1 /* gunicollate.c - Collation
3 * Copyright 2001,2005 Red Hat, Inc.
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.
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.
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.
25 #ifdef __STDC_ISO_10646__
30 #include "gunicodeprivate.h"
35 * @str1: a UTF-8 encoded string
36 * @str2: a UTF-8 encoded string
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.
45 * Return value: < 0 if @str1 compares before @str2,
46 * 0 if they compare equal, > 0 if @str1 compares after @str2.
49 g_utf8_collate (const gchar *str1,
54 #ifdef __STDC_ISO_10646__
59 g_return_val_if_fail (str1 != NULL, 0);
60 g_return_val_if_fail (str2 != NULL, 0);
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);
65 result = wcscoll ((wchar_t *)str1_norm, (wchar_t *)str2_norm);
70 #else /* !__STDC_ISO_10646__ */
76 g_return_val_if_fail (str1 != NULL, 0);
77 g_return_val_if_fail (str2 != NULL, 0);
79 str1_norm = g_utf8_normalize (str1, -1, G_NORMALIZE_ALL_COMPOSE);
80 str2_norm = g_utf8_normalize (str2, -1, G_NORMALIZE_ALL_COMPOSE);
82 if (g_get_charset (&charset))
84 result = strcoll (str1_norm, str2_norm);
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);
91 if (str1_locale && str2_locale)
92 result = strcoll (str1_locale, str2_locale);
98 result = strcmp (str1_norm, str2_norm);
100 g_free (str1_locale);
101 g_free (str2_locale);
107 #endif /* __STDC_ISO_10646__ */
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.
117 * The following routine is taken (with modification) from GNU
118 * libc's strxfrm routine:
120 * Copyright (C) 1995-1999,2000,2001 Free Software Foundation, Inc.
121 * Written by Ulrich Drepper <drepper@cygnus.com>, 1995.
124 utf8_encode (char *buf, wchar_t val)
138 for (step = 2; step < 6; ++step)
139 if ((val & (~(guint32)0 << (5 * step + 1))) == 0)
145 *buf = (unsigned char) (~0xff >> step);
149 buf[step] = 0x80 | (val & 0x3f);
159 #endif /* __STDC_ISO_10646__ */
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.
166 * Converts a string into a collation key that can be compared
167 * with other collation keys produced by the same function using
169 * The results of comparing the collation keys of two strings
170 * with strcmp() will always be the same as
171 * comparing the two original keys with g_utf8_collate().
173 * Return value: a newly allocated string. This string should
174 * be freed with g_free() when you are done with it.
177 g_utf8_collate_key (const gchar *str,
183 #ifdef __STDC_ISO_10646__
188 size_t result_len = 0;
190 g_return_val_if_fail (str != NULL, NULL);
192 str_norm = _g_utf8_normalize_wc (str, len, G_NORMALIZE_ALL_COMPOSE);
194 setlocale (LC_COLLATE, "");
196 xfrm_len = wcsxfrm (NULL, (wchar_t *)str_norm, 0);
197 result_wc = g_new (wchar_t, xfrm_len + 1);
198 wcsxfrm (result_wc, (wchar_t *)str_norm, xfrm_len + 1);
200 for (i=0; i < xfrm_len; i++)
201 result_len += utf8_encode (NULL, result_wc[i]);
203 result = g_malloc (result_len + 1);
205 for (i=0; i < xfrm_len; i++)
206 result_len += utf8_encode (result + result_len, result_wc[i]);
208 result[result_len] = '\0';
214 #else /* !__STDC_ISO_10646__ */
216 const gchar *charset;
219 g_return_val_if_fail (str != NULL, NULL);
221 str_norm = g_utf8_normalize (str, len, G_NORMALIZE_ALL_COMPOSE);
223 if (g_get_charset (&charset))
225 xfrm_len = strxfrm (NULL, str_norm, 0);
226 result = g_malloc (xfrm_len + 1);
227 strxfrm (result, str_norm, xfrm_len + 1);
231 gchar *str_locale = g_convert (str_norm, -1, charset, "UTF-8", NULL, NULL, NULL);
235 xfrm_len = strxfrm (NULL, str_locale, 0);
236 if (xfrm_len < 0 || xfrm_len >= G_MAXINT - 2)
244 result = g_malloc (xfrm_len + 2);
246 strxfrm (result + 1, str_locale, xfrm_len + 1);
252 xfrm_len = strlen (str_norm);
253 result = g_malloc (xfrm_len + 2);
255 memcpy (result + 1, str_norm, xfrm_len);
256 result[xfrm_len+1] = '\0';
261 #endif /* __STDC_ISO_10646__ */
266 /* This is a collation key that is very very likely to sort before any
267 collation key that libc strxfrm generates. We use this before any
268 special case (dot or number) to make sure that its sorted before
271 #define COLLATION_SENTINEL "\1\1\1"
274 * g_utf8_collate_key_for_filename:
275 * @str: a UTF-8 encoded string.
