Merge pull request #801 from swiftlang/swift61-repeated-case-insensitivity

[6.1] Fix repeated case-insensitive ASCII match

Author: natecook1000

Sources/_StringProcessing/ByteCodeGen.swift

@@ -773,11 +773,19 @@ fileprivate extension Compiler.ByteCodeGen {
     case .atom(let atom):
       switch atom {
       case .char(let c):
-        // Single scalar ascii value character
-        guard let val = c._singleScalarAsciiValue else {
-          return false
+        if options.isCaseInsensitive && c.isCased {
+          // Cased character with case-insensitive matching; match only as an ASCII bitset
+          guard let bitset = DSLTree.CustomCharacterClass(members: [.atom(atom)]).asAsciiBitset(options) else {
+            return false
+          }
+          builder.buildQuantify(bitset: bitset, kind, minTrips, maxExtraTrips, isScalarSemantics: isScalarSemantics)
+        } else {
+          // Uncased character OR case-sensitive matching; match as a single scalar ascii value character
+          guard let val = c._singleScalarAsciiValue else {
+            return false
+          }
+          builder.buildQuantify(asciiChar: val, kind, minTrips, maxExtraTrips, isScalarSemantics: isScalarSemantics)
         }
-        builder.buildQuantify(asciiChar: val, kind, minTrips, maxExtraTrips, isScalarSemantics: isScalarSemantics)
 
       case .any:
         builder.buildQuantifyAny(

Tests/RegexTests/MatchTests.swift

@@ -2193,6 +2193,41 @@ extension RegexTests {
       ("cafe", true),
       ("CaFe", true),
       ("EfAc", true))
+    
+    matchTest(
+      #"(?i)a+b"#,
+      ("ab", true),
+      ("Ab", true),
+      ("aB", true),
+      ("AB", true),
+      ("AaAab", true),
+      ("aaaAB", true))
+    matchTest(
+      #"^(?i)a?b$"#,
+      ("ab", true),
+      ("Ab", true),
+      ("aB", true),
+      ("AB", true),
+      ("aaB", false),
+      ("b", true),
+      ("B", true))
+    matchTest(
+      #"^(?i)[a]?b$"#,
+      ("ab", true),
+      ("Ab", true),
+      ("aB", true),
+      ("AB", true),
+      ("b", true),
+      ("B", true))
+    matchTest(
+      #"^(?i)a{2,4}b$"#,
+      ("ab", false),
+      ("Ab", false),
+      ("AaB", true),
+      ("aAB", true),
+      ("aAaB", true),
+      ("aAaAB", true),
+      ("AaAaAB", false))
   }
 
   func testNonSemanticWhitespace() {