This is from our WIP2 Bluetooth branch which is about a year out of date.

I copied files to other location, then created this branch and used WinMerge to merge in all related changes,

Without hopefully undoing any of the other changes

Turns out they are a lot different than other PS3 controllers.

When plugged into USB, it appears like these units do not send
any reports.  They do allow us to do pairing (which is differnt)
than other PS3, as well as set the Bulb LED color, which I now
have it alternating colors every few seconds and also it tries
to pair it with a BT dongle

When connected to Bluetooth, it now is setup to generate reports,
as well as again set the bulb color and rumble.

So far it maps the 3 button bytes into the buttons field and then
simply copies most all of the data down into Axis, starting where
the Trigger button is.

JoystickBT.ino was updated, to know about this controller and as
mentioned above in the USB case it tries to pair and plays with the
color of bulb.

In the BT case it tries to display data.  I did a quick and dirty
change of the Accel/Gyro like stuff which is probably totally wrong.

Pressing some of the buttons will set bulb color and pressing trigger
will set rumble.

PS3 Motion controller - very little mapping of axis

Thanks mostly to Mike (@mjs513) - a WIP to support some of the digitizer tablets that are not properly supported by Mouse.

Like Wacom Intous...

Turn off the debug support, so if pulled in to Master other users can use it without needing to setup Serial1

Looks like some devices when ask for extended inquriy results instead might return results with RSSI... So needed to process those as well

Trying out using the extended Inquiry data search instead of the real simple one.  This allows us to hopefully get the name of the device as part of this, which now allows the PS4 to know it is a PS4 on the Pair type constructor.

Still WIP to understand XBox one...

Support to pair PS3 if it knows the BTADDR and user presses select and PS button and plugged in USB.

Remembers remote name, and tries to get it when PAIR version of code is running.  (Currently fails).

Does some mapping of PS3 Bluetooth data to match USB HID data.

Warning: as part of debuggigng why joystickType value was changing, I changed the class to make it a function you call instead of just a member variable.  Should mark it inline...  I think I caught all of the places in example code that was accessing it.  If we find we want ability for user probram to update it, should add another member to explicityly set it.

I have some PS3 support working... Hope I did not break PS4.... ;)

Also This assumes that you did a bind of the dongle to PS3 external, no support yet, I did this on RPI3...

Rumble needs work!

Hopefully did not break other things...

Find that the PS4 requires you to ask to connect versus some/most other devices will do the request...

Also found issue without printing, where adding a delay(1) in a couple of places appears to help when Serial1 printing is not on

I put in hopefully the start of Bluetooth Mouse.

However the one mouse I have is not even trying to talk to the dongle, so not have not seen any data to verify that it is correct and to see the actual data returned...

There are some of the BT dongles that have a device type of 0xff instead of the correct one, so added table of VID:PID to check against, currently only one in list.  If it is this pair the Bluetooth system will claim it.

PS4 starting to work, Still some rumble issues.

Also some WIP parts for caching remote name

Some of these changes also came from @mjs513

Hopefully not everything screwed up... :D

The first gigabyte fix, was to always force the keyboard into boot protocol mode, which appeared to work well for all of the keyboards I tried.

Unforationally it caused issues with some wireless keyboard/mouse combo packages, including some logitech and some microsoft keyboards.

This version instead, I created a list (currently only one vendor id), of vendor id:product id, that need to be forced into boot mode.  Knowing that if there is one keyboard vendor with this N key rollover feature, there are probably others.  So while this list can be expanded, this requires users to update library.  So added a method forceBootProtocol method to keyboard class that allows the user to try it out without needing to update the library sources.

The mouse test program has example of either setting it before the keyboard is initiated, as well as another way that when a connection is found it asks for the vendor id, and if it is the one of my gigabyte keyboard, it calls the function then.

While doing this I found some of the query functions like idVendor on the keyboard class would not compile as the keyboard object has multiple inheritence and both base classes have this.  So like the joystick objects I added these to the top level function and these functions will direct the calls to the approrpriate places.

Support for using the Microsoft XBox 360 wireless receiver for windows
to be able to communicate with XBox 360 wireless controllers.

In theory this should support up to 4 controllers.  So far I have only
tested with one as I only have one.

Added a joystick only test that should support up to 4 joystick objects.

Still working on PS4

Again still WIP - but now have the Bluetooth object talking through a keyboard object...

With this update, I can now talk to the keyboard without having to repair each time.

Still WIP

Code is still WIP, needs lots of reorg and fix up and lots more to figure out.

But started working toward the Bluetooth HID devices and have the messages going back and forth from my BT dongle to a small BT keyboard.
It is finally showing keys being pressed and the like.

Code needs lots of work.  Like does pairing everytime, need to make that optional.

Also need to figure out where to split off the HID stuff from Root BT stuff.  ...

Plus it mainly prints lots of debug information.

WIP - I captured the USB packet information for the startup of a Bluetooth adapter.  I am now in the process of understanding the data that is being sent back and forth and have the device now partially configured.

Nothing works yet!

Changed from claim at device level to interface level.  such that the Microsoft Wireless controller can support multiple joysticks...

Tested with one joystick as I only have 1 (I think)...

WIP - So far i am getting the basic values for 1 XBox 360 controller.  The device should handle up to 4 devices, but so far hard coded to only handle 1...

Added getSysExArrayLength(), changed msg_sysex_len to uint16_t

Added support for Joysticks, that on some of them we can not receive more of the Joysticks axis.  So far mainly on PS3 and PS4.  So you can now get the Gyro/accel stuff.  You get feedback on the DS4 touch area, PS3 you can get the pressure values on several of the buttons.  LIke the RT/LT ones.

In addition added some support for Rumble.  So far it appears to be working somewhat on the PS3 and PS4 controllers.  On the PS4 you can also set the RGB LED light values and on the PS3 you can set the 4 LEDS on the front., which are normally used to say which controller it is.