--- parser3/src/main/pa_charset.C 2003/03/06 08:21:09 1.33.2.13 +++ parser3/src/main/pa_charset.C 2003/09/25 09:15:03 1.37 @@ -5,7 +5,7 @@ Author: Alexander Petrosyan(http://paf.design.ru) */ -static const char* IDENT_CHARSET_C="$Date: 2003/03/06 08:21:09 $"; +static const char* IDENT_CHARSET_C="$Date: 2003/09/25 09:15:03 $"; #include "pa_charset.h" #include "pa_charsets.h" @@ -68,18 +68,15 @@ static void element2case(unsigned char f // methods extern "C" unsigned char pcre_default_tables[]; // pcre/chartables.c -Charset::Charset(StringPtr aname, StringPtr afile_spec): - fname(aname), - fname_cstr(aname->cstr()) { - - for(char *c=fname_cstr; *c; c++) - *c = toupper(*c); +Charset::Charset(Request_charsets* charsets, const String::Body ANAME, const String* afile_spec): + FNAME(ANAME), + FNAME_CSTR(ANAME.cstrm()) { if(afile_spec) { fisUTF8=false; - load_definition(afile_spec); + load_definition(*charsets, *afile_spec); #ifdef XML - addEncoding(fname_cstr); + addEncoding(FNAME_CSTR); #endif } else { fisUTF8=true; @@ -88,18 +85,11 @@ Charset::Charset(StringPtr aname, String } #ifdef XML - initTranscoder(fname, fname_cstr); -#endif -} - -Charset::~Charset() { - // @todonow unregister encodings -#ifdef XML - // not deleting transcoder, that's not our business + initTranscoder(FNAME, FNAME_CSTR); #endif } -void Charset::load_definition(StringPtr afile_spec) { +void Charset::load_definition(Request_charsets& charsets, const String& afile_spec) { // pcre_tables // lowcase, flipcase, bits digit+word+whitespace, masks @@ -117,8 +107,7 @@ void Charset::load_definition(StringPtr tables.toTableSize++; // loading text - Pool pool_for_load_only; - char *data=file_read_text(pool_for_load_only, *UTF8_charset, afile_spec); + char *data=file_read_text(charsets, afile_spec); // ignore header getrow(&data); @@ -148,7 +137,7 @@ void Charset::load_definition(StringPtr // charset if(tables.toTableSize>MAX_CHARSET_UNI_CODES) throw Exception("parser.runtime", - afile_spec, + &afile_spec, "charset must contain not more then %d unicode values", MAX_CHARSET_UNI_CODES); XMLCh unicode=(XMLCh)to_wchar_code(cell); @@ -185,59 +174,44 @@ void Charset::sort_ToTable() { } static XMLByte xlatOneTo(const XMLCh toXlat, - const Charset::Tables& tables, - XMLByte not_found) { - unsigned int lowOfs = 0; - unsigned int hiOfs = tables.toTableSize - 1; - XMLByte curByte = 0; - do { - // Calc the mid point of the low and high offset. - const unsigned int midOfs =((hiOfs - lowOfs) / 2)+lowOfs; - - // If our test char is greater than the mid point char, then - // we move up to the upper half. Else we move to the lower - // half. If its equal, then its our guy. - if(toXlat>tables.toTable[midOfs].intCh) - lowOfs = midOfs; + const Charset::Tables& tables, + XMLByte not_found) { + unsigned int lowOfs = 0; + unsigned int hiOfs = tables.toTableSize - 1; + XMLByte curByte = 0; + do { + // Calc the mid point of the low and high offset. + const unsigned int midOfs =((hiOfs - lowOfs) / 2)+lowOfs; + + // If our test char is greater than the mid point char, then + // we move up to the upper half. Else we move to the lower + // half. If its equal, then its our guy. + if(toXlat>tables.toTable[midOfs].intCh) + lowOfs = midOfs; else if(toXlat 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; + // use the replacement character + *outPtr++= '?'; + srcPtr++; + continue; + } - // Return the characters read - toFillLen = outPtr - toFill; + // 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; + } + + // Update the bytes eaten + srcLen = srcPtr - srcData; + + // Return the characters read + toFillLen = outPtr - toFill; + //return srcPtr==srcEnd?