[libSyntax] Add a reference counted version of OwnedString

ahoppen · ahoppen · commit ac512d43415a · 2018-08-13T15:37:53.000-07:00
We cannot use unowned strings for token texts of incrementally parsed
syntax trees since the source buffer to which reused nodes refer will
have been freed for reused nodes. Always copying the token text whenever
OwnedString is passed is too expensive. A reference counted copy of the
string allows us to keep the token's string alive across incremental
parses while eliminating unnecessary copies.
diff --git a/include/swift/Basic/OwnedString.h b/include/swift/Basic/OwnedString.h
@@ -22,117 +22,89 @@
 
 #include "llvm/ADT/IntrusiveRefCntPtr.h"
 #include "llvm/ADT/StringRef.h"
+#include "llvm/Support/TrailingObjects.h"
 
 using llvm::StringRef;
 
 namespace swift {
 
-enum class StringOwnership {
-  /// An OwnedString holds a weak reference to the underlying string storage
-  /// and will never attempt to free it.
-  Unowned,
-
-  /// An OwnedString has its own copy of the underlying string storage and
-  /// will free the storage upon its destruction.
-  Copied,
-};
-
 /// Holds a string - either statically allocated or dynamically allocated
 /// and owned by this type.
 class OwnedString {
-  const char *Data;
-  size_t Length;
-  StringOwnership Ownership = StringOwnership::Unowned;
-
-  void release() {
-    if (Ownership == StringOwnership::Copied)
-      free(const_cast<char *>(Data));
-  }
+  /// An owner that keeps the buffer of a ref counted \c OwnedString alive.
+  class TextOwner final : public llvm::ThreadSafeRefCountedBase<TextOwner>,
+                          public llvm::TrailingObjects<TextOwner, char> {
+    TextOwner(StringRef Text) {
+      std::uninitialized_copy(Text.begin(), Text.end(),
+                              getTrailingObjects<char>());
+    }
 
-  void initialize(const char* Data, size_t Length, StringOwnership Ownership) {
-    this->Length = Length;
-    this->Ownership = Ownership;
-    if (Ownership == StringOwnership::Copied && Data) {
-      char *substring = static_cast<char *>(malloc(Length + 1));
-      assert(substring && "expected successful malloc of copy");
+  public:
+    static TextOwner *make(StringRef Text) {
+      auto size = totalSizeToAlloc<char>(Text.size());
+      void *data = ::operator new(size);
+      return new (data) TextOwner(Text);
+    }
 
-      memcpy(substring, Data, Length);
-      substring[Length] = '\0';
+    const char *getText() const { return getTrailingObjects<char>(); }
+  };
 
-      this->Data = substring;
-    }
-    else
-      this->Data = Data;
-  }
-  OwnedString(const char* Data, size_t Length, StringOwnership Ownership) {
-    initialize(Data, Length, Ownership);
-  }
-public:
-  OwnedString(): OwnedString(nullptr, 0, StringOwnership::Unowned) {}
+  /// The text this owned string represents
+  StringRef Text;
 
-  OwnedString(const char *Data, size_t Length):
-    OwnedString(Data, Length, StringOwnership::Copied) {}
+  /// In case of a ref counted string an owner that keeps the buffer \c Text
+  /// references alive.
+  llvm::IntrusiveRefCntPtr<TextOwner> OwnedPtr;
 
-  OwnedString(StringRef Str) : OwnedString(Str.data(), Str.size()) {}
+  OwnedString(StringRef Text, llvm::IntrusiveRefCntPtr<TextOwner> OwnedPtr)
+      : Text(Text), OwnedPtr(OwnedPtr) {}
 
-  OwnedString(const char *Data) : OwnedString(StringRef(Data)) {}
+public:
+  OwnedString() : OwnedString(/*Text=*/StringRef(), /*OwnedPtr=*/nullptr) {}
 
-  OwnedString(const OwnedString &Other):
-    OwnedString(Other.Data, Other.Length, Other.Ownership) {}
+  /// Create a ref counted \c OwnedString that is initialized with the text of
+  /// the given \c StringRef.
+  OwnedString(StringRef Str) : OwnedString(makeRefCounted(Str)) {}
 
-  OwnedString(OwnedString &&Other): Data(Other.Data), Length(Other.Length),
-      Ownership(Other.Ownership) {
-    Other.Data = nullptr;
-    Other.Ownership = StringOwnership::Unowned;
-  }
+  /// Create a ref counted \c OwnedString that is initialized with the text of
+  /// the given buffer.
+  OwnedString(const char *Str) : OwnedString(StringRef(Str)) {}
 
