Annotation of parser3/src/main/pa_charset.C, revision 1.69
1.1 paf 1: /** @file
2: Parser: Charset connection implementation.
3:
1.68 misha 4: Copyright(c) 2001-2009 ArtLebedev Group (http://www.artlebedev.com)
1.4 paf 5: Author: Alexander Petrosyan<paf@design.ru>(http://paf.design.ru)
1.27 paf 6: */
1.1 paf 7:
1.69 ! misha 8: static const char * const IDENT_CHARSET_C="$Date: 2009-04-15 04:51:40 $";
1.1 paf 9:
10: #include "pa_charset.h"
1.35 paf 11: #include "pa_charsets.h"
1.1 paf 12:
13: #ifdef XML
1.8 paf 14: #include "libxml/encoding.h"
1.1 paf 15: #endif
16:
1.46 paf 17: //#define PA_PATCHED_LIBXML_BACKWARD
1.67 misha 18:
19: // reduce memory usage by pre-calculation utf-8 string length
1.60 misha 20: #define PRECALCULATE_DEST_LENGTH
1.46 paf 21:
1.38 paf 22: // globals
23:
24: Charset::UTF8CaseTable::Rec UTF8CaseToUpperRecords[]={
25: #include "utf8-to-upper.inc"
26: };
27: Charset::UTF8CaseTable UTF8CaseToUpper={
28: sizeof(UTF8CaseToUpperRecords)/sizeof(Charset::UTF8CaseTable::Rec),
29: UTF8CaseToUpperRecords};
30:
31: Charset::UTF8CaseTable::Rec UTF8CaseToLowerRecords[]={
32: #include "utf8-to-lower.inc"
33: };
34: Charset::UTF8CaseTable UTF8CaseToLower={
35: sizeof(UTF8CaseToLowerRecords)/sizeof(Charset::UTF8CaseTable::Rec),
36: UTF8CaseToLowerRecords};
37:
1.1 paf 38: // helpers
39:
40: inline void prepare_case_tables(unsigned char *tables) {
41: unsigned char *lcc_table=tables+lcc_offset;
42: unsigned char *fcc_table=tables+fcc_offset;
43: for(int i=0; i<0x100; i++)
1.53 paf 44: lcc_table[i]=fcc_table[i]=(unsigned char)i;
1.1 paf 45: }
46: inline void cstr2ctypes(unsigned char *tables, const unsigned char *cstr,
47: unsigned char bit) {
48: unsigned char *ctypes_table=tables+ctypes_offset;
49: ctypes_table[0]=bit;
50: for(; *cstr; cstr++) {
51: unsigned char c=*cstr;
52: ctypes_table[c]|=bit;
53: }
54: }
1.35 paf 55: inline unsigned int to_wchar_code(const char* cstr) {
1.1 paf 56: if(!cstr || !*cstr)
57: return 0;
58: if(cstr[1]==0)
1.4 paf 59: return(unsigned int)(unsigned char)cstr[0];
1.1 paf 60:
61: char *error_pos;
1.4 paf 62: return(unsigned int)strtol(cstr, &error_pos, 0);
1.1 paf 63: }
1.35 paf 64: inline bool to_bool(const char* cstr) {
1.1 paf 65: return cstr && *cstr!=0;
66: }
67: static void element2ctypes(unsigned char c, bool belongs,
68: unsigned char *tables, unsigned char bit, int group_offset=-1) {
69: if(!belongs)
70: return;
71:
72: unsigned char *ctypes_table=tables+ctypes_offset;
73:
74: ctypes_table[c]|=bit;
75: if(group_offset>=0)
1.4 paf 76: tables[cbits_offset+group_offset+c/8] |= 1<<(c%8);
1.1 paf 77: }
78: static void element2case(unsigned char from, unsigned char to,
79: unsigned char *tables) {
80: if(!to)
81: return;
82:
83: unsigned char *lcc_table=tables+lcc_offset;
84: unsigned char *fcc_table=tables+fcc_offset;
85: lcc_table[from]=to;
86: fcc_table[from]=to; fcc_table[to]=from;
87: }
88:
89: // methods
90:
1.37 paf 91: Charset::Charset(Request_charsets* charsets, const String::Body ANAME, const String* afile_spec):
1.35 paf 92: FNAME(ANAME),
93: FNAME_CSTR(ANAME.cstrm()) {
1.7 paf 94:
1.35 paf 95: if(afile_spec) {
1.1 paf 96: fisUTF8=false;
1.35 paf 97: load_definition(*charsets, *afile_spec);
1.1 paf 98: #ifdef XML
1.35 paf 99: addEncoding(FNAME_CSTR);
1.1 paf 100: #endif
101: } else {
102: fisUTF8=true;
1.4 paf 103: // grab default onces [for UTF-8 so to be able to make a-z =>A-Z
1.65 misha 104: memcpy(pcre_tables, _pcre_default_tables, sizeof(pcre_tables));
1.1 paf 105: }
106:
107: #ifdef XML
1.35 paf 108: initTranscoder(FNAME, FNAME_CSTR);
1.1 paf 109: #endif
110: }
111:
1.35 paf 112: void Charset::load_definition(Request_charsets& charsets, const String& afile_spec) {
1.1 paf 113: // pcre_tables
114: // lowcase, flipcase, bits digit+word+whitespace, masks
115:
116: // must not move this inside of prepare_case_tables
117: // don't know the size there
118: memset(pcre_tables, 0, sizeof(pcre_tables));
119: prepare_case_tables(pcre_tables);
1.4 paf 120: cstr2ctypes(pcre_tables,(const unsigned char *)"*+?{^.$|()[", ctype_meta);
1.1 paf 121:
122: // charset
1.35 paf 123: memset(&tables, 0, sizeof(tables));
1.1 paf 124:
125: // loading text
1.35 paf 126: char *data=file_read_text(charsets, afile_spec);
1.1 paf 127:
128: // ignore header
129: getrow(&data);
130:
131: // parse cells
132: char *row;
1.42 paf 133: while((row=getrow(&data))) {
1.1 paf 134: // remove empty&comment lines
135: if(!*row || *row=='#')
136: continue;
137:
138: // char white-space digit hex-digit letter word lowercase unicode1 unicode2
1.53 paf 139: unsigned char c=0;
1.1 paf 140: char *cell;
1.42 paf 141: for(int column=0; (cell=lsplit(&row, '\t')); column++) {
1.1 paf 142: switch(column) {
1.53 paf 143: case 0: c=(unsigned char)to_wchar_code(cell); break;
1.1 paf 144: // pcre_tables
145: case 1: element2ctypes(c, to_bool(cell), pcre_tables, ctype_space, cbit_space); break;
146: case 2: element2ctypes(c, to_bool(cell), pcre_tables, ctype_digit, cbit_digit); break;
147: case 3: element2ctypes(c, to_bool(cell), pcre_tables, ctype_xdigit); break;
148: case 4: element2ctypes(c, to_bool(cell), pcre_tables, ctype_letter); break;
149: case 5: element2ctypes(c, to_bool(cell), pcre_tables, ctype_word, cbit_word); break;
1.53 paf 150: case 6: element2case(c, (unsigned char)to_wchar_code(cell), pcre_tables); break;
1.1 paf 151: case 7:
152: case 8:
153: // charset
1.10 paf 154: if(tables.toTableSize>MAX_CHARSET_UNI_CODES)
1.56 misha 155: throw Exception(PARSER_RUNTIME,
1.35 paf 156: &afile_spec,
1.1 paf 157: "charset must contain not more then %d unicode values", MAX_CHARSET_UNI_CODES);
