Question

I’m trying to write a universal text editor which can open and display ANSI and Unicode in EditControl. Do I need to repeatedly call ReadFile() if I determine that the text is ANSI? Can’t figure out how to perform this task. My attempt below does not work, it displays ‘?’ characters in EditControl.

LARGE_INTEGER fSize;
        GetFileSizeEx(hFile,&fSize);

        int bufferLen = fSize.QuadPart/sizeof(TCHAR)+1;
        TCHAR* buffer = new TCHAR[bufferLen];       
        buffer[0] = _T('\0');

        DWORD wasRead = 0;
        ReadFile(hFile,buffer,fSize.QuadPart,&wasRead,NULL);        
        buffer[wasRead/sizeof(TCHAR)] = _T('\0');

        if(!IsTextUnicode(buffer,bufferLen,NULL))
        {                       
            CHAR* ansiBuffer = new CHAR[bufferLen];
            ansiBuffer[0] = '\0';
            WideCharToMultiByte(CP_ACP,0,buffer,bufferLen,ansiBuffer,bufferLen,NULL,NULL);
            SetWindowTextA(edit,ansiBuffer);
            delete[]ansiBuffer;
        }
        else
            SetWindowText(edit,buffer);

        CloseHandle(hFile);
        delete[]buffer;

Answer 1

There are a few buffer length errors and oddities, but here’s your big problem. You call WideCharToMultiByte incorrectly. That is meant to receive UTF-16 encoded text as input. But when IsTextUnicode returns false that means that the buffer is not UTF-16 encoded.

The following is basically what you need:

if(!IsTextUnicode(buffer,bufferLen*sizeof(TCHAR),NULL))
    SetWindowTextA(edit,(char*)buffer);

Note that I’ve fixed the length parameter to IsTextUnicode.

For what it is worth, I think I’d read in to a buffer of char. That would remove the need for the sizeof(TCHAR). In fact I’d stop using TCHAR altogether. This program should be Unicode all the way – TCHAR is what you use when you compile for both NT and 9x variants of Windows. You aren’t compiling for 9x anymore I imagine.

So I’d probably code it like this:

char* buffer = new char[filesize+2];//+2 for UTF-16 null terminator
DWORD wasRead = 0;
ReadFile(hFile, buffer, filesize, &wasRead, NULL);        
//add error checking for ReadFile, including that wasRead == filesize
buffer[filesize] = '\0';
buffer[filesize+1] = '\0';
if (IsTextUnicode(buffer, filesize, NULL))
    SetWindowText(edit, (wchar_t*)buffer);
else
    SetWindowTextA(edit, buffer);
delete[] buffer;

Note also that this code makes no allowance for the possibility of receiving UTF-8 encoded text. If you want to handle that you’d need to take your char buffer and send to through MultiByteToWideChar using CP_UTF8.