1 /* split.c -- split a file into pieces.
2 Copyright (C) 88, 91, 1995-2003 Free Software Foundation, Inc.
4 This program is free software; you can redistribute it and/or modify
5 it under the terms of the GNU General Public License as published by
6 the Free Software Foundation; either version 2, or (at your option)
9 This program is distributed in the hope that it will be useful,
10 but WITHOUT ANY WARRANTY; without even the implied warranty of
11 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
12 GNU General Public License for more details.
14 You should have received a copy of the GNU General Public License
15 along with this program; if not, write to the Free Software Foundation,
16 Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */
18 /* By tege@sics.se, with rms.
21 * Implement -t CHAR or -t REGEX to specify break characters other
28 #include <sys/types.h>
33 #include "full-read.h"
34 #include "full-write.h"
37 #include "safe-read.h"
40 /* The official name of this program (e.g., no `g' prefix). */
41 #define PROGRAM_NAME "split"
43 #define WRITTEN_BY _("Written by Torbjorn Granlund and Richard M. Stallman.")
45 #define DEFAULT_SUFFIX_LENGTH 2
47 /* The name this program was run with. */
50 /* Base name of output files. */
51 static char const *outbase;
53 /* Name of output files. */
56 /* Pointer to the end of the prefix in OUTFILE.
57 Suffixes are inserted here. */
58 static char *outfile_mid;
60 /* Length of OUTFILE's suffix. */
61 static size_t suffix_length = DEFAULT_SUFFIX_LENGTH;
63 /* Alphabet of characters to use in suffix. */
64 static char const *suffix_alphabet = "abcdefghijklmnopqrstuvwxyz";
66 /* Name of input file. May be "-". */
69 /* Descriptor on which input file is open. */
70 static int input_desc;
72 /* Descriptor on which output file is open. */
73 static int output_desc;
75 /* If nonzero, print a diagnostic on standard error just before each
76 output file is opened. */
79 static struct option const longopts[] =
81 {"bytes", required_argument, NULL, 'b'},
82 {"lines", required_argument, NULL, 'l'},
83 {"line-bytes", required_argument, NULL, 'C'},
84 {"suffix-length", required_argument, NULL, 'a'},
85 {"numeric-suffixes", no_argument, NULL, 'd'},
86 {"verbose", no_argument, &verbose, 1},
87 {GETOPT_HELP_OPTION_DECL},
88 {GETOPT_VERSION_OPTION_DECL},
96 fprintf (stderr, _("Try `%s --help' for more information.\n"),
101 Usage: %s [OPTION] [INPUT [PREFIX]]\n\
105 Output fixed-size pieces of INPUT to PREFIXaa, PREFIXab, ...; default\n\
106 PREFIX is `x'. With no INPUT, or when INPUT is -, read standard input.\n\
110 Mandatory arguments to long options are mandatory for short options too.\n\
112 fprintf (stdout, _("\
113 -a, --suffix-length=N use suffixes of length N (default %d)\n\
114 -b, --bytes=SIZE put SIZE bytes per output file\n\
115 -C, --line-bytes=SIZE put at most SIZE bytes of lines per output file\n\
116 -d, --numeric-suffixes use numeric suffixes instead of alphabetic\n\
117 -l, --lines=NUMBER put NUMBER lines per output file\n\
118 "), DEFAULT_SUFFIX_LENGTH);
120 --verbose print a diagnostic to standard error just\n\
121 before each output file is opened\n\
123 fputs (HELP_OPTION_DESCRIPTION, stdout);
124 fputs (VERSION_OPTION_DESCRIPTION, stdout);
127 SIZE may have a multiplier suffix: b for 512, k for 1K, m for 1 Meg.\n\
129 printf (_("\nReport bugs to <%s>.\n"), PACKAGE_BUGREPORT);
131 exit (status == 0 ? EXIT_SUCCESS : EXIT_FAILURE);
134 /* Compute the next sequential output file name and store it into the
138 next_file_name (void)
140 /* Index in suffix_alphabet of each character in the suffix. */
141 static size_t *sufindex;
145 /* Allocate and initialize the first file name. */
147 size_t outbase_length = strlen (outbase);
148 size_t outfile_length = outbase_length + suffix_length;
149 if (outfile_length + 1 < outbase_length)
151 outfile = xmalloc (outfile_length + 1);
152 outfile_mid = outfile + outbase_length;
153 memcpy (outfile, outbase, outbase_length);
154 memset (outfile_mid, suffix_alphabet[0], suffix_length);
155 outfile[outfile_length] = 0;
156 sufindex = xcalloc (suffix_length, sizeof (size_t));
158 #if ! _POSIX_NO_TRUNC && HAVE_PATHCONF && defined _PC_NAME_MAX
159 /* POSIX requires that if the output file name is too long for
160 its directory, `split' must fail without creating any files.