158:
159: XMLCh unicode=(XMLCh)to_wchar_code(cell);
160: if(!unicode && column==7/*unicode1 column*/)
161: unicode=(XMLCh)c;
162: if(unicode) {
1.10 paf 163: if(!tables.fromTable[c])
164: tables.fromTable[c]=unicode;
165: tables.toTable[tables.toTableSize].intCh=unicode;
166: tables.toTable[tables.toTableSize].extCh=(XMLByte)c;
167: tables.toTableSize++;
1.1 paf 168: }
169: break;
170: }
171: }
172: };
173:
174: // sort by the Unicode code point
175: sort_ToTable();
176: }
177:
178: static int sort_cmp_Trans_rec_intCh(const void *a, const void *b) {
179: return
1.38 paf 180: static_cast<const Charset::Tables::Rec *>(a)->intCh-
181: static_cast<const Charset::Tables::Rec *>(b)->intCh;
1.1 paf 182: }
183:
184: void Charset::sort_ToTable() {
1.10 paf 185: _qsort(tables.toTable, tables.toTableSize, sizeof(*tables.toTable),
1.1 paf 186: sort_cmp_Trans_rec_intCh);
187: //FILE *f=fopen("c:\\temp\\a", "wb");
1.10 paf 188: //fwrite(tables.toTable, tables.toTableSize, sizeof(*tables.toTable), f);
1.1 paf 189: //fclose(f);
190: }
191:
1.60 misha 192: // @todo: precache for spedup searching
1.10 paf 193: static XMLByte xlatOneTo(const XMLCh toXlat,
1.35 paf 194: const Charset::Tables& tables,
195: XMLByte not_found) {
1.39 paf 196: int lo = 0;
197: int hi = tables.toTableSize - 1;
198: while(lo<=hi) {
1.35 paf 199: // Calc the mid point of the low and high offset.
1.39 paf 200: const unsigned int i = (lo + hi) / 2;
201:
202: XMLCh cur=tables.toTable[i].intCh;
203: if(toXlat==cur)
204: return tables.toTable[i].extCh;
205: if(toXlat>cur)
206: lo = i+1;
1.1 paf 207: else
1.39 paf 208: hi = i-1;
209: }
1.35 paf 210:
211: return not_found;
1.1 paf 212: }
213:
1.35 paf 214: String::C Charset::transcode(const String::C src,
215: const Charset& source_charset,
216: const Charset& dest_charset) {
217: if(!src.length)
218: return String::C("", 0);
1.4 paf 219:
1.1 paf 220: switch((source_charset.isUTF8()?0x10:0x00)|(dest_charset.isUTF8()?0x01:0x00)) {
221: default: // 0x00
1.35 paf 222: return source_charset.transcodeToCharset(src, dest_charset);
1.1 paf 223: case 0x01:
1.35 paf 224: return source_charset.transcodeToUTF8(src);
1.1 paf 225: case 0x10:
1.35 paf 226: return dest_charset.transcodeFromUTF8(src);
1.1 paf 227: case 0x11:
1.35 paf 228: return src;
1.1 paf 229: }
230: }
231:
232: // ---------------------------------------------------------------------------
233: // Local static data
234: //
235: // gUTFBytes
236: // A list of counts of trailing bytes for each initial byte in the input.
237: //
238: // gUTFOffsets
239: // A list of values to offset each result char type, according to how
240: // many source bytes when into making it.
241: //
242: // gFirstByteMark
243: // A list of values to mask onto the first byte of an encoded sequence,
244: // indexed by the number of bytes used to create the sequence.
245: // ---------------------------------------------------------------------------
246: static const XMLByte gUTFBytes[0x100] = {
247: 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0
248: , 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0
249: , 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0
250: , 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0
251: , 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0
252: , 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0
253: , 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0
254: , 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0
255: , 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0
256: , 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0
257: , 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0
258: , 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0
259: , 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1
260: , 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1
261: , 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2
262: , 3, 3, 3, 3, 3, 3, 3, 3, 4, 4, 4, 4, 5, 5, 5, 5
263: };
264:
265: static const uint gUTFOffsets[6] = {
266: 0, 0x3080, 0xE2080, 0x3C82080, 0xFA082080, 0x82082080
267: };
268:
269: static const XMLByte gFirstByteMark[7] = {
270: 0x00, 0x00, 0xC0, 0xE0, 0xF0, 0xF8, 0xFC
271: };
272:
1.35 paf 273: static int transcodeToUTF8(const XMLByte* srcData, size_t& srcLen,
274: XMLByte *toFill, size_t& toFillLen,
275: const Charset::Tables& tables) {
1.11 paf 276: const XMLByte* srcPtr=srcData;
277: const XMLByte* srcEnd=srcData+srcLen;
278: XMLByte* outPtr=toFill;
279: XMLByte* outEnd=toFill+toFillLen;
1.1 paf 280:
1.35 paf 281: while(srcPtr<srcEnd) {
282: uint curVal = tables.fromTable[*srcPtr];
1.1 paf 283: if(!curVal) {
1.35 paf 284: // use the replacement character
285: *outPtr++= '?';
286: srcPtr++;
287: continue;
288: }
1.1 paf 289:
1.35 paf 290: // Figure out how many bytes we need
291: unsigned int encodedBytes;
292: if(curVal<0x80)
293: encodedBytes = 1;
294: else if(curVal<0x800)
295: encodedBytes = 2;
296: else if(curVal<0x10000)
297: encodedBytes = 3;
298: else if(curVal<0x200000)
299: encodedBytes = 4;
300: else if(curVal<0x4000000)
301: encodedBytes = 5;
302: else if(curVal<= 0x7FFFFFFF)
303: encodedBytes = 6;
304: else {
305: // use the replacement character
306: *outPtr++= '?';
307: srcPtr++;
308: continue;
309: }
1.11 paf 310:
1.35 paf 311: // If we cannot fully get this char into the output buffer
312: if (outPtr + encodedBytes > outEnd)
313: break;
314:
315: // We can do it, so update the source index
316: srcPtr++;
317:
318: // And spit out the bytes. We spit them out in reverse order
319: // here, so bump up the output pointer and work down as we go.
