xref: /freebsd/usr.bin/sed/main.c (revision 1e413cf93298b5b97441a21d9a50fdcd0ee9945e)
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 <libgen.h>
56 #include <limits.h>
57 #include <locale.h>
58 #include <regex.h>
59 #include <stddef.h>
60 #include <stdio.h>
61 #include <stdlib.h>
62 #include <string.h>
63 #include <unistd.h>
64 
65 #include "defs.h"
66 #include "extern.h"
67 
68 /*
69  * Linked list of units (strings and files) to be compiled
70  */
71 struct s_compunit {
72 	struct s_compunit *next;
73 	enum e_cut {CU_FILE, CU_STRING} type;
74 	char *s;			/* Pointer to string or fname */
75 };
76 
77 /*
78  * Linked list pointer to compilation units and pointer to current
79  * next pointer.
80  */
81 static struct s_compunit *script, **cu_nextp = &script;
82 
83 /*
84  * Linked list of files to be processed
85  */
86 struct s_flist {
87 	char *fname;
88 	struct s_flist *next;
89 };
90 
91 /*
92  * Linked list pointer to files and pointer to current
93  * next pointer.
94  */
95 static struct s_flist *files, **fl_nextp = &files;
96 
97 FILE *infile;			/* Current input file */
98 FILE *outfile;			/* Current output file */
99 
100 int aflag, eflag, nflag;
101 int rflags = 0;
102 static int rval;		/* Exit status */
103 
104 static int ispan;		/* Whether inplace editing spans across files */
105 
106 /*
107  * Current file and line number; line numbers restart across compilation
108  * units, but span across input files.  The latter is optional if editing
109  * in place.
110  */
111 const char *fname;		/* File name. */
112 const char *outfname;		/* Output file name */
113 static char oldfname[PATH_MAX];	/* Old file name (for in-place editing) */
114 static char tmpfname[PATH_MAX];	/* Temporary file name (for in-place editing) */
115 static const char *inplace;	/* Inplace edit file extension. */
116 u_long linenum;
117 
118 static void add_compunit(enum e_cut, char *);
119 static void add_file(char *);
120 static void usage(void);
121 
122 int
123 main(int argc, char *argv[])
124 {
125 	int c, fflag;
126 	char *temp_arg;
127 
128 	(void) setlocale(LC_ALL, "");
129 
130 	fflag = 0;
131 	inplace = NULL;
132 
133 	while ((c = getopt(argc, argv, "EI:ae:f:i:ln")) != -1)
134 		switch (c) {
135 		case 'E':
136 			rflags = REG_EXTENDED;
137 			break;
138 		case 'I':
139 			inplace = optarg;
140 			ispan = 1;	/* span across input files */
141 			break;
142 		case 'a':
143 			aflag = 1;
144 			break;
145 		case 'e':
146 			eflag = 1;
147 			if ((temp_arg = malloc(strlen(optarg) + 2)) == NULL)
148 				err(1, "malloc");
149 			strcpy(temp_arg, optarg);
150 			strcat(temp_arg, "\n");
151 			add_compunit(CU_STRING, temp_arg);
152 			break;
153 		case 'f':
154 			fflag = 1;
155 			add_compunit(CU_FILE, optarg);
156 			break;
157 		case 'i':
158 			inplace = optarg;
159 			ispan = 0;	/* don't span across input files */
160 			break;
161 		case 'l':
162 			if(setlinebuf(stdout) != 0)
163 				warnx("setlinebuf() failed");
164 			break;
165 		case 'n':
166 			nflag = 1;
167 			break;
168 		default:
169 		case '?':
170 			usage();
171 		}
172 	argc -= optind;
173 	argv += optind;
174 
175 	/* First usage case; script is the first arg */
176 	if (!eflag && !fflag && *argv) {
177 		add_compunit(CU_STRING, *argv);
178 		argv++;
179 	}
180 
181 	compile();
182 
183 	/* Continue with first and start second usage */
184 	if (*argv)
185 		for (; *argv; argv++)
186 			add_file(*argv);
187 	else
188 		add_file(NULL);
189 	process();
190 	cfclose(prog, NULL);
191 	if (fclose(stdout))
192 		err(1, "stdout");
193 	exit(rval);
194 }
195 
196 static void
197 usage(void)
198 {
199 	(void)fprintf(stderr, "%s\n%s\n",
200 		"usage: sed script [-Ealn] [-i extension] [file ...]",
201 		"       sed [-Ealn] [-i extension] [-e script] ... [-f script_file] ... [file ...]");
202 	exit(1);
203 }
204 
205 /*
206  * Like fgets, but go through the chain of compilation units chaining them
207  * together.  Empty strings and files are ignored.
