aboutsummaryrefslogtreecommitdiff
path: root/coreutils/cut.c
diff options
context:
space:
mode:
authorMark Whitley <markw@lineo.com>2000-08-02 18:30:11 +0000
committerMark Whitley <markw@lineo.com>2000-08-02 18:30:11 +0000
commit807f0fd637ecc25e60c81e41ad969c177c7df327 (patch)
treef1c0645b6dc9c6c87e41280b822bc3b230e1f168 /coreutils/cut.c
parentd5fa3e3e9a8a1548b2e405ca9684bebbf946fd4f (diff)
downloadbusybox-w32-807f0fd637ecc25e60c81e41ad969c177c7df327.tar.gz
busybox-w32-807f0fd637ecc25e60c81e41ad969c177c7df327.tar.bz2
busybox-w32-807f0fd637ecc25e60c81e41ad969c177c7df327.zip
Brand, new version of cut. This fixes the bugs in the old cut, is smaller, and
is a sight easier to understand than the Minix cut.
Diffstat (limited to 'coreutils/cut.c')
-rw-r--r--coreutils/cut.c479
1 files changed, 171 insertions, 308 deletions
diff --git a/coreutils/cut.c b/coreutils/cut.c
index 4907ed935..e90a0e2e9 100644
--- a/coreutils/cut.c
+++ b/coreutils/cut.c
@@ -1,11 +1,8 @@
1/* vi: set sw=4 ts=4: */
2/* 1/*
3 * cut implementation for busybox 2 * cut.c - minimalist version of cut
4 * 3 *
5 * Copyright (c) Michael J. Holme 4 * Copyright (C) 1999,2000 by Lineo, inc.
6 * 5 * Written by Mark Whitley <markw@lineo.com>, <markw@enol.com>
7 * This version of cut is adapted from Minix cut and was modified
8 * by Erik Andersen <andersee@debian.org> to be used in busybox.
9 * 6 *
10 * This program is free software; you can redistribute it and/or modify 7 * This program is free software; you can redistribute it and/or modify
11 * it under the terms of the GNU General Public License as published by 8 * it under the terms of the GNU General Public License as published by
@@ -20,351 +17,217 @@
20 * You should have received a copy of the GNU General Public License 17 * You should have received a copy of the GNU General Public License
21 * along with this program; if not, write to the Free Software 18 * along with this program; if not, write to the Free Software
22 * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA 19 * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
23 * 20 *
24 * Original copyright notice is retained at the end of this file.
25 */ 21 */
26 22
27#include "internal.h" 23#include <stdio.h>
28#include <sys/types.h> 24#include <stdlib.h>
29#include <ctype.h> 25#include <unistd.h> /* getopt */
30#include <string.h> 26#include <string.h>
27#include <ctype.h>
31#include <errno.h> 28#include <errno.h>
32#include <stdlib.h> 29#include "internal.h"
33#include <stdio.h>
34#define BB_DECLARE_EXTERN
35#define bb_need_help
36#include "messages.c"
37
38#define MAX_FIELD 80 /* Pointers to the beginning of each field
39 * are stored in columns[], if a line holds
40 * more than MAX_FIELD columns the array
41 * boundary is exceed. But unlikely at 80 */
42
43#define MAX_ARGS 32 /* Maximum number of fields following -f or
44 * -c switches */
45int args[MAX_ARGS * 2];
46int num_args;
47
48/* Lots of new defines, should easen maintainance... */
49#define DUMP_STDIN 0 /* define for mode: no options */
50#define OPTIONF 1 /* define for mode: option -f */
51#define OPTIONC 2 /* define for mode: option -c */
52#define OPTIONB 3 /* define for mode: option -b */
