1 /* Cycript - Optimizing JavaScript Compiler/Runtime
2 * Copyright (C) 2009-2015 Jay Freeman (saurik)
5 /* GNU Affero General Public License, Version 3 {{{ */
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.
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.
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/>.
22 /* XXX: supposedly I will be screwed on very very long multi-line comments and need to replace these with a manual lexer. http://websrv.cs.fsu.edu/~engelen/courses/COP5621/Pr2.pdf */
26 #if defined(__clang__)
27 #pragma clang diagnostic push
28 #pragma clang diagnostic ignored "-Wunknown-pragmas"
29 #pragma clang diagnostic ignored "-Wdeprecated-register"
32 #define YYLTYPE CYLocation
33 #include "Cycript.tab.hh"
34 typedef cy::parser::token tk;
36 #include "Highlight.hpp"
38 #define YY_EXTRA_TYPE CYDriver *
40 // do /not/ fold token to the return: this is a macro and the ordering is dependent
41 #define F(value, highlight) do { \
44 yyextra->no_.AtImplementation = false; \
46 yyextra->no_.Function = false; \
47 yyextra->no_.OpenBrace = false; \
48 yylval->highlight_ = highlight; \
53 #define Y $pool.strmemdup(yytext, yyleng)
55 #define I(type, Type, value, highlight) do { \
56 yylval->type ## _ = A CY ## Type; \
57 F(value, highlight); \
60 #define T yylval->newline_ = yyextra->state_ == CYNewLine; BEGIN(Div);
61 #define C T yyextra->state_ = CYClear;
62 #define R T yyextra->state_ = CYRestricted;
65 if (yyextra->state_ != CYNewLine) { \
66 if (yyextra->state_ != CYRestricted) \
67 yyextra->state_ = CYNewLine; \
69 yyextra->state_ = CYClear; \
70 F(tk::NewLine, hi::Nothing); \
75 if (const char *nl = reinterpret_cast<const char *>(memchr(yytext, '\n', yyleng))) { \
80 left = yyleng - (nl - yytext) - 1; \
81 nl = reinterpret_cast<const char *>(memchr(nl + 1, '\n', left)); \
82 } while (nl != NULL); \
84 yylloc->end.lines(lines); \
85 yylloc->end.columns(left); \
92 yylloc->end.columns(yyleng); \
96 if (yyextra->commented_) { \
97 I(comment, Comment(Y), tk::Comment, hi::Comment); \
101 #define E(message) { \
102 CYDriver::Error error; \
103 error.location_ = *yylloc; \
104 error.message_ = "syntax error, " message; \
105 yyextra->errors_.push_back(error); \
110 if (c >= '0' && c <= '9')
112 if (c >= 'a' && c <= 'f')
114 if (c >= 'A' && c <= 'F')
119 #define YY_INPUT(data, value, size) { \
120 if (yyextra->data_.eof()) \
123 yyextra->data_.read(data, size); \
124 size_t copy(yyextra->data_.gcount()); \
125 value = copy == 0 ? YY_NULL : copy; \
133 %option bison-locations
143 %option never-interactive
148 Escape \\[\\'"bfnrtv]|\\0|\\x[0-9a-fA-F]{2}|\\u[0-9a-fA-F]{4}|\\\n
150 IdentifierStart [a-zA-Z$_]
151 IdentifierPart [a-zA-Z$_0-9]
154 BackslashSequence \\{NonTerminator}
155 RegularExpressionFirstChar [^\n*\\/]|{BackslashSequence}
156 RegularExpressionChar [^\n\\/]|{BackslashSequence}
157 RegularExpressionFlags {IdentifierPart}*
158 RegularExpressionChars {RegularExpressionChar}*
159 RegularExpressionBody {RegularExpressionFirstChar}{RegularExpressionChars}
162 XMLNameStart [a-zA-Z_:]
163 XMLNamePart [a-zA-Z0-9.-_:]
164 XMLName {XMLNameStart}{XMLNamePart}*
