cache VS upload-artifact

Since we launched this last year, GitHub released a v2 of their internal cache API [0], based on Twirp [1] of all things, so we adapted to that. Interestingly that Twirp service also receives Actions artifacts, though we have not intercepted those today given that you likely still want them to appear in the GitHub UI / be accessible from the GitHub API.

[0] https://github.com/actions/cache/discussions/1510

[1] https://github.com/twitchtv/twirp

Use the official actions/cache to implement the cache process.

Since GitHub offers a cache storage back-end supported by Docker, we use the gha cache type so that the cache storage is as close to our runners as possible. The mode=max option instructs Docker to cache even the intermediate build layers, not only those exported to the final image. And we also give our build image some (arbitrary) name.

We can use another amazing action provided by GitHub, actions/cache to cache our dependencies from our first job run and use that cache for subsequent jobs.

> with no persistent storage

There's https://github.com/actions/cache though?

Use Cache Actions: GitHub Actions provides cache actions that simplify caching implementation. The @actions/cache JavaScript library is a popular choice for managing caching in workflows. It offers flexible options for storing and retrieving cache artifacts based on keys, scopes, and paths.

GitHub maintain a set of repos called actions. One of which is called cache.

By utilizing the actions/cache action action, we implemented a strategy to store and retrieve dependencies, preventing redundant installations.

I do think it's good practice to enable caching, such that your script doesn't hit RubyGems / pip / npm / etc every time it runs.

That way at least the automation activity stays entirely within the GitHub / Azure network.

It looks like you can do that for Ruby by adding this:

https://github.com/actions/cache/blob/master/examples.md#rub...

    - uses: ruby/setup-ruby@v1

I have a small gripe that I think exemplifies a bigger problem. actions/upload-artifact strips executable permissions from binaries (1). The fact they fucked this up in the first place, and six years later haven't fixed it, gives me zero confidence in the team managing their platform. And when I'm picking a CI/CD service, I want reliability and correctness. GH has neither.

When it takes all of a day to self host your own task runner on a laptop in your office and have better uptime, lower cost, better performance, and more correct implementations, you have to ask why anyone chooses GHA. I guess the hello-world is convincing enough for some people.

(1) https://github.com/actions/upload-artifact/issues/38

Github's response was pretty bad. They changed the upload-artifacts workflow to no longer include hidden files _in a minor release_ and that broke thousands of workflows that required hidden files.

Github issue with angry comments: https://github.com/actions/upload-artifact/issues/602

upload-artifact

Here, the concept of "upload" is a little confusing. What the GitHub action upload-artifact does is to "copy" the jar file to a publicly accessible folder.

The content of the build destination folder folder needs is saved and transferred to the following jobs in the workflow. We do this with the Github actions actions/upload-artifact

The action for uploading the artifact doesn't do too much, but it takes care of all the nuance around GitHub Pages artifacts specifically. You can view the action's source here. It will tar the path (provided by the with option) and then call the upload artifact action. The artifact's name is github-pages and has a 1 day expiration. This artifact has the name and format required for the deploy action. It all just works (so far).