Home | History | Annotate | Line # | Download | only in sed
main.c revision 1.23
      1 /*	$NetBSD: main.c,v 1.23 2014/06/06 01:45:11 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 1.23 2014/06/06 01:45:11 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 #ifndef lint
     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);
    131 
    132 int
    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::lnr")) != -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 		default:
    185 		case '?':
    186 			usage();
    187 		}
    188 	argc -= optind;
    189 	argv += optind;
    190 
    191 	/* First usage case; script is the first arg */
    192 	if (!eflag && !fflag && *argv) {
    193 		add_compunit(CU_STRING, *argv);
    194 		argv++;
    195 	}
    196 
    197 	compile();
    198 
    199 	/* Continue with first and start second usage */
    200 	if (*argv)
    201 		for (; *argv; argv++)
    202 			add_file(*argv);
    203 	else
    204 		add_file(NULL);
    205 	process();
    206 	cfclose(prog, NULL);
    207 	if (fclose(stdout))
    208 		err(1, "stdout");
    209 	exit(rval);
    210 }
    211 
    212 static void
    213 usage(void)
    214 {
    215 	(void)fprintf(stderr, "%s\n%s\n",
    216 		"usage: sed script [-Ealn] [-i extension] [file ...]",
    217 		"       sed [-Ealn] [-i extension] [-e script] ... [-f script_file] ... [file ...]");
    218 	exit(1);
    219 }
    220 
    221 /*
    222  * Like fgets, but go through the chain of compilation units chaining them
    223  * together.  Empty strings and files are ignored.
    224  */
    225 char *
    226 cu_fgets(char *buf, int n, int *more)
    227 {
    228 	static enum {ST_EOF, ST_FILE, ST_STRING} state = ST_EOF;
    229 	static FILE *f;		/* Current open file */
    230 	static char *s;		/* Current pointer inside string */
    231 	static char string_ident[30];
    232 	char *p;
    233 
    234 again:
    235 	switch (state) {
    236 	case ST_EOF:
    237 		if (script == NULL) {
    238 			if (more != NULL)
    239 				*more = 0;
    240 			return (NULL);
    241 		}
    242 		linenum = 0;
    243 		switch (script->type) {
    244 		case CU_FILE:
    245 			if ((f = fopen(script->s, "r")) == NULL)
    246 				err(1, "%s", script->s);
    247 			fname = script->s;
    248 			state = ST_FILE;
    249 			goto again;
    250 		case CU_STRING:
    251 			if (((size_t)snprintf(string_ident,
    252 			    sizeof(string_ident), "\"%s\"", script->s)) >=
    253 			    sizeof(string_ident) - 1)
    254 				(void)strcpy(string_ident +
    255 				    sizeof(string_ident) - 6, " ...\"");
    256 			fname = string_ident;
    257 			s = script->s;
    258 			state = ST_STRING;
    259 			goto again;
    260 		}
    261 	case ST_FILE:
    262 		if ((p = fgets(buf, n, f)) != NULL) {
    263 			linenum++;
    264 			if (linenum == 1 && buf[0] == '#' && buf[1] == 'n')
    265 				nflag = 1;
    266 			if (more != NULL)
    267 				*more = !feof(f);
    268 			return (p);
    269 		}
    270 		script = script->next;
    271 		(void)fclose(f);
    272 		state = ST_EOF;
    273 		goto again;
    274 	case ST_STRING:
    275 		if (linenum == 0 && s[0] == '#' && s[1] == 'n')
    276 			nflag = 1;
    277 		p = buf;
    278 		for (;;) {
    279 			if (n-- <= 1) {
    280 				*p = '\0';
    281 				linenum++;
    282 				if (more != NULL)
    283 					*more = 1;
    284 				return (buf);
    285 			}
    286 			switch (*s) {
    287 			case '\0':
    288 				state = ST_EOF;
    289 				if (s == script->s) {
    290 					script = script->next;
    291 					goto again;
    292 				} else {
    293 					script = script->next;
    294 					*p = '\0';
    295 					linenum++;
    296 					if (more != NULL)
    297 						*more = 0;
    298 					return (buf);
    299 				}
    300 			case '\n':
    301 				*p++ = '\n';
    302 				*p = '\0';
    303 				s++;
    304 				linenum++;
    305 				if (more != NULL)
    306 					*more = 0;
    307 				return (buf);
    308 			default:
    309 				*p++ = *s++;
    310 			}
    311 		}
    312 	}
    313 	/* NOTREACHED */
    314 	return (NULL);
    315 }
    316 
    317 /*
    318  * Like fgets, but go through the list of files chaining them together.
    319  * Set len to the length of the line.
    320  */
    321 int
    322 mf_fgets(SPACE *sp, enum e_spflag spflag)
    323 {
    324 	struct stat sb;
    325 	size_t len;
    326 	static char *p = NULL;
    327 	static size_t plen = 0;
    328 	int c;
    329 	static int firstfile;
    330 
    331 	if (infile == NULL) {
    332 		/* stdin? */
    333 		if (files->fname == NULL) {
    334 			if (inplace != NULL)
    335 				errx(1, "-I or -i may not be used with stdin");
    336 			infile = stdin;
    337 			fname = "stdin";
    338 			outfile = stdout;
    339 			outfname = "stdout";
    340 		}
    341 		firstfile = 1;
    342 	}
    343 
    344 	for (;;) {
    345 		if (infile != NULL && (c = getc(infile)) != EOF) {
    346 			(void)ungetc(c, infile);
    347 			break;
    348 		}
    349 		/* If we are here then either eof or no files are open yet */
    350 		if (infile == stdin) {
    351 			sp->len = 0;
    352 			return (0);
    353 		}
    354 		if (infile != NULL) {
    355 			fclose(infile);
    356 			if (*oldfname != '\0') {
    357 				/* if there was a backup file, remove it */
    358 				unlink(oldfname);
    359 				/*
    360 				 * Backup the original.  Note that hard links
    361 				 * are not supported on all filesystems.
