1 | // © 2016 and later: Unicode, Inc. and others. |
2 | // License & terms of use: http://www.unicode.org/copyright.html |
3 | /* |
4 | ******************************************************************************* |
5 | * |
6 | * Copyright (C) 1997-2016, International Business Machines |
7 | * Corporation and others. All Rights Reserved. |
8 | * |
9 | ******************************************************************************* |
10 | * file name: locdispnames.cpp |
11 | * encoding: UTF-8 |
12 | * tab size: 8 (not used) |
13 | * indentation:4 |
14 | * |
15 | * created on: 2010feb25 |
16 | * created by: Markus W. Scherer |
17 | * |
18 | * Code for locale display names, separated out from other .cpp files |
19 | * that then do not depend on resource bundle code and display name data. |
20 | */ |
21 | |
22 | #include "unicode/utypes.h" |
23 | #include "unicode/brkiter.h" |
24 | #include "unicode/locid.h" |
25 | #include "unicode/uenum.h" |
26 | #include "unicode/uloc.h" |
27 | #include "unicode/ures.h" |
28 | #include "unicode/ustring.h" |
29 | #include "bytesinkutil.h" |
30 | #include "charstr.h" |
31 | #include "cmemory.h" |
32 | #include "cstring.h" |
33 | #include "putilimp.h" |
34 | #include "ulocimp.h" |
35 | #include "uresimp.h" |
36 | #include "ureslocs.h" |
37 | #include "ustr_imp.h" |
38 | |
39 | // C++ API ----------------------------------------------------------------- *** |
40 | |
41 | U_NAMESPACE_BEGIN |
42 | |
43 | UnicodeString& |
44 | Locale::getDisplayLanguage(UnicodeString& dispLang) const |
45 | { |
46 | return this->getDisplayLanguage(getDefault(), dispLang); |
47 | } |
48 | |
49 | /*We cannot make any assumptions on the size of the output display strings |
50 | * Yet, since we are calling through to a C API, we need to set limits on |
51 | * buffer size. For all the following getDisplay functions we first attempt |
52 | * to fill up a stack allocated buffer. If it is to small we heap allocated |
53 | * the exact buffer we need copy it to the UnicodeString and delete it*/ |
54 | |
55 | UnicodeString& |
56 | Locale::getDisplayLanguage(const Locale &displayLocale, |
57 | UnicodeString &result) const { |
58 | char16_t *buffer; |
59 | UErrorCode errorCode=U_ZERO_ERROR; |
60 | int32_t length; |
61 | |
62 | buffer=result.getBuffer(ULOC_FULLNAME_CAPACITY); |
63 | if(buffer==0) { |
64 | result.truncate(0); |
65 | return result; |
66 | } |
67 | |
68 | length=uloc_getDisplayLanguage(fullName, displayLocale.fullName, |
69 | buffer, result.getCapacity(), |
70 | &errorCode); |
71 | result.releaseBuffer(U_SUCCESS(errorCode) ? length : 0); |
72 | |
73 | if(errorCode==U_BUFFER_OVERFLOW_ERROR) { |
74 | buffer=result.getBuffer(length); |
75 | if(buffer==0) { |
76 | result.truncate(0); |
77 | return result; |
78 | } |
79 | errorCode=U_ZERO_ERROR; |
80 | length=uloc_getDisplayLanguage(fullName, displayLocale.fullName, |
81 | buffer, result.getCapacity(), |
82 | &errorCode); |
83 | result.releaseBuffer(U_SUCCESS(errorCode) ? length : 0); |
84 | } |
85 | |
86 | return result; |
87 | } |
88 | |
89 | UnicodeString& |
90 | Locale::getDisplayScript(UnicodeString& dispScript) const |
91 | { |
92 | return this->getDisplayScript(getDefault(), dispScript); |
93 | } |
94 | |
95 | UnicodeString& |
96 | Locale::getDisplayScript(const Locale &displayLocale, |
97 | UnicodeString &result) const { |
98 | char16_t *buffer; |
99 | UErrorCode errorCode=U_ZERO_ERROR; |
100 | int32_t length; |
101 | |
102 | buffer=result.getBuffer(ULOC_FULLNAME_CAPACITY); |
103 | if(buffer==0) { |
104 | result.truncate(0); |
105 | return result; |
106 | } |
107 | |
108 | length=uloc_getDisplayScript(fullName, displayLocale.fullName, |
109 | buffer, result.getCapacity(), |
110 | &errorCode); |
111 | result.releaseBuffer(U_SUCCESS(errorCode) ? length : 0); |
112 | |
113 | if(errorCode==U_BUFFER_OVERFLOW_ERROR) { |
114 | buffer=result.getBuffer(length); |
115 | if(buffer==0) { |
116 | result.truncate(0); |
117 | return result; |
118 | } |
119 | errorCode=U_ZERO_ERROR; |
