EyeLink - fix wait_for_*

[wkruijne]

See comments for commit 778ac13 . Fixes wait_for_events

Please evaluate whether this was corrected properly! I haven't been able to test this, because I do not have straightforward acces to an Eyelink-tracker atm.

Notes:

• This does NOT affect the non-native event detection, because pylinks sample() methods will always return the newest sample available, AFAIK.
• These functions now return a timestamp that is in sync with clock.time() . This is consistent with the non-native event detection return value and consistent with the opensesame eyelink plugins. However, it might, without warnings, disrupt programs that presently make use of eyelink-time (though I can't imagine there will be any)
• This also implements a 'prrivate' function _get_eyelink_clock_async() . One could argue that you should omit the _ and make it a function available to the user (this is what we did for opensesame). However, the function might only make sense for the eyelink -- breaking the consistency of the class across trackers. Plus I honestly can't think of a good use case for it.

[esdalmaijer]

Brilliant work, thanks a lot!

On your last note: I do see the value of potentially using the clock asynchrony, but only in highly specific cases. Advanced users should be able to type an extra _ to access the function.