Thu, 22 Jan 2015 13:21:57 +0100
Incorporate requested changes from Mozilla in review:
https://bugzilla.mozilla.org/show_bug.cgi?id=1123480#c6
michael@0 | 1 | /* |
michael@0 | 2 | ******************************************************************************* |
michael@0 | 3 | * Copyright (C) 1997-2013, International Business Machines Corporation and * |
michael@0 | 4 | * others. All Rights Reserved. * |
michael@0 | 5 | ******************************************************************************* |
michael@0 | 6 | * |
michael@0 | 7 | * File SMPDTFMT.CPP |
michael@0 | 8 | * |
michael@0 | 9 | * Modification History: |
michael@0 | 10 | * |
michael@0 | 11 | * Date Name Description |
michael@0 | 12 | * 02/19/97 aliu Converted from java. |
michael@0 | 13 | * 03/31/97 aliu Modified extensively to work with 50 locales. |
michael@0 | 14 | * 04/01/97 aliu Added support for centuries. |
michael@0 | 15 | * 07/09/97 helena Made ParsePosition into a class. |
michael@0 | 16 | * 07/21/98 stephen Added initializeDefaultCentury. |
michael@0 | 17 | * Removed getZoneIndex (added in DateFormatSymbols) |
michael@0 | 18 | * Removed subParseLong |
michael@0 | 19 | * Removed chk |
michael@0 | 20 | * 02/22/99 stephen Removed character literals for EBCDIC safety |
michael@0 | 21 | * 10/14/99 aliu Updated 2-digit year parsing so that only "00" thru |
michael@0 | 22 | * "99" are recognized. {j28 4182066} |
michael@0 | 23 | * 11/15/99 weiv Added support for week of year/day of week format |
michael@0 | 24 | ******************************************************************************** |
michael@0 | 25 | */ |
michael@0 | 26 | |
michael@0 | 27 | #define ZID_KEY_MAX 128 |
michael@0 | 28 | |
michael@0 | 29 | #include "unicode/utypes.h" |
michael@0 | 30 | |
michael@0 | 31 | #if !UCONFIG_NO_FORMATTING |
michael@0 | 32 | |
michael@0 | 33 | #include "unicode/smpdtfmt.h" |
michael@0 | 34 | #include "unicode/dtfmtsym.h" |
michael@0 | 35 | #include "unicode/ures.h" |
michael@0 | 36 | #include "unicode/msgfmt.h" |
michael@0 | 37 | #include "unicode/calendar.h" |
michael@0 | 38 | #include "unicode/gregocal.h" |
michael@0 | 39 | #include "unicode/timezone.h" |
michael@0 | 40 | #include "unicode/decimfmt.h" |
michael@0 | 41 | #include "unicode/dcfmtsym.h" |
michael@0 | 42 | #include "unicode/uchar.h" |
michael@0 | 43 | #include "unicode/uniset.h" |
michael@0 | 44 | #include "unicode/ustring.h" |
michael@0 | 45 | #include "unicode/basictz.h" |
michael@0 | 46 | #include "unicode/simpletz.h" |
michael@0 | 47 | #include "unicode/rbtz.h" |
michael@0 | 48 | #include "unicode/tzfmt.h" |
michael@0 | 49 | #include "unicode/utf16.h" |
michael@0 | 50 | #include "unicode/vtzone.h" |
michael@0 | 51 | #include "unicode/udisplaycontext.h" |
michael@0 | 52 | #include "olsontz.h" |
michael@0 | 53 | #include "patternprops.h" |
michael@0 | 54 | #include "fphdlimp.h" |
michael@0 | 55 | #include "gregoimp.h" |
michael@0 | 56 | #include "hebrwcal.h" |
michael@0 | 57 | #include "cstring.h" |
michael@0 | 58 | #include "uassert.h" |
michael@0 | 59 | #include "cmemory.h" |
michael@0 | 60 | #include "umutex.h" |
michael@0 | 61 | #include <float.h> |
michael@0 | 62 | #include "smpdtfst.h" |
michael@0 | 63 | |
michael@0 | 64 | #if defined( U_DEBUG_CALSVC ) || defined (U_DEBUG_CAL) |
michael@0 | 65 | #include <stdio.h> |
michael@0 | 66 | #endif |
michael@0 | 67 | |
michael@0 | 68 | // ***************************************************************************** |
michael@0 | 69 | // class SimpleDateFormat |
michael@0 | 70 | // ***************************************************************************** |
michael@0 | 71 | |
michael@0 | 72 | U_NAMESPACE_BEGIN |
michael@0 | 73 | |
michael@0 | 74 | static const UChar PATTERN_CHAR_BASE = 0x40; |
michael@0 | 75 | |
michael@0 | 76 | /** |
michael@0 | 77 | * Last-resort string to use for "GMT" when constructing time zone strings. |
michael@0 | 78 | */ |
michael@0 | 79 | // For time zones that have no names, use strings GMT+minutes and |
michael@0 | 80 | // GMT-minutes. For instance, in France the time zone is GMT+60. |
michael@0 | 81 | // Also accepted are GMT+H:MM or GMT-H:MM. |
michael@0 | 82 | // Currently not being used |
michael@0 | 83 | //static const UChar gGmt[] = {0x0047, 0x004D, 0x0054, 0x0000}; // "GMT" |
michael@0 | 84 | //static const UChar gGmtPlus[] = {0x0047, 0x004D, 0x0054, 0x002B, 0x0000}; // "GMT+" |
michael@0 | 85 | //static const UChar gGmtMinus[] = {0x0047, 0x004D, 0x0054, 0x002D, 0x0000}; // "GMT-" |
michael@0 | 86 | //static const UChar gDefGmtPat[] = {0x0047, 0x004D, 0x0054, 0x007B, 0x0030, 0x007D, 0x0000}; /* GMT{0} */ |
michael@0 | 87 | //static const UChar gDefGmtNegHmsPat[] = {0x002D, 0x0048, 0x0048, 0x003A, 0x006D, 0x006D, 0x003A, 0x0073, 0x0073, 0x0000}; /* -HH:mm:ss */ |
michael@0 | 88 | //static const UChar gDefGmtNegHmPat[] = {0x002D, 0x0048, 0x0048, 0x003A, 0x006D, 0x006D, 0x0000}; /* -HH:mm */ |
michael@0 | 89 | //static const UChar gDefGmtPosHmsPat[] = {0x002B, 0x0048, 0x0048, 0x003A, 0x006D, 0x006D, 0x003A, 0x0073, 0x0073, 0x0000}; /* +HH:mm:ss */ |
michael@0 | 90 | //static const UChar gDefGmtPosHmPat[] = {0x002B, 0x0048, 0x0048, 0x003A, 0x006D, 0x006D, 0x0000}; /* +HH:mm */ |
michael@0 | 91 | //static const UChar gUt[] = {0x0055, 0x0054, 0x0000}; // "UT" |
michael@0 | 92 | //static const UChar gUtc[] = {0x0055, 0x0054, 0x0043, 0x0000}; // "UT" |
michael@0 | 93 | |
michael@0 | 94 | typedef enum GmtPatSize { |
michael@0 | 95 | kGmtLen = 3, |
michael@0 | 96 | kGmtPatLen = 6, |
michael@0 | 97 | kNegHmsLen = 9, |
michael@0 | 98 | kNegHmLen = 6, |
michael@0 | 99 | kPosHmsLen = 9, |
michael@0 | 100 | kPosHmLen = 6, |
michael@0 | 101 | kUtLen = 2, |
michael@0 | 102 | kUtcLen = 3 |
michael@0 | 103 | } GmtPatSize; |
michael@0 | 104 | |
michael@0 | 105 | // Stuff needed for numbering system overrides |
michael@0 | 106 | |
michael@0 | 107 | typedef enum OvrStrType { |
michael@0 | 108 | kOvrStrDate = 0, |
michael@0 | 109 | kOvrStrTime = 1, |
michael@0 | 110 | kOvrStrBoth = 2 |
michael@0 | 111 | } OvrStrType; |
michael@0 | 112 | |
michael@0 | 113 | static const UDateFormatField kDateFields[] = { |
michael@0 | 114 | UDAT_YEAR_FIELD, |
michael@0 | 115 | UDAT_MONTH_FIELD, |
michael@0 | 116 | UDAT_DATE_FIELD, |
michael@0 | 117 | UDAT_DAY_OF_YEAR_FIELD, |
michael@0 | 118 | UDAT_DAY_OF_WEEK_IN_MONTH_FIELD, |
michael@0 | 119 | UDAT_WEEK_OF_YEAR_FIELD, |
michael@0 | 120 | UDAT_WEEK_OF_MONTH_FIELD, |
michael@0 | 121 | UDAT_YEAR_WOY_FIELD, |
michael@0 | 122 | UDAT_EXTENDED_YEAR_FIELD, |
michael@0 | 123 | UDAT_JULIAN_DAY_FIELD, |
michael@0 | 124 | UDAT_STANDALONE_DAY_FIELD, |
michael@0 | 125 | UDAT_STANDALONE_MONTH_FIELD, |
michael@0 | 126 | UDAT_QUARTER_FIELD, |
michael@0 | 127 | UDAT_STANDALONE_QUARTER_FIELD, |
michael@0 | 128 | UDAT_YEAR_NAME_FIELD }; |
michael@0 | 129 | static const int8_t kDateFieldsCount = 15; |
michael@0 | 130 | |
michael@0 | 131 | static const UDateFormatField kTimeFields[] = { |
michael@0 | 132 | UDAT_HOUR_OF_DAY1_FIELD, |
michael@0 | 133 | UDAT_HOUR_OF_DAY0_FIELD, |
michael@0 | 134 | UDAT_MINUTE_FIELD, |
michael@0 | 135 | UDAT_SECOND_FIELD, |
michael@0 | 136 | UDAT_FRACTIONAL_SECOND_FIELD, |
michael@0 | 137 | UDAT_HOUR1_FIELD, |
michael@0 | 138 | UDAT_HOUR0_FIELD, |
michael@0 | 139 | UDAT_MILLISECONDS_IN_DAY_FIELD, |
michael@0 | 140 | UDAT_TIMEZONE_RFC_FIELD, |
michael@0 | 141 | UDAT_TIMEZONE_LOCALIZED_GMT_OFFSET_FIELD }; |
michael@0 | 142 | static const int8_t kTimeFieldsCount = 10; |
michael@0 | 143 | |
michael@0 | 144 | |
michael@0 | 145 | // This is a pattern-of-last-resort used when we can't load a usable pattern out |
michael@0 | 146 | // of a resource. |
michael@0 | 147 | static const UChar gDefaultPattern[] = |
michael@0 | 148 | { |
michael@0 | 149 | 0x79, 0x79, 0x79, 0x79, 0x4D, 0x4D, 0x64, 0x64, 0x20, 0x68, 0x68, 0x3A, 0x6D, 0x6D, 0x20, 0x61, 0 |
michael@0 | 150 | }; /* "yyyyMMdd hh:mm a" */ |
michael@0 | 151 | |
michael@0 | 152 | // This prefix is designed to NEVER MATCH real text, in order to |
michael@0 | 153 | // suppress the parsing of negative numbers. Adjust as needed (if |
michael@0 | 154 | // this becomes valid Unicode). |
michael@0 | 155 | static const UChar SUPPRESS_NEGATIVE_PREFIX[] = {0xAB00, 0}; |
michael@0 | 156 | |
michael@0 | 157 | /** |
michael@0 | 158 | * These are the tags we expect to see in normal resource bundle files associated |
michael@0 | 159 | * with a locale. |
michael@0 | 160 | */ |
michael@0 | 161 | static const char gDateTimePatternsTag[]="DateTimePatterns"; |
michael@0 | 162 | |
michael@0 | 163 | //static const UChar gEtcUTC[] = {0x45, 0x74, 0x63, 0x2F, 0x55, 0x54, 0x43, 0x00}; // "Etc/UTC" |
michael@0 | 164 | static const UChar QUOTE = 0x27; // Single quote |
michael@0 | 165 | |
michael@0 | 166 | /* |
michael@0 | 167 | * The field range check bias for each UDateFormatField. |
michael@0 | 168 | * The bias is added to the minimum and maximum values |
michael@0 | 169 | * before they are compared to the parsed number. |
michael@0 | 170 | * For example, the calendar stores zero-based month numbers |
michael@0 | 171 | * but the parsed month numbers start at 1, so the bias is 1. |
michael@0 | 172 | * |
michael@0 | 173 | * A value of -1 means that the value is not checked. |
michael@0 | 174 | */ |
michael@0 | 175 | static const int32_t gFieldRangeBias[] = { |
michael@0 | 176 | -1, // 'G' - UDAT_ERA_FIELD |
michael@0 | 177 | -1, // 'y' - UDAT_YEAR_FIELD |
michael@0 | 178 | 1, // 'M' - UDAT_MONTH_FIELD |
michael@0 | 179 | 0, // 'd' - UDAT_DATE_FIELD |
michael@0 | 180 | -1, // 'k' - UDAT_HOUR_OF_DAY1_FIELD |
michael@0 | 181 | -1, // 'H' - UDAT_HOUR_OF_DAY0_FIELD |
michael@0 | 182 | 0, // 'm' - UDAT_MINUTE_FIELD |
michael@0 | 183 | 0, // 's' - UDAT_SEOND_FIELD |
michael@0 | 184 | -1, // 'S' - UDAT_FRACTIONAL_SECOND_FIELD (0-999?) |
michael@0 | 185 | -1, // 'E' - UDAT_DAY_OF_WEEK_FIELD (1-7?) |
michael@0 | 186 | -1, // 'D' - UDAT_DAY_OF_YEAR_FIELD (1 - 366?) |
michael@0 | 187 | -1, // 'F' - UDAT_DAY_OF_WEEK_IN_MONTH_FIELD (1-5?) |
michael@0 | 188 | -1, // 'w' - UDAT_WEEK_OF_YEAR_FIELD (1-52?) |
michael@0 | 189 | -1, // 'W' - UDAT_WEEK_OF_MONTH_FIELD (1-5?) |
michael@0 | 190 | -1, // 'a' - UDAT_AM_PM_FIELD |
michael@0 | 191 | -1, // 'h' - UDAT_HOUR1_FIELD |
michael@0 | 192 | -1, // 'K' - UDAT_HOUR0_FIELD |
michael@0 | 193 | -1, // 'z' - UDAT_TIMEZONE_FIELD |
michael@0 | 194 | -1, // 'Y' - UDAT_YEAR_WOY_FIELD |
michael@0 | 195 | -1, // 'e' - UDAT_DOW_LOCAL_FIELD |
michael@0 | 196 | -1, // 'u' - UDAT_EXTENDED_YEAR_FIELD |
michael@0 | 197 | -1, // 'g' - UDAT_JULIAN_DAY_FIELD |
michael@0 | 198 | -1, // 'A' - UDAT_MILLISECONDS_IN_DAY_FIELD |
michael@0 | 199 | -1, // 'Z' - UDAT_TIMEZONE_RFC_FIELD |
michael@0 | 200 | -1, // 'v' - UDAT_TIMEZONE_GENERIC_FIELD |
michael@0 | 201 | 0, // 'c' - UDAT_STANDALONE_DAY_FIELD |
michael@0 | 202 | 1, // 'L' - UDAT_STANDALONE_MONTH_FIELD |
michael@0 | 203 | -1, // 'Q' - UDAT_QUARTER_FIELD (1-4?) |
michael@0 | 204 | -1, // 'q' - UDAT_STANDALONE_QUARTER_FIELD |
michael@0 | 205 | -1 // 'V' - UDAT_TIMEZONE_SPECIAL_FIELD |
michael@0 | 206 | -1, // 'U' - UDAT_YEAR_NAME_FIELD |
michael@0 | 207 | -1, // 'O' - UDAT_TIMEZONE_LOCALIZED_GMT_OFFSET_FIELD |
michael@0 | 208 | -1, // 'X' - UDAT_TIMEZONE_ISO_FIELD |
michael@0 | 209 | -1, // 'x' - UDAT_TIMEZONE_ISO_LOCAL_FIELD |
michael@0 | 210 | }; |
michael@0 | 211 | |
michael@0 | 212 | // When calendar uses hebr numbering (i.e. he@calendar=hebrew), |
michael@0 | 213 | // offset the years within the current millenium down to 1-999 |
michael@0 | 214 | static const int32_t HEBREW_CAL_CUR_MILLENIUM_START_YEAR = 5000; |
michael@0 | 215 | static const int32_t HEBREW_CAL_CUR_MILLENIUM_END_YEAR = 6000; |
michael@0 | 216 | |
michael@0 | 217 | static UMutex LOCK = U_MUTEX_INITIALIZER; |
michael@0 | 218 | |
michael@0 | 219 | UOBJECT_DEFINE_RTTI_IMPLEMENTATION(SimpleDateFormat) |
michael@0 | 220 | |
michael@0 | 221 | //---------------------------------------------------------------------- |
michael@0 | 222 | |
michael@0 | 223 | SimpleDateFormat::~SimpleDateFormat() |
michael@0 | 224 | { |
michael@0 | 225 | delete fSymbols; |
michael@0 | 226 | if (fNumberFormatters) { |
michael@0 | 227 | uprv_free(fNumberFormatters); |
michael@0 | 228 | } |
michael@0 | 229 | if (fTimeZoneFormat) { |
michael@0 | 230 | delete fTimeZoneFormat; |
michael@0 | 231 | } |
michael@0 | 232 | |
michael@0 | 233 | while (fOverrideList) { |
michael@0 | 234 | NSOverride *cur = fOverrideList; |
michael@0 | 235 | fOverrideList = cur->next; |
michael@0 | 236 | delete cur->nf; |
michael@0 | 237 | uprv_free(cur); |
michael@0 | 238 | } |
michael@0 | 239 | } |
michael@0 | 240 | |
michael@0 | 241 | //---------------------------------------------------------------------- |
michael@0 | 242 | |
michael@0 | 243 | SimpleDateFormat::SimpleDateFormat(UErrorCode& status) |
michael@0 | 244 | : fLocale(Locale::getDefault()), |
michael@0 | 245 | fSymbols(NULL), |
michael@0 | 246 | fTimeZoneFormat(NULL), |
michael@0 | 247 | fNumberFormatters(NULL), |
michael@0 | 248 | fOverrideList(NULL), |
michael@0 | 249 | fCapitalizationContext(UDISPCTX_CAPITALIZATION_NONE) |
michael@0 | 250 | { |
michael@0 | 251 | setBooleanAttribute(UDAT_PARSE_ALLOW_WHITESPACE, true, status).setBooleanAttribute(UDAT_PARSE_ALLOW_NUMERIC, true, status); |
michael@0 | 252 | construct(kShort, (EStyle) (kShort + kDateOffset), fLocale, status); |
michael@0 | 253 | initializeDefaultCentury(); |
michael@0 | 254 | } |
michael@0 | 255 | |
michael@0 | 256 | //---------------------------------------------------------------------- |
michael@0 | 257 | |
michael@0 | 258 | SimpleDateFormat::SimpleDateFormat(const UnicodeString& pattern, |
michael@0 | 259 | UErrorCode &status) |
michael@0 | 260 | : fPattern(pattern), |
michael@0 | 261 | fLocale(Locale::getDefault()), |
michael@0 | 262 | fSymbols(NULL), |
michael@0 | 263 | fTimeZoneFormat(NULL), |
michael@0 | 264 | fNumberFormatters(NULL), |
michael@0 | 265 | fOverrideList(NULL), |
michael@0 | 266 | fCapitalizationContext(UDISPCTX_CAPITALIZATION_NONE) |
michael@0 | 267 | { |
michael@0 | 268 | fDateOverride.setToBogus(); |
michael@0 | 269 | fTimeOverride.setToBogus(); |
michael@0 | 270 | setBooleanAttribute(UDAT_PARSE_ALLOW_WHITESPACE, true, status).setBooleanAttribute(UDAT_PARSE_ALLOW_NUMERIC, true, status); |
michael@0 | 271 | initializeSymbols(fLocale, initializeCalendar(NULL,fLocale,status), status); |
michael@0 | 272 | initialize(fLocale, status); |
michael@0 | 273 | initializeDefaultCentury(); |
michael@0 | 274 | |
michael@0 | 275 | } |
michael@0 | 276 | //---------------------------------------------------------------------- |
michael@0 | 277 | |
michael@0 | 278 | SimpleDateFormat::SimpleDateFormat(const UnicodeString& pattern, |
michael@0 | 279 | const UnicodeString& override, |
michael@0 | 280 | UErrorCode &status) |
michael@0 | 281 | : fPattern(pattern), |
michael@0 | 282 | fLocale(Locale::getDefault()), |
michael@0 | 283 | fSymbols(NULL), |
michael@0 | 284 | fTimeZoneFormat(NULL), |
michael@0 | 285 | fNumberFormatters(NULL), |
michael@0 | 286 | fOverrideList(NULL), |
michael@0 | 287 | fCapitalizationContext(UDISPCTX_CAPITALIZATION_NONE) |
michael@0 | 288 | { |
michael@0 | 289 | fDateOverride.setTo(override); |
michael@0 | 290 | fTimeOverride.setToBogus(); |
michael@0 | 291 | setBooleanAttribute(UDAT_PARSE_ALLOW_WHITESPACE, true, status).setBooleanAttribute(UDAT_PARSE_ALLOW_NUMERIC, true, status); |
michael@0 | 292 | initializeSymbols(fLocale, initializeCalendar(NULL,fLocale,status), status); |
michael@0 | 293 | initialize(fLocale, status); |
michael@0 | 294 | initializeDefaultCentury(); |
michael@0 | 295 | |
michael@0 | 296 | processOverrideString(fLocale,override,kOvrStrBoth,status); |
michael@0 | 297 | |
michael@0 | 298 | } |
michael@0 | 299 | |
michael@0 | 300 | //---------------------------------------------------------------------- |
michael@0 | 301 | |
michael@0 | 302 | SimpleDateFormat::SimpleDateFormat(const UnicodeString& pattern, |
michael@0 | 303 | const Locale& locale, |
michael@0 | 304 | UErrorCode& status) |
michael@0 | 305 | : fPattern(pattern), |
michael@0 | 306 | fLocale(locale), |
michael@0 | 307 | fTimeZoneFormat(NULL), |
michael@0 | 308 | fNumberFormatters(NULL), |
michael@0 | 309 | fOverrideList(NULL), |
michael@0 | 310 | fCapitalizationContext(UDISPCTX_CAPITALIZATION_NONE) |
michael@0 | 311 | { |
michael@0 | 312 | |
michael@0 | 313 | fDateOverride.setToBogus(); |
michael@0 | 314 | fTimeOverride.setToBogus(); |
michael@0 | 315 | setBooleanAttribute(UDAT_PARSE_ALLOW_WHITESPACE, true, status).setBooleanAttribute(UDAT_PARSE_ALLOW_NUMERIC, true, status); |
michael@0 | 316 | |
michael@0 | 317 | initializeSymbols(fLocale, initializeCalendar(NULL,fLocale,status), status); |
michael@0 | 318 | initialize(fLocale, status); |
michael@0 | 319 | initializeDefaultCentury(); |
michael@0 | 320 | } |
michael@0 | 321 | |
michael@0 | 322 | //---------------------------------------------------------------------- |
michael@0 | 323 | |
michael@0 | 324 | SimpleDateFormat::SimpleDateFormat(const UnicodeString& pattern, |
michael@0 | 325 | const UnicodeString& override, |
michael@0 | 326 | const Locale& locale, |
michael@0 | 327 | UErrorCode& status) |
michael@0 | 328 | : fPattern(pattern), |
michael@0 | 329 | fLocale(locale), |
michael@0 | 330 | fTimeZoneFormat(NULL), |
michael@0 | 331 | fNumberFormatters(NULL), |
michael@0 | 332 | fOverrideList(NULL), |
michael@0 | 333 | fCapitalizationContext(UDISPCTX_CAPITALIZATION_NONE) |
michael@0 | 334 | { |
michael@0 | 335 | |
michael@0 | 336 | fDateOverride.setTo(override); |
michael@0 | 337 | fTimeOverride.setToBogus(); |
michael@0 | 338 | setBooleanAttribute(UDAT_PARSE_ALLOW_WHITESPACE, true, status).setBooleanAttribute(UDAT_PARSE_ALLOW_NUMERIC, true, status); |
michael@0 | 339 | |
michael@0 | 340 | initializeSymbols(fLocale, initializeCalendar(NULL,fLocale,status), status); |
michael@0 | 341 | initialize(fLocale, status); |
michael@0 | 342 | initializeDefaultCentury(); |
michael@0 | 343 | |
michael@0 | 344 | processOverrideString(locale,override,kOvrStrBoth,status); |
michael@0 | 345 | |
michael@0 | 346 | } |
michael@0 | 347 | |
michael@0 | 348 | //---------------------------------------------------------------------- |
michael@0 | 349 | |
michael@0 | 350 | SimpleDateFormat::SimpleDateFormat(const UnicodeString& pattern, |
michael@0 | 351 | DateFormatSymbols* symbolsToAdopt, |
michael@0 | 352 | UErrorCode& status) |
michael@0 | 353 | : fPattern(pattern), |
michael@0 | 354 | fLocale(Locale::getDefault()), |
michael@0 | 355 | fSymbols(symbolsToAdopt), |
michael@0 | 356 | fTimeZoneFormat(NULL), |
michael@0 | 357 | fNumberFormatters(NULL), |
michael@0 | 358 | fOverrideList(NULL), |
michael@0 | 359 | fCapitalizationContext(UDISPCTX_CAPITALIZATION_NONE) |
michael@0 | 360 | { |
michael@0 | 361 | |
michael@0 | 362 | fDateOverride.setToBogus(); |
michael@0 | 363 | fTimeOverride.setToBogus(); |
michael@0 | 364 | setBooleanAttribute(UDAT_PARSE_ALLOW_WHITESPACE, true, status).setBooleanAttribute(UDAT_PARSE_ALLOW_NUMERIC, true, status); |
michael@0 | 365 | |
michael@0 | 366 | initializeCalendar(NULL,fLocale,status); |
michael@0 | 367 | initialize(fLocale, status); |
michael@0 | 368 | initializeDefaultCentury(); |
michael@0 | 369 | } |
michael@0 | 370 | |
michael@0 | 371 | //---------------------------------------------------------------------- |
michael@0 | 372 | |
michael@0 | 373 | SimpleDateFormat::SimpleDateFormat(const UnicodeString& pattern, |
michael@0 | 374 | const DateFormatSymbols& symbols, |
michael@0 | 375 | UErrorCode& status) |
michael@0 | 376 | : fPattern(pattern), |
michael@0 | 377 | fLocale(Locale::getDefault()), |
michael@0 | 378 | fSymbols(new DateFormatSymbols(symbols)), |
michael@0 | 379 | fTimeZoneFormat(NULL), |
michael@0 | 380 | fNumberFormatters(NULL), |
michael@0 | 381 | fOverrideList(NULL), |
michael@0 | 382 | fCapitalizationContext(UDISPCTX_CAPITALIZATION_NONE) |
michael@0 | 383 | { |
michael@0 | 384 | |
michael@0 | 385 | fDateOverride.setToBogus(); |
michael@0 | 386 | fTimeOverride.setToBogus(); |
michael@0 | 387 | setBooleanAttribute(UDAT_PARSE_ALLOW_WHITESPACE, true, status).setBooleanAttribute(UDAT_PARSE_ALLOW_NUMERIC, true, status); |
michael@0 | 388 | |
michael@0 | 389 | initializeCalendar(NULL, fLocale, status); |
michael@0 | 390 | initialize(fLocale, status); |
michael@0 | 391 | initializeDefaultCentury(); |
michael@0 | 392 | } |
michael@0 | 393 | |
michael@0 | 394 | //---------------------------------------------------------------------- |
michael@0 | 395 | |
michael@0 | 396 | // Not for public consumption; used by DateFormat |
michael@0 | 397 | SimpleDateFormat::SimpleDateFormat(EStyle timeStyle, |
michael@0 | 398 | EStyle dateStyle, |
michael@0 | 399 | const Locale& locale, |
michael@0 | 400 | UErrorCode& status) |
michael@0 | 401 | : fLocale(locale), |
michael@0 | 402 | fSymbols(NULL), |
michael@0 | 403 | fTimeZoneFormat(NULL), |
michael@0 | 404 | fNumberFormatters(NULL), |
michael@0 | 405 | fOverrideList(NULL), |
michael@0 | 406 | fCapitalizationContext(UDISPCTX_CAPITALIZATION_NONE) |
michael@0 | 407 | { |
michael@0 | 408 | setBooleanAttribute(UDAT_PARSE_ALLOW_WHITESPACE, true, status).setBooleanAttribute(UDAT_PARSE_ALLOW_NUMERIC, true, status); |
michael@0 | 409 | construct(timeStyle, dateStyle, fLocale, status); |
michael@0 | 410 | if(U_SUCCESS(status)) { |
michael@0 | 411 | initializeDefaultCentury(); |
michael@0 | 412 | } |
michael@0 | 413 | } |
michael@0 | 414 | |
michael@0 | 415 | //---------------------------------------------------------------------- |
michael@0 | 416 | |
michael@0 | 417 | /** |
michael@0 | 418 | * Not for public consumption; used by DateFormat. This constructor |
michael@0 | 419 | * never fails. If the resource data is not available, it uses the |
michael@0 | 420 | * the last resort symbols. |
michael@0 | 421 | */ |
michael@0 | 422 | SimpleDateFormat::SimpleDateFormat(const Locale& locale, |
michael@0 | 423 | UErrorCode& status) |
michael@0 | 424 | : fPattern(gDefaultPattern), |
michael@0 | 425 | fLocale(locale), |
michael@0 | 426 | fSymbols(NULL), |
michael@0 | 427 | fTimeZoneFormat(NULL), |
michael@0 | 428 | fNumberFormatters(NULL), |
michael@0 | 429 | fOverrideList(NULL), |
michael@0 | 430 | fCapitalizationContext(UDISPCTX_CAPITALIZATION_NONE) |
michael@0 | 431 | { |
michael@0 | 432 | if (U_FAILURE(status)) return; |
michael@0 | 433 | initializeSymbols(fLocale, initializeCalendar(NULL, fLocale, status),status); |
michael@0 | 434 | if (U_FAILURE(status)) |
michael@0 | 435 | { |
michael@0 | 436 | status = U_ZERO_ERROR; |
