xref: /dflybsd-src/lib/libc/locale/collate.c (revision 17ea22213f86a5c5966c1e6bf8e95f022ebb92b9) 
1 /*-
2  * Copyright (c) 1995 Alex Tatmanjants <alex@elvisti.kiev.ua>
3  *		at Electronni Visti IA, Kiev, Ukraine.
4  *			All rights reserved.
5  *
6  * Redistribution and use in source and binary forms, with or without
7  * modification, are permitted provided that the following conditions
8  * are met:
9  * 1. Redistributions of source code must retain the above copyright
10  *    notice, this list of conditions and the following disclaimer.
11  * 2. Redistributions in binary form must reproduce the above copyright
12  *    notice, this list of conditions and the following disclaimer in the
13  *    documentation and/or other materials provided with the distribution.
14  *
15  * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND
16  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
17  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
18  * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR BE LIABLE
19  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
20  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
21  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
22  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
23  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
24  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
25  * SUCH DAMAGE.
26  *
27  * $FreeBSD: src/lib/libc/locale/collate.c,v 1.21.2.4 2002/10/11 10:36:47 ache Exp $
28  * $DragonFly: src/lib/libc/locale/collate.c,v 1.3 2005/01/31 22:29:31 dillon Exp $
29  */
30 
31 #include "namespace.h"
32 #include <arpa/inet.h>
33 #include <rune.h>
34 #include <stdio.h>
35 #include <stdlib.h>
36 #include <string.h>
37 #include <errno.h>
38 #include <unistd.h>
39 #include <sysexits.h>
40 #include "un-namespace.h"
41 
42 #include "collate.h"
43 #include "setlocale.h"
44 #include "ldpart.h"
45 
46 int __collate_load_error = 1;
47 int __collate_substitute_nontrivial;
48 
49 u_char __collate_substitute_table[UCHAR_MAX + 1][STR_LEN];
50 struct __collate_st_char_pri __collate_char_pri_table[UCHAR_MAX + 1];
51 struct __collate_st_chain_pri *__collate_chain_pri_table;
52 
53 void __collate_err(int ex, const char *f) __dead2;
54 
55 int
56 __collate_load_tables(const char *encoding)
57 {
58 	FILE *fp;
59 	int i, saverr, chains;
60 	uint32_t u32;
61 	char strbuf[STR_LEN], buf[PATH_MAX];
62 	void *TMP_substitute_table, *TMP_char_pri_table, *TMP_chain_pri_table;
63 	static char collate_encoding[ENCODING_LEN + 1];
64 
65 	/* 'encoding' must be already checked. */
66 	if (strcmp(encoding, "C") == 0 || strcmp(encoding, "POSIX") == 0) {
67 		__collate_load_error = 1;
68 		return (_LDP_CACHE);
69 	}
70 
71 	/*
72 	 * If the locale name is the same as our cache, use the cache.
73 	 */
74 	if (strcmp(encoding, collate_encoding) == 0) {
75 		__collate_load_error = 0;
76 		return (_LDP_CACHE);
77 	}
78 
79 	/*
80 	 * Slurp the locale file into the cache.
81 	 */
82 
83 	/* 'PathLocale' must be already set & checked. */
84 	/* Range checking not needed, encoding has fixed size */
85 	(void)strcpy(buf, _PathLocale);
86 	(void)strcat(buf, "/");
87 	(void)strcat(buf, encoding);
88 	(void)strcat(buf, "/LC_COLLATE");
89 	if ((fp = fopen(buf, "r")) == NULL)
90 		return (_LDP_ERROR);
91 
92 	if (fread(strbuf, sizeof(strbuf), 1, fp) != 1) {
93 		saverr = errno;
94 		(void)fclose(fp);
95 		errno = saverr;
96 		return (_LDP_ERROR);
97 	}
98 	chains = -1;
99 	if (strcmp(strbuf, COLLATE_VERSION) == 0)
100 		chains = 0;
101 	else if (strcmp(strbuf, COLLATE_VERSION1_1) == 0)
102 		chains = 1;
103 	if (chains < 0) {
104 		(void)fclose(fp);
105 		errno = EFTYPE;
106 		return (_LDP_ERROR);
107 	}
108 	if (chains) {
109 		if (fread(&u32, sizeof(u32), 1, fp) != 1) {
110 			saverr = errno;
111 			(void)fclose(fp);
112 			errno = saverr;
113 			return (_LDP_ERROR);
114 		}
115 		if ((chains = (int)ntohl(u32)) < 1) {
116 			(void)fclose(fp);
117 			errno = EFTYPE;
118 			return (_LDP_ERROR);
119 		}
120 	} else
121 		chains = TABLE_SIZE;
122 
123 	if ((TMP_substitute_table =
124 	     malloc(sizeof(__collate_substitute_table))) == NULL) {
125 		saverr = errno;
126 		(void)fclose(fp);
127 		errno = saverr;
128 		return (_LDP_ERROR);
129 	}
130 	if ((TMP_char_pri_table =
131 	     malloc(sizeof(__collate_char_pri_table))) == NULL) {
132 		saverr = errno;
133 		free(TMP_substitute_table);
134 		(void)fclose(fp);
135 		errno = saverr;