-  OwnedString& operator=(const OwnedString &Other) {
-    if (&Other != this) {
-      release();
-      initialize(Other.Data, Other.Length, Other.Ownership);
-    }
-    return *this;
+  /// Create an \c OwnedString that references the given string. The
+  /// \c OwnedString will not take ownership of that buffer and will assume that
+  /// the buffer outlives its lifetime.
+  static OwnedString makeUnowned(StringRef Str) {
+    return OwnedString(Str, /*OwnedPtr=*/nullptr);
   }
 
-  OwnedString& operator=(OwnedString &&Other) {
-    if (&Other != this) {
-      release();
-      this->Data = Other.Data;
-      this->Length = Other.Length;
-      this->Ownership = Other.Ownership;
-      Other.Ownership = StringOwnership::Unowned;
-      Other.Data = nullptr;
+  /// Create an \c OwnedString that keeps its contents in a reference counted
+  /// buffer. The contents of \p Str will be copied initially and are allowed to
+  /// be disposed after the \c OwnedString has been created.
+  static OwnedString makeRefCounted(StringRef Str) {
+    if (Str.empty()) {
+      // Copying an empty string doesn't make sense. Just create an unowned
+      // string that points to the empty string.
+      return makeUnowned(Str);
+    } else {
+      llvm::IntrusiveRefCntPtr<TextOwner> OwnedPtr(TextOwner::make(Str));
+      return OwnedString(StringRef(OwnedPtr->getText(), Str.size()),
+                         std::move(OwnedPtr));
     }
-    return *this;
-  }
-
-  OwnedString copy() const {
-    return OwnedString(Data, Length, StringOwnership::Copied);
   }
 
   /// Returns the length of the string in bytes.
-  size_t size() const {
-    return Length;
-  }
+  size_t size() const { return Text.size(); }
 
   /// Returns true if the length is 0.
-  bool empty() const {
-    return Length == 0;
-  }
+  bool empty() const { return size() == 0; }
 
   /// Returns a StringRef to the underlying data. No copy is made and no
   /// ownership changes take place.
-  StringRef str() const {
-    return StringRef { Data, Length };
-  }
+  StringRef str() const { return Text; }
 
   bool operator==(const OwnedString &Right) const {
     return str() == Right.str();
   }
-
-  ~OwnedString() {
-    release();
-  }
 };
 
 } // end namespace swift
diff --git a/include/swift/Syntax/RawSyntax.h b/include/swift/Syntax/RawSyntax.h
@@ -404,12 +404,17 @@ class RawSyntax final
     return static_cast<tok>(Bits.Token.TokenKind);
   }
 
-  /// Return the text of the token.
-  StringRef getTokenText() const {
+  /// Return the text of the token as an \c OwnedString. Keeping a reference to
+  /// this string will keep it alive even if the syntax node gets freed.
+  OwnedString getOwnedTokenText() const {
     assert(isToken());
-    return getTrailingObjects<OwnedString>()->str();
+    return *getTrailingObjects<OwnedString>();
   }
 
+  /// Return the text of the token as a reference. The referenced buffer may
+  /// disappear when the syntax node gets freed.
+  StringRef getTokenText() const { return getOwnedTokenText().str(); }
+
   /// Return the leading trivia list of the token.
   ArrayRef<TriviaPiece> getLeadingTrivia() const {
     assert(isToken());
@@ -434,7 +439,7 @@ class RawSyntax final
   /// trivia instead.
   RC<RawSyntax>
   withLeadingTrivia(ArrayRef<TriviaPiece> NewLeadingTrivia) const {
-    return make(getTokenKind(), getTokenText(), NewLeadingTrivia,
+    return make(getTokenKind(), getOwnedTokenText(), NewLeadingTrivia,
                 getTrailingTrivia(), getPresence());
   }
 
@@ -446,7 +451,7 @@ class RawSyntax final
   /// trivia instead.
   RC<RawSyntax>
   withTrailingTrivia(ArrayRef<TriviaPiece> NewTrailingTrivia) const {
-    return make(getTokenKind(), getTokenText(), getLeadingTrivia(),
+    return make(getTokenKind(), getOwnedTokenText(), getLeadingTrivia(),
                 NewTrailingTrivia, getPresence());
   }
 
