test: stop four JSON tests taking their verdict from the checkout - #548

Fixes the four TestJsonFormat failures on the push run for #542, all reading skip where they assert pass.

Nothing regressed — the tests were measuring the repository they run in

Each of the four supplies a message (on stdin or in a file), which makes it a prospective commit. _resolve_current_author then does:

if context.stdin_text is not None or context.commit_file is not None:
    return get_git_config_value("user.name") or get_commit_info("an")

Both sources are ambient. A GitHub runner configures no git identity — no workflow in this repository sets one — so the fallback always decides, and #542 was a dependabot merge: HEAD's author was dependabot[bot], which cchk.toml lists in [commit] ignore_authors. Every commit check skipped, and overall status is skip when they all do.

Reproduced

A clone pinned to 9f12a63 with the global and system git config disabled, which is what the runner looks like:

HEAD author : dependabot[bot]
user.name   : ''        ->  status: skip     (the four failures, exactly)
user.name   : set       ->  status: pass     (why laptops and PR runs were green)

So this was never about #542's contents, and it will recur on the next bot-authored merge to main — dependabot, pre-commit-ci, copilot or coderabbitai.

The fix

The four tests take a new pinned_author fixture that fixes both identity sources, so the verdict comes from the message under test. Worth noting: the two tests in the class that kept passing are exactly the two that already patched get_commit_info for unrelated reasons — the same pin, arrived at by accident.

A fixture that pins this would also hide the fallback everywhere it applies, so the behaviour is now asserted directly rather than left ambient: a new test drives an unconfigured identity with a bot as HEAD's author and expects every check to skip, exit code still 0. What silently decided other tests' results is now a contract of its own.

Checks

In the reproduction clone: 4 failed before, 7 passed after. Full suite unchanged in both a configured and an unconfigured environment (581 passed; the one unrelated failure, test_load_config_file_permission_error, fails only because my sandbox runs as root, where chmod 000 does not deny access — it fails identically on unmodified main and passes in CI).

Not changed

Whether the fallback should apply to ignore_authors for a prospective message is a separate design question — the author of the previous commit is arguably irrelevant to a message not yet written. Changing it would alter real behaviour (a bot's PR checked with -m on a runner with no git identity would stop being exempt), so this PR only stops the tests depending on it and leaves the call to you.

Generated by Claude Code