Home | History | Annotate | Line # | Download | only in split
split.c revision 1.24 
      1 /*	$NetBSD: split.c,v 1.24 2008/07/21 14:19:26 lukem Exp $	*/
      2 
      3 /*
      4  * Copyright (c) 1987, 1993, 1994
      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. Neither the name of the University nor the names of its contributors
     16  *    may be used to endorse or promote products derived from this software
     17  *    without specific prior written permission.
     18  *
     19  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
     20  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
     21  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
     22  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
     23  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
     24  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
     25  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
     26  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
     27  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
     28  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
     29  * SUCH DAMAGE.
     30  */
     31 
     32 #include <sys/cdefs.h>
     33 #ifndef lint
     34 __COPYRIGHT("@(#) Copyright (c) 1987, 1993, 1994\
     35  The Regents of the University of California.  All rights reserved.");
     36 #endif /* not lint */
     37 
     38 #ifndef lint
     39 #if 0
     40 static char sccsid[] = "@(#)split.c	8.3 (Berkeley) 4/25/94";
     41 #endif
     42 __RCSID("$NetBSD: split.c,v 1.24 2008/07/21 14:19:26 lukem Exp $");
     43 #endif /* not lint */
     44 
     45 #include <sys/param.h>
     46 
     47 #include <ctype.h>
     48 #include <err.h>
     49 #include <errno.h>
     50 #include <fcntl.h>
     51 #include <stdio.h>
     52 #include <stdlib.h>
     53 #include <string.h>
     54 #include <unistd.h>
     55 
     56 #define DEFLINE	1000		/* Default num lines per file. */
     57 
     58 static int file_open;		/* If a file open. */
     59 static int ifd = STDIN_FILENO, ofd = -1; /* Input/output file descriptors. */
     60 static char *fname;		/* File name prefix. */
     61 static size_t sfxlen = 2;		/* suffix length. */
     62 
     63 int  main(int, char **);
     64 static void newfile(void);
     65 static void split1(off_t, int);
     66 static void split2(off_t);
     67 static void split3(off_t);
     68 static void usage(void) __dead;
     69 static size_t bigwrite(int, void const *, size_t);
     70 
     71 int
     72 main(int argc, char *argv[])
     73 {
     74 	int ch;
     75 	char *ep, *p;
     76 	char const *base;
     77 	off_t bytecnt = 0;	/* Byte count to split on. */
     78 	off_t numlines = 0;	/* Line count to split on. */
     79 	off_t chunks = 0;	/* Number of chunks to split into. */
     80 
     81 	while ((ch = getopt(argc, argv, "0123456789b:l:a:n:")) != -1)
     82 		switch (ch) {
     83 		case '0': case '1': case '2': case '3': case '4':
     84 		case '5': case '6': case '7': case '8': case '9':
     85 			/*
     86 			 * Undocumented kludge: split was originally designed
     87 			 * to take a number after a dash.
     88 			 */
     89 			if (numlines == 0) {
     90 				p = argv[optind - 1];
     91 				if (p[0] == '-' && p[1] == ch && !p[2])
     92 					p++;
     93 				else
     94 					p = argv[optind] + 1;
     95 				numlines = strtoull(p, &ep, 10);
     96 				if (numlines == 0 || *ep != '\0')
     97 					errx(1, "%s: illegal line count.", p);
     98 			}
     99 			break;
    100 		case 'b':		/* Byte count. */
    101 			if (!isdigit((unsigned char)optarg[0]) ||
    102 			    (bytecnt = strtoull(optarg, &ep, 10)) == 0 ||
    103 			    (*ep != '\0' && *ep != 'k' && *ep != 'm'))
    104 				errx(1, "%s: illegal byte count.", optarg);
    105 			if (*ep == 'k')
