Be careful, you are leaking the transcoder 't' and the array 'sizes'.
Also, you could store the transcoder in a global
variable to avoid allocating/deallocating for each operation.
Alberto
At 23.09 16/07/2007 +0200, Kuisathaverat wrote:
thanks i make a function that transcode from ISO8859-1
and replace X("") for it , i send the code to future
people that have the same problem
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <ctype.h>
#include <sstream>
#include <iostream>
#include "xercesc/util/PlatformUtils.hpp"
#include "xercesc/util/TransService.hpp"
#include <xercesc/util/XMLString.hpp>
#include <xalanc/Include/PlatformDefinitions.hpp>
#include <xercesc/util/PlatformUtils.hpp>
#include <xercesc/framework/LocalFileInputSource.hpp>
#include <xercesc/framework/LocalFileFormatTarget.hpp>
#include <xercesc/framework/MemBufInputSource.hpp>
#include
<xalanc/XercesParserLiaison/XercesDocumentWrapper.hpp>
#include
<xalanc/XercesParserLiaison/XercesWrapperNavigator.hpp>
#include <xercesc/dom/DOM.hpp>
#include <xercesc/dom/DOMDocument.hpp>
#include <xercesc/sax/SAXParseException.hpp>
#include <xercesc/util/XMLUniDefs.hpp>
#include <xercesc/util/XMLUni.hpp>
#include <xercesc/framework/StdInInputSource.hpp>
#include <xercesc/framework/StdOutFormatTarget.hpp>
#include <xercesc/framework/XMLBuffer.hpp>
#include <xercesc/dom/DOMElement.hpp>
#include <xercesc/dom/DOMWriter.hpp>
#include <xercesc/dom/DOMNodeList.hpp>
#include <xercesc/dom/DOMNode.hpp>
#include <xercesc/parsers/XercesDOMParser.hpp>
#include <xercesc/parsers/SAXParser.hpp>
#include <xercesc/util/OutOfMemoryException.hpp>
#include <xercesc/sax/InputSource.hpp>
#include <locale.h>
using namespace std;
using namespace XERCES_CPP_NAMESPACE;
#define X(p) xercesc::XMLString::transcode(p)
XMLCh *transcodeFromISO88591(string sCadena){
XMLTransService::Codes failReason;
XMLTranscoder* t =
XMLPlatformUtils::fgTransService->makeNewTranscoderFor("ISO8859-1",
failReason,
16*1024,
xercesc::XMLPlatformUtils::fgMemoryManager);
unsigned int srcLen=sCadena.length();
int destLen=srcLen+10;
unsigned char* sizes = new unsigned char[destLen+1];
XMLCh *dest=0;
dest=(XMLCh*)calloc(destLen+1,sizeof(XMLCh));
unsigned int eaten=0;
unsigned int endDest=0;
endDest=t->transcodeFrom((const
XMLByte*)sCadena.c_str(),srcLen,(XMLCh
*)dest,destLen,eaten,sizes);
return dest;
}
void initXerces(){
// Initialize the XML4C2 system
try
{
xercesc::XMLPlatformUtils::Initialize();
}
catch(const XMLException &toCatch)
{
XERCES_STD_QUALIFIER cerr << "Error during Xerces-c
Initialization.\n"
<< " Exception message:"
<< X(toCatch.getMessage()) << XERCES_STD_QUALIFIER
endl;
throw std::runtime_error("Can No ini Xerces");
}
}
void endXerces(){
// Initialize the XML4C2 system
try
{
xercesc::XMLPlatformUtils::Terminate();
}
catch(const XMLException &toCatch)
{
XERCES_STD_QUALIFIER cerr << "Error during Xerces-c
Initialization.\n"
<< " Exception message:"
<< X(toCatch.getMessage()) << XERCES_STD_QUALIFIER
endl;
throw std::runtime_error("Can No ini Xerces");
}
}
DOMDocument *createDocument(string sRootElem, string
sEncoding){
DOMImplementation* impl =
DOMImplementationRegistry::getDOMImplementation(transcodeFromISO88591("Core"));
DOMDocument* pXMLDoc = NULL;
if (impl != NULL)
{
try
{
pXMLDoc = impl->createDocument(
0,
// root element namespace
URI.
