1/*
2*******************************************************************************
3* Copyright (C) 1997-2015, International Business Machines Corporation and    *
4* others. All Rights Reserved.                                                *
5*******************************************************************************
6*
7* File SMPDTFMT.CPP
8*
9* Modification History:
10*
11*   Date        Name        Description
12*   02/19/97    aliu        Converted from java.
13*   03/31/97    aliu        Modified extensively to work with 50 locales.
14*   04/01/97    aliu        Added support for centuries.
15*   07/09/97    helena      Made ParsePosition into a class.
16*   07/21/98    stephen     Added initializeDefaultCentury.
17*                             Removed getZoneIndex (added in DateFormatSymbols)
18*                             Removed subParseLong
19*                             Removed chk
20*  02/22/99     stephen     Removed character literals for EBCDIC safety
21*   10/14/99    aliu        Updated 2-digit year parsing so that only "00" thru
22*                           "99" are recognized. {j28 4182066}
23*   11/15/99    weiv        Added support for week of year/day of week format
24********************************************************************************
25*/
26
27#define ZID_KEY_MAX 128
28
29#include "unicode/utypes.h"
30
31#if !UCONFIG_NO_FORMATTING
32#include "unicode/smpdtfmt.h"
33#include "unicode/dtfmtsym.h"
34#include "unicode/ures.h"
35#include "unicode/msgfmt.h"
36#include "unicode/calendar.h"
37#include "unicode/gregocal.h"
38#include "unicode/timezone.h"
39#include "unicode/decimfmt.h"
40#include "unicode/dcfmtsym.h"
41#include "unicode/uchar.h"
42#include "unicode/uniset.h"
43#include "unicode/ustring.h"
44#include "unicode/basictz.h"
45#include "unicode/simpletz.h"
46#include "unicode/rbtz.h"
47#include "unicode/tzfmt.h"
48#include "unicode/utf16.h"
49#include "unicode/vtzone.h"
50#include "unicode/udisplaycontext.h"
51#include "unicode/brkiter.h"
52#include "olsontz.h"
53#include "patternprops.h"
54#include "fphdlimp.h"
55#include "gregoimp.h"
56#include "hebrwcal.h"
57#include "cstring.h"
58#include "uassert.h"
59#include "cmemory.h"
60#include "umutex.h"
61#include <float.h>
62#include "smpdtfst.h"
63#include "sharednumberformat.h"
64#include "ustr_imp.h"
65
66#if defined( U_DEBUG_CALSVC ) || defined (U_DEBUG_CAL)
67#include <stdio.h>
68#endif
69
70// *****************************************************************************
71// class SimpleDateFormat
72// *****************************************************************************
73
74U_NAMESPACE_BEGIN
75
76/**
77 * Last-resort string to use for "GMT" when constructing time zone strings.
78 */
79// For time zones that have no names, use strings GMT+minutes and
80// GMT-minutes. For instance, in France the time zone is GMT+60.
81// Also accepted are GMT+H:MM or GMT-H:MM.
82// Currently not being used
83//static const UChar gGmt[]      = {0x0047, 0x004D, 0x0054, 0x0000};         // "GMT"
84//static const UChar gGmtPlus[]  = {0x0047, 0x004D, 0x0054, 0x002B, 0x0000}; // "GMT+"
85//static const UChar gGmtMinus[] = {0x0047, 0x004D, 0x0054, 0x002D, 0x0000}; // "GMT-"
86//static const UChar gDefGmtPat[]       = {0x0047, 0x004D, 0x0054, 0x007B, 0x0030, 0x007D, 0x0000}; /* GMT{0} */
87//static const UChar gDefGmtNegHmsPat[] = {0x002D, 0x0048, 0x0048, 0x003A, 0x006D, 0x006D, 0x003A, 0x0073, 0x0073, 0x0000}; /* -HH:mm:ss */
88//static const UChar gDefGmtNegHmPat[]  = {0x002D, 0x0048, 0x0048, 0x003A, 0x006D, 0x006D, 0x0000}; /* -HH:mm */
89//static const UChar gDefGmtPosHmsPat[] = {0x002B, 0x0048, 0x0048, 0x003A, 0x006D, 0x006D, 0x003A, 0x0073, 0x0073, 0x0000}; /* +HH:mm:ss */
90//static const UChar gDefGmtPosHmPat[]  = {0x002B, 0x0048, 0x0048, 0x003A, 0x006D, 0x006D, 0x0000}; /* +HH:mm */
91//static const UChar gUt[]       = {0x0055, 0x0054, 0x0000};  // "UT"
92//static const UChar gUtc[]      = {0x0055, 0x0054, 0x0043, 0x0000};  // "UT"
93
94typedef enum GmtPatSize {
95    kGmtLen = 3,
96    kGmtPatLen = 6,
97    kNegHmsLen = 9,
98    kNegHmLen = 6,
99    kPosHmsLen = 9,
100    kPosHmLen = 6,
101    kUtLen = 2,
102    kUtcLen = 3
103} GmtPatSize;
104
105// Stuff needed for numbering system overrides
106
107typedef enum OvrStrType {
108    kOvrStrDate = 0,
109    kOvrStrTime = 1,
110    kOvrStrBoth = 2
111} OvrStrType;
112
113static const UDateFormatField kDateFields[] = {
114    UDAT_YEAR_FIELD,
115    UDAT_MONTH_FIELD,
116    UDAT_DATE_FIELD,
117    UDAT_DAY_OF_YEAR_FIELD,
118    UDAT_DAY_OF_WEEK_IN_MONTH_FIELD,
119    UDAT_WEEK_OF_YEAR_FIELD,
120    UDAT_WEEK_OF_MONTH_FIELD,
121    UDAT_YEAR_WOY_FIELD,
122    UDAT_EXTENDED_YEAR_FIELD,
123    UDAT_JULIAN_DAY_FIELD,
124    UDAT_STANDALONE_DAY_FIELD,
125    UDAT_STANDALONE_MONTH_FIELD,
126    UDAT_QUARTER_FIELD,
127    UDAT_STANDALONE_QUARTER_FIELD,
128    UDAT_YEAR_NAME_FIELD,
129    UDAT_RELATED_YEAR_FIELD };
130static const int8_t kDateFieldsCount = 16;
131
132static const UDateFormatField kTimeFields[] = {
133    UDAT_HOUR_OF_DAY1_FIELD,
134    UDAT_HOUR_OF_DAY0_FIELD,
135    UDAT_MINUTE_FIELD,
136    UDAT_SECOND_FIELD,
137    UDAT_FRACTIONAL_SECOND_FIELD,
138    UDAT_HOUR1_FIELD,
139    UDAT_HOUR0_FIELD,
140    UDAT_MILLISECONDS_IN_DAY_FIELD,
141    UDAT_TIMEZONE_RFC_FIELD,
142    UDAT_TIMEZONE_LOCALIZED_GMT_OFFSET_FIELD };
143static const int8_t kTimeFieldsCount = 10;
144
145
146// This is a pattern-of-last-resort used when we can't load a usable pattern out
147// of a resource.
148static const UChar gDefaultPattern[] =
149{
150    0x79, 0x79, 0x79, 0x79, 0x4D, 0x4D, 0x64, 0x64, 0x20, 0x68, 0x68, 0x3A, 0x6D, 0x6D, 0x20, 0x61, 0
151};  /* "yyyyMMdd hh:mm a" */
152
153// This prefix is designed to NEVER MATCH real text, in order to
154// suppress the parsing of negative numbers.  Adjust as needed (if
155// this becomes valid Unicode).
156static const UChar SUPPRESS_NEGATIVE_PREFIX[] = {0xAB00, 0};
157
158/**
159 * These are the tags we expect to see in normal resource bundle files associated
160 * with a locale.
161 */
162static const char gDateTimePatternsTag[]="DateTimePatterns";
163
164//static const UChar gEtcUTC[] = {0x45, 0x74, 0x63, 0x2F, 0x55, 0x54, 0x43, 0x00}; // "Etc/UTC"
165static const UChar QUOTE = 0x27; // Single quote
166
167/*
168 * The field range check bias for each UDateFormatField.
169 * The bias is added to the minimum and maximum values
170 * before they are compared to the parsed number.
171 * For example, the calendar stores zero-based month numbers
172 * but the parsed month numbers start at 1, so the bias is 1.
173 *
174 * A value of -1 means that the value is not checked.
175 */
176static const int32_t gFieldRangeBias[] = {
177    -1,  // 'G' - UDAT_ERA_FIELD
178    -1,  // 'y' - UDAT_YEAR_FIELD
179     1,  // 'M' - UDAT_MONTH_FIELD
180     0,  // 'd' - UDAT_DATE_FIELD
181    -1,  // 'k' - UDAT_HOUR_OF_DAY1_FIELD
182    -1,  // 'H' - UDAT_HOUR_OF_DAY0_FIELD
183     0,  // 'm' - UDAT_MINUTE_FIELD
184     0,  // 's' - UDAT_SECOND_FIELD
185    -1,  // 'S' - UDAT_FRACTIONAL_SECOND_FIELD (0-999?)
186    -1,  // 'E' - UDAT_DAY_OF_WEEK_FIELD (1-7?)
187    -1,  // 'D' - UDAT_DAY_OF_YEAR_FIELD (1 - 366?)
188    -1,  // 'F' - UDAT_DAY_OF_WEEK_IN_MONTH_FIELD (1-5?)
189    -1,  // 'w' - UDAT_WEEK_OF_YEAR_FIELD (1-52?)
190    -1,  // 'W' - UDAT_WEEK_OF_MONTH_FIELD (1-5?)
191    -1,  // 'a' - UDAT_AM_PM_FIELD
192    -1,  // 'h' - UDAT_HOUR1_FIELD
193    -1,  // 'K' - UDAT_HOUR0_FIELD
194    -1,  // 'z' - UDAT_TIMEZONE_FIELD
195    -1,  // 'Y' - UDAT_YEAR_WOY_FIELD
196    -1,  // 'e' - UDAT_DOW_LOCAL_FIELD
197    -1,  // 'u' - UDAT_EXTENDED_YEAR_FIELD
198    -1,  // 'g' - UDAT_JULIAN_DAY_FIELD
199    -1,  // 'A' - UDAT_MILLISECONDS_IN_DAY_FIELD
200    -1,  // 'Z' - UDAT_TIMEZONE_RFC_FIELD
201    -1,  // 'v' - UDAT_TIMEZONE_GENERIC_FIELD
202     0,  // 'c' - UDAT_STANDALONE_DAY_FIELD
203     1,  // 'L' - UDAT_STANDALONE_MONTH_FIELD
204    -1,  // 'Q' - UDAT_QUARTER_FIELD (1-4?)
205    -1,  // 'q' - UDAT_STANDALONE_QUARTER_FIELD
206    -1,  // 'V' - UDAT_TIMEZONE_SPECIAL_FIELD
207    -1,  // 'U' - UDAT_YEAR_NAME_FIELD
208    -1,  // 'O' - UDAT_TIMEZONE_LOCALIZED_GMT_OFFSET_FIELD
209    -1,  // 'X' - UDAT_TIMEZONE_ISO_FIELD
210    -1,  // 'x' - UDAT_TIMEZONE_ISO_LOCAL_FIELD
211    -1,  // 'r' - UDAT_RELATED_YEAR_FIELD
212#if UDAT_HAS_PATTERN_CHAR_FOR_TIME_SEPARATOR
213    -1,  // ':' - UDAT_TIME_SEPARATOR_FIELD
214#else
215    -1,  // (no pattern character currently) - UDAT_TIME_SEPARATOR_FIELD
216#endif
217};
218
219// When calendar uses hebr numbering (i.e. he@calendar=hebrew),
220// offset the years within the current millenium down to 1-999
221static const int32_t HEBREW_CAL_CUR_MILLENIUM_START_YEAR = 5000;
222static const int32_t HEBREW_CAL_CUR_MILLENIUM_END_YEAR = 6000;
223
224static UMutex LOCK = U_MUTEX_INITIALIZER;
225
226UOBJECT_DEFINE_RTTI_IMPLEMENTATION(SimpleDateFormat)
227
228SimpleDateFormat::NSOverride::~NSOverride() {
229    if (snf != NULL) {
230        snf->removeRef();
231    }
232}
233
234
235void SimpleDateFormat::NSOverride::free() {
236    NSOverride *cur = this;
237    while (cur) {
238        NSOverride *next = cur->next;
239        delete cur;
240        cur = next;
241    }
242}
243
244// no matter what the locale's default number format looked like, we want
245// to modify it so that it doesn't use thousands separators, doesn't always
246// show the decimal point, and recognizes integers only when parsing
247static void fixNumberFormatForDates(NumberFormat &nf) {
248    nf.setGroupingUsed(FALSE);
249    DecimalFormat* decfmt = dynamic_cast<DecimalFormat*>(&nf);
250    if (decfmt != NULL) {
251        decfmt->setDecimalSeparatorAlwaysShown(FALSE);
252    }
253    nf.setParseIntegerOnly(TRUE);
254    nf.setMinimumFractionDigits(0); // To prevent "Jan 1.00, 1997.00"
255}
256
257static const SharedNumberFormat *createSharedNumberFormat(
258        NumberFormat *nfToAdopt) {
259    fixNumberFormatForDates(*nfToAdopt);
260    const SharedNumberFormat *result = new SharedNumberFormat(nfToAdopt);
261    if (result == NULL) {
262        delete nfToAdopt;
263    }
264    return result;
265}
266
267static const SharedNumberFormat *createSharedNumberFormat(
268        const Locale &loc, UErrorCode &status) {
269    NumberFormat *nf = NumberFormat::createInstance(loc, status);
270    if (U_FAILURE(status)) {
271        return NULL;
272    }
273    const SharedNumberFormat *result = createSharedNumberFormat(nf);
274    if (result == NULL) {
275        status = U_MEMORY_ALLOCATION_ERROR;
276    }
277    return result;
278}
279
280static const SharedNumberFormat **allocSharedNumberFormatters() {
281    const SharedNumberFormat **result = (const SharedNumberFormat**)
282            uprv_malloc(UDAT_FIELD_COUNT * sizeof(const SharedNumberFormat*));
283    if (result == NULL) {
284        return NULL;
285    }
286    for (int32_t i = 0; i < UDAT_FIELD_COUNT; ++i) {
287        result[i] = NULL;
288    }
289    return result;
290}
291
292static void freeSharedNumberFormatters(const SharedNumberFormat ** list) {
293    for (int32_t i = 0; i < UDAT_FIELD_COUNT; ++i) {
294        SharedObject::clearPtr(list[i]);
295    }
296    uprv_free(list);
297}
298
299const NumberFormat *SimpleDateFormat::getNumberFormatByIndex(
300        UDateFormatField index) const {
301    if (fSharedNumberFormatters == NULL ||
302        fSharedNumberFormatters[index] == NULL) {
303        return fNumberFormat;
304    }
305    return &(**fSharedNumberFormatters[index]);
306}
307
308class SimpleDateFormatMutableNFNode {
309 public:
310    const NumberFormat *key;
311    NumberFormat *value;
312    SimpleDateFormatMutableNFNode()
313            : key(NULL), value(NULL) { }
314    ~SimpleDateFormatMutableNFNode() {
315        delete value;
316    }
317 private:
318    SimpleDateFormatMutableNFNode(const SimpleDateFormatMutableNFNode &);
319    SimpleDateFormatMutableNFNode &operator=(const SimpleDateFormatMutableNFNode &);
320};
321
322// Single threaded cache of non const NumberFormats. Designed to be stack
323// allocated and used for a single format call.
324class SimpleDateFormatMutableNFs : public UMemory {
325 public:
326    SimpleDateFormatMutableNFs() {
327    }
328
329    // Returns a non-const clone of nf which can be safely modified.
330    // Subsequent calls with same nf will return the same non-const clone.
331    // This object maintains ownership of all returned non-const
332    // NumberFormat objects. On memory allocation error returns NULL.
333    // Caller must check for NULL return value.
334    NumberFormat *get(const NumberFormat *nf) {
335        if (nf == NULL) {
336            return NULL;
337        }
338        int32_t idx = 0;
339        while (nodes[idx].value) {
340            if (nf == nodes[idx].key) {
341                return nodes[idx].value;
342            }
343            ++idx;
344        }
345        U_ASSERT(idx < UDAT_FIELD_COUNT);
346        nodes[idx].key = nf;
347        nodes[idx].value = (NumberFormat *) nf->clone();
348        return nodes[idx].value;
349    }
350 private:
351    // +1 extra for sentinel. If each field had its own NumberFormat, this
352    // cache would have to allocate UDAT_FIELD_COUNT mutable versions worst
353    // case.
354    SimpleDateFormatMutableNFNode nodes[UDAT_FIELD_COUNT + 1];
355    SimpleDateFormatMutableNFs(const SimpleDateFormatMutableNFs &);
356    SimpleDateFormatMutableNFs &operator=(const SimpleDateFormatMutableNFs &);
357};
358
359//----------------------------------------------------------------------
360
361SimpleDateFormat::~SimpleDateFormat()
362{
363    delete fSymbols;
364    if (fSharedNumberFormatters) {
365        freeSharedNumberFormatters(fSharedNumberFormatters);
366    }
367    if (fTimeZoneFormat) {
368        delete fTimeZoneFormat;
369    }
370
371#if !UCONFIG_NO_BREAK_ITERATION
372    delete fCapitalizationBrkIter;
373#endif
374}
375
376//----------------------------------------------------------------------
377
378SimpleDateFormat::SimpleDateFormat(UErrorCode& status)
379  :   fLocale(Locale::getDefault()),
380      fSymbols(NULL),
381      fTimeZoneFormat(NULL),
382      fSharedNumberFormatters(NULL),
383      fCapitalizationBrkIter(NULL)
384{
385    initializeBooleanAttributes();
386    construct(kShort, (EStyle) (kShort + kDateOffset), fLocale, status);
387    initializeDefaultCentury();
388}
389
390//----------------------------------------------------------------------
391
392SimpleDateFormat::SimpleDateFormat(const UnicodeString& pattern,
393                                   UErrorCode &status)
394:   fPattern(pattern),
395    fLocale(Locale::getDefault()),
396    fSymbols(NULL),
397    fTimeZoneFormat(NULL),
398    fSharedNumberFormatters(NULL),
399    fCapitalizationBrkIter(NULL)
400{
401    fDateOverride.setToBogus();
402    fTimeOverride.setToBogus();
403    initializeBooleanAttributes();
404    initializeCalendar(NULL,fLocale,status);
405    fSymbols = DateFormatSymbols::createForLocale(fLocale, status);
406    initialize(fLocale, status);
407    initializeDefaultCentury();
408
409}
410//----------------------------------------------------------------------
411
412SimpleDateFormat::SimpleDateFormat(const UnicodeString& pattern,
413                                   const UnicodeString& override,
414                                   UErrorCode &status)
415:   fPattern(pattern),
416    fLocale(Locale::getDefault()),
417    fSymbols(NULL),
418    fTimeZoneFormat(NULL),
419    fSharedNumberFormatters(NULL),
420    fCapitalizationBrkIter(NULL)
421{
422    fDateOverride.setTo(override);
423    fTimeOverride.setToBogus();
424    initializeBooleanAttributes();
425    initializeCalendar(NULL,fLocale,status);
426    fSymbols = DateFormatSymbols::createForLocale(fLocale, status);
427    initialize(fLocale, status);
428    initializeDefaultCentury();
429
430    processOverrideString(fLocale,override,kOvrStrBoth,status);
431
432}
433
434//----------------------------------------------------------------------
435
436SimpleDateFormat::SimpleDateFormat(const UnicodeString& pattern,
437                                   const Locale& locale,
438                                   UErrorCode& status)
439:   fPattern(pattern),
440    fLocale(locale),
441    fTimeZoneFormat(NULL),
442    fSharedNumberFormatters(NULL),
443    fCapitalizationBrkIter(NULL)
444{
445
446    fDateOverride.setToBogus();
447    fTimeOverride.setToBogus();
448    initializeBooleanAttributes();
449
450    initializeCalendar(NULL,fLocale,status);
451    fSymbols = DateFormatSymbols::createForLocale(fLocale, status);
452    initialize(fLocale, status);
453    initializeDefaultCentury();
454}
455
456//----------------------------------------------------------------------
457
458SimpleDateFormat::SimpleDateFormat(const UnicodeString& pattern,
459                                   const UnicodeString& override,
460                                   const Locale& locale,
461                                   UErrorCode& status)
462:   fPattern(pattern),
463    fLocale(locale),
464    fTimeZoneFormat(NULL),
465    fSharedNumberFormatters(NULL),
466    fCapitalizationBrkIter(NULL)
467{
468
469    fDateOverride.setTo(override);
470    fTimeOverride.setToBogus();
471    initializeBooleanAttributes();
472
473    initializeCalendar(NULL,fLocale,status);
474    fSymbols = DateFormatSymbols::createForLocale(fLocale, status);
475    initialize(fLocale, status);
476    initializeDefaultCentury();
477
478    processOverrideString(locale,override,kOvrStrBoth,status);
479
480}
481
482//----------------------------------------------------------------------
483
484SimpleDateFormat::SimpleDateFormat(const UnicodeString& pattern,
485                                   DateFormatSymbols* symbolsToAdopt,
486                                   UErrorCode& status)
487:   fPattern(pattern),
488    fLocale(Locale::getDefault()),
489    fSymbols(symbolsToAdopt),
490    fTimeZoneFormat(NULL),
491    fSharedNumberFormatters(NULL),
492    fCapitalizationBrkIter(NULL)
493{
494
495    fDateOverride.setToBogus();
496    fTimeOverride.setToBogus();
497    initializeBooleanAttributes();
498
499    initializeCalendar(NULL,fLocale,status);
500    initialize(fLocale, status);
501    initializeDefaultCentury();
502}
503
504//----------------------------------------------------------------------
505
506SimpleDateFormat::SimpleDateFormat(const UnicodeString& pattern,
507                                   const DateFormatSymbols& symbols,
508                                   UErrorCode& status)
509:   fPattern(pattern),
510    fLocale(Locale::getDefault()),
511    fSymbols(new DateFormatSymbols(symbols)),
512    fTimeZoneFormat(NULL),
513    fSharedNumberFormatters(NULL),
514    fCapitalizationBrkIter(NULL)
515{
516
517    fDateOverride.setToBogus();
518    fTimeOverride.setToBogus();
519    initializeBooleanAttributes();
520
521    initializeCalendar(NULL, fLocale, status);
522    initialize(fLocale, status);
523    initializeDefaultCentury();
524}
525
526//----------------------------------------------------------------------
527
528// Not for public consumption; used by DateFormat
529SimpleDateFormat::SimpleDateFormat(EStyle timeStyle,
530                                   EStyle dateStyle,
531                                   const Locale& locale,
532                                   UErrorCode& status)
533:   fLocale(locale),
534    fSymbols(NULL),
535    fTimeZoneFormat(NULL),
536    fSharedNumberFormatters(NULL),
537    fCapitalizationBrkIter(NULL)
538{
539    initializeBooleanAttributes();
540    construct(timeStyle, dateStyle, fLocale, status);
541    if(U_SUCCESS(status)) {
542      initializeDefaultCentury();
543    }
544}
545
546//----------------------------------------------------------------------
547
548/**
549 * Not for public consumption; used by DateFormat.  This constructor
550 * never fails.  If the resource data is not available, it uses the
551 * the last resort symbols.
