1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
|
/* tac - concatenate and print files in reverse
Copyright (C) 1988-1991, 1995-2005 Free Software Foundation, Inc.
This program is free software; you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
the Free Software Foundation; either version 2, or (at your option)
any later version.
This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU General Public License for more details.
You should have received a copy of the GNU General Public License
along with this program; if not, write to the Free Software Foundation,
Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */
/* Written by Jay Lepreau (lepreau@cs.utah.edu).
GNU enhancements by David MacKenzie (djm@gnu.ai.mit.edu). */
/* Copy each FILE, or the standard input if none are given or when a
FILE name of "-" is encountered, to the standard output with the
order of the records reversed. The records are separated by
instances of a string, or a newline if none is given. By default, the
separator string is attached to the end of the record that it
follows in the file.
Options:
-b, --before The separator is attached to the beginning
of the record that it precedes in the file.
-r, --regex The separator is a regular expression.
-s, --separator=separator Use SEPARATOR as the record separator.
To reverse a file byte by byte, use (in bash, ksh, or sh):
tac -r -s '.\|
' file */
#include <config.h>
#include <stdio.h>
#include <getopt.h>
#include <sys/types.h>
#include "system.h"
#include <regex.h>
#include "error.h"
#include "quote.h"
#include "quotearg.h"
#include "safe-read.h"
#include "unistd-safer.h"
/* The official name of this program (e.g., no `g' prefix). */
#define PROGRAM_NAME "tac"
#define AUTHORS "Jay Lepreau", "David MacKenzie"
#if defined __MSDOS__ || defined _WIN32
/* Define this to non-zero on systems for which the regular mechanism
(of unlinking an open file and expecting to be able to write, seek
back to the beginning, then reread it) doesn't work. E.g., on Windows
and DOS systems. */
# define DONT_UNLINK_WHILE_OPEN 1
#endif
#ifndef DEFAULT_TMPDIR
# define DEFAULT_TMPDIR "/tmp"
#endif
/* The number of bytes per atomic read. */
#define INITIAL_READSIZE 8192
/* The number of bytes per atomic write. */
#define WRITESIZE 8192
/* The name this program was run with. */
char *program_name;
/* The string that separates the records of the file. */
static char *separator;
/* True if we have ever read standard input. */
static bool have_read_stdin = false;
/* If true, print `separator' along with the record preceding it
in the file; otherwise with the record following it. */
static bool separator_ends_record;
/* 0 if `separator' is to be matched as a regular expression;
otherwise, the length of `separator', used as a sentinel to
stop the search. */
static size_t sentinel_length;
/* The length of a match with `separator'. If `sentinel_length' is 0,
`match_length' is computed every time a match succeeds;
otherwise, it is simply the length of `separator'. */
static size_t match_length;
/* The input buffer. */
static char *G_buffer;
/* The number of bytes to read at once into `buffer'. */
static size_t read_size;
/* The size of `buffer'. This is read_size * 2 + sentinel_length + 2.
The extra 2 bytes allow `past_end' to have a value beyond the
end of `G_buffer' and `match_start' to run off the front of `G_buffer'. */
static size_t G_buffer_size;
/* The compiled regular expression representing `separator'. */
static struct re_pattern_buffer compiled_separator;
static struct option const longopts[] =
{
{"before", no_argument, NULL, 'b'},
{"regex", no_argument, NULL, 'r'},
{"separator", required_argument, NULL, 's'},
{GETOPT_HELP_OPTION_DECL},
{GETOPT_VERSION_OPTION_DECL},
{NULL, 0, NULL, 0}
};
void
usage (int status)
{
if (status != EXIT_SUCCESS)
fprintf (stderr, _("Try `%s --help' for more information.\n"),
program_name);
else
{
printf (_("\
Usage: %s [OPTION]... [FILE]...\n\
"),
program_name);
fputs (_("\
Write each FILE to standard output, last line first.\n\
With no FILE, or when FILE is -, read standard input.\n\
\n\
"), stdout);
fputs (_("\
Mandatory arguments to long options are mandatory for short options too.\n\
"), stdout);
fputs (_("\
-b, --before attach the separator before instead of after\n\
-r, --regex interpret the separator as a regular expression\n\
-s, --separator=STRING use STRING as the separator instead of newline\n\
"), stdout);
fputs (HELP_OPTION_DESCRIPTION, stdout);
fputs (VERSION_OPTION_DESCRIPTION, stdout);
printf (_("\nReport bugs to <%s>.\n"), PACKAGE_BUGREPORT);
}
exit (status);
}
/* Print the characters from START to PAST_END - 1.
