Eli Zaretskii wrote:

> Only if that encoding matches the current user's codepage.

No. NO. *NO*. Try. I already showed how and i can repeat it AGAIN:

#include <stdio.h>
int main (int argc, char ** argv) {
     printf("argv[1]: %s\n", argv[1]);
     return 0;
}

gcc.EXE (GCC) 3.4.5 (mingw-vista special r3)

gcc -o test test.c

emacs -Q

M-x eval-expression
(setq coding-system-for-write 'utf-8)
M-x grep
test.exe "Šešios žąsys su šešiais žąsyčiais" >> out.txt

M-x eval-expression
(setq coding-system-for-write 'cp1257)
M-x grep
test.exe "Šešios žąsys su šešiais žąsyčiais" >> out.txt


M-x eval-expression
(setq coding-system-for-write 'utf-8)
M-x grep
test.exe "С новым годом Ели Заретски!" >> out.txt


M-x eval-expression
(setq coding-system-for-write 'cp1251)
M-x grep
test.exe "С новым годом Ели Заретски!" >> out.txt


M-x eval-expression
(setq coding-system-for-write 'utf-8)
M-x grep
test.exe "Χρήση εθνικών και ειδικών χαρακτήρων" >> out.txt

M-x eval-expression
(setq coding-system-for-write 'iso-8859-7)
M-x grep
test.exe "Χρήση εθνικών και ειδικών χαρακτήρων" >> out.txt


File 'out.txt' attached.