Eric Andersen | c9e7024 | 2003-06-20 09:16:00 +0000 | [diff] [blame] | 1 | /* vi: set sw=8 ts=8: */ |
Erik Andersen | 7ab9c7e | 2000-05-12 19:41:47 +0000 | [diff] [blame] | 2 | /* |
Mark Whitley | 807f0fd | 2000-08-02 18:30:11 +0000 | [diff] [blame] | 3 | * cut.c - minimalist version of cut |
Erik Andersen | 7ab9c7e | 2000-05-12 19:41:47 +0000 | [diff] [blame] | 4 | * |
Eric Andersen | 8ec10a9 | 2001-01-27 09:33:39 +0000 | [diff] [blame] | 5 | * Copyright (C) 1999,2000,2001 by Lineo, inc. |
Eric Andersen | c7bda1c | 2004-03-15 08:29:22 +0000 | [diff] [blame] | 6 | * Written by Mark Whitley <markw@codepoet.org> |
Erik Andersen | 7ab9c7e | 2000-05-12 19:41:47 +0000 | [diff] [blame] | 7 | * |
| 8 | * This program is free software; you can redistribute it and/or modify |
| 9 | * it under the terms of the GNU General Public License as published by |
| 10 | * the Free Software Foundation; either version 2 of the License, or |
| 11 | * (at your option) any later version. |
| 12 | * |
| 13 | * This program is distributed in the hope that it will be useful, |
| 14 | * but WITHOUT ANY WARRANTY; without even the implied warranty of |
| 15 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU |
| 16 | * General Public License for more details. |
| 17 | * |
| 18 | * You should have received a copy of the GNU General Public License |
| 19 | * along with this program; if not, write to the Free Software |
| 20 | * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA |
Mark Whitley | 807f0fd | 2000-08-02 18:30:11 +0000 | [diff] [blame] | 21 | * |
Erik Andersen | 7ab9c7e | 2000-05-12 19:41:47 +0000 | [diff] [blame] | 22 | */ |
| 23 | |
Erik Andersen | 7ab9c7e | 2000-05-12 19:41:47 +0000 | [diff] [blame] | 24 | #include <stdio.h> |
Mark Whitley | 807f0fd | 2000-08-02 18:30:11 +0000 | [diff] [blame] | 25 | #include <stdlib.h> |
Eric Andersen | 8876fb2 | 2003-06-20 09:01:58 +0000 | [diff] [blame] | 26 | #include <unistd.h> |
Mark Whitley | 807f0fd | 2000-08-02 18:30:11 +0000 | [diff] [blame] | 27 | #include <string.h> |
Mark Whitley | b696763 | 2001-05-18 23:04:51 +0000 | [diff] [blame] | 28 | #include <limits.h> |
Eric Andersen | 3570a34 | 2000-09-25 21:45:58 +0000 | [diff] [blame] | 29 | #include "busybox.h" |
Erik Andersen | 7ab9c7e | 2000-05-12 19:41:47 +0000 | [diff] [blame] | 30 | |
| 31 | |
Mark Whitley | b696763 | 2001-05-18 23:04:51 +0000 | [diff] [blame] | 32 | /* option vars */ |
Eric Andersen | 8876fb2 | 2003-06-20 09:01:58 +0000 | [diff] [blame] | 33 | static const char optstring[] = "b:c:f:d:sn"; |
| 34 | #define OPT_BYTE_FLGS 1 |
| 35 | #define OPT_CHAR_FLGS 2 |
| 36 | #define OPT_FIELDS_FLGS 4 |
| 37 | #define OPT_DELIM_FLGS 8 |
| 38 | #define OPT_SUPRESS_FLGS 16 |
| 39 | static char part; /* (b)yte, (c)har, (f)ields */ |
| 40 | static unsigned int supress_non_delimited_lines; |
Mark Whitley | b696763 | 2001-05-18 23:04:51 +0000 | [diff] [blame] | 41 | static char delim = '\t'; /* delimiter, default is tab */ |
| 42 | |
| 43 | struct cut_list { |
| 44 | int startpos; |
| 45 | int endpos; |
