xref: /dflybsd-src/usr.bin/sed/main.c (revision 71c738b2adedcfcf5ae73cc1db9d2b6480bf5135)
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  * 3. 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  * @(#) Copyright (c) 1992, 1993 The Regents of the University of California.  All rights reserved.
34  * @(#)main.c	8.2 (Berkeley) 1/3/94
35  * $FreeBSD: src/usr.bin/sed/main.c,v 1.41 2008/02/09 09:12:02 dwmalone Exp $
36  * $DragonFly: src/usr.bin/sed/main.c,v 1.4 2008/04/08 13:23:38 swildner Exp $
37  */
38 
39 #include <sys/types.h>
40 #include <sys/mman.h>
41 #include <sys/param.h>
42 #include <sys/stat.h>
43 
44 #include <err.h>
45 #include <errno.h>
46 #include <fcntl.h>
47 #include <libgen.h>
48 #include <limits.h>
49 #include <locale.h>
50 #include <regex.h>
51 #include <stddef.h>
52 #include <stdio.h>
53 #include <stdlib.h>
54 #include <string.h>
55 #include <unistd.h>
56 
57 #include "defs.h"
58 #include "extern.h"
59 
60 /*
61  * Linked list of units (strings and files) to be compiled
62  */
63 struct s_compunit {
64 	struct s_compunit *next;
65 	enum e_cut {CU_FILE, CU_STRING} type;
66 	char *s;			/* Pointer to string or fname */
67 };
68 
69 /*
70  * Linked list pointer to compilation units and pointer to current
71  * next pointer.
72  */
73 static struct s_compunit *script, **cu_nextp = &script;
74 
75 /*
76  * Linked list of files to be processed
77  */
78 struct s_flist {
79 	char *fname;
80 	struct s_flist *next;
81 };
82 
83 /*
84  * Linked list pointer to files and pointer to current
85  * next pointer.
86  */
87 static struct s_flist *files, **fl_nextp = &files;
88 
89 FILE *infile;			/* Current input file */
90 FILE *outfile;			/* Current output file */
91 
92 int aflag, eflag, nflag;
93 int rflags = 0;
94 static int rval;		/* Exit status */
95 
96 static int ispan;		/* Whether inplace editing spans across files */
97 
98 /*
99  * Current file and line number; line numbers restart across compilation
100  * units, but span across input files.  The latter is optional if editing
101  * in place.
102  */
103 const char *fname;		/* File name. */
104 const char *outfname;		/* Output file name */
105 static char oldfname[PATH_MAX];	/* Old file name (for in-place editing) */
106 static char tmpfname[PATH_MAX];	/* Temporary file name (for in-place editing) */
107 static const char *inplace;	/* Inplace edit file extension. */
108 u_long linenum;
109 
110 static void add_compunit(enum e_cut, char *);
111 static void add_file(char *);
112 static void usage(void);
113 
114 int
115 main(int argc, char *argv[])
116 {
117 	int c, fflag;
118 	char *temp_arg;
119 
120 	(void) setlocale(LC_ALL, "");
121 
122 	fflag = 0;
123 	inplace = NULL;
124 
125 	while ((c = getopt(argc, argv, "EI:ae:f:i:lnr")) != -1)
126 		switch (c) {
127 		case 'r':		/* GNU sed compat */
128 		case 'E':
129 			rflags = REG_EXTENDED;
130 			break;
131 		case 'I':
132 			inplace = optarg;
133 			ispan = 1;	/* span across input files */
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 			ispan = 0;	/* don't span across input files */
153 			break;
154 		case 'l':
155 			if(setlinebuf(stdout) != 0)
156 				warnx("setlinebuf() failed");
157 			break;
158 		case 'n':
159 			nflag = 1;
160 			break;
161 		default:
162 		case '?':
163 			usage();
164 		}
165 	argc -= optind;
166 	argv += optind;
167 
168 	/* First usage case; script is the first arg */
169 	if (!eflag && !fflag && *argv) {
170 		add_compunit(CU_STRING, *argv);
171 		argv++;
172 	}
173 
174 	compile();
175 
176 	/* Continue with first and start second usage */
177 	if (*argv)
178 		for (; *argv; argv++)
179 			add_file(*argv);
180 	else
181 		add_file(NULL);
182 	process();
183 	cfclose(prog, NULL);
184 	if (fclose(stdout))
185 		err(1, "stdout");
186 	exit(rval);
187 }
188 
189 static void
190 usage(void)
191 {
192 	(void)fprintf(stderr, "%s\n%s\n",
193 		"usage: sed script [-Ealn] [-i extension] [file ...]",
194 		"       sed [-Ealn] [-i extension] [-e script] ... [-f script_file] ... [file ...]");
195 	exit(1);
196 }
197 
198 /*
199  * Like fgets, but go through the chain of compilation units chaining them
200  * together.  Empty strings and files are ignored.
