Thread: structure woes equals undesired output

okay... i'm having a little trouble with the output i'm receiving. in the "context" column (read comment header below) at the end of each line are some control characters (ASCII values 01, 02, and 03).

you'll really have to compile the code and review the output to understand. and if anyone can add any improvements, that would be great too.

i've been fumbling with this for two days with no luck. thanks.

Code:

/*
 * FREQ.C -- blah blah
 *
 *  given input of "this_is_a_test" freq.c will output to a file
 *
 *	char	freq	context
 *	----	----	-------
 *	_ 	3	a, i, t
 *	a	1	_
 *	e	1	s
 *	i	2	s
 *	h	1	i
 *	s	3	_, t
 *	t	3	e, h
 *
 * the "context" column represents characters that appear to the right
 * of each character processed.
 *
 */

#include <errno.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>

#define TABLE_SIZE 256

struct data {
	unsigned char achar;
	unsigned char context[TABLE_SIZE-1];
	unsigned long int count;
};

struct data symbol[TABLE_SIZE-1];

void init_symbol_array();
void get_stats(char *);
void print_stats(char *);
int main(int, char *[]);

/*
 * Kinda obvious, but zero out array
 */
void
init_symbol_array(void)
{
	int i, j;

	for (i = 0; i < TABLE_SIZE; i++) {
		symbol[i].achar = i;
		symbol[i].count = 0;
		for (j = 0; j < TABLE_SIZE; j++)
			symbol[i].context[j] = 0;
	}
}


/*
 * Get frequency count of each character
 */
void
get_stats(char *file)
{
	int c, i;
	unsigned char buf[8192];
	FILE *fp;

	if((fp = fopen(file, "rb")) == NULL) {
		(void)fprintf(stderr, "\nerror: %s: Can't open '%s'\n",
		    strerror(ENOENT), file);
		exit(EXIT_FAILURE);
	}

	while ((c = fread(buf, 1, 8192, fp)) > 0) {
		for (i = 0; i < c; i++) {
			symbol[(int)buf[i]].count++;
			symbol[(int)buf[i]].context[(int)buf[i+1]] = (int)buf[i+1];
		}
	}

	fclose(fp);
}


/*
 * Print symbol stats to output file
 */
void
print_stats(char *file)
{
	FILE *fp;
	int i, j;

	if((fp = fopen(file, "w")) == NULL) {
		(void)fprintf(stderr, "\nerror: %s: Can't write to '%s'\n",
		    strerror(ENOENT), file);
		exit(EXIT_FAILURE);
	}

	fprintf(fp, "char\tfreq\tcontext\n");
	fprintf(fp, "----\t----\t-------");

	for (i = 0; i < TABLE_SIZE; i++) {
		if (symbol[i].count > 0) {
			fprintf(fp, "\n%c\t%lu\t", symbol[i].achar, symbol[i].count);
			for (j = 0; j < TABLE_SIZE; j++) {
				if (symbol[i].context[j] > 0)
					fprintf(fp, "%c", symbol[i].context[j]);
			}
		}
	}

	fclose(fp);
}


/*
 * This is where it all happens; the main function.
 */
int
main(int argc, char *argv[])
{

	if(argv[2] == NULL)
		argv[2] = "output.txt";

	init_symbol_array();
	printf("\nGathering statistics... ");
	get_stats(argv[1]);
	printf("done\n");
	print_stats(argv[2]);

	exit(EXIT_SUCCESS);
}

I see lots of buffer overruns and off-by-one errors

This seems better

Code:

/*
 * FREQ.C -- blah blah
 *
 *  given input of "this_is_a_test" freq.c will output to a file
 *
 *	char	freq	context
 *	----	----	-------
 *	_ 	3	a, i, t
 *	a	1	_
 *	e	1	s
 *	i	2	s
 *	h	1	i
 *	s	3	_, t
 *	t	3	e, h
 *
 * the "context" column represents characters that appear to the right
 * of each character processed.
 *
 */

#include <errno.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>

#define TABLE_SIZE 256

struct data {
	unsigned char achar;
	unsigned char context[TABLE_SIZE];
	unsigned long int count;
};

struct data symbol[TABLE_SIZE];

void init_symbol_array();
void get_stats(char *);
void print_stats(char *);
int main(int, char *[]);

/*
 * Kinda obvious, but zero out array
 */
void
init_symbol_array(void)
{
	int i, j;

	for (i = 0; i < TABLE_SIZE; i++) {
		symbol[i].achar = i;
		symbol[i].count = 0;
		for (j = 0; j < TABLE_SIZE; j++)
			symbol[i].context[j] = 0;
	}
}


/*
 * Get frequency count of each character
 */
void
get_stats(char *file)
{
	int c, i;
	unsigned char buf[8192];
	FILE *fp;

	if((fp = fopen(file, "rb")) == NULL) {
		(void)fprintf(stderr, "\nerror: %s: Can't open '%s'\n",
		    strerror(ENOENT), file);
		exit(EXIT_FAILURE);
	}

	while ((c = fread(buf, 1, sizeof(buf), fp)) > 0) {
		for (i = 0; i < c-1; i++) { /* nothing to the right of the last char */
			symbol[(int)buf[i]].count++;
			symbol[(int)buf[i]].context[(int)buf[i+1]] = (int)buf[i+1];
		}
	}

	fclose(fp);
}


/*
 * Print symbol stats to output file
 */
void
print_stats(char *file)
{
	FILE *fp;
	int i, j;

	if((fp = fopen(file, "w")) == NULL) {
		(void)fprintf(stderr, "\nerror: %s: Can't write to '%s'\n",
		    strerror(ENOENT), file);    /* use strerror(errno) */
		exit(EXIT_FAILURE);
	}

	fprintf(fp, "char\tfreq\tcontext\n");
	fprintf(fp, "----\t----\t-------\n");

	for (i = 0; i < TABLE_SIZE; i++) {
		if (symbol[i].count > 0) {
			fprintf(fp, "%c\t%lu\t", symbol[i].achar, symbol[i].count);
			for (j = 0; j < TABLE_SIZE; j++) {
				if (symbol[i].context[j] > 0)
					fprintf(fp, "%c", symbol[i].context[j]);
			}
            fprintf(fp,"\n");
		}
	}

	fclose(fp);
}


/*
 * This is where it all happens; the main function.
 */
int
main(int argc, char *argv[])
{

	if(argv[2] == NULL)
		argv[2] = "output.txt";

	init_symbol_array();
	printf("\nGathering statistics... ");
	get_stats(argv[1]);
	printf("done\n");
	print_stats(argv[2]);

	exit(EXIT_SUCCESS);
}

perfect!

thanks salem. i thought i had tried (c - 1) in the for loop. well, works now.

Didn't spot this first time around

int main(int, char *[]);

There is no need to prototype main
It is incorrect to even try IIRC.