|
|
| version 1.35, 2003/07/24 11:31:23 | version 1.60, 2008/07/15 12:53:10 |
|---|---|
| Line 1 | Line 1 |
| /** @file | /** @file |
| Parser: Charset connection implementation. | Parser: Charset connection implementation. |
| Copyright(c) 2001-2003 ArtLebedev Group (http://www.artlebedev.com) | Copyright(c) 2001-2005 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) |
| */ | */ |
| static const char* IDENT_CHARSET_C="$Date$"; | static const char * const IDENT_CHARSET_C="$Date$"; |
| #include "pa_charset.h" | #include "pa_charset.h" |
| #include "pa_charsets.h" | #include "pa_charsets.h" |
| Line 14 static const char* IDENT_CHARSET_C="$Dat | Line 14 static const char* IDENT_CHARSET_C="$Dat |
| #include "libxml/encoding.h" | #include "libxml/encoding.h" |
| #endif | #endif |
| //#define PA_PATCHED_LIBXML_BACKWARD | |
| #define PRECALCULATE_DEST_LENGTH | |
| // globals | |
| Charset::UTF8CaseTable::Rec UTF8CaseToUpperRecords[]={ | |
| #include "utf8-to-upper.inc" | |
| }; | |
| Charset::UTF8CaseTable UTF8CaseToUpper={ | |
| sizeof(UTF8CaseToUpperRecords)/sizeof(Charset::UTF8CaseTable::Rec), | |
| UTF8CaseToUpperRecords}; | |
| Charset::UTF8CaseTable::Rec UTF8CaseToLowerRecords[]={ | |
| #include "utf8-to-lower.inc" | |
| }; | |
| Charset::UTF8CaseTable UTF8CaseToLower={ | |
| sizeof(UTF8CaseToLowerRecords)/sizeof(Charset::UTF8CaseTable::Rec), | |
| UTF8CaseToLowerRecords}; | |
| // helpers | // helpers |
| inline void prepare_case_tables(unsigned char *tables) { | inline void prepare_case_tables(unsigned char *tables) { |
| 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 68 static void element2case(unsigned char f | Line 87 static void element2case(unsigned char f |
| // methods | // methods |
| extern "C" unsigned char pcre_default_tables[]; // pcre/chartables.c | extern "C" unsigned char pcre_default_tables[]; // pcre/chartables.c |
| Charset::Charset(Request_charsets* charsets, const StringBody ANAME, const String* afile_spec): | Charset::Charset(Request_charsets* charsets, const String::Body ANAME, const String* afile_spec): |
| FNAME(ANAME), | FNAME(ANAME), |
| FNAME_CSTR(ANAME.cstrm()) { | FNAME_CSTR(ANAME.cstrm()) { |
| Line 101 void Charset::load_definition(Request_ch | Line 120 void Charset::load_definition(Request_ch |
| // charset | // charset |
| memset(&tables, 0, sizeof(tables)); | memset(&tables, 0, sizeof(tables)); |
| // 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(charsets, afile_spec); | char *data=file_read_text(charsets, afile_spec); |
| Line 114 void Charset::load_definition(Request_ch | Line 129 void Charset::load_definition(Request_ch |
| // 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("parser.runtime", | throw Exception(PARSER_RUNTIME, |
| &afile_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); |
| Line 161 void Charset::load_definition(Request_ch | Line 176 void Charset::load_definition(Request_ch |
| 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 173 void Charset::sort_ToTable() { | Line 188 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, |
| XMLByte not_found) { | XMLByte not_found) { |
| unsigned int lowOfs = 0; | int lo = 0; |
| unsigned int hiOfs = tables.toTableSize - 1; | int hi = tables.toTableSize - 1; |
| XMLByte curByte = 0; | while(lo<=hi) { |
| do { | |
| // 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 not_found; | return not_found; |
