--- parser3/src/main/pa_charset.C 2012/03/16 09:24:13 1.90 +++ parser3/src/main/pa_charset.C 2025/06/28 15:42:00 1.117 @@ -1,21 +1,23 @@ /** @file Parser: Charset connection implementation. - Copyright (c) 2001-2012 Art. Lebedev Studio (http://www.artlebedev.com) - Author: Alexander Petrosyan(http://paf.design.ru) + Copyright (c) 2001-2024 Art. Lebedev Studio (http://www.artlebedev.com) + Authors: Konstantin Morshnev , Alexandr Petrosian */ #include "pa_charset.h" #include "pa_charsets.h" -volatile const char * IDENT_PA_CHARSET_C="$Id: pa_charset.C,v 1.90 2012/03/16 09:24:13 moko Exp $" IDENT_PA_CHARSET_H; +// we are using some pcre_internal.h stuff as well +#include "../lib/pcre/pa_pcre_internal.h" + +volatile const char * IDENT_PA_CHARSET_C="$Id: pa_charset.C,v 1.117 2025/06/28 15:42:00 moko Exp $" IDENT_PA_CHARSET_H; #ifdef XML +#include "libxml/xmlmemory.h" #include "libxml/encoding.h" #endif -//#define PA_PATCHED_LIBXML_BACKWARD - // reduce memory usage by pre-calculation utf-8 string length #define PRECALCULATE_DEST_LENGTH @@ -43,8 +45,7 @@ inline void prepare_case_tables(unsigned for(int i=0; i<0x100; i++) lcc_table[i]=fcc_table[i]=(unsigned char)i; } -inline void cstr2ctypes(unsigned char *tables, const unsigned char *cstr, - unsigned char bit) { +inline void cstr2ctypes(unsigned char *tables, const unsigned char *cstr, unsigned char bit) { unsigned char *ctypes_table=tables+ctypes_offset; ctypes_table[0]=bit; for(; *cstr; cstr++) { @@ -58,14 +59,12 @@ inline unsigned int to_wchar_code(const if(cstr[1]==0) return(unsigned int)(unsigned char)cstr[0]; - char *error_pos; - return(unsigned int)strtol(cstr, &error_pos, 0); + return pa_atoui(cstr,0); } inline bool to_bool(const char* cstr) { return cstr && *cstr!=0; } -static void element2ctypes(unsigned char c, bool belongs, - unsigned char *tables, unsigned char bit, int group_offset=-1) { +static void element2ctypes(unsigned char c, bool belongs, unsigned char *tables, unsigned char bit, int group_offset=-1) { if(!belongs) return; @@ -75,8 +74,7 @@ static void element2ctypes(unsigned char if(group_offset>=0) tables[cbits_offset+group_offset+c/8] |= 1<<(c%8); } -static void element2case(unsigned char from, unsigned char to, - unsigned char *tables) { +static void element2case(unsigned char from, unsigned char to, unsigned char *tables) { if(!to) return; @@ -86,30 +84,48 @@ static void element2case(unsigned char f fcc_table[from]=to; fcc_table[to]=from; } +inline XMLByte *append_hex_8(XMLByte *dest, unsigned char c, const char* prefix=0) { + if(prefix) { + strcpy((char *)dest, prefix); + dest+=strlen(prefix); + } + *dest++=hex_digits[c >> 4]; + *dest++=hex_digits[c & 0x0F]; + return dest; +} + +inline XMLByte *append_hex_16(XMLByte *dest, unsigned int c, const char* prefix=0) { + if(prefix) { + strcpy((char *)dest, prefix); + dest+=strlen(prefix); + } + *dest++=hex_digits[(c >> 12) & 0x0F]; + *dest++=hex_digits[(c >> 8) & 0x0F]; + *dest++=hex_digits[(c >> 4) & 0x0F]; + *dest++=hex_digits[(c) & 0x0F]; + return dest; +} + // methods -Charset::Charset(Request_charsets* charsets, const String::Body ANAME, const String* afile_spec): +Charset::Charset(Request_charsets* acharsets, const String::Body ANAME, const String* afile_spec): FNAME(ANAME), FNAME_CSTR(ANAME.cstrm()) { if(afile_spec) { fisUTF8=false; - load_definition(*charsets, *afile_spec); + load_definition(*acharsets, *afile_spec); #ifdef XML addEncoding(FNAME_CSTR); #endif } else { fisUTF8=true; // 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, pa_pcre_default_tables, sizeof(pcre_tables)); } - -#ifdef XML - initTranscoder(FNAME, FNAME_CSTR); -#endif } -void Charset::load_definition(Request_charsets& charsets, const String& afile_spec) { +void Charset::load_definition(Request_charsets& acharsets, const