I'm making a couple async API callouts and may throw a custom error depending on the outcome. I'm deleting Objects from S3.
try {
await s3.deleteObject(bucketParams);
//S3 API doesn't provide resp on if obj successfully deleted. Therefore, check that it doesn't exist afterwards to verify instead
const err = await s3.headObject(bucketParams);
if (err & err.code === 'NotFound') return { code:200, message:`${key} successfully deleted` }
throw `Error deleting ${key}`;
} catch (error) {
throw new Error(500, error);
}
So the main catch can handle any exceptions thrown by those couple API callouts (error with network, bad key, bad authorization, etc..)
However, if the object didn't actually get deleted (so if it still exists)..I throw a custom error to get reported back to end user.
My question is if there's a better pattern to just throwing a custom error in your try and then having it get slurped up by your catch.
Thanks!
You can do a lot to be honest, you just need to work out what you want to capture. Heres a basic example. If you were using TypeScript you would get more power and type safety
class S3ObjectNotFound extends Error {
// Add other attributes you might like
code
constructor(message) {
super(message)
this.name = 's3/object-id-not-found'
this.code = 404
}
}
const someFunction = async (bucketParmas, key) => {
try {
try {
await s3.deleteObject(bucketParams)
} catch (s3Error) {
throw new S3ObjectNotFound(`Object with id ${bucketParmas} was not found`)
}
const err = await s3.headObject(bucketParams)
if (err && err.code === 'NotFound') {
return { code: 200, message: `${key} successfully deleted` }
}
throw AnotherErrorYouCouldMake()
} catch (error) {
throw new Error(500, error)
}
}
someFunction({...YourParams}, 'YourKey')