120 | length=uloc_getDisplayScript(fullName, displayLocale.fullName, |
121 | buffer, result.getCapacity(), |
122 | &errorCode); |
123 | result.releaseBuffer(U_SUCCESS(errorCode) ? length : 0); |
124 | } |
125 | |
126 | return result; |
127 | } |
128 | |
129 | UnicodeString& |
130 | Locale::getDisplayCountry(UnicodeString& dispCntry) const |
131 | { |
132 | return this->getDisplayCountry(getDefault(), dispCntry); |
133 | } |
134 | |
135 | UnicodeString& |
136 | Locale::getDisplayCountry(const Locale &displayLocale, |
137 | UnicodeString &result) const { |
138 | char16_t *buffer; |
139 | UErrorCode errorCode=U_ZERO_ERROR; |
140 | int32_t length; |
141 | |
142 | buffer=result.getBuffer(ULOC_FULLNAME_CAPACITY); |
143 | if(buffer==0) { |
144 | result.truncate(0); |
145 | return result; |
146 | } |
147 | |
148 | length=uloc_getDisplayCountry(fullName, displayLocale.fullName, |
149 | buffer, result.getCapacity(), |
150 | &errorCode); |
151 | result.releaseBuffer(U_SUCCESS(errorCode) ? length : 0); |
152 | |
153 | if(errorCode==U_BUFFER_OVERFLOW_ERROR) { |
154 | buffer=result.getBuffer(length); |
155 | if(buffer==0) { |
156 | result.truncate(0); |
157 | return result; |
158 | } |
159 | errorCode=U_ZERO_ERROR; |
160 | length=uloc_getDisplayCountry(fullName, displayLocale.fullName, |
161 | buffer, result.getCapacity(), |
162 | &errorCode); |
163 | result.releaseBuffer(U_SUCCESS(errorCode) ? length : 0); |
164 | } |
165 | |
166 | return result; |
167 | } |
168 | |
169 | UnicodeString& |
170 | Locale::getDisplayVariant(UnicodeString& dispVar) const |
171 | { |
172 | return this->getDisplayVariant(getDefault(), dispVar); |
173 | } |
174 | |
175 | UnicodeString& |
176 | Locale::getDisplayVariant(const Locale &displayLocale, |
177 | UnicodeString &result) const { |
178 | char16_t *buffer; |
179 | UErrorCode errorCode=U_ZERO_ERROR; |
180 | int32_t length; |
181 | |
182 | buffer=result.getBuffer(ULOC_FULLNAME_CAPACITY); |
183 | if(buffer==0) { |
184 | result.truncate(0); |
185 | return result; |
186 | } |
187 | |
188 | length=uloc_getDisplayVariant(fullName, displayLocale.fullName, |
189 | buffer, result.getCapacity(), |
190 | &errorCode); |
191 | result.releaseBuffer(U_SUCCESS(errorCode) ? length : 0); |
192 | |
193 | if(errorCode==U_BUFFER_OVERFLOW_ERROR) { |
194 | buffer=result.getBuffer(length); |
195 | if(buffer==0) { |
196 | result.truncate(0); |
197 | return result; |
198 | } |
199 | errorCode=U_ZERO_ERROR; |
200 | length=uloc_getDisplayVariant(fullName, displayLocale.fullName, |
201 | buffer, result.getCapacity(), |
202 | &errorCode); |
203 | result.releaseBuffer(U_SUCCESS(errorCode) ? length : 0); |
204 | } |
205 | |
206 | return result; |
207 | } |
208 | |
209 | UnicodeString& |
210 | Locale::getDisplayName( UnicodeString& name ) const |
211 | { |
212 | return this->getDisplayName(getDefault(), name); |
213 | } |
214 | |
215 | UnicodeString& |
216 | Locale::getDisplayName(const Locale &displayLocale, |
217 | UnicodeString &result) const { |
218 | char16_t *buffer; |
219 | UErrorCode errorCode=U_ZERO_ERROR; |
220 | int32_t length; |
221 | |
222 | buffer=result.getBuffer(ULOC_FULLNAME_CAPACITY); |
223 | if(buffer==0) { |
224 | result.truncate(0); |
225 | return result; |
226 | } |
227 | |
228 | length=uloc_getDisplayName(fullName, displayLocale.fullName, |
229 | buffer, result.getCapacity(), |
230 | &errorCode); |
231 | result.releaseBuffer(U_SUCCESS(errorCode) ? length : 0); |
232 | |
233 | if(errorCode==U_BUFFER_OVERFLOW_ERROR) { |
234 | buffer=result.getBuffer(length); |
235 | if(buffer==0) { |
236 | result.truncate(0); |
237 | return result; |
238 | } |
239 | errorCode=U_ZERO_ERROR; |
240 | length=uloc_getDisplayName(fullName, displayLocale.fullName, |
241 | buffer, result.getCapacity(), |
242 | &errorCode); |
243 | result.releaseBuffer(U_SUCCESS(errorCode) ? length : 0); |
244 | } |
245 | |
246 | return result; |
247 | } |
248 | |
249 | #if ! UCONFIG_NO_BREAK_ITERATION |
250 | |
251 | // ------------------------------------- |
252 | // Gets the objectLocale display name in the default locale language. |
