Thank you (Dr) Bob,
after searching the net on reciprocity I have come to some conclusions (maybe an expert pinholes already knows all that):
1) reciprocity failure tables have been calculated for night-shots where contrast is much higher (the road is dark but a lamp can be very bright). Probably the table should be reconsidered in the light of this (no pun intended);
2) tri-x is not the best film for B&W pinhole as it has a very bad reciprocity curve. Tabular grain films will in theory perform better. Fuji across (impossible to find here in Italy) with no reciprocity up to very long time and tmax with very few reciprocity in the range 1 to 100 seconds. I'll make some test with t-max. Oddly Ilford gives the same reciprocity for delta, hp and fp