I tried to printf with some accented characters such as á é í ó ú:
printf("my name is Seán\n");
The text editor in the DEVC++ IDE displays them fine - i.e the source code looks fine.
I guess I need some library other than stdio.h and maybe some variant of the normal printf.
I'm using IDE Bloodshed DEVC running on Windows XP.
Windows console is generally considered badly broken regarding to character encodings. You can read about this problem here, for example.
The problem is that Windows generally uses the ANSI codepage (which is assuming you are in Western Europe or America Windows-1252), but the console uses the OEM codepage (CP850 under the same assumption).
You have several options:
CharToOem()). The drawback is that if the user redirects the output to a file (> file.txt) and opens the file with e.g. Notepad, he will see it wrong.chcp 1252.wprintf(): You need the TTF console font anyway.Perhaps the best is to use Unicode.
Here's how...
First, manually set your console font to "Consolas" or "Lucida Console" or whichever True-Type Unicode font you can choose ("Raster fonts" may not work, those aren't Unicode fonts, although they may include characters you're interested in).
Next, set the console code page to 65001 (UTF-8) with SetConsoleOutputCP(CP_UTF8).
Then convert your text to UTF-8 (if it's not yet in UTF-8) using WideCharToMultiByte(CP_UTF8, ...).
Finally, call WriteConsoleA() to output the UTF-8 text.
Here's a little function that does all these things for you, it's an "improved" variant of wprintf():
int _wprintf(const wchar_t* format, ...)
{
int r;
static int utf8ModeSet = 0;
static wchar_t* bufWchar = NULL;
static size_t bufWcharCount = 256;
static char* bufMchar = NULL;
static size_t bufMcharCount = 256;
va_list vl;
int mcharCount = 0;
if (utf8ModeSet == 0)
{
if (!SetConsoleOutputCP(CP_UTF8))
{
DWORD err = GetLastError();
fprintf(stderr, "SetConsoleOutputCP(CP_UTF8) failed with error 0x%X\n", err);
utf8ModeSet = -1;
}
else
{
utf8ModeSet = 1;
}
}
if (utf8ModeSet != 1)
{
va_start(vl, format);
r = vwprintf(format, vl);
va_end(vl);
return r;
}
if (bufWchar == NULL)
{
if ((bufWchar = malloc(bufWcharCount * sizeof(wchar_t))) == NULL)
{
return -1;
}
}
for (;;)
{
va_start(vl, format);
r = vswprintf(bufWchar, bufWcharCount, format, vl);
va_end(vl);
if (r < 0)
{
break;
}
if (r + 2 <= bufWcharCount)
{
break;
}
free(bufWchar);
if ((bufWchar = malloc(bufWcharCount * sizeof(wchar_t) * 2)) == NULL)
{
return -1;
}
bufWcharCount *= 2;
}
if (r > 0)
{
if (bufMchar == NULL)
{
if ((bufMchar = malloc(bufMcharCount)) == NULL)
{
return -1;
}
}
for (;;)
{
mcharCount = WideCharToMultiByte(CP_UTF8,
0,
bufWchar,
-1,
bufMchar,
bufMcharCount,
NULL,
NULL);
if (mcharCount > 0)
{
break;
}
if (GetLastError() != ERROR_INSUFFICIENT_BUFFER)
{
return -1;
}
free(bufMchar);
if ((bufMchar = malloc(bufMcharCount * 2)) == NULL)
