2000-02-08 19:58:47 +00:00
|
|
|
/* vi: set sw=4 ts=4: */
|
2000-01-15 22:28:50 +00:00
|
|
|
/*
|
2003-03-19 09:13:01 +00:00
|
|
|
* wc implementation for busybox
|
2000-01-15 22:28:50 +00:00
|
|
|
*
|
2003-03-19 09:13:01 +00:00
|
|
|
* Copyright (C) 2003 Manuel Novoa III <mjn3@codepoet.org>
|
2000-01-15 22:28:50 +00:00
|
|
|
*
|
2006-07-12 07:56:04 +00:00
|
|
|
* Licensed under GPLv2 or later, see file LICENSE in this tarball for details.
|
2000-01-15 22:28:50 +00:00
|
|
|
*/
|
|
|
|
|
2003-03-19 09:13:01 +00:00
|
|
|
/* 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.
|
2004-03-15 08:29:22 +00:00
|
|
|
* Problems in the previous busybox implementation (besides bloat) included:
|
2003-03-19 09:13:01 +00:00
|
|
|
* 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 &&
|
2007-07-21 15:08:09 +00:00
|
|
|
* (dd ibs=1k skip=1 count=0 &> /dev/null; wc -c) < /tmp/testfile
|
2003-03-19 09:13:01 +00:00
|
|
|
*
|
|
|
|
* for which 'wc -c' should output '0'.
|
|
|
|
*/
|
|
|
|
|
2007-05-26 19:00:18 +00:00
|
|
|
#include "libbb.h"
|
2000-01-15 22:28:50 +00:00
|
|
|
|
2009-10-22 20:28:08 +00:00
|
|
|
#if !ENABLE_LOCALE_SUPPORT
|
|
|
|
# undef isprint
|
|
|
|
# undef isspace
|
|
|
|
# define isprint(c) ((unsigned)((c) - 0x20) <= (0x7e - 0x20))
|
|
|
|
# define isspace(c) ((c) == ' ')
|
2003-03-19 09:13:01 +00:00
|
|
|
#endif
|
|
|
|
|
2006-09-29 23:41:59 +00:00
|
|
|
#if ENABLE_FEATURE_WC_LARGE
|
2009-10-22 20:28:08 +00:00
|
|
|
# define COUNT_T unsigned long long
|
|
|
|
# define COUNT_FMT "llu"
|
2006-09-29 23:41:59 +00:00
|
|
|
#else
|
2009-10-22 20:28:08 +00:00
|
|
|
# define COUNT_T unsigned
|
|
|
|
# define COUNT_FMT "u"
|
2006-09-29 23:41:59 +00:00
|
|
|
#endif
|
2006-09-29 23:41:04 +00:00
|
|
|
|
2003-03-19 09:13:01 +00:00
|
|
|
enum {
|
|
|
|
WC_LINES = 0,
|
|
|
|
WC_WORDS = 1,
|
|
|
|
WC_CHARS = 2,
|
|
|
|
WC_LENGTH = 3
|
2001-11-21 09:17:00 +00:00
|
|
|
};
|
2000-01-15 22:28:50 +00:00
|
|
|
|
2007-10-11 10:05:36 +00:00
|
|
|
int wc_main(int argc, char **argv) MAIN_EXTERNALLY_VISIBLE;
|
2008-07-05 09:18:54 +00:00
|
|
|
int wc_main(int argc UNUSED_PARAM, char **argv)
|
2000-01-15 22:28:50 +00:00
|
|
|
{
|
2009-11-18 09:48:09 +00:00
|
|
|
const char *arg;
|
2008-02-19 00:38:10 +00:00
|
|
|
const char *start_fmt = " %9"COUNT_FMT + 1;
|
2006-09-29 23:41:04 +00:00
|
|
|
const char *fname_fmt = " %s\n";
|
|
|
|
COUNT_T *pcounts;
|
|
|
|
COUNT_T counts[4];
|
|
|
|
COUNT_T totals[4];
