Home | History | Annotate | Download | only in common
      1 //  2016 and later: Unicode, Inc. and others.
      2 // License & terms of use: http://www.unicode.org/copyright.html
      3 /*
      4 **********************************************************************
      5 *   Copyright (C) 2000-2015, International Business Machines
      6 *   Corporation and others.  All Rights Reserved.
      7 **********************************************************************
      8 *   file name:  ucnvlat1.cpp
      9 *   encoding:   UTF-8
     10 *   tab size:   8 (not used)
     11 *   indentation:4
     12 *
     13 *   created on: 2000feb07
     14 *   created by: Markus W. Scherer
     15 */
     16 
     17 #include "unicode/utypes.h"
     18 
     19 #if !UCONFIG_NO_CONVERSION
     20 
     21 #include "unicode/ucnv.h"
     22 #include "unicode/uset.h"
     23 #include "unicode/utf8.h"
     24 #include "ucnv_bld.h"
     25 #include "ucnv_cnv.h"
     26 #include "ustr_imp.h"
     27 
     28 /* control optimizations according to the platform */
     29 #define LATIN1_UNROLL_FROM_UNICODE 1
     30 
     31 /* ISO 8859-1 --------------------------------------------------------------- */
     32 
     33 /* This is a table-less and callback-less version of ucnv_MBCSSingleToBMPWithOffsets(). */
     34 U_CDECL_BEGIN
     35 static void U_CALLCONV
     36 _Latin1ToUnicodeWithOffsets(UConverterToUnicodeArgs *pArgs,
     37                             UErrorCode *pErrorCode) {
     38     const uint8_t *source;
     39     UChar *target;
     40     int32_t targetCapacity, length;
     41     int32_t *offsets;
     42 
     43     int32_t sourceIndex;
     44 
     45     /* set up the local pointers */
     46     source=(const uint8_t *)pArgs->source;
     47     target=pArgs->target;
     48     targetCapacity=(int32_t)(pArgs->targetLimit-pArgs->target);
     49     offsets=pArgs->offsets;
     50 
     51     sourceIndex=0;
     52 
     53     /*
     54      * since the conversion here is 1:1 UChar:uint8_t, we need only one counter
     55      * for the minimum of the sourceLength and targetCapacity
     56      */
     57     length=(int32_t)((const uint8_t *)pArgs->sourceLimit-source);
     58     if(length<=targetCapacity) {
     59         targetCapacity=length;
     60     } else {
     61         /* target will be full */
     62         *pErrorCode=U_BUFFER_OVERFLOW_ERROR;
     63         length=targetCapacity;
     64     }
     65 
     66     if(targetCapacity>=8) {
     67         /* This loop is unrolled for speed and improved pipelining. */
     68         int32_t count, loops;
     69 
     70         loops=count=targetCapacity>>3;
     71         length=targetCapacity&=0x7;
     72         do {
     73             target[0]=source[0];
     74             target[1]=source[1];
     75             target[2]=source[2];
     76             target[3]=source[3];
     77             target[4]=source[4];
     78             target[5]=source[5];
     79             target[6]=source[6];
     80             target[7]=source[7];
     81             target+=8;
     82             source+=8;
     83         } while(--count>0);
     84 
     85         if(offsets!=NULL) {
     86             do {
     87                 offsets[0]=sourceIndex++;
     88                 offsets[1]=sourceIndex++;
     89                 offsets[2]=sourceIndex++;
     90                 offsets[3]=sourceIndex++;
     91                 offsets[4]=sourceIndex++;
     92                 offsets[5]=sourceIndex++;
     93                 offsets[6]=sourceIndex++;
     94                 offsets[7]=sourceIndex++;
     95                 offsets+=8;
     96             } while(--loops>0);
     97         }
     98     }
     99 
    100     /* conversion loop */
    101     while(targetCapacity>0) {
