Take wchar_t and put into char? - c++

Question

i've tried a few things and haven't yet been able to figure out how to get const wchar_t *text (shown bellow) to pass into the variable StoreText (shown below). What am i doing wrong?
void KeyboardComplete(int localClientNum, const wchar_t *text, unsigned int len)
{
char* StoreText = text; //This is where error occurs
}

Answer 1

You cannot directly assign a wchar_t* to a char*, as they are different and incompatible data types.
If StoreText needs to point at the same memory address that text is pointing at, such as if you are planning on looping through the individual bytes of the text data, then a simple type-cast will suffice:
char* StoreText = (char*)text;
However, if StoreText is expected to point to its own separate copy of the character data, then you would need to convert the wide character data into narrow character data instead. Such as by:
using the WideCharToMultiByte() function on Windows:
void KeyboardComplete(int localClientNum, const wchar_t *text, unsigned int len)
{
int StoreTextLen = 1 + WideCharToMultiByte(CP_ACP, 0, text, len, NULL, 0, NULL, NULL);
std::vector<char> StoreTextBuffer(StoreTextLen);
WideCharToMultiByte(CP_ACP, 0, text, len, &StoreTextBuffer[0], StoreTextLen, NULL, NULL);
char* StoreText = &StoreText[0];
//...
}
using the std::wcsrtombs() function:
#include <cwchar>
void KeyboardComplete(int localClientNum, const wchar_t *text, unsigned int len)
{
std::mbstate_t state = std::mbstate_t();
int StoreTextLen = 1 + std::wcsrtombs(NULL, &text, 0, &state);
std::vector<char> StoreTextBuffer(StoreTextLen);
std::wcsrtombs(&StoreTextBuffer[0], &text, StoreTextLen, &state);
char *StoreText = &StoreTextBuffer[0];
//...
}
using the std::wstring_convert class (C++11 and later):
#include <locale>
void KeyboardComplete(int localClientNum, const wchar_t *text, unsigned int len)
{
std::wstring_convert<std::codecvt<wchar_t, char, std::mbstate_t>> conv;
std::string StoreTextBuffer = conv.to_bytes(text, text+len);
char *StoreText = &StoreTextBuffer[0];
//...
}
using similar conversions from the ICONV or ICU library.

Answer 2

First of all, for strings you should use std::wstring/std::string instead of raw pointers.
The C++11 Locale (http://en.cppreference.com/w/cpp/locale) library can be used to convert wide string to narrow string.
I wrote a wrapper function below and have used it for years. Hope it will be helpful to you, too.
#include <string>
#include <locale>
#include <codecvt>
std::string WstringToString(const std::wstring & wstr, const std::locale & loc /*= std::locale()*/)
{
std::string buf(wstr.size(), 0);
std::use_facet<std::ctype<wchar_t>>(loc).narrow(wstr.c_str(), wstr.c_str() + wstr.size(), '?', &buf[0]);
return buf;
}

Answer 3

wchar_t is a wide character. It is typically 16 or 32 bits per character, but this is system dependent.
char is a good ol' CHAR_BIT-sized data type. Again, how big it is is system dependent. Most likely it's going to be one byte, but I can't think of a reason why CHAR_BIT can't be 16 or 32 bits, making it the same size as wchar_t.
If they are different sizes, a direct assignment is doomed. For example an 8 bit char will see 2 characters, and quite likely 2 completely unrelated characters, for every 1 character in a 16 bit wchar_t. This would be bad.
Second, even if they are the same size, they may have different encodings. For example, the numeric value assigned to the letter 'A' may be different for the char and the wchar_t. It could be 65 in char and 16640 in wchar_t.
To make any sense in the different data type char and wchar_t will need to be translated to the other's encoding. std::wstring_convert will often perform this translation for you, but look into the locale library for more complicated translations. Both require a compiler supporting C++11 or better. In previous C++ Standards, a small army of functions provided conversion support. Third party libraries such as Boost::locale are helpful to unify and provide wider support.
Conversion functions are supplied by the operating system to translate between the encoding used by the OS and other common encodings.

Answer 4

You have to do a cast, you can do this:
char* StoreText = (char*)text;
I think this may work.
But you can use the wcstombs function of cstdlib library.
char someText[12];
wcstombs(StoreText,text, 12);
Last parameter most be a number of byte available in the array pointed.