| OLD | NEW |
|---|
| (Empty) | |
| 1 /* |
| 2 ********************************************************************** |
| 3 * Copyright (C) 2000-2007, International Business Machines |
| 4 * Corporation and others. All Rights Reserved. |
| 5 ********************************************************************** |
| 6 * file name: ucnvlat1.cpp |
| 7 * encoding: US-ASCII |
| 8 * tab size: 8 (not used) |
| 9 * indentation:4 |
| 10 * |
| 11 * created on: 2000feb07 |
| 12 * created by: Markus W. Scherer |
| 13 */ |
| 14 |
| 15 #include "unicode/utypes.h" |
| 16 |
| 17 #if !UCONFIG_NO_CONVERSION |
| 18 |
| 19 #include "unicode/ucnv.h" |
| 20 #include "unicode/uset.h" |
| 21 #include "ucnv_bld.h" |
| 22 #include "ucnv_cnv.h" |
| 23 |
| 24 /* control optimizations according to the platform */ |
| 25 #define LATIN1_UNROLL_FROM_UNICODE 1 |
| 26 |
| 27 /* ISO 8859-1 --------------------------------------------------------------- */ |
| 28 |
| 29 /* This is a table-less and callback-less version of ucnv_MBCSSingleToBMPWithOff
sets(). */ |
| 30 static void |
| 31 _Latin1ToUnicodeWithOffsets(UConverterToUnicodeArgs *pArgs, |
| 32 UErrorCode *pErrorCode) { |
| 33 const uint8_t *source; |
| 34 UChar *target; |
| 35 int32_t targetCapacity, length; |
| 36 int32_t *offsets; |
| 37 |
| 38 int32_t sourceIndex; |
| 39 |
| 40 /* set up the local pointers */ |
| 41 source=(const uint8_t *)pArgs->source; |
| 42 target=pArgs->target; |
| 43 targetCapacity=(int32_t)(pArgs->targetLimit-pArgs->target); |
| 44 offsets=pArgs->offsets; |
| 45 |
| 46 sourceIndex=0; |
| 47 |
| 48 /* |
| 49 * since the conversion here is 1:1 UChar:uint8_t, we need only one counter |
| 50 * for the minimum of the sourceLength and targetCapacity |
| 51 */ |
| 52 length=(int32_t)((const uint8_t *)pArgs->sourceLimit-source); |
| 53 if(length<=targetCapacity) { |
| 54 targetCapacity=length; |
| 55 } else { |
| 56 /* target will be full */ |
| 57 *pErrorCode=U_BUFFER_OVERFLOW_ERROR; |
| 58 length=targetCapacity; |
| 59 } |
| 60 |
| 61 if(targetCapacity>=8) { |
| 62 /* This loop is unrolled for speed and improved pipelining. */ |
| 63 int32_t count, loops; |
| 64 |
| 65 loops=count=targetCapacity>>3; |
| 66 length=targetCapacity&=0x7; |
| 67 do { |
| 68 target[0]=source[0]; |
| 69 target[1]=source[1]; |
| 70 target[2]=source[2]; |
| 71 target[3]=source[3]; |
| 72 target[4]=source[4]; |
| 73 target[5]=source[5]; |
| 74 target[6]=source[6]; |
| 75 target[7]=source[7]; |
| 76 target+=8; |
| 77 source+=8; |
| 78 } while(--count>0); |
| 79 |
| 80 if(offsets!=NULL) { |
| 81 do { |
| 82 offsets[0]=sourceIndex++; |
| 83 offsets[1]=sourceIndex++; |
| 84 offsets[2]=sourceIndex++; |
| 85 offsets[3]=sourceIndex++; |
| 86 offsets[4]=sourceIndex++; |
| 87 offsets[5]=sourceIndex++; |
| 88 offsets[6]=sourceIndex++; |
| 89 offsets[7]=sourceIndex++; |
| 90 offsets+=8; |
