busybox/coreutils/wc.c

/* vi: set sw=4 ts=4: */
/*
 * wc implementation for busybox
 *
 * Copyright (C) 2003  Manuel Novoa III  <mjn3@codepoet.org>
 *
 * Licensed under GPLv2 or later, see file LICENSE in this tarball for details.
 */

/* BB_AUDIT SUSv3 _NOT_ compliant -- option -m is not currently supported. */
/* http://www.opengroup.org/onlinepubs/007904975/utilities/wc.html */

/* Mar 16, 2003      Manuel Novoa III   (mjn3@codepoet.org)
 *
 * Rewritten to fix a number of problems and do some size optimizations.
 * Problems in the previous busybox implementation (besides bloat) included:
 *  1) broken 'wc -c' optimization (read note below)
 *  2) broken handling of '-' args
 *  3) no checking of ferror on EOF returns
 *  4) isprint() wasn't considered when word counting.
 *
 * TODO:
 *
 * When locale support is enabled, count multibyte chars in the '-m' case.
 *
 * NOTES:
 *
 * The previous busybox wc attempted an optimization using stat for the
 * case of counting chars only.  I omitted that because it was broken.
 * It didn't take into account the possibility of input coming from a
 * pipe, or input from a file with file pointer not at the beginning.
 *
 * To implement such a speed optimization correctly, not only do you
 * need the size, but also the file position.  Note also that the
 * file position may be past the end of file.  Consider the example
 * (adapted from example in gnu wc.c)
 *
 *      echo hello > /tmp/testfile &&
 *      (dd ibs=1k skip=1 count=0 &> /dev/null; wc -c) < /tmp/testfile
 *
 * for which 'wc -c' should output '0'.
 */

#include "libbb.h"

#if !ENABLE_LOCALE_SUPPORT
# undef isprint
# undef isspace
# define isprint(c) ((unsigned)((c) - 0x20) <= (0x7e - 0x20))
# define isspace(c) ((c) == ' ')
#endif

#if ENABLE_FEATURE_WC_LARGE
# define COUNT_T unsigned long long
# define COUNT_FMT "llu"
#else
# define COUNT_T unsigned
# define COUNT_FMT "u"
#endif

enum {
	WC_LINES	= 0,
	WC_WORDS	= 1,
	WC_CHARS	= 2,
	WC_LENGTH	= 3
};

int wc_main(int argc, char **argv) MAIN_EXTERNALLY_VISIBLE;
int wc_main(int argc UNUSED_PARAM, char **argv)
{
	const char *arg;
	const char *start_fmt = " %9"COUNT_FMT + 1;
	const char *fname_fmt = " %s\n";
	COUNT_T *pcounts;
	COUNT_T counts[4];
	COUNT_T totals[4];
	int num_files;
	smallint status = EXIT_SUCCESS;
	unsigned print_type;

	print_type = getopt32(argv, "lwcL");

	if (print_type == 0) {
		print_type = (1 << WC_LINES) | (1 << WC_WORDS) | (1 << WC_CHARS);
	}

	argv += optind;
	if (!argv[0]) {
		*--argv = (char *) bb_msg_standard_input;
		fname_fmt = "\n";
	}
	if (!argv[1]) { /* zero or one filename? */
		if (!((print_type-1) & print_type)) /* exactly one option? */
			start_fmt = "%"COUNT_FMT;
	}

	memset(totals, 0, sizeof(totals));

	pcounts = counts;

	num_files = 0;
	while ((arg = *argv++) != 0) {
		FILE *fp;
		const char *s;
		unsigned u;
		unsigned linepos;
		smallint in_word;

		++num_files;
		fp = fopen_or_warn_stdin(arg);
		if (!fp) {
			status = EXIT_FAILURE;
			continue;
		}

		memset(counts, 0, sizeof(counts));
		linepos = 0;
		in_word = 0;

		do {
			int c;
			/* Our -w doesn't match GNU wc exactly... oh well */

			++counts[WC_CHARS];
			c = getc(fp);
			if (c == EOF) {
				if (ferror(fp)) {
					bb_simple_perror_msg(arg);
					status = EXIT_FAILURE;
				}
				--counts[WC_CHARS];
				goto DO_EOF;		/* Treat an EOF as '\r'. */
			}
			if (isprint_asciionly(c)) {
				++linepos;
				if (!isspace(c)) {
					in_word = 1;
					continue;
				}
			} else if ((unsigned)(c - 9) <= 4) {
				/* \t  9
				 * \n 10
				 * \v 11
				 * \f 12
				 * \r 13
				 */
				if (c == '\t') {
					linepos = (linepos | 7) + 1;
				} else {			/* '\n', '\r', '\f', or '\v' */
 DO_EOF:
					if (linepos > counts[WC_LENGTH]) {
						counts[WC_LENGTH] = linepos;
					}
					if (c == '\n') {
						++counts[WC_LINES];
					}
					if (c != '\v') {
						linepos = 0;
					}
				}
			} else {
				continue;
			}

