Scanner.lpp.in

   1 /* Cycript - The Truly Universal Scripting Language
   2  * Copyright (C) 2009-2016  Jay Freeman (saurik)
   3 */
   4
   5 /* GNU Affero General Public License, Version 3 {{{ */
   6 /*
   7  * This program is free software: you can redistribute it and/or modify
   8  * it under the terms of the GNU Affero General Public License as published by
   9  * the Free Software Foundation, either version 3 of the License, or
  10  * (at your option) any later version.
  11
  12  * This program is distributed in the hope that it will be useful,
  13  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  14  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  15  * GNU Affero General Public License for more details.
  16
  17  * You should have received a copy of the GNU Affero General Public License
  18  * along with this program.  If not, see <http://www.gnu.org/licenses/>.
  19 **/
  20 /* }}} */
  21
  22 %top{
  23 #if defined(__clang__)
  24 #pragma clang diagnostic push
  25 #pragma clang diagnostic ignored "-Wdeprecated-register"
  26 #pragma clang diagnostic ignored "-Wunused-function"
  27 #pragma clang diagnostic ignored "-Wunused-variable"
  28 #else
  29 #pragma GCC diagnostic push
  30 #pragma GCC diagnostic ignored "-Wsign-compare"
  31 #pragma GCC diagnostic ignored "-Wunused-function"
  32 #pragma GCC diagnostic ignored "-Wunused-variable"
  33 #endif
  34 }
  35
  36 %{
  37
  38 #define YYLTYPE CYLocation
  39 #include "Parser.hpp"
  40 typedef cy::parser::token tk;
  41
  42 #include "Highlight.hpp"
  43
  44 #include "IdentifierStart.h"
  45 #include "IdentifierContinue.h"
  46
  47 #define YY_EXTRA_TYPE CYDriver *
  48
  49 #define F(value, highlight) do { \
  50     BEGIN(yyextra->template_.top() ? DivOrTemplateTail : Div); \
  51     yylval->highlight_ = highlight; \
  52     return value; \
  53 } while (false)
  54
  55 #define S(stack) do { \
  56     if (yyextra->stack.size() == 1) \
  57         E("invalid brace") \
  58     yyextra->stack.pop(); \
  59 } while (false)
  60
  61 #define P yyextra->pool_
  62 #define A new(P)
  63 #define Y P.strmemdup(yytext, yyleng)
  64
  65 #define I(type, Type, value, highlight) do { \
  66     yylval->semantic_.type ## _ = A CY ## Type; \
  67     yylval->semantic_.type ## _->location_ = *yylloc; \
  68     F(value, highlight); \
  69 } while (false)
  70
  71 #define N \
  72     if (yyextra->last_) { \
  73         yyextra->last_ = false; \
  74         F(tk::NewLine, hi::Nothing); \
  75     }
  76
  77 #define V(more) { \
  78     if (const char *nl = reinterpret_cast<const char *>(memchr(yytext, '\n', yyleng))) { \
  79         unsigned lines(0); \
  80         size_t left; \
  81         do { \
  82             ++lines; \
  83             left = yyleng - (nl - yytext) - 1; \
  84             nl = reinterpret_cast<const char *>(memchr(nl + 1, '\n', left)); \
  85         } while (nl != NULL); \
  86         yylloc->step(); \
  87         yylloc->end.Lines(lines); \
  88         yylloc->end.Columns(left); \
  89         more \
  90     } else L \
  91 }
  92
