1 /* split.c -- split a file into pieces.
2 Copyright (C) 88, 91, 1995-2005 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., 51 Franklin Street, Fifth Floor, 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"
38 #include "safe-read.h"
39 #include "unistd-safer.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 true, print a diagnostic on standard error just before each
78 output file is opened. */
81 /* For long options that have no equivalent short option, use a
82 non-character as a pseudo short option, starting with CHAR_MAX + 1. */
85 VERBOSE_OPTION = CHAR_MAX + 1
88 static struct option const longopts[] =
90 {"bytes", required_argument, NULL, 'b'},
91 {"lines", required_argument, NULL, 'l'},
92 {"line-bytes", required_argument, NULL, 'C'},
93 {"suffix-length", required_argument, NULL, 'a'},
94 {"numeric-suffixes", no_argument, NULL, 'd'},
95 {"verbose", no_argument, NULL, VERBOSE_OPTION},
96 {GETOPT_HELP_OPTION_DECL},
97 {GETOPT_VERSION_OPTION_DECL},
104 if (status != EXIT_SUCCESS)
105 fprintf (stderr, _("Try `%s --help' for more information.\n"),
110 Usage: %s [OPTION] [INPUT [PREFIX]]\n\
114 Output fixed-size pieces of INPUT to PREFIXaa, PREFIXab, ...; default\n\
115 size is 1000 lines, and default PREFIX is `x'. With no INPUT, or when INPUT\n\
116 is -, read standard input.\n\
120 Mandatory arguments to long options are mandatory for short options too.\n\
122 fprintf (stdout, _("\
123 -a, --suffix-length=N use suffixes of length N (default %d)\n\
124 -b, --bytes=SIZE put SIZE bytes per output file\n\
125 -C, --line-bytes=SIZE put at most SIZE bytes of lines per output file\n\
126 -d, --numeric-suffixes use numeric suffixes instead of alphabetic\n\
127 -l, --lines=NUMBER put NUMBER lines per output file\n\
128 "), DEFAULT_SUFFIX_LENGTH);
130 --verbose print a diagnostic to standard error just\n\
131 before each output file is opened\n\
133 fputs (HELP_OPTION_DESCRIPTION, stdout);
134 fputs (VERSION_OPTION_DESCRIPTION, stdout);
137 SIZE may have a multiplier suffix: b for 512, k for 1K, m for 1 Meg.\n\
139 printf (_("\nReport bugs to <%s>.\n"), PACKAGE_BUGREPORT);
144 /* Compute the next sequential output file name and store it into the
148 next_file_name (void)
150 /* Index in suffix_alphabet of each character in the suffix. */
151 static size_t *sufindex;
155 /* Allocate and initialize the first file name. */
157 size_t outbase_length = strlen (outbase);
158 size_t outfile_length = outbase_length + suffix_length;
159 if (outfile_length + 1 < outbase_length)
161 outfile = xmalloc (outfile_length + 1);
162 outfile_mid = outfile + outbase_length;
163 memcpy (outfile, outbase, outbase_length);
164 memset (outfile_mid, suffix_alphabet[0], suffix_length);
165 outfile[outfile_length] = 0;
166 sufindex = xcalloc (suffix_length, sizeof *sufindex);
168 #if ! _POSIX_NO_TRUNC && HAVE_PATHCONF && defined _PC_NAME_MAX
169 /* POSIX requires that if the output file name is too long for
170 its directory, `split' must fail without creating any files.
171 This must be checked for explicitly on operating systems that
172 silently truncate file names. */
174 char *dir = dir_name (outfile);
175 long name_max = pathconf (dir, _PC_NAME_MAX);
176 if (0 <= name_max && name_max < base_len (base_name (outfile)))
177 error (EXIT_FAILURE, ENAMETOOLONG, "%s", outfile);
184 /* Increment the suffix in place, if possible. */
186 size_t i = suffix_length;
190 outfile_mid[i] = suffix_alphabet[sufindex[i]];
194 outfile_mid[i] = suffix_alphabet[sufindex[i]];
196 error (EXIT_FAILURE, 0, _("Output file suffixes exhausted"));
200 /* Write BYTES bytes at BP to an output file.
