1 | // © 2016 and later: Unicode, Inc. and others. |
2 | // License & terms of use: http://www.unicode.org/copyright.html |
3 | /* |
4 | ******************************************************************************* |
5 | * Copyright (C) 2011-2015, International Business Machines Corporation and |
6 | * others. All Rights Reserved. |
7 | ******************************************************************************* |
8 | */ |
9 | |
10 | #include "unicode/utypes.h" |
11 | |
12 | #if !UCONFIG_NO_FORMATTING |
13 | |
14 | #include "unicode/calendar.h" |
15 | #include "unicode/tzfmt.h" |
16 | #include "unicode/numsys.h" |
17 | #include "unicode/strenum.h" |
18 | #include "unicode/uchar.h" |
19 | #include "unicode/udat.h" |
20 | #include "unicode/ustring.h" |
21 | #include "unicode/utf16.h" |
22 | #include "bytesinkutil.h" |
23 | #include "charstr.h" |
24 | #include "tzgnames.h" |
25 | #include "cmemory.h" |
26 | #include "cstring.h" |
27 | #include "putilimp.h" |
28 | #include "uassert.h" |
29 | #include "ucln_in.h" |
30 | #include "ulocimp.h" |
31 | #include "umutex.h" |
32 | #include "uresimp.h" |
33 | #include "ureslocs.h" |
34 | #include "uvector.h" |
35 | #include "zonemeta.h" |
36 | #include "tznames_impl.h" // TextTrieMap |
37 | #include "patternprops.h" |
38 | |
39 | U_NAMESPACE_BEGIN |
40 | |
41 | // Bit flags used by the parse method. |
42 | // The order must match UTimeZoneFormatStyle enum. |
43 | #define ISO_Z_STYLE_FLAG 0x0080 |
44 | #define ISO_LOCAL_STYLE_FLAG 0x0100 |
45 | static const int16_t STYLE_PARSE_FLAGS[] = { |
46 | 0x0001, // UTZFMT_STYLE_GENERIC_LOCATION, |
47 | 0x0002, // UTZFMT_STYLE_GENERIC_LONG, |
48 | 0x0004, // UTZFMT_STYLE_GENERIC_SHORT, |
49 | 0x0008, // UTZFMT_STYLE_SPECIFIC_LONG, |
50 | 0x0010, // UTZFMT_STYLE_SPECIFIC_SHORT, |
51 | 0x0020, // UTZFMT_STYLE_LOCALIZED_GMT, |
52 | 0x0040, // UTZFMT_STYLE_LOCALIZED_GMT_SHORT, |
53 | ISO_Z_STYLE_FLAG, // UTZFMT_STYLE_ISO_BASIC_SHORT, |
54 | ISO_LOCAL_STYLE_FLAG, // UTZFMT_STYLE_ISO_BASIC_LOCAL_SHORT, |
55 | ISO_Z_STYLE_FLAG, // UTZFMT_STYLE_ISO_BASIC_FIXED, |
56 | ISO_LOCAL_STYLE_FLAG, // UTZFMT_STYLE_ISO_BASIC_LOCAL_FIXED, |
57 | ISO_Z_STYLE_FLAG, // UTZFMT_STYLE_ISO_BASIC_FULL, |
58 | ISO_LOCAL_STYLE_FLAG, // UTZFMT_STYLE_ISO_BASIC_LOCAL_FULL, |
59 | ISO_Z_STYLE_FLAG, // UTZFMT_STYLE_ISO_EXTENDED_FIXED, |
60 | ISO_LOCAL_STYLE_FLAG, // UTZFMT_STYLE_ISO_EXTENDED_LOCAL_FIXED, |
61 | ISO_Z_STYLE_FLAG, // UTZFMT_STYLE_ISO_EXTENDED_FULL, |
62 | ISO_LOCAL_STYLE_FLAG, // UTZFMT_STYLE_ISO_EXTENDED_LOCAL_FULL, |
63 | 0x0200, // UTZFMT_STYLE_ZONE_ID, |
64 | 0x0400, // UTZFMT_STYLE_ZONE_ID_SHORT, |
65 | 0x0800 // UTZFMT_STYLE_EXEMPLAR_LOCATION |
66 | }; |
67 | |
68 | static const char gZoneStringsTag[] = "zoneStrings" ; |
69 | static const char gGmtFormatTag[]= "gmtFormat" ; |
70 | static const char gGmtZeroFormatTag[] = "gmtZeroFormat" ; |
71 | static const char gHourFormatTag[]= "hourFormat" ; |
72 | |
73 | static const UChar TZID_GMT[] = {0x0045, 0x0074, 0x0063, 0x002F, 0x0047, 0x004D, 0x0054, 0}; // Etc/GMT |
74 | static const UChar UNKNOWN_ZONE_ID[] = { |
75 | 0x0045, 0x0074, 0x0063, 0x002F, 0x0055, 0x006E, 0x006B, 0x006E, 0x006F, 0x0077, 0x006E, 0}; // Etc/Unknown |
76 | static const UChar UNKNOWN_SHORT_ZONE_ID[] = {0x0075, 0x006E, 0x006B, 0}; // unk |
77 | static const UChar UNKNOWN_LOCATION[] = {0x0055, 0x006E, 0x006B, 0x006E, 0x006F, 0x0077, 0x006E, 0}; // Unknown |
78 | |
79 | static const UChar DEFAULT_GMT_PATTERN[] = {0x0047, 0x004D, 0x0054, 0x007B, 0x0030, 0x007D, 0}; // GMT{0} |
80 | //static const UChar DEFAULT_GMT_ZERO[] = {0x0047, 0x004D, 0x0054, 0}; // GMT |
81 | static const UChar DEFAULT_GMT_POSITIVE_HM[] = {0x002B, 0x0048, 0x003A, 0x006D, 0x006D, 0}; // +H:mm |
82 | static const UChar DEFAULT_GMT_POSITIVE_HMS[] = {0x002B, 0x0048, 0x003A, 0x006D, 0x006D, 0x003A, 0x0073, 0x0073, 0}; // +H:mm:ss |
83 | static const UChar DEFAULT_GMT_NEGATIVE_HM[] = {0x002D, 0x0048, 0x003A, 0x006D, 0x006D, 0}; // -H:mm |
84 | static const UChar DEFAULT_GMT_NEGATIVE_HMS[] = {0x002D, 0x0048, 0x003A, 0x006D, 0x006D, 0x003A, 0x0073, 0x0073, 0}; // -H:mm:ss |
85 | static const UChar DEFAULT_GMT_POSITIVE_H[] = {0x002B, 0x0048, 0}; // +H |
86 | static const UChar DEFAULT_GMT_NEGATIVE_H[] = {0x002D, 0x0048, 0}; // -H |
87 | |
88 | static const UChar32 DEFAULT_GMT_DIGITS[] = { |
89 | 0x0030, 0x0031, 0x0032, 0x0033, 0x0034, |
90 | 0x0035, 0x0036, 0x0037, 0x0038, 0x0039 |
91 | }; |
92 | |
93 | static const UChar DEFAULT_GMT_OFFSET_SEP = 0x003A; // ':' |
94 | |
95 | static const UChar ARG0[] = {0x007B, 0x0030, 0x007D}; // "{0}" |
96 | static const int32_t ARG0_LEN = 3; |
97 | |
98 | static const UChar DEFAULT_GMT_OFFSET_MINUTE_PATTERN[] = {0x006D, 0x006D, 0}; // "mm" |
99 | static const UChar DEFAULT_GMT_OFFSET_SECOND_PATTERN[] = {0x0073, 0x0073, 0}; // "ss" |
100 | |
101 | static const UChar ALT_GMT_STRINGS[][4] = { |
102 | {0x0047, 0x004D, 0x0054, 0}, // GMT |
103 | {0x0055, 0x0054, 0x0043, 0}, // UTC |
104 | {0x0055, 0x0054, 0, 0}, // UT |
105 | {0, 0, 0, 0} |
106 | }; |
107 | |
108 | // Order of GMT offset pattern parsing, *_HMS must be evaluated first |
109 | // because *_HM is most likely a substring of *_HMS |
110 | static const int32_t PARSE_GMT_OFFSET_TYPES[] = { |
111 | UTZFMT_PAT_POSITIVE_HMS, |
112 | UTZFMT_PAT_NEGATIVE_HMS, |
113 | UTZFMT_PAT_POSITIVE_HM, |
114 | UTZFMT_PAT_NEGATIVE_HM, |
115 | UTZFMT_PAT_POSITIVE_H, |
116 | UTZFMT_PAT_NEGATIVE_H, |
117 | -1 |
118 | }; |
119 | |
120 | static const UChar SINGLEQUOTE = 0x0027; |
121 | static const UChar PLUS = 0x002B; |
122 | static const UChar MINUS = 0x002D; |
123 | static const UChar ISO8601_UTC = 0x005A; // 'Z' |
124 | static const UChar ISO8601_SEP = 0x003A; // ':' |
125 | |
126 | static const int32_t MILLIS_PER_HOUR = 60 * 60 * 1000; |
127 | static const int32_t MILLIS_PER_MINUTE = 60 * 1000; |
128 | static const int32_t MILLIS_PER_SECOND = 1000; |
129 | |
130 | // Maximum offset (exclusive) in millisecond supported by offset formats |
131 | static int32_t MAX_OFFSET = 24 * MILLIS_PER_HOUR; |
132 | |
133 | // Maximum values for GMT offset fields |
134 | static const int32_t MAX_OFFSET_HOUR = 23; |
135 | static const int32_t MAX_OFFSET_MINUTE = 59; |
136 | static const int32_t MAX_OFFSET_SECOND = 59; |
137 | |
138 | static const int32_t UNKNOWN_OFFSET = 0x7FFFFFFF; |
139 | |
140 | static const int32_t ALL_SIMPLE_NAME_TYPES = UTZNM_LONG_STANDARD | UTZNM_LONG_DAYLIGHT | UTZNM_SHORT_STANDARD | UTZNM_SHORT_DAYLIGHT | UTZNM_EXEMPLAR_LOCATION; |
141 | static const int32_t ALL_GENERIC_NAME_TYPES = UTZGNM_LOCATION | UTZGNM_LONG | UTZGNM_SHORT; |
142 | |
143 | #define DIGIT_VAL(c) (0x0030 <= (c) && (c) <= 0x0039 ? (c) - 0x0030 : -1) |
144 | #define MAX_OFFSET_DIGITS 6 |
145 | |
146 | // Time Zone ID/Short ID trie |
147 | static TextTrieMap *gZoneIdTrie = NULL; |
148 | static icu::UInitOnce gZoneIdTrieInitOnce = U_INITONCE_INITIALIZER; |
149 | |
150 | static TextTrieMap *gShortZoneIdTrie = NULL; |
151 | static icu::UInitOnce gShortZoneIdTrieInitOnce = U_INITONCE_INITIALIZER; |
152 | |
153 | static UMutex gLock; |
154 | |
155 | U_CDECL_BEGIN |
156 | /** |
157 | * Cleanup callback func |
158 | */ |
159 | static UBool U_CALLCONV tzfmt_cleanup(void) |
160 | { |
161 | if (gZoneIdTrie != NULL) { |
162 | delete gZoneIdTrie; |
163 | } |
164 | gZoneIdTrie = NULL; |
165 | gZoneIdTrieInitOnce.reset(); |
166 | |
167 | if (gShortZoneIdTrie != NULL) { |
168 | delete gShortZoneIdTrie; |
169 | } |
170 | gShortZoneIdTrie = NULL; |
171 | gShortZoneIdTrieInitOnce.reset(); |
172 | |
173 | return TRUE; |
174 | } |
175 | U_CDECL_END |
176 | |
177 | // ------------------------------------------------------------------ |
178 | // GMTOffsetField |
179 | // |
180 | // This class represents a localized GMT offset pattern |
181 | // item and used by TimeZoneFormat |
182 | // ------------------------------------------------------------------ |
183 | class GMTOffsetField : public UMemory { |
184 | public: |
185 | enum FieldType { |
186 | TEXT = 0, |
187 | HOUR = 1, |
188 | MINUTE = 2, |
189 | SECOND = 4 |
190 | }; |
191 | |
192 | virtual ~GMTOffsetField(); |
193 | |
194 | static GMTOffsetField* createText(const UnicodeString& text, UErrorCode& status); |
195 | static GMTOffsetField* createTimeField(FieldType type, uint8_t width, UErrorCode& status); |
196 | static UBool isValid(FieldType type, int32_t width); |
197 | static FieldType getTypeByLetter(UChar ch); |
198 | |
199 | FieldType getType() const; |
200 | uint8_t getWidth() const; |
201 | const UChar* getPatternText(void) const; |
202 | |
203 | private: |
204 | UChar* fText; |
205 | FieldType fType; |
206 | uint8_t fWidth; |
207 | |
208 | GMTOffsetField(); |
209 | }; |
210 | |
211 | GMTOffsetField::GMTOffsetField() |
212 | : fText(NULL), fType(TEXT), fWidth(0) { |
213 | } |
214 | |
215 | GMTOffsetField::~GMTOffsetField() { |
216 | if (fText) { |
217 | uprv_free(fText); |
218 | } |
219 | } |
220 | |
221 | GMTOffsetField* |
222 | GMTOffsetField::createText(const UnicodeString& text, UErrorCode& status) { |
223 | if (U_FAILURE(status)) { |
224 | return NULL; |
225 | } |
226 | GMTOffsetField* result = new GMTOffsetField(); |
227 | if (result == NULL) { |
228 | status = U_MEMORY_ALLOCATION_ERROR; |
229 | return NULL; |
230 | } |
231 | |
232 | int32_t len = text.length(); |
233 | result->fText = (UChar*)uprv_malloc((len + 1) * sizeof(UChar)); |
234 | if (result->fText == NULL) { |
235 | status = U_MEMORY_ALLOCATION_ERROR; |
236 | delete result; |
237 | return NULL; |
238 | } |
239 | u_strncpy(result->fText, text.getBuffer(), len); |
240 | result->fText[len] = 0; |
241 | result->fType = TEXT; |
242 | |
243 | return result; |
244 | } |
245 | |
246 | GMTOffsetField* |
247 | GMTOffsetField::createTimeField(FieldType type, uint8_t width, UErrorCode& status) { |
248 | U_ASSERT(type != TEXT); |
249 | if (U_FAILURE(status)) { |
250 | return NULL; |
251 | } |
252 | GMTOffsetField* result = new GMTOffsetField(); |
253 | if (result == NULL) { |
254 | status = U_MEMORY_ALLOCATION_ERROR; |
255 | return NULL; |
256 | } |
257 | |
258 | result->fType = type; |
259 | result->fWidth = width; |
260 | |
261 | return result; |
262 | } |
263 | |
264 | UBool |
265 | GMTOffsetField::isValid(FieldType type, int32_t width) { |
266 | switch (type) { |
267 | case HOUR: |
268 | return (width == 1 || width == 2); |
269 | case MINUTE: |
270 | case SECOND: |
271 | return (width == 2); |
272 | default: |
273 | UPRV_UNREACHABLE; |
274 | } |
275 | return (width > 0); |
276 | } |
277 | |
278 | GMTOffsetField::FieldType |
279 | GMTOffsetField::getTypeByLetter(UChar ch) { |
280 | if (ch == 0x0048 /* H */) { |
281 | return HOUR; |
282 | } else if (ch == 0x006D /* m */) { |
283 | return MINUTE; |
284 | } else if (ch == 0x0073 /* s */) { |
285 | return SECOND; |
286 | } |
287 | return TEXT; |
288 | } |
289 | |
290 | inline GMTOffsetField::FieldType |
291 | GMTOffsetField::getType() const { |
292 | return fType; |
293 | } |
294 | |
295 | inline uint8_t |
296 | GMTOffsetField::getWidth() const { |
297 | return fWidth; |
298 | } |
299 | |
300 | inline const UChar* |
301 | GMTOffsetField::getPatternText(void) const { |
302 | return fText; |
303 | } |
304 | |
305 | |
306 | U_CDECL_BEGIN |
307 | static void U_CALLCONV |
308 | deleteGMTOffsetField(void *obj) { |
309 | delete static_cast<GMTOffsetField *>(obj); |
310 | } |
311 | U_CDECL_END |
312 | |
313 | |
314 | // ------------------------------------------------------------------ |
315 | // TimeZoneFormat |
316 | // ------------------------------------------------------------------ |
317 | UOBJECT_DEFINE_RTTI_IMPLEMENTATION(TimeZoneFormat) |
318 | |
319 | TimeZoneFormat::TimeZoneFormat(const Locale& locale, UErrorCode& status) |
320 | : fLocale(locale), fTimeZoneNames(NULL), fTimeZoneGenericNames(NULL), |
321 | fDefParseOptionFlags(0), fTZDBTimeZoneNames(NULL) { |
322 | |
323 | for (int32_t i = 0; i < UTZFMT_PAT_COUNT; i++) { |
324 | fGMTOffsetPatternItems[i] = NULL; |
325 | } |
326 | |
327 | const char* region = fLocale.getCountry(); |
328 | int32_t regionLen = static_cast<int32_t>(uprv_strlen(region)); |
329 | if (regionLen == 0) { |
330 | CharString loc; |
331 | { |
332 | CharStringByteSink sink(&loc); |
333 | ulocimp_addLikelySubtags(fLocale.getName(), sink, &status); |
334 | } |
335 | |
336 | regionLen = uloc_getCountry(loc.data(), fTargetRegion, sizeof(fTargetRegion), &status); |
337 | if (U_SUCCESS(status)) { |
338 | fTargetRegion[regionLen] = 0; |
339 | } else { |
340 | return; |
341 | } |
342 | } else if (regionLen < (int32_t)sizeof(fTargetRegion)) { |
343 | uprv_strcpy(fTargetRegion, region); |
344 | } else { |
345 | fTargetRegion[0] = 0; |
346 | } |
347 | |
348 | fTimeZoneNames = TimeZoneNames::createInstance(locale, status); |
349 | // fTimeZoneGenericNames is lazily instantiated |
350 | if (U_FAILURE(status)) { |
351 | return; |
352 | } |
353 | |
