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 = NULL; |
58 | |
59 | if (!U_FAILURE(*err)) { |
60 | int32_t resLen = 0; |
61 | const UChar* s = NULL; |
62 | UErrorCode tmpErr = U_ZERO_ERROR; |
63 | icu::LocalUResourceBundlePointer subtags(ures_openDirect(NULL, "likelySubtags" , &tmpErr)); |
64 | if (U_SUCCESS(tmpErr)) { |
65 | icu::CharString und; |
66 | if (localeID != NULL) { |
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 NULL; |
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/ouput 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 NULL 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 NULL, 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 NULL, 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 == NULL) { |
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 != NULL) { |
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 != NULL) { |
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 NULL 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 | NULL, |
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 == NULL || |
458 | lang == NULL || |
459 | langLength == NULL || |
460 | script == NULL || |
461 | scriptLength == NULL || |
462 | region == NULL || |
463 | regionLength == NULL) { |
464 | goto error; |
465 | } |
466 | |
467 | subtagLength = ulocimp_getLanguage(position, lang, *langLength, &position); |
468 | u_terminateChars(lang, *langLength, subtagLength, err); |
469 | |
470 | /* |
471 | * Note that we explicit consider U_STRING_NOT_TERMINATED_WARNING |
472 | * to be an error, because it indicates the user-supplied tag is |
473 | * not well-formed. |
474 | */ |
475 | if(U_FAILURE(*err)) { |
476 | goto error; |
477 | } |
478 | |
479 | *langLength = subtagLength; |
480 | |
481 | /* |
482 | * If no language was present, use the empty string instead. |
483 | * Otherwise, move past any separator. |
484 | */ |
485 | if (_isIDSeparator(*position)) { |
486 | ++position; |
487 | } |
488 | |
489 | subtagLength = ulocimp_getScript(position, script, *scriptLength, &position); |
490 | u_terminateChars(script, *scriptLength, subtagLength, err); |
491 | |
492 | if(U_FAILURE(*err)) { |
493 | goto error; |
494 | } |
495 | |
496 | *scriptLength = subtagLength; |
497 | |
498 | if (*scriptLength > 0) { |
499 | if (uprv_strnicmp(script, unknownScript, *scriptLength) == 0) { |
500 | /** |
501 | * If the script part is the "unknown" script, then don't return it. |
502 | **/ |
503 | *scriptLength = 0; |
504 | } |
505 | |
506 | /* |
507 | * Move past any separator. |
508 | */ |
509 | if (_isIDSeparator(*position)) { |
510 | ++position; |
511 | } |
512 | } |
513 | |
514 | subtagLength = ulocimp_getCountry(position, region, *regionLength, &position); |
515 | u_terminateChars(region, *regionLength, subtagLength, err); |
516 | |
517 | if(U_FAILURE(*err)) { |
518 | goto error; |
519 | } |
520 | |
521 | *regionLength = subtagLength; |
522 | |
523 | if (*regionLength > 0) { |
524 | if (uprv_strnicmp(region, unknownRegion, *regionLength) == 0) { |
525 | /** |
526 | * If the region part is the "unknown" region, then don't return it. |