transcodeFromISO88591(sRootElem), // root
element name
0); // document type object
(DTD).
pXMLDoc->setEncoding(transcodeFromISO88591(sEncoding));
//
pXMLDoc->setActualEncoding(X(sEncoding.c_str()));
// pXMLDoc->setVersion(X("1.1"));
DOMText* pText =
pXMLDoc->createTextNode(transcodeFromISO88591("níñó"));
pXMLDoc->getDocumentElement()->appendChild(pText);
}
catch (...)
{
throw std::runtime_error("Error: creating
document");
}
} // (inpl != NULL)
else
{
throw std::runtime_error("Error: no
implementation");
}
return pXMLDoc;
}
DOMElement *createNode(DOMDocument *pXMLDoc,string
sName){
DOMElement* pNode = NULL;
try{
pNode =
pXMLDoc->createElement(transcodeFromISO88591(sName));
pXMLDoc->getDocumentElement()->appendChild(pNode);
}catch(...){}
return pNode;
}
void setAttr(DOMElement *pNodeElem,string sName,string
sValue){
pNodeElem->setAttribute(transcodeFromISO88591(sName),transcodeFromISO88591(sValue));
}
bool saveFile(DOMDocument *pXMLDoc,string
sFileName,string sEncoding){
bool bRet = true;
try
{
// get a serializer, an instance of DOMWriter
DOMImplementation *impl =
pXMLDoc->getImplementation();
DOMWriter *theSerializer =
impl->createDOMWriter();
// set user specified output encoding
theSerializer->setEncoding(transcodeFromISO88591(sEncoding));
// set feature if the serializer supports
the feature/mode
// if
(theSerializer->canSetFeature(XMLUni::fgDOMWRTEntities,
true))
//
theSerializer->setFeature(XMLUni::fgDOMWRTEntities,
true);
// if
(theSerializer->canSetFeature(XMLUni::fgDOMWRTNormalizeCharacters,
true))
//
theSerializer->setFeature(XMLUni::fgDOMWRTNormalizeCharacters,
true);
// if
(theSerializer->canSetFeature(XMLUni::fgDOMWRTSplitCdataSections,
true))
//
theSerializer->setFeature(XMLUni::fgDOMWRTSplitCdataSections,
true);
// if
(theSerializer->canSetFeature(XMLUni::fgDOMWRTDiscardDefaultContent,
true))
//
theSerializer->setFeature(XMLUni::fgDOMWRTDiscardDefaultContent,
true);
if
(theSerializer->canSetFeature(XMLUni::fgDOMWRTFormatPrettyPrint,
true))
theSerializer->setFeature(XMLUni::fgDOMWRTFormatPrettyPrint,
true);
// if
(theSerializer->canSetFeature(XMLUni::fgDOMWRTBOM,
true))
// theSerializer->setFeature(XMLUni::fgDOMWRTBOM,
true);
//
// Plug in a format target to receive the
resultant
// XML stream from the serializer.
//
// StdOutFormatTarget prints the resultant
XML stream
// to stdout once it receives any thing
from the serializer.
//
XMLFormatTarget *myFormTarget;
// myFormTarget = new
LocalFileFormatTarget(sFileName.c_str());
myFormTarget = new StdOutFormatTarget();
// get the DOM representation
pXMLDoc->normalizeDocument();
DOMNode *doc = (DOMNode
*)pXMLDoc;
//
// do the serialization through
DOMWriter::writeNode();
//
theSerializer->writeNode(myFormTarget, *doc);
delete theSerializer;
//
// Filter, formatTarget and error handler
// are NOT owned by the serializer.
