eisenwave wrote: @arsenm it will map directly to an LLVM intrinsic once there is a floating-point version of it, but I wanted to keep this PR focused, so that's not provided yet.
The pure integer side of it doesn't really need an intrinsic or could be broken down into smaller ones perhaps, but if we're going to have an intrinsic for a saturating cast anyway, we may as well cover int-to-int. https://github.com/llvm/llvm-project/pull/227603 _______________________________________________ cfe-commits mailing list [email protected] https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