| } | } |
| Line 305 static int transcodeToUTF8(const XMLByte | Line 318 static int transcodeToUTF8(const XMLByte |
| // here, so bump up the output pointer and work down as we go. | // here, so bump up the output pointer and work down as we go. |
| outPtr+= encodedBytes; | outPtr+= encodedBytes; |
| switch(encodedBytes) { | switch(encodedBytes) { |
| case 6: *--outPtr = XMLByte((curVal | 0x80UL) & 0xBFUL); | case 6: *--outPtr = XMLByte((curVal | 0x80UL) & 0xBFUL); |
| curVal>>= 6; | curVal>>= 6; |
| case 5: *--outPtr = XMLByte((curVal | 0x80UL) & 0xBFUL); | case 5: *--outPtr = XMLByte((curVal | 0x80UL) & 0xBFUL); |
| curVal>>= 6; | curVal>>= 6; |
| case 4: *--outPtr = XMLByte((curVal | 0x80UL) & 0xBFUL); | case 4: *--outPtr = XMLByte((curVal | 0x80UL) & 0xBFUL); |
| curVal>>= 6; | curVal>>= 6; |
| case 3: *--outPtr = XMLByte((curVal | 0x80UL) & 0xBFUL); | case 3: *--outPtr = XMLByte((curVal | 0x80UL) & 0xBFUL); |
| curVal>>= 6; | curVal>>= 6; |
| case 2: *--outPtr = XMLByte((curVal | 0x80UL) & 0xBFUL); | case 2: *--outPtr = XMLByte((curVal | 0x80UL) & 0xBFUL); |
| curVal>>= 6; | curVal>>= 6; |
| case 1: *--outPtr = XMLByte(curVal | gFirstByteMark[encodedBytes]); | case 1: *--outPtr = XMLByte(curVal | gFirstByteMark[encodedBytes]); |
| } | } |
| // Add the encoded bytes back in again to indicate we've eaten them | // Add the encoded bytes back in again to indicate we've eaten them |
| Line 353 static int transcodeFromUTF8(const XMLBy | Line 366 static int transcodeFromUTF8(const XMLBy |
| const XMLByte firstByte =* srcPtr; | const XMLByte firstByte =* srcPtr; |
| // Special-case ASCII, which is a leading byte value of<= 127 | // Special-case ASCII, which is a leading byte value of<= 127 |
| if(firstByte<= 127) { | if(firstByte<=127) { |
| *outPtr++= firstByte; | *outPtr++= firstByte; |
| srcPtr++; | srcPtr++; |
| continue; | continue; |
| Line 382 static int transcodeFromUTF8(const XMLBy | Line 395 static int transcodeFromUTF8(const XMLBy |
| default: | default: |
| throw Exception(0, | throw Exception(0, |
| 0, | 0, |
| "transcodeFromUTF8 error: wrong trailingBytes value(%d)", trailingBytes); | "transcodeFromUTF8 error: wrong trailingBytes value(%d)", trailingBytes); // never |
| } | } |
| tmpVal-=gUTFOffsets[trailingBytes]; | tmpVal-=gUTFOffsets[trailingBytes]; |
| Line 392 static int transcodeFromUTF8(const XMLBy | Line 405 static int transcodeFromUTF8(const XMLBy |
| if(!(tmpVal & 0xFFFF0000)) { | if(!(tmpVal & 0xFFFF0000)) { |
| if(XMLByte xlat=xlatOneTo(tmpVal, tables, 0)) | if(XMLByte xlat=xlatOneTo(tmpVal, tables, 0)) |
| *outPtr++=xlat; | *outPtr++=xlat; |
| else | else { |
| outPtr+=sprintf((char *)outPtr, "&#%d;", tmpVal); // &#decimal; | outPtr+=sprintf((char *)outPtr, "&#%u;", tmpVal); // &#decimal; |
| } else | } |
| throw Exception(0, | } else { |
| 0, | const XMLByte* recoverPtr=srcPtr-trailingBytes-1; |
| "transcodeFromUTF8 error: too big tmpVal(0x%08X)", tmpVal); | for(uint i=0; i<=trailingBytes; i++) |
| outPtr+=sprintf((char*)outPtr, "%%%02X", *recoverPtr++); | |
| } | |
| } | } |
| // Update the bytes eaten | // Update the bytes eaten |
| Line 417 of ocetes consumed. | Line 432 of ocetes consumed. |
| return 0; | return 0; |
| } | } |
| /// @todo not so memory-hungry with prescan | static bool is_escaped(char c){ |
| return | |
