diplodocus.org Git - nmh/blob - sbr/check_charset.c

   1
   2 /*
   3  * check_charset.c -- routines for character sets
   4  *
   5  * $Id$
   6  *
   7  * This code is Copyright (c) 2002, by the authors of nmh.  See the
   8  * COPYRIGHT file in the root directory of the nmh distribution for
   9  * complete copyright information.
  10  */
  11
  12 #include <h/mh.h>
  13 #ifdef HAVE_LANGINFO_H
  14 # include <langinfo.h>
  15 #endif
  16
  17
  18 /*
  19  * Get the current character set
  20  */
  21 char *
  22 get_charset ()
  23 {
  24     char *charset = getenv ("MM_CHARSET");
  25 #if defined(HAVE_NL_LANGINFO) && defined(CODESET)
  26     if (!charset)
  27         charset = norm_charmap(nl_langinfo (CODESET));
  28 #endif
  29     return charset;
  30 }
  31
  32
  33 /*
  34  * Check if we can display a given character set natively.
  35  * We are passed the length of the initial part of the
  36  * string to check, since we want to allow the name of the
  37  * character set to be a substring of a larger string.
  38  */
  39
  40 int
  41 check_charset (char *str, int len)
  42 {
  43     static char *mm_charset = NULL;
  44     static char *alt_charset = NULL;
  45     static int mm_len;
  46     static int alt_len;
  47
  48     /* Cache the name of our default character set */
  49     if (!mm_charset) {
  50         if (!(mm_charset = get_charset ()))
  51             mm_charset = "US-ASCII";
  52         mm_len = strlen (mm_charset);
  53
  54         /* US-ASCII is a subset of the ISO-8859-X and UTF-8 character sets */
  55         if (!strncasecmp("ISO-8859-", mm_charset, 9) ||
  56                 !strcasecmp("UTF-8", mm_charset)) {
  57             alt_charset = "US-ASCII";
  58             alt_len = strlen (alt_charset);
  59         }
  60     }
  61
  62     /* Check if character set is OK */
  63     if ((len == mm_len) && !strncasecmp(str, mm_charset, mm_len))
  64         return 1;
  65     if (alt_charset && (len == alt_len) && !strncasecmp(str, alt_charset, alt_len))
  66         return 1;
  67
  68     return 0;
  69 }
  70
  71
  72 /*
  73  * Return the name of the character set we are
  74  * using for 8bit text.
  75  */
  76 char *
  77 write_charset_8bit (void)
  78 {
  79     static char *mm_charset = NULL;
  80
  81     /*
  82      * Cache the name of the character set to
  83      * use for 8bit text.
  84      */
  85     if (!mm_charset && !(mm_charset = get_charset ()))
  86             mm_charset = "x-unknown";
  87
  88     return mm_charset;
  89 }