remove time.Sleep from itests and make use of a more deterministic approach

[ZZiigguurraatt]

There are a number of places where time.Sleep is used in itests:

• lightning-terminal/accounts/store_test.go

  Line 125 in 66b0f15

  time.Sleep(5 * time.Millisecond)

• lightning-terminal/itest/assets_test.go

  Line 2017 in 66b0f15

  time.Sleep(time.Millisecond * 250)

• lightning-terminal/itest/litd_accounts_test.go

  Lines 442 to 446 in 66b0f15

  	// No idea why it makes a difference whether we wait before calling
  	// s.Recv() or not, but it does. Without the sleep, the test will fail
  	// with "insufficient local balance"... ¯\_(ツ)_/¯
  	// Probably something weird within lnd itself.
  	time.Sleep(time.Second)

• lightning-terminal/itest/litd_custom_channels_test.go

  Lines 239 to 242 in 66b0f15

  	// Sleep for a second to make sure the balances fully propagated before
  	// we make the payment. Otherwise, we'll make an RFQ order with a max
  	// amount of zero.
  	time.Sleep(time.Second * 1)

• lightning-terminal/itest/litd_custom_channels_test.go

  Line 1385 in 66b0f15

  time.Sleep(time.Second * 1)

• lightning-terminal/itest/litd_custom_channels_test.go

  Line 1411 in 66b0f15

  time.Sleep(time.Second * 1)

• lightning-terminal/itest/test_harness.go

  Line 204 in 66b0f15

  time.Sleep(slowMineDelay)

These sleeps should be removed and replaced with a more deterministic approach. This will increase the speed of testing, the results will be more predictable and not dependent on the loading of the machine running the tests.