1 //===--- TokenAnnotator.h - Format C++ code ---------------------*- C++ -*-===// 2 // 3 // The LLVM Compiler Infrastructure 4 // 5 // This file is distributed under the University of Illinois Open Source 6 // License. See LICENSE.TXT for details. 7 // 8 //===----------------------------------------------------------------------===// 9 /// 10 /// \file 11 /// \brief This file implements a token annotator, i.e. creates 12 /// \c AnnotatedTokens out of \c FormatTokens with required extra information. 13 /// 14 //===----------------------------------------------------------------------===// 15 16 #ifndef LLVM_CLANG_LIB_FORMAT_TOKENANNOTATOR_H 17 #define LLVM_CLANG_LIB_FORMAT_TOKENANNOTATOR_H 18 19 #include "UnwrappedLineParser.h" 20 #include "clang/Format/Format.h" 21 #include <string> 22 23 namespace clang { 24 class SourceManager; 25 26 namespace format { 27 28 enum LineType { 29 LT_Invalid, 30 LT_ImportStatement, 31 LT_ObjCDecl, // An @interface, @implementation, or @protocol line. 32 LT_ObjCMethodDecl, 33 LT_ObjCProperty, // An @property line. 34 LT_Other, 35 LT_PreprocessorDirective, 36 LT_VirtualFunctionDecl 37 }; 38 39 class AnnotatedLine { 40 public: AnnotatedLine(const UnwrappedLine & Line)41 AnnotatedLine(const UnwrappedLine &Line) 42 : First(Line.Tokens.front().Tok), Level(Line.Level), 43 InPPDirective(Line.InPPDirective), 44 MustBeDeclaration(Line.MustBeDeclaration), MightBeFunctionDecl(false), 45 IsMultiVariableDeclStmt(false), Affected(false), 46 LeadingEmptyLinesAffected(false), ChildrenAffected(false) { 47 assert(!Line.Tokens.empty()); 48 49 // Calculate Next and Previous for all tokens. Note that we must overwrite 50 // Next and Previous for every token, as previous formatting runs might have 51 // left them in a different state. 52 First->Previous = nullptr; 53 FormatToken *Current = First; 54 for (std::list<UnwrappedLineNode>::const_iterator I = ++Line.Tokens.begin(), 55 E = Line.Tokens.end(); 56 I != E; ++I) { 57 const UnwrappedLineNode &Node = *I; 58 Current->Next = I->Tok; 59 I->Tok->Previous = Current; 60 Current = Current->Next; 61 Current->Children.clear(); 62 for (const auto &Child : Node.Children) { 63 Children.push_back(new AnnotatedLine(Child)); 64 Current->Children.push_back(Children.back()); 65 } 66 } 67 Last = Current; 68 Last->Next = nullptr; 69 } 70 ~AnnotatedLine()71 ~AnnotatedLine() { 72 for (unsigned i = 0, e = Children.size(); i != e; ++i) { 73 delete Children[i]; 74 } 75 FormatToken *Current = First; 76 while (Current) { 77 Current->Children.clear(); 78 Current->Role.reset(); 79 Current = Current->Next; 80 } 81 } 82 83 /// \c true if this line starts with the given tokens in order, ignoring 84 /// comments. startsWith(Ts...Tokens)85 template <typename... Ts> bool startsWith(Ts... Tokens) const { 86 return First && First->startsSequence(Tokens...); 87 } 88 89 /// \c true if this line ends with the given tokens in reversed order, 90 /// ignoring comments. 91 /// For example, given tokens [T1, T2, T3, ...], the function returns true if 92 /// this line is like "... T3 T2 T1". endsWith(Ts...Tokens)93 template <typename... Ts> bool endsWith(Ts... Tokens) const { 94 return Last && Last->endsSequence(Tokens...); 95 } 96 97 /// \c true if this line looks like a function definition instead of a 98 /// function declaration. Asserts MightBeFunctionDecl. mightBeFunctionDefinition()99 bool mightBeFunctionDefinition() const { 100 assert(MightBeFunctionDecl); 101 // FIXME: Line.Last points to other characters than tok::semi 102 // and tok::lbrace. 103 return !Last->isOneOf(tok::semi, tok::comment); 104 } 105 106 FormatToken *First; 107 FormatToken *Last; 108 109 SmallVector<AnnotatedLine *, 0> Children; 110 111 LineType Type; 112 unsigned Level; 113 bool InPPDirective; 114 bool MustBeDeclaration; 115 bool MightBeFunctionDecl; 116 bool IsMultiVariableDeclStmt; 117 118 /// \c True if this line should be formatted, i.e. intersects directly or 119 /// indirectly with one of the input ranges. 120 bool Affected; 121 122 /// \c True if the leading empty lines of this line intersect with one of the 123 /// input ranges. 124 bool LeadingEmptyLinesAffected; 125 126 /// \c True if a one of this line's children intersects with an input range. 127 bool ChildrenAffected; 128 129 private: 130 // Disallow copying. 131 AnnotatedLine(const AnnotatedLine &) = delete; 132 void operator=(const AnnotatedLine &) = delete; 133 }; 134 135 /// \brief Determines extra information about the tokens comprising an 136 /// \c UnwrappedLine. 137 class TokenAnnotator { 138 public: TokenAnnotator(const FormatStyle & Style,const AdditionalKeywords & Keywords)139 TokenAnnotator(const FormatStyle &Style, const AdditionalKeywords &Keywords) 140 : Style(Style), Keywords(Keywords) {} 141 142 /// \brief Adapts the indent levels of comment lines to the indent of the 143 /// subsequent line. 144 // FIXME: Can/should this be done in the UnwrappedLineParser? 145 void setCommentLineLevels(SmallVectorImpl<AnnotatedLine *> &Lines); 146 147 void annotate(AnnotatedLine &Line); 148 void calculateFormattingInformation(AnnotatedLine &Line); 149 150 private: 151 /// \brief Calculate the penalty for splitting before \c Tok. 152 unsigned splitPenalty(const AnnotatedLine &Line, const FormatToken &Tok, 153 bool InFunctionDecl); 154 155 bool spaceRequiredBetween(const AnnotatedLine &Line, const FormatToken &Left, 156 const FormatToken &Right); 157 158 bool spaceRequiredBefore(const AnnotatedLine &Line, const FormatToken &Tok); 159 160 bool mustBreakBefore(const AnnotatedLine &Line, const FormatToken &Right); 161 162 bool canBreakBefore(const AnnotatedLine &Line, const FormatToken &Right); 163 164 bool mustBreakForReturnType(const AnnotatedLine &Line) const; 165 166 void printDebugInfo(const AnnotatedLine &Line); 167 168 void calculateUnbreakableTailLengths(AnnotatedLine &Line); 169 170 const FormatStyle &Style; 171 172 const AdditionalKeywords &Keywords; 173 }; 174 175 } // end namespace format 176 } // end namespace clang 177 178 #endif 179