The bugs in rtesthyp are related to:
(%i35) hgfred([-4],[1/2],z^2/2);
(%o35) hermite(8,z/sqrt(2))/1680
(%i36) ev(%,nouns);
(%o36) z^8/105-4*z^6/15+2*z^4-4*z^2+1
Changing rtesthyp test 13 from ratsimp(hgfred([-4],[1/2],z^2/2) -
hermite(8,z/sqrt(2))/1680) to ratsimp(''(hgfred([-4],[1/2],z^2/2)) -
hermite(8,z/sqrt(2))/1680) allows this test to pass. What is best, I
don't know; hgfred([-4],[1/2],z^2/2) --> hermite(8,z/sqrt(2))/1680
might be nicer than hgfred([-4],[1/2],z^2/2) --> explicit 8th degree
polynomial.
With SBCL, I get errors for rtesthyp.mac, problems (13 14 66) and
rtest_hypgeo.mac, problems: (42 142 225); test 42 is the difference
between pochhammer(1,i) and i!. There is a note in rtesthypeo about
test 226 (that actually seems to be 225) that says it might be a bug.
For share_testsuite, Error summary:
test_zeilberger.mac, problems: (47 52)
rtest_simplify_sum.mac, problems: (2 3 4 5 6 7 8 11 13 14 15 16 19 21 22)
rtest_levin.mac, problem: (77)
The errors in rtest_simplify_sum might be due to a weirdness in my Maxima.
Barton