177 <RegExp>\/{RegularExpressionBody}\/{RegularExpressionFlags} L C I(literal, RegEx(Y), tk::RegularExpressionLiteral, hi::Constant);
178 <RegExp>\/{RegularExpressionBody}?\\? L E("unterminated regex")
184 /* http://ostermiller.org/findcomment.html */
185 /* XXX: unify these two rules using !? */
186 \/\*!([^*]|[\r\n]|(\*+([^*/]|[\r\n])))*\*+\/ V() C I(comment, Comment(Y), tk::Comment, hi::Comment);
187 \/\*([^*]|[\r\n]|(\*+([^*/]|[\r\n])))*\*+\/ V(N) M
188 \/\*([^*]|[\r\n]|(\*+([^*/]|[\r\n])))*\** V() E("invalid comment")
191 <RegExp>"<>" L F(tk::LeftRight, hi::Structure);
192 <XMLContent>"</>" L F(tk::LeftSlashRight, hi::Structure);
194 <RegExp,XMLContent>\<!\[CDATA\[(\n|[^[]|\[[^[]|\[\[[^>])*]]> V() F(tk::XMLCDATA, hi::Constant);
195 <RegExp,XMLContent>\<!--(\n|[^-]|-[^-])*--> V() F(tk::XMLComment, hi::Comment);
196 <RegExp,XMLContent>\<?(\n|[^?]|\?[^>])*?> V() F(tk::XMLPI, hi::Meta);
198 <XMLTag>"=" L F(tk::Equal, hi::Structure);
199 <XMLTag>">" L F(tk::Right, hi::Structure);
200 <XMLTag>"/>" L F(tk::SlashRight, hi::Structure);
201 <XMLTag>"{" L F(tk::OpenBrace, hi::Structure);
203 <XMLTag>\"(\n|[^"])*\"|'(\n|[^'])*' V() F(tk::XMLAttributeValue, hi::Constant);
204 <XMLTag>{XMLName} L F(tk::XMLName, hi::Identifier);
205 <XMLTag>[ \t\r\n] V() F(tk::XMLWhitespace, hi::Nothing);
207 <XMLContent>"{" L F(tk::OpenBrace, hi::Structure);
208 <XMLContent>"<" L F(tk::Left, hi::Structure);
209 <XMLContent>"</" L F(tk::LeftSlash, hi::Structure);
212 "..." L C F(tk::PeriodPeriodPeriod, hi::Meta);
213 ".." L E("invalid operator")
216 "::" L C F(tk::ColonColon, hi::Operator);
217 ".." L C F(tk::PeriodPeriod, hi::Operator);
220 @begin E4X ObjectiveC
221 "@" L C F(tk::At, hi::Operator);
222 "#" L C F(tk::Pound, hi::Operator);
225 "&" L C F(tk::Ampersand, hi::Operator);
226 "&&" L C F(tk::AmpersandAmpersand, hi::Operator);
227 "&=" L C F(tk::AmpersandEqual, hi::Operator);
228 "^" L C F(tk::Carrot, hi::Operator);
229 "^=" L C F(tk::CarrotEqual, hi::Operator);
230 "=" L C F(tk::Equal, hi::Operator);
231 "==" L C F(tk::EqualEqual, hi::Operator);
232 "===" L C F(tk::EqualEqualEqual, hi::Operator);
233 "=>" L C F(tk::EqualRight, hi::Operator);
234 "!" L C F(tk::Exclamation, hi::Operator);
235 "!=" L C F(tk::ExclamationEqual, hi::Operator);
236 "!==" L C F(tk::ExclamationEqualEqual, hi::Operator);
237 "-" L C F(tk::Hyphen, hi::Operator);
238 "-=" L C F(tk::HyphenEqual, hi::Operator);
239 "--" L C F(yylval->newline_ ? tk::HyphenHyphen_ : tk::HyphenHyphen, hi::Operator);
240 "->" L C F(tk::HyphenRight, hi::Operator);
241 "<" L C F(tk::Left, hi::Operator);
242 "<=" L C F(tk::LeftEqual, hi::Operator);
243 "<<" L C F(tk::LeftLeft, hi::Operator);
244 "<<=" L C F(tk::LeftLeftEqual, hi::Operator);
245 "%" L C F(tk::Percent, hi::Operator);
246 "%=" L C F(tk::PercentEqual, hi::Operator);
247 "." L C F(tk::Period, hi::Operator);
248 "|" L C F(tk::Pipe, hi::Operator);
249 "|=" L C F(tk::PipeEqual, hi::Operator);
250 "||" L C F(tk::PipePipe, hi::Operator);
251 "+" L C F(tk::Plus, hi::Operator);
252 "+=" L C F(tk::PlusEqual, hi::Operator);
253 "++" L C F(yylval->newline_ ? tk::PlusPlus_ : tk::PlusPlus, hi::Operator);
254 ">" L C F(tk::Right, hi::Operator);
255 ">=" L C F(tk::RightEqual, hi::Operator);
256 ">>" L C F(tk::RightRight, hi::Operator);
