Bug Summary

File:root/firefox-clang/intl/icu/source/i18n/smpdtfmt.cpp
Warning:line 2032, column 13
Called C++ object pointer is null

Annotated Source Code

Press '?' to see keyboard shortcuts

clang -cc1 -cc1 -triple x86_64-pc-linux-gnu -analyze -disable-free -clear-ast-before-backend -disable-llvm-verifier -discard-value-names -main-file-name smpdtfmt.cpp -analyzer-checker=core -analyzer-checker=apiModeling -analyzer-checker=unix -analyzer-checker=deadcode -analyzer-checker=cplusplus -analyzer-checker=security.insecureAPI.UncheckedReturn -analyzer-checker=security.insecureAPI.getpw -analyzer-checker=security.insecureAPI.gets -analyzer-checker=security.insecureAPI.mktemp -analyzer-checker=security.insecureAPI.mkstemp -analyzer-checker=security.insecureAPI.vfork -analyzer-checker=nullability.NullPassedToNonnull -analyzer-checker=nullability.NullReturnedFromNonnull -analyzer-output plist -w -setup-static-analyzer -analyzer-config-compatibility-mode=true -mrelocation-model pic -pic-level 2 -fhalf-no-semantic-interposition -mframe-pointer=all -relaxed-aliasing -ffp-contract=off -fno-rounding-math -mconstructor-aliases -funwind-tables=2 -target-cpu x86-64 -tune-cpu generic -debugger-tuning=gdb -fdebug-compilation-dir=/root/firefox-clang/obj-x86_64-pc-linux-gnu/config/external/icu/i18n -fcoverage-compilation-dir=/root/firefox-clang/obj-x86_64-pc-linux-gnu/config/external/icu/i18n -resource-dir /usr/lib/llvm-21/lib/clang/21 -include /root/firefox-clang/config/gcc_hidden.h -include /root/firefox-clang/obj-x86_64-pc-linux-gnu/mozilla-config.h -I /root/firefox-clang/obj-x86_64-pc-linux-gnu/dist/system_wrappers -U _FORTIFY_SOURCE -D _FORTIFY_SOURCE=2 -D _GLIBCXX_ASSERTIONS -D DEBUG=1 -D U_I18N_IMPLEMENTATION -D _LIBCPP_DISABLE_DEPRECATION_WARNINGS -D U_USING_ICU_NAMESPACE=0 -D U_NO_DEFAULT_INCLUDE_UTF_HEADERS=1 -D U_HIDE_OBSOLETE_UTF_OLD_H=1 -D UCONFIG_NO_LEGACY_CONVERSION -D UCONFIG_NO_TRANSLITERATION -D UCONFIG_NO_REGULAR_EXPRESSIONS -D UCONFIG_NO_BREAK_ITERATION -D UCONFIG_NO_IDNA -D UCONFIG_NO_MF2 -D U_CHARSET_IS_UTF8 -D UNISTR_FROM_CHAR_EXPLICIT=explicit -D UNISTR_FROM_STRING_EXPLICIT=explicit -D U_ENABLE_DYLOAD=0 -D U_DEBUG=1 -I /root/firefox-clang/config/external/icu/i18n -I /root/firefox-clang/obj-x86_64-pc-linux-gnu/config/external/icu/i18n -I /root/firefox-clang/intl/icu/source/common -I /root/firefox-clang/mfbt/double-conversion -I /root/firefox-clang/obj-x86_64-pc-linux-gnu/dist/include -I /root/firefox-clang/obj-x86_64-pc-linux-gnu/dist/include/nspr -I /root/firefox-clang/obj-x86_64-pc-linux-gnu/dist/include/nss -D MOZILLA_CLIENT -internal-isystem /usr/lib/gcc/x86_64-linux-gnu/14/../../../../include/c++/14 -internal-isystem /usr/lib/gcc/x86_64-linux-gnu/14/../../../../include/x86_64-linux-gnu/c++/14 -internal-isystem /usr/lib/gcc/x86_64-linux-gnu/14/../../../../include/c++/14/backward -internal-isystem /usr/lib/llvm-21/lib/clang/21/include -internal-isystem /usr/local/include -internal-isystem /usr/lib/gcc/x86_64-linux-gnu/14/../../../../x86_64-linux-gnu/include -internal-externc-isystem /usr/include/x86_64-linux-gnu -internal-externc-isystem /include -internal-externc-isystem /usr/include -O2 -Wno-error=pessimizing-move -Wno-error=large-by-value-copy=128 -Wno-error=implicit-int-float-conversion -Wno-error=thread-safety-analysis -Wno-error=tautological-type-limit-compare -Wno-invalid-offsetof -Wno-range-loop-analysis -Wno-deprecated-anon-enum-enum-conversion -Wno-deprecated-enum-enum-conversion -Wno-deprecated-this-capture -Wno-inline-new-delete -Wno-error=deprecated-declarations -Wno-error=array-bounds -Wno-error=free-nonheap-object -Wno-error=atomic-alignment -Wno-error=deprecated-builtins -Wno-psabi -Wno-error=builtin-macro-redefined -Wno-vla-cxx-extension -Wno-unknown-warning-option -Wno-comma -Wno-implicit-const-int-float-conversion -Wno-macro-redefined -Wno-microsoft-include -Wno-tautological-unsigned-enum-zero-compare -Wno-unreachable-code-loop-increment -Wno-unreachable-code-return -fdeprecated-macro -ferror-limit 19 -fstrict-flex-arrays=1 -stack-protector 2 -fstack-clash-protection -ftrivial-auto-var-init=pattern -fgnuc-version=4.2.1 -fskip-odr-check-in-gmf -fno-sized-deallocation -fno-aligned-allocation -vectorize-loops -vectorize-slp -analyzer-checker optin.performance.Padding -analyzer-output=html -analyzer-config stable-report-filename=true -faddrsig -D__GCC_HAVE_DWARF2_CFI_ASM=1 -o /tmp/scan-build-2025-06-30-093548-1913035-1 -x c++ /root/firefox-clang/intl/icu/source/i18n/smpdtfmt.cpp
1// © 2016 and later: Unicode, Inc. and others.
2// License & terms of use: http://www.unicode.org/copyright.html
3/*
4*******************************************************************************
5* Copyright (C) 1997-2016, International Business Machines Corporation and *
6* others. All Rights Reserved. *
7*******************************************************************************
8*
9* File SMPDTFMT.CPP
10*
11* Modification History:
12*
13* Date Name Description
14* 02/19/97 aliu Converted from java.
15* 03/31/97 aliu Modified extensively to work with 50 locales.
16* 04/01/97 aliu Added support for centuries.
17* 07/09/97 helena Made ParsePosition into a class.
18* 07/21/98 stephen Added initializeDefaultCentury.
19* Removed getZoneIndex (added in DateFormatSymbols)
20* Removed subParseLong
21* Removed chk
22* 02/22/99 stephen Removed character literals for EBCDIC safety
23* 10/14/99 aliu Updated 2-digit year parsing so that only "00" thru
24* "99" are recognized. {j28 4182066}
25* 11/15/99 weiv Added support for week of year/day of week format
26********************************************************************************
27*/
28
29#define ZID_KEY_MAX128 128
30
31#include "unicode/utypes.h"
32
33#if !UCONFIG_NO_FORMATTING0
34#include "unicode/smpdtfmt.h"
35#include "unicode/dtfmtsym.h"
36#include "unicode/ures.h"
37#include "unicode/msgfmt.h"
38#include "unicode/calendar.h"
39#include "unicode/gregocal.h"
40#include "unicode/timezone.h"
41#include "unicode/decimfmt.h"
42#include "unicode/dcfmtsym.h"
43#include "unicode/uchar.h"
44#include "unicode/uniset.h"
45#include "unicode/ustring.h"
46#include "unicode/basictz.h"
47#include "unicode/simpleformatter.h"
48#include "unicode/simplenumberformatter.h"
49#include "unicode/simpletz.h"
50#include "unicode/rbtz.h"
51#include "unicode/tzfmt.h"
52#include "unicode/ucasemap.h"
53#include "unicode/utf16.h"
54#include "unicode/vtzone.h"
55#include "unicode/udisplaycontext.h"
56#include "unicode/brkiter.h"
57#include "unicode/rbnf.h"
58#include "unicode/dtptngen.h"
59#include "uresimp.h"
60#include "olsontz.h"
61#include "patternprops.h"
62#include "fphdlimp.h"
63#include "hebrwcal.h"
64#include "cstring.h"
65#include "uassert.h"
66#include "cmemory.h"
67#include "umutex.h"
68#include "mutex.h"
69#include <float.h>
70#include "smpdtfst.h"
71#include "sharednumberformat.h"
72#include "ucasemap_imp.h"
73#include "ustr_imp.h"
74#include "charstr.h"
75#include "uvector.h"
76#include "cstr.h"
77#include "dayperiodrules.h"
78#include "tznames_impl.h" // ZONE_NAME_U16_MAX
79#include "number_utypes.h"
80#include "chnsecal.h"
81#include "dangical.h"
82#include "japancal.h"
83#include <typeinfo>
84
85#if defined( U_DEBUG_CALSVC ) || defined (U_DEBUG_CAL)
86#include <stdio.h>
87#endif
88
89// *****************************************************************************
90// class SimpleDateFormat
91// *****************************************************************************
92
93U_NAMESPACE_BEGINnamespace icu_77 {
94
95/**
96 * Last-resort string to use for "GMT" when constructing time zone strings.
97 */
98// For time zones that have no names, use strings GMT+minutes and
99// GMT-minutes. For instance, in France the time zone is GMT+60.
100// Also accepted are GMT+H:MM or GMT-H:MM.
101// Currently not being used
102//static const char16_t gGmt[] = {0x0047, 0x004D, 0x0054, 0x0000}; // "GMT"
103//static const char16_t gGmtPlus[] = {0x0047, 0x004D, 0x0054, 0x002B, 0x0000}; // "GMT+"
104//static const char16_t gGmtMinus[] = {0x0047, 0x004D, 0x0054, 0x002D, 0x0000}; // "GMT-"
105//static const char16_t gDefGmtPat[] = {0x0047, 0x004D, 0x0054, 0x007B, 0x0030, 0x007D, 0x0000}; /* GMT{0} */
106//static const char16_t gDefGmtNegHmsPat[] = {0x002D, 0x0048, 0x0048, 0x003A, 0x006D, 0x006D, 0x003A, 0x0073, 0x0073, 0x0000}; /* -HH:mm:ss */
107//static const char16_t gDefGmtNegHmPat[] = {0x002D, 0x0048, 0x0048, 0x003A, 0x006D, 0x006D, 0x0000}; /* -HH:mm */
108//static const char16_t gDefGmtPosHmsPat[] = {0x002B, 0x0048, 0x0048, 0x003A, 0x006D, 0x006D, 0x003A, 0x0073, 0x0073, 0x0000}; /* +HH:mm:ss */
109//static const char16_t gDefGmtPosHmPat[] = {0x002B, 0x0048, 0x0048, 0x003A, 0x006D, 0x006D, 0x0000}; /* +HH:mm */
110//static const char16_t gUt[] = {0x0055, 0x0054, 0x0000}; // "UT"
111//static const char16_t gUtc[] = {0x0055, 0x0054, 0x0043, 0x0000}; // "UT"
112
113typedef enum GmtPatSize {
114 kGmtLen = 3,
115 kGmtPatLen = 6,
116 kNegHmsLen = 9,
117 kNegHmLen = 6,
118 kPosHmsLen = 9,
119 kPosHmLen = 6,
120 kUtLen = 2,
121 kUtcLen = 3
122} GmtPatSize;
123
124// Stuff needed for numbering system overrides
125
126typedef enum OvrStrType {
127 kOvrStrDate = 0,
128 kOvrStrTime = 1,
129 kOvrStrBoth = 2
130} OvrStrType;
131
132static const UDateFormatField kDateFields[] = {
133 UDAT_YEAR_FIELD,
134 UDAT_MONTH_FIELD,
135 UDAT_DATE_FIELD,
136 UDAT_DAY_OF_YEAR_FIELD,
137 UDAT_DAY_OF_WEEK_IN_MONTH_FIELD,
138 UDAT_WEEK_OF_YEAR_FIELD,
139 UDAT_WEEK_OF_MONTH_FIELD,
140 UDAT_YEAR_WOY_FIELD,
141 UDAT_EXTENDED_YEAR_FIELD,
142 UDAT_JULIAN_DAY_FIELD,
143 UDAT_STANDALONE_DAY_FIELD,
144 UDAT_STANDALONE_MONTH_FIELD,
145 UDAT_QUARTER_FIELD,
146 UDAT_STANDALONE_QUARTER_FIELD,
147 UDAT_YEAR_NAME_FIELD,
148 UDAT_RELATED_YEAR_FIELD };
149static const int8_t kDateFieldsCount = 16;
150
151static const UDateFormatField kTimeFields[] = {
152 UDAT_HOUR_OF_DAY1_FIELD,
153 UDAT_HOUR_OF_DAY0_FIELD,
154 UDAT_MINUTE_FIELD,
155 UDAT_SECOND_FIELD,
156 UDAT_FRACTIONAL_SECOND_FIELD,
157 UDAT_HOUR1_FIELD,
158 UDAT_HOUR0_FIELD,
159 UDAT_MILLISECONDS_IN_DAY_FIELD,
160 UDAT_TIMEZONE_RFC_FIELD,
161 UDAT_TIMEZONE_LOCALIZED_GMT_OFFSET_FIELD };
162static const int8_t kTimeFieldsCount = 10;
163
164
165// This is a pattern-of-last-resort used when we can't load a usable pattern out
166// of a resource.
167static const char16_t gDefaultPattern[] =
168{
169 0x79, 0x4D, 0x4D, 0x64, 0x64, 0x20, 0x68, 0x68, 0x3A, 0x6D, 0x6D, 0x20, 0x61, 0
170}; /* "yMMdd hh:mm a" */
171
172// This prefix is designed to NEVER MATCH real text, in order to
173// suppress the parsing of negative numbers. Adjust as needed (if
174// this becomes valid Unicode).
175static const char16_t SUPPRESS_NEGATIVE_PREFIX[] = {0xAB00, 0};
176
177/**
178 * These are the tags we expect to see in normal resource bundle files associated
179 * with a locale.
180 */
181static const char16_t QUOTE = 0x27; // Single quote
182
183/*
184 * The field range check bias for each UDateFormatField.
185 * The bias is added to the minimum and maximum values
186 * before they are compared to the parsed number.
187 * For example, the calendar stores zero-based month numbers
188 * but the parsed month numbers start at 1, so the bias is 1.
189 *
190 * A value of -1 means that the value is not checked.
191 */
192static const int32_t gFieldRangeBias[] = {
193 -1, // 'G' - UDAT_ERA_FIELD
194 -1, // 'y' - UDAT_YEAR_FIELD
195 1, // 'M' - UDAT_MONTH_FIELD
196 0, // 'd' - UDAT_DATE_FIELD
197 -1, // 'k' - UDAT_HOUR_OF_DAY1_FIELD
198 -1, // 'H' - UDAT_HOUR_OF_DAY0_FIELD
199 0, // 'm' - UDAT_MINUTE_FIELD
200 0, // 's' - UDAT_SECOND_FIELD
201 -1, // 'S' - UDAT_FRACTIONAL_SECOND_FIELD (0-999?)
202 -1, // 'E' - UDAT_DAY_OF_WEEK_FIELD (1-7?)
203 -1, // 'D' - UDAT_DAY_OF_YEAR_FIELD (1 - 366?)
204 -1, // 'F' - UDAT_DAY_OF_WEEK_IN_MONTH_FIELD (1-5?)
205 -1, // 'w' - UDAT_WEEK_OF_YEAR_FIELD (1-52?)
206 -1, // 'W' - UDAT_WEEK_OF_MONTH_FIELD (1-5?)
207 -1, // 'a' - UDAT_AM_PM_FIELD
208 -1, // 'h' - UDAT_HOUR1_FIELD
209 -1, // 'K' - UDAT_HOUR0_FIELD
210 -1, // 'z' - UDAT_TIMEZONE_FIELD
211 -1, // 'Y' - UDAT_YEAR_WOY_FIELD
212 -1, // 'e' - UDAT_DOW_LOCAL_FIELD
213 -1, // 'u' - UDAT_EXTENDED_YEAR_FIELD
214 -1, // 'g' - UDAT_JULIAN_DAY_FIELD
215 -1, // 'A' - UDAT_MILLISECONDS_IN_DAY_FIELD
216 -1, // 'Z' - UDAT_TIMEZONE_RFC_FIELD
217 -1, // 'v' - UDAT_TIMEZONE_GENERIC_FIELD
218 0, // 'c' - UDAT_STANDALONE_DAY_FIELD
219 1, // 'L' - UDAT_STANDALONE_MONTH_FIELD
220 -1, // 'Q' - UDAT_QUARTER_FIELD (1-4?)
221 -1, // 'q' - UDAT_STANDALONE_QUARTER_FIELD
222 -1, // 'V' - UDAT_TIMEZONE_SPECIAL_FIELD
223 -1, // 'U' - UDAT_YEAR_NAME_FIELD
224 -1, // 'O' - UDAT_TIMEZONE_LOCALIZED_GMT_OFFSET_FIELD
225 -1, // 'X' - UDAT_TIMEZONE_ISO_FIELD
226 -1, // 'x' - UDAT_TIMEZONE_ISO_LOCAL_FIELD
227 -1, // 'r' - UDAT_RELATED_YEAR_FIELD
228#if UDAT_HAS_PATTERN_CHAR_FOR_TIME_SEPARATOR0
229 -1, // ':' - UDAT_TIME_SEPARATOR_FIELD
230#else
231 -1, // (no pattern character currently) - UDAT_TIME_SEPARATOR_FIELD
232#endif
233};
234
235// When calendar uses hebr numbering (i.e. he@calendar=hebrew),
236// offset the years within the current millennium down to 1-999
237static const int32_t HEBREW_CAL_CUR_MILLENIUM_START_YEAR = 5000;
238static const int32_t HEBREW_CAL_CUR_MILLENIUM_END_YEAR = 6000;
239
240/**
241 * Maximum range for detecting daylight offset of a time zone when parsed time zone
242 * string indicates it's daylight saving time, but the detected time zone does not
243 * observe daylight saving time at the parsed date.
244 */
245static const double MAX_DAYLIGHT_DETECTION_RANGE = 30*365*24*60*60*1000.0;
246
247static UMutex LOCK;
248
249UOBJECT_DEFINE_RTTI_IMPLEMENTATION(SimpleDateFormat)UClassID SimpleDateFormat::getStaticClassID() { static char classID
= 0; return (UClassID)&classID; } UClassID SimpleDateFormat
::getDynamicClassID() const { return SimpleDateFormat::getStaticClassID
(); }
250
251SimpleDateFormat::NSOverride::~NSOverride() {
252 if (snf != nullptr) {
253 snf->removeRef();
254 }
255}
256
257
258void SimpleDateFormat::NSOverride::free() {
259 NSOverride *cur = this;
260 while (cur) {
261 NSOverride *next_temp = cur->next;
262 delete cur;
263 cur = next_temp;
264 }
265}
266
267// no matter what the locale's default number format looked like, we want
268// to modify it so that it doesn't use thousands separators, doesn't always
269// show the decimal point, and recognizes integers only when parsing
270static void fixNumberFormatForDates(NumberFormat &nf) {
271 nf.setGroupingUsed(false);
272 DecimalFormat* decfmt = dynamic_cast<DecimalFormat*>(&nf);
273 if (decfmt != nullptr) {
274 decfmt->setDecimalSeparatorAlwaysShown(false);
275 }
276 nf.setParseIntegerOnly(true);
277 nf.setMinimumFractionDigits(0); // To prevent "Jan 1.00, 1997.00"
278}
279
280static const SharedNumberFormat *createSharedNumberFormat(
281 NumberFormat *nfToAdopt) {
282 fixNumberFormatForDates(*nfToAdopt);
283 const SharedNumberFormat *result = new SharedNumberFormat(nfToAdopt);
284 if (result == nullptr) {
285 delete nfToAdopt;
286 }
287 return result;
288}
289
290static const SharedNumberFormat *createSharedNumberFormat(
291 const Locale &loc, UErrorCode &status) {
292 NumberFormat *nf = NumberFormat::createInstance(loc, status);
293 if (U_FAILURE(status)) {
294 return nullptr;
295 }
296 const SharedNumberFormat *result = createSharedNumberFormat(nf);
297 if (result == nullptr) {
298 status = U_MEMORY_ALLOCATION_ERROR;
299 }
300 return result;
301}
302
303static const SharedNumberFormat **allocSharedNumberFormatters() {
304 const SharedNumberFormat** result = static_cast<const SharedNumberFormat**>(
305 uprv_mallocuprv_malloc_77(UDAT_FIELD_COUNT * sizeof(const SharedNumberFormat*)));
306 if (result == nullptr) {
307 return nullptr;
308 }
309 for (int32_t i = 0; i < UDAT_FIELD_COUNT; ++i) {
310 result[i] = nullptr;
311 }
312 return result;
313}
314
315static void freeSharedNumberFormatters(const SharedNumberFormat ** list) {
316 for (int32_t i = 0; i < UDAT_FIELD_COUNT; ++i) {
317 SharedObject::clearPtr(list[i]);
318 }
319 uprv_freeuprv_free_77(list);
320}
321
322const NumberFormat *SimpleDateFormat::getNumberFormatByIndex(
323 UDateFormatField index) const {
324 if (fSharedNumberFormatters == nullptr ||
325 fSharedNumberFormatters[index] == nullptr) {
326 return fNumberFormat;
327 }
328 return &(**fSharedNumberFormatters[index]);
329}
330
331//----------------------------------------------------------------------
332
333SimpleDateFormat::~SimpleDateFormat()
334{
335 delete fSymbols;
336 if (fSharedNumberFormatters) {
337 freeSharedNumberFormatters(fSharedNumberFormatters);
338 }
339 delete fTimeZoneFormat;
340 delete fSimpleNumberFormatter;
341
342#if !UCONFIG_NO_BREAK_ITERATION1
343 delete fCapitalizationBrkIter;
344#endif
345}
346
347//----------------------------------------------------------------------
348
349SimpleDateFormat::SimpleDateFormat(UErrorCode& status)
350 : fLocale(Locale::getDefault())
351{
352 initializeBooleanAttributes();
353 construct(kShort, static_cast<EStyle>(kShort + kDateOffset), fLocale, status);
354 initializeDefaultCentury();
355}
356
357//----------------------------------------------------------------------
358
359SimpleDateFormat::SimpleDateFormat(const UnicodeString& pattern,
360 UErrorCode &status)
361: fPattern(pattern),
362 fLocale(Locale::getDefault())
363{
364 fDateOverride.setToBogus();
365 fTimeOverride.setToBogus();
366 initializeBooleanAttributes();
367 initializeCalendar(nullptr,fLocale,status);
368 fSymbols = DateFormatSymbols::createForLocale(fLocale, status);
369 initialize(fLocale, status);
370 initializeDefaultCentury();
371
372}
373//----------------------------------------------------------------------
374
375SimpleDateFormat::SimpleDateFormat(const UnicodeString& pattern,
376 const UnicodeString& override,
377 UErrorCode &status)
378: fPattern(pattern),
379 fLocale(Locale::getDefault())
380{
381 fDateOverride.setTo(override);
382 fTimeOverride.setToBogus();
383 initializeBooleanAttributes();
384 initializeCalendar(nullptr,fLocale,status);
385 fSymbols = DateFormatSymbols::createForLocale(fLocale, status);
386 initialize(fLocale, status);
387 initializeDefaultCentury();
388
389 processOverrideString(fLocale,override,kOvrStrBoth,status);
390
391}
392
393//----------------------------------------------------------------------
394
395SimpleDateFormat::SimpleDateFormat(const UnicodeString& pattern,
396 const Locale& locale,
397 UErrorCode& status)
398: fPattern(pattern),
399 fLocale(locale)
400{
401
402 fDateOverride.setToBogus();
403 fTimeOverride.setToBogus();
404 initializeBooleanAttributes();
405
406 initializeCalendar(nullptr,fLocale,status);
407 fSymbols = DateFormatSymbols::createForLocale(fLocale, status);
408 initialize(fLocale, status);
409 initializeDefaultCentury();
410}
411
412//----------------------------------------------------------------------
413
414SimpleDateFormat::SimpleDateFormat(const UnicodeString& pattern,
415 const UnicodeString& override,
416 const Locale& locale,
417 UErrorCode& status)
418: fPattern(pattern),
419 fLocale(locale)
420{
421
422 fDateOverride.setTo(override);
423 fTimeOverride.setToBogus();
424 initializeBooleanAttributes();
425
426 initializeCalendar(nullptr,fLocale,status);
427 fSymbols = DateFormatSymbols::createForLocale(fLocale, status);
428 initialize(fLocale, status);
429 initializeDefaultCentury();
430
431 processOverrideString(locale,override,kOvrStrBoth,status);
432
433}
434
435//----------------------------------------------------------------------
436
437SimpleDateFormat::SimpleDateFormat(const UnicodeString& pattern,
438 DateFormatSymbols* symbolsToAdopt,
439 UErrorCode& status)
440: fPattern(pattern),
441 fLocale(Locale::getDefault()),
442 fSymbols(symbolsToAdopt)
443{
444
445 fDateOverride.setToBogus();
446 fTimeOverride.setToBogus();
447 initializeBooleanAttributes();
448
449 initializeCalendar(nullptr,fLocale,status);
450 initialize(fLocale, status);
451 initializeDefaultCentury();
452}
453
454//----------------------------------------------------------------------
455
456SimpleDateFormat::SimpleDateFormat(const UnicodeString& pattern,
457 const DateFormatSymbols& symbols,
458 UErrorCode& status)
459: fPattern(pattern),
460 fLocale(Locale::getDefault()),
461 fSymbols(new DateFormatSymbols(symbols))
462{
463
464 fDateOverride.setToBogus();
465 fTimeOverride.setToBogus();
466 initializeBooleanAttributes();
467
468 initializeCalendar(nullptr, fLocale, status);
469 initialize(fLocale, status);
470 initializeDefaultCentury();
471}
472
473//----------------------------------------------------------------------
474
475// Not for public consumption; used by DateFormat
476SimpleDateFormat::SimpleDateFormat(EStyle timeStyle,
477 EStyle dateStyle,
478 const Locale& locale,
479 UErrorCode& status)
480: fLocale(locale)
481{
482 initializeBooleanAttributes();
483 construct(timeStyle, dateStyle, fLocale, status);
484 if(U_SUCCESS(status)) {
485 initializeDefaultCentury();
486 }
487}
488
489//----------------------------------------------------------------------
490
491/**
492 * Not for public consumption; used by DateFormat. This constructor
493 * never fails. If the resource data is not available, it uses the
494 * the last resort symbols.
