);
3529 }
3530
3531 return matchString(text, start, UCAL_FIELD_COUNT /* => nothing to set */, data, count, NULL, cal);
3532 }
3533
3534 case UDAT_AM_PM_MIDNIGHT_NOON_FIELD:
3535 {
3536 U_ASSERT(dayPeriod != NULL);
3537 int32_t ampmStart = subParse(text, start, 0x61, count,
3538 obeyCount, allowNegative, ambiguousYear, saveHebrewMonth, cal,
3539 patLoc, numericLeapMonthFormatter, tzTimeType, mutableNFs);
3540
3541 if (ampmStart > 0) {
3542 return ampmStart;
3543 } else {
3544 int32_t newStart = 0;
3545
3546 // Only match the first two strings from the day period strings array.
3547 if (getBooleanAttribute(UDAT_PARSE_MULTIPLE_PATTERNS_FOR_MATCH, status) || count == 3) {
3548 if ((newStart = matchDayPeriodStrings(text, start, fSymbols->fAbbreviatedDayPeriods,
3549 2, *dayPeriod)) > 0) {
3550 return newStart;
3551 }
3552 }
3553 if (getBooleanAttribute(UDAT_PARSE_MULTIPLE_PATTERNS_FOR_MATCH, status) || count == 5) {
3554 if ((newStart = matchDayPeriodStrings(text, start, fSymbols->fNarrowDayPeriods,
3555 2, *dayPeriod)) > 0) {
3556 return newStart;
3557 }
3558 }
3559 // count == 4, but allow other counts
3560 if (getBooleanAttribute(UDAT_PARSE_MULTIPLE_PATTERNS_FOR_MATCH, status)) {
3561 if ((newStart = matchDayPeriodStrings(text, start, fSymbols->fWideDayPeriods,
3562 2, *dayPeriod)) > 0) {
3563 return newStart;
3564 }
3565 }
3566
3567 return -start;
3568 }
3569 }
3570
3571 case UDAT_FLEXIBLE_DAY_PERIOD_FIELD:
3572 {
3573 U_ASSERT(dayPeriod != NULL);
3574 int32_t newStart = 0;
3575
3576 if (getBooleanAttribute(UDAT_PARSE_MULTIPLE_PATTERNS_FOR_MATCH, status) || count == 3) {
3577 if ((newStart = matchDayPeriodStrings(text, start, fSymbols->fAbbreviatedDayPeriods,
3578 fSymbols->fAbbreviatedDayPeriodsCount, *dayPeriod)) > 0) {
3579 return newStart;
3580 }
3581 }
3582 if (getBooleanAttribute(UDAT_PARSE_MULTIPLE_PATTERNS_FOR_MATCH, status) || count == 5) {
3583 if ((newStart = matchDayPeriodStrings(text, start, fSymbols->fNarrowDayPeriods,
3584 fSymbols->fNarrowDayPeriodsCount, *dayPeriod)) > 0) {
3585 return newStart;
3586 }
3587 }
3588 if (getBooleanAttribute(UDAT_PARSE_MULTIPLE_PATTERNS_FOR_MATCH, status) || count == 4) {
3589 if ((newStart = matchDayPeriodStrings(text, start, fSymbols->fWideDayPeriods,
3590 fSymbols->fWideDayPeriodsCount, *dayPeriod)) > 0) {
3591 return newStart;
3592 }
3593 }
3594
3595 return -start;
3596 }
3597
3598 default:
3599 // Handle "generic" fields
3600 // this is now handled below, outside the switch block
3601 break;
3602 }
3603 // Handle "generic" fields:
3604 // switch default case now handled here (outside switch block) to allow
3605 // parsing of some string fields as digits for lenient case
3606
3607 int32_t parseStart = pos.getIndex();
3608 const UnicodeString* src;
3609 if (obeyCount) {
3610 if ((start+count) > text.length()) {
3611 return -start;
3612 }
3613 text.extractBetween(0, start + count, temp);
3614 src = &temp;
3615 } else {
3616 src = &text;
3617 }
3618 parseInt(*src, number, pos, allowNegative,currentNumberFormat);
3619 if (pos.getIndex() != parseStart) {
3620 int32_t value = number.getLong();
3621
3622 // Don't need suffix processing here (as in number processing at the beginning of the function);
3623 // the new fields being handled as numeric values (month, weekdays, quarters) should not have suffixes.
