meillo@14: /*	$NetBSD: cut.c,v 1.29 2014/02/03 20:22:19 wiz Exp $	*/
meillo@14: 
meillo@14: /*
meillo@14:  * Copyright (c) 1989, 1993
meillo@14:  *	The Regents of the University of California.  All rights reserved.
meillo@14:  *
meillo@14:  * This code is derived from software contributed to Berkeley by
meillo@14:  * Adam S. Moskowitz of Menlo Consulting and Marciano Pitargue.
meillo@14:  *
meillo@14:  * Redistribution and use in source and binary forms, with or without
meillo@14:  * modification, are permitted provided that the following conditions
meillo@14:  * are met:
meillo@14:  * 1. Redistributions of source code must retain the above copyright
meillo@14:  *    notice, this list of conditions and the following disclaimer.
meillo@14:  * 2. Redistributions in binary form must reproduce the above copyright
meillo@14:  *    notice, this list of conditions and the following disclaimer in the
meillo@14:  *    documentation and/or other materials provided with the distribution.
meillo@14:  * 3. Neither the name of the University nor the names of its contributors
meillo@14:  *    may be used to endorse or promote products derived from this software
meillo@14:  *    without specific prior written permission.
meillo@14:  *
meillo@14:  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
meillo@14:  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
meillo@14:  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
meillo@14:  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
meillo@14:  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
meillo@14:  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
meillo@14:  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
meillo@14:  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
meillo@14:  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
meillo@14:  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
meillo@14:  * SUCH DAMAGE.
meillo@14:  */
meillo@14: 
meillo@14: #include <sys/cdefs.h>
meillo@14: #ifndef lint
meillo@14: __COPYRIGHT("@(#) Copyright (c) 1989, 1993\
meillo@14:  The Regents of the University of California.  All rights reserved.");
meillo@14: #endif /* not lint */
meillo@14: 
meillo@14: #ifndef lint
meillo@14: #if 0
meillo@14: static char sccsid[] = "@(#)cut.c	8.3 (Berkeley) 5/4/95";
meillo@14: #endif
meillo@14: __RCSID("$NetBSD: cut.c,v 1.29 2014/02/03 20:22:19 wiz Exp $");
meillo@14: #endif /* not lint */
meillo@14: 
meillo@14: #include <ctype.h>
meillo@14: #include <err.h>
meillo@14: #include <errno.h>
meillo@14: #include <limits.h>
meillo@14: #include <locale.h>
meillo@14: #include <stdio.h>
meillo@14: #include <stdlib.h>
meillo@14: #include <string.h>
meillo@14: #include <unistd.h>
meillo@14: #include <util.h>
meillo@14: #include <wchar.h>
meillo@14: #include <sys/param.h>
meillo@14: 
meillo@14: static int bflag;
meillo@14: static int	cflag;
meillo@14: static char	dchar;
meillo@14: static int	dflag;
meillo@14: static int	fflag;
meillo@14: static int	sflag;
meillo@14: 
meillo@14: static void	b_cut(FILE *, const char *);
meillo@14: static void	c_cut(FILE *, const char *);
meillo@14: static void	f_cut(FILE *, const char *);
meillo@14: static void	get_list(char *);
meillo@14: static void	usage(void) __dead;
meillo@14: 
meillo@14: int
meillo@14: main(int argc, char *argv[])
meillo@14: {
meillo@14: 	FILE *fp;
meillo@14: 	void (*fcn)(FILE *, const char *);
meillo@14: 	int ch, rval;
meillo@14: 
meillo@14: 	fcn = NULL;
meillo@14: 	(void)setlocale(LC_ALL, "");
meillo@14: 
meillo@14: 	dchar = '\t';			/* default delimiter is \t */
