1 /* split.c -- split a file into pieces.
2 Copyright (C) 1988, 1991, 1995-2006 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 3 of the License, or
7 (at your option) any later version.
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, see <http://www.gnu.org/licenses/>. */
17 /* By tege@sics.se, with rms.
20 * Implement -t CHAR or -t REGEX to specify break characters other
27 #include <sys/types.h>
31 #include "fd-reopen.h"
33 #include "getpagesize.h"
34 #include "full-read.h"
35 #include "full-write.h"
38 #include "safe-read.h"
41 /* The official name of this program (e.g., no `g' prefix). */
42 #define PROGRAM_NAME "split"
44 #define AUTHORS "Torbjorn Granlund", "Richard M. Stallman"
46 #define DEFAULT_SUFFIX_LENGTH 2
48 /* The name this program was run with. */
51 /* Base name of output files. */
52 static char const *outbase;
54 /* Name of output files. */
57 /* Pointer to the end of the prefix in OUTFILE.
58 Suffixes are inserted here. */
59 static char *outfile_mid;
61 /* Length of OUTFILE's suffix. */
62 static size_t suffix_length = DEFAULT_SUFFIX_LENGTH;
64 /* Alphabet of characters to use in suffix. */
65 static char const *suffix_alphabet = "abcdefghijklmnopqrstuvwxyz";
67 /* Name of input file. May be "-". */
70 /* Descriptor on which output file is open. */
71 static int output_desc;
73 /* If true, print a diagnostic on standard error just before each
74 output file is opened. */
77 /* For long options that have no equivalent short option, use a
78 non-character as a pseudo short option, starting with CHAR_MAX + 1. */
81 VERBOSE_OPTION = CHAR_MAX + 1
84 static struct option const longopts[] =
86 {"bytes", required_argument, NULL, 'b'},
87 {"lines", required_argument, NULL, 'l'},
88 {"line-bytes", required_argument, NULL, 'C'},
89 {"suffix-length", required_argument, NULL, 'a'},
90 {"numeric-suffixes", no_argument, NULL, 'd'},
91 {"verbose", no_argument, NULL, VERBOSE_OPTION},
92 {GETOPT_HELP_OPTION_DECL},
93 {GETOPT_VERSION_OPTION_DECL},
100 if (status != EXIT_SUCCESS)
101 fprintf (stderr, _("Try `%s --help' for more information.\n"),
106 Usage: %s [OPTION] [INPUT [PREFIX]]\n\
110 Output fixed-size pieces of INPUT to PREFIXaa, PREFIXab, ...; default\n\
111 size is 1000 lines, and default PREFIX is `x'. With no INPUT, or when INPUT\n\
112 is -, read standard input.\n\
116 Mandatory arguments to long options are mandatory for short options too.\n\
118 fprintf (stdout, _("\
119 -a, --suffix-length=N use suffixes of length N (default %d)\n\
120 -b, --bytes=SIZE put SIZE bytes per output file\n\
121 -C, --line-bytes=SIZE put at most SIZE bytes of lines per output file\n\
122 -d, --numeric-suffixes use numeric suffixes instead of alphabetic\n\
123 -l, --lines=NUMBER put NUMBER lines per output file\n\
124 "), DEFAULT_SUFFIX_LENGTH);
126 --verbose print a diagnostic to standard error just\n\
127 before each output file is opened\n\
129 fputs (HELP_OPTION_DESCRIPTION, stdout);
130 fputs (VERSION_OPTION_DESCRIPTION, stdout);
133 SIZE may have a multiplier suffix:\n\
134 b 512, kB 1000, K 1024, MB 1000*1000, M 1024*1024,\n\
135 GB 1000*1000*1000, G 1024*1024*1024, and so on for T, P, E, Z, Y.\n\
137 emit_bug_reporting_address ();
142 /* Compute the next sequential output file name and store it into the
146 next_file_name (void)
148 /* Index in suffix_alphabet of each character in the suffix. */
149 static size_t *sufindex;
153 /* Allocate and initialize the first file name. */
155 size_t outbase_length = strlen (outbase);
156 size_t outfile_length = outbase_length + suffix_length;
157 if (outfile_length + 1 < outbase_length)
159 outfile = xmalloc (outfile_length + 1);
160 outfile_mid = outfile + outbase_length;
161 memcpy (outfile, outbase, outbase_length);
162 memset (outfile_mid, suffix_alphabet[0], suffix_length);
163 outfile[outfile_length] = 0;
164 sufindex = xcalloc (suffix_length, sizeof *sufindex);
166 #if ! _POSIX_NO_TRUNC && HAVE_PATHCONF && defined _PC_NAME_MAX
167 /* POSIX requires that if the output file name is too long for
168 its directory, `split' must fail without creating any files.
