xref: /freebsd/usr.bin/sed/main.c (revision f39bffc62c1395bde25d152c7f68fdf7cbaab414)
1 /*-
2  * SPDX-License-Identifier: BSD-3-Clause
3  *
4  * Copyright (c) 2013 Johann 'Myrkraverk' Oskarsson.
5  * Copyright (c) 1992 Diomidis Spinellis.
6  * Copyright (c) 1992, 1993
7  *	The Regents of the University of California.  All rights reserved.
8  *
9  * This code is derived from software contributed to Berkeley by
10  * Diomidis Spinellis of Imperial College, University of London.
11  *
12  * Redistribution and use in source and binary forms, with or without
13  * modification, are permitted provided that the following conditions
14  * are met:
15  * 1. Redistributions of source code must retain the above copyright
16  *    notice, this list of conditions and the following disclaimer.
17  * 2. Redistributions in binary form must reproduce the above copyright
18  *    notice, this list of conditions and the following disclaimer in the
19  *    documentation and/or other materials provided with the distribution.
20  * 3. Neither the name of the University nor the names of its contributors
21  *    may be used to endorse or promote products derived from this software
22  *    without specific prior written permission.
23  *
24  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
25  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
26  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
27  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
28  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
29  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
30  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
31  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
32  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
33  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
34  * SUCH DAMAGE.
35  */
36 
37 #include <sys/cdefs.h>
38 __FBSDID("$FreeBSD$");
39 
40 #ifndef lint
41 static const char copyright[] =
42 "@(#) Copyright (c) 1992, 1993\n\
43 	The Regents of the University of California.  All rights reserved.\n";
44 #endif
45 
46 #ifndef lint
47 static const char sccsid[] = "@(#)main.c	8.2 (Berkeley) 1/3/94";
48 #endif
49 
50 #include <sys/types.h>
51 #include <sys/mman.h>
52 #include <sys/param.h>
53 #include <sys/stat.h>
54 
55 #include <err.h>
56 #include <errno.h>
57 #include <fcntl.h>
58 #include <libgen.h>
59 #include <limits.h>
60 #include <locale.h>
61 #include <regex.h>
62 #include <stddef.h>
63 #include <stdio.h>
64 #include <stdlib.h>
65 #include <string.h>
66 #include <unistd.h>
67 
68 #include "defs.h"
69 #include "extern.h"
70 
71 /*
72  * Linked list of units (strings and files) to be compiled
73  */
74 struct s_compunit {
75 	struct s_compunit *next;
76 	enum e_cut {CU_FILE, CU_STRING} type;
77 	char *s;			/* Pointer to string or fname */
78 };
79 
80 /*
81  * Linked list pointer to compilation units and pointer to current
82  * next pointer.
83  */
84 static struct s_compunit *script, **cu_nextp = &script;
85 
86 /*
87  * Linked list of files to be processed
88  */
89 struct s_flist {
90 	char *fname;
91 	struct s_flist *next;
92 };
93 
94 /*
95  * Linked list pointer to files and pointer to current
96  * next pointer.
97  */
98 static struct s_flist *files, **fl_nextp = &files;
99 
100 FILE *infile;			/* Current input file */
101 FILE *outfile;			/* Current output file */
102 
103 int aflag, eflag, nflag;
104 int rflags = 0;
105 static int rval;		/* Exit status */
106 
107 static int ispan;		/* Whether inplace editing spans across files */
108 
109 /*
110  * Current file and line number; line numbers restart across compilation
111  * units, but span across input files.  The latter is optional if editing
112  * in place.
