Files
OpenFrontIO/tests/client/render/gl/UnitPassSmoothing.test.ts
T
Evan f4db4a33c8 Send nukes as motion plans and render them smoothly per frame (#4255)
## Summary

Follow-up to #4244's payload work: nukes were the last per-tick movers
flooding the worker → main update stream.

- **Core**: nuke trajectories are fully determined at launch
(precomputed parabola), so `NukeExecution` now records a `GridPathPlan`
when the nuke is built — same mechanism trade ships use — and the client
derives the position each tick. Per-tick `UnitUpdate`s for nukes in
flight are suppressed; only targetable flips and deletion
(interception/detonation) still emit. This covers atom bombs, hydrogen
bombs, and MIRV warheads (dozens of per-tick movers per MIRV
separation).
- The plan path replays a separate pathfinder rather than reusing the
stored trajectory array: the curve's cached points don't advance exactly
one index per tick, and the plan must match the movement pathfinder's
exact per-tick tile sequence.
  - `startTick` accounts for MIRV warheads' staggered `waitTicks`.
- **Render**: `UnitPass.drawMissiles` now lerps each nuke's instance
position `lastPos→pos` by wall-clock progress through the current tick,
so nukes glide along their arc at render framerate instead of jumping
once per 100ms tick. Both endpoints are real simulated positions — the
rendered nuke trails the sim by at most one tick and settles exactly on
it when ticks stop. Plan-driven units sync `lastPos` on path-stall ticks
so the lerp never replays a segment. Shells keep their existing
two-instance trail; SAM missiles are unchanged.

## Test plan

- New `tests/nukes/NukeMotionPlan.test.ts`: tick-exact alignment between
the recorded plan and core nuke position over the whole flight
(mirroring `GameView.advanceMotionPlannedUnits` math), `waitTicks`
offset, and that no per-tick unit updates are emitted in flight except
targetable flips and deletion.
- Full suite passes (1452 + 65), tsc/eslint/prettier clean.
- Verified in-game (headless Chromium, real WebGL): atom bomb arcs from
silo to target with the client position driven by the plan, missile
sprite renders intact while the smoothing rewrites the instance buffer
every frame, detonation FX land at the target.

🤖 Generated with [Claude Code](https://claude.com/claude-code)

---------

Co-authored-by: Claude Fable 5 <noreply@anthropic.com>
2026-06-13 13:59:03 -07:00