//
delete myFormTarget;
}
catch (const OutOfMemoryException&)
{
XERCES_STD_QUALIFIER cerr << "OutOfMemoryException"
<< XERCES_STD_QUALIFIER endl;
bRet=false;
}
catch (XMLException& e)
{
XERCES_STD_QUALIFIER cerr << "An error occurred
during creation of output transcoder. Msg is:"
<< XERCES_STD_QUALIFIER endl
<< X(e.getMessage()) << XERCES_STD_QUALIFIER endl;
bRet=false;
}
return bRet;
}
int main(int argc, char *argv[])
{
initXerces();
DOMDocument *pXMLDoc =
createDocument("DocRoot","UTF-8");
DOMElement *pNodeElem=0;
pNodeElem = createNode(pXMLDoc,"Node");
setAttr(pNodeElem,"id","1");
setAttr(pNodeElem,"desc","níñó");
saveFile(pXMLDoc,"/tmp/sample.xml","UTF-8");
saveFile(pXMLDoc,"/tmp/sample.xml","ISO-8859-1");
saveFile(pXMLDoc,"/tmp/sample.xml","ISO-8859-15");
endXerces();
return EXIT_SUCCESS;
}
--- Alberto Massari <[EMAIL PROTECTED]>
escribió:
> You should not use X("ñíñó"), as
> XMLString::transcode assumes that the given
> string is written in the current locale; if you
> are not using a Spanish machine you will get
> garbage Unicode characters (probably, '?') that
> each output encoding will render with a character
> entity. You should use the appropriate transcoder
> to convert your strings into Unicode.
>
> Alberto
>
> At 13.26 14/07/2007 +0200, Kuisathaverat wrote:
> >I have used DOMWriter->WriteNode() to write
> documens
> >that previusly i parse and modify , with this
> >documents i no have problems with the output
> >characters in iso8859-1 or iso8859-15 or utf-8
> >encoding but when i make the document with
> >DOMImplementation the especial characteras as á or
> ñ
> >will lost in ther write to disk.
> >I make a short example to ilustrate it. The correct
> >output was :
> ><DocRoot>níñó
> ><Node id="1" desc="níñó" />
> ></DocRoot>
> >
> >but i obtain :
> >
> >in UTF-8
> >
> ><?xml version="1.0" encoding="UTF-8"
> standalone="no"
> >?>
> ><DocRoot xmlns="">����
> >
> > <Node desc="����"
> >id="1"/>
> >
> ></DocRoot>
> >
> >in iso8859-1
> >
> ><?xml version="1.0" encoding="ISO-8859-1"
> >standalone="no" ?>
> ><DocRoot xmlns="">����
> >
> > <Node desc="����"
> >id="1"/>
> >
> ></DocRoot>
> >
> >in iso8859-15
> >
> ><?xml version="1.0" encoding="ISO-8859-15"
> >standalone="no" ?>
> ><DocRoot xmlns="">����
> >
> > <Node desc="����"
> >id="1"/>
> >
> ></DocRoot>
> >
> >I know that i do some thin wrong but i do no know ,
> i
> >search in mailllist but i no found solution or
> similar
> >problem.
