ATL / WTL / STL

HRESULT GetDocumentContent(IWebBrowser2 *pWebBrowser2, CComBSTR &bstrHTMLText)
{
			//C: Clear the bstr output parameter.
	bstrHTMLText.Empty();
			//C: Get the WebBrowser's document object
	CComPtr<IDispatch> pDisp;
	HRESULT hr = pWebBrowser2->get_Document(&pDisp);
	if (FAILED(hr))
	{
		return hr;
	}
			//C: Verify that what we get is a pointer to a IHTMLDocument2 
			//   interface. 
	CComQIPtr<IHTMLDocument2, &IID_IHTMLDocument2> spHTML = pDisp;
			//C: Extract the source code of the document
	if (spHTML)
	{
		CHAR	*szData = NULL;
		OLECHAR *htmlText = NULL;
		try
		{
			//C: Query for the persist stream that holds the original data.
			CComQIPtr<IPersistStreamInit, &IID_IPersistStreamInit> spData = spHTML;
			CComPtr<IStream>	spStream;
			//C: Create an IStream object to receive the data.
			::CreateStreamOnHGlobal(NULL, TRUE, &spStream);
			//C: Save the data to this new stream memory.
			spData->Save(spStream ,FALSE);

			STATSTG stat;
			spStream->Stat(&stat, STATFLAG_NONAME);
			
			szData = new CHAR[stat.cbSize.LowPart];
			ULONG length;

			LARGE_INTEGER pos;
			pos.QuadPart = (__int64)0;
			spStream->Seek(pos, STREAM_SEEK_SET, NULL);
			spStream->Read(szData, stat.cbSize.LowPart, &length);

			htmlText = new OLECHAR[length];
			::MultiByteToWideChar(CP_ACP, 0, szData, length, htmlText, length);
			bstrHTMLText = htmlText;
		}
		catch (...)
		{
			//C: No current operations.
		}
			//C: Free resources.
		delete[] htmlText;
		delete[] szData;
	}
	else
	{
			//C: The document is not an HTML page.
			//C: No current operations.
	}
 	return S_OK;  
}