527 | **/ |
528 | *regionLength = 0; |
529 | } |
530 | } else if (*position != 0 && *position != '@') { |
531 | /* back up over consumed trailing separator */ |
532 | --position; |
533 | } |
534 | |
535 | exit: |
536 | |
537 | return (int32_t)(position - localeID); |
538 | |
539 | error: |
540 | |
541 | /** |
542 | * If we get here, we have no explicit error, it's the result of an |
543 | * illegal argument. |
544 | **/ |
545 | if (!U_FAILURE(*err)) { |
546 | *err = U_ILLEGAL_ARGUMENT_ERROR; |
547 | } |
548 | |
549 | goto exit; |
550 | } |
551 | |
552 | static UBool U_CALLCONV |
553 | createLikelySubtagsString( |
554 | const char* lang, |
555 | int32_t langLength, |
556 | const char* script, |
557 | int32_t scriptLength, |
558 | const char* region, |
559 | int32_t regionLength, |
560 | const char* variants, |
561 | int32_t variantsLength, |
562 | icu::ByteSink& sink, |
563 | UErrorCode* err) { |
564 | /** |
565 | * ULOC_FULLNAME_CAPACITY will provide enough capacity |
566 | * that we can build a string that contains the language, |
567 | * script and region code without worrying about overrunning |
568 | * the user-supplied buffer. |
569 | **/ |
570 | char likelySubtagsBuffer[ULOC_FULLNAME_CAPACITY]; |
571 | |
572 | if(U_FAILURE(*err)) { |
573 | goto error; |
574 | } |
575 | |
576 | /** |
577 | * Try the language with the script and region first. |
578 | **/ |
579 | if (scriptLength > 0 && regionLength > 0) { |
580 | |
581 | const char* likelySubtags = NULL; |
582 | |
583 | icu::CharString tagBuffer; |
584 | { |
585 | icu::CharStringByteSink sink(&tagBuffer); |
586 | createTagString( |
587 | lang, |
588 | langLength, |
589 | script, |
590 | scriptLength, |
591 | region, |
592 | regionLength, |
593 | NULL, |
594 | 0, |
595 | sink, |
596 | err); |
597 | } |
598 | if(U_FAILURE(*err)) { |
599 | goto error; |
600 | } |
601 | |
602 | likelySubtags = |
603 | findLikelySubtags( |
604 | tagBuffer.data(), |
605 | likelySubtagsBuffer, |
606 | sizeof(likelySubtagsBuffer), |
607 | err); |
608 | if(U_FAILURE(*err)) { |
609 | goto error; |
610 | } |
611 | |
612 | if (likelySubtags != NULL) { |
613 | /* Always use the language tag from the |
614 | maximal string, since it may be more |
615 | specific than the one provided. */ |
616 | createTagStringWithAlternates( |
617 | NULL, |
618 | 0, |
619 | NULL, |
620 | 0, |
621 | NULL, |
622 | 0, |
623 | variants, |
624 | variantsLength, |
625 | likelySubtags, |
626 | sink, |
627 | err); |
628 | return TRUE; |
629 | } |
630 | } |
631 | |
632 | /** |
633 | * Try the language with just the script. |
634 | **/ |
635 | if (scriptLength > 0) { |
636 | |
637 | const char* likelySubtags = NULL; |
638 | |
639 | icu::CharString tagBuffer; |
640 | { |
641 | icu::CharStringByteSink sink(&tagBuffer); |
642 | createTagString( |
643 | lang, |
644 | langLength, |
645 | script, |
646 | scriptLength, |
647 | NULL, |
648 | 0, |
649 | NULL, |
650 | 0, |
651 | sink, |
652 | err); |
653 | } |
654 | if(U_FAILURE(*err)) { |
655 | goto error; |
656 | } |
657 | |
658 | likelySubtags = |