552 */
553SimpleDateFormat::SimpleDateFormat(const Locale& locale,
554                                   UErrorCode& status)
555:   fPattern(gDefaultPattern),
556    fLocale(locale),
557    fSymbols(NULL),
558    fTimeZoneFormat(NULL),
559    fSharedNumberFormatters(NULL),
560    fCapitalizationBrkIter(NULL)
561{
562    if (U_FAILURE(status)) return;
563    initializeBooleanAttributes();
564    initializeCalendar(NULL, fLocale, status);
565    fSymbols = DateFormatSymbols::createForLocale(fLocale, status);
566    if (U_FAILURE(status))
567    {
568        status = U_ZERO_ERROR;
569        delete fSymbols;
570        // This constructor doesn't fail; it uses last resort data
571        fSymbols = new DateFormatSymbols(status);
572        /* test for NULL */
573        if (fSymbols == 0) {
574            status = U_MEMORY_ALLOCATION_ERROR;
575            return;
576        }
577    }
578
579    fDateOverride.setToBogus();
580    fTimeOverride.setToBogus();
581
582    initialize(fLocale, status);
583    if(U_SUCCESS(status)) {
584      initializeDefaultCentury();
585    }
586}
587
588//----------------------------------------------------------------------
589
590SimpleDateFormat::SimpleDateFormat(const SimpleDateFormat& other)
591:   DateFormat(other),
592    fLocale(other.fLocale),
593    fSymbols(NULL),
594    fTimeZoneFormat(NULL),
595    fSharedNumberFormatters(NULL),
596    fCapitalizationBrkIter(NULL)
597{
598    initializeBooleanAttributes();
599    *this = other;
600}
601
602//----------------------------------------------------------------------
603
604SimpleDateFormat& SimpleDateFormat::operator=(const SimpleDateFormat& other)
605{
606    if (this == &other) {
607        return *this;
608    }
609    DateFormat::operator=(other);
610    fDateOverride = other.fDateOverride;
611    fTimeOverride = other.fTimeOverride;
612
613    delete fSymbols;
614    fSymbols = NULL;
615
616    if (other.fSymbols)
617        fSymbols = new DateFormatSymbols(*other.fSymbols);
618
619    fDefaultCenturyStart         = other.fDefaultCenturyStart;
620    fDefaultCenturyStartYear     = other.fDefaultCenturyStartYear;
621    fHaveDefaultCentury          = other.fHaveDefaultCentury;
622
623    fPattern = other.fPattern;
624
625    // TimeZoneFormat in ICU4C only depends on a locale for now
626    if (fLocale != other.fLocale) {
627        delete fTimeZoneFormat;
628        fTimeZoneFormat = NULL; // forces lazy instantiation with the other locale
629        fLocale = other.fLocale;
630    }
631
632#if !UCONFIG_NO_BREAK_ITERATION
633    if (other.fCapitalizationBrkIter != NULL) {
634        fCapitalizationBrkIter = (other.fCapitalizationBrkIter)->clone();
635    }
636#endif
637
638    if (fSharedNumberFormatters != NULL) {
639        freeSharedNumberFormatters(fSharedNumberFormatters);
640        fSharedNumberFormatters = NULL;
641    }
642    if (other.fSharedNumberFormatters != NULL) {
643        fSharedNumberFormatters = allocSharedNumberFormatters();
644        if (fSharedNumberFormatters) {
645            for (int32_t i = 0; i < UDAT_FIELD_COUNT; ++i) {
646                SharedObject::copyPtr(
647                        other.fSharedNumberFormatters[i],
648                        fSharedNumberFormatters[i]);
649            }
650        }
651    }
652
653    return *this;
654}
655
656//----------------------------------------------------------------------
657
658Format*
659SimpleDateFormat::clone() const
660{
661    return new SimpleDateFormat(*this);
662}
663
664//----------------------------------------------------------------------
665
666UBool
667SimpleDateFormat::operator==(const Format& other) const
668{
669    if (DateFormat::operator==(other)) {
670        // The DateFormat::operator== check for fCapitalizationContext equality above
671        //   is sufficient to check equality of all derived context-related data.
672        // DateFormat::operator== guarantees following cast is safe
673        SimpleDateFormat* that = (SimpleDateFormat*)&other;
674        return (fPattern             == that->fPattern &&
675                fSymbols             != NULL && // Check for pathological object
676                that->fSymbols       != NULL && // Check for pathological object
677                *fSymbols            == *that->fSymbols &&
678                fHaveDefaultCentury  == that->fHaveDefaultCentury &&
679                fDefaultCenturyStart == that->fDefaultCenturyStart);
680    }
681    return FALSE;
682}
683
684//----------------------------------------------------------------------
685
686void SimpleDateFormat::construct(EStyle timeStyle,
687                                 EStyle dateStyle,
688                                 const Locale& locale,
689                                 UErrorCode& status)
690{
691    // called by several constructors to load pattern data from the resources
692    if (U_FAILURE(status)) return;
693
694    // We will need the calendar to know what type of symbols to load.
695    initializeCalendar(NULL, locale, status);
696    if (U_FAILURE(status)) return;
697
698    CalendarData calData(locale, fCalendar?fCalendar->getType():NULL, status);
699    UResourceBundle *dateTimePatterns = calData.getByKey(gDateTimePatternsTag, status);
700    UResourceBundle *currentBundle;
701
702    if (U_FAILURE(status)) return;
703
704    if (ures_getSize(dateTimePatterns) <= kDateTime)
705    {
706        status = U_INVALID_FORMAT_ERROR;
707        return;
708    }
709
710    setLocaleIDs(ures_getLocaleByType(dateTimePatterns, ULOC_VALID_LOCALE, &status),
711                 ures_getLocaleByType(dateTimePatterns, ULOC_ACTUAL_LOCALE, &status));
712
713    // create a symbols object from the locale
714    fSymbols = DateFormatSymbols::createForLocale(locale, status);
715    if (U_FAILURE(status)) return;
716    /* test for NULL */
717    if (fSymbols == 0) {
718        status = U_MEMORY_ALLOCATION_ERROR;
719        return;
720    }
721
722    const UChar *resStr,*ovrStr;
723    int32_t resStrLen,ovrStrLen = 0;
724    fDateOverride.setToBogus();
725    fTimeOverride.setToBogus();
726
727    // if the pattern should include both date and time information, use the date/time
728    // pattern string as a guide to tell use how to glue together the appropriate date
729    // and time pattern strings.  The actual gluing-together is handled by a convenience
730    // method on MessageFormat.
731    if ((timeStyle != kNone) && (dateStyle != kNone))
732    {
733        Formattable timeDateArray[2];
734
735        // use Formattable::adoptString() so that we can use fastCopyFrom()
736        // instead of Formattable::setString()'s unaware, safe, deep string clone
737        // see Jitterbug 2296
738
739        currentBundle = ures_getByIndex(dateTimePatterns, (int32_t)timeStyle, NULL, &status);
740        if (U_FAILURE(status)) {
741           status = U_INVALID_FORMAT_ERROR;
742           return;
743        }
744        switch (ures_getType(currentBundle)) {
745            case URES_STRING: {
746               resStr = ures_getString(currentBundle, &resStrLen, &status);
747               break;
748            }
749            case URES_ARRAY: {
750               resStr = ures_getStringByIndex(currentBundle, 0, &resStrLen, &status);
751               ovrStr = ures_getStringByIndex(currentBundle, 1, &ovrStrLen, &status);
752               fTimeOverride.setTo(TRUE, ovrStr, ovrStrLen);
753               break;
754            }
755            default: {
756               status = U_INVALID_FORMAT_ERROR;
757               ures_close(currentBundle);
758               return;
759            }
760        }
761        ures_close(currentBundle);
762
763        UnicodeString *tempus1 = new UnicodeString(TRUE, resStr, resStrLen);
764        // NULL pointer check
765        if (tempus1 == NULL) {
766            status = U_MEMORY_ALLOCATION_ERROR;
767            return;
768        }
769        timeDateArray[0].adoptString(tempus1);
770
771        currentBundle = ures_getByIndex(dateTimePatterns, (int32_t)dateStyle, NULL, &status);
772        if (U_FAILURE(status)) {
773           status = U_INVALID_FORMAT_ERROR;
774           return;
775        }
776        switch (ures_getType(currentBundle)) {
777            case URES_STRING: {
778               resStr = ures_getString(currentBundle, &resStrLen, &status);
779               break;
780            }
781            case URES_ARRAY: {
782               resStr = ures_getStringByIndex(currentBundle, 0, &resStrLen, &status);
783               ovrStr = ures_getStringByIndex(currentBundle, 1, &ovrStrLen, &status);
784               fDateOverride.setTo(TRUE, ovrStr, ovrStrLen);
785               break;
786            }
787            default: {
788               status = U_INVALID_FORMAT_ERROR;
789               ures_close(currentBundle);
790               return;
791            }
792        }
793        ures_close(currentBundle);
794
795        UnicodeString *tempus2 = new UnicodeString(TRUE, resStr, resStrLen);
796        // Null pointer check
797        if (tempus2 == NULL) {
798            status = U_MEMORY_ALLOCATION_ERROR;
799            return;
800        }
801        timeDateArray[1].adoptString(tempus2);
802
803        int32_t glueIndex = kDateTime;
804        int32_t patternsSize = ures_getSize(dateTimePatterns);
805        if (patternsSize >= (kDateTimeOffset + kShort + 1)) {
806            // Get proper date time format
807            glueIndex = (int32_t)(kDateTimeOffset + (dateStyle - kDateOffset));
808        }
809
810        resStr = ures_getStringByIndex(dateTimePatterns, glueIndex, &resStrLen, &status);
811        MessageFormat::format(UnicodeString(TRUE, resStr, resStrLen), timeDateArray, 2, fPattern, status);
812    }
813    // if the pattern includes just time data or just date date, load the appropriate
814    // pattern string from the resources
815    // setTo() - see DateFormatSymbols::assignArray comments
816    else if (timeStyle != kNone) {
817        currentBundle = ures_getByIndex(dateTimePatterns, (int32_t)timeStyle, NULL, &status);
818        if (U_FAILURE(status)) {
819           status = U_INVALID_FORMAT_ERROR;
820           return;
821        }
822        switch (ures_getType(currentBundle)) {
823            case URES_STRING: {
824               resStr = ures_getString(currentBundle, &resStrLen, &status);
825               break;
826            }
827            case URES_ARRAY: {
828               resStr = ures_getStringByIndex(currentBundle, 0, &resStrLen, &status);
829               ovrStr = ures_getStringByIndex(currentBundle, 1, &ovrStrLen, &status);
830               fDateOverride.setTo(TRUE, ovrStr, ovrStrLen);
831               break;
832            }
833            default: {
834               status = U_INVALID_FORMAT_ERROR;
835                ures_close(currentBundle);
836               return;
837            }
838        }
839        fPattern.setTo(TRUE, resStr, resStrLen);
840        ures_close(currentBundle);
841    }
842    else if (dateStyle != kNone) {
843        currentBundle = ures_getByIndex(dateTimePatterns, (int32_t)dateStyle, NULL, &status);
844        if (U_FAILURE(status)) {
845           status = U_INVALID_FORMAT_ERROR;
846           return;
847        }
848        switch (ures_getType(currentBundle)) {
849            case URES_STRING: {
850               resStr = ures_getString(currentBundle, &resStrLen, &status);
851               break;
852            }
853            case URES_ARRAY: {
854               resStr = ures_getStringByIndex(currentBundle, 0, &resStrLen, &status);
855               ovrStr = ures_getStringByIndex(currentBundle, 1, &ovrStrLen, &status);
856               fDateOverride.setTo(TRUE, ovrStr, ovrStrLen);
857               break;
858            }
859            default: {
860               status = U_INVALID_FORMAT_ERROR;
861               ures_close(currentBundle);
862               return;
863            }
864        }
865        fPattern.setTo(TRUE, resStr, resStrLen);
866        ures_close(currentBundle);
867    }
868
869    // and if it includes _neither_, that's an error
870    else
871        status = U_INVALID_FORMAT_ERROR;
872
873    // finally, finish initializing by creating a Calendar and a NumberFormat
874    initialize(locale, status);
875}
876
877//----------------------------------------------------------------------
878
879Calendar*
880SimpleDateFormat::initializeCalendar(TimeZone* adoptZone, const Locale& locale, UErrorCode& status)
881{
882    if(!U_FAILURE(status)) {
883        fCalendar = Calendar::createInstance(adoptZone?adoptZone:TimeZone::createDefault(), locale, status);
884    }
885    return fCalendar;
886}
887
888void
889SimpleDateFormat::initialize(const Locale& locale,
890                             UErrorCode& status)
891{
892    if (U_FAILURE(status)) return;
893
894    // We don't need to check that the row count is >= 1, since all 2d arrays have at
895    // least one row
896    fNumberFormat = NumberFormat::createInstance(locale, status);
897    if (fNumberFormat != NULL && U_SUCCESS(status))
898    {
899        fixNumberFormatForDates(*fNumberFormat);
900        //fNumberFormat->setLenient(TRUE); // Java uses a custom DateNumberFormat to format/parse
901
902        initNumberFormatters(locale,status);
903
904    }
905    else if (U_SUCCESS(status))
906    {
907        status = U_MISSING_RESOURCE_ERROR;
908    }
909}
910
911/* Initialize the fields we use to disambiguate ambiguous years. Separate
912 * so we can call it from readObject().
913 */
914void SimpleDateFormat::initializeDefaultCentury()
915{
916  if(fCalendar) {
917    fHaveDefaultCentury = fCalendar->haveDefaultCentury();
918    if(fHaveDefaultCentury) {
919      fDefaultCenturyStart = fCalendar->defaultCenturyStart();
920      fDefaultCenturyStartYear = fCalendar->defaultCenturyStartYear();
921    } else {
922      fDefaultCenturyStart = DBL_MIN;
923      fDefaultCenturyStartYear = -1;
924    }
925  }
926}
927
928/*
929 * Initialize the boolean attributes. Separate so we can call it from all constructors.
930 */
931void SimpleDateFormat::initializeBooleanAttributes()
932{
933    UErrorCode status = U_ZERO_ERROR;
934
935    setBooleanAttribute(UDAT_PARSE_ALLOW_WHITESPACE, true, status);
936    setBooleanAttribute(UDAT_PARSE_ALLOW_NUMERIC, true, status);
937    setBooleanAttribute(UDAT_PARSE_PARTIAL_LITERAL_MATCH, true, status);
938    setBooleanAttribute(UDAT_PARSE_MULTIPLE_PATTERNS_FOR_MATCH, true, status);
939}
940
941/* Define one-century window into which to disambiguate dates using
942 * two-digit years. Make public in JDK 1.2.
943 */
944void SimpleDateFormat::parseAmbiguousDatesAsAfter(UDate startDate, UErrorCode& status)
945{
946    if(U_FAILURE(status)) {
947        return;
948    }
949    if(!fCalendar) {
950      status = U_ILLEGAL_ARGUMENT_ERROR;
951      return;
952    }
953
954    fCalendar->setTime(startDate, status);
955    if(U_SUCCESS(status)) {
956        fHaveDefaultCentury = TRUE;
957        fDefaultCenturyStart = startDate;
958        fDefaultCenturyStartYear = fCalendar->get(UCAL_YEAR, status);
959    }
960}
961
962//----------------------------------------------------------------------
963
964UnicodeString&
965SimpleDateFormat::format(Calendar& cal, UnicodeString& appendTo, FieldPosition& pos) const
966{
967  UErrorCode status = U_ZERO_ERROR;
968  FieldPositionOnlyHandler handler(pos);
969  return _format(cal, appendTo, handler, status);
970}
971
972//----------------------------------------------------------------------
973
974UnicodeString&
975SimpleDateFormat::format(Calendar& cal, UnicodeString& appendTo,
976                         FieldPositionIterator* posIter, UErrorCode& status) const
977{
978  FieldPositionIteratorHandler handler(posIter, status);
979  return _format(cal, appendTo, handler, status);
980}
981
982//----------------------------------------------------------------------
983
984UnicodeString&
985SimpleDateFormat::_format(Calendar& cal, UnicodeString& appendTo,
986                            FieldPositionHandler& handler, UErrorCode& status) const
987{
988    if ( U_FAILURE(status) ) {
989       return appendTo;
990    }
991    Calendar* workCal = &cal;
992    Calendar* calClone = NULL;
993    if (&cal != fCalendar && uprv_strcmp(cal.getType(), fCalendar->getType()) != 0) {
994        // Different calendar type
995        // We use the time and time zone from the input calendar, but
996        // do not use the input calendar for field calculation.
997        calClone = fCalendar->clone();
998        if (calClone != NULL) {
999            UDate t = cal.getTime(status);
1000            calClone->setTime(t, status);
1001            calClone->setTimeZone(cal.getTimeZone());
1002            workCal = calClone;
1003        } else {
1004            status = U_MEMORY_ALLOCATION_ERROR;
1005            return appendTo;
1006        }
1007    }
1008
1009    UBool inQuote = FALSE;
1010    UChar prevCh = 0;
1011    int32_t count = 0;
1012    int32_t fieldNum = 0;
1013    UDisplayContext capitalizationContext = getContext(UDISPCTX_TYPE_CAPITALIZATION, status);
1014
1015    // Create temporary cache of mutable number format objects. This way
1016    // subFormat won't have to clone the const NumberFormat for each field.
1017    // if several fields share the same NumberFormat, which will almost
1018    // always be the case, this is a big save.
1019    SimpleDateFormatMutableNFs mutableNFs;
1020    // loop through the pattern string character by character
1021    for (int32_t i = 0; i < fPattern.length() && U_SUCCESS(status); ++i) {
1022        UChar ch = fPattern[i];
1023
1024        // Use subFormat() to format a repeated pattern character
1025        // when a different pattern or non-pattern character is seen
1026        if (ch != prevCh && count > 0) {
1027            subFormat(appendTo, prevCh, count, capitalizationContext, fieldNum++, handler, *workCal, mutableNFs, status);
1028            count = 0;
1029        }
1030        if (ch == QUOTE) {
1031            // Consecutive single quotes are a single quote literal,
1032            // either outside of quotes or between quotes
1033            if ((i+1) < fPattern.length() && fPattern[i+1] == QUOTE) {
1034                appendTo += (UChar)QUOTE;
1035                ++i;
1036            } else {
1037                inQuote = ! inQuote;
1038            }
1039        }
1040        else if (!inQuote && isSyntaxChar(ch)) {
1041            // ch is a date-time pattern character to be interpreted
1042            // by subFormat(); count the number of times it is repeated
1043            prevCh = ch;
1044            ++count;
1045        }
1046        else {
1047            // Append quoted characters and unquoted non-pattern characters
1048            appendTo += ch;
1049        }
1050    }
1051
1052    // Format the last item in the pattern, if any
1053    if (count > 0) {
1054        subFormat(appendTo, prevCh, count, capitalizationContext, fieldNum++, handler, *workCal, mutableNFs, status);
1055    }
1056
1057    if (calClone != NULL) {
1058        delete calClone;
1059    }
1060
1061    return appendTo;
1062}
1063
1064//----------------------------------------------------------------------
1065
1066/* Map calendar field into calendar field level.
