Decoding quoted-printable messages in Swift

Question

I have a quoted-printable string such as "The cost would be =C2=A31,000". How do I convert this to "The cost would be £1,000".

I'm just converting text manually at the moment and this doesn't cover all cases. I'm sure there is just one line of code that will help with this.

Here is my code:

func decodeUTF8(message: String) -> String
{
    var newMessage = message.stringByReplacingOccurrencesOfString("=2E", withString: ".", options: NSStringCompareOptions.LiteralSearch, range: nil)
    newMessage = newMessage.stringByReplacingOccurrencesOfString("=E2=80=A2", withString: "•", options: NSStringCompareOptions.LiteralSearch, range: nil)
    newMessage = newMessage.stringByReplacingOccurrencesOfString("=C2=A3", withString: "£", options: NSStringCompareOptions.LiteralSearch, range: nil)
    newMessage = newMessage.stringByReplacingOccurrencesOfString("=A3", withString: "£", options: NSStringCompareOptions.LiteralSearch, range: nil)
    newMessage = newMessage.stringByReplacingOccurrencesOfString("=E2=80=9C", withString: "\"", options: NSStringCompareOptions.LiteralSearch, range: nil)
    newMessage = newMessage.stringByReplacingOccurrencesOfString("=E2=80=A6", withString: "…", options: NSStringCompareOptions.LiteralSearch, range: nil)
    newMessage = newMessage.stringByReplacingOccurrencesOfString("=E2=80=9D", withString: "\"", options: NSStringCompareOptions.LiteralSearch, range: nil)
    newMessage = newMessage.stringByReplacingOccurrencesOfString("=92", withString: "'", options: NSStringCompareOptions.LiteralSearch, range: nil)
    newMessage = newMessage.stringByReplacingOccurrencesOfString("=3D", withString: "=", options: NSStringCompareOptions.LiteralSearch, range: nil)
    newMessage = newMessage.stringByReplacingOccurrencesOfString("=20", withString: "", options: NSStringCompareOptions.LiteralSearch, range: nil)
    newMessage = newMessage.stringByReplacingOccurrencesOfString("=E2=80=99", withString: "'", options: NSStringCompareOptions.LiteralSearch, range: nil)

    return newMessage
}

Thanks

Answer 1

An easy way would be to utilize the (NS)String method stringByRemovingPercentEncoding for this purpose. This was observed in decoding quoted-printables , so the first solution is mainly a translation of the answers in that thread to Swift.

The idea is to replace the quoted-printable "=NN" encoding by the percent encoding "%NN" and then use the existing method to remove the percent encoding.

Continuation lines are handled separately. Also, percent characters in the input string must be encoded first, otherwise they would be treated as the leading character in a percent encoding.

func decodeQuotedPrintable(message : String) -> String? {
    return message
        .stringByReplacingOccurrencesOfString("=\r\n", withString: "")
        .stringByReplacingOccurrencesOfString("=\n", withString: "")
        .stringByReplacingOccurrencesOfString("%", withString: "%25")
        .stringByReplacingOccurrencesOfString("=", withString: "%")
        .stringByRemovingPercentEncoding
}

The function returns an optional string which is nil for invalid input. Invalid input can be:

• A "=" character which is not followed by two hexadecimal digits, eg "=XX".
• A "=NN" sequence which does not decode to a valid UTF-8 sequence, eg "=E2=64".

Examples:

if let decoded = decodeQuotedPrintable("=C2=A31,000") {
    print(decoded) // £1,000
}

if let decoded = decodeQuotedPrintable("=E2=80=9CHello =E2=80=A6 world!=E2=80=9D") {
    print(decoded) // “Hello … world!”
}

---

Update 1: The above code assumes that the message uses the UTF-8 encoding for quoting non-ASCII characters, as in most of your examples: C2 A3 is the UTF-8 encoding for "£", E2 80 A4 is the UTF-8 encoding for … .

If the input is "Rub=E9n" then the message is using the Windows-1252 encoding. To decode that correctly, you have to replace

.stringByRemovingPercentEncoding

by

.stringByReplacingPercentEscapesUsingEncoding(NSWindowsCP1252StringEncoding)

There are also ways to detect the encoding from a "Content-Type" header field, compare eg https://stackoverflow.com/a/32051684/1187415 .

---

Update 2: The stringByReplacingPercentEscapesUsingEncoding method is marked as deprecated, so the above code will always generate a compiler warning. Unfortunately, it seems that no alternative method has been provided by Apple.

So here is a new, completely self-contained decoding method which does not cause any compiler warning. This time I have written it as an extension method for String . Explaining comments are in the code.

extension String {

    /// Returns a new string made by removing in the `String` all "soft line
    /// breaks" and replacing all quoted-printable escape sequences with the
    /// matching characters as determined by a given encoding. 
    /// - parameter encoding:     A string encoding. The default is UTF-8.
    /// - returns:                The decoded string, or `nil` for invalid input.

    func decodeQuotedPrintable(encoding enc : NSStringEncoding = NSUTF8StringEncoding) -> String? {

        // Handle soft line breaks, then replace quoted-printable escape sequences. 
        return self
            .stringByReplacingOccurrencesOfString("=\r\n", withString: "")
            .stringByReplacingOccurrencesOfString("=\n", withString: "")
            .decodeQuotedPrintableSequences(enc)
    }

    /// Helper function doing the real work.
    /// Decode all "=HH" sequences with respect to the given encoding.

    private func decodeQuotedPrintableSequences(enc : NSStringEncoding) -> String? {

        var result = ""
        var position = startIndex

        // Find the next "=" and copy characters preceding it to the result:
        while let range = rangeOfString("=", range: position ..< endIndex) {
            result.appendContentsOf(self[position ..< range.startIndex])
            position = range.startIndex

