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: loclikely.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 likely and minimized locale subtags, separated out from other .cpp files |
19 | * that then do not depend on resource bundle code and likely-subtags data. |
20 | */ |
21 | |
22 | #include "unicode/bytestream.h" |
23 | #include "unicode/utypes.h" |
24 | #include "unicode/locid.h" |
25 | #include "unicode/putil.h" |
26 | #include "unicode/uchar.h" |
27 | #include "unicode/uloc.h" |
28 | #include "unicode/ures.h" |
29 | #include "unicode/uscript.h" |
30 | #include "bytesinkutil.h" |
31 | #include "charstr.h" |
32 | #include "cmemory.h" |
33 | #include "cstring.h" |
34 | #include "ulocimp.h" |
35 | #include "ustr_imp.h" |
36 | |
37 | /** |
38 | * These are the canonical strings for unknown languages, scripts and regions. |
39 | **/ |
40 | static const char* const unknownLanguage = "und" ; |
41 | static const char* const unknownScript = "Zzzz" ; |
42 | static const char* const unknownRegion = "ZZ" ; |
43 | |
44 | /** |
45 | * This function looks for the localeID in the likelySubtags resource. |
46 | * |
47 | * @param localeID The tag to find. |
48 | * @param buffer A buffer to hold the matching entry |
49 | * @param bufferLength The length of the output buffer |
50 | * @return A pointer to "buffer" if found, or a null pointer if not. |
51 | */ |
52 | static const char* U_CALLCONV |
53 | findLikelySubtags(const char* localeID, |
54 | char* buffer, |
55 | int32_t bufferLength, |
56 | UErrorCode* err) { |
57 | const char* result = nullptr; |
58 | |
59 | if (!U_FAILURE(*err)) { |
60 | int32_t resLen = 0; |
61 | const char16_t* s = nullptr; |
62 | UErrorCode tmpErr = U_ZERO_ERROR; |
63 | icu::LocalUResourceBundlePointer subtags(ures_openDirect(nullptr, "likelySubtags" , &tmpErr)); |
64 | if (U_SUCCESS(tmpErr)) { |
65 | icu::CharString und; |
66 | if (localeID != nullptr) { |
67 | if (*localeID == '\0') { |
68 | localeID = unknownLanguage; |
69 | } else if (*localeID == '_') { |
70 | und.append(unknownLanguage, *err); |
71 | und.append(localeID, *err); |
72 | if (U_FAILURE(*err)) { |
73 | return nullptr; |
74 | } |
75 | localeID = und.data(); |
76 | } |
77 | } |
78 | s = ures_getStringByKey(subtags.getAlias(), localeID, &resLen, &tmpErr); |
79 | |
80 | if (U_FAILURE(tmpErr)) { |
81 | /* |
82 | * If a resource is missing, it's not really an error, it's |
83 | * just that we don't have any data for that particular locale ID. |
84 | */ |
85 | if (tmpErr != U_MISSING_RESOURCE_ERROR) { |
86 | *err = tmpErr; |
87 | } |
88 | } |
89 | else if (resLen >= bufferLength) { |
90 | /* The buffer should never overflow. */ |
91 | *err = U_INTERNAL_PROGRAM_ERROR; |
92 | } |
93 | else { |
94 | u_UCharsToChars(s, buffer, resLen + 1); |
95 | if (resLen >= 3 && |
96 | uprv_strnicmp(buffer, unknownLanguage, 3) == 0 && |
97 | (resLen == 3 || buffer[3] == '_')) { |
98 | uprv_memmove(buffer, buffer + 3, resLen - 3 + 1); |
99 | } |
100 | result = buffer; |
101 | } |
102 | } else { |
103 | *err = tmpErr; |
104 | } |
105 | } |
106 | |
107 | return result; |
108 | } |
109 | |
110 | /** |
111 | * Append a tag to a buffer, adding the separator if necessary. The buffer |
112 | * must be large enough to contain the resulting tag plus any separator |
113 | * necessary. The tag must not be a zero-length string. |
114 | * |
115 | * @param tag The tag to add. |
116 | * @param tagLength The length of the tag. |
117 | * @param buffer The output buffer. |
118 | * @param bufferLength The length of the output buffer. This is an input/output parameter. |
119 | **/ |
120 | static void U_CALLCONV |
121 | appendTag( |
122 | const char* tag, |
123 | int32_t tagLength, |
124 | char* buffer, |
125 | int32_t* bufferLength, |
126 | UBool withSeparator) { |
127 | |
128 | if (withSeparator) { |
129 | buffer[*bufferLength] = '_'; |
130 | ++(*bufferLength); |
131 | } |
132 | |
133 | uprv_memmove( |
134 | &buffer[*bufferLength], |
135 | tag, |
136 | tagLength); |
137 | |
138 | *bufferLength += tagLength; |
139 | } |
140 | |
141 | /** |
142 | * Create a tag string from the supplied parameters. The lang, script and region |
143 | * parameters may be nullptr pointers. If they are, their corresponding length parameters |
144 | * must be less than or equal to 0. |
145 | * |
146 | * If any of the language, script or region parameters are empty, and the alternateTags |
147 | * parameter is not nullptr, it will be parsed for potential language, script and region tags |
148 | * to be used when constructing the new tag. If the alternateTags parameter is nullptr, or |
149 | * it contains no language tag, the default tag for the unknown language is used. |
150 | * |
151 | * If the length of the new string exceeds the capacity of the output buffer, |
152 | * the function copies as many bytes to the output buffer as it can, and returns |
153 | * the error U_BUFFER_OVERFLOW_ERROR. |
154 | * |
155 | * If an illegal argument is provided, the function returns the error |
156 | * U_ILLEGAL_ARGUMENT_ERROR. |