|
2009-11-18 09:48:09 +00:00
|
|
|
int num_files;
|
2007-01-20 21:28:36 +00:00
|
|
|
smallint status = EXIT_SUCCESS;
|
2006-09-29 23:41:59 +00:00
|
|
|
unsigned print_type;
|
2004-03-15 08:29:22 +00:00
|
|
|
|
2007-08-18 15:32:12 +00:00
|
|
|
print_type = getopt32(argv, "lwcL");
|
2004-03-15 08:29:22 +00:00
|
|
|
|
2003-03-19 09:13:01 +00:00
|
|
|
if (print_type == 0) {
|
|
|
|
print_type = (1 << WC_LINES) | (1 << WC_WORDS) | (1 << WC_CHARS);
|
2001-11-21 12:46:36 +00:00
|
|
|
}
|
2004-03-15 08:29:22 +00:00
|
|
|
|
2003-03-19 09:13:01 +00:00
|
|
|
argv += optind;
|
2006-09-29 23:41:04 +00:00
|
|
|
if (!argv[0]) {
|
2003-03-19 09:13:01 +00:00
|
|
|
*--argv = (char *) bb_msg_standard_input;
|
2006-09-29 23:41:04 +00:00
|
|
|
fname_fmt = "\n";
|
|
|
|
if (!((print_type-1) & print_type)) /* exactly one option? */
|
|
|
|
start_fmt = "%"COUNT_FMT;
|
2001-11-21 12:46:36 +00:00
|
|
|
}
|
2004-03-15 08:29:22 +00:00
|
|
|
|
2003-03-19 09:13:01 +00:00
|
|
|
memset(totals, 0, sizeof(totals));
|
2004-03-15 08:29:22 +00:00
|
|
|
|
2003-03-19 09:13:01 +00:00
|
|
|
pcounts = counts;
|
2004-03-15 08:29:22 +00:00
|
|
|
|
2009-11-18 09:48:09 +00:00
|
|
|
num_files = 0;
|
2006-09-29 23:41:04 +00:00
|
|
|
while ((arg = *argv++) != 0) {
|
2009-11-18 09:48:09 +00:00
|
|
|
FILE *fp;
|
|
|
|
const char *s;
|
|
|
|
unsigned u;
|
|
|
|
unsigned linepos;
|
|
|
|
smallint in_word;
|
|
|
|
|
2003-03-19 09:13:01 +00:00
|
|
|
++num_files;
|
2006-10-26 23:25:17 +00:00
|
|
|
fp = fopen_or_warn_stdin(arg);
|
2006-09-29 23:41:04 +00:00
|
|
|
if (!fp) {
|
2003-03-19 09:13:01 +00:00
|
|
|
status = EXIT_FAILURE;
|
|
|
|
continue;
|
2001-11-21 12:46:36 +00:00
|
|
|
}
|
2004-03-15 08:29:22 +00:00
|
|
|
|
2003-03-19 09:13:01 +00:00
|
|
|
memset(counts, 0, sizeof(counts));
|
|
|
|
linepos = 0;
|
|
|
|
in_word = 0;
|
2004-03-15 08:29:22 +00:00
|
|
|
|
2003-03-19 09:13:01 +00:00
|
|
|
do {
|
2009-11-18 09:48:09 +00:00
|
|
|
int c;
|
2006-09-29 23:41:59 +00:00
|
|
|
/* Our -w doesn't match GNU wc exactly... oh well */
|
|
|
|
|
2003-03-19 09:13:01 +00:00
|
|
|
++counts[WC_CHARS];
|
|
|
|
c = getc(fp);
|
2009-11-18 09:48:09 +00:00
|
|
|
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'. */
|
|
|
|
}
|
2009-11-20 18:14:19 +00:00
|
|
|
if (isprint_asciionly(c)) {
|
2003-03-19 09:13:01 +00:00
|
|
|
++linepos;
|
2009-10-22 20:28:08 +00:00
|
|
|
if (!isspace(c)) {
|
2003-03-19 09:13:01 +00:00
|
|
|
in_word = 1;
|
|
|
|
continue;