    102         *target++=*source++;
    103         --targetCapacity;
    104     }
    105 
    106     /* write back the updated pointers */
    107     pArgs->source=(const char *)source;
    108     pArgs->target=target;
    109 
    110     /* set offsets */
    111     if(offsets!=NULL) {
    112         while(length>0) {
    113             *offsets++=sourceIndex++;
    114             --length;
    115         }
    116         pArgs->offsets=offsets;
    117     }
    118 }
    119 
    120 /* This is a table-less and callback-less version of ucnv_MBCSSingleGetNextUChar(). */
    121 static UChar32 U_CALLCONV
    122 _Latin1GetNextUChar(UConverterToUnicodeArgs *pArgs,
    123                     UErrorCode *pErrorCode) {
    124     const uint8_t *source=(const uint8_t *)pArgs->source;
    125     if(source<(const uint8_t *)pArgs->sourceLimit) {
    126         pArgs->source=(const char *)(source+1);
    127         return *source;
    128     }
    129 
    130     /* no output because of empty input */
    131     *pErrorCode=U_INDEX_OUTOFBOUNDS_ERROR;
    132     return 0xffff;
    133 }
    134 
    135 /* This is a table-less version of ucnv_MBCSSingleFromBMPWithOffsets(). */
    136 static void U_CALLCONV
    137 _Latin1FromUnicodeWithOffsets(UConverterFromUnicodeArgs *pArgs,
    138                               UErrorCode *pErrorCode) {
    139     UConverter *cnv;
    140     const UChar *source, *sourceLimit;
    141     uint8_t *target, *oldTarget;
    142     int32_t targetCapacity, length;
    143     int32_t *offsets;
    144 
    145     UChar32 cp;
    146     UChar c, max;
    147 
    148     int32_t sourceIndex;
    149 
    150     /* set up the local pointers */
    151     cnv=pArgs->converter;
    152     source=pArgs->source;
    153     sourceLimit=pArgs->sourceLimit;
    154     target=oldTarget=(uint8_t *)pArgs->target;
    155     targetCapacity=(int32_t)(pArgs->targetLimit-pArgs->target);
    156     offsets=pArgs->offsets;
    157 
    158     if(cnv->sharedData==&_Latin1Data) {
    159         max=0xff; /* Latin-1 */
    160     } else {
    161         max=0x7f; /* US-ASCII */
    162     }
    163 
    164     /* get the converter state from UConverter */
    165     cp=cnv->fromUChar32;
    166 
    167     /* sourceIndex=-1 if the current character began in the previous buffer */
    168     sourceIndex= cp==0 ? 0 : -1;
    169 
    170     /*
    171      * since the conversion here is 1:1 UChar:uint8_t, we need only one counter
    172      * for the minimum of the sourceLength and targetCapacity
    173      */
    174     length=(int32_t)(sourceLimit-source);
    175     if(length<targetCapacity) {
    176         targetCapacity=length;
    177     }
    178 
    179     /* conversion loop */
    180     if(cp!=0 && targetCapacity>0) {
    181         goto getTrail;
    182     }
    183 
    184 #if LATIN1_UNROLL_FROM_UNICODE
    185     /* unroll the loop with the most common case */
    186     if(targetCapacity>=16) {
    187         int32_t count, loops;
    188         UChar u, oredChars;
    189 
    190         loops=count=targetCapacity>>4;
    191         do {
    192             oredChars=u=*source++;
    193             *target++=(uint8_t)u;
    194             oredChars|=u=*source++;
    195             *target++=(uint8_t)u;
    196             oredChars|=u=*source++;
    197             *target++=(uint8_t)u;
    198             oredChars|=u=*source++;
    199             *target++=(uint8_t)u;
    200             oredChars|=u=*source++;
    201             *target++=(uint8_t)u;
    202             oredChars|=u=*source++;
    203             *target++=(uint8_t)u;
    204             oredChars|=u=*source++;
    205             *target++=(uint8_t)u;
    206             oredChars|=u=*source++;
    207             *target++=(uint8_t)u;