			counts[WC_WORDS] += in_word;
			in_word = 0;
			if (c == EOF) {
				break;
			}
		} while (1);

		if (totals[WC_LENGTH] < counts[WC_LENGTH]) {
			totals[WC_LENGTH] = counts[WC_LENGTH];
		}
		totals[WC_LENGTH] -= counts[WC_LENGTH];

		fclose_if_not_stdin(fp);

 OUTPUT:
		/* coreutils wc tries hard to print pretty columns
		 * (saves results for all files, find max col len etc...)
		 * we won't try that hard, it will bloat us too much */
		s = start_fmt;
		u = 0;
		do {
			if (print_type & (1 << u)) {
				printf(s, pcounts[u]);
				s = " %9"COUNT_FMT; /* Ok... restore the leading space. */
			}
			totals[u] += pcounts[u];
		} while (++u < 4);
		printf(fname_fmt, arg);
	}

	/* If more than one file was processed, we want the totals.  To save some
	 * space, we set the pcounts ptr to the totals array.  This has the side
	 * effect of trashing the totals array after outputting it, but that's
	 * irrelavent since we no longer need it. */
	if (num_files > 1) {
		num_files = 0;				/* Make sure we don't get here again. */
		arg = "total";
		pcounts = totals;
		--argv;
		goto OUTPUT;
	}