157 | * |
158 | * Note that this function can return the warning U_STRING_NOT_TERMINATED_WARNING if |
159 | * the tag string fits in the output buffer, but the null terminator doesn't. |
160 | * |
161 | * @param lang The language tag to use. |
162 | * @param langLength The length of the language tag. |
163 | * @param script The script tag to use. |
164 | * @param scriptLength The length of the script tag. |
165 | * @param region The region tag to use. |
166 | * @param regionLength The length of the region tag. |
167 | * @param trailing Any trailing data to append to the new tag. |
168 | * @param trailingLength The length of the trailing data. |
169 | * @param alternateTags A string containing any alternate tags. |
170 | * @param sink The output sink receiving the tag string. |
171 | * @param err A pointer to a UErrorCode for error reporting. |
172 | **/ |
173 | static void U_CALLCONV |
174 | createTagStringWithAlternates( |
175 | const char* lang, |
176 | int32_t langLength, |
177 | const char* script, |
178 | int32_t scriptLength, |
179 | const char* region, |
180 | int32_t regionLength, |
181 | const char* trailing, |
182 | int32_t trailingLength, |
183 | const char* alternateTags, |
184 | icu::ByteSink& sink, |
185 | UErrorCode* err) { |
186 | |
187 | if (U_FAILURE(*err)) { |
188 | goto error; |
189 | } |
190 | else if (langLength >= ULOC_LANG_CAPACITY || |
191 | scriptLength >= ULOC_SCRIPT_CAPACITY || |
192 | regionLength >= ULOC_COUNTRY_CAPACITY) { |
193 | goto error; |
194 | } |
195 | else { |
196 | /** |
197 | * ULOC_FULLNAME_CAPACITY will provide enough capacity |
198 | * that we can build a string that contains the language, |
199 | * script and region code without worrying about overrunning |
200 | * the user-supplied buffer. |
201 | **/ |
202 | char tagBuffer[ULOC_FULLNAME_CAPACITY]; |
203 | int32_t tagLength = 0; |
204 | UBool regionAppended = false; |
205 | |
206 | if (langLength > 0) { |
207 | appendTag( |
208 | lang, |
209 | langLength, |
210 | tagBuffer, |
211 | &tagLength, |
212 | /*withSeparator=*/false); |
213 | } |
214 | else if (alternateTags == nullptr) { |
215 | /* |
216 | * Use the empty string for an unknown language, if |
217 | * we found no language. |
218 | */ |
219 | } |
220 | else { |
221 | /* |
222 | * Parse the alternateTags string for the language. |
223 | */ |
224 | char alternateLang[ULOC_LANG_CAPACITY]; |
225 | int32_t alternateLangLength = sizeof(alternateLang); |
226 | |
227 | alternateLangLength = |
228 | uloc_getLanguage( |
229 | alternateTags, |
230 | alternateLang, |
231 | alternateLangLength, |
232 | err); |
233 | if(U_FAILURE(*err) || |
234 | alternateLangLength >= ULOC_LANG_CAPACITY) { |
235 | goto error; |
236 | } |
237 | else if (alternateLangLength == 0) { |
238 | /* |
239 | * Use the empty string for an unknown language, if |
240 | * we found no language. |
241 | */ |
242 | } |
243 | else { |
244 | appendTag( |
245 | alternateLang, |
246 | alternateLangLength, |
247 | tagBuffer, |
248 | &tagLength, |
249 | /*withSeparator=*/false); |
250 | } |
251 | } |
252 | |
253 | if (scriptLength > 0) { |
254 | appendTag( |
255 | script, |
256 | scriptLength, |
257 | tagBuffer, |
258 | &tagLength, |
259 | /*withSeparator=*/true); |
260 | } |
261 | else if (alternateTags != nullptr) { |
262 | /* |
263 | * Parse the alternateTags string for the script. |
264 | */ |
265 | char alternateScript[ULOC_SCRIPT_CAPACITY]; |
266 | |
267 | const int32_t alternateScriptLength = |
268 | uloc_getScript( |
269 | alternateTags, |
270 | alternateScript, |
271 | sizeof(alternateScript), |
272 | err); |
273 | |
274 | if (U_FAILURE(*err) || |
275 | alternateScriptLength >= ULOC_SCRIPT_CAPACITY) { |
276 | goto error; |
277 | } |
278 | else if (alternateScriptLength > 0) { |
279 | appendTag( |
280 | alternateScript, |
281 | alternateScriptLength, |
282 | tagBuffer, |
283 | &tagLength, |
284 | /*withSeparator=*/true); |
285 | } |
286 | } |
287 | |
288 | if (regionLength > 0) { |
289 | appendTag( |
290 | region, |
291 | regionLength, |
292 | tagBuffer, |
293 | &tagLength, |
294 | /*withSeparator=*/true); |
295 | |
296 | regionAppended = true; |
297 | } |
298 | else if (alternateTags != nullptr) { |
299 | /* |
300 | * Parse the alternateTags string for the region. |
301 | */ |
302 | char alternateRegion[ULOC_COUNTRY_CAPACITY]; |
303 | |
304 | const int32_t alternateRegionLength = |
305 | uloc_getCountry( |
306 | alternateTags, |
307 | alternateRegion, |
308 | sizeof(alternateRegion), |
309 | err); |
310 | if (U_FAILURE(*err) || |
311 | alternateRegionLength >= ULOC_COUNTRY_CAPACITY) { |
312 | goto error; |
313 | } |
314 | else if (alternateRegionLength > 0) { |
315 | appendTag( |
316 | alternateRegion, |
317 | alternateRegionLength, |
318 | tagBuffer, |
319 | &tagLength, |
320 | /*withSeparator=*/true); |
321 | |
322 | regionAppended = true; |
323 | } |
324 | } |
325 | |
326 | /** |
327 | * Copy the partial tag from our internal buffer to the supplied |
328 | * target. |
329 | **/ |
330 | sink.Append(tagBuffer, tagLength); |
