Annotation of parser3/src/main/pa_charset.C, revision 1.83
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:
1.83 ! misha 8: static const char * const IDENT_CHARSET_C="$Date: 2009-10-05 11:38:12 $";
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) {
1.80 misha 196: int lo = 0;
197: int hi = tables.toTableSize - 1;
1.39 paf 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] = {
1.80 misha 266: 0, 0x3080, 0xE2080, 0x3C82080, 0xFA082080, 0x82082080
1.1 paf 267: };
268:
269: static const XMLByte gFirstByteMark[7] = {
1.80 misha 270: 0x00, 0x00, 0xC0, 0xE0, 0xF0, 0xF8, 0xFC
1.1 paf 271: };
272:
1.71 misha 273: static int transcodeToUTF8(const XMLByte* srcData, int& srcLen,
274: XMLByte *toFill, int& 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]
1.71 misha 356: static int transcodeFromUTF8(const XMLByte* srcData, int& srcLen,
357: XMLByte* toFill, int& 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:
1.70 misha 436: static bool isEscaped(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:
1.70 misha 449: // read one UTF8 char and return length of this char (in bytes)
450: static unsigned int readUTF8Char(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:
1.70 misha 484: // skip UTF8 char and return length of this char (in bytes)
485: static unsigned int skipUTF8Char(const XMLByte*& srcPtr, const XMLByte* srcEnd){
1.62 misha 486: if(!srcPtr || !*srcPtr || srcPtr>=srcEnd)
487: return 0;
488:
1.63 misha 489: unsigned int trailingBytes=gUTFBytes[*srcPtr]+1;
490: srcPtr+=trailingBytes;
1.62 misha 491:
492: return trailingBytes;
1.61 misha 493: }
494:
1.70 misha 495: // read non-UTF8 char, and return number of bytes needed for store this char in UTF8
1.61 misha 496: static unsigned int readChar(const XMLByte*& srcPtr, const XMLByte* srcEnd, XMLByte& firstByte, XMLCh& UTF8Char, const Charset::Tables& tables){
1.60 misha 497: if(!srcPtr || !*srcPtr || srcPtr>=srcEnd)
498: return 0;
499:
500: firstByte=*srcPtr++;
501: UTF8Char=tables.fromTable[firstByte];
502:
503: if(UTF8Char<0x80)
504: return 1;
505: else if(UTF8Char<0x800)
506: return 2;
507: else if(UTF8Char<0x10000)
508: return 3;
509: else if(UTF8Char<0x200000)
510: return 4;
511: else if(UTF8Char<0x4000000)
512: return 5;
513: else if(UTF8Char<= 0x7FFFFFFF)
514: return 6;
515:
516: // will use the replacement character '?'
517: firstByte=0;
518: return 1;
519: }
520:
1.70 misha 521: static int escapeUTF8(const XMLByte* srcData, size_t& srcLen,
1.80 misha 522: XMLByte* toFill, size_t& toFillLen) {
1.60 misha 523: const XMLByte* srcPtr=srcData;
524: XMLByte* outPtr=toFill;
525: XMLByte* outEnd=toFill+toFillLen;
526: // loop until we either run out of input data, or room to store
1.81 misha 527: for(UTF8_string_iterator i((XMLByte *)srcPtr, srcLen); (outPtr < outEnd) && i.has_next(); ){
528: if(i.getCharSize()==1){
529: if(isEscaped(i.getFirstByte())) // %XX
530: outPtr+=sprintf((char*)outPtr, "%%%02X", i.getFirstByte());
1.60 misha 531: else
1.81 misha 532: *outPtr++=i.getFirstByte();
1.60 misha 533: } else
1.81 misha 534: outPtr+=sprintf((char*)outPtr, "%%u%04X", i.next()); // %uXXXX
1.60 misha 535: }
536: // Update the bytes eaten
537: srcLen=srcPtr-srcData;
538:
539: // Return the characters read
540: toFillLen=outPtr-toFill;
541:
542: return 0;
543: }
544:
545: static int escape(const XMLByte* srcData, size_t& srcLen,
546: XMLByte *toFill, size_t& toFillLen,
547: const Charset::Tables& tables) {
548: const XMLByte* srcPtr=srcData;
549: const XMLByte* srcEnd=srcData+srcLen;
550: XMLByte* outPtr=toFill;
551: XMLByte firstByte;
552: XMLCh UTF8Char;
553: uint charSize;
554:
555: while(charSize=readChar(srcPtr, srcEnd, firstByte, UTF8Char, tables)){
556: if(charSize==1){
557: if(firstByte){
1.70 misha 558: if(isEscaped(firstByte)) // %XX
1.60 misha 559: outPtr+=sprintf((char*)outPtr, "%%%02X", firstByte);
560: else
561: *outPtr++=firstByte;
562: } else // add replacement char '?'
