CRT
a0659e24ac
Two rules were missing, and each one had been quietly failing for a while. The first is structural. Every check in this repo asked about surfaces -- winding, coplanarity, triangle counts -- and none of them could see a piece of a model hanging in the air with nothing holding it up. A chest lid separated from its chest passed everything we had. scripts/support.ts voxelises the model, finds the connected components of occupied space and requires each one to reach the floor; scripts/audit.ts runs that across default, min, max, both sweep ends, actions and the midpoint of a morph, so a variation cannot break what the default gets right. That audit started at 22 failing cases: the bell and the sign had no supporting structure at all and now have a frame and a post, the lantern was in four pieces, the sack's collar floated above its body, the basket's fruit hung level with the rim instead of resting on the base, and the ladder's rungs were sized by a formula that disagreed with where the rails actually were. It now reports 0. The second is that we had no target. The feedback loop was an offline render of our own output, which can only answer "is this broken", never "is this right". Given a reference photograph of a cask, five defects were obvious in seconds that three rounds of self-review had missed: four hoops where there should be six, no rivets at all, a chime cut so shallow it did not read, and a stave count low enough to leave a visibly polygonal silhouette. All four are fixed here. Two things the rivets taught us, both kept as comments rather than as scars: a stud that tapers outward tilts its side normals back towards the axis, which the radial check correctly reports as reversed winding -- a real rivet head flares, so the honest shape is also the passing one. And once seated deep enough to look right, the stud's buried inner cap falls inside the shell the check inspects, so it is not generated at all, exactly as bandGeometry already declines to build the inside of a hoop. 27 models, 16973 triangles, all checks pass, 0 floating cases.
116 lines
3.8 KiB
TypeScript
116 lines
3.8 KiB
TypeScript
/**
|
|
* @medieval-kit/wooden-ladder
|
|
*
|
|
* Two rails with rungs between them. The cheapest model in the kit and one of
|
|
* the highest in scene value: it suggests vertical movement in a scene.
|
|
*
|
|
* The rungs go INTO the rails (a housed joint), so no surface sits on the same
|
|
* plane as the surface of the rails.
|
|
*/
|
|
import { Color } from 'three'
|
|
|
|
import {
|
|
chamferedBoxGeometry,
|
|
createKitModel,
|
|
jitter,
|
|
MEDIEVAL_PALETTE,
|
|
mergeColoured,
|
|
} from '../core/index.ts'
|
|
|
|
export interface WoodenLadderConfig {
|
|
readonly height: number
|
|
/** Distance between the rails (metres). */
|
|
readonly width: number
|
|
readonly rungCount: number
|
|
/** How much the rails narrow towards the top. 0 = parallel. */
|
|
readonly taper: number
|
|
readonly seed: number
|
|
}
|
|
|
|
export const woodenLadderDefaults: WoodenLadderConfig = {
|
|
height: 2.2,
|
|
width: 0.42,
|
|
rungCount: 8,
|
|
taper: 0.18,
|
|
seed: 4,
|
|
}
|
|
|
|
export type WoodenLadderParts = 'rails' | 'rungs'
|
|
|
|
export function createModel(overrides: Partial<WoodenLadderConfig> = {}) {
|
|
return createKitModel<WoodenLadderConfig, 'oak', WoodenLadderParts>({
|
|
id: 'wooden-ladder',
|
|
defaults: woodenLadderDefaults,
|
|
slots: ['oak'],
|
|
build: ({ config, random }) => {
|
|
const tint = new Color()
|
|
const shade = (lift = 0): Color => {
|
|
tint.copy(MEDIEVAL_PALETTE.oak)
|
|
tint.offsetHSL(jitter(random, 0.012), jitter(random, 0.05), lift + jitter(random, 0.055))
|
|
return tint
|
|
}
|
|
|
|
const railThickness = config.width * 0.1
|
|
const half = config.width / 2
|
|
|
|
// The rails converge towards the top; this single detail stops the ladder
|
|
// from being "two boards" and makes it a ladder.
|
|
const lean = half * config.taper
|
|
/**
|
|
* Half-gap between the rails at height fraction `t`.
|
|
*
|
|
* This exists because the rung length and the rail placement used to be
|
|
* two separate expressions that disagreed. The rails are built upright
|
|
* and then rotated about Z, so at height y each one sits at
|
|
* `side·half − y·sin(θ)`; the rung length assumed `half − lean·t`, which
|
|
* narrows where the real rails widen. At the default taper the error was
|
|
* small enough to hide. At taper 0.4 on a 5 m ladder the top rungs came
|
|
* out half the length of the gap they were supposed to span and floated
|
|
* between the rails.
|
|
*
|
|
* Both now read from here, so they cannot drift apart again.
|
|
*/
|
|
const railHalfAt = (t: number): number => {
|
|
const y = -config.height / 2 + t * config.height
|
|
return half - y * Math.sin(lean / config.height)
|
|
}
|
|
const rails = [-1, 1].map((side) => {
|
|
const rail = chamferedBoxGeometry(
|
|
[railThickness, railThickness * 1.35],
|
|
[railThickness * 0.85, railThickness * 1.15],
|
|
config.height,
|
|
railThickness * 0.16,
|
|
[0, 0, 0],
|
|
shade(),
|
|
)
|
|
// Lean it slightly around the Z axis: bottom end out, top end in.
|
|
rail.rotateZ((side * -lean) / config.height)
|
|
rail.translate(side * half, 0, 0)
|
|
return rail
|
|
})
|
|
|
|
const rungs = []
|
|
const count = Math.max(2, config.rungCount)
|
|
for (let i = 0; i < count; i += 1) {
|
|
const t = (i + 0.5) / count
|
|
const y = -config.height / 2 + t * config.height
|
|
// The rung is a little LONGER than the rail gap at that height, so that
|
|
// its ends stay inside the rails.
|
|
const span = railHalfAt(t) * 2 + railThickness * 0.9
|
|
rungs.push(chamferedBoxGeometry(
|
|
[span, railThickness * 1.05],
|
|
[span, railThickness * 1.05],
|
|
railThickness * 0.8,
|
|
railThickness * 0.16,
|
|
[0, y, 0],
|
|
shade(0.03),
|
|
))
|
|
}
|
|
|
|
return {
|
|
rails: { slot: 'oak', geometry: mergeColoured(rails) },
|
|
rungs: { slot: 'oak', geometry: mergeColoured(rungs) },
|
|
}
|
|
},
|
|
}, overrides)
|
|
}
|