| !(c<=127 | |
| && ( | |
| ((c>='0') && (c<='9')) | |
| || ((c>='A') && (c<='Z')) | |
| || ((c>='a') && (c<='z')) | |
| || strchr("*@-_+./", c)!=0 | |
| )); | |
| } | |
| // read one utf8 character, return number of bytes needed for store it | |
| static unsigned int readChar(const XMLByte*& srcPtr, const XMLByte*& srcEnd, XMLByte& firstByte, XMLCh& UTF8Char){ | |
| 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; | |
| } | |
| // read char, return number of bytes needed for store it as 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; | |
| } | |
| static int escape(const XMLByte* srcData, size_t& srcLen, | |
| XMLByte* toFill, size_t& toFillLen) { | |
| const XMLByte* srcPtr=srcData; | |
| const XMLByte* srcEnd=srcData+srcLen; | |
| XMLByte* outPtr=toFill; | |
| XMLByte* outEnd=toFill+toFillLen; | |
| XMLByte firstByte; | |
| XMLCh UTF8Char; | |
| uint charSize; | |
| // loop until we either run out of input data, or room to store | |
| while((outPtr < outEnd) && (charSize=readChar(srcPtr, srcEnd, firstByte, UTF8Char))){ | |
| if(charSize==1){ | |
| if(is_escaped(firstByte)) // %XX | |
| outPtr+=sprintf((char*)outPtr, "%%%02X", firstByte); | |
| else | |
| *outPtr++=firstByte; | |
| } else | |
| outPtr+=sprintf((char*)outPtr, "%%u%04X", UTF8Char); // %uXXXX | |
| } | |
| // Update the bytes eaten | |
| srcLen=srcPtr-srcData; | |
| // Return the characters read | |
| toFillLen=outPtr-toFill; | |
| return 0; | |
| } | |
| static int escape(const XMLByte* srcData, size_t& srcLen, | |
| XMLByte *toFill, size_t& toFillLen, | |
| const Charset::Tables& tables) { | |
| const XMLByte* srcPtr=srcData; | |
| const XMLByte* srcEnd=srcData+srcLen; | |
| XMLByte* outPtr=toFill; | |
| //XMLByte* outEnd=toFill+toFillLen; | |
| XMLByte firstByte; | |
| XMLCh UTF8Char; | |
| uint charSize; | |
| while(charSize=readChar(srcPtr, srcEnd, firstByte, UTF8Char, tables)){ | |
| if(charSize==1){ | |
| if(firstByte){ | |
| if(is_escaped(firstByte)) // %XX | |
| outPtr+=sprintf((char*)outPtr, "%%%02X", firstByte); | |
| else | |
| *outPtr++=firstByte; | |
| } else // add replacement char '?' | |
| *outPtr++='?'; | |
| } else | |
| outPtr+=sprintf((char*)outPtr, "%%u%04X", UTF8Char); // %uXXXX | |
| } | |
| // Update the bytes eaten | |
| srcLen = srcPtr - srcData; | |
| // Return the characters read | |
| toFillLen = outPtr - toFill; | |
| return 0; | |
| } | |
| 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); | |
| #ifdef PRECALCULATE_DEST_LENGTH | |
| size_t dest_length=0; | |
| const XMLByte* srcPtr=(XMLByte*)src.str; | |
| const XMLByte* srcEnd=srcPtr+src_length; | |
| XMLByte firstByte; | |
| XMLCh UTF8Char; | |
| if(source_charset.isUTF8()){ | |
| while(uint charSize=readChar(srcPtr, srcEnd, firstByte, UTF8Char)){ | |
| if(charSize==1) | |
| dest_length+=!is_escaped(firstByte)?1:3/*%XX*/; | |
| else | |
| dest_length+=6; // '%uXXXX' | |
| } | |
| } else { | |
| while(uint charSize=readChar(srcPtr, srcEnd, firstByte, UTF8Char, source_charset.tables)){ | |
| if(charSize==1) | |
| dest_length+=(!firstByte/*replacement char '?'*/ || !is_escaped(firstByte))?1:3/*'%XX'*/; | |
| else | |
| dest_length+=6; // '%uXXXX' | |
| } | |
| } | |
| #else | |
| size_t dest_length=src_length*6; // enough for %uXXXX but too memory-hungry | |
| #endif | |
| //throw Exception(0,0,"%u",dest_length); | |
