Merge pull request #597 from rctcwyvrn/backreference-in-scalar-mode

Fix matching of backreferences in scalar mode

Author: rctcwyvrn

Sources/_StringProcessing/ByteCodeGen.swift

@@ -84,7 +84,8 @@ fileprivate extension Compiler.ByteCodeGen {
       try emitBackreference(ref.ast)
 
     case let .symbolicReference(id):
-      builder.buildUnresolvedReference(id: id)
+      builder.buildUnresolvedReference(
+        id: id, isScalarMode: options.semanticLevel == .unicodeScalar)
 
     case let .changeMatchingOptions(optionSequence):
       if !hasEmittedFirstMatchableAtom {
@@ -143,9 +144,11 @@ fileprivate extension Compiler.ByteCodeGen {
       guard let i = n.value else {
         throw Unreachable("Expected a value")
       }
-      builder.buildBackreference(.init(i))
+      builder.buildBackreference(
+        .init(i), isScalarMode: options.semanticLevel == .unicodeScalar)
     case .named(let name):
-      try builder.buildNamedReference(name)
+      try builder.buildNamedReference(
+        name, isScalarMode: options.semanticLevel == .unicodeScalar)
     case .relative:
       throw Unsupported("Backreference kind: \(ref)")
     }

Sources/_StringProcessing/Engine/InstPayload.swift

@@ -240,14 +240,20 @@ extension Instruction.Payload {
     interpret()
   }
 
+  init(capture: CaptureRegister, isScalarMode: Bool) {
+    self.init(isScalarMode ? 1 : 0, capture)
+  }
+  var captureAndMode: (isScalarMode: Bool, CaptureRegister) {
+    let pair: (UInt64, CaptureRegister) = interpretPair()
+    return (pair.0 == 1, pair.1)
+  }
   init(capture: CaptureRegister) {
     self.init(capture)
   }
   var capture: CaptureRegister {
     interpret()
   }
 
-
   // MARK: Packed operand payloads
 
   init(immediate: UInt64, int: IntRegister) {

Sources/_StringProcessing/Engine/MEBuilder.swift

@@ -292,22 +292,23 @@ extension MEProgram.Builder {
   }
 
   mutating func buildBackreference(
-    _ cap: CaptureRegister
+    _ cap: CaptureRegister,
+    isScalarMode: Bool
   ) {
     instructions.append(
-      .init(.backreference, .init(capture: cap)))
+      .init(.backreference, .init(capture: cap, isScalarMode: isScalarMode)))
   }
 
-  mutating func buildUnresolvedReference(id: ReferenceID) {
-    buildBackreference(.init(0))
+  mutating func buildUnresolvedReference(id: ReferenceID, isScalarMode: Bool) {
+    buildBackreference(.init(0), isScalarMode: isScalarMode)
     unresolvedReferences[id, default: []].append(lastInstructionAddress)
   }
 
-  mutating func buildNamedReference(_ name: String) throws {
+  mutating func buildNamedReference(_ name: String, isScalarMode: Bool) throws {
     guard let index = captureList.indexOfCapture(named: name) else {
       throw RegexCompilationError.uncapturedReference
     }
-    buildBackreference(.init(index))
+    buildBackreference(.init(index), isScalarMode: isScalarMode)
   }
 
   // TODO: Mutating because of fail address fixup, drop when
@@ -456,8 +457,10 @@ fileprivate extension MEProgram.Builder {
         throw RegexCompilationError.uncapturedReference
       }
       for use in uses {
+        let (isScalarMode, _) = instructions[use.rawValue].payload.captureAndMode
         instructions[use.rawValue] =
-          Instruction(.backreference, .init(capture: .init(offset)))
+          Instruction(.backreference,
+            .init(capture: .init(offset), isScalarMode: isScalarMode))
       }
     }
   }

Sources/_StringProcessing/Engine/Processor.swift

@@ -231,9 +231,17 @@ extension Processor {
 
   // Match against the current input prefix. Returns whether
   // it succeeded vs signaling an error.
-  mutating func matchSeq<C: Collection>(
-    _ seq: C
-  ) -> Bool where C.Element == Input.Element {
+  mutating func matchSeq(
+    _ seq: Substring,
+    isScalarMode: Bool
+  ) -> Bool  {
+    if isScalarMode {
+      for s in seq.unicodeScalars {
+        guard matchScalar(s, boundaryCheck: false) else { return false }
+      }
+      return true
+    }
+
     for e in seq {
       guard match(e) else { return false }
     }
@@ -584,8 +592,9 @@ extension Processor {
       }
 
     case .backreference:
+      let (isScalarMode, capture) = payload.captureAndMode
       let capNum = Int(
-        asserting: payload.capture.rawValue)
+        asserting: capture.rawValue)
       guard capNum < storedCaptures.count else {
         fatalError("Should this be an assert?")
       }
@@ -597,7 +606,7 @@ extension Processor {
         signalFailure()
         return
       }
-      if matchSeq(input[range]) {
+      if matchSeq(input[range], isScalarMode: isScalarMode) {
         controller.step()
       }
 

Tests/RegexTests/MatchTests.swift

@@ -1643,6 +1643,11 @@ extension RegexTests {
       (input: "123x23", match: "23x23"),
       xfail: true)
 
+    // Backreferences in scalar mode
+    // In scalar mode the backreference should not match
+    firstMatchTest(#"(.+)\1"#, input: "ée\u{301}", match: "ée\u{301}")
+    firstMatchTest(#"(.+)\1"#, input: "ée\u{301}", match: nil, semanticLevel: .unicodeScalar)
+
     // Backreferences in lookaheads
     firstMatchTests(
       #"^(?=.*(.)(.)\2\1).+$"#,