github.com/johnnyeven/libtools@v0.0.0-20191126065708-61829c1adf46/third_party/mlir/lib/Parser/Token.cpp (about) 1 //===- Token.cpp - MLIR Token Implementation ------------------------------===// 2 // 3 // Copyright 2019 The MLIR Authors. 4 // 5 // Licensed under the Apache License, Version 2.0 (the "License"); 6 // you may not use this file except in compliance with the License. 7 // You may obtain a copy of the License at 8 // 9 // http://www.apache.org/licenses/LICENSE-2.0 10 // 11 // Unless required by applicable law or agreed to in writing, software 12 // distributed under the License is distributed on an "AS IS" BASIS, 13 // WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. 14 // See the License for the specific language governing permissions and 15 // limitations under the License. 16 // ============================================================================= 17 // 18 // This file implements the Token class for the MLIR textual form. 19 // 20 //===----------------------------------------------------------------------===// 21 22 #include "Token.h" 23 #include "llvm/ADT/StringExtras.h" 24 using namespace mlir; 25 using llvm::SMLoc; 26 using llvm::SMRange; 27 28 SMLoc Token::getLoc() const { return SMLoc::getFromPointer(spelling.data()); } 29 30 SMLoc Token::getEndLoc() const { 31 return SMLoc::getFromPointer(spelling.data() + spelling.size()); 32 } 33 34 SMRange Token::getLocRange() const { return SMRange(getLoc(), getEndLoc()); } 35 36 /// For an integer token, return its value as an unsigned. If it doesn't fit, 37 /// return None. 38 Optional<unsigned> Token::getUnsignedIntegerValue() const { 39 bool isHex = spelling.size() > 1 && spelling[1] == 'x'; 40 41 unsigned result = 0; 42 if (spelling.getAsInteger(isHex ? 0 : 10, result)) 43 return None; 44 return result; 45 } 46 47 /// For an integer token, return its value as a uint64_t. If it doesn't fit, 48 /// return None. 49 Optional<uint64_t> Token::getUInt64IntegerValue() const { 50 bool isHex = spelling.size() > 1 && spelling[1] == 'x'; 51 52 uint64_t result = 0; 53 if (spelling.getAsInteger(isHex ? 0 : 10, result)) 54 return None; 55 return result; 56 } 57 58 /// For a floatliteral, return its value as a double. Return None if the value 59 /// underflows or overflows. 60 Optional<double> Token::getFloatingPointValue() const { 61 double result = 0; 62 if (spelling.getAsDouble(result)) 63 return None; 64 return result; 65 } 66 67 /// For an inttype token, return its bitwidth. 68 Optional<unsigned> Token::getIntTypeBitwidth() const { 69 unsigned result = 0; 70 if (spelling[1] == '0' || spelling.drop_front().getAsInteger(10, result) || 71 result == 0) 72 return None; 73 return result; 74 } 75 76 /// Given a 'string' token, return its value, including removing the quote 77 /// characters and unescaping the contents of the string. The lexer has already 78 /// verified that this token is valid. 79 std::string Token::getStringValue() const { 80 assert(getKind() == string); 81 // Start by dropping the quotes. 82 StringRef bytes = getSpelling().drop_front().drop_back(); 83 84 std::string result; 85 result.reserve(bytes.size()); 86 for (unsigned i = 0, e = bytes.size(); i != e;) { 87 auto c = bytes[i++]; 88 if (c != '\\') { 89 result.push_back(c); 90 continue; 91 } 92 93 assert(i + 1 < e && "invalid string should be caught by lexer"); 94 auto c1 = bytes[i++]; 95 switch (c1) { 96 case '"': 97 case '\\': 98 result.push_back(c1); 99 continue; 100 case 'n': 101 result.push_back('\n'); 102 continue; 103 case 't': 104 result.push_back('\t'); 105 continue; 106 default: 107 break; 108 } 109 110 assert(i + 1 <= e && "invalid string should be caught by lexer"); 111 auto c2 = bytes[i++]; 112 113 assert(llvm::isHexDigit(c1) && llvm::isHexDigit(c2) && "invalid escape"); 114 result.push_back((llvm::hexDigitValue(c1) << 4) | llvm::hexDigitValue(c2)); 115 } 116 117 return result; 118 } 119 120 /// Given a hash_identifier token like #123, try to parse the number out of 121 /// the identifier, returning None if it is a named identifier like #x or 122 /// if the integer doesn't fit. 123 Optional<unsigned> Token::getHashIdentifierNumber() const { 124 assert(getKind() == hash_identifier); 125 unsigned result = 0; 126 if (spelling.drop_front().getAsInteger(10, result)) 127 return None; 128 return result; 129 } 130 131 /// Given a punctuation or keyword token kind, return the spelling of the 132 /// token as a string. Warning: This will abort on markers, identifiers and 133 /// literal tokens since they have no fixed spelling. 134 StringRef Token::getTokenSpelling(Kind kind) { 135 switch (kind) { 136 default: 137 llvm_unreachable("This token kind has no fixed spelling"); 138 #define TOK_PUNCTUATION(NAME, SPELLING) \ 139 case NAME: \ 140 return SPELLING; 141 #define TOK_OPERATOR(NAME, SPELLING) \ 142 case NAME: \ 143 return SPELLING; 144 #define TOK_KEYWORD(SPELLING) \ 145 case kw_##SPELLING: \ 146 return #SPELLING; 147 #include "TokenKinds.def" 148 } 149 } 150 151 /// Return true if this is one of the keyword token kinds (e.g. kw_if). 152 bool Token::isKeyword() const { 153 switch (kind) { 154 default: 155 return false; 156 #define TOK_KEYWORD(SPELLING) \ 157 case kw_##SPELLING: \ 158 return true; 159 #include "TokenKinds.def" 160 } 161 }