| 91 } while(--loops>0); |
| 92 } |
| 93 } |
| 94 |
| 95 /* conversion loop */ |
| 96 while(targetCapacity>0) { |
| 97 *target++=*source++; |
| 98 --targetCapacity; |
| 99 } |
| 100 |
| 101 /* write back the updated pointers */ |
| 102 pArgs->source=(const char *)source; |
| 103 pArgs->target=target; |
| 104 |
| 105 /* set offsets */ |
| 106 if(offsets!=NULL) { |
| 107 while(length>0) { |
| 108 *offsets++=sourceIndex++; |
| 109 --length; |
| 110 } |
| 111 pArgs->offsets=offsets; |
| 112 } |
| 113 } |
| 114 |
| 115 /* This is a table-less and callback-less version of ucnv_MBCSSingleGetNextUChar
(). */ |
| 116 static UChar32 |
| 117 _Latin1GetNextUChar(UConverterToUnicodeArgs *pArgs, |
| 118 UErrorCode *pErrorCode) { |
| 119 const uint8_t *source=(const uint8_t *)pArgs->source; |
| 120 if(source<(const uint8_t *)pArgs->sourceLimit) { |
| 121 pArgs->source=(const char *)(source+1); |
| 122 return *source; |
| 123 } |
| 124 |
| 125 /* no output because of empty input */ |
| 126 *pErrorCode=U_INDEX_OUTOFBOUNDS_ERROR; |
| 127 return 0xffff; |
| 128 } |
| 129 |
| 130 /* This is a table-less version of ucnv_MBCSSingleFromBMPWithOffsets(). */ |
| 131 static void |
| 132 _Latin1FromUnicodeWithOffsets(UConverterFromUnicodeArgs *pArgs, |
| 133 UErrorCode *pErrorCode) { |
| 134 UConverter *cnv; |
| 135 const UChar *source, *sourceLimit; |
| 136 uint8_t *target, *oldTarget; |
| 137 int32_t targetCapacity, length; |
| 138 int32_t *offsets; |
| 139 |
| 140 UChar32 cp; |
| 141 UChar c, max; |
| 142 |
| 143 int32_t sourceIndex; |
| 144 |
| 145 /* set up the local pointers */ |
| 146 cnv=pArgs->converter; |
| 147 source=pArgs->source; |
| 148 sourceLimit=pArgs->sourceLimit; |
| 149 target=oldTarget=(uint8_t *)pArgs->target; |
| 150 targetCapacity=(int32_t)(pArgs->targetLimit-pArgs->target); |
| 151 offsets=pArgs->offsets; |
| 152 |
| 153 if(cnv->sharedData==&_Latin1Data) { |
| 154 max=0xff; /* Latin-1 */ |
| 155 } else { |
| 156 max=0x7f; /* US-ASCII */ |
| 157 } |
| 158 |
| 159 /* get the converter state from UConverter */ |
| 160 cp=cnv->fromUChar32; |
| 161 |
| 162 /* sourceIndex=-1 if the current character began in the previous buffer */ |
| 163 sourceIndex= cp==0 ? 0 : -1; |
| 164 |
| 165 /* |
| 166 * since the conversion here is 1:1 UChar:uint8_t, we need only one counter |
| 167 * for the minimum of the sourceLength and targetCapacity |
| 168 */ |
| 169 length=(int32_t)(sourceLimit-source); |
| 170 if(length<targetCapacity) { |
| 171 targetCapacity=length; |
| 172 } |
| 173 |
| 174 /* conversion loop */ |
| 175 if(cp!=0 && targetCapacity>0) { |
| 176 goto getTrail; |
| 177 } |
| 178 |
| 179 #if LATIN1_UNROLL_FROM_UNICODE |
| 180 /* unroll the loop with the most common case */ |
| 181 if(targetCapacity>=16) { |
| 182 int32_t count, loops; |
| 183 UChar u, oredChars; |
| 184 |
| 185 loops=count=targetCapacity>>4; |
| 186 do { |
