Ascii value of EOF

[utkarshjadhav]

its well known that EOF has value -1.
but does it have any ASCII value???

[utkarshjadhav]

And other question (very basic and funny one)--
when we finish writing a file...How and when EOF gets associated with it..??
if EOF is only -1 (and nothing else)
What will happen while reading another -1 (which is within the file)?

[weaknessforcats]

Who said EOF is -1?

If EOF is -1 then the code should use -1 and avoid all the mystery. But EOF is not guaranteed to be -1. It may be anything so you use EOF in your code.

The -1 is an assumption. Like the guy who thinks FALSE is 0 (except in Visual Basic where it's -1). Avoid assumptions.

[utkarshjadhav]

@ weaknessforcats:
But for the following program in DevC++,in the file demo.c--

#include<stdio.h>
int main()
{
        int c=EOF;
        printf("%d",c);
        getch();
        return(0); 
 
}
 

It showed -1...

[weaknessforcats]

Maybe I wasn't clear.

EOF has to be a number larger than the largest number of bytes that you can put in a file. Otherwise, it can't be EOF.

In your compiler's case -1 was picked because no file can have -1 bytes in it. Therefore, the value of EOF can't be confused with any valid position in the file.

But -2 would also work. As would -222 or -123434566. The fact that your compiler uses -1 is only interesting. But your code better use EOF or it may not work when compiled on someone else's compiler.

[johny10151981]

in visual studio EOF is defined like below:

#define EOF (-1)

Same in Linux (checked on CentOS)

[Oralloy]

There is a "ASCII" EOF indicator.

It's called FS - file separator, value 0x1C.

There is also EOT - end of transmission, value 0x04, entered as ctrl-D under the Unix/Linux command line.

And, of course, PC-DOS/Windoze uses SUB - substitute, value 0x1a, entered as ctrl-Z, as it's command line end of file character.

So, pick one.....

The problem is that all of these are representable as ASCII codes which may quite reasonably appear in a binary data stream.

That's why the special token, EOF, is used in source code. The value of -1 is just handy and obvious.

Also, since ASCII is a seven bit code, it's still a useful value when used in 8-bit byte encoding. Probably not something that concerns you, though.

Did you observe that functions like getc() return integers? This keeps the -1 EOF value from conflicting with the 8 bit value 0xff, which may be valid in the data stream.

Does that help?