/*
* uniq implementation for busybox
*
- * Copyright (C) 2003 Manuel Novoa III <mjn3@codepoet.org>
- *
- * Licensed under GPL v2, see file LICENSE in this tarball for details.
+ * Copyright (C) 2005 Manuel Novoa III <mjn3@codepoet.org>
*
+ * Licensed under GPLv2 or later, see file LICENSE in this source tree.
*/
+//config:config UNIQ
+//config: bool "uniq (4.8 kb)"
+//config: default y
+//config: help
+//config: uniq is used to remove duplicate lines from a sorted file.
+
+//applet:IF_UNIQ(APPLET(uniq, BB_DIR_USR_BIN, BB_SUID_DROP))
+
+//kbuild:lib-$(CONFIG_UNIQ) += uniq.o
/* BB_AUDIT SUSv3 compliant */
/* http://www.opengroup.org/onlinepubs/007904975/utilities/uniq.html */
-#include <stdio.h>
-#include <stdlib.h>
-#include <string.h>
-#include <ctype.h>
-#include <unistd.h>
-#include "busybox.h"
-#include "libcoreutils/coreutils.h"
+//usage:#define uniq_trivial_usage
+//usage: "[-cdu][-f,s,w N] [INPUT [OUTPUT]]"
+//usage:#define uniq_full_usage "\n\n"
+//usage: "Discard duplicate lines\n"
+//usage: "\n -c Prefix lines by the number of occurrences"
+//usage: "\n -d Only print duplicate lines"
+//usage: "\n -u Only print unique lines"
+//usage: "\n -i Ignore case"
+//usage: "\n -f N Skip first N fields"
+//usage: "\n -s N Skip first N chars (after any skipped fields)"
+//usage: "\n -w N Compare N characters in line"
+//usage:
+//usage:#define uniq_example_usage
+//usage: "$ echo -e \"a\\na\\nb\\nc\\nc\\na\" | sort | uniq\n"
+//usage: "a\n"
+//usage: "b\n"
+//usage: "c\n"
+
+#include "libbb.h"
+
+int uniq_main(int argc, char **argv) MAIN_EXTERNALLY_VISIBLE;
+int uniq_main(int argc UNUSED_PARAM, char **argv)
+{
+ const char *input_filename;
+ unsigned skip_fields, skip_chars, max_chars;
+ unsigned opt;
+ char *cur_line;
+ const char *cur_compare;
+
+ enum {
+ OPT_c = 0x1,
+ OPT_d = 0x2, /* print only dups */
+ OPT_u = 0x4, /* print only uniq */
+ OPT_f = 0x8,
+ OPT_s = 0x10,
+ OPT_w = 0x20,
+ OPT_i = 0x40,
+ };
-/* The extra data is flags to make -d and -u switch each other off */
-static const char uniq_opts[] = "cudf:s:\0\7\3\5\1\2\4";
+ skip_fields = skip_chars = 0;
+ max_chars = INT_MAX;
-#define SHOW_COUNT 1
-#define SHOW_UNIQUE 2
-#define SHOW_DUPLICATE 4
+ opt = getopt32(argv, "cduf:+s:+w:+i", &skip_fields, &skip_chars, &max_chars);
+ argv += optind;
-static FILE *open_arg(char **argv, char *mode)
-{
- char *n=*argv;
-
- return (n && *n != '-' && n[1]) ? bb_xfopen(n, mode) :
- *mode=='r' ? stdin : stdout;
-}
+ input_filename = argv[0];
+ if (input_filename) {
+ const char *output;
+ if (input_filename[0] != '-' || input_filename[1]) {
+ close(STDIN_FILENO); /* == 0 */
+ xopen(input_filename, O_RDONLY); /* fd will be 0 */
+ }
+ output = argv[1];
+ if (output) {
+ if (argv[2])
+ bb_show_usage();
+ if (output[0] != '-' || output[1]) {
+ // Won't work with "uniq - FILE" and closed stdin:
+ //close(STDOUT_FILENO);
+ //xopen(output, O_WRONLY | O_CREAT | O_TRUNC);