3624
3625 if (!getBooleanAttribute(UDAT_PARSE_ALLOW_NUMERIC, status)) {
3626 // Check the range of the value
3627 int32_t bias = gFieldRangeBias[patternCharIndex];
3628 if (bias >= 0 && (value > cal.getMaximum(field) + bias || value < cal.getMinimum(field) + bias)) {
3629 return -start;
3630 }
3631 }
3632
3633 // For the following, need to repeat some of the "if (gotNumber)" code above:
3634 // UDAT_[STANDALONE_]MONTH_FIELD, UDAT_DOW_LOCAL_FIELD, UDAT_STANDALONE_DAY_FIELD,
3635 // UDAT_[STANDALONE_]QUARTER_FIELD
3636 switch (patternCharIndex) {
3637 case UDAT_MONTH_FIELD:
3638 // See notes under UDAT_MONTH_FIELD case above
3639 if (!strcmp(cal.getType(),"hebrew")) {
3640 HebrewCalendar *hc = (HebrewCalendar*)&cal;
3641 if (cal.isSet(UCAL_YEAR)) {
3642 UErrorCode status = U_ZERO_ERROR;
3643 if (!hc->isLeapYear(hc->get(UCAL_YEAR,status)) && value >= 6) {
3644 cal.set(UCAL_MONTH, value);
3645 } else {
3646 cal.set(UCAL_MONTH, value - 1);
3647 }
3648 } else {
3649 saveHebrewMonth = value;
3650 }
3651 } else {
3652 cal.set(UCAL_MONTH, value - 1);
3653 }
3654 break;
3655 case UDAT_STANDALONE_MONTH_FIELD:
3656 cal.set(UCAL_MONTH, value - 1);
3657 break;
3658 case UDAT_DOW_LOCAL_FIELD:
3659 case UDAT_STANDALONE_DAY_FIELD:
3660 cal.set(UCAL_DOW_LOCAL, value);
3661 break;
3662 case UDAT_QUARTER_FIELD:
3663 case UDAT_STANDALONE_QUARTER_FIELD:
3664 cal.set(UCAL_MONTH, (value - 1) * 3);
3665 break;
3666 case UDAT_RELATED_YEAR_FIELD:
3667 cal.setRelatedYear(value);
3668 break;
3669 default:
3670 cal.set(field, value);
3671 break;
3672 }
3673 return pos.getIndex();
3674 }
3675 return -start;
3676 }
3677
3678 /**
3679 * Parse an integer using fNumberFormat. This method is semantically
3680 * const, but actually may modify fNumberFormat.
3681 */
parseInt(const UnicodeString & text,Formattable & number,ParsePosition & pos,UBool allowNegative,NumberFormat * fmt) const3682 void SimpleDateFormat::parseInt(const UnicodeString& text,
3683 Formattable& number,
3684 ParsePosition& pos,
3685 UBool allowNegative,
3686 NumberFormat *fmt) const {
3687 parseInt(text, number, -1, pos, allowNegative,fmt);
3688 }
3689
3690 /**
3691 * Parse an integer using fNumberFormat up to maxDigits.