michael@0 | 437 | delete fSymbols; |
michael@0 | 438 | // This constructor doesn't fail; it uses last resort data |
michael@0 | 439 | fSymbols = new DateFormatSymbols(status); |
michael@0 | 440 | /* test for NULL */ |
michael@0 | 441 | if (fSymbols == 0) { |
michael@0 | 442 | status = U_MEMORY_ALLOCATION_ERROR; |
michael@0 | 443 | return; |
michael@0 | 444 | } |
michael@0 | 445 | } |
michael@0 | 446 | |
michael@0 | 447 | fDateOverride.setToBogus(); |
michael@0 | 448 | fTimeOverride.setToBogus(); |
michael@0 | 449 | setBooleanAttribute(UDAT_PARSE_ALLOW_WHITESPACE, true, status).setBooleanAttribute(UDAT_PARSE_ALLOW_NUMERIC, true, status); |
michael@0 | 450 | |
michael@0 | 451 | initialize(fLocale, status); |
michael@0 | 452 | if(U_SUCCESS(status)) { |
michael@0 | 453 | initializeDefaultCentury(); |
michael@0 | 454 | } |
michael@0 | 455 | } |
michael@0 | 456 | |
michael@0 | 457 | //---------------------------------------------------------------------- |
michael@0 | 458 | |
michael@0 | 459 | SimpleDateFormat::SimpleDateFormat(const SimpleDateFormat& other) |
michael@0 | 460 | : DateFormat(other), |
michael@0 | 461 | fLocale(other.fLocale), |
michael@0 | 462 | fSymbols(NULL), |
michael@0 | 463 | fTimeZoneFormat(NULL), |
michael@0 | 464 | fNumberFormatters(NULL), |
michael@0 | 465 | fOverrideList(NULL), |
michael@0 | 466 | fCapitalizationContext(UDISPCTX_CAPITALIZATION_NONE) |
michael@0 | 467 | { |
michael@0 | 468 | UErrorCode status = U_ZERO_ERROR; |
michael@0 | 469 | setBooleanAttribute(UDAT_PARSE_ALLOW_WHITESPACE, true, status).setBooleanAttribute(UDAT_PARSE_ALLOW_NUMERIC, true, status); |
michael@0 | 470 | *this = other; |
michael@0 | 471 | } |
michael@0 | 472 | |
michael@0 | 473 | //---------------------------------------------------------------------- |
michael@0 | 474 | |
michael@0 | 475 | SimpleDateFormat& SimpleDateFormat::operator=(const SimpleDateFormat& other) |
michael@0 | 476 | { |
michael@0 | 477 | if (this == &other) { |
michael@0 | 478 | return *this; |
michael@0 | 479 | } |
michael@0 | 480 | DateFormat::operator=(other); |
michael@0 | 481 | |
michael@0 | 482 | delete fSymbols; |
michael@0 | 483 | fSymbols = NULL; |
michael@0 | 484 | |
michael@0 | 485 | if (other.fSymbols) |
michael@0 | 486 | fSymbols = new DateFormatSymbols(*other.fSymbols); |
michael@0 | 487 | |
michael@0 | 488 | fDefaultCenturyStart = other.fDefaultCenturyStart; |
michael@0 | 489 | fDefaultCenturyStartYear = other.fDefaultCenturyStartYear; |
michael@0 | 490 | fHaveDefaultCentury = other.fHaveDefaultCentury; |
michael@0 | 491 | |
michael@0 | 492 | fPattern = other.fPattern; |
michael@0 | 493 | |
michael@0 | 494 | // TimeZoneFormat in ICU4C only depends on a locale for now |
michael@0 | 495 | if (fLocale != other.fLocale) { |
michael@0 | 496 | delete fTimeZoneFormat; |
michael@0 | 497 | fTimeZoneFormat = NULL; // forces lazy instantiation with the other locale |
michael@0 | 498 | fLocale = other.fLocale; |
michael@0 | 499 | } |
michael@0 | 500 | |
michael@0 | 501 | fCapitalizationContext = other.fCapitalizationContext; |
michael@0 | 502 | |
michael@0 | 503 | return *this; |
michael@0 | 504 | } |
michael@0 | 505 | |
michael@0 | 506 | //---------------------------------------------------------------------- |
michael@0 | 507 | |
michael@0 | 508 | Format* |
michael@0 | 509 | SimpleDateFormat::clone() const |
michael@0 | 510 | { |
michael@0 | 511 | return new SimpleDateFormat(*this); |
michael@0 | 512 | } |
michael@0 | 513 | |
michael@0 | 514 | //---------------------------------------------------------------------- |
michael@0 | 515 | |
michael@0 | 516 | UBool |
michael@0 | 517 | SimpleDateFormat::operator==(const Format& other) const |
michael@0 | 518 | { |
michael@0 | 519 | if (DateFormat::operator==(other)) { |
michael@0 | 520 | // DateFormat::operator== guarantees following cast is safe |
michael@0 | 521 | SimpleDateFormat* that = (SimpleDateFormat*)&other; |
michael@0 | 522 | return (fPattern == that->fPattern && |
michael@0 | 523 | fSymbols != NULL && // Check for pathological object |
michael@0 | 524 | that->fSymbols != NULL && // Check for pathological object |
michael@0 | 525 | *fSymbols == *that->fSymbols && |
michael@0 | 526 | fHaveDefaultCentury == that->fHaveDefaultCentury && |
michael@0 | 527 | fDefaultCenturyStart == that->fDefaultCenturyStart && |
michael@0 | 528 | fCapitalizationContext == that->fCapitalizationContext); |
michael@0 | 529 | } |
michael@0 | 530 | return FALSE; |
michael@0 | 531 | } |
michael@0 | 532 | |
michael@0 | 533 | //---------------------------------------------------------------------- |
michael@0 | 534 | |
michael@0 | 535 | void SimpleDateFormat::construct(EStyle timeStyle, |
michael@0 | 536 | EStyle dateStyle, |
michael@0 | 537 | const Locale& locale, |
michael@0 | 538 | UErrorCode& status) |
michael@0 | 539 | { |
michael@0 | 540 | // called by several constructors to load pattern data from the resources |
michael@0 | 541 | if (U_FAILURE(status)) return; |
michael@0 | 542 | |
michael@0 | 543 | // We will need the calendar to know what type of symbols to load. |
michael@0 | 544 | initializeCalendar(NULL, locale, status); |
michael@0 | 545 | if (U_FAILURE(status)) return; |
michael@0 | 546 | |
michael@0 | 547 | CalendarData calData(locale, fCalendar?fCalendar->getType():NULL, status); |
michael@0 | 548 | UResourceBundle *dateTimePatterns = calData.getByKey(gDateTimePatternsTag, status); |
michael@0 | 549 | UResourceBundle *currentBundle; |
michael@0 | 550 | |
michael@0 | 551 | if (U_FAILURE(status)) return; |
michael@0 | 552 | |
michael@0 | 553 | if (ures_getSize(dateTimePatterns) <= kDateTime) |
michael@0 | 554 | { |
michael@0 | 555 | status = U_INVALID_FORMAT_ERROR; |
michael@0 | 556 | return; |
michael@0 | 557 | } |
michael@0 | 558 | |
michael@0 | 559 | setLocaleIDs(ures_getLocaleByType(dateTimePatterns, ULOC_VALID_LOCALE, &status), |
michael@0 | 560 | ures_getLocaleByType(dateTimePatterns, ULOC_ACTUAL_LOCALE, &status)); |
michael@0 | 561 | |
michael@0 | 562 | // create a symbols object from the locale |
michael@0 | 563 | initializeSymbols(locale,fCalendar, status); |
michael@0 | 564 | if (U_FAILURE(status)) return; |
michael@0 | 565 | /* test for NULL */ |
michael@0 | 566 | if (fSymbols == 0) { |
michael@0 | 567 | status = U_MEMORY_ALLOCATION_ERROR; |
michael@0 | 568 | return; |
michael@0 | 569 | } |
michael@0 | 570 | |
michael@0 | 571 | const UChar *resStr,*ovrStr; |
michael@0 | 572 | int32_t resStrLen,ovrStrLen = 0; |
michael@0 | 573 | fDateOverride.setToBogus(); |
michael@0 | 574 | fTimeOverride.setToBogus(); |
michael@0 | 575 | |
michael@0 | 576 | // if the pattern should include both date and time information, use the date/time |
michael@0 | 577 | // pattern string as a guide to tell use how to glue together the appropriate date |
michael@0 | 578 | // and time pattern strings. The actual gluing-together is handled by a convenience |
michael@0 | 579 | // method on MessageFormat. |
michael@0 | 580 | if ((timeStyle != kNone) && (dateStyle != kNone)) |
michael@0 | 581 | { |
michael@0 | 582 | Formattable timeDateArray[2]; |
michael@0 | 583 | |
michael@0 | 584 | // use Formattable::adoptString() so that we can use fastCopyFrom() |
michael@0 | 585 | // instead of Formattable::setString()'s unaware, safe, deep string clone |
michael@0 | 586 | // see Jitterbug 2296 |
michael@0 | 587 | |
michael@0 | 588 | currentBundle = ures_getByIndex(dateTimePatterns, (int32_t)timeStyle, NULL, &status); |
michael@0 | 589 | if (U_FAILURE(status)) { |
michael@0 | 590 | status = U_INVALID_FORMAT_ERROR; |
michael@0 | 591 | return; |
michael@0 | 592 | } |
michael@0 | 593 | switch (ures_getType(currentBundle)) { |
michael@0 | 594 | case URES_STRING: { |
michael@0 | 595 | resStr = ures_getString(currentBundle, &resStrLen, &status); |
michael@0 | 596 | break; |
michael@0 | 597 | } |
michael@0 | 598 | case URES_ARRAY: { |
michael@0 | 599 | resStr = ures_getStringByIndex(currentBundle, 0, &resStrLen, &status); |
michael@0 | 600 | ovrStr = ures_getStringByIndex(currentBundle, 1, &ovrStrLen, &status); |
michael@0 | 601 | fTimeOverride.setTo(TRUE, ovrStr, ovrStrLen); |
michael@0 | 602 | break; |
michael@0 | 603 | } |
michael@0 | 604 | default: { |
michael@0 | 605 | status = U_INVALID_FORMAT_ERROR; |
michael@0 | 606 | ures_close(currentBundle); |
michael@0 | 607 | return; |
michael@0 | 608 | } |
michael@0 | 609 | } |
michael@0 | 610 | ures_close(currentBundle); |
michael@0 | 611 | |
michael@0 | 612 | UnicodeString *tempus1 = new UnicodeString(TRUE, resStr, resStrLen); |
michael@0 | 613 | // NULL pointer check |
michael@0 | 614 | if (tempus1 == NULL) { |
michael@0 | 615 | status = U_MEMORY_ALLOCATION_ERROR; |
michael@0 | 616 | return; |
michael@0 | 617 | } |
michael@0 | 618 | timeDateArray[0].adoptString(tempus1); |
michael@0 | 619 | |
michael@0 | 620 | currentBundle = ures_getByIndex(dateTimePatterns, (int32_t)dateStyle, NULL, &status); |
michael@0 | 621 | if (U_FAILURE(status)) { |
michael@0 | 622 | status = U_INVALID_FORMAT_ERROR; |
michael@0 | 623 | return; |
michael@0 | 624 | } |
michael@0 | 625 | switch (ures_getType(currentBundle)) { |
michael@0 | 626 | case URES_STRING: { |
michael@0 | 627 | resStr = ures_getString(currentBundle, &resStrLen, &status); |
michael@0 | 628 | break; |
michael@0 | 629 | } |
michael@0 | 630 | case URES_ARRAY: { |
michael@0 | 631 | resStr = ures_getStringByIndex(currentBundle, 0, &resStrLen, &status); |
michael@0 | 632 | ovrStr = ures_getStringByIndex(currentBundle, 1, &ovrStrLen, &status); |
michael@0 | 633 | fDateOverride.setTo(TRUE, ovrStr, ovrStrLen); |
michael@0 | 634 | break; |
michael@0 | 635 | } |
michael@0 | 636 | default: { |
michael@0 | 637 | status = U_INVALID_FORMAT_ERROR; |
michael@0 | 638 | ures_close(currentBundle); |
michael@0 | 639 | return; |
michael@0 | 640 | } |
michael@0 | 641 | } |
michael@0 | 642 | ures_close(currentBundle); |
michael@0 | 643 | |
michael@0 | 644 | UnicodeString *tempus2 = new UnicodeString(TRUE, resStr, resStrLen); |
michael@0 | 645 | // Null pointer check |
michael@0 | 646 | if (tempus2 == NULL) { |
michael@0 | 647 | status = U_MEMORY_ALLOCATION_ERROR; |
michael@0 | 648 | return; |
michael@0 | 649 | } |
michael@0 | 650 | timeDateArray[1].adoptString(tempus2); |
michael@0 | 651 | |
michael@0 | 652 | int32_t glueIndex = kDateTime; |
michael@0 | 653 | int32_t patternsSize = ures_getSize(dateTimePatterns); |
michael@0 | 654 | if (patternsSize >= (kDateTimeOffset + kShort + 1)) { |
michael@0 | 655 | // Get proper date time format |
michael@0 | 656 | glueIndex = (int32_t)(kDateTimeOffset + (dateStyle - kDateOffset)); |
michael@0 | 657 | } |
michael@0 | 658 | |
michael@0 | 659 | resStr = ures_getStringByIndex(dateTimePatterns, glueIndex, &resStrLen, &status); |
michael@0 | 660 | MessageFormat::format(UnicodeString(TRUE, resStr, resStrLen), timeDateArray, 2, fPattern, status); |
michael@0 | 661 | } |
michael@0 | 662 | // if the pattern includes just time data or just date date, load the appropriate |
michael@0 | 663 | // pattern string from the resources |
michael@0 | 664 | // setTo() - see DateFormatSymbols::assignArray comments |
michael@0 | 665 | else if (timeStyle != kNone) { |
michael@0 | 666 | currentBundle = ures_getByIndex(dateTimePatterns, (int32_t)timeStyle, NULL, &status); |
michael@0 | 667 | if (U_FAILURE(status)) { |
michael@0 | 668 | status = U_INVALID_FORMAT_ERROR; |
michael@0 | 669 | return; |
michael@0 | 670 | } |
michael@0 | 671 | switch (ures_getType(currentBundle)) { |
michael@0 | 672 | case URES_STRING: { |
michael@0 | 673 | resStr = ures_getString(currentBundle, &resStrLen, &status); |
michael@0 | 674 | break; |
michael@0 | 675 | } |
michael@0 | 676 | case URES_ARRAY: { |
michael@0 | 677 | resStr = ures_getStringByIndex(currentBundle, 0, &resStrLen, &status); |
michael@0 | 678 | ovrStr = ures_getStringByIndex(currentBundle, 1, &ovrStrLen, &status); |
michael@0 | 679 | fDateOverride.setTo(TRUE, ovrStr, ovrStrLen); |
michael@0 | 680 | break; |
michael@0 | 681 | } |
michael@0 | 682 | default: { |
michael@0 | 683 | status = U_INVALID_FORMAT_ERROR; |
michael@0 | 684 | ures_close(currentBundle); |
michael@0 | 685 | return; |
michael@0 | 686 | } |
michael@0 | 687 | } |
michael@0 | 688 | fPattern.setTo(TRUE, resStr, resStrLen); |
michael@0 | 689 | ures_close(currentBundle); |
michael@0 | 690 | } |
michael@0 | 691 | else if (dateStyle != kNone) { |
michael@0 | 692 | currentBundle = ures_getByIndex(dateTimePatterns, (int32_t)dateStyle, NULL, &status); |
michael@0 | 693 | if (U_FAILURE(status)) { |
michael@0 | 694 | status = U_INVALID_FORMAT_ERROR; |
michael@0 | 695 | return; |
michael@0 | 696 | } |
michael@0 | 697 | switch (ures_getType(currentBundle)) { |
michael@0 | 698 | case URES_STRING: { |
michael@0 | 699 | resStr = ures_getString(currentBundle, &resStrLen, &status); |
michael@0 | 700 | break; |
michael@0 | 701 | } |
michael@0 | 702 | case URES_ARRAY: { |
michael@0 | 703 | resStr = ures_getStringByIndex(currentBundle, 0, &resStrLen, &status); |
michael@0 | 704 | ovrStr = ures_getStringByIndex(currentBundle, 1, &ovrStrLen, &status); |
michael@0 | 705 | fDateOverride.setTo(TRUE, ovrStr, ovrStrLen); |
michael@0 | 706 | break; |
michael@0 | 707 | } |
michael@0 | 708 | default: { |
michael@0 | 709 | status = U_INVALID_FORMAT_ERROR; |
michael@0 | 710 | ures_close(currentBundle); |
michael@0 | 711 | return; |
michael@0 | 712 | } |
michael@0 | 713 | } |
michael@0 | 714 | fPattern.setTo(TRUE, resStr, resStrLen); |
michael@0 | 715 | ures_close(currentBundle); |
michael@0 | 716 | } |
michael@0 | 717 | |
michael@0 | 718 | // and if it includes _neither_, that's an error |
michael@0 | 719 | else |
michael@0 | 720 | status = U_INVALID_FORMAT_ERROR; |
michael@0 | 721 | |
michael@0 | 722 | // finally, finish initializing by creating a Calendar and a NumberFormat |
michael@0 | 723 | initialize(locale, status); |
michael@0 | 724 | } |
michael@0 | 725 | |
michael@0 | 726 | //---------------------------------------------------------------------- |
michael@0 | 727 | |
michael@0 | 728 | Calendar* |
michael@0 | 729 | SimpleDateFormat::initializeCalendar(TimeZone* adoptZone, const Locale& locale, UErrorCode& status) |
michael@0 | 730 | { |
michael@0 | 731 | if(!U_FAILURE(status)) { |
michael@0 | 732 | fCalendar = Calendar::createInstance(adoptZone?adoptZone:TimeZone::createDefault(), locale, status); |
michael@0 | 733 | } |
michael@0 | 734 | if (U_SUCCESS(status) && fCalendar == NULL) { |
michael@0 | 735 | status = U_MEMORY_ALLOCATION_ERROR; |
michael@0 | 736 | } |
michael@0 | 737 | return fCalendar; |
michael@0 | 738 | } |
michael@0 | 739 | |
michael@0 | 740 | void |
michael@0 | 741 | SimpleDateFormat::initializeSymbols(const Locale& locale, Calendar* calendar, UErrorCode& status) |
michael@0 | 742 | { |
michael@0 | 743 | if(U_FAILURE(status)) { |
michael@0 | 744 | fSymbols = NULL; |
michael@0 | 745 | } else { |
michael@0 | 746 | // pass in calendar type - use NULL (default) if no calendar set (or err). |
michael@0 | 747 | fSymbols = new DateFormatSymbols(locale, calendar?calendar->getType() :NULL , status); |
michael@0 | 748 | // Null pointer check |
michael@0 | 749 | if (fSymbols == NULL) { |
michael@0 | 750 | status = U_MEMORY_ALLOCATION_ERROR; |
michael@0 | 751 | return; |
michael@0 | 752 | } |
michael@0 | 753 | } |
michael@0 | 754 | } |
michael@0 | 755 | |
michael@0 | 756 | void |
michael@0 | 757 | SimpleDateFormat::initialize(const Locale& locale, |
michael@0 | 758 | UErrorCode& status) |
michael@0 | 759 | { |
michael@0 | 760 | if (U_FAILURE(status)) return; |
michael@0 | 761 | |
michael@0 | 762 | // We don't need to check that the row count is >= 1, since all 2d arrays have at |
michael@0 | 763 | // least one row |
michael@0 | 764 | fNumberFormat = NumberFormat::createInstance(locale, status); |
michael@0 | 765 | if (fNumberFormat != NULL && U_SUCCESS(status)) |
michael@0 | 766 | { |
michael@0 | 767 | // no matter what the locale's default number format looked like, we want |
michael@0 | 768 | // to modify it so that it doesn't use thousands separators, doesn't always |
michael@0 | 769 | // show the decimal point, and recognizes integers only when parsing |
michael@0 | 770 | |
michael@0 | 771 | fNumberFormat->setGroupingUsed(FALSE); |
michael@0 | 772 | DecimalFormat* decfmt = dynamic_cast<DecimalFormat*>(fNumberFormat); |
michael@0 | 773 | if (decfmt != NULL) { |
michael@0 | 774 | decfmt->setDecimalSeparatorAlwaysShown(FALSE); |
michael@0 | 775 | } |
michael@0 | 776 | fNumberFormat->setParseIntegerOnly(TRUE); |
michael@0 | 777 | fNumberFormat->setMinimumFractionDigits(0); // To prevent "Jan 1.00, 1997.00" |
michael@0 | 778 | |
michael@0 | 779 | //fNumberFormat->setLenient(TRUE); // Java uses a custom DateNumberFormat to format/parse |
michael@0 | 780 | |
michael@0 | 781 | initNumberFormatters(locale,status); |
michael@0 | 782 | |
michael@0 | 783 | } |
michael@0 | 784 | else if (U_SUCCESS(status)) |
michael@0 | 785 | { |
michael@0 | 786 | status = U_MISSING_RESOURCE_ERROR; |
michael@0 | 787 | } |
michael@0 | 788 | } |
michael@0 | 789 | |
michael@0 | 790 | /* Initialize the fields we use to disambiguate ambiguous years. Separate |
michael@0 | 791 | * so we can call it from readObject(). |
michael@0 | 792 | */ |
michael@0 | 793 | void SimpleDateFormat::initializeDefaultCentury() |
michael@0 | 794 | { |
michael@0 | 795 | if(fCalendar) { |
michael@0 | 796 | fHaveDefaultCentury = fCalendar->haveDefaultCentury(); |
michael@0 | 797 | if(fHaveDefaultCentury) { |
michael@0 | 798 | fDefaultCenturyStart = fCalendar->defaultCenturyStart(); |
michael@0 | 799 | fDefaultCenturyStartYear = fCalendar->defaultCenturyStartYear(); |
michael@0 | 800 | } else { |
michael@0 | 801 | fDefaultCenturyStart = DBL_MIN; |
michael@0 | 802 | fDefaultCenturyStartYear = -1; |
michael@0 | 803 | } |
michael@0 | 804 | } |
michael@0 | 805 | } |
michael@0 | 806 | |
michael@0 | 807 | /* Define one-century window into which to disambiguate dates using |
michael@0 | 808 | * two-digit years. Make public in JDK 1.2. |
michael@0 | 809 | */ |
michael@0 | 810 | void SimpleDateFormat::parseAmbiguousDatesAsAfter(UDate startDate, UErrorCode& status) |
michael@0 | 811 | { |
michael@0 | 812 | if(U_FAILURE(status)) { |
michael@0 | 813 | return; |
michael@0 | 814 | } |
michael@0 | 815 | if(!fCalendar) { |
michael@0 | 816 | status = U_ILLEGAL_ARGUMENT_ERROR; |
michael@0 | 817 | return; |
michael@0 | 818 | } |
michael@0 | 819 | |
michael@0 | 820 | fCalendar->setTime(startDate, status); |
michael@0 | 821 | if(U_SUCCESS(status)) { |
michael@0 | 822 | fHaveDefaultCentury = TRUE; |
michael@0 | 823 | fDefaultCenturyStart = startDate; |
michael@0 | 824 | fDefaultCenturyStartYear = fCalendar->get(UCAL_YEAR, status); |
michael@0 | 825 | } |
michael@0 | 826 | } |
michael@0 | 827 | |
michael@0 | 828 | //---------------------------------------------------------------------- |
michael@0 | 829 | |
michael@0 | 830 | UnicodeString& |
michael@0 | 831 | SimpleDateFormat::format(Calendar& cal, UnicodeString& appendTo, FieldPosition& pos) const |
michael@0 | 832 | { |
michael@0 | 833 | UErrorCode status = U_ZERO_ERROR; |
michael@0 | 834 | FieldPositionOnlyHandler handler(pos); |
michael@0 | 835 | return _format(cal, appendTo, handler, status); |
michael@0 | 836 | } |
michael@0 | 837 | |
michael@0 | 838 | //---------------------------------------------------------------------- |
michael@0 | 839 | |
michael@0 | 840 | UnicodeString& |
michael@0 | 841 | SimpleDateFormat::format(Calendar& cal, UnicodeString& appendTo, |
michael@0 | 842 | FieldPositionIterator* posIter, UErrorCode& status) const |
michael@0 | 843 | { |
michael@0 | 844 | FieldPositionIteratorHandler handler(posIter, status); |
michael@0 | 845 | return _format(cal, appendTo, handler, status); |
michael@0 | 846 | } |
michael@0 | 847 | |
michael@0 | 848 | //---------------------------------------------------------------------- |
michael@0 | 849 | |
michael@0 | 850 | UnicodeString& |
michael@0 | 851 | SimpleDateFormat::_format(Calendar& cal, UnicodeString& appendTo, |
michael@0 | 852 | FieldPositionHandler& handler, UErrorCode& status) const |
michael@0 | 853 | { |
michael@0 | 854 | if ( U_FAILURE(status) ) { |
michael@0 | 855 | return appendTo; |
michael@0 | 856 | } |
michael@0 | 857 | Calendar* workCal = &cal; |
michael@0 | 858 | Calendar* calClone = NULL; |
michael@0 | 859 | if (&cal != fCalendar && uprv_strcmp(cal.getType(), fCalendar->getType()) != 0) { |
michael@0 | 860 | // Different calendar type |
michael@0 | 861 | // We use the time and time zone from the input calendar, but |
michael@0 | 862 | // do not use the input calendar for field calculation. |
michael@0 | 863 | calClone = fCalendar->clone(); |
michael@0 | 864 | if (calClone != NULL) { |
michael@0 | 865 | UDate t = cal.getTime(status); |
michael@0 | 866 | calClone->setTime(t, status); |
michael@0 | 867 | calClone->setTimeZone(cal.getTimeZone()); |
michael@0 | 868 | workCal = calClone; |
michael@0 | 869 | } else { |
michael@0 | 870 | status = U_MEMORY_ALLOCATION_ERROR; |
michael@0 | 871 | return appendTo; |
michael@0 | 872 | } |
michael@0 | 873 | } |
michael@0 | 874 | |
michael@0 | 875 | UBool inQuote = FALSE; |
michael@0 | 876 | UChar prevCh = 0; |
michael@0 | 877 | int32_t count = 0; |
michael@0 | 878 | int32_t fieldNum = 0; |
michael@0 | 879 | |
michael@0 | 880 | // loop through the pattern string character by character |
michael@0 | 881 | for (int32_t i = 0; i < fPattern.length() && U_SUCCESS(status); ++i) { |
michael@0 | 882 | UChar ch = fPattern[i]; |
michael@0 | 883 | |
michael@0 | 884 | // Use subFormat() to format a repeated pattern character |
michael@0 | 885 | // when a different pattern or non-pattern character is seen |
michael@0 | 886 | if (ch != prevCh && count > 0) { |
michael@0 | 887 | subFormat(appendTo, prevCh, count, fCapitalizationContext, fieldNum++, handler, *workCal, status); |
michael@0 | 888 | count = 0; |
michael@0 | 889 | } |
michael@0 | 890 | if (ch == QUOTE) { |
michael@0 | 891 | // Consecutive single quotes are a single quote literal, |
michael@0 | 892 | // either outside of quotes or between quotes |
michael@0 | 893 | if ((i+1) < fPattern.length() && fPattern[i+1] == QUOTE) { |
michael@0 | 894 | appendTo += (UChar)QUOTE; |
michael@0 | 895 | ++i; |
michael@0 | 896 | } else { |
michael@0 | 897 | inQuote = ! inQuote; |
michael@0 | 898 | } |
michael@0 | 899 | } |
michael@0 | 900 | else if ( ! inQuote && ((ch >= 0x0061 /*'a'*/ && ch <= 0x007A /*'z'*/) |
michael@0 | 901 | || (ch >= 0x0041 /*'A'*/ && ch <= 0x005A /*'Z'*/))) { |
