float separationn

[Wolfgang Draxinger]

This is not really C/C++ related, but I think this is also a place to ask:
I'm currently working on a compressed 3D Object interchange format.
With C/C++ its no problem to raw-write the IEEE standard floats the data 
consists of into files. Now I'm writing some export scripts for blender 
in python. My problem is, that the python float object doesn't provide a 
raw-data access. What I need now, is some possibility to construct the 
IEEE float format from data returned by the standard math functions.
I think frexp() is a good start, however this still returns a float 
value. I'd like to have the raw bits...

Some ideas?

Thanks
Wolfgang Draxinger