53#define NOTSET 0 /* option not selected */
54#define SET 1 /* option selected */
55#define OPTIONS 1 /*define option -s */
56/* Defines for the warnings */
57#define DELIMITER_NOT_APPLICABLE 0
58#define OVERRIDING_PREVIOUS_MODE 1
59#define OPTION_NOT_APPLICABLE 2
60#define UNKNOWN_OPTION 3
61#define FILE_NOT_READABLE 4
62/* Defines for the fatal errors */
63#define SYNTAX_ERROR 101
64#define POSITION_ERROR 102
65#define LINE_TO_LONG_ERROR 103
66#define RANGE_ERROR 104
67#define MAX_FIELDS_EXEEDED_ERROR 105
68#define MAX_ARGS_EXEEDED_ERROR 106
69
70
71int mode; /* 0 = dump stdin to stdout, 1=-f, 2=-c */
72char delim = '\t'; /* default delimiting character */
73FILE *fd;
74char line[BUFSIZ];
75int exit_status;
76int option = 0; /* for -s option */
77
78int cut_main(int argc, char **argv);
79void warn(int warn_number, char *option);
80void cuterror(int err);
81void get_args(void);
82void cut(void);
83 30
84void warn(int warn_number, char *option)
85{
86 static char *warn_msg[] = {
87 "Option -%s allowed only with -f\n",
88 "-%s overrides earlier option\n",
89 "-%s not allowed in current mode\n",
90 "Cannot open %s\n"
91 };
92 31
93 errorMsg(warn_msg[warn_number], option); 32/* globals from other files */
94 exit_status = warn_number + 1; 33extern int optind;
34extern char *optarg;
95 35
96}
97
98void cuterror(int err)
99{
100 static char *err_mes[] = {
101 "syntax error\n",
102 "position must be >0\n",
103 "line longer than BUFSIZ\n",
104 "range must not decrease from left to right\n",
105 "MAX_FIELD exceeded\n",
106 "MAX_ARGS exceeded\n"
107 };
108 36
109 errorMsg(err_mes[err - 101]); 37/* globals in this file only */
110 exit(err); 38static char part = 0; /* (b)yte, (c)har, (f)ields */
111} 39static int startpos = 1;
40static int endpos = -1;
41static char delim = '\t'; /* delimiter, default is tab */
42static unsigned int supress_non_delimited_lines = 0;
112 43
113 44
114void get_args() 45/*
46 * decompose_list() - parses a list and puts values into startpos and endpos.
47 * valid list formats: N, N-, N-M, -M
48 */
49static void decompose_list(const char *list)
115{ 50{
116 int i = 0; 51 unsigned int nminus = 0;
117 int arg_ptr = 0; 52 char *ptr;
118 int flag;
119 53
120 num_args = 0; 54 /* the list must contain only digits and no more than one minus sign */
121 do { 55 for (ptr = (char *)list; *ptr; ptr++) {
122 if (num_args == MAX_ARGS) 56 if (!isdigit(*ptr) && *ptr != '-') {
123 cuterror(MAX_ARGS_EXEEDED_ERROR); 57 fatalError("invalid byte or field list\n");
124 if (!isdigit(line[i]) && line[i] != '-') 58 }
125 cuterror(SYNTAX_ERROR); 59 if (*ptr == '-') {
60 nminus++;
61 if (nminus > 1) {
62 fatalError("invalid byte or field list\n");
63 }
64 }
65 }
126 66
127 args[arg_ptr] = 1; 67 /* handle single value 'N' case */
128 args[arg_ptr + 1] = BUFSIZ; 68 if (nminus == 0) {
129 flag = 1; 69 startpos = strtol(list, &ptr, 10);
70 if (startpos == 0) {
71 fatalError("missing list of fields\n");
72 }
73 endpos = startpos;
74 }
75 /* handle multi-value cases */
76 else if (nminus == 1) {
77 /* handle 'N-' case */
78 if (list[strlen(list) - 1] == '-') {
79 startpos = strtol(list, &ptr, 10);
80 }
81 /* handle '-M' case */
