Commit graph

2 commits

Author SHA1 Message Date
Erik
c1a84cbe0c fix(D.2b): wield is ContainerId-based — drop the WielderId write (code review)
Code-quality review (needs-changes) on Task 3: WieldItemOptimistic wrote
item.WielderId directly, but RollbackMove (via MoveItem) never cleared it
→ a wield-rollback left a pack item with a stale WielderId=player.

Root-cause fix (vs the reviewer's snapshot-WielderId suggestion): acdream's
existing WieldObject 0x0023 confirm models a wielded item as ContainerId=
wielder + equip=mask and does NOT touch WielderId. So the optimistic path
must match — drop the WielderId write entirely. Optimistic state now equals
the confirmed state, rollback fully restores through MoveItem alone, and the
stale-state class is structurally eliminated. The paperdoll's
(WielderId==p || ContainerId==p) filter still matches optimistic wields via
ContainerId (login-equipped items match via WielderId from their CreateObject).

Also: + the wield+move outstanding-count combo test (spec §8), MoveItem doc
note that it doesn't manage WielderId, and the test pins WielderId==0 post-
optimistic-wield to document the model. Core 74/74 green.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-06-22 22:15:18 +02:00
Erik
cf472bd8db docs(D.2b): paperdoll Slice 1 implementation plan
Six TDD tasks: (1) de-risk probe (equip slots resolve to UiItemList) →
(2) correct EquipMask to canonical retail INVENTORY_LOC + numeric-pin
test → (3) WieldItemOptimistic + equip-aware rollback snapshot →
(4) ConfirmMove on the WieldObject 0x0023 echo → (5) PaperdollController
+ tests + divergence rows → (6) GameWindow wiring + full-suite gate.
Then the user visual gate. Full no-placeholder code; all signatures
verified against source.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-06-22 21:50:04 +02:00