354 | const UChar* gmtPattern = NULL; |
355 | const UChar* hourFormats = NULL; |
356 | |
357 | UResourceBundle *zoneBundle = ures_open(U_ICUDATA_ZONE, locale.getName(), &status); |
358 | UResourceBundle *zoneStringsArray = ures_getByKeyWithFallback(zoneBundle, gZoneStringsTag, NULL, &status); |
359 | if (U_SUCCESS(status)) { |
360 | const UChar* resStr; |
361 | int32_t len; |
362 | resStr = ures_getStringByKeyWithFallback(zoneStringsArray, gGmtFormatTag, &len, &status); |
363 | if (len > 0) { |
364 | gmtPattern = resStr; |
365 | } |
366 | resStr = ures_getStringByKeyWithFallback(zoneStringsArray, gGmtZeroFormatTag, &len, &status); |
367 | if (len > 0) { |
368 | fGMTZeroFormat.setTo(TRUE, resStr, len); |
369 | } |
370 | resStr = ures_getStringByKeyWithFallback(zoneStringsArray, gHourFormatTag, &len, &status); |
371 | if (len > 0) { |
372 | hourFormats = resStr; |
373 | } |
374 | ures_close(zoneStringsArray); |
375 | ures_close(zoneBundle); |
376 | } |
377 | |
378 | if (gmtPattern == NULL) { |
379 | gmtPattern = DEFAULT_GMT_PATTERN; |
380 | } |
381 | initGMTPattern(UnicodeString(TRUE, gmtPattern, -1), status); |
382 | |
383 | UBool useDefaultOffsetPatterns = TRUE; |
384 | if (hourFormats) { |
385 | UChar *sep = u_strchr(hourFormats, (UChar)0x003B /* ';' */); |
386 | if (sep != NULL) { |
387 | UErrorCode tmpStatus = U_ZERO_ERROR; |
388 | fGMTOffsetPatterns[UTZFMT_PAT_POSITIVE_HM].setTo(FALSE, hourFormats, (int32_t)(sep - hourFormats)); |
389 | fGMTOffsetPatterns[UTZFMT_PAT_NEGATIVE_HM].setTo(TRUE, sep + 1, -1); |
390 | expandOffsetPattern(fGMTOffsetPatterns[UTZFMT_PAT_POSITIVE_HM], fGMTOffsetPatterns[UTZFMT_PAT_POSITIVE_HMS], tmpStatus); |
391 | expandOffsetPattern(fGMTOffsetPatterns[UTZFMT_PAT_NEGATIVE_HM], fGMTOffsetPatterns[UTZFMT_PAT_NEGATIVE_HMS], tmpStatus); |
392 | truncateOffsetPattern(fGMTOffsetPatterns[UTZFMT_PAT_POSITIVE_HM], fGMTOffsetPatterns[UTZFMT_PAT_POSITIVE_H], tmpStatus); |
393 | truncateOffsetPattern(fGMTOffsetPatterns[UTZFMT_PAT_NEGATIVE_HM], fGMTOffsetPatterns[UTZFMT_PAT_NEGATIVE_H], tmpStatus); |
394 | if (U_SUCCESS(tmpStatus)) { |
395 | useDefaultOffsetPatterns = FALSE; |
396 | } |
397 | } |
398 | } |
399 | if (useDefaultOffsetPatterns) { |
400 | fGMTOffsetPatterns[UTZFMT_PAT_POSITIVE_H].setTo(TRUE, DEFAULT_GMT_POSITIVE_H, -1); |
401 | fGMTOffsetPatterns[UTZFMT_PAT_POSITIVE_HM].setTo(TRUE, DEFAULT_GMT_POSITIVE_HM, -1); |
402 | fGMTOffsetPatterns[UTZFMT_PAT_POSITIVE_HMS].setTo(TRUE, DEFAULT_GMT_POSITIVE_HMS, -1); |
403 | fGMTOffsetPatterns[UTZFMT_PAT_NEGATIVE_H].setTo(TRUE, DEFAULT_GMT_NEGATIVE_H, -1); |
404 | fGMTOffsetPatterns[UTZFMT_PAT_NEGATIVE_HM].setTo(TRUE, DEFAULT_GMT_NEGATIVE_HM, -1); |
405 | fGMTOffsetPatterns[UTZFMT_PAT_NEGATIVE_HMS].setTo(TRUE, DEFAULT_GMT_NEGATIVE_HMS, -1); |
406 | } |
407 | initGMTOffsetPatterns(status); |
408 | |
409 | NumberingSystem* ns = NumberingSystem::createInstance(locale, status); |
410 | UBool useDefDigits = TRUE; |
411 | if (ns && !ns->isAlgorithmic()) { |
412 | UnicodeString digits = ns->getDescription(); |
413 | useDefDigits = !toCodePoints(digits, fGMTOffsetDigits, 10); |
414 | } |
415 | if (useDefDigits) { |
416 | uprv_memcpy(fGMTOffsetDigits, DEFAULT_GMT_DIGITS, sizeof(UChar32) * 10); |
417 | } |
418 | delete ns; |
419 | } |
420 | |
421 | TimeZoneFormat::TimeZoneFormat(const TimeZoneFormat& other) |
422 | : Format(other), fTimeZoneNames(NULL), fTimeZoneGenericNames(NULL), |
423 | fTZDBTimeZoneNames(NULL) { |
424 | |
425 | for (int32_t i = 0; i < UTZFMT_PAT_COUNT; i++) { |
426 | fGMTOffsetPatternItems[i] = NULL; |
427 | } |
428 | *this = other; |
429 | } |
430 | |
431 | |
432 | TimeZoneFormat::~TimeZoneFormat() { |
433 | delete fTimeZoneNames; |
434 | delete fTimeZoneGenericNames; |
435 | delete fTZDBTimeZoneNames; |
436 | for (int32_t i = 0; i < UTZFMT_PAT_COUNT; i++) { |
437 | delete fGMTOffsetPatternItems[i]; |
438 | } |
439 | } |
440 | |
441 | TimeZoneFormat& |
442 | TimeZoneFormat::operator=(const TimeZoneFormat& other) { |
443 | if (this == &other) { |
444 | return *this; |
445 | } |
446 | |
447 | delete fTimeZoneNames; |
448 | delete fTimeZoneGenericNames; |
449 | fTimeZoneGenericNames = NULL; |
450 | delete fTZDBTimeZoneNames; |
451 | fTZDBTimeZoneNames = NULL; |
452 | |
453 | fLocale = other.fLocale; |
454 | uprv_memcpy(fTargetRegion, other.fTargetRegion, sizeof(fTargetRegion)); |
455 | |
456 | fTimeZoneNames = other.fTimeZoneNames->clone(); |
457 | if (other.fTimeZoneGenericNames) { |
458 | // TODO: this test has dubious thread safety. |
459 | fTimeZoneGenericNames = other.fTimeZoneGenericNames->clone(); |
460 | } |
461 | |
462 | fGMTPattern = other.fGMTPattern; |
463 | fGMTPatternPrefix = other.fGMTPatternPrefix; |
464 | fGMTPatternSuffix = other.fGMTPatternSuffix; |
465 | |
466 | UErrorCode status = U_ZERO_ERROR; |
467 | for (int32_t i = 0; i < UTZFMT_PAT_COUNT; i++) { |
468 | fGMTOffsetPatterns[i] = other.fGMTOffsetPatterns[i]; |
469 | delete fGMTOffsetPatternItems[i]; |
470 | fGMTOffsetPatternItems[i] = NULL; |
471 | } |
472 | initGMTOffsetPatterns(status); |
473 | U_ASSERT(U_SUCCESS(status)); |
474 | |
475 | fGMTZeroFormat = other.fGMTZeroFormat; |
476 | |
477 | uprv_memcpy(fGMTOffsetDigits, other.fGMTOffsetDigits, sizeof(fGMTOffsetDigits)); |
478 | |
479 | fDefParseOptionFlags = other.fDefParseOptionFlags; |
480 | |
481 | return *this; |
482 | } |
483 | |
484 | |
485 | UBool |
486 | TimeZoneFormat::operator==(const Format& other) const { |
487 | TimeZoneFormat* tzfmt = (TimeZoneFormat*)&other; |
488 | |
489 | UBool isEqual = |
490 | fLocale == tzfmt->fLocale |
491 | && fGMTPattern == tzfmt->fGMTPattern |
492 | && fGMTZeroFormat == tzfmt->fGMTZeroFormat |
493 | && *fTimeZoneNames == *tzfmt->fTimeZoneNames; |
494 | |
495 | for (int32_t i = 0; i < UTZFMT_PAT_COUNT && isEqual; i++) { |
496 | isEqual = fGMTOffsetPatterns[i] == tzfmt->fGMTOffsetPatterns[i]; |
497 | } |
498 | for (int32_t i = 0; i < 10 && isEqual; i++) { |
499 | isEqual = fGMTOffsetDigits[i] == tzfmt->fGMTOffsetDigits[i]; |
500 | } |
501 | // TODO |
502 | // Check fTimeZoneGenericNames. For now, |
503 | // if fTimeZoneNames is same, fTimeZoneGenericNames should |
504 | // be also equivalent. |
505 | return isEqual; |
506 | } |
507 | |
508 | TimeZoneFormat* |
509 | TimeZoneFormat::clone() const { |
510 | return new TimeZoneFormat(*this); |
511 | } |
512 | |
513 | TimeZoneFormat* U_EXPORT2 |
514 | TimeZoneFormat::createInstance(const Locale& locale, UErrorCode& status) { |
515 | TimeZoneFormat* tzfmt = new TimeZoneFormat(locale, status); |
516 | if (U_SUCCESS(status)) { |
517 | return tzfmt; |
518 | } |
519 | delete tzfmt; |
520 | return NULL; |
521 | } |
522 | |
523 | // ------------------------------------------------------------------ |
524 | // Setter and Getter |
525 | |
526 | const TimeZoneNames* |
527 | TimeZoneFormat::getTimeZoneNames() const { |
528 | return (const TimeZoneNames*)fTimeZoneNames; |
529 | } |
530 | |
531 | void |
532 | TimeZoneFormat::adoptTimeZoneNames(TimeZoneNames *tznames) { |
533 | delete fTimeZoneNames; |
534 | fTimeZoneNames = tznames; |
535 | |
536 | // TODO - We should also update fTimeZoneGenericNames |
537 | } |
538 | |
539 | void |
540 | TimeZoneFormat::setTimeZoneNames(const TimeZoneNames &tznames) { |
541 | delete fTimeZoneNames; |
542 | fTimeZoneNames = tznames.clone(); |
543 | |
544 | // TODO - We should also update fTimeZoneGenericNames |
545 | } |
546 | |
547 | void |
548 | TimeZoneFormat::setDefaultParseOptions(uint32_t flags) { |
549 | fDefParseOptionFlags = flags; |
550 | } |
551 | |
552 | uint32_t |
553 | TimeZoneFormat::getDefaultParseOptions(void) const { |
554 | return fDefParseOptionFlags; |
555 | } |
556 | |
557 | |
558 | UnicodeString& |
559 | TimeZoneFormat::getGMTPattern(UnicodeString& pattern) const { |
560 | return pattern.setTo(fGMTPattern); |
561 | } |
562 | |
563 | void |
564 | TimeZoneFormat::setGMTPattern(const UnicodeString& pattern, UErrorCode& status) { |
565 | initGMTPattern(pattern, status); |
566 | } |
567 | |
568 | UnicodeString& |
569 | TimeZoneFormat::getGMTOffsetPattern(UTimeZoneFormatGMTOffsetPatternType type, UnicodeString& pattern) const { |
570 | return pattern.setTo(fGMTOffsetPatterns[type]); |
571 | } |
572 | |
573 | void |
574 | TimeZoneFormat::setGMTOffsetPattern(UTimeZoneFormatGMTOffsetPatternType type, const UnicodeString& pattern, UErrorCode& status) { |
575 | if (U_FAILURE(status)) { |
576 | return; |
577 | } |
578 | if (pattern == fGMTOffsetPatterns[type]) { |
579 | // No need to reset |
580 | return; |
581 | } |
582 | |
583 | OffsetFields required = FIELDS_HM; |
584 | switch (type) { |
585 | case UTZFMT_PAT_POSITIVE_H: |
586 | case UTZFMT_PAT_NEGATIVE_H: |
587 | required = FIELDS_H; |
588 | break; |
589 | case UTZFMT_PAT_POSITIVE_HM: |
590 | case UTZFMT_PAT_NEGATIVE_HM: |
591 | required = FIELDS_HM; |
592 | break; |
593 | case UTZFMT_PAT_POSITIVE_HMS: |
594 | case UTZFMT_PAT_NEGATIVE_HMS: |
595 | required = FIELDS_HMS; |
596 | break; |
597 | default: |
598 | UPRV_UNREACHABLE; |
599 | } |
600 | |
601 | UVector* patternItems = parseOffsetPattern(pattern, required, status); |
602 | if (patternItems == NULL) { |
603 | return; |
604 | } |
605 | |
606 | fGMTOffsetPatterns[type].setTo(pattern); |
607 | delete fGMTOffsetPatternItems[type]; |
608 | fGMTOffsetPatternItems[type] = patternItems; |
609 | checkAbuttingHoursAndMinutes(); |
610 | } |
611 | |
612 | UnicodeString& |
613 | TimeZoneFormat::getGMTOffsetDigits(UnicodeString& digits) const { |
614 | digits.remove(); |
615 | for (int32_t i = 0; i < 10; i++) { |
616 | digits.append(fGMTOffsetDigits[i]); |
617 | } |
618 | return digits; |
619 | } |
620 | |
621 | void |
622 | TimeZoneFormat::setGMTOffsetDigits(const UnicodeString& digits, UErrorCode& status) { |
623 | if (U_FAILURE(status)) { |
624 | return; |
625 | } |
626 | UChar32 digitArray[10]; |
627 | if (!toCodePoints(digits, digitArray, 10)) { |
628 | status = U_ILLEGAL_ARGUMENT_ERROR; |
629 | return; |
630 | } |
631 | uprv_memcpy(fGMTOffsetDigits, digitArray, sizeof(UChar32)*10); |
632 | } |
633 | |
634 | UnicodeString& |
635 | TimeZoneFormat::getGMTZeroFormat(UnicodeString& gmtZeroFormat) const { |
636 | return gmtZeroFormat.setTo(fGMTZeroFormat); |
637 | } |
638 | |
639 | void |
640 | TimeZoneFormat::setGMTZeroFormat(const UnicodeString& gmtZeroFormat, UErrorCode& status) { |
641 | if (U_SUCCESS(status)) { |
642 | if (gmtZeroFormat.isEmpty()) { |
643 | status = U_ILLEGAL_ARGUMENT_ERROR; |
644 | } else if (gmtZeroFormat != fGMTZeroFormat) { |
645 | fGMTZeroFormat.setTo(gmtZeroFormat); |
646 | } |
647 | } |
648 | } |
649 | |
650 | // ------------------------------------------------------------------ |
651 | // Format and Parse |
652 | |
653 | UnicodeString& |
654 | TimeZoneFormat::format(UTimeZoneFormatStyle style, const TimeZone& tz, UDate date, |
655 | UnicodeString& name, UTimeZoneFormatTimeType* timeType /* = NULL */) const { |
656 | if (timeType) { |
657 | *timeType = UTZFMT_TIME_TYPE_UNKNOWN; |
658 | } |
659 | |
660 | UBool noOffsetFormatFallback = FALSE; |
661 | |
662 | switch (style) { |
663 | case UTZFMT_STYLE_GENERIC_LOCATION: |
664 | formatGeneric(tz, UTZGNM_LOCATION, date, name); |
665 | break; |
666 | case UTZFMT_STYLE_GENERIC_LONG: |
667 | formatGeneric(tz, UTZGNM_LONG, date, name); |
668 | break; |
669 | case UTZFMT_STYLE_GENERIC_SHORT: |
670 | formatGeneric(tz, UTZGNM_SHORT, date, name); |
671 | break; |
672 | case UTZFMT_STYLE_SPECIFIC_LONG: |
673 | formatSpecific(tz, UTZNM_LONG_STANDARD, UTZNM_LONG_DAYLIGHT, date, name, timeType); |
674 | break; |
675 | case UTZFMT_STYLE_SPECIFIC_SHORT: |
676 | formatSpecific(tz, UTZNM_SHORT_STANDARD, UTZNM_SHORT_DAYLIGHT, date, name, timeType); |
677 | break; |
678 | |
679 | case UTZFMT_STYLE_ZONE_ID: |
680 | tz.getID(name); |
681 | noOffsetFormatFallback = TRUE; |
682 | break; |
683 | case UTZFMT_STYLE_ZONE_ID_SHORT: |
684 | { |
685 | const UChar* shortID = ZoneMeta::getShortID(tz); |
686 | if (shortID == NULL) { |
687 | shortID = UNKNOWN_SHORT_ZONE_ID; |
688 | } |
689 | name.setTo(shortID, -1); |
690 | } |
691 | noOffsetFormatFallback = TRUE; |
692 | break; |
693 | |
694 | case UTZFMT_STYLE_EXEMPLAR_LOCATION: |
695 | formatExemplarLocation(tz, name); |
696 | noOffsetFormatFallback = TRUE; |
697 | break; |
698 | |
699 | default: |
700 | // will be handled below |
701 | break; |
702 | } |
703 | |
704 | if (name.isEmpty() && !noOffsetFormatFallback) { |
705 | UErrorCode status = U_ZERO_ERROR; |
706 | int32_t rawOffset, dstOffset; |
707 | tz.getOffset(date, FALSE, rawOffset, dstOffset, status); |
708 | int32_t offset = rawOffset + dstOffset; |
709 | if (U_SUCCESS(status)) { |
710 | switch (style) { |
711 | case UTZFMT_STYLE_GENERIC_LOCATION: |
712 | case UTZFMT_STYLE_GENERIC_LONG: |
713 | case UTZFMT_STYLE_SPECIFIC_LONG: |
714 | case UTZFMT_STYLE_LOCALIZED_GMT: |
715 | formatOffsetLocalizedGMT(offset, name, status); |
