source: trunk/coreutils/src/split.c@ 2603

Last change on this file since 2603 was 2554, checked in by bird, 20 years ago

coretuils-5.94
File size: 14.6 KB
Line 
1/* split.c -- split a file into pieces.
2 Copyright (C) 88, 91, 1995-2005 Free Software Foundation, Inc.
3
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)
7 any later version.
8
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.
13
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 02110-1301, USA. */
17
18
19/* By [email protected], with rms.
20
21 To do:
22 * Implement -t CHAR or -t REGEX to specify break characters other
23 than newline. */
24
25#include <config.h>
26
27#include <stdio.h>
28#include <getopt.h>
29#include <sys/types.h>
30
31#include "system.h"
32#include "dirname.h"
33#include "error.h"
34#include "fd-reopen.h"
35#include "fcntl--.h"
36#include "getpagesize.h"
37#include "full-read.h"
38#include "full-write.h"
39#include "inttostr.h"
40#include "quote.h"
41#include "safe-read.h"
42#include "xstrtol.h"
43
44/* The official name of this program (e.g., no `g' prefix). */
45#define PROGRAM_NAME "split"
46
47#define AUTHORS "Torbjorn Granlund", "Richard M. Stallman"
48
49#define DEFAULT_SUFFIX_LENGTH 2
50
51/* The name this program was run with. */
52char *program_name;
53
54/* Base name of output files. */
55static char const *outbase;
56
57/* Name of output files. */
58static char *outfile;
59
60/* Pointer to the end of the prefix in OUTFILE.
61 Suffixes are inserted here. */
62static char *outfile_mid;
63
64/* Length of OUTFILE's suffix. */
65static size_t suffix_length = DEFAULT_SUFFIX_LENGTH;
66
67/* Alphabet of characters to use in suffix. */
68static char const *suffix_alphabet = "abcdefghijklmnopqrstuvwxyz";
69
70/* Name of input file. May be "-". */
71static char *infile;
72
73/* Descriptor on which output file is open. */
74static int output_desc;
75
76/* If true, print a diagnostic on standard error just before each
77 output file is opened. */
78static bool verbose;
79
80/* For long options that have no equivalent short option, use a
81 non-character as a pseudo short option, starting with CHAR_MAX + 1. */
82enum
83{
84 VERBOSE_OPTION = CHAR_MAX + 1
85};
86
87static struct option const longopts[] =
88{
89 {"bytes", required_argument, NULL, 'b'},
90 {"lines", required_argument, NULL, 'l'},
91 {"line-bytes", required_argument, NULL, 'C'},
92 {"suffix-length", required_argument, NULL, 'a'},
93 {"numeric-suffixes", no_argument, NULL, 'd'},
94 {"verbose", no_argument, NULL, VERBOSE_OPTION},
95 {GETOPT_HELP_OPTION_DECL},
96 {GETOPT_VERSION_OPTION_DECL},
97 {NULL, 0, NULL, 0}
98};
99
100void
101usage (int status)
102{
103 if (status != EXIT_SUCCESS)
104 fprintf (stderr, _("Try `%s --help' for more information.\n"),
105 program_name);
106 else
107 {
108 printf (_("\
109Usage: %s [OPTION] [INPUT [PREFIX]]\n\
110"),
111 program_name);
112 fputs (_("\
113Output fixed-size pieces of INPUT to PREFIXaa, PREFIXab, ...; default\n\
114size is 1000 lines, and default PREFIX is `x'. With no INPUT, or when INPUT\n\
115is -, read standard input.\n\
116\n\
117"), stdout);
118 fputs (_("\
119Mandatory arguments to long options are mandatory for short options too.\n\
120"), stdout);
121 fprintf (stdout, _("\
122 -a, --suffix-length=N use suffixes of length N (default %d)\n\
