How in swift to convert Int16 to two UInt8 Bytes - swift

I have some binary data that encodes a two byte value as a signed integer.
bytes[1] = 255 // 0xFF
bytes[2] = 251 // 0xF1
Decoding
This is fairly easy - I can extract an Int16 value from these bytes with:
Int16(bytes[1]) << 8 | Int16(bytes[2])
Encoding
This is where I'm running into issues. Most of my data spec called for UInt and that is easy but I'm having trouble extracting the two bytes that make up an Int16
let nv : Int16 = -15
UInt8(nv >> 8) // fail
UInt8(nv) // fail
Question
How would I extract the two bytes that make up an Int16 value

You should work with unsigned integers:
let bytes: [UInt8] = [255, 251]
let uInt16Value = UInt16(bytes[0]) << 8 | UInt16(bytes[1])
let uInt8Value0 = uInt16Value >> 8
let uInt8Value1 = UInt8(uInt16Value & 0x00ff)
If you want to convert UInt16 to bit equivalent Int16 then you can do it with specific initializer:
let int16Value: Int16 = -15
let uInt16Value = UInt16(bitPattern: int16Value)
And vice versa:
let uInt16Value: UInt16 = 65000
let int16Value = Int16(bitPattern: uInt16Value)
In your case:
let nv: Int16 = -15
let uNv = UInt16(bitPattern: nv)
UInt8(uNv >> 8)
UInt8(uNv & 0x00ff)

You could use init(truncatingBitPattern: Int16) initializer:
let nv: Int16 = -15
UInt8(truncatingBitPattern: nv >> 8) // -> 255
UInt8(truncatingBitPattern: nv) // -> 241

I would just do this:
let a = UInt8(nv >> 8 & 0x00ff) // 255
let b = UInt8(nv & 0x00ff) // 241

extension Int16 {
var twoBytes : [UInt8] {
let unsignedSelf = UInt16(bitPattern: self)
return [UInt8(truncatingIfNeeded: unsignedSelf >> 8),
UInt8(truncatingIfNeeded: unsignedSelf)]
}
}
var test : Int16 = -15
test.twoBytes // [255, 241]

Related

CheckSum8 Modulo 256 Swift

I have an array of UInt8 and I want to calculate CheckSum8 Modulo 256.
If sum of bytes is less than 255, checkSum function returns correct value.
e.g
let bytes1 : [UInt8] = [1, 0xa1]
let validCheck = checkSum(data : bytes1) // 162 = 0xa2
let bytes : [UInt8] = [6, 0xB1, 27,0xc5,0xf5,0x9d]
let invalidCheck = checkSum(data : bytes) // 41
Below function returns 41 but expected checksum is 35.
func checkSum(data: [UInt8]) -> UInt8 {
var sum = 0
for i in 0..<data.count {
sum += Int(data[i])
}
let retVal = sum & 0xff
return UInt8(retVal)
}
Your checkSum method is largely right. If you want, you could simplify it to:
func checkSum(_ values: [UInt8]) -> UInt8 {
let result = values.reduce(0) { ($0 + UInt32($1)) & 0xff }
return UInt8(result)
}
You point out a web site that reports the checksum8 for 06B127c5f59d is 35.
The problem is that in your array has 27, not 0x27. If you have hexadecimal values, you always need the 0x prefix for each value in your array literal (or, technically, at least if the value is larger than 9).
So, consider:
let values: [UInt8] = [0x06, 0xB1, 0x27, 0xc5, 0xf5, 0x9d]
let result = checkSum(values)
That’s 53. If you want to see that in hexadecimal (like that site you referred to):
let hex = String(result, radix: 16)
That shows us that the checksum is 0x35 in hexadecimal.

Swift: Byte array to decimal value

In my project, I communicate with a bluetooth device, the bluetooth device must send me a timestamp second, I received in byte:
[2,6,239]
When I convert converted to a string:
let payloadString = payload.map {
String(format: "%02x", $0)
}
Output:
["02", "06","ef"]
When I converted from the website 0206ef = 132847 seconds
How can I directly convert my aray [2,6,239] in second (= 132847 seconds)?
And if it's complicated then translate my array ["02", "06,"ef"] in second (= 132847 seconds)
The payload contains the bytes of the binary representation of the value.
You convert it back to the value by shifting each byte into its corresponding position:
let payload: [UInt8] = [2, 6, 239]
let value = Int(payload[0]) << 16 + Int(payload[1]) << 8 + Int(payload[2])
print(value) // 132847
The important point is to convert the bytes to integers before shifting, otherwise an overflow error would occur. Alternatively,
with multiplication:
let value = (Int(payload[0]) * 256 + Int(payload[1])) * 256 + Int(payload[2])
or
let value = payload.reduce(0) { $0 * 256 + Int($1) }
The last approach works with an arbitrary number of bytes – as long as
the result fits into an Int. For 4...8 bytes you better choose UInt64
to avoid overflow errors:
let value = payload.reduce(0) { $0 * 256 + UInt64($1) }
payloadString string can be reduced to hexStr and then converted to decimal
var payload = [2,6,239];
let payloadString = payload.map {
String(format: "%02x", $0)
}
//let hexStr = payloadString.reduce(""){$0 + $1}
let hexStr = payloadString.joined()
if let value = UInt64(hexStr, radix: 16) {
print(value)//132847
}