  93 #define R yylloc->end.Columns(yyleng);
  94 #define L yylloc->step(); R
  95
  96 #define H(value, highlight) do { \
  97     if (yyextra->highlight_) \
  98         F(value, highlight); \
  99 } while (false)
 100
 101 #define M \
 102     H(tk::Comment, hi::Comment);
 103
 104 #define E(message) { \
 105     CYDriver::Error error; \
 106     error.location_ = *yylloc; \
 107     error.message_ = "syntax error, " message; \
 108     yyextra->errors_.push_back(error); \
 109     yyterminate(); \
 110 }
 111
 112 int X(char c) {
 113     if (c >= '0' && c <= '9')
 114         return c - '0';
 115     if (c >= 'a' && c <= 'f')
 116         return c - 'a' + 10;
 117     if (c >= 'A' && c <= 'F')
 118         return c - 'A' + 10;
 119     return -1;
 120 }
 121
 122 template <typename Type_>
 123 static void U(Type_ &local, unsigned point) {
 124     if (false) {
 125     } else if (point < 0x000080) {
 126         *local++ = point;
 127     } else if (point < 0x000800) {
 128         *local++ = 0xc0 | point >> 0x06 & 0x1f;
 129         goto one;
 130     } else if (point < 0x010000) {
 131         *local++ = 0xe0 | point >> 0x0c & 0x0f;
 132         goto two;
 133     } else if (point < 0x110000) {
 134         *local++ = 0xf0 | point >> 0x12 & 0x07;
 135         *local++ = 0x80 | point >> 0x0c & 0x3f;
 136       two:
 137         *local++ = 0x80 | point >> 0x06 & 0x3f;
 138       one:
 139         *local++ = 0x80 | point >> 0x00 & 0x3f;
 140     } else _assert(false);
 141 }
 142
 143 static unsigned U(char *&local, const char *text, yy_size_t &i) {
 144     unsigned point;
 145
 146     char next(text[++i]);
 147     if (next != '{') {
 148         point = X(text[i + 0]) << 12 | X(text[i + 1]) << 8 | X(text[i + 2]) << 4 | X(text[i + 3]);
 149         i += 3;
 150     } else {
 151         point = 0;
 152         for (;;) {
 153             next = text[++i];
 154             if (next == '}')
 155                 break;
 156             point = (point << 4) | X(next);
 157         }
 158     }
 159
 160     U(local, point);
 161     return point;
 162 }
 163
 164 #define CYLexBufferPoint(point) do { \
 165     std::back_insert_iterator<std::vector<char> > inserter(yyextra->buffer_); \
 166     U(inserter, point); \
 167 } while (false)
 168
 169 #define CYLexBufferUnit(value) do { \
 170     yyextra->buffer_.push_back(value); \
 171 } while (false)
 172
 173 #define CYLexBufferUnits(data, size) do { \
 174     yyextra->buffer_.insert(yyextra->buffer_.end(), data, data + size); \
 175 } while (false)
 176
 177 #define CYLexBufferStart(condition) do { \
 178     yyextra->buffer_.clear(); \
 179     yy_push_state(condition, yyscanner); \
 180 } while (false)
 181
 182 #define CYLexBufferEnd(type, Type, value, highlight) do { \
 183     yy_pop_state(yyscanner); \
 184     I(type, Type(P.strmemdup(yyextra->buffer_.data(), yyextra->buffer_.size()), yyextra->buffer_.size()), value, highlight); \
 185 } while (false)
 186
 187 #define YY_INPUT(data, value, size) do { \
 188     value = yyextra->data_.sgetn(data, size) ?: YY_NULL; \
 189 } while (false)
 190
 191 %}
 192
 193 %option prefix="cy"
