git-spice: format

[GaetanLepage]

Things done

cc @vinnymeller

• Built on platform:
  • x86_64-linux
  • aarch64-linux
  • x86_64-darwin
  • aarch64-darwin
• Tested, as applicable:
  • NixOS tests in nixos/tests.
  • Package tests at passthru.tests.
  • Tests in lib/tests or pkgs/test for functions and "core" functionality.
• Ran nixpkgs-review on this PR. See nixpkgs-review usage.
• Tested basic functionality of all binary files, usually in ./result/bin/.
• Nixpkgs Release Notes
  • Package update: when the change is major or breaking.
• NixOS Release Notes
  • Module addition: when adding a new NixOS module.
  • Module update: when the change is significant.
• Fits CONTRIBUTING.md, pkgs/README.md, maintainers/README.md and other READMEs.

---

Add a 👍 reaction to pull requests you find important.

[emilazy]

@wolfgangwalther I fear we are going to break the branch many times over the next few weeks from stale CI results. Do you know how difficult it would be to get a merge queue going just for formatting/parse checks at first? 🙏

[jfly]

@wolfgangwalther I fear we are going to break the branch many times over the next few weeks from stale CI results

Is there a central place where this issue is being discussed?

[emilazy]

Not really. This PR came up on Matrix. The issue is that successful formatting checks still allow merges into the branch even after the formatting check has switched to use the new version, so you can merge “green” PRs that break the branch. A merge queue would fix that.

[emilazy]

re:

Unless there's some way to "expire" all these known stale checks?

I suppose we could change the name of the required job again, at the expense of annoying people a bunch. Changes that touch existing code in a way that could cause issues will become conflicted, so it’s only new code that is particularly problematic.

[jfly]

I suppose we could change the name of the required job again, at the expense of annoying people a bunch.

Neat trick! Wouldn't a merge queue also annoy a bunch of people in this scenario? That is, the annoying thing we did is changing the formatter. Merge queue/living with stale checks/proactively "expiring" stale checks are all just various ways of shifting that pain around.

[emilazy]

The difference is that a merge queue will only reject PRs that are actually problematic; changing the required checks will reject everything. (Which also means that breaking the branch is more likely, because people will feel compelled to override it more.)

[wolfgangwalther]

I fear we are going to break the branch many times over the next few weeks from stale CI results. Do you know how difficult it would be to get a merge queue going just for formatting/parse checks at first? 🙏

Here we go: #431146