]>
Commit | Line | Data |
---|---|---|
4388f060 A |
1 | /* |
2 | ******************************************************************************* | |
57a6839d A |
3 | * Copyright (C) 2011-2013, International Business Machines Corporation and |
4 | * others. All Rights Reserved. | |
4388f060 A |
5 | ******************************************************************************* |
6 | */ | |
7 | ||
8 | #include "unicode/utypes.h" | |
9 | ||
10 | #if !UCONFIG_NO_FORMATTING | |
11 | ||
12 | #include "unicode/calendar.h" | |
13 | #include "unicode/tzfmt.h" | |
14 | #include "unicode/numsys.h" | |
15 | #include "unicode/uchar.h" | |
16 | #include "unicode/udat.h" | |
17 | #include "tzgnames.h" | |
18 | #include "cmemory.h" | |
19 | #include "cstring.h" | |
20 | #include "putilimp.h" | |
21 | #include "uassert.h" | |
22 | #include "ucln_in.h" | |
23 | #include "umutex.h" | |
24 | #include "uresimp.h" | |
25 | #include "ureslocs.h" | |
26 | #include "uvector.h" | |
27 | #include "zonemeta.h" | |
51004dcb | 28 | #include "tznames_impl.h" // TextTrieMap |
4388f060 A |
29 | |
30 | U_NAMESPACE_BEGIN | |
31 | ||
51004dcb A |
32 | // Bit flags used by the parse method. |
33 | // The order must match UTimeZoneFormatStyle enum. | |
34 | #define ISO_Z_STYLE_FLAG 0x0080 | |
35 | #define ISO_LOCAL_STYLE_FLAG 0x0100 | |
36 | static const int16_t STYLE_PARSE_FLAGS[] = { | |
37 | 0x0001, // UTZFMT_STYLE_GENERIC_LOCATION, | |
38 | 0x0002, // UTZFMT_STYLE_GENERIC_LONG, | |
39 | 0x0004, // UTZFMT_STYLE_GENERIC_SHORT, | |
40 | 0x0008, // UTZFMT_STYLE_SPECIFIC_LONG, | |
41 | 0x0010, // UTZFMT_STYLE_SPECIFIC_SHORT, | |
42 | 0x0020, // UTZFMT_STYLE_LOCALIZED_GMT, | |
43 | 0x0040, // UTZFMT_STYLE_LOCALIZED_GMT_SHORT, | |
44 | ISO_Z_STYLE_FLAG, // UTZFMT_STYLE_ISO_BASIC_SHORT, | |
45 | ISO_LOCAL_STYLE_FLAG, // UTZFMT_STYLE_ISO_BASIC_LOCAL_SHORT, | |
46 | ISO_Z_STYLE_FLAG, // UTZFMT_STYLE_ISO_BASIC_FIXED, | |
47 | ISO_LOCAL_STYLE_FLAG, // UTZFMT_STYLE_ISO_BASIC_LOCAL_FIXED, | |
48 | ISO_Z_STYLE_FLAG, // UTZFMT_STYLE_ISO_BASIC_FULL, | |
49 | ISO_LOCAL_STYLE_FLAG, // UTZFMT_STYLE_ISO_BASIC_LOCAL_FULL, | |
50 | ISO_Z_STYLE_FLAG, // UTZFMT_STYLE_ISO_EXTENDED_FIXED, | |
51 | ISO_LOCAL_STYLE_FLAG, // UTZFMT_STYLE_ISO_EXTENDED_LOCAL_FIXED, | |
52 | ISO_Z_STYLE_FLAG, // UTZFMT_STYLE_ISO_EXTENDED_FULL, | |
53 | ISO_LOCAL_STYLE_FLAG, // UTZFMT_STYLE_ISO_EXTENDED_LOCAL_FULL, | |
54 | 0x0200, // UTZFMT_STYLE_ZONE_ID, | |
55 | 0x0400, // UTZFMT_STYLE_ZONE_ID_SHORT, | |
56 | 0x0800 // UTZFMT_STYLE_EXEMPLAR_LOCATION | |
57 | }; | |
58 | ||
4388f060 A |
59 | static const char gZoneStringsTag[] = "zoneStrings"; |
60 | static const char gGmtFormatTag[]= "gmtFormat"; | |
61 | static const char gGmtZeroFormatTag[] = "gmtZeroFormat"; | |
62 | static const char gHourFormatTag[]= "hourFormat"; | |
63 | ||
64 | static const UChar TZID_GMT[] = {0x0045, 0x0074, 0x0063, 0x002F, 0x0047, 0x004D, 0x0054, 0}; // Etc/GMT | |
51004dcb A |
65 | static const UChar UNKNOWN_ZONE_ID[] = { |
66 | 0x0045, 0x0074, 0x0063, 0x002F, 0x0055, 0x006E, 0x006B, 0x006E, 0x006F, 0x0077, 0x006E, 0}; // Etc/Unknown | |
67 | static const UChar UNKNOWN_SHORT_ZONE_ID[] = {0x0075, 0x006E, 0x006B, 0}; // unk | |
68 | static const UChar UNKNOWN_LOCATION[] = {0x0055, 0x006E, 0x006B, 0x006E, 0x006F, 0x0077, 0x006E, 0}; // Unknown | |
4388f060 A |
69 | |
70 | static const UChar DEFAULT_GMT_PATTERN[] = {0x0047, 0x004D, 0x0054, 0x007B, 0x0030, 0x007D, 0}; // GMT{0} | |
51004dcb A |
71 | //static const UChar DEFAULT_GMT_ZERO[] = {0x0047, 0x004D, 0x0054, 0}; // GMT |
72 | static const UChar DEFAULT_GMT_POSITIVE_HM[] = {0x002B, 0x0048, 0x003A, 0x006D, 0x006D, 0}; // +H:mm | |
73 | static const UChar DEFAULT_GMT_POSITIVE_HMS[] = {0x002B, 0x0048, 0x003A, 0x006D, 0x006D, 0x003A, 0x0073, 0x0073, 0}; // +H:mm:ss | |
74 | static const UChar DEFAULT_GMT_NEGATIVE_HM[] = {0x002D, 0x0048, 0x003A, 0x006D, 0x006D, 0}; // -H:mm | |
75 | static const UChar DEFAULT_GMT_NEGATIVE_HMS[] = {0x002D, 0x0048, 0x003A, 0x006D, 0x006D, 0x003A, 0x0073, 0x0073, 0}; // -H:mm:ss | |
76 | static const UChar DEFAULT_GMT_POSITIVE_H[] = {0x002B, 0x0048, 0}; // +H | |
77 | static const UChar DEFAULT_GMT_NEGATIVE_H[] = {0x002D, 0x0048, 0}; // -H | |
4388f060 A |
78 | |
79 | static const UChar32 DEFAULT_GMT_DIGITS[] = { | |
80 | 0x0030, 0x0031, 0x0032, 0x0033, 0x0034, | |
81 | 0x0035, 0x0036, 0x0037, 0x0038, 0x0039 | |
82 | }; | |
83 | ||
84 | static const UChar DEFAULT_GMT_OFFSET_SEP = 0x003A; // ':' | |
85 | ||
86 | static const UChar ARG0[] = {0x007B, 0x0030, 0x007D}; // "{0}" | |
51004dcb | 87 | static const int32_t ARG0_LEN = 3; |
4388f060 A |
88 | |
89 | static const UChar DEFAULT_GMT_OFFSET_MINUTE_PATTERN[] = {0x006D, 0x006D, 0}; // "mm" | |
90 | static const UChar DEFAULT_GMT_OFFSET_SECOND_PATTERN[] = {0x0073, 0x0073, 0}; // "ss" | |
91 | ||
92 | static const UChar ALT_GMT_STRINGS[][4] = { | |
93 | {0x0047, 0x004D, 0x0054, 0}, // GMT | |
94 | {0x0055, 0x0054, 0x0043, 0}, // UTC | |
95 | {0x0055, 0x0054, 0, 0}, // UT | |
96 | {0, 0, 0, 0} | |
97 | }; | |
98 | ||
99 | // Order of GMT offset pattern parsing, *_HMS must be evaluated first | |
100 | // because *_HM is most likely a substring of *_HMS | |
101 | static const int32_t PARSE_GMT_OFFSET_TYPES[] = { | |
102 | UTZFMT_PAT_POSITIVE_HMS, | |
103 | UTZFMT_PAT_NEGATIVE_HMS, | |
104 | UTZFMT_PAT_POSITIVE_HM, | |
105 | UTZFMT_PAT_NEGATIVE_HM, | |
51004dcb A |
106 | UTZFMT_PAT_POSITIVE_H, |
107 | UTZFMT_PAT_NEGATIVE_H, | |
4388f060 A |
108 | -1 |
109 | }; | |
110 | ||
111 | static const UChar SINGLEQUOTE = 0x0027; | |
112 | static const UChar PLUS = 0x002B; | |
113 | static const UChar MINUS = 0x002D; | |
114 | static const UChar ISO8601_UTC = 0x005A; // 'Z' | |
115 | static const UChar ISO8601_SEP = 0x003A; // ':' | |
116 | ||
117 | static const int32_t MILLIS_PER_HOUR = 60 * 60 * 1000; | |
118 | static const int32_t MILLIS_PER_MINUTE = 60 * 1000; | |
119 | static const int32_t MILLIS_PER_SECOND = 1000; | |
120 | ||
121 | // Maximum offset (exclusive) in millisecond supported by offset formats | |
122 | static int32_t MAX_OFFSET = 24 * MILLIS_PER_HOUR; | |
123 | ||
124 | // Maximum values for GMT offset fields | |
125 | static const int32_t MAX_OFFSET_HOUR = 23; | |
126 | static const int32_t MAX_OFFSET_MINUTE = 59; | |
127 | static const int32_t MAX_OFFSET_SECOND = 59; | |
128 | ||
129 | static const int32_t UNKNOWN_OFFSET = 0x7FFFFFFF; | |
130 | ||
51004dcb | 131 | static const int32_t ALL_SIMPLE_NAME_TYPES = UTZNM_LONG_STANDARD | UTZNM_LONG_DAYLIGHT | UTZNM_SHORT_STANDARD | UTZNM_SHORT_DAYLIGHT | UTZNM_EXEMPLAR_LOCATION; |
4388f060 A |
132 | static const int32_t ALL_GENERIC_NAME_TYPES = UTZGNM_LOCATION | UTZGNM_LONG | UTZGNM_SHORT; |
133 | ||
4388f060 A |
134 | #define DIGIT_VAL(c) (0x0030 <= (c) && (c) <= 0x0039 ? (c) - 0x0030 : -1) |
135 | #define MAX_OFFSET_DIGITS 6 | |
136 | ||
51004dcb A |
137 | // Time Zone ID/Short ID trie |
138 | static TextTrieMap *gZoneIdTrie = NULL; | |
57a6839d | 139 | static icu::UInitOnce gZoneIdTrieInitOnce = U_INITONCE_INITIALIZER; |
51004dcb A |
140 | |
141 | static TextTrieMap *gShortZoneIdTrie = NULL; | |
57a6839d | 142 | static icu::UInitOnce gShortZoneIdTrieInitOnce = U_INITONCE_INITIALIZER; |
51004dcb A |
143 | |
144 | static UMutex gLock = U_MUTEX_INITIALIZER; | |
145 | ||
146 | U_CDECL_BEGIN | |
147 | /** | |
148 | * Cleanup callback func | |
149 | */ | |
150 | static UBool U_CALLCONV tzfmt_cleanup(void) | |
151 | { | |
152 | if (gZoneIdTrie != NULL) { | |
153 | delete gZoneIdTrie; | |
154 | } | |
155 | gZoneIdTrie = NULL; | |
57a6839d | 156 | gZoneIdTrieInitOnce.reset(); |
51004dcb A |
157 | |
158 | if (gShortZoneIdTrie != NULL) { | |
159 | delete gShortZoneIdTrie; | |
160 | } | |
161 | gShortZoneIdTrie = NULL; | |
57a6839d | 162 | gShortZoneIdTrieInitOnce.reset(); |
51004dcb A |
163 | |
164 | return TRUE; | |
165 | } | |
166 | U_CDECL_END | |
4388f060 A |
167 | |
168 | // ------------------------------------------------------------------ | |
169 | // GMTOffsetField | |
170 | // | |
171 | // This class represents a localized GMT offset pattern | |
172 | // item and used by TimeZoneFormat | |
173 | // ------------------------------------------------------------------ | |
174 | class GMTOffsetField : public UMemory { | |
175 | public: | |
176 | enum FieldType { | |
177 | TEXT = 0, | |
178 | HOUR = 1, | |
179 | MINUTE = 2, | |
180 | SECOND = 4 | |
181 | }; | |
182 | ||
183 | virtual ~GMTOffsetField(); | |
184 | ||
185 | static GMTOffsetField* createText(const UnicodeString& text, UErrorCode& status); | |
186 | static GMTOffsetField* createTimeField(FieldType type, uint8_t width, UErrorCode& status); | |
187 | static UBool isValid(FieldType type, int32_t width); | |
188 | static FieldType getTypeByLetter(UChar ch); | |
189 | ||
190 | FieldType getType() const; | |
191 | uint8_t getWidth() const; | |
192 | const UChar* getPatternText(void) const; | |
193 | ||
194 | private: | |
195 | UChar* fText; | |
196 | FieldType fType; | |
197 | uint8_t fWidth; | |
198 | ||
199 | GMTOffsetField(); | |
200 | }; | |
201 | ||
202 | GMTOffsetField::GMTOffsetField() | |
203 | : fText(NULL), fType(TEXT), fWidth(0) { | |
204 | } | |
205 | ||
206 | GMTOffsetField::~GMTOffsetField() { | |
207 | if (fText) { | |
208 | uprv_free(fText); | |
209 | } | |
210 | } | |
211 | ||
212 | GMTOffsetField* | |
213 | GMTOffsetField::createText(const UnicodeString& text, UErrorCode& status) { | |
214 | if (U_FAILURE(status)) { | |
215 | return NULL; | |
216 | } | |
217 | GMTOffsetField* result = new GMTOffsetField(); | |
218 | if (result == NULL) { | |
219 | status = U_MEMORY_ALLOCATION_ERROR; | |
220 | return NULL; | |
221 | } | |
222 | ||
223 | int32_t len = text.length(); | |
224 | result->fText = (UChar*)uprv_malloc((len + 1) * sizeof(UChar)); | |
225 | if (result->fText == NULL) { | |
226 | status = U_MEMORY_ALLOCATION_ERROR; | |
227 | delete result; | |
228 | return NULL; | |
229 | } | |
230 | u_strncpy(result->fText, text.getBuffer(), len); | |
231 | result->fText[len] = 0; | |
232 | result->fType = TEXT; | |
233 | ||
234 | return result; | |
235 | } | |
236 | ||
237 | GMTOffsetField* | |
238 | GMTOffsetField::createTimeField(FieldType type, uint8_t width, UErrorCode& status) { | |
239 | U_ASSERT(type != TEXT); | |
240 | if (U_FAILURE(status)) { | |
241 | return NULL; | |
242 | } | |
243 | GMTOffsetField* result = new GMTOffsetField(); | |
244 | if (result == NULL) { | |
245 | status = U_MEMORY_ALLOCATION_ERROR; | |
246 | return NULL; | |
247 | } | |
248 | ||
249 | result->fType = type; | |
250 | result->fWidth = width; | |
251 | ||
252 | return result; | |
253 | } | |
254 | ||
255 | UBool | |
256 | GMTOffsetField::isValid(FieldType type, int32_t width) { | |
257 | switch (type) { | |
258 | case HOUR: | |
259 | return (width == 1 || width == 2); | |
260 | case MINUTE: | |
261 | case SECOND: | |
262 | return (width == 2); | |
263 | default: | |
264 | U_ASSERT(FALSE); | |
265 | } | |
266 | return (width > 0); | |
267 | } | |
268 | ||
269 | GMTOffsetField::FieldType | |
270 | GMTOffsetField::getTypeByLetter(UChar ch) { | |
271 | if (ch == 0x0048 /* H */) { | |
272 | return HOUR; | |
273 | } else if (ch == 0x006D /* m */) { | |
274 | return MINUTE; | |
275 | } else if (ch == 0x0073 /* s */) { | |
276 | return SECOND; | |
277 | } | |
278 | return TEXT; | |
279 | } | |
280 | ||
281 | inline GMTOffsetField::FieldType | |
282 | GMTOffsetField::getType() const { | |
283 | return fType; | |
284 | } | |
285 | ||
286 | inline uint8_t | |
287 | GMTOffsetField::getWidth() const { | |
288 | return fWidth; | |
289 | } | |
290 | ||
291 | inline const UChar* | |
292 | GMTOffsetField::getPatternText(void) const { | |
293 | return fText; | |
294 | } | |
295 | ||
296 | ||
297 | U_CDECL_BEGIN | |
298 | static void U_CALLCONV | |
299 | deleteGMTOffsetField(void *obj) { | |
300 | delete static_cast<GMTOffsetField *>(obj); | |
301 | } | |
302 | U_CDECL_END | |
303 | ||
304 | ||
305 | // ------------------------------------------------------------------ | |
306 | // TimeZoneFormat | |
307 | // ------------------------------------------------------------------ | |
308 | UOBJECT_DEFINE_RTTI_IMPLEMENTATION(TimeZoneFormat) | |
309 | ||
310 | TimeZoneFormat::TimeZoneFormat(const Locale& locale, UErrorCode& status) | |
51004dcb | 311 | : fLocale(locale), fTimeZoneNames(NULL), fTimeZoneGenericNames(NULL), fDefParseOptionFlags(0) { |
4388f060 | 312 | |
51004dcb | 313 | for (int32_t i = 0; i < UTZFMT_PAT_COUNT; i++) { |
4388f060 A |
314 | fGMTOffsetPatternItems[i] = NULL; |
315 | } | |
316 | ||
317 | const char* region = fLocale.getCountry(); | |
318 | int32_t regionLen = uprv_strlen(region); | |
319 | if (regionLen == 0) { | |
320 | char loc[ULOC_FULLNAME_CAPACITY]; | |
321 | uloc_addLikelySubtags(fLocale.getName(), loc, sizeof(loc), &status); | |
322 | ||
323 | regionLen = uloc_getCountry(loc, fTargetRegion, sizeof(fTargetRegion), &status); | |
324 | if (U_SUCCESS(status)) { | |
325 | fTargetRegion[regionLen] = 0; | |
326 | } else { | |
327 | return; | |
328 | } | |
329 | } else if (regionLen < (int32_t)sizeof(fTargetRegion)) { | |
330 | uprv_strcpy(fTargetRegion, region); | |
331 | } else { | |
332 | fTargetRegion[0] = 0; | |
333 | } | |
