[Request] D-Pad Support for XInput/DirectInput hat switches on gamepads/controllers

[Lv44EnderShaman]

Build: 0.5.1-master-c4dbe0a
OS: Windows 7 Home Edition 64-bit
Replicated with: Logitech Gamepad F310 (detected as Logitech Dual Action)

While the button mapping feature of the emulator supports gamepads' analog sticks as valid inputs for the in-game directional-pad, it is this user's notice that the emulator does not support mapping the Logitech F310's directional-pad as valid button inputs. This was replicated in both the gamepad's XInput and DirectInput modes. It is my theory that the gamepad maps its directional pad as the equivalent of hat switch signals, which mGBA to this date has had no sign of detecting for directional pad input. It may be a simple enough fix, if possible: Just add detection for hat switches from both XInput and DirectInput gamepads.

[endrift]

1. Hats do work, they just aren't remappable. Up = Up, Left = Left, Right = Right, Down = Down. This is currently hardcoded.
2. It's not a simple fix, as they show up completely separate from axes and buttons, so I'd need to add an entire new codepath through the controller codebase, which isn't exactly the simplest thing.

[endrift]

...it looks like I removed hat support pending a rewrite and then forgot to put it back. I should do that.

[Nicholas-Steel]

Yep, can confirm HAT support is missing. Thankfully the Logitech F310, F510 and F710 have a button that will swap the analog stick functionality with the DPad functionality.

[PrincessPleco]

Strange. I'm using the latest test build and my DPad works. I'm using the Logitech F310 as well and have "mode" turned off. :S

EDIT: Oh, but I'm using Windows 10. Could that have anything to do with it at all?

[Nicholas-Steel]

Windows 10 has nothing to do with it, the functionality will be re-implemented in v0.6.0 so it isn't too surprising if the functionality has already been restored in the source code.

[PrincessPleco]

Is it an old version? Because I've confirmed that the last five development builds (from the 11th to the 21st in January) all have working DPads on the Logitech F310 without using the mode button.

[Nicholas-Steel]

Did you not read this bug report topic? It doesn't work in v0.5.1 and Endrift has made it clear it also does not work in v0.5.2 (I confirmed this too but didn't clarify which version I tested). Endrift has also made it clear that it will be re-implemented at some point between v0.5.2 and v0.6.0 and you have confirmed that this has already happened.

Go back to v0.5.2 to see for yourself that support was missing. I don't know why you would want to do this when the developer has already stated that the support was previously removed and accidentally never re-implemented though.

[endrift]

Regardless, it's back now. See f4a61f9.

[PrincessPleco]

I'm sorry. I didn't mean to offend you. But I did check for myself and the DPad did work in the test builds on my machine. I made no error on that. I don't get why you felt it necessary to insult me over it when I've only been honest. I thought my input might help endrift solve the problem faster.

Because you seem so sure I'm wrong, I tested both v0.5.1 and v0.5.2 for myself. My DPad works in both versions with and without the mode button activated. I can't explain to you why it does, but it does. (aside: I don't have JoyToKey on)

Regardless, it's good that whatever broken thing existed was fixed, though it never affected my controller for some reason. Perhaps endrift has some explanation for this over why one person would be affected but not another? It could be useful for people to better test things in the future. Also I'm curious. :3

EDIT: For posterity, here's how my controller was set in all of those mgba versions I tried:

[gba.input-profile.XInput Controller #1]
keyR=5
keyL=4
axisUpAxis=-1
gyroSensitivity=2e+009
keyStart=7
keyLeft=2
keyDown=1
axisLeftAxis=-0
keyB=2
keyUp=0
axisRightValue=12288
axisRightAxis=+0
axisUpValue=-12416
axisDownAxis=+1
gyroAxisX=0
tiltAxisX=2
axisDownValue=12160
keyA=0
keyRight=3
gyroAxisY=1
axisLeftValue=-12288
keySelect=6
tiltAxisY=3

[Nicholas-Steel]

I was just saying it was a bit ingenuous to claim the problem doesn't exist when testing a Test Build that was made after the version that the bug was reported for, because the problem is likely already fixed. In your case though it's likely the problem only cropped up in the Windows version of the program which would explain why you don't notice it in v0.5.2 either (or the program is recognizing and using the gamepad configuration settings from a pre-existing config file generated by test builds you were using).

[PrincessPleco]

On the contrary: I was pointing out a discrepancy between my observation and another's. I couldn't reproduce the error. I don't see how pointing out my own observation is naive when it can be useful to check for differences between systems to find potential issues.

And alas: I just tested using a completely fresh and unedited config file and v0.5.1 still has working DPad for me. I also only use the Windows version. So the mystery of why it always works for me continues. :S (I'd like to know why, but it's really not that important, I suppose...)