The most promising method capable of providing accurate measurements of the first and second normal-stress differences in shear flows at shear rates typical of polymer processing is the so-called hole-pressure method, but its use has not been as widespread as would be expected, namely due to the experimental difficulties associated with performing such experiments accurately. In this work, we use a small-scale modular slit die to assess the practical utility of the method for in-line monitoring of polymer melt flow. We provide a quantitative analysis of intrinsic error sources and use state-of-the-art data acquisition tools to minimize errors associated with pressure transducers. Our results demonstrate that the method can be used to accurately measure the viscosity and first normal-stress difference in melts but probably not the second normal-stress difference because the intrinsic errors are too high, even when the influence of all the potential error sources is minimized or eliminated.