#158 OPT_DOUBLE_PRECISION build is broken

[Grégoire Payen de La Garanderie]

Hello,

The flag OPT_DOUBLE_PRECISION introduced in version 2.4 is broken.
When a daeFloat variable is read from a .dae file, it is read as a single-precision floating point variable then stored in a double-precision floating point which leads to weird values.

In daeAtomicType.cpp:331
_scanFormat = "%g"

I believe it should be:
#ifndef COLLADA_DOM_DAEFLOAT_IS64
_scanFormat = "%lg"
#else
_scanFormat = "%g"
#endif

Thanks