1 // Copyright (C) 2016 and later: Unicode, Inc. and others. 2 // License & terms of use: http://www.unicode.org/copyright.html 3 /* 4 ******************************************************************************* 5 * Copyright (C) 2010-2015, International Business Machines 6 * Corporation and others. All Rights Reserved. 7 ******************************************************************************* 8 * file name: uts46.cpp 9 * encoding: US-ASCII 10 * tab size: 8 (not used) 11 * indentation:4 12 * 13 * created on: 2010mar09 14 * created by: Markus W. Scherer 15 */ 16 17 #include "unicode/utypes.h" 18 19 #if !UCONFIG_NO_IDNA 20 21 #include "unicode/idna.h" 22 #include "unicode/normalizer2.h" 23 #include "unicode/uscript.h" 24 #include "unicode/ustring.h" 25 #include "unicode/utf16.h" 26 #include "cmemory.h" 27 #include "cstring.h" 28 #include "punycode.h" 29 #include "ubidi_props.h" 30 #include "ustr_imp.h" 31 32 // Note about tests for UIDNA_ERROR_DOMAIN_NAME_TOO_LONG: 33 // 34 // The domain name length limit is 255 octets in an internal DNS representation 35 // where the last ("root") label is the empty label 36 // represented by length byte 0 alone. 37 // In a conventional string, this translates to 253 characters, or 254 38 // if there is a trailing dot for the root label. 39 40 U_NAMESPACE_BEGIN 41 42 // Severe errors which usually result in a U+FFFD replacement character in the result string. 43 const uint32_t severeErrors= 44 UIDNA_ERROR_LEADING_COMBINING_MARK| 45 UIDNA_ERROR_DISALLOWED| 46 UIDNA_ERROR_PUNYCODE| 47 UIDNA_ERROR_LABEL_HAS_DOT| 48 UIDNA_ERROR_INVALID_ACE_LABEL; 49 50 static inline UBool 51 isASCIIString(const UnicodeString &dest) { 52 const UChar *s=dest.getBuffer(); 53 const UChar *limit=s+dest.length(); 54 while(s<limit) { 55 if(*s++>0x7f) { 56 return FALSE; 57 } 58 } 59 return TRUE; 60 } 61 62 static UBool 63 isASCIIOkBiDi(const UChar *s, int32_t length); 64 65 static UBool 66 isASCIIOkBiDi(const char *s, int32_t length); 67 68 // IDNA class default implementations -------------------------------------- *** 69 70 IDNA::~IDNA() {} 71 72 void 73 IDNA::labelToASCII_UTF8(StringPiece label, ByteSink &dest, 74 IDNAInfo &info, UErrorCode &errorCode) const { 75 if(U_SUCCESS(errorCode)) { 76 UnicodeString destString; 77 labelToASCII(UnicodeString::fromUTF8(label), destString, 78 info, errorCode).toUTF8(dest); 79 } 80 } 81 82 void 83 IDNA::labelToUnicodeUTF8(StringPiece label, ByteSink &dest, 84 IDNAInfo &info, UErrorCode &errorCode) const { 85 if(U_SUCCESS(errorCode)) { 86 UnicodeString destString; 87 labelToUnicode(UnicodeString::fromUTF8(label), destString, 88 info, errorCode).toUTF8(dest); 89 } 90 } 91 92 void 93 IDNA::nameToASCII_UTF8(StringPiece name, ByteSink &dest, 94 IDNAInfo &info, UErrorCode &errorCode) const { 95 if(U_SUCCESS(errorCode)) { 96 UnicodeString destString; 97 nameToASCII(UnicodeString::fromUTF8(name), destString, 98 info, errorCode).toUTF8(dest); 99 } 100 } 101 102 void 103 IDNA::nameToUnicodeUTF8(StringPiece name, ByteSink &dest, 104 IDNAInfo &info, UErrorCode &errorCode) const { 105 if(U_SUCCESS(errorCode)) { 106 UnicodeString destString; 107 nameToUnicode(UnicodeString::fromUTF8(name), destString, 108 info, errorCode).toUTF8(dest); 109 } 110 } 111 112 // UTS46 class declaration ------------------------------------------------- *** 113 114 class UTS46 : public IDNA { 115 public: 116 UTS46(uint32_t options, UErrorCode &errorCode); 117 virtual ~UTS46(); 118 119 virtual UnicodeString & 120 labelToASCII(const UnicodeString &label, UnicodeString &dest, 121 IDNAInfo &info, UErrorCode &errorCode) const; 122 123 virtual UnicodeString & 124 labelToUnicode(const UnicodeString &label, UnicodeString &dest, 125 IDNAInfo &info, UErrorCode &errorCode) const; 126 127 virtual UnicodeString & 128 nameToASCII(const UnicodeString &name, UnicodeString &dest, 129 IDNAInfo &info, UErrorCode &errorCode) const; 130 131 virtual UnicodeString & 132 nameToUnicode(const UnicodeString &name, UnicodeString &dest, 133 IDNAInfo &info, UErrorCode &errorCode) const; 134 135 virtual void 136 labelToASCII_UTF8(StringPiece label, ByteSink &dest, 137 IDNAInfo &info, UErrorCode &errorCode) const; 138 139 virtual void 140 labelToUnicodeUTF8(StringPiece label, ByteSink &dest, 141 IDNAInfo &info, UErrorCode &errorCode) const; 142 143 virtual void 144 nameToASCII_UTF8(StringPiece name, ByteSink &dest, 145 IDNAInfo &info, UErrorCode &errorCode) const; 146 147 virtual void 148 nameToUnicodeUTF8(StringPiece name, ByteSink &dest, 149 IDNAInfo &info, UErrorCode &errorCode) const; 150 151 private: 152 UnicodeString & 153 process(const UnicodeString &src, 154 UBool isLabel, UBool toASCII, 155 UnicodeString &dest, 156 IDNAInfo &info, UErrorCode &errorCode) const; 157 158 void 159 processUTF8(StringPiece src, 160 UBool isLabel, UBool toASCII, 161 ByteSink &dest, 162 IDNAInfo &info, UErrorCode &errorCode) const; 163 164 UnicodeString & 165 processUnicode(const UnicodeString &src, 166 int32_t labelStart, int32_t mappingStart, 167 UBool isLabel, UBool toASCII, 168 UnicodeString &dest, 169 IDNAInfo &info, UErrorCode &errorCode) const; 170 171 // returns the new dest.length() 172 int32_t 173 mapDevChars(UnicodeString &dest, int32_t labelStart, int32_t mappingStart, 174 UErrorCode &errorCode) const; 175 176 // returns the new label length 177 int32_t 178 processLabel(UnicodeString &dest, 179 int32_t labelStart, int32_t labelLength, 180 UBool toASCII, 181 IDNAInfo &info, UErrorCode &errorCode) const; 182 int32_t 183 markBadACELabel(UnicodeString &dest, 184 int32_t labelStart, int32_t labelLength, 185 UBool toASCII, IDNAInfo &info, UErrorCode &errorCode) const; 186 187 void 188 checkLabelBiDi(const UChar *label, int32_t labelLength, IDNAInfo &info) const; 189 190 UBool 191 isLabelOkContextJ(const UChar *label, int32_t labelLength) const; 192 193 void 194 checkLabelContextO(const UChar *label, int32_t labelLength, IDNAInfo &info) const; 195 196 const Normalizer2 &uts46Norm2; // uts46.nrm 197 uint32_t options; 198 }; 199 200 IDNA * 201 IDNA::createUTS46Instance(uint32_t options, UErrorCode &errorCode) { 202 if(U_SUCCESS(errorCode)) { 203 IDNA *idna=new