Claude
b6529719dd
The far sheet ring was already gone from the runtime; this removes what was left of it: - UVoxelSettings::bFarSheetRing / FarSheetSpanLevels, the voxel.TestFarSheetRing switch, and their log/trace/dump fields. - The sheet branch of VoxelClipmapDesiredTiles (FParameters sheet fields, OuterShell sheet level, sheet ring emission). Output for the runtime configuration (sheet off) is unchanged: same rings, same alignment. - Tests: VerticalStrateReach now asserts the sheet-free selector (its expected counts were stale since the aligned-ring change85a803b: LOD0 343 -> 512); ClipmapNoOverlap drops the sheet sweep cases and the frozen7486c29oracle's sheet branch. New expected values were computed with a Python port of the selector that reproduces the test's existing 635/594487 and 508/74295 measurements exactly; all 5,145 sweep cases pass (no overlap, full coverage). - Tools/VoxelForgeTest.ps1 no longer passes voxel.TestFarSheetRing or reports far_sheet_ring. The sheet_tile dump field is kept (the parity compare reads it). Also removes settings nothing reads: CeilingViewMultiplier, CeilingBandChunks (pre-clipmap ceiling band) and DensityVolumeMaxTasks (the volume fill runs on its own thread). Saved assets that still carry these values load normally; the values are ignored. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01URDGLsu7c4nunqNFPJEFPU