495 */
496SimpleDateFormat::SimpleDateFormat(const Locale& locale,
497 UErrorCode& status)
498: fPattern(gDefaultPattern),
499 fLocale(locale)
500{
501 if (U_FAILURE(status)) return;
502 initializeBooleanAttributes();
503 initializeCalendar(nullptr, fLocale, status);
504 fSymbols = DateFormatSymbols::createForLocale(fLocale, status);
505 if (U_FAILURE(status))
506 {
507 status = U_ZERO_ERROR;
508 delete fSymbols;
509 // This constructor doesn't fail; it uses last resort data
510 fSymbols = new DateFormatSymbols(status);
511 /* test for nullptr */
512 if (fSymbols == nullptr) {
513 status = U_MEMORY_ALLOCATION_ERROR;
514 return;
515 }
516 }
517
518 fDateOverride.setToBogus();
519 fTimeOverride.setToBogus();
520
521 initialize(fLocale, status);
522 if(U_SUCCESS(status)) {
523 initializeDefaultCentury();
524 }
525}
526
527//----------------------------------------------------------------------
528
529SimpleDateFormat::SimpleDateFormat(const SimpleDateFormat& other)
530: DateFormat(other),
531 fLocale(other.fLocale)
532{
533 initializeBooleanAttributes();
534 *this = other;
535}
536
537//----------------------------------------------------------------------
538
539SimpleDateFormat& SimpleDateFormat::operator=(const SimpleDateFormat& other)
540{
541 if (this == &other) {
542 return *this;
543 }
544
545 // fSimpleNumberFormatter references fNumberFormatter, delete it
546 // before we call the = operator which may invalidate fNumberFormatter
547 delete fSimpleNumberFormatter;
548 fSimpleNumberFormatter = nullptr;
549
550 DateFormat::operator=(other);
551 fDateOverride = other.fDateOverride;
552 fTimeOverride = other.fTimeOverride;
553
554 delete fSymbols;
555 fSymbols = nullptr;
556
557 if (other.fSymbols)
558 fSymbols = new DateFormatSymbols(*other.fSymbols);
559
560 fDefaultCenturyStart = other.fDefaultCenturyStart;
561 fDefaultCenturyStartYear = other.fDefaultCenturyStartYear;
562 fHaveDefaultCentury = other.fHaveDefaultCentury;
563
564 fPattern = other.fPattern;
565 fHasMinute = other.fHasMinute;
566 fHasSecond = other.fHasSecond;
567
568 fLocale = other.fLocale;
569
570 // TimeZoneFormat can now be set independently via setter.
571 // If it is nullptr, it will be lazily initialized from locale.
572 delete fTimeZoneFormat;
573 fTimeZoneFormat = nullptr;
574 TimeZoneFormat *otherTZFormat;
575 {
576 // Synchronization is required here, when accessing other.fTimeZoneFormat,
577 // because another thread may be concurrently executing other.tzFormat(),
578 // a logically const function that lazily creates other.fTimeZoneFormat.
579 //
580 // Without synchronization, reordered memory writes could allow us
581 // to see a non-null fTimeZoneFormat before the object itself was
582 // fully initialized. In case of a race, it doesn't matter whether
583 // we see a null or a fully initialized other.fTimeZoneFormat,
584 // only that we avoid seeing a partially initialized object.
585 //
586 // Once initialized, no const function can modify fTimeZoneFormat,
587 // meaning that once we have safely grabbed the other.fTimeZoneFormat
588 // pointer, continued synchronization is not required to use it.
589 Mutex m(&LOCK);
590 otherTZFormat = other.fTimeZoneFormat;
591 }
592 if (otherTZFormat) {
593 fTimeZoneFormat = new TimeZoneFormat(*otherTZFormat);
594 }
595
596#if !UCONFIG_NO_BREAK_ITERATION1
597 if (other.fCapitalizationBrkIter != nullptr) {
598 fCapitalizationBrkIter = (other.fCapitalizationBrkIter)->clone();
599 }
600#endif
601
602 if (fSharedNumberFormatters != nullptr) {
603 freeSharedNumberFormatters(fSharedNumberFormatters);
604 fSharedNumberFormatters = nullptr;
605 }
606 if (other.fSharedNumberFormatters != nullptr) {
607 fSharedNumberFormatters = allocSharedNumberFormatters();
608 if (fSharedNumberFormatters) {
609 for (int32_t i = 0; i < UDAT_FIELD_COUNT; ++i) {
610 SharedObject::copyPtr(
611 other.fSharedNumberFormatters[i],
612 fSharedNumberFormatters[i]);
613 }
614 }
615 }
616
617 UErrorCode localStatus = U_ZERO_ERROR;
618 // SimpleNumberFormatter does not have a copy constructor. Furthermore,
619 // it references data from an internal field, fNumberFormatter,
620 // so we must rematerialize that reference after copying over the number formatter.
621 initSimpleNumberFormatter(localStatus);
622 return *this;
623}
624
625//----------------------------------------------------------------------
626
627SimpleDateFormat*
628SimpleDateFormat::clone() const
629{
630 return new SimpleDateFormat(*this);
631}
632
633//----------------------------------------------------------------------
634
635bool
636SimpleDateFormat::operator==(const Format& other) const
637{
638 if (DateFormat::operator==(other)) {
639 // The DateFormat::operator== check for fCapitalizationContext equality above
640 // is sufficient to check equality of all derived context-related data.
641 // DateFormat::operator== guarantees following cast is safe
642 SimpleDateFormat* that = (SimpleDateFormat*)&other;
643 return (fPattern == that->fPattern &&
644 fSymbols != nullptr && // Check for pathological object
645 that->fSymbols != nullptr && // Check for pathological object
646 *fSymbols == *that->fSymbols &&
647 fHaveDefaultCentury == that->fHaveDefaultCentury &&
648 fDefaultCenturyStart == that->fDefaultCenturyStart);
649 }
650 return false;
651}
652
653//----------------------------------------------------------------------
654static const char16_t* timeSkeletons[4] = {
655 u"jmmsszzzz", // kFull
656 u"jmmssz", // kLong
657 u"jmmss", // kMedium
658 u"jmm", // kShort
659};
660
661void SimpleDateFormat::construct(EStyle timeStyle,
662 EStyle dateStyle,
663 const Locale& locale,
664 UErrorCode& status)
665{
666 // called by several constructors to load pattern data from the resources
667 if (U_FAILURE(status)) return;
668
669 // We will need the calendar to know what type of symbols to load.
670 initializeCalendar(nullptr, locale, status);
671 if (U_FAILURE(status)) return;
672
673 // Load date time patterns directly from resources.
674 const char* cType = fCalendar ? fCalendar->getType() : nullptr;
675 LocalUResourceBundlePointer bundle(ures_openures_open_77(nullptr, locale.getBaseName(), &status));
676 if (U_FAILURE(status)) return;
677
678 UBool cTypeIsGregorian = true;
679 LocalUResourceBundlePointer dateTimePatterns;
680 if (cType != nullptr && uprv_strcmp(cType, "gregorian"):: strcmp(cType, "gregorian") != 0) {
681 CharString resourcePath("calendar/", status);
682 resourcePath.append(cType, status).append("/DateTimePatterns", status);
683 dateTimePatterns.adoptInstead(
684 ures_getByKeyWithFallbackures_getByKeyWithFallback_77(bundle.getAlias(), resourcePath.data(),
685 (UResourceBundle*)nullptr, &status));
686 cTypeIsGregorian = false;
687 }
688
689 // Check for "gregorian" fallback.
690 if (cTypeIsGregorian || status == U_MISSING_RESOURCE_ERROR) {
691 status = U_ZERO_ERROR;
692 dateTimePatterns.adoptInstead(
693 ures_getByKeyWithFallbackures_getByKeyWithFallback_77(bundle.getAlias(),
694 "calendar/gregorian/DateTimePatterns",
695 (UResourceBundle*)nullptr, &status));
696 }
697 if (U_FAILURE(status)) return;
698
699 LocalUResourceBundlePointer currentBundle;
700
701 if (ures_getSizeures_getSize_77(dateTimePatterns.getAlias()) <= kDateTime)
702 {
703 status = U_INVALID_FORMAT_ERROR;
704 return;
705 }
706
707 setLocaleIDs(ures_getLocaleByTypeures_getLocaleByType_77(dateTimePatterns.getAlias(), ULOC_VALID_LOCALE, &status),
708 ures_getLocaleByTypeures_getLocaleByType_77(dateTimePatterns.getAlias(), ULOC_ACTUAL_LOCALE, &status));
709
710 // create a symbols object from the locale
711 fSymbols = DateFormatSymbols::createForLocale(locale, status);
712 if (U_FAILURE(status)) return;
713 /* test for nullptr */
714 if (fSymbols == nullptr) {
715 status = U_MEMORY_ALLOCATION_ERROR;
716 return;
717 }
718
719 const char16_t *resStr,*ovrStr;
720 int32_t resStrLen,ovrStrLen = 0;
721 fDateOverride.setToBogus();
722 fTimeOverride.setToBogus();
723
724 UnicodeString timePattern;
725 if (timeStyle >= kFull && timeStyle <= kShort) {
726 bool hasRgOrHcSubtag = false;
727 // also use DTPG if the locale has the "rg" or "hc" ("hours") subtag-- even if the overriding region
728 // or hour cycle is the same as the one we get by default, we go through the DateTimePatternGenerator
729 UErrorCode dummyErr1 = U_ZERO_ERROR, dummyErr2 = U_ZERO_ERROR;
730 if (locale.getKeywordValue("rg", nullptr, 0, dummyErr1) > 0 || locale.getKeywordValue("hours", nullptr, 0, dummyErr2) > 0) {
731 hasRgOrHcSubtag = true;
732 }
733
734 const char* baseLocID = locale.getBaseName();
735 if (baseLocID != nullptr && uprv_strcmp(baseLocID,"und"):: strcmp(baseLocID, "und")!=0) {
736 UErrorCode useStatus = U_ZERO_ERROR;
737 Locale baseLoc(baseLocID);
738 Locale validLoc(getLocale(ULOC_VALID_LOCALE, useStatus));
739 if (hasRgOrHcSubtag || (U_SUCCESS(useStatus) && validLoc!=baseLoc)) {
740 bool useDTPG = hasRgOrHcSubtag;
741 const char* baseReg = baseLoc.getCountry(); // empty string if no region
742 if ((baseReg != nullptr && baseReg[0] != 0 &&
743 uprv_strncmp(baseReg,validLoc.getCountry(),ULOC_COUNTRY_CAPACITY):: strncmp(baseReg, validLoc.getCountry(), 4)!=0)
744 || uprv_strncmp(baseLoc.getLanguage(),validLoc.getLanguage(),ULOC_LANG_CAPACITY):: strncmp(baseLoc.getLanguage(), validLoc.getLanguage(), 12)!=0) {
745 // use DTPG if
746 // * baseLoc has a region and validLoc does not have the same one (or has none), OR
747 // * validLoc has a different language code than baseLoc
748 // * the original locale has the rg or hc subtag
749 useDTPG = true;
750 }
751 if (useDTPG) {
752 // The standard time formats may have the wrong time cycle, because:
753 // the valid locale differs in important ways (region, language) from
754 // the base locale.
755 // We could *also* check whether they do actually have a mismatch with
756 // the time cycle preferences for the region, but that is a lot more
757 // work for little or no additional benefit, since just going ahead
758 // and always synthesizing the time format as per the following should
759 // create a locale-appropriate pattern with cycle that matches the
760 // region preferences anyway.
761 LocalPointer<DateTimePatternGenerator> dtpg(DateTimePatternGenerator::createInstanceNoStdPat(locale, useStatus));
762 if (U_SUCCESS(useStatus)) {
763 UnicodeString timeSkeleton(true, timeSkeletons[timeStyle], -1);
764 timePattern = dtpg->getBestPattern(timeSkeleton, useStatus);
765 }
766 }
767 }
768 }
769 }
770
771 // if the pattern should include both date and time information, use the date/time
772 // pattern string as a guide to tell use how to glue together the appropriate date
773 // and time pattern strings.
774 if ((timeStyle != kNone) && (dateStyle != kNone))
775 {
776 UnicodeString tempus1(timePattern);
777 if (tempus1.length() == 0) {
778 currentBundle.adoptInstead(
779 ures_getByIndexures_getByIndex_77(dateTimePatterns.getAlias(), static_cast<int32_t>(timeStyle), nullptr, &status));
780 if (U_FAILURE(status)) {
781 status = U_INVALID_FORMAT_ERROR;
782 return;
783 }
784 switch (ures_getTypeures_getType_77(currentBundle.getAlias())) {
785 case URES_STRING: {
786 resStr = ures_getStringures_getString_77(currentBundle.getAlias(), &resStrLen, &status);
787 break;
788 }
789 case URES_ARRAY: {
790 resStr = ures_getStringByIndexures_getStringByIndex_77(currentBundle.getAlias(), 0, &resStrLen, &status);
791 ovrStr = ures_getStringByIndexures_getStringByIndex_77(currentBundle.getAlias(), 1, &ovrStrLen, &status);
792 fTimeOverride.setTo(true, ovrStr, ovrStrLen);
793 break;
794 }
795 default: {
796 status = U_INVALID_FORMAT_ERROR;
797 return;
798 }
799 }
800
801 tempus1.setTo(true, resStr, resStrLen);
802 }
803
804 currentBundle.adoptInstead(
805 ures_getByIndexures_getByIndex_77(dateTimePatterns.getAlias(), static_cast<int32_t>(dateStyle), nullptr, &status));
806 if (U_FAILURE(status)) {
807 status = U_INVALID_FORMAT_ERROR;
808 return;
809 }
810 switch (ures_getTypeures_getType_77(currentBundle.getAlias())) {
811 case URES_STRING: {
812 resStr = ures_getStringures_getString_77(currentBundle.getAlias(), &resStrLen, &status);
813 break;
814 }
815 case URES_ARRAY: {
816 resStr = ures_getStringByIndexures_getStringByIndex_77(currentBundle.getAlias(), 0, &resStrLen, &status);
817 ovrStr = ures_getStringByIndexures_getStringByIndex_77(currentBundle.getAlias(), 1, &ovrStrLen, &status);
818 fDateOverride.setTo(true, ovrStr, ovrStrLen);
819 break;
820 }
821 default: {
822 status = U_INVALID_FORMAT_ERROR;
823 return;
824 }
825 }
826
827 UnicodeString tempus2(true, resStr, resStrLen);
828
829 // Currently, for compatibility with pre-CLDR-42 data, we default to the "atTime"
830 // combining patterns. Depending on guidance in CLDR 42 spec and on DisplayOptions,
831 // we may change this.
832 LocalUResourceBundlePointer dateAtTimePatterns;
833 if (!cTypeIsGregorian) {
834 CharString resourcePath("calendar/", status);
835 resourcePath.append(cType, status).append("/DateTimePatterns%atTime", status);
836 dateAtTimePatterns.adoptInstead(
837 ures_getByKeyWithFallbackures_getByKeyWithFallback_77(bundle.getAlias(), resourcePath.data(),
838 nullptr, &status));
839 }
840 if (cTypeIsGregorian || status == U_MISSING_RESOURCE_ERROR) {
841 status = U_ZERO_ERROR;
842 dateAtTimePatterns.adoptInstead(
843 ures_getByKeyWithFallbackures_getByKeyWithFallback_77(bundle.getAlias(),
844 "calendar/gregorian/DateTimePatterns%atTime",
845 nullptr, &status));
846 }
847 if (U_SUCCESS(status) && ures_getSizeures_getSize_77(dateAtTimePatterns.getAlias()) >= 4) {
848 resStr = ures_getStringByIndexures_getStringByIndex_77(dateAtTimePatterns.getAlias(), dateStyle - kDateOffset, &resStrLen, &status);
849 } else {
850 status = U_ZERO_ERROR;
851 int32_t glueIndex = kDateTime;
852 int32_t patternsSize = ures_getSizeures_getSize_77(dateTimePatterns.getAlias());
853 if (patternsSize >= (kDateTimeOffset + kShort + 1)) {
854 // Get proper date time format
855 glueIndex = static_cast<int32_t>(kDateTimeOffset + (dateStyle - kDateOffset));
856 }
857
858 resStr = ures_getStringByIndexures_getStringByIndex_77(dateTimePatterns.getAlias(), glueIndex, &resStrLen, &status);
859 }
860 SimpleFormatter(UnicodeString(true, resStr, resStrLen), 2, 2, status).
861 format(tempus1, tempus2, fPattern, status);
862 }
863 // if the pattern includes just time data or just date date, load the appropriate
864 // pattern string from the resources
865 // setTo() - see DateFormatSymbols::assignArray comments
866 else if (timeStyle != kNone) {
867 fPattern.setTo(timePattern);
868 if (fPattern.length() == 0) {
869 currentBundle.adoptInstead(
870 ures_getByIndexures_getByIndex_77(dateTimePatterns.getAlias(), static_cast<int32_t>(timeStyle), nullptr, &status));
871 if (U_FAILURE(status)) {
872 status = U_INVALID_FORMAT_ERROR;
873 return;
874 }
875 switch (ures_getTypeures_getType_77(currentBundle.getAlias())) {
876 case URES_STRING: {
877 resStr = ures_getStringures_getString_77(currentBundle.getAlias(), &resStrLen, &status);
878 break;
879 }
880 case URES_ARRAY: {
881 resStr = ures_getStringByIndexures_getStringByIndex_77(currentBundle.getAlias(), 0, &resStrLen, &status);
882 ovrStr = ures_getStringByIndexures_getStringByIndex_77(currentBundle.getAlias(), 1, &ovrStrLen, &status);
883 fDateOverride.setTo(true, ovrStr, ovrStrLen);
884 break;
885 }
886 default: {
887 status = U_INVALID_FORMAT_ERROR;
888 return;
889 }
890 }
891 fPattern.setTo(true, resStr, resStrLen);
892 }
893 }
894 else if (dateStyle != kNone) {
895 currentBundle.adoptInstead(
896 ures_getByIndexures_getByIndex_77(dateTimePatterns.getAlias(), static_cast<int32_t>(dateStyle), nullptr, &status));
897 if (U_FAILURE(status)) {
898 status = U_INVALID_FORMAT_ERROR;
899 return;
900 }
901 switch (ures_getTypeures_getType_77(currentBundle.getAlias())) {
902 case URES_STRING: {
903 resStr = ures_getStringures_getString_77(currentBundle.getAlias(), &resStrLen, &status);
904 break;
905 }
906 case URES_ARRAY: {
907 resStr = ures_getStringByIndexures_getStringByIndex_77(currentBundle.getAlias(), 0, &resStrLen, &status);
908 ovrStr = ures_getStringByIndexures_getStringByIndex_77(currentBundle.getAlias(), 1, &ovrStrLen, &status);
909 fDateOverride.setTo(true, ovrStr, ovrStrLen);
910 break;
911 }
912 default: {
913 status = U_INVALID_FORMAT_ERROR;
914 return;
915 }
916 }
917 fPattern.setTo(true, resStr, resStrLen);
918 }
919
920 // and if it includes _neither_, that's an error
921 else
922 status = U_INVALID_FORMAT_ERROR;
923
924 // finally, finish initializing by creating a Calendar and a NumberFormat
925 initialize(locale, status);
926}
927
928//----------------------------------------------------------------------
929
930Calendar*
931SimpleDateFormat::initializeCalendar(TimeZone* adoptZone, const Locale& locale, UErrorCode& status)
932{
933 if(!U_FAILURE(status)) {
934 fCalendar = Calendar::createInstance(
935 adoptZone ? adoptZone : TimeZone::forLocaleOrDefault(locale), locale, status);
936 }
937 return fCalendar;
938}
939
940void
941SimpleDateFormat::initialize(const Locale& locale,
942 UErrorCode& status)
943{
944 if (U_FAILURE(status)) return;
945
946 parsePattern(); // Need this before initNumberFormatters(), to set fHasHanYearChar
947
948 // Simple-minded hack to force Gannen year numbering for ja@calendar=japanese
949 // if format is non-numeric (includes 年) and fDateOverride is not already specified.
950 // Now this does get updated if applyPattern subsequently changes the pattern type.
951 if (fDateOverride.isBogus() && fHasHanYearChar &&
952 fCalendar != nullptr &&
953 typeid(*fCalendar) == typeid(JapaneseCalendar) &&
954 uprv_strcmp(fLocale.getLanguage(),"ja"):: strcmp(fLocale.getLanguage(), "ja") == 0) {
955 fDateOverride.setTo(u"y=jpanyear", -1);
956 }
957
958 // We don't need to check that the row count is >= 1, since all 2d arrays have at
959 // least one row
960 fNumberFormat = NumberFormat::createInstance(locale, status);
961 if (fNumberFormat != nullptr && U_SUCCESS(status))
962 {
963 fixNumberFormatForDates(*fNumberFormat);
964 //fNumberFormat->setLenient(true); // Java uses a custom DateNumberFormat to format/parse
965
966 initNumberFormatters(locale, status);
967 initSimpleNumberFormatter(status);
968
969 }
970 else if (U_SUCCESS(status))
971 {
972 status = U_MISSING_RESOURCE_ERROR;
973 }
974}
975
976/* Initialize the fields we use to disambiguate ambiguous years. Separate
977 * so we can call it from readObject().
978 */
979void SimpleDateFormat::initializeDefaultCentury()
980{
981 if(fCalendar) {
982 fHaveDefaultCentury = fCalendar->haveDefaultCentury();
983 if(fHaveDefaultCentury) {
984 fDefaultCenturyStart = fCalendar->defaultCenturyStart();
985 fDefaultCenturyStartYear = fCalendar->defaultCenturyStartYear();
986 } else {
987 fDefaultCenturyStart = DBL_MIN2.2250738585072014e-308;
988 fDefaultCenturyStartYear = -1;
989 }
990 }
991}
992
993/*
994 * Initialize the boolean attributes. Separate so we can call it from all constructors.
995 */
996void SimpleDateFormat::initializeBooleanAttributes()
997{
998 UErrorCode status = U_ZERO_ERROR;
999
1000 setBooleanAttribute(UDAT_PARSE_ALLOW_WHITESPACE, true, status);
1001 setBooleanAttribute(UDAT_PARSE_ALLOW_NUMERIC, true, status);
1002 setBooleanAttribute(UDAT_PARSE_PARTIAL_LITERAL_MATCH, true, status);
1003 setBooleanAttribute(UDAT_PARSE_MULTIPLE_PATTERNS_FOR_MATCH, true, status);
1004}
1005
1006/* Define one-century window into which to disambiguate dates using
1007 * two-digit years. Make public in JDK 1.2.
1008 */
1009void SimpleDateFormat::parseAmbiguousDatesAsAfter(UDate startDate, UErrorCode& status)
1010{
1011 if(U_FAILURE(status)) {
1012 return;
1013 }
1014 if(!fCalendar) {
1015 status = U_ILLEGAL_ARGUMENT_ERROR;
1016 return;
1017 }
1018
1019 fCalendar->setTime(startDate, status);
1020 if(U_SUCCESS(status)) {
1021 fHaveDefaultCentury = true;
1022 fDefaultCenturyStart = startDate;
1023 fDefaultCenturyStartYear = fCalendar->get(UCAL_YEAR, status);
1024 }
1025}
1026
1027//----------------------------------------------------------------------
1028
1029UnicodeString&
1030SimpleDateFormat::format(Calendar& cal, UnicodeString& appendTo, FieldPosition& pos) const
1031{
1032 UErrorCode status = U_ZERO_ERROR;
1033 FieldPositionOnlyHandler handler(pos);
1034 return _format(cal, appendTo, handler, status);
1035}
1036
1037//----------------------------------------------------------------------
1038
1039UnicodeString&
1040SimpleDateFormat::format(Calendar& cal, UnicodeString& appendTo,
1041 FieldPositionIterator* posIter, UErrorCode& status) const
1042{
1043 FieldPositionIteratorHandler handler(posIter, status);
1044 return _format(cal, appendTo, handler, status);
1045}
1046
1047//----------------------------------------------------------------------
1048
1049UnicodeString&
1050SimpleDateFormat::_format(Calendar& cal, UnicodeString& appendTo,
1051 FieldPositionHandler& handler, UErrorCode& status) const
1052{
1053 if ( U_FAILURE(status) ) {
1054 return appendTo;
1055 }
1056 Calendar* workCal = &cal;
1057 Calendar* calClone = nullptr;
1058 if (&cal != fCalendar && typeid(cal) != typeid(*fCalendar)) {
1059 // Different calendar type
1060 // We use the time and time zone from the input calendar, but
1061 // do not use the input calendar for field calculation.
