/* vi: set sw=4 ts=4: */
/*
- * Mini wc implementation for busybox
+ * wc implementation for busybox
*
- * by Edward Betts <edward@debian.org>
+ * Copyright (C) 2003 Manuel Novoa III <mjn3@codepoet.org>
*
- * This program is free software; you can redistribute it and/or modify
- * it under the terms of the GNU General Public License as published by
- * the Free Software Foundation; either version 2 of the License, or
- * (at your option) any later version.
+ * Licensed under GPLv2 or later, see file LICENSE in this source tree.
+ */
+/* 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.
+ *
+ * NOTES:
*
- * This program is distributed in the hope that it will be useful,
- * but WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
- * General Public License for more details.
+ * 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.
*
- * You should have received a copy of the GNU General Public License
- * along with this program; if not, write to the Free Software
- * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
+ * 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'.
*/
+//config:config WC
+//config: bool "wc (4.4 kb)"
+//config: default y
+//config: help
+//config: wc is used to print the number of bytes, words, and lines,
+//config: in specified files.
+//config:
+//config:config FEATURE_WC_LARGE
+//config: bool "Support very large counts"
+//config: default y
+//config: depends on WC
+//config: help
+//config: Use "unsigned long long" for counter variables.
+
+//applet:IF_WC(APPLET(wc, BB_DIR_USR_BIN, BB_SUID_DROP))
-#include "internal.h"
-#include <stdio.h>
+//kbuild:lib-$(CONFIG_WC) += wc.o
-static const char wc_usage[] = "wc [OPTION]... [FILE]...\n\n"
- "Print line, word, and byte counts for each FILE, and a total line if\n"
- "more than one FILE is specified. With no FILE, read standard input.\n"
- "\t-c\tprint the byte counts\n"
- "\t-l\tprint the newline counts\n"
+/* BB_AUDIT SUSv3 compliant. */
+/* http://www.opengroup.org/onlinepubs/007904975/utilities/wc.html */
+
+#include "libbb.h"
+#include "unicode.h"
+
+#if !ENABLE_LOCALE_SUPPORT
+# undef isprint
+# undef isspace
+# define isprint(c) ((unsigned)((c) - 0x20) <= (0x7e - 0x20))
+# define isspace(c) ((c) == ' ')
+#endif
+
+#if ENABLE_FEATURE_WC_LARGE
+# define COUNT_T unsigned long long
+# define COUNT_FMT "llu"
+#else
+# define COUNT_T unsigned
+# define COUNT_FMT "u"
+#endif
+
+/* We support -m even when UNICODE_SUPPORT is off,
+ * we just don't advertise it in help text,
+ * since it is the same as -c in this case.
+ */
- "\t-L\tprint the length of the longest line\n"
- "\t-w\tprint the word counts\n";
+//usage:#define wc_trivial_usage
+//usage: "[-c"IF_UNICODE_SUPPORT("m")"lwL] [FILE]..."
+//usage:
+//usage:#define wc_full_usage "\n\n"
+//usage: "Count lines, words, and bytes for each FILE (or stdin)\n"
+//usage: "\n -c Count bytes"
+//usage: IF_UNICODE_SUPPORT(
+//usage: "\n -m Count characters"
+//usage: )
+//usage: "\n -l Count newlines"
+//usage: "\n -w Count words"
+//usage: "\n -L Print longest line length"
+//usage:
+//usage:#define wc_example_usage
+//usage: "$ wc /etc/passwd\n"
+//usage: " 31 46 1365 /etc/passwd\n"
-static int total_lines, total_words, total_chars, max_length;
-static int print_lines, print_words, print_chars, print_length;
+/* Order is important if we want to be compatible with
+ * column order in "wc -cmlwL" output:
+ */
+enum {
+ WC_LINES = 0, /* -l */
+ WC_WORDS = 1, /* -w */
+ WC_UNICHARS = 2, /* -m */
+ WC_BYTES = 3, /* -c */
+ WC_LENGTH = 4, /* -L */
+ NUM_WCS = 5,
+};
-void print_counts(int lines, int words, int chars, int length,
- const char *name)