334 | ||
335 | fTimeZoneNames = TimeZoneNames::createInstance(locale, status); | |
336 | // fTimeZoneGenericNames is lazily instantiated | |
51004dcb A |
337 | if (U_FAILURE(status)) { |
338 | return; | |
339 | } | |
4388f060 A |
340 | |
341 | const UChar* gmtPattern = NULL; | |
342 | const UChar* hourFormats = NULL; | |
343 | ||
344 | UResourceBundle *zoneBundle = ures_open(U_ICUDATA_ZONE, locale.getName(), &status); | |
345 | UResourceBundle *zoneStringsArray = ures_getByKeyWithFallback(zoneBundle, gZoneStringsTag, NULL, &status); | |
346 | if (U_SUCCESS(status)) { | |
347 | const UChar* resStr; | |
348 | int32_t len; | |
349 | resStr = ures_getStringByKeyWithFallback(zoneStringsArray, gGmtFormatTag, &len, &status); | |
350 | if (len > 0) { | |
351 | gmtPattern = resStr; | |
352 | } | |
353 | resStr = ures_getStringByKeyWithFallback(zoneStringsArray, gGmtZeroFormatTag, &len, &status); | |
354 | if (len > 0) { | |
355 | fGMTZeroFormat.setTo(TRUE, resStr, len); | |
356 | } | |
357 | resStr = ures_getStringByKeyWithFallback(zoneStringsArray, gHourFormatTag, &len, &status); | |
358 | if (len > 0) { | |
359 | hourFormats = resStr; | |
360 | } | |
361 | ures_close(zoneStringsArray); | |
362 | ures_close(zoneBundle); | |
363 | } | |
364 | ||
365 | if (gmtPattern == NULL) { | |
366 | gmtPattern = DEFAULT_GMT_PATTERN; | |
367 | } | |
368 | initGMTPattern(UnicodeString(gmtPattern, -1), status); | |
369 | ||
51004dcb | 370 | UBool useDefaultOffsetPatterns = TRUE; |
4388f060 A |
371 | if (hourFormats) { |
372 | UChar *sep = u_strchr(hourFormats, (UChar)0x003B /* ';' */); | |
373 | if (sep != NULL) { | |
51004dcb | 374 | UErrorCode tmpStatus = U_ZERO_ERROR; |
4388f060 A |
375 | fGMTOffsetPatterns[UTZFMT_PAT_POSITIVE_HM].setTo(FALSE, hourFormats, (int32_t)(sep - hourFormats)); |
376 | fGMTOffsetPatterns[UTZFMT_PAT_NEGATIVE_HM].setTo(TRUE, sep + 1, -1); | |
51004dcb A |
377 | expandOffsetPattern(fGMTOffsetPatterns[UTZFMT_PAT_POSITIVE_HM], fGMTOffsetPatterns[UTZFMT_PAT_POSITIVE_HMS], tmpStatus); |
378 | expandOffsetPattern(fGMTOffsetPatterns[UTZFMT_PAT_NEGATIVE_HM], fGMTOffsetPatterns[UTZFMT_PAT_NEGATIVE_HMS], tmpStatus); | |
379 | truncateOffsetPattern(fGMTOffsetPatterns[UTZFMT_PAT_POSITIVE_HM], fGMTOffsetPatterns[UTZFMT_PAT_POSITIVE_H], tmpStatus); | |
380 | truncateOffsetPattern(fGMTOffsetPatterns[UTZFMT_PAT_NEGATIVE_HM], fGMTOffsetPatterns[UTZFMT_PAT_NEGATIVE_H], tmpStatus); | |
381 | if (U_SUCCESS(tmpStatus)) { | |
382 | useDefaultOffsetPatterns = FALSE; | |
383 | } | |
4388f060 A |
384 | } |
385 | } | |
51004dcb A |
386 | if (useDefaultOffsetPatterns) { |
387 | fGMTOffsetPatterns[UTZFMT_PAT_POSITIVE_H].setTo(TRUE, DEFAULT_GMT_POSITIVE_H, -1); | |
4388f060 A |
388 | fGMTOffsetPatterns[UTZFMT_PAT_POSITIVE_HM].setTo(TRUE, DEFAULT_GMT_POSITIVE_HM, -1); |
389 | fGMTOffsetPatterns[UTZFMT_PAT_POSITIVE_HMS].setTo(TRUE, DEFAULT_GMT_POSITIVE_HMS, -1); | |
51004dcb | 390 | fGMTOffsetPatterns[UTZFMT_PAT_NEGATIVE_H].setTo(TRUE, DEFAULT_GMT_NEGATIVE_H, -1); |
4388f060 A |
391 | fGMTOffsetPatterns[UTZFMT_PAT_NEGATIVE_HM].setTo(TRUE, DEFAULT_GMT_NEGATIVE_HM, -1); |
392 | fGMTOffsetPatterns[UTZFMT_PAT_NEGATIVE_HMS].setTo(TRUE, DEFAULT_GMT_NEGATIVE_HMS, -1); | |
393 | } | |
394 | initGMTOffsetPatterns(status); | |
395 | ||
396 | NumberingSystem* ns = NumberingSystem::createInstance(locale, status); | |
397 | UBool useDefDigits = TRUE; | |
398 | if (ns && !ns->isAlgorithmic()) { | |
399 | UnicodeString digits = ns->getDescription(); | |
400 | useDefDigits = !toCodePoints(digits, fGMTOffsetDigits, 10); | |
401 | } | |
402 | if (useDefDigits) { | |
403 | uprv_memcpy(fGMTOffsetDigits, DEFAULT_GMT_DIGITS, sizeof(UChar32) * 10); | |
404 | } | |
405 | delete ns; | |
406 | } | |
407 | ||
408 | TimeZoneFormat::TimeZoneFormat(const TimeZoneFormat& other) | |
409 | : Format(other), fTimeZoneNames(NULL), fTimeZoneGenericNames(NULL) { | |
410 | ||
51004dcb | 411 | for (int32_t i = 0; i < UTZFMT_PAT_COUNT; i++) { |
4388f060 A |
412 | fGMTOffsetPatternItems[i] = NULL; |
413 | } | |
414 | *this = other; | |
415 | } | |
416 | ||
417 | ||
418 | TimeZoneFormat::~TimeZoneFormat() { | |
419 | delete fTimeZoneNames; | |
420 | delete fTimeZoneGenericNames; | |
51004dcb | 421 | for (int32_t i = 0; i < UTZFMT_PAT_COUNT; i++) { |
4388f060 A |
422 | delete fGMTOffsetPatternItems[i]; |
423 | } | |
4388f060 A |
424 | } |
425 | ||
426 | TimeZoneFormat& | |
427 | TimeZoneFormat::operator=(const TimeZoneFormat& other) { | |
428 | if (this == &other) { | |
429 | return *this; | |
430 | } | |
431 | ||
432 | delete fTimeZoneNames; | |
433 | delete fTimeZoneGenericNames; | |
434 | fTimeZoneGenericNames = NULL; | |
435 | ||
436 | fLocale = other.fLocale; | |
437 | uprv_memcpy(fTargetRegion, other.fTargetRegion, sizeof(fTargetRegion)); | |
438 | ||
439 | fTimeZoneNames = other.fTimeZoneNames->clone(); | |
440 | if (other.fTimeZoneGenericNames) { | |
57a6839d | 441 | // TODO: this test has dubious thread safety. |
4388f060 A |
442 | fTimeZoneGenericNames = other.fTimeZoneGenericNames->clone(); |
443 | } | |
444 | ||
445 | fGMTPattern = other.fGMTPattern; | |
446 | fGMTPatternPrefix = other.fGMTPatternPrefix; | |
447 | fGMTPatternSuffix = other.fGMTPatternSuffix; | |
448 | ||
449 | UErrorCode status = U_ZERO_ERROR; | |
51004dcb | 450 | for (int32_t i = 0; i < UTZFMT_PAT_COUNT; i++) { |
4388f060 A |
451 | fGMTOffsetPatterns[i] = other.fGMTOffsetPatterns[i]; |
452 | delete fGMTOffsetPatternItems[i]; | |
453 | } | |
454 | initGMTOffsetPatterns(status); | |
455 | U_ASSERT(U_SUCCESS(status)); | |
456 | ||
457 | fGMTZeroFormat = other.fGMTZeroFormat; | |
458 | ||
459 | uprv_memcpy(fGMTOffsetDigits, other.fGMTOffsetDigits, sizeof(fGMTOffsetDigits)); | |
460 | ||
461 | fDefParseOptionFlags = other.fDefParseOptionFlags; | |
462 | ||
463 | return *this; | |
464 | } | |
465 | ||
466 | ||
467 | UBool | |
468 | TimeZoneFormat::operator==(const Format& other) const { | |
469 | TimeZoneFormat* tzfmt = (TimeZoneFormat*)&other; | |
470 | ||
471 | UBool isEqual = | |
472 | fLocale == tzfmt->fLocale | |
473 | && fGMTPattern == tzfmt->fGMTPattern | |
474 | && fGMTZeroFormat == tzfmt->fGMTZeroFormat | |
475 | && *fTimeZoneNames == *tzfmt->fTimeZoneNames; | |
476 | ||
51004dcb | 477 | for (int32_t i = 0; i < UTZFMT_PAT_COUNT && isEqual; i++) { |
4388f060 A |
478 | isEqual = fGMTOffsetPatterns[i] == tzfmt->fGMTOffsetPatterns[i]; |
479 | } | |
480 | for (int32_t i = 0; i < 10 && isEqual; i++) { | |
481 | isEqual = fGMTOffsetDigits[i] == tzfmt->fGMTOffsetDigits[i]; | |
482 | } | |
483 | // TODO | |
484 | // Check fTimeZoneGenericNames. For now, | |
485 | // if fTimeZoneNames is same, fTimeZoneGenericNames should | |
486 | // be also equivalent. | |
487 | return isEqual; | |
488 | } | |
489 | ||
490 | Format* | |
491 | TimeZoneFormat::clone() const { | |
492 | return new TimeZoneFormat(*this); | |
493 | } | |
494 | ||
495 | TimeZoneFormat* U_EXPORT2 | |
496 | TimeZoneFormat::createInstance(const Locale& locale, UErrorCode& status) { | |
497 | TimeZoneFormat* tzfmt = new TimeZoneFormat(locale, status); | |
498 | if (U_SUCCESS(status)) { | |
499 | return tzfmt; | |
500 | } | |
501 | delete tzfmt; | |
502 | return NULL; | |
503 | } | |
504 | ||
505 | // ------------------------------------------------------------------ | |
506 | // Setter and Getter | |
507 | ||
508 | const TimeZoneNames* | |
509 | TimeZoneFormat::getTimeZoneNames() const { | |
510 | return (const TimeZoneNames*)fTimeZoneNames; | |
511 | } | |
512 | ||
513 | void | |
514 | TimeZoneFormat::adoptTimeZoneNames(TimeZoneNames *tznames) { | |
515 | delete fTimeZoneNames; | |
516 | fTimeZoneNames = tznames; | |
517 | ||
518 | // TODO - We should also update fTimeZoneGenericNames | |
519 | } | |
520 | ||
521 | void | |
522 | TimeZoneFormat::setTimeZoneNames(const TimeZoneNames &tznames) { | |
523 | delete fTimeZoneNames; | |
524 | fTimeZoneNames = tznames.clone(); | |
525 | ||
526 | // TODO - We should also update fTimeZoneGenericNames | |
527 | } | |
528 | ||
529 | void | |
51004dcb | 530 | TimeZoneFormat::setDefaultParseOptions(uint32_t flags) { |
4388f060 A |
531 | fDefParseOptionFlags = flags; |
532 | } | |
533 | ||
51004dcb | 534 | uint32_t |
4388f060 A |
535 | TimeZoneFormat::getDefaultParseOptions(void) const { |
536 | return fDefParseOptionFlags; | |
537 | } | |
538 | ||
539 | ||
540 | UnicodeString& | |
541 | TimeZoneFormat::getGMTPattern(UnicodeString& pattern) const { | |
542 | return pattern.setTo(fGMTPattern); | |
543 | } | |
544 | ||
545 | void | |
546 | TimeZoneFormat::setGMTPattern(const UnicodeString& pattern, UErrorCode& status) { | |
547 | initGMTPattern(pattern, status); | |
548 | } | |
549 | ||
550 | UnicodeString& | |
551 | TimeZoneFormat::getGMTOffsetPattern(UTimeZoneFormatGMTOffsetPatternType type, UnicodeString& pattern) const { | |
552 | return pattern.setTo(fGMTOffsetPatterns[type]); | |
553 | } | |
554 | ||
555 | void | |
556 | TimeZoneFormat::setGMTOffsetPattern(UTimeZoneFormatGMTOffsetPatternType type, const UnicodeString& pattern, UErrorCode& status) { | |
557 | if (U_FAILURE(status)) { | |
558 | return; | |
559 | } | |
560 | if (pattern == fGMTOffsetPatterns[type]) { | |
561 | // No need to reset | |
562 | return; | |
563 | } | |
564 | ||
51004dcb A |
565 | OffsetFields required = FIELDS_HM; |
566 | switch (type) { | |
567 | case UTZFMT_PAT_POSITIVE_H: | |
568 | case UTZFMT_PAT_NEGATIVE_H: | |
569 | required = FIELDS_H; | |
570 | break; | |
571 | case UTZFMT_PAT_POSITIVE_HM: | |
572 | case UTZFMT_PAT_NEGATIVE_HM: | |
573 | required = FIELDS_HM; | |
574 | break; | |
575 | case UTZFMT_PAT_POSITIVE_HMS: | |
576 | case UTZFMT_PAT_NEGATIVE_HMS: | |
577 | required = FIELDS_HMS; | |
578 | break; | |
579 | default: | |
580 | U_ASSERT(FALSE); | |
581 | break; | |
582 | } | |
4388f060 A |
583 | |
584 | UVector* patternItems = parseOffsetPattern(pattern, required, status); | |
585 | if (patternItems == NULL) { | |
586 | return; | |
587 | } | |
588 | ||
589 | fGMTOffsetPatterns[type].setTo(pattern); | |
590 | delete fGMTOffsetPatternItems[type]; | |
591 | fGMTOffsetPatternItems[type] = patternItems; | |
51004dcb | 592 | checkAbuttingHoursAndMinutes(); |
4388f060 A |
593 | } |
594 | ||
595 | UnicodeString& | |
596 | TimeZoneFormat::getGMTOffsetDigits(UnicodeString& digits) const { | |
597 | digits.remove(); | |
598 | for (int32_t i = 0; i < 10; i++) { | |
599 | digits.append(fGMTOffsetDigits[i]); | |
600 | } | |
601 | return digits; | |
602 | } | |
603 | ||
604 | void | |
605 | TimeZoneFormat::setGMTOffsetDigits(const UnicodeString& digits, UErrorCode& status) { | |
606 | if (U_FAILURE(status)) { | |
607 | return; | |
608 | } | |
609 | UChar32 digitArray[10]; | |
610 | if (!toCodePoints(digits, digitArray, 10)) { | |
611 | status = U_ILLEGAL_ARGUMENT_ERROR; | |
612 | return; | |
613 | } | |
614 | uprv_memcpy(fGMTOffsetDigits, digitArray, sizeof(UChar32)*10); | |
615 | } | |
616 | ||
617 | UnicodeString& | |
618 | TimeZoneFormat::getGMTZeroFormat(UnicodeString& gmtZeroFormat) const { | |
619 | return gmtZeroFormat.setTo(fGMTZeroFormat); | |
620 | } | |
621 | ||
622 | void | |
623 | TimeZoneFormat::setGMTZeroFormat(const UnicodeString& gmtZeroFormat, UErrorCode& status) { | |
624 | if (U_SUCCESS(status)) { | |
625 | if (gmtZeroFormat.isEmpty()) { | |
626 | status = U_ILLEGAL_ARGUMENT_ERROR; | |
627 | } else if (gmtZeroFormat != fGMTZeroFormat) { | |
628 | fGMTZeroFormat.setTo(gmtZeroFormat); | |
629 | } | |
630 | } | |
631 | } | |
632 | ||
633 | // ------------------------------------------------------------------ | |
634 | // Format and Parse | |
635 | ||
636 | UnicodeString& | |
637 | TimeZoneFormat::format(UTimeZoneFormatStyle style, const TimeZone& tz, UDate date, | |
638 | UnicodeString& name, UTimeZoneFormatTimeType* timeType /* = NULL */) const { | |
639 | if (timeType) { | |
640 | *timeType = UTZFMT_TIME_TYPE_UNKNOWN; | |
641 | } | |
57a6839d A |
642 | |
643 | UBool noOffsetFormatFallback = FALSE; | |
644 | ||
4388f060 A |
645 | switch (style) { |
646 | case UTZFMT_STYLE_GENERIC_LOCATION: | |
647 | formatGeneric(tz, UTZGNM_LOCATION, date, name); | |
648 | break; | |
649 | case UTZFMT_STYLE_GENERIC_LONG: | |
650 | formatGeneric(tz, UTZGNM_LONG, date, name); | |
651 | break; | |
652 | case UTZFMT_STYLE_GENERIC_SHORT: | |
653 | formatGeneric(tz, UTZGNM_SHORT, date, name); | |
654 | break; | |
655 | case UTZFMT_STYLE_SPECIFIC_LONG: | |
656 | formatSpecific(tz, UTZNM_LONG_STANDARD, UTZNM_LONG_DAYLIGHT, date, name, timeType); | |
657 | break; | |
658 | case UTZFMT_STYLE_SPECIFIC_SHORT: | |
659 | formatSpecific(tz, UTZNM_SHORT_STANDARD, UTZNM_SHORT_DAYLIGHT, date, name, timeType); | |
660 | break; | |
57a6839d A |
661 | |
662 | case UTZFMT_STYLE_ZONE_ID: | |
663 | tz.getID(name); | |
664 | noOffsetFormatFallback = TRUE; | |
665 | break; | |
666 | case UTZFMT_STYLE_ZONE_ID_SHORT: | |
667 | { | |
668 | const UChar* shortID = ZoneMeta::getShortID(tz); | |
669 | if (shortID == NULL) { | |
670 | shortID = UNKNOWN_SHORT_ZONE_ID; | |
671 | } | |
672 | name.setTo(shortID, -1); | |
673 | } | |
674 | noOffsetFormatFallback = TRUE; | |
675 | break; | |
676 | ||
677 | case UTZFMT_STYLE_EXEMPLAR_LOCATION: | |
678 | formatExemplarLocation(tz, name); | |
