Adds the command line option --float-encoding=e, where e can be either real or ieee32.
The current mode with uninterpreted floats remains the default, since both modes have disadvantages (reals are not a sound representation of floats, and ieee floating point reasoning is incredibly slow).
When floating point operations are used in the default uninterpreted mode, a warning is issued.
Adds the command line option
--float-encoding=e, whereecan be eitherrealorieee32. The current mode with uninterpreted floats remains the default, since both modes have disadvantages (reals are not a sound representation of floats, and ieee floating point reasoning is incredibly slow). When floating point operations are used in the default uninterpreted mode, a warning is issued.This addresses issue #158.
This change is