201  */
202 char *
203 cu_fgets(char *buf, int n, int *more)
204 {
205 	static enum {ST_EOF, ST_FILE, ST_STRING} state = ST_EOF;
206 	static FILE *f;		/* Current open file */
207 	static char *s;		/* Current pointer inside string */
208 	static char string_ident[30];
209 	char *p;
210 
211 again:
212 	switch (state) {
213 	case ST_EOF:
214 		if (script == NULL) {
215 			if (more != NULL)
216 				*more = 0;
217 			return (NULL);
218 		}
219 		linenum = 0;
220 		switch (script->type) {
221 		case CU_FILE:
222 			if ((f = fopen(script->s, "r")) == NULL)
223 				err(1, "%s", script->s);
224 			fname = script->s;
225 			state = ST_FILE;
226 			goto again;
227 		case CU_STRING:
228 			if (((size_t)snprintf(string_ident,
229 			    sizeof(string_ident), "\"%s\"", script->s)) >=
230 			    sizeof(string_ident) - 1)
231 				(void)strcpy(string_ident +
232 				    sizeof(string_ident) - 6, " ...\"");
233 			fname = string_ident;
234 			s = script->s;
235 			state = ST_STRING;
236 			goto again;
237 		}
238 	case ST_FILE:
239 		if ((p = fgets(buf, n, f)) != NULL) {
240 			linenum++;
241 			if (linenum == 1 && buf[0] == '#' && buf[1] == 'n')
242 				nflag = 1;
243 			if (more != NULL)
244 				*more = !feof(f);
245 			return (p);
246 		}
247 		script = script->next;
248 		(void)fclose(f);
249 		state = ST_EOF;
250 		goto again;
251 	case ST_STRING:
252 		if (linenum == 0 && s[0] == '#' && s[1] == 'n')
253 			nflag = 1;
254 		p = buf;
255 		for (;;) {
256 			if (n-- <= 1) {
257 				*p = '\0';
258 				linenum++;
259 				if (more != NULL)
260 					*more = 1;
261 				return (buf);
262 			}
263 			switch (*s) {
264 			case '\0':
265 				state = ST_EOF;
266 				if (s == script->s) {
267 					script = script->next;
268 					goto again;
269 				} else {
270 					script = script->next;
271 					*p = '\0';
272 					linenum++;
273 					if (more != NULL)
274 						*more = 0;
275 					return (buf);
276 				}
277 			case '\n':
278 				*p++ = '\n';
279 				*p = '\0';
280 				s++;
281 				linenum++;
282 				if (more != NULL)
283 					*more = 0;
284 				return (buf);
285 			default:
286 				*p++ = *s++;
287 			}
288 		}
289 	}
290 	/* NOTREACHED */
291 	return (NULL);
292 }
293 
294 /*
295  * Like fgets, but go through the list of files chaining them together.
296  * Set len to the length of the line.