 194 %option bison-bridge
 195 %option bison-locations
 196 %option nodefault
 197 %option noyywrap
 198 %option noyylineno
 199 %option nounput
 200 %option nounistd
 201 %option 8bit
 202 %option backup
 203 %option batch
 204 %option never-interactive
 205 %option pointer
 206 %option reentrant
 207 %option stack
 208
 209 U1 [\x00-\x7f]
 210 U0 [\x80-\xbf]
 211 U2 [\xc2-\xdf]
 212 U3 [\xe0-\xef]
 213 U4 [\xf0-\xf4]
 214 UN [\xc0-\xc1\xf5-\xff]
 215 UE {U1}|{U2}|{U3}|{U4}|{UN}
 216
 217 HexDigit [0-9a-fA-F]
 218 LineTerminatorSequence \r?\n|\r|\xe2\x80[\xa8\xa9]
 219 WhiteSpace [\x09\x0b\x0c\x20]|\xc2\xa0|\xef\xbb\xbf
 220 UnicodeEscape \\u({HexDigit}{4}|\{{HexDigit}+\})
 221
 222 @include NotLineTerminator.l
 223 NoneTerminatorCharacter [^\r\n\x80-\xff]|{NotLineTerminator}
 224 RegExCharacter [^/[\\]{-}[\r\n\x80-\xff]|{NotLineTerminator}
 225 RegClsCharacter [^]\\]{-}[\r\n\x80-\xff]|{NotLineTerminator}
 226 CommentCharacter [^*/]{-}[\r\n\x80-\xff]|{NotLineTerminator}
 227 SingleCharacter [^'\\]{-}[\r\n\x80-\xff]|{NotLineTerminator}
 228 DoubleCharacter [^"\\]{-}[\r\n\x80-\xff]|{NotLineTerminator}
 229 PlateCharacter [^$`\\]{-}[\r\n\x80-\xff]|{NotLineTerminator}
 230
 231 @include UnicodeIDStart.l
 232 @include UnicodeIDContinue.l
 233 IdentifierMore [$_]
 234
 235 UnicodeStart {IdentifierMore}|{UnicodeIDStart}
 236 UnicodePart {IdentifierMore}|\xe2\x80[\x8c\x8d]|{UnicodeIDContinue}
 237 UnicodeScrap {U2}|{U3}{U0}{0,1}|{U4}{U0}{0,2}|{UN}|{U0}
 238 UnicodeError ({U2}|{U3}{U0}{0,1}|{U4}{U0}{0,2}){UE}|{UN}|{U0}
 239
 240 IdentifierStart {UnicodeStart}|{UnicodeEscape}
 241 IdentifierPart {UnicodePart}|{UnicodeEscape}
 242 IdentifierFail {UnicodeError}|\\(u({HexDigit}{0,3}|\{{HexDigit}*))?
 243 IdentifierScrap {IdentifierPart}*{IdentifierFail}?
 244
 245 RegularExpressionBackslashSequence \\{NoneTerminatorCharacter}
 246 RegularExpressionClassChars ({RegClsCharacter}|{RegularExpressionBackslashSequence})*
 247
 248 @begin E4X
 249 XMLNameStart [a-zA-Z_:]
 250 XMLNamePart [a-zA-Z0-9.-_:]
 251 XMLName {XMLNameStart}{XMLNamePart}*
 252 @end
 253
 254 %x RegularExpression
 255 %x MultiLine
 256
 257 %x LegacySingleString
 258 %x LegacyDoubleString
 259
 260 %x StrictSingleString
 261 %x StrictDoubleString
 262 %x StrictAccentString
 263
 264 %s Div
 265 %s DivOrTemplateTail
 266
 267 @begin E4X
 268 %x XMLContent
 269 %x XMLTag
 270 @end
 271
 272 %%
 273
 274     /* RegEx {{{ */
 275 <RegularExpression>{
 276     \/{UnicodePart}* R CYLexBufferUnits(yytext, yyleng); CYLexBufferEnd(literal, RegEx, tk::RegularExpressionLiteral_, hi::Constant);
 277     \/{UnicodePart}*{UnicodeError} R E("invalid character");
 278
 279     {RegExCharacter}+ R CYLexBufferUnits(yytext, yyleng);
 280
 281     {RegularExpressionBackslashSequence} R CYLexBufferUnits(yytext, yyleng);
 282     \\ R E("invalid escape")
 283
 284     (\\|{RegExCharacter}+)?{LineTerminatorSequence} R E("invalid newline");
 285     (\\|{RegExCharacter}+)?{UnicodeScrap} R E("invalid character");
 286
 287     "["{RegularExpressionClassChars}"]" R CYLexBufferUnits(yytext, yyleng);
 288     "["{RegularExpressionClassChars}\\? R E("invalid class");
 289
 290     "["{RegularExpressionClassChars}\\?{LineTerminatorSequence} R E("invalid newline");
 291     "["{RegularExpressionClassChars}\\?{UnicodeScrap} R E("invalid character");
 292
 293     <<EOF>> R E("unterminated regex")
 294 }
 295     /* }}} */
 296     /* Comment {{{ */
 297     /* XXX: maybe fold LineTerminatorSequence into these definitions */
 298 #!{NoneTerminatorCharacter}* L M
 299 \/\/{NoneTerminatorCharacter}* L M
 300 (#!|\/\/){NoneTerminatorCharacter}*{UnicodeError} L E("invalid character");