3692 */
parseInt(const UnicodeString & text,Formattable & number,int32_t maxDigits,ParsePosition & pos,UBool allowNegative,NumberFormat * fmt) const3693 void SimpleDateFormat::parseInt(const UnicodeString& text,
3694 Formattable& number,
3695 int32_t maxDigits,
3696 ParsePosition& pos,
3697 UBool allowNegative,
3698 NumberFormat *fmt) const {
3699 UnicodeString oldPrefix;
3700 DecimalFormat* df = NULL;
3701 if (!allowNegative && (df = dynamic_cast<DecimalFormat*>(fmt)) != NULL) {
3702 df->getNegativePrefix(oldPrefix);
3703 df->setNegativePrefix(UnicodeString(TRUE, SUPPRESS_NEGATIVE_PREFIX, -1));
3704 }
3705 int32_t oldPos = pos.getIndex();
3706 fmt->parse(text, number, pos);
3707 if (df != NULL) {
3708 df->setNegativePrefix(oldPrefix);
3709 }
3710
3711 if (maxDigits > 0) {
3712 // adjust the result to fit into
3713 // the maxDigits and move the position back
3714 int32_t nDigits = pos.getIndex() - oldPos;
3715 if (nDigits > maxDigits) {
3716 int32_t val = number.getLong();
3717 nDigits -= maxDigits;
3718 while (nDigits > 0) {
3719 val /= 10;
3720 nDigits--;
3721 }
3722 pos.setIndex(oldPos + maxDigits);
3723 number.setLong(val);
3724 }
3725 }
3726 }
3727
3728 //----------------------------------------------------------------------
3729
translatePattern(const UnicodeString & originalPattern,UnicodeString & translatedPattern,const UnicodeString & from,const UnicodeString & to,UErrorCode & status)3730 void SimpleDateFormat::translatePattern(const UnicodeString& originalPattern,
3731 UnicodeString& translatedPattern,
3732 const UnicodeString& from,
3733 const UnicodeString& to,
3734 UErrorCode& status)
3735 {
3736 // run through the pattern and convert any pattern symbols from the version
3737 // in "from" to the corresponding character in "to". This code takes
3738 // quoted strings into account (it doesn't try to translate them), and it signals
3739 // an error if a particular "pattern character" doesn't appear in "from".
3740 // Depending on the values of "from" and "to" this can convert from generic
3741 // to localized patterns or localized to generic.
3742 if (U_FAILURE(status)) {
3743 return;
3744 }
3745
3746 translatedPattern.remove();
3747 UBool inQuote = FALSE;
3748 for (int32_t i = 0; i < originalPattern.length(); ++i) {
3749 UChar c = originalPattern[i];
3750 if (inQuote) {
3751 if (c == QUOTE) {
3752 inQuote = FALSE;
3753 }
3754 } else {
3755 if (c == QUOTE) {
3756 inQuote = TRUE;
3757 } else if (isSyntaxChar(c)) {
3758 int32_t ci = from.indexOf(c);
3759 if (ci == -1) {
3760 status = U_INVALID_FORMAT_ERROR;
3761 return;
3762 }
3763 c = to[ci];
3764 }
3765 }
3766 translatedPattern += c;
3767 }
3768 if (inQuote) {
3769 status = U_INVALID_FORMAT_ERROR;
3770 return;
3771 }
3772 }
3773
3774 //----------------------------------------------------------------------
3775
3776 UnicodeString&
toPattern(UnicodeString & result) const3777 SimpleDateFormat::toPattern(UnicodeString& result) const
3778 {
3779 result = fPattern;
3780 return result;
3781 }
3782
3783 //----------------------------------------------------------------------
3784
3785 UnicodeString&
toLocalizedPattern(UnicodeString & result,UErrorCode & status) const3786 SimpleDateFormat::toLocalizedPattern(UnicodeString& result,
3787 UErrorCode& status) const
3788 {
3789 translatePattern(fPattern, result,
3790 UnicodeString(DateFormatSymbols::getPatternUChars()),
3791 fSymbols->fLocalPatternChars, status);
3792 return result;
3793 }
3794
3795 //----------------------------------------------------------------------
3796
3797 void
applyPattern(const UnicodeString & pattern)3798 SimpleDateFormat::applyPattern(const UnicodeString& pattern)
3799 {
3800 fPattern = pattern;
3801 parsePattern();
3802 }
3803
3804 //----------------------------------------------------------------------
3805
3806 void
applyLocalizedPattern(const UnicodeString & pattern,UErrorCode & status)3807 SimpleDateFormat::applyLocalizedPattern(const UnicodeString& pattern,
3808 UErrorCode &status)
3809 {
3810 translatePattern(pattern, fPattern,
3811 fSymbols->fLocalPatternChars,
3812 UnicodeString(DateFormatSymbols::getPatternUChars()), status);
3813 }
3814
3815 //----------------------------------------------------------------------
