replaced tests timeouts with flush helper

[rafapaezbas]

This PR replaces the timeouts in tests that could be replaced by the flush helper. It does not replace back off timeouts because the logic for does tests do not allow the replacement.

[mafintosh]

CI is failing

[HDegroote]

Yeah, I think we got lucky and discovered another flaky test relying on TIMEOUTS (it's not a test which was modified in this PR)

hyperswarm/test/swarm.js

Lines 360 to 402 in 2229393

	test('one server, one client - correct deduplication when a client connection is destroyed', async (t) => {
	const { bootstrap } = await createTestnet(3, t.teardown)
	const swarm1 = new Hyperswarm({ bootstrap, backoffs: BACKOFFS, jitter: 0 })
	const swarm2 = new Hyperswarm({ bootstrap, backoffs: BACKOFFS, jitter: 0 })
	let clientConnections = 0
	let serverConnections = 0
	let clientData = 0
	let serverData = 0
	const RECONNECT_TIMEOUT = CONNECTION_TIMEOUT * 4
	swarm1.on('connection', (conn) => {
	serverConnections++
	conn.on('error', noop)
	conn.on('data', () => serverData++)
	conn.write('hello world')
	})
	swarm2.on('connection', (conn) => {
	clientConnections++
	conn.on('error', noop)
	conn.on('data', () => clientData++)
	conn.write('hello world')
	if (clientConnections === 1) setTimeout(() => conn.destroy(), 100) // Destroy the first client connection
	})
	const topic = Buffer.alloc(32).fill('hello world')
	await swarm1.join(topic, { server: true, client: false }).flushed()
	swarm2.join(topic, { server: false, client: true })
	await swarm2.flush()
	await timeout(RECONNECT_TIMEOUT) // Wait for the first connection to be destroyed/reestablished.
	t.is(clientConnections, 2)
	t.is(serverConnections, 2)
	t.is(clientData, 2)
	t.is(serverData, 2)
	await swarm1.destroy()
	await swarm2.destroy()
	})

[HDegroote]

Looks good to me. The flaky test discussed above needs to be fixed though, but can be in another PR

Edit: managed to reproduce the flakiness: https://github.com/holepunchto/hyperswarm/actions/runs/7170541056/job/19523507101 . It's pretty rare, 80 repetitions before it hit. Probably fixable by using the flush helper of this PR--will look into that once this is merged