]>
Commit | Line | Data |
---|---|---|
f3c0d7a5 A |
1 | // © 2016 and later: Unicode, Inc. and others. |
2 | // License & terms of use: http://www.unicode.org/copyright.html | |
46f4442e A |
3 | /* |
4 | ******************************************************************************* | |
2ca993e8 | 5 | * Copyright (C) 2007-2016, International Business Machines Corporation and |
46f4442e A |
6 | * others. All Rights Reserved. |
7 | ******************************************************************************* | |
8 | * | |
9 | * File DTPTNGEN.CPP | |
10 | * | |
11 | ******************************************************************************* | |
12 | */ | |
13 | ||
14 | #include "unicode/utypes.h" | |
15 | #if !UCONFIG_NO_FORMATTING | |
16 | ||
17 | #include "unicode/datefmt.h" | |
18 | #include "unicode/decimfmt.h" | |
19 | #include "unicode/dtfmtsym.h" | |
20 | #include "unicode/dtptngen.h" | |
2ca993e8 | 21 | #include "unicode/simpleformatter.h" |
46f4442e A |
22 | #include "unicode/smpdtfmt.h" |
23 | #include "unicode/udat.h" | |
24 | #include "unicode/udatpg.h" | |
25 | #include "unicode/uniset.h" | |
26 | #include "unicode/uloc.h" | |
27 | #include "unicode/ures.h" | |
28 | #include "unicode/ustring.h" | |
29 | #include "unicode/rep.h" | |
30 | #include "cpputils.h" | |
46f4442e | 31 | #include "mutex.h" |
2ca993e8 | 32 | #include "umutex.h" |
46f4442e A |
33 | #include "cmemory.h" |
34 | #include "cstring.h" | |
35 | #include "locbased.h" | |
46f4442e | 36 | #include "hash.h" |
2ca993e8 | 37 | #include "uhash.h" |
46f4442e A |
38 | #include "uresimp.h" |
39 | #include "dtptngen_impl.h" | |
2ca993e8 A |
40 | #include "ucln_in.h" |
41 | #include "charstr.h" | |
f3c0d7a5 | 42 | #include "uassert.h" |
46f4442e A |
43 | |
44 | #if U_CHARSET_FAMILY==U_EBCDIC_FAMILY | |
45 | /** | |
729e4ab9 | 46 | * If we are on EBCDIC, use an iterator which will |
46f4442e A |
47 | * traverse the bundles in ASCII order. |
48 | */ | |
49 | #define U_USE_ASCII_BUNDLE_ITERATOR | |
50 | #define U_SORT_ASCII_BUNDLE_ITERATOR | |
51 | #endif | |
52 | ||
53 | #if defined(U_USE_ASCII_BUNDLE_ITERATOR) | |
54 | ||
55 | #include "unicode/ustring.h" | |
56 | #include "uarrsort.h" | |
57 | ||
58 | struct UResAEntry { | |
59 | UChar *key; | |
60 | UResourceBundle *item; | |
61 | }; | |
62 | ||
63 | struct UResourceBundleAIterator { | |
64 | UResourceBundle *bund; | |
65 | UResAEntry *entries; | |
66 | int32_t num; | |
67 | int32_t cursor; | |
68 | }; | |
69 | ||
70 | /* Must be C linkage to pass function pointer to the sort function */ | |
71 | ||
4388f060 A |
72 | U_CDECL_BEGIN |
73 | ||
729e4ab9 | 74 | static int32_t U_CALLCONV |
46f4442e A |
75 | ures_a_codepointSort(const void *context, const void *left, const void *right) { |
76 | //CompareContext *cmp=(CompareContext *)context; | |
77 | return u_strcmp(((const UResAEntry *)left)->key, | |
78 | ((const UResAEntry *)right)->key); | |
79 | } | |
80 | ||
4388f060 | 81 | U_CDECL_END |
46f4442e A |
82 | |
83 | static void ures_a_open(UResourceBundleAIterator *aiter, UResourceBundle *bund, UErrorCode *status) { | |
84 | if(U_FAILURE(*status)) { | |
85 | return; | |
86 | } | |
87 | aiter->bund = bund; | |
88 | aiter->num = ures_getSize(aiter->bund); | |
89 | aiter->cursor = 0; | |
90 | #if !defined(U_SORT_ASCII_BUNDLE_ITERATOR) | |
91 | aiter->entries = NULL; | |
92 | #else | |
93 | aiter->entries = (UResAEntry*)uprv_malloc(sizeof(UResAEntry)*aiter->num); | |
94 | for(int i=0;i<aiter->num;i++) { | |
95 | aiter->entries[i].item = ures_getByIndex(aiter->bund, i, NULL, status); | |
96 | const char *akey = ures_getKey(aiter->entries[i].item); | |
97 | int32_t len = uprv_strlen(akey)+1; | |
98 | aiter->entries[i].key = (UChar*)uprv_malloc(len*sizeof(UChar)); | |
99 | u_charsToUChars(akey, aiter->entries[i].key, len); | |
100 | } | |
101 | uprv_sortArray(aiter->entries, aiter->num, sizeof(UResAEntry), ures_a_codepointSort, NULL, TRUE, status); | |
102 | #endif | |
103 | } | |
104 | ||
105 | static void ures_a_close(UResourceBundleAIterator *aiter) { | |
106 | #if defined(U_SORT_ASCII_BUNDLE_ITERATOR) | |
107 | for(int i=0;i<aiter->num;i++) { | |
108 | uprv_free(aiter->entries[i].key); | |
109 | ures_close(aiter->entries[i].item); | |
110 | } | |
111 | #endif | |
112 | } | |
113 | ||
114 | static const UChar *ures_a_getNextString(UResourceBundleAIterator *aiter, int32_t *len, const char **key, UErrorCode *err) { | |
115 | #if !defined(U_SORT_ASCII_BUNDLE_ITERATOR) | |
116 | return ures_getNextString(aiter->bund, len, key, err); | |
117 | #else | |
118 | if(U_FAILURE(*err)) return NULL; | |
119 | UResourceBundle *item = aiter->entries[aiter->cursor].item; | |
120 | const UChar* ret = ures_getString(item, len, err); | |
121 | *key = ures_getKey(item); | |
122 | aiter->cursor++; | |
123 | return ret; | |
124 | #endif | |
125 | } | |
126 | ||
127 | ||
128 | #endif | |
129 | ||
130 | ||
131 | U_NAMESPACE_BEGIN | |
132 | ||
46f4442e A |
133 | // ***************************************************************************** |
134 | // class DateTimePatternGenerator | |
135 | // ***************************************************************************** | |
136 | static const UChar Canonical_Items[] = { | |
729e4ab9 A |
137 | // GyQMwWEdDFHmsSv |
138 | CAP_G, LOW_Y, CAP_Q, CAP_M, LOW_W, CAP_W, CAP_E, LOW_D, CAP_D, CAP_F, | |
46f4442e A |
139 | CAP_H, LOW_M, LOW_S, CAP_S, LOW_V, 0 |
140 | }; | |
141 | ||
142 | static const dtTypeElem dtTypes[] = { | |
143 | // patternChar, field, type, minLen, weight | |
144 | {CAP_G, UDATPG_ERA_FIELD, DT_SHORT, 1, 3,}, | |
145 | {CAP_G, UDATPG_ERA_FIELD, DT_LONG, 4, 0}, | |
146 | {LOW_Y, UDATPG_YEAR_FIELD, DT_NUMERIC, 1, 20}, | |
147 | {CAP_Y, UDATPG_YEAR_FIELD, DT_NUMERIC + DT_DELTA, 1, 20}, | |
148 | {LOW_U, UDATPG_YEAR_FIELD, DT_NUMERIC + 2*DT_DELTA, 1, 20}, | |
57a6839d | 149 | {LOW_R, UDATPG_YEAR_FIELD, DT_NUMERIC + 3*DT_DELTA, 1, 20}, |
4388f060 A |
150 | {CAP_U, UDATPG_YEAR_FIELD, DT_SHORT, 1, 3}, |
151 | {CAP_U, UDATPG_YEAR_FIELD, DT_LONG, 4, 0}, | |
152 | {CAP_U, UDATPG_YEAR_FIELD, DT_NARROW, 5, 0}, | |
46f4442e A |
153 | {CAP_Q, UDATPG_QUARTER_FIELD, DT_NUMERIC, 1, 2}, |
154 | {CAP_Q, UDATPG_QUARTER_FIELD, DT_SHORT, 3, 0}, | |
155 | {CAP_Q, UDATPG_QUARTER_FIELD, DT_LONG, 4, 0}, | |
729e4ab9 A |
156 | {LOW_Q, UDATPG_QUARTER_FIELD, DT_NUMERIC + DT_DELTA, 1, 2}, |
157 | {LOW_Q, UDATPG_QUARTER_FIELD, DT_SHORT + DT_DELTA, 3, 0}, | |
158 | {LOW_Q, UDATPG_QUARTER_FIELD, DT_LONG + DT_DELTA, 4, 0}, | |
46f4442e A |
159 | {CAP_M, UDATPG_MONTH_FIELD, DT_NUMERIC, 1, 2}, |
160 | {CAP_M, UDATPG_MONTH_FIELD, DT_SHORT, 3, 0}, | |
161 | {CAP_M, UDATPG_MONTH_FIELD, DT_LONG, 4, 0}, | |
162 | {CAP_M, UDATPG_MONTH_FIELD, DT_NARROW, 5, 0}, | |
163 | {CAP_L, UDATPG_MONTH_FIELD, DT_NUMERIC + DT_DELTA, 1, 2}, | |
164 | {CAP_L, UDATPG_MONTH_FIELD, DT_SHORT - DT_DELTA, 3, 0}, | |
165 | {CAP_L, UDATPG_MONTH_FIELD, DT_LONG - DT_DELTA, 4, 0}, | |
166 | {CAP_L, UDATPG_MONTH_FIELD, DT_NARROW - DT_DELTA, 5, 0}, | |
4388f060 | 167 | {LOW_L, UDATPG_MONTH_FIELD, DT_NUMERIC + DT_DELTA, 1, 1}, |
46f4442e A |
168 | {LOW_W, UDATPG_WEEK_OF_YEAR_FIELD, DT_NUMERIC, 1, 2}, |
169 | {CAP_W, UDATPG_WEEK_OF_MONTH_FIELD, DT_NUMERIC + DT_DELTA, 1, 0}, | |
46f4442e A |
170 | {CAP_E, UDATPG_WEEKDAY_FIELD, DT_SHORT, 1, 3}, |
171 | {CAP_E, UDATPG_WEEKDAY_FIELD, DT_LONG, 4, 0}, | |
172 | {CAP_E, UDATPG_WEEKDAY_FIELD, DT_NARROW, 5, 0}, | |
173 | {LOW_C, UDATPG_WEEKDAY_FIELD, DT_NUMERIC + 2*DT_DELTA, 1, 2}, | |
174 | {LOW_C, UDATPG_WEEKDAY_FIELD, DT_SHORT - 2*DT_DELTA, 3, 0}, | |
175 | {LOW_C, UDATPG_WEEKDAY_FIELD, DT_LONG - 2*DT_DELTA, 4, 0}, | |
176 | {LOW_C, UDATPG_WEEKDAY_FIELD, DT_NARROW - 2*DT_DELTA, 5, 0}, | |
729e4ab9 A |
177 | {LOW_E, UDATPG_WEEKDAY_FIELD, DT_NUMERIC + DT_DELTA, 1, 2}, // LOW_E is currently not used in CLDR data, should not be canonical |
178 | {LOW_E, UDATPG_WEEKDAY_FIELD, DT_SHORT - DT_DELTA, 3, 0}, | |
179 | {LOW_E, UDATPG_WEEKDAY_FIELD, DT_LONG - DT_DELTA, 4, 0}, | |
180 | {LOW_E, UDATPG_WEEKDAY_FIELD, DT_NARROW - DT_DELTA, 5, 0}, | |
46f4442e A |
181 | {LOW_D, UDATPG_DAY_FIELD, DT_NUMERIC, 1, 2}, |
182 | {CAP_D, UDATPG_DAY_OF_YEAR_FIELD, DT_NUMERIC + DT_DELTA, 1, 3}, | |
183 | {CAP_F, UDATPG_DAY_OF_WEEK_IN_MONTH_FIELD, DT_NUMERIC + 2*DT_DELTA, 1, 0}, | |
184 | {LOW_G, UDATPG_DAY_FIELD, DT_NUMERIC + 3*DT_DELTA, 1, 20}, // really internal use, so we don't care | |
185 | {LOW_A, UDATPG_DAYPERIOD_FIELD, DT_SHORT, 1, 0}, | |
186 | {CAP_H, UDATPG_HOUR_FIELD, DT_NUMERIC + 10*DT_DELTA, 1, 2}, // 24 hour | |
57a6839d | 187 | {LOW_K, UDATPG_HOUR_FIELD, DT_NUMERIC + 11*DT_DELTA, 1, 2}, // 24 hour |
46f4442e | 188 | {LOW_H, UDATPG_HOUR_FIELD, DT_NUMERIC, 1, 2}, // 12 hour |
57a6839d | 189 | {CAP_K, UDATPG_HOUR_FIELD, DT_NUMERIC + DT_DELTA, 1, 2}, // 12 hour |
46f4442e A |
190 | {LOW_M, UDATPG_MINUTE_FIELD, DT_NUMERIC, 1, 2}, |
191 | {LOW_S, UDATPG_SECOND_FIELD, DT_NUMERIC, 1, 2}, | |
192 | {CAP_S, UDATPG_FRACTIONAL_SECOND_FIELD, DT_NUMERIC + DT_DELTA, 1, 1000}, | |
193 | {CAP_A, UDATPG_SECOND_FIELD, DT_NUMERIC + 2*DT_DELTA, 1, 1000}, | |
194 | {LOW_V, UDATPG_ZONE_FIELD, DT_SHORT - 2*DT_DELTA, 1, 0}, | |
195 | {LOW_V, UDATPG_ZONE_FIELD, DT_LONG - 2*DT_DELTA, 4, 0}, | |
196 | {LOW_Z, UDATPG_ZONE_FIELD, DT_SHORT, 1, 3}, | |
197 | {LOW_Z, UDATPG_ZONE_FIELD, DT_LONG, 4, 0}, | |
57a6839d | 198 | {CAP_Z, UDATPG_ZONE_FIELD, DT_NARROW - DT_DELTA, 1, 3}, |
46f4442e | 199 | {CAP_Z, UDATPG_ZONE_FIELD, DT_LONG - DT_DELTA, 4, 0}, |
57a6839d A |
200 | {CAP_Z, UDATPG_ZONE_FIELD, DT_SHORT - DT_DELTA, 5, 0}, |
201 | {CAP_O, UDATPG_ZONE_FIELD, DT_SHORT - DT_DELTA, 1, 0}, | |
202 | {CAP_O, UDATPG_ZONE_FIELD, DT_LONG - DT_DELTA, 4, 0}, | |
203 | {CAP_V, UDATPG_ZONE_FIELD, DT_SHORT - DT_DELTA, 1, 0}, | |
204 | {CAP_V, UDATPG_ZONE_FIELD, DT_LONG - DT_DELTA, 2, 0}, | |
205 | {CAP_X, UDATPG_ZONE_FIELD, DT_NARROW - DT_DELTA, 1, 0}, | |
206 | {CAP_X, UDATPG_ZONE_FIELD, DT_SHORT - DT_DELTA, 2, 0}, | |
207 | {CAP_X, UDATPG_ZONE_FIELD, DT_LONG - DT_DELTA, 4, 0}, | |
208 | {LOW_X, UDATPG_ZONE_FIELD, DT_NARROW - DT_DELTA, 1, 0}, | |
209 | {LOW_X, UDATPG_ZONE_FIELD, DT_SHORT - DT_DELTA, 2, 0}, | |
210 | {LOW_X, UDATPG_ZONE_FIELD, DT_LONG - DT_DELTA, 4, 0}, | |
2ca993e8 A |
211 | {LOW_J, UDATPG_HOUR_FIELD, DT_NUMERIC, 1, 2}, // 12/24 hour |
212 | {CAP_J, UDATPG_HOUR_FIELD, DT_NUMERIC, 1, 2}, // 12/24 hour no AM/PM | |
729e4ab9 | 213 | {0, UDATPG_FIELD_COUNT, 0, 0, 0} , // last row of dtTypes[] |
46f4442e A |
214 | }; |
215 | ||
216 | static const char* const CLDR_FIELD_APPEND[] = { | |
217 | "Era", "Year", "Quarter", "Month", "Week", "*", "Day-Of-Week", "Day", "*", "*", "*", | |
218 | "Hour", "Minute", "Second", "*", "Timezone" | |
219 | }; | |
220 | ||
221 | static const char* const CLDR_FIELD_NAME[] = { | |
51004dcb | 222 | "era", "year", "quarter", "month", "week", "*", "weekday", "*", "*", "day", "dayperiod", |
46f4442e | 223 | "hour", "minute", "second", "*", "zone" |
729e4ab9 | 224 | }; |
46f4442e | 225 | |
46f4442e A |
226 | // For appendItems |
227 | static const UChar UDATPG_ItemFormat[]= {0x7B, 0x30, 0x7D, 0x20, 0x251C, 0x7B, 0x32, 0x7D, 0x3A, | |
228 | 0x20, 0x7B, 0x31, 0x7D, 0x2524, 0}; // {0} \u251C{2}: {1}\u2524 | |
229 | ||
51004dcb | 230 | //static const UChar repeatedPatterns[6]={CAP_G, CAP_E, LOW_Z, LOW_V, CAP_Q, 0}; // "GEzvQ" |
46f4442e A |
231 | |
232 | static const char DT_DateTimePatternsTag[]="DateTimePatterns"; | |
233 | static const char DT_DateTimeCalendarTag[]="calendar"; | |
234 | static const char DT_DateTimeGregorianTag[]="gregorian"; | |
235 | static const char DT_DateTimeAppendItemsTag[]="appendItems"; | |
236 | static const char DT_DateTimeFieldsTag[]="fields"; | |
237 | static const char DT_DateTimeAvailableFormatsTag[]="availableFormats"; | |
238 | //static const UnicodeString repeatedPattern=UnicodeString(repeatedPatterns); | |
239 | ||
240 | UOBJECT_DEFINE_RTTI_IMPLEMENTATION(DateTimePatternGenerator) | |
241 | UOBJECT_DEFINE_RTTI_IMPLEMENTATION(DTSkeletonEnumeration) | |
242 | UOBJECT_DEFINE_RTTI_IMPLEMENTATION(DTRedundantEnumeration) | |
243 | ||
244 | DateTimePatternGenerator* U_EXPORT2 | |
245 | DateTimePatternGenerator::createInstance(UErrorCode& status) { | |
246 | return createInstance(Locale::getDefault(), status); | |
247 | } | |
248 | ||
249 | DateTimePatternGenerator* U_EXPORT2 | |
250 | DateTimePatternGenerator::createInstance(const Locale& locale, UErrorCode& status) { | |
46f4442e | 251 | if (U_FAILURE(status)) { |
b331163b | 252 | return NULL; |
46f4442e | 253 | } |
2ca993e8 A |
254 | LocalPointer<DateTimePatternGenerator> result( |
255 | new DateTimePatternGenerator(locale, status), status); | |
256 | return U_SUCCESS(status) ? result.orphan() : NULL; | |
46f4442e A |
257 | } |
258 | ||