716 | break; |
717 | |
718 | case UTZFMT_STYLE_GENERIC_SHORT: |
719 | case UTZFMT_STYLE_SPECIFIC_SHORT: |
720 | case UTZFMT_STYLE_LOCALIZED_GMT_SHORT: |
721 | formatOffsetShortLocalizedGMT(offset, name, status); |
722 | break; |
723 | |
724 | case UTZFMT_STYLE_ISO_BASIC_SHORT: |
725 | formatOffsetISO8601Basic(offset, TRUE, TRUE, TRUE, name, status); |
726 | break; |
727 | |
728 | case UTZFMT_STYLE_ISO_BASIC_LOCAL_SHORT: |
729 | formatOffsetISO8601Basic(offset, FALSE, TRUE, TRUE, name, status); |
730 | break; |
731 | |
732 | case UTZFMT_STYLE_ISO_BASIC_FIXED: |
733 | formatOffsetISO8601Basic(offset, TRUE, FALSE, TRUE, name, status); |
734 | break; |
735 | |
736 | case UTZFMT_STYLE_ISO_BASIC_LOCAL_FIXED: |
737 | formatOffsetISO8601Basic(offset, FALSE, FALSE, TRUE, name, status); |
738 | break; |
739 | |
740 | case UTZFMT_STYLE_ISO_EXTENDED_FIXED: |
741 | formatOffsetISO8601Extended(offset, TRUE, FALSE, TRUE, name, status); |
742 | break; |
743 | |
744 | case UTZFMT_STYLE_ISO_EXTENDED_LOCAL_FIXED: |
745 | formatOffsetISO8601Extended(offset, FALSE, FALSE, TRUE, name, status); |
746 | break; |
747 | |
748 | case UTZFMT_STYLE_ISO_BASIC_FULL: |
749 | formatOffsetISO8601Basic(offset, TRUE, FALSE, FALSE, name, status); |
750 | break; |
751 | |
752 | case UTZFMT_STYLE_ISO_BASIC_LOCAL_FULL: |
753 | formatOffsetISO8601Basic(offset, FALSE, FALSE, FALSE, name, status); |
754 | break; |
755 | |
756 | case UTZFMT_STYLE_ISO_EXTENDED_FULL: |
757 | formatOffsetISO8601Extended(offset, TRUE, FALSE, FALSE, name, status); |
758 | break; |
759 | |
760 | case UTZFMT_STYLE_ISO_EXTENDED_LOCAL_FULL: |
761 | formatOffsetISO8601Extended(offset, FALSE, FALSE, FALSE, name, status); |
762 | break; |
763 | |
764 | default: |
765 | // UTZFMT_STYLE_ZONE_ID, UTZFMT_STYLE_ZONE_ID_SHORT, UTZFMT_STYLE_EXEMPLAR_LOCATION |
766 | break; |
767 | } |
768 | |
769 | if (timeType) { |
770 | *timeType = (dstOffset != 0) ? UTZFMT_TIME_TYPE_DAYLIGHT : UTZFMT_TIME_TYPE_STANDARD; |
771 | } |
772 | } |
773 | } |
774 | |
775 | return name; |
776 | } |
777 | |
778 | UnicodeString& |
779 | TimeZoneFormat::format(const Formattable& obj, UnicodeString& appendTo, |
780 | FieldPosition& pos, UErrorCode& status) const { |
781 | if (U_FAILURE(status)) { |
782 | return appendTo; |
783 | } |
784 | UDate date = Calendar::getNow(); |
785 | if (obj.getType() == Formattable::kObject) { |
786 | const UObject* formatObj = obj.getObject(); |
787 | const TimeZone* tz = dynamic_cast<const TimeZone*>(formatObj); |
788 | if (tz == NULL) { |
789 | const Calendar* cal = dynamic_cast<const Calendar*>(formatObj); |
790 | if (cal != NULL) { |
791 | tz = &cal->getTimeZone(); |
792 | date = cal->getTime(status); |
793 | } |
794 | } |
795 | if (tz != NULL) { |
796 | int32_t rawOffset, dstOffset; |
797 | tz->getOffset(date, FALSE, rawOffset, dstOffset, status); |
798 | UChar buf[ZONE_NAME_U16_MAX]; |
799 | UnicodeString result(buf, 0, UPRV_LENGTHOF(buf)); |
800 | formatOffsetLocalizedGMT(rawOffset + dstOffset, result, status); |
801 | if (U_SUCCESS(status)) { |
802 | appendTo.append(result); |
803 | if (pos.getField() == UDAT_TIMEZONE_FIELD) { |
804 | pos.setBeginIndex(0); |
805 | pos.setEndIndex(result.length()); |
806 | } |
807 | } |
808 | } |
809 | } |
810 | return appendTo; |
811 | } |
812 | |
813 | TimeZone* |
814 | TimeZoneFormat::parse(UTimeZoneFormatStyle style, const UnicodeString& text, ParsePosition& pos, |
815 | UTimeZoneFormatTimeType* timeType /*= NULL*/) const { |
816 | return parse(style, text, pos, getDefaultParseOptions(), timeType); |
817 | } |
818 | |
819 | TimeZone* |
820 | TimeZoneFormat::parse(UTimeZoneFormatStyle style, const UnicodeString& text, ParsePosition& pos, |
821 | int32_t parseOptions, UTimeZoneFormatTimeType* timeType /* = NULL */) const { |
822 | if (timeType) { |
823 | *timeType = UTZFMT_TIME_TYPE_UNKNOWN; |
824 | } |
825 | |
826 | int32_t startIdx = pos.getIndex(); |
827 | int32_t maxPos = text.length(); |
828 | int32_t offset; |
829 | |
830 | // Styles using localized GMT format as fallback |
831 | UBool fallbackLocalizedGMT = |
832 | (style == UTZFMT_STYLE_SPECIFIC_LONG || style == UTZFMT_STYLE_GENERIC_LONG || style == UTZFMT_STYLE_GENERIC_LOCATION); |
833 | UBool fallbackShortLocalizedGMT = |
834 | (style == UTZFMT_STYLE_SPECIFIC_SHORT || style == UTZFMT_STYLE_GENERIC_SHORT); |
835 | |
836 | int32_t evaluated = 0; // bit flags representing already evaluated styles |
837 | ParsePosition tmpPos(startIdx); |
838 | |
839 | int32_t parsedOffset = UNKNOWN_OFFSET; // stores successfully parsed offset for later use |
840 | int32_t parsedPos = -1; // stores successfully parsed offset position for later use |
841 | |
842 | // Try localized GMT format first if necessary |
843 | if (fallbackLocalizedGMT || fallbackShortLocalizedGMT) { |
844 | UBool hasDigitOffset = FALSE; |
845 | offset = parseOffsetLocalizedGMT(text, tmpPos, fallbackShortLocalizedGMT, &hasDigitOffset); |
846 | if (tmpPos.getErrorIndex() == -1) { |
847 | // Even when the input text was successfully parsed as a localized GMT format text, |
848 | // we may still need to evaluate the specified style if - |
849 | // 1) GMT zero format was used, and |
850 | // 2) The input text was not completely processed |
851 | if (tmpPos.getIndex() == maxPos || hasDigitOffset) { |
852 | pos.setIndex(tmpPos.getIndex()); |
853 | return createTimeZoneForOffset(offset); |
854 | } |
855 | parsedOffset = offset; |
856 | parsedPos = tmpPos.getIndex(); |
857 | } |
858 | // Note: For now, no distinction between long/short localized GMT format in the parser. |
859 | // This might be changed in future. |
860 | // evaluated |= (fallbackLocalizedGMT ? STYLE_PARSE_FLAGS[UTZFMT_STYLE_LOCALIZED_GMT] : STYLE_PARSE_FLAGS[UTZFMT_STYLE_LOCALIZED_GMT_SHORT]); |
861 | evaluated |= STYLE_PARSE_FLAGS[UTZFMT_STYLE_LOCALIZED_GMT] | STYLE_PARSE_FLAGS[UTZFMT_STYLE_LOCALIZED_GMT_SHORT]; |
862 | } |
863 | |
864 | UErrorCode status = U_ZERO_ERROR; |
865 | UChar tzIDBuf[32]; |
866 | UnicodeString tzID(tzIDBuf, 0, UPRV_LENGTHOF(tzIDBuf)); |
867 | |
868 | UBool parseTZDBAbbrev = ((parseOptions & UTZFMT_PARSE_OPTION_TZ_DATABASE_ABBREVIATIONS) != 0); |
869 | |
870 | // Try the specified style |
871 | switch (style) { |
872 | case UTZFMT_STYLE_LOCALIZED_GMT: |
873 | { |
874 | tmpPos.setIndex(startIdx); |
875 | tmpPos.setErrorIndex(-1); |
876 | |
877 | offset = parseOffsetLocalizedGMT(text, tmpPos); |
878 | if (tmpPos.getErrorIndex() == -1) { |
879 | pos.setIndex(tmpPos.getIndex()); |
880 | return createTimeZoneForOffset(offset); |
881 | } |
882 | |
883 | // Note: For now, no distinction between long/short localized GMT format in the parser. |
884 | // This might be changed in future. |
885 | evaluated |= STYLE_PARSE_FLAGS[UTZFMT_STYLE_LOCALIZED_GMT_SHORT]; |
886 | |
887 | break; |
888 | } |
889 | case UTZFMT_STYLE_LOCALIZED_GMT_SHORT: |
890 | { |
891 | tmpPos.setIndex(startIdx); |
892 | tmpPos.setErrorIndex(-1); |
893 | |
894 | offset = parseOffsetShortLocalizedGMT(text, tmpPos); |
895 | if (tmpPos.getErrorIndex() == -1) { |
896 | pos.setIndex(tmpPos.getIndex()); |
897 | return createTimeZoneForOffset(offset); |
898 | } |
899 | |
900 | // Note: For now, no distinction between long/short localized GMT format in the parser. |
901 | // This might be changed in future. |
902 | evaluated |= STYLE_PARSE_FLAGS[UTZFMT_STYLE_LOCALIZED_GMT]; |
903 | |
904 | break; |
905 | } |
906 | case UTZFMT_STYLE_ISO_BASIC_SHORT: |
907 | case UTZFMT_STYLE_ISO_BASIC_FIXED: |
908 | case UTZFMT_STYLE_ISO_BASIC_FULL: |
909 | case UTZFMT_STYLE_ISO_EXTENDED_FIXED: |
910 | case UTZFMT_STYLE_ISO_EXTENDED_FULL: |
911 | { |
912 | tmpPos.setIndex(startIdx); |
913 | tmpPos.setErrorIndex(-1); |
914 | |
915 | offset = parseOffsetISO8601(text, tmpPos); |
916 | if (tmpPos.getErrorIndex() == -1) { |
917 | pos.setIndex(tmpPos.getIndex()); |
918 | return createTimeZoneForOffset(offset); |
919 | } |
920 | |
921 | break; |
922 | } |
923 | |
924 | case UTZFMT_STYLE_ISO_BASIC_LOCAL_SHORT: |
925 | case UTZFMT_STYLE_ISO_BASIC_LOCAL_FIXED: |
926 | case UTZFMT_STYLE_ISO_BASIC_LOCAL_FULL: |
927 | case UTZFMT_STYLE_ISO_EXTENDED_LOCAL_FIXED: |
928 | case UTZFMT_STYLE_ISO_EXTENDED_LOCAL_FULL: |
929 | { |
930 | tmpPos.setIndex(startIdx); |
931 | tmpPos.setErrorIndex(-1); |
932 | |
933 | // Exclude the case of UTC Indicator "Z" here |
934 | UBool hasDigitOffset = FALSE; |
935 | offset = parseOffsetISO8601(text, tmpPos, FALSE, &hasDigitOffset); |
936 | if (tmpPos.getErrorIndex() == -1 && hasDigitOffset) { |
937 | pos.setIndex(tmpPos.getIndex()); |
938 | return createTimeZoneForOffset(offset); |
939 | } |
940 | |
941 | break; |
942 | } |
943 | |
944 | case UTZFMT_STYLE_SPECIFIC_LONG: |
945 | case UTZFMT_STYLE_SPECIFIC_SHORT: |
946 | { |
947 | // Specific styles |
948 | int32_t nameTypes = 0; |
949 | if (style == UTZFMT_STYLE_SPECIFIC_LONG) { |
950 | nameTypes = (UTZNM_LONG_STANDARD | UTZNM_LONG_DAYLIGHT); |
951 | } else { |
952 | U_ASSERT(style == UTZFMT_STYLE_SPECIFIC_SHORT); |
953 | nameTypes = (UTZNM_SHORT_STANDARD | UTZNM_SHORT_DAYLIGHT); |
954 | } |
955 | LocalPointer<TimeZoneNames::MatchInfoCollection> specificMatches(fTimeZoneNames->find(text, startIdx, nameTypes, status)); |
956 | if (U_FAILURE(status)) { |
957 | pos.setErrorIndex(startIdx); |
958 | return NULL; |
959 | } |
960 | if (!specificMatches.isNull()) { |
961 | int32_t matchIdx = -1; |
962 | int32_t matchPos = -1; |
963 | for (int32_t i = 0; i < specificMatches->size(); i++) { |
964 | matchPos = startIdx + specificMatches->getMatchLengthAt(i); |
965 | if (matchPos > parsedPos) { |
966 | matchIdx = i; |
967 | parsedPos = matchPos; |
968 | } |
969 | } |
970 | if (matchIdx >= 0) { |
971 | if (timeType) { |
972 | *timeType = getTimeType(specificMatches->getNameTypeAt(matchIdx)); |
973 | } |
974 | pos.setIndex(matchPos); |
975 | getTimeZoneID(specificMatches.getAlias(), matchIdx, tzID); |
976 | U_ASSERT(!tzID.isEmpty()); |
977 | return TimeZone::createTimeZone(tzID); |
978 | } |
979 | } |
980 | |
981 | if (parseTZDBAbbrev && style == UTZFMT_STYLE_SPECIFIC_SHORT) { |
982 | U_ASSERT((nameTypes & UTZNM_SHORT_STANDARD) != 0); |
983 | U_ASSERT((nameTypes & UTZNM_SHORT_DAYLIGHT) != 0); |
984 | |
985 | const TZDBTimeZoneNames *tzdbTimeZoneNames = getTZDBTimeZoneNames(status); |
986 | if (U_SUCCESS(status)) { |
987 | LocalPointer<TimeZoneNames::MatchInfoCollection> tzdbNameMatches( |
988 | tzdbTimeZoneNames->find(text, startIdx, nameTypes, status)); |
989 | if (U_FAILURE(status)) { |
990 | pos.setErrorIndex(startIdx); |
991 | return NULL; |
992 | } |
993 | if (!tzdbNameMatches.isNull()) { |
994 | int32_t matchIdx = -1; |
995 | int32_t matchPos = -1; |
996 | for (int32_t i = 0; i < tzdbNameMatches->size(); i++) { |
997 | matchPos = startIdx + tzdbNameMatches->getMatchLengthAt(i); |
998 | if (matchPos > parsedPos) { |
999 | matchIdx = i; |
1000 | parsedPos = matchPos; |
1001 | } |
1002 | } |
1003 | if (matchIdx >= 0) { |
1004 | if (timeType) { |
1005 | *timeType = getTimeType(tzdbNameMatches->getNameTypeAt(matchIdx)); |
1006 | } |
1007 | pos.setIndex(matchPos); |
1008 | getTimeZoneID(tzdbNameMatches.getAlias(), matchIdx, tzID); |
1009 | U_ASSERT(!tzID.isEmpty()); |
1010 | return TimeZone::createTimeZone(tzID); |
1011 | } |
1012 | } |
1013 | } |
1014 | } |
1015 | break; |
1016 | } |
1017 | case UTZFMT_STYLE_GENERIC_LONG: |
1018 | case UTZFMT_STYLE_GENERIC_SHORT: |
1019 | case UTZFMT_STYLE_GENERIC_LOCATION: |
1020 | { |
1021 | int32_t genericNameTypes = 0; |
1022 | switch (style) { |
1023 | case UTZFMT_STYLE_GENERIC_LOCATION: |
1024 | genericNameTypes = UTZGNM_LOCATION; |
1025 | break; |
1026 | |
1027 | case UTZFMT_STYLE_GENERIC_LONG: |
1028 | genericNameTypes = UTZGNM_LONG | UTZGNM_LOCATION; |
1029 | break; |
1030 | |
1031 | case UTZFMT_STYLE_GENERIC_SHORT: |
1032 | genericNameTypes = UTZGNM_SHORT | UTZGNM_LOCATION; |
1033 | break; |
1034 | |
1035 | default: |
1036 | UPRV_UNREACHABLE; |
1037 | } |
1038 | |
1039 | int32_t len = 0; |
1040 | UTimeZoneFormatTimeType tt = UTZFMT_TIME_TYPE_UNKNOWN; |
1041 | const TimeZoneGenericNames *gnames = getTimeZoneGenericNames(status); |
1042 | if (U_SUCCESS(status)) { |
1043 | len = gnames->findBestMatch(text, startIdx, genericNameTypes, tzID, tt, status); |
1044 | } |
1045 | if (U_FAILURE(status)) { |
1046 | pos.setErrorIndex(startIdx); |
1047 | return NULL; |
1048 | } |
1049 | if (len > 0) { |
1050 | // Found a match |
1051 | if (timeType) { |
1052 | *timeType = tt; |
1053 | } |
1054 | pos.setIndex(startIdx + len); |
1055 | U_ASSERT(!tzID.isEmpty()); |
1056 | return TimeZone::createTimeZone(tzID); |
1057 | } |
1058 | |
1059 | break; |
1060 | } |
1061 | case UTZFMT_STYLE_ZONE_ID: |
1062 | { |
1063 | tmpPos.setIndex(startIdx); |
1064 | tmpPos.setErrorIndex(-1); |
1065 | |
1066 | parseZoneID(text, tmpPos, tzID); |
1067 | if (tmpPos.getErrorIndex() == -1) { |
1068 | pos.setIndex(tmpPos.getIndex()); |
