Regression in `hidApiLibrary.hid_open_path(path)` v 0.8 about hid4java HOT 6 CLOSED

Success! With the correct udev rule it finally worked:

KERNEL=="hidraw*", ATTRS{idVendor}=="XXXX", ATTRS{idProduct}=="YYYY", TAG+="uaccess"

(where XXXX and YYYY are, of course, the corresponding vendor and product ids). This is from https://github.com/libsdl-org/SDL/blob/main/src/hidapi/udev/69-hid.rules. Other udev recommendations floating around (ATTRS{busnum}="X") didn't work for me.

from hid4java.

Thanks for reporting this. That code hasn't changed beyond documentation since inception so I'm inclined to suspect the underlying compilation of the hidapi library.

Can you give some more detail on the following:

• which library is being used (libusb or hidraw (default))
• verification that the udev settings are correct
• exact platform (e.g. linux-aarch64)

from hid4java.

• Hidraw (This might be the difference, because in Julia I'm using the build that used libusb backend)
• udev settings should be fine, as it works in Julia on the same computer
• linux-x86-64

from hid4java.

If you have time, could you review issue #97 which provides a lot of information regarding Ubuntu and hidraw/libusb.

from hid4java.

I didn't see anything related in #97, but I found out that in version 0.7 linux-x86-64\libhidapi-libusb.so and linux-x86-64\libhidapi.so are binary equal and are actually the libusb backend build. So the "regression" stems from the fact there was a build bug in 0.7 which resulted in only one linux backend for 64bit (not 32!): libusb.

Finding out why I can't open devices using the linux backend will need to be investigated further.

from hid4java.

I'm glad you got to the bottom of this issue. Was the wiki Troubleshooting section any help? See https://github.com/gary-rowe/hid4java/wiki/Troubleshooting#how-usb-devices-are-mapped-in-linux

from hid4java.