blob: 2918e43b6e45e8214a39ff1f9888e211ab87d582 [file] [log] [blame]
Chris Lattnera59e8772009-06-21 07:19:10 +00001//===- 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
Daniel Dunbar9a7e2cc2009-07-27 21:49:56 +000017#include "llvm/ADT/StringRef.h"
Daniel Dunbardbd692a2009-07-20 20:01:54 +000018#include "llvm/MC/MCAsmLexer.h"
Chris Lattnera59e8772009-06-21 07:19:10 +000019#include "llvm/Support/DataTypes.h"
20#include <string>
21#include <cassert>
22
23namespace llvm {
24class MemoryBuffer;
25class SourceMgr;
26class SMLoc;
27
Daniel Dunbar3f872332009-07-28 16:08:33 +000028/// AsmToken - Target independent representation for an assembler token.
29struct AsmToken {
30 enum TokenKind {
Chris Lattnera59e8772009-06-21 07:19:10 +000031 // Markers
32 Eof, Error,
33
Chris Lattner10a907d2009-06-21 19:56:35 +000034 // String values.
Chris Lattnera59e8772009-06-21 07:19:10 +000035 Identifier,
Chris Lattner4651bca2009-06-21 19:21:25 +000036 Register,
Chris Lattner10a907d2009-06-21 19:56:35 +000037 String,
38
39 // Integer values.
Daniel Dunbar3f872332009-07-28 16:08:33 +000040 Integer,
Chris Lattnera59e8772009-06-21 07:19:10 +000041
Chris Lattner10a907d2009-06-21 19:56:35 +000042 // No-value.
Chris Lattner4651bca2009-06-21 19:21:25 +000043 EndOfStatement,
Chris Lattnera59e8772009-06-21 07:19:10 +000044 Colon,
Chris Lattner74ec1a32009-06-22 06:32:03 +000045 Plus, Minus, Tilde,
Chris Lattner4651bca2009-06-21 19:21:25 +000046 Slash, // '/'
47 LParen, RParen,
Daniel Dunbar475839e2009-06-29 20:37:27 +000048 Star, Comma, Dollar, Equal, EqualEqual,
Chris Lattner8dfbe6c2009-06-23 05:57:07 +000049
Daniel Dunbar475839e2009-06-29 20:37:27 +000050 Pipe, PipePipe, Caret,
51 Amp, AmpAmp, Exclaim, ExclaimEqual, Percent,
52 Less, LessEqual, LessLess, LessGreater,
53 Greater, GreaterEqual, GreaterGreater
Chris Lattnera59e8772009-06-21 07:19:10 +000054 };
Chris Lattnera59e8772009-06-21 07:19:10 +000055
Daniel Dunbar3f872332009-07-28 16:08:33 +000056 TokenKind Kind;
Daniel Dunbarcb358b62009-07-28 03:00:54 +000057
58 /// A reference to the entire token contents; this is always a pointer into
59 /// a memory buffer owned by the source manager.
60 StringRef Str;
61
62 int64_t IntVal;
63
64public:
65 AsmToken() {}
Daniel Dunbar3f872332009-07-28 16:08:33 +000066 AsmToken(TokenKind _Kind, const StringRef &_Str, int64_t _IntVal = 0)
Daniel Dunbarcb358b62009-07-28 03:00:54 +000067 : Kind(_Kind), Str(_Str), IntVal(_IntVal) {}
68
Daniel Dunbar3f872332009-07-28 16:08:33 +000069 TokenKind getKind() const { return Kind; }
70 bool is(TokenKind K) const { return Kind == K; }
71 bool isNot(TokenKind K) const { return Kind != K; }
Daniel Dunbarcb358b62009-07-28 03:00:54 +000072
73 SMLoc getLoc() const;
74
Daniel Dunbar419aded2009-07-28 16:38:40 +000075 /// getString - Get the string for the current token, this includes all
76 /// characters (for example, the quotes on strings) in the token.
77 ///
78 /// The returned StringRef points into the source manager's memory buffer, and
79 /// is safe to store across calls to Lex().
Daniel Dunbarcb358b62009-07-28 03:00:54 +000080 StringRef getString() const { return Str; }
81
Daniel Dunbar3f872332009-07-28 16:08:33 +000082 // FIXME: Don't compute this in advance, it makes every token larger, and is
83 // also not generally what we want (it is nicer for recovery etc. to lex 123br
84 // as a single token, then diagnose as an invalid number).