    208             oredChars|=u=*source++;
    209             *target++=(uint8_t)u;
    210             oredChars|=u=*source++;
    211             *target++=(uint8_t)u;
    212             oredChars|=u=*source++;
    213             *target++=(uint8_t)u;
    214             oredChars|=u=*source++;
    215             *target++=(uint8_t)u;
    216             oredChars|=u=*source++;
    217             *target++=(uint8_t)u;
    218             oredChars|=u=*source++;
    219             *target++=(uint8_t)u;
    220             oredChars|=u=*source++;
    221             *target++=(uint8_t)u;
    222             oredChars|=u=*source++;
    223             *target++=(uint8_t)u;
    224 
    225             /* were all 16 entries really valid? */
    226             if(oredChars>max) {
    227                 /* no, return to the first of these 16 */
    228                 source-=16;
    229                 target-=16;
    230                 break;
    231             }
    232         } while(--count>0);
    233         count=loops-count;
    234         targetCapacity-=16*count;
    235 
    236         if(offsets!=NULL) {
    237             oldTarget+=16*count;
    238             while(count>0) {
    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                 *offsets++=sourceIndex++;
    251                 *offsets++=sourceIndex++;
    252                 *offsets++=sourceIndex++;
    253                 *offsets++=sourceIndex++;
    254                 *offsets++=sourceIndex++;
    255                 --count;
    256             }
    257         }
    258     }
    259 #endif
    260 
    261     /* conversion loop */
    262     c=0;
    263     while(targetCapacity>0 && (c=*source++)<=max) {
    264         /* convert the Unicode code point */
    265         *target++=(uint8_t)c;
    266         --targetCapacity;
    267     }
    268 
    269     if(c>max) {
    270         cp=c;
    271         if(!U_IS_SURROGATE(cp)) {
    272             /* callback(unassigned) */
    273         } else if(U_IS_SURROGATE_LEAD(cp)) {
    274 getTrail:
    275             if(source<sourceLimit) {
    276                 /* test the following code unit */
    277                 UChar trail=*source;
    278                 if(U16_IS_TRAIL(trail)) {
    279                     ++source;
    280                     cp=U16_GET_SUPPLEMENTARY(cp, trail);
    281                     /* this codepage does not map supplementary code points */
    282                     /* callback(unassigned) */
    283                 } else {
    284                     /* this is an unmatched lead code unit (1st surrogate) */
    285                     /* callback(illegal) */
    286                 }
    287             } else {
    288                 /* no more input */
    289                 cnv->fromUChar32=cp;
    290                 goto noMoreInput;
    291             }
    292         } else {
    293             /* this is an unmatched trail code unit (2nd surrogate) */
    294             /* callback(illegal) */
    295         }
    296 
    297         *pErrorCode= U_IS_SURROGATE(cp) ? U_ILLEGAL_CHAR_FOUND : U_INVALID_CHAR_FOUND;
    298         cnv->fromUChar32=cp;
    299     }
    300 noMoreInput:
    301 
    302     /* set offsets since the start */
    303     if(offsets!=NULL) {
    304         size_t count=target-oldTarget;
    305         while(count>0) {
    306             *offsets++=sourceIndex++;
    307             --count;
    308         }
    309     }
    310 
    311     if(U_SUCCESS(*pErrorCode) && source<sourceLimit && target>=(uint8_t *)pArgs->targetLimit) {
    312         /* target is full */
    313         *pErrorCode=U_BUFFER_OVERFLOW_ERROR;
    314     }
    315 
    316     /* write back the updated pointers */
    317     pArgs->source=source;
    318     pArgs->target=(char *)target;