meillo@14: 
meillo@14: 	/* Since we don't support multi-byte characters, the -c and -b
meillo@14: 	   options are equivalent, and the -n option is meaningless. */
meillo@14: 	while ((ch = getopt(argc, argv, "b:c:d:f:sn")) != -1)
meillo@14: 		switch(ch) {
meillo@14: 		case 'b':
meillo@14: 			fcn = b_cut;
meillo@14: 			get_list(optarg);
meillo@14: 			bflag = 1;
meillo@14: 			break;
meillo@14: 		case 'c':
meillo@14: 			fcn = c_cut;
meillo@14: 			get_list(optarg);
meillo@14: 			cflag = 1;
meillo@14: 			break;
meillo@14: 		case 'd':
meillo@14: 			dchar = *optarg;
meillo@14: 			dflag = 1;
meillo@14: 			break;
meillo@14: 		case 'f':
meillo@14: 			get_list(optarg);
meillo@14: 			fcn = f_cut;
meillo@14: 			fflag = 1;
meillo@14: 			break;
meillo@14: 		case 's':
meillo@14: 			sflag = 1;
meillo@14: 			break;
meillo@14: 		case 'n':
meillo@14: 			break;
meillo@14: 		case '?':
meillo@14: 		default:
meillo@14: 			usage();
meillo@14: 		}
meillo@14: 	argc -= optind;
meillo@14: 	argv += optind;
meillo@14: 
meillo@14: 	if (fflag) {
meillo@14: 		if (cflag || bflag)
meillo@14: 			usage();
meillo@14: 	} else if ((!cflag && !bflag) || dflag || sflag)
meillo@14: 		usage();
meillo@14: 	else if (bflag && cflag)
meillo@14: 		usage();
meillo@14: 
meillo@14: 	rval = 0;
meillo@14: 	if (*argv)
meillo@14: 		for (; *argv; ++argv) {
meillo@14: 			if (strcmp(*argv, "-") == 0)
meillo@14: 				fcn(stdin, "stdin");
meillo@14: 			else {
meillo@14: 				if ((fp = fopen(*argv, "r"))) {
meillo@14: 					fcn(fp, *argv);
meillo@14: 					(void)fclose(fp);
meillo@14: 				} else {
meillo@14: 					rval = 1;
meillo@14: 					warn("%s", *argv);
meillo@14: 				}
meillo@14: 			}
meillo@14: 		}
meillo@14: 	else
meillo@14: 		fcn(stdin, "stdin");
meillo@14: 	return(rval);
meillo@14: }
meillo@14: 
meillo@14: static size_t autostart, autostop, maxval;
meillo@14: 
meillo@14: static char *positions = NULL;
meillo@14: static size_t numpositions = 0;
meillo@14: #define ALLOC_CHUNK	_POSIX2_LINE_MAX	/* malloc granularity */
meillo@14: 
meillo@14: static void
meillo@14: get_list(char *list)
meillo@14: {
meillo@14: 	size_t setautostart, start, stop;
meillo@14: 	char *pos;
meillo@14: 	char *p;
meillo@14: 
meillo@14: 	if (positions == NULL) {
meillo@14: 		numpositions = ALLOC_CHUNK;
meillo@14: 		positions = ecalloc(numpositions, sizeof(*positions));
meillo@14: 	}
meillo@14: 
meillo@14: 	/*
meillo@14: 	 * set a byte in the positions array to indicate if a field or
meillo@14: 	 * column is to be selected; use +1, it's 1-based, not 0-based.
meillo@14: 	 * This parser is less restrictive than the Draft 9 POSIX spec.
meillo@14: 	 * POSIX doesn't allow lists that aren't in increasing order or
meillo@14: 	 * overlapping lists.  We also handle "-3-5" although there's no
meillo@14: 	 * real reason to.
meillo@14: 	 */
meillo@14: 	for (; (p = strtok(list, ", \t")) != NULL; list = NULL) {
meillo@14: 		setautostart = start = stop = 0;
meillo@14: 		if (*p == '-') {
meillo@14: 			++p;
meillo@14: 			setautostart = 1;
meillo@14: 		}
meillo@14: 		if (isdigit((unsigned char)*p)) {
meillo@14: 			start = stop = strtol(p, &p, 10);
meillo@14: 			if (setautostart && start > autostart)
meillo@14: 				autostart = start;
meillo@14: 		}
meillo@14: 		if (*p == '-') {
meillo@14: 			if (isdigit((unsigned char)p[1]))
meillo@14: 				stop = strtol(p + 1, &p, 10);
meillo@14: 			if (*p == '-') {
meillo@14: 				++p;
meillo@14: 				if (!autostop || autostop > stop)
meillo@14: 					autostop = stop;
meillo@14: 			}
meillo@14: 		}
meillo@14: 		if (*p)
meillo@14: 			errx(1, "[-bcf] list: illegal list value");
