Skip to content

Latest commit

 

History

History
119 lines (76 loc) · 4.65 KB

CONTRIBUTING.md

File metadata and controls

119 lines (76 loc) · 4.65 KB

How to contribute to Kuzzle

Here are a few rules and guidelines to follow if you want to contribute to Kuzzle and, more importantly, if you want to see your pull requests accepted by Kuzzle team.

Coding style

We use most of the NPM Coding Style rules, except for these ones:

  • Semicolons at the end of lines
  • 'Comma first' rule is not followed

Guidelines

  • Prefer async/await or promises instead of callbacks as often as you can
    • Except for methods invoked before the funnel module: ALWAYS use callbacks there to prevent event loop saturation (i.e. mostly methods handling network connections)
  • Always add/update the corresponding unit and/or functional tests. We won't accept non-tested pull requests.
  • Documentation and comments are more important than code: comment your code, use jsdoc for every new function, add or update markdown documentation if need be. We won't accept undocumented pull requests.
  • Similar to the previous rule: documentation is important, but also is code readability. Write self-describing code.

General rules and principles we'd like you to follow

Tools

For development only, we built a specific Docker Compose file: docker-compose.yml. You can use it to profile, debug, test a variable on the fly, add breakpoints and so on, thanks to chrome-devtools.
Check the logs at the start of Kuzzle using the development docker image to get the appropriate debug URL.

How to run the development stack (needs Docker 1.10+ and Docker Compose 1.8+):

# clone this repository
git clone git@github.com:kuzzleio/kuzzle.git
cd kuzzle

# Start a kuzzle cluster with development tools enabled
docker compose up

ENOSPC error

On some Linux environments, you may get ENOSPC errors from the filesystem watcher, because of limits set too low.

If that happens, simply raise the limits on the number of files that can be watched:

sudo sysctl -w fs.inotify.max_user_watches=524288

That configuration change will last until the next reboot.

To make it permanent, add the following line to your /etc/sysctl.conf file:

fs.inotify.max_user_watches=524288

You can now access the Kuzzle HTTP/WebSocket API through the following URL: http://localhost:7512. This is the entrypoint for the loadbalancer: API requests are then forwarded to kuzzle individual kuzzle nodes (round-robin).

For development purposes, nodes can be accessed individually:

Node no. HTTP/WebSocket port MQTT port Chrome Inspect Port
1 17510 1883 9229
2 17511 11883 9230
3 17512 11884 9231

Everytime a modification is detected in the source files, the nodes are automatically restarted.

Kuzzle over SSL

The development stack include a endpoint to access Kuzzle API through SSL on port 7443.

The certificates are privately signed, using provided CA certificate.
Domains accepted:

  • localhost
  • *.kuzzle.loc

You'll need to import the CA certificate to your browser and possibly your system local authorities to make it verified. Once done, your browser should not complain when reaching https://localhost:7443.
The CA certificate is here: docker/nginx/kuzzleCA.crt

Using node.js, for instance when using the sdk, you'll need to pass the CA cert using the NODE_EXTRA_CA_CERTS environment variable:

NODE_EXTRA_CA_CERTS=/path/to/certificate/kuzzleCA.crt wscat -c wss://localhost:7443

Create a plugin

See our plugins documentation

About Mac M1

First of all make sure that you have at least 4GB of ram allocated to your vm docker desktop and that it is running.

Run the following command to install all the dependencies in your container:

docker compose run kuzzle_node_1 npm ci

Finally, run the command docker compose up to start your Kuzzle stack.

Launching tests suits

Unit tests

npm run test:unit

Functional tests

KUZZLE_FUNCTIONAL_TESTS="test:functional:websocket" NODE_VERSION="20" ./.ci/scripts/run-test-cluster.sh