|
1 /* |
|
2 ******************************************************************************* |
|
3 * |
|
4 * Copyright (C) 2003-2013, International Business Machines |
|
5 * Corporation and others. All Rights Reserved. |
|
6 * |
|
7 ******************************************************************************* |
|
8 * file name: usprep.cpp |
|
9 * encoding: US-ASCII |
|
10 * tab size: 8 (not used) |
|
11 * indentation:4 |
|
12 * |
|
13 * created on: 2003jul2 |
|
14 * created by: Ram Viswanadha |
|
15 */ |
|
16 |
|
17 #include "unicode/utypes.h" |
|
18 |
|
19 #if !UCONFIG_NO_IDNA |
|
20 |
|
21 #include "unicode/usprep.h" |
|
22 |
|
23 #include "unicode/unorm.h" |
|
24 #include "unicode/ustring.h" |
|
25 #include "unicode/uchar.h" |
|
26 #include "unicode/uversion.h" |
|
27 #include "umutex.h" |
|
28 #include "cmemory.h" |
|
29 #include "sprpimpl.h" |
|
30 #include "ustr_imp.h" |
|
31 #include "uhash.h" |
|
32 #include "cstring.h" |
|
33 #include "udataswp.h" |
|
34 #include "ucln_cmn.h" |
|
35 #include "ubidi_props.h" |
|
36 |
|
37 U_NAMESPACE_USE |
|
38 |
|
39 U_CDECL_BEGIN |
|
40 |
|
41 /* |
|
42 Static cache for already opened StringPrep profiles |
|
43 */ |
|
44 static UHashtable *SHARED_DATA_HASHTABLE = NULL; |
|
45 static icu::UInitOnce gSharedDataInitOnce; |
|
46 |
|
47 static UMutex usprepMutex = U_MUTEX_INITIALIZER; |
|
48 |
|
49 /* format version of spp file */ |
|
50 //static uint8_t formatVersion[4]={ 0, 0, 0, 0 }; |
|
51 |
|
52 /* the Unicode version of the sprep data */ |
|
53 static UVersionInfo dataVersion={ 0, 0, 0, 0 }; |
|
54 |
|
55 /* Profile names must be aligned to UStringPrepProfileType */ |
|
56 static const char * const PROFILE_NAMES[] = { |
|
57 "rfc3491", /* USPREP_RFC3491_NAMEPREP */ |
|
58 "rfc3530cs", /* USPREP_RFC3530_NFS4_CS_PREP */ |
|
59 "rfc3530csci", /* USPREP_RFC3530_NFS4_CS_PREP_CI */ |
|
60 "rfc3491", /* USPREP_RFC3530_NSF4_CIS_PREP */ |
|
61 "rfc3530mixp", /* USPREP_RFC3530_NSF4_MIXED_PREP_PREFIX */ |
|
62 "rfc3491", /* USPREP_RFC3530_NSF4_MIXED_PREP_SUFFIX */ |
|
63 "rfc3722", /* USPREP_RFC3722_ISCSI */ |
|
64 "rfc3920node", /* USPREP_RFC3920_NODEPREP */ |
|
65 "rfc3920res", /* USPREP_RFC3920_RESOURCEPREP */ |
|
66 "rfc4011", /* USPREP_RFC4011_MIB */ |
|
67 "rfc4013", /* USPREP_RFC4013_SASLPREP */ |
|
68 "rfc4505", /* USPREP_RFC4505_TRACE */ |
|
69 "rfc4518", /* USPREP_RFC4518_LDAP */ |
|
70 "rfc4518ci", /* USPREP_RFC4518_LDAP_CI */ |
|
71 }; |
|
72 |
|
73 static UBool U_CALLCONV |
|
74 isSPrepAcceptable(void * /* context */, |
|
75 const char * /* type */, |
|
76 const char * /* name */, |
|
77 const UDataInfo *pInfo) { |
|
78 if( |
|
79 pInfo->size>=20 && |
|
80 pInfo->isBigEndian==U_IS_BIG_ENDIAN && |
|
81 pInfo->charsetFamily==U_CHARSET_FAMILY && |
|
82 pInfo->dataFormat[0]==0x53 && /* dataFormat="SPRP" */ |
|
83 pInfo->dataFormat[1]==0x50 && |
|
84 pInfo->dataFormat[2]==0x52 && |
|
85 pInfo->dataFormat[3]==0x50 && |
|
86 pInfo->formatVersion[0]==3 && |
|
87 pInfo->formatVersion[2]==UTRIE_SHIFT && |
|
88 pInfo->formatVersion[3]==UTRIE_INDEX_SHIFT |
|
89 ) { |
|
90 //uprv_memcpy(formatVersion, pInfo->formatVersion, 4); |
|
91 uprv_memcpy(dataVersion, pInfo->dataVersion, 4); |
|
92 return TRUE; |
|
93 } else { |
|
94 return FALSE; |
|
95 } |
|
96 } |
|
97 |
|
98 static int32_t U_CALLCONV |
|
99 getSPrepFoldingOffset(uint32_t data) { |
|
100 |
|
101 return (int32_t)data; |
|
102 |
|
103 } |
|
104 |
|
105 /* hashes an entry */ |
|
106 static int32_t U_CALLCONV |
|
107 hashEntry(const UHashTok parm) { |
|
108 UStringPrepKey *b = (UStringPrepKey *)parm.pointer; |
|
109 UHashTok namekey, pathkey; |
|
110 namekey.pointer = b->name; |
|
111 pathkey.pointer = b->path; |
|
112 return uhash_hashChars(namekey)+37*uhash_hashChars(pathkey); |
|
113 } |
|
114 |