1069 | return TimeZone::createTimeZone(tzID); |
1070 | } |
1071 | break; |
1072 | } |
1073 | case UTZFMT_STYLE_ZONE_ID_SHORT: |
1074 | { |
1075 | tmpPos.setIndex(startIdx); |
1076 | tmpPos.setErrorIndex(-1); |
1077 | |
1078 | parseShortZoneID(text, tmpPos, tzID); |
1079 | if (tmpPos.getErrorIndex() == -1) { |
1080 | pos.setIndex(tmpPos.getIndex()); |
1081 | return TimeZone::createTimeZone(tzID); |
1082 | } |
1083 | break; |
1084 | } |
1085 | case UTZFMT_STYLE_EXEMPLAR_LOCATION: |
1086 | { |
1087 | tmpPos.setIndex(startIdx); |
1088 | tmpPos.setErrorIndex(-1); |
1089 | |
1090 | parseExemplarLocation(text, tmpPos, tzID); |
1091 | if (tmpPos.getErrorIndex() == -1) { |
1092 | pos.setIndex(tmpPos.getIndex()); |
1093 | return TimeZone::createTimeZone(tzID); |
1094 | } |
1095 | break; |
1096 | } |
1097 | } |
1098 | evaluated |= STYLE_PARSE_FLAGS[style]; |
1099 | |
1100 | |
1101 | if (parsedPos > startIdx) { |
1102 | // When the specified style is one of SPECIFIC_XXX or GENERIC_XXX, we tried to parse the input |
1103 | // as localized GMT format earlier. If parsedOffset is positive, it means it was successfully |
1104 | // parsed as localized GMT format, but offset digits were not detected (more specifically, GMT |
1105 | // zero format). Then, it tried to find a match within the set of display names, but could not |
1106 | // find a match. At this point, we can safely assume the input text contains the localized |
1107 | // GMT format. |
1108 | U_ASSERT(parsedOffset != UNKNOWN_OFFSET); |
1109 | pos.setIndex(parsedPos); |
1110 | return createTimeZoneForOffset(parsedOffset); |
1111 | } |
1112 | |
1113 | // Failed to parse the input text as the time zone format in the specified style. |
1114 | // Check the longest match among other styles below. |
1115 | UChar parsedIDBuf[32]; |
1116 | UnicodeString parsedID(parsedIDBuf, 0, UPRV_LENGTHOF(parsedIDBuf)); |
1117 | UTimeZoneFormatTimeType parsedTimeType = UTZFMT_TIME_TYPE_UNKNOWN; |
1118 | |
1119 | U_ASSERT(parsedPos < 0); |
1120 | U_ASSERT(parsedOffset == UNKNOWN_OFFSET); |
1121 | |
1122 | // ISO 8601 |
1123 | if (parsedPos < maxPos && |
1124 | ((evaluated & ISO_Z_STYLE_FLAG) == 0 || (evaluated & ISO_LOCAL_STYLE_FLAG) == 0)) { |
1125 | tmpPos.setIndex(startIdx); |
1126 | tmpPos.setErrorIndex(-1); |
1127 | |
1128 | UBool hasDigitOffset = FALSE; |
1129 | offset = parseOffsetISO8601(text, tmpPos, FALSE, &hasDigitOffset); |
1130 | if (tmpPos.getErrorIndex() == -1) { |
1131 | if (tmpPos.getIndex() == maxPos || hasDigitOffset) { |
1132 | pos.setIndex(tmpPos.getIndex()); |
1133 | return createTimeZoneForOffset(offset); |
1134 | } |
1135 | // Note: When ISO 8601 format contains offset digits, it should not |
1136 | // collide with other formats. However, ISO 8601 UTC format "Z" (single letter) |
1137 | // may collide with other names. In this case, we need to evaluate other names. |
1138 | if (parsedPos < tmpPos.getIndex()) { |
1139 | parsedOffset = offset; |
1140 | parsedID.setToBogus(); |
1141 | parsedTimeType = UTZFMT_TIME_TYPE_UNKNOWN; |
1142 | parsedPos = tmpPos.getIndex(); |
1143 | U_ASSERT(parsedPos == startIdx + 1); // only when "Z" is used |
1144 | } |
1145 | } |
1146 | } |
1147 | |
1148 | // Localized GMT format |
1149 | if (parsedPos < maxPos && |
1150 | (evaluated & STYLE_PARSE_FLAGS[UTZFMT_STYLE_LOCALIZED_GMT]) == 0) { |
1151 | tmpPos.setIndex(startIdx); |
1152 | tmpPos.setErrorIndex(-1); |
1153 | |
1154 | UBool hasDigitOffset = FALSE; |
1155 | offset = parseOffsetLocalizedGMT(text, tmpPos, FALSE, &hasDigitOffset); |
1156 | if (tmpPos.getErrorIndex() == -1) { |
1157 | if (tmpPos.getIndex() == maxPos || hasDigitOffset) { |
1158 | pos.setIndex(tmpPos.getIndex()); |
1159 | return createTimeZoneForOffset(offset); |
1160 | } |
1161 | // Evaluate other names - see the comment earlier in this method. |
1162 | if (parsedPos < tmpPos.getIndex()) { |
1163 | parsedOffset = offset; |
1164 | parsedID.setToBogus(); |
1165 | parsedTimeType = UTZFMT_TIME_TYPE_UNKNOWN; |
1166 | parsedPos = tmpPos.getIndex(); |
1167 | } |
1168 | } |
1169 | } |
1170 | |
1171 | if (parsedPos < maxPos && |
1172 | (evaluated & STYLE_PARSE_FLAGS[UTZFMT_STYLE_LOCALIZED_GMT_SHORT]) == 0) { |
1173 | tmpPos.setIndex(startIdx); |
1174 | tmpPos.setErrorIndex(-1); |
1175 | |
1176 | UBool hasDigitOffset = FALSE; |
1177 | offset = parseOffsetLocalizedGMT(text, tmpPos, TRUE, &hasDigitOffset); |
1178 | if (tmpPos.getErrorIndex() == -1) { |
1179 | if (tmpPos.getIndex() == maxPos || hasDigitOffset) { |
1180 | pos.setIndex(tmpPos.getIndex()); |
1181 | return createTimeZoneForOffset(offset); |
1182 | } |
1183 | // Evaluate other names - see the comment earlier in this method. |
1184 | if (parsedPos < tmpPos.getIndex()) { |
1185 | parsedOffset = offset; |
1186 | parsedID.setToBogus(); |
1187 | parsedTimeType = UTZFMT_TIME_TYPE_UNKNOWN; |
1188 | parsedPos = tmpPos.getIndex(); |
1189 | } |
1190 | } |
1191 | } |
1192 | |
1193 | // When ParseOption.ALL_STYLES is available, we also try to look all possible display names and IDs. |
1194 | // For example, when style is GENERIC_LONG, "EST" (SPECIFIC_SHORT) is never |
1195 | // used for America/New_York. With parseAllStyles true, this code parses "EST" |
1196 | // as America/New_York. |
1197 | |
1198 | // Note: Adding all possible names into the trie used by the implementation is quite heavy operation, |
1199 | // which we want to avoid normally (note that we cache the trie, so this is applicable to the |
1200 | // first time only as long as the cache does not expire). |
1201 | |
1202 | if (parseOptions & UTZFMT_PARSE_OPTION_ALL_STYLES) { |
1203 | // Try all specific names and exemplar location names |
1204 | if (parsedPos < maxPos) { |
1205 | LocalPointer<TimeZoneNames::MatchInfoCollection> specificMatches(fTimeZoneNames->find(text, startIdx, ALL_SIMPLE_NAME_TYPES, status)); |
1206 | if (U_FAILURE(status)) { |
1207 | pos.setErrorIndex(startIdx); |
1208 | return NULL; |
1209 | } |
1210 | int32_t specificMatchIdx = -1; |
1211 | int32_t matchPos = -1; |
1212 | if (!specificMatches.isNull()) { |
1213 | for (int32_t i = 0; i < specificMatches->size(); i++) { |
1214 | if (startIdx + specificMatches->getMatchLengthAt(i) > matchPos) { |
1215 | specificMatchIdx = i; |
1216 | matchPos = startIdx + specificMatches->getMatchLengthAt(i); |
1217 | } |
1218 | } |
1219 | } |
1220 | if (parsedPos < matchPos) { |
1221 | U_ASSERT(specificMatchIdx >= 0); |
1222 | parsedPos = matchPos; |
1223 | getTimeZoneID(specificMatches.getAlias(), specificMatchIdx, parsedID); |
1224 | parsedTimeType = getTimeType(specificMatches->getNameTypeAt(specificMatchIdx)); |
1225 | parsedOffset = UNKNOWN_OFFSET; |
1226 | } |
1227 | } |
1228 | if (parseTZDBAbbrev && parsedPos < maxPos && (evaluated & STYLE_PARSE_FLAGS[UTZFMT_STYLE_SPECIFIC_SHORT]) == 0) { |
1229 | const TZDBTimeZoneNames *tzdbTimeZoneNames = getTZDBTimeZoneNames(status); |
1230 | if (U_SUCCESS(status)) { |
1231 | LocalPointer<TimeZoneNames::MatchInfoCollection> tzdbNameMatches( |
1232 | tzdbTimeZoneNames->find(text, startIdx, ALL_SIMPLE_NAME_TYPES, status)); |
1233 | if (U_FAILURE(status)) { |
1234 | pos.setErrorIndex(startIdx); |
1235 | return NULL; |
1236 | } |
1237 | int32_t tzdbNameMatchIdx = -1; |
1238 | int32_t matchPos = -1; |
1239 | if (!tzdbNameMatches.isNull()) { |
1240 | for (int32_t i = 0; i < tzdbNameMatches->size(); i++) { |
1241 | if (startIdx + tzdbNameMatches->getMatchLengthAt(i) > matchPos) { |
1242 | tzdbNameMatchIdx = i; |
1243 | matchPos = startIdx + tzdbNameMatches->getMatchLengthAt(i); |
1244 | } |
1245 | } |
1246 | } |
1247 | if (parsedPos < matchPos) { |
1248 | U_ASSERT(tzdbNameMatchIdx >= 0); |
1249 | parsedPos = matchPos; |
1250 | getTimeZoneID(tzdbNameMatches.getAlias(), tzdbNameMatchIdx, parsedID); |
1251 | parsedTimeType = getTimeType(tzdbNameMatches->getNameTypeAt(tzdbNameMatchIdx)); |
1252 | parsedOffset = UNKNOWN_OFFSET; |
1253 | } |
1254 | } |
1255 | } |
1256 | // Try generic names |
1257 | if (parsedPos < maxPos) { |
1258 | int32_t genMatchLen = -1; |
1259 | UTimeZoneFormatTimeType tt = UTZFMT_TIME_TYPE_UNKNOWN; |
1260 | |
1261 | const TimeZoneGenericNames *gnames = getTimeZoneGenericNames(status); |
1262 | if (U_SUCCESS(status)) { |
1263 | genMatchLen = gnames->findBestMatch(text, startIdx, ALL_GENERIC_NAME_TYPES, tzID, tt, status); |
1264 | } |
1265 | if (U_FAILURE(status)) { |
1266 | pos.setErrorIndex(startIdx); |
1267 | return NULL; |
1268 | } |
1269 | |
1270 | if (genMatchLen > 0 && parsedPos < startIdx + genMatchLen) { |
1271 | parsedPos = startIdx + genMatchLen; |
1272 | parsedID.setTo(tzID); |
1273 | parsedTimeType = tt; |
1274 | parsedOffset = UNKNOWN_OFFSET; |
1275 | } |
1276 | } |
1277 | |
1278 | // Try time zone ID |
1279 | if (parsedPos < maxPos && (evaluated & STYLE_PARSE_FLAGS[UTZFMT_STYLE_ZONE_ID]) == 0) { |
1280 | tmpPos.setIndex(startIdx); |
1281 | tmpPos.setErrorIndex(-1); |
1282 | |
1283 | parseZoneID(text, tmpPos, tzID); |
1284 | if (tmpPos.getErrorIndex() == -1 && parsedPos < tmpPos.getIndex()) { |
1285 | parsedPos = tmpPos.getIndex(); |
1286 | parsedID.setTo(tzID); |
1287 | parsedTimeType = UTZFMT_TIME_TYPE_UNKNOWN; |
1288 | parsedOffset = UNKNOWN_OFFSET; |
1289 | } |
1290 | } |
1291 | // Try short time zone ID |
1292 | if (parsedPos < maxPos && (evaluated & STYLE_PARSE_FLAGS[UTZFMT_STYLE_ZONE_ID]) == 0) { |
1293 | tmpPos.setIndex(startIdx); |
1294 | tmpPos.setErrorIndex(-1); |
1295 | |
1296 | parseShortZoneID(text, tmpPos, tzID); |
1297 | if (tmpPos.getErrorIndex() == -1 && parsedPos < tmpPos.getIndex()) { |
1298 | parsedPos = tmpPos.getIndex(); |
1299 | parsedID.setTo(tzID); |
1300 | parsedTimeType = UTZFMT_TIME_TYPE_UNKNOWN; |
1301 | parsedOffset = UNKNOWN_OFFSET; |
1302 | } |
1303 | } |
1304 | } |
1305 | |
1306 | if (parsedPos > startIdx) { |
1307 | // Parsed successfully |
1308 | TimeZone* parsedTZ; |
1309 | if (parsedID.length() > 0) { |
1310 | parsedTZ = TimeZone::createTimeZone(parsedID); |
1311 | } else { |
1312 | U_ASSERT(parsedOffset != UNKNOWN_OFFSET); |
1313 | parsedTZ = createTimeZoneForOffset(parsedOffset); |
1314 | } |
1315 | if (timeType) { |
1316 | *timeType = parsedTimeType; |
1317 | } |
1318 | pos.setIndex(parsedPos); |
1319 | return parsedTZ; |
1320 | } |
1321 | |
1322 | pos.setErrorIndex(startIdx); |
1323 | return NULL; |
1324 | } |
1325 | |
1326 | void |
1327 | TimeZoneFormat::parseObject(const UnicodeString& source, Formattable& result, |
1328 | ParsePosition& parse_pos) const { |
1329 | result.adoptObject(parse(UTZFMT_STYLE_GENERIC_LOCATION, source, parse_pos, UTZFMT_PARSE_OPTION_ALL_STYLES)); |
1330 | } |
1331 | |
1332 | |
1333 | // ------------------------------------------------------------------ |
1334 | // Private zone name format/parse implementation |
1335 | |
1336 | UnicodeString& |
1337 | TimeZoneFormat::formatGeneric(const TimeZone& tz, int32_t genType, UDate date, UnicodeString& name) const { |
1338 | UErrorCode status = U_ZERO_ERROR; |
1339 | const TimeZoneGenericNames* gnames = getTimeZoneGenericNames(status); |
1340 | if (U_FAILURE(status)) { |
1341 | name.setToBogus(); |
1342 | return name; |
1343 | } |
1344 | |
1345 | if (genType == UTZGNM_LOCATION) { |
1346 | const UChar* canonicalID = ZoneMeta::getCanonicalCLDRID(tz); |
1347 | if (canonicalID == NULL) { |
1348 | name.setToBogus(); |
1349 | return name; |
1350 | } |
1351 | return gnames->getGenericLocationName(UnicodeString(TRUE, canonicalID, -1), name); |
1352 | } |
1353 | return gnames->getDisplayName(tz, (UTimeZoneGenericNameType)genType, date, name); |
1354 | } |
1355 | |
1356 | UnicodeString& |
1357 | TimeZoneFormat::formatSpecific(const TimeZone& tz, UTimeZoneNameType stdType, UTimeZoneNameType dstType, |
1358 | UDate date, UnicodeString& name, UTimeZoneFormatTimeType *timeType) const { |
1359 | if (fTimeZoneNames == NULL) { |
1360 | name.setToBogus(); |
1361 | return name; |
1362 | } |
1363 | |
1364 | UErrorCode status = U_ZERO_ERROR; |
1365 | UBool isDaylight = tz.inDaylightTime(date, status); |
1366 | const UChar* canonicalID = ZoneMeta::getCanonicalCLDRID(tz); |
1367 | |
1368 | if (U_FAILURE(status) || canonicalID == NULL) { |
1369 | name.setToBogus(); |
1370 | return name; |
1371 | } |
1372 | |
1373 | if (isDaylight) { |
1374 | fTimeZoneNames->getDisplayName(UnicodeString(TRUE, canonicalID, -1), dstType, date, name); |
1375 | } else { |
1376 | fTimeZoneNames->getDisplayName(UnicodeString(TRUE, canonicalID, -1), stdType, date, name); |
1377 | } |
1378 | |
1379 | if (timeType && !name.isEmpty()) { |
1380 | *timeType = isDaylight ? UTZFMT_TIME_TYPE_DAYLIGHT : UTZFMT_TIME_TYPE_STANDARD; |
1381 | } |
1382 | return name; |
1383 | } |
1384 | |
1385 | const TimeZoneGenericNames* |
1386 | TimeZoneFormat::getTimeZoneGenericNames(UErrorCode& status) const { |
1387 | if (U_FAILURE(status)) { |
1388 | return NULL; |
1389 | } |
1390 | |
1391 | umtx_lock(&gLock); |
1392 | if (fTimeZoneGenericNames == NULL) { |