| 46 | }; |
| 47 | |
| 48 | static const int BOL = 0; |
| 49 | static const int EOL = INT_MAX; |
| 50 | static const int NON_RANGE = -1; |
| 51 | |
| 52 | static struct cut_list *cut_lists = NULL; /* growable array holding a series of lists */ |
| 53 | static unsigned int nlists = 0; /* number of elements in above list */ |
| 54 | |
| 55 | |
| 56 | static int cmpfunc(const void *a, const void *b) |
| 57 | { |
| 58 | struct cut_list *la = (struct cut_list *)a; |
| 59 | struct cut_list *lb = (struct cut_list *)b; |
| 60 | |
| 61 | if (la->startpos > lb->startpos) |
| 62 | return 1; |
| 63 | if (la->startpos < lb->startpos) |
| 64 | return -1; |
| 65 | return 0; |
| 66 | } |
Erik Andersen | 7ab9c7e | 2000-05-12 19:41:47 +0000 | [diff] [blame] | 67 | |
Erik Andersen | 7ab9c7e | 2000-05-12 19:41:47 +0000 | [diff] [blame] | 68 | |
| 69 | /* |
Mark Whitley | b696763 | 2001-05-18 23:04:51 +0000 | [diff] [blame] | 70 | * parse_lists() - parses a list and puts values into startpos and endpos. |
Eric Andersen | c7bda1c | 2004-03-15 08:29:22 +0000 | [diff] [blame] | 71 | * valid list formats: N, N-, N-M, -M |
Eric Andersen | aff114c | 2004-04-14 17:51:38 +0000 | [diff] [blame] | 72 | * more than one list can be separated by commas |
Erik Andersen | 7ab9c7e | 2000-05-12 19:41:47 +0000 | [diff] [blame] | 73 | */ |
Mark Whitley | b696763 | 2001-05-18 23:04:51 +0000 | [diff] [blame] | 74 | static void parse_lists(char *lists) |
Mark Whitley | 807f0fd | 2000-08-02 18:30:11 +0000 | [diff] [blame] | 75 | { |
Mark Whitley | b696763 | 2001-05-18 23:04:51 +0000 | [diff] [blame] | 76 | char *ltok = NULL; |
| 77 | char *ntok = NULL; |
| 78 | char *junk; |
| 79 | int s = 0, e = 0; |
Mark Whitley | 807f0fd | 2000-08-02 18:30:11 +0000 | [diff] [blame] | 80 | |
Eric Andersen | aff114c | 2004-04-14 17:51:38 +0000 | [diff] [blame] | 81 | /* take apart the lists, one by one (they are separated with commas */ |
Mark Whitley | b696763 | 2001-05-18 23:04:51 +0000 | [diff] [blame] | 82 | while ((ltok = strsep(&lists, ",")) != NULL) { |
| 83 | |
| 84 | /* it's actually legal to pass an empty list */ |
| 85 | if (strlen(ltok) == 0) |
| 86 | continue; |
| 87 | |
| 88 | /* get the start pos */ |
| 89 | ntok = strsep(<ok, "-"); |
| 90 | if (ntok == NULL) { |
| 91 | fprintf(stderr, "Help ntok is null for starting position! What do I do?\n"); |
| 92 | } else if (strlen(ntok) == 0) { |
| 93 | s = BOL; |
| 94 | } else { |
| 95 | s = strtoul(ntok, &junk, 10); |
| 96 | if(*junk != '\0' || s < 0) |
Manuel Novoa III | cad5364 | 2003-03-19 09:13:01 +0000 | [diff] [blame] | 97 | bb_error_msg_and_die("invalid byte or field list"); |
Eric Andersen | c7bda1c | 2004-03-15 08:29:22 +0000 | [diff] [blame] | 98 | |
Mark Whitley | b696763 | 2001-05-18 23:04:51 +0000 | [diff] [blame] | 99 | /* account for the fact that arrays are zero based, while the user |
| 100 | * expects the first char on the line to be char # 1 */ |
| 101 | if (s != 0) |
| 102 | s--; |
| 103 | } |
| 104 | |
| 105 | /* get the end pos */ |
| 106 | ntok = strsep(<ok, "-"); |