679 | noOffsetFormatFallback = TRUE; | |
680 | break; | |
681 | ||
51004dcb | 682 | default: |
4388f060 A |
683 | // will be handled below |
684 | break; | |
685 | } | |
686 | ||
57a6839d | 687 | if (name.isEmpty() && !noOffsetFormatFallback) { |
4388f060 A |
688 | UErrorCode status = U_ZERO_ERROR; |
689 | int32_t rawOffset, dstOffset; | |
690 | tz.getOffset(date, FALSE, rawOffset, dstOffset, status); | |
51004dcb | 691 | int32_t offset = rawOffset + dstOffset; |
4388f060 A |
692 | if (U_SUCCESS(status)) { |
693 | switch (style) { | |
51004dcb A |
694 | case UTZFMT_STYLE_GENERIC_LOCATION: |
695 | case UTZFMT_STYLE_GENERIC_LONG: | |
696 | case UTZFMT_STYLE_SPECIFIC_LONG: | |
697 | case UTZFMT_STYLE_LOCALIZED_GMT: | |
698 | formatOffsetLocalizedGMT(offset, name, status); | |
699 | break; | |
700 | ||
701 | case UTZFMT_STYLE_GENERIC_SHORT: | |
702 | case UTZFMT_STYLE_SPECIFIC_SHORT: | |
703 | case UTZFMT_STYLE_LOCALIZED_GMT_SHORT: | |
704 | formatOffsetShortLocalizedGMT(offset, name, status); | |
4388f060 | 705 | break; |
51004dcb A |
706 | |
707 | case UTZFMT_STYLE_ISO_BASIC_SHORT: | |
708 | formatOffsetISO8601Basic(offset, TRUE, TRUE, TRUE, name, status); | |
4388f060 | 709 | break; |
51004dcb A |
710 | |
711 | case UTZFMT_STYLE_ISO_BASIC_LOCAL_SHORT: | |
712 | formatOffsetISO8601Basic(offset, FALSE, TRUE, TRUE, name, status); | |
713 | break; | |
714 | ||
715 | case UTZFMT_STYLE_ISO_BASIC_FIXED: | |
716 | formatOffsetISO8601Basic(offset, TRUE, FALSE, TRUE, name, status); | |
717 | break; | |
718 | ||
719 | case UTZFMT_STYLE_ISO_BASIC_LOCAL_FIXED: | |
720 | formatOffsetISO8601Basic(offset, FALSE, FALSE, TRUE, name, status); | |
721 | break; | |
722 | ||
723 | case UTZFMT_STYLE_ISO_EXTENDED_FIXED: | |
724 | formatOffsetISO8601Extended(offset, TRUE, FALSE, TRUE, name, status); | |
725 | break; | |
726 | ||
727 | case UTZFMT_STYLE_ISO_EXTENDED_LOCAL_FIXED: | |
728 | formatOffsetISO8601Extended(offset, FALSE, FALSE, TRUE, name, status); | |
729 | break; | |
730 | ||
731 | case UTZFMT_STYLE_ISO_BASIC_FULL: | |
732 | formatOffsetISO8601Basic(offset, TRUE, FALSE, FALSE, name, status); | |
733 | break; | |
734 | ||
735 | case UTZFMT_STYLE_ISO_BASIC_LOCAL_FULL: | |
736 | formatOffsetISO8601Basic(offset, FALSE, FALSE, FALSE, name, status); | |
737 | break; | |
738 | ||
739 | case UTZFMT_STYLE_ISO_EXTENDED_FULL: | |
740 | formatOffsetISO8601Extended(offset, TRUE, FALSE, FALSE, name, status); | |
741 | break; | |
742 | ||
743 | case UTZFMT_STYLE_ISO_EXTENDED_LOCAL_FULL: | |
744 | formatOffsetISO8601Extended(offset, FALSE, FALSE, FALSE, name, status); | |
745 | break; | |
746 | ||
57a6839d A |
747 | default: |
748 | // UTZFMT_STYLE_ZONE_ID, UTZFMT_STYLE_ZONE_ID_SHORT, UTZFMT_STYLE_EXEMPLAR_LOCATION | |
749 | break; | |
4388f060 | 750 | } |
57a6839d | 751 | |
4388f060 A |
752 | if (timeType) { |
753 | *timeType = (dstOffset != 0) ? UTZFMT_TIME_TYPE_DAYLIGHT : UTZFMT_TIME_TYPE_STANDARD; | |
754 | } | |
755 | } | |
4388f060 A |
756 | } |
757 | ||
758 | return name; | |
759 | } | |
760 | ||
761 | UnicodeString& | |
762 | TimeZoneFormat::format(const Formattable& obj, UnicodeString& appendTo, | |
763 | FieldPosition& pos, UErrorCode& status) const { | |
764 | if (U_FAILURE(status)) { | |
765 | return appendTo; | |
766 | } | |
767 | UDate date = Calendar::getNow(); | |
768 | if (obj.getType() == Formattable::kObject) { | |
769 | const UObject* formatObj = obj.getObject(); | |
770 | const TimeZone* tz = dynamic_cast<const TimeZone*>(formatObj); | |
771 | if (tz == NULL) { | |
772 | const Calendar* cal = dynamic_cast<const Calendar*>(formatObj); | |
773 | if (cal != NULL) { | |
774 | tz = &cal->getTimeZone(); | |
775 | date = cal->getTime(status); | |
776 | } | |
777 | } | |
778 | if (tz != NULL) { | |
779 | int32_t rawOffset, dstOffset; | |
780 | tz->getOffset(date, FALSE, rawOffset, dstOffset, status); | |
781 | UnicodeString result; | |
782 | formatOffsetLocalizedGMT(rawOffset + dstOffset, result, status); | |
783 | if (U_SUCCESS(status)) { | |
784 | appendTo.append(result); | |
785 | if (pos.getField() == UDAT_TIMEZONE_FIELD) { | |
786 | pos.setBeginIndex(0); | |
787 | pos.setEndIndex(result.length()); | |
788 | } | |
789 | } | |
790 | } | |
791 | } | |
792 | return appendTo; | |
793 | } | |
794 | ||
795 | TimeZone* | |
796 | TimeZoneFormat::parse(UTimeZoneFormatStyle style, const UnicodeString& text, ParsePosition& pos, | |
797 | UTimeZoneFormatTimeType* timeType /*= NULL*/) const { | |
798 | return parse(style, text, pos, getDefaultParseOptions(), timeType); | |
799 | } | |
800 | ||
801 | TimeZone* | |
802 | TimeZoneFormat::parse(UTimeZoneFormatStyle style, const UnicodeString& text, ParsePosition& pos, | |
803 | int32_t parseOptions, UTimeZoneFormatTimeType* timeType /* = NULL */) const { | |
804 | if (timeType) { | |
805 | *timeType = UTZFMT_TIME_TYPE_UNKNOWN; | |
806 | } | |
807 | ||
808 | int32_t startIdx = pos.getIndex(); | |
809 | int32_t maxPos = text.length(); | |
810 | int32_t offset; | |
811 | ||
51004dcb A |
812 | // Styles using localized GMT format as fallback |
813 | UBool fallbackLocalizedGMT = | |
814 | (style == UTZFMT_STYLE_SPECIFIC_LONG || style == UTZFMT_STYLE_GENERIC_LONG || style == UTZFMT_STYLE_GENERIC_LOCATION); | |
815 | UBool fallbackShortLocalizedGMT = | |
816 | (style == UTZFMT_STYLE_SPECIFIC_SHORT || style == UTZFMT_STYLE_GENERIC_SHORT); | |
4388f060 | 817 | |
51004dcb | 818 | int32_t evaluated = 0; // bit flags representing already evaluated styles |
4388f060 A |
819 | ParsePosition tmpPos(startIdx); |
820 | ||
821 | int32_t parsedOffset = UNKNOWN_OFFSET; // stores successfully parsed offset for later use | |
822 | int32_t parsedPos = -1; // stores successfully parsed offset position for later use | |
823 | ||
824 | // Try localized GMT format first if necessary | |
51004dcb | 825 | if (fallbackLocalizedGMT || fallbackShortLocalizedGMT) { |
4388f060 | 826 | UBool hasDigitOffset = FALSE; |
51004dcb | 827 | offset = parseOffsetLocalizedGMT(text, tmpPos, fallbackShortLocalizedGMT, &hasDigitOffset); |
4388f060 A |
828 | if (tmpPos.getErrorIndex() == -1) { |
829 | // Even when the input text was successfully parsed as a localized GMT format text, | |
830 | // we may still need to evaluate the specified style if - | |
831 | // 1) GMT zero format was used, and | |
832 | // 2) The input text was not completely processed | |
833 | if (tmpPos.getIndex() == maxPos || hasDigitOffset) { | |
834 | pos.setIndex(tmpPos.getIndex()); | |
835 | return createTimeZoneForOffset(offset); | |
836 | } | |
837 | parsedOffset = offset; | |
838 | parsedPos = tmpPos.getIndex(); | |
839 | } | |
51004dcb A |
840 | // Note: For now, no distinction between long/short localized GMT format in the parser. |
841 | // This might be changed in future. | |
842 | // evaluated |= (fallbackLocalizedGMT ? STYLE_PARSE_FLAGS[UTZFMT_STYLE_LOCALIZED_GMT] : STYLE_PARSE_FLAGS[UTZFMT_STYLE_LOCALIZED_GMT_SHORT]); | |
843 | evaluated |= STYLE_PARSE_FLAGS[UTZFMT_STYLE_LOCALIZED_GMT] | STYLE_PARSE_FLAGS[UTZFMT_STYLE_LOCALIZED_GMT_SHORT]; | |
4388f060 A |
844 | } |
845 | ||
846 | UErrorCode status = U_ZERO_ERROR; | |
847 | UnicodeString tzID; | |
4388f060 A |
848 | |
849 | // Try the specified style | |
850 | switch (style) { | |
51004dcb | 851 | case UTZFMT_STYLE_LOCALIZED_GMT: |
4388f060 | 852 | { |
51004dcb A |
853 | tmpPos.setIndex(startIdx); |
854 | tmpPos.setErrorIndex(-1); | |
855 | ||
856 | offset = parseOffsetLocalizedGMT(text, tmpPos); | |
4388f060 A |
857 | if (tmpPos.getErrorIndex() == -1) { |
858 | pos.setIndex(tmpPos.getIndex()); | |
859 | return createTimeZoneForOffset(offset); | |
860 | } | |
4388f060 | 861 | |
51004dcb A |
862 | // Note: For now, no distinction between long/short localized GMT format in the parser. |
863 | // This might be changed in future. | |
864 | evaluated |= STYLE_PARSE_FLAGS[UTZFMT_STYLE_LOCALIZED_GMT_SHORT]; | |
865 | ||
866 | break; | |
867 | } | |
868 | case UTZFMT_STYLE_LOCALIZED_GMT_SHORT: | |
4388f060 | 869 | { |
51004dcb A |
870 | tmpPos.setIndex(startIdx); |
871 | tmpPos.setErrorIndex(-1); | |
872 | ||
873 | offset = parseOffsetShortLocalizedGMT(text, tmpPos); | |
4388f060 A |
874 | if (tmpPos.getErrorIndex() == -1) { |
875 | pos.setIndex(tmpPos.getIndex()); | |
876 | return createTimeZoneForOffset(offset); | |
877 | } | |
4388f060 | 878 | |
51004dcb A |
879 | // Note: For now, no distinction between long/short localized GMT format in the parser. |
880 | // This might be changed in future. | |
881 | evaluated |= STYLE_PARSE_FLAGS[UTZFMT_STYLE_LOCALIZED_GMT]; | |
882 | ||
883 | break; | |
884 | } | |
885 | case UTZFMT_STYLE_ISO_BASIC_SHORT: | |
886 | case UTZFMT_STYLE_ISO_BASIC_FIXED: | |
887 | case UTZFMT_STYLE_ISO_BASIC_FULL: | |
888 | case UTZFMT_STYLE_ISO_EXTENDED_FIXED: | |
889 | case UTZFMT_STYLE_ISO_EXTENDED_FULL: | |
4388f060 | 890 | { |
51004dcb A |
891 | tmpPos.setIndex(startIdx); |
892 | tmpPos.setErrorIndex(-1); | |
893 | ||
4388f060 A |
894 | offset = parseOffsetISO8601(text, tmpPos); |
895 | if (tmpPos.getErrorIndex() == -1) { | |
896 | pos.setIndex(tmpPos.getIndex()); | |
897 | return createTimeZoneForOffset(offset); | |
898 | } | |
51004dcb A |
899 | |
900 | break; | |
901 | } | |
902 | ||
903 | case UTZFMT_STYLE_ISO_BASIC_LOCAL_SHORT: | |
904 | case UTZFMT_STYLE_ISO_BASIC_LOCAL_FIXED: | |
905 | case UTZFMT_STYLE_ISO_BASIC_LOCAL_FULL: | |
906 | case UTZFMT_STYLE_ISO_EXTENDED_LOCAL_FIXED: | |
907 | case UTZFMT_STYLE_ISO_EXTENDED_LOCAL_FULL: | |
908 | { | |
909 | tmpPos.setIndex(startIdx); | |
910 | tmpPos.setErrorIndex(-1); | |
911 | ||
912 | // Exclude the case of UTC Indicator "Z" here | |
913 | UBool hasDigitOffset = FALSE; | |
914 | offset = parseOffsetISO8601(text, tmpPos, FALSE, &hasDigitOffset); | |
915 | if (tmpPos.getErrorIndex() == -1 && hasDigitOffset) { | |
916 | pos.setIndex(tmpPos.getIndex()); | |
917 | return createTimeZoneForOffset(offset); | |
918 | } | |
919 | ||
920 | break; | |
4388f060 | 921 | } |
4388f060 A |
922 | |
923 | case UTZFMT_STYLE_SPECIFIC_LONG: | |
924 | case UTZFMT_STYLE_SPECIFIC_SHORT: | |
925 | { | |
926 | // Specific styles | |
927 | int32_t nameTypes = 0; | |
928 | if (style == UTZFMT_STYLE_SPECIFIC_LONG) { | |
929 | nameTypes = (UTZNM_LONG_STANDARD | UTZNM_LONG_DAYLIGHT); | |
930 | } else { | |
931 | U_ASSERT(style == UTZFMT_STYLE_SPECIFIC_SHORT); | |
932 | nameTypes = (UTZNM_SHORT_STANDARD | UTZNM_SHORT_DAYLIGHT); | |
933 | } | |
934 | LocalPointer<TimeZoneNames::MatchInfoCollection> specificMatches(fTimeZoneNames->find(text, startIdx, nameTypes, status)); | |
935 | if (U_FAILURE(status)) { | |
936 | pos.setErrorIndex(startIdx); | |
937 | return NULL; | |
938 | } | |
939 | if (!specificMatches.isNull()) { | |
940 | int32_t matchIdx = -1; | |
941 | int32_t matchPos = -1; | |
942 | for (int32_t i = 0; i < specificMatches->size(); i++) { | |
943 | matchPos = startIdx + specificMatches->getMatchLengthAt(i); | |
944 | if (matchPos > parsedPos) { | |
945 | matchIdx = i; | |
946 | parsedPos = matchPos; | |
947 | } | |
948 | } | |
949 | if (matchIdx >= 0) { | |
950 | if (timeType) { | |
951 | *timeType = getTimeType(specificMatches->getNameTypeAt(matchIdx)); | |
952 | } | |
953 | pos.setIndex(matchPos); | |
954 | getTimeZoneID(specificMatches.getAlias(), matchIdx, tzID); | |
955 | U_ASSERT(!tzID.isEmpty()); | |
956 | return TimeZone::createTimeZone(tzID); | |
957 | } | |
958 | } | |
51004dcb | 959 | break; |
4388f060 | 960 | } |
4388f060 A |
961 | case UTZFMT_STYLE_GENERIC_LONG: |
962 | case UTZFMT_STYLE_GENERIC_SHORT: | |
963 | case UTZFMT_STYLE_GENERIC_LOCATION: | |
964 | { | |
965 | int32_t genericNameTypes = 0; | |
966 | switch (style) { | |
967 | case UTZFMT_STYLE_GENERIC_LOCATION: | |
968 | genericNameTypes = UTZGNM_LOCATION; | |
969 | break; | |
970 | ||
971 | case UTZFMT_STYLE_GENERIC_LONG: | |
972 | genericNameTypes = UTZGNM_LONG | UTZGNM_LOCATION; | |
973 | break; | |
974 | ||
975 | case UTZFMT_STYLE_GENERIC_SHORT: | |
976 | genericNameTypes = UTZGNM_SHORT | UTZGNM_LOCATION; | |
977 | break; | |
978 | ||
979 | default: | |
980 | U_ASSERT(FALSE); | |
981 | } | |
982 | ||
983 | int32_t len = 0; | |
51004dcb | 984 | UTimeZoneFormatTimeType tt = UTZFMT_TIME_TYPE_UNKNOWN; |
4388f060 A |
985 | const TimeZoneGenericNames *gnames = getTimeZoneGenericNames(status); |
986 | if (U_SUCCESS(status)) { | |
51004dcb | 987 | len = gnames->findBestMatch(text, startIdx, genericNameTypes, tzID, tt, status); |
4388f060 A |
988 | } |
989 | if (U_FAILURE(status)) { | |
990 | pos.setErrorIndex(startIdx); | |
991 | return NULL; | |
992 | } | |
993 | if (len > 0) { | |
994 | // Found a match | |
995 | if (timeType) { | |
51004dcb | 996 | *timeType = tt; |
4388f060 A |
997 | } |
998 | pos.setIndex(startIdx + len); | |
999 | U_ASSERT(!tzID.isEmpty()); | |
1000 | return TimeZone::createTimeZone(tzID); | |
1001 | } | |
51004dcb A |
1002 | |
1003 | break; | |
1004 | } | |
1005 | case UTZFMT_STYLE_ZONE_ID: | |
1006 | { | |
1007 | tmpPos.setIndex(startIdx); | |
1008 | tmpPos.setErrorIndex(-1); | |
1009 | ||
1010 | parseZoneID(text, tmpPos, tzID); | |
1011 | if (tmpPos.getErrorIndex() == -1) { | |
1012 | pos.setIndex(tmpPos.getIndex()); | |
1013 | return TimeZone::createTimeZone(tzID); | |
1014 | } | |
