*
* Copyright (C) 2003 Manuel Novoa III <mjn3@codepoet.org>
*
- * Licensed under GPLv2 or later, see file LICENSE in this tarball for details.
+ * Licensed under GPLv2 or later, see file LICENSE in this source tree.
*/
-/* BB_AUDIT SUSv3 _NOT_ compliant -- option -m is not currently supported. */
+/* BB_AUDIT SUSv3 compliant. */
/* http://www.opengroup.org/onlinepubs/007904975/utilities/wc.html */
/* Mar 16, 2003 Manuel Novoa III (mjn3@codepoet.org)
* 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
* (adapted from example in gnu wc.c)
*
* echo hello > /tmp/testfile &&
- * (dd ibs=1k skip=1 count=0 &> /dev/null ; wc -c) < /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"
+#include "unicode.h"
+
+#if !ENABLE_LOCALE_SUPPORT
+# undef isprint
+# undef isspace
+# define isprint(c) ((unsigned)((c) - 0x20) <= (0x7e - 0x20))
+# define isspace(c) ((c) == ' ')
+#endif
-#include <stdio.h>
-#include <stdlib.h>
-#include <string.h>
-#include <unistd.h>
-#include "busybox.h"
-
-#ifdef CONFIG_LOCALE_SUPPORT
-#include <locale.h>
-#include <ctype.h>
-#define isspace_given_isprint(c) isspace(c)
+#if ENABLE_FEATURE_WC_LARGE
+# define COUNT_T unsigned long long
+# define COUNT_FMT "llu"
#else
-#undef isspace
-#undef isprint
-#define isspace(c) ((((c) == ' ') || (((unsigned int)((c) - 9)) <= (13 - 9))))
-#define isprint(c) (((unsigned int)((c) - 0x20)) <= (0x7e - 0x20))
-#define isspace_given_isprint(c) ((c) == ' ')
+# define COUNT_T unsigned
+# define COUNT_FMT "u"
#endif
-enum {
- WC_LINES = 0,
- WC_WORDS = 1,
- WC_CHARS = 2,
- WC_LENGTH = 3
-};
-
-/* Note: If this changes, remember to change the initialization of
- * 'name' in wc_main. It needs to point to the terminating nul. */
-static const char wc_opts[] = "lwcL"; /* READ THE WARNING ABOVE! */
+/* We support -m even when UNICODE_SUPPORT is off,
+ * we just don't advertise it in help text,
+ * since it is the same as -c in this case.
+ */
+//usage:#define wc_trivial_usage
+//usage: "[-c"IF_UNICODE_SUPPORT("m")"lwL] [FILE]..."
+//usage:
+//usage:#define wc_full_usage "\n\n"
+//usage: "Count lines, words, and bytes for each FILE (or stdin)\n"
+//usage: "\n -c Count bytes"
+//usage: IF_UNICODE_SUPPORT(
+//usage: "\n -m Count characters"
+//usage: )
+//usage: "\n -l Count newlines"
+//usage: "\n -w Count words"
+//usage: "\n -L Print longest line length"
+//usage:
+//usage:#define wc_example_usage
+//usage: "$ wc /etc/passwd\n"
+//usage: " 31 46 1365 /etc/passwd\n"
+
+/* Order is important if we want to be compatible with
+ * column order in "wc -cmlwL" output:
+ */
enum {
- OP_INC_LINE = 1, /* OP_INC_LINE must be 1. */
- OP_SPACE = 2,
- OP_NEWLINE = 4,
- OP_TAB = 8,
- OP_NUL = 16,
+ WC_LINES = 0, /* -l */
+ WC_WORDS = 1, /* -w */
+ WC_UNICHARS = 2, /* -m */
+ WC_BYTES = 3, /* -c */
+ WC_LENGTH = 4, /* -L */
+ NUM_WCS = 5,
};
-/* Note: If fmt_str changes, the offsets to 's' in the OUTPUT section
- * will need to be updated. */
-static const char fmt_str[] = " %7u\0 %s\n";
-static const char total_str[] = "total";
-
-int wc_main(int argc, char **argv)
+int wc_main(int argc, char **argv) MAIN_EXTERNALLY_VISIBLE;
+int wc_main(int argc UNUSED_PARAM, char **argv)
{
- FILE *fp;
- const char *s;
- unsigned int *pcounts;
- unsigned int counts[4];
- unsigned int totals[4];
- unsigned int linepos;
- unsigned int u;
- int num_files = 0;
- int c;
- char status = EXIT_SUCCESS;
