1 /* split.c -- split a file into pieces.
2 Copyright (C) 88, 91, 1995-2004 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 "getpagesize.h"
34 #include "full-read.h"
35 #include "full-write.h"
39 #include "safe-read.h"
42 /* The official name of this program (e.g., no `g' prefix). */
43 #define PROGRAM_NAME "split"
45 #define AUTHORS "Torbjorn Granlund", "Richard M. Stallman"
47 #define DEFAULT_SUFFIX_LENGTH 2
49 /* The name this program was run with. */
52 /* Base name of output files. */
53 static char const *outbase;
55 /* Name of output files. */
58 /* Pointer to the end of the prefix in OUTFILE.
59 Suffixes are inserted here. */
60 static char *outfile_mid;
62 /* Length of OUTFILE's suffix. */
63 static size_t suffix_length = DEFAULT_SUFFIX_LENGTH;
65 /* Alphabet of characters to use in suffix. */
66 static char const *suffix_alphabet = "abcdefghijklmnopqrstuvwxyz";
68 /* Name of input file. May be "-". */
71 /* Descriptor on which input file is open. */
72 static int input_desc;
74 /* Descriptor on which output file is open. */
75 static int output_desc;
77 /* If nonzero, print a diagnostic on standard error just before each
78 output file is opened. */
81 static struct option const longopts[] =
83 {"bytes", required_argument, NULL, 'b'},
84 {"lines", required_argument, NULL, 'l'},
85 {"line-bytes", required_argument, NULL, 'C'},
86 {"suffix-length", required_argument, NULL, 'a'},
87 {"numeric-suffixes", no_argument, NULL, 'd'},
88 {"verbose", no_argument, &verbose, 1},
89 {GETOPT_HELP_OPTION_DECL},
90 {GETOPT_VERSION_OPTION_DECL},
97 if (status != EXIT_SUCCESS)
98 fprintf (stderr, _("Try `%s --help' for more information.\n"),
103 Usage: %s [OPTION] [INPUT [PREFIX]]\n\
107 Output fixed-size pieces of INPUT to PREFIXaa, PREFIXab, ...; default\n\
108 PREFIX is `x'. With no INPUT, or when INPUT is -, read standard input.\n\
112 Mandatory arguments to long options are mandatory for short options too.\n\
114 fprintf (stdout, _("\
115 -a, --suffix-length=N use suffixes of length N (default %d)\n\
116 -b, --bytes=SIZE put SIZE bytes per output file\n\
117 -C, --line-bytes=SIZE put at most SIZE bytes of lines per output file\n\
118 -d, --numeric-suffixes use numeric suffixes instead of alphabetic\n\
119 -l, --lines=NUMBER put NUMBER lines per output file\n\
120 "), DEFAULT_SUFFIX_LENGTH);
122 --verbose print a diagnostic to standard error just\n\
123 before each output file is opened\n\
125 fputs (HELP_OPTION_DESCRIPTION, stdout);
126 fputs (VERSION_OPTION_DESCRIPTION, stdout);
129 SIZE may have a multiplier suffix: b for 512, k for 1K, m for 1 Meg.\n\
131 printf (_("\nReport bugs to <%s>.\n"), PACKAGE_BUGREPORT);
136 /* Compute the next sequential output file name and store it into the
140 next_file_name (void)
142 /* Index in suffix_alphabet of each character in the suffix. */
143 static size_t *sufindex;
147 /* Allocate and initialize the first file name. */
149 size_t outbase_length = strlen (outbase);
150 size_t outfile_length = outbase_length + suffix_length;
151 if (outfile_length + 1 < outbase_length)
153 outfile = xmalloc (outfile_length + 1);
154 outfile_mid = outfile + outbase_length;
155 memcpy (outfile, outbase, outbase_length);
156 memset (outfile_mid, suffix_alphabet[0], suffix_length);
157 outfile[outfile_length] = 0;
158 sufindex = xcalloc (suffix_length, sizeof *sufindex);
160 #if ! _POSIX_NO_TRUNC && HAVE_PATHCONF && defined _PC_NAME_MAX
161 /* POSIX requires that if the output file name is too long for
162 its directory, `split' must fail without creating any files.
163 This must be checked for explicitly on operating systems that
164 silently truncate file names. */
166 char *dir = dir_name (outfile);
167 long name_max = pathconf (dir, _PC_NAME_MAX);
168 if (0 <= name_max && name_max < base_len (base_name (outfile)))
169 error (EXIT_FAILURE, ENAMETOOLONG, "%s", outfile);
176 /* Increment the suffix in place, if possible. */
178 size_t i = suffix_length;
182 outfile_mid[i] = suffix_alphabet[sufindex[i]];
186 outfile_mid[i] = suffix_alphabet[sufindex[i]];
188 error (EXIT_FAILURE, 0, _("Output file suffixes exhausted"));
192 /* Write BYTES bytes at BP to an output file.
