1 /*
2 **********************************************************************
3 *   Copyright (C) 2000-2015, International Business Machines
4 *   Corporation and others.  All Rights Reserved.
5 **********************************************************************
6 *   file name:  ucnvhz.c
7 *   encoding:   US-ASCII
8 *   tab size:   8 (not used)
9 *   indentation:4
10 *
11 *   created on: 2000oct16
12 *   created by: Ram Viswanadha
13 *   10/31/2000  Ram     Implemented offsets logic function
14 *
15 */
16 
17 #include "unicode/utypes.h"
18 
19 #if !UCONFIG_NO_CONVERSION && !UCONFIG_NO_LEGACY_CONVERSION && !UCONFIG_ONLY_HTML_CONVERSION
20 
21 #include "cmemory.h"
22 #include "unicode/ucnv.h"
23 #include "unicode/ucnv_cb.h"
24 #include "unicode/uset.h"
25 #include "unicode/utf16.h"
26 #include "ucnv_bld.h"
27 #include "ucnv_cnv.h"
28 #include "ucnv_imp.h"
29 
30 #define UCNV_TILDE 0x7E          /* ~ */
31 #define UCNV_OPEN_BRACE 0x7B     /* { */
32 #define UCNV_CLOSE_BRACE 0x7D   /* } */
33 #define SB_ESCAPE    "\x7E\x7D"
34 #define DB_ESCAPE    "\x7E\x7B"
35 #define TILDE_ESCAPE "\x7E\x7E"
36 #define ESC_LEN       2
37 
38 
39 #define CONCAT_ESCAPE_MACRO( args, targetIndex,targetLength,strToAppend, err, len,sourceIndex){                             \
40     while(len-->0){                                                                                                         \
41         if(targetIndex < targetLength){                                                                                     \
42             args->target[targetIndex] = (unsigned char) *strToAppend;                                                       \
43             if(args->offsets!=NULL){                                                                                        \
44                 *(offsets++) = sourceIndex-1;                                                                               \
45             }                                                                                                               \
46             targetIndex++;                                                                                                  \
47         }                                                                                                                   \
48         else{                                                                                                               \
49             args->converter->charErrorBuffer[(int)args->converter->charErrorBufferLength++] = (unsigned char) *strToAppend; \
50             *err =U_BUFFER_OVERFLOW_ERROR;                                                                                  \
51         }                                                                                                                   \
52         strToAppend++;                                                                                                      \
53     }                                                                                                                       \
54 }
55 
56 
57 typedef struct{
58     UConverter* gbConverter;
59     int32_t targetIndex;
60     int32_t sourceIndex;
61     UBool isEscapeAppended;
62     UBool isStateDBCS;
63     UBool isTargetUCharDBCS;
64     UBool isEmptySegment;
65 }UConverterDataHZ;
66 
67 
68 
69 static void
_HZOpen(UConverter * cnv,UConverterLoadArgs * pArgs,UErrorCode * errorCode)70 _HZOpen(UConverter *cnv, UConverterLoadArgs *pArgs, UErrorCode *errorCode){
71     UConverter *gbConverter;
72     if(pArgs->onlyTestIsLoadable) {
73         ucnv_canCreateConverter("GBK", errorCode);  /* errorCode carries result */
74         return;
75     }
76     gbConverter = ucnv_open("GBK", errorCode);
77     if(U_FAILURE(*errorCode)) {
78         return;
79     }
80     cnv->toUnicodeStatus = 0;
81     cnv->fromUnicodeStatus= 0;
82     cnv->mode=0;
83     cnv->fromUChar32=0x0000;
84     cnv->extraInfo = uprv_calloc(1, sizeof(UConverterDataHZ));
85     if(cnv->extraInfo != NULL){
86         ((UConverterDataHZ*)cnv->extraInfo)->gbConverter = gbConverter;
87     }
88     else {
89         ucnv_close(gbConverter);
90         *errorCode = U_MEMORY_ALLOCATION_ERROR;
91         return;
92     }
93 }
94 
95 static void
