cut.c revision 99115
122347Spst/* 222347Spst * Copyright (c) 1989, 1993 329964Sache * The Regents of the University of California. All rights reserved. 429964Sache * 522347Spst * This code is derived from software contributed to Berkeley by 622347Spst * Adam S. Moskowitz of Menlo Consulting and Marciano Pitargue. 722347Spst * 822347Spst * Redistribution and use in source and binary forms, with or without 922347Spst * modification, are permitted provided that the following conditions 1022347Spst * are met: 1122347Spst * 1. Redistributions of source code must retain the above copyright 1222347Spst * notice, this list of conditions and the following disclaimer. 1322347Spst * 2. Redistributions in binary form must reproduce the above copyright 1422347Spst * notice, this list of conditions and the following disclaimer in the 1522347Spst * documentation and/or other materials provided with the distribution. 1622347Spst * 3. All advertising materials mentioning features or use of this software 1729964Sache * must display the following acknowledgement: 1829964Sache * This product includes software developed by the University of 1929964Sache * California, Berkeley and its contributors. 2029964Sache * 4. Neither the name of the University nor the names of its contributors 2129964Sache * may be used to endorse or promote products derived from this software 2229964Sache * without specific prior written permission. 2329964Sache * 2429964Sache * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND 2522347Spst * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 2622347Spst * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 2722347Spst * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE 2822347Spst * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 2922347Spst * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 3022347Spst * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 3122347Spst * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 3222347Spst * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 3322347Spst * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 3422347Spst * SUCH DAMAGE. 3522347Spst */ 3622347Spst 3722347Spst#ifndef lint 3822347Spststatic const char copyright[] = 3922347Spst"@(#) Copyright (c) 1989, 1993\n\ 4022347Spst The Regents of the University of California. All rights reserved.\n"; 4122347Spststatic const char sccsid[] = "@(#)cut.c 8.3 (Berkeley) 5/4/95"; 4222347Spst#endif /* not lint */ 4322347Spst#include <sys/cdefs.h> 4422347Spst__FBSDID("$FreeBSD: head/usr.bin/cut/cut.c 99115 2002-06-30 05:34:21Z obrien $"); 4529964Sache 4629964Sache#include <ctype.h> 4729964Sache#include <err.h> 4829964Sache#include <limits.h> 4929964Sache#include <locale.h> 5029964Sache#include <stdio.h> 5129964Sache#include <stdlib.h> 5229964Sache#include <string.h> 5329964Sache#include <unistd.h> 5429964Sache 5529964Sacheint bflag; 5629964Sacheint cflag; 5729964Sachechar dchar; 5829964Sacheint dflag; 5929964Sacheint fflag; 6029964Sacheint nflag; 6129964Sacheint sflag; 6229964Sache 6329964Sachevoid b_n_cut(FILE *, const char *); 6429964Sachevoid c_cut(FILE *, const char *); 6529964Sachevoid f_cut(FILE *, const char *); 6629964Sachevoid get_list(char *); 6729964Sacheint main(int, char **); 6829964Sachevoid needpos(size_t); 6929964Sachestatic void usage(void); 7029964Sache 7129964Sacheint 7229964Sachemain(argc, argv) 7322347Spst int argc; 7429964Sache char *argv[]; 7529964Sache{ 7622347Spst FILE *fp; 7722347Spst void (*fcn)(FILE *, const char *); 7822347Spst int ch, rval; 7922347Spst 8022347Spst setlocale(LC_ALL, ""); 8122347Spst 8222347Spst fcn = NULL; 8322347Spst dchar = '\t'; /* default delimiter is \t */ 8422347Spst 8522347Spst /* 8622347Spst * Since we don't support multi-byte characters, the -c and -b 8722347Spst * options are equivalent. 