From b311a75884408193fe2d4611b33057519e437c3a Mon Sep 17 00:00:00 2001
From: Nate Cook <natecook@apple.com>
Date: Mon, 6 Jan 2025 15:02:01 -0600
Subject: [PATCH 1/2] Fix repeated case-insensitive ASCII match

The optimized bytecode for matching repetition of a single character overlooks
case insensitivity. This resolves that by falling back to use an ASCII bitset
when doing a case-insensitive match of a cased character.

Fixes #785.
---
 Sources/_StringProcessing/ByteCodeGen.swift | 16 ++++++++++++----
 Tests/RegexTests/MatchTests.swift           | 19 +++++++++++++++++++
 2 files changed, 31 insertions(+), 4 deletions(-)

diff --git a/Sources/_StringProcessing/ByteCodeGen.swift b/Sources/_StringProcessing/ByteCodeGen.swift
index ae8a1c8b..6a00a0df 100644
--- a/Sources/_StringProcessing/ByteCodeGen.swift
+++ b/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(
diff --git a/Tests/RegexTests/MatchTests.swift b/Tests/RegexTests/MatchTests.swift
index 1b5c67e0..256372b3 100644
--- a/Tests/RegexTests/MatchTests.swift
+++ b/Tests/RegexTests/MatchTests.swift
@@ -2193,6 +2193,25 @@ extension RegexTests {
       ("cafe", true),
       ("CaFe", true),
       ("EfAc", true))
+    
+    matchTest(
+      #"(?i)a+b"#,
+      ("ab", true),
+      ("Ab", true),
+      ("aB", true),
+      ("AB", true))
+    matchTest(
+      #"(?i)a?b"#,
+      ("ab", true),
+      ("Ab", true),
+      ("aB", true),
+      ("AB", true))
+    matchTest(
+      #"(?i)[a]?b"#,
+      ("ab", true),
+      ("Ab", true),
+      ("aB", true),
+      ("AB", true))
   }
 
   func testNonSemanticWhitespace() {

From cf27df1d897f156716bdc4abd1c628720a6003db Mon Sep 17 00:00:00 2001
From: Nate Cook <natecook@apple.com>
Date: Tue, 7 Jan 2025 17:37:11 -0600
Subject: [PATCH 2/2] Expand test coverage

---
 Tests/RegexTests/MatchTests.swift | 26 +++++++++++++++++++++-----
 1 file changed, 21 insertions(+), 5 deletions(-)

diff --git a/Tests/RegexTests/MatchTests.swift b/Tests/RegexTests/MatchTests.swift
index 256372b3..c52560d6 100644
--- a/Tests/RegexTests/MatchTests.swift
+++ b/Tests/RegexTests/MatchTests.swift
@@ -2199,19 +2199,35 @@ extension RegexTests {
       ("ab", true),
       ("Ab", true),
       ("aB", true),
-      ("AB", true))
+      ("AB", true),
+      ("AaAab", true),
+      ("aaaAB", true))
     matchTest(
-      #"(?i)a?b"#,
+      #"^(?i)a?b$"#,
       ("ab", true),
       ("Ab", true),
       ("aB", true),
-      ("AB", true))
+      ("AB", true),
+      ("aaB", false),
+      ("b", true),
+      ("B", true))
     matchTest(
-      #"(?i)[a]?b"#,
+      #"^(?i)[a]?b$"#,
       ("ab", true),
       ("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() {