563: *outPtr++='?';
564: } else
565: outPtr+=sprintf((char*)outPtr, "%%u%04X", UTF8Char); // %uXXXX
566: }
567:
568: // Update the bytes eaten
569: srcLen = srcPtr - srcData;
570:
571: // Return the characters read
572: toFillLen = outPtr - toFill;
573:
574: return 0;
575: }
576:
577:
578: String::C Charset::escape(const String::C src, const Charset& source_charset){
579: size_t src_length=src.length;
580: if(!src_length)
581: return String::C("", 0);
582:
583: #ifdef PRECALCULATE_DEST_LENGTH
584: size_t dest_length=0;
585:
586: if(source_charset.isUTF8()){
1.81 misha 587: for(UTF8_string_iterator i((XMLByte *)src.str, src_length); i.has_next(); ){
588: if(i.getCharSize()==1)
589: dest_length+=!isEscaped(i.getFirstByte())?1/*as-is*/:3/*%XX*/;
1.60 misha 590: else
1.81 misha 591: dest_length+=6/*%uXXXX*/;
1.60 misha 592: }
593: } else {
1.81 misha 594: const XMLByte* srcPtr=(XMLByte*)src.str;
595: const XMLByte* srcEnd=srcPtr+src_length;
596: XMLByte firstByte;
597: XMLCh UTF8Char;
1.60 misha 598: while(uint charSize=readChar(srcPtr, srcEnd, firstByte, UTF8Char, source_charset.tables)){
599: if(charSize==1)
1.70 misha 600: dest_length+=(!firstByte/*replacement char '?'*/ || !isEscaped(firstByte))?1:3/*'%XX'*/;
1.60 misha 601: else
602: dest_length+=6; // '%uXXXX'
603: }
604: }
605: #else
606: size_t dest_length=src_length*6; // enough for %uXXXX but too memory-hungry
607: #endif
608:
609: //throw Exception(0,0,"%u",dest_length);
610:
611: #ifndef NDEBUG
612: size_t saved_dest_length=dest_length;
613: #endif
614: XMLByte *dest_body=new(PointerFreeGC) XMLByte[dest_length+1/*for terminator*/];
615:
616: int status;
617: if(source_charset.isUTF8()){
1.70 misha 618: status=::escapeUTF8((XMLByte *)src.str, src_length, dest_body, dest_length);
1.60 misha 619: } else {
620: status=::escape((XMLByte *)src.str, src_length, dest_body, dest_length, source_charset.tables);
621: }
622:
623: if(status<0)
624: throw Exception(0,
625: 0,
1.70 misha 626: "Charset::escape buffer overflow");
1.60 misha 627:
628: assert(dest_length<=saved_dest_length);
629: dest_body[dest_length]=0; // terminator
630: return String::C((char*)dest_body, dest_length);
631: }
632:
1.64 misha 633: String::Body Charset::escape(const String::Body src, const Charset& source_charset) {
634: const char *src_ptr=src.cstr();
1.83 ! misha 635: size_t src_size=src.length();
1.64 misha 636:
1.70 misha 637: String::C dest=Charset::escape(String::C(src_ptr, src_size), source_charset);
1.64 misha 638:
1.77 misha 639: return String::Body(dest.length ? dest.str:0);
1.64 misha 640: }
641:
642: String& Charset::escape(const String& src, const Charset& source_charset) {
1.72 misha 643: if(src.is_empty())
1.73 misha 644: return *new String();
1.64 misha 645:
