# `ExUnit`
[🔗](https://github.com/elixir-lang/elixir/blob/v1.20.0-rc.4/lib/ex_unit/lib/ex_unit.ex#L5)

Unit testing framework for Elixir.

## Example

A basic setup for ExUnit is shown below:

    # File: assertion_test.exs

    # 1) Start ExUnit.
    ExUnit.start()

    # 2) Create a new test module and use "ExUnit.Case".
    defmodule AssertionTest do
      # 3) Note that we pass "async: true", this runs the tests in the
      #    test module concurrently with other test modules. The
      #    individual tests within each test module are still run serially.
      use ExUnit.Case, async: true

      # 4) Use the "test" macro instead of "def" for clarity.
      test "the truth" do
        assert true
      end
    end

To run the tests above, run the file using `elixir` from the
command line. Assuming you named the file `assertion_test.exs`,
you can run it as:

    $ elixir assertion_test.exs

## Case, Callbacks and Assertions

See `ExUnit.Case` and `ExUnit.Callbacks` for more information
about defining test cases and setting up callbacks.

The `ExUnit.Assertions` module contains a set of macros to
generate assertions with appropriate error messages.

## Integration with Mix

Mix is the project management and build tool for Elixir. Invoking `mix test`
from the command line will run the tests in each file matching the pattern
`*_test.exs` found in the `test` directory of your project.

You must create a `test_helper.exs` file inside the
`test` directory and put the code common to all tests there.

The minimum example of a `test_helper.exs` file would be:

    # test/test_helper.exs
    ExUnit.start()

Mix will load the `test_helper.exs` file before executing the tests.
It is not necessary to `require` the `test_helper.exs` file in your test
files. Run `mix help test` for more information.

# `configure_opts`

```elixir
@type configure_opts() :: [
  {:assert_receive_timeout, non_neg_integer()}
  | {:autorun, boolean()}
  | {:capture_log, boolean() | [{:level, Logger.level()}]}
  | {:colors,
     enabled: boolean(),
     success: atom(),
     invalid: atom(),
     skipped: atom(),
     failure: atom(),
     error_info: atom(),
     extra_info: atom(),
     location_info: [atom()],
     diff_insert: atom(),
     diff_insert_whitespace: IO.ANSI.ansidata(),
     diff_delete: atom(),
     diff_delete_whitespace: IO.ANSI.ansidata()}
  | {:exclude, atom() | [atom() | {atom(), any()}]}
  | {:exit_status, non_neg_integer()}
  | {:failures_manifest_path, String.t()}
  | {:formatters, [module()]}
  | {:include, atom() | [atom() | {atom(), any()}]}
  | {:max_cases, pos_integer()}
  | {:max_failures, pos_integer() | :infinity}
  | {:only_test_ids, [test_id()]}
  | {:rand_algorithm, atom()}
  | {:refute_receive_timeout, non_neg_integer()}
  | {:seed, non_neg_integer()}
  | {:slowest, non_neg_integer()}
  | {:slowest_modules, non_neg_integer()}
  | {:stacktrace_depth, non_neg_integer()}
  | {:timeout, pos_integer()}
  | {:trace, boolean()}
  | {:test_location_relative_path, String.t()}
  | {atom(), term()}
]
```

Configuration options for ExUnit.

See `configure/1` for detailed documentation of each option.

# `failed`

```elixir
@type failed() :: [{Exception.kind(), reason :: term(), Exception.stacktrace()}]
```

The error state returned by `ExUnit.Test` and `ExUnit.TestModule`

# `state`

```elixir
@type state() ::
  nil
  | {:excluded, binary()}
  | {:failed, failed()}
  | {:invalid, ExUnit.TestModule.t()}
  | {:skipped, binary()}
```

All tests start with a state of `nil`.

