Patch Set 1:
Is there any reasonable way someone could implement/extend BigInt?
We don't require our users to be reasonable.
Another issue is that this code isn't actually particularly efficient for negative numbers.
The internal representation is signed magnitude, not two's complement, so we have to do conversion to expose a two's complement value.
So, maybe it's not the best API to use. Converting from unsigned Uint32List to signed Uint8List is a non-trivial conversion, it's not just "giving access to the underlying data".
s this what we want?
The best you could do is construct/destruct via arithmetic.
So, don't assume that this is *anyhing* other than an experiment (and I uploaded it only because I couldn't build it locally because my repository has some issue).