AbstractUnitRange for BitArray/BitVector indexing

[mkitti]

Several methods could be generalized to work with AbstractUnitRange rather than just UnitRange. There are now several AbstractUnitRange subtypes:

julia> subtypes(AbstractUnitRange)
4-element Vector{Any}:
 Base.IdentityUnitRange
 Base.OneTo
 Base.Slice
 UnitRange

This pull request adapts several methods to work on BitArray / BitVector and AbstractUnitRange.

[johnnychen94]

Can you also share more information about this change? For example, what was wrong with the previous status. The code change itself doesn't explain the background very well.

It's also a good practice to have associated test cases with every non-trivial functionality change.

I also see that you just opened the other two similar PR #41805, #41807 and #41809. The idea behind those changes are also uncleared to me.

[mkitti]

The impetus for the series of PRs stems from
#39241 (comment)

Basically if we have more Base.OneTo around is the Base ecosystem ready for it?

[johnnychen94]

For that, I don't have any objection. But please follow the previous PR #39896 as an example, prepare for more explanation and sufficient test cases so that when someone with privilege jumps in, he doesn't need to ask you to do these and we don't need to wait for one more round of review.

[mkitti]

Thanks @johnnychen94 . I just haven't had time to write it up just yet and mainly focusing on that range PR at the moment. This PR and the other related ones are mainly just to scout ahead for issues that may create in the future. There seems to be some complicated issues with the string part.