82 else if (list[0] == '-') {
83 endpos = strtol(&list[1], NULL, 10);
84 }
85 /* handle 'N-M' case */
86 else {
87 startpos = strtol(list, &ptr, 10);
88 endpos = strtol(ptr+1, &ptr, 10);
89 }
130 90
131 while (line[i] != ',' && line[i] != 0) { 91 /* a sanity check */
132 if (isdigit(line[i])) { 92 if (startpos == 0) {
133 args[arg_ptr] = 0; 93 startpos = 1;
134 while (isdigit(line[i]))
135 args[arg_ptr] = 10 * args[arg_ptr] + line[i++] - '0';
136 if (!args[arg_ptr])
137 cuterror(POSITION_ERROR);
138 arg_ptr++;
139 }
140 if (line[i] == '-') {
141 arg_ptr |= 1;
142 i++;
143 flag = 0;
144 }
145 } 94 }
146 if (flag && arg_ptr & 1)
147 args[arg_ptr] = args[arg_ptr - 1];
148 if (args[num_args * 2] > args[num_args * 2 + 1])
149 cuterror(RANGE_ERROR);
150 num_args++;
151 arg_ptr = num_args * 2;
152 } 95 }
153 while (line[i++]);
154} 96}
155 97
156 98
157void cut() 99/*
100 * snippy-snip
101 */
102static void cut_file(FILE *file)
158{ 103{
159 int i, j, length, maxcol=0; 104 char *line;
160 char *columns[MAX_FIELD]; 105
161 106 /* go through every line in the file */
162 while (fgets(line, BUFSIZ, fd)) { 107 for (line = NULL; (line = get_line_from_file(file)) != NULL; free(line)) {
163 maxcol=0; 108
164 length = strlen(line) - 1; 109 /* cut based on chars/bytes */
165 *(line + length) = 0; 110 if (part == 'c' || part == 'b') {
166 switch (mode) { 111 int i;
167 case DUMP_STDIN: 112 /* a valid end position has been specified */
168 printf("%s", line); 113 if (endpos > 0) {
169 break; 114 for (i = startpos-1; i < endpos; i++) {
170 case OPTIONF: 115 fputc(line[i], stdout);
171 columns[maxcol++] = line;
172 for (i = 0; i < length; i++) {
173 if (*(line + i) == delim) {
174 *(line + i) = 0;
175 if (maxcol == MAX_FIELD)
176 cuterror(MAX_FIELDS_EXEEDED_ERROR);
177 columns[maxcol] = line + i + 1;
178 maxcol++;
179 } 116 }
117 fputc('\n', stdout);
180 } 118 }
181 if (maxcol != 1) { 119 /* otherwise, just go to the end of the line */
182 for (i = 0; i < num_args; i++) { 120 else {
183 for (j = args[i * 2]; j <= args[i * 2 + 1]; j++) 121 for (i = startpos-1; line[i]; i++) {
184 if (j <= maxcol) { 122 fputc(line[i], stdout);
185
186 printf("%s", columns[j - 1]);
187
188 if (i != num_args - 1 || j != args[i * 2 + 1])
189 putchar(delim);
190 }
191 } 123 }
192 } else if (option != OPTIONS) {
193 printf("%s",line);
194 } 124 }
195 break; 125 }
196 case OPTIONC: 126 /* cut based on fields */
197 for (i = 0; i < num_args; i++) { 127 else if (part == 'f') {
198 for (j = args[i * 2]; 128 char *ptr;
199 j <= (args[i * 2 + 1] > 129 char *start = line;
200 length ? length : args[i * 2 + 1]); j++) 130 unsigned int delims_hit = 0;
201 putchar(*(line + j - 1)); 131
132 for (ptr = line; (ptr = strchr(ptr, delim)) != NULL; ptr++) {
133 delims_hit++;
134 if (delims_hit == (startpos - 1)) {
135 start = ptr+1;
136 }
137 if (delims_hit == endpos) {
138 break;
139 }
140 }
141 /* we didn't hit any delimeters */
142 if (delims_hit == 0 && !supress_non_delimited_lines) {
143 fputs(line, stdout);
144 }
145 /* we =did= hit some delimiters */
146 else if (delims_hit > 0) {
147 /* we have a fixed end point */
148 if (ptr) {
149 while (start < ptr) {