259 | DateTimePatternGenerator* U_EXPORT2 | |
260 | DateTimePatternGenerator::createEmptyInstance(UErrorCode& status) { | |
261 | DateTimePatternGenerator *result = new DateTimePatternGenerator(status); | |
262 | if (result == NULL) { | |
263 | status = U_MEMORY_ALLOCATION_ERROR; | |
264 | } | |
265 | if (U_FAILURE(status)) { | |
266 | delete result; | |
267 | result = NULL; | |
268 | } | |
269 | return result; | |
270 | } | |
271 | ||
729e4ab9 | 272 | DateTimePatternGenerator::DateTimePatternGenerator(UErrorCode &status) : |
46f4442e A |
273 | skipMatcher(NULL), |
274 | fAvailableFormatKeyHash(NULL) | |
275 | { | |
276 | fp = new FormatParser(); | |
277 | dtMatcher = new DateTimeMatcher(); | |
278 | distanceInfo = new DistanceInfo(); | |
279 | patternMap = new PatternMap(); | |
280 | if (fp == NULL || dtMatcher == NULL || distanceInfo == NULL || patternMap == NULL) { | |
281 | status = U_MEMORY_ALLOCATION_ERROR; | |
282 | } | |
283 | } | |
284 | ||
285 | DateTimePatternGenerator::DateTimePatternGenerator(const Locale& locale, UErrorCode &status) : | |
286 | skipMatcher(NULL), | |
287 | fAvailableFormatKeyHash(NULL) | |
288 | { | |
289 | fp = new FormatParser(); | |
290 | dtMatcher = new DateTimeMatcher(); | |
291 | distanceInfo = new DistanceInfo(); | |
292 | patternMap = new PatternMap(); | |
293 | if (fp == NULL || dtMatcher == NULL || distanceInfo == NULL || patternMap == NULL) { | |
294 | status = U_MEMORY_ALLOCATION_ERROR; | |
295 | } | |
296 | else { | |
297 | initData(locale, status); | |
298 | } | |
299 | } | |
300 | ||
729e4ab9 | 301 | DateTimePatternGenerator::DateTimePatternGenerator(const DateTimePatternGenerator& other) : |
46f4442e A |
302 | UObject(), |
303 | skipMatcher(NULL), | |
304 | fAvailableFormatKeyHash(NULL) | |
305 | { | |
306 | fp = new FormatParser(); | |
729e4ab9 | 307 | dtMatcher = new DateTimeMatcher(); |
46f4442e A |
308 | distanceInfo = new DistanceInfo(); |
309 | patternMap = new PatternMap(); | |
310 | *this=other; | |
311 | } | |
312 | ||
313 | DateTimePatternGenerator& | |
314 | DateTimePatternGenerator::operator=(const DateTimePatternGenerator& other) { | |
b331163b A |
315 | // reflexive case |
316 | if (&other == this) { | |
317 | return *this; | |
318 | } | |
46f4442e A |
319 | pLocale = other.pLocale; |
320 | fDefaultHourFormatChar = other.fDefaultHourFormatChar; | |
321 | *fp = *(other.fp); | |
322 | dtMatcher->copyFrom(other.dtMatcher->skeleton); | |
323 | *distanceInfo = *(other.distanceInfo); | |
324 | dateTimeFormat = other.dateTimeFormat; | |
325 | decimal = other.decimal; | |
326 | // NUL-terminate for the C API. | |
327 | dateTimeFormat.getTerminatedBuffer(); | |
328 | decimal.getTerminatedBuffer(); | |
329 | delete skipMatcher; | |
330 | if ( other.skipMatcher == NULL ) { | |
331 | skipMatcher = NULL; | |
332 | } | |
333 | else { | |
334 | skipMatcher = new DateTimeMatcher(*other.skipMatcher); | |
335 | } | |
336 | for (int32_t i=0; i< UDATPG_FIELD_COUNT; ++i ) { | |
337 | appendItemFormats[i] = other.appendItemFormats[i]; | |
338 | appendItemNames[i] = other.appendItemNames[i]; | |
339 | // NUL-terminate for the C API. | |
340 | appendItemFormats[i].getTerminatedBuffer(); | |
341 | appendItemNames[i].getTerminatedBuffer(); | |
342 | } | |
343 | UErrorCode status = U_ZERO_ERROR; | |
344 | patternMap->copyFrom(*other.patternMap, status); | |
345 | copyHashtable(other.fAvailableFormatKeyHash, status); | |
346 | return *this; | |
347 | } | |
348 | ||
349 | ||
350 | UBool | |
351 | DateTimePatternGenerator::operator==(const DateTimePatternGenerator& other) const { | |
352 | if (this == &other) { | |
353 | return TRUE; | |
354 | } | |
355 | if ((pLocale==other.pLocale) && (patternMap->equals(*other.patternMap)) && | |
356 | (dateTimeFormat==other.dateTimeFormat) && (decimal==other.decimal)) { | |
357 | for ( int32_t i=0 ; i<UDATPG_FIELD_COUNT; ++i ) { | |
358 | if ((appendItemFormats[i] != other.appendItemFormats[i]) || | |
359 | (appendItemNames[i] != other.appendItemNames[i]) ) { | |
360 | return FALSE; | |
361 | } | |
362 | } | |
363 | return TRUE; | |
364 | } | |
365 | else { | |
366 | return FALSE; | |
367 | } | |
368 | } | |
369 | ||
370 | UBool | |
371 | DateTimePatternGenerator::operator!=(const DateTimePatternGenerator& other) const { | |
372 | return !operator==(other); | |
373 | } | |
374 | ||
375 | DateTimePatternGenerator::~DateTimePatternGenerator() { | |
376 | if (fAvailableFormatKeyHash!=NULL) { | |
377 | delete fAvailableFormatKeyHash; | |
378 | } | |
729e4ab9 A |
379 | |
380 | if (fp != NULL) delete fp; | |
46f4442e A |
381 | if (dtMatcher != NULL) delete dtMatcher; |
382 | if (distanceInfo != NULL) delete distanceInfo; | |
383 | if (patternMap != NULL) delete patternMap; | |
384 | if (skipMatcher != NULL) delete skipMatcher; | |
385 | } | |
386 | ||
2ca993e8 A |
387 | namespace { |
388 | ||
389 | UInitOnce initOnce = U_INITONCE_INITIALIZER; | |
390 | UHashtable *localeToAllowedHourFormatsMap = NULL; | |
391 | ||
392 | // Value deleter for hashmap. | |
f3c0d7a5 | 393 | U_CFUNC void U_CALLCONV deleteAllowedHourFormats(void *ptr) { |
2ca993e8 A |
394 | uprv_free(ptr); |
395 | } | |
396 | ||
397 | // Close hashmap at cleanup. | |
f3c0d7a5 | 398 | U_CFUNC UBool U_CALLCONV allowedHourFormatsCleanup() { |
2ca993e8 A |
399 | uhash_close(localeToAllowedHourFormatsMap); |
400 | return TRUE; | |
401 | } | |
402 | ||
403 | enum AllowedHourFormat{ | |
404 | ALLOWED_HOUR_FORMAT_UNKNOWN = -1, | |
405 | ALLOWED_HOUR_FORMAT_h, | |
406 | ALLOWED_HOUR_FORMAT_H, | |
407 | ALLOWED_HOUR_FORMAT_hb, | |
408 | ALLOWED_HOUR_FORMAT_Hb, | |
409 | ALLOWED_HOUR_FORMAT_hB, | |
410 | ALLOWED_HOUR_FORMAT_HB | |
411 | }; | |
412 | ||
413 | } // namespace | |
414 | ||
46f4442e A |
415 | void |
416 | DateTimePatternGenerator::initData(const Locale& locale, UErrorCode &status) { | |
417 | //const char *baseLangName = locale.getBaseName(); // unused | |
729e4ab9 | 418 | |
46f4442e A |
419 | skipMatcher = NULL; |
420 | fAvailableFormatKeyHash=NULL; | |
f3c0d7a5 | 421 | addCanonicalItems(status); |
46f4442e | 422 | addICUPatterns(locale, status); |
729e4ab9 | 423 | addCLDRData(locale, status); |
46f4442e A |
424 | setDateTimeFromCalendar(locale, status); |
425 | setDecimalSymbols(locale, status); | |
2ca993e8 A |
426 | umtx_initOnce(initOnce, loadAllowedHourFormatsData, status); |
427 | getAllowedHourFormats(locale, status); | |
46f4442e A |
428 | } // DateTimePatternGenerator::initData |
429 | ||
2ca993e8 A |
430 | namespace { |
431 | ||
f3c0d7a5 | 432 | struct AllowedHourFormatsSink : public ResourceSink { |
2ca993e8 | 433 | // Initialize sub-sinks. |
f3c0d7a5 | 434 | AllowedHourFormatsSink() {} |
2ca993e8 A |
435 | virtual ~AllowedHourFormatsSink(); |
436 | ||
f3c0d7a5 A |
437 | virtual void put(const char *key, ResourceValue &value, UBool /*noFallback*/, |
438 | UErrorCode &errorCode) { | |
439 | ResourceTable timeData = value.getTable(errorCode); | |
440 | if (U_FAILURE(errorCode)) { return; } | |
441 | for (int32_t i = 0; timeData.getKeyAndValue(i, key, value); ++i) { | |
442 | const char *regionOrLocale = key; | |
443 | ResourceTable formatList = value.getTable(errorCode); | |
444 | if (U_FAILURE(errorCode)) { return; } | |
445 | for (int32_t j = 0; formatList.getKeyAndValue(j, key, value); ++j) { | |
446 | if (uprv_strcmp(key, "allowed") == 0) { // Ignore "preferred" list. | |
447 | LocalMemory<int32_t> list; | |
448 | int32_t length; | |
449 | if (value.getType() == URES_STRING) { | |
450 | if (list.allocateInsteadAndReset(2) == NULL) { | |
451 | errorCode = U_MEMORY_ALLOCATION_ERROR; | |
452 | return; | |
453 | } | |
454 | list[0] = getHourFormatFromUnicodeString(value.getUnicodeString(errorCode)); | |
455 | length = 1; | |
456 | } | |
457 | else { | |
458 | ResourceArray allowedFormats = value.getArray(errorCode); | |
459 | length = allowedFormats.getSize(); | |
460 | if (list.allocateInsteadAndReset(length + 1) == NULL) { | |
461 | errorCode = U_MEMORY_ALLOCATION_ERROR; | |
462 | return; | |
463 | } | |
464 | for (int32_t k = 0; k < length; ++k) { | |
465 | allowedFormats.getValue(k, value); | |
466 | list[k] = getHourFormatFromUnicodeString(value.getUnicodeString(errorCode)); | |
467 | } | |
468 | } | |
469 | list[length] = ALLOWED_HOUR_FORMAT_UNKNOWN; | |
470 | uhash_put(localeToAllowedHourFormatsMap, | |
471 | const_cast<char *>(regionOrLocale), list.orphan(), &errorCode); | |
472 | if (U_FAILURE(errorCode)) { return; } | |
2ca993e8 A |
473 | } |
474 | } | |
2ca993e8 | 475 | } |
f3c0d7a5 | 476 | } |
2ca993e8 | 477 | |
f3c0d7a5 | 478 | AllowedHourFormat getHourFormatFromUnicodeString(const UnicodeString &s) { |
2ca993e8 A |
479 | if (s.length() == 1) { |
480 | if (s[0] == LOW_H) { return ALLOWED_HOUR_FORMAT_h; } | |
481 | if (s[0] == CAP_H) { return ALLOWED_HOUR_FORMAT_H; } | |
482 | } else if (s.length() == 2) { | |
483 | if (s[0] == LOW_H && s[1] == LOW_B) { return ALLOWED_HOUR_FORMAT_hb; } | |
484 | if (s[0] == CAP_H && s[1] == LOW_B) { return ALLOWED_HOUR_FORMAT_Hb; } | |
485 | if (s[0] == LOW_H && s[1] == CAP_B) { return ALLOWED_HOUR_FORMAT_hB; } | |
486 | if (s[0] == CAP_H && s[1] == CAP_B) { return ALLOWED_HOUR_FORMAT_HB; } | |
487 | } | |
488 | ||
489 | return ALLOWED_HOUR_FORMAT_UNKNOWN; | |
490 | } | |
491 | }; | |
492 | ||
493 | } // namespace | |
494 | ||
495 | AllowedHourFormatsSink::~AllowedHourFormatsSink() {} | |
2ca993e8 | 496 | |
f3c0d7a5 | 497 | U_CFUNC void U_CALLCONV DateTimePatternGenerator::loadAllowedHourFormatsData(UErrorCode &status) { |
2ca993e8 A |
498 | if (U_FAILURE(status)) { return; } |
499 | localeToAllowedHourFormatsMap = uhash_open( | |
500 | uhash_hashChars, uhash_compareChars, NULL, &status); | |
501 | uhash_setValueDeleter(localeToAllowedHourFormatsMap, deleteAllowedHourFormats); | |
502 | LocalUResourceBundlePointer rb(ures_openDirect(NULL, "supplementalData", &status)); | |
503 | ||
504 | AllowedHourFormatsSink sink; | |
505 | // TODO: Currently in the enumeration each table allocates a new array. | |
506 | // Try to reduce the number of memory allocations. Consider storing a | |
507 | // UVector32 with the concatenation of all of the sub-arrays, put the start index | |
508 | // into the hashmap, store 6 single-value sub-arrays right at the beginning of the | |
509 | // vector (at index enum*2) for easy data sharing, copy sub-arrays into runtime | |
510 | // object. Remember to clean up the vector, too. | |
f3c0d7a5 | 511 | ures_getAllItemsWithFallback(rb.getAlias(), "timeData", sink, status); |
2ca993e8 A |
512 | |
513 | ucln_i18n_registerCleanup(UCLN_I18N_ALLOWED_HOUR_FORMATS, allowedHourFormatsCleanup); | |
514 | } | |
515 | ||
516 | void DateTimePatternGenerator::getAllowedHourFormats(const Locale &locale, UErrorCode &status) { | |
517 | if (U_FAILURE(status)) { return; } | |
2ca993e8 A |
518 | const char *localeID = locale.getName(); |
519 | char maxLocaleID[ULOC_FULLNAME_CAPACITY]; | |
520 | int32_t length = uloc_addLikelySubtags(localeID, maxLocaleID, ULOC_FULLNAME_CAPACITY, &status); | |
521 | if (U_FAILURE(status)) { | |
522 | return; | |
523 | } else if (length == ULOC_FULLNAME_CAPACITY) { // no room for NUL | |
524 | status = U_BUFFER_OVERFLOW_ERROR; | |
525 | return; | |
526 | } | |
527 | Locale maxLocale = Locale(maxLocaleID); | |
528 | ||
529 | const char *country = maxLocale.getCountry(); | |
530 | if (*country == '\0') { country = "001"; } | |
531 | const char *language = maxLocale.getLanguage(); | |
532 | ||
533 | CharString langCountry; | |
534 | langCountry.append(language, uprv_strlen(language), status); | |
535 | langCountry.append('_', status); | |
536 | langCountry.append(country, uprv_strlen(country), status); | |
537 | ||
538 | int32_t *allowedFormats; | |
539 | allowedFormats = (int32_t *)uhash_get(localeToAllowedHourFormatsMap, langCountry.data()); | |
540 | if (allowedFormats == NULL) { | |
541 | allowedFormats = (int32_t *)uhash_get(localeToAllowedHourFormatsMap, const_cast<char *>(country)); | |
542 | } | |
543 | ||
544 | if (allowedFormats != NULL) { // Lookup is successful | |
545 | for (int32_t i = 0; i < UPRV_LENGTHOF(fAllowedHourFormats); ++i) { | |
546 | fAllowedHourFormats[i] = allowedFormats[i]; | |
547 | if (allowedFormats[i] == ALLOWED_HOUR_FORMAT_UNKNOWN) { | |
548 | break; | |
549 | } | |
550 | } | |
551 | } else { // Lookup failed, twice | |
552 | fAllowedHourFormats[0] = ALLOWED_HOUR_FORMAT_H; | |
553 | fAllowedHourFormats[1] = ALLOWED_HOUR_FORMAT_UNKNOWN; | |
554 | } | |
555 | } | |
556 | ||
46f4442e A |
557 | UnicodeString |
558 | DateTimePatternGenerator::getSkeleton(const UnicodeString& pattern, UErrorCode& | |
559 | /*status*/) { | |
2ca993e8 A |
560 | FormatParser fp; |
561 | DateTimeMatcher matcher; | |
