xref: /freebsd/usr.bin/sed/main.c (revision a35d88931c87cfe6bd38f01d7bad22140b3b38f3)
1 /*-
2  * Copyright (c) 1992 Diomidis Spinellis.
3  * Copyright (c) 1992, 1993
4  *	The Regents of the University of California.  All rights reserved.
5  *
6  * This code is derived from software contributed to Berkeley by
7  * Diomidis Spinellis of Imperial College, University of London.
8  *
9  * Redistribution and use in source and binary forms, with or without
10  * modification, are permitted provided that the following conditions
11  * are met:
12  * 1. Redistributions of source code must retain the above copyright
13  *    notice, this list of conditions and the following disclaimer.
14  * 2. Redistributions in binary form must reproduce the above copyright
15  *    notice, this list of conditions and the following disclaimer in the
16  *    documentation and/or other materials provided with the distribution.
17  * 4. Neither the name of the University nor the names of its contributors
18  *    may be used to endorse or promote products derived from this software
19  *    without specific prior written permission.
20  *
21  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``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 REGENTS OR CONTRIBUTORS 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 #ifndef lint
38 static const char copyright[] =
39 "@(#) Copyright (c) 1992, 1993\n\
40 	The Regents of the University of California.  All rights reserved.\n";
41 #endif
42 
43 #ifndef lint
44 static const char sccsid[] = "@(#)main.c	8.2 (Berkeley) 1/3/94";
45 #endif
46 
47 #include <sys/types.h>
48 #include <sys/mman.h>
49 #include <sys/param.h>
50 #include <sys/stat.h>
51 
52 #include <err.h>
53 #include <errno.h>
54 #include <fcntl.h>
55 #include <limits.h>
56 #include <locale.h>
57 #include <regex.h>
58 #include <stddef.h>
59 #include <stdio.h>
60 #include <stdlib.h>
61 #include <string.h>
62 #include <unistd.h>
63 
64 #include "defs.h"
65 #include "extern.h"
66 
67 /*
68  * Linked list of units (strings and files) to be compiled
69  */
70 struct s_compunit {
71 	struct s_compunit *next;
72 	enum e_cut {CU_FILE, CU_STRING} type;
73 	char *s;			/* Pointer to string or fname */
74 };
75 
76 /*
77  * Linked list pointer to compilation units and pointer to current
78  * next pointer.
79  */
80 static struct s_compunit *script, **cu_nextp = &script;
81 
82 /*
83  * Linked list of files to be processed
84  */
85 struct s_flist {
86 	char *fname;
87 	struct s_flist *next;
88 };
89 
90 /*
91  * Linked list pointer to files and pointer to current
92  * next pointer.
93  */
94 static struct s_flist *files, **fl_nextp = &files;
95 
96 FILE *infile;			/* Current input file */
97 FILE *outfile;			/* Current output file */
98 
99 int aflag, eflag, nflag;
100 int rflags = 0;
101 static int rval;		/* Exit status */
102 
103 /*
104  * Current file and line number; line numbers restart across compilation
105  * units, but span across input files.
106  */
107 const char *fname;		/* File name. */
108 const char *outfname;		/* Output file name */
109 static char oldfname[PATH_MAX];	/* Old file name (for in-place editing) */
110 static char tmpfname[PATH_MAX];	/* Temporary file name (for in-place editing) */
111 const char *inplace;		/* Inplace edit file extension. */
112 u_long linenum;
113 
114 static void add_compunit(enum e_cut, char *);
115 static void add_file(char *);
116 static int inplace_edit(char **);
117 static void usage(void);
118 
119 int
120 main(int argc, char *argv[])
121 {
122 	int c, fflag;
123 	char *temp_arg;
124 
125 	(void) setlocale(LC_ALL, "");
126 
127 	fflag = 0;
128 	inplace = NULL;
129 
130 	while ((c = getopt(argc, argv, "Eae:f:i:n")) != -1)
131 		switch (c) {
132 		case 'E':
133 			rflags = REG_EXTENDED;
134 			break;
135 		case 'a':
136 			aflag = 1;
137 			break;
138 		case 'e':
139 			eflag = 1;
140 			if ((temp_arg = malloc(strlen(optarg) + 2)) == NULL)
141 				err(1, "malloc");
142 			strcpy(temp_arg, optarg);
143 			strcat(temp_arg, "\n");
144 			add_compunit(CU_STRING, temp_arg);
145 			break;
146 		case 'f':
147 			fflag = 1;
148 			add_compunit(CU_FILE, optarg);
149 			break;
150 		case 'i':
151 			inplace = optarg;
152 			break;
153 		case 'n':
154 			nflag = 1;
155 			break;
156 		default:
157 		case '?':
158 			usage();
159 		}
160 	argc -= optind;
161 	argv += optind;
162 
163 	/* First usage case; script is the first arg */
164 	if (!eflag && !fflag && *argv) {
165 		add_compunit(CU_STRING, *argv);
166 		argv++;
167 	}
168 
169 	compile();
170 
171 	/* Continue with first and start second usage */
172 	if (*argv)
173 		for (; *argv; argv++)
174 			add_file(*argv);
175 	else
176 		add_file(NULL);
177 	process();
178 	cfclose(prog, NULL);
179 	if (fclose(stdout))
180 		err(1, "stdout");
181 	exit(rval);
182 }
183 
184 static void
185 usage(void)
186 {
187 	(void)fprintf(stderr, "%s\n%s\n",
188 		"usage: sed script [-Ean] [-i extension] [file ...]",
189 		"       sed [-an] [-i extension] [-e script] ... [-f script_file] ... [file ...]");
190 	exit(1);
191 }
192 
193 /*
194  * Like fgets, but go through the chain of compilation units chaining them
195  * together.  Empty strings and files are ignored.