320: outPtr+= encodedBytes;
321: switch(encodedBytes) {
1.60 misha 322: case 6: *--outPtr = XMLByte((curVal | 0x80UL) & 0xBFUL);
323: curVal>>= 6;
324: case 5: *--outPtr = XMLByte((curVal | 0x80UL) & 0xBFUL);
325: curVal>>= 6;
326: case 4: *--outPtr = XMLByte((curVal | 0x80UL) & 0xBFUL);
327: curVal>>= 6;
328: case 3: *--outPtr = XMLByte((curVal | 0x80UL) & 0xBFUL);
329: curVal>>= 6;
330: case 2: *--outPtr = XMLByte((curVal | 0x80UL) & 0xBFUL);
331: curVal>>= 6;
332: case 1: *--outPtr = XMLByte(curVal | gFirstByteMark[encodedBytes]);
1.35 paf 333: }
334:
335: // Add the encoded bytes back in again to indicate we've eaten them
336: outPtr+= encodedBytes;
337: }
338:
339: // Update the bytes eaten
340: srcLen = srcPtr - srcData;
341:
342: // Return the characters read
343: toFillLen = outPtr - toFill;
344:
1.29 paf 345: //return srcPtr==srcEnd?(int)toFillLen:-1;
346: /*
347: xmlCharEncodingInputFunc
348: Returns :
349: the number of byte written, or -1 by lack of space, or -2 if the transcoding failed. The value of inlen after return is the
350: number of octets consumed as the return value is positive, else unpredictiable. The value of outlen after return is the number
351: of ocetes consumed.
352: */
353: return 0;
1.1 paf 354: }
1.26 paf 355: /// @todo digital entites only when xml/html output [at output in html/xml mode, in html part of a letter]
1.35 paf 356: static int transcodeFromUTF8(const XMLByte* srcData, size_t& srcLen,
357: XMLByte* toFill, size_t& toFillLen,
358: const Charset::Tables& tables) {
1.11 paf 359: const XMLByte* srcPtr=srcData;
360: const XMLByte* srcEnd=srcData+srcLen;
361: XMLByte* outPtr=toFill;
362: XMLByte* outEnd=toFill+toFillLen;
1.1 paf 363:
1.35 paf 364: // We now loop until we either run out of input data, or room to store
365: while ((srcPtr < srcEnd) && (outPtr < outEnd)) {
366: // Get the next leading byte out
367: const XMLByte firstByte =* srcPtr;
368:
369: // Special-case ASCII, which is a leading byte value of<= 127
1.60 misha 370: if(firstByte<=127) {
1.35 paf 371: *outPtr++= firstByte;
372: srcPtr++;
373: continue;
374: }
375:
376: // See how many trailing src bytes this sequence is going to require
377: const unsigned int trailingBytes = gUTFBytes[firstByte];
378:
379: // If there are not enough source bytes to do this one, then we
380: // are done. Note that we done>= here because we are implicitly
381: // counting the 1 byte we get no matter what.
382: if(srcPtr+trailingBytes>= srcEnd)
383: break;
384:
385: // Looks ok, so lets build up the value
386: uint tmpVal=0;
387: switch(trailingBytes) {
388: case 5: tmpVal+=*srcPtr++; tmpVal<<=6;
389: case 4: tmpVal+=*srcPtr++; tmpVal<<=6;
390: case 3: tmpVal+=*srcPtr++; tmpVal<<=6;
391: case 2: tmpVal+=*srcPtr++; tmpVal<<=6;
392: case 1: tmpVal+=*srcPtr++; tmpVal<<=6;
393: case 0: tmpVal+=*srcPtr++;
394: break;
395:
396: default:
397: throw Exception(0,
398: 0,
1.49 paf 399: "transcodeFromUTF8 error: wrong trailingBytes value(%d)", trailingBytes); // never
1.35 paf 400: }
401: tmpVal-=gUTFOffsets[trailingBytes];
402:
403: // If it will fit into a single char, then put it in. Otherwise
404: // fail [*encode it as a surrogate pair. If its not valid, use the
405: // replacement char.*]
406: if(!(tmpVal & 0xFFFF0000)) {
1.25 paf 407: if(XMLByte xlat=xlatOneTo(tmpVal, tables, 0))
408: *outPtr++=xlat;
1.49 paf 409: else {
1.50 paf 410: outPtr+=sprintf((char *)outPtr, "&#%u;", tmpVal); // &#decimal;
1.49 paf 411: }
412: } else {
413: const XMLByte* recoverPtr=srcPtr-trailingBytes-1;
414: for(uint i=0; i<=trailingBytes; i++)
415: outPtr+=sprintf((char*)outPtr, "%%%02X", *recoverPtr++);
416: }
1.1 paf 417: }
1.35 paf 418:
419: // Update the bytes eaten
420: srcLen = srcPtr - srcData;
421:
422: // Return the characters read
423: toFillLen = outPtr - toFill;
1.11 paf 424:
1.29 paf 425: //return srcPtr==srcEnd?(int)toFillLen:-1;
426: /*
427: xmlCharEncodingOutputFunc
428: Returns :
429: the number of byte written, or -1 by lack of space, or -2 if the transcoding failed. The value of inlen after return is the
430: number of octets consumed as the return value is positive, else unpredictiable. The value of outlen after return is the number
431: of ocetes consumed.
