xref: /freebsd/usr.bin/split/split.c (revision 22cf89c938886d14f5796fc49f9f020c23ea8eaf)
1 /*-
2  * SPDX-License-Identifier: BSD-3-Clause
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 static const char copyright[] =
35 "@(#) Copyright (c) 1987, 1993, 1994\n\
36 	The Regents of the University of California.  All rights reserved.\n";
37 #endif
38 
39 #ifndef lint
40 static const char sccsid[] = "@(#)split.c	8.2 (Berkeley) 4/16/94";
41 #endif
42 
43 #include <sys/param.h>
44 #include <sys/stat.h>
45 
46 #include <ctype.h>
47 #include <err.h>
48 #include <errno.h>
49 #include <fcntl.h>
50 #include <inttypes.h>
51 #include <libutil.h>
52 #include <limits.h>
53 #include <locale.h>
54 #include <stdbool.h>
55 #include <stdint.h>
56 #include <stdio.h>
57 #include <stdlib.h>
58 #include <string.h>
59 #include <unistd.h>
60 #include <regex.h>
61 #include <sysexits.h>
62 
63 #define DEFLINE	1000			/* Default num lines per file. */
64 
65 static off_t	 bytecnt;		/* Byte count to split on. */
66 static off_t	 chunks = 0;		/* Chunks count to split into. */
67 static bool      clobber = true;        /* Whether to overwrite existing output files. */
68 static long	 numlines;		/* Line count to split on. */
69 static int	 file_open;		/* If a file open. */
70 static int	 ifd = -1, ofd = -1;	/* Input/output file descriptors. */
71 static char	 fname[MAXPATHLEN];	/* File name prefix. */
72 static regex_t	 rgx;
73 static int	 pflag;
74 static bool	 dflag;
75 static long	 sufflen = 2;		/* File name suffix length. */
76 static int	 autosfx = 1;		/* Whether to auto-extend the suffix length. */
77 
78 static void newfile(void);
79 static void split1(void);
80 static void split2(void);
81 static void split3(void);
82 static void usage(void) __dead2;
83 
84 int
85 main(int argc, char **argv)
86 {
87 	const char *p;
88 	char *ep;
89 	int ch, error;
90 
91 	setlocale(LC_ALL, "");
92 
93 	dflag = false;
94 	while ((ch = getopt(argc, argv, "0::1::2::3::4::5::6::7::8::9::a:b:cdl:n:p:")) != -1)
95 		switch (ch) {
96 		case '0': case '1': case '2': case '3': case '4':
97 		case '5': case '6': case '7': case '8': case '9':
98 			/*
99 			 * Undocumented kludge: split was originally designed
100 			 * to take a number after a dash.
101 			 */
102 			if (numlines != 0)
103 				usage();
104 			numlines = ch - '0';
105 			p = optarg ? optarg : "";
106 			while (numlines >= 0 && *p >= '0' && *p <= '9')
107 				numlines = numlines * 10 + *p++ - '0';
108 			if (numlines <= 0 || *p != '\0')
109 				errx(EX_USAGE, "%c%s: illegal line count", ch,
110 				    optarg ? optarg : "");
111 			break;
112 		case 'a':		/* Suffix length */
113 			if ((sufflen = strtol(optarg, &ep, 10)) <= 0 || *ep)
114 				errx(EX_USAGE,
115 				    "%s: illegal suffix length", optarg);
116 			autosfx = 0;
117 			break;
118 		case 'b':		/* Byte count. */
119 			errno = 0;
120 			error = expand_number(optarg, &bytecnt);
121 			if (error == -1)
122 				errx(EX_USAGE, "%s: offset too large", optarg);
123 			break;
124 		case 'c':               /* Continue, don't overwrite output files. */
125 			clobber = false;
126 			break;
127 		case 'd':		/* Decimal suffix */
128 			dflag = true;
129 			break;
130 		case 'l':		/* Line count. */
131 			if (numlines != 0)
132 				usage();
133 			if ((numlines = strtol(optarg, &ep, 10)) <= 0 || *ep)
134 				errx(EX_USAGE,
135 				    "%s: illegal line count", optarg);
136 			break;
137 		case 'n':		/* Chunks. */
138 			if (!isdigit((unsigned char)optarg[0]) ||
139 			    (chunks = (size_t)strtoul(optarg, &ep, 10)) == 0 ||
140 			    *ep != '\0') {
141 				errx(EX_USAGE, "%s: illegal number of chunks",
142 				     optarg);
143 			}
144 			break;
145 
146 		case 'p':		/* pattern matching. */
147 			if (regcomp(&rgx, optarg, REG_EXTENDED|REG_NOSUB) != 0)
148 				errx(EX_USAGE, "%s: illegal regexp", optarg);
149 			pflag = 1;
150 			break;
151 		default:
152 			usage();
153 		}
154 	argv += optind;
155 	argc -= optind;
156 
157 	if (argc > 0) {			/* Input file. */
158 		if (strcmp(*argv, "-") == 0)
159 			ifd = STDIN_FILENO;
