Error after reading all words from a .txt

This program reads words from a dictionary text file and writes all of the words that end in 'man' to another file. I know this is probably a very strange way to do this but it's what I have after trying it a few ways.

It works, and writes all the words to the file, but it crashes after reading through all of the words in the dictionary file. I don't know why (ms vc++ 6.0).

Code:

if ((dictfile = fopen("/aacraig/dict.txt", "r"))==NULL) {
		printf("Cannot open this file\n");
		exit(1);
	}

	

	while(!feof(dictfile)) { 
	
		next_char = 0;
		count=0;
		for (i=0; next_char!=10; i++) {        /* read one line in */
			fscanf(dictfile, "%c", &next_char);
			array[i] = next_char;
			count++;
		}


		//printf("count is %d\n", count);
		wordfile = fopen("/aacraig/wordfile.txt", "a");

		if ((array[count - 2] == 'n') && (array[count - 3] == 'a') 
									  && (array[count - 4] == 'm'))
			for(i=0; i<count; i++)
				fprintf(wordfile, "%c", array[i]);
		

		printf("\n");
		fclose(wordfile);

	}

	fclose(dictfile);

>>while(!feof(dictfile))
This is not the way to control a read loop. You should use the return code from the read function to determine when you've reached EOF.

Code:

for (i=0; next_char!=10; i++) {        /* read one line in */
  fscanf(dictfile, "%c", &next_char);
  array[i] = next_char;
  count++;
}

This looks like a clear case for a buffer overflow. What happens when you reach EOF? (See my first point)

>>array[count - 2]
And if count == 0, where are you indexing too?!

>>/* read one line in */
If you want a line reading function, use fgets().

I got it to work like this, but it's really yucky and unsafe and all kinds of nasty stuff that you can imagine :-)

Code:

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

int main(void)
{
  int i;
  int count;
  char next_char;
  char array[10000];
  FILE *dictfile, *wordfile;

  if ((dictfile = fopen("input.txt", "r"))==NULL) {
    printf("Cannot open this file\n");
    exit(1);
  }
  
  while(!feof(dictfile)) {
    next_char = 0;
    count=0;

    for (i=0; next_char!='\n' && next_char != EOF; i++) {        /* read one line in */
      if (fscanf(dictfile, "%c", &next_char) != 1)
      {
        break;
      }
      array[i] = next_char;
      count++;
    }
    
    //printf("count is %d\n", count);
    wordfile = stdout;
    
    if ((array[count - 2] == 'n') && (array[count - 3] == 'a') 
      && (array[count - 4] == 'm'))
    {
      for(i=0; i<count; i++)
      {
        fprintf(wordfile, "%c", array[i]);
      }
    }
    
    printf("\n");
  }
  
  fclose(dictfile);

  return 0;
}

This way would be way better :-)

Code:

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

int main(void)
{
  char array[10000];
  FILE *dictfile, *wordfile = stdout;

  if ((dictfile = fopen("input.txt", "r")) == 0)
  {
    return 1;
  }
  
  while(fgets(array, sizeof array, dictfile) != 0)
  {
    if (strstr(array, "man") != 0)
    {
      fprintf(wordfile, "%s", array);
    }
  }
  
  fclose(dictfile);

  return 0;
}

>>for (i=0; next_char!='\n' && next_char != EOF; i++)
>> if (fscanf(dictfile, "%c", &next_char) != 1)
next_char won't get to be EOF as fscanf() won't populate it so.

>>next_char won't get to be EOF as fscanf() won't populate it so.
I plead insanity for doing that in the first place and then I plead stupidity for leaving it there when I posted :-)

Your 2nd program finds all words with man in it, like manual, etc. I guess that's what my problem with fgets() is. I can read all day what it does ( http://www.rt.com/man/fgets.3.html ) but I dont really know what it does

Can you somehow break what fgets reads into an array so I can work with the last characters of that line? (obviously possible, I just dont know how).

Thanks Hammer, I realized your bottom 2 points, and will have to read about the top 2.

>>Your 2nd program finds all words with man in it, like manual, etc.
Sorry, I thought you wanted any word with man in it, try this :-)

Code:

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

int main(void)
{
  char array[10000];
  FILE *dictfile, *wordfile = stdout;

  if ((dictfile = fopen("input.txt", "r")) == 0)
  {
    return 1;
  }
 
  while(fgets(array, sizeof array, dictfile) != 0)
  {
    char *p = strchr(array, '\n');

    if (p != 0)
    {
      *p = '\0';
    }

    if (strcmp(array + (strlen(array)-3), "man") == 0)
    {
      fprintf(wordfile, "%s\n", array);
    }
  }
 
  fclose(dictfile);

  return 0;
}

>>work with the last characters of that line?
They're the ones just before the terminating \0 character

Here's a quick hack, showing how you might be able to test the last 3 chars of a string. Remember, if you use fgets(), get rid of, or at least account for the possible presence of a newline character in the buffer.

Code:

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

int main (void)
{
  char a[][20] = {"I am a man", "No youre not", "OK"};
  int i, len;
  
  for (i = 0; i < 3; i++)
  {
    len = strlen (a[i]);
    if (len > 3)
    {
      len -= 3;
      if (strcmp(&a[i][len], "man") == 0)
        printf ("found it\n");
      else
        printf ("didn't find it\n");
    }
    else printf ("Line too short\n");
  }
    
  
  return 0;
}


/*
 Program output
  found it
  didn't find it
  Line too short
 */

[edit]Doh, beat by miles

[edit2]
>>(strlen(array)-3),
What if strlen() returns 0?

>>What if strlen() returns 0?
You test it before using it of course. I'll just say that's an exercise for the reader and let you wonder if I meant to do it for simplicity or I forgot :-)

ps: If it was bugging you, I forgot :-)

Thanks , will play with these. I'll have more newb questions soon. This stuff is finally starting to be fun.