Home

Awesome

Result Monad

ci Maven Central Kotlin License

Usage

Kotlin DSL:

repositories {
    mavenCentral()
}

dependencies {
    implementation("org.erwinkok.result:result-monad:$latest")
}

Introduction

Suppose you have the following method signature:

fun createItem(name: string): String {
    ...
} 

This tells you everything to know, right? No, not really. It doesn't say anything about the error conditions it can return or throw. Of course, you can add it to the documentation, but the source code itself should be the documentation (i.e. if writing documentation as a comment, it can deviate over time from the real code). So the code itself should be as readable and understandable as possible.

In Java (and other languages) you can specify that the method can throw an exception (by adding, for example, a throws IllegalArgumentException).

However, when the caller deals with the exception:

try {
    doSomething()
    doSomethingElse()
    finalizeTheThing()
    doSomeCleanup()
} catch (e: IllegalArgumentException) {
    ...
}

You don't know which method threw the exception. This is especially cumbersome if, for example, doSometing() opens a socket and doSomethingElse() throws an exception. Normally, doSomeCleanup() would close the socket, but if an exception was thrown, it would never get there. Of course, you can handle this with a flag socketOpened or something and in the catch-clause you can check this flag and close the socket there. But this looks, in my opinion, rather ugly and is error-prone. That is where the result-monad comes in.

Result Monad

In its generic form, a result monad has two result values: one value for the success case, and an error for the failure case:

fun doSomething(): Result<...ok value ..., ...error...  > {
    ...
}

The actual implementation of this method can return either a 'normal' result value, by using Ok(...), or an error value, by using Err(...). For example:

fun divideNumbers(a: Int, b: Int): Result<Float, Error> {
    if (b == 0) {
        return Err("Division by zero is not allowed")
    }
    return Ok(a / b)
}

Callers can check the returning Result value whether an error condition is raised. If not, it can get the ok-value. Otherwise, it can handle the error case:

val result = divideNumber(n1, n2)
    .getOrElse {
        // Handle division by zero
        log.error { "Division by zero: ${it.message}" }
        return Err(it) // Return the error to its caller.
    }
log.info { "n1 divided by n2 is $result" }

Chaining calls

Because Result has two sub-types, one for the value and one for the error, we can act upon these two different cases. For example by doing something when Ok was returned, and something else when Err was returned. map, for example, applies the given lambda when an Ok value was returned, or just passes the error along when Err was returned.

val result = divideNumber(n1, n2)
    .map { it * 2 }

The example above means that when divideNumber returns Ok the value is multiplied by 2. If an Err was returned, the multiplication is not performed, but the error original error is returned.

Of course, this can be chained with multiple calls.

Inspiration, based on

The idea of a result monad is not new, some other programming languages have something similar already present, like: Go Lang and Rust.

There are already some libraries available in Kotlin, in particular the excellent kotlin-result project created by Michael Bull: kotlin-result where this project is based on. The main difference between his project and this one is that the error case always has an Error as object:

kotlin-result:

fun doSomething(): Result<String, Error> {

}

This project:

fun doSomething(): Result<String> {

}

This saves typing.

The Error object is a small wrapper around a string and is in itself an Exception. We can instantiate an Error object by either a string or an exception. Because of this, the Err() can be created more compact:

kotlin-result:

return Err(IllegalArgumentException("The argument is wrong!"))

This project:

return Err("The argument is wrong!")

Remember, an Error is a small wrapper around a string, and the error case of the Result is always Error. This means, the Exception type is lost. However, this can be solved by using a fixed Error:

val illegalArgumentError = Error("The argument provided is illegal")
...
return Err(illegalArgumentError)

...and then check the result for an error, and if an error occurs, check for the specific error:

    val result = doSomething()
    .getOrElse {
        if (it === illegalArgumentError) {
            // handle illegal argument
        } else {
            // handle other errors
        }
    }

Contributing

Bug reports and pull requests are welcome on GitHub.

Contact

If you want to contact me, please write an e-mail to: erwin-kok@gmx.com

License

This project is licensed under the BSD-3-Clause license, see LICENSE file for more details.

Since this project is partly based on the awesome kotlin-result project, created by Michael Bull, I add that license here as well:

Copyright (c) 2017-2022 Michael Bull (https://www.michael-bull.com)

Permission to use, copy, modify, and/or distribute this software for any
purpose with or without fee is hereby granted, provided that the above
copyright notice and this permission notice appear in all copies.

THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.

kotlin-result is available at: https://github.com/michaelbull/kotlin-result