[ODE] 0.8-rc1 does not build with double precision and GIMPACT

[dranders@xerotope.net]

Hello,

When I use "./configure --enable-double-precision --with-trimesh=gimpact",
and then build with make, I get the following error:

collision_trimesh_gimpact.cpp: In function 'void
dGeomTriMeshGetTriangle(dxGeom*, int, dReal (*)[4], dReal (*)[4], dReal
(*)[4])':
collision_trimesh_gimpact.cpp:494: error: cannot convert 'dReal*' to
'float*' for argument '3' to 'void
gim_trimesh_get_triangle_vertices(GIM_TRIMESH*, long unsigned int, float*,
float*, float*)'
collision_trimesh_gimpact.cpp: In function 'void
dGeomTriMeshGetPoint(dxGeom*, int, dReal, dReal, dReal*)':
collision_trimesh_gimpact.cpp:505: error: cannot convert 'dReal*' to
'float*' for argument '3' to 'void
gim_trimesh_get_triangle_vertices(GIM_TRIMESH*, long unsigned int, float*,
float*, float*)'

Looks like gim_math.h #defines a GREAL as a float, and does not check
compiler flags to switch between double and float.

Thanks,

Dean