xref: /freebsd/lib/libc/string/strcoll.c (revision fed1ca4b719c56c930f2259d80663cd34be812bb)
1 /*-
2  * Copyright 2010 Nexenta Systems, Inc.  All rights reserved.
3  * Copyright (c) 1995 Alex Tatmanjants <alex@elvisti.kiev.ua>
4  *		at Electronni Visti IA, Kiev, Ukraine.
5  *			All rights reserved.
6  *
7  * Copyright (c) 2011 The FreeBSD Foundation
8  * All rights reserved.
9  * Portions of this software were developed by David Chisnall
10  * under sponsorship from the FreeBSD Foundation.
11  *
12  * Redistribution and use in source and binary forms, with or without
13  * modification, are permitted provided that the following conditions
14  * are met:
15  * 1. Redistributions of source code must retain the above copyright
16  *    notice, this list of conditions and the following disclaimer.
17  * 2. Redistributions in binary form must reproduce the above copyright
18  *    notice, this list of conditions and the following disclaimer in the
19  *    documentation and/or other materials provided with the distribution.
20  *
21  * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND
22  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
23  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
24  * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR BE LIABLE
25  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
26  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
27  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
28  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
29  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
30  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
31  * SUCH DAMAGE.
32  */
33 
34 #include <sys/cdefs.h>
35 __FBSDID("$FreeBSD$");
36 
37 #include <stdlib.h>
38 #include <string.h>
39 #include <errno.h>
40 #include <wchar.h>
41 #include "collate.h"
42 
43 
44 /*
45  * In order to properly handle multibyte locales, its easiest to just
46  * convert to wide characters and then use wcscoll.  However if an
47  * error occurs, we gracefully fall back to simple strcmp.  Caller
48  * should check errno.
49  */
50 int
51 strcoll_l(const char *s, const char *s2, locale_t locale)
52 {
53 	int ret;
54 	wchar_t *t1 = NULL, *t2 = NULL;
55 	wchar_t *w1 = NULL, *w2 = NULL;
56 	const char *cs1, *cs2;
57 	mbstate_t mbs1;
58 	mbstate_t mbs2;
59 	size_t sz1, sz2;
60 
61 	memset(&mbs1, 0, sizeof (mbstate_t));
62 	memset(&mbs2, 0, sizeof (mbstate_t));
63 
64 	/*
65 	 * The mbsrtowcs_l function can set the src pointer to null upon
66 	 * failure, so it should act on a copy to avoid:
67 	 *   - sending null pointer to strcmp
68 	 *   - having strcoll/strcoll_l change *s or *s2 to null
69 	 */
70 	cs1 = s;
71 	cs2 = s2;
72 
73 	FIX_LOCALE(locale);
74 	struct xlocale_collate *table =
75 		(struct xlocale_collate*)locale->components[XLC_COLLATE];
76 
77 	if (table->__collate_load_error)
78 		goto error;
79 
80 	sz1 = strlen(s) + 1;
81 	sz2 = strlen(s2) + 1;
82 
83 	/*
84 	 * Simple assumption: conversion to wide format is strictly
85 	 * reducing, i.e. a single byte (or multibyte character)
86 	 * cannot result in multiple wide characters.
87 	 */
88 	if ((t1 = malloc(sz1 * sizeof (wchar_t))) == NULL)
89 		goto error;
90 	w1 = t1;
91 	if ((t2 = malloc(sz2 * sizeof (wchar_t))) == NULL)
92 		goto error;
93 	w2 = t2;
94 
95 	if ((mbsrtowcs_l(w1, &cs1, sz1, &mbs1, locale)) == (size_t)-1)
96 		goto error;
97 
98 	if ((mbsrtowcs_l(w2, &cs2, sz2, &mbs2, locale)) == (size_t)-1)
99 		goto error;
100 
101 	ret = wcscoll_l(w1, w2, locale);
102 	free(t1);
103 	free(t2);
104 
105 	return (ret);
106 
107 error:
108 	free(t1);
109 	free(t2);
110 	return (strcmp(s, s2));
111 }
112 
113 int
114 strcoll(const char *s, const char *s2)
115 {
116 	return strcoll_l(s, s2, __get_locale());
117 }
118 
119