| #ifndef NDEBUG | |
| size_t saved_dest_length=dest_length; | |
| #endif | |
| XMLByte *dest_body=new(PointerFreeGC) XMLByte[dest_length+1/*for terminator*/]; | |
| int status; | |
| if(source_charset.isUTF8()){ | |
| status=::escape((XMLByte *)src.str, src_length, dest_body, dest_length); | |
| } else { | |
| status=::escape((XMLByte *)src.str, src_length, dest_body, dest_length, source_charset.tables); | |
| } | |
| if(status<0) | |
| throw Exception(0, | |
| 0, | |
| "Charset::escapeString buffer overflow"); | |
| assert(dest_length<=saved_dest_length); | |
| dest_body[dest_length]=0; // terminator | |
| return String::C((char*)dest_body, dest_length); | |
| } | |
| const String::C Charset::transcodeToUTF8(const String::C src) const { | const String::C Charset::transcodeToUTF8(const String::C src) const { |
| size_t src_length=src.length; | size_t src_length=src.length; |
| size_t dest_length=src.length*6/*so that surly enough, max utf8 seq len=6*/; | |
| #ifdef PRECALCULATE_DEST_LENGTH | |
| size_t 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 | |
| size_t dest_length=src_length*6; // so that surly enough (max utf8 seq len=6) but too memory-hyngry | |
| #endif | |
| //throw Exception(0,0,"%u",dest_length); | |
| #ifndef NDEBUG | #ifndef NDEBUG |
| size_t saved_dest_length=dest_length; | size_t saved_dest_length=dest_length; |
| #endif | #endif |
| Line 430 const String::C Charset::transcodeToUTF8 | Line 650 const String::C Charset::transcodeToUTF8 |
| (XMLByte *)src.str, src_length, | (XMLByte *)src.str, src_length, |
| dest_body, dest_length, | dest_body, dest_length, |
| tables)<0) | tables)<0) |
| throw(0, 0, | throw Exception(0, |
| 0, | 0, |
| "Charset::transcodeToUTF8 buffer overflow"); | "Charset::transcodeToUTF8 buffer overflow"); |
| assert(dest_length<=saved_dest_length); dest_body[dest_length]=0; // terminator | assert(dest_length<=saved_dest_length); |
| dest_body[dest_length]=0; // terminator | |
| return String::C((char*)dest_body, dest_length); | 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; | |
| } | |
| // not found | |
| return src; | |
| } | |
| static void store_UTF8(XMLCh src, XMLByte*& outPtr){ | |
| if(!src) { | |
| // use the replacement character | |
| *outPtr++= '?'; | |
| return; | |
| } | |
| // Figure out how many bytes we need | |
| 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; | |
| } | |
| // 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((src | 0x80UL) & 0xBFUL); | |
| src>>= 6; | |
| 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 { | const String::C Charset::transcodeFromUTF8(const String::C src) const { |
| size_t src_length=src.length; | size_t src_length=src.length; |
| size_t dest_length=src.length*6/*so that surly enough, "ÿ" has max ratio */; | |
| #ifdef PRECALCULATE_DEST_LENGTH | |
| size_t 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)){ | |
| if(charSize==1) | |
| dest_length++; | |
| else | |
| dest_length+=(UTF8Char & 0xFFFF0000) | |
| ?charSize*3 // '%XX' for each byte | |
| :(xlatOneTo(UTF8Char, tables, 0)!=0) | |
| ?1 // can convert it to single char | |
| :getDecNumLength(UTF8Char)+3; // &#XX; - &#XXXXX; | |
| } | |
| #else | |
| // so that surly enough, "&#XXX;" has max ratio (huh? 8 bytes needed for '&#XXXXX;') | |
| size_t dest_length=src_length*6; | |
| #endif | |
| //throw Exception(0,0,"%u",dest_length); | |