1062 calClone = fCalendar->clone();
1063 if (calClone != nullptr) {
1064 UDate t = cal.getTime(status);
1065 calClone->setTime(t, status);
1066 calClone->setTimeZone(cal.getTimeZone());
1067 workCal = calClone;
1068 } else {
1069 status = U_MEMORY_ALLOCATION_ERROR;
1070 return appendTo;
1071 }
1072 }
1073
1074 UBool inQuote = false;
1075 char16_t prevCh = 0;
1076 int32_t count = 0;
1077 int32_t fieldNum = 0;
1078 UDisplayContext capitalizationContext = getContext(UDISPCTX_TYPE_CAPITALIZATION, status);
1079
1080 // loop through the pattern string character by character
1081 int32_t patternLength = fPattern.length();
1082 for (int32_t i = 0; i < patternLength && U_SUCCESS(status); ++i) {
1083 char16_t ch = fPattern[i];
1084
1085 // Use subFormat() to format a repeated pattern character
1086 // when a different pattern or non-pattern character is seen
1087 if (ch != prevCh && count > 0) {
1088 subFormat(appendTo, prevCh, count, capitalizationContext, fieldNum++,
1089 prevCh, handler, *workCal, status);
1090 count = 0;
1091 }
1092 if (ch == QUOTE) {
1093 // Consecutive single quotes are a single quote literal,
1094 // either outside of quotes or between quotes
1095 if ((i+1) < patternLength && fPattern[i+1] == QUOTE) {
1096 appendTo += QUOTE;
1097 ++i;
1098 } else {
1099 inQuote = ! inQuote;
1100 }
1101 }
1102 else if (!inQuote && isSyntaxChar(ch)) {
1103 // ch is a date-time pattern character to be interpreted
1104 // by subFormat(); count the number of times it is repeated
1105 prevCh = ch;
1106 ++count;
1107 }
1108 else {
1109 // Append quoted characters and unquoted non-pattern characters
1110 appendTo += ch;
1111 }
1112 }
1113
1114 // Format the last item in the pattern, if any
1115 if (count > 0) {
1116 subFormat(appendTo, prevCh, count, capitalizationContext, fieldNum++,
1117 prevCh, handler, *workCal, status);
1118 }
1119
1120 delete calClone;
1121
1122 return appendTo;
1123}
1124
1125//----------------------------------------------------------------------
1126
1127/* Map calendar field into calendar field level.
1128 * the larger the level, the smaller the field unit.
1129 * For example, UCAL_ERA level is 0, UCAL_YEAR level is 10,
1130 * UCAL_MONTH level is 20.
1131 * NOTE: if new fields adds in, the table needs to update.
1132 */
1133const int32_t
1134SimpleDateFormat::fgCalendarFieldToLevel[] =
1135{
1136 /*GyM*/ 0, 10, 20,
1137 /*wW*/ 20, 30,
1138 /*dDEF*/ 30, 20, 30, 30,
1139 /*ahHm*/ 40, 50, 50, 60,
1140 /*sS*/ 70, 80,
1141 /*z?Y*/ 0, 0, 10,
1142 /*eug*/ 30, 10, 0,
1143 /*A?.*/ 40, 0, 0
1144};
1145
1146int32_t SimpleDateFormat::getLevelFromChar(char16_t ch) {
1147 // Map date field LETTER into calendar field level.
1148 // the larger the level, the smaller the field unit.
1149 // NOTE: if new fields adds in, the table needs to update.
1150 static const int32_t mapCharToLevel[] = {
1151 -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
1152 //
1153 -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
1154 // ! " # $ % & ' ( ) * + , - . /
1155 -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
1156#if UDAT_HAS_PATTERN_CHAR_FOR_TIME_SEPARATOR0
1157 // 0 1 2 3 4 5 6 7 8 9 : ; < = > ?
1158 -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, 0, -1, -1, -1, -1, -1,
1159#else
1160 // 0 1 2 3 4 5 6 7 8 9 : ; < = > ?
1161 -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
1162#endif
1163 // @ A B C D E F G H I J K L M N O
1164 -1, 40, -1, -1, 20, 30, 30, 0, 50, -1, -1, 50, 20, 20, -1, 0,
1165 // P Q R S T U V W X Y Z [ \ ] ^ _
1166 -1, 20, -1, 80, -1, 10, 0, 30, 0, 10, 0, -1, -1, -1, -1, -1,
1167 // ` a b c d e f g h i j k l m n o
1168 -1, 40, -1, 30, 30, 30, -1, 0, 50, -1, -1, 50, 0, 60, -1, -1,
1169 // p q r s t u v w x y z { | } ~
1170 -1, 20, 10, 70, -1, 10, 0, 20, 0, 10, 0, -1, -1, -1, -1, -1
1171 };
1172
1173 return ch < UPRV_LENGTHOF(mapCharToLevel)(int32_t)(sizeof(mapCharToLevel)/sizeof((mapCharToLevel)[0])) ? mapCharToLevel[ch] : -1;
1174}
1175
1176UBool SimpleDateFormat::isSyntaxChar(char16_t ch) {
1177 static const UBool mapCharToIsSyntax[] = {
1178 //
1179 false, false, false, false, false, false, false, false,
1180 //
1181 false, false, false, false, false, false, false, false,
1182 //
1183 false, false, false, false, false, false, false, false,
1184 //
1185 false, false, false, false, false, false, false, false,
1186 // ! " # $ % & '
1187 false, false, false, false, false, false, false, false,
1188 // ( ) * + , - . /
1189 false, false, false, false, false, false, false, false,
1190 // 0 1 2 3 4 5 6 7
1191 false, false, false, false, false, false, false, false,
1192#if UDAT_HAS_PATTERN_CHAR_FOR_TIME_SEPARATOR0
1193 // 8 9 : ; < = > ?
1194 false, false, true, false, false, false, false, false,
1195#else
1196 // 8 9 : ; < = > ?
1197 false, false, false, false, false, false, false, false,
1198#endif
1199 // @ A B C D E F G
1200 false, true, true, true, true, true, true, true,
1201 // H I J K L M N O
1202 true, true, true, true, true, true, true, true,
1203 // P Q R S T U V W
1204 true, true, true, true, true, true, true, true,
1205 // X Y Z [ \ ] ^ _
1206 true, true, true, false, false, false, false, false,
1207 // ` a b c d e f g
1208 false, true, true, true, true, true, true, true,
1209 // h i j k l m n o
1210 true, true, true, true, true, true, true, true,
1211 // p q r s t u v w
1212 true, true, true, true, true, true, true, true,
1213 // x y z { | } ~
1214 true, true, true, false, false, false, false, false
1215 };
1216
1217 return ch < UPRV_LENGTHOF(mapCharToIsSyntax)(int32_t)(sizeof(mapCharToIsSyntax)/sizeof((mapCharToIsSyntax
)[0])) ? mapCharToIsSyntax[ch] : false;
1218}
1219
1220// Map index into pattern character string to Calendar field number.
1221const UCalendarDateFields
1222SimpleDateFormat::fgPatternIndexToCalendarField[] =
1223{
1224 /*GyM*/ UCAL_ERA, UCAL_YEAR, UCAL_MONTH,
1225 /*dkH*/ UCAL_DATE, UCAL_HOUR_OF_DAY, UCAL_HOUR_OF_DAY,
1226 /*msS*/ UCAL_MINUTE, UCAL_SECOND, UCAL_MILLISECOND,
1227 /*EDF*/ UCAL_DAY_OF_WEEK, UCAL_DAY_OF_YEAR, UCAL_DAY_OF_WEEK_IN_MONTH,
1228 /*wWa*/ UCAL_WEEK_OF_YEAR, UCAL_WEEK_OF_MONTH, UCAL_AM_PM,
1229 /*hKz*/ UCAL_HOUR, UCAL_HOUR, UCAL_ZONE_OFFSET,
1230 /*Yeu*/ UCAL_YEAR_WOY, UCAL_DOW_LOCAL, UCAL_EXTENDED_YEAR,
1231 /*gAZ*/ UCAL_JULIAN_DAY, UCAL_MILLISECONDS_IN_DAY, UCAL_ZONE_OFFSET,
1232 /*v*/ UCAL_ZONE_OFFSET,
1233 /*c*/ UCAL_DOW_LOCAL,
1234 /*L*/ UCAL_MONTH,
1235 /*Q*/ UCAL_MONTH,
1236 /*q*/ UCAL_MONTH,
1237 /*V*/ UCAL_ZONE_OFFSET,
1238 /*U*/ UCAL_YEAR,
1239 /*O*/ UCAL_ZONE_OFFSET,
1240 /*Xx*/ UCAL_ZONE_OFFSET, UCAL_ZONE_OFFSET,
1241 /*r*/ UCAL_EXTENDED_YEAR,
1242 /*bB*/ UCAL_FIELD_COUNT, UCAL_FIELD_COUNT, // no mappings to calendar fields
1243#if UDAT_HAS_PATTERN_CHAR_FOR_TIME_SEPARATOR0
1244 /*:*/ UCAL_FIELD_COUNT, /* => no useful mapping to any calendar field */
1245#else
1246 /*no pattern char for UDAT_TIME_SEPARATOR_FIELD*/ UCAL_FIELD_COUNT, /* => no useful mapping to any calendar field */
1247#endif
1248};
1249
1250// Map index into pattern character string to DateFormat field number
1251const UDateFormatField
1252SimpleDateFormat::fgPatternIndexToDateFormatField[] = {
1253 /*GyM*/ UDAT_ERA_FIELD, UDAT_YEAR_FIELD, UDAT_MONTH_FIELD,
1254 /*dkH*/ UDAT_DATE_FIELD, UDAT_HOUR_OF_DAY1_FIELD, UDAT_HOUR_OF_DAY0_FIELD,
1255 /*msS*/ UDAT_MINUTE_FIELD, UDAT_SECOND_FIELD, UDAT_FRACTIONAL_SECOND_FIELD,
1256 /*EDF*/ UDAT_DAY_OF_WEEK_FIELD, UDAT_DAY_OF_YEAR_FIELD, UDAT_DAY_OF_WEEK_IN_MONTH_FIELD,
1257 /*wWa*/ UDAT_WEEK_OF_YEAR_FIELD, UDAT_WEEK_OF_MONTH_FIELD, UDAT_AM_PM_FIELD,
1258 /*hKz*/ UDAT_HOUR1_FIELD, UDAT_HOUR0_FIELD, UDAT_TIMEZONE_FIELD,
1259 /*Yeu*/ UDAT_YEAR_WOY_FIELD, UDAT_DOW_LOCAL_FIELD, UDAT_EXTENDED_YEAR_FIELD,
1260 /*gAZ*/ UDAT_JULIAN_DAY_FIELD, UDAT_MILLISECONDS_IN_DAY_FIELD, UDAT_TIMEZONE_RFC_FIELD,
1261 /*v*/ UDAT_TIMEZONE_GENERIC_FIELD,
1262 /*c*/ UDAT_STANDALONE_DAY_FIELD,
1263 /*L*/ UDAT_STANDALONE_MONTH_FIELD,
1264 /*Q*/ UDAT_QUARTER_FIELD,
1265 /*q*/ UDAT_STANDALONE_QUARTER_FIELD,
1266 /*V*/ UDAT_TIMEZONE_SPECIAL_FIELD,
1267 /*U*/ UDAT_YEAR_NAME_FIELD,
1268 /*O*/ UDAT_TIMEZONE_LOCALIZED_GMT_OFFSET_FIELD,
1269 /*Xx*/ UDAT_TIMEZONE_ISO_FIELD, UDAT_TIMEZONE_ISO_LOCAL_FIELD,
1270 /*r*/ UDAT_RELATED_YEAR_FIELD,
1271 /*bB*/ UDAT_AM_PM_MIDNIGHT_NOON_FIELD, UDAT_FLEXIBLE_DAY_PERIOD_FIELD,
1272#if UDAT_HAS_PATTERN_CHAR_FOR_TIME_SEPARATOR0
1273 /*:*/ UDAT_TIME_SEPARATOR_FIELD,
1274#else
1275 /*no pattern char for UDAT_TIME_SEPARATOR_FIELD*/ UDAT_TIME_SEPARATOR_FIELD,
1276#endif
1277};
1278
1279//----------------------------------------------------------------------
1280
1281/**
1282 * Append symbols[value] to dst. Make sure the array index is not out
1283 * of bounds.
1284 */
1285static inline void
1286_appendSymbol(UnicodeString& dst,
1287 int32_t value,
1288 const UnicodeString* symbols,
1289 int32_t symbolsCount) {
1290 U_ASSERT(0 <= value && value < symbolsCount)(static_cast <bool> (0 <= value && value <
symbolsCount) ? void (0) : __assert_fail ("0 <= value && value < symbolsCount"
, __builtin_FILE (), __builtin_LINE (), __extension__ __PRETTY_FUNCTION__
));
1291 if (0 <= value && value < symbolsCount) {
1292 dst += symbols[value];
1293 }
1294}
1295
1296static inline void
1297_appendSymbolWithMonthPattern(UnicodeString& dst, int32_t value, const UnicodeString* symbols, int32_t symbolsCount,
1298 const UnicodeString* monthPattern, UErrorCode& status) {
1299 U_ASSERT(0 <= value && value < symbolsCount)(static_cast <bool> (0 <= value && value <
symbolsCount) ? void (0) : __assert_fail ("0 <= value && value < symbolsCount"
, __builtin_FILE (), __builtin_LINE (), __extension__ __PRETTY_FUNCTION__
));
1300 if (0 <= value && value < symbolsCount) {
1301 if (monthPattern == nullptr) {
1302 dst += symbols[value];
1303 } else {
1304 SimpleFormatter(*monthPattern, 1, 1, status).format(symbols[value], dst, status);
1305 }
1306 }
1307}
1308
1309//----------------------------------------------------------------------
1310
1311void
1312SimpleDateFormat::initSimpleNumberFormatter(UErrorCode &status) {
1313 if (U_FAILURE(status)) {
1314 return;
1315 }
1316 const auto* df = dynamic_cast<const DecimalFormat*>(fNumberFormat);
1317 if (df == nullptr) {
1318 return;
1319 }
1320 const DecimalFormatSymbols* syms = df->getDecimalFormatSymbols();
1321 if (syms == nullptr) {
1322 return;
1323 }
1324 fSimpleNumberFormatter = new number::SimpleNumberFormatter(
1325 number::SimpleNumberFormatter::forLocaleAndSymbolsAndGroupingStrategy(
1326 fLocale, *syms, UNUM_GROUPING_OFF, status
1327 )
1328 );
1329 if (fSimpleNumberFormatter == nullptr) {
1330 status = U_MEMORY_ALLOCATION_ERROR;
1331 }
1332}
1333
1334void
1335SimpleDateFormat::initNumberFormatters(const Locale &locale,UErrorCode &status) {
1336 if (U_FAILURE(status)) {
1337 return;
1338 }
1339 if ( fDateOverride.isBogus() && fTimeOverride.isBogus() ) {
1340 return;
1341 }
1342 umtx_lockumtx_lock_77(&LOCK);
1343 if (fSharedNumberFormatters == nullptr) {
1344 fSharedNumberFormatters = allocSharedNumberFormatters();
1345 if (fSharedNumberFormatters == nullptr) {
1346 status = U_MEMORY_ALLOCATION_ERROR;
1347 }
1348 }
1349 umtx_unlockumtx_unlock_77(&LOCK);
1350
1351 if (U_FAILURE(status)) {
1352 return;
1353 }
1354
1355 processOverrideString(locale,fDateOverride,kOvrStrDate,status);
1356 processOverrideString(locale,fTimeOverride,kOvrStrTime,status);
1357}
1358
1359void
1360SimpleDateFormat::processOverrideString(const Locale &locale, const UnicodeString &str, int8_t type, UErrorCode &status) {
1361 if (str.isBogus() || U_FAILURE(status)) {
1362 return;
1363 }
1364
1365 int32_t start = 0;
1366 int32_t len;
1367 UnicodeString nsName;
1368 UnicodeString ovrField;
1369 UBool moreToProcess = true;
1370 NSOverride *overrideList = nullptr;
1371
1372 while (moreToProcess) {
1373 int32_t delimiterPosition = str.indexOf(static_cast<char16_t>(ULOC_KEYWORD_ITEM_SEPARATOR_UNICODE0x3B), start);
1374 if (delimiterPosition == -1) {
1375 moreToProcess = false;
1376 len = str.length() - start;
1377 } else {
1378 len = delimiterPosition - start;
1379 }
1380 UnicodeString currentString(str,start,len);
1381 int32_t equalSignPosition = currentString.indexOf(static_cast<char16_t>(ULOC_KEYWORD_ASSIGN_UNICODE0x3D), 0);
1382 if (equalSignPosition == -1) { // Simple override string such as "hebrew"
1383 nsName.setTo(currentString);
1384 ovrField.setToBogus();
1385 } else { // Field specific override string such as "y=hebrew"
1386 nsName.setTo(currentString,equalSignPosition+1);
1387 ovrField.setTo(currentString,0,1); // We just need the first character.
1388 }
1389
1390 int32_t nsNameHash = nsName.hashCode();
1391 // See if the numbering system is in the override list, if not, then add it.
1392 NSOverride *curr = overrideList;
1393 const SharedNumberFormat *snf = nullptr;
1394 UBool found = false;
1395 while ( curr && !found ) {
1396 if ( curr->hash == nsNameHash ) {
1397 snf = curr->snf;
1398 found = true;
1399 }
1400 curr = curr->next;
1401 }
1402
1403 if (!found) {
1404 LocalPointer<NSOverride> cur(new NSOverride);
1405 if (!cur.isNull()) {
1406 char kw[ULOC_KEYWORD_AND_VALUES_CAPACITY100];
1407 uprv_strcpy(kw,"numbers="):: strcpy(kw, "numbers=");
1408 nsName.extract(0,len,kw+8,ULOC_KEYWORD_AND_VALUES_CAPACITY100-8,US_INVicu::UnicodeString::kInvariant);
1409
1410 Locale ovrLoc(locale.getLanguage(),locale.getCountry(),locale.getVariant(),kw);
1411 cur->hash = nsNameHash;
1412 cur->next = overrideList;
1413 SharedObject::copyPtr(
1414 createSharedNumberFormat(ovrLoc, status), cur->snf);
1415 if (U_FAILURE(status)) {
1416 if (overrideList) {
1417 overrideList->free();
1418 }
1419 return;
1420 }
1421 snf = cur->snf;
1422 overrideList = cur.orphan();
1423 } else {
1424 status = U_MEMORY_ALLOCATION_ERROR;
1425 if (overrideList) {
1426 overrideList->free();
1427 }
1428 return;
1429 }
1430 }
1431
1432 // Now that we have an appropriate number formatter, fill in the appropriate spaces in the
1433 // number formatters table.
1434 if (ovrField.isBogus()) {
1435 switch (type) {
1436 case kOvrStrDate:
1437 case kOvrStrBoth: {
1438 for ( int8_t i=0 ; i<kDateFieldsCount; i++ ) {
1439 SharedObject::copyPtr(snf, fSharedNumberFormatters[kDateFields[i]]);
1440 }
1441 if (type==kOvrStrDate) {
1442 break;
1443 }
1444 U_FALLTHROUGH[[clang::fallthrough]];
1445 }
1446 case kOvrStrTime : {
1447 for ( int8_t i=0 ; i<kTimeFieldsCount; i++ ) {
1448 SharedObject::copyPtr(snf, fSharedNumberFormatters[kTimeFields[i]]);
1449 }
1450 break;
1451 }
1452 }
1453 } else {
1454 // if the pattern character is unrecognized, signal an error and bail out
1455 UDateFormatField patternCharIndex =
1456 DateFormatSymbols::getPatternCharIndex(ovrField.charAt(0));
1457 if (patternCharIndex == UDAT_FIELD_COUNT) {
1458 status = U_INVALID_FORMAT_ERROR;
1459 if (overrideList) {
1460 overrideList->free();
1461 }
1462 return;
1463 }
1464 SharedObject::copyPtr(snf, fSharedNumberFormatters[patternCharIndex]);
1465 }
1466
1467 start = delimiterPosition + 1;
1468 }
1469 if (overrideList) {
1470 overrideList->free();
1471 }
1472}
1473
1474//---------------------------------------------------------------------
1475void
1476SimpleDateFormat::subFormat(UnicodeString &appendTo,
1477 char16_t ch,
1478 int32_t count,
1479 UDisplayContext capitalizationContext,
1480 int32_t fieldNum,
1481 char16_t fieldToOutput,
1482 FieldPositionHandler& handler,
1483 Calendar& cal,
1484 UErrorCode& status) const
1485{
1486 static const int32_t maxIntCount = 10;
1487 static const UnicodeString hebr(u"hebr");
1488
1489 if (U_FAILURE(status)) {
1
Taking false branch
→
1490 return;
1491 }
1492
1493 // this function gets called by format() to produce the appropriate substitution
1494 // text for an individual pattern symbol (e.g., "HH" or "yyyy")
1495
1496 UDateFormatField patternCharIndex = DateFormatSymbols::getPatternCharIndex(ch);
1497 int32_t beginOffset = appendTo.length();
1498 DateFormatSymbols::ECapitalizationContextUsageType capContextUsageType = DateFormatSymbols::kCapContextUsageOther;
1499
1500 // if the pattern character is unrecognized, signal an error and dump out
1501 if (patternCharIndex == UDAT_FIELD_COUNT)
2
←
Assuming 'patternCharIndex' is not equal to UDAT_FIELD_COUNT
→
3
←
Taking false branch
→
1502 {
1503 if (ch != 0x6C) { // pattern char 'l' (SMALL LETTER L) just gets ignored
1504 status = U_INVALID_FORMAT_ERROR;
1505 }
1506 return;
1507 }
1508
1509 UCalendarDateFields field = fgPatternIndexToCalendarField[patternCharIndex];
1510 int32_t value = 0;
1511 // Don't get value unless it is useful
1512 if (field < UCAL_FIELD_COUNT) {
4
←
Assuming 'field' is >= UCAL_FIELD_COUNT
→
5
←
Taking false branch
→
1513 value = (patternCharIndex != UDAT_RELATED_YEAR_FIELD)? cal.get(field, status): cal.getRelatedYear(status);
1514 if (U_FAILURE(status)) {
1515 return;
1516 }
1517 }
1518
1519 const NumberFormat *currentNumberFormat = getNumberFormatByIndex(patternCharIndex);
1520 if (currentNumberFormat == nullptr) {
6
←
Taking false branch
→
1521 status = U_INTERNAL_PROGRAM_ERROR;
1522 return;
1523 }
1524
1525 switch (patternCharIndex) {
7
←
Control jumps to 'case UDAT_FLEXIBLE_DAY_PERIOD_FIELD:' at line 1948
→
1526
1527 // for any "G" symbol, write out the appropriate era string
1528 // "GGGG" is wide era name, "GGGGG" is narrow era name, anything else is abbreviated name
1529 case UDAT_ERA_FIELD:
1530 {
1531 const char* type = cal.getType();
1532 if (strcmp(type, "chinese") == 0 ||
1533 strcmp(type, "dangi") == 0) {
1534 zeroPaddingNumber(currentNumberFormat,appendTo, value, 1, 9); // as in ICU4J
1535 } else {
1536 if (count == 5) {
1537 _appendSymbol(appendTo, value, fSymbols->fNarrowEras, fSymbols->fNarrowErasCount);
1538 capContextUsageType = DateFormatSymbols::kCapContextUsageEraNarrow;
1539 } else if (count == 4) {
1540 _appendSymbol(appendTo, value, fSymbols->fEraNames, fSymbols->fEraNamesCount);
1541 capContextUsageType = DateFormatSymbols::kCapContextUsageEraWide;
1542 } else {
1543 _appendSymbol(appendTo, value, fSymbols->fEras, fSymbols->fErasCount);
1544 capContextUsageType = DateFormatSymbols::kCapContextUsageEraAbbrev;
1545 }
1546 }
1547 }
1548 break;
1549
1550 case UDAT_YEAR_NAME_FIELD:
1551 if (fSymbols->fShortYearNames != nullptr && value <= fSymbols->fShortYearNamesCount) {
1552 // the Calendar YEAR field runs 1 through 60 for cyclic years
1553 _appendSymbol(appendTo, value - 1, fSymbols->fShortYearNames, fSymbols->fShortYearNamesCount);
1554 break;
1555 }
1556 // else fall through to numeric year handling, do not break here
1557 U_FALLTHROUGH[[clang::fallthrough]];
1558
1559 // OLD: for "yyyy", write out the whole year; for "yy", write out the last 2 digits
1560 // NEW: UTS#35:
1561//Year y yy yyy yyyy yyyyy
1562//AD 1 1 01 001 0001 00001
1563//AD 12 12 12 012 0012 00012
1564//AD 123 123 23 123 0123 00123
1565//AD 1234 1234 34 1234 1234 01234
1566//AD 12345 12345 45 12345 12345 12345
1567 case UDAT_YEAR_FIELD:
1568 case UDAT_YEAR_WOY_FIELD:
1569 if (fDateOverride.compare(hebr)==0 && value>HEBREW_CAL_CUR_MILLENIUM_START_YEAR && value<HEBREW_CAL_CUR_MILLENIUM_END_YEAR) {
1570 value-=HEBREW_CAL_CUR_MILLENIUM_START_YEAR;
1571 }
1572 if(count == 2)
1573 zeroPaddingNumber(currentNumberFormat, appendTo, value, 2, 2);
1574 else
1575 zeroPaddingNumber(currentNumberFormat, appendTo, value, count, maxIntCount);
1576 break;
1577
1578 // for "MMMM"/"LLLL", write out the whole month name, for "MMM"/"LLL", write out the month
1579 // abbreviation, for "M"/"L" or "MM"/"LL", write out the month as a number with the
1580 // appropriate number of digits
1581 // for "MMMMM"/"LLLLL", use the narrow form
1582 case UDAT_MONTH_FIELD:
1583 case UDAT_STANDALONE_MONTH_FIELD:
1584 if (typeid(cal) == typeid(HebrewCalendar)) {
1585 if (HebrewCalendar::isLeapYear(cal.get(UCAL_YEAR,status)) && value == 6 && count >= 3 )
1586 value = 13; // Show alternate form for Adar II in leap years in Hebrew calendar.
1587 if (!HebrewCalendar::isLeapYear(cal.get(UCAL_YEAR,status)) && value >= 6 && count < 3 )
1588 value--; // Adjust the month number down 1 in Hebrew non-leap years, i.e. Adar is 6, not 7.