If START is NULL, just flush the buffer. */
static void
output (const char *start, const char *past_end)
{
static char buffer[WRITESIZE];
static size_t bytes_in_buffer = 0;
size_t bytes_to_add = past_end - start;
size_t bytes_available = WRITESIZE - bytes_in_buffer;
if (start == 0)
{
fwrite (buffer, 1, bytes_in_buffer, stdout);
bytes_in_buffer = 0;
return;
}
/* Write out as many full buffers as possible. */
while (bytes_to_add >= bytes_available)
{
memcpy (buffer + bytes_in_buffer, start, bytes_available);
bytes_to_add -= bytes_available;
start += bytes_available;
fwrite (buffer, 1, WRITESIZE, stdout);
bytes_in_buffer = 0;
bytes_available = WRITESIZE;
}
memcpy (buffer + bytes_in_buffer, start, bytes_to_add);
bytes_in_buffer += bytes_to_add;
}
/* Print in reverse the file open on descriptor FD for reading FILE.
Return true if successful. */
static bool
tac_seekable (int input_fd, const char *file)
{
/* Pointer to the location in `G_buffer' where the search for
the next separator will begin. */
char *match_start;
/* Pointer to one past the rightmost character in `G_buffer' that
has not been printed yet. */
char *past_end;
/* Length of the record growing in `G_buffer'. */
size_t saved_record_size;
/* Offset in the file of the next read. */
off_t file_pos;
/* True if `output' has not been called yet for any file.
Only used when the separator is attached to the preceding record. */
bool first_time = true;
char first_char = *separator; /* Speed optimization, non-regexp. */
char *separator1 = separator + 1; /* Speed optimization, non-regexp. */
size_t match_length1 = match_length - 1; /* Speed optimization, non-regexp. */
struct re_registers regs;
/* Find the size of the input file. */
file_pos = lseek (input_fd, (off_t) 0, SEEK_END);
if (file_pos < 1)
return true; /* It's an empty file. */
/* Arrange for the first read to lop off enough to leave the rest of the
file a multiple of `read_size'. Since `read_size' can change, this may
not always hold during the program run, but since it usually will, leave
it here for i/o efficiency (page/sector boundaries and all that).
Note: the efficiency gain has not been verified. */
saved_record_size = file_pos % read_size;
if (saved_record_size == 0)
saved_record_size = read_size;
file_pos -= saved_record_size;
/* `file_pos' now points to the start of the last (probably partial) block
in the input file. */
if (lseek (input_fd, file_pos, SEEK_SET) < 0)
error (0, errno, _("%s: seek failed"), quotearg_colon (file));
if (safe_read (input_fd, G_buffer, saved_record_size) != saved_record_size)
{
error (0, errno, _("%s: read error"), quotearg_colon (file));
return false;
}
match_start = past_end = G_buffer + saved_record_size;
/* For non-regexp search, move past impossible positions for a match. */
if (sentinel_length)
match_start -= match_length1;
for (;;)
{
/* Search backward from `match_start' - 1 to `G_buffer' for a match
with `separator'; for speed, use strncmp if `separator' contains no
metacharacters.
If the match succeeds, set `match_start' to point to the start of
the match and `match_length' to the length of the match.
