wc/historical/busybox-wc_comments_stripped.c

#include "libbb.h"
#include "unicode.h"

/* Copyright (C) 2003  Manuel Novoa III  <mjn3@codepoet.org>
/*
/* Licensed under GPLv2 or later, see file LICENSE in this source tree.
#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

enum {
	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;
int wc_main(int argc UNUSED_PARAM, char **argv)
{
	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_BYTES);
	}

	argv += optind;
	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;

	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;
		}

		memset(counts, 0, sizeof(counts));
		linepos = 0;
		in_word = 0;

		while (1) {
			int c;

			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'. */
			}

			++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) {
				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;
			}
		}

		fclose_if_not_stdin(fp);

		if (totals[WC_LENGTH] < counts[WC_LENGTH]) {
			totals[WC_LENGTH] = counts[WC_LENGTH];
		}
		totals[WC_LENGTH] -= counts[WC_LENGTH];

 OUTPUT:
		s = start_fmt;
		u = 0;
		do {
			if (print_type & (1 << u)) {
				printf(s, pcounts[u]);
				s = " %9"COUNT_FMT; /* Ok... restore the leading space. */
			}
			totals[u] += pcounts[u];
		} while (++u < NUM_WCS);
		printf(fname_fmt, arg);
	}

	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);
}