michael@0 | 902 | // ch is a date-time pattern character to be interpreted |
michael@0 | 903 | // by subFormat(); count the number of times it is repeated |
michael@0 | 904 | prevCh = ch; |
michael@0 | 905 | ++count; |
michael@0 | 906 | } |
michael@0 | 907 | else { |
michael@0 | 908 | // Append quoted characters and unquoted non-pattern characters |
michael@0 | 909 | appendTo += ch; |
michael@0 | 910 | } |
michael@0 | 911 | } |
michael@0 | 912 | |
michael@0 | 913 | // Format the last item in the pattern, if any |
michael@0 | 914 | if (count > 0) { |
michael@0 | 915 | subFormat(appendTo, prevCh, count, fCapitalizationContext, fieldNum++, handler, *workCal, status); |
michael@0 | 916 | } |
michael@0 | 917 | |
michael@0 | 918 | if (calClone != NULL) { |
michael@0 | 919 | delete calClone; |
michael@0 | 920 | } |
michael@0 | 921 | |
michael@0 | 922 | return appendTo; |
michael@0 | 923 | } |
michael@0 | 924 | |
michael@0 | 925 | //---------------------------------------------------------------------- |
michael@0 | 926 | |
michael@0 | 927 | /* Map calendar field into calendar field level. |
michael@0 | 928 | * the larger the level, the smaller the field unit. |
michael@0 | 929 | * For example, UCAL_ERA level is 0, UCAL_YEAR level is 10, |
michael@0 | 930 | * UCAL_MONTH level is 20. |
michael@0 | 931 | * NOTE: if new fields adds in, the table needs to update. |
michael@0 | 932 | */ |
michael@0 | 933 | const int32_t |
michael@0 | 934 | SimpleDateFormat::fgCalendarFieldToLevel[] = |
michael@0 | 935 | { |
michael@0 | 936 | /*GyM*/ 0, 10, 20, |
michael@0 | 937 | /*wW*/ 20, 30, |
michael@0 | 938 | /*dDEF*/ 30, 20, 30, 30, |
michael@0 | 939 | /*ahHm*/ 40, 50, 50, 60, |
michael@0 | 940 | /*sS..*/ 70, 80, |
michael@0 | 941 | /*z?Y*/ 0, 0, 10, |
michael@0 | 942 | /*eug*/ 30, 10, 0, |
michael@0 | 943 | /*A*/ 40 |
michael@0 | 944 | }; |
michael@0 | 945 | |
michael@0 | 946 | |
michael@0 | 947 | /* Map calendar field LETTER into calendar field level. |
michael@0 | 948 | * the larger the level, the smaller the field unit. |
michael@0 | 949 | * NOTE: if new fields adds in, the table needs to update. |
michael@0 | 950 | */ |
michael@0 | 951 | const int32_t |
michael@0 | 952 | SimpleDateFormat::fgPatternCharToLevel[] = { |
michael@0 | 953 | // A B C D E F G H I J K L M N O |
michael@0 | 954 | -1, 40, -1, -1, 20, 30, 30, 0, 50, -1, -1, 50, 20, 20, -1, 0, |
michael@0 | 955 | // P Q R S T U V W X Y Z |
michael@0 | 956 | -1, 20, -1, 80, -1, 10, 0, 30, 0, 10, 0, -1, -1, -1, -1, -1, |
michael@0 | 957 | // a b c d e f g h i j k l m n o |
michael@0 | 958 | -1, 40, -1, 30, 30, 30, -1, 0, 50, -1, -1, 50, -1, 60, -1, -1, |
michael@0 | 959 | // p q r s t u v w x y z |
michael@0 | 960 | -1, 20, -1, 70, -1, 10, 0, 20, 0, 10, 0, -1, -1, -1, -1, -1 |
michael@0 | 961 | }; |
michael@0 | 962 | |
michael@0 | 963 | |
michael@0 | 964 | // Map index into pattern character string to Calendar field number. |
michael@0 | 965 | const UCalendarDateFields |
michael@0 | 966 | SimpleDateFormat::fgPatternIndexToCalendarField[] = |
michael@0 | 967 | { |
michael@0 | 968 | /*GyM*/ UCAL_ERA, UCAL_YEAR, UCAL_MONTH, |
michael@0 | 969 | /*dkH*/ UCAL_DATE, UCAL_HOUR_OF_DAY, UCAL_HOUR_OF_DAY, |
michael@0 | 970 | /*msS*/ UCAL_MINUTE, UCAL_SECOND, UCAL_MILLISECOND, |
michael@0 | 971 | /*EDF*/ UCAL_DAY_OF_WEEK, UCAL_DAY_OF_YEAR, UCAL_DAY_OF_WEEK_IN_MONTH, |
michael@0 | 972 | /*wWa*/ UCAL_WEEK_OF_YEAR, UCAL_WEEK_OF_MONTH, UCAL_AM_PM, |
michael@0 | 973 | /*hKz*/ UCAL_HOUR, UCAL_HOUR, UCAL_ZONE_OFFSET, |
michael@0 | 974 | /*Yeu*/ UCAL_YEAR_WOY, UCAL_DOW_LOCAL, UCAL_EXTENDED_YEAR, |
michael@0 | 975 | /*gAZ*/ UCAL_JULIAN_DAY, UCAL_MILLISECONDS_IN_DAY, UCAL_ZONE_OFFSET, |
michael@0 | 976 | /*v*/ UCAL_ZONE_OFFSET, |
michael@0 | 977 | /*c*/ UCAL_DOW_LOCAL, |
michael@0 | 978 | /*L*/ UCAL_MONTH, |
michael@0 | 979 | /*Q*/ UCAL_MONTH, |
michael@0 | 980 | /*q*/ UCAL_MONTH, |
michael@0 | 981 | /*V*/ UCAL_ZONE_OFFSET, |
michael@0 | 982 | /*U*/ UCAL_YEAR, |
michael@0 | 983 | /*O*/ UCAL_ZONE_OFFSET, |
michael@0 | 984 | /*Xx*/ UCAL_ZONE_OFFSET, UCAL_ZONE_OFFSET, |
michael@0 | 985 | }; |
michael@0 | 986 | |
michael@0 | 987 | // Map index into pattern character string to DateFormat field number |
michael@0 | 988 | const UDateFormatField |
michael@0 | 989 | SimpleDateFormat::fgPatternIndexToDateFormatField[] = { |
michael@0 | 990 | /*GyM*/ UDAT_ERA_FIELD, UDAT_YEAR_FIELD, UDAT_MONTH_FIELD, |
michael@0 | 991 | /*dkH*/ UDAT_DATE_FIELD, UDAT_HOUR_OF_DAY1_FIELD, UDAT_HOUR_OF_DAY0_FIELD, |
michael@0 | 992 | /*msS*/ UDAT_MINUTE_FIELD, UDAT_SECOND_FIELD, UDAT_FRACTIONAL_SECOND_FIELD, |
michael@0 | 993 | /*EDF*/ UDAT_DAY_OF_WEEK_FIELD, UDAT_DAY_OF_YEAR_FIELD, UDAT_DAY_OF_WEEK_IN_MONTH_FIELD, |
michael@0 | 994 | /*wWa*/ UDAT_WEEK_OF_YEAR_FIELD, UDAT_WEEK_OF_MONTH_FIELD, UDAT_AM_PM_FIELD, |
michael@0 | 995 | /*hKz*/ UDAT_HOUR1_FIELD, UDAT_HOUR0_FIELD, UDAT_TIMEZONE_FIELD, |
michael@0 | 996 | /*Yeu*/ UDAT_YEAR_WOY_FIELD, UDAT_DOW_LOCAL_FIELD, UDAT_EXTENDED_YEAR_FIELD, |
michael@0 | 997 | /*gAZ*/ UDAT_JULIAN_DAY_FIELD, UDAT_MILLISECONDS_IN_DAY_FIELD, UDAT_TIMEZONE_RFC_FIELD, |
michael@0 | 998 | /*v*/ UDAT_TIMEZONE_GENERIC_FIELD, |
michael@0 | 999 | /*c*/ UDAT_STANDALONE_DAY_FIELD, |
michael@0 | 1000 | /*L*/ UDAT_STANDALONE_MONTH_FIELD, |
michael@0 | 1001 | /*Q*/ UDAT_QUARTER_FIELD, |
michael@0 | 1002 | /*q*/ UDAT_STANDALONE_QUARTER_FIELD, |
michael@0 | 1003 | /*V*/ UDAT_TIMEZONE_SPECIAL_FIELD, |
michael@0 | 1004 | /*U*/ UDAT_YEAR_NAME_FIELD, |
michael@0 | 1005 | /*O*/ UDAT_TIMEZONE_LOCALIZED_GMT_OFFSET_FIELD, |
michael@0 | 1006 | /*Xx*/ UDAT_TIMEZONE_ISO_FIELD, UDAT_TIMEZONE_ISO_LOCAL_FIELD, |
michael@0 | 1007 | }; |
michael@0 | 1008 | |
michael@0 | 1009 | //---------------------------------------------------------------------- |
michael@0 | 1010 | |
michael@0 | 1011 | /** |
michael@0 | 1012 | * Append symbols[value] to dst. Make sure the array index is not out |
michael@0 | 1013 | * of bounds. |
michael@0 | 1014 | */ |
michael@0 | 1015 | static inline void |
michael@0 | 1016 | _appendSymbol(UnicodeString& dst, |
michael@0 | 1017 | int32_t value, |
michael@0 | 1018 | const UnicodeString* symbols, |
michael@0 | 1019 | int32_t symbolsCount) { |
michael@0 | 1020 | U_ASSERT(0 <= value && value < symbolsCount); |
michael@0 | 1021 | if (0 <= value && value < symbolsCount) { |
michael@0 | 1022 | dst += symbols[value]; |
michael@0 | 1023 | } |
michael@0 | 1024 | } |
michael@0 | 1025 | |
michael@0 | 1026 | static inline void |
michael@0 | 1027 | _appendSymbolWithMonthPattern(UnicodeString& dst, int32_t value, const UnicodeString* symbols, int32_t symbolsCount, |
michael@0 | 1028 | const UnicodeString* monthPattern, UErrorCode& status) { |
michael@0 | 1029 | U_ASSERT(0 <= value && value < symbolsCount); |
michael@0 | 1030 | if (0 <= value && value < symbolsCount) { |
michael@0 | 1031 | if (monthPattern == NULL) { |
michael@0 | 1032 | dst += symbols[value]; |
michael@0 | 1033 | } else { |
michael@0 | 1034 | Formattable monthName((const UnicodeString&)(symbols[value])); |
michael@0 | 1035 | MessageFormat::format(*monthPattern, &monthName, 1, dst, status); |
michael@0 | 1036 | } |
michael@0 | 1037 | } |
michael@0 | 1038 | } |
michael@0 | 1039 | |
michael@0 | 1040 | //---------------------------------------------------------------------- |
michael@0 | 1041 | void |
michael@0 | 1042 | SimpleDateFormat::initNumberFormatters(const Locale &locale,UErrorCode &status) { |
michael@0 | 1043 | if (U_FAILURE(status)) { |
michael@0 | 1044 | return; |
michael@0 | 1045 | } |
michael@0 | 1046 | if ( fDateOverride.isBogus() && fTimeOverride.isBogus() ) { |
michael@0 | 1047 | return; |
michael@0 | 1048 | } |
michael@0 | 1049 | umtx_lock(&LOCK); |
michael@0 | 1050 | if (fNumberFormatters == NULL) { |
michael@0 | 1051 | fNumberFormatters = (NumberFormat**)uprv_malloc(UDAT_FIELD_COUNT * sizeof(NumberFormat*)); |
michael@0 | 1052 | if (fNumberFormatters) { |
michael@0 | 1053 | for (int32_t i = 0; i < UDAT_FIELD_COUNT; i++) { |
michael@0 | 1054 | fNumberFormatters[i] = fNumberFormat; |
michael@0 | 1055 | } |
michael@0 | 1056 | } else { |
michael@0 | 1057 | status = U_MEMORY_ALLOCATION_ERROR; |
michael@0 | 1058 | } |
michael@0 | 1059 | } |
michael@0 | 1060 | umtx_unlock(&LOCK); |
michael@0 | 1061 | |
michael@0 | 1062 | processOverrideString(locale,fDateOverride,kOvrStrDate,status); |
michael@0 | 1063 | processOverrideString(locale,fTimeOverride,kOvrStrTime,status); |
michael@0 | 1064 | |
michael@0 | 1065 | } |
michael@0 | 1066 | |
michael@0 | 1067 | void |
michael@0 | 1068 | SimpleDateFormat::processOverrideString(const Locale &locale, const UnicodeString &str, int8_t type, UErrorCode &status) { |
michael@0 | 1069 | if (str.isBogus()) { |
michael@0 | 1070 | return; |
michael@0 | 1071 | } |
michael@0 | 1072 | int32_t start = 0; |
michael@0 | 1073 | int32_t len; |
michael@0 | 1074 | UnicodeString nsName; |
michael@0 | 1075 | UnicodeString ovrField; |
michael@0 | 1076 | UBool moreToProcess = TRUE; |
michael@0 | 1077 | |
michael@0 | 1078 | while (moreToProcess) { |
michael@0 | 1079 | int32_t delimiterPosition = str.indexOf((UChar)ULOC_KEYWORD_ITEM_SEPARATOR_UNICODE,start); |
michael@0 | 1080 | if (delimiterPosition == -1) { |
michael@0 | 1081 | moreToProcess = FALSE; |
michael@0 | 1082 | len = str.length() - start; |
michael@0 | 1083 | } else { |
michael@0 | 1084 | len = delimiterPosition - start; |
michael@0 | 1085 | } |
michael@0 | 1086 | UnicodeString currentString(str,start,len); |
michael@0 | 1087 | int32_t equalSignPosition = currentString.indexOf((UChar)ULOC_KEYWORD_ASSIGN_UNICODE,0); |
michael@0 | 1088 | if (equalSignPosition == -1) { // Simple override string such as "hebrew" |
michael@0 | 1089 | nsName.setTo(currentString); |
michael@0 | 1090 | ovrField.setToBogus(); |
michael@0 | 1091 | } else { // Field specific override string such as "y=hebrew" |
michael@0 | 1092 | nsName.setTo(currentString,equalSignPosition+1); |
michael@0 | 1093 | ovrField.setTo(currentString,0,1); // We just need the first character. |
michael@0 | 1094 | } |
michael@0 | 1095 | |
michael@0 | 1096 | int32_t nsNameHash = nsName.hashCode(); |
michael@0 | 1097 | // See if the numbering system is in the override list, if not, then add it. |
michael@0 | 1098 | NSOverride *cur = fOverrideList; |
michael@0 | 1099 | NumberFormat *nf = NULL; |
michael@0 | 1100 | UBool found = FALSE; |
michael@0 | 1101 | while ( cur && !found ) { |
michael@0 | 1102 | if ( cur->hash == nsNameHash ) { |
michael@0 | 1103 | nf = cur->nf; |
michael@0 | 1104 | found = TRUE; |
michael@0 | 1105 | } |
michael@0 | 1106 | cur = cur->next; |
michael@0 | 1107 | } |
michael@0 | 1108 | |
michael@0 | 1109 | if (!found) { |
michael@0 | 1110 | cur = (NSOverride *)uprv_malloc(sizeof(NSOverride)); |
michael@0 | 1111 | if (cur) { |
michael@0 | 1112 | char kw[ULOC_KEYWORD_AND_VALUES_CAPACITY]; |
michael@0 | 1113 | uprv_strcpy(kw,"numbers="); |
michael@0 | 1114 | nsName.extract(0,len,kw+8,ULOC_KEYWORD_AND_VALUES_CAPACITY-8,US_INV); |
michael@0 | 1115 | |
michael@0 | 1116 | Locale ovrLoc(locale.getLanguage(),locale.getCountry(),locale.getVariant(),kw); |
michael@0 | 1117 | nf = NumberFormat::createInstance(ovrLoc,status); |
michael@0 | 1118 | |
michael@0 | 1119 | // no matter what the locale's default number format looked like, we want |
michael@0 | 1120 | // to modify it so that it doesn't use thousands separators, doesn't always |
michael@0 | 1121 | // show the decimal point, and recognizes integers only when parsing |
michael@0 | 1122 | |
michael@0 | 1123 | if (U_SUCCESS(status)) { |
michael@0 | 1124 | nf->setGroupingUsed(FALSE); |
michael@0 | 1125 | DecimalFormat* decfmt = dynamic_cast<DecimalFormat*>(nf); |
michael@0 | 1126 | if (decfmt != NULL) { |
michael@0 | 1127 | decfmt->setDecimalSeparatorAlwaysShown(FALSE); |
michael@0 | 1128 | } |
michael@0 | 1129 | nf->setParseIntegerOnly(TRUE); |
michael@0 | 1130 | nf->setMinimumFractionDigits(0); // To prevent "Jan 1.00, 1997.00" |
michael@0 | 1131 | |
michael@0 | 1132 | cur->nf = nf; |
michael@0 | 1133 | cur->hash = nsNameHash; |
michael@0 | 1134 | cur->next = fOverrideList; |
michael@0 | 1135 | fOverrideList = cur; |
michael@0 | 1136 | } |
michael@0 | 1137 | else { |
michael@0 | 1138 | // clean up before returning |
michael@0 | 1139 | if (cur != NULL) { |
michael@0 | 1140 | uprv_free(cur); |
michael@0 | 1141 | } |
michael@0 | 1142 | return; |
michael@0 | 1143 | } |
michael@0 | 1144 | |
michael@0 | 1145 | } else { |
michael@0 | 1146 | status = U_MEMORY_ALLOCATION_ERROR; |
michael@0 | 1147 | return; |
michael@0 | 1148 | } |
michael@0 | 1149 | } |
michael@0 | 1150 | |
michael@0 | 1151 | // Now that we have an appropriate number formatter, fill in the appropriate spaces in the |
michael@0 | 1152 | // number formatters table. |
michael@0 | 1153 | |
michael@0 | 1154 | if (ovrField.isBogus()) { |
michael@0 | 1155 | switch (type) { |
michael@0 | 1156 | case kOvrStrDate: |
michael@0 | 1157 | case kOvrStrBoth: { |
michael@0 | 1158 | for ( int8_t i=0 ; i<kDateFieldsCount; i++ ) { |
michael@0 | 1159 | fNumberFormatters[kDateFields[i]] = nf; |
michael@0 | 1160 | } |
michael@0 | 1161 | if (type==kOvrStrDate) { |
michael@0 | 1162 | break; |
michael@0 | 1163 | } |
michael@0 | 1164 | } |
michael@0 | 1165 | case kOvrStrTime : { |
michael@0 | 1166 | for ( int8_t i=0 ; i<kTimeFieldsCount; i++ ) { |
michael@0 | 1167 | fNumberFormatters[kTimeFields[i]] = nf; |
michael@0 | 1168 | } |
michael@0 | 1169 | break; |
michael@0 | 1170 | } |
michael@0 | 1171 | } |
michael@0 | 1172 | } else { |
michael@0 | 1173 | // if the pattern character is unrecognized, signal an error and bail out |
michael@0 | 1174 | UDateFormatField patternCharIndex = |
michael@0 | 1175 | DateFormatSymbols::getPatternCharIndex(ovrField.charAt(0)); |
michael@0 | 1176 | if (patternCharIndex == UDAT_FIELD_COUNT) { |
michael@0 | 1177 | status = U_INVALID_FORMAT_ERROR; |
michael@0 | 1178 | return; |
michael@0 | 1179 | } |
michael@0 | 1180 | |
michael@0 | 1181 | // Set the number formatter in the table |
michael@0 | 1182 | fNumberFormatters[patternCharIndex] = nf; |
michael@0 | 1183 | } |
michael@0 | 1184 | |
michael@0 | 1185 | start = delimiterPosition + 1; |
michael@0 | 1186 | } |
michael@0 | 1187 | } |
michael@0 | 1188 | |
michael@0 | 1189 | //--------------------------------------------------------------------- |
michael@0 | 1190 | void |
michael@0 | 1191 | SimpleDateFormat::subFormat(UnicodeString &appendTo, |
michael@0 | 1192 | UChar ch, |
michael@0 | 1193 | int32_t count, |
michael@0 | 1194 | UDisplayContext capitalizationContext, |
michael@0 | 1195 | int32_t fieldNum, |
michael@0 | 1196 | FieldPositionHandler& handler, |
michael@0 | 1197 | Calendar& cal, |
michael@0 | 1198 | UErrorCode& status) const |
michael@0 | 1199 | { |
michael@0 | 1200 | if (U_FAILURE(status)) { |
michael@0 | 1201 | return; |
michael@0 | 1202 | } |
michael@0 | 1203 | |
michael@0 | 1204 | // this function gets called by format() to produce the appropriate substitution |
michael@0 | 1205 | // text for an individual pattern symbol (e.g., "HH" or "yyyy") |
michael@0 | 1206 | |
michael@0 | 1207 | UDateFormatField patternCharIndex = DateFormatSymbols::getPatternCharIndex(ch); |
michael@0 | 1208 | const int32_t maxIntCount = 10; |
michael@0 | 1209 | int32_t beginOffset = appendTo.length(); |
michael@0 | 1210 | NumberFormat *currentNumberFormat; |
michael@0 | 1211 | DateFormatSymbols::ECapitalizationContextUsageType capContextUsageType = DateFormatSymbols::kCapContextUsageOther; |
michael@0 | 1212 | |
michael@0 | 1213 | UBool isHebrewCalendar = (uprv_strcmp(cal.getType(),"hebrew") == 0); |
michael@0 | 1214 | UBool isChineseCalendar = (uprv_strcmp(cal.getType(),"chinese") == 0 || uprv_strcmp(cal.getType(),"dangi") == 0); |
michael@0 | 1215 | |
michael@0 | 1216 | // if the pattern character is unrecognized, signal an error and dump out |
michael@0 | 1217 | if (patternCharIndex == UDAT_FIELD_COUNT) |
michael@0 | 1218 | { |
michael@0 | 1219 | if (ch != 0x6C) { // pattern char 'l' (SMALL LETTER L) just gets ignored |
michael@0 | 1220 | status = U_INVALID_FORMAT_ERROR; |
michael@0 | 1221 | } |
michael@0 | 1222 | return; |
michael@0 | 1223 | } |
michael@0 | 1224 | |
michael@0 | 1225 | UCalendarDateFields field = fgPatternIndexToCalendarField[patternCharIndex]; |
michael@0 | 1226 | int32_t value = cal.get(field, status); |
michael@0 | 1227 | if (U_FAILURE(status)) { |
michael@0 | 1228 | return; |
michael@0 | 1229 | } |
michael@0 | 1230 | |
michael@0 | 1231 | currentNumberFormat = getNumberFormatByIndex(patternCharIndex); |
michael@0 | 1232 | UnicodeString hebr("hebr", 4, US_INV); |
michael@0 | 1233 | |
michael@0 | 1234 | switch (patternCharIndex) { |
michael@0 | 1235 | |
michael@0 | 1236 | // for any "G" symbol, write out the appropriate era string |
michael@0 | 1237 | // "GGGG" is wide era name, "GGGGG" is narrow era name, anything else is abbreviated name |
michael@0 | 1238 | case UDAT_ERA_FIELD: |
michael@0 | 1239 | if (isChineseCalendar) { |
michael@0 | 1240 | zeroPaddingNumber(currentNumberFormat,appendTo, value, 1, 9); // as in ICU4J |
michael@0 | 1241 | } else { |
michael@0 | 1242 | if (count == 5) { |
michael@0 | 1243 | _appendSymbol(appendTo, value, fSymbols->fNarrowEras, fSymbols->fNarrowErasCount); |
michael@0 | 1244 | capContextUsageType = DateFormatSymbols::kCapContextUsageEraNarrow; |
michael@0 | 1245 | } else if (count == 4) { |
michael@0 | 1246 | _appendSymbol(appendTo, value, fSymbols->fEraNames, fSymbols->fEraNamesCount); |
michael@0 | 1247 | capContextUsageType = DateFormatSymbols::kCapContextUsageEraWide; |
michael@0 | 1248 | } else { |
michael@0 | 1249 | _appendSymbol(appendTo, value, fSymbols->fEras, fSymbols->fErasCount); |
michael@0 | 1250 | capContextUsageType = DateFormatSymbols::kCapContextUsageEraAbbrev; |
michael@0 | 1251 | } |
michael@0 | 1252 | } |
michael@0 | 1253 | break; |
michael@0 | 1254 | |
michael@0 | 1255 | case UDAT_YEAR_NAME_FIELD: |
michael@0 | 1256 | if (fSymbols->fShortYearNames != NULL && value <= fSymbols->fShortYearNamesCount) { |
michael@0 | 1257 | // the Calendar YEAR field runs 1 through 60 for cyclic years |
michael@0 | 1258 | _appendSymbol(appendTo, value - 1, fSymbols->fShortYearNames, fSymbols->fShortYearNamesCount); |
michael@0 | 1259 | break; |
michael@0 | 1260 | } |
michael@0 | 1261 | // else fall through to numeric year handling, do not break here |
michael@0 | 1262 | |
michael@0 | 1263 | // OLD: for "yyyy", write out the whole year; for "yy", write out the last 2 digits |
michael@0 | 1264 | // NEW: UTS#35: |
michael@0 | 1265 | //Year y yy yyy yyyy yyyyy |
michael@0 | 1266 | //AD 1 1 01 001 0001 00001 |
michael@0 | 1267 | //AD 12 12 12 012 0012 00012 |
michael@0 | 1268 | //AD 123 123 23 123 0123 00123 |
michael@0 | 1269 | //AD 1234 1234 34 1234 1234 01234 |
michael@0 | 1270 | //AD 12345 12345 45 12345 12345 12345 |
michael@0 | 1271 | case UDAT_YEAR_FIELD: |
michael@0 | 1272 | case UDAT_YEAR_WOY_FIELD: |
michael@0 | 1273 | if (fDateOverride.compare(hebr)==0 && value>HEBREW_CAL_CUR_MILLENIUM_START_YEAR && value<HEBREW_CAL_CUR_MILLENIUM_END_YEAR) { |
michael@0 | 1274 | value-=HEBREW_CAL_CUR_MILLENIUM_START_YEAR; |
michael@0 | 1275 | } |
michael@0 | 1276 | if(count == 2) |
michael@0 | 1277 | zeroPaddingNumber(currentNumberFormat, appendTo, value, 2, 2); |
michael@0 | 1278 | else |
michael@0 | 1279 | zeroPaddingNumber(currentNumberFormat, appendTo, value, count, maxIntCount); |
michael@0 | 1280 | break; |
michael@0 | 1281 | |
michael@0 | 1282 | // for "MMMM"/"LLLL", write out the whole month name, for "MMM"/"LLL", write out the month |
michael@0 | 1283 | // abbreviation, for "M"/"L" or "MM"/"LL", write out the month as a number with the |
michael@0 | 1284 | // appropriate number of digits |
michael@0 | 1285 | // for "MMMMM"/"LLLLL", use the narrow form |
michael@0 | 1286 | case UDAT_MONTH_FIELD: |
michael@0 | 1287 | case UDAT_STANDALONE_MONTH_FIELD: |
michael@0 | 1288 | if ( isHebrewCalendar ) { |
michael@0 | 1289 | HebrewCalendar *hc = (HebrewCalendar*)&cal; |
michael@0 | 1290 | if (hc->isLeapYear(hc->get(UCAL_YEAR,status)) && value == 6 && count >= 3 ) |
michael@0 | 1291 | value = 13; // Show alternate form for Adar II in leap years in Hebrew calendar. |
michael@0 | 1292 | if (!hc->isLeapYear(hc->get(UCAL_YEAR,status)) && value >= 6 && count < 3 ) |
michael@0 | 1293 | value--; // Adjust the month number down 1 in Hebrew non-leap years, i.e. Adar is 6, not 7. |
michael@0 | 1294 | } |
michael@0 | 1295 | { |
michael@0 | 1296 | int32_t isLeapMonth = (fSymbols->fLeapMonthPatterns != NULL && fSymbols->fLeapMonthPatternsCount >= DateFormatSymbols::kMonthPatternsCount)? |
michael@0 | 1297 | cal.get(UCAL_IS_LEAP_MONTH, status): 0; |
michael@0 | 1298 | // should consolidate the next section by using arrays of pointers & counts for the right symbols... |
michael@0 | 1299 | if (count == 5) { |
michael@0 | 1300 | if (patternCharIndex == UDAT_MONTH_FIELD) { |
michael@0 | 1301 | _appendSymbolWithMonthPattern(appendTo, value, fSymbols->fNarrowMonths, fSymbols->fNarrowMonthsCount, |
michael@0 | 1302 | (isLeapMonth!=0)? &(fSymbols->fLeapMonthPatterns[DateFormatSymbols::kLeapMonthPatternFormatNarrow]): NULL, status); |
michael@0 | 1303 | } else { |
michael@0 | 1304 | _appendSymbolWithMonthPattern(appendTo, value, fSymbols->fStandaloneNarrowMonths, fSymbols->fStandaloneNarrowMonthsCount, |
michael@0 | 1305 | (isLeapMonth!=0)? &(fSymbols->fLeapMonthPatterns[DateFormatSymbols::kLeapMonthPatternStandaloneNarrow]): NULL, status); |
michael@0 | 1306 | } |
michael@0 | 1307 | capContextUsageType = DateFormatSymbols::kCapContextUsageMonthNarrow; |
michael@0 | 1308 | } else if (count == 4) { |
michael@0 | 1309 | if (patternCharIndex == UDAT_MONTH_FIELD) { |
michael@0 | 1310 | _appendSymbolWithMonthPattern(appendTo, value, fSymbols->fMonths, fSymbols->fMonthsCount, |
michael@0 | 1311 | (isLeapMonth!=0)? &(fSymbols->fLeapMonthPatterns[DateFormatSymbols::kLeapMonthPatternFormatWide]): NULL, status); |
michael@0 | 1312 | capContextUsageType = DateFormatSymbols::kCapContextUsageMonthFormat; |
michael@0 | 1313 | } else { |
michael@0 | 1314 | _appendSymbolWithMonthPattern(appendTo, value, fSymbols->fStandaloneMonths, fSymbols->fStandaloneMonthsCount, |
michael@0 | 1315 | (isLeapMonth!=0)? &(fSymbols->fLeapMonthPatterns[DateFormatSymbols::kLeapMonthPatternStandaloneWide]): NULL, status); |
michael@0 | 1316 | capContextUsageType = DateFormatSymbols::kCapContextUsageMonthStandalone; |
michael@0 | 1317 | } |
michael@0 | 1318 | } else if (count == 3) { |
michael@0 | 1319 | if (patternCharIndex == UDAT_MONTH_FIELD) { |
