String Manipulations using TCHAR Library

Introduction

To write Win32 application or MFC application, it may happen that you want to add Unicode support. Unicode gives you the freedom to include texts from any language in your application. On the contrary despite the advantages of Unicode you may decide that you will never use Unicode in your program. In both cases, TCHAR library will help you a lot. 

Tips and Tricks 

If you want to use Unicode in project properties, select "Use Unicode Character Set" as character set. From now on, you can easily use Unicode strings in your code wrapping with a macro. The same macro works fine even if you selected not to use "Unicode Character Set". Code remains same while you have the advantage. 

To use a Unicode string as a function parameter, wrap it with macro _T or TEXT macro. For example:

AfxMessageBox(_T("You clicked it!"));  

_T() and TEXT() are macros from tchar header file. TCHAR library automatically maps functions to Unicode when Unicode is defined. Using TCHAR library helps us to move code to multibyte stream (or unicode) whenever required. Try to avoid primitive data type char array or char *. This is because before using them in your controls, you have to convert them. Repetitive conversion may be tedious. 

Use TCHAR instead of char and use TCHAR* instead of char*. TCHAR* can be written LPTSTR. For const TCHAR*, you may write LPCTSTR which is required when a string is passed as an argument to a function where modification should be restricted.

To calculate length of strings, use _tcslen function:

len = _tcslen(str); 

To compare strings instead of strcmp and strncmp, use _tcscmp and _tcsncmp. Here’s an example:

if (!_tcsncmp(line, _T("desiredtext"), 11))
    AfxMessageBox(_T(“Got desired text.”));

For copying string _tcscpy_s and _tcsncpy_s instead of strcpy and strncpy:

_tcsncpy_s(dest, srcstr, 20);

Note, strpy or _tcscpy_s, don’t put a null after copying the string so remember to set null after copying the string when required.

For string concatenation, use _tcscat_s and _tcsncat_s instead of strcat and strncat. Here, the 2^nd parameter is the size of the destination string.

_tcsncat_s(timestamp, 20, str, i);

For splitting tokens, use _tcstok_s instead of strtok.

LPTSTR  next_token;
token = _tcstok_s(str, delim, &next_token);

To convert a string to integer, you can use _ttoi() function.

CString str = _T("10");
CString temp;
temp.Format(_T(" length: %d"), _ttoi(str));
AfxMessageBox(str+temp);

When I didn't know about this handy function I wrote the following function to do the same task. Having a look at the function may help you as an example code: 

int GetEquivValue(TCHAR ch, int base) {
    int diff = _T('a') - _T('A');

    // Invalid base
    if (base >= 16)
        return 0;

    if (base == 16) {
        // make upper case if lowercase
        if (ch >= _T('a') && ch <= _T('z'))
            ch -= diff;

        diff = _T('A') - _T('0');

        if (ch >= _T('A') && ch <= _T('F'))
            return (ch-diff);
    }

    if (ch >= _T('0') && ch <= _T('9'))
        return (ch - _T('0'));
    return 0;
}

int SAatoib(LPTSTR  str, int base) {
    int res = 0;
    int i, len, tmp;

    len = _tcslen(str);
    for (i=0; i<len; i++) {
        tmp = GetEquivValue(str[i], base);
        res = res*base + tmp;
    }
    return res;
} 

Using these functions to convert a string to decimal number, you have to write like this: 

int decVal = SAatoib((LPTSTR )NumStr, 10);  

To convert a string to hexadecimal number, you have to write like this: 

int decVal = SAatoib((LPTSTR )NumStr, 16);  

To convert a string to octal number, you have to write: 

int decVal = SAatoib((LPTSTR )NumStr, 8);  

To format the string like printf, you can use CString::Format function. For example, to get IP address: 

unsigned long int ipsegval[4] = {1, 2, 3, 4};
CString ip;
ip.Format(_T("%u.%u.%u.%u"), ipsegval[0], ipsegval[1], ipsegval[2], ipsegval[3]);

For example, you have a CString. But you need it as const char* or char* The cast to LPCTSTR from CString is a valid one, but cast to const char* isn’t valid. But there is a way. Here’s an example:

CString ipaddrstr(_T(“1.2.3.4”));
CStringA ipaddrstrA(ipaddrstr);
ipaddr = inet_addr(ipaddrstrA);

Note: inet_addr function requires char *.

I hope this post helps beginners. 

History 

• 16th January, 2012: Version 4 
• 11^th May, 2011: Version 3
• 9^th May, 2011: Version 2
• 6^th May, 2011: Initial version