| 107 | if (ntok == NULL) { |
| 108 | e = NON_RANGE; |
| 109 | } else if (strlen(ntok) == 0) { |
| 110 | e = EOL; |
| 111 | } else { |
| 112 | e = strtoul(ntok, &junk, 10); |
| 113 | if(*junk != '\0' || e < 0) |
Manuel Novoa III | cad5364 | 2003-03-19 09:13:01 +0000 | [diff] [blame] | 114 | bb_error_msg_and_die("invalid byte or field list"); |
Mark Whitley | b696763 | 2001-05-18 23:04:51 +0000 | [diff] [blame] | 115 | /* if the user specified and end position of 0, that means "til the |
| 116 | * end of the line */ |
| 117 | if (e == 0) |
| 118 | e = INT_MAX; |
| 119 | e--; /* again, arrays are zero based, lines are 1 based */ |
| 120 | if (e == s) |
| 121 | e = NON_RANGE; |
| 122 | } |
| 123 | |
| 124 | /* if there's something left to tokenize, the user past an invalid list */ |
| 125 | if (ltok) |
Manuel Novoa III | cad5364 | 2003-03-19 09:13:01 +0000 | [diff] [blame] | 126 | bb_error_msg_and_die("invalid byte or field list"); |
Eric Andersen | c7bda1c | 2004-03-15 08:29:22 +0000 | [diff] [blame] | 127 | |
Mark Whitley | b696763 | 2001-05-18 23:04:51 +0000 | [diff] [blame] | 128 | /* add the new list */ |
| 129 | cut_lists = xrealloc(cut_lists, sizeof(struct cut_list) * (++nlists)); |
| 130 | cut_lists[nlists-1].startpos = s; |
| 131 | cut_lists[nlists-1].endpos = e; |
| 132 | } |
| 133 | |
| 134 | /* make sure we got some cut positions out of all that */ |
| 135 | if (nlists == 0) |
Manuel Novoa III | cad5364 | 2003-03-19 09:13:01 +0000 | [diff] [blame] | 136 | bb_error_msg_and_die("missing list of positions"); |
Mark Whitley | b696763 | 2001-05-18 23:04:51 +0000 | [diff] [blame] | 137 | |
| 138 | /* now that the lists are parsed, we need to sort them to make life easier |
| 139 | * on us when it comes time to print the chars / fields / lines */ |
| 140 | qsort(cut_lists, nlists, sizeof(struct cut_list), cmpfunc); |
| 141 | |
| 142 | } |
| 143 | |
| 144 | |
| 145 | static void cut_line_by_chars(const char *line) |
| 146 | { |
| 147 | int c, l; |
| 148 | /* set up a list so we can keep track of what's been printed */ |
| 149 | char *printed = xcalloc(strlen(line), sizeof(char)); |
| 150 | |
| 151 | /* print the chars specified in each cut list */ |
| 152 | for (c = 0; c < nlists; c++) { |
| 153 | l = cut_lists[c].startpos; |
| 154 | while (l < strlen(line)) { |
| 155 | if (!printed[l]) { |
| 156 | putchar(line[l]); |
| 157 | printed[l] = 'X'; |
Mark Whitley | 807f0fd | 2000-08-02 18:30:11 +0000 | [diff] [blame] | 158 | } |
Mark Whitley | b696763 | 2001-05-18 23:04:51 +0000 | [diff] [blame] | 159 | l++; |
| 160 | if (cut_lists[c].endpos == NON_RANGE || l > cut_lists[c].endpos) |
| 161 | break; |
| 162 | } |
| 163 | } |
| 164 | putchar('\n'); /* cuz we were handed a chomped line */ |
| 165 | free(printed); |
| 166 | } |
| 167 | |
| 168 | |
| 169 | static void cut_line_by_fields(char *line) |
| 170 | { |
| 171 | int c, f; |
| 172 | int ndelim = -1; /* zero-based / one-based problem */ |
| 173 | int nfields_printed = 0; |
| 174 | char *field = NULL; |
| 175 | char d[2] = { delim, 0 }; |
