src/loader/Tokenizer.cpp

   1 #include "Tokenizer.h"
   2
   3 #include <istream>
   4
   5 namespace loader {
   6
   7 bool Tokenizer::HasMore() {
   8         if (isPutback) return true;
   9         ScanSpace();
  10         if (!in) return false;
  11
  12         putback = ReadToken();
  13         isPutback = true;
  14         if (!skipComments || putback.type != Token::COMMENT) return true;
  15
  16         while (in && putback.type == Token::COMMENT) {
  17                 putback = ReadToken();
  18                 ScanSpace();
  19         }
  20         return putback.type != Token::COMMENT;
  21 }
  22
  23 void Tokenizer::ScanSpace() {
  24         std::istream::char_type c;
  25         in.get(c);
  26         while (in && std::isspace(c)) {
  27                 if (c == '\n') {
  28                         ++line;
  29                 }
  30                 in.get(c);
  31         }
  32         if (in) {
  33                 in.putback(c);
  34         }
  35 }
  36
  37 void Tokenizer::Putback(const Token &t) {
  38         if (isPutback) {
  39                 throw LexerError(line, "Tokenizer: double putback not supported");
  40         } else {
  41                 putback = t;
  42                 isPutback = true;
  43         }
  44 }
  45
  46 const Tokenizer::Token &Tokenizer::Peek() {
  47         if (!isPutback) {
  48                 putback = GetNext();
  49                 isPutback = true;
  50         }
  51         return putback;
  52 }
  53
  54 Tokenizer::Token Tokenizer::GetNext() {
  55         if (!HasMore()) {
  56                 throw LexerError(line, "read beyond last token");
  57         }
  58         if (isPutback) {
  59                 isPutback = false;
  60                 return putback;
  61         } else {
  62                 return ReadToken();
  63         }
  64 }
  65
  66 Tokenizer::Token Tokenizer::ReadToken() {
  67         ScanSpace();
  68         std::istream::char_type c;
  69         in.get(c);
  70         switch (c) {
  71                 case Token::ANGLE_BRACKET_OPEN:
  72                 case Token::ANGLE_BRACKET_CLOSE:
  73                 case Token::COLON:
  74                 case Token::COMMA:
  75                 case Token::BRACKET_OPEN:
  76                 case Token::BRACKET_CLOSE:
  77                 case Token::PARENTHESIS_OPEN:
  78                 case Token::PARENTHESIS_CLOSE:
  79                 case Token::COMMAND:
  80                 case Token::REGISTER:
  81                         return Token ((Token::Type) c);
  82                 case '<': {
  83                         std::istream::char_type c2;
  84                         in.get(c2);
  85                         if (c2 == '<') {
  86                                 return Token(Token::SCRIPT_BEGIN);
  87                         } else {
  88                                 in.putback(c2);
  89                                 return Token(Token::CHEVRON_OPEN);
  90                         }
  91                 }
  92                 case '>': {
  93                         std::istream::char_type c2;
  94                         in.get(c2);
  95                         if (c2 == '>') {
  96                                 return Token(Token::SCRIPT_END);
  97                         } else {
  98                                 in.putback(c2);
  99                                 return Token(Token::CHEVRON_CLOSE);
 100                         }
 101                 }
 102                 case '+': case '-':
 103                 case '0': case '1': case '2': case '3': case '4':
 104                 case '5': case '6': case '7': case '8': case '9':
 105                         in.putback(c);
 106                         return ReadNumber();
 107                 case '"':
 108                         in.putback(c);
 109                         return ReadString();
 110                 case '/':
 111                         {
 112                                 std::istream::char_type c2;
 113                                 in.get(c2);
 114                                 if (c2 == '/') {
 115                                         return ReadComment();
 116                                 } else if (c2 == '*') {
 117                                         return ReadMultilineComment();
 118                                 } else {
 119                                         throw LexerError(line, std::string("Tokenizer: cannot parse token: ") + c + c2 + ": expected / or *");
 120                                 }
 121                         }
 122                         break;
 123                 default:
 124                         in.putback(c);
 125                         {
