SF-0033: Implement String.Encoding.ianaName and String.Encoding(ianaName:).

Sources/FoundationEssentials/String/CMakeLists.txt

@@ -15,10 +15,12 @@
 target_sources(FoundationEssentials PRIVATE
     BidirectionalCollection.swift
     BuiltInUnicodeScalarSet.swift
+    IANACharsetNames.swift
     RegexPatternCache.swift
     String+Bridging.swift
     String+Comparison.swift
     String+Encoding.swift
+    String+Encoding+Names.swift
     String+EndianAdaptorSequence.swift
     String+Essentials.swift
     String+IO.swift

Sources/FoundationEssentials/String/IANACharsetNames.swift

@@ -0,0 +1,213 @@
+//===----------------------------------------------------------------------===//
+//
+// This source file is part of the Swift.org open source project
+//
+// Copyright (c) 2025 Apple Inc. and the Swift project authors
+// Licensed under Apache License v2.0 with Runtime Library Exception
+//
+// See https://swift.org/LICENSE.txt for license information
+// See https://swift.org/CONTRIBUTORS.txt for the list of Swift project authors
+//
+//===----------------------------------------------------------------------===//
+
+
+// WARNING: DO NOT EDIT THIS FILE DIRECTLY.
+//          This is auto-generated by `update-iana-charset-names`.
+
+
+extension IANACharset {
+    /// IANA Charset `US-ASCII`.
+    static let usASCII = IANACharset(
+        preferredMIMEName: "US-ASCII",
+        name: "US-ASCII",
+        aliases: [
+            "iso-ir-6",
+            "ANSI_X3.4-1968",
+            "ANSI_X3.4-1986",
+            "ISO_646.irv:1991",
+            "ISO646-US",
+            "US-ASCII",
+            "us",
+            "IBM367",
+            "cp367",
+            "csASCII",
+        ]
+    )
+
+    /// IANA Charset `ISO-8859-1`.
+    static let iso8859_1 = IANACharset(
+        preferredMIMEName: "ISO-8859-1",
+        name: "ISO_8859-1:1987",
+        aliases: [
+            "iso-ir-100",
+            "ISO_8859-1",
+            "ISO-8859-1",
+            "latin1",
+            "l1",
+            "IBM819",
+            "CP819",
+            "csISOLatin1",
+        ]
+    )
+
+    /// IANA Charset `ISO-8859-2`.
+    static let iso8859_2 = IANACharset(
+        preferredMIMEName: "ISO-8859-2",
+        name: "ISO_8859-2:1987",
+        aliases: [
+            "iso-ir-101",
+            "ISO_8859-2",
+            "ISO-8859-2",
+            "latin2",
+            "l2",
+            "csISOLatin2",
+        ]
+    )
+
+    /// IANA Charset `Shift_JIS`.
+    static let shiftJIS = IANACharset(
+        preferredMIMEName: "Shift_JIS",
+        name: "Shift_JIS",
+        aliases: [
+            "MS_Kanji",
+            "csShiftJIS",
+        ]
+    )
+
+    /// IANA Charset `EUC-JP`.
+    static let eucJP = IANACharset(
+        preferredMIMEName: "EUC-JP",
+        name: "Extended_UNIX_Code_Packed_Format_for_Japanese",
+        aliases: [
+            "csEUCPkdFmtJapanese",
