Cloud integration

[eliot-akira]

Features to integrate the pipeline with Cloud server.

• Upload versioned zip file to the cloud server
• Update the plugin on the cloud plugin page

Related:

• https://github.com/TangibleInc/example-plugin-with-cloud/issues/1
• Licensing feature updater#5

@ethelynmatias

[eliot-akira]

OK, I imagine this will take some effort, we can see how it goes. The current pipeline actions are performed by these two files.

• before-release.ts
• after-release.ts

They're run by the GitHub Actions workflow in the user plugin, like in the Example Plugin with Cloud.

• .github/workflows/release.yml

---

The new features can be added in afterRelease(). At the bottom of the function, you can see how it dispatches a POST request to log the Git commit or version tag. You can dispatch additional requests to the Cloud server.

We'll probably need to experiment a bit in a feature branch, to figure out how to achieve the two actions described (upload zip and update plugin page).

How about this?

• You can create a branch feature/cloud in this repo and start a pull request as a draft.
• Then the example plugin can be updated to run the pipeline from that branch. We can push test commits and version tags in the plugin as we develop the pipeline further.

Here's where the Pipeline module is downloaded in the GitHub Actions workflow.

• https://github.com/TangibleInc/example-plugin-with-cloud/blob/main/.github/workflows/release.yml#L19-L20

To clone the feature branch:

git clone --single-branch --branch feature/cloud https://github.com/tangibleinc/pipeline

Please feel free to make direct commits to the example plugin as needed.

[eliot-akira]

We'll need to get the Cloud ID of the plugin being deployed.

For this I suggest we define a new optional property in package.json of the user plugin.

{
  "tangible": {
    "cloud": "..ID here.."
  }
}

Having the ID in JSON format will make it easier for the pipeline to get it, instead of from a PHP file.

The pipeline script and its tests are written in TypeScript and runs on Bun. I'll need to add that info in the docs.

[eliot-akira]

To upload the plugin zip file, here's a couple of possible solutions.

• Post the Git event metadata with the zip file URL to the server, which downloads it. This is limited to public repos unless we set up the server to have Git/SSH access to private repos.
• Use a command like scp or rsync in the pipeline to upload it directly via SSH.
  • Deploying to a server via SSH and Rsync in a Github Action
  • Using secrets in GitHub Actions

There are pros/cons to either approach. The first is simpler for the pipeline and gives full control to the server, but maybe difficult to work with private repos. The second is how previous versions of the Bitbucket pipeline worked. It's a more common approach perhaps, but I feel it gives too much power to the pipeline.

Maybe a third option: the pipeline posts the Git event metadata, and the server responds with a temporary URL to upload the zip file, such as to an S3-compatible data storage. This keeps the setup simple(r) without involving SSH between the pipeline and the server. But maybe a bit complex to coordinate multi-step actions.

[ethelynmatias]

@eliot-akira and @ghen-arvesu I will be working on this tomorrow set up a pipeline same as the function on the Bitbucket.

[ethelynmatias]

@eliot-akira and @fintina-madalin added the plugin cloud id on the package.json under the example plugin.

Also edited the release.yml from this plugin to run the feature/cloud branch

[eliot-akira]

edited the release.yml from this plugin to run the feature/cloud branch

OK, so this branch is meant for us to develop the Cloud integration in the pipeline. We can work on it in a pull request until it's ready, then merge it to main branch for all plugins with Cloud ID.

The pipeline's release scripts are written in Node.js, as you see here:

• https://github.com/TangibleInc/pipeline/blob/main/after-release.ts

It depends how much of the new feature you'd be comfortable writing - if not, we can discuss the deploy strategy to coordinate with the Cloud server, and I'll implement the pipeline script.