How to decode a UTF16 string into a Unicode character

Question

An device encodes a string "" as "\uD83E\uDD1B\uD83C\uDFFD". The hexadecimal numbers represented in this string are from the UTF-16 hex encoding of the character. The Unicode code point U+1F91B, U+1F3FD gets its numbers from the UTF-32 hex encoding.

Taking this later one, in Swift we can do a literal like this "\u{1F91B}\u{1F3FD}" and we will get the character "" as expected.

How can I convert from the UTF-16 hex string "\uD83E\uDD1B\uD83C\uDFFD" to get the ""?

I've tried taking the string and converting it to [UInt32] array of 32 bit integers and then using that to create unicode scalars, but this only works for Unicode characters that can be expressed int a single UTF-32 code point.

Here is the source code I'm using.

extension String {

    func decodeBlock() -> String {
        let strings = self.components(separatedBy: "\\u")
        var scalars : [UInt32] = []
    
        var value: UInt32 = 0
        for string in strings {
            print(string)
            let scanner = Scanner(string: string)
            if scanner.scanHexInt32(&value) {
                scalars.append(value)
            }
        }
    
        let utf32chars = scalars
        var str = ""
        var generator = utf32chars.makeIterator()
        var utf32 : UTF32 = UTF32()
        var done = false
        while !done {
            let r = utf32.decode(&generator)
            switch (r) {
            case . emptyInput:
                done = true
            case .scalarValue(let val):
                str.append(Character(val))
            case .error:
                return "$"
            }
        }
        return str
    

        return self
    }
}

It is adapted from the code in an answer to a similar question. https://stackoverflow.com/a/41412056/731773

The source of the encoded string is the org.apache.commons.lang.StringEscapeUtils escapeJava function which can be found here.

Answer 1

This is a little bit of a cheat but UTF-16 happens to be the encoding used by NSString so you can borrow the methods of NSString to achieve it:

extension String {
    func decodeBlock() -> String? {
        var chars = [unichar]()

        for substr in self.components(separatedBy: "\\u") where !substr.isEmpty {
            if let value = UInt16(substr, radix: 16) {
                chars.append(value)
            } else {
                return nil
            }
        }

        return NSString(characters: chars, length: chars.count) as String
    }
}

if let decoded = "\\uD83E\\uDD1B\\uD83C\\uDFFD".decodeBlock() {
    print(decoded)
} else {
    print("Cannot decode")
}

Answer 2

This is definitely cheating since it just uses the built in method in JavaScript, but it works.

func decode() -> String{
    // getting a JSContext
    let context = JSContext()
    let encodedString = self
    // defining a JavaScript function
    let jsFunctionText = "var decode = function(encodedString) {\n" +
        "var r = /\\\\u([\\d\\w]{4})/gi;\n" +
        " x = encodedString\n" +
        "x = x.replace(r, function (match, grp) {\n" +
        "     return String.fromCharCode(parseInt(grp, 16)); } );\n" +
        " x = unescape(x);\n" +
        " return x\n" +
    "}"
    //    print(jsFunctionText)
    context!.evaluateScript(jsFunctionText)!

    // calling a JavaScript function
    let jsFunction = context?.objectForKeyedSubscript("decode")

    let decodedValue = jsFunction?.call(withArguments: [encodedString]);
    if let decodedString = decodedValue?.toString() {
        return decodedString
    } else {
        return self
    }
}