Try but don't catch: Elegant error handling with TypeScript
Problem
Tired of throw? No longer checking each return value via If to see whether the value is really present? Tired of boilerplate code with try/catch? You are looking for an implementation of the Result/Optional/Maybe monad? Or have you heard about railway oriented programming and want to try it out?
Solution
If you are using TypeScript, there is a nice tool for you – neverthrow! As the name suggests, it lets you write code that no longer throws or catches errors. Instead, there is a Result type that can take the states Ok or Err. The best thing about it, in the sense of railway oriented programming, is that you can "chain" different Results together and switch between the states, i.e. switch to the Err state or even "recover" from the Err state back to the Ok state. As if that weren't enough, the library also provides interoperability with Promise through the ResultAsync type.
Example
function hasKeys<T extends object>(item: T, keys: Array<keyof T>): Result<T, string> {
const requiredKeysArePresent = keys.every((key) => key in item);
return requiredKeysArePresent ? ok(item) : err('Object to validate has missing keys');
}// Using the synchronous Result API
async function storeCoffee(candidate?: Coffee): Promise<void> {
const validationResult: Result<Coffee, string> = exists(candidate)
// Only if Ok is returned, the code is continued in "andThen"
.andThen((existingCandidate) => hasKeys(existingCandidate, ['name', 'origin']));
if (validationResult.isOk()) {
await saveCoffeeInDb(validationResult.value);
}
}// Converting a Promise to ResultAsync
function get<T>(url: string): ResultAsync<T, ApiError> {
return fromPromise(
httpClient.get<T>(url, { responseType: 'json' }).then((result) => result.body),
(e) => mapToApiError(e)); // Replacement for the "catch" method of Promise
}function getCoffeeList(): Promise<Coffee[]> {
return get<Coffee[]>('api.jambit.com/coffee')
.mapErr((e: ApiError) => {
logger.error(`Could not get coffee list because of: ${e}`);
return [coffeeA, coffeeB]; // Recover with fallback value
}).match((coffeeList: Coffee[]) => {
logger.info(`There are ${coffeeList.length} types of coffee available.`);
return coffeeList;
}, (fallbackList) => fallbackList));
}
Further Aspects
---
Author: Robert Gruner / Software Engineer / Office Leipzig
Download Toilet Paper #150: Try but don't catch: Elegant error handling with TypeScript (PDF)
Want to write the next ToiletPaper? Apply at jambit!
- Previous page
- To overview
- Next page
