struct PickerRange<T: Comparable> {
var start: T
var end: T
var step: T
var length: T {
// Binary operator '/' cannot be applied to two 'T' operands.
return max(start, end) - min(start,end) / step
}
}
Error is generated when creating computed properties. How to fix?
Simply specify one of the Numeric
protocols.
For floating point precision:
struct PickerRange<T:FloatingPoint> {
var start: T
var end: T
var step: T
var length: T {
return max(start, end) - (min(start,end) / step)
}
}
SignedInteger
, UnsignedInteger
or other Numeric
protocols are also an option.
// Comparable
Neither of these require of you to specify Comparable
additionally, so specifying one of these is suffice for the case.
EDIT:
How to make one structure for Int and Double?
Essentially, you can, by specifying SignedNumeric:
struct PickerRange<T:SignedNumeric & Comparable> {
var start: T
var end: T
var step: T
}
// Treated as PickerRange<Double>
let a = PickerRange(start: 1.0, end: 5.0, step: 1)
// Treated as PickerRange<Int>
let b = PickerRange(start: 1, end: 5, step: 1)
However, you have specialized length
property, which includes /
operator not supported by SignedNumeric
. A workaround would consist in type-checking/force-casting to a type you have initialized your struct with:
struct PickerRange<T:SignedNumeric & Comparable> {
var start: T
var end: T
var step: T
var length: T {
if T.self is Double.Type {
return (max(start, end) as! Double) - ((min(start, end) as! Double) / (step as! Double)) as! T
} else if T.self is Int.Type {
return (max(start, end) as! Int) - ((min(start, end) as! Int) / (step as! Int)) as! T
} else {
return 0
}
}
}
let a = PickerRange(start: 1.0, end: 5.0, step: 1)
let b = PickerRange(start: 1, end: 10, step: 1)
print(a.length)
print(b.length)