tools/llvm-mc/AsmLexer.h

   1 //===- AsmLexer.h - Lexer for Assembly 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 declares the lexer for assembly files.
  11 //
  12 //===----------------------------------------------------------------------===//
  13
  14 #ifndef ASMLEXER_H
  15 #define ASMLEXER_H
  16
  17 #include "llvm/Support/DataTypes.h"
  18 #include <string>
  19 #include <cassert>
  20
  21 namespace llvm {
  22 class MemoryBuffer;
  23 class SourceMgr;
  24 class SMLoc;
  25
  26 namespace asmtok {
  27   enum TokKind {
  28     // Markers
  29     Eof, Error,
  30
  31     // String values.
  32     Identifier,
  33     Register,
  34     String,
  35
  36     // Integer values.
  37     IntVal,
  38
  39     // No-value.
  40     EndOfStatement,
  41     Colon,
  42     Plus, Minus, Tilde,
  43     Slash,    // '/'
  44     LParen, RParen,
  45     Star, Comma, Dollar, Equal, EqualEqual,
  46
  47     Pipe, PipePipe, Caret,
  48     Amp, AmpAmp, Exclaim, ExclaimEqual, Percent,
  49     Less, LessEqual, LessLess, LessGreater,
  50     Greater, GreaterEqual, GreaterGreater
  51   };
  52 }
  53
  54 /// AsmLexer - Lexer class for assembly files.
  55 class AsmLexer {
  56   SourceMgr &SrcMgr;
  57
  58   const char *CurPtr;
  59   const MemoryBuffer *CurBuf;
  60   // A llvm::StringSet<>, which provides uniqued and null-terminated strings.
  61   void *TheStringSet;
  62
  63   // Information about the current token.
  64   const char *TokStart;
  65   asmtok::TokKind CurKind;
  66   const char *CurStrVal;  // This is valid for Identifier.
  67   int64_t CurIntVal;
  68
  69   /// CurBuffer - This is the current buffer index we're lexing from as managed
  70   /// by the SourceMgr object.
  71   int CurBuffer;
  72
  73   void operator=(const AsmLexer&); // DO NOT IMPLEMENT
  74   AsmLexer(const AsmLexer&);       // DO NOT IMPLEMENT
  75 public:
  76   AsmLexer(SourceMgr &SrcMgr);
  77   ~AsmLexer();
  78
  79   asmtok::TokKind Lex() {
  80     return CurKind = LexToken();
  81   }
  82
  83   asmtok::TokKind getKind() const { return CurKind; }
  84   bool is(asmtok::TokKind K) const { return CurKind == K; }
  85   bool isNot(asmtok::TokKind K) const { return CurKind != K; }
  86
  87   const char *getCurStrVal() const {
  88     assert((CurKind == asmtok::Identifier || CurKind == asmtok::Register ||
  89             CurKind == asmtok::String) &&
  90            "This token doesn't have a string value");
  91     return CurStrVal;
  92   }
  93   int64_t getCurIntVal() const {
  94     assert(CurKind == asmtok::IntVal && "This token isn't an integer");
  95     return CurIntVal;
  96   }
  97
  98   SMLoc getLoc() const;
  99
 100   /// EnterIncludeFile - Enter the specified file. This returns true on failure.
 101   bool EnterIncludeFile(const std::string &Filename);
 102
 103   void PrintMessage(SMLoc Loc, const std::string &Msg, const char *Type) const;
 104
 105 private:
 106   int getNextChar();
 107   asmtok::TokKind ReturnError(const char *Loc, const std::string &Msg);
 108
 109   /// LexToken - Read the next token and return its code.
 110   asmtok::TokKind LexToken();
 111   asmtok::TokKind LexIdentifier();
 112   asmtok::TokKind LexPercent();
 113   asmtok::TokKind LexSlash();
 114   asmtok::TokKind LexLineComment();
 115   asmtok::TokKind LexDigit();
 116   asmtok::TokKind LexQuote();
 117 };
 118
 119 } // end namespace llvm
 120
 121 #endif