Phase 3 / Task 2: Extract LOD and predictive-loading coordinator

[MichaelFisher1997]

Goal

Split render-distance and predictive-loading logic out of WorldStreamer.

Scope

• src/world/world_streamer.zig
• src/world/lod_manager.zig

Work

• Move PlayerMovement and priority weighting into a dedicated coordinator
• Isolate lod_manager interactions behind a small API
• Keep render-distance clamping behavior unchanged

Acceptance criteria

• LOD decisions are no longer embedded in chunk queue code
• Changing render distance still triggers the expected rescan behavior

Parallel-safe

Yes

[github-actions]

This is a Feature Request / Refactoring Task focused on architectural cleanup.

Analysis:

The issue requests extracting a coordinator to decouple LOD decisions from chunk queue code. Looking at the current implementation:

1. PlayerMovement (lines 74-130 in world_streamer.zig) is embedded directly in chunk prioritization via priorityWeight() calls at lines 291 and 325.

2. Render-distance clamping is duplicated in three places (updateStreaming() lines 275, 316, and processUnloads() line 420), all using the same pattern:

   const render_dist = if (self.lod_manager) |mgr| @min(self.render_distance, mgr.config.getRadii()[0]) else self.render_distance;

3. LOD interaction in updateStreaming() (lines 351-358) directly calls lod_mgr.update() with raw velocity calculations, tightly coupling the streaming logic to LOD internals.

Scope Assessment: Well-contained - only two files involved, and the acceptance criteria are clear and testable. The refactoring is additive (creating a new coordinator) rather than destructive, which reduces risk.

Implementation suggestion: A StreamingCoordinator struct that encapsulates:

• PlayerMovement state and priorityWeight() logic
• Centralized render-distance clamping via getEffectiveRenderDistance()
• Isolated LOD manager interface with a small API (e.g., isChunkCovered(), getLODLevel())

The WorldStreamer would then delegate priority calculations and LOD queries to this coordinator, satisfying the acceptance criteria that "LOD decisions are no longer embedded in chunk queue code."

No sub-issues needed - this is a single, well-defined refactoring task.

opencode session  |  github run