[clang-format] Add option to insert braces after control statements

Adds a new option InsertBraces to insert the optional braces after
if, else, for, while, and do in C++.

Differential Revision: https://reviews.llvm.org/D120217

clang/docs/ClangFormatStyleOptions.rst

@@ -2756,6 +2756,39 @@ the configuration (without a prefix: ``Auto``).
      LoooooooooooooooooooooooooooooooooooooooongReturnType
      LoooooooooooooooooooooooooooooooongFunctionDeclaration();
 
+**InsertBraces** (``Boolean``) :versionbadge:`clang-format 15`
+  Insert braces after control statements (``if``, ``else``, ``for``, ``do``,
+  and ``while``) in C++ unless the control statements are inside macro
+  definitions or the braces would enclose preprocessor directives.
+
+  .. warning:: 
+
+   Setting this option to `true` could lead to incorrect code formatting due
+   to clang-format's lack of complete semantic information. As such, extra
+   care should be taken to review code changes made by this option.
+
+  .. code-block:: c++
+
+    false:                                    true:
+
+    if (isa<FunctionDecl>(D))        vs.      if (isa<FunctionDecl>(D)) {
+      handleFunctionDecl(D);                    handleFunctionDecl(D);
+    else if (isa<VarDecl>(D))                 } else if (isa<VarDecl>(D)) {
+      handleVarDecl(D);                         handleVarDecl(D);
+    else                                      } else {
+      return;                                   return;
+                                              }
+
+    while (i--)                      vs.      while (i--) {
+      for (auto *A : D.attrs())                 for (auto *A : D.attrs()) {
+        handleAttr(A);                            handleAttr(A);
+                                                }
+                                              }
+
+    do                               vs.      do {
+      --i;                                      --i;
+    while (i);                                } while (i);
+
 **InsertTrailingCommas** (``TrailingCommaStyle``) :versionbadge:`clang-format 12`
   If set to ``TCS_Wrapped`` will insert trailing commas in container
   literals (arrays and objects) that wrap across multiple lines.

clang/docs/ReleaseNotes.rst

@@ -188,6 +188,9 @@ clang-format
 
 - Changed ``BreakBeforeConceptDeclarations`` from ``Boolean`` to an enum.
 
+- Option ``InsertBraces`` has been added to insert optional braces after control
+  statements.
+
 libclang
 --------
 

clang/include/clang/Format/Format.h

@@ -2571,6 +2571,38 @@ struct FormatStyle {
   /// \version 3.7
   bool IndentWrappedFunctionNames;
 
+  /// Insert braces after control statements (``if``, ``else``, ``for``, ``do``,
+  /// and ``while``) in C++ unless the control statements are inside macro
+  /// definitions or the braces would enclose preprocessor directives.
+  /// \warning
+  ///  Setting this option to `true` could lead to incorrect code formatting due
+  ///  to clang-format's lack of complete semantic information. As such, extra
+  ///  care should be taken to review code changes made by this option.
+  /// \endwarning
+  /// \code
+  ///   false:                                    true:
+  ///
+  ///   if (isa<FunctionDecl>(D))        vs.      if (isa<FunctionDecl>(D)) {
+  ///     handleFunctionDecl(D);                    handleFunctionDecl(D);
+  ///   else if (isa<VarDecl>(D))                 } else if (isa<VarDecl>(D)) {
+  ///     handleVarDecl(D);                         handleVarDecl(D);
+  ///   else                                      } else {
+  ///     return;                                   return;
+  ///                                             }
+  ///
+  ///   while (i--)                      vs.      while (i--) {
+  ///     for (auto *A : D.attrs())                 for (auto *A : D.attrs()) {
+  ///       handleAttr(A);                            handleAttr(A);
+  ///                                               }
+  ///                                             }
+  ///
+  ///   do                               vs.      do {
+  ///     --i;                                      --i;
+  ///   while (i);                                } while (i);
+  /// \endcode
+  /// \version 15
+  bool InsertBraces;
+
   /// A vector of prefixes ordered by the desired groups for Java imports.
   ///
   /// One group's prefix can be a subset of another - the longest prefix is

