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 <string>
  19 #include <vector>
  20 #include <cassert>
  21
  22 namespace llvm {
  23 class MemoryBuffer;
  24 class SourceMgr;
  25 class SMLoc;
  26 class Twine;
  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, XCast, XSubst,
  49     XForEach, XHead, XTail, XEmpty, XIf, XEq,
  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   SourceMgr &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   /// Dependencies - This is the list of all included files.
  76   std::vector<std::string> Dependencies;
  77
  78 public:
  79   TGLexer(SourceMgr &SrcMgr);
  80   ~TGLexer() {}
  81
  82   tgtok::TokKind Lex() {
  83     return CurCode = LexToken();
  84   }
  85
  86   const std::vector<std::string> &getDependencies() const {
  87     return Dependencies;
  88   }
  89
  90   tgtok::TokKind getCode() const { return CurCode; }
  91
  92   const std::string &getCurStrVal() const {
  93     assert((CurCode == tgtok::Id || CurCode == tgtok::StrVal ||
  94             CurCode == tgtok::VarName || CurCode == tgtok::CodeFragment) &&
  95            "This token doesn't have a string value");
  96     return CurStrVal;
  97   }
  98   int64_t getCurIntVal() const {
  99     assert(CurCode == tgtok::IntVal && "This token isn't an integer");
 100     return CurIntVal;
 101   }
 102
 103   SMLoc getLoc() const;
 104
 105 private:
 106   /// LexToken - Read the next token and return its code.
 107   tgtok::TokKind LexToken();
 108
 109   tgtok::TokKind ReturnError(const char *Loc, const Twine &Msg);
 110
 111   int getNextChar();
 112   void SkipBCPLComment();
 113   bool SkipCComment();
 114   tgtok::TokKind LexIdentifier();
 115   bool LexInclude();
 116   tgtok::TokKind LexString();
 117   tgtok::TokKind LexVarName();
 118   tgtok::TokKind LexNumber();
 119   tgtok::TokKind LexBracket();
 120   tgtok::TokKind LexExclaim();
 121 };
 122
 123 } // end namespace llvm
 124
 125 #endif