1 //===--- UnwrappedLineParser.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 contains the declaration of the UnwrappedLineParser, 12 /// which turns a stream of tokens into UnwrappedLines. 13 /// 14 //===----------------------------------------------------------------------===// 15 16 #ifndef LLVM_CLANG_LIB_FORMAT_UNWRAPPEDLINEPARSER_H 17 #define LLVM_CLANG_LIB_FORMAT_UNWRAPPEDLINEPARSER_H 18 19 #include "FormatToken.h" 20 #include "clang/Basic/IdentifierTable.h" 21 #include "clang/Format/Format.h" 22 #include <list> 23 #include <stack> 24 25 namespace clang { 26 namespace format { 27 28 struct UnwrappedLineNode; 29 30 /// \brief An unwrapped line is a sequence of \c Token, that we would like to 31 /// put on a single line if there was no column limit. 32 /// 33 /// This is used as a main interface between the \c UnwrappedLineParser and the 34 /// \c UnwrappedLineFormatter. The key property is that changing the formatting 35 /// within an unwrapped line does not affect any other unwrapped lines. 36 struct UnwrappedLine { 37 UnwrappedLine(); 38 39 // FIXME: Don't use std::list here. 40 /// \brief The \c Tokens comprising this \c UnwrappedLine. 41 std::list<UnwrappedLineNode> Tokens; 42 43 /// \brief The indent level of the \c UnwrappedLine. 44 unsigned Level; 45 46 /// \brief Whether this \c UnwrappedLine is part of a preprocessor directive. 47 bool InPPDirective; 48 49 bool MustBeDeclaration; 50 }; 51 52 class UnwrappedLineConsumer { 53 public: 54 virtual ~UnwrappedLineConsumer() {} 55 virtual void consumeUnwrappedLine(const UnwrappedLine &Line) = 0; 56 virtual void finishRun() = 0; 57 }; 58 59 class FormatTokenSource; 60 61 class UnwrappedLineParser { 62 public: 63 UnwrappedLineParser(const FormatStyle &Style, 64 const AdditionalKeywords &Keywords, 65 ArrayRef<FormatToken *> Tokens, 66 UnwrappedLineConsumer &Callback); 67 68 void parse(); 69 70 private: 71 void reset(); 72 void parseFile(); 73 void parseLevel(bool HasOpeningBrace); 74 void parseBlock(bool MustBeDeclaration, bool AddLevel = true, 75 bool MunchSemi = true); 76 void parseChildBlock(); 77 void parsePPDirective(); 78 void parsePPDefine(); 79 void parsePPIf(bool IfDef); 80 void parsePPElIf(); 81 void parsePPElse(); 82 void parsePPEndIf(); 83 void parsePPUnknown(); 84 void readTokenWithJavaScriptASI(); 85 void parseStructuralElement(); 86 bool tryToParseBracedList(); 87 bool parseBracedList(bool ContinueOnSemicolons = false); 88 void parseParens(); 89 void parseSquare(); 90 void parseIfThenElse(); 91 void parseTryCatch(); 92 void parseForOrWhileLoop(); 93 void parseDoWhile(); 94 void parseLabel(); 95 void parseCaseLabel(); 96 void parseSwitch(); 97 void parseNamespace(); 98 void parseNew(); 99 void parseAccessSpecifier(); 100 bool parseEnum(); 101 void parseJavaEnumBody(); 102 void parseRecord(); 103 void parseObjCProtocolList(); 104 void parseObjCUntilAtEnd(); 105 void parseObjCInterfaceOrImplementation(); 106 void parseObjCProtocol(); 107 void parseJavaScriptEs6ImportExport(); 108 bool tryToParseLambda(); 109 bool tryToParseLambdaIntroducer(); 110 void tryToParseJSFunction(); 111 void addUnwrappedLine(); 112 bool eof() const; 113 void nextToken(); 114 const FormatToken *getPreviousToken(); 115 void readToken(); 116 void flushComments(bool NewlineBeforeNext); 117 void pushToken(FormatToken *Tok); 118 void calculateBraceTypes(bool ExpectClassBody = false); 119 120 // Marks a conditional compilation edge (for example, an '#if', '#ifdef', 121 // '#else' or merge conflict marker). If 'Unreachable' is true, assumes 122 // this branch either cannot be taken (for example '#if false'), or should 123 // not be taken in this round. 