562 | PtnSkeleton localSkeleton; | |
563 | matcher.set(pattern, &fp, localSkeleton); | |
564 | return localSkeleton.getSkeleton(); | |
565 | } | |
566 | ||
567 | UnicodeString | |
568 | DateTimePatternGenerator::staticGetSkeleton( | |
569 | const UnicodeString& pattern, UErrorCode& /*status*/) { | |
570 | FormatParser fp; | |
571 | DateTimeMatcher matcher; | |
572 | PtnSkeleton localSkeleton; | |
573 | matcher.set(pattern, &fp, localSkeleton); | |
574 | return localSkeleton.getSkeleton(); | |
46f4442e A |
575 | } |
576 | ||
577 | UnicodeString | |
578 | DateTimePatternGenerator::getBaseSkeleton(const UnicodeString& pattern, UErrorCode& /*status*/) { | |
2ca993e8 A |
579 | FormatParser fp; |
580 | DateTimeMatcher matcher; | |
581 | PtnSkeleton localSkeleton; | |
582 | matcher.set(pattern, &fp, localSkeleton); | |
583 | return localSkeleton.getBaseSkeleton(); | |
584 | } | |
585 | ||
586 | UnicodeString | |
587 | DateTimePatternGenerator::staticGetBaseSkeleton( | |
588 | const UnicodeString& pattern, UErrorCode& /*status*/) { | |
589 | FormatParser fp; | |
590 | DateTimeMatcher matcher; | |
591 | PtnSkeleton localSkeleton; | |
592 | matcher.set(pattern, &fp, localSkeleton); | |
593 | return localSkeleton.getBaseSkeleton(); | |
46f4442e A |
594 | } |
595 | ||
596 | void | |
597 | DateTimePatternGenerator::addICUPatterns(const Locale& locale, UErrorCode& status) { | |
f3c0d7a5 | 598 | if (U_FAILURE(status)) { return; } |
46f4442e A |
599 | UnicodeString dfPattern; |
600 | UnicodeString conflictingString; | |
46f4442e A |
601 | DateFormat* df; |
602 | ||
46f4442e A |
603 | // Load with ICU patterns |
604 | for (int32_t i=DateFormat::kFull; i<=DateFormat::kShort; i++) { | |
605 | DateFormat::EStyle style = (DateFormat::EStyle)i; | |
606 | df = DateFormat::createDateInstance(style, locale); | |
729e4ab9 A |
607 | SimpleDateFormat* sdf; |
608 | if (df != NULL && (sdf = dynamic_cast<SimpleDateFormat*>(df)) != NULL) { | |
f3c0d7a5 A |
609 | sdf->toPattern(dfPattern); |
610 | addPattern(dfPattern, FALSE, conflictingString, status); | |
46f4442e A |
611 | } |
612 | // TODO Maybe we should return an error when the date format isn't simple. | |
613 | delete df; | |
f3c0d7a5 | 614 | if (U_FAILURE(status)) { return; } |
46f4442e A |
615 | |
616 | df = DateFormat::createTimeInstance(style, locale); | |
729e4ab9 | 617 | if (df != NULL && (sdf = dynamic_cast<SimpleDateFormat*>(df)) != NULL) { |
f3c0d7a5 A |
618 | sdf->toPattern(dfPattern); |
619 | addPattern(dfPattern, FALSE, conflictingString, status); | |
620 | ||
621 | // TODO: C++ and Java are inconsistent (see #12568). | |
622 | // C++ uses MEDIUM, but Java uses SHORT. | |
623 | if ( i==DateFormat::kShort && !dfPattern.isEmpty() ) { | |
624 | consumeShortTimePattern(dfPattern, status); | |
46f4442e A |
625 | } |
626 | } | |
627 | // TODO Maybe we should return an error when the date format isn't simple. | |
628 | delete df; | |
f3c0d7a5 | 629 | if (U_FAILURE(status)) { return; } |
46f4442e A |
630 | } |
631 | } | |
632 | ||
633 | void | |
634 | DateTimePatternGenerator::hackTimes(const UnicodeString& hackPattern, UErrorCode& status) { | |
46f4442e | 635 | UnicodeString conflictingString; |
729e4ab9 | 636 | |
46f4442e A |
637 | fp->set(hackPattern); |
638 | UnicodeString mmss; | |
639 | UBool gotMm=FALSE; | |
640 | for (int32_t i=0; i<fp->itemNumber; ++i) { | |
641 | UnicodeString field = fp->items[i]; | |
642 | if ( fp->isQuoteLiteral(field) ) { | |
643 | if ( gotMm ) { | |
644 | UnicodeString quoteLiteral; | |
645 | fp->getQuoteLiteral(quoteLiteral, &i); | |
646 | mmss += quoteLiteral; | |
647 | } | |
648 | } | |
649 | else { | |
650 | if (fp->isPatternSeparator(field) && gotMm) { | |
651 | mmss+=field; | |
652 | } | |
653 | else { | |
654 | UChar ch=field.charAt(0); | |
655 | if (ch==LOW_M) { | |
656 | gotMm=TRUE; | |
657 | mmss+=field; | |
658 | } | |
659 | else { | |
660 | if (ch==LOW_S) { | |
661 | if (!gotMm) { | |
662 | break; | |
663 | } | |
664 | mmss+= field; | |
4388f060 | 665 | addPattern(mmss, FALSE, conflictingString, status); |
46f4442e A |
666 | break; |
667 | } | |
668 | else { | |
669 | if (gotMm || ch==LOW_Z || ch==CAP_Z || ch==LOW_V || ch==CAP_V) { | |
670 | break; | |
671 | } | |
672 | } | |
673 | } | |
674 | } | |
675 | } | |
676 | } | |
677 | } | |
678 | ||
679 | #define ULOC_LOCALE_IDENTIFIER_CAPACITY (ULOC_FULLNAME_CAPACITY + 1 + ULOC_KEYWORD_AND_VALUES_CAPACITY) | |
680 | ||
681 | static const UChar hourFormatChars[] = { CAP_H, LOW_H, CAP_K, LOW_K, 0 }; // HhKk, the hour format characters | |
682 | ||
683 | void | |
f3c0d7a5 A |
684 | DateTimePatternGenerator::getCalendarTypeToUse(const Locale& locale, CharString& destination, UErrorCode& err) { |
685 | destination.clear().append(DT_DateTimeGregorianTag, -1, err); // initial default | |
46f4442e | 686 | if ( U_SUCCESS(err) ) { |
f3c0d7a5 | 687 | char localeWithCalendarKey[ULOC_LOCALE_IDENTIFIER_CAPACITY]; |
46f4442e | 688 | // obtain a locale that always has the calendar key value that should be used |
f3c0d7a5 A |
689 | ures_getFunctionalEquivalent( |
690 | localeWithCalendarKey, | |
691 | ULOC_LOCALE_IDENTIFIER_CAPACITY, | |
692 | NULL, | |
693 | "calendar", | |
694 | "calendar", | |
695 | locale.getName(), | |
696 | NULL, | |
697 | FALSE, | |
698 | &err); | |
46f4442e A |
699 | localeWithCalendarKey[ULOC_LOCALE_IDENTIFIER_CAPACITY-1] = 0; // ensure null termination |
700 | // now get the calendar key value from that locale | |
f3c0d7a5 A |
701 | char calendarType[ULOC_KEYWORDS_CAPACITY]; |
702 | int32_t calendarTypeLen = uloc_getKeywordValue( | |
703 | localeWithCalendarKey, | |
704 | "calendar", | |
705 | calendarType, | |
706 | ULOC_KEYWORDS_CAPACITY, | |
707 | &err); | |
46f4442e | 708 | if (U_SUCCESS(err) && calendarTypeLen < ULOC_KEYWORDS_CAPACITY) { |
f3c0d7a5 A |
709 | destination.clear().append(calendarType, -1, err); |
710 | if (U_FAILURE(err)) { return; } | |
46f4442e A |
711 | } |
712 | err = U_ZERO_ERROR; | |
713 | } | |
f3c0d7a5 | 714 | } |
46f4442e | 715 | |
f3c0d7a5 A |
716 | void |
717 | DateTimePatternGenerator::consumeShortTimePattern(const UnicodeString& shortTimePattern, | |
718 | UErrorCode& status) { | |
719 | ||
720 | // set fDefaultHourFormatChar to the hour format character from this pattern | |
721 | int32_t tfIdx, tfLen = shortTimePattern.length(); | |
722 | UBool ignoreChars = FALSE; | |
723 | for (tfIdx = 0; tfIdx < tfLen; tfIdx++) { | |
724 | UChar tfChar = shortTimePattern.charAt(tfIdx); | |
725 | if ( tfChar == SINGLE_QUOTE ) { | |
726 | ignoreChars = !ignoreChars; // toggle (handle quoted literals & '' for single quote) | |
727 | } else if ( !ignoreChars && u_strchr(hourFormatChars, tfChar) != NULL ) { | |
728 | fDefaultHourFormatChar = tfChar; | |
729 | break; | |
46f4442e | 730 | } |
729e4ab9 | 731 | } |
729e4ab9 | 732 | |
f3c0d7a5 A |
733 | // HACK for hh:ss |
734 | hackTimes(shortTimePattern, status); | |
735 | } | |
736 | ||
737 | struct DateTimePatternGenerator::AppendItemFormatsSink : public ResourceSink { | |
738 | ||
739 | // Destination for data, modified via setters. | |
740 | DateTimePatternGenerator& dtpg; | |
741 | ||
742 | AppendItemFormatsSink(DateTimePatternGenerator& _dtpg) : dtpg(_dtpg) {} | |
743 | virtual ~AppendItemFormatsSink(); | |
744 | ||
745 | virtual void put(const char *key, ResourceValue &value, UBool /*noFallback*/, | |
746 | UErrorCode &errorCode) { | |
747 | ResourceTable itemsTable = value.getTable(errorCode); | |
748 | if (U_FAILURE(errorCode)) { return; } | |
749 | for (int32_t i = 0; itemsTable.getKeyAndValue(i, key, value); ++i) { | |
750 | UDateTimePatternField field = dtpg.getAppendFormatNumber(key); | |
751 | if (field == UDATPG_FIELD_COUNT) { continue; } | |
752 | const UnicodeString& valueStr = value.getUnicodeString(errorCode); | |
753 | if (dtpg.getAppendItemFormat(field).isEmpty() && !valueStr.isEmpty()) { | |
754 | dtpg.setAppendItemFormat(field, valueStr); | |
755 | } | |
46f4442e A |
756 | } |
757 | } | |
729e4ab9 | 758 | |
f3c0d7a5 A |
759 | void fillInMissing() { |
760 | UnicodeString defaultItemFormat(TRUE, UDATPG_ItemFormat, UPRV_LENGTHOF(UDATPG_ItemFormat)-1); // Read-only alias. | |
761 | for (int32_t i = 0; i < UDATPG_FIELD_COUNT; i++) { | |
762 | UDateTimePatternField field = (UDateTimePatternField)i; | |
763 | if (dtpg.getAppendItemFormat(field).isEmpty()) { | |
764 | dtpg.setAppendItemFormat(field, defaultItemFormat); | |
765 | } | |
46f4442e A |
766 | } |
767 | } | |
f3c0d7a5 | 768 | }; |
46f4442e | 769 | |
f3c0d7a5 A |
770 | struct DateTimePatternGenerator::AppendItemNamesSink : public ResourceSink { |
771 | ||
772 | // Destination for data, modified via setters. | |
773 | DateTimePatternGenerator& dtpg; | |
774 | ||
775 | AppendItemNamesSink(DateTimePatternGenerator& _dtpg) : dtpg(_dtpg) {} | |
776 | virtual ~AppendItemNamesSink(); | |
777 | ||
778 | virtual void put(const char *key, ResourceValue &value, UBool /*noFallback*/, | |
779 | UErrorCode &errorCode) { | |
780 | ResourceTable itemsTable = value.getTable(errorCode); | |
781 | if (U_FAILURE(errorCode)) { return; } | |
782 | for (int32_t i = 0; itemsTable.getKeyAndValue(i, key, value); ++i) { | |
783 | UDateTimePatternField field = dtpg.getAppendNameNumber(key); | |
784 | if (field == UDATPG_FIELD_COUNT) { continue; } | |
785 | ResourceTable detailsTable = value.getTable(errorCode); | |
786 | if (U_FAILURE(errorCode)) { return; } | |
787 | for (int32_t j = 0; detailsTable.getKeyAndValue(j, key, value); ++j) { | |
788 | if (uprv_strcmp(key, "dn") != 0) { continue; } | |
789 | const UnicodeString& valueStr = value.getUnicodeString(errorCode); | |
790 | if (dtpg.getAppendItemName(field).isEmpty() && !valueStr.isEmpty()) { | |
791 | dtpg.setAppendItemName(field, valueStr); | |
46f4442e | 792 | } |
f3c0d7a5 | 793 | break; |
4388f060 | 794 | } |
46f4442e | 795 | } |
f3c0d7a5 A |
796 | } |
797 | ||
798 | void fillInMissing() { | |
799 | for (int32_t i = 0; i < UDATPG_FIELD_COUNT; i++) { | |
800 | UDateTimePatternField field = (UDateTimePatternField)i; | |
801 | UnicodeString& valueStr = dtpg.getMutableAppendItemName(field); | |
802 | if (valueStr.isEmpty()) { | |
803 | valueStr = CAP_F; | |
804 | U_ASSERT(i < 20); | |
805 | if (i < 10) { | |
806 | // F0, F1, ..., F9 | |
807 | valueStr += (UChar)(i+0x30); | |
808 | } else { | |
809 | // F10, F11, ... | |
810 | valueStr += (UChar)0x31; | |
811 | valueStr += (UChar)(i-10 + 0x30); | |
812 | } | |
813 | // NUL-terminate for the C API. | |
814 | valueStr.getTerminatedBuffer(); | |
4388f060 A |
815 | } |
816 | } | |
f3c0d7a5 A |
817 | } |
818 | }; | |
819 | ||
820 | struct DateTimePatternGenerator::AvailableFormatsSink : public ResourceSink { | |
821 | ||
822 | // Destination for data, modified via setters. | |
823 | DateTimePatternGenerator& dtpg; | |
824 | ||
825 | // Temporary variable, required for calling addPatternWithSkeleton. | |
826 | UnicodeString conflictingPattern; | |
827 | ||
828 | AvailableFormatsSink(DateTimePatternGenerator& _dtpg) : dtpg(_dtpg) {} | |
829 | virtual ~AvailableFormatsSink(); | |
830 | ||
831 | virtual void put(const char *key, ResourceValue &value, UBool isRoot, | |
832 | UErrorCode &errorCode) { | |
833 | ResourceTable itemsTable = value.getTable(errorCode); | |
834 | if (U_FAILURE(errorCode)) { return; } | |
835 | for (int32_t i = 0; itemsTable.getKeyAndValue(i, key, value); ++i) { | |
836 | const UnicodeString formatKey(key, -1, US_INV); | |
837 | if (!dtpg.isAvailableFormatSet(formatKey) ) { | |
838 | dtpg.setAvailableFormat(formatKey, errorCode); | |
839 | // Add pattern with its associated skeleton. Override any duplicate | |
840 | // derived from std patterns, but not a previous availableFormats entry: | |
841 | const UnicodeString& formatValue = value.getUnicodeString(errorCode); | |
842 | conflictingPattern.remove(); | |
843 | dtpg.addPatternWithSkeleton(formatValue, &formatKey, !isRoot, conflictingPattern, errorCode); | |
4388f060 A |
844 | } |
845 | } | |
2ca993e8 | 846 | } |
f3c0d7a5 | 847 | }; |
46f4442e | 848 | |
f3c0d7a5 A |
849 | // Virtual destructors must be defined out of line. |
850 | DateTimePatternGenerator::AppendItemFormatsSink::~AppendItemFormatsSink() {} | |
851 | DateTimePatternGenerator::AppendItemNamesSink::~AppendItemNamesSink() {} | |
852 | DateTimePatternGenerator::AvailableFormatsSink::~AvailableFormatsSink() {} | |
853 | ||
854 | void | |
855 | DateTimePatternGenerator::addCLDRData(const Locale& locale, UErrorCode& errorCode) { | |
856 | if (U_FAILURE(errorCode)) { return; } | |
857 | UnicodeString rbPattern, value, field; | |
858 | CharString path; | |
859 | ||
860 | LocalUResourceBundlePointer rb(ures_open(NULL, locale.getName(), &errorCode)); | |
861 | if (U_FAILURE(errorCode)) { return; } | |