+ xmove_fd(xopen(output, O_WRONLY | O_CREAT | O_TRUNC), STDOUT_FILENO);
+ }
+ }
+ }
-int uniq_main(int argc, char **argv)
-{
- FILE *in, *out;
- unsigned long dups, skip_fields, skip_chars, i;
- const char *oldline, *oldskipped, *line, *skipped, *input_filename;
- int opt;
- int uniq_flags = SHOW_UNIQUE | SHOW_DUPLICATE;
+ cur_compare = cur_line = NULL; /* prime the pump */
- skip_fields = skip_chars = 0;
+ do {
+ unsigned i;
+ unsigned long dups;
+ char *old_line;
+ const char *old_compare;
- while ((opt = getopt(argc, argv, uniq_opts)) > 0) {
- if (opt == 'f') skip_fields = bb_xgetularg10(optarg);
- else if (opt == 's') skip_chars = bb_xgetularg10(optarg);
-
- /* This bit uses the extra data at the end of uniq_opts to make
- * -d and -u switch each other off in a very small amount of space */
-
- else if ((line = strchr(uniq_opts, opt)) != NULL) {
- uniq_flags &= line[8];
- uniq_flags |= line[11];
- } else bb_show_usage();
- }
+ old_line = cur_line;
+ old_compare = cur_compare;
+ dups = 0;
- input_filename = *(argv += optind);
-
- in = open_arg(argv, "r");
- if (*argv) ++argv;
- out = open_arg(argv, "w");
- if (*argv && argv[1]) bb_show_usage();
-
- line = skipped = NULL;
-
-NOT_DUPLICATE:
- oldline = line;
- oldskipped = skipped;
- dups = 0;
-
- /* gnu uniq ignores newlines */
- while ((line = bb_get_chomped_line_from_file(in)) != NULL) {
- skipped = line;
- for (i=skip_fields ; i ; i--) {
- skipped = bb_skip_whitespace(skipped);
- while (*skipped && !isspace(*skipped)) ++skipped;
- }
- for (i = skip_chars ; *skipped && i ; i--) ++skipped;
- if (oldline) {
- if (!strcmp(oldskipped, skipped)) {
- ++dups; /* Note: Testing for overflow seems excessive. */
- continue;
+ /* gnu uniq ignores newlines */
+ while ((cur_line = xmalloc_fgetline(stdin)) != NULL) {
+ cur_compare = cur_line;
+ for (i = skip_fields; i; i--) {
+ cur_compare = skip_whitespace(cur_compare);
+ cur_compare = skip_non_whitespace(cur_compare);
}
-DO_LAST:
- if (uniq_flags & (dups ? SHOW_DUPLICATE : SHOW_UNIQUE)) {
- bb_fprintf(out, "\0%7d " + (uniq_flags & SHOW_COUNT), dups + 1);
- bb_fprintf(out, "%s\n", oldline);
+ for (i = skip_chars; *cur_compare && i; i--) {
+ ++cur_compare;
}
- free((void *)oldline);
+
+ if (!old_line)
+ break;
+ if ((opt & OPT_i)
+ ? strncasecmp(old_compare, cur_compare, max_chars)
+ : strncmp(old_compare, cur_compare, max_chars)
+ ) {
+ break;
+ }
+
+ free(cur_line);
+ ++dups; /* testing for overflow seems excessive */
}
- goto NOT_DUPLICATE;
- }
- if (oldline) goto DO_LAST;
+ if (old_line) {
+ if (!(opt & (OPT_d << !!dups))) { /* (if dups, opt & OPT_u) */
+ if (opt & OPT_c) {
+ /* %7lu matches GNU coreutils 6.9 */
+ printf("%7lu ", dups + 1);
+ }
+ puts(old_line);
+ }
+ free(old_line);
+ }
+ } while (cur_line);
- bb_xferror(in, input_filename);
+ die_if_ferror(stdin, input_filename);
- bb_fflush_stdout_and_exit(EXIT_SUCCESS);
+ fflush_stdout_and_exit(EXIT_SUCCESS);
}
projects / oweals / busybox.git / blobdiff