UTS46(options, errorCode); 204 if(idna==NULL) { 205 errorCode=U_MEMORY_ALLOCATION_ERROR; 206 } else if(U_FAILURE(errorCode)) { 207 delete idna; 208 idna=NULL; 209 } 210 return idna; 211 } else { 212 return NULL; 213 } 214 } 215 216 // UTS46 implementation ---------------------------------------------------- *** 217 218 UTS46::UTS46(uint32_t opt, UErrorCode &errorCode) 219 : uts46Norm2(*Normalizer2::getInstance(NULL, "uts46", UNORM2_COMPOSE, errorCode)), 220 options(opt) {} 221 222 UTS46::~UTS46() {} 223 224 UnicodeString & 225 UTS46::labelToASCII(const UnicodeString &label, UnicodeString &dest, 226 IDNAInfo &info, UErrorCode &errorCode) const { 227 return process(label, TRUE, TRUE, dest, info, errorCode); 228 } 229 230 UnicodeString & 231 UTS46::labelToUnicode(const UnicodeString &label, UnicodeString &dest, 232 IDNAInfo &info, UErrorCode &errorCode) const { 233 return process(label, TRUE, FALSE, dest, info, errorCode); 234 } 235 236 UnicodeString & 237 UTS46::nameToASCII(const UnicodeString &name, UnicodeString &dest, 238 IDNAInfo &info, UErrorCode &errorCode) const { 239 process(name, FALSE, TRUE, dest, info, errorCode); 240 if( dest.length()>=254 && (info.errors&UIDNA_ERROR_DOMAIN_NAME_TOO_LONG)==0 && 241 isASCIIString(dest) && 242 (dest.length()>254 || dest[253]!=0x2e) 243 ) { 244 info.errors|=UIDNA_ERROR_DOMAIN_NAME_TOO_LONG; 245 } 246 return dest; 247 } 248 249 UnicodeString & 250 UTS46::nameToUnicode(const UnicodeString &name, UnicodeString &dest, 251 IDNAInfo &info, UErrorCode &errorCode) const { 252 return process(name, FALSE, FALSE, dest, info, errorCode); 253 } 254 255 void 256 UTS46::labelToASCII_UTF8(StringPiece label, ByteSink &dest, 257 IDNAInfo &info, UErrorCode &errorCode) const { 258 processUTF8(label, TRUE, TRUE, dest, info, errorCode); 259 } 260 261 void 262 UTS46::labelToUnicodeUTF8(StringPiece label, ByteSink &dest, 263 IDNAInfo &info, UErrorCode &errorCode) const { 264 processUTF8(label, TRUE, FALSE, dest, info, errorCode); 265 } 266 267 void 268 UTS46::nameToASCII_UTF8(StringPiece name, ByteSink &dest, 269 IDNAInfo &info, UErrorCode &errorCode) const { 270 processUTF8(name, FALSE, TRUE, dest, info, errorCode); 271 } 272 273 void 274 UTS46::nameToUnicodeUTF8(StringPiece name, ByteSink &dest, 275 IDNAInfo &info, UErrorCode &errorCode) const { 276 processUTF8(name, FALSE, FALSE, dest, info, errorCode); 277 } 278 279 // UTS #46 data for ASCII characters. 280 // The normalizer (using uts46.nrm) maps uppercase ASCII letters to lowercase 281 // and passes through all other ASCII characters. 282 // If UIDNA_USE_STD3_RULES is set, then non-LDH characters are disallowed 283 // using this data. 284 // The ASCII fastpath also uses this data. 285 // Values: -1=disallowed 0==valid 1==mapped (lowercase) 286 static const int8_t asciiData[128]={ 287 -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, 288 -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, 289 // 002D..002E; valid # HYPHEN-MINUS..FULL STOP 290 -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, 0, 0, -1, 291 // 0030..0039; valid # DIGIT ZERO..DIGIT NINE 292 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, -1, -1, -1, -1, -1, -1, 293 // 0041..005A; mapped # LATIN CAPITAL LETTER A..LATIN CAPITAL LETTER Z 294 -1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 295 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, -1, -1, -1, -1, -1, 296 // 0061..007A; valid # LATIN SMALL LETTER A..LATIN SMALL LETTER Z 297 -1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 298 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, -1, -1, -1, -1, -1 299 }; 300 301 UnicodeString & 302 UTS46::process(const UnicodeString &src, 303 UBool isLabel, UBool toASCII, 304 UnicodeString &dest, 305 IDNAInfo &info, UErrorCode &errorCode) const { 306 // uts46Norm2.normalize() would do all of this error checking and setup, 307 // but with the ASCII fastpath we do not always call it, and do not 308 // call it first. 309 if(U_FAILURE(errorCode)) { 310 dest.setToBogus(); 311 return dest; 312 } 313 const UChar *srcArray=src.getBuffer(); 314 if(&dest==&src || srcArray==NULL) { 315 errorCode=U_ILLEGAL_ARGUMENT_ERROR; 316 dest.setToBogus(); 317 return dest; 318 } 319 // Arguments are fine, reset output values. 320 dest.remove(); 321 info.reset(); 322 int32_t srcLength=src.length(); 323 if(srcLength==0) { 324 info.errors|=UIDNA_ERROR_EMPTY_LABEL; 325 return dest; 326 } 327 UChar *destArray=dest.getBuffer(srcLength); 328 if(destArray==NULL) { 329 errorCode=U_MEMORY_ALLOCATION_ERROR; 330 return dest; 331 } 332 // ASCII fastpath 333 UBool disallowNonLDHDot=(options&UIDNA_USE_STD3_RULES)!=0; 334 int32_t labelStart=0; 335 int32_t i; 336 for(i=0;; ++i) { 337 if(i==srcLength) { 338 if(toASCII) { 339 if((i-labelStart)>63) { 340 info.labelErrors|=UIDNA_ERROR_LABEL_TOO_LONG; 341 } 342 // There is a trailing dot if labelStart==i. 343 if(!isLabel && i>=254 && (i>254 || labelStart<i)) { 344 info.errors|=UIDNA_ERROR_DOMAIN_NAME_TOO_LONG; 345 } 346 } 347 info.errors|=info.labelErrors; 348 dest.releaseBuffer(i); 349 return dest; 350 } 351 UChar c=srcArray[i]; 352 if(c>0x7f) { 353 break; 354 } 355 int cData=asciiData[c]; 356 if(cData>0) { 357 destArray[i]=c+0x20; // Lowercase an uppercase ASCII letter. 358 } else if(cData<0 && disallowNonLDHDot) { 359 break; // Replacing with U+FFFD can be complicated for toASCII. 360 } else { 361 destArray[i]=c; 362 if(c==0x2d) { // hyphen 363 if(i==(labelStart+3) && srcArray[i-1]==0x2d) { 364 // "??--..." is Punycode or forbidden. 365 ++i; // '-' was copied to dest already 366 break; 367 } 368 if(i==labelStart) { 369 // label starts with "-" 370 info.labelErrors|=UIDNA_ERROR_LEADING_HYPHEN; 371 } 372 if((i+1)==srcLength || srcArray[i+1]==0x2e) { 373 // label ends with "-" 374 info.labelErrors|=UIDNA_ERROR_TRAILING_HYPHEN; 375 } 376 } else if(c==0x2e) { // dot 377 if(isLabel) { 378 // Replacing with U+FFFD can be complicated for toASCII. 