            // Decode one or more successive "=HH" sequences to a byte array:
            let bytes = NSMutableData()
            repeat {
                let hexCode = self[position.advancedBy(1) ..< position.advancedBy(3, limit: endIndex)]
                if hexCode.characters.count < 2 {
                    return nil // Incomplete hex code
                }
                guard var byte = UInt8(hexCode, radix: 16) else {
                    return nil // Invalid hex code
                }
                bytes.appendBytes(&byte, length: 1)
                position = position.advancedBy(3)
            } while position != endIndex && self[position] == "="

            // Convert the byte array to a string, and append it to the result:
            guard let dec = String(data: bytes, encoding: enc) else {
                return nil // Decoded bytes not valid in the given encoding
            }
            result.appendContentsOf(dec)
        }

        // Copy remaining characters to the result:
        result.appendContentsOf(self[position ..< endIndex])

        return result
    }
}

Example usage:

if let decoded = "=C2=A31,000".decodeQuotedPrintable() {
    print(decoded) // £1,000
}

if let decoded = "=E2=80=9CHello =E2=80=A6 world!=E2=80=9D".decodeQuotedPrintable() {
    print(decoded) // “Hello … world!”
}

if let decoded = "Rub=E9n".decodeQuotedPrintable(encoding: NSWindowsCP1252StringEncoding) {
    print(decoded) // Rubén
}

---

Update for Swift 4 (and later):

extension String {

    /// Returns a new string made by removing in the `String` all "soft line
    /// breaks" and replacing all quoted-printable escape sequences with the
    /// matching characters as determined by a given encoding.
    /// - parameter encoding:     A string encoding. The default is UTF-8.
    /// - returns:                The decoded string, or `nil` for invalid input.

    func decodeQuotedPrintable(encoding enc : String.Encoding = .utf8) -> String? {

        // Handle soft line breaks, then replace quoted-printable escape sequences.
        return self
            .replacingOccurrences(of: "=\r\n", with: "")
            .replacingOccurrences(of: "=\n", with: "")
            .decodeQuotedPrintableSequences(encoding: enc)
    }

    /// Helper function doing the real work.
    /// Decode all "=HH" sequences with respect to the given encoding.

    private func decodeQuotedPrintableSequences(encoding enc : String.Encoding) -> String? {

        var result = ""
        var position = startIndex

        // Find the next "=" and copy characters preceding it to the result:
        while let range = range(of: "=", range: position..<endIndex) {
            result.append(contentsOf: self[position ..< range.lowerBound])
            position = range.lowerBound

            // Decode one or more successive "=HH" sequences to a byte array:
            var bytes = Data()
            repeat {
                let hexCode = self[position...].dropFirst().prefix(2)
                if hexCode.count < 2 {
                    return nil // Incomplete hex code
                }
                guard let byte = UInt8(hexCode, radix: 16) else {
                    return nil // Invalid hex code
                }
                bytes.append(byte)
                position = index(position, offsetBy: 3)
            } while position != endIndex && self[position] == "="

            // Convert the byte array to a string, and append it to the result:
            guard let dec = String(data: bytes, encoding: enc) else {
                return nil // Decoded bytes not valid in the given encoding
            }
            result.append(contentsOf: dec)
        }

        // Copy remaining characters to the result:
        result.append(contentsOf: self[position ..< endIndex])

        return result
    }
}

Example usage:

if let decoded = "=C2=A31,000".decodeQuotedPrintable() {
    print(decoded) // £1,000
}

if let decoded = "=E2=80=9CHello =E2=80=A6 world!=E2=80=9D".decodeQuotedPrintable() {
    print(decoded) // “Hello … world!”
}

if let decoded = "Rub=E9n".decodeQuotedPrintable(encoding: .windowsCP1252) {
    print(decoded) // Rubén
}

Answer 2

Unfortunately, I'm a bit late with my answer. It might be helpful for the others though.

var string = "The cost would be =C2=A31,000"

var finalString: String? = nil

if let regEx = try? NSRegularExpression(pattern: "={1}?([a-f0-9]{2}?)", options: NSRegularExpressionOptions.CaseInsensitive)
{
    let intermediatePercentEscapedString = regEx.stringByReplacingMatchesInString(string, options: NSMatchingOptions.WithTransparentBounds, range: NSMakeRange(0, string.characters.count), withTemplate: "%$1")
    print(intermediatePercentEscapedString)
    finalString = intermediatePercentEscapedString.stringByRemovingPercentEncoding
    print(finalString)
}

Answer 3

这种编码称为'quoted-printable'，你需要做的是使用ASCII编码将字符串转换为NSData，然后迭代数据，用字节/字符0xA3替换所有3个符号的方，如'= A3'，然后使用NSUTF8StringEncoding将结果数据转换为字符串。

Answer 4

In order to give an applicable solution, a few more information is required. So, I will make some assumptions.

In an HTML or Mail message for example, you can apply one or more encodings to some kind of source data. For example, you could encode a binary file eg an png file with base64 and then zip it. The order is important.

In your example as you say, the source data is a String and has been encoded via UTF-8.

In a HTPP message, your Content-Type is thus text/plain; charset = UTF-8 text/plain; charset = UTF-8 . In your example there seems also an additional encoding applied, a "Content-Transfer-Encoding": possibly Content-transfer-encoding is quoted-printable or base64 (not sure about that, though).

In order to revert it back, you would need to apply the corresponding decodings in reverse order.

Hint :

You can view the headers ( Contente-type and Content-Transfer-Encoding ) of a mail message when viewing the raw source of the mail.

Answer 5

您还可以查看此工作解决方案 - https://github.com/dunkelstern/QuotedPrintable

let result = QuotedPrintable.decode(string: quoted)