1393 | TimeZoneFormat *nonConstThis = const_cast<TimeZoneFormat *>(this); |
1394 | nonConstThis->fTimeZoneGenericNames = TimeZoneGenericNames::createInstance(fLocale, status); |
1395 | } |
1396 | umtx_unlock(&gLock); |
1397 | |
1398 | return fTimeZoneGenericNames; |
1399 | } |
1400 | |
1401 | const TZDBTimeZoneNames* |
1402 | TimeZoneFormat::getTZDBTimeZoneNames(UErrorCode& status) const { |
1403 | if (U_FAILURE(status)) { |
1404 | return NULL; |
1405 | } |
1406 | |
1407 | umtx_lock(&gLock); |
1408 | if (fTZDBTimeZoneNames == NULL) { |
1409 | TZDBTimeZoneNames *tzdbNames = new TZDBTimeZoneNames(fLocale); |
1410 | if (tzdbNames == NULL) { |
1411 | status = U_MEMORY_ALLOCATION_ERROR; |
1412 | } else { |
1413 | TimeZoneFormat *nonConstThis = const_cast<TimeZoneFormat *>(this); |
1414 | nonConstThis->fTZDBTimeZoneNames = tzdbNames; |
1415 | } |
1416 | } |
1417 | umtx_unlock(&gLock); |
1418 | |
1419 | return fTZDBTimeZoneNames; |
1420 | } |
1421 | |
1422 | UnicodeString& |
1423 | TimeZoneFormat::formatExemplarLocation(const TimeZone& tz, UnicodeString& name) const { |
1424 | UChar locationBuf[ZONE_NAME_U16_MAX]; |
1425 | UnicodeString location(locationBuf, 0, UPRV_LENGTHOF(locationBuf)); |
1426 | const UChar* canonicalID = ZoneMeta::getCanonicalCLDRID(tz); |
1427 | |
1428 | if (canonicalID) { |
1429 | fTimeZoneNames->getExemplarLocationName(UnicodeString(TRUE, canonicalID, -1), location); |
1430 | } |
1431 | if (location.length() > 0) { |
1432 | name.setTo(location); |
1433 | } else { |
1434 | // Use "unknown" location |
1435 | fTimeZoneNames->getExemplarLocationName(UnicodeString(TRUE, UNKNOWN_ZONE_ID, -1), location); |
1436 | if (location.length() > 0) { |
1437 | name.setTo(location); |
1438 | } else { |
1439 | // last resort |
1440 | name.setTo(UNKNOWN_LOCATION, -1); |
1441 | } |
1442 | } |
1443 | return name; |
1444 | } |
1445 | |
1446 | |
1447 | // ------------------------------------------------------------------ |
1448 | // Zone offset format and parse |
1449 | |
1450 | UnicodeString& |
1451 | TimeZoneFormat::formatOffsetISO8601Basic(int32_t offset, UBool useUtcIndicator, UBool isShort, UBool ignoreSeconds, |
1452 | UnicodeString& result, UErrorCode& status) const { |
1453 | return formatOffsetISO8601(offset, TRUE, useUtcIndicator, isShort, ignoreSeconds, result, status); |
1454 | } |
1455 | |
1456 | UnicodeString& |
1457 | TimeZoneFormat::formatOffsetISO8601Extended(int32_t offset, UBool useUtcIndicator, UBool isShort, UBool ignoreSeconds, |
1458 | UnicodeString& result, UErrorCode& status) const { |
1459 | return formatOffsetISO8601(offset, FALSE, useUtcIndicator, isShort, ignoreSeconds, result, status); |
1460 | } |
1461 | |
1462 | UnicodeString& |
1463 | TimeZoneFormat::formatOffsetLocalizedGMT(int32_t offset, UnicodeString& result, UErrorCode& status) const { |
1464 | return formatOffsetLocalizedGMT(offset, FALSE, result, status); |
1465 | } |
1466 | |
1467 | UnicodeString& |
1468 | TimeZoneFormat::formatOffsetShortLocalizedGMT(int32_t offset, UnicodeString& result, UErrorCode& status) const { |
1469 | return formatOffsetLocalizedGMT(offset, TRUE, result, status); |
1470 | } |
1471 | |
1472 | int32_t |
1473 | TimeZoneFormat::parseOffsetISO8601(const UnicodeString& text, ParsePosition& pos) const { |
1474 | return parseOffsetISO8601(text, pos, FALSE); |
1475 | } |
1476 | |
1477 | int32_t |
1478 | TimeZoneFormat::parseOffsetLocalizedGMT(const UnicodeString& text, ParsePosition& pos) const { |
1479 | return parseOffsetLocalizedGMT(text, pos, FALSE, NULL); |
1480 | } |
1481 | |
1482 | int32_t |
1483 | TimeZoneFormat::parseOffsetShortLocalizedGMT(const UnicodeString& text, ParsePosition& pos) const { |
1484 | return parseOffsetLocalizedGMT(text, pos, TRUE, NULL); |
1485 | } |
1486 | |
1487 | // ------------------------------------------------------------------ |
1488 | // Private zone offset format/parse implementation |
1489 | |
1490 | UnicodeString& |
1491 | TimeZoneFormat::formatOffsetISO8601(int32_t offset, UBool isBasic, UBool useUtcIndicator, |
1492 | UBool isShort, UBool ignoreSeconds, UnicodeString& result, UErrorCode& status) const { |
1493 | if (U_FAILURE(status)) { |
1494 | result.setToBogus(); |
1495 | return result; |
1496 | } |
1497 | int32_t absOffset = offset < 0 ? -offset : offset; |
1498 | if (useUtcIndicator && (absOffset < MILLIS_PER_SECOND || (ignoreSeconds && absOffset < MILLIS_PER_MINUTE))) { |
1499 | result.setTo(ISO8601_UTC); |
1500 | return result; |
1501 | } |
1502 | |
1503 | OffsetFields minFields = isShort ? FIELDS_H : FIELDS_HM; |
1504 | OffsetFields maxFields = ignoreSeconds ? FIELDS_HM : FIELDS_HMS; |
1505 | UChar sep = isBasic ? 0 : ISO8601_SEP; |
1506 | |
1507 | // Note: FIELDS_HMS as maxFields is a CLDR/ICU extension. ISO 8601 specification does |
1508 | // not support seconds field. |
1509 | |
1510 | if (absOffset >= MAX_OFFSET) { |
1511 | result.setToBogus(); |
1512 | status = U_ILLEGAL_ARGUMENT_ERROR; |
1513 | return result; |
1514 | } |
1515 | |
1516 | int fields[3]; |
1517 | fields[0] = absOffset / MILLIS_PER_HOUR; |
1518 | absOffset = absOffset % MILLIS_PER_HOUR; |
1519 | fields[1] = absOffset / MILLIS_PER_MINUTE; |
1520 | absOffset = absOffset % MILLIS_PER_MINUTE; |
1521 | fields[2] = absOffset / MILLIS_PER_SECOND; |
1522 | |
1523 | U_ASSERT(fields[0] >= 0 && fields[0] <= MAX_OFFSET_HOUR); |
1524 | U_ASSERT(fields[1] >= 0 && fields[1] <= MAX_OFFSET_MINUTE); |
1525 | U_ASSERT(fields[2] >= 0 && fields[2] <= MAX_OFFSET_SECOND); |
1526 | |
1527 | int32_t lastIdx = maxFields; |
1528 | while (lastIdx > minFields) { |
1529 | if (fields[lastIdx] != 0) { |
1530 | break; |
1531 | } |
1532 | lastIdx--; |
1533 | } |
1534 | |
1535 | UChar sign = PLUS; |
1536 | if (offset < 0) { |
1537 | // if all output fields are 0s, do not use negative sign |
1538 | for (int32_t idx = 0; idx <= lastIdx; idx++) { |
1539 | if (fields[idx] != 0) { |
1540 | sign = MINUS; |
1541 | break; |
1542 | } |
1543 | } |
1544 | } |
1545 | result.setTo(sign); |
1546 | |
1547 | for (int32_t idx = 0; idx <= lastIdx; idx++) { |
1548 | if (sep && idx != 0) { |
1549 | result.append(sep); |
1550 | } |
1551 | result.append((UChar)(0x0030 + fields[idx]/10)); |
1552 | result.append((UChar)(0x0030 + fields[idx]%10)); |
1553 | } |
1554 | |
1555 | return result; |
1556 | } |
1557 | |
1558 | UnicodeString& |
1559 | TimeZoneFormat::formatOffsetLocalizedGMT(int32_t offset, UBool isShort, UnicodeString& result, UErrorCode& status) const { |
1560 | if (U_FAILURE(status)) { |
1561 | result.setToBogus(); |
1562 | return result; |
1563 | } |
1564 | if (offset <= -MAX_OFFSET || offset >= MAX_OFFSET) { |
1565 | result.setToBogus(); |
1566 | status = U_ILLEGAL_ARGUMENT_ERROR; |
1567 | return result; |
1568 | } |
1569 | |
1570 | if (offset == 0) { |
1571 | result.setTo(fGMTZeroFormat); |
1572 | return result; |
1573 | } |
1574 | |
1575 | UBool positive = TRUE; |
1576 | if (offset < 0) { |
1577 | offset = -offset; |
1578 | positive = FALSE; |
1579 | } |
1580 | |
1581 | int32_t offsetH = offset / MILLIS_PER_HOUR; |
1582 | offset = offset % MILLIS_PER_HOUR; |
1583 | int32_t offsetM = offset / MILLIS_PER_MINUTE; |
1584 | offset = offset % MILLIS_PER_MINUTE; |
1585 | int32_t offsetS = offset / MILLIS_PER_SECOND; |
1586 | |
1587 | U_ASSERT(offsetH <= MAX_OFFSET_HOUR && offsetM <= MAX_OFFSET_MINUTE && offsetS <= MAX_OFFSET_SECOND); |
1588 | |
1589 | const UVector* offsetPatternItems = NULL; |
1590 | if (positive) { |
1591 | if (offsetS != 0) { |
1592 | offsetPatternItems = fGMTOffsetPatternItems[UTZFMT_PAT_POSITIVE_HMS]; |
1593 | } else if (offsetM != 0 || !isShort) { |
1594 | offsetPatternItems = fGMTOffsetPatternItems[UTZFMT_PAT_POSITIVE_HM]; |
1595 | } else { |
1596 | offsetPatternItems = fGMTOffsetPatternItems[UTZFMT_PAT_POSITIVE_H]; |
1597 | } |
1598 | } else { |
1599 | if (offsetS != 0) { |
1600 | offsetPatternItems = fGMTOffsetPatternItems[UTZFMT_PAT_NEGATIVE_HMS]; |
1601 | } else if (offsetM != 0 || !isShort) { |
1602 | offsetPatternItems = fGMTOffsetPatternItems[UTZFMT_PAT_NEGATIVE_HM]; |
1603 | } else { |
1604 | offsetPatternItems = fGMTOffsetPatternItems[UTZFMT_PAT_NEGATIVE_H]; |
1605 | } |
1606 | } |
1607 | |
1608 | U_ASSERT(offsetPatternItems != NULL); |
1609 | |
1610 | // Building the GMT format string |
1611 | result.setTo(fGMTPatternPrefix); |
1612 | |
1613 | for (int32_t i = 0; i < offsetPatternItems->size(); i++) { |
1614 | const GMTOffsetField* item = (GMTOffsetField*)offsetPatternItems->elementAt(i); |
1615 | GMTOffsetField::FieldType type = item->getType(); |
1616 | |
1617 | switch (type) { |
1618 | case GMTOffsetField::TEXT: |
1619 | result.append(item->getPatternText(), -1); |
1620 | break; |
1621 | |
1622 | case GMTOffsetField::HOUR: |
1623 | appendOffsetDigits(result, offsetH, (isShort ? 1 : 2)); |
1624 | break; |
1625 | |
1626 | case GMTOffsetField::MINUTE: |
1627 | appendOffsetDigits(result, offsetM, 2); |
1628 | break; |
1629 | |
1630 | case GMTOffsetField::SECOND: |
1631 | appendOffsetDigits(result, offsetS, 2); |
1632 | break; |
1633 | } |
1634 | } |
1635 | |
1636 | result.append(fGMTPatternSuffix); |
1637 | return result; |
1638 | } |
1639 | |
1640 | int32_t |
1641 | TimeZoneFormat::parseOffsetISO8601(const UnicodeString& text, ParsePosition& pos, UBool extendedOnly, UBool* hasDigitOffset /* = NULL */) const { |
1642 | if (hasDigitOffset) { |
1643 | *hasDigitOffset = FALSE; |
1644 | } |
1645 | int32_t start = pos.getIndex(); |
1646 | if (start >= text.length()) { |
1647 | pos.setErrorIndex(start); |
1648 | return 0; |
1649 | } |
1650 | |
1651 | UChar firstChar = text.charAt(start); |
1652 | if (firstChar == ISO8601_UTC || firstChar == (UChar)(ISO8601_UTC + 0x20)) { |
1653 | // "Z" (or "z") - indicates UTC |
1654 | pos.setIndex(start + 1); |
1655 | return 0; |
1656 | } |
1657 | |
1658 | int32_t sign = 1; |
1659 | if (firstChar == PLUS) { |
1660 | sign = 1; |
1661 | } else if (firstChar == MINUS) { |
1662 | sign = -1; |
1663 | } else { |
1664 | // Not an ISO 8601 offset string |
1665 | pos.setErrorIndex(start); |
1666 | return 0; |
1667 | } |
1668 | ParsePosition posOffset(start + 1); |
1669 | int32_t offset = parseAsciiOffsetFields(text, posOffset, ISO8601_SEP, FIELDS_H, FIELDS_HMS); |
1670 | if (posOffset.getErrorIndex() == -1 && !extendedOnly && (posOffset.getIndex() - start <= 3)) { |
1671 | // If the text is successfully parsed as extended format with the options above, it can be also parsed |
1672 | // as basic format. For example, "0230" can be parsed as offset 2:00 (only first digits are valid for |
1673 | // extended format), but it can be parsed as offset 2:30 with basic format. We use longer result. |
1674 | ParsePosition posBasic(start + 1); |
1675 | int32_t tmpOffset = parseAbuttingAsciiOffsetFields(text, posBasic, FIELDS_H, FIELDS_HMS, FALSE); |
1676 | if (posBasic.getErrorIndex() == -1 && posBasic.getIndex() > posOffset.getIndex()) { |
1677 | offset = tmpOffset; |
1678 | posOffset.setIndex(posBasic.getIndex()); |
1679 | } |
1680 | } |
1681 | |
1682 | if (posOffset.getErrorIndex() != -1) { |
1683 | pos.setErrorIndex(start); |
1684 | return 0; |
1685 | } |
1686 | |
1687 | pos.setIndex(posOffset.getIndex()); |
1688 | if (hasDigitOffset) { |
1689 | *hasDigitOffset = TRUE; |
1690 | } |
1691 | return sign * offset; |
1692 | } |
1693 | |
1694 | int32_t |
1695 | TimeZoneFormat::parseOffsetLocalizedGMT(const UnicodeString& text, ParsePosition& pos, UBool isShort, UBool* hasDigitOffset) const { |
1696 | int32_t start = pos.getIndex(); |
1697 | int32_t offset = 0; |
1698 | int32_t parsedLength = 0; |
1699 | |
1700 | if (hasDigitOffset) { |
1701 | *hasDigitOffset = FALSE; |
1702 | } |
1703 | |
1704 | offset = parseOffsetLocalizedGMTPattern(text, start, isShort, parsedLength); |
1705 | |
1706 | // For now, parseOffsetLocalizedGMTPattern handles both long and short |
1707 | // formats, no matter isShort is true or false. This might be changed in future |
1708 | // when strict parsing is necessary, or different set of patterns are used for |
1709 | // short/long formats. |
1710 | #if 0 |
1711 | if (parsedLength == 0) { |
1712 | offset = parseOffsetLocalizedGMTPattern(text, start, !isShort, parsedLength); |
1713 | } |
1714 | #endif |
1715 | |
1716 | if (parsedLength > 0) { |
1717 | if (hasDigitOffset) { |
1718 | *hasDigitOffset = TRUE; |
1719 | } |
1720 | pos.setIndex(start + parsedLength); |
1721 | return offset; |
1722 | } |
1723 | |
1724 | // Try the default patterns |
1725 | offset = parseOffsetDefaultLocalizedGMT(text, start, parsedLength); |
1726 | if (parsedLength > 0) { |
1727 | if (hasDigitOffset) { |
1728 | *hasDigitOffset = TRUE; |
1729 | } |
1730 | pos.setIndex(start + parsedLength); |
1731 | return offset; |
1732 | } |
1733 | |
1734 | // Check if this is a GMT zero format |
1735 | if (text.caseCompare(start, fGMTZeroFormat.length(), fGMTZeroFormat, 0) == 0) { |
