Merge pull request #21392 from milseman/5_0_4_bit_string

[5.0] [String] Last-minute ABI adjustment: 4-bit discriminator

Author: AnnaZaks

stdlib/public/SwiftShims/CoreFoundationShims.h

@@ -134,9 +134,9 @@ _swift_stdlib_NSStringCStringUsingEncodingTrampoline(id _Nonnull obj,
 SWIFT_RUNTIME_STDLIB_API
 __swift_uint8_t
 _swift_stdlib_NSStringGetCStringTrampoline(id _Nonnull obj,
-                                         _swift_shims_UInt8 *buffer,
-                                         _swift_shims_CFIndex maxLength,
-                                         unsigned long encoding);
+                                           _swift_shims_UInt8 *_Nonnull buffer,
+                                           _swift_shims_CFIndex maxLength,
+                                           unsigned long encoding);
 
 SWIFT_RUNTIME_STDLIB_API
 __swift_uintptr_t

stdlib/public/core/SmallString.swift

@@ -72,18 +72,9 @@ extension _SmallString {
     }
   }
 
-  @inlinable
-  internal var discriminator: _StringObject.Discriminator {
-    @inline(__always) get {
-      let value = _storage.1 &>> _StringObject.Nibbles.discriminatorShift
-      return _StringObject.Discriminator(UInt8(truncatingIfNeeded: value))
-    }
-    @inline(__always) set {
-      _storage.1 &= _StringObject.Nibbles.largeAddressMask
-      _storage.1 |= (
-        UInt64(truncatingIfNeeded: newValue._value)
-          &<< _StringObject.Nibbles.discriminatorShift)
-    }
+  @inlinable @inline(__always)
+  internal var rawDiscriminatedObject: UInt64 {
+    return _storage.1
   }
 
   @inlinable
@@ -96,7 +87,7 @@ extension _SmallString {
   @inlinable
   internal var count: Int {
     @inline(__always) get {
-      return discriminator.smallCount
+      return _StringObject.getSmallCount(fromRaw: rawDiscriminatedObject)
     }
   }
 
@@ -108,27 +99,22 @@ extension _SmallString {
   @inlinable
   internal var isASCII: Bool {
     @inline(__always) get {
-      return discriminator.smallIsASCII
+      return _StringObject.getSmallIsASCII(fromRaw: rawDiscriminatedObject)
     }
   }
 
   // Give raw, nul-terminated code units. This is only for limited internal
   // usage: it always clears the discriminator and count (in case it's full)
-  @inlinable
+  @inlinable @inline(__always)
   internal var zeroTerminatedRawCodeUnits: RawBitPattern {
-    @inline(__always) get {
-      return (
-        self._storage.0,
-        self._storage.1 & _StringObject.Nibbles.largeAddressMask)
-    }
+    let smallStringCodeUnitMask: UInt64 = 0x00FF_FFFF_FFFF_FFFF
+    return (self._storage.0, self._storage.1 & smallStringCodeUnitMask)
   }
 
-  @inlinable
   internal func computeIsASCII() -> Bool {
-    // TODO(String micro-performance): Evaluate other expressions, e.g. | first
     let asciiMask: UInt64 = 0x8080_8080_8080_8080
     let raw = zeroTerminatedRawCodeUnits
-    return (raw.0 & asciiMask == 0) && (raw.1 & asciiMask == 0)
+    return (raw.0 | raw.1) & asciiMask == 0
   }
 }
 
@@ -220,7 +206,7 @@ extension _SmallString {
 
   // Overwrite stored code units, including uninitialized. `f` should return the
   // new count.
-  @inlinable @inline(__always)
+  @inline(__always)
   internal mutating func withMutableCapacity(
     _ f: (UnsafeMutableBufferPointer<UInt8>) throws -> Int
   ) rethrows {
@@ -231,14 +217,28 @@ extension _SmallString {
       return try f(UnsafeMutableBufferPointer(
         start: ptr, count: _SmallString.capacity))
     }
-
     _internalInvariant(len <= _SmallString.capacity)
-    discriminator = .small(withCount: len, isASCII: self.computeIsASCII())
+
+    let (leading, trailing) = self.zeroTerminatedRawCodeUnits
+    self = _SmallString(leading: leading, trailing: trailing, count: len)
   }
 }
 
 // Creation
 extension _SmallString {
+  @inlinable @inline(__always)
+  internal init(leading: UInt64, trailing: UInt64, count: Int) {
+    _internalInvariant(count <= _SmallString.capacity)
+
+    let isASCII = (leading | trailing) & 0x8080_8080_8080_8080 == 0
+    let countAndDiscriminator = UInt64(truncatingIfNeeded: count) &<< 56
+                              | _StringObject.Nibbles.small(isASCII: isASCII)
+    _internalInvariant(trailing & countAndDiscriminator == 0)
+
+    self.init(raw: (leading, trailing | countAndDiscriminator))
+    _internalInvariant(self.count == count)
+  }
+
   // Direct from UTF-8
   @inlinable @inline(__always)
   internal init?(_ input: UnsafeBufferPointer<UInt8>) {
@@ -251,11 +251,7 @@ extension _SmallString {
     let leading = _bytesToUInt64(ptr, Swift.min(input.count, 8))
     let trailing = count > 8 ? _bytesToUInt64(ptr + 8, count &- 8) : 0
 