646: return *new String(escape((String::Body)src, source_charset), String::L_CLEAN);
647: }
1.60 misha 648:
1.35 paf 649: const String::C Charset::transcodeToUTF8(const String::C src) const {
1.71 misha 650: int src_length=src.length;
1.60 misha 651:
652: #ifdef PRECALCULATE_DEST_LENGTH
1.71 misha 653: int dest_length=0;
1.60 misha 654: const XMLByte* srcPtr=(XMLByte*)src.str;
655: const XMLByte* srcEnd=srcPtr+src_length;
1.69 misha 656: XMLByte firstByte;
657: XMLCh UTF8Char;
658: while(uint charSize=readChar(srcPtr, srcEnd, firstByte, UTF8Char, tables))
1.60 misha 659: dest_length+=charSize;
660: #else
1.71 misha 661: int dest_length=src_length*6; // so that surly enough (max utf8 seq len=6) but too memory-hyngry
1.60 misha 662: #endif
663:
664: //throw Exception(0,0,"%u",dest_length);
665:
1.35 paf 666: #ifndef NDEBUG
1.71 misha 667: int saved_dest_length=dest_length;
1.35 paf 668: #endif
669: XMLByte *dest_body=new(PointerFreeGC) XMLByte[dest_length+1/*for terminator*/];
1.11 paf 670:
671: if(::transcodeToUTF8(
1.35 paf 672: (XMLByte *)src.str, src_length,
673: dest_body, dest_length,
1.11 paf 674: tables)<0)
1.43 paf 675: throw Exception(0,
1.10 paf 676: 0,
1.11 paf 677: "Charset::transcodeToUTF8 buffer overflow");
1.10 paf 678:
1.60 misha 679: assert(dest_length<=saved_dest_length);
680: dest_body[dest_length]=0; // terminator
1.35 paf 681: return String::C((char*)dest_body, dest_length);
1.10 paf 682: }
1.38 paf 683:
684: static XMLCh change_case_UTF8(const XMLCh src, const Charset::UTF8CaseTable& table) {
1.80 misha 685: int lo = 0;
686: int hi = table.size - 1;
1.39 paf 687: while(lo<=hi) {
1.38 paf 688: // Calc the mid point of the low and high offset.
1.39 paf 689: const unsigned int i = (lo + hi) / 2;
690:
691: XMLCh cur=table.records[i].from;
692: if(src==cur)
693: return table.records[i].to;
694: if(src>cur)
695: lo = i+1;
1.38 paf 696: else
1.39 paf 697: hi = i-1;
698: }
699:
700: // not found
1.38 paf 701: return src;
702: }
703:
1.58 misha 704: static void store_UTF8(XMLCh src, XMLByte*& outPtr){
1.38 paf 705: if(!src) {
706: // use the replacement character
707: *outPtr++= '?';
708: return;
709: }
710:
711: // Figure out how many bytes we need
712: unsigned int encodedBytes;
713: if(src<0x80)
714: encodedBytes = 1;
715: else if(src<0x800)
716: encodedBytes = 2;
717: else if(src<0x10000)
718: encodedBytes = 3;
719: else if(src<0x200000)
720: encodedBytes = 4;
721: else if(src<0x4000000)
722: encodedBytes = 5;
723: else if(src<= 0x7FFFFFFF)
724: encodedBytes = 6;
725: else {
726: // use the replacement character
727: *outPtr++= '?';
728: return;
729: }
730:
731: // And spit out the bytes. We spit them out in reverse order
732: // here, so bump up the output pointer and work down as we go.