862 | ||
863 | CharString calendarTypeToUse; // to be filled in with the type to use, if all goes well | |
864 | getCalendarTypeToUse(locale, calendarTypeToUse, errorCode); | |
865 | if (U_FAILURE(errorCode)) { return; } | |
866 | ||
867 | // Local err to ignore resource not found exceptions | |
868 | UErrorCode err = U_ZERO_ERROR; | |
869 | ||
870 | // Load append item formats. | |
871 | AppendItemFormatsSink appendItemFormatsSink(*this); | |
872 | path.clear() | |
873 | .append(DT_DateTimeCalendarTag, errorCode) | |
874 | .append('/', errorCode) | |
875 | .append(calendarTypeToUse, errorCode) | |
876 | .append('/', errorCode) | |
877 | .append(DT_DateTimeAppendItemsTag, errorCode); // i.e., calendar/xxx/appendItems | |
878 | if (U_FAILURE(errorCode)) { return; } | |
879 | ures_getAllItemsWithFallback(rb.getAlias(), path.data(), appendItemFormatsSink, err); | |
880 | appendItemFormatsSink.fillInMissing(); | |
881 | ||
882 | // Load CLDR item names. | |
883 | err = U_ZERO_ERROR; | |
884 | AppendItemNamesSink appendItemNamesSink(*this); | |
885 | ures_getAllItemsWithFallback(rb.getAlias(), DT_DateTimeFieldsTag, appendItemNamesSink, err); | |
886 | appendItemNamesSink.fillInMissing(); | |
887 | ||
888 | // Load the available formats from CLDR. | |
889 | err = U_ZERO_ERROR; | |
890 | initHashtable(errorCode); | |
891 | if (U_FAILURE(errorCode)) { return; } | |
892 | AvailableFormatsSink availableFormatsSink(*this); | |
893 | path.clear() | |
894 | .append(DT_DateTimeCalendarTag, errorCode) | |
895 | .append('/', errorCode) | |
896 | .append(calendarTypeToUse, errorCode) | |
897 | .append('/', errorCode) | |
898 | .append(DT_DateTimeAvailableFormatsTag, errorCode); // i.e., calendar/xxx/availableFormats | |
899 | if (U_FAILURE(errorCode)) { return; } | |
900 | ures_getAllItemsWithFallback(rb.getAlias(), path.data(), availableFormatsSink, err); | |
46f4442e A |
901 | } |
902 | ||
903 | void | |
904 | DateTimePatternGenerator::initHashtable(UErrorCode& err) { | |
905 | if (fAvailableFormatKeyHash!=NULL) { | |
906 | return; | |
907 | } | |
908 | if ((fAvailableFormatKeyHash = new Hashtable(FALSE, err))==NULL) { | |
909 | err=U_MEMORY_ALLOCATION_ERROR; | |
910 | return; | |
911 | } | |
912 | } | |
913 | ||
46f4442e A |
914 | void |
915 | DateTimePatternGenerator::setAppendItemFormat(UDateTimePatternField field, const UnicodeString& value) { | |
916 | appendItemFormats[field] = value; | |
917 | // NUL-terminate for the C API. | |
918 | appendItemFormats[field].getTerminatedBuffer(); | |
919 | } | |
920 | ||
921 | const UnicodeString& | |
922 | DateTimePatternGenerator::getAppendItemFormat(UDateTimePatternField field) const { | |
923 | return appendItemFormats[field]; | |
924 | } | |
925 | ||
926 | void | |
927 | DateTimePatternGenerator::setAppendItemName(UDateTimePatternField field, const UnicodeString& value) { | |
928 | appendItemNames[field] = value; | |
929 | // NUL-terminate for the C API. | |
930 | appendItemNames[field].getTerminatedBuffer(); | |
931 | } | |
932 | ||
933 | const UnicodeString& | |
f3c0d7a5 A |
934 | DateTimePatternGenerator::getAppendItemName(UDateTimePatternField field) const { |
935 | return appendItemNames[field]; | |
936 | } | |
937 | ||
938 | UnicodeString& | |
939 | DateTimePatternGenerator::getMutableAppendItemName(UDateTimePatternField field) { | |
46f4442e A |
940 | return appendItemNames[field]; |
941 | } | |
942 | ||
943 | void | |
944 | DateTimePatternGenerator::getAppendName(UDateTimePatternField field, UnicodeString& value) { | |
945 | value = SINGLE_QUOTE; | |
946 | value += appendItemNames[field]; | |
947 | value += SINGLE_QUOTE; | |
948 | } | |
949 | ||
950 | UnicodeString | |
951 | DateTimePatternGenerator::getBestPattern(const UnicodeString& patternForm, UErrorCode& status) { | |
729e4ab9 A |
952 | return getBestPattern(patternForm, UDATPG_MATCH_NO_OPTIONS, status); |
953 | } | |
954 | ||
955 | UnicodeString | |
956 | DateTimePatternGenerator::getBestPattern(const UnicodeString& patternForm, UDateTimePatternMatchOptions options, UErrorCode& status) { | |
46f4442e A |
957 | const UnicodeString *bestPattern=NULL; |
958 | UnicodeString dtFormat; | |
959 | UnicodeString resultPattern; | |
57a6839d | 960 | int32_t flags = kDTPGNoFlags; |
46f4442e A |
961 | |
962 | int32_t dateMask=(1<<UDATPG_DAYPERIOD_FIELD) - 1; | |
963 | int32_t timeMask=(1<<UDATPG_FIELD_COUNT) - 1 - dateMask; | |
729e4ab9 | 964 | |
2ca993e8 | 965 | // Replace hour metacharacters 'j', 'C' and 'J', set flags as necessary |
46f4442e | 966 | UnicodeString patternFormCopy = UnicodeString(patternForm); |
57a6839d A |
967 | UChar hourFormatSkeletonCharForLowJ = fDefaultHourFormatChar; |
968 | switch (options & UADATPG_FORCE_HOUR_CYCLE_MASK) { | |
969 | case UADATPG_FORCE_12_HOUR_CYCLE: hourFormatSkeletonCharForLowJ = LOW_H; break; | |
970 | case UADATPG_FORCE_24_HOUR_CYCLE: hourFormatSkeletonCharForLowJ = CAP_H; break; | |
971 | default: break; | |
972 | } | |
973 | int32_t patPos, patLen = patternFormCopy.length(); | |
974 | UBool inQuoted = FALSE; | |
975 | for (patPos = 0; patPos < patLen; patPos++) { | |
976 | UChar patChr = patternFormCopy.charAt(patPos); | |
977 | if (patChr == SINGLE_QUOTE) { | |
978 | inQuoted = !inQuoted; | |
979 | } else if (!inQuoted) { | |
980 | if (patChr == LOW_J) { | |
981 | patternFormCopy.setCharAt(patPos, hourFormatSkeletonCharForLowJ); | |
2ca993e8 A |
982 | } else if (patChr == CAP_C) { |
983 | AllowedHourFormat preferred; | |
984 | if (fAllowedHourFormats[0] != ALLOWED_HOUR_FORMAT_UNKNOWN) { | |
985 | preferred = (AllowedHourFormat)fAllowedHourFormats[0]; | |
986 | } else { | |
987 | status = U_INVALID_FORMAT_ERROR; | |
988 | return UnicodeString(); | |
989 | } | |
990 | ||
991 | if (preferred == ALLOWED_HOUR_FORMAT_H || preferred == ALLOWED_HOUR_FORMAT_HB || preferred == ALLOWED_HOUR_FORMAT_Hb) { | |
992 | patternFormCopy.setCharAt(patPos, CAP_H); | |
993 | } else { | |
994 | patternFormCopy.setCharAt(patPos, LOW_H); | |
995 | } | |
996 | ||
997 | if (preferred == ALLOWED_HOUR_FORMAT_HB || preferred == ALLOWED_HOUR_FORMAT_hB) { | |
998 | flags |= kDTPGSkeletonUsesCapB; | |
999 | } else if (preferred == ALLOWED_HOUR_FORMAT_Hb || preferred == ALLOWED_HOUR_FORMAT_hb) { | |
1000 | flags |= kDTPGSkeletonUsesLowB; | |
1001 | } | |
57a6839d A |
1002 | } else if (patChr == CAP_J) { |
1003 | // Get pattern for skeleton with H, then in adjustFieldTypes | |
1004 | // replace hour pattern characters as necessary. | |
1005 | patternFormCopy.setCharAt(patPos, CAP_H); | |
1006 | flags |= kDTPGSkeletonUsesCapJ; | |
1007 | } | |
1008 | } | |
1009 | } | |
729e4ab9 | 1010 | |
46f4442e A |
1011 | resultPattern.remove(); |
1012 | dtMatcher->set(patternFormCopy, fp); | |
1013 | const PtnSkeleton* specifiedSkeleton=NULL; | |
1014 | bestPattern=getBestRaw(*dtMatcher, -1, distanceInfo, &specifiedSkeleton); | |
1015 | if ( distanceInfo->missingFieldMask==0 && distanceInfo->extraFieldMask==0 ) { | |
57a6839d | 1016 | resultPattern = adjustFieldTypes(*bestPattern, specifiedSkeleton, flags, options); |
46f4442e A |
1017 | |
1018 | return resultPattern; | |
1019 | } | |
1020 | int32_t neededFields = dtMatcher->getFieldMask(); | |
57a6839d A |
1021 | UnicodeString datePattern=getBestAppending(neededFields & dateMask, flags, options); |
1022 | UnicodeString timePattern=getBestAppending(neededFields & timeMask, flags, options); | |
46f4442e A |
1023 | if (datePattern.length()==0) { |
1024 | if (timePattern.length()==0) { | |
1025 | resultPattern.remove(); | |
1026 | } | |
1027 | else { | |
1028 | return timePattern; | |
1029 | } | |
1030 | } | |
1031 | if (timePattern.length()==0) { | |
1032 | return datePattern; | |
1033 | } | |
1034 | resultPattern.remove(); | |
1035 | status = U_ZERO_ERROR; | |
1036 | dtFormat=getDateTimeFormat(); | |
2ca993e8 | 1037 | SimpleFormatter(dtFormat, 2, 2, status).format(timePattern, datePattern, resultPattern, status); |
46f4442e A |
1038 | return resultPattern; |
1039 | } | |
1040 | ||
1041 | UnicodeString | |
729e4ab9 A |
1042 | DateTimePatternGenerator::replaceFieldTypes(const UnicodeString& pattern, |
1043 | const UnicodeString& skeleton, | |
1044 | UErrorCode& status) { | |
1045 | return replaceFieldTypes(pattern, skeleton, UDATPG_MATCH_NO_OPTIONS, status); | |
1046 | } | |
1047 | ||
1048 | UnicodeString | |
1049 | DateTimePatternGenerator::replaceFieldTypes(const UnicodeString& pattern, | |
1050 | const UnicodeString& skeleton, | |
1051 | UDateTimePatternMatchOptions options, | |
46f4442e A |
1052 | UErrorCode& /*status*/) { |
1053 | dtMatcher->set(skeleton, fp); | |
57a6839d | 1054 | UnicodeString result = adjustFieldTypes(pattern, NULL, kDTPGNoFlags, options); |
46f4442e A |
1055 | return result; |
1056 | } | |
1057 | ||
1058 | void | |
1059 | DateTimePatternGenerator::setDecimal(const UnicodeString& newDecimal) { | |
1060 | this->decimal = newDecimal; | |
1061 | // NUL-terminate for the C API. | |
1062 | this->decimal.getTerminatedBuffer(); | |
1063 | } | |
1064 | ||
1065 | const UnicodeString& | |
1066 | DateTimePatternGenerator::getDecimal() const { | |
1067 | return decimal; | |
1068 | } | |
1069 | ||
1070 | void | |
f3c0d7a5 A |
1071 | DateTimePatternGenerator::addCanonicalItems(UErrorCode& status) { |
1072 | if (U_FAILURE(status)) { return; } | |
46f4442e | 1073 | UnicodeString conflictingPattern; |
46f4442e A |
1074 | |
1075 | for (int32_t i=0; i<UDATPG_FIELD_COUNT; i++) { | |
f3c0d7a5 A |
1076 | if (Canonical_Items[i] > 0) { |
1077 | addPattern(UnicodeString(Canonical_Items[i]), FALSE, conflictingPattern, status); | |
1078 | } | |
1079 | if (U_FAILURE(status)) { return; } | |
46f4442e A |
1080 | } |
1081 | } | |
1082 | ||
1083 | void | |
1084 | DateTimePatternGenerator::setDateTimeFormat(const UnicodeString& dtFormat) { | |
1085 | dateTimeFormat = dtFormat; | |
1086 | // NUL-terminate for the C API. | |
1087 | dateTimeFormat.getTerminatedBuffer(); | |
1088 | } | |
1089 | ||
1090 | const UnicodeString& | |
1091 | DateTimePatternGenerator::getDateTimeFormat() const { | |
1092 | return dateTimeFormat; | |
1093 | } | |
1094 | ||
1095 | void | |
1096 | DateTimePatternGenerator::setDateTimeFromCalendar(const Locale& locale, UErrorCode& status) { | |
1097 | const UChar *resStr; | |
1098 | int32_t resStrLen = 0; | |
729e4ab9 | 1099 | |
46f4442e | 1100 | Calendar* fCalendar = Calendar::createInstance(locale, status); |
f3c0d7a5 A |
1101 | if (U_FAILURE(status)) { return; } |
1102 | ||
1103 | LocalUResourceBundlePointer calData(ures_open(NULL, locale.getBaseName(), &status)); | |
1104 | ures_getByKey(calData.getAlias(), DT_DateTimeCalendarTag, calData.getAlias(), &status); | |
1105 | ||
1106 | LocalUResourceBundlePointer dateTimePatterns; | |
1107 | if (fCalendar != NULL && fCalendar->getType() != NULL && *fCalendar->getType() != '\0' | |
1108 | && uprv_strcmp(fCalendar->getType(), DT_DateTimeGregorianTag) != 0) { | |
1109 | dateTimePatterns.adoptInstead(ures_getByKeyWithFallback(calData.getAlias(), fCalendar->getType(), | |
1110 | NULL, &status)); | |
1111 | ures_getByKeyWithFallback(dateTimePatterns.getAlias(), DT_DateTimePatternsTag, | |
1112 | dateTimePatterns.getAlias(), &status); | |
1113 | } | |
1114 | ||
1115 | if (dateTimePatterns.isNull() || status == U_MISSING_RESOURCE_ERROR) { | |
1116 | status = U_ZERO_ERROR; | |
1117 | dateTimePatterns.adoptInstead(ures_getByKeyWithFallback(calData.getAlias(), DT_DateTimeGregorianTag, | |
1118 | dateTimePatterns.orphan(), &status)); | |
1119 | ures_getByKeyWithFallback(dateTimePatterns.getAlias(), DT_DateTimePatternsTag, | |
1120 | dateTimePatterns.getAlias(), &status); | |
1121 | } | |
1122 | if (U_FAILURE(status)) { return; } | |
46f4442e | 1123 | |
f3c0d7a5 | 1124 | if (ures_getSize(dateTimePatterns.getAlias()) <= DateFormat::kDateTime) |
46f4442e A |
1125 | { |
1126 | status = U_INVALID_FORMAT_ERROR; | |
1127 | return; | |
1128 | } | |
f3c0d7a5 | 1129 | resStr = ures_getStringByIndex(dateTimePatterns.getAlias(), (int32_t)DateFormat::kDateTime, &resStrLen, &status); |
46f4442e | 1130 | setDateTimeFormat(UnicodeString(TRUE, resStr, resStrLen)); |
729e4ab9 | 1131 | |
46f4442e A |
1132 | delete fCalendar; |
1133 | } | |
1134 | ||
1135 | void | |
1136 | DateTimePatternGenerator::setDecimalSymbols(const Locale& locale, UErrorCode& status) { | |
1137 | DecimalFormatSymbols dfs = DecimalFormatSymbols(locale, status); | |
1138 | if(U_SUCCESS(status)) { | |
1139 | decimal = dfs.getSymbol(DecimalFormatSymbols::kDecimalSeparatorSymbol); | |
1140 | // NUL-terminate for the C API. | |
1141 | decimal.getTerminatedBuffer(); | |
1142 | } | |
1143 | } | |
1144 | ||
1145 | UDateTimePatternConflict | |
1146 | DateTimePatternGenerator::addPattern( | |
