Merge pull request #14356 from milseman/what_a_character

[string] Minor cleanup on Character

Author: milseman

stdlib/public/core/Character.swift

@@ -136,9 +136,7 @@ extension Character
   @_inlineable // FIXME(sil-serialize-all)
   @effects(readonly)
   public init(_builtinUnicodeScalarLiteral value: Builtin.Int32) {
-    self = Character(
-      String._fromWellFormedCodeUnitSequence(
-        UTF32.self, input: CollectionOfOne(UInt32(value))))
+    self.init(Unicode.Scalar(_builtinUnicodeScalarLiteral: value))
   }
 
   // Inlining ensures that the whole constructor can be folded away to a single
@@ -154,22 +152,22 @@ extension Character
     let utf8 = UnsafeBufferPointer(
       start: UnsafePointer<Unicode.UTF8.CodeUnit>(start),
       count: Int(utf8CodeUnitCount))
-    
+
     if utf8.count == 1 {
       _representation = .smallUTF16(
         Builtin.zext_Int8_Int63(utf8.first._unsafelyUnwrappedUnchecked._value))
       return
     }
 
-  FastPath: 
+  FastPath:
     repeat {
       var shift = 0
       let maxShift = 64 - 16
       var bits: UInt64 = 0
-      
+
       for s8 in Unicode._ParsingIterator(
         codeUnits: utf8.makeIterator(), parser: UTF8.ForwardParser()) {
-        
+
         let s16
           = UTF16.transcode(s8, from: UTF8.self)._unsafelyUnwrappedUnchecked
 
@@ -186,7 +184,7 @@ extension Character
       return
     }
     while false
-    
+
     // For anything that doesn't fit in 63 bits, build the large
     // representation.
     self = Character(_largeRepresentationString:
@@ -227,11 +225,10 @@ extension Character
         | UInt64(utf16[3]) &<< 48
       _representation = .smallUTF16(Builtin.trunc_Int64_Int63(bits._value))
     default:
-      _representation = Character(
-        _largeRepresentationString: String(_StringGuts(utf16)))._representation
+      _representation = .large(_StringGuts(utf16)._extractNativeStorage())
     }
   }
-  
+
   /// Creates a character with the specified value.
   ///
   /// Do not call this initalizer directly. It is used by the compiler when
@@ -485,7 +482,7 @@ extension Character : Equatable {
         if l == r { return true }
       }
     }
-    
+
     // FIXME(performance): constructing two temporary strings is extremely
     // wasteful and inefficient.
     return String(lhs) == String(rhs)

stdlib/public/core/String.swift

@@ -25,17 +25,17 @@ public protocol StringProtocol
 
   associatedtype UTF8View : /*Bidirectional*/Collection
   where UTF8View.Element == UInt8 // Unicode.UTF8.CodeUnit
-  
+
   associatedtype UTF16View : BidirectionalCollection
   where UTF16View.Element == UInt16 // Unicode.UTF16.CodeUnit
 
   associatedtype UnicodeScalarView : BidirectionalCollection
   where UnicodeScalarView.Element == Unicode.Scalar
-  
+
   var utf8: UTF8View { get }
   var utf16: UTF16View { get }
   var unicodeScalars: UnicodeScalarView { get }
-  
+
 #if _runtime(_ObjC)
   func hasPrefix(_ prefix: String) -> Bool
   func hasSuffix(_ prefix: String) -> Bool
@@ -63,7 +63,7 @@ public protocol StringProtocol
   /// - Parameter nullTerminatedUTF8: A pointer to a sequence of contiguous,
   ///   UTF-8 encoded bytes ending just before the first zero byte.
   init(cString nullTerminatedUTF8: UnsafePointer<CChar>)
-  
+
   /// Creates a string from the null-terminated sequence of bytes at the given
   /// pointer.
   ///
@@ -76,7 +76,7 @@ public protocol StringProtocol
   init<Encoding: Unicode.Encoding>(
     decodingCString nullTerminatedCodeUnits: UnsafePointer<Encoding.CodeUnit>,
     as sourceEncoding: Encoding.Type)
-  
+
   /// Calls the given closure with a pointer to the contents of the string,
   /// represented as a null-terminated sequence of UTF-8 code units.
   ///
@@ -151,7 +151,7 @@ internal protocol _SwiftStringView {
   /// A `String`, having the same contents as `self`, that may be unsuitable for
   /// long-term storage.
   var _ephemeralContent : String { get }
-  
+
   /// A `String`, having the same contents as `self`, that is suitable for
   /// long-term storage.
   //
@@ -206,8 +206,8 @@ extension String : _SwiftStringView {
 @_versioned // FIXME(sil-serialize-all)
 internal func _withCString<
   Source : Collection,
-  SourceEncoding : Unicode.Encoding, 
-  TargetEncoding : Unicode.Encoding, 
+  SourceEncoding : Unicode.Encoding,
+  TargetEncoding : Unicode.Encoding,
   Result
 >(
   encodedAs targetEncoding: TargetEncoding.Type,
@@ -227,8 +227,8 @@ where Source.Iterator.Element == SourceEncoding.CodeUnit {
 @_semantics("optimize.sil.specialize.generic.partial.never")
 internal func _withCStringAndLength<
   Source : Collection,
-  SourceEncoding : Unicode.Encoding, 
-  TargetEncoding : Unicode.Encoding, 
+  SourceEncoding : Unicode.Encoding,
+  TargetEncoding : Unicode.Encoding,
   Result
 >(
   encodedAs targetEncoding: TargetEncoding.Type,
@@ -383,7 +383,7 @@ extension String {
       repairIllFormedSequences: true)
     self = result!
   }
-  
+
   /// Creates a string from the null-terminated sequence of bytes at the given
   /// pointer.
   ///
@@ -404,7 +404,7 @@ extension String {
     )
     self.init(decoding: codeUnits, as: sourceEncoding)
   }
-  
+
   /// Calls the given closure with a pointer to the contents of the string,
   /// represented as a null-terminated sequence of code units.
   ///
@@ -1063,7 +1063,7 @@ extension String {
     start: UnsafeMutablePointer<UTF8.CodeUnit>,
     utf8CodeUnitCount: Int
   ) {
-    resultStorage.initialize(to: 
+    resultStorage.initialize(to:
       String._fromWellFormedCodeUnitSequence(
         UTF8.self,
         input: UnsafeBufferPointer(start: start, count: utf8CodeUnitCount)))
@@ -1356,7 +1356,7 @@ extension String {
     return _nativeUnicodeUppercaseString(self)
 #endif
   }
-  
+
   /// Creates an instance from the description of a given
   /// `LosslessStringConvertible` instance.
   @_inlineable // FIXME(sil-serialize-all)

stdlib/public/core/StringBridge.swift

@@ -81,7 +81,7 @@ internal func _cocoaStringSlice(
   _ target: _CocoaString, _ bounds: Range<Int>
 ) -> _CocoaString {
   let cfSelf: _swift_shims_CFStringRef = target
-  
+
   _sanityCheck(
     _swift_stdlib_CFStringGetCharactersPtr(cfSelf) == nil,
     "Known contiguously stored strings should already be converted to Swift")