| 187 oredChars=u=*source++; |
| 188 *target++=(uint8_t)u; |
| 189 oredChars|=u=*source++; |
| 190 *target++=(uint8_t)u; |
| 191 oredChars|=u=*source++; |
| 192 *target++=(uint8_t)u; |
| 193 oredChars|=u=*source++; |
| 194 *target++=(uint8_t)u; |
| 195 oredChars|=u=*source++; |
| 196 *target++=(uint8_t)u; |
| 197 oredChars|=u=*source++; |
| 198 *target++=(uint8_t)u; |
| 199 oredChars|=u=*source++; |
| 200 *target++=(uint8_t)u; |
| 201 oredChars|=u=*source++; |
| 202 *target++=(uint8_t)u; |
| 203 oredChars|=u=*source++; |
| 204 *target++=(uint8_t)u; |
| 205 oredChars|=u=*source++; |
| 206 *target++=(uint8_t)u; |
| 207 oredChars|=u=*source++; |
| 208 *target++=(uint8_t)u; |
| 209 oredChars|=u=*source++; |
| 210 *target++=(uint8_t)u; |
| 211 oredChars|=u=*source++; |
| 212 *target++=(uint8_t)u; |
| 213 oredChars|=u=*source++; |
| 214 *target++=(uint8_t)u; |
| 215 oredChars|=u=*source++; |
| 216 *target++=(uint8_t)u; |
| 217 oredChars|=u=*source++; |
| 218 *target++=(uint8_t)u; |
| 219 |
| 220 /* were all 16 entries really valid? */ |
| 221 if(oredChars>max) { |
| 222 /* no, return to the first of these 16 */ |
| 223 source-=16; |
| 224 target-=16; |
| 225 break; |
| 226 } |
| 227 } while(--count>0); |
| 228 count=loops-count; |
| 229 targetCapacity-=16*count; |
| 230 |
| 231 if(offsets!=NULL) { |
| 232 oldTarget+=16*count; |
| 233 while(count>0) { |
| 234 *offsets++=sourceIndex++; |
| 235 *offsets++=sourceIndex++; |
| 236 *offsets++=sourceIndex++; |
| 237 *offsets++=sourceIndex++; |
| 238 *offsets++=sourceIndex++; |
| 239 *offsets++=sourceIndex++; |
| 240 *offsets++=sourceIndex++; |
| 241 *offsets++=sourceIndex++; |
| 242 *offsets++=sourceIndex++; |
| 243 *offsets++=sourceIndex++; |
| 244 *offsets++=sourceIndex++; |
| 245 *offsets++=sourceIndex++; |
| 246 *offsets++=sourceIndex++; |
| 247 *offsets++=sourceIndex++; |
| 248 *offsets++=sourceIndex++; |
| 249 *offsets++=sourceIndex++; |
| 250 --count; |
| 251 } |
| 252 } |
| 253 } |
| 254 #endif |
| 255 |
| 256 /* conversion loop */ |
| 257 c=0; |
| 258 while(targetCapacity>0 && (c=*source++)<=max) { |
| 259 /* convert the Unicode code point */ |
| 260 *target++=(uint8_t)c; |
| 261 --targetCapacity; |
| 262 } |
| 263 |
| 264 if(c>max) { |
| 265 cp=c; |
| 266 if(!U_IS_SURROGATE(cp)) { |
| 267 /* callback(unassigned) */ |
| 268 } else if(U_IS_SURROGATE_LEAD(cp)) { |
| 269 getTrail: |
| 270 if(source<sourceLimit) { |
| 271 /* test the following code unit */ |
| 272 UChar trail=*source; |
| 273 if(U16_IS_TRAIL(trail)) { |
| 274 ++source; |
| 275 cp=U16_GET_SUPPLEMENTARY(cp, trail); |
| 276 /* this codepage does not map supplementary code points */ |
| 277 /* callback(unassigned) */ |
| 278 } else { |
| 279 /* this is an unmatched lead code unit (1st surrogate) */ |
| 280 /* callback(illegal) */ |
| 281 } |
| 282 } else { |
| 283 /* no more input */ |
| 284 cnv->fromUChar32=cp; |