160 		else if ((ifd = open(*argv, O_RDONLY, 0)) < 0)
161 			err(EX_NOINPUT, "%s", *argv);
162 		++argv;
163 		--argc;
164 	}
165 	if (argc > 0) {			/* File name prefix. */
166 		if (strlcpy(fname, *argv, sizeof(fname)) >= sizeof(fname))
167 			errx(EX_USAGE, "file name prefix is too long: %s",
168 			    *argv);
169 		++argv;
170 		--argc;
171 	}
172 	if (argc > 0)
173 		usage();
174 
175 	if (strlen(fname) + (unsigned long)sufflen >= sizeof(fname))
176 		errx(EX_USAGE, "suffix is too long");
177 	if (pflag && (numlines != 0 || bytecnt != 0 || chunks != 0))
178 		usage();
179 
180 	if (numlines == 0)
181 		numlines = DEFLINE;
182 	else if (bytecnt != 0 || chunks != 0)
183 		usage();
184 
185 	if (bytecnt && chunks)
186 		usage();
187 
188 	if (ifd == -1)				/* Stdin by default. */
189 		ifd = 0;
190 
191 	if (bytecnt) {
192 		split1();
193 		exit (0);
194 	} else if (chunks) {
195 		split3();
196 		exit (0);
197 	}
198 	split2();
199 	if (pflag)
200 		regfree(&rgx);
201 	exit(0);
202 }
203 
204 /*
205  * split1 --
206  *	Split the input by bytes.
207  */
208 static void
209 split1(void)
210 {
211 	static char bfr[MAXBSIZE];
212 	off_t bcnt;
213 	char *C;
214 	ssize_t dist, len;
215 	int nfiles;
216 
217 	nfiles = 0;
218 
219 	for (bcnt = 0;;)
220 		switch ((len = read(ifd, bfr, sizeof(bfr)))) {
221 		case 0:
222 			exit(0);
223 		case -1:
224 			err(EX_IOERR, "read");
225 			/* NOTREACHED */
226 		default:
227 			if (!file_open) {
228 				if (!chunks || (nfiles < chunks)) {
229 					newfile();
230 					nfiles++;
231 				}
232 			}
233 			if (bcnt + len >= bytecnt) {
234 				dist = bytecnt - bcnt;
235 				if (write(ofd, bfr, dist) != dist)
236 					err(EX_IOERR, "write");
237 				len -= dist;
238 				for (C = bfr + dist; len >= bytecnt;
239 				    len -= bytecnt, C += bytecnt) {
240 					if (!chunks || (nfiles < chunks)) {
241 					newfile();
242 						nfiles++;
243 					}
244 					if (write(ofd,
245 					    C, bytecnt) != bytecnt)
246 						err(EX_IOERR, "write");
247 				}
248 				if (len != 0) {
249 					if (!chunks || (nfiles < chunks)) {
250 					newfile();
251 						nfiles++;
252 					}
253 					if (write(ofd, C, len) != len)
254 						err(EX_IOERR, "write");
255 				} else
256 					file_open = 0;
257 				bcnt = len;
258 			} else {
259 				bcnt += len;
260 				if (write(ofd, bfr, len) != len)
261 					err(EX_IOERR, "write");
262 			}
263 		}
264 }
265 
266 /*
267  * split2 --
268  *	Split the input by lines.
269  */
270 static void
271 split2(void)
272 {
273 	char *buf;
274 	size_t bufsize;
275 	ssize_t len;
276 	long lcnt = 0;
277 	FILE *infp;
278 
279 	buf = NULL;
280 	bufsize = 0;
281 
282 	/* Stick a stream on top of input file descriptor */
283 	if ((infp = fdopen(ifd, "r")) == NULL)
284 		err(EX_NOINPUT, "fdopen");
285 
286 	/* Process input one line at a time */
287 	while ((errno = 0, len = getline(&buf, &bufsize, infp)) > 0) {
288 		/* Check if we need to start a new file */
289 		if (pflag) {
290 			regmatch_t pmatch;
291 
292 			pmatch.rm_so = 0;
293 			pmatch.rm_eo = len - 1;
294 			if (regexec(&rgx, buf, 0, &pmatch, REG_STARTEND) == 0)
295 				newfile();
296 		} else if (lcnt++ == numlines) {
297 			newfile();
298 			lcnt = 1;
299 		}
300 
301 		/* Open output file if needed */
302 		if (!file_open)
303 			newfile();
304 
305 		/* Write out line */
306 		if (write(ofd, buf, len) != len)
307 			err(EX_IOERR, "write");
308 	}
309 
310 	/* EOF or error? */
311 	if ((len == -1 && errno != 0) || ferror(infp))
312 		err(EX_IOERR, "read");
313 	else
314 		exit(0);
315 }
316 
317 /*
318  * split3 --
319  *	Split the input into specified number of chunks
320  */
321 static void
322 split3(void)
323 {
324 	struct stat sb;
325 
326 	if (fstat(ifd, &sb) == -1) {
327 		err(1, "stat");
328 		/* NOTREACHED */
329 	}
330 
331 	if (chunks > sb.st_size) {
332 		errx(1, "can't split into more than %d files",
333 		    (int)sb.st_size);
334 		/* NOTREACHED */
335 	}
336 
337 	bytecnt = sb.st_size / chunks;
338 	split1();
339 }
340 
341 
342 /*
343  * newfile --
344  *	Open a new output file.
