From: Chris Lattner Date: Sun, 9 Mar 2008 02:55:12 +0000 (+0000) Subject: split the MacroArgs class out of TokenLexer.cpp/h into X-Git-Url: https://granicus.if.org/sourcecode?a=commitdiff_plain;h=e5c8ffe09f8fec836b90f23c2a3b996bfc5b09c4;p=clang split the MacroArgs class out of TokenLexer.cpp/h into MacroArgs.cpp/h git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@48075 91177308-0d34-0410-b5e6-96231b3b80d8 --- diff --git a/Lex/MacroArgs.cpp b/Lex/MacroArgs.cpp new file mode 100644 index 0000000000..bebf5cd2d8 --- /dev/null +++ b/Lex/MacroArgs.cpp @@ -0,0 +1,224 @@ +//===--- TokenLexer.cpp - Lex from a token stream -------------------------===// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// +// +// This file implements the TokenLexer interface. +// +//===----------------------------------------------------------------------===// + +#include "MacroArgs.h" +#include "clang/Lex/MacroInfo.h" +#include "clang/Lex/Preprocessor.h" +#include "clang/Basic/Diagnostic.h" +using namespace clang; + +/// MacroArgs ctor function - This destroys the vector passed in. +MacroArgs *MacroArgs::create(const MacroInfo *MI, + const Token *UnexpArgTokens, + unsigned NumToks, bool VarargsElided) { + assert(MI->isFunctionLike() && + "Can't have args for an object-like macro!"); + + // Allocate memory for the MacroArgs object with the lexer tokens at the end. + MacroArgs *Result = (MacroArgs*)malloc(sizeof(MacroArgs) + + NumToks*sizeof(Token)); + // Construct the macroargs object. + new (Result) MacroArgs(NumToks, VarargsElided); + + // Copy the actual unexpanded tokens to immediately after the result ptr. + if (NumToks) + memcpy(const_cast(Result->getUnexpArgument(0)), + UnexpArgTokens, NumToks*sizeof(Token)); + + return Result; +} + +/// destroy - Destroy and deallocate the memory for this object. +/// +void MacroArgs::destroy() { + // Run the dtor to deallocate the vectors. + this->~MacroArgs(); + // Release the memory for the object. + free(this); +} + + +/// getArgLength - Given a pointer to an expanded or unexpanded argument, +/// return the number of tokens, not counting the EOF, that make up the +/// argument. +unsigned MacroArgs::getArgLength(const Token *ArgPtr) { + unsigned NumArgTokens = 0; + for (; ArgPtr->isNot(tok::eof); ++ArgPtr) + ++NumArgTokens; + return NumArgTokens; +} + + +/// getUnexpArgument - Return the unexpanded tokens for the specified formal. +/// +const Token *MacroArgs::getUnexpArgument(unsigned Arg) const { + // The unexpanded argument tokens start immediately after the MacroArgs object + // in memory. + const Token *Start = (const Token *)(this+1); + const Token *Result = Start; + // Scan to find Arg. + for (; Arg; ++Result) { + assert(Result < Start+NumUnexpArgTokens && "Invalid arg #"); + if (Result->is(tok::eof)) + --Arg; + } + return Result; +} + + +/// ArgNeedsPreexpansion - If we can prove that the argument won't be affected +/// by pre-expansion, return false. Otherwise, conservatively return true. +bool MacroArgs::ArgNeedsPreexpansion(const Token *ArgTok, + Preprocessor &PP) const { + // If there are no identifiers in the argument list, or if the identifiers are + // known to not be macros, pre-expansion won't modify it. + for (; ArgTok->isNot(tok::eof); ++ArgTok) + if (IdentifierInfo *II = ArgTok->getIdentifierInfo()) { + if (II->hasMacroDefinition() && PP.getMacroInfo(II)->isEnabled()) + // Return true even though the macro could be a function-like macro + // without a following '(' token. + return true; + } + return false; +} + +/// getPreExpArgument - Return the pre-expanded form of the specified +/// argument. +const std::vector & +MacroArgs::getPreExpArgument(unsigned