xref: /minix/usr.bin/sed/main.c (revision 0a6a1f1d)
1 /*	$NetBSD: main.c,v 1.34 2015/03/12 12:40:41 christos Exp $	*/
2 
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 #if HAVE_NBTOOL_CONFIG_H
38 #include "nbtool_config.h"
39 #endif
40 
41 #include <sys/cdefs.h>
42 __RCSID("$NetBSD: main.c,v 1.34 2015/03/12 12:40:41 christos Exp $");
43 #ifdef __FBSDID
44 __FBSDID("$FreeBSD: head/usr.bin/sed/main.c 252231 2013-06-26 04:14:19Z pfg $");
45 #endif
46 
47 #ifndef lint
48 __COPYRIGHT("@(#) Copyright (c) 1992, 1993\
49 	The Regents of the University of California.  All rights reserved.");
50 #endif
51 
52 #if 0
53 static const char sccsid[] = "@(#)main.c	8.2 (Berkeley) 1/3/94";
54 #endif
55 
56 #include <sys/types.h>
57 #include <sys/mman.h>
58 #include <sys/param.h>
59 #include <sys/stat.h>
60 
61 #include <err.h>
62 #include <errno.h>
63 #include <fcntl.h>
64 #include <libgen.h>
65 #include <limits.h>
66 #include <locale.h>
67 #include <regex.h>
68 #include <stddef.h>
69 #define _WITH_GETLINE
70 #include <stdio.h>
71 #include <stdlib.h>
72 #include <string.h>
73 #include <unistd.h>
74 
75 #include "defs.h"
76 #include "extern.h"
77 
78 /*
79  * Linked list of units (strings and files) to be compiled
80  */
81 struct s_compunit {
82 	struct s_compunit *next;
83 	enum e_cut {CU_FILE, CU_STRING} type;
84 	char *s;			/* Pointer to string or fname */
85 };
86 
87 /*
88  * Linked list pointer to compilation units and pointer to current
89  * next pointer.
90  */
91 static struct s_compunit *script, **cu_nextp = &script;
92 
93 /*
94  * Linked list of files to be processed
95  */
96 struct s_flist {
97 	char *fname;
98 	struct s_flist *next;
99 };
100 
101 /*
102  * Linked list pointer to files and pointer to current
103  * next pointer.
104  */
105 static struct s_flist *files, **fl_nextp = &files;
106 
107 FILE *infile;			/* Current input file */
108 FILE *outfile;			/* Current output file */
109 
110 int aflag, eflag, nflag;
111 int rflags = 0;
112 static int rval;		/* Exit status */
113 
114 static int ispan;		/* Whether inplace editing spans across files */
115 
116 /*
117  * Current file and line number; line numbers restart across compilation
118  * units, but span across input files.  The latter is optional if editing
119  * in place.
