docs/cut

annotate code/cut.c__netbsd.2014-02-03 @ 14:21ad1c1548c4

find changesets by author, revision, files, or words in the commit message
Code ausgewaehlter Implementierungen eingefuegt Das Datum entspricht dem Dateiaenderungsdatum.
author markus schnalke <meillo@marmaro.de>
date Tue, 12 May 2015 06:46:59 +0200
parents
children
rev   line source
meillo@14 1 /* $NetBSD: cut.c,v 1.29 2014/02/03 20:22:19 wiz Exp $ */
meillo@14 2
meillo@14 3 /*
meillo@14 4 * Copyright (c) 1989, 1993
meillo@14 5 * The Regents of the University of California. All rights reserved.
meillo@14 6 *
meillo@14 7 * This code is derived from software contributed to Berkeley by
meillo@14 8 * Adam S. Moskowitz of Menlo Consulting and Marciano Pitargue.
meillo@14 9 *
meillo@14 10 * Redistribution and use in source and binary forms, with or without
meillo@14 11 * modification, are permitted provided that the following conditions
meillo@14 12 * are met:
meillo@14 13 * 1. Redistributions of source code must retain the above copyright
meillo@14 14 * notice, this list of conditions and the following disclaimer.
meillo@14 15 * 2. Redistributions in binary form must reproduce the above copyright
meillo@14 16 * notice, this list of conditions and the following disclaimer in the
meillo@14 17 * documentation and/or other materials provided with the distribution.
meillo@14 18 * 3. Neither the name of the University nor the names of its contributors
meillo@14 19 * may be used to endorse or promote products derived from this software
meillo@14 20 * without specific prior written permission.
meillo@14 21 *
meillo@14 22 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
meillo@14 23 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
meillo@14 24 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
meillo@14 25 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
meillo@14 26 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
meillo@14 27 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
meillo@14 28 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
meillo@14 29 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
meillo@14 30 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
meillo@14 31 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
meillo@14 32 * SUCH DAMAGE.
meillo@14 33 */
meillo@14 34
meillo@14 35 #include <sys/cdefs.h>
meillo@14 36 #ifndef lint
meillo@14 37 __COPYRIGHT("@(#) Copyright (c) 1989, 1993\
meillo@14 38 The Regents of the University of California. All rights reserved.");
meillo@14 39 #endif /* not lint */
meillo@14 40
meillo@14 41 #ifndef lint
meillo@14 42 #if 0
meillo@14 43 static char sccsid[] = "@(#)cut.c 8.3 (Berkeley) 5/4/95";
meillo@14 44 #endif
meillo@14 45 __RCSID("$NetBSD: cut.c,v 1.29 2014/02/03 20:22:19 wiz Exp $");
meillo@14 46 #endif /* not lint */
meillo@14 47
meillo@14 48 #include <ctype.h>
meillo@14 49 #include <err.h>
meillo@14 50 #include <errno.h>
meillo@14 51 #include <limits.h>
meillo@14 52 #include <locale.h>
meillo@14 53 #include <stdio.h>
meillo@14 54 #include <stdlib.h>
meillo@14 55 #include <string.h>
meillo@14 56 #include <unistd.h>
meillo@14 57 #include <util.h>
meillo@14 58 #include <wchar.h>
meillo@14 59 #include <sys/param.h>
meillo@14 60
meillo@14 61 static int bflag;
meillo@14 62 static int cflag;
meillo@14 63 static char dchar;
meillo@14 64 static int dflag;
meillo@14 65 static int fflag;
meillo@14 66 static int sflag;
meillo@14 67