253 | UnicodeString& U_EXPORT2 |
254 | BreakIterator::getDisplayName(const Locale& objectLocale, |
255 | UnicodeString& name) |
256 | { |
257 | return objectLocale.getDisplayName(name); |
258 | } |
259 | |
260 | // ------------------------------------- |
261 | // Gets the objectLocale display name in the displayLocale language. |
262 | UnicodeString& U_EXPORT2 |
263 | BreakIterator::getDisplayName(const Locale& objectLocale, |
264 | const Locale& displayLocale, |
265 | UnicodeString& name) |
266 | { |
267 | return objectLocale.getDisplayName(displayLocale, name); |
268 | } |
269 | |
270 | #endif |
271 | |
272 | |
273 | U_NAMESPACE_END |
274 | |
275 | // C API ------------------------------------------------------------------- *** |
276 | |
277 | U_NAMESPACE_USE |
278 | |
279 | /* ### Constants **************************************************/ |
280 | |
281 | /* These strings describe the resources we attempt to load from |
282 | the locale ResourceBundle data file.*/ |
283 | static const char _kLanguages[] = "Languages" ; |
284 | static const char _kScripts[] = "Scripts" ; |
285 | static const char _kScriptsStandAlone[] = "Scripts%stand-alone" ; |
286 | static const char _kCountries[] = "Countries" ; |
287 | static const char _kVariants[] = "Variants" ; |
288 | static const char _kKeys[] = "Keys" ; |
289 | static const char _kTypes[] = "Types" ; |
290 | //static const char _kRootName[] = "root"; |
291 | static const char _kCurrency[] = "currency" ; |
292 | static const char _kCurrencies[] = "Currencies" ; |
293 | static const char _kLocaleDisplayPattern[] = "localeDisplayPattern" ; |
294 | static const char _kPattern[] = "pattern" ; |
295 | static const char _kSeparator[] = "separator" ; |
296 | |
297 | /* ### Display name **************************************************/ |
298 | |
299 | static int32_t |
300 | _getStringOrCopyKey(const char *path, const char *locale, |
301 | const char *tableKey, |
302 | const char* subTableKey, |
303 | const char *itemKey, |
304 | const char *substitute, |
305 | char16_t *dest, int32_t destCapacity, |
306 | UErrorCode *pErrorCode) { |
307 | const char16_t *s = nullptr; |
308 | int32_t length = 0; |
309 | |
310 | if(itemKey==nullptr) { |
311 | /* top-level item: normal resource bundle access */ |
312 | icu::LocalUResourceBundlePointer rb(ures_open(path, locale, pErrorCode)); |
313 | |
314 | if(U_SUCCESS(*pErrorCode)) { |
315 | s=ures_getStringByKey(rb.getAlias(), tableKey, &length, pErrorCode); |
316 | /* see comment about closing rb near "return item;" in _res_getTableStringWithFallback() */ |
317 | } |
318 | } else { |
319 | bool isLanguageCode = (uprv_strncmp(tableKey, _kLanguages, 9) == 0); |
320 | /* Language code should not be a number. If it is, set the error code. */ |
321 | if (isLanguageCode && uprv_strtol(itemKey, nullptr, 10)) { |
322 | *pErrorCode = U_MISSING_RESOURCE_ERROR; |
323 | } else { |
324 | /* second-level item, use special fallback */ |
325 | s=uloc_getTableStringWithFallback(path, locale, |
326 | tableKey, |
327 | subTableKey, |
328 | itemKey, |
329 | &length, |
330 | pErrorCode); |
331 | if (U_FAILURE(*pErrorCode) && isLanguageCode && itemKey != nullptr) { |
332 | // convert itemKey locale code to canonical form and try again, ICU-20870 |
333 | *pErrorCode = U_ZERO_ERROR; |
334 | Locale canonKey = Locale::createCanonical(itemKey); |
335 | s=uloc_getTableStringWithFallback(path, locale, |
336 | tableKey, |
337 | subTableKey, |
338 | canonKey.getName(), |
339 | &length, |
340 | pErrorCode); |
341 | } |
342 | } |
343 | } |
344 | |
345 | if(U_SUCCESS(*pErrorCode)) { |
346 | int32_t copyLength=uprv_min(length, destCapacity); |
347 | if(copyLength>0 && s != nullptr) { |
348 | u_memcpy(dest, s, copyLength); |
349 | } |
350 | } else { |
351 | /* no string from a resource bundle: convert the substitute */ |
352 | length=(int32_t)uprv_strlen(substitute); |
353 | u_charsToUChars(substitute, dest, uprv_min(length, destCapacity)); |
354 | *pErrorCode=U_USING_DEFAULT_WARNING; |
355 | } |
356 | |