1147 | const UnicodeString& pattern, | |
1148 | UBool override, | |
1149 | UnicodeString &conflictingPattern, | |
1150 | UErrorCode& status) | |
1151 | { | |
1152 | return addPatternWithSkeleton(pattern, NULL, override, conflictingPattern, status); | |
1153 | } | |
1154 | ||
1155 | // For DateTimePatternGenerator::addPatternWithSkeleton - | |
1156 | // If skeletonToUse is specified, then an availableFormats entry is being added. In this case: | |
1157 | // 1. We pass that skeleton to matcher.set instead of having it derive a skeleton from the pattern. | |
1158 | // 2. If the new entry's skeleton or basePattern does match an existing entry but that entry also had a skeleton specified | |
1159 | // (i.e. it was also from availableFormats), then the new entry does not override it regardless of the value of the override | |
1160 | // parameter. This prevents later availableFormats entries from a parent locale overriding earlier ones from the actual | |
1161 | // specified locale. However, availableFormats entries *should* override entries with matching skeleton whose skeleton was | |
1162 | // derived (i.e. entries derived from the standard date/time patters for the specified locale). | |
1163 | // 3. When adding the pattern (patternMap->add), we set a new boolean to indicate that the added entry had a | |
1164 | // specified skeleton (which sets a new field in the PtnElem in the PatternMap). | |
1165 | UDateTimePatternConflict | |
1166 | DateTimePatternGenerator::addPatternWithSkeleton( | |
1167 | const UnicodeString& pattern, | |
1168 | const UnicodeString* skeletonToUse, | |
1169 | UBool override, | |
1170 | UnicodeString& conflictingPattern, | |
1171 | UErrorCode& status) | |
1172 | { | |
1173 | ||
1174 | UnicodeString basePattern; | |
1175 | PtnSkeleton skeleton; | |
1176 | UDateTimePatternConflict conflictingStatus = UDATPG_NO_CONFLICT; | |
1177 | ||
1178 | DateTimeMatcher matcher; | |
1179 | if ( skeletonToUse == NULL ) { | |
1180 | matcher.set(pattern, fp, skeleton); | |
1181 | matcher.getBasePattern(basePattern); | |
1182 | } else { | |
4388f060 | 1183 | matcher.set(*skeletonToUse, fp, skeleton); // no longer trims skeleton fields to max len 3, per #7930 |
46f4442e A |
1184 | matcher.getBasePattern(basePattern); // or perhaps instead: basePattern = *skeletonToUse; |
1185 | } | |
51004dcb A |
1186 | // We only care about base conflicts - and replacing the pattern associated with a base - if: |
1187 | // 1. the conflicting previous base pattern did *not* have an explicit skeleton; in that case the previous | |
1188 | // base + pattern combination was derived from either (a) a canonical item, (b) a standard format, or | |
1189 | // (c) a pattern specified programmatically with a previous call to addPattern (which would only happen | |
1190 | // if we are getting here from a subsequent call to addPattern). | |
1191 | // 2. a skeleton is specified for the current pattern, but override=false; in that case we are checking | |
1192 | // availableFormats items from root, which should not override any previous entry with the same base. | |
46f4442e A |
1193 | UBool entryHadSpecifiedSkeleton; |
1194 | const UnicodeString *duplicatePattern = patternMap->getPatternFromBasePattern(basePattern, entryHadSpecifiedSkeleton); | |
51004dcb | 1195 | if (duplicatePattern != NULL && (!entryHadSpecifiedSkeleton || (skeletonToUse != NULL && !override))) { |
46f4442e A |
1196 | conflictingStatus = UDATPG_BASE_CONFLICT; |
1197 | conflictingPattern = *duplicatePattern; | |
51004dcb | 1198 | if (!override) { |
46f4442e A |
1199 | return conflictingStatus; |
1200 | } | |
1201 | } | |
51004dcb A |
1202 | // The only time we get here with override=true and skeletonToUse!=null is when adding availableFormats |
1203 | // items from CLDR data. In that case, we don't want an item from a parent locale to replace an item with | |
1204 | // same skeleton from the specified locale, so skip the current item if skeletonWasSpecified is true for | |
1205 | // the previously-specified conflicting item. | |
46f4442e A |
1206 | const PtnSkeleton* entrySpecifiedSkeleton = NULL; |
1207 | duplicatePattern = patternMap->getPatternFromSkeleton(skeleton, &entrySpecifiedSkeleton); | |
1208 | if (duplicatePattern != NULL ) { | |
1209 | conflictingStatus = UDATPG_CONFLICT; | |
1210 | conflictingPattern = *duplicatePattern; | |
1211 | if (!override || (skeletonToUse != NULL && entrySpecifiedSkeleton != NULL)) { | |
1212 | return conflictingStatus; | |
1213 | } | |
1214 | } | |
1215 | patternMap->add(basePattern, skeleton, pattern, skeletonToUse != NULL, status); | |
1216 | if(U_FAILURE(status)) { | |
1217 | return conflictingStatus; | |
1218 | } | |
729e4ab9 | 1219 | |
46f4442e A |
1220 | return UDATPG_NO_CONFLICT; |
1221 | } | |
1222 | ||
1223 | ||
1224 | UDateTimePatternField | |
1225 | DateTimePatternGenerator::getAppendFormatNumber(const char* field) const { | |
1226 | for (int32_t i=0; i<UDATPG_FIELD_COUNT; ++i ) { | |
1227 | if (uprv_strcmp(CLDR_FIELD_APPEND[i], field)==0) { | |
1228 | return (UDateTimePatternField)i; | |
1229 | } | |
1230 | } | |
1231 | return UDATPG_FIELD_COUNT; | |
1232 | } | |
1233 | ||
1234 | UDateTimePatternField | |
1235 | DateTimePatternGenerator::getAppendNameNumber(const char* field) const { | |
1236 | for (int32_t i=0; i<UDATPG_FIELD_COUNT; ++i ) { | |
1237 | if (uprv_strcmp(CLDR_FIELD_NAME[i],field)==0) { | |
1238 | return (UDateTimePatternField)i; | |
1239 | } | |
1240 | } | |
1241 | return UDATPG_FIELD_COUNT; | |
1242 | } | |
1243 | ||
1244 | const UnicodeString* | |
1245 | DateTimePatternGenerator::getBestRaw(DateTimeMatcher& source, | |
1246 | int32_t includeMask, | |
1247 | DistanceInfo* missingFields, | |
1248 | const PtnSkeleton** specifiedSkeletonPtr) { | |
1249 | int32_t bestDistance = 0x7fffffff; | |
1250 | DistanceInfo tempInfo; | |
1251 | const UnicodeString *bestPattern=NULL; | |
1252 | const PtnSkeleton* specifiedSkeleton=NULL; | |
1253 | ||
1254 | PatternMapIterator it; | |
1255 | for (it.set(*patternMap); it.hasNext(); ) { | |
1256 | DateTimeMatcher trial = it.next(); | |
1257 | if (trial.equals(skipMatcher)) { | |
1258 | continue; | |
1259 | } | |
1260 | int32_t distance=source.getDistance(trial, includeMask, tempInfo); | |
1261 | if (distance<bestDistance) { | |
1262 | bestDistance=distance; | |
1263 | bestPattern=patternMap->getPatternFromSkeleton(*trial.getSkeletonPtr(), &specifiedSkeleton); | |
1264 | missingFields->setTo(tempInfo); | |
1265 | if (distance==0) { | |
1266 | break; | |
1267 | } | |
1268 | } | |
1269 | } | |
1270 | ||
1271 | // If the best raw match had a specified skeleton and that skeleton was requested by the caller, | |
1272 | // then return it too. This generally happens when the caller needs to pass that skeleton | |
1273 | // through to adjustFieldTypes so the latter can do a better job. | |
1274 | if (bestPattern && specifiedSkeletonPtr) { | |
1275 | *specifiedSkeletonPtr = specifiedSkeleton; | |
1276 | } | |
1277 | return bestPattern; | |
1278 | } | |
1279 | ||
1280 | UnicodeString | |
1281 | DateTimePatternGenerator::adjustFieldTypes(const UnicodeString& pattern, | |
1282 | const PtnSkeleton* specifiedSkeleton, | |
57a6839d | 1283 | int32_t flags, |
729e4ab9 | 1284 | UDateTimePatternMatchOptions options) { |
46f4442e A |
1285 | UnicodeString newPattern; |
1286 | fp->set(pattern); | |
1287 | for (int32_t i=0; i < fp->itemNumber; i++) { | |
1288 | UnicodeString field = fp->items[i]; | |
1289 | if ( fp->isQuoteLiteral(field) ) { | |
1290 | ||
1291 | UnicodeString quoteLiteral; | |
1292 | fp->getQuoteLiteral(quoteLiteral, &i); | |
1293 | newPattern += quoteLiteral; | |
1294 | } | |
1295 | else { | |
1296 | if (fp->isPatternSeparator(field)) { | |
1297 | newPattern+=field; | |
1298 | continue; | |
1299 | } | |
1300 | int32_t canonicalIndex = fp->getCanonicalIndex(field); | |
1301 | if (canonicalIndex < 0) { | |
1302 | newPattern+=field; | |
1303 | continue; // don't adjust | |
1304 | } | |
1305 | const dtTypeElem *row = &dtTypes[canonicalIndex]; | |
1306 | int32_t typeValue = row->field; | |
2ca993e8 A |
1307 | |
1308 | // Handle special day periods. | |
1309 | if (typeValue == UDATPG_DAYPERIOD_FIELD && flags != 0) { | |
1310 | UChar c = NONE; // '0' | |
1311 | if (flags & kDTPGSkeletonUsesCapB) { c = CAP_B; } | |
1312 | if (flags & kDTPGSkeletonUsesLowB) { c = LOW_B; } | |
1313 | ||
1314 | if (c != NONE) { | |
1315 | for (int32_t i = 0; i < field.length(); ++i) | |
1316 | field.setCharAt(i, c); | |
1317 | } | |
1318 | } | |
1319 | ||
57a6839d | 1320 | if ((flags & kDTPGFixFractionalSeconds) != 0 && typeValue == UDATPG_SECOND_FIELD) { |
f3c0d7a5 A |
1321 | field += decimal; |
1322 | dtMatcher->skeleton.original.appendFieldTo(UDATPG_FRACTIONAL_SECOND_FIELD, field); | |
729e4ab9 | 1323 | } else if (dtMatcher->skeleton.type[typeValue]!=0) { |
46f4442e A |
1324 | // Here: |
1325 | // - "reqField" is the field from the originally requested skeleton, with length | |
1326 | // "reqFieldLen". | |
1327 | // - "field" is the field from the found pattern. | |
1328 | // | |
1329 | // The adjusted field should consist of characters from the originally requested | |
729e4ab9 | 1330 | // skeleton, except in the case of UDATPG_HOUR_FIELD or UDATPG_MONTH_FIELD or |
4388f060 A |
1331 | // UDATPG_WEEKDAY_FIELD or UDATPG_YEAR_FIELD, in which case it should consist |
1332 | // of characters from the found pattern. | |
46f4442e A |
1333 | // |
1334 | // The length of the adjusted field (adjFieldLen) should match that in the originally | |
729e4ab9 A |
1335 | // requested skeleton, except that in the following cases the length of the adjusted field |
1336 | // should match that in the found pattern (i.e. the length of this pattern field should | |
1337 | // not be adjusted): | |
1338 | // 1. typeValue is UDATPG_HOUR_FIELD/MINUTE/SECOND and the corresponding bit in options is | |
1339 | // not set (ticket #7180). Note, we may want to implement a similar change for other | |
1340 | // numeric fields (MM, dd, etc.) so the default behavior is to get locale preference for | |
1341 | // field length, but options bits can be used to override this. | |
1342 | // 2. There is a specified skeleton for the found pattern and one of the following is true: | |
1343 | // a) The length of the field in the skeleton (skelFieldLen) is equal to reqFieldLen. | |
1344 | // b) The pattern field is numeric and the skeleton field is not, or vice versa. | |
46f4442e | 1345 | |
f3c0d7a5 A |
1346 | UChar reqFieldChar = dtMatcher->skeleton.original.getFieldChar(typeValue); |
1347 | int32_t reqFieldLen = dtMatcher->skeleton.original.getFieldLength(typeValue); | |
1348 | if (reqFieldChar == CAP_E && reqFieldLen < 3) | |
729e4ab9 | 1349 | reqFieldLen = 3; // 1-3 for E are equivalent to 3 for c,e |
46f4442e | 1350 | int32_t adjFieldLen = reqFieldLen; |
729e4ab9 A |
1351 | if ( (typeValue==UDATPG_HOUR_FIELD && (options & UDATPG_MATCH_HOUR_FIELD_LENGTH)==0) || |
1352 | (typeValue==UDATPG_MINUTE_FIELD && (options & UDATPG_MATCH_MINUTE_FIELD_LENGTH)==0) || | |
1353 | (typeValue==UDATPG_SECOND_FIELD && (options & UDATPG_MATCH_SECOND_FIELD_LENGTH)==0) ) { | |
1354 | adjFieldLen = field.length(); | |
1355 | } else if (specifiedSkeleton) { | |
f3c0d7a5 | 1356 | int32_t skelFieldLen = specifiedSkeleton->original.getFieldLength(typeValue); |
46f4442e A |
1357 | UBool patFieldIsNumeric = (row->type > 0); |
1358 | UBool skelFieldIsNumeric = (specifiedSkeleton->type[typeValue] > 0); | |
1359 | if (skelFieldLen == reqFieldLen || (patFieldIsNumeric && !skelFieldIsNumeric) || (skelFieldIsNumeric && !patFieldIsNumeric)) { | |
1360 | // don't adjust the field length in the found pattern | |
1361 | adjFieldLen = field.length(); | |
1362 | } | |
1363 | } | |
f3c0d7a5 A |
1364 | UChar c = (typeValue!= UDATPG_HOUR_FIELD |
1365 | && typeValue!= UDATPG_MONTH_FIELD | |
1366 | && typeValue!= UDATPG_WEEKDAY_FIELD | |
1367 | && (typeValue!= UDATPG_YEAR_FIELD || reqFieldChar==CAP_Y)) | |
1368 | ? reqFieldChar | |
1369 | : field.charAt(0); | |
57a6839d A |
1370 | if (typeValue == UDATPG_HOUR_FIELD && (flags & kDTPGSkeletonUsesCapJ) != 0) { |
1371 | c = fDefaultHourFormatChar; | |
1372 | switch (options & UADATPG_FORCE_HOUR_CYCLE_MASK) { | |
1373 | case UADATPG_FORCE_12_HOUR_CYCLE: | |
1374 | if (c == CAP_H || c == LOW_K) { | |
1375 | // Have 24-hour cycle, change to 12-hour cycle. | |
1376 | // Should have better way to pick 'h' or 'K'. | |
1377 | c = LOW_H; | |
1378 | }; | |
1379 | break; | |
1380 | case UADATPG_FORCE_24_HOUR_CYCLE: | |
1381 | if (c == LOW_H || c == CAP_K) { | |
1382 | // Have 12-hour cycle, change to 24-hour cycle. | |
1383 | // Should have better way to pick 'H' or 'k'. | |