 301
 302 \/\* L yy_push_state(MultiLine, yyscanner);
 303
 304 <MultiLine>{
 305     \**\*\/ R yy_pop_state(yyscanner); M N
 306     \**{LineTerminatorSequence} yylloc->end.Lines(); yyextra->last_ = true;
 307     \**{CommentCharacter}|\/ R
 308
 309     \**{UnicodeScrap} R E("invalid character");
 310     \**\* R E("invalid comment");
 311
 312     <<EOF>> R E("invalid comment")
 313 }
 314     /* }}} */
 315     /* Element {{{ */
 316 @begin E4X
 317 <RegExp>"<>"      L F(tk::LeftRight, hi::Structure);
 318 <XMLContent>"</>" L F(tk::LeftSlashRight, hi::Structure);
 319
 320 <RegExp,XMLContent>\<!\[CDATA\[(\n|[^[]|\[[^[]|\[\[[^>])*]]> V() F(tk::XMLCDATA, hi::Constant);
 321 <RegExp,XMLContent>\<!--(\n|[^-]|-[^-])*--> V() F(tk::XMLComment, hi::Comment);
 322 <RegExp,XMLContent>\<?(\n|[^?]|\?[^>])*?> V() F(tk::XMLPI, hi::Meta);
 323
 324 <XMLTag>"="  L F(tk::Equal, hi::Structure);
 325 <XMLTag>">"  L F(tk::Right, hi::Structure);
 326 <XMLTag>"/>" L F(tk::SlashRight, hi::Structure);
 327 <XMLTag>"{"  L F(tk::OpenBrace, hi::Structure);
 328
 329 <XMLTag>\"(\n|[^"])*\"|'(\n|[^'])*' V() F(tk::XMLAttributeValue, hi::Constant);
 330 <XMLTag>{XMLName} L F(tk::XMLName, hi::Identifier);
 331 <XMLTag>[ \t\r\n] V() F(tk::XMLWhitespace, hi::Nothing);
 332
 333 <XMLContent>"{"  L F(tk::OpenBrace, hi::Structure);
 334 <XMLContent>"<"  L F(tk::Left, hi::Structure);
 335 <XMLContent>"</" L F(tk::LeftSlash, hi::Structure);
 336 @end
 337     /* }}} */
 338     /* Operator {{{ */
 339 "..."  L F(tk::PeriodPeriodPeriod, hi::Meta);
 340 ".."   L E("invalid operator")
 341
 342 @begin E4X
 343 ".."   L F(tk::PeriodPeriod, hi::Operator);
 344 @end
 345
 346 @begin E4X ObjectiveC
 347 "@"    L F(tk::At, hi::Operator);
 348 @end
 349
 350 "&"    L F(tk::Ampersand, hi::Operator);
 351 "&&"   L F(tk::AmpersandAmpersand, hi::Operator);
 352 "&="   L F(tk::AmpersandEqual, hi::Operator);
 353 "^"    L F(tk::Carrot, hi::Operator);
 354 "^="   L F(tk::CarrotEqual, hi::Operator);
 355 "="    L F(tk::Equal, hi::Operator);
 356 "=="   L F(tk::EqualEqual, hi::Operator);
 357 "==="  L F(tk::EqualEqualEqual, hi::Operator);
 358 "=>"   L F(tk::EqualRight, hi::Operator);
 359 "!"    L F(tk::Exclamation, hi::Operator);
 360 "!="   L F(tk::ExclamationEqual, hi::Operator);
 361 "!=="  L F(tk::ExclamationEqualEqual, hi::Operator);
 362 "-"    L F(tk::Hyphen, hi::Operator);
 363 "-="   L F(tk::HyphenEqual, hi::Operator);
 364 "--"   L F(tk::HyphenHyphen, hi::Operator);
 365 "->"   L F(tk::HyphenRight, hi::Operator);
 366 "<"    L F(tk::Left, hi::Operator);
 367 "<="   L F(tk::LeftEqual, hi::Operator);
 368 "<<"   L F(tk::LeftLeft, hi::Operator);
 369 "<<="  L F(tk::LeftLeftEqual, hi::Operator);
 370 "%"    L F(tk::Percent, hi::Operator);
 371 "%="   L F(tk::PercentEqual, hi::Operator);
 372 "."    L F(tk::Period, hi::Operator);
 373 "|"    L F(tk::Pipe, hi::Operator);
 374 "|="   L F(tk::PipeEqual, hi::Operator);
 375 "||"   L F(tk::PipePipe, hi::Operator);
 376 "+"    L F(tk::Plus, hi::Operator);
 377 "+="   L F(tk::PlusEqual, hi::Operator);
 378 "++"   L F(tk::PlusPlus, hi::Operator);
 379 ">"    L F(tk::Right, hi::Operator);
 380 ">="   L F(tk::RightEqual, hi::Operator);
 381 ">>"   L F(tk::RightRight, hi::Operator);
 382 ">>="  L F(tk::RightRightEqual, hi::Operator);
 383 ">>>"  L F(tk::RightRightRight, hi::Operator);
 384 ">>>=" L F(tk::RightRightRightEqual, hi::Operator);
 385 "*"    L F(tk::Star, hi::Operator);
 386 "*="   L F(tk::StarEqual, hi::Operator);
 387 "~"    L F(tk::Tilde, hi::Operator);
 388
 389 "/"  L F(tk::Slash, hi::Operator);
 390 "/=" L F(tk::SlashEqual, hi::Operator);
 391