_HZClose(UConverter * cnv)96 _HZClose(UConverter *cnv){
97     if(cnv->extraInfo != NULL) {
98         ucnv_close (((UConverterDataHZ *) (cnv->extraInfo))->gbConverter);
99         if(!cnv->isExtraLocal) {
100             uprv_free(cnv->extraInfo);
101         }
102         cnv->extraInfo = NULL;
103     }
104 }
105 
106 static void
_HZReset(UConverter * cnv,UConverterResetChoice choice)107 _HZReset(UConverter *cnv, UConverterResetChoice choice){
108     if(choice<=UCNV_RESET_TO_UNICODE) {
109         cnv->toUnicodeStatus = 0;
110         cnv->mode=0;
111         if(cnv->extraInfo != NULL){
112             ((UConverterDataHZ*)cnv->extraInfo)->isStateDBCS = FALSE;
113             ((UConverterDataHZ*)cnv->extraInfo)->isEmptySegment = FALSE;
114         }
115     }
116     if(choice!=UCNV_RESET_TO_UNICODE) {
117         cnv->fromUnicodeStatus= 0;
118         cnv->fromUChar32=0x0000;
119         if(cnv->extraInfo != NULL){
120             ((UConverterDataHZ*)cnv->extraInfo)->isEscapeAppended = FALSE;
121             ((UConverterDataHZ*)cnv->extraInfo)->targetIndex = 0;
122             ((UConverterDataHZ*)cnv->extraInfo)->sourceIndex = 0;
123             ((UConverterDataHZ*)cnv->extraInfo)->isTargetUCharDBCS = FALSE;
124         }
125     }
126 }
127 
128 /**************************************HZ Encoding*************************************************
129 * Rules for HZ encoding
130 *
131 *   In ASCII mode, a byte is interpreted as an ASCII character, unless a
132 *   '~' is encountered. The character '~' is an escape character. By
133 *   convention, it must be immediately followed ONLY by '~', '{' or '\n'
134 *   (<LF>), with the following special meaning.
135 
136 *   1. The escape sequence '~~' is interpreted as a '~'.
137 *   2. The escape-to-GB sequence '~{' switches the mode from ASCII to GB.
138 *   3. The escape sequence '~\n' is a line-continuation marker to be
139 *     consumed with no output produced.
140 *   In GB mode, characters are interpreted two bytes at a time as (pure)
141 *   GB codes until the escape-from-GB code '~}' is read. This code
142 *   switches the mode from GB back to ASCII.  (Note that the escape-
143 *   from-GB code '~}' ($7E7D) is outside the defined GB range.)
144 *
145 *   Source: RFC 1842
146 *
147 *   Note that the formal syntax in RFC 1842 is invalid. I assume that the
148 *   intended definition of single-byte-segment is as follows (pedberg):
149 *   single-byte-segment = single-byte-seq 1*single-byte-char
150 */
151 
152 
153 static void
UConverter_toUnicode_HZ_OFFSETS_LOGIC(UConverterToUnicodeArgs * args,UErrorCode * err)154 UConverter_toUnicode_HZ_OFFSETS_LOGIC(UConverterToUnicodeArgs *args,
155                                                             UErrorCode* err){
156     char tempBuf[2];
157     const char *mySource = ( char *) args->source;
158     UChar *myTarget = args->target;
159     const char *mySourceLimit = args->sourceLimit;
160     UChar32 targetUniChar = 0x0000;
161     int32_t mySourceChar = 0x0000;
162     UConverterDataHZ* myData=(UConverterDataHZ*)(args->converter->extraInfo);
163     tempBuf[0]=0;
164     tempBuf[1]=0;
165 
166     /* Calling code already handles this situation. */
167     /*if ((args->converter == NULL) || (args->targetLimit < args->target) || (mySourceLimit < args->source)){
168         *err = U_ILLEGAL_ARGUMENT_ERROR;
169         return;
170     }*/
171 
172     while(mySource< mySourceLimit){
173 
174         if(myTarget < args->targetLimit){
175 
176             mySourceChar= (unsigned char) *mySource++;
177 
178             if(args->converter->mode == UCNV_TILDE) {
179                 /* second byte after ~ */
180                 args->converter->mode=0;
181                 switch(mySourceChar) {
182                 case 0x0A:
183                     /* no output for ~\n (line-continuation marker) */
184                     continue;
185                 case UCNV_TILDE:
186                     if(args->offsets) {
187                         args->offsets[myTarget - args->target]=(int32_t)(mySource - args->source - 2);
188                     }
189                     *(myTarget++)=(UChar)mySourceChar;
190                     myData->isEmptySegment = FALSE;
191                     continue;