169 This must be checked for explicitly on operating systems that
170 silently truncate file names. */
172 char *dir = dir_name (outfile);
173 long name_max = pathconf (dir, _PC_NAME_MAX);
174 if (0 <= name_max && name_max < base_len (last_component (outfile)))
175 error (EXIT_FAILURE, ENAMETOOLONG, "%s", outfile);
182 /* Increment the suffix in place, if possible. */
184 size_t i = suffix_length;
188 outfile_mid[i] = suffix_alphabet[sufindex[i]];
192 outfile_mid[i] = suffix_alphabet[sufindex[i]];
194 error (EXIT_FAILURE, 0, _("Output file suffixes exhausted"));
198 /* Write BYTES bytes at BP to an output file.
199 If NEW_FILE_FLAG is true, open the next output file.
200 Otherwise add to the same output file already in use. */
203 cwrite (bool new_file_flag, const char *bp, size_t bytes)
207 if (output_desc >= 0 && close (output_desc) < 0)
208 error (EXIT_FAILURE, errno, "%s", outfile);
212 fprintf (stderr, _("creating file %s\n"), quote (outfile));
213 output_desc = open (outfile,
214 O_WRONLY | O_CREAT | O_TRUNC | O_BINARY,
215 (S_IRUSR | S_IWUSR | S_IRGRP | S_IWGRP
216 | S_IROTH | S_IWOTH));
218 error (EXIT_FAILURE, errno, "%s", outfile);
220 if (full_write (output_desc, bp, bytes) != bytes)
221 error (EXIT_FAILURE, errno, "%s", outfile);
224 /* Split into pieces of exactly N_BYTES bytes.
225 Use buffer BUF, whose size is BUFSIZE. */
228 bytes_split (uintmax_t n_bytes, char *buf, size_t bufsize)
231 bool new_file_flag = true;
233 uintmax_t to_write = n_bytes;
238 n_read = full_read (STDIN_FILENO, buf, bufsize);
239 if (n_read == SAFE_READ_ERROR)
240 error (EXIT_FAILURE, errno, "%s", infile);
245 if (to_read < to_write)
247 if (to_read) /* do not write 0 bytes! */
249 cwrite (new_file_flag, bp_out, to_read);
251 new_file_flag = false;
258 cwrite (new_file_flag, bp_out, w);
261 new_file_flag = true;
266 while (n_read == bufsize);
269 /* Split into pieces of exactly N_LINES lines.
270 Use buffer BUF, whose size is BUFSIZE. */
273 lines_split (uintmax_t n_lines, char *buf, size_t bufsize)
276 char *bp, *bp_out, *eob;
277 bool new_file_flag = true;
282 n_read = full_read (STDIN_FILENO, buf, bufsize);
283 if (n_read == SAFE_READ_ERROR)
284 error (EXIT_FAILURE, errno, "%s", infile);
290 bp = memchr (bp, '\n', eob - bp + 1);
293 if (eob != bp_out) /* do not write 0 bytes! */
295 size_t len = eob - bp_out;
296 cwrite (new_file_flag, bp_out, len);
297 new_file_flag = false;
305 cwrite (new_file_flag, bp_out, bp - bp_out);
307 new_file_flag = true;
312 while (n_read == bufsize);
315 /* Split into pieces that are as large as possible while still not more
316 than N_BYTES bytes, and are split on line boundaries except
317 where lines longer than N_BYTES bytes occur.
