What would that function do? Fully copy/paste the original impl and ignore the errors in the body? Call the original function and ignore the error? In each of those scenarios you’d still need syntax for “call function and ignore error”, so you’re right back where you started.

Unless you’re proposing writing alternative “error-ignoring” versions of everything in the std::io namespace, your idea doesn’t even make sense.