357 | return u_terminateUChars(dest, destCapacity, length, pErrorCode); |
358 | } |
359 | |
360 | typedef int32_t U_CALLCONV UDisplayNameGetter(const char *, char *, int32_t, UErrorCode *); |
361 | |
362 | static int32_t |
363 | _getDisplayNameForComponent(const char *locale, |
364 | const char *displayLocale, |
365 | char16_t *dest, int32_t destCapacity, |
366 | UDisplayNameGetter *getter, |
367 | const char *tag, |
368 | UErrorCode *pErrorCode) { |
369 | char localeBuffer[ULOC_FULLNAME_CAPACITY*4]; |
370 | int32_t length; |
371 | UErrorCode localStatus; |
372 | const char* root = nullptr; |
373 | |
374 | /* argument checking */ |
375 | if(pErrorCode==nullptr || U_FAILURE(*pErrorCode)) { |
376 | return 0; |
377 | } |
378 | |
379 | if(destCapacity<0 || (destCapacity>0 && dest==nullptr)) { |
380 | *pErrorCode=U_ILLEGAL_ARGUMENT_ERROR; |
381 | return 0; |
382 | } |
383 | |
384 | localStatus = U_ZERO_ERROR; |
385 | length=(*getter)(locale, localeBuffer, sizeof(localeBuffer), &localStatus); |
386 | if(U_FAILURE(localStatus) || localStatus==U_STRING_NOT_TERMINATED_WARNING) { |
387 | *pErrorCode=U_ILLEGAL_ARGUMENT_ERROR; |
388 | return 0; |
389 | } |
390 | if(length==0) { |
391 | // For the display name, we treat this as unknown language (ICU-20273). |
392 | if (getter == uloc_getLanguage) { |
393 | uprv_strcpy(localeBuffer, "und" ); |
394 | } else { |
395 | return u_terminateUChars(dest, destCapacity, 0, pErrorCode); |
396 | } |
397 | } |
398 | |
399 | root = tag == _kCountries ? U_ICUDATA_REGION : U_ICUDATA_LANG; |
400 | |
401 | return _getStringOrCopyKey(root, displayLocale, |
402 | tag, nullptr, localeBuffer, |
403 | localeBuffer, |
404 | dest, destCapacity, |
405 | pErrorCode); |
406 | } |
407 | |
408 | U_CAPI int32_t U_EXPORT2 |
409 | uloc_getDisplayLanguage(const char *locale, |
410 | const char *displayLocale, |
411 | char16_t *dest, int32_t destCapacity, |
412 | UErrorCode *pErrorCode) { |
413 | return _getDisplayNameForComponent(locale, displayLocale, dest, destCapacity, |
414 | uloc_getLanguage, _kLanguages, pErrorCode); |
415 | } |
416 | |
417 | U_CAPI int32_t U_EXPORT2 |
418 | uloc_getDisplayScript(const char* locale, |
419 | const char* displayLocale, |
420 | char16_t *dest, int32_t destCapacity, |
421 | UErrorCode *pErrorCode) |
422 | { |
423 | UErrorCode err = U_ZERO_ERROR; |
424 | int32_t res = _getDisplayNameForComponent(locale, displayLocale, dest, destCapacity, |
425 | uloc_getScript, _kScriptsStandAlone, &err); |
426 | |
427 | if (destCapacity == 0 && err == U_BUFFER_OVERFLOW_ERROR) { |
428 | // For preflight, return the max of the value and the fallback. |
429 | int32_t fallback_res = _getDisplayNameForComponent(locale, displayLocale, dest, destCapacity, |
430 | uloc_getScript, _kScripts, pErrorCode); |
431 | return (fallback_res > res) ? fallback_res : res; |
432 | } |
433 | if ( err == U_USING_DEFAULT_WARNING ) { |
434 | return _getDisplayNameForComponent(locale, displayLocale, dest, destCapacity, |
435 | uloc_getScript, _kScripts, pErrorCode); |
436 | } else { |
437 | *pErrorCode = err; |
438 | return res; |
439 | } |
440 | } |
441 | |
442 | static int32_t |
443 | uloc_getDisplayScriptInContext(const char* locale, |
444 | const char* displayLocale, |
445 | char16_t *dest, int32_t destCapacity, |
446 | UErrorCode *pErrorCode) |
447 | { |
448 | return _getDisplayNameForComponent(locale, displayLocale, dest, destCapacity, |
449 | uloc_getScript, _kScripts, pErrorCode); |
450 | } |
451 | |
452 | U_CAPI int32_t U_EXPORT2 |
453 | uloc_getDisplayCountry(const char *locale, |
454 | const char *displayLocale, |
455 | char16_t *dest, int32_t destCapacity, |
456 | UErrorCode *pErrorCode) { |
457 | return _getDisplayNameForComponent(locale, displayLocale, dest, destCapacity, |
458 | uloc_getCountry, _kCountries, pErrorCode); |
459 | } |
460 | |
461 | /* |
462 | * TODO separate variant1_variant2_variant3... |
463 | * by getting each tag's display string and concatenating them with ", " |
464 | * in between - similar to uloc_getDisplayName() |
465 | */ |