Otherwise, make `match_start' < `G_buffer'. */
if (sentinel_length == 0)
{
ptrdiff_t i = match_start - G_buffer;
int ret;
if (! (INT_MIN < i && i <= INT_MAX))
error (EXIT_FAILURE, 0, _("record too large"));
ret = re_search (&compiled_separator, G_buffer, i, i - 1, -i, ®s);
if (ret == -1)
match_start = G_buffer - 1;
else if (ret == -2)
{
error (EXIT_FAILURE, 0,
_("error in regular expression search"));
}
else
{
match_start = G_buffer + regs.start[0];
match_length = regs.end[0] - regs.start[0];
}
}
else
{
/* `match_length' is constant for non-regexp boundaries. */
while (*--match_start != first_char
|| (match_length1 && strncmp (match_start + 1, separator1,
match_length1)))
/* Do nothing. */ ;
}
/* Check whether we backed off the front of `G_buffer' without finding
a match for `separator'. */
if (match_start < G_buffer)
{
if (file_pos == 0)
{
/* Hit the beginning of the file; print the remaining record. */
output (G_buffer, past_end);
return true;
}
saved_record_size = past_end - G_buffer;
if (saved_record_size > read_size)
{
/* `G_buffer_size' is about twice `read_size', so since
we want to read in another `read_size' bytes before
the data already in `G_buffer', we need to increase
`G_buffer_size'. */
char *newbuffer;
size_t offset = sentinel_length ? sentinel_length : 1;
ptrdiff_t match_start_offset = match_start - G_buffer;
ptrdiff_t past_end_offset = past_end - G_buffer;
size_t old_G_buffer_size = G_buffer_size;
read_size *= 2;
G_buffer_size = read_size * 2 + sentinel_length + 2;
if (G_buffer_size < old_G_buffer_size)
xalloc_die ();
newbuffer = xrealloc (G_buffer - offset, G_buffer_size);
newbuffer += offset;
/* Adjust the pointers for the new buffer location. */
match_start = newbuffer + match_start_offset;
past_end = newbuffer + past_end_offset;
G_buffer = newbuffer;
}
/* Back up to the start of the next bufferfull of the file. */
if (file_pos >= read_size)
file_pos -= read_size;
else
{
read_size = file_pos;
file_pos = 0;
}
if (lseek (input_fd, file_pos, SEEK_SET) < 0)
error (0, errno, _("%s: seek failed"), quotearg_colon (file));
/* Shift the pending record data right to make room for the new.
The source and destination regions probably overlap. */
memmove (G_buffer + read_size, G_buffer, saved_record_size);
past_end = G_buffer + read_size + saved_record_size;
/* For non-regexp searches, avoid unneccessary scanning. */
if (sentinel_length)
match_start = G_buffer + read_size;
else
match_start = past_end;
if (safe_read (input_fd, G_buffer, read_size) != read_size)
{
error (0, errno, _("%s: read error"), quotearg_colon (file));
return false;
}
}
else
{
/* Found a match of `separator'. */
if (separator_ends_record)
{
char *match_end = match_start + match_length;
/* If this match of `separator' isn't at the end of the
file, print the record. */
if (!first_time || match_end != past_end)
output (match_end, past_end);
past_end = match_end;
first_time = false;
}
else
{
output (match_start, past_end);
past_end = match_start;
}
/* For non-regex matching, we can back up. */
if (sentinel_length > 0)
match_start -= match_length - 1;
}
}
}
#if DONT_UNLINK_WHILE_OPEN
/* FIXME-someday: remove all of this DONT_UNLINK_WHILE_OPEN junk.
Using atexit like this is wrong, since it can fail
when called e.g. 32 or more times.
But this isn't a big deal, since the code is used only on WOE/DOS
systems, and few people invoke tac on that many nonseekable files. */
static const char *file_to_remove;
static FILE *fp_to_close;
static void
unlink_tempfile (void)
{
fclose (fp_to_close);
unlink (file_to_remove);
}
static void
record_or_unlink_tempfile (char const *fn, FILE *fp)
{
if (!file_to_remove)
{
file_to_remove = fn;
fp_to_close = fp;
atexit (unlink_tempfile);
}
}
#else
static void
record_or_unlink_tempfile (char const *fn, FILE *fp ATTRIBUTE_UNUSED)
{
unlink (fn);
}
#endif
/* Copy from file descriptor INPUT_FD (corresponding to the named FILE) to
a temporary file, and set *G_TMP and *G_TEMPFILE to the resulting stream
and file name. Return true if successful. */
static bool
copy_to_temp (FILE **g_tmp, char **g_tempfile, int input_fd, char const *file)
{
static char *template = NULL;
static char *tempdir;
char *tempfile;
FILE *tmp;
int fd;
if (template == NULL)
{
char const * const Template = "%s/tacXXXXXX";
tempdir = getenv ("TMPDIR");
if (tempdir == NULL)
tempdir = DEFAULT_TMPDIR;
/* Subtract 2 for `%s' and add 1 for the trailing NUL byte. */
template = xmalloc (strlen (tempdir) + strlen (Template) - 2 + 1);
sprintf (template, Template, tempdir);
}
/* FIXME: there's a small window between a successful mkstemp call
and the unlink that's performed by record_or_unlink_tempfile.
