RSpec matchers

Adding matchers to project

Add this line to your application's Gemfile:

group :test do
  gem 'ruby_event_store-rspec'
end

Event matchers

be_event

The be_event matcher enables you to make expectations on a domain event. It exposes fluent interface.

OrderPlaced  = Class.new(RubyEventStore::Event)
domain_event = OrderPlaced.new(
  data: {
    order_id: 42,
    net_value: BigDecimal.new("1999.0")
  },
  metadata: {
    remote_ip: '1.2.3.4'
  }
)

expect(domain_event)
  .to(be_an_event(OrderPlaced)
    .with_data(order_id: 42, net_value: BigDecimal.new("1999.0"))
    .with_metadata(remote_ip: '1.2.3.4'))

By default the behaviour of with_data and with_metadata is not strict, that is the expectation is met when all specified values for keys match. Additional data or metadata that is not specified to be expected does not change the outcome.

domain_event = OrderPlaced.new(
  data: {
    order_id: 42,
    net_value: BigDecimal.new("1999.0")
  }
)

# this would pass even though data contains also net_value
expect(domain_event).to be_an_event(OrderPlaced).with_data(order_id: 42)

This matcher is both composable and accepting built-in matchers as a part of an expectation.

expect(domain_event).to be_an_event(OrderPlaced).with_data(order_id: kind_of(Integer))
expect([domain_event]).to include(an_event(OrderPlaced))

If you depend on matching the exact data or metadata, there's a strict modifier.

domain_event = OrderPlaced.new(
  data: {
    order_id: 42,
    net_value: BigDecimal.new("1999.0")
  }
)

# this would fail as data contains unexpected net_value
expect(domain_event).to be_an_event(OrderPlaced).with_data(order_id: 42).strict

Mind that strict makes both with_data and with_metadata behave in a stricter way. If you need to mix both, i.e. strict data but non-strict metadata then consider composing matchers.

expect(domain_event)
  .to(be_event(OrderPlaced).with_data(order_id: 42, net_value: BigDecimal.new("1999.0")).strict
    .and(an_event(OrderPlaced).with_metadata(timestamp: kind_of(Time)))

You may have noticed the same matcher being referenced as be_event, be_an_event and an_event. There's also just event. Use whichever reads better grammatically.

Event store matchers

have_published

Use this matcher to target event_store and reading from streams specifically. In a simplest form it would read all streams forward and check whether the expectation holds true. Its behaviour can be best compared to the include matcher — it is satisfied by at least one element present in the collection. You're encouraged to compose it with be_event.

event_store = RubyEventStore::Client.new(repository: RubyEventStore::InMemoryRepository.new)
event_store.publish(OrderPlaced.new(data: { order_id: 42 }))

expect(event_store).to have_published(an_event(OrderPlaced))

Expectation can be narrowed to the specific stream.

event_store = RubyEventStore::Client.new(repository: RubyEventStore::InMemoryRepository.new)
event_store.publish(OrderPlaced.new(data: { order_id: 42 }), stream_name: "Order$42")

expect(event_store).to have_published(an_event(OrderPlaced)).in_stream("Order$42")

It is sometimes important to ensure that specific amount of events of given type have been published. Luckily there's a modifier to cover that usecase.

expect(event_store).not_to have_published(an_event(OrderPlaced)).once
expect(event_store).to have_published(an_event(OrderPlaced)).exactly(2).times

You can make expectation on several events at once.

expect(event_store).to have_published(
  an_event(OrderPlaced),
  an_event(OrderExpired).with_data(expired_at: be_between(Date.yesterday, Date.tomorrow))
)

You can also make expectation to ensure that expected list of events is exact actual list of events (and in the same order) using strict modifier.

expect(event_store).to have_published(
  an_event(OrderPlaced),
  an_event(OrderExpired).with_data(expired_at: be_between(Date.yesterday, Date.tomorrow))
).strict

