--- parser3/src/main/pa_charset.C 2008/07/16 17:07:16 1.61 +++ parser3/src/main/pa_charset.C 2009/10/04 22:02:57 1.81 @@ -1,11 +1,11 @@ /** @file Parser: Charset connection implementation. - Copyright(c) 2001-2005 ArtLebedev Group (http://www.artlebedev.com) + Copyright(c) 2001-2009 ArtLebedev Group (http://www.artlebedev.com) Author: Alexander Petrosyan(http://paf.design.ru) */ -static const char * const IDENT_CHARSET_C="$Date: 2008/07/16 17:07:16 $"; +static const char * const IDENT_CHARSET_C="$Date: 2009/10/04 22:02:57 $"; #include "pa_charset.h" #include "pa_charsets.h" @@ -15,6 +15,8 @@ static const char * const IDENT_CHARSET_ #endif //#define PA_PATCHED_LIBXML_BACKWARD + +// reduce memory usage by pre-calculation utf-8 string length #define PRECALCULATE_DEST_LENGTH // globals @@ -86,7 +88,6 @@ static void element2case(unsigned char f // methods -extern "C" unsigned char pcre_default_tables[]; // pcre/chartables.c Charset::Charset(Request_charsets* charsets, const String::Body ANAME, const String* afile_spec): FNAME(ANAME), FNAME_CSTR(ANAME.cstrm()) { @@ -100,7 +101,7 @@ Charset::Charset(Request_charsets* chars } 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, _pcre_default_tables, sizeof(pcre_tables)); } #ifdef XML @@ -192,8 +193,8 @@ void Charset::sort_ToTable() { static XMLByte xlatOneTo(const XMLCh toXlat, const Charset::Tables& tables, XMLByte not_found) { - int lo = 0; - int hi = tables.toTableSize - 1; + int lo = 0; + int hi = tables.toTableSize - 1; while(lo<=hi) { // Calc the mid point of the low and high offset. const unsigned int i = (lo + hi) / 2; @@ -262,16 +263,16 @@ static const XMLByte gUTFBytes[0x100] = }; static const uint gUTFOffsets[6] = { - 0, 0x3080, 0xE2080, 0x3C82080, 0xFA082080, 0x82082080 + 0, 0x3080, 0xE2080, 0x3C82080, 0xFA082080, 0x82082080 }; static const XMLByte gFirstByteMark[7] = { - 0x00, 0x00, 0xC0, 0xE0, 0xF0, 0xF8, 0xFC + 0x00, 0x00, 0xC0, 0xE0, 0xF0, 0xF8, 0xFC }; -static int transcodeToUTF8(const XMLByte* srcData, size_t& srcLen, - XMLByte *toFill, size_t& 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; @@ -352,9 +353,9 @@ 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, size_t& srcLen, - XMLByte* toFill, size_t& 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; @@ -432,19 +433,21 @@ of ocetes consumed. return 0; } -static bool is_escaped(char c){ +static bool isEscaped(XMLByte c){ return - !(c<=127 - && ( - ((c>='0') && (c<='9')) - || ((c>='A') && (c<='Z')) - || ((c>='a') && (c<='z')) - || strchr("*@-_+./", c)!=0 - )); + !( + (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){ +// read one UTF8 char and return length of this char (in bytes) +static unsigned int readUTF8Char(const XMLByte*& srcPtr, const XMLByte* srcEnd, XMLByte& firstByte, XMLCh& UTF8Char){ if(!srcPtr || !*srcPtr || srcPtr>=srcEnd) return 0; @@ -478,13 +481,18 @@ static unsigned int readChar(const XMLBy return trailingBytes+1; } -static unsigned int readChar(const XMLByte*& srcPtr, const XMLByte* srcEnd){ - XMLByte firstByte; - XMLCh UTF8Char; - return readChar(srcPtr, srcEnd, firstByte, UTF8Char); +// 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 char, return number of bytes needed for store it as UTF8 +// read non-UTF8 char, and return number of bytes needed for store 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; @@ -510,27 +518,21 @@ static unsigned int readChar(const XMLBy return 1; } -static int escape(const XMLByte* srcData, size_t& srcLen, - XMLByte* toFill, size_t& toFillLen) { +static int escapeUTF8(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); + for(UTF8_string_iterator i((XMLByte *)srcPtr, srcLen); (outPtr < outEnd) && i.has_next(); ){ + if(i.getCharSize()==1){ + if(isEscaped(i.getFirstByte())) // %XX + outPtr+=sprintf((char*)outPtr, "%%%02X", i.getFirstByte()); else - *outPtr++=firstByte; + *outPtr++=i.getFirstByte(); } else - outPtr+=sprintf((char*)outPtr, "%%u%04X", UTF8Char); // %uXXXX + outPtr+=sprintf((char*)outPtr, "%%u%04X", i.next()); // %uXXXX } - // Update the bytes eaten srcLen=srcPtr-srcData; @@ -546,7 +548,6 @@ static int escape(const XMLByte* srcData const XMLByte* srcPtr=srcData; const XMLByte* srcEnd=srcData+srcLen; XMLByte* outPtr=toFill; - //XMLByte* outEnd=toFill+toFillLen; XMLByte firstByte; XMLCh UTF8Char; uint charSize; @@ -554,7 +555,7 @@ static int escape(const XMLByte* srcData while(charSize=readChar(srcPtr, srcEnd, firstByte, UTF8Char, tables)){ if(charSize==1){ if(firstByte){ - if(is_escaped(firstByte)) // %XX + if(isEscaped(firstByte)) // %XX outPtr+=sprintf((char*)outPtr, "%%%02X", firstByte); else *outPtr++=firstByte; @@ -581,22 +582,22 @@ String::C Charset::escape(const String:: #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*/; + for(UTF8_string_iterator i((XMLByte *)src.str, src_length); i.has_next(); ){ + if(i.getCharSize()==1) + dest_length+=!isEscaped(i.getFirstByte())?1/*as-is*/:3/*%XX*/; else - dest_length+=6; // '%uXXXX' + dest_length+=6/*%uXXXX*/; } } else { + const XMLByte* srcPtr=(XMLByte*)src.str; + const XMLByte* srcEnd=srcPtr+src_length; + XMLByte firstByte; + XMLCh UTF8Char; 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'*/; + dest_length+=(!firstByte/*replacement char '?'*/ || !isEscaped(firstByte))?1:3/*'%XX'*/; else dest_length+=6; // '%uXXXX' } @@ -614,7 +615,7 @@ String::C Charset::escape(const String:: int status; if(source_charset.isUTF8()){ - status=::escape((XMLByte *)src.str, src_length, dest_body, dest_length); + status=::escapeUTF8((XMLByte *)src.str, src_length, dest_body, dest_length); } else { status=::escape((XMLByte *)src.str, src_length, dest_body, dest_length, source_charset.tables); } @@ -622,33 +623,48 @@ String::C Charset::escape(const String:: if(status<0) throw Exception(0, 0, - "Charset::escapeString buffer overflow"); + "Charset::escape buffer overflow"); assert(dest_length<=saved_dest_length); 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=strlen(src_ptr); + + 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); +} const String::C Charset::transcodeToUTF8(const String::C src) const { - size_t src_length=src.length; + int src_length=src.length; #ifdef PRECALCULATE_DEST_LENGTH - size_t dest_length=0; + int dest_length=0; const XMLByte* srcPtr=(XMLByte*)src.str; const XMLByte* srcEnd=srcPtr+src_length; - XMLByte firstByte; - XMLCh UTF8Char; + 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 + int 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 - size_t saved_dest_length=dest_length; + int saved_dest_length=dest_length; #endif XMLByte *dest_body=new(PointerFreeGC) XMLByte[dest_length+1/*for terminator*/]; @@ -666,8 +682,8 @@ const String::C Charset::transcodeToUTF8 } static XMLCh change_case_UTF8(const XMLCh src, const Charset::UTF8CaseTable& table) { - int lo = 0; - int hi = table.size - 1; + 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; @@ -806,38 +822,30 @@ static size_t getDecNumLength(XMLCh UTF8 } const String::C Charset::transcodeFromUTF8(const String::C src) const { - size_t src_length=src.length; - #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) + 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+=(UTF8Char & 0xFFFF0000) - ?charSize*3 // '%XX' for each byte - :(xlatOneTo(UTF8Char, tables, 0)!=0) + 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(UTF8Char)+3; // &#XX; - &#XXXXX; + :getDecNumLength(i.next())+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; + int dest_length=src.length*6; #endif - //throw Exception(0,0,"%u",dest_length); - #ifndef NDEBUG - size_t saved_dest_length=dest_length; + 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, + (XMLByte *)src.str, (int&)src.length, dest_body, dest_length, tables)<0) throw Exception(0, @@ -851,7 +859,7 @@ const String::C Charset::transcodeFromUT /// transcode using both charsets const String::C Charset::transcodeToCharset(const String::C src, - const Charset& dest_charset) const { + const Charset& dest_charset) const { if(&dest_charset==this) return src; else { @@ -890,16 +898,16 @@ static const Charset::Tables* tables[MAX unsigned char *out, int *outlen, \ const unsigned char *in, int *inlen, void*) { \ return transcodeToUTF8( \ - in, *(size_t*)inlen, \ - out, *(size_t*)outlen, \ + 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, *(size_t*)inlen, \ - out, *(size_t*)outlen, \ + in, *inlen, \ + out, *outlen, \ *tables[i]); \ } @@ -910,16 +918,16 @@ static const Charset::Tables* tables[MAX unsigned char *out, int *outlen, \ const unsigned char *in, int *inlen) { \ return transcodeToUTF8( \ - in, *(size_t*)inlen, \ - out, *(size_t*)outlen, \ + in, *inlen, \ + out, *outlen, \ *tables[i]); \ } \ static int xml256CharEncodingOutputFunc##i( \ unsigned char *out, int *outlen, \ const unsigned char *in, int *inlen) { \ return transcodeFromUTF8( \ - in, *(size_t*)inlen, \ - out, *(size_t*)outlen, \ + in, *inlen, \ + out, *outlen, \ *tables[i]); \ } @@ -1015,7 +1023,7 @@ String::C Charset::transcode_cstr(const } const String& Charset::transcode(const xmlChar* s) { String::C cstr=transcode_cstr(s); - return *new String(cstr.str, cstr.length, true); + return *new String(cstr.str, String::L_TAINTED); } /// @test less memory using -maybe- xmlParserInputBufferCreateMem @@ -1057,15 +1065,14 @@ xmlChar* Charset::transcode_buf2xchar(co assert(outlen<=saved_outlen); out[outlen]=0; return out; } -xmlChar* Charset::transcode(const String& s) { - const char* cstr=s.cstr(String::L_UNSPECIFIED); - return transcode_buf2xchar(cstr, strlen(cstr)); +xmlChar* Charset::transcode(const String& s) { + String::Body sbody=s.cstr_to_string_body_untaint(String::L_AS_IS); + return transcode_buf2xchar(sbody.cstr(), sbody.length()); } -xmlChar* Charset::transcode(const String::Body s) { - const char* cstr=s.cstr(); - return transcode_buf2xchar(cstr, s.length()); +xmlChar* Charset::transcode(const String::Body s) { + return transcode_buf2xchar(s.cstr(), s.length()); } #endif @@ -1076,18 +1083,16 @@ String::Body Charset::transcode(const St const char *src_ptr=src.cstr(); size_t src_size=strlen(src_ptr); - String::C dest=Charset::transcode(String::C(src_ptr, src_size), - source_transcoder, - dest_transcoder); + String::C dest=Charset::transcode(String::C(src_ptr, src_size), source_transcoder, dest_transcoder); - return String::Body(dest.str, dest.length); + return String::Body(dest.length ? dest.str:0); } String& Charset::transcode(const String& src, const Charset& source_transcoder, const Charset& dest_transcoder) { - if(!src.length()) - return *new String("", 0, false); + if(src.is_empty()) + return *new String(); return *new String(transcode((String::Body)src, source_transcoder, dest_transcoder), String::L_CLEAN); } @@ -1105,7 +1110,7 @@ struct Transcode_pair_info { const Charset* dest_transcoder; }; #endif -static void transcode_pair(const String::Body /*akey*/, +static void transcode_pair(HashStringValue::key_type /*akey*/, String::Body& avalue, Transcode_pair_info* info) { avalue=Charset::transcode(avalue, @@ -1122,7 +1127,7 @@ void Charset::transcode(HashStringString size_t getUTF8BytePos(const XMLByte* srcBegin, const XMLByte* srcEnd, size_t charPos){ const XMLByte* ptr=srcBegin; - while(charPos-- && readChar(ptr, srcEnd)); + while(charPos-- && skipUTF8Char(ptr, srcEnd)); return ptr-srcBegin; } @@ -1131,7 +1136,7 @@ size_t getUTF8CharPos(const XMLByte* src size_t charPos=0; const XMLByte* ptr=srcBegin; const XMLByte* ptrEnd=srcBegin+bytePos; - while(readChar(ptr, srcEnd)){ + while(skipUTF8Char(ptr, srcEnd)){ if(ptr>ptrEnd) return charPos; charPos++; @@ -1145,8 +1150,13 @@ size_t getUTF8CharPos(const XMLByte* src size_t lengthUTF8(const XMLByte* srcBegin, const XMLByte* srcEnd){ size_t size=0; - while(readChar(srcBegin, srcEnd)) + while(skipUTF8Char(srcBegin, srcEnd)) size++; return size; } + +bool UTF8_string_iterator::has_next(){ + fcharSize=readUTF8Char(fsrcPtr, fsrcEnd, ffirstByte, fUTF8Char); + return fcharSize!=0; +}