Otherwise we stomp programmatic additions

Turns out that my test for the device wasn't actually being called.
Then it turned out that I wasn't setting the match_all flag.
Then it turned out that the match_all flag wasn't being applied to RotateCmd.

Needed for exposing up through intiface central

Have the utility methods build the smallest unit possible.

Also fix a spelling mistake in the config

Thanks to @denialtek for the heads up

* Edeny (117)
* Echo Neo (117)
* Tammy Pro (S38A)

There turned out to be 2 issues:
1. DeviceInfo returned `Ok;` which broke the version checking logic
An easy fix with some validation
2. This device doesn't accept write_without_response
Harder, but corrected by building in fallback support to the BTLEPlug
hardware adapter: at write time, we check that the characteristic
supports the write mode and if not, we emit a warning and fall back
to other mode.

This change will prevent spurious characters from needing the be added to user configs

This is a little device by the Hismith brand Auxfun to add remote
and app control to their low spec fucking machines.

Fixes #556

Fixes #510

* Alex Neo 2
* Emma Neo (bad config)
* Phoenix Neo 2
* Hannes Neo

And using the same app and protocol:
* Fantasy Cup Theodore

Move module up to intiface-engine, as that's the only place we use it.
Anyone else who will use this probably has their own impl, so this will
provide us extra freedom to randomly change whatever we want while not
having to rev the library.

Introduces ButtplugTestServer for tests, which is mostly the same
as remote server. Could probably be paired down and cleaned up but eh.

Since we'll be using these to build user config files

Fixes #555

Also adding name for Emma Neo

Fixed the name of the Svakom Lite Neo too

This is a BT2 serial rotation controller that doesn't
have a reverse mode, so it's mapped as a scalar rotator.

The 2 outputs are not indivually controlled.

The device registers 2 RFCOMM ports, only the second seems
to allow connections.

Whilst I'm able to write to the device, I'm not getting any
response, but sirah on Discord confirmed that it worked for
them.

No tests, as I don't see an easy way to test serial devices.

Broke when I removed docs

Fixes #534

These devices don't send name advertisments properly (not sure why), so we were
using the advertised service 0xffb0, but that conflicted with the Hismith devices.

This fix uses the manufacturer data: I'm hoping the company will be rare 0x026c,
because only the first 2 bytes of the actual data array seem to be consistent...

Also reverting d923726 which added a duplicate Chorus definition as a
WeVibe4 (it was discovered that the Chorus was an oddball and got its
own protocol implementation in 4a2fb619)

Bumps [json5](https://github.com/json5/json5) from 2.2.0 to 2.2.2.
- [Release notes](https://github.com/json5/json5/releases)
- [Changelog](https://github.com/json5/json5/blob/main/CHANGELOG.md)
- [Commits](https://github.com/json5/json5/compare/v2.2.0...v2.2.2)

---
updated-dependencies:
- dependency-name: json5
dependency-type: indirect
...

Signed-off-by: dependabot[bot] <support@github.com>

Bumps [json5](https://github.com/json5/json5) from 2.2.0 to 2.2.2.
- [Release notes](https://github.com/json5/json5/releases)
- [Changelog](https://github.com/json5/json5/blob/main/CHANGELOG.md)
- [Commits](https://github.com/json5/json5/compare/v2.2.0...v2.2.2)

---
updated-dependencies:
- dependency-name: json5
dependency-type: indirect
...

Signed-off-by: dependabot[bot] <support@github.com>