Twinned from https://github.com/cfsh/ninefold/issues/499 by tools/beads/import_github.py's reconcile pass.
Suggestions from ninefold-reviewer's review of #492.
All three are below the review gate's severity floor — #492 is CONVERGED and none of these blocked it. Filed so they are captured rather than carried on the thread.
1. Unused import in the F1 regression test.
tools/tests/test_rebuild_f5.py:233 — import unittest.mock as mock inside test_the_git_dir_probe_is_absolute_and_survives_a_foreign_cwd. The test uses os.chdir; mock is never referenced. Same class as the item already tracked in #479.
2. The harness scenario count disagrees with itself.
tools/pr/rebuild_f5.py:28 says the port was pinned "against a 13-scenario characterization harness"; tools/tests/test_rebuild_f5.py:6 and #492's body both say 14. fast_forward_solo was added in the fix commit, so 14 looks like the final number and the module docstring is the stale one. It matters slightly more than a normal doc nit because the harness is deleted — that docstring is now the only surviving description of how the port was pinned.
3. The shim's set line drifted from its siblings.
tools/pr/rebuild_f5.sh:9 is set -Eeuo pipefail. merge_pr.sh, probe_gate.sh, thread.sh, dispatch.sh and app_token.sh are all set -euo pipefail. -E propagates ERR traps and a two-line exec shim installs none, so it is inert — but these shims are meant to be identical boilerplate, and a difference invites the next reader to hunt for a reason that isn't there.
Not filed here: the Gh.pr_list fail-open question #492's body raises is already tracked as #493.
None.
No comments.