andykaylor wrote:
> Given that there are already similar functions in the `Type` class -- e.g.
> `has{Signed,Unsigned,}IntegerRepresentation`, `hasFloatingRepresentation` --
> it seems a bit inconsistent the definition of `hasBooleanRepresentation`. I
> would expect that "vector of booleans" to have `hasBooleanRepresentation`
> returning `true`, and I would expect `_Atomic(<type with boolean
> representation>)` to have `hasBooleanRepresentation` returning `false`. Any
> thoughts?
Yes, I also noticed that when I was moving this. I wasn't sure which way it
should go, so I opted to keep the code the same as it was in its original
location. There is already a precedent (`hasPointerRepresentation`) for not
including vectors. Also, vectors of Booleans have different considerations that
scalar Booleans. For instance, they are stored differently in memory.
Maybe the function should have a different name to avoid confusion?
https://github.com/llvm/llvm-project/pull/134159
_______________________________________________
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
