Remove dubious looking FP to int code that was ifdef'ed out. Doesn't

look like a TODO is needed anymore.

src/CodeGen_ARM.cpp

@@ -1153,27 +1153,13 @@ void CodeGen_ARM::visit(const Cast *op) {
         }
     }
 
-    // TODO(<issue needed>): Resolve whether this is still needed.
-#if 1
     // LLVM fptoui generates fcvtzs or fcvtzu in inconsistent way
     if (op->value.type().is_float() && op->type.is_int_or_uint()) {
         if (Value *v = call_overloaded_intrin(op->type, "fp_to_int", {op->value})) {
             value = v;
             return;
         }
     }
-#else
-    // LLVM fptoui generates fcvtzs if src is fp16 scalar else fcvtzu.
-    // To avoid that, we use neon intrinsic explicitly.
-    if (is_float16_and_has_feature(op->value.type())) {
-        if (op->type.is_int_or_uint() && op->type.bits() == 16) {
-            value = call_overloaded_intrin(op->type, "fp_to_int", {op->value});
-            if (value) {
-                return;
-            }
-        }
-    }
-#endif
 
     CodeGen_Posix::visit(op);
 }

src/CodeGen_LLVM.cpp

@@ -4771,7 +4771,7 @@ Value *CodeGen_LLVM::call_intrin(const llvm::Type *result_type, int intrin_lanes
             bool formal_is_fixed = isa<FixedVectorType>(formal_param_type);
 
             // Apparently the bitcast in the else branch below can
-            // change the scalar type and vector length togehter so
+            // change the scalar type and vector length together so
             // long as the total bits are the same. E.g. on HVX,
             // <128 x i16> to <64 x i32>.  This is probably a bug, but
             // it seems to be allowed so it is also supported in the