257 ">>=" L C F(tk::RightRightEqual, hi::Operator);
258 ">>>" L C F(tk::RightRightRight, hi::Operator);
259 ">>>=" L C F(tk::RightRightRightEqual, hi::Operator);
260 "*" L C F(tk::Star, hi::Operator);
261 "*=" L C F(tk::StarEqual, hi::Operator);
262 "~" L C F(tk::Tilde, hi::Operator);
264 <Div>"/" L C F(tk::Slash, hi::Operator);
265 <Div>"/=" L C F(tk::SlashEqual, hi::Operator);
267 ":" L C F(tk::Colon, hi::Structure);
268 "," L C F(tk::Comma, hi::Structure);
269 "?" L C F(tk::Question, hi::Structure);
270 ";" L C F(tk::SemiColon, hi::Structure);
272 "(" L C F(tk::OpenParen, hi::Structure);
273 ")" L C F(tk::CloseParen, hi::Structure);
275 "{" L C F(yyextra->no_.OpenBrace ? tk::OpenBrace__ : yylval->newline_ ? tk::OpenBrace_ : tk::OpenBrace, hi::Structure);
276 "}" L C F(tk::CloseBrace, hi::Structure);
278 "[" L C F(tk::OpenBracket, hi::Structure);
279 "]" L C F(tk::CloseBracket, hi::Structure);
281 "@error" L C F(tk::AtError, hi::Error);
284 "@class" L C F(tk::AtClass, hi::Meta);
288 "typedef" L C I(identifier, Identifier("typedef"), tk::Typedef, hi::Meta);
289 "unsigned" L C I(identifier, Identifier("unsigned"), tk::Unsigned, hi::Type);
290 "signed" L C I(identifier, Identifier("signed"), tk::Signed, hi::Type);
291 "extern" L C I(identifier, Identifier("extern"), tk::Extern, hi::Type);
295 "@encode" L C F(tk::AtEncode, hi::Meta);
299 "@end" L C F(tk::AtEnd, hi::Meta);
300 "@implementation" L C F(yyextra->no_.AtImplementation ? tk::AtImplementation_ : tk::AtImplementation, hi::Meta);
301 "@import" L C F(tk::AtImport, hi::Special);
302 "@selector" L C F(tk::AtSelector, hi::Meta);
304 "NULL" L C I(identifier, Identifier("NULL"), tk::Identifier_, hi::Constant);
305 "nil" L C I(identifier, Identifier("nil"), tk::Identifier_, hi::Constant);
306 "YES" L C I(identifier, Identifier("YES"), tk::Yes, hi::Constant);
307 "NO" L C I(identifier, Identifier("NO"), tk::No, hi::Constant);
309 "bool" L C I(identifier, Identifier("bool"), tk::Identifier_, hi::Type);
310 "BOOL" L C I(identifier, Identifier("BOOL"), tk::Identifier_, hi::Type);
311 "id" L C I(identifier, Identifier("id"), tk::Identifier_, hi::Type);
312 "SEL" L C I(identifier, Identifier("SEL"), tk::Identifier_, hi::Type);
315 "undefined" L C I(identifier, Identifier("undefined"), tk::Identifier_, hi::Operator);
317 "false" L C I(false, False(), tk::False, hi::Constant);
318 "null" L C I(null, Null(), tk::Null, hi::Constant);
319 "true" L C I(true, True(), tk::True, hi::Constant);
321 "auto" L C I(word, Word("auto"), tk::Auto, hi::Meta);
322 "break" L R I(word, Word("break"), tk::Break, hi::Control);
323 "case" L C I(word, Word("case"), tk::Case, hi::Control);
324 "catch" L C I(word, Word("catch"), tk::Catch, hi::Control);
325 "continue" L R I(word, Word("continue"), tk::Continue, hi::Control);
326 "default" L C I(word, Word("default"), tk::Default, hi::Control);
327 "delete" L C I(word, Word("delete"), tk::Delete, hi::Operator);
328 "do" L C I(word, Word("do"), tk::Do, hi::Control);
329 "else" L C I(word, Word("else"), tk::Else, hi::Control);
330 "finally" L C I(word, Word("finally"), tk::Finally, hi::Control);
331 "for" L C I(word, Word("for"), tk::For, hi::Control);
