xref: /freebsd/usr.bin/split/split.c (revision 5ca8e32633c4ffbbcd6762e5888b6a4ba0708c6c)
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/param.h>
33 #include <sys/stat.h>
34 
35 #include <ctype.h>
36 #include <err.h>
37 #include <errno.h>
38 #include <fcntl.h>
39 #include <inttypes.h>
40 #include <libutil.h>
41 #include <limits.h>
42 #include <locale.h>
43 #include <stdbool.h>
44 #include <stdint.h>
45 #include <stdio.h>
46 #include <stdlib.h>
47 #include <string.h>
48 #include <unistd.h>
49 #include <regex.h>
50 #include <sysexits.h>
51 
52 #define DEFLINE	1000			/* Default num lines per file. */
53 
54 static off_t	 bytecnt;		/* Byte count to split on. */
55 static long	 chunks;		/* Chunks count to split into. */
56 static bool      clobber = true;        /* Whether to overwrite existing output files. */
57 static long	 numlines;		/* Line count to split on. */
58 static int	 file_open;		/* If a file open. */
59 static int	 ifd = -1, ofd = -1;	/* Input/output file descriptors. */
60 static char	 fname[MAXPATHLEN];	/* File name prefix. */
61 static regex_t	 rgx;
62 static int	 pflag;
63 static bool	 dflag;
64 static long	 sufflen = 2;		/* File name suffix length. */
65 static bool	 autosfx = true;	/* Whether to auto-extend the suffix length. */
66 
67 static void newfile(void);
68 static void split1(void);
69 static void split2(void);
70 static void split3(void);
71 static void usage(void) __dead2;
72 
73 int
74 main(int argc, char **argv)
75 {
76 	char errbuf[64];
77 	const char *p, *errstr;
78 	int ch, error;
79 
80 	setlocale(LC_ALL, "");
81 
82 	dflag = false;
83 	while ((ch = getopt(argc, argv, "0::1::2::3::4::5::6::7::8::9::a:b:cdl:n:p:")) != -1)
84 		switch (ch) {
85 		case '0': case '1': case '2': case '3': case '4':
86 		case '5': case '6': case '7': case '8': case '9':
87 			/*
88 			 * Undocumented kludge: split was originally designed
89 			 * to take a number after a dash.
90 			 */
91 			if (numlines != 0)
92 				usage();
93 			numlines = ch - '0';
94 			p = optarg ? optarg : "";
95 			while (numlines >= 0 && *p >= '0' && *p <= '9')
96 				numlines = numlines * 10 + *p++ - '0';
97 			if (numlines <= 0 || *p != '\0')
98 				errx(EX_USAGE, "%c%s: line count is invalid",
99 				    ch, optarg ? optarg : "");
100 			break;
101 		case 'a':		/* Suffix length */
102 			sufflen = strtonum(optarg, 0, INT_MAX, &errstr);
103 			if (errstr != NULL) {
104 				errx(EX_USAGE, "%s: suffix length is %s",
105 				    optarg, errstr);
106 			}
107 			if (sufflen == 0) {
108 				sufflen = 2;
109 				autosfx = true;
110 			} else {
111 				autosfx = false;
112 			}
113 			break;
114 		case 'b':		/* Byte count. */
115 			if (expand_number(optarg, &bytecnt) != 0) {
116 				errx(EX_USAGE, "%s: byte count is invalid",
117 				    optarg);
118 			}
119 			break;
120 		case 'c':               /* Continue, don't overwrite output files. */
121 			clobber = false;
122 			break;
123 		case 'd':		/* Decimal suffix */
124 			dflag = true;
125 			break;
126 		case 'l':		/* Line count. */
127 			if (numlines != 0)
128 				usage();
129 			numlines = strtonum(optarg, 1, LONG_MAX, &errstr);
130 			if (errstr != NULL) {
131 				errx(EX_USAGE, "%s: line count is %s",
132 				    optarg, errstr);
133 			}
134 			break;
135 		case 'n':		/* Chunks. */
136 			chunks = strtonum(optarg, 1, LONG_MAX, &errstr);
137 			if (errstr != NULL) {
138 				errx(EX_USAGE, "%s: number of chunks is %s",
139 				    optarg, errstr);
140 			}
141 			break;
142 
143 		case 'p':		/* pattern matching. */
144 			error = regcomp(&rgx, optarg, REG_EXTENDED|REG_NOSUB);
145 			if (error != 0) {
146 				regerror(error, &rgx, errbuf, sizeof(errbuf));
147 				errx(EX_USAGE, "%s: regex is invalid: %s",
148 				    optarg, errbuf);
149 			}
150 			pflag = 1;
151 			break;
152 		default:
153 			usage();
154 		}
155 	argv += optind;
156 	argc -= optind;
157 
158 	if (argc > 0) {			/* Input file. */
159 		if (strcmp(*argv, "-") == 0)
160 			ifd = STDIN_FILENO;
161 		else if ((ifd = open(*argv, O_RDONLY, 0)) < 0)