1015 | break; | |
1016 | } | |
1017 | case UTZFMT_STYLE_ZONE_ID_SHORT: | |
1018 | { | |
1019 | tmpPos.setIndex(startIdx); | |
1020 | tmpPos.setErrorIndex(-1); | |
1021 | ||
1022 | parseShortZoneID(text, tmpPos, tzID); | |
1023 | if (tmpPos.getErrorIndex() == -1) { | |
1024 | pos.setIndex(tmpPos.getIndex()); | |
1025 | return TimeZone::createTimeZone(tzID); | |
1026 | } | |
1027 | break; | |
1028 | } | |
1029 | case UTZFMT_STYLE_EXEMPLAR_LOCATION: | |
1030 | { | |
1031 | tmpPos.setIndex(startIdx); | |
1032 | tmpPos.setErrorIndex(-1); | |
1033 | ||
1034 | parseExemplarLocation(text, tmpPos, tzID); | |
1035 | if (tmpPos.getErrorIndex() == -1) { | |
1036 | pos.setIndex(tmpPos.getIndex()); | |
1037 | return TimeZone::createTimeZone(tzID); | |
1038 | } | |
1039 | break; | |
4388f060 | 1040 | } |
4388f060 | 1041 | } |
51004dcb | 1042 | evaluated |= STYLE_PARSE_FLAGS[style]; |
4388f060 A |
1043 | |
1044 | ||
1045 | if (parsedPos > startIdx) { | |
1046 | // When the specified style is one of SPECIFIC_XXX or GENERIC_XXX, we tried to parse the input | |
1047 | // as localized GMT format earlier. If parsedOffset is positive, it means it was successfully | |
1048 | // parsed as localized GMT format, but offset digits were not detected (more specifically, GMT | |
1049 | // zero format). Then, it tried to find a match within the set of display names, but could not | |
1050 | // find a match. At this point, we can safely assume the input text contains the localized | |
1051 | // GMT format. | |
1052 | U_ASSERT(parsedOffset != UNKNOWN_OFFSET); | |
1053 | pos.setIndex(parsedPos); | |
1054 | return createTimeZoneForOffset(parsedOffset); | |
1055 | } | |
1056 | ||
1057 | // Failed to parse the input text as the time zone format in the specified style. | |
1058 | // Check the longest match among other styles below. | |
51004dcb A |
1059 | UnicodeString parsedID; |
1060 | UTimeZoneFormatTimeType parsedTimeType = UTZFMT_TIME_TYPE_UNKNOWN; | |
1061 | ||
4388f060 A |
1062 | U_ASSERT(parsedPos < 0); |
1063 | U_ASSERT(parsedOffset == UNKNOWN_OFFSET); | |
4388f060 A |
1064 | |
1065 | // ISO 8601 | |
51004dcb A |
1066 | if (parsedPos < maxPos && |
1067 | ((evaluated & ISO_Z_STYLE_FLAG) == 0 || (evaluated & ISO_LOCAL_STYLE_FLAG) == 0)) { | |
1068 | tmpPos.setIndex(startIdx); | |
1069 | tmpPos.setErrorIndex(-1); | |
1070 | ||
4388f060 A |
1071 | UBool hasDigitOffset = FALSE; |
1072 | offset = parseOffsetISO8601(text, tmpPos, FALSE, &hasDigitOffset); | |
1073 | if (tmpPos.getErrorIndex() == -1) { | |
1074 | if (tmpPos.getIndex() == maxPos || hasDigitOffset) { | |
1075 | pos.setIndex(tmpPos.getIndex()); | |
1076 | return createTimeZoneForOffset(offset); | |
1077 | } | |
1078 | // Note: When ISO 8601 format contains offset digits, it should not | |
51004dcb A |
1079 | // collide with other formats. However, ISO 8601 UTC format "Z" (single letter) |
1080 | // may collide with other names. In this case, we need to evaluate other names. | |
1081 | if (parsedPos < tmpPos.getIndex()) { | |
1082 | parsedOffset = offset; | |
1083 | parsedID.setToBogus(); | |
1084 | parsedTimeType = UTZFMT_TIME_TYPE_UNKNOWN; | |
1085 | parsedPos = tmpPos.getIndex(); | |
1086 | U_ASSERT(parsedPos == startIdx + 1); // only when "Z" is used | |
1087 | } | |
4388f060 | 1088 | } |
51004dcb A |
1089 | } |
1090 | ||
1091 | // Localized GMT format | |
1092 | if (parsedPos < maxPos && | |
1093 | (evaluated & STYLE_PARSE_FLAGS[UTZFMT_STYLE_LOCALIZED_GMT]) == 0) { | |
4388f060 A |
1094 | tmpPos.setIndex(startIdx); |
1095 | tmpPos.setErrorIndex(-1); | |
4388f060 | 1096 | |
51004dcb A |
1097 | UBool hasDigitOffset = FALSE; |
1098 | offset = parseOffsetLocalizedGMT(text, tmpPos, FALSE, &hasDigitOffset); | |
1099 | if (tmpPos.getErrorIndex() == -1) { | |
1100 | if (tmpPos.getIndex() == maxPos || hasDigitOffset) { | |
1101 | pos.setIndex(tmpPos.getIndex()); | |
1102 | return createTimeZoneForOffset(offset); | |
1103 | } | |
1104 | // Evaluate other names - see the comment earlier in this method. | |
1105 | if (parsedPos < tmpPos.getIndex()) { | |
1106 | parsedOffset = offset; | |
1107 | parsedID.setToBogus(); | |
1108 | parsedTimeType = UTZFMT_TIME_TYPE_UNKNOWN; | |
1109 | parsedPos = tmpPos.getIndex(); | |
1110 | } | |
1111 | } | |
1112 | } | |
4388f060 | 1113 | |
51004dcb A |
1114 | if (parsedPos < maxPos && |
1115 | (evaluated & STYLE_PARSE_FLAGS[UTZFMT_STYLE_LOCALIZED_GMT_SHORT]) == 0) { | |
1116 | tmpPos.setIndex(startIdx); | |
1117 | tmpPos.setErrorIndex(-1); | |
4388f060 | 1118 | |
4388f060 | 1119 | UBool hasDigitOffset = FALSE; |
51004dcb | 1120 | offset = parseOffsetLocalizedGMT(text, tmpPos, TRUE, &hasDigitOffset); |
4388f060 A |
1121 | if (tmpPos.getErrorIndex() == -1) { |
1122 | if (tmpPos.getIndex() == maxPos || hasDigitOffset) { | |
1123 | pos.setIndex(tmpPos.getIndex()); | |
1124 | return createTimeZoneForOffset(offset); | |
1125 | } | |
1126 | // Evaluate other names - see the comment earlier in this method. | |
51004dcb A |
1127 | if (parsedPos < tmpPos.getIndex()) { |
1128 | parsedOffset = offset; | |
1129 | parsedID.setToBogus(); | |
1130 | parsedTimeType = UTZFMT_TIME_TYPE_UNKNOWN; | |
1131 | parsedPos = tmpPos.getIndex(); | |
1132 | } | |
4388f060 A |
1133 | } |
1134 | } | |
1135 | ||
51004dcb | 1136 | // When ParseOption.ALL_STYLES is available, we also try to look all possible display names and IDs. |
4388f060 A |
1137 | // For example, when style is GENERIC_LONG, "EST" (SPECIFIC_SHORT) is never |
1138 | // used for America/New_York. With parseAllStyles true, this code parses "EST" | |
1139 | // as America/New_York. | |
1140 | ||
1141 | // Note: Adding all possible names into the trie used by the implementation is quite heavy operation, | |
1142 | // which we want to avoid normally (note that we cache the trie, so this is applicable to the | |
1143 | // first time only as long as the cache does not expire). | |
51004dcb | 1144 | |
4388f060 | 1145 | if (parseOptions & UTZFMT_PARSE_OPTION_ALL_STYLES) { |
51004dcb A |
1146 | // Try all specific names and exemplar location names |
1147 | if (parsedPos < maxPos) { | |
1148 | LocalPointer<TimeZoneNames::MatchInfoCollection> specificMatches(fTimeZoneNames->find(text, startIdx, ALL_SIMPLE_NAME_TYPES, status)); | |
1149 | if (U_FAILURE(status)) { | |
1150 | pos.setErrorIndex(startIdx); | |
1151 | return NULL; | |
1152 | } | |
1153 | int32_t specificMatchIdx = -1; | |
4388f060 | 1154 | int32_t matchPos = -1; |
51004dcb A |
1155 | if (!specificMatches.isNull()) { |
1156 | for (int32_t i = 0; i < specificMatches->size(); i++) { | |
1157 | if (startIdx + specificMatches->getMatchLengthAt(i) > matchPos) { | |
1158 | specificMatchIdx = i; | |
1159 | matchPos = startIdx + specificMatches->getMatchLengthAt(i); | |
1160 | } | |
4388f060 A |
1161 | } |
1162 | } | |
51004dcb A |
1163 | if (parsedPos < matchPos) { |
1164 | U_ASSERT(specificMatchIdx >= 0); | |
1165 | parsedPos = matchPos; | |
1166 | getTimeZoneID(specificMatches.getAlias(), specificMatchIdx, parsedID); | |
1167 | parsedTimeType = getTimeType(specificMatches->getNameTypeAt(specificMatchIdx)); | |
1168 | parsedOffset = UNKNOWN_OFFSET; | |
1169 | } | |
4388f060 | 1170 | } |
51004dcb | 1171 | // Try generic names |
4388f060 | 1172 | if (parsedPos < maxPos) { |
51004dcb A |
1173 | int32_t genMatchLen = -1; |
1174 | UTimeZoneFormatTimeType tt = UTZFMT_TIME_TYPE_UNKNOWN; | |
1175 | ||
4388f060 A |
1176 | const TimeZoneGenericNames *gnames = getTimeZoneGenericNames(status); |
1177 | if (U_SUCCESS(status)) { | |
51004dcb | 1178 | genMatchLen = gnames->findBestMatch(text, startIdx, ALL_GENERIC_NAME_TYPES, tzID, tt, status); |
4388f060 A |
1179 | } |
1180 | if (U_FAILURE(status)) { | |
1181 | pos.setErrorIndex(startIdx); | |
1182 | return NULL; | |
1183 | } | |
51004dcb A |
1184 | |
1185 | if (parsedPos < startIdx + genMatchLen) { | |
1186 | parsedPos = startIdx + genMatchLen; | |
1187 | parsedID.setTo(tzID); | |
1188 | parsedTimeType = tt; | |
1189 | parsedOffset = UNKNOWN_OFFSET; | |
1190 | } | |
4388f060 | 1191 | } |
51004dcb A |
1192 | |
1193 | // Try time zone ID | |
1194 | if (parsedPos < maxPos && (evaluated & STYLE_PARSE_FLAGS[UTZFMT_STYLE_ZONE_ID]) == 0) { | |
1195 | tmpPos.setIndex(startIdx); | |
1196 | tmpPos.setErrorIndex(-1); | |
1197 | ||
1198 | parseZoneID(text, tmpPos, tzID); | |
1199 | if (tmpPos.getErrorIndex() == -1 && parsedPos < tmpPos.getIndex()) { | |
1200 | parsedPos = tmpPos.getIndex(); | |
1201 | parsedID.setTo(tzID); | |
1202 | parsedTimeType = UTZFMT_TIME_TYPE_UNKNOWN; | |
1203 | parsedOffset = UNKNOWN_OFFSET; | |
4388f060 | 1204 | } |
51004dcb A |
1205 | } |
1206 | // Try short time zone ID | |
1207 | if (parsedPos < maxPos && (evaluated & STYLE_PARSE_FLAGS[UTZFMT_STYLE_ZONE_ID]) == 0) { | |
1208 | tmpPos.setIndex(startIdx); | |
1209 | tmpPos.setErrorIndex(-1); | |
1210 | ||
1211 | parseShortZoneID(text, tmpPos, tzID); | |
1212 | if (tmpPos.getErrorIndex() == -1 && parsedPos < tmpPos.getIndex()) { | |
1213 | parsedPos = tmpPos.getIndex(); | |
1214 | parsedID.setTo(tzID); | |
1215 | parsedTimeType = UTZFMT_TIME_TYPE_UNKNOWN; | |
1216 | parsedOffset = UNKNOWN_OFFSET; | |
4388f060 | 1217 | } |
4388f060 A |
1218 | } |
1219 | } | |
1220 | ||
1221 | if (parsedPos > startIdx) { | |
51004dcb A |
1222 | // Parsed successfully |
1223 | TimeZone* parsedTZ; | |
1224 | if (parsedID.length() > 0) { | |
1225 | parsedTZ = TimeZone::createTimeZone(parsedID); | |
1226 | } else { | |
1227 | U_ASSERT(parsedOffset != UNKNOWN_OFFSET); | |
1228 | parsedTZ = createTimeZoneForOffset(parsedOffset); | |
1229 | } | |
1230 | if (timeType) { | |
1231 | *timeType = parsedTimeType; | |
1232 | } | |
4388f060 | 1233 | pos.setIndex(parsedPos); |
51004dcb | 1234 | return parsedTZ; |
4388f060 A |
1235 | } |
1236 | ||
1237 | pos.setErrorIndex(startIdx); | |
1238 | return NULL; | |
1239 | } | |
1240 | ||
1241 | void | |
1242 | TimeZoneFormat::parseObject(const UnicodeString& source, Formattable& result, | |
1243 | ParsePosition& parse_pos) const { | |
1244 | result.adoptObject(parse(UTZFMT_STYLE_GENERIC_LOCATION, source, parse_pos, UTZFMT_PARSE_OPTION_ALL_STYLES)); | |
1245 | } | |
1246 | ||
1247 | ||
1248 | // ------------------------------------------------------------------ | |
1249 | // Private zone name format/parse implementation | |
1250 | ||
1251 | UnicodeString& | |
1252 | TimeZoneFormat::formatGeneric(const TimeZone& tz, int32_t genType, UDate date, UnicodeString& name) const { | |
1253 | UErrorCode status = U_ZERO_ERROR; | |
1254 | const TimeZoneGenericNames* gnames = getTimeZoneGenericNames(status); | |
1255 | if (U_FAILURE(status)) { | |
1256 | name.setToBogus(); | |
1257 | return name; | |
1258 | } | |
1259 | ||
1260 | if (genType == UTZGNM_LOCATION) { | |
1261 | const UChar* canonicalID = ZoneMeta::getCanonicalCLDRID(tz); | |
1262 | if (canonicalID == NULL) { | |
1263 | name.setToBogus(); | |
1264 | return name; | |
1265 | } | |
1266 | return gnames->getGenericLocationName(UnicodeString(canonicalID), name); | |
1267 | } | |
1268 | return gnames->getDisplayName(tz, (UTimeZoneGenericNameType)genType, date, name); | |
1269 | } | |
1270 | ||
1271 | UnicodeString& | |
1272 | TimeZoneFormat::formatSpecific(const TimeZone& tz, UTimeZoneNameType stdType, UTimeZoneNameType dstType, | |
1273 | UDate date, UnicodeString& name, UTimeZoneFormatTimeType *timeType) const { | |
1274 | if (fTimeZoneNames == NULL) { | |
1275 | name.setToBogus(); | |
1276 | return name; | |
1277 | } | |
1278 | ||
1279 | UErrorCode status = U_ZERO_ERROR; | |
1280 | UBool isDaylight = tz.inDaylightTime(date, status); | |
1281 | const UChar* canonicalID = ZoneMeta::getCanonicalCLDRID(tz); | |
1282 | ||
1283 | if (U_FAILURE(status) || canonicalID == NULL) { | |
1284 | name.setToBogus(); | |
1285 | return name; | |
1286 | } | |
1287 | ||
1288 | if (isDaylight) { | |
1289 | fTimeZoneNames->getDisplayName(UnicodeString(canonicalID), dstType, date, name); | |
1290 | } else { | |
1291 | fTimeZoneNames->getDisplayName(UnicodeString(canonicalID), stdType, date, name); | |
1292 | } | |
1293 | ||
1294 | if (timeType && !name.isEmpty()) { | |
1295 | *timeType = isDaylight ? UTZFMT_TIME_TYPE_DAYLIGHT : UTZFMT_TIME_TYPE_STANDARD; | |
1296 | } | |
1297 | return name; | |
1298 | } | |
1299 | ||
1300 | const TimeZoneGenericNames* | |
1301 | TimeZoneFormat::getTimeZoneGenericNames(UErrorCode& status) const { | |
1302 | if (U_FAILURE(status)) { | |
1303 | return NULL; | |
1304 | } | |
1305 | ||
57a6839d A |
1306 | umtx_lock(&gLock); |
1307 | if (fTimeZoneGenericNames == NULL) { | |
4388f060 | 1308 | TimeZoneFormat *nonConstThis = const_cast<TimeZoneFormat *>(this); |
57a6839d | 1309 | nonConstThis->fTimeZoneGenericNames = TimeZoneGenericNames::createInstance(fLocale, status); |
4388f060 | 1310 | } |
57a6839d | 1311 | umtx_unlock(&gLock); |
4388f060 A |
1312 | |
1313 | return fTimeZoneGenericNames; | |
1314 | } | |
1315 | ||
51004dcb A |
1316 | UnicodeString& |
1317 | TimeZoneFormat::formatExemplarLocation(const TimeZone& tz, UnicodeString& name) const { | |
1318 | UnicodeString location; | |
1319 | const UChar* canonicalID = ZoneMeta::getCanonicalCLDRID(tz); | |
1320 | ||
1321 | if (canonicalID) { | |
1322 | fTimeZoneNames->getExemplarLocationName(UnicodeString(canonicalID), location); | |
