Problem with REC body

I have a rather complicated geometry (which took considerable effort to
construct) that ran well (and good results) with the 2008.3 version of
Fluka. I'm checking out the 2011.2.2 version and have hit a problem
with the REC geometry.

The relevant geometry statement is
REC Pipe1 30.4439 0.0 316.9884 -11.5236 0.0 81.9942
               0.0 13.0 0.0 16.4384 0.0 2.31027983
(The last number had 4 decimal places in the original, but has been
adjusted here to achieve perpendicularity on my calculator.)

The error in the output file is
   REC Pipe1 37 0.30443900E+02 0.00000000E+00
0.31698840E+03 -0.11523600E+02 0.00000000E+00 0.81994200E+02 521
                           0.00000000E+00 0.13000000E+02
0.00000000E+00 0.16438400E+02 0.00000000E+00 0.23102798E+01
  Abort called from BODYGN reason REC HEIGHT NOT PERP. TO 2ND SEMIAXIS
Run stopped!
  STOP REC HEIGHT NOT PERP. TO 2ND SEMIAXIS

Does Fluka now require perpendicularity in double precision? Why isn't
it scaled, sensibly, to the accuracy of the input?

Or is there a bug?

The what(2) parameter on the GEOBEGIN card, even with values up to 100.,
appears to have no effect.

The peculiar numbers here reflect an 8 degree rotation about Y. It
would be nice for the code to make an adjustment to ensure the condition
it wants. Yes, a rotation transformation can be used. But that leads
to a run-time penalty, and I have many bodies that will need to be rotated.

A test input file is attached.

Thank you,
Joe Comfort