1 /* cut - remove parts of lines of files
2 Copyright (C) 1984, 1997, 1998, 1999, 2000, 2001 by David M. Ihnat
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 /* Written by David Ihnat. */
20 /* POSIX changes, bug fixes, long-named options, and cleanup
21 by David MacKenzie <djm@gnu.ai.mit.edu>.
23 Rewrite cut_fields and cut_bytes -- Jim Meyering. */
30 #include <sys/types.h>
36 /* The official name of this program (e.g., no `g' prefix). */
37 #define PROGRAM_NAME "cut"
39 #define AUTHORS N_ ("David Ihnat, David MacKenzie, and Jim Meyering")
41 #define FATAL_ERROR(Message) \
44 error (0, 0, (Message)); \
49 /* Append LOW, HIGH to the list RP of range pairs, allocating additional
50 space if necessary. Update local variable N_RP. When allocating,
51 update global variable N_RP_ALLOCATED. */
53 #define ADD_RANGE_PAIR(rp, low, high) \
56 if (n_rp >= n_rp_allocated) \
58 n_rp_allocated *= 2; \
59 (rp) = (struct range_pair *) xrealloc ((char *) (rp), \
60 n_rp_allocated * sizeof (*(rp))); \
62 rp[n_rp].lo = (low); \
63 rp[n_rp].hi = (high); \
74 /* This buffer is used to support the semantics of the -s option
75 (or lack of same) when the specified field list includes (does
76 not include) the first field. In both of those cases, the entire
77 first field must be read into this buffer to determine whether it
78 is followed by a delimiter or a newline before any of it may be
79 output. Otherwise, cut_fields can do the job without using this
81 static char *field_1_buffer;
83 /* The number of bytes allocated for FIELD_1_BUFFER. */
84 static size_t field_1_bufsize;
86 /* The largest field or byte index used as an endpoint of a closed
87 or degenerate range specification; this doesn't include the starting
88 index of right-open-ended ranges. For example, with either range spec
89 `2-5,9-', `2-3,5,9-' this variable would be set to 5. */
90 static unsigned int max_range_endpoint;
92 /* If nonzero, this is the index of the first field in a range that goes
94 static unsigned int eol_range_start;
96 /* In byte mode, which bytes to output.
97 In field mode, which DELIM-separated fields to output.
98 Both bytes and fields are numbered starting with 1,
99 so the zeroth element of this array is unused.
100 A field or byte K has been selected if
101 (K <= MAX_RANGE_ENDPOINT and PRINTABLE_FIELD[K])
102 || (EOL_RANGE_START > 0 && K >= EOL_RANGE_START). */
103 static int *printable_field;
109 /* Output characters that are in the given bytes. */
112 /* Output the given delimeter-separated fields. */
116 /* The name this program was run with. */
119 static enum operating_mode operating_mode;
121 /* If nonzero do not output lines containing no delimeter characters.
