|
|
| version 1.10, 2001/12/28 18:12:30 | version 1.85, 2010/08/27 02:53:24 |
|---|---|
| Line 1 | Line 1 |
| /** @file | /** @file |
| Parser: Charset connection implementation. | Parser: Charset connection implementation. |
| Copyright(c) 2001 ArtLebedev Group(http://www.artlebedev.com) | Copyright(c) 2001-2009 ArtLebedev Group (http://www.artlebedev.com) |
| Author: Alexander Petrosyan<paf@design.ru>(http://paf.design.ru) | Author: Alexander Petrosyan<paf@design.ru>(http://paf.design.ru) |
| $Id$ | |
| */ | */ |
| static const char * const IDENT_CHARSET_C="$Date$"; | |
| #include "pa_charset.h" | #include "pa_charset.h" |
| #include "pa_charsets.h" | |
| #ifdef XML | #ifdef XML |
| #include "libxml/encoding.h" | #include "libxml/encoding.h" |
| #endif | #endif |
| // globals | //#define PA_PATCHED_LIBXML_BACKWARD |
| // reduce memory usage by pre-calculation utf-8 string length | |
| #define PRECALCULATE_DEST_LENGTH | |
| // globals | |
| // consts | Charset::UTF8CaseTable::Rec UTF8CaseToUpperRecords[]={ |
| #include "utf8-to-upper.inc" | |
| }; | |
| Charset::UTF8CaseTable UTF8CaseToUpper={ | |
| sizeof(UTF8CaseToUpperRecords)/sizeof(Charset::UTF8CaseTable::Rec), | |
| UTF8CaseToUpperRecords}; | |
| #define MAX_CHARSET_UNI_CODES 500 | Charset::UTF8CaseTable::Rec UTF8CaseToLowerRecords[]={ |
| #include "utf8-to-lower.inc" | |
| }; | |
| Charset::UTF8CaseTable UTF8CaseToLower={ | |
| sizeof(UTF8CaseToLowerRecords)/sizeof(Charset::UTF8CaseTable::Rec), | |
| UTF8CaseToLowerRecords}; | |
| // helpers | // helpers |
| Line 26 inline void prepare_case_tables(unsigned | Line 41 inline void prepare_case_tables(unsigned |
| unsigned char *lcc_table=tables+lcc_offset; | unsigned char *lcc_table=tables+lcc_offset; |
| unsigned char *fcc_table=tables+fcc_offset; | unsigned char *fcc_table=tables+fcc_offset; |
| for(int i=0; i<0x100; i++) | for(int i=0; i<0x100; i++) |
| lcc_table[i]=fcc_table[i]=i; | lcc_table[i]=fcc_table[i]=(unsigned char)i; |
| } | } |
| inline void cstr2ctypes(unsigned char *tables, const unsigned char *cstr, | inline void cstr2ctypes(unsigned char *tables, const unsigned char *cstr, |
| unsigned char bit) { | unsigned char bit) { |
| Line 37 inline void cstr2ctypes(unsigned char *t | Line 52 inline void cstr2ctypes(unsigned char *t |
| ctypes_table[c]|=bit; | ctypes_table[c]|=bit; |
| } | } |
| } | } |
| inline unsigned int to_wchar_code(const char *cstr) { | inline unsigned int to_wchar_code(const char* cstr) { |
| if(!cstr || !*cstr) | if(!cstr || !*cstr) |
| return 0; | return 0; |
| if(cstr[1]==0) | if(cstr[1]==0) |
| Line 46 inline unsigned int to_wchar_code(const | Line 61 inline unsigned int to_wchar_code(const |
| char *error_pos; | char *error_pos; |
| return(unsigned int)strtol(cstr, &error_pos, 0); | return(unsigned int)strtol(cstr, &error_pos, 0); |
| } | } |
| inline bool to_bool(const char *cstr) { | inline bool to_bool(const char* cstr) { |
| return cstr && *cstr!=0; | return cstr && *cstr!=0; |
| } | } |
| static void element2ctypes(unsigned char c, bool belongs, | static void element2ctypes(unsigned char c, bool belongs, |
| Line 73 static void element2case(unsigned char f | Line 88 static void element2case(unsigned char f |
| // methods | // methods |
| extern "C" unsigned char pcre_default_tables[]; // pcre/chartables.c | Charset::Charset(Request_charsets* charsets, const String::Body ANAME, const String* afile_spec): |
| Charset::Charset(Pool& apool, const String& aname, const String *request_file_spec): Pooled(apool), | FNAME(ANAME), |
| fname(aname) { | FNAME_CSTR(ANAME.cstrm()) { |
| char *name_cstr=fname.cstr(); | |
| for(char *c=name_cstr; *c; c++) | |
| *c = toupper(*c); | |
| if(request_file_spec) { | if(afile_spec) { |
| fisUTF8=false; | fisUTF8=false; |
| loadDefinition(*request_file_spec); | load_definition(*charsets, *afile_spec); |
| #ifdef XML | #ifdef XML |
| addEncoding(name_cstr); | addEncoding(FNAME_CSTR); |
| #endif | #endif |
| } else { | } else { |
| fisUTF8=true; | fisUTF8=true; |
| // grab default onces [for UTF-8 so to be able to make a-z =>A-Z | // grab default onces [for UTF-8 so to be able to make a-z =>A-Z |
| memcpy(pcre_tables, pcre_default_tables, sizeof(pcre_tables)); | memcpy(pcre_tables, _pcre_default_tables, sizeof(pcre_tables)); |
| } | } |
| #ifdef XML | #ifdef XML |
| initTranscoder(&aname, name_cstr); | initTranscoder(FNAME, FNAME_CSTR); |
| #endif | |
| } | |
| Charset::~Charset() { | |
| #ifdef XML | |
| // not deleting transcoder, that's not our business | |
| #endif | #endif |
| } | } |
| void Charset::loadDefinition(const String& request_file_spec) { | void Charset::load_definition(Request_charsets& charsets, const String& afile_spec) { |
| // pcre_tables | // pcre_tables |
| // lowcase, flipcase, bits digit+word+whitespace, masks | // lowcase, flipcase, bits digit+word+whitespace, masks |
| Line 115 void Charset::loadDefinition(const Strin | Line 120 void Charset::loadDefinition(const Strin |
| cstr2ctypes(pcre_tables,(const unsigned char *)"*+?{^.$|()[", ctype_meta); | cstr2ctypes(pcre_tables,(const unsigned char *)"*+?{^.$|()[", ctype_meta); |
| // charset | // charset |
| memset(tables.fromTable, 0, sizeof(tables.fromTable)); | memset(&tables, 0, sizeof(tables)); |
| tables.toTable=(Charset_TransRec *)calloc(sizeof(Charset_TransRec)*MAX_CHARSET_UNI_CODES); | |
| tables.toTableSize=0; | |
| // strangly vital | |
| tables.toTable[tables.toTableSize].intCh=0; | |
| tables.toTable[tables.toTableSize].extCh=(XMLByte)0; | |
| tables.toTableSize++; | |
| // loading text | // loading text |
| char *data=file_read_text(pool(), request_file_spec); | char *data=file_read_text(charsets, afile_spec); |
| // ignore header | // ignore header |
| getrow(&data); | getrow(&data); |
| // parse cells | // parse cells |
| char *row; | char *row; |
| while(row=getrow(&data)) { | while((row=getrow(&data))) { |
| // remove empty&comment lines | // remove empty&comment lines |
| if(!*row || *row=='#') | if(!*row || *row=='#') |
| continue; | continue; |
| // char white-space digit hex-digit letter word lowercase unicode1 unicode2 | // char white-space digit hex-digit letter word lowercase unicode1 unicode2 |
| unsigned int c=0; | unsigned char c=0; |
| char *cell; | char *cell; |
| for(int column=0; cell=lsplit(&row, '\t'); column++) { | for(int column=0; (cell=lsplit(&row, '\t')); column++) { |
| switch(column) { | switch(column) { |
| case 0: c=to_wchar_code(cell); break; | case 0: c=(unsigned char)to_wchar_code(cell); break; |
| // pcre_tables | // pcre_tables |