{
return -1;
}
bufMcharCount *= 2;
}
}
if (mcharCount > 1)
{
DWORD numberOfCharsWritten, consoleMode;
if (GetConsoleMode(GetStdHandle(STD_OUTPUT_HANDLE), &consoleMode))
{
fflush(stdout);
if (!WriteConsoleA(GetStdHandle(STD_OUTPUT_HANDLE),
bufMchar,
mcharCount - 1,
&numberOfCharsWritten,
NULL))
{
return -1;
}
}
else
{
if (fputs(bufMchar, stdout) == EOF)
{
return -1;
}
}
}
return r;
}
Following tests this function:
_wprintf(L"\xA0\xA1\xA2\xA3\xA4\xA5\xA6\xA7"
L"\xA8\xA9\xAA\xAB\xAC\xAD\xAE\xAF"
L"\xB0\xB1\xB2\xB3\xB4\xB5\xB6\xB7"
L"\xB8\xB9\xBA\xBB\xBC\xBD\xBE\xBF"
L"\n"
L"\xC0\xC1\xC2\xC3\xC4\xC5\xC6\xC7"
L"\xC8\xC9\xCA\xCB\xCC\xCD\xCE\xCF"
L"\xD0\xD1\xD2\xD3\xD4\xD5\xD6\xD7"
L"\xD8\xD9\xDA\xDB\xDC\xDD\xDE\xDF"
L"\n"
L"\xE0\xE1\xE2\xE3\xE4\xE5\xE6\xE7"
L"\xE8\xE9\xEA\xEB\xEC\xED\xEE\xEF"
L"\xF0\xF1\xF2\xF3\xF4\xF5\xF6\xF7"
L"\xF8\xF9\xFA\xFB\xFC\xFD\xFE\xFF"
L"\n");
_wprintf(L"\x391\x392\x393\x394\x395\x396\x397"
L"\x398\x399\x39A\x39B\x39C\x39D\x39E\x39F"
L"\x3A0\x3A1\x3A2\x3A3\x3A4\x3A5\x3A6\x3A7"
L"\x3A8\x3A9\x3AA\x3AB\x3AC\x3AD\x3AE\x3AF\x3B0"
L"\n"
L"\x3B1\x3B2\x3B3\x3B4\x3B5\x3B6\x3B7"
L"\x3B8\x3B9\x3BA\x3BB\x3BC\x3BD\x3BE\x3BF"
L"\x3C0\x3C1\x3C2\x3C3\x3C4\x3C5\x3C6\x3C7"
L"\x3C8\x3C9\x3CA\x3CB\x3CC\x3CD\x3CE"
L"\n");
_wprintf(L"\x410\x411\x412\x413\x414\x415\x401\x416\x417"
L"\x418\x419\x41A\x41B\x41C\x41D\x41E\x41F"
L"\x420\x421\x422\x423\x424\x425\x426\x427"
L"\x428\x429\x42A\x42B\x42C\x42D\x42E\x42F"
L"\n"
L"\x430\x431\x432\x433\x434\x435\x451\x436\x437"
L"\x438\x439\x43A\x43B\x43C\x43D\x43E\x43F"
L"\x440\x441\x442\x443\x444\x445\x446\x447"
L"\x448\x449\x44A\x44B\x44C\x44D\x44E\x44F"
L"\n");
And should result in the following text in the console:
¡¢£¤¥¦§¨©ª«¬®¯°±²³´µ¶·¸¹º»¼½¾¿
ÀÁÂÃÄÅÆÇÈÉÊËÌÍÎÏÐÑÒÓÔÕÖ×ØÙÚÛÜÝÞß
àáâãäåæçèéêëìíîïðñòóôõö÷øùúûüýþÿ
ΑΒΓΔΕΖΗΘΙΚΛΜΝΞΟΠΡΣΤΥΦΧΨΩΪΫάέήίΰ
αβγδεζηθικλμνξοπρςστυφχψωϊϋόύώ
АБВГДЕЁЖЗИЙКЛМНОПРСТУФХЦЧШЩЪЫЬЭЮЯ
абвгдеёжзийклмнопрстуфхцчшщъыьэюя
I do not know the encoding in which your IDE stores non-ASCII characters in .c/.cpp files and I do not know what your compiler does when encounters non-ASCII characters. This part you should figure out yourself.
As long as you supply to _wprintf() properly encoded UTF-16 text or call WriteConsoleA() with properly encoded UTF-8 text, things should work.
P.S. Some gory details about console fonts can be found here.
The Windows-1252 (also known as "ANSI") character set used by Windows console mode is not the same as that used by GUI applications. Hence the IDE representation differs from the runtime representation.
A quick-and-dirty solution for your example is:
printf("my name is Se\xe9n\n");
Most solutions to this problem are flawed one way or another and the simplest solution for Windows applications that need extensive multi-language localisation is to write them as GUI apps using Unicode.
