On constructivity and the Rosser property: a closer look at some Godelean proofs