busybox/coreutils/cut.c

/* vi: set sw=4 ts=4: */
/*
 * cut.c - minimalist version of cut
 *
 * Copyright (C) 1999,2000,2001 by Lineo, inc.
 * Written by Mark Whitley <markw@codepoet.org>
 * debloated by Bernhard Reutner-Fischer
 *
 * Licensed under GPLv2 or later, see file LICENSE in this tarball for details.
 */

#include "libbb.h"

/* This is a NOEXEC applet. Be very careful! */


/* option vars */
static const char optstring[] ALIGN1 = "b:c:f:d:sn";
#define CUT_OPT_BYTE_FLGS     (1 << 0)
#define CUT_OPT_CHAR_FLGS     (1 << 1)
#define CUT_OPT_FIELDS_FLGS   (1 << 2)
#define CUT_OPT_DELIM_FLGS    (1 << 3)
#define CUT_OPT_SUPPRESS_FLGS (1 << 4)

struct cut_list {
	int startpos;
	int endpos;
};

enum {
	BOL = 0,
	EOL = INT_MAX,
	NON_RANGE = -1
};

static int cmpfunc(const void *a, const void *b)
{
	return (((struct cut_list *) a)->startpos -
			((struct cut_list *) b)->startpos);

}

static void cut_file(FILE *file, char delim, const struct cut_list *cut_lists, unsigned nlists)
{
	char *line;
	unsigned linenum = 0;	/* keep these zero-based to be consistent */

	/* go through every line in the file */
	while ((line = xmalloc_fgetline(file)) != NULL) {

		/* set up a list so we can keep track of what's been printed */
		int linelen = strlen(line);
		char *printed = xzalloc(linelen + 1);
		char *orig_line = line;
		unsigned cl_pos = 0;
		int spos;

		/* cut based on chars/bytes XXX: only works when sizeof(char) == byte */
		if (option_mask32 & (CUT_OPT_CHAR_FLGS | CUT_OPT_BYTE_FLGS)) {
			/* print the chars specified in each cut list */
			for (; cl_pos < nlists; cl_pos++) {
				spos = cut_lists[cl_pos].startpos;
				while (spos < linelen) {
					if (!printed[spos]) {
						printed[spos] = 'X';
						putchar(line[spos]);
					}
					spos++;
					if (spos > cut_lists[cl_pos].endpos
					/* NON_RANGE is -1, so if below is true,
					 * the above was true too (spos is >= 0) */
					/* || cut_lists[cl_pos].endpos == NON_RANGE */
					) {
						break;
					}
				}
			}
		} else if (delim == '\n') {	/* cut by lines */
			spos = cut_lists[cl_pos].startpos;

			/* get out if we have no more lists to process or if the lines
			 * are lower than what we're interested in */
			if (((int)linenum < spos) || (cl_pos >= nlists))
				goto next_line;

			/* if the line we're looking for is lower than the one we were
			 * passed, it means we displayed it already, so move on */
			while (spos < (int)linenum) {
				spos++;
				/* go to the next list if we're at the end of this one */
				if (spos > cut_lists[cl_pos].endpos
				 || cut_lists[cl_pos].endpos == NON_RANGE
				) {
					cl_pos++;
					/* get out if there's no more lists to process */
					if (cl_pos >= nlists)
						goto next_line;
					spos = cut_lists[cl_pos].startpos;
					/* get out if the current line is lower than the one
					 * we just became interested in */
					if ((int)linenum < spos)
						goto next_line;
				}
			}

			/* If we made it here, it means we've found the line we're
			 * looking for, so print it */
			puts(line);
			goto next_line;
		} else {		/* cut by fields */
			int ndelim = -1;	/* zero-based / one-based problem */
			int nfields_printed = 0;
			char *field = NULL;
			const char delimiter[2] = { delim, 0 };

			/* does this line contain any delimiters? */
			if (strchr(line, delim) == NULL) {
				if (!(option_mask32 & CUT_OPT_SUPPRESS_FLGS))
					puts(line);
				goto next_line;
			}