733: outPtr+= encodedBytes;
734: switch(encodedBytes) {
735: case 6: *--outPtr = XMLByte((src | 0x80UL) & 0xBFUL);
736: src>>= 6;
737: case 5: *--outPtr = XMLByte((src | 0x80UL) & 0xBFUL);
738: src>>= 6;
739: case 4: *--outPtr = XMLByte((src | 0x80UL) & 0xBFUL);
740: src>>= 6;
741: case 3: *--outPtr = XMLByte((src | 0x80UL) & 0xBFUL);
742: src>>= 6;
743: case 2: *--outPtr = XMLByte((src | 0x80UL) & 0xBFUL);
744: src>>= 6;
745: case 1: *--outPtr = XMLByte(src | gFirstByteMark[encodedBytes]);
746: }
747:
748: // Add the encoded bytes back in again to indicate we've eaten them
749: outPtr+= encodedBytes;
750: }
751:
752: static void change_case_UTF8(XMLCh src, XMLByte*& outPtr,
753: const Charset::UTF8CaseTable& table) {
754: store_UTF8(change_case_UTF8(src, table), outPtr);
755: };
1.44 paf 756: void change_case_UTF8(const XMLByte* srcData, size_t srcLen,
757: XMLByte* toFill, size_t toFillLen,
758: const Charset::UTF8CaseTable& table) {
1.38 paf 759: const XMLByte* srcPtr=srcData;
1.44 paf 760: const XMLByte* srcEnd=srcData+srcLen;
1.38 paf 761: XMLByte* outPtr=toFill;
1.44 paf 762: XMLByte* outEnd=toFill+toFillLen;
763:
764: // We now loop until we either run out of input data, or room to store
765: while ((srcPtr < srcEnd) && (outPtr < outEnd)) {
766: // Get the next leading byte out
767: const XMLByte firstByte =* srcPtr;
1.38 paf 768:
1.60 misha 769: if(firstByte<=127) {
1.38 paf 770: change_case_UTF8(firstByte, outPtr, table);
771: srcPtr++;
772: continue;
773: }
774:
775: // See how many trailing src bytes this sequence is going to require
776: const unsigned int trailingBytes = gUTFBytes[firstByte];
777:
778: // Looks ok, so lets build up the value
779: uint tmpVal=0;
780: switch(trailingBytes) {
781: case 5: tmpVal+=*srcPtr++; tmpVal<<=6;
782: case 4: tmpVal+=*srcPtr++; tmpVal<<=6;
783: case 3: tmpVal+=*srcPtr++; tmpVal<<=6;
784: case 2: tmpVal+=*srcPtr++; tmpVal<<=6;
785: case 1: tmpVal+=*srcPtr++; tmpVal<<=6;
786: case 0: tmpVal+=*srcPtr++;
787: break;
788:
789: default:
790: throw Exception(0,
791: 0,
792: "change_case_UTF8 error: wrong trailingBytes value(%d)", trailingBytes);
793: }
794: tmpVal-=gUTFOffsets[trailingBytes];
795:
796: // If it will fit into a single char, then put it in. Otherwise
797: // fail [*encode it as a surrogate pair. If its not valid, use the
798: // replacement char.*]
799: if(!(tmpVal & 0xFFFF0000))
800: change_case_UTF8(tmpVal, outPtr, table);
801: else
802: throw Exception(0,
803: 0,
804: "change_case_UTF8 error: too big tmpVal(0x%08X)", tmpVal);
805: }
806:
807: if(srcPtr!=outPtr)
808: throw Exception(0,
809: 0,
810: "change_case_UTF8 error: end pointers do not match");
811: }
812:
1.60 misha 813: static size_t getDecNumLength(XMLCh UTF8Char){
814: return
815: (UTF8Char < 100)
816: ?2
817: :(UTF8Char < 1000)
818: ?3
819: :(UTF8Char < 10000)
820: ?4
821: :5;
822: }
1.38 paf 823:
1.35 paf 824: const String::C Charset::transcodeFromUTF8(const String::C src) const {
1.82 misha 825: int src_length=src.length;
