glib/gunicollate.c

   1 /* gunicollate.c - Collation
   2  *
   3  *  Copyright 2001 Red Hat, Inc.
   4  *
   5  * The Gnome Library is free software; you can redistribute it and/or
   6  * modify it under the terms of the GNU Lesser General Public License as
   7  * published by the Free Software Foundation; either version 2 of the
   8  * License, or (at your option) any later version.
   9  *
  10  * The Gnome Library is distributed in the hope that it will be useful,
  11  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  12  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  13  * Lesser General Public License for more details.
  14  *
  15  * You should have received a copy of the GNU Lesser General Public
  16  * License along with the Gnome Library; see the file COPYING.LIB.  If not,
  17  * write to the Free Software Foundation, Inc., 59 Temple Place - Suite 330,
  18  *   Boston, MA 02111-1307, USA.
  19  */
  20
  21 #include "config.h"
  22
  23 #include <locale.h>
  24 #include <string.h>
  25 #ifdef __STDC_ISO_10646__
  26 #include <wchar.h>
  27 #endif
  28
  29 #include "glib.h"
  30 #include "gunicodeprivate.h"
  31 #include "galias.h"
  32
  33 /**
  34  * g_utf8_collate:
  35  * @str1: a UTF-8 encoded string
  36  * @str2: a UTF-8 encoded string
  37  *
  38  * Compares two strings for ordering using the linguistically
  39  * correct rules for the current locale. When sorting a large
  40  * number of strings, it will be significantly faster to
  41  * obtain collation keys with g_utf8_collate_key() and
  42  * compare the keys with strcmp() when
  43  * sorting instead of sorting the original strings.
  44  *
  45  * Return value: &lt; 0 if @str1 compares before @str2,
  46  *   0 if they compare equal, &gt; 0 if @str1 compares after @str2.
  47  **/
  48 gint
  49 g_utf8_collate (const gchar *str1,
  50                 const gchar *str2)
  51 {
  52   gint result;
  53
  54 #ifdef __STDC_ISO_10646__
  55
  56   gunichar *str1_norm;
  57   gunichar *str2_norm;
  58
  59   g_return_val_if_fail (str1 != NULL, 0);
  60   g_return_val_if_fail (str2 != NULL, 0);
  61
  62   str1_norm = _g_utf8_normalize_wc (str1, -1, G_NORMALIZE_ALL_COMPOSE);
  63   str2_norm = _g_utf8_normalize_wc (str2, -1, G_NORMALIZE_ALL_COMPOSE);
  64
  65   result = wcscoll ((wchar_t *)str1_norm, (wchar_t *)str2_norm);
  66
  67   g_free (str1_norm);
  68   g_free (str2_norm);
  69
  70 #else /* !__STDC_ISO_10646__ */
  71
  72   const gchar *charset;
  73   gchar *str1_norm;
  74   gchar *str2_norm;
  75
  76   g_return_val_if_fail (str1 != NULL, 0);
  77   g_return_val_if_fail (str2 != NULL, 0);
  78
  79   str1_norm = g_utf8_normalize (str1, -1, G_NORMALIZE_ALL_COMPOSE);
  80   str2_norm = g_utf8_normalize (str2, -1, G_NORMALIZE_ALL_COMPOSE);
  81
  82   if (g_get_charset (&charset))
  83     {
  84       result = strcoll (str1_norm, str2_norm);
  85     }
  86   else
  87     {
  88       gchar *str1_locale = g_convert (str1_norm, -1, charset, "UTF-8", NULL, NULL, NULL);
  89       gchar *str2_locale = g_convert (str2_norm, -1, charset, "UTF-8", NULL, NULL, NULL);
  90
  91       if (str1_locale && str2_locale)
  92         result =  strcoll (str1_locale, str2_locale);
  93       else if (str1_locale)
  94         result = -1;
  95       else if (str2_locale)
  96         result = 1;
  97       else
  98         result = strcmp (str1_norm, str2_norm);
  99
 100       g_free (str1_locale);
 101       g_free (str2_locale);
 102     }
 103
 104   g_free (str1_norm);
 105   g_free (str2_norm);
 106
 107 #endif /* __STDC_ISO_10646__ */
 108
 109   return result;
 110 }
 111
 112 #ifdef __STDC_ISO_10646__
 113 /* We need UTF-8 encoding of numbers to encode the weights if
 114  * we are using wcsxfrm. However, we aren't encoding Unicode
 115  * characters, so we can't simply use g_unichar_to_utf8.
 116  *
 117  * The following routine is taken (with modification) from GNU
 118  * libc's strxfrm routine:
 119  *
 120  * Copyright (C) 1995-1999,2000,2001 Free Software Foundation, Inc.
 121  * Written by Ulrich Drepper <drepper@cygnus.com>, 1995.