Arg, Preprocessor &PP) { + assert(Arg < NumUnexpArgTokens && "Invalid argument number!"); + + // If we have already computed this, return it. + if (PreExpArgTokens.empty()) + PreExpArgTokens.resize(NumUnexpArgTokens); + + std::vector &Result = PreExpArgTokens[Arg]; + if (!Result.empty()) return Result; + + const Token *AT = getUnexpArgument(Arg); + unsigned NumToks = getArgLength(AT)+1; // Include the EOF. + + // Otherwise, we have to pre-expand this argument, populating Result. To do + // this, we set up a fake TokenLexer to lex from the unexpanded argument + // list. With this installed, we lex expanded tokens until we hit the EOF + // token at the end of the unexp list. + PP.EnterTokenStream(AT, NumToks); + + // Lex all of the macro-expanded tokens into Result. + do { + Result.push_back(Token()); + PP.Lex(Result.back()); + } while (Result.back().isNot(tok::eof)); + + // Pop the token stream off the top of the stack. We know that the internal + // pointer inside of it is to the "end" of the token stream, but the stack + // will not otherwise be popped until the next token is lexed. The problem is + // that the token may be lexed sometime after the vector of tokens itself is + // destroyed, which would be badness. + PP.RemoveTopOfLexerStack(); + return Result; +} + + +/// StringifyArgument - Implement C99 6.10.3.2p2, converting a sequence of +/// tokens into the literal string token that should be produced by the C # +/// preprocessor operator. If Charify is true, then it should be turned into +/// a character literal for the Microsoft charize (#@) extension. +/// +Token MacroArgs::StringifyArgument(const Token *ArgToks, + Preprocessor &PP, bool Charify) { + Token Tok; + Tok.startToken(); + Tok.setKind(tok::string_literal); + + const Token *ArgTokStart = ArgToks; + + // Stringify all the tokens. + std::string Result = "\""; + // FIXME: Optimize this loop to not use std::strings. + bool isFirst = true; + for (; ArgToks->isNot(tok::eof); ++ArgToks) { + const Token &Tok = *ArgToks; + if (!isFirst && (Tok.hasLeadingSpace() || Tok.isAtStartOfLine())) + Result += ' '; + isFirst = false; + + // If this is a string or character constant, escape the token as specified + // by 6.10.3.2p2. + if (Tok.is(tok::string_literal) || // "foo" + Tok.is(tok::wide_string_literal) || // L"foo" + Tok.is(tok::char_constant)) { // 'x' and L'x'. + Result += Lexer::Stringify(PP.getSpelling(Tok)); + } else { + // Otherwise, just append the token. + Result += PP.getSpelling(Tok); + } + } + + // If the last character of the string is a \, and if it isn't escaped, this + // is an invalid string literal, diagnose it as specified in C99. + if (Result[Result.size()-1] == '\\') { + // Count the number of consequtive \ characters. If even, then they are + // just escaped backslashes, otherwise it's an error. + unsigned FirstNonSlash = Result.size()-2; + // Guaranteed to find the starting " if nothing else. + while (Result[FirstNonSlash] == '\\') + --FirstNonSlash; + if ((Result.size()-1-FirstNonSlash) & 1) { + // Diagnose errors for things like: #define F(X) #X / F(\) + PP.Diag(ArgToks[-1], diag::pp_invalid_string_literal); + Result.erase(Result.end()-1); // remove one of the \'s. + } + } + Result += '"'; + + // If this is the charify operation and the result is not a legal character + // constant, diagnose it. + if (Charify) { + // First step, turn double quotes into single quotes: + Result[0] = '\''; + Result[Result.size()-1] = '\''; + + // Check for bogus character. + bool isBad = false; + if (Result.size() == 3) { + isBad = Result[1] == '\''; // ''' is not legal. '\' already fixed above. + } else { + isBad = (Result.size() != 