xref: /freebsd/usr.bin/tr/str.c (revision ae477ca7da55f76d28859e1bd01cd1051e36f28f)
1 /*-
2  * SPDX-License-Identifier: BSD-3-Clause
3  *
4  * Copyright (c) 1991, 1993
5  *	The Regents of the University of California.  All rights reserved.
6  *
7  * Redistribution and use in source and binary forms, with or without
8  * modification, are permitted provided that the following conditions
9  * are met:
10  * 1. Redistributions of source code must retain the above copyright
11  *    notice, this list of conditions and the following disclaimer.
12  * 2. Redistributions in binary form must reproduce the above copyright
13  *    notice, this list of conditions and the following disclaimer in the
14  *    documentation and/or other materials provided with the distribution.
15  * 3. Neither the name of the University nor the names of its contributors
16  *    may be used to endorse or promote products derived from this software
17  *    without specific prior written permission.
18  *
19  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
20  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
21  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
22  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
23  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
24  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
25  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
26  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
27  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
28  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
29  * SUCH DAMAGE.
30  */
31 
32 
33 
34 #include <sys/types.h>
35 
36 #include <ctype.h>
37 #include <err.h>
38 #include <errno.h>
39 #include <stddef.h>
40 #include <stdio.h>
41 #include <stdlib.h>
42 #include <string.h>
43 #include <wchar.h>
44 #include <wctype.h>
45 
46 #include "extern.h"
47 
48 static int      backslash(STR *, int *);
49 static int	bracket(STR *);
50 static void	genclass(STR *);
51 static void	genequiv(STR *);
52 static int      genrange(STR *, int);
53 static void	genseq(STR *);
54 
55 wint_t
56 next(STR *s)
57 {
58 	int is_octal;
59 	wint_t ch;
60 	wchar_t wch;
61 	size_t clen;
62 
63 	switch (s->state) {
64 	case EOS:
65 		return (0);
66 	case INFINITE:
67 		return (1);
68 	case NORMAL:
69 		switch (*s->str) {
70 		case '\0':
71 			s->state = EOS;
72 			return (0);
73 		case '\\':
74 			s->lastch = backslash(s, &is_octal);
75 			break;
76 		case '[':
77 			if (bracket(s))
78 				return (next(s));
79 			/* FALLTHROUGH */
80 		default:
81 			clen = mbrtowc(&wch, s->str, MB_LEN_MAX, NULL);
82 			if (clen == (size_t)-1 || clen == (size_t)-2 ||
83 			    clen == 0)
84 				errc(1, EILSEQ, NULL);
85 			is_octal = 0;
86 			s->lastch = wch;
87 			s->str += clen;
88 			break;
89 		}
90 
91 		/* We can start a range at any time. */
92 		if (s->str[0] == '-' && genrange(s, is_octal))
93 			return (next(s));
94 		return (1);
95 	case RANGE:
96 		if (s->cnt-- == 0) {
97 			s->state = NORMAL;
98 			return (next(s));
99 		}
100 		++s->lastch;
101 		return (1);
102 	case SEQUENCE:
103 		if (s->cnt-- == 0) {
104 			s->state = NORMAL;
105 			return (next(s));
106 		}
107 		return (1);
108 	case CCLASS:
109 	case CCLASS_UPPER:
110 	case CCLASS_LOWER:
111 		s->cnt++;
112 		ch = nextwctype(s->lastch, s->cclass);
113 		if (ch == -1) {
114 			s->state = NORMAL;
115 			return (next(s));
116 		}
117 		s->lastch = ch;
118 		return (1);
119 	case SET:
120 		if ((ch = s->set[s->cnt++]) == OOBCH) {
121 			s->state = NORMAL;
