#define _UNICODE ... C and C++, works on nearly every compiler I've used. It's a switch for certain defines and macros in the libraries.Originally Posted by msh
#define UNICODE ... Windows, automatically switches between ascii and unicode versions of almost every function call documented in the SDK.
I did not realize we were on a non-windows setup... sorry about that.
Code:#include <stdio.h> #include <wchar.h> #include <string.h> #include <stdlib.h> #include <locale.h> #include <limits.h> /* taken from libc manual */ size_t mbslen (const char *s) { mbstate_t state; size_t result = 0; size_t nbytes; memset(&state, '\0', sizeof(state)); while ((nbytes = mbrlen(s, MB_LEN_MAX, &state)) > 0) { if (nbytes >= (size_t)-2) return (size_t)-1; s += nbytes; ++result; } return result; } int main(void) { char input[80] = {0}; wchar_t output[80] = {0}; #if 1 const char *l = setlocale(LC_CTYPE, ""); if (!l) perror("setlocale"); else printf("locale = %s\n", l); #else puts("Using C locale"); #endif printf("Give me something: "); scanf("%s", input); int n1 = mbstowcs(output, input, 80); if (n1 == -1) perror("mbstowcs"); int n2 = wcslen(output); int n3 = strlen(input); int n4 = mbslen(input); if (n4 == -1) perror("mbslen"); printf("mbstowcs(output,input)=%d\n" "wcslen(output)=%d\n" "strlen(input)=%d\n" "mbslen(input)=%d\n", n1, n2, n3, n4); if (*input) { printf("input = "); const char *p = input; for (; *p; ++p) printf("0x%02X,", (unsigned char)*p); puts("\b "); } if (*output) { printf("output = "); const wchar_t *p = output; for (; *p; ++p) printf("0x%X,", (unsigned)*p); puts("\b "); } return 0; }Code:locale = en_US.utf8 Give me something: œ∑Ω mbstowcs(output,input)=3 wcslen(output)=3 strlen(input)=7 mbslen(input)=3 input = 0xC5,0x93,0xE2,0x88,0x91,0xCE,0xA9 output = 0x153,0x2211,0x3A9On my Fedora 14 box, not calling setlocale causes mbstowcs to fail - due to bytes that don't belong to the C locale.Code:Using C locale Give me something: œ∑Ω mbstowcs: Invalid or incomplete multibyte or wide character mbslen: Invalid or incomplete multibyte or wide character mbstowcs(output,input)=-1 wcslen(output)=0 strlen(input)=7 mbslen(input)=-1 input = 0xC5,0x93,0xE2,0x88,0x91,0xCE,0xA9
On tabstop's box, mbstowcs seems to just copy each char to a wchar_t piecemeal under the C locale.
gg
Hey, a big thanks to all of you for taking the time to test and post code or otherwise offer suggestions! And, yes, I did get what I was after.Here's the final version, tested and working, exactly as I intended:
It works for any Unicode character -- tested on some characters from non-Latin scripts far outside the ASCII 7-bit range. It returns, as I intended, the Unicode code point (in decimal form) of any Unicode character, and the Unicode character corresponding to any Unicode code point typed in.Code:#include <stdio.h> #include <wchar.h> #include <locale.h> int main (int argc, const char * argv[]) { setlocale(LC_ALL, ""); wchar_t UnicodeChar=0; int c2=0; wprintf(L"Type any Unicode character here:"); wscanf(L"%lc",&UnicodeChar); wprintf(L"Type any Unicode numeric value:"); scanf("%d",&c2); wprintf(L"\nThe character '%lc' has a Unicode numeric value of '%d.'\nThe numeric value '%d' represents the Unicode character '%lc.'",UnicodeChar,UnicodeChar,c2,c2); return 0; }
Yes, Codeplug, you are right. Thanks for pointing out that issue, corrected by cas's modification to the second wscanf:
Thanks for that, cas!
Thanks, tabstop, for reminding me to run code in the Mac terminal, not just in Xcode: I was getting nothing but failures building and running my code in Xcode! But compiling and running my code in terminal worked. Of course, that begs the question: Why can't Xcode reproduce the success in Terminal? Probably an Xcode configuration issue, I guess.
