Float mantissa, negative integer exponent

GTBecker · Post by **GTBecker** » 15 February 2010, 12:32 PM

console.writeline&#40;cstr&#40;&#40;10.0^2.0&#41;&#41;&#41;  ' yields 100.0
console.writeline&#40;cstr&#40;&#40;10.0^-2.0&#41;&#41;&#41; ' yields 10.0E-3 = 0.01
console.writeline&#40;cstr&#40;&#40;10^2&#41;&#41;&#41;      ' yields 100
console.writeline&#40;cstr&#40;&#40;10^-2&#41;&#41;&#41;     ' yields 0 since cint&#40;0.01&#41;=0
console.writeline&#40;cstr&#40;&#40;10.0^2&#41;&#41;&#41;    ' yields 100.0
console.writeline&#40;cstr&#40;&#40;10.0^-2&#41;&#41;&#41;   ' yields &.& ?

Post by **dkinzer** » 15 February 2010, 15:33 PM

GTBecker wrote:

Code: Select all

console.writeline&#40;cstr&#40;&#40;10.0^-2&#41;&#41;&#41;   ' yields &.& ?

This occurs because integral constants are represented internally as unsigned 32-bit values so the effect is the same as 10.0 ^ &Hfffffffe. The fact that the negative sign is present is stored along with the constant so this probably can be made to work as you were expecting it to work.

In the mean time, you can force the desired behavior by making it clear that you want the exponent treated as a signed value, e.g.

Code: Select all

console.writeline&#40;cstr&#40;&#40;10.0^CInt&#40;-2&#41;&#41;&#41;&#41;

ZBasic Forum

Float mantissa, negative integer exponent

Float mantissa, negative integer exponent

Re: Float mantissa, negative integer exponent