Passing a SAFEARRAY of bytes to load function of MSXML - c++

Question

If any kind soul out there please go through following source and tell me why MSXML "load" function fails to load this XML.
Here i'm trying to load a UTF-8 encoded XML using MSXML parser's "load" function. And i have a BSTR [UTF-16 encoded] as an argument, so i'm trying to convert it, into a SAFEARRAY of bytes so that i can passed it into the "load" function of MSXML. But the issue is load function failed to load this XML. If anyone could provide a solution to i would be really grateful.
#include <windows.h>
#include <objsafe.h>
#include <objbase.h>
#include <atlbase.h>
#include <string>
#include <comutil.h>
#include <msxml2.h>
#include <iostream>
using namespace std;
#define STATUS_SUCCESS 0
#define STATUS_FAIL -1
long LoadXmlData(BSTR xmlDoc)
{
HRESULT hr = S_OK;
CComPtr <IXMLDOMDocument> xmlDomDoc = NULL;
CComPtr <IXMLDOMElement> docRoot = NULL;
VARIANT_BOOL isParseSucess = FALSE;
CoInitialize(NULL);
hr = xmlDomDoc.CoCreateInstance(__uuidof(DOMDocument30));
if (FAILED(hr))
{
return STATUS_FAIL;
}
BYTE HUGEP *pByte;
int len = WideCharToMultiByte(CP_UTF8, 0, xmlDoc, -1, NULL, 0, NULL, NULL);
SAFEARRAYBOUND rgsabound[1];
rgsabound[0].cElements = len;
rgsabound[0].lLbound = 0;
SAFEARRAY* psa = SafeArrayCreate(VT_UI1, 1, rgsabound);
if (psa != NULL)
{
hr = SafeArrayAccessData(psa, (void HUGEP**)&pByte);
if (!FAILED(hr))
{
if (len > 0)
{
WideCharToMultiByte(CP_UTF8, 0, xmlDoc, -1, (LPSTR)&pByte[0], len, NULL, NULL);
//cout << "Converted Byte Array: " << pByte << endl << endl;
}
else
{
return STATUS_FAIL;
}
SafeArrayUnaccessData(psa);
}
}
VARIANT v;
VariantInit(&v);
V_VT(&v) = VT_ARRAY | VT_UI1;
V_ARRAY(&v) = psa;
hr = xmlDomDoc->load(v, &isParseSucess);
//hr = xmlDomDoc->loadXML(xmlDoc, &isParseSucess); //can't use this function because XML is encoded in UTF-8
if (FAILED(hr) || (!isParseSucess))
{
return STATUS_FAIL;
}
else
{
return STATUS_SUCCESS;
}
}
int main()
{
BSTR xmlDoc = SysAllocString(L"<?xml version=\"1.0\" encoding=\"UTF-8\"?> <response> </response> ");
long ret = LoadXmlData(xmlDoc);
if (ret == STATUS_SUCCESS)
{
cout << "MSXML: loading the XML succeeded";
}
else
{
cout << "MSXML: loading the XML failed";
}
//string str;
//getline(cin, str);
return 0;
}
ps: If anyone try to compile this source you may get a link error first time, add comsuppw.lib as a linker dependency in VS settings. And XML is UTF-8 encoded, so i can't use "LoadXML" function in MSXML.

Answer 1

XML files are UTF-8 but there is no conversion needed here because Windows functions will automatically convert between UTF-16 and UTF-8 when needed (unless input/output is BYTE*...)
BSTR needs cleanup. You can use CComBSTR(str) which has automatic cleanup.
You can use const wchar_t* to pass strings to your own functions. BSTR is required for COM etc.
As noted in comments, IXMLDOMDocument::load expects a file name as input. Use IXMLDOMDocument::loadXML in this case.
#include <iostream>
#include <windows.h>
#include <atlbase.h>
#include <msxml2.h>
long LoadXmlData(const wchar_t* content)
{
HRESULT hr = S_FALSE;
CComPtr<IXMLDOMDocument> doc = NULL;
CComPtr<IXMLDOMElement> docRoot = NULL;
hr = doc.CoCreateInstance(__uuidof(DOMDocument30));
if(FAILED(hr))
return S_FALSE;
VARIANT_BOOL is_success = FALSE;
CComBSTR arg(content);
hr = doc->loadXML(arg, &is_success);
//if you are loading from a file:
//hr = doc->load(CComVariant(L"c:\\test\\test.xml"), &is_success);
if(FAILED(hr) || !is_success)
return S_FALSE;
//if save is needed:
//doc->save(CComVariant(L"c:\\test\\test.xml"));
return S_OK;
}
int main()
{
CoInitialize(NULL);
//ελληνική γλώσσα Greek text for testing
CComBSTR data(L"<?xml version=\"1.0\" encoding=\"UTF-8\"?><response>ελληνική γλώσσα</response>");
long ret = LoadXmlData(data);
if(ret == S_OK)
std::cout << "MSXML: loading the XML succeeded\n";
else
std::cout << "MSXML: loading the XML failed\n";
CoUninitialize();
return 0;
}