1384 | c = CAP_H; | |
1385 | }; | |
1386 | break; | |
1387 | default: | |
1388 | break; | |
1389 | } | |
1390 | } | |
46f4442e A |
1391 | field.remove(); |
1392 | for (int32_t i=adjFieldLen; i>0; --i) { | |
1393 | field+=c; | |
1394 | } | |
46f4442e | 1395 | } |
729e4ab9 | 1396 | newPattern+=field; |
46f4442e A |
1397 | } |
1398 | } | |
1399 | return newPattern; | |
1400 | } | |
1401 | ||
1402 | UnicodeString | |
57a6839d | 1403 | DateTimePatternGenerator::getBestAppending(int32_t missingFields, int32_t flags, UDateTimePatternMatchOptions options) { |
729e4ab9 | 1404 | UnicodeString resultPattern, tempPattern; |
46f4442e A |
1405 | UErrorCode err=U_ZERO_ERROR; |
1406 | int32_t lastMissingFieldMask=0; | |
1407 | if (missingFields!=0) { | |
1408 | resultPattern=UnicodeString(); | |
1409 | const PtnSkeleton* specifiedSkeleton=NULL; | |
1410 | tempPattern = *getBestRaw(*dtMatcher, missingFields, distanceInfo, &specifiedSkeleton); | |
57a6839d | 1411 | resultPattern = adjustFieldTypes(tempPattern, specifiedSkeleton, flags, options); |
46f4442e A |
1412 | if ( distanceInfo->missingFieldMask==0 ) { |
1413 | return resultPattern; | |
1414 | } | |
1415 | while (distanceInfo->missingFieldMask!=0) { // precondition: EVERY single field must work! | |
1416 | if ( lastMissingFieldMask == distanceInfo->missingFieldMask ) { | |
1417 | break; // cannot find the proper missing field | |
1418 | } | |
1419 | if (((distanceInfo->missingFieldMask & UDATPG_SECOND_AND_FRACTIONAL_MASK)==UDATPG_FRACTIONAL_MASK) && | |
1420 | ((missingFields & UDATPG_SECOND_AND_FRACTIONAL_MASK) == UDATPG_SECOND_AND_FRACTIONAL_MASK)) { | |
57a6839d | 1421 | resultPattern = adjustFieldTypes(resultPattern, specifiedSkeleton, flags | kDTPGFixFractionalSeconds, options); |
46f4442e A |
1422 | distanceInfo->missingFieldMask &= ~UDATPG_FRACTIONAL_MASK; |
1423 | continue; | |
1424 | } | |
1425 | int32_t startingMask = distanceInfo->missingFieldMask; | |
1426 | tempPattern = *getBestRaw(*dtMatcher, distanceInfo->missingFieldMask, distanceInfo, &specifiedSkeleton); | |
57a6839d | 1427 | tempPattern = adjustFieldTypes(tempPattern, specifiedSkeleton, flags, options); |
46f4442e A |
1428 | int32_t foundMask=startingMask& ~distanceInfo->missingFieldMask; |
1429 | int32_t topField=getTopBitNumber(foundMask); | |
1430 | UnicodeString appendName; | |
1431 | getAppendName((UDateTimePatternField)topField, appendName); | |
2ca993e8 A |
1432 | const UnicodeString *values[3] = { |
1433 | &resultPattern, | |
1434 | &tempPattern, | |
1435 | &appendName | |
46f4442e | 1436 | }; |
2ca993e8 A |
1437 | SimpleFormatter(appendItemFormats[topField], 2, 3, err). |
1438 | formatAndReplace(values, 3, resultPattern, NULL, 0, err); | |
46f4442e A |
1439 | lastMissingFieldMask = distanceInfo->missingFieldMask; |
1440 | } | |
1441 | } | |
729e4ab9 | 1442 | return resultPattern; |
46f4442e A |
1443 | } |
1444 | ||
1445 | int32_t | |
1446 | DateTimePatternGenerator::getTopBitNumber(int32_t foundMask) { | |
1447 | if ( foundMask==0 ) { | |
1448 | return 0; | |
1449 | } | |
1450 | int32_t i=0; | |
1451 | while (foundMask!=0) { | |
1452 | foundMask >>=1; | |
1453 | ++i; | |
1454 | } | |
1455 | if (i-1 >UDATPG_ZONE_FIELD) { | |
1456 | return UDATPG_ZONE_FIELD; | |
1457 | } | |
729e4ab9 | 1458 | else |
46f4442e A |
1459 | return i-1; |
1460 | } | |
1461 | ||
1462 | void | |
1463 | DateTimePatternGenerator::setAvailableFormat(const UnicodeString &key, UErrorCode& err) | |
1464 | { | |
1465 | fAvailableFormatKeyHash->puti(key, 1, err); | |
1466 | } | |
1467 | ||
1468 | UBool | |
1469 | DateTimePatternGenerator::isAvailableFormatSet(const UnicodeString &key) const { | |
1470 | return (UBool)(fAvailableFormatKeyHash->geti(key) == 1); | |
1471 | } | |
1472 | ||
1473 | void | |
1474 | DateTimePatternGenerator::copyHashtable(Hashtable *other, UErrorCode &status) { | |
1475 | ||
1476 | if (other == NULL) { | |
1477 | return; | |
1478 | } | |
1479 | if (fAvailableFormatKeyHash != NULL) { | |
1480 | delete fAvailableFormatKeyHash; | |
1481 | fAvailableFormatKeyHash = NULL; | |
1482 | } | |
1483 | initHashtable(status); | |
1484 | if(U_FAILURE(status)){ | |
1485 | return; | |
1486 | } | |
b331163b | 1487 | int32_t pos = UHASH_FIRST; |
46f4442e A |
1488 | const UHashElement* elem = NULL; |
1489 | // walk through the hash table and create a deep clone | |
1490 | while((elem = other->nextElement(pos))!= NULL){ | |
1491 | const UHashTok otherKeyTok = elem->key; | |
1492 | UnicodeString* otherKey = (UnicodeString*)otherKeyTok.pointer; | |
1493 | fAvailableFormatKeyHash->puti(*otherKey, 1, status); | |
1494 | if(U_FAILURE(status)){ | |
1495 | return; | |
1496 | } | |
1497 | } | |
1498 | } | |
1499 | ||
1500 | StringEnumeration* | |
1501 | DateTimePatternGenerator::getSkeletons(UErrorCode& status) const { | |
1502 | StringEnumeration* skeletonEnumerator = new DTSkeletonEnumeration(*patternMap, DT_SKELETON, status); | |
1503 | return skeletonEnumerator; | |
1504 | } | |
1505 | ||
1506 | const UnicodeString& | |
1507 | DateTimePatternGenerator::getPatternForSkeleton(const UnicodeString& skeleton) const { | |
1508 | PtnElem *curElem; | |
729e4ab9 | 1509 | |
46f4442e A |
1510 | if (skeleton.length() ==0) { |
1511 | return emptyString; | |
729e4ab9 | 1512 | } |
46f4442e A |
1513 | curElem = patternMap->getHeader(skeleton.charAt(0)); |
1514 | while ( curElem != NULL ) { | |
1515 | if ( curElem->skeleton->getSkeleton()==skeleton ) { | |
1516 | return curElem->pattern; | |
1517 | } | |
1518 | curElem=curElem->next; | |
1519 | } | |
1520 | return emptyString; | |
1521 | } | |
1522 | ||
1523 | StringEnumeration* | |
1524 | DateTimePatternGenerator::getBaseSkeletons(UErrorCode& status) const { | |
1525 | StringEnumeration* baseSkeletonEnumerator = new DTSkeletonEnumeration(*patternMap, DT_BASESKELETON, status); | |
1526 | return baseSkeletonEnumerator; | |
1527 | } | |
1528 | ||
1529 | StringEnumeration* | |
1530 | DateTimePatternGenerator::getRedundants(UErrorCode& status) { | |
1531 | StringEnumeration* output = new DTRedundantEnumeration(); | |
1532 | const UnicodeString *pattern; | |
1533 | PatternMapIterator it; | |
1534 | for (it.set(*patternMap); it.hasNext(); ) { | |
1535 | DateTimeMatcher current = it.next(); | |
1536 | pattern = patternMap->getPatternFromSkeleton(*(it.getSkeleton())); | |
1537 | if ( isCanonicalItem(*pattern) ) { | |
1538 | continue; | |
1539 | } | |
1540 | if ( skipMatcher == NULL ) { | |
1541 | skipMatcher = new DateTimeMatcher(current); | |
1542 | } | |
1543 | else { | |
1544 | *skipMatcher = current; | |
1545 | } | |
1546 | UnicodeString trial = getBestPattern(current.getPattern(), status); | |
729e4ab9 | 1547 | if (trial == *pattern) { |
46f4442e A |
1548 | ((DTRedundantEnumeration *)output)->add(*pattern, status); |
1549 | } | |
1550 | if (current.equals(skipMatcher)) { | |
1551 | continue; | |
1552 | } | |
1553 | } | |
1554 | return output; | |
1555 | } | |
1556 | ||
1557 | UBool | |
1558 | DateTimePatternGenerator::isCanonicalItem(const UnicodeString& item) const { | |
1559 | if ( item.length() != 1 ) { | |
1560 | return FALSE; | |
1561 | } | |
1562 | for (int32_t i=0; i<UDATPG_FIELD_COUNT; ++i) { | |
1563 | if (item.charAt(0)==Canonical_Items[i]) { | |
1564 | return TRUE; | |
1565 | } | |
1566 | } | |
1567 | return FALSE; | |
1568 | } | |
1569 | ||
1570 | ||
1571 | DateTimePatternGenerator* | |
1572 | DateTimePatternGenerator::clone() const { | |
1573 | return new DateTimePatternGenerator(*this); | |
1574 | } | |
1575 | ||
1576 | PatternMap::PatternMap() { | |
1577 | for (int32_t i=0; i < MAX_PATTERN_ENTRIES; ++i ) { | |
1578 | boot[i]=NULL; | |
1579 | } | |
1580 | isDupAllowed = TRUE; | |
1581 | } | |
1582 | ||
1583 | void | |
1584 | PatternMap::copyFrom(const PatternMap& other, UErrorCode& status) { | |
1585 | this->isDupAllowed = other.isDupAllowed; | |
1586 | for (int32_t bootIndex=0; bootIndex<MAX_PATTERN_ENTRIES; ++bootIndex ) { | |
1587 | PtnElem *curElem, *otherElem, *prevElem=NULL; | |
1588 | otherElem = other.boot[bootIndex]; | |
1589 | while (otherElem!=NULL) { | |
1590 | if ((curElem = new PtnElem(otherElem->basePattern, otherElem->pattern))==NULL) { | |
1591 | // out of memory | |
1592 | status = U_MEMORY_ALLOCATION_ERROR; | |
1593 | return; | |
1594 | } | |
1595 | if ( this->boot[bootIndex]== NULL ) { | |
1596 | this->boot[bootIndex] = curElem; | |
1597 | } | |
1598 | if ((curElem->skeleton=new PtnSkeleton(*(otherElem->skeleton))) == NULL ) { | |
1599 | // out of memory | |
1600 | status = U_MEMORY_ALLOCATION_ERROR; | |
1601 | return; | |
1602 | } | |
b331163b | 1603 | curElem->skeletonWasSpecified = otherElem->skeletonWasSpecified; |
46f4442e A |
1604 | if (prevElem!=NULL) { |
1605 | prevElem->next=curElem; | |
1606 | } | |
1607 | curElem->next=NULL; | |
1608 | prevElem = curElem; | |
1609 | otherElem = otherElem->next; | |
1610 | } | |
1611 | ||
1612 | } | |
1613 | } | |
1614 | ||
729e4ab9 | 1615 | PtnElem* |
46f4442e A |
1616 | PatternMap::getHeader(UChar baseChar) { |
1617 | PtnElem* curElem; | |
729e4ab9 | 1618 | |
46f4442e A |
1619 | if ( (baseChar >= CAP_A) && (baseChar <= CAP_Z) ) { |
1620 | curElem = boot[baseChar-CAP_A]; | |
1621 | } | |
1622 | else { | |
1623 | if ( (baseChar >=LOW_A) && (baseChar <= LOW_Z) ) { | |
1624 | curElem = boot[26+baseChar-LOW_A]; | |
1625 | } | |
1626 | else { | |
1627 | return NULL; | |
1628 | } | |
1629 | } | |
1630 | return curElem; | |
1631 | } | |
729e4ab9 | 1632 | |
46f4442e A |
1633 | PatternMap::~PatternMap() { |
1634 | for (int32_t i=0; i < MAX_PATTERN_ENTRIES; ++i ) { | |
1635 | if (boot[i]!=NULL ) { | |
1636 | delete boot[i]; | |
1637 | boot[i]=NULL; | |
1638 | } | |
1639 | } | |
1640 | } // PatternMap destructor | |
1641 | ||
1642 | void | |
1643 | PatternMap::add(const UnicodeString& basePattern, | |
1644 | const PtnSkeleton& skeleton, | |
1645 | const UnicodeString& value,// mapped pattern value | |
1646 | UBool skeletonWasSpecified, | |
1647 | UErrorCode &status) { | |
1648 | UChar baseChar = basePattern.charAt(0); | |
1649 | PtnElem *curElem, *baseElem; | |
1650 | status = U_ZERO_ERROR; | |
1651 | ||
1652 | // the baseChar must be A-Z or a-z | |
1653 | if ((baseChar >= CAP_A) && (baseChar <= CAP_Z)) { | |
1654 | baseElem = boot[baseChar-CAP_A]; | |
1655 | } | |
1656 | else { | |
1657 | if ((baseChar >=LOW_A) && (baseChar <= LOW_Z)) { | |
1658 | baseElem = boot[26+baseChar-LOW_A]; | |
1659 | } | |
1660 | else { | |
1661 | status = U_ILLEGAL_CHARACTER; | |
1662 | return; | |
1663 | } | |
1664 | } | |
1665 | ||
1666 | if (baseElem == NULL) { | |
1667 | if ((curElem = new PtnElem(basePattern, value)) == NULL ) { | |
1668 | // out of memory | |
1669 | status = U_MEMORY_ALLOCATION_ERROR; | |
1670 | return; | |
1671 | } | |
1672 | if (baseChar >= LOW_A) { | |
1673 | boot[26 + (baseChar-LOW_A)] = curElem; | |
1674 | } | |
1675 | else { | |
1676 | boot[baseChar-CAP_A] = curElem; | |
1677 | } | |
1678 | curElem->skeleton = new PtnSkeleton(skeleton); | |
1679 | curElem->skeletonWasSpecified = skeletonWasSpecified; | |
1680 | } | |
1681 | if ( baseElem != NULL ) { | |
1682 | curElem = getDuplicateElem(basePattern, skeleton, baseElem); | |
1683 | ||
1684 | if (curElem == NULL) { | |
1685 | // add new element to the list. | |
1686 | curElem = baseElem; | |
1687 | while( curElem -> next != NULL ) | |
1688 | { | |
1689 | curElem = curElem->next; | |
1690 | } | |
1691 | if ((curElem->next = new PtnElem(basePattern, value)) == NULL ) { | |
1692 | // out of memory | |
1693 | status = U_MEMORY_ALLOCATION_ERROR; | |
1694 | return; | |
1695 | } | |
1696 | curElem=curElem->next; | |
1697 | curElem->skeleton = new PtnSkeleton(skeleton); | |
1698 | curElem->skeletonWasSpecified = skeletonWasSpecified; | |
1699 | } | |
1700 | else { | |
1701 | // Pattern exists in the list already. | |
1702 | if ( !isDupAllowed ) { | |
1703 | return; | |
1704 | } | |
1705 | // Overwrite the value. | |
1706 | curElem->pattern = value; | |
51004dcb A |
1707 | // It was a bug that we were not doing the following previously, |
1708 | // though that bug hid other problems by making things partly work. | |
1709 | curElem->skeletonWasSpecified = skeletonWasSpecified; | |
46f4442e A |
1710 | } |
1711 | } | |
1712 | } // PatternMap::add | |
1713 | ||
1714 | // Find the pattern from the given basePattern string. | |
1715 | const UnicodeString * | |
1716 | PatternMap::getPatternFromBasePattern(UnicodeString& basePattern, UBool& skeletonWasSpecified) { // key to search for | |
1717 | PtnElem *curElem; | |
1718 | ||
1719 | if ((curElem=getHeader(basePattern.charAt(0)))==NULL) { | |
1720 | return NULL; // no match | |
1721 | } | |
1722 | ||
1723 | do { | |
1724 | if ( basePattern.compare(curElem->basePattern)==0 ) { | |