| case 1: element2ctypes(c, to_bool(cell), pcre_tables, ctype_space, cbit_space); break; | case 1: element2ctypes(c, to_bool(cell), pcre_tables, ctype_space, cbit_space); break; |
| case 2: element2ctypes(c, to_bool(cell), pcre_tables, ctype_digit, cbit_digit); break; | case 2: element2ctypes(c, to_bool(cell), pcre_tables, ctype_digit, cbit_digit); break; |
| case 3: element2ctypes(c, to_bool(cell), pcre_tables, ctype_xdigit); break; | case 3: element2ctypes(c, to_bool(cell), pcre_tables, ctype_xdigit); break; |
| case 4: element2ctypes(c, to_bool(cell), pcre_tables, ctype_letter); break; | case 4: element2ctypes(c, to_bool(cell), pcre_tables, ctype_letter); break; |
| case 5: element2ctypes(c, to_bool(cell), pcre_tables, ctype_word, cbit_word); break; | case 5: element2ctypes(c, to_bool(cell), pcre_tables, ctype_word, cbit_word); break; |
| case 6: element2case(c, to_wchar_code(cell), pcre_tables); break; | case 6: element2case(c, (unsigned char)to_wchar_code(cell), pcre_tables); break; |
| case 7: | case 7: |
| case 8: | case 8: |
| // charset | // charset |
| if(tables.toTableSize>MAX_CHARSET_UNI_CODES) | if(tables.toTableSize>MAX_CHARSET_UNI_CODES) |
| throw Exception(0, 0, | throw Exception(PARSER_RUNTIME, |
| &request_file_spec, | &afile_spec, |
| "charset must contain not more then %d unicode values", MAX_CHARSET_UNI_CODES); | "charset must contain not more then %d unicode values", MAX_CHARSET_UNI_CODES); |
| XMLCh unicode=(XMLCh)to_wchar_code(cell); | XMLCh unicode=(XMLCh)to_wchar_code(cell); |
| Line 178 void Charset::loadDefinition(const Strin | Line 177 void Charset::loadDefinition(const Strin |
| static int sort_cmp_Trans_rec_intCh(const void *a, const void *b) { | static int sort_cmp_Trans_rec_intCh(const void *a, const void *b) { |
| return | return |
| static_cast<const Charset_TransRec *>(a)->intCh- | static_cast<const Charset::Tables::Rec *>(a)->intCh- |
| static_cast<const Charset_TransRec *>(b)->intCh; | static_cast<const Charset::Tables::Rec *>(b)->intCh; |
| } | } |
| void Charset::sort_ToTable() { | void Charset::sort_ToTable() { |
| Line 190 void Charset::sort_ToTable() { | Line 189 void Charset::sort_ToTable() { |
| //fclose(f); | //fclose(f); |
| } | } |
| // @todo: precache for spedup searching | |
| static XMLByte xlatOneTo(const XMLCh toXlat, | static XMLByte xlatOneTo(const XMLCh toXlat, |
| const Charset::Tables& tables) { | const Charset::Tables& tables, |
| unsigned int lowOfs = 0; | XMLByte not_found) { |
| unsigned int hiOfs = tables.toTableSize - 1; | int lo = 0; |
| XMLByte curByte = 0; | int hi = tables.toTableSize - 1; |
| do { | while(lo<=hi) { |
| // Calc the mid point of the low and high offset. | // Calc the mid point of the low and high offset. |
| const unsigned int midOfs =((hiOfs - lowOfs) / 2)+lowOfs; | const unsigned int i = (lo + hi) / 2; |
| // If our test char is greater than the mid point char, then | XMLCh cur=tables.toTable[i].intCh; |
| // we move up to the upper half. Else we move to the lower | if(toXlat==cur) |
| // half. If its equal, then its our guy. | return tables.toTable[i].extCh; |
| if(toXlat>tables.toTable[midOfs].intCh) | if(toXlat>cur) |
| lowOfs = midOfs; | lo = i+1; |
| else if(toXlat<tables.toTable[midOfs].intCh) | |
| hiOfs = midOfs; | |
| else | else |
| return tables.toTable[midOfs].extCh; | hi = i-1; |
| } while(lowOfs+1<hiOfs); | } |
| return '?'; | return not_found; |
| } | } |
| void Charset::transcode(Pool& pool, | String::C Charset::transcode(const String::C src, |
| const Charset& source_charset, const void *source_body, size_t source_content_length, | const Charset& source_charset, |
| const Charset& dest_charset, const void *& dest_body, size_t& dest_content_length | const Charset& dest_charset) { |
| ) { | if(!src.length) |
| if(!source_content_length) { | return String::C("", 0); |
| dest_body=0; | |
| dest_content_length=0; | |
| return; | |
| } | |
| switch((source_charset.isUTF8()?0x10:0x00)|(dest_charset.isUTF8()?0x01:0x00)) { | switch((source_charset.isUTF8()?0x10:0x00)|(dest_charset.isUTF8()?0x01:0x00)) { |
| default: // 0x00 | default: // 0x00 |
| source_charset.transcodeToCharset(pool, dest_charset, | return source_charset.transcodeToCharset(src, dest_charset); |
| source_body, source_content_length, | |
| dest_body, dest_content_length); | |
| break; | |
| case 0x01: | case 0x01: |
| source_charset.transcodeToUTF8(pool, | return source_charset.transcodeToUTF8(src); |
| source_body, source_content_length, | |
| dest_body, dest_content_length); | |
| break; | |
| case 0x10: | case 0x10: |
| dest_charset.transcodeFromUTF8(pool, | return dest_charset.transcodeFromUTF8(src); |
| source_body, source_content_length, | |
| dest_body, dest_content_length); | |
| break; | |
| case 0x11: | case 0x11: |
| dest_body=source_body; | return src; |
| dest_content_length=source_content_length; | |
| break; | |
| } | } |
| } | } |
| Line 280 static const XMLByte gUTFBytes[0x100] = | Line 263 static const XMLByte gUTFBytes[0x100] = |
| }; | }; |
| static const uint gUTFOffsets[6] = { | static const uint gUTFOffsets[6] = { |
| 0, 0x3080, 0xE2080, 0x3C82080, 0xFA082080, 0x82082080 | 0, 0x3080, 0xE2080, 0x3C82080, 0xFA082080, 0x82082080 |
| }; | }; |
| static const XMLByte gFirstByteMark[7] = { | static const XMLByte gFirstByteMark[7] = { |
| 0x00, 0x00, 0xC0, 0xE0, 0xF0, 0xF8, 0xFC | 0x00, 0x00, 0xC0, 0xE0, 0xF0, 0xF8, 0xFC |
| }; | }; |
| static size_t transcodeToUTF8( | static int transcodeToUTF8(const XMLByte* srcData, int& srcLen, |
| const void *source_body, size_t source_content_length, | XMLByte *toFill, int& toFillLen, |
| XMLByte *dest_body, size_t dest_buf_length, | const Charset::Tables& tables) { |
| const Charset::Tables& tables) { | const XMLByte* srcPtr=srcData; |
| const XMLByte* srcPtr=(const XMLByte*)source_body; | const XMLByte* srcEnd=srcData+srcLen; |
| const XMLByte* srcEnd=(const XMLByte*)source_body+source_content_length; | XMLByte* outPtr=toFill; |
| XMLByte* outPtr=dest_body; | XMLByte* outEnd=toFill+toFillLen; |
| XMLByte* outEnd=dest_body+dest_buf_length; | |
| while(srcPtr<srcEnd) { | while(srcPtr<srcEnd) { |
| uint curVal = tables.fromTable[*srcPtr]; | uint curVal = tables.fromTable[*srcPtr]; |
| if(!curVal) { | if(!curVal) { |
| // use the replacement character | // use the replacement character |
| *outPtr++= '?'; | *outPtr++= '?'; |
| srcPtr++; | srcPtr++; |
| continue; | continue; |
| } | } |
| // Figure out how many bytes we need | |
| unsigned int encodedBytes; | |
| if(curVal<0x80) | |
| encodedBytes = 1; | |
| else if(curVal<0x800) | |
| encodedBytes = 2; | |
