Can the hearth share the village’s friendly colors? The first native bonfire was rejected. A new local reconstruction brings copper logs and sage, teal and cream stones into the game, ready for another visual review.
Originalnotiz und Messwerte (Englisch)
The owner rejected the first generated bonfire’s pale flattened appearance. Its native frame and rejected test build remain preserved; the prior approved hut player was restored while making the replacement. A new built-in imagegen concept informed another actual TripoSR mesh, rather than serving as implementation proof. Blender v3 showed washout/dark texture seams, v4 smoothed and recolored the mesh, v5/v6 explored coherent stone colors but retained 12 tiny boundary edges, and v7 closes the generated solid before simplification and validates the saved/exported source. v4’s original zero-nonmanifold figure was stale because it preceded simplification; later versions and this note correct that claim. No old evidence was overwritten.
Measured 16 October 2026 Europe/Zurich: (1) the replacement raw mesh’s 28,176 triangles become 4,606 in saved v7, from generation report and Blender manifest/source; (2) 188 raw boundary edges become zero nonmanifold edges after v7 cleanup, checked afresh in the exporter; (3) 13/13 isolated Unity Editor checks passed at 00:28 Zurich for source, dimensions/material, unchanged colliders, flames/light, paused idle/validated use and picking coverage; (4) the scoped routine/home-life/evidence/journal-privacy/workshop Python suites passed 67 tests in 4.07 seconds. These checks do not establish natural activity frequency, native mouse picking or sustained performance.
The final native source is unity-bonfire-source-905f9c190896-20261015222920, Blender SHA256 9fe83f91bf2e971aea86a851e809f82eb5fc421f4f75d84369a1e7033604ea8b. The normal player is open, simulation paused and background disabled; no inference or resident action was started. The solid uses one packed texture/material, while the pre-existing flame/light behavior remains separate and is raised to clear the taller solid. Existing navigation, model budgets, Needle gates, huts and tree studies were not changed by this increment. Prior build/source assets were preserved, with no pruning. The replacement is visibly more colorful but still has softened/fused log transitions, lacks the concept’s crisp cut-end rings and has artificial stone color-sector boundaries. It is integrated for review, not owner-accepted art.
Evidence: first rejected actual native frame 20261015T221555158425-ba319f4e1f18; new concept 20261016T002104247757-6bb4d84b4ed6; actual raw CPU mesh preview 20261016T002344361202-ee7f285dd704 (corrected resolution-128 label; initial incorrect resolution-64 sidecar stays private/unapproved). Intermediate Blender front/underside pairs: v3 20261016T002415058682-c60e5f5064b6 and 20261016T002420438092-5047c33378f1 (washout/seams); v4 20261016T002458568822-41f5a0f0988d and 20261016T002505275474-cad6f12318af (smooth palette); v5 20261016T002542223326-564df428a904 and 20261016T002547695771-ce1b5e4eee11 (coherent colors); v6 20261016T002620136084-2c72993e41d0 and 20261016T002626090182-b374707e2bd9 (boundary-repair attempt); v7 20261016T002724727562-4fbfc7ddd68a and 20261016T002729672296-0d598ad4a571 (closed solid). Final actual native close/garden views are 20261015T223030821638-30c020d64204 and 20261015T223101501090-a3383bd41065, with prior garden baseline 20261015T221252196020-b70707df8d01. Every published image and sidecar was individually inspected and privacy reviewed, which is separate from artistic acceptance. Details: docs/bonfire-asset.md. Next is owner visual feedback on the actual replacement.