192                 case UCNV_OPEN_BRACE:
193                 case UCNV_CLOSE_BRACE:
194                     myData->isStateDBCS = (mySourceChar == UCNV_OPEN_BRACE);
195                     if (myData->isEmptySegment) {
196                         myData->isEmptySegment = FALSE; /* we are handling it, reset to avoid future spurious errors */
197                         *err = U_ILLEGAL_ESCAPE_SEQUENCE;
198                         args->converter->toUCallbackReason = UCNV_IRREGULAR;
199                         args->converter->toUBytes[0] = UCNV_TILDE;
200                         args->converter->toUBytes[1] = mySourceChar;
201                         args->converter->toULength = 2;
202                         args->target = myTarget;
203                         args->source = mySource;
204                         return;
205                     }
206                     myData->isEmptySegment = TRUE;
207                     continue;
208                 default:
209                      /* if the first byte is equal to TILDE and the trail byte
210                      * is not a valid byte then it is an error condition
211                      */
212                     /*
213                      * Ticket 5691: consistent illegal sequences:
214                      * - We include at least the first byte in the illegal sequence.
215                      * - If any of the non-initial bytes could be the start of a character,
216                      *   we stop the illegal sequence before the first one of those.
217                      */
218                     myData->isEmptySegment = FALSE; /* different error here, reset this to avoid spurious future error */
219                     *err = U_ILLEGAL_ESCAPE_SEQUENCE;
220                     args->converter->toUBytes[0] = UCNV_TILDE;
221                     if( myData->isStateDBCS ?
222                             (0x21 <= mySourceChar && mySourceChar <= 0x7e) :
223                             mySourceChar <= 0x7f
224                     ) {
225                         /* The current byte could be the start of a character: Back it out. */
226                         args->converter->toULength = 1;
227                         --mySource;
228                     } else {
229                         /* Include the current byte in the illegal sequence. */
230                         args->converter->toUBytes[1] = mySourceChar;
231                         args->converter->toULength = 2;
232                     }
233                     args->target = myTarget;
234                     args->source = mySource;
235                     return;
236                 }
237             } else if(myData->isStateDBCS) {
238                 if(args->converter->toUnicodeStatus == 0x00){
239                     /* lead byte */
240                     if(mySourceChar == UCNV_TILDE) {
241                         args->converter->mode = UCNV_TILDE;
242                     } else {
243                         /* add another bit to distinguish a 0 byte from not having seen a lead byte */
244                         args->converter->toUnicodeStatus = (uint32_t) (mySourceChar | 0x100);
245                         myData->isEmptySegment = FALSE; /* the segment has something, either valid or will produce a different error, so reset this */
246                     }
247                     continue;
248                 }
249                 else{
250                     /* trail byte */
251                     int leadIsOk, trailIsOk;
252                     uint32_t leadByte = args->converter->toUnicodeStatus & 0xff;
253                     targetUniChar = 0xffff;
254                     /*
255                      * Ticket 5691: consistent illegal sequences:
256                      * - We include at least the first byte in the illegal sequence.
257                      * - If any of the non-initial bytes could be the start of a character,
258                      *   we stop the illegal sequence before the first one of those.
259                      *
260                      * In HZ DBCS, if the second byte is in the 21..7e range,
261                      * we report only the first byte as the illegal sequence.
262                      * Otherwise we convert or report the pair of bytes.