+int wc_main(int argc, char **argv) MAIN_EXTERNALLY_VISIBLE;
+int wc_main(int argc UNUSED_PARAM, char **argv)
{
- char const *space = "";
+ 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_lines) {
- printf("%7d", lines);
- space = " ";
+ if (print_type == 0) {
+ print_type = (1 << WC_LINES) | (1 << WC_WORDS) | (1 << WC_BYTES);
}
- if (print_words) {
- printf("%s%7d", space, words);
- space = " ";
+
+ argv += optind;
+ if (!argv[0]) {
+ *--argv = (char *) bb_msg_standard_input;
+ fname_fmt = "\n";
}
- if (print_chars) {
- printf("%s%7d", space, chars);
- space = " ";
+ if (!argv[1]) { /* zero or one filename? */
+ if (!((print_type-1) & print_type)) /* exactly one option? */
+ start_fmt = "%"COUNT_FMT;
}
- if (print_length)
- printf("%s%7d", space, length);
- if (*name)
- printf(" %s", name);
- putchar('\n');
-}
-static void wc_file(FILE * file, const char *name)
-{
- int lines, words, chars, length;
- int in_word = 0, linepos = 0;
- int c;
-
- lines = words = chars = length = 0;
- while ((c = getc(file)) != EOF) {
- chars++;
- switch (c) {
- case '\n':
- lines++;
- case '\r':
- case '\f':
- if (linepos > length)
- length = linepos;
- linepos = 0;
- goto word_separator;
- case '\t':
- linepos += 8 - (linepos % 8);
- goto word_separator;
- case ' ':
- linepos++;
- case '\v':
- word_separator:
- if (in_word) {
- in_word = 0;
- words++;
- }
- break;
- default:
- linepos++;
- in_word = 1;
- break;
+ memset(totals, 0, sizeof(totals));
+
+ pcounts = counts;
+
+ num_files = 0;
+ while ((arg = *argv++) != NULL) {
+ 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;
}
- }
- if (linepos > length)
- length = linepos;
- if (in_word)
- words++;
- print_counts(lines, words, chars, length, name);
- total_lines += lines;
- total_words += words;
- total_chars += chars;
- if (length > max_length)
- max_length = length;
- fclose(file);
- fflush(stdout);
-}
-int wc_main(int argc, char **argv)
-{
- FILE *file;
+ memset(counts, 0, sizeof(counts));
+ linepos = 0;
+ in_word = 0;
- total_lines = total_words = total_chars = max_length = 0;
- print_lines = print_words = print_chars = print_length = 0;
+ while (1) {
+ int c;
+ /* Our -w doesn't match GNU wc exactly... oh well */
- while (--argc && **(++argv) == '-') {
- while (*++(*argv))
- switch (**argv) {
- case 'c':
- print_chars = 1;
- break;
- case 'l':
- print_lines = 1;
- break;
- case 'L':
- print_length = 1;
- break;
- case 'w':
- print_words = 1;
+ c = getc(fp);
+ 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(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;
- default:
- usage(wc_usage);
}
- }
+ }
+
+ fclose_if_not_stdin(fp);
- if (!print_lines && !print_words && !print_chars && !print_length)
- print_lines = print_words = print_chars = 1;
-
- if (argc == 0) {
- wc_file(stdin, "");
- exit(TRUE);
- } else if (argc == 1) {
- file = fopen(*argv, "r");
- if (file == NULL) {
- perror(*argv);
- exit(FALSE);
+ if (totals[WC_LENGTH] < counts[WC_LENGTH]) {
+ totals[WC_LENGTH] = counts[WC_LENGTH];
}
- wc_file(file, *argv);
- } else {
- while (argc-- > 0 && *argv != '\0' && strlen(*argv)) {
- file = fopen(*argv, "r");
- if (file == NULL) {
- perror(*argv);
- exit(FALSE);
+ totals[WC_LENGTH] -= counts[WC_LENGTH];
+
+ 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)) {
+ printf(s, pcounts[u]);
+ s = " %9"COUNT_FMT; /* Ok... restore the leading space. */
}
- wc_file(file, *argv);
- argv++;
- }
- print_counts(total_lines, total_words, total_chars,
- max_length, "total");
+ totals[u] += pcounts[u];
+ } while (++u < NUM_WCS);
+ printf(fname_fmt, arg);
}
- exit(TRUE);
+
+ /* 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);
}