1323 | } | |
1324 | if (location.length() > 0) { | |
1325 | name.setTo(location); | |
1326 | } else { | |
1327 | // Use "unknown" location | |
1328 | fTimeZoneNames->getExemplarLocationName(UnicodeString(UNKNOWN_ZONE_ID), location); | |
1329 | if (location.length() > 0) { | |
1330 | name.setTo(location); | |
1331 | } else { | |
1332 | // last resort | |
1333 | name.setTo(UNKNOWN_LOCATION, -1); | |
1334 | } | |
1335 | } | |
1336 | return name; | |
1337 | } | |
1338 | ||
1339 | ||
4388f060 A |
1340 | // ------------------------------------------------------------------ |
1341 | // Zone offset format and parse | |
1342 | ||
1343 | UnicodeString& | |
51004dcb A |
1344 | TimeZoneFormat::formatOffsetISO8601Basic(int32_t offset, UBool useUtcIndicator, UBool isShort, UBool ignoreSeconds, |
1345 | UnicodeString& result, UErrorCode& status) const { | |
1346 | return formatOffsetISO8601(offset, TRUE, useUtcIndicator, isShort, ignoreSeconds, result, status); | |
1347 | } | |
1348 | ||
1349 | UnicodeString& | |
1350 | TimeZoneFormat::formatOffsetISO8601Extended(int32_t offset, UBool useUtcIndicator, UBool isShort, UBool ignoreSeconds, | |
1351 | UnicodeString& result, UErrorCode& status) const { | |
1352 | return formatOffsetISO8601(offset, FALSE, useUtcIndicator, isShort, ignoreSeconds, result, status); | |
1353 | } | |
1354 | ||
1355 | UnicodeString& | |
1356 | TimeZoneFormat::formatOffsetLocalizedGMT(int32_t offset, UnicodeString& result, UErrorCode& status) const { | |
1357 | return formatOffsetLocalizedGMT(offset, FALSE, result, status); | |
1358 | } | |
1359 | ||
1360 | UnicodeString& | |
1361 | TimeZoneFormat::formatOffsetShortLocalizedGMT(int32_t offset, UnicodeString& result, UErrorCode& status) const { | |
1362 | return formatOffsetLocalizedGMT(offset, TRUE, result, status); | |
1363 | } | |
1364 | ||
1365 | int32_t | |
1366 | TimeZoneFormat::parseOffsetISO8601(const UnicodeString& text, ParsePosition& pos) const { | |
1367 | return parseOffsetISO8601(text, pos, FALSE); | |
1368 | } | |
1369 | ||
1370 | int32_t | |
1371 | TimeZoneFormat::parseOffsetLocalizedGMT(const UnicodeString& text, ParsePosition& pos) const { | |
1372 | return parseOffsetLocalizedGMT(text, pos, FALSE, NULL); | |
1373 | } | |
1374 | ||
1375 | int32_t | |
1376 | TimeZoneFormat::parseOffsetShortLocalizedGMT(const UnicodeString& text, ParsePosition& pos) const { | |
1377 | return parseOffsetLocalizedGMT(text, pos, TRUE, NULL); | |
1378 | } | |
1379 | ||
1380 | // ------------------------------------------------------------------ | |
1381 | // Private zone offset format/parse implementation | |
1382 | ||
1383 | UnicodeString& | |
1384 | TimeZoneFormat::formatOffsetISO8601(int32_t offset, UBool isBasic, UBool useUtcIndicator, | |
1385 | UBool isShort, UBool ignoreSeconds, UnicodeString& result, UErrorCode& status) const { | |
4388f060 A |
1386 | if (U_FAILURE(status)) { |
1387 | result.setToBogus(); | |
1388 | return result; | |
1389 | } | |
51004dcb A |
1390 | int32_t absOffset = offset < 0 ? -offset : offset; |
1391 | if (useUtcIndicator && (absOffset < MILLIS_PER_SECOND || (ignoreSeconds && absOffset < MILLIS_PER_MINUTE))) { | |
1392 | result.setTo(ISO8601_UTC); | |
1393 | return result; | |
1394 | } | |
1395 | ||
1396 | OffsetFields minFields = isShort ? FIELDS_H : FIELDS_HM; | |
1397 | OffsetFields maxFields = ignoreSeconds ? FIELDS_HM : FIELDS_HMS; | |
1398 | UChar sep = isBasic ? 0 : ISO8601_SEP; | |
1399 | ||
1400 | // Note: FIELDS_HMS as maxFields is a CLDR/ICU extension. ISO 8601 specification does | |
1401 | // not support seconds field. | |
1402 | ||
1403 | if (absOffset >= MAX_OFFSET) { | |
4388f060 A |
1404 | result.setToBogus(); |
1405 | status = U_ILLEGAL_ARGUMENT_ERROR; | |
1406 | return result; | |
1407 | } | |
1408 | ||
51004dcb A |
1409 | int fields[3]; |
1410 | fields[0] = absOffset / MILLIS_PER_HOUR; | |
1411 | absOffset = absOffset % MILLIS_PER_HOUR; | |
1412 | fields[1] = absOffset / MILLIS_PER_MINUTE; | |
1413 | absOffset = absOffset % MILLIS_PER_MINUTE; | |
1414 | fields[2] = absOffset / MILLIS_PER_SECOND; | |
1415 | ||
1416 | U_ASSERT(fields[0] >= 0 && fields[0] <= MAX_OFFSET_HOUR); | |
1417 | U_ASSERT(fields[1] >= 0 && fields[1] <= MAX_OFFSET_MINUTE); | |
1418 | U_ASSERT(fields[2] >= 0 && fields[2] <= MAX_OFFSET_SECOND); | |
1419 | ||
1420 | int32_t lastIdx = maxFields; | |
1421 | while (lastIdx > minFields) { | |
1422 | if (fields[lastIdx] != 0) { | |
1423 | break; | |
1424 | } | |
1425 | lastIdx--; | |
1426 | } | |
1427 | ||
1428 | UChar sign = PLUS; | |
1429 | if (offset < 0) { | |
1430 | // if all output fields are 0s, do not use negative sign | |
1431 | for (int32_t idx = 0; idx <= lastIdx; idx++) { | |
1432 | if (fields[idx] != 0) { | |
1433 | sign = MINUS; | |
1434 | break; | |
1435 | } | |
1436 | } | |
1437 | } | |
1438 | result.setTo(sign); | |
1439 | ||
1440 | for (int32_t idx = 0; idx <= lastIdx; idx++) { | |
1441 | if (sep && idx != 0) { | |
1442 | result.append(sep); | |
1443 | } | |
1444 | result.append((UChar)(0x0030 + fields[idx]/10)); | |
1445 | result.append((UChar)(0x0030 + fields[idx]%10)); | |
1446 | } | |
1447 | ||
1448 | return result; | |
4388f060 A |
1449 | } |
1450 | ||
1451 | UnicodeString& | |
51004dcb | 1452 | TimeZoneFormat::formatOffsetLocalizedGMT(int32_t offset, UBool isShort, UnicodeString& result, UErrorCode& status) const { |
4388f060 A |
1453 | if (U_FAILURE(status)) { |
1454 | result.setToBogus(); | |
1455 | return result; | |
1456 | } | |
1457 | if (offset <= -MAX_OFFSET || offset >= MAX_OFFSET) { | |
1458 | result.setToBogus(); | |
1459 | status = U_ILLEGAL_ARGUMENT_ERROR; | |
1460 | return result; | |
1461 | } | |
1462 | ||
1463 | if (offset == 0) { | |
1464 | result.setTo(fGMTZeroFormat); | |
1465 | return result; | |
1466 | } | |
1467 | ||
1468 | UBool positive = TRUE; | |
1469 | if (offset < 0) { | |
1470 | offset = -offset; | |
1471 | positive = FALSE; | |
1472 | } | |
1473 | ||
1474 | int32_t offsetH = offset / MILLIS_PER_HOUR; | |
1475 | offset = offset % MILLIS_PER_HOUR; | |
1476 | int32_t offsetM = offset / MILLIS_PER_MINUTE; | |
1477 | offset = offset % MILLIS_PER_MINUTE; | |
1478 | int32_t offsetS = offset / MILLIS_PER_SECOND; | |
1479 | ||
1480 | U_ASSERT(offsetH <= MAX_OFFSET_HOUR && offsetM <= MAX_OFFSET_MINUTE && offsetS <= MAX_OFFSET_SECOND); | |
1481 | ||
1482 | const UVector* offsetPatternItems = NULL; | |
1483 | if (positive) { | |
51004dcb A |
1484 | if (offsetS != 0) { |
1485 | offsetPatternItems = fGMTOffsetPatternItems[UTZFMT_PAT_POSITIVE_HMS]; | |
1486 | } else if (offsetM != 0 || !isShort) { | |
1487 | offsetPatternItems = fGMTOffsetPatternItems[UTZFMT_PAT_POSITIVE_HM]; | |
1488 | } else { | |
1489 | offsetPatternItems = fGMTOffsetPatternItems[UTZFMT_PAT_POSITIVE_H]; | |
1490 | } | |
4388f060 | 1491 | } else { |
51004dcb A |
1492 | if (offsetS != 0) { |
1493 | offsetPatternItems = fGMTOffsetPatternItems[UTZFMT_PAT_NEGATIVE_HMS]; | |
1494 | } else if (offsetM != 0 || !isShort) { | |
1495 | offsetPatternItems = fGMTOffsetPatternItems[UTZFMT_PAT_NEGATIVE_HM]; | |
1496 | } else { | |
1497 | offsetPatternItems = fGMTOffsetPatternItems[UTZFMT_PAT_NEGATIVE_H]; | |
1498 | } | |
4388f060 A |
1499 | } |
1500 | ||
1501 | U_ASSERT(offsetPatternItems != NULL); | |
1502 | ||
1503 | // Building the GMT format string | |
1504 | result.setTo(fGMTPatternPrefix); | |
1505 | ||
1506 | for (int32_t i = 0; i < offsetPatternItems->size(); i++) { | |
1507 | const GMTOffsetField* item = (GMTOffsetField*)offsetPatternItems->elementAt(i); | |
1508 | GMTOffsetField::FieldType type = item->getType(); | |
1509 | ||
1510 | switch (type) { | |
1511 | case GMTOffsetField::TEXT: | |
1512 | result.append(item->getPatternText(), -1); | |
1513 | break; | |
1514 | ||
1515 | case GMTOffsetField::HOUR: | |
51004dcb | 1516 | appendOffsetDigits(result, offsetH, (isShort ? 1 : 2)); |
4388f060 A |
1517 | break; |
1518 | ||
1519 | case GMTOffsetField::MINUTE: | |
51004dcb | 1520 | appendOffsetDigits(result, offsetM, 2); |
4388f060 A |
1521 | break; |
1522 | ||
1523 | case GMTOffsetField::SECOND: | |
51004dcb | 1524 | appendOffsetDigits(result, offsetS, 2); |
4388f060 A |
1525 | break; |
1526 | } | |
1527 | } | |
1528 | ||
1529 | result.append(fGMTPatternSuffix); | |
1530 | return result; | |
1531 | } | |
1532 | ||
4388f060 A |
1533 | int32_t |
1534 | TimeZoneFormat::parseOffsetISO8601(const UnicodeString& text, ParsePosition& pos, UBool extendedOnly, UBool* hasDigitOffset /* = NULL */) const { | |
1535 | if (hasDigitOffset) { | |
1536 | *hasDigitOffset = FALSE; | |
1537 | } | |
1538 | int32_t start = pos.getIndex(); | |
1539 | if (start >= text.length()) { | |
1540 | pos.setErrorIndex(start); | |
1541 | return 0; | |
1542 | } | |
1543 | ||
1544 | UChar firstChar = text.charAt(start); | |
1545 | if (firstChar == ISO8601_UTC || firstChar == (UChar)(ISO8601_UTC + 0x20)) { | |
1546 | // "Z" (or "z") - indicates UTC | |
1547 | pos.setIndex(start + 1); | |
1548 | return 0; | |
1549 | } | |
1550 | ||
1551 | int32_t sign = 1; | |
1552 | if (firstChar == PLUS) { | |
1553 | sign = 1; | |
1554 | } else if (firstChar == MINUS) { | |
1555 | sign = -1; | |
1556 | } else { | |
1557 | // Not an ISO 8601 offset string | |
1558 | pos.setErrorIndex(start); | |
1559 | return 0; | |
1560 | } | |
1561 | ParsePosition posOffset(start + 1); | |
51004dcb | 1562 | int32_t offset = parseAsciiOffsetFields(text, posOffset, ISO8601_SEP, FIELDS_H, FIELDS_HMS); |
4388f060 A |
1563 | if (posOffset.getErrorIndex() == -1 && !extendedOnly && (posOffset.getIndex() - start <= 3)) { |
1564 | // If the text is successfully parsed as extended format with the options above, it can be also parsed | |
1565 | // as basic format. For example, "0230" can be parsed as offset 2:00 (only first digits are valid for | |
1566 | // extended format), but it can be parsed as offset 2:30 with basic format. We use longer result. | |
1567 | ParsePosition posBasic(start + 1); | |
1568 | int32_t tmpOffset = parseAbuttingAsciiOffsetFields(text, posBasic, FIELDS_H, FIELDS_HMS, FALSE); | |
1569 | if (posBasic.getErrorIndex() == -1 && posBasic.getIndex() > posOffset.getIndex()) { | |
1570 | offset = tmpOffset; | |
1571 | posOffset.setIndex(posBasic.getIndex()); | |
1572 | } | |
1573 | } | |
1574 | ||
1575 | if (posOffset.getErrorIndex() != -1) { | |
1576 | pos.setErrorIndex(start); | |
1577 | return 0; | |
1578 | } | |
1579 | ||
1580 | pos.setIndex(posOffset.getIndex()); | |
1581 | if (hasDigitOffset) { | |
1582 | *hasDigitOffset = TRUE; | |
1583 | } | |
1584 | return sign * offset; | |
1585 | } | |
1586 | ||
1587 | int32_t | |
51004dcb | 1588 | TimeZoneFormat::parseOffsetLocalizedGMT(const UnicodeString& text, ParsePosition& pos, UBool isShort, UBool* hasDigitOffset) const { |
4388f060 | 1589 | int32_t start = pos.getIndex(); |
4388f060 | 1590 | int32_t offset = 0; |
51004dcb | 1591 | int32_t parsedLength = 0; |
4388f060 A |
1592 | |
1593 | if (hasDigitOffset) { | |
1594 | *hasDigitOffset = FALSE; | |
1595 | } | |
1596 | ||
51004dcb | 1597 | offset = parseOffsetLocalizedGMTPattern(text, start, isShort, parsedLength); |
4388f060 | 1598 | |
51004dcb A |
1599 | // For now, parseOffsetLocalizedGMTPattern handles both long and short |
1600 | // formats, no matter isShort is true or false. This might be changed in future | |
1601 | // when strict parsing is necessary, or different set of patterns are used for | |
1602 | // short/long formats. | |
1603 | #if 0 | |
1604 | if (parsedLength == 0) { | |
1605 | offset = parseOffsetLocalizedGMTPattern(text, start, !isShort, parsedLength); | |
1606 | } | |
1607 | #endif | |
4388f060 | 1608 | |
51004dcb | 1609 | if (parsedLength > 0) { |
4388f060 A |
1610 | if (hasDigitOffset) { |
1611 | *hasDigitOffset = TRUE; | |
1612 | } | |
51004dcb | 1613 | pos.setIndex(start + parsedLength); |
4388f060 A |
1614 | return offset; |
1615 | } | |
1616 | ||
1617 | // Try the default patterns | |
4388f060 A |
1618 | offset = parseOffsetDefaultLocalizedGMT(text, start, parsedLength); |
1619 | if (parsedLength > 0) { | |
1620 | if (hasDigitOffset) { | |
1621 | *hasDigitOffset = TRUE; | |
1622 | } | |
1623 | pos.setIndex(start + parsedLength); | |
1624 | return offset; | |
1625 | } | |
1626 | ||
1627 | // Check if this is a GMT zero format | |
1628 | if (text.caseCompare(start, fGMTZeroFormat.length(), fGMTZeroFormat, 0) == 0) { | |
1629 | pos.setIndex(start + fGMTZeroFormat.length()); | |
1630 | return 0; | |
1631 | } | |
1632 | ||
1633 | // Check if this is a default GMT zero format | |
1634 | for (int32_t i = 0; ALT_GMT_STRINGS[i][0] != 0; i++) { | |
1635 | const UChar* defGMTZero = ALT_GMT_STRINGS[i]; | |
1636 | int32_t defGMTZeroLen = u_strlen(defGMTZero); | |
1637 | if (text.caseCompare(start, defGMTZeroLen, defGMTZero, 0) == 0) { | |
1638 | pos.setIndex(start + defGMTZeroLen); | |
1639 | return 0; | |
1640 | } | |
1641 | } | |
1642 | ||
1643 | // Nothing matched | |
1644 | pos.setErrorIndex(start); | |
1645 | return 0; | |
1646 | } | |
1647 | ||
1648 | int32_t | |
51004dcb A |
1649 | TimeZoneFormat::parseOffsetLocalizedGMTPattern(const UnicodeString& text, int32_t start, UBool /*isShort*/, int32_t& parsedLen) const { |