263                      */
264                     leadIsOk = (uint8_t)(leadByte - 0x21) <= (0x7d - 0x21);
265                     trailIsOk = (uint8_t)(mySourceChar - 0x21) <= (0x7e - 0x21);
266                     if (leadIsOk && trailIsOk) {
267                         tempBuf[0] = (char) (leadByte+0x80) ;
268                         tempBuf[1] = (char) (mySourceChar+0x80);
269                         targetUniChar = ucnv_MBCSSimpleGetNextUChar(myData->gbConverter->sharedData,
270                             tempBuf, 2, args->converter->useFallback);
271                         mySourceChar= (leadByte << 8) | mySourceChar;
272                     } else if (trailIsOk) {
273                         /* report a single illegal byte and continue with the following DBCS starter byte */
274                         --mySource;
275                         mySourceChar = (int32_t)leadByte;
276                     } else {
277                         /* report a pair of illegal bytes if the second byte is not a DBCS starter */
278                         /* add another bit so that the code below writes 2 bytes in case of error */
279                         mySourceChar= 0x10000 | (leadByte << 8) | mySourceChar;
280                     }
281                     args->converter->toUnicodeStatus =0x00;
282                 }
283             }
284             else{
285                 if(mySourceChar == UCNV_TILDE) {
286                     args->converter->mode = UCNV_TILDE;
287                     continue;
288                 } else if(mySourceChar <= 0x7f) {
289                     targetUniChar = (UChar)mySourceChar;  /* ASCII */
290                     myData->isEmptySegment = FALSE; /* the segment has something valid */
291                 } else {
292                     targetUniChar = 0xffff;
293                     myData->isEmptySegment = FALSE; /* different error here, reset this to avoid spurious future error */
294                 }
295             }
296             if(targetUniChar < 0xfffe){
297                 if(args->offsets) {
298                     args->offsets[myTarget - args->target]=(int32_t)(mySource - args->source - 1-(myData->isStateDBCS));
299                 }
300 
301                 *(myTarget++)=(UChar)targetUniChar;
302             }
303             else /* targetUniChar>=0xfffe */ {
304                 if(targetUniChar == 0xfffe){
305                     *err = U_INVALID_CHAR_FOUND;
306                 }
307                 else{
308                     *err = U_ILLEGAL_CHAR_FOUND;
309                 }
310                 if(mySourceChar > 0xff){
311                     args->converter->toUBytes[0] = (uint8_t)(mySourceChar >> 8);
312                     args->converter->toUBytes[1] = (uint8_t)mySourceChar;
313                     args->converter->toULength=2;
314                 }
315                 else{
316                     args->converter->toUBytes[0] = (uint8_t)mySourceChar;
317                     args->converter->toULength=1;
318                 }
319                 break;
320             }
321         }
322         else{
323             *err =U_BUFFER_OVERFLOW_ERROR;
324             break;
325         }
326     }
327 
328     args->target = myTarget;
329     args->source = mySource;
330 }
331 
332 
333 static void
UConverter_fromUnicode_HZ_OFFSETS_LOGIC(UConverterFromUnicodeArgs * args,UErrorCode * err)334 UConverter_fromUnicode_HZ_OFFSETS_LOGIC (UConverterFromUnicodeArgs * args,
335                                                       UErrorCode * err){
336     const UChar *mySource = args->source;
337     char *myTarget = args->target;
338     int32_t* offsets = args->offsets;
339     int32_t mySourceIndex = 0;
340     int32_t myTargetIndex = 0;
341     int32_t targetLength = (int32_t)(args->targetLimit - myTarget);
342     int32_t mySourceLength = (int32_t)(args->sourceLimit - args->source);
343     uint32_t targetUniChar = 0x0000;
344     UChar32 mySourceChar = 0x0000;
345     UConverterDataHZ *myConverterData=(UConverterDataHZ*)args->converter->extraInfo;
346     UBool isTargetUCharDBCS = (UBool) myConverterData->isTargetUCharDBCS;
347     UBool oldIsTargetUCharDBCS;
348     int len =0;
349     const char* escSeq=NULL;
350 
351     /* Calling code already handles this situation. */