659 | findLikelySubtags( |
660 | tagBuffer.data(), |
661 | likelySubtagsBuffer, |
662 | sizeof(likelySubtagsBuffer), |
663 | err); |
664 | if(U_FAILURE(*err)) { |
665 | goto error; |
666 | } |
667 | |
668 | if (likelySubtags != NULL) { |
669 | /* Always use the language tag from the |
670 | maximal string, since it may be more |
671 | specific than the one provided. */ |
672 | createTagStringWithAlternates( |
673 | NULL, |
674 | 0, |
675 | NULL, |
676 | 0, |
677 | region, |
678 | regionLength, |
679 | variants, |
680 | variantsLength, |
681 | likelySubtags, |
682 | sink, |
683 | err); |
684 | return TRUE; |
685 | } |
686 | } |
687 | |
688 | /** |
689 | * Try the language with just the region. |
690 | **/ |
691 | if (regionLength > 0) { |
692 | |
693 | const char* likelySubtags = NULL; |
694 | |
695 | icu::CharString tagBuffer; |
696 | { |
697 | icu::CharStringByteSink sink(&tagBuffer); |
698 | createTagString( |
699 | lang, |
700 | langLength, |
701 | NULL, |
702 | 0, |
703 | region, |
704 | regionLength, |
705 | NULL, |
706 | 0, |
707 | sink, |
708 | err); |
709 | } |
710 | if(U_FAILURE(*err)) { |
711 | goto error; |
712 | } |
713 | |
714 | likelySubtags = |
715 | findLikelySubtags( |
716 | tagBuffer.data(), |
717 | likelySubtagsBuffer, |
718 | sizeof(likelySubtagsBuffer), |
719 | err); |
720 | if(U_FAILURE(*err)) { |
721 | goto error; |
722 | } |
723 | |
724 | if (likelySubtags != NULL) { |
725 | /* Always use the language tag from the |
726 | maximal string, since it may be more |
727 | specific than the one provided. */ |
728 | createTagStringWithAlternates( |
729 | NULL, |
730 | 0, |
731 | script, |
732 | scriptLength, |
733 | NULL, |
734 | 0, |
735 | variants, |
736 | variantsLength, |
737 | likelySubtags, |
738 | sink, |
739 | err); |
740 | return TRUE; |
741 | } |
742 | } |
743 | |
744 | /** |
745 | * Finally, try just the language. |
746 | **/ |
747 | { |
748 | const char* likelySubtags = NULL; |
749 | |
750 | icu::CharString tagBuffer; |
751 | { |
752 | icu::CharStringByteSink sink(&tagBuffer); |
753 | createTagString( |
754 | lang, |
755 | langLength, |
756 | NULL, |
757 | 0, |
758 | NULL, |
759 | 0, |
760 | NULL, |
761 | 0, |
762 | sink, |
763 | err); |
764 | } |
765 | if(U_FAILURE(*err)) { |
766 | goto error; |
767 | } |
768 | |
769 | likelySubtags = |
770 | findLikelySubtags( |
771 | tagBuffer.data(), |
772 | likelySubtagsBuffer, |
773 | sizeof(likelySubtagsBuffer), |
774 | err); |
775 | if(U_FAILURE(*err)) { |
776 | goto error; |
777 | } |
778 | |
779 | if (likelySubtags != NULL) { |
780 | /* Always use the language tag from the |
781 | maximal string, since it may be more |
782 | specific than the one provided. */ |
783 | createTagStringWithAlternates( |
784 | NULL, |
785 | 0, |
786 | script, |
787 | scriptLength, |
788 | region, |
789 | regionLength, |
790 | variants, |
791 | variantsLength, |
792 | likelySubtags, |
793 | sink, |
794 | err); |
795 | return TRUE; |
796 | } |
797 | } |
798 | |
799 | return FALSE; |
800 | |
801 | error: |
802 | |
803 | if (!U_FAILURE(*err)) { |
804 | *err = U_ILLEGAL_ARGUMENT_ERROR; |
