ipc_helpers: Amend floating-point type in Pop<double> specialization

Currently, this overload isn't used, so this wasn't actually hit in any
code, only the float overload is used.