xref: /freebsd/bin/cat/cat.c (revision 8a4217aacf57330755501a349d0ea662d4880386)
1 /*-
2  * SPDX-License-Identifier: BSD-3-Clause
3  *
4  * Copyright (c) 1989, 1993
5  *	The Regents of the University of California.  All rights reserved.
6  *
7  * This code is derived from software contributed to Berkeley by
8  * Kevin Fall.
9  *
10  * Redistribution and use in source and binary forms, with or without
11  * modification, are permitted provided that the following conditions
12  * are met:
13  * 1. Redistributions of source code must retain the above copyright
14  *    notice, this list of conditions and the following disclaimer.
15  * 2. Redistributions in binary form must reproduce the above copyright
16  *    notice, this list of conditions and the following disclaimer in the
17  *    documentation and/or other materials provided with the distribution.
18  * 3. Neither the name of the University nor the names of its contributors
19  *    may be used to endorse or promote products derived from this software
20  *    without specific prior written permission.
21  *
22  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
23  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
24  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
25  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
26  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
27  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
28  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
29  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
30  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
31  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
32  * SUCH DAMAGE.
33  */
34 
35 #if 0
36 #ifndef lint
37 static char const copyright[] =
38 "@(#) Copyright (c) 1989, 1993\n\
39 	The Regents of the University of California.  All rights reserved.\n";
40 #endif /* not lint */
41 #endif
42 
43 #ifndef lint
44 #if 0
45 static char sccsid[] = "@(#)cat.c	8.2 (Berkeley) 4/27/95";
46 #endif
47 #endif /* not lint */
48 #include <sys/cdefs.h>
49 __FBSDID("$FreeBSD$");
50 
51 #include <sys/param.h>
52 #include <sys/stat.h>
53 #ifndef NO_UDOM_SUPPORT
54 #include <sys/socket.h>
55 #include <sys/un.h>
56 #include <netdb.h>
57 #endif
58 
59 #include <ctype.h>
60 #include <err.h>
61 #include <errno.h>
62 #include <fcntl.h>
63 #include <locale.h>
64 #include <stddef.h>
65 #include <stdio.h>
66 #include <stdlib.h>
67 #include <string.h>
68 #include <unistd.h>
69 #include <wchar.h>
70 #include <wctype.h>
71 
72 static int bflag, eflag, lflag, nflag, sflag, tflag, vflag;
73 static int rval;
74 static const char *filename;
75 
76 static void usage(void) __dead2;
77 static void scanfiles(char *argv[], int cooked);
78 static void cook_cat(FILE *);
79 static void raw_cat(int);
80 
81 #ifndef NO_UDOM_SUPPORT
82 static int udom_open(const char *path, int flags);
83 #endif
84 
85 /*
86  * Memory strategy threshold, in pages: if physmem is larger than this,
87  * use a large buffer.
88  */
89 #define	PHYSPAGES_THRESHOLD (32 * 1024)
90 
91 /* Maximum buffer size in bytes - do not allow it to grow larger than this. */
92 #define	BUFSIZE_MAX (2 * 1024 * 1024)
93 
94 /*
95  * Small (default) buffer size in bytes. It's inefficient for this to be
96  * smaller than MAXPHYS.