201 If NEW_FILE_FLAG is true, open the next output file.
202 Otherwise add to the same output file already in use. */
205 cwrite (bool new_file_flag, const char *bp, size_t bytes)
209 if (output_desc >= 0 && close (output_desc) < 0)
210 error (EXIT_FAILURE, errno, "%s", outfile);
214 fprintf (stderr, _("creating file `%s'\n"), outfile);
215 output_desc = fd_safer (open (outfile,
216 O_WRONLY | O_CREAT | O_TRUNC | O_BINARY,
217 (S_IRUSR | S_IWUSR | S_IRGRP | S_IWGRP
218 | S_IROTH | S_IWOTH)));
220 error (EXIT_FAILURE, errno, "%s", outfile);
222 if (full_write (output_desc, bp, bytes) != bytes)
223 error (EXIT_FAILURE, errno, "%s", outfile);
226 /* Split into pieces of exactly N_BYTES bytes.
227 Use buffer BUF, whose size is BUFSIZE. */
230 bytes_split (uintmax_t n_bytes, char *buf, size_t bufsize)
233 bool new_file_flag = true;
235 uintmax_t to_write = n_bytes;
240 n_read = full_read (input_desc, buf, bufsize);
241 if (n_read == SAFE_READ_ERROR)
242 error (EXIT_FAILURE, errno, "%s", infile);
247 if (to_read < to_write)
249 if (to_read) /* do not write 0 bytes! */
251 cwrite (new_file_flag, bp_out, to_read);
253 new_file_flag = false;
260 cwrite (new_file_flag, bp_out, w);
263 new_file_flag = true;
268 while (n_read == bufsize);
271 /* Split into pieces of exactly N_LINES lines.
272 Use buffer BUF, whose size is BUFSIZE. */
275 lines_split (uintmax_t n_lines, char *buf, size_t bufsize)
278 char *bp, *bp_out, *eob;
279 bool new_file_flag = true;
284 n_read = full_read (input_desc, buf, bufsize);
285 if (n_read == SAFE_READ_ERROR)
286 error (EXIT_FAILURE, errno, "%s", infile);
292 bp = memchr (bp, '\n', eob - bp + 1);
295 if (eob != bp_out) /* do not write 0 bytes! */
297 size_t len = eob - bp_out;
298 cwrite (new_file_flag, bp_out, len);
299 new_file_flag = false;
307 cwrite (new_file_flag, bp_out, bp - bp_out);
309 new_file_flag = true;
314 while (n_read == bufsize);
317 /* Split into pieces that are as large as possible while still not more
318 than N_BYTES bytes, and are split on line boundaries except
319 where lines longer than N_BYTES bytes occur.
320 FIXME: Allow N_BYTES to be any uintmax_t value, and don't require a
321 buffer of size N_BYTES, in case N_BYTES is very large. */
324 line_bytes_split (size_t n_bytes)
329 size_t n_buffered = 0;
330 char *buf = xmalloc (n_bytes);
334 /* Fill up the full buffer size from the input file. */
336 n_read = full_read (input_desc, buf + n_buffered, n_bytes - n_buffered);
337 if (n_read == SAFE_READ_ERROR)
338 error (EXIT_FAILURE, errno, "%s", infile);
340 n_buffered += n_read;
341 if (n_buffered != n_bytes)
344 /* Find where to end this chunk. */
345 bp = buf + n_buffered;
346 if (n_buffered == n_bytes)
348 while (bp > buf && bp[-1] != '\n')
352 /* If chunk has no newlines, use all the chunk. */
354 bp = buf + n_buffered;
356 /* Output the chars as one output file. */
357 cwrite (true, buf, bp - buf);
359 /* Discard the chars we just output; move rest of chunk
360 down to be the start of the next chunk. Source and
361 destination probably overlap. */
362 n_buffered -= bp - buf;
364 memmove (buf, bp, n_buffered);
370 #define FAIL_ONLY_ONE_WAY() \
373 error (0, 0, _("cannot split in more than one way")); \
374 usage (EXIT_FAILURE); \
379 main (int argc, char **argv)
381 struct stat stat_buf;
384 type_undef, type_bytes, type_byteslines, type_lines, type_digits
385 } split_type = type_undef;
386 size_t in_blk_size; /* optimal block size of input file device */