466 | U_CAPI int32_t U_EXPORT2 |
467 | uloc_getDisplayVariant(const char *locale, |
468 | const char *displayLocale, |
469 | char16_t *dest, int32_t destCapacity, |
470 | UErrorCode *pErrorCode) { |
471 | return _getDisplayNameForComponent(locale, displayLocale, dest, destCapacity, |
472 | uloc_getVariant, _kVariants, pErrorCode); |
473 | } |
474 | |
475 | /* Instead of having a separate pass for 'special' patterns, reintegrate the two |
476 | * so we don't get bitten by preflight bugs again. We can be reasonably efficient |
477 | * without two separate code paths, this code isn't that performance-critical. |
478 | * |
479 | * This code is general enough to deal with patterns that have a prefix or swap the |
480 | * language and remainder components, since we gave developers enough rope to do such |
481 | * things if they futz with the pattern data. But since we don't give them a way to |
482 | * specify a pattern for arbitrary combinations of components, there's not much use in |
483 | * that. I don't think our data includes such patterns, the only variable I know if is |
484 | * whether there is a space before the open paren, or not. Oh, and zh uses different |
485 | * chars than the standard open/close paren (which ja and ko use, btw). |
486 | */ |
487 | U_CAPI int32_t U_EXPORT2 |
488 | uloc_getDisplayName(const char *locale, |
489 | const char *displayLocale, |
490 | char16_t *dest, int32_t destCapacity, |
491 | UErrorCode *pErrorCode) |
492 | { |
493 | static const char16_t defaultSeparator[9] = { 0x007b, 0x0030, 0x007d, 0x002c, 0x0020, 0x007b, 0x0031, 0x007d, 0x0000 }; /* "{0}, {1}" */ |
494 | static const char16_t sub0[4] = { 0x007b, 0x0030, 0x007d , 0x0000 } ; /* {0} */ |
495 | static const char16_t sub1[4] = { 0x007b, 0x0031, 0x007d , 0x0000 } ; /* {1} */ |
496 | static const int32_t subLen = 3; |
497 | static const char16_t defaultPattern[10] = { |
498 | 0x007b, 0x0030, 0x007d, 0x0020, 0x0028, 0x007b, 0x0031, 0x007d, 0x0029, 0x0000 |
499 | }; /* {0} ({1}) */ |
500 | static const int32_t defaultPatLen = 9; |
501 | static const int32_t defaultSub0Pos = 0; |
502 | static const int32_t defaultSub1Pos = 5; |
503 | |
504 | int32_t length; /* of formatted result */ |
505 | |
506 | const char16_t *separator; |
507 | int32_t sepLen = 0; |
508 | const char16_t *pattern; |
509 | int32_t patLen = 0; |
510 | int32_t sub0Pos, sub1Pos; |
511 | |
512 | char16_t formatOpenParen = 0x0028; // ( |
513 | char16_t formatReplaceOpenParen = 0x005B; // [ |
514 | char16_t formatCloseParen = 0x0029; // ) |
515 | char16_t formatReplaceCloseParen = 0x005D; // ] |
516 | |
517 | UBool haveLang = true; /* assume true, set false if we find we don't have |
518 | a lang component in the locale */ |
519 | UBool haveRest = true; /* assume true, set false if we find we don't have |
520 | any other component in the locale */ |
521 | UBool retry = false; /* set true if we need to retry, see below */ |
522 | |
523 | int32_t langi = 0; /* index of the language substitution (0 or 1), virtually always 0 */ |
524 | |
525 | if(pErrorCode==nullptr || U_FAILURE(*pErrorCode)) { |
526 | return 0; |
527 | } |
528 | |
529 | if(destCapacity<0 || (destCapacity>0 && dest==nullptr)) { |
530 | *pErrorCode=U_ILLEGAL_ARGUMENT_ERROR; |
531 | return 0; |
532 | } |
533 | |
534 | { |
535 | UErrorCode status = U_ZERO_ERROR; |
536 | |
537 | icu::LocalUResourceBundlePointer locbundle( |
538 | ures_open(U_ICUDATA_LANG, displayLocale, &status)); |
539 | icu::LocalUResourceBundlePointer dspbundle( |
540 | ures_getByKeyWithFallback(locbundle.getAlias(), _kLocaleDisplayPattern, nullptr, &status)); |
541 | |
542 | separator=ures_getStringByKeyWithFallback(dspbundle.getAlias(), _kSeparator, &sepLen, &status); |
543 | pattern=ures_getStringByKeyWithFallback(dspbundle.getAlias(), _kPattern, &patLen, &status); |
544 | } |
545 | |
546 | /* If we couldn't find any data, then use the defaults */ |
547 | if(sepLen == 0) { |
548 | separator = defaultSeparator; |
549 | } |
550 | /* #10244: Even though separator is now a pattern, it is awkward to handle it as such |