|
115 /* compares two entries */ |
|
116 static UBool U_CALLCONV |
|
117 compareEntries(const UHashTok p1, const UHashTok p2) { |
|
118 UStringPrepKey *b1 = (UStringPrepKey *)p1.pointer; |
|
119 UStringPrepKey *b2 = (UStringPrepKey *)p2.pointer; |
|
120 UHashTok name1, name2, path1, path2; |
|
121 name1.pointer = b1->name; |
|
122 name2.pointer = b2->name; |
|
123 path1.pointer = b1->path; |
|
124 path2.pointer = b2->path; |
|
125 return ((UBool)(uhash_compareChars(name1, name2) & |
|
126 uhash_compareChars(path1, path2))); |
|
127 } |
|
128 |
|
129 static void |
|
130 usprep_unload(UStringPrepProfile* data){ |
|
131 udata_close(data->sprepData); |
|
132 } |
|
133 |
|
134 static int32_t |
|
135 usprep_internal_flushCache(UBool noRefCount){ |
|
136 UStringPrepProfile *profile = NULL; |
|
137 UStringPrepKey *key = NULL; |
|
138 int32_t pos = -1; |
|
139 int32_t deletedNum = 0; |
|
140 const UHashElement *e; |
|
141 |
|
142 /* |
|
143 * if shared data hasn't even been lazy evaluated yet |
|
144 * return 0 |
|
145 */ |
|
146 umtx_lock(&usprepMutex); |
|
147 if (SHARED_DATA_HASHTABLE == NULL) { |
|
148 umtx_unlock(&usprepMutex); |
|
149 return 0; |
|
150 } |
|
151 |
|
152 /*creates an enumeration to iterate through every element in the table */ |
|
153 while ((e = uhash_nextElement(SHARED_DATA_HASHTABLE, &pos)) != NULL) |
|
154 { |
|
155 profile = (UStringPrepProfile *) e->value.pointer; |
|
156 key = (UStringPrepKey *) e->key.pointer; |
|
157 |
|
158 if ((noRefCount== FALSE && profile->refCount == 0) || |
|
159 noRefCount== TRUE) { |
|
160 deletedNum++; |
|
161 uhash_removeElement(SHARED_DATA_HASHTABLE, e); |
|
162 |
|
163 /* unload the data */ |
|
164 usprep_unload(profile); |
|
165 |
|
166 if(key->name != NULL) { |
|
167 uprv_free(key->name); |
|
168 key->name=NULL; |
|
169 } |
|
170 if(key->path != NULL) { |
|
171 uprv_free(key->path); |
|
172 key->path=NULL; |
|
173 } |
|
174 uprv_free(profile); |
|
175 uprv_free(key); |
|
176 } |
|
177 |
|
178 } |
|
179 umtx_unlock(&usprepMutex); |
|
180 |
|
181 return deletedNum; |
|
182 } |
|
183 |
|
184 /* Works just like ucnv_flushCache() |
|
185 static int32_t |
|
186 usprep_flushCache(){ |
|
187 return usprep_internal_flushCache(FALSE); |
|
188 } |
|
189 */ |
|
190 |
|
191 static UBool U_CALLCONV usprep_cleanup(void){ |
|
192 if (SHARED_DATA_HASHTABLE != NULL) { |
|
193 usprep_internal_flushCache(TRUE); |
|
194 if (SHARED_DATA_HASHTABLE != NULL && uhash_count(SHARED_DATA_HASHTABLE) == 0) { |
|
195 uhash_close(SHARED_DATA_HASHTABLE); |
|
196 SHARED_DATA_HASHTABLE = NULL; |
|
197 } |
|
198 } |
|
199 gSharedDataInitOnce.reset(); |
|
200 return (SHARED_DATA_HASHTABLE == NULL); |
|
201 } |
|
202 U_CDECL_END |
|
203 |
|
204 |
|
205 /** Initializes the cache for resources */ |
|
206 static void U_CALLCONV |
|
207 createCache(UErrorCode &status) { |
|
208 SHARED_DATA_HASHTABLE = uhash_open(hashEntry, compareEntries, NULL, &status); |
|
209 if (U_FAILURE(status)) { |
|
210 SHARED_DATA_HASHTABLE = NULL; |
|
211 } |
|
212 ucln_common_registerCleanup(UCLN_COMMON_USPREP, usprep_cleanup); |
|
213 } |
|
214 |
|
215 static void |
|
216 initCache(UErrorCode *status) { |
|
217 umtx_initOnce(gSharedDataInitOnce, &createCache, *status); |
|
218 } |
|
219 |
|
220 static UBool U_CALLCONV |
|
221 loadData(UStringPrepProfile* profile, |
|
222 const char* path, |
|
223 const char* name, |
|
224 const char* type, |
|
225 UErrorCode* errorCode) { |
|
226 /* load Unicode SPREP data from file */ |
|
227 UTrie _sprepTrie={ 0,0,0,0,0,0,0 }; |
|
228 UDataMemory *dataMemory; |
|
229 const int32_t *p=NULL; |
|
230 const uint8_t *pb; |
|
231 UVersionInfo normUnicodeVersion; |
|
232 int32_t normUniVer, sprepUniVer, normCorrVer; |
|
233 |
|
234 if(errorCode==NULL || U_FAILURE(*errorCode)) { |
