Behat is used to execute User Acceptance Tests written in Gherkin. These tests are stored in tests/features/
.
To develop acceptance tests locally, install Vagrant. This allows you to run tests against different versions of PHP. Find the line "PHPVERSION=7.4" and change it to the version of PHP you want to use. Put @wip in the feature you are working on, so that it doesn't interfere with Github CI.
Then run
vagant up
Once it has run once, you can fire off further runs with
vagant ssh
cd /vagrant/tests
make test
You can also use the following commands:
make verbosetest - show progress during testing, useful to identify issues in your tests make test-wip - only run the features marked "@wip" make testall - run all tests
$ curl -sS https://getcomposer.org/installer | php
$ mv composer.phar /usr/local/bin/composer
You can read the official documentation here
To install behat and other dependencies run the following command on the project root:
$ composer install
run the following command:
$ cp default.behat.yml behat.yml
Edit behat.yml file and customize your base_url
. The base_url
determines which phpList server you wish to test. This could be a local development copy running on your machine, or a remotely hosted server.
Create a database, but keep it empty
$ mysqladmin -uroot -p create phplisttestdb
$ mysqladmin -uroot -p -e "grant all on phplisttestdb.* to phplist@localhost identified by 'testpassword'"
Edit your phplist config.php
file to use these details
Some tests can only be run once before they change the system state upon which they depend. These tests should only be run once, or only in automated testing environments like Github actions. Execute these and all other tests together by running the following command from within your phpList 3 code root directory:
$ vendor/bin/behat
Usually you will want to avoid running the first time tests and run all the others instead. To do so execute the following command from within your phpList 3 code root directory:
$ vendor/bin/behat --tags '~@initialise'
This will run all tests except for those with the initialise
tag.
When you run it locally or using Vagrant, and a test fails, it will drop a screenshot of the page where it failed in tests/output/screenshots
That way you can find out why it failed and update your test.
Add writing tests documentation.