How to translate push tokens between NSData and NSString representations in Swift? - ios

Question

I have used the following Objective C routines for years, to convert a NSData push token into a NSString (for use by a web side push service), and the inverse, to take a known NSString version of the token and recreate the NSData representation. Now, I find the need for the exact same capabilities, but in Swift.
The dataToHex Objective C code essentially uses printf formatting:
- (NSString *)dataToHex:(NSData *)data
{
NSMutableString *str = [NSMutableString stringWithCapacity:100];
const unsigned char *p = [data bytes];
NSUInteger len = [data length];
for(int i=0; i<len; ++i) {
[str appendFormat:#"%02.2X", p[i]];
}
return str;
}
The inverse translation is:
- (NSData *)hexToData:(NSString *)str
{
const char *ptr = [str cStringUsingEncoding:NSASCIIStringEncoding];
NSUInteger len = [str length]/2;
NSMutableData *data = [NSMutableData dataWithCapacity:len];
while(len--) {
char num[5] = (char[]){ '0', 'x', 0, 0, 0 };
num[2] = *ptr++;
num[3] = *ptr++;
uint8_t n = (uint8_t)strtol(num, NULL, 0);
[data appendBytes:&n length:1];
}
return data;
}
By "cleverly" overwriting a two byes in an ASCII array, the "0xXX" string is converted into a byte, which is then appended to the mutable data object.
Now that I'm coding in Swift, I need the same capabilities but have not found any posts with anything like the code above in Swift.

Answer 1

Converting from the NSData representation as supplied by iOS matches the Objective C code almost line for line:
func dataToHex(data: NSData) -> String
{
var str: String = String()
let p = UnsafePointer<UInt8>(data.bytes)
let len = data.length
for var i=0; i<len; ++i {
str += String(format: "%02.2X", p[i])
}
return str
}
However, given an NSString object, the conversion back to a NSData object is a bit harder. You might need to do this if you are testing in the Simulator, have a string token from a real device, and need it to say register with a service.
The first approach I took tried to replicate code I used before, by creating a string with character pairs, and calling strtol:
func hexToData0(str: NSString) -> NSData {
let len = str.length/2
var data = NSMutableData(capacity:len)!
var num: [Int8] = [ 0, 0, 0 ]
let ptr = str.cStringUsingEncoding(NSUTF8StringEncoding)
for var i=0; i<len; ++i {
num[0] = ptr[i*2+0]
num[1] = ptr[i*2+1]
var n = UInt8 ( strtol(&num, nil, 16) )
data.appendBytes(&n, length:1)
}
return data;
}
I just felt the strtol was a bit of a hack, so I did the same using NSScanner that about the same code size, while most likely less efficient:
func hexToData1(str: NSString) -> NSData {
var data = NSMutableData(capacity: str.length/2)!
for var i = 0; i<str.length; i+=2 {
let r = NSRange(location: i, length: 2)
let s = str.substringWithRange(r)
let sc = NSScanner(string: s)
var val: UInt32 = 0
let ret = sc.scanHexInt(&val)
if ret {
var b = UInt8(val)
data.appendBytes(&b, length: 1)
} else {
assert(false, "Yikes!")
}
}
return data
}
Then, it occurred to me that I could do it all in Swift, no Darwin or Foundation needed, at the expense of a few more lines of code:
// Swift 4
func hexToData(str: String) -> Data {
let len = str.count/2
var data = Data(capacity:len)
let ptr = str.cString(using: String.Encoding.utf8)!
for i in 0..<len {
var num: UInt8 = 0
var multi: UInt8 = 16;
for j in 0..<2 {
let c: UInt8 = UInt8(ptr[i*2+j])
var offset: UInt8 = 0
switch c {
case 48...57: // '0'-'9'
offset = 48
case 65...70: // 'A'-'F'
offset = 65 - 10 // 10 since 'A' is 10, not 0
case 97...102: // 'a'-'f'
offset = 97 - 10 // 10 since 'a' is 10, not 0
default:
assert(false)
}
num += (c - offset)*multi
multi = 1
}
data.append(num)
}
return data;
}
I'm using the final hexToData in my code.