    319     pArgs->offsets=offsets;
    320 }
    321 
    322 /* Convert UTF-8 to Latin-1. Adapted from ucnv_SBCSFromUTF8(). */
    323 static void U_CALLCONV
    324 ucnv_Latin1FromUTF8(UConverterFromUnicodeArgs *pFromUArgs,
    325                     UConverterToUnicodeArgs *pToUArgs,
    326                     UErrorCode *pErrorCode) {
    327     UConverter *utf8;
    328     const uint8_t *source, *sourceLimit;
    329     uint8_t *target;
    330     int32_t targetCapacity;
    331 
    332     UChar32 c;
    333     uint8_t b, t1;
    334 
    335     /* set up the local pointers */
    336     utf8=pToUArgs->converter;
    337     source=(uint8_t *)pToUArgs->source;
    338     sourceLimit=(uint8_t *)pToUArgs->sourceLimit;
    339     target=(uint8_t *)pFromUArgs->target;
    340     targetCapacity=(int32_t)(pFromUArgs->targetLimit-pFromUArgs->target);
    341 
    342     /* get the converter state from the UTF-8 UConverter */
    343     c=(UChar32)utf8->toUnicodeStatus;
    344     if(c!=0 && source<sourceLimit) {
    345         if(targetCapacity==0) {
    346             *pErrorCode=U_BUFFER_OVERFLOW_ERROR;
    347             return;
    348         } else if(c>=0xc2 && c<=0xc3 && (t1=(uint8_t)(*source-0x80)) <= 0x3f) {
    349             ++source;
    350             *target++=(uint8_t)(((c&3)<<6)|t1);
    351             --targetCapacity;
    352 
    353             utf8->toUnicodeStatus=0;
    354             utf8->toULength=0;
    355         } else {
    356             /* complicated, illegal or unmappable input: fall back to the pivoting implementation */
    357             *pErrorCode=U_USING_DEFAULT_WARNING;
    358             return;
    359         }
    360     }
    361 
    362     /*
    363      * Make sure that the last byte sequence before sourceLimit is complete
    364      * or runs into a lead byte.
    365      * In the conversion loop compare source with sourceLimit only once
    366      * per multi-byte character.
    367      * For Latin-1, adjust sourceLimit only for 1 trail byte because
    368      * the conversion loop handles at most 2-byte sequences.
    369      */
    370     if(source<sourceLimit && U8_IS_LEAD(*(sourceLimit-1))) {
    371         --sourceLimit;
    372     }
    373 
    374     /* conversion loop */
    375     while(source<sourceLimit) {
    376         if(targetCapacity>0) {
    377             b=*source++;
    378             if(U8_IS_SINGLE(b)) {
    379                 /* convert ASCII */
    380                 *target++=(uint8_t)b;
    381                 --targetCapacity;
    382             } else if( /* handle U+0080..U+00FF inline */
    383                        b>=0xc2 && b<=0xc3 &&
    384                        (t1=(uint8_t)(*source-0x80)) <= 0x3f
    385             ) {
    386                 ++source;
    387                 *target++=(uint8_t)(((b&3)<<6)|t1);
    388                 --targetCapacity;
    389             } else {
    390                 /* complicated, illegal or unmappable input: fall back to the pivoting implementation */
    391                 pToUArgs->source=(char *)(source-1);
    392                 pFromUArgs->target=(char *)target;
    393                 *pErrorCode=U_USING_DEFAULT_WARNING;
    394                 return;
    395             }
    396         } else {
    397             /* target is full */
    398             *pErrorCode=U_BUFFER_OVERFLOW_ERROR;
    399             break;
    400         }
    401     }
    402 
    403     /*
    404      * The sourceLimit may have been adjusted before the conversion loop
    405      * to stop before a truncated sequence.
    406      * If so, then collect the truncated sequence now.
    407      * For Latin-1, there is at most exactly one lead byte because of the
    408      * smaller sourceLimit adjustment logic.