diff --git a/include/swift/Syntax/Serialization/SyntaxDeserialization.h b/include/swift/Syntax/Serialization/SyntaxDeserialization.h
@@ -162,9 +162,9 @@ template <> struct MappingTraits<swift::RC<swift::RawSyntax>> {
       StringRef nodeIdString;
       in.mapRequired("id", nodeIdString);
       unsigned nodeId = std::atoi(nodeIdString.data());
-      value =
-          swift::RawSyntax::make(tokenKind, text, leadingTrivia, trailingTrivia,
-                                 presence, /*Arena=*/nullptr, nodeId);
+      value = swift::RawSyntax::make(
+          tokenKind, swift::OwnedString::makeRefCounted(text), leadingTrivia,
+          trailingTrivia, presence, /*Arena=*/nullptr, nodeId);
     } else {
       swift::SyntaxKind kind;
       in.mapRequired("kind", kind);
diff --git a/include/swift/Syntax/Trivia.h.gyb b/include/swift/Syntax/Trivia.h.gyb
@@ -454,7 +454,8 @@ struct MappingTraits<swift::syntax::TriviaPiece> {
 % else:
       StringRef text;
       in.mapRequired("value", text);
-      return swift::syntax::TriviaPiece(kind, text);
+      return swift::syntax::TriviaPiece(
+          kind, swift::OwnedString::makeRefCounted(text));
 % end
       break;
     }
diff --git a/lib/Parse/Lexer.cpp b/lib/Parse/Lexer.cpp
@@ -2182,7 +2182,8 @@ void Lexer::lexImpl() {
       size_t BOMLen = ContentStart - BufferStart;
       assert(BOMLen == 3 && "UTF-8 BOM is 3 bytes");
       // Add UTF-8 BOM to LeadingTrivia.
-      LeadingTrivia.push_back(TriviaPiece::garbageText({CurPtr, BOMLen}));
+      auto Text = OwnedString::makeRefCounted(StringRef(CurPtr, BOMLen));
+      LeadingTrivia.push_back(TriviaPiece::garbageText(Text));
       CurPtr += BOMLen;
     }
     NextToken.setAtStartOfLine(true);
@@ -2407,18 +2408,18 @@ void Lexer::lexTrivia(syntax::Trivia &Pieces, bool IsForTrailingTrivia) {
       bool isDocComment = CurPtr[1] == '/';
       skipSlashSlashComment(/*EatNewline=*/false);
       size_t Length = CurPtr - TriviaStart;
-      Pieces.push_back(isDocComment
-                           ? TriviaPiece::docLineComment({TriviaStart, Length})
-                           : TriviaPiece::lineComment({TriviaStart, Length}));
+      auto Text = OwnedString::makeRefCounted(StringRef(TriviaStart, Length));
+      Pieces.push_back(isDocComment ? TriviaPiece::docLineComment(Text)
+                                    : TriviaPiece::lineComment(Text));
       goto Restart;
     } else if (*CurPtr == '*') {
       // '/* ... */' comment.
       bool isDocComment = CurPtr[1] == '*';
       skipSlashStarComment();
       size_t Length = CurPtr - TriviaStart;
-      Pieces.push_back(isDocComment
-                           ? TriviaPiece::docBlockComment({TriviaStart, Length})
-                           : TriviaPiece::blockComment({TriviaStart, Length}));
+      auto Text = OwnedString::makeRefCounted(StringRef(TriviaStart, Length));
+      Pieces.push_back(isDocComment ? TriviaPiece::docBlockComment(Text)
+                                    : TriviaPiece::blockComment(Text));
       goto Restart;
     }
     break;
@@ -2430,7 +2431,8 @@ void Lexer::lexTrivia(syntax::Trivia &Pieces, bool IsForTrailingTrivia) {
         diagnose(TriviaStart, diag::lex_hashbang_not_allowed);
       skipHashbang(/*EatNewline=*/false);
       size_t Length = CurPtr - TriviaStart;
-      Pieces.push_back(TriviaPiece::garbageText({TriviaStart, Length}));
+      auto Text = OwnedString::makeRefCounted(StringRef(TriviaStart, Length));
+      Pieces.push_back(TriviaPiece::garbageText(Text));
       goto Restart;
     }
     break;
@@ -2439,7 +2441,8 @@ void Lexer::lexTrivia(syntax::Trivia &Pieces, bool IsForTrailingTrivia) {
     if (tryLexConflictMarker(/*EatNewline=*/false)) {
       // Conflict marker.
       size_t Length = CurPtr - TriviaStart;
-      Pieces.push_back(TriviaPiece::garbageText({TriviaStart, Length}));
+      auto Text = OwnedString::makeRefCounted(StringRef(TriviaStart, Length));
+      Pieces.push_back(TriviaPiece::garbageText(Text));
       goto Restart;
     }
     break;
@@ -2448,7 +2451,8 @@ void Lexer::lexTrivia(syntax::Trivia &Pieces, bool IsForTrailingTrivia) {
     case NulCharacterKind::Embedded: {
       diagnoseEmbeddedNul(Diags, CurPtr - 1);
       size_t Length = CurPtr - TriviaStart;
-      Pieces.push_back(TriviaPiece::garbageText({TriviaStart, Length}));
+      auto Text = OwnedString::makeRefCounted(StringRef(TriviaStart, Length));
+      Pieces.push_back(TriviaPiece::garbageText(Text));
       goto Restart;
     }
     case NulCharacterKind::CodeCompletion:
@@ -2494,7 +2498,8 @@ void Lexer::lexTrivia(syntax::Trivia &Pieces, bool IsForTrailingTrivia) {
     }
 
