chmod_1, chmod_2, chmod_3 and lrshift_1

[Paul Thomas]

Steve and FX,

chmod_2.f90 and chmod_3.f90 are both failing under FC5/Athlon.  The 
failure occurs for

  i = chmod (n, "a-w")
  if (i == 0) then
    if (access(n,"w") == 0 .or. access(n,"W") == 0) call abort
  end if

It looks to me as if the function form suffers from the same 
unreliability as that in g77 and remarked upon in its manual; presumably 
for the same reason?

I would vote for expunging these tests from the testsuite.

BTW I trust that the intrinsic.texi entries will follow sometime?

All the best

Paul