			/* process each list on this line, for as long as we've got
			 * a line to process */
			for (; cl_pos < nlists && line; cl_pos++) {
				spos = cut_lists[cl_pos].startpos;
				do {
					/* find the field we're looking for */
					while (line && ndelim < spos) {
						field = strsep(&line, delimiter);
						ndelim++;
					}

					/* we found it, and it hasn't been printed yet */
					if (field && ndelim == spos && !printed[ndelim]) {
						/* if this isn't our first time through, we need to
						 * print the delimiter after the last field that was
						 * printed */
						if (nfields_printed > 0)
							putchar(delim);
						fputs(field, stdout);
						printed[ndelim] = 'X';
						nfields_printed++;	/* shouldn't overflow.. */
					}

					spos++;

					/* keep going as long as we have a line to work with,
					 * this is a list, and we're not at the end of that
					 * list */
				} while (spos <= cut_lists[cl_pos].endpos && line
						&& cut_lists[cl_pos].endpos != NON_RANGE);
			}
		}
		/* if we printed anything at all, we need to finish it with a
		 * newline cuz we were handed a chomped line */
		putchar('\n');
 next_line:
		linenum++;
		free(printed);
		free(orig_line);
	}
}

int cut_main(int argc, char **argv) MAIN_EXTERNALLY_VISIBLE;
int cut_main(int argc UNUSED_PARAM, char **argv)
{
	/* growable array holding a series of lists */
	struct cut_list *cut_lists = NULL;
	unsigned nlists = 0;	/* number of elements in above list */
	char delim = '\t';	/* delimiter, default is tab */
	char *sopt, *ltok;
	unsigned opt;

	opt_complementary = "b--bcf:c--bcf:f--bcf";
	opt = getopt32(argv, optstring, &sopt, &sopt, &sopt, &ltok);
//	argc -= optind;
	argv += optind;
	if (!(opt & (CUT_OPT_BYTE_FLGS | CUT_OPT_CHAR_FLGS | CUT_OPT_FIELDS_FLGS)))
		bb_error_msg_and_die("expected a list of bytes, characters, or fields");

	if (opt & CUT_OPT_DELIM_FLGS) {
		if (ltok[0] && ltok[1]) { /* more than 1 char? */
			bb_error_msg_and_die("the delimiter must be a single character");
		}
		delim = ltok[0];
	}

	/*  non-field (char or byte) cutting has some special handling */
	if (!(opt & CUT_OPT_FIELDS_FLGS)) {
		static const char _op_on_field[] ALIGN1 = " only when operating on fields";

		if (opt & CUT_OPT_SUPPRESS_FLGS) {
			bb_error_msg_and_die
				("suppressing non-delimited lines makes sense%s",
				 _op_on_field);
		}
		if (delim != '\t') {
			bb_error_msg_and_die
				("a delimiter may be specified%s", _op_on_field);
		}
	}

	/*
	 * parse list and put values into startpos and endpos.
	 * valid list formats: N, N-, N-M, -M
	 * more than one list can be separated by commas
	 */
	{
		char *ntok;
		int s = 0, e = 0;

		/* take apart the lists, one by one (they are separated with commas) */
		while ((ltok = strsep(&sopt, ",")) != NULL) {

			/* it's actually legal to pass an empty list */
			if (!ltok[0])
				continue;

			/* get the start pos */
			ntok = strsep(&ltok, "-");
			if (!ntok[0]) {
				s = BOL;
			} else {
				s = xatoi_u(ntok);
				/* account for the fact that arrays are zero based, while
				 * the user expects the first char on the line to be char #1 */
				if (s != 0)
					s--;
			}

			/* get the end pos */
			if (ltok == NULL) {
				e = NON_RANGE;
			} else if (!ltok[0]) {
				e = EOL;
			} else {
				e = xatoi_u(ltok);
				/* if the user specified and end position of 0,
				 * that means "til the end of the line" */
				if (e == 0)
					e = EOL;
				e--;	/* again, arrays are zero based, lines are 1 based */
				if (e == s)
					e = NON_RANGE;
			}