1725 | skeletonWasSpecified = curElem->skeletonWasSpecified; | |
1726 | return &(curElem->pattern); | |
1727 | } | |
1728 | curElem=curElem->next; | |
1729 | }while (curElem != NULL); | |
1730 | ||
1731 | return NULL; | |
1732 | } // PatternMap::getFromBasePattern | |
1733 | ||
1734 | ||
1735 | // Find the pattern from the given skeleton. | |
1736 | // At least when this is called from getBestRaw & addPattern (in which case specifiedSkeletonPtr is non-NULL), | |
1737 | // the comparison should be based on skeleton.original (which is unique and tied to the distance measurement in bestRaw) | |
1738 | // and not skeleton.baseOriginal (which is not unique); otherwise we may pick a different skeleton than the one with the | |
1739 | // optimum distance value in getBestRaw. When this is called from public getRedundants (specifiedSkeletonPtr is NULL), | |
1740 | // for now it will continue to compare based on baseOriginal so as not to change the behavior unnecessarily. | |
1741 | const UnicodeString * | |
1742 | PatternMap::getPatternFromSkeleton(PtnSkeleton& skeleton, const PtnSkeleton** specifiedSkeletonPtr) { // key to search for | |
1743 | PtnElem *curElem; | |
1744 | ||
1745 | if (specifiedSkeletonPtr) { | |
1746 | *specifiedSkeletonPtr = NULL; | |
1747 | } | |
1748 | ||
1749 | // find boot entry | |
f3c0d7a5 | 1750 | UChar baseChar = skeleton.getFirstChar(); |
46f4442e A |
1751 | if ((curElem=getHeader(baseChar))==NULL) { |
1752 | return NULL; // no match | |
1753 | } | |
1754 | ||
1755 | do { | |
f3c0d7a5 | 1756 | UBool equal; |
46f4442e | 1757 | if (specifiedSkeletonPtr != NULL) { // called from DateTimePatternGenerator::getBestRaw or addPattern, use original |
f3c0d7a5 | 1758 | equal = curElem->skeleton->original == skeleton.original; |
46f4442e | 1759 | } else { // called from DateTimePatternGenerator::getRedundants, use baseOriginal |
f3c0d7a5 | 1760 | equal = curElem->skeleton->baseOriginal == skeleton.baseOriginal; |
46f4442e | 1761 | } |
f3c0d7a5 | 1762 | if (equal) { |
46f4442e A |
1763 | if (specifiedSkeletonPtr && curElem->skeletonWasSpecified) { |
1764 | *specifiedSkeletonPtr = curElem->skeleton; | |
1765 | } | |
1766 | return &(curElem->pattern); | |
1767 | } | |
1768 | curElem=curElem->next; | |
1769 | }while (curElem != NULL); | |
1770 | ||
1771 | return NULL; | |
729e4ab9 | 1772 | } |
46f4442e A |
1773 | |
1774 | UBool | |
1775 | PatternMap::equals(const PatternMap& other) { | |
1776 | if ( this==&other ) { | |
1777 | return TRUE; | |
1778 | } | |
1779 | for (int32_t bootIndex=0; bootIndex<MAX_PATTERN_ENTRIES; ++bootIndex ) { | |
1780 | if ( boot[bootIndex]==other.boot[bootIndex] ) { | |
1781 | continue; | |
1782 | } | |
1783 | if ( (boot[bootIndex]==NULL)||(other.boot[bootIndex]==NULL) ) { | |
1784 | return FALSE; | |
1785 | } | |
1786 | PtnElem *otherElem = other.boot[bootIndex]; | |
1787 | PtnElem *myElem = boot[bootIndex]; | |
1788 | while ((otherElem!=NULL) || (myElem!=NULL)) { | |
1789 | if ( myElem == otherElem ) { | |
1790 | break; | |
1791 | } | |
1792 | if ((otherElem==NULL) || (myElem==NULL)) { | |
1793 | return FALSE; | |
1794 | } | |
1795 | if ( (myElem->basePattern != otherElem->basePattern) || | |
1796 | (myElem->pattern != otherElem->pattern) ) { | |
1797 | return FALSE; | |
1798 | } | |
1799 | if ((myElem->skeleton!=otherElem->skeleton)&& | |
1800 | !myElem->skeleton->equals(*(otherElem->skeleton))) { | |
1801 | return FALSE; | |
1802 | } | |
1803 | myElem = myElem->next; | |
1804 | otherElem=otherElem->next; | |
1805 | } | |
1806 | } | |
1807 | return TRUE; | |
1808 | } | |
1809 | ||
1810 | // find any key existing in the mapping table already. | |
1811 | // return TRUE if there is an existing key, otherwise return FALSE. | |
1812 | PtnElem* | |
1813 | PatternMap::getDuplicateElem( | |
1814 | const UnicodeString &basePattern, | |
1815 | const PtnSkeleton &skeleton, | |
1816 | PtnElem *baseElem) { | |
1817 | PtnElem *curElem; | |
1818 | ||
1819 | if ( baseElem == (PtnElem *)NULL ) { | |
1820 | return (PtnElem*)NULL; | |
1821 | } | |
1822 | else { | |
1823 | curElem = baseElem; | |
1824 | } | |
1825 | do { | |
1826 | if ( basePattern.compare(curElem->basePattern)==0 ) { | |
1827 | UBool isEqual=TRUE; | |
1828 | for (int32_t i=0; i<UDATPG_FIELD_COUNT; ++i) { | |
1829 | if (curElem->skeleton->type[i] != skeleton.type[i] ) { | |
1830 | isEqual=FALSE; | |
1831 | break; | |
1832 | } | |
1833 | } | |
1834 | if (isEqual) { | |
1835 | return curElem; | |
1836 | } | |
1837 | } | |
1838 | curElem = curElem->next; | |
1839 | } while( curElem != (PtnElem *)NULL ); | |
1840 | ||
1841 | // end of the list | |
1842 | return (PtnElem*)NULL; | |
1843 | ||
1844 | } // PatternMap::getDuplicateElem | |
1845 | ||
1846 | DateTimeMatcher::DateTimeMatcher(void) { | |
1847 | } | |
1848 | ||
4388f060 A |
1849 | DateTimeMatcher::~DateTimeMatcher() {} |
1850 | ||
46f4442e A |
1851 | DateTimeMatcher::DateTimeMatcher(const DateTimeMatcher& other) { |
1852 | copyFrom(other.skeleton); | |
1853 | } | |
1854 | ||
1855 | ||
1856 | void | |
1857 | DateTimeMatcher::set(const UnicodeString& pattern, FormatParser* fp) { | |
1858 | PtnSkeleton localSkeleton; | |
1859 | return set(pattern, fp, localSkeleton); | |
1860 | } | |
1861 | ||
1862 | void | |
1863 | DateTimeMatcher::set(const UnicodeString& pattern, FormatParser* fp, PtnSkeleton& skeletonResult) { | |
1864 | int32_t i; | |
1865 | for (i=0; i<UDATPG_FIELD_COUNT; ++i) { | |
f3c0d7a5 | 1866 | skeletonResult.type[i] = NONE; |
46f4442e A |
1867 | } |
1868 | fp->set(pattern); | |
1869 | for (i=0; i < fp->itemNumber; i++) { | |
f3c0d7a5 A |
1870 | const UnicodeString& value = fp->items[i]; |
1871 | if ( value.charAt(0) == LOW_A ) { | |
46f4442e A |
1872 | continue; // skip 'a' |
1873 | } | |
1874 | ||
f3c0d7a5 | 1875 | if ( fp->isQuoteLiteral(value) ) { |
46f4442e A |
1876 | UnicodeString quoteLiteral; |
1877 | fp->getQuoteLiteral(quoteLiteral, &i); | |
1878 | continue; | |
1879 | } | |
f3c0d7a5 | 1880 | int32_t canonicalIndex = fp->getCanonicalIndex(value); |
46f4442e A |
1881 | if (canonicalIndex < 0 ) { |
1882 | continue; | |
1883 | } | |
1884 | const dtTypeElem *row = &dtTypes[canonicalIndex]; | |
f3c0d7a5 A |
1885 | int32_t field = row->field; |
1886 | skeletonResult.original.populate(field, value); | |
46f4442e | 1887 | UChar repeatChar = row->patternChar; |
4388f060 | 1888 | int32_t repeatCount = row->minLen; // #7930 removes cap at 3 |
f3c0d7a5 A |
1889 | skeletonResult.baseOriginal.populate(field, repeatChar, repeatCount); |
1890 | int16_t subField = row->type; | |
46f4442e | 1891 | if ( row->type > 0) { |
f3c0d7a5 | 1892 | subField += value.length(); |
46f4442e | 1893 | } |
f3c0d7a5 | 1894 | skeletonResult.type[field] = subField; |
46f4442e A |
1895 | } |
1896 | copyFrom(skeletonResult); | |
1897 | } | |
1898 | ||
1899 | void | |
1900 | DateTimeMatcher::getBasePattern(UnicodeString &result ) { | |
1901 | result.remove(); // Reset the result first. | |
f3c0d7a5 | 1902 | skeleton.baseOriginal.appendTo(result); |
46f4442e A |
1903 | } |
1904 | ||
1905 | UnicodeString | |
1906 | DateTimeMatcher::getPattern() { | |
1907 | UnicodeString result; | |
f3c0d7a5 | 1908 | return skeleton.original.appendTo(result); |
46f4442e A |
1909 | } |
1910 | ||
1911 | int32_t | |
1912 | DateTimeMatcher::getDistance(const DateTimeMatcher& other, int32_t includeMask, DistanceInfo& distanceInfo) { | |
1913 | int32_t result=0; | |
1914 | distanceInfo.clear(); | |
1915 | for (int32_t i=0; i<UDATPG_FIELD_COUNT; ++i ) { | |
1916 | int32_t myType = (includeMask&(1<<i))==0 ? 0 : skeleton.type[i]; | |
1917 | int32_t otherType = other.skeleton.type[i]; | |
1918 | if (myType==otherType) { | |
1919 | continue; | |
1920 | } | |
1921 | if (myType==0) {// and other is not | |
1922 | result += EXTRA_FIELD; | |
1923 | distanceInfo.addExtra(i); | |
1924 | } | |
1925 | else { | |
1926 | if (otherType==0) { | |
1927 | result += MISSING_FIELD; | |
1928 | distanceInfo.addMissing(i); | |
1929 | } | |
1930 | else { | |
1931 | result += abs(myType - otherType); | |
1932 | } | |
1933 | } | |
1934 | ||
1935 | } | |
1936 | return result; | |
1937 | } | |
1938 | ||
1939 | void | |
1940 | DateTimeMatcher::copyFrom(const PtnSkeleton& newSkeleton) { | |
f3c0d7a5 | 1941 | skeleton.copyFrom(newSkeleton); |
46f4442e A |
1942 | } |
1943 | ||
1944 | void | |
1945 | DateTimeMatcher::copyFrom() { | |
1946 | // same as clear | |
f3c0d7a5 | 1947 | skeleton.clear(); |
46f4442e A |
1948 | } |
1949 | ||
1950 | UBool | |
1951 | DateTimeMatcher::equals(const DateTimeMatcher* other) const { | |
f3c0d7a5 A |
1952 | if (other==NULL) { return FALSE; } |
1953 | return skeleton.original == other->skeleton.original; | |
46f4442e A |
1954 | } |
1955 | ||
1956 | int32_t | |
1957 | DateTimeMatcher::getFieldMask() { | |
1958 | int32_t result=0; | |
1959 | ||
1960 | for (int32_t i=0; i<UDATPG_FIELD_COUNT; ++i) { | |
1961 | if (skeleton.type[i]!=0) { | |
1962 | result |= (1<<i); | |
1963 | } | |
1964 | } | |
1965 | return result; | |
1966 | } | |
1967 | ||
1968 | PtnSkeleton* | |
1969 | DateTimeMatcher::getSkeletonPtr() { | |
1970 | return &skeleton; | |
1971 | } | |
1972 | ||
1973 | FormatParser::FormatParser () { | |
1974 | status = START; | |
1975 | itemNumber=0; | |
1976 | } | |
1977 | ||
1978 | ||
1979 | FormatParser::~FormatParser () { | |
1980 | } | |
1981 | ||
1982 | ||
1983 | // Find the next token with the starting position and length | |
1984 | // Note: the startPos may | |
1985 | FormatParser::TokenStatus | |
1986 | FormatParser::setTokens(const UnicodeString& pattern, int32_t startPos, int32_t *len) { | |
1987 | int32_t curLoc = startPos; | |
1988 | if ( curLoc >= pattern.length()) { | |
1989 | return DONE; | |
1990 | } | |
1991 | // check the current char is between A-Z or a-z | |
1992 | do { | |
1993 | UChar c=pattern.charAt(curLoc); | |
1994 | if ( (c>=CAP_A && c<=CAP_Z) || (c>=LOW_A && c<=LOW_Z) ) { | |
1995 | curLoc++; | |
1996 | } | |
1997 | else { | |
1998 | startPos = curLoc; | |
1999 | *len=1; | |
2000 | return ADD_TOKEN; | |
2001 | } | |
2002 | ||
2003 | if ( pattern.charAt(curLoc)!= pattern.charAt(startPos) ) { | |
2004 | break; // not the same token | |
2005 | } | |
2006 | } while(curLoc <= pattern.length()); | |
2007 | *len = curLoc-startPos; | |
2008 | return ADD_TOKEN; | |
2009 | } | |
2010 | ||
2011 | void | |
2012 | FormatParser::set(const UnicodeString& pattern) { | |
2013 | int32_t startPos=0; | |
2014 | TokenStatus result=START; | |
2015 | int32_t len=0; | |
2016 | itemNumber =0; | |
2017 | ||
2018 | do { | |
2019 | result = setTokens( pattern, startPos, &len ); | |
2020 | if ( result == ADD_TOKEN ) | |
2021 | { | |
2022 | items[itemNumber++] = UnicodeString(pattern, startPos, len ); | |
2023 | startPos += len; | |
2024 | } | |
2025 | else { | |
2026 | break; | |
2027 | } | |
2028 | } while (result==ADD_TOKEN && itemNumber < MAX_DT_TOKEN); | |
2029 | } | |
2030 | ||
2031 | int32_t | |
729e4ab9 | 2032 | FormatParser::getCanonicalIndex(const UnicodeString& s, UBool strict) { |
46f4442e | 2033 | int32_t len = s.length(); |
729e4ab9 A |
2034 | if (len == 0) { |
2035 | return -1; | |
2036 | } | |
46f4442e | 2037 | UChar ch = s.charAt(0); |
46f4442e | 2038 | |
729e4ab9 A |
2039 | // Verify that all are the same character. |
2040 | for (int32_t l = 1; l < len; l++) { | |
2041 | if (ch != s.charAt(l)) { | |
2042 | return -1; | |
2043 | } | |
2044 | } | |
2045 | int32_t i = 0; | |
2046 | int32_t bestRow = -1; | |
f3c0d7a5 | 2047 | while (dtTypes[i].patternChar != 0x0000) { |
729e4ab9 | 2048 | if ( dtTypes[i].patternChar != ch ) { |
46f4442e A |
2049 | ++i; |
2050 | continue; | |
2051 | } | |
729e4ab9 A |
2052 | bestRow = i; |
2053 | if (dtTypes[i].patternChar != dtTypes[i+1].patternChar) { | |
46f4442e A |
2054 | return i; |
2055 | } | |
2056 | if (dtTypes[i+1].minLen <= len) { | |
2057 | ++i; | |
2058 | continue; | |
2059 | } | |
2060 | return i; | |
2061 | } | |
729e4ab9 | 2062 | return strict ? -1 : bestRow; |
46f4442e A |
2063 | } |
2064 | ||
2065 | UBool | |
2ca993e8 | 2066 | FormatParser::isQuoteLiteral(const UnicodeString& s) { |
46f4442e A |
2067 | return (UBool)(s.charAt(0)==SINGLE_QUOTE); |
2068 | } | |
2069 | ||
2070 | // This function aussumes the current itemIndex points to the quote literal. | |
2071 | // Please call isQuoteLiteral prior to this function. | |
2072 | void | |
2073 | FormatParser::getQuoteLiteral(UnicodeString& quote, int32_t *itemIndex) { | |
2074 | int32_t i=*itemIndex; | |
729e4ab9 | 2075 | |
46f4442e A |
2076 | quote.remove(); |
2077 | if (items[i].charAt(0)==SINGLE_QUOTE) { | |
2078 | quote += items[i]; | |
2079 | ++i; | |
2080 | } | |
2081 | while ( i < itemNumber ) { | |