123 -b, --bytes=SIZE put SIZE bytes per output file\n\
124 -C, --line-bytes=SIZE put at most SIZE bytes of lines per output file\n\
125 -d, --numeric-suffixes use numeric suffixes instead of alphabetic\n\
126 -l, --lines=NUMBER put NUMBER lines per output file\n\
127"), DEFAULT_SUFFIX_LENGTH);
128 fputs (_("\
129 --verbose print a diagnostic to standard error just\n\
130 before each output file is opened\n\
131"), stdout);
132 fputs (HELP_OPTION_DESCRIPTION, stdout);
133 fputs (VERSION_OPTION_DESCRIPTION, stdout);
134 fputs (_("\
135\n\
136SIZE may have a multiplier suffix: b for 512, k for 1K, m for 1 Meg.\n\
137"), stdout);
138 printf (_("\nReport bugs to <%s>.\n"), PACKAGE_BUGREPORT);
139 }
140 exit (status);
141}
142
143/* Compute the next sequential output file name and store it into the
144 string `outfile'. */
145
146static void
147next_file_name (void)
148{
149 /* Index in suffix_alphabet of each character in the suffix. */
150 static size_t *sufindex;
151
152 if (! outfile)
153 {
154 /* Allocate and initialize the first file name. */
155
156 size_t outbase_length = strlen (outbase);
157 size_t outfile_length = outbase_length + suffix_length;
158 if (outfile_length + 1 < outbase_length)
159 xalloc_die ();
160 outfile = xmalloc (outfile_length + 1);
161 outfile_mid = outfile + outbase_length;
162 memcpy (outfile, outbase, outbase_length);
163 memset (outfile_mid, suffix_alphabet[0], suffix_length);
164 outfile[outfile_length] = 0;
165 sufindex = xcalloc (suffix_length, sizeof *sufindex);
166
167#if ! _POSIX_NO_TRUNC && HAVE_PATHCONF && defined _PC_NAME_MAX
168 /* POSIX requires that if the output file name is too long for
169 its directory, `split' must fail without creating any files.
170 This must be checked for explicitly on operating systems that
171 silently truncate file names. */
172 {
173 char *dir = dir_name (outfile);
174 long name_max = pathconf (dir, _PC_NAME_MAX);
175 if (0 <= name_max && name_max < base_len (base_name (outfile)))
176 error (EXIT_FAILURE, ENAMETOOLONG, "%s", outfile);
177 free (dir);
178 }
179#endif
180 }
181 else
182 {
183 /* Increment the suffix in place, if possible. */
184
185 size_t i = suffix_length;
186 while (i-- != 0)
187 {
188 sufindex[i]++;
189 outfile_mid[i] = suffix_alphabet[sufindex[i]];
190 if (outfile_mid[i])
191 return;
192 sufindex[i] = 0;
193 outfile_mid[i] = suffix_alphabet[sufindex[i]];
194 }
195 error (EXIT_FAILURE, 0, _("Output file suffixes exhausted"));
196 }
197}
198
199/* Write BYTES bytes at BP to an output file.
200 If NEW_FILE_FLAG is true, open the next output file.
201 Otherwise add to the same output file already in use. */
202
203static void
204cwrite (bool new_file_flag, const char *bp, size_t bytes)
205{
206 if (new_file_flag)
207 {
208 if (output_desc >= 0 && close (output_desc) < 0)
209 error (EXIT_FAILURE, errno, "%s", outfile);
210
211 next_file_name ();
212 if (verbose)
213 fprintf (stderr, _("creating file %s\n"), quote (outfile));
214 output_desc = open (outfile,
215 O_WRONLY | O_CREAT | O_TRUNC | O_BINARY,
216 (S_IRUSR | S_IWUSR | S_IRGRP | S_IWGRP
217 | S_IROTH | S_IWOTH));
218 if (output_desc < 0)
219 error (EXIT_FAILURE, errno, "%s", outfile);
220 }
221 if (full_write (output_desc, bp, bytes) != bytes)
222 error (EXIT_FAILURE, errno, "%s", outfile);
223}
224
225/* Split into pieces of exactly N_BYTES bytes.