318 FIXME: Allow N_BYTES to be any uintmax_t value, and don't require a
319 buffer of size N_BYTES, in case N_BYTES is very large. */
322 line_bytes_split (size_t n_bytes)
327 size_t n_buffered = 0;
328 char *buf = xmalloc (n_bytes);
332 /* Fill up the full buffer size from the input file. */
334 n_read = full_read (STDIN_FILENO, buf + n_buffered, n_bytes - n_buffered);
335 if (n_read == SAFE_READ_ERROR)
336 error (EXIT_FAILURE, errno, "%s", infile);
338 n_buffered += n_read;
339 if (n_buffered != n_bytes)
346 /* Find where to end this chunk. */
347 bp = buf + n_buffered;
348 if (n_buffered == n_bytes)
350 while (bp > buf && bp[-1] != '\n')
354 /* If chunk has no newlines, use all the chunk. */
356 bp = buf + n_buffered;
358 /* Output the chars as one output file. */
359 cwrite (true, buf, bp - buf);
361 /* Discard the chars we just output; move rest of chunk
362 down to be the start of the next chunk. Source and
363 destination probably overlap. */
364 n_buffered -= bp - buf;
366 memmove (buf, bp, n_buffered);
372 #define FAIL_ONLY_ONE_WAY() \
375 error (0, 0, _("cannot split in more than one way")); \
376 usage (EXIT_FAILURE); \
381 main (int argc, char **argv)
383 struct stat stat_buf;
386 type_undef, type_bytes, type_byteslines, type_lines, type_digits
387 } split_type = type_undef;
388 size_t in_blk_size; /* optimal block size of input file device */
389 char *buf; /* file i/o buffer */
390 size_t page_size = getpagesize ();
392 static char const multipliers[] = "bEGKkMmPTYZ0";
394 int digits_optind = 0;
396 initialize_main (&argc, &argv);
397 program_name = argv[0];
398 setlocale (LC_ALL, "");
399 bindtextdomain (PACKAGE, LOCALEDIR);
400 textdomain (PACKAGE);
402 atexit (close_stdout);
404 /* Parse command line options. */
411 /* This is the argv-index of the option we will read next. */
412 int this_optind = optind ? optind : 1;
414 c = getopt_long (argc, argv, "0123456789C:a:b:dl:", longopts, NULL);
423 if (xstrtoul (optarg, NULL, 10, &tmp, "") != LONGINT_OK
424 || SIZE_MAX / sizeof (size_t) < tmp)
426 error (0, 0, _("%s: invalid suffix length"), optarg);
427 usage (EXIT_FAILURE);
434 if (split_type != type_undef)
435 FAIL_ONLY_ONE_WAY ();
436 split_type = type_bytes;
437 if (xstrtoumax (optarg, NULL, 10, &n_units, multipliers) != LONGINT_OK
440 error (0, 0, _("%s: invalid number of bytes"), optarg);
441 usage (EXIT_FAILURE);
446 if (split_type != type_undef)
447 FAIL_ONLY_ONE_WAY ();
448 split_type = type_lines;
449 if (xstrtoumax (optarg, NULL, 10, &n_units, "") != LONGINT_OK
452 error (0, 0, _("%s: invalid number of lines"), optarg);
453 usage (EXIT_FAILURE);
458 if (split_type != type_undef)
459 FAIL_ONLY_ONE_WAY ();
460 split_type = type_byteslines;
461 if (xstrtoumax (optarg, NULL, 10, &n_units, multipliers) != LONGINT_OK
462 || n_units == 0 || SIZE_MAX < n_units)
464 error (0, 0, _("%s: invalid number of bytes"), optarg);
465 usage (EXIT_FAILURE);
479 if (split_type == type_undef)
481 split_type = type_digits;
484 if (split_type != type_undef && split_type != type_digits)
485 FAIL_ONLY_ONE_WAY ();
486 if (digits_optind != 0 && digits_optind != this_optind)
487 n_units = 0; /* More than one number given; ignore other. */
488 digits_optind = this_optind;
489 if (!DECIMAL_DIGIT_ACCUMULATE (n_units, c - '0', uintmax_t))
491 char buffer[INT_BUFSIZE_BOUND (uintmax_t)];
492 error (EXIT_FAILURE, 0,
493 _("line count option -%s%c... is too large"),
494 umaxtostr (n_units, buffer), c);
499 suffix_alphabet = "0123456789";
506 case_GETOPT_HELP_CHAR;
508 case_GETOPT_VERSION_CHAR (PROGRAM_NAME, AUTHORS);
511 usage (EXIT_FAILURE);
515 /* Handle default case. */
516 if (split_type == type_undef)
518 split_type = type_lines;
524 error (0, 0, _("invalid number of lines: 0"));
525 usage (EXIT_FAILURE);
528 /* Get out the filename arguments. */
531 infile = argv[optind++];
534 outbase = argv[optind++];
538 error (0, 0, _("extra operand %s"), quote (argv[optind]));
539 usage (EXIT_FAILURE);
542 /* Open the input file. */
543 if (! STREQ (infile, "-")
544 && fd_reopen (STDIN_FILENO, infile, O_RDONLY, 0) < 0)
545 error (EXIT_FAILURE, errno, _("cannot open %s for reading"),
548 /* Binary I/O is safer when bytecounts are used. */
549 if (O_BINARY && ! isatty (STDIN_FILENO))
550 freopen (NULL, "rb", stdin);
552 /* No output file is open now. */
555 /* Get the optimal block size of input device and make a buffer. */
557 if (fstat (STDIN_FILENO, &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 (STDIN_FILENO) != 0)
583 error (EXIT_FAILURE, errno, "%s", infile);
584 if (output_desc >= 0 && close (output_desc) < 0)
585 error (EXIT_FAILURE, errno, "%s", outfile);