3816
3817 const DateFormatSymbols*
getDateFormatSymbols() const3818 SimpleDateFormat::getDateFormatSymbols() const
3819 {
3820 return fSymbols;
3821 }
3822
3823 //----------------------------------------------------------------------
3824
3825 void
adoptDateFormatSymbols(DateFormatSymbols * newFormatSymbols)3826 SimpleDateFormat::adoptDateFormatSymbols(DateFormatSymbols* newFormatSymbols)
3827 {
3828 delete fSymbols;
3829 fSymbols = newFormatSymbols;
3830 }
3831
3832 //----------------------------------------------------------------------
3833 void
setDateFormatSymbols(const DateFormatSymbols & newFormatSymbols)3834 SimpleDateFormat::setDateFormatSymbols(const DateFormatSymbols& newFormatSymbols)
3835 {
3836 delete fSymbols;
3837 fSymbols = new DateFormatSymbols(newFormatSymbols);
3838 }
3839
3840 //----------------------------------------------------------------------
3841 const TimeZoneFormat*
getTimeZoneFormat(void) const3842 SimpleDateFormat::getTimeZoneFormat(void) const {
3843 return (const TimeZoneFormat*)tzFormat();
3844 }
3845
3846 //----------------------------------------------------------------------
3847 void
adoptTimeZoneFormat(TimeZoneFormat * timeZoneFormatToAdopt)3848 SimpleDateFormat::adoptTimeZoneFormat(TimeZoneFormat* timeZoneFormatToAdopt)
3849 {
3850 delete fTimeZoneFormat;
3851 fTimeZoneFormat = timeZoneFormatToAdopt;
3852 }
3853
3854 //----------------------------------------------------------------------
3855 void
setTimeZoneFormat(const TimeZoneFormat & newTimeZoneFormat)3856 SimpleDateFormat::setTimeZoneFormat(const TimeZoneFormat& newTimeZoneFormat)
3857 {
3858 delete fTimeZoneFormat;
3859 fTimeZoneFormat = new TimeZoneFormat(newTimeZoneFormat);
3860 }
3861
3862 //----------------------------------------------------------------------
3863
3864
adoptCalendar(Calendar * calendarToAdopt)3865 void SimpleDateFormat::adoptCalendar(Calendar* calendarToAdopt)
3866 {
3867 UErrorCode status = U_ZERO_ERROR;
3868 Locale calLocale(fLocale);
3869 calLocale.setKeywordValue("calendar", calendarToAdopt->getType(), status);
3870 DateFormatSymbols *newSymbols =
3871 DateFormatSymbols::createForLocale(calLocale, status);
3872 if (U_FAILURE(status)) {
3873 return;
3874 }
3875 DateFormat::adoptCalendar(calendarToAdopt);
3876 delete fSymbols;
3877 fSymbols = newSymbols;
3878 initializeDefaultCentury(); // we need a new century (possibly)
3879 }
3880
3881
3882 //----------------------------------------------------------------------
3883
3884
3885 // override the DateFormat implementation in order to
3886 // lazily initialize fCapitalizationBrkIter
3887 void
setContext(UDisplayContext value,UErrorCode & status)3888 SimpleDateFormat::setContext(UDisplayContext value, UErrorCode& status)
3889 {
3890 DateFormat::setContext(value, status);
3891 #if !UCONFIG_NO_BREAK_ITERATION
3892 if (U_SUCCESS(status)) {
3893 if ( fCapitalizationBrkIter == NULL && (value==UDISPCTX_CAPITALIZATION_FOR_BEGINNING_OF_SENTENCE ||
3894 value==UDISPCTX_CAPITALIZATION_FOR_UI_LIST_OR_MENU || value==UDISPCTX_CAPITALIZATION_FOR_STANDALONE) ) {
3895 UErrorCode status = U_ZERO_ERROR;
3896 fCapitalizationBrkIter = BreakIterator::createSentenceInstance(fLocale, status);
3897 if (U_FAILURE(status)) {
3898 delete fCapitalizationBrkIter;
3899 fCapitalizationBrkIter = NULL;
3900 }
3901 }
3902 }
3903 #endif
3904 }
3905
3906
3907 //----------------------------------------------------------------------
3908
3909
3910 UBool
isFieldUnitIgnored(UCalendarDateFields field) const3911 SimpleDateFormat::isFieldUnitIgnored(UCalendarDateFields field) const {
3912 return isFieldUnitIgnored(fPattern, field);
3913 }
3914
3915
3916 UBool
isFieldUnitIgnored(const UnicodeString & pattern,UCalendarDateFields field)3917 SimpleDateFormat::isFieldUnitIgnored(const UnicodeString& pattern,
3918 UCalendarDateFields field) {
3919 int32_t fieldLevel = fgCalendarFieldToLevel[field];
3920 int32_t level;
3921 UChar ch;
3922 UBool inQuote = FALSE;
3923 UChar prevCh = 0;
3924 int32_t count = 0;
3925
3926 for (int32_t i = 0; i < pattern.length(); ++i) {
3927 ch = pattern[i];
3928 if (ch != prevCh && count > 0) {
3929 level = getLevelFromChar(prevCh);
3930 // the larger the level, the smaller the field unit.
