We report a single-lens/single-source microlensing event designated as OGLE-2019-BLG-1058. For this event, the short timescale ($sim 2.5$ days) and very fast lens-source relative proper motion ($mu_{rm rel} sim 17.6, {rm mas, yr^{-1}}$) suggest that this isolated lens is a free-floating planet (FFP) candidate located in the disk of our Galaxy. Because this is a high-magnification event that could have a nearby lens, we have the opportunity to measure the terrestrial microlens parallax (TPRX). We find a TPRX signal consistent with a disk FFP, but at low significance. A direct measurement of the source proper motion ($mathbf{mu}_{rm S}$) shows that the large $mu_{rm rel}$ is due to an extreme $mathbf{mu}_{rm S}$, and thus, the lens is consistent with being a very low-mass star in the bulge and the TPRX measurement is likely spurious. We show how a precise measurement of $mathbf{mu}_{rm S}$ with the mean properties of the bulge proper motion distribution would have given the opposite result, i.e., provided supporting evidence for an FFP in the disk and the TPRX measurement. Because the conditions for producing TPRX (i.e., a nearby disk lens) will also tend to produce a large $mu_{rm rel}$, this case demonstrates how $mathbf{mu}_{rm S}$ measurements in general provide a strong test of TPRX signals, which Gould et al. (2013) showed were an important probe of FFP candidates.