| else if(curVal<0x10000) | |
| encodedBytes = 3; | |
| else if(curVal<0x200000) | |
| encodedBytes = 4; | |
| else if(curVal<0x4000000) | |
| encodedBytes = 5; | |
| else if(curVal<= 0x7FFFFFFF) | |
| encodedBytes = 6; | |
| else { | |
| // use the replacement character | |
| *outPtr++= '?'; | |
| srcPtr++; | |
| continue; | |
| } | |
| // If we cannot fully get this char into the output buffer | |
| if (outPtr + encodedBytes > outEnd) | |
| break; | |
| // We can do it, so update the source index | |
| srcPtr++; | |
| // And spit out the bytes. We spit them out in reverse order | |
| // here, so bump up the output pointer and work down as we go. | |
| outPtr+= encodedBytes; | |
| switch(encodedBytes) { | |
| case 6: *--outPtr = XMLByte((curVal | 0x80UL) & 0xBFUL); | |
| curVal>>= 6; | |
| case 5: *--outPtr = XMLByte((curVal | 0x80UL) & 0xBFUL); | |
| curVal>>= 6; | |
| case 4: *--outPtr = XMLByte((curVal | 0x80UL) & 0xBFUL); | |
| curVal>>= 6; | |
| case 3: *--outPtr = XMLByte((curVal | 0x80UL) & 0xBFUL); | |
| curVal>>= 6; | |
| case 2: *--outPtr = XMLByte((curVal | 0x80UL) & 0xBFUL); | |
| curVal>>= 6; | |
| case 1: *--outPtr = XMLByte(curVal | gFirstByteMark[encodedBytes]); | |
| } | |
| // Add the encoded bytes back in again to indicate we've eaten them | |
| outPtr+= encodedBytes; | |
| } | |
| // eaten | |
| return outPtr-dest_body; | |
| } | |
| static size_t transcodeFromUTF8( | |
| const void *source_body, size_t source_content_length, | |
| XMLByte* dest_body, size_t dest_buf_length, | |
| const Charset::Tables& tables) { | |
| const XMLByte* srcPtr=(const XMLByte*)source_body; | |
| const XMLByte* srcEnd=(const XMLByte*)source_body+source_content_length; | |
| XMLByte* outPtr=dest_body; | |
| XMLByte* outEnd=dest_body+dest_buf_length; | |
| // We now loop until we either run out of input data, or room to store | |
| while ((srcPtr < srcEnd) && (outPtr < outEnd)) { | |
| // Get the next leading byte out | |
| const XMLByte firstByte = *srcPtr; | |
| // Special-case ASCII, which is a leading byte value of<= 127 | |
| if(firstByte<= 127) { | |
| *outPtr++= firstByte; | |
| srcPtr++; | |
| continue; | |
| } | |
| // See how many trailing src bytes this sequence is going to require | |
| const unsigned int trailingBytes = gUTFBytes[firstByte]; | |
| // If there are not enough source bytes to do this one, then we | |
| // are done. Note that we done>= here because we are implicitly | |
| // counting the 1 byte we get no matter what. | |
| if(srcPtr+trailingBytes>= srcEnd) | |
| break; | |
| // Looks ok, so lets build up the value | |
| uint tmpVal=0; | |
| switch(trailingBytes) { | |
| case 5: tmpVal+=*srcPtr++; tmpVal<<=6; | |
| case 4: tmpVal+=*srcPtr++; tmpVal<<=6; | |
| case 3: tmpVal+=*srcPtr++; tmpVal<<=6; | |
| case 2: tmpVal+=*srcPtr++; tmpVal<<=6; | |
| case 1: tmpVal+=*srcPtr++; tmpVal<<=6; | |
| case 0: tmpVal+=*srcPtr++; | |
| break; | |
| default: | // Figure out how many bytes we need |
| throw Exception(0, 0, | unsigned int encodedBytes; |
| 0, | if(curVal<0x80) |
| "transcodeFromUTF8 error: wrong trailingBytes value(%d)", trailingBytes); | encodedBytes = 1; |
| } | else if(curVal<0x800) |
| tmpVal-=gUTFOffsets[trailingBytes]; | encodedBytes = 2; |
| else if(curVal<0x10000) | |
| // If it will fit into a single char, then put it in. Otherwise | encodedBytes = 3; |
| // fail [*encode it as a surrogate pair. If its not valid, use the | else if(curVal<0x200000) |
| // replacement char.*] | encodedBytes = 4; |
| if(!(tmpVal & 0xFFFF0000)) | else if(curVal<0x4000000) |
| *outPtr++= xlatOneTo(tmpVal, tables); | encodedBytes = 5; |
| else | else if(curVal<= 0x7FFFFFFF) |
| throw Exception(0, 0, | encodedBytes = 6; |
| else { | |
| // use the replacement character | |
| *outPtr++= '?'; | |
| srcPtr++; | |
| continue; | |
| } | |
| // If we cannot fully get this char into the output buffer | |
| if (outPtr + encodedBytes > outEnd) | |
| break; | |
| // We can do it, so update the source index | |
| srcPtr++; | |
| // And spit out the bytes. We spit them out in reverse order | |
| // here, so bump up the output pointer and work down as we go. | |
| outPtr+= encodedBytes; | |
| switch(encodedBytes) { | |
| case 6: *--outPtr = XMLByte((curVal | 0x80UL) & 0xBFUL); | |
| curVal>>= 6; | |
| case 5: *--outPtr = XMLByte((curVal | 0x80UL) & 0xBFUL); | |
| curVal>>= 6; | |
| case 4: *--outPtr = XMLByte((curVal | 0x80UL) & 0xBFUL); | |
| curVal>>= 6; | |
| case 3: *--outPtr = XMLByte((curVal | 0x80UL) & 0xBFUL); | |
| curVal>>= 6; | |
| case 2: *--outPtr = XMLByte((curVal | 0x80UL) & 0xBFUL); | |
| curVal>>= 6; | |
| case 1: *--outPtr = XMLByte(curVal | gFirstByteMark[encodedBytes]); | |
| } | |
| // Add the encoded bytes back in again to indicate we've eaten them | |
| outPtr+= encodedBytes; | |
| } | |
| // Update the bytes eaten | |
| srcLen = srcPtr - srcData; | |
| // Return the characters read | |
| toFillLen = outPtr - toFill; | |
| //return srcPtr==srcEnd?(int)toFillLen:-1; | |
| /* | |
| xmlCharEncodingInputFunc | |
| Returns : | |
| 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 | |
| number of octets consumed as the return value is positive, else unpredictiable. The value of outlen after return is the number | |
| of ocetes consumed. | |
| */ | |
| return 0; | |
| } | |
| /// @todo digital entites only when xml/html output [at output in html/xml mode, in html part of a letter] | |
| static int transcodeFromUTF8(const XMLByte* srcData, int& srcLen, | |
| XMLByte* toFill, int& toFillLen, | |
| const Charset::Tables& tables) { | |
| const XMLByte* srcPtr=srcData; | |
| const XMLByte* srcEnd=srcData+srcLen; | |
| XMLByte* outPtr=toFill; | |
| XMLByte* outEnd=toFill+toFillLen; | |
| // We now loop until we either run out of input data, or room to store | |
| while ((srcPtr < srcEnd) && (outPtr < outEnd)) { | |
| // Get the next leading byte out | |
| const XMLByte firstByte =* srcPtr; | |
| // Special-case ASCII, which is a leading byte value of<= 127 | |
| if(firstByte<=127) { | |
| *outPtr++= firstByte; | |
| srcPtr++; | |
| continue; | |
| } | |
| // See how many trailing src bytes this sequence is going to require | |
| const unsigned int trailingBytes = gUTFBytes[firstByte]; | |
| // If there are not enough source bytes to do this one, then we | |
| // are done. Note that we done>= here because we are implicitly | |
| // counting the 1 byte we get no matter what. | |
| if(srcPtr+trailingBytes>= srcEnd) | |
| break; | |
| // Looks ok, so lets build up the value | |
| uint tmpVal=0; | |
| switch(trailingBytes) { | |
| case 5: tmpVal+=*srcPtr++; tmpVal<<=6; | |