|
235 return 0; |
|
236 } |
|
237 |
|
238 /* open the data outside the mutex block */ |
|
239 //TODO: change the path |
|
240 dataMemory=udata_openChoice(path, type, name, isSPrepAcceptable, NULL, errorCode); |
|
241 if(U_FAILURE(*errorCode)) { |
|
242 return FALSE; |
|
243 } |
|
244 |
|
245 p=(const int32_t *)udata_getMemory(dataMemory); |
|
246 pb=(const uint8_t *)(p+_SPREP_INDEX_TOP); |
|
247 utrie_unserialize(&_sprepTrie, pb, p[_SPREP_INDEX_TRIE_SIZE], errorCode); |
|
248 _sprepTrie.getFoldingOffset=getSPrepFoldingOffset; |
|
249 |
|
250 |
|
251 if(U_FAILURE(*errorCode)) { |
|
252 udata_close(dataMemory); |
|
253 return FALSE; |
|
254 } |
|
255 |
|
256 /* in the mutex block, set the data for this process */ |
|
257 umtx_lock(&usprepMutex); |
|
258 if(profile->sprepData==NULL) { |
|
259 profile->sprepData=dataMemory; |
|
260 dataMemory=NULL; |
|
261 uprv_memcpy(&profile->indexes, p, sizeof(profile->indexes)); |
|
262 uprv_memcpy(&profile->sprepTrie, &_sprepTrie, sizeof(UTrie)); |
|
263 } else { |
|
264 p=(const int32_t *)udata_getMemory(profile->sprepData); |
|
265 } |
|
266 umtx_unlock(&usprepMutex); |
|
267 /* initialize some variables */ |
|
268 profile->mappingData=(uint16_t *)((uint8_t *)(p+_SPREP_INDEX_TOP)+profile->indexes[_SPREP_INDEX_TRIE_SIZE]); |
|
269 |
|
270 u_getUnicodeVersion(normUnicodeVersion); |
|
271 normUniVer = (normUnicodeVersion[0] << 24) + (normUnicodeVersion[1] << 16) + |
|
272 (normUnicodeVersion[2] << 8 ) + (normUnicodeVersion[3]); |
|
273 sprepUniVer = (dataVersion[0] << 24) + (dataVersion[1] << 16) + |
|
274 (dataVersion[2] << 8 ) + (dataVersion[3]); |
|
275 normCorrVer = profile->indexes[_SPREP_NORM_CORRECTNS_LAST_UNI_VERSION]; |
|
276 |
|
277 if(U_FAILURE(*errorCode)){ |
|
278 udata_close(dataMemory); |
|
279 return FALSE; |
|
280 } |
|
281 if( normUniVer < sprepUniVer && /* the Unicode version of SPREP file must be less than the Unicode Vesion of the normalization data */ |
|
282 normUniVer < normCorrVer && /* the Unicode version of the NormalizationCorrections.txt file should be less than the Unicode Vesion of the normalization data */ |
|
283 ((profile->indexes[_SPREP_OPTIONS] & _SPREP_NORMALIZATION_ON) > 0) /* normalization turned on*/ |
|
284 ){ |
|
285 *errorCode = U_INVALID_FORMAT_ERROR; |
|
286 udata_close(dataMemory); |
|
287 return FALSE; |
|
288 } |
|
289 profile->isDataLoaded = TRUE; |
|
290 |
|
291 /* if a different thread set it first, then close the extra data */ |
|
292 if(dataMemory!=NULL) { |
|
293 udata_close(dataMemory); /* NULL if it was set correctly */ |
|
294 } |
|
295 |
|
296 |
|
297 return profile->isDataLoaded; |
|
298 } |
|
299 |
|
300 static UStringPrepProfile* |
|
301 usprep_getProfile(const char* path, |
|
302 const char* name, |
|
303 UErrorCode *status){ |
|
304 |
|
305 UStringPrepProfile* profile = NULL; |
|
306 |
|
307 initCache(status); |
|
308 |
|
309 if(U_FAILURE(*status)){ |
|
310 return NULL; |
|
311 } |
|
312 |
|
313 UStringPrepKey stackKey; |
|
314 /* |
|
315 * const is cast way to save malloc, strcpy and free calls |
|
316 * we use the passed in pointers for fetching the data from the |
|
317 * hash table which is safe |
|
318 */ |
|
319 stackKey.name = (char*) name; |
|
320 stackKey.path = (char*) path; |
|
321 |
|
322 /* fetch the data from the cache */ |
|
323 umtx_lock(&usprepMutex); |
|
324 profile = (UStringPrepProfile*) (uhash_get(SHARED_DATA_HASHTABLE,&stackKey)); |
|
325 if(profile != NULL) { |
|
326 profile->refCount++; |
|
327 } |
|
328 umtx_unlock(&usprepMutex); |
|
329 |
|
330 if(profile == NULL) { |
|
331 /* else load the data and put the data in the cache */ |
|
332 LocalMemory<UStringPrepProfile> newProfile; |
|
333 if(newProfile.allocateInsteadAndReset() == NULL) { |
|
