--- parser3/src/main/pa_charset.C 2001/12/17 17:59:51 1.4 +++ parser3/src/main/pa_charset.C 2003/03/06 12:21:19 1.33.2.15 @@ -1,31 +1,19 @@ /** @file Parser: Charset connection implementation. - Copyright(c) 2001 ArtLebedev Group(http://www.artlebedev.com) + Copyright(c) 2001-2003 ArtLebedev Group (http://www.artlebedev.com) Author: Alexander Petrosyan(http://paf.design.ru) - - $Id: pa_charset.C,v 1.4 2001/12/17 17:59:51 paf Exp $ */ +static const char* IDENT_CHARSET_C="$Date: 2003/03/06 12:21:19 $"; + #include "pa_charset.h" -//#include "pa_exception.h" -//#include "pa_common.h" -//#include "pa_threads.h" +#include "pa_charsets.h" #ifdef XML -# include -# include -# include -# include +#include "libxml/encoding.h" #endif -// globals - - -// consts - -#define MAX_CHARSET_UNI_CODES 500 - // helpers inline void prepare_case_tables(unsigned char *tables) { @@ -43,7 +31,7 @@ inline void cstr2ctypes(unsigned char *t 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) return 0; if(cstr[1]==0) @@ -52,7 +40,7 @@ inline unsigned int to_wchar_code(const char *error_pos; 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; } static void element2ctypes(unsigned char c, bool belongs, @@ -77,78 +65,21 @@ static void element2case(unsigned char f fcc_table[from]=to; fcc_table[to]=from; } -#ifdef XML -template class ENameMapFor2 : public ENameMap { -public: - // ----------------------------------------------------------------------- - // Constructors and Destructor - // ----------------------------------------------------------------------- - ENameMapFor2( - const XMLCh* const encodingName - , const XMLCh* const fromTable - , const XMLTransService::TransRec* const toTable - , const unsigned int toTableSize - ): ENameMap(encodingName), - ffromTable(fromTable), - ftoTable(toTable), - ftoTableSize(toTableSize) {} - - // ----------------------------------------------------------------------- - // Implementation of virtual factory method - // ----------------------------------------------------------------------- - virtual XMLTranscoder* makeNew(const unsigned int blockSize) const { - return new TType( - getKey(), - blockSize, - ffromTable, - ftoTable, ftoTableSize); - } -private: - const XMLCh* const ffromTable; - const XMLTransService::TransRec* const ftoTable; - const unsigned int ftoTableSize; - -private: - // ----------------------------------------------------------------------- - // Unimplemented constructors and operators - // ----------------------------------------------------------------------- - ENameMapFor2(); - ENameMapFor2(const ENameMapFor2&); - void operator=(const ENameMapFor2&); -}; - -class XML256TableTranscoder2 : public XML256TableTranscoder { -public : - XML256TableTranscoder2( - const XMLCh* const encodingName - , const unsigned int blockSize - , const XMLCh* const fromTable - , const XMLTransService::TransRec* const toTable - , const unsigned int toTableSize - ) : XML256TableTranscoder(encodingName, blockSize, fromTable, toTable, toTableSize) {} - -private: - XML256TableTranscoder2(); - XML256TableTranscoder2(const XML256TableTranscoder2&); - void operator=(const XML256TableTranscoder2&); -}; -#endif - // methods extern "C" unsigned char pcre_default_tables[]; // pcre/chartables.c -Charset::Charset(Pool& apool, const String& aname, const String *file_spec): Pooled(apool), - fname(apool) { - // fname - char *name_cstr=(char *)malloc(aname.size()+1); - memcpy(name_cstr, aname.cstr(String::UL_AS_IS), aname.size()+1); - fname.APPEND_CLEAN(name_cstr, aname.size(), 0, 0); +Charset::Charset(StringPtr aname, StringPtr afile_spec): + fname(aname), + fname_cstr(aname->cstr()) { - if(file_spec) { + for(char *c=fname_cstr; *c; c++) + *c = toupper(*c); + + if(afile_spec) { fisUTF8=false; - loadDefinition(*file_spec); + load_definition(afile_spec); #ifdef XML - addEncoding(name_cstr); + addEncoding(fname_cstr); #endif } else { fisUTF8=true; @@ -157,17 +88,18 @@ Charset::Charset(Pool& apool, const Stri } #ifdef XML - initTranscoder(&aname, name_cstr); + initTranscoder(fname, fname_cstr); #endif } Charset::~Charset() { + // @todonow unregister encodings #ifdef XML - delete transcoder; + // not deleting transcoder, that's not our business #endif } -void Charset::loadDefinition(const String& file_spec) { +void Charset::load_definition(StringPtr afile_spec) { // pcre_tables // lowcase, flipcase, bits digit+word+whitespace, masks @@ -178,17 +110,15 @@ void Charset::loadDefinition(const Strin cstr2ctypes(pcre_tables,(const unsigned char *)"*+?{^.$|()[", ctype_meta); // charset - memset(fromTable, 0, sizeof(fromTable)); - toTable=(XMLTransService::TransRec *)calloc( - sizeof(XMLTransService::TransRec)*MAX_CHARSET_UNI_CODES); - toTableSize=0; + memset(&tables, 0, sizeof(tables)); // strangly vital - toTable[toTableSize].intCh=0; - toTable[toTableSize].extCh=(XMLByte)0; - toTableSize++; + tables.toTable[tables.toTableSize].intCh=0; + tables.toTable[tables.toTableSize].extCh=(XMLByte)0; + tables.toTableSize++; // loading text - char *data=file_read_text(pool(), file_spec); + Pool pool_for_load_only; + char *data=file_read_text(pool_for_load_only, *UTF8_charset, afile_spec); // ignore header getrow(&data); @@ -216,20 +146,20 @@ void Charset::loadDefinition(const Strin case 7: case 8: // charset - if(toTableSize>MAX_CHARSET_UNI_CODES) - throw Exception(0, 0, - &file_spec, + 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); XMLCh unicode=(XMLCh)to_wchar_code(cell); if(!unicode && column==7/*unicode1 column*/) unicode=(XMLCh)c; if(unicode) { - if(!fromTable[c]) - fromTable[c]=unicode; - toTable[toTableSize].intCh=unicode; - toTable[toTableSize].extCh=(XMLByte)c; - toTableSize++; + if(!tables.fromTable[c]) + tables.fromTable[c]=unicode; + tables.toTable[tables.toTableSize].intCh=unicode; + tables.toTable[tables.toTableSize].extCh=(XMLByte)c; + tables.toTableSize++; } break; } @@ -240,36 +170,6 @@ void Charset::loadDefinition(const Strin sort_ToTable(); } -#ifdef XML -void Charset::addEncoding(const char *name_cstr) { - // addEncoding - XalanDOMString sencoding(name_cstr); - const XMLCh* const auto_encoding_cstr=sencoding.c_str(); - int size=sizeof(XMLCh)*(sencoding.size()+1); - XMLCh* pool_encoding_cstr=(XMLCh*)malloc(size); - memcpy(pool_encoding_cstr, auto_encoding_cstr, size); - XMLString::upperCase(pool_encoding_cstr); - - XMLPlatformUtils::fgTransService->addEncoding( - pool_encoding_cstr, - new ENameMapFor2( - pool_encoding_cstr - , fromTable - , toTable - , toTableSize - )); -} - -void Charset::initTranscoder(const String *source, const char *name_cstr) { - XMLTransService::Codes resValue; - transcoder=XMLPlatformUtils::fgTransService->makeNewTranscoderFor(name_cstr, resValue, 60); - if(!transcoder) - throw Exception(0, 0, - source, - "unsupported