122 Otherwise, all such lines are printed. This option is valid only
124 static int suppress_non_delimited;
126 /* The delimeter character for field mode. */
129 /* The length of output_delimiter_string. */
130 static size_t output_delimiter_length;
132 /* The output field separator string. Defaults to the 1-character
133 string consisting of the input delimiter. */
134 static char *output_delimiter_string;
136 /* Nonzero if we have ever read standard input. */
137 static int have_read_stdin;
139 /* For long options that have no equivalent short option, use a
140 non-character as a pseudo short option, starting with CHAR_MAX + 1. */
143 OUTPUT_DELIMITER_OPTION = CHAR_MAX + 1
146 static struct option const longopts[] =
148 {"bytes", required_argument, 0, 'b'},
149 {"characters", required_argument, 0, 'c'},
150 {"fields", required_argument, 0, 'f'},
151 {"delimiter", required_argument, 0, 'd'},
152 {"only-delimited", no_argument, 0, 's'},
153 {"output-delimiter", required_argument, 0, OUTPUT_DELIMITER_OPTION},
154 {GETOPT_HELP_OPTION_DECL},
155 {GETOPT_VERSION_OPTION_DECL},
163 fprintf (stderr, _("Try `%s --help' for more information.\n"),
168 Usage: %s [OPTION]... [FILE]...\n\
172 Print selected parts of lines from each FILE to standard output.\n\
174 Mandatory arguments to long options are mandatory for short options too.\n\
175 -b, --bytes=LIST output only these bytes\n\
176 -c, --characters=LIST output only these characters\n\
177 -d, --delimiter=DELIM use DELIM instead of TAB for field delimiter\n\
178 -f, --fields=LIST output only these fields; also print any line\n\
179 that contains no delimiter character, unless\n\
180 the -s option is specified\n\
182 -s, --only-delimited do not print lines not containing delimiters\n\
183 --output-delimiter=STRING use STRING as the output delimiter\n\
184 the default is to use the input delimiter\n\
185 --help display this help and exit\n\
186 --version output version information and exit\n\
188 Use one, and only one of -b, -c or -f. Each LIST is made up of one\n\
189 range, or many ranges separated by commas. Each range is one of:\n\
191 N N'th byte, character or field, counted from 1\n\
192 N- from N'th byte, character or field, to end of line\n\
193 N-M from N'th to M'th (included) byte, character or field\n\
194 -M from first to M'th (included) byte, character or field\n\
196 With no FILE, or when FILE is -, read standard input.\n\
198 puts (_("\nReport bugs to <bug-textutils@gnu.org>."));
200 exit (status == 0 ? EXIT_SUCCESS : EXIT_FAILURE);
204 print_kth (unsigned int k)
206 return ((0 < eol_range_start && eol_range_start <= k)
207 || (k <= max_range_endpoint && printable_field[k]));
210 /* Given the list of field or byte range specifications FIELDSTR, set
211 MAX_RANGE_ENDPOINT and allocate and initialize the PRINTABLE_FIELD
212 array. If there is a right-open-ended range, set EOL_RANGE_START
213 to its starting index. FIELDSTR should be composed of one or more
214 numbers or ranges of numbers, separated by blanks or commas.
215 Incomplete ranges may be given: `-m' means `1-m'; `n-' means `n'
216 through end of line. Return nonzero if FIELDSTR contains at least
217 one field specification, zero otherwise. */
219 /* FIXME-someday: What if the user wants to cut out the 1,000,000-th field
220 of some huge input file? This function shouldn't have to alloate a table
221 of a million ints just so we can test every field < 10^6 with an array
222 dereference. Instead, consider using a dynamic hash table. It would be
223 simpler and nearly as good a solution to use a 32K x 4-byte table with
224 one bit per field index instead of a whole `int' per index. */
227 set_fields (const char *fieldstr)
229 unsigned int initial = 1; /* Value of first number in a range. */
230 unsigned int value = 0; /* If nonzero, a number being accumulated. */
231 int dash_found = 0; /* Nonzero if a '-' is found in this field. */
232 int field_found = 0; /* Non-zero if at least one field spec
233 has been processed. */
235 struct range_pair *rp;
237 unsigned int n_rp_allocated;
242 rp = (struct range_pair *) xmalloc (n_rp_allocated * sizeof (*rp));
244 /* Collect and store in RP the range end points.
245 It also sets EOL_RANGE_START if appropriate. */
249 if (*fieldstr == '-')
251 /* Starting a range. */
253 FATAL_ERROR (_("invalid byte or field list"));
265 else if (*fieldstr == ',' || ISBLANK (*fieldstr) || *fieldstr == '\0')
267 /* Ending the string, or this field/byte sublist. */
272 /* A range. Possibilites: -n, m-n, n-.
