A sane way of using mocks in Elixir. It borrows a lot from both Meck & Mox! Thanks @eproxus & @josevalim
Just add mimic to your list of dependencies in mix.exs:
def deps do
[
{:mimic, "~> 0.2", only: :test}
]
end
If :applications
key is defined inside your mix.exs
or you run mix test --no-start
, you probably want to add Application.ensure_all_started(:mimic)
in your test_helper.exs
Modules need to be prepared so that they can be used.
You must first call copy
in your test_helper.exs
for
each module that may have the behaviour changed.
Mimic.copy(Calculator)
ExUnit.start()
Calling copy
will not change the behaviour of the module.
The user must call stub/1
, stub/3
, expect/4
or reject/1
so that the functions can
behave differently.
Then for the actual tests one could use it like this:
use ExUnit.Case, async: true
import Mimic
# Make sure mocks are verified when the test exits
setup :verify_on_exit!
test "invokes add once and mult twice" do
Calculator
|> stub(:add, fn x, y -> :stub end)
|> expect(:add, fn x, y -> x + y end)
|> expect(:mult, 2, fn x, y -> x * y end)
assert Calculator.add(2, 3) == 5
assert Calculator.mult(2, 3) == 6
assert Calculator.add(2, 3) == :stub
end
stub/1
will change every module function to throw an exception if called.
stub(Calculator)
** (Mimic.UnexpectedCallError) Stub! Unexpected call to Calculator.add(3, 7) from #PID<0.187.0>
code: assert Calculator.add(3, 7) == 10
stub/3
changes a specific function to behave differently. If the function is not called no verification error will happen.
expect/4
changes a specific function and it works like a queue of operations. It has precedence over stubs and if not called a verification error will be thrown.
If the same function is called with expect/4
the order will be respected:
Calculator
|> stub(:add, fn _x, _y -> :stub end)
|> expect(:add, fn _, _ -> :expected_1 end)
|> expect(:add, fn _, _ -> :expected_2 end)
assert Calculator.add(1, 1) == :expected_1
assert Calculator.add(1, 1) == :expected_2
assert Calculator.add(1, 1) == :stub
expect/4
has an optional parameter which is the amount of calls expected:
Calculator
|> expect(:add, 2, fn x, y -> {:add, x, y} end)
assert Calculator.add(1, 3) == {:add, 1, 3}
assert Calculator.add(4, 5) == {:add, 4, 5}
One may want to reject calls to a specific function. reject/1
can be used to achieved this behaviour.
reject(&Calculator.add/2)
assert_raise Mimic.UnexpectedCallError, fn -> Calculator.add(4, 2) end
The default mode is private which means that only the process and explicitly allowed process will see the different behaviour.
Calling allow/2
will permit a different pid to call the stubs and expects from the original process.
Global mode can be used with set_mimic_global
like this:
setup :set_mimic_global
test "invokes add and mult" do
Calculator
|> expect(:add, fn x, y -> x + y end)
|> expect(:mult, fn x, y -> x * y end)
Task.async(fn ->
assert Calculator.add(2, 3) == 5
assert Calculator.mult(2, 3) == 6
end)
|> Task.await
end
This means that all processes will get the same behaviour
defined with expect & stub. This option is simpler but tests running
concurrently will have undefined behaviour. It is important to run with async: false
.
One could use :set_mimic_from_context
instead of using :set_mimic_global
or :set_mimic_private
. It will be private if async: true
, global otherwise.
After calling Mimic.copy(MyModule)
, calls to functions belonging to this module will first go through an ETS table to check which pid sees what (stubs, expects or call original).
It is really fast but it won't be as fast as calling a no-op function. Here's a very simple benchmark:
defmodule Enumerator do
def to_list(x, y), do: Enum.to_list(x..y)
end
Benchmarking Enumerator.to_list(1, 100)
:
Name ips average deviation median 99th %
mimic 116.00 K 8.62 μs ±729.13% 5 μs 29 μs
original 19.55 K 51.15 μs ±302.46% 34 μs 264 μs
Comparison:
mimic 116.00 K
original 19.55 K - 5.93x slower
Benchmarking Enumerator.to_list(1, 250)
:
Name ips average deviation median 99th %
original 131.49 K 7.61 μs ±167.90% 7 μs 16 μs
mimic 105.47 K 9.48 μs ±145.21% 9 μs 27 μs
Comparison:
original 131.49 K
mimic 105.47 K - 1.25x slower
There's a small fixed price to pay when mimic is used but it is unnoticeable for tests purposes.
Thanks to @jamesotron and @alissonsales for all the help! 🎉