379 ++i; // '.' was copied to dest already 380 break; 381 } 382 if(i==labelStart) { 383 info.labelErrors|=UIDNA_ERROR_EMPTY_LABEL; 384 } 385 if(toASCII && (i-labelStart)>63) { 386 info.labelErrors|=UIDNA_ERROR_LABEL_TOO_LONG; 387 } 388 info.errors|=info.labelErrors; 389 info.labelErrors=0; 390 labelStart=i+1; 391 } 392 } 393 } 394 info.errors|=info.labelErrors; 395 dest.releaseBuffer(i); 396 processUnicode(src, labelStart, i, isLabel, toASCII, dest, info, errorCode); 397 if( info.isBiDi && U_SUCCESS(errorCode) && (info.errors&severeErrors)==0 && 398 (!info.isOkBiDi || (labelStart>0 && !isASCIIOkBiDi(dest.getBuffer(), labelStart))) 399 ) { 400 info.errors|=UIDNA_ERROR_BIDI; 401 } 402 return dest; 403 } 404 405 void 406 UTS46::processUTF8(StringPiece src, 407 UBool isLabel, UBool toASCII, 408 ByteSink &dest, 409 IDNAInfo &info, UErrorCode &errorCode) const { 410 if(U_FAILURE(errorCode)) { 411 return; 412 } 413 const char *srcArray=src.data(); 414 int32_t srcLength=src.length(); 415 if(srcArray==NULL && srcLength!=0) { 416 errorCode=U_ILLEGAL_ARGUMENT_ERROR; 417 return; 418 } 419 // Arguments are fine, reset output values. 420 info.reset(); 421 if(srcLength==0) { 422 info.errors|=UIDNA_ERROR_EMPTY_LABEL; 423 dest.Flush(); 424 return; 425 } 426 UnicodeString destString; 427 int32_t labelStart=0; 428 if(srcLength<=256) { // length of stackArray[] 429 // ASCII fastpath 430 char stackArray[256]; 431 int32_t destCapacity; 432 char *destArray=dest.GetAppendBuffer(srcLength, srcLength+20, 433 stackArray, UPRV_LENGTHOF(stackArray), &destCapacity); 434 UBool disallowNonLDHDot=(options&UIDNA_USE_STD3_RULES)!=0; 435 int32_t i; 436 for(i=0;; ++i) { 437 if(i==srcLength) { 438 if(toASCII) { 439 if((i-labelStart)>63) { 440 info.labelErrors|=UIDNA_ERROR_LABEL_TOO_LONG; 441 } 442 // There is a trailing dot if labelStart==i. 443 if(!isLabel && i>=254 && (i>254 || labelStart<i)) { 444 info.errors|=UIDNA_ERROR_DOMAIN_NAME_TOO_LONG; 445 } 446 } 447 info.errors|=info.labelErrors; 448 dest.Append(destArray, i); 449 dest.Flush(); 450 return; 451 } 452 char c=srcArray[i]; 453 if((int8_t)c<0) { // (uint8_t)c>0x7f 454 break; 455 } 456 int cData=asciiData[(int)c]; // Cast: gcc warns about indexing with a char. 457 if(cData>0) { 458 destArray[i]=c+0x20; // Lowercase an uppercase ASCII letter. 459 } else if(cData<0 && disallowNonLDHDot) { 460 break; // Replacing with U+FFFD can be complicated for toASCII. 461 } else { 462 destArray[i]=c; 463 if(c==0x2d) { // hyphen 464 if(i==(labelStart+3) && srcArray[i-1]==0x2d) { 465 // "??--..." is Punycode or forbidden. 466 break; 467 } 468 if(i==labelStart) { 469 // label starts with "-" 470 info.labelErrors|=UIDNA_ERROR_LEADING_HYPHEN; 471 } 472 if((i+1)==srcLength || srcArray[i+1]==0x2e) { 473 // label ends with "-" 474 info.labelErrors|=UIDNA_ERROR_TRAILING_HYPHEN; 475 } 476 } else if(c==0x2e) { // dot 477 if(isLabel) { 478 break; // Replacing with U+FFFD can be complicated for toASCII. 479 } 480 if(i==labelStart) { 481 info.labelErrors|=UIDNA_ERROR_EMPTY_LABEL; 482 } 483 if(toASCII && (i-labelStart)>63) { 484 info.labelErrors|=UIDNA_ERROR_LABEL_TOO_LONG; 485 } 486 info.errors|=info.labelErrors; 487 info.labelErrors=0; 488 labelStart=i+1; 489 } 490 } 491 } 492 info.errors|=info.labelErrors; 493 // Convert the processed ASCII prefix of the current label to UTF-16. 494 int32_t mappingStart=i-labelStart; 495 destString=UnicodeString::fromUTF8(StringPiece(destArray+labelStart, mappingStart)); 496 // Output the previous ASCII labels and process the rest of src in UTF-16. 497 dest.Append(destArray, labelStart); 498 processUnicode(UnicodeString::fromUTF8(StringPiece(src, labelStart)), 0, mappingStart, 499 isLabel, toASCII, 500 destString, info, errorCode); 501 } else { 502 // src is too long for the ASCII fastpath implementation. 503 processUnicode(UnicodeString::fromUTF8(src), 0, 0, 504 isLabel, toASCII, 505 destString, info, errorCode); 506 } 507 destString.toUTF8(dest); // calls dest.Flush() 508 if(toASCII && !isLabel) { 509 // length==labelStart==254 means that there is a trailing dot (ok) and 510 // destString is empty (do not index at 253-labelStart). 511 int32_t length=labelStart+destString.length(); 512 if( length>=254 && isASCIIString(destString) && 513 (length>254 || 514 (labelStart<254 && destString[253-labelStart]!=0x2e)) 515 ) { 516 info.errors|=UIDNA_ERROR_DOMAIN_NAME_TOO_LONG; 517 } 518 } 519 if( info.isBiDi && U_SUCCESS(errorCode) && (info.errors&severeErrors)==0 && 520 (!info.isOkBiDi || (labelStart>0 && !isASCIIOkBiDi(srcArray, labelStart))) 521 ) { 522 info.errors|=UIDNA_ERROR_BIDI; 523 } 524 } 525 526 UnicodeString & 527 UTS46::processUnicode(const UnicodeString &src, 528 int32_t labelStart, int32_t mappingStart, 529 UBool isLabel, UBool toASCII, 530 UnicodeString &dest, 531 IDNAInfo &info, UErrorCode &errorCode) const { 532 if(mappingStart==0) { 533 uts46Norm2.normalize(src, dest, errorCode); 534 } else { 535 uts46Norm2.normalizeSecondAndAppend(dest, src.tempSubString(mappingStart), errorCode); 536 } 537 if(U_FAILURE(errorCode)) { 538 return dest; 539 } 540 UBool doMapDevChars= 541 toASCII ? (options&UIDNA_NONTRANSITIONAL_TO_ASCII)==0 : 542 (options&UIDNA_NONTRANSITIONAL_TO_UNICODE)==0; 543 const UChar *destArray=dest.getBuffer(); 544 int32_t destLength=dest.length(); 545 int32_t labelLimit=labelStart; 546 while(labelLimit<destLength) { 547 UChar c=destArray[labelLimit]; 548 if(c==0x2e && !isLabel) { 549 int32_t labelLength=labelLimit-labelStart; 550 int32_t newLength=processLabel(dest, labelStart, labelLength, 551 toASCII, info, errorCode); 552 info.errors|=info.labelErrors; 553 info.labelErrors=0; 554 if(U_FAILURE(errorCode)) { 555 return dest; 556 } 557 destArray=dest.getBuffer(); 558 destLength+=newLength-labelLength; 559 labelLimit=labelStart+=newLength+1; 560 } else if(0xdf<=c && c<=0x200d && (c==0xdf || c==0x3c2 || c>=0x200c)) { 561 info.isTransDiff=TRUE; 562 if(doMapDevChars) { 563 destLength=mapDevChars(dest, labelStart, labelLimit, errorCode); 564 if(U_FAILURE(errorCode)) { 565 return dest; 566 } 567 destArray=dest.getBuffer(); 568 // Do not increment labelLimit in case c was removed. 569 // All deviation characters have been mapped, no need to check for them again. 