3931 if (fieldLevel <= level) {
3932 return FALSE;
3933 }
3934 count = 0;
3935 }
3936 if (ch == QUOTE) {
3937 if ((i+1) < pattern.length() && pattern[i+1] == QUOTE) {
3938 ++i;
3939 } else {
3940 inQuote = ! inQuote;
3941 }
3942 }
3943 else if (!inQuote && isSyntaxChar(ch)) {
3944 prevCh = ch;
3945 ++count;
3946 }
3947 }
3948 if (count > 0) {
3949 // last item
3950 level = getLevelFromChar(prevCh);
3951 if (fieldLevel <= level) {
3952 return FALSE;
3953 }
3954 }
3955 return TRUE;
3956 }
3957
3958 //----------------------------------------------------------------------
3959
3960 const Locale&
getSmpFmtLocale(void) const3961 SimpleDateFormat::getSmpFmtLocale(void) const {
3962 return fLocale;
3963 }
3964
3965 //----------------------------------------------------------------------
3966
3967 int32_t
checkIntSuffix(const UnicodeString & text,int32_t start,int32_t patLoc,UBool isNegative) const3968 SimpleDateFormat::checkIntSuffix(const UnicodeString& text, int32_t start,
3969 int32_t patLoc, UBool isNegative) const {
3970 // local variables
3971 UnicodeString suf;
3972 int32_t patternMatch;
3973 int32_t textPreMatch;
3974 int32_t textPostMatch;
3975
3976 // check that we are still in range
3977 if ( (start > text.length()) ||
3978 (start < 0) ||
3979 (patLoc < 0) ||
3980 (patLoc > fPattern.length())) {
3981 // out of range, don't advance location in text
3982 return start;
3983 }
3984
3985 // get the suffix
3986 DecimalFormat* decfmt = dynamic_cast<DecimalFormat*>(fNumberFormat);
3987 if (decfmt != NULL) {
3988 if (isNegative) {
3989 suf = decfmt->getNegativeSuffix(suf);
3990 }
3991 else {
3992 suf = decfmt->getPositiveSuffix(suf);
3993 }
3994 }
3995
3996 // check for suffix
3997 if (suf.length() <= 0) {
3998 return start;
3999 }
4000
4001 // check suffix will be encountered in the pattern
4002 patternMatch = compareSimpleAffix(suf,fPattern,patLoc);
4003
4004 // check if a suffix will be encountered in the text
4005 textPreMatch = compareSimpleAffix(suf,text,start);
4006
4007 // check if a suffix was encountered in the text
4008 textPostMatch = compareSimpleAffix(suf,text,start-suf.length());
4009
4010 // check for suffix match
4011 if ((textPreMatch >= 0) && (patternMatch >= 0) && (textPreMatch == patternMatch)) {
4012 return start;
4013 }
4014 else if ((textPostMatch >= 0) && (patternMatch >= 0) && (textPostMatch == patternMatch)) {
4015 return start - suf.length();
4016 }
4017
4018 // should not get here
4019 return start;
4020 }
4021
4022 //----------------------------------------------------------------------
4023
4024 int32_t
compareSimpleAffix(const UnicodeString & affix,const UnicodeString & input,int32_t pos) const4025 SimpleDateFormat::compareSimpleAffix(const UnicodeString& affix,
4026 const UnicodeString& input,
4027 int32_t pos) const {
4028 int32_t start = pos;
4029 for (int32_t i=0; i<affix.length(); ) {
4030 UChar32 c = affix.char32At(i);
4031 int32_t len = U16_LENGTH(c);
4032 if (PatternProps::isWhiteSpace(c)) {
4033 // We may have a pattern like: \u200F \u0020
4034 // and input text like: \u200F \u0020
4035 // Note that U+200F and U+0020 are Pattern_White_Space but only
4036 // U+0020 is UWhiteSpace. So we have to first do a direct
4037 // match of the run of Pattern_White_Space in the pattern,
4038 // then match any extra characters.