226 Use buffer BUF, whose size is BUFSIZE. */
227
228static void
229bytes_split (uintmax_t n_bytes, char *buf, size_t bufsize)
230{
231 size_t n_read;
232 bool new_file_flag = true;
233 size_t to_read;
234 uintmax_t to_write = n_bytes;
235 char *bp_out;
236
237 do
238 {
239 n_read = full_read (STDIN_FILENO, buf, bufsize);
240 if (n_read == SAFE_READ_ERROR)
241 error (EXIT_FAILURE, errno, "%s", infile);
242 bp_out = buf;
243 to_read = n_read;
244 for (;;)
245 {
246 if (to_read < to_write)
247 {
248 if (to_read) /* do not write 0 bytes! */
249 {
250 cwrite (new_file_flag, bp_out, to_read);
251 to_write -= to_read;
252 new_file_flag = false;
253 }
254 break;
255 }
256 else
257 {
258 size_t w = to_write;
259 cwrite (new_file_flag, bp_out, w);
260 bp_out += w;
261 to_read -= w;
262 new_file_flag = true;
263 to_write = n_bytes;
264 }
265 }
266 }
267 while (n_read == bufsize);
268}
269
270/* Split into pieces of exactly N_LINES lines.
271 Use buffer BUF, whose size is BUFSIZE. */
272
273static void
274lines_split (uintmax_t n_lines, char *buf, size_t bufsize)
275{
276 size_t n_read;
277 char *bp, *bp_out, *eob;
278 bool new_file_flag = true;
279 uintmax_t n = 0;
280
281 do
282 {
283 n_read = full_read (STDIN_FILENO, buf, bufsize);
284 if (n_read == SAFE_READ_ERROR)
285 error (EXIT_FAILURE, errno, "%s", infile);
286 bp = bp_out = buf;
287 eob = bp + n_read;
288 *eob = '\n';
289 for (;;)
290 {
291 bp = memchr (bp, '\n', eob - bp + 1);
292 if (bp == eob)
293 {
294 if (eob != bp_out) /* do not write 0 bytes! */
295 {
296 size_t len = eob - bp_out;
297 cwrite (new_file_flag, bp_out, len);
298 new_file_flag = false;
299 }
300 break;
301 }
302
303 ++bp;
304 if (++n >= n_lines)
305 {
306 cwrite (new_file_flag, bp_out, bp - bp_out);
307 bp_out = bp;
308 new_file_flag = true;
309 n = 0;
310 }
311 }
312 }
313 while (n_read == bufsize);
314}
315
316
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. */
322
323static void
324line_bytes_split (size_t n_bytes)
325{
326 size_t n_read;
327 char *bp;
328 bool eof = false;
329 size_t n_buffered = 0;
330 char *buf = xmalloc (n_bytes);
331
332 do
333 {
334 /* Fill up the full buffer size from the input file. */
335
336 n_read = full_read (STDIN_FILENO, buf + n_buffered, n_bytes - n_buffered);
337 if (n_read == SAFE_READ_ERROR)
338 error (EXIT_FAILURE, errno, "%s", infile);
339
340 n_buffered += n_read;
341 if (n_buffered != n_bytes)
342 eof = true;
343
344 /* Find where to end this chunk. */
345 bp = buf + n_buffered;
346 if (n_buffered == n_bytes)
347 {
348 while (bp > buf && bp[-1] != '\n')
349 bp--;
350 }
351
352 /* If chunk has no newlines, use all the chunk. */
353 if (bp == buf)
354 bp = buf + n_buffered;
355
356 /* Output the chars as one output file. */
357 cwrite (true, buf, bp - buf);
358
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;
363 if (n_buffered > 0)
364 memmove (buf, bp, n_buffered);
365 }
366 while (!eof);
367 free (buf);
368}
369
370#define FAIL_ONLY_ONE_WAY() \
371 do \
372 { \
373 error (0, 0, _("cannot split in more than one way")); \
374 usage (EXIT_FAILURE); \
375 } \
376 while (0)
377
378int
379main (int argc, char **argv)
380{
381 struct stat stat_buf;
382 enum
383 {
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 ();
389 uintmax_t n_units;
390 int c;
391 int digits_optind = 0;
392
393 initialize_main (&argc, &argv);
394 program_name = argv[0];
395 setlocale (LC_ALL, "");
396 bindtextdomain (PACKAGE, LOCALEDIR);
397 textdomain (PACKAGE);
398
399 atexit (close_stdout);
400
401 /* Parse command line options. */
402
403 infile = "-";
404 outbase = "x";
405
406 while (1)
