v0.503

Fixed a small bug in EDmain_convex which gave an error for diffraction IRs of order 2 and higher. Also made a small change in EDwedge2nd which can give a substantial speedup: the function had three lines which cleared huge variables (to maximise the memory available) but that clearing took much time, so for one test case: a number of receivers for a 64-sided polygon disc, the saving in calculation time was 56%. So, those clear commands were commented out.