michael@0 | 1320 | _appendSymbolWithMonthPattern(appendTo, value, fSymbols->fShortMonths, fSymbols->fShortMonthsCount, |
michael@0 | 1321 | (isLeapMonth!=0)? &(fSymbols->fLeapMonthPatterns[DateFormatSymbols::kLeapMonthPatternFormatAbbrev]): NULL, status); |
michael@0 | 1322 | capContextUsageType = DateFormatSymbols::kCapContextUsageMonthFormat; |
michael@0 | 1323 | } else { |
michael@0 | 1324 | _appendSymbolWithMonthPattern(appendTo, value, fSymbols->fStandaloneShortMonths, fSymbols->fStandaloneShortMonthsCount, |
michael@0 | 1325 | (isLeapMonth!=0)? &(fSymbols->fLeapMonthPatterns[DateFormatSymbols::kLeapMonthPatternStandaloneAbbrev]): NULL, status); |
michael@0 | 1326 | capContextUsageType = DateFormatSymbols::kCapContextUsageMonthStandalone; |
michael@0 | 1327 | } |
michael@0 | 1328 | } else { |
michael@0 | 1329 | UnicodeString monthNumber; |
michael@0 | 1330 | zeroPaddingNumber(currentNumberFormat,monthNumber, value + 1, count, maxIntCount); |
michael@0 | 1331 | _appendSymbolWithMonthPattern(appendTo, 0, &monthNumber, 1, |
michael@0 | 1332 | (isLeapMonth!=0)? &(fSymbols->fLeapMonthPatterns[DateFormatSymbols::kLeapMonthPatternNumeric]): NULL, status); |
michael@0 | 1333 | } |
michael@0 | 1334 | } |
michael@0 | 1335 | break; |
michael@0 | 1336 | |
michael@0 | 1337 | // for "k" and "kk", write out the hour, adjusting midnight to appear as "24" |
michael@0 | 1338 | case UDAT_HOUR_OF_DAY1_FIELD: |
michael@0 | 1339 | if (value == 0) |
michael@0 | 1340 | zeroPaddingNumber(currentNumberFormat,appendTo, cal.getMaximum(UCAL_HOUR_OF_DAY) + 1, count, maxIntCount); |
michael@0 | 1341 | else |
michael@0 | 1342 | zeroPaddingNumber(currentNumberFormat,appendTo, value, count, maxIntCount); |
michael@0 | 1343 | break; |
michael@0 | 1344 | |
michael@0 | 1345 | case UDAT_FRACTIONAL_SECOND_FIELD: |
michael@0 | 1346 | // Fractional seconds left-justify |
michael@0 | 1347 | { |
michael@0 | 1348 | currentNumberFormat->setMinimumIntegerDigits((count > 3) ? 3 : count); |
michael@0 | 1349 | currentNumberFormat->setMaximumIntegerDigits(maxIntCount); |
michael@0 | 1350 | if (count == 1) { |
michael@0 | 1351 | value /= 100; |
michael@0 | 1352 | } else if (count == 2) { |
michael@0 | 1353 | value /= 10; |
michael@0 | 1354 | } |
michael@0 | 1355 | FieldPosition p(0); |
michael@0 | 1356 | currentNumberFormat->format(value, appendTo, p); |
michael@0 | 1357 | if (count > 3) { |
michael@0 | 1358 | currentNumberFormat->setMinimumIntegerDigits(count - 3); |
michael@0 | 1359 | currentNumberFormat->format((int32_t)0, appendTo, p); |
michael@0 | 1360 | } |
michael@0 | 1361 | } |
michael@0 | 1362 | break; |
michael@0 | 1363 | |
michael@0 | 1364 | // for "ee" or "e", use local numeric day-of-the-week |
michael@0 | 1365 | // for "EEEEEE" or "eeeeee", write out the short day-of-the-week name |
michael@0 | 1366 | // for "EEEEE" or "eeeee", write out the narrow day-of-the-week name |
michael@0 | 1367 | // for "EEEE" or "eeee", write out the wide day-of-the-week name |
michael@0 | 1368 | // for "EEE" or "EE" or "E" or "eee", write out the abbreviated day-of-the-week name |
michael@0 | 1369 | case UDAT_DOW_LOCAL_FIELD: |
michael@0 | 1370 | if ( count < 3 ) { |
michael@0 | 1371 | zeroPaddingNumber(currentNumberFormat,appendTo, value, count, maxIntCount); |
michael@0 | 1372 | break; |
michael@0 | 1373 | } |
michael@0 | 1374 | // fall through to EEEEE-EEE handling, but for that we don't want local day-of-week, |
michael@0 | 1375 | // we want standard day-of-week, so first fix value to work for EEEEE-EEE. |
michael@0 | 1376 | value = cal.get(UCAL_DAY_OF_WEEK, status); |
michael@0 | 1377 | if (U_FAILURE(status)) { |
michael@0 | 1378 | return; |
michael@0 | 1379 | } |
michael@0 | 1380 | // fall through, do not break here |
michael@0 | 1381 | case UDAT_DAY_OF_WEEK_FIELD: |
michael@0 | 1382 | if (count == 5) { |
michael@0 | 1383 | _appendSymbol(appendTo, value, fSymbols->fNarrowWeekdays, |
michael@0 | 1384 | fSymbols->fNarrowWeekdaysCount); |
michael@0 | 1385 | capContextUsageType = DateFormatSymbols::kCapContextUsageDayNarrow; |
michael@0 | 1386 | } else if (count == 4) { |
michael@0 | 1387 | _appendSymbol(appendTo, value, fSymbols->fWeekdays, |
michael@0 | 1388 | fSymbols->fWeekdaysCount); |
michael@0 | 1389 | capContextUsageType = DateFormatSymbols::kCapContextUsageDayFormat; |
michael@0 | 1390 | } else if (count == 6) { |
michael@0 | 1391 | _appendSymbol(appendTo, value, fSymbols->fShorterWeekdays, |
michael@0 | 1392 | fSymbols->fShorterWeekdaysCount); |
michael@0 | 1393 | capContextUsageType = DateFormatSymbols::kCapContextUsageDayFormat; |
michael@0 | 1394 | } else { |
michael@0 | 1395 | _appendSymbol(appendTo, value, fSymbols->fShortWeekdays, |
michael@0 | 1396 | fSymbols->fShortWeekdaysCount); |
michael@0 | 1397 | capContextUsageType = DateFormatSymbols::kCapContextUsageDayFormat; |
michael@0 | 1398 | } |
michael@0 | 1399 | break; |
michael@0 | 1400 | |
michael@0 | 1401 | // for "ccc", write out the abbreviated day-of-the-week name |
michael@0 | 1402 | // for "cccc", write out the wide day-of-the-week name |
michael@0 | 1403 | // for "ccccc", use the narrow day-of-the-week name |
michael@0 | 1404 | // for "ccccc", use the short day-of-the-week name |
michael@0 | 1405 | case UDAT_STANDALONE_DAY_FIELD: |
michael@0 | 1406 | if ( count < 3 ) { |
michael@0 | 1407 | zeroPaddingNumber(currentNumberFormat,appendTo, value, 1, maxIntCount); |
michael@0 | 1408 | break; |
michael@0 | 1409 | } |
michael@0 | 1410 | // fall through to alpha DOW handling, but for that we don't want local day-of-week, |
michael@0 | 1411 | // we want standard day-of-week, so first fix value. |
michael@0 | 1412 | value = cal.get(UCAL_DAY_OF_WEEK, status); |
michael@0 | 1413 | if (U_FAILURE(status)) { |
michael@0 | 1414 | return; |
michael@0 | 1415 | } |
michael@0 | 1416 | if (count == 5) { |
michael@0 | 1417 | _appendSymbol(appendTo, value, fSymbols->fStandaloneNarrowWeekdays, |
michael@0 | 1418 | fSymbols->fStandaloneNarrowWeekdaysCount); |
michael@0 | 1419 | capContextUsageType = DateFormatSymbols::kCapContextUsageDayNarrow; |
michael@0 | 1420 | } else if (count == 4) { |
michael@0 | 1421 | _appendSymbol(appendTo, value, fSymbols->fStandaloneWeekdays, |
michael@0 | 1422 | fSymbols->fStandaloneWeekdaysCount); |
michael@0 | 1423 | capContextUsageType = DateFormatSymbols::kCapContextUsageDayStandalone; |
michael@0 | 1424 | } else if (count == 6) { |
michael@0 | 1425 | _appendSymbol(appendTo, value, fSymbols->fStandaloneShorterWeekdays, |
michael@0 | 1426 | fSymbols->fStandaloneShorterWeekdaysCount); |
michael@0 | 1427 | capContextUsageType = DateFormatSymbols::kCapContextUsageDayStandalone; |
michael@0 | 1428 | } else { // count == 3 |
michael@0 | 1429 | _appendSymbol(appendTo, value, fSymbols->fStandaloneShortWeekdays, |
michael@0 | 1430 | fSymbols->fStandaloneShortWeekdaysCount); |
michael@0 | 1431 | capContextUsageType = DateFormatSymbols::kCapContextUsageDayStandalone; |
michael@0 | 1432 | } |
michael@0 | 1433 | break; |
michael@0 | 1434 | |
michael@0 | 1435 | // for and "a" symbol, write out the whole AM/PM string |
michael@0 | 1436 | case UDAT_AM_PM_FIELD: |
michael@0 | 1437 | _appendSymbol(appendTo, value, fSymbols->fAmPms, |
michael@0 | 1438 | fSymbols->fAmPmsCount); |
michael@0 | 1439 | break; |
michael@0 | 1440 | |
michael@0 | 1441 | // for "h" and "hh", write out the hour, adjusting noon and midnight to show up |
michael@0 | 1442 | // as "12" |
michael@0 | 1443 | case UDAT_HOUR1_FIELD: |
michael@0 | 1444 | if (value == 0) |
michael@0 | 1445 | zeroPaddingNumber(currentNumberFormat,appendTo, cal.getLeastMaximum(UCAL_HOUR) + 1, count, maxIntCount); |
michael@0 | 1446 | else |
michael@0 | 1447 | zeroPaddingNumber(currentNumberFormat,appendTo, value, count, maxIntCount); |
michael@0 | 1448 | break; |
michael@0 | 1449 | |
michael@0 | 1450 | case UDAT_TIMEZONE_FIELD: // 'z' |
michael@0 | 1451 | case UDAT_TIMEZONE_RFC_FIELD: // 'Z' |
michael@0 | 1452 | case UDAT_TIMEZONE_GENERIC_FIELD: // 'v' |
michael@0 | 1453 | case UDAT_TIMEZONE_SPECIAL_FIELD: // 'V' |
michael@0 | 1454 | case UDAT_TIMEZONE_LOCALIZED_GMT_OFFSET_FIELD: // 'O' |
michael@0 | 1455 | case UDAT_TIMEZONE_ISO_FIELD: // 'X' |
michael@0 | 1456 | case UDAT_TIMEZONE_ISO_LOCAL_FIELD: // 'x' |
michael@0 | 1457 | { |
michael@0 | 1458 | UnicodeString zoneString; |
michael@0 | 1459 | const TimeZone& tz = cal.getTimeZone(); |
michael@0 | 1460 | UDate date = cal.getTime(status); |
michael@0 | 1461 | if (U_SUCCESS(status)) { |
michael@0 | 1462 | if (patternCharIndex == UDAT_TIMEZONE_FIELD) { |
michael@0 | 1463 | if (count < 4) { |
michael@0 | 1464 | // "z", "zz", "zzz" |
michael@0 | 1465 | tzFormat()->format(UTZFMT_STYLE_SPECIFIC_SHORT, tz, date, zoneString); |
michael@0 | 1466 | capContextUsageType = DateFormatSymbols::kCapContextUsageMetazoneShort; |
michael@0 | 1467 | } else { |
michael@0 | 1468 | // "zzzz" or longer |
michael@0 | 1469 | tzFormat()->format(UTZFMT_STYLE_SPECIFIC_LONG, tz, date, zoneString); |
michael@0 | 1470 | capContextUsageType = DateFormatSymbols::kCapContextUsageMetazoneLong; |
michael@0 | 1471 | } |
michael@0 | 1472 | } |
michael@0 | 1473 | else if (patternCharIndex == UDAT_TIMEZONE_RFC_FIELD) { |
michael@0 | 1474 | if (count < 4) { |
michael@0 | 1475 | // "Z" |
michael@0 | 1476 | tzFormat()->format(UTZFMT_STYLE_ISO_BASIC_LOCAL_FULL, tz, date, zoneString); |
michael@0 | 1477 | } else if (count == 5) { |
michael@0 | 1478 | // "ZZZZZ" |
michael@0 | 1479 | tzFormat()->format(UTZFMT_STYLE_ISO_EXTENDED_FULL, tz, date, zoneString); |
michael@0 | 1480 | } else { |
michael@0 | 1481 | // "ZZ", "ZZZ", "ZZZZ" |
michael@0 | 1482 | tzFormat()->format(UTZFMT_STYLE_LOCALIZED_GMT, tz, date, zoneString); |
michael@0 | 1483 | } |
michael@0 | 1484 | } |
michael@0 | 1485 | else if (patternCharIndex == UDAT_TIMEZONE_GENERIC_FIELD) { |
michael@0 | 1486 | if (count == 1) { |
michael@0 | 1487 | // "v" |
michael@0 | 1488 | tzFormat()->format(UTZFMT_STYLE_GENERIC_SHORT, tz, date, zoneString); |
michael@0 | 1489 | capContextUsageType = DateFormatSymbols::kCapContextUsageMetazoneShort; |
michael@0 | 1490 | } else if (count == 4) { |
michael@0 | 1491 | // "vvvv" |
michael@0 | 1492 | tzFormat()->format(UTZFMT_STYLE_GENERIC_LONG, tz, date, zoneString); |
michael@0 | 1493 | capContextUsageType = DateFormatSymbols::kCapContextUsageMetazoneLong; |
michael@0 | 1494 | } |
michael@0 | 1495 | } |
michael@0 | 1496 | else if (patternCharIndex == UDAT_TIMEZONE_SPECIAL_FIELD) { |
michael@0 | 1497 | if (count == 1) { |
michael@0 | 1498 | // "V" |
michael@0 | 1499 | tzFormat()->format(UTZFMT_STYLE_ZONE_ID_SHORT, tz, date, zoneString); |
michael@0 | 1500 | } else if (count == 2) { |
michael@0 | 1501 | // "VV" |
michael@0 | 1502 | tzFormat()->format(UTZFMT_STYLE_ZONE_ID, tz, date, zoneString); |
michael@0 | 1503 | } else if (count == 3) { |
michael@0 | 1504 | // "VVV" |
michael@0 | 1505 | tzFormat()->format(UTZFMT_STYLE_EXEMPLAR_LOCATION, tz, date, zoneString); |
michael@0 | 1506 | } else if (count == 4) { |
michael@0 | 1507 | // "VVVV" |
michael@0 | 1508 | tzFormat()->format(UTZFMT_STYLE_GENERIC_LOCATION, tz, date, zoneString); |
michael@0 | 1509 | capContextUsageType = DateFormatSymbols::kCapContextUsageZoneLong; |
michael@0 | 1510 | } |
michael@0 | 1511 | } |
michael@0 | 1512 | else if (patternCharIndex == UDAT_TIMEZONE_LOCALIZED_GMT_OFFSET_FIELD) { |
michael@0 | 1513 | if (count == 1) { |
michael@0 | 1514 | // "O" |
michael@0 | 1515 | tzFormat()->format(UTZFMT_STYLE_LOCALIZED_GMT_SHORT, tz, date, zoneString); |
michael@0 | 1516 | } else if (count == 4) { |
michael@0 | 1517 | // "OOOO" |
michael@0 | 1518 | tzFormat()->format(UTZFMT_STYLE_LOCALIZED_GMT, tz, date, zoneString); |
michael@0 | 1519 | } |
michael@0 | 1520 | } |
michael@0 | 1521 | else if (patternCharIndex == UDAT_TIMEZONE_ISO_FIELD) { |
michael@0 | 1522 | if (count == 1) { |
michael@0 | 1523 | // "X" |
michael@0 | 1524 | tzFormat()->format(UTZFMT_STYLE_ISO_BASIC_SHORT, tz, date, zoneString); |
michael@0 | 1525 | } else if (count == 2) { |
michael@0 | 1526 | // "XX" |
michael@0 | 1527 | tzFormat()->format(UTZFMT_STYLE_ISO_BASIC_FIXED, tz, date, zoneString); |
michael@0 | 1528 | } else if (count == 3) { |
michael@0 | 1529 | // "XXX" |
michael@0 | 1530 | tzFormat()->format(UTZFMT_STYLE_ISO_EXTENDED_FIXED, tz, date, zoneString); |
michael@0 | 1531 | } else if (count == 4) { |
michael@0 | 1532 | // "XXXX" |
michael@0 | 1533 | tzFormat()->format(UTZFMT_STYLE_ISO_BASIC_FULL, tz, date, zoneString); |
michael@0 | 1534 | } else if (count == 5) { |
michael@0 | 1535 | // "XXXXX" |
michael@0 | 1536 | tzFormat()->format(UTZFMT_STYLE_ISO_EXTENDED_FULL, tz, date, zoneString); |
michael@0 | 1537 | } |
michael@0 | 1538 | } |
michael@0 | 1539 | else if (patternCharIndex == UDAT_TIMEZONE_ISO_LOCAL_FIELD) { |
michael@0 | 1540 | if (count == 1) { |
michael@0 | 1541 | // "x" |
michael@0 | 1542 | tzFormat()->format(UTZFMT_STYLE_ISO_BASIC_LOCAL_SHORT, tz, date, zoneString); |
michael@0 | 1543 | } else if (count == 2) { |
michael@0 | 1544 | // "xx" |
michael@0 | 1545 | tzFormat()->format(UTZFMT_STYLE_ISO_BASIC_LOCAL_FIXED, tz, date, zoneString); |
michael@0 | 1546 | } else if (count == 3) { |
michael@0 | 1547 | // "xxx" |
michael@0 | 1548 | tzFormat()->format(UTZFMT_STYLE_ISO_EXTENDED_LOCAL_FIXED, tz, date, zoneString); |
michael@0 | 1549 | } else if (count == 4) { |
michael@0 | 1550 | // "xxxx" |
michael@0 | 1551 | tzFormat()->format(UTZFMT_STYLE_ISO_BASIC_LOCAL_FULL, tz, date, zoneString); |
michael@0 | 1552 | } else if (count == 5) { |
michael@0 | 1553 | // "xxxxx" |
michael@0 | 1554 | tzFormat()->format(UTZFMT_STYLE_ISO_EXTENDED_LOCAL_FULL, tz, date, zoneString); |
michael@0 | 1555 | } |
michael@0 | 1556 | } |
michael@0 | 1557 | else { |
michael@0 | 1558 | U_ASSERT(FALSE); |
michael@0 | 1559 | } |
michael@0 | 1560 | } |
michael@0 | 1561 | appendTo += zoneString; |
michael@0 | 1562 | } |
michael@0 | 1563 | break; |
michael@0 | 1564 | |
michael@0 | 1565 | case UDAT_QUARTER_FIELD: |
michael@0 | 1566 | if (count >= 4) |
michael@0 | 1567 | _appendSymbol(appendTo, value/3, fSymbols->fQuarters, |
michael@0 | 1568 | fSymbols->fQuartersCount); |
michael@0 | 1569 | else if (count == 3) |
michael@0 | 1570 | _appendSymbol(appendTo, value/3, fSymbols->fShortQuarters, |
michael@0 | 1571 | fSymbols->fShortQuartersCount); |
michael@0 | 1572 | else |
michael@0 | 1573 | zeroPaddingNumber(currentNumberFormat,appendTo, (value/3) + 1, count, maxIntCount); |
michael@0 | 1574 | break; |
michael@0 | 1575 | |
michael@0 | 1576 | case UDAT_STANDALONE_QUARTER_FIELD: |
michael@0 | 1577 | if (count >= 4) |
michael@0 | 1578 | _appendSymbol(appendTo, value/3, fSymbols->fStandaloneQuarters, |
michael@0 | 1579 | fSymbols->fStandaloneQuartersCount); |
michael@0 | 1580 | else if (count == 3) |
michael@0 | 1581 | _appendSymbol(appendTo, value/3, fSymbols->fStandaloneShortQuarters, |
michael@0 | 1582 | fSymbols->fStandaloneShortQuartersCount); |
michael@0 | 1583 | else |
michael@0 | 1584 | zeroPaddingNumber(currentNumberFormat,appendTo, (value/3) + 1, count, maxIntCount); |
michael@0 | 1585 | break; |
michael@0 | 1586 | |
michael@0 | 1587 | |
michael@0 | 1588 | // all of the other pattern symbols can be formatted as simple numbers with |
michael@0 | 1589 | // appropriate zero padding |
michael@0 | 1590 | default: |
michael@0 | 1591 | zeroPaddingNumber(currentNumberFormat,appendTo, value, count, maxIntCount); |
michael@0 | 1592 | break; |
michael@0 | 1593 | } |
michael@0 | 1594 | #if !UCONFIG_NO_BREAK_ITERATION |
michael@0 | 1595 | if (fieldNum == 0) { |
michael@0 | 1596 | // first field, check to see whether we need to titlecase it |
michael@0 | 1597 | UBool titlecase = FALSE; |
michael@0 | 1598 | switch (capitalizationContext) { |
michael@0 | 1599 | case UDISPCTX_CAPITALIZATION_FOR_BEGINNING_OF_SENTENCE: |
michael@0 | 1600 | titlecase = TRUE; |
michael@0 | 1601 | break; |
michael@0 | 1602 | case UDISPCTX_CAPITALIZATION_FOR_UI_LIST_OR_MENU: |
michael@0 | 1603 | titlecase = fSymbols->fCapitalization[capContextUsageType][0]; |
michael@0 | 1604 | break; |
michael@0 | 1605 | case UDISPCTX_CAPITALIZATION_FOR_STANDALONE: |
michael@0 | 1606 | titlecase = fSymbols->fCapitalization[capContextUsageType][1]; |
michael@0 | 1607 | break; |
michael@0 | 1608 | default: |
michael@0 | 1609 | // titlecase = FALSE; |
michael@0 | 1610 | break; |
michael@0 | 1611 | } |
michael@0 | 1612 | if (titlecase) { |
michael@0 | 1613 | UnicodeString firstField(appendTo, beginOffset); |
michael@0 | 1614 | firstField.toTitle(NULL, fLocale, U_TITLECASE_NO_LOWERCASE | U_TITLECASE_NO_BREAK_ADJUSTMENT); |
michael@0 | 1615 | appendTo.replaceBetween(beginOffset, appendTo.length(), firstField); |
michael@0 | 1616 | } |
michael@0 | 1617 | } |
michael@0 | 1618 | #endif |
michael@0 | 1619 | |
michael@0 | 1620 | handler.addAttribute(fgPatternIndexToDateFormatField[patternCharIndex], beginOffset, appendTo.length()); |
michael@0 | 1621 | } |
michael@0 | 1622 | |
michael@0 | 1623 | //---------------------------------------------------------------------- |
michael@0 | 1624 | |
michael@0 | 1625 | NumberFormat * |
michael@0 | 1626 | SimpleDateFormat::getNumberFormatByIndex(UDateFormatField index) const { |
michael@0 | 1627 | if (fNumberFormatters != NULL) { |
michael@0 | 1628 | return fNumberFormatters[index]; |
michael@0 | 1629 | } else { |
michael@0 | 1630 | return fNumberFormat; |
michael@0 | 1631 | } |
michael@0 | 1632 | } |
michael@0 | 1633 | |
michael@0 | 1634 | //---------------------------------------------------------------------- |
michael@0 | 1635 | void |
michael@0 | 1636 | SimpleDateFormat::zeroPaddingNumber(NumberFormat *currentNumberFormat,UnicodeString &appendTo, |
michael@0 | 1637 | int32_t value, int32_t minDigits, int32_t maxDigits) const |
michael@0 | 1638 | { |
michael@0 | 1639 | if (currentNumberFormat!=NULL) { |
michael@0 | 1640 | FieldPosition pos(0); |
michael@0 | 1641 | |
michael@0 | 1642 | currentNumberFormat->setMinimumIntegerDigits(minDigits); |
michael@0 | 1643 | currentNumberFormat->setMaximumIntegerDigits(maxDigits); |
michael@0 | 1644 | currentNumberFormat->format(value, appendTo, pos); // 3rd arg is there to speed up processing |
michael@0 | 1645 | } |
michael@0 | 1646 | } |
michael@0 | 1647 | |
michael@0 | 1648 | //---------------------------------------------------------------------- |
michael@0 | 1649 | |
michael@0 | 1650 | /** |
michael@0 | 1651 | * Return true if the given format character, occuring count |
michael@0 | 1652 | * times, represents a numeric field. |
michael@0 | 1653 | */ |
michael@0 | 1654 | UBool SimpleDateFormat::isNumeric(UChar formatChar, int32_t count) { |
michael@0 | 1655 | return DateFormatSymbols::isNumericPatternChar(formatChar, count); |
michael@0 | 1656 | } |
michael@0 | 1657 | |
michael@0 | 1658 | UBool |
michael@0 | 1659 | SimpleDateFormat::isAtNumericField(const UnicodeString &pattern, int32_t patternOffset) { |
michael@0 | 1660 | if (patternOffset >= pattern.length()) { |
michael@0 | 1661 | // not at any field |
michael@0 | 1662 | return FALSE; |
michael@0 | 1663 | } |
michael@0 | 1664 | UChar ch = pattern.charAt(patternOffset); |
michael@0 | 1665 | UDateFormatField f = DateFormatSymbols::getPatternCharIndex(ch); |
michael@0 | 1666 | if (f == UDAT_FIELD_COUNT) { |
michael@0 | 1667 | // not at any field |
michael@0 | 1668 | return FALSE; |
michael@0 | 1669 | } |
michael@0 | 1670 | int32_t i = patternOffset; |
michael@0 | 1671 | while (pattern.charAt(++i) == ch) {} |
michael@0 | 1672 | return DateFormatSymbols::isNumericField(f, i - patternOffset); |
michael@0 | 1673 | } |
michael@0 | 1674 | |
michael@0 | 1675 | UBool |
michael@0 | 1676 | SimpleDateFormat::isAfterNonNumericField(const UnicodeString &pattern, int32_t patternOffset) { |
michael@0 | 1677 | if (patternOffset <= 0) { |
michael@0 | 1678 | // not after any field |
michael@0 | 1679 | return FALSE; |
michael@0 | 1680 | } |
michael@0 | 1681 | UChar ch = pattern.charAt(--patternOffset); |
michael@0 | 1682 | UDateFormatField f = DateFormatSymbols::getPatternCharIndex(ch); |
michael@0 | 1683 | if (f == UDAT_FIELD_COUNT) { |
michael@0 | 1684 | // not after any field |
michael@0 | 1685 | return FALSE; |
michael@0 | 1686 | } |
michael@0 | 1687 | int32_t i = patternOffset; |
michael@0 | 1688 | while (pattern.charAt(--i) == ch) {} |
michael@0 | 1689 | return !DateFormatSymbols::isNumericField(f, patternOffset - i); |
michael@0 | 1690 | } |
michael@0 | 1691 | |
michael@0 | 1692 | void |
michael@0 | 1693 | SimpleDateFormat::parse(const UnicodeString& text, Calendar& cal, ParsePosition& parsePos) const |
michael@0 | 1694 | { |
michael@0 | 1695 | UErrorCode status = U_ZERO_ERROR; |
michael@0 | 1696 | int32_t pos = parsePos.getIndex(); |
michael@0 | 1697 | int32_t start = pos; |
michael@0 | 1698 | |
michael@0 | 1699 | UBool ambiguousYear[] = { FALSE }; |
michael@0 | 1700 | int32_t saveHebrewMonth = -1; |
michael@0 | 1701 | int32_t count = 0; |
michael@0 | 1702 | |
michael@0 | 1703 | // hack, reset tztype, cast away const |
michael@0 | 1704 | ((SimpleDateFormat*)this)->tztype = UTZFMT_TIME_TYPE_UNKNOWN; |
michael@0 | 1705 | |
michael@0 | 1706 | // For parsing abutting numeric fields. 'abutPat' is the |
michael@0 | 1707 | // offset into 'pattern' of the first of 2 or more abutting |
michael@0 | 1708 | // numeric fields. 'abutStart' is the offset into 'text' |
michael@0 | 1709 | // where parsing the fields begins. 'abutPass' starts off as 0 |
michael@0 | 1710 | // and increments each time we try to parse the fields. |
michael@0 | 1711 | int32_t abutPat = -1; // If >=0, we are in a run of abutting numeric fields |
michael@0 | 1712 | int32_t abutStart = 0; |
michael@0 | 1713 | int32_t abutPass = 0; |
michael@0 | 1714 | UBool inQuote = FALSE; |
michael@0 | 1715 | |
michael@0 | 1716 | MessageFormat * numericLeapMonthFormatter = NULL; |
michael@0 | 1717 | |
michael@0 | 1718 | Calendar* calClone = NULL; |
michael@0 | 1719 | Calendar *workCal = &cal; |
michael@0 | 1720 | if (&cal != fCalendar && uprv_strcmp(cal.getType(), fCalendar->getType()) != 0) { |
michael@0 | 1721 | // Different calendar type |
michael@0 | 1722 | // We use the time/zone from the input calendar, but |
michael@0 | 1723 | // do not use the input calendar for field calculation. |
michael@0 | 1724 | calClone = fCalendar->clone(); |
michael@0 | 1725 | if (calClone != NULL) { |
michael@0 | 1726 | calClone->setTime(cal.getTime(status),status); |
michael@0 | 1727 | if (U_FAILURE(status)) { |
michael@0 | 1728 | goto ExitParse; |
michael@0 | 1729 | } |