8822347Spst */ 8922347Spst while ((ch = getopt(argc, argv, "b:c:d:f:sn")) != -1) 9022347Spst switch(ch) { 9122347Spst case 'b': 9222347Spst fcn = c_cut; 9322347Spst get_list(optarg); 9422347Spst bflag = 1; 9522347Spst break; 9622347Spst case 'c': 9722347Spst fcn = c_cut; 9822347Spst get_list(optarg); 9922347Spst cflag = 1; 10022347Spst break; 10122347Spst case 'd': 10222347Spst dchar = *optarg; 10322347Spst dflag = 1; 10422347Spst break; 10522347Spst case 'f': 10622347Spst get_list(optarg); 10722347Spst fcn = f_cut; 10822347Spst fflag = 1; 10922347Spst break; 11022347Spst case 's': 11122347Spst sflag = 1; 11222347Spst break; 11322347Spst case 'n': 11422347Spst nflag = 1; 11522347Spst break; 11622347Spst case '?': 11722347Spst default: 11822347Spst usage(); 11922347Spst } 12022347Spst argc -= optind; 12122347Spst argv += optind; 12222347Spst 12322347Spst if (fflag) { 12422347Spst if (bflag || cflag || nflag) 12522347Spst usage(); 12622347Spst } else if (!(bflag || cflag) || dflag || sflag) 12722347Spst usage(); 12822347Spst else if (!bflag && nflag) 12922347Spst usage(); 13022347Spst 13122347Spst if (nflag) 13222347Spst fcn = b_n_cut; 13322347Spst 13422347Spst rval = 0; 13522347Spst if (*argv) 13622347Spst for (; *argv; ++argv) { 13722347Spst if (strcmp(*argv, "-") == 0) 13822347Spst fcn(stdin, "stdin"); 13922347Spst else { 14022347Spst if (!(fp = fopen(*argv, "r"))) { 14122347Spst warn("%s", *argv); 14222347Spst rval = 1; 14322347Spst continue; 14422347Spst } 14522347Spst fcn(fp, *argv); 14622347Spst (void)fclose(fp); 14722347Spst } 14822347Spst } 14922347Spst else 15022347Spst fcn(stdin, "stdin"); 15122347Spst exit(rval); 15222347Spst} 15322347Spst 15422347Spstsize_t autostart, autostop, maxval; 15522347Spst 15622347Spstchar *positions; 15722347Spst 15822347Spstvoid 15922347Spstget_list(list) 16022347Spst char *list; 16122347Spst{ 16222347Spst size_t setautostart, start, stop; 16322347Spst char *pos; 16422347Spst char *p; 16522347Spst 16622347Spst /* 16722347Spst * set a byte in the positions array to indicate if a field or 16822347Spst * column is to be selected; use +1, it's 1-based, not 0-based. 16922347Spst * This parser is less restrictive than the Draft 9 POSIX spec. 17022347Spst * POSIX doesn't allow lists that aren't in increasing order or 17122347Spst * overlapping lists. We also handle "-3-5" although there's no 17222347Spst * real reason too. 17322347Spst */ 17422347Spst for (; (p = strsep(&list, ", \t")) != NULL;) { 17522347Spst setautostart = start = stop = 0; 17622347Spst if (*p == '-') { 17722347Spst ++p; 17822347Spst setautostart = 1; 17922347Spst } 18022347Spst if (isdigit((unsigned char)*p)) { 18122347Spst start = stop = strtol(p, &p, 10); 18222347Spst if (setautostart && start > autostart) 18322347Spst autostart = start; 18422347Spst } 18522347Spst if (*p == '-') { 18622347Spst if (isdigit((unsigned char)p[1])) 18722347Spst stop = strtol(p + 1, &p, 10); 18822347Spst if (*p == '-') { 18922347Spst ++p; 19022347Spst