retil-issues

retil-issues

Superpowers for validation and error handling

Getting Started

yarn add retil-issues

import { useIssues, useValidator } from 'retil-issues'

2-minute primer

The useIssues() hook manages a list of issues related to a data object.

const [issues, addIssues, clearIssues] = useIssues({ email, password })

With this line, you now have a access to a list of your form's issues:

// Log any issues with the email field to the console
console.log(issues)

But where do these issues come from? You may have guessed that it has something to do with that addIssues variable. But in fact, there are two different ways that you add issues: validation, and server responses.

useValidator()

The first, and simplest, is to use the useValidator hook. This hook accepts an addIssues function (as returned by useIssues) as its first argument, and a validator function as its second argument. It returns a tuple with a function to trigger validation, and a a function to remove any issues.

The validator function is simple - it takes the data to be validated, and returns an object mapping paths to arrays of issues.

const [validate, clear] = useValidator(addIssues, ({ email, password }) => ({
  email: [
    !email && "Please enter your email",
    !isValidEmail(email) && "That doesn't look like a valid email",
  ],
  password: [
    !password && "Please enter your password",
  ]
}))

The returned validate function can be called to run the validator and add any resulting issues to your issues object. This function returns a promise to a boolean, which will resolve to true if the data looks good and issue-free.

This is great for form handlers, where you'll want validate your data before hitting the server.

const handleSubmit = async () => {
  const isValid = await validate()
  if (isValid) {
    // ...
  }
}

Of course, you don't always want to wait until your form is submitted to help your user out with a validation hint. That's why the validate function optionally accepts a path which you'd like to validate.

This makes it perfect for use in the onBlur handler or an input -- it allows you to validate a field as soon as the user taps out of it.

const input = (
  <input
    // Validate just the `email` field whenever focus moves somewhere else
    onBlur={() => validate('email')}
    onChange={event => setEmail(event.target.value)}
    value={email}
  >
)

addIssues()

Not all issues with user input can be detected client side. For example, say you're building a sign-up form with an email field. While you can detect missing emails on the client, what you can't detect is duplicate emails.

If your server returns a message saying "that email has already been used", then you'll want to display it the user -- but you'll also want to clean it up once the user changes the email. This is easy to achieve by calling addIssues() directly:

const handleSubmit = async () => {
  // Note that `validate` only validates the form -- it doesn't check for
  // issues added from a previous submit.
  if (await validate()) {
    clearIssues()

    const serverIssues = await submitToServer(data)

    // Add the issues returned from the server. They'll be automatically removed
    // as the relevant input data changes.
    addIssues(serverIssues)
  }
}