Home » Project database » Core » The Open Toolkit library » Issues
AndyKorth's picture

No mouse input from OpenTK.Input.Mouse.GetState() on Mac

Posted Friday, 21 December, 2012 - 21:26 by AndyKorth
Project:The Open Toolkit library
Version:1.x-dev
Component:Code
Category:bug report
Priority:critical
Assigned:Unassigned
Status:in progress (review)
Description

Mouse input via OpenTK.Input.Mouse.GetState() doesn't work on Mac OS X. It always reports no buttons pressed and a position of (0,0).

I have commit a replication example here:
https://github.com/andykorth/opentk/blob/master/Source/Examples/OpenTK/T...

The short of it is really just:

if(OpenTK.Input.Mouse.GetState()[MouseButton.Left]){
                Console.WriteLine("You clicked the left mouse button!");
}

I'd appreciate it if anyone on other platforms could say if I've got my replication correct.. ie: This should work, right?

I'll be looking for a fix next. Note that this input method is suggested in the docs here: http://www.opentk.com/doc/input


Comments

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.
AndyKorth's picture

#1

Posted Friday, 21 December, 2012 - 22:52 by AndyKorth
Title:No mouse input on Mac» No mouse input from OpenTK.Input.Mouse.GetState() on Mac

This is probably the same issue as http://www.opentk.com/node/2800

In HIDInput.cs, 

        void DeviceValueReceived(IntPtr context, IOReturn res, IntPtr sender, IOHIDValueRef val)
        {
            MouseState mouse;
            KeyboardState keyboard;
            if (MouseDevices.TryGetValue(sender, out mouse))
            {
                MouseDevices[sender] = UpdateMouse(mouse, val);
            }
            else if (KeyboardDevices.TryGetValue(sender, out keyboard))
            {
                KeyboardDevices[sender] = UpdateKeyboard(keyboard, val);
            }else{
                Debug.Print ("Device {0:x} not found in list of keyboards or mice", sender);
            }
        }

MouseDevices (and keyboard devices) is populated from:

      void DeviceAdded(IntPtr context, IOReturn res, IntPtr sender, IOHIDDeviceRef device)
  ...
                        MouseDevices.Add(device, state);
..

The device key isn't sent to the DeviceValueReceived callback. (You can check in /System/Library/Frameworks/IOKit.framework/Headers/hid/IOHIDBase.h) Instead I think the intention was to register different callbacks for each device - specifically different callbacks for the mouse and keyboard, instead of trying to determine which input was from which device inside the DeviceValueReceived callback.

Jase's solution of using the context variable to identify the device is pretty good, it requires a lot less touching of the code. Will keep you folks up to date.

AndyKorth's picture

#2

Posted Friday, 21 December, 2012 - 22:58 by AndyKorth

In the interest of being through, this method of accessing the keyboard doesn't work:

 if(OpenTK.Input.Keyboard.GetState()[Key.A]){
                Console.WriteLine("The A key is down!");
            }

As expected, for the same reason.

AndyKorth's picture

#3

Posted Friday, 21 December, 2012 - 23:12 by AndyKorth

I have fixed the mouse issue, and I have mouse events in OpenTK! Very exciting :D

Patch:
http://kortham.net/temp/OpenTKIssue3242.txt

And it's been pushed here: https://github.com/andykorth/opentk

In this commit:
https://github.com/andykorth/opentk/commit/39c3f3f54324b80262d10fc5af823...

AndyKorth's picture

#4

Posted Saturday, 22 December, 2012 - 00:45 by AndyKorth
Status:open» in progress (review)

I thought there were some other keyboard issues, but after a lot more testing than I should have needed to do, it seems like the above commit has fixed keyboard as well. For the sake of documenting what I found:

Quote:

Press A on the keyboard and hold it, and it generates these four events:
IOHIDValueGetIntegerValue: 4 IOHIDElementGetUsagePage: KeyboardOrKeypad IOHIDElementGetUsage: -1
IOHIDValueGetIntegerValue: 0 IOHIDElementGetUsagePage: KeyboardOrKeypad IOHIDElementGetUsage: 1
IOHIDValueGetIntegerValue: 1 IOHIDElementGetUsagePage: KeyboardOrKeypad IOHIDElementGetUsage: 4
IOHIDValueGetIntegerValue: 4 IOHIDElementGetUsagePage: KeyboardOrKeypad IOHIDElementGetUsage: -1

Release A and these four events are generated:
IOHIDValueGetIntegerValue: 0 IOHIDElementGetUsagePage: KeyboardOrKeypad IOHIDElementGetUsage: -1
IOHIDValueGetIntegerValue: 0 IOHIDElementGetUsagePage: KeyboardOrKeypad IOHIDElementGetUsage: 1
IOHIDValueGetIntegerValue: 0 IOHIDElementGetUsagePage: KeyboardOrKeypad IOHIDElementGetUsage: 4
IOHIDValueGetIntegerValue: 0 IOHIDElementGetUsagePage: KeyboardOrKeypad IOHIDElementGetUsage: -1

So I'm going to just ignore the usage of -1.

Here's the change to quiet down some of the error printing:
https://github.com/andykorth/opentk/commit/c98d9e4503cf8f9498a17ddd2d37b...

References: http://www.usb.org/developers/devclass_docs/Hut1_11.pdf and HID headers.