276 * @len: length of @str, in bytes, or -1 if @str is nul-terminated.
278 * Converts a string into a collation key that can be compared
279 * with other collation keys produced by the same function using strcmp().
281 * In order to sort filenames correctly, this function treats the dot '.'
282 * as a special case. Most dictionary orderings seem to consider it
283 * insignificant, thus producing the ordering "event.c" "eventgenerator.c"
284 * "event.h" instead of "event.c" "event.h" "eventgenerator.c". Also, we
285 * would like to treat numbers intelligently so that "file1" "file10" "file5"
286 * is sorted as "file1" "file5" "file10".
288 * Return value: a newly allocated string. This string should
289 * be freed with g_free() when you are done with it.
294 g_utf8_collate_key_for_filename (const gchar *str,
309 * Split the filename into collatable substrings which do
310 * not contain [.0-9] and special-cased substrings. The collatable
311 * substrings are run through the normal g_utf8_collate_key() and the
312 * resulting keys are concatenated with keys generated from the
313 * special-cased substrings.
315 * Special cases: Dots are handled by replacing them with '\1' which
316 * implies that short dot-delimited substrings are before long ones,
323 * Numbers are handled by prepending to each number d-1 superdigits
324 * where d = number of digits in the number and SUPERDIGIT is a
325 * character with an integer value higher than any digit (for instance
326 * ':'). This ensures that single-digit numbers are sorted before
327 * double-digit numbers which in turn are sorted separately from
328 * triple-digit numbers, etc. To avoid strange side-effects when
329 * sorting strings that already contain SUPERDIGITs, a '\2'
330 * is also prepended, like this
336 * file\2::100 (file100)
337 * file:foo (file:foo)
339 * This has the side-effect of sorting numbers before everything else (except
340 * dots), but this is probably OK.
342 * Leading digits are ignored when doing the above. To discriminate
343 * numbers which differ only in the number of leading digits, we append
344 * the number of leading digits as a byte at the very end of the collation
347 * To try avoid conflict with any collation key sequence generated by libc we
348 * start each switch to a special cased part with a sentinel that hopefully
349 * will sort before anything libc will generate.
355 result = g_string_sized_new (len * 2);
356 append = g_string_sized_new (0);
360 /* No need to use utf8 functions, since we're only looking for ascii chars */
361 for (prev = p = str; p < end; p++)
368 collate_key = g_utf8_collate_key (prev, p - prev);
369 g_string_append (result, collate_key);
370 g_free (collate_key);
373 g_string_append (result, COLLATION_SENTINEL "\1");
391 collate_key = g_utf8_collate_key (prev, p - prev);
392 g_string_append (result, collate_key);
393 g_free (collate_key);
396 g_string_append (result, COLLATION_SENTINEL "\2");
400 /* write d-1 colons */
414 if (*p == '0' && !digits)
416 else if (g_ascii_isdigit(*p))
420 /* count an all-zero sequence as
421 * one digit plus leading zeros
434 g_string_append_c (result, ':');
438 if (leading_zeros > 0)
440 g_string_append_c (append, (char)leading_zeros);
441 prev += leading_zeros;
444 /* write the number itself */
445 g_string_append_len (result, prev, p - prev);
448 --p; /* go one step back to avoid disturbing outer loop */
452 /* other characters just accumulate */
459 collate_key = g_utf8_collate_key (prev, p - prev);
460 g_string_append (result, collate_key);
461 g_free (collate_key);
464 g_string_append (result, append->str);
465 g_string_free (append, TRUE);
467 return g_string_free (result, FALSE);
471 #define __G_UNICOLLATE_C__
472 #include "galiasdef.c"