570 doMapDevChars=FALSE; 571 } else { 572 ++labelLimit; 573 } 574 } else { 575 ++labelLimit; 576 } 577 } 578 // Permit an empty label at the end (0<labelStart==labelLimit==destLength is ok) 579 // but not an empty label elsewhere nor a completely empty domain name. 580 // processLabel() sets UIDNA_ERROR_EMPTY_LABEL when labelLength==0. 581 if(0==labelStart || labelStart<labelLimit) { 582 processLabel(dest, labelStart, labelLimit-labelStart, 583 toASCII, info, errorCode); 584 info.errors|=info.labelErrors; 585 } 586 return dest; 587 } 588 589 int32_t 590 UTS46::mapDevChars(UnicodeString &dest, int32_t labelStart, int32_t mappingStart, 591 UErrorCode &errorCode) const { 592 if(U_FAILURE(errorCode)) { 593 return 0; 594 } 595 int32_t length=dest.length(); 596 UChar *s=dest.getBuffer(dest[mappingStart]==0xdf ? length+1 : length); 597 if(s==NULL) { 598 errorCode=U_MEMORY_ALLOCATION_ERROR; 599 return length; 600 } 601 int32_t capacity=dest.getCapacity(); 602 UBool didMapDevChars=FALSE; 603 int32_t readIndex=mappingStart, writeIndex=mappingStart; 604 do { 605 UChar c=s[readIndex++]; 606 switch(c) { 607 case 0xdf: 608 // Map sharp s to ss. 609 didMapDevChars=TRUE; 610 s[writeIndex++]=0x73; // Replace sharp s with first s. 611 // Insert second s and account for possible buffer reallocation. 612 if(writeIndex==readIndex) { 613 if(length==capacity) { 614 dest.releaseBuffer(length); 615 s=dest.getBuffer(length+1); 616 if(s==NULL) { 617 errorCode=U_MEMORY_ALLOCATION_ERROR; 618 return length; 619 } 620 capacity=dest.getCapacity(); 621 } 622 u_memmove(s+writeIndex+1, s+writeIndex, length-writeIndex); 623 ++readIndex; 624 } 625 s[writeIndex++]=0x73; 626 ++length; 627 break; 628 case 0x3c2: // Map final sigma to nonfinal sigma. 629 didMapDevChars=TRUE; 630 s[writeIndex++]=0x3c3; 631 break; 632 case 0x200c: // Ignore/remove ZWNJ. 633 case 0x200d: // Ignore/remove ZWJ. 634 didMapDevChars=TRUE; 635 --length; 636 break; 637 default: 638 // Only really necessary if writeIndex was different from readIndex. 639 s[writeIndex++]=c; 640 break; 641 } 642 } while(writeIndex<length); 643 dest.releaseBuffer(length); 644 if(didMapDevChars) { 645 // Mapping deviation characters might have resulted in an un-NFC string. 646 // We could use either the NFC or the UTS #46 normalizer. 647 // By using the UTS #46 normalizer again, we avoid having to load a second .nrm data file. 648 UnicodeString normalized; 649 uts46Norm2.normalize(dest.tempSubString(labelStart), normalized, errorCode); 650 if(U_SUCCESS(errorCode)) { 651 dest.replace(labelStart, 0x7fffffff, normalized); 652 if(dest.isBogus()) { 653 errorCode=U_MEMORY_ALLOCATION_ERROR; 654 } 655 return dest.length(); 656 } 657 } 658 return length; 659 } 660 661 // Some non-ASCII characters are equivalent to sequences with 662 // non-LDH ASCII characters. To find them: 663 // grep disallowed_STD3_valid IdnaMappingTable.txt (or uts46.txt) 664 static inline UBool 665 isNonASCIIDisallowedSTD3Valid(UChar32 c) { 666 return c==0x2260 || c==0x226E || c==0x226F; 667 } 668 669 // Replace the label in dest with the label string, if the label was modified. 670 // If &label==&dest then the label was modified in-place and labelLength 671 // is the new label length, different from label.length(). 672 // If &label!=&dest then labelLength==label.length(). 673 // Returns labelLength (= the new label length). 674 static int32_t 675 replaceLabel(UnicodeString &dest, int32_t destLabelStart, int32_t destLabelLength, 676 const UnicodeString &label, int32_t labelLength, UErrorCode &errorCode) { 677 if(U_FAILURE(errorCode)) { 678 return 0; 679 } 680 if(&label!=&dest) { 681 dest.replace(destLabelStart, destLabelLength, label); 682 if(dest.isBogus()) { 683 errorCode=U_MEMORY_ALLOCATION_ERROR; 684 return 0; 685 } 686 } 687 return labelLength; 688 } 689 690 int32_t 691 UTS46::processLabel(UnicodeString &dest, 692 int32_t labelStart, int32_t labelLength, 693 UBool toASCII, 694 IDNAInfo &info, UErrorCode &errorCode) const { 695 if(U_FAILURE(errorCode)) { 696 return 0; 697 } 698 UnicodeString fromPunycode; 699 UnicodeString *labelString; 700 const UChar *label=dest.getBuffer()+labelStart; 701 int32_t destLabelStart=labelStart; 702 int32_t destLabelLength=labelLength; 703 UBool wasPunycode; 704 if(labelLength>=4 && label[0]==0x78 && label[1]==0x6e && label[2]==0x2d && label[3]==0x2d) { 705 // Label starts with "xn--", try to un-Punycode it. 706 wasPunycode=TRUE; 707 UChar *unicodeBuffer=fromPunycode.getBuffer(-1); // capacity==-1: most labels should fit 708 if(unicodeBuffer==NULL) { 709 // Should never occur if we used capacity==-1 which uses the internal buffer. 710 errorCode=U_MEMORY_ALLOCATION_ERROR; 711 return labelLength; 712 } 713 UErrorCode punycodeErrorCode=U_ZERO_ERROR; 714 int32_t unicodeLength=u_strFromPunycode(label+4, labelLength-4, 715 unicodeBuffer, fromPunycode.getCapacity(), 716 NULL, &punycodeErrorCode); 717 if(punycodeErrorCode==U_BUFFER_OVERFLOW_ERROR) { 718 fromPunycode.releaseBuffer(0); 719 unicodeBuffer=fromPunycode.getBuffer(unicodeLength); 720 if(unicodeBuffer==NULL) { 721 errorCode=U_MEMORY_ALLOCATION_ERROR; 722 return labelLength; 723 } 724 punycodeErrorCode=U_ZERO_ERROR; 725 unicodeLength=u_strFromPunycode(label+4, labelLength-4, 726 unicodeBuffer, fromPunycode.getCapacity(), 727 NULL, &punycodeErrorCode); 728 } 729 fromPunycode.releaseBuffer(unicodeLength); 730 if(U_FAILURE(punycodeErrorCode)) { 731 info.labelErrors|=UIDNA_ERROR_PUNYCODE; 732 return markBadACELabel(dest, labelStart, labelLength, toASCII, info, errorCode); 733 } 734 // Check for NFC, and for characters that are not 735 // valid or deviation characters according to the normalizer. 736 // If there is something wrong, then the string will change. 737 // Note that the normalizer passes through non-LDH ASCII and deviation characters. 738 // Deviation characters are ok in Punycode even in transitional processing. 