120  */
121 const char *fname;		/* File name. */
122 const char *outfname;		/* Output file name */
123 static char oldfname[PATH_MAX];	/* Old file name (for in-place editing) */
124 static char tmpfname[PATH_MAX];	/* Temporary file name (for in-place editing) */
125 static const char *inplace;	/* Inplace edit file extension. */
126 u_long linenum;
127 
128 static void add_compunit(enum e_cut, char *);
129 static void add_file(char *);
130 static void usage(void) __dead;
131 
132 int
main(int argc,char * argv[])133 main(int argc, char *argv[])
134 {
135 	int c, fflag;
136 	char *temp_arg;
137 
138 	setprogname(argv[0]);
139 	(void) setlocale(LC_ALL, "");
140 
141 	fflag = 0;
142 	inplace = NULL;
143 
144 	while ((c = getopt(argc, argv, "EI::ae:f:i::lnru")) != -1)
145 		switch (c) {
146 		case 'r':		/* Gnu sed compat */
147 		case 'E':
148 			rflags = REG_EXTENDED;
149 			break;
150 		case 'I':
151 			inplace = optarg ? optarg : __UNCONST("");
152 			ispan = 1;	/* span across input files */
153 			break;
154 		case 'a':
155 			aflag = 1;
156 			break;
157 		case 'e':
158 			eflag = 1;
159 			temp_arg = xmalloc(strlen(optarg) + 2);
160 			strcpy(temp_arg, optarg);
161 			strcat(temp_arg, "\n");
162 			add_compunit(CU_STRING, temp_arg);
163 			break;
164 		case 'f':
165 			fflag = 1;
166 			add_compunit(CU_FILE, optarg);
167 			break;
168 		case 'i':
169 			inplace = optarg ? optarg : __UNCONST("");
170 			ispan = 0;	/* don't span across input files */
171 			break;
172 		case 'l':
173 #ifdef _IOLBF
174 			c = setvbuf(stdout, NULL, _IOLBF, 0);
175 #else
176 			c = setlinebuf(stdout);
177 #endif
178 			if (c)
179 				warn("setting line buffered output failed");
180 			break;
181 		case 'n':
182 			nflag = 1;
183 			break;
184 		case 'u':
185 #ifdef _IONBF
186 			c = setvbuf(stdout, NULL, _IONBF, 0);
187 #else
188 			c = -1;
189 			errno = EOPNOTSUPP;
190 #endif
191 			if (c)
192 				warn("setting unbuffered output failed");
193 			break;
194 		default:
195 		case '?':
196 			usage();
197 		}
198 	argc -= optind;
199 	argv += optind;
200 
201 	/* First usage case; script is the first arg */
202 	if (!eflag && !fflag && *argv) {
203 		add_compunit(CU_STRING, *argv);
204 		argv++;
205 	}
206 
207 	compile();
208 
209 	/* Continue with first and start second usage */
210 	if (*argv)
211 		for (; *argv; argv++)
212 			add_file(*argv);
213 	else
214 		add_file(NULL);
215 	process();
216 	cfclose(prog, NULL);
217 	if (fclose(stdout))
218 		err(1, "stdout");
219 	exit(rval);
220 }
221 
222 static void
usage(void)223 usage(void)
224 {
225 	(void)fprintf(stderr,
226 	    "Usage:  %s [-aElnru] command [file ...]\n"
227 	    "\t%s [-aElnru] [-e command] [-f command_file] [-I[extension]]\n"
228 	    "\t    [-i[extension]] [file ...]\n", getprogname(), getprogname());
229 	exit(1);
230 }
231 
232 /*
233  * Like fgets, but go through the chain of compilation units chaining them
234  * together.  Empty strings and files are ignored.
235  */
236 char *
cu_fgets(char * buf,int n,int * more)237 cu_fgets(char *buf, int n, int *more)
238 {
239 	static enum {ST_EOF, ST_FILE, ST_STRING} state = ST_EOF;
240 	static FILE *f;		/* Current open file */
241 	static char *s;		/* Current pointer inside string */