> >
> >it is the code
> >
> >#include <stdio.h>
> >#include <stdlib.h>
> >#include <string.h>
> >#include <ctype.h>
> >#include <sstream>
> >#include <iostream>
> >#include "xercesc/util/PlatformUtils.hpp"
> >#include "xercesc/util/TransService.hpp"
> >#include <xercesc/util/XMLString.hpp>
> >#include <xalanc/Include/PlatformDefinitions.hpp>
> >
> >#include <xercesc/util/PlatformUtils.hpp>
> >#include
> <xercesc/framework/LocalFileInputSource.hpp>
> >#include
> <xercesc/framework/LocalFileFormatTarget.hpp>
> >#include <xercesc/framework/MemBufInputSource.hpp>
> >#include
>
><xalanc/XercesParserLiaison/XercesDocumentWrapper.hpp>
> >#include
>
><xalanc/XercesParserLiaison/XercesWrapperNavigator.hpp>
> >#include <xercesc/dom/DOM.hpp>
> >#include <xercesc/dom/DOMDocument.hpp>
> >#include <xercesc/sax/SAXParseException.hpp>
> >#include <xercesc/util/XMLUniDefs.hpp>
> >#include <xercesc/util/XMLUni.hpp>
> >#include <xercesc/framework/StdInInputSource.hpp>
> >#include <xercesc/framework/StdOutFormatTarget.hpp>
> >#include <xercesc/framework/XMLBuffer.hpp>
> >#include <xercesc/dom/DOMElement.hpp>
> >#include <xercesc/dom/DOMWriter.hpp>
> >#include <xercesc/dom/DOMNodeList.hpp>
> >#include <xercesc/dom/DOMNode.hpp>
> >#include <xercesc/parsers/XercesDOMParser.hpp>
> >#include <xercesc/parsers/SAXParser.hpp>
> >#include <xercesc/util/OutOfMemoryException.hpp>
> >#include <xercesc/sax/InputSource.hpp>
> >
> >using namespace std;
> >using namespace XERCES_CPP_NAMESPACE;
> >
> >
> >#define X(p) xercesc::XMLString::transcode(p)
> >
> > void initXerces(){
> > // Initialize the XML4C2 system
> > try
> > {
> >
> xercesc::XMLPlatformUtils::Initialize();
> > }
> >
> > catch(const XMLException &toCatch)
> > {
> > XERCES_STD_QUALIFIER cerr <<
> "Error during Xerces-c
> >Initialization.\n"
> > << " Exception
> message:"
> > <<
> X(toCatch.getMessage()) <<
> >XERCES_STD_QUALIFIER endl;
> > throw std::runtime_error("Can No
> ini Xerces");
> > }
> >
> > }
> >
> >void endXerces(){
> > // Initialize the XML4C2 system
> > try
> > {
> >
> xercesc::XMLPlatformUtils::Terminate();
> > }
> >
> > catch(const XMLException &toCatch)
> > {
> > XERCES_STD_QUALIFIER cerr <<
> "Error during Xerces-c
> >Initialization.\n"
> > << " Exception
> message:"
> > <<
> X(toCatch.getMessage()) <<
> >XERCES_STD_QUALIFIER endl;
> > throw std::runtime_error("Can No
> ini Xerces");
> > }
> >
> > }
> >DOMDocument *createDocument(string sRootElem,
> string
> >sEncoding){
> >DOMImplementation* impl =
>
>DOMImplementationRegistry::getDOMImplementation(X("Core"));
> >DOMDocument* pXMLDoc = NULL;
> > if (impl != NULL)
> > {
> > try
> > {
> > pXMLDoc =
> impl->createDocument(
> > 0,
> > // root element namespace
> >URI.
> > X(sRootElem.c_str()),
> // root element name
> > 0); //
> document type object
> >(DTD).
> >
> pXMLDoc->setEncoding(X(sEncoding.c_str()));
> >//
> >pXMLDoc->setActualEncoding(X(sEncoding.c_str()));
> >//
> pXMLDoc->setVersion(X("1.1"));
> > DOMText* pText =
> >pXMLDoc->createTextNode(X("ñíñó"));
> >
> pXMLDoc->getDocumentElement()->appendChild(pText);
> > }
> > catch (...)
> > {
> > throw
> std::runtime_error("Error: creating
> >document");
> > }
> > } // (inpl != NULL)
> > else
>
=== message truncated ===
Un Saludo
kuisathaverat
Solo hay 10 tipos de personas, las que saben binario y las que no
http://kuisat.no-ip.org
http://www.fortunecity.com/campus/physics/114
http://kuisat.infierno.org
ICQ #9368869
Linux User #107185
______________________________________________
LLama Gratis a cualquier PC del Mundo.
Llamadas a fijos y móviles desde 1 céntimo por minuto.
http://es.voice.yahoo.com
