OLD | NEW |
| (Empty) |
1 #!/usr/bin/python | |
2 # Copyright (c) 2011, the Dart project authors. Please see the AUTHORS file | |
3 # for details. All rights reserved. Use of this source code is governed by a | |
4 # BSD-style license that can be found in the LICENSE file. | |
5 | |
6 import re | |
7 import subprocess | |
8 import tempfile | |
9 | |
10 from pegparser import * | |
11 | |
12 # IDL grammar variants. | |
13 WEBIDL_SYNTAX = 0 | |
14 WEBKIT_SYNTAX = 1 | |
15 FREMONTCUT_SYNTAX = 2 | |
16 | |
17 | |
18 class IDLParser(object): | |
19 """IDLParser is a PEG based IDL files parser.""" | |
20 | |
21 def __init__(self, syntax=WEBIDL_SYNTAX): | |
22 """Constructor. | |
23 | |
24 Initializes the IDLParser by defining the grammar and initializing | |
25 a PEGParserinstance. | |
26 | |
27 Args: | |
28 syntax -- supports either WEBIDL_SYNTAX (0) or WEBKIT_SYNTAX (1) | |
29 """ | |
30 self._syntax = syntax | |
31 self._pegparser = PegParser(self._idl_grammar(), | |
32 self._whitespace_grammar(), | |
33 strings_are_tokens=True) | |
34 | |
35 def _idl_grammar(self): | |
36 """Returns the PEG grammar for IDL parsing.""" | |
37 | |
38 # utilities: | |
39 def syntax_switch(w3c_syntax, webkit_syntax, fremontcut_syntax=None): | |
40 """Returns w3c_syntax or web_syntax, depending on the current | |
41 configuration. | |
42 """ | |
43 if self._syntax == WEBIDL_SYNTAX: | |
44 return w3c_syntax | |
45 elif self._syntax == WEBKIT_SYNTAX: | |
46 return webkit_syntax | |
47 elif self._syntax == FREMONTCUT_SYNTAX: | |
48 if fremontcut_syntax is not None: | |
49 return fremontcut_syntax | |
50 return w3c_syntax | |
51 else: | |
52 raise RuntimeError('unsupported IDL syntax %s' % syntax) | |
53 | |
54 # The following grammar is based on the Web IDL's LL(1) grammar | |
55 # (specified in: http://dev.w3.org/2006/webapi/WebIDL/#idl-grammar). | |
56 # It is adjusted to PEG grammar, as well as to also support | |
57 # WebKit IDL and FremontCut grammar. | |
58 | |
59 ###################### BEGIN GRAMMAR ##################### | |
60 | |
61 def Id(): | |
62 return re.compile(r'[\w\_]+') | |
63 | |
64 def _Definitions(): | |
65 return MAYBE(MANY(_Definition)) | |
66 | |
67 def _Definition(): | |
68 return syntax_switch( | |
69 # Web IDL: | |
70 OR(Module, Interface, ExceptionDef, TypeDef, ImplStmt, | |
71 ValueTypeDef, Const), | |
72 # WebKit: | |
73 OR(Module, Interface)) | |
74 | |
75 def Module(): | |
76 return syntax_switch( | |
77 # Web IDL: | |
78 [MAYBE(ExtAttrs), 'module', Id, '{', _Definitions, '}', | |
79 MAYBE(';')], | |
80 # WebKit: | |
81 ['module', MAYBE(ExtAttrs), Id, '{', _Definitions, '}', | |
82 MAYBE(';')], | |
83 # FremontCut: | |
84 [MAYBE(_Annotations), MAYBE(ExtAttrs), 'module', Id, | |
85 '{', _Definitions, '}', MAYBE(';')]) | |
86 | |
87 def Interface(): | |
88 return syntax_switch( | |
89 # Web IDL: | |
90 [MAYBE(ExtAttrs), 'interface', Id, MAYBE(_ParentInterfaces), | |
91 MAYBE(['{', MAYBE(MANY(_Member)), '}']), ';'], | |
92 # WebKit: | |
93 [MAYBE(ExtAttrs), OR('interface', 'exception'), MAYBE(ExtAttrs), Id, MAY
BE(_ParentInterfaces), | |
94 MAYBE(['{', MAYBE(MANY(_Member)), '}']), MAYBE(';')], | |
95 # FremontCut: | |
96 [MAYBE(_Annotations), MAYBE(ExtAttrs), 'interface', | |
97 Id, MAYBE(_ParentInterfaces), MAYBE(['{', MAYBE(MANY(_Member)), | |
98 '}']), ';']) | |
99 | |
