it's easy to determine the shape (and size, thus effective ƒ/stop) of a pinhole using a scanner or a microscope. Regarding how much it matters, well, I used to have a 4x5 pinhole camera that had a pinhole with an assymetric and rather undefined shape (crude needle job) and the negatives were good for contact prints and quite acceptable for moderate enlargements. In the most simple case, assuming an infinitesimally thin plate for the pinhole, the distribution of corner vignetting will directly follow the shape of the pinhole.