| #ifndef NDEBUG | #ifndef NDEBUG |
| size_t saved_dest_length=dest_length; | size_t saved_dest_length=dest_length; |
| #endif | #endif |
| Line 449 const String::C Charset::transcodeFromUT | Line 834 const String::C Charset::transcodeFromUT |
| (XMLByte *)src.str, src_length, | (XMLByte *)src.str, src_length, |
| dest_body, dest_length, | dest_body, dest_length, |
| tables)<0) | tables)<0) |
| throw(0, 0, | throw Exception(0, |
| 0, | 0, |
| "Charset::transcodeFromUTF8 buffer overflow"); | "Charset::transcodeFromUTF8 buffer overflow"); |
| assert(dest_length<=saved_dest_length); dest_body[dest_length]=0; // terminator | assert(dest_length<=saved_dest_length); |
| dest_body[dest_length]=0; // terminator | |
| return String::C((char*)dest_body, dest_length); | return String::C((char*)dest_body, dest_length); |
| } | } |
| Line 480 const String::C Charset::transcodeToChar | Line 866 const String::C Charset::transcodeToChar |
| } | } |
| } | } |
| void Charset::store_Char(XMLByte*& outPtr, XMLCh src, XMLByte not_found){ | |
| if(isUTF8()) | |
| store_UTF8(src, outPtr); | |
| else if(char ch=xlatOneTo(src, tables, not_found)) | |
| *outPtr++=ch; | |
| } | |
| #ifdef XML | #ifdef XML |
| static const Charset::Tables* tables[MAX_CHARSETS]; | 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, *(size_t*)inlen, \ | |
| out, *(size_t*)outlen, \ | |
| *tables[i]); \ | |
| } \ | |
| static int xml256CharEncodingOutputFunc##i( \ | |
| unsigned char *out, int *outlen, \ | |
| const unsigned char *in, int *inlen, void*) { \ | |
| return transcodeFromUTF8( \ | |
| in, *(size_t*)inlen, \ | |
| out, *(size_t*)outlen, \ | |
| *tables[i]); \ | |
| } | |
| #else | |
| #define declareXml256ioFuncs(i) \ | #define declareXml256ioFuncs(i) \ |
| static int xml256CharEncodingInputFunc##i( \ | static int xml256CharEncodingInputFunc##i( \ |
| unsigned char *out, int *outlen, \ | unsigned char *out, int *outlen, \ |
| Line 502 static const Charset::Tables* tables[MAX | Line 917 static const Charset::Tables* tables[MAX |
| *tables[i]); \ | *tables[i]); \ |
| } | } |
| #endif | |
| declareXml256ioFuncs(0) declareXml256ioFuncs(1) | declareXml256ioFuncs(0) declareXml256ioFuncs(1) |
| declareXml256ioFuncs(2) declareXml256ioFuncs(3) | declareXml256ioFuncs(2) declareXml256ioFuncs(3) |
| declareXml256ioFuncs(4) declareXml256ioFuncs(5) | declareXml256ioFuncs(4) declareXml256ioFuncs(5) |
| Line 531 void Charset::addEncoding(char *name_cst | Line 949 void Charset::addEncoding(char *name_cst |
| "already allocated %d handlers, no space for new encoding '%s'", | "already allocated %d handlers, no space for new encoding '%s'", |
| MAX_CHARSETS, name_cstr); | MAX_CHARSETS, name_cstr); |
| xmlCharEncodingHandler* handler=new(PointerFreeGC) xmlCharEncodingHandler; | xmlCharEncodingHandler* handler=new(UseGC) xmlCharEncodingHandler; |
| { | { |
| handler->name=name_cstr; | handler->name=name_cstr; |
| handler->input=inputFuncs[handlers_count]; | handler->input=inputFuncs[handlers_count]; |
| Line 544 void Charset::addEncoding(char *name_cst | Line 962 void Charset::addEncoding(char *name_cst |
| } | } |
| void Charset::initTranscoder(const StringBody NAME, const char* name_cstr) { | void Charset::initTranscoder(const String::Body NAME, const char* name_cstr) { |