1.60 misha 826: #ifdef PRECALCULATE_DEST_LENGTH
1.71 misha 827: int dest_length=0;
1.82 misha 828: for(UTF8_string_iterator i((XMLByte *)src.str, src_length); i.has_next(); ){
1.81 misha 829: if(i.getCharSize()==1)
1.60 misha 830: dest_length++;
831: else
1.81 misha 832: dest_length+=(i.next() & 0xFFFF0000)
833: ?i.getCharSize()*3 // '%XX' for each byte
834: :(xlatOneTo(i.next(), tables, 0)!=0)
1.60 misha 835: ?1 // can convert it to single char
1.81 misha 836: :getDecNumLength(i.next())+3; // &#XX; - &#XXXXX;
1.60 misha 837: }
838: #else
839: // so that surly enough, "&#XXX;" has max ratio (huh? 8 bytes needed for '&#XXXXX;')
1.82 misha 840: int dest_length=src_length*6;
1.60 misha 841: #endif
842:
1.35 paf 843: #ifndef NDEBUG
1.71 misha 844: int saved_dest_length=dest_length;
1.35 paf 845: #endif
846: XMLByte *dest_body=new(PointerFreeGC) XMLByte[dest_length+1/*for terminator*/];
1.11 paf 847:
848: if(::transcodeFromUTF8(
1.82 misha 849: (XMLByte *)src.str, src_length,
1.35 paf 850: dest_body, dest_length,
1.11 paf 851: tables)<0)
1.43 paf 852: throw Exception(0,
1.10 paf 853: 0,
1.35 paf 854: "Charset::transcodeFromUTF8 buffer overflow");
1.10 paf 855:
1.60 misha 856: assert(dest_length<=saved_dest_length);
857: dest_body[dest_length]=0; // terminator
1.35 paf 858: return String::C((char*)dest_body, dest_length);
1.1 paf 859: }
860:
861: /// transcode using both charsets
1.35 paf 862: const String::C Charset::transcodeToCharset(const String::C src,
1.80 misha 863: const Charset& dest_charset) const {
1.35 paf 864: if(&dest_charset==this)
865: return src;
866: else {
867: size_t dest_length=src.length;
868: XMLByte* dest_body=new(PointerFreeGC) XMLByte[dest_length+1/*for terminator*/];
869:
870: XMLByte* output=dest_body;
871: const XMLByte* input=(XMLByte *)src.str;
872: while(XMLCh c=*input++) {
873: XMLCh curVal = tables.fromTable[c];
874: *output++=curVal?
875: xlatOneTo(curVal, dest_charset.tables, '?') // OK
876: :'?'; // use the replacement character
1.6 paf 877: }
1.1 paf 878:
1.35 paf 879: dest_body[dest_length]=0; // terminator
880: return String::C((char*)dest_body, dest_length);
1.6 paf 881: }
1.1 paf 882: }
883:
1.58 misha 884: void Charset::store_Char(XMLByte*& outPtr, XMLCh src, XMLByte not_found){
1.59 misha 885: if(isUTF8())
1.58 misha 886: store_UTF8(src, outPtr);
1.59 misha 887: else if(char ch=xlatOneTo(src, tables, not_found))
1.58 misha 888: *outPtr++=ch;
1.57 misha 889: }
890:
1.1 paf 891: #ifdef XML
1.10 paf 892:
1.35 paf 893: static const Charset::Tables* tables[MAX_CHARSETS];
894:
1.46 paf 895: #ifdef PA_PATCHED_LIBXML_BACKWARD
896:
897: #define declareXml256ioFuncs(i) \
898: static int xml256CharEncodingInputFunc##i( \
899: unsigned char *out, int *outlen, \
900: const unsigned char *in, int *inlen, void*) { \
901: return transcodeToUTF8( \
1.71 misha 902: in, *inlen, \
903: out, *outlen, \
1.46 paf 904: *tables[i]); \
905: } \
906: static int xml256CharEncodingOutputFunc##i( \
907: unsigned char *out, int *outlen, \
908: const unsigned char *in, int *inlen, void*) { \
