Worldgen Phase 2: preserve ocean, coast, and river priority rules

[MichaelFisher1997]

Parent: #646

Phase

Phase 2: Multi-parameter biome selection

Goal

Ensure the richer selector does not regress hard structural rules for ocean, beaches/coasts, rivers, and frozen rivers.

Tasks

• Keep ocean decision driven primarily by continentalness and sea-level structure.
• Keep beach/coastal selection narrow and slope/height aware.
• Preserve river/frozen river override behavior or replace it with an equivalent explicit priority layer.
• Add tests around threshold boundaries.

Acceptance Criteria

• Ocean biomes do not leak inland above sea level.
• Inland water does not become sand beach by default.
• River/frozen river behavior remains deterministic and climate-aware.
• Threshold tests cover edge cases around ocean/coast/river boundaries.

Dependencies

• Requires multi-parameter selector shape.

Parallelization

Can run in parallel with biome migration and LOD consistency after selector shape is in place.

[github-actions]

This is a Feature Request (Phase 2 of a biome selection enhancement). The issue is well-scoped with clear acceptance criteria.

Analysis:

1. Current ocean/coast/river priority rules exist but are scattered:

   • Ocean: Hard decision in height_sampler.zig:227 (continentalness < 0.35) - bypasses biome selection entirely
   • Beach: BIOME_POINTS continentalness band 0.35-0.42 + slope ≤ 2 in biome_registry.zig:188-190
   • River: Override in biome_selector.zig:52-66 (river_mask > 0.5, height < 120)
   • Frozen river: Temperature ≤ 0.20 in biome_selector.zig:63

2. "Multi-parameter selector shape" dependency is undefined - No such component found in the codebase. Parent issue Worldgen: biome terrain modifiers and multi-parameter biome selection #646 context is not available in code. This needs clarification from the author.

3. Missing test coverage for boundary cases:

   • No tests at exact threshold transitions (e.g., continentalness = 0.35)
   • No integration tests for the full selection pipeline

4. Potential risk: If a richer selector is introduced without explicit priority layering, ocean biomes could leak inland (acceptance criterion Roadmapv3 #1) if the hard HeightSampler decision is removed or bypassed.

Suggestion: Before implementation, clarify what "multi-parameter selector shape" means in #646. The existing priority rules are already well-implemented via BIOME_POINTS continentalness ranges and the HeightSampler ocean gate - Phase 2 may only need additive multi-parameter scoring without modifying these hard constraints.

opencode session  |  github run