1650 | int32_t idx = start; | |
4388f060 | 1651 | int32_t offset = 0; |
51004dcb A |
1652 | UBool parsed = FALSE; |
1653 | ||
1654 | do { | |
1655 | // Prefix part | |
1656 | int32_t len = fGMTPatternPrefix.length(); | |
1657 | if (len > 0 && text.caseCompare(idx, len, fGMTPatternPrefix, 0) != 0) { | |
1658 | // prefix match failed | |
1659 | break; | |
1660 | } | |
1661 | idx += len; | |
1662 | ||
1663 | // Offset part | |
1664 | offset = parseOffsetFields(text, idx, FALSE, len); | |
1665 | if (len == 0) { | |
1666 | // offset field match failed | |
1667 | break; | |
1668 | } | |
1669 | idx += len; | |
1670 | ||
1671 | len = fGMTPatternSuffix.length(); | |
1672 | if (len > 0 && text.caseCompare(idx, len, fGMTPatternSuffix, 0) != 0) { | |
1673 | // no suffix match | |
1674 | break; | |
1675 | } | |
1676 | idx += len; | |
1677 | parsed = TRUE; | |
1678 | } while (FALSE); | |
1679 | ||
1680 | parsedLen = parsed ? idx - start : 0; | |
1681 | return offset; | |
1682 | } | |
1683 | ||
1684 | int32_t | |
1685 | TimeZoneFormat::parseOffsetFields(const UnicodeString& text, int32_t start, UBool /*isShort*/, int32_t& parsedLen) const { | |
1686 | int32_t outLen = 0; | |
1687 | int32_t offset = 0; | |
1688 | int32_t sign = 1; | |
4388f060 A |
1689 | |
1690 | parsedLen = 0; | |
1691 | ||
51004dcb A |
1692 | int32_t offsetH, offsetM, offsetS; |
1693 | offsetH = offsetM = offsetS = 0; | |
1694 | ||
4388f060 A |
1695 | for (int32_t patidx = 0; PARSE_GMT_OFFSET_TYPES[patidx] >= 0; patidx++) { |
1696 | int32_t gmtPatType = PARSE_GMT_OFFSET_TYPES[patidx]; | |
4388f060 A |
1697 | UVector* items = fGMTOffsetPatternItems[gmtPatType]; |
1698 | U_ASSERT(items != NULL); | |
4388f060 | 1699 | |
51004dcb A |
1700 | outLen = parseOffsetFieldsWithPattern(text, start, items, FALSE, offsetH, offsetM, offsetS); |
1701 | if (outLen > 0) { | |
1702 | sign = (gmtPatType == UTZFMT_PAT_POSITIVE_H || gmtPatType == UTZFMT_PAT_POSITIVE_HM || gmtPatType == UTZFMT_PAT_POSITIVE_HMS) ? | |
1703 | 1 : -1; | |
4388f060 A |
1704 | break; |
1705 | } | |
1706 | } | |
1707 | ||
51004dcb A |
1708 | if (outLen > 0 && fAbuttingOffsetHoursAndMinutes) { |
1709 | // When hours field is sabutting minutes field, | |
1710 | // the parse result above may not be appropriate. | |
1711 | // For example, "01020" is parsed as 01:02: above, | |
1712 | // but it should be parsed as 00:10:20. | |
1713 | int32_t tmpLen = 0; | |
1714 | int32_t tmpSign = 1; | |
1715 | int32_t tmpH, tmpM, tmpS; | |
1716 | ||
1717 | for (int32_t patidx = 0; PARSE_GMT_OFFSET_TYPES[patidx] >= 0; patidx++) { | |
1718 | int32_t gmtPatType = PARSE_GMT_OFFSET_TYPES[patidx]; | |
1719 | UVector* items = fGMTOffsetPatternItems[gmtPatType]; | |
1720 | U_ASSERT(items != NULL); | |
4388f060 | 1721 | |
51004dcb A |
1722 | // forcing parse to use single hour digit |
1723 | tmpLen = parseOffsetFieldsWithPattern(text, start, items, TRUE, tmpH, tmpM, tmpS); | |
1724 | if (tmpLen > 0) { | |
1725 | tmpSign = (gmtPatType == UTZFMT_PAT_POSITIVE_H || gmtPatType == UTZFMT_PAT_POSITIVE_HM || gmtPatType == UTZFMT_PAT_POSITIVE_HMS) ? | |
1726 | 1 : -1; | |
1727 | break; | |
1728 | } | |
1729 | } | |
1730 | if (tmpLen > outLen) { | |
1731 | // Better parse result with single hour digit | |
1732 | outLen = tmpLen; | |
1733 | sign = tmpSign; | |
1734 | offsetH = tmpH; | |
1735 | offsetM = tmpM; | |
1736 | offsetS = tmpS; | |
1737 | } | |
1738 | } | |
1739 | ||
1740 | if (outLen > 0) { | |
1741 | offset = ((((offsetH * 60) + offsetM) * 60) + offsetS) * 1000 * sign; | |
1742 | parsedLen = outLen; | |
4388f060 A |
1743 | } |
1744 | ||
1745 | return offset; | |
1746 | } | |
1747 | ||
51004dcb A |
1748 | int32_t |
1749 | TimeZoneFormat::parseOffsetFieldsWithPattern(const UnicodeString& text, int32_t start, | |
1750 | UVector* patternItems, UBool forceSingleHourDigit, int32_t& hour, int32_t& min, int32_t& sec) const { | |
1751 | UBool failed = FALSE; | |
1752 | int32_t offsetH, offsetM, offsetS; | |
1753 | offsetH = offsetM = offsetS = 0; | |
1754 | int32_t idx = start; | |
1755 | ||
1756 | for (int32_t i = 0; i < patternItems->size(); i++) { | |
57a6839d | 1757 | int32_t len = 0; |
51004dcb A |
1758 | const GMTOffsetField* field = (const GMTOffsetField*)patternItems->elementAt(i); |
1759 | GMTOffsetField::FieldType fieldType = field->getType(); | |
1760 | if (fieldType == GMTOffsetField::TEXT) { | |
1761 | const UChar* patStr = field->getPatternText(); | |
1762 | len = u_strlen(patStr); | |
1763 | if (text.caseCompare(idx, len, patStr, 0) != 0) { | |
1764 | failed = TRUE; | |
1765 | break; | |
1766 | } | |
1767 | idx += len; | |
1768 | } else { | |
1769 | if (fieldType == GMTOffsetField::HOUR) { | |
1770 | uint8_t maxDigits = forceSingleHourDigit ? 1 : 2; | |
1771 | offsetH = parseOffsetFieldWithLocalizedDigits(text, idx, 1, maxDigits, 0, MAX_OFFSET_HOUR, len); | |
1772 | } else if (fieldType == GMTOffsetField::MINUTE) { | |
1773 | offsetM = parseOffsetFieldWithLocalizedDigits(text, idx, 2, 2, 0, MAX_OFFSET_MINUTE, len); | |
1774 | } else if (fieldType == GMTOffsetField::SECOND) { | |
1775 | offsetS = parseOffsetFieldWithLocalizedDigits(text, idx, 2, 2, 0, MAX_OFFSET_SECOND, len); | |
1776 | } | |
1777 | ||
1778 | if (len == 0) { | |
1779 | failed = TRUE; | |
1780 | break; | |
1781 | } | |
1782 | idx += len; | |
1783 | } | |
1784 | } | |
1785 | ||
1786 | if (failed) { | |
1787 | hour = min = sec = 0; | |
1788 | return 0; | |
1789 | } | |
1790 | ||
1791 | hour = offsetH; | |
1792 | min = offsetM; | |
1793 | sec = offsetS; | |
1794 | ||
1795 | return idx - start; | |
1796 | } | |
1797 | ||
4388f060 A |
1798 | int32_t |
1799 | TimeZoneFormat::parseAbuttingOffsetFields(const UnicodeString& text, int32_t start, int32_t& parsedLen) const { | |
1800 | int32_t digits[MAX_OFFSET_DIGITS]; | |
1801 | int32_t parsed[MAX_OFFSET_DIGITS]; // accumulative offsets | |
1802 | ||
1803 | // Parse digits into int[] | |
1804 | int32_t idx = start; | |
1805 | int32_t len = 0; | |
1806 | int32_t numDigits = 0; | |
1807 | for (int32_t i = 0; i < MAX_OFFSET_DIGITS; i++) { | |
1808 | digits[i] = parseSingleLocalizedDigit(text, idx, len); | |
1809 | if (digits[i] < 0) { | |
1810 | break; | |
1811 | } | |
1812 | idx += len; | |
1813 | parsed[i] = idx - start; | |
1814 | numDigits++; | |
1815 | } | |
1816 | ||
1817 | if (numDigits == 0) { | |
1818 | parsedLen = 0; | |
1819 | return 0; | |
1820 | } | |
1821 | ||
1822 | int32_t offset = 0; | |
1823 | while (numDigits > 0) { | |
1824 | int32_t hour = 0; | |
1825 | int32_t min = 0; | |
1826 | int32_t sec = 0; | |
1827 | ||
1828 | U_ASSERT(numDigits > 0 && numDigits <= MAX_OFFSET_DIGITS); | |
1829 | switch (numDigits) { | |
1830 | case 1: // H | |
1831 | hour = digits[0]; | |
1832 | break; | |
1833 | case 2: // HH | |
1834 | hour = digits[0] * 10 + digits[1]; | |
1835 | break; | |
1836 | case 3: // Hmm | |
1837 | hour = digits[0]; | |
1838 | min = digits[1] * 10 + digits[2]; | |
1839 | break; | |
1840 | case 4: // HHmm | |
1841 | hour = digits[0] * 10 + digits[1]; | |
1842 | min = digits[2] * 10 + digits[3]; | |
1843 | break; | |
1844 | case 5: // Hmmss | |
1845 | hour = digits[0]; | |
1846 | min = digits[1] * 10 + digits[2]; | |
1847 | sec = digits[3] * 10 + digits[4]; | |
1848 | break; | |
1849 | case 6: // HHmmss | |
1850 | hour = digits[0] * 10 + digits[1]; | |
1851 | min = digits[2] * 10 + digits[3]; | |
1852 | sec = digits[4] * 10 + digits[5]; | |
1853 | break; | |
1854 | } | |
1855 | if (hour <= MAX_OFFSET_HOUR && min <= MAX_OFFSET_MINUTE && sec <= MAX_OFFSET_SECOND) { | |
1856 | // found a valid combination | |
1857 | offset = hour * MILLIS_PER_HOUR + min * MILLIS_PER_MINUTE + sec * MILLIS_PER_SECOND; | |
1858 | parsedLen = parsed[numDigits - 1]; | |
1859 | break; | |
1860 | } | |
1861 | numDigits--; | |
1862 | } | |
1863 | return offset; | |
1864 | } | |
1865 | ||
1866 | int32_t | |
1867 | TimeZoneFormat::parseOffsetDefaultLocalizedGMT(const UnicodeString& text, int start, int32_t& parsedLen) const { | |
1868 | int32_t idx = start; | |
1869 | int32_t offset = 0; | |
1870 | int32_t parsed = 0; | |
1871 | ||
1872 | do { | |
1873 | // check global default GMT alternatives | |
1874 | int32_t gmtLen = 0; | |
1875 | ||
1876 | for (int32_t i = 0; ALT_GMT_STRINGS[i][0] != 0; i++) { | |
1877 | const UChar* gmt = ALT_GMT_STRINGS[i]; | |
1878 | int32_t len = u_strlen(gmt); | |
1879 | if (text.caseCompare(start, len, gmt, 0) == 0) { | |
1880 | gmtLen = len; | |
1881 | break; | |
1882 | } | |
1883 | } | |
1884 | if (gmtLen == 0) { | |
1885 | break; | |
1886 | } | |
1887 | idx += gmtLen; | |
1888 | ||
1889 | // offset needs a sign char and a digit at minimum | |
1890 | if (idx + 1 >= text.length()) { | |
1891 | break; | |
1892 | } | |
1893 | ||
1894 | // parse sign | |
1895 | int32_t sign = 1; | |
1896 | UChar c = text.charAt(idx); | |
1897 | if (c == PLUS) { | |
1898 | sign = 1; | |
1899 | } else if (c == MINUS) { | |
1900 | sign = -1; | |
1901 | } else { | |
1902 | break; | |
1903 | } | |
1904 | idx++; | |
1905 | ||
1906 | // offset part | |
1907 | // try the default pattern with the separator first | |
1908 | int32_t lenWithSep = 0; | |
1909 | int32_t offsetWithSep = parseDefaultOffsetFields(text, idx, DEFAULT_GMT_OFFSET_SEP, lenWithSep); | |
1910 | if (lenWithSep == text.length() - idx) { | |
1911 | // maximum match | |
1912 | offset = offsetWithSep * sign; | |
1913 | idx += lenWithSep; | |
1914 | } else { | |
1915 | // try abutting field pattern | |
1916 | int32_t lenAbut = 0; | |
1917 | int32_t offsetAbut = parseAbuttingOffsetFields(text, idx, lenAbut); | |
1918 | ||
1919 | if (lenWithSep > lenAbut) { | |
1920 | offset = offsetWithSep * sign; | |
1921 | idx += lenWithSep; | |
1922 | } else { | |
1923 | offset = offsetAbut * sign; | |
1924 | idx += lenAbut; | |
1925 | } | |
1926 | } | |
1927 | parsed = idx - start; | |
1928 | } while (false); | |
1929 | ||
1930 | parsedLen = parsed; | |
1931 | return offset; | |
1932 | } | |
1933 | ||
1934 | int32_t | |
1935 | TimeZoneFormat::parseDefaultOffsetFields(const UnicodeString& text, int32_t start, UChar separator, int32_t& parsedLen) const { | |
1936 | int32_t max = text.length(); | |
1937 | int32_t idx = start; | |
1938 | int32_t len = 0; | |
1939 | int32_t hour = 0, min = 0, sec = 0; | |
1940 | ||
1941 | parsedLen = 0; | |
1942 | ||
1943 | do { | |
1944 | hour = parseOffsetFieldWithLocalizedDigits(text, idx, 1, 2, 0, MAX_OFFSET_HOUR, len); | |
1945 | if (len == 0) { | |
1946 | break; | |
1947 | } | |
1948 | idx += len; | |
1949 | ||
1950 | if (idx + 1 < max && text.charAt(idx) == separator) { | |
1951 | min = parseOffsetFieldWithLocalizedDigits(text, idx + 1, 2, 2, 0, MAX_OFFSET_MINUTE, len); | |
1952 | if (len == 0) { | |
1953 | break; | |
1954 | } | |
1955 | idx += (1 + len); | |
1956 | ||
1957 | if (idx + 1 < max && text.charAt(idx) == separator) { | |
1958 | sec = parseOffsetFieldWithLocalizedDigits(text, idx + 1, 2, 2, 0, MAX_OFFSET_SECOND, len); | |
1959 | if (len == 0) { | |
1960 | break; | |
1961 | } | |
1962 | idx += (1 + len); | |
1963 | } | |
1964 | } | |
1965 | } while (FALSE); | |
1966 | ||
1967 | if (idx == start) { | |
1968 | return 0; | |
1969 | } | |
1970 | ||
1971 | parsedLen = idx - start; | |
1972 | return hour * MILLIS_PER_HOUR + min * MILLIS_PER_MINUTE + sec * MILLIS_PER_SECOND; | |
1973 | } | |
1974 | ||
1975 | int32_t | |
1976 | TimeZoneFormat::parseOffsetFieldWithLocalizedDigits(const UnicodeString& text, int32_t start, uint8_t minDigits, uint8_t maxDigits, uint16_t minVal, uint16_t maxVal, int32_t& parsedLen) const { | |
1977 | parsedLen = 0; | |
1978 | ||
1979 | int32_t decVal = 0; | |
1980 | int32_t numDigits = 0; | |
1981 | int32_t idx = start; | |
1982 | int32_t digitLen = 0; | |
1983 | ||
1984 | while (idx < text.length() && numDigits < maxDigits) { | |
1985 | int32_t digit = parseSingleLocalizedDigit(text, idx, digitLen); | |
1986 | if (digit < 0) { | |
1987 | break; | |
1988 | } | |
1989 | int32_t tmpVal = decVal * 10 + digit; | |
1990 | if (tmpVal > maxVal) { | |
1991 | break; | |
1992 | } | |
1993 | decVal = tmpVal; | |
1994 | numDigits++; | |
1995 | idx += digitLen; | |
1996 | } | |
1997 | ||
1998 | // Note: maxVal is checked in the while loop | |
1999 | if (numDigits < minDigits || decVal < minVal) { | |
2000 | decVal = -1; | |
2001 | numDigits = 0; | |
2002 | } else { | |
2003 | parsedLen = idx - start; | |
2004 | } | |
2005 | ||
2006 | return decVal; | |
2007 | } | |
2008 | ||
2009 | int32_t | |
2010 | TimeZoneFormat::parseSingleLocalizedDigit(const UnicodeString& text, int32_t start, int32_t& len) const { | |
2011 | int32_t digit = -1; | |
2012 | len = 0; | |
2013 | if (start < text.length()) { | |
2014 | UChar32 cp = text.char32At(start); | |
2015 | ||
2016 | // First, try digits configured for this instance | |
2017 | for (int32_t i = 0; i < 10; i++) { | |
2018 | if (cp == fGMTOffsetDigits[i]) { | |
2019 | digit = i; | |
2020 | break; | |
2021 | } | |
2022 | } | |
2023 | // If failed, check if this is a Unicode digit | |
2024 | if (digit < 0) { | |
2025 | int32_t tmp = u_charDigitValue(cp); | |
2026 | digit = (tmp >= 0 && tmp <= 9) ? tmp : -1; | |
2027 | } | |
2028 | ||
2029 | if (digit >= 0) { | |
2030 | int32_t next = text.moveIndex32(start, 1); | |