    409      */
    410     if(U_SUCCESS(*pErrorCode) && source<(sourceLimit=(uint8_t *)pToUArgs->sourceLimit)) {
    411         utf8->toUnicodeStatus=utf8->toUBytes[0]=b=*source++;
    412         utf8->toULength=1;
    413         utf8->mode=U8_COUNT_BYTES(b);
    414     }
    415 
    416     /* write back the updated pointers */
    417     pToUArgs->source=(char *)source;
    418     pFromUArgs->target=(char *)target;
    419 }
    420 
    421 static void U_CALLCONV
    422 _Latin1GetUnicodeSet(const UConverter *cnv,
    423                      const USetAdder *sa,
    424                      UConverterUnicodeSet which,
    425                      UErrorCode *pErrorCode) {
    426     (void)cnv;
    427     (void)which;
    428     (void)pErrorCode;
    429     sa->addRange(sa->set, 0, 0xff);
    430 }
    431 U_CDECL_END
    432 
    433 
    434 static const UConverterImpl _Latin1Impl={
    435     UCNV_LATIN_1,
    436 
    437     NULL,
    438     NULL,
    439 
    440     NULL,
    441     NULL,
    442     NULL,
    443 
    444     _Latin1ToUnicodeWithOffsets,
    445     _Latin1ToUnicodeWithOffsets,
    446     _Latin1FromUnicodeWithOffsets,
    447     _Latin1FromUnicodeWithOffsets,
    448     _Latin1GetNextUChar,
    449 
    450     NULL,
    451     NULL,
    452     NULL,
    453     NULL,
    454     _Latin1GetUnicodeSet,
    455 
    456     NULL,
    457     ucnv_Latin1FromUTF8
    458 };
    459 
    460 static const UConverterStaticData _Latin1StaticData={
    461     sizeof(UConverterStaticData),
    462     "ISO-8859-1",
    463     819, UCNV_IBM, UCNV_LATIN_1, 1, 1,
    464     { 0x1a, 0, 0, 0 }, 1, FALSE, FALSE,
    465     0,
    466     0,
    467     { 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0 } /* reserved */
    468 };
    469 
    470 const UConverterSharedData _Latin1Data=
    471         UCNV_IMMUTABLE_SHARED_DATA_INITIALIZER(&_Latin1StaticData, &_Latin1Impl);
    472 
    473 /* US-ASCII ----------------------------------------------------------------- */
    474 
    475 U_CDECL_BEGIN
    476 /* This is a table-less version of ucnv_MBCSSingleToBMPWithOffsets(). */
    477 static void U_CALLCONV
    478 _ASCIIToUnicodeWithOffsets(UConverterToUnicodeArgs *pArgs,
    479                            UErrorCode *pErrorCode) {
    480     const uint8_t *source, *sourceLimit;
    481     UChar *target, *oldTarget;
    482     int32_t targetCapacity, length;
    483     int32_t *offsets;
    484 
    485     int32_t sourceIndex;
    486 
    487     uint8_t c;
    488 
    489     /* set up the local pointers */
    490     source=(const uint8_t *)pArgs->source;
    491     sourceLimit=(const uint8_t *)pArgs->sourceLimit;
    492     target=oldTarget=pArgs->target;
    493     targetCapacity=(int32_t)(pArgs->targetLimit-pArgs->target);
    494     offsets=pArgs->offsets;
    495 
    496     /* sourceIndex=-1 if the current character began in the previous buffer */
    497     sourceIndex=0;
    498 
    499     /*
    500      * since the conversion here is 1:1 UChar:uint8_t, we need only one counter
    501      * for the minimum of the sourceLength and targetCapacity
    502      */
    503     length=(int32_t)(sourceLimit-source);
    504     if(length<targetCapacity) {
    505         targetCapacity=length;
    506     }
    507 
    508     if(targetCapacity>=8) {
    509         /* This loop is unrolled for speed and improved pipelining. */
    510         int32_t count, loops;
    511         UChar oredChars;
    512 
    513         loops=count=targetCapacity>>3;
    514         do {
    515             oredChars=target[0]=source[0];
    516             oredChars|=target[1]=source[1];
    517             oredChars|=target[2]=source[2];
    518             oredChars|=target[3]=source[3];
    519             oredChars|=target[4]=source[4];
    520             oredChars|=target[5]=source[5];
    521             oredChars|=target[6]=source[6];