michael@0 | 1730 | calClone->setTimeZone(cal.getTimeZone()); |
michael@0 | 1731 | workCal = calClone; |
michael@0 | 1732 | } else { |
michael@0 | 1733 | status = U_MEMORY_ALLOCATION_ERROR; |
michael@0 | 1734 | goto ExitParse; |
michael@0 | 1735 | } |
michael@0 | 1736 | } |
michael@0 | 1737 | |
michael@0 | 1738 | if (fSymbols->fLeapMonthPatterns != NULL && fSymbols->fLeapMonthPatternsCount >= DateFormatSymbols::kMonthPatternsCount) { |
michael@0 | 1739 | numericLeapMonthFormatter = new MessageFormat(fSymbols->fLeapMonthPatterns[DateFormatSymbols::kLeapMonthPatternNumeric], fLocale, status); |
michael@0 | 1740 | if (numericLeapMonthFormatter == NULL) { |
michael@0 | 1741 | status = U_MEMORY_ALLOCATION_ERROR; |
michael@0 | 1742 | goto ExitParse; |
michael@0 | 1743 | } else if (U_FAILURE(status)) { |
michael@0 | 1744 | goto ExitParse; // this will delete numericLeapMonthFormatter |
michael@0 | 1745 | } |
michael@0 | 1746 | } |
michael@0 | 1747 | |
michael@0 | 1748 | for (int32_t i=0; i<fPattern.length(); ++i) { |
michael@0 | 1749 | UChar ch = fPattern.charAt(i); |
michael@0 | 1750 | |
michael@0 | 1751 | // Handle alphabetic field characters. |
michael@0 | 1752 | if (!inQuote && ((ch >= 0x41 && ch <= 0x5A) || (ch >= 0x61 && ch <= 0x7A))) { // [A-Za-z] |
michael@0 | 1753 | int32_t fieldPat = i; |
michael@0 | 1754 | |
michael@0 | 1755 | // Count the length of this field specifier |
michael@0 | 1756 | count = 1; |
michael@0 | 1757 | while ((i+1)<fPattern.length() && |
michael@0 | 1758 | fPattern.charAt(i+1) == ch) { |
michael@0 | 1759 | ++count; |
michael@0 | 1760 | ++i; |
michael@0 | 1761 | } |
michael@0 | 1762 | |
michael@0 | 1763 | if (isNumeric(ch, count)) { |
michael@0 | 1764 | if (abutPat < 0) { |
michael@0 | 1765 | // Determine if there is an abutting numeric field. |
michael@0 | 1766 | // Record the start of a set of abutting numeric fields. |
michael@0 | 1767 | if (isAtNumericField(fPattern, i + 1)) { |
michael@0 | 1768 | abutPat = fieldPat; |
michael@0 | 1769 | abutStart = pos; |
michael@0 | 1770 | abutPass = 0; |
michael@0 | 1771 | } |
michael@0 | 1772 | } |
michael@0 | 1773 | } else { |
michael@0 | 1774 | abutPat = -1; // End of any abutting fields |
michael@0 | 1775 | } |
michael@0 | 1776 | |
michael@0 | 1777 | // Handle fields within a run of abutting numeric fields. Take |
michael@0 | 1778 | // the pattern "HHmmss" as an example. We will try to parse |
michael@0 | 1779 | // 2/2/2 characters of the input text, then if that fails, |
michael@0 | 1780 | // 1/2/2. We only adjust the width of the leftmost field; the |
michael@0 | 1781 | // others remain fixed. This allows "123456" => 12:34:56, but |
michael@0 | 1782 | // "12345" => 1:23:45. Likewise, for the pattern "yyyyMMdd" we |
michael@0 | 1783 | // try 4/2/2, 3/2/2, 2/2/2, and finally 1/2/2. |
michael@0 | 1784 | if (abutPat >= 0) { |
michael@0 | 1785 | // If we are at the start of a run of abutting fields, then |
michael@0 | 1786 | // shorten this field in each pass. If we can't shorten |
michael@0 | 1787 | // this field any more, then the parse of this set of |
michael@0 | 1788 | // abutting numeric fields has failed. |
michael@0 | 1789 | if (fieldPat == abutPat) { |
michael@0 | 1790 | count -= abutPass++; |
michael@0 | 1791 | if (count == 0) { |
michael@0 | 1792 | status = U_PARSE_ERROR; |
michael@0 | 1793 | goto ExitParse; |
michael@0 | 1794 | } |
michael@0 | 1795 | } |
michael@0 | 1796 | |
michael@0 | 1797 | pos = subParse(text, pos, ch, count, |
michael@0 | 1798 | TRUE, FALSE, ambiguousYear, saveHebrewMonth, *workCal, i, numericLeapMonthFormatter); |
michael@0 | 1799 | |
michael@0 | 1800 | // If the parse fails anywhere in the run, back up to the |
michael@0 | 1801 | // start of the run and retry. |
michael@0 | 1802 | if (pos < 0) { |
michael@0 | 1803 | i = abutPat - 1; |
michael@0 | 1804 | pos = abutStart; |
michael@0 | 1805 | continue; |
michael@0 | 1806 | } |
michael@0 | 1807 | } |
michael@0 | 1808 | |
michael@0 | 1809 | // Handle non-numeric fields and non-abutting numeric |
michael@0 | 1810 | // fields. |
michael@0 | 1811 | else if (ch != 0x6C) { // pattern char 'l' (SMALL LETTER L) just gets ignored |
michael@0 | 1812 | int32_t s = subParse(text, pos, ch, count, |
michael@0 | 1813 | FALSE, TRUE, ambiguousYear, saveHebrewMonth, *workCal, i, numericLeapMonthFormatter); |
michael@0 | 1814 | |
michael@0 | 1815 | if (s == -pos-1) { |
michael@0 | 1816 | // era not present, in special cases allow this to continue |
michael@0 | 1817 | // from the position where the era was expected |
michael@0 | 1818 | s = pos; |
michael@0 | 1819 | |
michael@0 | 1820 | if (i+1 < fPattern.length()) { |
michael@0 | 1821 | // move to next pattern character |
michael@0 | 1822 | UChar ch = fPattern.charAt(i+1); |
michael@0 | 1823 | |
michael@0 | 1824 | // check for whitespace |
michael@0 | 1825 | if (PatternProps::isWhiteSpace(ch)) { |
michael@0 | 1826 | i++; |
michael@0 | 1827 | // Advance over run in pattern |
michael@0 | 1828 | while ((i+1)<fPattern.length() && |
michael@0 | 1829 | PatternProps::isWhiteSpace(fPattern.charAt(i+1))) { |
michael@0 | 1830 | ++i; |
michael@0 | 1831 | } |
michael@0 | 1832 | } |
michael@0 | 1833 | } |
michael@0 | 1834 | } |
michael@0 | 1835 | else if (s <= 0) { |
michael@0 | 1836 | status = U_PARSE_ERROR; |
michael@0 | 1837 | goto ExitParse; |
michael@0 | 1838 | } |
michael@0 | 1839 | pos = s; |
michael@0 | 1840 | } |
michael@0 | 1841 | } |
michael@0 | 1842 | |
michael@0 | 1843 | // Handle literal pattern characters. These are any |
michael@0 | 1844 | // quoted characters and non-alphabetic unquoted |
michael@0 | 1845 | // characters. |
michael@0 | 1846 | else { |
michael@0 | 1847 | |
michael@0 | 1848 | abutPat = -1; // End of any abutting fields |
michael@0 | 1849 | |
michael@0 | 1850 | if (! matchLiterals(fPattern, i, text, pos, getBooleanAttribute(UDAT_PARSE_ALLOW_WHITESPACE, status))) { |
michael@0 | 1851 | status = U_PARSE_ERROR; |
michael@0 | 1852 | goto ExitParse; |
michael@0 | 1853 | } |
michael@0 | 1854 | } |
michael@0 | 1855 | } |
michael@0 | 1856 | |
michael@0 | 1857 | // Special hack for trailing "." after non-numeric field. |
michael@0 | 1858 | if (text.charAt(pos) == 0x2e && getBooleanAttribute(UDAT_PARSE_ALLOW_WHITESPACE, status)) { |
michael@0 | 1859 | // only do if the last field is not numeric |
michael@0 | 1860 | if (isAfterNonNumericField(fPattern, fPattern.length())) { |
michael@0 | 1861 | pos++; // skip the extra "." |
michael@0 | 1862 | } |
michael@0 | 1863 | } |
michael@0 | 1864 | |
michael@0 | 1865 | // At this point the fields of Calendar have been set. Calendar |
michael@0 | 1866 | // will fill in default values for missing fields when the time |
michael@0 | 1867 | // is computed. |
michael@0 | 1868 | |
michael@0 | 1869 | parsePos.setIndex(pos); |
michael@0 | 1870 | |
michael@0 | 1871 | // This part is a problem: When we call parsedDate.after, we compute the time. |
michael@0 | 1872 | // Take the date April 3 2004 at 2:30 am. When this is first set up, the year |
michael@0 | 1873 | // will be wrong if we're parsing a 2-digit year pattern. It will be 1904. |
michael@0 | 1874 | // April 3 1904 is a Sunday (unlike 2004) so it is the DST onset day. 2:30 am |
michael@0 | 1875 | // is therefore an "impossible" time, since the time goes from 1:59 to 3:00 am |
michael@0 | 1876 | // on that day. It is therefore parsed out to fields as 3:30 am. Then we |
michael@0 | 1877 | // add 100 years, and get April 3 2004 at 3:30 am. Note that April 3 2004 is |
michael@0 | 1878 | // a Saturday, so it can have a 2:30 am -- and it should. [LIU] |
michael@0 | 1879 | /* |
michael@0 | 1880 | UDate parsedDate = calendar.getTime(); |
michael@0 | 1881 | if( ambiguousYear[0] && !parsedDate.after(fDefaultCenturyStart) ) { |
michael@0 | 1882 | calendar.add(Calendar.YEAR, 100); |
michael@0 | 1883 | parsedDate = calendar.getTime(); |
michael@0 | 1884 | } |
michael@0 | 1885 | */ |
michael@0 | 1886 | // Because of the above condition, save off the fields in case we need to readjust. |
michael@0 | 1887 | // The procedure we use here is not particularly efficient, but there is no other |
michael@0 | 1888 | // way to do this given the API restrictions present in Calendar. We minimize |
michael@0 | 1889 | // inefficiency by only performing this computation when it might apply, that is, |
michael@0 | 1890 | // when the two-digit year is equal to the start year, and thus might fall at the |
michael@0 | 1891 | // front or the back of the default century. This only works because we adjust |
michael@0 | 1892 | // the year correctly to start with in other cases -- see subParse(). |
michael@0 | 1893 | if (ambiguousYear[0] || tztype != UTZFMT_TIME_TYPE_UNKNOWN) // If this is true then the two-digit year == the default start year |
michael@0 | 1894 | { |
michael@0 | 1895 | // We need a copy of the fields, and we need to avoid triggering a call to |
michael@0 | 1896 | // complete(), which will recalculate the fields. Since we can't access |
michael@0 | 1897 | // the fields[] array in Calendar, we clone the entire object. This will |
michael@0 | 1898 | // stop working if Calendar.clone() is ever rewritten to call complete(). |
michael@0 | 1899 | Calendar *copy; |
michael@0 | 1900 | if (ambiguousYear[0]) { |
michael@0 | 1901 | copy = cal.clone(); |
michael@0 | 1902 | // Check for failed cloning. |
michael@0 | 1903 | if (copy == NULL) { |
michael@0 | 1904 | status = U_MEMORY_ALLOCATION_ERROR; |
michael@0 | 1905 | goto ExitParse; |
michael@0 | 1906 | } |
michael@0 | 1907 | UDate parsedDate = copy->getTime(status); |
michael@0 | 1908 | // {sfb} check internalGetDefaultCenturyStart |
michael@0 | 1909 | if (fHaveDefaultCentury && (parsedDate < fDefaultCenturyStart)) { |
michael@0 | 1910 | // We can't use add here because that does a complete() first. |
michael@0 | 1911 | cal.set(UCAL_YEAR, fDefaultCenturyStartYear + 100); |
michael@0 | 1912 | } |
michael@0 | 1913 | delete copy; |
michael@0 | 1914 | } |
michael@0 | 1915 | |
michael@0 | 1916 | if (tztype != UTZFMT_TIME_TYPE_UNKNOWN) { |
michael@0 | 1917 | copy = cal.clone(); |
michael@0 | 1918 | // Check for failed cloning. |
michael@0 | 1919 | if (copy == NULL) { |
michael@0 | 1920 | status = U_MEMORY_ALLOCATION_ERROR; |
michael@0 | 1921 | goto ExitParse; |
michael@0 | 1922 | } |
michael@0 | 1923 | const TimeZone & tz = cal.getTimeZone(); |
michael@0 | 1924 | BasicTimeZone *btz = NULL; |
michael@0 | 1925 | |
michael@0 | 1926 | if (dynamic_cast<const OlsonTimeZone *>(&tz) != NULL |
michael@0 | 1927 | || dynamic_cast<const SimpleTimeZone *>(&tz) != NULL |
michael@0 | 1928 | || dynamic_cast<const RuleBasedTimeZone *>(&tz) != NULL |
michael@0 | 1929 | || dynamic_cast<const VTimeZone *>(&tz) != NULL) { |
michael@0 | 1930 | btz = (BasicTimeZone*)&tz; |
michael@0 | 1931 | } |
michael@0 | 1932 | |
michael@0 | 1933 | // Get local millis |
michael@0 | 1934 | copy->set(UCAL_ZONE_OFFSET, 0); |
michael@0 | 1935 | copy->set(UCAL_DST_OFFSET, 0); |
michael@0 | 1936 | UDate localMillis = copy->getTime(status); |
michael@0 | 1937 | |
michael@0 | 1938 | // Make sure parsed time zone type (Standard or Daylight) |
michael@0 | 1939 | // matches the rule used by the parsed time zone. |
michael@0 | 1940 | int32_t raw, dst; |
michael@0 | 1941 | if (btz != NULL) { |
michael@0 | 1942 | if (tztype == UTZFMT_TIME_TYPE_STANDARD) { |
michael@0 | 1943 | btz->getOffsetFromLocal(localMillis, |
michael@0 | 1944 | BasicTimeZone::kStandard, BasicTimeZone::kStandard, raw, dst, status); |
michael@0 | 1945 | } else { |
michael@0 | 1946 | btz->getOffsetFromLocal(localMillis, |
michael@0 | 1947 | BasicTimeZone::kDaylight, BasicTimeZone::kDaylight, raw, dst, status); |
michael@0 | 1948 | } |
michael@0 | 1949 | } else { |
michael@0 | 1950 | // No good way to resolve ambiguous time at transition, |
michael@0 | 1951 | // but following code work in most case. |
michael@0 | 1952 | tz.getOffset(localMillis, TRUE, raw, dst, status); |
michael@0 | 1953 | } |
michael@0 | 1954 | |
michael@0 | 1955 | // Now, compare the results with parsed type, either standard or daylight saving time |
michael@0 | 1956 | int32_t resolvedSavings = dst; |
michael@0 | 1957 | if (tztype == UTZFMT_TIME_TYPE_STANDARD) { |
michael@0 | 1958 | if (dst != 0) { |
michael@0 | 1959 | // Override DST_OFFSET = 0 in the result calendar |
michael@0 | 1960 | resolvedSavings = 0; |
michael@0 | 1961 | } |
michael@0 | 1962 | } else { // tztype == TZTYPE_DST |
michael@0 | 1963 | if (dst == 0) { |
michael@0 | 1964 | if (btz != NULL) { |
michael@0 | 1965 | UDate time = localMillis + raw; |
michael@0 | 1966 | // We use the nearest daylight saving time rule. |
michael@0 | 1967 | TimeZoneTransition beforeTrs, afterTrs; |
michael@0 | 1968 | UDate beforeT = time, afterT = time; |
michael@0 | 1969 | int32_t beforeSav = 0, afterSav = 0; |
michael@0 | 1970 | UBool beforeTrsAvail, afterTrsAvail; |
michael@0 | 1971 | |
michael@0 | 1972 | // Search for DST rule before or on the time |
michael@0 | 1973 | while (TRUE) { |
michael@0 | 1974 | beforeTrsAvail = btz->getPreviousTransition(beforeT, TRUE, beforeTrs); |
michael@0 | 1975 | if (!beforeTrsAvail) { |
michael@0 | 1976 | break; |
michael@0 | 1977 | } |
michael@0 | 1978 | beforeT = beforeTrs.getTime() - 1; |
michael@0 | 1979 | beforeSav = beforeTrs.getFrom()->getDSTSavings(); |
michael@0 | 1980 | if (beforeSav != 0) { |
michael@0 | 1981 | break; |
michael@0 | 1982 | } |
michael@0 | 1983 | } |
michael@0 | 1984 | |
michael@0 | 1985 | // Search for DST rule after the time |
michael@0 | 1986 | while (TRUE) { |
michael@0 | 1987 | afterTrsAvail = btz->getNextTransition(afterT, FALSE, afterTrs); |
michael@0 | 1988 | if (!afterTrsAvail) { |
michael@0 | 1989 | break; |
michael@0 | 1990 | } |
michael@0 | 1991 | afterT = afterTrs.getTime(); |
michael@0 | 1992 | afterSav = afterTrs.getTo()->getDSTSavings(); |
michael@0 | 1993 | if (afterSav != 0) { |
michael@0 | 1994 | break; |
michael@0 | 1995 | } |
michael@0 | 1996 | } |
michael@0 | 1997 | |
michael@0 | 1998 | if (beforeTrsAvail && afterTrsAvail) { |
michael@0 | 1999 | if (time - beforeT > afterT - time) { |
michael@0 | 2000 | resolvedSavings = afterSav; |
michael@0 | 2001 | } else { |
michael@0 | 2002 | resolvedSavings = beforeSav; |
michael@0 | 2003 | } |
michael@0 | 2004 | } else if (beforeTrsAvail && beforeSav != 0) { |
michael@0 | 2005 | resolvedSavings = beforeSav; |
michael@0 | 2006 | } else if (afterTrsAvail && afterSav != 0) { |
michael@0 | 2007 | resolvedSavings = afterSav; |
michael@0 | 2008 | } else { |
michael@0 | 2009 | resolvedSavings = btz->getDSTSavings(); |
michael@0 | 2010 | } |
michael@0 | 2011 | } else { |
michael@0 | 2012 | resolvedSavings = tz.getDSTSavings(); |
michael@0 | 2013 | } |
michael@0 | 2014 | if (resolvedSavings == 0) { |
michael@0 | 2015 | // final fallback |
michael@0 | 2016 | resolvedSavings = U_MILLIS_PER_HOUR; |
michael@0 | 2017 | } |
michael@0 | 2018 | } |
michael@0 | 2019 | } |
michael@0 | 2020 | cal.set(UCAL_ZONE_OFFSET, raw); |
michael@0 | 2021 | cal.set(UCAL_DST_OFFSET, resolvedSavings); |
michael@0 | 2022 | delete copy; |
michael@0 | 2023 | } |
michael@0 | 2024 | } |
michael@0 | 2025 | ExitParse: |
michael@0 | 2026 | // Set the parsed result if local calendar is used |
michael@0 | 2027 | // instead of the input calendar |
michael@0 | 2028 | if (U_SUCCESS(status) && workCal != &cal) { |
michael@0 | 2029 | cal.setTimeZone(workCal->getTimeZone()); |
michael@0 | 2030 | cal.setTime(workCal->getTime(status), status); |
michael@0 | 2031 | } |
michael@0 | 2032 | |
michael@0 | 2033 | if (numericLeapMonthFormatter != NULL) { |
michael@0 | 2034 | delete numericLeapMonthFormatter; |
michael@0 | 2035 | } |
michael@0 | 2036 | if (calClone != NULL) { |
michael@0 | 2037 | delete calClone; |
michael@0 | 2038 | } |
michael@0 | 2039 | |
michael@0 | 2040 | // If any Calendar calls failed, we pretend that we |
michael@0 | 2041 | // couldn't parse the string, when in reality this isn't quite accurate-- |
michael@0 | 2042 | // we did parse it; the Calendar calls just failed. |
michael@0 | 2043 | if (U_FAILURE(status)) { |
michael@0 | 2044 | parsePos.setErrorIndex(pos); |
michael@0 | 2045 | parsePos.setIndex(start); |
michael@0 | 2046 | } |
michael@0 | 2047 | } |
michael@0 | 2048 | |
michael@0 | 2049 | //---------------------------------------------------------------------- |
michael@0 | 2050 | |
michael@0 | 2051 | static UBool |
michael@0 | 2052 | newBestMatchWithOptionalDot(const UnicodeString &lcaseText, |
michael@0 | 2053 | const UnicodeString &data, |
michael@0 | 2054 | UnicodeString &bestMatchName, |
michael@0 | 2055 | int32_t &bestMatchLength); |
michael@0 | 2056 | |
michael@0 | 2057 | int32_t SimpleDateFormat::matchQuarterString(const UnicodeString& text, |
michael@0 | 2058 | int32_t start, |
michael@0 | 2059 | UCalendarDateFields field, |
michael@0 | 2060 | const UnicodeString* data, |
michael@0 | 2061 | int32_t dataCount, |
michael@0 | 2062 | Calendar& cal) const |
michael@0 | 2063 | { |
michael@0 | 2064 | int32_t i = 0; |
michael@0 | 2065 | int32_t count = dataCount; |
michael@0 | 2066 | |
michael@0 | 2067 | // There may be multiple strings in the data[] array which begin with |
michael@0 | 2068 | // the same prefix (e.g., Cerven and Cervenec (June and July) in Czech). |
michael@0 | 2069 | // We keep track of the longest match, and return that. Note that this |
michael@0 | 2070 | // unfortunately requires us to test all array elements. |
michael@0 | 2071 | int32_t bestMatchLength = 0, bestMatch = -1; |
michael@0 | 2072 | UnicodeString bestMatchName; |
michael@0 | 2073 | |
michael@0 | 2074 | // {sfb} kludge to support case-insensitive comparison |
michael@0 | 2075 | // {markus 2002oct11} do not just use caseCompareBetween because we do not know |
michael@0 | 2076 | // the length of the match after case folding |
michael@0 | 2077 | // {alan 20040607} don't case change the whole string, since the length |
michael@0 | 2078 | // can change |
michael@0 | 2079 | // TODO we need a case-insensitive startsWith function |
michael@0 | 2080 | UnicodeString lcaseText; |
michael@0 | 2081 | text.extract(start, INT32_MAX, lcaseText); |
michael@0 | 2082 | lcaseText.foldCase(); |
michael@0 | 2083 | |
michael@0 | 2084 | for (; i < count; ++i) |
michael@0 | 2085 | { |
michael@0 | 2086 | // Always compare if we have no match yet; otherwise only compare |
michael@0 | 2087 | // against potentially better matches (longer strings). |
michael@0 | 2088 | |
michael@0 | 2089 | if (newBestMatchWithOptionalDot(lcaseText, data[i], bestMatchName, bestMatchLength)) { |
michael@0 | 2090 | bestMatch = i; |
michael@0 | 2091 | } |
michael@0 | 2092 | } |
michael@0 | 2093 | if (bestMatch >= 0) |
michael@0 | 2094 | { |
michael@0 | 2095 | cal.set(field, bestMatch * 3); |
michael@0 | 2096 | |
michael@0 | 2097 | // Once we have a match, we have to determine the length of the |
michael@0 | 2098 | // original source string. This will usually be == the length of |
michael@0 | 2099 | // the case folded string, but it may differ (e.g. sharp s). |
michael@0 | 2100 | |
michael@0 | 2101 | // Most of the time, the length will be the same as the length |
michael@0 | 2102 | // of the string from the locale data. Sometimes it will be |
michael@0 | 2103 | // different, in which case we will have to figure it out by |
michael@0 | 2104 | // adding a character at a time, until we have a match. We do |
michael@0 | 2105 | // this all in one loop, where we try 'len' first (at index |
michael@0 | 2106 | // i==0). |
michael@0 | 2107 | int32_t len = bestMatchName.length(); // 99+% of the time |
michael@0 | 2108 | int32_t n = text.length() - start; |
michael@0 | 2109 | for (i=0; i<=n; ++i) { |
michael@0 | 2110 | int32_t j=i; |
michael@0 | 2111 | if (i == 0) { |
michael@0 | 2112 | j = len; |
michael@0 | 2113 | } else if (i == len) { |
michael@0 | 2114 | continue; // already tried this when i was 0 |
michael@0 | 2115 | } |
michael@0 | 2116 | text.extract(start, j, lcaseText); |
michael@0 | 2117 | lcaseText.foldCase(); |
michael@0 | 2118 | if (bestMatchName == lcaseText) { |
michael@0 | 2119 | return start + j; |
michael@0 | 2120 | } |
michael@0 | 2121 | } |
michael@0 | 2122 | } |
michael@0 | 2123 | |
michael@0 | 2124 | return -start; |
michael@0 | 2125 | } |
michael@0 | 2126 | |
michael@0 | 2127 | //---------------------------------------------------------------------- |
michael@0 | 2128 | UBool SimpleDateFormat::matchLiterals(const UnicodeString &pattern, |
michael@0 | 2129 | int32_t &patternOffset, |
michael@0 | 2130 | const UnicodeString &text, |
michael@0 | 2131 | int32_t &textOffset, |
michael@0 | 2132 | UBool lenient) |
michael@0 | 2133 | { |
michael@0 | 2134 | UBool inQuote = FALSE; |
michael@0 | 2135 | UnicodeString literal; |
michael@0 | 2136 | int32_t i = patternOffset; |
michael@0 | 2137 | |
michael@0 | 2138 | // scan pattern looking for contiguous literal characters |
michael@0 | 2139 | for ( ; i < pattern.length(); i += 1) { |
michael@0 | 2140 | UChar ch = pattern.charAt(i); |
michael@0 | 2141 | |
michael@0 | 2142 | if (!inQuote && ((ch >= 0x41 && ch <= 0x5A) || (ch >= 0x61 && ch <= 0x7A))) { // unquoted [A-Za-z] |
michael@0 | 2143 | break; |
michael@0 | 2144 | } |
michael@0 | 2145 | |
michael@0 | 2146 | if (ch == QUOTE) { |
michael@0 | 2147 | // Match a quote literal ('') inside OR outside of quotes |
michael@0 | 2148 | if ((i + 1) < pattern.length() && pattern.charAt(i + 1) == QUOTE) { |
michael@0 | 2149 | i += 1; |
michael@0 | 2150 | } else { |
michael@0 | 2151 | inQuote = !inQuote; |
michael@0 | 2152 | continue; |
michael@0 | 2153 | } |
michael@0 | 2154 | } |
michael@0 | 2155 | |
michael@0 | 2156 | literal += ch; |
michael@0 | 2157 | } |
michael@0 | 2158 | |
michael@0 | 2159 | // at this point, literal contains the literal text |
michael@0 | 2160 | // and i is the index of the next non-literal pattern character. |
michael@0 | 2161 | int32_t p; |
michael@0 | 2162 | int32_t t = textOffset; |
michael@0 | 2163 | |
michael@0 | 2164 | if (lenient) { |
michael@0 | 2165 | // trim leading, trailing whitespace from |
michael@0 | 2166 | // the literal text |
michael@0 | 2167 | literal.trim(); |
michael@0 | 2168 | |
michael@0 | 2169 | // ignore any leading whitespace in the text |
michael@0 | 2170 | while (t < text.length() && u_isWhitespace(text.charAt(t))) { |
michael@0 | 2171 | t += 1; |
michael@0 | 2172 | } |
michael@0 | 2173 | } |
michael@0 | 2174 | |
michael@0 | 2175 | for (p = 0; p < literal.length() && t < text.length();) { |
michael@0 | 2176 | UBool needWhitespace = FALSE; |
michael@0 | 2177 | |