+            "EUC-JP",
+        ]
+    )
+
+    /// IANA Charset `ISO-2022-JP`.
+    static let iso2022JP = IANACharset(
+        preferredMIMEName: "ISO-2022-JP",
+        name: "ISO-2022-JP",
+        aliases: [
+            "csISO2022JP",
+        ]
+    )
+
+    /// IANA Charset `UTF-8`.
+    static let utf8 = IANACharset(
+        preferredMIMEName: nil,
+        name: "UTF-8",
+        aliases: [
+            "csUTF8",
+        ]
+    )
+
+    /// IANA Charset `UTF-16BE`.
+    static let utf16BE = IANACharset(
+        preferredMIMEName: nil,
+        name: "UTF-16BE",
+        aliases: [
+            "csUTF16BE",
+        ]
+    )
+
+    /// IANA Charset `UTF-16LE`.
+    static let utf16LE = IANACharset(
+        preferredMIMEName: nil,
+        name: "UTF-16LE",
+        aliases: [
+            "csUTF16LE",
+        ]
+    )
+
+    /// IANA Charset `UTF-16`.
+    static let utf16 = IANACharset(
+        preferredMIMEName: nil,
+        name: "UTF-16",
+        aliases: [
+            "csUTF16",
+        ]
+    )
+
+    /// IANA Charset `UTF-32`.
+    static let utf32 = IANACharset(
+        preferredMIMEName: nil,
+        name: "UTF-32",
+        aliases: [
+            "csUTF32",
+        ]
+    )
+
+    /// IANA Charset `UTF-32BE`.
+    static let utf32BE = IANACharset(
+        preferredMIMEName: nil,
+        name: "UTF-32BE",
+        aliases: [
+            "csUTF32BE",
+        ]
+    )
+
+    /// IANA Charset `UTF-32LE`.
+    static let utf32LE = IANACharset(
+        preferredMIMEName: nil,
+        name: "UTF-32LE",
+        aliases: [
+            "csUTF32LE",
+        ]
+    )
+
+    /// IANA Charset `macintosh`.
+    static let macintosh = IANACharset(
+        preferredMIMEName: nil,
+        name: "macintosh",
+        aliases: [
+            "mac",
+            "csMacintosh",
+        ]
+    )
+
+    /// IANA Charset `windows-1250`.
+    static let windows1250 = IANACharset(
+        preferredMIMEName: nil,
+        name: "windows-1250",
+        aliases: [
+            "cswindows1250",
+        ]
+    )
+
+    /// IANA Charset `windows-1251`.
+    static let windows1251 = IANACharset(
+        preferredMIMEName: nil,
+        name: "windows-1251",
+        aliases: [
+            "cswindows1251",
+        ]
+    )
+
+    /// IANA Charset `windows-1252`.
+    static let windows1252 = IANACharset(
+        preferredMIMEName: nil,
+        name: "windows-1252",
+        aliases: [
+            "cswindows1252",
+        ]
+    )
+
+    /// IANA Charset `windows-1253`.
+    static let windows1253 = IANACharset(
+        preferredMIMEName: nil,
+        name: "windows-1253",
+        aliases: [
+            "cswindows1253",
+        ]
+    )
+
+    /// IANA Charset `windows-1254`.
+    static let windows1254 = IANACharset(
+        preferredMIMEName: nil,
+        name: "windows-1254",
+        aliases: [
+            "cswindows1254",
+        ]
+    )
+}