| ftranscoder=xmlFindCharEncodingHandler(name_cstr); | ftranscoder=xmlFindCharEncodingHandler(name_cstr); |
| transcoder(NAME); // check right way | transcoder(NAME); // check right way |
| } | } |
| xmlCharEncodingHandler& Charset::transcoder(const StringBody NAME) { | xmlCharEncodingHandler& Charset::transcoder(const String::Body NAME) { |
| if(!ftranscoder) | if(!ftranscoder) |
| throw Exception("parser.runtime", | throw Exception(PARSER_RUNTIME, |
| new String(NAME, String::L_TAINTED), | new String(NAME, String::L_TAINTED), |
| "unsupported encoding"); | "unsupported encoding"); |
| return *ftranscoder; | return *ftranscoder; |
| } | } |
| String::C Charset::transcode_cstr(xmlChar* s) { | String::C Charset::transcode_cstr(const xmlChar* s) { |
| if(!s) | if(!s) |
| return String::C("", 0); | return String::C("", 0); |
| int inlen=strlen((const char*)s); | int inlen=strlen((const char*)s); |
| int outlen=inlen; // max | int outlen=inlen*6/*strlen("ÿ")*/; // max |
| #ifndef NDEBUG | #ifndef NDEBUG |
| int saved_outlen=outlen; | int saved_outlen=outlen; |
| #endif | #endif |
| Line 572 String::C Charset::transcode_cstr(xmlCha | Line 990 String::C Charset::transcode_cstr(xmlCha |
| if(xmlCharEncodingOutputFunc output=transcoder(FNAME).output) { | if(xmlCharEncodingOutputFunc output=transcoder(FNAME).output) { |
| error=output( | error=output( |
| (unsigned char*)out, &outlen, | (unsigned char*)out, &outlen, |
| (const unsigned char*)s, &inlen); | (const unsigned char*)s, &inlen |
| #ifdef PA_PATCHED_LIBXML_BACKWARD | |
| ,0 | |
| #endif | |
| ); | |
| } else { | } else { |
| memcpy(out, s, outlen=inlen); | memcpy(out, s, outlen=inlen); |
| error=0; | error=0; |
| Line 585 String::C Charset::transcode_cstr(xmlCha | Line 1007 String::C Charset::transcode_cstr(xmlCha |
| assert(outlen<=saved_outlen); out[outlen]=0; | assert(outlen<=saved_outlen); out[outlen]=0; |
| return String::C(out, outlen); | return String::C(out, outlen); |
| } | } |
| const String& Charset::transcode(xmlChar* s) { | const String& Charset::transcode(const xmlChar* s) { |
| String::C cstr=transcode_cstr(s); | |
| return *new String(cstr.str, cstr.length, true); | |
| } | |
| String::C Charset::transcode_cstr(GdomeDOMString* s) { | |
| return s?transcode_cstr(BAD_CAST s->str) | |
| :String::C("", 0); | |
| } | |
| const String& Charset::transcode(GdomeDOMString* s) { | |
| String::C cstr=transcode_cstr(s); | String::C cstr=transcode_cstr(s); |
| return *new String(cstr.str, cstr.length, true); | return *new String(cstr.str, cstr.length, true); |
| } | } |
| Line 607 xmlChar* Charset::transcode_buf2xchar(co | Line 1021 xmlChar* Charset::transcode_buf2xchar(co |
| int saved_outlen; | int saved_outlen; |
| #endif | #endif |
| if(xmlCharEncodingInputFunc input=transcoder(FNAME).input) { | if(xmlCharEncodingInputFunc input=transcoder(FNAME).input) { |
| outlen=buf_size*6/*max*/; | outlen=buf_size*6/*max UTF8 bytes per char*/; |
| #ifndef NDEBUG | #ifndef NDEBUG |
| saved_outlen=outlen; | saved_outlen=outlen; |
| #endif | #endif |
| out=(xmlChar*)xmlMalloc(outlen+1); | out=(xmlChar*)xmlMalloc(outlen+1); |
| error=input( | error=input( |
| out, &outlen, | out, &outlen, |
| (const unsigned char*)buf, (int*)&buf_size); | (const unsigned char*)buf, (int*)&buf_size |
| #ifdef PA_PATCHED_LIBXML_BACKWARD | |
| ,0 | |
| #endif | |
| ); | |
| } else { | } else { |
| outlen=buf_size; | outlen=buf_size; |