1589 }
1590 {
1591 int32_t isLeapMonth = (fSymbols->fLeapMonthPatterns != nullptr && fSymbols->fLeapMonthPatternsCount >= DateFormatSymbols::kMonthPatternsCount)?
1592 cal.get(UCAL_IS_LEAP_MONTH, status): 0;
1593 // should consolidate the next section by using arrays of pointers & counts for the right symbols...
1594 if (count == 5) {
1595 if (patternCharIndex == UDAT_MONTH_FIELD) {
1596 _appendSymbolWithMonthPattern(appendTo, value, fSymbols->fNarrowMonths, fSymbols->fNarrowMonthsCount,
1597 (isLeapMonth!=0)? &(fSymbols->fLeapMonthPatterns[DateFormatSymbols::kLeapMonthPatternFormatNarrow]): nullptr, status);
1598 } else {
1599 _appendSymbolWithMonthPattern(appendTo, value, fSymbols->fStandaloneNarrowMonths, fSymbols->fStandaloneNarrowMonthsCount,
1600 (isLeapMonth!=0)? &(fSymbols->fLeapMonthPatterns[DateFormatSymbols::kLeapMonthPatternStandaloneNarrow]): nullptr, status);
1601 }
1602 capContextUsageType = DateFormatSymbols::kCapContextUsageMonthNarrow;
1603 } else if (count == 4) {
1604 if (patternCharIndex == UDAT_MONTH_FIELD) {
1605 _appendSymbolWithMonthPattern(appendTo, value, fSymbols->fMonths, fSymbols->fMonthsCount,
1606 (isLeapMonth!=0)? &(fSymbols->fLeapMonthPatterns[DateFormatSymbols::kLeapMonthPatternFormatWide]): nullptr, status);
1607 capContextUsageType = DateFormatSymbols::kCapContextUsageMonthFormat;
1608 } else {
1609 _appendSymbolWithMonthPattern(appendTo, value, fSymbols->fStandaloneMonths, fSymbols->fStandaloneMonthsCount,
1610 (isLeapMonth!=0)? &(fSymbols->fLeapMonthPatterns[DateFormatSymbols::kLeapMonthPatternStandaloneWide]): nullptr, status);
1611 capContextUsageType = DateFormatSymbols::kCapContextUsageMonthStandalone;
1612 }
1613 } else if (count == 3) {
1614 if (patternCharIndex == UDAT_MONTH_FIELD) {
1615 _appendSymbolWithMonthPattern(appendTo, value, fSymbols->fShortMonths, fSymbols->fShortMonthsCount,
1616 (isLeapMonth!=0)? &(fSymbols->fLeapMonthPatterns[DateFormatSymbols::kLeapMonthPatternFormatAbbrev]): nullptr, status);
1617 capContextUsageType = DateFormatSymbols::kCapContextUsageMonthFormat;
1618 } else {
1619 _appendSymbolWithMonthPattern(appendTo, value, fSymbols->fStandaloneShortMonths, fSymbols->fStandaloneShortMonthsCount,
1620 (isLeapMonth!=0)? &(fSymbols->fLeapMonthPatterns[DateFormatSymbols::kLeapMonthPatternStandaloneAbbrev]): nullptr, status);
1621 capContextUsageType = DateFormatSymbols::kCapContextUsageMonthStandalone;
1622 }
1623 } else {
1624 UnicodeString monthNumber;
1625 zeroPaddingNumber(currentNumberFormat,monthNumber, value + 1, count, maxIntCount);
1626 _appendSymbolWithMonthPattern(appendTo, 0, &monthNumber, 1,
1627 (isLeapMonth!=0)? &(fSymbols->fLeapMonthPatterns[DateFormatSymbols::kLeapMonthPatternNumeric]): nullptr, status);
1628 }
1629 }
1630 break;
1631
1632 // for "k" and "kk", write out the hour, adjusting midnight to appear as "24"
1633 case UDAT_HOUR_OF_DAY1_FIELD:
1634 if (value == 0)
1635 zeroPaddingNumber(currentNumberFormat,appendTo, cal.getMaximum(UCAL_HOUR_OF_DAY) + 1, count, maxIntCount);
1636 else
1637 zeroPaddingNumber(currentNumberFormat,appendTo, value, count, maxIntCount);
1638 break;
1639
1640 case UDAT_FRACTIONAL_SECOND_FIELD:
1641 // Fractional seconds left-justify
1642 {
1643 int32_t minDigits = (count > 3) ? 3 : count;
1644 if (count == 1) {
1645 value /= 100;
1646 } else if (count == 2) {
1647 value /= 10;
1648 }
1649 zeroPaddingNumber(currentNumberFormat, appendTo, value, minDigits, maxIntCount);
1650 if (count > 3) {
1651 zeroPaddingNumber(currentNumberFormat, appendTo, 0, count - 3, maxIntCount);
1652 }
1653 }
1654 break;
1655
1656 // for "ee" or "e", use local numeric day-of-the-week
1657 // for "EEEEEE" or "eeeeee", write out the short day-of-the-week name
1658 // for "EEEEE" or "eeeee", write out the narrow day-of-the-week name
1659 // for "EEEE" or "eeee", write out the wide day-of-the-week name
1660 // for "EEE" or "EE" or "E" or "eee", write out the abbreviated day-of-the-week name
1661 case UDAT_DOW_LOCAL_FIELD:
1662 if ( count < 3 ) {
1663 zeroPaddingNumber(currentNumberFormat,appendTo, value, count, maxIntCount);
1664 break;
1665 }
1666 // fall through to EEEEE-EEE handling, but for that we don't want local day-of-week,
1667 // we want standard day-of-week, so first fix value to work for EEEEE-EEE.
1668 value = cal.get(UCAL_DAY_OF_WEEK, status);
1669 if (U_FAILURE(status)) {
1670 return;
1671 }
1672 // fall through, do not break here
1673 U_FALLTHROUGH[[clang::fallthrough]];
1674 case UDAT_DAY_OF_WEEK_FIELD:
1675 if (count == 5) {
1676 _appendSymbol(appendTo, value, fSymbols->fNarrowWeekdays,
1677 fSymbols->fNarrowWeekdaysCount);
1678 capContextUsageType = DateFormatSymbols::kCapContextUsageDayNarrow;
1679 } else if (count == 4) {
1680 _appendSymbol(appendTo, value, fSymbols->fWeekdays,
1681 fSymbols->fWeekdaysCount);
1682 capContextUsageType = DateFormatSymbols::kCapContextUsageDayFormat;
1683 } else if (count == 6) {
1684 _appendSymbol(appendTo, value, fSymbols->fShorterWeekdays,
1685 fSymbols->fShorterWeekdaysCount);
1686 capContextUsageType = DateFormatSymbols::kCapContextUsageDayFormat;
1687 } else {
1688 _appendSymbol(appendTo, value, fSymbols->fShortWeekdays,
1689 fSymbols->fShortWeekdaysCount);
1690 capContextUsageType = DateFormatSymbols::kCapContextUsageDayFormat;
1691 }
1692 break;
1693
1694 // for "ccc", write out the abbreviated day-of-the-week name
1695 // for "cccc", write out the wide day-of-the-week name
1696 // for "ccccc", use the narrow day-of-the-week name
1697 // for "ccccc", use the short day-of-the-week name
1698 case UDAT_STANDALONE_DAY_FIELD:
1699 if ( count < 3 ) {
1700 zeroPaddingNumber(currentNumberFormat,appendTo, value, 1, maxIntCount);
1701 break;
1702 }
1703 // fall through to alpha DOW handling, but for that we don't want local day-of-week,
1704 // we want standard day-of-week, so first fix value.
1705 value = cal.get(UCAL_DAY_OF_WEEK, status);
1706 if (U_FAILURE(status)) {
1707 return;
1708 }
1709 if (count == 5) {
1710 _appendSymbol(appendTo, value, fSymbols->fStandaloneNarrowWeekdays,
1711 fSymbols->fStandaloneNarrowWeekdaysCount);
1712 capContextUsageType = DateFormatSymbols::kCapContextUsageDayNarrow;
1713 } else if (count == 4) {
1714 _appendSymbol(appendTo, value, fSymbols->fStandaloneWeekdays,
1715 fSymbols->fStandaloneWeekdaysCount);
1716 capContextUsageType = DateFormatSymbols::kCapContextUsageDayStandalone;
1717 } else if (count == 6) {
1718 _appendSymbol(appendTo, value, fSymbols->fStandaloneShorterWeekdays,
1719 fSymbols->fStandaloneShorterWeekdaysCount);
1720 capContextUsageType = DateFormatSymbols::kCapContextUsageDayStandalone;
1721 } else { // count == 3
1722 _appendSymbol(appendTo, value, fSymbols->fStandaloneShortWeekdays,
1723 fSymbols->fStandaloneShortWeekdaysCount);
1724 capContextUsageType = DateFormatSymbols::kCapContextUsageDayStandalone;
1725 }
1726 break;
1727
1728 // for "a" symbol, write out the whole AM/PM string
1729 case UDAT_AM_PM_FIELD:
1730 if (count < 5) {
1731 _appendSymbol(appendTo, value, fSymbols->fAmPms,
1732 fSymbols->fAmPmsCount);
1733 } else {
1734 _appendSymbol(appendTo, value, fSymbols->fNarrowAmPms,
1735 fSymbols->fNarrowAmPmsCount);
1736 }
1737 break;
1738
1739 // if we see pattern character for UDAT_TIME_SEPARATOR_FIELD (none currently defined),
1740 // write out the time separator string. Leave support in for future definition.
1741 case UDAT_TIME_SEPARATOR_FIELD:
1742 {
1743 UnicodeString separator;
1744 appendTo += fSymbols->getTimeSeparatorString(separator);
1745 }
1746 break;
1747
1748 // for "h" and "hh", write out the hour, adjusting noon and midnight to show up
1749 // as "12"
1750 case UDAT_HOUR1_FIELD:
1751 if (value == 0)
1752 zeroPaddingNumber(currentNumberFormat,appendTo, cal.getLeastMaximum(UCAL_HOUR) + 1, count, maxIntCount);
1753 else
1754 zeroPaddingNumber(currentNumberFormat,appendTo, value, count, maxIntCount);
1755 break;
1756
1757 case UDAT_TIMEZONE_FIELD: // 'z'
1758 case UDAT_TIMEZONE_RFC_FIELD: // 'Z'
1759 case UDAT_TIMEZONE_GENERIC_FIELD: // 'v'
1760 case UDAT_TIMEZONE_SPECIAL_FIELD: // 'V'
1761 case UDAT_TIMEZONE_LOCALIZED_GMT_OFFSET_FIELD: // 'O'
1762 case UDAT_TIMEZONE_ISO_FIELD: // 'X'
1763 case UDAT_TIMEZONE_ISO_LOCAL_FIELD: // 'x'
1764 {
1765 char16_t zsbuf[ZONE_NAME_U16_MAX128];
1766 UnicodeString zoneString(zsbuf, 0, UPRV_LENGTHOF(zsbuf)(int32_t)(sizeof(zsbuf)/sizeof((zsbuf)[0])));
1767 const TimeZone& tz = cal.getTimeZone();
1768 UDate date = cal.getTime(status);
1769 const TimeZoneFormat *tzfmt = tzFormat(status);
1770 if (U_SUCCESS(status)) {
1771 switch (patternCharIndex) {
1772 case UDAT_TIMEZONE_FIELD:
1773 if (count < 4) {
1774 // "z", "zz", "zzz"
1775 tzfmt->format(UTZFMT_STYLE_SPECIFIC_SHORT, tz, date, zoneString);
1776 capContextUsageType = DateFormatSymbols::kCapContextUsageMetazoneShort;
1777 } else {
1778 // "zzzz" or longer
1779 tzfmt->format(UTZFMT_STYLE_SPECIFIC_LONG, tz, date, zoneString);
1780 capContextUsageType = DateFormatSymbols::kCapContextUsageMetazoneLong;
1781 }
1782 break;
1783 case UDAT_TIMEZONE_RFC_FIELD:
1784 if (count < 4) {
1785 // "Z"
1786 tzfmt->format(UTZFMT_STYLE_ISO_BASIC_LOCAL_FULL, tz, date, zoneString);
1787 } else if (count == 5) {
1788 // "ZZZZZ"
1789 tzfmt->format(UTZFMT_STYLE_ISO_EXTENDED_FULL, tz, date, zoneString);
1790 } else {
1791 // "ZZ", "ZZZ", "ZZZZ"
1792 tzfmt->format(UTZFMT_STYLE_LOCALIZED_GMT, tz, date, zoneString);
1793 }
1794 break;
1795 case UDAT_TIMEZONE_GENERIC_FIELD:
1796 if (count == 1) {
1797 // "v"
1798 tzfmt->format(UTZFMT_STYLE_GENERIC_SHORT, tz, date, zoneString);
1799 capContextUsageType = DateFormatSymbols::kCapContextUsageMetazoneShort;
1800 } else if (count == 4) {
1801 // "vvvv"
1802 tzfmt->format(UTZFMT_STYLE_GENERIC_LONG, tz, date, zoneString);
1803 capContextUsageType = DateFormatSymbols::kCapContextUsageMetazoneLong;
1804 }
1805 break;
1806 case UDAT_TIMEZONE_SPECIAL_FIELD:
1807 if (count == 1) {
1808 // "V"
1809 tzfmt->format(UTZFMT_STYLE_ZONE_ID_SHORT, tz, date, zoneString);
1810 } else if (count == 2) {
1811 // "VV"
1812 tzfmt->format(UTZFMT_STYLE_ZONE_ID, tz, date, zoneString);
1813 } else if (count == 3) {
1814 // "VVV"
1815 tzfmt->format(UTZFMT_STYLE_EXEMPLAR_LOCATION, tz, date, zoneString);
1816 } else if (count == 4) {
1817 // "VVVV"
1818 tzfmt->format(UTZFMT_STYLE_GENERIC_LOCATION, tz, date, zoneString);
1819 capContextUsageType = DateFormatSymbols::kCapContextUsageZoneLong;
1820 }
1821 break;
1822 case UDAT_TIMEZONE_LOCALIZED_GMT_OFFSET_FIELD:
1823 if (count == 1) {
1824 // "O"
1825 tzfmt->format(UTZFMT_STYLE_LOCALIZED_GMT_SHORT, tz, date, zoneString);
1826 } else if (count == 4) {
1827 // "OOOO"
1828 tzfmt->format(UTZFMT_STYLE_LOCALIZED_GMT, tz, date, zoneString);
1829 }
1830 break;
1831 case UDAT_TIMEZONE_ISO_FIELD:
1832 if (count == 1) {
1833 // "X"
1834 tzfmt->format(UTZFMT_STYLE_ISO_BASIC_SHORT, tz, date, zoneString);
1835 } else if (count == 2) {
1836 // "XX"
1837 tzfmt->format(UTZFMT_STYLE_ISO_BASIC_FIXED, tz, date, zoneString);
1838 } else if (count == 3) {
1839 // "XXX"
1840 tzfmt->format(UTZFMT_STYLE_ISO_EXTENDED_FIXED, tz, date, zoneString);
1841 } else if (count == 4) {
1842 // "XXXX"
1843 tzfmt->format(UTZFMT_STYLE_ISO_BASIC_FULL, tz, date, zoneString);
1844 } else if (count == 5) {
1845 // "XXXXX"
1846 tzfmt->format(UTZFMT_STYLE_ISO_EXTENDED_FULL, tz, date, zoneString);
1847 }
1848 break;
1849 case UDAT_TIMEZONE_ISO_LOCAL_FIELD:
1850 if (count == 1) {
1851 // "x"
1852 tzfmt->format(UTZFMT_STYLE_ISO_BASIC_LOCAL_SHORT, tz, date, zoneString);
1853 } else if (count == 2) {
1854 // "xx"
1855 tzfmt->format(UTZFMT_STYLE_ISO_BASIC_LOCAL_FIXED, tz, date, zoneString);
1856 } else if (count == 3) {
1857 // "xxx"
1858 tzfmt->format(UTZFMT_STYLE_ISO_EXTENDED_LOCAL_FIXED, tz, date, zoneString);
1859 } else if (count == 4) {
1860 // "xxxx"
1861 tzfmt->format(UTZFMT_STYLE_ISO_BASIC_LOCAL_FULL, tz, date, zoneString);
1862 } else if (count == 5) {
1863 // "xxxxx"
1864 tzfmt->format(UTZFMT_STYLE_ISO_EXTENDED_LOCAL_FULL, tz, date, zoneString);
1865 }
1866 break;
1867 default:
1868 UPRV_UNREACHABLE_EXITabort();
1869 }
1870 }
1871 appendTo += zoneString;
1872 }
1873 break;
1874
1875 case UDAT_QUARTER_FIELD:
1876 if (count >= 5)
1877 _appendSymbol(appendTo, value/3, fSymbols->fNarrowQuarters,
1878 fSymbols->fNarrowQuartersCount);
1879 else if (count == 4)
1880 _appendSymbol(appendTo, value/3, fSymbols->fQuarters,
1881 fSymbols->fQuartersCount);
1882 else if (count == 3)
1883 _appendSymbol(appendTo, value/3, fSymbols->fShortQuarters,
1884 fSymbols->fShortQuartersCount);
1885 else
1886 zeroPaddingNumber(currentNumberFormat,appendTo, (value/3) + 1, count, maxIntCount);
1887 break;
1888
1889 case UDAT_STANDALONE_QUARTER_FIELD:
1890 if (count >= 5)
1891 _appendSymbol(appendTo, value/3, fSymbols->fStandaloneNarrowQuarters,
1892 fSymbols->fStandaloneNarrowQuartersCount);
1893 else if (count == 4)
1894 _appendSymbol(appendTo, value/3, fSymbols->fStandaloneQuarters,
1895 fSymbols->fStandaloneQuartersCount);
1896 else if (count == 3)
1897 _appendSymbol(appendTo, value/3, fSymbols->fStandaloneShortQuarters,
1898 fSymbols->fStandaloneShortQuartersCount);
1899 else
1900 zeroPaddingNumber(currentNumberFormat,appendTo, (value/3) + 1, count, maxIntCount);
1901 break;
1902
1903 case UDAT_AM_PM_MIDNIGHT_NOON_FIELD:
1904 {
1905 const UnicodeString *toAppend = nullptr;
1906 int32_t hour = cal.get(UCAL_HOUR_OF_DAY, status);
1907
1908 // Note: "midnight" can be ambiguous as to whether it refers to beginning of day or end of day.
1909 // For ICU 57 output of "midnight" is temporarily suppressed.
1910
1911 // For "midnight" and "noon":
1912 // Time, as displayed, must be exactly noon or midnight.
1913 // This means minutes and seconds, if present, must be zero.
1914 if ((/*hour == 0 ||*/ hour == 12) &&
1915 (!fHasMinute || cal.get(UCAL_MINUTE, status) == 0) &&
1916 (!fHasSecond || cal.get(UCAL_SECOND, status) == 0)) {
1917 // Stealing am/pm value to use as our array index.
1918 // It works out: am/midnight are both 0, pm/noon are both 1,
1919 // 12 am is 12 midnight, and 12 pm is 12 noon.
1920 int32_t val = cal.get(UCAL_AM_PM, status);
1921
1922 if (count <= 3) {
1923 toAppend = &fSymbols->fAbbreviatedDayPeriods[val];
1924 } else if (count == 4 || count > 5) {
1925 toAppend = &fSymbols->fWideDayPeriods[val];
1926 } else { // count == 5
1927 toAppend = &fSymbols->fNarrowDayPeriods[val];
1928 }
1929 }
1930
1931 // toAppend is nullptr if time isn't exactly midnight or noon (as displayed).
1932 // toAppend is bogus if time is midnight or noon, but no localized string exists.
1933 // In either case, fall back to am/pm.
1934 if (toAppend == nullptr || toAppend->isBogus()) {
1935 // Reformat with identical arguments except ch, now changed to 'a'.
1936 // We are passing a different fieldToOutput because we want to add
1937 // 'b' to field position. This makes this fallback stable when
1938 // there is a data change on locales.
1939 subFormat(appendTo, u'a', count, capitalizationContext, fieldNum, u'b', handler, cal, status);
1940 return;
1941 } else {
1942 appendTo += *toAppend;
1943 }
1944
1945 break;
1946 }
1947
1948 case UDAT_FLEXIBLE_DAY_PERIOD_FIELD:
1949 {
1950 // TODO: Maybe fetch the DayperiodRules during initialization (instead of at the first
1951 // loading of an instance) if a relevant pattern character (b or B) is used.
1952 const DayPeriodRules *ruleSet = DayPeriodRules::getInstance(this->getSmpFmtLocale(), status);
1953 if (U_FAILURE(status)) {
8
←
Taking false branch
→
1954 // Data doesn't conform to spec, therefore loading failed.
1955 break;
1956 }
1957 if (ruleSet == nullptr) {
9
←
Assuming the condition is false
→
10
←
Taking false branch
→
1958 // Data doesn't exist for the locale we're looking for.
1959 // Falling back to am/pm.
1960 // We are passing a different fieldToOutput because we want to add
1961 // 'B' to field position. This makes this fallback stable when
1962 // there is a data change on locales.
1963 subFormat(appendTo, u'a', count, capitalizationContext, fieldNum, u'B', handler, cal, status);
1964 return;
1965 }
1966
1967 // Get current display time.
1968 int32_t hour = cal.get(UCAL_HOUR_OF_DAY, status);
1969 int32_t minute = 0;
1970 if (fHasMinute) {
11
←
Assuming field 'fHasMinute' is 0
→
12
←
Taking false branch
→
1971 minute = cal.get(UCAL_MINUTE, status);
1972 }
1973 int32_t second = 0;
1974 if (fHasSecond) {
13
←
Assuming field 'fHasSecond' is 0
→
14
←
Taking false branch
→
1975 second = cal.get(UCAL_SECOND, status);
1976 }
1977
1978 // Determine day period.
1979 DayPeriodRules::DayPeriod periodType;
1980 if (hour == 0 && minute == 0 && second == 0 && ruleSet->hasMidnight()) {
15
←
Assuming 'hour' is not equal to 0
→
1981 periodType = DayPeriodRules::DAYPERIOD_MIDNIGHT;
1982 } else if (hour == 12 && minute == 0 && second == 0 && ruleSet->hasNoon()) {
16
←
Assuming 'hour' is not equal to 12
→
1983 periodType = DayPeriodRules::DAYPERIOD_NOON;
1984 } else {
1985 periodType = ruleSet->getDayPeriodForHour(hour);
1986 }
1987
1988 // Rule set exists, therefore periodType can't be UNKNOWN.
1989 // Get localized string.
1990 U_ASSERT(periodType != DayPeriodRules::DAYPERIOD_UNKNOWN)(static_cast <bool> (periodType != DayPeriodRules::DAYPERIOD_UNKNOWN
) ? void (0) : __assert_fail ("periodType != DayPeriodRules::DAYPERIOD_UNKNOWN"
, __builtin_FILE (), __builtin_LINE (), __extension__ __PRETTY_FUNCTION__
));
17
←
Assuming 'periodType' is not equal to DAYPERIOD_UNKNOWN
→
18
←
'?' condition is true
→
1991 UnicodeString *toAppend = nullptr;
1992 int32_t index;
1993
1994 // Note: "midnight" can be ambiguous as to whether it refers to beginning of day or end of day.
1995 // For ICU 57 output of "midnight" is temporarily suppressed.
1996
1997 if (periodType != DayPeriodRules::DAYPERIOD_AM &&
19
←
Assuming 'periodType' is not equal to DAYPERIOD_AM
→
22
←
Taking true branch
→
1998 periodType != DayPeriodRules::DAYPERIOD_PM &&
20
←
Assuming 'periodType' is not equal to DAYPERIOD_PM
→
1999 periodType != DayPeriodRules::DAYPERIOD_MIDNIGHT) {
21
←
Assuming 'periodType' is not equal to DAYPERIOD_MIDNIGHT
→
2000 index = static_cast<int32_t>(periodType);
2001 if (count <= 3) {
23
←
Assuming 'count' is <= 3
→
24
←
Taking true branch
→
2002 toAppend = &fSymbols->fAbbreviatedDayPeriods[index]; // i.e. short
25
←
Value assigned to 'toAppend'
→
2003 } else if (count == 4 || count > 5) {
2004 toAppend = &fSymbols->fWideDayPeriods[index];
2005 } else { // count == 5
2006 toAppend = &fSymbols->fNarrowDayPeriods[index];
2007 }
2008 }
2009
2010 // Fallback schedule:
2011 // Midnight/Noon -> General Periods -> AM/PM.
2012
2013 // Midnight/Noon -> General Periods.
2014 if ((toAppend == nullptr || toAppend->isBogus()) &&
26
←
Assuming pointer value is null
→
2015 (periodType
26.1
'periodType' is not equal to DAYPERIOD_MIDNIGHT
 == DayPeriodRules::DAYPERIOD_MIDNIGHT ||
2016 periodType == DayPeriodRules::DAYPERIOD_NOON)) {
27
←
Assuming 'periodType' is not equal to DAYPERIOD_NOON
→
2017 periodType = ruleSet->getDayPeriodForHour(hour);
2018 index = static_cast<int32_t>(periodType);
2019
2020 if (count <= 3) {
2021 toAppend = &fSymbols->fAbbreviatedDayPeriods[index]; // i.e. short
2022 } else if (count == 4 || count > 5) {
2023 toAppend = &fSymbols->fWideDayPeriods[index];
2024 } else { // count == 5
2025 toAppend = &fSymbols->fNarrowDayPeriods[index];
2026 }
2027 }
2028
2029 // General Periods -> AM/PM.
2030 if (periodType
27.1
'periodType' is not equal to DAYPERIOD_AM
 == DayPeriodRules::DAYPERIOD_AM ||
2031 periodType
27.2
'periodType' is not equal to DAYPERIOD_PM
 == DayPeriodRules::DAYPERIOD_PM ||
2032 toAppend->isBogus()) {
28
←
Called C++ object pointer is null
2033 // We are passing a different fieldToOutput because we want to add
2034 // 'B' to field position iterator. This makes this fallback stable when
2035 // there is a data change on locales.