Last but not least, you can specify reading starting point for matcher.

expect(event_store).to have_published(
  an_event(OrderExpired)
).from(order_placed.event_id)

If there's a usecase not covered by examples above or you need a different set of events to make expectations on you can always resort to a more verbose approach and skip have_published.

expect(event_store.read.stream("OrderAuditLog$42").limit(2)).to eq([
  an_event(OrderPlaced),
  an_event(OrderExpired)
])

publish

This matcher is similar to have_published one, but targets only events published in given execution block.

event_store = RubyEventStore::Client.new(repository: RubyEventStore::InMemoryRepository.new)
expect {
  event_store.publish(OrderPlaced.new(data: { order_id: 42 }))
}.to publish(an_event(OrderPlaced)).in(event_store)

Expectation can be narrowed to the specific stream.

event_store = RubyEventStore::Client.new(repository: RubyEventStore::InMemoryRepository.new)
expect {
  event_store.publish(OrderPlaced.new(data: { order_id: 42 }), stream_name: "Order$42")
}.to publish(an_event(OrderPlaced)).in(event_store).in_stream("Order$42")

You can make expectation on several events at once.

expect {
  # ...tested code here
}.to publish(
  an_event(OrderPlaced),
  an_event(OrderExpired).with_data(expired_at: be_between(Date.yesterday, Date.tomorrow))
).in(event_store)

have_subscribed_to_events

Use this matcher to make sure that a handler has or has not subscribed to event types in target event_store.

Ensuring handler is subscribed to given event types:

event_store = RubyEventStore::Client.new(repository: RubyEventStore::InMemoryRepository.new)

expect(Handler).to have_subscribed_to_events(FooEvent, BarEvent).in(event_store)

Checking if handler does not subscribe to any of given event types:

expect(Handler).not_to have_subscribed_to_events(FooEvent, BarEvent).in(event_store)

Aggregate root matchers

The matchers described below are intended to be used on aggregate root gem.

To explain the usage of matchers sample aggregate class is defined:

OrderSubmitted = Class.new(RubyEventStore::Event)
OrderExpired   = Class.new(RubyEventStore::Event)


class Order
  include AggregateRoot
  HasBeenAlreadySubmitted = Class.new(StandardError)
  HasExpired              = Class.new(StandardError)

  def initialize
    self.state = :new
    # any other code here
  end

  def submit
    raise HasBeenAlreadySubmitted if state == :submitted
    raise HasExpired if state == :expired
    apply OrderSubmitted.new(data: {delivery_date: Time.now + 24.hours})
  end

  def expire
    apply OrderExpired.new
  end

  private
  attr_accessor :state

  on OrderSubmitted do |event|
    self.state = :submitted
  end

  on OrderExpired do |event|
    self.state = :expired
  end
end

The matchers behaviour is almost identical to have_published and publish counterparts, except the concept of stream. Expecations are made against internal unpublished events collection.

have_applied

This matcher check if an expected event has been applied in aggregate_root object.

aggregate_root = Order.new
aggregate_root.submit

expect(aggregate_root).to have_applied(event(OrderSubmitted))

You could define expectations how many events have been applied by using:

expect(aggregate_root).to have_applied(event(OrderSubmitted)).once
expect(aggregate_root).to have_applied(event(OrderSubmitted)).exactly(3).times

With strict option it checks if only expected events have been applied.

expect(aggregate_root).to have_applied(event(OrderSubmitted)).strict

apply

This matcher is similar to have_applied. It check if expected event is applied in given aggregate_root object but only during execution of code block.

aggregate_root = Order.new
aggregate_root.submit

expect {
  aggregate_root.expire
}.to apply(event(OrderExpired)).in(aggregate_root)

With strict option it checks if only expected events have been applied in given execution block.

aggregate_root = Order.new
aggregate_root.submit

expect {
  aggregate_root.expire
}.to apply(event(OrderExpired)).in(aggregate_root).strict