113  */
114 const char *fname;		/* File name. */
115 const char *outfname;		/* Output file name */
116 static char oldfname[PATH_MAX];	/* Old file name (for in-place editing) */
117 static char tmpfname[PATH_MAX];	/* Temporary file name (for in-place editing) */
118 static const char *inplace;	/* Inplace edit file extension. */
119 u_long linenum;
120 
121 static void add_compunit(enum e_cut, char *);
122 static void add_file(char *);
123 static void usage(void);
124 
125 int
126 main(int argc, char *argv[])
127 {
128 	int c, fflag;
129 	char *temp_arg;
130 
131 	(void) setlocale(LC_ALL, "");
132 
133 	fflag = 0;
134 	inplace = NULL;
135 
136 	while ((c = getopt(argc, argv, "EI:ae:f:i:lnru")) != -1)
137 		switch (c) {
138 		case 'r':		/* Gnu sed compat */
139 		case 'E':
140 			rflags = REG_EXTENDED;
141 			break;
142 		case 'I':
143 			inplace = optarg;
144 			ispan = 1;	/* span across input files */
145 			break;
146 		case 'a':
147 			aflag = 1;
148 			break;
149 		case 'e':
150 			eflag = 1;
151 			if ((temp_arg = malloc(strlen(optarg) + 2)) == NULL)
152 				err(1, "malloc");
153 			strcpy(temp_arg, optarg);
154 			strcat(temp_arg, "\n");
155 			add_compunit(CU_STRING, temp_arg);
156 			break;
157 		case 'f':
158 			fflag = 1;
159 			add_compunit(CU_FILE, optarg);
160 			break;
161 		case 'i':
162 			inplace = optarg;
163 			ispan = 0;	/* don't span across input files */
164 			break;
165 		case 'l':
166 			if(setvbuf(stdout, NULL, _IOLBF, 0) != 0)
167 				warnx("setting line buffered output failed");
168 			break;
169 		case 'n':
170 			nflag = 1;
171 			break;
172 		case 'u':
173 			if(setvbuf(stdout, NULL, _IONBF, 0) != 0)
174 				warnx("setting unbuffered output failed");
175 			break;
176 		default:
177 		case '?':
178 			usage();
179 		}
180 	argc -= optind;
181 	argv += optind;
182 
183 	/* First usage case; script is the first arg */
184 	if (!eflag && !fflag && *argv) {
185 		add_compunit(CU_STRING, *argv);
186 		argv++;
187 	}
188 
189 	compile();
190 
191 	/* Continue with first and start second usage */
192 	if (*argv)
193 		for (; *argv; argv++)
194 			add_file(*argv);
195 	else
196 		add_file(NULL);
197 	process();
198 	cfclose(prog, NULL);
199 	if (fclose(stdout))
200 		err(1, "stdout");
201 	exit(rval);
202 }
203 
204 static void
205 usage(void)
206 {
207 	(void)fprintf(stderr,
208 	    "usage: %s script [-Ealnru] [-i extension] [file ...]\n"
209 	    "\t%s [-Ealnu] [-i extension] [-e script] ... [-f script_file]"
210 	    " ... [file ...]\n", getprogname(), getprogname());
211 	exit(1);
212 }
213 
214 /*
215  * Like fgets, but go through the chain of compilation units chaining them
216  * together.  Empty strings and files are ignored.
217  */
218 char *
219 cu_fgets(char *buf, int n, int *more)
220 {
221 	static enum {ST_EOF, ST_FILE, ST_STRING} state = ST_EOF;
222 	static FILE *f;		/* Current open file */
223 	static char *s;		/* Current pointer inside string */
224 	static char string_ident[30];
225 	char *p;
226 
227 again:
228 	switch (state) {
229 	case ST_EOF:
230 		if (script == NULL) {
231 			if (more != NULL)
232 				*more = 0;
233 			return (NULL);
234 		}
235 		linenum = 0;
236 		switch (script->type) {
237 		case CU_FILE:
238 			if ((f = fopen(script->s, "r")) == NULL)
239 				err(1, "%s", script->s);
240 			fname = script->s;
241 			state = ST_FILE;
242 			goto again;
243 		case CU_STRING:
244 			if (((size_t)snprintf(string_ident,
245 			    sizeof(string_ident), "\"%s\"", script->s)) >=
246 			    sizeof(string_ident) - 1)
247 				(void)strcpy(string_ident +
248 				    sizeof(string_ident) - 6, " ...\"");
249 			fname = string_ident;
250 			s = script->s;
251 			state = ST_STRING;
252 			goto again;
253 		}
254 	case ST_FILE:
255 		if ((p = fgets(buf, n, f)) != NULL) {
256 			linenum++;
257 			if (linenum == 1 && buf[0] == '#' && buf[1] == 'n')
258 				nflag = 1;
259 			if (more != NULL)
260 				*more = !feof(f);
261 			return (p);
262 		}
263 		script = script->next;
264 		(void)fclose(f);
265 		state = ST_EOF;
266 		goto again;
267 	case ST_STRING:
268 		if (linenum == 0 && s[0] == '#' && s[1] == 'n')
269 			nflag = 1;
270 		p = buf;
271 		for (;;) {
272 			if (n-- <= 1) {
273 				*p = '\0';
274 				linenum++;
275 				if (more != NULL)
276 					*more = 1;
277 				return (buf);
278 			}
279 			switch (*s) {
280 			case '\0':
281 				state = ST_EOF;
282 				if (s == script->s) {
283 					script = script->next;
284 					goto again;
285 				} else {
286 					script = script->next;
287 					*p = '\0';
288 					linenum++;
289 					if (more != NULL)
290 						*more = 0;
291 					return (buf);
292 				}
293 			case '\n':
294 				*p++ = '\n';
295 				*p = '\0';
296 				s++;
297 				linenum++;
298 				if (more != NULL)
299 					*more = 0;
300 				return (buf);
301 			default:
302 				*p++ = *s++;
303 			}
304 		}
305 	}
306 	/* NOTREACHED */
307 	return (NULL);
308 }
309 
310 /*
311  * Like fgets, but go through the list of files chaining them together.