909: return transcodeFromUTF8( \
1.71 misha 910: in, *inlen, \
911: out, *outlen, \
1.46 paf 912: *tables[i]); \
913: }
914:
915: #else
916:
1.35 paf 917: #define declareXml256ioFuncs(i) \
918: static int xml256CharEncodingInputFunc##i( \
919: unsigned char *out, int *outlen, \
920: const unsigned char *in, int *inlen) { \
921: return transcodeToUTF8( \
1.71 misha 922: in, *inlen, \
923: out, *outlen, \
1.35 paf 924: *tables[i]); \
925: } \
926: static int xml256CharEncodingOutputFunc##i( \
927: unsigned char *out, int *outlen, \
928: const unsigned char *in, int *inlen) { \
929: return transcodeFromUTF8( \
1.71 misha 930: in, *inlen, \
931: out, *outlen, \
1.35 paf 932: *tables[i]); \
933: }
934:
1.46 paf 935: #endif
936:
937:
1.35 paf 938: declareXml256ioFuncs(0) declareXml256ioFuncs(1)
939: declareXml256ioFuncs(2) declareXml256ioFuncs(3)
940: declareXml256ioFuncs(4) declareXml256ioFuncs(5)
941: declareXml256ioFuncs(6) declareXml256ioFuncs(7)
942: declareXml256ioFuncs(8) declareXml256ioFuncs(9)
943:
944: static xmlCharEncodingInputFunc inputFuncs[MAX_CHARSETS]={
945: xml256CharEncodingInputFunc0, xml256CharEncodingInputFunc1,
946: xml256CharEncodingInputFunc2, xml256CharEncodingInputFunc3,
947: xml256CharEncodingInputFunc4, xml256CharEncodingInputFunc5,
948: xml256CharEncodingInputFunc6, xml256CharEncodingInputFunc7,
949: xml256CharEncodingInputFunc8, xml256CharEncodingInputFunc9
950: };
951: static xmlCharEncodingOutputFunc outputFuncs[MAX_CHARSETS]={
952: xml256CharEncodingOutputFunc0, xml256CharEncodingOutputFunc1,
953: xml256CharEncodingOutputFunc2, xml256CharEncodingOutputFunc3,
954: xml256CharEncodingOutputFunc4, xml256CharEncodingOutputFunc5,
955: xml256CharEncodingOutputFunc6, xml256CharEncodingOutputFunc7,
956: xml256CharEncodingOutputFunc8, xml256CharEncodingOutputFunc9
957: };
958: static size_t handlers_count=0;
1.10 paf 959:
960: void Charset::addEncoding(char *name_cstr) {
1.35 paf 961: if(handlers_count==MAX_CHARSETS)
962: throw Exception(0,
963: 0,
964: "already allocated %d handlers, no space for new encoding '%s'",
965: MAX_CHARSETS, name_cstr);
966:
1.45 paf 967: xmlCharEncodingHandler* handler=new(UseGC) xmlCharEncodingHandler;
1.35 paf 968: {
969: handler->name=name_cstr;
970: handler->input=inputFuncs[handlers_count];
971: handler->output=outputFuncs[handlers_count];
972: ::tables[handlers_count]=&tables;
973: handlers_count++;
974: }
1.10 paf 975:
976: xmlRegisterCharEncodingHandler(handler);
1.35 paf 977:
1.10 paf 978: }
979:
1.37 paf 980: void Charset::initTranscoder(const String::Body NAME, const char* name_cstr) {
1.15 paf 981: ftranscoder=xmlFindCharEncodingHandler(name_cstr);
1.35 paf 982: transcoder(NAME); // check right way
1.15 paf 983: }
984:
1.37 paf 985: xmlCharEncodingHandler& Charset::transcoder(const String::Body NAME) {
1.15 paf 986: if(!ftranscoder)
1.56 misha 987: throw Exception(PARSER_RUNTIME,
1.35 paf 988: new String(NAME, String::L_TAINTED),
1.10 paf 989: "unsupported encoding");