273 In any case, `initial' contains the start of the range. */
276 /* `n-'. From `initial' to end of line. */
277 eol_range_start = initial;
282 /* `m-n' or `-n' (1-n). */
284 FATAL_ERROR (_("invalid byte or field list"));
286 /* Is there already a range going to end of line? */
287 if (eol_range_start != 0)
289 /* Yes. Is the new sequence already contained
290 in the old one? If so, no processing is
292 if (initial < eol_range_start)
294 /* No, the new sequence starts before the
295 old. Does the old range going to end of line
296 extend into the new range? */
297 if (value + 1 >= eol_range_start)
299 /* Yes. Simply move the end of line marker. */
300 eol_range_start = initial;
304 /* No. A simple range, before and disjoint from
305 the range going to end of line. Fill it. */
306 ADD_RANGE_PAIR (rp, initial, value);
309 /* In any case, some fields were selected. */
315 /* There is no range going to end of line. */
316 ADD_RANGE_PAIR (rp, initial, value);
324 /* A simple field number, not a range. */
325 ADD_RANGE_PAIR (rp, value, value);
330 if (*fieldstr == '\0')
337 else if (ISDIGIT (*fieldstr))
339 /* FIXME: detect overflow? */
340 value = 10 * value + *fieldstr - '0';
344 FATAL_ERROR (_("invalid byte or field list"));
347 max_range_endpoint = 0;
348 for (i = 0; i < n_rp; i++)
350 if (rp[i].hi > max_range_endpoint)
351 max_range_endpoint = rp[i].hi;
354 /* Allocate an array large enough so that it may be indexed by
355 the field numbers corresponding to all finite ranges
356 (i.e. `2-6' or `-4', but not `5-') in FIELDSTR. */
358 printable_field = (int *) xmalloc ((max_range_endpoint + 1) * sizeof (int));
359 memset (printable_field, 0, (max_range_endpoint + 1) * sizeof (int));
361 /* Set the array entries corresponding to integers in the ranges of RP. */
362 for (i = 0; i < n_rp; i++)
365 for (j = rp[i].lo; j <= rp[i].hi; j++)
367 printable_field[j] = 1;
376 /* Read from stream STREAM, printing to standard output any selected bytes. */
379 cut_bytes (FILE *stream)
381 unsigned int byte_idx; /* Number of chars in the line so far. */
386 register int c; /* Each character from the file. */
404 if (print_kth (byte_idx))
412 /* Read from stream STREAM, printing to standard output any selected fields. */
415 cut_fields (FILE *stream)
418 unsigned int field_idx;
419 int found_any_selected_field;
420 int buffer_first_field;
423 found_any_selected_field = 0;
427 empty_input = (c == EOF);
431 /* To support the semantics of the -s flag, we may have to buffer
432 all of the first field to determine whether it is `delimited.'
433 But that is unnecessary if all non-delimited lines must be printed
434 and the first field has been selected, or if non-delimited lines
435 must be suppressed and the first field has *not* been selected.
436 That is because a non-delimited line has exactly one field. */
437 buffer_first_field = (suppress_non_delimited ^ !print_kth (1));
441 if (field_idx == 1 && buffer_first_field)
445 len = getstr (&field_1_buffer, &field_1_bufsize, stream,
449 if (ferror (stream) || feof (stream))
456 /* If the first field extends to the end of line (it is not
457 delimited) and we are printing all non-delimited lines,
459 if ((unsigned char) field_1_buffer[len - 1] != delim)
461 if (suppress_non_delimited)
467 fwrite (field_1_buffer, sizeof (char), len, stdout);
468 /* Make sure the output line is newline terminated. */
469 if (field_1_buffer[len - 1] != '\n')
476 /* Print the field, but not the trailing delimiter. */
477 fwrite (field_1_buffer, sizeof (char), len - 1, stdout);
478 found_any_selected_field = 1;
485 if (print_kth (field_idx))
487 if (found_any_selected_field)
489 fwrite (output_delimiter_string, sizeof (char),
490 output_delimiter_length, stdout);
492 found_any_selected_field = 1;
494 while ((c = getc (stream)) != delim && c != '\n' && c != EOF)
501 while ((c = getc (stream)) != delim && c != '\n' && c != EOF)
520 else if (c == '\n' || c == EOF)
522 if (found_any_selected_field
523 || (!empty_input && !(suppress_non_delimited && field_idx == 1)))
528 found_any_selected_field = 0;
534 cut_stream (FILE *stream)
536 if (operating_mode == byte_mode)
542 /* Process file FILE to standard output.
