1 //===- LLLexer.cpp - Lexer for .ll Files ----------------------------------===//
5 // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
7 //===----------------------------------------------------------------------===//
11 //===----------------------------------------------------------------------===//
37 //===----------------------------------------------------------------------===//
39 //===----------------------------------------------------------------------===//
41 // atoull - Convert an ascii string of decimal digits into the unsigned long
43 // because we know that the input will be matched by a suitable regex...
50     Result += *Buffer-'0';  in atoull()
77   if (End - Buffer >= 16) {  in HexToIntPair()
93 /// FP80HexToIntPair - translate an 80 bit FP80 number (20 hexits) into
112 // UnEscapeLexed - Run through the specified buffer and change \xx codes to the
121       if (BIn < EndBuffer-1 && BIn[1] == '\\') {  in UnEscapeLexed()
124       } else if (BIn < EndBuffer-2 &&  in UnEscapeLexed()
137   Str.resize(BOut-Buffer);  in UnEscapeLexed()
140 /// isLabelChar - Return true for [-a-zA-Z$._0-9].
142   return isalnum(static_cast<unsigned char>(C)) || C == '-' || C == '$' ||  in isLabelChar()
146 /// isLabelTail - Return true if this pointer points to a valid end of a label.
155 //===----------------------------------------------------------------------===//
157 //===----------------------------------------------------------------------===//
170     // A nul character in the stream is either the end of the current buffer or  in getNextChar()
171     // a random nul in the file.  Disambiguate that here.  in getNextChar()
172     if (CurPtr-1 != CurBuf.end())  in getNextChar()
176     --CurPtr;  // Another call to lex will return EOF again.  in getNextChar()
188       // Handle letters: [a-zA-Z_]  in LexToken()
209         StrVal.assign(TokStart, CurPtr-1);  in LexToken()
228     case '-':  in LexToken()
255 ///   GlobalVar   @[-a-zA-Z$._][-a-zA-Z$._0-9]*
256 ///   GlobalVarID @[0-9]+
264     StrVal.assign(TokStart, CurPtr - 1);  in LexDollar()
280         StrVal.assign(TokStart + 2, CurPtr - 1);  in LexDollar()
291   // Handle ComdatVarName: $[-a-zA-Z$._][-a-zA-Z$._0-9]*  in LexDollar()
298 /// ReadString - Read a string until the closing quote.
309       StrVal.assign(Start, CurPtr-1);  in ReadString()
316 /// ReadVarName - Read the rest of a token containing a variable name.
320       CurPtr[0] == '-' || CurPtr[0] == '$' ||  in ReadVarName()
324            CurPtr[0] == '-' || CurPtr[0] == '$' ||  in ReadVarName()
334 // Lex an ID: [0-9]+. On success, the ID is stored in UIntVal and Token is
363         StrVal.assign(TokStart+2, CurPtr-1);  in LexVar()
374   // Handle VarName: [-a-zA-Z$._][-a-zA-Z$._0-9]*  in LexVar()
378   // Handle VarID: [0-9]+  in LexVar()
382 /// Lex all tokens that start with a % character.
384 ///   LocalVar   ::= %[-a-zA-Z$._][-a-zA-Z$._0-9]*
385 ///   LocalVarID ::= %[0-9]+
390 /// Lex all tokens that start with a " character.
411 /// Lex all tokens that start with a ! character.
415   // Lex a metadata name as a MetadataVar.  in LexExclaim()
417       CurPtr[0] == '-' || CurPtr[0] == '$' ||  in LexExclaim()
421            CurPtr[0] == '-' || CurPtr[0] == '$' ||  in LexExclaim()
432 /// Lex all tokens that start with a ^ character.
433 ///    SummaryID ::= ^[0-9]+
435   // Handle SummaryID: ^[0-9]+  in LexCaret()
439 /// Lex all tokens that start with a # character.
440 ///    AttrGrpID ::= #[0-9]+
443   // Handle AttrGrpID: #[0-9]+  in LexHash()
449 /// Lex a label, integer type, keyword, or hexadecimal integer constant.