2036 subFormat(appendTo, u'a', count, capitalizationContext, fieldNum, u'B', handler, cal, status);
2037 return;
2038 }
2039 else {
2040 appendTo += *toAppend;
2041 }
2042
2043 break;
2044 }
2045
2046 // all of the other pattern symbols can be formatted as simple numbers with
2047 // appropriate zero padding
2048 default:
2049 zeroPaddingNumber(currentNumberFormat,appendTo, value, count, maxIntCount);
2050 break;
2051 }
2052#if !UCONFIG_NO_BREAK_ITERATION1
2053 // if first field, check to see whether we need to and are able to titlecase it
2054 if (fieldNum == 0 && fCapitalizationBrkIter != nullptr && appendTo.length() > beginOffset &&
2055 u_isloweru_islower_77(appendTo.char32At(beginOffset))) {
2056 UBool titlecase = false;
2057 switch (capitalizationContext) {
2058 case UDISPCTX_CAPITALIZATION_FOR_BEGINNING_OF_SENTENCE:
2059 titlecase = true;
2060 break;
2061 case UDISPCTX_CAPITALIZATION_FOR_UI_LIST_OR_MENU:
2062 titlecase = fSymbols->fCapitalization[capContextUsageType][0];
2063 break;
2064 case UDISPCTX_CAPITALIZATION_FOR_STANDALONE:
2065 titlecase = fSymbols->fCapitalization[capContextUsageType][1];
2066 break;
2067 default:
2068 // titlecase = false;
2069 break;
2070 }
2071 if (titlecase) {
2072 BreakIterator* const mutableCapitalizationBrkIter = fCapitalizationBrkIter->clone();
2073 UnicodeString firstField(appendTo, beginOffset);
2074 firstField.toTitle(mutableCapitalizationBrkIter, fLocale, U_TITLECASE_NO_LOWERCASE0x100 | U_TITLECASE_NO_BREAK_ADJUSTMENT0x200);
2075 appendTo.replaceBetween(beginOffset, appendTo.length(), firstField);
2076 delete mutableCapitalizationBrkIter;
2077 }
2078 }
2079#endif
2080
2081 handler.addAttribute(DateFormatSymbols::getPatternCharIndex(fieldToOutput), beginOffset, appendTo.length());
2082}
2083
2084//----------------------------------------------------------------------
2085
2086void SimpleDateFormat::adoptNumberFormat(NumberFormat *formatToAdopt) {
2087 // Null out the fast formatter, it references fNumberFormat which we're
2088 // about to invalidate
2089 delete fSimpleNumberFormatter;
2090 fSimpleNumberFormatter = nullptr;
2091
2092 fixNumberFormatForDates(*formatToAdopt);
2093 delete fNumberFormat;
2094 fNumberFormat = formatToAdopt;
2095
2096 // We successfully set the default number format. Now delete the overrides
2097 // (can't fail).
2098 if (fSharedNumberFormatters) {
2099 freeSharedNumberFormatters(fSharedNumberFormatters);
2100 fSharedNumberFormatters = nullptr;
2101 }
2102
2103 // Recompute fSimpleNumberFormatter if necessary
2104 UErrorCode localStatus = U_ZERO_ERROR;
2105 initSimpleNumberFormatter(localStatus);
2106}
2107
2108void SimpleDateFormat::adoptNumberFormat(const UnicodeString& fields, NumberFormat *formatToAdopt, UErrorCode &status){
2109 fixNumberFormatForDates(*formatToAdopt);
2110 LocalPointer<NumberFormat> fmt(formatToAdopt);
2111 if (U_FAILURE(status)) {
2112 return;
2113 }
2114
2115 // We must ensure fSharedNumberFormatters is allocated.
2116 if (fSharedNumberFormatters == nullptr) {
2117 fSharedNumberFormatters = allocSharedNumberFormatters();
2118 if (fSharedNumberFormatters == nullptr) {
2119 status = U_MEMORY_ALLOCATION_ERROR;
2120 return;
2121 }
2122 }
2123 const SharedNumberFormat *newFormat = createSharedNumberFormat(fmt.orphan());
2124 if (newFormat == nullptr) {
2125 status = U_MEMORY_ALLOCATION_ERROR;
2126 return;
2127 }
2128 for (int i=0; i<fields.length(); i++) {
2129 char16_t field = fields.charAt(i);
2130 // if the pattern character is unrecognized, signal an error and bail out
2131 UDateFormatField patternCharIndex = DateFormatSymbols::getPatternCharIndex(field);
2132 if (patternCharIndex == UDAT_FIELD_COUNT) {
2133 status = U_INVALID_FORMAT_ERROR;
2134 newFormat->deleteIfZeroRefCount();
2135 return;
2136 }
2137
2138 // Set the number formatter in the table
2139 SharedObject::copyPtr(
2140 newFormat, fSharedNumberFormatters[patternCharIndex]);
2141 }
2142 newFormat->deleteIfZeroRefCount();
2143}
2144
2145const NumberFormat *
2146SimpleDateFormat::getNumberFormatForField(char16_t field) const {
2147 UDateFormatField index = DateFormatSymbols::getPatternCharIndex(field);
2148 if (index == UDAT_FIELD_COUNT) {
2149 return nullptr;
2150 }
2151 return getNumberFormatByIndex(index);
2152}
2153
2154//----------------------------------------------------------------------
2155void
2156SimpleDateFormat::zeroPaddingNumber(
2157 const NumberFormat *currentNumberFormat,
2158 UnicodeString &appendTo,
2159 int32_t value, int32_t minDigits, int32_t maxDigits) const
2160{
2161
2162 if (currentNumberFormat == fNumberFormat && fSimpleNumberFormatter) {
2163 // Can use fast path
2164 // We create UFormattedNumberData ourselves to avoid a heap allocation
2165 // and corresponding free. Set the pointer to null afterwards to prevent
2166 // the implementation from attempting to free it.
2167 UErrorCode localStatus = U_ZERO_ERROR;
2168 number::impl::UFormattedNumberData data;
2169 data.quantity.setToLong(value);
2170 number::SimpleNumber number(&data, localStatus);
2171 number.setMinimumIntegerDigits(minDigits, localStatus);
2172 number.setMaximumIntegerDigits(maxDigits, localStatus);
2173
2174 number::FormattedNumber result = fSimpleNumberFormatter->format(std::move(number), localStatus);
2175 if (U_FAILURE(localStatus)) {
2176 result.fData = nullptr;
2177 return;
2178 }
2179 UnicodeStringAppendable appendable(appendTo);
2180 result.appendTo(appendable, localStatus);
2181 result.fData = nullptr;
2182 return;
2183 }
2184
2185 // Check for RBNF (no clone necessary)
2186 const auto* rbnf = dynamic_cast<const RuleBasedNumberFormat*>(currentNumberFormat);
2187 if (rbnf != nullptr) {
2188 FieldPosition pos(FieldPosition::DONT_CARE);
2189 rbnf->format(value, appendTo, pos); // 3rd arg is there to speed up processing
2190 return;
2191 }
2192
2193 // Fall back to slow path (clone and mutate the NumberFormat)
2194 if (currentNumberFormat != nullptr) {
2195 FieldPosition pos(FieldPosition::DONT_CARE);
2196 LocalPointer<NumberFormat> nf(currentNumberFormat->clone());
2197 nf->setMinimumIntegerDigits(minDigits);
2198 nf->setMaximumIntegerDigits(maxDigits);
2199 nf->format(value, appendTo, pos); // 3rd arg is there to speed up processing
2200 }
2201}
2202
2203//----------------------------------------------------------------------
2204
2205/**
2206 * Return true if the given format character, occurring count
2207 * times, represents a numeric field.
2208 */
2209UBool SimpleDateFormat::isNumeric(char16_t formatChar, int32_t count) {
2210 return DateFormatSymbols::isNumericPatternChar(formatChar, count);
2211}
2212
2213UBool
2214SimpleDateFormat::isAtNumericField(const UnicodeString &pattern, int32_t patternOffset) {
2215 if (patternOffset >= pattern.length()) {
2216 // not at any field
2217 return false;
2218 }
2219 char16_t ch = pattern.charAt(patternOffset);
2220 UDateFormatField f = DateFormatSymbols::getPatternCharIndex(ch);
2221 if (f == UDAT_FIELD_COUNT) {
2222 // not at any field
2223 return false;
2224 }
2225 int32_t i = patternOffset;
2226 while (pattern.charAt(++i) == ch) {}
2227 return DateFormatSymbols::isNumericField(f, i - patternOffset);
2228}
2229
2230UBool
2231SimpleDateFormat::isAfterNonNumericField(const UnicodeString &pattern, int32_t patternOffset) {
2232 if (patternOffset <= 0) {
2233 // not after any field
2234 return false;
2235 }
2236 char16_t ch = pattern.charAt(--patternOffset);
2237 UDateFormatField f = DateFormatSymbols::getPatternCharIndex(ch);
2238 if (f == UDAT_FIELD_COUNT) {
2239 // not after any field
2240 return false;
2241 }
2242 int32_t i = patternOffset;
2243 while (pattern.charAt(--i) == ch) {}
2244 return !DateFormatSymbols::isNumericField(f, patternOffset - i);
2245}
2246
2247void
2248SimpleDateFormat::parse(const UnicodeString& text, Calendar& cal, ParsePosition& parsePos) const
2249{
2250 UErrorCode status = U_ZERO_ERROR;
2251 int32_t pos = parsePos.getIndex();
2252 if(parsePos.getIndex() < 0) {
2253 parsePos.setErrorIndex(0);
2254 return;
2255 }
2256 int32_t start = pos;
2257
2258 // Hold the day period until everything else is parsed, because we need
2259 // the hour to interpret time correctly.
2260 int32_t dayPeriodInt = -1;
2261
2262 UBool ambiguousYear[] = { false };
2263 int32_t saveHebrewMonth = -1;
2264 int32_t count = 0;
2265 UTimeZoneFormatTimeType tzTimeType = UTZFMT_TIME_TYPE_UNKNOWN;
2266
2267 // For parsing abutting numeric fields. 'abutPat' is the
2268 // offset into 'pattern' of the first of 2 or more abutting
2269 // numeric fields. 'abutStart' is the offset into 'text'
2270 // where parsing the fields begins. 'abutPass' starts off as 0
2271 // and increments each time we try to parse the fields.
2272 int32_t abutPat = -1; // If >=0, we are in a run of abutting numeric fields
2273 int32_t abutStart = 0;
2274 int32_t abutPass = 0;
2275 UBool inQuote = false;
2276
2277 MessageFormat * numericLeapMonthFormatter = nullptr;
2278
2279 Calendar* calClone = nullptr;
2280 Calendar *workCal = &cal;
2281 if (&cal != fCalendar && typeid(cal) != typeid(*fCalendar)) {
2282 // Different calendar type
2283 // We use the time/zone from the input calendar, but
2284 // do not use the input calendar for field calculation.
2285 calClone = fCalendar->clone();
2286 if (calClone != nullptr) {
2287 calClone->setTime(cal.getTime(status),status);
2288 if (U_FAILURE(status)) {
2289 goto ExitParse;
2290 }
2291 calClone->setTimeZone(cal.getTimeZone());
2292 workCal = calClone;
2293 } else {
2294 status = U_MEMORY_ALLOCATION_ERROR;
2295 goto ExitParse;
2296 }
2297 }
2298
2299 if (fSymbols->fLeapMonthPatterns != nullptr && fSymbols->fLeapMonthPatternsCount >= DateFormatSymbols::kMonthPatternsCount) {
2300 numericLeapMonthFormatter = new MessageFormat(fSymbols->fLeapMonthPatterns[DateFormatSymbols::kLeapMonthPatternNumeric], fLocale, status);
2301 if (numericLeapMonthFormatter == nullptr) {
2302 status = U_MEMORY_ALLOCATION_ERROR;
2303 goto ExitParse;
2304 } else if (U_FAILURE(status)) {
2305 goto ExitParse; // this will delete numericLeapMonthFormatter
2306 }
2307 }
2308
2309 for (int32_t i=0; i<fPattern.length(); ++i) {
2310 char16_t ch = fPattern.charAt(i);
2311
2312 // Handle alphabetic field characters.
2313 if (!inQuote && isSyntaxChar(ch)) {
2314 int32_t fieldPat = i;
2315
2316 // Count the length of this field specifier
2317 count = 1;
2318 while ((i+1)<fPattern.length() &&
2319 fPattern.charAt(i+1) == ch) {
2320 ++count;
2321 ++i;
2322 }
2323
2324 if (isNumeric(ch, count)) {
2325 if (abutPat < 0) {
2326 // Determine if there is an abutting numeric field.
2327 // Record the start of a set of abutting numeric fields.
2328 if (isAtNumericField(fPattern, i + 1)) {
2329 abutPat = fieldPat;
2330 abutStart = pos;
2331 abutPass = 0;
2332 }
2333 }
2334 } else {
2335 abutPat = -1; // End of any abutting fields
2336 }
2337
2338 // Handle fields within a run of abutting numeric fields. Take
2339 // the pattern "HHmmss" as an example. We will try to parse
2340 // 2/2/2 characters of the input text, then if that fails,
2341 // 1/2/2. We only adjust the width of the leftmost field; the
2342 // others remain fixed. This allows "123456" => 12:34:56, but
2343 // "12345" => 1:23:45. Likewise, for the pattern "yyyyMMdd" we
2344 // try 4/2/2, 3/2/2, 2/2/2, and finally 1/2/2.
2345 if (abutPat >= 0) {
2346 // If we are at the start of a run of abutting fields, then
2347 // shorten this field in each pass. If we can't shorten
2348 // this field any more, then the parse of this set of
2349 // abutting numeric fields has failed.
2350 if (fieldPat == abutPat) {
2351 count -= abutPass++;
2352 if (count == 0) {
2353 status = U_PARSE_ERROR;
2354 goto ExitParse;
2355 }
2356 }
2357
2358 pos = subParse(text, pos, ch, count,
2359 true, false, ambiguousYear, saveHebrewMonth, *workCal, i, numericLeapMonthFormatter, &tzTimeType);
2360
2361 // If the parse fails anywhere in the run, back up to the
2362 // start of the run and retry.
2363 if (pos < 0) {
2364 i = abutPat - 1;
2365 pos = abutStart;
2366 continue;
2367 }
2368 }
2369
2370 // Handle non-numeric fields and non-abutting numeric
2371 // fields.
2372 else if (ch != 0x6C) { // pattern char 'l' (SMALL LETTER L) just gets ignored
2373 int32_t s = subParse(text, pos, ch, count,
2374 false, true, ambiguousYear, saveHebrewMonth, *workCal, i, numericLeapMonthFormatter, &tzTimeType, &dayPeriodInt);
2375
2376 if (s == -pos-1) {
2377 // era not present, in special cases allow this to continue
2378 // from the position where the era was expected
2379 s = pos;
2380
2381 if (i+1 < fPattern.length()) {
2382 // move to next pattern character
2383 char16_t c = fPattern.charAt(i+1);
2384
2385 // check for whitespace
2386 if (PatternProps::isWhiteSpace(c)) {
2387 i++;
2388 // Advance over run in pattern
2389 while ((i+1)<fPattern.length() &&
2390 PatternProps::isWhiteSpace(fPattern.charAt(i+1))) {
2391 ++i;
2392 }
2393 }
2394 }
2395 }
2396 else if (s <= 0) {
2397 status = U_PARSE_ERROR;
2398 goto ExitParse;
2399 }
2400 pos = s;
2401 }
2402 }
2403
2404 // Handle literal pattern characters. These are any
2405 // quoted characters and non-alphabetic unquoted
2406 // characters.
2407 else {
2408
2409 abutPat = -1; // End of any abutting fields
2410
2411 if (! matchLiterals(fPattern, i, text, pos, getBooleanAttribute(UDAT_PARSE_ALLOW_WHITESPACE, status), getBooleanAttribute(UDAT_PARSE_PARTIAL_LITERAL_MATCH, status), isLenient())) {
2412 status = U_PARSE_ERROR;
2413 goto ExitParse;
2414 }
2415 }
2416 }
2417
2418 // Special hack for trailing "." after non-numeric field.
2419 if (text.charAt(pos) == 0x2e && getBooleanAttribute(UDAT_PARSE_ALLOW_WHITESPACE, status)) {
2420 // only do if the last field is not numeric
2421 if (isAfterNonNumericField(fPattern, fPattern.length())) {
2422 pos++; // skip the extra "."
2423 }
2424 }
2425
2426 // If dayPeriod is set, use it in conjunction with hour-of-day to determine am/pm.
2427 if (dayPeriodInt >= 0) {
2428 DayPeriodRules::DayPeriod dayPeriod = static_cast<DayPeriodRules::DayPeriod>(dayPeriodInt);
2429 const DayPeriodRules *ruleSet = DayPeriodRules::getInstance(this->getSmpFmtLocale(), status);
2430
2431 if (!cal.isSet(UCAL_HOUR) && !cal.isSet(UCAL_HOUR_OF_DAY)) {
2432 // If hour is not set, set time to the midpoint of current day period, overwriting
2433 // minutes if it's set.
2434 double midPoint = ruleSet->getMidPointForDayPeriod(dayPeriod, status);
2435
2436 // If we can't get midPoint we do nothing.
2437 if (U_SUCCESS(status)) {
2438 // Truncate midPoint toward zero to get the hour.
2439 // Any leftover means it was a half-hour.
2440 int32_t midPointHour = static_cast<int32_t>(midPoint);
2441 int32_t midPointMinute = (midPoint - midPointHour) > 0 ? 30 : 0;
2442
2443 // No need to set am/pm because hour-of-day is set last therefore takes precedence.
2444 cal.set(UCAL_HOUR_OF_DAY, midPointHour);
2445 cal.set(UCAL_MINUTE, midPointMinute);
2446 }
2447 } else {
2448 int hourOfDay;
2449
2450 if (cal.isSet(UCAL_HOUR_OF_DAY)) { // Hour is parsed in 24-hour format.
2451 hourOfDay = cal.get(UCAL_HOUR_OF_DAY, status);
2452 } else { // Hour is parsed in 12-hour format.
2453 hourOfDay = cal.get(UCAL_HOUR, status);
2454 // cal.get() turns 12 to 0 for 12-hour time; change 0 to 12
2455 // so 0 unambiguously means a 24-hour time from above.
2456 if (hourOfDay == 0) { hourOfDay = 12; }
2457 }
2458 U_ASSERT(0 <= hourOfDay && hourOfDay <= 23)(static_cast <bool> (0 <= hourOfDay && hourOfDay
<= 23) ? void (0) : __assert_fail ("0 <= hourOfDay && hourOfDay <= 23"
, __builtin_FILE (), __builtin_LINE (), __extension__ __PRETTY_FUNCTION__
));
2459
2460
2461 // If hour-of-day is 0 or 13 thru 23 then input time in unambiguously in 24-hour format.
2462 if (hourOfDay == 0 || (13 <= hourOfDay && hourOfDay <= 23)) {
2463 // Make hour-of-day take precedence over (hour + am/pm) by setting it again.
2464 cal.set(UCAL_HOUR_OF_DAY, hourOfDay);
2465 } else {
2466 // We have a 12-hour time and need to choose between am and pm.
2467 // Behave as if dayPeriod spanned 6 hours each way from its center point.
2468 // This will parse correctly for consistent time + period (e.g. 10 at night) as
2469 // well as provide a reasonable recovery for inconsistent time + period (e.g.
2470 // 9 in the afternoon).
2471
2472 // Assume current time is in the AM.
2473 // - Change 12 back to 0 for easier handling of 12am.
2474 // - Append minutes as fractional hours because e.g. 8:15 and 8:45 could be parsed
2475 // into different half-days if center of dayPeriod is at 14:30.
2476 // - cal.get(MINUTE) will return 0 if MINUTE is unset, which works.
2477 if (hourOfDay == 12) { hourOfDay = 0; }
2478 double currentHour = hourOfDay + (cal.get(UCAL_MINUTE, status)) / 60.0;
2479 double midPointHour = ruleSet->getMidPointForDayPeriod(dayPeriod, status);
2480
2481 if (U_SUCCESS(status)) {
2482 double hoursAheadMidPoint = currentHour - midPointHour;
2483
2484 // Assume current time is in the AM.
2485 if (-6 <= hoursAheadMidPoint && hoursAheadMidPoint < 6) {
2486 // Assumption holds; set time as such.
2487 cal.set(UCAL_AM_PM, 0);
2488 } else {
2489 cal.set(UCAL_AM_PM, 1);
2490 }
2491 }
2492 }
2493 }
2494 }
2495
2496 // At this point the fields of Calendar have been set. Calendar
2497 // will fill in default values for missing fields when the time
2498 // is computed.
2499
2500 parsePos.setIndex(pos);
2501
2502 // This part is a problem: When we call parsedDate.after, we compute the time.
2503 // Take the date April 3 2004 at 2:30 am. When this is first set up, the year
2504 // will be wrong if we're parsing a 2-digit year pattern. It will be 1904.
2505 // April 3 1904 is a Sunday (unlike 2004) so it is the DST onset day. 2:30 am
2506 // is therefore an "impossible" time, since the time goes from 1:59 to 3:00 am
2507 // on that day. It is therefore parsed out to fields as 3:30 am. Then we
2508 // add 100 years, and get April 3 2004 at 3:30 am. Note that April 3 2004 is
2509 // a Saturday, so it can have a 2:30 am -- and it should. [LIU]
2510 /*
2511 UDate parsedDate = calendar.getTime();
2512 if( ambiguousYear[0] && !parsedDate.after(fDefaultCenturyStart) ) {
2513 calendar.add(Calendar.YEAR, 100);
2514 parsedDate = calendar.getTime();
2515 }
2516 */
2517 // Because of the above condition, save off the fields in case we need to readjust.
2518 // The procedure we use here is not particularly efficient, but there is no other
2519 // way to do this given the API restrictions present in Calendar. We minimize
2520 // inefficiency by only performing this computation when it might apply, that is,
2521 // when the two-digit year is equal to the start year, and thus might fall at the
2522 // front or the back of the default century. This only works because we adjust
2523 // the year correctly to start with in other cases -- see subParse().
2524 if (ambiguousYear[0] || tzTimeType != UTZFMT_TIME_TYPE_UNKNOWN) // If this is true then the two-digit year == the default start year
2525 {
2526 // We need a copy of the fields, and we need to avoid triggering a call to
2527 // complete(), which will recalculate the fields. Since we can't access
2528 // the fields[] array in Calendar, we clone the entire object. This will
2529 // stop working if Calendar.clone() is ever rewritten to call complete().
2530 Calendar *copy;
2531 if (ambiguousYear[0]) {
2532 copy = cal.clone();
2533 // Check for failed cloning.
2534 if (copy == nullptr) {
2535 status = U_MEMORY_ALLOCATION_ERROR;
2536 goto ExitParse;
2537 }
2538 UDate parsedDate = copy->getTime(status);
2539 // {sfb} check internalGetDefaultCenturyStart
2540 if (fHaveDefaultCentury && (parsedDate < fDefaultCenturyStart)) {
2541 // We can't use add here because that does a complete() first.
2542 cal.set(UCAL_YEAR, fDefaultCenturyStartYear + 100);
2543 }
2544 delete copy;
2545 }
2546
2547 if (tzTimeType != UTZFMT_TIME_TYPE_UNKNOWN) {
2548 copy = cal.clone();
2549 // Check for failed cloning.
2550 if (copy == nullptr) {
2551 status = U_MEMORY_ALLOCATION_ERROR;
2552 goto ExitParse;
2553 }
2554 const TimeZone & tz = cal.getTimeZone();
2555 BasicTimeZone *btz = nullptr;
2556
2557 if (dynamic_cast<const OlsonTimeZone *>(&tz) != nullptr
2558 || dynamic_cast<const SimpleTimeZone *>(&tz) != nullptr
2559 || dynamic_cast<const RuleBasedTimeZone *>(&tz) != nullptr
2560 || dynamic_cast<const VTimeZone *>(&tz) != nullptr) {
2561 btz = (BasicTimeZone*)&tz;
2562 }
2563
2564 // Get local millis
2565 copy->set(UCAL_ZONE_OFFSET, 0);
2566 copy->set(UCAL_DST_OFFSET, 0);
2567 UDate localMillis = copy->getTime(status);
2568
2569 // Make sure parsed time zone type (Standard or Daylight)
2570 // matches the rule used by the parsed time zone.
2571 int32_t raw, dst;
2572 if (btz != nullptr) {
2573 if (tzTimeType == UTZFMT_TIME_TYPE_STANDARD) {
2574 btz->getOffsetFromLocal(localMillis,
2575 UCAL_TZ_LOCAL_STANDARD_FORMER, UCAL_TZ_LOCAL_STANDARD_LATTER, raw, dst, status);
2576 } else {
2577 btz->getOffsetFromLocal(localMillis,
2578 UCAL_TZ_LOCAL_DAYLIGHT_FORMER, UCAL_TZ_LOCAL_DAYLIGHT_LATTER, raw, dst, status);
2579 }
2580 } else {
2581 // No good way to resolve ambiguous time at transition,
2582 // but following code work in most case.
2583 tz.getOffset(localMillis, true, raw, dst, status);
2584 }
2585
2586 // Now, compare the results with parsed type, either standard or daylight saving time
2587 int32_t resolvedSavings = dst;
2588 if (tzTimeType == UTZFMT_TIME_TYPE_STANDARD) {
2589 if (dst != 0) {
2590 // Override DST_OFFSET = 0 in the result calendar
2591 resolvedSavings = 0;
2592 }
2593 } else { // tztype == TZTYPE_DST
2594 if (dst == 0) {
2595 if (btz != nullptr) {
2596 // This implementation resolves daylight saving time offset
2597 // closest rule after the given time.