A finished test can be in one of five states:

  1. Passed (also represented by `nil`)
  2. Failed
  3. Skipped (via @tag :skip)
  4. Excluded (via :exclude filters)
  5. Invalid (when setup_all fails)

# `suite_result`

```elixir
@type suite_result() :: %{
  excluded: non_neg_integer(),
  failures: non_neg_integer(),
  skipped: non_neg_integer(),
  total: non_neg_integer()
}
```

A map representing the results of running a test suite

# `test_id`

```elixir
@type test_id() :: {module(), name :: atom()}
```

# `after_suite`
*since 1.8.0* 

```elixir
@spec after_suite((suite_result() -> any())) :: :ok
```

Sets a callback to be executed after the completion of a test suite.

Callbacks set with `after_suite/1` must accept a single argument, which is a
map containing the results of the test suite's execution.

If `after_suite/1` is called multiple times, the callbacks will be called in
reverse order. In other words, the last callback set will be the first to be
called.

# `async_run`
*since 1.12.0* 

```elixir
@spec async_run() :: Task.t()
```

Starts tests asynchronously while test cases are still loading.

It returns a task that must be given to `await_run/0` when a result
is desired.

# `await_run`
*since 1.12.0* 

```elixir
@spec await_run(Task.t()) :: suite_result()
```

Awaits for a test suite that has been started with `async_run/0`.

# `configuration`

```elixir
@spec configuration() :: Keyword.t()
```

Returns ExUnit configuration.

For the available configuration options, see `configure/1`.

# `configure`

```elixir
@spec configure(configure_opts()) :: :ok
```

Configures ExUnit.

## Options

ExUnit supports the following options:

  * `:assert_receive_timeout` - the timeout to be used on `assert_receive`
    calls in milliseconds, defaults to `100`;

  * `:autorun` - if ExUnit should run by default on exit. Defaults to `true`;

  * `:capture_log` - if ExUnit should default to keeping track of log messages
    and print them on test failure. Can be overridden for individual tests via
    `@tag capture_log: false`. This can also be configured to a specific level
    with `capture_log: [level: LEVEL]`, to capture all logs but only keep those
    above `LEVEL`. Note that `on_exit` and `setup_all` callbacks may still log,
    as they run outside of the testing process. To silent those, you can use
    `ExUnit.CaptureLog.capture_log/2` or consider disabling logging altogether.

  * `:colors` - a keyword list of color options to be used by some formatters:
    * `:enabled` - boolean option to enable colors, defaults to `IO.ANSI.enabled?/0`;

    * `:success` - success message (defaults to `:green`)
    * `:invalid` - invalid test message (defaults to `:yellow`)
    * `:skipped` - skipped test message (defaults to `:yellow`)
    * `:failure` - failed test message (defaults to `:red`)
    * `:error_info` - display of actual error (defaults to `:red`)
    * `:extra_info` - additional information (defaults to `:cyan`)
    * `:location_info` - filename and tags (defaults to `[:bright, :black]`)
    * `:diff_insert` - color of the insertions on diffs, defaults to `:green`;
    * `:diff_insert_whitespace` - color of the whitespace insertions on diffs,
      defaults to `IO.ANSI.color_background(2, 0, 0)`;
    * `:diff_delete` - color of the deletions on diffs, defaults to `:red`;
    * `:diff_delete_whitespace` - color of the whitespace deletions on diffs,
      defaults to `IO.ANSI.color_background(0, 2, 0)`;

  * `:exclude` - specifies which tests are run by skipping tests that match the
    filter. For more information, see the "Tags" and "Filters" sections in the
    documentation for `ExUnit.Case`;

  * `:exit_status` - specifies an alternate exit status to use when the test
    suite fails. Defaults to `2`;

  * `:failures_manifest_path` - specifies a path to the file used to store failures
    between runs;

  * `:formatters` - the formatters that will print results,
    defaults to `[ExUnit.CLIFormatter]`;

  * `:include` - specifies which tests are run by skipping tests that do not
    match the filter. Keep in mind that all tests are included by default, so unless they are
    excluded first, the `:include` option has no effect. To only run the tests
    that match the `:include` filter, exclude the `:test` tag first.
    For more information, see the "Tags" and "Filters" sections in the
    documentation for `ExUnit.Case`;

  * `:max_cases` - maximum number of tests to run in parallel. Only tests from
    different modules run in parallel. It defaults to `System.schedulers_online * 2`
    to optimize both CPU-bound and IO-bound tests;