1067 * the larger the level, the smaller the field unit.
1068 * For example, UCAL_ERA level is 0, UCAL_YEAR level is 10,
1069 * UCAL_MONTH level is 20.
1070 * NOTE: if new fields adds in, the table needs to update.
1071 */
1072const int32_t
1073SimpleDateFormat::fgCalendarFieldToLevel[] =
1074{
1075    /*GyM*/ 0, 10, 20,
1076    /*wW*/ 20, 30,
1077    /*dDEF*/ 30, 20, 30, 30,
1078    /*ahHm*/ 40, 50, 50, 60,
1079    /*sS*/ 70, 80,
1080    /*z?Y*/ 0, 0, 10,
1081    /*eug*/ 30, 10, 0,
1082    /*A?.*/ 40, 0, 0
1083};
1084
1085int32_t SimpleDateFormat::getLevelFromChar(UChar ch) {
1086    // Map date field LETTER into calendar field level.
1087    // the larger the level, the smaller the field unit.
1088    // NOTE: if new fields adds in, the table needs to update.
1089    static const int32_t mapCharToLevel[] = {
1090            -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
1091        //
1092            -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
1093        //       !   "   #   $   %   &   '   (   )   *   +   ,   -   .   /
1094            -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
1095#if UDAT_HAS_PATTERN_CHAR_FOR_TIME_SEPARATOR
1096        //   0   1   2   3   4   5   6   7   8   9   :   ;   <   =   >   ?
1097            -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,  0, -1, -1, -1, -1, -1,
1098#else
1099        //   0   1   2   3   4   5   6   7   8   9   :   ;   <   =   >   ?
1100            -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
1101#endif
1102        //   @   A   B   C   D   E   F   G   H   I   J   K   L   M   N   O
1103            -1, 40, -1, -1, 20, 30, 30,  0, 50, -1, -1, 50, 20, 20, -1,  0,
1104        //   P   Q   R   S   T   U   V   W   X   Y   Z   [   \   ]   ^   _
1105            -1, 20, -1, 80, -1, 10,  0, 30,  0, 10,  0, -1, -1, -1, -1, -1,
1106        //   `   a   b   c   d   e   f   g   h   i   j   k   l   m   n   o
1107            -1, 40, -1, 30, 30, 30, -1,  0, 50, -1, -1, 50,  0, 60, -1, -1,
1108        //   p   q   r   s   t   u   v   w   x   y   z   {   |   }   ~
1109            -1, 20, 10, 70, -1, 10,  0, 20,  0, 10,  0, -1, -1, -1, -1, -1
1110    };
1111
1112    return ch < UPRV_LENGTHOF(mapCharToLevel) ? mapCharToLevel[ch] : -1;
1113}
1114
1115UBool SimpleDateFormat::isSyntaxChar(UChar ch) {
1116    static const UBool mapCharToIsSyntax[] = {
1117        //
1118        FALSE, FALSE, FALSE, FALSE, FALSE, FALSE, FALSE, FALSE,
1119        //
1120        FALSE, FALSE, FALSE, FALSE, FALSE, FALSE, FALSE, FALSE,
1121        //
1122        FALSE, FALSE, FALSE, FALSE, FALSE, FALSE, FALSE, FALSE,
1123        //
1124        FALSE, FALSE, FALSE, FALSE, FALSE, FALSE, FALSE, FALSE,
1125        //         !      "      #      $      %      &      '
1126        FALSE, FALSE, FALSE, FALSE, FALSE, FALSE, FALSE, FALSE,
1127        //  (      )      *      +      ,      -      .      /
1128        FALSE, FALSE, FALSE, FALSE, FALSE, FALSE, FALSE, FALSE,
1129        //  0      1      2      3      4      5      6      7
1130        FALSE, FALSE, FALSE, FALSE, FALSE, FALSE, FALSE, FALSE,
1131#if UDAT_HAS_PATTERN_CHAR_FOR_TIME_SEPARATOR
1132        //  8      9      :      ;      <      =      >      ?
1133        FALSE, FALSE,  TRUE, FALSE, FALSE, FALSE, FALSE, FALSE,
1134#else
1135        //  8      9      :      ;      <      =      >      ?
1136        FALSE, FALSE, FALSE, FALSE, FALSE, FALSE, FALSE, FALSE,
1137#endif
1138        //  @      A      B      C      D      E      F      G
1139        FALSE,  TRUE,  TRUE,  TRUE,  TRUE,  TRUE,  TRUE,  TRUE,
1140        //  H      I      J      K      L      M      N      O
1141         TRUE,  TRUE,  TRUE,  TRUE,  TRUE,  TRUE,  TRUE,  TRUE,
1142        //  P      Q      R      S      T      U      V      W
1143         TRUE,  TRUE,  TRUE,  TRUE,  TRUE,  TRUE,  TRUE,  TRUE,
1144        //  X      Y      Z      [      \      ]      ^      _
1145         TRUE,  TRUE,  TRUE, FALSE, FALSE, FALSE, FALSE, FALSE,
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    };
1155
1156    return ch < UPRV_LENGTHOF(mapCharToIsSyntax) ? mapCharToIsSyntax[ch] : FALSE;
1157}
1158
1159// Map index into pattern character string to Calendar field number.
1160const UCalendarDateFields
1161SimpleDateFormat::fgPatternIndexToCalendarField[] =
1162{
1163    /*GyM*/ UCAL_ERA, UCAL_YEAR, UCAL_MONTH,
1164    /*dkH*/ UCAL_DATE, UCAL_HOUR_OF_DAY, UCAL_HOUR_OF_DAY,
1165    /*msS*/ UCAL_MINUTE, UCAL_SECOND, UCAL_MILLISECOND,
1166    /*EDF*/ UCAL_DAY_OF_WEEK, UCAL_DAY_OF_YEAR, UCAL_DAY_OF_WEEK_IN_MONTH,
1167    /*wWa*/ UCAL_WEEK_OF_YEAR, UCAL_WEEK_OF_MONTH, UCAL_AM_PM,
1168    /*hKz*/ UCAL_HOUR, UCAL_HOUR, UCAL_ZONE_OFFSET,
1169    /*Yeu*/ UCAL_YEAR_WOY, UCAL_DOW_LOCAL, UCAL_EXTENDED_YEAR,
1170    /*gAZ*/ UCAL_JULIAN_DAY, UCAL_MILLISECONDS_IN_DAY, UCAL_ZONE_OFFSET,
1171    /*v*/   UCAL_ZONE_OFFSET,
1172    /*c*/   UCAL_DOW_LOCAL,
1173    /*L*/   UCAL_MONTH,
1174    /*Q*/   UCAL_MONTH,
1175    /*q*/   UCAL_MONTH,
1176    /*V*/   UCAL_ZONE_OFFSET,
1177    /*U*/   UCAL_YEAR,
1178    /*O*/   UCAL_ZONE_OFFSET,
1179    /*Xx*/  UCAL_ZONE_OFFSET, UCAL_ZONE_OFFSET,
1180    /*r*/   UCAL_EXTENDED_YEAR,
1181#if UDAT_HAS_PATTERN_CHAR_FOR_TIME_SEPARATOR
1182    /*:*/   UCAL_FIELD_COUNT, /* => no useful mapping to any calendar field */
1183#else
1184    /*no pattern char for UDAT_TIME_SEPARATOR_FIELD*/   UCAL_FIELD_COUNT, /* => no useful mapping to any calendar field */
1185#endif
1186};
1187
1188// Map index into pattern character string to DateFormat field number
1189const UDateFormatField
1190SimpleDateFormat::fgPatternIndexToDateFormatField[] = {
1191    /*GyM*/ UDAT_ERA_FIELD, UDAT_YEAR_FIELD, UDAT_MONTH_FIELD,
1192    /*dkH*/ UDAT_DATE_FIELD, UDAT_HOUR_OF_DAY1_FIELD, UDAT_HOUR_OF_DAY0_FIELD,
1193    /*msS*/ UDAT_MINUTE_FIELD, UDAT_SECOND_FIELD, UDAT_FRACTIONAL_SECOND_FIELD,
1194    /*EDF*/ UDAT_DAY_OF_WEEK_FIELD, UDAT_DAY_OF_YEAR_FIELD, UDAT_DAY_OF_WEEK_IN_MONTH_FIELD,
1195    /*wWa*/ UDAT_WEEK_OF_YEAR_FIELD, UDAT_WEEK_OF_MONTH_FIELD, UDAT_AM_PM_FIELD,
1196    /*hKz*/ UDAT_HOUR1_FIELD, UDAT_HOUR0_FIELD, UDAT_TIMEZONE_FIELD,
1197    /*Yeu*/ UDAT_YEAR_WOY_FIELD, UDAT_DOW_LOCAL_FIELD, UDAT_EXTENDED_YEAR_FIELD,
1198    /*gAZ*/ UDAT_JULIAN_DAY_FIELD, UDAT_MILLISECONDS_IN_DAY_FIELD, UDAT_TIMEZONE_RFC_FIELD,
1199    /*v*/   UDAT_TIMEZONE_GENERIC_FIELD,
1200    /*c*/   UDAT_STANDALONE_DAY_FIELD,
1201    /*L*/   UDAT_STANDALONE_MONTH_FIELD,
1202    /*Q*/   UDAT_QUARTER_FIELD,
1203    /*q*/   UDAT_STANDALONE_QUARTER_FIELD,
1204    /*V*/   UDAT_TIMEZONE_SPECIAL_FIELD,
1205    /*U*/   UDAT_YEAR_NAME_FIELD,
1206    /*O*/   UDAT_TIMEZONE_LOCALIZED_GMT_OFFSET_FIELD,
1207    /*Xx*/  UDAT_TIMEZONE_ISO_FIELD, UDAT_TIMEZONE_ISO_LOCAL_FIELD,
1208    /*r*/   UDAT_RELATED_YEAR_FIELD,
1209#if UDAT_HAS_PATTERN_CHAR_FOR_TIME_SEPARATOR
1210    /*:*/   UDAT_TIME_SEPARATOR_FIELD,
1211#else
1212    /*no pattern char for UDAT_TIME_SEPARATOR_FIELD*/   UDAT_TIME_SEPARATOR_FIELD,
1213#endif
1214};
1215
1216//----------------------------------------------------------------------
1217
1218/**
1219 * Append symbols[value] to dst.  Make sure the array index is not out
1220 * of bounds.
1221 */
1222static inline void
1223_appendSymbol(UnicodeString& dst,
1224              int32_t value,
1225              const UnicodeString* symbols,
1226              int32_t symbolsCount) {
1227    U_ASSERT(0 <= value && value < symbolsCount);
1228    if (0 <= value && value < symbolsCount) {
1229        dst += symbols[value];
1230    }
1231}
1232
1233static inline void
1234_appendSymbolWithMonthPattern(UnicodeString& dst, int32_t value, const UnicodeString* symbols, int32_t symbolsCount,
1235              const UnicodeString* monthPattern, UErrorCode& status) {
1236    U_ASSERT(0 <= value && value < symbolsCount);
1237    if (0 <= value && value < symbolsCount) {
1238        if (monthPattern == NULL) {
1239            dst += symbols[value];
1240        } else {
1241            Formattable monthName((const UnicodeString&)(symbols[value]));
1242            MessageFormat::format(*monthPattern, &monthName, 1, dst, status);
1243        }
1244    }
1245}
1246
1247//----------------------------------------------------------------------
1248void
1249SimpleDateFormat::initNumberFormatters(const Locale &locale,UErrorCode &status) {
1250    if (U_FAILURE(status)) {
1251        return;
1252    }
1253    if ( fDateOverride.isBogus() && fTimeOverride.isBogus() ) {
1254        return;
1255    }
1256    umtx_lock(&LOCK);
1257    if (fSharedNumberFormatters == NULL) {
1258        fSharedNumberFormatters = allocSharedNumberFormatters();
1259        if (fSharedNumberFormatters == NULL) {
1260            status = U_MEMORY_ALLOCATION_ERROR;
1261        }
1262    }
1263    umtx_unlock(&LOCK);
1264
1265    if (U_FAILURE(status)) {
1266        return;
1267    }
1268
1269    processOverrideString(locale,fDateOverride,kOvrStrDate,status);
1270    processOverrideString(locale,fTimeOverride,kOvrStrTime,status);
1271}
1272
1273void
1274SimpleDateFormat::processOverrideString(const Locale &locale, const UnicodeString &str, int8_t type, UErrorCode &status) {
1275    if (str.isBogus() || U_FAILURE(status)) {
1276        return;
1277    }
1278
1279    int32_t start = 0;
1280    int32_t len;
1281    UnicodeString nsName;
1282    UnicodeString ovrField;
1283    UBool moreToProcess = TRUE;
1284    NSOverride *overrideList = NULL;
1285
1286    while (moreToProcess) {
1287        int32_t delimiterPosition = str.indexOf((UChar)ULOC_KEYWORD_ITEM_SEPARATOR_UNICODE,start);
1288        if (delimiterPosition == -1) {
1289            moreToProcess = FALSE;
1290            len = str.length() - start;
1291        } else {
1292            len = delimiterPosition - start;
1293        }
1294        UnicodeString currentString(str,start,len);
1295        int32_t equalSignPosition = currentString.indexOf((UChar)ULOC_KEYWORD_ASSIGN_UNICODE,0);
1296        if (equalSignPosition == -1) { // Simple override string such as "hebrew"
1297            nsName.setTo(currentString);
1298            ovrField.setToBogus();
1299        } else { // Field specific override string such as "y=hebrew"
1300            nsName.setTo(currentString,equalSignPosition+1);
1301            ovrField.setTo(currentString,0,1); // We just need the first character.
1302        }
1303
1304        int32_t nsNameHash = nsName.hashCode();
1305        // See if the numbering system is in the override list, if not, then add it.
1306        NSOverride *cur = overrideList;
1307        const SharedNumberFormat *snf = NULL;
1308        UBool found = FALSE;
1309        while ( cur && !found ) {
1310            if ( cur->hash == nsNameHash ) {
1311                snf = cur->snf;
1312                found = TRUE;
1313            }
1314            cur = cur->next;
1315        }
1316
1317        if (!found) {
1318           LocalPointer<NSOverride> cur(new NSOverride);
1319           if (!cur.isNull()) {
1320               char kw[ULOC_KEYWORD_AND_VALUES_CAPACITY];
1321               uprv_strcpy(kw,"numbers=");
1322               nsName.extract(0,len,kw+8,ULOC_KEYWORD_AND_VALUES_CAPACITY-8,US_INV);
1323
1324               Locale ovrLoc(locale.getLanguage(),locale.getCountry(),locale.getVariant(),kw);
1325               cur->hash = nsNameHash;
1326               cur->next = overrideList;
1327               SharedObject::copyPtr(
1328                       createSharedNumberFormat(ovrLoc, status), cur->snf);
1329               if (U_FAILURE(status)) {
1330                   if (overrideList) {
1331                       overrideList->free();
1332                   }
1333                   return;
1334               }
1335               snf = cur->snf;
1336               overrideList = cur.orphan();
1337           } else {
1338               status = U_MEMORY_ALLOCATION_ERROR;
1339               if (overrideList) {
1340                   overrideList->free();
1341               }
1342               return;
1343           }
1344        }
1345
1346        // Now that we have an appropriate number formatter, fill in the appropriate spaces in the
1347        // number formatters table.
1348        if (ovrField.isBogus()) {
1349            switch (type) {
1350                case kOvrStrDate:
1351                case kOvrStrBoth: {
1352                    for ( int8_t i=0 ; i<kDateFieldsCount; i++ ) {
1353                        SharedObject::copyPtr(snf, fSharedNumberFormatters[kDateFields[i]]);
1354                    }
1355                    if (type==kOvrStrDate) {
1356                        break;
1357                    }
1358                }
1359                case kOvrStrTime : {
1360                    for ( int8_t i=0 ; i<kTimeFieldsCount; i++ ) {
1361                        SharedObject::copyPtr(snf, fSharedNumberFormatters[kTimeFields[i]]);
1362                    }
1363                    break;
1364                }
1365            }
1366        } else {
1367           // if the pattern character is unrecognized, signal an error and bail out
1368           UDateFormatField patternCharIndex =
1369              DateFormatSymbols::getPatternCharIndex(ovrField.charAt(0));
1370           if (patternCharIndex == UDAT_FIELD_COUNT) {
1371               status = U_INVALID_FORMAT_ERROR;
1372               if (overrideList) {
1373                   overrideList->free();
1374               }
1375               return;
1376           }
1377           SharedObject::copyPtr(snf, fSharedNumberFormatters[patternCharIndex]);
1378        }
1379
1380        start = delimiterPosition + 1;
1381    }
1382    if (overrideList) {
1383        overrideList->free();
1384    }
1385}
1386
1387//---------------------------------------------------------------------
1388void
1389SimpleDateFormat::subFormat(UnicodeString &appendTo,
1390                            UChar ch,
1391                            int32_t count,
1392                            UDisplayContext capitalizationContext,
1393                            int32_t fieldNum,
1394                            FieldPositionHandler& handler,
1395                            Calendar& cal,
1396                            SimpleDateFormatMutableNFs &mutableNFs,
1397                            UErrorCode& status) const
1398{
1399    if (U_FAILURE(status)) {
1400        return;
1401    }
1402
1403    // this function gets called by format() to produce the appropriate substitution
1404    // text for an individual pattern symbol (e.g., "HH" or "yyyy")
1405
1406    UDateFormatField patternCharIndex = DateFormatSymbols::getPatternCharIndex(ch);
1407    const int32_t maxIntCount = 10;
1408    int32_t beginOffset = appendTo.length();
1409    NumberFormat *currentNumberFormat;
1410    DateFormatSymbols::ECapitalizationContextUsageType capContextUsageType = DateFormatSymbols::kCapContextUsageOther;
1411
1412    UBool isHebrewCalendar = (uprv_strcmp(cal.getType(),"hebrew") == 0);
1413    UBool isChineseCalendar = (uprv_strcmp(cal.getType(),"chinese") == 0 || uprv_strcmp(cal.getType(),"dangi") == 0);
1414
1415    // if the pattern character is unrecognized, signal an error and dump out
1416    if (patternCharIndex == UDAT_FIELD_COUNT)
1417    {
1418        if (ch != 0x6C) { // pattern char 'l' (SMALL LETTER L) just gets ignored
1419            status = U_INVALID_FORMAT_ERROR;
1420        }
1421        return;
1422    }
1423
1424    UCalendarDateFields field = fgPatternIndexToCalendarField[patternCharIndex];
1425    int32_t value = 0;
1426    // Don't get value unless it is useful
1427    if (field < UCAL_FIELD_COUNT) {
1428        value = (patternCharIndex != UDAT_RELATED_YEAR_FIELD)? cal.get(field, status): cal.getRelatedYear(status);
1429    }
1430    if (U_FAILURE(status)) {
1431        return;
1432    }
1433
1434    currentNumberFormat = mutableNFs.get(getNumberFormatByIndex(patternCharIndex));
1435    if (currentNumberFormat == NULL) {
1436        status = U_MEMORY_ALLOCATION_ERROR;
1437        return;
1438    }
1439    UnicodeString hebr("hebr", 4, US_INV);
1440
1441    switch (patternCharIndex) {
1442
1443    // for any "G" symbol, write out the appropriate era string
1444    // "GGGG" is wide era name, "GGGGG" is narrow era name, anything else is abbreviated name
1445    case UDAT_ERA_FIELD:
1446        if (isChineseCalendar) {
1447            zeroPaddingNumber(currentNumberFormat,appendTo, value, 1, 9); // as in ICU4J
1448        } else {
1449            if (count == 5) {
1450                _appendSymbol(appendTo, value, fSymbols->fNarrowEras, fSymbols->fNarrowErasCount);
1451                capContextUsageType = DateFormatSymbols::kCapContextUsageEraNarrow;
1452            } else if (count == 4) {
1453                _appendSymbol(appendTo, value, fSymbols->fEraNames, fSymbols->fEraNamesCount);
1454                capContextUsageType = DateFormatSymbols::kCapContextUsageEraWide;
1455            } else {
1456                _appendSymbol(appendTo, value, fSymbols->fEras, fSymbols->fErasCount);
1457                capContextUsageType = DateFormatSymbols::kCapContextUsageEraAbbrev;
1458            }
1459        }
1460        break;
1461
1462     case UDAT_YEAR_NAME_FIELD:
1463        if (fSymbols->fShortYearNames != NULL && value <= fSymbols->fShortYearNamesCount) {
1464            // the Calendar YEAR field runs 1 through 60 for cyclic years
1465            _appendSymbol(appendTo, value - 1, fSymbols->fShortYearNames, fSymbols->fShortYearNamesCount);
1466            break;
1467        }
1468        // else fall through to numeric year handling, do not break here
1469
1470   // OLD: for "yyyy", write out the whole year; for "yy", write out the last 2 digits
1471    // NEW: UTS#35:
1472//Year         y     yy     yyy     yyyy     yyyyy
1473//AD 1         1     01     001     0001     00001
1474//AD 12       12     12     012     0012     00012
1475//AD 123     123     23     123     0123     00123
1476//AD 1234   1234     34    1234     1234     01234
1477//AD 12345 12345     45   12345    12345     12345
1478    case UDAT_YEAR_FIELD:
1479    case UDAT_YEAR_WOY_FIELD:
1480        if (fDateOverride.compare(hebr)==0 && value>HEBREW_CAL_CUR_MILLENIUM_START_YEAR && value<HEBREW_CAL_CUR_MILLENIUM_END_YEAR) {
1481            value-=HEBREW_CAL_CUR_MILLENIUM_START_YEAR;
1482        }
1483        if(count == 2)
1484            zeroPaddingNumber(currentNumberFormat, appendTo, value, 2, 2);
1485        else
1486            zeroPaddingNumber(currentNumberFormat, appendTo, value, count, maxIntCount);
1487        break;
1488
1489    // for "MMMM"/"LLLL", write out the whole month name, for "MMM"/"LLL", write out the month
1490    // abbreviation, for "M"/"L" or "MM"/"LL", write out the month as a number with the
1491    // appropriate number of digits
1492    // for "MMMMM"/"LLLLL", use the narrow form
1493    case UDAT_MONTH_FIELD:
1494    case UDAT_STANDALONE_MONTH_FIELD:
1495        if ( isHebrewCalendar ) {
1496           HebrewCalendar *hc = (HebrewCalendar*)&cal;
1497           if (hc->isLeapYear(hc->get(UCAL_YEAR,status)) && value == 6 && count >= 3 )
1498               value = 13; // Show alternate form for Adar II in leap years in Hebrew calendar.
