> >>Agree, we should be consistent with the JLS, so throwing exceptions in case > >>of overflow isn't the right thing to do. > > > >"ArithmeticException" is currently used for that purpose in many methods of > >"MathUtils". > > > For floating point values? Which methods?
No. But from the start, my question was: Shouldn't CM be consistent with itself? In the same way that "MathUtils" contains methods (taking integer arguments) that use "ArithmeticException" to signal overflow (whereas Java will happily wrap to negative values), I suggested that the other methods (taking floating point arguments) could be consistent with the choice of being safer than the JLS. Gilles --------------------------------------------------------------------- To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org For additional commands, e-mail: dev-h...@commons.apache.org