			/* add the new list */
			cut_lists = xrealloc_vector(cut_lists, 4, nlists);
			/* NB: startpos is always >= 0,
			 * while endpos may be = NON_RANGE (-1) */
			cut_lists[nlists].startpos = s;
			cut_lists[nlists].endpos = e;
			nlists++;
		}

		/* make sure we got some cut positions out of all that */
		if (nlists == 0)
			bb_error_msg_and_die("missing list of positions");

		/* now that the lists are parsed, we need to sort them to make life
		 * easier on us when it comes time to print the chars / fields / lines
		 */
		qsort(cut_lists, nlists, sizeof(struct cut_list), cmpfunc);
	}

	{
		int retval = EXIT_SUCCESS;

		if (!*argv)
			*--argv = (char *)"-";

		do {
			FILE *file = fopen_or_warn_stdin(*argv);
			if (!file) {
				retval = EXIT_FAILURE;
				continue;
			}
			cut_file(file, delim, cut_lists, nlists);
			fclose_if_not_stdin(file);
		} while (*++argv);

		if (ENABLE_FEATURE_CLEAN_UP)
			free(cut_lists);
		fflush_stdout_and_exit(retval);
	}
}
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 Reutner-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 */
18	static 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
25	struct cut_list {
26	int startpos;
27	int endpos;
28	};
29
30	enum {
31	BOL = 0,
32	EOL = INT_MAX,
33	NON_RANGE = -1
34	};
35
36	static int cmpfunc(const void a, const void b)
37	{
38	return (((struct cut_list *) a)->startpos -
39	((struct cut_list *) b)->startpos);
40
41	}
42
43	static void cut_file(FILE file, char delim, const struct cut_list cut_lists, unsigned nlists)
44	{
45	char *line;
46	unsigned linenum = 0; /* keep these zero-based to be consistent */
47
48	/* go through every line in the file */
49	while ((line = xmalloc_fgetline(file)) != NULL) {
50
51	/* set up a list so we can keep track of what's been printed */
52	int linelen = strlen(line);
53	char *printed = xzalloc(linelen + 1);
54	char *orig_line = line;
55	unsigned cl_pos = 0;
56	int spos;
57
58	/* cut based on chars/bytes XXX: only works when sizeof(char) == byte */
59	if (option_mask32 & (CUT_OPT_CHAR_FLGS \| CUT_OPT_BYTE_FLGS)) {
60	/* print the chars specified in each cut list */
61	for (; cl_pos < nlists; cl_pos++) {
62	spos = cut_lists[cl_pos].startpos;
63	while (spos < linelen) {
64	if (!printed[spos]) {
65	printed[spos] = 'X';
66	putchar(line[spos]);
67	}
68	spos++;
69	if (spos > cut_lists[cl_pos].endpos
70	/* NON_RANGE is -1, so if below is true,
71	* the above was true too (spos is >= 0) */
72	/* \|\| cut_lists[cl_pos].endpos == NON_RANGE */
73	) {
74	break;
75	}
76	}
77	}
78	} else if (delim == '\n') { /* cut by lines */
79	spos = cut_lists[cl_pos].startpos;
80
81	/* get out if we have no more lists to process or if the lines
82	* are lower than what we're interested in */
83	if (((int)linenum < spos) \|\| (cl_pos >= nlists))
84	goto next_line;
85
86	/* if the line we're looking for is lower than the one we were
87	* passed, it means we displayed it already, so move on */
88	while (spos < (int)linenum) {
89	spos++;
90	/* go to the next list if we're at the end of this one */
91	if (spos > cut_lists[cl_pos].endpos
92	\|\| cut_lists[cl_pos].endpos == NON_RANGE
93	) {
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 ((int)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
165	int cut_main(int argc, char **argv) MAIN_EXTERNALLY_VISIBLE;
166	int cut_main(int argc UNUSED_PARAM, char **argv)