332 "function" L C I(word, Word("function"), yyextra->no_.Function ? tk::Function_ : tk::Function, hi::Meta);
333 "if" L C I(word, Word("if"), tk::If, hi::Control);
334 "in" L C I(word, Word("in"), yyextra->in_.top() ? tk::In_ : tk::In, hi::Operator);
335 "instanceof" L C I(word, Word("instanceof"), tk::InstanceOf, hi::Operator);
336 "new" L C I(word, Word("new"), tk::New, hi::Operator);
337 "return" L R I(word, Word("return"), tk::Return, hi::Control);
338 "switch" L C I(word, Word("switch"), tk::Switch, hi::Control);
339 "this" L C I(this, This(), tk::This, hi::Constant);
340 "throw" L R I(word, Word("throw"), tk::Throw, hi::Control);
341 "try" L C I(word, Word("try"), tk::Try, hi::Control);
342 "typeof" L C I(word, Word("typeof"), tk::TypeOf, hi::Operator);
343 "var" L C I(word, Word("var"), tk::Var, hi::Meta);
344 "void" L C I(word, Word("void"), tk::Void, hi::Operator);
345 "while" L C I(word, Word("while"), tk::While, hi::Control);
346 "with" L C I(word, Word("with"), tk::With, hi::Control);
348 "debugger" L C I(word, Word("debugger"), tk::Debugger, hi::Meta);
350 "const" L C I(word, Word("const"), tk::Const, hi::Meta);
352 "class" L C I(word, Word("class"), tk::Class, hi::Meta);
353 "enum" L C I(word, Word("enum"), tk::Enum, hi::Meta);
354 "export" L C I(word, Word("export"), tk::Export, hi::Meta);
355 "extends" L C I(word, Word("extends"), tk::Extends, hi::Meta);
356 "import" L C I(word, Word("import"), tk::Import, hi::Meta);
357 "super" L C I(word, Word("super"), tk::Super, hi::Constant);
359 "implements" L C I(identifier, Identifier("implements"), tk::Implements, hi::Meta);
360 "interface" L C I(identifier, Identifier("interface"), tk::Interface, hi::Meta);
361 "package" L C I(identifier, Identifier("package"), tk::Package, hi::Meta);
362 "private" L C I(identifier, Identifier("private"), tk::Private, hi::Meta);
363 "protected" L C I(identifier, Identifier("protected"), tk::Protected, hi::Meta);
364 "public" L C I(identifier, Identifier("public"), tk::Public, hi::Meta);
365 "static" L C I(identifier, Identifier("static"), tk::Static, hi::Meta);
367 "abstract" L C I(identifier, Identifier("abstract"), tk::Abstract, hi::Meta);
368 "boolean" L C I(identifier, Identifier("boolean"), tk::Boolean, hi::Type);
369 "byte" L C I(identifier, Identifier("byte"), tk::Byte, hi::Type);
370 "char" L C I(identifier, Identifier("char"), tk::Char, hi::Type);
371 "double" L C I(identifier, Identifier("double"), tk::Double, hi::Type);
372 "final" L C I(identifier, Identifier("final"), tk::Final, hi::Meta);
373 "float" L C I(identifier, Identifier("float"), tk::Float, hi::Type);
374 "goto" L C I(identifier, Identifier("goto"), tk::Goto, hi::Control);
375 "int" L C I(identifier, Identifier("int"), tk::Int, hi::Type);
376 "long" L C I(identifier, Identifier("long"), tk::Long, hi::Type);
377 "native" L C I(identifier, Identifier("native"), tk::Native, hi::Meta);
378 "short" L C I(identifier, Identifier("short"), tk::Short, hi::Type);
379 "synchronized" L C I(identifier, Identifier("synchronized"), tk::Synchronized, hi::Meta);
380 "throws" L C I(identifier, Identifier("throws"), tk::Throws, hi::Meta);
381 "transient" L C I(identifier, Identifier("transient"), tk::Transient, hi::Meta);