331 | |
332 | if (trailingLength > 0) { |
333 | if (*trailing != '@') { |
334 | sink.Append("_" , 1); |
335 | if (!regionAppended) { |
336 | /* extra separator is required */ |
337 | sink.Append("_" , 1); |
338 | } |
339 | } |
340 | |
341 | /* |
342 | * Copy the trailing data into the supplied buffer. |
343 | */ |
344 | sink.Append(trailing, trailingLength); |
345 | } |
346 | |
347 | return; |
348 | } |
349 | |
350 | error: |
351 | |
352 | /** |
353 | * An overflow indicates the locale ID passed in |
354 | * is ill-formed. If we got here, and there was |
355 | * no previous error, it's an implicit overflow. |
356 | **/ |
357 | if (*err == U_BUFFER_OVERFLOW_ERROR || |
358 | U_SUCCESS(*err)) { |
359 | *err = U_ILLEGAL_ARGUMENT_ERROR; |
360 | } |
361 | } |
362 | |
363 | /** |
364 | * Create a tag string from the supplied parameters. The lang, script and region |
365 | * parameters may be nullptr pointers. If they are, their corresponding length parameters |
366 | * must be less than or equal to 0. If the lang parameter is an empty string, the |
367 | * default value for an unknown language is written to the output buffer. |
368 | * |
369 | * If the length of the new string exceeds the capacity of the output buffer, |
370 | * the function copies as many bytes to the output buffer as it can, and returns |
371 | * the error U_BUFFER_OVERFLOW_ERROR. |
372 | * |
373 | * If an illegal argument is provided, the function returns the error |
374 | * U_ILLEGAL_ARGUMENT_ERROR. |
375 | * |
376 | * @param lang The language tag to use. |
377 | * @param langLength The length of the language tag. |
378 | * @param script The script tag to use. |
379 | * @param scriptLength The length of the script tag. |
380 | * @param region The region tag to use. |
381 | * @param regionLength The length of the region tag. |
382 | * @param trailing Any trailing data to append to the new tag. |
383 | * @param trailingLength The length of the trailing data. |
384 | * @param sink The output sink receiving the tag string. |
385 | * @param err A pointer to a UErrorCode for error reporting. |
386 | **/ |
387 | static void U_CALLCONV |
388 | createTagString( |
389 | const char* lang, |
390 | int32_t langLength, |
391 | const char* script, |
392 | int32_t scriptLength, |
393 | const char* region, |
394 | int32_t regionLength, |
395 | const char* trailing, |
396 | int32_t trailingLength, |
397 | icu::ByteSink& sink, |
398 | UErrorCode* err) |
399 | { |
400 | createTagStringWithAlternates( |
401 | lang, |
402 | langLength, |
403 | script, |
404 | scriptLength, |
405 | region, |
406 | regionLength, |
407 | trailing, |
408 | trailingLength, |
409 | nullptr, |
410 | sink, |
411 | err); |
412 | } |
413 | |
414 | /** |
415 | * Parse the language, script, and region subtags from a tag string, and copy the |
416 | * results into the corresponding output parameters. The buffers are null-terminated, |
417 | * unless overflow occurs. |
418 | * |
419 | * The langLength, scriptLength, and regionLength parameters are input/output |
420 | * parameters, and must contain the capacity of their corresponding buffers on |
421 | * input. On output, they will contain the actual length of the buffers, not |
422 | * including the null terminator. |
423 | * |
424 | * If the length of any of the output subtags exceeds the capacity of the corresponding |
425 | * buffer, the function copies as many bytes to the output buffer as it can, and returns |
426 | * the error U_BUFFER_OVERFLOW_ERROR. It will not parse any more subtags once overflow |
427 | * occurs. |
428 | * |
429 | * If an illegal argument is provided, the function returns the error |
430 | * U_ILLEGAL_ARGUMENT_ERROR. |
431 | * |
432 | * @param localeID The locale ID to parse. |
433 | * @param lang The language tag buffer. |
434 | * @param langLength The length of the language tag. |
435 | * @param script The script tag buffer. |
436 | * @param scriptLength The length of the script tag. |
437 | * @param region The region tag buffer. |
438 | * @param regionLength The length of the region tag. |
439 | * @param err A pointer to a UErrorCode for error reporting. |
440 | * @return The number of chars of the localeID parameter consumed. |
441 | **/ |
442 | static int32_t U_CALLCONV |
443 | parseTagString( |
444 | const char* localeID, |
445 | char* lang, |
446 | int32_t* langLength, |
447 | char* script, |
448 | int32_t* scriptLength, |
449 | char* region, |
450 | int32_t* regionLength, |
451 | UErrorCode* err) |
452 | { |
453 | const char* position = localeID; |
454 | int32_t subtagLength = 0; |
455 | |
456 | if(U_FAILURE(*err) || |
457 | localeID == nullptr || |
458 | lang == nullptr || |
459 | langLength == nullptr || |
460 | script == nullptr || |
461 | scriptLength == nullptr || |
462 | region == nullptr || |
463 | regionLength == nullptr) { |
464 | goto error; |
465 | } |
466 | |
467 | subtagLength = ulocimp_getLanguage(position, &position, *err).extract(lang, *langLength, *err); |
468 | |
469 | /* |
