Error

Errors modeled as data.

Installation

def deps do
  [
    {:error, "~> 0.1.0"}
  ]
end

Creation

Errors come in two kinds:

:domain errors, which are part of your business model, or Domain in DDD parlance. You can use the map in the second argument position to supply whatever extra details you find useful:

    Error.domain(:invalid_username, %{
     provided_username: "alex",
     unmet_requirement: "Must start with capital letter"}
    })

:infra errors, which represent failures of the infrastructure or computation substrate. Similarly, you can supply a map as the second argument, with arbitrary details of your choosing.

    Error.infra(:db_down, %{retried_count: 5})

Usage

Errors should be created at the place where they occur. Domain errors can be later converted to user-facing messages, in the presentation logic.

Infra errors, on the other hand, most often translate to a "500 error" and the only piece of information that an end-user needs to know is that "Something went wrong". At the same time, you want to log these errors extensively so that you can investigate the issue later.

Errors provide a convenient to_map function, so that your view logic doesn't need to know about the Error type.

Pattern-matching

If you'd like to pattern-match on errors, it's possible to match directly on the Error.InfraError and Error.DomainError structs, however it's strongly discouraged.
The best way of testing for domain or infra errors are dedicated guards. Import the guard is_error, or the specific guards is_infra_error and is_domain_error and use them in the match clauses:


    import Error, only: [is_error: 1]
    case something() do
        {:ok, %MyItem{} = i} -> handle(i)
        {:error, e} when is_error(e) -> Error.reason(e)
        {:error, :some_atom} -> other_path()
    end

Source