the 1/c -> 1*c optimisation is fine for x86 floats too
[libfirm] / include /
drwxr-xr-x   ..
drwxr-xr-x - libfirm