470 | * Note that we explicit consider U_STRING_NOT_TERMINATED_WARNING |
471 | * to be an error, because it indicates the user-supplied tag is |
472 | * not well-formed. |
473 | */ |
474 | if(U_FAILURE(*err)) { |
475 | goto error; |
476 | } |
477 | |
478 | *langLength = subtagLength; |
479 | |
480 | /* |
481 | * If no language was present, use the empty string instead. |
482 | * Otherwise, move past any separator. |
483 | */ |
484 | if (_isIDSeparator(*position)) { |
485 | ++position; |
486 | } |
487 | |
488 | subtagLength = ulocimp_getScript(position, &position, *err).extract(script, *scriptLength, *err); |
489 | |
490 | if(U_FAILURE(*err)) { |
491 | goto error; |
492 | } |
493 | |
494 | *scriptLength = subtagLength; |
495 | |
496 | if (*scriptLength > 0) { |
497 | if (uprv_strnicmp(script, unknownScript, *scriptLength) == 0) { |
498 | /** |
499 | * If the script part is the "unknown" script, then don't return it. |
500 | **/ |
501 | *scriptLength = 0; |
502 | } |
503 | |
504 | /* |
505 | * Move past any separator. |
506 | */ |
507 | if (_isIDSeparator(*position)) { |
508 | ++position; |
509 | } |
510 | } |
511 | |
512 | subtagLength = ulocimp_getCountry(position, &position, *err).extract(region, *regionLength, *err); |
513 | |
514 | if(U_FAILURE(*err)) { |
515 | goto error; |
516 | } |
517 | |
518 | *regionLength = subtagLength; |
519 | |
520 | if (*regionLength > 0) { |
521 | if (uprv_strnicmp(region, unknownRegion, *regionLength) == 0) { |
522 | /** |
523 | * If the region part is the "unknown" region, then don't return it. |
524 | **/ |
525 | *regionLength = 0; |
526 | } |
527 | } else if (*position != 0 && *position != '@') { |
528 | /* back up over consumed trailing separator */ |
529 | --position; |
530 | } |
531 | |
532 | exit: |
533 | |
534 | return (int32_t)(position - localeID); |
535 | |
536 | error: |
537 | |
538 | /** |
539 | * If we get here, we have no explicit error, it's the result of an |
540 | * illegal argument. |
541 | **/ |
542 | if (!U_FAILURE(*err)) { |
543 | *err = U_ILLEGAL_ARGUMENT_ERROR; |
544 | } |
545 | |
546 | goto exit; |
547 | } |
548 | |
549 | static UBool U_CALLCONV |
550 | createLikelySubtagsString( |
551 | const char* lang, |
552 | int32_t langLength, |
553 | const char* script, |
554 | int32_t scriptLength, |
555 | const char* region, |
556 | int32_t regionLength, |
557 | const char* variants, |
558 | int32_t variantsLength, |
559 | icu::ByteSink& sink, |
560 | UErrorCode* err) { |
561 | /** |
562 | * ULOC_FULLNAME_CAPACITY will provide enough capacity |
563 | * that we can build a string that contains the language, |
564 | * script and region code without worrying about overrunning |
565 | * the user-supplied buffer. |
566 | **/ |
567 | char likelySubtagsBuffer[ULOC_FULLNAME_CAPACITY]; |
568 | |
569 | if(U_FAILURE(*err)) { |
570 | goto error; |
571 | } |
572 | |
573 | /** |
574 | * Try the language with the script and region first. |
575 | **/ |
576 | if (scriptLength > 0 && regionLength > 0) { |
577 | |
578 | const char* likelySubtags = nullptr; |
579 | |
580 | icu::CharString tagBuffer; |
581 | { |
582 | icu::CharStringByteSink sink(&tagBuffer); |
583 | createTagString( |
584 | lang, |
585 | langLength, |
586 | script, |
587 | scriptLength, |
588 | region, |
589 | regionLength, |
590 | nullptr, |
591 | 0, |
592 | sink, |
593 | err); |
594 | } |
595 | if(U_FAILURE(*err)) { |
596 | goto error; |
597 | } |
598 | |
599 | likelySubtags = |
600 | findLikelySubtags( |
601 | tagBuffer.data(), |
602 | likelySubtagsBuffer, |
603 | sizeof(likelySubtagsBuffer), |
604 | err); |
605 | if(U_FAILURE(*err)) { |
606 | goto error; |
607 | } |
608 | |
609 | if (likelySubtags != nullptr) { |
610 | /* Always use the language tag from the |
611 | maximal string, since it may be more |
612 | specific than the one provided. */ |
613 | createTagStringWithAlternates( |
614 | nullptr, |
615 | 0, |
616 | nullptr, |
617 | 0, |
618 | nullptr, |
619 | 0, |
620 | variants, |
621 | variantsLength, |
622 | likelySubtags, |
623 | sink, |
624 | err); |
625 | return true; |
626 | } |
627 | } |
628 | |
629 | /** |
630 | * Try the language with just the script. |
631 | **/ |
632 | if (scriptLength > 0) { |
633 | |
634 | const char* likelySubtags = nullptr; |
635 | |
636 | icu::CharString tagBuffer; |
637 | { |
638 | icu::CharStringByteSink sink(&tagBuffer); |
639 | createTagString( |
640 | lang, |
641 | langLength, |
642 | script, |
643 | scriptLength, |
644 | nullptr, |
645 | 0, |
646 | nullptr, |
647 | 0, |
648 | sink, |
649 | err); |
650 | } |
651 | if(U_FAILURE(*err)) { |
652 | goto error; |
653 | } |
654 | |
655 | likelySubtags = |
656 | findLikelySubtags( |
657 | tagBuffer.data(), |
658 | likelySubtagsBuffer, |
659 | sizeof(likelySubtagsBuffer), |
660 | err); |
661 | if(U_FAILURE(*err)) { |
662 | goto error; |
663 | } |
664 | |
665 | if (likelySubtags != nullptr) { |
666 | /* Always use the language tag from the |
667 | maximal string, since it may be more |
668 | specific than the one provided. */ |
669 | createTagStringWithAlternates( |
670 | nullptr, |
