5.5 Unit Testing🔗ℹ

• In ~is mode, evaluates the expected_expr or expected_body, then checks that the original result is not an exception and is equal by ==.

• In ~is_now mode, check like ~is mode, but using is_now.

• In ~is_a mode, checks that the original result is not an exception and satisfies annot.

• In ~prints_like mode, evaluates the expected_expr or expected_body, then checks that the original result is not an exception and has the same printed form as the expected result.

• In ~matches mode, checks that the original result is not an exception, that the number of result values matches the number of supplied expected_binds, and that each value matches the corresponding expected_bind.

• In ~prints mode, evaluates the expected_expr or expected_body in a mode that captures output to Port.Output.current() to a string, then checks that there is no exception and the output string is the same as the expected result.

• In ~throws mode, obtains one or more strings (as multiple values) by evaluating expected_body or expected_expr, then checks that original body or expr threw an exception and that each string is contained in the exception message.

• In ~completes mode, checks merely that the original result is not an exception. A #void is implicitly added to the end of body or expr, so it could end with a definition.

If ~eval is present, then body is quoted to be parsed and evaluated using eval at run time. A typical use of ~eval is to test syntax errors. When evaluator_expr is provided, then its result is installed as the current evaluator for eval; otherwise, an evaluator is created with Evaluator.make_rhombus. The body sequence is evaluated as an interactive form (i.e., eval is called with ~as_interactive: #true).

Providing multiple expr expected_result groups in a single check form is the same as providing each group to a separate check form. The ~eval mode is not supported in the shorthand forms of check.

The evaluation of an expr or body sequence is wrapped with a prompt for the default continuation prompt tag.

When a test fails, a failure message is printed to the current error port, but using the current error display handler when a source location is available. No output is printed for a successful test.

> check 1+1 ~is 2

> check 1+1 ~is 3

check: failed

got: 2

expected: 3 Port.Output(#’string)

> check:     1+1     ~is 2

> check:     1+1 ~is 2     1+1 ~is 3

check: failed

got: 2

expected: 3 Port.Output(#’string)

> check:     1+1     ~is_a Int

> check:     1+1     ~is_a String

check: failed

got: 2

expected: satisfying String Port.Output(#’string)

> check:     [1+1, 1+2]     ~matches [_ :: Int, ...]

> check:     '$(1+1)'     ~prints_like '2'

> check:     1+"a"     ~throws "expected: Number"

check: failed

got: exception +: value does not satisfy annotation

annotation: Number

value: "a"

expected: exception "expected: Number" Port.Output(#’string)

> check:     + // oops: causes syntax error for overall `check` form     ~throws "infix operator without preceding argument"

+: infix operator without preceding argument Port.Output(#’string)

> check:     ~eval // lets `check` confirm the expected syntax error     +     ~throws "infix operator without preceding argument"

> check:     ~eval     +     ~throws values("+", "infix operator without preceding argument")