2082 | if ( items[i].charAt(0)==SINGLE_QUOTE ) { | |
2083 | if ( (i+1<itemNumber) && (items[i+1].charAt(0)==SINGLE_QUOTE)) { | |
2084 | // two single quotes e.g. 'o''clock' | |
2085 | quote += items[i++]; | |
2086 | quote += items[i++]; | |
2087 | continue; | |
2088 | } | |
2089 | else { | |
2090 | quote += items[i]; | |
2091 | break; | |
2092 | } | |
2093 | } | |
2094 | else { | |
2095 | quote += items[i]; | |
2096 | } | |
2097 | ++i; | |
2098 | } | |
2099 | *itemIndex=i; | |
2100 | } | |
2101 | ||
2102 | UBool | |
2103 | FormatParser::isPatternSeparator(UnicodeString& field) { | |
2104 | for (int32_t i=0; i<field.length(); ++i ) { | |
2105 | UChar c= field.charAt(i); | |
2106 | if ( (c==SINGLE_QUOTE) || (c==BACKSLASH) || (c==SPACE) || (c==COLON) || | |
2107 | (c==QUOTATION_MARK) || (c==COMMA) || (c==HYPHEN) ||(items[i].charAt(0)==DOT) ) { | |
2108 | continue; | |
2109 | } | |
2110 | else { | |
2111 | return FALSE; | |
2112 | } | |
2113 | } | |
2114 | return TRUE; | |
2115 | } | |
2116 | ||
4388f060 A |
2117 | DistanceInfo::~DistanceInfo() {} |
2118 | ||
46f4442e A |
2119 | void |
2120 | DistanceInfo::setTo(DistanceInfo &other) { | |
2121 | missingFieldMask = other.missingFieldMask; | |
2122 | extraFieldMask= other.extraFieldMask; | |
2123 | } | |
2124 | ||
2125 | PatternMapIterator::PatternMapIterator() { | |
2126 | bootIndex = 0; | |
2127 | nodePtr = NULL; | |
2128 | patternMap=NULL; | |
2129 | matcher= new DateTimeMatcher(); | |
2130 | } | |
2131 | ||
2132 | ||
2133 | PatternMapIterator::~PatternMapIterator() { | |
2134 | delete matcher; | |
2135 | } | |
2136 | ||
2137 | void | |
2138 | PatternMapIterator::set(PatternMap& newPatternMap) { | |
2139 | this->patternMap=&newPatternMap; | |
2140 | } | |
2141 | ||
729e4ab9 | 2142 | PtnSkeleton* |
46f4442e A |
2143 | PatternMapIterator::getSkeleton() { |
2144 | if ( nodePtr == NULL ) { | |
2145 | return NULL; | |
2146 | } | |
2147 | else { | |
2148 | return nodePtr->skeleton; | |
2149 | } | |
2150 | } | |
2151 | ||
2152 | UBool | |
2153 | PatternMapIterator::hasNext() { | |
2154 | int32_t headIndex=bootIndex; | |
2155 | PtnElem *curPtr=nodePtr; | |
2156 | ||
2157 | if (patternMap==NULL) { | |
2158 | return FALSE; | |
2159 | } | |
2160 | while ( headIndex < MAX_PATTERN_ENTRIES ) { | |
2161 | if ( curPtr != NULL ) { | |
2162 | if ( curPtr->next != NULL ) { | |
2163 | return TRUE; | |
2164 | } | |
2165 | else { | |
2166 | headIndex++; | |
2167 | curPtr=NULL; | |
2168 | continue; | |
2169 | } | |
2170 | } | |
2171 | else { | |
2172 | if ( patternMap->boot[headIndex] != NULL ) { | |
2173 | return TRUE; | |
2174 | } | |
2175 | else { | |
2176 | headIndex++; | |
2177 | continue; | |
2178 | } | |
2179 | } | |
2180 | ||
2181 | } | |
2182 | return FALSE; | |
2183 | } | |
2184 | ||
2185 | DateTimeMatcher& | |
2186 | PatternMapIterator::next() { | |
2187 | while ( bootIndex < MAX_PATTERN_ENTRIES ) { | |
2188 | if ( nodePtr != NULL ) { | |
2189 | if ( nodePtr->next != NULL ) { | |
2190 | nodePtr = nodePtr->next; | |
2191 | break; | |
2192 | } | |
2193 | else { | |
2194 | bootIndex++; | |
2195 | nodePtr=NULL; | |
2196 | continue; | |
2197 | } | |
2198 | } | |
2199 | else { | |
2200 | if ( patternMap->boot[bootIndex] != NULL ) { | |
2201 | nodePtr = patternMap->boot[bootIndex]; | |
2202 | break; | |
2203 | } | |
2204 | else { | |
2205 | bootIndex++; | |
2206 | continue; | |
2207 | } | |
2208 | } | |
2209 | } | |
2210 | if (nodePtr!=NULL) { | |
2211 | matcher->copyFrom(*nodePtr->skeleton); | |
2212 | } | |
2213 | else { | |
2214 | matcher->copyFrom(); | |
2215 | } | |
2216 | return *matcher; | |
2217 | } | |
2218 | ||
f3c0d7a5 A |
2219 | |
2220 | SkeletonFields::SkeletonFields() { | |
2221 | // Set initial values to zero | |
2222 | clear(); | |
46f4442e A |
2223 | } |
2224 | ||
f3c0d7a5 A |
2225 | void SkeletonFields::clear() { |
2226 | uprv_memset(chars, 0, sizeof(chars)); | |
2227 | uprv_memset(lengths, 0, sizeof(lengths)); | |
2228 | } | |
46f4442e | 2229 | |
f3c0d7a5 A |
2230 | void SkeletonFields::copyFrom(const SkeletonFields& other) { |
2231 | uprv_memcpy(chars, other.chars, sizeof(chars)); | |
2232 | uprv_memcpy(lengths, other.lengths, sizeof(lengths)); | |
2233 | } | |
2234 | ||
2235 | void SkeletonFields::clearField(int32_t field) { | |
2236 | chars[field] = 0; | |
2237 | lengths[field] = 0; | |
2238 | } | |
2239 | ||
2240 | UChar SkeletonFields::getFieldChar(int32_t field) const { | |
2241 | return chars[field]; | |
2242 | } | |
2243 | ||
2244 | int32_t SkeletonFields::getFieldLength(int32_t field) const { | |
2245 | return lengths[field]; | |
2246 | } | |
2247 | ||
2248 | void SkeletonFields::populate(int32_t field, const UnicodeString& value) { | |
2249 | populate(field, value.charAt(0), value.length()); | |
2250 | } | |
2251 | ||
2252 | void SkeletonFields::populate(int32_t field, UChar ch, int32_t length) { | |
2253 | chars[field] = (int8_t) ch; | |
2254 | lengths[field] = (int8_t) length; | |
2255 | } | |
2256 | ||
2257 | UBool SkeletonFields::isFieldEmpty(int32_t field) const { | |
2258 | return lengths[field] == 0; | |
2259 | } | |
2260 | ||
2261 | UnicodeString& SkeletonFields::appendTo(UnicodeString& string) const { | |
2262 | for (int32_t i = 0; i < UDATPG_FIELD_COUNT; ++i) { | |
2263 | appendFieldTo(i, string); | |
46f4442e | 2264 | } |
f3c0d7a5 | 2265 | return string; |
46f4442e A |
2266 | } |
2267 | ||
f3c0d7a5 A |
2268 | UnicodeString& SkeletonFields::appendFieldTo(int32_t field, UnicodeString& string) const { |
2269 | UChar ch(chars[field]); | |
2270 | int32_t length = (int32_t) lengths[field]; | |
2271 | ||
2272 | for (int32_t i=0; i<length; i++) { | |
2273 | string += ch; | |
2274 | } | |
2275 | return string; | |
2276 | } | |
2277 | ||
2278 | UChar SkeletonFields::getFirstChar() const { | |
2279 | for (int32_t i = 0; i < UDATPG_FIELD_COUNT; ++i) { | |
2280 | if (lengths[i] != 0) { | |
2281 | return chars[i]; | |
46f4442e A |
2282 | } |
2283 | } | |
f3c0d7a5 A |
2284 | return '\0'; |
2285 | } | |
2286 | ||
2287 | ||
2288 | PtnSkeleton::PtnSkeleton() { | |
2289 | } | |
2290 | ||
2291 | PtnSkeleton::PtnSkeleton(const PtnSkeleton& other) { | |
2292 | copyFrom(other); | |
2293 | } | |
2294 | ||
2295 | void PtnSkeleton::copyFrom(const PtnSkeleton& other) { | |
2296 | uprv_memcpy(type, other.type, sizeof(type)); | |
2297 | original.copyFrom(other.original); | |
2298 | baseOriginal.copyFrom(other.baseOriginal); | |
2299 | } | |
2300 | ||
2301 | void PtnSkeleton::clear() { | |
2302 | uprv_memset(type, 0, sizeof(type)); | |
2303 | original.clear(); | |
2304 | baseOriginal.clear(); | |
2305 | } | |
2306 | ||
2307 | UBool | |
2308 | PtnSkeleton::equals(const PtnSkeleton& other) const { | |
2309 | return (original == other.original) | |
2310 | && (baseOriginal == other.baseOriginal) | |
2311 | && (uprv_memcmp(type, other.type, sizeof(type)) == 0); | |
46f4442e A |
2312 | } |
2313 | ||
2314 | UnicodeString | |
f3c0d7a5 | 2315 | PtnSkeleton::getSkeleton() const { |
46f4442e | 2316 | UnicodeString result; |
f3c0d7a5 | 2317 | return original.appendTo(result); |
46f4442e A |
2318 | } |
2319 | ||
2320 | UnicodeString | |
f3c0d7a5 | 2321 | PtnSkeleton::getBaseSkeleton() const { |
46f4442e | 2322 | UnicodeString result; |
f3c0d7a5 A |
2323 | return baseOriginal.appendTo(result); |
2324 | } | |
46f4442e | 2325 | |
f3c0d7a5 A |
2326 | UChar |
2327 | PtnSkeleton::getFirstChar() const { | |
2328 | return baseOriginal.getFirstChar(); | |
46f4442e A |
2329 | } |
2330 | ||
2331 | PtnSkeleton::~PtnSkeleton() { | |
2332 | } | |
2333 | ||
729e4ab9 | 2334 | PtnElem::PtnElem(const UnicodeString &basePat, const UnicodeString &pat) : |
46f4442e A |
2335 | basePattern(basePat), |
2336 | skeleton(NULL), | |
2337 | pattern(pat), | |
2338 | next(NULL) | |
2339 | { | |
2340 | } | |
2341 | ||
2342 | PtnElem::~PtnElem() { | |
2343 | ||
2344 | if (next!=NULL) { | |
2345 | delete next; | |
2346 | } | |
2347 | delete skeleton; | |
2348 | } | |
2349 | ||
2350 | DTSkeletonEnumeration::DTSkeletonEnumeration(PatternMap &patternMap, dtStrEnum type, UErrorCode& status) { | |
2351 | PtnElem *curElem; | |
2352 | PtnSkeleton *curSkeleton; | |
2353 | UnicodeString s; | |
2354 | int32_t bootIndex; | |
2355 | ||
2356 | pos=0; | |
2357 | fSkeletons = new UVector(status); | |
2358 | if (U_FAILURE(status)) { | |
2359 | delete fSkeletons; | |
2360 | return; | |
2361 | } | |
2362 | for (bootIndex=0; bootIndex<MAX_PATTERN_ENTRIES; ++bootIndex ) { | |
2363 | curElem = patternMap.boot[bootIndex]; | |
2364 | while (curElem!=NULL) { | |
2365 | switch(type) { | |
2366 | case DT_BASESKELETON: | |
2367 | s=curElem->basePattern; | |
2368 | break; | |
2369 | case DT_PATTERN: | |
2370 | s=curElem->pattern; | |
2371 | break; | |
2372 | case DT_SKELETON: | |
2373 | curSkeleton=curElem->skeleton; | |
2374 | s=curSkeleton->getSkeleton(); | |
2375 | break; | |
2376 | } | |
2377 | if ( !isCanonicalItem(s) ) { | |
2378 | fSkeletons->addElement(new UnicodeString(s), status); | |
2379 | if (U_FAILURE(status)) { | |
2380 | delete fSkeletons; | |
2381 | fSkeletons = NULL; | |
2382 | return; | |
2383 | } | |
2384 | } | |
2385 | curElem = curElem->next; | |
2386 | } | |
2387 | } | |
2388 | if ((bootIndex==MAX_PATTERN_ENTRIES) && (curElem!=NULL) ) { | |
2389 | status = U_BUFFER_OVERFLOW_ERROR; | |
2390 | } | |
2391 | } | |
2392 | ||
2393 | const UnicodeString* | |
2394 | DTSkeletonEnumeration::snext(UErrorCode& status) { | |
2395 | if (U_SUCCESS(status) && pos < fSkeletons->size()) { | |
2396 | return (const UnicodeString*)fSkeletons->elementAt(pos++); | |
2397 | } | |
2398 | return NULL; | |
2399 | } | |
2400 | ||
2401 | void | |
2402 | DTSkeletonEnumeration::reset(UErrorCode& /*status*/) { | |
2403 | pos=0; | |
2404 | } | |
2405 | ||
2406 | int32_t | |
2407 | DTSkeletonEnumeration::count(UErrorCode& /*status*/) const { | |
2408 | return (fSkeletons==NULL) ? 0 : fSkeletons->size(); | |
2409 | } | |
2410 | ||
2411 | UBool | |
2412 | DTSkeletonEnumeration::isCanonicalItem(const UnicodeString& item) { | |
2413 | if ( item.length() != 1 ) { | |
2414 | return FALSE; | |
2415 | } | |
2416 | for (int32_t i=0; i<UDATPG_FIELD_COUNT; ++i) { | |
2417 | if (item.charAt(0)==Canonical_Items[i]) { | |
2418 | return TRUE; | |
2419 | } | |
2420 | } | |
2421 | return FALSE; | |
2422 | } | |
2423 | ||
2424 | DTSkeletonEnumeration::~DTSkeletonEnumeration() { | |
2425 | UnicodeString *s; | |
2426 | for (int32_t i=0; i<fSkeletons->size(); ++i) { | |
2427 | if ((s=(UnicodeString *)fSkeletons->elementAt(i))!=NULL) { | |
2428 | delete s; | |
2429 | } | |
2430 | } | |
2431 | delete fSkeletons; | |
2432 | } | |
2433 | ||
2434 | DTRedundantEnumeration::DTRedundantEnumeration() { | |
2435 | pos=0; | |
2436 | fPatterns = NULL; | |
2437 | } | |
2438 | ||
2439 | void | |
2440 | DTRedundantEnumeration::add(const UnicodeString& pattern, UErrorCode& status) { | |
2441 | if (U_FAILURE(status)) return; | |
2442 | if (fPatterns == NULL) { | |
2443 | fPatterns = new UVector(status); | |
2444 | if (U_FAILURE(status)) { | |
2445 | delete fPatterns; | |
2446 | fPatterns = NULL; | |
2447 | return; | |
2448 | } | |
2449 | } | |
2450 | fPatterns->addElement(new UnicodeString(pattern), status); | |
2451 | if (U_FAILURE(status)) { | |
2452 | delete fPatterns; | |
2453 | fPatterns = NULL; | |
2454 | return; | |
2455 | } | |
2456 | } | |
2457 | ||
2458 | const UnicodeString* | |
2459 | DTRedundantEnumeration::snext(UErrorCode& status) { | |
2460 | if (U_SUCCESS(status) && pos < fPatterns->size()) { | |
2461 | return (const UnicodeString*)fPatterns->elementAt(pos++); | |
2462 | } | |
2463 | return NULL; | |
2464 | } | |
2465 | ||
2466 | void | |
2467 | DTRedundantEnumeration::reset(UErrorCode& /*status*/) { | |
2468 | pos=0; | |
2469 | } | |
2470 | ||
2471 | int32_t | |
2472 | DTRedundantEnumeration::count(UErrorCode& /*status*/) const { | |
2473 | return (fPatterns==NULL) ? 0 : fPatterns->size(); | |
2474 | } | |
2475 | ||
2476 | UBool | |
2477 | DTRedundantEnumeration::isCanonicalItem(const UnicodeString& item) { | |
2478 | if ( item.length() != 1 ) { | |
2479 | return FALSE; | |
2480 | } | |
2481 | for (int32_t i=0; i<UDATPG_FIELD_COUNT; ++i) { | |
2482 | if (item.charAt(0)==Canonical_Items[i]) { | |
2483 | return TRUE; | |
2484 | } | |
2485 | } | |
2486 | return FALSE; | |
2487 | } | |
2488 | ||
2489 | DTRedundantEnumeration::~DTRedundantEnumeration() { | |
2490 | UnicodeString *s; | |
2491 | for (int32_t i=0; i<fPatterns->size(); ++i) { | |
2492 | if ((s=(UnicodeString *)fPatterns->elementAt(i))!=NULL) { | |
2493 | delete s; | |
2494 | } | |
2495 | } | |
2496 | delete fPatterns; | |
2497 | } | |
2498 | ||
2499 | U_NAMESPACE_END | |
2500 | ||
2501 | ||
2502 | #endif /* #if !UCONFIG_NO_FORMATTING */ | |
2503 | ||
2504 | //eof |