432: */
433: return 0;
1.10 paf 434: }
435:
1.67 misha 436: static bool is_escaped(XMLByte c){
1.60 misha 437: return
1.66 misha 438: !(
439: (c<=127)
440: && (
441: ((c>='0') && (c<='9'))
442: || ((c>='A') && (c<='Z'))
443: || ((c>='a') && (c<='z'))
444: || strchr("*@-_+./", c)!=0
445: )
446: );
1.60 misha 447: }
448:
449: // read one utf8 character, return number of bytes needed for store it
1.61 misha 450: static unsigned int readChar(const XMLByte*& srcPtr, const XMLByte* srcEnd, XMLByte& firstByte, XMLCh& UTF8Char){
1.60 misha 451: if(!srcPtr || !*srcPtr || srcPtr>=srcEnd)
452: return 0;
453:
454: firstByte=*srcPtr;
455:
456: if(firstByte<=127){
457: UTF8Char=firstByte;
458: srcPtr++;
459: return 1;
460: }
461:
462: unsigned int trailingBytes=gUTFBytes[firstByte];
463:
464: if(srcPtr+trailingBytes>=srcEnd){
465: return 0; // not enough bytes in source string for reading
466: }
467:
468: uint tmpVal=0;
469: switch(trailingBytes){
470: case 5: tmpVal+=*srcPtr++; tmpVal<<=6;
471: case 4: tmpVal+=*srcPtr++; tmpVal<<=6;
472: case 3: tmpVal+=*srcPtr++; tmpVal<<=6;
473: case 2: tmpVal+=*srcPtr++; tmpVal<<=6;
474: case 1: tmpVal+=*srcPtr++; tmpVal<<=6;
475: case 0: tmpVal+=*srcPtr++;
476: }
477:
478: tmpVal-=gUTFOffsets[trailingBytes];
479: UTF8Char=tmpVal;
480:
481: return trailingBytes+1;
482: }
483:
1.62 misha 484: static unsigned int skipChar(const XMLByte*& srcPtr, const XMLByte* srcEnd){
485: if(!srcPtr || !*srcPtr || srcPtr>=srcEnd)
486: return 0;
487:
1.63 misha 488: unsigned int trailingBytes=gUTFBytes[*srcPtr]+1;
489: srcPtr+=trailingBytes;
1.62 misha 490:
491: return trailingBytes;
1.61 misha 492: }
493:
1.60 misha 494: // read char, return number of bytes needed for store it as UTF8
1.61 misha 495: static unsigned int readChar(const XMLByte*& srcPtr, const XMLByte* srcEnd, XMLByte& firstByte, XMLCh& UTF8Char, const Charset::Tables& tables){
1.60 misha 496: if(!srcPtr || !*srcPtr || srcPtr>=srcEnd)
497: return 0;
498:
499: firstByte=*srcPtr++;
500: UTF8Char=tables.fromTable[firstByte];
501:
502: if(UTF8Char<0x80)
503: return 1;
504: else if(UTF8Char<0x800)
505: return 2;
506: else if(UTF8Char<0x10000)
507: return 3;
508: else if(UTF8Char<0x200000)
509: return 4;
510: else if(UTF8Char<0x4000000)
511: return 5;
512: else if(UTF8Char<= 0x7FFFFFFF)
513: return 6;
514:
515: // will use the replacement character '?'
516: firstByte=0;
517: return 1;
518: }
519:
520: static int escape(const XMLByte* srcData, size_t& srcLen,
521: XMLByte* toFill, size_t& toFillLen) {
522: const XMLByte* srcPtr=srcData;
523: const XMLByte* srcEnd=srcData+srcLen;
524: XMLByte* outPtr=toFill;
525: XMLByte* outEnd=toFill+toFillLen;
526: XMLByte firstByte;
527: XMLCh UTF8Char;
528: uint charSize;
529:
530: // loop until we either run out of input data, or room to store
531: while((outPtr < outEnd) && (charSize=readChar(srcPtr, srcEnd, firstByte, UTF8Char))){
532: if(charSize==1){
533: if(is_escaped(firstByte)) // %XX
534: outPtr+=sprintf((char*)outPtr, "%%%02X", firstByte);
535: else
536: *outPtr++=firstByte;
537: } else
538: outPtr+=sprintf((char*)outPtr, "%%u%04X", UTF8Char); // %uXXXX
539: }
540:
541: // Update the bytes eaten
542: srcLen=srcPtr-srcData;
543:
544: // Return the characters read
545: toFillLen=outPtr-toFill;
546:
547: return 0;
548: }
549:
550: static int escape(const XMLByte* srcData, size_t& srcLen,
551: XMLByte *toFill, size_t& toFillLen,
552: const Charset::Tables& tables) {
553: const XMLByte* srcPtr=srcData;
554: const XMLByte* srcEnd=srcData+srcLen;
555: XMLByte* outPtr=toFill;
556: //XMLByte* outEnd=toFill+toFillLen;
557: XMLByte firstByte;
558: XMLCh UTF8Char;
559: uint charSize;
560:
561: while(charSize=readChar(srcPtr, srcEnd, firstByte, UTF8Char, tables)){
562: if(charSize==1){
563: if(firstByte){
564: if(is_escaped(firstByte)) // %XX
565: outPtr+=sprintf((char*)outPtr, "%%%02X", firstByte);
566: else
567: *outPtr++=firstByte;
568: } else // add replacement char '?'
