Home | History | Annotate | Line # | Download | only in wc
wc.c revision 1.16 
      1 /*	$NetBSD: wc.c,v 1.16 1999/02/14 18:03:18 mjacob Exp $	*/
      2 
      3 /*
      4  * Copyright (c) 1980, 1987, 1991, 1993
      5  *	The Regents of the University of California.  All rights reserved.
      6  *
      7  * Redistribution and use in source and binary forms, with or without
      8  * modification, are permitted provided that the following conditions
      9  * are met:
     10  * 1. Redistributions of source code must retain the above copyright
     11  *    notice, this list of conditions and the following disclaimer.
     12  * 2. Redistributions in binary form must reproduce the above copyright
     13  *    notice, this list of conditions and the following disclaimer in the
     14  *    documentation and/or other materials provided with the distribution.
     15  * 3. All advertising materials mentioning features or use of this software
     16  *    must display the following acknowledgement:
     17  *	This product includes software developed by the University of
     18  *	California, Berkeley and its contributors.
     19  * 4. Neither the name of the University nor the names of its contributors
     20  *    may be used to endorse or promote products derived from this software
     21  *    without specific prior written permission.
     22  *
     23  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
     24  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
     25  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
     26  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
     27  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
     28  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
     29  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
     30  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
     31  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
     32  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
     33  * SUCH DAMAGE.
     34  */
     35 
     36 #include <sys/cdefs.h>
     37 #ifndef lint
     38 __COPYRIGHT("@(#) Copyright (c) 1980, 1987, 1991, 1993\n\
     39 	The Regents of the University of California.  All rights reserved.\n");
     40 #endif /* not lint */
     41 
     42 #ifndef lint
     43 #if 0
     44 static char sccsid[] = "@(#)wc.c	8.2 (Berkeley) 5/2/95";
     45 #else
     46 __RCSID("$NetBSD: wc.c,v 1.16 1999/02/14 18:03:18 mjacob Exp $");
     47 #endif
     48 #endif /* not lint */
     49 
     50 /* wc line, word and char count */
     51 
     52 #include <sys/param.h>
     53 #include <sys/stat.h>
     54 #include <sys/types.h>
     55 
     56 #include <fcntl.h>
     57 #include <unistd.h>
     58 #include <errno.h>
     59 #include <stdio.h>
     60 
     61 #include <stdlib.h>
     62 #include <string.h>
     63 #include <locale.h>
     64 #include <ctype.h>
     65 #include <errno.h>
     66 #include <sys/param.h>
     67 #include <sys/stat.h>
     68 #include <sys/file.h>
     69 #include <unistd.h>
     70 #include <err.h>
     71 
     72 static u_int64_t	tlinect, twordct, tcharct;
     73 static int		doline, doword, dochar;
     74 static int 		rval = 0;
     75 static char *qfmt;
     76 
     77 static void	cnt __P((char *));
     78 static void	print_counts __P((u_int64_t, u_int64_t, u_int64_t, char *));
     79 static void	usage __P((void));
     80 int	main __P((int, char *[]));
     81 
     82 int
     83 main(argc, argv)
     84 	int argc;
     85 	char *argv[];
     86 {
     87 	int ch;
     88 
     89 	setlocale(LC_ALL, "");
     90 	if (sizeof (u_int64_t) > 4)
     91 		qfmt = " %7u";
     92 	else
     93 		qfmt = " %7qu";
     94 
     95 	while ((ch = getopt(argc, argv, "lwcm")) != -1)
     96 		switch((char)ch) {
     97 		case 'l':
     98 			doline = 1;
     99 			break;
    100 		case 'w':
    101 			doword = 1;
    102 			break;
    103 		case 'c':
    104 		case 'm':
    105 			dochar = 1;
    106 			break;
    107 		case '?':
    108 		default:
    109 			usage();
    110 		}
    111 	argv += optind;
    112 	argc -= optind;
    113 
    114 	/* Wc's flags are on by default. */
    115 	if (doline + doword + dochar == 0)
    116 		doline = doword = dochar = 1;
    117 
    118 	if (!*argv) {
    119 		cnt(NULL);
    120 	} else {
    121 		int dototal = (argc > 1);
    122 
    123 		do {
    124 			cnt(*argv);
    125 		} while(*++argv);
    126 
    127 		if (dototal) {
    128 			print_counts(tlinect, twordct, tcharct, "total");
    129 		}
    130 	}
    131 
    132 	exit(rval);
    133 }
    134 
    135 static void
    136 cnt(file)
    137 	char *file;
    138 {
    139 	u_char *C;
    140 	short gotsp;
    141 	int len;
    142 	u_int64_t linect, wordct, charct;
    143 	struct stat sb;
    144 	int fd;
    145 	u_char buf[MAXBSIZE];
    146 
    147 	linect = wordct = charct = 0;
    148 	if (file) {
    149 		if ((fd = open(file, O_RDONLY, 0)) < 0) {
    150 			warn("%s", file);
    151 			rval = 1;
    152 			return;
    153 		}
    154 	} else  {
    155 		fd = STDIN_FILENO;
    156 	}
    157 
    158 	if (!doword) {
    159 		/*
    160 		 * line counting is split out because it's a lot
    161 		 * faster to get lines than to get words, since
    162 		 * the word count requires some logic.