1499           if (!hc->isLeapYear(hc->get(UCAL_YEAR,status)) && value >= 6 && count < 3 )
1500               value--; // Adjust the month number down 1 in Hebrew non-leap years, i.e. Adar is 6, not 7.
1501        }
1502        {
1503            int32_t isLeapMonth = (fSymbols->fLeapMonthPatterns != NULL && fSymbols->fLeapMonthPatternsCount >= DateFormatSymbols::kMonthPatternsCount)?
1504                        cal.get(UCAL_IS_LEAP_MONTH, status): 0;
1505            // should consolidate the next section by using arrays of pointers & counts for the right symbols...
1506            if (count == 5) {
1507                if (patternCharIndex == UDAT_MONTH_FIELD) {
1508                    _appendSymbolWithMonthPattern(appendTo, value, fSymbols->fNarrowMonths, fSymbols->fNarrowMonthsCount,
1509                            (isLeapMonth!=0)? &(fSymbols->fLeapMonthPatterns[DateFormatSymbols::kLeapMonthPatternFormatNarrow]): NULL, status);
1510                } else {
1511                    _appendSymbolWithMonthPattern(appendTo, value, fSymbols->fStandaloneNarrowMonths, fSymbols->fStandaloneNarrowMonthsCount,
1512                            (isLeapMonth!=0)? &(fSymbols->fLeapMonthPatterns[DateFormatSymbols::kLeapMonthPatternStandaloneNarrow]): NULL, status);
1513                }
1514                capContextUsageType = DateFormatSymbols::kCapContextUsageMonthNarrow;
1515            } else if (count == 4) {
1516                if (patternCharIndex == UDAT_MONTH_FIELD) {
1517                    _appendSymbolWithMonthPattern(appendTo, value, fSymbols->fMonths, fSymbols->fMonthsCount,
1518                            (isLeapMonth!=0)? &(fSymbols->fLeapMonthPatterns[DateFormatSymbols::kLeapMonthPatternFormatWide]): NULL, status);
1519                    capContextUsageType = DateFormatSymbols::kCapContextUsageMonthFormat;
1520                } else {
1521                    _appendSymbolWithMonthPattern(appendTo, value, fSymbols->fStandaloneMonths, fSymbols->fStandaloneMonthsCount,
1522                            (isLeapMonth!=0)? &(fSymbols->fLeapMonthPatterns[DateFormatSymbols::kLeapMonthPatternStandaloneWide]): NULL, status);
1523                    capContextUsageType = DateFormatSymbols::kCapContextUsageMonthStandalone;
1524                }
1525            } else if (count == 3) {
1526                if (patternCharIndex == UDAT_MONTH_FIELD) {
1527                    _appendSymbolWithMonthPattern(appendTo, value, fSymbols->fShortMonths, fSymbols->fShortMonthsCount,
1528                            (isLeapMonth!=0)? &(fSymbols->fLeapMonthPatterns[DateFormatSymbols::kLeapMonthPatternFormatAbbrev]): NULL, status);
1529                    capContextUsageType = DateFormatSymbols::kCapContextUsageMonthFormat;
1530                } else {
1531                    _appendSymbolWithMonthPattern(appendTo, value, fSymbols->fStandaloneShortMonths, fSymbols->fStandaloneShortMonthsCount,
1532                            (isLeapMonth!=0)? &(fSymbols->fLeapMonthPatterns[DateFormatSymbols::kLeapMonthPatternStandaloneAbbrev]): NULL, status);
1533                    capContextUsageType = DateFormatSymbols::kCapContextUsageMonthStandalone;
1534                }
1535            } else {
1536                UnicodeString monthNumber;
1537                zeroPaddingNumber(currentNumberFormat,monthNumber, value + 1, count, maxIntCount);
1538                _appendSymbolWithMonthPattern(appendTo, 0, &monthNumber, 1,
1539                        (isLeapMonth!=0)? &(fSymbols->fLeapMonthPatterns[DateFormatSymbols::kLeapMonthPatternNumeric]): NULL, status);
1540            }
1541        }
1542        break;
1543
1544    // for "k" and "kk", write out the hour, adjusting midnight to appear as "24"
1545    case UDAT_HOUR_OF_DAY1_FIELD:
1546        if (value == 0)
1547            zeroPaddingNumber(currentNumberFormat,appendTo, cal.getMaximum(UCAL_HOUR_OF_DAY) + 1, count, maxIntCount);
1548        else
1549            zeroPaddingNumber(currentNumberFormat,appendTo, value, count, maxIntCount);
1550        break;
1551
1552    case UDAT_FRACTIONAL_SECOND_FIELD:
1553        // Fractional seconds left-justify
1554        {
1555            currentNumberFormat->setMinimumIntegerDigits((count > 3) ? 3 : count);
1556            currentNumberFormat->setMaximumIntegerDigits(maxIntCount);
1557            if (count == 1) {
1558                value /= 100;
1559            } else if (count == 2) {
1560                value /= 10;
1561            }
1562            FieldPosition p(0);
1563            currentNumberFormat->format(value, appendTo, p);
1564            if (count > 3) {
1565                currentNumberFormat->setMinimumIntegerDigits(count - 3);
1566                currentNumberFormat->format((int32_t)0, appendTo, p);
1567            }
1568        }
1569        break;
1570
1571    // for "ee" or "e", use local numeric day-of-the-week
1572    // for "EEEEEE" or "eeeeee", write out the short day-of-the-week name
1573    // for "EEEEE" or "eeeee", write out the narrow day-of-the-week name
1574    // for "EEEE" or "eeee", write out the wide day-of-the-week name
1575    // for "EEE" or "EE" or "E" or "eee", write out the abbreviated day-of-the-week name
1576    case UDAT_DOW_LOCAL_FIELD:
1577        if ( count < 3 ) {
1578            zeroPaddingNumber(currentNumberFormat,appendTo, value, count, maxIntCount);
1579            break;
1580        }
1581        // fall through to EEEEE-EEE handling, but for that we don't want local day-of-week,
1582        // we want standard day-of-week, so first fix value to work for EEEEE-EEE.
1583        value = cal.get(UCAL_DAY_OF_WEEK, status);
1584        if (U_FAILURE(status)) {
1585            return;
1586        }
1587        // fall through, do not break here
1588    case UDAT_DAY_OF_WEEK_FIELD:
1589        if (count == 5) {
1590            _appendSymbol(appendTo, value, fSymbols->fNarrowWeekdays,
1591                          fSymbols->fNarrowWeekdaysCount);
1592            capContextUsageType = DateFormatSymbols::kCapContextUsageDayNarrow;
1593        } else if (count == 4) {
1594            _appendSymbol(appendTo, value, fSymbols->fWeekdays,
1595                          fSymbols->fWeekdaysCount);
1596            capContextUsageType = DateFormatSymbols::kCapContextUsageDayFormat;
1597        } else if (count == 6) {
1598            _appendSymbol(appendTo, value, fSymbols->fShorterWeekdays,
1599                          fSymbols->fShorterWeekdaysCount);
1600            capContextUsageType = DateFormatSymbols::kCapContextUsageDayFormat;
1601        } else {
1602            _appendSymbol(appendTo, value, fSymbols->fShortWeekdays,
1603                          fSymbols->fShortWeekdaysCount);
1604            capContextUsageType = DateFormatSymbols::kCapContextUsageDayFormat;
1605        }
1606        break;
1607
1608    // for "ccc", write out the abbreviated day-of-the-week name
1609    // for "cccc", write out the wide day-of-the-week name
1610    // for "ccccc", use the narrow day-of-the-week name
1611    // for "ccccc", use the short day-of-the-week name
1612    case UDAT_STANDALONE_DAY_FIELD:
1613        if ( count < 3 ) {
1614            zeroPaddingNumber(currentNumberFormat,appendTo, value, 1, maxIntCount);
1615            break;
1616        }
1617        // fall through to alpha DOW handling, but for that we don't want local day-of-week,
1618        // we want standard day-of-week, so first fix value.
1619        value = cal.get(UCAL_DAY_OF_WEEK, status);
1620        if (U_FAILURE(status)) {
1621            return;
1622        }
1623        if (count == 5) {
1624            _appendSymbol(appendTo, value, fSymbols->fStandaloneNarrowWeekdays,
1625                          fSymbols->fStandaloneNarrowWeekdaysCount);
1626            capContextUsageType = DateFormatSymbols::kCapContextUsageDayNarrow;
1627        } else if (count == 4) {
1628            _appendSymbol(appendTo, value, fSymbols->fStandaloneWeekdays,
1629                          fSymbols->fStandaloneWeekdaysCount);
1630            capContextUsageType = DateFormatSymbols::kCapContextUsageDayStandalone;
1631        } else if (count == 6) {
1632            _appendSymbol(appendTo, value, fSymbols->fStandaloneShorterWeekdays,
1633                          fSymbols->fStandaloneShorterWeekdaysCount);
1634            capContextUsageType = DateFormatSymbols::kCapContextUsageDayStandalone;
1635        } else { // count == 3
1636            _appendSymbol(appendTo, value, fSymbols->fStandaloneShortWeekdays,
1637                          fSymbols->fStandaloneShortWeekdaysCount);
1638            capContextUsageType = DateFormatSymbols::kCapContextUsageDayStandalone;
1639        }
1640        break;
1641
1642    // for "a" symbol, write out the whole AM/PM string
1643    case UDAT_AM_PM_FIELD:
1644        if (count < 5) {
1645            _appendSymbol(appendTo, value, fSymbols->fAmPms,
1646                          fSymbols->fAmPmsCount);
1647        } else {
1648            _appendSymbol(appendTo, value, fSymbols->fNarrowAmPms,
1649                          fSymbols->fNarrowAmPmsCount);
1650        }
1651        break;
1652
1653    // if we see pattern character for UDAT_TIME_SEPARATOR_FIELD (none currently defined),
1654    // write out the time separator string. Leave support in for future definition.
1655    case UDAT_TIME_SEPARATOR_FIELD:
1656        {
1657            UnicodeString separator;
1658            appendTo += fSymbols->getTimeSeparatorString(separator);
1659        }
1660        break;
1661
1662    // for "h" and "hh", write out the hour, adjusting noon and midnight to show up
1663    // as "12"
1664    case UDAT_HOUR1_FIELD:
1665        if (value == 0)
1666            zeroPaddingNumber(currentNumberFormat,appendTo, cal.getLeastMaximum(UCAL_HOUR) + 1, count, maxIntCount);
1667        else
1668            zeroPaddingNumber(currentNumberFormat,appendTo, value, count, maxIntCount);
1669        break;
1670
1671    case UDAT_TIMEZONE_FIELD: // 'z'
1672    case UDAT_TIMEZONE_RFC_FIELD: // 'Z'
1673    case UDAT_TIMEZONE_GENERIC_FIELD: // 'v'
1674    case UDAT_TIMEZONE_SPECIAL_FIELD: // 'V'
1675    case UDAT_TIMEZONE_LOCALIZED_GMT_OFFSET_FIELD: // 'O'
1676    case UDAT_TIMEZONE_ISO_FIELD: // 'X'
1677    case UDAT_TIMEZONE_ISO_LOCAL_FIELD: // 'x'
1678        {
1679            UChar zsbuf[64];
1680            UnicodeString zoneString(zsbuf, 0, UPRV_LENGTHOF(zsbuf));
1681            const TimeZone& tz = cal.getTimeZone();
1682            UDate date = cal.getTime(status);
1683            if (U_SUCCESS(status)) {
1684                if (patternCharIndex == UDAT_TIMEZONE_FIELD) {
1685                    if (count < 4) {
1686                        // "z", "zz", "zzz"
1687                        tzFormat()->format(UTZFMT_STYLE_SPECIFIC_SHORT, tz, date, zoneString);
1688                        capContextUsageType = DateFormatSymbols::kCapContextUsageMetazoneShort;
1689                    } else {
1690                        // "zzzz" or longer
1691                        tzFormat()->format(UTZFMT_STYLE_SPECIFIC_LONG, tz, date, zoneString);
1692                        capContextUsageType = DateFormatSymbols::kCapContextUsageMetazoneLong;
1693                    }
1694                }
1695                else if (patternCharIndex == UDAT_TIMEZONE_RFC_FIELD) {
1696                    if (count < 4) {
1697                        // "Z"
1698                        tzFormat()->format(UTZFMT_STYLE_ISO_BASIC_LOCAL_FULL, tz, date, zoneString);
1699                    } else if (count == 5) {
1700                        // "ZZZZZ"
1701                        tzFormat()->format(UTZFMT_STYLE_ISO_EXTENDED_FULL, tz, date, zoneString);
1702                    } else {
1703                        // "ZZ", "ZZZ", "ZZZZ"
1704                        tzFormat()->format(UTZFMT_STYLE_LOCALIZED_GMT, tz, date, zoneString);
1705                    }
1706                }
1707                else if (patternCharIndex == UDAT_TIMEZONE_GENERIC_FIELD) {
1708                    if (count == 1) {
1709                        // "v"
1710                        tzFormat()->format(UTZFMT_STYLE_GENERIC_SHORT, tz, date, zoneString);
1711                        capContextUsageType = DateFormatSymbols::kCapContextUsageMetazoneShort;
1712                    } else if (count == 4) {
1713                        // "vvvv"
1714                        tzFormat()->format(UTZFMT_STYLE_GENERIC_LONG, tz, date, zoneString);
1715                        capContextUsageType = DateFormatSymbols::kCapContextUsageMetazoneLong;
1716                    }
1717                }
1718                else if (patternCharIndex == UDAT_TIMEZONE_SPECIAL_FIELD) {
1719                    if (count == 1) {
1720                        // "V"
1721                        tzFormat()->format(UTZFMT_STYLE_ZONE_ID_SHORT, tz, date, zoneString);
1722                    } else if (count == 2) {
1723                        // "VV"
1724                        tzFormat()->format(UTZFMT_STYLE_ZONE_ID, tz, date, zoneString);
1725                    } else if (count == 3) {
1726                        // "VVV"
1727                        tzFormat()->format(UTZFMT_STYLE_EXEMPLAR_LOCATION, tz, date, zoneString);
1728                    } else if (count == 4) {
1729                        // "VVVV"
1730                        tzFormat()->format(UTZFMT_STYLE_GENERIC_LOCATION, tz, date, zoneString);
1731                        capContextUsageType = DateFormatSymbols::kCapContextUsageZoneLong;
1732                    }
1733                }
1734                else if (patternCharIndex == UDAT_TIMEZONE_LOCALIZED_GMT_OFFSET_FIELD) {
1735                    if (count == 1) {
1736                        // "O"
1737                        tzFormat()->format(UTZFMT_STYLE_LOCALIZED_GMT_SHORT, tz, date, zoneString);
1738                    } else if (count == 4) {
1739                        // "OOOO"
1740                        tzFormat()->format(UTZFMT_STYLE_LOCALIZED_GMT, tz, date, zoneString);
1741                    }
1742                }
1743                else if (patternCharIndex == UDAT_TIMEZONE_ISO_FIELD) {
1744                    if (count == 1) {
1745                        // "X"
1746                        tzFormat()->format(UTZFMT_STYLE_ISO_BASIC_SHORT, tz, date, zoneString);
1747                    } else if (count == 2) {
1748                        // "XX"
1749                        tzFormat()->format(UTZFMT_STYLE_ISO_BASIC_FIXED, tz, date, zoneString);
1750                    } else if (count == 3) {
1751                        // "XXX"
1752                        tzFormat()->format(UTZFMT_STYLE_ISO_EXTENDED_FIXED, tz, date, zoneString);
1753                    } else if (count == 4) {
1754                        // "XXXX"
1755                        tzFormat()->format(UTZFMT_STYLE_ISO_BASIC_FULL, tz, date, zoneString);
1756                    } else if (count == 5) {
1757                        // "XXXXX"
1758                        tzFormat()->format(UTZFMT_STYLE_ISO_EXTENDED_FULL, tz, date, zoneString);
1759                    }
1760                }
1761                else if (patternCharIndex == UDAT_TIMEZONE_ISO_LOCAL_FIELD) {
1762                    if (count == 1) {
1763                        // "x"
1764                        tzFormat()->format(UTZFMT_STYLE_ISO_BASIC_LOCAL_SHORT, tz, date, zoneString);
1765                    } else if (count == 2) {
1766                        // "xx"
1767                        tzFormat()->format(UTZFMT_STYLE_ISO_BASIC_LOCAL_FIXED, tz, date, zoneString);
1768                    } else if (count == 3) {
1769                        // "xxx"
1770                        tzFormat()->format(UTZFMT_STYLE_ISO_EXTENDED_LOCAL_FIXED, tz, date, zoneString);
1771                    } else if (count == 4) {
1772                        // "xxxx"
1773                        tzFormat()->format(UTZFMT_STYLE_ISO_BASIC_LOCAL_FULL, tz, date, zoneString);
1774                    } else if (count == 5) {
1775                        // "xxxxx"
1776                        tzFormat()->format(UTZFMT_STYLE_ISO_EXTENDED_LOCAL_FULL, tz, date, zoneString);
1777                    }
1778                }
1779                else {
1780                    U_ASSERT(FALSE);
1781                }
1782            }
1783            appendTo += zoneString;
1784        }
1785        break;
1786
1787    case UDAT_QUARTER_FIELD:
1788        if (count >= 4)
1789            _appendSymbol(appendTo, value/3, fSymbols->fQuarters,
1790                          fSymbols->fQuartersCount);
1791        else if (count == 3)
1792            _appendSymbol(appendTo, value/3, fSymbols->fShortQuarters,
1793                          fSymbols->fShortQuartersCount);
1794        else
1795            zeroPaddingNumber(currentNumberFormat,appendTo, (value/3) + 1, count, maxIntCount);
1796        break;
1797
1798    case UDAT_STANDALONE_QUARTER_FIELD:
1799        if (count >= 4)
1800            _appendSymbol(appendTo, value/3, fSymbols->fStandaloneQuarters,
1801                          fSymbols->fStandaloneQuartersCount);
1802        else if (count == 3)
1803            _appendSymbol(appendTo, value/3, fSymbols->fStandaloneShortQuarters,
1804                          fSymbols->fStandaloneShortQuartersCount);
1805        else
1806            zeroPaddingNumber(currentNumberFormat,appendTo, (value/3) + 1, count, maxIntCount);
1807        break;
1808
1809
1810    // all of the other pattern symbols can be formatted as simple numbers with
1811    // appropriate zero padding
1812    default:
1813        zeroPaddingNumber(currentNumberFormat,appendTo, value, count, maxIntCount);
1814        break;
1815    }
1816#if !UCONFIG_NO_BREAK_ITERATION
1817    // if first field, check to see whether we need to and are able to titlecase it
1818    if (fieldNum == 0 && u_islower(appendTo.char32At(beginOffset)) && fCapitalizationBrkIter != NULL) {
1819        UBool titlecase = FALSE;
1820        switch (capitalizationContext) {
1821            case UDISPCTX_CAPITALIZATION_FOR_BEGINNING_OF_SENTENCE:
1822                titlecase = TRUE;
1823                break;
1824            case UDISPCTX_CAPITALIZATION_FOR_UI_LIST_OR_MENU:
1825                titlecase = fSymbols->fCapitalization[capContextUsageType][0];
1826                break;
1827            case UDISPCTX_CAPITALIZATION_FOR_STANDALONE:
1828                titlecase = fSymbols->fCapitalization[capContextUsageType][1];
1829                break;
1830            default:
1831                // titlecase = FALSE;
1832                break;
1833        }
1834        if (titlecase) {
1835            UnicodeString firstField(appendTo, beginOffset);
1836            firstField.toTitle(fCapitalizationBrkIter, fLocale, U_TITLECASE_NO_LOWERCASE | U_TITLECASE_NO_BREAK_ADJUSTMENT);
1837            appendTo.replaceBetween(beginOffset, appendTo.length(), firstField);
1838        }
1839    }
1840#endif
1841
1842    handler.addAttribute(fgPatternIndexToDateFormatField[patternCharIndex], beginOffset, appendTo.length());
1843}
1844
1845//----------------------------------------------------------------------
1846
1847void SimpleDateFormat::adoptNumberFormat(NumberFormat *formatToAdopt) {
1848    fixNumberFormatForDates(*formatToAdopt);
1849    delete fNumberFormat;
1850    fNumberFormat = formatToAdopt;
1851
1852    // We successfully set the default number format. Now delete the overrides
1853    // (can't fail).
