nan [was: grftest failure]

[David Relson]

At 01:40 PM 1/5/03, Clint Adams wrote:

> > modify t.frame so it doesn't delete the output directories and then 
> compare
> > the t.grftest runs.
>
>checks is from make check, and grftest is run manually.
>
>--- checks.10032.20030105T133701/tests/grftest.out      2003-01-05 
>13:37:44.000000000 -0500
>+++ grftest.20030105/tests/grftest.out  2003-01-05 13:30:41.000000000 -0500
>@@ -1,9 +1,9 @@
>  id     graham       robinson     fisher       tristate
>- 1   S 1.000000   S 0.613817   S 1.000000   u 0.500000
>- 2   S 1.000000   S 0.685438   S 1.000000   S 9.40e-03
>- 3   S 1.000000   S 0.671075   S 1.000000   S 1.72e-05
>- 4   S 1.000000   S 0.671587   S 1.000000   S 1.47e-07
>- 5   S 1.000000   S 0.703862   S 1.000000   S 0.00e+00
>- 6   S 1.000000   S 0.686800   S 1.000000   S 7.75e-12
>- 7   S 1.000000   S 0.739067   S 1.000000   S 0.00e+00
>- 8   S 1.000000   S 0.736257   S 1.000000   S 0.00e+00
>+ 1   S 1.000000   h 0.538130   h 0.500000   u 0.500000
>+ 2   S 1.000000   S 0.625327   S 0.990601   S 9.40e-03
>+ 3   S 1.000000   S 0.640989   S 0.999983   S 1.72e-05
>+ 4   S 1.000000   S 0.694330   S 1.000000   S 1.47e-07
>+ 5   S 1.000000   S 0.723848   S 1.000000   S 0.00e+00
>+ 6   S 1.000000   S 0.690170   S 1.000000   S 7.75e-12
>+ 7   S 1.000000   S 0.745375   S 1.000000   S 0.00e+00
>+ 8   S 1.000000   S 0.784853   S 1.000000   S 0.00e+00

I'm revising my thinking ...  Right answers from manual execution, wrong 
answers from make check.

How many versions of bogofilter on the machine?  Could it be a pathing issue?