[5.7] Update API for congruence with Unicode proposal

Sources/RegexBuilder/CharacterClass.swift

@@ -60,6 +60,10 @@ extension RegexComponent where Self == CharacterClass {
   public static var anyGrapheme: CharacterClass {
     .init(unconverted: .anyGrapheme)
   }
+
+  public static var anyUnicodeScalar: CharacterClass {
+    .init(unconverted: .anyUnicodeScalar)
+  }
 
   public static var whitespace: CharacterClass {
     .init(unconverted: .whitespace)

Sources/RegexBuilder/DSL.swift

@@ -94,40 +94,20 @@ extension UnicodeScalar: RegexComponent {
 
 // Note: Quantifiers are currently gyb'd.
 
-/// Specifies how much to attempt to match when using a quantifier.
-@available(SwiftStdlib 5.7, *)
-public struct QuantificationBehavior {
-  internal enum Kind {
-    case eagerly
-    case reluctantly
-    case possessively
-  }
-
-  var kind: Kind
-
-  internal var astKind: AST.Quantification.Kind {
-    switch kind {
-    case .eagerly: return .eager
-    case .reluctantly: return .reluctant
-    case .possessively: return .possessive
-    }
-  }
-}
-
 extension DSLTree.Node {
   /// Generates a DSLTree node for a repeated range of the given DSLTree node.
   /// Individual public API functions are in the generated Variadics.swift file.
   @available(SwiftStdlib 5.7, *)
   static func repeating(
     _ range: Range<Int>,
-    _ behavior: QuantificationBehavior?,
+    _ behavior: RegexRepetitionBehavior?,
     _ node: DSLTree.Node
   ) -> DSLTree.Node {
     // TODO: Throw these as errors
     assert(range.lowerBound >= 0, "Cannot specify a negative lower bound")
     assert(!range.isEmpty, "Cannot specify an empty range")
 
-    let kind: DSLTree.QuantificationKind = behavior.map { .explicit($0.astKind) } ?? .default
+    let kind: DSLTree.QuantificationKind = behavior.map { .explicit($0.dslTreeKind) } ?? .default
 
     switch (range.lowerBound, range.upperBound) {
     case (0, Int.max): // 0...
@@ -147,26 +127,6 @@ extension DSLTree.Node {
   }
 }
 
-@available(SwiftStdlib 5.7, *)
-extension QuantificationBehavior {
-  /// Match as much of the input string as possible, backtracking when
-  /// necessary.
-  public static var eagerly: QuantificationBehavior {
-    .init(kind: .eagerly)
-  }
-
-  /// Match as little of the input string as possible, expanding the matched
-  /// region as necessary to complete a match.
-  public static var reluctantly: QuantificationBehavior {
-    .init(kind: .reluctantly)
-  }
-
-  /// Match as much of the input string as possible, performing no backtracking.
-  public static var possessively: QuantificationBehavior {
-    .init(kind: .possessively)
-  }
-}
-
 @available(SwiftStdlib 5.7, *)
 public struct OneOrMore<Output>: _BuiltinRegexComponent {
   public var regex: Regex<Output>

Sources/VariadicsGenerator/VariadicsGenerator.swift

@@ -378,9 +378,9 @@ struct VariadicsGenerator: ParsableCommand {
         \(params.disfavored)\
         public init<\(params.genericParams)>(
           _ component: Component,
-          _ behavior: QuantificationBehavior? = nil
+          _ behavior: RegexRepetitionBehavior? = nil
         ) \(params.whereClauseForInit) {
-          let kind: DSLTree.QuantificationKind = behavior.map { .explicit($0.astKind) } ?? .default
+          let kind: DSLTree.QuantificationKind = behavior.map { .explicit($0.dslTreeKind) } ?? .default
           self.init(node: .quantification(.\(kind.astQuantifierAmount), kind, component.regex.root))
         }
       }
@@ -390,10 +390,10 @@ struct VariadicsGenerator: ParsableCommand {
         \(defaultAvailableAttr)
         \(params.disfavored)\
         public init<\(params.genericParams)>(
-          _ behavior: QuantificationBehavior? = nil,
+          _ behavior: RegexRepetitionBehavior? = nil,
           @\(concatBuilderName) _ component: () -> Component
         ) \(params.whereClauseForInit) {
-          let kind: DSLTree.QuantificationKind = behavior.map { .explicit($0.astKind) } ?? .default
+          let kind: DSLTree.QuantificationKind = behavior.map { .explicit($0.dslTreeKind) } ?? .default
           self.init(node: .quantification(.\(kind.astQuantifierAmount), kind, component().regex.root))
         }
       }
@@ -509,7 +509,7 @@ struct VariadicsGenerator: ParsableCommand {
         public init<\(params.genericParams), R: RangeExpression>(
           _ component: Component,
           _ expression: R,
-          _ behavior: QuantificationBehavior? = nil
+          _ behavior: RegexRepetitionBehavior? = nil
         ) \(params.repeatingWhereClause) {
           self.init(node: .repeating(expression.relative(to: 0..<Int.max), behavior, component.regex.root))
         }