    106 				bytecnt *= 1024;
    107 			else if (*ep == 'm')
    108 				bytecnt *= 1024 * 1024;
    109 			break;
    110 		case 'l':		/* Line count. */
    111 			if (numlines != 0)
    112 				usage();
    113 			if (!isdigit((unsigned char)optarg[0]) ||
    114 			    (numlines = strtoull(optarg, &ep, 10)) == 0 ||
    115 			    *ep != '\0')
    116 				errx(1, "%s: illegal line count.", optarg);
    117 			break;
    118 		case 'a':		/* Suffix length. */
    119 			if (!isdigit((unsigned char)optarg[0]) ||
    120 			    (sfxlen = (size_t)strtoul(optarg, &ep, 10)) == 0 ||
    121 			    *ep != '\0')
    122 				errx(1, "%s: illegal suffix length.", optarg);
    123 			break;
    124 		case 'n':		/* Chunks. */
    125 			if (!isdigit((unsigned char)optarg[0]) ||
    126 			    (chunks = (size_t)strtoul(optarg, &ep, 10)) == 0 ||
    127 			    *ep != '\0')
    128 				errx(1, "%s: illegal number of chunks.", optarg);
    129 			break;
    130 		default:
    131 			usage();
    132 		}
    133 	argv += optind;
    134 	argc -= optind;
    135 
    136 	if (*argv != NULL) {
    137 		if (strcmp(*argv, "-") != 0 &&
    138 		    (ifd = open(*argv, O_RDONLY, 0)) < 0)
    139 			err(1, "%s", *argv);
    140 		++argv;
    141 	}
    142 
    143 
    144 	base = (*argv != NULL) ? *argv++ : "x";
    145 	if ((fname = malloc(strlen(base) + sfxlen + 1)) == NULL)
    146 		err(EXIT_FAILURE, NULL);
    147 	(void)strcpy(fname, base);		/* File name prefix. */
    148 
    149 	if (*argv != NULL)
    150 		usage();
    151 
    152 	if (numlines == 0)
    153 		numlines = DEFLINE;
    154 	else if (bytecnt || chunks)
    155 		usage();
    156 
    157 	if (bytecnt && chunks)
    158 		usage();
    159 
    160 	if (bytecnt)
    161 		split1(bytecnt, 0);
    162 	else if (chunks)
    163 		split3(chunks);
    164 	else
    165 		split2(numlines);
    166 
    167 	return 0;
    168 }
    169 
    170 /*
    171  * split1 --
    172  *	Split the input by bytes.
    173  */
    174 static void
    175 split1(off_t bytecnt, int maxcnt)
    176 {
    177 	off_t bcnt;
    178 	ssize_t dist, len;
    179 	char *C;
    180 	char bfr[MAXBSIZE];
    181 	int nfiles;
    182 
    183 	nfiles = 0;
    184 
    185 	for (bcnt = 0;;)
    186 		switch (len = read(ifd, bfr, MAXBSIZE)) {
    187 		case 0:
    188 			exit(0);
    189 			/* NOTREACHED */
    190 		case -1:
    191 			err(1, "read");
    192 			/* NOTREACHED */
    193 		default:
    194 			if (!file_open) {
    195 				if (!maxcnt || (nfiles < maxcnt)) {
    196 					newfile();
    197 					nfiles++;
    198 					file_open = 1;
    199 				}
    200 			}
    201 			if (bcnt + len >= bytecnt) {
    202 				/* LINTED: bytecnt - bcnt <= len */
    203 				dist = bytecnt - bcnt;
    204 				if (bigwrite(ofd, bfr, dist) != dist)
    205 					err(1, "write");
    206 				len -= dist;
    207 				for (C = bfr + dist; len >= bytecnt;
    208 				    /* LINTED: bytecnt <= len */
    209 				    len -= bytecnt, C += bytecnt) {
    210 					if (!maxcnt || (nfiles < maxcnt)) {
    211 						newfile();
    212 						nfiles++;
    213 					}
    214 					/* LINTED: as above */
    215 					if (bigwrite(ofd,
    216 					    C, bytecnt) != bytecnt)
    217 						err(1, "write");
    218 				}
    219 				if (len) {
    220 					if (!maxcnt || (nfiles < maxcnt)) {
    221 						newfile();
    222 						nfiles++;
    223 					}
    224 					/* LINTED: len >= 0 */
    225 					if (bigwrite(ofd, C, len) != len)
    226 						err(1, "write");
    227 				} else
    228 					file_open = 0;
    229 				bcnt = len;
    230 			} else {
    231 				bcnt += len;
    232 				/* LINTED: len >= 0 */
    233 				if (bigwrite(ofd, bfr, len) != len)
    234 					err(1, "write");
    235 			}
    236 		}
    237 }
    238 
    239 /*
    240  * split2 --
    241  *	Split the input by lines.