| case 4: tmpVal+=*srcPtr++; tmpVal<<=6; | |
| case 3: tmpVal+=*srcPtr++; tmpVal<<=6; | |
| case 2: tmpVal+=*srcPtr++; tmpVal<<=6; | |
| case 1: tmpVal+=*srcPtr++; tmpVal<<=6; | |
| case 0: tmpVal+=*srcPtr++; | |
| break; | |
| default: | |
| throw Exception(0, | |
| 0, | 0, |
| "transcodeFromUTF8 error: too big tmpVal(0x%08X)", tmpVal); | "transcodeFromUTF8 error: wrong trailingBytes value(%d)", trailingBytes); // never |
| } | |
| tmpVal-=gUTFOffsets[trailingBytes]; | |
| // If it will fit into a single char, then put it in. Otherwise | |
| // fail [*encode it as a surrogate pair. If its not valid, use the | |
| // replacement char.*] | |
| if(!(tmpVal & 0xFFFF0000)) { | |
| if(XMLByte xlat=xlatOneTo(tmpVal, tables, 0)) | |
| *outPtr++=xlat; | |
| else { | |
| outPtr+=sprintf((char *)outPtr, "&#%u;", tmpVal); // &#decimal; | |
| } | |
| } else { | |
| const XMLByte* recoverPtr=srcPtr-trailingBytes-1; | |
| for(uint i=0; i<=trailingBytes; i++) | |
| outPtr+=sprintf((char*)outPtr, "%%%02X", *recoverPtr++); | |
| } | |
| } | } |
| // Update the bytes eaten | |
| srcLen = srcPtr - srcData; | |
| // Return the characters read | |
| toFillLen = outPtr - toFill; | |
| // eaten | //return srcPtr==srcEnd?(int)toFillLen:-1; |
| return outPtr-dest_body; | /* |
| xmlCharEncodingOutputFunc | |
| Returns : | |
| 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 | |
| number of octets consumed as the return value is positive, else unpredictiable. The value of outlen after return is the number | |
| of ocetes consumed. | |
| */ | |
| return 0; | |
| } | } |
| /// @todo not so memory-hungry with prescan | static bool need_escape(XMLByte c){ |
| /// @test buf overflow | return |
| void Charset::transcodeToUTF8(Pool& pool, | !( |
| const void *source_body, size_t source_content_length, | (c<=127) |
| const void *& adest_body, size_t& adest_content_length) const { | && ( |
| size_t dest_buf_length=source_content_length*6/*so that surly enough*/; | ((c>='0') && (c<='9')) |
| XMLByte *dest_body=(XMLByte*)pool.malloc(dest_buf_length); | || ((c>='A') && (c<='Z')) |
| || ((c>='a') && (c<='z')) | |
| size_t dest_content_length=::transcodeToUTF8( | || strchr("*@-_+./", c)!=0 |
| source_body, source_content_length, | ) |
| dest_body, dest_buf_length, | ); |
| tables); | } |
| /*if(dest_content_length<dest_buf_length) | |
| throw(0, 0, | // read one UTF8 char and return length of this char (in bytes) |
| 0, | static unsigned int readUTF8Char(const XMLByte*& srcPtr, const XMLByte* srcEnd, XMLByte& firstByte, XMLCh& UTF8Char){ |
| "Charset::transcodeToUTF8 error");*/ | if(!srcPtr || !*srcPtr || srcPtr>=srcEnd) |
| return 0; | |
| firstByte=*srcPtr; | |
| if(firstByte<=127){ | |
| UTF8Char=firstByte; | |
| srcPtr++; | |
| return 1; | |
| } | |
| unsigned int trailingBytes=gUTFBytes[firstByte]; | |
| if(srcPtr+trailingBytes>=srcEnd){ | |
| return 0; // not enough bytes in source string for reading | |
| } | |
| uint tmpVal=0; | |
| switch(trailingBytes){ | |
| case 5: tmpVal+=*srcPtr++; tmpVal<<=6; | |
| case 4: tmpVal+=*srcPtr++; tmpVal<<=6; | |
| case 3: tmpVal+=*srcPtr++; tmpVal<<=6; | |
| case 2: tmpVal+=*srcPtr++; tmpVal<<=6; | |
| case 1: tmpVal+=*srcPtr++; tmpVal<<=6; | |
| case 0: tmpVal+=*srcPtr++; | |
| } | |
| tmpVal-=gUTFOffsets[trailingBytes]; | |
| UTF8Char=tmpVal; | |
| return trailingBytes+1; | |
| } | |
| // skip UTF8 char and return length of this char (in bytes) | |
| static unsigned int skipUTF8Char(const XMLByte*& srcPtr, const XMLByte* srcEnd){ | |
| if(!srcPtr || !*srcPtr || srcPtr>=srcEnd) | |
| return 0; | |
| unsigned int trailingBytes=gUTFBytes[*srcPtr]+1; | |
| srcPtr+=trailingBytes; | |
| return trailingBytes; | |
| } | |
| // read non-UTF8 char, and return number of bytes needed for storing this char in UTF8 | |
| static unsigned int readChar(const XMLByte*& srcPtr, const XMLByte* srcEnd, XMLByte& firstByte, XMLCh& UTF8Char, const Charset::Tables& tables){ | |
| if(!srcPtr || !*srcPtr || srcPtr>=srcEnd) | |
| return 0; | |
| firstByte=*srcPtr++; | |
| UTF8Char=tables.fromTable[firstByte]; | |
| if(UTF8Char<0x80) | |
| return 1; | |
| else if(UTF8Char<0x800) | |
| return 2; | |
| else if(UTF8Char<0x10000) | |
| return 3; | |
| else if(UTF8Char<0x200000) | |
| return 4; | |
| else if(UTF8Char<0x4000000) | |
| return 5; | |
| else if(UTF8Char<= 0x7FFFFFFF) | |
| return 6; | |
| // will use the replacement character '?' | |
| firstByte=0; | |
| return 1; | |
| } | |
| size_t Charset::calc_escaped_length_UTF8(XMLByte* src, size_t src_length){ | |
| size_t dest_length=0; | |
| for(UTF8_string_iterator i(src, src_length); i.has_next(); ){ | |
| if(i.getCharSize()==1) | |
| dest_length+=!need_escape(i.getFirstByte())?1/*as-is*/:3/*%XX*/; | |
| else | |
| dest_length+=6; // %uXXXX | |
| } | |
| return dest_length; | |
| } | |
| size_t Charset::calc_escaped_length(XMLByte* src, size_t src_length, const Charset::Tables& tables){ | |
| size_t dest_length=0; | |
| const XMLByte* src_ptr=src; | |
| const XMLByte* src_end=src_ptr+src_length; | |
| XMLByte first_byte; | |
| XMLCh UTF8_char; | |
| while(uint char_size=readChar(src_ptr, src_end, first_byte, UTF8_char, tables)){ | |
| if(char_size==1) | |
| dest_length+=(!first_byte/*replacement char '?'*/ || !need_escape(first_byte))?1:3/*'%XX'*/; | |
| else | |
| dest_length+=6; // %uXXXX | |
| } | |
| return dest_length; | |
| } | |
| size_t Charset::calc_escaped_length(const String::C src, const Charset& source_charset){ | |
| size_t src_length=src.length; | |
| if(!src_length) | |
| return 0; | |
| #ifdef PRECALCULATE_DEST_LENGTH | |
| if(source_charset.isUTF8()) | |
| return calc_escaped_length_UTF8((XMLByte *)src.str, src_length); | |
| else | |
| return calc_escaped_length((XMLByte *)src.str, src_length, source_charset.tables); | |
| #else | |
| return src_length*6; // enough for %uXXXX but too memory-hungry | |
| #endif | |
| } | |
| #define escape_char(dest_ptr, char_size, first_byte, UTF8_char) \ | |
| if(char_size==1) \ | |
| if(first_byte){ \ | |
| if(need_escape(first_byte)) \ | |
| dest_ptr+=sprintf((char*)dest_ptr, "%%%02X", first_byte); /* %XX */ \ | |
| else \ | |
| *dest_ptr++=first_byte; /*as is*/ \ | |
| } else \ | |
| *dest_ptr++='?'; /* replacement char '?' */ \ | |
| else \ | |
| dest_ptr+=sprintf((char*)dest_ptr, "%%u%04X", UTF8_char); // %uXXXX | |
| size_t Charset::escape_UTF8(const XMLByte* src, size_t src_length, XMLByte* dest) { | |
| XMLByte* dest_ptr=dest; | |
| // loop until we either run out of input data | |
| for(UTF8_string_iterator i((XMLByte *)src, src_length); i.has_next(); ) | |
