XINPUT TEST BLOODSTONE WINDOWS

Unity is combining XInput devices and HID XInput devices into a single list, and is still attempting to make an association between HID devices and their XInput id's. I see that Unity will support > 4 XInput controllers (not possible using XInput), and ones that are not associated with an XInput device will still function, except Axes 8 and 9 do not function, as expected because they're HID devices. The symptoms have changed, but the underlying cause is still the same from what I can tell. I can still reproduce issues related to the HID->XInput association which Unity is still doing, though not the exact same issues as before. Only 4 XInput controllers can be used at once.Ĭlick to expand.Actually, I just did testing on 2019.3.0b2 on the legacy input system. In this case, XInput tells you one thing - this is a gamepad. You must work within the limitations of the underlying input APIs. Rewired will never attempt this kind of 'fusion' to work around Microsoft's built-in limitation, and I certainly hope Unity doesn't try to do this in their new input system. That's why you will see many complaints on the forum about XInput gamepads triggers responding on the wrong device when multiple devices are attached in some scenarios. All attempts to do this fail at some point or another with multiple controllers attached. However, read the forums and you will see many, many problems that come from this 'fusion' because there simply is no reliable way to fuse these devices because Microsoft gives you no way to associate an XInput device id to a HID device.

XINPUT TEST BLOODSTONE WINDOWS

Unity has been combining XInput and HID for XInput controllers for years on Windows using the L/R trigger values from XInput and the name and other elements from the HID device. This is an XInput limitation and cannot be worked around reliably. This is one of the most common questions I get from Rewired users.

It may be that this is confusing plover? Or plover is listening to the slave devices directly? I don't know.Click to expand.The console is not using XInput to display that information, it's using information from the HID device. Whenever the other event node gets active an XIDeviceChangedEvent is sent to notify the client. The second cause would be that plover is doing something incorrect with the keyboard events. Given your comment about xev that's unlikely. I can't verify that but you could easily check this by running both evemus side-to-side and check for release events before you release any physical key. So there's only two causes I can identify here: one is that the keyboard sends release events whenever a key on the other node is down or more than 6 keys are pressed.

And combined state from devices has always been required functionality (so you can e.g. That's not surprising since we don't have any limits on how many keys can be pressed at once in the server. Replaying the events attached here with an added delay between key press and key releases shows that xinput does correctly receive the 8 key presses before the key releases. This isn't a bug though, this is expected behaviour in this situation. It's identical to having multiple keyboards connected.

So to X it looks like you're pressing 6 of the keys on one device, then 2 on the other. Ok, so the thing here is: these are two separate devices, sending the key events separately.

Xinput test bloodstone