Sources/_RegexParser/Regex/AST/MatchingOptions.swift

@@ -41,7 +41,11 @@ extension AST {
       case graphemeClusterSemantics // X
       case unicodeScalarSemantics   // u
       case byteSemantics            // b
+
+      // Swift-only default possessive quantifier
+      case possessiveByDefault      // t.b.d.
     }
+
     public var kind: Kind
     public var location: SourceLocation
 

Sources/_StringProcessing/ByteCodeGen.swift

@@ -371,9 +371,7 @@ extension Compiler.ByteCodeGen {
     case .syntax(let kind):
       updatedKind = kind.applying(options)
     case .default:
-      updatedKind = options.isReluctantByDefault
-        ? .reluctant
-        : .eager
+      updatedKind = options.defaultQuantificationKind
     }
 
     let (low, high) = amount.bounds

Sources/_StringProcessing/MatchingOptions.swift

@@ -22,6 +22,9 @@ struct MatchingOptions {
     // Must contain exactly one of each mutually exclusive group
     assert(stack.last!.intersection(.textSegmentOptions).rawValue.nonzeroBitCount == 1)
     assert(stack.last!.intersection(.semanticMatchingLevels).rawValue.nonzeroBitCount == 1)
+
+    // Must contain at most one quantifier behavior
+    assert(stack.last!.intersection(.repetitionBehaviors).rawValue.nonzeroBitCount <= 1)
   }
 }
 
@@ -63,6 +66,16 @@ extension MatchingOptions {
     stack.last!.contains(.reluctantByDefault)
   }
 
+  var defaultQuantificationKind: AST.Quantification.Kind {
+    if stack.last!.contains(.possessiveByDefault) {
+      return .possessive
+    } else if stack.last!.contains(.reluctantByDefault) {
+      return .reluctant
+    } else {
+      return .eager
+    }
+  }
+
   var dotMatchesNewline: Bool {
     stack.last!.contains(.singleLine)
   }
@@ -150,6 +163,9 @@ extension MatchingOptions {
     case unicodeScalarSemantics
     case byteSemantics
 
+    // Swift-only default possessive quantifier
+    case possessiveByDefault
+
     init?(_ astKind: AST.MatchingOption.Kind) {
       switch astKind {
       case .caseInsensitive:
@@ -184,6 +200,8 @@ extension MatchingOptions {
         self = .unicodeScalarSemantics
       case .byteSemantics:
         self = .byteSemantics
+      case .possessiveByDefault:
+        self = .possessiveByDefault
 
       // Whitespace options are only relevant during parsing, not compilation.
       case .extended, .extraExtended:
@@ -219,6 +237,9 @@ extension MatchingOptions {
       if Self.textSegmentOptions.contains(opt.representation) {
         remove(.textSegmentOptions)
       }
+      if Self.repetitionBehaviors.contains(opt.representation) {
+        remove(.repetitionBehaviors)
+      }
 
       insert(opt.representation)
     }
@@ -241,6 +262,9 @@ extension MatchingOptions {
         guard let opt = Option(opt.kind) else {
           continue
         }
+        if Self.repetitionBehaviors.contains(opt.representation) {
+          remove(.repetitionBehaviors)
+        }
         remove(opt.representation)
       }
     }
@@ -274,7 +298,15 @@ extension MatchingOptions.Representation {
   static var semanticMatchingLevels: Self {
     [.graphemeClusterSemantics, .unicodeScalarSemantics, .byteSemantics]
   }
-
+
+  // Quantification behavior options
+  static var reluctantByDefault: Self { .init(.reluctantByDefault) }
+  static var possessiveByDefault: Self { .init(.possessiveByDefault) }
+
+  static var repetitionBehaviors: Self {
+    [.reluctantByDefault, .possessiveByDefault]
+  }
+
   /// The default set of options.
   static var `default`: Self {
     [.graphemeClusterSemantics, .textSegmentGraphemeMode]

Sources/_StringProcessing/Regex/Options.swift

@@ -58,8 +58,8 @@ extension RegexComponent {
   /// Returns a regular expression where the start and end of input
   /// anchors (`^` and `$`) also match against the start and end of a line.
   ///
-  /// This method corresponds to applying the `m` option in a regular
-  /// expression literal. For this behavior in the `RegexBuilder` syntax, see
+  /// This method corresponds to applying the `m` option in regex syntax. For
+  /// this behavior in the `RegexBuilder` syntax, see
   /// ``Anchor.startOfLine``, ``Anchor.endOfLine``, ``Anchor.startOfInput``,
   /// and ``Anchor.endOfInput``.
   ///
@@ -69,16 +69,22 @@ extension RegexComponent {
     wrapInOption(.multiline, addingIf: matchLineEndings)
   }
 