| 285 goto noMoreInput; |
| 286 } |
| 287 } else { |
| 288 /* this is an unmatched trail code unit (2nd surrogate) */ |
| 289 /* callback(illegal) */ |
| 290 } |
| 291 |
| 292 *pErrorCode= U_IS_SURROGATE(cp) ? U_ILLEGAL_CHAR_FOUND : U_INVALID_CHAR_
FOUND; |
| 293 cnv->fromUChar32=cp; |
| 294 } |
| 295 noMoreInput: |
| 296 |
| 297 /* set offsets since the start */ |
| 298 if(offsets!=NULL) { |
| 299 size_t count=target-oldTarget; |
| 300 while(count>0) { |
| 301 *offsets++=sourceIndex++; |
| 302 --count; |
| 303 } |
| 304 } |
| 305 |
| 306 if(U_SUCCESS(*pErrorCode) && source<sourceLimit && target>=(uint8_t *)pArgs-
>targetLimit) { |
| 307 /* target is full */ |
| 308 *pErrorCode=U_BUFFER_OVERFLOW_ERROR; |
| 309 } |
| 310 |
| 311 /* write back the updated pointers */ |
| 312 pArgs->source=source; |
| 313 pArgs->target=(char *)target; |
| 314 pArgs->offsets=offsets; |
| 315 } |
| 316 |
| 317 /* Convert UTF-8 to Latin-1. Adapted from ucnv_SBCSFromUTF8(). */ |
| 318 static void |
| 319 ucnv_Latin1FromUTF8(UConverterFromUnicodeArgs *pFromUArgs, |
| 320 UConverterToUnicodeArgs *pToUArgs, |
| 321 UErrorCode *pErrorCode) { |
| 322 UConverter *utf8; |
| 323 const uint8_t *source, *sourceLimit; |
| 324 uint8_t *target; |
| 325 int32_t targetCapacity; |
| 326 |
| 327 UChar32 c; |
| 328 uint8_t b, t1; |
| 329 |
| 330 /* set up the local pointers */ |
| 331 utf8=pToUArgs->converter; |
| 332 source=(uint8_t *)pToUArgs->source; |
| 333 sourceLimit=(uint8_t *)pToUArgs->sourceLimit; |
| 334 target=(uint8_t *)pFromUArgs->target; |
| 335 targetCapacity=(int32_t)(pFromUArgs->targetLimit-pFromUArgs->target); |
| 336 |
| 337 /* get the converter state from the UTF-8 UConverter */ |
| 338 c=(UChar32)utf8->toUnicodeStatus; |
| 339 if(c!=0 && source<sourceLimit) { |
| 340 if(targetCapacity==0) { |
| 341 *pErrorCode=U_BUFFER_OVERFLOW_ERROR; |
| 342 return; |
| 343 } else if(c>=0xc2 && c<=0xc3 && (t1=(uint8_t)(*source-0x80)) <= 0x3f) { |
| 344 ++source; |
| 345 *target++=(uint8_t)(((c&3)<<6)|t1); |
| 346 --targetCapacity; |
| 347 |
| 348 utf8->toUnicodeStatus=0; |
| 349 utf8->toULength=0; |
| 350 } else { |
| 351 /* complicated, illegal or unmappable input: fall back to the pivoti
ng implementation */ |
| 352 *pErrorCode=U_USING_DEFAULT_WARNING; |
| 353 return; |
| 354 } |
| 355 } |
| 356 |
| 357 /* |
| 358 * Make sure that the last byte sequence before sourceLimit is complete |
| 359 * or runs into a lead byte. |
| 360 * In the conversion loop compare source with sourceLimit only once |
| 361 * per multi-byte character. |
| 362 * For Latin-1, adjust sourceLimit only for 1 trail byte because |
| 363 * the conversion loop handles at most 2-byte sequences. |
| 364 */ |
| 365 if(source<sourceLimit && U8_IS_LEAD(*(sourceLimit-1))) { |
| 366 --sourceLimit; |
| 367 } |
| 368 |
| 369 /* conversion loop */ |
| 370 while(source<sourceLimit) { |