188 lines
5.7 KiB
TypeScript

/**
* Render-side nuke smoothing (UnitPass).
*
* `applyMissileSmoothing` lerps each recorded segment's lastPos→pos by
* wall-clock progress through the current tick and rewrites the missile
* instance buffer's x/y, leaving the packed per-instance bytes untouched.
* `flickerHashByte` reproduces (CPU-side) the per-instance flicker phase the
* vertex shader used to derive from its rendered position.
*
* The GL pass is exercised directly via its prototype with stubbed GL calls —
* the real method runs, only the two WebGL calls it makes are captured.
*/
import { afterEach, describe, expect, it, vi } from "vitest";
import {
flickerHashByte,
UnitPass,
} from "../../../../src/client/render/gl/passes/UnitPass";
const FLOATS_PER_INSTANCE = 4;
const TICK_MS = 100;
interface SmoothingHarness {
pass: UnitPass;
f32: Float32Array;
bufferSubData: ReturnType<typeof vi.fn>;
}
/**
* Build a UnitPass instance without its GL-heavy constructor, wiring up only
* the fields applyMissileSmoothing touches.
*/
function makeSmoothingHarness(
segs: number[],
instanceCount: number,
nowMs: number,
lastUpdateMs: number,
): SmoothingHarness {
const f32 = new Float32Array(instanceCount * FLOATS_PER_INSTANCE);
const bufferSubData = vi.fn();
const gl = {
ARRAY_BUFFER: 0x8892,
bindBuffer: vi.fn(),
bufferSubData,
};
vi.spyOn(performance, "now").mockReturnValue(nowMs);
const pass = Object.create(UnitPass.prototype) as UnitPass;
Object.assign(pass, {
gl,
smoothSegs: segs,
missileCount: instanceCount,
missileBuf: { float32: f32, buffer: {} },
lastUnitsUpdateMs: lastUpdateMs,
tickIntervalMs: TICK_MS,
});
return { pass, f32, bufferSubData };
}
function runSmoothing(h: SmoothingHarness): void {
(
h.pass as unknown as { applyMissileSmoothing(): void }
).applyMissileSmoothing();
}
afterEach(() => {
vi.restoreAllMocks();
});
describe("UnitPass.applyMissileSmoothing", () => {
it("lerps lastPos→pos by wall-clock progress through the tick", () => {
// instance 2; lastPos (10,20) → pos (30,40); 50ms into a 100ms tick.
const h = makeSmoothingHarness([2, 10, 20, 30, 40], 5, 1050, 1000);
runSmoothing(h);
const off = 2 * FLOATS_PER_INSTANCE;
expect(h.f32[off + 0]).toBeCloseTo(20); // 10 + (30-10)*0.5
expect(h.f32[off + 1]).toBeCloseTo(30); // 20 + (40-20)*0.5
});
it("sits at lastPos at the start of the tick (alpha 0)", () => {
const h = makeSmoothingHarness([0, 10, 20, 30, 40], 1, 1000, 1000);
runSmoothing(h);
expect(h.f32[0]).toBeCloseTo(10);
expect(h.f32[1]).toBeCloseTo(20);
});
it("clamps alpha to 1 so a stalled tick settles exactly on pos", () => {
// 200ms elapsed into a 100ms tick → alpha would be 2, clamped to 1.
const h = makeSmoothingHarness([0, 10, 20, 30, 40], 1, 1200, 1000);
runSmoothing(h);
expect(h.f32[0]).toBeCloseTo(30);
expect(h.f32[1]).toBeCloseTo(40);
});
it("writes each segment to its own instance slot and leaves others alone", () => {
const h = makeSmoothingHarness([1, 0, 0, 8, 16], 3, 1050, 1000);
// Mark instance 0's slot; it has no segment and must be untouched.
h.f32[0] = 111;
h.f32[1] = 222;
runSmoothing(h);
expect(h.f32[0]).toBe(111);
expect(h.f32[1]).toBe(222);
const off = 1 * FLOATS_PER_INSTANCE;
expect(h.f32[off + 0]).toBeCloseTo(4); // 0 + 8*0.5
expect(h.f32[off + 1]).toBeCloseTo(8); // 0 + 16*0.5
});
it("overwrites only x/y, preserving the packed ownerID/atlas/flags floats", () => {
const h = makeSmoothingHarness([0, 10, 20, 30, 40], 1, 1050, 1000);
h.f32[2] = 999; // ownerID slot
h.f32[3] = 888; // packed atlasIdx/flags/flickerHash slot
runSmoothing(h);
expect(h.f32[2]).toBe(999);
expect(h.f32[3]).toBe(888);
});
it("handles multiple smoothed nukes in one pass", () => {
const h = makeSmoothingHarness(
[0, 0, 0, 10, 10, 2, 100, 100, 200, 200],
3,
1050,
1000,
);
runSmoothing(h);
expect(h.f32[0]).toBeCloseTo(5);
expect(h.f32[1]).toBeCloseTo(5);
expect(h.f32[2 * FLOATS_PER_INSTANCE + 0]).toBeCloseTo(150);
expect(h.f32[2 * FLOATS_PER_INSTANCE + 1]).toBeCloseTo(150);
});
it("re-uploads exactly the active missile-instance float range", () => {
const h = makeSmoothingHarness([0, 10, 20, 30, 40], 4, 1050, 1000);
runSmoothing(h);
expect(h.bufferSubData).toHaveBeenCalledTimes(1);
const args = h.bufferSubData.mock.calls[0];
// gl.bufferSubData(ARRAY_BUFFER, dstOffset, srcData, srcOffset, length)
expect(args[1]).toBe(0);
expect(args[3]).toBe(0);
expect(args[4]).toBe(4 * FLOATS_PER_INSTANCE);
});
it("does nothing when there are no smoothed segments", () => {
const h = makeSmoothingHarness([], 3, 1050, 1000);
runSmoothing(h);
expect(h.bufferSubData).not.toHaveBeenCalled();
});
});
describe("flickerHashByte", () => {
const fract = (x: number, y: number) => {
const v = x * 0.1731 + y * 0.3179;
return v - Math.floor(v);
};
it("returns an integer byte in [0, 254]", () => {
for (const [x, y] of [
[0, 0],
[3, 5],
[40, 80],
[127, 255],
]) {
const h = flickerHashByte(x, y);
expect(Number.isInteger(h)).toBe(true);
expect(h).toBeGreaterThanOrEqual(0);
expect(h).toBeLessThanOrEqual(254);
}
});
it("is deterministic for a given tile position", () => {
expect(flickerHashByte(12, 34)).toBe(flickerHashByte(12, 34));
});
it("matches the shader's fract(worldX*0.1731 + worldY*0.3179) phase", () => {
for (const [x, y] of [
[3, 5],
[17, 42],
[80, 9],
]) {
// byte/255 should reproduce the original fract phase to within one step.
expect(flickerHashByte(x, y) / 255).toBeCloseTo(fract(x, y), 2);
}
});
});