805 | } |
806 | |
807 | return FALSE; |
808 | } |
809 | |
810 | #define CHECK_TRAILING_VARIANT_SIZE(trailing, trailingLength) UPRV_BLOCK_MACRO_BEGIN { \ |
811 | int32_t count = 0; \ |
812 | int32_t i; \ |
813 | for (i = 0; i < trailingLength; i++) { \ |
814 | if (trailing[i] == '-' || trailing[i] == '_') { \ |
815 | count = 0; \ |
816 | if (count > 8) { \ |
817 | goto error; \ |
818 | } \ |
819 | } else if (trailing[i] == '@') { \ |
820 | break; \ |
821 | } else if (count > 8) { \ |
822 | goto error; \ |
823 | } else { \ |
824 | count++; \ |
825 | } \ |
826 | } \ |
827 | } UPRV_BLOCK_MACRO_END |
828 | |
829 | static void |
830 | _uloc_addLikelySubtags(const char* localeID, |
831 | icu::ByteSink& sink, |
832 | UErrorCode* err) { |
833 | char lang[ULOC_LANG_CAPACITY]; |
834 | int32_t langLength = sizeof(lang); |
835 | char script[ULOC_SCRIPT_CAPACITY]; |
836 | int32_t scriptLength = sizeof(script); |
837 | char region[ULOC_COUNTRY_CAPACITY]; |
838 | int32_t regionLength = sizeof(region); |
839 | const char* trailing = "" ; |
840 | int32_t trailingLength = 0; |
841 | int32_t trailingIndex = 0; |
842 | UBool success = FALSE; |
843 | |
844 | if(U_FAILURE(*err)) { |
845 | goto error; |
846 | } |
847 | if (localeID == NULL) { |
848 | goto error; |
849 | } |
850 | |
851 | trailingIndex = parseTagString( |
852 | localeID, |
853 | lang, |
854 | &langLength, |
855 | script, |
856 | &scriptLength, |
857 | region, |
858 | ®ionLength, |
859 | err); |
860 | if(U_FAILURE(*err)) { |
861 | /* Overflow indicates an illegal argument error */ |
862 | if (*err == U_BUFFER_OVERFLOW_ERROR) { |
863 | *err = U_ILLEGAL_ARGUMENT_ERROR; |
864 | } |
865 | |
866 | goto error; |
867 | } |
868 | |
869 | /* Find the length of the trailing portion. */ |
870 | while (_isIDSeparator(localeID[trailingIndex])) { |
871 | trailingIndex++; |
872 | } |
873 | trailing = &localeID[trailingIndex]; |
874 | trailingLength = (int32_t)uprv_strlen(trailing); |
875 | |
876 | CHECK_TRAILING_VARIANT_SIZE(trailing, trailingLength); |
877 | |
878 | success = |
879 | createLikelySubtagsString( |
880 | lang, |
881 | langLength, |
882 | script, |
883 | scriptLength, |
884 | region, |
885 | regionLength, |
886 | trailing, |
887 | trailingLength, |
888 | sink, |
889 | err); |
890 | |
891 | if (!success) { |
892 | const int32_t localIDLength = (int32_t)uprv_strlen(localeID); |
893 | |
894 | /* |
895 | * If we get here, we need to return localeID. |
896 | */ |
897 | sink.Append(localeID, localIDLength); |
898 | } |
899 | |
900 | return; |
901 | |
902 | error: |
903 | |
904 | if (!U_FAILURE(*err)) { |
905 | *err = U_ILLEGAL_ARGUMENT_ERROR; |
906 | } |
907 | } |
908 | |
909 | static void |
910 | _uloc_minimizeSubtags(const char* localeID, |
911 | icu::ByteSink& sink, |
912 | UErrorCode* err) { |
913 | icu::CharString maximizedTagBuffer; |
914 | |
915 | char lang[ULOC_LANG_CAPACITY]; |
916 | int32_t langLength = sizeof(lang); |
917 | char script[ULOC_SCRIPT_CAPACITY]; |
918 | int32_t scriptLength = sizeof(script); |
919 | char region[ULOC_COUNTRY_CAPACITY]; |
