xref: /freebsd/usr.bin/wc/wc.c (revision 56b17de1e8360fe131d425de20b5e75ff3ea897c)
1 /*-
2  * SPDX-License-Identifier: BSD-3-Clause
3  *
4  * Copyright (c) 1980, 1987, 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 #include <sys/capsicum.h>
33 #include <sys/param.h>
34 #include <sys/stat.h>
35 
36 #include <capsicum_helpers.h>
37 #include <ctype.h>
38 #include <errno.h>
39 #include <fcntl.h>
40 #include <locale.h>
41 #include <stdbool.h>
42 #include <stdint.h>
43 #include <stdio.h>
44 #include <stdlib.h>
45 #include <string.h>
46 #include <unistd.h>
47 #include <wchar.h>
48 #include <wctype.h>
49 #include <libxo/xo.h>
50 
51 #include <libcasper.h>
52 #include <casper/cap_fileargs.h>
53 
54 #define WC_XO_VERSION "1"
55 
56 static const char *stdin_filename = "stdin";
57 
58 static fileargs_t *fa;
59 static uintmax_t tlinect, twordct, tcharct, tlongline;
60 static bool doline, doword, dochar, domulti, dolongline;
61 static volatile sig_atomic_t siginfo;
62 static xo_handle_t *stderr_handle;
63 
64 static void	show_cnt(const char *file, uintmax_t linect, uintmax_t wordct,
65 		    uintmax_t charct, uintmax_t llct);
66 static int	cnt(const char *);
67 static void	usage(void);
68 
69 static void
70 siginfo_handler(int sig __unused)
71 {
72 
73 	siginfo = 1;
74 }
75 
76 static void
77 reset_siginfo(void)
78 {
79 
80 	signal(SIGINFO, SIG_DFL);
81 	siginfo = 0;
82 }
83 
84 int
85 main(int argc, char *argv[])
86 {
87 	int ch, errors, total;
88 	cap_rights_t rights;
89 
90 	(void) setlocale(LC_CTYPE, "");
91 
92 	argc = xo_parse_args(argc, argv);
93 	if (argc < 0)
94 		exit(EXIT_FAILURE);
95 
96 	while ((ch = getopt(argc, argv, "clmwL")) != -1)
97 		switch((char)ch) {
98 		case 'l':
99 			doline = true;
100 			break;
101 		case 'w':
102 			doword = true;
103 			break;
104 		case 'c':
105 			dochar = true;
106 			domulti = false;
107 			break;
108 		case 'L':
109 			dolongline = true;
110 			break;
111 		case 'm':
112 			domulti = true;
113 			dochar = false;
114 			break;
115 		case '?':
116 		default:
117 			usage();
118 		}
119 	argv += optind;
120 	argc -= optind;
121 
122 	fa = fileargs_init(argc, argv, O_RDONLY, 0,
123 	    cap_rights_init(&rights, CAP_READ, CAP_FSTAT), FA_OPEN);
124 	if (fa == NULL)
125 		xo_err(EXIT_FAILURE, "Unable to initialize casper");
126 	caph_cache_catpages();
127 	if (caph_limit_stdio() < 0)
128 		xo_err(EXIT_FAILURE, "Unable to limit stdio");
129 	if (caph_enter_casper() < 0)
130 		xo_err(EXIT_FAILURE, "Unable to enter capability mode");
131 
132 	/* Wc's flags are on by default. */
133 	if (!(doline || doword || dochar || domulti || dolongline))
134 		doline = doword = dochar = true;
135 
136 	stderr_handle = xo_create_to_file(stderr, XO_STYLE_TEXT, 0);
137 
138 	xo_set_version(WC_XO_VERSION);
139 	xo_open_container("wc");
140 	xo_open_list("file");
141 
142 	(void)signal(SIGINFO, siginfo_handler);
143 	errors = 0;
144 	total = 0;
145 	if (argc == 0) {
146 		xo_open_instance("file");
147 		if (cnt(NULL) != 0)
148 			++errors;
149 		xo_close_instance("file");
150 	} else {
151 		while (argc--) {
152 			xo_open_instance("file");
153 			if (cnt(*argv++) != 0)
154 				++errors;
155 			xo_close_instance("file");
156 			++total;
157 		}
158 	}
159 
160 	xo_close_list("file");
161 
162 	if (total > 1) {
163 		xo_open_container("total");
164 		show_cnt("total", tlinect, twordct, tcharct, tlongline);
165 		xo_close_container("total");
166 	}
167 
168 	fileargs_free(fa);
169 	xo_close_container("wc");
170 	if (xo_finish() < 0)
171 		xo_err(EXIT_FAILURE, "stdout");
172 	exit(errors == 0 ? EXIT_SUCCESS : EXIT_FAILURE);
173 }
174 
175 static void
176 show_cnt(const char *file, uintmax_t linect, uintmax_t wordct,
177     uintmax_t charct, uintmax_t llct)
178 {
179 	xo_handle_t *xop;
180 
181 	if (!siginfo)
182 		xop = NULL;
183 	else {
184 		xop = stderr_handle;
185 		siginfo = 0;
186 	}
187 
188 	if (doline)
189 		xo_emit_h(xop, " {:lines/%7ju/%ju}", linect);
190 	if (doword)
191 		xo_emit_h(xop, " {:words/%7ju/%ju}", wordct);
192 	if (dochar || domulti)
193 		xo_emit_h(xop, " {:characters/%7ju/%ju}", charct);
194 	if (dolongline)
195 		xo_emit_h(xop, " {:long-lines/%7ju/%ju}", llct);
196 	if (file != stdin_filename)
197 		xo_emit_h(xop, " {:filename/%s}\n", file);
198 	else
199 		xo_emit_h(xop, "\n");
200 }
201 
202 static int
203 cnt(const char *file)
204 {
205 	static char buf[MAXBSIZE];
206 	struct stat sb;
207 	mbstate_t mbs;
208 	const char *p;
209 	uintmax_t linect, wordct, charct, llct, tmpll;
210 	ssize_t len;
211 	size_t clen;
212 	int fd;
213 	wchar_t wch;
214 	bool gotsp, warned;
215 
216 	linect = wordct = charct = llct = tmpll = 0;
217 	if (file == NULL) {
218 		fd = STDIN_FILENO;
219 		file = stdin_filename;
220 	} else if ((fd = fileargs_open(fa, file)) < 0) {
221 		xo_warn("%s: open", file);
222 		return (1);
223 	}
224 	if (doword || (domulti && MB_CUR_MAX != 1))
225 		goto word;
226 	/*
227 	 * If all we need is the number of characters and it's a regular file,
228 	 * just stat it.