 122  */
 123 static inline int
 124 utf8_encode (char *buf, wchar_t val)
 125 {
 126   int retval;
 127
 128   if (val < 0x80)
 129     {
 130       if (buf)
 131         *buf++ = (char) val;
 132       retval = 1;
 133     }
 134   else
 135     {
 136       int step;
 137
 138       for (step = 2; step < 6; ++step)
 139         if ((val & (~(guint32)0 << (5 * step + 1))) == 0)
 140           break;
 141       retval = step;
 142
 143       if (buf)
 144         {
 145           *buf = (unsigned char) (~0xff >> step);
 146           --step;
 147           do
 148             {
 149               buf[step] = 0x80 | (val & 0x3f);
 150               val >>= 6;
 151             }
 152           while (--step > 0);
 153           *buf |= val;
 154         }
 155     }
 156
 157   return retval;
 158 }
 159 #endif /* __STDC_ISO_10646__ */
 160
 161 /**
 162  * g_utf8_collate_key:
 163  * @str: a UTF-8 encoded string.
 164  * @len: length of @str, in bytes, or -1 if @str is nul-terminated.
 165  *
 166  * Converts a string into a collation key that can be compared
 167  * with other collation keys using strcmp().
 168  * The results of comparing the collation keys of two strings
 169  * with strcmp() will always be the same as
 170  * comparing the two original keys with g_utf8_collate().
 171  *
 172  * Return value: a newly allocated string. This string should
 173  *   be freed with g_free() when you are done with it.
 174  **/
 175 gchar *
 176 g_utf8_collate_key (const gchar *str,
 177                     gssize       len)
 178 {
 179   gchar *result;
 180   size_t xfrm_len;
 181
 182 #ifdef __STDC_ISO_10646__
 183
 184   gunichar *str_norm;
 185   wchar_t *result_wc;
 186   size_t i;
 187   size_t result_len = 0;
 188
 189   g_return_val_if_fail (str != NULL, NULL);
 190
 191   str_norm = _g_utf8_normalize_wc (str, len, G_NORMALIZE_ALL_COMPOSE);
 192
 193   setlocale (LC_COLLATE, "");
 194
 195   xfrm_len = wcsxfrm (NULL, (wchar_t *)str_norm, 0);
 196   result_wc = g_new (wchar_t, xfrm_len + 1);
 197   wcsxfrm (result_wc, (wchar_t *)str_norm, xfrm_len + 1);
 198
 199   for (i=0; i < xfrm_len; i++)
 200     result_len += utf8_encode (NULL, result_wc[i]);
 201
 202   result = g_malloc (result_len + 1);
 203   result_len = 0;
 204   for (i=0; i < xfrm_len; i++)
 205     result_len += utf8_encode (result + result_len, result_wc[i]);
 206
 207   result[result_len] = '\0';
 208
 209   g_free (result_wc);
 210   g_free (str_norm);
 211
 212   return result;
 213 #else /* !__STDC_ISO_10646__ */
 214
 215   const gchar *charset;
 216   gchar *str_norm;
 217
 218   g_return_val_if_fail (str != NULL, NULL);
 219
 220   str_norm = g_utf8_normalize (str, len, G_NORMALIZE_ALL_COMPOSE);
 221
 222   if (g_get_charset (&charset))
 223     {
 224       xfrm_len = strxfrm (NULL, str_norm, 0);
 225       result = g_malloc (xfrm_len + 1);
 226       strxfrm (result, str_norm, xfrm_len + 1);
 227     }
 228   else
 229     {
 230       gchar *str_locale = g_convert (str_norm, -1, charset, "UTF-8", NULL, NULL, NULL);
 231
 232       if (str_locale)
 233         {
 234           xfrm_len = strxfrm (NULL, str_locale, 0);
 235           if (xfrm_len < 0 || xfrm_len >= G_MAXINT - 2)
 236             {
 237               g_free (str_locale);
 238               str_locale = NULL;
 239             }
 240         }
 241       if (str_locale)
 242         {
 243           result = g_malloc (xfrm_len + 2);
 244           result[0] = 'A';
 245           strxfrm (result + 1, str_locale, xfrm_len + 1);
 246
 247           g_free (str_locale);
 248         }
 249       else
 250         {
 251           xfrm_len = strlen (str_norm);
 252           result = g_malloc (xfrm_len + 2);
 253           result[0] = 'B';
 254           memcpy (result + 1, str_norm, xfrm_len);
 255           result[xfrm_len+1] = '\0';
 256         }
 257     }
 258
 259   g_free (str_norm);
 260 #endif /* __STDC_ISO_10646__ */
 261
 262   return result;
 263 }
 264
 265 #define __G_UNICOLLATE_C__
 266 #include "galiasdef.c"