usr.bin/csplit/csplit.c

   1 /*      $NetBSD: csplit.c,v 1.4 2007/12/15 19:44:49 perry Exp $ */
   2 /*      $FreeBSD: src/usr.bin/csplit/csplit.c,v 1.9 2004/03/22 11:15:03 tjr Exp$        */
   3
   4 /*-
   5  * Copyright (c) 2002 Tim J. Robbins.
   6  * All rights reserved.
   7  *
   8  * Redistribution and use in source and binary forms, with or without
   9  * modification, are permitted provided that the following conditions
  10  * are met:
  11  * 1. Redistributions of source code must retain the above copyright
  12  *    notice, this list of conditions and the following disclaimer.
  13  * 2. Redistributions in binary form must reproduce the above copyright
  14  *    notice, this list of conditions and the following disclaimer in the
  15  *    documentation and/or other materials provided with the distribution.
  16  *
  17  * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
  18  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
  19  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
  20  * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
  21  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
  22  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
  23  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
  24  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
  25  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
  26  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
  27  * SUCH DAMAGE.
  28  */
  29
  30 /*
  31  * csplit -- split files based on context
  32  *
  33  * This utility splits its input into numbered output files by line number
  34  * or by a regular expression. Regular expression matches have an optional
  35  * offset with them, allowing the split to occur a specified number of
  36  * lines before or after the match.
  37  *
  38  * To handle negative offsets, we stop reading when the match occurs and
  39  * store the offset that the file should have been split at, then use
  40  * this output file as input until all the "overflowed" lines have been read.
  41  * The file is then closed and truncated to the correct length.
  42  *
  43  * We assume that the output files can be seeked upon (ie. they cannot be
  44  * symlinks to named pipes or character devices), but make no such
  45  * assumption about the input.
  46  */
  47
  48 #include <sys/cdefs.h>
  49 #ifndef lint
  50 __RCSID("$NetBSD: csplit.c,v 1.4 2007/12/15 19:44:49 perry Exp $");
  51 #endif
  52
  53 #include <sys/types.h>
  54
  55 #include <ctype.h>
  56 #include <err.h>
  57 #include <errno.h>
  58 #include <limits.h>
  59 #include <locale.h>
  60 #include <regex.h>
  61 #include <signal.h>
  62 #include <stdint.h>
  63 #include <stdio.h>
  64 #include <stdlib.h>
  65 #include <string.h>
  66 #include <unistd.h>
  67 #include <util.h>
  68
  69 static void      cleanup(void);
  70 static void      do_lineno(const char *);
  71 static void      do_rexp(const char *);
  72 static char     *get_line(void);
  73 static void      handlesig(int);
  74 static FILE     *newfile(void);
  75 static void      toomuch(FILE *, long);
  76 static void      usage(void) __dead;
  77
  78 /*
  79  * Command line options
  80  */
  81 const char *prefix;             /* File name prefix */
  82 long     sufflen;               /* Number of decimal digits for suffix */
  83 int      sflag;                 /* Suppress output of file names */
  84 int      kflag;                 /* Keep output if error occurs */
  85
  86 /*
  87  * Other miscellaneous globals (XXX too many)
  88  */
  89 long     lineno;                /* Current line number in input file */
  90 long     reps;                  /* Number of repetitions for this pattern */
  91 long     nfiles;                /* Number of files output so far */
  92 long     maxfiles;              /* Maximum number of files we can create */
  93 char     currfile[PATH_MAX];    /* Current output file */
  94 const char *infn;               /* Name of the input file */
  95 FILE    *infile;                /* Input file handle */
  96 FILE    *overfile;              /* Overflow file for toomuch() */
  97 off_t    truncofs;              /* Offset this file should be truncated at */
  98 int      doclean;               /* Should cleanup() remove output? */
  99
 100 int
 101 main(int argc, char *argv[])
 102 {
 103         struct sigaction sa;
 104         long i;
 105         int ch;
 106         const char *expr;
 107         char *ep, *p;
