act VS combine-prs-workflow

When it comes to the alternatives to tmate, there is another great debugging tool that you could check out. It is called act and it allows you to run GitHub Actions code on your local machine making debugging even easier. It has its own limitations and some learning curve but overall it is another tool you should use if you can’t fix the CI bugs by connecting directly into the running action with the tmate.

Link: https://github.com/nektos/act

Could you upload your build of GitHub's runner image to Docker Hub?

This would be quite useful for users of other GitHub Actions clones like act [0].

[0]: https://github.com/nektos/act

> These kinds of commit messages are typically an indicator of a broken process where somebody needs to commit to see something happen, like a deployment or build process, and aren't able to assert that stuff works locally.

This is one of my biggest pet peeves with services like github actions. Something running locally like "act" [1] isn't sufficient because it doesn't have everything github has and is extra friction anyway to get everyone to use it for testing.

[1] https://github.com/nektos/act

View on GitHub

If you use Github actions, act is incredibly useful. It can be used to test your GH actions, but also serves as an interface for running tasks locally.

That's something that often is difficult or basically impossible. Except for maybe GitHub actions through Act (https://github.com/nektos/act). I'd still lean to something in the yaml sphere if it eventually would be used in deployment pipelines and such. For example a solution incorporating ansible.

It also seems to me that the argument you make is mostly focused on the building step? Earthly certainly seems focused on that aspect.

I feel I'm being trolled, but I'll bite and accept the resulting downvotes

I don't think treating every mention of act as an opportunity for airing of personal grievances is helpful in a discussion when there's already ample reports of people's concrete issues with it, had one looked at the 800 issues in its repo https://github.com/nektos/act/issues?q=is%3Aissue or the 239 from gitea's for https://gitea.com/gitea/act_runner/issues or whatever is going on with Forgejo's fork https://code.forgejo.org/forgejo/act .

But, as for me specifically, there are two and a half answers: I wanted to run VSCodium's build locally, which act for sure puked about. Then, while trying to troubleshoot that, I thought I'd try something simpler and have it run the lint job from act's own repo <https://github.com/nektos/act/blob/1252e551b8672b1e16dc8835d...> to rule out "you're holding it wrong" type junk. It died with

  [checks/lint]     Failure - Main actions/setup-go@v3

https://twitter.com/mitsuhiko/status/1720410479141487099 :

> GitHub Actions currently charges $0.16 per minute* for the macOS M1 Runners. That comes out to $84,096 for 1 machine year*

GitHub Runner is written in Go; it fetches tasks from GitHub Actions and posts the results back to the Pull Request that spawned the build.

nektos/act is how Gitea Actions builds GitHub Actions workflow YAML build definition documents. https://github.com/nektos/act

https://twitter.com/MatthewCroughan/status/17200423527675700... :

> This is the macOS Ventura installer running in 30 VMs, in 30 #nix derivations at once. It gets the installer from Apple, automates the installation using Tesseract OCR and TCL Expect scripts. This is to test the repeatability. A single function call `makeDarwinImage`.

With a Multi-Stage Dockerfile/Containerfild, you can have a dev environment like xcode or gcc+make in the first stage that builds the package, and then the second stage the package is installed and tested, and then the package is signed and published to a package repo / app store / OCI container image repository.

SLSA now specifies builders for signing things correctly in CI builds with keys in RAM on the build workers.

"Build your own SLSA 3+ provenance builder on GitHub Actions" https://slsa.dev/blog/2023/08/bring-your-own-builder-github

To address inefficiency caused by separate PRs, a workflow was designed to join them automatically into one big PR. However, it was unable to deal with lockfile conflicts. PRs that caused conflict in the Combine PRs job, were omitted and you had to add them manually anyway. It spared some time, but the developer experience was still far from being perfect.

Another pitfall I ran into recently with a workflow I've been working on [1]: Checks and CI that are made with GitHub Actions are reported to the new Checks API, while some (all?) external services report to their old Statuses API. This makes it needlessly difficult to ascertain whether a PR/branch is "green" or not. They finally decided to create a "statusRollUp" that combines the state of the two APIs, but it's not available in their REST api, only their GraphQL API.

[1] https://github.com/hrvey/combine-prs-workflow/

GitHub Actions can do some neat things. I got tired of waiting for Dependabot (tool that makes automatic PRs to update your middleware, acquired by GitHub) to add an option to group PRs together (it opens a separate PR for each dependency that can be updated, so merging and re-running CI can take a long time) so I scratched my own itch and made a workflow that merges their PRs together: https://github.com/hrvey/combine-prs-workflow Been running it for a year now, and still pretty happy with it.