739 // In the code further below, if we find non-LDH ASCII and we have UIDNA_USE_STD3_RULES 740 // then we will set UIDNA_ERROR_INVALID_ACE_LABEL there too. 741 UBool isValid=uts46Norm2.isNormalized(fromPunycode, errorCode); 742 if(U_FAILURE(errorCode)) { 743 return labelLength; 744 } 745 if(!isValid) { 746 info.labelErrors|=UIDNA_ERROR_INVALID_ACE_LABEL; 747 return markBadACELabel(dest, labelStart, labelLength, toASCII, info, errorCode); 748 } 749 labelString=&fromPunycode; 750 label=fromPunycode.getBuffer(); 751 labelStart=0; 752 labelLength=fromPunycode.length(); 753 } else { 754 wasPunycode=FALSE; 755 labelString=&dest; 756 } 757 // Validity check 758 if(labelLength==0) { 759 info.labelErrors|=UIDNA_ERROR_EMPTY_LABEL; 760 return replaceLabel(dest, destLabelStart, destLabelLength, 761 *labelString, labelLength, errorCode); 762 } 763 // labelLength>0 764 if(labelLength>=4 && label[2]==0x2d && label[3]==0x2d) { 765 // label starts with "??--" 766 info.labelErrors|=UIDNA_ERROR_HYPHEN_3_4; 767 } 768 if(label[0]==0x2d) { 769 // label starts with "-" 770 info.labelErrors|=UIDNA_ERROR_LEADING_HYPHEN; 771 } 772 if(label[labelLength-1]==0x2d) { 773 // label ends with "-" 774 info.labelErrors|=UIDNA_ERROR_TRAILING_HYPHEN; 775 } 776 // If the label was not a Punycode label, then it was the result of 777 // mapping, normalization and label segmentation. 778 // If the label was in Punycode, then we mapped it again above 779 // and checked its validity. 780 // Now we handle the STD3 restriction to LDH characters (if set) 781 // and we look for U+FFFD which indicates disallowed characters 782 // in a non-Punycode label or U+FFFD itself in a Punycode label. 783 // We also check for dots which can come from the input to a single-label function. 784 // Ok to cast away const because we own the UnicodeString. 785 UChar *s=(UChar *)label; 786 const UChar *limit=label+labelLength; 787 UChar oredChars=0; 788 // If we enforce STD3 rules, then ASCII characters other than LDH and dot are disallowed. 789 UBool disallowNonLDHDot=(options&UIDNA_USE_STD3_RULES)!=0; 790 do { 791 UChar c=*s; 792 if(c<=0x7f) { 793 if(c==0x2e) { 794 info.labelErrors|=UIDNA_ERROR_LABEL_HAS_DOT; 795 *s=0xfffd; 796 } else if(disallowNonLDHDot && asciiData[c]<0) { 797 info.labelErrors|=UIDNA_ERROR_DISALLOWED; 798 *s=0xfffd; 799 } 800 } else { 801 oredChars|=c; 802 if(disallowNonLDHDot && isNonASCIIDisallowedSTD3Valid(c)) { 803 info.labelErrors|=UIDNA_ERROR_DISALLOWED; 804 *s=0xfffd; 805 } else if(c==0xfffd) { 806 info.labelErrors|=UIDNA_ERROR_DISALLOWED; 807 } 808 } 809 ++s; 810 } while(s<limit); 811 // Check for a leading combining mark after other validity checks 812 // so that we don't report UIDNA_ERROR_DISALLOWED for the U+FFFD from here. 813 UChar32 c; 814 int32_t cpLength=0; 815 // "Unsafe" is ok because unpaired surrogates were mapped to U+FFFD. 816 U16_NEXT_UNSAFE(label, cpLength, c); 817 if((U_GET_GC_MASK(c)&U_GC_M_MASK)!=0) { 818 info.labelErrors|=UIDNA_ERROR_LEADING_COMBINING_MARK; 819 labelString->replace(labelStart, cpLength, (UChar)0xfffd); 820 label=labelString->getBuffer()+labelStart; 821 labelLength+=1-cpLength; 822 if(labelString==&dest) { 823 destLabelLength=labelLength; 824 } 825 } 826 if((info.labelErrors&severeErrors)==0) { 827 // Do contextual checks only if we do not have U+FFFD from a severe error 828 // because U+FFFD can make these checks fail. 829 if((options&UIDNA_CHECK_BIDI)!=0 && (!info.isBiDi || info.isOkBiDi)) { 830 checkLabelBiDi(label, labelLength, info); 831 } 832 if( (options&UIDNA_CHECK_CONTEXTJ)!=0 && (oredChars&0x200c)==0x200c && 833 !isLabelOkContextJ(label, labelLength) 834 ) { 835 info.labelErrors|=UIDNA_ERROR_CONTEXTJ; 836 } 837 if((options&UIDNA_CHECK_CONTEXTO)!=0 && oredChars>=0xb7) { 838 checkLabelContextO(label, labelLength, info); 839 } 840 if(toASCII) { 841 if(wasPunycode) { 842 // Leave a Punycode label unchanged if it has no severe errors. 843 if(destLabelLength>63) { 844 info.labelErrors|=UIDNA_ERROR_LABEL_TOO_LONG; 845 } 846 return destLabelLength; 847 } else if(oredChars>=0x80) { 848 // Contains non-ASCII characters. 849 UnicodeString punycode; 850 UChar *buffer=punycode.getBuffer(63); // 63==maximum DNS label length 851 if(buffer==NULL) { 852 errorCode=U_MEMORY_ALLOCATION_ERROR; 853 return destLabelLength; 854 } 855 buffer[0]=0x78; // Write "xn--". 856 buffer[1]=0x6e; 857 buffer[2]=0x2d; 858 buffer[3]=0x2d; 859 int32_t punycodeLength=u_strToPunycode(label, labelLength, 860 buffer+4, punycode.getCapacity()-4, 861 NULL, &errorCode); 862 if(errorCode==U_BUFFER_OVERFLOW_ERROR) { 863 errorCode=U_ZERO_ERROR; 864 punycode.releaseBuffer(4); 865 buffer=punycode.getBuffer(4+punycodeLength); 866 if(buffer==NULL) { 867 errorCode=U_MEMORY_ALLOCATION_ERROR; 868 return destLabelLength; 869 } 870 punycodeLength=u_strToPunycode(label, labelLength, 871 buffer+4, punycode.getCapacity()-4, 872 NULL, &errorCode); 873 } 874 punycodeLength+=4; 875 punycode.releaseBuffer(punycodeLength); 876 if(U_FAILURE(errorCode)) { 877 return destLabelLength; 878 } 879 if(punycodeLength>63) { 880 info.labelErrors|=UIDNA_ERROR_LABEL_TOO_LONG; 881 } 882 return replaceLabel(dest, destLabelStart, destLabelLength, 883 punycode, punycodeLength, errorCode); 884 } else { 885 // all-ASCII label 886 if(labelLength>63) { 887 info.labelErrors|=UIDNA_ERROR_LABEL_TOO_LONG; 888 } 889 } 890 } 891 } else { 892 // If a Punycode label has severe errors, 893 // then leave it but make sure it does not look valid. 894 if(wasPunycode) { 895 info.labelErrors|=UIDNA_ERROR_INVALID_ACE_LABEL; 896 return markBadACELabel(dest, destLabelStart, destLabelLength, toASCII, info, errorCode); 897 } 898 } 899 return replaceLabel(dest, destLabelStart, destLabelLength, 900 *labelString, labelLength, errorCode); 901 } 902 903 // Make sure an ACE label does not look valid. 904 // Append U+FFFD if the label has only LDH characters. 905 // If UIDNA_USE_STD3_RULES, also replace disallowed ASCII characters with U+FFFD. 906 int32_t 907 UTS46::markBadACELabel(UnicodeString &dest, 908 int32_t labelStart, int32_t labelLength, 909 UBool toASCII, IDNAInfo &info, UErrorCode &errorCode) const { 910 if(U_FAILURE(errorCode)) { 911 return 0; 912 } 913 UBool disallowNonLDHDot=(options&UIDNA_USE_STD3_RULES)!=0; 914 UBool isASCII=TRUE; 915 UBool onlyLDH=TRUE; 916 const UChar *label=dest.getBuffer()+labelStart; 917 // Ok to cast away const because we own the UnicodeString. 918 UChar *s=(UChar *)label+4; // After the initial "xn--". 