161 This must be checked for explicitly on operating systems that
162 silently truncate file names. */
164 char *dir = dir_name (outfile);
165 long name_max = pathconf (dir, _PC_NAME_MAX);
166 if (0 <= name_max && name_max < base_len (base_name (outfile)))
167 error (EXIT_FAILURE, ENAMETOOLONG, "%s", outfile);
174 /* Increment the suffix in place, if possible. */
176 size_t i = suffix_length;
180 outfile_mid[i] = suffix_alphabet[sufindex[i]];
184 outfile_mid[i] = suffix_alphabet[sufindex[i]];
186 error (EXIT_FAILURE, 0, _("Output file suffixes exhausted"));
190 /* Write BYTES bytes at BP to an output file.
191 If NEW_FILE_FLAG is nonzero, open the next output file.
192 Otherwise add to the same output file already in use. */
195 cwrite (int new_file_flag, const char *bp, size_t bytes)
199 if (output_desc >= 0 && close (output_desc) < 0)
200 error (EXIT_FAILURE, errno, "%s", outfile);
204 fprintf (stderr, _("creating file `%s'\n"), outfile);
205 output_desc = open (outfile,
206 O_WRONLY | O_CREAT | O_TRUNC | O_BINARY, 0666);
208 error (EXIT_FAILURE, errno, "%s", outfile);
210 if (full_write (output_desc, bp, bytes) != bytes)
211 error (EXIT_FAILURE, errno, "%s", outfile);
214 /* Split into pieces of exactly N_BYTES bytes.
215 Use buffer BUF, whose size is BUFSIZE. */
218 bytes_split (uintmax_t n_bytes, char *buf, size_t bufsize)
221 int new_file_flag = 1;
223 uintmax_t to_write = n_bytes;
228 n_read = full_read (input_desc, buf, bufsize);
229 if (n_read == SAFE_READ_ERROR)
230 error (EXIT_FAILURE, errno, "%s", infile);
235 if (to_read < to_write)
237 if (to_read) /* do not write 0 bytes! */
239 cwrite (new_file_flag, bp_out, to_read);
248 cwrite (new_file_flag, bp_out, w);
256 while (n_read == bufsize);
259 /* Split into pieces of exactly N_LINES lines.
260 Use buffer BUF, whose size is BUFSIZE. */
263 lines_split (uintmax_t n_lines, char *buf, size_t bufsize)
266 char *bp, *bp_out, *eob;
267 int new_file_flag = 1;
272 n_read = full_read (input_desc, buf, bufsize);
273 if (n_read == SAFE_READ_ERROR)
274 error (EXIT_FAILURE, errno, "%s", infile);
280 bp = memchr (bp, '\n', eob - bp + 1);
283 if (eob != bp_out) /* do not write 0 bytes! */
285 size_t len = eob - bp_out;
286 cwrite (new_file_flag, bp_out, len);
295 cwrite (new_file_flag, bp_out, bp - bp_out);
302 while (n_read == bufsize);
305 /* Split into pieces that are as large as possible while still not more
306 than N_BYTES bytes, and are split on line boundaries except
307 where lines longer than N_BYTES bytes occur.
308 FIXME: Allow N_BYTES to be any uintmax_t value, and don't require a
309 buffer of size N_BYTES, in case N_BYTES is very large. */
312 line_bytes_split (size_t n_bytes)
317 size_t n_buffered = 0;
318 char *buf = xmalloc (n_bytes);
322 /* Fill up the full buffer size from the input file. */
324 n_read = full_read (input_desc, buf + n_buffered, n_bytes - n_buffered);
325 if (n_read == SAFE_READ_ERROR)
326 error (EXIT_FAILURE, errno, "%s", infile);
328 n_buffered += n_read;
329 if (n_buffered != n_bytes)
332 /* Find where to end this chunk. */
333 bp = buf + n_buffered;
334 if (n_buffered == n_bytes)
336 while (bp > buf && bp[-1] != '\n')
340 /* If chunk has no newlines, use all the chunk. */
342 bp = buf + n_buffered;
344 /* Output the chars as one output file. */
345 cwrite (1, buf, bp - buf);
347 /* Discard the chars we just output; move rest of chunk
348 down to be the start of the next chunk. Source and
349 destination probably overlap. */
350 n_buffered -= bp - buf;
352 memmove (buf, bp, n_buffered);
358 #define FAIL_ONLY_ONE_WAY() \
361 error (0, 0, _("cannot split in more than one way")); \
362 usage (EXIT_FAILURE); \
367 main (int argc, char **argv)
369 struct stat stat_buf;
372 type_undef, type_bytes, type_byteslines, type_lines, type_digits
373 } split_type = type_undef;