michael@0 | 2178 | while (p < literal.length() && PatternProps::isWhiteSpace(literal.charAt(p))) { |
michael@0 | 2179 | needWhitespace = TRUE; |
michael@0 | 2180 | p += 1; |
michael@0 | 2181 | } |
michael@0 | 2182 | |
michael@0 | 2183 | if (needWhitespace) { |
michael@0 | 2184 | int32_t tStart = t; |
michael@0 | 2185 | |
michael@0 | 2186 | while (t < text.length()) { |
michael@0 | 2187 | UChar tch = text.charAt(t); |
michael@0 | 2188 | |
michael@0 | 2189 | if (!u_isUWhiteSpace(tch) && !PatternProps::isWhiteSpace(tch)) { |
michael@0 | 2190 | break; |
michael@0 | 2191 | } |
michael@0 | 2192 | |
michael@0 | 2193 | t += 1; |
michael@0 | 2194 | } |
michael@0 | 2195 | |
michael@0 | 2196 | // TODO: should we require internal spaces |
michael@0 | 2197 | // in lenient mode? (There won't be any |
michael@0 | 2198 | // leading or trailing spaces) |
michael@0 | 2199 | if (!lenient && t == tStart) { |
michael@0 | 2200 | // didn't find matching whitespace: |
michael@0 | 2201 | // an error in strict mode |
michael@0 | 2202 | return FALSE; |
michael@0 | 2203 | } |
michael@0 | 2204 | |
michael@0 | 2205 | // In strict mode, this run of whitespace |
michael@0 | 2206 | // may have been at the end. |
michael@0 | 2207 | if (p >= literal.length()) { |
michael@0 | 2208 | break; |
michael@0 | 2209 | } |
michael@0 | 2210 | } |
michael@0 | 2211 | |
michael@0 | 2212 | if (t >= text.length() || literal.charAt(p) != text.charAt(t)) { |
michael@0 | 2213 | // Ran out of text, or found a non-matching character: |
michael@0 | 2214 | // OK in lenient mode, an error in strict mode. |
michael@0 | 2215 | if (lenient) { |
michael@0 | 2216 | if (t == textOffset && text.charAt(t) == 0x2e && |
michael@0 | 2217 | isAfterNonNumericField(pattern, patternOffset)) { |
michael@0 | 2218 | // Lenient mode and the literal input text begins with a "." and |
michael@0 | 2219 | // we are after a non-numeric field: We skip the "." |
michael@0 | 2220 | ++t; |
michael@0 | 2221 | continue; // Do not update p. |
michael@0 | 2222 | } |
michael@0 | 2223 | break; |
michael@0 | 2224 | } |
michael@0 | 2225 | |
michael@0 | 2226 | return FALSE; |
michael@0 | 2227 | } |
michael@0 | 2228 | ++p; |
michael@0 | 2229 | ++t; |
michael@0 | 2230 | } |
michael@0 | 2231 | |
michael@0 | 2232 | // At this point if we're in strict mode we have a complete match. |
michael@0 | 2233 | // If we're in lenient mode we may have a partial match, or no |
michael@0 | 2234 | // match at all. |
michael@0 | 2235 | if (p <= 0) { |
michael@0 | 2236 | // no match. Pretend it matched a run of whitespace |
michael@0 | 2237 | // and ignorables in the text. |
michael@0 | 2238 | const UnicodeSet *ignorables = NULL; |
michael@0 | 2239 | UDateFormatField patternCharIndex = DateFormatSymbols::getPatternCharIndex(pattern.charAt(i)); |
michael@0 | 2240 | if (patternCharIndex != UDAT_FIELD_COUNT) { |
michael@0 | 2241 | ignorables = SimpleDateFormatStaticSets::getIgnorables(patternCharIndex); |
michael@0 | 2242 | } |
michael@0 | 2243 | |
michael@0 | 2244 | for (t = textOffset; t < text.length(); t += 1) { |
michael@0 | 2245 | UChar ch = text.charAt(t); |
michael@0 | 2246 | |
michael@0 | 2247 | if (ignorables == NULL || !ignorables->contains(ch)) { |
michael@0 | 2248 | break; |
michael@0 | 2249 | } |
michael@0 | 2250 | } |
michael@0 | 2251 | } |
michael@0 | 2252 | |
michael@0 | 2253 | // if we get here, we've got a complete match. |
michael@0 | 2254 | patternOffset = i - 1; |
michael@0 | 2255 | textOffset = t; |
michael@0 | 2256 | |
michael@0 | 2257 | return TRUE; |
michael@0 | 2258 | } |
michael@0 | 2259 | |
michael@0 | 2260 | //---------------------------------------------------------------------- |
michael@0 | 2261 | |
michael@0 | 2262 | int32_t SimpleDateFormat::matchString(const UnicodeString& text, |
michael@0 | 2263 | int32_t start, |
michael@0 | 2264 | UCalendarDateFields field, |
michael@0 | 2265 | const UnicodeString* data, |
michael@0 | 2266 | int32_t dataCount, |
michael@0 | 2267 | const UnicodeString* monthPattern, |
michael@0 | 2268 | Calendar& cal) const |
michael@0 | 2269 | { |
michael@0 | 2270 | int32_t i = 0; |
michael@0 | 2271 | int32_t count = dataCount; |
michael@0 | 2272 | |
michael@0 | 2273 | if (field == UCAL_DAY_OF_WEEK) i = 1; |
michael@0 | 2274 | |
michael@0 | 2275 | // There may be multiple strings in the data[] array which begin with |
michael@0 | 2276 | // the same prefix (e.g., Cerven and Cervenec (June and July) in Czech). |
michael@0 | 2277 | // We keep track of the longest match, and return that. Note that this |
michael@0 | 2278 | // unfortunately requires us to test all array elements. |
michael@0 | 2279 | int32_t bestMatchLength = 0, bestMatch = -1; |
michael@0 | 2280 | UnicodeString bestMatchName; |
michael@0 | 2281 | int32_t isLeapMonth = 0; |
michael@0 | 2282 | |
michael@0 | 2283 | // {sfb} kludge to support case-insensitive comparison |
michael@0 | 2284 | // {markus 2002oct11} do not just use caseCompareBetween because we do not know |
michael@0 | 2285 | // the length of the match after case folding |
michael@0 | 2286 | // {alan 20040607} don't case change the whole string, since the length |
michael@0 | 2287 | // can change |
michael@0 | 2288 | // TODO we need a case-insensitive startsWith function |
michael@0 | 2289 | UnicodeString lcaseText; |
michael@0 | 2290 | text.extract(start, INT32_MAX, lcaseText); |
michael@0 | 2291 | lcaseText.foldCase(); |
michael@0 | 2292 | |
michael@0 | 2293 | for (; i < count; ++i) |
michael@0 | 2294 | { |
michael@0 | 2295 | // Always compare if we have no match yet; otherwise only compare |
michael@0 | 2296 | // against potentially better matches (longer strings). |
michael@0 | 2297 | |
michael@0 | 2298 | if (newBestMatchWithOptionalDot(lcaseText, data[i], bestMatchName, bestMatchLength)) { |
michael@0 | 2299 | bestMatch = i; |
michael@0 | 2300 | isLeapMonth = 0; |
michael@0 | 2301 | } |
michael@0 | 2302 | |
michael@0 | 2303 | if (monthPattern != NULL) { |
michael@0 | 2304 | UErrorCode status = U_ZERO_ERROR; |
michael@0 | 2305 | UnicodeString leapMonthName; |
michael@0 | 2306 | Formattable monthName((const UnicodeString&)(data[i])); |
michael@0 | 2307 | MessageFormat::format(*monthPattern, &monthName, 1, leapMonthName, status); |
michael@0 | 2308 | if (U_SUCCESS(status)) { |
michael@0 | 2309 | if (newBestMatchWithOptionalDot(lcaseText, leapMonthName, bestMatchName, bestMatchLength)) { |
michael@0 | 2310 | bestMatch = i; |
michael@0 | 2311 | isLeapMonth = 1; |
michael@0 | 2312 | } |
michael@0 | 2313 | } |
michael@0 | 2314 | } |
michael@0 | 2315 | } |
michael@0 | 2316 | if (bestMatch >= 0) |
michael@0 | 2317 | { |
michael@0 | 2318 | // Adjustment for Hebrew Calendar month Adar II |
michael@0 | 2319 | if (!strcmp(cal.getType(),"hebrew") && field==UCAL_MONTH && bestMatch==13) { |
michael@0 | 2320 | cal.set(field,6); |
michael@0 | 2321 | } |
michael@0 | 2322 | else { |
michael@0 | 2323 | if (field == UCAL_YEAR) { |
michael@0 | 2324 | bestMatch++; // only get here for cyclic year names, which match 1-based years 1-60 |
michael@0 | 2325 | } |
michael@0 | 2326 | cal.set(field, bestMatch); |
michael@0 | 2327 | } |
michael@0 | 2328 | if (monthPattern != NULL) { |
michael@0 | 2329 | cal.set(UCAL_IS_LEAP_MONTH, isLeapMonth); |
michael@0 | 2330 | } |
michael@0 | 2331 | |
michael@0 | 2332 | // Once we have a match, we have to determine the length of the |
michael@0 | 2333 | // original source string. This will usually be == the length of |
michael@0 | 2334 | // the case folded string, but it may differ (e.g. sharp s). |
michael@0 | 2335 | |
michael@0 | 2336 | // Most of the time, the length will be the same as the length |
michael@0 | 2337 | // of the string from the locale data. Sometimes it will be |
michael@0 | 2338 | // different, in which case we will have to figure it out by |
michael@0 | 2339 | // adding a character at a time, until we have a match. We do |
michael@0 | 2340 | // this all in one loop, where we try 'len' first (at index |
michael@0 | 2341 | // i==0). |
michael@0 | 2342 | int32_t len = bestMatchName.length(); // 99+% of the time |
michael@0 | 2343 | int32_t n = text.length() - start; |
michael@0 | 2344 | for (i=0; i<=n; ++i) { |
michael@0 | 2345 | int32_t j=i; |
michael@0 | 2346 | if (i == 0) { |
michael@0 | 2347 | j = len; |
michael@0 | 2348 | } else if (i == len) { |
michael@0 | 2349 | continue; // already tried this when i was 0 |
michael@0 | 2350 | } |
michael@0 | 2351 | text.extract(start, j, lcaseText); |
michael@0 | 2352 | lcaseText.foldCase(); |
michael@0 | 2353 | if (bestMatchName == lcaseText) { |
michael@0 | 2354 | return start + j; |
michael@0 | 2355 | } |
michael@0 | 2356 | } |
michael@0 | 2357 | } |
michael@0 | 2358 | |
michael@0 | 2359 | return -start; |
michael@0 | 2360 | } |
michael@0 | 2361 | |
michael@0 | 2362 | static UBool |
michael@0 | 2363 | newBestMatchWithOptionalDot(const UnicodeString &lcaseText, |
michael@0 | 2364 | const UnicodeString &data, |
michael@0 | 2365 | UnicodeString &bestMatchName, |
michael@0 | 2366 | int32_t &bestMatchLength) { |
michael@0 | 2367 | UnicodeString lcase; |
michael@0 | 2368 | lcase.fastCopyFrom(data).foldCase(); |
michael@0 | 2369 | int32_t length = lcase.length(); |
michael@0 | 2370 | if (length <= bestMatchLength) { |
michael@0 | 2371 | // data cannot provide a better match. |
michael@0 | 2372 | return FALSE; |
michael@0 | 2373 | } |
michael@0 | 2374 | |
michael@0 | 2375 | if (lcaseText.compareBetween(0, length, lcase, 0, length) == 0) { |
michael@0 | 2376 | // normal match |
michael@0 | 2377 | bestMatchName = lcase; |
michael@0 | 2378 | bestMatchLength = length; |
michael@0 | 2379 | return TRUE; |
michael@0 | 2380 | } |
michael@0 | 2381 | if (lcase.charAt(--length) == 0x2e) { |
michael@0 | 2382 | if (lcaseText.compareBetween(0, length, lcase, 0, length) == 0) { |
michael@0 | 2383 | // The input text matches the data except for data's trailing dot. |
michael@0 | 2384 | bestMatchName = lcase; |
michael@0 | 2385 | bestMatchName.truncate(length); |
michael@0 | 2386 | bestMatchLength = length; |
michael@0 | 2387 | return TRUE; |
michael@0 | 2388 | } |
michael@0 | 2389 | } |
michael@0 | 2390 | return FALSE; |
michael@0 | 2391 | } |
michael@0 | 2392 | |
michael@0 | 2393 | //---------------------------------------------------------------------- |
michael@0 | 2394 | |
michael@0 | 2395 | void |
michael@0 | 2396 | SimpleDateFormat::set2DigitYearStart(UDate d, UErrorCode& status) |
michael@0 | 2397 | { |
michael@0 | 2398 | parseAmbiguousDatesAsAfter(d, status); |
michael@0 | 2399 | } |
michael@0 | 2400 | |
michael@0 | 2401 | /** |
michael@0 | 2402 | * Private member function that converts the parsed date strings into |
michael@0 | 2403 | * timeFields. Returns -start (for ParsePosition) if failed. |
michael@0 | 2404 | * @param text the time text to be parsed. |
michael@0 | 2405 | * @param start where to start parsing. |
michael@0 | 2406 | * @param ch the pattern character for the date field text to be parsed. |
michael@0 | 2407 | * @param count the count of a pattern character. |
michael@0 | 2408 | * @return the new start position if matching succeeded; a negative number |
michael@0 | 2409 | * indicating matching failure, otherwise. |
michael@0 | 2410 | */ |
michael@0 | 2411 | int32_t SimpleDateFormat::subParse(const UnicodeString& text, int32_t& start, UChar ch, int32_t count, |
michael@0 | 2412 | UBool obeyCount, UBool allowNegative, UBool ambiguousYear[], int32_t& saveHebrewMonth, Calendar& cal, |
michael@0 | 2413 | int32_t patLoc, MessageFormat * numericLeapMonthFormatter) const |
michael@0 | 2414 | { |
michael@0 | 2415 | Formattable number; |
michael@0 | 2416 | int32_t value = 0; |
michael@0 | 2417 | int32_t i; |
michael@0 | 2418 | int32_t ps = 0; |
michael@0 | 2419 | UErrorCode status = U_ZERO_ERROR; |
michael@0 | 2420 | ParsePosition pos(0); |
michael@0 | 2421 | UDateFormatField patternCharIndex = DateFormatSymbols::getPatternCharIndex(ch); |
michael@0 | 2422 | NumberFormat *currentNumberFormat; |
michael@0 | 2423 | UnicodeString temp; |
michael@0 | 2424 | UBool gotNumber = FALSE; |
michael@0 | 2425 | |
michael@0 | 2426 | #if defined (U_DEBUG_CAL) |
michael@0 | 2427 | //fprintf(stderr, "%s:%d - [%c] st=%d \n", __FILE__, __LINE__, (char) ch, start); |
michael@0 | 2428 | #endif |
michael@0 | 2429 | |
michael@0 | 2430 | if (patternCharIndex == UDAT_FIELD_COUNT) { |
michael@0 | 2431 | return -start; |
michael@0 | 2432 | } |
michael@0 | 2433 | |
michael@0 | 2434 | currentNumberFormat = getNumberFormatByIndex(patternCharIndex); |
michael@0 | 2435 | UCalendarDateFields field = fgPatternIndexToCalendarField[patternCharIndex]; |
michael@0 | 2436 | UnicodeString hebr("hebr", 4, US_INV); |
michael@0 | 2437 | |
michael@0 | 2438 | if (numericLeapMonthFormatter != NULL) { |
michael@0 | 2439 | numericLeapMonthFormatter->setFormats((const Format **)¤tNumberFormat, 1); |
michael@0 | 2440 | } |
michael@0 | 2441 | UBool isChineseCalendar = (uprv_strcmp(cal.getType(),"chinese") == 0 || uprv_strcmp(cal.getType(),"dangi") == 0); |
michael@0 | 2442 | |
michael@0 | 2443 | // If there are any spaces here, skip over them. If we hit the end |
michael@0 | 2444 | // of the string, then fail. |
michael@0 | 2445 | for (;;) { |
michael@0 | 2446 | if (start >= text.length()) { |
michael@0 | 2447 | return -start; |
michael@0 | 2448 | } |
michael@0 | 2449 | UChar32 c = text.char32At(start); |
michael@0 | 2450 | if (!u_isUWhiteSpace(c) /*||*/ && !PatternProps::isWhiteSpace(c)) { |
michael@0 | 2451 | break; |
michael@0 | 2452 | } |
michael@0 | 2453 | start += U16_LENGTH(c); |
michael@0 | 2454 | } |
michael@0 | 2455 | pos.setIndex(start); |
michael@0 | 2456 | |
michael@0 | 2457 | // We handle a few special cases here where we need to parse |
michael@0 | 2458 | // a number value. We handle further, more generic cases below. We need |
michael@0 | 2459 | // to handle some of them here because some fields require extra processing on |
michael@0 | 2460 | // the parsed value. |
michael@0 | 2461 | if (patternCharIndex == UDAT_HOUR_OF_DAY1_FIELD || // k |
michael@0 | 2462 | patternCharIndex == UDAT_HOUR_OF_DAY0_FIELD || // H |
michael@0 | 2463 | patternCharIndex == UDAT_HOUR1_FIELD || // h |
michael@0 | 2464 | patternCharIndex == UDAT_HOUR0_FIELD || // K |
michael@0 | 2465 | (patternCharIndex == UDAT_DOW_LOCAL_FIELD && count <= 2) || // e |
michael@0 | 2466 | (patternCharIndex == UDAT_STANDALONE_DAY_FIELD && count <= 2) || // c |
michael@0 | 2467 | (patternCharIndex == UDAT_MONTH_FIELD && count <= 2) || // M |
michael@0 | 2468 | (patternCharIndex == UDAT_STANDALONE_MONTH_FIELD && count <= 2) || // L |
michael@0 | 2469 | (patternCharIndex == UDAT_QUARTER_FIELD && count <= 2) || // Q |
michael@0 | 2470 | (patternCharIndex == UDAT_STANDALONE_QUARTER_FIELD && count <= 2) || // q |
michael@0 | 2471 | patternCharIndex == UDAT_YEAR_FIELD || // y |
michael@0 | 2472 | patternCharIndex == UDAT_YEAR_WOY_FIELD || // Y |
michael@0 | 2473 | patternCharIndex == UDAT_YEAR_NAME_FIELD || // U (falls back to numeric) |
michael@0 | 2474 | (patternCharIndex == UDAT_ERA_FIELD && isChineseCalendar) || // G |
michael@0 | 2475 | patternCharIndex == UDAT_FRACTIONAL_SECOND_FIELD) // S |
michael@0 | 2476 | { |
michael@0 | 2477 | int32_t parseStart = pos.getIndex(); |
michael@0 | 2478 | // It would be good to unify this with the obeyCount logic below, |
michael@0 | 2479 | // but that's going to be difficult. |
michael@0 | 2480 | const UnicodeString* src; |
michael@0 | 2481 | |
michael@0 | 2482 | UBool parsedNumericLeapMonth = FALSE; |
michael@0 | 2483 | if (numericLeapMonthFormatter != NULL && (patternCharIndex == UDAT_MONTH_FIELD || patternCharIndex == UDAT_STANDALONE_MONTH_FIELD)) { |
michael@0 | 2484 | int32_t argCount; |
michael@0 | 2485 | Formattable * args = numericLeapMonthFormatter->parse(text, pos, argCount); |
michael@0 | 2486 | if (args != NULL && argCount == 1 && pos.getIndex() > parseStart && args[0].isNumeric()) { |
michael@0 | 2487 | parsedNumericLeapMonth = TRUE; |
michael@0 | 2488 | number.setLong(args[0].getLong()); |
michael@0 | 2489 | cal.set(UCAL_IS_LEAP_MONTH, 1); |
michael@0 | 2490 | delete[] args; |
michael@0 | 2491 | } else { |
michael@0 | 2492 | pos.setIndex(parseStart); |
michael@0 | 2493 | cal.set(UCAL_IS_LEAP_MONTH, 0); |
michael@0 | 2494 | } |
michael@0 | 2495 | } |
michael@0 | 2496 | |
michael@0 | 2497 | if (!parsedNumericLeapMonth) { |
michael@0 | 2498 | if (obeyCount) { |
michael@0 | 2499 | if ((start+count) > text.length()) { |
michael@0 | 2500 | return -start; |
michael@0 | 2501 | } |
michael@0 | 2502 | |
michael@0 | 2503 | text.extractBetween(0, start + count, temp); |
michael@0 | 2504 | src = &temp; |
michael@0 | 2505 | } else { |
michael@0 | 2506 | src = &text; |
michael@0 | 2507 | } |
michael@0 | 2508 | |
michael@0 | 2509 | parseInt(*src, number, pos, allowNegative,currentNumberFormat); |
michael@0 | 2510 | } |
michael@0 | 2511 | |
michael@0 | 2512 | int32_t txtLoc = pos.getIndex(); |
michael@0 | 2513 | |
michael@0 | 2514 | if (txtLoc > parseStart) { |
michael@0 | 2515 | value = number.getLong(); |
michael@0 | 2516 | gotNumber = TRUE; |
michael@0 | 2517 | |
michael@0 | 2518 | // suffix processing |
michael@0 | 2519 | if (value < 0 ) { |
michael@0 | 2520 | txtLoc = checkIntSuffix(text, txtLoc, patLoc+1, TRUE); |
michael@0 | 2521 | if (txtLoc != pos.getIndex()) { |
michael@0 | 2522 | value *= -1; |
michael@0 | 2523 | } |
michael@0 | 2524 | } |
michael@0 | 2525 | else { |
michael@0 | 2526 | txtLoc = checkIntSuffix(text, txtLoc, patLoc+1, FALSE); |
michael@0 | 2527 | } |
michael@0 | 2528 | |
michael@0 | 2529 | if (!getBooleanAttribute(UDAT_PARSE_ALLOW_WHITESPACE, status)) { |
michael@0 | 2530 | // Check the range of the value |
michael@0 | 2531 | int32_t bias = gFieldRangeBias[patternCharIndex]; |
michael@0 | 2532 | if (bias >= 0 && (value > cal.getMaximum(field) + bias || value < cal.getMinimum(field) + bias)) { |
michael@0 | 2533 | return -start; |
michael@0 | 2534 | } |
michael@0 | 2535 | } |
michael@0 | 2536 | |
michael@0 | 2537 | pos.setIndex(txtLoc); |
michael@0 | 2538 | } |
michael@0 | 2539 | } |
michael@0 | 2540 | |
michael@0 | 2541 | // Make sure that we got a number if |
michael@0 | 2542 | // we want one, and didn't get one |
michael@0 | 2543 | // if we don't want one. |
michael@0 | 2544 | switch (patternCharIndex) { |
michael@0 | 2545 | case UDAT_HOUR_OF_DAY1_FIELD: |
michael@0 | 2546 | case UDAT_HOUR_OF_DAY0_FIELD: |
michael@0 | 2547 | case UDAT_HOUR1_FIELD: |
michael@0 | 2548 | case UDAT_HOUR0_FIELD: |
michael@0 | 2549 | // special range check for hours: |
michael@0 | 2550 | if (value < 0 || value > 24) { |
michael@0 | 2551 | return -start; |
michael@0 | 2552 | } |
michael@0 | 2553 | |
michael@0 | 2554 | // fall through to gotNumber check |
michael@0 | 2555 | |
michael@0 | 2556 | case UDAT_YEAR_FIELD: |
michael@0 | 2557 | case UDAT_YEAR_WOY_FIELD: |
michael@0 | 2558 | case UDAT_FRACTIONAL_SECOND_FIELD: |
michael@0 | 2559 | // these must be a number |
michael@0 | 2560 | if (! gotNumber) { |
michael@0 | 2561 | return -start; |
michael@0 | 2562 | } |
michael@0 | 2563 | |
michael@0 | 2564 | break; |
michael@0 | 2565 | |
michael@0 | 2566 | default: |
michael@0 | 2567 | // we check the rest of the fields below. |
michael@0 | 2568 | break; |
michael@0 | 2569 | } |
michael@0 | 2570 | |
michael@0 | 2571 | switch (patternCharIndex) { |
michael@0 | 2572 | case UDAT_ERA_FIELD: |
michael@0 | 2573 | if (isChineseCalendar) { |
michael@0 | 2574 | if (!gotNumber) { |
michael@0 | 2575 | return -start; |
michael@0 | 2576 | } |
michael@0 | 2577 | cal.set(UCAL_ERA, value); |
michael@0 | 2578 | return pos.getIndex(); |
michael@0 | 2579 | } |
michael@0 | 2580 | if (count == 5) { |
michael@0 | 2581 | ps = matchString(text, start, UCAL_ERA, fSymbols->fNarrowEras, fSymbols->fNarrowErasCount, NULL, cal); |
michael@0 | 2582 | } else if (count == 4) { |
michael@0 | 2583 | ps = matchString(text, start, UCAL_ERA, fSymbols->fEraNames, fSymbols->fEraNamesCount, NULL, cal); |
michael@0 | 2584 | } else { |
michael@0 | 2585 | ps = matchString(text, start, UCAL_ERA, fSymbols->fEras, fSymbols->fErasCount, NULL, cal); |
michael@0 | 2586 | } |
michael@0 | 2587 | |
michael@0 | 2588 | // check return position, if it equals -start, then matchString error |
michael@0 | 2589 | // special case the return code so we don't necessarily fail out until we |
michael@0 | 2590 | // verify no year information also |
michael@0 | 2591 | if (ps == -start) |
michael@0 | 2592 | ps--; |
michael@0 | 2593 | |
michael@0 | 2594 | return ps; |
michael@0 | 2595 | |
michael@0 | 2596 | case UDAT_YEAR_FIELD: |
michael@0 | 2597 | // If there are 3 or more YEAR pattern characters, this indicates |
michael@0 | 2598 | // that the year value is to be treated literally, without any |
michael@0 | 2599 | // two-digit year adjustments (e.g., from "01" to 2001). Otherwise |
michael@0 | 2600 | // we made adjustments to place the 2-digit year in the proper |
michael@0 | 2601 | // century, for parsed strings from "00" to "99". Any other string |
michael@0 | 2602 | // is treated literally: "2250", "-1", "1", "002". |
michael@0 | 2603 | if (fDateOverride.compare(hebr)==0 && value < 1000) { |
michael@0 | 2604 | value += HEBREW_CAL_CUR_MILLENIUM_START_YEAR; |
michael@0 | 2605 | } else if ((pos.getIndex() - start) == 2 && !isChineseCalendar |
michael@0 | 2606 | && u_isdigit(text.charAt(start)) |
michael@0 | 2607 | && u_isdigit(text.charAt(start+1))) |
michael@0 | 2608 | { |
michael@0 | 2609 | // only adjust year for patterns less than 3. |
michael@0 | 2610 | if(count < 3) { |
michael@0 | 2611 | // Assume for example that the defaultCenturyStart is 6/18/1903. |
michael@0 | 2612 | // This means that two-digit years will be forced into the range |
michael@0 | 2613 | // 6/18/1903 to 6/17/2003. As a result, years 00, 01, and 02 |
michael@0 | 2614 | // correspond to 2000, 2001, and 2002. Years 04, 05, etc. correspond |
michael@0 | 2615 | // to 1904, 1905, etc. If the year is 03, then it is 2003 if the |
michael@0 | 2616 | // other fields specify a date before 6/18, or 1903 if they specify a |
michael@0 | 2617 | // date afterwards. As a result, 03 is an ambiguous year. All other |
michael@0 | 2618 | // two-digit years are unambiguous. |
michael@0 | 2619 | if(fHaveDefaultCentury) { // check if this formatter even has a pivot year |
michael@0 | 2620 | int32_t ambiguousTwoDigitYear = fDefaultCenturyStartYear % 100; |
michael@0 | 2621 | ambiguousYear[0] = (value == ambiguousTwoDigitYear); |
michael@0 | 2622 | value += (fDefaultCenturyStartYear/100)*100 + |
