You'll notice we wrapped docker-compose
earlier, but you'll generally want to use this to interact with the app.
bundle exec rails console
→docker-compose run web rails console
bundle exec rake db:migrate
→docker-compose run web rake db:migrate
bundle exec rake db:test:prepare
→docker-compose run web rake db:test:prepare
- and so on...
We do have a couple other helper scripts:
bin/setup
→ sets up the appbin/migrate
→ migrates your databasebin/reseed
→ re-runsrake db:seed_fu
to re-seed your database
You'll also notice that the test
container we mentioned above is running guard
. This means that file changes will be observed and tests re-run on those files.
You can certainly run docker-console run test rspec spec
, but guard
can help you by constantly watching for failing specs.
Need to stop the containers? Either Ctrl+C
or in a seperate prompt run docker-compose stop
.
To start the services again you can run docker-compose up
, or docker-compose start
to start the containers in a detached state.
If you ever need to rebuild you can run docker-compose up --build
. Unless you've destroyed your Docker container images, this should be faster than the first run.
You can use git
as you normally would, either on your own host machine or in Docker's web
container.
The Code Corps API is intended to work alongside a client written in Ember. For that purpose, the Rails application exposes all of its API endpoints behind an api.
subdomain.
On the Ember client side of things, we use ember-cli-deploy
with a redis
plugin to deploy the client application to redis. Multiple revisions are maintained this way.
Any server request pointing to the main domain and not the api.
subdomain is redirected to the API's ember_index_controller#index
. There, depending on the remainder of the request path and the current environment, a specific revision of the Ember app's index.html
is retrieved from redis and rendered. This index.html
can be:
- the development revision, if the current environment is development
- a specific deployed revision in production if the request contains a revision parameter in SHORT_UUID format
- the latest deployed revision in production if the request does not contain a revision parameter
- a plain text string containing "INDEX NOT FOUND" if a revision was specified, but the key for the specified revision was not found by redis
Because the app is running foreman
, debugging using pry
won't work the same way. If you want to use pry
, you'll need to debug remotely.
Add binding.remote_pry
where you want to pause:
class UsersController < ApplicationController
def index
binding.remote_pry
...
end
end
Load a page or make a request that triggers the code. Connect to the session:
$ docker-compose run web pry-remote