162 			err(EX_NOINPUT, "%s", *argv);
163 		++argv;
164 		--argc;
165 	}
166 	if (argc > 0) {			/* File name prefix. */
167 		if (strlcpy(fname, *argv, sizeof(fname)) >= sizeof(fname)) {
168 			errx(EX_USAGE, "%s: file name prefix is too long",
169 			    *argv);
170 		}
171 		++argv;
172 		--argc;
173 	}
174 	if (argc > 0)
175 		usage();
176 
177 	if (strlen(fname) + (unsigned long)sufflen >= sizeof(fname))
178 		errx(EX_USAGE, "suffix is too long");
179 	if (pflag && (numlines != 0 || bytecnt != 0 || chunks != 0))
180 		usage();
181 
182 	if (numlines == 0)
183 		numlines = DEFLINE;
184 	else if (bytecnt != 0 || chunks != 0)
185 		usage();
186 
187 	if (bytecnt != 0 && chunks != 0)
188 		usage();
189 
190 	if (ifd == -1)				/* Stdin by default. */
191 		ifd = 0;
192 
193 	if (bytecnt != 0) {
194 		split1();
195 		exit (0);
196 	} else if (chunks != 0) {
197 		split3();
198 		exit (0);
199 	}
200 	split2();
201 	if (pflag)
202 		regfree(&rgx);
203 	exit(0);
204 }
205 
206 /*
207  * split1 --
208  *	Split the input by bytes.
209  */
210 static void
211 split1(void)
212 {
213 	static char bfr[MAXBSIZE];
214 	off_t bcnt;
215 	char *C;
216 	ssize_t dist, len;
217 	int nfiles;
218 
219 	nfiles = 0;
220 
221 	for (bcnt = 0;;)
222 		switch ((len = read(ifd, bfr, sizeof(bfr)))) {
223 		case 0:
224 			exit(0);
225 		case -1:
226 			err(EX_IOERR, "read");
227 			/* NOTREACHED */
228 		default:
229 			if (!file_open) {
230 				if (chunks == 0 || nfiles < chunks) {
231 					newfile();
232 					nfiles++;
233 				}
234 			}
235 			if (bcnt + len >= bytecnt) {
236 				dist = bytecnt - bcnt;
237 				if (write(ofd, bfr, dist) != dist)
238 					err(EX_IOERR, "write");
239 				len -= dist;
240 				for (C = bfr + dist; len >= bytecnt;
241 				     len -= bytecnt, C += bytecnt) {
242 					if (chunks == 0 || nfiles < chunks) {
243 						newfile();
244 						nfiles++;
245 					}
246 					if (write(ofd, C, bytecnt) != bytecnt)
247 						err(EX_IOERR, "write");
248 				}
249 				if (len != 0) {
250 					if (chunks == 0 || nfiles < chunks) {
251 						newfile();
252 						nfiles++;
253 					}
254 					if (write(ofd, C, len) != len)
255 						err(EX_IOERR, "write");
256 				} else {
257 					file_open = 0;
258 				}
259 				bcnt = len;
260 			} else {
261 				bcnt += len;
262 				if (write(ofd, bfr, len) != len)
263 					err(EX_IOERR, "write");
264 			}
265 		}
266 }
267 
268 /*
269  * split2 --
270  *	Split the input by lines.
271  */
272 static void
273 split2(void)
274 {
275 	char *buf;
276 	size_t bufsize;
277 	ssize_t len;
278 	long lcnt = 0;
279 	FILE *infp;
280 
281 	buf = NULL;
282 	bufsize = 0;
283 
284 	/* Stick a stream on top of input file descriptor */
285 	if ((infp = fdopen(ifd, "r")) == NULL)
286 		err(EX_NOINPUT, "fdopen");
287 
288 	/* Process input one line at a time */
289 	while ((errno = 0, len = getline(&buf, &bufsize, infp)) > 0) {
290 		/* Check if we need to start a new file */
291 		if (pflag) {
292 			regmatch_t pmatch;
293 
294 			pmatch.rm_so = 0;
295 			pmatch.rm_eo = len - 1;
296 			if (regexec(&rgx, buf, 0, &pmatch, REG_STARTEND) == 0)
297 				newfile();
298 		} else if (lcnt++ == numlines) {
299 			newfile();
300 			lcnt = 1;
301 		}
302 
303 		/* Open output file if needed */
304 		if (!file_open)
305 			newfile();
306 
307 		/* Write out line */
308 		if (write(ofd, buf, len) != len)
309 			err(EX_IOERR, "write");
310 	}
311 
312 	/* EOF or error? */
313 	if ((len == -1 && errno != 0) || ferror(infp))
314 		err(EX_IOERR, "read");
315 	else
316 		exit(0);
317 }
318 
319 /*
320  * split3 --
321  *	Split the input into specified number of chunks
322  */
323 static void
324 split3(void)
325 {
326 	struct stat sb;
327 
328 	if (fstat(ifd, &sb) == -1) {
329 		err(1, "stat");
330 		/* NOTREACHED */
331 	}
332 
333 	if (chunks > sb.st_size) {
334 		errx(1, "can't split into more than %d files",
335 		    (int)sb.st_size);
336 		/* NOTREACHED */
337 	}
338 
339 	bytecnt = sb.st_size / chunks;
340 	split1();
341 }
342 
343 
344 /*
345  * newfile --
346  *	Open a new output file.