	fflush_stdout_and_exit(status);
}
1	niro	532	/* vi: set sw=4 ts=4: */
2			/*
3			* wc implementation for busybox
4			*
5			* Copyright (C) 2003 Manuel Novoa III <mjn3@codepoet.org>
6			*
7			* Licensed under GPLv2 or later, see file LICENSE in this tarball for details.
8			*/
9
10			/* BB_AUDIT SUSv3 _NOT_ compliant -- option -m is not currently supported. */
11			/* http://www.opengroup.org/onlinepubs/007904975/utilities/wc.html */
12
13			/* Mar 16, 2003 Manuel Novoa III (mjn3@codepoet.org)
14			*
15			* Rewritten to fix a number of problems and do some size optimizations.
16			* Problems in the previous busybox implementation (besides bloat) included:
17			* 1) broken 'wc -c' optimization (read note below)
18			* 2) broken handling of '-' args
19			* 3) no checking of ferror on EOF returns
20			* 4) isprint() wasn't considered when word counting.
21			*
22			* TODO:
23			*
24			* When locale support is enabled, count multibyte chars in the '-m' case.
25			*
26			* NOTES:
27			*
28			* The previous busybox wc attempted an optimization using stat for the
29			* case of counting chars only. I omitted that because it was broken.
30			* It didn't take into account the possibility of input coming from a
31			* pipe, or input from a file with file pointer not at the beginning.
32			*
33			* To implement such a speed optimization correctly, not only do you
34			* need the size, but also the file position. Note also that the
35			* file position may be past the end of file. Consider the example
36			* (adapted from example in gnu wc.c)
37			*
38			* echo hello > /tmp/testfile &&
39	niro	816	* (dd ibs=1k skip=1 count=0 &> /dev/null; wc -c) < /tmp/testfile
40	niro	532	*
41			* for which 'wc -c' should output '0'.
42			*/
43
44	niro	816	#include "libbb.h"
45	niro	532
46	niro	984	#if !ENABLE_LOCALE_SUPPORT
47			# undef isprint
48			# undef isspace
49			# define isprint(c) ((unsigned)((c) - 0x20) <= (0x7e - 0x20))
50			# define isspace(c) ((c) == ' ')
51	niro	532	#endif
52
53			#if ENABLE_FEATURE_WC_LARGE
54	niro	984	# define COUNT_T unsigned long long
55			# define COUNT_FMT "llu"
56	niro	532	#else
57	niro	984	# define COUNT_T unsigned
58			# define COUNT_FMT "u"
59	niro	532	#endif
60
61			enum {
62			WC_LINES = 0,
63			WC_WORDS = 1,
64			WC_CHARS = 2,
65			WC_LENGTH = 3
66			};
67
68	niro	816	int wc_main(int argc, char **argv) MAIN_EXTERNALLY_VISIBLE;
69			int wc_main(int argc UNUSED_PARAM, char **argv)
70	niro	532	{
71	niro	984	const char *arg;
72	niro	816	const char *start_fmt = " %9"COUNT_FMT + 1;
73	niro	532	const char *fname_fmt = " %s\n";
74			COUNT_T *pcounts;
75			COUNT_T counts[4];
76			COUNT_T totals[4];
77	niro	984	int num_files;
78	niro	816	smallint status = EXIT_SUCCESS;
79	niro	532	unsigned print_type;
80
81	niro	816	print_type = getopt32(argv, "lwcL");
82	niro	532
83			if (print_type == 0) {
84			print_type = (1 << WC_LINES) \| (1 << WC_WORDS) \| (1 << WC_CHARS);
85			}
86
87			argv += optind;
88			if (!argv[0]) {
89			--argv = (char ) bb_msg_standard_input;
90			fname_fmt = "\n";
91	niro	984	}
92			if (!argv[1]) { /* zero or one filename? */
93	niro	532	if (!((print_type-1) & print_type)) /* exactly one option? */
94			start_fmt = "%"COUNT_FMT;
95			}
96
97			memset(totals, 0, sizeof(totals));
98
99			pcounts = counts;
100
101	niro	984	num_files = 0;
102	niro	532	while ((arg = *argv++) != 0) {
103	niro	984	FILE *fp;
104			const char *s;
105			unsigned u;
106			unsigned linepos;
107			smallint in_word;
108
109	niro	532	++num_files;
110			fp = fopen_or_warn_stdin(arg);
111			if (!fp) {
112			status = EXIT_FAILURE;
113			continue;
114			}
115
116			memset(counts, 0, sizeof(counts));
117			linepos = 0;
118			in_word = 0;
119
120			do {
121	niro	984	int c;
122	niro	532	/* Our -w doesn't match GNU wc exactly... oh well */
123
124			++counts[WC_CHARS];
125			c = getc(fp);
126	niro	984	if (c == EOF) {
127			if (ferror(fp)) {
128			bb_simple_perror_msg(arg);
129			status = EXIT_FAILURE;
130			}
131			--counts[WC_CHARS];
132			goto DO_EOF; /* Treat an EOF as '\r'. */
133			}
134			if (isprint_asciionly(c)) {
135	niro	532	++linepos;
136	niro	984	if (!isspace(c)) {
137	niro	532	in_word = 1;
138			continue;
139			}
140	niro	984	} else if ((unsigned)(c - 9) <= 4) {
141	niro	532	/* \t 9
142			* \n 10
143			* \v 11
144			* \f 12
145			* \r 13
146			*/
147			if (c == '\t') {
148			linepos = (linepos \| 7) + 1;
149			} else { /* '\n', '\r', '\f', or '\v' */
150	niro	984	DO_EOF:
151	niro	532	if (linepos > counts[WC_LENGTH]) {
152			counts[WC_LENGTH] = linepos;
153			}
154			if (c == '\n') {
155			++counts[WC_LINES];
156			}
157			if (c != '\v') {
158			linepos = 0;
159			}
160			}
161			} else {
162			continue;
163			}
164
165			counts[WC_WORDS] += in_word;
166			in_word = 0;
167			if (c == EOF) {
168			break;
169			}
170			} while (1);
171
172			if (totals[WC_LENGTH] < counts[WC_LENGTH]) {
173			totals[WC_LENGTH] = counts[WC_LENGTH];
174			}
175			totals[WC_LENGTH] -= counts[WC_LENGTH];
176
177			fclose_if_not_stdin(fp);
178
179	niro	984	OUTPUT:
180	niro	532	/* coreutils wc tries hard to print pretty columns
181			* (saves results for all files, find max col len etc...)
182			* we won't try that hard, it will bloat us too much */
183			s = start_fmt;
184			u = 0;
185			do {
186			if (print_type & (1 << u)) {
187			printf(s, pcounts[u]);
188			s = " %9"COUNT_FMT; /* Ok... restore the leading space. */
189			}
190			totals[u] += pcounts[u];
191			} while (++u < 4);
192			printf(fname_fmt, arg);
193			}
194
195			/* If more than one file was processed, we want the totals. To save some
196			* space, we set the pcounts ptr to the totals array. This has the side
197			* effect of trashing the totals array after outputting it, but that's
198			* irrelavent since we no longer need it. */
199			if (num_files > 1) {
200			num_files = 0; /* Make sure we don't get here again. */
201			arg = "total";
202			pcounts = totals;
203			--argv;
204			goto OUTPUT;
205			}
206
207			fflush_stdout_and_exit(status);
208			}