1/* vi: set sw=4 ts=4: */
2/*
3 * cut.c - minimalist version of cut
4 *
5 * Copyright (C) 1999,2000,2001 by Lineo, inc.
6 * Written by Mark Whitley <markw@codepoet.org>
7 * debloated by Bernhard Fischer
8 *
9 * Licensed under GPLv2 or later, see file LICENSE in this tarball for details.
10 */
11
12#include "libbb.h"
13
14/* This is a NOEXEC applet. Be very careful! */
15
16
17/* option vars */
18static const char optstring[] ALIGN1 = "b:c:f:d:sn";
19#define CUT_OPT_BYTE_FLGS	(1<<0)
20#define CUT_OPT_CHAR_FLGS	(1<<1)
21#define CUT_OPT_FIELDS_FLGS	(1<<2)
22#define CUT_OPT_DELIM_FLGS	(1<<3)
23#define CUT_OPT_SUPPRESS_FLGS (1<<4)
24
25static char delim = '\t';	/* delimiter, default is tab */
26
27struct cut_list {
28	int startpos;
29	int endpos;
30};
31
32enum {
33	BOL = 0,
34	EOL = INT_MAX,
35	NON_RANGE = -1
36};
37
38/* growable array holding a series of lists */
39static struct cut_list *cut_lists;
40static unsigned int nlists;	/* number of elements in above list */
41
42
43static int cmpfunc(const void *a, const void *b)
44{
45	return (((struct cut_list *) a)->startpos -
46			((struct cut_list *) b)->startpos);
47
48}
49
50static void cut_file(FILE * file)
51{
52	char *line = NULL;
53	unsigned int linenum = 0;	/* keep these zero-based to be consistent */
54
55	/* go through every line in the file */
56	while ((line = xmalloc_getline(file)) != NULL) {
57
58		/* set up a list so we can keep track of what's been printed */
59		char * printed = xzalloc(strlen(line) * sizeof(char));
60		char * orig_line = line;
61		unsigned int cl_pos = 0;
62		int spos;
63
64		if (option_mask32 & (CUT_OPT_CHAR_FLGS | CUT_OPT_BYTE_FLGS)) {
65			/* print the chars specified in each cut list */
66			for (; cl_pos < nlists; cl_pos++) {
67				spos = cut_lists[cl_pos].startpos;
68				while (spos < strlen(line)) {
69					if (!printed[spos]) {
70						printed[spos] = 'X';
71						putchar(line[spos]);
72					}
73					spos++;
74					if (spos > cut_lists[cl_pos].endpos
75						|| cut_lists[cl_pos].endpos == NON_RANGE)
76						break;
77				}
78			}
79		} else if (delim == '\n') {	/* cut by lines */
80			spos = cut_lists[cl_pos].startpos;
81
82			/* get out if we have no more lists to process or if the lines
83			 * are lower than what we're interested in */
84			if (linenum < spos || cl_pos >= nlists)
85				goto next_line;
86
87			/* if the line we're looking for is lower than the one we were
88			 * passed, it means we displayed it already, so move on */
89			while (spos < linenum) {
90				spos++;
91				/* go to the next list if we're at the end of this one */
92				if (spos > cut_lists[cl_pos].endpos
93					|| cut_lists[cl_pos].endpos == NON_RANGE) {
94					cl_pos++;
95					/* get out if there's no more lists to process */
96					if (cl_pos >= nlists)
97						goto next_line;
98					spos = cut_lists[cl_pos].startpos;
99					/* get out if the current line is lower than the one
100					 * we just became interested in */
101					if (linenum < spos)
102						goto next_line;
103				}
104			}
105
106			/* If we made it here, it means we've found the line we're
107			 * looking for, so print it */
108			puts(line);
109			goto next_line;
110		} else {		/* cut by fields */
111			int ndelim = -1;	/* zero-based / one-based problem */
112			int nfields_printed = 0;
113			char *field = NULL;
114			const char delimiter[2] = { delim, 0 };
115
116			/* does this line contain any delimiters? */
117			if (strchr(line, delim) == NULL) {
118				if (!(option_mask32 & CUT_OPT_SUPPRESS_FLGS))
119					puts(line);
120				goto next_line;
121			}
122
123			/* process each list on this line, for as long as we've got
124			 * a line to process */
125			for (; cl_pos < nlists && line; cl_pos++) {
126				spos = cut_lists[cl_pos].startpos;
127				do {
128					/* find the field we're looking for */
129					while (line && ndelim < spos) {
130						field = strsep(&line, delimiter);
131						ndelim++;
132					}
133
134					/* we found it, and it hasn't been printed yet */
135					if (field && ndelim == spos && !printed[ndelim]) {
136						/* if this isn't our first time through, we need to
137						 * print the delimiter after the last field that was
138						 * printed */
139						if (nfields_printed > 0)
140							putchar(delim);
141						fputs(field, stdout);
142						printed[ndelim] = 'X';