208  */
209 char *
210 cu_fgets(char *buf, int n, int *more)
211 {
212 	static enum {ST_EOF, ST_FILE, ST_STRING} state = ST_EOF;
213 	static FILE *f;		/* Current open file */
214 	static char *s;		/* Current pointer inside string */
215 	static char string_ident[30];
216 	char *p;
217 
218 again:
219 	switch (state) {
220 	case ST_EOF:
221 		if (script == NULL) {
222 			if (more != NULL)
223 				*more = 0;
224 			return (NULL);
225 		}
226 		linenum = 0;
227 		switch (script->type) {
228 		case CU_FILE:
229 			if ((f = fopen(script->s, "r")) == NULL)
230 				err(1, "%s", script->s);
231 			fname = script->s;
232 			state = ST_FILE;
233 			goto again;
234 		case CU_STRING:
235 			if ((snprintf(string_ident,
236 			    sizeof(string_ident), "\"%s\"", script->s)) >=
237 			    sizeof(string_ident) - 1)
238 				(void)strcpy(string_ident +
239 				    sizeof(string_ident) - 6, " ...\"");
240 			fname = string_ident;
241 			s = script->s;
242 			state = ST_STRING;
243 			goto again;
244 		}
245 	case ST_FILE:
246 		if ((p = fgets(buf, n, f)) != NULL) {
247 			linenum++;
248 			if (linenum == 1 && buf[0] == '#' && buf[1] == 'n')
249 				nflag = 1;
250 			if (more != NULL)
251 				*more = !feof(f);
252 			return (p);
253 		}
254 		script = script->next;
255 		(void)fclose(f);
256 		state = ST_EOF;
257 		goto again;
258 	case ST_STRING:
259 		if (linenum == 0 && s[0] == '#' && s[1] == 'n')
260 			nflag = 1;
261 		p = buf;
262 		for (;;) {
263 			if (n-- <= 1) {
264 				*p = '\0';
265 				linenum++;
266 				if (more != NULL)
267 					*more = 1;
268 				return (buf);
269 			}
270 			switch (*s) {
271 			case '\0':
272 				state = ST_EOF;
273 				if (s == script->s) {
274 					script = script->next;
275 					goto again;
276 				} else {
277 					script = script->next;
278 					*p = '\0';
279 					linenum++;
280 					if (more != NULL)
281 						*more = 0;
282 					return (buf);
283 				}
284 			case '\n':
285 				*p++ = '\n';
286 				*p = '\0';
287 				s++;
288 				linenum++;
289 				if (more != NULL)
290 					*more = 0;
291 				return (buf);
292 			default:
293 				*p++ = *s++;
294 			}
295 		}
296 	}
297 	/* NOTREACHED */
298 	return (NULL);
299 }
300 
301 /*
302  * Like fgets, but go through the list of files chaining them together.
303  * Set len to the length of the line.