543 Return 0 if successful, 1 if not. */
546 cut_file (char *file)
550 if (STREQ (file, "-"))
557 stream = fopen (file, "r");
560 error (0, errno, "%s", file);
569 error (0, errno, "%s", file);
572 if (STREQ (file, "-"))
573 clearerr (stream); /* Also clear EOF. */
574 else if (fclose (stream) == EOF)
576 error (0, errno, "%s", file);
583 main (int argc, char **argv)
585 int optc, exit_status = 0;
586 int delim_specified = 0;
588 program_name = argv[0];
589 setlocale (LC_ALL, "");
590 bindtextdomain (PACKAGE, LOCALEDIR);
591 textdomain (PACKAGE);
593 atexit (close_stdout);
595 operating_mode = undefined_mode;
597 /* By default, all non-delimited lines are printed. */
598 suppress_non_delimited = 0;
603 while ((optc = getopt_long (argc, argv, "b:c:d:f:ns", longopts, NULL)) != -1)
612 /* Build the byte list. */
613 if (operating_mode != undefined_mode)
614 FATAL_ERROR (_("only one type of list may be specified"));
615 operating_mode = byte_mode;
616 if (set_fields (optarg) == 0)
617 FATAL_ERROR (_("missing list of positions"));
621 /* Build the field list. */
622 if (operating_mode != undefined_mode)
623 FATAL_ERROR (_("only one type of list may be specified"));
624 operating_mode = field_mode;
625 if (set_fields (optarg) == 0)
626 FATAL_ERROR (_("missing list of fields"));
631 /* Interpret -d '' to mean `use the NUL byte as the delimiter.' */
632 if (optarg[0] != '\0' && optarg[1] != '\0')
633 FATAL_ERROR (_("the delimiter must be a single character"));
634 delim = (unsigned char) optarg[0];
638 case OUTPUT_DELIMITER_OPTION:
639 /* Interpret --output-delimiter='' to mean
640 `use the NUL byte as the delimiter.' */
641 output_delimiter_length = (optarg[0] == '\0'
642 ? 1 : strlen (optarg));
643 output_delimiter_string = xstrdup (optarg);
650 suppress_non_delimited = 1;
653 case_GETOPT_HELP_CHAR;
655 case_GETOPT_VERSION_CHAR (PROGRAM_NAME, AUTHORS);
662 if (operating_mode == undefined_mode)
663 FATAL_ERROR (_("you must specify a list of bytes, characters, or fields"));
665 if (delim != '\0' && operating_mode != field_mode)
666 FATAL_ERROR (_("a delimiter may be specified only when operating on fields"));
668 if (suppress_non_delimited && operating_mode != field_mode)
669 FATAL_ERROR (_("suppressing non-delimited lines makes sense\n\
670 \tonly when operating on fields"));
672 if (!delim_specified)
675 if (output_delimiter_string == NULL)
677 static char dummy[2];
680 output_delimiter_string = dummy;
681 output_delimiter_length = 1;
685 exit_status |= cut_file ("-");
687 for (; optind < argc; optind++)
688 exit_status |= cut_file (argv[optind]);
690 if (have_read_stdin && fclose (stdin) == EOF)
692 error (0, errno, "-");
696 exit (exit_status == 0 ? EXIT_SUCCESS : EXIT_FAILURE);