| escape_char(dest_ptr, i.getCharSize(), i.getFirstByte(), i.next()) | |
| return dest_ptr - dest; | |
| } | |
| size_t Charset::escape(const XMLByte* src, size_t src_length, XMLByte* dest, const Charset::Tables& tables) { | |
| const XMLByte* src_ptr=src; | |
| const XMLByte* src_end=src+src_length; | |
| XMLByte* dest_ptr=dest; | |
| XMLByte first_byte; | |
| XMLCh UTF8_char; | |
| uint char_size; | |
| while(char_size=readChar(src_ptr, src_end, first_byte, UTF8_char, tables)) | |
| escape_char(dest_ptr, char_size, first_byte, UTF8_char) | |
| return dest_ptr - dest; | |
| } | |
| String::C Charset::escape(const String::C src, const Charset& source_charset){ | |
| size_t src_length=src.length; | |
| if(!src_length) | |
| return String::C("", 0); | |
| size_t dest_calculated_length=calc_escaped_length(src, source_charset); | |
| XMLByte *dest_body=new(PointerFreeGC) XMLByte[dest_calculated_length+1/*terminator*/]; | |
| size_t dest_length; | |
| if(source_charset.isUTF8()) | |
| dest_length=escape_UTF8((XMLByte *)src.str, src_length, dest_body); | |
| else | |
| dest_length=escape((XMLByte *)src.str, src_length, dest_body, source_charset.tables); | |
| if(dest_length>dest_calculated_length) | |
| throw Exception(0, 0, "Charset::escape buffer overflow"); | |
| dest_body[dest_length]=0; // terminator | |
| return String::C((char*)dest_body, dest_length); | |
| } | |
| String::Body Charset::escape(const String::Body src, const Charset& source_charset) { | |
| const char *src_ptr=src.cstr(); | |
| size_t src_size=src.length(); | |
| String::C dest=Charset::escape(String::C(src_ptr, src_size), source_charset); | |
| return String::Body(dest.length ? dest.str:0); | |
| } | |
| String& Charset::escape(const String& src, const Charset& source_charset) { | |
| if(src.is_empty()) | |
| return *new String(); | |
| return *new String(escape((String::Body)src, source_charset), String::L_CLEAN); | |
| } | |
| inline bool need_json_escape(unsigned char c){ | |
| return strchr("\n\"\\/\t\r\b\f", c)!=0; | |
| } | |
| size_t Charset::calc_JSON_escaped_length_UTF8(XMLByte* src, size_t src_length){ | |
| size_t dest_length=0; | |
| for(UTF8_string_iterator i(src, src_length); i.has_next(); ){ | |
| if(i.getCharSize()==1) | |
| dest_length+=need_json_escape(i.getFirstByte()) ? 2 : 1; | |
| else | |
| dest_length+=6; // \uXXXX | |
| } | |
| return dest_length; | |
| } | |
| size_t Charset::calc_JSON_escaped_length(XMLByte* src, size_t src_length, const Charset::Tables& tables){ | |
| const XMLByte* src_ptr=src; | |
| const XMLByte* src_end=src_ptr+src_length; | |
| XMLByte first_byte; | |
| XMLCh UTF8_char; | |
| size_t dest_length=0; | |
| while(uint char_size=readChar(src_ptr, src_end, first_byte, UTF8_char, tables)){ | |
| if(char_size==1) | |
| dest_length+=(!first_byte/*replacement char '?'*/ || !need_json_escape(first_byte))? 1 : 2; | |
| else | |
| dest_length+=6; // \uXXXX | |
| } | |
| return dest_length; | |
| } | |
| size_t Charset::calc_JSON_escaped_length(const String::C src, const Charset& source_charset){ | |
| size_t src_length=src.length; | |
| if(!src_length) | |
| return 0; | |
| #ifdef PRECALCULATE_DEST_LENGTH | |
| if(source_charset.isUTF8()) | |
| return calc_JSON_escaped_length_UTF8((XMLByte *)src.str, src_length); | |
| else | |
| return calc_JSON_escaped_length((XMLByte *)src.str, src_length, source_charset.tables); | |
| #else | |
| return src_length*6; // enough for \uXXXX but too memory-hungry | |
| #endif | |
| } | |
| #define escape_char_JSON(dest_ptr, char_size, first_byte, UTF8_char) \ | |
| if(char_size==1) \ | |
| switch(first_byte){ \ | |
| case '\n': *dest_ptr++='\\'; *dest_ptr++='n'; break; \ | |
| case '"' : *dest_ptr++='\\'; *dest_ptr++='"'; break; \ | |
| case '\\': *dest_ptr++='\\'; *dest_ptr++='\\'; break; \ | |
| case '/' : *dest_ptr++='\\'; *dest_ptr++='/'; break; \ | |
| case '\t': *dest_ptr++='\\'; *dest_ptr++='t'; break; \ | |
| case '\r': *dest_ptr++='\\'; *dest_ptr++='r'; break; \ | |
| case '\b': *dest_ptr++='\\'; *dest_ptr++='b'; break; \ | |
| case '\f': *dest_ptr++='\\'; *dest_ptr++='f'; break; \ | |
| case 0 : *dest_ptr++='?'; break; /*replacement char*/ \ | |
| default : *dest_ptr++=first_byte; \ | |
| } \ | |
| else \ | |
| dest_ptr+=sprintf((char*)dest_ptr, "\\u%04X", UTF8_char); // \uXXXX | |
| size_t Charset::escape_JSON_UTF8(const XMLByte* src, size_t src_length, XMLByte* dest) { | |
| XMLByte* dest_ptr=dest; | |
| // loop until we either run out of input data | |
| for(UTF8_string_iterator i((XMLByte *)src, src_length); i.has_next(); ) | |
| escape_char_JSON(dest_ptr, i.getCharSize(), i.getFirstByte(), i.next()) | |
| // return | return dest_ptr - dest; |
| adest_body=dest_body; | } |
| adest_content_length=dest_content_length; | |
| } | size_t Charset::escape_JSON(const XMLByte* src, size_t src_length, XMLByte* dest, const Charset::Tables& tables) { |
| /// @test buf overflow | const XMLByte* src_ptr=src; |
| void Charset::transcodeFromUTF8(Pool& pool, | const XMLByte* src_end=src+src_length; |
| const void *source_body, size_t source_content_length, | XMLByte* dest_ptr=dest; |
| const void *& adest_body, size_t& adest_content_length) const { | |
| size_t dest_buf_length=source_content_length/*surly enough*/; | XMLByte first_byte; |
| XMLByte *dest_body=(XMLByte*)pool.malloc(dest_buf_length); | XMLCh UTF8_char; |
| uint char_size; | |
| size_t dest_content_length=::transcodeFromUTF8( | |
| source_body, source_content_length, | while(char_size=readChar(src_ptr, src_end, first_byte, UTF8_char, tables)) |
| dest_body, dest_buf_length, | escape_char_JSON(dest_ptr, char_size, first_byte, UTF8_char) |
| tables); | |
| /* if(dest_content_length!=dest_buf_length) | return dest_ptr - dest; |
| throw(0, 0, | } |
| String::C Charset::escape_JSON(const String::C src, const Charset& source_charset){ | |
| size_t src_length=src.length; | |
| if(!src_length) | |
| return String::C("", 0); | |
| size_t dest_calculated_length=calc_JSON_escaped_length(src, source_charset); | |
| XMLByte *dest_body=new(PointerFreeGC) XMLByte[dest_calculated_length+1/*terminator*/]; | |
| size_t dest_length; | |
| if(source_charset.isUTF8()) | |
| dest_length=escape_JSON_UTF8((XMLByte *)src.str, src_length, dest_body); | |
| else | |
| dest_length=escape_JSON((XMLByte *)src.str, src_length, dest_body, source_charset.tables); | |
| if(dest_length>dest_calculated_length) | |
| throw Exception(0, 0, "Charset::escape_JSON buffer overflow"); | |
| dest_body[dest_length]=0; // terminator | |
| return String::C((char*)dest_body, dest_length); | |
