I recently stumbled upon this interesting piece of code from one of Shopify libraries for Node.js. It makes use of an interesting way of error handling with a switch-case statement.
try {
// ...
} catch (e) {
switch (true) {
case e instanceof Shopify.Errors.InvalidOAuthError:
res.status(400);
res.send(e.message);
break;
case e instanceof Shopify.Errors.CookieNotFound:
case e instanceof Shopify.Errors.SessionNotFound:
// This is likely because the OAuth session cookie expired before the merchant approved the request
res.redirect(`/auth?shop=${req.query.shop}`);
break;
default:
res.status(500);
res.send(e.message);
break;
}
}
It's not necessarily shorter than its if-else ladder counterpart and in most cases only makes sense if you're dealing with a library built with various Error classes. Also the performance and readability aspects are up for debate. What do you think?
Top comments (2)
I use it all the time. You can also use it to test the conditions of anything you like. Often makes for more readable code
I find it difficult to apply such patterns (especially
e instaceof <class>
) because in JS it's not obvious what the errors that you have to catch and handle could be. In other language such as C#, Java, etc. thethows
signature gives you this information, but in JS you have to hope it's documented somewhere.