97  */
98 #define	BUFSIZE_SMALL (MAXPHYS)
99 
100 int
101 main(int argc, char *argv[])
102 {
103 	int ch;
104 	struct flock stdout_lock;
105 
106 	setlocale(LC_CTYPE, "");
107 
108 	while ((ch = getopt(argc, argv, "belnstuv")) != -1)
109 		switch (ch) {
110 		case 'b':
111 			bflag = nflag = 1;	/* -b implies -n */
112 			break;
113 		case 'e':
114 			eflag = vflag = 1;	/* -e implies -v */
115 			break;
116 		case 'l':
117 			lflag = 1;
118 			break;
119 		case 'n':
120 			nflag = 1;
121 			break;
122 		case 's':
123 			sflag = 1;
124 			break;
125 		case 't':
126 			tflag = vflag = 1;	/* -t implies -v */
127 			break;
128 		case 'u':
129 			setbuf(stdout, NULL);
130 			break;
131 		case 'v':
132 			vflag = 1;
133 			break;
134 		default:
135 			usage();
136 		}
137 	argv += optind;
138 
139 	if (lflag) {
140 		stdout_lock.l_len = 0;
141 		stdout_lock.l_start = 0;
142 		stdout_lock.l_type = F_WRLCK;
143 		stdout_lock.l_whence = SEEK_SET;
144 		if (fcntl(STDOUT_FILENO, F_SETLKW, &stdout_lock) == -1)
145 			err(EXIT_FAILURE, "stdout");
146 	}
147 
148 	if (bflag || eflag || nflag || sflag || tflag || vflag)
149 		scanfiles(argv, 1);
150 	else
151 		scanfiles(argv, 0);
152 	if (fclose(stdout))
153 		err(1, "stdout");
154 	exit(rval);
155 	/* NOTREACHED */
156 }
157 
158 static void
159 usage(void)
160 {
161 
162 	fprintf(stderr, "usage: cat [-belnstuv] [file ...]\n");
163 	exit(1);
164 	/* NOTREACHED */
165 }
166 
167 static void
168 scanfiles(char *argv[], int cooked)
169 {
170 	int fd, i;
171 	char *path;
172 	FILE *fp;
173 
174 	i = 0;
175 	fd = -1;
176 	while ((path = argv[i]) != NULL || i == 0) {
177 		if (path == NULL || strcmp(path, "-") == 0) {
178 			filename = "stdin";
179 			fd = STDIN_FILENO;
180 		} else {
181 			filename = path;
182 			fd = open(path, O_RDONLY);
183 #ifndef NO_UDOM_SUPPORT
184 			if (fd < 0 && errno == EOPNOTSUPP)
185 				fd = udom_open(path, O_RDONLY);
186 #endif
187 		}
188 		if (fd < 0) {
189 			warn("%s", path);
190 			rval = 1;
191 		} else if (cooked) {
192 			if (fd == STDIN_FILENO)
193 				cook_cat(stdin);
194 			else {
195 				fp = fdopen(fd, "r");
196 				cook_cat(fp);
197 				fclose(fp);
198 			}
199 		} else {
200 			raw_cat(fd);
201 			if (fd != STDIN_FILENO)
202 				close(fd);
203 		}
204 		if (path == NULL)
205 			break;
206 		++i;
207 	}
208 }
209 
210 static void
211 cook_cat(FILE *fp)
212 {
213 	int ch, gobble, line, prev;
214 	wint_t wch;
215 
216 	/* Reset EOF condition on stdin. */
217 	if (fp == stdin && feof(stdin))
218 		clearerr(stdin);
219 
220 	line = gobble = 0;
221 	for (prev = '\n'; (ch = getc(fp)) != EOF; prev = ch) {
222 		if (prev == '\n') {
223 			if (sflag) {
224 				if (ch == '\n') {
225 					if (gobble)
226 						continue;
227 					gobble = 1;
228 				} else
229 					gobble = 0;
230 			}
231 			if (nflag) {
232 				if (!bflag || ch != '\n') {
233 					(void)fprintf(stdout, "%6d\t", ++line);
234 					if (ferror(stdout))
235 						break;
236 				} else if (eflag) {
237 					(void)fprintf(stdout, "%6s\t", "");
238 					if (ferror(stdout))
239 						break;
240 				}
241 			}
242 		}
243 		if (ch == '\n') {
244 			if (eflag && putchar('$') == EOF)
245 				break;
246 		} else if (ch == '\t') {
247 			if (tflag) {
248 				if (putchar('^') == EOF || putchar('I') == EOF)
249 					break;
250 				continue;
251 			}
252 		} else if (vflag) {
253 			(void)ungetc(ch, fp);
254 			/*
255 			 * Our getwc(3) doesn't change file position
256 			 * on error.