clang/lib/Format/Format.cpp

@@ -768,6 +768,7 @@ template <> struct MappingTraits<FormatStyle> {
     IO.mapOptional("IndentWidth", Style.IndentWidth);
     IO.mapOptional("IndentWrappedFunctionNames",
                    Style.IndentWrappedFunctionNames);
+    IO.mapOptional("InsertBraces", Style.InsertBraces);
     IO.mapOptional("InsertTrailingCommas", Style.InsertTrailingCommas);
     IO.mapOptional("JavaImportGroups", Style.JavaImportGroups);
     IO.mapOptional("JavaScriptQuotes", Style.JavaScriptQuotes);
@@ -1223,6 +1224,7 @@ FormatStyle getLLVMStyle(FormatStyle::LanguageKind Language) {
   LLVMStyle.IndentWrappedFunctionNames = false;
   LLVMStyle.IndentWidth = 2;
   LLVMStyle.PPIndentWidth = -1;
+  LLVMStyle.InsertBraces = false;
   LLVMStyle.InsertTrailingCommas = FormatStyle::TCS_None;
   LLVMStyle.JavaScriptQuotes = FormatStyle::JSQS_Leave;
   LLVMStyle.JavaScriptWrapImports = true;
@@ -1661,7 +1663,7 @@ ParseError validateQualifierOrder(FormatStyle *Style) {
     return ParseError::DuplicateQualifierSpecified;
   }
 
-  // Ensure the list has 'type' in it
+  // Ensure the list has 'type' in it.
   auto type = std::find(Style->QualifierOrder.begin(),
                         Style->QualifierOrder.end(), "type");
   if (type == Style->QualifierOrder.end())
@@ -1821,6 +1823,48 @@ class BracesRemover : public TokenAnalyzer {
   }
 };
 
+class BracesInserter : public TokenAnalyzer {
+public:
+  BracesInserter(const Environment &Env, const FormatStyle &Style)
+      : TokenAnalyzer(Env, Style) {}
+
+  std::pair<tooling::Replacements, unsigned>
+  analyze(TokenAnnotator &Annotator,
+          SmallVectorImpl<AnnotatedLine *> &AnnotatedLines,
+          FormatTokenLexer &Tokens) override {
+    AffectedRangeMgr.computeAffectedLines(AnnotatedLines);
+    tooling::Replacements Result;
+    insertBraces(AnnotatedLines, Result);
+    return {Result, 0};
+  }
+
+private:
+  void insertBraces(SmallVectorImpl<AnnotatedLine *> &Lines,
+                    tooling::Replacements &Result) {
+    const auto &SourceMgr = Env.getSourceManager();
+    for (AnnotatedLine *Line : Lines) {
+      insertBraces(Line->Children, Result);
+      if (!Line->Affected)
+        continue;
+      for (FormatToken *Token = Line->First; Token && !Token->Finalized;
+           Token = Token->Next) {
+        if (Token->BraceCount == 0)
+          continue;
+        std::string Brace;
+        if (Token->BraceCount < 0) {
+          assert(Token->BraceCount == -1);
+          Brace = '{';
+        } else {
+          Brace = std::string(Token->BraceCount, '}');
+        }
+        Token->BraceCount = 0;
+        const auto Start = Token->Tok.getEndLoc();
+        cantFail(Result.add(tooling::Replacement(SourceMgr, Start, 0, Brace)));
+      }
+    }
+  }
+};
+
 class JavaScriptRequoter : public TokenAnalyzer {
 public:
   JavaScriptRequoter(const Environment &Env, const FormatStyle &Style)
@@ -3133,6 +3177,11 @@ reformat(const FormatStyle &Style, StringRef Code,
     });
   }
 
+  if (Style.isCpp() && Style.InsertBraces)
+    Passes.emplace_back([&](const Environment &Env) {
+      return BracesInserter(Env, Expanded).process();
+    });
+
   if (Style.isCpp() && Style.RemoveBracesLLVM)
     Passes.emplace_back([&](const Environment &Env) {
       return BracesRemover(Env, Expanded).process();

clang/lib/Format/FormatToken.h

@@ -489,6 +489,12 @@ struct FormatToken {
   /// Is optional and can be removed.
   bool Optional = false;
 