 392 ":"    L F(tk::Colon, hi::Structure);
 393 "::"   L F(tk::ColonColon, hi::Structure);
 394 ","    L F(tk::Comma, hi::Structure);
 395 "?"    L F(tk::Question, hi::Structure);
 396 ";"    L F(tk::SemiColon, hi::Structure);
 397 "#"    L F(tk::Pound, hi::Operator);
 398
 399 "("    L F(tk::OpenParen, hi::Structure);
 400 ")"    L F(tk::CloseParen, hi::Structure);
 401
 402 "{"    L yyextra->template_.push(false); F(tk::OpenBrace, hi::Structure);
 403 <Div>"}" L S(template_); F(tk::CloseBrace, hi::Structure);
 404
 405 "["    L F(tk::OpenBracket, hi::Structure);
 406 "]"    L F(tk::CloseBracket, hi::Structure);
 407     /* }}} */
 408     /* Keyword {{{ */
 409 "@error"          L F(tk::At_error_, hi::Error);
 410
 411 @begin Java
 412 "@class"          L F(tk::At_class_, hi::Meta);
 413 @end
 414
 415 @begin C
 416 "@encode"         L F(tk::At_encode_, hi::Meta);
 417 @end
 418
 419 @begin ObjectiveC
 420 "@end"            L F(tk::At_end_, hi::Meta);
 421 "@false"          L F(tk::At_false_, hi::Constant);
 422 "@implementation" L F(tk::At_implementation_, hi::Meta);
 423 "@import"         L F(tk::At_import_, hi::Special);
 424 "@NO"             L F(tk::At_NO_, hi::Constant);
 425 "@null"           L F(tk::At_null_, hi::Constant);
 426 "@selector"       L F(tk::At_selector_, hi::Meta);
 427 "@true"           L F(tk::At_true_, hi::Constant);
 428 "@YES"            L F(tk::At_YES_, hi::Constant);
 429 @end
 430
 431 @({UnicodeStart}{UnicodePart}*{UnicodeError}?|{UnicodeError}) L E("invalid keyword")
 432     /* }}} */
 433     /* Highlight {{{ */
 434 "undefined"       L F(tk::_undefined_, hi::Operator);
 435
 436 @begin ObjectiveC
 437 "bool"            L F(tk::_bool_, hi::Type);
 438 "BOOL"            L F(tk::_BOOL_, hi::Type);
 439 "id"              L F(tk::_id_, hi::Type);
 440 "nil"             L F(tk::_nil_, hi::Constant);
 441 "NULL"            L F(tk::_NULL_, hi::Constant);
 442 "SEL"             L F(tk::_SEL_, hi::Type);
 443 @end
 444     /* }}} */
 445     /* Reserved {{{ */
 446 "abstract"        L /*FII*/ F(tk::_abstract_, hi::Meta);
 447 "as"              L /*III*/ F(tk::_as_, hi::Meta);
 448 "await"           L /*II?*/ F(tk::_await_, hi::Meta);
 449 "boolean"         L /*FII*/ F(tk::_boolean_, hi::Type);
 450 "break"           L /*KKK*/ F(tk::_break_, hi::Control);
 451 "byte"            L /*FII*/ F(tk::_byte_, hi::Type);
 452 "case"            L /*KKK*/ F(tk::_case_, hi::Control);
 453 "catch"           L /*KKK*/ F(tk::_catch_, hi::Control);
 454 "char"            L /*FII*/ F(tk::_char_, hi::Type);
 455 "class"           L /*FFK*/ F(tk::_class_, hi::Meta);
 456 "const"           L /*FFK*/ F(tk::_const_, hi::Meta);
 457 "constructor"     L /*III*/ F(tk::_constructor_, hi::Special);
 458 "continue"        L /*KKK*/ F(tk::_continue_, hi::Control);
 459 "debugger"        L /*FKK*/ F(tk::_debugger_, hi::Meta);
 460 "default"         L /*KKK*/ F(tk::_default_, hi::Control);
 461 "delete"          L /*KKK*/ F(tk::_delete_, hi::Operator);
 462 "do"              L /*KKK*/ F(tk::_do_, hi::Control);
 463 "double"          L /*FII*/ F(tk::_double_, hi::Type);
 464 "else"            L /*KKK*/ F(tk::_else_, hi::Control);
 465 "enum"            L /*FFF*/ F(tk::_enum_, hi::Meta);
 466 "export"          L /*FFK*/ F(tk::_export_, hi::Meta);
 467 "extends"         L /*FFK*/ F(tk::_extends_, hi::Meta);
 468 "eval"            L /*III*/ F(tk::_eval_, hi::Special);
 469 "false"           L /*LLL*/ F(tk::_false_, hi::Constant);
 470 "final"           L /*FII*/ F(tk::_final_, hi::Meta);
 471 "finally"         L /*KKK*/ F(tk::_finally_, hi::Control);
 472 "float"           L /*FII*/ F(tk::_float_, hi::Type);
