llvm/AsmParser/LLLexer.h

0001 //===- LLLexer.h - Lexer for LLVM Assembly Files ----------------*- C++ -*-===//
0002 //
0003 // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
0004 // See https://llvm.org/LICENSE.txt for license information.
0005 // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
0006 //
0007 //===----------------------------------------------------------------------===//
0008 //
0009 // This class represents the Lexer for .ll files.
0010 //
0011 //===----------------------------------------------------------------------===//
0012
0013 #ifndef LLVM_ASMPARSER_LLLEXER_H
0014 #define LLVM_ASMPARSER_LLLEXER_H
0015
0016 #include "LLToken.h"
0017 #include "llvm/ADT/APFloat.h"
0018 #include "llvm/ADT/APSInt.h"
0019 #include "llvm/Support/SMLoc.h"
0020 #include <string>
0021
0022 namespace llvm {
0023   class Type;
0024   class SMDiagnostic;
0025   class SourceMgr;
0026   class LLVMContext;
0027
0028   class LLLexer {
0029     const char *CurPtr;
0030     StringRef CurBuf;
0031
0032     enum class ErrorPriority {
0033       None,   // No error message present.
0034       Parser, // Errors issued by parser.
0035       Lexer,  // Errors issued by lexer.
0036     };
0037
0038     struct ErrorInfo {
0039       ErrorPriority Priority = ErrorPriority::None;
0040       SMDiagnostic &Error;
0041
0042       explicit ErrorInfo(SMDiagnostic &Error) : Error(Error) {}
0043     } ErrorInfo;
0044
0045     SourceMgr &SM;
0046     LLVMContext &Context;
0047
0048     // Information about the current token.
0049     const char *TokStart;
0050     lltok::Kind CurKind;
0051     std::string StrVal;
0052     unsigned UIntVal = 0;
0053     Type *TyVal = nullptr;
0054     APFloat APFloatVal{0.0};
0055     APSInt APSIntVal{0};
0056
0057     // When false (default), an identifier ending in ':' is a label token.
0058     // When true, the ':' is treated as a separate token.
0059     bool IgnoreColonInIdentifiers = false;
0060
0061   public:
0062     explicit LLLexer(StringRef StartBuf, SourceMgr &SM, SMDiagnostic &,
0063                      LLVMContext &C);
0064
0065     lltok::Kind Lex() {
0066       return CurKind = LexToken();
0067     }
0068
0069     typedef SMLoc LocTy;
0070     LocTy getLoc() const { return SMLoc::getFromPointer(TokStart); }
0071     lltok::Kind getKind() const { return CurKind; }
0072     const std::string &getStrVal() const { return StrVal; }
0073     Type *getTyVal() const { return TyVal; }
0074     unsigned getUIntVal() const { return UIntVal; }
0075     const APSInt &getAPSIntVal() const { return APSIntVal; }
0076     const APFloat &getAPFloatVal() const { return APFloatVal; }
0077
0078     void setIgnoreColonInIdentifiers(bool val) {
0079       IgnoreColonInIdentifiers = val;
0080     }
0081
0082     // This returns true as a convenience for the parser functions that return
0083     // true on error.
0084     bool ParseError(LocTy ErrorLoc, const Twine &Msg) {
0085       Error(ErrorLoc, Msg, ErrorPriority::Parser);
0086       return true;
0087     }
0088     bool ParseError(const Twine &Msg) { return ParseError(getLoc(), Msg); }
0089
0090     void Warning(LocTy WarningLoc, const Twine &Msg) const;
0091     void Warning(const Twine &Msg) const { return Warning(getLoc(), Msg); }
0092
0093   private:
0094     lltok::Kind LexToken();
0095
0096     int getNextChar();
0097     void SkipLineComment();
0098     bool SkipCComment();
0099     lltok::Kind ReadString(lltok::Kind kind);
0100     bool ReadVarName();
0101
0102     lltok::Kind LexIdentifier();
0103     lltok::Kind LexDigitOrNegative();
0104     lltok::Kind LexPositive();
0105     lltok::Kind LexAt();
0106     lltok::Kind LexDollar();
0107     lltok::Kind LexExclaim();
0108     lltok::Kind LexPercent();
0109     lltok::Kind LexUIntID(lltok::Kind Token);
0110     lltok::Kind LexVar(lltok::Kind Var, lltok::Kind VarID);
0111     lltok::Kind LexQuote();
0112     lltok::Kind Lex0x();
0113     lltok::Kind LexHash();
0114     lltok::Kind LexCaret();
0115
0116     uint64_t atoull(const char *Buffer, const char *End);
0117     uint64_t HexIntToVal(const char *Buffer, const char *End);
0118     void HexToIntPair(const char *Buffer, const char *End, uint64_t Pair[2]);
0119     void FP80HexToIntPair(const char *Buffer, const char *End,
0120                           uint64_t Pair[2]);
0121
0122     void Error(LocTy ErrorLoc, const Twine &Msg, ErrorPriority Origin);
0123
0124     void LexError(LocTy ErrorLoc, const Twine &Msg) {
0125       Error(ErrorLoc, Msg, ErrorPriority::Lexer);
0126     }
0127     void LexError(const Twine &Msg) { LexError(getLoc(), Msg); }
0128   };
0129 } // end namespace llvm
0130
0131 #endif