671 | 0, |
672 | nullptr, |
673 | 0, |
674 | region, |
675 | regionLength, |
676 | variants, |
677 | variantsLength, |
678 | likelySubtags, |
679 | sink, |
680 | err); |
681 | return true; |
682 | } |
683 | } |
684 | |
685 | /** |
686 | * Try the language with just the region. |
687 | **/ |
688 | if (regionLength > 0) { |
689 | |
690 | const char* likelySubtags = nullptr; |
691 | |
692 | icu::CharString tagBuffer; |
693 | { |
694 | icu::CharStringByteSink sink(&tagBuffer); |
695 | createTagString( |
696 | lang, |
697 | langLength, |
698 | nullptr, |
699 | 0, |
700 | region, |
701 | regionLength, |
702 | nullptr, |
703 | 0, |
704 | sink, |
705 | err); |
706 | } |
707 | if(U_FAILURE(*err)) { |
708 | goto error; |
709 | } |
710 | |
711 | likelySubtags = |
712 | findLikelySubtags( |
713 | tagBuffer.data(), |
714 | likelySubtagsBuffer, |
715 | sizeof(likelySubtagsBuffer), |
716 | err); |
717 | if(U_FAILURE(*err)) { |
718 | goto error; |
719 | } |
720 | |
721 | if (likelySubtags != nullptr) { |
722 | /* Always use the language tag from the |
723 | maximal string, since it may be more |
724 | specific than the one provided. */ |
725 | createTagStringWithAlternates( |
726 | nullptr, |
727 | 0, |
728 | script, |
729 | scriptLength, |
730 | nullptr, |
731 | 0, |
732 | variants, |
733 | variantsLength, |
734 | likelySubtags, |
735 | sink, |
736 | err); |
737 | return true; |
738 | } |
739 | } |
740 | |
741 | /** |
742 | * Finally, try just the language. |
743 | **/ |
744 | { |
745 | const char* likelySubtags = nullptr; |
746 | |
747 | icu::CharString tagBuffer; |
748 | { |
749 | icu::CharStringByteSink sink(&tagBuffer); |
750 | createTagString( |
751 | lang, |
752 | langLength, |
753 | nullptr, |
754 | 0, |
755 | nullptr, |
756 | 0, |
757 | nullptr, |
758 | 0, |
759 | sink, |
760 | err); |
761 | } |
762 | if(U_FAILURE(*err)) { |
763 | goto error; |
764 | } |
765 | |
766 | likelySubtags = |
767 | findLikelySubtags( |
768 | tagBuffer.data(), |
769 | likelySubtagsBuffer, |
770 | sizeof(likelySubtagsBuffer), |
771 | err); |
772 | if(U_FAILURE(*err)) { |
773 | goto error; |
774 | } |
775 | |
776 | if (likelySubtags != nullptr) { |
777 | /* Always use the language tag from the |
778 | maximal string, since it may be more |
779 | specific than the one provided. */ |
780 | createTagStringWithAlternates( |
781 | nullptr, |
782 | 0, |
783 | script, |
784 | scriptLength, |
785 | region, |
786 | regionLength, |
787 | variants, |
788 | variantsLength, |
789 | likelySubtags, |
790 | sink, |
791 | err); |
792 | return true; |
793 | } |
794 | } |
795 | |
796 | return false; |
797 | |
798 | error: |
799 | |
800 | if (!U_FAILURE(*err)) { |
801 | *err = U_ILLEGAL_ARGUMENT_ERROR; |
802 | } |
803 | |
804 | return false; |
805 | } |
806 | |
807 | #define CHECK_TRAILING_VARIANT_SIZE(trailing, trailingLength) UPRV_BLOCK_MACRO_BEGIN { \ |
808 | int32_t count = 0; \ |
809 | int32_t i; \ |
810 | for (i = 0; i < trailingLength; i++) { \ |
811 | if (trailing[i] == '-' || trailing[i] == '_') { \ |
812 | count = 0; \ |
813 | if (count > 8) { \ |
814 | goto error; \ |
815 | } \ |
816 | } else if (trailing[i] == '@') { \ |
817 | break; \ |
818 | } else if (count > 8) { \ |
819 | goto error; \ |
820 | } else { \ |
821 | count++; \ |
822 | } \ |
823 | } \ |
824 | } UPRV_BLOCK_MACRO_END |
825 | |
826 | static UBool |
827 | _uloc_addLikelySubtags(const char* localeID, |
828 | icu::ByteSink& sink, |
829 | UErrorCode* err) { |
830 | char lang[ULOC_LANG_CAPACITY]; |
831 | int32_t langLength = sizeof(lang); |
832 | char script[ULOC_SCRIPT_CAPACITY]; |
833 | int32_t scriptLength = sizeof(script); |
834 | char region[ULOC_COUNTRY_CAPACITY]; |
835 | int32_t regionLength = sizeof(region); |
836 | const char* trailing = "" ; |
837 | int32_t trailingLength = 0; |
838 | int32_t trailingIndex = 0; |
839 | UBool success = false; |
840 | |
841 | if(U_FAILURE(*err)) { |
842 | goto error; |
843 | } |
844 | if (localeID == nullptr) { |
845 | goto error; |
846 | } |
847 | |
848 | trailingIndex = parseTagString( |
849 | localeID, |
850 | lang, |
851 | &langLength, |
852 | script, |
853 | &scriptLength, |
854 | region, |
855 | ®ionLength, |
856 | err); |
857 | if(U_FAILURE(*err)) { |
858 | /* Overflow indicates an illegal argument error */ |
859 | if (*err == U_BUFFER_OVERFLOW_ERROR) { |
860 | *err = U_ILLEGAL_ARGUMENT_ERROR; |
861 | } |
862 | |
863 | goto error; |
864 | } |
865 | |
866 | /* Find the length of the trailing portion. */ |
867 | while (_isIDSeparator(localeID[trailingIndex])) { |
868 | trailingIndex++; |
869 | } |
870 | trailing = &localeID[trailingIndex]; |
871 | trailingLength = (int32_t)uprv_strlen(trailing); |
872 | |
873 | CHECK_TRAILING_VARIANT_SIZE(trailing, trailingLength); |
874 | |
875 | success = |
876 | createLikelySubtagsString( |
877 | lang, |
878 | langLength, |
879 | script, |
880 | scriptLength, |
881 | region, |
882 | regionLength, |
883 | trailing, |
884 | trailingLength, |
885 | sink, |
886 | err); |
887 | |
888 | if (!success) { |