2031 | len = next - start; | |
2032 | } | |
2033 | } | |
2034 | return digit; | |
2035 | } | |
2036 | ||
2037 | UnicodeString& | |
2038 | TimeZoneFormat::formatOffsetWithAsciiDigits(int32_t offset, UChar sep, OffsetFields minFields, OffsetFields maxFields, UnicodeString& result) { | |
2039 | U_ASSERT(maxFields >= minFields); | |
2040 | U_ASSERT(offset > -MAX_OFFSET && offset < MAX_OFFSET); | |
2041 | ||
2042 | UChar sign = PLUS; | |
2043 | if (offset < 0) { | |
2044 | sign = MINUS; | |
2045 | offset = -offset; | |
2046 | } | |
2047 | result.setTo(sign); | |
2048 | ||
2049 | int fields[3]; | |
2050 | fields[0] = offset / MILLIS_PER_HOUR; | |
2051 | offset = offset % MILLIS_PER_HOUR; | |
2052 | fields[1] = offset / MILLIS_PER_MINUTE; | |
2053 | offset = offset % MILLIS_PER_MINUTE; | |
2054 | fields[2] = offset / MILLIS_PER_SECOND; | |
2055 | ||
2056 | U_ASSERT(fields[0] >= 0 && fields[0] <= MAX_OFFSET_HOUR); | |
2057 | U_ASSERT(fields[1] >= 0 && fields[1] <= MAX_OFFSET_MINUTE); | |
2058 | U_ASSERT(fields[2] >= 0 && fields[2] <= MAX_OFFSET_SECOND); | |
2059 | ||
2060 | int32_t lastIdx = maxFields; | |
2061 | while (lastIdx > minFields) { | |
2062 | if (fields[lastIdx] != 0) { | |
2063 | break; | |
2064 | } | |
2065 | lastIdx--; | |
2066 | } | |
2067 | ||
2068 | for (int32_t idx = 0; idx <= lastIdx; idx++) { | |
2069 | if (sep && idx != 0) { | |
2070 | result.append(sep); | |
2071 | } | |
2072 | result.append((UChar)(0x0030 + fields[idx]/10)); | |
2073 | result.append((UChar)(0x0030 + fields[idx]%10)); | |
2074 | } | |
2075 | ||
2076 | return result; | |
2077 | } | |
2078 | ||
2079 | int32_t | |
2080 | TimeZoneFormat::parseAbuttingAsciiOffsetFields(const UnicodeString& text, ParsePosition& pos, OffsetFields minFields, OffsetFields maxFields, UBool fixedHourWidth) { | |
2081 | int32_t start = pos.getIndex(); | |
2082 | ||
2083 | int32_t minDigits = 2 * (minFields + 1) - (fixedHourWidth ? 0 : 1); | |
2084 | int32_t maxDigits = 2 * (maxFields + 1); | |
2085 | ||
2086 | U_ASSERT(maxDigits <= MAX_OFFSET_DIGITS); | |
2087 | ||
51004dcb | 2088 | int32_t digits[MAX_OFFSET_DIGITS] = {}; |
4388f060 A |
2089 | int32_t numDigits = 0; |
2090 | int32_t idx = start; | |
2091 | while (numDigits < maxDigits && idx < text.length()) { | |
2092 | UChar uch = text.charAt(idx); | |
2093 | int32_t digit = DIGIT_VAL(uch); | |
2094 | if (digit < 0) { | |
2095 | break; | |
2096 | } | |
2097 | digits[numDigits] = digit; | |
2098 | numDigits++; | |
2099 | idx++; | |
2100 | } | |
2101 | ||
2102 | if (fixedHourWidth && (numDigits & 1)) { | |
2103 | // Fixed digits, so the number of digits must be even number. Truncating. | |
2104 | numDigits--; | |
2105 | } | |
2106 | ||
2107 | if (numDigits < minDigits) { | |
2108 | pos.setErrorIndex(start); | |
2109 | return 0; | |
2110 | } | |
2111 | ||
2112 | int32_t hour = 0, min = 0, sec = 0; | |
2113 | UBool bParsed = FALSE; | |
2114 | while (numDigits >= minDigits) { | |
2115 | switch (numDigits) { | |
2116 | case 1: //H | |
2117 | hour = digits[0]; | |
2118 | break; | |
2119 | case 2: //HH | |
2120 | hour = digits[0] * 10 + digits[1]; | |
2121 | break; | |
2122 | case 3: //Hmm | |
2123 | hour = digits[0]; | |
2124 | min = digits[1] * 10 + digits[2]; | |
2125 | break; | |
2126 | case 4: //HHmm | |
2127 | hour = digits[0] * 10 + digits[1]; | |
2128 | min = digits[2] * 10 + digits[3]; | |
2129 | break; | |
2130 | case 5: //Hmmss | |
2131 | hour = digits[0]; | |
2132 | min = digits[1] * 10 + digits[2]; | |
2133 | sec = digits[3] * 10 + digits[4]; | |
2134 | break; | |
2135 | case 6: //HHmmss | |
2136 | hour = digits[0] * 10 + digits[1]; | |
2137 | min = digits[2] * 10 + digits[3]; | |
2138 | sec = digits[4] * 10 + digits[5]; | |
2139 | break; | |
2140 | } | |
2141 | ||
2142 | if (hour <= MAX_OFFSET_HOUR && min <= MAX_OFFSET_MINUTE && sec <= MAX_OFFSET_SECOND) { | |
2143 | // Successfully parsed | |
2144 | bParsed = true; | |
2145 | break; | |
2146 | } | |
2147 | ||
2148 | // Truncating | |
2149 | numDigits -= (fixedHourWidth ? 2 : 1); | |
2150 | hour = min = sec = 0; | |
2151 | } | |
2152 | ||
2153 | if (!bParsed) { | |
2154 | pos.setErrorIndex(start); | |
2155 | return 0; | |
2156 | } | |
2157 | pos.setIndex(start + numDigits); | |
2158 | return ((((hour * 60) + min) * 60) + sec) * 1000; | |
2159 | } | |
2160 | ||
2161 | int32_t | |
51004dcb | 2162 | TimeZoneFormat::parseAsciiOffsetFields(const UnicodeString& text, ParsePosition& pos, UChar sep, OffsetFields minFields, OffsetFields maxFields) { |
4388f060 A |
2163 | int32_t start = pos.getIndex(); |
2164 | int32_t fieldVal[] = {0, 0, 0}; | |
2165 | int32_t fieldLen[] = {0, -1, -1}; | |
2166 | for (int32_t idx = start, fieldIdx = 0; idx < text.length() && fieldIdx <= maxFields; idx++) { | |
2167 | UChar c = text.charAt(idx); | |
2168 | if (c == sep) { | |
51004dcb A |
2169 | if (fieldIdx == 0) { |
2170 | if (fieldLen[0] == 0) { | |
2171 | // no hours field | |
2172 | break; | |
2173 | } | |
4388f060 | 2174 | // 1 digit hour, move to next field |
4388f060 | 2175 | } else { |
51004dcb A |
2176 | if (fieldLen[fieldIdx] != -1) { |
2177 | // premature minute or seconds field | |
2178 | break; | |
2179 | } | |
2180 | fieldLen[fieldIdx] = 0; | |
4388f060 A |
2181 | } |
2182 | continue; | |
51004dcb A |
2183 | } else if (fieldLen[fieldIdx] == -1) { |
2184 | // no separator after 2 digit field | |
2185 | break; | |
4388f060 A |
2186 | } |
2187 | int32_t digit = DIGIT_VAL(c); | |
2188 | if (digit < 0) { | |
2189 | // not a digit | |
2190 | break; | |
2191 | } | |
2192 | fieldVal[fieldIdx] = fieldVal[fieldIdx] * 10 + digit; | |
2193 | fieldLen[fieldIdx]++; | |
2194 | if (fieldLen[fieldIdx] >= 2) { | |
2195 | // parsed 2 digits, move to next field | |
2196 | fieldIdx++; | |
2197 | } | |
2198 | } | |
2199 | ||
2200 | int32_t offset = 0; | |
2201 | int32_t parsedLen = 0; | |
2202 | int32_t parsedFields = -1; | |
2203 | do { | |
2204 | // hour | |
51004dcb | 2205 | if (fieldLen[0] == 0) { |
4388f060 A |
2206 | break; |
2207 | } | |
2208 | if (fieldVal[0] > MAX_OFFSET_HOUR) { | |
4388f060 A |
2209 | offset = (fieldVal[0] / 10) * MILLIS_PER_HOUR; |
2210 | parsedFields = FIELDS_H; | |
2211 | parsedLen = 1; | |
2212 | break; | |
2213 | } | |
2214 | offset = fieldVal[0] * MILLIS_PER_HOUR; | |
2215 | parsedLen = fieldLen[0]; | |
2216 | parsedFields = FIELDS_H; | |
2217 | ||
2218 | // minute | |
2219 | if (fieldLen[1] != 2 || fieldVal[1] > MAX_OFFSET_MINUTE) { | |
2220 | break; | |
2221 | } | |
2222 | offset += fieldVal[1] * MILLIS_PER_MINUTE; | |
2223 | parsedLen += (1 + fieldLen[1]); | |
2224 | parsedFields = FIELDS_HM; | |
2225 | ||
2226 | // second | |
2227 | if (fieldLen[2] != 2 || fieldVal[2] > MAX_OFFSET_SECOND) { | |
2228 | break; | |
2229 | } | |
2230 | offset += fieldVal[2] * MILLIS_PER_SECOND; | |
2231 | parsedLen += (1 + fieldLen[2]); | |
2232 | parsedFields = FIELDS_HMS; | |
2233 | } while (false); | |
2234 | ||
2235 | if (parsedFields < minFields) { | |
2236 | pos.setErrorIndex(start); | |
2237 | return 0; | |
2238 | } | |
2239 | ||
2240 | pos.setIndex(start + parsedLen); | |
2241 | return offset; | |
2242 | } | |
2243 | ||
2244 | void | |
2245 | TimeZoneFormat::appendOffsetDigits(UnicodeString& buf, int32_t n, uint8_t minDigits) const { | |
2246 | U_ASSERT(n >= 0 && n < 60); | |
2247 | int32_t numDigits = n >= 10 ? 2 : 1; | |
2248 | for (int32_t i = 0; i < minDigits - numDigits; i++) { | |
2249 | buf.append(fGMTOffsetDigits[0]); | |
2250 | } | |
2251 | if (numDigits == 2) { | |
2252 | buf.append(fGMTOffsetDigits[n / 10]); | |
2253 | } | |
2254 | buf.append(fGMTOffsetDigits[n % 10]); | |
2255 | } | |
2256 | ||
2257 | // ------------------------------------------------------------------ | |
2258 | // Private misc | |
2259 | void | |
2260 | TimeZoneFormat::initGMTPattern(const UnicodeString& gmtPattern, UErrorCode& status) { | |
2261 | if (U_FAILURE(status)) { | |
2262 | return; | |
2263 | } | |
2264 | // This implementation not perfect, but sufficient practically. | |
2265 | int32_t idx = gmtPattern.indexOf(ARG0, ARG0_LEN, 0); | |
2266 | if (idx < 0) { | |
2267 | status = U_ILLEGAL_ARGUMENT_ERROR; | |
2268 | return; | |
2269 | } | |
2270 | fGMTPattern.setTo(gmtPattern); | |
2271 | unquote(gmtPattern.tempSubString(0, idx), fGMTPatternPrefix); | |
2272 | unquote(gmtPattern.tempSubString(idx + ARG0_LEN), fGMTPatternSuffix); | |
2273 | } | |
2274 | ||
2275 | UnicodeString& | |
2276 | TimeZoneFormat::unquote(const UnicodeString& pattern, UnicodeString& result) { | |
2277 | if (pattern.indexOf(SINGLEQUOTE) < 0) { | |
2278 | result.setTo(pattern); | |
2279 | return result; | |
2280 | } | |
2281 | result.remove(); | |
2282 | UBool isPrevQuote = FALSE; | |
2283 | UBool inQuote = FALSE; | |
2284 | for (int32_t i = 0; i < pattern.length(); i++) { | |
2285 | UChar c = pattern.charAt(i); | |
2286 | if (c == SINGLEQUOTE) { | |
2287 | if (isPrevQuote) { | |
2288 | result.append(c); | |
2289 | isPrevQuote = FALSE; | |
2290 | } else { | |
2291 | isPrevQuote = TRUE; | |
2292 | } | |
2293 | inQuote = !inQuote; | |
2294 | } else { | |
2295 | isPrevQuote = FALSE; | |
2296 | result.append(c); | |
2297 | } | |
2298 | } | |
2299 | return result; | |
2300 | } | |
2301 | ||
2302 | UVector* | |
2303 | TimeZoneFormat::parseOffsetPattern(const UnicodeString& pattern, OffsetFields required, UErrorCode& status) { | |
2304 | if (U_FAILURE(status)) { | |
2305 | return NULL; | |
2306 | } | |
2307 | UVector* result = new UVector(deleteGMTOffsetField, NULL, status); | |
2308 | if (result == NULL) { | |
2309 | status = U_MEMORY_ALLOCATION_ERROR; | |
2310 | return NULL; | |
2311 | } | |
2312 | ||
2313 | int32_t checkBits = 0; | |
2314 | UBool isPrevQuote = FALSE; | |
2315 | UBool inQuote = FALSE; | |
2316 | UnicodeString text; | |
2317 | GMTOffsetField::FieldType itemType = GMTOffsetField::TEXT; | |
2318 | int32_t itemLength = 1; | |
2319 | ||
2320 | for (int32_t i = 0; i < pattern.length(); i++) { | |
2321 | UChar ch = pattern.charAt(i); | |
2322 | if (ch == SINGLEQUOTE) { | |
2323 | if (isPrevQuote) { | |
2324 | text.append(SINGLEQUOTE); | |
2325 | isPrevQuote = FALSE; | |
2326 | } else { | |
2327 | isPrevQuote = TRUE; | |
2328 | if (itemType != GMTOffsetField::TEXT) { | |
2329 | if (GMTOffsetField::isValid(itemType, itemLength)) { | |
2330 | GMTOffsetField* fld = GMTOffsetField::createTimeField(itemType, (uint8_t)itemLength, status); | |
2331 | result->addElement(fld, status); | |
2332 | if (U_FAILURE(status)) { | |
2333 | break; | |
2334 | } | |
2335 | } else { | |
2336 | status = U_ILLEGAL_ARGUMENT_ERROR; | |
2337 | break; | |
2338 | } | |
2339 | itemType = GMTOffsetField::TEXT; | |
2340 | } | |
2341 | } | |
2342 | inQuote = !inQuote; | |
2343 | } else { | |
2344 | isPrevQuote = FALSE; | |
2345 | if (inQuote) { | |
2346 | text.append(ch); | |
2347 | } else { | |
2348 | GMTOffsetField::FieldType tmpType = GMTOffsetField::getTypeByLetter(ch); | |
2349 | if (tmpType != GMTOffsetField::TEXT) { | |
2350 | // an offset time pattern character | |
2351 | if (tmpType == itemType) { | |
2352 | itemLength++; | |
2353 | } else { | |
2354 | if (itemType == GMTOffsetField::TEXT) { | |
2355 | if (text.length() > 0) { | |
2356 | GMTOffsetField* textfld = GMTOffsetField::createText(text, status); | |
2357 | result->addElement(textfld, status); | |
2358 | if (U_FAILURE(status)) { | |
2359 | break; | |
2360 | } | |
2361 | text.remove(); | |
2362 | } | |
2363 | } else { | |
2364 | if (GMTOffsetField::isValid(itemType, itemLength)) { | |
2365 | GMTOffsetField* fld = GMTOffsetField::createTimeField(itemType, itemLength, status); | |
2366 | result->addElement(fld, status); | |
2367 | if (U_FAILURE(status)) { | |
2368 | break; | |
2369 | } | |
2370 | } else { | |
2371 | status = U_ILLEGAL_ARGUMENT_ERROR; | |
2372 | break; | |
2373 | } | |
2374 | } | |
2375 | itemType = tmpType; | |
2376 | itemLength = 1; | |
2377 | checkBits |= tmpType; | |
2378 | } | |
2379 | } else { | |
2380 | // a string literal | |
2381 | if (itemType != GMTOffsetField::TEXT) { | |
2382 | if (GMTOffsetField::isValid(itemType, itemLength)) { | |
2383 | GMTOffsetField* fld = GMTOffsetField::createTimeField(itemType, itemLength, status); | |
2384 | result->addElement(fld, status); | |
2385 | if (U_FAILURE(status)) { | |
2386 | break; | |
2387 | } | |
2388 | } else { | |
2389 | status = U_ILLEGAL_ARGUMENT_ERROR; | |
2390 | break; | |
2391 | } | |
2392 | itemType = GMTOffsetField::TEXT; | |
2393 | } | |
2394 | text.append(ch); | |
2395 | } | |
2396 | } | |
2397 | } | |
2398 | } | |
2399 | // handle last item | |
2400 | if (U_SUCCESS(status)) { | |
2401 | if (itemType == GMTOffsetField::TEXT) { | |
2402 | if (text.length() > 0) { | |
2403 | GMTOffsetField* tfld = GMTOffsetField::createText(text, status); | |
2404 | result->addElement(tfld, status); | |
2405 | } | |
2406 | } else { | |
2407 | if (GMTOffsetField::isValid(itemType, itemLength)) { | |
2408 | GMTOffsetField* fld = GMTOffsetField::createTimeField(itemType, itemLength, status); | |
2409 | result->addElement(fld, status); | |
2410 | } else { | |
2411 | status = U_ILLEGAL_ARGUMENT_ERROR; | |
2412 | } | |
2413 | } | |
2414 | ||
2415 | // Check all required fields are set | |
2416 | if (U_SUCCESS(status)) { | |
2417 | int32_t reqBits = 0; | |