| 371 if(targetCapacity>0) { |
| 372 b=*source++; |
| 373 if((int8_t)b>=0) { |
| 374 /* convert ASCII */ |
| 375 *target++=(uint8_t)b; |
| 376 --targetCapacity; |
| 377 } else if( /* handle U+0080..U+00FF inline */ |
| 378 b>=0xc2 && b<=0xc3 && |
| 379 (t1=(uint8_t)(*source-0x80)) <= 0x3f |
| 380 ) { |
| 381 ++source; |
| 382 *target++=(uint8_t)(((b&3)<<6)|t1); |
| 383 --targetCapacity; |
| 384 } else { |
| 385 /* complicated, illegal or unmappable input: fall back to the pi
voting implementation */ |
| 386 pToUArgs->source=(char *)(source-1); |
| 387 pFromUArgs->target=(char *)target; |
| 388 *pErrorCode=U_USING_DEFAULT_WARNING; |
| 389 return; |
| 390 } |
| 391 } else { |
| 392 /* target is full */ |
| 393 *pErrorCode=U_BUFFER_OVERFLOW_ERROR; |
| 394 break; |
| 395 } |
| 396 } |
| 397 |
| 398 /* |
| 399 * The sourceLimit may have been adjusted before the conversion loop |
| 400 * to stop before a truncated sequence. |
| 401 * If so, then collect the truncated sequence now. |
| 402 * For Latin-1, there is at most exactly one lead byte because of the |
| 403 * smaller sourceLimit adjustment logic. |
| 404 */ |
| 405 if(U_SUCCESS(*pErrorCode) && source<(sourceLimit=(uint8_t *)pToUArgs->source
Limit)) { |
| 406 utf8->toUnicodeStatus=utf8->toUBytes[0]=b=*source++; |
| 407 utf8->toULength=1; |
| 408 utf8->mode=utf8_countTrailBytes[b]+1; |
| 409 } |
| 410 |
| 411 /* write back the updated pointers */ |
| 412 pToUArgs->source=(char *)source; |
| 413 pFromUArgs->target=(char *)target; |
| 414 } |
| 415 |
| 416 static void |
| 417 _Latin1GetUnicodeSet(const UConverter *cnv, |
| 418 const USetAdder *sa, |
| 419 UConverterUnicodeSet which, |
| 420 UErrorCode *pErrorCode) { |
| 421 sa->addRange(sa->set, 0, 0xff); |
| 422 } |
| 423 |
| 424 static const UConverterImpl _Latin1Impl={ |
| 425 UCNV_LATIN_1, |
| 426 |
| 427 NULL, |
| 428 NULL, |
| 429 |
| 430 NULL, |
| 431 NULL, |
| 432 NULL, |
| 433 |
| 434 _Latin1ToUnicodeWithOffsets, |
| 435 _Latin1ToUnicodeWithOffsets, |
| 436 _Latin1FromUnicodeWithOffsets, |
| 437 _Latin1FromUnicodeWithOffsets, |
| 438 _Latin1GetNextUChar, |
| 439 |
| 440 NULL, |
| 441 NULL, |
| 442 NULL, |
| 443 NULL, |
| 444 _Latin1GetUnicodeSet, |
| 445 |
| 446 NULL, |
| 447 ucnv_Latin1FromUTF8 |
| 448 }; |
| 449 |
| 450 static const UConverterStaticData _Latin1StaticData={ |
| 451 sizeof(UConverterStaticData), |
| 452 "ISO-8859-1", |
| 453 819, UCNV_IBM, UCNV_LATIN_1, 1, 1, |
| 454 { 0x1a, 0, 0, 0 }, 1, FALSE, FALSE, |
| 455 0, |
| 456 0, |
| 457 { 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0 } /* reserved */ |
| 458 }; |
| 459 |
| 460 const UConverterSharedData _Latin1Data={ |
| 461 sizeof(UConverterSharedData), ~((uint32_t) 0), |
| 462 NULL, NULL, &_Latin1StaticData, FALSE, &_Latin1Impl, |
| 463 0 |
| 464 }; |
| 465 |
| 466 /* US-ASCII ----------------------------------------------------------------- */ |