 108         FILE *ofp;
 109
 110         (void)setlocale(LC_ALL, "");
 111
 112         kflag = sflag = 0;
 113         prefix = "xx";
 114         sufflen = 2;
 115         while ((ch = getopt(argc, argv, "ksf:n:")) > 0) {
 116                 switch (ch) {
 117                 case 'f':
 118                         prefix = optarg;
 119                         break;
 120                 case 'k':
 121                         kflag = 1;
 122                         break;
 123                 case 'n':
 124                         errno = 0;
 125                         sufflen = strtol(optarg, &ep, 10);
 126                         if (sufflen <= 0 || *ep != '\0' || errno != 0)
 127                                 errx(1, "%s: bad suffix length", optarg);
 128                         break;
 129                 case 's':
 130                         sflag = 1;
 131                         break;
 132                 default:
 133                         usage();
 134                         /*NOTREACHED*/
 135                 }
 136         }
 137
 138         if (sufflen + strlen(prefix) >= PATH_MAX)
 139                 errx(1, "name too long");
 140
 141         argc -= optind;
 142         argv += optind;
 143
 144         if ((infn = *argv++) == NULL)
 145                 usage();
 146         if (strcmp(infn, "-") == 0) {
 147                 infile = stdin;
 148                 infn = "stdin";
 149         } else if ((infile = fopen(infn, "r")) == NULL)
 150                 err(1, "%s", infn);
 151
 152         if (!kflag) {
 153                 doclean = 1;
 154                 (void)atexit(cleanup);
 155                 sa.sa_flags = 0;
 156                 sa.sa_handler = handlesig;
 157                 (void)sigemptyset(&sa.sa_mask);
 158                 (void)sigaddset(&sa.sa_mask, SIGHUP);
 159                 (void)sigaddset(&sa.sa_mask, SIGINT);
 160                 (void)sigaddset(&sa.sa_mask, SIGTERM);
 161                 (void)sigaction(SIGHUP, &sa, NULL);
 162                 (void)sigaction(SIGINT, &sa, NULL);
 163                 (void)sigaction(SIGTERM, &sa, NULL);
 164         }
 165
 166         lineno = 0;
 167         nfiles = 0;
 168         truncofs = 0;
 169         overfile = NULL;
 170
 171         /* Ensure 10^sufflen < LONG_MAX. */
 172         for (maxfiles = 1, i = 0; i < sufflen; i++) {
 173                 if (maxfiles > LONG_MAX / 10)
 174                         errx(1, "%ld: suffix too long (limit %ld)",
 175                             sufflen, i);
 176                 maxfiles *= 10;
 177         }
 178
 179         /* Create files based on supplied patterns. */
 180         while (nfiles < maxfiles - 1 && (expr = *argv++) != NULL) {
 181                 /* Look ahead & see if this pattern has any repetitions. */
 182                 if (*argv != NULL && **argv == '{') {
 183                         errno = 0;
 184                         reps = strtol(*argv + 1, &ep, 10);
 185                         if (reps < 0 || *ep != '}' || errno != 0)
 186                                 errx(1, "%s: bad repetition count", *argv + 1);
 187                         argv++;
 188                 } else
 189                         reps = 0;
 190
 191                 if (*expr == '/' || *expr == '%') {
 192                         do
 193                                 do_rexp(expr);
 194                         while (reps-- != 0 && nfiles < maxfiles - 1);
 195                 } else if (isdigit((unsigned char)*expr))
 196                         do_lineno(expr);
 197                 else
 198                         errx(1, "%s: unrecognised pattern", expr);
 199         }
 200
 201         /* Copy the rest into a new file. */
 202         if (!feof(infile)) {
 203                 ofp = newfile();
 204                 while ((p = get_line()) != NULL && fputs(p, ofp) == 0)
 205                         ;
 206                 if (!sflag)
 207                         (void)printf("%jd\n", (intmax_t)ftello(ofp));
 208                 if (fclose(ofp) != 0)
 209                         err(1, "%s", currfile);
 210         }
 211
 212         toomuch(NULL, 0L);
 213         doclean = 0;
 214
 215         return (0);
 216 }
 217
 218 static void
 219 usage(void)
 220 {
 221
 222         (void)fprintf(stderr,
 223 "Usage: %s [-ks] [-f prefix] [-n number] file args ...\n", getprogname());
 224         exit(1);
 225 }
 226
 227 static void
 228 handlesig(int sig)