1736 | pos.setIndex(start + fGMTZeroFormat.length()); |
1737 | return 0; |
1738 | } |
1739 | |
1740 | // Check if this is a default GMT zero format |
1741 | for (int32_t i = 0; ALT_GMT_STRINGS[i][0] != 0; i++) { |
1742 | const UChar* defGMTZero = ALT_GMT_STRINGS[i]; |
1743 | int32_t defGMTZeroLen = u_strlen(defGMTZero); |
1744 | if (text.caseCompare(start, defGMTZeroLen, defGMTZero, 0) == 0) { |
1745 | pos.setIndex(start + defGMTZeroLen); |
1746 | return 0; |
1747 | } |
1748 | } |
1749 | |
1750 | // Nothing matched |
1751 | pos.setErrorIndex(start); |
1752 | return 0; |
1753 | } |
1754 | |
1755 | int32_t |
1756 | TimeZoneFormat::parseOffsetLocalizedGMTPattern(const UnicodeString& text, int32_t start, UBool /*isShort*/, int32_t& parsedLen) const { |
1757 | int32_t idx = start; |
1758 | int32_t offset = 0; |
1759 | UBool parsed = FALSE; |
1760 | |
1761 | do { |
1762 | // Prefix part |
1763 | int32_t len = fGMTPatternPrefix.length(); |
1764 | if (len > 0 && text.caseCompare(idx, len, fGMTPatternPrefix, 0) != 0) { |
1765 | // prefix match failed |
1766 | break; |
1767 | } |
1768 | idx += len; |
1769 | |
1770 | // Offset part |
1771 | offset = parseOffsetFields(text, idx, FALSE, len); |
1772 | if (len == 0) { |
1773 | // offset field match failed |
1774 | break; |
1775 | } |
1776 | idx += len; |
1777 | |
1778 | len = fGMTPatternSuffix.length(); |
1779 | if (len > 0 && text.caseCompare(idx, len, fGMTPatternSuffix, 0) != 0) { |
1780 | // no suffix match |
1781 | break; |
1782 | } |
1783 | idx += len; |
1784 | parsed = TRUE; |
1785 | } while (FALSE); |
1786 | |
1787 | parsedLen = parsed ? idx - start : 0; |
1788 | return offset; |
1789 | } |
1790 | |
1791 | int32_t |
1792 | TimeZoneFormat::parseOffsetFields(const UnicodeString& text, int32_t start, UBool /*isShort*/, int32_t& parsedLen) const { |
1793 | int32_t outLen = 0; |
1794 | int32_t offset = 0; |
1795 | int32_t sign = 1; |
1796 | |
1797 | parsedLen = 0; |
1798 | |
1799 | int32_t offsetH, offsetM, offsetS; |
1800 | offsetH = offsetM = offsetS = 0; |
1801 | |
1802 | for (int32_t patidx = 0; PARSE_GMT_OFFSET_TYPES[patidx] >= 0; patidx++) { |
1803 | int32_t gmtPatType = PARSE_GMT_OFFSET_TYPES[patidx]; |
1804 | UVector* items = fGMTOffsetPatternItems[gmtPatType]; |
1805 | U_ASSERT(items != NULL); |
1806 | |
1807 | outLen = parseOffsetFieldsWithPattern(text, start, items, FALSE, offsetH, offsetM, offsetS); |
1808 | if (outLen > 0) { |
1809 | sign = (gmtPatType == UTZFMT_PAT_POSITIVE_H || gmtPatType == UTZFMT_PAT_POSITIVE_HM || gmtPatType == UTZFMT_PAT_POSITIVE_HMS) ? |
1810 | 1 : -1; |
1811 | break; |
1812 | } |
1813 | } |
1814 | |
1815 | if (outLen > 0 && fAbuttingOffsetHoursAndMinutes) { |
1816 | // When hours field is sabutting minutes field, |
1817 | // the parse result above may not be appropriate. |
1818 | // For example, "01020" is parsed as 01:02: above, |
1819 | // but it should be parsed as 00:10:20. |
1820 | int32_t tmpLen = 0; |
1821 | int32_t tmpSign = 1; |
1822 | int32_t tmpH = 0; |
1823 | int32_t tmpM = 0; |
1824 | int32_t tmpS = 0; |
1825 | |
1826 | for (int32_t patidx = 0; PARSE_GMT_OFFSET_TYPES[patidx] >= 0; patidx++) { |
1827 | int32_t gmtPatType = PARSE_GMT_OFFSET_TYPES[patidx]; |
1828 | UVector* items = fGMTOffsetPatternItems[gmtPatType]; |
1829 | U_ASSERT(items != NULL); |
1830 | |
1831 | // forcing parse to use single hour digit |
1832 | tmpLen = parseOffsetFieldsWithPattern(text, start, items, TRUE, tmpH, tmpM, tmpS); |
1833 | if (tmpLen > 0) { |
1834 | tmpSign = (gmtPatType == UTZFMT_PAT_POSITIVE_H || gmtPatType == UTZFMT_PAT_POSITIVE_HM || gmtPatType == UTZFMT_PAT_POSITIVE_HMS) ? |
1835 | 1 : -1; |
1836 | break; |
1837 | } |
1838 | } |
1839 | if (tmpLen > outLen) { |
1840 | // Better parse result with single hour digit |
1841 | outLen = tmpLen; |
1842 | sign = tmpSign; |
1843 | offsetH = tmpH; |
1844 | offsetM = tmpM; |
1845 | offsetS = tmpS; |
1846 | } |
1847 | } |
1848 | |
1849 | if (outLen > 0) { |
1850 | offset = ((((offsetH * 60) + offsetM) * 60) + offsetS) * 1000 * sign; |
1851 | parsedLen = outLen; |
1852 | } |
1853 | |
1854 | return offset; |
1855 | } |
1856 | |
1857 | int32_t |
1858 | TimeZoneFormat::parseOffsetFieldsWithPattern(const UnicodeString& text, int32_t start, |
1859 | UVector* patternItems, UBool forceSingleHourDigit, int32_t& hour, int32_t& min, int32_t& sec) const { |
1860 | UBool failed = FALSE; |
1861 | int32_t offsetH, offsetM, offsetS; |
1862 | offsetH = offsetM = offsetS = 0; |
1863 | int32_t idx = start; |
1864 | |
1865 | for (int32_t i = 0; i < patternItems->size(); i++) { |
1866 | int32_t len = 0; |
1867 | const GMTOffsetField* field = (const GMTOffsetField*)patternItems->elementAt(i); |
1868 | GMTOffsetField::FieldType fieldType = field->getType(); |
1869 | if (fieldType == GMTOffsetField::TEXT) { |
1870 | const UChar* patStr = field->getPatternText(); |
1871 | len = u_strlen(patStr); |
1872 | if (i == 0) { |
1873 | // When TimeZoneFormat parse() is called from SimpleDateFormat, |
1874 | // leading space characters might be truncated. If the first pattern text |
1875 | // starts with such character (e.g. Bidi control), then we need to |
1876 | // skip the leading space charcters. |
1877 | if (idx < text.length() && !PatternProps::isWhiteSpace(text.char32At(idx))) { |
1878 | while (len > 0) { |
1879 | UChar32 ch; |
1880 | int32_t chLen; |
1881 | U16_GET(patStr, 0, 0, len, ch); |
1882 | if (PatternProps::isWhiteSpace(ch)) { |
1883 | chLen = U16_LENGTH(ch); |
1884 | len -= chLen; |
1885 | patStr += chLen; |
1886 | } |
1887 | else { |
1888 | break; |
1889 | } |
1890 | } |
1891 | } |
1892 | } |
1893 | if (text.caseCompare(idx, len, patStr, 0) != 0) { |
1894 | failed = TRUE; |
1895 | break; |
1896 | } |
1897 | idx += len; |
1898 | } else { |
1899 | if (fieldType == GMTOffsetField::HOUR) { |
1900 | uint8_t maxDigits = forceSingleHourDigit ? 1 : 2; |
1901 | offsetH = parseOffsetFieldWithLocalizedDigits(text, idx, 1, maxDigits, 0, MAX_OFFSET_HOUR, len); |
1902 | } else if (fieldType == GMTOffsetField::MINUTE) { |
1903 | offsetM = parseOffsetFieldWithLocalizedDigits(text, idx, 2, 2, 0, MAX_OFFSET_MINUTE, len); |
1904 | } else if (fieldType == GMTOffsetField::SECOND) { |
1905 | offsetS = parseOffsetFieldWithLocalizedDigits(text, idx, 2, 2, 0, MAX_OFFSET_SECOND, len); |
1906 | } |
1907 | |
1908 | if (len == 0) { |
1909 | failed = TRUE; |
1910 | break; |
1911 | } |
1912 | idx += len; |
1913 | } |
1914 | } |
1915 | |
1916 | if (failed) { |
1917 | hour = min = sec = 0; |
1918 | return 0; |
1919 | } |
1920 | |
1921 | hour = offsetH; |
1922 | min = offsetM; |
1923 | sec = offsetS; |
1924 | |
1925 | return idx - start; |
1926 | } |
1927 | |
1928 | int32_t |
1929 | TimeZoneFormat::parseAbuttingOffsetFields(const UnicodeString& text, int32_t start, int32_t& parsedLen) const { |
1930 | int32_t digits[MAX_OFFSET_DIGITS]; |
1931 | int32_t parsed[MAX_OFFSET_DIGITS]; // accumulative offsets |
1932 | |
1933 | // Parse digits into int[] |
1934 | int32_t idx = start; |
1935 | int32_t len = 0; |
1936 | int32_t numDigits = 0; |
1937 | for (int32_t i = 0; i < MAX_OFFSET_DIGITS; i++) { |
1938 | digits[i] = parseSingleLocalizedDigit(text, idx, len); |
1939 | if (digits[i] < 0) { |
1940 | break; |
1941 | } |
1942 | idx += len; |
1943 | parsed[i] = idx - start; |
1944 | numDigits++; |
1945 | } |
1946 | |
1947 | if (numDigits == 0) { |
1948 | parsedLen = 0; |
1949 | return 0; |
1950 | } |
1951 | |
1952 | int32_t offset = 0; |
1953 | while (numDigits > 0) { |
1954 | int32_t hour = 0; |
1955 | int32_t min = 0; |
1956 | int32_t sec = 0; |
1957 | |
1958 | U_ASSERT(numDigits > 0 && numDigits <= MAX_OFFSET_DIGITS); |
1959 | switch (numDigits) { |
1960 | case 1: // H |
1961 | hour = digits[0]; |
1962 | break; |
1963 | case 2: // HH |
1964 | hour = digits[0] * 10 + digits[1]; |
1965 | break; |
1966 | case 3: // Hmm |
1967 | hour = digits[0]; |
1968 | min = digits[1] * 10 + digits[2]; |
1969 | break; |
1970 | case 4: // HHmm |
1971 | hour = digits[0] * 10 + digits[1]; |
1972 | min = digits[2] * 10 + digits[3]; |
1973 | break; |
1974 | case 5: // Hmmss |
1975 | hour = digits[0]; |
1976 | min = digits[1] * 10 + digits[2]; |
1977 | sec = digits[3] * 10 + digits[4]; |
1978 | break; |
1979 | case 6: // HHmmss |
1980 | hour = digits[0] * 10 + digits[1]; |
1981 | min = digits[2] * 10 + digits[3]; |
1982 | sec = digits[4] * 10 + digits[5]; |
1983 | break; |
1984 | } |
1985 | if (hour <= MAX_OFFSET_HOUR && min <= MAX_OFFSET_MINUTE && sec <= MAX_OFFSET_SECOND) { |
1986 | // found a valid combination |
1987 | offset = hour * MILLIS_PER_HOUR + min * MILLIS_PER_MINUTE + sec * MILLIS_PER_SECOND; |
1988 | parsedLen = parsed[numDigits - 1]; |
1989 | break; |
1990 | } |
1991 | numDigits--; |
1992 | } |
1993 | return offset; |
1994 | } |
1995 | |
1996 | int32_t |
1997 | TimeZoneFormat::parseOffsetDefaultLocalizedGMT(const UnicodeString& text, int start, int32_t& parsedLen) const { |
1998 | int32_t idx = start; |
1999 | int32_t offset = 0; |
2000 | int32_t parsed = 0; |
2001 | |
2002 | do { |
2003 | // check global default GMT alternatives |
2004 | int32_t gmtLen = 0; |
2005 | |
2006 | for (int32_t i = 0; ALT_GMT_STRINGS[i][0] != 0; i++) { |
2007 | const UChar* gmt = ALT_GMT_STRINGS[i]; |
2008 | int32_t len = u_strlen(gmt); |
2009 | if (text.caseCompare(start, len, gmt, 0) == 0) { |
2010 | gmtLen = len; |
2011 | break; |
2012 | } |
2013 | } |
2014 | if (gmtLen == 0) { |
2015 | break; |
2016 | } |
2017 | idx += gmtLen; |
2018 | |
2019 | // offset needs a sign char and a digit at minimum |
2020 | if (idx + 1 >= text.length()) { |
2021 | break; |
2022 | } |
2023 | |
2024 | // parse sign |
2025 | int32_t sign = 1; |
2026 | UChar c = text.charAt(idx); |
2027 | if (c == PLUS) { |
2028 | sign = 1; |
2029 | } else if (c == MINUS) { |
2030 | sign = -1; |
2031 | } else { |
2032 | break; |
2033 | } |
2034 | idx++; |
2035 | |
2036 | // offset part |
2037 | // try the default pattern with the separator first |
2038 | int32_t lenWithSep = 0; |
2039 | int32_t offsetWithSep = parseDefaultOffsetFields(text, idx, DEFAULT_GMT_OFFSET_SEP, lenWithSep); |
2040 | if (lenWithSep == text.length() - idx) { |
2041 | // maximum match |
2042 | offset = offsetWithSep * sign; |
2043 | idx += lenWithSep; |
2044 | } else { |
2045 | // try abutting field pattern |
2046 | int32_t lenAbut = 0; |
2047 | int32_t offsetAbut = parseAbuttingOffsetFields(text, idx, lenAbut); |
2048 | |
2049 | if (lenWithSep > lenAbut) { |
2050 | offset = offsetWithSep * sign; |
2051 | idx += lenWithSep; |
2052 | } else { |
2053 | offset = offsetAbut * sign; |
2054 | idx += lenAbut; |
2055 | } |
2056 | } |
2057 | parsed = idx - start; |
2058 | } while (false); |
2059 | |
2060 | parsedLen = parsed; |
2061 | return offset; |
2062 | } |
2063 | |
2064 | int32_t |
2065 | TimeZoneFormat::parseDefaultOffsetFields(const UnicodeString& text, int32_t start, UChar separator, int32_t& parsedLen) const { |
2066 | int32_t max = text.length(); |
2067 | int32_t idx = start; |
2068 | int32_t len = 0; |
2069 | int32_t hour = 0, min = 0, sec = 0; |
2070 | |
2071 | parsedLen = 0; |
2072 | |
2073 | do { |
2074 | hour = parseOffsetFieldWithLocalizedDigits(text, idx, 1, 2, 0, MAX_OFFSET_HOUR, len); |
2075 | if (len == 0) { |
2076 | break; |
2077 | } |
2078 | idx += len; |
2079 | |
2080 | if (idx + 1 < max && text.charAt(idx) == separator) { |
2081 | min = parseOffsetFieldWithLocalizedDigits(text, idx + 1, 2, 2, 0, MAX_OFFSET_MINUTE, len); |
2082 | if (len == 0) { |
2083 | break; |
2084 | } |
2085 | idx += (1 + len); |
2086 | |
2087 | if (idx + 1 < max && text.charAt(idx) == separator) { |
2088 | sec = parseOffsetFieldWithLocalizedDigits(text, idx + 1, 2, 2, 0, MAX_OFFSET_SECOND, len); |
2089 | if (len == 0) { |
2090 | break; |
2091 | } |
2092 | idx += (1 + len); |
2093 | } |
2094 | } |
2095 | } while (FALSE); |
2096 | |
2097 | if (idx == start) { |
2098 | return 0; |
2099 | } |
2100 | |
2101 | parsedLen = idx - start; |
2102 | return hour * MILLIS_PER_HOUR + min * MILLIS_PER_MINUTE + sec * MILLIS_PER_SECOND; |
2103 | } |
2104 | |
2105 | int32_t |
2106 | TimeZoneFormat::parseOffsetFieldWithLocalizedDigits(const UnicodeString& text, int32_t start, uint8_t minDigits, uint8_t maxDigits, uint16_t minVal, uint16_t maxVal, int32_t& parsedLen) const { |
2107 | parsedLen = 0; |
2108 | |
2109 | int32_t decVal = 0; |
2110 | int32_t numDigits = 0; |
2111 | int32_t idx = start; |
2112 | int32_t digitLen = 0; |
2113 | |
2114 | while (idx < text.length() && numDigits < maxDigits) { |
2115 | int32_t digit = parseSingleLocalizedDigit(text, idx, digitLen); |
2116 | if (digit < 0) { |
2117 | break; |
2118 | } |
2119 | int32_t tmpVal = decVal * 10 + digit; |
2120 | if (tmpVal > maxVal) { |
2121 | break; |
2122 | } |
2123 | decVal = tmpVal; |
2124 | numDigits++; |
2125 | idx += digitLen; |
2126 | } |
2127 | |
2128 | // Note: maxVal is checked in the while loop |
