fix(force-simulation): Fix a bug that would sometimes keep a simulation running indefinitely

[regexident]

I discovered the bug while experimenting with an approach of adjusting the alphaTarget/alphaDecay dynamically, based on the simulation's total kinetic energy (i.e. sum of node velocities): the simulation just kept running, despite alpha having dropped far, far below 0.001 already.

Any change to alphaTarget/alphaDecay caused the simulation to get restarted, even if alpha had already fallen below alphaMin.

There are a couple of places where runOrResumeSimulation(); gets called, but currently only one of them is guarded by a alpha >= alphaMin check, so I move the check into runOrResumeSimulation() itself.

The bug is trivially reproducible by doing something like this:

function handleTick(e: { alpha: number; alphaTarget: number }) {
    // the re-fires the simulation, but due to `alphaTarget < alphaMin` it should still terminate:
    alphaTarget = Math.random() * alphaMin;
    console.log(alpha, alphaTarget, alpha >= alphaMin);
}

<ForceSimulation {alphaTarget} onTick={handleTick} ...>...</ForceSimulation>

[changeset-bot]

🦋 Changeset detected

Latest commit: 950214d

The changes in this PR will be included in the next version bump.

This PR includes changesets to release 1 package

Name	Type
layerchart	Patch

Not sure what this means? Click here to learn what changesets are.

Click here if you're a maintainer who wants to add another changeset to this PR

[github-actions]

built with Refined Cloudflare Pages Action

⚡ Cloudflare Pages Deployment

Name	Status	Preview	Last Commit
layerchart	✅ Ready (View Log)	Visit Preview	950214d

[pkg-pr-new]

Open in StackBlitz

npm i https://pkg.pr.new/layerchart@546

commit: 950214d

[techniq]

Nice catch! I noticed this the other day while I was experimenting with something. Thanks for the fix!

[techniq]

@regexident care to run pnpm format to apply prettier formatting and make CI happy?