334 *status = U_MEMORY_ALLOCATION_ERROR; |
|
335 return NULL; |
|
336 } |
|
337 |
|
338 /* load the data */ |
|
339 if(!loadData(newProfile.getAlias(), path, name, _SPREP_DATA_TYPE, status) || U_FAILURE(*status) ){ |
|
340 return NULL; |
|
341 } |
|
342 |
|
343 /* get the options */ |
|
344 newProfile->doNFKC = (UBool)((newProfile->indexes[_SPREP_OPTIONS] & _SPREP_NORMALIZATION_ON) > 0); |
|
345 newProfile->checkBiDi = (UBool)((newProfile->indexes[_SPREP_OPTIONS] & _SPREP_CHECK_BIDI_ON) > 0); |
|
346 |
|
347 if(newProfile->checkBiDi) { |
|
348 newProfile->bdp = ubidi_getSingleton(); |
|
349 } |
|
350 |
|
351 LocalMemory<UStringPrepKey> key; |
|
352 LocalMemory<char> keyName; |
|
353 LocalMemory<char> keyPath; |
|
354 if( key.allocateInsteadAndReset() == NULL || |
|
355 keyName.allocateInsteadAndCopy(uprv_strlen(name)+1) == NULL || |
|
356 (path != NULL && |
|
357 keyPath.allocateInsteadAndCopy(uprv_strlen(path)+1) == NULL) |
|
358 ) { |
|
359 *status = U_MEMORY_ALLOCATION_ERROR; |
|
360 usprep_unload(newProfile.getAlias()); |
|
361 return NULL; |
|
362 } |
|
363 |
|
364 umtx_lock(&usprepMutex); |
|
365 // If another thread already inserted the same key/value, refcount and cleanup our thread data |
|
366 profile = (UStringPrepProfile*) (uhash_get(SHARED_DATA_HASHTABLE,&stackKey)); |
|
367 if(profile != NULL) { |
|
368 profile->refCount++; |
|
369 usprep_unload(newProfile.getAlias()); |
|
370 } |
|
371 else { |
|
372 /* initialize the key members */ |
|
373 key->name = keyName.orphan(); |
|
374 uprv_strcpy(key->name, name); |
|
375 if(path != NULL){ |
|
376 key->path = keyPath.orphan(); |
|
377 uprv_strcpy(key->path, path); |
|
378 } |
|
379 profile = newProfile.orphan(); |
|
380 |
|
381 /* add the data object to the cache */ |
|
382 profile->refCount = 1; |
|
383 uhash_put(SHARED_DATA_HASHTABLE, key.orphan(), profile, status); |
|
384 } |
|
385 umtx_unlock(&usprepMutex); |
|
386 } |
|
387 |
|
388 return profile; |
|
389 } |
|
390 |
|
391 U_CAPI UStringPrepProfile* U_EXPORT2 |
|
392 usprep_open(const char* path, |
|
393 const char* name, |
|
394 UErrorCode* status){ |
|
395 |
|
396 if(status == NULL || U_FAILURE(*status)){ |
|
397 return NULL; |
|
398 } |
|
399 |
|
400 /* initialize the profile struct members */ |
|
401 return usprep_getProfile(path,name,status); |
|
402 } |
|
403 |
|
404 U_CAPI UStringPrepProfile* U_EXPORT2 |
|
405 usprep_openByType(UStringPrepProfileType type, |
|
406 UErrorCode* status) { |
|
407 if(status == NULL || U_FAILURE(*status)){ |
|
408 return NULL; |
|
409 } |
|
410 int32_t index = (int32_t)type; |
|
411 if (index < 0 || index >= (int32_t)(sizeof(PROFILE_NAMES)/sizeof(PROFILE_NAMES[0]))) { |
|
412 *status = U_ILLEGAL_ARGUMENT_ERROR; |
|
413 return NULL; |
|
414 } |
|
415 return usprep_open(NULL, PROFILE_NAMES[index], status); |
|
416 } |
|
417 |
|
418 U_CAPI void U_EXPORT2 |
|
419 usprep_close(UStringPrepProfile* profile){ |
|
420 if(profile==NULL){ |
|
421 return; |
|
422 } |
|
423 |
|
424 umtx_lock(&usprepMutex); |
|
425 /* decrement the ref count*/ |
|
426 if(profile->refCount > 0){ |
|
427 profile->refCount--; |
|
428 } |
|
429 umtx_unlock(&usprepMutex); |
|
430 |
|
431 } |
|
432 |
|
433 U_CFUNC void |
|
434 uprv_syntaxError(const UChar* rules, |
|
435 int32_t pos, |
|
436 int32_t rulesLen, |
|
437 UParseError* parseError){ |
|
438 if(parseError == NULL){ |
|
439 return; |
|
440 } |
|
441 parseError->offset = pos; |
|
442 parseError->line = 0 ; // we are not using line numbers |
|
443 |
|
444 // for pre-context |
|
445 int32_t start = (pos < U_PARSE_CONTEXT_LEN)? 0 : (pos - (U_PARSE_CONTEXT_LEN-1)); |
|
446 int32_t limit = pos; |
|
447 |
|
448 u_memcpy(parseError->preContext,rules+start,limit-start); |
|
449 //null terminate the buffer |
|
450 parseError->preContext[limit-start] = 0; |