889 | const int32_t localIDLength = (int32_t)uprv_strlen(localeID); |
890 | |
891 | /* |
892 | * If we get here, we need to return localeID. |
893 | */ |
894 | sink.Append(localeID, localIDLength); |
895 | } |
896 | |
897 | return success; |
898 | |
899 | error: |
900 | |
901 | if (!U_FAILURE(*err)) { |
902 | *err = U_ILLEGAL_ARGUMENT_ERROR; |
903 | } |
904 | return false; |
905 | } |
906 | |
907 | // Add likely subtags to the sink |
908 | // return true if the value in the sink is produced by a match during the lookup |
909 | // return false if the value in the sink is the same as input because there are |
910 | // no match after the lookup. |
911 | static UBool _ulocimp_addLikelySubtags(const char*, icu::ByteSink&, UErrorCode*); |
912 | |
913 | static void |
914 | _uloc_minimizeSubtags(const char* localeID, |
915 | icu::ByteSink& sink, |
916 | UErrorCode* err) { |
917 | icu::CharString maximizedTagBuffer; |
918 | |
919 | char lang[ULOC_LANG_CAPACITY]; |
920 | int32_t langLength = sizeof(lang); |
921 | char script[ULOC_SCRIPT_CAPACITY]; |
922 | int32_t scriptLength = sizeof(script); |
923 | char region[ULOC_COUNTRY_CAPACITY]; |
924 | int32_t regionLength = sizeof(region); |
925 | const char* trailing = "" ; |
926 | int32_t trailingLength = 0; |
927 | int32_t trailingIndex = 0; |
928 | UBool successGetMax = false; |
929 | |
930 | if(U_FAILURE(*err)) { |
931 | goto error; |
932 | } |
933 | else if (localeID == nullptr) { |
934 | goto error; |
935 | } |
936 | |
937 | trailingIndex = |
938 | parseTagString( |
939 | localeID, |
940 | lang, |
941 | &langLength, |
942 | script, |
943 | &scriptLength, |
944 | region, |
945 | ®ionLength, |
946 | err); |
947 | if(U_FAILURE(*err)) { |
948 | |
949 | /* Overflow indicates an illegal argument error */ |
950 | if (*err == U_BUFFER_OVERFLOW_ERROR) { |
951 | *err = U_ILLEGAL_ARGUMENT_ERROR; |
952 | } |
953 | |
954 | goto error; |
955 | } |
956 | |
957 | /* Find the spot where the variants or the keywords begin, if any. */ |
958 | while (_isIDSeparator(localeID[trailingIndex])) { |
959 | trailingIndex++; |
960 | } |
961 | trailing = &localeID[trailingIndex]; |
962 | trailingLength = (int32_t)uprv_strlen(trailing); |
963 | |
964 | CHECK_TRAILING_VARIANT_SIZE(trailing, trailingLength); |
965 | |
966 | { |
967 | icu::CharString base; |
968 | { |
969 | icu::CharStringByteSink baseSink(&base); |
970 | createTagString( |
971 | lang, |
972 | langLength, |
973 | script, |
974 | scriptLength, |
975 | region, |
976 | regionLength, |
977 | nullptr, |
978 | 0, |
979 | baseSink, |
980 | err); |
981 | } |
982 | |
983 | /** |
984 | * First, we need to first get the maximization |
985 | * from AddLikelySubtags. |
986 | **/ |
987 | { |
988 | icu::CharStringByteSink maxSink(&maximizedTagBuffer); |
989 | successGetMax = _ulocimp_addLikelySubtags(base.data(), maxSink, err); |
990 | } |
991 | } |
992 | |
993 | if(U_FAILURE(*err)) { |
994 | goto error; |
995 | } |
996 | |
997 | if (!successGetMax) { |
998 | /** |
999 | * If we got here, return the locale ID parameter unchanged. |
1000 | **/ |
1001 | const int32_t localeIDLength = (int32_t)uprv_strlen(localeID); |
1002 | sink.Append(localeID, localeIDLength); |
1003 | return; |
1004 | } |
1005 | |
1006 | // In the following, the lang, script, region are referring to those in |
1007 | // the maximizedTagBuffer, not the one in the localeID. |
1008 | langLength = sizeof(lang); |
1009 | scriptLength = sizeof(script); |
1010 | regionLength = sizeof(region); |
1011 | parseTagString( |
1012 | maximizedTagBuffer.data(), |
1013 | lang, |
1014 | &langLength, |
1015 | script, |
1016 | &scriptLength, |
1017 | region, |
1018 | ®ionLength, |
1019 | err); |
1020 | if(U_FAILURE(*err)) { |
1021 | goto error; |
1022 | } |
1023 | |
1024 | /** |
1025 | * Start first with just the language. |
1026 | **/ |
1027 | { |
1028 | icu::CharString tagBuffer; |
1029 | { |
1030 | icu::CharStringByteSink tagSink(&tagBuffer); |
1031 | createLikelySubtagsString( |
1032 | lang, |
1033 | langLength, |
1034 | nullptr, |
1035 | 0, |
1036 | nullptr, |
1037 | 0, |
1038 | nullptr, |
1039 | 0, |
1040 | tagSink, |
1041 | err); |
1042 | } |
1043 | |
1044 | if(U_FAILURE(*err)) { |
1045 | goto error; |
1046 | } |
1047 | else if (!tagBuffer.isEmpty() && |
1048 | uprv_strnicmp( |
1049 | maximizedTagBuffer.data(), |
1050 | tagBuffer.data(), |
1051 | tagBuffer.length()) == 0) { |
1052 | |
1053 | createTagString( |
1054 | lang, |
1055 | langLength, |
1056 | nullptr, |
1057 | 0, |
1058 | nullptr, |
1059 | 0, |
1060 | trailing, |
1061 | trailingLength, |
1062 | sink, |
1063 | err); |
1064 | return; |
1065 | } |
1066 | } |
1067 | |
1068 | /** |
1069 | * Next, try the language and region. |
1070 | **/ |
1071 | if (regionLength > 0) { |
1072 | |
1073 | icu::CharString tagBuffer; |
1074 | { |
1075 | icu::CharStringByteSink tagSink(&tagBuffer); |
1076 | createLikelySubtagsString( |
1077 | lang, |
1078 | langLength, |
1079 | nullptr, |
1080 | 0, |
1081 | region, |
1082 | regionLength, |
1083 | nullptr, |
1084 | 0, |
1085 | tagSink, |
1086 | err); |
1087 | } |