387 char *buf; /* file i/o buffer */
388 size_t page_size = getpagesize ();
391 int digits_optind = 0;
393 initialize_main (&argc, &argv);
394 program_name = argv[0];
395 setlocale (LC_ALL, "");
396 bindtextdomain (PACKAGE, LOCALEDIR);
397 textdomain (PACKAGE);
399 atexit (close_stdout);
401 /* Parse command line options. */
408 /* This is the argv-index of the option we will read next. */
409 int this_optind = optind ? optind : 1;
411 c = getopt_long (argc, argv, "0123456789C:a:b:dl:", longopts, NULL);
420 if (xstrtoul (optarg, NULL, 10, &tmp, "") != LONGINT_OK
421 || SIZE_MAX / sizeof (size_t) < tmp)
423 error (0, 0, _("%s: invalid suffix length"), optarg);
424 usage (EXIT_FAILURE);
431 if (split_type != type_undef)
432 FAIL_ONLY_ONE_WAY ();
433 split_type = type_bytes;
434 if (xstrtoumax (optarg, NULL, 10, &n_units, "bkm") != LONGINT_OK
437 error (0, 0, _("%s: invalid number of bytes"), optarg);
438 usage (EXIT_FAILURE);
443 if (split_type != type_undef)
444 FAIL_ONLY_ONE_WAY ();
445 split_type = type_lines;
446 if (xstrtoumax (optarg, NULL, 10, &n_units, "") != LONGINT_OK
449 error (0, 0, _("%s: invalid number of lines"), optarg);
450 usage (EXIT_FAILURE);
455 if (split_type != type_undef)
456 FAIL_ONLY_ONE_WAY ();
457 split_type = type_byteslines;
458 if (xstrtoumax (optarg, NULL, 10, &n_units, "bkm") != LONGINT_OK
459 || n_units == 0 || SIZE_MAX < n_units)
461 error (0, 0, _("%s: invalid number of bytes"), optarg);
462 usage (EXIT_FAILURE);
476 if (split_type == type_undef)
478 split_type = type_digits;
481 if (split_type != type_undef && split_type != type_digits)
482 FAIL_ONLY_ONE_WAY ();
483 if (digits_optind != 0 && digits_optind != this_optind)
484 n_units = 0; /* More than one number given; ignore other. */
485 digits_optind = this_optind;
486 if (!DECIMAL_DIGIT_ACCUMULATE (n_units, c - '0', UINTMAX_MAX))
488 char buffer[INT_BUFSIZE_BOUND (uintmax_t)];
489 error (EXIT_FAILURE, 0,
490 _("line count option -%s%c... is too large"),
491 umaxtostr (n_units, buffer), c);
496 suffix_alphabet = "0123456789";
503 case_GETOPT_HELP_CHAR;
505 case_GETOPT_VERSION_CHAR (PROGRAM_NAME, AUTHORS);
508 usage (EXIT_FAILURE);
512 /* Handle default case. */
513 if (split_type == type_undef)
515 split_type = type_lines;
521 error (0, 0, _("invalid number of lines: 0"));
522 usage (EXIT_FAILURE);
525 /* Get out the filename arguments. */
528 infile = argv[optind++];
531 outbase = argv[optind++];
535 error (0, 0, _("extra operand %s"), quote (argv[optind]));
536 usage (EXIT_FAILURE);
539 /* Open the input file. */
540 if (STREQ (infile, "-"))
541 input_desc = STDIN_FILENO;
544 input_desc = open (infile, O_RDONLY);
546 error (EXIT_FAILURE, errno, "%s", infile);
548 /* Binary I/O is safer when bytecounts are used. */
549 SET_BINARY (input_desc);
551 /* No output file is open now. */
554 /* Get the optimal block size of input device and make a buffer. */
556 if (fstat (input_desc, &stat_buf) < 0)
557 error (EXIT_FAILURE, errno, "%s", infile);
558 in_blk_size = ST_BLKSIZE (stat_buf);
560 buf = ptr_align (xmalloc (in_blk_size + 1 + page_size - 1), page_size);
566 lines_split (n_units, buf, in_blk_size);
570 bytes_split (n_units, buf, in_blk_size);
573 case type_byteslines:
574 line_bytes_split (n_units);
581 if (close (input_desc) < 0)
582 error (EXIT_FAILURE, errno, "%s", infile);
583 if (output_desc >= 0 && close (output_desc) < 0)
584 error (EXIT_FAILURE, errno, "%s", outfile);