450 ///    Label           [-a-zA-Z$._0-9]+:
451 ///    IntegerType     i[0-9]+
453 ///    HexIntConstant  [us]0x[0-9A-Fa-f]+
456   const char *IntEnd = CurPtr[-1] == 'i' ? nullptr : StartChar;  in LexIdentifier()
468   // If we stopped due to a colon, unless we were directed to ignore it,  in LexIdentifier()
469   // this really is a label.  in LexIdentifier()
471     StrVal.assign(StartChar-1, CurPtr++);  in LexIdentifier()
475   // Otherwise, this wasn't a label.  If this was valid as an integer type,  in LexIdentifier()
490   // Otherwise, this was a letter sequence.  See which keyword this is.  in LexIdentifier()
493   --StartChar;  in LexIdentifier()
494   StringRef Keyword(StartChar, CurPtr - StartChar);  in LexIdentifier()
514   KEYWORD(weak); // Use as a linkage, and a modifier for "cmpxchg".  in LexIdentifier()
674   // ninf already a keyword  in LexIdentifier()
679   // sub already a keyword  in LexIdentifier()
719   // Use-list order directives.  in LexIdentifier()
978   // Check for [us]0x[0-9A-Fa-f]+ which are Hexadecimal constant generated by  in LexIdentifier()
979   // the CFE to avoid forcing it to deal with 64-bit numbers.  in LexIdentifier()
983     int len = CurPtr-TokStart-3;  in LexIdentifier()
1010 /// Lex all tokens that start with a 0x prefix, knowing they match and are not
1012 ///    HexFPConstant     0x[0-9A-Fa-f]+
1013 ///    HexFP80Constant   0xK[0-9A-Fa-f]+
1014 ///    HexFP128Constant  0xL[0-9A-Fa-f]+
1015 ///    HexPPC128Constant 0xM[0-9A-Fa-f]+
1016 ///    HexHalfConstant   0xH[0-9A-Fa-f]+
1017 ///    HexBFloatConstant 0xR[0-9A-Fa-f]+
1039     // HexFPConstant - Floating point constant represented in IEEE format as a  in Lex0x()
1051     // F80HexFPConstant - x87 long double in hexadecimal format (10 bytes)  in Lex0x()
1056     // F128HexFPConstant - IEEE 128-bit in hexadecimal format (16 bytes)  in Lex0x()
1061     // PPC128HexFPConstant - PowerPC 128-bit in hexadecimal format (16 bytes)  in Lex0x()
1077 /// Lex tokens for a label or a numeric constant, possibly starting with -.
1078 ///    Label             [-a-zA-Z$._0-9]+:
1079 ///    NInteger          -[0-9]+
1080 ///    FPConstant        [-+]?[0-9]+[.][0-9]*([eE][-+]?[0-9]+)?
1081 ///    PInteger          [0-9]+
1082 ///    HexFPConstant     0x[0-9A-Fa-f]+
1083 ///    HexFP80Constant   0xK[0-9A-Fa-f]+
1084 ///    HexFP128Constant  0xL[0-9A-Fa-f]+
1085 ///    HexPPC128Constant 0xM[0-9A-Fa-f]+
1087   // If the letter after the negative is not a number, this is probably a label.  in LexDigitOrNegative()
1090     // Okay, this is not a number after the -, it's probably a label.  in LexDigitOrNegative()
1092       StrVal.assign(TokStart, End-1);  in LexDigitOrNegative()
1100   // At this point, it is either a label, int or fp constant.  in LexDigitOrNegative()
1106   // Check if this is a fully-numeric label:  in LexDigitOrNegative()
1116   // Check to see if this really is a string label, e.g. "-1:".  in LexDigitOrNegative()
1119       StrVal.assign(TokStart, End-1);  in LexDigitOrNegative()
1125   // If the next character is a '.', then it is a fp value, otherwise its  in LexDigitOrNegative()
1130     APSIntVal = APSInt(StringRef(TokStart, CurPtr - TokStart));  in LexDigitOrNegative()
1136   // Skip over [0-9]*([eE][-+]?[0-9]+)?  in LexDigitOrNegative()
1141         ((CurPtr[1] == '-' || CurPtr[1] == '+') &&  in LexDigitOrNegative()
1149                        StringRef(TokStart, CurPtr - TokStart));  in LexDigitOrNegative()
1153 /// Lex a floating point constant starting with +.
1154 ///    FPConstant  [-+]?[0-9]+[.][0-9]*([eE][-+]?[0-9]+)?
1156   // If the letter after the negative is a number, this is probably not a  in LexPositive()
1165   // At this point, we need a '.'.  in LexPositive()
1173   // Skip over [0-9]*([eE][-+]?[0-9]+)?  in LexPositive()
1178         ((CurPtr[1] == '-' || CurPtr[1] == '+') &&  in LexPositive()
1186                        StringRef(TokStart, CurPtr - TokStart));  in LexPositive()