919 const UChar *limit=label+labelLength; 920 do { 921 UChar c=*s; 922 if(c<=0x7f) { 923 if(c==0x2e) { 924 info.labelErrors|=UIDNA_ERROR_LABEL_HAS_DOT; 925 *s=0xfffd; 926 isASCII=onlyLDH=FALSE; 927 } else if(asciiData[c]<0) { 928 onlyLDH=FALSE; 929 if(disallowNonLDHDot) { 930 *s=0xfffd; 931 isASCII=FALSE; 932 } 933 } 934 } else { 935 isASCII=onlyLDH=FALSE; 936 } 937 } while(++s<limit); 938 if(onlyLDH) { 939 dest.insert(labelStart+labelLength, (UChar)0xfffd); 940 if(dest.isBogus()) { 941 errorCode=U_MEMORY_ALLOCATION_ERROR; 942 return 0; 943 } 944 ++labelLength; 945 } else { 946 if(toASCII && isASCII && labelLength>63) { 947 info.labelErrors|=UIDNA_ERROR_LABEL_TOO_LONG; 948 } 949 } 950 return labelLength; 951 } 952 953 const uint32_t L_MASK=U_MASK(U_LEFT_TO_RIGHT); 954 const uint32_t R_AL_MASK=U_MASK(U_RIGHT_TO_LEFT)|U_MASK(U_RIGHT_TO_LEFT_ARABIC); 955 const uint32_t L_R_AL_MASK=L_MASK|R_AL_MASK; 956 957 const uint32_t R_AL_AN_MASK=R_AL_MASK|U_MASK(U_ARABIC_NUMBER); 958 959 const uint32_t EN_AN_MASK=U_MASK(U_EUROPEAN_NUMBER)|U_MASK(U_ARABIC_NUMBER); 960 const uint32_t R_AL_EN_AN_MASK=R_AL_MASK|EN_AN_MASK; 961 const uint32_t L_EN_MASK=L_MASK|U_MASK(U_EUROPEAN_NUMBER); 962 963 const uint32_t ES_CS_ET_ON_BN_NSM_MASK= 964 U_MASK(U_EUROPEAN_NUMBER_SEPARATOR)| 965 U_MASK(U_COMMON_NUMBER_SEPARATOR)| 966 U_MASK(U_EUROPEAN_NUMBER_TERMINATOR)| 967 U_MASK(U_OTHER_NEUTRAL)| 968 U_MASK(U_BOUNDARY_NEUTRAL)| 969 U_MASK(U_DIR_NON_SPACING_MARK); 970 const uint32_t L_EN_ES_CS_ET_ON_BN_NSM_MASK=L_EN_MASK|ES_CS_ET_ON_BN_NSM_MASK; 971 const uint32_t R_AL_AN_EN_ES_CS_ET_ON_BN_NSM_MASK=R_AL_MASK|EN_AN_MASK|ES_CS_ET_ON_BN_NSM_MASK; 972 973 // We scan the whole label and check both for whether it contains RTL characters 974 // and whether it passes the BiDi Rule. 975 // In a BiDi domain name, all labels must pass the BiDi Rule, but we might find 976 // that a domain name is a BiDi domain name (has an RTL label) only after 977 // processing several earlier labels. 978 void 979 UTS46::checkLabelBiDi(const UChar *label, int32_t labelLength, IDNAInfo &info) const { 980 // IDNA2008 BiDi rule 981 // Get the directionality of the first character. 982 UChar32 c; 983 int32_t i=0; 984 U16_NEXT_UNSAFE(label, i, c); 985 uint32_t firstMask=U_MASK(u_charDirection(c)); 986 // 1. The first character must be a character with BIDI property L, R 987 // or AL. If it has the R or AL property, it is an RTL label; if it 988 // has the L property, it is an LTR label. 989 if((firstMask&~L_R_AL_MASK)!=0) { 990 info.isOkBiDi=FALSE; 991 } 992 // Get the directionality of the last non-NSM character. 993 uint32_t lastMask; 994 for(;;) { 995 if(i>=labelLength) { 996 lastMask=firstMask; 997 break; 998 } 999 U16_PREV_UNSAFE(label, labelLength, c); 1000 UCharDirection dir=u_charDirection(c); 1001 if(dir!=U_DIR_NON_SPACING_MARK) { 1002 lastMask=U_MASK(dir); 1003 break; 1004 } 1005 } 1006 // 3. In an RTL label, the end of the label must be a character with 1007 // BIDI property R, AL, EN or AN, followed by zero or more 1008 // characters with BIDI property NSM. 1009 // 6. In an LTR label, the end of the label must be a character with 1010 // BIDI property L or EN, followed by zero or more characters with 1011 // BIDI property NSM. 1012 if( (firstMask&L_MASK)!=0 ? 1013 (lastMask&~L_EN_MASK)!=0 : 1014 (lastMask&~R_AL_EN_AN_MASK)!=0 1015 ) { 1016 info.isOkBiDi=FALSE; 1017 } 1018 // Get the directionalities of the intervening characters. 1019 uint32_t mask=0; 1020 while(i<labelLength) { 1021 U16_NEXT_UNSAFE(label, i, c); 1022 mask|=U_MASK(u_charDirection(c)); 1023 } 1024 if(firstMask&L_MASK) { 1025 // 5. In an LTR label, only characters with the BIDI properties L, EN, 1026 // ES, CS, ET, ON, BN and NSM are allowed. 1027 if((mask&~L_EN_ES_CS_ET_ON_BN_NSM_MASK)!=0) { 1028 info.isOkBiDi=FALSE; 1029 } 1030 } else { 1031 // 2. In an RTL label, only characters with the BIDI properties R, AL, 1032 // AN, EN, ES, CS, ET, ON, BN and NSM are allowed. 1033 if((mask&~R_AL_AN_EN_ES_CS_ET_ON_BN_NSM_MASK)!=0) { 1034 info.isOkBiDi=FALSE; 1035 } 1036 // 4. In an RTL label, if an EN is present, no AN may be present, and 1037 // vice versa. 1038 if((mask&EN_AN_MASK)==EN_AN_MASK) { 1039 info.isOkBiDi=FALSE; 1040 } 1041 } 1042 // An RTL label is a label that contains at least one character of type 1043 // R, AL or AN. [...] 1044 // A "BIDI domain name" is a domain name that contains at least one RTL 1045 // label. [...] 1046 // The following rule, consisting of six conditions, applies to labels 1047 // in BIDI domain names. 1048 if(((firstMask|mask|lastMask)&R_AL_AN_MASK)!=0) { 1049 info.isBiDi=TRUE; 1050 } 1051 } 1052 1053 // Special code for the ASCII prefix of a BiDi domain name. 1054 // The ASCII prefix is all-LTR. 1055 1056 // IDNA2008 BiDi rule, parts relevant to ASCII labels: 1057 // 1. The first character must be a character with BIDI property L [...] 1058 // 5. In an LTR label, only characters with the BIDI properties L, EN, 1059 // ES, CS, ET, ON, BN and NSM are allowed. 1060 // 6. In an LTR label, the end of the label must be a character with 1061 // BIDI property L or EN [...] 1062 1063 // UTF-16 version, called for mapped ASCII prefix. 1064 // Cannot contain uppercase A-Z. 1065 // s[length-1] must be the trailing dot. 1066 static UBool 1067 isASCIIOkBiDi(const UChar *s, int32_t length) { 1068 int32_t labelStart=0; 1069 for(int32_t i=0; i<length; ++i) { 1070 UChar c=s[i]; 1071 if(c==0x2e) { // dot 1072 if(i>labelStart) { 1073 c=s[i-1]; 1074 if(!(0x61<=c && c<=0x7a) && !(0x30<=c && c<=0x39)) { 1075 // Last character in the label is not an L or EN. 1076 return FALSE; 1077 } 1078 } 1079 labelStart=i+1; 1080 } else if(i==labelStart) { 1081 if(!(0x61<=c && c<=0x7a)) { 1082 // First character in the label is not an L. 1083 return FALSE; 1084 } 1085 } else { 1086 if(c<=0x20 && (c>=0x1c || (9<=c && c<=0xd))) { 1087 // Intermediate character in the label is a B, S or WS. 1088 return FALSE; 1089 } 1090 } 1091 } 1092 return TRUE; 1093 } 1094 1095 // UTF-8 version, called for source ASCII prefix. 1096 // Can contain uppercase A-Z. 1097 // s[length-1] must be the trailing dot. 1098 static UBool 1099 isASCIIOkBiDi(const char *s, int32_t length) { 1100 int32_t labelStart=0; 1101 for(int32_t i=0; i<length; ++i) { 1102 char c=s[i]; 1103 if(c==0x2e) { // dot 1104 if(i>labelStart) { 1105 c=s[i-1]; 1106 if(!