Detect Code page using IMultiLanguage2 Interface (MS VC++ (Win32))

[xoinki]

Hi all,

I have this sample data in a .txt file,

-- Ne doit pas faire
-- é
select PARC.CONTYP
, PARC.CONSTA
, count(*)
from ParcNonCede PARC
, PLAN P
and PARC.CONSTA not in ('Stock', 'Détourné')
group by CONTYP, CONSTA
;

I am saving this in UTF-8 (not necessarily using Notepad).

Now I have this sample code to detect the Codepage of a buffer as below..
void
ExamineData(String &codepage, String &buffer)
{
IMultiLanguage2 *mlang;
DetectEncodingInfo info = {0};
MIMECPINFO codepageinfo;
int length = 0, cnt = 1;
String codepagestr; // String Datatype is my class
unsigned int len =1;

if(buffer.GetString())
length = buffer.GetLength();

HRESULT hr = S_OK;
SUCCEEDED(CoInitialize(NULL)); // init COM

hr = CoCreateInstance(CLSID_CMultiLanguage, NULL, CLSCTX_INPROC_SERVER, IID_IMultiLanguage2,(void **)&mlang);

hr = mlang->DetectInputCodepage(0,0, (wyChar *)buffer.GetString(), &length, &info, &cnt);

if(SUCCEEDED(hr))
{
hr = mlang->GetCodePageInfo(info.nCodePage, info.nLangID, &codepageinfo);
if(SUCCEEDED(hr))
codepage.SetAs(codepageinfo.wszDescription);
}

mlang->Release();
CoUninitialize();
}

Now this Function Detects the Codepage for the Buffer as "Western European (Windows)" even though it is an UTF-8. But Notepad is able to correctly detect it as UTF-8.

I am Not able to come to a conclusion whether it is a problem with the code or with the API.

Please Help me.

Thanks in Advance,
Xoinki

[weaknessforcats]

This looks odd:

HRESULT hr = S_OK;
SUCCEEDED(CoInitialize(NULL)); // init COM

Shouldn't it be:

HRESULT hr = CoInitialize(NULL); // init COM
if (!SUCCEEDED(hr))
{
    //bail. COM did not initialize
}
 

Further, you should be using CoInitializeEx. Otherwise, your are forced to marshal and that slows everything down. Probably, you should code:

HRESULT hr = CoInitializeEx(NULL, COINIT_MULTITHREADED ); // init COM
if (!SUCCEEDED(hr))
{
    //bail. COM did not initialize
}
 

Then here:

hr = CoCreateInstance(CLSID_CMultiLanguage, NULL, CLSCTX_INPROC_SERVER, IID_IMultiLanguage2,(void **)&mlang);

You don't check that your instance was created. You need to:

hr = CoCreateInstance(CLSID_CMultiLanguage, NULL, CLSCTX_INPROC_SERVER, IID_IMultiLanguage2,(void **)&mlang);
if (!SUCCEEDED(hr))
{
     //bail the instance was not created
}
 

This also looks odd:

hr = mlang->DetectInputCodepage(0,0, (wyChar *)buffer.GetString(), &length, &info, &cnt);

This is the protptype for DetectInputCodePage:

HRESULT DetectInputCodepage(          DWORD dwFlag,
    DWORD dwPrefWinCodePage,
    CHAR *pSrcStr,
    INT *pcSrcSize,
    DetectEncodingInfo *lpEncoding,
    INT *pnScores
);
 

Your third argument is a wyChar* and not a CHAR* as far as I can see. The fifth argument is a DetectEncodingInfo pointer. DetectEncodingInfo is a struct and you have

DetectEncodingInfo info = {0};

but you never use the results from DetectinoutCodePage().

[xoinki]

Hi I changed the Code as u suggested..

<Code>
void
ExamineData(String &codepage, String &buffer)
{
IMultiLanguage2 *mlang;
DetectEncodingInfo info;
MIMECPINFO codepageinfo;
int length = 0, cnt = 1;
String codepagestr;
unsigned int len =1;

if(buffer.GetString())
length = buffer.GetLength();

HRESULT hr = CoInitializeEx(NULL, COINIT_MULTITHREADED); // init COM
if(SUCCEEDED(hr) == false)
{
codepage.SetAs("Unicode (UTF-8)");
return;
}


hr = CoCreateInstance(CLSID_CMultiLanguage, NULL, CLSCTX_INPROC_SERVER, IID_IMultiLanguage2,(void **)&mlang);

if(SUCCEEDED(hr) == false)
{
codepage.SetAs("Unicode (UTF-8)");
CoUninitialize();
return;
}

hr = mlang->DetectInputCodepage(0,0, (char *)buffer.GetString(), &length, &info, &cnt);

if(SUCCEEDED(hr) == true)
{
hr = mlang->GetCodePageInfo(info.nCodePage, info.nLangID, &codepageinfo);

if(SUCCEEDED(hr) == true)
codepage.SetAs(codepageinfo.wszDescription);
}
else
{
codepage.SetAs("Unicode (UTF-8)");
mlang->Release();
CoUninitialize();
return;
}

mlang->Release();
CoUninitialize();
}
</code>

It is still detecting Utf-8 as "western Europen (Windows)" for the buffer i have mentioned above in my post.

Also I am saving the above file (with sample accented characters given above in m,y post ) through a program. so i am not writing any BOM. if i open the same file in notepad - saveas- select UTF-8 from the combo-- save.. then Codepage is detected properly since notepad inserts BOM.. without BOM this API is detecting properly? or is it a mistake in my code?
please help me..
Thnx in advance
Xoinki