4 || Result[1] != '\\'); // Not '\x' + } + + if (isBad) { + PP.Diag(ArgTokStart[0], diag::err_invalid_character_to_charify); + Result = "' '"; // Use something arbitrary, but legal. + } + } + + Tok.setLength(Result.size()); + Tok.setLocation(PP.CreateString(&Result[0], Result.size())); + return Tok; +} + +/// getStringifiedArgument - Compute, cache, and return the specified argument +/// that has been 'stringified' as required by the # operator. +const Token &MacroArgs::getStringifiedArgument(unsigned ArgNo, + Preprocessor &PP) { + assert(ArgNo < NumUnexpArgTokens && "Invalid argument number!"); + if (StringifiedArgs.empty()) { + StringifiedArgs.resize(getNumArguments()); + memset(&StringifiedArgs[0], 0, + sizeof(StringifiedArgs[0])*getNumArguments()); + } + if (StringifiedArgs[ArgNo].isNot(tok::string_literal)) + StringifiedArgs[ArgNo] = StringifyArgument(getUnexpArgument(ArgNo), PP); + return StringifiedArgs[ArgNo]; +} diff --git a/Lex/MacroArgs.h b/Lex/MacroArgs.h new file mode 100644 index 0000000000..4b22fa18aa --- /dev/null +++ b/Lex/MacroArgs.h @@ -0,0 +1,109 @@ +//===--- MacroArgs.h - Formal argument info for Macros ----------*- C++ -*-===// +// +// The LLVM Compiler Infrastructure +// +// This file is distributed under the University of Illinois Open Source +// License. See LICENSE.TXT for details. +// +//===----------------------------------------------------------------------===// +// +// This file defines the MacroArgs interface. +// +//===----------------------------------------------------------------------===// + +#ifndef LLVM_CLANG_MACROARGS_H +#define LLVM_CLANG_MACROARGS_H + +#include + +namespace clang { + class MacroInfo; + class Preprocessor; + class Token; + +/// MacroArgs - An instance of this class captures information about +/// the formal arguments specified to a function-like macro invocation. +class MacroArgs { + /// NumUnexpArgTokens - The number of raw, unexpanded tokens for the + /// arguments. All of the actual argument tokens are allocated immediately + /// after the MacroArgs object in memory. This is all of the arguments + /// concatenated together, with 'EOF' markers at the end of each argument. + unsigned NumUnexpArgTokens; + + /// PreExpArgTokens - Pre-expanded tokens for arguments that need them. Empty + /// if not yet computed. This includes the EOF marker at the end of the + /// stream. + std::vector > PreExpArgTokens; + + /// StringifiedArgs - This contains arguments in 'stringified' form. If the + /// stringified form of an argument has not yet been computed, this is empty. + std::vector StringifiedArgs; + + /// VarargsElided - True if this is a C99 style varargs macro invocation and + /// there was no argument specified for the "..." argument. If the argument + /// was specified (even empty) or this isn't a C99 style varargs function, or + /// if in strict mode and the C99 varargs macro had only a ... argument, this + /// is false. + bool VarargsElided; + + MacroArgs(unsigned NumToks, bool varargsElided) + : NumUnexpArgTokens(NumToks), VarargsElided(varargsElided) {} + ~MacroArgs() {} +public: + /// MacroArgs ctor function - Create a new MacroArgs object with the specified + /// macro and argument info. + static MacroArgs *create(const MacroInfo *MI, + const Token *UnexpArgTokens, + unsigned NumArgTokens, bool VarargsElided); + + /// destroy - Destroy and deallocate the memory for this object. + /// + void destroy(); + + /// ArgNeedsPreexpansion - If we can prove that the argument won't be affected + /// by pre-expansion, return false. Otherwise, conservatively return true. + bool ArgNeedsPreexpansion(const Token *ArgTok, Preprocessor &PP) const; + + /// getUnexpArgument - Return a pointer to the first token of the unexpanded + /// token list for the specified formal. + /// + const Token *getUnexpArgument(unsigned Arg) const; + + /// getArgLength - Given a pointer to an expanded or unexpanded argument, + /// return the number of tokens, not counting the EOF, that make up the + /// argument. + static unsigned getArgLength(const Token *ArgPtr); + + /// getPreExpArgument - Return the pre-expanded form of the specified + /// argument. + const std::vector & + getPreExpArgument(unsigned Arg, Preprocessor &PP); + + /// getStringifiedArgument - Compute, cache, and return the specified argument + /// that has been 'stringified' as required by the # operator. + const Token &getStringifiedArgument(unsigned ArgNo, Preprocessor &PP); + + /// getNumArguments - Return the number of arguments passed into this macro + /// invocation. + unsigned getNumArguments() const { return NumUnexpArgTokens; } + + + /// isVarargsElidedUse - Return true if this is a C99 style varargs macro + /// invocation and there was no argument specified for the "..." argument. If + /// the argument was specified (even empty) or this isn't a C99 style varargs + /// function, or if in strict mode and the C99 varargs macro had only a ... + /// argument, this returns false. + bool isVarargsElidedUse() const { return VarargsElided; } + + /// StringifyArgument - Implement C99 6.10.3.2p2, converting a sequence of + /// tokens into the literal string token that should be produced by the C # + /// preprocessor operator. If Charify is true, then it should be turned into + /// a character literal for the Microsoft charize (#@) extension. + /// + static Token StringifyArgument(const Token *ArgToks, + Preprocessor &PP, bool Charify = false); +}; + +} // end namespace clang + +#endif diff --git a/Lex/Preprocessor.cpp b/Lex/Preprocessor.cpp index 1cc4e6ecea..d8865441b0 100644 --- a/Lex/Preprocessor.cpp +++ b/Lex/Preprocessor.cpp @@ -26,6 +26,7 @@ //===----------------------------------------------------------------------===// #include "clang/Lex/Preprocessor.h" +#include "MacroArgs.h" #include "clang/Lex/HeaderSearch.h" #include "clang/Lex/MacroInfo.h" #include "clang/Lex/PPCallbacks.h" diff --git a/Lex/TokenLexer.cpp b/Lex/TokenLexer.cpp index 66549fdacc..2a5d6c5cf5 100644 --- a/Lex/TokenLexer.cpp +++ b/Lex/TokenLexer.cpp @@ -12,6 +12,7 @@ //===----------------------------------------------------------------------===// #include "clang/Lex/TokenLexer.h" +#include "MacroArgs.h" #include "clang/Lex/MacroInfo.h" #include "clang/Lex/Preprocessor.h" #include "clang/Basic/SourceManager.h" @@ -19,218 +20,6 @@ #include "llvm/ADT/SmallVector.h" using namespace clang; -//===----------------------------------------------------------------------===// -// MacroArgs Implementation -//===----------------------------------------------------------------------===// - -/// MacroArgs ctor function - This destroys the vector passed in. -MacroArgs *MacroArgs::create(const MacroInfo *MI, - const Token *UnexpArgTokens, - unsigned NumToks, bool VarargsElided) { - assert(MI->isFunctionLike() && - "Can't have args for an object-like macro!"); - - // Allocate memory for the MacroArgs object with the lexer tokens at the end. - MacroArgs *Result = (MacroArgs*)malloc(sizeof(MacroArgs) + - NumToks*sizeof(Token)); - // Construct the macroargs object. - new (Result) MacroArgs(NumToks, VarargsElided); - - // Copy the actual unexpanded tokens to immediately after the result ptr. - if (NumToks) - memcpy(const_cast(Result->getUnexpArgument(0)), - UnexpArgTokens, NumToks*sizeof(Token)); - - return Result; -} - -/// destroy - Destroy and deallocate