2129 | if (numDigits < minDigits || decVal < minVal) { |
2130 | decVal = -1; |
2131 | numDigits = 0; |
2132 | } else { |
2133 | parsedLen = idx - start; |
2134 | } |
2135 | |
2136 | return decVal; |
2137 | } |
2138 | |
2139 | int32_t |
2140 | TimeZoneFormat::parseSingleLocalizedDigit(const UnicodeString& text, int32_t start, int32_t& len) const { |
2141 | int32_t digit = -1; |
2142 | len = 0; |
2143 | if (start < text.length()) { |
2144 | UChar32 cp = text.char32At(start); |
2145 | |
2146 | // First, try digits configured for this instance |
2147 | for (int32_t i = 0; i < 10; i++) { |
2148 | if (cp == fGMTOffsetDigits[i]) { |
2149 | digit = i; |
2150 | break; |
2151 | } |
2152 | } |
2153 | // If failed, check if this is a Unicode digit |
2154 | if (digit < 0) { |
2155 | int32_t tmp = u_charDigitValue(cp); |
2156 | digit = (tmp >= 0 && tmp <= 9) ? tmp : -1; |
2157 | } |
2158 | |
2159 | if (digit >= 0) { |
2160 | int32_t next = text.moveIndex32(start, 1); |
2161 | len = next - start; |
2162 | } |
2163 | } |
2164 | return digit; |
2165 | } |
2166 | |
2167 | UnicodeString& |
2168 | TimeZoneFormat::formatOffsetWithAsciiDigits(int32_t offset, UChar sep, OffsetFields minFields, OffsetFields maxFields, UnicodeString& result) { |
2169 | U_ASSERT(maxFields >= minFields); |
2170 | U_ASSERT(offset > -MAX_OFFSET && offset < MAX_OFFSET); |
2171 | |
2172 | UChar sign = PLUS; |
2173 | if (offset < 0) { |
2174 | sign = MINUS; |
2175 | offset = -offset; |
2176 | } |
2177 | result.setTo(sign); |
2178 | |
2179 | int fields[3]; |
2180 | fields[0] = offset / MILLIS_PER_HOUR; |
2181 | offset = offset % MILLIS_PER_HOUR; |
2182 | fields[1] = offset / MILLIS_PER_MINUTE; |
2183 | offset = offset % MILLIS_PER_MINUTE; |
2184 | fields[2] = offset / MILLIS_PER_SECOND; |
2185 | |
2186 | U_ASSERT(fields[0] >= 0 && fields[0] <= MAX_OFFSET_HOUR); |
2187 | U_ASSERT(fields[1] >= 0 && fields[1] <= MAX_OFFSET_MINUTE); |
2188 | U_ASSERT(fields[2] >= 0 && fields[2] <= MAX_OFFSET_SECOND); |
2189 | |
2190 | int32_t lastIdx = maxFields; |
2191 | while (lastIdx > minFields) { |
2192 | if (fields[lastIdx] != 0) { |
2193 | break; |
2194 | } |
2195 | lastIdx--; |
2196 | } |
2197 | |
2198 | for (int32_t idx = 0; idx <= lastIdx; idx++) { |
2199 | if (sep && idx != 0) { |
2200 | result.append(sep); |
2201 | } |
2202 | result.append((UChar)(0x0030 + fields[idx]/10)); |
2203 | result.append((UChar)(0x0030 + fields[idx]%10)); |
2204 | } |
2205 | |
2206 | return result; |
2207 | } |
2208 | |
2209 | int32_t |
2210 | TimeZoneFormat::parseAbuttingAsciiOffsetFields(const UnicodeString& text, ParsePosition& pos, OffsetFields minFields, OffsetFields maxFields, UBool fixedHourWidth) { |
2211 | int32_t start = pos.getIndex(); |
2212 | |
2213 | int32_t minDigits = 2 * (minFields + 1) - (fixedHourWidth ? 0 : 1); |
2214 | int32_t maxDigits = 2 * (maxFields + 1); |
2215 | |
2216 | U_ASSERT(maxDigits <= MAX_OFFSET_DIGITS); |
2217 | |
2218 | int32_t digits[MAX_OFFSET_DIGITS] = {}; |
2219 | int32_t numDigits = 0; |
2220 | int32_t idx = start; |
2221 | while (numDigits < maxDigits && idx < text.length()) { |
2222 | UChar uch = text.charAt(idx); |
2223 | int32_t digit = DIGIT_VAL(uch); |
2224 | if (digit < 0) { |
2225 | break; |
2226 | } |
2227 | digits[numDigits] = digit; |
2228 | numDigits++; |
2229 | idx++; |
2230 | } |
2231 | |
2232 | if (fixedHourWidth && (numDigits & 1)) { |
2233 | // Fixed digits, so the number of digits must be even number. Truncating. |
2234 | numDigits--; |
2235 | } |
2236 | |
2237 | if (numDigits < minDigits) { |
2238 | pos.setErrorIndex(start); |
2239 | return 0; |
2240 | } |
2241 | |
2242 | int32_t hour = 0, min = 0, sec = 0; |
2243 | UBool bParsed = FALSE; |
2244 | while (numDigits >= minDigits) { |
2245 | switch (numDigits) { |
2246 | case 1: //H |
2247 | hour = digits[0]; |
2248 | break; |
2249 | case 2: //HH |
2250 | hour = digits[0] * 10 + digits[1]; |
2251 | break; |
2252 | case 3: //Hmm |
2253 | hour = digits[0]; |
2254 | min = digits[1] * 10 + digits[2]; |
2255 | break; |
2256 | case 4: //HHmm |
2257 | hour = digits[0] * 10 + digits[1]; |
2258 | min = digits[2] * 10 + digits[3]; |
2259 | break; |
2260 | case 5: //Hmmss |
2261 | hour = digits[0]; |
2262 | min = digits[1] * 10 + digits[2]; |
2263 | sec = digits[3] * 10 + digits[4]; |
2264 | break; |
2265 | case 6: //HHmmss |
2266 | hour = digits[0] * 10 + digits[1]; |
2267 | min = digits[2] * 10 + digits[3]; |
2268 | sec = digits[4] * 10 + digits[5]; |
2269 | break; |
2270 | } |
2271 | |
2272 | if (hour <= MAX_OFFSET_HOUR && min <= MAX_OFFSET_MINUTE && sec <= MAX_OFFSET_SECOND) { |
2273 | // Successfully parsed |
2274 | bParsed = true; |
2275 | break; |
2276 | } |
2277 | |
2278 | // Truncating |
2279 | numDigits -= (fixedHourWidth ? 2 : 1); |
2280 | hour = min = sec = 0; |
2281 | } |
2282 | |
2283 | if (!bParsed) { |
2284 | pos.setErrorIndex(start); |
2285 | return 0; |
2286 | } |
2287 | pos.setIndex(start + numDigits); |
2288 | return ((((hour * 60) + min) * 60) + sec) * 1000; |
2289 | } |
2290 | |
2291 | int32_t |
2292 | TimeZoneFormat::parseAsciiOffsetFields(const UnicodeString& text, ParsePosition& pos, UChar sep, OffsetFields minFields, OffsetFields maxFields) { |
2293 | int32_t start = pos.getIndex(); |
2294 | int32_t fieldVal[] = {0, 0, 0}; |
2295 | int32_t fieldLen[] = {0, -1, -1}; |
2296 | for (int32_t idx = start, fieldIdx = 0; idx < text.length() && fieldIdx <= maxFields; idx++) { |
2297 | UChar c = text.charAt(idx); |
2298 | if (c == sep) { |
2299 | if (fieldIdx == 0) { |
2300 | if (fieldLen[0] == 0) { |
2301 | // no hours field |
2302 | break; |
2303 | } |
2304 | // 1 digit hour, move to next field |
2305 | } else { |
2306 | if (fieldLen[fieldIdx] != -1) { |
2307 | // premature minute or seconds field |
2308 | break; |
2309 | } |
2310 | fieldLen[fieldIdx] = 0; |
2311 | } |
2312 | continue; |
2313 | } else if (fieldLen[fieldIdx] == -1) { |
2314 | // no separator after 2 digit field |
2315 | break; |
2316 | } |
2317 | int32_t digit = DIGIT_VAL(c); |
2318 | if (digit < 0) { |
2319 | // not a digit |
2320 | break; |
2321 | } |
2322 | fieldVal[fieldIdx] = fieldVal[fieldIdx] * 10 + digit; |
2323 | fieldLen[fieldIdx]++; |
2324 | if (fieldLen[fieldIdx] >= 2) { |
2325 | // parsed 2 digits, move to next field |
2326 | fieldIdx++; |
2327 | } |
2328 | } |
2329 | |
2330 | int32_t offset = 0; |
2331 | int32_t parsedLen = 0; |
2332 | int32_t parsedFields = -1; |
2333 | do { |
2334 | // hour |
2335 | if (fieldLen[0] == 0) { |
2336 | break; |
2337 | } |
2338 | if (fieldVal[0] > MAX_OFFSET_HOUR) { |
2339 | offset = (fieldVal[0] / 10) * MILLIS_PER_HOUR; |
2340 | parsedFields = FIELDS_H; |
2341 | parsedLen = 1; |
2342 | break; |
2343 | } |
2344 | offset = fieldVal[0] * MILLIS_PER_HOUR; |
2345 | parsedLen = fieldLen[0]; |
2346 | parsedFields = FIELDS_H; |
2347 | |
2348 | // minute |
2349 | if (fieldLen[1] != 2 || fieldVal[1] > MAX_OFFSET_MINUTE) { |
2350 | break; |
2351 | } |
2352 | offset += fieldVal[1] * MILLIS_PER_MINUTE; |
2353 | parsedLen += (1 + fieldLen[1]); |
2354 | parsedFields = FIELDS_HM; |
2355 | |
2356 | // second |
2357 | if (fieldLen[2] != 2 || fieldVal[2] > MAX_OFFSET_SECOND) { |
2358 | break; |
2359 | } |
2360 | offset += fieldVal[2] * MILLIS_PER_SECOND; |
2361 | parsedLen += (1 + fieldLen[2]); |
2362 | parsedFields = FIELDS_HMS; |
2363 | } while (false); |
2364 | |
2365 | if (parsedFields < minFields) { |
2366 | pos.setErrorIndex(start); |
2367 | return 0; |
2368 | } |
2369 | |
2370 | pos.setIndex(start + parsedLen); |
2371 | return offset; |
2372 | } |
2373 | |
2374 | void |
2375 | TimeZoneFormat::appendOffsetDigits(UnicodeString& buf, int32_t n, uint8_t minDigits) const { |
2376 | U_ASSERT(n >= 0 && n < 60); |
2377 | int32_t numDigits = n >= 10 ? 2 : 1; |
2378 | for (int32_t i = 0; i < minDigits - numDigits; i++) { |
2379 | buf.append(fGMTOffsetDigits[0]); |
2380 | } |
2381 | if (numDigits == 2) { |
2382 | buf.append(fGMTOffsetDigits[n / 10]); |
2383 | } |
2384 | buf.append(fGMTOffsetDigits[n % 10]); |
2385 | } |
2386 | |
2387 | // ------------------------------------------------------------------ |
2388 | // Private misc |
2389 | void |
2390 | TimeZoneFormat::initGMTPattern(const UnicodeString& gmtPattern, UErrorCode& status) { |
2391 | if (U_FAILURE(status)) { |
2392 | return; |
2393 | } |
2394 | // This implementation not perfect, but sufficient practically. |
2395 | int32_t idx = gmtPattern.indexOf(ARG0, ARG0_LEN, 0); |
2396 | if (idx < 0) { |
2397 | status = U_ILLEGAL_ARGUMENT_ERROR; |
2398 | return; |
2399 | } |
2400 | fGMTPattern.setTo(gmtPattern); |
2401 | unquote(gmtPattern.tempSubString(0, idx), fGMTPatternPrefix); |
2402 | unquote(gmtPattern.tempSubString(idx + ARG0_LEN), fGMTPatternSuffix); |
2403 | } |
2404 | |
2405 | UnicodeString& |
2406 | TimeZoneFormat::unquote(const UnicodeString& pattern, UnicodeString& result) { |
2407 | if (pattern.indexOf(SINGLEQUOTE) < 0) { |
2408 | result.setTo(pattern); |
2409 | return result; |
2410 | } |
2411 | result.remove(); |
2412 | UBool isPrevQuote = FALSE; |
2413 | UBool inQuote = FALSE; |
2414 | for (int32_t i = 0; i < pattern.length(); i++) { |
2415 | UChar c = pattern.charAt(i); |
2416 | if (c == SINGLEQUOTE) { |
2417 | if (isPrevQuote) { |
2418 | result.append(c); |
2419 | isPrevQuote = FALSE; |
2420 | } else { |
2421 | isPrevQuote = TRUE; |
2422 | } |
2423 | inQuote = !inQuote; |
2424 | } else { |
2425 | isPrevQuote = FALSE; |
2426 | result.append(c); |
2427 | } |
2428 | } |
2429 | return result; |
2430 | } |
2431 | |
2432 | UVector* |
2433 | TimeZoneFormat::parseOffsetPattern(const UnicodeString& pattern, OffsetFields required, UErrorCode& status) { |
2434 | if (U_FAILURE(status)) { |
2435 | return NULL; |
2436 | } |
2437 | UVector* result = new UVector(deleteGMTOffsetField, NULL, status); |
2438 | if (result == NULL) { |
2439 | status = U_MEMORY_ALLOCATION_ERROR; |
2440 | return NULL; |
2441 | } |
2442 | |
2443 | int32_t checkBits = 0; |
2444 | UBool isPrevQuote = FALSE; |
2445 | UBool inQuote = FALSE; |
2446 | UChar textBuf[32]; |
2447 | UnicodeString text(textBuf, 0, UPRV_LENGTHOF(textBuf)); |
2448 | GMTOffsetField::FieldType itemType = GMTOffsetField::TEXT; |
2449 | int32_t itemLength = 1; |
2450 | |
2451 | for (int32_t i = 0; i < pattern.length(); i++) { |
2452 | UChar ch = pattern.charAt(i); |
2453 | if (ch == SINGLEQUOTE) { |
2454 | if (isPrevQuote) { |
2455 | text.append(SINGLEQUOTE); |
2456 | isPrevQuote = FALSE; |
2457 | } else { |
2458 | isPrevQuote = TRUE; |
2459 | if (itemType != GMTOffsetField::TEXT) { |
2460 | if (GMTOffsetField::isValid(itemType, itemLength)) { |
2461 | GMTOffsetField* fld = GMTOffsetField::createTimeField(itemType, static_cast<uint8_t>(itemLength), status); |
2462 | result->addElement(fld, status); |
2463 | if (U_FAILURE(status)) { |
2464 | break; |
2465 | } |
2466 | } else { |
2467 | status = U_ILLEGAL_ARGUMENT_ERROR; |
2468 | break; |
2469 | } |
2470 | itemType = GMTOffsetField::TEXT; |
2471 | } |
2472 | } |
2473 | inQuote = !inQuote; |
2474 | } else { |
2475 | isPrevQuote = FALSE; |
2476 | if (inQuote) { |
2477 | text.append(ch); |
2478 | } else { |
2479 | GMTOffsetField::FieldType tmpType = GMTOffsetField::getTypeByLetter(ch); |
2480 | if (tmpType != GMTOffsetField::TEXT) { |
2481 | // an offset time pattern character |
2482 | if (tmpType == itemType) { |
2483 | itemLength++; |
2484 | } else { |
2485 | if (itemType == GMTOffsetField::TEXT) { |
2486 | if (text.length() > 0) { |
2487 | GMTOffsetField* textfld = GMTOffsetField::createText(text, status); |
2488 | result->addElement(textfld, status); |
2489 | if (U_FAILURE(status)) { |
2490 | break; |
2491 | } |
2492 | text.remove(); |
2493 | } |
2494 | } else { |
2495 | if (GMTOffsetField::isValid(itemType, itemLength)) { |
2496 | GMTOffsetField* fld = GMTOffsetField::createTimeField(itemType, static_cast<uint8_t>(itemLength), status); |
2497 | result->addElement(fld, status); |
2498 | if (U_FAILURE(status)) { |
2499 | break; |
2500 | } |
2501 | } else { |
2502 | status = U_ILLEGAL_ARGUMENT_ERROR; |
2503 | break; |
2504 | } |
2505 | } |
2506 | itemType = tmpType; |
2507 | itemLength = 1; |
2508 | checkBits |= tmpType; |
2509 | } |
2510 | } else { |
2511 | // a string literal |
2512 | if (itemType != GMTOffsetField::TEXT) { |
2513 | if (GMTOffsetField::isValid(itemType, itemLength)) { |
2514 | GMTOffsetField* fld = GMTOffsetField::createTimeField(itemType, static_cast<uint8_t>(itemLength), status); |
2515 | result->addElement(fld, status); |
2516 | if (U_FAILURE(status)) { |
2517 | break; |
2518 | } |
2519 | } else { |
2520 | status = U_ILLEGAL_ARGUMENT_ERROR; |
2521 | break; |
2522 | } |
2523 | itemType = GMTOffsetField::TEXT; |
2524 | } |
2525 | text.append(ch); |
2526 | } |
2527 | } |
2528 | } |
2529 | } |
2530 | // handle last item |
2531 | if (U_SUCCESS(status)) { |
2532 | if (itemType == GMTOffsetField::TEXT) { |
2533 | if (text.length() > 0) { |
