2 * Copyright (c) 2002 Tim J. Robbins.
5 * Redistribution and use in source and binary forms, with or without
6 * modification, are permitted provided that the following conditions
8 * 1. Redistributions of source code must retain the above copyright
9 * notice, this list of conditions and the following disclaimer.
10 * 2. Redistributions in binary form must reproduce the above copyright
11 * notice, this list of conditions and the following disclaimer in the
12 * documentation and/or other materials provided with the distribution.
14 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
15 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
16 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
17 * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
18 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
19 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
20 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
21 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
22 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
23 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
26 * $FreeBSD: src/usr.bin/csplit/csplit.c,v 1.9 2004/03/22 11:15:03 tjr Exp $
30 * csplit -- split files based on context
32 * This utility splits its input into numbered output files by line number
33 * or by a regular expression. Regular expression matches have an optional
34 * offset with them, allowing the split to occur a specified number of
35 * lines before or after the match.
37 * To handle negative offsets, we stop reading when the match occurs and
38 * store the offset that the file should have been split at, then use
39 * this output file as input until all the "overflowed" lines have been read.
40 * The file is then closed and truncated to the correct length.
42 * We assume that the output files can be seeked upon (ie. they cannot be
43 * symlinks to named pipes or character devices), but make no such
44 * assumption about the input.
47 #include <sys/types.h>
62 static void cleanup(void);
63 static void do_lineno(const char *);
64 static void do_rexp(const char *);
65 static char *get_line(void);
66 static void handlesig(int) __dead2;
67 static FILE *newfile(void);
68 static void toomuch(FILE *, long);
69 static void usage(void) __dead2;
72 * Command line options
74 static const char *prefix; /* File name prefix */
75 static long sufflen; /* Number of decimal digits for suffix */
76 static int sflag; /* Suppress output of file names */
77 static int kflag; /* Keep output if error occurs */
80 * Other miscellaneous globals (XXX too many)
82 static long lineno; /* Current line number in input file */
83 static long reps; /* Number of repetitions for this pattern */
84 static long nfiles; /* Number of files output so far */
85 static long maxfiles; /* Maximum number of files we can create */
86 static char currfile[PATH_MAX]; /* Current output file */
87 static const char *infn; /* Name of the input file */
88 static FILE *infile; /* Input file handle */
89 static FILE *overfile; /* Overflow file for toomuch() */
90 static off_t truncofs; /* Offset this file should be truncated at */
91 static int doclean; /* Should cleanup() remove output? */
94 main(int argc, char *argv[])
103 setlocale(LC_ALL, "");
108 while ((ch = getopt(argc, argv, "ksf:n:")) > 0) {
118 sufflen = strtol(optarg, &ep, 10);
119 if (sufflen <= 0 || *ep != '\0' || errno != 0)
120 errx(1, "%s: bad suffix length", optarg);
131 if (sufflen + strlen(prefix) >= PATH_MAX)
132 errx(1, "name too long");
137 if ((infn = *argv++) == NULL)
139 if (strcmp(infn, "-") == 0) {
142 } else if ((infile = fopen(infn, "r")) == NULL)
149 sa.sa_handler = handlesig;
150 sigemptyset(&sa.sa_mask);
151 sigaddset(&sa.sa_mask, SIGHUP);
152 sigaddset(&sa.sa_mask, SIGINT);
153 sigaddset(&sa.sa_mask, SIGTERM);
154 sigaction(SIGHUP, &sa, NULL);
155 sigaction(SIGINT, &sa, NULL);
156 sigaction(SIGTERM, &sa, NULL);
164 /* Ensure 10^sufflen < LONG_MAX. */
165 for (maxfiles = 1, i = 0; i < sufflen; i++) {
166 if (maxfiles > LONG_MAX / 10)
167 errx(1, "%ld: suffix too long (limit %ld)",
172 /* Create files based on supplied patterns. */
173 while (nfiles < maxfiles - 1 && (expr = *argv++) != NULL) {
174 /* Look ahead & see if this pattern has any repetitions. */
175 if (*argv != NULL && **argv == '{') {
177 reps = strtol(*argv + 1, &ep, 10);
178 if (reps < 0 || *ep != '}' || errno != 0)
179 errx(1, "%s: bad repetition count", *argv + 1);
184 if (*expr == '/' || *expr == '%') {
187 while (reps-- != 0 && nfiles < maxfiles - 1);
188 } else if (isdigit((unsigned char)*expr))
191 errx(1, "%s: unrecognised pattern", expr);
194 /* Copy the rest into a new file. */
197 while ((p = get_line()) != NULL && fputs(p, ofp) == 0)
200 printf("%jd\n", (intmax_t)ftello(ofp));
201 if (fclose(ofp) != 0)
202 err(1, "%s", currfile);
216 "usage: csplit [-ks] [-f prefix] [-n number] file args ...\n");
221 handlesig(int sig __unused)
223 const char msg[] = "csplit: caught signal, cleaning up\n";
225 write(STDERR_FILENO, msg, sizeof(msg) - 1);
230 /* Create a new output file. */
236 if ((size_t)snprintf(currfile, sizeof(currfile), "%s%0*ld", prefix,
237 (int)sufflen, nfiles) >= sizeof(currfile))
238 errc(1, ENAMETOOLONG, NULL);
239 if ((fp = fopen(currfile, "w+")) == NULL)
240 err(1, "%s", currfile);
246 /* Remove partial output, called before exiting. */
250 char fnbuf[PATH_MAX];
257 * NOTE: One cannot portably assume to be able to call snprintf()
258 * from inside a signal handler. It does, however, appear to be safe
259 * to do on FreeBSD. The solution to this problem is worse than the
263 for (i = 0; i < nfiles; i++) {
264 snprintf(fnbuf, sizeof(fnbuf), "%s%0*ld", prefix,
270 /* Read a line from the input into a static buffer. */
274 static char lbuf[LINE_MAX];
277 src = overfile != NULL ? overfile : infile;
279 again: if (fgets(lbuf, sizeof(lbuf), src) == NULL) {
280 if (src == overfile) {
293 /* Conceptually rewind the input (as obtained by get_line()) back `n' lines. */
295 toomuch(FILE *ofp, long n)
300 if (overfile != NULL) {
302 * Truncate the previous file we overflowed into back to
303 * the correct length, close it.