meillo@14: 		if (!stop || !start)
meillo@14: 			errx(1, "[-bcf] list: values may not include zero");
meillo@14: 		if (stop + 1 > numpositions) {
meillo@14: 			size_t newsize;
meillo@14: 			newsize = roundup(stop + 1, ALLOC_CHUNK);
meillo@14: 			positions = erealloc(positions, newsize);
meillo@14: 			(void)memset(positions + numpositions, 0,
meillo@14: 			    newsize - numpositions);
meillo@14: 			numpositions = newsize;
meillo@14: 		}
meillo@14: 		if (maxval < stop)
meillo@14: 			maxval = stop;
meillo@14: 		for (pos = positions + start; start++ <= stop; pos++)
meillo@14: 			*pos = 1;
meillo@14: 	}
meillo@14: 
meillo@14: 	/* overlapping ranges */
meillo@14: 	if (autostop && maxval > autostop)
meillo@14: 		maxval = autostop;
meillo@14: 
meillo@14: 	/* set autostart */
meillo@14: 	if (autostart)
meillo@14: 		(void)memset(positions + 1, '1', autostart);
meillo@14: }
meillo@14: 
meillo@14: static void
meillo@14: /*ARGSUSED*/
meillo@14: f_cut(FILE *fp, const char *fname __unused)
meillo@14: {
meillo@14: 	int ch, field, isdelim;
meillo@14: 	char *pos, *p, sep;
meillo@14: 	int output;
meillo@14: 	size_t len;
meillo@14: 	char *lbuf, *tbuf;
meillo@14: 
meillo@14: 	for (sep = dchar, tbuf = NULL; (lbuf = fgetln(fp, &len)) != NULL;) {
meillo@14: 		output = 0;
meillo@14: 		if (lbuf[len - 1] != '\n') {
meillo@14: 			/* no newline at the end of the last line so add one */
meillo@14: 			if ((tbuf = (char *)malloc(len + 1)) == NULL)
meillo@14: 				err(1, NULL);
meillo@14: 			(void)memcpy(tbuf, lbuf, len);
meillo@14: 			tbuf[len++] = '\n';
meillo@14: 			lbuf = tbuf;
meillo@14: 		}
meillo@14: 		for (isdelim = 0, p = lbuf;; ++p) {
meillo@14: 			ch = *p;
meillo@14: 			/* this should work if newline is delimiter */
meillo@14: 			if (ch == sep)
meillo@14: 				isdelim = 1;
meillo@14: 			if (ch == '\n') {
meillo@14: 				if (!isdelim && !sflag)
meillo@14: 					(void)fwrite(lbuf, len, 1, stdout);
meillo@14: 				break;
meillo@14: 			}
meillo@14: 		}
meillo@14: 		if (!isdelim)
meillo@14: 			continue;
meillo@14: 
meillo@14: 		pos = positions + 1;
meillo@14: 		for (field = maxval, p = lbuf; field; --field, ++pos) {
meillo@14: 			if (*pos) {
meillo@14: 				if (output++)
meillo@14: 					(void)putchar(sep);
meillo@14: 				while ((ch = *p++) != '\n' && ch != sep)
meillo@14: 					(void)putchar(ch);
meillo@14: 			} else {
meillo@14: 				while ((ch = *p++) != '\n' && ch != sep)
meillo@14: 					continue;
meillo@14: 			}
meillo@14: 			if (ch == '\n')
meillo@14: 				break;
meillo@14: 		}
meillo@14: 		if (ch != '\n') {
meillo@14: 			if (autostop) {
meillo@14: 				if (output)
meillo@14: 					(void)putchar(sep);
meillo@14: 				for (; (ch = *p) != '\n'; ++p)
meillo@14: 					(void)putchar(ch);
meillo@14: 			} else
meillo@14: 				for (; (ch = *p) != '\n'; ++p);
meillo@14: 		}
meillo@14: 		(void)putchar('\n');
meillo@14: 		if (tbuf) {
meillo@14: 			free(tbuf);
meillo@14: 			tbuf = NULL;
meillo@14: 		}
meillo@14: 	}
meillo@14: 	if (tbuf)
meillo@14: 		free(tbuf);
meillo@14: }
meillo@14: 
meillo@14: static void
meillo@14: usage(void)
meillo@14: {
meillo@14: 	(void)fprintf(stderr, "usage:\tcut -b list [-n] [file ...]\n"
meillo@14: 	    "\tcut -c list [file ...]\n"
meillo@14: 	    "\tcut -f list [-d string] [-s] [file ...]\n");
meillo@14: 	exit(1);
meillo@14: }
meillo@14: 
meillo@14: /* make b_put(): */
meillo@14: #define CUT_BYTE 1
meillo@14: #include "x_cut.c"
meillo@14: #undef CUT_BYTE
meillo@14: 
meillo@14: /* make c_put(): */
meillo@14: #define CUT_BYTE 0
meillo@14: #include "x_cut.c"
meillo@14: #undef CUT_BYTE