`deepErrorX` behavior for `Bit` and `BitVector n` are differenct

[lmbollen]

AFAIK there's always been discussion on how undefinedness should work especially for BitVector n (whether it should produce an X exception or not)

But regardless of that I think that at least BitVector n and Bit should behave the same. Right now we have the following behavior:

ghci> deepErrorX "" :: Bit
*** Exception: X: 
CallStack (from HasCallStack):
  errorX, called at src/Clash/Sized/Internal/BitVector.hs:300:16 in clash-prelude-1.9.0-67fee7029a5493706dc3a40bc6492fe6aff05d8584aa3f56c3d3f55011874849:Clash.Sized.Internal.BitVector
  deepErrorX, called at <interactive>:36:1 in interactive:Ghci21
ghci> deepErrorX "" :: BitVector 8
0b...._....
ghci> Clash.Sized.Internal.BitVector.Bit 1 0 :: Bit
.
ghci> Clash.Sized.Internal.BitVector.BV 1 0 :: BitVector 8
0b0000_000.

I feel like whether or not undefined bits cause an X exception, at least Bit and BitVector n should behave the same

[DigitalBrains1]

I think the idea is that BitVector has a spine whereas Bit does not. So for BitVector, we have a defined "spine" with undefined individual bits, whereas a bottoming Bit is just a bottoming Bit.

I would definitely support defining deepErrorX and ensureSpine for Bit such that it is also just Bit 1 0 / . ; that's my opinion.