     size_t Length = CurPtr - TriviaStart;
-    Pieces.push_back(TriviaPiece::garbageText({TriviaStart, Length}));
+    auto Text = OwnedString::makeRefCounted(StringRef(TriviaStart, Length));
+    Pieces.push_back(TriviaPiece::garbageText(Text));
     goto Restart;
   }
   // Reset the cursor.
diff --git a/lib/Parse/Parser.cpp b/lib/Parse/Parser.cpp
@@ -312,15 +312,15 @@ swift::tokenizeWithTrivia(const LangOptions &LangOpts, const SourceManager &SM,
   syntax::AbsolutePosition RunningPos;
 
   tokenize(
-      LangOpts, SM, BufferID, Offset, EndOffset,
-      Diags,
+      LangOpts, SM, BufferID, Offset, EndOffset, Diags,
       CommentRetentionMode::AttachToNextToken, TriviaRetentionMode::WithTrivia,
       /*TokenizeInterpolatedString=*/false,
       /*SplitTokens=*/ArrayRef<Token>(),
       [&](const Token &Tok, const Trivia &LeadingTrivia,
           const Trivia &TrailingTrivia) {
+        auto Text = OwnedString::makeRefCounted(Tok.getText());
         auto ThisToken =
-            RawSyntax::make(Tok.getKind(), Tok.getText(), LeadingTrivia.Pieces,
+            RawSyntax::make(Tok.getKind(), Text, LeadingTrivia.Pieces,
                             TrailingTrivia.Pieces, SourcePresence::Present);
 
         auto ThisTokenPos = ThisToken->accumulateAbsolutePosition(RunningPos);
diff --git a/lib/Parse/SyntaxParsingContext.cpp b/lib/Parse/SyntaxParsingContext.cpp
@@ -163,9 +163,9 @@ void SyntaxParsingContext::addToken(Token &Tok, Trivia &LeadingTrivia,
     return;
 
   auto &Arena = getArena();
-  addRawSyntax(RawSyntax::getToken(Arena, Tok.getKind(), Tok.getText(),
-                                   LeadingTrivia.Pieces,
-                                   TrailingTrivia.Pieces));
+  auto Text = OwnedString::makeRefCounted(Tok.getText());
+  addRawSyntax(RawSyntax::getToken(
+      Arena, Tok.getKind(), Text, LeadingTrivia.Pieces, TrailingTrivia.Pieces));
 }
 
 /// Add Syntax to the parts.
@@ -313,7 +313,7 @@ void finalizeSourceFile(RootContextData &RootData,
   }
 
   if (!EOFToken)
-    EOFToken = RawSyntax::missing(tok::eof, "");
+    EOFToken = RawSyntax::missing(tok::eof, OwnedString::makeUnowned(""));
 
   auto newRaw = SyntaxFactory::createRaw(
       SyntaxKind::SourceFile,
@@ -352,7 +352,8 @@ void SyntaxParsingContext::synthesize(tok Kind, StringRef Text) {
     return;
   if (Text.empty())
     Text = getTokenText(Kind);
-  getStorage().push_back(RawSyntax::missing(Kind, Text));
+  auto OwnedText = OwnedString::makeRefCounted(Text);
+  getStorage().push_back(RawSyntax::missing(Kind, OwnedText));
 }
 
 void SyntaxParsingContext::synthesize(SyntaxKind Kind) {
diff --git a/lib/Syntax/SyntaxFactory.cpp.gyb b/lib/Syntax/SyntaxFactory.cpp.gyb
diff --git a/unittests/Basic/OwnedStringTest.cpp b/unittests/Basic/OwnedStringTest.cpp
diff --git a/utils/gyb_syntax_support/__init__.py b/utils/gyb_syntax_support/__init__.py
