Description
Found by loop52 lane l1683: just gen reorders about 15 unrelated failure-mode rows in BLUEPRINT-SCHEMA.md / fielddocs.json between runs because failure modes are sorted unstably. just gen-check can then flake or produce noisy diffs.
Acceptance Criteria
- #1 Two consecutive just gen runs produce byte-identical BLUEPRINT-SCHEMA.md and fielddocs.json
- #2 A test with ties in the sort key is red at base and green after
Definition of Done
- #1 just check (fmt-check, lint, gen-check, env-check, docs-check, test, race, hygiene, ui-check, compose-check, helm-test, lab-check, signal-fidelity)
- #2 just gen (only if a blueprint field, construct/workload config struct, or a skill under plugins/synthkit/skills/ changed)
- #3 just dump — inventory diffed against signals/
Implementation Plan
Compose the accepted total-order comparator candidate, review once, run integrated gate, land exact green tree.
Implementation Notes
Landed total-order comparator and permutation regression in 4b24a9f, pushed as 13e4258 after the exact clean-tree composed just check passed remotely. CodeRabbit complete: zero findings, all three paths reviewed. Real schema generator run twice with identical artifact hashes. No emitted signal or blueprint field behavior changed.
Final Summary
Deterministic failure-mode rows now sort by the complete dedup tuple (name, axis, help). The tied-row regression was red before the comparator change, green after; two generator runs and the full integrated gate passed.