257 			 */
258 			if ((wch = getwc(fp)) == WEOF) {
259 				if (ferror(fp) && errno == EILSEQ) {
260 					clearerr(fp);
261 					/* Resync attempt. */
262 					memset(&fp->_mbstate, 0, sizeof(mbstate_t));
263 					if ((ch = getc(fp)) == EOF)
264 						break;
265 					wch = ch;
266 					goto ilseq;
267 				} else
268 					break;
269 			}
270 			if (!iswascii(wch) && !iswprint(wch)) {
271 ilseq:
272 				if (putchar('M') == EOF || putchar('-') == EOF)
273 					break;
274 				wch = toascii(wch);
275 			}
276 			if (iswcntrl(wch)) {
277 				ch = toascii(wch);
278 				ch = (ch == '\177') ? '?' : (ch | 0100);
279 				if (putchar('^') == EOF || putchar(ch) == EOF)
280 					break;
281 				continue;
282 			}
283 			if (putwchar(wch) == WEOF)
284 				break;
285 			ch = -1;
286 			continue;
287 		}
288 		if (putchar(ch) == EOF)
289 			break;
290 	}
291 	if (ferror(fp)) {
292 		warn("%s", filename);
293 		rval = 1;
294 		clearerr(fp);
295 	}
296 	if (ferror(stdout))
297 		err(1, "stdout");
298 }
299 
300 static void
301 raw_cat(int rfd)
302 {
303 	int off, wfd;
304 	ssize_t nr, nw;
305 	static size_t bsize;
306 	static char *buf = NULL;
307 	struct stat sbuf;
308 
309 	wfd = fileno(stdout);
310 	if (buf == NULL) {
311 		if (fstat(wfd, &sbuf))
312 			err(1, "stdout");
313 		if (S_ISREG(sbuf.st_mode)) {
314 			/* If there's plenty of RAM, use a large copy buffer */
315 			if (sysconf(_SC_PHYS_PAGES) > PHYSPAGES_THRESHOLD)
316 				bsize = MIN(BUFSIZE_MAX, MAXPHYS * 8);
317 			else
318 				bsize = BUFSIZE_SMALL;
319 		} else
320 			bsize = MAX(sbuf.st_blksize,
321 			    (blksize_t)sysconf(_SC_PAGESIZE));
322 		if ((buf = malloc(bsize)) == NULL)
323 			err(1, "malloc() failure of IO buffer");
324 	}
325 	while ((nr = read(rfd, buf, bsize)) > 0)
326 		for (off = 0; nr; nr -= nw, off += nw)
327 			if ((nw = write(wfd, buf + off, (size_t)nr)) < 0)
328 				err(1, "stdout");
329 	if (nr < 0) {
330 		warn("%s", filename);
331 		rval = 1;
332 	}
333 }
334 
335 #ifndef NO_UDOM_SUPPORT
336 
337 static int
338 udom_open(const char *path, int flags)
339 {
340 	struct addrinfo hints, *res, *res0;
341 	char rpath[PATH_MAX];
342 	int fd = -1;
343 	int error;
344 
345 	/*
346 	 * Construct the unix domain socket address and attempt to connect.
347 	 */
348 	bzero(&hints, sizeof(hints));
349 	hints.ai_family = AF_LOCAL;
350 	if (realpath(path, rpath) == NULL)
351 		return (-1);
352 	error = getaddrinfo(rpath, NULL, &hints, &res0);
353 	if (error) {
354 		warn("%s", gai_strerror(error));
355 		errno = EINVAL;
356 		return (-1);
357 	}
358 	for (res = res0; res != NULL; res = res->ai_next) {
359 		fd = socket(res->ai_family, res->ai_socktype,
360 		    res->ai_protocol);
361 		if (fd < 0) {
362 			freeaddrinfo(res0);
363 			return (-1);
364 		}
365 		error = connect(fd, res->ai_addr, res->ai_addrlen);
366 		if (error == 0)
367 			break;
368 		else {
369 			close(fd);
370 			fd = -1;
371 		}
372 	}
373 	freeaddrinfo(res0);
374 
375 	/*
376 	 * handle the open flags by shutting down appropriate directions
377 	 */
378 	if (fd >= 0) {
379 		switch(flags & O_ACCMODE) {
380 		case O_RDONLY:
381 			if (shutdown(fd, SHUT_WR) == -1)
382 				warn(NULL);
383 			break;
384 		case O_WRONLY:
385 			if (shutdown(fd, SHUT_RD) == -1)
386 				warn(NULL);
387 			break;
388 		default:
389 			break;
390 		}
391 	}
392 	return (fd);
393 }
394 
395 #endif
396