| #ifndef NDEBUG | #ifndef NDEBUG |
| Line 633 xmlChar* Charset::transcode_buf2xchar(co | Line 1051 xmlChar* Charset::transcode_buf2xchar(co |
| assert(outlen<=saved_outlen); out[outlen]=0; | assert(outlen<=saved_outlen); out[outlen]=0; |
| return out; | return out; |
| } | } |
| GdomeDOMString_auto_ptr Charset::transcode_buf2dom(const char* buf, size_t buf_size) { | xmlChar* Charset::transcode(const String& s) { |
| return GdomeDOMString_auto_ptr(transcode_buf2xchar(buf, buf_size)); | |
| } | |
| GdomeDOMString_auto_ptr Charset::transcode(const String& s) { | |
| const char* cstr=s.cstr(String::L_UNSPECIFIED); | const char* cstr=s.cstr(String::L_UNSPECIFIED); |
| return transcode_buf2dom(cstr, strlen(cstr)); | return transcode_buf2xchar(cstr, strlen(cstr)); |
| } | } |
| GdomeDOMString_auto_ptr Charset::transcode(const StringBody s) { | xmlChar* Charset::transcode(const String::Body s) { |
| const char* cstr=s.cstr(); | const char* cstr=s.cstr(); |
| return transcode_buf2dom(cstr, s.length()); | return transcode_buf2xchar(cstr, s.length()); |
| } | } |
| #endif | |
| StringBody Charset::transcode(const StringBody src, | String::Body Charset::transcode(const String::Body src, |
| const Charset& source_transcoder, | const Charset& source_transcoder, |
| const Charset& dest_transcoder) { | const Charset& dest_transcoder) { |
| Line 658 StringBody Charset::transcode(const Stri | Line 1074 StringBody Charset::transcode(const Stri |
| source_transcoder, | source_transcoder, |
| dest_transcoder); | dest_transcoder); |
| return StringBody(dest.str, dest.length); | return String::Body(dest.str, dest.length); |
| } | } |
| String& Charset::transcode(const String& src, | String& Charset::transcode(const String& src, |
| Line 667 String& Charset::transcode(const String& | Line 1083 String& Charset::transcode(const String& |
| if(!src.length()) | if(!src.length()) |
| return *new String("", 0, false); | return *new String("", 0, false); |
| return *new String(transcode((StringBody)src, source_transcoder, dest_transcoder), String::L_CLEAN); | return *new String(transcode((String::Body)src, source_transcoder, dest_transcoder), String::L_CLEAN); |
| } | } |
| void Charset::transcode(ArrayString& src, | void Charset::transcode(ArrayString& src, |
| Line 683 struct Transcode_pair_info { | Line 1099 struct Transcode_pair_info { |
| const Charset* dest_transcoder; | const Charset* dest_transcoder; |
| }; | }; |
| #endif | #endif |
| static void transcode_pair(const StringBody akey, | static void transcode_pair(const String::Body /*akey*/, |
| StringBody& avalue, | String::Body& avalue, |
| Transcode_pair_info* info) { | Transcode_pair_info* info) { |
| avalue=Charset::transcode(avalue, | avalue=Charset::transcode(avalue, |
| *info->source_transcoder, | *info->source_transcoder, |
| Line 694 void Charset::transcode(HashStringString | Line 1110 void Charset::transcode(HashStringString |
| const Charset& source_transcoder, | const Charset& source_transcoder, |
| const Charset& dest_transcoder) { | const Charset& dest_transcoder) { |
| Transcode_pair_info info={&source_transcoder, &dest_transcoder}; | Transcode_pair_info info={&source_transcoder, &dest_transcoder}; |
| src.for_each_ref(transcode_pair, &info); | src.for_each_ref<Transcode_pair_info*>(transcode_pair, &info); |
| } | } |
| #endif |