Splitting a UInt16 to 2 UInt8 bytes and getting the hexa string of both. Swift

I need 16383 to be converted to 7F7F but I can only get this to be converted to 3fff or 77377.
I can convert 8192 to hexadecimal string 4000 which is essentially the same thing.
If I use let firstHexa = String(format:"%02X", a) It stops at 3fff hexadecimal for the first number and and 2000 hexadecimal for the second number. here is my code
public func intToHexString(_ int: Int16) -> String {
var encodedHexa: String = ""
if int >= -8192 && int <= 8191 {
let int16 = int + 8192
//convert to two unsigned Int8 bytes
let a = UInt8(int16 >> 8 & 0x00ff)
let b = UInt8(int16 & 0x00ff)
//convert the 2 bytes to hexadecimals
let first1Hexa = String(a, radix: 8 )
let second2Hexa = String(b, radix: 8)
let firstHexa = String(format:"%02X", a)
let secondHexa = String(format:"%02X", b)
//combine the 2 hexas into 1 string with 4 characters...adding 0 to the beggining if only 1 character.
if firstHexa.count == 1 {
let appendedFHexa = "0" + firstHexa
encodedHexa = appendedFHexa + secondHexa
} else if secondHexa.count == 1 {
let appendedSHexa = "0" + secondHexa
encodedHexa = firstHexa + appendedSHexa
} else {
encodedHexa = firstHexa + secondHexa
}
}
return encodedHexa
}
Please help ma'ams and sirs! Thanks.
From your test cases, it seems like your values are 7 bits per byte.
You want 8192 to convert to 4000.
You want 16383 to convert to 7F7F.
Note that:
(0x7f << 7) + 0x7f == 16383
Given that:
let a = UInt8((int16 >> 7) & 0x7f)
let b = UInt8(int16 & 0x7f)
let result = String(format: "%02X%02X", a , b)
This gives:
"4000" for 8128
"7F7F" for 16383
To reverse the process:
let str = "7F7F"
let value = Int(str, radix: 16)!
let result = ((value >> 8) & 0x7f) << 7 + (value & 0x7f)
print(result) // 16383