100 def _Member(): | |
101 return syntax_switch( | |
102 # Web IDL: | |
103 OR(Const, Attribute, Operation, ExtAttrs), | |
104 # WebKit: | |
105 OR(Const, Attribute, Operation), | |
106 # FremontCut: | |
107 OR(Const, Attribute, Operation)) | |
108 | |
109 # Interface inheritance: | |
110 def _ParentInterfaces(): | |
111 return [':', MANY(ParentInterface, separator=',')] | |
112 | |
113 def ParentInterface(): | |
114 return syntax_switch( | |
115 # Web IDL: | |
116 [InterfaceType], | |
117 # WebKit: | |
118 [InterfaceType], | |
119 # FremontCut: | |
120 [MAYBE(_Annotations), InterfaceType]) | |
121 | |
122 # TypeDef (Web IDL): | |
123 def TypeDef(): | |
124 return ['typedef', Type, Id, ';'] | |
125 | |
126 # TypeDef (Old-school W3C IDLs) | |
127 def ValueTypeDef(): | |
128 return ['valuetype', Id, Type, ';'] | |
129 | |
130 # Implements Statement (Web IDL): | |
131 def ImplStmt(): | |
132 return [ImplStmtImplementor, 'implements', ImplStmtImplemented, | |
133 ';'] | |
134 | |
135 def ImplStmtImplementor(): | |
136 return ScopedName | |
137 | |
138 def ImplStmtImplemented(): | |
139 return ScopedName | |
140 | |
141 # Constants: | |
142 def Const(): | |
143 return syntax_switch( | |
144 # Web IDL: | |
145 [MAYBE(ExtAttrs), 'const', Type, Id, '=', ConstExpr, ';'], | |
146 # WebKit: | |
147 [MAYBE(ExtAttrs), 'const', Type, Id, '=', ConstExpr, ';'], | |
148 # FremontCut: | |
149 [MAYBE(_Annotations), MAYBE(ExtAttrs), 'const', Type, Id, '=', | |
150 ConstExpr, ';']) | |
151 | |
152 def ConstExpr(): | |
153 return OR(_BooleanLiteral, | |
154 _IntegerLiteral, | |
155 _FloatLiteral) | |
156 | |
157 def _BooleanLiteral(): | |
158 return re.compile(r'true|false') | |
159 | |
160 def _IntegerLiteral(): | |
161 return OR(re.compile(r'(0x)?[0-9ABCDEF]+'), | |
162 re.compile(r'[0-9]+')) | |
163 | |
164 def _FloatLiteral(): | |
165 return re.compile(r'[0-9]+\.[0-9]*') | |
166 | |
167 # Attributes: | |
168 def Attribute(): | |
169 return syntax_switch( | |
170 # Web IDL: | |
171 [MAYBE(ExtAttrs), MAYBE(Stringifier), MAYBE(ReadOnly), | |
172 'attribute', Type, Id, MAYBE(_AttrRaises), ';'], | |
173 # WebKit: | |
174 [ | |
175 MAYBE(Stringifier), | |
176 MAYBE(ExtAttrs), MAYBE(Static), MAYBE(ReadOnly), 'attribute', MAYBE(Ex
tAttrs), | |
177 Type, Id, MAYBE(_AttrRaises), ';'], | |
178 # FremontCut: | |
179 [MAYBE(_Annotations), MAYBE(ExtAttrs), | |
180 MAYBE(_AttrGetterSetter), MAYBE(Stringifier), MAYBE(ReadOnly), | |
181 'attribute', Type, Id, MAYBE(_AttrRaises), ';']) | |
182 | |
183 def _AttrRaises(): | |
184 return syntax_switch( | |
185 # Web IDL: | |
186 MANY(OR(GetRaises, SetRaises)), | |
187 # WebKit: | |
188 MANY(OR(GetRaises, SetRaises, Raises), separator=',')) | |
189 | |
190 # Special fremontcut feature: | |
191 def _AttrGetterSetter(): | |
192 return OR(AttrGetter, AttrSetter) | |
193 | |
194 def AttrGetter(): | |
195 return 'getter' | |
196 | |
197 def AttrSetter(): | |
198 return 'setter' | |
199 | |
200 def ReadOnly(): | |
201 return 'readonly' | |
202 | |
203 def GetRaises(): | |
204 return syntax_switch( | |
205 # Web IDL: | |
206 ['getraises', '(', _ScopedNames, ')'], | |
207 # WebKit: | |
208 ['getter', 'raises', '(', _ScopedNames, ')']) | |
209 | |
210 def SetRaises(): | |
211 return syntax_switch( | |
212 # Web IDL: | |
213 ['setraises', '(', _ScopedNames, ')'], | |
214 # WebKit: | |
