Brake: fix infinite run out

[ronso0]

Multiple times I experienced that engine.brake() would slow down the deck until the waveforms halted but play was never disabled.
I suspected a broken controller button (faulty debounce filter) to initiate brake again, but I could also reproduce it with a MIDI Through script a few times.

There is at least one possibility that the rampTo rate was crossed but scratchProcess didn't stop, thus slowing down to zero and accelerating again very, very slowly:

debug [Controller]    old 0
debug [Controller]    new 1.69772e-311
debug [Controller]    fabs 0.01
debug [Controller] .
debug [Controller] .
debug [Controller]    old 1.69772e-311
debug [Controller]    new 5.08986e-311
debug [Controller]    fabs 0.01
debug [Controller] .
debug [Controller] .
debug [Controller]    old 5.08986e-311
debug [Controller]    new 1.01731e-310
debug [Controller]    fabs 0.01
...
debug [Controller]    old 1.85776e-307
debug [Controller]    new 1.88052e-307
debug [Controller]    fabs 0.01
debug [Controller] .
debug [Controller] .
debug [Controller]    old 1.88052e-307
debug [Controller]    new 1.90341e-307
debug [Controller]    fabs 0.01
debug [Controller] .
debug [Controller] .
debug [Controller]    old 1.90341e-307
debug [Controller]    new 1.92641e-307
debug [Controller]    fabs 0.01

I fixed that

That and some more simplification and I couldn't reproduce the bug anymore 🤷‍♂️

[ronso0]

Draft until I can confirm it's fixed with some 'real life' tests with my controller.

[daschuer]

Suggested change

	const double kBrakeRampToRate = 0.01;
	constexpr double kBrakeRampToRate = 0.01;

[daschuer]

Thank you for investigating.
If you like, you may split out the useful refactoring. I think that can be merged immediately since it does not change the behavior.

[ronso0]

I don't mind if the refactring has to wait.
I have to start over anyway because I didn't consider that brake() can also be a spinback :|

Also, the code hasn't changed since 2.3 AFAICT so I'll rebase this onto 2.3

[ronso0]

Uff, we didn't test spinback after my most recent brake fixes #1085
It's broken 😬

• spinback accelerates infinitely
• this prevents spinback from starting when the deck is stopped
• this makes it runs way to long because brake rampTo is positive while spinback is running with negative rate
• spinback while softStart'ing accelerates the track to warp 10 until it hits the end

I'll prepare a fix for both issues.

[ronso0]

Superseded by #4708 (2.3)