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