| } | |
| String::Body Charset::escape_JSON(const String::Body src, const Charset& source_charset) { | |
| const char *src_ptr=src.cstr(); | |
| size_t src_size=src.length(); | |
| String::C dest=Charset::escape_JSON(String::C(src_ptr, src_size), source_charset); | |
| return String::Body(dest.length ? dest.str:0); | |
| } | |
| String& Charset::escape_JSON(const String& src, const Charset& source_charset) { | |
| if(src.is_empty()) | |
| return *new String(); | |
| return *new String(escape_JSON((String::Body)src, source_charset), String::L_CLEAN); | |
| } | |
| const String::C Charset::transcodeToUTF8(const String::C src) const { | |
| int src_length=src.length; | |
| #ifdef PRECALCULATE_DEST_LENGTH | |
| int dest_length=0; | |
| const XMLByte* srcPtr=(XMLByte*)src.str; | |
| const XMLByte* srcEnd=srcPtr+src_length; | |
| XMLByte firstByte; | |
| XMLCh UTF8Char; | |
| while(uint charSize=readChar(srcPtr, srcEnd, firstByte, UTF8Char, tables)) | |
| dest_length+=charSize; | |
| #else | |
| int dest_length=src_length*6; // so that surly enough (max utf8 seq len=6) but too memory-hungry | |
| #endif | |
| #ifndef NDEBUG | |
| int saved_dest_length=dest_length; | |
| #endif | |
| XMLByte *dest_body=new(PointerFreeGC) XMLByte[dest_length+1/*for terminator*/]; | |
| if(::transcodeToUTF8( | |
| (XMLByte *)src.str, src_length, | |
| dest_body, dest_length, | |
| tables)<0) | |
| throw Exception(0, | |
| 0, | 0, |
| "Charset::transcodeToUTF8 error");*/ | "Charset::transcodeToUTF8 buffer overflow"); |
| assert(dest_length<=saved_dest_length); | |
| dest_body[dest_length]=0; // terminator | |
| return String::C((char*)dest_body, dest_length); | |
| } | |
| static XMLCh change_case_UTF8(const XMLCh src, const Charset::UTF8CaseTable& table) { | |
| int lo = 0; | |
| int hi = table.size - 1; | |
| while(lo<=hi) { | |
| // Calc the mid point of the low and high offset. | |
| const unsigned int i = (lo + hi) / 2; | |
| XMLCh cur=table.records[i].from; | |
| if(src==cur) | |
| return table.records[i].to; | |
| if(src>cur) | |
| lo = i+1; | |
| else | |
| hi = i-1; | |
| } | |
| // return | // not found |
| adest_body=dest_body; | return src; |
| adest_content_length=dest_content_length; | |
| } | } |
| /// transcode using both charsets | static void store_UTF8(XMLCh src, XMLByte*& outPtr){ |
| void Charset::transcodeToCharset(Pool& pool, | if(!src) { |
| const Charset& dest_charset, | // use the replacement character |
| const void *source_body, size_t source_content_length, | *outPtr++= '?'; |
| const void *& adest_body, size_t& adest_content_length) const { | return; |
| if(&dest_charset==this) { | } |
| adest_body=source_body; | |
| adest_content_length=source_content_length; | |
| } else { | |
| size_t dest_content_length=source_content_length; | |
| unsigned char *dest_body=(unsigned char *)pool.malloc(dest_content_length); | |
| const XMLByte* srcPtr=(const XMLByte*)source_body; | // Figure out how many bytes we need |
| const XMLByte* srcEnd=(const XMLByte*)source_body+source_content_length; | unsigned int encodedBytes; |
| if(src<0x80) | |
| encodedBytes = 1; | |
| else if(src<0x800) | |
| encodedBytes = 2; | |
| else if(src<0x10000) | |
| encodedBytes = 3; | |
| else if(src<0x200000) | |
| encodedBytes = 4; | |
| else if(src<0x4000000) | |
| encodedBytes = 5; | |
| else if(src<= 0x7FFFFFFF) | |
| encodedBytes = 6; | |
| else { | |
| // use the replacement character | |
| *outPtr++= '?'; | |
| return; | |
| } | |
| for(XMLByte* outPtr=dest_body; srcPtr<srcEnd; srcPtr++) { | // And spit out the bytes. We spit them out in reverse order |
| XMLCh curVal = tables.fromTable[*srcPtr]; | // here, so bump up the output pointer and work down as we go. |
| if(curVal) | outPtr+= encodedBytes; |
| *outPtr++=xlatOneTo(curVal, dest_charset.tables); | switch(encodedBytes) { |
| else { | case 6: *--outPtr = XMLByte((src | 0x80UL) & 0xBFUL); |
| // use the replacement character | src>>= 6; |
| *outPtr++= '?'; | case 5: *--outPtr = XMLByte((src | 0x80UL) & 0xBFUL); |
| } | src>>= 6; |
| case 4: *--outPtr = XMLByte((src | 0x80UL) & 0xBFUL); | |
| src>>= 6; | |
| case 3: *--outPtr = XMLByte((src | 0x80UL) & 0xBFUL); | |
| src>>= 6; | |
| case 2: *--outPtr = XMLByte((src | 0x80UL) & 0xBFUL); | |
| src>>= 6; | |
| case 1: *--outPtr = XMLByte(src | gFirstByteMark[encodedBytes]); | |
| } | |
| // Add the encoded bytes back in again to indicate we've eaten them | |
| outPtr+= encodedBytes; | |
| } | |
| static void change_case_UTF8(XMLCh src, XMLByte*& outPtr, | |
| const Charset::UTF8CaseTable& table) { | |
| store_UTF8(change_case_UTF8(src, table), outPtr); | |
| }; | |
| void change_case_UTF8(const XMLByte* srcData, size_t srcLen, | |
| XMLByte* toFill, size_t toFillLen, | |
| const Charset::UTF8CaseTable& table) { | |
| const XMLByte* srcPtr=srcData; | |
| const XMLByte* srcEnd=srcData+srcLen; | |
| XMLByte* outPtr=toFill; | |
| XMLByte* outEnd=toFill+toFillLen; | |
| // We now loop until we either run out of input data, or room to store | |
| while ((srcPtr < srcEnd) && (outPtr < outEnd)) { | |
| // Get the next leading byte out | |
| const XMLByte firstByte =* srcPtr; | |
| if(firstByte<=127) { | |
| change_case_UTF8(firstByte, outPtr, table); | |
| srcPtr++; | |
| continue; | |
| } | |
| // See how many trailing src bytes this sequence is going to require | |
| const unsigned int trailingBytes = gUTFBytes[firstByte]; | |
| // Looks ok, so lets build up the value | |
| uint tmpVal=0; | |
| switch(trailingBytes) { | |
| case 5: tmpVal+=*srcPtr++; tmpVal<<=6; | |
| case 4: tmpVal+=*srcPtr++; tmpVal<<=6; | |
| case 3: tmpVal+=*srcPtr++; tmpVal<<=6; | |
| case 2: tmpVal+=*srcPtr++; tmpVal<<=6; | |
| case 1: tmpVal+=*srcPtr++; tmpVal<<=6; | |
| case 0: tmpVal+=*srcPtr++; | |
| break; | |
| default: | |
| throw Exception(0, | |
| 0, | |
| "change_case_UTF8 error: wrong trailingBytes value(%d)", trailingBytes); | |
| } | |
| tmpVal-=gUTFOffsets[trailingBytes]; | |
| // If it will fit into a single char, then put it in. Otherwise | |
| // fail [*encode it as a surrogate pair. If its not valid, use the | |
| // replacement char.*] | |
| if(!(tmpVal & 0xFFFF0000)) | |
| change_case_UTF8(tmpVal, outPtr, table); | |
| else | |
| throw Exception(0, | |
| 0, | |
| "change_case_UTF8 error: too big tmpVal(0x%08X)", tmpVal); | |
| } | |
| if(srcPtr!=outPtr) | |
| throw Exception(0, | |
| 0, | |
| "change_case_UTF8 error: end pointers do not match"); | |
| } | |
| static size_t getDecNumLength(XMLCh UTF8Char){ | |
| return | |
| (UTF8Char < 100) | |
| ?2 | |
| :(UTF8Char < 1000) | |
| ?3 | |