352     /*if ((args->converter == NULL) || (args->targetLimit < myTarget) || (args->sourceLimit < args->source)){
353         *err = U_ILLEGAL_ARGUMENT_ERROR;
354         return;
355     }*/
356     if(args->converter->fromUChar32!=0 && myTargetIndex < targetLength) {
357         goto getTrail;
358     }
359     /*writing the char to the output stream */
360     while (mySourceIndex < mySourceLength){
361         targetUniChar = missingCharMarker;
362         if (myTargetIndex < targetLength){
363 
364             mySourceChar = (UChar) mySource[mySourceIndex++];
365 
366 
367             oldIsTargetUCharDBCS = isTargetUCharDBCS;
368             if(mySourceChar ==UCNV_TILDE){
369                 /*concatEscape(args, &myTargetIndex, &targetLength,"\x7E\x7E",err,2,&mySourceIndex);*/
370                 len = ESC_LEN;
371                 escSeq = TILDE_ESCAPE;
372                 CONCAT_ESCAPE_MACRO(args, myTargetIndex, targetLength, escSeq,err,len,mySourceIndex);
373                 continue;
374             } else if(mySourceChar <= 0x7f) {
375                 targetUniChar = mySourceChar;
376             } else {
377                 int32_t length= ucnv_MBCSFromUChar32(myConverterData->gbConverter->sharedData,
378                     mySourceChar,&targetUniChar,args->converter->useFallback);
379                 /* we can only use lead bytes 21..7D and trail bytes 21..7E */
380                 if( length == 2 &&
381                     (uint16_t)(targetUniChar - 0xa1a1) <= (0xfdfe - 0xa1a1) &&
382                     (uint8_t)(targetUniChar - 0xa1) <= (0xfe - 0xa1)
383                 ) {
384                     targetUniChar -= 0x8080;
385                 } else {
386                     targetUniChar = missingCharMarker;
387                 }
388             }
389             if (targetUniChar != missingCharMarker){
390                myConverterData->isTargetUCharDBCS = isTargetUCharDBCS = (UBool)(targetUniChar>0x00FF);
391                  if(oldIsTargetUCharDBCS != isTargetUCharDBCS || !myConverterData->isEscapeAppended ){
392                     /*Shifting from a double byte to single byte mode*/
393                     if(!isTargetUCharDBCS){
394                         len =ESC_LEN;
395                         escSeq = SB_ESCAPE;
396                         CONCAT_ESCAPE_MACRO(args, myTargetIndex, targetLength, escSeq,err,len,mySourceIndex);
397                         myConverterData->isEscapeAppended = TRUE;
398                     }
399                     else{ /* Shifting from a single byte to double byte mode*/
400                         len =ESC_LEN;
401                         escSeq = DB_ESCAPE;
402                         CONCAT_ESCAPE_MACRO(args, myTargetIndex, targetLength, escSeq,err,len,mySourceIndex);
403                         myConverterData->isEscapeAppended = TRUE;
404 
405                     }
406                 }
407 
408                 if(isTargetUCharDBCS){
409                     if( myTargetIndex <targetLength){
410                         myTarget[myTargetIndex++] =(char) (targetUniChar >> 8);
411                         if(offsets){
412                             *(offsets++) = mySourceIndex-1;
413                         }
414                         if(myTargetIndex < targetLength){
415                             myTarget[myTargetIndex++] =(char) targetUniChar;
416                             if(offsets){
417                                 *(offsets++) = mySourceIndex-1;
418                             }
419                         }else{
420                             args->converter->charErrorBuffer[args->converter->charErrorBufferLength++] = (char) targetUniChar;
421                             *err = U_BUFFER_OVERFLOW_ERROR;
422                         }
423                     }else{
424                         args->converter->charErrorBuffer[args->converter->charErrorBufferLength++] =(char) (targetUniChar >> 8);
425                         args->converter->charErrorBuffer[args->converter->charErrorBufferLength++] = (char) targetUniChar;
426                         *err = U_BUFFER_OVERFLOW_ERROR;
427                     }
428 
429                 }else{
430                     if( myTargetIndex <targetLength){
431                         myTarget[myTargetIndex++] = (char) (targetUniChar );
