tRPC•3y ago•

14 replies

Vanilla Client Error Handling

What is the right way to handle errors when using the vanilla client?

If I setup a client like so:

import type { Router } from '../../../server/src/routers'

export const trpcClient = createTRPCProxyClient<Router>({
    links: [
        httpBatchLink({
            url: `${env.SERVER_URI}/trpc`
        })
    ]
})

import type { Router } from '../../../server/src/routers'

export const trpcClient = createTRPCProxyClient<Router>({
    links: [
        httpBatchLink({
            url: `${env.SERVER_URI}/trpc`
        })
    ]
})

And execute a mutation like so:

const user = await trpcClient.user.create.mutate({ email: string; password: string })

const user = await trpcClient.user.create.mutate({ email: string; password: string })

How should I handle errors? If I try to directly catch when calling the mutate function, the error is typed as

any

any

const user = await trpcClient.user.create.mutate({ email: string; password: string }).catch((error) => {
  // error is any
}

const user = await trpcClient.user.create.mutate({ email: string; password: string }).catch((error) => {
  // error is any
}

Is there a way to type narrow the error at runtime? On the server side, I am using the

errorFormatter

errorFormatter

to add a custom field to my errors:

const t = initTRPC.context<Context>().create({
    isDev: env.NODE_ENV !== 'production',
    errorFormatter: ({ shape, error }) => {
        return {
            ...shape,
            data: {
                ...shape.data,
                customError: error.cause instanceof CustomError ? error.cause.customCode : null
            }
        }
    }
})

const t = initTRPC.context<Context>().create({
    isDev: env.NODE_ENV !== 'production',
    errorFormatter: ({ shape, error }) => {
        return {
            ...shape,
            data: {
                ...shape.data,
                customError: error.cause instanceof CustomError ? error.cause.customCode : null
            }
        }
    }
})

My current approach will be to use a runtime validation library like zod to check if this custom field exists on the error

const user = await trpcClient.user.create.mutate({ email: string; password: string }).catch((error) => {
  const parsedError = errorSchema.safeParse(error)
  if (parsedError.isSuccess) {
    console.error(parsedError.data.data.customError)
  } else {
    console.error(error)
  }
}

const user = await trpcClient.user.create.mutate({ email: string; password: string }).catch((error) => {
  const parsedError = errorSchema.safeParse(error)
  if (parsedError.isSuccess) {
    console.error(parsedError.data.data.customError)
  } else {
    console.error(error)
  }
}

But this is clearly not ideal. Is this the right/best way to approach error handling with the vanilla client, or is there a better approach?

Solution

There is actually a docs page on this, does that answer the question? https://trpc.io/docs/client/vanilla/infer-types#infer-trpcclienterror-types

Inferring Types | tRPC

It is often useful to access the types of your API within your clients. For this purpose, you are able to infer the types contained in your AppRouter.

Jump to solution

Vanilla Client Error Handling

What is the right way to handle errors when using the vanilla client?

If I setup a client like so:

import type { Router } from '../../../server/src/routers'

export const trpcClient = createTRPCProxyClient<Router>({
    links: [
        httpBatchLink({
            url: `${env.SERVER_URI}/trpc`
        })
    ]
})

import type { Router } from '../../../server/src/routers'

export const trpcClient = createTRPCProxyClient<Router>({
    links: [
        httpBatchLink({
            url: `${env.SERVER_URI}/trpc`
        })
    ]
})

And execute a mutation like so:

const user = await trpcClient.user.create.mutate({ email: string; password: string })

const user = await trpcClient.user.create.mutate({ email: string; password: string })

How should I handle errors? If I try to directly catch when calling the mutate function, the error is typed as

any

any

const user = await trpcClient.user.create.mutate({ email: string; password: string }).catch((error) => {
  // error is any
}

const user = await trpcClient.user.create.mutate({ email: string; password: string }).catch((error) => {
  // error is any
}

Is there a way to type narrow the error at runtime? On the server side, I am using the

errorFormatter

errorFormatter

to add a custom field to my errors:

const t = initTRPC.context<Context>().create({
    isDev: env.NODE_ENV !== 'production',
    errorFormatter: ({ shape, error }) => {
        return {
            ...shape,
            data: {
                ...shape.data,
                customError: error.cause instanceof CustomError ? error.cause.customCode : null
            }
        }
    }
})

const t = initTRPC.context<Context>().create({
    isDev: env.NODE_ENV !== 'production',
    errorFormatter: ({ shape, error }) => {
        return {
            ...shape,
            data: {
                ...shape.data,
                customError: error.cause instanceof CustomError ? error.cause.customCode : null
            }
        }
    }
})

My current approach will be to use a runtime validation library like zod to check if this custom field exists on the error

const user = await trpcClient.user.create.mutate({ email: string; password: string }).catch((error) => {
  const parsedError = errorSchema.safeParse(error)
  if (parsedError.isSuccess) {
    console.error(parsedError.data.data.customError)
  } else {
    console.error(error)
  }
}

const user = await trpcClient.user.create.mutate({ email: string; password: string }).catch((error) => {
  const parsedError = errorSchema.safeParse(error)
  if (parsedError.isSuccess) {
    console.error(parsedError.data.data.customError)
  } else {
    console.error(error)
  }
}

But this is clearly not ideal. Is this the right/best way to approach error handling with the vanilla client, or is there a better approach?

Solution

There is actually a docs page on this, does that answer the question? https://trpc.io/docs/client/vanilla/infer-types#infer-trpcclienterror-types

Inferring Types | tRPC

It is often useful to access the types of your API within your clients. For this purpose, you are able to infer the types contained in your AppRouter.

Jump to solution

Vanilla Client Error Handling

Similar Threads

Vanilla Client Error Handling

Similar Threads

Similar Threads

Similar Threads