Merge pull request swiftlang#22663 from milseman/indigenous_characters

[String] Naturalize Character

Author: milseman

stdlib/public/core/Character.swift

@@ -81,6 +81,10 @@ extension Character {
   internal func _invariantCheck() {
     _internalInvariant(_str.count == 1)
     _internalInvariant(_str._guts.isFastUTF8)
+
+    // TODO(@eject): Switch to a helper property on StringObject/StringGuts.
+    _internalInvariant(
+      _str._guts.isSmall || _str._guts._object._countAndFlags.isTailAllocated)
   }
   #endif // INTERNAL_CHECKS_ENABLED
 }
@@ -173,7 +177,17 @@ extension Character :
       "Can't form a Character from an empty String")
     _debugPrecondition(s.index(after: s.startIndex) == s.endIndex,
       "Can't form a Character from a String containing more than one extended grapheme cluster")
-    self.init(unchecked: s)
+
+    // TODO(@eject): Switch to a helper property on StringObject/StringGuts.
+    if _fastPath(
+      s._guts.isSmall || s._guts._object._countAndFlags.isTailAllocated
+    ) {
+      self.init(unchecked: s)
+      return
+    }
+
+    // TODO(@eject): Outline this
+    self.init(unchecked: s._withUTF8 { String._uncheckedFromUTF8($0) })
   }
 }
 

stdlib/public/core/StringObject.swift

@@ -617,9 +617,10 @@ extension _StringObject {
  isNativelyStored: set for native stored strings
    - `largeAddressBits` holds an instance of `_StringStorage`.
    - I.e. the start of the code units is at the stored address + `nativeBias`
- isTailAllocated: start of the code units is at the stored address + `nativeBias`
+ isTailAllocated: contiguous UTF-8 code units starts at address + `nativeBias`
    - `isNativelyStored` always implies `isTailAllocated`, but not vice versa
       (e.g. literals)
+   - `isTailAllocated` always implies `isFastUTF8`
  TBD: Reserved for future usage
    - Setting a TBD bit to 1 must be semantically equivalent to 0
    - I.e. it can only be used to "cache" fast-path information in the future
@@ -1073,6 +1074,9 @@ extension _StringObject {
     } else {
       _internalInvariant(isLarge)
       _internalInvariant(largeCount == count)
+      if _countAndFlags.isTailAllocated {
+        _internalInvariant(providesFastUTF8)
+      }
       if providesFastUTF8 && largeFastIsTailAllocated {
         _internalInvariant(!isSmall)
         _internalInvariant(!largeIsCocoa)

test/stdlib/StringBridge.swift

@@ -75,5 +75,29 @@ StringBridgeTests.test("Tagged NSString") {
 #endif // not 32bit
 }
 
+func returnOne<T>(_ t: T) -> Int { return 1 }
+StringBridgeTests.test("Character from NSString") {
+  // NOTE: Using hard-coded literals to directly construct NSStrings
+  let ns1 = "A" as NSString
+  let ns2 = "A\u{301}" as NSString
+  let ns3 = "𓁹͇͈͉͍͎͊͋͌ͧͨͩͪͫͬͭͮ͏̛͓͔͕͖͙͚̗̘̙̜̹̺̻̼͐͑͒͗͛ͣͤͥͦ̽̾̿̀́͂̓̈́͆ͧͨͩͪͫͬͭͮ͘̚͜͟͢͝͞͠͡ͅ" as NSString
+
+  let c1 = Character(ns1 as String)
+  let c2 = Character(ns2 as String)
+  let c3 = Character(ns3 as String)
+
+  expectEqual("A", String(c1))
+  expectNotNil(String(c1).utf8.withContiguousStorageIfAvailable(returnOne))
+
+  expectEqual("A\u{301}", String(c2))
+  expectNotNil(String(c2).utf8.withContiguousStorageIfAvailable(returnOne))
+  expectNil((ns2 as String).utf8.withContiguousStorageIfAvailable(returnOne))
+
+  expectEqual("𓁹͇͈͉͍͎͊͋͌ͧͨͩͪͫͬͭͮ͏̛͓͔͕͖͙͚̗̘̙̜̹̺̻̼͐͑͒͗͛ͣͤͥͦ̽̾̿̀́͂̓̈́͆ͧͨͩͪͫͬͭͮ͘̚͜͟͢͝͞͠͡ͅ", String(c3))
+  expectNotNil(String(c3).utf8.withContiguousStorageIfAvailable(returnOne))
+  expectNil((ns3 as String).utf8.withContiguousStorageIfAvailable(returnOne))
+}
+
+
 runAllTests()