Hitching Post

A showcase piece, not an example. Procedural timber hitching post: a 125 mm square post seated in a closed iron shoe, one cross-rail through the post, carried on two raking knee braces and capped with iron end sleeves, a pyramidal cap with eaves, two rings hung through eyes under the rail, and a horseshoe nailed heels-up to the post's face. Then the shipped pipeline: unique-cell UVs, Cycles high-to-low normal bake, LOD chain, convex collider, Unity glTF export.

It asserts budget conformance of the generated result. It does not witness an API contract. "It rendered without error" is not a check.

Intended size: post 0.125 m square and 1.16 m of timber, cap 0.096 m, cross-rail 0.50 m (0.506 m over its end caps). Outer AABB 0.506 × 0.149 × 1.246 m.

Composes skills mesh-editing-and-bmesh, bake-high-to-low, depsgraph-and-evaluated-data, engine-export-presets, and snippets bake_normal_high_to_low.py, setup_bake_target_image.py, lod_chain.py / decimate_to_budget.py, convex_hull_collider.py, export_preset_unity.py (helpers copied, not imported as a package). Hygiene combinatorics match examples/mesh-hygiene-audit (copied, not imported). Coplanar-pair counting matches showcase/signpost.

Budgets#

Declared as named constants; every gate recomputes from the mesh, materials, UVs, evaluated LOD, collider, or export file.

Base triangles rose from 1598 to 1680. The old rings were faceted tori clipped into the arm; the new eyes, shanks, and 24-segment hung rings replace that mesh. Cylinder caps are triangulated so the n-gon budget stays at 0.

DECIMATE COLLAPSE triangle counts are not identical across series — the gate is a ratio band, not an exact count. On this mesh the three binaries agreed. Bake pixels are stochastic; the gate is has_data plus operator FINISHED, not byte-identity. Construction is closed-form; the only RNG is the seeded per-piece wood tone. Bevel inputs are sorted by edge index, so the face order is the same on every run; a Python set of edges handed to the bevel had made it vary. Export byte counts differ on 5.2.1 (glTF serializer), not a gated axis. Euler is 14 (12 before the horseshoe added one more closed shell) and is not gated to 2.

Horseshoe#

A lone post with a stub rail read as the plainest object in the showcase lineup. A horseshoe nailed heels-up to the post's front face is what a real hitching post carries, and it is geometry, so it is budgeted: a flat iron bar (17 mm wide, 7.5 mm thick) swept around 290° of an arc, heels capped, centred on the post between the upper band and the rail and clear of the rings. Its back face bites the post face by 1.5 mm (HS_BITE); its face stands 7.5 mm proud. The shell classifier sets it apart from the hung rings, which are the same flat-loop shape class, by where it sits (on the post's centre line, not at ±EYE_X). Horseshoe seat reads the bite and the stand-off against the post's own front face off the mesh; --float-horseshoe pulls it 4 mm off the face and exits 18 (bite −0.0025 m). It moved the budgets it should: base triangles 1680 → 1852 (inside the declared 1400–2400) and metal faces 742 → 828. The outer AABB, collider (80 tris) and every other seat are unchanged. The hero key light rose from 660 to 960 W so the still sits in the calibration luma band.

Knee braces and end caps#

After the horseshoe the post was still among the simplest objects in the asset-sheet lineup: a stub rail pinned through a post reads as a cross, not as a load-bearing hitch. The rail now sits on two raking knee braces, one under each arm, as a hitch rail that takes a pulling horse needs. Each brace is a 40 × 50 mm strut from the post's side face at 0.66 m up to the rail's underside, housed 22 mm into both (BRACE_BITE), and it stops at x = 0.090, inboard of the hung rings. The ring–wood overlap gate now counts the braces as wood. Both rail ends carry an iron sleeve 36 mm long, with bevelled edges and 5 mm walls, standing 3 mm past the end grain.

seat_audit classifies the new shells by position. A wood shell off the post's centre line is a brace; the cap sits on it. A metal shell out at the rail ends is an end cap, not a band. Brace seat asserts exactly two braces, each overlapping the post and the rail, and two caps gripping the rail. --float-braces drops each brace head 36 mm below the rail and exits 18 (brace_arm 0). The additions moved the budgets they should: base triangles 1852 → 2156 (inside the declared 1400–2400), wood faces 114 → 222, metal faces 828 → 880. The outer AABB widened by the caps' 3 mm standoff per end, from 0.500 to 0.506 m, and OUTER_SIZE follows the measurement. The collider went from 80 to 72 triangles.

Bands, wood and stage#

The two iron bands were sized from half - grip: an inner half-width of 53 mm and an outer one of 61 mm, against a 62.5 mm post face. They sat inside the post, and only the chamfered corners broke the surface, showing as small black slits up the post. They now wrap it: the inner face bites 1.5 mm into the post (BAND_BITE), so each band stands 6.5 mm proud. Band seat asserts it; --sunk-bands restores the old bands and exits 18 (proud −0.0015 m).

The post, rail and cap each carry their own tone and grain, and the iron is rusted. The stage grid grew from 14 to 60 m, because the wall's left edge showed as a bright band in the corner of the hero. The temp .glb is removed after its size is measured.

Each falsifier violates exactly one named budget. Proven on Blender 4.5.11 LTS, 5.1.2, and 5.2.1 LTS (the binaries' own --version):

Default exit is 0 on all three. --clip-ring also overlaps the rail; the centerline gate is the one that fires.

Run#

blender --background --python hitching_post.py --
blender --background --python hitching_post.py -- --skip-decimate
blender --background --python hitching_post.py -- --stray-vert
blender --background --python hitching_post.py -- --twin-sole
blender --background --python hitching_post.py -- --lift-z
blender --background --python hitching_post.py -- --clip-ring
blender --background --python hitching_post.py -- --short-post
blender --background --python hitching_post.py -- --sunk-bands
blender --background --python hitching_post.py -- --float-horseshoe
blender --background --python hitching_post.py -- --float-braces
blender --background --python hitching_post.py -- --output preview.webp --engine cycles

Smoke does not pass --output or a falsifier.

Exit codes#

File-local. 9 is a valid check code. 10 is reserved for gallery_framing.check_framing on the --output path.