304  */
305 int
306 mf_fgets(SPACE *sp, enum e_spflag spflag)
307 {
308 	struct stat sb;
309 	size_t len;
310 	char *p;
311 	int c;
312 	static int firstfile;
313 
314 	if (infile == NULL) {
315 		/* stdin? */
316 		if (files->fname == NULL) {
317 			if (inplace != NULL)
318 				errx(1, "-I or -i may not be used with stdin");
319 			infile = stdin;
320 			fname = "stdin";
321 			outfile = stdout;
322 			outfname = "stdout";
323 		}
324 		firstfile = 1;
325 	}
326 
327 	for (;;) {
328 		if (infile != NULL && (c = getc(infile)) != EOF) {
329 			(void)ungetc(c, infile);
330 			break;
331 		}
332 		/* If we are here then either eof or no files are open yet */
333 		if (infile == stdin) {
334 			sp->len = 0;
335 			return (0);
336 		}
337 		if (infile != NULL) {
338 			fclose(infile);
339 			if (*oldfname != '\0') {
340 				if (rename(fname, oldfname) != 0) {
341 					warn("rename()");
342 					unlink(tmpfname);
343 					exit(1);
344 				}
345 				*oldfname = '\0';
346 			}
347 			if (*tmpfname != '\0') {
348 				if (outfile != NULL && outfile != stdout)
349 					fclose(outfile);
350 				outfile = NULL;
351 				rename(tmpfname, fname);
352 				*tmpfname = '\0';
353 			}
354 			outfname = NULL;
355 		}
356 		if (firstfile == 0)
357 			files = files->next;
358 		else
359 			firstfile = 0;
360 		if (files == NULL) {
361 			sp->len = 0;
362 			return (0);
363 		}
364 		fname = files->fname;
365 		if (inplace != NULL) {
366 			if (lstat(fname, &sb) != 0)
367 				err(1, "%s", fname);
368 			if (!(sb.st_mode & S_IFREG))
369 				errx(1, "%s: %s %s", fname,
370 				    "in-place editing only",
371 				    "works for regular files");
372 			if (*inplace != '\0') {
373 				strlcpy(oldfname, fname,
374 				    sizeof(oldfname));
375 				len = strlcat(oldfname, inplace,
376 				    sizeof(oldfname));
377 				if (len > sizeof(oldfname))
378 					errx(1, "%s: name too long", fname);
379 			}
380 			len = snprintf(tmpfname, sizeof(tmpfname),
381 			    "%s/.!%ld!%s", dirname(fname), (long)getpid(),
382 			    basename(fname));
383 			if (len >= sizeof(tmpfname))
384 				errx(1, "%s: name too long", fname);
385 			unlink(tmpfname);
386 			if ((outfile = fopen(tmpfname, "w")) == NULL)
387 				err(1, "%s", fname);
388 			fchown(fileno(outfile), sb.st_uid, sb.st_gid);
389 			fchmod(fileno(outfile), sb.st_mode & ALLPERMS);
390 			outfname = tmpfname;
391 			if (!ispan) {
392 				linenum = 0;
393 				resetstate();
394 			}
395 		} else {
396 			outfile = stdout;
397 			outfname = "stdout";
398 		}
399 		if ((infile = fopen(fname, "r")) == NULL) {
400 			warn("%s", fname);
401 			rval = 1;
402 			continue;
403 		}
404 	}
405 	/*
406 	 * We are here only when infile is open and we still have something
407 	 * to read from it.
408 	 *
409 	 * Use fgetln so that we can handle essentially infinite input data.
410 	 * Can't use the pointer into the stdio buffer as the process space
411 	 * because the ungetc() can cause it to move.
412 	 */
413 	p = fgetln(infile, &len);
414 	if (ferror(infile))
415 		errx(1, "%s: %s", fname, strerror(errno ? errno : EIO));
416 	if (len != 0 && p[len - 1] == '\n')
417 		len--;
418 	cspace(sp, p, len, spflag);
419 
420 	linenum++;
421 
422 	return (1);
423 }
424 
425 /*
426  * Add a compilation unit to the linked list
427  */
428 static void
429 add_compunit(enum e_cut type, char *s)
430 {
431 	struct s_compunit *cu;
432 
433 	if ((cu = malloc(sizeof(struct s_compunit))) == NULL)
434 		err(1, "malloc");
435 	cu->type = type;
436 	cu->s = s;
437 	cu->next = NULL;
438 	*cu_nextp = cu;
439 	cu_nextp = &cu->next;
440 }
441 
442 /*
443  * Add a file to the linked list
444  */
445 static void
446 add_file(char *s)
447 {
448 	struct s_flist *fp;
449 
450 	if ((fp = malloc(sizeof(struct s_flist))) == NULL)
451 		err(1, "malloc");
452 	fp->next = NULL;
453 	*fl_nextp = fp;
454 	fp->fname = s;
455 	fl_nextp = &fp->next;
456 }
457 
458 int
459 lastline(void)
460 {
461 	int ch;
462 
463 	if (files->next != NULL && (inplace == NULL || ispan))
464 		return (0);
465 	if ((ch = getc(infile)) == EOF)
466 		return (1);
467 	ungetc(ch, infile);
468 	return (0);
469 }
470