Fix scalar mode for quoted sequences in character class

Previously we would only match entire characters.
Update to use the generic Character consumer logic
that can handle scalar semantic mode.

rdar://97209131

Author: hamishknight

Sources/_StringProcessing/ConsumerInterface.swift

@@ -63,7 +63,7 @@ extension DSLTree._AST.Atom {
 extension Character {
   func generateConsumer(
     _ opts: MatchingOptions
-  ) throws -> MEProgram.ConsumeFunction? {
+  ) throws -> MEProgram.ConsumeFunction {
     let isCaseInsensitive = opts.isCaseInsensitive
     switch opts.semanticLevel {
     case .graphemeCluster:
@@ -437,21 +437,17 @@ extension DSLTree.CustomCharacterClass.Member {
         }
         return rhs(input, bounds)
       }
-    case .quotedLiteral(let s):
-      if opts.isCaseInsensitive {
-        return { input, bounds in
-          guard s.lowercased()._contains(input[bounds.lowerBound].lowercased()) else {
-            return nil
-          }
-          return input.index(after: bounds.lowerBound)
-        }
-      } else {
-        return { input, bounds in
-          guard s.contains(input[bounds.lowerBound]) else {
-            return nil
+    case .quotedLiteral(let str):
+      let consumers = try str.map {
+        try $0.generateConsumer(opts)
+      }
+      return { input, bounds in
+        for fn in consumers {
+          if let idx = fn(input, bounds) {
+            return idx
           }
-          return input.index(after: bounds.lowerBound)
         }
+        return nil
       }
     case .trivia:
       // TODO: Should probably strip this earlier...

Tests/RegexTests/MatchTests.swift

@@ -191,6 +191,7 @@ func firstMatchTests(
   enableTracing: Bool = false,
   dumpAST: Bool = false,
   xfail: Bool = false,
+  semanticLevel: RegexSemanticLevel = .graphemeCluster,
   file: StaticString = #filePath,
   line: UInt = #line
 ) {
@@ -203,6 +204,7 @@ func firstMatchTests(
       enableTracing: enableTracing,
       dumpAST: dumpAST,
       xfail: xfail,
+      semanticLevel: semanticLevel,
       file: file,
       line: line)
   }
@@ -685,6 +687,50 @@ extension RegexTests {
       ("a\u{301}", true),
       semanticLevel: .unicodeScalar)
 
+    // Scalar matching in quoted sequences.
+    firstMatchTests(
+      "[\\Qe\u{301}\\E]",
+      ("e", nil),
+      ("E", nil),
+      ("\u{301}", nil),
+      (eDecomposed, eDecomposed),
+      (eComposed, eComposed),
+      ("E\u{301}", nil),
+      ("\u{C9}", nil)
+    )
+    firstMatchTests(
+      "[\\Qe\u{301}\\E]",
+      ("e", "e"),
+      ("E", nil),
+      ("\u{301}", "\u{301}"),
+      (eDecomposed, "e"),
+      (eComposed, nil),
+      ("E\u{301}", "\u{301}"),
+      ("\u{C9}", nil),
+      semanticLevel: .unicodeScalar
+    )
+    firstMatchTests(
+      "(?i)[\\Qe\u{301}\\E]",
+      ("e", nil),
+      ("E", nil),
+      ("\u{301}", nil),
+      (eDecomposed, eDecomposed),
+      (eComposed, eComposed),
+      ("E\u{301}", "E\u{301}"),
+      ("\u{C9}", "\u{C9}")
+    )
+    firstMatchTests(
+      "(?i)[\\Qe\u{301}\\E]",
+      ("e", "e"),
+      ("E", "E"),
+      ("\u{301}", "\u{301}"),
+      (eDecomposed, "e"),
+      (eComposed, nil),
+      ("E\u{301}", "E"),
+      ("\u{C9}", nil),
+      semanticLevel: .unicodeScalar
+    )
+
     firstMatchTest("[-]", input: "123-abcxyz", match: "-")
 
     // These are metacharacters in certain contexts, but normal characters