1/* 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 *      (dd ibs=1k skip=1 count=0 &> /dev/null; wc -c) < /tmp/testfile
40 *
41 * for which 'wc -c' should output '0'.
42 */
43
44#include "libbb.h"
45
46#if ENABLE_LOCALE_SUPPORT
47#define isspace_given_isprint(c) isspace(c)
48#else
49#undef isspace
50#undef isprint
51#define isspace(c) ((((c) == ' ') || (((unsigned int)((c) - 9)) <= (13 - 9))))
52#define isprint(c) (((unsigned int)((c) - 0x20)) <= (0x7e - 0x20))
53#define isspace_given_isprint(c) ((c) == ' ')
54#endif
55
56#if ENABLE_FEATURE_WC_LARGE
57#define COUNT_T unsigned long long
58#define COUNT_FMT "llu"
59#else
60#define COUNT_T unsigned
61#define COUNT_FMT "u"
62#endif
63
64enum {
65	WC_LINES	= 0,
66	WC_WORDS	= 1,
67	WC_CHARS	= 2,
68	WC_LENGTH	= 3
69};
70
71int wc_main(int argc, char **argv);
72int wc_main(int argc, char **argv)
73{
74	FILE *fp;
75	const char *s, *arg;
76	const char *start_fmt = "%9"COUNT_FMT;
77	const char *fname_fmt = " %s\n";
78	COUNT_T *pcounts;
79	COUNT_T counts[4];
80	COUNT_T totals[4];
81	unsigned linepos;
82	unsigned u;
83	int num_files = 0;
84	int c;
85	smallint status = EXIT_SUCCESS;
86	smallint in_word;
87	unsigned print_type;
88
89	print_type = getopt32(argv, "lwcL");
90
91	if (print_type == 0) {
92		print_type = (1 << WC_LINES) | (1 << WC_WORDS) | (1 << WC_CHARS);
93	}
94
95	argv += optind;
96	if (!argv[0]) {
97		*--argv = (char *) bb_msg_standard_input;
98		fname_fmt = "\n";
99		if (!((print_type-1) & print_type)) /* exactly one option? */
100			start_fmt = "%"COUNT_FMT;
101	}
102
103	memset(totals, 0, sizeof(totals));
104
105	pcounts = counts;
106
107	while ((arg = *argv++) != 0) {
108		++num_files;
109		fp = fopen_or_warn_stdin(arg);
110		if (!fp) {
111			status = EXIT_FAILURE;
112			continue;
113		}
114
115		memset(counts, 0, sizeof(counts));
116		linepos = 0;
117		in_word = 0;
118
119		do {
120			/* Our -w doesn't match GNU wc exactly... oh well */
121
122			++counts[WC_CHARS];
123			c = getc(fp);
124			if (isprint(c)) {
125				++linepos;
126				if (!isspace_given_isprint(c)) {
127					in_word = 1;
128					continue;
129				}
130			} else if (((unsigned int)(c - 9)) <= 4) {
131				/* \t  9
132				 * \n 10
133				 * \v 11
134				 * \f 12
135				 * \r 13
136				 */
137				if (c == '\t') {
138					linepos = (linepos | 7) + 1;
139				} else {			/* '\n', '\r', '\f', or '\v' */
140				DO_EOF:
141					if (linepos > counts[WC_LENGTH]) {
142						counts[WC_LENGTH] = linepos;
143					}
144					if (c == '\n') {
145						++counts[WC_LINES];
146					}
147					if (c != '\v') {
148						linepos = 0;
149					}
150				}
151			} else if (c == EOF) {
152				if (ferror(fp)) {
153					bb_perror_msg("%s", arg);
154					status = EXIT_FAILURE;
155				}
156				--counts[WC_CHARS];
157				goto DO_EOF;		/* Treat an EOF as '\r'. */
158			} else {
159				continue;
160			}
161
162			counts[WC_WORDS] += in_word;
163			in_word = 0;
164			if (c == EOF) {
165				break;
166			}
167		} while (1);
168
169		if (totals[WC_LENGTH] < counts[WC_LENGTH]) {
170			totals[WC_LENGTH] = counts[WC_LENGTH];
171		}
172		totals[WC_LENGTH] -= counts[WC_LENGTH];
173
174		fclose_if_not_stdin(fp);
175
176	OUTPUT:
177		/* coreutils wc tries hard to print pretty columns
178		 * (saves results for all files, find max col len etc...)
179		 * we won't try that hard, it will bloat us too much */
180		s = start_fmt;
181		u = 0;
182		do {
183			if (print_type & (1 << u)) {
184				printf(s, pcounts[u]);
185				s = " %9"COUNT_FMT; /* Ok... restore the leading space. */
186			}
187			totals[u] += pcounts[u];
188		} while (++u < 4);
189		printf(fname_fmt, arg);
190	}
191
192	/* If more than one file was processed, we want the totals.  To save some
193	 * space, we set the pcounts ptr to the totals array.  This has the side
194	 * effect of trashing the totals array after outputting it, but that's
195	 * irrelavent since we no longer need it. */
196	if (num_files > 1) {
197		num_files = 0;				/* Make sure we don't get here again. */
198		arg = "total";
199		pcounts = totals;
200		--argv;
201		goto OUTPUT;
202	}
203
204	fflush_stdout_and_exit(status);
205}
206