1.35 paf 990: return *ftranscoder;
1.10 paf 991: }
992:
1.54 paf 993: String::C Charset::transcode_cstr(const xmlChar* s) {
1.13 paf 994: if(!s)
1.35 paf 995: return String::C("", 0);
1.8 paf 996:
1.35 paf 997: int inlen=strlen((const char*)s);
1.51 paf 998: int outlen=inlen*6/*strlen("ÿ")*/; // max
1.35 paf 999: #ifndef NDEBUG
1000: int saved_outlen=outlen;
1001: #endif
1002: char *out=new(PointerFreeGC) char[outlen+1];
1.8 paf 1003:
1.30 paf 1004: int error;
1.35 paf 1005: if(xmlCharEncodingOutputFunc output=transcoder(FNAME).output) {
1.30 paf 1006: error=output(
1.17 paf 1007: (unsigned char*)out, &outlen,
1.46 paf 1008: (const unsigned char*)s, &inlen
1009: #ifdef PA_PATCHED_LIBXML_BACKWARD
1010: ,0
1011: #endif
1012: );
1.30 paf 1013: } else {
1014: memcpy(out, s, outlen=inlen);
1015: error=0;
1016: }
1017: if(error<0)
1.23 paf 1018: throw Exception(0,
1.8 paf 1019: 0,
1.30 paf 1020: "transcode_cstr failed (%d)", error);
1.8 paf 1021:
1.35 paf 1022: assert(outlen<=saved_outlen); out[outlen]=0;
1023: return String::C(out, outlen);
1.14 paf 1024: }
1.54 paf 1025: const String& Charset::transcode(const xmlChar* s) {
1.35 paf 1026: String::C cstr=transcode_cstr(s);
1.75 misha 1027: return *new String(cstr.str, String::L_TAINTED);
1.1 paf 1028: }
1029:
1.8 paf 1030: /// @test less memory using -maybe- xmlParserInputBufferCreateMem
1.35 paf 1031: xmlChar* Charset::transcode_buf2xchar(const char* buf, size_t buf_size) {
1032: xmlChar* out;
1.30 paf 1033: int outlen;
1034: int error;
1.35 paf 1035: #ifndef NDEBUG
1036: int saved_outlen;
1037: #endif
1038: if(xmlCharEncodingInputFunc input=transcoder(FNAME).input) {
1.51 paf 1039: outlen=buf_size*6/*max UTF8 bytes per char*/;
1.35 paf 1040: #ifndef NDEBUG
1041: saved_outlen=outlen;
1042: #endif
1.47 paf 1043: out=(xmlChar*)xmlMalloc(outlen+1);
1.30 paf 1044: error=input(
1.17 paf 1045: out, &outlen,
1.46 paf 1046: (const unsigned char*)buf, (int*)&buf_size
1047: #ifdef PA_PATCHED_LIBXML_BACKWARD
1048: ,0
1049: #endif
1050: );
1.30 paf 1051: } else {
1052: outlen=buf_size;
1.35 paf 1053: #ifndef NDEBUG
1054: saved_outlen=outlen;
1055: #endif
1056: out=(xmlChar*)xmlMalloc(outlen+1);
1.30 paf 1057: memcpy(out, buf, outlen);
1058: error=0;
1059: }
1.17 paf 1060:
1.30 paf 1061: if(error<0)
1.23 paf 1062: throw Exception(0,
1.8 paf 1063: 0,
1.30 paf 1064: "transcode_buf failed (%d)", error);
1.8 paf 1065:
1.35 paf 1066: assert(outlen<=saved_outlen); out[outlen]=0;
1067: return out;
1.24 paf 1068: }
1.1 paf 1069:
1.79 misha 1070: xmlChar* Charset::transcode(const String& s) {
1071: String::Body sbody=s.cstr_to_string_body_untaint(String::L_AS_IS);
1072: return transcode_buf2xchar(sbody.cstr(), sbody.length());
1.1 paf 1073: }
1.35 paf 1074:
1.79 misha 1075: xmlChar* Charset::transcode(const String::Body s) {
1076: return transcode_buf2xchar(s.cstr(), s.length());
1.35 paf 1077: }
1.36 paf 1078: #endif
1.34 paf 1079:
1.37 paf 1080: String::Body Charset::transcode(const String::Body src,
1.34 paf 1081: const Charset& source_transcoder,
