This is a breaking change - see BREAKING-CHANGES.txt

While the affected callbacks are cancelled before the referenced
state is deleted, we have had user reports that they can still be
accessed by the cancelled callbacks causing crashes. After only
finding warnings that WinRT AsyncCallback cancellation is not a
guaranteed thing, we saw it best to wrap the pointers.

The current minimum-supported Xcode (10.1) includes the iOS 12.1 SDK, so
APIs from iOS 10 will always be available.

f1b6bbc921 erroneously replaced

    callback = newCallback;
    previousCallback = callback;

with

    previousCallback = std::exchange (callback, newCallback);

It looks like previousCallback is intended to hold the last active
callback, so the previous version was correct after all.

By defining JUCE_IOS_AUDIO_EXPLICIT_SAMPLERATES, the iOS audio device
will always use the requested samplerates instead of querying the
current audio device for the samplerates it supports. This is useful
because certain hardware (such as the Focusrite iTrack Dock) takes a
long time to set new samplerates, which can end up freezing the main
thread for significant lengths of time.

This approach is inspired by the AUM app for iOS, which appears to
provide a fixed list of "allowed" samplerates, rather than querying the
device for its allowed samplerates.

When AudioDeviceStop() is called from a thread other than the IO thread there are no guarantees about the IOProc being called before the thread actually terminates. Moving this call to the audioCallback() method ensures that the device will be stopped immediately and we can remove a check that was taking a few seconds to close devices.