1
0
Fork 0
mirror of https://github.com/ruby/ruby.git synced 2022-11-09 12:17:21 -05:00
ruby--ruby/missing/langinfo.c

149 lines
4.8 KiB
C
Raw Normal View History

/* -*- c-file-style: "gnu" -*- */
/*
* This is a quick-and-dirty emulator of the nl_langinfo(CODESET)
* function defined in the Single Unix Specification for those systems
* (FreeBSD, etc.) that don't have one yet. It behaves as if it had
* been called after setlocale(LC_CTYPE, ""), that is it looks at
* the locale environment variables.
*
* http://www.opengroup.org/onlinepubs/7908799/xsh/langinfo.h.html
*
* Please extend it as needed and suggest improvements to the author.
* This emulator will hopefully become redundant soon as
* nl_langinfo(CODESET) becomes more widely implemented.
*
* Since the proposed Li18nux encoding name registry is still not mature,
* the output follows the MIME registry where possible:
*
* http://www.iana.org/assignments/character-sets
*
* A possible autoconf test for the availability of nl_langinfo(CODESET)
* can be found in
*
* http://www.cl.cam.ac.uk/~mgk25/unicode.html#activate
*
* Markus.Kuhn@cl.cam.ac.uk -- 2002-03-11
* Permission to use, copy, modify, and distribute this software
* for any purpose and without fee is hereby granted. The author
* disclaims all warranties with regard to this software.
*
* Latest version:
*
* http://www.cl.cam.ac.uk/~mgk25/ucs/langinfo.c
*/
#include "ruby/missing.h"
#include <stdlib.h>
#include <string.h>
#if defined _WIN32 || defined __CYGWIN__
#include <windows.h>
#if defined _WIN32
#define strncasecmp strnicmp
#endif
#endif
#ifdef HAVE_LANGINFO_H
#include "langinfo.h"
#endif
#define C_CODESET "US-ASCII" /* Return this as the encoding of the
* C/POSIX locale. Could as well one day
* become "UTF-8". */
#if defined _WIN32 || defined __CYGWIN__
#define JA_CODESET "Windows-31J"
#else
#define JA_CODESET "EUC-JP"
#endif
#define digit(x) ((x) >= '0' && (x) <= '9')
#define strstart(s, n) (strncasecmp(s, n, strlen(n)) == 0)
static char buf[16];
const char *
nl_langinfo_codeset(void)
{
const char *l, *p;
int n;
if (((l = getenv("LC_ALL")) && *l) ||
((l = getenv("LC_CTYPE")) && *l) ||
((l = getenv("LANG")) && *l)) {
/* check standardized locales */
if (!strcmp(l, "C") || !strcmp(l, "POSIX"))
return C_CODESET;
/* check for encoding name fragment */
p = strchr(l, '.');
if (!p++) p = l;
if (strstart(p, "UTF"))
return "UTF-8";
if ((n = 5, strstart(p, "8859-")) || (n = 9, strstart(p, "ISO-8859-"))) {
if (digit(p[n])) {
p += n;
memcpy(buf, "ISO-8859-\0\0", 12);
buf[9] = *p++;
if (digit(*p)) buf[10] = *p++;
return buf;
}
}
if (strstart(p, "KOI8-R")) return "KOI8-R";
if (strstart(p, "KOI8-U")) return "KOI8-U";
if (strstart(p, "620")) return "TIS-620";
if (strstart(p, "2312")) return "GB2312";
if (strstart(p, "HKSCS")) return "Big5HKSCS"; /* no MIME charset */
if (strstart(p, "BIG5")) return "Big5";
if (strstart(p, "GBK")) return "GBK"; /* no MIME charset */
if (strstart(p, "18030")) return "GB18030"; /* no MIME charset */
if (strstart(p, "Shift_JIS") || strstart(p, "SJIS")) return "Windows-31J";
/* check for conclusive modifier */
if (strstart(p, "euro")) return "ISO-8859-15";
/* check for language (and perhaps country) codes */
if (strstart(l, "zh_TW")) return "Big5";
if (strstart(l, "zh_HK")) return "Big5HKSCS"; /* no MIME charset */
if (strstart(l, "zh")) return "GB2312";
if (strstart(l, "ja")) return JA_CODESET;
if (strstart(l, "ko")) return "EUC-KR";
if (strstart(l, "ru")) return "KOI8-R";
if (strstart(l, "uk")) return "KOI8-U";
if (strstart(l, "pl") || strstart(l, "hr") ||
strstart(l, "hu") || strstart(l, "cs") ||
strstart(l, "sk") || strstart(l, "sl")) return "ISO-8859-2";
if (strstart(l, "eo") || strstart(l, "mt")) return "ISO-8859-3";
if (strstart(l, "el")) return "ISO-8859-7";
if (strstart(l, "he")) return "ISO-8859-8";
if (strstart(l, "tr")) return "ISO-8859-9";
if (strstart(l, "th")) return "TIS-620"; /* or ISO-8859-11 */
if (strstart(l, "lt")) return "ISO-8859-13";
if (strstart(l, "cy")) return "ISO-8859-14";
if (strstart(l, "ro")) return "ISO-8859-2"; /* or ISO-8859-16 */
if (strstart(l, "am") || strstart(l, "vi")) return "UTF-8";
/* Send me further rules if you like, but don't forget that we are
* *only* interested in locale naming conventions on platforms
* that do not already provide an nl_langinfo(CODESET) implementation. */
}
return NULL;
}
#ifdef HAVE_LANGINFO_H
char *nl_langinfo(nl_item item)
{
const char *codeset;
if (item != CODESET)
return NULL;
codeset = nl_langinfo_codeset();
if (!codeset) codeset = C_CODESET;
return (char *)codeset;
}
#endif
/* For a demo, compile with "gcc -W -Wall -o langinfo -D TEST langinfo.c" */
#ifdef TEST
#include <stdio.h>
int main()
{
printf("%s\n", nl_langinfo(CODESET));
return 0;
}
#endif