michael@0 | 2623 | (value < ambiguousTwoDigitYear ? 100 : 0); |
michael@0 | 2624 | } |
michael@0 | 2625 | } |
michael@0 | 2626 | } |
michael@0 | 2627 | cal.set(UCAL_YEAR, value); |
michael@0 | 2628 | |
michael@0 | 2629 | // Delayed checking for adjustment of Hebrew month numbers in non-leap years. |
michael@0 | 2630 | if (saveHebrewMonth >= 0) { |
michael@0 | 2631 | HebrewCalendar *hc = (HebrewCalendar*)&cal; |
michael@0 | 2632 | if (!hc->isLeapYear(value) && saveHebrewMonth >= 6) { |
michael@0 | 2633 | cal.set(UCAL_MONTH,saveHebrewMonth); |
michael@0 | 2634 | } else { |
michael@0 | 2635 | cal.set(UCAL_MONTH,saveHebrewMonth-1); |
michael@0 | 2636 | } |
michael@0 | 2637 | saveHebrewMonth = -1; |
michael@0 | 2638 | } |
michael@0 | 2639 | return pos.getIndex(); |
michael@0 | 2640 | |
michael@0 | 2641 | case UDAT_YEAR_WOY_FIELD: |
michael@0 | 2642 | // Comment is the same as for UDAT_Year_FIELDs - look above |
michael@0 | 2643 | if (fDateOverride.compare(hebr)==0 && value < 1000) { |
michael@0 | 2644 | value += HEBREW_CAL_CUR_MILLENIUM_START_YEAR; |
michael@0 | 2645 | } else if ((pos.getIndex() - start) == 2 |
michael@0 | 2646 | && u_isdigit(text.charAt(start)) |
michael@0 | 2647 | && u_isdigit(text.charAt(start+1)) |
michael@0 | 2648 | && fHaveDefaultCentury ) |
michael@0 | 2649 | { |
michael@0 | 2650 | int32_t ambiguousTwoDigitYear = fDefaultCenturyStartYear % 100; |
michael@0 | 2651 | ambiguousYear[0] = (value == ambiguousTwoDigitYear); |
michael@0 | 2652 | value += (fDefaultCenturyStartYear/100)*100 + |
michael@0 | 2653 | (value < ambiguousTwoDigitYear ? 100 : 0); |
michael@0 | 2654 | } |
michael@0 | 2655 | cal.set(UCAL_YEAR_WOY, value); |
michael@0 | 2656 | return pos.getIndex(); |
michael@0 | 2657 | |
michael@0 | 2658 | case UDAT_YEAR_NAME_FIELD: |
michael@0 | 2659 | if (fSymbols->fShortYearNames != NULL) { |
michael@0 | 2660 | int32_t newStart = matchString(text, start, UCAL_YEAR, fSymbols->fShortYearNames, fSymbols->fShortYearNamesCount, NULL, cal); |
michael@0 | 2661 | if (newStart > 0) { |
michael@0 | 2662 | return newStart; |
michael@0 | 2663 | } |
michael@0 | 2664 | } |
michael@0 | 2665 | if (gotNumber && (getBooleanAttribute(UDAT_PARSE_ALLOW_NUMERIC,status) || value > fSymbols->fShortYearNamesCount)) { |
michael@0 | 2666 | cal.set(UCAL_YEAR, value); |
michael@0 | 2667 | return pos.getIndex(); |
michael@0 | 2668 | } |
michael@0 | 2669 | return -start; |
michael@0 | 2670 | |
michael@0 | 2671 | case UDAT_MONTH_FIELD: |
michael@0 | 2672 | case UDAT_STANDALONE_MONTH_FIELD: |
michael@0 | 2673 | if (gotNumber) // i.e., M or MM. |
michael@0 | 2674 | { |
michael@0 | 2675 | // When parsing month numbers from the Hebrew Calendar, we might need to adjust the month depending on whether |
michael@0 | 2676 | // 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 |
michael@0 | 2677 | // the year is parsed. |
michael@0 | 2678 | if (!strcmp(cal.getType(),"hebrew")) { |
michael@0 | 2679 | HebrewCalendar *hc = (HebrewCalendar*)&cal; |
michael@0 | 2680 | if (cal.isSet(UCAL_YEAR)) { |
michael@0 | 2681 | UErrorCode status = U_ZERO_ERROR; |
michael@0 | 2682 | if (!hc->isLeapYear(hc->get(UCAL_YEAR,status)) && value >= 6) { |
michael@0 | 2683 | cal.set(UCAL_MONTH, value); |
michael@0 | 2684 | } else { |
michael@0 | 2685 | cal.set(UCAL_MONTH, value - 1); |
michael@0 | 2686 | } |
michael@0 | 2687 | } else { |
michael@0 | 2688 | saveHebrewMonth = value; |
michael@0 | 2689 | } |
michael@0 | 2690 | } else { |
michael@0 | 2691 | // Don't want to parse the month if it is a string |
michael@0 | 2692 | // while pattern uses numeric style: M/MM, L/LL |
michael@0 | 2693 | // [We computed 'value' above.] |
michael@0 | 2694 | cal.set(UCAL_MONTH, value - 1); |
michael@0 | 2695 | } |
michael@0 | 2696 | return pos.getIndex(); |
michael@0 | 2697 | } else { |
michael@0 | 2698 | // count >= 3 // i.e., MMM/MMMM, LLL/LLLL |
michael@0 | 2699 | // Want to be able to parse both short and long forms. |
michael@0 | 2700 | // Try count == 4 first: |
michael@0 | 2701 | UnicodeString * wideMonthPat = NULL; |
michael@0 | 2702 | UnicodeString * shortMonthPat = NULL; |
michael@0 | 2703 | if (fSymbols->fLeapMonthPatterns != NULL && fSymbols->fLeapMonthPatternsCount >= DateFormatSymbols::kMonthPatternsCount) { |
michael@0 | 2704 | if (patternCharIndex==UDAT_MONTH_FIELD) { |
michael@0 | 2705 | wideMonthPat = &fSymbols->fLeapMonthPatterns[DateFormatSymbols::kLeapMonthPatternFormatWide]; |
michael@0 | 2706 | shortMonthPat = &fSymbols->fLeapMonthPatterns[DateFormatSymbols::kLeapMonthPatternFormatAbbrev]; |
michael@0 | 2707 | } else { |
michael@0 | 2708 | wideMonthPat = &fSymbols->fLeapMonthPatterns[DateFormatSymbols::kLeapMonthPatternStandaloneWide]; |
michael@0 | 2709 | shortMonthPat = &fSymbols->fLeapMonthPatterns[DateFormatSymbols::kLeapMonthPatternStandaloneAbbrev]; |
michael@0 | 2710 | } |
michael@0 | 2711 | } |
michael@0 | 2712 | int32_t newStart = 0; |
michael@0 | 2713 | if (patternCharIndex==UDAT_MONTH_FIELD) { |
michael@0 | 2714 | newStart = matchString(text, start, UCAL_MONTH, fSymbols->fMonths, fSymbols->fMonthsCount, wideMonthPat, cal); // try MMMM |
michael@0 | 2715 | if (newStart > 0) { |
michael@0 | 2716 | return newStart; |
michael@0 | 2717 | } |
michael@0 | 2718 | newStart = matchString(text, start, UCAL_MONTH, fSymbols->fShortMonths, fSymbols->fShortMonthsCount, shortMonthPat, cal); // try MMM |
michael@0 | 2719 | } else { |
michael@0 | 2720 | newStart = matchString(text, start, UCAL_MONTH, fSymbols->fStandaloneMonths, fSymbols->fStandaloneMonthsCount, wideMonthPat, cal); // try LLLL |
michael@0 | 2721 | if (newStart > 0) { |
michael@0 | 2722 | return newStart; |
michael@0 | 2723 | } |
michael@0 | 2724 | newStart = matchString(text, start, UCAL_MONTH, fSymbols->fStandaloneShortMonths, fSymbols->fStandaloneShortMonthsCount, shortMonthPat, cal); // try LLL |
michael@0 | 2725 | } |
michael@0 | 2726 | if (newStart > 0 || !getBooleanAttribute(UDAT_PARSE_ALLOW_NUMERIC, status)) // currently we do not try to parse MMMMM/LLLLL: #8860 |
michael@0 | 2727 | return newStart; |
michael@0 | 2728 | // else we allowing parsing as number, below |
michael@0 | 2729 | } |
michael@0 | 2730 | break; |
michael@0 | 2731 | |
michael@0 | 2732 | case UDAT_HOUR_OF_DAY1_FIELD: |
michael@0 | 2733 | // [We computed 'value' above.] |
michael@0 | 2734 | if (value == cal.getMaximum(UCAL_HOUR_OF_DAY) + 1) |
michael@0 | 2735 | value = 0; |
michael@0 | 2736 | |
michael@0 | 2737 | // fall through to set field |
michael@0 | 2738 | |
michael@0 | 2739 | case UDAT_HOUR_OF_DAY0_FIELD: |
michael@0 | 2740 | cal.set(UCAL_HOUR_OF_DAY, value); |
michael@0 | 2741 | return pos.getIndex(); |
michael@0 | 2742 | |
michael@0 | 2743 | case UDAT_FRACTIONAL_SECOND_FIELD: |
michael@0 | 2744 | // Fractional seconds left-justify |
michael@0 | 2745 | i = pos.getIndex() - start; |
michael@0 | 2746 | if (i < 3) { |
michael@0 | 2747 | while (i < 3) { |
michael@0 | 2748 | value *= 10; |
michael@0 | 2749 | i++; |
michael@0 | 2750 | } |
michael@0 | 2751 | } else { |
michael@0 | 2752 | int32_t a = 1; |
michael@0 | 2753 | while (i > 3) { |
michael@0 | 2754 | a *= 10; |
michael@0 | 2755 | i--; |
michael@0 | 2756 | } |
michael@0 | 2757 | value /= a; |
michael@0 | 2758 | } |
michael@0 | 2759 | cal.set(UCAL_MILLISECOND, value); |
michael@0 | 2760 | return pos.getIndex(); |
michael@0 | 2761 | |
michael@0 | 2762 | case UDAT_DOW_LOCAL_FIELD: |
michael@0 | 2763 | if (gotNumber) // i.e., e or ee |
michael@0 | 2764 | { |
michael@0 | 2765 | // [We computed 'value' above.] |
michael@0 | 2766 | cal.set(UCAL_DOW_LOCAL, value); |
michael@0 | 2767 | return pos.getIndex(); |
michael@0 | 2768 | } |
michael@0 | 2769 | // else for eee-eeeee fall through to handling of EEE-EEEEE |
michael@0 | 2770 | // fall through, do not break here |
michael@0 | 2771 | case UDAT_DAY_OF_WEEK_FIELD: |
michael@0 | 2772 | { |
michael@0 | 2773 | // Want to be able to parse both short and long forms. |
michael@0 | 2774 | // Try count == 4 (EEEE) wide first: |
michael@0 | 2775 | int32_t newStart = 0; |
michael@0 | 2776 | if ((newStart = matchString(text, start, UCAL_DAY_OF_WEEK, |
michael@0 | 2777 | fSymbols->fWeekdays, fSymbols->fWeekdaysCount, NULL, cal)) > 0) |
michael@0 | 2778 | return newStart; |
michael@0 | 2779 | // EEEE wide failed, now try EEE abbreviated |
michael@0 | 2780 | else if ((newStart = matchString(text, start, UCAL_DAY_OF_WEEK, |
michael@0 | 2781 | fSymbols->fShortWeekdays, fSymbols->fShortWeekdaysCount, NULL, cal)) > 0) |
michael@0 | 2782 | return newStart; |
michael@0 | 2783 | // EEE abbreviated failed, now try EEEEEE short |
michael@0 | 2784 | else if ((newStart = matchString(text, start, UCAL_DAY_OF_WEEK, |
michael@0 | 2785 | fSymbols->fShorterWeekdays, fSymbols->fShorterWeekdaysCount, NULL, cal)) > 0) |
michael@0 | 2786 | return newStart; |
michael@0 | 2787 | // EEEEEE short failed, now try EEEEE narrow |
michael@0 | 2788 | else if ((newStart = matchString(text, start, UCAL_DAY_OF_WEEK, |
michael@0 | 2789 | fSymbols->fNarrowWeekdays, fSymbols->fNarrowWeekdaysCount, NULL, cal)) > 0) |
michael@0 | 2790 | return newStart; |
michael@0 | 2791 | else if (!getBooleanAttribute(UDAT_PARSE_ALLOW_NUMERIC, status) || patternCharIndex == UDAT_DAY_OF_WEEK_FIELD) |
michael@0 | 2792 | return newStart; |
michael@0 | 2793 | // else we allowing parsing as number, below |
michael@0 | 2794 | } |
michael@0 | 2795 | break; |
michael@0 | 2796 | |
michael@0 | 2797 | case UDAT_STANDALONE_DAY_FIELD: |
michael@0 | 2798 | { |
michael@0 | 2799 | if (gotNumber) // c or cc |
michael@0 | 2800 | { |
michael@0 | 2801 | // [We computed 'value' above.] |
michael@0 | 2802 | cal.set(UCAL_DOW_LOCAL, value); |
michael@0 | 2803 | return pos.getIndex(); |
michael@0 | 2804 | } |
michael@0 | 2805 | // Want to be able to parse both short and long forms. |
michael@0 | 2806 | // Try count == 4 (cccc) first: |
michael@0 | 2807 | int32_t newStart = 0; |
michael@0 | 2808 | if ((newStart = matchString(text, start, UCAL_DAY_OF_WEEK, |
michael@0 | 2809 | fSymbols->fStandaloneWeekdays, fSymbols->fStandaloneWeekdaysCount, NULL, cal)) > 0) |
michael@0 | 2810 | return newStart; |
michael@0 | 2811 | else if ((newStart = matchString(text, start, UCAL_DAY_OF_WEEK, |
michael@0 | 2812 | fSymbols->fStandaloneShortWeekdays, fSymbols->fStandaloneShortWeekdaysCount, NULL, cal)) > 0) |
michael@0 | 2813 | return newStart; |
michael@0 | 2814 | else if ((newStart = matchString(text, start, UCAL_DAY_OF_WEEK, |
michael@0 | 2815 | fSymbols->fStandaloneShorterWeekdays, fSymbols->fStandaloneShorterWeekdaysCount, NULL, cal)) > 0) |
michael@0 | 2816 | return newStart; |
michael@0 | 2817 | else if (!getBooleanAttribute(UDAT_PARSE_ALLOW_NUMERIC, status)) |
michael@0 | 2818 | return newStart; |
michael@0 | 2819 | // else we allowing parsing as number, below |
michael@0 | 2820 | } |
michael@0 | 2821 | break; |
michael@0 | 2822 | |
michael@0 | 2823 | case UDAT_AM_PM_FIELD: |
michael@0 | 2824 | return matchString(text, start, UCAL_AM_PM, fSymbols->fAmPms, fSymbols->fAmPmsCount, NULL, cal); |
michael@0 | 2825 | |
michael@0 | 2826 | case UDAT_HOUR1_FIELD: |
michael@0 | 2827 | // [We computed 'value' above.] |
michael@0 | 2828 | if (value == cal.getLeastMaximum(UCAL_HOUR)+1) |
michael@0 | 2829 | value = 0; |
michael@0 | 2830 | |
michael@0 | 2831 | // fall through to set field |
michael@0 | 2832 | |
michael@0 | 2833 | case UDAT_HOUR0_FIELD: |
michael@0 | 2834 | cal.set(UCAL_HOUR, value); |
michael@0 | 2835 | return pos.getIndex(); |
michael@0 | 2836 | |
michael@0 | 2837 | case UDAT_QUARTER_FIELD: |
michael@0 | 2838 | if (gotNumber) // i.e., Q or QQ. |
michael@0 | 2839 | { |
michael@0 | 2840 | // Don't want to parse the month if it is a string |
michael@0 | 2841 | // while pattern uses numeric style: Q or QQ. |
michael@0 | 2842 | // [We computed 'value' above.] |
michael@0 | 2843 | cal.set(UCAL_MONTH, (value - 1) * 3); |
michael@0 | 2844 | return pos.getIndex(); |
michael@0 | 2845 | } else { |
michael@0 | 2846 | // count >= 3 // i.e., QQQ or QQQQ |
michael@0 | 2847 | // Want to be able to parse both short and long forms. |
michael@0 | 2848 | // Try count == 4 first: |
michael@0 | 2849 | int32_t newStart = 0; |
michael@0 | 2850 | |
michael@0 | 2851 | if ((newStart = matchQuarterString(text, start, UCAL_MONTH, |
michael@0 | 2852 | fSymbols->fQuarters, fSymbols->fQuartersCount, cal)) > 0) |
michael@0 | 2853 | return newStart; |
michael@0 | 2854 | else if ((newStart = matchQuarterString(text, start, UCAL_MONTH, |
michael@0 | 2855 | fSymbols->fShortQuarters, fSymbols->fShortQuartersCount, cal)) > 0) |
michael@0 | 2856 | return newStart; |
michael@0 | 2857 | else if (!getBooleanAttribute(UDAT_PARSE_ALLOW_NUMERIC, status)) |
michael@0 | 2858 | return newStart; |
michael@0 | 2859 | // else we allowing parsing as number, below |
michael@0 | 2860 | } |
michael@0 | 2861 | break; |
michael@0 | 2862 | |
michael@0 | 2863 | case UDAT_STANDALONE_QUARTER_FIELD: |
michael@0 | 2864 | if (gotNumber) // i.e., q or qq. |
michael@0 | 2865 | { |
michael@0 | 2866 | // Don't want to parse the month if it is a string |
michael@0 | 2867 | // while pattern uses numeric style: q or q. |
michael@0 | 2868 | // [We computed 'value' above.] |
michael@0 | 2869 | cal.set(UCAL_MONTH, (value - 1) * 3); |
michael@0 | 2870 | return pos.getIndex(); |
michael@0 | 2871 | } else { |
michael@0 | 2872 | // count >= 3 // i.e., qqq or qqqq |
michael@0 | 2873 | // Want to be able to parse both short and long forms. |
michael@0 | 2874 | // Try count == 4 first: |
michael@0 | 2875 | int32_t newStart = 0; |
michael@0 | 2876 | |
michael@0 | 2877 | if ((newStart = matchQuarterString(text, start, UCAL_MONTH, |
michael@0 | 2878 | fSymbols->fStandaloneQuarters, fSymbols->fStandaloneQuartersCount, cal)) > 0) |
michael@0 | 2879 | return newStart; |
michael@0 | 2880 | else if ((newStart = matchQuarterString(text, start, UCAL_MONTH, |
michael@0 | 2881 | fSymbols->fStandaloneShortQuarters, fSymbols->fStandaloneShortQuartersCount, cal)) > 0) |
michael@0 | 2882 | return newStart; |
michael@0 | 2883 | else if (!getBooleanAttribute(UDAT_PARSE_ALLOW_NUMERIC, status)) |
michael@0 | 2884 | return newStart; |
michael@0 | 2885 | // else we allowing parsing as number, below |
michael@0 | 2886 | } |
michael@0 | 2887 | break; |
michael@0 | 2888 | |
michael@0 | 2889 | case UDAT_TIMEZONE_FIELD: // 'z' |
michael@0 | 2890 | { |
michael@0 | 2891 | UTimeZoneFormatTimeType tzTimeType = UTZFMT_TIME_TYPE_UNKNOWN; |
michael@0 | 2892 | UTimeZoneFormatStyle style = (count < 4) ? UTZFMT_STYLE_SPECIFIC_SHORT : UTZFMT_STYLE_SPECIFIC_LONG; |
michael@0 | 2893 | TimeZone *tz = tzFormat()->parse(style, text, pos, &tzTimeType); |
michael@0 | 2894 | if (tz != NULL) { |
michael@0 | 2895 | ((SimpleDateFormat*)this)->tztype = tzTimeType; |
michael@0 | 2896 | cal.adoptTimeZone(tz); |
michael@0 | 2897 | return pos.getIndex(); |
michael@0 | 2898 | } |
michael@0 | 2899 | } |
michael@0 | 2900 | break; |
michael@0 | 2901 | case UDAT_TIMEZONE_RFC_FIELD: // 'Z' |
michael@0 | 2902 | { |
michael@0 | 2903 | UTimeZoneFormatTimeType tzTimeType = UTZFMT_TIME_TYPE_UNKNOWN; |
michael@0 | 2904 | UTimeZoneFormatStyle style = (count < 4) ? |
michael@0 | 2905 | UTZFMT_STYLE_ISO_BASIC_LOCAL_FULL : ((count == 5) ? UTZFMT_STYLE_ISO_EXTENDED_FULL: UTZFMT_STYLE_LOCALIZED_GMT); |
michael@0 | 2906 | TimeZone *tz = tzFormat()->parse(style, text, pos, &tzTimeType); |
michael@0 | 2907 | if (tz != NULL) { |
michael@0 | 2908 | ((SimpleDateFormat*)this)->tztype = tzTimeType; |
michael@0 | 2909 | cal.adoptTimeZone(tz); |
michael@0 | 2910 | return pos.getIndex(); |
michael@0 | 2911 | } |
michael@0 | 2912 | return -start; |
michael@0 | 2913 | } |
michael@0 | 2914 | case UDAT_TIMEZONE_GENERIC_FIELD: // 'v' |
michael@0 | 2915 | { |
michael@0 | 2916 | UTimeZoneFormatTimeType tzTimeType = UTZFMT_TIME_TYPE_UNKNOWN; |
michael@0 | 2917 | UTimeZoneFormatStyle style = (count < 4) ? UTZFMT_STYLE_GENERIC_SHORT : UTZFMT_STYLE_GENERIC_LONG; |
michael@0 | 2918 | TimeZone *tz = tzFormat()->parse(style, text, pos, &tzTimeType); |
michael@0 | 2919 | if (tz != NULL) { |
michael@0 | 2920 | ((SimpleDateFormat*)this)->tztype = tzTimeType; |
michael@0 | 2921 | cal.adoptTimeZone(tz); |
michael@0 | 2922 | return pos.getIndex(); |
michael@0 | 2923 | } |
michael@0 | 2924 | return -start; |
michael@0 | 2925 | } |
michael@0 | 2926 | case UDAT_TIMEZONE_SPECIAL_FIELD: // 'V' |
michael@0 | 2927 | { |
michael@0 | 2928 | UTimeZoneFormatTimeType tzTimeType = UTZFMT_TIME_TYPE_UNKNOWN; |
michael@0 | 2929 | UTimeZoneFormatStyle style; |
michael@0 | 2930 | switch (count) { |
michael@0 | 2931 | case 1: |
michael@0 | 2932 | style = UTZFMT_STYLE_ZONE_ID_SHORT; |
michael@0 | 2933 | break; |
michael@0 | 2934 | case 2: |
michael@0 | 2935 | style = UTZFMT_STYLE_ZONE_ID; |
michael@0 | 2936 | break; |
michael@0 | 2937 | case 3: |
michael@0 | 2938 | style = UTZFMT_STYLE_EXEMPLAR_LOCATION; |
michael@0 | 2939 | break; |
michael@0 | 2940 | default: |
michael@0 | 2941 | style = UTZFMT_STYLE_GENERIC_LOCATION; |
michael@0 | 2942 | break; |
michael@0 | 2943 | } |
michael@0 | 2944 | TimeZone *tz = tzFormat()->parse(style, text, pos, &tzTimeType); |
michael@0 | 2945 | if (tz != NULL) { |
michael@0 | 2946 | ((SimpleDateFormat*)this)->tztype = tzTimeType; |
michael@0 | 2947 | cal.adoptTimeZone(tz); |
michael@0 | 2948 | return pos.getIndex(); |
michael@0 | 2949 | } |
michael@0 | 2950 | return -start; |
michael@0 | 2951 | } |
michael@0 | 2952 | case UDAT_TIMEZONE_LOCALIZED_GMT_OFFSET_FIELD: // 'O' |
michael@0 | 2953 | { |
michael@0 | 2954 | UTimeZoneFormatTimeType tzTimeType = UTZFMT_TIME_TYPE_UNKNOWN; |
michael@0 | 2955 | UTimeZoneFormatStyle style = (count < 4) ? UTZFMT_STYLE_LOCALIZED_GMT_SHORT : UTZFMT_STYLE_LOCALIZED_GMT; |
michael@0 | 2956 | TimeZone *tz = tzFormat()->parse(style, text, pos, &tzTimeType); |
michael@0 | 2957 | if (tz != NULL) { |
michael@0 | 2958 | ((SimpleDateFormat*)this)->tztype = tzTimeType; |
michael@0 | 2959 | cal.adoptTimeZone(tz); |
michael@0 | 2960 | return pos.getIndex(); |
michael@0 | 2961 | } |
michael@0 | 2962 | return -start; |
michael@0 | 2963 | } |
michael@0 | 2964 | case UDAT_TIMEZONE_ISO_FIELD: // 'X' |
michael@0 | 2965 | { |
michael@0 | 2966 | UTimeZoneFormatTimeType tzTimeType = UTZFMT_TIME_TYPE_UNKNOWN; |
michael@0 | 2967 | UTimeZoneFormatStyle style; |
michael@0 | 2968 | switch (count) { |
michael@0 | 2969 | case 1: |
michael@0 | 2970 | style = UTZFMT_STYLE_ISO_BASIC_SHORT; |
michael@0 | 2971 | break; |
michael@0 | 2972 | case 2: |
michael@0 | 2973 | style = UTZFMT_STYLE_ISO_BASIC_FIXED; |
michael@0 | 2974 | break; |
michael@0 | 2975 | case 3: |
michael@0 | 2976 | style = UTZFMT_STYLE_ISO_EXTENDED_FIXED; |
michael@0 | 2977 | break; |
michael@0 | 2978 | case 4: |
michael@0 | 2979 | style = UTZFMT_STYLE_ISO_BASIC_FULL; |
michael@0 | 2980 | break; |
michael@0 | 2981 | default: |
michael@0 | 2982 | style = UTZFMT_STYLE_ISO_EXTENDED_FULL; |
michael@0 | 2983 | break; |
michael@0 | 2984 | } |
michael@0 | 2985 | TimeZone *tz = tzFormat()->parse(style, text, pos, &tzTimeType); |
michael@0 | 2986 | if (tz != NULL) { |
michael@0 | 2987 | ((SimpleDateFormat*)this)->tztype = tzTimeType; |
michael@0 | 2988 | cal.adoptTimeZone(tz); |
michael@0 | 2989 | return pos.getIndex(); |
michael@0 | 2990 | } |
michael@0 | 2991 | return -start; |
michael@0 | 2992 | } |
michael@0 | 2993 | case UDAT_TIMEZONE_ISO_LOCAL_FIELD: // 'x' |
michael@0 | 2994 | { |
michael@0 | 2995 | UTimeZoneFormatTimeType tzTimeType = UTZFMT_TIME_TYPE_UNKNOWN; |
michael@0 | 2996 | UTimeZoneFormatStyle style; |
michael@0 | 2997 | switch (count) { |
michael@0 | 2998 | case 1: |
michael@0 | 2999 | style = UTZFMT_STYLE_ISO_BASIC_LOCAL_SHORT; |
michael@0 | 3000 | break; |
michael@0 | 3001 | case 2: |
michael@0 | 3002 | style = UTZFMT_STYLE_ISO_BASIC_LOCAL_FIXED; |
michael@0 | 3003 | break; |
michael@0 | 3004 | case 3: |
michael@0 | 3005 | style = UTZFMT_STYLE_ISO_EXTENDED_LOCAL_FIXED; |
michael@0 | 3006 | break; |
michael@0 | 3007 | case 4: |
michael@0 | 3008 | style = UTZFMT_STYLE_ISO_BASIC_LOCAL_FULL; |
michael@0 | 3009 | break; |
michael@0 | 3010 | default: |
michael@0 | 3011 | style = UTZFMT_STYLE_ISO_EXTENDED_LOCAL_FULL; |
michael@0 | 3012 | break; |
michael@0 | 3013 | } |
michael@0 | 3014 | TimeZone *tz = tzFormat()->parse(style, text, pos, &tzTimeType); |
michael@0 | 3015 | if (tz != NULL) { |
michael@0 | 3016 | ((SimpleDateFormat*)this)->tztype = tzTimeType; |
michael@0 | 3017 | cal.adoptTimeZone(tz); |
michael@0 | 3018 | return pos.getIndex(); |
michael@0 | 3019 | } |
michael@0 | 3020 | return -start; |
michael@0 | 3021 | } |
michael@0 | 3022 | |
michael@0 | 3023 | default: |
michael@0 | 3024 | // Handle "generic" fields |
michael@0 | 3025 | // this is now handled below, outside the switch block |
michael@0 | 3026 | break; |
michael@0 | 3027 | } |
michael@0 | 3028 | // Handle "generic" fields: |
michael@0 | 3029 | // switch default case now handled here (outside switch block) to allow |
michael@0 | 3030 | // parsing of some string fields as digits for lenient case |
michael@0 | 3031 | |
michael@0 | 3032 | int32_t parseStart = pos.getIndex(); |
michael@0 | 3033 | const UnicodeString* src; |
michael@0 | 3034 | if (obeyCount) { |
michael@0 | 3035 | if ((start+count) > text.length()) { |
michael@0 | 3036 | return -start; |
michael@0 | 3037 | } |
michael@0 | 3038 | text.extractBetween(0, start + count, temp); |
michael@0 | 3039 | src = &temp; |
michael@0 | 3040 | } else { |
michael@0 | 3041 | src = &text; |
michael@0 | 3042 | } |
michael@0 | 3043 | parseInt(*src, number, pos, allowNegative,currentNumberFormat); |
michael@0 | 3044 | if (pos.getIndex() != parseStart) { |
michael@0 | 3045 | int32_t value = number.getLong(); |
michael@0 | 3046 | |
michael@0 | 3047 | // Don't need suffix processing here (as in number processing at the beginning of the function); |
michael@0 | 3048 | // the new fields being handled as numeric values (month, weekdays, quarters) should not have suffixes. |
michael@0 | 3049 | |
michael@0 | 3050 | if (!getBooleanAttribute(UDAT_PARSE_ALLOW_NUMERIC, status)) { |
michael@0 | 3051 | // Check the range of the value |
michael@0 | 3052 | int32_t bias = gFieldRangeBias[patternCharIndex]; |
michael@0 | 3053 | if (bias >= 0 && (value > cal.getMaximum(field) + bias || value < cal.getMinimum(field) + bias)) { |
michael@0 | 3054 | return -start; |
michael@0 | 3055 | } |
michael@0 | 3056 | } |
michael@0 | 3057 | |
michael@0 | 3058 | // For the following, need to repeat some of the "if (gotNumber)" code above: |
michael@0 | 3059 | // UDAT_[STANDALONE_]MONTH_FIELD, UDAT_DOW_LOCAL_FIELD, UDAT_STANDALONE_DAY_FIELD, |