242 	static char string_ident[30];
243 	char *p;
244 
245 again:
246 	switch (state) {
247 	case ST_EOF:
248 		if (script == NULL) {
249 			if (more != NULL)
250 				*more = 0;
251 			return (NULL);
252 		}
253 		linenum = 0;
254 		switch (script->type) {
255 		case CU_FILE:
256 			if ((f = fopen(script->s, "r")) == NULL)
257 				err(1, "%s", script->s);
258 			fname = script->s;
259 			state = ST_FILE;
260 			goto again;
261 		case CU_STRING:
262 			if (((size_t)snprintf(string_ident,
263 			    sizeof(string_ident), "\"%s\"", script->s)) >=
264 			    sizeof(string_ident) - 1)
265 				(void)strcpy(string_ident +
266 				    sizeof(string_ident) - 6, " ...\"");
267 			fname = string_ident;
268 			s = script->s;
269 			state = ST_STRING;
270 			goto again;
271 		}
272 	case ST_FILE:
273 		if ((p = fgets(buf, n, f)) != NULL) {
274 			linenum++;
275 			if (linenum == 1 && buf[0] == '#' && buf[1] == 'n')
276 				nflag = 1;
277 			if (more != NULL)
278 				*more = !feof(f);
279 			return (p);
280 		}
281 		script = script->next;
282 		(void)fclose(f);
283 		state = ST_EOF;
284 		goto again;
285 	case ST_STRING:
286 		if (linenum == 0 && s[0] == '#' && s[1] == 'n')
287 			nflag = 1;
288 		p = buf;
289 		for (;;) {
290 			if (n-- <= 1) {
291 				*p = '\0';
292 				linenum++;
293 				if (more != NULL)
294 					*more = 1;
295 				return (buf);
296 			}
297 			switch (*s) {
298 			case '\0':
299 				state = ST_EOF;
300 				if (s == script->s) {
301 					script = script->next;
302 					goto again;
303 				} else {
304 					script = script->next;
305 					*p = '\0';
306 					linenum++;
307 					if (more != NULL)
308 						*more = 0;
309 					return (buf);
310 				}
311 			case '\n':
312 				*p++ = '\n';
313 				*p = '\0';
314 				s++;
315 				linenum++;
316 				if (more != NULL)
317 					*more = 0;
318 				return (buf);
319 			default:
320 				*p++ = *s++;
321 			}
322 		}
323 	}
324 	/* NOTREACHED */
325 	return (NULL);
326 }
327 
328 /*
329  * Like fgets, but go through the list of files chaining them together.
330  * Set len to the length of the line.
331  */
332 int
mf_fgets(SPACE * sp,enum e_spflag spflag)333 mf_fgets(SPACE *sp, enum e_spflag spflag)
334 {
335 	struct stat sb;
336 	size_t len;
337 	static char *p = NULL;
338 	static size_t plen = 0;
339 	int c;
340 	static int firstfile;
341 
342 	if (infile == NULL) {
343 		/* stdin? */
344 		if (files->fname == NULL) {
345 			if (inplace != NULL)
346 				errx(1, "-I or -i may not be used with stdin");
347 			infile = stdin;
348 			fname = "stdin";
349 			outfile = stdout;
350 			outfname = "stdout";
351 		}
352 		firstfile = 1;
353 	}
354 
355 	for (;;) {
356 		if (infile != NULL && (c = getc(infile)) != EOF) {
357 			(void)ungetc(c, infile);
358 			break;
359 		}
360 		/* If we are here then either eof or no files are open yet */
361 		if (infile == stdin) {
362 			sp->len = 0;
363 			return (0);
364 		}
365 		if (infile != NULL) {
366 			fclose(infile);
367 			if (*oldfname != '\0') {
368 				/* if there was a backup file, remove it */
369 				unlink(oldfname);
370 				/*
371 				 * Backup the original.  Note that hard links
372 				 * are not supported on all filesystems.