345  */
346 static void
347 newfile(void)
348 {
349 	long i, maxfiles, tfnum;
350 	static long fnum;
351 	static char *fpnt;
352 	char beg, end;
353 	int pattlen;
354 	int flags = O_WRONLY | O_CREAT | O_TRUNC;
355 
356 	if (!clobber)
357 		flags |= O_EXCL;
358 
359 	if (ofd == -1) {
360 		if (fname[0] == '\0') {
361 			fname[0] = 'x';
362 			fpnt = fname + 1;
363 		} else {
364 			fpnt = fname + strlen(fname);
365 		}
366 	} else if (close(ofd) != 0)
367 		err(1, "%s", fname);
368 
369 	again:
370 	if (dflag) {
371 		beg = '0';
372 		end = '9';
373 	}
374 	else {
375 		beg = 'a';
376 		end = 'z';
377 	}
378 	pattlen = end - beg + 1;
379 
380 	/*
381 	 * If '-a' is not specified, then we automatically expand the
382 	 * suffix length to accomodate splitting all input.  We do this
383 	 * by moving the suffix pointer (fpnt) forward and incrementing
384 	 * sufflen by one, thereby yielding an additional two characters
385 	 * and allowing all output files to sort such that 'cat *' yields
386 	 * the input in order.  I.e., the order is '... xyy xyz xzaaa
387 	 * xzaab ... xzyzy, xzyzz, xzzaaaa, xzzaaab' and so on.
388 	 */
389 	if (!dflag && autosfx && (fpnt[0] == 'y') &&
390 			strspn(fpnt+1, "z") == strlen(fpnt+1)) {
391 		fpnt = fname + strlen(fname) - sufflen;
392 		fpnt[sufflen + 2] = '\0';
393 		fpnt[0] = end;
394 		fpnt[1] = beg;
395 
396 		/*  Basename | Suffix
397 		 *  before:
398 		 *  x        | yz
399 		 *  after:
400 		 *  xz       | a.. */
401 		fpnt++;
402 		sufflen++;
403 
404 		/* Reset so we start back at all 'a's in our extended suffix. */
405 		fnum = 0;
406 	}
407 
408 	/* maxfiles = pattlen^sufflen, but don't use libm. */
409 	for (maxfiles = 1, i = 0; i < sufflen; i++)
410 		if (LONG_MAX / pattlen < maxfiles)
411 			errx(EX_USAGE, "suffix is too long (max %ld)", i);
412 		else
413 			maxfiles *= pattlen;
414 
415 	if (fnum == maxfiles)
416 		errx(EX_DATAERR, "too many files");
417 
418 	/* Generate suffix of sufflen letters */
419 	tfnum = fnum;
420 	i = sufflen - 1;
421 	do {
422 		fpnt[i] = tfnum % pattlen + beg;
423 		tfnum /= pattlen;
424 	} while (i-- > 0);
425 	fpnt[sufflen] = '\0';
426 
427 	++fnum;
428 	if ((ofd = open(fname, flags, DEFFILEMODE)) < 0) {
429 		if (!clobber && errno == EEXIST)
430 			goto again;
431 		err(EX_IOERR, "%s", fname);
432 	}
433 	file_open = 1;
434 }
435 
436 static void
437 usage(void)
438 {
439 	(void)fprintf(stderr,
440 "usage: split [-cd] [-l line_count] [-a suffix_length] [file [prefix]]\n"
441 "       split [-cd] -b byte_count[K|k|M|m|G|g] [-a suffix_length] [file [prefix]]\n"
442 "       split [-cd] -n chunk_count [-a suffix_length] [file [prefix]]\n"
443 "       split [-cd] -p pattern [-a suffix_length] [file [prefix]]\n");
444 	exit(EX_USAGE);
445 }
446