test(M00): diagnose waterfall tile ownership
Work-Package: M00-QAR-TILE-OWNERSHIP-001 Agent: sindo-main-codex Tests: strict five-point terrain probe, coordination and documentation gates Fidelity: tile 30_49 ownership is healthy; exact ray miss classified as triangle seam
This commit is contained in:
@@ -0,0 +1,78 @@
|
||||
# M00-QAR-TILE-OWNERSHIP-001 — Waterfall terrain tile ownership
|
||||
|
||||
<!-- OPENWC_CLAIM:M00-QAR-TILE-OWNERSHIP-001:sindo-main-codex:2026-07-13 -->
|
||||
|
||||
## Ownership
|
||||
|
||||
- Target: M00
|
||||
- Program: QAR
|
||||
- Owner/Agent ID: sindo-main-codex
|
||||
- Branch: `work/sindo-main-codex/m00-waterfall-tile-ownership`
|
||||
- Lease expires UTC: 2026-07-13
|
||||
- Integrator: milestone integrator
|
||||
|
||||
## Outcome
|
||||
|
||||
Identify the runtime transition that prevents waterfall tile `30_49` from exposing a terrain mesh to the height probe.
|
||||
|
||||
## Non-goals
|
||||
|
||||
- Changing streaming budgets or terrain behavior.
|
||||
- Rebuilding caches or modifying extracted assets.
|
||||
- Adding a production terrain query service.
|
||||
|
||||
## Paths
|
||||
|
||||
- Exclusive: terrain probe runtime ownership diagnostics
|
||||
- Shared/hotspots: renderer baseline documentation
|
||||
- Generated/ignored: local probe reports and caches
|
||||
|
||||
## Contracts and data
|
||||
|
||||
- Public API/events: additive diagnostic JSON fields
|
||||
- Schema/format version: terrain report remains schema 1
|
||||
- Migration/compatibility: additive fields only
|
||||
- Consumers: M00 fidelity diagnosis
|
||||
|
||||
## Dependencies
|
||||
|
||||
- Requires: merged terrain-height probe and local cache inventory
|
||||
- Blocks: waterfall terrain ownership classification
|
||||
- External state: local extracted/cache data
|
||||
|
||||
## Verification
|
||||
|
||||
- Commands: isolated waterfall probe, coordination and documentation gates
|
||||
- Fixtures: checkpoint tile `30_49`
|
||||
- Fidelity evidence: runtime state correlated with build 12340 waterfall viewpoint
|
||||
- Performance budget: offline diagnostic
|
||||
|
||||
## Documentation deliverables
|
||||
|
||||
- Inline public API docs: diagnostic output fields
|
||||
- Module specification: verification/source map if behavior changes
|
||||
- Data-flow diagram: update ownership transition if needed
|
||||
- Sequence/state/dependency diagrams: document observed tile transition
|
||||
- Source map/status updates: baseline findings
|
||||
|
||||
## Simplicity and naming
|
||||
|
||||
- Important names introduced: `available`, `queued_index`, `loading`, `state_present`, `mesh_source`
|
||||
- Simplest considered solution: inspect existing loader registries read-only
|
||||
- Rejected complexity/abstractions: new tracing framework
|
||||
- Unavoidable complexity and justification: none
|
||||
- Measured optimization evidence: not applicable
|
||||
|
||||
## Status
|
||||
|
||||
- State: ready
|
||||
- Done: raw/cache inventory, isolated runtime state, mesh AABB/local probe and nearby sampling diagnosis
|
||||
- Next: integrator review; placement/composition remains the actual paired-camera gap
|
||||
- Blocked by:
|
||||
|
||||
## Handoff
|
||||
|
||||
- Commit: branch HEAD
|
||||
- Results: tile 30_49 ownership and meshes are healthy; exact ray misses a triangle seam/edge, while a 2-unit offset samples terrain at 113.872
|
||||
- Remaining risks: nearby estimate is diagnostic and must not become a gameplay terrain-query contract
|
||||
- Documentation updated: `docs/RENDER_BASELINE.md`, `docs/modules/world-renderer.md`
|
||||
Reference in New Issue
Block a user