| :(UTF8Char < 10000) | |
| ?4 | |
| :5; | |
| } | |
| const String::C Charset::transcodeFromUTF8(const String::C src) const { | |
| int src_length=src.length; | |
| #ifdef PRECALCULATE_DEST_LENGTH | |
| int dest_length=0; | |
| for(UTF8_string_iterator i((XMLByte *)src.str, src_length); i.has_next(); ){ | |
| if(i.getCharSize()==1) | |
| dest_length++; | |
| else | |
| dest_length+=(i.next() & 0xFFFF0000) | |
| ?i.getCharSize()*3 // '%XX' for each byte | |
| :(xlatOneTo(i.next(), tables, 0)!=0) | |
| ?1 // can convert it to single char | |
| :getDecNumLength(i.next())+3; // &#XX; - &#XXXXX; | |
| } | |
| #else | |
| // so that surly enough, "&#XXX;" has max ratio (huh? 8 bytes needed for '&#XXXXX;') | |
| int dest_length=src_length*6; | |
| #endif | |
| #ifndef NDEBUG | |
| int saved_dest_length=dest_length; | |
| #endif | |
| XMLByte *dest_body=new(PointerFreeGC) XMLByte[dest_length+1/*for terminator*/]; | |
| if(::transcodeFromUTF8( | |
| (XMLByte *)src.str, src_length, | |
| dest_body, dest_length, | |
| tables)<0) | |
| throw Exception(0, | |
| 0, | |
| "Charset::transcodeFromUTF8 buffer overflow"); | |
| assert(dest_length<=saved_dest_length); | |
| dest_body[dest_length]=0; // terminator | |
| return String::C((char*)dest_body, dest_length); | |
| } | |
| /// transcode using both charsets | |
| const String::C Charset::transcodeToCharset(const String::C src, | |
| const Charset& dest_charset) const { | |
| if(&dest_charset==this) | |
| return src; | |
| else { | |
| size_t dest_length=src.length; | |
| XMLByte* dest_body=new(PointerFreeGC) XMLByte[dest_length+1/*for terminator*/]; | |
| XMLByte* output=dest_body; | |
| const XMLByte* input=(XMLByte *)src.str; | |
| while(XMLCh c=*input++) { | |
| XMLCh curVal = tables.fromTable[c]; | |
| *output++=curVal? | |
| xlatOneTo(curVal, dest_charset.tables, '?') // OK | |
| :'?'; // use the replacement character | |
| } | } |
| adest_body=dest_body; | dest_body[dest_length]=0; // terminator |
| adest_content_length=dest_content_length; | return String::C((char*)dest_body, dest_length); |
| } | } |
| } | } |
| #ifdef XML | void Charset::store_Char(XMLByte*& outPtr, XMLCh src, XMLByte not_found){ |
| static int xml256CharEncodingInputFunc ( | if(isUTF8()) |
| unsigned char *out, | store_UTF8(src, outPtr); |
| int *outlen, | else if(char ch=xlatOneTo(src, tables, not_found)) |
| const unsigned char *in, | *outPtr++=ch; |
| int *inlen, | |
| void *info) { | |
| return transcodeToUTF8( | |
| in, *inlen, | |
| out, *(unsigned int*)outlen, | |
| *(const Charset::Tables *)info); | |
| } | |
| static int xml256CharEncodingOutputFunc ( | |
| unsigned char *out, | |
| int *outlen, | |
| const unsigned char *in, | |
| int *inlen, | |
| void *info) { | |
| return transcodeFromUTF8( | |
| in, *inlen, | |
| out, *outlen, | |
| *(const Charset::Tables *)info); | |
| } | } |
| #ifdef XML | |
| static const Charset::Tables* tables[MAX_CHARSETS]; | |
| #ifdef PA_PATCHED_LIBXML_BACKWARD | |
| #define declareXml256ioFuncs(i) \ | |
| static int xml256CharEncodingInputFunc##i( \ | |
| unsigned char *out, int *outlen, \ | |
| const unsigned char *in, int *inlen, void*) { \ | |
| return transcodeToUTF8( \ | |
| in, *inlen, \ | |
| out, *outlen, \ | |
| *tables[i]); \ | |
| } \ | |
| static int xml256CharEncodingOutputFunc##i( \ | |
| unsigned char *out, int *outlen, \ | |
| const unsigned char *in, int *inlen, void*) { \ | |
| return transcodeFromUTF8( \ | |
| in, *inlen, \ | |
| out, *outlen, \ | |
| *tables[i]); \ | |
| } | |
| #else | |
| #define declareXml256ioFuncs(i) \ | |
| static int xml256CharEncodingInputFunc##i( \ | |
| unsigned char *out, int *outlen, \ | |
| const unsigned char *in, int *inlen) { \ | |
| return transcodeToUTF8( \ | |
| in, *inlen, \ | |
| out, *outlen, \ | |
| *tables[i]); \ | |
| } \ | |
| static int xml256CharEncodingOutputFunc##i( \ | |
| unsigned char *out, int *outlen, \ | |
| const unsigned char *in, int *inlen) { \ | |
| return transcodeFromUTF8( \ | |
| in, *inlen, \ | |
| out, *outlen, \ | |
| *tables[i]); \ | |
| } | |
| #endif | |
| declareXml256ioFuncs(0) declareXml256ioFuncs(1) | |
| declareXml256ioFuncs(2) declareXml256ioFuncs(3) | |
| declareXml256ioFuncs(4) declareXml256ioFuncs(5) | |
| declareXml256ioFuncs(6) declareXml256ioFuncs(7) | |
| declareXml256ioFuncs(8) declareXml256ioFuncs(9) | |
| static xmlCharEncodingInputFunc inputFuncs[MAX_CHARSETS]={ | |
| xml256CharEncodingInputFunc0, xml256CharEncodingInputFunc1, | |
| xml256CharEncodingInputFunc2, xml256CharEncodingInputFunc3, | |
| xml256CharEncodingInputFunc4, xml256CharEncodingInputFunc5, | |
| xml256CharEncodingInputFunc6, xml256CharEncodingInputFunc7, | |
| xml256CharEncodingInputFunc8, xml256CharEncodingInputFunc9 | |
| }; | |
| static xmlCharEncodingOutputFunc outputFuncs[MAX_CHARSETS]={ | |
| xml256CharEncodingOutputFunc0, xml256CharEncodingOutputFunc1, | |
| xml256CharEncodingOutputFunc2, xml256CharEncodingOutputFunc3, | |
| xml256CharEncodingOutputFunc4, xml256CharEncodingOutputFunc5, | |
| xml256CharEncodingOutputFunc6, xml256CharEncodingOutputFunc7, | |
| xml256CharEncodingOutputFunc8, xml256CharEncodingOutputFunc9 | |
| }; | |
| static size_t handlers_count=0; | |
| void Charset::addEncoding(char *name_cstr) { | void Charset::addEncoding(char *name_cstr) { |
| xmlCharEncodingHandler *handler= | if(handlers_count==MAX_CHARSETS) |
| (xmlCharEncodingHandler *)malloc(sizeof(xmlCharEncodingHandler)); | throw Exception(0, |
| handler->name=name_cstr; | 0, |
| handler->input=xml256CharEncodingInputFunc; handler->inputInfo=&tables; | "already allocated %d handlers, no space for new encoding '%s'", |
| handler->output=xml256CharEncodingOutputFunc; handler->outputInfo=&tables; | MAX_CHARSETS, name_cstr); |
| xmlCharEncodingHandler* handler=new(UseGC) xmlCharEncodingHandler; | |
| { | |
| handler->name=name_cstr; | |
| handler->input=inputFuncs[handlers_count]; | |
| handler->output=outputFuncs[handlers_count]; | |
| ::tables[handlers_count]=&tables; | |
| handlers_count++; | |
| } | |
| xmlRegisterCharEncodingHandler(handler); | xmlRegisterCharEncodingHandler(handler); |
| } | |
| void Charset::initTranscoder(const String::Body NAME, const char* name_cstr) { | |
| ftranscoder=xmlFindCharEncodingHandler(name_cstr); | |
| transcoder(NAME); // check right way | |
| } | } |
| void Charset::initTranscoder(const String *source, const char *name_cstr) { | xmlCharEncodingHandler& Charset::transcoder(const String::Body NAME) { |
| transcoder=xmlFindCharEncodingHandler(name_cstr); | if(!ftranscoder) |
| if(!transcoder) | throw Exception(PARSER_RUNTIME, |