| 467 |
| 468 /* This is a table-less version of ucnv_MBCSSingleToBMPWithOffsets(). */ |
| 469 static void |
| 470 _ASCIIToUnicodeWithOffsets(UConverterToUnicodeArgs *pArgs, |
| 471 UErrorCode *pErrorCode) { |
| 472 const uint8_t *source, *sourceLimit; |
| 473 UChar *target, *oldTarget; |
| 474 int32_t targetCapacity, length; |
| 475 int32_t *offsets; |
| 476 |
| 477 int32_t sourceIndex; |
| 478 |
| 479 uint8_t c; |
| 480 |
| 481 /* set up the local pointers */ |
| 482 source=(const uint8_t *)pArgs->source; |
| 483 sourceLimit=(const uint8_t *)pArgs->sourceLimit; |
| 484 target=oldTarget=pArgs->target; |
| 485 targetCapacity=(int32_t)(pArgs->targetLimit-pArgs->target); |
| 486 offsets=pArgs->offsets; |
| 487 |
| 488 /* sourceIndex=-1 if the current character began in the previous buffer */ |
| 489 sourceIndex=0; |
| 490 |
| 491 /* |
| 492 * since the conversion here is 1:1 UChar:uint8_t, we need only one counter |
| 493 * for the minimum of the sourceLength and targetCapacity |
| 494 */ |
| 495 length=(int32_t)(sourceLimit-source); |
| 496 if(length<targetCapacity) { |
| 497 targetCapacity=length; |
| 498 } |
| 499 |
| 500 if(targetCapacity>=8) { |
| 501 /* This loop is unrolled for speed and improved pipelining. */ |
| 502 int32_t count, loops; |
| 503 UChar oredChars; |
| 504 |
| 505 loops=count=targetCapacity>>3; |
| 506 do { |
| 507 oredChars=target[0]=source[0]; |
| 508 oredChars|=target[1]=source[1]; |
| 509 oredChars|=target[2]=source[2]; |
| 510 oredChars|=target[3]=source[3]; |
| 511 oredChars|=target[4]=source[4]; |
| 512 oredChars|=target[5]=source[5]; |
| 513 oredChars|=target[6]=source[6]; |
| 514 oredChars|=target[7]=source[7]; |
| 515 |
| 516 /* were all 16 entries really valid? */ |
| 517 if(oredChars>0x7f) { |
| 518 /* no, return to the first of these 16 */ |
| 519 break; |
| 520 } |
| 521 source+=8; |
| 522 target+=8; |
| 523 } while(--count>0); |
| 524 count=loops-count; |
| 525 targetCapacity-=count*8; |
| 526 |
| 527 if(offsets!=NULL) { |
| 528 oldTarget+=count*8; |
| 529 while(count>0) { |
| 530 offsets[0]=sourceIndex++; |
| 531 offsets[1]=sourceIndex++; |
| 532 offsets[2]=sourceIndex++; |
| 533 offsets[3]=sourceIndex++; |
| 534 offsets[4]=sourceIndex++; |
| 535 offsets[5]=sourceIndex++; |
| 536 offsets[6]=sourceIndex++; |
| 537 offsets[7]=sourceIndex++; |
| 538 offsets+=8; |
| 539 --count; |
| 540 } |
| 541 } |
| 542 } |
| 543 |
| 544 /* conversion loop */ |
| 545 c=0; |
| 546 while(targetCapacity>0 && (c=*source++)<=0x7f) { |
| 547 *target++=c; |
| 548 --targetCapacity; |
| 549 } |
| 550 |
| 551 if(c>0x7f) { |
| 552 /* callback(illegal); copy the current bytes to toUBytes[] */ |
| 553 UConverter *cnv=pArgs->converter; |
| 554 cnv->toUBytes[0]=c; |
| 555 cnv->toULength=1; |
| 556 *pErrorCode=U_ILLEGAL_CHAR_FOUND; |
| 557 } else if(source<sourceLimit && target>=pArgs->targetLimit) { |