If we're interrupted in that interval, this code fails to remove
the temporary file. On systems that define DONT_UNLINK_WHILE_OPEN,
the window is much larger -- it extends to the atexit-called
unlink_tempfile.
FIXME: clean up upon fatal signal. Don't block them, in case
$TMPFILE is a remote file system. */
tempfile = template;
fd = mkstemp (template);
if (fd < 0)
{
error (0, errno, _("cannot create temporary file %s"), quote (tempfile));
return false;
}
if ((fd = fd_safer (fd)) < 0 || ! (tmp = fdopen (fd, "w+")))
{
error (0, errno, _("cannot open %s for writing"), quote (tempfile));
close (fd);
unlink (tempfile);
return false;
}
record_or_unlink_tempfile (tempfile, tmp);
while (1)
{
size_t bytes_read = safe_read (input_fd, G_buffer, read_size);
if (bytes_read == 0)
break;
if (bytes_read == SAFE_READ_ERROR)
{
error (0, errno, _("%s: read error"), quotearg_colon (file));
goto Fail;
}
if (fwrite (G_buffer, 1, bytes_read, tmp) != bytes_read)
{
error (0, errno, _("%s: write error"), quotearg_colon (tempfile));
goto Fail;
}
}
if (fflush (tmp) != 0)
{
error (0, errno, _("%s: write error"), quotearg_colon (tempfile));
goto Fail;
}
SET_BINARY (fileno (tmp));
*g_tmp = tmp;
*g_tempfile = tempfile;
return true;
Fail:
fclose (tmp);
return false;
}
/* Copy INPUT_FD to a temporary, then tac that file.
Return true if successful. */
static bool
tac_nonseekable (int input_fd, const char *file)
{
FILE *tmp_stream;
char *tmp_file;
return (copy_to_temp (&tmp_stream, &tmp_file, input_fd, file)
&& tac_seekable (fileno (tmp_stream), tmp_file));
}
/* Print FILE in reverse, copying it to a temporary
file first if it is not seekable.
Return true if successful. */
static bool
tac_file (const char *filename)
{
bool ok;
off_t file_size;
int fd;
bool is_stdin = STREQ (filename, "-");
if (is_stdin)
{
have_read_stdin = true;
fd = STDIN_FILENO;
filename = _("standard input");
}
else
{
fd = open (filename, O_RDONLY);
if (fd < 0)
{
error (0, errno, _("cannot open %s for reading"), quote (filename));
return false;
}
}
/* We need binary I/O, since `tac' relies
on `lseek' and byte counts.
Binary output will leave the lines' ends (NL or
CR/LF) intact when the output is a disk file.
Writing a file with CR/LF pairs at end of lines in
text mode has no visible effect on console output,
since two CRs in a row are just like one CR. */
SET_BINARY2 (fd, STDOUT_FILENO);
file_size = lseek (fd, (off_t) 0, SEEK_END);
ok = (0 <= file_size
? tac_seekable (fd, filename)
: tac_nonseekable (fd, filename));
if (!is_stdin && close (fd) != 0)
{
error (0, errno, _("%s: read error"), quotearg_colon (filename));
ok = false;
}
return ok;
}
#if 0
/* BUF_END points one byte past the end of the buffer to be searched. */
/* FIXME: describe */
static void
tac_mem (const char *buf, size_t n_bytes, FILE *out)
{
const char *nl;
const char *bol;
if (n_bytes == 0)
return;
nl = memrchr (buf, buf + n_bytes, '\n');
bol = (nl == NULL ? buf : nl + 1);
/* If the last line of the input file has no terminating newline,
treat it as a special case. */
if (bol < buf + n_bytes)
{
/* Print out the line from bol to end of input. */
fwrite (bol, 1, (buf + n_bytes) - bol, out);
/* Add a newline here. Otherwise, the first and second lines
of output would appear to have been joined. */
fputc ('\n', out);
}
while ((nl = memrchr (buf, bol - 1, '\n')) != NULL)
{
/* Output the line (which includes a trailing newline)
from NL+1 to BOL-1. */
fwrite (nl + 1, 1, bol - (nl + 1), out);
bol = nl + 1;
}
/* If there's anything left, output the last line: BUF .. BOL-1.