meillo@14 68 static void b_cut(FILE *, const char *);
meillo@14 69 static void c_cut(FILE *, const char *);
meillo@14 70 static void f_cut(FILE *, const char *);
meillo@14 71 static void get_list(char *);
meillo@14 72 static void usage(void) __dead;
meillo@14 73
meillo@14 74 int
meillo@14 75 main(int argc, char *argv[])
meillo@14 76 {
meillo@14 77 FILE *fp;
meillo@14 78 void (*fcn)(FILE *, const char *);
meillo@14 79 int ch, rval;
meillo@14 80
meillo@14 81 fcn = NULL;
meillo@14 82 (void)setlocale(LC_ALL, "");
meillo@14 83
meillo@14 84 dchar = '\t'; /* default delimiter is \t */
meillo@14 85
meillo@14 86 /* Since we don't support multi-byte characters, the -c and -b
meillo@14 87 options are equivalent, and the -n option is meaningless. */
meillo@14 88 while ((ch = getopt(argc, argv, "b:c:d:f:sn")) != -1)
meillo@14 89 switch(ch) {
meillo@14 90 case 'b':
meillo@14 91 fcn = b_cut;
meillo@14 92 get_list(optarg);
meillo@14 93 bflag = 1;
meillo@14 94 break;
meillo@14 95 case 'c':
meillo@14 96 fcn = c_cut;
meillo@14 97 get_list(optarg);
meillo@14 98 cflag = 1;
meillo@14 99 break;
meillo@14 100 case 'd':
meillo@14 101 dchar = *optarg;
meillo@14 102 dflag = 1;
meillo@14 103 break;
meillo@14 104 case 'f':
meillo@14 105 get_list(optarg);
meillo@14 106 fcn = f_cut;
meillo@14 107 fflag = 1;
meillo@14 108 break;
meillo@14 109 case 's':
meillo@14 110 sflag = 1;
meillo@14 111 break;
meillo@14 112 case 'n':
meillo@14 113 break;
meillo@14 114 case '?':
meillo@14 115 default:
meillo@14 116 usage();
meillo@14 117 }
meillo@14 118 argc -= optind;
meillo@14 119 argv += optind;
meillo@14 120
meillo@14 121 if (fflag) {
meillo@14 122 if (cflag || bflag)
meillo@14 123 usage();
meillo@14 124 } else if ((!cflag && !bflag) || dflag || sflag)
meillo@14 125 usage();
meillo@14 126 else if (bflag && cflag)
meillo@14 127 usage();
meillo@14 128
meillo@14 129 rval = 0;
meillo@14 130 if (*argv)
meillo@14 131 for (; *argv; ++argv) {
meillo@14 132 if (strcmp(*argv, "-") == 0)
meillo@14 133 fcn(stdin, "stdin");
meillo@14 134 else {
meillo@14 135 if ((fp = fopen(*argv, "r"))) {
meillo@14 136 fcn(fp, *argv);
meillo@14 137 (void)fclose(fp);
meillo@14 138 } else {
meillo@14 139 rval = 1;
meillo@14 140 warn("%s", *argv);
meillo@14 141 }
meillo@14 142 }
meillo@14 143 }
meillo@14 144 else
meillo@14 145 fcn(stdin, "stdin");
meillo@14 146 return(rval);
meillo@14 147 }
meillo@14 148
meillo@14 149 static size_t autostart, autostop, maxval;
meillo@14 150
meillo@14 151 static char *positions = NULL;
meillo@14 152 static size_t numpositions = 0;
meillo@14 153 #define ALLOC_CHUNK _POSIX2_LINE_MAX /* malloc granularity */
meillo@14 154
meillo@14 155 static void
meillo@14 156 get_list(char *list)
meillo@14 157 {
meillo@14 158 size_t setautostart, start, stop;
meillo@14 159 char *pos;
meillo@14 160 char *p;
meillo@14 161
meillo@14 162 if (positions == NULL) {
meillo@14 163 numpositions = ALLOC_CHUNK;
meillo@14 164 positions = ecalloc(numpositions, sizeof(*positions));
meillo@14 165 }
meillo@14 166
meillo@14 167 /*
meillo@14 168 * set a byte in the positions array to indicate if a field or
meillo@14 169 * column is to be selected; use +1, it's 1-based, not 0-based.
meillo@14 170 * This parser is less restrictive than the Draft 9 POSIX spec.
meillo@14 171 * POSIX doesn't allow lists that aren't in increasing order or
meillo@14 172 * overlapping lists. We also handle "-3-5" although there's no
meillo@14 173 * real reason to.