373 				 */
374 				if ((link(fname, oldfname) != 0) &&
375 				   (rename(fname, oldfname) != 0)) {
376 					warn("rename()");
377 					if (*tmpfname)
378 						unlink(tmpfname);
379 					exit(1);
380 				}
381 				*oldfname = '\0';
382 			}
383 			if (*tmpfname != '\0') {
384 				if (outfile != NULL && outfile != stdout)
385 					if (fclose(outfile) != 0) {
386 						warn("fclose()");
387 						unlink(tmpfname);
388 						exit(1);
389 					}
390 				outfile = NULL;
391 				if (rename(tmpfname, fname) != 0) {
392 					/* this should not happen really! */
393 					warn("rename()");
394 					unlink(tmpfname);
395 					exit(1);
396 				}
397 				*tmpfname = '\0';
398 			}
399 			outfname = NULL;
400 		}
401 		if (firstfile == 0)
402 			files = files->next;
403 		else
404 			firstfile = 0;
405 		if (files == NULL) {
406 			sp->len = 0;
407 			return (0);
408 		}
409 		fname = files->fname;
410 		if (inplace != NULL) {
411 			if (lstat(fname, &sb) != 0)
412 				err(1, "%s", fname);
413 			if (!(sb.st_mode & S_IFREG))
414 				errx(1, "%s: %s %s", fname,
415 				    "in-place editing only",
416 				    "works for regular files");
417 			if (*inplace != '\0') {
418 				strlcpy(oldfname, fname,
419 				    sizeof(oldfname));
420 				len = strlcat(oldfname, inplace,
421 				    sizeof(oldfname));
422 				if (len > sizeof(oldfname))
423 					errx(1, "%s: name too long", fname);
424 			}
425 			char d_name[PATH_MAX], f_name[PATH_MAX];
426 			(void)strlcpy(d_name, fname, sizeof(d_name));
427 			(void)strlcpy(f_name, fname, sizeof(f_name));
428 			len = (size_t)snprintf(tmpfname, sizeof(tmpfname),
429 			    "%s/.!%ld!%s", dirname(d_name), (long)getpid(),
430 			    basename(f_name));
431 			if (len >= sizeof(tmpfname))
432 				errx(1, "%s: name too long", fname);
433 			unlink(tmpfname);
434 			if (outfile != NULL && outfile != stdout)
435 				fclose(outfile);
436 			if ((outfile = fopen(tmpfname, "w")) == NULL)
437 				err(1, "%s", fname);
438 			fchown(fileno(outfile), sb.st_uid, sb.st_gid);
439 			fchmod(fileno(outfile), sb.st_mode & ALLPERMS);
440 			outfname = tmpfname;
441 			if (!ispan) {
442 				linenum = 0;
443 				resetstate();
444 			}
445 		} else {
446 			outfile = stdout;
447 			outfname = "stdout";
448 		}
449 		if ((infile = fopen(fname, "r")) == NULL) {
450 			warn("%s", fname);
451 			rval = 1;
452 			continue;
453 		}
454 	}
455 	/*
456 	 * We are here only when infile is open and we still have something
457 	 * to read from it.
458 	 *
459 	 * Use getline() so that we can handle essentially infinite input
460 	 * data.  The p and plen are static so each invocation gives
461 	 * getline() the same buffer which is expanded as needed.
462 	 */
463 	ssize_t slen = getline(&p, &plen, infile);
464 	if (slen == -1)
465 		err(1, "%s", fname);
466 	if (slen != 0 && p[slen - 1] == '\n')
467 		slen--;
468 	cspace(sp, p, (size_t)slen, spflag);
469 
470 	linenum++;
471 
472 	return (1);
473 }
474 
475 /*
476  * Add a compilation unit to the linked list
477  */
478 static void
add_compunit(enum e_cut type,char * s)479 add_compunit(enum e_cut type, char *s)
480 {
481 	struct s_compunit *cu;
482 
483 	cu = xmalloc(sizeof(struct s_compunit));
484 	cu->type = type;
485 	cu->s = s;
486 	cu->next = NULL;
487 	*cu_nextp = cu;
488 	cu_nextp = &cu->next;
489 }
490 
491 /*
492  * Add a file to the linked list
493  */
494 static void
add_file(char * s)495 add_file(char *s)
496 {
497 	struct s_flist *fp;
498 
499 	fp = xmalloc(sizeof(struct s_flist));
500 	fp->next = NULL;
501 	*fl_nextp = fp;
502 	fp->fname = s;
503 	fl_nextp = &fp->next;
504 }
505 
506 int
lastline(void)507 lastline(void)
508 {
509 	int ch;
510 
511 	if (files->next != NULL && (inplace == NULL || ispan))
512 		return (0);
513 	if ((ch = getc(infile)) == EOF)
514 		return (1);
515 	ungetc(ch, infile);
516 	return (0);
517 }
518