- char in_word;
- char print_type;
-
- print_type = bb_getopt_ulflags(argc, argv, wc_opts);
+ const char *arg;
+ const char *start_fmt = " %9"COUNT_FMT + 1;
+ const char *fname_fmt = " %s\n";
+ COUNT_T *pcounts;
+ COUNT_T counts[NUM_WCS];
+ COUNT_T totals[NUM_WCS];
+ int num_files;
+ smallint status = EXIT_SUCCESS;
+ unsigned print_type;
+
+ init_unicode();
+
+ print_type = getopt32(argv, "lwmcL");
if (print_type == 0) {
- print_type = (1 << WC_LINES) | (1 << WC_WORDS) | (1 << WC_CHARS);
+ print_type = (1 << WC_LINES) | (1 << WC_WORDS) | (1 << WC_BYTES);
}
argv += optind;
- if (!*argv) {
+ 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;
- do {
+ num_files = 0;
+ while ((arg = *argv++) != NULL) {
+ FILE *fp;
+ const char *s;
+ unsigned u;
+ unsigned linepos;
+ smallint in_word;
+
++num_files;
- if (!(fp = bb_wfopen_input(*argv))) {
+ fp = fopen_or_warn_stdin(arg);
+ if (!fp) {
status = EXIT_FAILURE;
continue;
}
linepos = 0;
in_word = 0;
- do {
- ++counts[WC_CHARS];
+ while (1) {
+ int c;
+ /* Our -w doesn't match GNU wc exactly... oh well */
+
c = getc(fp);
- if (isprint(c)) {
+ if (c == EOF) {
+ if (ferror(fp)) {
+ bb_simple_perror_msg(arg);
+ status = EXIT_FAILURE;
+ }
+ goto DO_EOF; /* Treat an EOF as '\r'. */
+ }
+
+ /* Cater for -c and -m */
+ ++counts[WC_BYTES];
+ if (unicode_status != UNICODE_ON /* every byte is a new char */
+ || (c & 0xc0) != 0x80 /* it isn't a 2nd+ byte of a Unicode char */
+ ) {
+ ++counts[WC_UNICHARS];
+ }
+
+ if (isprint_asciionly(c)) { /* FIXME: not unicode-aware */
++linepos;
- if (!isspace_given_isprint(c)) {
+ if (!isspace(c)) {
in_word = 1;
continue;
}
- } else if (((unsigned int)(c - 9)) <= 4) {
+ } else if ((unsigned)(c - 9) <= 4) {
/* \t 9
* \n 10
* \v 11
*/
if (c == '\t') {
linepos = (linepos | 7) + 1;
- } else { /* '\n', '\r', '\f', or '\v' */
- DO_EOF:
+ } else { /* '\n', '\r', '\f', or '\v' */
+ DO_EOF:
if (linepos > counts[WC_LENGTH]) {
counts[WC_LENGTH] = linepos;
}
linepos = 0;
}
}
- } else if (c == EOF) {
- if (ferror(fp)) {
- bb_perror_msg("%s", *argv);
- status = EXIT_FAILURE;
- }
- --counts[WC_CHARS];
- goto DO_EOF; /* Treat an EOF as '\r'. */
} else {
continue;
}
if (c == EOF) {
break;
}
- } while (1);
+ }
+
+ fclose_if_not_stdin(fp);
if (totals[WC_LENGTH] < counts[WC_LENGTH]) {
totals[WC_LENGTH] = counts[WC_LENGTH];
}
totals[WC_LENGTH] -= counts[WC_LENGTH];
- bb_fclose_nonstdin(fp);
-
- OUTPUT:
- s = fmt_str + 1; /* Skip the leading space on 1st pass. */
+ OUTPUT:
+ /* coreutils wc tries hard to print pretty columns
+ * (saves results for all files, finds 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)) {
- bb_printf(s, pcounts[u]);
- s = fmt_str; /* Ok... restore the leading space. */
+ printf(s, pcounts[u]);
+ s = " %9"COUNT_FMT; /* Ok... restore the leading space. */
}
totals[u] += pcounts[u];
- } while (++u < 4);
-
- s += 8; /* Set the format to the empty string. */
-
- if (*argv != bb_msg_standard_input) {
- s -= 3; /* We have a name, so do %s conversion. */
- }
- bb_printf(s, *argv);
-
- } while (*++argv);
+ } while (++u < NUM_WCS);
+ 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. */
- *--argv = (char *) total_str;
+ num_files = 0; /* Make sure we don't get here again. */
+ arg = "total";
pcounts = totals;
+ --argv;
goto OUTPUT;
}
- bb_fflush_stdout_and_exit(status);
+ fflush_stdout_and_exit(status);
}