407 {
408 /* This is the argv-index of the option we will read next. */
409 int this_optind = optind ? optind : 1;
410
411 c = getopt_long (argc, argv, "0123456789C:a:b:dl:", longopts, NULL);
412 if (c == -1)
413 break;
414
415 switch (c)
416 {
417 case 'a':
418 {
419 unsigned long tmp;
420 if (xstrtoul (optarg, NULL, 10, &tmp, "") != LONGINT_OK
421 || SIZE_MAX / sizeof (size_t) < tmp)
422 {
423 error (0, 0, _("%s: invalid suffix length"), optarg);
424 usage (EXIT_FAILURE);
425 }
426 suffix_length = tmp;
427 }
428 break;
429
430 case 'b':
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
435 || n_units == 0)
436 {
437 error (0, 0, _("%s: invalid number of bytes"), optarg);
438 usage (EXIT_FAILURE);
439 }
440 break;
441
442 case 'l':
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
447 || n_units == 0)
448 {
449 error (0, 0, _("%s: invalid number of lines"), optarg);
450 usage (EXIT_FAILURE);
451 }
452 break;
453
454 case 'C':
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)
460 {
461 error (0, 0, _("%s: invalid number of bytes"), optarg);
462 usage (EXIT_FAILURE);
463 }
464 break;
465
466 case '0':
467 case '1':
468 case '2':
469 case '3':
470 case '4':
471 case '5':
472 case '6':
473 case '7':
474 case '8':
475 case '9':
476 if (split_type == type_undef)
477 {
478 split_type = type_digits;
479 n_units = 0;
480 }
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_t))
487 {
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);
492 }
493 break;
494
495 case 'd':
496 suffix_alphabet = "0123456789";
497 break;
498
499 case VERBOSE_OPTION:
500 verbose = true;
501 break;
502
503 case_GETOPT_HELP_CHAR;
504
505 case_GETOPT_VERSION_CHAR (PROGRAM_NAME, AUTHORS);
506
507 default:
508 usage (EXIT_FAILURE);
509 }
510 }
511
512 /* Handle default case. */
513 if (split_type == type_undef)
514 {
515 split_type = type_lines;
516 n_units = 1000;
517 }
518
519 if (n_units == 0)
520 {
521 error (0, 0, _("invalid number of lines: 0"));
522 usage (EXIT_FAILURE);
523 }
524
525 /* Get out the filename arguments. */
526
527 if (optind < argc)
528 infile = argv[optind++];
529
530 if (optind < argc)
531 outbase = argv[optind++];
532
533 if (optind < argc)
534 {
535 error (0, 0, _("extra operand %s"), quote (argv[optind]));
536 usage (EXIT_FAILURE);
537 }
538
539 /* Open the input file. */
540 if (! STREQ (infile, "-")
541 && fd_reopen (STDIN_FILENO, infile, O_RDONLY, 0) < 0)
542 error (EXIT_FAILURE, errno, _("cannot open %s for reading"),
543 quote (infile));
544
545 /* Binary I/O is safer when bytecounts are used. */
546 if (O_BINARY && ! isatty (STDIN_FILENO))
547 freopen (NULL, "rb", stdin);
548
549 /* No output file is open now. */
550 output_desc = -1;
551
552 /* Get the optimal block size of input device and make a buffer. */
553
554 if (fstat (STDIN_FILENO, &stat_buf) != 0)
555 error (EXIT_FAILURE, errno, "%s", infile);
556 in_blk_size = ST_BLKSIZE (stat_buf);
557
558 buf = ptr_align (xmalloc (in_blk_size + 1 + page_size - 1), page_size);
559
560 switch (split_type)
561 {
562 case type_digits:
563 case type_lines:
564 lines_split (n_units, buf, in_blk_size);
565 break;
566
567 case type_bytes:
568 bytes_split (n_units, buf, in_blk_size);
569 break;
570
571 case type_byteslines:
572 line_bytes_split (n_units);
573 break;
574
575 default:
576 abort ();
577 }
578
579 if (close (STDIN_FILENO) != 0)
580 error (EXIT_FAILURE, errno, "%s", infile);
581 if (output_desc >= 0 && close (output_desc) < 0)
582 error (EXIT_FAILURE, errno, "%s", outfile);
583
584 exit (EXIT_SUCCESS);
585}
Note: See TracBrowser for help on using the repository browser.