569: *outPtr++='?';
570: } else
571: outPtr+=sprintf((char*)outPtr, "%%u%04X", UTF8Char); // %uXXXX
572: }
573:
574: // Update the bytes eaten
575: srcLen = srcPtr - srcData;
576:
577: // Return the characters read
578: toFillLen = outPtr - toFill;
579:
580: return 0;
581: }
582:
583:
584: String::C Charset::escape(const String::C src, const Charset& source_charset){
585: size_t src_length=src.length;
586: if(!src_length)
587: return String::C("", 0);
588:
589: #ifdef PRECALCULATE_DEST_LENGTH
590: size_t dest_length=0;
591: const XMLByte* srcPtr=(XMLByte*)src.str;
592: const XMLByte* srcEnd=srcPtr+src_length;
593: XMLByte firstByte;
594: XMLCh UTF8Char;
595:
596: if(source_charset.isUTF8()){
597: while(uint charSize=readChar(srcPtr, srcEnd, firstByte, UTF8Char)){
598: if(charSize==1)
599: dest_length+=!is_escaped(firstByte)?1:3/*%XX*/;
600: else
601: dest_length+=6; // '%uXXXX'
602: }
603: } else {
604: while(uint charSize=readChar(srcPtr, srcEnd, firstByte, UTF8Char, source_charset.tables)){
605: if(charSize==1)
606: dest_length+=(!firstByte/*replacement char '?'*/ || !is_escaped(firstByte))?1:3/*'%XX'*/;
607: else
608: dest_length+=6; // '%uXXXX'
609: }
610: }
611: #else
612: size_t dest_length=src_length*6; // enough for %uXXXX but too memory-hungry
613: #endif
614:
615: //throw Exception(0,0,"%u",dest_length);
616:
617: #ifndef NDEBUG
618: size_t saved_dest_length=dest_length;
619: #endif
620: XMLByte *dest_body=new(PointerFreeGC) XMLByte[dest_length+1/*for terminator*/];
621:
622: int status;
623: if(source_charset.isUTF8()){
624: status=::escape((XMLByte *)src.str, src_length, dest_body, dest_length);
625: } else {
626: status=::escape((XMLByte *)src.str, src_length, dest_body, dest_length, source_charset.tables);
627: }
628:
629: if(status<0)
630: throw Exception(0,
631: 0,
632: "Charset::escapeString buffer overflow");
633:
634: assert(dest_length<=saved_dest_length);
635: dest_body[dest_length]=0; // terminator
636: return String::C((char*)dest_body, dest_length);
637: }
638:
1.64 misha 639: String::Body Charset::escape(const String::Body src, const Charset& source_charset) {
640: const char *src_ptr=src.cstr();
641: size_t src_size=strlen(src_ptr);
642:
643: String::C dest=Charset::escape(String::C(src_ptr, src_size),
644: source_charset);
645:
646: return String::Body(dest.str, dest.length);
647: }
648:
649: String& Charset::escape(const String& src, const Charset& source_charset) {
650: if(!src.length())
651: return *new String("", 0, false);
652:
653: return *new String(escape((String::Body)src, source_charset), String::L_CLEAN);
654: }
1.60 misha 655:
1.35 paf 656: const String::C Charset::transcodeToUTF8(const String::C src) const {
657: size_t src_length=src.length;
1.60 misha 658:
659: #ifdef PRECALCULATE_DEST_LENGTH
660: size_t dest_length=0;
661: const XMLByte* srcPtr=(XMLByte*)src.str;
662: const XMLByte* srcEnd=srcPtr+src_length;
1.69 ! misha 663: XMLByte firstByte;
! 664: XMLCh UTF8Char;
! 665: while(uint charSize=readChar(srcPtr, srcEnd, firstByte, UTF8Char, tables))
1.60 misha 666: dest_length+=charSize;
667: #else
668: size_t dest_length=src_length*6; // so that surly enough (max utf8 seq len=6) but too memory-hyngry
669: #endif
670:
671: //throw Exception(0,0,"%u",dest_length);
672:
1.35 paf 673: #ifndef NDEBUG
674: size_t saved_dest_length=dest_length;
675: #endif
676: XMLByte *dest_body=new(PointerFreeGC) XMLByte[dest_length+1/*for terminator*/];
1.11 paf 677:
678: if(::transcodeToUTF8(
1.35 paf 679: (XMLByte *)src.str, src_length,
680: dest_body, dest_length,
1.11 paf 681: tables)<0)
1.43 paf 682: throw Exception(0,
1.10 paf 683: 0,
1.11 paf 684: "Charset::transcodeToUTF8 buffer overflow");
1.10 paf 685:
1.60 misha 686: assert(dest_length<=saved_dest_length);
687: dest_body[dest_length]=0; // terminator
1.35 paf 688: return String::C((char*)dest_body, dest_length);
1.10 paf 689: }
1.38 paf 690:
691: static XMLCh change_case_UTF8(const XMLCh src, const Charset::UTF8CaseTable& table) {
1.39 paf 692: int lo = 0;
693: int hi = table.size - 1;
694: while(lo<=hi) {
1.38 paf 695: // Calc the mid point of the low and high offset.
1.39 paf 696: const unsigned int i = (lo + hi) / 2;
697:
698: XMLCh cur=table.records[i].from;
699: if(src==cur)
700: return table.records[i].to;
701: if(src>cur)
702: lo = i+1;
1.38 paf 703: else
1.39 paf 704: hi = i-1;
705: }
706:
707: // not found
1.38 paf 708: return src;
709: }
710:
1.58 misha 711: static void store_UTF8(XMLCh src, XMLByte*& outPtr){
1.38 paf 712: if(!src) {
713: // use the replacement character
714: *outPtr++= '?';
715: return;
716: }
717:
718: // Figure out how many bytes we need
719: unsigned int encodedBytes;
720: if(src<0x80)
721: encodedBytes = 1;
722: else if(src<0x800)
723: encodedBytes = 2;
724: else if(src<0x10000)
725: encodedBytes = 3;
726: else if(src<0x200000)
727: encodedBytes = 4;
728: else if(src<0x4000000)
729: encodedBytes = 5;
730: else if(src<= 0x7FFFFFFF)
731: encodedBytes = 6;
732: else {
733: // use the replacement character
734: *outPtr++= '?';
735: return;
736: }
737:
738: // And spit out the bytes. We spit them out in reverse order
739: // here, so bump up the output pointer and work down as we go.
