Same file with differente size fprintf on linux

Admin 1
Viewed 80

I can't understand how EOF works and HOW does my pc know the size of a file.

For example I made 3 files, test0, test1 and test2. I created test0 and test1 in this way:

int main(void) {

    /* */
    FILE* fp0 = fopen("/home/kevin/eclipse-workspaceC/FileMangment/src/test0", "w+");
    FILE* fp1 = fopen("/home/kevin/eclipse-workspaceC/FileMangment/src/test1", "w+");


    if (!fp0 || !fp1) {
        fprintf(stderr, "file pointer points to NULL\n");
        return 1;
    }

    fprintf(fp0, "abcd");
    fprintf(fp1, "abcd\n");

    fclose(fp0);
    fclose(fp1);

    return 0;
}

Then I created test2 with, right click -> new documents and inside the file I wrote:

abcd

Without a space nor a new line at the end. Then on the bash I ran the command: wc -c test*; file test* and the output:

4 test0
5 test1
5 test2
14 total
test0: ASCII text, with no line terminators
test1: ASCII text
test2: ASCII text

When I go to watch the files with gedit they seems identical(no new lines founded).

I know that \n is newline and not EOF (And I really can't understand what is this this EOF), so HOW does my pc know when the file ends? If the disk block size is 4K why does it stop after the correct number of bytes and it does not continue to read whatever is comes after?

Why do I see tons of codes that parses a file in this way:

while((c=fgetc(infile))!=EOF){
putchar(c);
}

So this MEANS that there is SOMEWHERE a character called (EOF)??????? But I can't find it in ascii table. And I found that is value is -1.... Very very very confusing...

Is my pc parsing every file for know is size or it's written somewhere?

Somewhere in my computer the must be: 

your file at position xyz has the size of blahblah...

So WHERE is this place? or what should I google? Because I Googled something like 3Gozzilion of things like: where is my file size stored ecc... and did't find anything. I Just want to know where is this information stored.

eof filesystems unix linux c
2 Answers

On UNIX each file corresponds to an inode number of the file system. In that inode you can find several information about the file, including the file size. You can find more information here : Overview of the Linux file system

user_88754780

On a unix system when you issue a command like 

ls -l

the output is something similar to this

-rw-r--r--    1 user group       393 Nov  3 22:02 README.txt

this includes file permissions, owner, group size, created time, file name etc etc etc.

at a file system level this information is stored in an inode viz a structure representing the file meta-data, infact each file itself is represented by an inode number.

at a very high level every read call increments the cursor-position to read bytes. and at some point this will reach size of file. in which case read call may not even return the requested number of files. for example

observe this read call below, where we are trying to read size number of bytes from file represented by file descriptor fd.

int nbytes = read( fd, buf, size );

however it is perfectly possible that nbytes can be < size in which case most of the times next subsequent read call may fail with EOF ( if end of file is reached) . however there are other scenarios as well where nbytes may not be equal to size.

user_18536050
Related