432                         if(offsets){
433                             *(offsets++) = mySourceIndex-1;
434                         }
435 
436                     }else{
437                         args->converter->charErrorBuffer[args->converter->charErrorBufferLength++] = (char) targetUniChar;
438                         *err = U_BUFFER_OVERFLOW_ERROR;
439                     }
440                 }
441 
442             }
443             else{
444                 /* oops.. the code point is unassigned */
445                 /*Handle surrogates */
446                 /*check if the char is a First surrogate*/
447                 if(U16_IS_SURROGATE(mySourceChar)) {
448                     if(U16_IS_SURROGATE_LEAD(mySourceChar)) {
449                         args->converter->fromUChar32=mySourceChar;
450 getTrail:
451                         /*look ahead to find the trail surrogate*/
452                         if(mySourceIndex <  mySourceLength) {
453                             /* test the following code unit */
454                             UChar trail=(UChar) args->source[mySourceIndex];
455                             if(U16_IS_TRAIL(trail)) {
456                                 ++mySourceIndex;
457                                 mySourceChar=U16_GET_SUPPLEMENTARY(args->converter->fromUChar32, trail);
458                                 args->converter->fromUChar32=0x00;
459                                 /* there are no surrogates in GB2312*/
460                                 *err = U_INVALID_CHAR_FOUND;
461                                 /* exit this condition tree */
462                             } else {
463                                 /* this is an unmatched lead code unit (1st surrogate) */
464                                 /* callback(illegal) */
465                                 *err=U_ILLEGAL_CHAR_FOUND;
466                             }
467                         } else {
468                             /* no more input */
469                             *err = U_ZERO_ERROR;
470                         }
471                     } else {
472                         /* this is an unmatched trail code unit (2nd surrogate) */
473                         /* callback(illegal) */
474                         *err=U_ILLEGAL_CHAR_FOUND;
475                     }
476                 } else {
477                     /* callback(unassigned) for a BMP code point */
478                     *err = U_INVALID_CHAR_FOUND;
479                 }
480 
481                 args->converter->fromUChar32=mySourceChar;
482                 break;
483             }
484         }
485         else{
486             *err = U_BUFFER_OVERFLOW_ERROR;
487             break;
488         }
489         targetUniChar=missingCharMarker;
490     }
491 
492     args->target += myTargetIndex;
493     args->source += mySourceIndex;
494     myConverterData->isTargetUCharDBCS = isTargetUCharDBCS;
495 }
496 
497 static void
_HZ_WriteSub(UConverterFromUnicodeArgs * args,int32_t offsetIndex,UErrorCode * err)498 _HZ_WriteSub(UConverterFromUnicodeArgs *args, int32_t offsetIndex, UErrorCode *err) {
499     UConverter *cnv = args->converter;
500     UConverterDataHZ *convData=(UConverterDataHZ *) cnv->extraInfo;
501     char *p;
502     char buffer[4];
503     p = buffer;
504 
505     if( convData->isTargetUCharDBCS){
506         *p++= UCNV_TILDE;
507         *p++= UCNV_CLOSE_BRACE;
508         convData->isTargetUCharDBCS=FALSE;
509     }
510     *p++= (char)cnv->subChars[0];
511 
512     ucnv_cbFromUWriteBytes(args,
513                            buffer, (int32_t)(p - buffer),
514                            offsetIndex, err);
515 }
516 
517 /*
518  * Structure for cloning an HZ converter into a single memory block.
519  * ucnv_safeClone() of the HZ converter will align the entire cloneHZStruct,
520  * and then ucnv_safeClone() of the sub-converter may additionally align
521  * subCnv inside the cloneHZStruct, for which we need the deadSpace after
522  * subCnv. This is because UAlignedMemory may be larger than the actually
523  * necessary alignment size for the platform.
524  * The other cloneHZStruct fields will not be moved around,
525  * and are aligned properly with cloneHZStruct's alignment.