143						nfields_printed++;	/* shouldn't overflow.. */
144					}
145
146					spos++;
147
148					/* keep going as long as we have a line to work with,
149					 * this is a list, and we're not at the end of that
150					 * list */
151				} while (spos <= cut_lists[cl_pos].endpos && line
152						 && cut_lists[cl_pos].endpos != NON_RANGE);
153			}
154		}
155		/* if we printed anything at all, we need to finish it with a
156		 * newline cuz we were handed a chomped line */
157		putchar('\n');
158 next_line:
159		linenum++;
160		free(printed);
161		free(orig_line);
162	}
163}
164
165static const char _op_on_field[] ALIGN1 = " only when operating on fields";
166
167int cut_main(int argc, char **argv);
168int cut_main(int argc, char **argv)
169{
170	char *sopt, *ltok;
171
172	opt_complementary = "b--bcf:c--bcf:f--bcf";
173	getopt32(argv, optstring, &sopt, &sopt, &sopt, &ltok);
174//	argc -= optind;
175	argv += optind;
176	if (!(option_mask32 & (CUT_OPT_BYTE_FLGS | CUT_OPT_CHAR_FLGS | CUT_OPT_FIELDS_FLGS)))
177		bb_error_msg_and_die("expected a list of bytes, characters, or fields");
178
179	if (option_mask32 & CUT_OPT_DELIM_FLGS) {
180		if (strlen(ltok) > 1) {
181			bb_error_msg_and_die("the delimiter must be a single character");
182		}
183		delim = ltok[0];
184	}
185
186	/*  non-field (char or byte) cutting has some special handling */
187	if (!(option_mask32 & CUT_OPT_FIELDS_FLGS)) {
188		if (option_mask32 & CUT_OPT_SUPPRESS_FLGS) {
189			bb_error_msg_and_die
190				("suppressing non-delimited lines makes sense%s",
191				 _op_on_field);
192		}
193		if (delim != '\t') {
194			bb_error_msg_and_die
195				("a delimiter may be specified%s", _op_on_field);
196		}
197	}
198
199	/*
200	 * parse list and put values into startpos and endpos.
201	 * valid list formats: N, N-, N-M, -M
202	 * more than one list can be separated by commas
203	 */
204	{
205		char *ntok;
206		int s = 0, e = 0;
207
208		/* take apart the lists, one by one (they are separated with commas */
209		while ((ltok = strsep(&sopt, ",")) != NULL) {
210
211			/* it's actually legal to pass an empty list */
212			if (strlen(ltok) == 0)
213				continue;
214
215			/* get the start pos */
216			ntok = strsep(&ltok, "-");
217			if (ntok == NULL) {
218				bb_error_msg
219					("internal error: ntok is null for start pos!?\n");
220			} else if (strlen(ntok) == 0) {
221				s = BOL;
222			} else {
223				s = xatoi_u(ntok);
224				/* account for the fact that arrays are zero based, while
225				 * the user expects the first char on the line to be char #1 */
226				if (s != 0)
227					s--;
228			}
229
230			/* get the end pos */
231			ntok = strsep(&ltok, "-");
232			if (ntok == NULL) {
233				e = NON_RANGE;
234			} else if (strlen(ntok) == 0) {
235				e = EOL;
236			} else {
237				e = xatoi_u(ntok);
238				/* if the user specified and end position of 0, that means "til the
239				 * end of the line */
240				if (e == 0)
241					e = EOL;
242				e--;	/* again, arrays are zero based, lines are 1 based */
243				if (e == s)
244					e = NON_RANGE;
245			}
246
247			/* if there's something left to tokenize, the user passed
248			 * an invalid list */
249			if (ltok)
250				bb_error_msg_and_die("invalid byte or field list");
251
252			/* add the new list */
253			cut_lists = xrealloc(cut_lists, sizeof(struct cut_list) * (++nlists));
254			cut_lists[nlists-1].startpos = s;
255			cut_lists[nlists-1].endpos = e;
256		}
257
258		/* make sure we got some cut positions out of all that */
259		if (nlists == 0)
260			bb_error_msg_and_die("missing list of positions");
261
262		/* now that the lists are parsed, we need to sort them to make life
263		 * easier on us when it comes time to print the chars / fields / lines
264		 */
265		qsort(cut_lists, nlists, sizeof(struct cut_list), cmpfunc);
266	}
267
268	/* argv[0..argc-1] should be names of file to process. If no
269	 * files were specified or '-' was specified, take input from stdin.
270	 * Otherwise, we process all the files specified. */
271	if (argv[0] == NULL || LONE_DASH(argv[0])) {
272		cut_file(stdin);
273	} else {
274		FILE *file;
275
276		do {
277			file = fopen_or_warn(argv[0], "r");
278			if (file) {
279				cut_file(file);
280				fclose(file);
281			}
282		} while (*++argv);
283	}
284	if (ENABLE_FEATURE_CLEAN_UP)
285		free(cut_lists);
286	return EXIT_SUCCESS;
287}
288