305 if (fflush(overfile) != 0)
307 if (ftruncate(fileno(overfile), truncofs) != 0)
309 if (fclose(overfile) != 0)
315 /* Just tidying up */
321 * Wind the overflow file backwards to `n' lines before the
325 if (ftello(ofp) < (off_t)sizeof(buf))
328 fseeko(ofp, -(off_t)sizeof(buf), SEEK_CUR);
330 errx(1, "%s: can't seek", currfile);
331 if ((nread = fread(buf, 1, sizeof(buf), ofp)) == 0)
332 errx(1, "can't read overflowed output");
333 if (fseeko(ofp, -(off_t)nread, SEEK_CUR) != 0)
334 err(1, "%s", currfile);
335 for (i = 1; i <= nread; i++)
336 if (buf[nread - i] == '\n' && n-- == 0)
338 if (ftello(ofp) == 0)
341 if (fseeko(ofp, nread - i + 1, SEEK_CUR) != 0)
342 err(1, "%s", currfile);
345 * get_line() will read from here. Next call will truncate to
346 * truncofs in this file.
349 truncofs = ftello(overfile);
352 /* Handle splits for /regexp/ and %regexp% patterns. */
354 do_rexp(const char *expr)
360 char *ecopy, *ep, *p, *pofs, *re;
363 if ((ecopy = strdup(expr)) == NULL)
367 if ((pofs = strrchr(ecopy, *expr)) == NULL || pofs[-1] == '\\')
368 errx(1, "%s: missing trailing %c", expr, *expr);
373 ofs = strtol(pofs, &ep, 10);
374 if (*ep != '\0' || errno != 0)
375 errx(1, "%s: bad offset", pofs);
379 if (regcomp(&cre, re, REG_BASIC|REG_NOSUB) != 0)
380 errx(1, "%s: bad regular expression", re);
383 /* /regexp/: Save results to a file. */
386 /* %regexp%: Make a temporary file for overflow. */
387 if ((ofp = tmpfile()) == NULL)
391 /* Read and output lines until we get a match. */
393 while ((p = get_line()) != NULL) {
394 if (fputs(p, ofp) != 0)
396 if (!first && regexec(&cre, p, 0, NULL, 0) == 0)
402 errx(1, "%s: no match", re);
406 * Negative (or zero) offset: throw back any lines we should
410 toomuch(ofp, -ofs + 1);
411 nwritten = (intmax_t)truncofs;
413 nwritten = (intmax_t)ftello(ofp);
416 * Positive offset: copy the requested number of lines
419 while (--ofs > 0 && (p = get_line()) != NULL)
422 nwritten = (intmax_t)ftello(ofp);
423 if (fclose(ofp) != 0)
424 err(1, "%s", currfile);
427 if (!sflag && *expr == '/')
428 printf("%jd\n", nwritten);
434 /* Handle splits based on line number. */
436 do_lineno(const char *expr)
438 long lastline, tgtline;
443 tgtline = strtol(expr, &ep, 10);
444 if (tgtline <= 0 || errno != 0 || *ep != '\0')
445 errx(1, "%s: bad line number", expr);
447 if (lastline <= lineno)
448 errx(1, "%s: can't go backwards", expr);
450 while (nfiles < maxfiles - 1) {
452 while (lineno + 1 != lastline) {
453 if ((p = get_line()) == NULL)
454 errx(1, "%ld: out of range", lastline);
455 if (fputs(p, ofp) != 0)
459 printf("%jd\n", (intmax_t)ftello(ofp));
460 if (fclose(ofp) != 0)
461 err(1, "%s", currfile);