Skip to content

Latest commit

 

History

History
129 lines (93 loc) · 3.96 KB

README.md

File metadata and controls

129 lines (93 loc) · 3.96 KB
Raw

Functional tests

Functional tests reproduces a use case or reproduce a bug or ensure a fix.

Functional tests runs on CI by calling tests_functcional/runtests.py script this script is called when you run make test_functional in the root of the repository.

This script walks through the tests_functional directory and runs all then for every subdirectory it looks for one of the following files in order:

  • Makefile
  • test.sh
  • test.py
  • app.py
  • program.py
  • manage.py

The first found file will be executed via subprocess.check_call and the execution context will be

  • cwd: The subdirectory and then it will try to run also from the parent folder.
  • env: os.environ + variables defined on env.toml file if it exists.

How to assert in a test?

Makefile or test.sh

If the test is executed on the shell level, you can ensure it returns 0 as exit code.

Ex: On a makefile or test.sh you can do python file.py | grep "something" and it will return 0 if the grep found the string.

test.py or app.py

Tests executed on Python level is executed via python filename.py so anything that returns 0 as exit code will be considered a success.

Ex: assert settings.KEY == "value" will return 0 as exit code if the assertion is true.

manage.py (Django)

On a Django application, if no other test file is found such as (Makefile, test.sh) then runner will call python manage.py test with no arguments.

NOTE to use pytest or more flexibility on command call please call it from the Makefile (test target) or test.sh, the only important thing is that the command returns 0 as exit code on success.

Envvars

Environment variables are very important for Dynaconf so most of the functional tests will need to customize those there are 2 ways:

  • Write a .env file and then pass load_dotenv=True to Dynaconf contructor.
  • Use env.toml file to define environment variables, IMPORTANT that this file has only upper case keys and all values are strings: THING = "value".

Skipping functional tests

ON the skipfile.toml you can define a list of folder names to skip depending on the platform.

nt = ["tests_to_skip_on_windows"]
posix = ["tests_to_skip_on_linux"]
macos = ["tests_to_skip_on_macos"]

Those names matches os.name for the platform.

How to add a new test?

NOTE: Give your functional test example folder a meaningful name, for example test_django_with_yaml/ or test_feature_x_works_with_y_enabled/

  1. Create a new folder under tests_functional:

    $ mkdir tests_functional/my_test

    IMPORTANT!!!! if your test is to reproduce an issue, please add it to tests_functional/issues/ folder, for example tests_functional/issues/999_launching_a_rocket/

  2. Go to that folder:

    $ cd tests_functional/my_test
# or
$ cd tests_functional/issues/999_launching_a_rocket

  3. Create an app.py:

    from dynaconf import Dynaconf
settings = Dynaconf(
    envvar_prefix="MYAPP",
    settings_files=["settings.toml"],
    load_dotenv=True,
    # add parameters matching your use case
)

assert settings.KEY == "value"
# Add things needed to assert your test case

    NOTE Ensure you have the needed assertions at the end of the file.

  4. Create needed artifacts for the test to run:

    $ touch settings.toml
$ touch .env

    Optionally

    $ touch env.toml

  5. If you need more flexibility to call the command or you have to pass extra envvars, do extra checks or connect services please use a Makefile or a test.sh

    $ touch Makefile
    test:
    python app.py

    or

    $ touch test.sh
    #!/usr/bin/env bash
python app.py | grep "something"

TIP If you need multiple variations of the same structure with fewer changes, you can create subfolders with the same structure.