| 176 | char *printed; |
| 177 | |
| 178 | /* test the easy case first: does this line contain any delimiters? */ |
| 179 | if (strchr(line, delim) == NULL) { |
| 180 | if (!supress_non_delimited_lines) |
| 181 | puts(line); |
| 182 | return; |
| 183 | } |
| 184 | |
| 185 | /* set up a list so we can keep track of what's been printed */ |
| 186 | printed = xcalloc(strlen(line), sizeof(char)); |
| 187 | |
| 188 | /* process each list on this line, for as long as we've got a line to process */ |
| 189 | for (c = 0; c < nlists && line; c++) { |
| 190 | f = cut_lists[c].startpos; |
| 191 | do { |
| 192 | |
| 193 | /* find the field we're looking for */ |
| 194 | while (line && ndelim < f) { |
| 195 | field = strsep(&line, d); |
| 196 | ndelim++; |
| 197 | } |
| 198 | |
| 199 | /* we found it, and it hasn't been printed yet */ |
| 200 | if (field && ndelim == f && !printed[ndelim]) { |
| 201 | /* if this isn't our first time through, we need to print the |
| 202 | * delimiter after the last field that was printed */ |
| 203 | if (nfields_printed > 0) |
| 204 | putchar(delim); |
| 205 | fputs(field, stdout); |
| 206 | printed[ndelim] = 'X'; |
| 207 | nfields_printed++; |
| 208 | } |
| 209 | |
| 210 | f++; |
| 211 | |
| 212 | /* keep going as long as we have a line to work with, this is a |
| 213 | * list, and we're not at the end of that list */ |
| 214 | } while (line && cut_lists[c].endpos != NON_RANGE && f <= cut_lists[c].endpos); |
| 215 | } |
| 216 | |
| 217 | /* if we printed anything at all, we need to finish it with a newline cuz |
| 218 | * we were handed a chomped line */ |
| 219 | putchar('\n'); |
| 220 | |
| 221 | free(printed); |
| 222 | } |
| 223 | |
| 224 | |
| 225 | static void cut_file_by_lines(const char *line, unsigned int linenum) |
| 226 | { |
| 227 | static int c = 0; |
| 228 | static int l = -1; |
Eric Andersen | c7bda1c | 2004-03-15 08:29:22 +0000 | [diff] [blame] | 229 | |
Mark Whitley | b696763 | 2001-05-18 23:04:51 +0000 | [diff] [blame] | 230 | /* I can't initialize this above cuz the "initializer isn't |
| 231 | * constant" *sigh* */ |
| 232 | if (l == -1) |
| 233 | l = cut_lists[c].startpos; |
| 234 | |
| 235 | /* get out if we have no more lists to process or if the lines are lower |
| 236 | * than what we're interested in */ |
| 237 | if (c >= nlists || linenum < l) |
| 238 | return; |
| 239 | |
| 240 | /* if the line we're looking for is lower than the one we were passed, it |
| 241 | * means we displayed it already, so move on */ |
| 242 | while (l < linenum) { |
| 243 | l++; |
| 244 | /* move on to the next list if we're at the end of this one */ |
| 245 | if (cut_lists[c].endpos == NON_RANGE || l > cut_lists[c].endpos) { |
| 246 | c++; |
| 247 | /* get out if there's no more lists to process */ |
| 248 | if (c >= nlists) |
| 249 | return; |
| 250 | l = cut_lists[c].startpos; |
| 251 | /* get out if the current line is lower than the one we just became |
| 252 | * interested in */ |
| 253 | if (linenum < l) |
| 254 | return; |
