Skip to main content

FAQ

Features​

What is the Context type parameter for ?​

Context represents any metadata which a message transport layer provides to the API implementation layer.

Since a message transport layer can be anything the context is a generic type parameter which is then defined by specific message transport implementations. For example:

  • HTTP transport provides request headers and query parameters
  • AMQP transport provides correlation identifier and delivery mode
  • Homing pigeon transport provides name and color of the bird

Is it possible to bind individual functions as opposed to classes/trait methods ?​

Not at the moment.

This feature was considered since it would fit well with overall functional style of the library. However, this would introduce the following additional complexity which did not seem worth it:

  • Function parameter names would have to be extracted via additional compile-time introspection since function values in Scala do not preserve them
  • Bind method variants for all function parameter numbers would have to be created for all entry point classes

How is Automorh different from a gRPC library ?​

It has similar goals but achieves them in a different way.

  • External API definition using special interface definition language is not needed since API bindings are generated from the code
  • Additional code generation build step is not needed since API bindings are generated as part of the standard compilation step
  • Supports manipulation of transport layer metadata
  • It is effect system agnostic and thus allows combining any effect type with any transport layer
  • Implements multiple RPC protocols instead of just one, in fact implementing gRPC for Automorph is also possible

Integration​

Can I use Automorph with my existing effect system ?​

Yes.

All Scala effect systems are supported. Please see the relevant examples.

Can I use Automorph with my existing JSON parser ?​

Quite possibly.

Several Scala and Java JSON parsers are supported. Please see the relevant examples.

Additional data serialization layer support can be added if the following conditions are met:

  • The underlying data format must support arbitrarily nested structures of basic data types.
  • Given library must provide intermediate representation of structured data (i.e. document object model).

Can I use Automorph with my existing HTTP server ?​

Almost certainly.

Many Scala and Java HTTP server libraries are supported. Please see the relevant server examples and endpoint examples.

Integrating with unsupported server of any kind involves the following steps as demonstrated in the unsupported server example:

  • Create an instance of LocalEndpoint
    • Supply the desired effect system and default request context value (use () if no request context is needed).
  • Create an RPC endpoint
  • In the HTTP server request handling code
    • Pass the request body to processRequest() method of the RPC endpoint handler
    • Send the response body returned by processRequest() to the client

Can I use Automorph with my existing HTTP client ?​

Most likely.

Many Scala and Java HTTP client libraries are supported. Please see the relevant examples.

Web-RPC​

When to use Web-RPC ?​

In case any of the following remote API concerns need to be addressed with minimal effort:

  • Caching GET requests
  • Using URLs to pass arguments
  • External constraints requiring a simple REST-style API with RPC semantics

In other situations it is probably better to use an established remote call protocol such as:

Can Web-RPC be used in without a specific Web-RPC library ?​

Yes. Any REST client or server library will suffice. However, using a specific Web-RPC library minimizes the implementation effort.

Why call it a REST-style protocol when it is conceptually unrelated to REST ?​

To illustrate that it provides remote API authors with a solution with capabilities equivalent to and compatible with typical REST-style protocols but does so by openly embracing RPC principles.