2418 | switch (required) { | |
2419 | case FIELDS_H: | |
2420 | reqBits = GMTOffsetField::HOUR; | |
2421 | break; | |
2422 | case FIELDS_HM: | |
2423 | reqBits = GMTOffsetField::HOUR | GMTOffsetField::MINUTE; | |
2424 | break; | |
2425 | case FIELDS_HMS: | |
2426 | reqBits = GMTOffsetField::HOUR | GMTOffsetField::MINUTE | GMTOffsetField::SECOND; | |
2427 | break; | |
2428 | } | |
2429 | if (checkBits == reqBits) { | |
2430 | // all required fields are set, no extra fields | |
2431 | return result; | |
2432 | } | |
2433 | } | |
2434 | } | |
2435 | ||
2436 | // error | |
2437 | delete result; | |
2438 | return NULL; | |
2439 | } | |
2440 | ||
2441 | UnicodeString& | |
51004dcb A |
2442 | TimeZoneFormat::expandOffsetPattern(const UnicodeString& offsetHM, UnicodeString& result, UErrorCode& status) { |
2443 | result.setToBogus(); | |
2444 | if (U_FAILURE(status)) { | |
2445 | return result; | |
2446 | } | |
4388f060 A |
2447 | U_ASSERT(u_strlen(DEFAULT_GMT_OFFSET_MINUTE_PATTERN) == 2); |
2448 | ||
2449 | int32_t idx_mm = offsetHM.indexOf(DEFAULT_GMT_OFFSET_MINUTE_PATTERN, 2, 0); | |
2450 | if (idx_mm < 0) { | |
51004dcb A |
2451 | // Bad time zone hour pattern data |
2452 | status = U_ILLEGAL_ARGUMENT_ERROR; | |
4388f060 A |
2453 | return result; |
2454 | } | |
2455 | ||
2456 | UnicodeString sep; | |
51004dcb | 2457 | int32_t idx_H = offsetHM.tempSubString(0, idx_mm).lastIndexOf((UChar)0x0048 /* H */); |
4388f060 A |
2458 | if (idx_H >= 0) { |
2459 | sep = offsetHM.tempSubString(idx_H + 1, idx_mm - (idx_H + 1)); | |
2460 | } | |
2461 | result.setTo(offsetHM.tempSubString(0, idx_mm + 2)); | |
2462 | result.append(sep); | |
2463 | result.append(DEFAULT_GMT_OFFSET_SECOND_PATTERN, -1); | |
2464 | result.append(offsetHM.tempSubString(idx_mm + 2)); | |
2465 | return result; | |
2466 | } | |
2467 | ||
51004dcb A |
2468 | UnicodeString& |
2469 | TimeZoneFormat::truncateOffsetPattern(const UnicodeString& offsetHM, UnicodeString& result, UErrorCode& status) { | |
2470 | result.setToBogus(); | |
2471 | if (U_FAILURE(status)) { | |
2472 | return result; | |
2473 | } | |
2474 | U_ASSERT(u_strlen(DEFAULT_GMT_OFFSET_MINUTE_PATTERN) == 2); | |
2475 | ||
2476 | int32_t idx_mm = offsetHM.indexOf(DEFAULT_GMT_OFFSET_MINUTE_PATTERN, 2, 0); | |
2477 | if (idx_mm < 0) { | |
2478 | // Bad time zone hour pattern data | |
2479 | status = U_ILLEGAL_ARGUMENT_ERROR; | |
2480 | return result; | |
2481 | } | |
2482 | UChar HH[] = {0x0048, 0x0048}; | |
2483 | int32_t idx_HH = offsetHM.tempSubString(0, idx_mm).lastIndexOf(HH, 2, 0); | |
2484 | if (idx_HH >= 0) { | |
2485 | return result.setTo(offsetHM.tempSubString(0, idx_HH + 2)); | |
2486 | } | |
2487 | int32_t idx_H = offsetHM.tempSubString(0, idx_mm).lastIndexOf((UChar)0x0048, 0); | |
2488 | if (idx_H >= 0) { | |
2489 | return result.setTo(offsetHM.tempSubString(0, idx_H + 1)); | |
2490 | } | |
2491 | // Bad time zone hour pattern data | |
2492 | status = U_ILLEGAL_ARGUMENT_ERROR; | |
2493 | return result; | |
2494 | } | |
2495 | ||
4388f060 A |
2496 | void |
2497 | TimeZoneFormat::initGMTOffsetPatterns(UErrorCode& status) { | |
51004dcb | 2498 | for (int32_t type = 0; type < UTZFMT_PAT_COUNT; type++) { |
4388f060 | 2499 | switch (type) { |
51004dcb A |
2500 | case UTZFMT_PAT_POSITIVE_H: |
2501 | case UTZFMT_PAT_NEGATIVE_H: | |
2502 | fGMTOffsetPatternItems[type] = parseOffsetPattern(fGMTOffsetPatterns[type], FIELDS_H, status); | |
2503 | break; | |
4388f060 A |
2504 | case UTZFMT_PAT_POSITIVE_HM: |
2505 | case UTZFMT_PAT_NEGATIVE_HM: | |
2506 | fGMTOffsetPatternItems[type] = parseOffsetPattern(fGMTOffsetPatterns[type], FIELDS_HM, status); | |
2507 | break; | |
2508 | case UTZFMT_PAT_POSITIVE_HMS: | |
2509 | case UTZFMT_PAT_NEGATIVE_HMS: | |
2510 | fGMTOffsetPatternItems[type] = parseOffsetPattern(fGMTOffsetPatterns[type], FIELDS_HMS, status); | |
2511 | break; | |
2512 | } | |
2513 | } | |
51004dcb A |
2514 | checkAbuttingHoursAndMinutes(); |
2515 | } | |
2516 | ||
2517 | void | |
2518 | TimeZoneFormat::checkAbuttingHoursAndMinutes() { | |
2519 | fAbuttingOffsetHoursAndMinutes= FALSE; | |
2520 | for (int32_t type = 0; type < UTZFMT_PAT_COUNT; type++) { | |
2521 | UBool afterH = FALSE; | |
2522 | UVector *items = fGMTOffsetPatternItems[type]; | |
2523 | for (int32_t i = 0; i < items->size(); i++) { | |
2524 | const GMTOffsetField* item = (GMTOffsetField*)items->elementAt(i); | |
2525 | GMTOffsetField::FieldType type = item->getType(); | |
2526 | if (type != GMTOffsetField::TEXT) { | |
2527 | if (afterH) { | |
2528 | fAbuttingOffsetHoursAndMinutes = TRUE; | |
2529 | break; | |
2530 | } else if (type == GMTOffsetField::HOUR) { | |
2531 | afterH = TRUE; | |
2532 | } | |
2533 | } else if (afterH) { | |
2534 | break; | |
2535 | } | |
2536 | } | |
2537 | if (fAbuttingOffsetHoursAndMinutes) { | |
2538 | break; | |
2539 | } | |
2540 | } | |
4388f060 A |
2541 | } |
2542 | ||
2543 | UBool | |
2544 | TimeZoneFormat::toCodePoints(const UnicodeString& str, UChar32* codeArray, int32_t size) { | |
2545 | int32_t count = str.countChar32(); | |
2546 | if (count != size) { | |
2547 | return FALSE; | |
2548 | } | |
2549 | ||
2550 | for (int32_t idx = 0, start = 0; idx < size; idx++) { | |
2551 | codeArray[idx] = str.char32At(start); | |
2552 | start = str.moveIndex32(start, 1); | |
2553 | } | |
2554 | ||
2555 | return TRUE; | |
2556 | } | |
2557 | ||
2558 | TimeZone* | |
2559 | TimeZoneFormat::createTimeZoneForOffset(int32_t offset) const { | |
2560 | if (offset == 0) { | |
2561 | // when offset is 0, we should use "Etc/GMT" | |
2562 | return TimeZone::createTimeZone(UnicodeString(TZID_GMT)); | |
2563 | } | |
2564 | return ZoneMeta::createCustomTimeZone(offset); | |
2565 | } | |
2566 | ||
2567 | UTimeZoneFormatTimeType | |
2568 | TimeZoneFormat::getTimeType(UTimeZoneNameType nameType) { | |
2569 | switch (nameType) { | |
2570 | case UTZNM_LONG_STANDARD: | |
2571 | case UTZNM_SHORT_STANDARD: | |
2572 | return UTZFMT_TIME_TYPE_STANDARD; | |
2573 | ||
2574 | case UTZNM_LONG_DAYLIGHT: | |
2575 | case UTZNM_SHORT_DAYLIGHT: | |
2576 | return UTZFMT_TIME_TYPE_DAYLIGHT; | |
2577 | ||
2578 | default: | |
2579 | U_ASSERT(FALSE); | |
2580 | } | |
2581 | return UTZFMT_TIME_TYPE_UNKNOWN; | |
2582 | } | |
2583 | ||
2584 | UnicodeString& | |
2585 | TimeZoneFormat::getTimeZoneID(const TimeZoneNames::MatchInfoCollection* matches, int32_t idx, UnicodeString& tzID) const { | |
2586 | if (!matches->getTimeZoneIDAt(idx, tzID)) { | |
2587 | UnicodeString mzID; | |
2588 | if (matches->getMetaZoneIDAt(idx, mzID)) { | |
2589 | fTimeZoneNames->getReferenceZoneID(mzID, fTargetRegion, tzID); | |
2590 | } | |
2591 | } | |
2592 | return tzID; | |
2593 | } | |
2594 | ||
51004dcb A |
2595 | |
2596 | class ZoneIdMatchHandler : public TextTrieMapSearchResultHandler { | |
2597 | public: | |
2598 | ZoneIdMatchHandler(); | |
2599 | virtual ~ZoneIdMatchHandler(); | |
2600 | ||
2601 | UBool handleMatch(int32_t matchLength, const CharacterNode *node, UErrorCode &status); | |
2602 | const UChar* getID(); | |
2603 | int32_t getMatchLen(); | |
2604 | private: | |
2605 | int32_t fLen; | |
2606 | const UChar* fID; | |
2607 | }; | |
2608 | ||
2609 | ZoneIdMatchHandler::ZoneIdMatchHandler() | |
2610 | : fLen(0), fID(NULL) { | |
2611 | } | |
2612 | ||
2613 | ZoneIdMatchHandler::~ZoneIdMatchHandler() { | |
2614 | } | |
2615 | ||
2616 | UBool | |
2617 | ZoneIdMatchHandler::handleMatch(int32_t matchLength, const CharacterNode *node, UErrorCode &status) { | |
2618 | if (U_FAILURE(status)) { | |
2619 | return FALSE; | |
2620 | } | |
2621 | if (node->hasValues()) { | |
2622 | const UChar* id = (const UChar*)node->getValue(0); | |
2623 | if (id != NULL) { | |
2624 | if (fLen < matchLength) { | |
2625 | fID = id; | |
2626 | fLen = matchLength; | |
2627 | } | |
2628 | } | |
2629 | } | |
2630 | return TRUE; | |
2631 | } | |
2632 | ||
2633 | const UChar* | |
2634 | ZoneIdMatchHandler::getID() { | |
2635 | return fID; | |
2636 | } | |
2637 | ||
2638 | int32_t | |
2639 | ZoneIdMatchHandler::getMatchLen() { | |
2640 | return fLen; | |
2641 | } | |
2642 | ||
57a6839d A |
2643 | |
2644 | static void U_CALLCONV initZoneIdTrie(UErrorCode &status) { | |
2645 | U_ASSERT(gZoneIdTrie == NULL); | |
2646 | ucln_i18n_registerCleanup(UCLN_I18N_TIMEZONEFORMAT, tzfmt_cleanup); | |
2647 | gZoneIdTrie = new TextTrieMap(TRUE, NULL); // No deleter, because values are pooled by ZoneMeta | |
2648 | if (gZoneIdTrie == NULL) { | |
2649 | status = U_MEMORY_ALLOCATION_ERROR; | |
2650 | return; | |
2651 | } | |
2652 | StringEnumeration *tzenum = TimeZone::createEnumeration(); | |
2653 | const UnicodeString *id; | |
2654 | while ((id = tzenum->snext(status))) { | |
2655 | const UChar* uid = ZoneMeta::findTimeZoneID(*id); | |
2656 | if (uid) { | |
2657 | gZoneIdTrie->put(uid, const_cast<UChar *>(uid), status); | |
2658 | } | |
2659 | } | |
2660 | delete tzenum; | |
2661 | } | |
2662 | ||
2663 | ||
51004dcb A |
2664 | UnicodeString& |
2665 | TimeZoneFormat::parseZoneID(const UnicodeString& text, ParsePosition& pos, UnicodeString& tzID) const { | |
2666 | UErrorCode status = U_ZERO_ERROR; | |
57a6839d | 2667 | umtx_initOnce(gZoneIdTrieInitOnce, &initZoneIdTrie, status); |
51004dcb A |
2668 | |
2669 | int32_t start = pos.getIndex(); | |
2670 | int32_t len = 0; | |
2671 | tzID.setToBogus(); | |
2672 | ||
57a6839d | 2673 | if (U_SUCCESS(status)) { |
51004dcb A |
2674 | LocalPointer<ZoneIdMatchHandler> handler(new ZoneIdMatchHandler()); |
2675 | gZoneIdTrie->search(text, start, handler.getAlias(), status); | |
2676 | len = handler->getMatchLen(); | |
2677 | if (len > 0) { | |
2678 | tzID.setTo(handler->getID(), -1); | |
2679 | } | |
2680 | } | |
2681 | ||
2682 | if (len > 0) { | |
2683 | pos.setIndex(start + len); | |
2684 | } else { | |
2685 | pos.setErrorIndex(start); | |
2686 | } | |
2687 | ||
2688 | return tzID; | |
2689 | } | |
2690 | ||
57a6839d A |
2691 | static void U_CALLCONV initShortZoneIdTrie(UErrorCode &status) { |
2692 | U_ASSERT(gShortZoneIdTrie == NULL); | |
2693 | ucln_i18n_registerCleanup(UCLN_I18N_TIMEZONEFORMAT, tzfmt_cleanup); | |
2694 | StringEnumeration *tzenum = TimeZone::createTimeZoneIDEnumeration(UCAL_ZONE_TYPE_CANONICAL, NULL, NULL, status); | |
2695 | if (U_SUCCESS(status)) { | |
2696 | gShortZoneIdTrie = new TextTrieMap(TRUE, NULL); // No deleter, because values are pooled by ZoneMeta | |
2697 | if (gShortZoneIdTrie == NULL) { | |
2698 | status = U_MEMORY_ALLOCATION_ERROR; | |
2699 | } else { | |
2700 | const UnicodeString *id; | |
2701 | while ((id = tzenum->snext(status))) { | |
2702 | const UChar* uID = ZoneMeta::findTimeZoneID(*id); | |
2703 | const UChar* shortID = ZoneMeta::getShortID(*id); | |
2704 | if (shortID && uID) { | |
2705 | gShortZoneIdTrie->put(shortID, const_cast<UChar *>(uID), status); | |
51004dcb | 2706 | } |
51004dcb A |
2707 | } |
2708 | } | |
51004dcb | 2709 | } |
57a6839d A |
2710 | delete tzenum; |
2711 | } | |
2712 | ||
2713 | ||
2714 | UnicodeString& | |
2715 | TimeZoneFormat::parseShortZoneID(const UnicodeString& text, ParsePosition& pos, UnicodeString& tzID) const { | |
2716 | UErrorCode status = U_ZERO_ERROR; | |
2717 | umtx_initOnce(gShortZoneIdTrieInitOnce, &initShortZoneIdTrie, status); | |
51004dcb A |
2718 | |
2719 | int32_t start = pos.getIndex(); | |
2720 | int32_t len = 0; | |
2721 | tzID.setToBogus(); | |
2722 | ||
57a6839d | 2723 | if (U_SUCCESS(status)) { |
51004dcb A |
2724 | LocalPointer<ZoneIdMatchHandler> handler(new ZoneIdMatchHandler()); |
2725 | gShortZoneIdTrie->search(text, start, handler.getAlias(), status); | |
2726 | len = handler->getMatchLen(); | |
2727 | if (len > 0) { | |
2728 | tzID.setTo(handler->getID(), -1); | |
2729 | } | |
2730 | } | |
2731 | ||
2732 | if (len > 0) { | |
2733 | pos.setIndex(start + len); | |
2734 | } else { | |
2735 | pos.setErrorIndex(start); | |
2736 | } | |
2737 | ||
2738 | return tzID; | |
2739 | } | |
2740 | ||
2741 | ||
2742 | UnicodeString& | |
2743 | TimeZoneFormat::parseExemplarLocation(const UnicodeString& text, ParsePosition& pos, UnicodeString& tzID) const { | |
2744 | int32_t startIdx = pos.getIndex(); | |
2745 | int32_t parsedPos = -1; | |
2746 | tzID.setToBogus(); | |
2747 | ||
2748 | UErrorCode status = U_ZERO_ERROR; | |
2749 | LocalPointer<TimeZoneNames::MatchInfoCollection> exemplarMatches(fTimeZoneNames->find(text, startIdx, UTZNM_EXEMPLAR_LOCATION, status)); | |
2750 | if (U_FAILURE(status)) { | |
2751 | pos.setErrorIndex(startIdx); | |
2752 | return tzID; | |
2753 | } | |
2754 | int32_t matchIdx = -1; | |
2755 | if (!exemplarMatches.isNull()) { | |
2756 | for (int32_t i = 0; i < exemplarMatches->size(); i++) { | |
2757 | if (startIdx + exemplarMatches->getMatchLengthAt(i) > parsedPos) { | |
2758 | matchIdx = i; | |
2759 | parsedPos = startIdx + exemplarMatches->getMatchLengthAt(i); | |
2760 | } | |
2761 | } | |
2762 | if (parsedPos > 0) { | |
2763 | pos.setIndex(parsedPos); | |
2764 | getTimeZoneID(exemplarMatches.getAlias(), matchIdx, tzID); | |
2765 | } | |
2766 | } | |
2767 | ||
2768 | if (tzID.length() == 0) { | |
2769 | pos.setErrorIndex(startIdx); | |
2770 | } | |
2771 | ||
2772 | return tzID; | |
2773 | } | |
2774 | ||
4388f060 A |
2775 | U_NAMESPACE_END |
2776 | ||
2777 | #endif |