122 			return (next(s));
123 		}
124 		s->lastch = ch;
125 		return (1);
126 	default:
127 		return (0);
128 	}
129 	/* NOTREACHED */
130 }
131 
132 static int
133 bracket(STR *s)
134 {
135 	char *p;
136 
137 	switch (s->str[1]) {
138 	case ':':				/* "[:class:]" */
139 		if ((p = strchr(s->str + 2, ']')) == NULL)
140 			return (0);
141 		if (*(p - 1) != ':' || p - s->str < 4)
142 			goto repeat;
143 		*(p - 1) = '\0';
144 		s->str += 2;
145 		genclass(s);
146 		s->str = p + 1;
147 		return (1);
148 	case '=':				/* "[=equiv=]" */
149 		if (s->str[2] == '\0' || (p = strchr(s->str + 3, ']')) == NULL)
150 			return (0);
151 		if (*(p - 1) != '=' || p - s->str < 4)
152 			goto repeat;
153 		s->str += 2;
154 		genequiv(s);
155 		return (1);
156 	default:				/* "[\###*n]" or "[#*n]" */
157 	repeat:
158 		if ((p = strpbrk(s->str + 2, "*]")) == NULL)
159 			return (0);
160 		if (p[0] != '*' || strchr(p, ']') == NULL)
161 			return (0);
162 		s->str += 1;
163 		genseq(s);
164 		return (1);
165 	}
166 	/* NOTREACHED */
167 }
168 
169 static void
170 genclass(STR *s)
171 {
172 
173 	if ((s->cclass = wctype(s->str)) == 0)
174 		errx(1, "unknown class %s", s->str);
175 	s->cnt = 0;
176 	s->lastch = -1;		/* incremented before check in next() */
177 	if (strcmp(s->str, "upper") == 0)
178 		s->state = CCLASS_UPPER;
179 	else if (strcmp(s->str, "lower") == 0)
180 		s->state = CCLASS_LOWER;
181 	else
182 		s->state = CCLASS;
183 }
184 
185 static void
186 genequiv(STR *s)
187 {
188 	int i, p, pri;
189 	char src[2], dst[3];
190 	size_t clen;
191 	wchar_t wc;
192 
193 	if (*s->str == '\\') {
194 		s->equiv[0] = backslash(s, NULL);
195 		if (*s->str != '=')
196 			errx(1, "misplaced equivalence equals sign");
197 		s->str += 2;
198 	} else {
199 		clen = mbrtowc(&wc, s->str, MB_LEN_MAX, NULL);
200 		if (clen == (size_t)-1 || clen == (size_t)-2 || clen == 0)
201 			errc(1, EILSEQ, NULL);
202 		s->equiv[0] = wc;
203 		if (s->str[clen] != '=')
204 			errx(1, "misplaced equivalence equals sign");
205 		s->str += clen + 2;
206 	}
207 
208 	/*
209 	 * Calculate the set of all characters in the same equivalence class
210 	 * as the specified character (they will have the same primary
211 	 * collation weights).
212 	 * XXX Knows too much about how strxfrm() is implemented. Assumes
213 	 * it fills the string with primary collation weight bytes. Only one-
214 	 * to-one mappings are supported.
215 	 * XXX Equivalence classes not supported in multibyte locales.
216 	 */
217 	src[0] = (char)s->equiv[0];
218 	src[1] = '\0';
219 	if (MB_CUR_MAX == 1 && strxfrm(dst, src, sizeof(dst)) == 1) {
220 		pri = (unsigned char)*dst;
221 		for (p = 1, i = 1; i < NCHARS_SB; i++) {
222 			*src = i;
223 			if (strxfrm(dst, src, sizeof(dst)) == 1 && pri &&
224 			    pri == (unsigned char)*dst)
225 				s->equiv[p++] = i;
226 		}
227 		s->equiv[p] = OOBCH;
228 	}
229 
230 	s->cnt = 0;
231 	s->state = SET;
232 	s->set = s->equiv;
233 }
234 
235 static int
236 genrange(STR *s, int was_octal)
237 {
238 	int stopval, octal;
239 	char *savestart;
240 	int n, cnt, *p;
241 	size_t clen;
242 	wchar_t wc;
243 
244 	octal = 0;
245 	savestart = s->str;
246 	if (*++s->str == '\\')
247 		stopval = backslash(s, &octal);
248 	else {
249 		clen = mbrtowc(&wc, s->str, MB_LEN_MAX, NULL);
250 		if (clen == (size_t)-1 || clen == (size_t)-2)
251 			errc(1, EILSEQ, NULL);
252 		stopval = wc;
253 		s->str += clen;
254 	}
255 	/*
256 	 * XXX Characters are not ordered according to collating sequence in
257 	 * multibyte locales.