the memory for this object. -/// -void MacroArgs::destroy() { - // Run the dtor to deallocate the vectors. - this->~MacroArgs(); - // Release the memory for the object. - free(this); -} - - -/// getArgLength - Given a pointer to an expanded or unexpanded argument, -/// return the number of tokens, not counting the EOF, that make up the -/// argument. -unsigned MacroArgs::getArgLength(const Token *ArgPtr) { - unsigned NumArgTokens = 0; - for (; ArgPtr->isNot(tok::eof); ++ArgPtr) - ++NumArgTokens; - return NumArgTokens; -} - - -/// getUnexpArgument - Return the unexpanded tokens for the specified formal. -/// -const Token *MacroArgs::getUnexpArgument(unsigned Arg) const { - // The unexpanded argument tokens start immediately after the MacroArgs object - // in memory. - const Token *Start = (const Token *)(this+1); - const Token *Result = Start; - // Scan to find Arg. - for (; Arg; ++Result) { - assert(Result < Start+NumUnexpArgTokens && "Invalid arg #"); - if (Result->is(tok::eof)) - --Arg; - } - return Result; -} - - -/// ArgNeedsPreexpansion - If we can prove that the argument won't be affected -/// by pre-expansion, return false. Otherwise, conservatively return true. -bool MacroArgs::ArgNeedsPreexpansion(const Token *ArgTok, - Preprocessor &PP) const { - // If there are no identifiers in the argument list, or if the identifiers are - // known to not be macros, pre-expansion won't modify it. - for (; ArgTok->isNot(tok::eof); ++ArgTok) - if (IdentifierInfo *II = ArgTok->getIdentifierInfo()) { - if (II->hasMacroDefinition() && PP.getMacroInfo(II)->isEnabled()) - // Return true even though the macro could be a function-like macro - // without a following '(' token. - return true; - } - return false; -} - -/// getPreExpArgument - Return the pre-expanded form of the specified -/// argument. -const std::vector & -MacroArgs::getPreExpArgument(unsigned Arg, Preprocessor &PP) { - assert(Arg < NumUnexpArgTokens && "Invalid argument number!"); - - // If we have already computed this, return it. - if (PreExpArgTokens.empty()) - PreExpArgTokens.resize(NumUnexpArgTokens); - - std::vector &Result = PreExpArgTokens[Arg]; - if (!Result.empty()) return Result; - - const Token *AT = getUnexpArgument(Arg); - unsigned NumToks = getArgLength(AT)+1; // Include the EOF. - - // Otherwise, we have to pre-expand this argument, populating Result. To do - // this, we set up a fake TokenLexer to lex from the unexpanded argument - // list. With this installed, we lex expanded tokens until we hit the EOF - // token at the end of the unexp list. - PP.EnterTokenStream(AT, NumToks); - - // Lex all of the macro-expanded tokens into Result. - do { - Result.push_back(Token()); - PP.Lex(Result.back()); - } while (Result.back().isNot(tok::eof)); - - // Pop the token stream off the top of the stack. We know that the internal - // pointer inside of it is to the "end" of the token stream, but the stack - // will not otherwise be popped until the next token is lexed. The problem is - // that the token may be lexed sometime after the vector of tokens itself is - // destroyed, which would be badness. - PP.RemoveTopOfLexerStack(); - return Result; -} - - -/// StringifyArgument - Implement C99 6.10.3.2p2, converting a sequence of -/// tokens into the literal string token that should be produced by the C # -/// preprocessor operator. -/// -static Token StringifyArgument(const Token *ArgToks, - Preprocessor &PP, bool Charify = false) { - Token Tok; - Tok.startToken(); - Tok.setKind(tok::string_literal); - - const Token *ArgTokStart = ArgToks; - - // Stringify all the tokens. - std::string Result = "\""; - // FIXME: Optimize this loop to not use std::strings. - bool isFirst = true; - for (; ArgToks->isNot(tok::eof); ++ArgToks) { - const Token &Tok = *ArgToks; - if (!isFirst && (Tok.hasLeadingSpace() || Tok.isAtStartOfLine())) - Result += ' '; - isFirst = false; - - // If this is a string or character constant, escape the token as specified - // by 6.10.3.2p2. - if (Tok.is(tok::string_literal) || // "foo" - Tok.is(tok::wide_string_literal) || // L"foo" - Tok.is(tok::char_constant)) { // 'x' and L'x'. - Result += Lexer::Stringify(PP.getSpelling(Tok)); - } else { - // Otherwise, just append the token. - Result += PP.getSpelling(Tok); - } - } - - // If the last character of the string is a \, and if it isn't escaped, this - // is an invalid string literal, diagnose it as specified in C99. - if (Result[Result.size()-1] == '\\') { - // Count the number of consequtive \ characters. If even, then they are - // just escaped backslashes, otherwise it's an error. - unsigned FirstNonSlash = Result.size()-2; - // Guaranteed to find the starting " if nothing else. - while (Result[FirstNonSlash] == '\\') - --FirstNonSlash; - if ((Result.size()-1-FirstNonSlash) & 1) { - // Diagnose errors for things like: #define F(X) #X / F(\) - PP.Diag(ArgToks[-1], diag::pp_invalid_string_literal); - Result.erase(Result.end()-1); // remove one of the \'s. - } - } - Result += '"'; - - // If this is the charify operation and the result is not a legal character - // constant, diagnose it. - if (Charify) { - // First step, turn double quotes into single quotes: - Result[0] = '\''; - Result[Result.size()-1] = '\''; - - // Check for bogus character. - bool isBad = false; - if (Result.size() == 3) { - isBad = Result[1] == '\''; // ''' is not legal. '\' already fixed above. - } else { - isBad = (Result.size() != 4 || Result[1] != '\\'); // Not '\x' - } - - if (isBad) { - PP.Diag(ArgTokStart[0], diag::err_invalid_character_to_charify); - Result = "' '"; // Use something arbitrary, but legal. - } - } - - Tok.setLength(Result.size()); - Tok.setLocation(PP.CreateString(&Result[0], Result.size())); - return Tok; -} - -/// getStringifiedArgument - Compute, cache, and return the specified argument -/// that has been 'stringified' as required by the # operator. -const Token &MacroArgs::getStringifiedArgument(unsigned ArgNo, - Preprocessor &PP) { - assert(ArgNo < NumUnexpArgTokens && "Invalid argument number!"); - if (StringifiedArgs.empty()) { - StringifiedArgs.resize(getNumArguments()); - memset(&StringifiedArgs[0], 0, - sizeof(StringifiedArgs[0])*getNumArguments()); - } - if (StringifiedArgs[ArgNo].isNot(tok::string_literal)) - StringifiedArgs[ArgNo] = StringifyArgument(getUnexpArgument(ArgNo), PP); - return StringifiedArgs[ArgNo]; -} - -//===----------------------------------------------------------------------===// -// TokenLexer Implementation -//===----------------------------------------------------------------------===// /// Create a TokenLexer for the specified macro with the specified actual /// arguments. Note that this ctor takes ownership of the ActualArgs pointer. @@ -329,7 +118,8 @@ void TokenLexer::ExpandFunctionArguments() { Res = ActualArgs->getStringifiedArgument(ArgNo, PP); else { // 'charify': don't bother caching these. - Res = StringifyArgument(ActualArgs->getUnexpArgument(ArgNo), PP, true); + Res = MacroArgs::StringifyArgument(ActualArgs->getUnexpArgument(ArgNo), + PP, true); } // The stringified/charified string leading space flag gets set to match diff --git a/clang.xcodeproj/project.pbxproj b/clang.xcodeproj/project.pbxproj index 31068b7e8d..e265404f40 100644 --- a/clang.xcodeproj/project.pbxproj +++ b/clang.xcodeproj/project.pbxproj @@ -103,6 +103,7 @@ DE75EDF10B06880E0020CF81 /* Type.cpp in Sources */ = {isa = PBXBuildFile; fileRef = DE75EDF00B06880E0020CF81 /* Type.cpp */; }; DE85CD4B0D8378320070E26E /* Directives.cpp in Sources */ = {isa = PBXBuildFile; fileRef = DE85CD4A0D8378320070E26E /* Directives.cpp */; }; DE85CD810D8380B10070E26E /* TokenLexer.cpp in Sources */ = {isa = PBXBuildFile; fileRef = DE85CD800D8380B10070E26E /* TokenLexer.cpp */; }; + DE85CDA30D8383B20070E26E /* MacroArgs.cpp in Sources */ = {isa = PBXBuildFile; fileRef = DE85CDA20D8383B20070E26E /* MacroArgs.cpp */; }; DE928B130C05659200231DA4 /* ModuleBuilder.cpp in Sources */ = {isa = PBXBuildFile; fileRef = DE928B120C05659200231DA4 /* ModuleBuilder.cpp */; }; DE928B200C0565B000231DA4 /* ModuleBuilder.h in CopyFiles */ = {isa = PBXBuildFile; fileRef = DE928B1F0C0565B000231DA4 /* ModuleBuilder.h */; }; DE928B7D0C0A615100231DA4 /* CodeGenModule.h in CopyFiles */ = {isa = PBXBuildFile; fileRef = DE928B7C0C0A615100231DA4 /* CodeGenModule.h */; }; @@ -365,6 +366,8 @@ DE85CD4A0D8378320070E26E /* Directives.cpp */ = {isa = PBXFileReference; fileEncoding = 4; lastKnownFileType = sourcecode.cpp.cpp; path = Directives.cpp; sourceTree = ""; }; DE85CD800D8380B10070E26E /* TokenLexer.cpp */ = {isa = PBXFileReference; fileEncoding = 4; lastKnownFileType = sourcecode.cpp.cpp; path = TokenLexer.cpp; sourceTree = ""; }; DE85CD840D8380F20070E26E /* TokenLexer.h */ = {isa = PBXFileReference; fileEncoding = 4; lastKnownFileType = sourcecode.c.h; path = TokenLexer.h; sourceTree = ""; }; + DE85CD9E0D8382DD0070E26E /* MacroArgs.h */ = {isa = PBXFileReference; fileEncoding = 4; lastKnownFileType = sourcecode.c.h; path = MacroArgs.h; sourceTree = ""; }; + DE85CDA20D8383B20070E26E /* MacroArgs.cpp */ = {isa = PBXFileReference; fileEncoding = 4; lastKnownFileType = sourcecode.cpp.cpp; path = MacroArgs.cpp; sourceTree = ""; }; DE928B120C05659200231DA4 /* ModuleBuilder.cpp */ = {isa = PBXFileReference; fileEncoding = 30; lastKnownFileType = sourcecode.cpp.cpp; name = ModuleBuilder.cpp; path = CodeGen/ModuleBuilder.cpp; sourceTree = ""; }; DE928B1F0C0565B000231DA4 /* ModuleBuilder.h */ = {isa = PBXFileReference; fileEncoding = 30; lastKnownFileType = sourcecode.c.h; name = ModuleBuilder.h; path = clang/CodeGen/ModuleBuilder.h; sourceTree = ""; }; DE928B7C0C0A615100231DA4 /* CodeGenModule.h */ = {isa = PBXFileReference; fileEncoding = 30; lastKnownFileType = sourcecode.c.h; name = CodeGenModule.h; path = CodeGen/CodeGenModule.h; sourceTree = ""; }; @@ -816,6 +819,8 @@ DE344B530AE5E46C00DBC861 /* HeaderSearch.cpp */, DED7D79E0A5242E6003AD0FB /* Lexer.cpp */, 1A869AA70BA21ABA008DA07A /* LiteralSupport.cpp */, + DE85CD9E0D8382DD0070E26E /* MacroArgs.h */, + DE85CDA20D8383B20070E26E /* MacroArgs.cpp */, DED7D7A00A5242E6003AD0FB /* MacroInfo.cpp */, DED7D7A20A5242E6003AD0FB /* PPExpressions.cpp */, DED7D7A30A5242E6003AD0FB /* Pragma.cpp */, @@ -983,6 +988,7 @@ 35D55B280D81D8C60092E734 /* CFRefCount.cpp in Sources */, DE85CD4B0D8378320070E26E /* Directives.cpp in Sources */, DE85CD810D8380B10070E26E /* TokenLexer.cpp in Sources */, + DE85CDA30D8383B20070E26E /* MacroArgs.cpp in Sources */, ); runOnlyForDeploymentPostprocessing = 0; }; diff --git a/include/clang/Lex/TokenLexer.h b/include/clang/Lex/TokenLexer.h index 12ef4ca8d7..0ecb4ec45d 100644 --- a/include/clang/Lex/TokenLexer.h +++ b/include/clang/Lex/TokenLexer.h @@ -7,7 +7,7 @@ // //===----------------------------------------------------------------------===// // -// This file defines the TokenLexer and MacroArgs interfaces. +// This file defines the TokenLexer interface. // //===----------------------------------------------------------------------===// @@ -15,88 +15,12 @@ #define LLVM_CLANG_TOKENLEXER_H #include "clang/Basic/SourceLocation.h" -#include namespace clang { class MacroInfo; class Preprocessor; class Token; - -/// MacroArgs - An instance of this class captures information about -/// the formal arguments specified to a function-like macro invocation. -class MacroArgs { - /// NumUnexpArgTokens - The number of raw, unexpanded tokens for the - /// arguments. All of the actual argument tokens are allocated immediately - /// after the MacroArgs object in memory. This is all of the arguments - /// concatenated together, with 'EOF' markers at the end of each argument. - unsigned NumUnexpArgTokens; - - /// PreExpArgTokens - Pre-expanded tokens for arguments that need them. Empty - /// if not yet computed. This includes the EOF marker at the end of the - /// stream. - std::vector > PreExpArgTokens; - - /// StringifiedArgs - This contains arguments in 'stringified' form. If the - /// stringified form of an argument has not yet been computed, this is empty. - std::vector StringifiedArgs; - - /// VarargsElided - True if this is a C99 style varargs macro invocation and - /// there was no argument specified for the "..." argument. If the argument - /// was specified (even empty) or this isn't a C99 style varargs function, or - /// if in strict mode and the C99 varargs macro had only a ... argument, this - /// is false. - bool VarargsElided; - - MacroArgs(unsigned NumToks, bool varargsElided) - : NumUnexpArgTokens(NumToks), VarargsElided(varargsElided) {} - ~MacroArgs() {} -public: - /// MacroArgs ctor function - Create a new MacroArgs object with the specified - /// macro and argument info. - static MacroArgs *create(const MacroInfo *MI, - const Token *UnexpArgTokens, - unsigned NumArgTokens, bool VarargsElided); - - /// destroy - Destroy and deallocate the memory for this object. - /// - void destroy(); - - /// ArgNeedsPreexpansion - If we can prove that the argument won't be affected - /// by pre-expansion, return false. Otherwise, conservatively return true. - bool ArgNeedsPreexpansion(const Token *ArgTok, Preprocessor &PP) const; - - /// getUnexpArgument - Return a pointer to the first token of the unexpanded - /// token list for the specified formal. - /// - const Token *getUnexpArgument(unsigned Arg) const; - - /// getArgLength - Given a pointer to an expanded or unexpanded argument, - /// return the number of tokens, not counting the EOF, that make up the - /// argument. - static unsigned getArgLength(const Token *ArgPtr); - - /// getPreExpArgument - Return the pre-expanded form of the specified - /// argument. - const std::vector & - getPreExpArgument(unsigned Arg, Preprocessor &PP); - - /// getStringifiedArgument - Compute, cache, and return the specified argument - /// that has been 'stringified' as required by the # operator. - const Token &getStringifiedArgument(unsigned ArgNo, Preprocessor &PP); - - /// getNumArguments - Return the number of arguments passed into this macro - /// invocation. - unsigned getNumArguments() const { return NumUnexpArgTokens; } - - - /// isVarargsElidedUse - Return true if this is a C99 style varargs macro - /// invocation and there was no argument specified for the "..." argument. If - /// the argument was specified (even empty) or this isn't a C99 style varargs - /// function, or if in strict mode and the C99 varargs macro had only a ... - /// argument, this returns false. - bool isVarargsElidedUse() const { return VarargsElided; } -}; - + class MacroArgs; /// TokenLexer - This implements a lexer that returns token from a macro body /// or token stream instead of lexing from a character buffer. This is used for