PR #716 follow-ups: the legacy-path comment names a file #716 never writes nf-58y ← Beads

open priority 2 task unassigned

Twinned from https://github.com/cfsh/ninefold/issues/730 by tools/beads/import_github.py's reconcile pass.

Suggestions from ninefold-reviewer's review of #716. Unclaimed, and it does not block that PR — it converged.

One line, and it is the tenth site of the class this thread has now fixed at nine (F2 the docstring's "Where the baseline lives"; F3 notes/machines.md, notes/merging.md, render_report's ⚠ NAME THE AGE, the wipeout arm's ⚠ UNWRAPPED; F5 the docstring's opening paragraph; F6 the three filed as #729 and fixed in a957ca7).

The site. tools/pr/probe_gate.py:133, the closing sentence of the LEGACY_BASELINE_PATH comment:

> Nothing is migrated or deleted: a Windows box that starts merging again writes main-windows.json beside it.

baseline_path("windows") returns probe-baseline/main-windows-<host>.json. main-windows.json is the pre-F1 spelling — the per-platform path the PR started with, before the host key was added. The code is right and only the comment is wrong: the module header (:11, :48), baseline_path's own docstring and read_baseline_blob's all say main-<platform>-<host>.json, and test_each_platform_gets_its_own_path asserts "main-windows-" with the trailing hyphen precisely so the host segment cannot be dropped.

⚠️ Where it sits is what makes it worth an issue. It is the last line before LEGACY_BASELINE_PATH = "probe-baseline/main.json", so a reader hits it exactly when deciding what the legacy file's successor is called — and the wrong name differs from the legacy one by a *platform* suffix alone. That reads as "the baselines are per platform", which is the belief #716 F1 exists to correct.

The same sentence is in #716's body ("Four decisions worth your eye", second ⚠). That half is raised on the PR as a B2 item, because a body edit moves no sha.

Deliberately NOT in scope here. Two nearby lines still say "per platform" and are correct as written — named so a sweep for this class does not churn on them:

Why not on the PR. It is a comment fix, the behaviour is correct, and pushing it moves the sha — which buys another full Tier-0 pass (#678 measures what that costs) on a PR that is governance-perimeter and already waiting on @aedanpope. Fold it into whatever next touches this file.

Dependencies

None.

Comments

No comments.

Add a comment