utils/TableGen/TGLexer.h

   1 //===- TGLexer.h - Lexer for TableGen Files ---------------------*- 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 class represents the Lexer for tablegen files.
  11 //
  12 //===----------------------------------------------------------------------===//
  13
  14 #ifndef TGLEXER_H
  15 #define TGLEXER_H
  16
  17 #include "llvm/Support/DataTypes.h"
  18 #include <vector>
  19 #include <string>
  20 #include <iosfwd>
  21 #include <cassert>
  22
  23 namespace llvm {
  24 class MemoryBuffer;
  25 class TGSourceMgr;
  26 class TGLoc;
  27
  28 namespace tgtok {
  29   enum TokKind {
  30     // Markers
  31     Eof, Error,
  32
  33     // Tokens with no info.
  34     minus, plus,        // - +
  35     l_square, r_square, // [ ]
  36     l_brace, r_brace,   // { }
  37     l_paren, r_paren,   // ( )
  38     less, greater,      // < >
  39     colon, semi,        // ; :
  40     comma, period,      // , .
  41     equal, question,    // = ?
  42
  43     // Keywords.
  44     Bit, Bits, Class, Code, Dag, Def, Defm, Field, In, Int, Let, List,
  45     MultiClass, String,
  46
  47     // !keywords.
  48     XConcat, XSRA, XSRL, XSHL, XStrConcat, XNameConcat, XCast, XSubst,
  49     XForEach, XCar, XCdr, XNull, XIf,
  50
  51     // Integer value.
  52     IntVal,
  53
  54     // String valued tokens.
  55     Id, StrVal, VarName, CodeFragment
  56   };
  57 }
  58
  59 /// TGLexer - TableGen Lexer class.
  60 class TGLexer {
  61   TGSourceMgr &SrcMgr;
  62
  63   const char *CurPtr;
  64   const MemoryBuffer *CurBuf;
  65
  66   // Information about the current token.
  67   const char *TokStart;
  68   tgtok::TokKind CurCode;
  69   std::string CurStrVal;  // This is valid for ID, STRVAL, VARNAME, CODEFRAGMENT
  70   int64_t CurIntVal;      // This is valid for INTVAL.
  71
  72   /// CurBuffer - This is the current buffer index we're lexing from as managed
  73   /// by the SourceMgr object.
  74   int CurBuffer;
  75
  76   // IncludeDirectories - This is the list of directories we should search for
  77   // include files in.
  78   std::vector<std::string> IncludeDirectories;
  79 public:
  80   TGLexer(TGSourceMgr &SrcMgr);
  81   ~TGLexer() {}
  82
  83   void setIncludeDirs(const std::vector<std::string> &Dirs) {
  84     IncludeDirectories = Dirs;
  85   }
  86
  87   tgtok::TokKind Lex() {
  88     return CurCode = LexToken();
  89   }
  90
  91   tgtok::TokKind getCode() const { return CurCode; }
  92
  93   const std::string &getCurStrVal() const {
  94     assert((CurCode == tgtok::Id || CurCode == tgtok::StrVal ||
  95             CurCode == tgtok::VarName || CurCode == tgtok::CodeFragment) &&
  96            "This token doesn't have a string value");
  97     return CurStrVal;
  98   }
  99   int64_t getCurIntVal() const {
 100     assert(CurCode == tgtok::IntVal && "This token isn't an integer");
 101     return CurIntVal;
 102   }
 103
 104   TGLoc getLoc() const;
 105
 106   void PrintError(const char *Loc, const std::string &Msg) const;
 107   void PrintError(TGLoc Loc, const std::string &Msg) const;
 108
 109 private:
 110   /// LexToken - Read the next token and return its code.
 111   tgtok::TokKind LexToken();
 112
 113   tgtok::TokKind ReturnError(const char *Loc, const std::string &Msg);
 114
 115   int getNextChar();
 116   void SkipBCPLComment();
 117   bool SkipCComment();
 118   tgtok::TokKind LexIdentifier();
 119   bool LexInclude();
 120   tgtok::TokKind LexString();
 121   tgtok::TokKind LexVarName();
 122   tgtok::TokKind LexNumber();
 123   tgtok::TokKind LexBracket();
 124   tgtok::TokKind LexExclaim();
 125 };
 126
 127 } // end namespace llvm
 128
 129 #endif