312  * Set len to the length of the line.
313  */
314 int
315 mf_fgets(SPACE *sp, enum e_spflag spflag)
316 {
317 	struct stat sb;
318 	ssize_t len;
319 	char *dirbuf, *basebuf;
320 	static char *p = NULL;
321 	static size_t plen = 0;
322 	int c;
323 	static int firstfile;
324 
325 	if (infile == NULL) {
326 		/* stdin? */
327 		if (files->fname == NULL) {
328 			if (inplace != NULL)
329 				errx(1, "-I or -i may not be used with stdin");
330 			infile = stdin;
331 			fname = "stdin";
332 			outfile = stdout;
333 			outfname = "stdout";
334 		}
335 		firstfile = 1;
336 	}
337 
338 	for (;;) {
339 		if (infile != NULL && (c = getc(infile)) != EOF) {
340 			(void)ungetc(c, infile);
341 			break;
342 		}
343 		/* If we are here then either eof or no files are open yet */
344 		if (infile == stdin) {
345 			sp->len = 0;
346 			return (0);
347 		}
348 		if (infile != NULL) {
349 			fclose(infile);
350 			if (*oldfname != '\0') {
351 				/* if there was a backup file, remove it */
352 				unlink(oldfname);
353 				/*
354 				 * Backup the original.  Note that hard links
355 				 * are not supported on all filesystems.
356 				 */
357 				if ((link(fname, oldfname) != 0) &&
358 				   (rename(fname, oldfname) != 0)) {
359 					warn("rename()");
360 					if (*tmpfname)
361 						unlink(tmpfname);
362 					exit(1);
363 				}
364 				*oldfname = '\0';
365 			}
366 			if (*tmpfname != '\0') {
367 				if (outfile != NULL && outfile != stdout)
368 					if (fclose(outfile) != 0) {
369 						warn("fclose()");
370 						unlink(tmpfname);
371 						exit(1);
372 					}
373 				outfile = NULL;
374 				if (rename(tmpfname, fname) != 0) {
375 					/* this should not happen really! */
376 					warn("rename()");
377 					unlink(tmpfname);
378 					exit(1);
379 				}
380 				*tmpfname = '\0';
381 			}
382 			outfname = NULL;
383 		}
384 		if (firstfile == 0)
385 			files = files->next;
386 		else
387 			firstfile = 0;
388 		if (files == NULL) {
389 			sp->len = 0;
390 			return (0);
391 		}
392 		fname = files->fname;
393 		if (inplace != NULL) {
394 			if (lstat(fname, &sb) != 0)
395 				err(1, "%s", fname);
396 			if (!S_ISREG(sb.st_mode))
397 				errx(1, "%s: %s %s", fname,
398 				    "in-place editing only",
399 				    "works for regular files");
400 			if (*inplace != '\0') {
401 				strlcpy(oldfname, fname,
402 				    sizeof(oldfname));
403 				len = strlcat(oldfname, inplace,
404 				    sizeof(oldfname));
405 				if (len > (ssize_t)sizeof(oldfname))
406 					errx(1, "%s: name too long", fname);
407 			}
408 			if ((dirbuf = strdup(fname)) == NULL ||
409 			    (basebuf = strdup(fname)) == NULL)
410 				err(1, "strdup");
411 			len = snprintf(tmpfname, sizeof(tmpfname),
412 			    "%s/.!%ld!%s", dirname(dirbuf), (long)getpid(),
413 			    basename(basebuf));
414 			free(dirbuf);
415 			free(basebuf);
416 			if (len >= (ssize_t)sizeof(tmpfname))
417 				errx(1, "%s: name too long", fname);
