FA-87238 / Physics integrator stability / Member archive

Cached normal impulse ignores the warm-start factor · case 03

Changing the warm factor has no effect on normal impulses and damping settings misbehave.

Member previewVariant 3 · 3 implementations · 8 checks per implementation

Case contract

solve(vn0, vt0, w, mu, iterations, cached, warm): several contact points share one relative normal and tangent velocity with inverse mass w. Warm start each point with warm*cached [jn, jt] and apply it to the velocities. Each iteration visits points in order: friction first (accumulated jt clamped to |jt|<=mu*jn), then normal (accumulated jn clamped >=0). Return [jn list, jt list, vn, vt] rounded to 6.

Why this case matters

Game and robotics physics loops depend on integrator update order, step control and stabilization terms; a wrong decision point turns a stable simulation into drifting or exploding motion.

One recorded failure

Sample boundary fixture

This sample comes from the broken implementation of a controlled reproducer.

Boundary fixture	Actual	Expected	Outcome
case 0	`[[0.285354, 0.189, 1.244], [0.194041, 0.12852, 0.84592], 0.0, -0.01976]`	`[[0.571954, 0.1512, 0.9952], [0.388929, 0.102816, 0.676736], 0.0, -0.01976]`	Failed

MEMBER ARCHIVE

The complete case is available to members.

This record includes three runnable implementations, regression fixtures, execution results, and source hashes.

Member access is invitation-based. Sign in with your invited account to inspect the sources.