124 void conditionalCompilationCondition(bool Unreachable); 125 void conditionalCompilationStart(bool Unreachable); 126 void conditionalCompilationAlternative(); 127 void conditionalCompilationEnd(); 128 129 bool isOnNewLine(const FormatToken &FormatTok); 130 131 // FIXME: We are constantly running into bugs where Line.Level is incorrectly 132 // subtracted from beyond 0. Introduce a method to subtract from Line.Level 133 // and use that everywhere in the Parser. 134 std::unique_ptr<UnwrappedLine> Line; 135 136 // Comments are sorted into unwrapped lines by whether they are in the same 137 // line as the previous token, or not. If not, they belong to the next token. 138 // Since the next token might already be in a new unwrapped line, we need to 139 // store the comments belonging to that token. 140 SmallVector<FormatToken *, 1> CommentsBeforeNextToken; 141 FormatToken *FormatTok; 142 bool MustBreakBeforeNextToken; 143 144 // The parsed lines. Only added to through \c CurrentLines. 145 SmallVector<UnwrappedLine, 8> Lines; 146 147 // Preprocessor directives are parsed out-of-order from other unwrapped lines. 148 // Thus, we need to keep a list of preprocessor directives to be reported 149 // after an unwarpped line that has been started was finished. 150 SmallVector<UnwrappedLine, 4> PreprocessorDirectives; 151 152 // New unwrapped lines are added via CurrentLines. 153 // Usually points to \c &Lines. While parsing a preprocessor directive when 154 // there is an unfinished previous unwrapped line, will point to 155 // \c &PreprocessorDirectives. 156 SmallVectorImpl<UnwrappedLine> *CurrentLines; 157 158 // We store for each line whether it must be a declaration depending on 159 // whether we are in a compound statement or not. 160 std::vector<bool> DeclarationScopeStack; 161 162 const FormatStyle &Style; 163 const AdditionalKeywords &Keywords; 164 165 FormatTokenSource *Tokens; 166 UnwrappedLineConsumer &Callback; 167 168 // FIXME: This is a temporary measure until we have reworked the ownership 169 // of the format tokens. The goal is to have the actual tokens created and 170 // owned outside of and handed into the UnwrappedLineParser. 171 ArrayRef<FormatToken *> AllTokens; 172 173 // Represents preprocessor branch type, so we can find matching 174 // #if/#else/#endif directives. 175 enum PPBranchKind { 176 PP_Conditional, // Any #if, #ifdef, #ifndef, #elif, block outside #if 0 177 PP_Unreachable // #if 0 or a conditional preprocessor block inside #if 0 178 }; 179 180 // Keeps a stack of currently active preprocessor branching directives. 181 SmallVector<PPBranchKind, 16> PPStack; 182 183 // The \c UnwrappedLineParser re-parses the code for each combination 184 // of preprocessor branches that can be taken. 185 // To that end, we take the same branch (#if, #else, or one of the #elif 186 // branches) for each nesting level of preprocessor branches. 187 // \c PPBranchLevel stores the current nesting level of preprocessor 188 // branches during one pass over the code. 189 int PPBranchLevel; 190 191 // Contains the current branch (#if, #else or one of the #elif branches) 192 // for each nesting level. 193 SmallVector<int, 8> PPLevelBranchIndex; 194 195 // Contains the maximum number of branches at each nesting level. 196 SmallVector<int, 8> PPLevelBranchCount; 197 198 // Contains the number of branches per nesting level we are currently 199 // in while parsing a preprocessor branch sequence. 200 // This is used to update PPLevelBranchCount at the end of a branch 201 // sequence. 202 std::stack<int> PPChainBranchIndex; 203 204 friend class ScopedLineState; 205 friend class CompoundStatementIndenter; 206 }; 207 208 struct UnwrappedLineNode { 209 UnwrappedLineNode() : Tok(nullptr) {} 210 UnwrappedLineNode(FormatToken *Tok) : Tok(Tok) {} 211 212 FormatToken *Tok; 213 SmallVector<UnwrappedLine, 0> Children; 214 }; 215 216 inline UnwrappedLine::UnwrappedLine() 217 : Level(0), InPPDirective(false), MustBeDeclaration(false) {} 218 219 } // end namespace format 220 } // end namespace clang 221 222 #endif 223