In Objective-C the code looked liked this and worked flawlessly,
NSInteger random = arc4random_uniform(99) + 1
NSData *data = [NSData dataWithBytes:& ra
There are a couple of things to consider when extracting an integer value from a Data
stream. Signedness and Endianess. So I came up with a function in an extension to Data
that infers Signedness from the type of integer you want to extract and passes Endianess and Index
as parameters. The types of integers that can be extracted are all that conform to FixedWidthInteger
protocol.
Reminder: This function will not check if the Index
range is inside the bounds of the Data
buffer so it may crash depending on the size of the type being extracted in relation to the end of the buffer.
extension Data {
enum Endianness {
case BigEndian
case LittleEndian
}
func scanValue(at index: Data.Index, endianess: Endianness) -> T {
let number: T = self.subdata(in: index...size).withUnsafeBytes({ $0.pointee })
switch endianess {
case .BigEndian:
return number.bigEndian
case .LittleEndian:
return number.littleEndian
}
}
}
Example:
let data = Data(bytes: [0xFF,0x1F,0x1F,0xFF])
let number1 = data.scanValue(at: 0, endianess: .LittleEndian) as UInt16
let number2 = data.scanValue(at: 0, endianess: .BigEndian) as UInt16
let number3: Int16 = data.scanValue(at: 2, endianess: .LittleEndian)
let number4: Int16 = data.scanValue(at: 2, endianess: .BigEndian)
Results:
number1 is 8191
number2 is 65311
number3 is -225
number4 is 8191
Observe the function calls to see how the type to be extracted is inferred. Of course Endianess doesn't make sense for Int8
or UInt8
, but the function works as expected.
Values can later be cast to Int
if needed.