Closed marius311 closed 4 years ago
Seems reasonable to me. Not sure who wrote this code but @maleadt might want to check it looks ok.
Sorry, forgot about this. LGTM!
I know its not much, but any chance of a release including this fix? The adapt(NamedTuple())
bug causes me to need to be on master of this package and makes reproducing the necessary environment harder.
Sure!
I simplified the code for adapting Tuples and NamedTuples without breaking any of the tests (I'm not seeing any other obvious impact on inferability or allocations). I also fixed the fact that previously
adapt(NamedTuple())
was returningTuple()
instead ofNamedTuple()
(and added a test for this).