740: outPtr+= encodedBytes;
741: switch(encodedBytes) {
742: case 6: *--outPtr = XMLByte((src | 0x80UL) & 0xBFUL);
743: src>>= 6;
744: case 5: *--outPtr = XMLByte((src | 0x80UL) & 0xBFUL);
745: src>>= 6;
746: case 4: *--outPtr = XMLByte((src | 0x80UL) & 0xBFUL);
747: src>>= 6;
748: case 3: *--outPtr = XMLByte((src | 0x80UL) & 0xBFUL);
749: src>>= 6;
750: case 2: *--outPtr = XMLByte((src | 0x80UL) & 0xBFUL);
751: src>>= 6;
752: case 1: *--outPtr = XMLByte(src | gFirstByteMark[encodedBytes]);
753: }
754:
755: // Add the encoded bytes back in again to indicate we've eaten them
756: outPtr+= encodedBytes;
757: }
758:
759: static void change_case_UTF8(XMLCh src, XMLByte*& outPtr,
760: const Charset::UTF8CaseTable& table) {
761: store_UTF8(change_case_UTF8(src, table), outPtr);
762: };
1.44 paf 763: void change_case_UTF8(const XMLByte* srcData, size_t srcLen,
764: XMLByte* toFill, size_t toFillLen,
765: const Charset::UTF8CaseTable& table) {
1.38 paf 766: const XMLByte* srcPtr=srcData;
1.44 paf 767: const XMLByte* srcEnd=srcData+srcLen;
1.38 paf 768: XMLByte* outPtr=toFill;
1.44 paf 769: XMLByte* outEnd=toFill+toFillLen;
770:
771: // We now loop until we either run out of input data, or room to store
772: while ((srcPtr < srcEnd) && (outPtr < outEnd)) {
773: // Get the next leading byte out
774: const XMLByte firstByte =* srcPtr;
1.38 paf 775:
1.60 misha 776: if(firstByte<=127) {
1.38 paf 777: change_case_UTF8(firstByte, outPtr, table);
778: srcPtr++;
779: continue;
780: }
781:
782: // See how many trailing src bytes this sequence is going to require
783: const unsigned int trailingBytes = gUTFBytes[firstByte];
784:
785: // Looks ok, so lets build up the value
786: uint tmpVal=0;
787: switch(trailingBytes) {
788: case 5: tmpVal+=*srcPtr++; tmpVal<<=6;
789: case 4: tmpVal+=*srcPtr++; tmpVal<<=6;
790: case 3: tmpVal+=*srcPtr++; tmpVal<<=6;
791: case 2: tmpVal+=*srcPtr++; tmpVal<<=6;
792: case 1: tmpVal+=*srcPtr++; tmpVal<<=6;
793: case 0: tmpVal+=*srcPtr++;
794: break;
795:
796: default:
797: throw Exception(0,
798: 0,
799: "change_case_UTF8 error: wrong trailingBytes value(%d)", trailingBytes);
800: }
801: tmpVal-=gUTFOffsets[trailingBytes];
802:
803: // If it will fit into a single char, then put it in. Otherwise
804: // fail [*encode it as a surrogate pair. If its not valid, use the
805: // replacement char.*]
806: if(!(tmpVal & 0xFFFF0000))
807: change_case_UTF8(tmpVal, outPtr, table);
808: else
809: throw Exception(0,
810: 0,
811: "change_case_UTF8 error: too big tmpVal(0x%08X)", tmpVal);
812: }
813:
814: if(srcPtr!=outPtr)
815: throw Exception(0,
816: 0,
817: "change_case_UTF8 error: end pointers do not match");
818: }
819:
1.60 misha 820: static size_t getDecNumLength(XMLCh UTF8Char){
821: return
822: (UTF8Char < 100)
823: ?2
824: :(UTF8Char < 1000)
825: ?3
826: :(UTF8Char < 10000)
827: ?4
828: :5;
829: }
1.38 paf 830:
1.35 paf 831: const String::C Charset::transcodeFromUTF8(const String::C src) const {
832: size_t src_length=src.length;
1.60 misha 833:
834: #ifdef PRECALCULATE_DEST_LENGTH
835: size_t dest_length=0;
836: const XMLByte* srcPtr=(XMLByte*)src.str;
837: const XMLByte* srcEnd=srcPtr+src_length;
838: XMLByte firstByte;
839: XMLCh UTF8Char;
840: while(uint charSize=readChar(srcPtr, srcEnd, firstByte, UTF8Char)){
841: if(charSize==1)
842: dest_length++;
843: else
844: dest_length+=(UTF8Char & 0xFFFF0000)
845: ?charSize*3 // '%XX' for each byte
846: :(xlatOneTo(UTF8Char, tables, 0)!=0)
847: ?1 // can convert it to single char
848: :getDecNumLength(UTF8Char)+3; // &#XX; - &#XXXXX;
849: }
850: #else
851: // so that surly enough, "&#XXX;" has max ratio (huh? 8 bytes needed for '&#XXXXX;')
852: size_t dest_length=src_length*6;
853: #endif
854:
855: //throw Exception(0,0,"%u",dest_length);
856:
1.35 paf 857: #ifndef NDEBUG
858: size_t saved_dest_length=dest_length;
859: #endif
860: XMLByte *dest_body=new(PointerFreeGC) XMLByte[dest_length+1/*for terminator*/];
1.11 paf 861:
862: if(::transcodeFromUTF8(
1.35 paf 863: (XMLByte *)src.str, src_length,
864: dest_body, dest_length,
1.11 paf 865: tables)<0)
1.43 paf 866: throw Exception(0,
1.10 paf 867: 0,
1.35 paf 868: "Charset::transcodeFromUTF8 buffer overflow");
1.10 paf 869:
1.60 misha 870: assert(dest_length<=saved_dest_length);
871: dest_body[dest_length]=0; // terminator
1.35 paf 872: return String::C((char*)dest_body, dest_length);
1.1 paf 873: }
874:
875: /// transcode using both charsets
1.35 paf 876: const String::C Charset::transcodeToCharset(const String::C src,
877: const Charset& dest_charset) const {
878: if(&dest_charset==this)
879: return src;
880: else {
881: size_t dest_length=src.length;
882: XMLByte* dest_body=new(PointerFreeGC) XMLByte[dest_length+1/*for terminator*/];
883:
884: XMLByte* output=dest_body;
885: const XMLByte* input=(XMLByte *)src.str;
886: while(XMLCh c=*input++) {
887: XMLCh curVal = tables.fromTable[c];
888: *output++=curVal?