|
451 |
|
452 // for post-context; include error rules[pos] |
|
453 start = pos; |
|
454 limit = start + (U_PARSE_CONTEXT_LEN-1); |
|
455 if (limit > rulesLen) { |
|
456 limit = rulesLen; |
|
457 } |
|
458 if (start < rulesLen) { |
|
459 u_memcpy(parseError->postContext,rules+start,limit-start); |
|
460 } |
|
461 //null terminate the buffer |
|
462 parseError->postContext[limit-start]= 0; |
|
463 } |
|
464 |
|
465 |
|
466 static inline UStringPrepType |
|
467 getValues(uint16_t trieWord, int16_t& value, UBool& isIndex){ |
|
468 |
|
469 UStringPrepType type; |
|
470 if(trieWord == 0){ |
|
471 /* |
|
472 * Initial value stored in the mapping table |
|
473 * just return USPREP_TYPE_LIMIT .. so that |
|
474 * the source codepoint is copied to the destination |
|
475 */ |
|
476 type = USPREP_TYPE_LIMIT; |
|
477 isIndex =FALSE; |
|
478 value = 0; |
|
479 }else if(trieWord >= _SPREP_TYPE_THRESHOLD){ |
|
480 type = (UStringPrepType) (trieWord - _SPREP_TYPE_THRESHOLD); |
|
481 isIndex =FALSE; |
|
482 value = 0; |
|
483 }else{ |
|
484 /* get the type */ |
|
485 type = USPREP_MAP; |
|
486 /* ascertain if the value is index or delta */ |
|
487 if(trieWord & 0x02){ |
|
488 isIndex = TRUE; |
|
489 value = trieWord >> 2; //mask off the lower 2 bits and shift |
|
490 }else{ |
|
491 isIndex = FALSE; |
|
492 value = (int16_t)trieWord; |
|
493 value = (value >> 2); |
|
494 } |
|
495 |
|
496 if((trieWord>>2) == _SPREP_MAX_INDEX_VALUE){ |
|
497 type = USPREP_DELETE; |
|
498 isIndex =FALSE; |
|
499 value = 0; |
|
500 } |
|
501 } |
|
502 return type; |
|
503 } |
|
504 |
|
505 |
|
506 |
|
507 static int32_t |
|
508 usprep_map( const UStringPrepProfile* profile, |
|
509 const UChar* src, int32_t srcLength, |
|
510 UChar* dest, int32_t destCapacity, |
|
511 int32_t options, |
|
512 UParseError* parseError, |
|
513 UErrorCode* status ){ |
|
514 |
|
515 uint16_t result; |
|
516 int32_t destIndex=0; |
|
517 int32_t srcIndex; |
|
518 UBool allowUnassigned = (UBool) ((options & USPREP_ALLOW_UNASSIGNED)>0); |
|
519 UStringPrepType type; |
|
520 int16_t value; |
|
521 UBool isIndex; |
|
522 const int32_t* indexes = profile->indexes; |
|
523 |
|
524 // no error checking the caller check for error and arguments |
|
525 // no string length check the caller finds out the string length |
|
526 |
|
527 for(srcIndex=0;srcIndex<srcLength;){ |
|
528 UChar32 ch; |
|
529 |
|
530 U16_NEXT(src,srcIndex,srcLength,ch); |
|
531 |
|
532 result=0; |
|
533 |
|
534 UTRIE_GET16(&profile->sprepTrie,ch,result); |
|
535 |
|
536 type = getValues(result, value, isIndex); |
|
537 |
|
538 // check if the source codepoint is unassigned |
|
539 if(type == USPREP_UNASSIGNED && allowUnassigned == FALSE){ |
|
540 |
|
541 uprv_syntaxError(src,srcIndex-U16_LENGTH(ch), srcLength,parseError); |
|
542 *status = U_STRINGPREP_UNASSIGNED_ERROR; |
|
543 return 0; |
|
544 |
|
545 }else if(type == USPREP_MAP){ |
|
546 |
|
547 int32_t index, length; |
|
548 |
|
549 if(isIndex){ |
|
550 index = value; |
|
551 if(index >= indexes[_SPREP_ONE_UCHAR_MAPPING_INDEX_START] && |
|
552 index < indexes[_SPREP_TWO_UCHARS_MAPPING_INDEX_START]){ |
|
553 length = 1; |
|
554 }else if(index >= indexes[_SPREP_TWO_UCHARS_MAPPING_INDEX_START] && |
|
555 index < indexes[_SPREP_THREE_UCHARS_MAPPING_INDEX_START]){ |
|
556 length = 2; |
|
557 }else if(index >= indexes[_SPREP_THREE_UCHARS_MAPPING_INDEX_START] && |
|
558 index < indexes[_SPREP_FOUR_UCHARS_MAPPING_INDEX_START]){ |
|
559 length = 3; |
|
560 }else{ |
|
561 length = profile->mappingData[index++]; |
|
562 |
|
563 } |
|
564 |
|
565 /* copy mapping to destination */ |
|
566 for(int32_t i=0; i< length; i++){ |
|
567 if(destIndex < destCapacity ){ |
|
568 dest[destIndex] = profile->mappingData[index+i]; |
|
569 } |
|