2598 UDate baseTime = localMillis + raw;
2599 UDate time = baseTime;
2600 UDate limit = baseTime + MAX_DAYLIGHT_DETECTION_RANGE;
2601 TimeZoneTransition trs;
2602 UBool trsAvail;
2603
2604 // Search for DST rule after the given time
2605 while (time < limit) {
2606 trsAvail = btz->getNextTransition(time, false, trs);
2607 if (!trsAvail) {
2608 break;
2609 }
2610 resolvedSavings = trs.getTo()->getDSTSavings();
2611 if (resolvedSavings != 0) {
2612 break;
2613 }
2614 time = trs.getTime();
2615 }
2616
2617 if (resolvedSavings == 0) {
2618 // If no DST rule after the given time was found, search for
2619 // DST rule before.
2620 time = baseTime;
2621 limit = baseTime - MAX_DAYLIGHT_DETECTION_RANGE;
2622 while (time > limit) {
2623 trsAvail = btz->getPreviousTransition(time, true, trs);
2624 if (!trsAvail) {
2625 break;
2626 }
2627 resolvedSavings = trs.getFrom()->getDSTSavings();
2628 if (resolvedSavings != 0) {
2629 break;
2630 }
2631 time = trs.getTime() - 1;
2632 }
2633
2634 if (resolvedSavings == 0) {
2635 resolvedSavings = btz->getDSTSavings();
2636 }
2637 }
2638 } else {
2639 resolvedSavings = tz.getDSTSavings();
2640 }
2641 if (resolvedSavings == 0) {
2642 // final fallback
2643 resolvedSavings = U_MILLIS_PER_HOUR(3600000);
2644 }
2645 }
2646 }
2647 cal.set(UCAL_ZONE_OFFSET, raw);
2648 cal.set(UCAL_DST_OFFSET, resolvedSavings);
2649 delete copy;
2650 }
2651 }
2652ExitParse:
2653 // Set the parsed result if local calendar is used
2654 // instead of the input calendar
2655 if (U_SUCCESS(status) && workCal != &cal) {
2656 cal.setTimeZone(workCal->getTimeZone());
2657 cal.setTime(workCal->getTime(status), status);
2658 }
2659
2660 delete numericLeapMonthFormatter;
2661 delete calClone;
2662
2663 // If any Calendar calls failed, we pretend that we
2664 // couldn't parse the string, when in reality this isn't quite accurate--
2665 // we did parse it; the Calendar calls just failed.
2666 if (U_FAILURE(status)) {
2667 parsePos.setErrorIndex(pos);
2668 parsePos.setIndex(start);
2669 }
2670}
2671
2672//----------------------------------------------------------------------
2673
2674static int32_t
2675matchStringWithOptionalDot(const UnicodeString &text,
2676 int32_t index,
2677 const UnicodeString &data);
2678
2679int32_t SimpleDateFormat::matchQuarterString(const UnicodeString& text,
2680 int32_t start,
2681 UCalendarDateFields field,
2682 const UnicodeString* data,
2683 int32_t dataCount,
2684 Calendar& cal) const
2685{
2686 int32_t i = 0;
2687 int32_t count = dataCount;
2688
2689 // There may be multiple strings in the data[] array which begin with
2690 // the same prefix (e.g., Cerven and Cervenec (June and July) in Czech).
2691 // We keep track of the longest match, and return that. Note that this
2692 // unfortunately requires us to test all array elements.
2693 int32_t bestMatchLength = 0, bestMatch = -1;
2694 UnicodeString bestMatchName;
2695
2696 for (; i < count; ++i) {
2697 int32_t matchLength = 0;
2698 if ((matchLength = matchStringWithOptionalDot(text, start, data[i])) > bestMatchLength) {
2699 bestMatchLength = matchLength;
2700 bestMatch = i;
2701 }
2702 }
2703
2704 if (bestMatch >= 0) {
2705 cal.set(field, bestMatch * 3);
2706 return start + bestMatchLength;
2707 }
2708
2709 return -start;
2710}
2711
2712int32_t SimpleDateFormat::matchDayPeriodStrings(const UnicodeString& text, int32_t start,
2713 const UnicodeString* data, int32_t dataCount,
2714 int32_t &dayPeriod) const
2715{
2716
2717 int32_t bestMatchLength = 0, bestMatch = -1;
2718
2719 for (int32_t i = 0; i < dataCount; ++i) {
2720 int32_t matchLength = 0;
2721 if ((matchLength = matchStringWithOptionalDot(text, start, data[i])) > bestMatchLength) {
2722 bestMatchLength = matchLength;
2723 bestMatch = i;
2724 }
2725 }
2726
2727 if (bestMatch >= 0) {
2728 dayPeriod = bestMatch;
2729 return start + bestMatchLength;
2730 }
2731
2732 return -start;
2733}
2734
2735//----------------------------------------------------------------------
2736UBool SimpleDateFormat::matchLiterals(const UnicodeString &pattern,
2737 int32_t &patternOffset,
2738 const UnicodeString &text,
2739 int32_t &textOffset,
2740 UBool whitespaceLenient,
2741 UBool partialMatchLenient,
2742 UBool oldLeniency)
2743{
2744 UBool inQuote = false;
2745 UnicodeString literal;
2746 int32_t i = patternOffset;
2747
2748 // scan pattern looking for contiguous literal characters
2749 for ( ; i < pattern.length(); i += 1) {
2750 char16_t ch = pattern.charAt(i);
2751
2752 if (!inQuote && isSyntaxChar(ch)) {
2753 break;
2754 }
2755
2756 if (ch == QUOTE) {
2757 // Match a quote literal ('') inside OR outside of quotes
2758 if ((i + 1) < pattern.length() && pattern.charAt(i + 1) == QUOTE) {
2759 i += 1;
2760 } else {
2761 inQuote = !inQuote;
2762 continue;
2763 }
2764 }
2765
2766 literal += ch;
2767 }
2768
2769 // at this point, literal contains the literal text
2770 // and i is the index of the next non-literal pattern character.
2771 int32_t p;
2772 int32_t t = textOffset;
2773
2774 if (whitespaceLenient) {
2775 // trim leading, trailing whitespace from
2776 // the literal text
2777 literal.trim();
2778
2779 // ignore any leading whitespace in the text
2780 while (t < text.length() && u_isWhitespaceu_isWhitespace_77(text.charAt(t))) {
2781 t += 1;
2782 }
2783 }
2784
2785 for (p = 0; p < literal.length() && t < text.length();) {
2786 UBool needWhitespace = false;
2787
2788 while (p < literal.length() && PatternProps::isWhiteSpace(literal.charAt(p))) {
2789 needWhitespace = true;
2790 p += 1;
2791 }
2792
2793 if (needWhitespace) {
2794 int32_t tStart = t;
2795
2796 while (t < text.length()) {
2797 char16_t tch = text.charAt(t);
2798
2799 if (!u_isUWhiteSpaceu_isUWhiteSpace_77(tch) && !PatternProps::isWhiteSpace(tch)) {
2800 break;
2801 }
2802
2803 t += 1;
2804 }
2805
2806 // TODO: should we require internal spaces
2807 // in lenient mode? (There won't be any
2808 // leading or trailing spaces)
2809 if (!whitespaceLenient && t == tStart) {
2810 // didn't find matching whitespace:
2811 // an error in strict mode
2812 return false;
2813 }
2814
2815 // In strict mode, this run of whitespace
2816 // may have been at the end.
2817 if (p >= literal.length()) {
2818 break;
2819 }
2820 }
2821 if (t >= text.length() || literal.charAt(p) != text.charAt(t)) {
2822 // Ran out of text, or found a non-matching character:
2823 // OK in lenient mode, an error in strict mode.
2824 if (whitespaceLenient) {
2825 if (t == textOffset && text.charAt(t) == 0x2e &&
2826 isAfterNonNumericField(pattern, patternOffset)) {
2827 // Lenient mode and the literal input text begins with a "." and
2828 // we are after a non-numeric field: We skip the "."
2829 ++t;
2830 continue; // Do not update p.
2831 }
2832 // if it is actual whitespace and we're whitespace lenient it's OK
2833
2834 char16_t wsc = text.charAt(t);
2835 if(PatternProps::isWhiteSpace(wsc)) {
2836 // Lenient mode and it's just whitespace we skip it
2837 ++t;
2838 continue; // Do not update p.
2839 }
2840 }
2841 // hack around oldleniency being a bit of a catch-all bucket and we're just adding support specifically for partial matches
2842 if(partialMatchLenient && oldLeniency) {
2843 break;
2844 }
2845
2846 return false;
2847 }
2848 ++p;
2849 ++t;
2850 }
2851
2852 // At this point if we're in strict mode we have a complete match.
2853 // If we're in lenient mode we may have a partial match, or no
2854 // match at all.
2855 if (p <= 0) {
2856 // no match. Pretend it matched a run of whitespace
2857 // and ignorables in the text.
2858 const UnicodeSet *ignorables = nullptr;
2859 UDateFormatField patternCharIndex = DateFormatSymbols::getPatternCharIndex(pattern.charAt(i));
2860 if (patternCharIndex != UDAT_FIELD_COUNT) {
2861 ignorables = SimpleDateFormatStaticSets::getIgnorables(patternCharIndex);
2862 }
2863
2864 for (t = textOffset; t < text.length(); t += 1) {
2865 char16_t ch = text.charAt(t);
2866
2867 if (ignorables == nullptr || !ignorables->contains(ch)) {
2868 break;
2869 }
2870 }
2871 }
2872
2873 // if we get here, we've got a complete match.
2874 patternOffset = i - 1;
2875 textOffset = t;
2876
2877 return true;
2878}
2879
2880//----------------------------------------------------------------------
2881// check both wide and abbrev months.
2882// Does not currently handle monthPattern.
2883// UCalendarDateFields field = UCAL_MONTH
2884
2885int32_t SimpleDateFormat::matchAlphaMonthStrings(const UnicodeString& text,
2886 int32_t start,
2887 const UnicodeString* wideData,
2888 const UnicodeString* shortData,
2889 int32_t dataCount,
2890 Calendar& cal) const
2891{
2892 int32_t i;
2893 int32_t bestMatchLength = 0, bestMatch = -1;
2894
2895 for (i = 0; i < dataCount; ++i) {
2896 int32_t matchLen = 0;
2897 if ((matchLen = matchStringWithOptionalDot(text, start, wideData[i])) > bestMatchLength) {
2898 bestMatch = i;
2899 bestMatchLength = matchLen;
2900 }
2901 }
2902 for (i = 0; i < dataCount; ++i) {
2903 int32_t matchLen = 0;
2904 if ((matchLen = matchStringWithOptionalDot(text, start, shortData[i])) > bestMatchLength) {
2905 bestMatch = i;
2906 bestMatchLength = matchLen;
2907 }
2908 }
2909
2910 if (bestMatch >= 0) {
2911 // Adjustment for Hebrew Calendar month Adar II
2912 if (typeid(cal) == typeid(HebrewCalendar) && bestMatch==13) {
2913 cal.set(UCAL_MONTH,6);
2914 } else {
2915 cal.set(UCAL_MONTH, bestMatch);
2916 }
2917 return start + bestMatchLength;
2918 }
2919
2920 return -start;
2921}
2922
2923//----------------------------------------------------------------------
2924
2925int32_t SimpleDateFormat::matchString(const UnicodeString& text,
2926 int32_t start,
2927 UCalendarDateFields field,
2928 const UnicodeString* data,
2929 int32_t dataCount,
2930 const UnicodeString* monthPattern,
2931 Calendar& cal) const
2932{
2933 int32_t i = 0;
2934 int32_t count = dataCount;
2935
2936 if (field == UCAL_DAY_OF_WEEK) i = 1;
2937
2938 // There may be multiple strings in the data[] array which begin with
2939 // the same prefix (e.g., Cerven and Cervenec (June and July) in Czech).
2940 // We keep track of the longest match, and return that. Note that this
2941 // unfortunately requires us to test all array elements.
2942 // But this does not really work for cases such as Chuvash in which
2943 // May is "ҫу" and August is "ҫурла"/"ҫур.", hence matchAlphaMonthStrings.
2944 int32_t bestMatchLength = 0, bestMatch = -1;
2945 UnicodeString bestMatchName;
2946 int32_t isLeapMonth = 0;
2947
2948 for (; i < count; ++i) {
2949 int32_t matchLen = 0;
2950 if ((matchLen = matchStringWithOptionalDot(text, start, data[i])) > bestMatchLength) {
2951 bestMatch = i;
2952 bestMatchLength = matchLen;
2953 }
2954
2955 if (monthPattern != nullptr) {
2956 UErrorCode status = U_ZERO_ERROR;
2957 UnicodeString leapMonthName;
2958 SimpleFormatter(*monthPattern, 1, 1, status).format(data[i], leapMonthName, status);
2959 if (U_SUCCESS(status)) {
2960 if ((matchLen = matchStringWithOptionalDot(text, start, leapMonthName)) > bestMatchLength) {
2961 bestMatch = i;
2962 bestMatchLength = matchLen;
2963 isLeapMonth = 1;
2964 }
2965 }
2966 }
2967 }
2968
2969 if (bestMatch >= 0) {
2970 if (field < UCAL_FIELD_COUNT) {
2971 // Adjustment for Hebrew Calendar month Adar II
2972 if (typeid(cal) == typeid(HebrewCalendar) && field==UCAL_MONTH && bestMatch==13) {
2973 cal.set(field,6);
2974 } else {
2975 if (field == UCAL_YEAR) {
2976 bestMatch++; // only get here for cyclic year names, which match 1-based years 1-60
2977 }
2978 cal.set(field, bestMatch);
2979 }
2980 if (monthPattern != nullptr) {
2981 cal.set(UCAL_IS_LEAP_MONTH, isLeapMonth);
2982 }
2983 }
2984
2985 return start + bestMatchLength;
2986 }
2987
2988 return -start;
2989}
2990
2991static int32_t
2992matchStringWithOptionalDot(const UnicodeString &text,
2993 int32_t index,
2994 const UnicodeString &data) {
2995 UErrorCode sts = U_ZERO_ERROR;
2996 int32_t matchLenText = 0;
2997 int32_t matchLenData = 0;
2998
2999 u_caseInsensitivePrefixMatchu_caseInsensitivePrefixMatch_77(text.getBuffer() + index, text.length() - index,
3000 data.getBuffer(), data.length(),
3001 0 /* default case option */,
3002 &matchLenText, &matchLenData,
3003 &sts);
3004 U_ASSERT (U_SUCCESS(sts))(static_cast <bool> (U_SUCCESS(sts)) ? void (0) : __assert_fail
("U_SUCCESS(sts)", __builtin_FILE (), __builtin_LINE (), __extension__
__PRETTY_FUNCTION__));
3005
3006 if (matchLenData == data.length() /* normal match */
3007 || (data.charAt(data.length() - 1) == 0x2e
3008 && matchLenData == data.length() - 1 /* match without trailing dot */)) {
3009 return matchLenText;
3010 }
3011
3012 return 0;
3013}
3014
3015//----------------------------------------------------------------------
3016
3017void
3018SimpleDateFormat::set2DigitYearStart(UDate d, UErrorCode& status)
3019{
3020 parseAmbiguousDatesAsAfter(d, status);
3021}
3022
3023/**
3024 * Private member function that converts the parsed date strings into
3025 * timeFields. Returns -start (for ParsePosition) if failed.
3026 */
3027int32_t SimpleDateFormat::subParse(const UnicodeString& text, int32_t& start, char16_t ch, int32_t count,
3028 UBool obeyCount, UBool allowNegative, UBool ambiguousYear[], int32_t& saveHebrewMonth, Calendar& cal,
3029 int32_t patLoc, MessageFormat * numericLeapMonthFormatter, UTimeZoneFormatTimeType *tzTimeType,
3030 int32_t *dayPeriod) const
3031{
3032 Formattable number;
3033 int32_t value = 0;
3034 int32_t i;
3035 int32_t ps = 0;
3036 UErrorCode status = U_ZERO_ERROR;
3037 ParsePosition pos(0);
3038 UDateFormatField patternCharIndex = DateFormatSymbols::getPatternCharIndex(ch);
3039 const NumberFormat *currentNumberFormat;
3040 UnicodeString temp;
3041 UBool gotNumber = false;
3042
3043#if defined (U_DEBUG_CAL)
3044 //fprintf(stderr, "%s:%d - [%c] st=%d \n", __FILE__, __LINE__, (char) ch, start);
3045#endif
3046
3047 if (patternCharIndex == UDAT_FIELD_COUNT) {
3048 return -start;
3049 }
3050
3051 currentNumberFormat = getNumberFormatByIndex(patternCharIndex);
3052 if (currentNumberFormat == nullptr) {
3053 return -start;
3054 }
3055 UCalendarDateFields field = fgPatternIndexToCalendarField[patternCharIndex]; // UCAL_FIELD_COUNT if irrelevant
3056 UnicodeString hebr("hebr", 4, US_INVicu::UnicodeString::kInvariant);
3057
3058 if (numericLeapMonthFormatter != nullptr) {
3059 numericLeapMonthFormatter->setFormats(reinterpret_cast<const Format**>(&currentNumberFormat), 1);
3060 }
3061
3062 // If there are any spaces here, skip over them. If we hit the end
3063 // of the string, then fail.
3064 for (;;) {
3065 if (start >= text.length()) {
3066 return -start;
3067 }
3068 UChar32 c = text.char32At(start);
3069 if (!u_isUWhiteSpaceu_isUWhiteSpace_77(c) /*||*/ && !PatternProps::isWhiteSpace(c)) {
3070 break;
3071 }
3072 start += U16_LENGTH(c)((uint32_t)(c)<=0xffff ? 1 : 2);
3073 }
3074 pos.setIndex(start);
3075
3076 UBool isChineseCalendar = typeid(cal) == typeid(ChineseCalendar) ||
3077 typeid(cal) == typeid(DangiCalendar);
3078 // We handle a few special cases here where we need to parse
3079 // a number value. We handle further, more generic cases below. We need
3080 // to handle some of them here because some fields require extra processing on
3081 // the parsed value.
3082 if (patternCharIndex == UDAT_HOUR_OF_DAY1_FIELD || // k
3083 patternCharIndex == UDAT_HOUR_OF_DAY0_FIELD || // H
3084 patternCharIndex == UDAT_HOUR1_FIELD || // h
3085 patternCharIndex == UDAT_HOUR0_FIELD || // K
3086 (patternCharIndex == UDAT_DOW_LOCAL_FIELD && count <= 2) || // e
3087 (patternCharIndex == UDAT_STANDALONE_DAY_FIELD && count <= 2) || // c
3088 (patternCharIndex == UDAT_MONTH_FIELD && count <= 2) || // M
3089 (patternCharIndex == UDAT_STANDALONE_MONTH_FIELD && count <= 2) || // L
3090 (patternCharIndex == UDAT_QUARTER_FIELD && count <= 2) || // Q
3091 (patternCharIndex == UDAT_STANDALONE_QUARTER_FIELD && count <= 2) || // q
3092 patternCharIndex == UDAT_YEAR_FIELD || // y
3093 patternCharIndex == UDAT_YEAR_WOY_FIELD || // Y
3094 patternCharIndex == UDAT_YEAR_NAME_FIELD || // U (falls back to numeric)
3095 (patternCharIndex == UDAT_ERA_FIELD && isChineseCalendar) || // G
3096 patternCharIndex == UDAT_FRACTIONAL_SECOND_FIELD) // S
3097 {
3098 int32_t parseStart = pos.getIndex();
3099 // It would be good to unify this with the obeyCount logic below,
3100 // but that's going to be difficult.
3101 const UnicodeString* src;
3102
3103 UBool parsedNumericLeapMonth = false;
3104 if (numericLeapMonthFormatter != nullptr && (patternCharIndex == UDAT_MONTH_FIELD || patternCharIndex == UDAT_STANDALONE_MONTH_FIELD)) {
3105 int32_t argCount;
3106 Formattable * args = numericLeapMonthFormatter->parse(text, pos, argCount);
3107 if (args != nullptr && argCount == 1 && pos.getIndex() > parseStart && args[0].isNumeric()) {
3108 parsedNumericLeapMonth = true;
3109 number.setLong(args[0].getLong());
3110 cal.set(UCAL_IS_LEAP_MONTH, 1);
3111 delete[] args;
3112 } else {
3113 pos.setIndex(parseStart);
3114 cal.set(UCAL_IS_LEAP_MONTH, 0);
3115 }
3116 }
3117
3118 if (!parsedNumericLeapMonth) {
3119 if (obeyCount) {
3120 if ((start+count) > text.length()) {
3121 return -start;
3122 }
3123
3124 text.extractBetween(0, start + count, temp);
3125 src = &temp;
3126 } else {
3127 src = &text;
3128 }
3129
3130 parseInt(*src, number, pos, allowNegative,currentNumberFormat);
3131 }
3132
3133 int32_t txtLoc = pos.getIndex();
3134
3135 if (txtLoc > parseStart) {
3136 value = number.getLong();
3137 gotNumber = true;
3138
3139 // suffix processing
3140 if (value < 0 ) {
3141 txtLoc = checkIntSuffix(text, txtLoc, patLoc+1, true);
3142 if (txtLoc != pos.getIndex()) {
3143 value *= -1;
3144 }
3145 }
3146 else {
3147 txtLoc = checkIntSuffix(text, txtLoc, patLoc+1, false);
3148 }
3149
3150 if (!getBooleanAttribute(UDAT_PARSE_ALLOW_WHITESPACE, status)) {
3151 // Check the range of the value
3152 int32_t bias = gFieldRangeBias[patternCharIndex];
3153 if (bias >= 0 && (value > cal.getMaximum(field) + bias || value < cal.getMinimum(field) + bias)) {
3154 return -start;
3155 }
3156 }
3157
3158 pos.setIndex(txtLoc);
3159 }
3160 }
3161
3162 // Make sure that we got a number if
3163 // we want one, and didn't get one
3164 // if we don't want one.
3165 switch (patternCharIndex) {
3166 case UDAT_HOUR_OF_DAY1_FIELD:
3167 case UDAT_HOUR_OF_DAY0_FIELD:
3168 case UDAT_HOUR1_FIELD:
3169 case UDAT_HOUR0_FIELD:
3170 // special range check for hours:
3171 if (value < 0 || value > 24) {
3172 return -start;
3173 }
3174
3175 // fall through to gotNumber check
3176 U_FALLTHROUGH[[clang::fallthrough]];
3177 case UDAT_YEAR_FIELD:
3178 case UDAT_YEAR_WOY_FIELD:
3179 case UDAT_FRACTIONAL_SECOND_FIELD:
3180 // these must be a number
3181 if (! gotNumber) {
3182 return -start;
3183 }
3184
3185 break;
3186
3187 default:
3188 // we check the rest of the fields below.
3189 break;
3190 }
3191
3192 switch (patternCharIndex) {
3193 case UDAT_ERA_FIELD:
3194 if (isChineseCalendar) {
3195 if (!gotNumber) {
3196 return -start;
3197 }
3198 cal.set(UCAL_ERA, value);
3199 return pos.getIndex();
3200 }
3201 if (count == 5) {
3202 ps = matchString(text, start, UCAL_ERA, fSymbols->fNarrowEras, fSymbols->fNarrowErasCount, nullptr, cal);
3203 } else if (count == 4) {
3204 ps = matchString(text, start, UCAL_ERA, fSymbols->fEraNames, fSymbols->fEraNamesCount, nullptr, cal);
3205 } else {
3206 ps = matchString(text, start, UCAL_ERA, fSymbols->fEras, fSymbols->fErasCount, nullptr, cal);
3207 }
3208
3209 // check return position, if it equals -start, then matchString error
3210 // special case the return code so we don't necessarily fail out until we
3211 // verify no year information also
3212 if (ps == -start)
3213 ps--;
3214
3215 return ps;
3216
3217 case UDAT_YEAR_FIELD:
3218 // If there are 3 or more YEAR pattern characters, this indicates
3219 // that the year value is to be treated literally, without any
3220 // two-digit year adjustments (e.g., from "01" to 2001). Otherwise
3221 // we made adjustments to place the 2-digit year in the proper
3222 // century, for parsed strings from "00" to "99". Any other string
3223 // is treated literally: "2250", "-1", "1", "002".
3224 if (fDateOverride.compare(hebr)==0 && value < 1000) {
3225 value += HEBREW_CAL_CUR_MILLENIUM_START_YEAR;
3226 } else if (text.moveIndex32(start, 2) == pos.getIndex() && !isChineseCalendar
3227 && u_isdigitu_isdigit_77(text.char32At(start))
3228 && u_isdigitu_isdigit_77(text.char32At(text.moveIndex32(start, 1))))
3229 {
3230 // only adjust year for patterns less than 3.
3231 if(count < 3) {
3232 // Assume for example that the defaultCenturyStart is 6/18/1903.
3233 // This means that two-digit years will be forced into the range
3234 // 6/18/1903 to 6/17/2003. As a result, years 00, 01, and 02
3235 // correspond to 2000, 2001, and 2002. Years 04, 05, etc. correspond
3236 // to 1904, 1905, etc. If the year is 03, then it is 2003 if the
3237 // other fields specify a date before 6/18, or 1903 if they specify a
3238 // date afterwards. As a result, 03 is an ambiguous year. All other
3239 // two-digit years are unambiguous.
3240 if(fHaveDefaultCentury) { // check if this formatter even has a pivot year
3241 int32_t ambiguousTwoDigitYear = fDefaultCenturyStartYear % 100;
3242 ambiguousYear[0] = (value == ambiguousTwoDigitYear);
3243 value += (fDefaultCenturyStartYear/100)*100 +
3244 (value < ambiguousTwoDigitYear ? 100 : 0);
3245 }
3246 }
3247 }
3248 cal.set(UCAL_YEAR, value);
3249
3250 // Delayed checking for adjustment of Hebrew month numbers in non-leap years.
