Reputation: 27275
I'm currently doing this:
let floatData: [UInt8] = [0x00, 0x3F, 0xC0, 0x00, 0x00, 0x66, 0x00, 0x00, 0x3F, 0xC0, 0x00, 0x00, 0x66, 0x00, 0x00, 0x3F, 0xC0, 0x00, 0x00, 0x66, 0x00]
floatData.withUnsafeBufferPointer { (bytes: UnsafeBufferPointer<UInt8>) -> Float in
let byteOffset = 1
var floatBytes : UInt32 = (UInt32(bytes[byteOffset]) << 24)
floatBytes |= (UInt32(bytes[byteOffset+1]) << 16)
floatBytes |= (UInt32(bytes[byteOffset+2]) << 8)
floatBytes |= UInt32(bytes[byteOffset+3])
return Float(bitPattern: floatBytes)
}
Which works fine but is a little cumbersome. I've tried casting my bytes to a UnsafeBufferPointer<UInt32>
but the compiler doesn't like that.
Is there a cleaner way to extract this bit pattern that I'm missing?
Thanks
Upvotes: 1
Views: 181
Reputation: 539745
A possible solution:
let byteOffset = 1
let result = Data(bytes: floatData[byteOffset..<byteOffset+4]).withUnsafeBytes {
(ptr: UnsafePointer<UInt32>) in
Float(bitPattern: UInt32(bigEndian: ptr.pointee))
}
print(result) // 1.5
The Data
object is created from an array slice, and the
closure is called with a pointer to the data.
ptr.pointee
dereferences the data, and UInt32(bigEndian:)
converts it to host byte order.
Upvotes: 3