flaky

About

Flaky is a plugin for nose or py.test that automatically reruns flaky tests.

Ideally, tests reliably pass or fail, but sometimes test fixtures must rely on components that aren't 100% reliable. With flaky, instead of removing those tests or marking them to \@skip, they can be automatically retried.

For more information about flaky, see this presentation.

Marking tests flaky

To mark a test as flaky, simply import flaky and decorate the test with \@flaky:

``` {.sourceCode .python} from flaky import flaky

``` {.sourceCode .python}
@flaky
def test_something_that_usually_passes(self):
    value_to_double = 21
    result = get_result_from_flaky_doubler(value_to_double)
    self.assertEqual(result, value_to_double * 2, 'Result doubled incorrectly.')

By default, flaky will retry a failing test once, but that behavior can be overridden by passing values to the flaky decorator. It accepts two parameters: max_runs, and min_passes; flaky will run tests up to max_runs times, until it has succeeded min_passes times. Once a test passes min_passes times, it's considered a success; once it has been run max_runs times without passing min_passes times, it's considered a failure.

``` {.sourceCode .python} @flaky(max_runs=3, min_passes=2) def test_something_that_usually_passes(self): “"”This test must pass twice, and it can be run up to three times.””” value_to_double = 21 result = get_result_from_flaky_doubler(value_to_double) self.assertEqual(result, value_to_double * 2, ‘Result doubled incorrectly.’)

### Marking a class flaky

In addition to marking a single test flaky, entire test cases can be
marked flaky:


``` {.sourceCode .python}
@flaky
class TestMultipliers(TestCase):
    def test_flaky_doubler(self):
        value_to_double = 21
        result = get_result_from_flaky_doubler(value_to_double)
        self.assertEqual(result, value_to_double * 2, 'Result doubled incorrectly.')

    @flaky(max_runs=3)
    def test_flaky_tripler(self):
        value_to_triple = 14
        result = get_result_from_flaky_tripler(value_to_triple)
        self.assertEqual(result, value_to_triple * 3, 'Result tripled incorrectly.')

The \@flaky class decorator will mark test_flaky_doubler as flaky, but it won't override the 3 max_runs for test_flaky_tripler (from the decorator on that test method).

Don't rerun certain types of failures

Depending on your tests, some failures are obviously not due to flakiness. Instead of rerunning after those failures, you can specify a filter function that can tell flaky to fail the test right away.

``` {.sourceCode .python} def is_not_crash(err, *args): return not issubclass(err[0], ProductCrashedError)

@flaky def test_something(): raise ProductCrashedError

@flaky(rerun_filter=is_not_crash) def test_something_else(): raise ProductCrashedError

Flaky will run [test\_something]{.title-ref} twice, but will only run
[test\_something\_else]{.title-ref} once.

It can also be used to incur a delay between test retries:


``` {.sourceCode .python}
import time

def delay_rerun(*args):
    time.sleep(1)
    return True

@flaky(rerun_filter=delay_rerun)
def test_something_else():
    ...

Activating the plugin

Like any nose plugin, flaky can be activated via the command line:

``` {.sourceCode .console} nosetests –with-flaky

With py.test, flaky will automatically run. It can, however be disabled
via the command line:


``` {.sourceCode .console}
py.test -p no:flaky

Command line arguments

No Flaky Report

Pass --no-flaky-report to suppress the report at the end of the run detailing flaky test results.

Shorter Flaky Report

Pass --no-success-flaky-report to suppress information about successful flaky tests.

Force Flaky

Pass --force-flaky to treat all tests as flaky.

Pass --max-runs=MAX_RUNS and/or --min-passes=MIN_PASSES to control the behavior of flaky if --force-flaky is specified. Flaky decorators on individual tests will override these defaults.

Additional usage examples are in the code - see test/test_example.py

Installation

To install, simply:

``` {.sourceCode .console} pip install flaky

Compatibility
-------------

Flaky is tested with the following test runners and options:

-   Nosetests. Doctests cannot be marked flaky.
-   Py.test. Works with [pytest-xdist]{.title-ref} but not with the
    [\--boxed]{.title-ref} option. Doctests cannot be marked flaky.

Contributing
------------

See
[CONTRIBUTING.rst](https://github.com/box/flaky/blob/master/CONTRIBUTING.rst).

### Setup

Create a virtual environment and install packages -


``` {.sourceCode .console}
mkvirtualenv flaky
pip install -r requirements-dev.txt

Testing

Run all tests using -

``` {.sourceCode .console} tox

```

The tox tests include code style checks via pep8 and pylint.

Copyright and License

Copyright 2015 Box, Inc. All rights reserved.

Licensed under the Apache License, Version 2.0 (the "License");
you may not use this file except in compliance with the License.
You may obtain a copy of the License at

   http://www.apache.org/licenses/LICENSE-2.0

Unless required by applicable law or agreed to in writing, software
distributed under the License is distributed on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
See the License for the specific language governing permissions and
limitations under the License.

Contributing

All contributions are welcome to this project.

Contributor License Agreement

Before a contribution can be merged into this project, please fill out the Contributor License Agreement (CLA) located at:

http://box.github.io/cla

To learn more about CLAs and why they are important to open source projects, please see the Wikipedia entry.

How to contribute

• File an issue - if you found a bug, want to request an enhancement, or want to implement something (bug fix or feature).
• Send a pull request - if you want to contribute code. Please be sure to file an issue first.

Pull request best practices

We want to accept your pull requests. Please follow these steps:

Step 1: File an issue

Before writing any code, please file an issue stating the problem you want to solve or the feature you want to implement. This allows us to give you feedback before you spend any time writing code. There may be a known limitation that can't be addressed, or a bug that has already been fixed in a different way. The issue allows us to communicate and figure out if it's worth your time to write a bunch of code for the project.

Step 2: Fork this repository in GitHub

This will create your own copy of our repository.

Step 3: Add the upstream source

The upstream source is the project under the Box organization on GitHub. To add an upstream source for this project, type:

``` {.sourceCode .console} git remote add upstream git@github.com:box/flaky.git

This will come in useful later.

### Step 4: Create a feature branch

Create a branch with a descriptive name, such as `add-search`.

### Step 5: Push your feature branch to your fork

As you develop code, continue to push code to your remote feature
branch. Please make sure to include the issue number you\'re addressing
in your commit message, such as:


``` {.sourceCode .console}
git commit -am "Adding search (fixes #123)"

This helps us out by allowing us to track which issue your commit relates to.

Keep a separate feature branch for each issue you want to address.

Step 6: Rebase

Before sending a pull request, rebase against upstream, such as:

``` {.sourceCode .console} git fetch upstream git rebase upstream/master

```

This will add your changes on top of what's already in upstream, minimizing merge issues.

Step 7: Run the tests

Make sure that all tests are passing before submitting a pull request.

Step 8: Send the pull request

Send the pull request from your feature branch to us. Be sure to include a description that lets us know what work you did.

Keep in mind that we like to see one issue addressed per pull request, as this helps keep our git history clean and we can more easily track down issues.

Finally, please add a note in HISTORY.rst under the Upcoming section detailing what's new in your change. These will become the release notes for the next release. In addition, feel free to add yourself to AUTHORS.rst if you aren't already listed.