Convert lptstr to char*

Question

Would anyone happen to know how to convert type LPTSTR to char * in C++?

Answer 1

Depends if it is Unicode or not it appears. LPTSTR is char* if not Unicode, or w_char* if so.

Discussed better here (accepted answer worth reading)

Answer 2

Here are a lot of ways to do this. MFC or ATL's CString, ATL macros, or Win32 API.

LPTSTR szString = _T("Testing");
char* pBuffer;

You can use ATL macros to convert:

USES_CONVERSION;
pBuffer = T2A(szString);

CString:

CStringA cstrText(szString);

or the Win32 API WideCharToMultiByte if UNICODE is defined.

Answer 3

If your compiler Character Setting is set to Unicode Character Set, then LPTSTR will be interpreted as wchar_t*. In that case Unicode to Multibyte character conversion is required.
(In Visual Studio, setting is located at Project Properties\Configuration Properties\General\Character Set)

The sample code below should give an idea:

#include <windows.h>

/* string consisting of several Asian characters */
LPTSTR wcsString = L"\u9580\u961c\u9640\u963f\u963b\u9644";
//LPTSTR wcsString = L"OnlyAsciiCharacters";

char* encode(const wchar_t* wstr, unsigned int codePage)
{
    int sizeNeeded = WideCharToMultiByte(codePage, 0, wstr, -1, NULL, 0, NULL, NULL);
    char* encodedStr = new char[sizeNeeded];
    WideCharToMultiByte(codePage, 0, wstr, -1, encodedStr, sizeNeeded, NULL, NULL);
    return encodedStr;
}

wchar_t* decode(const char* encodedStr, unsigned int codePage)
{
   int sizeNeeded = MultiByteToWideChar(codePage, 0, encodedStr, -1, NULL, 0);
   wchar_t* decodedStr = new wchar_t[sizeNeeded ];
   MultiByteToWideChar(codePage, 0, encodedStr, -1, decodedStr, sizeNeeded );
   return decodedStr;
}

int main(int argc, char* argv[])
{
   char* str = encode(wcsString, CP_UTF8); //UTF-8 encoding
   wchar_t* wstr = decode(str, CP_UTF8);
   //If the wcsString is UTF-8 encodable, then this comparison will result to true.
   //(As i remember some of the Chinese dialects cannot be UTF-8 encoded 
   bool ok = memcmp(wstr, wcsString, sizeof(wchar_t) * wcslen(wcsString)) == 0; 
   delete str;
   delete wstr;

   str = encode(wcsString, 20127); //US-ASCII (7-bit) encoding
   wstr = decode(str, 20127);
   //If there were non-ascii characters existing on wcsString, 
   //we cannot return back, since some of the data is lost
   ok = memcmp(wstr, wcsString, sizeof(wchar_t) * wcslen(wcsString)) == 0; 
   delete str;
   delete wstr;
}

On the other hand, if your compiler Character Setting is set to Multibyte, then LPTSTR will be interpreted as char*.

In that case:

LPTSTR x = "test";
char* y;
y = x;

Also see:

Another discussion about wchar_t conversion: How do you properly use WideCharToMultiByte
MSDN Article: http://msdn.microsoft.com/en-us/library/dd374130(v=vs.85).aspx
Valid Code Page Identifiers: http://msdn.microsoft.com/en-us/library/dd317756(v=vs.85).aspx

Answer 4

char * pCopy = NULL;
if (sizeof(TCHAR) == sizeof(char))
{
    size_t size = strlen(pOriginal);
    pCopy = new char[size + 1];
    strcpy(pCopy, pOriginal);
}
else
{
    size_t size = wcstombs(NULL, pOriginal, 0);
    pCopy = new char[size + 1];
    wcstombs(pCopy, pOriginal, size + 1);
}

Answer 5

OK, so lets say that you HAVE to use Unicode. And you use some functions like LookupAccountSid, that are required for your program to function - but they return LPTSTR for important information you NEED to process as a string (for whatever reason - it's programming, stuff like this happens)

Now, if you were using multibyte - this wouldn't be an issue. But there is a way to solve it. This is my method and is admittedly sloppy. But nonetheless you should be able to see how it works.