    522             oredChars|=target[7]=source[7];
    523 
    524             /* were all 16 entries really valid? */
    525             if(oredChars>0x7f) {
    526                 /* no, return to the first of these 16 */
    527                 break;
    528             }
    529             source+=8;
    530             target+=8;
    531         } while(--count>0);
    532         count=loops-count;
    533         targetCapacity-=count*8;
    534 
    535         if(offsets!=NULL) {
    536             oldTarget+=count*8;
    537             while(count>0) {
    538                 offsets[0]=sourceIndex++;
    539                 offsets[1]=sourceIndex++;
    540                 offsets[2]=sourceIndex++;
    541                 offsets[3]=sourceIndex++;
    542                 offsets[4]=sourceIndex++;
    543                 offsets[5]=sourceIndex++;
    544                 offsets[6]=sourceIndex++;
    545                 offsets[7]=sourceIndex++;
    546                 offsets+=8;
    547                 --count;
    548             }
    549         }
    550     }
    551 
    552     /* conversion loop */
    553     c=0;
    554     while(targetCapacity>0 && (c=*source++)<=0x7f) {
    555         *target++=c;
    556         --targetCapacity;
    557     }
    558 
    559     if(c>0x7f) {
    560         /* callback(illegal); copy the current bytes to toUBytes[] */
    561         UConverter *cnv=pArgs->converter;
    562         cnv->toUBytes[0]=c;
    563         cnv->toULength=1;
    564         *pErrorCode=U_ILLEGAL_CHAR_FOUND;
    565     } else if(source<sourceLimit && target>=pArgs->targetLimit) {
    566         /* target is full */
    567         *pErrorCode=U_BUFFER_OVERFLOW_ERROR;
    568     }
    569 
    570     /* set offsets since the start */
    571     if(offsets!=NULL) {
    572         size_t count=target-oldTarget;
    573         while(count>0) {
    574             *offsets++=sourceIndex++;
    575             --count;
    576         }
    577     }
    578 
    579     /* write back the updated pointers */
    580     pArgs->source=(const char *)source;
    581     pArgs->target=target;
    582     pArgs->offsets=offsets;
    583 }
    584 
    585 /* This is a table-less version of ucnv_MBCSSingleGetNextUChar(). */
    586 static UChar32 U_CALLCONV
    587 _ASCIIGetNextUChar(UConverterToUnicodeArgs *pArgs,
    588                    UErrorCode *pErrorCode) {
    589     const uint8_t *source;
    590     uint8_t b;
    591 
    592     source=(const uint8_t *)pArgs->source;
    593     if(source<(const uint8_t *)pArgs->sourceLimit) {
    594         b=*source++;
    595         pArgs->source=(const char *)source;
    596         if(b<=0x7f) {
    597             return b;
    598         } else {
    599             UConverter *cnv=pArgs->converter;
    600             cnv->toUBytes[0]=b;
    601             cnv->toULength=1;
    602             *pErrorCode=U_ILLEGAL_CHAR_FOUND;
    603             return 0xffff;
    604         }
    605     }
    606 
    607     /* no output because of empty input */
    608     *pErrorCode=U_INDEX_OUTOFBOUNDS_ERROR;
    609     return 0xffff;
    610 }
    611 
    612 /* "Convert" UTF-8 to US-ASCII: Validate and copy. */
    613 static void U_CALLCONV
    614 ucnv_ASCIIFromUTF8(UConverterFromUnicodeArgs *pFromUArgs,
    615                    UConverterToUnicodeArgs *pToUArgs,
    616                    UErrorCode *pErrorCode) {
    617     const uint8_t *source, *sourceLimit;
    618     uint8_t *target;
    619     int32_t targetCapacity, length;
    620 
    621     uint8_t c;
    622 
    623     if(pToUArgs->converter->toUnicodeStatus!=0) {
    624         /* no handling of partial UTF-8 characters here, fall back to pivoting */
    625         *pErrorCode=U_USING_DEFAULT_WARNING;
    626         return;
    627     }
    628 
    629     /* set up the local pointers */
    630     source=(const uint8_t *)pToUArgs->source;
