Merge pull request #31 from orlandos-nl/feature/jo/simplify-unicode-d…

…ecode-loop Simplify the unicode decode loop, thereby fixing some Unicode decode errors
orlandos-nl · Jun 27, 2023 · 412a017 · 412a017
2 parents c530e46 + 89a3fd1
commit 412a017
Show file tree

Hide file tree

Showing 2 changed files with 13 additions and 8 deletions.
diff --git a/Sources/IkigaJSON/Core/Bounds.swift b/Sources/IkigaJSON/Core/Bounds.swift
@@ -41,7 +41,6 @@ internal struct Bounds {
 
         // If we can't take a shortcut by decoding immediately thanks to an escaping character
         if escaping || unicode {
-            var length = Int(self.length)
             var i = 0
             var unicodes = [UInt16]()
 
@@ -53,7 +52,7 @@ internal struct Bounds {
                 }
             }
 
-            next: while i < length {
+            next: while i < data.count {
                 let byte = data[i]
 
                 unescape: if escaping {
@@ -68,7 +67,6 @@ internal struct Bounds {
 
                     // Remove the backslash and translate the next character
                     data.remove(at: i)
-                    length = length &- 1
 
                     switch data[i] {
                     case .backslash, .solidus, .quote:
@@ -82,8 +80,7 @@ internal struct Bounds {
                     case .u:
                         // `\u` indicates a unicode character
                         data.remove(at: i)
-                        length = length &- 1
-                        let unicode = try decodeUnicode(from: &data, offset: &i, length: &length)
+                        let unicode = try decodeUnicode(from: &data, offset: &i)
                         unicodes.append(unicode)
 
                         // Continue explicitly, so that we do not trigger the unicode 'flush' flow
@@ -159,9 +156,9 @@ internal struct Bounds {
 
 struct UTF8ParsingError: Error {}
 
-fileprivate func decodeUnicode(from data: inout Data, offset: inout Int, length: inout Int) throws -> UInt16 {
+fileprivate func decodeUnicode(from data: inout Data, offset: inout Int) throws -> UInt16 {
     let hexCharacters = 4
-    guard length - offset >= hexCharacters else {
+    guard data.count - offset >= hexCharacters else {
         throw UTF8ParsingError()
     }
 
@@ -174,7 +171,6 @@ fileprivate func decodeUnicode(from data: inout Data, offset: inout Int, length:
         throw UTF8ParsingError()
     }
 
-    length -= hexCharacters
     var unicode: UInt16 = 0
     unicode += UInt16(hex0) << 12
     unicode += UInt16(hex1) << 8

diff --git a/Tests/IkigaJSONTests/JSONTests.swift b/Tests/IkigaJSONTests/JSONTests.swift
@@ -50,6 +50,15 @@ final class IkigaJSONTests: XCTestCase {
             XCTAssertEqual(result, ["simple": "ßhello world", "complex": "👩‍👩‍👧‍👧hello world"])
         }
     }
+
+    func testEscapedUnicodeWeis() throws {
+        do {
+            let json: Data = #"{"foo":"\u0022wei\u00DF\u0022"}"#.data(using: .utf8)!
+
+            let result = try IkigaJSONDecoder().decode([String: String].self, from: json)
+            XCTAssertEqual(result, ["foo": #""weiß""#])
+        }
+    }
 
     func testPropertyWrapper() throws {
         @propertyWrapper struct FluentPropertyTest<Value: Codable & Equatable>: Codable, Equatable {