meillo@14 174 */
meillo@14 175 for (; (p = strtok(list, ", \t")) != NULL; list = NULL) {
meillo@14 176 setautostart = start = stop = 0;
meillo@14 177 if (*p == '-') {
meillo@14 178 ++p;
meillo@14 179 setautostart = 1;
meillo@14 180 }
meillo@14 181 if (isdigit((unsigned char)*p)) {
meillo@14 182 start = stop = strtol(p, &p, 10);
meillo@14 183 if (setautostart && start > autostart)
meillo@14 184 autostart = start;
meillo@14 185 }
meillo@14 186 if (*p == '-') {
meillo@14 187 if (isdigit((unsigned char)p[1]))
meillo@14 188 stop = strtol(p + 1, &p, 10);
meillo@14 189 if (*p == '-') {
meillo@14 190 ++p;
meillo@14 191 if (!autostop || autostop > stop)
meillo@14 192 autostop = stop;
meillo@14 193 }
meillo@14 194 }
meillo@14 195 if (*p)
meillo@14 196 errx(1, "[-bcf] list: illegal list value");
meillo@14 197 if (!stop || !start)
meillo@14 198 errx(1, "[-bcf] list: values may not include zero");
meillo@14 199 if (stop + 1 > numpositions) {
meillo@14 200 size_t newsize;
meillo@14 201 newsize = roundup(stop + 1, ALLOC_CHUNK);
meillo@14 202 positions = erealloc(positions, newsize);
meillo@14 203 (void)memset(positions + numpositions, 0,
meillo@14 204 newsize - numpositions);
meillo@14 205 numpositions = newsize;
meillo@14 206 }
meillo@14 207 if (maxval < stop)
meillo@14 208 maxval = stop;
meillo@14 209 for (pos = positions + start; start++ <= stop; pos++)
meillo@14 210 *pos = 1;
meillo@14 211 }
meillo@14 212
meillo@14 213 /* overlapping ranges */
meillo@14 214 if (autostop && maxval > autostop)
meillo@14 215 maxval = autostop;
meillo@14 216
meillo@14 217 /* set autostart */
meillo@14 218 if (autostart)
meillo@14 219 (void)memset(positions + 1, '1', autostart);
meillo@14 220 }
meillo@14 221
meillo@14 222 static void
meillo@14 223 /*ARGSUSED*/
meillo@14 224 f_cut(FILE *fp, const char *fname __unused)
meillo@14 225 {
meillo@14 226 int ch, field, isdelim;
meillo@14 227 char *pos, *p, sep;
meillo@14 228 int output;
meillo@14 229 size_t len;
meillo@14 230 char *lbuf, *tbuf;
meillo@14 231
meillo@14 232 for (sep = dchar, tbuf = NULL; (lbuf = fgetln(fp, &len)) != NULL;) {
meillo@14 233 output = 0;
meillo@14 234 if (lbuf[len - 1] != '\n') {
meillo@14 235 /* no newline at the end of the last line so add one */
meillo@14 236 if ((tbuf = (char *)malloc(len + 1)) == NULL)
meillo@14 237 err(1, NULL);
meillo@14 238 (void)memcpy(tbuf, lbuf, len);
meillo@14 239 tbuf[len++] = '\n';
meillo@14 240 lbuf = tbuf;
meillo@14 241 }
meillo@14 242 for (isdelim = 0, p = lbuf;; ++p) {
meillo@14 243 ch = *p;
meillo@14 244 /* this should work if newline is delimiter */
meillo@14 245 if (ch == sep)
meillo@14 246 isdelim = 1;
meillo@14 247 if (ch == '\n') {
meillo@14 248 if (!isdelim && !sflag)
meillo@14 249 (void)fwrite(lbuf, len, 1, stdout);
meillo@14 250 break;
meillo@14 251 }
meillo@14 252 }
meillo@14 253 if (!isdelim)
meillo@14 254 continue;
meillo@14 255
meillo@14 256 pos = positions + 1;
meillo@14 257 for (field = maxval, p = lbuf; field; --field, ++pos) {
meillo@14 258 if (*pos) {
meillo@14 259 if (output++)
meillo@14 260 (void)putchar(sep);
meillo@14 261 while ((ch = *p++) != '\n' && ch != sep)
meillo@14 262 (void)putchar(ch);
meillo@14 263 } else {
meillo@14 264 while ((ch = *p++) != '\n' && ch != sep)
meillo@14 265 continue;
meillo@14 266 }
meillo@14 267 if (ch == '\n')
meillo@14 268 break;
meillo@14 269 }
meillo@14 270 if (ch != '\n') {
meillo@14 271 if (autostop) {
meillo@14 272 if (output)
meillo@14 273 (void)putchar(sep);
meillo@14 274 for (; (ch = *p) != '\n'; ++p)
meillo@14 275 (void)putchar(ch);
meillo@14 276 } else
meillo@14 277 for (; (ch = *p) != '\n'; ++p);
meillo@14 278 }
meillo@14 279 (void)putchar('\n');
meillo@14 280 if (tbuf) {
meillo@14 281 free(tbuf);
meillo@14 282 tbuf = NULL;
meillo@14 283 }
meillo@14 284 }
meillo@14 285 if (tbuf)
meillo@14 286 free(tbuf);
meillo@14 287 }
meillo@14 288
meillo@14 289 static void
meillo@14 290 usage(void)
meillo@14 291 {
meillo@14 292 (void)fprintf(stderr, "usage:\tcut -b list [-n] [file ...]\n"
meillo@14 293 "\tcut -c list [file ...]\n"
meillo@14 294 "\tcut -f list [-d string] [-s] [file ...]\n");
meillo@14 295 exit(1);
meillo@14 296 }
meillo@14 297
meillo@14 298 /* make b_put(): */
meillo@14 299 #define CUT_BYTE 1
meillo@14 300 #include "x_cut.c"
meillo@14 301 #undef CUT_BYTE
meillo@14 302
meillo@14 303 /* make c_put(): */
meillo@14 304 #define CUT_BYTE 0
meillo@14 305 #include "x_cut.c"
meillo@14 306 #undef CUT_BYTE