1088 | |
1089 | if(U_FAILURE(*err)) { |
1090 | goto error; |
1091 | } |
1092 | else if (!tagBuffer.isEmpty() && |
1093 | uprv_strnicmp( |
1094 | maximizedTagBuffer.data(), |
1095 | tagBuffer.data(), |
1096 | tagBuffer.length()) == 0) { |
1097 | |
1098 | createTagString( |
1099 | lang, |
1100 | langLength, |
1101 | nullptr, |
1102 | 0, |
1103 | region, |
1104 | regionLength, |
1105 | trailing, |
1106 | trailingLength, |
1107 | sink, |
1108 | err); |
1109 | return; |
1110 | } |
1111 | } |
1112 | |
1113 | /** |
1114 | * Finally, try the language and script. This is our last chance, |
1115 | * since trying with all three subtags would only yield the |
1116 | * maximal version that we already have. |
1117 | **/ |
1118 | if (scriptLength > 0) { |
1119 | icu::CharString tagBuffer; |
1120 | { |
1121 | icu::CharStringByteSink tagSink(&tagBuffer); |
1122 | createLikelySubtagsString( |
1123 | lang, |
1124 | langLength, |
1125 | script, |
1126 | scriptLength, |
1127 | nullptr, |
1128 | 0, |
1129 | nullptr, |
1130 | 0, |
1131 | tagSink, |
1132 | err); |
1133 | } |
1134 | |
1135 | if(U_FAILURE(*err)) { |
1136 | goto error; |
1137 | } |
1138 | else if (!tagBuffer.isEmpty() && |
1139 | uprv_strnicmp( |
1140 | maximizedTagBuffer.data(), |
1141 | tagBuffer.data(), |
1142 | tagBuffer.length()) == 0) { |
1143 | |
1144 | createTagString( |
1145 | lang, |
1146 | langLength, |
1147 | script, |
1148 | scriptLength, |
1149 | nullptr, |
1150 | 0, |
1151 | trailing, |
1152 | trailingLength, |
1153 | sink, |
1154 | err); |
1155 | return; |
1156 | } |
1157 | } |
1158 | |
1159 | { |
1160 | /** |
1161 | * If we got here, return the max + trail. |
1162 | **/ |
1163 | createTagString( |
1164 | lang, |
1165 | langLength, |
1166 | script, |
1167 | scriptLength, |
1168 | region, |
1169 | regionLength, |
1170 | trailing, |
1171 | trailingLength, |
1172 | sink, |
1173 | err); |
1174 | return; |
1175 | } |
1176 | |
1177 | error: |
1178 | |
1179 | if (!U_FAILURE(*err)) { |
1180 | *err = U_ILLEGAL_ARGUMENT_ERROR; |
1181 | } |
1182 | } |
1183 | |
1184 | static int32_t |
1185 | do_canonicalize(const char* localeID, |
1186 | char* buffer, |
1187 | int32_t bufferCapacity, |
1188 | UErrorCode* err) |
1189 | { |
1190 | int32_t canonicalizedSize = uloc_canonicalize( |
1191 | localeID, |
1192 | buffer, |
1193 | bufferCapacity, |
1194 | err); |
1195 | |
1196 | if (*err == U_STRING_NOT_TERMINATED_WARNING || |
1197 | *err == U_BUFFER_OVERFLOW_ERROR) { |
1198 | return canonicalizedSize; |
1199 | } |
1200 | else if (U_FAILURE(*err)) { |
1201 | |
1202 | return -1; |
1203 | } |
1204 | else { |
1205 | return canonicalizedSize; |
1206 | } |
1207 | } |
1208 | |
1209 | U_CAPI int32_t U_EXPORT2 |
1210 | uloc_addLikelySubtags(const char* localeID, |
1211 | char* maximizedLocaleID, |
1212 | int32_t maximizedLocaleIDCapacity, |
1213 | UErrorCode* status) { |
1214 | if (U_FAILURE(*status)) { |
1215 | return 0; |
1216 | } |
1217 | |
1218 | icu::CheckedArrayByteSink sink( |
1219 | maximizedLocaleID, maximizedLocaleIDCapacity); |
1220 | |
1221 | ulocimp_addLikelySubtags(localeID, sink, status); |
1222 | int32_t reslen = sink.NumberOfBytesAppended(); |
1223 | |
1224 | if (U_FAILURE(*status)) { |
1225 | return sink.Overflowed() ? reslen : -1; |
1226 | } |
1227 | |
1228 | if (sink.Overflowed()) { |
1229 | *status = U_BUFFER_OVERFLOW_ERROR; |
1230 | } else { |
1231 | u_terminateChars( |
1232 | maximizedLocaleID, maximizedLocaleIDCapacity, reslen, status); |
1233 | } |
1234 | |
1235 | return reslen; |
1236 | } |
1237 | |
1238 | static UBool |
1239 | _ulocimp_addLikelySubtags(const char* localeID, |
1240 | icu::ByteSink& sink, |
1241 | UErrorCode* status) { |
1242 | PreflightingLocaleIDBuffer localeBuffer; |
1243 | do { |
1244 | localeBuffer.requestedCapacity = do_canonicalize(localeID, localeBuffer.getBuffer(), |
1245 | localeBuffer.getCapacity(), status); |
1246 | } while (localeBuffer.needToTryAgain(status)); |
1247 | |
1248 | if (U_SUCCESS(*status)) { |
1249 | return _uloc_addLikelySubtags(localeBuffer.getBuffer(), sink, status); |
1250 | } else { |
1251 | return false; |
1252 | } |
1253 | } |
1254 | |
1255 | U_CAPI void U_EXPORT2 |
1256 | ulocimp_addLikelySubtags(const char* localeID, |
1257 | icu::ByteSink& sink, |
1258 | UErrorCode* status) { |
1259 | _ulocimp_addLikelySubtags(localeID, sink, status); |
1260 | } |
1261 | |
1262 | U_CAPI int32_t U_EXPORT2 |
1263 | uloc_minimizeSubtags(const char* localeID, |
1264 | char* minimizedLocaleID, |
1265 | int32_t minimizedLocaleIDCapacity, |
1266 | UErrorCode* status) { |
1267 | if (U_FAILURE(*status)) { |
1268 | return 0; |
1269 | } |
1270 | |
1271 | icu::CheckedArrayByteSink sink( |
1272 | minimizedLocaleID, minimizedLocaleIDCapacity); |
1273 | |
1274 | ulocimp_minimizeSubtags(localeID, sink, status); |
1275 | int32_t reslen = sink.NumberOfBytesAppended(); |
1276 | |
1277 | if (U_FAILURE(*status)) { |
1278 | return sink.Overflowed() ? reslen : -1; |
1279 | } |
1280 | |
1281 | if (sink.Overflowed()) { |
1282 | *status = U_BUFFER_OVERFLOW_ERROR; |
