Khronos stopped supporting .spec files in 2013. They're now over a year
old without support. As such it's not worth us continuing to support
them either.
If the legacy keyboard device receives a key down event with IsRepeat set it
will only raise it's own key down event if it's KeyRepeat field is set to true.
This is as documented, regression casused by refactoring. Fixes issue #201.
Also change the GameWindowState example to show setting of KeyRepeat to true
and false and how that changes the event counts for the legacy and new keyboard
devices.
When running under NUnit GetEntryAssembly returns null. In this case we instead
search through the AppDomain for an assembly that matches the AppDomain name.
Fixes warnings [#61] by disabling unused field warnings for two structures in
HidProtocol. These fields aren't currently used by OpenTK but the stuctures are
used in native marshalling so must match the documented structures perfectly.
Issue reported at http://www.opentk.com/node/3805
We must not throw exceptions from a finalizer, as this leads to the
runtime forcibly taking down the application.
WinFactory.Dispose() could crash with a NRE when the joystick driver has
not been initialized. Fixed by checking for null before disposing the
input driver.
HIDInput.Dispose() would remove elements from a DeviceCollection during
iteration. This results in an exception.
We are now using a plain for-loop instead.
`DeviceCollection.GetEnumerator()` now returns a struct IEnumerable<T>
directly to avoid boxing.
Additionally, we can now use `DeviceCollection[int]` as a shortcut to
`FromIndex(int)`.
Joystick elements (axes, buttons, hats) are now reported in the same
order as SDL2. This fixes potential mismatches in the GamePad
configuration database.
Additionally, elements are now counted correctly (duplicate elements no
longer count towards the total.)
Starting with SDL 2.0.4, its xinput bindings are working correctly.
Previous versions would return an invalid zero guid and use incorrect
mapping - this has now been fixed.