570 destIndex++; /* for pre-flighting */ |
|
571 } |
|
572 continue; |
|
573 }else{ |
|
574 // subtract the delta to arrive at the code point |
|
575 ch -= value; |
|
576 } |
|
577 |
|
578 }else if(type==USPREP_DELETE){ |
|
579 // just consume the codepoint and contine |
|
580 continue; |
|
581 } |
|
582 //copy the code point into destination |
|
583 if(ch <= 0xFFFF){ |
|
584 if(destIndex < destCapacity ){ |
|
585 dest[destIndex] = (UChar)ch; |
|
586 } |
|
587 destIndex++; |
|
588 }else{ |
|
589 if(destIndex+1 < destCapacity ){ |
|
590 dest[destIndex] = U16_LEAD(ch); |
|
591 dest[destIndex+1] = U16_TRAIL(ch); |
|
592 } |
|
593 destIndex +=2; |
|
594 } |
|
595 |
|
596 } |
|
597 |
|
598 return u_terminateUChars(dest, destCapacity, destIndex, status); |
|
599 } |
|
600 |
|
601 |
|
602 static int32_t |
|
603 usprep_normalize( const UChar* src, int32_t srcLength, |
|
604 UChar* dest, int32_t destCapacity, |
|
605 UErrorCode* status ){ |
|
606 return unorm_normalize( |
|
607 src, srcLength, |
|
608 UNORM_NFKC, UNORM_UNICODE_3_2, |
|
609 dest, destCapacity, |
|
610 status); |
|
611 } |
|
612 |
|
613 |
|
614 /* |
|
615 1) Map -- For each character in the input, check if it has a mapping |
|
616 and, if so, replace it with its mapping. |
|
617 |
|
618 2) Normalize -- Possibly normalize the result of step 1 using Unicode |
|
619 normalization. |
|
620 |
|
621 3) Prohibit -- Check for any characters that are not allowed in the |
|
622 output. If any are found, return an error. |
|
623 |
|
624 4) Check bidi -- Possibly check for right-to-left characters, and if |
|
625 any are found, make sure that the whole string satisfies the |
|
626 requirements for bidirectional strings. If the string does not |
|
627 satisfy the requirements for bidirectional strings, return an |
|
628 error. |
|
629 [Unicode3.2] defines several bidirectional categories; each character |
|
630 has one bidirectional category assigned to it. For the purposes of |
|
631 the requirements below, an "RandALCat character" is a character that |
|
632 has Unicode bidirectional categories "R" or "AL"; an "LCat character" |
|
633 is a character that has Unicode bidirectional category "L". Note |
|
634 |
|
635 |
|
636 that there are many characters which fall in neither of the above |
|
637 definitions; Latin digits (<U+0030> through <U+0039>) are examples of |
|
638 this because they have bidirectional category "EN". |
|
639 |
|
640 In any profile that specifies bidirectional character handling, all |
|
641 three of the following requirements MUST be met: |
|
642 |
|
643 1) The characters in section 5.8 MUST be prohibited. |
|
644 |
|
645 2) If a string contains any RandALCat character, the string MUST NOT |
|
646 contain any LCat character. |
|
647 |
|
648 3) If a string contains any RandALCat character, a RandALCat |
|
649 character MUST be the first character of the string, and a |
|
650 RandALCat character MUST be the last character of the string. |
|
651 */ |
|
652 |
|
653 #define MAX_STACK_BUFFER_SIZE 300 |
|
654 |
|
655 |
|
656 U_CAPI int32_t U_EXPORT2 |
|
657 usprep_prepare( const UStringPrepProfile* profile, |
|
658 const UChar* src, int32_t srcLength, |
|
659 UChar* dest, int32_t destCapacity, |
|
660 int32_t options, |
|
661 UParseError* parseError, |
|
662 UErrorCode* status ){ |
|
663 |
|
664 // check error status |
|
665 if(status == NULL || U_FAILURE(*status)){ |
|
666 return 0; |
|
667 } |
|
668 |
|
669 //check arguments |
|
670 if(profile==NULL || src==NULL || srcLength<-1 || (dest==NULL && destCapacity!=0)) { |
|
671 *status=U_ILLEGAL_ARGUMENT_ERROR; |
|
672 return 0; |
|
673 } |
|
674 |
|
675 UChar b1Stack[MAX_STACK_BUFFER_SIZE], b2Stack[MAX_STACK_BUFFER_SIZE]; |
|
676 UChar *b1 = b1Stack, *b2 = b2Stack; |
|
