Simple EOF question

Hi,
I am currently studying K&R and I was looking at the example as follows:

Code:

---

#include <stdio.h>

main ()
{
    int c;

    c = getchar();
    while (c != EOF)
    {
        putchar(c);
        c = getchar();
    }
}

---

I am a little confused about the EOF. I checked and I am pretty sure that my EOF is -1 but when I input -1 in this program, the program doesn't end. -1 is treated just like the other characters. Am I missing something?

http://faq.cprogramming.com/cgi-bin/...&id=1043284351

Likely your program expects redirected input:

Code:

---

c:\myprog.exe < file.txt

---

If you want to read until the end of input, when you are reading from stdin (i.e. if the user is typing). You want to look for end of line, or from a string (char-array) a NULL-value (0).

If you type "-1", it will read the character - and 1, which have ascii-values which are not -1 either together or in their parts.

Instead you could look for 10 or '\n' (that is if you expect typed input):

Code:

---

    while (c != '\n')
    {
        putchar(c);
        c = getchar();
    }

---

Thanks for the replies. I am still a little confused though. I thought because EOF is defined in stdio.h as -1, that when I type in -1 the program should end.

Your suggestions work of course and I think understand you explanation about ascii values but I wonder why nearly every example in Kernighan and Ritchie uses EOF if it doesn't work.

Quote:

---

if it doesn't work

---

it works
EOF is not what was read from input stream - it is artifitially defined value that cannot be in any case read from input stream, and thus was selected to notify the caller that there is nothing left to read...

to imitate the EOF condition from the keyboard you can try to use Ctrl-Z combination on Windows or Ctrl-D on *nix system...

Thanks for the reply. I'm not really sure about the reasoning but I guess I'll try to just accept it as a fact.
Why is EOF defined as -1 in stdio.h though?

Ctrl-C worked for me. Thanks for the suggestion.

Quote:

---

Why is EOF defined as -1 in stdio.h though?

---

possible return values of fgetc are 0 - 255 (in case of success)
closes impossible values are -1 and 256
because errors in general are notified with negative values - I think -1 was choosen to indicate the read error

Thanks again.

> Ctrl-C worked for me. Thanks for the suggestion.
No, that just kills the program. It is not an orderly exit.

> Why is EOF defined as -1 in stdio.h though?
All chars are small positive integers, so it kinda made sense (I guess) to pick a value which could be never equal to a char no matter how large a char value got.

ctrl-d or ctrl-z will work, but exactly how they work depend on your terminal driver and the I/O code in the C runtime library provided by your compiler.
Normally you press the control key sequence at the start of a line, and sometimes you have to press return afterwards (if the input is buffered).

If you try to press the control sequence at any other time, then some systems need you to press the same sequence again (eg. ctrl-z ctrl-z).

Expermiment with this in a simple program to see how your specific implementation behaves.

Code:

---

if ( ch == EOF ) {
  printf( "yes, that was EOF\n" );
}

---

Thanks for the reply and the clarification. I'm not quite sure what to do with that program yu gave me. Should I use it combination with the previous program or as a standalone. I tried it as a standalone but was told i had to declare c so I created the following program but nothing happens.

Code:

---

#include <stdio.h>

main()
{
    int ch;

    if ( ch == EOF )
    {
        printf( "yes, that was EOF\n" );
    }

}

---

Put it in a loop which prints say the decimal, hex and character value of every character received.

Quote:

---

All chars are small positive integers, so it kinda made sense (I guess) to pick a value which could be never equal to a char no matter how large a char value got.

---

Interestingly, C99 defines EOF as a macro "which expands to an integer constant expression, with type int and a negative value, that is returned by several functions to indicate end-of-file". There is no footnote that gives the exact value required. So, EOF is typically -1 since it makes sense, but it could actually be any negative int constant?

It could be any integer constant not equal to any valid character I suppose. My guess is that it started off as a literal -1, then came the EOF symbol which typically was assigned the value -1 just to avoid any surprises with any old code using a magic number.