526  */
527 struct cloneHZStruct
528 {
529     UConverter cnv;
530     UConverter subCnv;
531     UAlignedMemory deadSpace;
532     UConverterDataHZ mydata;
533 };
534 
535 
536 static UConverter *
_HZ_SafeClone(const UConverter * cnv,void * stackBuffer,int32_t * pBufferSize,UErrorCode * status)537 _HZ_SafeClone(const UConverter *cnv,
538               void *stackBuffer,
539               int32_t *pBufferSize,
540               UErrorCode *status)
541 {
542     struct cloneHZStruct * localClone;
543     int32_t size, bufferSizeNeeded = sizeof(struct cloneHZStruct);
544 
545     if (U_FAILURE(*status)){
546         return 0;
547     }
548 
549     if (*pBufferSize == 0){ /* 'preflighting' request - set needed size into *pBufferSize */
550         *pBufferSize = bufferSizeNeeded;
551         return 0;
552     }
553 
554     localClone = (struct cloneHZStruct *)stackBuffer;
555     /* ucnv.c/ucnv_safeClone() copied the main UConverter already */
556 
557     uprv_memcpy(&localClone->mydata, cnv->extraInfo, sizeof(UConverterDataHZ));
558     localClone->cnv.extraInfo = &localClone->mydata;
559     localClone->cnv.isExtraLocal = TRUE;
560 
561     /* deep-clone the sub-converter */
562     size = (int32_t)(sizeof(UConverter) + sizeof(UAlignedMemory)); /* include size of padding */
563     ((UConverterDataHZ*)localClone->cnv.extraInfo)->gbConverter =
564         ucnv_safeClone(((UConverterDataHZ*)cnv->extraInfo)->gbConverter, &localClone->subCnv, &size, status);
565 
566     return &localClone->cnv;
567 }
568 
569 static void
_HZ_GetUnicodeSet(const UConverter * cnv,const USetAdder * sa,UConverterUnicodeSet which,UErrorCode * pErrorCode)570 _HZ_GetUnicodeSet(const UConverter *cnv,
571                   const USetAdder *sa,
572                   UConverterUnicodeSet which,
573                   UErrorCode *pErrorCode) {
574     /* HZ converts all of ASCII */
575     sa->addRange(sa->set, 0, 0x7f);
576 
577     /* add all of the code points that the sub-converter handles */
578     ucnv_MBCSGetFilteredUnicodeSetForUnicode(
579         ((UConverterDataHZ*)cnv->extraInfo)->gbConverter->sharedData,
580         sa, which, UCNV_SET_FILTER_HZ,
581         pErrorCode);
582 }
583 
584 static const UConverterImpl _HZImpl={
585 
586     UCNV_HZ,
587 
588     NULL,
589     NULL,
590 
591     _HZOpen,
592     _HZClose,
593     _HZReset,
594 
595     UConverter_toUnicode_HZ_OFFSETS_LOGIC,
596     UConverter_toUnicode_HZ_OFFSETS_LOGIC,
597     UConverter_fromUnicode_HZ_OFFSETS_LOGIC,
598     UConverter_fromUnicode_HZ_OFFSETS_LOGIC,
599     NULL,
600 
601     NULL,
602     NULL,
603     _HZ_WriteSub,
604     _HZ_SafeClone,
605     _HZ_GetUnicodeSet
606 };
607 
608 static const UConverterStaticData _HZStaticData={
609     sizeof(UConverterStaticData),
610         "HZ",
611          0,
612          UCNV_IBM,
613          UCNV_HZ,
614          1,
615          4,
616         { 0x1a, 0, 0, 0 },
617         1,
618         FALSE,
619         FALSE,
620         0,
621         0,
622         { 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0 }, /* reserved */
623 
624 };
625 
626 
627 const UConverterSharedData _HZData={
628     sizeof(UConverterSharedData),
629         ~((uint32_t) 0),
630         NULL,
631         NULL,
632         &_HZStaticData,
633         FALSE,
634         &_HZImpl,
635         0
636 };
637 
638 #endif /* #if !UCONFIG_NO_CONVERSION && !UCONFIG_NO_LEGACY_CONVERSION && !UCONFIG_ONLY_HTML_CONVERSION */
639