| throw Exception(0, 0, | new String(NAME, String::L_TAINTED), |
| source, | |
| "unsupported encoding"); | "unsupported encoding"); |
| return *ftranscoder; | |
| } | } |
| const char *Charset::transcode_cstr(GdomeDOMString *s) { | String::C Charset::transcode_cstr(const xmlChar* s) { |
| if(!transcoder) | if(!s) |
| throw Exception(0, 0, | return String::C("", 0); |
| 0, | |
| "transcode_cstr no transcoder"); | |
| int inlen=gdome_str_length(s); | int inlen=strlen((const char*)s); |
| int outlen=inlen+1; // max | int outlen=inlen*6/*strlen("ÿ")*/; // max |
| char *out=(char *)malloc(outlen*sizeof(char)); | #ifndef NDEBUG |
| int saved_outlen=outlen; | |
| int size=transcoder->output( | #endif |
| (unsigned char*)out, &outlen, | char *out=new(PointerFreeGC) char[outlen+1]; |
| (const unsigned char*)s->str, &inlen, | |
| transcoder->outputInfo); | int error; |
| if(size<0) | if(xmlCharEncodingOutputFunc output=transcoder(FNAME).output) { |
| throw Exception(0, 0, | error=output( |
| (unsigned char*)out, &outlen, | |
| (const unsigned char*)s, &inlen | |
| #ifdef PA_PATCHED_LIBXML_BACKWARD | |
| ,0 | |
| #endif | |
| ); | |
| } else { | |
| memcpy(out, s, outlen=inlen); | |
| error=0; | |
| } | |
| if(error<0) | |
| throw Exception(0, | |
| 0, | 0, |
| "transcode_cstr failed (%d)", size); | "transcode_cstr failed (%d)", error); |
| out[size]=0; | assert(outlen<=saved_outlen); out[outlen]=0; |
| return out; | return String::C(out, outlen); |
| } | } |
| String& Charset::transcode(GdomeDOMString *s) { | const String& Charset::transcode(const xmlChar* s) { |
| return *NEW String(pool(), transcode_cstr(s)); | String::C cstr=transcode_cstr(s); |
| return *new String(cstr.str, String::L_TAINTED); | |
| } | } |
| /// @test less memory using -maybe- xmlParserInputBufferCreateMem | /// @test less memory using -maybe- xmlParserInputBufferCreateMem |
| GdomeDOMString *Charset::transcode_buf(const char *buf, size_t buf_size) { | xmlChar* Charset::transcode_buf2xchar(const char* buf, size_t buf_size) { |
| if(!transcoder) | xmlChar* out; |
| throw Exception(0, 0, | int outlen; |
| int error; | |
| #ifndef NDEBUG | |
| int saved_outlen; | |
| #endif | |
| if(xmlCharEncodingInputFunc input=transcoder(FNAME).input) { | |
| outlen=buf_size*6/*max UTF8 bytes per char*/; | |
| #ifndef NDEBUG | |
| saved_outlen=outlen; | |
| #endif | |
| out=(xmlChar*)xmlMalloc(outlen+1); | |
| error=input( | |
| out, &outlen, | |
| (const unsigned char*)buf, (int*)&buf_size | |
| #ifdef PA_PATCHED_LIBXML_BACKWARD | |
| ,0 | |
| #endif | |
| ); | |
| } else { | |
| outlen=buf_size; | |
| #ifndef NDEBUG | |
| saved_outlen=outlen; | |
| #endif | |
| out=(xmlChar*)xmlMalloc(outlen+1); | |
| memcpy(out, buf, outlen); | |
| error=0; | |
| } | |
| if(error<0) | |
| throw Exception(0, | |
| 0, | 0, |
| "transcode_buf no transcoder"); | "transcode_buf failed (%d)", error); |
| int outlen=buf_size*6/*max*/+1; | assert(outlen<=saved_outlen); out[outlen]=0; |
| unsigned char *out=(unsigned char*)malloc(outlen*sizeof(unsigned char)); | return out; |
| int size=transcoder->input( | } |
| out, &outlen, | |
| (const unsigned char *)buf, (int *)&buf_size, | |
| transcoder->inputInfo); | |
| if(size<0) | |
| throw Exception(0, 0, | |
| 0, | |
| "transcode_buf failed (%d)", size); | |
| out[size]=0; | xmlChar* Charset::transcode(const String& s) { |
| return gdome_str_mkref_own((gchar*)out); | String::Body sbody=s.cstr_to_string_body_untaint(String::L_AS_IS); |
| return transcode_buf2xchar(sbody.cstr(), sbody.length()); | |
| } | } |
| GdomeDOMString *Charset::transcode(const String& s) { | |
| const char *cstr=s.cstr(String::UL_UNSPECIFIED); | |
| return transcode_buf(cstr, strlen(cstr)); | xmlChar* Charset::transcode(const String::Body s) { |
| return transcode_buf2xchar(s.cstr(), s.length()); | |
| } | } |
| #endif | #endif |
| String::Body Charset::transcode(const String::Body src, | |
| const Charset& source_transcoder, | |
| const Charset& dest_transcoder) { | |
| const char *src_ptr=src.cstr(); | |
| size_t src_size=src.length(); | |
| String::C dest=Charset::transcode(String::C(src_ptr, src_size), source_transcoder, dest_transcoder); | |
| return String::Body(dest.length ? dest.str:0); | |
| } | |
| String& Charset::transcode(const String& src, | |
| const Charset& source_transcoder, | |
| const Charset& dest_transcoder) { | |
| if(src.is_empty()) | |
| return *new String(); | |
| return *new String(transcode((String::Body)src, source_transcoder, dest_transcoder), String::L_CLEAN); | |
| } | |
| void Charset::transcode(ArrayString& src, | |
| const Charset& source_transcoder, | |
| const Charset& dest_transcoder) { | |
| for(size_t i=0; i<src.count(); i++) | |
| src.put(i, &transcode(*src[i], source_transcoder, dest_transcoder)); | |
| } | |
| #ifndef DOXYGEN | |
| struct Transcode_pair_info { | |
| const Charset* source_transcoder; | |
| const Charset* dest_transcoder; | |
| }; | |
| #endif | |
| static void transcode_pair(HashStringValue::key_type /*akey*/, | |
| String::Body& avalue, | |
| Transcode_pair_info* info) { | |
| avalue=Charset::transcode(avalue, | |
| *info->source_transcoder, | |
| *info->dest_transcoder); | |
| } | |
| void Charset::transcode(HashStringString& src, | |
| const Charset& source_transcoder, | |
| const Charset& dest_transcoder) { | |
| Transcode_pair_info info={&source_transcoder, &dest_transcoder}; | |
| src.for_each_ref<Transcode_pair_info*>(transcode_pair, &info); | |
| } | |
| size_t getUTF8BytePos(const XMLByte* srcBegin, const XMLByte* srcEnd, size_t charPos){ | |
| const XMLByte* ptr=srcBegin; | |
| while(charPos-- && skipUTF8Char(ptr, srcEnd)); | |
| return ptr-srcBegin; | |
| } | |
| size_t getUTF8CharPos(const XMLByte* srcBegin, const XMLByte* srcEnd, size_t bytePos){ | |
| size_t charPos=0; | |
| const XMLByte* ptr=srcBegin; | |
| const XMLByte* ptrEnd=srcBegin+bytePos; | |
| while(skipUTF8Char(ptr, srcEnd)){ | |
| if(ptr>ptrEnd) | |
| return charPos; | |
| charPos++; | |
| } | |
| // scan till end but position in bytes still too low | |
| throw Exception(0, | |
| 0, | |
| "Error convertion byte pos to char pos"); | |
| } | |
| size_t lengthUTF8(const XMLByte* srcBegin, const XMLByte* srcEnd){ | |
| size_t size=0; | |
| while(skipUTF8Char(srcBegin, srcEnd)) | |
| size++; | |
| return size; | |
| } | |
| unsigned int lengthUTF8Char(const XMLByte c){ | |
| return gUTFBytes[c]+1; | |
| } | |
| bool UTF8_string_iterator::has_next(){ | |
| fcharSize=readUTF8Char(fsrcPtr, fsrcEnd, ffirstByte, fUTF8Char); | |
| return fcharSize!=0; | |
| } |