1.35 paf 1082: const Charset& dest_transcoder) {
1.34 paf 1083:
1.35 paf 1084: const char *src_ptr=src.cstr();
1.83 ! misha 1085: size_t src_size=src.length();
1.34 paf 1086:
1.77 misha 1087: String::C dest=Charset::transcode(String::C(src_ptr, src_size), source_transcoder, dest_transcoder);
1.34 paf 1088:
1.77 misha 1089: return String::Body(dest.length ? dest.str:0);
1.35 paf 1090: }
1091:
1092: String& Charset::transcode(const String& src,
1093: const Charset& source_transcoder,
1094: const Charset& dest_transcoder) {
1.72 misha 1095: if(src.is_empty())
1.73 misha 1096: return *new String();
1.34 paf 1097:
1.37 paf 1098: return *new String(transcode((String::Body)src, source_transcoder, dest_transcoder), String::L_CLEAN);
1.34 paf 1099: }
1100:
1.35 paf 1101: void Charset::transcode(ArrayString& src,
1.34 paf 1102: const Charset& source_transcoder,
1.35 paf 1103: const Charset& dest_transcoder) {
1104: for(size_t i=0; i<src.count(); i++)
1105: src.put(i, &transcode(*src[i], source_transcoder, dest_transcoder));
1.34 paf 1106: }
1107:
1108: #ifndef DOXYGEN
1109: struct Transcode_pair_info {
1110: const Charset* source_transcoder;
1111: const Charset* dest_transcoder;
1112: };
1113: #endif
1.76 misha 1114: static void transcode_pair(HashStringValue::key_type /*akey*/,
1.37 paf 1115: String::Body& avalue,
1.35 paf 1116: Transcode_pair_info* info) {
1117: avalue=Charset::transcode(avalue,
1118: *info->source_transcoder,
1119: *info->dest_transcoder);
1.34 paf 1120: }
1.61 misha 1121:
1.35 paf 1122: void Charset::transcode(HashStringString& src,
1.34 paf 1123: const Charset& source_transcoder,
1.35 paf 1124: const Charset& dest_transcoder) {
1125: Transcode_pair_info info={&source_transcoder, &dest_transcoder};
1.55 paf 1126: src.for_each_ref<Transcode_pair_info*>(transcode_pair, &info);
1.34 paf 1127: }
1.61 misha 1128:
1129: size_t getUTF8BytePos(const XMLByte* srcBegin, const XMLByte* srcEnd, size_t charPos){
1130: const XMLByte* ptr=srcBegin;
1.70 misha 1131: while(charPos-- && skipUTF8Char(ptr, srcEnd));
1.61 misha 1132:
1133: return ptr-srcBegin;
1134: }
1135:
1136: size_t getUTF8CharPos(const XMLByte* srcBegin, const XMLByte* srcEnd, size_t bytePos){
1137: size_t charPos=0;
1138: const XMLByte* ptr=srcBegin;
1139: const XMLByte* ptrEnd=srcBegin+bytePos;
1.70 misha 1140: while(skipUTF8Char(ptr, srcEnd)){
1.61 misha 1141: if(ptr>ptrEnd)
1142: return charPos;
1143: charPos++;
1144: }
1145:
1146: // scan till end but position in bytes still too low
1147: throw Exception(0,
1148: 0,
1149: "Error convertion byte pos to char pos");
1150: }
1151:
1152: size_t lengthUTF8(const XMLByte* srcBegin, const XMLByte* srcEnd){
1153: size_t size=0;
1.70 misha 1154: while(skipUTF8Char(srcBegin, srcEnd))
1.61 misha 1155: size++;
1156:
1157: return size;
1158: }
1.80 misha 1159:
1160: bool UTF8_string_iterator::has_next(){
1161: fcharSize=readUTF8Char(fsrcPtr, fsrcEnd, ffirstByte, fUTF8Char);
1162: return fcharSize!=0;
1163: }
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