677 int32_t b1Len, b2Len=0, |
|
678 b1Capacity = MAX_STACK_BUFFER_SIZE , |
|
679 b2Capacity = MAX_STACK_BUFFER_SIZE; |
|
680 uint16_t result; |
|
681 int32_t b2Index = 0; |
|
682 UCharDirection direction=U_CHAR_DIRECTION_COUNT, firstCharDir=U_CHAR_DIRECTION_COUNT; |
|
683 UBool leftToRight=FALSE, rightToLeft=FALSE; |
|
684 int32_t rtlPos =-1, ltrPos =-1; |
|
685 |
|
686 //get the string length |
|
687 if(srcLength == -1){ |
|
688 srcLength = u_strlen(src); |
|
689 } |
|
690 // map |
|
691 b1Len = usprep_map(profile, src, srcLength, b1, b1Capacity, options, parseError, status); |
|
692 |
|
693 if(*status == U_BUFFER_OVERFLOW_ERROR){ |
|
694 // redo processing of string |
|
695 /* we do not have enough room so grow the buffer*/ |
|
696 b1 = (UChar*) uprv_malloc(b1Len * U_SIZEOF_UCHAR); |
|
697 if(b1==NULL){ |
|
698 *status = U_MEMORY_ALLOCATION_ERROR; |
|
699 goto CLEANUP; |
|
700 } |
|
701 |
|
702 *status = U_ZERO_ERROR; // reset error |
|
703 |
|
704 b1Len = usprep_map(profile, src, srcLength, b1, b1Len, options, parseError, status); |
|
705 |
|
706 } |
|
707 |
|
708 // normalize |
|
709 if(profile->doNFKC == TRUE){ |
|
710 b2Len = usprep_normalize(b1,b1Len, b2,b2Capacity,status); |
|
711 |
|
712 if(*status == U_BUFFER_OVERFLOW_ERROR){ |
|
713 // redo processing of string |
|
714 /* we do not have enough room so grow the buffer*/ |
|
715 b2 = (UChar*) uprv_malloc(b2Len * U_SIZEOF_UCHAR); |
|
716 if(b2==NULL){ |
|
717 *status = U_MEMORY_ALLOCATION_ERROR; |
|
718 goto CLEANUP; |
|
719 } |
|
720 |
|
721 *status = U_ZERO_ERROR; // reset error |
|
722 |
|
723 b2Len = usprep_normalize(b1,b1Len, b2,b2Len,status); |
|
724 |
|
725 } |
|
726 |
|
727 }else{ |
|
728 b2 = b1; |
|
729 b2Len = b1Len; |
|
730 } |
|
731 |
|
732 |
|
733 if(U_FAILURE(*status)){ |
|
734 goto CLEANUP; |
|
735 } |
|
736 |
|
737 UChar32 ch; |
|
738 UStringPrepType type; |
|
739 int16_t value; |
|
740 UBool isIndex; |
|
741 |
|
742 // Prohibit and checkBiDi in one pass |
|
743 for(b2Index=0; b2Index<b2Len;){ |
|
744 |
|
745 ch = 0; |
|
746 |
|
747 U16_NEXT(b2, b2Index, b2Len, ch); |
|
748 |
|
749 UTRIE_GET16(&profile->sprepTrie,ch,result); |
|
750 |
|
751 type = getValues(result, value, isIndex); |
|
752 |
|
753 if( type == USPREP_PROHIBITED || |
|
754 ((result < _SPREP_TYPE_THRESHOLD) && (result & 0x01) /* first bit says it the code point is prohibited*/) |
|
755 ){ |
|
756 *status = U_STRINGPREP_PROHIBITED_ERROR; |
|
757 uprv_syntaxError(b1, b2Index-U16_LENGTH(ch), b2Len, parseError); |
|
758 goto CLEANUP; |
|
759 } |
|
760 |
|
761 if(profile->checkBiDi) { |
|
762 direction = ubidi_getClass(profile->bdp, ch); |
|
763 if(firstCharDir == U_CHAR_DIRECTION_COUNT){ |
|
764 firstCharDir = direction; |
|
765 } |
|
766 if(direction == U_LEFT_TO_RIGHT){ |
|
767 leftToRight = TRUE; |
|
768 ltrPos = b2Index-1; |
|
769 } |
|
770 if(direction == U_RIGHT_TO_LEFT || direction == U_RIGHT_TO_LEFT_ARABIC){ |
|
771 rightToLeft = TRUE; |
|
772 rtlPos = b2Index-1; |
|
773 } |
|
774 } |
|
775 } |
|
776 if(profile->checkBiDi == TRUE){ |
|
777 // satisfy 2 |
|
778 if( leftToRight == TRUE && rightToLeft == TRUE){ |
|
779 *status = U_STRINGPREP_CHECK_BIDI_ERROR; |
|
780 uprv_syntaxError(b2,(rtlPos>ltrPos) ? rtlPos : ltrPos, b2Len, parseError); |
|
781 goto CLEANUP; |
|
782 } |
|
783 |
|
784 //satisfy 3 |
|
785 if( rightToLeft == TRUE && |
|
786 !((firstCharDir == U_RIGHT_TO_LEFT || firstCharDir == U_RIGHT_TO_LEFT_ARABIC) && |
|
787 (direction == U_RIGHT_TO_LEFT || direction == U_RIGHT_TO_LEFT_ARABIC)) |
|
788 ){ |
|
789 *status = U_STRINGPREP_CHECK_BIDI_ERROR; |
|
790 uprv_syntaxError(b2, rtlPos, b2Len, parseError); |
|
791 return FALSE; |
|
792 } |
|
793 } |
|
794 if(b2Len>0 && b2Len <= destCapacity){ |
|
795 uprv_memmove(dest,b2, b2Len*U_SIZEOF_UCHAR); |
|
