Dart backend for cocoon

This folder contains a Dart based backend for Cocoon.

Building and running

Prerequisites

Install the Google Cloud Developer Tools Command Line Interface (gcloud). Then initialize it and authenticate yourself by running:

gcloud auth login
gcloud init

Install Flutter

export PATH="$PATH":"path/to/flutter/bin/"
flutter upgrade
flutter pub get
export PATH="$PATH":"path/to/flutter/bin/cache/dart-sdk/bin/"

Running the tests

$ dart test

Running codegen

JSON

To update the JSON serialization generated code, run:

$ dart pub run build_runner build

Any updates should be checked into source control.

Protobuf

To update the Protocol Buffer generated code:

Download and install the protocol buffer compiler (protoc). Once installed, update your PATH to include the path to the protoc binary.
On Linux, you may be able to use sudo apt-get install protocol-compiler to install. On macOS, you may be able to use brew install protobuf
Install the protoc_plugin Dart package. Once installed, update your PATH to include the path to the protoc_plugin/bin directory (or $HOME/.pub-cache/bin if you used pub global activate protoc_plugin).

Run the following command:

$ protoc \
    --plugin=/path/to/protoc_plugin/bin/ \
    --dart_out=. \
    lib/src/path/to/file.proto

Remove the unused generated files:
```
$ find . -regex '.*\.$pbjson\|pbserver$\.dart' -delete
```
(you can remove the *.pbenum.dart files too, except for protobuffers that actually define enums, like build_status_response.proto)

Generating cloud datastore indexes

To update the indexes in the App Engine project, run:

$ gcloud datastore indexes create index.yaml

Local development

Using physical machine

Setting up the environment

export COCOON_USE_IN_MEMORY_CACHE=true

This environment is needed as you don't have access to the remote redis instance during local development.

Starting server

export COCOON_USE_IN_MEMORY_CACHE=true
dart bin/server.dart

If you see Serving requests at 0.0.0.0:8080 the dev server is working.

Using Docker

Running a local development instance

Once you've installed Docker and have the docker command-line tool in your path, then you can use the following commands to build, run, stop, and kill a local development instance.

# Build the docker image
$ docker build -t local .

# Start the local container, clearing the console buffer and tailing the logs
$ container_id="$(docker run -d -p 8080:8080 local)" && \
    clear && \
    printf '\e[3J' && \
    docker logs $container_id -f

# Stop the local Docker container
$ docker container ls|grep local|tr -s ' '|cut -d' ' -f1|xargs docker container stop

# Remove the local Docker image
$ docker images|grep local|tr -s ' '|cut -d' ' -f3|xargs docker rmi -f

ssh into instance

$ docker exec -it <container name> /bin/bash

Deploying a release to App Engine

Auto-deploy

Cocoon auto deployment has been set up via Google Cloud Build daily on Workdays.

[Manual-deploy(go/cocoon-cloud-build#manual-deploy)

Using the cloud build

This is easy to deploy if you simply want a new version based on the latest commit. Open Cloud Build dashboard and click run in the push-master trigger (example)

Using a cocoon checkout Let PROJECT_ID be the Google Cloud Project ID and VERSION be the version you're deploying to App Engine. Visit https://console.cloud.google.com/appengine/versions?project=flutter-dashboard for the list of current versions.

$ dart dev/deploy.dart --version version-$(git rev-parse --short HEAD) --project flutter-dashboard

The deploy script will build the Flutter project and copy it over for deployment. Then it will use the Google Cloud CLI to deploy the project to AppEngine.

For more options run:

$ dart dev/deploy.dart --help

Branching support for flutter repo

Add targeted branches in dev/branches.txt, based on which cocoon API filters targeted branches and then runs tests on those branches. With tests running against different branches, the frontend then supports listing commits on a specific branch (defaulting to master).

Once the PR has merged, go to https://flutter-dashboard.appspot.com/api/flush-cache?key=flutterBranches to flush the cache and have the latest change take effect. Otherwise, it can take 12 hours for it to propagate. Ensure you are signed into cocoon dashboard otherwise it will throw a 403.