claenups for previous commit

[oweals/busybox.git] / coreutils / wc.c

diff --git a/coreutils/wc.c b/coreutils/wc.c
index ae38fd5fe3d755dce1f05de6ab6de653e19c2f6b..c74b7a65f35f43ff967a22805f619470d1a1e586 100644 (file)
--- a/coreutils/wc.c
+++ b/coreutils/wc.c
@@ -6,10 +6,6 @@
  *
  * Licensed under GPLv2 or later, see file LICENSE in this source tree.
  */
-
-/* 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.
@@ -19,10 +15,6 @@
  *  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
@@ -40,8 +32,29 @@
  *
  * 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))
+
+//kbuild:lib-$(CONFIG_WC) += wc.o
+
+/* 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
@@ -58,12 +71,38 @@
 # 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.
+ */
+
+//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 {
-       WC_LINES  = 0,
-       WC_WORDS  = 1,
-       WC_CHARS  = 2,
-       WC_LENGTH = 3,
-       NUM_WCS   = 4,
+       WC_LINES    = 0, /* -l */
+       WC_WORDS    = 1, /* -w */
+       WC_UNICHARS = 2, /* -m */
+       WC_BYTES    = 3, /* -c */
+       WC_LENGTH   = 4, /* -L */
+       NUM_WCS     = 5,
 };
 
 int wc_main(int argc, char **argv) MAIN_EXTERNALLY_VISIBLE;
@@ -79,10 +118,12 @@ int wc_main(int argc UNUSED_PARAM, char **argv)
        smallint status = EXIT_SUCCESS;
        unsigned print_type;
 
-       print_type = getopt32(argv, "lwcL");
+       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;
@@ -128,11 +169,18 @@ int wc_main(int argc UNUSED_PARAM, char **argv)
                                        bb_simple_perror_msg(arg);
                                        status = EXIT_FAILURE;
                                }
-                               goto DO_EOF;            /* Treat an EOF as '\r'. */
+                               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];
                        }
-                       ++counts[WC_CHARS];
 
-                       if (isprint_asciionly(c)) {
+                       if (isprint_asciionly(c)) { /* FIXME: not unicode-aware */
                                ++linepos;
                                if (!isspace(c)) {
                                        in_word = 1;
@@ -147,7 +195,7 @@ int wc_main(int argc UNUSED_PARAM, char **argv)
                                 */
                                if (c == '\t') {
                                        linepos = (linepos | 7) + 1;
-                               } else {                        /* '\n', '\r', '\f', or '\v' */
+                               } else {  /* '\n', '\r', '\f', or '\v' */
  DO_EOF:
                                        if (linepos > counts[WC_LENGTH]) {
                                                counts[WC_LENGTH] = linepos;
@@ -198,7 +246,7 @@ int wc_main(int argc UNUSED_PARAM, char **argv)
         * 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. */
+               num_files = 0;  /* Make sure we don't get here again. */
                arg = "total";
                pcounts = totals;
                --argv;