150 fputc(*start, stdout);
151 start++;
152 }
153 fputc('\n', stdout);
154 }
155 /* or we're just going til the end of the line */
156 else {
157 while (*start) {
158 fputc(*start, stdout);
159 start++;
160 }
161 }
202 } 162 }
203 } 163 }
204 if (maxcol != 1)
205 putchar('\n');
206 } 164 }
207} 165}
208 166
209int cut_main(int argc, char **argv) 167extern int cut_main(int argc, char **argv)
210{ 168{
211 int i = 1; 169 int opt;
212 int numberFilenames = 0;
213 170
214 while (i < argc) { 171 while ((opt = getopt(argc, argv, "b:c:d:f:ns")) > 0) {
215 if (argv[i][0] == '-') { 172 switch (opt) {
216 switch (argv[i++][1]) {
217 case 'd':
218 if (mode == OPTIONC || mode == OPTIONB)
219 warn(DELIMITER_NOT_APPLICABLE, "d");
220 if (argc > i)
221 delim = argv[i++][0];
222 else
223 cuterror(SYNTAX_ERROR);
224 break;
225 case 'f':
226 sprintf(line, "%s", argv[i++]);
227 if (mode == OPTIONC || mode == OPTIONB)
228 warn(OVERRIDING_PREVIOUS_MODE, "f");
229 mode = OPTIONF;
230 break;
231 case 'b': 173 case 'b':
232 sprintf(line, "%s", argv[i++]);
233 if (mode == OPTIONF || mode == OPTIONC)
234 warn(OVERRIDING_PREVIOUS_MODE, "b");
235 mode = OPTIONB;
236 break;
237 case 'c': 174 case 'c':
238 sprintf(line, "%s", argv[i++]); 175 case 'f':
239 if (mode == OPTIONF || mode == OPTIONB) 176 /* make sure they didn't ask for two types of lists */
240 warn(OVERRIDING_PREVIOUS_MODE, "c"); 177 if (part != 0) {
241 mode = OPTIONC; 178 fatalError("only one type of list may be specified");
179 }
180 part = (char)opt;
181 decompose_list(optarg);
242 break; 182 break;
243 case 's': 183 case 'd':
244 option = OPTIONS; 184 if (strlen(optarg) > 1) {
245 185 fatalError("the delimiter must be a single character\n");
186 }
187 delim = optarg[0];
246 break; 188 break;
247 case '\0': /* - means: read from stdin */ 189 case 'n':
248 numberFilenames++; 190 /* no-op */
249 break; 191 break;
250 case 'n': /* needed for Posix, but no effect here */ 192 case 's':
251 if (mode != OPTIONB) 193 supress_non_delimited_lines++;
252 warn(OPTION_NOT_APPLICABLE, "n");
253 break; 194 break;
254 default:
255 warn(UNKNOWN_OPTION, &(argv[i - 1][1]));
256 }
257 } else {
258 i++;
259 numberFilenames++;
260 } 195 }
261 } 196 }
262 197
263/* Here follow the checks, if the selected options are reasonable. */ 198 if (part == 0) {
264 if (mode == OPTIONB) /* since in Minix char := byte */ 199 fatalError("you must specify a list of bytes, characters, or fields\n");
265 mode = OPTIONC; 200 }
266 201
267 if (mode != OPTIONF && option == OPTIONS) 202 if (supress_non_delimited_lines && part != 'f') {
268 warn(DELIMITER_NOT_APPLICABLE,"s"); 203 fatalError("suppressing non-delimited lines makes sense
269 get_args(); 204 only when operating on fields\n");
270 if (numberFilenames != 0) { 205 }
271 i = 1; 206
272 while (i < argc) { 207 if (delim != '\t' && part != 'f') {
273 if (argv[i][0] == '-') { 208 fatalError("a delimiter may be specified only when operating on fields\n");
274 switch (argv[i][1]) { 209 }
275 case 'f': 210
276 case 'c': 211 /* argv[(optind)..(argc-1)] should be names of file to process. If no
277 case 'b': 212 * files were specified or '-' was specified, take input from stdin.
