1/* -*- c-file-style: "gnu" -*- */
2/*
3 * This is a quick-and-dirty emulator of the nl_langinfo(CODESET)
4 * function defined in the Single Unix Specification for those systems
5 * (FreeBSD, etc.) that don't have one yet. It behaves as if it had
6 * been called after setlocale(LC_CTYPE, ""), that is it looks at
7 * the locale environment variables.
8 *
9 * http://www.opengroup.org/onlinepubs/7908799/xsh/langinfo.h.html
10 *
11 * Please extend it as needed and suggest improvements to the author.
12 * This emulator will hopefully become redundant soon as
13 * nl_langinfo(CODESET) becomes more widely implemented.
14 *
15 * Since the proposed Li18nux encoding name registry is still not mature,
16 * the output follows the MIME registry where possible:
17 *
18 *   http://www.iana.org/assignments/character-sets
19 *
20 * A possible autoconf test for the availability of nl_langinfo(CODESET)
21 * can be found in
22 *
23 *   http://www.cl.cam.ac.uk/~mgk25/unicode.html#activate
24 *
25 * Markus.Kuhn@cl.cam.ac.uk -- 2002-03-11
26 * Permission to use, copy, modify, and distribute this software
27 * for any purpose and without fee is hereby granted. The author
28 * disclaims all warranties with regard to this software.
29 *
30 * Latest version:
31 *
32 *   http://www.cl.cam.ac.uk/~mgk25/ucs/langinfo.c
33 */
34
35#include "ruby/missing.h"
36#include <stdlib.h>
37#include <string.h>
38#if defined _WIN32 || defined __CYGWIN__
39#include <windows.h>
40#if defined _WIN32 && !defined strncasecmp
41#define strncasecmp strnicmp
42#endif
43#endif
44#ifdef HAVE_LANGINFO_H
45#include "langinfo.h"
46#endif
47
48#define C_CODESET "US-ASCII"     /* Return this as the encoding of the
49				  * C/POSIX locale. Could as well one day
50				  * become "UTF-8". */
51
52#if defined _WIN32 || defined __CYGWIN__
53#define JA_CODESET "Windows-31J"
54#else
55#define JA_CODESET "EUC-JP"
56#endif
57
58#define digit(x) ((x) >= '0' && (x) <= '9')
59#define strstart(s, n) (strncasecmp((s), (n), strlen(n)) == 0)
60
61static char buf[16];
62
63const char *
64nl_langinfo_codeset(void)
65{
66  const char *l, *p;
67  int n;
68
69  if (((l = getenv("LC_ALL"))   && *l) ||
70      ((l = getenv("LC_CTYPE")) && *l) ||
71      ((l = getenv("LANG"))     && *l)) {
72    /* check standardized locales */
73    if (!strcmp(l, "C") || !strcmp(l, "POSIX"))
74      return C_CODESET;
75    /* check for encoding name fragment */
76    p = strchr(l, '.');
77    if (!p++) p = l;
78    if (strstart(p, "UTF"))
79	return "UTF-8";
80    if ((n = 5, strstart(p, "8859-")) || (n = 9, strstart(p, "ISO-8859-"))) {
81      if (digit(p[n])) {
82	p += n;
83	memcpy(buf, "ISO-8859-\0\0", 12);
84	buf[9] = *p++;
85	if (digit(*p)) buf[10] = *p++;
86	return buf;
87      }
88    }
89    if (strstart(p, "KOI8-R")) return "KOI8-R";
90    if (strstart(p, "KOI8-U")) return "KOI8-U";
91    if (strstart(p, "620")) return "TIS-620";
92    if (strstart(p, "2312")) return "GB2312";
93    if (strstart(p, "HKSCS")) return "Big5HKSCS";   /* no MIME charset */
94    if (strstart(p, "BIG5")) return "Big5";
95    if (strstart(p, "GBK")) return "GBK";           /* no MIME charset */
96    if (strstart(p, "18030")) return "GB18030";     /* no MIME charset */
97    if (strstart(p, "Shift_JIS") || strstart(p, "SJIS")) return "Windows-31J";
98    /* check for conclusive modifier */
99    if (strstart(p, "euro")) return "ISO-8859-15";
100    /* check for language (and perhaps country) codes */
101    if (strstart(l, "zh_TW")) return "Big5";
102    if (strstart(l, "zh_HK")) return "Big5HKSCS";   /* no MIME charset */
103    if (strstart(l, "zh")) return "GB2312";
104    if (strstart(l, "ja")) return JA_CODESET;
105    if (strstart(l, "ko")) return "EUC-KR";
106    if (strstart(l, "ru")) return "KOI8-R";
107    if (strstart(l, "uk")) return "KOI8-U";
108    if (strstart(l, "pl") || strstart(l, "hr") ||
109	strstart(l, "hu") || strstart(l, "cs") ||
110	strstart(l, "sk") || strstart(l, "sl")) return "ISO-8859-2";
111    if (strstart(l, "eo") || strstart(l, "mt")) return "ISO-8859-3";
112    if (strstart(l, "el")) return "ISO-8859-7";
113    if (strstart(l, "he")) return "ISO-8859-8";
114    if (strstart(l, "tr")) return "ISO-8859-9";
115    if (strstart(l, "th")) return "TIS-620";      /* or ISO-8859-11 */
116    if (strstart(l, "lt")) return "ISO-8859-13";
117    if (strstart(l, "cy")) return "ISO-8859-14";
118    if (strstart(l, "ro")) return "ISO-8859-2";   /* or ISO-8859-16 */
119    if (strstart(l, "am") || strstart(l, "vi")) return "UTF-8";
120    /* Send me further rules if you like, but don't forget that we are
121     * *only* interested in locale naming conventions on platforms
122     * that do not already provide an nl_langinfo(CODESET) implementation. */
123  }
124  return NULL;
125}
126
127#ifdef HAVE_LANGINFO_H
128char *nl_langinfo(nl_item item)
129{
130  const char *codeset;
131  if (item != CODESET)
132    return NULL;
133  codeset = nl_langinfo_codeset();
134  if (!codeset) codeset = C_CODESET;
135  return (char *)codeset;
136}
137#endif
138
139/* For a demo, compile with "gcc -W -Wall -o langinfo -D TEST langinfo.c" */
140
141#ifdef TEST
142#include <stdio.h>
143int main()
144{
145  printf("%s\n", nl_langinfo(CODESET));
146  return 0;
147}
148#endif
149