if (!autostop || autostop > stop) 19122347Spst autostop = stop; 19222347Spst } 19322347Spst } 19422347Spst if (*p) 19522347Spst errx(1, "[-cf] list: illegal list value"); 19622347Spst if (!stop || !start) 19722347Spst errx(1, "[-cf] list: values may not include zero"); 19822347Spst if (maxval < stop) { 19922347Spst maxval = stop; 20022347Spst needpos(maxval + 1); 20122347Spst } 20222347Spst for (pos = positions + start; start++ <= stop; *pos++ = 1); 20322347Spst } 20422347Spst 20522347Spst /* overlapping ranges */ 20622347Spst if (autostop && maxval > autostop) { 20722347Spst maxval = autostop; 20822347Spst needpos(maxval + 1); 20922347Spst } 21022347Spst 21122347Spst /* set autostart */ 21222347Spst if (autostart) 21322347Spst memset(positions + 1, '1', autostart); 21422347Spst} 21522347Spst 21622347Spstvoid 21722347Spstneedpos(size_t n) 21822347Spst{ 21922347Spst static size_t npos; 22022347Spst size_t oldnpos; 22122347Spst 22222347Spst /* Grow the positions array to at least the specified size. */ 22322347Spst if (n > npos) { 22422347Spst oldnpos = npos; 22522347Spst if (npos == 0) 22622347Spst npos = n; 22722347Spst while (n > npos) 22822347Spst npos *= 2; 22922347Spst if ((positions = realloc(positions, npos)) == NULL) 23022347Spst err(1, "realloc"); 23122347Spst memset((char *)positions + oldnpos, 0, npos - oldnpos); 23222347Spst } 23322347Spst} 23422347Spst 23522347Spst/* 23622347Spst * Cut based on byte positions, taking care not to split multibyte characters. 23722347Spst * Although this function also handles the case where -n is not specified, 23822347Spst * c_cut() ought to be much faster. 23922347Spst */ 24022347Spstvoid 24122347Spstb_n_cut(fp, fname) 24222347Spst FILE *fp; 24322347Spst const char *fname; 24422347Spst{ 24522347Spst size_t col, i, lbuflen; 24622347Spst char *lbuf; 24722347Spst int canwrite, clen, warned; 24822347Spst 24922347Spst warned = 0; 25022347Spst while ((lbuf = fgetln(fp, &lbuflen)) != NULL) { 25122347Spst for (col = 0; lbuflen > 0; col += clen) { 25222347Spst if ((clen = mblen(lbuf, lbuflen)) < 0) { 25322347Spst if (!warned) { 25422347Spst warn("%s", fname); 25522347Spst warned = 1; 25622347Spst } 25722347Spst clen = 1; 25822347Spst } 25922347Spst if (clen == 0 || *lbuf == '\n') 26022347Spst break; 26122347Spst if (col < maxval && !positions[1 + col]) { 26222347Spst /* 26322347Spst * Print the character if (1) after an initial 26422347Spst * segment of un-selected bytes, the rest of 26522347Spst * it is selected, and (2) the last byte is 26622347Spst * selected. 26722347Spst */ 26822347Spst i = col; 26922347Spst while (i < col + clen && i < maxval && 27022347Spst !positions[1 + i]) 27129964Sache i++; 27229964Sache canwrite = i < col + clen; 27329964Sache for (; i < col + clen && i < maxval; i++) 27429964Sache canwrite &= positions[1 + i]; 27529964Sache if (canwrite) 27629964Sache fwrite(lbuf, 1, clen, stdout); 27729964Sache } else { 27829964Sache /* 27929964Sache * Print the character if all of it has 28029964Sache * been selected. 