196  */
197 char *
198 cu_fgets(char *buf, int n, int *more)
199 {
200 	static enum {ST_EOF, ST_FILE, ST_STRING} state = ST_EOF;
201 	static FILE *f;		/* Current open file */
202 	static char *s;		/* Current pointer inside string */
203 	static char string_ident[30];
204 	char *p;
205 
206 again:
207 	switch (state) {
208 	case ST_EOF:
209 		if (script == NULL) {
210 			if (more != NULL)
211 				*more = 0;
212 			return (NULL);
213 		}
214 		linenum = 0;
215 		switch (script->type) {
216 		case CU_FILE:
217 			if ((f = fopen(script->s, "r")) == NULL)
218 				err(1, "%s", script->s);
219 			fname = script->s;
220 			state = ST_FILE;
221 			goto again;
222 		case CU_STRING:
223 			if ((snprintf(string_ident,
224 			    sizeof(string_ident), "\"%s\"", script->s)) >=
225 			    sizeof(string_ident) - 1)
226 				(void)strcpy(string_ident +
227 				    sizeof(string_ident) - 6, " ...\"");
228 			fname = string_ident;
229 			s = script->s;
230 			state = ST_STRING;
231 			goto again;
232 		}
233 	case ST_FILE:
234 		if ((p = fgets(buf, n, f)) != NULL) {
235 			linenum++;
236 			if (linenum == 1 && buf[0] == '#' && buf[1] == 'n')
237 				nflag = 1;
238 			if (more != NULL)
239 				*more = !feof(f);
240 			return (p);
241 		}
242 		script = script->next;
243 		(void)fclose(f);
244 		state = ST_EOF;
245 		goto again;
246 	case ST_STRING:
247 		if (linenum == 0 && s[0] == '#' && s[1] == 'n')
248 			nflag = 1;
249 		p = buf;
250 		for (;;) {
251 			if (n-- <= 1) {
252 				*p = '\0';
253 				linenum++;
254 				if (more != NULL)
255 					*more = 1;
256 				return (buf);
257 			}
258 			switch (*s) {
259 			case '\0':
260 				state = ST_EOF;
261 				if (s == script->s) {
262 					script = script->next;
263 					goto again;
264 				} else {
265 					script = script->next;
266 					*p = '\0';
267 					linenum++;
268 					if (more != NULL)
269 						*more = 0;
270 					return (buf);
271 				}
272 			case '\n':
273 				*p++ = '\n';
274 				*p = '\0';
275 				s++;
276 				linenum++;
277 				if (more != NULL)
278 					*more = 0;
279 				return (buf);
280 			default:
281 				*p++ = *s++;
282 			}
283 		}
284 	}
285 	/* NOTREACHED */
286 	return (NULL);
287 }
288 
289 /*
290  * Like fgets, but go through the list of files chaining them together.
291  * Set len to the length of the line.