Reading Data into a Struct in Swift

I'm trying to read bare Data into a Swift 4 struct using the withUnsafeBytes method. The problem
The network UDP packet has this format:
data: 0102 0A00 0000 0B00 0000
01 : 1 byte : majorVersion (decimal 01)
02 : 1 byte : minorVersion (decimal 02)
0A00 0000 : 4 bytes: applicationHostId (decimal 10)
0B00 0000 : 4 bytes: versionNumber (decimal 11)
Then I have an extension on Data that takes a start and the length of bytes to read
extension Data {
func scanValue<T>(start: Int, length: Int) -> T {
return self.subdata(in: start..<start+length).withUnsafeBytes { $0.pointee }
}
}
This works correctly when reading the values one by one:
// correctly read as decimal "1"
let majorVersion: UInt8 = data.scanValue(start: 0, length: 1)
// correctly read as decimal "2"
let minorVersion: UInt8 = data.scanValue(start: 1, length: 1)
// correctly read as decimal "10"
let applicationHostId: UInt32 = data.scanValue(start: 2, length: 4)
// correctly read as decimal "11"
let versionNumber: UInt32 = data.scanValue(start: 6, length: 4)
Then I created a struct that represents the entire packet as follows
struct XPLBeacon {
var majorVersion: UInt8 // 1 Byte
var minorVersion: UInt8 // 1 Byte
var applicationHostId: UInt32 // 4 Bytes
var versionNumber: UInt32 // 4 Bytes
}
But when I read the data directly into the structure I have some issues:
var beacon: XPLBeacon = data.scanValue(start: 0, length: data.count)
// correctly read as decimal "1"
beacon.majorVersion
// correctly read as decimal "2"
beacon.minorVersion
// not correctly read
beacon.applicationHostId
// not correctly read
beacon.versionNumber
I it supposed to work to parse an entire struct like this?
Reading the entire structure from the data does not work because
the struct members are padded to their natural boundary. The
memory layout of struct XPLBeacon is
A B x x C C C C D D D D
where
offset member
0 A - majorVersion (UInt8)
1 B - minorVersion (UInt8)
2 x x - padding
4 C C C C - applicationHostId (UInt32)
8 D D D D - versionNumber (UInt32)
and the padding is inserted so that the UInt32 members are
aligned to memory addresses which are a multiple of their size. This is
also confirmed by
print(MemoryLayout<XPLBeacon>.size) // 12
(For more information about alignment in Swift, see
Type Layout).
If you read the entire data into the struct then the bytes are assigned
as follows
01 02 0A 00 00 00 0B 00 00 00
A B x x C C C C D D D D
which explains why major/minorVersion are correct, but applicationHostId and versionNumber
are wrong. Reading all members separately from the data is the correct solution.
Since Swift 3 Data conforms to RandomAccessCollection, MutableCollection, RangeReplaceableCollection. So you can simply create a custom initializer to initialise your struct properties as follow:
struct XPLBeacon {
let majorVersion, minorVersion: UInt8 // 1 + 1 = 2 Bytes
let applicationHostId, versionNumber: UInt32 // 4 + 4 = 8 Bytes
init(data: Data) {
self.majorVersion = data[0]
self.minorVersion = data[1]
self.applicationHostId = data
.subdata(in: 2..<6)
.withUnsafeBytes { $0.load(as: UInt32.self) }
self.versionNumber = data
.subdata(in: 6..<10)
.withUnsafeBytes { $0.load(as: UInt32.self) }
}
}
var data = Data([0x01,0x02, 0x0A, 0x00, 0x00, 0x00, 0x0B, 0x00, 0x00,0x00])
print(data as NSData) // "{length = 10, bytes = 0x01020a0000000b000000}\n" <01020a00 00000b00 0000>
let beacon = XPLBeacon(data: data)
beacon.majorVersion // 1
beacon.minorVersion // 2
beacon.applicationHostId // 10
beacon.versionNumber // 11
Following on Leo Dabus answer, I created a slightly more readable constructor:
extension Data {
func object<T>(at index: Index) -> T {
subdata(in: index ..< index.advanced(by: MemoryLayout<T>.size))
.withUnsafeBytes { $0.load(as: T.self) }
}
}
struct XPLBeacon {
var majorVersion: UInt8
var minorVersion: UInt8
var applicationHostId: UInt32
var versionNumber: UInt32
init(data: Data) {
var index = data.startIndex
majorVersion = data.object(at: index)
index += MemoryLayout.size(ofValue: majorVersion)
minorVersion = data.object(at: index)
index += MemoryLayout.size(ofValue: minorVersion)
applicationHostId = data.object(at: index)
index += MemoryLayout.size(ofValue: applicationHostId)
versionNumber = data.object(at: index)
}
}
What is not part of this is of course the checking for the correctness of the data. As other mentioned in comments, this could be done either by having a failable init method or by throwing an Error.

Swift concatenating Int16 and Int32 vars into a multi-byte entity

I am writing code to read and write file using a rigidly defined binary format. I have a mixture of Int16 and Int32 values, which are defined as a 240 byte structure. How do I concatenate these values into this single structure?
here is an example, you can custom your own init method, use loop if there two many repeated length
public struct HistoryRecord {
public let sequence: UInt8
let impFlag: UInt8
let mode: UInt8
let power: UInt16
let utc: UInt32
let weight: UInt16
let impedance: UInt16
private let sequenceLength = 1
private let modeLength = 1
private let powerLength = 2
private let utcLength = 4
private let weightLength = 2
private let impedanceLength = 2
public init() {
sequence = 0
impFlag = 0
mode = 0
power = 0
utc = 0
weight = 0
impedance = 0
}
public init?(_ data: NSData) {
let needLength = sequenceLength + modeLength + powerLength + utcLength + weightLength + impedanceLength
guard data.length == needLength else {
return nil
}
var bufferUInt8: UInt8
var bufferUInt16: UInt16
var bufferUInt32: UInt32
bufferUInt8 = 0
data.getBytes(&bufferUInt8, range: NSMakeRange(0, sequenceLength))
sequence = bufferUInt8
bufferUInt8 = 0
data.getBytes(&bufferUInt8, range: NSMakeRange(sequenceLength, modeLength))
impFlag = (bufferUInt8 >> 4) & 0x00ff
mode = bufferUInt8 & 0x00ff
bufferUInt16 = 0
data.getBytes(&bufferUInt16, range: NSMakeRange(sequenceLength + modeLength, powerLength))
power = bufferUInt16.bigEndian
bufferUInt32 = 0
data.getBytes(&bufferUInt32, range: NSMakeRange(sequenceLength + modeLength + powerLength, utcLength))
utc = bufferUInt32.bigEndian
bufferUInt16 = 0
data.getBytes(&bufferUInt16, range: NSMakeRange(sequenceLength + modeLength + powerLength + utcLength, weightLength))
weight = bufferUInt16.bigEndian
bufferUInt16 = 0
data.getBytes(&bufferUInt16, range: NSMakeRange(sequenceLength + modeLength + powerLength + utcLength + weightLength, impedanceLength))
impedance = bufferUInt16.bigEndian