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. All advertising materials mentioning features or use of this software
meillo@14:  *    must display the following acknowledgement:
meillo@14:  *	This product includes software developed by the University of
meillo@14:  *	California, Berkeley and its contributors.
meillo@14:  * 4. 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: #ifndef lint
meillo@14: static char copyright[] =
meillo@14: "@(#) Copyright (c) 1989, 1993\n\
meillo@14: 	The Regents of the University of California.  All rights reserved.\n";
meillo@14: #endif /* not lint */
meillo@14: 
meillo@14: #ifndef lint
meillo@14: static char sccsid[] = "@(#)cut.c	8.1 (Berkeley) 6/6/93";
meillo@14: #endif /* not lint */
meillo@14: 
meillo@14: #include <ctype.h>
meillo@14: #include <errno.h>
meillo@14: #include <limits.h>
meillo@14: #include <stdio.h>
meillo@14: #include <stdlib.h>
meillo@14: #include <string.h>
meillo@14: 
meillo@14: int	cflag;
meillo@14: char	dchar;
meillo@14: int	dflag;
meillo@14: int	fflag;
meillo@14: int	sflag;
meillo@14: 
meillo@14: void	c_cut __P((FILE *, char *));
meillo@14: void	err __P((const char *, ...));
meillo@14: void	f_cut __P((FILE *, char *));
meillo@14: void	get_list __P((char *));
meillo@14: void	usage __P((void));
meillo@14: 
meillo@14: int
meillo@14: main(argc, argv)
meillo@14: 	int argc;
meillo@14: 	char *argv[];
meillo@14: {
meillo@14: 	FILE *fp;
meillo@14: 	void (*fcn) __P((FILE *, char *));
meillo@14: 	int ch;
meillo@14: 
meillo@14: 	dchar = '\t';			/* default delimiter is \t */
meillo@14: 
meillo@14: 	while ((ch = getopt(argc, argv, "c:d:f:s")) != EOF)
meillo@14: 		switch(ch) {
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 '?':
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)
meillo@14: 			usage();
meillo@14: 	} else if (!cflag || dflag || sflag)
meillo@14: 		usage();
meillo@14: 
meillo@14: 	if (*argv)
meillo@14: 		for (; *argv; ++argv) {
meillo@14: 			if (!(fp = fopen(*argv, "r")))
meillo@14: 				err("%s: %s\n", *argv, strerror(errno));
meillo@14: 			fcn(fp, *argv);
meillo@14: 			(void)fclose(fp);
meillo@14: 		}
meillo@14: 	else
meillo@14: 		fcn(stdin, "stdin");
meillo@14: 	exit(0);
meillo@14: }
meillo@14: 
meillo@14: int autostart, autostop, maxval;
meillo@14: 
meillo@14: char positions[_POSIX2_LINE_MAX + 1];
meillo@14: 
meillo@14: void
meillo@14: get_list(list)
meillo@14: 	char *list;
meillo@14: {
meillo@14: 	register int setautostart, start, stop;
meillo@14: 	register char *pos;
meillo@14: 	char *p;
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 too.
meillo@14: 	 */
meillo@14: 	for (; p = strtok(list, ", \t"); 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(*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(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: 			err("[-cf] list: illegal list value\n");
meillo@14: 		if (!stop || !start)
meillo@14: 			err("[-cf] list: values may not include zero\n");
meillo@14: 		if (stop > _POSIX2_LINE_MAX)
meillo@14: 			err("[-cf] list: %d too large (max %d)\n",
meillo@14: 			    stop, _POSIX2_LINE_MAX);
meillo@14: 		if (maxval < stop)
meillo@14: 			maxval = stop;
meillo@14: 		for (pos = positions + start; start++ <= stop; *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: 		memset(positions + 1, '1', autostart);
meillo@14: }
meillo@14: 
meillo@14: /* ARGSUSED */
meillo@14: void
meillo@14: c_cut(fp, fname)
meillo@14: 	FILE *fp;
meillo@14: 	char *fname;
meillo@14: {
meillo@14: 	register int ch, col;
meillo@14: 	register char *pos;
meillo@14: 
meillo@14: 	for (;;) {
meillo@14: 		pos = positions + 1;
meillo@14: 		for (col = maxval; col; --col) {
meillo@14: 			if ((ch = getc(fp)) == EOF)
meillo@14: 				return;
meillo@14: 			if (ch == '\n')
meillo@14: 				break;
meillo@14: 			if (*pos++)
meillo@14: 				(void)putchar(ch);
meillo@14: 		}
meillo@14: 		if (ch != '\n')
meillo@14: 			if (autostop)
meillo@14: 				while ((ch = getc(fp)) != EOF && ch != '\n')
meillo@14: 					(void)putchar(ch);
meillo@14: 			else
meillo@14: 				while ((ch = getc(fp)) != EOF && ch != '\n');
meillo@14: 		(void)putchar('\n');
meillo@14: 	}
meillo@14: }
meillo@14: 
meillo@14: void
meillo@14: f_cut(fp, fname)
meillo@14: 	FILE *fp;
meillo@14: 	char *fname;
meillo@14: {
meillo@14: 	register int ch, field, isdelim;
meillo@14: 	register char *pos, *p, sep;
meillo@14: 	int output;
meillo@14: 	char lbuf[_POSIX2_LINE_MAX + 1];
meillo@14: 
meillo@14: 	for (sep = dchar, output = 0; fgets(lbuf, sizeof(lbuf), fp);) {
meillo@14: 		for (isdelim = 0, p = lbuf;; ++p) {
meillo@14: 			if (!(ch = *p))
meillo@14: 				err("%s: line too long.\n", fname);
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)printf("%s", lbuf);
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: 			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: 		(void)putchar('\n');
meillo@14: 	}
meillo@14: }
meillo@14: 
meillo@14: void
meillo@14: usage()
meillo@14: {
meillo@14: 	(void)fprintf(stderr,
meillo@14: "usage:\tcut -c list [file1 ...]\n\tcut -f list [-s] [-d delim] [file ...]\n");
meillo@14: 	exit(1);
meillo@14: }
meillo@14: 
meillo@14: #if __STDC__
meillo@14: #include <stdarg.h>
meillo@14: #else
meillo@14: #include <varargs.h>
meillo@14: #endif
meillo@14: 
meillo@14: void
meillo@14: #if __STDC__
meillo@14: err(const char *fmt, ...)
meillo@14: #else
meillo@14: err(fmt, va_alist)
meillo@14: 	char *fmt;
meillo@14:         va_dcl
meillo@14: #endif
meillo@14: {
meillo@14: 	va_list ap;
meillo@14: #if __STDC__
meillo@14: 	va_start(ap, fmt);
meillo@14: #else
meillo@14: 	va_start(ap);
meillo@14: #endif
meillo@14: 	(void)fprintf(stderr, "cut: ");
meillo@14: 	(void)vfprintf(stderr, fmt, ap);
meillo@14: 	va_end(ap);
meillo@14: 	(void)fprintf(stderr, "\n");
meillo@14: 	exit(1);
meillo@14: 	/* NOTREACHED */
meillo@14: }