229 	 */
230 	if (doline == 0 && dolongline == 0) {
231 		if (fstat(fd, &sb)) {
232 			xo_warn("%s: fstat", file);
233 			(void)close(fd);
234 			return (1);
235 		}
236 		/* pseudo-filesystems advertize a zero size */
237 		if (S_ISREG(sb.st_mode) && sb.st_size > 0) {
238 			reset_siginfo();
239 			charct = sb.st_size;
240 			show_cnt(file, linect, wordct, charct, llct);
241 			tcharct += charct;
242 			(void)close(fd);
243 			return (0);
244 		}
245 	}
246 	/*
247 	 * For files we can't stat, or if we need line counting, slurp the
248 	 * file.  Line counting is split out because it's a lot faster to get
249 	 * lines than to get words, since the word count requires locale
250 	 * handling.
251 	 */
252 	while ((len = read(fd, buf, sizeof(buf))) != 0) {
253 		if (len < 0) {
254 			xo_warn("%s: read", file);
255 			(void)close(fd);
256 			return (1);
257 		}
258 		if (siginfo)
259 			show_cnt(file, linect, wordct, charct, llct);
260 		charct += len;
261 		if (doline || dolongline) {
262 			for (p = buf; len > 0; --len, ++p) {
263 				if (*p == '\n') {
264 					if (tmpll > llct)
265 						llct = tmpll;
266 					tmpll = 0;
267 					++linect;
268 				} else {
269 					tmpll++;
270 				}
271 			}
272 		}
273 	}
274 	reset_siginfo();
275 	if (doline)
276 		tlinect += linect;
277 	if (dochar)
278 		tcharct += charct;
279 	if (dolongline && llct > tlongline)
280 		tlongline = llct;
281 	show_cnt(file, linect, wordct, charct, llct);
282 	(void)close(fd);
283 	return (0);
284 
285 	/* Do it the hard way... */
286 word:	gotsp = true;
287 	warned = false;
288 	memset(&mbs, 0, sizeof(mbs));
289 	while ((len = read(fd, buf, sizeof(buf))) != 0) {
290 		if (len < 0) {
291 			xo_warn("%s: read", file);
292 			(void)close(fd);
293 			return (1);
294 		}
295 		p = buf;
296 		while (len > 0) {
297 			if (siginfo)
298 				show_cnt(file, linect, wordct, charct, llct);
299 			if (!domulti || MB_CUR_MAX == 1) {
300 				clen = 1;
301 				wch = (unsigned char)*p;
302 			} else if ((clen = mbrtowc(&wch, p, len, &mbs)) == 0) {
303 				clen = 1;
304 			} else if (clen == (size_t)-1) {
305 				if (!warned) {
306 					errno = EILSEQ;
307 					xo_warn("%s", file);
308 					warned = true;
309 				}
310 				memset(&mbs, 0, sizeof(mbs));
311 				clen = 1;
312 				wch = (unsigned char)*p;
313 			} else if (clen == (size_t)-2) {
314 				break;
315 			}
316 			charct++;
317 			if (wch != L'\n')
318 				tmpll++;
319 			len -= clen;
320 			p += clen;
321 			if (wch == L'\n') {
322 				if (tmpll > llct)
323 					llct = tmpll;
324 				tmpll = 0;
325 				++linect;
326 			}
327 			if (iswspace(wch)) {
328 				gotsp = true;
329 			} else if (gotsp) {
330 				gotsp = false;
331 				++wordct;
332 			}
333 		}
334 	}
335 	reset_siginfo();
336 	if (domulti && MB_CUR_MAX > 1) {
337 		if (mbrtowc(NULL, NULL, 0, &mbs) == (size_t)-1 && !warned)
338 			xo_warn("%s", file);
339 	}
340 	if (doline)
341 		tlinect += linect;
342 	if (doword)
343 		twordct += wordct;
344 	if (dochar || domulti)
345 		tcharct += charct;
346 	if (dolongline && llct > tlongline)
347 		tlongline = llct;
348 	show_cnt(file, linect, wordct, charct, llct);
349 	(void)close(fd);
350 	return (0);
351 }
352 
353 static void
354 usage(void)
355 {
356 	xo_error("usage: wc [-Lclmw] [file ...]\n");
357 	exit(EXIT_FAILURE);
358 }
359