Instructions for running on Cloud Functions for Firebase

Similar to https://github.com/probot/probot/issues/149 this framework seems like a great fit for a serverless workload. In fact the code block in the main README looks very similar to the hello world example for Cloud Functions:

Probot:

module.exports = robot => {
  robot.on('issues.opened', async context => {
    const issueComment = context.issue({ body: 'Thanks for opening this issue!' })
    return context.github.issues.createComment(issueComment)
  })
}

Cloud Functions HTTP Example:

exports.myRobotFunction = functions.https.onRequest((req, res) => {
  // ...
});

Would you be open to my adding documentation on how to deploy a Probot on Cloud Functions for Firebase in this repository? I assume I’ll have to get around the async/await issues mentioned in #149 but that’s likely not an impossible hurdle.

cc @abehaskins @codingdoug

[samtstern]

@JasonEtco thanks for the update, that proposed integration looks perfect! I’ll comment if I see anything/

[tcbyrd]

@JasonEtco What I’m hoping we can accomplish with any cloud SDKs is to use inherent knowledge of what’s available in the execution environment to avoid bringing in any dependencies that are already available.

A few examples:

• A Firebase developer will deploy functions from a Firebase project, which means they probably already have the firebase package as a dependency.
• A Lambda function’s execution environment includes the necessary libraries for connecting to other AWS services.
• Azure functions have bindings available to various other Azure services.

The way I see it working, deploying a Probot serverless function is basically a way to add the GitHub APIs to those platforms, so any Probot function you deploy would be able to talk to GitHub as easily as it talks to other native platform services.