920 | int32_t regionLength = sizeof(region); |
921 | const char* trailing = "" ; |
922 | int32_t trailingLength = 0; |
923 | int32_t trailingIndex = 0; |
924 | |
925 | if(U_FAILURE(*err)) { |
926 | goto error; |
927 | } |
928 | else if (localeID == NULL) { |
929 | goto error; |
930 | } |
931 | |
932 | trailingIndex = |
933 | parseTagString( |
934 | localeID, |
935 | lang, |
936 | &langLength, |
937 | script, |
938 | &scriptLength, |
939 | region, |
940 | ®ionLength, |
941 | err); |
942 | if(U_FAILURE(*err)) { |
943 | |
944 | /* Overflow indicates an illegal argument error */ |
945 | if (*err == U_BUFFER_OVERFLOW_ERROR) { |
946 | *err = U_ILLEGAL_ARGUMENT_ERROR; |
947 | } |
948 | |
949 | goto error; |
950 | } |
951 | |
952 | /* Find the spot where the variants or the keywords begin, if any. */ |
953 | while (_isIDSeparator(localeID[trailingIndex])) { |
954 | trailingIndex++; |
955 | } |
956 | trailing = &localeID[trailingIndex]; |
957 | trailingLength = (int32_t)uprv_strlen(trailing); |
958 | |
959 | CHECK_TRAILING_VARIANT_SIZE(trailing, trailingLength); |
960 | |
961 | { |
962 | icu::CharString base; |
963 | { |
964 | icu::CharStringByteSink sink(&base); |
965 | createTagString( |
966 | lang, |
967 | langLength, |
968 | script, |
969 | scriptLength, |
970 | region, |
971 | regionLength, |
972 | NULL, |
973 | 0, |
974 | sink, |
975 | err); |
976 | } |
977 | |
978 | /** |
979 | * First, we need to first get the maximization |
980 | * from AddLikelySubtags. |
981 | **/ |
982 | { |
983 | icu::CharStringByteSink sink(&maximizedTagBuffer); |
984 | ulocimp_addLikelySubtags(base.data(), sink, err); |
985 | } |
986 | } |
987 | |
988 | if(U_FAILURE(*err)) { |
989 | goto error; |
990 | } |
991 | |
992 | /** |
993 | * Start first with just the language. |
994 | **/ |
995 | { |
996 | icu::CharString tagBuffer; |
997 | { |
998 | icu::CharStringByteSink sink(&tagBuffer); |
999 | createLikelySubtagsString( |
1000 | lang, |
1001 | langLength, |
1002 | NULL, |
1003 | 0, |
1004 | NULL, |
1005 | 0, |
1006 | NULL, |
1007 | 0, |
1008 | sink, |
1009 | err); |
1010 | } |
1011 | |
1012 | if(U_FAILURE(*err)) { |
1013 | goto error; |
1014 | } |
1015 | else if (!tagBuffer.isEmpty() && uprv_strnicmp( |
1016 | maximizedTagBuffer.data(), |
1017 | tagBuffer.data(), |
1018 | tagBuffer.length()) == 0) { |
1019 | |
1020 | createTagString( |
1021 | lang, |
1022 | langLength, |
1023 | NULL, |
1024 | 0, |
1025 | NULL, |
1026 | 0, |
1027 | trailing, |
1028 | trailingLength, |
1029 | sink, |
1030 | err); |
1031 | return; |
1032 | } |
1033 | } |
1034 | |
1035 | /** |
1036 | * Next, try the language and region. |
1037 | **/ |
1038 | if (regionLength > 0) { |
1039 | |
1040 | icu::CharString tagBuffer; |
1041 | { |
1042 | icu::CharStringByteSink sink(&tagBuffer); |
1043 | createLikelySubtagsString( |
1044 | lang, |
1045 | langLength, |
1046 | NULL, |
1047 | 0, |
1048 | region, |
1049 | regionLength, |
1050 | NULL, |
1051 | 0, |
1052 | sink, |
1053 | err); |
1054 | } |
1055 | |
1056 | if(U_FAILURE(*err)) { |
1057 | goto error; |
1058 | } |
1059 | else if (uprv_strnicmp( |