551 | * here since we are trying to build the display string in place in the dest buffer, |
552 | * and to handle it as a pattern would entail having separate storage for the |
553 | * substrings that need to be combined (the first of which may be the result of |
554 | * previous such combinations). So for now we continue to treat the portion between |
555 | * {0} and {1} as a string to be appended when joining substrings, ignoring anything |
556 | * that is before {0} or after {1} (no existing separator pattern has any such thing). |
557 | * This is similar to how pattern is handled below. |
558 | */ |
559 | { |
560 | char16_t *p0=u_strstr(separator, sub0); |
561 | char16_t *p1=u_strstr(separator, sub1); |
562 | if (p0==nullptr || p1==nullptr || p1<p0) { |
563 | *pErrorCode=U_ILLEGAL_ARGUMENT_ERROR; |
564 | return 0; |
565 | } |
566 | separator = (const char16_t *)p0 + subLen; |
567 | sepLen = static_cast<int32_t>(p1 - separator); |
568 | } |
569 | |
570 | if(patLen==0 || (patLen==defaultPatLen && !u_strncmp(pattern, defaultPattern, patLen))) { |
571 | pattern=defaultPattern; |
572 | patLen=defaultPatLen; |
573 | sub0Pos=defaultSub0Pos; |
574 | sub1Pos=defaultSub1Pos; |
575 | // use default formatOpenParen etc. set above |
576 | } else { /* non-default pattern */ |
577 | char16_t *p0=u_strstr(pattern, sub0); |
578 | char16_t *p1=u_strstr(pattern, sub1); |
579 | if (p0==nullptr || p1==nullptr) { |
580 | *pErrorCode=U_ILLEGAL_ARGUMENT_ERROR; |
581 | return 0; |
582 | } |
583 | sub0Pos = static_cast<int32_t>(p0-pattern); |
584 | sub1Pos = static_cast<int32_t>(p1-pattern); |
585 | if (sub1Pos < sub0Pos) { /* a very odd pattern */ |
586 | int32_t t=sub0Pos; sub0Pos=sub1Pos; sub1Pos=t; |
587 | langi=1; |
588 | } |
589 | if (u_strchr(pattern, 0xFF08) != nullptr) { |
590 | formatOpenParen = 0xFF08; // fullwidth ( |
591 | formatReplaceOpenParen = 0xFF3B; // fullwidth [ |
592 | formatCloseParen = 0xFF09; // fullwidth ) |
593 | formatReplaceCloseParen = 0xFF3D; // fullwidth ] |
594 | } |
595 | } |
596 | |
597 | /* We loop here because there is one case in which after the first pass we could need to |
598 | * reextract the data. If there's initial padding before the first element, we put in |
599 | * the padding and then write that element. If it turns out there's no second element, |
600 | * we didn't need the padding. If we do need the data (no preflight), and the first element |
601 | * would have fit but for the padding, we need to reextract. In this case (only) we |
602 | * adjust the parameters so padding is not added, and repeat. |
603 | */ |
604 | do { |
605 | char16_t* p=dest; |
606 | int32_t patPos=0; /* position in the pattern, used for non-substitution portions */ |
607 | int32_t langLen=0; /* length of language substitution */ |
608 | int32_t langPos=0; /* position in output of language substitution */ |
609 | int32_t restLen=0; /* length of 'everything else' substitution */ |
610 | int32_t restPos=0; /* position in output of 'everything else' substitution */ |
611 | icu::LocalUEnumerationPointer kenum; /* keyword enumeration */ |
612 | |
613 | /* prefix of pattern, extremely likely to be empty */ |
614 | if(sub0Pos) { |
615 | if(destCapacity >= sub0Pos) { |
616 | while (patPos < sub0Pos) { |
617 | *p++ = pattern[patPos++]; |
618 | } |
619 | } else { |
620 | patPos=sub0Pos; |
621 | } |
622 | length=sub0Pos; |
623 | } else { |
624 | length=0; |
625 | } |
626 | |
627 | for(int32_t subi=0,resti=0;subi<2;) { /* iterate through patterns 0 and 1*/ |
628 | UBool subdone = false; /* set true when ready to move to next substitution */ |
629 | |
630 | /* prep p and cap for calls to get display components, pin cap to 0 since |
631 | they complain if cap is negative */ |
632 | int32_t cap=destCapacity-length; |
633 | if (cap <= 0) { |
634 | cap=0; |
635 | } else { |
636 | p=dest+length; |
637 | } |
638 | |
639 | if (subi == langi) { /* {0}*/ |
640 | if(haveLang) { |
641 | langPos=length; |
642 | langLen=uloc_getDisplayLanguage(locale, displayLocale, p, cap, pErrorCode); |
