Why do i have to input EOF 3 times when using fget

2020-07-27 02:06发布

问题:

So basically I want to copy everything i write to stdin (including newline char) to string for hash purposes. I managed to accomplish that and made small code to represent my problem.

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

#define BUFFERSIZE 10000

int main()
{
char *myStr = calloc(1,1);
char buffer[BUFFERSIZE];

while( fgets(buffer, BUFFERSIZE , stdin) != NULL ){
  myStr = realloc(myStr, strlen(myStr)+1+strlen(buffer) );
  strcat( myStr, buffer );
}
printf("\n%s\n",myStr);

}

everything works when I enter some text then press ENTER and after I call EOF.

But when I start program enter "a" then I try to call EOF (using Ctrl Z + (Windows cmd prompt), Ctrl D (Linux)) I have to do it three times for program to actually break the loop. I was expecting maximum of 2 times.

Can someone explain how using EOF, stdin and fgets works? Or should I use something else (for example getline)? I am sorry if I am not clear about my problem, just ask anything you need.

Thank you.

回答1:

First of all, ^Z or ^D are control characters that mean something to the terminal you are using, and sometimes that means for the terminal to signal end-of-file condition.

Anyway, your three keypresses are processed by the terminal to take the following actions, after entering text:

  1. Flush the input (i.e. send the characters that have been input so far from the terminal to the program - by default this doesn't happen as the terminal uses line buffering)
  2. Set end-of-file condition
  3. Set end-of-file condition again

Inside your program that corresponds to:

  1. Nothing happens: even though a is received, fgets keeps reading until end-of-file or newline
  2. fgets completes because of end-of file. However it does not return NULL because characters were read, "a" to be specific.
  3. fgets completes because of end-of-file, and returns NULL because there were no characters read.