Mark Whitley | 807f0fd | 2000-08-02 18:30:11 +0000 | [diff] [blame] | 255 | } |
| 256 | } |
| 257 | |
Mark Whitley | b696763 | 2001-05-18 23:04:51 +0000 | [diff] [blame] | 258 | /* If we made it here, it means we've found the line we're looking for, so print it */ |
| 259 | puts(line); |
Mark Whitley | 807f0fd | 2000-08-02 18:30:11 +0000 | [diff] [blame] | 260 | } |
Erik Andersen | 7ab9c7e | 2000-05-12 19:41:47 +0000 | [diff] [blame] | 261 | |
| 262 | |
Mark Whitley | 807f0fd | 2000-08-02 18:30:11 +0000 | [diff] [blame] | 263 | /* |
| 264 | * snippy-snip |
| 265 | */ |
| 266 | static void cut_file(FILE *file) |
| 267 | { |
Mark Whitley | b696763 | 2001-05-18 23:04:51 +0000 | [diff] [blame] | 268 | char *line = NULL; |
| 269 | unsigned int linenum = 0; /* keep these zero-based to be consistent */ |
Mark Whitley | 807f0fd | 2000-08-02 18:30:11 +0000 | [diff] [blame] | 270 | |
| 271 | /* go through every line in the file */ |
Manuel Novoa III | cad5364 | 2003-03-19 09:13:01 +0000 | [diff] [blame] | 272 | while ((line = bb_get_chomped_line_from_file(file)) != NULL) { |
Mark Whitley | b696763 | 2001-05-18 23:04:51 +0000 | [diff] [blame] | 273 | |
| 274 | /* cut based on chars/bytes XXX: only works when sizeof(char) == byte */ |
Eric Andersen | 8876fb2 | 2003-06-20 09:01:58 +0000 | [diff] [blame] | 275 | if ((part & (OPT_CHAR_FLGS | OPT_BYTE_FLGS))) |
Mark Whitley | b696763 | 2001-05-18 23:04:51 +0000 | [diff] [blame] | 276 | cut_line_by_chars(line); |
| 277 | |
Mark Whitley | 807f0fd | 2000-08-02 18:30:11 +0000 | [diff] [blame] | 278 | /* cut based on fields */ |
Eric Andersen | 8876fb2 | 2003-06-20 09:01:58 +0000 | [diff] [blame] | 279 | else { |
Mark Whitley | b696763 | 2001-05-18 23:04:51 +0000 | [diff] [blame] | 280 | if (delim == '\n') |
| 281 | cut_file_by_lines(line, linenum); |
| 282 | else |
| 283 | cut_line_by_fields(line); |
Mark Whitley | 807f0fd | 2000-08-02 18:30:11 +0000 | [diff] [blame] | 284 | } |
Mark Whitley | b696763 | 2001-05-18 23:04:51 +0000 | [diff] [blame] | 285 | |
| 286 | linenum++; |
| 287 | free(line); |
Mark Whitley | 807f0fd | 2000-08-02 18:30:11 +0000 | [diff] [blame] | 288 | } |
| 289 | } |
| 290 | |
Mark Whitley | b696763 | 2001-05-18 23:04:51 +0000 | [diff] [blame] | 291 | |
Mark Whitley | 807f0fd | 2000-08-02 18:30:11 +0000 | [diff] [blame] | 292 | extern int cut_main(int argc, char **argv) |
| 293 | { |
Eric Andersen | 8876fb2 | 2003-06-20 09:01:58 +0000 | [diff] [blame] | 294 | unsigned long opt; |
| 295 | char *sopt, *sdopt; |
Mark Whitley | 807f0fd | 2000-08-02 18:30:11 +0000 | [diff] [blame] | 296 | |
"Vladimir N. Oleynik" | f704b27 | 2005-10-14 09:56:52 +0000 | [diff] [blame] | 297 | bb_opt_complementally = "b--bcf:c--bcf:f--bcf"; |
Eric Andersen | 8876fb2 | 2003-06-20 09:01:58 +0000 | [diff] [blame] | 298 | opt = bb_getopt_ulflags(argc, argv, optstring, &sopt, &sopt, &sopt, &sdopt); |
| 299 | part = opt & (OPT_BYTE_FLGS|OPT_CHAR_FLGS|OPT_FIELDS_FLGS); |
| 300 | if(part == 0) |
| 301 | bb_error_msg_and_die("you must specify a list of bytes, characters, or fields"); |