889: xlatOneTo(curVal, dest_charset.tables, '?') // OK
890: :'?'; // use the replacement character
1.6 paf 891: }
1.1 paf 892:
1.35 paf 893: dest_body[dest_length]=0; // terminator
894: return String::C((char*)dest_body, dest_length);
1.6 paf 895: }
1.1 paf 896: }
897:
1.58 misha 898: void Charset::store_Char(XMLByte*& outPtr, XMLCh src, XMLByte not_found){
1.59 misha 899: if(isUTF8())
1.58 misha 900: store_UTF8(src, outPtr);
1.59 misha 901: else if(char ch=xlatOneTo(src, tables, not_found))
1.58 misha 902: *outPtr++=ch;
1.57 misha 903: }
904:
1.1 paf 905: #ifdef XML
1.10 paf 906:
1.35 paf 907: static const Charset::Tables* tables[MAX_CHARSETS];
908:
1.46 paf 909: #ifdef PA_PATCHED_LIBXML_BACKWARD
910:
911: #define declareXml256ioFuncs(i) \
912: static int xml256CharEncodingInputFunc##i( \
913: unsigned char *out, int *outlen, \
914: const unsigned char *in, int *inlen, void*) { \
915: return transcodeToUTF8( \
916: in, *(size_t*)inlen, \
917: out, *(size_t*)outlen, \
918: *tables[i]); \
919: } \
920: static int xml256CharEncodingOutputFunc##i( \
921: unsigned char *out, int *outlen, \
922: const unsigned char *in, int *inlen, void*) { \
923: return transcodeFromUTF8( \
924: in, *(size_t*)inlen, \
925: out, *(size_t*)outlen, \
926: *tables[i]); \
927: }
928:
929: #else
930:
1.35 paf 931: #define declareXml256ioFuncs(i) \
932: static int xml256CharEncodingInputFunc##i( \
933: unsigned char *out, int *outlen, \
934: const unsigned char *in, int *inlen) { \
935: return transcodeToUTF8( \
936: in, *(size_t*)inlen, \
937: out, *(size_t*)outlen, \
938: *tables[i]); \
939: } \
940: static int xml256CharEncodingOutputFunc##i( \
941: unsigned char *out, int *outlen, \
942: const unsigned char *in, int *inlen) { \
943: return transcodeFromUTF8( \
944: in, *(size_t*)inlen, \
945: out, *(size_t*)outlen, \
946: *tables[i]); \
947: }
948:
1.46 paf 949: #endif
950:
951:
1.35 paf 952: declareXml256ioFuncs(0) declareXml256ioFuncs(1)
953: declareXml256ioFuncs(2) declareXml256ioFuncs(3)
954: declareXml256ioFuncs(4) declareXml256ioFuncs(5)
955: declareXml256ioFuncs(6) declareXml256ioFuncs(7)
956: declareXml256ioFuncs(8) declareXml256ioFuncs(9)
957:
958: static xmlCharEncodingInputFunc inputFuncs[MAX_CHARSETS]={
959: xml256CharEncodingInputFunc0, xml256CharEncodingInputFunc1,
960: xml256CharEncodingInputFunc2, xml256CharEncodingInputFunc3,
961: xml256CharEncodingInputFunc4, xml256CharEncodingInputFunc5,
962: xml256CharEncodingInputFunc6, xml256CharEncodingInputFunc7,
963: xml256CharEncodingInputFunc8, xml256CharEncodingInputFunc9
964: };
965: static xmlCharEncodingOutputFunc outputFuncs[MAX_CHARSETS]={
966: xml256CharEncodingOutputFunc0, xml256CharEncodingOutputFunc1,
967: xml256CharEncodingOutputFunc2, xml256CharEncodingOutputFunc3,
968: xml256CharEncodingOutputFunc4, xml256CharEncodingOutputFunc5,
969: xml256CharEncodingOutputFunc6, xml256CharEncodingOutputFunc7,
970: xml256CharEncodingOutputFunc8, xml256CharEncodingOutputFunc9
971: };
972: static size_t handlers_count=0;
1.10 paf 973:
974: void Charset::addEncoding(char *name_cstr) {
1.35 paf 975: if(handlers_count==MAX_CHARSETS)
976: throw Exception(0,
977: 0,
978: "already allocated %d handlers, no space for new encoding '%s'",
979: MAX_CHARSETS, name_cstr);
980:
1.45 paf 981: xmlCharEncodingHandler* handler=new(UseGC) xmlCharEncodingHandler;
1.35 paf 982: {
983: handler->name=name_cstr;
984: handler->input=inputFuncs[handlers_count];
985: handler->output=outputFuncs[handlers_count];
986: ::tables[handlers_count]=&tables;
987: handlers_count++;
988: }
1.10 paf 989:
990: xmlRegisterCharEncodingHandler(handler);
1.35 paf 991:
1.10 paf 992: }
993:
1.37 paf 994: void Charset::initTranscoder(const String::Body NAME, const char* name_cstr) {
1.15 paf 995: ftranscoder=xmlFindCharEncodingHandler(name_cstr);
1.35 paf 996: transcoder(NAME); // check right way
1.15 paf 997: }
998:
1.37 paf 999: xmlCharEncodingHandler& Charset::transcoder(const String::Body NAME) {
1.15 paf 1000: if(!ftranscoder)
1.56 misha 1001: throw Exception(PARSER_RUNTIME,
1.35 paf 1002: new String(NAME, String::L_TAINTED),
1.10 paf 1003: "unsupported encoding");
1.35 paf 1004: return *ftranscoder;
1.10 paf 1005: }
1006:
1.54 paf 1007: String::C Charset::transcode_cstr(const xmlChar* s) {
1.13 paf 1008: if(!s)
1.35 paf 1009: return String::C("", 0);
1.8 paf 1010:
1.35 paf 1011: int inlen=strlen((const char*)s);
1.51 paf 1012: int outlen=inlen*6/*strlen("ÿ")*/; // max
1.35 paf 1013: #ifndef NDEBUG
1014: int saved_outlen=outlen;
1015: #endif
1016: char *out=new(PointerFreeGC) char[outlen+1];
1.8 paf 1017:
1.30 paf 1018: int error;
1.35 paf 1019: if(xmlCharEncodingOutputFunc output=transcoder(FNAME).output) {
1.30 paf 1020: error=output(
1.17 paf 1021: (unsigned char*)out, &outlen,
1.46 paf 1022: (const unsigned char*)s, &inlen
1023: #ifdef PA_PATCHED_LIBXML_BACKWARD
1024: ,0
1025: #endif
1026: );
1.30 paf 1027: } else {
1028: memcpy(out, s, outlen=inlen);
1029: error=0;
1030: }
1031: if(error<0)
1.23 paf 1032: throw Exception(0,
1.8 paf 1033: 0,
1.30 paf 1034: "transcode_cstr failed (%d)", error);