    362 				 */
    363 				if ((link(fname, oldfname) != 0) &&
    364 				   (rename(fname, oldfname) != 0)) {
    365 					warn("rename()");
    366 					if (*tmpfname)
    367 						unlink(tmpfname);
    368 					exit(1);
    369 				}
    370 				*oldfname = '\0';
    371 			}
    372 			if (*tmpfname != '\0') {
    373 				if (outfile != NULL && outfile != stdout)
    374 					if (fclose(outfile) != 0) {
    375 						warn("fclose()");
    376 						unlink(tmpfname);
    377 						exit(1);
    378 					}
    379 				outfile = NULL;
    380 				if (rename(tmpfname, fname) != 0) {
    381 					/* this should not happen really! */
    382 					warn("rename()");
    383 					unlink(tmpfname);
    384 					exit(1);
    385 				}
    386 				*tmpfname = '\0';
    387 			}
    388 			outfname = NULL;
    389 		}
    390 		if (firstfile == 0)
    391 			files = files->next;
    392 		else
    393 			firstfile = 0;
    394 		if (files == NULL) {
    395 			sp->len = 0;
    396 			return (0);
    397 		}
    398 		fname = files->fname;
    399 		if (inplace != NULL) {
    400 			if (lstat(fname, &sb) != 0)
    401 				err(1, "%s", fname);
    402 			if (!(sb.st_mode & S_IFREG))
    403 				errx(1, "%s: %s %s", fname,
    404 				    "in-place editing only",
    405 				    "works for regular files");
    406 			if (*inplace != '\0') {
    407 				strlcpy(oldfname, fname,
    408 				    sizeof(oldfname));
    409 				len = strlcat(oldfname, inplace,
    410 				    sizeof(oldfname));
    411 				if (len > sizeof(oldfname))
    412 					errx(1, "%s: name too long", fname);
    413 			}
    414 			char d_name[PATH_MAX], f_name[PATH_MAX];
    415 			(void)strlcpy(d_name, fname, sizeof(d_name));
    416 			(void)strlcpy(f_name, fname, sizeof(f_name));
    417 			len = (size_t)snprintf(tmpfname, sizeof(tmpfname),
    418 			    "%s/.!%ld!%s", dirname(d_name), (long)getpid(),
    419 			    basename(f_name));
    420 			if (len >= sizeof(tmpfname))
    421 				errx(1, "%s: name too long", fname);
    422 			unlink(tmpfname);
    423 			if ((outfile = fopen(tmpfname, "w")) == NULL)
    424 				err(1, "%s", fname);
    425 			fchown(fileno(outfile), sb.st_uid, sb.st_gid);
    426 			fchmod(fileno(outfile), sb.st_mode & ALLPERMS);
    427 			outfname = tmpfname;
    428 			if (!ispan) {
    429 				linenum = 0;
    430 				resetstate();
    431 			}
    432 		} else {
    433 			outfile = stdout;
    434 			outfname = "stdout";
    435 		}
    436 		if ((infile = fopen(fname, "r")) == NULL) {
    437 			warn("%s", fname);
    438 			rval = 1;
    439 			continue;
    440 		}
    441 	}
    442 	/*
    443 	 * We are here only when infile is open and we still have something
    444 	 * to read from it.
    445 	 *
    446 	 * Use getline() so that we can handle essentially infinite input
    447 	 * data.  The p and plen are static so each invocation gives
    448 	 * getline() the same buffer which is expanded as needed.
    449 	 */
    450 	ssize_t slen = getline(&p, &plen, infile);
    451 	if (slen == -1)
    452 		err(1, "%s", fname);
    453 	if (slen != 0 && p[slen - 1] == '\n')
    454 		slen--;
    455 	cspace(sp, p, (size_t)slen, spflag);
    456 
    457 	linenum++;
    458 
    459 	return (1);
    460 }
    461 
    462 /*
    463  * Add a compilation unit to the linked list
    464  */
    465 static void
    466 add_compunit(enum e_cut type, char *s)
    467 {
    468 	struct s_compunit *cu;
    469 
    470 	cu = xmalloc(sizeof(struct s_compunit));
    471 	cu->type = type;
    472 	cu->s = s;
    473 	cu->next = NULL;
    474 	*cu_nextp = cu;
    475 	cu_nextp = &cu->next;
    476 }
    477 
    478 /*
    479  * Add a file to the linked list
    480  */
    481 static void
    482 add_file(char *s)
    483 {
    484 	struct s_flist *fp;
    485 
    486 	fp = xmalloc(sizeof(struct s_flist));
    487 	fp->next = NULL;
    488 	*fl_nextp = fp;
    489 	fp->fname = s;
    490 	fl_nextp = &fp->next;
    491 }
    492 
    493 int
    494 lastline(void)
    495 {
    496 	int ch;
    497 
    498 	if (files->next != NULL && (inplace == NULL || ispan))
    499 		return (0);
    500 	if ((ch = getc(infile)) == EOF)
    501 		return (1);
    502 	ungetc(ch, infile);
    503 	return (0);
    504 }
    505