is it bad practise to use "void function()" in c in terms of error handling-CodePudding

Function return types are useful for returning values that indicate if the function did what it was supposed to do.

However if error handling is done within the function (for example fprintf(stderr, errormessage) and exit(EXIT_FAILURE)) that means one could define a function with the void return type instead.

So my question is: is it bad practice in C to have a program with multiple functions with void return types while handling errors in the way stated above in contrast to returning a number?

I ask because I am writing a program and noticed my header files look like this and each function uses printing to stderr and exit():

void function1();
void function2();
void function3();
void function4();
void function5();
void function6();
void function7();
void function8();
void function9();

CodePudding user response：

In general, returning void isn't bad if your errors are handled internally. The point is for all errors to be handled somewhere. Error codes are for the things that the caller needs to handle.

Speaking to your particular case, terminating the entire program when an error occurs isn't really error handling. It's just an extremely fragile program. You're not giving the rest of your program a chance to clean up after itself, so you can end up with leaked resources, things left in unsafe states, etc. The only time an error should terminate your program is if there's no possible way to handle it gracefully (and there almost always is). The error "handling" technique you describe might be okay for very short toy programs but for anything beyond that, you need to write real error handling code.

CodePudding user response：

That depends on how many exceptions you will need to handle. But I would highly recommend you to use the errno variable which will always store the number of the last error. It will return -1 if something went wrong, and 0 if not. You can easily use it after invoking some function and/or system calls, it will automatically set the value according to the output/value of the last function that was invoked. You can also check which type/the name of the error that was thrown, for this, the errno.h header offers a huge collection of errors names that you can combine with your project requirements. For example, after trying to read a file, you can check if some error occurred and which type of error. Check the code bellow.

#include<errno.h> //you need it.
#include<stdio.h>


int main(){
    FILE *fd;
    if (fopen("/myfile.itsextension", someflaghere){
         if (errno ==  ENOENT){
            // ENOENT stands for: No such file or directory
           //do something here if it is the error
         }
         //if it is something else
         errno("Error:"); it output in the console Error: some message according to the error that occurred.
    }
}

You can check more about errno at the manpage website.

If you really need to create your own errors handler, just make sure you create it as a header.

I hope it helps you.