8bitdo micro on Windows does not display as a gamepad using al_set_joystick_mappings (directx)

[connorjclark]

more of #1697 ... but now on Windows and with this controller:

03000000c82d00002090000000000000,8BitDo Micro,a:b1,b:b0,back:b10,dpdown:h0.4,dpleft:h0.8,dpright:h0.2,dpup:h0.1,guide:b12,leftshoulder:b6,lefttrigger:b8,leftx:a0,lefty:a1,rightshoulder:b7,righttrigger:b9,rightx:a3,righty:a4,start:b11,x:b4,y:b3,platform:Windows,

This is how it appears with and without the controller database:

I've no idea what the 2nd/3rd sticks are, and the 1st stick shouldn't have any accelerator.

But as far as the controller database goes: fill_standard_gamepad in wjoydxnu.cpp shows that the mapping has 4 axes, but allegro detects only 3. The number of buttons/hats are matching.

[connorjclark]

For Stadia, all the sticks actually make sense (with and without database, same result) err, no: the second stick's horizontal scale is treated as the "accelerometer" portion of the first stick, while the vertical portion is treated as the horizontal component of the "brake" stick. weird

Fails to identify as a standard gamepad for the same reason: one axis short.

[connorjclark]

this seems just like what I was fixing over in #1463. I was looking for a reason that PR could be closed, but I just updated it with main to try this now and it does improve the detection enough for the controller database mapping to kick in:

so I'll get that PR cleaned up now.