Mike Frysinger | 348e84c | 2005-05-11 00:39:03 +0000 | [diff] [blame] | 302 | if(opt & BB_GETOPT_ERROR) |
Eric Andersen | c9e7024 | 2003-06-20 09:16:00 +0000 | [diff] [blame] | 303 | bb_error_msg_and_die("only one type of list may be specified"); |
Eric Andersen | 8876fb2 | 2003-06-20 09:01:58 +0000 | [diff] [blame] | 304 | parse_lists(sopt); |
| 305 | if((opt & (OPT_DELIM_FLGS))) { |
| 306 | if (strlen(sdopt) > 1) { |
Eric Andersen | c9e7024 | 2003-06-20 09:16:00 +0000 | [diff] [blame] | 307 | bb_error_msg_and_die("the delimiter must be a single character"); |
| 308 | } |
Eric Andersen | 8876fb2 | 2003-06-20 09:01:58 +0000 | [diff] [blame] | 309 | delim = sdopt[0]; |
Mark Whitley | 807f0fd | 2000-08-02 18:30:11 +0000 | [diff] [blame] | 310 | } |
Eric Andersen | 8876fb2 | 2003-06-20 09:01:58 +0000 | [diff] [blame] | 311 | supress_non_delimited_lines = opt & OPT_SUPRESS_FLGS; |
Mark Whitley | 807f0fd | 2000-08-02 18:30:11 +0000 | [diff] [blame] | 312 | |
Mark Whitley | b696763 | 2001-05-18 23:04:51 +0000 | [diff] [blame] | 313 | /* non-field (char or byte) cutting has some special handling */ |
Eric Andersen | 8876fb2 | 2003-06-20 09:01:58 +0000 | [diff] [blame] | 314 | if (part != OPT_FIELDS_FLGS) { |
Mark Whitley | b696763 | 2001-05-18 23:04:51 +0000 | [diff] [blame] | 315 | if (supress_non_delimited_lines) { |
Manuel Novoa III | cad5364 | 2003-03-19 09:13:01 +0000 | [diff] [blame] | 316 | bb_error_msg_and_die("suppressing non-delimited lines makes sense" |
Mark Whitley | b696763 | 2001-05-18 23:04:51 +0000 | [diff] [blame] | 317 | " only when operating on fields"); |
| 318 | } |
Eric Andersen | 8876fb2 | 2003-06-20 09:01:58 +0000 | [diff] [blame] | 319 | if (delim != '\t') { |
Manuel Novoa III | cad5364 | 2003-03-19 09:13:01 +0000 | [diff] [blame] | 320 | bb_error_msg_and_die("a delimiter may be specified only when operating on fields"); |
Mark Whitley | b696763 | 2001-05-18 23:04:51 +0000 | [diff] [blame] | 321 | } |
Mark Whitley | 807f0fd | 2000-08-02 18:30:11 +0000 | [diff] [blame] | 322 | } |
| 323 | |
| 324 | /* argv[(optind)..(argc-1)] should be names of file to process. If no |
| 325 | * files were specified or '-' was specified, take input from stdin. |
| 326 | * Otherwise, we process all the files specified. */ |
| 327 | if (argv[optind] == NULL || (strcmp(argv[optind], "-") == 0)) { |
| 328 | cut_file(stdin); |
| 329 | } |
| 330 | else { |
| 331 | int i; |
| 332 | FILE *file; |
| 333 | for (i = optind; i < argc; i++) { |
Manuel Novoa III | cad5364 | 2003-03-19 09:13:01 +0000 | [diff] [blame] | 334 | file = bb_wfopen(argv[i], "r"); |
Eric Andersen | 3450636 | 2001-08-02 05:02:46 +0000 | [diff] [blame] | 335 | if(file) { |
Mark Whitley | 807f0fd | 2000-08-02 18:30:11 +0000 | [diff] [blame] | 336 | cut_file(file); |
| 337 | fclose(file); |
| 338 | } |
| 339 | } |
| 340 | } |
| 341 | |
Matt Kraai | 3e856ce | 2000-12-01 02:55:13 +0000 | [diff] [blame] | 342 | return EXIT_SUCCESS; |
Mark Whitley | 807f0fd | 2000-08-02 18:30:11 +0000 | [diff] [blame] | 343 | } |