3251 if (saveHebrewMonth >= 0) {
3252 HebrewCalendar *hc = (HebrewCalendar*)&cal;
3253 if (!hc->isLeapYear(value) && saveHebrewMonth >= 6) {
3254 cal.set(UCAL_MONTH,saveHebrewMonth);
3255 } else {
3256 cal.set(UCAL_MONTH,saveHebrewMonth-1);
3257 }
3258 saveHebrewMonth = -1;
3259 }
3260 return pos.getIndex();
3261
3262 case UDAT_YEAR_WOY_FIELD:
3263 // Comment is the same as for UDAT_Year_FIELDs - look above
3264 if (fDateOverride.compare(hebr)==0 && value < 1000) {
3265 value += HEBREW_CAL_CUR_MILLENIUM_START_YEAR;
3266 } else if (text.moveIndex32(start, 2) == pos.getIndex()
3267 && u_isdigitu_isdigit_77(text.char32At(start))
3268 && u_isdigitu_isdigit_77(text.char32At(text.moveIndex32(start, 1)))
3269 && fHaveDefaultCentury )
3270 {
3271 int32_t ambiguousTwoDigitYear = fDefaultCenturyStartYear % 100;
3272 ambiguousYear[0] = (value == ambiguousTwoDigitYear);
3273 value += (fDefaultCenturyStartYear/100)*100 +
3274 (value < ambiguousTwoDigitYear ? 100 : 0);
3275 }
3276 cal.set(UCAL_YEAR_WOY, value);
3277 return pos.getIndex();
3278
3279 case UDAT_YEAR_NAME_FIELD:
3280 if (fSymbols->fShortYearNames != nullptr) {
3281 int32_t newStart = matchString(text, start, UCAL_YEAR, fSymbols->fShortYearNames, fSymbols->fShortYearNamesCount, nullptr, cal);
3282 if (newStart > 0) {
3283 return newStart;
3284 }
3285 }
3286 if (gotNumber && (getBooleanAttribute(UDAT_PARSE_ALLOW_NUMERIC,status) || value > fSymbols->fShortYearNamesCount)) {
3287 cal.set(UCAL_YEAR, value);
3288 return pos.getIndex();
3289 }
3290 return -start;
3291
3292 case UDAT_MONTH_FIELD:
3293 case UDAT_STANDALONE_MONTH_FIELD:
3294 if (gotNumber) // i.e., M or MM.
3295 {
3296 // When parsing month numbers from the Hebrew Calendar, we might need to adjust the month depending on whether
3297 // 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
3298 // the year is parsed.
3299 if (typeid(cal) == typeid(HebrewCalendar)) {
3300 HebrewCalendar *hc = (HebrewCalendar*)&cal;
3301 if (cal.isSet(UCAL_YEAR)) {
3302 UErrorCode monthStatus = U_ZERO_ERROR;
3303 if (!hc->isLeapYear(hc->get(UCAL_YEAR, monthStatus)) && value >= 6) {
3304 cal.set(UCAL_MONTH, value);
3305 } else {
3306 cal.set(UCAL_MONTH, value - 1);
3307 }
3308 } else {
3309 saveHebrewMonth = value;
3310 }
3311 } else {
3312 // Don't want to parse the month if it is a string
3313 // while pattern uses numeric style: M/MM, L/LL
3314 // [We computed 'value' above.]
3315 cal.set(UCAL_MONTH, value - 1);
3316 }
3317 return pos.getIndex();
3318 } else {
3319 // count >= 3 // i.e., MMM/MMMM, LLL/LLLL
3320 // Want to be able to parse both short and long forms.
3321 // Try count == 4 first:
3322 UnicodeString * wideMonthPat = nullptr;
3323 UnicodeString * shortMonthPat = nullptr;
3324 if (fSymbols->fLeapMonthPatterns != nullptr && fSymbols->fLeapMonthPatternsCount >= DateFormatSymbols::kMonthPatternsCount) {
3325 if (patternCharIndex==UDAT_MONTH_FIELD) {
3326 wideMonthPat = &fSymbols->fLeapMonthPatterns[DateFormatSymbols::kLeapMonthPatternFormatWide];
3327 shortMonthPat = &fSymbols->fLeapMonthPatterns[DateFormatSymbols::kLeapMonthPatternFormatAbbrev];
3328 } else {
3329 wideMonthPat = &fSymbols->fLeapMonthPatterns[DateFormatSymbols::kLeapMonthPatternStandaloneWide];
3330 shortMonthPat = &fSymbols->fLeapMonthPatterns[DateFormatSymbols::kLeapMonthPatternStandaloneAbbrev];
3331 }
3332 }
3333 int32_t newStart = 0;
3334 if (patternCharIndex==UDAT_MONTH_FIELD) {
3335 if(getBooleanAttribute(UDAT_PARSE_MULTIPLE_PATTERNS_FOR_MATCH, status) && count>=3 && count <=4 &&
3336 fSymbols->fLeapMonthPatterns==nullptr && fSymbols->fMonthsCount==fSymbols->fShortMonthsCount) {
3337 // single function to check both wide and short, an experiment
3338 newStart = matchAlphaMonthStrings(text, start, fSymbols->fMonths, fSymbols->fShortMonths, fSymbols->fMonthsCount, cal); // try MMMM,MMM
3339 if (newStart > 0) {
3340 return newStart;
3341 }
3342 }
3343 if(getBooleanAttribute(UDAT_PARSE_MULTIPLE_PATTERNS_FOR_MATCH, status) || count == 4) {
3344 newStart = matchString(text, start, UCAL_MONTH, fSymbols->fMonths, fSymbols->fMonthsCount, wideMonthPat, cal); // try MMMM
3345 if (newStart > 0) {
3346 return newStart;
3347 }
3348 }
3349 if(getBooleanAttribute(UDAT_PARSE_MULTIPLE_PATTERNS_FOR_MATCH, status) || count == 3) {
3350 newStart = matchString(text, start, UCAL_MONTH, fSymbols->fShortMonths, fSymbols->fShortMonthsCount, shortMonthPat, cal); // try MMM
3351 }
3352 } else {
3353 if(getBooleanAttribute(UDAT_PARSE_MULTIPLE_PATTERNS_FOR_MATCH, status) && count>=3 && count <=4 &&
3354 fSymbols->fLeapMonthPatterns==nullptr && fSymbols->fStandaloneMonthsCount==fSymbols->fStandaloneShortMonthsCount) {
3355 // single function to check both wide and short, an experiment
3356 newStart = matchAlphaMonthStrings(text, start, fSymbols->fStandaloneMonths, fSymbols->fStandaloneShortMonths, fSymbols->fStandaloneMonthsCount, cal); // try MMMM,MMM
3357 if (newStart > 0) {
3358 return newStart;
3359 }
3360 }
3361 if(getBooleanAttribute(UDAT_PARSE_MULTIPLE_PATTERNS_FOR_MATCH, status) || count == 4) {
3362 newStart = matchString(text, start, UCAL_MONTH, fSymbols->fStandaloneMonths, fSymbols->fStandaloneMonthsCount, wideMonthPat, cal); // try LLLL
3363 if (newStart > 0) {
3364 return newStart;
3365 }
3366 }
3367 if(getBooleanAttribute(UDAT_PARSE_MULTIPLE_PATTERNS_FOR_MATCH, status) || count == 3) {
3368 newStart = matchString(text, start, UCAL_MONTH, fSymbols->fStandaloneShortMonths, fSymbols->fStandaloneShortMonthsCount, shortMonthPat, cal); // try LLL
3369 }
3370 }
3371 if (newStart > 0 || !getBooleanAttribute(UDAT_PARSE_ALLOW_NUMERIC, status)) // currently we do not try to parse MMMMM/LLLLL: #8860
3372 return newStart;
3373 // else we allowing parsing as number, below
3374 }
3375 break;
3376
3377 case UDAT_HOUR_OF_DAY1_FIELD:
3378 // [We computed 'value' above.]
3379 if (value == cal.getMaximum(UCAL_HOUR_OF_DAY) + 1)
3380 value = 0;
3381
3382 // fall through to set field
3383 U_FALLTHROUGH[[clang::fallthrough]];
3384 case UDAT_HOUR_OF_DAY0_FIELD:
3385 cal.set(UCAL_HOUR_OF_DAY, value);
3386 return pos.getIndex();
3387
3388 case UDAT_FRACTIONAL_SECOND_FIELD:
3389 // Fractional seconds left-justify
3390 i = countDigits(text, start, pos.getIndex());
3391 if (i < 3) {
3392 while (i < 3) {
3393 value *= 10;
3394 i++;
3395 }
3396 } else {
3397 int32_t a = 1;
3398 while (i > 3) {
3399 a *= 10;
3400 i--;
3401 }
3402 value /= a;
3403 }
3404 cal.set(UCAL_MILLISECOND, value);
3405 return pos.getIndex();
3406
3407 case UDAT_DOW_LOCAL_FIELD:
3408 if (gotNumber) // i.e., e or ee
3409 {
3410 // [We computed 'value' above.]
3411 cal.set(UCAL_DOW_LOCAL, value);
3412 return pos.getIndex();
3413 }
3414 // else for eee-eeeee fall through to handling of EEE-EEEEE
3415 // fall through, do not break here
3416 U_FALLTHROUGH[[clang::fallthrough]];
3417 case UDAT_DAY_OF_WEEK_FIELD:
3418 {
3419 // Want to be able to parse both short and long forms.
3420 // Try count == 4 (EEEE) wide first:
3421 int32_t newStart = 0;
3422 if(getBooleanAttribute(UDAT_PARSE_MULTIPLE_PATTERNS_FOR_MATCH, status) || count == 4) {
3423 if ((newStart = matchString(text, start, UCAL_DAY_OF_WEEK,
3424 fSymbols->fWeekdays, fSymbols->fWeekdaysCount, nullptr, cal)) > 0)
3425 return newStart;
3426 }
3427 // EEEE wide failed, now try EEE abbreviated
3428 if(getBooleanAttribute(UDAT_PARSE_MULTIPLE_PATTERNS_FOR_MATCH, status) || count == 3) {
3429 if ((newStart = matchString(text, start, UCAL_DAY_OF_WEEK,
3430 fSymbols->fShortWeekdays, fSymbols->fShortWeekdaysCount, nullptr, cal)) > 0)
3431 return newStart;
3432 }
3433 // EEE abbreviated failed, now try EEEEEE short
3434 if(getBooleanAttribute(UDAT_PARSE_MULTIPLE_PATTERNS_FOR_MATCH, status) || count == 6) {
3435 if ((newStart = matchString(text, start, UCAL_DAY_OF_WEEK,
3436 fSymbols->fShorterWeekdays, fSymbols->fShorterWeekdaysCount, nullptr, cal)) > 0)
3437 return newStart;
3438 }
3439 // EEEEEE short failed, now try EEEEE narrow
3440 if(getBooleanAttribute(UDAT_PARSE_MULTIPLE_PATTERNS_FOR_MATCH, status) || count == 5) {
3441 if ((newStart = matchString(text, start, UCAL_DAY_OF_WEEK,
3442 fSymbols->fNarrowWeekdays, fSymbols->fNarrowWeekdaysCount, nullptr, cal)) > 0)
3443 return newStart;
3444 }
3445 if (!getBooleanAttribute(UDAT_PARSE_ALLOW_NUMERIC, status) || patternCharIndex == UDAT_DAY_OF_WEEK_FIELD)
3446 return newStart;
3447 // else we allowing parsing as number, below
3448 }
3449 break;
3450
3451 case UDAT_STANDALONE_DAY_FIELD:
3452 {
3453 if (gotNumber) // c or cc
3454 {
3455 // [We computed 'value' above.]
3456 cal.set(UCAL_DOW_LOCAL, value);
3457 return pos.getIndex();
3458 }
3459 // Want to be able to parse both short and long forms.
3460 // Try count == 4 (cccc) first:
3461 int32_t newStart = 0;
3462 if(getBooleanAttribute(UDAT_PARSE_MULTIPLE_PATTERNS_FOR_MATCH, status) || count == 4) {
3463 if ((newStart = matchString(text, start, UCAL_DAY_OF_WEEK,
3464 fSymbols->fStandaloneWeekdays, fSymbols->fStandaloneWeekdaysCount, nullptr, cal)) > 0)
3465 return newStart;
3466 }
3467 if(getBooleanAttribute(UDAT_PARSE_MULTIPLE_PATTERNS_FOR_MATCH, status) || count == 3) {
3468 if ((newStart = matchString(text, start, UCAL_DAY_OF_WEEK,
3469 fSymbols->fStandaloneShortWeekdays, fSymbols->fStandaloneShortWeekdaysCount, nullptr, cal)) > 0)
3470 return newStart;
3471 }
3472 if(getBooleanAttribute(UDAT_PARSE_MULTIPLE_PATTERNS_FOR_MATCH, status) || count == 6) {
3473 if ((newStart = matchString(text, start, UCAL_DAY_OF_WEEK,
3474 fSymbols->fStandaloneShorterWeekdays, fSymbols->fStandaloneShorterWeekdaysCount, nullptr, cal)) > 0)
3475 return newStart;
3476 }
3477 if (!getBooleanAttribute(UDAT_PARSE_ALLOW_NUMERIC, status))
3478 return newStart;
3479 // else we allowing parsing as number, below
3480 }
3481 break;
3482
3483 case UDAT_AM_PM_FIELD:
3484 {
3485 // optionally try both wide/abbrev and narrow forms
3486 int32_t newStart = 0;
3487 // try wide/abbrev
3488 if( getBooleanAttribute(UDAT_PARSE_MULTIPLE_PATTERNS_FOR_MATCH, status) || count < 5 ) {
3489 if ((newStart = matchString(text, start, UCAL_AM_PM, fSymbols->fAmPms, fSymbols->fAmPmsCount, nullptr, cal)) > 0) {
3490 return newStart;
3491 }
3492 }
3493 // try narrow
3494 if( getBooleanAttribute(UDAT_PARSE_MULTIPLE_PATTERNS_FOR_MATCH, status) || count >= 5 ) {
3495 if ((newStart = matchString(text, start, UCAL_AM_PM, fSymbols->fNarrowAmPms, fSymbols->fNarrowAmPmsCount, nullptr, cal)) > 0) {
3496 return newStart;
3497 }
3498 }
3499 // no matches for given options
3500 return -start;
3501 }
3502
3503 case UDAT_HOUR1_FIELD:
3504 // [We computed 'value' above.]
3505 if (value == cal.getLeastMaximum(UCAL_HOUR)+1)
3506 value = 0;
3507
3508 // fall through to set field
3509 U_FALLTHROUGH[[clang::fallthrough]];
3510 case UDAT_HOUR0_FIELD:
3511 cal.set(UCAL_HOUR, value);
3512 return pos.getIndex();
3513
3514 case UDAT_QUARTER_FIELD:
3515 if (gotNumber) // i.e., Q or QQ.
3516 {
3517 // Don't want to parse the month if it is a string
3518 // while pattern uses numeric style: Q or QQ.
3519 // [We computed 'value' above.]
3520 cal.set(UCAL_MONTH, (value - 1) * 3);
3521 return pos.getIndex();
3522 } else {
3523 // count >= 3 // i.e., QQQ or QQQQ
3524 // Want to be able to parse short, long, and narrow forms.
3525 // Try count == 4 first:
3526 int32_t newStart = 0;
3527
3528 if(getBooleanAttribute(UDAT_PARSE_MULTIPLE_PATTERNS_FOR_MATCH, status) || count == 4) {
3529 if ((newStart = matchQuarterString(text, start, UCAL_MONTH,
3530 fSymbols->fQuarters, fSymbols->fQuartersCount, cal)) > 0)
3531 return newStart;
3532 }
3533 if(getBooleanAttribute(UDAT_PARSE_MULTIPLE_PATTERNS_FOR_MATCH, status) || count == 3) {
3534 if ((newStart = matchQuarterString(text, start, UCAL_MONTH,
3535 fSymbols->fShortQuarters, fSymbols->fShortQuartersCount, cal)) > 0)
3536 return newStart;
3537 }
3538 if(getBooleanAttribute(UDAT_PARSE_MULTIPLE_PATTERNS_FOR_MATCH, status) || count == 5) {
3539 if ((newStart = matchQuarterString(text, start, UCAL_MONTH,
3540 fSymbols->fNarrowQuarters, fSymbols->fNarrowQuartersCount, cal)) > 0)
3541 return newStart;
3542 }
3543 if (!getBooleanAttribute(UDAT_PARSE_ALLOW_NUMERIC, status))
3544 return newStart;
3545 // else we allowing parsing as number, below
3546 if(!getBooleanAttribute(UDAT_PARSE_MULTIPLE_PATTERNS_FOR_MATCH, status))
3547 return -start;
3548 }
3549 break;
3550
3551 case UDAT_STANDALONE_QUARTER_FIELD:
3552 if (gotNumber) // i.e., q or qq.
3553 {
3554 // Don't want to parse the month if it is a string
3555 // while pattern uses numeric style: q or q.
3556 // [We computed 'value' above.]
3557 cal.set(UCAL_MONTH, (value - 1) * 3);
3558 return pos.getIndex();
3559 } else {
3560 // count >= 3 // i.e., qqq or qqqq
3561 // Want to be able to parse both short and long forms.
3562 // Try count == 4 first:
3563 int32_t newStart = 0;
3564
3565 if(getBooleanAttribute(UDAT_PARSE_MULTIPLE_PATTERNS_FOR_MATCH, status) || count == 4) {
3566 if ((newStart = matchQuarterString(text, start, UCAL_MONTH,
3567 fSymbols->fStandaloneQuarters, fSymbols->fStandaloneQuartersCount, cal)) > 0)
3568 return newStart;
3569 }
3570 if(getBooleanAttribute(UDAT_PARSE_MULTIPLE_PATTERNS_FOR_MATCH, status) || count == 3) {
3571 if ((newStart = matchQuarterString(text, start, UCAL_MONTH,
3572 fSymbols->fStandaloneShortQuarters, fSymbols->fStandaloneShortQuartersCount, cal)) > 0)
3573 return newStart;
3574 }
3575 if(getBooleanAttribute(UDAT_PARSE_MULTIPLE_PATTERNS_FOR_MATCH, status) || count == 5) {
3576 if ((newStart = matchQuarterString(text, start, UCAL_MONTH,
3577 fSymbols->fStandaloneNarrowQuarters, fSymbols->fStandaloneNarrowQuartersCount, cal)) > 0)
3578 return newStart;
3579 }
3580 if (!getBooleanAttribute(UDAT_PARSE_ALLOW_NUMERIC, status))
3581 return newStart;
3582 // else we allowing parsing as number, below
3583 if(!getBooleanAttribute(UDAT_PARSE_MULTIPLE_PATTERNS_FOR_MATCH, status))
3584 return -start;
3585 }
3586 break;
3587
3588 case UDAT_TIMEZONE_FIELD: // 'z'
3589 {
3590 UTimeZoneFormatStyle style = (count < 4) ? UTZFMT_STYLE_SPECIFIC_SHORT : UTZFMT_STYLE_SPECIFIC_LONG;
3591 const TimeZoneFormat *tzfmt = tzFormat(status);
3592 if (U_SUCCESS(status)) {
3593 TimeZone *tz = tzfmt->parse(style, text, pos, tzTimeType);
3594 if (tz != nullptr) {
3595 cal.adoptTimeZone(tz);
3596 return pos.getIndex();
3597 }
3598 }
3599 return -start;
3600 }
3601 break;
3602 case UDAT_TIMEZONE_RFC_FIELD: // 'Z'
3603 {
3604 UTimeZoneFormatStyle style = (count < 4) ?
3605 UTZFMT_STYLE_ISO_BASIC_LOCAL_FULL : ((count == 5) ? UTZFMT_STYLE_ISO_EXTENDED_FULL: UTZFMT_STYLE_LOCALIZED_GMT);
3606 const TimeZoneFormat *tzfmt = tzFormat(status);
3607 if (U_SUCCESS(status)) {
3608 TimeZone *tz = tzfmt->parse(style, text, pos, tzTimeType);
3609 if (tz != nullptr) {
3610 cal.adoptTimeZone(tz);
3611 return pos.getIndex();
3612 }
3613 }
3614 return -start;
3615 }
3616 case UDAT_TIMEZONE_GENERIC_FIELD: // 'v'
3617 {
3618 UTimeZoneFormatStyle style = (count < 4) ? UTZFMT_STYLE_GENERIC_SHORT : UTZFMT_STYLE_GENERIC_LONG;
3619 const TimeZoneFormat *tzfmt = tzFormat(status);
3620 if (U_SUCCESS(status)) {
3621 TimeZone *tz = tzfmt->parse(style, text, pos, tzTimeType);
3622 if (tz != nullptr) {
3623 cal.adoptTimeZone(tz);
3624 return pos.getIndex();
3625 }
3626 }
3627 return -start;
3628 }
3629 case UDAT_TIMEZONE_SPECIAL_FIELD: // 'V'
3630 {
3631 UTimeZoneFormatStyle style;
3632 switch (count) {
3633 case 1:
3634 style = UTZFMT_STYLE_ZONE_ID_SHORT;
3635 break;
3636 case 2:
3637 style = UTZFMT_STYLE_ZONE_ID;
3638 break;
3639 case 3:
3640 style = UTZFMT_STYLE_EXEMPLAR_LOCATION;
3641 break;
3642 default:
3643 style = UTZFMT_STYLE_GENERIC_LOCATION;
3644 break;
3645 }
3646 const TimeZoneFormat *tzfmt = tzFormat(status);
3647 if (U_SUCCESS(status)) {
3648 TimeZone *tz = tzfmt->parse(style, text, pos, tzTimeType);
3649 if (tz != nullptr) {
3650 cal.adoptTimeZone(tz);
3651 return pos.getIndex();
3652 }
3653 }
3654 return -start;
3655 }
3656 case UDAT_TIMEZONE_LOCALIZED_GMT_OFFSET_FIELD: // 'O'
3657 {
3658 UTimeZoneFormatStyle style = (count < 4) ? UTZFMT_STYLE_LOCALIZED_GMT_SHORT : UTZFMT_STYLE_LOCALIZED_GMT;
3659 const TimeZoneFormat *tzfmt = tzFormat(status);
3660 if (U_SUCCESS(status)) {
3661 TimeZone *tz = tzfmt->parse(style, text, pos, tzTimeType);
3662 if (tz != nullptr) {
3663 cal.adoptTimeZone(tz);
3664 return pos.getIndex();
3665 }
3666 }
3667 return -start;
3668 }
3669 case UDAT_TIMEZONE_ISO_FIELD: // 'X'
3670 {
3671 UTimeZoneFormatStyle style;
3672 switch (count) {
3673 case 1:
3674 style = UTZFMT_STYLE_ISO_BASIC_SHORT;
3675 break;
3676 case 2:
3677 style = UTZFMT_STYLE_ISO_BASIC_FIXED;
3678 break;
3679 case 3:
3680 style = UTZFMT_STYLE_ISO_EXTENDED_FIXED;
3681 break;
3682 case 4:
3683 style = UTZFMT_STYLE_ISO_BASIC_FULL;
3684 break;
3685 default:
3686 style = UTZFMT_STYLE_ISO_EXTENDED_FULL;
3687 break;
3688 }
3689 const TimeZoneFormat *tzfmt = tzFormat(status);
3690 if (U_SUCCESS(status)) {
3691 TimeZone *tz = tzfmt->parse(style, text, pos, tzTimeType);
3692 if (tz != nullptr) {
3693 cal.adoptTimeZone(tz);
3694 return pos.getIndex();
3695 }
3696 }
3697 return -start;
3698 }
3699 case UDAT_TIMEZONE_ISO_LOCAL_FIELD: // 'x'
3700 {
3701 UTimeZoneFormatStyle style;
3702 switch (count) {
3703 case 1:
3704 style = UTZFMT_STYLE_ISO_BASIC_LOCAL_SHORT;
3705 break;
3706 case 2:
3707 style = UTZFMT_STYLE_ISO_BASIC_LOCAL_FIXED;
3708 break;
3709 case 3:
3710 style = UTZFMT_STYLE_ISO_EXTENDED_LOCAL_FIXED;
3711 break;
3712 case 4:
3713 style = UTZFMT_STYLE_ISO_BASIC_LOCAL_FULL;
3714 break;
3715 default:
3716 style = UTZFMT_STYLE_ISO_EXTENDED_LOCAL_FULL;
3717 break;
3718 }
3719 const TimeZoneFormat *tzfmt = tzFormat(status);
3720 if (U_SUCCESS(status)) {
3721 TimeZone *tz = tzfmt->parse(style, text, pos, tzTimeType);
3722 if (tz != nullptr) {
3723 cal.adoptTimeZone(tz);
3724 return pos.getIndex();
3725 }
3726 }
3727 return -start;
3728 }
3729 // currently no pattern character is defined for UDAT_TIME_SEPARATOR_FIELD
3730 // so we should not get here. Leave support in for future definition.
3731 case UDAT_TIME_SEPARATOR_FIELD:
3732 {
3733 static const char16_t def_sep = DateFormatSymbols::DEFAULT_TIME_SEPARATOR;
3734 static const char16_t alt_sep = DateFormatSymbols::ALTERNATE_TIME_SEPARATOR;
3735
3736 // Try matching a time separator.
3737 int32_t count_sep = 1;
3738 UnicodeString data[3];
3739 fSymbols->getTimeSeparatorString(data[0]);
3740
3741 // Add the default, if different from the locale.
3742 if (data[0].compare(&def_sep, 1) != 0) {
3743 data[count_sep++].setTo(def_sep);
3744 }
3745
3746 // If lenient, add also the alternate, if different from the locale.
3747 if (isLenient() && data[0].compare(&alt_sep, 1) != 0) {
3748 data[count_sep++].setTo(alt_sep);
3749 }
3750
3751 return matchString(text, start, UCAL_FIELD_COUNT /* => nothing to set */, data, count_sep, nullptr, cal);
3752 }
3753
3754 case UDAT_AM_PM_MIDNIGHT_NOON_FIELD:
3755 {
3756 U_ASSERT(dayPeriod != nullptr)(static_cast <bool> (dayPeriod != nullptr) ? void (0) :
__assert_fail ("dayPeriod != nullptr", __builtin_FILE (), __builtin_LINE
(), __extension__ __PRETTY_FUNCTION__));
3757 int32_t ampmStart = subParse(text, start, 0x61, count,
3758 obeyCount, allowNegative, ambiguousYear, saveHebrewMonth, cal,
3759 patLoc, numericLeapMonthFormatter, tzTimeType);
3760
3761 if (ampmStart > 0) {
3762 return ampmStart;
3763 } else {
3764 int32_t newStart = 0;
3765
3766 // Only match the first two strings from the day period strings array.
