Pending Tests

If it bothers us to see the integration test continue to fail while we write the unit tests that will make it pass, RSpec allows us to specify a test as pending or to skip it altogether. In RSpec, any it method defined without a block is considered to be “pending.”

it "bends steel in its bare hands"

The :pending argument

We can temporarily mark an it or describe block as pending by adding :pending as a second argument after the string as shown below:

it "bends steel in its bare hands", :pending do
  #anything
end

The pending method

Alternatively, we can use the method pending in the spec:

it "bends steel in its bare hands" do
  pending "not implemented yet"
end

Pending tests failures

In RSpec all pending specs are actually run if there is code in the block part of the spec. The code is executed, with any failure in the pending spec treated as a pending result rather than a failure result. However, if the code in the pending spec passes, we’ll get an error that effectively means, “We said this was pending, but lo and behold, it works. Maybe it’s not actually pending anymore, so please remove the pending status.”

Skipping tests

If we want the spec to not run, which means that we do not check for whether the spec passes, we employ the preceding syntax but use skip instead of pending. Alternatively, we can prefix the method name with x, as in xit or xdescribe. A skipped test will not run, meaning that we won’t get any notification if the test suddenly starts to pass.