Support text segment boundary anchors (#178)

This enables the `\y` and `\Y` anchors in regex literals and
`Anchor.textSegmentBoundary` in the DSL.

Note: This also includes `UnicodeScalar` conformance to
`RegexProtocol`, which acts like Unicode scalar literals in
regex literals.

Author: natecook1000

Sources/_StringProcessing/ByteCodeGen.swift

@@ -99,12 +99,16 @@ extension Compiler.ByteCodeGen {
       }
 
     case .textSegment:
-      // This we should be able to do!
-      throw Unsupported(#"\y (text segment)"#)
+      builder.buildAssert { (input, pos, _) in
+        // FIXME: Grapheme or word based on options
+        input.isOnGraphemeClusterBoundary(pos)
+      }
 
     case .notTextSegment:
-      // This we should be able to do!
-      throw Unsupported(#"\Y (not text segment)"#)
+      builder.buildAssert { (input, pos, _) in
+        // FIXME: Grapheme or word based on options
+        !input.isOnGraphemeClusterBoundary(pos)
+      }
 
     case .startOfLine:
       builder.buildAssert { (input, pos, bounds) in

Sources/_StringProcessing/RegexDSL/DSL.swift

@@ -17,25 +17,31 @@ extension String: RegexProtocol {
   public typealias Match = Substring
 
   public var regex: Regex<Match> {
-    let atoms = self.map { atom(.char($0)) }
-    return .init(ast: concat(atoms))
+    .init(node: .quotedLiteral(self))
   }
 }
 
 extension Substring: RegexProtocol {
   public typealias Match = Substring
 
   public var regex: Regex<Match> {
-    let atoms = self.map { atom(.char($0)) }
-    return .init(ast: concat(atoms))
+    .init(node: .quotedLiteral(String(self)))
   }
 }
 
 extension Character: RegexProtocol {
   public typealias Match = Substring
 
   public var regex: Regex<Match> {
-    .init(ast: atom(.char(self)))
+    .init(node: .atom(.char(self)))
+  }
+}
+
+extension UnicodeScalar: RegexProtocol {
+  public typealias Match = Substring
+
+  public var regex: Regex<Match> {
+    .init(node: .atom(.scalar(self)))
   }
 }
 

Tests/RegexTests/MatchTests.swift

@@ -877,7 +877,8 @@ extension RegexTests {
       #"\d+\b"#,
       ("123", "123"),
       (" 123", "123"),
-      ("123 456", "123"))
+      ("123 456", "123"),
+      ("123A 456", "456"))
     firstMatchTests(
       #"\d+\b\s\b\d+"#,
       ("123", nil),
@@ -893,7 +894,18 @@ extension RegexTests {
     // TODO: \G and \K
 
     // TODO: Oniguruma \y and \Y
-
+    firstMatchTests(
+      #"\u{65}"#,             // Scalar 'e' is present in both:
+      ("Cafe\u{301}", "e"),   // composed and
+      ("Sol Cafe", "e"))      // standalone
+    firstMatchTests(
+      #"\u{65}\y"#,           // Grapheme boundary assertion
+      ("Cafe\u{301}", nil),
+      ("Sol Cafe", "e"))
+    firstMatchTests(
+      #"\u{65}\Y"#,           // Grapheme non-boundary assertion
+      ("Cafe\u{301}", "e"),
+      ("Sol Cafe", nil))
   }
 
   func testMatchGroups() {

Tests/RegexTests/RegexDSLTests.swift

@@ -280,6 +280,16 @@ class RegexDSLTests: XCTestCase {
       Anchor.endOfLine
     }
 
+    try _testDSLCaptures(
+      ("Cafe\u{301}", nil),
+      ("Cafe", "Cafe"),
+      matchType: Substring.self, ==)
+    {
+      oneOrMore(.word)
+      UnicodeScalar("e")
+      Anchor.textSegmentBoundary
+    }
+
     try _testDSLCaptures(
       ("aaaaa1", "aaaaa1"),
       ("aaaaa2", nil),