|
|
|
|
}
|
2009-10-22 20:28:08 +00:00
|
|
|
} else if ((unsigned)(c - 9) <= 4) {
|
2003-03-19 09:13:01 +00:00
|
|
|
/* \t 9
|
|
|
|
* \n 10
|
|
|
|
* \v 11
|
|
|
|
* \f 12
|
|
|
|
* \r 13
|
|
|
|
*/
|
|
|
|
if (c == '\t') {
|
|
|
|
linepos = (linepos | 7) + 1;
|
|
|
|
} else { /* '\n', '\r', '\f', or '\v' */
|
2009-11-18 09:48:09 +00:00
|
|
|
DO_EOF:
|
2003-03-19 09:13:01 +00:00
|
|
|
if (linepos > counts[WC_LENGTH]) {
|
|
|
|
counts[WC_LENGTH] = linepos;
|
|
|
|
}
|
|
|
|
if (c == '\n') {
|
|
|
|
++counts[WC_LINES];
|
|
|
|
}
|
|
|
|
if (c != '\v') {
|
|
|
|
linepos = 0;
|
|
|
|
}
|
|
|
|
}
|
2001-11-21 10:26:28 +00:00
|
|
|
} else {
|
2003-03-19 09:13:01 +00:00
|
|
|
continue;
|
|
|
|
}
|
2004-03-15 08:29:22 +00:00
|
|
|
|
2003-03-19 09:13:01 +00:00
|
|
|
counts[WC_WORDS] += in_word;
|
|
|
|
in_word = 0;
|
|
|
|
if (c == EOF) {
|
|
|
|
break;
|
2001-11-21 10:26:28 +00:00
|
|
|
}
|
2003-03-19 09:13:01 +00:00
|
|
|
} while (1);
|
2004-03-15 08:29:22 +00:00
|
|
|
|
2003-03-19 09:13:01 +00:00
|
|
|
if (totals[WC_LENGTH] < counts[WC_LENGTH]) {
|
|
|
|
totals[WC_LENGTH] = counts[WC_LENGTH];
|
2001-11-21 12:46:36 +00:00
|
|
|
}
|
2003-03-19 09:13:01 +00:00
|
|
|
totals[WC_LENGTH] -= counts[WC_LENGTH];
|
2004-03-15 08:29:22 +00:00
|
|
|
|
2006-10-26 23:25:17 +00:00
|
|
|
fclose_if_not_stdin(fp);
|
2004-03-15 08:29:22 +00:00
|
|
|
|
2009-11-18 09:48:09 +00:00
|
|
|
OUTPUT:
|
2006-09-29 23:41:04 +00:00
|
|
|
/* 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;
|
2003-03-19 09:13:01 +00:00
|
|
|
u = 0;
|
|
|
|
do {
|
|
|
|
if (print_type & (1 << u)) {
|
2006-10-26 23:21:47 +00:00
|
|
|
printf(s, pcounts[u]);
|
2006-09-29 23:41:04 +00:00
|
|
|
s = " %9"COUNT_FMT; /* Ok... restore the leading space. */
|
2003-03-19 09:13:01 +00:00
|
|
|
}
|
|
|
|
totals[u] += pcounts[u];
|
|
|
|
} while (++u < 4);
|
2006-10-26 23:21:47 +00:00
|
|
|
printf(fname_fmt, arg);
|
2006-09-29 23:41:04 +00:00
|
|
|
}
|
2004-03-15 08:29:22 +00:00
|
|
|
|
2003-03-19 09:13:01 +00:00
|
|
|
/* 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. */
|
2006-09-29 23:41:04 +00:00
|
|
|
arg = "total";
|
2003-03-19 09:13:01 +00:00
|
|
|
pcounts = totals;
|
2006-09-29 23:41:04 +00:00
|
|
|
--argv;
|
2003-03-19 09:13:01 +00:00
|
|
|
goto OUTPUT;
|
2000-07-20 00:08:10 +00:00
|
|
|
}
|
2004-03-15 08:29:22 +00:00
|
|
|
|
2006-10-26 23:21:47 +00:00
|
|
|
fflush_stdout_and_exit(status);
|
2000-01-15 22:28:50 +00:00
|
|
|
}
|