1283 | } else { |
1284 | u_terminateChars( |
1285 | minimizedLocaleID, minimizedLocaleIDCapacity, reslen, status); |
1286 | } |
1287 | |
1288 | return reslen; |
1289 | } |
1290 | |
1291 | U_CAPI void U_EXPORT2 |
1292 | ulocimp_minimizeSubtags(const char* localeID, |
1293 | icu::ByteSink& sink, |
1294 | UErrorCode* status) { |
1295 | PreflightingLocaleIDBuffer localeBuffer; |
1296 | do { |
1297 | localeBuffer.requestedCapacity = do_canonicalize(localeID, localeBuffer.getBuffer(), |
1298 | localeBuffer.getCapacity(), status); |
1299 | } while (localeBuffer.needToTryAgain(status)); |
1300 | |
1301 | _uloc_minimizeSubtags(localeBuffer.getBuffer(), sink, status); |
1302 | } |
1303 | |
1304 | // Pairs of (language subtag, + or -) for finding out fast if common languages |
1305 | // are LTR (minus) or RTL (plus). |
1306 | static const char LANG_DIR_STRING[] = |
1307 | "root-en-es-pt-zh-ja-ko-de-fr-it-ar+he+fa+ru-nl-pl-th-tr-" ; |
1308 | |
1309 | // Implemented here because this calls ulocimp_addLikelySubtags(). |
1310 | U_CAPI UBool U_EXPORT2 |
1311 | uloc_isRightToLeft(const char *locale) { |
1312 | UErrorCode errorCode = U_ZERO_ERROR; |
1313 | char script[8]; |
1314 | int32_t scriptLength = uloc_getScript(locale, script, UPRV_LENGTHOF(script), &errorCode); |
1315 | if (U_FAILURE(errorCode) || errorCode == U_STRING_NOT_TERMINATED_WARNING || |
1316 | scriptLength == 0) { |
1317 | // Fastpath: We know the likely scripts and their writing direction |
1318 | // for some common languages. |
1319 | errorCode = U_ZERO_ERROR; |
1320 | char lang[8]; |
1321 | int32_t langLength = uloc_getLanguage(locale, lang, UPRV_LENGTHOF(lang), &errorCode); |
1322 | if (U_FAILURE(errorCode) || errorCode == U_STRING_NOT_TERMINATED_WARNING) { |
1323 | return false; |
1324 | } |
1325 | if (langLength > 0) { |
1326 | const char* langPtr = uprv_strstr(LANG_DIR_STRING, lang); |
1327 | if (langPtr != nullptr) { |
1328 | switch (langPtr[langLength]) { |
1329 | case '-': return false; |
1330 | case '+': return true; |
1331 | default: break; // partial match of a longer code |
1332 | } |
1333 | } |
1334 | } |
1335 | // Otherwise, find the likely script. |
1336 | errorCode = U_ZERO_ERROR; |
1337 | icu::CharString likely; |
1338 | { |
1339 | icu::CharStringByteSink sink(&likely); |
1340 | ulocimp_addLikelySubtags(locale, sink, &errorCode); |
1341 | } |
1342 | if (U_FAILURE(errorCode) || errorCode == U_STRING_NOT_TERMINATED_WARNING) { |
1343 | return false; |
1344 | } |
1345 | scriptLength = uloc_getScript(likely.data(), script, UPRV_LENGTHOF(script), &errorCode); |
1346 | if (U_FAILURE(errorCode) || errorCode == U_STRING_NOT_TERMINATED_WARNING || |
1347 | scriptLength == 0) { |
1348 | return false; |
1349 | } |
1350 | } |
1351 | UScriptCode scriptCode = (UScriptCode)u_getPropertyValueEnum(UCHAR_SCRIPT, script); |
1352 | return uscript_isRightToLeft(scriptCode); |
1353 | } |
1354 | |
1355 | U_NAMESPACE_BEGIN |
1356 | |
1357 | UBool |
1358 | Locale::isRightToLeft() const { |
1359 | return uloc_isRightToLeft(getBaseName()); |
1360 | } |
1361 | |
1362 | U_NAMESPACE_END |
1363 | |
1364 | // The following must at least allow for rg key value (6) plus terminator (1). |
1365 | #define ULOC_RG_BUFLEN 8 |
1366 | |
1367 | U_CAPI int32_t U_EXPORT2 |
1368 | ulocimp_getRegionForSupplementalData(const char *localeID, UBool inferRegion, |
1369 | char *region, int32_t regionCapacity, UErrorCode* status) { |
1370 | if (U_FAILURE(*status)) { |
1371 | return 0; |
1372 | } |
1373 | char rgBuf[ULOC_RG_BUFLEN]; |
1374 | UErrorCode rgStatus = U_ZERO_ERROR; |
1375 | |
1376 | // First check for rg keyword value |
1377 | int32_t rgLen = uloc_getKeywordValue(localeID, "rg" , rgBuf, ULOC_RG_BUFLEN, &rgStatus); |
1378 | if (U_FAILURE(rgStatus) || rgLen != 6) { |
1379 | rgLen = 0; |
1380 | } else { |
1381 | // rgBuf guaranteed to be zero terminated here, with text len 6 |
1382 | char *rgPtr = rgBuf; |
1383 | for (; *rgPtr!= 0; rgPtr++) { |
1384 | *rgPtr = uprv_toupper(*rgPtr); |
1385 | } |
1386 | rgLen = (uprv_strcmp(rgBuf+2, "ZZZZ" ) == 0)? 2: 0; |
1387 | } |
1388 | |
1389 | if (rgLen == 0) { |
1390 | // No valid rg keyword value, try for unicode_region_subtag |
1391 | rgLen = uloc_getCountry(localeID, rgBuf, ULOC_RG_BUFLEN, status); |
1392 | if (U_FAILURE(*status)) { |
1393 | rgLen = 0; |
1394 | } else if (rgLen == 0 && inferRegion) { |
1395 | // no unicode_region_subtag but inferRegion true, try likely subtags |
1396 | rgStatus = U_ZERO_ERROR; |
1397 | icu::CharString locBuf; |
1398 | { |
1399 | icu::CharStringByteSink sink(&locBuf); |
1400 | ulocimp_addLikelySubtags(localeID, sink, &rgStatus); |
1401 | } |
1402 | if (U_SUCCESS(rgStatus)) { |
1403 | rgLen = uloc_getCountry(locBuf.data(), rgBuf, ULOC_RG_BUFLEN, status); |
1404 | if (U_FAILURE(*status)) { |
1405 | rgLen = 0; |
1406 | } |
1407 | } |
1408 | } |
1409 | } |
1410 | |
1411 | rgBuf[rgLen] = 0; |
1412 | uprv_strncpy(region, rgBuf, regionCapacity); |
1413 | return u_terminateChars(region, regionCapacity, rgLen, status); |
1414 | } |
1415 | |
1416 | |