[string] String(decoding:as:) fast path for withContiguousStorageIfAvailable

stdlib/public/core/ContiguouslyStored.swift

@@ -10,6 +10,10 @@
 //
 //===----------------------------------------------------------------------===//
 
+// NOTE: The below is necessary for fast String initialization from untyped
+// memory. When we add Collection.withContiguousRawStorageIfAvailabe(), we can
+// deprecate this functionality.
+
 @usableFromInline
 internal protocol _HasContiguousBytes {
   func withUnsafeBytes<R>(

stdlib/public/core/String.swift

@@ -388,6 +388,22 @@ extension String {
 }
 
 extension String {
+  // This force type-casts element to UInt8, since we cannot currently
+  // communicate to the type checker that we proved this with our dynamic
+  // check in String(decoding:as:).
+  @_alwaysEmitIntoClient
+  @inline(never) // slow-path
+  private static func _fromNonContiguousUnsafeBitcastUTF8Repairing<
+    C: Collection
+  >(_ input: C) -> (result: String, repairsMade: Bool) {
+    _internalInvariant(C.Element.self == UInt8.self)
+    return Array(input).withUnsafeBufferPointer {
+      let raw = UnsafeRawBufferPointer($0)
+      return String._fromUTF8Repairing(raw.bindMemory(to: UInt8.self))
+    }
+  }
+
+
   /// Creates a string from the given Unicode code units in the specified
   /// encoding.
   ///
@@ -407,8 +423,27 @@ extension String {
       return
     }
 
+    // Fast path for user-defined Collections and typed contiguous collections.
+    //
+    // Note: this comes first, as the optimizer nearly always has insight into
+    // wCSIA, but cannot prove that a type does not have conformance to
+    // _HasContiguousBytes.
+    if let str = codeUnits.withContiguousStorageIfAvailable({
+      (buffer: UnsafeBufferPointer<C.Element>) -> String in
+      Builtin.onFastPath() // encourage SIL Optimizer to inline this closure :-(
+      let rawBufPtr = UnsafeRawBufferPointer(buffer)
+      return String._fromUTF8Repairing(
+        UnsafeBufferPointer(
+          start: rawBufPtr.baseAddress?.assumingMemoryBound(to: UInt8.self),
+          count: rawBufPtr.count)).0
+    }) {
+      self = str
+      return
+    }
+
+    // Fast path for untyped raw storage and known stdlib types
     if let contigBytes = codeUnits as? _HasContiguousBytes,
-       contigBytes._providesContiguousBytesNoCopy
+      contigBytes._providesContiguousBytesNoCopy
     {
       self = contigBytes.withUnsafeBytes { rawBufPtr in
         return String._fromUTF8Repairing(
@@ -419,15 +454,9 @@ extension String {
       return
     }
 
-    // Just copying to an Array is significantly faster than performing
-    // generic operations
-    self = Array(codeUnits).withUnsafeBufferPointer {
-      let raw = UnsafeRawBufferPointer($0)
-      return String._fromUTF8Repairing(raw.bindMemory(to: UInt8.self)).0
-    }
-    return
+    self = String._fromNonContiguousUnsafeBitcastUTF8Repairing(codeUnits).0
   }
-  
+
   /// Creates a new string with the specified capacity in UTF-8 code units, and
   /// then calls the given closure with a buffer covering the string's
   /// uninitialized memory.
@@ -484,7 +513,7 @@ extension String {
       initializingUTF8With: initializer
     )
   }
-  
+
   @inline(__always)
   internal init(
     _uninitializedCapacity capacity: Int,
@@ -503,7 +532,7 @@ extension String {
       }
       return
     }
-    
+
     self = try String._fromLargeUTF8Repairing(
       uninitializedCapacity: capacity,
       initializingWith: initializer)
@@ -968,20 +997,20 @@ extension _StringGutsSlice {
     var outputBuffer = outputBuffer
     var icuInputBuffer = icuInputBuffer
     var icuOutputBuffer = icuOutputBuffer
-  
+
     var index = range.lowerBound
     let cachedEndIndex = range.upperBound
-  
+
     var hasBufferOwnership = false
-    
+
     defer {
       if hasBufferOwnership {
         outputBuffer.deallocate()
         icuInputBuffer.deallocate()
         icuOutputBuffer.deallocate()
       }
     }
-    
+
     while index < cachedEndIndex {
       let result = _foreignNormalize(
         readIndex: index,
@@ -1017,17 +1046,17 @@ internal func _fastWithNormalizedCodeUnitsImpl(
 
   var index = String.Index(_encodedOffset: 0)
   let cachedEndIndex = String.Index(_encodedOffset: sourceBuffer.count)
-  
+
   var hasBufferOwnership = false
-  
+
   defer {
     if hasBufferOwnership {
       outputBuffer.deallocate()
       icuInputBuffer.deallocate()
       icuOutputBuffer.deallocate()
     }
   }
-  
+
   while index < cachedEndIndex {
     let result = _fastNormalize(
       readIndex: index,