193 If NEW_FILE_FLAG is true, open the next output file.
194 Otherwise add to the same output file already in use. */
197 cwrite (bool new_file_flag, const char *bp, size_t bytes)
201 if (output_desc >= 0 && close (output_desc) < 0)
202 error (EXIT_FAILURE, errno, "%s", outfile);
206 fprintf (stderr, _("creating file `%s'\n"), outfile);
207 output_desc = open (outfile,
208 O_WRONLY | O_CREAT | O_TRUNC | O_BINARY, 0666);
210 error (EXIT_FAILURE, errno, "%s", outfile);
212 if (full_write (output_desc, bp, bytes) != bytes)
213 error (EXIT_FAILURE, errno, "%s", outfile);
216 /* Split into pieces of exactly N_BYTES bytes.
217 Use buffer BUF, whose size is BUFSIZE. */
220 bytes_split (uintmax_t n_bytes, char *buf, size_t bufsize)
223 bool new_file_flag = true;
225 uintmax_t to_write = n_bytes;
230 n_read = full_read (input_desc, buf, bufsize);
231 if (n_read == SAFE_READ_ERROR)
232 error (EXIT_FAILURE, errno, "%s", infile);
237 if (to_read < to_write)
239 if (to_read) /* do not write 0 bytes! */
241 cwrite (new_file_flag, bp_out, to_read);
243 new_file_flag = false;
250 cwrite (new_file_flag, bp_out, w);
253 new_file_flag = true;
258 while (n_read == bufsize);
261 /* Split into pieces of exactly N_LINES lines.
262 Use buffer BUF, whose size is BUFSIZE. */
265 lines_split (uintmax_t n_lines, char *buf, size_t bufsize)
268 char *bp, *bp_out, *eob;
269 bool new_file_flag = true;
274 n_read = full_read (input_desc, buf, bufsize);
275 if (n_read == SAFE_READ_ERROR)
276 error (EXIT_FAILURE, errno, "%s", infile);
282 bp = memchr (bp, '\n', eob - bp + 1);
285 if (eob != bp_out) /* do not write 0 bytes! */
287 size_t len = eob - bp_out;
288 cwrite (new_file_flag, bp_out, len);
289 new_file_flag = false;
297 cwrite (new_file_flag, bp_out, bp - bp_out);
299 new_file_flag = true;
304 while (n_read == bufsize);
307 /* Split into pieces that are as large as possible while still not more
308 than N_BYTES bytes, and are split on line boundaries except
309 where lines longer than N_BYTES bytes occur.
310 FIXME: Allow N_BYTES to be any uintmax_t value, and don't require a
311 buffer of size N_BYTES, in case N_BYTES is very large. */
314 line_bytes_split (size_t n_bytes)
319 size_t n_buffered = 0;
320 char *buf = xmalloc (n_bytes);
324 /* Fill up the full buffer size from the input file. */
326 n_read = full_read (input_desc, buf + n_buffered, n_bytes - n_buffered);
327 if (n_read == SAFE_READ_ERROR)
328 error (EXIT_FAILURE, errno, "%s", infile);
330 n_buffered += n_read;
331 if (n_buffered != n_bytes)
334 /* Find where to end this chunk. */
335 bp = buf + n_buffered;
336 if (n_buffered == n_bytes)
338 while (bp > buf && bp[-1] != '\n')
342 /* If chunk has no newlines, use all the chunk. */
344 bp = buf + n_buffered;
346 /* Output the chars as one output file. */
347 cwrite (true, buf, bp - buf);
349 /* Discard the chars we just output; move rest of chunk
350 down to be the start of the next chunk. Source and
351 destination probably overlap. */
352 n_buffered -= bp - buf;
354 memmove (buf, bp, n_buffered);
360 #define FAIL_ONLY_ONE_WAY() \
363 error (0, 0, _("cannot split in more than one way")); \
364 usage (EXIT_FAILURE); \
369 main (int argc, char **argv)
371 struct stat stat_buf;
374 type_undef, type_bytes, type_byteslines, type_lines, type_digits
375 } split_type = type_undef;
376 size_t in_blk_size; /* optimal block size of input file device */
377 char *buf; /* file i/o buffer */
378 size_t page_size = getpagesize ();
381 int digits_optind = 0;
383 initialize_main (&argc, &argv);
384 program_name = argv[0];
385 setlocale (LC_ALL, "");
386 bindtextdomain (PACKAGE, LOCALEDIR);