-    let isASCII = (leading | trailing) & 0x8080_8080_8080_8080 == 0
-    let discriminator = _StringObject.Discriminator.small(
-      withCount: count,
-      isASCII: isASCII)
-    self.init(raw: (leading, trailing | discriminator.rawBits))
+    self.init(leading: leading, trailing: trailing, count: count)
   }
 
   @usableFromInline // @testable
@@ -273,13 +269,8 @@ extension _SmallString {
     }
     _internalInvariant(writeIdx == totalCount)
 
-    let isASCII = base.isASCII && other.isASCII
-    let discriminator = _StringObject.Discriminator.small(
-      withCount: totalCount,
-      isASCII: isASCII)
-
     let (leading, trailing) = result.zeroTerminatedRawCodeUnits
-    self.init(raw: (leading, trailing | discriminator.rawBits))
+    self.init(leading: leading, trailing: trailing, count: totalCount)
   }
 }
 

stdlib/public/core/StringBridge.swift

@@ -2,7 +2,7 @@
 //
 // This source file is part of the Swift.org open source project
 //
-// Copyright (c) 2014 - 2017 Apple Inc. and the Swift project authors
+// Copyright (c) 2014 - 2018 Apple Inc. and the Swift project authors
 // Licensed under Apache License v2.0 with Runtime Library Exception
 //
 // See https://swift.org/LICENSE.txt for license information
@@ -47,8 +47,8 @@ internal func _stdlib_binary_CFStringGetCharactersPtr(
     mutating: _swift_stdlib_CFStringGetCharactersPtr(source))
 }
 
-/// Copies a slice of a _CocoaString into contiguous storage of
-/// sufficient capacity.
+/// Copies a slice of a _CocoaString into contiguous storage of sufficient
+/// capacity.
 @_effects(releasenone)
 internal func _cocoaStringCopyCharacters(
   from source: _CocoaString,
@@ -81,54 +81,46 @@ internal func _cocoaStringCompare(
 @_effects(readonly)
 internal func _cocoaHashString(
   _ string: _CocoaString
-  ) -> UInt {
+) -> UInt {
   return _swift_stdlib_CFStringHashNSString(string)
 }
 
 @_effects(readonly)
 internal func _cocoaHashASCIIBytes(
-  _ bytes: UnsafePointer<UInt8>,
-  length: Int
-  ) -> UInt {
+  _ bytes: UnsafePointer<UInt8>, length: Int
+) -> UInt {
   return _swift_stdlib_CFStringHashCString(bytes, length)
 }
 
 // These "trampolines" are effectively objc_msgSend_super.
-// They bypass our implementations to use NSString's
+// They bypass our implementations to use NSString's.
 
 @_effects(readonly)
 internal func _cocoaCStringUsingEncodingTrampoline(
-  _ string: _CocoaString,
-  _ encoding: UInt)
-  -> UnsafePointer<UInt8>? {
-    return _swift_stdlib_NSStringCStringUsingEncodingTrampoline(
-      string,
-      encoding)
+  _ string: _CocoaString, _ encoding: UInt
+) -> UnsafePointer<UInt8>? {
+  return _swift_stdlib_NSStringCStringUsingEncodingTrampoline(string, encoding)
 }
 
-
-
-
 @_effects(releasenone)
 internal func _cocoaGetCStringTrampoline(
-                             _ string: _CocoaString,
-                             _ buffer: UnsafeMutablePointer<UInt8>,
-                             _ maxLength: Int,
-                             _ encoding: UInt)
-  -> Int8 {
-    return Int8(_swift_stdlib_NSStringGetCStringTrampoline(string,
-                                                    buffer,
-                                                    maxLength,
-                                                    encoding))
+  _ string: _CocoaString,
+  _ buffer: UnsafeMutablePointer<UInt8>,
+  _ maxLength: Int,
+  _ encoding: UInt
+) -> Int8 {
+  return Int8(_swift_stdlib_NSStringGetCStringTrampoline(
+    string, buffer, maxLength, encoding))
 }
 
 //
-// Conversion from NSString to Swift's native representation
+// Conversion from NSString to Swift's native representation.
 //
 
 private var kCFStringEncodingASCII : _swift_shims_CFStringEncoding {
   @inline(__always) get { return 0x0600 }
 }
+
 private var kCFStringEncodingUTF8 : _swift_shims_CFStringEncoding {
   @inline(__always) get { return 0x8000100 }
 }
@@ -149,12 +141,12 @@ internal enum _KnownCocoaString {
   @inline(__always)
   init(_ str: _CocoaString) {
 