382 "volatile" L C I(identifier, Identifier("volatile"), tk::Volatile, hi::Meta);
384 "let" L C I(identifier, Identifier("let"), tk::Let, hi::Meta);
385 "yield" L C I(identifier, Identifier("yield"), tk::Yield, hi::Control);
387 "each" L C I(identifier, Identifier("each"), tk::Each, hi::Control);
388 "of" L C I(identifier, Identifier("of"), tk::Of, hi::Operator);
391 "namespace" L C I(identifier, Identifier("namespace"), tk::Namespace, hi::Meta);
392 "xml" L C I(identifier, Identifier("xml"), tk::XML, hi::Meta);
395 {IdentifierStart}{IdentifierPart}* L C I(identifier, Identifier(Y), tk::Identifier_, hi::Identifier);
398 0[xX][0-9a-fA-F]+ L C I(number, Number(strtoull(yytext + 2, NULL, 16)), tk::NumericLiteral, hi::Constant);
399 0[0-7]+ L C I(number, Number(strtoull(yytext + 1, NULL, 8)), tk::NumericLiteral, hi::Constant);
400 0[bB][0-1]+ L C I(number, Number(strtoull(yytext + 2, NULL, 2)), tk::NumericLiteral, hi::Constant);
402 (\.[0-9]+|(0|[1-9][0-9]*)(\.[0-9]*)?)([eE][+-]?[0-9]+)? L C I(number, Number(strtod(yytext, NULL)), tk::NumericLiteral, hi::Constant);
403 (\.[0-9]+|(0|[1-9][0-9]*)(\.[0-9]*)?)[eE][+-]?{IdentifierPart}* L E("invalid exponent")
404 (\.?[0-9]|(0|[1-9][0-9]*)?\.){IdentifierPart}* L E("invalid number")
406 \"([^"\\\n]|{Escape})*\"|'([^'\\\n]|{Escape})*' L C {
407 char *value(A char[yyleng]);
410 for (yy_size_t i(1), e(yyleng - 1); i != e; ++i) {
411 char next(yytext[i]);
413 if (yytext[i] == '\\')
414 switch (next = yytext[++i]) {
416 case '\\': next = '\\'; break;
417 case '\'': next = '\''; break;
418 case '"': next = '"'; break;
419 case 'b': next = '\b'; break;
420 case 'f': next = '\f'; break;
421 case 'n': next = '\n'; break;
422 case 'r': next = '\r'; break;
423 case 't': next = '\t'; break;
424 case 'v': next = '\v'; break;
425 case '0': next = '\0'; break;
428 next = H(yytext[i + 1]) << 4 | H(yytext[i + 2]);
437 I(string, String(value, local - value), tk::StringLiteral, hi::Constant);
440 (\"([^"\\\n]|{Escape})*|'([^'\\\n]|{Escape})*)(\\(x.{0,2}|u.{0,4})?)? L E("invalid escape")
442 \r?\n|\r|\xe2\x80[\xa8\xa9] yylloc->step(); yylloc->end.lines(); N
446 <<EOF>> if (yyextra->auto_) { yyextra->auto_ = false; F(tk::AutoComplete, hi::Nothing); } L yyterminate();
448 @{IdentifierPart}+|\xe2.|. L E("unknown token")
452 void CYDriver::ScannerInit() {
453 cylex_init(&scanner_);
454 cyset_extra(this, scanner_);
457 void CYDriver::ScannerDestroy() {
458 cylex_destroy(scanner_);
461 CYDriver::Condition CYDriver::GetCondition() {
462 switch (yy_top_state(scanner_)) {
464 return RegExpCondition;
467 return XMLContentCondition;
469 return XMLTagCondition;
476 void CYDriver::SetCondition(Condition condition) {
477 struct yyguts_t *yyg(reinterpret_cast<struct yyguts_t *>(scanner_));
480 case RegExpCondition:
484 case XMLContentCondition:
487 case XMLTagCondition:
496 void CYDriver::PushCondition(Condition condition) {
498 case RegExpCondition:
499 yy_push_state(RegExp, scanner_);
502 case XMLContentCondition:
503 yy_push_state(XMLContent, scanner_);
505 case XMLTagCondition:
506 yy_push_state(XMLTag, scanner_);
514 void CYDriver::PopCondition() {
515 yy_pop_state(scanner_);
518 #if defined(__clang__)
519 #pragma clang diagnostic pop