1854    if (fSharedNumberFormatters) {
1855        freeSharedNumberFormatters(fSharedNumberFormatters);
1856        fSharedNumberFormatters = NULL;
1857    }
1858}
1859
1860void SimpleDateFormat::adoptNumberFormat(const UnicodeString& fields, NumberFormat *formatToAdopt, UErrorCode &status){
1861    fixNumberFormatForDates(*formatToAdopt);
1862    LocalPointer<NumberFormat> fmt(formatToAdopt);
1863    if (U_FAILURE(status)) {
1864        return;
1865    }
1866
1867    // We must ensure fSharedNumberFormatters is allocated.
1868    if (fSharedNumberFormatters == NULL) {
1869        fSharedNumberFormatters = allocSharedNumberFormatters();
1870        if (fSharedNumberFormatters == NULL) {
1871            status = U_MEMORY_ALLOCATION_ERROR;
1872            return;
1873        }
1874    }
1875    const SharedNumberFormat *newFormat = createSharedNumberFormat(fmt.orphan());
1876    if (newFormat == NULL) {
1877        status = U_MEMORY_ALLOCATION_ERROR;
1878        return;
1879    }
1880    for (int i=0; i<fields.length(); i++) {
1881        UChar field = fields.charAt(i);
1882        // if the pattern character is unrecognized, signal an error and bail out
1883        UDateFormatField patternCharIndex = DateFormatSymbols::getPatternCharIndex(field);
1884        if (patternCharIndex == UDAT_FIELD_COUNT) {
1885            status = U_INVALID_FORMAT_ERROR;
1886            newFormat->deleteIfZeroRefCount();
1887            return;
1888        }
1889
1890        // Set the number formatter in the table
1891        SharedObject::copyPtr(
1892                newFormat, fSharedNumberFormatters[patternCharIndex]);
1893    }
1894    newFormat->deleteIfZeroRefCount();
1895}
1896
1897const NumberFormat *
1898SimpleDateFormat::getNumberFormatForField(UChar field) const {
1899    UDateFormatField index = DateFormatSymbols::getPatternCharIndex(field);
1900    if (index == UDAT_FIELD_COUNT) {
1901        return NULL;
1902    }
1903    return getNumberFormatByIndex(index);
1904}
1905
1906//----------------------------------------------------------------------
1907void
1908SimpleDateFormat::zeroPaddingNumber(
1909        NumberFormat *currentNumberFormat,
1910        UnicodeString &appendTo,
1911        int32_t value, int32_t minDigits, int32_t maxDigits) const
1912{
1913    if (currentNumberFormat!=NULL) {
1914        FieldPosition pos(0);
1915
1916        currentNumberFormat->setMinimumIntegerDigits(minDigits);
1917        currentNumberFormat->setMaximumIntegerDigits(maxDigits);
1918        currentNumberFormat->format(value, appendTo, pos);  // 3rd arg is there to speed up processing
1919    }
1920}
1921
1922//----------------------------------------------------------------------
1923
1924/**
1925 * Return true if the given format character, occuring count
1926 * times, represents a numeric field.
1927 */
1928UBool SimpleDateFormat::isNumeric(UChar formatChar, int32_t count) {
1929    return DateFormatSymbols::isNumericPatternChar(formatChar, count);
1930}
1931
1932UBool
1933SimpleDateFormat::isAtNumericField(const UnicodeString &pattern, int32_t patternOffset) {
1934    if (patternOffset >= pattern.length()) {
1935        // not at any field
1936        return FALSE;
1937    }
1938    UChar ch = pattern.charAt(patternOffset);
1939    UDateFormatField f = DateFormatSymbols::getPatternCharIndex(ch);
1940    if (f == UDAT_FIELD_COUNT) {
1941        // not at any field
1942        return FALSE;
1943    }
1944    int32_t i = patternOffset;
1945    while (pattern.charAt(++i) == ch) {}
1946    return DateFormatSymbols::isNumericField(f, i - patternOffset);
1947}
1948
1949UBool
1950SimpleDateFormat::isAfterNonNumericField(const UnicodeString &pattern, int32_t patternOffset) {
1951    if (patternOffset <= 0) {
1952        // not after any field
1953        return FALSE;
1954    }
1955    UChar ch = pattern.charAt(--patternOffset);
1956    UDateFormatField f = DateFormatSymbols::getPatternCharIndex(ch);
1957    if (f == UDAT_FIELD_COUNT) {
1958        // not after any field
1959        return FALSE;
1960    }
1961    int32_t i = patternOffset;
1962    while (pattern.charAt(--i) == ch) {}
1963    return !DateFormatSymbols::isNumericField(f, patternOffset - i);
1964}
1965
1966void
1967SimpleDateFormat::parse(const UnicodeString& text, Calendar& cal, ParsePosition& parsePos) const
1968{
1969    UErrorCode status = U_ZERO_ERROR;
1970    int32_t pos = parsePos.getIndex();
1971    if(parsePos.getIndex() < 0) {
1972        parsePos.setErrorIndex(0);
1973        return;
1974    }
1975    int32_t start = pos;
1976
1977
1978    UBool ambiguousYear[] = { FALSE };
1979    int32_t saveHebrewMonth = -1;
1980    int32_t count = 0;
1981    UTimeZoneFormatTimeType tzTimeType = UTZFMT_TIME_TYPE_UNKNOWN;
1982    SimpleDateFormatMutableNFs mutableNFs;
1983
1984    // For parsing abutting numeric fields. 'abutPat' is the
1985    // offset into 'pattern' of the first of 2 or more abutting
1986    // numeric fields.  'abutStart' is the offset into 'text'
1987    // where parsing the fields begins. 'abutPass' starts off as 0
1988    // and increments each time we try to parse the fields.
1989    int32_t abutPat = -1; // If >=0, we are in a run of abutting numeric fields
1990    int32_t abutStart = 0;
1991    int32_t abutPass = 0;
1992    UBool inQuote = FALSE;
1993
1994    MessageFormat * numericLeapMonthFormatter = NULL;
1995
1996    Calendar* calClone = NULL;
1997    Calendar *workCal = &cal;
1998    if (&cal != fCalendar && uprv_strcmp(cal.getType(), fCalendar->getType()) != 0) {
1999        // Different calendar type
2000        // We use the time/zone from the input calendar, but
2001        // do not use the input calendar for field calculation.
2002        calClone = fCalendar->clone();
2003        if (calClone != NULL) {
2004            calClone->setTime(cal.getTime(status),status);
2005            if (U_FAILURE(status)) {
2006                goto ExitParse;
2007            }
2008            calClone->setTimeZone(cal.getTimeZone());
2009            workCal = calClone;
2010        } else {
2011            status = U_MEMORY_ALLOCATION_ERROR;
2012            goto ExitParse;
2013        }
2014    }
2015
2016    if (fSymbols->fLeapMonthPatterns != NULL && fSymbols->fLeapMonthPatternsCount >= DateFormatSymbols::kMonthPatternsCount) {
2017        numericLeapMonthFormatter = new MessageFormat(fSymbols->fLeapMonthPatterns[DateFormatSymbols::kLeapMonthPatternNumeric], fLocale, status);
2018        if (numericLeapMonthFormatter == NULL) {
2019             status = U_MEMORY_ALLOCATION_ERROR;
2020             goto ExitParse;
2021        } else if (U_FAILURE(status)) {
2022             goto ExitParse; // this will delete numericLeapMonthFormatter
2023        }
2024    }
2025
2026    for (int32_t i=0; i<fPattern.length(); ++i) {
2027        UChar ch = fPattern.charAt(i);
2028
2029        // Handle alphabetic field characters.
2030        if (!inQuote && isSyntaxChar(ch)) {
2031            int32_t fieldPat = i;
2032
2033            // Count the length of this field specifier
2034            count = 1;
2035            while ((i+1)<fPattern.length() &&
2036                   fPattern.charAt(i+1) == ch) {
2037                ++count;
2038                ++i;
2039            }
2040
2041            if (isNumeric(ch, count)) {
2042                if (abutPat < 0) {
2043                    // Determine if there is an abutting numeric field.
2044                    // Record the start of a set of abutting numeric fields.
2045                    if (isAtNumericField(fPattern, i + 1)) {
2046                        abutPat = fieldPat;
2047                        abutStart = pos;
2048                        abutPass = 0;
2049                    }
2050                }
2051            } else {
2052                abutPat = -1; // End of any abutting fields
2053            }
2054
2055            // Handle fields within a run of abutting numeric fields.  Take
2056            // the pattern "HHmmss" as an example. We will try to parse
2057            // 2/2/2 characters of the input text, then if that fails,
2058            // 1/2/2.  We only adjust the width of the leftmost field; the
2059            // others remain fixed.  This allows "123456" => 12:34:56, but
2060            // "12345" => 1:23:45.  Likewise, for the pattern "yyyyMMdd" we
2061            // try 4/2/2, 3/2/2, 2/2/2, and finally 1/2/2.
2062            if (abutPat >= 0) {
2063                // If we are at the start of a run of abutting fields, then
2064                // shorten this field in each pass.  If we can't shorten
2065                // this field any more, then the parse of this set of
2066                // abutting numeric fields has failed.
2067                if (fieldPat == abutPat) {
2068                    count -= abutPass++;
2069                    if (count == 0) {
2070                        status = U_PARSE_ERROR;
2071                        goto ExitParse;
2072                    }
2073                }
2074
2075                pos = subParse(text, pos, ch, count,
2076                               TRUE, FALSE, ambiguousYear, saveHebrewMonth, *workCal, i, numericLeapMonthFormatter, &tzTimeType, mutableNFs);
2077
2078                // If the parse fails anywhere in the run, back up to the
2079                // start of the run and retry.
2080                if (pos < 0) {
2081                    i = abutPat - 1;
2082                    pos = abutStart;
2083                    continue;
2084                }
2085            }
2086
2087            // Handle non-numeric fields and non-abutting numeric
2088            // fields.
2089            else if (ch != 0x6C) { // pattern char 'l' (SMALL LETTER L) just gets ignored
2090                int32_t s = subParse(text, pos, ch, count,
2091                               FALSE, TRUE, ambiguousYear, saveHebrewMonth, *workCal, i, numericLeapMonthFormatter, &tzTimeType, mutableNFs);
2092
2093                if (s == -pos-1) {
2094                    // era not present, in special cases allow this to continue
2095                    // from the position where the era was expected
2096                    s = pos;
2097
2098                    if (i+1 < fPattern.length()) {
2099                        // move to next pattern character
2100                        UChar ch = fPattern.charAt(i+1);
2101
2102                        // check for whitespace
2103                        if (PatternProps::isWhiteSpace(ch)) {
2104                            i++;
2105                            // Advance over run in pattern
2106                            while ((i+1)<fPattern.length() &&
2107                                   PatternProps::isWhiteSpace(fPattern.charAt(i+1))) {
2108                                ++i;
2109                            }
2110                        }
2111                    }
2112                }
2113                else if (s <= 0) {
2114                    status = U_PARSE_ERROR;
2115                    goto ExitParse;
2116                }
2117                pos = s;
2118            }
2119        }
2120
2121        // Handle literal pattern characters.  These are any
2122        // quoted characters and non-alphabetic unquoted
2123        // characters.
2124        else {
2125
2126            abutPat = -1; // End of any abutting fields
2127
2128            if (! matchLiterals(fPattern, i, text, pos, getBooleanAttribute(UDAT_PARSE_ALLOW_WHITESPACE, status), getBooleanAttribute(UDAT_PARSE_PARTIAL_LITERAL_MATCH, status), isLenient())) {
2129                status = U_PARSE_ERROR;
2130                goto ExitParse;
2131            }
2132        }
2133    }
2134
2135    // Special hack for trailing "." after non-numeric field.
2136    if (text.charAt(pos) == 0x2e && getBooleanAttribute(UDAT_PARSE_ALLOW_WHITESPACE, status)) {
2137        // only do if the last field is not numeric
2138        if (isAfterNonNumericField(fPattern, fPattern.length())) {
2139            pos++; // skip the extra "."
2140        }
2141    }
2142
2143    // At this point the fields of Calendar have been set.  Calendar
2144    // will fill in default values for missing fields when the time
2145    // is computed.
2146
2147    parsePos.setIndex(pos);
2148
2149    // This part is a problem:  When we call parsedDate.after, we compute the time.
2150    // Take the date April 3 2004 at 2:30 am.  When this is first set up, the year
2151    // will be wrong if we're parsing a 2-digit year pattern.  It will be 1904.
2152    // April 3 1904 is a Sunday (unlike 2004) so it is the DST onset day.  2:30 am
2153    // is therefore an "impossible" time, since the time goes from 1:59 to 3:00 am
2154    // on that day.  It is therefore parsed out to fields as 3:30 am.  Then we
2155    // add 100 years, and get April 3 2004 at 3:30 am.  Note that April 3 2004 is
2156    // a Saturday, so it can have a 2:30 am -- and it should. [LIU]
2157    /*
2158        UDate parsedDate = calendar.getTime();
2159        if( ambiguousYear[0] && !parsedDate.after(fDefaultCenturyStart) ) {
2160            calendar.add(Calendar.YEAR, 100);
2161            parsedDate = calendar.getTime();
2162        }
2163    */
2164    // Because of the above condition, save off the fields in case we need to readjust.
2165    // The procedure we use here is not particularly efficient, but there is no other
2166    // way to do this given the API restrictions present in Calendar.  We minimize
2167    // inefficiency by only performing this computation when it might apply, that is,
2168    // when the two-digit year is equal to the start year, and thus might fall at the
2169    // front or the back of the default century.  This only works because we adjust
2170    // the year correctly to start with in other cases -- see subParse().
2171    if (ambiguousYear[0] || tzTimeType != UTZFMT_TIME_TYPE_UNKNOWN) // If this is true then the two-digit year == the default start year
2172    {
2173        // We need a copy of the fields, and we need to avoid triggering a call to
2174        // complete(), which will recalculate the fields.  Since we can't access
2175        // the fields[] array in Calendar, we clone the entire object.  This will
2176        // stop working if Calendar.clone() is ever rewritten to call complete().
2177        Calendar *copy;
2178        if (ambiguousYear[0]) {
2179            copy = cal.clone();
2180            // Check for failed cloning.
2181            if (copy == NULL) {
2182                status = U_MEMORY_ALLOCATION_ERROR;
2183                goto ExitParse;
2184            }
2185            UDate parsedDate = copy->getTime(status);
2186            // {sfb} check internalGetDefaultCenturyStart
2187            if (fHaveDefaultCentury && (parsedDate < fDefaultCenturyStart)) {
2188                // We can't use add here because that does a complete() first.
2189                cal.set(UCAL_YEAR, fDefaultCenturyStartYear + 100);
2190            }
2191            delete copy;
2192        }
2193
2194        if (tzTimeType != UTZFMT_TIME_TYPE_UNKNOWN) {
2195            copy = cal.clone();
2196            // Check for failed cloning.
2197            if (copy == NULL) {
2198                status = U_MEMORY_ALLOCATION_ERROR;
2199                goto ExitParse;
2200            }
2201            const TimeZone & tz = cal.getTimeZone();
2202            BasicTimeZone *btz = NULL;
2203
2204            if (dynamic_cast<const OlsonTimeZone *>(&tz) != NULL
2205                || dynamic_cast<const SimpleTimeZone *>(&tz) != NULL
2206                || dynamic_cast<const RuleBasedTimeZone *>(&tz) != NULL
2207                || dynamic_cast<const VTimeZone *>(&tz) != NULL) {
2208                btz = (BasicTimeZone*)&tz;
2209            }
2210
2211            // Get local millis
2212            copy->set(UCAL_ZONE_OFFSET, 0);
2213            copy->set(UCAL_DST_OFFSET, 0);
2214            UDate localMillis = copy->getTime(status);
2215
2216            // Make sure parsed time zone type (Standard or Daylight)
2217            // matches the rule used by the parsed time zone.
2218            int32_t raw, dst;
2219            if (btz != NULL) {
2220                if (tzTimeType == UTZFMT_TIME_TYPE_STANDARD) {
2221                    btz->getOffsetFromLocal(localMillis,
2222                        BasicTimeZone::kStandard, BasicTimeZone::kStandard, raw, dst, status);
2223                } else {
2224                    btz->getOffsetFromLocal(localMillis,
2225                        BasicTimeZone::kDaylight, BasicTimeZone::kDaylight, raw, dst, status);
2226                }
2227            } else {
2228                // No good way to resolve ambiguous time at transition,
2229                // but following code work in most case.
2230                tz.getOffset(localMillis, TRUE, raw, dst, status);
2231            }
2232
2233            // Now, compare the results with parsed type, either standard or daylight saving time
2234            int32_t resolvedSavings = dst;
2235            if (tzTimeType == UTZFMT_TIME_TYPE_STANDARD) {
2236                if (dst != 0) {
2237                    // Override DST_OFFSET = 0 in the result calendar
2238                    resolvedSavings = 0;
2239                }
2240            } else { // tztype == TZTYPE_DST
2241                if (dst == 0) {
2242                    if (btz != NULL) {
2243                        UDate time = localMillis + raw;
2244                        // We use the nearest daylight saving time rule.
2245                        TimeZoneTransition beforeTrs, afterTrs;
2246                        UDate beforeT = time, afterT = time;
2247                        int32_t beforeSav = 0, afterSav = 0;
2248                        UBool beforeTrsAvail, afterTrsAvail;
2249
2250                        // Search for DST rule before or on the time
2251                        while (TRUE) {
2252                            beforeTrsAvail = btz->getPreviousTransition(beforeT, TRUE, beforeTrs);
2253                            if (!beforeTrsAvail) {
2254                                break;
2255                            }
2256                            beforeT = beforeTrs.getTime() - 1;
2257                            beforeSav = beforeTrs.getFrom()->getDSTSavings();
2258                            if (beforeSav != 0) {
2259                                break;
2260                            }
2261                        }
2262
2263                        // Search for DST rule after the time
2264                        while (TRUE) {
2265                            afterTrsAvail = btz->getNextTransition(afterT, FALSE, afterTrs);
2266                            if (!afterTrsAvail) {
2267                                break;
2268                            }
2269                            afterT = afterTrs.getTime();
2270                            afterSav = afterTrs.getTo()->getDSTSavings();
2271                            if (afterSav != 0) {
2272                                break;
2273                            }
2274                        }
2275
2276                        if (beforeTrsAvail && afterTrsAvail) {
2277                            if (time - beforeT > afterT - time) {
2278                                resolvedSavings = afterSav;
2279                            } else {
2280                                resolvedSavings = beforeSav;
2281                            }
2282                        } else if (beforeTrsAvail && beforeSav != 0) {
2283                            resolvedSavings = beforeSav;
2284                        } else if (afterTrsAvail && afterSav != 0) {
2285                            resolvedSavings = afterSav;
2286                        } else {
2287                            resolvedSavings = btz->getDSTSavings();
2288                        }
2289                    } else {
2290                        resolvedSavings = tz.getDSTSavings();
2291                    }
2292                    if (resolvedSavings == 0) {
2293                        // final fallback
2294                        resolvedSavings = U_MILLIS_PER_HOUR;
2295                    }
2296                }
2297            }
2298            cal.set(UCAL_ZONE_OFFSET, raw);
2299            cal.set(UCAL_DST_OFFSET, resolvedSavings);
2300            delete copy;
2301        }
2302    }
2303ExitParse:
2304    // Set the parsed result if local calendar is used
2305    // instead of the input calendar
2306    if (U_SUCCESS(status) && workCal != &cal) {
2307        cal.setTimeZone(workCal->getTimeZone());
2308        cal.setTime(workCal->getTime(status), status);
2309    }
2310
2311    if (numericLeapMonthFormatter != NULL) {
2312        delete numericLeapMonthFormatter;
2313    }
2314    if (calClone != NULL) {
2315        delete calClone;
2316    }
2317
2318    // If any Calendar calls failed, we pretend that we
2319    // couldn't parse the string, when in reality this isn't quite accurate--
2320    // we did parse it; the Calendar calls just failed.