297  */
298 int
299 mf_fgets(SPACE *sp, enum e_spflag spflag)
300 {
301 	struct stat sb;
302 	size_t len;
303 	char *p;
304 	int c;
305 	static int firstfile;
306 
307 	if (infile == NULL) {
308 		/* stdin? */
309 		if (files->fname == NULL) {
310 			if (inplace != NULL)
311 				errx(1, "-I or -i may not be used with stdin");
312 			infile = stdin;
313 			fname = "stdin";
314 			outfile = stdout;
315 			outfname = "stdout";
316 		}
317 		firstfile = 1;
318 	}
319 
320 	for (;;) {
321 		if (infile != NULL && (c = getc(infile)) != EOF) {
322 			(void)ungetc(c, infile);
323 			break;
324 		}
325 		/* If we are here then either eof or no files are open yet */
326 		if (infile == stdin) {
327 			sp->len = 0;
328 			return (0);
329 		}
330 		if (infile != NULL) {
331 			fclose(infile);
332 			if (*oldfname != '\0') {
333 				if (rename(fname, oldfname) != 0) {
334 					warn("rename()");
335 					unlink(tmpfname);
336 					exit(1);
337 				}
338 				*oldfname = '\0';
339 			}
340 			if (*tmpfname != '\0') {
341 				if (outfile != NULL && outfile != stdout)
342 					fclose(outfile);
343 				outfile = NULL;
344 				rename(tmpfname, fname);
345 				*tmpfname = '\0';
346 			}
347 			outfname = NULL;
348 		}
349 		if (firstfile == 0)
350 			files = files->next;
351 		else
352 			firstfile = 0;
353 		if (files == NULL) {
354 			sp->len = 0;
355 			return (0);
356 		}
357 		fname = files->fname;
358 		if (inplace != NULL) {
359 			if (lstat(fname, &sb) != 0)
360 				err(1, "%s", fname);
361 			if (!(sb.st_mode & S_IFREG))
362 				errx(1, "%s: %s %s", fname,
363 				    "in-place editing only",
364 				    "works for regular files");
365 			if (*inplace != '\0') {
366 				strlcpy(oldfname, fname,
367 				    sizeof(oldfname));
368 				len = strlcat(oldfname, inplace,
369 				    sizeof(oldfname));
370 				if (len > sizeof(oldfname))
371 					errx(1, "%s: name too long", fname);
372 			}
373 			len = snprintf(tmpfname, sizeof(tmpfname),
374 			    "%s/.!%ld!%s", dirname(fname), (long)getpid(),
375 			    basename(fname));
376 			if (len >= sizeof(tmpfname))
377 				errx(1, "%s: name too long", fname);
378 			unlink(tmpfname);
379 			if ((outfile = fopen(tmpfname, "w")) == NULL)
380 				err(1, "%s", fname);
381 			fchown(fileno(outfile), sb.st_uid, sb.st_gid);
382 			fchmod(fileno(outfile), sb.st_mode & ALLPERMS);
383 			outfname = tmpfname;
384 			if (!ispan) {
385 				linenum = 0;
386 				resetstate();
387 			}
388 		} else {
389 			outfile = stdout;
390 			outfname = "stdout";
391 		}
392 		if ((infile = fopen(fname, "r")) == NULL) {
393 			warn("%s", fname);
394 			rval = 1;
395 			continue;
396 		}
397 	}
398 	/*
399 	 * We are here only when infile is open and we still have something
400 	 * to read from it.
401 	 *
402 	 * Use fgetln so that we can handle essentially infinite input data.
403 	 * Can't use the pointer into the stdio buffer as the process space
404 	 * because the ungetc() can cause it to move.
405 	 */
406 	p = fgetln(infile, &len);
407 	if (ferror(infile))
408 		errx(1, "%s: %s", fname, strerror(errno ? errno : EIO));
409 	if (len != 0 && p[len - 1] == '\n')
410 		len--;
411 	cspace(sp, p, len, spflag);
412 
413 	linenum++;
414 
415 	return (1);
416 }
417 
418 /*
419  * Add a compilation unit to the linked list
420  */
421 static void
422 add_compunit(enum e_cut type, char *s)
423 {
424 	struct s_compunit *cu;
425 
426 	if ((cu = malloc(sizeof(struct s_compunit))) == NULL)
427 		err(1, "malloc");
428 	cu->type = type;
429 	cu->s = s;
430 	cu->next = NULL;
431 	*cu_nextp = cu;
432 	cu_nextp = &cu->next;
433 }
434 
435 /*
436  * Add a file to the linked list
437  */
438 static void
439 add_file(char *s)
440 {
441 	struct s_flist *fp;
442 
443 	if ((fp = malloc(sizeof(struct s_flist))) == NULL)
444 		err(1, "malloc");
445 	fp->next = NULL;
446 	*fl_nextp = fp;
447 	fp->fname = s;
448 	fl_nextp = &fp->next;
449 }
450 
451 int
452 lastline(void)
453 {
454 	int ch;
455 
456 	if (files->next != NULL && (inplace == NULL || ispan))
457 		return (0);
458 	if ((ch = getc(infile)) == EOF)
459 		return (1);
460 	ungetc(ch, infile);
461 	return (0);
462 }
463