If you happened to come across a bug or want to suggest a feature, feel free to say something!
If you'd like to contribute code, the steps below will help you get up and running.
Resource
or @resource
is an instantiated Avo::Resource
object.
resource_name
: Pluralized, machine name, snake cased version of a resource. Ex: resource_name
for /avo/resources/team_memberships
path is going to be TeamMemberships
.
Model
or @model
: Active Record Model
object.
Since v 1.20
(Jan 2022) we started to change the model
name with record
where it fits. The model
should be the class (ex: User
, Comment
, etc.) and record
will be the instantiated model class with data (ex: user
is a record when doing user = User.find(1)
).
Please fork Avo and create a descriptive branch for your feature or fix. We usually use the feature/
, chore/
, or fix/
branch prefixes. This way, the PR gets automatically labelled.
NOTE: We're using docker-compose
to run Postgres. While we find this incredibly helpful, it's not a requirement, but you'll have to BYOD (Bring Your Own Database).
Once you pull the code down to your machine, modify spec/dummy/config/database.yml
as appropriate to your environment (no changes needed if you're using Docker). Do not commit these changes. From here, running bin/init
will get you up-and-running.
From the root directory you can run bin/dev
and that will start a foreman process for the rails server, jsbundling and cssbundling. Then, navigate to localhost:3030
and enjoy the app.
Run AVO_ADMIN_PASSWORD=secret bin/rails db:seed
to seed the database with dummy data and create a user for yourself with the email [email protected]
and password secret
.
You may want to evaluate your changes in the context of a real project. To do so without publishing your own version of the gem, follow these instructions.
In the other project, change the Gemfile
entry for avo
to point to your local clone of this repo. For example:
gem 'avo', path: '../avo'
Avo's assets will not show up by default, resulting in 404 errors on /avo-assets/avo.js
and /avo-assets/avo.css
. To avoid this, you need to compile the asset bundles, and symlink them into public/avo-assets
.
Create symlinks for compiled assets into public. You'll only need to do this once.
# `cd` into the root directory of this project.
ln -sf $(pwd)/app/assets/builds/avo.js public/avo-assets/avo.js
ln -sf $(pwd)/app/assets/builds/avo.css public/avo-assets/avo.css
After that, you'll need to compile the asset bundles any time you make changes to the JS or CSS code:
yarn build
When running tests you have two options. 1. running them on your local database or 2. to use a database in a docker environment.
Run the migration script bin/rails db:migrate
. Now you'll be able to run the test scripts below:
Copy the .env.test.sample
(cp spec/dummy/.env.test.sample spec/dummy/.env.test
) and update it with the proper credentials for your local database. Run the migration script bin/rails db:migrate
. Now you'll be able to run the test scripts below:
We've set up a few helpers to get you going:
- Run all tests (slow):
bin/test
- Run unit tests (fast):
bin/test unit
- Run system tests (slow):
bin/test system
- Run a particular spec file/test (fast):
bin/test ./spec/features/hq_spec.rb
You may want to run the tests on your docker container. To do that, update your .env.test
file with valid credentials for the docker setup like below. We do provide a docker-compose.yml
file for the testing DB, so you can do docker compose up
, but that's not our preferred method and you might need to do some tweaks. You'll then be able to run the migration script and the testing commands.
POSTGRES_HOST=localhost
POSTGRES_PORT=5433
POSTGRES_USERNAME=postgres
POSTGRES_PASSWORD=
We use appraisal to run tests against multiple versions of rails. When the gemfile gets updated you must also run bundle exec appraisal install
to update the versioned ones and commit them to the repo.
Please read the RELEASE.MD for release schedules.
In order to be able to test the custom Stimulus integration we need to have some kind of JS content injected in the dummy app. We also need not to have that content present in every Avo installation. That's why we ejected the _head.html.erb
partial that loads the avo_custom.js
file from the engine's build
directory. The build directory is not commited to git so we're all good on that. In order to have that file built and watche we added a cjs
(short for custom-js
) process in the Procfile.dev
(which is used only in development).
To keep track of the schema structure for the models run annotate --models --exclude fixtures
in the dummy app.