215 ['setter', 'raises', '(', _ScopedNames, ')']) | |
216 | |
217 # Operation: | |
218 def Operation(): | |
219 return syntax_switch( | |
220 # Web IDL: | |
221 [MAYBE(ExtAttrs), MAYBE(Static), MAYBE(Stringifier), MAYBE(_Specials), | |
222 ReturnType, MAYBE(Id), '(', _Arguments, ')', MAYBE(Raises), | |
223 ';'], | |
224 # WebKit: | |
225 [MAYBE(ExtAttrs), MAYBE(Static), | |
226 ReturnType, MAYBE(Id), '(', _Arguments, ')', | |
227 MAYBE(Raises), ';'], | |
228 # FremontCut: | |
229 [MAYBE(_Annotations), MAYBE(ExtAttrs), MAYBE(Static), MAYBE(Stringifier)
, | |
230 MAYBE(_Specials), ReturnType, MAYBE(Id), '(', _Arguments, ')', | |
231 MAYBE(Raises), ';']) | |
232 | |
233 def Static(): | |
234 return 'static' | |
235 | |
236 def _Specials(): | |
237 return MANY(Special) | |
238 | |
239 def Special(): | |
240 return re.compile(r'getter|setter|creator|deleter|caller') | |
241 | |
242 def Stringifier(): | |
243 return 'stringifier' | |
244 | |
245 def Raises(): | |
246 return ['raises', '(', _ScopedNames, ')'] | |
247 | |
248 # Operation arguments: | |
249 def _Arguments(): | |
250 return MAYBE(MANY(Argument, ',')) | |
251 | |
252 def Argument(): | |
253 return syntax_switch( | |
254 # Web IDL: | |
255 [MAYBE(ExtAttrs), MAYBE(Optional), MAYBE('in'), | |
256 MAYBE(Optional), Type, MAYBE(AnEllipsis), Id], | |
257 # WebKit: | |
258 [MAYBE(Optional), MAYBE('in'), MAYBE(Optional), | |
259 MAYBE(ExtAttrs), Type, Id]) | |
260 | |
261 def Optional(): | |
262 return 'optional' | |
263 | |
264 def AnEllipsis(): | |
265 return '...' | |
266 | |
267 # Exceptions (Web IDL). | |
268 def ExceptionDef(): | |
269 return ['exception', Id, '{', MAYBE(MANY(_ExceptionMember)), '}', | |
270 ';'] | |
271 | |
272 def _ExceptionMember(): | |
273 return OR(Const, ExceptionField, ExtAttrs) | |
274 | |
275 def ExceptionField(): | |
276 return [Type, Id, ';'] | |
277 | |
278 # Types: | |
279 def Type(): | |
280 return _Type | |
281 | |
282 def ReturnType(): | |
283 return OR(VoidType, _Type) | |
284 | |
285 def InterfaceType(): | |
286 return ScopedName | |
287 | |
288 def ArrayModifiers(): | |
289 return re.compile(r'(\[\])+') | |
290 | |
291 def _Type(): | |
292 return OR( | |
293 [OR(AnyType, ObjectType), MAYBE([ArrayModifiers, MAYBE(Nullable)])], | |
294 [_NullableNonArrayType(), MAYBE(ArrayModifiers), MAYBE(Nullable)]) | |
295 | |
296 def _NullableNonArrayType(): | |
297 return [OR(_IntegerType, BooleanType, OctetType, FloatType, | |
298 DoubleType, SequenceType, ScopedName)] | |
299 | |
300 def Nullable(): | |
301 return '?' | |
302 | |
303 def SequenceType(): | |
304 return ['sequence', '<', Type, '>'] | |
305 | |
306 def AnyType(): | |
307 return 'any' | |
308 | |
309 def ObjectType(): | |
310 return re.compile(r'(object|Object)\b') # both spellings. | |
311 | |
312 def VoidType(): | |
313 return 'void' | |
314 | |
315 def _IntegerType(): | |
316 return [MAYBE(Unsigned), OR(ByteType, IntType, LongLongType, | |
317 LongType, OctetType, ShortType)] | |
318 | |
319 def Unsigned(): | |
320 return 'unsigned' | |
321 | |
322 def ShortType(): | |
323 return 'short' | |
324 | |
325 def LongLongType(): | |
326 return ['long', 'long'] | |
327 | |
328 def LongType(): | |
329 return 'long' | |
330 | |
331 def IntType(): | |
332 return 'int' | |
333 | |
334 def ByteType(): | |
335 return 'byte' | |
336 | |
337 def OctetType(): | |