278 case 'd': 213 * Otherwise, we process all the files specified. */
279 i += 2; 214 if (argv[optind] == NULL || (strcmp(argv[optind], "-") == 0)) {
280 break; 215 cut_file(stdin);
281 case 'n': 216 }
282 case 'i': 217 else {
283 case 's': 218 int i;
284 i++; 219 FILE *file;
285 break; 220 for (i = optind; i < argc; i++) {
286 case '\0': 221 file = fopen(argv[i], "r");
287 fd = stdin; 222 if (file == NULL) {
288 i++; 223 /* errorMsg("%s: %s\n", argv[i], strerror(errno)); */
289 cut(); 224 fprintf(stderr, "%s: %s\n", argv[i], strerror(errno));
290 break;
291 default:
292 i++;
293 }
294 } else { 225 } else {
295 if ((fd = fopen(argv[i++], "r")) == NULL) { 226 cut_file(file);
296 warn(FILE_NOT_READABLE, argv[i - 1]); 227 fclose(file);
297 } else {
298 cut();
299 fclose(fd);
300 }
301 } 228 }
302 } 229 }
303 } else {
304 fd = stdin;
305 cut();
306 } 230 }
307 231
308 return(exit_status); 232 return 0;
309} 233}
310
311/* cut - extract columns from a file or stdin. Author: Michael J. Holme
312 *
313 * Copyright 1989, Michael John Holme, All rights reserved.
314 * This code may be freely distributed, provided that this notice
315 * remains intact.
316 *
317 * V1.1: 6th September 1989
318 *
319 * Bugs, criticisms, etc,
320 * c/o Mark Powell
321 * JANET sq79@uk.ac.liv
322 * ARPA sq79%liv.ac.uk@nsfnet-relay.ac.uk
323 * UUCP ...!mcvax!ukc!liv.ac.uk!sq79
324 *-------------------------------------------------------------------------
325 * Changed for POSIX1003.2/Draft10 conformance
326 * Thomas Brupbacher (tobr@mw.lpc.ethz.ch), September 1990.
327 * Changes:
328 * - separation of error messages ( stderr) and output (stdout).
329 * - support for -b and -n (no effect, -b acts as -c)
330 * - support for -s
331 *-------------------------------------------------------------------------
332 */
333
334/*
335 * Copyright (c) 1987,1997, Prentice Hall
336 * All rights reserved.
337 *
338 * Redistribution and use of the MINIX operating system in source and
339 * binary forms, with or without modification, are permitted provided
340 * that the following conditions are met:
341 *
342 * Redistributions of source code must retain the above copyright
343 * notice, this list of conditions and the following disclaimer.
344 *
345 * Redistributions in binary form must reproduce the above
346 * copyright notice, this list of conditions and the following
347 * disclaimer in the documentation and/or other materials provided
348 * with the distribution.
349 *
350 * Neither the name of Prentice Hall nor the names of the software
351 * authors or contributors may be used to endorse or promote
352 * products derived from this software without specific prior
353 * written permission.
354 *
355 * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS, AUTHORS, AND
356 * CONTRIBUTORS ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES,
357 * INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF
358 * MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
359 * IN NO EVENT SHALL PRENTICE HALL OR ANY AUTHORS OR CONTRIBUTORS BE
360 * LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
361 * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
362 * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR
363 * BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY,
364 * WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE
365 * OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE,
366 * EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
367 *
368 */
369
370
generated by cgit v1.2.3-55-g6feb (git 2.39.1) at 2025-03-17 23:51:44 +0000