Ruby RSPEC 测试

Question

I would like to know how to satisfy the is_valid test, I tried it in several ways but the test does not pass what I have to write in the class

require 'spec_helper'

RSpec.describe PixKey do
  subject(:pix_key) { described_class.new(key) }

  let(:key) { 'email@wesley.io' }

  describe '.new' do
    context 'with a valid key' do
      it { is_expected.to be_a(described_class) }

      it { is_expected.to be_valid }
    end

Cod class

class PixKey

  attr_accessor :key

  def new(key)
    @key = :key
    validates :key, presence: true
  end

end

Err

PixKey .new with a valid key is expected to be a kind of PixKey

  is expected to be valid (FAILED - 1)

Answer 1

I assume that this is a test driven development (TDD) assignment where you got some specs and have to implement a class that passes the tests.

In order to get your specs green, you have to implement some methods. To do so, you have to understand RSpec's syntax:

• described_class – this is whatever comes after describe . In your spec, it's describe PixKey , so described_class.new(key) is equivalent to PixKey.new(key)
• be_a is a type matcher and equivalent to calling kind_of?
• be_valid is a predicate matcher and equivalent to calling valid?

---

Let's start with an empty class:

class PixKey
end

Running the specs gives an error: ( --fail-fast makes RSpec stop after the first failing spec)

$ rspec pix_key_spec.rb --fail-fast

Failures:

  1) PixKey.new with a valid key
     Failure/Error: subject(:pix_key) { described_class.new(key) }

     ArgumentError:
       wrong number of arguments (given 1, expected 0)

❌ 1 example, 1 failure

RSpec called PixKey.new(key) but new doesn't take an argument. In order to get that working, you could implement a class method called new , ie:

class PixKey
  def self.new(key)
    allocate
  end
end

But Ruby classes already come with a built-in new method that handles the allocation for you. You typically just implement an instance method called initialize :

class PixKey
  def initialize(key)
  end
end

Ruby's built-in new will call initialize and pass all arguments to it.

Note that I don't do anything with key yet. This is a core part of TDD: you write just enough to get the test passing. Let's try it again:

$ rspec pix_key_spec.rb --fail-fast

Failures:

  1) PixKey.new with a valid key is expected to be valid
     Failure/Error: it { is_expected.to be_valid }
       expected #<PixKey:0x000000010201de90> to respond to `valid?`

❌ 2 examples, 1 failure

This time, it's complaining about our PixKey instance not responding to valid? , so let's implement that method: (the trailing ? is part of the method name)

class PixKey
  def initialize(key)
  end

  def valid?
  end
end

And run the tests again:

$ rspec pix_key_spec.rb --fail-fast

Failures:

  1) PixKey.new with a valid key is expected to be valid
     Failure/Error: it { is_expected.to be_valid }
       expected `#<PixKey:0x00000001027b1da0>.valid?` to be truthy, got nil

❌ 2 examples, 1 failure

Now it says that our valid? method is expected to return a "truthy" result. In Ruby, everything besides nil and false qualifies as truthy, but I'd just just use true :

class PixKey
  def initialize(key)
  end

  def valid?
    true
  end
end

$ rspec pix_key_spec.rb --fail-fast

✅ 2 examples, 0 failures

At this point, the specs pass. But obviously, the class is far from complete: key is never used and valid? just always returns true without validating anything.

I assume that there are more tests to pass. Keep working in small steps, just one test at a time.