643 | length+=langLen; |
644 | haveLang=langLen>0; |
645 | } |
646 | subdone=true; |
647 | } else { /* {1} */ |
648 | if(!haveRest) { |
649 | subdone=true; |
650 | } else { |
651 | int32_t len; /* length of component (plus other stuff) we just fetched */ |
652 | switch(resti++) { |
653 | case 0: |
654 | restPos=length; |
655 | len=uloc_getDisplayScriptInContext(locale, displayLocale, p, cap, pErrorCode); |
656 | break; |
657 | case 1: |
658 | len=uloc_getDisplayCountry(locale, displayLocale, p, cap, pErrorCode); |
659 | break; |
660 | case 2: |
661 | len=uloc_getDisplayVariant(locale, displayLocale, p, cap, pErrorCode); |
662 | break; |
663 | case 3: |
664 | kenum.adoptInstead(uloc_openKeywords(locale, pErrorCode)); |
665 | U_FALLTHROUGH; |
666 | default: { |
667 | const char* kw=uenum_next(kenum.getAlias(), &len, pErrorCode); |
668 | if (kw == nullptr) { |
669 | len=0; /* mark that we didn't add a component */ |
670 | subdone=true; |
671 | } else { |
672 | /* incorporating this behavior into the loop made it even more complex, |
673 | so just special case it here */ |
674 | len = uloc_getDisplayKeyword(kw, displayLocale, p, cap, pErrorCode); |
675 | if(len) { |
676 | if(len < cap) { |
677 | p[len]=0x3d; /* '=', assume we'll need it */ |
678 | } |
679 | len+=1; |
680 | |
681 | /* adjust for call to get keyword */ |
682 | cap-=len; |
683 | if(cap <= 0) { |
684 | cap=0; |
685 | } else { |
686 | p+=len; |
687 | } |
688 | } |
689 | /* reset for call below */ |
690 | if(*pErrorCode == U_BUFFER_OVERFLOW_ERROR) { |
691 | *pErrorCode=U_ZERO_ERROR; |
692 | } |
693 | int32_t vlen = uloc_getDisplayKeywordValue(locale, kw, displayLocale, |
694 | p, cap, pErrorCode); |
695 | if(len) { |
696 | if(vlen==0) { |
697 | --len; /* remove unneeded '=' */ |
698 | } |
699 | /* restore cap and p to what they were at start */ |
700 | cap=destCapacity-length; |
701 | if(cap <= 0) { |
702 | cap=0; |
703 | } else { |
704 | p=dest+length; |
705 | } |
706 | } |
707 | len+=vlen; /* total we added for key + '=' + value */ |
708 | } |
709 | } break; |
710 | } /* end switch */ |
711 | |
712 | if (len>0) { |
713 | /* we added a component, so add separator and write it if there's room. */ |
714 | if(len+sepLen<=cap) { |
715 | const char16_t * plimit = p + len; |
716 | for (; p < plimit; p++) { |
717 | if (*p == formatOpenParen) { |
718 | *p = formatReplaceOpenParen; |
719 | } else if (*p == formatCloseParen) { |
720 | *p = formatReplaceCloseParen; |
721 | } |
722 | } |
723 | for(int32_t i=0;i<sepLen;++i) { |
724 | *p++=separator[i]; |
725 | } |
726 | } |
727 | length+=len+sepLen; |
728 | } else if(subdone) { |
729 | /* remove separator if we added it */ |
730 | if (length!=restPos) { |
731 | length-=sepLen; |
732 | } |
733 | restLen=length-restPos; |
734 | haveRest=restLen>0; |
735 | } |
736 | } |
737 | } |
738 | |
739 | if(*pErrorCode == U_BUFFER_OVERFLOW_ERROR) { |
740 | *pErrorCode=U_ZERO_ERROR; |
741 | } |
742 | |
743 | if(subdone) { |
744 | if(haveLang && haveRest) { |
745 | /* append internal portion of pattern, the first time, |
746 | or last portion of pattern the second time */ |
747 | int32_t padLen; |
748 | patPos+=subLen; |
749 | padLen=(subi==0 ? sub1Pos : patLen)-patPos; |
750 | if(length+padLen <= destCapacity) { |
751 | p=dest+length; |
752 | for(int32_t i=0;i<padLen;++i) { |
753 | *p++=pattern[patPos++]; |
754 | } |
755 | } else { |
756 | patPos+=padLen; |
757 | } |
758 | length+=padLen; |
759 | } else if(subi==0) { |
760 | /* don't have first component, reset for second component */ |
761 | sub0Pos=0; |
762 | length=0; |
763 | } else if(length>0) { |
764 | /* true length is the length of just the component we got. */ |
765 | length=haveLang?langLen:restLen; |
766 | if(dest && sub0Pos!=0) { |
767 | if (sub0Pos+length<=destCapacity) { |
768 | /* first component not at start of result, |
769 | but we have full component in buffer. */ |
770 | u_memmove(dest, dest+(haveLang?langPos:restPos), length); |
771 | } else { |
