Record the Cut C follow-up and close the provenance schema cut
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
This commit is contained in:
parent
5e9634efcd
commit
46efccbec8
1 changed files with 10 additions and 0 deletions
|
|
@ -71,6 +71,16 @@ Progress (Self updates this in each cut's landing commit):
|
|||
It diverges from `ItemManager.Brand` when two factions sell one design.
|
||||
- `Brand`'s tie-break by record key has no test (`OrderBy` → `OrderByDescending`
|
||||
survives).
|
||||
- **Fixed** at `fc1c0526` and `5e9634ef`:
|
||||
- Census lists every maker selling a design; its output is byte-identical today.
|
||||
- A same-maker tie-break test was added.
|
||||
- 42 of 42 tests pass. `tests/mutation_tests.py` holds 19 entries: the control stays
|
||||
green and all 18 mutants are killed.
|
||||
- **Count correction:** two earlier reports counted the control as a kill. At `8fe09a23`
|
||||
the script held control + 16 mutants, not 17, and at `b2036fb8` control + 17, not 18.
|
||||
No mutant verdict changes.
|
||||
- **The schema cut is complete.** Remaining: the operator's play smoke, and merging
|
||||
`codex/item-provenance` together with `codex/cultcache-cutover`.
|
||||
- 115 records rewritten: 12,686,267 → 12,683,061 bytes, 0 ignored slots.
|
||||
- The `AetherDb` outputs, including census, are byte-identical: no design's maker
|
||||
disagreed with its products' maker.
|
||||
|
|
|
|||
Loading…
Reference in a new issue