CODEWITHSUNDEEP
swift
pistolcaffe
pistolcaffe

Reputation: 879

If the text contains emoji, Range is nil from swift

The text "Welcome my application..❣️" does not make sense during the NSRange and Range tests. If ❣️ is included, Range is returned as nil, and I wonder why.

func testA() {
        let testStr = "Welcome my application..❣️"

        let range = NSRange(location: 0, length: testStr.count)
        let wrapRange = Range(range, in: testStr)

        let testStrB = "Welcome my application.."

        let rangeB = NSRange(location: 0, length: testStrB.count)
        let wrapRangeB = Range(rangeB, in: testStrB)

        print("wrapRange: \(wrapRange) wrapRangeB: \(wrapRangeB)")
}

RESULT:

wrapRange: nil wrapRangeB: Optional(Range(Swift.String.Index(_rawBits: 1)..<Swift.String.Index(_rawBits: 1572864)))

Upvotes: 4

Views: 777

Answers (1)

Martin R
Martin R

Reputation: 540105

"❣️" is a single “extended grapheme cluster”, but two UTF-16 code units:

print("❣️".count)       // 1
print("❣️".utf16.count) // 2

NSRange counts UTF-16 code units (which are the “characters” in an NSString) , therefore the correct way to create an NSRange comprising the complete range of a Swift string is

let range = NSRange(location: 0, length: testStr.utf16.count)

or better (since Swift 4):

let range = NSRange(testStr.startIndex..., in: testStr)

Explanation: In your code (simplified here)

let testStr = "❣️"
let range = NSRange(location: 0, length: testStr.count)
print(range) // {0, 1}

creates an NSRange describing a single UTF-16 code unit. This cannot be converted to a Range<String.Index> in testStr because its first Character consists of two UTF-16 code units:

let wrapRange = Range(range, in: testStr)
print(wrapRange)  // nil

Upvotes: 8

Related Questions