4039 UBool literalMatch = FALSE;
4040 while (pos < input.length() &&
4041 input.char32At(pos) == c) {
4042 literalMatch = TRUE;
4043 i += len;
4044 pos += len;
4045 if (i == affix.length()) {
4046 break;
4047 }
4048 c = affix.char32At(i);
4049 len = U16_LENGTH(c);
4050 if (!PatternProps::isWhiteSpace(c)) {
4051 break;
4052 }
4053 }
4054
4055 // Advance over run in pattern
4056 i = skipPatternWhiteSpace(affix, i);
4057
4058 // Advance over run in input text
4059 // Must see at least one white space char in input,
4060 // unless we've already matched some characters literally.
4061 int32_t s = pos;
4062 pos = skipUWhiteSpace(input, pos);
4063 if (pos == s && !literalMatch) {
4064 return -1;
4065 }
4066
4067 // If we skip UWhiteSpace in the input text, we need to skip it in the pattern.
4068 // Otherwise, the previous lines may have skipped over text (such as U+00A0) that
4069 // is also in the affix.
4070 i = skipUWhiteSpace(affix, i);
4071 } else {
4072 if (pos < input.length() &&
4073 input.char32At(pos) == c) {
4074 i += len;
4075 pos += len;
4076 } else {
4077 return -1;
4078 }
4079 }
4080 }
4081 return pos - start;
4082 }
4083
4084 //----------------------------------------------------------------------
4085
4086 int32_t
skipPatternWhiteSpace(const UnicodeString & text,int32_t pos) const4087 SimpleDateFormat::skipPatternWhiteSpace(const UnicodeString& text, int32_t pos) const {
4088 const UChar* s = text.getBuffer();
4089 return (int32_t)(PatternProps::skipWhiteSpace(s + pos, text.length() - pos) - s);
4090 }
4091
4092 //----------------------------------------------------------------------
4093
4094 int32_t
skipUWhiteSpace(const UnicodeString & text,int32_t pos) const4095 SimpleDateFormat::skipUWhiteSpace(const UnicodeString& text, int32_t pos) const {
4096 while (pos < text.length()) {
4097 UChar32 c = text.char32At(pos);
4098 if (!u_isUWhiteSpace(c)) {
4099 break;
4100 }
4101 pos += U16_LENGTH(c);
4102 }
4103 return pos;
4104 }
4105
4106 //----------------------------------------------------------------------
4107
4108 // Lazy TimeZoneFormat instantiation, semantically const.
4109 TimeZoneFormat *
tzFormat() const4110 SimpleDateFormat::tzFormat() const {
4111 if (fTimeZoneFormat == NULL) {
4112 umtx_lock(&LOCK);
4113 {
4114 if (fTimeZoneFormat == NULL) {
4115 UErrorCode status = U_ZERO_ERROR;
4116 TimeZoneFormat *tzfmt = TimeZoneFormat::createInstance(fLocale, status);
4117 if (U_FAILURE(status)) {
4118 return NULL;
4119 }
4120
4121 const_cast<SimpleDateFormat *>(this)->fTimeZoneFormat = tzfmt;
4122 }
4123 }
4124 umtx_unlock(&LOCK);
4125 }
4126 return fTimeZoneFormat;
4127 }
4128
parsePattern()4129 void SimpleDateFormat::parsePattern() {
4130 fHasMinute = FALSE;
4131 fHasSecond = FALSE;
4132
4133 int len = fPattern.length();
4134 UBool inQuote = FALSE;
4135 for (int32_t i = 0; i < len; ++i) {
4136 UChar ch = fPattern[i];
4137 if (ch == QUOTE) {
4138 inQuote = !inQuote;
4139 }
4140 if (!inQuote) {
4141 if (ch == 0x6D) { // 0x6D == 'm'
4142 fHasMinute = TRUE;
4143 }
4144 if (ch == 0x73) { // 0x73 == 's'
4145 fHasSecond = TRUE;
4146 }
4147 }
4148 }
4149 }
4150
4151 U_NAMESPACE_END
4152
4153 #endif /* #if !UCONFIG_NO_FORMATTING */
4154
4155 //eof
4156