Declare in:/out: explicitly on method steps

  • Declare in:/out: on a method step explicitly - do not skip them just because a method step is a regular Ruby method with access to the whole instance, and could technically read or write instance state directly instead.

  • This keeps the data flow of the whole step chain visible from the step declarations alone. One glance at the list of steps tells you what each step consumes and produces, without opening its method body, regardless of whether that step is a method step or a service step.

    class ParseCents
  include ConvenientService::Standard::Config

  attr_reader :amount

  step :NormalizeAmount,
    in: :amount,
    out: :normalized_amount

  step :AssertValidAmount,
    in: :normalized_amount

  step :result,
    in: :normalized_amount,
    out: :cents

  def initialize(amount:)
    @amount = amount
  end

  def result
    success(cents: (normalized_amount.to_f * 100).round)
  end

  private

  def NormalizeAmount
    success(normalized_amount: amount.to_s.strip)
  end

  def AssertValidAmount
    return failure("Amount `#{normalized_amount}` is not a valid positive number") unless normalized_amount.match?(/\A\d+(\.\d{1,2})?\z/)

    success
  end
end

  • A method step's in: does not even have to match its method's parameter list - see what is method step loose call?

See also

Sources