3767 if (getBooleanAttribute(UDAT_PARSE_MULTIPLE_PATTERNS_FOR_MATCH, status) || count == 3) {
3768 if ((newStart = matchDayPeriodStrings(text, start, fSymbols->fAbbreviatedDayPeriods,
3769 2, *dayPeriod)) > 0) {
3770 return newStart;
3771 }
3772 }
3773 if (getBooleanAttribute(UDAT_PARSE_MULTIPLE_PATTERNS_FOR_MATCH, status) || count == 5) {
3774 if ((newStart = matchDayPeriodStrings(text, start, fSymbols->fNarrowDayPeriods,
3775 2, *dayPeriod)) > 0) {
3776 return newStart;
3777 }
3778 }
3779 // count == 4, but allow other counts
3780 if (getBooleanAttribute(UDAT_PARSE_MULTIPLE_PATTERNS_FOR_MATCH, status)) {
3781 if ((newStart = matchDayPeriodStrings(text, start, fSymbols->fWideDayPeriods,
3782 2, *dayPeriod)) > 0) {
3783 return newStart;
3784 }
3785 }
3786
3787 return -start;
3788 }
3789 }
3790
3791 case UDAT_FLEXIBLE_DAY_PERIOD_FIELD:
3792 {
3793 U_ASSERT(dayPeriod != nullptr)(static_cast <bool> (dayPeriod != nullptr) ? void (0) :
__assert_fail ("dayPeriod != nullptr", __builtin_FILE (), __builtin_LINE
(), __extension__ __PRETTY_FUNCTION__));
3794 int32_t newStart = 0;
3795
3796 if (getBooleanAttribute(UDAT_PARSE_MULTIPLE_PATTERNS_FOR_MATCH, status) || count == 3) {
3797 if ((newStart = matchDayPeriodStrings(text, start, fSymbols->fAbbreviatedDayPeriods,
3798 fSymbols->fAbbreviatedDayPeriodsCount, *dayPeriod)) > 0) {
3799 return newStart;
3800 }
3801 }
3802 if (getBooleanAttribute(UDAT_PARSE_MULTIPLE_PATTERNS_FOR_MATCH, status) || count == 5) {
3803 if ((newStart = matchDayPeriodStrings(text, start, fSymbols->fNarrowDayPeriods,
3804 fSymbols->fNarrowDayPeriodsCount, *dayPeriod)) > 0) {
3805 return newStart;
3806 }
3807 }
3808 if (getBooleanAttribute(UDAT_PARSE_MULTIPLE_PATTERNS_FOR_MATCH, status) || count == 4) {
3809 if ((newStart = matchDayPeriodStrings(text, start, fSymbols->fWideDayPeriods,
3810 fSymbols->fWideDayPeriodsCount, *dayPeriod)) > 0) {
3811 return newStart;
3812 }
3813 }
3814
3815 return -start;
3816 }
3817
3818 default:
3819 // Handle "generic" fields
3820 // this is now handled below, outside the switch block
3821 break;
3822 }
3823 // Handle "generic" fields:
3824 // switch default case now handled here (outside switch block) to allow
3825 // parsing of some string fields as digits for lenient case
3826
3827 int32_t parseStart = pos.getIndex();
3828 const UnicodeString* src;
3829 if (obeyCount) {
3830 if ((start+count) > text.length()) {
3831 return -start;
3832 }
3833 text.extractBetween(0, start + count, temp);
3834 src = &temp;
3835 } else {
3836 src = &text;
3837 }
3838 parseInt(*src, number, pos, allowNegative,currentNumberFormat);
3839 if (obeyCount && !isLenient() && pos.getIndex() < start + count) {
3840 return -start;
3841 }
3842 if (pos.getIndex() != parseStart) {
3843 int32_t val = number.getLong();
3844
3845 // Don't need suffix processing here (as in number processing at the beginning of the function);
3846 // the new fields being handled as numeric values (month, weekdays, quarters) should not have suffixes.
3847
3848 if (!getBooleanAttribute(UDAT_PARSE_ALLOW_NUMERIC, status)) {
3849 // Check the range of the value
3850 int32_t bias = gFieldRangeBias[patternCharIndex];
3851 if (bias >= 0 && (val > cal.getMaximum(field) + bias || val < cal.getMinimum(field) + bias)) {
3852 return -start;
3853 }
3854 }
3855
3856 // For the following, need to repeat some of the "if (gotNumber)" code above:
3857 // UDAT_[STANDALONE_]MONTH_FIELD, UDAT_DOW_LOCAL_FIELD, UDAT_STANDALONE_DAY_FIELD,
3858 // UDAT_[STANDALONE_]QUARTER_FIELD
3859 switch (patternCharIndex) {
3860 case UDAT_MONTH_FIELD:
3861 // See notes under UDAT_MONTH_FIELD case above
3862 if (typeid(cal) == typeid(HebrewCalendar)) {
3863 HebrewCalendar *hc = (HebrewCalendar*)&cal;
3864 if (cal.isSet(UCAL_YEAR)) {
3865 UErrorCode monthStatus = U_ZERO_ERROR;
3866 if (!hc->isLeapYear(hc->get(UCAL_YEAR, monthStatus)) && val >= 6) {
3867 cal.set(UCAL_MONTH, val);
3868 } else {
3869 cal.set(UCAL_MONTH, val - 1);
3870 }
3871 } else {
3872 saveHebrewMonth = val;
3873 }
3874 } else {
3875 cal.set(UCAL_MONTH, val - 1);
3876 }
3877 break;
3878 case UDAT_STANDALONE_MONTH_FIELD:
3879 cal.set(UCAL_MONTH, val - 1);
3880 break;
3881 case UDAT_DOW_LOCAL_FIELD:
3882 case UDAT_STANDALONE_DAY_FIELD:
3883 cal.set(UCAL_DOW_LOCAL, val);
3884 break;
3885 case UDAT_QUARTER_FIELD:
3886 case UDAT_STANDALONE_QUARTER_FIELD:
3887 cal.set(UCAL_MONTH, (val - 1) * 3);
3888 break;
3889 case UDAT_RELATED_YEAR_FIELD:
3890 cal.setRelatedYear(val);
3891 break;
3892 default:
3893 cal.set(field, val);
3894 break;
3895 }
3896 return pos.getIndex();
3897 }
3898 return -start;
3899}
3900
3901/**
3902 * Parse an integer using fNumberFormat. This method is semantically
3903 * const, but actually may modify fNumberFormat.
3904 */
3905void SimpleDateFormat::parseInt(const UnicodeString& text,
3906 Formattable& number,
3907 ParsePosition& pos,
3908 UBool allowNegative,
3909 const NumberFormat *fmt) const {
3910 parseInt(text, number, -1, pos, allowNegative,fmt);
3911}
3912
3913/**
3914 * Parse an integer using fNumberFormat up to maxDigits.
3915 */
3916void SimpleDateFormat::parseInt(const UnicodeString& text,
3917 Formattable& number,
3918 int32_t maxDigits,
3919 ParsePosition& pos,
3920 UBool allowNegative,
3921 const NumberFormat *fmt) const {
3922 UnicodeString oldPrefix;
3923 const auto* fmtAsDF = dynamic_cast<const DecimalFormat*>(fmt);
3924 LocalPointer<DecimalFormat> df;
3925 if (!allowNegative && fmtAsDF != nullptr) {
3926 df.adoptInstead(fmtAsDF->clone());
3927 if (df.isNull()) {
3928 // Memory allocation error
3929 return;
3930 }
3931 df->setNegativePrefix(UnicodeString(true, SUPPRESS_NEGATIVE_PREFIX, -1));
3932 fmt = df.getAlias();
3933 }
3934 int32_t oldPos = pos.getIndex();
3935 fmt->parse(text, number, pos);
3936
3937 if (maxDigits > 0) {
3938 // adjust the result to fit into
3939 // the maxDigits and move the position back
3940 int32_t nDigits = pos.getIndex() - oldPos;
3941 if (nDigits > maxDigits) {
3942 int32_t val = number.getLong();
3943 nDigits -= maxDigits;
3944 while (nDigits > 0) {
3945 val /= 10;
3946 nDigits--;
3947 }
3948 pos.setIndex(oldPos + maxDigits);
3949 number.setLong(val);
3950 }
3951 }
3952}
3953
3954int32_t SimpleDateFormat::countDigits(const UnicodeString& text, int32_t start, int32_t end) const {
3955 int32_t numDigits = 0;
3956 int32_t idx = start;
3957 while (idx < end) {
3958 UChar32 cp = text.char32At(idx);
3959 if (u_isdigitu_isdigit_77(cp)) {
3960 numDigits++;
3961 }
3962 idx += U16_LENGTH(cp)((uint32_t)(cp)<=0xffff ? 1 : 2);
3963 }
3964 return numDigits;
3965}
3966
3967//----------------------------------------------------------------------
3968
3969void SimpleDateFormat::translatePattern(const UnicodeString& originalPattern,
3970 UnicodeString& translatedPattern,
3971 const UnicodeString& from,
3972 const UnicodeString& to,
3973 UErrorCode& status)
3974{
3975 // run through the pattern and convert any pattern symbols from the version
3976 // in "from" to the corresponding character in "to". This code takes
3977 // quoted strings into account (it doesn't try to translate them), and it signals
3978 // an error if a particular "pattern character" doesn't appear in "from".
3979 // Depending on the values of "from" and "to" this can convert from generic
3980 // to localized patterns or localized to generic.
3981 if (U_FAILURE(status)) {
3982 return;
3983 }
3984
3985 translatedPattern.remove();
3986 UBool inQuote = false;
3987 for (int32_t i = 0; i < originalPattern.length(); ++i) {
3988 char16_t c = originalPattern[i];
3989 if (inQuote) {
3990 if (c == QUOTE) {
3991 inQuote = false;
3992 }
3993 } else {
3994 if (c == QUOTE) {
3995 inQuote = true;
3996 } else if (isSyntaxChar(c)) {
3997 int32_t ci = from.indexOf(c);
3998 if (ci == -1) {
3999 status = U_INVALID_FORMAT_ERROR;
4000 return;
4001 }
4002 c = to[ci];
4003 }
4004 }
4005 translatedPattern += c;
4006 }
4007 if (inQuote) {
4008 status = U_INVALID_FORMAT_ERROR;
4009 return;
4010 }
4011}
4012
4013//----------------------------------------------------------------------
4014
4015UnicodeString&
4016SimpleDateFormat::toPattern(UnicodeString& result) const
4017{
4018 result = fPattern;
4019 return result;
4020}
4021
4022//----------------------------------------------------------------------
4023
4024UnicodeString&
4025SimpleDateFormat::toLocalizedPattern(UnicodeString& result,
4026 UErrorCode& status) const
4027{
4028 translatePattern(fPattern, result,
4029 UnicodeString(DateFormatSymbols::getPatternUChars()),
4030 fSymbols->fLocalPatternChars, status);
4031 return result;
4032}
4033
4034//----------------------------------------------------------------------
4035
4036void
4037SimpleDateFormat::applyPattern(const UnicodeString& pattern)
4038{
4039 fPattern = pattern;
4040 parsePattern();
4041
4042 // Hack to update use of Gannen year numbering for ja@calendar=japanese -
4043 // use only if format is non-numeric (includes 年) and no other fDateOverride.
4044 if (fCalendar != nullptr && typeid(*fCalendar) == typeid(JapaneseCalendar) &&
4045 uprv_strcmp(fLocale.getLanguage(),"ja"):: strcmp(fLocale.getLanguage(), "ja") == 0) {
4046 if (fDateOverride==UnicodeString(u"y=jpanyear") && !fHasHanYearChar) {
4047 // Gannen numbering is set but new pattern should not use it, unset;
4048 // use procedure from adoptNumberFormat to clear overrides
4049 if (fSharedNumberFormatters) {
4050 freeSharedNumberFormatters(fSharedNumberFormatters);
4051 fSharedNumberFormatters = nullptr;
4052 }
4053 fDateOverride.setToBogus(); // record status
4054 } else if (fDateOverride.isBogus() && fHasHanYearChar) {
4055 // No current override (=> no Gannen numbering) but new pattern needs it;
4056 // use procedures from initNUmberFormatters / adoptNumberFormat
4057 umtx_lockumtx_lock_77(&LOCK);
4058 if (fSharedNumberFormatters == nullptr) {
4059 fSharedNumberFormatters = allocSharedNumberFormatters();
4060 }
4061 umtx_unlockumtx_unlock_77(&LOCK);
4062 if (fSharedNumberFormatters != nullptr) {
4063 Locale ovrLoc(fLocale.getLanguage(),fLocale.getCountry(),fLocale.getVariant(),"numbers=jpanyear");
4064 UErrorCode status = U_ZERO_ERROR;
4065 const SharedNumberFormat *snf = createSharedNumberFormat(ovrLoc, status);
4066 if (U_SUCCESS(status)) {
4067 // Now that we have an appropriate number formatter, fill in the
4068 // appropriate slot in the number formatters table.
4069 UDateFormatField patternCharIndex = DateFormatSymbols::getPatternCharIndex(u'y');
4070 SharedObject::copyPtr(snf, fSharedNumberFormatters[patternCharIndex]);
4071 snf->deleteIfZeroRefCount();
4072 fDateOverride.setTo(u"y=jpanyear", -1); // record status
4073 }
4074 }
4075 }
4076 }
4077}
4078
4079//----------------------------------------------------------------------
4080
4081void
4082SimpleDateFormat::applyLocalizedPattern(const UnicodeString& pattern,
4083 UErrorCode &status)
4084{
4085 translatePattern(pattern, fPattern,
4086 fSymbols->fLocalPatternChars,
4087 UnicodeString(DateFormatSymbols::getPatternUChars()), status);
4088}
4089
4090//----------------------------------------------------------------------
4091
4092const DateFormatSymbols*
4093SimpleDateFormat::getDateFormatSymbols() const
4094{
4095 return fSymbols;
4096}
4097
4098//----------------------------------------------------------------------
4099
4100void
4101SimpleDateFormat::adoptDateFormatSymbols(DateFormatSymbols* newFormatSymbols)
4102{
4103 delete fSymbols;
4104 fSymbols = newFormatSymbols;
4105}
4106
4107//----------------------------------------------------------------------
4108void
4109SimpleDateFormat::setDateFormatSymbols(const DateFormatSymbols& newFormatSymbols)
4110{
4111 delete fSymbols;
4112 fSymbols = new DateFormatSymbols(newFormatSymbols);
4113}
4114
4115//----------------------------------------------------------------------
4116const TimeZoneFormat*
4117SimpleDateFormat::getTimeZoneFormat() const {
4118 // TimeZoneFormat initialization might fail when out of memory.
4119 // If we always initialize TimeZoneFormat instance, we can return
4120 // such status there. For now, this implementation lazily instantiates
4121 // a TimeZoneFormat for performance optimization reasons, but cannot
4122 // propagate such error (probably just out of memory case) to the caller.
4123 UErrorCode status = U_ZERO_ERROR;
4124 return (const TimeZoneFormat*)tzFormat(status);
4125}
4126
4127//----------------------------------------------------------------------
4128void
4129SimpleDateFormat::adoptTimeZoneFormat(TimeZoneFormat* timeZoneFormatToAdopt)
4130{
4131 delete fTimeZoneFormat;
4132 fTimeZoneFormat = timeZoneFormatToAdopt;
4133}
4134
4135//----------------------------------------------------------------------
4136void
4137SimpleDateFormat::setTimeZoneFormat(const TimeZoneFormat& newTimeZoneFormat)
4138{
4139 delete fTimeZoneFormat;
4140 fTimeZoneFormat = new TimeZoneFormat(newTimeZoneFormat);
4141}
4142
4143//----------------------------------------------------------------------
4144
4145
4146void SimpleDateFormat::adoptCalendar(Calendar* calendarToAdopt)
4147{
4148 UErrorCode status = U_ZERO_ERROR;
4149 Locale calLocale(fLocale);
4150 calLocale.setKeywordValue("calendar", calendarToAdopt->getType(), status);
4151 DateFormatSymbols *newSymbols =
4152 DateFormatSymbols::createForLocale(calLocale, status);
4153 if (U_FAILURE(status)) {
4154 delete calendarToAdopt;
4155 return;
4156 }
4157 DateFormat::adoptCalendar(calendarToAdopt);
4158 delete fSymbols;
4159 fSymbols = newSymbols;
4160 initializeDefaultCentury(); // we need a new century (possibly)
4161}
4162
4163
4164//----------------------------------------------------------------------
4165
4166
4167// override the DateFormat implementation in order to
4168// lazily initialize fCapitalizationBrkIter
4169void
4170SimpleDateFormat::setContext(UDisplayContext value, UErrorCode& status)
4171{
4172 DateFormat::setContext(value, status);
4173#if !UCONFIG_NO_BREAK_ITERATION1
4174 if (U_SUCCESS(status)) {
4175 if ( fCapitalizationBrkIter == nullptr && (value==UDISPCTX_CAPITALIZATION_FOR_BEGINNING_OF_SENTENCE ||
4176 value==UDISPCTX_CAPITALIZATION_FOR_UI_LIST_OR_MENU || value==UDISPCTX_CAPITALIZATION_FOR_STANDALONE) ) {
4177 status = U_ZERO_ERROR;
4178 fCapitalizationBrkIter = BreakIterator::createSentenceInstance(fLocale, status);
4179 if (U_FAILURE(status)) {
4180 delete fCapitalizationBrkIter;
4181 fCapitalizationBrkIter = nullptr;
4182 }
4183 }
4184 }
4185#endif
4186}
4187
4188
4189//----------------------------------------------------------------------
4190
4191
4192UBool
4193SimpleDateFormat::isFieldUnitIgnored(UCalendarDateFields field) const {
4194 return isFieldUnitIgnored(fPattern, field);
4195}
4196
4197
4198UBool
4199SimpleDateFormat::isFieldUnitIgnored(const UnicodeString& pattern,
4200 UCalendarDateFields field) {
4201 int32_t fieldLevel = fgCalendarFieldToLevel[field];
4202 int32_t level;
4203 char16_t ch;
4204 UBool inQuote = false;
4205 char16_t prevCh = 0;
4206 int32_t count = 0;
4207
4208 for (int32_t i = 0; i < pattern.length(); ++i) {
4209 ch = pattern[i];
4210 if (ch != prevCh && count > 0) {
4211 level = getLevelFromChar(prevCh);
4212 // the larger the level, the smaller the field unit.
4213 if (fieldLevel <= level) {
4214 return false;
4215 }
4216 count = 0;
4217 }
4218 if (ch == QUOTE) {
4219 if ((i+1) < pattern.length() && pattern[i+1] == QUOTE) {
4220 ++i;
4221 } else {
4222 inQuote = ! inQuote;
4223 }
4224 }
4225 else if (!inQuote && isSyntaxChar(ch)) {
4226 prevCh = ch;
4227 ++count;
4228 }
4229 }
4230 if (count > 0) {
4231 // last item
4232 level = getLevelFromChar(prevCh);
4233 if (fieldLevel <= level) {
4234 return false;
4235 }
4236 }
4237 return true;
4238}
4239
4240//----------------------------------------------------------------------
4241
4242const Locale&
4243SimpleDateFormat::getSmpFmtLocale() const {
4244 return fLocale;
4245}
4246
4247//----------------------------------------------------------------------
4248
4249int32_t
4250SimpleDateFormat::checkIntSuffix(const UnicodeString& text, int32_t start,
4251 int32_t patLoc, UBool isNegative) const {
4252 // local variables
4253 UnicodeString suf;
4254 int32_t patternMatch;
4255 int32_t textPreMatch;
4256 int32_t textPostMatch;
4257
4258 // check that we are still in range
4259 if ( (start > text.length()) ||
4260 (start < 0) ||
4261 (patLoc < 0) ||
4262 (patLoc > fPattern.length())) {
4263 // out of range, don't advance location in text
4264 return start;
4265 }
4266
4267 // get the suffix
4268 DecimalFormat* decfmt = dynamic_cast<DecimalFormat*>(fNumberFormat);
4269 if (decfmt != nullptr) {
4270 if (isNegative) {
4271 suf = decfmt->getNegativeSuffix(suf);
4272 }
4273 else {
4274 suf = decfmt->getPositiveSuffix(suf);
4275 }
4276 }
4277
4278 // check for suffix
4279 if (suf.length() <= 0) {
4280 return start;
4281 }
4282
4283 // check suffix will be encountered in the pattern
4284 patternMatch = compareSimpleAffix(suf,fPattern,patLoc);
4285
4286 // check if a suffix will be encountered in the text
4287 textPreMatch = compareSimpleAffix(suf,text,start);
4288
4289 // check if a suffix was encountered in the text
4290 textPostMatch = compareSimpleAffix(suf,text,start-suf.length());
4291
4292 // check for suffix match
4293 if ((textPreMatch >= 0) && (patternMatch >= 0) && (textPreMatch == patternMatch)) {
4294 return start;
4295 }
4296 else if ((textPostMatch >= 0) && (patternMatch >= 0) && (textPostMatch == patternMatch)) {
4297 return start - suf.length();
4298 }
4299
4300 // should not get here
4301 return start;
4302}
4303
4304//----------------------------------------------------------------------
4305
4306int32_t
4307SimpleDateFormat::compareSimpleAffix(const UnicodeString& affix,
4308 const UnicodeString& input,
4309 int32_t pos) const {
4310 int32_t start = pos;
4311 for (int32_t i=0; i<affix.length(); ) {
4312 UChar32 c = affix.char32At(i);
4313 int32_t len = U16_LENGTH(c)((uint32_t)(c)<=0xffff ? 1 : 2);
4314 if (PatternProps::isWhiteSpace(c)) {
4315 // We may have a pattern like: \u200F \u0020
4316 // and input text like: \u200F \u0020
4317 // Note that U+200F and U+0020 are Pattern_White_Space but only
4318 // U+0020 is UWhiteSpace. So we have to first do a direct
4319 // match of the run of Pattern_White_Space in the pattern,
4320 // then match any extra characters.
4321 UBool literalMatch = false;
4322 while (pos < input.length() &&
4323 input.char32At(pos) == c) {
4324 literalMatch = true;
4325 i += len;
4326 pos += len;
4327 if (i == affix.length()) {
4328 break;
4329 }
4330 c = affix.char32At(i);
4331 len = U16_LENGTH(c)((uint32_t)(c)<=0xffff ? 1 : 2);
4332 if (!PatternProps::isWhiteSpace(c)) {
4333 break;
4334 }
4335 }
4336
4337 // Advance over run in pattern
4338 i = skipPatternWhiteSpace(affix, i);
4339
4340 // Advance over run in input text
4341 // Must see at least one white space char in input,
4342 // unless we've already matched some characters literally.
4343 int32_t s = pos;
4344 pos = skipUWhiteSpace(input, pos);
4345 if (pos == s && !literalMatch) {
4346 return -1;
4347 }
4348
4349 // If we skip UWhiteSpace in the input text, we need to skip it in the pattern.
4350 // Otherwise, the previous lines may have skipped over text (such as U+00A0) that
4351 // is also in the affix.
4352 i = skipUWhiteSpace(affix, i);
4353 } else {
4354 if (pos < input.length() &&
4355 input.char32At(pos) == c) {
4356 i += len;
4357 pos += len;
4358 } else {
4359 return -1;
4360 }
4361 }
4362 }
4363 return pos - start;
4364}
4365
4366//----------------------------------------------------------------------
4367
4368int32_t
4369SimpleDateFormat::skipPatternWhiteSpace(const UnicodeString& text, int32_t pos) const {
4370 const char16_t* s = text.getBuffer();
4371 return static_cast<int32_t>(PatternProps::skipWhiteSpace(s + pos, text.length() - pos) - s);
4372}
4373
4374//----------------------------------------------------------------------
4375
4376int32_t
4377SimpleDateFormat::skipUWhiteSpace(const UnicodeString& text, int32_t pos) const {
4378 while (pos < text.length()) {
4379 UChar32 c = text.char32At(pos);
4380 if (!u_isUWhiteSpaceu_isUWhiteSpace_77(c)) {
4381 break;
4382 }
4383 pos += U16_LENGTH(c)((uint32_t)(c)<=0xffff ? 1 : 2);
4384 }
4385 return pos;
4386}
4387
4388//----------------------------------------------------------------------
4389
4390// Lazy TimeZoneFormat instantiation, semantically const.
4391TimeZoneFormat *
4392SimpleDateFormat::tzFormat(UErrorCode &status) const {
4393 Mutex m(&LOCK);
4394 if (fTimeZoneFormat == nullptr && U_SUCCESS(status)) {
4395 const_cast<SimpleDateFormat *>(this)->fTimeZoneFormat =
4396 TimeZoneFormat::createInstance(fLocale, status);
4397 }
4398 return fTimeZoneFormat;
4399}
4400
4401void SimpleDateFormat::parsePattern() {
4402 fHasMinute = false;
4403 fHasSecond = false;
4404 fHasHanYearChar = false;
4405
4406 int len = fPattern.length();
4407 UBool inQuote = false;
4408 for (int32_t i = 0; i < len; ++i) {
4409 char16_t ch = fPattern[i];
4410 if (ch == QUOTE) {
4411 inQuote = !inQuote;
4412 }
4413 if (ch == 0x5E74) { // don't care whether this is inside quotes
4414 fHasHanYearChar = true;
4415 }
4416 if (!inQuote) {
4417 if (ch == 0x6D) { // 0x6D == 'm'
4418 fHasMinute = true;
4419 }
4420 if (ch == 0x73) { // 0x73 == 's'
4421 fHasSecond = true;
4422 }
4423 }
4424 }
4425}
4426
4427U_NAMESPACE_END}
4428
4429#endif /* #if !UCONFIG_NO_FORMATTING */
4430
4431//eof