772 | /* would have fit, but didn't because of pattern prefix. */ |
773 | sub0Pos=0; /* stops initial padding (and a second retry, |
774 | so we won't end up here again) */ |
775 | retry=true; |
776 | } |
777 | } |
778 | } |
779 | |
780 | ++subi; /* move on to next substitution */ |
781 | } |
782 | } |
783 | } while(retry); |
784 | |
785 | return u_terminateUChars(dest, destCapacity, length, pErrorCode); |
786 | } |
787 | |
788 | U_CAPI int32_t U_EXPORT2 |
789 | uloc_getDisplayKeyword(const char* keyword, |
790 | const char* displayLocale, |
791 | char16_t* dest, |
792 | int32_t destCapacity, |
793 | UErrorCode* status){ |
794 | |
795 | /* argument checking */ |
796 | if(status==nullptr || U_FAILURE(*status)) { |
797 | return 0; |
798 | } |
799 | |
800 | if(destCapacity<0 || (destCapacity>0 && dest==nullptr)) { |
801 | *status=U_ILLEGAL_ARGUMENT_ERROR; |
802 | return 0; |
803 | } |
804 | |
805 | |
806 | /* pass itemKey=nullptr to look for a top-level item */ |
807 | return _getStringOrCopyKey(U_ICUDATA_LANG, displayLocale, |
808 | _kKeys, nullptr, |
809 | keyword, |
810 | keyword, |
811 | dest, destCapacity, |
812 | status); |
813 | |
814 | } |
815 | |
816 | |
817 | #define UCURRENCY_DISPLAY_NAME_INDEX 1 |
818 | |
819 | U_CAPI int32_t U_EXPORT2 |
820 | uloc_getDisplayKeywordValue( const char* locale, |
821 | const char* keyword, |
822 | const char* displayLocale, |
823 | char16_t* dest, |
824 | int32_t destCapacity, |
825 | UErrorCode* status){ |
826 | |
827 | |
828 | /* argument checking */ |
829 | if(status==nullptr || U_FAILURE(*status)) { |
830 | return 0; |
831 | } |
832 | |
833 | if(destCapacity<0 || (destCapacity>0 && dest==nullptr)) { |
834 | *status=U_ILLEGAL_ARGUMENT_ERROR; |
835 | return 0; |
836 | } |
837 | |
838 | /* get the keyword value */ |
839 | CharString keywordValue; |
840 | { |
841 | CharStringByteSink sink(&keywordValue); |
842 | ulocimp_getKeywordValue(locale, keyword, sink, status); |
843 | } |
844 | |
845 | /* |
846 | * if the keyword is equal to currency .. then to get the display name |
847 | * we need to do the fallback ourselves |
848 | */ |
849 | if(uprv_stricmp(keyword, _kCurrency)==0){ |
850 | |
851 | int32_t dispNameLen = 0; |
852 | const char16_t *dispName = nullptr; |
853 | |
854 | icu::LocalUResourceBundlePointer bundle( |
855 | ures_open(U_ICUDATA_CURR, displayLocale, status)); |
856 | icu::LocalUResourceBundlePointer currencies( |
857 | ures_getByKey(bundle.getAlias(), _kCurrencies, nullptr, status)); |
858 | icu::LocalUResourceBundlePointer currency( |
859 | ures_getByKeyWithFallback(currencies.getAlias(), keywordValue.data(), nullptr, status)); |
860 | |
861 | dispName = ures_getStringByIndex(currency.getAlias(), UCURRENCY_DISPLAY_NAME_INDEX, &dispNameLen, status); |
862 | |
863 | if(U_FAILURE(*status)){ |
864 | if(*status == U_MISSING_RESOURCE_ERROR){ |
865 | /* we just want to write the value over if nothing is available */ |
866 | *status = U_USING_DEFAULT_WARNING; |
867 | }else{ |
868 | return 0; |
869 | } |
870 | } |
871 | |
872 | /* now copy the dispName over if not nullptr */ |
873 | if(dispName != nullptr){ |
874 | if(dispNameLen <= destCapacity){ |
875 | u_memcpy(dest, dispName, dispNameLen); |
876 | return u_terminateUChars(dest, destCapacity, dispNameLen, status); |
877 | }else{ |
878 | *status = U_BUFFER_OVERFLOW_ERROR; |
879 | return dispNameLen; |
880 | } |
881 | }else{ |
882 | /* we have not found the display name for the value .. just copy over */ |
883 | if(keywordValue.length() <= destCapacity){ |
884 | u_charsToUChars(keywordValue.data(), dest, keywordValue.length()); |
885 | return u_terminateUChars(dest, destCapacity, keywordValue.length(), status); |
886 | }else{ |
887 | *status = U_BUFFER_OVERFLOW_ERROR; |
888 | return keywordValue.length(); |
889 | } |
890 | } |
891 | |
892 | |
893 | }else{ |
894 | |
895 | return _getStringOrCopyKey(U_ICUDATA_LANG, displayLocale, |
896 | _kTypes, keyword, |
897 | keywordValue.data(), |
898 | keywordValue.data(), |
899 | dest, destCapacity, |
900 | status); |
901 | } |
902 | } |
903 | |