2321    if (U_FAILURE(status)) {
2322        parsePos.setErrorIndex(pos);
2323        parsePos.setIndex(start);
2324    }
2325}
2326
2327//----------------------------------------------------------------------
2328
2329static int32_t
2330matchStringWithOptionalDot(const UnicodeString &text,
2331                            int32_t index,
2332                            const UnicodeString &data);
2333
2334int32_t SimpleDateFormat::matchQuarterString(const UnicodeString& text,
2335                              int32_t start,
2336                              UCalendarDateFields field,
2337                              const UnicodeString* data,
2338                              int32_t dataCount,
2339                              Calendar& cal) const
2340{
2341    int32_t i = 0;
2342    int32_t count = dataCount;
2343
2344    // There may be multiple strings in the data[] array which begin with
2345    // the same prefix (e.g., Cerven and Cervenec (June and July) in Czech).
2346    // We keep track of the longest match, and return that.  Note that this
2347    // unfortunately requires us to test all array elements.
2348    int32_t bestMatchLength = 0, bestMatch = -1;
2349    UnicodeString bestMatchName;
2350
2351    for (; i < count; ++i) {
2352        int32_t matchLength = 0;
2353        if ((matchLength = matchStringWithOptionalDot(text, start, data[i])) > bestMatchLength) {
2354            bestMatchLength = matchLength;
2355            bestMatch = i;
2356        }
2357    }
2358
2359    if (bestMatch >= 0) {
2360        cal.set(field, bestMatch * 3);
2361        return start + bestMatchLength;
2362    }
2363
2364    return -start;
2365}
2366
2367//----------------------------------------------------------------------
2368UBool SimpleDateFormat::matchLiterals(const UnicodeString &pattern,
2369                                      int32_t &patternOffset,
2370                                      const UnicodeString &text,
2371                                      int32_t &textOffset,
2372                                      UBool whitespaceLenient,
2373                                      UBool partialMatchLenient,
2374                                      UBool oldLeniency)
2375{
2376    UBool inQuote = FALSE;
2377    UnicodeString literal;
2378    int32_t i = patternOffset;
2379
2380    // scan pattern looking for contiguous literal characters
2381    for ( ; i < pattern.length(); i += 1) {
2382        UChar ch = pattern.charAt(i);
2383
2384        if (!inQuote && isSyntaxChar(ch)) {
2385            break;
2386        }
2387
2388        if (ch == QUOTE) {
2389            // Match a quote literal ('') inside OR outside of quotes
2390            if ((i + 1) < pattern.length() && pattern.charAt(i + 1) == QUOTE) {
2391                i += 1;
2392            } else {
2393                inQuote = !inQuote;
2394                continue;
2395            }
2396        }
2397
2398        literal += ch;
2399    }
2400
2401    // at this point, literal contains the literal text
2402    // and i is the index of the next non-literal pattern character.
2403    int32_t p;
2404    int32_t t = textOffset;
2405
2406    if (whitespaceLenient) {
2407        // trim leading, trailing whitespace from
2408        // the literal text
2409        literal.trim();
2410
2411        // ignore any leading whitespace in the text
2412        while (t < text.length() && u_isWhitespace(text.charAt(t))) {
2413            t += 1;
2414        }
2415    }
2416
2417    for (p = 0; p < literal.length() && t < text.length();) {
2418        UBool needWhitespace = FALSE;
2419
2420        while (p < literal.length() && PatternProps::isWhiteSpace(literal.charAt(p))) {
2421            needWhitespace = TRUE;
2422            p += 1;
2423        }
2424
2425        if (needWhitespace) {
2426            int32_t tStart = t;
2427
2428            while (t < text.length()) {
2429                UChar tch = text.charAt(t);
2430
2431                if (!u_isUWhiteSpace(tch) && !PatternProps::isWhiteSpace(tch)) {
2432                    break;
2433                }
2434
2435                t += 1;
2436            }
2437
2438            // TODO: should we require internal spaces
2439            // in lenient mode? (There won't be any
2440            // leading or trailing spaces)
2441            if (!whitespaceLenient && t == tStart) {
2442                // didn't find matching whitespace:
2443                // an error in strict mode
2444                return FALSE;
2445            }
2446
2447            // In strict mode, this run of whitespace
2448            // may have been at the end.
2449            if (p >= literal.length()) {
2450                break;
2451            }
2452        }
2453        if (t >= text.length() || literal.charAt(p) != text.charAt(t)) {
2454            // Ran out of text, or found a non-matching character:
2455            // OK in lenient mode, an error in strict mode.
2456            if (whitespaceLenient) {
2457                if (t == textOffset && text.charAt(t) == 0x2e &&
2458                        isAfterNonNumericField(pattern, patternOffset)) {
2459                    // Lenient mode and the literal input text begins with a "." and
2460                    // we are after a non-numeric field: We skip the "."
2461                    ++t;
2462                    continue;  // Do not update p.
2463                }
2464                // if it is actual whitespace and we're whitespace lenient it's OK
2465
2466                UChar wsc = text.charAt(t);
2467                if(PatternProps::isWhiteSpace(wsc)) {
2468                    // Lenient mode and it's just whitespace we skip it
2469                    ++t;
2470                    continue;  // Do not update p.
2471                }
2472            }
2473            // hack around oldleniency being a bit of a catch-all bucket and we're just adding support specifically for paritial matches
2474            if(partialMatchLenient && oldLeniency) {
2475                break;
2476            }
2477
2478            return FALSE;
2479        }
2480        ++p;
2481        ++t;
2482    }
2483
2484    // At this point if we're in strict mode we have a complete match.
2485    // If we're in lenient mode we may have a partial match, or no
2486    // match at all.
2487    if (p <= 0) {
2488        // no match. Pretend it matched a run of whitespace
2489        // and ignorables in the text.
2490        const  UnicodeSet *ignorables = NULL;
2491        UDateFormatField patternCharIndex = DateFormatSymbols::getPatternCharIndex(pattern.charAt(i));
2492        if (patternCharIndex != UDAT_FIELD_COUNT) {
2493            ignorables = SimpleDateFormatStaticSets::getIgnorables(patternCharIndex);
2494        }
2495
2496        for (t = textOffset; t < text.length(); t += 1) {
2497            UChar ch = text.charAt(t);
2498
2499            if (ignorables == NULL || !ignorables->contains(ch)) {
2500                break;
2501            }
2502        }
2503    }
2504
2505    // if we get here, we've got a complete match.
2506    patternOffset = i - 1;
2507    textOffset = t;
2508
2509    return TRUE;
2510}
2511
2512//----------------------------------------------------------------------
2513
2514int32_t SimpleDateFormat::matchString(const UnicodeString& text,
2515                              int32_t start,
2516                              UCalendarDateFields field,
2517                              const UnicodeString* data,
2518                              int32_t dataCount,
2519                              const UnicodeString* monthPattern,
2520                              Calendar& cal) const
2521{
2522    int32_t i = 0;
2523    int32_t count = dataCount;
2524
2525    if (field == UCAL_DAY_OF_WEEK) i = 1;
2526
2527    // There may be multiple strings in the data[] array which begin with
2528    // the same prefix (e.g., Cerven and Cervenec (June and July) in Czech).
2529    // We keep track of the longest match, and return that.  Note that this
2530    // unfortunately requires us to test all array elements.
2531    int32_t bestMatchLength = 0, bestMatch = -1;
2532    UnicodeString bestMatchName;
2533    int32_t isLeapMonth = 0;
2534
2535    for (; i < count; ++i) {
2536        int32_t matchLen = 0;
2537        if ((matchLen = matchStringWithOptionalDot(text, start, data[i])) > bestMatchLength) {
2538            bestMatch = i;
2539            bestMatchLength = matchLen;
2540        }
2541
2542        if (monthPattern != NULL) {
2543            UErrorCode status = U_ZERO_ERROR;
2544            UnicodeString leapMonthName;
2545            Formattable monthName((const UnicodeString&)(data[i]));
2546            MessageFormat::format(*monthPattern, &monthName, 1, leapMonthName, status);
2547            if (U_SUCCESS(status)) {
2548                if ((matchLen = matchStringWithOptionalDot(text, start, leapMonthName)) > bestMatchLength) {
2549                    bestMatch = i;
2550                    bestMatchLength = matchLen;
2551                    isLeapMonth = 1;
2552                }
2553            }
2554        }
2555    }
2556
2557    if (bestMatch >= 0) {
2558        if (field < UCAL_FIELD_COUNT) {
2559            // Adjustment for Hebrew Calendar month Adar II
2560            if (!strcmp(cal.getType(),"hebrew") && field==UCAL_MONTH && bestMatch==13) {
2561                cal.set(field,6);
2562            } else {
2563                if (field == UCAL_YEAR) {
2564                    bestMatch++; // only get here for cyclic year names, which match 1-based years 1-60
2565                }
2566                cal.set(field, bestMatch);
2567            }
2568            if (monthPattern != NULL) {
2569                cal.set(UCAL_IS_LEAP_MONTH, isLeapMonth);
2570            }
2571        }
2572
2573        return start + bestMatchLength;
2574    }
2575
2576    return -start;
2577}
2578
2579static int32_t
2580matchStringWithOptionalDot(const UnicodeString &text,
2581                            int32_t index,
2582                            const UnicodeString &data) {
2583    UErrorCode sts = U_ZERO_ERROR;
2584    int32_t matchLenText = 0;
2585    int32_t matchLenData = 0;
2586
2587    u_caseInsensitivePrefixMatch(text.getBuffer() + index, text.length() - index,
2588                                 data.getBuffer(), data.length(),
2589                                 0 /* default case option */,
2590                                 &matchLenText, &matchLenData,
2591                                 &sts);
2592    U_ASSERT (U_SUCCESS(sts));
2593
2594    if (matchLenData == data.length() /* normal match */
2595        || (data.charAt(data.length() - 1) == 0x2e
2596            && matchLenData == data.length() - 1 /* match without trailing dot */)) {
2597        return matchLenText;
2598    }
2599
2600    return 0;
2601}
2602
2603//----------------------------------------------------------------------
2604
2605void
2606SimpleDateFormat::set2DigitYearStart(UDate d, UErrorCode& status)
2607{
2608    parseAmbiguousDatesAsAfter(d, status);
2609}
2610
2611/**
2612 * Private member function that converts the parsed date strings into
2613 * timeFields. Returns -start (for ParsePosition) if failed.
2614 */
2615int32_t SimpleDateFormat::subParse(const UnicodeString& text, int32_t& start, UChar ch, int32_t count,
2616                           UBool obeyCount, UBool allowNegative, UBool ambiguousYear[], int32_t& saveHebrewMonth, Calendar& cal,
2617                           int32_t patLoc, MessageFormat * numericLeapMonthFormatter, UTimeZoneFormatTimeType *tzTimeType, SimpleDateFormatMutableNFs &mutableNFs) const
2618{
2619    Formattable number;
2620    int32_t value = 0;
2621    int32_t i;
2622    int32_t ps = 0;
2623    UErrorCode status = U_ZERO_ERROR;
2624    ParsePosition pos(0);
2625    UDateFormatField patternCharIndex = DateFormatSymbols::getPatternCharIndex(ch);
2626    NumberFormat *currentNumberFormat;
2627    UnicodeString temp;
2628    UBool gotNumber = FALSE;
2629
2630#if defined (U_DEBUG_CAL)
2631    //fprintf(stderr, "%s:%d - [%c]  st=%d \n", __FILE__, __LINE__, (char) ch, start);
2632#endif
2633
2634    if (patternCharIndex == UDAT_FIELD_COUNT) {
2635        return -start;
2636    }
2637
2638    currentNumberFormat = mutableNFs.get(getNumberFormatByIndex(patternCharIndex));
2639    if (currentNumberFormat == NULL) {
2640        return -start;
2641    }
2642    UCalendarDateFields field = fgPatternIndexToCalendarField[patternCharIndex]; // UCAL_FIELD_COUNT if irrelevant
2643    UnicodeString hebr("hebr", 4, US_INV);
2644
2645    if (numericLeapMonthFormatter != NULL) {
2646        numericLeapMonthFormatter->setFormats((const Format **)&currentNumberFormat, 1);
2647    }
2648    UBool isChineseCalendar = (uprv_strcmp(cal.getType(),"chinese") == 0 || uprv_strcmp(cal.getType(),"dangi") == 0);
2649
2650    // If there are any spaces here, skip over them.  If we hit the end
2651    // of the string, then fail.
2652    for (;;) {
2653        if (start >= text.length()) {
2654            return -start;
2655        }
2656        UChar32 c = text.char32At(start);
2657        if (!u_isUWhiteSpace(c) /*||*/ && !PatternProps::isWhiteSpace(c)) {
2658            break;
2659        }
2660        start += U16_LENGTH(c);
2661    }
2662    pos.setIndex(start);
2663
2664    // We handle a few special cases here where we need to parse
2665    // a number value.  We handle further, more generic cases below.  We need
2666    // to handle some of them here because some fields require extra processing on
2667    // the parsed value.
2668    if (patternCharIndex == UDAT_HOUR_OF_DAY1_FIELD ||                       // k
2669        patternCharIndex == UDAT_HOUR_OF_DAY0_FIELD ||                       // H
2670        patternCharIndex == UDAT_HOUR1_FIELD ||                              // h
2671        patternCharIndex == UDAT_HOUR0_FIELD ||                              // K
2672        (patternCharIndex == UDAT_DOW_LOCAL_FIELD && count <= 2) ||          // e
2673        (patternCharIndex == UDAT_STANDALONE_DAY_FIELD && count <= 2) ||     // c
2674        (patternCharIndex == UDAT_MONTH_FIELD && count <= 2) ||              // M
2675        (patternCharIndex == UDAT_STANDALONE_MONTH_FIELD && count <= 2) ||   // L
2676        (patternCharIndex == UDAT_QUARTER_FIELD && count <= 2) ||            // Q
2677        (patternCharIndex == UDAT_STANDALONE_QUARTER_FIELD && count <= 2) || // q
2678        patternCharIndex == UDAT_YEAR_FIELD ||                               // y
2679        patternCharIndex == UDAT_YEAR_WOY_FIELD ||                           // Y
2680        patternCharIndex == UDAT_YEAR_NAME_FIELD ||                          // U (falls back to numeric)
2681        (patternCharIndex == UDAT_ERA_FIELD && isChineseCalendar) ||         // G
2682        patternCharIndex == UDAT_FRACTIONAL_SECOND_FIELD)                    // S
2683    {
2684        int32_t parseStart = pos.getIndex();
2685        // It would be good to unify this with the obeyCount logic below,
2686        // but that's going to be difficult.
2687        const UnicodeString* src;
2688
2689        UBool parsedNumericLeapMonth = FALSE;
2690        if (numericLeapMonthFormatter != NULL && (patternCharIndex == UDAT_MONTH_FIELD || patternCharIndex == UDAT_STANDALONE_MONTH_FIELD)) {
2691            int32_t argCount;
2692            Formattable * args = numericLeapMonthFormatter->parse(text, pos, argCount);
2693            if (args != NULL && argCount == 1 && pos.getIndex() > parseStart && args[0].isNumeric()) {
2694                parsedNumericLeapMonth = TRUE;
2695                number.setLong(args[0].getLong());
2696                cal.set(UCAL_IS_LEAP_MONTH, 1);
2697                delete[] args;
2698            } else {
2699                pos.setIndex(parseStart);
2700                cal.set(UCAL_IS_LEAP_MONTH, 0);
2701            }
2702        }
2703
2704        if (!parsedNumericLeapMonth) {
2705            if (obeyCount) {
2706                if ((start+count) > text.length()) {
2707                    return -start;
2708                }
2709
2710                text.extractBetween(0, start + count, temp);
2711                src = &temp;
2712            } else {
2713                src = &text;
2714            }
2715
2716            parseInt(*src, number, pos, allowNegative,currentNumberFormat);
2717        }
2718
2719        int32_t txtLoc = pos.getIndex();
2720
2721        if (txtLoc > parseStart) {
2722            value = number.getLong();
2723            gotNumber = TRUE;
2724
2725            // suffix processing
2726            if (value < 0 ) {
2727                txtLoc = checkIntSuffix(text, txtLoc, patLoc+1, TRUE);
2728                if (txtLoc != pos.getIndex()) {
2729                    value *= -1;
2730                }
2731            }
2732            else {
2733                txtLoc = checkIntSuffix(text, txtLoc, patLoc+1, FALSE);
2734            }
2735
2736            if (!getBooleanAttribute(UDAT_PARSE_ALLOW_WHITESPACE, status)) {
2737                // Check the range of the value
2738                int32_t bias = gFieldRangeBias[patternCharIndex];
2739                if (bias >= 0 && (value > cal.getMaximum(field) + bias || value < cal.getMinimum(field) + bias)) {
2740                    return -start;
2741                }
2742            }
2743
2744            pos.setIndex(txtLoc);
2745        }
2746    }
2747
2748    // Make sure that we got a number if
2749    // we want one, and didn't get one
2750    // if we don't want one.
2751    switch (patternCharIndex) {
2752        case UDAT_HOUR_OF_DAY1_FIELD:
2753        case UDAT_HOUR_OF_DAY0_FIELD:
2754        case UDAT_HOUR1_FIELD:
2755        case UDAT_HOUR0_FIELD:
2756            // special range check for hours:
2757            if (value < 0 || value > 24) {
2758                return -start;
2759            }
2760
2761            // fall through to gotNumber check
2762
2763        case UDAT_YEAR_FIELD:
2764        case UDAT_YEAR_WOY_FIELD:
2765        case UDAT_FRACTIONAL_SECOND_FIELD:
2766            // these must be a number
2767            if (! gotNumber) {
2768                return -start;
2769            }
2770
2771            break;
2772
2773        default:
2774            // we check the rest of the fields below.
2775            break;
2776    }
2777
2778    switch (patternCharIndex) {
2779    case UDAT_ERA_FIELD:
2780        if (isChineseCalendar) {
2781            if (!gotNumber) {
2782                return -start;
2783            }
2784            cal.set(UCAL_ERA, value);
2785            return pos.getIndex();
2786        }
2787        if (count == 5) {
2788            ps = matchString(text, start, UCAL_ERA, fSymbols->fNarrowEras, fSymbols->fNarrowErasCount, NULL, cal);
2789        } else if (count == 4) {
2790            ps = matchString(text, start, UCAL_ERA, fSymbols->fEraNames, fSymbols->fEraNamesCount, NULL, cal);
2791        } else {
2792            ps = matchString(text, start, UCAL_ERA, fSymbols->fEras, fSymbols->fErasCount, NULL, cal);
2793        }
2794
2795        // check return position, if it equals -start, then matchString error
2796        // special case the return code so we don't necessarily fail out until we
2797        // verify no year information also
2798        if (ps == -start)
2799            ps--;
2800
2801        return ps;
2802
2803    case UDAT_YEAR_FIELD:
2804        // If there are 3 or more YEAR pattern characters, this indicates
2805        // that the year value is to be treated literally, without any
2806        // two-digit year adjustments (e.g., from "01" to 2001).  Otherwise
2807        // we made adjustments to place the 2-digit year in the proper
2808        // century, for parsed strings from "00" to "99".  Any other string
2809        // is treated literally:  "2250", "-1", "1", "002".
2810        if (fDateOverride.compare(hebr)==0 && value < 1000) {
2811            value += HEBREW_CAL_CUR_MILLENIUM_START_YEAR;
2812        } else if ((pos.getIndex() - start) == 2 && !isChineseCalendar
2813            && u_isdigit(text.charAt(start))
2814            && u_isdigit(text.charAt(start+1)))
2815        {
2816            // only adjust year for patterns less than 3.
2817            if(count < 3) {
2818                // Assume for example that the defaultCenturyStart is 6/18/1903.
2819                // This means that two-digit years will be forced into the range
2820                // 6/18/1903 to 6/17/2003.  As a result, years 00, 01, and 02
2821                // correspond to 2000, 2001, and 2002.  Years 04, 05, etc. correspond
2822                // to 1904, 1905, etc.  If the year is 03, then it is 2003 if the
2823                // other fields specify a date before 6/18, or 1903 if they specify a
2824                // date afterwards.  As a result, 03 is an ambiguous year.  All other
2825                // two-digit years are unambiguous.
2826                if(fHaveDefaultCentury) { // check if this formatter even has a pivot year
2827                    int32_t ambiguousTwoDigitYear = fDefaultCenturyStartYear % 100;
2828                    ambiguousYear[0] = (value == ambiguousTwoDigitYear);
2829                    value += (fDefaultCenturyStartYear/100)*100 +
2830                            (value < ambiguousTwoDigitYear ? 100 : 0);
2831                }
2832            }
2833        }
2834        cal.set(UCAL_YEAR, value);
2835
2836        // Delayed checking for adjustment of Hebrew month numbers in non-leap years.