Daniel Dunbarcb358b62009-07-28 03:00:54 +000085 int64_t getIntVal() const {
Daniel Dunbar3f872332009-07-28 16:08:33 +000086 assert(Kind == Integer && "This token isn't an integer");
Daniel Dunbarcb358b62009-07-28 03:00:54 +000087 return IntVal;
88 }
89};
90
Chris Lattnera59e8772009-06-21 07:19:10 +000091/// AsmLexer - Lexer class for assembly files.
Daniel Dunbardbd692a2009-07-20 20:01:54 +000092class AsmLexer : public MCAsmLexer {
Chris Lattnera59e8772009-06-21 07:19:10 +000093 SourceMgr &SrcMgr;
94
95 const char *CurPtr;
96 const MemoryBuffer *CurBuf;
97
Chris Lattnera59e8772009-06-21 07:19:10 +000098 const char *TokStart;
Daniel Dunbarcb358b62009-07-28 03:00:54 +000099
100 /// The current token.
101 AsmToken CurTok;
Chris Lattnera59e8772009-06-21 07:19:10 +0000102
Daniel Dunbarcb358b62009-07-28 03:00:54 +0000103 /// This is the current buffer index we're lexing from as managed by the
104 /// SourceMgr object.
Chris Lattnera59e8772009-06-21 07:19:10 +0000105 int CurBuffer;
106
Chris Lattnerfaf32c12009-06-24 00:33:19 +0000107 void operator=(const AsmLexer&); // DO NOT IMPLEMENT
108 AsmLexer(const AsmLexer&); // DO NOT IMPLEMENT
Chris Lattnera59e8772009-06-21 07:19:10 +0000109public:
110 AsmLexer(SourceMgr &SrcMgr);
Chris Lattnerfaf32c12009-06-24 00:33:19 +0000111 ~AsmLexer();
Chris Lattnera59e8772009-06-21 07:19:10 +0000112
Daniel Dunbara3c924f2009-07-28 16:56:42 +0000113 const AsmToken &Lex() {
114 return CurTok = LexToken();
Chris Lattnera59e8772009-06-21 07:19:10 +0000115 }
116
Daniel Dunbar3f872332009-07-28 16:08:33 +0000117 AsmToken::TokenKind getKind() const { return CurTok.getKind(); }
118 bool is(AsmToken::TokenKind K) const { return CurTok.is(K); }
119 bool isNot(AsmToken::TokenKind K) const { return CurTok.isNot(K); }
Daniel Dunbar9a7e2cc2009-07-27 21:49:56 +0000120
Chris Lattnera59e8772009-06-21 07:19:10 +0000121 SMLoc getLoc() const;
Daniel Dunbar419aded2009-07-28 16:38:40 +0000122
123 /// getTok - Return a reference to the current (last) lexed token.
124 const AsmToken &getTok() const { return CurTok; }
Chris Lattnera59e8772009-06-21 07:19:10 +0000125
Chris Lattner8e25e2d2009-07-16 06:14:39 +0000126 /// EnterIncludeFile - Enter the specified file. This returns true on failure.
127 bool EnterIncludeFile(const std::string &Filename);
128
Daniel Dunbar3fb76832009-06-30 00:49:23 +0000129 void PrintMessage(SMLoc Loc, const std::string &Msg, const char *Type) const;
Chris Lattnera59e8772009-06-21 07:19:10 +0000130
131private:
132 int getNextChar();
Daniel Dunbarcb358b62009-07-28 03:00:54 +0000133 AsmToken ReturnError(const char *Loc, const std::string &Msg);
Chris Lattnera59e8772009-06-21 07:19:10 +0000134
135 /// LexToken - Read the next token and return its code.
Daniel Dunbarcb358b62009-07-28 03:00:54 +0000136 AsmToken LexToken();
137 AsmToken LexIdentifier();
138 AsmToken LexPercent();
139 AsmToken LexSlash();
140 AsmToken LexLineComment();
141 AsmToken LexDigit();
142 AsmToken LexQuote();
Chris Lattnera59e8772009-06-21 07:19:10 +0000143};
144
145} // end namespace llvm
146
147#endif