258 	 */
259 	if (octal || was_octal || MB_CUR_MAX > 1) {
260 		if (stopval < s->lastch) {
261 			s->str = savestart;
262 			return (0);
263 		}
264 		s->cnt = stopval - s->lastch + 1;
265 		s->state = RANGE;
266 		--s->lastch;
267 		return (1);
268 	}
269 	if (charcoll((const void *)&stopval, (const void *)&(s->lastch)) < 0) {
270 		s->str = savestart;
271 		return (0);
272 	}
273 	if ((s->set = p = malloc((NCHARS_SB + 1) * sizeof(int))) == NULL)
274 		err(1, "genrange() malloc");
275 	for (cnt = 0; cnt < NCHARS_SB; cnt++)
276 		if (charcoll((const void *)&cnt, (const void *)&(s->lastch)) >= 0 &&
277 		    charcoll((const void *)&cnt, (const void *)&stopval) <= 0)
278 			*p++ = cnt;
279 	*p = OOBCH;
280 	n = p - s->set;
281 
282 	s->cnt = 0;
283 	s->state = SET;
284 	if (n > 1)
285 		mergesort(s->set, n, sizeof(*(s->set)), charcoll);
286 	return (1);
287 }
288 
289 static void
290 genseq(STR *s)
291 {
292 	char *ep;
293 	wchar_t wc;
294 	size_t clen;
295 
296 	if (s->which == STRING1)
297 		errx(1, "sequences only valid in string2");
298 
299 	if (*s->str == '\\')
300 		s->lastch = backslash(s, NULL);
301 	else {
302 		clen = mbrtowc(&wc, s->str, MB_LEN_MAX, NULL);
303 		if (clen == (size_t)-1 || clen == (size_t)-2)
304 			errc(1, EILSEQ, NULL);
305 		s->lastch = wc;
306 		s->str += clen;
307 	}
308 	if (*s->str != '*')
309 		errx(1, "misplaced sequence asterisk");
310 
311 	switch (*++s->str) {
312 	case '\\':
313 		s->cnt = backslash(s, NULL);
314 		break;
315 	case ']':
316 		s->cnt = 0;
317 		++s->str;
318 		break;
319 	default:
320 		if (isdigit((u_char)*s->str)) {
321 			s->cnt = strtol(s->str, &ep, 0);
322 			if (*ep == ']') {
323 				s->str = ep + 1;
324 				break;
325 			}
326 		}
327 		errx(1, "illegal sequence count");
328 		/* NOTREACHED */
329 	}
330 
331 	s->state = s->cnt ? SEQUENCE : INFINITE;
332 }
333 
334 /*
335  * Translate \??? into a character.  Up to 3 octal digits, if no digits either
336  * an escape code or a literal character.
337  */
338 static int
339 backslash(STR *s, int *is_octal)
340 {
341 	int ch, cnt, val;
342 
343 	if (is_octal != NULL)
344 		*is_octal = 0;
345 	for (cnt = val = 0;;) {
346 		ch = (u_char)*++s->str;
347 		if (!isdigit(ch) || ch > '7')
348 			break;
349 		val = val * 8 + ch - '0';
350 		if (++cnt == 3) {
351 			++s->str;
352 			break;
353 		}
354 	}
355 	if (cnt) {
356 		if (is_octal != NULL)
357 			*is_octal = 1;
358 		return (val);
359 	}
360 	if (ch != '\0')
361 		++s->str;
362 	switch (ch) {
363 		case 'a':			/* escape characters */
364 			return ('\7');
365 		case 'b':
366 			return ('\b');
367 		case 'f':
368 			return ('\f');
369 		case 'n':
370 			return ('\n');
371 		case 'r':
372 			return ('\r');
373 		case 't':
374 			return ('\t');
375 		case 'v':
376 			return ('\13');
377 		case '\0':			/*  \" -> \ */
378 			s->state = EOS;
379 			return ('\\');
380 		default:			/* \x" -> x */
381 			return (ch);
382 	}
383 }
384