 229 {
 230         char msg[BUFSIZ];
 231         size_t len;
 232
 233         len = snprintf(msg, sizeof(msg), "%s: Caught %s, cleaning up\n",
 234             getprogname(), strsignal(sig));
 235         if (len < sizeof(msg))
 236                 (void)write(STDERR_FILENO, msg, len);
 237         cleanup();
 238         (void)raise_default_signal(sig);
 239         _exit(2);
 240 }
 241
 242 /* Create a new output file. */
 243 static FILE *
 244 newfile(void)
 245 {
 246         FILE *fp;
 247
 248         if ((size_t)snprintf(currfile, sizeof(currfile), "%s%0*ld", prefix,
 249             (int)sufflen, nfiles) >= sizeof(currfile))
 250                 errx(1, "%s: %s", currfile, strerror(ENAMETOOLONG));
 251         if ((fp = fopen(currfile, "w+")) == NULL)
 252                 err(1, "%s", currfile);
 253         nfiles++;
 254
 255         return (fp);
 256 }
 257
 258 /* Remove partial output, called before exiting. */
 259 static void
 260 cleanup(void)
 261 {
 262         char fnbuf[PATH_MAX];
 263         long i;
 264
 265         if (!doclean)
 266                 return;
 267
 268         /*
 269          * NOTE: One cannot portably assume to be able to call snprintf()
 270          * from inside a signal handler. It does, however, appear to be safe
 271          * to do on FreeBSD and NetBSD. The solution to this problem is worse
 272          * than the problem itself.
 273          */
 274
 275         for (i = 0; i < nfiles; i++) {
 276                 (void)snprintf(fnbuf, sizeof(fnbuf), "%s%0*ld", prefix,
 277                     (int)sufflen, i);
 278                 (void)unlink(fnbuf);
 279         }
 280 }
 281
 282 /* Read a line from the input into a static buffer. */
 283 static char *
 284 get_line(void)
 285 {
 286         static char lbuf[LINE_MAX];
 287         FILE *src;
 288
 289         src = overfile != NULL ? overfile : infile;
 290
 291 again: if (fgets(lbuf, sizeof(lbuf), src) == NULL) {
 292                 if (src == overfile) {
 293                         src = infile;
 294                         goto again;
 295                 }
 296                 return (NULL);
 297         }
 298         if (ferror(src))
 299                 err(1, "%s", infn);
 300         lineno++;
 301
 302         return (lbuf);
 303 }
 304
 305 /* Conceptually rewind the input (as obtained by get_line()) back `n' lines. */
 306 static void
 307 toomuch(FILE *ofp, long n)
 308 {
 309         char buf[BUFSIZ];
 310         size_t i, nread;
 311
 312         if (overfile != NULL) {
 313                 /*
 314                  * Truncate the previous file we overflowed into back to
 315                  * the correct length, close it.
 316                  */
 317                 if (fflush(overfile) != 0)
 318                         err(1, "overflow");
 319                 if (ftruncate(fileno(overfile), truncofs) != 0)
 320                         err(1, "overflow");
 321                 if (fclose(overfile) != 0)
 322                         err(1, "overflow");
 323                 overfile = NULL;
 324         }
 325
 326         if (n == 0)
 327                 /* Just tidying up */
 328                 return;
 329
 330         lineno -= n;
 331
 332         /*
 333          * Wind the overflow file backwards to `n' lines before the
 334          * current one.
 335          */
 336         do {
 337                 if (ftello(ofp) < (off_t)sizeof(buf))
 338                         rewind(ofp);
 339                 else
 340                         (void)fseeko(ofp, -(off_t)sizeof(buf), SEEK_CUR);
 341                 if (ferror(ofp))
 342                         errx(1, "%s: can't seek", currfile);
 343                 if ((nread = fread(buf, 1, sizeof(buf), ofp)) == 0)
 344                         errx(1, "can't read overflowed output");
 345                 if (fseeko(ofp, -(off_t)nread, SEEK_CUR) != 0)
 346                         err(1, "%s", currfile);
 347                 for (i = 1; i <= nread; i++)
 348                         if (buf[nread - i] == '\n' && n-- == 0)
 349                                 break;
 350                 if (ftello(ofp) == 0)
 351                         break;
 352         } while (n > 0);
 353         if (fseeko(ofp, (off_t)nread - i + 1, SEEK_CUR) != 0)
 354                 err(1, "%s", currfile);
 355
 356         /*
 357          * get_line() will read from here. Next call will truncate to
 358          * truncofs in this file.