michael@0 | 3060 | // UDAT_[STANDALONE_]QUARTER_FIELD |
michael@0 | 3061 | switch (patternCharIndex) { |
michael@0 | 3062 | case UDAT_MONTH_FIELD: |
michael@0 | 3063 | // See notes under UDAT_MONTH_FIELD case above |
michael@0 | 3064 | if (!strcmp(cal.getType(),"hebrew")) { |
michael@0 | 3065 | HebrewCalendar *hc = (HebrewCalendar*)&cal; |
michael@0 | 3066 | if (cal.isSet(UCAL_YEAR)) { |
michael@0 | 3067 | UErrorCode status = U_ZERO_ERROR; |
michael@0 | 3068 | if (!hc->isLeapYear(hc->get(UCAL_YEAR,status)) && value >= 6) { |
michael@0 | 3069 | cal.set(UCAL_MONTH, value); |
michael@0 | 3070 | } else { |
michael@0 | 3071 | cal.set(UCAL_MONTH, value - 1); |
michael@0 | 3072 | } |
michael@0 | 3073 | } else { |
michael@0 | 3074 | saveHebrewMonth = value; |
michael@0 | 3075 | } |
michael@0 | 3076 | } else { |
michael@0 | 3077 | cal.set(UCAL_MONTH, value - 1); |
michael@0 | 3078 | } |
michael@0 | 3079 | break; |
michael@0 | 3080 | case UDAT_STANDALONE_MONTH_FIELD: |
michael@0 | 3081 | cal.set(UCAL_MONTH, value - 1); |
michael@0 | 3082 | break; |
michael@0 | 3083 | case UDAT_DOW_LOCAL_FIELD: |
michael@0 | 3084 | case UDAT_STANDALONE_DAY_FIELD: |
michael@0 | 3085 | cal.set(UCAL_DOW_LOCAL, value); |
michael@0 | 3086 | break; |
michael@0 | 3087 | case UDAT_QUARTER_FIELD: |
michael@0 | 3088 | case UDAT_STANDALONE_QUARTER_FIELD: |
michael@0 | 3089 | cal.set(UCAL_MONTH, (value - 1) * 3); |
michael@0 | 3090 | break; |
michael@0 | 3091 | default: |
michael@0 | 3092 | cal.set(field, value); |
michael@0 | 3093 | break; |
michael@0 | 3094 | } |
michael@0 | 3095 | return pos.getIndex(); |
michael@0 | 3096 | } |
michael@0 | 3097 | return -start; |
michael@0 | 3098 | } |
michael@0 | 3099 | |
michael@0 | 3100 | /** |
michael@0 | 3101 | * Parse an integer using fNumberFormat. This method is semantically |
michael@0 | 3102 | * const, but actually may modify fNumberFormat. |
michael@0 | 3103 | */ |
michael@0 | 3104 | void SimpleDateFormat::parseInt(const UnicodeString& text, |
michael@0 | 3105 | Formattable& number, |
michael@0 | 3106 | ParsePosition& pos, |
michael@0 | 3107 | UBool allowNegative, |
michael@0 | 3108 | NumberFormat *fmt) const { |
michael@0 | 3109 | parseInt(text, number, -1, pos, allowNegative,fmt); |
michael@0 | 3110 | } |
michael@0 | 3111 | |
michael@0 | 3112 | /** |
michael@0 | 3113 | * Parse an integer using fNumberFormat up to maxDigits. |
michael@0 | 3114 | */ |
michael@0 | 3115 | void SimpleDateFormat::parseInt(const UnicodeString& text, |
michael@0 | 3116 | Formattable& number, |
michael@0 | 3117 | int32_t maxDigits, |
michael@0 | 3118 | ParsePosition& pos, |
michael@0 | 3119 | UBool allowNegative, |
michael@0 | 3120 | NumberFormat *fmt) const { |
michael@0 | 3121 | UnicodeString oldPrefix; |
michael@0 | 3122 | DecimalFormat* df = NULL; |
michael@0 | 3123 | if (!allowNegative && (df = dynamic_cast<DecimalFormat*>(fmt)) != NULL) { |
michael@0 | 3124 | df->getNegativePrefix(oldPrefix); |
michael@0 | 3125 | df->setNegativePrefix(UnicodeString(TRUE, SUPPRESS_NEGATIVE_PREFIX, -1)); |
michael@0 | 3126 | } |
michael@0 | 3127 | int32_t oldPos = pos.getIndex(); |
michael@0 | 3128 | fmt->parse(text, number, pos); |
michael@0 | 3129 | if (df != NULL) { |
michael@0 | 3130 | df->setNegativePrefix(oldPrefix); |
michael@0 | 3131 | } |
michael@0 | 3132 | |
michael@0 | 3133 | if (maxDigits > 0) { |
michael@0 | 3134 | // adjust the result to fit into |
michael@0 | 3135 | // the maxDigits and move the position back |
michael@0 | 3136 | int32_t nDigits = pos.getIndex() - oldPos; |
michael@0 | 3137 | if (nDigits > maxDigits) { |
michael@0 | 3138 | int32_t val = number.getLong(); |
michael@0 | 3139 | nDigits -= maxDigits; |
michael@0 | 3140 | while (nDigits > 0) { |
michael@0 | 3141 | val /= 10; |
michael@0 | 3142 | nDigits--; |
michael@0 | 3143 | } |
michael@0 | 3144 | pos.setIndex(oldPos + maxDigits); |
michael@0 | 3145 | number.setLong(val); |
michael@0 | 3146 | } |
michael@0 | 3147 | } |
michael@0 | 3148 | } |
michael@0 | 3149 | |
michael@0 | 3150 | //---------------------------------------------------------------------- |
michael@0 | 3151 | |
michael@0 | 3152 | void SimpleDateFormat::translatePattern(const UnicodeString& originalPattern, |
michael@0 | 3153 | UnicodeString& translatedPattern, |
michael@0 | 3154 | const UnicodeString& from, |
michael@0 | 3155 | const UnicodeString& to, |
michael@0 | 3156 | UErrorCode& status) |
michael@0 | 3157 | { |
michael@0 | 3158 | // run through the pattern and convert any pattern symbols from the version |
michael@0 | 3159 | // in "from" to the corresponding character ion "to". This code takes |
michael@0 | 3160 | // quoted strings into account (it doesn't try to translate them), and it signals |
michael@0 | 3161 | // an error if a particular "pattern character" doesn't appear in "from". |
michael@0 | 3162 | // Depending on the values of "from" and "to" this can convert from generic |
michael@0 | 3163 | // to localized patterns or localized to generic. |
michael@0 | 3164 | if (U_FAILURE(status)) |
michael@0 | 3165 | return; |
michael@0 | 3166 | |
michael@0 | 3167 | translatedPattern.remove(); |
michael@0 | 3168 | UBool inQuote = FALSE; |
michael@0 | 3169 | for (int32_t i = 0; i < originalPattern.length(); ++i) { |
michael@0 | 3170 | UChar c = originalPattern[i]; |
michael@0 | 3171 | if (inQuote) { |
michael@0 | 3172 | if (c == QUOTE) |
michael@0 | 3173 | inQuote = FALSE; |
michael@0 | 3174 | } |
michael@0 | 3175 | else { |
michael@0 | 3176 | if (c == QUOTE) |
michael@0 | 3177 | inQuote = TRUE; |
michael@0 | 3178 | else if ((c >= 0x0061 /*'a'*/ && c <= 0x007A) /*'z'*/ |
michael@0 | 3179 | || (c >= 0x0041 /*'A'*/ && c <= 0x005A /*'Z'*/)) { |
michael@0 | 3180 | int32_t ci = from.indexOf(c); |
michael@0 | 3181 | if (ci == -1) { |
michael@0 | 3182 | status = U_INVALID_FORMAT_ERROR; |
michael@0 | 3183 | return; |
michael@0 | 3184 | } |
michael@0 | 3185 | c = to[ci]; |
michael@0 | 3186 | } |
michael@0 | 3187 | } |
michael@0 | 3188 | translatedPattern += c; |
michael@0 | 3189 | } |
michael@0 | 3190 | if (inQuote) { |
michael@0 | 3191 | status = U_INVALID_FORMAT_ERROR; |
michael@0 | 3192 | return; |
michael@0 | 3193 | } |
michael@0 | 3194 | } |
michael@0 | 3195 | |
michael@0 | 3196 | //---------------------------------------------------------------------- |
michael@0 | 3197 | |
michael@0 | 3198 | UnicodeString& |
michael@0 | 3199 | SimpleDateFormat::toPattern(UnicodeString& result) const |
michael@0 | 3200 | { |
michael@0 | 3201 | result = fPattern; |
michael@0 | 3202 | return result; |
michael@0 | 3203 | } |
michael@0 | 3204 | |
michael@0 | 3205 | //---------------------------------------------------------------------- |
michael@0 | 3206 | |
michael@0 | 3207 | UnicodeString& |
michael@0 | 3208 | SimpleDateFormat::toLocalizedPattern(UnicodeString& result, |
michael@0 | 3209 | UErrorCode& status) const |
michael@0 | 3210 | { |
michael@0 | 3211 | translatePattern(fPattern, result, |
michael@0 | 3212 | UnicodeString(DateFormatSymbols::getPatternUChars()), |
michael@0 | 3213 | fSymbols->fLocalPatternChars, status); |
michael@0 | 3214 | return result; |
michael@0 | 3215 | } |
michael@0 | 3216 | |
michael@0 | 3217 | //---------------------------------------------------------------------- |
michael@0 | 3218 | |
michael@0 | 3219 | void |
michael@0 | 3220 | SimpleDateFormat::applyPattern(const UnicodeString& pattern) |
michael@0 | 3221 | { |
michael@0 | 3222 | fPattern = pattern; |
michael@0 | 3223 | } |
michael@0 | 3224 | |
michael@0 | 3225 | //---------------------------------------------------------------------- |
michael@0 | 3226 | |
michael@0 | 3227 | void |
michael@0 | 3228 | SimpleDateFormat::applyLocalizedPattern(const UnicodeString& pattern, |
michael@0 | 3229 | UErrorCode &status) |
michael@0 | 3230 | { |
michael@0 | 3231 | translatePattern(pattern, fPattern, |
michael@0 | 3232 | fSymbols->fLocalPatternChars, |
michael@0 | 3233 | UnicodeString(DateFormatSymbols::getPatternUChars()), status); |
michael@0 | 3234 | } |
michael@0 | 3235 | |
michael@0 | 3236 | //---------------------------------------------------------------------- |
michael@0 | 3237 | |
michael@0 | 3238 | const DateFormatSymbols* |
michael@0 | 3239 | SimpleDateFormat::getDateFormatSymbols() const |
michael@0 | 3240 | { |
michael@0 | 3241 | return fSymbols; |
michael@0 | 3242 | } |
michael@0 | 3243 | |
michael@0 | 3244 | //---------------------------------------------------------------------- |
michael@0 | 3245 | |
michael@0 | 3246 | void |
michael@0 | 3247 | SimpleDateFormat::adoptDateFormatSymbols(DateFormatSymbols* newFormatSymbols) |
michael@0 | 3248 | { |
michael@0 | 3249 | delete fSymbols; |
michael@0 | 3250 | fSymbols = newFormatSymbols; |
michael@0 | 3251 | } |
michael@0 | 3252 | |
michael@0 | 3253 | //---------------------------------------------------------------------- |
michael@0 | 3254 | void |
michael@0 | 3255 | SimpleDateFormat::setDateFormatSymbols(const DateFormatSymbols& newFormatSymbols) |
michael@0 | 3256 | { |
michael@0 | 3257 | delete fSymbols; |
michael@0 | 3258 | fSymbols = new DateFormatSymbols(newFormatSymbols); |
michael@0 | 3259 | } |
michael@0 | 3260 | |
michael@0 | 3261 | //---------------------------------------------------------------------- |
michael@0 | 3262 | const TimeZoneFormat* |
michael@0 | 3263 | SimpleDateFormat::getTimeZoneFormat(void) const { |
michael@0 | 3264 | return (const TimeZoneFormat*)tzFormat(); |
michael@0 | 3265 | } |
michael@0 | 3266 | |
michael@0 | 3267 | //---------------------------------------------------------------------- |
michael@0 | 3268 | void |
michael@0 | 3269 | SimpleDateFormat::adoptTimeZoneFormat(TimeZoneFormat* timeZoneFormatToAdopt) |
michael@0 | 3270 | { |
michael@0 | 3271 | delete fTimeZoneFormat; |
michael@0 | 3272 | fTimeZoneFormat = timeZoneFormatToAdopt; |
michael@0 | 3273 | } |
michael@0 | 3274 | |
michael@0 | 3275 | //---------------------------------------------------------------------- |
michael@0 | 3276 | void |
michael@0 | 3277 | SimpleDateFormat::setTimeZoneFormat(const TimeZoneFormat& newTimeZoneFormat) |
michael@0 | 3278 | { |
michael@0 | 3279 | delete fTimeZoneFormat; |
michael@0 | 3280 | fTimeZoneFormat = new TimeZoneFormat(newTimeZoneFormat); |
michael@0 | 3281 | } |
michael@0 | 3282 | |
michael@0 | 3283 | //---------------------------------------------------------------------- |
michael@0 | 3284 | |
michael@0 | 3285 | |
michael@0 | 3286 | void SimpleDateFormat::adoptCalendar(Calendar* calendarToAdopt) |
michael@0 | 3287 | { |
michael@0 | 3288 | UErrorCode status = U_ZERO_ERROR; |
michael@0 | 3289 | DateFormat::adoptCalendar(calendarToAdopt); |
michael@0 | 3290 | delete fSymbols; |
michael@0 | 3291 | fSymbols=NULL; |
michael@0 | 3292 | initializeSymbols(fLocale, fCalendar, status); // we need new symbols |
michael@0 | 3293 | initializeDefaultCentury(); // we need a new century (possibly) |
michael@0 | 3294 | } |
michael@0 | 3295 | |
michael@0 | 3296 | |
michael@0 | 3297 | //---------------------------------------------------------------------- |
michael@0 | 3298 | |
michael@0 | 3299 | |
michael@0 | 3300 | void SimpleDateFormat::setContext(UDisplayContext value, UErrorCode& status) |
michael@0 | 3301 | { |
michael@0 | 3302 | if (U_FAILURE(status)) |
michael@0 | 3303 | return; |
michael@0 | 3304 | if ( (UDisplayContextType)((uint32_t)value >> 8) == UDISPCTX_TYPE_CAPITALIZATION ) { |
michael@0 | 3305 | fCapitalizationContext = value; |
michael@0 | 3306 | } else { |
michael@0 | 3307 | status = U_ILLEGAL_ARGUMENT_ERROR; |
michael@0 | 3308 | } |
michael@0 | 3309 | } |
michael@0 | 3310 | |
michael@0 | 3311 | |
michael@0 | 3312 | //---------------------------------------------------------------------- |
michael@0 | 3313 | |
michael@0 | 3314 | |
michael@0 | 3315 | UDisplayContext SimpleDateFormat::getContext(UDisplayContextType type, UErrorCode& status) const |
michael@0 | 3316 | { |
michael@0 | 3317 | if (U_FAILURE(status)) |
michael@0 | 3318 | return (UDisplayContext)0; |
michael@0 | 3319 | if (type != UDISPCTX_TYPE_CAPITALIZATION) { |
michael@0 | 3320 | status = U_ILLEGAL_ARGUMENT_ERROR; |
michael@0 | 3321 | return (UDisplayContext)0; |
michael@0 | 3322 | } |
michael@0 | 3323 | return fCapitalizationContext; |
michael@0 | 3324 | } |
michael@0 | 3325 | |
michael@0 | 3326 | |
michael@0 | 3327 | //---------------------------------------------------------------------- |
michael@0 | 3328 | |
michael@0 | 3329 | |
michael@0 | 3330 | UBool |
michael@0 | 3331 | SimpleDateFormat::isFieldUnitIgnored(UCalendarDateFields field) const { |
michael@0 | 3332 | return isFieldUnitIgnored(fPattern, field); |
michael@0 | 3333 | } |
michael@0 | 3334 | |
michael@0 | 3335 | |
michael@0 | 3336 | UBool |
michael@0 | 3337 | SimpleDateFormat::isFieldUnitIgnored(const UnicodeString& pattern, |
michael@0 | 3338 | UCalendarDateFields field) { |
michael@0 | 3339 | int32_t fieldLevel = fgCalendarFieldToLevel[field]; |
michael@0 | 3340 | int32_t level; |
michael@0 | 3341 | UChar ch; |
michael@0 | 3342 | UBool inQuote = FALSE; |
michael@0 | 3343 | UChar prevCh = 0; |
michael@0 | 3344 | int32_t count = 0; |
michael@0 | 3345 | |
michael@0 | 3346 | for (int32_t i = 0; i < pattern.length(); ++i) { |
michael@0 | 3347 | ch = pattern[i]; |
michael@0 | 3348 | if (ch != prevCh && count > 0) { |
michael@0 | 3349 | level = fgPatternCharToLevel[prevCh - PATTERN_CHAR_BASE]; |
michael@0 | 3350 | // the larger the level, the smaller the field unit. |
michael@0 | 3351 | if ( fieldLevel <= level ) { |
michael@0 | 3352 | return FALSE; |
michael@0 | 3353 | } |
michael@0 | 3354 | count = 0; |
michael@0 | 3355 | } |
michael@0 | 3356 | if (ch == QUOTE) { |
michael@0 | 3357 | if ((i+1) < pattern.length() && pattern[i+1] == QUOTE) { |
michael@0 | 3358 | ++i; |
michael@0 | 3359 | } else { |
michael@0 | 3360 | inQuote = ! inQuote; |
michael@0 | 3361 | } |
michael@0 | 3362 | } |
michael@0 | 3363 | else if ( ! inQuote && ((ch >= 0x0061 /*'a'*/ && ch <= 0x007A /*'z'*/) |
michael@0 | 3364 | || (ch >= 0x0041 /*'A'*/ && ch <= 0x005A /*'Z'*/))) { |
michael@0 | 3365 | prevCh = ch; |
michael@0 | 3366 | ++count; |
michael@0 | 3367 | } |
michael@0 | 3368 | } |
michael@0 | 3369 | if ( count > 0 ) { |
michael@0 | 3370 | // last item |
michael@0 | 3371 | level = fgPatternCharToLevel[prevCh - PATTERN_CHAR_BASE]; |
michael@0 | 3372 | if ( fieldLevel <= level ) { |
michael@0 | 3373 | return FALSE; |
michael@0 | 3374 | } |
michael@0 | 3375 | } |
michael@0 | 3376 | return TRUE; |
michael@0 | 3377 | } |
michael@0 | 3378 | |
michael@0 | 3379 | //---------------------------------------------------------------------- |
michael@0 | 3380 | |
michael@0 | 3381 | const Locale& |
michael@0 | 3382 | SimpleDateFormat::getSmpFmtLocale(void) const { |
michael@0 | 3383 | return fLocale; |
michael@0 | 3384 | } |
michael@0 | 3385 | |
michael@0 | 3386 | //---------------------------------------------------------------------- |
michael@0 | 3387 | |
michael@0 | 3388 | int32_t |
michael@0 | 3389 | SimpleDateFormat::checkIntSuffix(const UnicodeString& text, int32_t start, |
michael@0 | 3390 | int32_t patLoc, UBool isNegative) const { |
michael@0 | 3391 | // local variables |
michael@0 | 3392 | UnicodeString suf; |
michael@0 | 3393 | int32_t patternMatch; |
michael@0 | 3394 | int32_t textPreMatch; |
michael@0 | 3395 | int32_t textPostMatch; |
michael@0 | 3396 | |
michael@0 | 3397 | // check that we are still in range |
michael@0 | 3398 | if ( (start > text.length()) || |
michael@0 | 3399 | (start < 0) || |
michael@0 | 3400 | (patLoc < 0) || |
michael@0 | 3401 | (patLoc > fPattern.length())) { |
michael@0 | 3402 | // out of range, don't advance location in text |
michael@0 | 3403 | return start; |
michael@0 | 3404 | } |
michael@0 | 3405 | |
michael@0 | 3406 | // get the suffix |
michael@0 | 3407 | DecimalFormat* decfmt = dynamic_cast<DecimalFormat*>(fNumberFormat); |
michael@0 | 3408 | if (decfmt != NULL) { |
michael@0 | 3409 | if (isNegative) { |
michael@0 | 3410 | suf = decfmt->getNegativeSuffix(suf); |
michael@0 | 3411 | } |
michael@0 | 3412 | else { |
michael@0 | 3413 | suf = decfmt->getPositiveSuffix(suf); |
michael@0 | 3414 | } |
michael@0 | 3415 | } |
michael@0 | 3416 | |
michael@0 | 3417 | // check for suffix |
michael@0 | 3418 | if (suf.length() <= 0) { |
michael@0 | 3419 | return start; |
michael@0 | 3420 | } |
michael@0 | 3421 | |
michael@0 | 3422 | // check suffix will be encountered in the pattern |
michael@0 | 3423 | patternMatch = compareSimpleAffix(suf,fPattern,patLoc); |
michael@0 | 3424 | |
michael@0 | 3425 | // check if a suffix will be encountered in the text |
michael@0 | 3426 | textPreMatch = compareSimpleAffix(suf,text,start); |
michael@0 | 3427 | |
michael@0 | 3428 | // check if a suffix was encountered in the text |
michael@0 | 3429 | textPostMatch = compareSimpleAffix(suf,text,start-suf.length()); |
michael@0 | 3430 | |
michael@0 | 3431 | // check for suffix match |
michael@0 | 3432 | if ((textPreMatch >= 0) && (patternMatch >= 0) && (textPreMatch == patternMatch)) { |
michael@0 | 3433 | return start; |
michael@0 | 3434 | } |
michael@0 | 3435 | else if ((textPostMatch >= 0) && (patternMatch >= 0) && (textPostMatch == patternMatch)) { |
michael@0 | 3436 | return start - suf.length(); |
michael@0 | 3437 | } |
michael@0 | 3438 | |
michael@0 | 3439 | // should not get here |
michael@0 | 3440 | return start; |
michael@0 | 3441 | } |
michael@0 | 3442 | |
michael@0 | 3443 | //---------------------------------------------------------------------- |
michael@0 | 3444 | |
michael@0 | 3445 | int32_t |
michael@0 | 3446 | SimpleDateFormat::compareSimpleAffix(const UnicodeString& affix, |
michael@0 | 3447 | const UnicodeString& input, |
michael@0 | 3448 | int32_t pos) const { |
michael@0 | 3449 | int32_t start = pos; |
michael@0 | 3450 | for (int32_t i=0; i<affix.length(); ) { |
michael@0 | 3451 | UChar32 c = affix.char32At(i); |
michael@0 | 3452 | int32_t len = U16_LENGTH(c); |
michael@0 | 3453 | if (PatternProps::isWhiteSpace(c)) { |
michael@0 | 3454 | // We may have a pattern like: \u200F \u0020 |
michael@0 | 3455 | // and input text like: \u200F \u0020 |
michael@0 | 3456 | // Note that U+200F and U+0020 are Pattern_White_Space but only |
michael@0 | 3457 | // U+0020 is UWhiteSpace. So we have to first do a direct |
michael@0 | 3458 | // match of the run of Pattern_White_Space in the pattern, |
michael@0 | 3459 | // then match any extra characters. |
michael@0 | 3460 | UBool literalMatch = FALSE; |
michael@0 | 3461 | while (pos < input.length() && |
michael@0 | 3462 | input.char32At(pos) == c) { |
michael@0 | 3463 | literalMatch = TRUE; |
michael@0 | 3464 | i += len; |
michael@0 | 3465 | pos += len; |
michael@0 | 3466 | if (i == affix.length()) { |
michael@0 | 3467 | break; |
michael@0 | 3468 | } |
michael@0 | 3469 | c = affix.char32At(i); |
michael@0 | 3470 | len = U16_LENGTH(c); |
michael@0 | 3471 | if (!PatternProps::isWhiteSpace(c)) { |
michael@0 | 3472 | break; |
michael@0 | 3473 | } |
michael@0 | 3474 | } |
michael@0 | 3475 | |
michael@0 | 3476 | // Advance over run in pattern |
michael@0 | 3477 | i = skipPatternWhiteSpace(affix, i); |
michael@0 | 3478 | |
michael@0 | 3479 | // Advance over run in input text |
michael@0 | 3480 | // Must see at least one white space char in input, |
michael@0 | 3481 | // unless we've already matched some characters literally. |
michael@0 | 3482 | int32_t s = pos; |
michael@0 | 3483 | pos = skipUWhiteSpace(input, pos); |
michael@0 | 3484 | if (pos == s && !literalMatch) { |
michael@0 | 3485 | return -1; |
michael@0 | 3486 | } |
michael@0 | 3487 | |
michael@0 | 3488 | // If we skip UWhiteSpace in the input text, we need to skip it in the pattern. |
michael@0 | 3489 | // Otherwise, the previous lines may have skipped over text (such as U+00A0) that |
michael@0 | 3490 | // is also in the affix. |
michael@0 | 3491 | i = skipUWhiteSpace(affix, i); |
michael@0 | 3492 | } else { |
michael@0 | 3493 | if (pos < input.length() && |
michael@0 | 3494 | input.char32At(pos) == c) { |
michael@0 | 3495 | i += len; |
michael@0 | 3496 | pos += len; |
michael@0 | 3497 | } else { |
michael@0 | 3498 | return -1; |
michael@0 | 3499 | } |
michael@0 | 3500 | } |
michael@0 | 3501 | } |
michael@0 | 3502 | return pos - start; |
michael@0 | 3503 | } |
michael@0 | 3504 | |
michael@0 | 3505 | //---------------------------------------------------------------------- |
michael@0 | 3506 | |
michael@0 | 3507 | int32_t |
michael@0 | 3508 | SimpleDateFormat::skipPatternWhiteSpace(const UnicodeString& text, int32_t pos) const { |
michael@0 | 3509 | const UChar* s = text.getBuffer(); |
michael@0 | 3510 | return (int32_t)(PatternProps::skipWhiteSpace(s + pos, text.length() - pos) - s); |
michael@0 | 3511 | } |
michael@0 | 3512 | |
michael@0 | 3513 | //---------------------------------------------------------------------- |
michael@0 | 3514 | |
michael@0 | 3515 | int32_t |
michael@0 | 3516 | SimpleDateFormat::skipUWhiteSpace(const UnicodeString& text, int32_t pos) const { |
michael@0 | 3517 | while (pos < text.length()) { |
michael@0 | 3518 | UChar32 c = text.char32At(pos); |
michael@0 | 3519 | if (!u_isUWhiteSpace(c)) { |
michael@0 | 3520 | break; |
michael@0 | 3521 | } |
michael@0 | 3522 | pos += U16_LENGTH(c); |
michael@0 | 3523 | } |
michael@0 | 3524 | return pos; |
michael@0 | 3525 | } |
michael@0 | 3526 | |
michael@0 | 3527 | //---------------------------------------------------------------------- |
michael@0 | 3528 | |
michael@0 | 3529 | // Lazy TimeZoneFormat instantiation, semantically const. |
michael@0 | 3530 | TimeZoneFormat * |
michael@0 | 3531 | SimpleDateFormat::tzFormat() const { |
michael@0 | 3532 | if (fTimeZoneFormat == NULL) { |
michael@0 | 3533 | umtx_lock(&LOCK); |
michael@0 | 3534 | { |
michael@0 | 3535 | if (fTimeZoneFormat == NULL) { |
michael@0 | 3536 | UErrorCode status = U_ZERO_ERROR; |
michael@0 | 3537 | TimeZoneFormat *tzfmt = TimeZoneFormat::createInstance(fLocale, status); |
michael@0 | 3538 | if (U_FAILURE(status)) { |
michael@0 | 3539 | return NULL; |
michael@0 | 3540 | } |
michael@0 | 3541 | |
michael@0 | 3542 | const_cast<SimpleDateFormat *>(this)->fTimeZoneFormat = tzfmt; |
michael@0 | 3543 | } |
michael@0 | 3544 | } |
michael@0 | 3545 | umtx_unlock(&LOCK); |
michael@0 | 3546 | } |
michael@0 | 3547 | return fTimeZoneFormat; |
michael@0 | 3548 | } |
michael@0 | 3549 | |
michael@0 | 3550 | U_NAMESPACE_END |
michael@0 | 3551 | |
michael@0 | 3552 | #endif /* #if !UCONFIG_NO_FORMATTING */ |
michael@0 | 3553 | |
michael@0 | 3554 | //eof |