const std::wstring &wstring = AcctName; // AcctName being my LPTSTR string
int size_needed = WideCharToMultiByte(CP_UTF8, 0, &wstring[0], (int)wstring.size(), NULL, 0, NULL, NULL);
std::string strTo(size_needed, 0);

WideCharToMultiByte(CP_UTF8, 0, & wstring[0], (int)wstring[0], &strTo[0], size_needed, NULL, NULL);

char* charUserName = new char[strTo.size() + 1];

// Set charUserName via copying
std::copy(strTo.begin(), strTo.end(), charUserName);
charUserName[strTo.size()] = '\0';

SetUPI(charUserName); // charUserName being my converted char * - 
// You don't need this last part - but this is an example of passing to method
// that takes a string

Any questions just ask. I realise this is an old post - but I like to post for people in the furture that come looking. (people like me)

Answer 6

I hope this helps someone, because it took me a while to figure out how to do it.

First of all, LPTSTR is of pointer type and it is basically equivalent to TCHAR* (assuming that <tchar.h> is included). Note that the size of TCHAR varies based of the character encoding type. i.e. if unicode is defined, TCHAR is equal to wchar_t, otherwise it is char.

Naturally, if you convert a wide character to a normal char, you can only keep the LSB and may lose some data. This was somehow irritating for me. so I wrote the following code. Its main advantage is doing the conversion without losing any data.

By the way, if you are okay with data loss, then wcstombs does the job.

#include <cstring>
#include <algorithm>
#include <tchar.h>

void lptstr2str(LPTSTR tch, char* &pch) // or (TCHAR* tch, char* &pch)
{
#ifndef UNICODE
    std::memcpy(pch, tch, strlen(tch) + 1);
#else
    size_t n =
        sizeof(TCHAR) / sizeof(char)* wcsnlen(tch, std::string::npos);
    pch = new char[n + 1];
    std::memcpy(pch, tch, n + 1);
    int len = n - std::count(pch, pch + n, NULL);
    std::remove(pch, pch + n, NULL);
    pch[len] = NULL;
#endif
}

Answer 7

I was missing some simple example so here it is:

(for me char* is identical to char[])

LPCTSTR myLPCTSTR = getLPCTSTR();
TCHAR myT[500];
wcscpy(myT,myLPCTSTR);
char myC[500];
sprintf(myC, "%S", myT);

Answer 8

no doubt many (eg, us unix folk) will recoil in horror at the mad-hatter Microserf doublespeak - "if your compiler is in Unicode mode, use LPWSTR or stick a "T_" in front of it, but only if it's a static string, which is the same as an "L", or use T2A() if using ATL, but that's now outdated, or use VARIANT but not if linking with COM/OLE"...).

The "if (sizeof(TCHAR) == sizeof(char))" listed on this page is a logical attempt at a nice solution, but it won't compile - either the if-true won't compile or if-false wont' compile, depending on your compiler flags (Aaargh!). For a write-and-forget portable solution you'll need to resort to the [too-generic named] UNICODE macro. I offer this adaptation of the previous code:

string mfc_to_zstring (CString &sref)
{
    char nojoy[65536];
    char *ptr, *psin = NULL;
    string sot;
    LPCTSTR p = sref;


#if UNICODE
    if (sizeof(TCHAR) != sizeof(char))
    {
        size_t n = wcstombs(NULL, p, 0);
        if (n > 65530)
        {
            psin = new char[n + 1];
            wcstombs(psin, p, n + 1);
            ptr = psin;
        }
        else
        {
            wcstombs(nojoy, p, n + 1);
            ptr = nojoy;
        }

        sot = ptr;
        if (psin != NULL)
            delete psin;
    }
    else
        { std::cerr << "Aaargh! Microsoft horror.\n"; exit(1); }
#else
    if (sizeof(TCHAR) == sizeof(char))
    {
        const char *ptr = p;
        sot = ptr;
    }
    else
      { std::cerr << "Aaargh! You should never see this line\n"; exit(1); }
#endif

    return sot;
}