796 } |
|
797 |
|
798 CLEANUP: |
|
799 if(b1!=b1Stack){ |
|
800 uprv_free(b1); |
|
801 b1=NULL; |
|
802 } |
|
803 |
|
804 if(b2!=b1Stack && b2!=b2Stack && b2!=b1 /* b1 should not be freed twice */){ |
|
805 uprv_free(b2); |
|
806 b2=NULL; |
|
807 } |
|
808 return u_terminateUChars(dest, destCapacity, b2Len, status); |
|
809 } |
|
810 |
|
811 |
|
812 /* data swapping ------------------------------------------------------------ */ |
|
813 |
|
814 U_CAPI int32_t U_EXPORT2 |
|
815 usprep_swap(const UDataSwapper *ds, |
|
816 const void *inData, int32_t length, void *outData, |
|
817 UErrorCode *pErrorCode) { |
|
818 const UDataInfo *pInfo; |
|
819 int32_t headerSize; |
|
820 |
|
821 const uint8_t *inBytes; |
|
822 uint8_t *outBytes; |
|
823 |
|
824 const int32_t *inIndexes; |
|
825 int32_t indexes[16]; |
|
826 |
|
827 int32_t i, offset, count, size; |
|
828 |
|
829 /* udata_swapDataHeader checks the arguments */ |
|
830 headerSize=udata_swapDataHeader(ds, inData, length, outData, pErrorCode); |
|
831 if(pErrorCode==NULL || U_FAILURE(*pErrorCode)) { |
|
832 return 0; |
|
833 } |
|
834 |
|
835 /* check data format and format version */ |
|
836 pInfo=(const UDataInfo *)((const char *)inData+4); |
|
837 if(!( |
|
838 pInfo->dataFormat[0]==0x53 && /* dataFormat="SPRP" */ |
|
839 pInfo->dataFormat[1]==0x50 && |
|
840 pInfo->dataFormat[2]==0x52 && |
|
841 pInfo->dataFormat[3]==0x50 && |
|
842 pInfo->formatVersion[0]==3 |
|
843 )) { |
|
844 udata_printError(ds, "usprep_swap(): data format %02x.%02x.%02x.%02x (format version %02x) is not recognized as StringPrep .spp data\n", |
|
845 pInfo->dataFormat[0], pInfo->dataFormat[1], |
|
846 pInfo->dataFormat[2], pInfo->dataFormat[3], |
|
847 pInfo->formatVersion[0]); |
|
848 *pErrorCode=U_UNSUPPORTED_ERROR; |
|
849 return 0; |
|
850 } |
|
851 |
|
852 inBytes=(const uint8_t *)inData+headerSize; |
|
853 outBytes=(uint8_t *)outData+headerSize; |
|
854 |
|
855 inIndexes=(const int32_t *)inBytes; |
|
856 |
|
857 if(length>=0) { |
|
858 length-=headerSize; |
|
859 if(length<16*4) { |
|
860 udata_printError(ds, "usprep_swap(): too few bytes (%d after header) for StringPrep .spp data\n", |
|
861 length); |
|
862 *pErrorCode=U_INDEX_OUTOFBOUNDS_ERROR; |
|
863 return 0; |
|
864 } |
|
865 } |
|
866 |
|
867 /* read the first 16 indexes (ICU 2.8/format version 3: _SPREP_INDEX_TOP==16, might grow) */ |
|
868 for(i=0; i<16; ++i) { |
|
869 indexes[i]=udata_readInt32(ds, inIndexes[i]); |
|
870 } |
|
871 |
|
872 /* calculate the total length of the data */ |
|
873 size= |
|
874 16*4+ /* size of indexes[] */ |
|
875 indexes[_SPREP_INDEX_TRIE_SIZE]+ |
|
876 indexes[_SPREP_INDEX_MAPPING_DATA_SIZE]; |
|
877 |
|
878 if(length>=0) { |
|
879 if(length<size) { |
|
880 udata_printError(ds, "usprep_swap(): too few bytes (%d after header) for all of StringPrep .spp data\n", |
|
881 length); |
|
882 *pErrorCode=U_INDEX_OUTOFBOUNDS_ERROR; |
|
883 return 0; |
|
884 } |
|
885 |
|
886 /* copy the data for inaccessible bytes */ |
|
887 if(inBytes!=outBytes) { |
|
888 uprv_memcpy(outBytes, inBytes, size); |
|
889 } |
|
890 |
|
891 offset=0; |
|
892 |
|
893 /* swap the int32_t indexes[] */ |
|
894 count=16*4; |
|
895 ds->swapArray32(ds, inBytes, count, outBytes, pErrorCode); |
|
896 offset+=count; |
|
897 |
|
898 /* swap the UTrie */ |
|
899 count=indexes[_SPREP_INDEX_TRIE_SIZE]; |
|
900 utrie_swap(ds, inBytes+offset, count, outBytes+offset, pErrorCode); |
|
901 offset+=count; |
|
902 |
|
903 /* swap the uint16_t mappingTable[] */ |
|
904 count=indexes[_SPREP_INDEX_MAPPING_DATA_SIZE]; |
|
905 ds->swapArray16(ds, inBytes+offset, count, outBytes+offset, pErrorCode); |
|
906 offset+=count; |
|
907 } |
|
908 |
|
909 return headerSize+size; |
|
910 } |
|
911 |
|
912 #endif /* #if !UCONFIG_NO_IDNA */ |