1060 | maximizedTagBuffer.data(), |
1061 | tagBuffer.data(), |
1062 | tagBuffer.length()) == 0) { |
1063 | |
1064 | createTagString( |
1065 | lang, |
1066 | langLength, |
1067 | NULL, |
1068 | 0, |
1069 | region, |
1070 | regionLength, |
1071 | trailing, |
1072 | trailingLength, |
1073 | sink, |
1074 | err); |
1075 | return; |
1076 | } |
1077 | } |
1078 | |
1079 | /** |
1080 | * Finally, try the language and script. This is our last chance, |
1081 | * since trying with all three subtags would only yield the |
1082 | * maximal version that we already have. |
1083 | **/ |
1084 | if (scriptLength > 0 && regionLength > 0) { |
1085 | icu::CharString tagBuffer; |
1086 | { |
1087 | icu::CharStringByteSink sink(&tagBuffer); |
1088 | createLikelySubtagsString( |
1089 | lang, |
1090 | langLength, |
1091 | script, |
1092 | scriptLength, |
1093 | NULL, |
1094 | 0, |
1095 | NULL, |
1096 | 0, |
1097 | sink, |
1098 | err); |
1099 | } |
1100 | |
1101 | if(U_FAILURE(*err)) { |
1102 | goto error; |
1103 | } |
1104 | else if (uprv_strnicmp( |
1105 | maximizedTagBuffer.data(), |
1106 | tagBuffer.data(), |
1107 | tagBuffer.length()) == 0) { |
1108 | |
1109 | createTagString( |
1110 | lang, |
1111 | langLength, |
1112 | script, |
1113 | scriptLength, |
1114 | NULL, |
1115 | 0, |
1116 | trailing, |
1117 | trailingLength, |
1118 | sink, |
1119 | err); |
1120 | return; |
1121 | } |
1122 | } |
1123 | |
1124 | { |
1125 | /** |
1126 | * If we got here, return the locale ID parameter. |
1127 | **/ |
1128 | const int32_t localeIDLength = (int32_t)uprv_strlen(localeID); |
1129 | sink.Append(localeID, localeIDLength); |
1130 | return; |
1131 | } |
1132 | |
1133 | error: |
1134 | |
1135 | if (!U_FAILURE(*err)) { |
1136 | *err = U_ILLEGAL_ARGUMENT_ERROR; |
1137 | } |
1138 | } |
1139 | |
1140 | static UBool |
1141 | do_canonicalize(const char* localeID, |
1142 | char* buffer, |
1143 | int32_t bufferCapacity, |
1144 | UErrorCode* err) |
1145 | { |
1146 | uloc_canonicalize( |
1147 | localeID, |
1148 | buffer, |
1149 | bufferCapacity, |
1150 | err); |
1151 | |
1152 | if (*err == U_STRING_NOT_TERMINATED_WARNING || |
1153 | *err == U_BUFFER_OVERFLOW_ERROR) { |
1154 | *err = U_ILLEGAL_ARGUMENT_ERROR; |
1155 | |
1156 | return FALSE; |
1157 | } |
1158 | else if (U_FAILURE(*err)) { |
1159 | |
1160 | return FALSE; |
1161 | } |
1162 | else { |
1163 | return TRUE; |
1164 | } |
1165 | } |
1166 | |
1167 | U_CAPI int32_t U_EXPORT2 |
1168 | uloc_addLikelySubtags(const char* localeID, |
1169 | char* maximizedLocaleID, |
1170 | int32_t maximizedLocaleIDCapacity, |
1171 | UErrorCode* status) { |
1172 | if (U_FAILURE(*status)) { |
1173 | return 0; |
1174 | } |
1175 | |
1176 | icu::CheckedArrayByteSink sink( |
1177 | maximizedLocaleID, maximizedLocaleIDCapacity); |
1178 | |
1179 | ulocimp_addLikelySubtags(localeID, sink, status); |
1180 | int32_t reslen = sink.NumberOfBytesAppended(); |
1181 | |
1182 | if (U_FAILURE(*status)) { |
1183 | return sink.Overflowed() ? reslen : -1; |
1184 | } |
1185 | |
1186 | if (sink.Overflowed()) { |
1187 | *status = U_BUFFER_OVERFLOW_ERROR; |
