[Mlir-commits] [mlir] [mlir][emitc] Add EmitC lowering for arith.trunci, arith.extsi, arith.extui (PR #91491)

[Corentin Ferry]

================
@@ -112,6 +112,78 @@ class CmpIOpConversion : public OpConversionPattern<arith::CmpIOp> {
   }
 };
 
+template <typename ArithOp, bool needsUnsigned>
----------------
cferry-AMD wrote:

`needsUnsigned` means that the cast op is to an unsigned type (non-sign-extending, guaranteed truncation prior to C++20). Perhaps `castToUnsigned` would be a better job :)

https://github.com/llvm/llvm-project/pull/91491