28129964Sache */ 28222347Spst canwrite = 1; 28322347Spst for (i = col; i < col + clen; i++) 28422347Spst if ((i >= maxval && !autostop) || 28522347Spst (i < maxval && !positions[1 + i])) { 28622347Spst canwrite = 0; 28722347Spst break; 28822347Spst } 28922347Spst if (canwrite) 29022347Spst fwrite(lbuf, 1, clen, stdout); 29122347Spst } 29222347Spst lbuf += clen; 29322347Spst lbuflen -= clen; 29422347Spst } 29522347Spst if (lbuflen > 0) 29622347Spst putchar('\n'); 29722347Spst } 29822347Spst} 29922347Spst 30022347Spstvoid 30122347Spstc_cut(fp, fname) 30222347Spst FILE *fp; 30322347Spst const char *fname __unused; 30422347Spst{ 30522347Spst int ch, col; 30622347Spst char *pos; 30722347Spst 30822347Spst ch = 0; 30922347Spst for (;;) { 31022347Spst pos = positions + 1; 31122347Spst for (col = maxval; col; --col) { 31222347Spst if ((ch = getc(fp)) == EOF) 31329964Sache return; 31429964Sache if (ch == '\n') 31529964Sache break; 31629964Sache if (*pos++) 31729964Sache (void)putchar(ch); 31829964Sache } 31929964Sache if (ch != '\n') { 32029964Sache if (autostop) 32129964Sache while ((ch = getc(fp)) != EOF && ch != '\n') 32229964Sache (void)putchar(ch); 32329964Sache else 32429964Sache while ((ch = getc(fp)) != EOF && ch != '\n'); 32529964Sache } 32629964Sache (void)putchar('\n'); 32729964Sache } 32829964Sache} 32929964Sache 33022347Spstvoid 33129964Sachef_cut(fp, fname) 33229964Sache FILE *fp; 33329964Sache const char *fname __unused; 33429964Sache{ 33529964Sache int ch, field, isdelim; 33629964Sache char *pos, *p, sep; 33729964Sache int output; 33829964Sache char *lbuf, *mlbuf; 33929964Sache size_t lbuflen; 34029964Sache 34129964Sache mlbuf = NULL; 34229964Sache for (sep = dchar; (lbuf = fgetln(fp, &lbuflen)) != NULL;) { 34329964Sache /* Assert EOL has a newline. */ 34429964Sache if (*(lbuf + lbuflen - 1) != '\n') { 34529964Sache /* Can't have > 1 line with no trailing newline. */ 34622347Spst mlbuf = malloc(lbuflen + 1); 34729964Sache if (mlbuf == NULL) 34829964Sache err(1, "malloc"); 34929964Sache memcpy(mlbuf, lbuf, lbuflen); 35029964Sache *(mlbuf + lbuflen) = '\n'; 35129964Sache lbuf = mlbuf; 35229964Sache } 35329964Sache output = 0; 35429964Sache for (isdelim = 0, p = lbuf;; ++p) { 35529964Sache ch = *p; 35629964Sache /* this should work if newline is delimiter */ 35729964Sache if (ch == sep) 35829964Sache isdelim = 1; 35929964Sache if (ch == '\n') { 36029964Sache if (!isdelim && !sflag) 36129964Sache (void)fwrite(lbuf, lbuflen, 1, stdout); 36229964Sache break; 36329964Sache } 36429964Sache } 36529964Sache if (!isdelim) 36629964Sache continue; 36729964Sache 36829964Sache pos = positions + 1; 36929964Sache for (field = maxval, p = lbuf; field; --field, ++pos) { 37029964Sache if (*pos) { 37129964Sache if (output++) 37229964Sache (void)putchar(sep); 37329964Sache while ((ch = *p++) != '\n' && ch != sep) 37429964Sache (void)putchar(ch); 37529964Sache } else { 37629964Sache while ((ch = *p++) != '\n' && ch != sep) 37729964Sache continue; 37829964Sache } 37929964Sache if (ch == '\n') 38029964Sache break; 38129964Sache } 38229964Sache if (ch != '\n') { 38322347Spst if (autostop) { 38422347Spst if (output) 38522347Spst (void)putchar(sep); 38622347Spst for (; (ch = *p) != '\n'; ++p) 38722347Spst (void)putchar(ch); 38822347Spst } else 38922347Spst for (; (ch = *p) != '\n'; ++p); 39022347Spst } 39122347Spst (void)putchar('\n'); 39222347Spst } 39322347Spst if (mlbuf != NULL) 39422347Spst free(mlbuf); 39522347Spst} 39622347Spst 39722347Spststatic void 39822347Spstusage() 39922347Spst{ 40022347Spst (void)fprintf(stderr, "%s\n%s\n%s\n", 40122347Spst "usage: cut -b list [-n] [file ...]", 40222347Spst " cut -c list [file ...]", 40322347Spst " cut -f list [-s] [-d delim] [file ...]"); 40422347Spst exit(1); 40522347Spst} 40622347Spst