2837        if (saveHebrewMonth >= 0) {
2838            HebrewCalendar *hc = (HebrewCalendar*)&cal;
2839            if (!hc->isLeapYear(value) && saveHebrewMonth >= 6) {
2840               cal.set(UCAL_MONTH,saveHebrewMonth);
2841            } else {
2842               cal.set(UCAL_MONTH,saveHebrewMonth-1);
2843            }
2844            saveHebrewMonth = -1;
2845        }
2846        return pos.getIndex();
2847
2848    case UDAT_YEAR_WOY_FIELD:
2849        // Comment is the same as for UDAT_Year_FIELDs - look above
2850        if (fDateOverride.compare(hebr)==0 && value < 1000) {
2851            value += HEBREW_CAL_CUR_MILLENIUM_START_YEAR;
2852        } else if ((pos.getIndex() - start) == 2
2853            && u_isdigit(text.charAt(start))
2854            && u_isdigit(text.charAt(start+1))
2855            && fHaveDefaultCentury )
2856        {
2857            int32_t ambiguousTwoDigitYear = fDefaultCenturyStartYear % 100;
2858            ambiguousYear[0] = (value == ambiguousTwoDigitYear);
2859            value += (fDefaultCenturyStartYear/100)*100 +
2860                (value < ambiguousTwoDigitYear ? 100 : 0);
2861        }
2862        cal.set(UCAL_YEAR_WOY, value);
2863        return pos.getIndex();
2864
2865    case UDAT_YEAR_NAME_FIELD:
2866        if (fSymbols->fShortYearNames != NULL) {
2867            int32_t newStart = matchString(text, start, UCAL_YEAR, fSymbols->fShortYearNames, fSymbols->fShortYearNamesCount, NULL, cal);
2868            if (newStart > 0) {
2869                return newStart;
2870            }
2871        }
2872        if (gotNumber && (getBooleanAttribute(UDAT_PARSE_ALLOW_NUMERIC,status) || value > fSymbols->fShortYearNamesCount)) {
2873            cal.set(UCAL_YEAR, value);
2874            return pos.getIndex();
2875        }
2876        return -start;
2877
2878    case UDAT_MONTH_FIELD:
2879    case UDAT_STANDALONE_MONTH_FIELD:
2880        if (gotNumber) // i.e., M or MM.
2881        {
2882            // When parsing month numbers from the Hebrew Calendar, we might need to adjust the month depending on whether
2883            // 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
2884            // the year is parsed.
2885            if (!strcmp(cal.getType(),"hebrew")) {
2886                HebrewCalendar *hc = (HebrewCalendar*)&cal;
2887                if (cal.isSet(UCAL_YEAR)) {
2888                   UErrorCode status = U_ZERO_ERROR;
2889                   if (!hc->isLeapYear(hc->get(UCAL_YEAR,status)) && value >= 6) {
2890                       cal.set(UCAL_MONTH, value);
2891                   } else {
2892                       cal.set(UCAL_MONTH, value - 1);
2893                   }
2894                } else {
2895                    saveHebrewMonth = value;
2896                }
2897            } else {
2898                // Don't want to parse the month if it is a string
2899                // while pattern uses numeric style: M/MM, L/LL
2900                // [We computed 'value' above.]
2901                cal.set(UCAL_MONTH, value - 1);
2902            }
2903            return pos.getIndex();
2904        } else {
2905            // count >= 3 // i.e., MMM/MMMM, LLL/LLLL
2906            // Want to be able to parse both short and long forms.
2907            // Try count == 4 first:
2908            UnicodeString * wideMonthPat = NULL;
2909            UnicodeString * shortMonthPat = NULL;
2910            if (fSymbols->fLeapMonthPatterns != NULL && fSymbols->fLeapMonthPatternsCount >= DateFormatSymbols::kMonthPatternsCount) {
2911                if (patternCharIndex==UDAT_MONTH_FIELD) {
2912                    wideMonthPat = &fSymbols->fLeapMonthPatterns[DateFormatSymbols::kLeapMonthPatternFormatWide];
2913                    shortMonthPat = &fSymbols->fLeapMonthPatterns[DateFormatSymbols::kLeapMonthPatternFormatAbbrev];
2914                } else {
2915                    wideMonthPat = &fSymbols->fLeapMonthPatterns[DateFormatSymbols::kLeapMonthPatternStandaloneWide];
2916                    shortMonthPat = &fSymbols->fLeapMonthPatterns[DateFormatSymbols::kLeapMonthPatternStandaloneAbbrev];
2917                }
2918            }
2919            int32_t newStart = 0;
2920            if (patternCharIndex==UDAT_MONTH_FIELD) {
2921                if(getBooleanAttribute(UDAT_PARSE_MULTIPLE_PATTERNS_FOR_MATCH, status) || count == 4) {
2922                    newStart = matchString(text, start, UCAL_MONTH, fSymbols->fMonths, fSymbols->fMonthsCount, wideMonthPat, cal); // try MMMM
2923                    if (newStart > 0) {
2924                        return newStart;
2925                    }
2926                }
2927                if(getBooleanAttribute(UDAT_PARSE_MULTIPLE_PATTERNS_FOR_MATCH, status) || count == 3) {
2928                    newStart = matchString(text, start, UCAL_MONTH, fSymbols->fShortMonths, fSymbols->fShortMonthsCount, shortMonthPat, cal); // try MMM
2929                }
2930            } else {
2931                if(getBooleanAttribute(UDAT_PARSE_MULTIPLE_PATTERNS_FOR_MATCH, status) || count == 4) {
2932                    newStart = matchString(text, start, UCAL_MONTH, fSymbols->fStandaloneMonths, fSymbols->fStandaloneMonthsCount, wideMonthPat, cal); // try LLLL
2933                    if (newStart > 0) {
2934                        return newStart;
2935                    }
2936                }
2937                if(getBooleanAttribute(UDAT_PARSE_MULTIPLE_PATTERNS_FOR_MATCH, status) || count == 3) {
2938                    newStart = matchString(text, start, UCAL_MONTH, fSymbols->fStandaloneShortMonths, fSymbols->fStandaloneShortMonthsCount, shortMonthPat, cal); // try LLL
2939                }
2940            }
2941            if (newStart > 0 || !getBooleanAttribute(UDAT_PARSE_ALLOW_NUMERIC, status))  // currently we do not try to parse MMMMM/LLLLL: #8860
2942                return newStart;
2943            // else we allowing parsing as number, below
2944        }
2945        break;
2946
2947    case UDAT_HOUR_OF_DAY1_FIELD:
2948        // [We computed 'value' above.]
2949        if (value == cal.getMaximum(UCAL_HOUR_OF_DAY) + 1)
2950            value = 0;
2951
2952        // fall through to set field
2953
2954    case UDAT_HOUR_OF_DAY0_FIELD:
2955        cal.set(UCAL_HOUR_OF_DAY, value);
2956        return pos.getIndex();
2957
2958    case UDAT_FRACTIONAL_SECOND_FIELD:
2959        // Fractional seconds left-justify
2960        i = pos.getIndex() - start;
2961        if (i < 3) {
2962            while (i < 3) {
2963                value *= 10;
2964                i++;
2965            }
2966        } else {
2967            int32_t a = 1;
2968            while (i > 3) {
2969                a *= 10;
2970                i--;
2971            }
2972            value /= a;
2973        }
2974        cal.set(UCAL_MILLISECOND, value);
2975        return pos.getIndex();
2976
2977    case UDAT_DOW_LOCAL_FIELD:
2978        if (gotNumber) // i.e., e or ee
2979        {
2980            // [We computed 'value' above.]
2981            cal.set(UCAL_DOW_LOCAL, value);
2982            return pos.getIndex();
2983        }
2984        // else for eee-eeeee fall through to handling of EEE-EEEEE
2985        // fall through, do not break here
2986    case UDAT_DAY_OF_WEEK_FIELD:
2987        {
2988            // Want to be able to parse both short and long forms.
2989            // Try count == 4 (EEEE) wide first:
2990            int32_t newStart = 0;
2991            if(getBooleanAttribute(UDAT_PARSE_MULTIPLE_PATTERNS_FOR_MATCH, status) || count == 4) {
2992                if ((newStart = matchString(text, start, UCAL_DAY_OF_WEEK,
2993                                          fSymbols->fWeekdays, fSymbols->fWeekdaysCount, NULL, cal)) > 0)
2994                    return newStart;
2995            }
2996            // EEEE wide failed, now try EEE abbreviated
2997            if(getBooleanAttribute(UDAT_PARSE_MULTIPLE_PATTERNS_FOR_MATCH, status) || count == 3) {
2998                if ((newStart = matchString(text, start, UCAL_DAY_OF_WEEK,
2999                                       fSymbols->fShortWeekdays, fSymbols->fShortWeekdaysCount, NULL, cal)) > 0)
3000                    return newStart;
3001            }
3002            // EEE abbreviated failed, now try EEEEEE short
3003            if(getBooleanAttribute(UDAT_PARSE_MULTIPLE_PATTERNS_FOR_MATCH, status) || count == 6) {
3004                if ((newStart = matchString(text, start, UCAL_DAY_OF_WEEK,
3005                                       fSymbols->fShorterWeekdays, fSymbols->fShorterWeekdaysCount, NULL, cal)) > 0)
3006                    return newStart;
3007            }
3008            // EEEEEE short failed, now try EEEEE narrow
3009            if(getBooleanAttribute(UDAT_PARSE_MULTIPLE_PATTERNS_FOR_MATCH, status) || count == 5) {
3010                if ((newStart = matchString(text, start, UCAL_DAY_OF_WEEK,
3011                                       fSymbols->fNarrowWeekdays, fSymbols->fNarrowWeekdaysCount, NULL, cal)) > 0)
3012                    return newStart;
3013            }
3014            if (!getBooleanAttribute(UDAT_PARSE_ALLOW_NUMERIC, status) || patternCharIndex == UDAT_DAY_OF_WEEK_FIELD)
3015                return newStart;
3016            // else we allowing parsing as number, below
3017        }
3018        break;
3019
3020    case UDAT_STANDALONE_DAY_FIELD:
3021        {
3022            if (gotNumber) // c or cc
3023            {
3024                // [We computed 'value' above.]
3025                cal.set(UCAL_DOW_LOCAL, value);
3026                return pos.getIndex();
3027            }
3028            // Want to be able to parse both short and long forms.
3029            // Try count == 4 (cccc) first:
3030            int32_t newStart = 0;
3031            if(getBooleanAttribute(UDAT_PARSE_MULTIPLE_PATTERNS_FOR_MATCH, status) || count == 4) {
3032                if ((newStart = matchString(text, start, UCAL_DAY_OF_WEEK,
3033                                      fSymbols->fStandaloneWeekdays, fSymbols->fStandaloneWeekdaysCount, NULL, cal)) > 0)
3034                    return newStart;
3035            }
3036            if(getBooleanAttribute(UDAT_PARSE_MULTIPLE_PATTERNS_FOR_MATCH, status) || count == 3) {
3037                if ((newStart = matchString(text, start, UCAL_DAY_OF_WEEK,
3038                                          fSymbols->fStandaloneShortWeekdays, fSymbols->fStandaloneShortWeekdaysCount, NULL, cal)) > 0)
3039                    return newStart;
3040            }
3041            if(getBooleanAttribute(UDAT_PARSE_MULTIPLE_PATTERNS_FOR_MATCH, status) || count == 6) {
3042                if ((newStart = matchString(text, start, UCAL_DAY_OF_WEEK,
3043                                          fSymbols->fStandaloneShorterWeekdays, fSymbols->fStandaloneShorterWeekdaysCount, NULL, cal)) > 0)
3044                    return newStart;
3045            }
3046            if (!getBooleanAttribute(UDAT_PARSE_ALLOW_NUMERIC, status))
3047                return newStart;
3048            // else we allowing parsing as number, below
3049        }
3050        break;
3051
3052    case UDAT_AM_PM_FIELD:
3053        {
3054            // optionally try both wide/abbrev and narrow forms
3055            int32_t newStart = 0;
3056            // try wide/abbrev
3057            if( getBooleanAttribute(UDAT_PARSE_MULTIPLE_PATTERNS_FOR_MATCH, status) || count < 5 ) {
3058                if ((newStart = matchString(text, start, UCAL_AM_PM, fSymbols->fAmPms, fSymbols->fAmPmsCount, NULL, cal)) > 0) {
3059                    return newStart;
3060                }
3061            }
3062            // try narrow
3063            if( getBooleanAttribute(UDAT_PARSE_MULTIPLE_PATTERNS_FOR_MATCH, status) || count >= 5 ) {
3064                if ((newStart = matchString(text, start, UCAL_AM_PM, fSymbols->fNarrowAmPms, fSymbols->fNarrowAmPmsCount, NULL, cal)) > 0) {
3065                    return newStart;
3066                }
3067            }
3068            // no matches for given options
3069            return -start;
3070        }
3071
3072    case UDAT_HOUR1_FIELD:
3073        // [We computed 'value' above.]
3074        if (value == cal.getLeastMaximum(UCAL_HOUR)+1)
3075            value = 0;
3076
3077        // fall through to set field
3078
3079    case UDAT_HOUR0_FIELD:
3080        cal.set(UCAL_HOUR, value);
3081        return pos.getIndex();
3082
3083    case UDAT_QUARTER_FIELD:
3084        if (gotNumber) // i.e., Q or QQ.
3085        {
3086            // Don't want to parse the month if it is a string
3087            // while pattern uses numeric style: Q or QQ.
3088            // [We computed 'value' above.]
3089            cal.set(UCAL_MONTH, (value - 1) * 3);
3090            return pos.getIndex();
3091        } else {
3092            // count >= 3 // i.e., QQQ or QQQQ
3093            // Want to be able to parse both short and long forms.
3094            // Try count == 4 first:
3095            int32_t newStart = 0;
3096
3097            if(getBooleanAttribute(UDAT_PARSE_MULTIPLE_PATTERNS_FOR_MATCH, status) || count == 4) {
3098                if ((newStart = matchQuarterString(text, start, UCAL_MONTH,
3099                                      fSymbols->fQuarters, fSymbols->fQuartersCount, cal)) > 0)
3100                    return newStart;
3101            }
3102            if(getBooleanAttribute(UDAT_PARSE_MULTIPLE_PATTERNS_FOR_MATCH, status) || count == 3) {
3103                if ((newStart = matchQuarterString(text, start, UCAL_MONTH,
3104                                          fSymbols->fShortQuarters, fSymbols->fShortQuartersCount, cal)) > 0)
3105                    return newStart;
3106            }
3107            if (!getBooleanAttribute(UDAT_PARSE_ALLOW_NUMERIC, status))
3108                return newStart;
3109            // else we allowing parsing as number, below
3110            if(!getBooleanAttribute(UDAT_PARSE_MULTIPLE_PATTERNS_FOR_MATCH, status))
3111                return -start;
3112        }
3113        break;
3114
3115    case UDAT_STANDALONE_QUARTER_FIELD:
3116        if (gotNumber) // i.e., q or qq.
3117        {
3118            // Don't want to parse the month if it is a string
3119            // while pattern uses numeric style: q or q.
3120            // [We computed 'value' above.]
3121            cal.set(UCAL_MONTH, (value - 1) * 3);
3122            return pos.getIndex();
3123        } else {
3124            // count >= 3 // i.e., qqq or qqqq
3125            // Want to be able to parse both short and long forms.
3126            // Try count == 4 first:
3127            int32_t newStart = 0;
3128
3129            if(getBooleanAttribute(UDAT_PARSE_MULTIPLE_PATTERNS_FOR_MATCH, status) || count == 4) {
3130                if ((newStart = matchQuarterString(text, start, UCAL_MONTH,
3131                                      fSymbols->fStandaloneQuarters, fSymbols->fStandaloneQuartersCount, cal)) > 0)
3132                    return newStart;
3133            }
3134            if(getBooleanAttribute(UDAT_PARSE_MULTIPLE_PATTERNS_FOR_MATCH, status) || count == 3) {
3135                if ((newStart = matchQuarterString(text, start, UCAL_MONTH,
3136                                          fSymbols->fStandaloneShortQuarters, fSymbols->fStandaloneShortQuartersCount, cal)) > 0)
3137                    return newStart;
3138            }
3139            if (!getBooleanAttribute(UDAT_PARSE_ALLOW_NUMERIC, status))
3140                return newStart;
3141            // else we allowing parsing as number, below
3142            if(!getBooleanAttribute(UDAT_PARSE_MULTIPLE_PATTERNS_FOR_MATCH, status))
3143                return -start;
3144        }
3145        break;
3146
3147    case UDAT_TIMEZONE_FIELD: // 'z'
3148        {
3149            UTimeZoneFormatStyle style = (count < 4) ? UTZFMT_STYLE_SPECIFIC_SHORT : UTZFMT_STYLE_SPECIFIC_LONG;
3150            TimeZone *tz  = tzFormat()->parse(style, text, pos, tzTimeType);
3151            if (tz != NULL) {
3152                cal.adoptTimeZone(tz);
3153                return pos.getIndex();
3154            }
3155        }
3156        break;
3157    case UDAT_TIMEZONE_RFC_FIELD: // 'Z'
3158        {
3159            UTimeZoneFormatStyle style = (count < 4) ?
3160                UTZFMT_STYLE_ISO_BASIC_LOCAL_FULL : ((count == 5) ? UTZFMT_STYLE_ISO_EXTENDED_FULL: UTZFMT_STYLE_LOCALIZED_GMT);
3161            TimeZone *tz  = tzFormat()->parse(style, text, pos, tzTimeType);
3162            if (tz != NULL) {
3163                cal.adoptTimeZone(tz);
3164                return pos.getIndex();
3165            }
3166            return -start;
3167        }
3168    case UDAT_TIMEZONE_GENERIC_FIELD: // 'v'
3169        {
3170            UTimeZoneFormatStyle style = (count < 4) ? UTZFMT_STYLE_GENERIC_SHORT : UTZFMT_STYLE_GENERIC_LONG;
3171            TimeZone *tz  = tzFormat()->parse(style, text, pos, tzTimeType);
3172            if (tz != NULL) {
3173                cal.adoptTimeZone(tz);
3174                return pos.getIndex();
3175            }
3176            return -start;
3177        }
3178    case UDAT_TIMEZONE_SPECIAL_FIELD: // 'V'
3179        {
3180            UTimeZoneFormatStyle style;
3181            switch (count) {
3182            case 1:
3183                style = UTZFMT_STYLE_ZONE_ID_SHORT;
3184                break;
3185            case 2:
3186                style = UTZFMT_STYLE_ZONE_ID;
3187                break;
3188            case 3:
3189                style = UTZFMT_STYLE_EXEMPLAR_LOCATION;
3190                break;
3191            default:
3192                style = UTZFMT_STYLE_GENERIC_LOCATION;
3193                break;
3194            }
3195            TimeZone *tz  = tzFormat()->parse(style, text, pos, tzTimeType);
3196            if (tz != NULL) {
3197                cal.adoptTimeZone(tz);
3198                return pos.getIndex();
3199            }
3200            return -start;
3201        }
3202    case UDAT_TIMEZONE_LOCALIZED_GMT_OFFSET_FIELD: // 'O'
3203        {
3204            UTimeZoneFormatStyle style = (count < 4) ? UTZFMT_STYLE_LOCALIZED_GMT_SHORT : UTZFMT_STYLE_LOCALIZED_GMT;
3205            TimeZone *tz  = tzFormat()->parse(style, text, pos, tzTimeType);
3206            if (tz != NULL) {
3207                cal.adoptTimeZone(tz);
3208                return pos.getIndex();
3209            }
3210            return -start;
3211        }
3212    case UDAT_TIMEZONE_ISO_FIELD: // 'X'
3213        {
3214            UTimeZoneFormatStyle style;
3215            switch (count) {
3216            case 1:
3217                style = UTZFMT_STYLE_ISO_BASIC_SHORT;
3218                break;
3219            case 2:
3220                style = UTZFMT_STYLE_ISO_BASIC_FIXED;
3221                break;
3222            case 3:
3223                style = UTZFMT_STYLE_ISO_EXTENDED_FIXED;
3224                break;
3225            case 4:
3226                style = UTZFMT_STYLE_ISO_BASIC_FULL;
3227                break;
3228            default:
3229                style = UTZFMT_STYLE_ISO_EXTENDED_FULL;
3230                break;
3231            }
3232            TimeZone *tz  = tzFormat()->parse(style, text, pos, tzTimeType);
3233            if (tz != NULL) {
3234                cal.adoptTimeZone(tz);
3235                return pos.getIndex();
3236            }
3237            return -start;
3238        }
3239    case UDAT_TIMEZONE_ISO_LOCAL_FIELD: // 'x'
3240        {
3241            UTimeZoneFormatStyle style;
3242            switch (count) {
3243            case 1:
3244                style = UTZFMT_STYLE_ISO_BASIC_LOCAL_SHORT;
3245                break;
3246            case 2:
3247                style = UTZFMT_STYLE_ISO_BASIC_LOCAL_FIXED;
3248                break;
3249            case 3:
3250                style = UTZFMT_STYLE_ISO_EXTENDED_LOCAL_FIXED;
3251                break;
3252            case 4:
3253                style = UTZFMT_STYLE_ISO_BASIC_LOCAL_FULL;
3254                break;
3255            default:
3256                style = UTZFMT_STYLE_ISO_EXTENDED_LOCAL_FULL;
3257                break;
3258            }
3259            TimeZone *tz  = tzFormat()->parse(style, text, pos, tzTimeType);
3260            if (tz != NULL) {
3261                cal.adoptTimeZone(tz);
3262                return pos.getIndex();
3263            }
3264            return -start;
3265        }
3266    // currently no pattern character is defined for UDAT_TIME_SEPARATOR_FIELD
3267    // so we should not get here. Leave support in for future definition.