2534 | GMTOffsetField* tfld = GMTOffsetField::createText(text, status); |
2535 | result->addElement(tfld, status); |
2536 | } |
2537 | } else { |
2538 | if (GMTOffsetField::isValid(itemType, itemLength)) { |
2539 | GMTOffsetField* fld = GMTOffsetField::createTimeField(itemType, static_cast<uint8_t>(itemLength), status); |
2540 | result->addElement(fld, status); |
2541 | } else { |
2542 | status = U_ILLEGAL_ARGUMENT_ERROR; |
2543 | } |
2544 | } |
2545 | |
2546 | // Check all required fields are set |
2547 | if (U_SUCCESS(status)) { |
2548 | int32_t reqBits = 0; |
2549 | switch (required) { |
2550 | case FIELDS_H: |
2551 | reqBits = GMTOffsetField::HOUR; |
2552 | break; |
2553 | case FIELDS_HM: |
2554 | reqBits = GMTOffsetField::HOUR | GMTOffsetField::MINUTE; |
2555 | break; |
2556 | case FIELDS_HMS: |
2557 | reqBits = GMTOffsetField::HOUR | GMTOffsetField::MINUTE | GMTOffsetField::SECOND; |
2558 | break; |
2559 | } |
2560 | if (checkBits == reqBits) { |
2561 | // all required fields are set, no extra fields |
2562 | return result; |
2563 | } |
2564 | } |
2565 | } |
2566 | |
2567 | // error |
2568 | delete result; |
2569 | return NULL; |
2570 | } |
2571 | |
2572 | UnicodeString& |
2573 | TimeZoneFormat::expandOffsetPattern(const UnicodeString& offsetHM, UnicodeString& result, UErrorCode& status) { |
2574 | result.setToBogus(); |
2575 | if (U_FAILURE(status)) { |
2576 | return result; |
2577 | } |
2578 | U_ASSERT(u_strlen(DEFAULT_GMT_OFFSET_MINUTE_PATTERN) == 2); |
2579 | |
2580 | int32_t idx_mm = offsetHM.indexOf(DEFAULT_GMT_OFFSET_MINUTE_PATTERN, 2, 0); |
2581 | if (idx_mm < 0) { |
2582 | // Bad time zone hour pattern data |
2583 | status = U_ILLEGAL_ARGUMENT_ERROR; |
2584 | return result; |
2585 | } |
2586 | |
2587 | UnicodeString sep; |
2588 | int32_t idx_H = offsetHM.tempSubString(0, idx_mm).lastIndexOf((UChar)0x0048 /* H */); |
2589 | if (idx_H >= 0) { |
2590 | sep = offsetHM.tempSubString(idx_H + 1, idx_mm - (idx_H + 1)); |
2591 | } |
2592 | result.setTo(offsetHM.tempSubString(0, idx_mm + 2)); |
2593 | result.append(sep); |
2594 | result.append(DEFAULT_GMT_OFFSET_SECOND_PATTERN, -1); |
2595 | result.append(offsetHM.tempSubString(idx_mm + 2)); |
2596 | return result; |
2597 | } |
2598 | |
2599 | UnicodeString& |
2600 | TimeZoneFormat::truncateOffsetPattern(const UnicodeString& offsetHM, UnicodeString& result, UErrorCode& status) { |
2601 | result.setToBogus(); |
2602 | if (U_FAILURE(status)) { |
2603 | return result; |
2604 | } |
2605 | U_ASSERT(u_strlen(DEFAULT_GMT_OFFSET_MINUTE_PATTERN) == 2); |
2606 | |
2607 | int32_t idx_mm = offsetHM.indexOf(DEFAULT_GMT_OFFSET_MINUTE_PATTERN, 2, 0); |
2608 | if (idx_mm < 0) { |
2609 | // Bad time zone hour pattern data |
2610 | status = U_ILLEGAL_ARGUMENT_ERROR; |
2611 | return result; |
2612 | } |
2613 | UChar HH[] = {0x0048, 0x0048}; |
2614 | int32_t idx_HH = offsetHM.tempSubString(0, idx_mm).lastIndexOf(HH, 2, 0); |
2615 | if (idx_HH >= 0) { |
2616 | return result.setTo(offsetHM.tempSubString(0, idx_HH + 2)); |
2617 | } |
2618 | int32_t idx_H = offsetHM.tempSubString(0, idx_mm).lastIndexOf((UChar)0x0048, 0); |
2619 | if (idx_H >= 0) { |
2620 | return result.setTo(offsetHM.tempSubString(0, idx_H + 1)); |
2621 | } |
2622 | // Bad time zone hour pattern data |
2623 | status = U_ILLEGAL_ARGUMENT_ERROR; |
2624 | return result; |
2625 | } |
2626 | |
2627 | void |
2628 | TimeZoneFormat::initGMTOffsetPatterns(UErrorCode& status) { |
2629 | for (int32_t type = 0; type < UTZFMT_PAT_COUNT; type++) { |
2630 | switch (type) { |
2631 | case UTZFMT_PAT_POSITIVE_H: |
2632 | case UTZFMT_PAT_NEGATIVE_H: |
2633 | fGMTOffsetPatternItems[type] = parseOffsetPattern(fGMTOffsetPatterns[type], FIELDS_H, status); |
2634 | break; |
2635 | case UTZFMT_PAT_POSITIVE_HM: |
2636 | case UTZFMT_PAT_NEGATIVE_HM: |
2637 | fGMTOffsetPatternItems[type] = parseOffsetPattern(fGMTOffsetPatterns[type], FIELDS_HM, status); |
2638 | break; |
2639 | case UTZFMT_PAT_POSITIVE_HMS: |
2640 | case UTZFMT_PAT_NEGATIVE_HMS: |
2641 | fGMTOffsetPatternItems[type] = parseOffsetPattern(fGMTOffsetPatterns[type], FIELDS_HMS, status); |
2642 | break; |
2643 | } |
2644 | } |
2645 | checkAbuttingHoursAndMinutes(); |
2646 | } |
2647 | |
2648 | void |
2649 | TimeZoneFormat::checkAbuttingHoursAndMinutes() { |
2650 | fAbuttingOffsetHoursAndMinutes= FALSE; |
2651 | for (int32_t type = 0; type < UTZFMT_PAT_COUNT; type++) { |
2652 | UBool afterH = FALSE; |
2653 | UVector *items = fGMTOffsetPatternItems[type]; |
2654 | for (int32_t i = 0; i < items->size(); i++) { |
2655 | const GMTOffsetField* item = (GMTOffsetField*)items->elementAt(i); |
2656 | GMTOffsetField::FieldType fieldType = item->getType(); |
2657 | if (fieldType != GMTOffsetField::TEXT) { |
2658 | if (afterH) { |
2659 | fAbuttingOffsetHoursAndMinutes = TRUE; |
2660 | break; |
2661 | } else if (fieldType == GMTOffsetField::HOUR) { |
2662 | afterH = TRUE; |
2663 | } |
2664 | } else if (afterH) { |
2665 | break; |
2666 | } |
2667 | } |
2668 | if (fAbuttingOffsetHoursAndMinutes) { |
2669 | break; |
2670 | } |
2671 | } |
2672 | } |
2673 | |
2674 | UBool |
2675 | TimeZoneFormat::toCodePoints(const UnicodeString& str, UChar32* codeArray, int32_t size) { |
2676 | int32_t count = str.countChar32(); |
2677 | if (count != size) { |
2678 | return FALSE; |
2679 | } |
2680 | |
2681 | for (int32_t idx = 0, start = 0; idx < size; idx++) { |
2682 | codeArray[idx] = str.char32At(start); |
2683 | start = str.moveIndex32(start, 1); |
2684 | } |
2685 | |
2686 | return TRUE; |
2687 | } |
2688 | |
2689 | TimeZone* |
2690 | TimeZoneFormat::createTimeZoneForOffset(int32_t offset) const { |
2691 | if (offset == 0) { |
2692 | // when offset is 0, we should use "Etc/GMT" |
2693 | return TimeZone::createTimeZone(UnicodeString(TRUE, TZID_GMT, -1)); |
2694 | } |
2695 | return ZoneMeta::createCustomTimeZone(offset); |
2696 | } |
2697 | |
2698 | UTimeZoneFormatTimeType |
2699 | TimeZoneFormat::getTimeType(UTimeZoneNameType nameType) { |
2700 | switch (nameType) { |
2701 | case UTZNM_LONG_STANDARD: |
2702 | case UTZNM_SHORT_STANDARD: |
2703 | return UTZFMT_TIME_TYPE_STANDARD; |
2704 | |
2705 | case UTZNM_LONG_DAYLIGHT: |
2706 | case UTZNM_SHORT_DAYLIGHT: |
2707 | return UTZFMT_TIME_TYPE_DAYLIGHT; |
2708 | |
2709 | default: |
2710 | return UTZFMT_TIME_TYPE_UNKNOWN; |
2711 | } |
2712 | } |
2713 | |
2714 | UnicodeString& |
2715 | TimeZoneFormat::getTimeZoneID(const TimeZoneNames::MatchInfoCollection* matches, int32_t idx, UnicodeString& tzID) const { |
2716 | if (!matches->getTimeZoneIDAt(idx, tzID)) { |
2717 | UChar mzIDBuf[32]; |
2718 | UnicodeString mzID(mzIDBuf, 0, UPRV_LENGTHOF(mzIDBuf)); |
2719 | if (matches->getMetaZoneIDAt(idx, mzID)) { |
2720 | fTimeZoneNames->getReferenceZoneID(mzID, fTargetRegion, tzID); |
2721 | } |
2722 | } |
2723 | return tzID; |
2724 | } |
2725 | |
2726 | |
2727 | class ZoneIdMatchHandler : public TextTrieMapSearchResultHandler { |
2728 | public: |
2729 | ZoneIdMatchHandler(); |
2730 | virtual ~ZoneIdMatchHandler(); |
2731 | |
2732 | UBool handleMatch(int32_t matchLength, const CharacterNode *node, UErrorCode &status); |
2733 | const UChar* getID(); |
2734 | int32_t getMatchLen(); |
2735 | private: |
2736 | int32_t fLen; |
2737 | const UChar* fID; |
2738 | }; |
2739 | |
2740 | ZoneIdMatchHandler::ZoneIdMatchHandler() |
2741 | : fLen(0), fID(NULL) { |
2742 | } |
2743 | |
2744 | ZoneIdMatchHandler::~ZoneIdMatchHandler() { |
2745 | } |
2746 | |
2747 | UBool |
2748 | ZoneIdMatchHandler::handleMatch(int32_t matchLength, const CharacterNode *node, UErrorCode &status) { |
2749 | if (U_FAILURE(status)) { |
2750 | return FALSE; |
2751 | } |
2752 | if (node->hasValues()) { |
2753 | const UChar* id = (const UChar*)node->getValue(0); |
2754 | if (id != NULL) { |
2755 | if (fLen < matchLength) { |
2756 | fID = id; |
2757 | fLen = matchLength; |
2758 | } |
2759 | } |
2760 | } |
2761 | return TRUE; |
2762 | } |
2763 | |
2764 | const UChar* |
2765 | ZoneIdMatchHandler::getID() { |
2766 | return fID; |
2767 | } |
2768 | |
2769 | int32_t |
2770 | ZoneIdMatchHandler::getMatchLen() { |
2771 | return fLen; |
2772 | } |
2773 | |
2774 | |
2775 | static void U_CALLCONV initZoneIdTrie(UErrorCode &status) { |
2776 | U_ASSERT(gZoneIdTrie == NULL); |
2777 | ucln_i18n_registerCleanup(UCLN_I18N_TIMEZONEFORMAT, tzfmt_cleanup); |
2778 | gZoneIdTrie = new TextTrieMap(TRUE, NULL); // No deleter, because values are pooled by ZoneMeta |
2779 | if (gZoneIdTrie == NULL) { |
2780 | status = U_MEMORY_ALLOCATION_ERROR; |
2781 | return; |
2782 | } |
2783 | StringEnumeration *tzenum = TimeZone::createEnumeration(); |
2784 | const UnicodeString *id; |
2785 | while ((id = tzenum->snext(status)) != NULL) { |
2786 | const UChar* uid = ZoneMeta::findTimeZoneID(*id); |
2787 | if (uid) { |
2788 | gZoneIdTrie->put(uid, const_cast<UChar *>(uid), status); |
2789 | } |
2790 | } |
2791 | delete tzenum; |
2792 | } |
2793 | |
2794 | |
2795 | UnicodeString& |
2796 | TimeZoneFormat::parseZoneID(const UnicodeString& text, ParsePosition& pos, UnicodeString& tzID) const { |
2797 | UErrorCode status = U_ZERO_ERROR; |
2798 | umtx_initOnce(gZoneIdTrieInitOnce, &initZoneIdTrie, status); |
2799 | |
2800 | int32_t start = pos.getIndex(); |
2801 | int32_t len = 0; |
2802 | tzID.setToBogus(); |
2803 | |
2804 | if (U_SUCCESS(status)) { |
2805 | LocalPointer<ZoneIdMatchHandler> handler(new ZoneIdMatchHandler()); |
2806 | gZoneIdTrie->search(text, start, handler.getAlias(), status); |
2807 | len = handler->getMatchLen(); |
2808 | if (len > 0) { |
2809 | tzID.setTo(handler->getID(), -1); |
2810 | } |
2811 | } |
2812 | |
2813 | if (len > 0) { |
2814 | pos.setIndex(start + len); |
2815 | } else { |
2816 | pos.setErrorIndex(start); |
2817 | } |
2818 | |
2819 | return tzID; |
2820 | } |
2821 | |
2822 | static void U_CALLCONV initShortZoneIdTrie(UErrorCode &status) { |
2823 | U_ASSERT(gShortZoneIdTrie == NULL); |
2824 | ucln_i18n_registerCleanup(UCLN_I18N_TIMEZONEFORMAT, tzfmt_cleanup); |
2825 | StringEnumeration *tzenum = TimeZone::createTimeZoneIDEnumeration(UCAL_ZONE_TYPE_CANONICAL, NULL, NULL, status); |
2826 | if (U_SUCCESS(status)) { |
2827 | gShortZoneIdTrie = new TextTrieMap(TRUE, NULL); // No deleter, because values are pooled by ZoneMeta |
2828 | if (gShortZoneIdTrie == NULL) { |
2829 | status = U_MEMORY_ALLOCATION_ERROR; |
2830 | } else { |
2831 | const UnicodeString *id; |
2832 | while ((id = tzenum->snext(status)) != NULL) { |
2833 | const UChar* uID = ZoneMeta::findTimeZoneID(*id); |
2834 | const UChar* shortID = ZoneMeta::getShortID(*id); |
2835 | if (shortID && uID) { |
2836 | gShortZoneIdTrie->put(shortID, const_cast<UChar *>(uID), status); |
2837 | } |
2838 | } |
2839 | } |
2840 | } |
2841 | delete tzenum; |
2842 | } |
2843 | |
2844 | |
2845 | UnicodeString& |
2846 | TimeZoneFormat::parseShortZoneID(const UnicodeString& text, ParsePosition& pos, UnicodeString& tzID) const { |
2847 | UErrorCode status = U_ZERO_ERROR; |
2848 | umtx_initOnce(gShortZoneIdTrieInitOnce, &initShortZoneIdTrie, status); |
2849 | |
2850 | int32_t start = pos.getIndex(); |
2851 | int32_t len = 0; |
2852 | tzID.setToBogus(); |
2853 | |
2854 | if (U_SUCCESS(status)) { |
2855 | LocalPointer<ZoneIdMatchHandler> handler(new ZoneIdMatchHandler()); |
2856 | gShortZoneIdTrie->search(text, start, handler.getAlias(), status); |
2857 | len = handler->getMatchLen(); |
2858 | if (len > 0) { |
2859 | tzID.setTo(handler->getID(), -1); |
2860 | } |
2861 | } |
2862 | |
2863 | if (len > 0) { |
2864 | pos.setIndex(start + len); |
2865 | } else { |
2866 | pos.setErrorIndex(start); |
2867 | } |
2868 | |
2869 | return tzID; |
2870 | } |
2871 | |
2872 | |
2873 | UnicodeString& |
2874 | TimeZoneFormat::parseExemplarLocation(const UnicodeString& text, ParsePosition& pos, UnicodeString& tzID) const { |
2875 | int32_t startIdx = pos.getIndex(); |
2876 | int32_t parsedPos = -1; |
2877 | tzID.setToBogus(); |
2878 | |
2879 | UErrorCode status = U_ZERO_ERROR; |
2880 | LocalPointer<TimeZoneNames::MatchInfoCollection> exemplarMatches(fTimeZoneNames->find(text, startIdx, UTZNM_EXEMPLAR_LOCATION, status)); |
2881 | if (U_FAILURE(status)) { |
2882 | pos.setErrorIndex(startIdx); |
2883 | return tzID; |
2884 | } |
2885 | int32_t matchIdx = -1; |
2886 | if (!exemplarMatches.isNull()) { |
2887 | for (int32_t i = 0; i < exemplarMatches->size(); i++) { |
2888 | if (startIdx + exemplarMatches->getMatchLengthAt(i) > parsedPos) { |
2889 | matchIdx = i; |
2890 | parsedPos = startIdx + exemplarMatches->getMatchLengthAt(i); |
2891 | } |
2892 | } |
2893 | if (parsedPos > 0) { |
2894 | pos.setIndex(parsedPos); |
2895 | getTimeZoneID(exemplarMatches.getAlias(), matchIdx, tzID); |
2896 | } |
2897 | } |
2898 | |
2899 | if (tzID.length() == 0) { |
2900 | pos.setErrorIndex(startIdx); |
2901 | } |
2902 | |
2903 | return tzID; |
2904 | } |
2905 | |
2906 | U_NAMESPACE_END |
2907 | |
2908 | #endif |
2909 | |