 473 "for"             L /*KKK*/ F(tk::_for_, hi::Control);
 474 "from"            L /*III*/ F(tk::_from_, hi::Meta);
 475 "function"        L /*KKK*/ F(tk::_function_, hi::Meta);
 476 "goto"            L /*FII*/ F(tk::_goto_, hi::Control);
 477 "get"             L /*III*/ F(tk::_get_, hi::Meta);
 478 "if"              L /*KKK*/ F(tk::_if_, hi::Control);
 479 "implements"      L /*FSS*/ F(tk::_implements_, hi::Meta);
 480 "import"          L /*FFK*/ F(tk::_import_, hi::Meta);
 481 "in"              L /*KKK*/ F(tk::_in_, hi::Operator);
 482 "Infinity"        L /*III*/ F(tk::_Infinity_, hi::Constant);
 483 "instanceof"      L /*KKK*/ F(tk::_instanceof_, hi::Operator);
 484 "int"             L /*FII*/ F(tk::_int_, hi::Type);
 485 "__int128"        L /*III*/ F(tk::___int128_, hi::Type);
 486 "interface"       L /*FSS*/ F(tk::_interface_, hi::Meta);
 487 "let"             L /*IS?*/ F(tk::_let_, hi::Meta);
 488 "long"            L /*FII*/ F(tk::_long_, hi::Type);
 489 "native"          L /*FII*/ F(tk::_native_, hi::Meta);
 490 "new"             L /*KKK*/ F(tk::_new_, hi::Operator);
 491 "null"            L /*LLL*/ F(tk::_null_, hi::Constant);
 492 "package"         L /*FSS*/ F(tk::_package_, hi::Meta);
 493 "private"         L /*FSS*/ F(tk::_private_, hi::Meta);
 494 "protected"       L /*FSS*/ F(tk::_protected_, hi::Meta);
 495 "__proto__"       L /*III*/ F(tk::___proto___, hi::Special);
 496 "prototype"       L /*III*/ F(tk::_prototype_, hi::Special);
 497 "public"          L /*FSS*/ F(tk::_public_, hi::Meta);
 498 "__restrict"      L /*III*/ F(tk::___restrict_, hi::Meta);
 499 "restrict"        L /*III*/ F(tk::_restrict_, hi::Meta);
 500 "return"          L /*KKK*/ F(tk::_return_, hi::Control);
 501 "set"             L /*III*/ F(tk::_set_, hi::Meta);
 502 "short"           L /*FII*/ F(tk::_short_, hi::Type);
 503 "static"          L /*FS?*/ F(tk::_static_, hi::Meta);
 504 "super"           L /*FFK*/ F(tk::_super_, hi::Constant);
 505 "switch"          L /*KKK*/ F(tk::_switch_, hi::Control);
 506 "synchronized"    L /*FII*/ F(tk::_synchronized_, hi::Meta);
 507 "target"          L /*III*/ F(tk::_target_, hi::Identifier);
 508 "this"            L /*KKK*/ F(tk::_this_, hi::Constant);
 509 "throw"           L /*KKK*/ F(tk::_throw_, hi::Control);
 510 "throws"          L /*FII*/ F(tk::_throws_, hi::Meta);
 511 "transient"       L /*FII*/ F(tk::_transient_, hi::Meta);
 512 "true"            L /*LLL*/ F(tk::_true_, hi::Constant);
 513 "try"             L /*KKK*/ F(tk::_try_, hi::Control);
 514 "typeid"          L /*III*/ F(tk::_typeid_, hi::Operator);
 515 "typeof"          L /*KKK*/ F(tk::_typeof_, hi::Operator);
 516 "var"             L /*KKK*/ F(tk::_var_, hi::Meta);
 517 "void"            L /*KKK*/ F(tk::_void_, hi::Operator);
 518 "volatile"        L /*FII*/ F(tk::_volatile_, hi::Meta);
 519 "while"           L /*KKK*/ F(tk::_while_, hi::Control);
 520 "with"            L /*KKK*/ F(tk::_with_, hi::Control);
 521 "yield"           L /*IS?*/ F(tk::_yield_, hi::Control);
 522
 523 "each"            L F(tk::_each_, hi::Control);
 524 "of"              L F(tk::_of_, hi::Operator);
 525
 526 @begin C
 527 "extern"          L F(tk::_extern_, hi::Type);
 528 "signed"          L F(tk::_signed_, hi::Type);
 529 "struct"          L F(tk::_struct_, hi::Meta);
 530 "typedef"         L F(tk::_typedef_, hi::Meta);
 531 "unsigned"        L F(tk::_unsigned_, hi::Type);
 532 @end
 533
 534 @begin ObjectiveC
 535 "NO"              L F(tk::_NO_, hi::Constant);
 536 "YES"             L F(tk::_YES_, hi::Constant);
 537 @end
 538
 539 @begin E4X
 540 "namespace"       L F(tk::_namespace_, hi::Meta);