338 return 'octet' | |
339 | |
340 def BooleanType(): | |
341 return 'boolean' | |
342 | |
343 def FloatType(): | |
344 return 'float' | |
345 | |
346 def DoubleType(): | |
347 return 'double' | |
348 | |
349 def _ScopedNames(): | |
350 return MANY(ScopedName, separator=',') | |
351 | |
352 def ScopedName(): | |
353 return re.compile(r'[\w\_\:\.\<\>]+') | |
354 | |
355 # Extended Attributes: | |
356 def ExtAttrs(): | |
357 return ['[', MAYBE(MANY(ExtAttr, ',')), ']'] | |
358 | |
359 def ExtAttr(): | |
360 return [Id, MAYBE(OR(['=', ExtAttrValue], ExtAttrArgList))] | |
361 | |
362 def ExtAttrValue(): | |
363 return OR(ExtAttrFunctionValue, re.compile(r'[\w&0-9:\-\| ]+')) | |
364 | |
365 def ExtAttrFunctionValue(): | |
366 return [Id, ExtAttrArgList] | |
367 | |
368 def ExtAttrArgList(): | |
369 return ['(', MAYBE(MANY(Argument, ',')), ')'] | |
370 | |
371 # Annotations - used in the FremontCut IDL grammar: | |
372 def _Annotations(): | |
373 return MANY(Annotation) | |
374 | |
375 def Annotation(): | |
376 return ['@', Id, MAYBE(_AnnotationBody)] | |
377 | |
378 def _AnnotationBody(): | |
379 return ['(', MAYBE(MANY(AnnotationArg, ',')), ')'] | |
380 | |
381 def AnnotationArg(): | |
382 return [Id, MAYBE(['=', AnnotationArgValue])] | |
383 | |
384 def AnnotationArgValue(): | |
385 return re.compile(r'[\w&0-9:/\-\.]+') | |
386 | |
387 ###################### END GRAMMAR ##################### | |
388 | |
389 # Return the grammar's root rule: | |
390 return MANY(_Definition) | |
391 | |
392 def _whitespace_grammar(self): | |
393 return OR(re.compile(r'\s+'), | |
394 re.compile(r'//.*'), | |
395 re.compile(r'#.*'), | |
396 re.compile(r'/\*.*?\*/', re.S)) | |
397 | |
398 def _pre_process(self, content, defines, includePaths): | |
399 """Pre-processes the content using gcc. | |
400 | |
401 WebKit IDLs require pre-processing by gcc. This is done by invoking | |
402 gcc in a sub-process and capturing the results. | |
403 | |
404 Returns: | |
405 The result of running gcc on the content. | |
406 | |
407 Args: | |
408 content -- text to process. | |
409 defines -- an array of pre-processor defines. | |
410 includePaths -- an array of path strings. | |
411 """ | |
412 # FIXME: Handle gcc not found, or any other processing errors | |
413 gcc = 'gcc' | |
414 cmd = [gcc, '-E', '-P', '-C', '-x', 'c++'] | |
415 for define in defines: | |
416 cmd.append('-D%s' % define) | |
417 cmd.append('-') | |
418 pipe = subprocess.Popen(cmd, stdin=subprocess.PIPE, | |
419 stdout=subprocess.PIPE, stderr=subprocess.PIPE) | |
420 (content, stderr) = pipe.communicate(content) | |
421 return content | |
422 | |
423 def parse(self, content, defines=[], includePaths=[]): | |
424 """Parse the give content string. | |
425 | |
426 The WebKit IDL syntax also allows gcc pre-processing instructions. | |
427 Lists of defined variables and include paths can be provided. | |
428 | |
429 Returns: | |
430 An abstract syntax tree (AST). | |
431 | |
432 Args: | |
433 content -- text to parse. | |
434 defines -- an array of pre-processor defines. | |
435 includePaths -- an array of path strings used by the | |
436 gcc pre-processor. | |
437 """ | |
438 if self._syntax == WEBKIT_SYNTAX: | |
439 content = self._pre_process(content, defines, includePaths) | |
440 | |
441 return self._pegparser.parse(content) | |
OLD | NEW |