347  */
348 static void
349 newfile(void)
350 {
351 	long i, maxfiles, tfnum;
352 	static long fnum;
353 	static char *fpnt;
354 	char beg, end;
355 	int pattlen;
356 	int flags = O_WRONLY | O_CREAT | O_TRUNC;
357 
358 	if (!clobber)
359 		flags |= O_EXCL;
360 
361 	if (ofd == -1) {
362 		if (fname[0] == '\0') {
363 			fname[0] = 'x';
364 			fpnt = fname + 1;
365 		} else {
366 			fpnt = fname + strlen(fname);
367 		}
368 	} else if (close(ofd) != 0)
369 		err(1, "%s", fname);
370 
371 	again:
372 	if (dflag) {
373 		beg = '0';
374 		end = '9';
375 	}
376 	else {
377 		beg = 'a';
378 		end = 'z';
379 	}
380 	pattlen = end - beg + 1;
381 
382 	/*
383 	 * If '-a' is not specified, then we automatically expand the
384 	 * suffix length to accomodate splitting all input.  We do this
385 	 * by moving the suffix pointer (fpnt) forward and incrementing
386 	 * sufflen by one, thereby yielding an additional two characters
387 	 * and allowing all output files to sort such that 'cat *' yields
388 	 * the input in order.  I.e., the order is '... xyy xyz xzaaa
389 	 * xzaab ... xzyzy, xzyzz, xzzaaaa, xzzaaab' and so on.
390 	 */
391 	if (!dflag && autosfx && (fpnt[0] == 'y') &&
392 			strspn(fpnt+1, "z") == strlen(fpnt+1)) {
393 		fpnt = fname + strlen(fname) - sufflen;
394 		fpnt[sufflen + 2] = '\0';
395 		fpnt[0] = end;
396 		fpnt[1] = beg;
397 
398 		/*  Basename | Suffix
399 		 *  before:
400 		 *  x        | yz
401 		 *  after:
402 		 *  xz       | a.. */
403 		fpnt++;
404 		sufflen++;
405 
406 		/* Reset so we start back at all 'a's in our extended suffix. */
407 		fnum = 0;
408 	}
409 
410 	/* maxfiles = pattlen^sufflen, but don't use libm. */
411 	for (maxfiles = 1, i = 0; i < sufflen; i++)
412 		if (LONG_MAX / pattlen < maxfiles)
413 			errx(EX_USAGE, "suffix is too long (max %ld)", i);
414 		else
415 			maxfiles *= pattlen;
416 
417 	if (fnum == maxfiles)
418 		errx(EX_DATAERR, "too many files");
419 
420 	/* Generate suffix of sufflen letters */
421 	tfnum = fnum;
422 	i = sufflen - 1;
423 	do {
424 		fpnt[i] = tfnum % pattlen + beg;
425 		tfnum /= pattlen;
426 	} while (i-- > 0);
427 	fpnt[sufflen] = '\0';
428 
429 	++fnum;
430 	if ((ofd = open(fname, flags, DEFFILEMODE)) < 0) {
431 		if (!clobber && errno == EEXIST)
432 			goto again;
433 		err(EX_IOERR, "%s", fname);
434 	}
435 	file_open = 1;
436 }
437 
438 static void
439 usage(void)
440 {
441 	(void)fprintf(stderr,
442 "usage: split [-cd] [-l line_count] [-a suffix_length] [file [prefix]]\n"
443 "       split [-cd] -b byte_count[K|k|M|m|G|g] [-a suffix_length] [file [prefix]]\n"
444 "       split [-cd] -n chunk_count [-a suffix_length] [file [prefix]]\n"
445 "       split [-cd] -p pattern [-a suffix_length] [file [prefix]]\n");
446 	exit(EX_USAGE);
447 }
448