    631     sourceLimit=(const uint8_t *)pToUArgs->sourceLimit;
    632     target=(uint8_t *)pFromUArgs->target;
    633     targetCapacity=(int32_t)(pFromUArgs->targetLimit-pFromUArgs->target);
    634 
    635     /*
    636      * since the conversion here is 1:1 uint8_t:uint8_t, we need only one counter
    637      * for the minimum of the sourceLength and targetCapacity
    638      */
    639     length=(int32_t)(sourceLimit-source);
    640     if(length<targetCapacity) {
    641         targetCapacity=length;
    642     }
    643 
    644     /* unroll the loop with the most common case */
    645     if(targetCapacity>=16) {
    646         int32_t count, loops;
    647         uint8_t oredChars;
    648 
    649         loops=count=targetCapacity>>4;
    650         do {
    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             oredChars|=*target++=*source++;
    660             oredChars|=*target++=*source++;
    661             oredChars|=*target++=*source++;
    662             oredChars|=*target++=*source++;
    663             oredChars|=*target++=*source++;
    664             oredChars|=*target++=*source++;
    665             oredChars|=*target++=*source++;
    666             oredChars|=*target++=*source++;
    667 
    668             /* were all 16 entries really valid? */
    669             if(oredChars>0x7f) {
    670                 /* no, return to the first of these 16 */
    671                 source-=16;
    672                 target-=16;
    673                 break;
    674             }
    675         } while(--count>0);
    676         count=loops-count;
    677         targetCapacity-=16*count;
    678     }
    679 
    680     /* conversion loop */
    681     c=0;
    682     while(targetCapacity>0 && (c=*source)<=0x7f) {
    683         ++source;
    684         *target++=c;
    685         --targetCapacity;
    686     }
    687 
    688     if(c>0x7f) {
    689         /* non-ASCII character, handle in standard converter */
    690         *pErrorCode=U_USING_DEFAULT_WARNING;
    691     } else if(source<sourceLimit && target>=(const uint8_t *)pFromUArgs->targetLimit) {
    692         /* target is full */
    693         *pErrorCode=U_BUFFER_OVERFLOW_ERROR;
    694     }
    695 
    696     /* write back the updated pointers */
    697     pToUArgs->source=(const char *)source;
    698     pFromUArgs->target=(char *)target;
    699 }
    700 
    701 static void U_CALLCONV
    702 _ASCIIGetUnicodeSet(const UConverter *cnv,
    703                     const USetAdder *sa,
    704                     UConverterUnicodeSet which,
    705                     UErrorCode *pErrorCode) {
    706     (void)cnv;
    707     (void)which;
    708     (void)pErrorCode;
    709     sa->addRange(sa->set, 0, 0x7f);
    710 }
    711 U_CDECL_END
    712 
    713 static const UConverterImpl _ASCIIImpl={
    714     UCNV_US_ASCII,
    715 
    716     NULL,
    717     NULL,
    718 
    719     NULL,
    720     NULL,
    721     NULL,
    722 
    723     _ASCIIToUnicodeWithOffsets,
    724     _ASCIIToUnicodeWithOffsets,
    725     _Latin1FromUnicodeWithOffsets,
    726     _Latin1FromUnicodeWithOffsets,
    727     _ASCIIGetNextUChar,
    728 
    729     NULL,
    730     NULL,
    731     NULL,
    732     NULL,
    733     _ASCIIGetUnicodeSet,
    734 
    735     NULL,
    736     ucnv_ASCIIFromUTF8
    737 };
    738 
    739 static const UConverterStaticData _ASCIIStaticData={
    740     sizeof(UConverterStaticData),
    741     "US-ASCII",
    742     367, UCNV_IBM, UCNV_US_ASCII, 1, 1,
    743     { 0x1a, 0, 0, 0 }, 1, FALSE, FALSE,
    744     0,
    745     0,
    746     { 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0 } /* reserved */
    747 };
    748 
    749 const UConverterSharedData _ASCIIData=
    750         UCNV_IMMUTABLE_SHARED_DATA_INITIALIZER(&_ASCIIStaticData, &_ASCIIImpl);
    751 
    752 #endif
    753