 126                                 Token t(ReadIdentifier());
 127                                 if (std::isupper(c)) {
 128                                         t.type = Token::TYPE_NAME;
 129                                 } else if (std::islower(c)) {
 130                                         CheckKeyword(t);
 131                                 } else {
 132                                         throw LexerError(line, std::string("Tokenizer: cannot parse token: ") + c);
 133                                 }
 134                                 return t;
 135                         }
 136         }
 137 }
 138
 139 Tokenizer::Token Tokenizer::ReadNumber() {
 140         Token t(Token::NUMBER);
 141         bool isNegative(false);
 142
 143         std::istream::char_type c;
 144         in.get(c);
 145         if (c == '-') {
 146                 isNegative = true;
 147         } else if (c != '+') {
 148                 in.putback(c);
 149         }
 150
 151         while (in.get(c)) {
 152                 if (!std::isdigit(c)) {
 153                         in.putback(c);
 154                         break;
 155                 }
 156                 t.number *= 10;
 157                 t.number += c - '0';
 158         }
 159
 160         if (isNegative) t.number *= -1;
 161
 162         return t;
 163 }
 164
 165 Tokenizer::Token Tokenizer::ReadString() {
 166         Token t(Token::STRING);
 167         bool escape(false);
 168
 169         std::istream::char_type c;
 170         in.get(c);
 171         if (c != '"') {
 172                 throw LexerError(line, "Tokenizer: strings must begin with '\"'");
 173         }
 174
 175         while (in.get(c)) {
 176                 if (escape) {
 177                         escape = false;
 178                         switch (c) {
 179                                 case 'n':
 180                                         t.str.push_back('\n');
 181                                         break;
 182                                 case 'r':
 183                                         t.str.push_back('\r');
 184                                         break;
 185                                 case 't':
 186                                         t.str.push_back('\t');
 187                                         break;
 188                                 default:
 189                                         t.str.push_back(c);
 190                                         break;
 191                         }
 192                 } else if (c == '"') {
 193                         break;
 194                 } else if (c == '\\') {
 195                         escape = true;
 196                 } else {
 197                         t.str.push_back(c);
 198                 }
 199         }
 200
 201         return t;
 202 }
 203
 204 Tokenizer::Token Tokenizer::ReadIdentifier() {
 205         Token t(Token::IDENTIFIER);
 206
 207         std::istream::char_type c;
 208         while (in.get(c)) {
 209                 if (std::isalnum(c) || c == '_') {
 210                         t.str.push_back(c);
 211                 } else {
 212                         in.putback(c);
 213                         break;
 214                 }
 215         }
 216
 217         return t;
 218 }
 219
 220 Tokenizer::Token Tokenizer::ReadComment() {
 221         std::istream::char_type c;
 222         while (in.get(c) && c != '\n');
 223         ++line;
 224         return Token(Token::COMMENT);
 225 }
 226
 227 Tokenizer::Token Tokenizer::ReadMultilineComment() {
 228         std::istream::char_type c;
 229         while (in.get(c)) {
 230                 if (c == '*') {
 231                         std::istream::char_type c2;
 232                         if (in.get(c2) && c2 == '/') {
 233                                 break;
 234                         }
 235                 } else if (c == '\n') {
 236                         ++line;
 237                 }
 238         }
 239         return Token(Token::COMMENT);
 240 }
 241
 242 bool Tokenizer::CheckKeyword(Token &t) {
 243         if (t.str == "export") {
 244                 t.type = Token::KEYWORD_EXPORT;
 245                 return true;
 246         } else if (t.str == "false") {
 247                 t.type = Token::KEYWORD_FALSE;
 248                 return true;
 249         } else if (t.str == "include") {
 250                 t.type = Token::KEYWORD_INCLUDE;
 251                 return true;
 252         } else if (t.str == "true") {
 253                 t.type = Token::KEYWORD_TRUE;
 254                 return true;
 255         } else {
 256                 return false;
 257         }
 258 }
 259
 260 }