167	{
168	/* growable array holding a series of lists */
169	struct cut_list *cut_lists = NULL;
170	unsigned nlists = 0; /* number of elements in above list */
171	char delim = '\t'; /* delimiter, default is tab */
172	char sopt, ltok;
173	unsigned opt;
174
175	opt_complementary = "b--bcf:c--bcf:f--bcf";
176	opt = getopt32(argv, optstring, &sopt, &sopt, &sopt, &ltok);
177	// argc -= optind;
178	argv += optind;
179	if (!(opt & (CUT_OPT_BYTE_FLGS \| CUT_OPT_CHAR_FLGS \| CUT_OPT_FIELDS_FLGS)))
180	bb_error_msg_and_die("expected a list of bytes, characters, or fields");
181
182	if (opt & CUT_OPT_DELIM_FLGS) {
183	if (ltok[0] && ltok[1]) { /* more than 1 char? */
184	bb_error_msg_and_die("the delimiter must be a single character");
185	}
186	delim = ltok[0];
187	}
188
189	/* non-field (char or byte) cutting has some special handling */
190	if (!(opt & CUT_OPT_FIELDS_FLGS)) {
191	static const char _op_on_field[] ALIGN1 = " only when operating on fields";
192
193	if (opt & CUT_OPT_SUPPRESS_FLGS) {
194	bb_error_msg_and_die
195	("suppressing non-delimited lines makes sense%s",
196	_op_on_field);
197	}
198	if (delim != '\t') {
199	bb_error_msg_and_die
200	("a delimiter may be specified%s", _op_on_field);
201	}
202	}
203
204	/*
205	* parse list and put values into startpos and endpos.
206	* valid list formats: N, N-, N-M, -M
207	* more than one list can be separated by commas
208	*/
209	{
210	char *ntok;
211	int s = 0, e = 0;
212
213	/* take apart the lists, one by one (they are separated with commas) */
214	while ((ltok = strsep(&sopt, ",")) != NULL) {
215
216	/* it's actually legal to pass an empty list */
217	if (!ltok[0])
218	continue;
219
220	/* get the start pos */
221	ntok = strsep(&ltok, "-");
222	if (!ntok[0]) {
223	s = BOL;
224	} else {
225	s = xatoi_u(ntok);
226	/* account for the fact that arrays are zero based, while
227	* the user expects the first char on the line to be char #1 */
228	if (s != 0)
229	s--;
230	}
231
232	/* get the end pos */
233	if (ltok == NULL) {
234	e = NON_RANGE;
235	} else if (!ltok[0]) {
236	e = EOL;
237	} else {
238	e = xatoi_u(ltok);
239	/* if the user specified and end position of 0,
240	* that means "til the end of the line" */
241	if (e == 0)
242	e = EOL;
243	e--; /* again, arrays are zero based, lines are 1 based */
244	if (e == s)
245	e = NON_RANGE;
246	}
247
248	/* add the new list */
249	cut_lists = xrealloc_vector(cut_lists, 4, nlists);
250	/* NB: startpos is always >= 0,
251	* while endpos may be = NON_RANGE (-1) */
252	cut_lists[nlists].startpos = s;
253	cut_lists[nlists].endpos = e;
254	nlists++;
255	}
256
257	/* make sure we got some cut positions out of all that */
258	if (nlists == 0)
259	bb_error_msg_and_die("missing list of positions");
260
261	/* now that the lists are parsed, we need to sort them to make life
262	* easier on us when it comes time to print the chars / fields / lines
263	*/
264	qsort(cut_lists, nlists, sizeof(struct cut_list), cmpfunc);
265	}
266
267	{
268	int retval = EXIT_SUCCESS;
269
270	if (!*argv)
271	--argv = (char )"-";
272
273	do {
274	FILE file = fopen_or_warn_stdin(argv);
275	if (!file) {
276	retval = EXIT_FAILURE;
277	continue;
278	}
279	cut_file(file, delim, cut_lists, nlists);
280	fclose_if_not_stdin(file);
281	} while (*++argv);
282
283	if (ENABLE_FEATURE_CLEAN_UP)
284	free(cut_lists);
285	fflush_stdout_and_exit(retval);
286	}
287	}