Skip to content

#920 follow-ups: release test should call each installer's own checks #921

Open

Description

@scuttletoby

Follow-ups from the two review lenses on #920 (merge-path: SAFE, state: SAFE). None block.

  1. .github/scripts/fork-ci-release.test.py:430-440: the heavy, model-guard, gh-apps and pc-agents checks repeat each installer's missing-file loop instead of calling it. Give each installer a small missing_sources helper (as desk has) and call that, so the test cannot drift from the installer.
  2. .github/scripts/fork-ci-release.test.py:403: the test runs only each installer's missing-file check, not its self-test/staging step. A copied script that starts loading a new file next to it would pass the test and fail the real install. Run desk's staging step and heavy's copy step on the release folder (both work with the test prefix).
  3. .github/scripts/fork-ci-release.test.py:475-476: a kit is only noticed when it has an sbin/ folder, and nothing checks each bootstrap's --source against the test's rows.
  4. .github/scripts/fork-ci-release.test.py:500: every failure says "add the path(s) to fork-ci-release.json paths", even when the cause is a missing test row or a check that did not run. Say which.

Each fix needs a test that fails without it.

🤖 Generated with Claude Code

Activity

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions