1 //===- TokenLexer.h - Lex from a token buffer -------------------*- 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 // This file defines the TokenLexer interface. 11 // 12 //===----------------------------------------------------------------------===// 13 14 #ifndef LLVM_CLANG_LEX_TOKENLEXER_H 15 #define LLVM_CLANG_LEX_TOKENLEXER_H 16 17 #include "clang/Basic/SourceLocation.h" 18 #include "llvm/ADT/ArrayRef.h" 19 20 namespace clang { 21 22 class MacroArgs; 23 class MacroInfo; 24 class Preprocessor; 25 class Token; 26 class VAOptExpansionContext; 27 28 /// TokenLexer - This implements a lexer that returns tokens from a macro body 29 /// or token stream instead of lexing from a character buffer. This is used for 30 /// macro expansion and _Pragma handling, for example. 31 class TokenLexer { 32 friend class Preprocessor; 33 34 /// The macro we are expanding from. This is null if expanding a token stream. 35 MacroInfo *Macro = nullptr; 36 37 /// The actual arguments specified for a function-like macro, or null. The 38 /// TokenLexer owns the pointed-to object. 39 MacroArgs *ActualArgs = nullptr; 40 41 /// The current preprocessor object we are expanding for. 42 Preprocessor &PP; 43 44 /// This is the pointer to an array of tokens that the macro is 45 /// defined to, with arguments expanded for function-like macros. If this is 46 /// a token stream, these are the tokens we are returning. This points into 47 /// the macro definition we are lexing from, a cache buffer that is owned by 48 /// the preprocessor, or some other buffer that we may or may not own 49 /// (depending on OwnsTokens). 50 /// Note that if it points into Preprocessor's cache buffer, the Preprocessor 51 /// may update the pointer as needed. 52 const Token *Tokens; 53 54 /// This is the length of the Tokens array. 55 unsigned NumTokens; 56 57 /// This is the index of the next token that Lex will return. 58 unsigned CurTokenIdx; 59 60 /// The source location range where this macro was expanded. 61 SourceLocation ExpandLocStart, ExpandLocEnd; 62 63 /// Source location pointing at the source location entry chunk that 64 /// was reserved for the current macro expansion. 65 SourceLocation MacroExpansionStart; 66 67 /// The offset of the macro expansion in the 68 /// "source location address space". 69 unsigned MacroStartSLocOffset; 70 71 /// Location of the macro definition. 72 SourceLocation MacroDefStart; 73 74 /// Length of the macro definition. 75 unsigned MacroDefLength; 76 77 /// Lexical information about the expansion point of the macro: the identifier 78 /// that the macro expanded from had these properties. 79 bool AtStartOfLine : 1; 80 bool HasLeadingSpace : 1; 81 82 // When this is true, the next token appended to the 83 // output list during function argument expansion will get a leading space, 84 // regardless of whether it had one to begin with or not. This is used for 85 // placemarker support. If still true after function argument expansion, the 86 // leading space will be applied to the first token following the macro 87 // expansion. 88 bool NextTokGetsSpace : 1; 89 90 /// This is true if this TokenLexer allocated the Tokens 91 /// array, and thus needs to free it when destroyed. For simple object-like 92 /// macros (for example) we just point into the token buffer of the macro 93 /// definition, we don't make a copy of it. 94 bool OwnsTokens : 1; 95 96 /// This is true when tokens lexed from the TokenLexer 97 /// should not be subject to further macro expansion. 98 bool DisableMacroExpansion : 1; 99 100 public: 101 /// Create a TokenLexer for the specified macro with the specified actual 102 /// arguments. Note that this ctor takes ownership of the ActualArgs pointer. 103 /// ILEnd specifies the location of the ')' for a function-like macro or the 104 /// identifier for an object-like macro. TokenLexer(Token & Tok,SourceLocation ILEnd,MacroInfo * MI,MacroArgs * ActualArgs,Preprocessor & pp)105 TokenLexer(Token &Tok, SourceLocation ILEnd, MacroInfo *MI, 106 MacroArgs *ActualArgs, Preprocessor &pp) 107 : PP(pp), OwnsTokens(false) { 108 Init(Tok, ILEnd, MI, ActualArgs); 109 } 110 111 /// Create a TokenLexer for the specified token stream. If 'OwnsTokens' is 112 /// specified, this takes ownership of the tokens and delete[]'s them when 113 /// the token lexer is empty. TokenLexer(const Token * TokArray,unsigned NumToks,bool DisableExpansion,bool ownsTokens,Preprocessor & pp)114 TokenLexer(const Token *TokArray, unsigned NumToks, bool DisableExpansion, 115 bool ownsTokens, Preprocessor &pp) 116 : PP(pp), OwnsTokens(false) { 117 Init(TokArray, NumToks, DisableExpansion, ownsTokens); 118 } 119 120 TokenLexer(const TokenLexer &) = delete; 121 TokenLexer &operator=(const TokenLexer &) = delete; ~TokenLexer()122 ~TokenLexer() { destroy(); } 123 124 /// Initialize this TokenLexer to expand from the specified macro 125 /// with the specified argument information. Note that this ctor takes 126 /// ownership of the ActualArgs pointer. ILEnd specifies the location of the 127 /// ')' for a function-like macro or the identifier for an object-like macro. 128 void Init(Token &Tok, SourceLocation ELEnd, MacroInfo *MI, 129 MacroArgs *Actuals); 130 131 /// Initialize this TokenLexer with the specified token stream. 132 /// This does not take ownership of the specified token vector. 133 /// 134 /// DisableExpansion is true when macro expansion of tokens lexed from this 135 /// stream should be disabled. 136 void Init(const Token *TokArray, unsigned NumToks, 137 bool DisableMacroExpansion, bool OwnsTokens); 138 139 /// If the next token lexed will pop this macro off the 140 /// expansion stack, return 2. If the next unexpanded token is a '(', return 141 /// 1, otherwise return 0. 142 unsigned isNextTokenLParen() const; 143 144 /// Lex and return a token from this macro stream. 145 bool Lex(Token &Tok); 146 147 /// isParsingPreprocessorDirective - Return true if we are in the middle of a 148 /// preprocessor directive. 149 bool isParsingPreprocessorDirective() const; 150 151 private: 152 void destroy(); 153 154 /// Return true if the next lex call will pop this macro off the include 155 /// stack. isAtEnd()156 bool isAtEnd() const { 157 return CurTokenIdx == NumTokens; 158 } 159 160 /// Concatenates the next (sub-)sequence of \p Tokens separated by '##' 161 /// starting with LHSTok - stopping when we encounter a token that is neither 162 /// '##' nor preceded by '##'. Places the result back into \p LHSTok and sets 163 /// \p CurIdx to point to the token following the last one that was pasted. 164 /// 165 /// Also performs the MSVC extension wide-literal token pasting involved with: 166 /// \code L #macro-arg. \endcode 167 /// 168 /// \param[in,out] LHSTok - Contains the token to the left of '##' in \p 169 /// Tokens upon entry and will contain the resulting concatenated Token upon 170 /// exit. 171 /// 172 /// \param[in] TokenStream - The stream of Tokens we are lexing from. 173 /// 174 /// \param[in,out] CurIdx - Upon entry, \pTokens[\pCurIdx] must equal '##' 175 /// (with the exception of the MSVC extension mentioned above). Upon exit, it 176 /// is set to the index of the token following the last token that was 177 /// concatenated together. 178 /// 179 /// \returns If this returns true, the caller should immediately return the 180 /// token. 181 bool pasteTokens(Token &LHSTok, ArrayRef<Token> TokenStream, 182 unsigned int &CurIdx); 183 184 /// Calls pasteTokens above, passing in the '*this' object's Tokens and 185 /// CurTokenIdx data members. 186 bool pasteTokens(Token &Tok); 187 188 189 /// Takes the tail sequence of tokens within ReplacementToks that represent 190 /// the just expanded __VA_OPT__ tokens (possibly zero tokens) and transforms 191 /// them into a string. \p VCtx is used to determine which token represents 192 /// the first __VA_OPT__ replacement token. 193 /// 194 /// \param[in,out] ResultToks - Contains the current Replacement Tokens 195 /// (prior to rescanning and token pasting), the tail end of which represents 196 /// the tokens just expanded through __VA_OPT__ processing. These (sub) 197 /// sequence of tokens are folded into one stringified token. 198 /// 199 /// \param[in] VCtx - contains relevant contextual information about the 200 /// state of the tokens around and including the __VA_OPT__ token, necessary 201 /// for stringification. 202 void stringifyVAOPTContents(SmallVectorImpl<Token> &ResultToks, 203 const VAOptExpansionContext &VCtx, 204 SourceLocation VAOPTClosingParenLoc); 205 206 /// Expand the arguments of a function-like macro so that we can quickly 207 /// return preexpanded tokens from Tokens. 208 void ExpandFunctionArguments(); 209 210 /// In microsoft compatibility mode, /##/ pastes 211 /// together to form a comment that comments out everything in the current 212 /// macro, other active macros, and anything left on the current physical 213 /// source line of the expanded buffer. Handle this by returning the 214 /// first token on the next line. 215 void HandleMicrosoftCommentPaste(Token &Tok, SourceLocation OpLoc); 216 217 /// If \p loc is a FileID and points inside the current macro 218 /// definition, returns the appropriate source location pointing at the 219 /// macro expansion source location entry. 220 SourceLocation getExpansionLocForMacroDefLoc(SourceLocation loc) const; 221 222 /// Creates SLocEntries and updates the locations of macro argument 223 /// tokens to their new expanded locations. 224 /// 225 /// \param ArgIdSpellLoc the location of the macro argument id inside the 226 /// macro definition. 227 void updateLocForMacroArgTokens(SourceLocation ArgIdSpellLoc, 228 Token *begin_tokens, Token *end_tokens); 229 230 /// Remove comma ahead of __VA_ARGS__, if present, according to compiler 231 /// dialect settings. Returns true if the comma is removed. 232 bool MaybeRemoveCommaBeforeVaArgs(SmallVectorImpl<Token> &ResultToks, 233 bool HasPasteOperator, 234 MacroInfo *Macro, unsigned MacroArgNo, 235 Preprocessor &PP); 236 237 void PropagateLineStartLeadingSpaceInfo(Token &Result); 238 }; 239 240 } // namespace clang 241 242 #endif // LLVM_CLANG_LEX_TOKENLEXER_H 243