| 558 /* target is full */ |
| 559 *pErrorCode=U_BUFFER_OVERFLOW_ERROR; |
| 560 } |
| 561 |
| 562 /* set offsets since the start */ |
| 563 if(offsets!=NULL) { |
| 564 size_t count=target-oldTarget; |
| 565 while(count>0) { |
| 566 *offsets++=sourceIndex++; |
| 567 --count; |
| 568 } |
| 569 } |
| 570 |
| 571 /* write back the updated pointers */ |
| 572 pArgs->source=(const char *)source; |
| 573 pArgs->target=target; |
| 574 pArgs->offsets=offsets; |
| 575 } |
| 576 |
| 577 /* This is a table-less version of ucnv_MBCSSingleGetNextUChar(). */ |
| 578 static UChar32 |
| 579 _ASCIIGetNextUChar(UConverterToUnicodeArgs *pArgs, |
| 580 UErrorCode *pErrorCode) { |
| 581 const uint8_t *source; |
| 582 uint8_t b; |
| 583 |
| 584 source=(const uint8_t *)pArgs->source; |
| 585 if(source<(const uint8_t *)pArgs->sourceLimit) { |
| 586 b=*source++; |
| 587 pArgs->source=(const char *)source; |
| 588 if(b<=0x7f) { |
| 589 return b; |
| 590 } else { |
| 591 UConverter *cnv=pArgs->converter; |
| 592 cnv->toUBytes[0]=b; |
| 593 cnv->toULength=1; |
| 594 *pErrorCode=U_ILLEGAL_CHAR_FOUND; |
| 595 return 0xffff; |
| 596 } |
| 597 } |
| 598 |
| 599 /* no output because of empty input */ |
| 600 *pErrorCode=U_INDEX_OUTOFBOUNDS_ERROR; |
| 601 return 0xffff; |
| 602 } |
| 603 |
| 604 /* "Convert" UTF-8 to US-ASCII: Validate and copy. */ |
| 605 static void |
| 606 ucnv_ASCIIFromUTF8(UConverterFromUnicodeArgs *pFromUArgs, |
| 607 UConverterToUnicodeArgs *pToUArgs, |
| 608 UErrorCode *pErrorCode) { |
| 609 const uint8_t *source, *sourceLimit; |
| 610 uint8_t *target; |
| 611 int32_t targetCapacity, length; |
| 612 |
| 613 uint8_t c; |
| 614 |
| 615 if(pToUArgs->converter->toUnicodeStatus!=0) { |
| 616 /* no handling of partial UTF-8 characters here, fall back to pivoting *
/ |
| 617 *pErrorCode=U_USING_DEFAULT_WARNING; |
| 618 return; |
| 619 } |
| 620 |
| 621 /* set up the local pointers */ |
| 622 source=(const uint8_t *)pToUArgs->source; |
| 623 sourceLimit=(const uint8_t *)pToUArgs->sourceLimit; |
| 624 target=(uint8_t *)pFromUArgs->target; |
| 625 targetCapacity=(int32_t)(pFromUArgs->targetLimit-pFromUArgs->target); |
| 626 |
| 627 /* |
| 628 * since the conversion here is 1:1 uint8_t:uint8_t, we need only one counte
r |
| 629 * for the minimum of the sourceLength and targetCapacity |
| 630 */ |
| 631 length=(int32_t)(sourceLimit-source); |
| 632 if(length<targetCapacity) { |
| 633 targetCapacity=length; |
| 634 } |
| 635 |
| 636 /* unroll the loop with the most common case */ |
| 637 if(targetCapacity>=16) { |
| 638 int32_t count, loops; |
| 639 uint8_t oredChars; |
| 640 |
| 641 loops=count=targetCapacity>>4; |
| 642 do { |
| 643 oredChars=*target++=*source++; |
| 644 oredChars|=*target++=*source++; |
| 645 oredChars|=*target++=*source++; |
| 646 oredChars|=*target++=*source++; |