String& afile_spec) { // pcre_tables // lowcase, flipcase, bits digit+word+whitespace, masks @@ -123,7 +139,7 @@ void Charset::load_definition(Request_ch memset(&tables, 0, sizeof(tables)); // loading text - char *data=file_read_text(charsets, afile_spec); + char *data=file_read_text(acharsets, afile_spec); // ignore header getrow(&data); @@ -152,9 +168,7 @@ void Charset::load_definition(Request_ch case 8: // charset if(tables.toTableSize>MAX_CHARSET_UNI_CODES) - throw Exception(PARSER_RUNTIME, - &afile_spec, - "charset must contain not more then %d unicode values", MAX_CHARSET_UNI_CODES); + throw Exception(PARSER_RUNTIME, &afile_spec, "charset must contain not more than %d unicode values", MAX_CHARSET_UNI_CODES); XMLCh unicode=(XMLCh)to_wchar_code(cell); if(!unicode && column==7/*unicode1 column*/) @@ -191,17 +205,11 @@ static int sort_cmp_Trans_rec_intCh(cons } void Charset::sort_ToTable() { - _qsort(tables.toTable, tables.toTableSize, sizeof(*tables.toTable), - sort_cmp_Trans_rec_intCh); - //FILE *f=fopen("c:\\temp\\a", "wb"); - //fwrite(tables.toTable, tables.toTableSize, sizeof(*tables.toTable), f); - //fclose(f); + qsort(tables.toTable, tables.toTableSize, sizeof(*tables.toTable), sort_cmp_Trans_rec_intCh); } // @todo: precache for spedup searching -static XMLByte xlatOneTo(const XMLCh toXlat, - const Charset::Tables& tables, - XMLByte not_found) { +static XMLByte xlatOneTo(const XMLCh toXlat, const Charset::Tables& tables, XMLByte not_found) { int lo = 0; int hi = tables.toTableSize - 1; while(lo<=hi) { @@ -220,9 +228,7 @@ static XMLByte xlatOneTo(const XMLCh toX return not_found; } -String::C Charset::transcode(const String::C src, - const Charset& source_charset, - const Charset& dest_charset) { +String::C Charset::transcode(const String::C src, const Charset& source_charset, const Charset& dest_charset) { if(!src.length) return String::C("", 0); @@ -279,9 +285,7 @@ static const XMLByte gFirstByteMark[7] = 0x00, 0x00, 0xC0, 0xE0, 0xF0, 0xF8, 0xFC }; -static int transcodeToUTF8(const XMLByte* srcData, int& srcLen, - XMLByte *toFill, int& toFillLen, - const Charset::Tables& tables) { +static int transcodeToUTF8(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; @@ -362,9 +366,7 @@ 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) { +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; @@ -403,9 +405,7 @@ static int transcodeFromUTF8(const XMLBy break; default: - throw Exception(0, - 0, - "transcodeFromUTF8 error: wrong trailingBytes value(%d)", trailingBytes); // never + throw Exception(0, 0, "transcodeFromUTF8 error: wrong trailingBytes value(%d)", trailingBytes); // never } tmpVal-=gUTFOffsets[trailingBytes]; @@ -572,13 +572,13 @@ size_t Charset::calc_escaped_length(cons if(char_size==1) \ if(first_byte){ \ if(need_escape(first_byte)) \ - dest_ptr+=sprintf((char*)dest_ptr, "%%%02X", first_byte); /* %XX */ \ + dest_ptr=append_hex_8(dest_ptr, 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 + dest_ptr=append_hex_16(dest_ptr, UTF8_char, "%u"); /* %uXXXX */ size_t Charset::escape_UTF8(const XMLByte* src, size_t src_length, XMLByte* dest) { @@ -645,9 +645,10 @@ size_t Charset::calc_JSON_escaped_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 + if(i.getCharSize()==1){ + XMLByte first_byte=i.getFirstByte(); + dest_length+=need_json_escape(first_byte) ? 2 : (first_byte < 0x20 && first_byte /* 0 replacement char is '?' */) ? 6 : 1; + } else dest_length+=6; // \uXXXX } @@ -662,7 +663,7 @@ size_t Charset::calc_JSON_escaped_length while(uint char_size=readChar(src, src_end, first_byte, UTF8_char, tables)){ if(char_size==1) - dest_length+=(!first_byte/*replacement char '?'*/ || !need_json_escape(first_byte))? 1 : 2; + dest_length+=need_json_escape(first_byte) ? 