(int)toFillLen:-1; /* xmlCharEncodingInputFunc @@ -366,73 +339,72 @@ 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, 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; - // 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, - Exception::undefined_source, - "transcodeFromUTF8 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)) { + // 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, + "transcodeFromUTF8 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)) { if(XMLByte xlat=xlatOneTo(tmpVal, tables, 0)) *outPtr++=xlat; else outPtr+=sprintf((char *)outPtr, "&#%d;", tmpVal); // &#decimal; } else throw Exception(0, - Exception::undefined_source, - "transcodeFromUTF8 error: too big tmpVal(0x%08X)", tmpVal); + 0, + "transcodeFromUTF8 error: too big tmpVal(0x%08X)", tmpVal); } - - // Update the bytes eaten - srcLen = srcPtr - srcData; - - // Return the characters read - toFillLen = outPtr - toFill; + + // Update the bytes eaten + srcLen = srcPtr - srcData; + + // Return the characters read + toFillLen = outPtr - toFill; //return srcPtr==srcEnd?(int)toFillLen:-1; /* @@ -446,134 +418,161 @@ of ocetes consumed. } /// @todo not so memory-hungry with prescan -void Charset::transcodeToUTF8(Pool& pool, - const void *source_body, size_t source_content_length, - const void *& adest_body, size_t& dest_content_length) const { - dest_content_length=source_content_length*6/*so that surly enough, max utf8 seq len=6*/; - XMLByte *dest_body=new(pool) XMLByte[dest_content_length]; +const String::C Charset::transcodeToUTF8(const String::C src) const { + size_t src_length=src.length; + size_t dest_length=src.length*6/*so that surly enough, max utf8 seq len=6*/; +#ifndef NDEBUG + size_t saved_dest_length=dest_length; +#endif + XMLByte *dest_body=new(PointerFreeGC) XMLByte[dest_length+1/*for terminator*/]; if(::transcodeToUTF8( - (XMLByte *)source_body, source_content_length, - dest_body, dest_content_length, + (XMLByte *)src.str, src_length, + dest_body, dest_length, tables)<0) throw(0, 0, 0, "Charset::transcodeToUTF8 buffer overflow"); - // return - adest_body=dest_body; + assert(dest_length<=saved_dest_length); dest_body[dest_length]=0; // terminator + return String::C((char*)dest_body, dest_length); } -void Charset::transcodeFromUTF8(Pool& pool, - const void *source_body, size_t source_content_length, - const void *& adest_body, size_t& dest_content_length) const { - dest_content_length=source_content_length*6/*so that surly enough, "ÿ" has max ratio */; - XMLByte *dest_body=new(pool) XMLByte[dest_content_length]; +const String::C Charset::transcodeFromUTF8(const String::C src) const { + size_t src_length=src.length; + size_t dest_length=src.length*6/*so that surly enough, "ÿ" has max ratio */; +#ifndef NDEBUG + size_t saved_dest_length=dest_length; +#endif + XMLByte *dest_body=new(PointerFreeGC) XMLByte[dest_length+1/*for terminator*/]; if(::transcodeFromUTF8( - (XMLByte *)source_body, source_content_length, - dest_body, dest_content_length, + (XMLByte *)src.str, src_length, + dest_body, dest_length, tables)<0) throw(0, 0, 0, - "Charset::transcodeToUTF8 buffer overflow"); + "Charset::transcodeFromUTF8 buffer overflow"); - // return - adest_body=dest_body; + assert(dest_length<=saved_dest_length); dest_body[dest_length]=0; // terminator + return String::C((char*)dest_body, dest_length); } /// transcode using both charsets -void Charset::transcodeToCharset(Pool& pool, - const Charset& dest_charset, - const void *source_body, size_t source_content_length, - const void *& adest_body, size_t& adest_content_length) const { - 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=new(pool) unsigned char[dest_content_length]; - - const XMLByte* srcPtr=(XMLByte *)source_body; - const XMLByte* srcEnd=(XMLByte *)source_body+source_content_length; - - for(XMLByte* outPtr=dest_body; srcPtrname=name_cstr; - handler->input=xml256CharEncodingInputFunc; handler->inputInfo=&tables; - handler->output=xml256CharEncodingOutputFunc; handler->outputInfo=&tables; + 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(PointerFreeGC) 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 *source, const char* name_cstr) { +void Charset::initTranscoder(const String::Body NAME, const char* name_cstr) { ftranscoder=xmlFindCharEncodingHandler(name_cstr); - transcoder(source); // check right way + transcoder(NAME); // check right way } -xmlCharEncodingHandler *Charset::transcoder(const String *source) { +xmlCharEncodingHandler& Charset::transcoder(const String::Body NAME) { if(!ftranscoder) throw Exception("parser.runtime", - source, + new String(NAME, String::L_TAINTED), "unsupported