1188 | } else { |
1189 | u_terminateChars( |
1190 | maximizedLocaleID, maximizedLocaleIDCapacity, reslen, status); |
1191 | } |
1192 | |
1193 | return reslen; |
1194 | } |
1195 | |
1196 | U_CAPI void U_EXPORT2 |
1197 | ulocimp_addLikelySubtags(const char* localeID, |
1198 | icu::ByteSink& sink, |
1199 | UErrorCode* status) { |
1200 | char localeBuffer[ULOC_FULLNAME_CAPACITY]; |
1201 | |
1202 | if (do_canonicalize(localeID, localeBuffer, sizeof localeBuffer, status)) { |
1203 | _uloc_addLikelySubtags(localeBuffer, sink, status); |
1204 | } |
1205 | } |
1206 | |
1207 | U_CAPI int32_t U_EXPORT2 |
1208 | uloc_minimizeSubtags(const char* localeID, |
1209 | char* minimizedLocaleID, |
1210 | int32_t minimizedLocaleIDCapacity, |
1211 | UErrorCode* status) { |
1212 | if (U_FAILURE(*status)) { |
1213 | return 0; |
1214 | } |
1215 | |
1216 | icu::CheckedArrayByteSink sink( |
1217 | minimizedLocaleID, minimizedLocaleIDCapacity); |
1218 | |
1219 | ulocimp_minimizeSubtags(localeID, sink, status); |
1220 | int32_t reslen = sink.NumberOfBytesAppended(); |
1221 | |
1222 | if (U_FAILURE(*status)) { |
1223 | return sink.Overflowed() ? reslen : -1; |
1224 | } |
1225 | |
1226 | if (sink.Overflowed()) { |
1227 | *status = U_BUFFER_OVERFLOW_ERROR; |
1228 | } else { |
1229 | u_terminateChars( |
1230 | minimizedLocaleID, minimizedLocaleIDCapacity, reslen, status); |
1231 | } |
1232 | |
1233 | return reslen; |
1234 | } |
1235 | |
1236 | U_CAPI void U_EXPORT2 |
1237 | ulocimp_minimizeSubtags(const char* localeID, |
1238 | icu::ByteSink& sink, |
1239 | UErrorCode* status) { |
1240 | char localeBuffer[ULOC_FULLNAME_CAPACITY]; |
1241 | |
1242 | if (do_canonicalize(localeID, localeBuffer, sizeof localeBuffer, status)) { |
1243 | _uloc_minimizeSubtags(localeBuffer, sink, status); |
1244 | } |
1245 | } |
1246 | |
1247 | // Pairs of (language subtag, + or -) for finding out fast if common languages |
1248 | // are LTR (minus) or RTL (plus). |
1249 | static const char LANG_DIR_STRING[] = |
1250 | "root-en-es-pt-zh-ja-ko-de-fr-it-ar+he+fa+ru-nl-pl-th-tr-" ; |
1251 | |
1252 | // Implemented here because this calls ulocimp_addLikelySubtags(). |
1253 | U_CAPI UBool U_EXPORT2 |
1254 | uloc_isRightToLeft(const char *locale) { |
1255 | UErrorCode errorCode = U_ZERO_ERROR; |
1256 | char script[8]; |
1257 | int32_t scriptLength = uloc_getScript(locale, script, UPRV_LENGTHOF(script), &errorCode); |
1258 | if (U_FAILURE(errorCode) || errorCode == U_STRING_NOT_TERMINATED_WARNING || |
1259 | scriptLength == 0) { |
1260 | // Fastpath: We know the likely scripts and their writing direction |
1261 | // for some common languages. |
1262 | errorCode = U_ZERO_ERROR; |
1263 | char lang[8]; |
1264 | int32_t langLength = uloc_getLanguage(locale, lang, UPRV_LENGTHOF(lang), &errorCode); |
1265 | if (U_FAILURE(errorCode) || errorCode == U_STRING_NOT_TERMINATED_WARNING) { |
1266 | return FALSE; |
1267 | } |
1268 | if (langLength > 0) { |
1269 | const char* langPtr = uprv_strstr(LANG_DIR_STRING, lang); |
1270 | if (langPtr != NULL) { |
1271 | switch (langPtr[langLength]) { |