2 : (first_byte < 0x20 && first_byte /* 0 replacement char is '?' */) ? 6 : 1; else dest_length+=6; // \uXXXX } @@ -696,10 +697,11 @@ size_t Charset::calc_JSON_escaped_length 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; \ + default : if(first_byte < 0x20) dest_ptr=append_hex_16(dest_ptr, UTF8_char, "\\u"); \ + else *dest_ptr++=first_byte; \ } \ else \ - dest_ptr+=sprintf((char*)dest_ptr, "\\u%04X", UTF8_char); // \uXXXX + dest_ptr=append_hex_16(dest_ptr, UTF8_char, "\\u"); // \uXXXX size_t Charset::escape_JSON_UTF8(const XMLByte* src, size_t src_length, XMLByte* dest) { @@ -782,9 +784,7 @@ const String::C Charset::transcodeToUTF8 (XMLByte *)src.str, src_length, dest_body, dest_length, tables)<0) - throw Exception(0, - 0, - "Charset::transcodeToUTF8 buffer overflow"); + throw Exception(0, 0, "Charset::transcodeToUTF8 buffer overflow"); assert(dest_length<=saved_dest_length); dest_body[dest_length]=0; // terminator @@ -859,13 +859,11 @@ static void store_UTF8(XMLCh src, XMLByt outPtr+= encodedBytes; } -static void change_case_UTF8(XMLCh src, XMLByte*& outPtr, - const Charset::UTF8CaseTable& table) { +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) { +} + +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; @@ -897,9 +895,7 @@ void change_case_UTF8(const XMLByte* src break; default: - throw Exception(0, - 0, - "change_case_UTF8 error: wrong trailingBytes value(%d)", trailingBytes); + throw Exception(0, 0, "change_case_UTF8 error: wrong trailingBytes value(%d)", trailingBytes); } tmpVal-=gUTFOffsets[trailingBytes]; @@ -909,15 +905,11 @@ void change_case_UTF8(const XMLByte* src if(!(tmpVal & 0xFFFF0000)) change_case_UTF8(tmpVal, outPtr, table); else - throw Exception(0, - 0, - "change_case_UTF8 error: too big tmpVal(0x%08X)", tmpVal); + 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"); + throw Exception(0, 0, "change_case_UTF8 error: end pointers do not match"); } static size_t getDecNumLength(XMLCh UTF8Char){ @@ -956,9 +948,7 @@ const String::C Charset::transcodeFromUT (XMLByte *)src.str, src_length, dest_body, dest_length, tables)<0) - throw Exception(0, - 0, - "Charset::transcodeFromUTF8 buffer overflow"); + throw Exception(0, 0, "Charset::transcodeFromUTF8 buffer overflow"); assert(dest_length<=saved_dest_length); dest_body[dest_length]=0; // terminator @@ -966,8 +956,7 @@ const String::C Charset::transcodeFromUT } /// transcode using both charsets -const String::C Charset::transcodeToCharset(const String::C src, - const Charset& dest_charset) const { +const String::C Charset::transcodeToCharset(const String::C src, const Charset& dest_charset) const { if(&dest_charset==this) return src; else { @@ -999,28 +988,6 @@ void Charset::store_Char(XMLByte*& outPt 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, \ @@ -1039,9 +1006,6 @@ static const Charset::Tables* tables[MAX *tables[i]); \ } -#endif - - declareXml256ioFuncs(0) declareXml256ioFuncs(1) declareXml256ioFuncs(2) declareXml256ioFuncs(3) declareXml256ioFuncs(4) declareXml256ioFuncs(5) @@ -1066,35 +1030,16 @@ static size_t handlers_count=0; void Charset::addEncoding(char *name_cstr) { if(handlers_count==MAX_CHARSETS) - throw Exception(0, - 0, - "already allocated %d handlers, no space for new encoding '%s'", - 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); - -} - -void Charset::initTranscoder(const String::Body NAME, const char* name_cstr) { - ftranscoder=xmlFindCharEncodingHandler(name_cstr); - transcoder(NAME); // check right way -} + throw Exception(0, 0, "already allocated %d handlers, no space for new encoding '%s'", MAX_CHARSETS, name_cstr); -xmlCharEncodingHandler& Charset::transcoder(const String::Body NAME) { - if(!ftranscoder) - throw Exception(PARSER_RUNTIME, - new String(NAME, String::L_TAINTED), - "unsupported encoding"); - return *ftranscoder; + ftranscoder_input=inputFuncs[handlers_count]; + ftranscoder_output=outputFuncs[handlers_count]; + ::tables[handlers_count++]=&tables; + + xmlCharEncodingHandler* handler=xmlNewCharEncodingHandler(name_cstr, ftranscoder_input, ftranscoder_output); + if(!handler) + throw Exception(0, new String(name_cstr, String::L_TAINTED), "unable to register XML encoding handler"); + xmlRegisterCharEncodingHandler(handler); } String::C Charset::transcode_cstr(const xmlChar* s) { @@ -1109,22 +1054,14 @@ String::C Charset::transcode_cstr(const char *out=new(PointerFreeGC) char[outlen+1]; int error; - if(xmlCharEncodingOutputFunc output=transcoder(FNAME).output) { - error=output( - (unsigned char*)out, &outlen, - (const unsigned char*)s, &inlen -#ifdef PA_PATCHED_LIBXML_BACKWARD - ,0 -#endif - ); + if(!fisUTF8) { + error=ftranscoder_output((unsigned char*)out, &outlen, (const unsigned char*)s, &inlen); } else { memcpy(out, s, outlen=inlen); error=0; } if(error<0) - throw Exception(0, - 0, - "transcode_cstr failed (%d)", error); + throw Exception(0, 0, "transcode_cstr failed (%d)", error); assert(outlen<=saved_outlen); out[outlen]=0; return String::C(out, outlen); @@ -1142,19 +1079,13 @@ xmlChar* Charset::transcode_buf2xchar(co #ifndef NDEBUG int saved_outlen; #endif - if(xmlCharEncodingInputFunc input=transcoder(FNAME).input) { + if(!fisUTF8) { 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 - ); + error=ftranscoder_input(out, &outlen, (const unsigned char*)buf, (int*)&buf_size); } else { outlen=buf_size; #ifndef NDEBUG @@ -1166,9 +1097,7 @@ xmlChar* Charset::transcode_buf2xchar(co } if(error<0) - throw Exception(0, - 0, - "transcode_buf failed (%d)", error); + throw Exception(0, 0, "transcode_buf failed (%d)", error); assert(outlen<=saved_outlen); out[outlen]=0; return out; @@ -1184,30 +1113,18 @@ xmlChar* Charset::transcode(const String } #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::Body Charset::transcode(const String::Body src, const Charset& source_transcoder, const Charset& dest_transcoder) { + return String::Body(Charset::transcode(String::C(src.cstr(), src.length()), source_transcoder, dest_transcoder)); } -String& Charset::transcode(const String& src, - const Charset& source_transcoder, - const Charset& dest_transcoder) { +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) { +void Charset::transcode(ArrayString& src, const Charset& source_transcoder, const Charset& dest_transcoder) { for(size_t i=0; isource_transcoder, - *info->dest_transcoder); +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) { +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); } @@ -1251,9 +1162,7 @@ size_t getUTF8CharPos(const XMLByte* src } // scan till end but position in bytes still too low - throw Exception(0, - 0, - "Error convertion byte pos to char pos"); + throw Exception(0, 0, "Error conversion byte pos to char pos"); } size_t lengthUTF8(const XMLByte* srcBegin, const XMLByte* srcEnd){ @@ -1268,6 +1177,45 @@ unsigned int lengthUTF8Char(const XMLByt return gUTFBytes[c]+1; } +const char *fixUTF8(const char *src){ + if(src && *src){ + size_t length=strlen(src); + + int error_offset; + if(pa_pcre_valid_utf((unsigned char *)src, length, &error_offset)){ + + char *result=(char *)pa_malloc_atomic(length+1); + char *dst=result; + + do { + + if(error_offset){ + memcpy(dst, src, error_offset); + dst+=error_offset; + + src+=error_offset; + length-=error_offset; + + } + + *dst++='?'; + src++; + length--; + + } while (length && pa_pcre_valid_utf((unsigned char *)src, length, &error_offset)); + + if(length){ + strcpy(dst, src); + } else { + *dst='\0'; + } + + return result; + } + } + return src; +} + bool UTF8_string_iterator::has_next(){ fcharSize=readUTF8Char(fsrcPtr, fsrcEnd, ffirstByte, fUTF8Char); return fcharSize!=0;