1.8 paf 1035:
1.35 paf 1036: assert(outlen<=saved_outlen); out[outlen]=0;
1037: return String::C(out, outlen);
1.14 paf 1038: }
1.54 paf 1039: const String& Charset::transcode(const xmlChar* s) {
1.35 paf 1040: String::C cstr=transcode_cstr(s);
1041: return *new String(cstr.str, cstr.length, true);
1.1 paf 1042: }
1043:
1.8 paf 1044: /// @test less memory using -maybe- xmlParserInputBufferCreateMem
1.35 paf 1045: xmlChar* Charset::transcode_buf2xchar(const char* buf, size_t buf_size) {
1046: xmlChar* out;
1.30 paf 1047: int outlen;
1048: int error;
1.35 paf 1049: #ifndef NDEBUG
1050: int saved_outlen;
1051: #endif
1052: if(xmlCharEncodingInputFunc input=transcoder(FNAME).input) {
1.51 paf 1053: outlen=buf_size*6/*max UTF8 bytes per char*/;
1.35 paf 1054: #ifndef NDEBUG
1055: saved_outlen=outlen;
1056: #endif
1.47 paf 1057: out=(xmlChar*)xmlMalloc(outlen+1);
1.30 paf 1058: error=input(
1.17 paf 1059: out, &outlen,
1.46 paf 1060: (const unsigned char*)buf, (int*)&buf_size
1061: #ifdef PA_PATCHED_LIBXML_BACKWARD
1062: ,0
1063: #endif
1064: );
1.30 paf 1065: } else {
1066: outlen=buf_size;
1.35 paf 1067: #ifndef NDEBUG
1068: saved_outlen=outlen;
1069: #endif
1070: out=(xmlChar*)xmlMalloc(outlen+1);
1.30 paf 1071: memcpy(out, buf, outlen);
1072: error=0;
1073: }
1.17 paf 1074:
1.30 paf 1075: if(error<0)
1.23 paf 1076: throw Exception(0,
1.8 paf 1077: 0,
1.30 paf 1078: "transcode_buf failed (%d)", error);
1.8 paf 1079:
1.35 paf 1080: assert(outlen<=saved_outlen); out[outlen]=0;
1081: return out;
1.24 paf 1082: }
1.54 paf 1083: xmlChar* Charset::transcode(const String& s) {
1.35 paf 1084: const char* cstr=s.cstr(String::L_UNSPECIFIED);
1.1 paf 1085:
1.54 paf 1086: return transcode_buf2xchar(cstr, strlen(cstr));
1.1 paf 1087: }
1.54 paf 1088: xmlChar* Charset::transcode(const String::Body s) {
1.35 paf 1089: const char* cstr=s.cstr();
1090:
1.54 paf 1091: return transcode_buf2xchar(cstr, s.length());
1.35 paf 1092: }
1.36 paf 1093: #endif
1.34 paf 1094:
1.37 paf 1095: String::Body Charset::transcode(const String::Body src,
1.34 paf 1096: const Charset& source_transcoder,
1.35 paf 1097: const Charset& dest_transcoder) {
1.34 paf 1098:
1.35 paf 1099: const char *src_ptr=src.cstr();
1.34 paf 1100: size_t src_size=strlen(src_ptr);
1101:
1.35 paf 1102: String::C dest=Charset::transcode(String::C(src_ptr, src_size),
1103: source_transcoder,
1104: dest_transcoder);
1.34 paf 1105:
1.37 paf 1106: return String::Body(dest.str, dest.length);
1.35 paf 1107: }
1108:
1109: String& Charset::transcode(const String& src,
1110: const Charset& source_transcoder,
1111: const Charset& dest_transcoder) {
1112: if(!src.length())
1113: return *new String("", 0, false);
1.34 paf 1114:
1.37 paf 1115: return *new String(transcode((String::Body)src, source_transcoder, dest_transcoder), String::L_CLEAN);
1.34 paf 1116: }
1117:
1.35 paf 1118: void Charset::transcode(ArrayString& src,
1.34 paf 1119: const Charset& source_transcoder,
1.35 paf 1120: const Charset& dest_transcoder) {
1121: for(size_t i=0; i<src.count(); i++)
1122: src.put(i, &transcode(*src[i], source_transcoder, dest_transcoder));
1.34 paf 1123: }
1124:
1125: #ifndef DOXYGEN
1126: struct Transcode_pair_info {
1127: const Charset* source_transcoder;
1128: const Charset* dest_transcoder;
1129: };
1130: #endif
1.40 paf 1131: static void transcode_pair(const String::Body /*akey*/,
1.37 paf 1132: String::Body& avalue,
1.35 paf 1133: Transcode_pair_info* info) {
1134: avalue=Charset::transcode(avalue,
1135: *info->source_transcoder,
1136: *info->dest_transcoder);
1.34 paf 1137: }
1.61 misha 1138:
1.35 paf 1139: void Charset::transcode(HashStringString& src,
1.34 paf 1140: const Charset& source_transcoder,
1.35 paf 1141: const Charset& dest_transcoder) {
1142: Transcode_pair_info info={&source_transcoder, &dest_transcoder};
1.55 paf 1143: src.for_each_ref<Transcode_pair_info*>(transcode_pair, &info);
1.34 paf 1144: }
1.61 misha 1145:
1146: size_t getUTF8BytePos(const XMLByte* srcBegin, const XMLByte* srcEnd, size_t charPos){
1147: const XMLByte* ptr=srcBegin;
1.62 misha 1148: while(charPos-- && skipChar(ptr, srcEnd));
1.61 misha 1149:
1150: return ptr-srcBegin;
1151: }
1152:
1153: size_t getUTF8CharPos(const XMLByte* srcBegin, const XMLByte* srcEnd, size_t bytePos){
1154: size_t charPos=0;
1155: const XMLByte* ptr=srcBegin;
1156: const XMLByte* ptrEnd=srcBegin+bytePos;
1.62 misha 1157: while(skipChar(ptr, srcEnd)){
1.61 misha 1158: if(ptr>ptrEnd)
1159: return charPos;
1160: charPos++;
1161: }
1162:
1163: // scan till end but position in bytes still too low
1164: throw Exception(0,
1165: 0,
1166: "Error convertion byte pos to char pos");
1167: }
1168:
1169: size_t lengthUTF8(const XMLByte* srcBegin, const XMLByte* srcEnd){
1170: size_t size=0;
1.62 misha 1171: while(skipChar(srcBegin, srcEnd))
1.61 misha 1172: size++;
1173:
1174: return size;
1175: }
E-mail:
![[BACK]](/cvsweb/icons/back.gif){kind=icon}
Return to pa_charset.C CVS log ![[TXT]](/cvsweb/icons/text.gif){kind=icon}
![[DIR]](/cvsweb/icons/dir.gif){kind=icon}
Up to [parser3project] / parser3 / src / main