-  /// Returns a regular expression where quantifiers are reluctant by default
-  /// instead of eager.
+  /// Returns a regular expression where quantifiers use the specified behavior
+  /// by default.
   ///
-  /// This method corresponds to applying the `U` option in a regular
-  /// expression literal.
+  /// This setting does not affect calls to quantifier methods, such as
+  /// `OneOrMore`, that include an explicit `behavior` parameter.
   ///
-  /// - Parameter useReluctantQuantifiers: A Boolean value indicating whether
-  ///   quantifiers should be reluctant by default.
-  public func reluctantQuantifiers(_ useReluctantQuantifiers: Bool = true) -> Regex<RegexOutput> {
-    wrapInOption(.reluctantByDefault, addingIf: useReluctantQuantifiers)
+  /// Passing `.eager` or `.reluctant` to this method corresponds to applying
+  /// the `(?-U)` or `(?U)` option in regex syntax, respectively.
+  ///
+  /// - Parameter behavior: The default behavior to use for quantifiers.
+  public func repetitionBehavior(_ behavior: RegexRepetitionBehavior) -> Regex<RegexOutput> {
+    if behavior == .possessive {
+      return wrapInOption(.possessiveByDefault, addingIf: true)
+    } else {
+      return wrapInOption(.reluctantByDefault, addingIf: behavior == .reluctant)
+    }
   }
 
   /// Returns a regular expression that matches with the specified semantic
@@ -183,6 +189,46 @@ public struct RegexWordBoundaryKind: Hashable {
   }
 }
 
+/// Specifies how much to attempt to match when using a quantifier.
+@available(SwiftStdlib 5.7, *)
+public struct RegexRepetitionBehavior: Hashable {
+  internal enum Kind {
+    case eager
+    case reluctant
+    case possessive
+  }
+
+  var kind: Kind
+
+  @_spi(RegexBuilder) public var dslTreeKind: AST.Quantification.Kind {
+    switch kind {
+    case .eager: return .eager
+    case .reluctant: return .reluctant
+    case .possessive: return .possessive
+    }
+  }
+}
+
+@available(SwiftStdlib 5.7, *)
+extension RegexRepetitionBehavior {
+  /// Match as much of the input string as possible, backtracking when
+  /// necessary.
+  public static var eager: Self {
+    .init(kind: .eager)
+  }
+
+  /// Match as little of the input string as possible, expanding the matched
+  /// region as necessary to complete a match.
+  public static var reluctant: Self {
+    .init(kind: .reluctant)
+  }
+
+  /// Match as much of the input string as possible, performing no backtracking.
+  public static var possessive: Self {
+    .init(kind: .possessive)
+  }
+}
+
 // MARK: - Helper method
 
 @available(SwiftStdlib 5.7, *)

Sources/_StringProcessing/_CharacterClassModel.swift

@@ -33,6 +33,8 @@ public struct _CharacterClassModel: Hashable {
     case any
     /// Any grapheme cluster
     case anyGrapheme
+    /// Any Unicode scalar
+    case anyScalar
     /// Character.isDigit
     case digit
     /// Character.isHexDigit
@@ -155,8 +157,12 @@ public struct _CharacterClassModel: Hashable {
     case .graphemeCluster:
       let c = str[i]
       var matched: Bool
+      var next = str.index(after: i)
       switch cc {
       case .any, .anyGrapheme: matched = true
+      case .anyScalar:
+        matched = true
+        next = str.unicodeScalars.index(after: i)
       case .digit:
         matched = c.isNumber && (c.isASCII || !options.usesASCIIDigits)
       case .hexDigit:
@@ -174,12 +180,13 @@ public struct _CharacterClassModel: Hashable {
       if isInverted {
         matched.toggle()
       }
-      return matched ? str.index(after: i) : nil
+      return matched ? next : nil
     case .unicodeScalar:
       let c = str.unicodeScalars[i]
       var matched: Bool
       switch cc {
       case .any: matched = true
+      case .anyScalar: matched = true
       case .anyGrapheme: fatalError("Not matched in this mode")
       case .digit:
         matched = c.properties.numericType != nil && (c.isASCII || !options.usesASCIIDigits)
@@ -224,6 +231,10 @@ extension _CharacterClassModel {
     .init(cc: .anyGrapheme, matchLevel: .graphemeCluster)
   }
 
+  public static var anyUnicodeScalar: _CharacterClassModel {
+    .init(cc: .any, matchLevel: .unicodeScalar)
+  }
+
   public static var whitespace: _CharacterClassModel {
     .init(cc: .whitespace, matchLevel: .graphemeCluster)
   }
@@ -275,6 +286,7 @@ extension _CharacterClassModel.Representation: CustomStringConvertible {
     switch self {
     case .any: return "<any>"
     case .anyGrapheme: return "<any grapheme>"
+    case .anyScalar: return "<any scalar>"
     case .digit: return "<digit>"
     case .hexDigit: return "<hex digit>"
     case .horizontalWhitespace: return "<horizontal whitespace>"
@@ -431,6 +443,7 @@ extension AST.Atom.EscapedBuiltin {
     case .notWordCharacter: return .word.inverted
 
     case .graphemeCluster: return .anyGrapheme
+    case .trueAnychar: return .anyUnicodeScalar
 
     default:
       return nil