+  /// Number of optional braces to be inserted after this token:
+  ///   -1: a single left brace
+  ///    0: no braces
+  ///   >0: number of right braces
+  int8_t BraceCount = 0;
+
   /// If this token starts a block, this contains all the unwrapped lines
   /// in it.
   SmallVector<AnnotatedLine *, 1> Children;

clang/lib/Format/UnwrappedLineParser.cpp

@@ -2300,6 +2300,53 @@ void UnwrappedLineParser::keepAncestorBraces() {
   NestedTooDeep.push_back(false);
 }
 
+static FormatToken *getLastNonComment(const UnwrappedLine &Line) {
+  for (const auto &Token : llvm::reverse(Line.Tokens))
+    if (Token.Tok->isNot(tok::comment))
+      return Token.Tok;
+
+  return nullptr;
+}
+
+void UnwrappedLineParser::parseUnbracedBody(bool CheckEOF) {
+  FormatToken *Tok = nullptr;
+
+  if (Style.InsertBraces && !Line->InPPDirective && !Line->Tokens.empty() &&
+      PreprocessorDirectives.empty()) {
+    Tok = getLastNonComment(*Line);
+    assert(Tok);
+    if (Tok->BraceCount < 0) {
+      assert(Tok->BraceCount == -1);
+      Tok = nullptr;
+    } else {
+      Tok->BraceCount = -1;
+    }
+  }
+
+  addUnwrappedLine();
+  ++Line->Level;
+  parseStructuralElement();
+
+  if (Tok) {
+    assert(!Line->InPPDirective);
+    Tok = nullptr;
+    for (const auto &L : llvm::reverse(*CurrentLines)) {
+      if (!L.InPPDirective) {
+        Tok = getLastNonComment(L);
+        if (Tok)
+          break;
+      }
+    }
+    assert(Tok);
+    ++Tok->BraceCount;
+  }
+
+  if (CheckEOF && FormatTok->is(tok::eof))
+    addUnwrappedLine();
+
+  --Line->Level;
+}
+
 static void markOptionalBraces(FormatToken *LeftBrace) {
   if (!LeftBrace)
     return;
@@ -2354,10 +2401,7 @@ FormatToken *UnwrappedLineParser::parseIfThenElse(IfStmtKind *IfKind,
     else
       NeedsUnwrappedLine = true;
   } else {
-    addUnwrappedLine();
-    ++Line->Level;
-    parseStructuralElement();
-    --Line->Level;
+    parseUnbracedBody();
   }
 
   bool KeepIfBraces = false;
@@ -2403,12 +2447,7 @@ FormatToken *UnwrappedLineParser::parseIfThenElse(IfStmtKind *IfKind,
       if (IsPrecededByComment)
         --Line->Level;
     } else {
-      addUnwrappedLine();
-      ++Line->Level;
-      parseStructuralElement();
-      if (FormatTok->is(tok::eof))
-        addUnwrappedLine();
-      --Line->Level;
+      parseUnbracedBody(/*CheckEOF=*/true);
     }
   } else {
     if (Style.RemoveBracesLLVM)
@@ -2654,10 +2693,7 @@ void UnwrappedLineParser::parseForOrWhileLoop() {
     }
     addUnwrappedLine();
   } else {
-    addUnwrappedLine();
-    ++Line->Level;
-    parseStructuralElement();
-    --Line->Level;
+    parseUnbracedBody();
   }
 
   if (Style.RemoveBracesLLVM)
@@ -2676,10 +2712,7 @@ void UnwrappedLineParser::parseDoWhile() {
     if (Style.BraceWrapping.BeforeWhile)
       addUnwrappedLine();
   } else {
-    addUnwrappedLine();
-    ++Line->Level;
-    parseStructuralElement();
-    --Line->Level;
+    parseUnbracedBody();
   }
 
   if (Style.RemoveBracesLLVM)

clang/lib/Format/UnwrappedLineParser.h

@@ -119,6 +119,7 @@ class UnwrappedLineParser {
   void parseParens(TokenType AmpAmpTokenType = TT_Unknown);
   void parseSquare(bool LambdaIntroducer = false);
   void keepAncestorBraces();
+  void parseUnbracedBody(bool CheckEOF = false);
   FormatToken *parseIfThenElse(IfStmtKind *IfKind, bool KeepBraces = false);
   void parseTryCatch();
   void parseForOrWhileLoop();