stdlib/public/core/StringCreate.swift

@@ -50,7 +50,7 @@ internal func _allASCII(_ input: UnsafeBufferPointer<UInt8>) -> Bool {
 }
 
 extension String {
-  
+
   internal static func _uncheckedFromASCII(
     _ input: UnsafeBufferPointer<UInt8>
   ) -> String {
@@ -61,15 +61,15 @@ extension String {
     let storage = __StringStorage.create(initializingFrom: input, isASCII: true)
     return storage.asString
   }
-  
+
   @usableFromInline
   internal static func _fromASCII(
     _ input: UnsafeBufferPointer<UInt8>
   ) -> String {
     _internalInvariant(_allASCII(input), "not actually ASCII")
     return _uncheckedFromASCII(input)
   }
-  
+
   internal static func _fromASCIIValidating(
     _ input: UnsafeBufferPointer<UInt8>
   ) -> String? {
@@ -101,7 +101,7 @@ extension String {
         return (repairUTF8(input, firstKnownBrokenRange: initialRange), true)
     }
   }
-  
+
   internal static func _fromLargeUTF8Repairing(
     uninitializedCapacity capacity: Int,
     initializingWith initializer: (
@@ -111,7 +111,7 @@ extension String {
     let result = try __StringStorage.create(
       uninitializedCodeUnitCapacity: capacity,
       initializingUncheckedUTF8With: initializer)
-    
+
     switch validateUTF8(result.codeUnits) {
     case .success(let info):
       result._updateCountAndFlags(
@@ -181,7 +181,7 @@ extension String {
 
     return contents.withUnsafeBufferPointer { String._uncheckedFromUTF8($0) }
   }
-  
+
   @inline(never) // slow path
   private static func _slowFromCodeUnits<
     Input: Collection,
@@ -209,7 +209,7 @@ extension String {
     let str = contents.withUnsafeBufferPointer { String._uncheckedFromUTF8($0) }
     return (str, repaired)
   }
-  
+
   @usableFromInline @inline(never) // can't be inlined w/out breaking ABI
   @_specialize(
     where Input == UnsafeBufferPointer<UInt8>, Encoding == Unicode.ASCII)
@@ -227,27 +227,39 @@ extension String {
     guard _fastPath(encoding == Unicode.ASCII.self) else {
       return _slowFromCodeUnits(input, encoding: encoding, repair: repair)
     }
-
-    var result:String? = nil
-
+
+    // Helper to simplify early returns
+    func resultOrSlow(_ resultOpt: String?) -> (String, repairsMade: Bool)? {
+      guard let result = resultOpt else {
+        return _slowFromCodeUnits(input, encoding: encoding, repair: repair)
+      }
+      return (result, repairsMade: false)
+    }
+
+    // Fast path for untyped raw storage and known stdlib types
     if let contigBytes = input as? _HasContiguousBytes,
       contigBytes._providesContiguousBytesNoCopy {
-      result = contigBytes.withUnsafeBytes { rawBufPtr in
+      return resultOrSlow(contigBytes.withUnsafeBytes { rawBufPtr in
         let buffer = UnsafeBufferPointer(
           start: rawBufPtr.baseAddress?.assumingMemoryBound(to: UInt8.self),
           count: rawBufPtr.count)
         return String._fromASCIIValidating(buffer)
-      }
-    } else {
-      result = Array(input).withUnsafeBufferPointer {
+      })
+    }
+
+    // Fast path for user-defined Collections
+    if let strOpt = input.withContiguousStorageIfAvailable({
+      (buffer: UnsafeBufferPointer<Input.Element>) -> String? in
+      return String._fromASCIIValidating(
+        UnsafeRawBufferPointer(buffer).bindMemory(to: UInt8.self))
+    }) {
+      return resultOrSlow(strOpt)
+    }
+
+    return resultOrSlow(Array(input).withUnsafeBufferPointer {
         let buffer = UnsafeRawBufferPointer($0).bindMemory(to: UInt8.self)
         return String._fromASCIIValidating(buffer)
-      }
-    }
-
-    return result != nil ?
-      (result!, repairsMade: false) :
-      _slowFromCodeUnits(input, encoding: encoding, repair: repair)
+      })
   }
 
   public // @testable

stdlib/public/core/Substring.swift

@@ -391,6 +391,7 @@ extension Substring.UTF8View: BidirectionalCollection {
   }
 
   @_alwaysEmitIntoClient
+  @inlinable
   public func withContiguousStorageIfAvailable<R>(
     _ body: (UnsafeBufferPointer<Element>) throws -> R
   ) rethrows -> R? {