 359          */
 360         overfile = ofp;
 361         truncofs = ftello(overfile);
 362 }
 363
 364 /* Handle splits for /regexp/ and %regexp% patterns. */
 365 static void
 366 do_rexp(const char *expr)
 367 {
 368         regex_t cre;
 369         intmax_t nwritten;
 370         long ofs;
 371         int first;
 372         char *ecopy, *ep, *p, *pofs, *re;
 373         FILE *ofp;
 374
 375         if ((ecopy = strdup(expr)) == NULL)
 376                 err(1, "strdup");
 377
 378         re = ecopy + 1;
 379         if ((pofs = strrchr(ecopy, *expr)) == NULL || pofs[-1] == '\\')
 380                 errx(1, "%s: missing trailing %c", expr, *expr);
 381         *pofs++ = '\0';
 382
 383         if (*pofs != '\0') {
 384                 errno = 0;
 385                 ofs = strtol(pofs, &ep, 10);
 386                 if (*ep != '\0' || errno != 0)
 387                         errx(1, "%s: bad offset", pofs);
 388         } else
 389                 ofs = 0;
 390
 391         if (regcomp(&cre, re, REG_BASIC|REG_NOSUB) != 0)
 392                 errx(1, "%s: bad regular expression", re);
 393
 394         if (*expr == '/')
 395                 /* /regexp/: Save results to a file. */
 396                 ofp = newfile();
 397         else {
 398                 /* %regexp%: Make a temporary file for overflow. */
 399                 if ((ofp = tmpfile()) == NULL)
 400                         err(1, "tmpfile");
 401         }
 402
 403         /* Read and output lines until we get a match. */
 404         first = 1;
 405         while ((p = get_line()) != NULL) {
 406                 if (fputs(p, ofp) != 0)
 407                         break;
 408                 if (!first && regexec(&cre, p, 0, NULL, 0) == 0)
 409                         break;
 410                 first = 0;
 411         }
 412
 413         if (p == NULL)
 414                 errx(1, "%s: no match", re);
 415
 416         if (ofs <= 0) {
 417                 /*
 418                  * Negative (or zero) offset: throw back any lines we should
 419                  * not have read yet.
 420                   */
 421                 if (p != NULL) {
 422                         toomuch(ofp, -ofs + 1);
 423                         nwritten = (intmax_t)truncofs;
 424                 } else
 425                         nwritten = (intmax_t)ftello(ofp);
 426         } else {
 427                 /*
 428                  * Positive offset: copy the requested number of lines
 429                  * after the match.
 430                  */
 431                 while (--ofs > 0 && (p = get_line()) != NULL)
 432                         if (fputs(p, ofp) != 0)
 433                                 break;
 434                 toomuch(NULL, 0L);
 435                 nwritten = (intmax_t)ftello(ofp);
 436                 if (fclose(ofp) != 0)
 437                         err(1, "%s", currfile);
 438         }
 439
 440         if (!sflag && *expr == '/')
 441                 (void)printf("%jd\n", nwritten);
 442
 443         regfree(&cre);
 444         free(ecopy);
 445 }
 446
 447 /* Handle splits based on line number. */
 448 static void
 449 do_lineno(const char *expr)
 450 {
 451         long lastline, tgtline;
 452         char *ep, *p;
 453         FILE *ofp;
 454
 455         errno = 0;
 456         tgtline = strtol(expr, &ep, 10);
 457         if (tgtline <= 0 || errno != 0 || *ep != '\0')
 458                 errx(1, "%s: bad line number", expr);
 459         lastline = tgtline;
 460         if (lastline <= lineno)
 461                 errx(1, "%s: can't go backwards", expr);
 462
 463         while (nfiles < maxfiles - 1) {
 464                 ofp = newfile();
 465                 while (lineno + 1 != lastline) {
 466                         if ((p = get_line()) == NULL)
 467                                 errx(1, "%ld: out of range", lastline);
 468                         if (fputs(p, ofp) != 0)
 469                                 break;
 470                 }
 471                 if (!sflag)
 472                         (void)printf("%jd\n", (intmax_t)ftello(ofp));
 473                 if (fclose(ofp) != 0)
 474                         err(1, "%s", currfile);
 475                 if (reps-- == 0)
 476                         break;
 477                 lastline += tgtline;
 478         }
 479 }