When the first byte of the input is a newline, there is nothing
left to do here. */
if (buf < bol)
fwrite (buf, 1, bol - buf, out);
/* FIXME: this is work in progress.... */
}
/* FIXME: describe */
static bool
tac_stdin_to_mem (void)
{
char *buf = NULL;
size_t bufsiz = 8 * BUFSIZ;
size_t delta = 8 * BUFSIZ;
size_t n_bytes = 0;
while (1)
{
size_t bytes_read;
char *new_buf = realloc (buf, bufsiz);
if (new_buf == NULL)
{
/* Write contents of buf to a temporary file, ... */
/* FIXME */
/* Free the buffer and fall back on the code that relies on a
temporary file. */
free (buf);
/* FIXME */
abort ();
}
buf = new_buf;
bytes_read = safe_read (STDIN_FILENO, buf + n_bytes, bufsiz - n_bytes);
if (bytes_read == 0)
break;
if (bytes_read == SAFE_READ_ERROR)
error (EXIT_FAILURE, errno, _("stdin: read error"));
n_bytes += bytes_read;
bufsiz += delta;
}
tac_mem (buf, n_bytes, stdout);
return true;
}
#endif
int
main (int argc, char **argv)
{
const char *error_message; /* Return value from re_compile_pattern. */
int optc;
bool ok;
size_t half_buffer_size;
/* Initializer for file_list if no file-arguments
were specified on the command line. */
static char const *const default_file_list[] = {"-", NULL};
char const *const *file;
initialize_main (&argc, &argv);
program_name = argv[0];
setlocale (LC_ALL, "");
bindtextdomain (PACKAGE, LOCALEDIR);
textdomain (PACKAGE);
atexit (close_stdout);
separator = "\n";
sentinel_length = 1;
separator_ends_record = true;
while ((optc = getopt_long (argc, argv, "brs:", longopts, NULL)) != -1)
{
switch (optc)
{
case 'b':
separator_ends_record = false;
break;
case 'r':
sentinel_length = 0;
break;
case 's':
separator = optarg;
if (*separator == 0)
error (EXIT_FAILURE, 0, _("separator cannot be empty"));
break;
case_GETOPT_HELP_CHAR;
case_GETOPT_VERSION_CHAR (PROGRAM_NAME, AUTHORS);
default:
usage (EXIT_FAILURE);
}
}
if (sentinel_length == 0)
{
compiled_separator.allocated = 100;
compiled_separator.buffer = xmalloc (compiled_separator.allocated);
compiled_separator.fastmap = xmalloc (256);
compiled_separator.translate = NULL;
error_message = re_compile_pattern (separator, strlen (separator),
&compiled_separator);
if (error_message)
error (EXIT_FAILURE, 0, "%s", error_message);
}
else
match_length = sentinel_length = strlen (separator);
read_size = INITIAL_READSIZE;
while (sentinel_length >= read_size / 2)
{
if (SIZE_MAX / 2 < read_size)
xalloc_die ();
read_size *= 2;
}
half_buffer_size = read_size + sentinel_length + 1;
G_buffer_size = 2 * half_buffer_size;
if (! (read_size < half_buffer_size && half_buffer_size < G_buffer_size))
xalloc_die ();
G_buffer = xmalloc (G_buffer_size);
if (sentinel_length)
{
strcpy (G_buffer, separator);
G_buffer += sentinel_length;
}
else
{
++G_buffer;
}
file = (optind < argc
? (char const *const *) &argv[optind]
: default_file_list);
{
size_t i;
ok = true;
for (i = 0; file[i]; ++i)
ok &= tac_file (file[i]);
}
/* Flush the output buffer. */
output ((char *) NULL, (char *) NULL);
if (have_read_stdin && close (STDIN_FILENO) < 0)
error (EXIT_FAILURE, errno, "-");
exit (ok ? EXIT_SUCCESS : EXIT_FAILURE);
}
|