I think you look at Serge lang analysis I. there you may get an answer gneeral enough and one which requires limited machinery. also look on net for macdonalds paper on Proof of satokes theorem. the conditions on f are very weeak if you employ gauge integral  in fact differential forms should be defined using integral .The prototype is efinition of divergence given in many physics texts.