(0x61<=c && c<=0x7a) && !(0x41<=c && c<=0x5a) && !(0x30<=c && c<=0x39)) { 1107 // Last character in the label is not an L or EN. 1108 return FALSE; 1109 } 1110 } 1111 labelStart=i+1; 1112 } else if(i==labelStart) { 1113 if(!(0x61<=c && c<=0x7a) && !(0x41<=c && c<=0x5a)) { 1114 // First character in the label is not an L. 1115 return FALSE; 1116 } 1117 } else { 1118 if(c<=0x20 && (c>=0x1c || (9<=c && c<=0xd))) { 1119 // Intermediate character in the label is a B, S or WS. 1120 return FALSE; 1121 } 1122 } 1123 } 1124 return TRUE; 1125 } 1126 1127 UBool 1128 UTS46::isLabelOkContextJ(const UChar *label, int32_t labelLength) const { 1129 const UBiDiProps *bdp=ubidi_getSingleton(); 1130 // [IDNA2008-Tables] 1131 // 200C..200D ; CONTEXTJ # ZERO WIDTH NON-JOINER..ZERO WIDTH JOINER 1132 for(int32_t i=0; i<labelLength; ++i) { 1133 if(label[i]==0x200c) { 1134 // Appendix A.1. ZERO WIDTH NON-JOINER 1135 // Rule Set: 1136 // False; 1137 // If Canonical_Combining_Class(Before(cp)) .eq. Virama Then True; 1138 // If RegExpMatch((Joining_Type:{L,D})(Joining_Type:T)*\u200C 1139 // (Joining_Type:T)*(Joining_Type:{R,D})) Then True; 1140 if(i==0) { 1141 return FALSE; 1142 } 1143 UChar32 c; 1144 int32_t j=i; 1145 U16_PREV_UNSAFE(label, j, c); 1146 if(uts46Norm2.getCombiningClass(c)==9) { 1147 continue; 1148 } 1149 // check precontext (Joining_Type:{L,D})(Joining_Type:T)* 1150 for(;;) { 1151 UJoiningType type=ubidi_getJoiningType(bdp, c); 1152 if(type==U_JT_TRANSPARENT) { 1153 if(j==0) { 1154 return FALSE; 1155 } 1156 U16_PREV_UNSAFE(label, j, c); 1157 } else if(type==U_JT_LEFT_JOINING || type==U_JT_DUAL_JOINING) { 1158 break; // precontext fulfilled 1159 } else { 1160 return FALSE; 1161 } 1162 } 1163 // check postcontext (Joining_Type:T)*(Joining_Type:{R,D}) 1164 for(j=i+1;;) { 1165 if(j==labelLength) { 1166 return FALSE; 1167 } 1168 U16_NEXT_UNSAFE(label, j, c); 1169 UJoiningType type=ubidi_getJoiningType(bdp, c); 1170 if(type==U_JT_TRANSPARENT) { 1171 // just skip this character 1172 } else if(type==U_JT_RIGHT_JOINING || type==U_JT_DUAL_JOINING) { 1173 break; // postcontext fulfilled 1174 } else { 1175 return FALSE; 1176 } 1177 } 1178 } else if(label[i]==0x200d) { 1179 // Appendix A.2. ZERO WIDTH JOINER (U+200D) 1180 // Rule Set: 1181 // False; 1182 // If Canonical_Combining_Class(Before(cp)) .eq. Virama Then True; 1183 if(i==0) { 1184 return FALSE; 1185 } 1186 UChar32 c; 1187 int32_t j=i; 1188 U16_PREV_UNSAFE(label, j, c); 1189 if(uts46Norm2.getCombiningClass(c)!=9) { 1190 return FALSE; 1191 } 1192 } 1193 } 1194 return TRUE; 1195 } 1196 1197 void 1198 UTS46::checkLabelContextO(const UChar *label, int32_t labelLength, IDNAInfo &info) const { 1199 int32_t labelEnd=labelLength-1; // inclusive 1200 int32_t arabicDigits=0; // -1 for 066x, +1 for 06Fx 1201 for(int32_t i=0; i<=labelEnd; ++i) { 1202 UChar32 c=label[i]; 1203 if(c<0xb7) { 1204 // ASCII fastpath 1205 } else if(c<=0x6f9) { 1206 if(c==0xb7) { 1207 // Appendix A.3. MIDDLE DOT (U+00B7) 1208 // Rule Set: 1209 // False; 1210 // If Before(cp) .eq. U+006C And 1211 // After(cp) .eq. U+006C Then True; 1212 if(!(0<i && label[i-1]==0x6c && 1213 i<labelEnd && label[i+1]==0x6c)) { 1214 info.labelErrors|=UIDNA_ERROR_CONTEXTO_PUNCTUATION; 1215 } 1216 } else if(c==0x375) { 1217 // Appendix A.4. GREEK LOWER NUMERAL SIGN (KERAIA) (U+0375) 1218 // Rule Set: 1219 // False; 1220 // If Script(After(cp)) .eq. Greek Then True; 1221 UScriptCode script=USCRIPT_INVALID_CODE; 1222 if(i<labelEnd) { 1223 UErrorCode errorCode=U_ZERO_ERROR; 1224 int32_t j=i+1; 1225 U16_NEXT(label, j, labelLength, c); 1226 script=uscript_getScript(c, &errorCode); 1227 } 1228 if(script!=USCRIPT_GREEK) { 1229 info.labelErrors|=UIDNA_ERROR_CONTEXTO_PUNCTUATION; 1230 } 1231 } else if(c==0x5f3 || c==0x5f4) { 1232 // Appendix A.5. HEBREW PUNCTUATION GERESH (U+05F3) 1233 // Rule Set: 1234 // False; 1235 // If Script(Before(cp)) .eq. Hebrew Then True; 1236 // 1237 // Appendix A.6. HEBREW PUNCTUATION GERSHAYIM (U+05F4) 1238 // Rule Set: 1239 // False; 1240 // If Script(Before(cp)) .eq. Hebrew Then True; 1241 UScriptCode script=USCRIPT_INVALID_CODE; 1242 if(0<i) { 1243 UErrorCode errorCode=U_ZERO_ERROR; 1244 int32_t j=i; 1245 U16_PREV(label, 0, j, c); 1246 script=uscript_getScript(c, &errorCode); 1247 } 1248 if(script!=USCRIPT_HEBREW) { 1249 info.labelErrors|=UIDNA_ERROR_CONTEXTO_PUNCTUATION; 1250 } 1251 } else if(0x660<=c /* && c<=0x6f9 */) { 1252 // Appendix A.8. ARABIC-INDIC DIGITS (0660..0669) 1253 // Rule Set: 1254 // True; 1255 // For All Characters: 1256 // If cp .in. 06F0..06F9 Then False; 1257 // End For; 1258 // 1259 // Appendix A.9. EXTENDED ARABIC-INDIC DIGITS (06F0..06F9) 1260 // Rule Set: 1261 // True; 1262 // For All Characters: 1263 // If cp .in. 0660..0669 Then False; 1264 // End For; 1265 if(c<=0x669) { 1266 if(arabicDigits>0) { 1267 info.labelErrors|=UIDNA_ERROR_CONTEXTO_DIGITS; 1268 } 1269 arabicDigits=-1; 1270 } else if(0x6f0<=c) { 1271 if(arabicDigits<0) { 1272 info.labelErrors|=UIDNA_ERROR_CONTEXTO_DIGITS; 1273 } 1274 arabicDigits=1; 1275 } 1276 } 1277 } else if(c==0x30fb) { 1278 // Appendix A.7. KATAKANA MIDDLE DOT (U+30FB) 1279 // Rule Set: 1280 // False; 1281 // For All Characters: 1282 // If Script(cp) .in. {Hiragana, Katakana, Han} Then True; 1283 // End For; 1284 UErrorCode errorCode=U_ZERO_ERROR; 1285 for(int j=0;;) { 1286 if(j>labelEnd) { 1287 info.labelErrors|=UIDNA_ERROR_CONTEXTO_PUNCTUATION; 1288 break; 1289 } 1290 U16_NEXT(label, j, labelLength, c); 1291 UScriptCode script=uscript_getScript(c, &errorCode); 1292 if(script==USCRIPT_HIRAGANA || script==USCRIPT_KATAKANA || script==USCRIPT_HAN) { 1293 break; 1294 } 1295 } 1296 } 1297 } 1298 } 1299 1300 U_NAMESPACE_END 1301 1302 // C API ------------------------------------------------------------------- *** 1303 1304 U_NAMESPACE_USE 1305 1306 U_CAPI UIDNA * U_EXPORT2 1307 uidna_openUTS46(uint32_t options, UErrorCode *pErrorCode) { 1308 return reinterpret_cast<UIDNA *>(IDNA::createUTS46Instance(options, *pErrorCode)); 1309 } 1310 1311 U_CAPI void U_EXPORT2 1312 uidna_close(UIDNA *idna) { 1313 delete reinterpret_cast<IDNA *>(idna); 1314 } 1315 1316 static UBool 1317 checkArgs(const void *label, int32_t length, 1318 void *dest, int32_t capacity, 1319 UIDNAInfo *pInfo, UErrorCode *pErrorCode) { 1320 if(U_FAILURE(*pErrorCode)) { 1321 return FALSE; 1322 } 1323 // sizeof(UIDNAInfo)=16 in the first API version. 