  * `:max_failures` - the suite stops evaluating tests when this number of test failures
    is reached. All tests within a module that fail when using the
    [`setup_all/1,2`](`ExUnit.Callbacks.setup_all/1`) callbacks
    are counted as failures. Defaults to `:infinity`;

  * `:only_test_ids` - a list of `{module_name, test_name}` tuples that limits
    what tests get run. This is typically used by Mix to filter which tests
    should run;

  * `:rand_algorithm` - algorithm to be used when generating the test seed.
    Available algorithms can be found in Erlang's
    [`:rand`](`:rand`) documentation (see
    [`:rand.builting_arg/0`](https://www.erlang.org/doc/apps/stdlib/rand.html#t:builtin_alg/0)).
    Available since v1.16.0. Before v1.16.0, the algorithm was hard-coded to
    `:exs1024`. On Elixir v1.16.0 and after, the default changed to `:exsss`;

  * `:refute_receive_timeout` - the timeout to be used on `refute_receive`
    calls in milliseconds, defaults to `100`;

  * `:seed` - an integer seed value to randomize the test suite. This seed
    is also mixed with the test module and name to create a new unique seed
    on every test, which is automatically fed into the [`:rand`](`:rand`) module.
    This provides randomness between tests, but predictable and reproducible
    results. A `:seed` of `0` will disable randomization and the tests in each
    file will always run in the order that they were defined in;

  * `:slowest` - prints timing information for the N slowest tests. Running
    ExUnit with slow test reporting automatically runs in `trace` mode. It
    is disabled by default;

  * `:slowest_modules` - prints timing information for the N slowest test modules. Running
    ExUnit with slow test reporting automatically runs in `trace` mode. It
    is disabled by default;

  * `:stacktrace_depth` - configures the stacktrace depth to be used
    on formatting and reporters, defaults to `20`;

  * `:timeout` - sets the timeout for the tests in milliseconds, defaults to `60_000`;

  * `:trace` - sets ExUnit into trace mode, this sets `:max_cases` to `1` and
    prints each test case and test while running. Note that in trace mode test
    timeouts will be ignored as timeout is set to `:infinity`;

  * `:test_location_relative_path` - the test location is the file:line information
    printed by tests as a shortcut to run a given test. When this value is set,
    the value is used as a prefix for the test itself. This is typically used by
    Mix to properly set-up umbrella projects;

Any arbitrary configuration can also be passed to `configure/1` or `start/1`,
and these options can then be used in places such as custom formatters. These
other options will be ignored by ExUnit itself.

# `fetch_test_supervisor`
*since 1.11.0* 

```elixir
@spec fetch_test_supervisor() :: {:ok, pid()} | :error
```

Fetches the test supervisor for the current test.

Returns `{:ok, supervisor_pid}` or `:error` if not called from the test process.

This is the same supervisor as used by `ExUnit.Callbacks.start_supervised/2`
and similar, see `ExUnit.Callbacks` module documentation for more information.

# `plural_rule`

```elixir
@spec plural_rule(binary()) :: binary()
```

Returns the pluralization for `word`.

If one is not registered, returns the word appended with an "s".

# `plural_rule`

```elixir
@spec plural_rule(binary(), binary()) :: :ok
```

Registers a `pluralization` for `word`.

If one is already registered, it is replaced.

# `run`

```elixir
@spec run([module()]) :: suite_result()
```

Runs the tests. It is invoked automatically
if ExUnit is started via `start/1`.

From Elixir v1.14, it accepts an optional list of modules to run
as part of the suite. This is often used to rerun modules already
loaded in memory.

Returns a map containing the total number of tests, the number
of failures, the number of excluded tests and the number of skipped tests.

# `start`

```elixir
@spec start(configure_opts()) :: :ok
```

Starts ExUnit and automatically runs tests right before the
VM terminates.

It accepts a set of `options` to configure `ExUnit`
(the same ones accepted by `configure/1`).

If you want to run tests manually, you can set the `:autorun` option
to `false` and use `run/0` to run tests.

---

*Consult [api-reference.md](api-reference.md) for complete listing*