[Mlir-commits] [llvm] [mlir] [LLVM] Refactor intrinsic validation (PR #194061)

[Durgadoss R]

durga4github wrote:

> Anyway, I first wanted to get a general feedback here and if this seems reasonable, we can maybe get this in and then work on improving the error messages generated.

This direction seems reasonable to me.

Is it possible to separate the VarArg absorption logic into a separate PR and land it independently?
>From what I see, this may require us to fold the VarArg failure into `NoMatchArg` itself or add a (temporary) new value in the enum.





https://github.com/llvm/llvm-project/pull/194061