    163 		 */
    164 		if (doline) {
    165 			while ((len = read(fd, buf, MAXBSIZE)) > 0) {
    166 				charct += len;
    167 				for (C = buf; len--; ++C)
    168 					if (*C == '\n')
    169 						++linect;
    170 			}
    171 			if (len == -1) {
    172 				warn ("%s", file);
    173 				rval = 1;
    174 			}
    175 		}
    176 
    177 		/*
    178 		 * if all we need is the number of characters and
    179 		 * it's a directory or a regular or linked file, just
    180 		 * stat the puppy.  We avoid testing for it not being
    181 		 * a special device in case someone adds a new type
    182 		 * of inode.
    183 		 */
    184 		else if (dochar) {
    185 			if (fstat(fd, &sb)) {
    186 				warn("%s", file);
    187 				rval = 1;
    188 			} else {
    189 				if (S_ISREG(sb.st_mode) ||
    190 				    S_ISLNK(sb.st_mode) ||
    191 				    S_ISDIR(sb.st_mode)) {
    192 					charct = sb.st_size;
    193 				} else {
    194 					while ((len = read(fd, buf, MAXBSIZE)) > 0)
    195 						charct += len;
    196 					if (len == -1) {
    197 						warn ("%s", file);
    198 						rval = 1;
    199 					}
    200 				}
    201 			}
    202 		}
    203 	}
    204 	else
    205 	{
    206 		/* do it the hard way... */
    207 		gotsp = 1;
    208 		while ((len = read(fd, buf, MAXBSIZE)) > 0) {
    209 			charct += len;
    210 			for (C = buf; len--; ++C) {
    211 				if (isspace(*C)) {
    212 					gotsp = 1;
    213 					if (*C == '\n') {
    214 						++linect;
    215 					}
    216 				} else {
    217 					/*
    218 					 * This line implements the POSIX
    219 					 * spec, i.e. a word is a "maximal
    220 					 * string of characters delimited by
    221 					 * whitespace."  Notice nothing was
    222 					 * said about a character being
    223 					 * printing or non-printing.
    224 					 */
    225 					if (gotsp) {
    226 						gotsp = 0;
    227 						++wordct;
    228 					}
    229 				}
    230 			}
    231 		}
    232 		if (len == -1) {
    233 			warn("%s", file);
    234 			rval = 1;
    235 		}
    236 	}
    237 
    238 	print_counts(linect, wordct, charct, file ? file : "");
    239 
    240 	/* don't bother checkint doline, doword, or dochar --- speeds
    241            up the common case */
    242 	tlinect += linect;
    243 	twordct += wordct;
    244 	tcharct += charct;
    245 
    246 	if (close(fd)) {
    247 		warn ("%s", file);
    248 		rval = 1;
    249 	}
    250 }
    251 
    252 static void
    253 print_counts(lines, words, chars, name)
    254 	u_int64_t lines;
    255 	u_int64_t words;
    256 	u_int64_t chars;
    257 	char *name;
    258 {
    259 
    260 	if (doline)
    261 		printf(qfmt, lines);
    262 	if (doword)
    263 		printf(qfmt, words);
    264 	if (dochar)
    265 		printf(qfmt, chars);
    266 
    267 	printf(" %s\n", name);
    268 }
    269 
    270 static void
    271 usage()
    272 {
    273 	(void)fprintf(stderr, "usage: wc [-clw] [files]\n");
    274 	exit(1);
    275 }
    276