| 647 oredChars|=*target++=*source++; |
| 648 oredChars|=*target++=*source++; |
| 649 oredChars|=*target++=*source++; |
| 650 oredChars|=*target++=*source++; |
| 651 oredChars|=*target++=*source++; |
| 652 oredChars|=*target++=*source++; |
| 653 oredChars|=*target++=*source++; |
| 654 oredChars|=*target++=*source++; |
| 655 oredChars|=*target++=*source++; |
| 656 oredChars|=*target++=*source++; |
| 657 oredChars|=*target++=*source++; |
| 658 oredChars|=*target++=*source++; |
| 659 |
| 660 /* were all 16 entries really valid? */ |
| 661 if(oredChars>0x7f) { |
| 662 /* no, return to the first of these 16 */ |
| 663 source-=16; |
| 664 target-=16; |
| 665 break; |
| 666 } |
| 667 } while(--count>0); |
| 668 count=loops-count; |
| 669 targetCapacity-=16*count; |
| 670 } |
| 671 |
| 672 /* conversion loop */ |
| 673 c=0; |
| 674 while(targetCapacity>0 && (c=*source)<=0x7f) { |
| 675 ++source; |
| 676 *target++=c; |
| 677 --targetCapacity; |
| 678 } |
| 679 |
| 680 if(c>0x7f) { |
| 681 /* non-ASCII character, handle in standard converter */ |
| 682 *pErrorCode=U_USING_DEFAULT_WARNING; |
| 683 } else if(source<sourceLimit && target>=(const uint8_t *)pFromUArgs->targetL
imit) { |
| 684 /* target is full */ |
| 685 *pErrorCode=U_BUFFER_OVERFLOW_ERROR; |
| 686 } |
| 687 |
| 688 /* write back the updated pointers */ |
| 689 pToUArgs->source=(const char *)source; |
| 690 pFromUArgs->target=(char *)target; |
| 691 } |
| 692 |
| 693 static void |
| 694 _ASCIIGetUnicodeSet(const UConverter *cnv, |
| 695 const USetAdder *sa, |
| 696 UConverterUnicodeSet which, |
| 697 UErrorCode *pErrorCode) { |
| 698 sa->addRange(sa->set, 0, 0x7f); |
| 699 } |
| 700 |
| 701 static const UConverterImpl _ASCIIImpl={ |
| 702 UCNV_US_ASCII, |
| 703 |
| 704 NULL, |
| 705 NULL, |
| 706 |
| 707 NULL, |
| 708 NULL, |
| 709 NULL, |
| 710 |
| 711 _ASCIIToUnicodeWithOffsets, |
| 712 _ASCIIToUnicodeWithOffsets, |
| 713 _Latin1FromUnicodeWithOffsets, |
| 714 _Latin1FromUnicodeWithOffsets, |
| 715 _ASCIIGetNextUChar, |
| 716 |
| 717 NULL, |
| 718 NULL, |
| 719 NULL, |
| 720 NULL, |
| 721 _ASCIIGetUnicodeSet, |
| 722 |
| 723 NULL, |
| 724 ucnv_ASCIIFromUTF8 |
| 725 }; |
| 726 |
| 727 static const UConverterStaticData _ASCIIStaticData={ |
| 728 sizeof(UConverterStaticData), |
| 729 "US-ASCII", |
| 730 367, UCNV_IBM, UCNV_US_ASCII, 1, 1, |
| 731 { 0x1a, 0, 0, 0 }, 1, FALSE, FALSE, |
| 732 0, |
| 733 0, |
| 734 { 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0 } /* reserved */ |
| 735 }; |
| 736 |
| 737 const UConverterSharedData _ASCIIData={ |
| 738 sizeof(UConverterSharedData), ~((uint32_t) 0), |
| 739 NULL, NULL, &_ASCIIStaticData, FALSE, &_ASCIIImpl, |
| 740 0 |
| 741 }; |
| 742 |
| 743 #endif |
| OLD | NEW |
|---|
Chromium Code Reviews
Issue 5516007:
Check in the pristine copy of ICU 4.6... (Closed)
Base URL: svn://chrome-svn/chrome/trunk/deps/third_party/