-    #if !(arch(i386) || arch(arm))
+#if !(arch(i386) || arch(arm))
     if _isObjCTaggedPointer(str) {
       self = .tagged
       return
     }
-    #endif
+#endif
 
     switch _unsafeAddressOfCocoaStringClass(str) {
     case unsafeBitCast(_StringStorage.self, to: UInt.self):
@@ -168,7 +160,7 @@ internal enum _KnownCocoaString {
 }
 
 #if !(arch(i386) || arch(arm))
-// Resiliently write a tagged cocoa string's contents into a buffer
+// Resiliently write a tagged _CocoaString's contents into a buffer.
 @_effects(releasenone) // @opaque
 internal func _bridgeTagged(
   _ cocoa: _CocoaString,
@@ -223,32 +215,33 @@ private func _getCocoaStringPointer(
 internal func _bridgeCocoaString(_ cocoaString: _CocoaString) -> _StringGuts {
   switch _KnownCocoaString(cocoaString) {
   case .storage:
-    return _unsafeUncheckedDowncast(cocoaString, to: _StringStorage.self).asString._guts
+    return _unsafeUncheckedDowncast(
+      cocoaString, to: _StringStorage.self).asString._guts
   case .shared:
-    return _unsafeUncheckedDowncast(cocoaString, to: _SharedStringStorage.self).asString._guts
+    return _unsafeUncheckedDowncast(
+      cocoaString, to: _SharedStringStorage.self).asString._guts
 #if !(arch(i386) || arch(arm))
   case .tagged:
-      return _StringGuts(_SmallString(taggedCocoa: cocoaString))
+    return _StringGuts(_SmallString(taggedCocoa: cocoaString))
 #endif
   case .cocoa:
-    // "copy" it into a value to be sure nobody will modify behind
-    // our backs.  In practice, when value is already immutable, this
+    // "Copy" it into a value to be sure nobody will modify behind
+    // our backs. In practice, when value is already immutable, this
     // just does a retain.
     //
     // TODO: Only in certain circumstances should we emit this call:
     //   1) If it's immutable, just retain it.
     //   2) If it's mutable with no associated information, then a copy must
     //      happen; might as well eagerly bridge it in.
     //   3) If it's mutable with associated information, must make the call
-    //
     let immutableCopy
       = _stdlib_binary_CFStringCreateCopy(cocoaString) as AnyObject
 
-    #if !(arch(i386) || arch(arm))
+#if !(arch(i386) || arch(arm))
     if _isObjCTaggedPointer(immutableCopy) {
       return _StringGuts(_SmallString(taggedCocoa: immutableCopy))
     }
-    #endif
+#endif
 
     let (fastUTF8, isASCII): (Bool, Bool)
     switch _getCocoaStringPointer(immutableCopy) {
@@ -288,9 +281,13 @@ extension String {
       }
     }
     if _guts._object.isImmortal {
+      // TODO: We'd rather emit a valid ObjC object statically than create a
+      // shared string class instance.
+      let gutsCountAndFlags = _guts._object._countAndFlags
       return _SharedStringStorage(
         immortal: _guts._object.fastUTF8.baseAddress!,
-        countAndFlags: _guts._object._countAndFlags)
+        countAndFlags: _StringObject.CountAndFlags(
+          sharedCount: _guts.count, isASCII: gutsCountAndFlags.isASCII))
     }
 
     _internalInvariant(_guts._object.hasObjCBridgeableObject,

stdlib/public/core/StringGuts.swift

@@ -61,8 +61,7 @@ extension _StringGuts {
   }
 
   internal init(_ storage: _SharedStringStorage) {
-    // TODO(cleanup): We should probably pass whole perf flags struct around
-    self.init(_StringObject(storage, isASCII: false))
+    self.init(_StringObject(storage))
   }
 
   internal init(
@@ -109,18 +108,15 @@ extension _StringGuts {
     @inline(__always) get { return isFastUTF8 && _object.isASCII }
   }
 
-  @inlinable
-  internal var isNFC: Bool  {
-    @inline(__always) get { return _object.isNFC }
-  }
+  @inline(__always)
+  internal var isNFC: Bool { return _object.isNFC }
 
-  @inlinable
-  internal var isNFCFastUTF8: Bool  {
+  @inline(__always)
+  internal var isNFCFastUTF8: Bool {
     // TODO(String micro-performance): Consider a dedicated bit for this
-    @inline(__always) get { return _object.isNFC && isFastUTF8 }
+    return _object.isNFC && isFastUTF8
   }
 
-  @inlinable
   internal var hasNativeStorage: Bool { return _object.hasNativeStorage }
 
   internal var hasSharedStorage: Bool { return _object.hasSharedStorage }