Sources/FoundationEssentials/String/String+Encoding+Names.swift

@@ -0,0 +1,164 @@
+//===----------------------------------------------------------------------===//
+//
+// This source file is part of the Swift.org open source project
+//
+// Copyright (c) 2025 Apple Inc. and the Swift project authors
+// Licensed under Apache License v2.0 with Runtime Library Exception
+//
+// See https://swift.org/LICENSE.txt for license information
+// See https://swift.org/CONTRIBUTORS.txt for the list of Swift project authors
+//
+//===----------------------------------------------------------------------===//
+
+
+// MARK: - Private extensions for parsing encoding names
+
+private extension UTF8.CodeUnit {
+    func _isASCIICaseinsensitivelyEqual(to other: UTF8.CodeUnit) -> Bool {
+        return switch self {
+        case other, other._uppercased, other._lowercased: true
+        default: false
+        }
+    }
+}
+
+private extension String {
+    func _isASCIICaseinsensitivelyEqual(to other: String) -> Bool {
+        let (myUTF8, otherUTF8) = (self.utf8, other.utf8)
+        var (myIndex, otherIndex) = (myUTF8.startIndex, otherUTF8.startIndex)
+        while myIndex < myUTF8.endIndex && otherIndex < otherUTF8.endIndex {
+            guard myUTF8[myIndex]._isASCIICaseinsensitivelyEqual(to: otherUTF8[otherIndex]) else {
+                return false
+            }
+
+            myUTF8.formIndex(after: &myIndex)
+            otherUTF8.formIndex(after: &otherIndex)
+        }
+        return myIndex == myUTF8.endIndex && otherIndex == otherUTF8.endIndex
+    }
+}
+
+
+// MARK: - IANA Charset Names
+
+/// Info about IANA Charset.
+internal struct IANACharset {
+    /// Preferred MIME Name
+    let preferredMIMEName: String?
+
+    /// The name of this charset
+    let name: String
+
+    /// The aliases of this charset
+    let aliases: Array<String>
+
+    var representativeName: String {
+        return preferredMIMEName ?? name
+    }
+
+    init(preferredMIMEName: String?, name: String, aliases: Array<String>) {
+        self.preferredMIMEName = preferredMIMEName
+        self.name = name
+        self.aliases = aliases
+    }
+
+    func matches(_ string: String) -> Bool {
+        if let preferredMIMEName = self.preferredMIMEName,
+           preferredMIMEName._isASCIICaseinsensitivelyEqual(to: string) {
+            return true
+        }
+        if name._isASCIICaseinsensitivelyEqual(to: string) {
+            return true
+        }
+        for alias in aliases {
+            if alias._isASCIICaseinsensitivelyEqual(to: string) {
+                return true
+            }
+        }
+        return false
+    }
+}
+
+
+// MARK: - `String.Encoding` Names
+
+extension String.Encoding {
+    private var _ianaCharset: IANACharset? {
+        switch self {
+        case .utf8: .utf8
+        case .ascii: .usASCII
+        case .japaneseEUC: .eucJP
+        case .isoLatin1: .iso8859_1
+        case .shiftJIS: .shiftJIS
+        case .isoLatin2: .iso8859_2
+        case .unicode: .utf16
+        case .windowsCP1251: .windows1251
+        case .windowsCP1252: .windows1252
+        case .windowsCP1253: .windows1253
+        case .windowsCP1254: .windows1254
+        case .windowsCP1250: .windows1250
+        case .iso2022JP: .iso2022JP
+        case .macOSRoman: .macintosh
+        case .utf16BigEndian: .utf16BE
+        case .utf16LittleEndian: .utf16LE
+        case .utf32: .utf32
+        case .utf32BigEndian: .utf32BE
+        case .utf32LittleEndian: .utf32LE
+        default: nil
+        }
+    }
+
+    /// The name of this encoding that is compatible with the one of the IANA registry "charset".
+    @available(FoundationPreview 6.3, *)
+    public var ianaName: String? {
+        return _ianaCharset?.representativeName
+    }
+
+    /// Creates an instance from the name of the IANA registry "charset".
+    ///
+    /// - Note: The given name is compared to each IANA "charset" name
+    ///         with ASCII case-insensitive collation
+    ///         to determine which encoding is suitable.
+    @available(FoundationPreview 6.3, *)
+    public init?(ianaName charsetName: String) {
+        let possibilities: [String.Encoding] = [
+            .utf8,
+            .ascii,
+            .japaneseEUC,
+            .isoLatin1,
+            .shiftJIS,
+            .isoLatin2,
+            .unicode, // .utf16
+            .windowsCP1251,
+            .windowsCP1252,
+            .windowsCP1253,
+            .windowsCP1254,
+            .windowsCP1250,
+            .iso2022JP,
+            .macOSRoman,
+            .utf16BigEndian,
+            .utf16LittleEndian,
+            .utf32,
+            .utf32BigEndian,
+            .utf32LittleEndian,
+        ]
+
+        func __determineEncoding() -> String.Encoding? {
+            for encoding in possibilities {
+                guard let ianaCharset = encoding._ianaCharset else {
+                    continue
+                }
+                if ianaCharset.matches(charsetName) {
+                    return encoding
+                }
+            }
+            return nil
+        }
+
+        guard let encoding = __determineEncoding() else {
+            return nil
+        }
+        self = encoding
+    }
+}
+