Fix support for external trackers via VMC

[Seneral]

This PR intends to fix the existing support for external tracking via VMC.

Known Issues:

• OSCRouter spams "Packet is too large for the buffer in use"
  • Buffer is just randomly small, data is NOT unexpectedly large
  • No apparent reason why buffer is downsized, might be data race
  • WONT FIX: JavaOSC is slated to be replaced with a simpler reimplementation, no point fixing every last issue with it rn
• OSC sometimes fails to listen when restarted due to refreshSettings
  • Causes OSC to stop working randomly, even when not in settings
  • WONT FIX: Hopefully addressed by osc reimplementation

Requires SlimeVR/SolarXR-Protocol#213

[Seneral]

The first few commits are in PR #1840
After that, this is ready for review, too

[Seneral]

Alright this should be just about everything needed for basic external tracker support via VMC.

There are a few things I'd change about the UI still:
I added a rescale tracking toggle to both OSC Router and VMC settings.
OSC Router really needs it, but it requires a VRM to be loaded in VMC settings. This is not obvious to the user.
The toggle in VMC is perhaps redundant, turning it off matches current behaviour without a VRM supplied at all.
I also don't like that it's not very obvious by just looking at the settings page why and when the VRM is needed.
I'd like to change it or associate it closely with that toggle somehow. Input appreciated.

Since IK Solver is now properly integrated (#142), tracking arms from controllers should now be deprecated IMO.
It's inferior in every way except if you absolutely need the hand position to match up, even if the IK solver determines the arm reach is insufficient.
This would also mean adding a warning to anybody that has "Force Arms From HMD" set to off - it should be on, the skeletal hierarchy should be one tree with head as the root, and IK solver resolving all positional constraints within (including hands).

The IK Solver should be able to handle trackers on any body part, though not optimally yet:

• Two serial IK chains (e.g. head-hip-feet) would not optimally converge, can likely be easily addressed once I have those trackers to test with
• Overconstrained scenarios (e.g. two slime arm trackers + positional hand tracker) may be unstable due to fixed lerp factor

Ideally, there'd be configuration in the GUI whether a given tracker needs positional calibration in IKConstraints.kt or not. Currently, an HMD on the head and hand-trackers are assumed to be aligned properly, all other trackers have a rudimentary offset calibration on reset (untested by me).

Finally, I'd consider not squashing some of these commits as they're quite varied and tracing back the reasoning for each of them would be painful without the context of the messages.
I'll open this for review after I address at least the remaining GUI issues.