encoding"); - return ftranscoder; + return *ftranscoder; } -CharPtr Charset::transcode_cstr(xmlChar *s) { +String::C Charset::transcode_cstr(xmlChar* s) { if(!s) - return ""; + return String::C("", 0); - int inlen=strlen((const char* )s); - int outlen=inlen+1; // max - char *out=new char[outlen]; + int inlen=strlen((const char*)s); + int outlen=inlen; // max +#ifndef NDEBUG + int saved_outlen=outlen; +#endif + char *out=new(PointerFreeGC) char[outlen+1]; int error; - if(xmlCharEncodingOutputFunc output=transcoder(0)->output) { + if(xmlCharEncodingOutputFunc output=transcoder(FNAME).output) { error=output( (unsigned char*)out, &outlen, - (const unsigned char*)s, &inlen, - transcoder(0)->outputInfo); + (const unsigned char*)s, &inlen); } else { memcpy(out, s, outlen=inlen); error=0; @@ -583,46 +582,45 @@ CharPtr Charset::transcode_cstr(xmlChar 0, "transcode_cstr failed (%d)", error); - out[outlen/*surely would be less then on input*/]=0; - return out; + assert(outlen<=saved_outlen); out[outlen]=0; + return String::C(out, outlen); } -String& Charset::transcode(xmlChar *s -#ifndef NO_STRING_ORIGIN - , const String *origin -#endif - ) { - String& result=*NEW String(pool()); - result.APPEND_CLEAN(transcode_cstr(s), 0/*auto-size*/, origin->origin().file, origin->origin().line); - return result; -} -const char* Charset::transcode_cstr(GdomeDOMString *s) { - return s?transcode_cstr(BAD_CAST s->str):""; -} -String& Charset::transcode(GdomeDOMString *s -#ifndef NO_STRING_ORIGIN - , const String *origin -#endif - ) { - String& result=*NEW String(pool()); - result.APPEND_CLEAN(transcode_cstr(s), 0/*auto-size*/, origin->origin().file, origin->origin().line); - return result; +const String& Charset::transcode(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); + return *new String(cstr.str, cstr.length, true); } /// @test less memory using -maybe- xmlParserInputBufferCreateMem -xmlChar *Charset::transcode_buf2xchar(Pool& pool, const char* buf, size_t buf_size) { - unsigned char *out; +xmlChar* Charset::transcode_buf2xchar(const char* buf, size_t buf_size) { + xmlChar* out; int outlen; int error; - if(xmlCharEncodingInputFunc input=transcoder(0)->input) { +#ifndef NDEBUG + int saved_outlen; +#endif + if(xmlCharEncodingInputFunc input=transcoder(FNAME).input) { outlen=buf_size*6/*max*/; - out=new(pool) unsigned char[outlen+1]; +#ifndef NDEBUG + saved_outlen=outlen; +#endif + out=(xmlChar*)xmlMalloc(outlen+1); error=input( out, &outlen, - (const unsigned char *)buf, (int *)&buf_size, - transcoder(0)->inputInfo); + (const unsigned char*)buf, (int*)&buf_size); } else { outlen=buf_size; - out=new(pool) unsigned char[outlen+1]; +#ifndef NDEBUG + saved_outlen=outlen; +#endif + out=(xmlChar*)xmlMalloc(outlen+1); memcpy(out, buf, outlen); error=0; } @@ -632,15 +630,70 @@ xmlChar *Charset::transcode_buf2xchar(Po 0, "transcode_buf failed (%d)", error); - out[outlen/*surely would be less then on input*/]=0; - return (xmlChar *)out; + assert(outlen<=saved_outlen); out[outlen]=0; + return out; } GdomeDOMString_auto_ptr Charset::transcode_buf2dom(const char* buf, size_t buf_size) { - return GdomeDOMString_auto_ptr((gchar*)transcode_buf2xchar(buf, buf_size)); + return GdomeDOMString_auto_ptr(transcode_buf2xchar(buf, buf_size)); } GdomeDOMString_auto_ptr Charset::transcode(const String& s) { - CharPtr cstr=s.cstr(String::UL_UNSPECIFIED); + const char* cstr=s.cstr(String::L_UNSPECIFIED); return transcode_buf2dom(cstr, strlen(cstr)); } +GdomeDOMString_auto_ptr Charset::transcode(const String::Body s) { + const char* cstr=s.cstr(); + + return transcode_buf2dom(cstr, s.length()); +} +#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=strlen(src_ptr); + + String::C dest=Charset::transcode(String::C(src_ptr, src_size), + source_transcoder, + dest_transcoder); + + return String::Body(dest.str, dest.length); +} + +String& Charset::transcode(const String& src, + const Charset& source_transcoder, + const Charset& dest_transcoder) { + if(!src.length()) + return *new String("", 0, false); + + 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; isource_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); +}