locale and wchar_t character constant

[Owner]

what's wrong with this code?

#include <locale.h>
#include <stdio.h>
#include <wchar.h>

main(){
setlocale(LC_CTYPE, "");
wchar_t a = L'ÇÑ';
putwchar(a);

}

result

test.c:7: warning: multi-character character constant

 

[Ben Bacarisse]

what's wrong with this code?

#include <locale.h>
#include <stdio.h>
#include <wchar.h>

main(){

int main(void)

setlocale(LC_CTYPE, "");
wchar_t a = L'한';
putwchar(a);

}

result

test.c:7: warning: multi-character character constant

It works for me (as I'd expect). I image that the particular multi-byte
sequence you have is not being recognised by the compiler.

 

[Owner]

int main(void)


It works for me (as I'd expect). I image that the particular multi-byte
sequence you have is not being recognised by the compiler.

So you mean I should switch my tcc compiler to better compiler?

 

[Keith Thompson]

what's wrong with this code?

#include <locale.h>
#include <stdio.h>
#include <wchar.h>

main(){

Should be "int main(void) {", but that's not the problem.

setlocale(LC_CTYPE, "");
wchar_t a = L'한';
putwchar(a);

}

result

test.c:7: warning: multi-character character constant

It's hard to tell. Your article appears to have an euc-kr encoding
(judging by the "Content-Type:" header on your article), but
that could have been introduced by some news software somewhere.
Is your original source file encoded in euc-kr?

Probably your compiler doesn't recognize the character encoding you
used in your source file, and is interpreting the bytes between the
single quotes as two 8-bit characters, producing a multi-character
character constant.

Note that a call to setlocale() (which occurs at run time) cannot
affect the compiler's interpretation of source characters.

See what your compiler's documentation says about source character
sets. You might hae better luck using UTF-8.

 

[Ben Bacarisse]

So you mean I should switch my tcc compiler to better compiler?

Not necessarily. There may be compiler options that will make it work
or, as Keith has suggested, you may be using an encoding that is not
recognised by tcc. It might even as a simple as setting an environment
variable during the compile.

However, tcc is a small project and multi-byte support may well have
been something that's been left to one side. There is a good mailing
tcc list: (e-mail address removed).

 

[Owner]

Should be "int main(void) {", but that's not the problem.


It's hard to tell. Your article appears to have an euc-kr encoding
(judging by the "Content-Type:" header on your article), but
that could have been introduced by some news software somewhere.
Is your original source file encoded in euc-kr?

yes, korean locale

 

[Owner]

what's wrong with this code?

#include <locale.h>
#include <stdio.h>
#include <wchar.h>

main(){
setlocale(LC_CTYPE, "");
wchar_t a = L'ÇÑ';
putwchar(a);

}

result

test.c:7: warning: multi-character character constant

Strangely this below code works then.

#include <locale.h>
#include <stdio.h>
#include <wchar.h>


main(){
wchar_t *a = L"ÇѱÛ";
setlocale(LC_ALL, "ko");

wprintf(L"%s",a);

}

 

[Keith Thompson]

Strangely this below code works then.

#include <locale.h>
#include <stdio.h>
#include <wchar.h>


main(){
wchar_t *a = L"한글";
setlocale(LC_ALL, "ko");

wprintf(L"%s",a);

}

Well, at least the compiler isn't going to warn about the second
case; you're using a string literal rather than a character constant,
so you can have as many characters/bytes/glyphs between them as
you like.

What does wcslen(a) return?

(Note that, in my newsreader, your string literal looks something like
L"[][]"
only fuzzier.)

 

[Owner]

Strangely this below code works then.

#include <locale.h>
#include <stdio.h>
#include <wchar.h>


main(){
wchar_t *a = L"ÇѱÛ";
setlocale(LC_ALL, "ko");

wprintf(L"%s",a);

}

Well, at least the compiler isn't going to warn about the second
case; you're using a string literal rather than a character constant,
so you can have as many characters/bytes/glyphs between them as
you like.

What does wcslen(a) return?

(Note that, in my newsreader, your string literal looks something like
L"[][]"
only fuzzier.)

Solved!

downloaded visual c++ express 2010 and compiled below code
with cl.exe and worked.

#include <locale.h>
#include <stdio.h>
#include <wchar.h>


main(){
wchar_t a = L'ÇÑ';
setlocale(LC_ALL, "kor");

putwchar(a);

}

 

[Owner]

Well, at least the compiler isn't going to warn about the second
case; you're using a string literal rather than a character constant,
so you can have as many characters/bytes/glyphs between them as
you like.

What does wcslen(a) return?

4

(Note that, in my newsreader, your string literal looks something like
L"[][]"
only fuzzier.)

supposed to be korean in 2 letters

 

[Owner]

4

and visual c++ express 2010 cl.exe returns 2 which is correct
length of the string.