[ODE] Some Visual Studio warnings fixed

[J. Perkins]

I went through and fixed any warnings reported by Visual Studio that
were straightforward. This was mostly upcasts to a size_t from an int,
explicit casts to dReal from a constant value, and adding explicit
boolean comparisions instead of assuming the value from an int result.

The warnings that remain are probably legitimate 64-bit portability
problems and double/float mismatches, almost all of which are in
OPCODE. I need double precision positions for some work I have in the
pipeline so I will have to address this at some point.

Jason