3268    case UDAT_TIME_SEPARATOR_FIELD: //
3269        {
3270            static const UChar def_sep = DateFormatSymbols::DEFAULT_TIME_SEPARATOR;
3271            static const UChar alt_sep = DateFormatSymbols::ALTERNATE_TIME_SEPARATOR;
3272
3273            // Try matching a time separator.
3274            int32_t count = 1;
3275            UnicodeString data[3];
3276            fSymbols->getTimeSeparatorString(data[0]);
3277
3278            // Add the default, if different from the locale.
3279            if (data[0].compare(&def_sep, 1) != 0) {
3280                data[count++].setTo(def_sep);
3281            }
3282
3283            // If lenient, add also the alternate, if different from the locale.
3284            if (isLenient() && data[0].compare(&alt_sep, 1) != 0) {
3285                data[count++].setTo(alt_sep);
3286            }
3287
3288            return matchString(text, start, UCAL_FIELD_COUNT /* => nothing to set */, data, count, NULL, cal);
3289        }
3290
3291    default:
3292        // Handle "generic" fields
3293        // this is now handled below, outside the switch block
3294        break;
3295    }
3296    // Handle "generic" fields:
3297    // switch default case now handled here (outside switch block) to allow
3298    // parsing of some string fields as digits for lenient case
3299
3300    int32_t parseStart = pos.getIndex();
3301    const UnicodeString* src;
3302    if (obeyCount) {
3303        if ((start+count) > text.length()) {
3304            return -start;
3305        }
3306        text.extractBetween(0, start + count, temp);
3307        src = &temp;
3308    } else {
3309        src = &text;
3310    }
3311    parseInt(*src, number, pos, allowNegative,currentNumberFormat);
3312    if (pos.getIndex() != parseStart) {
3313        int32_t value = number.getLong();
3314
3315        // Don't need suffix processing here (as in number processing at the beginning of the function);
3316        // the new fields being handled as numeric values (month, weekdays, quarters) should not have suffixes.
3317
3318        if (!getBooleanAttribute(UDAT_PARSE_ALLOW_NUMERIC, status)) {
3319            // Check the range of the value
3320            int32_t bias = gFieldRangeBias[patternCharIndex];
3321            if (bias >= 0 && (value > cal.getMaximum(field) + bias || value < cal.getMinimum(field) + bias)) {
3322                return -start;
3323            }
3324        }
3325
3326        // For the following, need to repeat some of the "if (gotNumber)" code above:
3327        // UDAT_[STANDALONE_]MONTH_FIELD, UDAT_DOW_LOCAL_FIELD, UDAT_STANDALONE_DAY_FIELD,
3328        // UDAT_[STANDALONE_]QUARTER_FIELD
3329        switch (patternCharIndex) {
3330        case UDAT_MONTH_FIELD:
3331            // See notes under UDAT_MONTH_FIELD case above
3332            if (!strcmp(cal.getType(),"hebrew")) {
3333                HebrewCalendar *hc = (HebrewCalendar*)&cal;
3334                if (cal.isSet(UCAL_YEAR)) {
3335                   UErrorCode status = U_ZERO_ERROR;
3336                   if (!hc->isLeapYear(hc->get(UCAL_YEAR,status)) && value >= 6) {
3337                       cal.set(UCAL_MONTH, value);
3338                   } else {
3339                       cal.set(UCAL_MONTH, value - 1);
3340                   }
3341                } else {
3342                    saveHebrewMonth = value;
3343                }
3344            } else {
3345                cal.set(UCAL_MONTH, value - 1);
3346            }
3347            break;
3348        case UDAT_STANDALONE_MONTH_FIELD:
3349            cal.set(UCAL_MONTH, value - 1);
3350            break;
3351        case UDAT_DOW_LOCAL_FIELD:
3352        case UDAT_STANDALONE_DAY_FIELD:
3353            cal.set(UCAL_DOW_LOCAL, value);
3354            break;
3355        case UDAT_QUARTER_FIELD:
3356        case UDAT_STANDALONE_QUARTER_FIELD:
3357             cal.set(UCAL_MONTH, (value - 1) * 3);
3358             break;
3359        case UDAT_RELATED_YEAR_FIELD:
3360            cal.setRelatedYear(value);
3361            break;
3362        default:
3363            cal.set(field, value);
3364            break;
3365        }
3366        return pos.getIndex();
3367    }
3368    return -start;
3369}
3370
3371/**
3372 * Parse an integer using fNumberFormat.  This method is semantically
3373 * const, but actually may modify fNumberFormat.
3374 */
3375void SimpleDateFormat::parseInt(const UnicodeString& text,
3376                                Formattable& number,
3377                                ParsePosition& pos,
3378                                UBool allowNegative,
3379                                NumberFormat *fmt) const {
3380    parseInt(text, number, -1, pos, allowNegative,fmt);
3381}
3382
3383/**
3384 * Parse an integer using fNumberFormat up to maxDigits.
3385 */
3386void SimpleDateFormat::parseInt(const UnicodeString& text,
3387                                Formattable& number,
3388                                int32_t maxDigits,
3389                                ParsePosition& pos,
3390                                UBool allowNegative,
3391                                NumberFormat *fmt) const {
3392    UnicodeString oldPrefix;
3393    DecimalFormat* df = NULL;
3394    if (!allowNegative && (df = dynamic_cast<DecimalFormat*>(fmt)) != NULL) {
3395        df->getNegativePrefix(oldPrefix);
3396        df->setNegativePrefix(UnicodeString(TRUE, SUPPRESS_NEGATIVE_PREFIX, -1));
3397    }
3398    int32_t oldPos = pos.getIndex();
3399    fmt->parse(text, number, pos);
3400    if (df != NULL) {
3401        df->setNegativePrefix(oldPrefix);
3402    }
3403
3404    if (maxDigits > 0) {
3405        // adjust the result to fit into
3406        // the maxDigits and move the position back
3407        int32_t nDigits = pos.getIndex() - oldPos;
3408        if (nDigits > maxDigits) {
3409            int32_t val = number.getLong();
3410            nDigits -= maxDigits;
3411            while (nDigits > 0) {
3412                val /= 10;
3413                nDigits--;
3414            }
3415            pos.setIndex(oldPos + maxDigits);
3416            number.setLong(val);
3417        }
3418    }
3419}
3420
3421//----------------------------------------------------------------------
3422
3423void SimpleDateFormat::translatePattern(const UnicodeString& originalPattern,
3424                                        UnicodeString& translatedPattern,
3425                                        const UnicodeString& from,
3426                                        const UnicodeString& to,
3427                                        UErrorCode& status)
3428{
3429    // run through the pattern and convert any pattern symbols from the version
3430    // in "from" to the corresponding character ion "to".  This code takes
3431    // quoted strings into account (it doesn't try to translate them), and it signals
3432    // an error if a particular "pattern character" doesn't appear in "from".
3433    // Depending on the values of "from" and "to" this can convert from generic
3434    // to localized patterns or localized to generic.
3435    if (U_FAILURE(status)) {
3436        return;
3437    }
3438
3439    translatedPattern.remove();
3440    UBool inQuote = FALSE;
3441    for (int32_t i = 0; i < originalPattern.length(); ++i) {
3442        UChar c = originalPattern[i];
3443        if (inQuote) {
3444            if (c == QUOTE) {
3445                inQuote = FALSE;
3446            }
3447        } else {
3448            if (c == QUOTE) {
3449                inQuote = TRUE;
3450            } else if (isSyntaxChar(c)) {
3451                int32_t ci = from.indexOf(c);
3452                if (ci == -1) {
3453                    status = U_INVALID_FORMAT_ERROR;
3454                    return;
3455                }
3456                c = to[ci];
3457            }
3458        }
3459        translatedPattern += c;
3460    }
3461    if (inQuote) {
3462        status = U_INVALID_FORMAT_ERROR;
3463        return;
3464    }
3465}
3466
3467//----------------------------------------------------------------------
3468
3469UnicodeString&
3470SimpleDateFormat::toPattern(UnicodeString& result) const
3471{
3472    result = fPattern;
3473    return result;
3474}
3475
3476//----------------------------------------------------------------------
3477
3478UnicodeString&
3479SimpleDateFormat::toLocalizedPattern(UnicodeString& result,
3480                                     UErrorCode& status) const
3481{
3482    translatePattern(fPattern, result,
3483                     UnicodeString(DateFormatSymbols::getPatternUChars()),
3484                     fSymbols->fLocalPatternChars, status);
3485    return result;
3486}
3487
3488//----------------------------------------------------------------------
3489
3490void
3491SimpleDateFormat::applyPattern(const UnicodeString& pattern)
3492{
3493    fPattern = pattern;
3494}
3495
3496//----------------------------------------------------------------------
3497
3498void
3499SimpleDateFormat::applyLocalizedPattern(const UnicodeString& pattern,
3500                                        UErrorCode &status)
3501{
3502    translatePattern(pattern, fPattern,
3503                     fSymbols->fLocalPatternChars,
3504                     UnicodeString(DateFormatSymbols::getPatternUChars()), status);
3505}
3506
3507//----------------------------------------------------------------------
3508
3509const DateFormatSymbols*
3510SimpleDateFormat::getDateFormatSymbols() const
3511{
3512    return fSymbols;
3513}
3514
3515//----------------------------------------------------------------------
3516
3517void
3518SimpleDateFormat::adoptDateFormatSymbols(DateFormatSymbols* newFormatSymbols)
3519{
3520    delete fSymbols;
3521    fSymbols = newFormatSymbols;
3522}
3523
3524//----------------------------------------------------------------------
3525void
3526SimpleDateFormat::setDateFormatSymbols(const DateFormatSymbols& newFormatSymbols)
3527{
3528    delete fSymbols;
3529    fSymbols = new DateFormatSymbols(newFormatSymbols);
3530}
3531
3532//----------------------------------------------------------------------
3533const TimeZoneFormat*
3534SimpleDateFormat::getTimeZoneFormat(void) const {
3535    return (const TimeZoneFormat*)tzFormat();
3536}
3537
3538//----------------------------------------------------------------------
3539void
3540SimpleDateFormat::adoptTimeZoneFormat(TimeZoneFormat* timeZoneFormatToAdopt)
3541{
3542    delete fTimeZoneFormat;
3543    fTimeZoneFormat = timeZoneFormatToAdopt;
3544}
3545
3546//----------------------------------------------------------------------
3547void
3548SimpleDateFormat::setTimeZoneFormat(const TimeZoneFormat& newTimeZoneFormat)
3549{
3550    delete fTimeZoneFormat;
3551    fTimeZoneFormat = new TimeZoneFormat(newTimeZoneFormat);
3552}
3553
3554//----------------------------------------------------------------------
3555
3556
3557void SimpleDateFormat::adoptCalendar(Calendar* calendarToAdopt)
3558{
3559  UErrorCode status = U_ZERO_ERROR;
3560  Locale calLocale(fLocale);
3561  calLocale.setKeywordValue("calendar", calendarToAdopt->getType(), status);
3562  DateFormatSymbols *newSymbols =
3563          DateFormatSymbols::createForLocale(calLocale, status);
3564  if (U_FAILURE(status)) {
3565      return;
3566  }
3567  DateFormat::adoptCalendar(calendarToAdopt);
3568  delete fSymbols;
3569  fSymbols = newSymbols;
3570  initializeDefaultCentury();  // we need a new century (possibly)
3571}
3572
3573
3574//----------------------------------------------------------------------
3575
3576
3577// override the DateFormat implementation in order to
3578// lazily initialize fCapitalizationBrkIter
3579void
3580SimpleDateFormat::setContext(UDisplayContext value, UErrorCode& status)
3581{
3582    DateFormat::setContext(value, status);
3583#if !UCONFIG_NO_BREAK_ITERATION
3584    if (U_SUCCESS(status)) {
3585        if ( fCapitalizationBrkIter == NULL && (value==UDISPCTX_CAPITALIZATION_FOR_BEGINNING_OF_SENTENCE ||
3586                value==UDISPCTX_CAPITALIZATION_FOR_UI_LIST_OR_MENU || value==UDISPCTX_CAPITALIZATION_FOR_STANDALONE) ) {
3587            UErrorCode status = U_ZERO_ERROR;
3588            fCapitalizationBrkIter = BreakIterator::createSentenceInstance(fLocale, status);
3589            if (U_FAILURE(status)) {
3590                delete fCapitalizationBrkIter;
3591                fCapitalizationBrkIter = NULL;
3592            }
3593        }
3594    }
3595#endif
3596}
3597
3598
3599//----------------------------------------------------------------------
3600
3601
3602UBool
3603SimpleDateFormat::isFieldUnitIgnored(UCalendarDateFields field) const {
3604    return isFieldUnitIgnored(fPattern, field);
3605}
3606
3607
3608UBool
3609SimpleDateFormat::isFieldUnitIgnored(const UnicodeString& pattern,
3610                                     UCalendarDateFields field) {
3611    int32_t fieldLevel = fgCalendarFieldToLevel[field];
3612    int32_t level;
3613    UChar ch;
3614    UBool inQuote = FALSE;
3615    UChar prevCh = 0;
3616    int32_t count = 0;
3617
3618    for (int32_t i = 0; i < pattern.length(); ++i) {
3619        ch = pattern[i];
3620        if (ch != prevCh && count > 0) {
3621            level = getLevelFromChar(prevCh);
3622            // the larger the level, the smaller the field unit.
3623            if (fieldLevel <= level) {
3624                return FALSE;
3625            }
3626            count = 0;
3627        }
3628        if (ch == QUOTE) {
3629            if ((i+1) < pattern.length() && pattern[i+1] == QUOTE) {
3630                ++i;
3631            } else {
3632                inQuote = ! inQuote;
3633            }
3634        }
3635        else if (!inQuote && isSyntaxChar(ch)) {
3636            prevCh = ch;
3637            ++count;
3638        }
3639    }
3640    if (count > 0) {
3641        // last item
3642        level = getLevelFromChar(prevCh);
3643        if (fieldLevel <= level) {
3644            return FALSE;
3645        }
3646    }
3647    return TRUE;
3648}
3649
3650//----------------------------------------------------------------------
3651
3652const Locale&
3653SimpleDateFormat::getSmpFmtLocale(void) const {
3654    return fLocale;
3655}
3656
3657//----------------------------------------------------------------------
3658
3659int32_t
3660SimpleDateFormat::checkIntSuffix(const UnicodeString& text, int32_t start,
3661                                 int32_t patLoc, UBool isNegative) const {
3662    // local variables
3663    UnicodeString suf;
3664    int32_t patternMatch;
3665    int32_t textPreMatch;
3666    int32_t textPostMatch;
3667
3668    // check that we are still in range
3669    if ( (start > text.length()) ||
3670         (start < 0) ||
3671         (patLoc < 0) ||
3672         (patLoc > fPattern.length())) {
3673        // out of range, don't advance location in text
3674        return start;
3675    }
3676
3677    // get the suffix
3678    DecimalFormat* decfmt = dynamic_cast<DecimalFormat*>(fNumberFormat);
3679    if (decfmt != NULL) {
3680        if (isNegative) {
3681            suf = decfmt->getNegativeSuffix(suf);
3682        }
3683        else {
3684            suf = decfmt->getPositiveSuffix(suf);
3685        }
3686    }
3687
3688    // check for suffix
3689    if (suf.length() <= 0) {
3690        return start;
3691    }
3692
3693    // check suffix will be encountered in the pattern
3694    patternMatch = compareSimpleAffix(suf,fPattern,patLoc);
3695
3696    // check if a suffix will be encountered in the text
3697    textPreMatch = compareSimpleAffix(suf,text,start);
3698
3699    // check if a suffix was encountered in the text
3700    textPostMatch = compareSimpleAffix(suf,text,start-suf.length());
3701
3702    // check for suffix match
3703    if ((textPreMatch >= 0) && (patternMatch >= 0) && (textPreMatch == patternMatch)) {
3704        return start;
3705    }
3706    else if ((textPostMatch >= 0) && (patternMatch >= 0) && (textPostMatch == patternMatch)) {
3707        return  start - suf.length();
3708    }
3709
3710    // should not get here
3711    return start;
3712}
3713
3714//----------------------------------------------------------------------
3715
3716int32_t
3717SimpleDateFormat::compareSimpleAffix(const UnicodeString& affix,
3718                   const UnicodeString& input,
3719                   int32_t pos) const {
3720    int32_t start = pos;
3721    for (int32_t i=0; i<affix.length(); ) {
3722        UChar32 c = affix.char32At(i);
3723        int32_t len = U16_LENGTH(c);
3724        if (PatternProps::isWhiteSpace(c)) {
3725            // We may have a pattern like: \u200F \u0020
3726            //        and input text like: \u200F \u0020
3727            // Note that U+200F and U+0020 are Pattern_White_Space but only
3728            // U+0020 is UWhiteSpace.  So we have to first do a direct
3729            // match of the run of Pattern_White_Space in the pattern,
3730            // then match any extra characters.
3731            UBool literalMatch = FALSE;
3732            while (pos < input.length() &&
3733                   input.char32At(pos) == c) {
3734                literalMatch = TRUE;
3735                i += len;
3736                pos += len;
3737                if (i == affix.length()) {
3738                    break;
3739                }
3740                c = affix.char32At(i);
3741                len = U16_LENGTH(c);
3742                if (!PatternProps::isWhiteSpace(c)) {
3743                    break;
3744                }
3745            }
3746
3747            // Advance over run in pattern
3748            i = skipPatternWhiteSpace(affix, i);
3749
3750            // Advance over run in input text
3751            // Must see at least one white space char in input,
3752            // unless we've already matched some characters literally.
3753            int32_t s = pos;
3754            pos = skipUWhiteSpace(input, pos);
3755            if (pos == s && !literalMatch) {
3756                return -1;
3757            }
3758
3759            // If we skip UWhiteSpace in the input text, we need to skip it in the pattern.
3760            // Otherwise, the previous lines may have skipped over text (such as U+00A0) that
3761            // is also in the affix.
3762            i = skipUWhiteSpace(affix, i);
3763        } else {
3764            if (pos < input.length() &&
3765                input.char32At(pos) == c) {
3766                i += len;
3767                pos += len;
3768            } else {
3769                return -1;
3770            }
3771        }
3772    }
3773    return pos - start;
3774}
3775
3776//----------------------------------------------------------------------
3777
3778int32_t
3779SimpleDateFormat::skipPatternWhiteSpace(const UnicodeString& text, int32_t pos) const {
3780    const UChar* s = text.getBuffer();
3781    return (int32_t)(PatternProps::skipWhiteSpace(s + pos, text.length() - pos) - s);
3782}
3783
3784//----------------------------------------------------------------------
3785
3786int32_t
3787SimpleDateFormat::skipUWhiteSpace(const UnicodeString& text, int32_t pos) const {
3788    while (pos < text.length()) {
3789        UChar32 c = text.char32At(pos);
3790        if (!u_isUWhiteSpace(c)) {
3791            break;
3792        }
3793        pos += U16_LENGTH(c);
3794    }
3795    return pos;
3796}
3797
3798//----------------------------------------------------------------------
3799
3800// Lazy TimeZoneFormat instantiation, semantically const.
3801TimeZoneFormat *
3802SimpleDateFormat::tzFormat() const {
3803    if (fTimeZoneFormat == NULL) {
3804        umtx_lock(&LOCK);
3805        {
3806            if (fTimeZoneFormat == NULL) {
3807                UErrorCode status = U_ZERO_ERROR;
3808                TimeZoneFormat *tzfmt = TimeZoneFormat::createInstance(fLocale, status);
3809                if (U_FAILURE(status)) {
3810                    return NULL;
3811                }
3812
3813                const_cast<SimpleDateFormat *>(this)->fTimeZoneFormat = tzfmt;
3814            }
3815        }
3816        umtx_unlock(&LOCK);
3817    }
3818    return fTimeZoneFormat;
3819}
3820
3821U_NAMESPACE_END
3822
3823#endif /* #if !UCONFIG_NO_FORMATTING */
3824
3825//eof
3826