    242  */
    243 static void
    244 split2(off_t numlines)
    245 {
    246 	off_t lcnt;
    247 	size_t bcnt;
    248 	ssize_t len;
    249 	char *Ce, *Cs;
    250 	char bfr[MAXBSIZE];
    251 
    252 	for (lcnt = 0;;)
    253 		switch (len = read(ifd, bfr, MAXBSIZE)) {
    254 		case 0:
    255 			exit(0);
    256 			/* NOTREACHED */
    257 		case -1:
    258 			err(1, "read");
    259 			/* NOTREACHED */
    260 		default:
    261 			if (!file_open) {
    262 				newfile();
    263 				file_open = 1;
    264 			}
    265 			for (Cs = Ce = bfr; len--; Ce++)
    266 				if (*Ce == '\n' && ++lcnt == numlines) {
    267 					bcnt = Ce - Cs + 1;
    268 					if (bigwrite(ofd, Cs, bcnt) != bcnt)
    269 						err(1, "write");
    270 					lcnt = 0;
    271 					Cs = Ce + 1;
    272 					if (len)
    273 						newfile();
    274 					else
    275 						file_open = 0;
    276 				}
    277 			if (Cs < Ce) {
    278 				bcnt = Ce - Cs;
    279 				if (bigwrite(ofd, Cs, bcnt) != bcnt)
    280 					err(1, "write");
    281 			}
    282 		}
    283 }
    284 
    285 /*
    286  * split3 --
    287  *	Split the input into specified number of chunks
    288  */
    289 static void
    290 split3(off_t chunks)
    291 {
    292 	struct stat sb;
    293 
    294 	if (fstat(ifd, &sb) == -1) {
    295 		err(1, "stat");
    296 		/* NOTREACHED */
    297 	}
    298 
    299 	if (chunks > sb.st_size) {
    300 		errx(1, "can't split into more than %d files",
    301 				(int)sb.st_size);
    302 		/* NOTREACHED */
    303 	}
    304 
    305 	split1(sb.st_size/chunks, chunks);
    306 }
    307 
    308 /*
    309  * newfile --
    310  *	Open a new output file.
    311  */
    312 static void
    313 newfile(void)
    314 {
    315 	static int fnum;
    316 	static char *fpnt;
    317 	int quot, i;
    318 
    319 	if (ofd == -1) {
    320 		fpnt = fname + strlen(fname);
    321 		fpnt[sfxlen] = '\0';
    322 	} else if (close(ofd) != 0)
    323 		err(1, "%s", fname);
    324 
    325 	quot = fnum;
    326 	for (i = sfxlen - 1; i >= 0; i--) {
    327 		fpnt[i] = quot % 26 + 'a';
    328 		quot = quot / 26;
    329 	}
    330 	if (quot > 0)
    331 		errx(1, "too many files.");
    332 	++fnum;
    333 	if ((ofd = open(fname, O_WRONLY | O_CREAT | O_TRUNC, DEFFILEMODE)) < 0)
    334 		err(1, "%s", fname);
    335 }
    336 
    337 static size_t
    338 bigwrite(int fd, const void *buf, size_t len)
    339 {
    340 	const char *ptr = buf;
    341 	size_t sofar = 0;
    342 	ssize_t w;
    343 
    344 	while (len != 0) {
    345 		if  ((w = write(fd, ptr, len)) == -1)
    346 			return sofar;
    347 		len -= w;
    348 		ptr += w;
    349 		sofar += w;
    350 	}
    351 	return sofar;
    352 }
    353 
    354 
    355 static void
    356 usage(void)
    357 {
    358 	(void)fprintf(stderr,
    359 "usage: %s [-b byte_count] [-l line_count] [-n chunk_count] [-a suffix_length] "
    360 "[file [prefix]]\n", getprogname());
    361 	exit(1);
    362 }
    363