387 textdomain (PACKAGE);
389 atexit (close_stdout);
391 /* Parse command line options. */
398 /* This is the argv-index of the option we will read next. */
399 int this_optind = optind ? optind : 1;
401 c = getopt_long (argc, argv, "0123456789C:a:b:dl:", longopts, NULL);
413 if (xstrtoul (optarg, NULL, 10, &tmp, "") != LONGINT_OK
414 || SIZE_MAX / sizeof (size_t) < tmp)
416 error (0, 0, _("%s: invalid suffix length"), optarg);
417 usage (EXIT_FAILURE);
424 if (split_type != type_undef)
425 FAIL_ONLY_ONE_WAY ();
426 split_type = type_bytes;
427 if (xstrtoumax (optarg, NULL, 10, &n_units, "bkm") != LONGINT_OK
430 error (0, 0, _("%s: invalid number of bytes"), optarg);
431 usage (EXIT_FAILURE);
436 if (split_type != type_undef)
437 FAIL_ONLY_ONE_WAY ();
438 split_type = type_lines;
439 if (xstrtoumax (optarg, NULL, 10, &n_units, "") != LONGINT_OK
442 error (0, 0, _("%s: invalid number of lines"), optarg);
443 usage (EXIT_FAILURE);
448 if (split_type != type_undef)
449 FAIL_ONLY_ONE_WAY ();
450 split_type = type_byteslines;
451 if (xstrtoumax (optarg, NULL, 10, &n_units, "bkm") != LONGINT_OK
452 || n_units == 0 || SIZE_MAX < n_units)
454 error (0, 0, _("%s: invalid number of bytes"), optarg);
455 usage (EXIT_FAILURE);
469 if (split_type == type_undef)
471 split_type = type_digits;
474 if (split_type != type_undef && split_type != type_digits)
475 FAIL_ONLY_ONE_WAY ();
476 if (digits_optind != 0 && digits_optind != this_optind)
477 n_units = 0; /* More than one number given; ignore other. */
478 digits_optind = this_optind;
479 if (UINTMAX_MAX / 10 < n_units
480 || n_units * 10 + c - '0' < n_units * 10)
482 char buffer[INT_BUFSIZE_BOUND (uintmax_t)];
483 error (EXIT_FAILURE, 0,
484 _("line count option -%s%c... is too large"),
485 umaxtostr (n_units, buffer), c);
487 n_units = n_units * 10 + c - '0';
491 suffix_alphabet = "0123456789";
494 case_GETOPT_HELP_CHAR;
496 case_GETOPT_VERSION_CHAR (PROGRAM_NAME, AUTHORS);
499 usage (EXIT_FAILURE);
503 if (digits_optind && 200112 <= posix2_version ())
505 char buffer[INT_BUFSIZE_BOUND (uintmax_t)];
506 char const *a = umaxtostr (n_units, buffer);
507 error (0, 0, _("`-%s' option is obsolete; use `-l %s'"), a, a);
508 usage (EXIT_FAILURE);
511 /* Handle default case. */
512 if (split_type == type_undef)
514 split_type = type_lines;
520 /* FIXME: be sure to remove this block when removing
521 support for obsolete options like `-10'. */
522 error (0, 0, _("invalid number of lines: 0"));
523 usage (EXIT_FAILURE);
526 /* Get out the filename arguments. */
529 infile = argv[optind++];
532 outbase = argv[optind++];
536 error (0, 0, _("extra operand %s"), quote (argv[optind]));
537 usage (EXIT_FAILURE);
540 /* Open the input file. */
541 if (STREQ (infile, "-"))
542 input_desc = STDIN_FILENO;
545 input_desc = open (infile, O_RDONLY);
547 error (EXIT_FAILURE, errno, "%s", infile);
549 /* Binary I/O is safer when bytecounts are used. */
550 SET_BINARY (input_desc);
552 /* No output file is open now. */
555 /* Get the optimal block size of input device and make a buffer. */
557 if (fstat (input_desc, &stat_buf) < 0)
558 error (EXIT_FAILURE, errno, "%s", infile);
559 in_blk_size = ST_BLKSIZE (stat_buf);
561 buf = ptr_align (xmalloc (in_blk_size + 1 + page_size - 1), page_size);
567 lines_split (n_units, buf, in_blk_size);
571 bytes_split (n_units, buf, in_blk_size);
574 case type_byteslines:
575 line_bytes_split (n_units);
582 if (close (input_desc) < 0)
583 error (EXIT_FAILURE, errno, "%s", infile);
584 if (output_desc >= 0 && close (output_desc) < 0)
585 error (EXIT_FAILURE, errno, "%s", outfile);