Why the use of iostream::eof inside a loop condition considered wrong?

Just because we haven't reached the EOF, doesn't mean the next read will succeed.

Consider you have a file that you read using file streams in C++. When writing a loop to read the file, if you are checking for stream.eof(), you're basically checking if the file has already reached eof. 

So you'd write the code like −

Example

#include<iostream>
#include<fstream>
using namespace std;

int main() {
   ifstream myFile("myfile.txt");
   string x;
   
   while(!myFile.eof()) {
      myFile >> x;
      // Need to check again if x is valid or eof
      if(x) {
         // Do something with x
      }
   }
}

Example

While when you use the stream directly in a loop, you'd not be checking the condition twice −

#include<iostream>
#include<fstream>
using namespace std;

int main() {
   ifstream myFile("myfile.txt");
   string x;
   while(myFile >> x) {
      // Do something with x
      // No checks needed!
   }
}