Twinned from https://github.com/cfsh/ninefold/issues/649 by tools/beads/import_github.py's reconcile pass.
Suggestions from ninefold-reviewer's review of #638. Advisory — #638 is CONVERGED and approved; this does not block it. #642 and #644 were the earlier rounds and are closed, both folded into the PR.
--board N changes which issue is read, but nothing downstream is told
main() builds args.board and reads that issue — and the empty-body refusal correctly says f"issue #{args.board} came back with an empty body". Every other place that names the board uses the module constant instead:
render() — print(f" board #{BOARD_ISSUE} — ..."), the digest's second line;as_json() — "board": BOARD_ISSUE;audit()'s UNREGISTERED detail — f"Add its row to {BOARD_URL}, or retire it.", where BOARD_URL is built from BOARD_ISSUE at import.So ./tools/office/audit_sessions.py --board 999 audits #999 and reports board #398 — N rows, emits "board": 398 in the JSON, and tells the reader to add rows to #398. The flag is documented in the module docstring, in --help, and in tools/office/README.md's usage block, so the invocation is expected rather than hypothetical.
Latent today: the default is 398 and nothing on the box passes another value. The shape is the one the tool otherwise refuses — a confident report that names the wrong thing — and main()'s refusal path already gets it right, which is what suggests an oversight rather than a decision.
Threading the resolved board through audit()/render()/as_json() (or dropping the flag) closes it. Either way a test that runs main(["--board", "999"]) and asserts the digest says #999 is the pin.
None.
No comments.