1324 if(pInfo==NULL || pInfo->size<16) { 1325 *pErrorCode=U_ILLEGAL_ARGUMENT_ERROR; 1326 return FALSE; 1327 } 1328 if( (label==NULL ? length!=0 : length<-1) || 1329 (dest==NULL ? capacity!=0 : capacity<0) || 1330 (dest==label && label!=NULL) 1331 ) { 1332 *pErrorCode=U_ILLEGAL_ARGUMENT_ERROR; 1333 return FALSE; 1334 } 1335 // Set all *pInfo bytes to 0 except for the size field itself. 1336 uprv_memset(&pInfo->size+1, 0, pInfo->size-sizeof(pInfo->size)); 1337 return TRUE; 1338 } 1339 1340 static void 1341 idnaInfoToStruct(IDNAInfo &info, UIDNAInfo *pInfo) { 1342 pInfo->isTransitionalDifferent=info.isTransitionalDifferent(); 1343 pInfo->errors=info.getErrors(); 1344 } 1345 1346 U_CAPI int32_t U_EXPORT2 1347 uidna_labelToASCII(const UIDNA *idna, 1348 const UChar *label, int32_t length, 1349 UChar *dest, int32_t capacity, 1350 UIDNAInfo *pInfo, UErrorCode *pErrorCode) { 1351 if(!checkArgs(label, length, dest, capacity, pInfo, pErrorCode)) { 1352 return 0; 1353 } 1354 UnicodeString src((UBool)(length<0), label, length); 1355 UnicodeString destString(dest, 0, capacity); 1356 IDNAInfo info; 1357 reinterpret_cast<const IDNA *>(idna)->labelToASCII(src, destString, info, *pErrorCode); 1358 idnaInfoToStruct(info, pInfo); 1359 return destString.extract(dest, capacity, *pErrorCode); 1360 } 1361 1362 U_CAPI int32_t U_EXPORT2 1363 uidna_labelToUnicode(const UIDNA *idna, 1364 const UChar *label, int32_t length, 1365 UChar *dest, int32_t capacity, 1366 UIDNAInfo *pInfo, UErrorCode *pErrorCode) { 1367 if(!checkArgs(label, length, dest, capacity, pInfo, pErrorCode)) { 1368 return 0; 1369 } 1370 UnicodeString src((UBool)(length<0), label, length); 1371 UnicodeString destString(dest, 0, capacity); 1372 IDNAInfo info; 1373 reinterpret_cast<const IDNA *>(idna)->labelToUnicode(src, destString, info, *pErrorCode); 1374 idnaInfoToStruct(info, pInfo); 1375 return destString.extract(dest, capacity, *pErrorCode); 1376 } 1377 1378 U_CAPI int32_t U_EXPORT2 1379 uidna_nameToASCII(const UIDNA *idna, 1380 const UChar *name, int32_t length, 1381 UChar *dest, int32_t capacity, 1382 UIDNAInfo *pInfo, UErrorCode *pErrorCode) { 1383 if(!checkArgs(name, length, dest, capacity, pInfo, pErrorCode)) { 1384 return 0; 1385 } 1386 UnicodeString src((UBool)(length<0), name, length); 1387 UnicodeString destString(dest, 0, capacity); 1388 IDNAInfo info; 1389 reinterpret_cast<const IDNA *>(idna)->nameToASCII(src, destString, info, *pErrorCode); 1390 idnaInfoToStruct(info, pInfo); 1391 return destString.extract(dest, capacity, *pErrorCode); 1392 } 1393 1394 U_CAPI int32_t U_EXPORT2 1395 uidna_nameToUnicode(const UIDNA *idna, 1396 const UChar *name, int32_t length, 1397 UChar *dest, int32_t capacity, 1398 UIDNAInfo *pInfo, UErrorCode *pErrorCode) { 1399 if(!checkArgs(name, length, dest, capacity, pInfo, pErrorCode)) { 1400 return 0; 1401 } 1402 UnicodeString src((UBool)(length<0), name, length); 1403 UnicodeString destString(dest, 0, capacity); 1404 IDNAInfo info; 1405 reinterpret_cast<const IDNA *>(idna)->nameToUnicode(src, destString, info, *pErrorCode); 1406 idnaInfoToStruct(info, pInfo); 1407 return destString.extract(dest, capacity, *pErrorCode); 1408 } 1409 1410 U_CAPI int32_t U_EXPORT2 1411 uidna_labelToASCII_UTF8(const UIDNA *idna, 1412 const char *label, int32_t length, 1413 char *dest, int32_t capacity, 1414 UIDNAInfo *pInfo, UErrorCode *pErrorCode) { 1415 if(!checkArgs(label, length, dest, capacity, pInfo, pErrorCode)) { 1416 return 0; 1417 } 1418 StringPiece src(label, length<0 ? uprv_strlen(label) : length); 1419 CheckedArrayByteSink sink(dest, capacity); 1420 IDNAInfo info; 1421 reinterpret_cast<const IDNA *>(idna)->labelToASCII_UTF8(src, sink, info, *pErrorCode); 1422 idnaInfoToStruct(info, pInfo); 1423 return u_terminateChars(dest, capacity, sink.NumberOfBytesAppended(), pErrorCode); 1424 } 1425 1426 U_CAPI int32_t U_EXPORT2 1427 uidna_labelToUnicodeUTF8(const UIDNA *idna, 1428 const char *label, int32_t length, 1429 char *dest, int32_t capacity, 1430 UIDNAInfo *pInfo, UErrorCode *pErrorCode) { 1431 if(!checkArgs(label, length, dest, capacity, pInfo, pErrorCode)) { 1432 return 0; 1433 } 1434 StringPiece src(label, length<0 ? uprv_strlen(label) : length); 1435 CheckedArrayByteSink sink(dest, capacity); 1436 IDNAInfo info; 1437 reinterpret_cast<const IDNA *>(idna)->labelToUnicodeUTF8(src, sink, info, *pErrorCode); 1438 idnaInfoToStruct(info, pInfo); 1439 return u_terminateChars(dest, capacity, sink.NumberOfBytesAppended(), pErrorCode); 1440 } 1441 1442 U_CAPI int32_t U_EXPORT2 1443 uidna_nameToASCII_UTF8(const UIDNA *idna, 1444 const char *name, int32_t length, 1445 char *dest, int32_t capacity, 1446 UIDNAInfo *pInfo, UErrorCode *pErrorCode) { 1447 if(!checkArgs(name, length, dest, capacity, pInfo, pErrorCode)) { 1448 return 0; 1449 } 1450 StringPiece src(name, length<0 ? uprv_strlen(name) : length); 1451 CheckedArrayByteSink sink(dest, capacity); 1452 IDNAInfo info; 1453 reinterpret_cast<const IDNA *>(idna)->nameToASCII_UTF8(src, sink, info, *pErrorCode); 1454 idnaInfoToStruct(info, pInfo); 1455 return u_terminateChars(dest, capacity, sink.NumberOfBytesAppended(), pErrorCode); 1456 } 1457 1458 U_CAPI int32_t U_EXPORT2 1459 uidna_nameToUnicodeUTF8(const UIDNA *idna, 1460 const char *name, int32_t length, 1461 char *dest, int32_t capacity, 1462 UIDNAInfo *pInfo, UErrorCode *pErrorCode) { 1463 if(!checkArgs(name, length, dest, capacity, pInfo, pErrorCode)) { 1464 return 0; 1465 } 1466 StringPiece src(name, length<0 ? uprv_strlen(name) : length); 1467 CheckedArrayByteSink sink(dest, capacity); 1468 IDNAInfo info; 1469 reinterpret_cast<const IDNA *>(idna)->nameToUnicodeUTF8(src, sink, info, *pErrorCode); 1470 idnaInfoToStruct(info, pInfo); 1471 return u_terminateChars(dest, capacity, sink.NumberOfBytesAppended(), pErrorCode); 1472 } 1473 1474 #endif // UCONFIG_NO_IDNA 1475