 541 "xml"             L F(tk::_xml_, hi::Meta);
 542 @end
 543     /* }}} */
 544     /* Identifier {{{ */
 545 {UnicodeStart}{UnicodePart}* L I(identifier, Identifier(Y), tk::Identifier_, hi::Identifier);
 546
 547 {IdentifierStart}{IdentifierPart}* L {
 548     char *value(A char[yyleng + 1]);
 549     char *local(value);
 550
 551     for (yy_size_t i(0), e(yyleng); i != e; ++i) {
 552         char next(yytext[i]);
 553         if (next != '\\')
 554             *local++ = next;
 555         else {
 556             bool (*is)(unsigned) = (i == 0 ? &IsIdentifierStart : &IsIdentifierContinue);
 557             unsigned point(U(local, yytext, ++i));
 558             if (!is(point))
 559                 E("invalid character");
 560         }
 561     }
 562
 563     *local = '\0';
 564     I(identifier, Identifier(value), tk::Identifier_, hi::Identifier);
 565 }
 566
 567 ({IdentifierStart}{IdentifierPart}*)?{IdentifierFail} L E("invalid identifier")
 568     /* }}} */
 569     /* Number {{{ */
 570 0[0-7]+ L I(number, Number(strtoull(yytext + 1, NULL, 8)), tk::NumericLiteral, hi::Constant);
 571 0[0-9]+ L I(number, Number(strtoull(yytext + 1, NULL, 10)), tk::NumericLiteral, hi::Constant);
 572
 573 0[xX][0-9a-fA-F]+ L I(number, Number(strtoull(yytext + 2, NULL, 16)), tk::NumericLiteral, hi::Constant);
 574 0[oO][0-7]+ L I(number, Number(strtoull(yytext + 2, NULL, 8)), tk::NumericLiteral, hi::Constant);
 575 0[bB][0-1]+ L I(number, Number(strtoull(yytext + 2, NULL, 2)), tk::NumericLiteral, hi::Constant);
 576
 577 (\.[0-9]+|(0|[1-9][0-9]*)(\.[0-9]*)?)([eE][+-]?[0-9]+)? L I(number, Number(strtod(yytext, NULL)), tk::NumericLiteral, hi::Constant);
 578 (\.[0-9]+|(0|[1-9][0-9]*)(\.[0-9]*)?)[eE][+-]?{IdentifierScrap} L E("invalid exponent")
 579 (\.?[0-9]|(0|[1-9][0-9]*)\.){IdentifierScrap} L E("invalid number")
 580     /* }}} */
 581     /* String {{{ */
 582 \' L CYLexBufferStart(LegacySingleString);
 583 <LegacySingleString,StrictSingleString>{
 584     \' R CYLexBufferEnd(string, String, tk::StringLiteral, hi::Constant);
 585     {SingleCharacter}+ R CYLexBufferUnits(yytext, yyleng);
 586     {SingleCharacter}*{LineTerminatorSequence} R E("invalid newline");
 587     {SingleCharacter}*{UnicodeScrap} R E("invalid character");
 588 }
 589
 590 \" L CYLexBufferStart(LegacyDoubleString);
 591 <LegacyDoubleString,StrictDoubleString>{
 592     \" R CYLexBufferEnd(string, String, tk::StringLiteral, hi::Constant);
 593     {DoubleCharacter}+ R CYLexBufferUnits(yytext, yyleng);
 594     {DoubleCharacter}*{LineTerminatorSequence} R E("invalid newline");
 595     {DoubleCharacter}*{UnicodeScrap} R E("invalid character");
 596 }
 597     /* }}} */
 598     /* Template {{{ */
 599 "`" L yyextra->tail_ = false; CYLexBufferStart(StrictAccentString);
 600 <DivOrTemplateTail>"}" L yyextra->tail_ = true; S(template_); CYLexBufferStart(StrictAccentString);
 601
 602 <StrictAccentString>{
 603     "`" R CYLexBufferEnd(string, String, yyextra->tail_ ? tk::TemplateTail : tk::NoSubstitutionTemplate, hi::Constant);
 604     "${" R yyextra->template_.push(true); CYLexBufferEnd(string, String, yyextra->tail_ ? tk::TemplateMiddle : tk::TemplateHead, hi::Constant);
 605
 606     "$" R CYLexBufferUnit('$');
 607
 608     {PlateCharacter}+ R CYLexBufferUnits(yytext, yyleng);
 609     {PlateCharacter}*{UnicodeScrap} R E("invalid character");
 610
 611     {PlateCharacter}*{LineTerminatorSequence} yylloc->end.Lines(); CYLexBufferUnits(yytext, yyleng);
 612     \\{LineTerminatorSequence} yylloc->end.Lines();
 613 }
 614     /* }}} */
 615     /* Escapes {{{ */
 616 <LegacySingleString,LegacyDoubleString>{