418 			unlink(tmpfname);
419 			if (outfile != NULL && outfile != stdout)
420 				fclose(outfile);
421 			if ((outfile = fopen(tmpfname, "w")) == NULL)
422 				err(1, "%s", fname);
423 			fchown(fileno(outfile), sb.st_uid, sb.st_gid);
424 			fchmod(fileno(outfile), sb.st_mode & ALLPERMS);
425 			outfname = tmpfname;
426 			if (!ispan) {
427 				linenum = 0;
428 				resetstate();
429 			}
430 		} else {
431 			outfile = stdout;
432 			outfname = "stdout";
433 		}
434 		if ((infile = fopen(fname, "r")) == NULL) {
435 			warn("%s", fname);
436 			rval = 1;
437 			continue;
438 		}
439 	}
440 	/*
441 	 * We are here only when infile is open and we still have something
442 	 * to read from it.
443 	 *
444 	 * Use getline() so that we can handle essentially infinite input
445 	 * data.  The p and plen are static so each invocation gives
446 	 * getline() the same buffer which is expanded as needed.
447 	 */
448 	len = getline(&p, &plen, infile);
449 	if (len == -1)
450 		err(1, "%s", fname);
451 	if (len != 0 && p[len - 1] == '\n') {
452 		sp->append_newline = 1;
453 		len--;
454 	} else if (!lastline()) {
455 		sp->append_newline = 1;
456 	} else {
457 		sp->append_newline = 0;
458 	}
459 	cspace(sp, p, len, spflag);
460 
461 	linenum++;
462 
463 	return (1);
464 }
465 
466 /*
467  * Add a compilation unit to the linked list
468  */
469 static void
470 add_compunit(enum e_cut type, char *s)
471 {
472 	struct s_compunit *cu;
473 
474 	if ((cu = malloc(sizeof(struct s_compunit))) == NULL)
475 		err(1, "malloc");
476 	cu->type = type;
477 	cu->s = s;
478 	cu->next = NULL;
479 	*cu_nextp = cu;
480 	cu_nextp = &cu->next;
481 }
482 
483 /*
484  * Add a file to the linked list
485  */
486 static void
487 add_file(char *s)
488 {
489 	struct s_flist *fp;
490 
491 	if ((fp = malloc(sizeof(struct s_flist))) == NULL)
492 		err(1, "malloc");
493 	fp->next = NULL;
494 	*fl_nextp = fp;
495 	fp->fname = s;
496 	fl_nextp = &fp->next;
497 }
498 
499 static int
500 next_files_have_lines(void)
501 {
502 	struct s_flist *file;
503 	FILE *file_fd;
504 	int ch;
505 
506 	file = files;
507 	while ((file = file->next) != NULL) {
508 		if ((file_fd = fopen(file->fname, "r")) == NULL)
509 			continue;
510 
511 		if ((ch = getc(file_fd)) != EOF) {
512 			/*
513 			 * This next file has content, therefore current
514 			 * file doesn't contains the last line.
515 			 */
516 			ungetc(ch, file_fd);
517 			fclose(file_fd);
518 			return (1);
519 		}
520 
521 		fclose(file_fd);
522 	}
523 
524 	return (0);
525 }
526 
527 int
528 lastline(void)
529 {
530 	int ch;
531 
532 	if (feof(infile)) {
533 		return !(
534 		    (inplace == NULL || ispan) &&
535 		    next_files_have_lines());
536 	}
537 	if ((ch = getc(infile)) == EOF) {
538 		return !(
539 		    (inplace == NULL || ispan) &&
540 		    next_files_have_lines());
541 	}
542 	ungetc(ch, infile);
543 	return (0);
544 }
545