neon-sprawl/docs/reviews/2026-05-23-NEO-54.md

59 lines
5.3 KiB
Markdown
Raw Permalink Blame History

This file contains ambiguous Unicode characters!

This file contains ambiguous Unicode characters that may be confused with others in your current locale. If your use case is intentional and legitimate, you can safely ignore this warning. Use the Escape button to highlight these characters.

# Code review — NEO-54 player inventory store + stack/slot rules
**Date:** 2026-05-23
**Scope:** Branch `NEO-54-player-inventory-store-stackslot-rules` · commits `e5ffa1b``3724a54` vs `origin/main`
**Base:** `origin/main`
## Verdict
**Approve with nits**
## Summary
NEO-54 delivers the planned split between **`IPlayerInventoryStore`** (in-memory + Postgres sparse rows, `V005` migration) and **`PlayerInventoryOperations`** (catalog-routed add/remove, all-or-nothing adds, stable deny reason codes). The engine mirrors NEO-38s store + operations pattern, chains DI from **`AddItemCatalog`**, and tests cover merge/stack max, bag full, equipment single slot, remove drain order, deny immutability, and Postgres persistence across a fresh factory. Docs (plan, E3.M3 slice, alignment table, README) match the implementation. Risk is moderate-low for merge: no HTTP yet, but persistence and rule correctness are well exercised; the main follow-ups are test hygiene (AAA), optional coverage gaps, and register footnote parity.
## Documentation checked
| Document | Result |
|----------|--------|
| [`docs/plans/NEO-54-implementation-plan.md`](../plans/NEO-54-implementation-plan.md) | **Matches** — 24 bag + 1 equipment, split store/engine, reason codes, `V005`, Postgres/in-memory parity tests, DI via catalog extensions, docs/README updates; acceptance checklist complete. |
| [`docs/decomposition/modules/E3_M3_ItemizationAndInventorySchema.md`](../decomposition/modules/E3_M3_ItemizationAndInventorySchema.md) | **Matches** — NEO-54 bullet under **Related implementation slices**; frozen roster/stack limits align with tests. |
| [`docs/decomposition/modules/documentation_and_implementation_alignment.md`](../decomposition/modules/documentation_and_implementation_alignment.md) | **Matches** — E3.M3 row notes NEO-54 landed; NEO-55 HTTP still planned. |
| [`docs/decomposition/modules/module_dependency_register.md`](../decomposition/modules/module_dependency_register.md) | **Matches** — E3.M3 **In Progress**; footnote lists NEO-50NEO-54. |
| [`docs/decomposition/modules/client_server_authority.md`](../decomposition/modules/client_server_authority.md) | **Matches** — server-authoritative inventory mutations; no client-trusted placement in this slice. |
| [`server/README.md`](../../server/README.md) | **Matches** — Player inventory store section: containers, reason codes, persistence note, NEO-55 deferral. |
Register/tracking: E3.M3 alignment row is updated appropriately; module **Status** remains **In Progress** until NEO-55+.
## Blocking issues
None.
## Suggestions
1. ~~**AAA in Postgres deny test** — In `PlayerInventoryPersistenceIntegrationTests.TryAddStack_WhenDenied_ShouldNotPersistPartialState`, move `Assert.Equal(PlayerInventorySnapshot.BagSlotCount, OccupiedBagSlotCount(afterDeny!))` from the first **Act** block into **Assert** (or a dedicated assert phase after both act steps), per [csharp-style](../../.cursor/rules/csharp-style.md#unit-and-integration-tests-arrange-act-assert).~~ **Done.** — removed in-Act assertion; slot count verified only in **Assert** via fresh factory read-back.
2. ~~**`StoreMissing` coverage (optional)** — `PlayerInventoryMutationKind.StoreMissing` is documented on `PlayerInventoryMutationOutcome` but untested. A small operations test for an unknown `playerId` (in-memory store without that bucket) would lock the contract before NEO-55 HTTP maps it.~~ **Done.**`TryAddStack_ForUnknownPlayer_ShouldReturnStoreMissing` in `PlayerInventoryOperationsTests.cs`.
3. ~~**Multi-slot add (optional)** — Add a test that places quantity **> `stackMax`** across **multiple new** empty slots (e.g. 41 × `field_stim_mk0` on empty bag → three occupied slots). Merge and bag-full paths are covered; this would explicitly guard the `FillEmptySlots` loop beyond single-slot adds.~~ **Done.**`TryAddStack_WhenQuantityExceedsStackMax_ShouldSpanMultipleEmptySlots` asserts 20+20+1 across three slots.
4. ~~**Dependency register footnote (optional)** — Extend the **E3.M3 note** in [`module_dependency_register.md`](../decomposition/modules/module_dependency_register.md) to mention **NEO-54** alongside NEO-50NEO-53 (same pattern as NEO-52/NEO-53 reviews).~~ **Done.**
## Nits
- ~~Nit: **`PlayerInventoryMutationOutcome.cs`** — enum and record parameters use **2-space** indent; neighboring outcome types (e.g. `SkillProgressionGrantApplyOutcome.cs`) use **4-space**. Format-only tidy.~~ **Done.**
- Nit: **`PostgresPlayerInventoryStore.TryReplaceSnapshot`** — delete-all + insert per mutation is correct for snapshot replace; concurrent writers are last-write-wins (same class of risk as other NEO-29-style stores). Fine for prototype; NEO-55 HTTP may want documenting or serializing per player if load tests show races.
- Nit: **`ContainerFromSlotKind`** — any non-`equipment` catalog value routes to **bag**. Safe for the frozen six-item catalog; if new `inventorySlotKind` values appear, consider an explicit deny rather than silent bag routing.
## Verification
```bash
cd server
dotnet test NeonSprawl.Server.Tests/NeonSprawl.Server.Tests.csproj --filter "FullyQualifiedName~PlayerInventory"
dotnet test NeonSprawl.Server.Tests/NeonSprawl.Server.Tests.csproj
```
With Postgres configured (`ConnectionStrings__NeonSprawl`), confirm `[RequirePostgresFact]` persistence tests run (not skipped). No Bruno/manual QA for this story (no HTTP).