1272 | case '-': return FALSE; |
1273 | case '+': return TRUE; |
1274 | default: break; // partial match of a longer code |
1275 | } |
1276 | } |
1277 | } |
1278 | // Otherwise, find the likely script. |
1279 | errorCode = U_ZERO_ERROR; |
1280 | icu::CharString likely; |
1281 | { |
1282 | icu::CharStringByteSink sink(&likely); |
1283 | ulocimp_addLikelySubtags(locale, sink, &errorCode); |
1284 | } |
1285 | if (U_FAILURE(errorCode) || errorCode == U_STRING_NOT_TERMINATED_WARNING) { |
1286 | return FALSE; |
1287 | } |
1288 | scriptLength = uloc_getScript(likely.data(), script, UPRV_LENGTHOF(script), &errorCode); |
1289 | if (U_FAILURE(errorCode) || errorCode == U_STRING_NOT_TERMINATED_WARNING || |
1290 | scriptLength == 0) { |
1291 | return FALSE; |
1292 | } |
1293 | } |
1294 | UScriptCode scriptCode = (UScriptCode)u_getPropertyValueEnum(UCHAR_SCRIPT, script); |
1295 | return uscript_isRightToLeft(scriptCode); |
1296 | } |
1297 | |
1298 | U_NAMESPACE_BEGIN |
1299 | |
1300 | UBool |
1301 | Locale::isRightToLeft() const { |
1302 | return uloc_isRightToLeft(getBaseName()); |
1303 | } |
1304 | |
1305 | U_NAMESPACE_END |
1306 | |
1307 | // The following must at least allow for rg key value (6) plus terminator (1). |
1308 | #define ULOC_RG_BUFLEN 8 |
1309 | |
1310 | U_CAPI int32_t U_EXPORT2 |
1311 | ulocimp_getRegionForSupplementalData(const char *localeID, UBool inferRegion, |
1312 | char *region, int32_t regionCapacity, UErrorCode* status) { |
1313 | if (U_FAILURE(*status)) { |
1314 | return 0; |
1315 | } |
1316 | char rgBuf[ULOC_RG_BUFLEN]; |
1317 | UErrorCode rgStatus = U_ZERO_ERROR; |
1318 | |
1319 | // First check for rg keyword value |
1320 | int32_t rgLen = uloc_getKeywordValue(localeID, "rg" , rgBuf, ULOC_RG_BUFLEN, &rgStatus); |
1321 | if (U_FAILURE(rgStatus) || rgLen != 6) { |
1322 | rgLen = 0; |
1323 | } else { |
1324 | // rgBuf guaranteed to be zero terminated here, with text len 6 |
1325 | char *rgPtr = rgBuf; |
1326 | for (; *rgPtr!= 0; rgPtr++) { |
1327 | *rgPtr = uprv_toupper(*rgPtr); |
1328 | } |
1329 | rgLen = (uprv_strcmp(rgBuf+2, "ZZZZ" ) == 0)? 2: 0; |
1330 | } |
1331 | |
1332 | if (rgLen == 0) { |
1333 | // No valid rg keyword value, try for unicode_region_subtag |
1334 | rgLen = uloc_getCountry(localeID, rgBuf, ULOC_RG_BUFLEN, status); |
1335 | if (U_FAILURE(*status)) { |
1336 | rgLen = 0; |
1337 | } else if (rgLen == 0 && inferRegion) { |
1338 | // no unicode_region_subtag but inferRegion TRUE, try likely subtags |
1339 | rgStatus = U_ZERO_ERROR; |
1340 | icu::CharString locBuf; |
1341 | { |
1342 | icu::CharStringByteSink sink(&locBuf); |
1343 | ulocimp_addLikelySubtags(localeID, sink, &rgStatus); |
1344 | } |
1345 | if (U_SUCCESS(rgStatus)) { |
1346 | rgLen = uloc_getCountry(locBuf.data(), rgBuf, ULOC_RG_BUFLEN, status); |
1347 | if (U_FAILURE(*status)) { |
1348 | rgLen = 0; |
1349 | } |
1350 | } |
1351 | } |
1352 | } |
1353 | |
1354 | rgBuf[rgLen] = 0; |
1355 | uprv_strncpy(region, rgBuf, regionCapacity); |
1356 | return u_terminateChars(region, regionCapacity, rgLen, status); |
1357 | } |
1358 | |
1359 | |