292  */
293 int
294 mf_fgets(SPACE *sp, enum e_spflag spflag)
295 {
296 	struct stat sb;
297 	size_t len;
298 	char *p;
299 	int c;
300 	static int firstfile;
301 
302 	if (infile == NULL) {
303 		/* stdin? */
304 		if (files->fname == NULL) {
305 			if (inplace != NULL)
306 				errx(1, "-i may not be used with stdin");
307 			infile = stdin;
308 			fname = "stdin";
309 			outfile = stdout;
310 			outfname = "stdout";
311 		}
312 		firstfile = 1;
313 	}
314 
315 	for (;;) {
316 		if (infile != NULL && (c = getc(infile)) != EOF) {
317 			(void)ungetc(c, infile);
318 			break;
319 		}
320 		/* If we are here then either eof or no files are open yet */
321 		if (infile == stdin) {
322 			sp->len = 0;
323 			return (0);
324 		}
325 		if (infile != NULL) {
326 			fclose(infile);
327 			if (*oldfname != '\0') {
328 				if (rename(fname, oldfname) != 0) {
329 					warn("rename()");
330 					unlink(tmpfname);
331 					exit(1);
332 				}
333 				*oldfname = '\0';
334 			}
335 			if (*tmpfname != '\0') {
336 				if (outfile != NULL && outfile != stdout)
337 					fclose(outfile);
338 				outfile = NULL;
339 				rename(tmpfname, fname);
340 				*tmpfname = '\0';
341 			}
342 			outfname = NULL;
343 		}
344 		if (firstfile == 0)
345 			files = files->next;
346 		else
347 			firstfile = 0;
348 		if (files == NULL) {
349 			sp->len = 0;
350 			return (0);
351 		}
352 		fname = files->fname;
353 		if (inplace != NULL) {
354 			if (lstat(fname, &sb) != 0)
355 				err(1, "%s", fname);
356 			if (!(sb.st_mode & S_IFREG))
357 				errx(1, "%s: %s %s", fname,
358 				    "in-place editing only",
359 				    "works for regular files");
360 			if (*inplace != '\0') {
361 				strlcpy(oldfname, fname,
362 				    sizeof(oldfname));
363 				len = strlcat(oldfname, inplace,
364 				    sizeof(oldfname));
365 				if (len > sizeof(oldfname))
366 					errx(1, "%s: name too long", fname);
367 			}
368 			len = snprintf(tmpfname, sizeof(tmpfname),
369 			    "%s/.!%ld!%s", dirname(fname), (long)getpid(),
370 			    basename(fname));
371 			if (len >= sizeof(tmpfname))
372 				errx(1, "%s: name too long", fname);
373 			unlink(tmpfname);
374 			if ((outfile = fopen(tmpfname, "w")) == NULL)
375 				err(1, "%s", fname);
376 			fchown(fileno(outfile), sb.st_uid, sb.st_gid);
377 			fchmod(fileno(outfile), sb.st_mode & ALLPERMS);
378 			outfname = tmpfname;
379 		} else {
380 			outfile = stdout;
381 			outfname = "stdout";
382 		}
383 		if ((infile = fopen(fname, "r")) == NULL) {
384 			warn("%s", fname);
385 			rval = 1;
386 			continue;
387 		}
388 	}
389 	/*
390 	 * We are here only when infile is open and we still have something
391 	 * to read from it.
392 	 *
393 	 * Use fgetln so that we can handle essentially infinite input data.
394 	 * Can't use the pointer into the stdio buffer as the process space
395 	 * because the ungetc() can cause it to move.
396 	 */
397 	p = fgetln(infile, &len);
398 	if (ferror(infile))
399 		errx(1, "%s: %s", fname, strerror(errno ? errno : EIO));
400 	if (len != 0 && p[len - 1] == '\n')
401 		len--;
402 	cspace(sp, p, len, spflag);
403 
404 	linenum++;
405 
406 	return (1);
407 }
408 
409 /*
410  * Add a compilation unit to the linked list
411  */
412 static void
413 add_compunit(enum e_cut type, char *s)
414 {
415 	struct s_compunit *cu;
416 
417 	if ((cu = malloc(sizeof(struct s_compunit))) == NULL)
418 		err(1, "malloc");
419 	cu->type = type;
420 	cu->s = s;
421 	cu->next = NULL;
422 	*cu_nextp = cu;
423 	cu_nextp = &cu->next;
424 }
425 
426 /*
427  * Add a file to the linked list
428  */
429 static void
430 add_file(char *s)
431 {
432 	struct s_flist *fp;
433 
434 	if ((fp = malloc(sizeof(struct s_flist))) == NULL)
435 		err(1, "malloc");
436 	fp->next = NULL;
437 	*fl_nextp = fp;
438 	fp->fname = s;
439 	fl_nextp = &fp->next;
440 }
441 
442 int
443 lastline(void)
444 {
445 	int ch;
446 
447 	if (files->next != NULL)
448 		return (0);
449 	if ((ch = getc(infile)) == EOF)
450 		return (1);
451 	ungetc(ch, infile);
452 	return (0);
453 }
454