Hello, error handling!

In the previous tutorial we looked at printing a simple message to the terminal. In this tutorial we'll expand on that example by adding basic error handling.

The need for error handling may seem redundant. Surely printing to the terminal can't fail? Well, it can! To showcase this, we'll start with our hello.inko from the previous tutorial and change it to the following:

import std.stdio (STDOUT)

class async Main {
  fn async main {
    STDOUT.new.print('Hello, world!').get
  }
}

The change made here compared to the previous tutorial is the addition of .get at the end of the print() line. We'll look into what this does later. Now run the program as follows:

inko run hello.inko | true

The output of this program should be something similar to the following:

Stack trace (the most recent call comes last):
  [...]/hello.inko:5 in main.Main.main
  [...]/std/src/std/result.inko:119 in std.result.Result.get
  [...]/std/src/std/process.inko:15 in std.process.panic
Process 'Main' (0x5637caa83150) panicked: Result.get expects an Ok(_), but an Error(_) is found

What happened here is that the print() failed to write to the standard output stream. The get method call turns such an error into a "panic". A panic is a type of error that terminates the program. Panics are the result of bugs in your code, and as such can't be handled at runtime. In a well written program, such errors shouldn't occur.

Writing to the terminal may fail for different reasons. We're not going to cover these cases as that's out of the scope of this tutorial. Instead, we'll explore how to prevent such errors from terminating our program.

In this case we have two options:

1. We ignore the error
2. We handle it somehow

Ignoring the error

We'll start with the first option: ignoring the error. To do so, change hello.inko to the following:

import std.stdio (STDOUT)

class async Main {
  fn async main {
    let _ = STDOUT.new.print('Hello, world!')
  }
}

Now re-run the program like so:

inko run hello.inko | true

If all went well, the program should run without printing anything to the terminal.

This may seem surprising at first, but is indeed the correct behaviour: the print() call still fails, but instead of calling get we assign the result to the variable _. By assigning the result of print() to _, the compiler won't emit any warnings because the result isn't used, nor will it emit any warnings because the variable assigned to isn't used.

If we run the program using just inko run hello.inko, we get the expected "Hello, world!" output, confirming our program still works.

Handling the error

The second and (almost always) better option is to handle the error in a way other than ignoring it.

For example, imagine a hypothetical method with the signature log(message: String), which logs the message to an external logging system (e.g. syslog). We can use this method to log the error instead of ignoring it:

import std.stdio (STDOUT)

class async Main {
  fn async main {
    match STDOUT.new.print('Hello, world!') {
      case Ok(_) -> {}
      case Error(e) -> log(e.to_string)
    }
  }
}

Here we don't do anything if the call to print succeeds, as there's nothing special to be done. If the call fails instead, we convert the error to a String and log it.