encoding"); -} -#endif - static int sort_cmp_Trans_rec_intCh(const void *a, const void *b) { return static_cast(a)->intCh- @@ -277,16 +177,18 @@ static int sort_cmp_Trans_rec_intCh(cons } void Charset::sort_ToTable() { - _qsort(toTable, toTableSize, sizeof(*toTable), + _qsort(tables.toTable, tables.toTableSize, sizeof(*tables.toTable), sort_cmp_Trans_rec_intCh); //FILE *f=fopen("c:\\temp\\a", "wb"); - //fwrite(toTable, toTableSize, sizeof(*toTable), f); + //fwrite(tables.toTable, tables.toTableSize, sizeof(*tables.toTable), f); //fclose(f); } -XMLByte Charset::xlatOneTo(const XMLCh toXlat) const { +static XMLByte xlatOneTo(const XMLCh toXlat, + const Charset::Tables& tables, + XMLByte not_found) { unsigned int lowOfs = 0; - unsigned int hiOfs = toTableSize - 1; + unsigned int hiOfs = tables.toTableSize - 1; XMLByte curByte = 0; do { // Calc the mid point of the low and high offset. @@ -295,19 +197,19 @@ XMLByte Charset::xlatOneTo(const XMLCh t // 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>toTable[midOfs].intCh) + if(toXlat>tables.toTable[midOfs].intCh) lowOfs = midOfs; - else if(toXlat outEnd) + break; // We can do it, so update the source index srcPtr++; @@ -449,24 +349,36 @@ void Charset::transcodeToUTF8(Pool& pool outPtr+= encodedBytes; } - // return - adest_body=dest_body; - adest_content_length=outPtr-dest_body; + // 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; } -void Charset::transcodeFromUTF8(Pool& pool, - const void *source_body, size_t source_content_length, - const void *& adest_body, size_t& adest_content_length) const { - size_t dest_content_length=0; - XMLByte *dest_body=(XMLByte*)pool.malloc(source_content_length/*surly enough*/); - - const XMLByte* srcPtr=(const XMLByte*)source_body; - const XMLByte* srcEnd=(const XMLByte*)source_body+source_content_length; - XMLByte* outPtr=dest_body; +/// @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) { + 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 - while(srcPtrtranscodeTo( - s.c_str(), s.length(), - dest, len, - charsEaten, - XMLTranscoder::UnRep_RepChar //UnRep_Throw - ); - dest[size]=0; - error=false; - } - } catch(XMLException& e) { - Exception::provide_source(pool(), 0, e); - } - return(const char *)dest; +static int xml256CharEncodingInputFunc ( + unsigned char *out, + int *outlen, + const unsigned char *in, + int *inlen, + void *info) { + return transcodeToUTF8( + in, *(size_t*)inlen, + out, *(size_t*)outlen, + *(const Charset::Tables *)info); +} + +static int xml256CharEncodingOutputFunc ( + unsigned char *out, + int *outlen, + const unsigned char *in, + int *inlen, + void *info) { + return transcodeFromUTF8( + in, *(size_t*)inlen, + out, *(size_t*)outlen, + *(const Charset::Tables *)info); } -String& Charset::transcode(const XalanDOMString& s) { - return *NEW String(pool(), transcode_cstr(s)); + + +void Charset::addEncoding(char *name_cstr) { + xmlCharEncodingHandler* handler=new xmlCharEncodingHandler; + fcreated_handler=xmlCharEncodingHandlerPtr(fcreated_handler); + + handler->name=name_cstr; + handler->input=xml256CharEncodingInputFunc; handler->inputInfo=&tables; + handler->output=xml256CharEncodingOutputFunc; handler->outputInfo=&tables; + + xmlRegisterCharEncodingHandler(handler); } -std::auto_ptrCharset::transcode_buf(const char *buf, size_t buf_size) { - unsigned int dest_size=0; - XMLCh* dest=(XMLCh *)malloc((buf_size+1)*sizeof(XMLCh)); - unsigned char *charSizes=(unsigned char *)malloc(buf_size*sizeof(unsigned char)); - XalanDOMString *result; - try { - if(transcoder) { - unsigned int bytesEaten; - unsigned int dest_size=transcoder->transcodeFrom( - (unsigned char *)buf, - (const unsigned int)buf_size, - dest,(const unsigned int)buf_size, - bytesEaten, - charSizes - ); - result=new XalanDOMString(dest, dest_size); - } - } catch(XMLException& e) { - Exception::provide_source(pool(), 0, e); - result=0; //calm, compiler +void Charset::initTranscoder(StringPtr source, const char* name_cstr) { + ftranscoder=xmlFindCharEncodingHandler(name_cstr); + transcoder(source); // check right way +} + +xmlCharEncodingHandler& Charset::transcoder(StringPtr source) { + if(!ftranscoder) + throw Exception("parser.runtime", + source, + "unsupported encoding"); + return *ftranscoder; +} + +CharPtr Charset::transcode_cstr(xmlChar* s) { + if(!s) + return ""; + + int inlen=strlen((const char* )s); + int outlen=inlen+1; // max + char *out=new char[outlen]; + + int error; + if(xmlCharEncodingOutputFunc output=transcoder(Exception::undefined_source).output) { + error=output( + (unsigned char*)out, &outlen, + (const unsigned char*)s, &inlen, + transcoder(Exception::undefined_source).outputInfo); + } else { + memcpy(out, s, outlen=inlen); + error=0; + } + if(error<0) + throw Exception(Exception::undefined_type, + Exception::undefined_source, + "transcode_cstr failed (%d)", error); + + out[outlen/*surely would be less then on input*/]=0; + return out; +} +StringPtr Charset::transcode(xmlChar* s +#ifndef NO_STRING_ORIGIN + , StringPtr origin +#endif + ) { + StringPtr result(new String()); + 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):""; +} +StringPtr Charset::transcode(GdomeDOMString* s +#ifndef NO_STRING_ORIGIN + , StringPtr origin +#endif + ) { + StringPtr result(new String()); + result->APPEND_CLEAN(transcode_cstr(s), 0/*auto-size*/, origin->origin().file, origin->origin().line); + return result; +} + +/// @test less memory using -maybe- xmlParserInputBufferCreateMem +xmlChar* Charset::transcode_buf2xchar(const char* buf, size_t buf_size) { + unsigned char* out; + int outlen; + int error; + if(xmlCharEncodingInputFunc input=transcoder(Exception::undefined_source).input) { + outlen=buf_size*6/*max*/; + out=(unsigned char*)malloc(sizeof(xmlChar)*(outlen+1)); + error=input( + out, &outlen, + (const unsigned char *)buf, (int *)&buf_size, + transcoder(Exception::undefined_source).inputInfo); + } else { + outlen=buf_size; + out=(unsigned char*)malloc(sizeof(xmlChar)*(outlen+1)); + memcpy(out, buf, outlen); + error=0; } - return std::auto_ptr(result); + if(error<0) + throw Exception(Exception::undefined_type, + Exception::undefined_source, + "transcode_buf failed (%d)", error); + + out[outlen/*surely would be less then on input*/]=0; + return (xmlChar*)out; +} +GdomeDOMString_auto_ptr Charset::transcode_buf2dom(const char* buf, size_t buf_size) { + return GdomeDOMString_auto_ptr((gchar*)transcode_buf2xchar(buf, buf_size)); } -std::auto_ptrCharset::transcode(const String& s) { - const char *cstr=s.cstr(String::UL_UNSPECIFIED); +GdomeDOMString_auto_ptr Charset::transcode(StringPtr s) { + CharPtr cstr=s->cstr(String::UL_UNSPECIFIED); - return transcode_buf(cstr, strlen(cstr)); + return transcode_buf2dom(cstr, strlen(cstr)); } #endif