136 		return (_LDP_ERROR);
137 	}
138 	if ((TMP_chain_pri_table =
139 	     malloc(sizeof(*__collate_chain_pri_table) * chains)) == NULL) {
140 		saverr = errno;
141 		free(TMP_substitute_table);
142 		free(TMP_char_pri_table);
143 		(void)fclose(fp);
144 		errno = saverr;
145 		return (_LDP_ERROR);
146 	}
147 
148 #define FREAD(a, b, c, d) \
149 { \
150 	if ( fread(a, b, c, d) != c) { \
151 		saverr = errno; \
152 		free(TMP_substitute_table); \
153 		free(TMP_char_pri_table); \
154 		free(TMP_chain_pri_table); \
155 		(void)fclose(d); \
156 		errno = saverr; \
157 		return (_LDP_ERROR); \
158 	} \
159 }
160 
161 	FREAD(TMP_substitute_table, sizeof(__collate_substitute_table), 1, fp);
162 	FREAD(TMP_char_pri_table, sizeof(__collate_char_pri_table), 1, fp);
163 	FREAD(TMP_chain_pri_table,
164 	      sizeof(*__collate_chain_pri_table), chains, fp);
165 	(void)fclose(fp);
166 
167 	(void)strcpy(collate_encoding, encoding);
168 	if (__collate_substitute_table_ptr != NULL)
169 		free(__collate_substitute_table_ptr);
170 	__collate_substitute_table_ptr = TMP_substitute_table;
171 	if (__collate_char_pri_table_ptr != NULL)
172 		free(__collate_char_pri_table_ptr);
173 	__collate_char_pri_table_ptr = TMP_char_pri_table;
174 	if (__collate_chain_pri_table != NULL)
175 		free(__collate_chain_pri_table);
176 	__collate_chain_pri_table = TMP_chain_pri_table;
177 
178 	__collate_substitute_nontrivial = 0;
179 	for (i = 0; i < UCHAR_MAX + 1; i++) {
180 		if (__collate_substitute_table[i][0] != i ||
181 		    __collate_substitute_table[i][1] != 0) {
182 			__collate_substitute_nontrivial = 1;
183 			break;
184 		}
185 	}
186 	__collate_load_error = 0;
187 
188 	return (_LDP_LOADED);
189 }
190 
191 u_char *
192 __collate_substitute(s)
193 	const u_char *s;
194 {
195 	int dest_len, len, nlen;
196 	int delta = strlen(s);
197 	u_char *dest_str = NULL;
198 
199 	if (s == NULL || *s == '\0')
200 		return (__collate_strdup(""));
201 	delta += delta / 8;
202 	dest_str = malloc(dest_len = delta);
203 	if (dest_str == NULL)
204 		__collate_err(EX_OSERR, __FUNCTION__);
205 	len = 0;
206 	while (*s) {
207 		nlen = len + strlen(__collate_substitute_table[*s]);
208 		if (dest_len <= nlen) {
209 			dest_str = reallocf(dest_str, dest_len = nlen + delta);
210 			if (dest_str == NULL)
211 				__collate_err(EX_OSERR, __FUNCTION__);
212 		}
213 		(void)strcpy(dest_str + len, __collate_substitute_table[*s++]);
214 		len = nlen;
215 	}
216 	return (dest_str);
217 }
218 
219 void
220 __collate_lookup(t, len, prim, sec)
221 	const u_char *t;
222 	int *len, *prim, *sec;
223 {
224 	struct __collate_st_chain_pri *p2;
225 
226 	*len = 1;
227 	*prim = *sec = 0;
228 	for (p2 = __collate_chain_pri_table; p2->str[0] != '\0'; p2++) {
229 		if (*t == p2->str[0] &&
230 		    strncmp(t, p2->str, strlen(p2->str)) == 0) {
231 			*len = strlen(p2->str);
232 			*prim = p2->prim;
233 			*sec = p2->sec;
234 			return;
235 		}
236 	}
237 	*prim = __collate_char_pri_table[*t].prim;
238 	*sec = __collate_char_pri_table[*t].sec;
239 }
240 
241 u_char *
242 __collate_strdup(s)
243 	u_char *s;
244 {
245 	u_char *t = strdup(s);
246 
247 	if (t == NULL)
248 		__collate_err(EX_OSERR, __FUNCTION__);
249 	return (t);
250 }
251 
252 void
253 __collate_err(int ex, const char *f)
254 {
255 	extern char *__progname;                /* Program name, from crt0. */
256 	const char *s;
257 	int serrno = errno;
258 
259 	s = __progname;
260 	_write(STDERR_FILENO, s, strlen(s));
261 	_write(STDERR_FILENO, ": ", 2);
262 	s = f;
263 	_write(STDERR_FILENO, s, strlen(s));
264 	_write(STDERR_FILENO, ": ", 2);
265 	s = strerror(serrno);
266 	_write(STDERR_FILENO, s, strlen(s));
267 	_write(STDERR_FILENO, "\n", 1);
268 	exit(ex);
269 }
270 
271 #ifdef COLLATE_DEBUG
272 void
273 __collate_print_tables()
274 {
275 	int i;
276 	struct __collate_st_chain_pri *p2;
277 
278 	printf("Substitute table:\n");
279 	for (i = 0; i < UCHAR_MAX + 1; i++)
280 	    if (i != *__collate_substitute_table[i])
281 		printf("\t'%c' --> \"%s\"\n", i,
282 		       __collate_substitute_table[i]);
283 	printf("Chain priority table:\n");
284 	for (p2 = __collate_chain_pri_table; p2->str[0] != '\0'; p2++)
285 		printf("\t\"%s\" : %d %d\n", p2->str, p2->prim, p2->sec);
286 	printf("Char priority table:\n");
287 	for (i = 0; i < UCHAR_MAX + 1; i++)
288 		printf("\t'%c' : %d %d\n", i, __collate_char_pri_table[i].prim,
289 		       __collate_char_pri_table[i].sec);
290 }
291 #endif
292