intel/debian/unstable/2.2.17 failures

[Brian Gough]

The monte failures are known problems -- so far I have not turned them
off as I thought it would be the wrong approach to get rid of a
warning without finding the cause

Changing 

  for (num_dim = 1; num_dim < 10; num_dim++)

to 

  for (num_dim = 1; num_dim < 9; num_dim++)

would be sufficient

The specfunc failure looks like the error estimated by the function is
slightly too small -- also not critical, although maybe Jerry will
increase the error estimate slightly.