M15 Gate 3: FOW Integration & Production Validation

[WolfCTOS]

Goal

Verify that the proven Dynamic FOW system works as an actual gameplay system and integrate the proven Gate 2 mechanism into the real M15 architecture.

Status

IN PROGRESS

Gate 1: PASS / CLOSED (#17)
Gate 2: PASS / CLOSED (#18)
Gate 3: NOT YET VERIFIED

Gate 2 closed the core causal loop:

FowModel VISIBLE → real coverage 1→0 → native sustained unmap → visual FOW → coverage 0→1 → VISIBLE-RETURN → normal rendering restored

The successful PoC-1c evidence is recorded in docs/milestones/M15_GATE2_HANDOFF.md and docs/milestones/M15_FOW_DISCOVERY.md §23.

Gate 3 Definition of Done

• Integrate the proven PoC mechanism into the real multi-cell M15 architecture.
• Remove/retire temporary PoC-1c runtime code after its mechanism is transferred and verified.
• Verify infantry, vehicles, and buildings as vision sources where applicable.
• Verify multiple simultaneous friendly vision sources and union coverage.
• Verify moving objects across visibility boundaries.
• Verify creation and destruction/loss of vision sources.
• Verify dynamic FOW over multiple cells, not only the PoC single-cell case.
• Verify previously explored terrain remains FOGGED rather than becoming SHROUDED.
• Verify enemy units/buildings do not remain visible outside current vision where required by the M15 DoD.
• Verify Lua visibility queries against the real M15 model.
• Verify CnCNet / multiplayer behavior and determinism.
• Check performance during normal gameplay.
• Check save/load if technically relevant.
• Check important special rendering cases already identified by M15 research.
• Document verified behavior and known limitations.

Constraints

• Do not reopen Gate 1 or Gate 2 without fresh contradictory evidence.
• Do not repeat completed native reverse engineering unless new evidence requires it.
• Do not reintroduce retired veteran-rendering, occupancy-watcher, or one-shot-unmap investigations without new evidence.
• Do not create a second fake battlefield state.
• Prefer the existing FowModel as truth and the native engine as the visual actuator.
• Keep the implementation KISS and evidence-driven.

Gate 3 Exit Criteria

Gate 3 result must be one of: VERIFIED / FAILED / INCONCLUSIVE.

Gate 3 is complete only when the above gameplay, integration, multiplayer/determinism, performance, and limitation evidence is documented.