374 size_t in_blk_size; /* optimal block size of input file device */
375 char *buf; /* file i/o buffer */
378 int digits_optind = 0;
380 initialize_main (&argc, &argv);
381 program_name = argv[0];
382 setlocale (LC_ALL, "");
383 bindtextdomain (PACKAGE, LOCALEDIR);
384 textdomain (PACKAGE);
386 atexit (close_stdout);
388 /* Parse command line options. */
395 /* This is the argv-index of the option we will read next. */
396 int this_optind = optind ? optind : 1;
398 c = getopt_long (argc, argv, "0123456789C:a:b:dl:", longopts, NULL);
410 if (xstrtoul (optarg, NULL, 10, &tmp, "") != LONGINT_OK
411 || SIZE_MAX / sizeof (size_t) < tmp)
413 error (0, 0, _("%s: invalid suffix length"), optarg);
414 usage (EXIT_FAILURE);
421 if (split_type != type_undef)
422 FAIL_ONLY_ONE_WAY ();
423 split_type = type_bytes;
424 if (xstrtoumax (optarg, NULL, 10, &n_units, "bkm") != LONGINT_OK
427 error (0, 0, _("%s: invalid number of bytes"), optarg);
428 usage (EXIT_FAILURE);
433 if (split_type != type_undef)
434 FAIL_ONLY_ONE_WAY ();
435 split_type = type_lines;
436 if (xstrtoumax (optarg, NULL, 10, &n_units, "") != LONGINT_OK
439 error (0, 0, _("%s: invalid number of lines"), optarg);
440 usage (EXIT_FAILURE);
445 if (split_type != type_undef)
446 FAIL_ONLY_ONE_WAY ();
447 split_type = type_byteslines;
448 if (xstrtoumax (optarg, NULL, 10, &n_units, "bkm") != LONGINT_OK
449 || n_units == 0 || SIZE_MAX < n_units)
451 error (0, 0, _("%s: invalid number of bytes"), optarg);
452 usage (EXIT_FAILURE);
466 if (split_type == type_undef)
468 split_type = type_digits;
471 if (split_type != type_undef && split_type != type_digits)
472 FAIL_ONLY_ONE_WAY ();
473 if (digits_optind != 0 && digits_optind != this_optind)
474 n_units = 0; /* More than one number given; ignore other. */
475 digits_optind = this_optind;
476 if (UINTMAX_MAX / 10 < n_units
477 || n_units * 10 + c - '0' < n_units * 10)
479 char buffer[INT_BUFSIZE_BOUND (uintmax_t)];
480 error (EXIT_FAILURE, 0,
481 _("line count option -%s%c... is too large"),
482 umaxtostr (n_units, buffer), c);
484 n_units = n_units * 10 + c - '0';
488 suffix_alphabet = "0123456789";
491 case_GETOPT_HELP_CHAR;
493 case_GETOPT_VERSION_CHAR (PROGRAM_NAME, WRITTEN_BY);
496 usage (EXIT_FAILURE);
500 if (digits_optind && 200112 <= posix2_version ())
502 char buffer[INT_BUFSIZE_BOUND (uintmax_t)];
503 char const *a = umaxtostr (n_units, buffer);
504 error (0, 0, _("`-%s' option is obsolete; use `-l %s'"), a, a);
505 usage (EXIT_FAILURE);
508 /* Handle default case. */
509 if (split_type == type_undef)
511 split_type = type_lines;
517 /* FIXME: be sure to remove this block when removing
518 support for obsolete options like `-10'. */
519 error (0, 0, _("invalid number of lines: 0"));
520 usage (EXIT_FAILURE);
523 /* Get out the filename arguments. */
526 infile = argv[optind++];
529 outbase = argv[optind++];
533 error (0, 0, _("too many arguments"));
534 usage (EXIT_FAILURE);
537 /* Open the input file. */
538 if (STREQ (infile, "-"))
539 input_desc = STDIN_FILENO;
542 input_desc = open (infile, O_RDONLY);
544 error (EXIT_FAILURE, errno, "%s", infile);
546 /* Binary I/O is safer when bytecounts are used. */
547 SET_BINARY (input_desc);
549 /* No output file is open now. */
552 /* Get the optimal block size of input device and make a buffer. */
554 if (fstat (input_desc, &stat_buf) < 0)
555 error (EXIT_FAILURE, errno, "%s", infile);
556 in_blk_size = ST_BLKSIZE (stat_buf);
558 buf = xmalloc (in_blk_size + 1);
564 lines_split (n_units, buf, in_blk_size);
568 bytes_split (n_units, buf, in_blk_size);
571 case type_byteslines:
572 line_bytes_split (n_units);
579 if (close (input_desc) < 0)
580 error (EXIT_FAILURE, errno, "%s", infile);
581 if (output_desc >= 0 && close (output_desc) < 0)
582 error (EXIT_FAILURE, errno, "%s", outfile);