 617     \\[0-3][0-7][0-7] R CYLexBufferPoint(X(yytext[1]) << 6 | X(yytext[2]) << 3 | X(yytext[3]));
 618     \\[0-7][0-7] R CYLexBufferUnit(X(yytext[1]) << 3 | X(yytext[2]));
 619     \\[0-7] R CYLexBufferUnit(X(yytext[1]));
 620 }
 621
 622 <StrictSingleString,StrictDoubleString,StrictAccentString>{
 623     \\0[0-7] R E("legacy escape");
 624     \\0 R CYLexBufferUnit('\0');
 625 }
 626
 627 <LegacySingleString,LegacyDoubleString,StrictSingleString,StrictDoubleString,StrictAccentString>{
 628     \\b R CYLexBufferUnit('\b');
 629     \\f R CYLexBufferUnit('\f');
 630     \\n R CYLexBufferUnit('\n');
 631     \\r R CYLexBufferUnit('\r');
 632     \\t R CYLexBufferUnit('\t');
 633     \\v R CYLexBufferUnit('\v');
 634
 635     \\x{HexDigit}{2} R CYLexBufferPoint(X(yytext[2]) << 4 | X(yytext[3]));
 636
 637     \\u{HexDigit}{4} R CYLexBufferPoint(X(yytext[2]) << 12 | X(yytext[3]) << 8 | X(yytext[4]) << 4 | X(yytext[5]));
 638
 639     \\u\{{HexDigit}+\} R {
 640         unsigned point(0);
 641         for (yy_size_t i(3); i != yyleng - 1; ++i)
 642             point = point << 4 | X(yytext[i]);
 643         CYLexBufferPoint(point);
 644     }
 645
 646     \\{LineTerminatorSequence} yylloc->end.Lines();
 647     \\{NoneTerminatorCharacter} R CYLexBufferUnits(yytext + 1, yyleng - 1);
 648     \\{UnicodeScrap} R E("invalid character");
 649
 650     \\(x{HexDigit}{0,1}|u({HexDigit}{0,3}|\{{HexDigit}*))? R E("invalid escape");
 651     <<EOF>> R E("invalid string");
 652 }
 653     /* }}} */
 654
 655 {LineTerminatorSequence} yylloc->step(); yylloc->end.Lines(); yyextra->last_ = true; N
 656 {WhiteSpace} L
 657 {U1}|{UnicodeScrap} L E("invalid character");
 658
 659 <<EOF>> if (yyextra->auto_) { yyextra->auto_ = false; F(tk::AutoComplete, hi::Nothing); } L yyterminate();
 660
 661 %%
 662
 663 #undef yyextra
 664 #define yyextra this
 665 #define yyscanner scanner_
 666
 667 void CYDriver::ScannerInit() {
 668     cylex_init(&scanner_);
 669     cyset_extra(this, scanner_);
 670 }
 671
 672 void CYDriver::ScannerDestroy() {
 673     cylex_destroy(scanner_);
 674 }
 675
 676 void CYDriver::SetRegEx(bool equal) {
 677     CYLexBufferStart(RegularExpression);
 678     CYLexBufferUnit('/');
 679     if (equal)
 680         CYLexBufferUnit('=');
 681 }
 682
 683 void CYDriver::SetCondition(Condition condition) {
 684     struct yyguts_t *yyg(reinterpret_cast<struct yyguts_t *>(scanner_));
 685
 686     switch (condition) {
 687 @begin E4X
 688         case XMLContentCondition:
 689             BEGIN(XMLContent);
 690             break;
 691         case XMLTagCondition:
 692             BEGIN(XMLTag);
 693             break;
 694 @end
 695         default:
 696             _assert(false);
 697     }
 698 }
 699
 700 void CYDriver::PushCondition(Condition condition) {
 701     switch (condition) {
 702 @begin E4X
 703         case XMLContentCondition:
 704             yy_push_state(XMLContent, scanner_);
 705             break;
 706         case XMLTagCondition:
 707             yy_push_state(XMLTag, scanner_);
 708             break;
 709 @end
 710         default:
 711             _assert(false);
 712     }
 713 }
 714
 715 void CYDriver::PopCondition() {
 716     yy_pop_state(scanner_);
 717 }
 718
 719 bool CYLexerHighlight(hi::Value &highlight, CYLocation &location, void *scanner) {
 720     YYSTYPE value;
 721     if (cylex(&value, &location, scanner) == 0)
 722         return false;
 723     highlight = value.highlight_;
 724     return true;
 725 }
 726
 727 #if defined(__clang__)
 728 #pragma clang diagnostic pop
 729 #else
 730 // must not pop -Wunused-function
 731 //#pragma GCC diagnostic pop
 732 #endif