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 ['interface', MAYBE(ExtAttrs), Id, MAYBE(_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 [MAYBE(Stringifier), MAYBE(ReadOnly), 'attribute', | |
175 MAYBE(ExtAttrs), Type, Id, MAYBE(_AttrRaises), ';'], | |
176 # FremontCut: | |
177 [MAYBE(_Annotations), MAYBE(ExtAttrs), | |
178 MAYBE(_AttrGetterSetter), MAYBE(Stringifier), MAYBE(ReadOnly), | |
179 'attribute', Type, Id, MAYBE(_AttrRaises), ';']) | |
180 | |
181 def _AttrRaises(): | |
182 return syntax_switch( | |
183 # Web IDL: | |
184 MANY(OR(GetRaises, SetRaises)), | |
185 # WebKit: | |
186 MANY(OR(GetRaises, SetRaises, Raises), separator=',')) | |
187 | |
188 # Special fremontcut feature: | |
189 def _AttrGetterSetter(): | |
190 return OR(AttrGetter, AttrSetter) | |
191 | |
192 def AttrGetter(): | |
193 return 'getter' | |
194 | |
195 def AttrSetter(): | |
196 return 'setter' | |
197 | |
198 def ReadOnly(): | |
199 return 'readonly' | |
200 | |
201 def GetRaises(): | |
202 return syntax_switch( | |
203 # Web IDL: | |
204 ['getraises', '(', _ScopedNames, ')'], | |
205 # WebKit: | |
206 ['getter', 'raises', '(', _ScopedNames, ')']) | |
207 | |
208 def SetRaises(): | |
209 return syntax_switch( | |
210 # Web IDL: | |
211 ['setraises', '(', _ScopedNames, ')'], | |
212 # WebKit: | |
213 ['setter', 'raises', '(', _ScopedNames, ')']) | |
214 | |
215 # Operation: | |
216 def Operation(): | |
217 return syntax_switch( | |
218 # Web IDL: | |
219 [MAYBE(ExtAttrs), MAYBE(Static), MAYBE(Stringifier), MAYBE(_Specials), | |
220 ReturnType, MAYBE(Id), '(', _Arguments, ')', MAYBE(Raises), | |
221 ';'], | |
222 # WebKit: | |
223 [MAYBE(ExtAttrs), MAYBE(Static), | |
224 ReturnType, MAYBE(Id), '(', _Arguments, ')', | |
225 MAYBE(Raises), ';'], | |
226 # FremontCut: | |
227 [MAYBE(_Annotations), MAYBE(ExtAttrs), MAYBE(Static), MAYBE(Stringifier)
, | |
228 MAYBE(_Specials), ReturnType, MAYBE(Id), '(', _Arguments, ')', | |
229 MAYBE(Raises), ';']) | |
230 | |
231 def Static(): | |
232 return 'static' | |
233 | |
234 def _Specials(): | |
235 return MANY(Special) | |
236 | |
237 def Special(): | |
238 return re.compile(r'getter|setter|creator|deleter|caller') | |
239 | |
240 def Stringifier(): | |
241 return 'stringifier' | |
242 | |
243 def Raises(): | |
244 return ['raises', '(', _ScopedNames, ')'] | |
245 | |
246 # Operation arguments: | |
247 def _Arguments(): | |
248 return MAYBE(MANY(Argument, ',')) | |
249 | |
250 def Argument(): | |
251 return syntax_switch( | |
252 # Web IDL: | |
253 [MAYBE(ExtAttrs), MAYBE(Optional), MAYBE('in'), | |
254 MAYBE(Optional), Type, MAYBE(AnEllipsis), Id], | |
255 # WebKit: | |
256 [MAYBE(Optional), MAYBE('in'), MAYBE(Optional), | |
257 MAYBE(ExtAttrs), Type, Id]) | |
258 | |
259 def Optional(): | |
260 return 'optional' | |
261 | |
262 def AnEllipsis(): | |
263 return '...' | |
264 | |
265 # Exceptions (Web IDL). | |
266 def ExceptionDef(): | |
267 return ['exception', Id, '{', MAYBE(MANY(_ExceptionMember)), '}', | |
268 ';'] | |
269 | |
270 def _ExceptionMember(): | |
271 return OR(Const, ExceptionField, ExtAttrs) | |
272 | |
273 def ExceptionField(): | |
274 return [Type, Id, ';'] | |
275 | |
276 # Types: | |
277 def Type(): | |
278 return _Type | |
279 | |
280 def ReturnType(): | |
281 return OR(VoidType, _Type) | |
282 | |
283 def InterfaceType(): | |
284 return ScopedName | |
285 | |
286 def _Type(): | |
287 return OR(AnyArrayType, AnyType, ObjectType, _NullableType) | |
288 | |
289 def _NullableType(): | |
290 return [OR(_IntegerType, BooleanType, OctetType, FloatType, | |
291 DoubleType, SequenceType, DOMStringArrayType, ScopedName), | |
292 MAYBE(Nullable)] | |
293 | |
294 def Nullable(): | |
295 return '?' | |
296 | |
297 def SequenceType(): | |
298 return ['sequence', '<', Type, '>'] | |
299 | |
300 def AnyType(): | |
301 return 'any' | |
302 | |
303 def AnyArrayType(): | |
304 # TODO(sra): Do more general handling of array types. | |
305 return 'any[]' | |
306 | |
307 def ObjectType(): | |
308 return re.compile(r'(object|Object)\b') # both spellings. | |
309 | |
310 def VoidType(): | |
311 return 'void' | |
312 | |
313 def _IntegerType(): | |
314 return [MAYBE(Unsigned), OR(ByteType, IntType, LongLongType, | |
315 LongType, OctetType, ShortType)] | |
316 | |
317 def Unsigned(): | |
318 return 'unsigned' | |
319 | |
320 def ShortType(): | |
321 return 'short' | |
322 | |
323 def LongLongType(): | |
324 return ['long', 'long'] | |
325 | |
326 def LongType(): | |
327 return 'long' | |
328 | |
329 def IntType(): | |
330 return 'int' | |
331 | |
332 def ByteType(): | |
333 return 'byte' | |
334 | |
335 def OctetType(): | |
336 return 'octet' | |
337 | |
338 def BooleanType(): | |
339 return 'boolean' | |
340 | |
341 def FloatType(): | |
342 return 'float' | |
343 | |
344 def DoubleType(): | |
345 return 'double' | |
346 | |
347 def _ScopedNames(): | |
348 return MANY(ScopedName, separator=',') | |
349 | |
350 def ScopedName(): | |
351 return re.compile(r'[\w\_\:\.\<\>]+') | |
352 | |
353 def DOMStringArrayType(): | |
354 return 'DOMString[]' | |
355 | |
356 # Extended Attributes: | |
357 def ExtAttrs(): | |
358 return ['[', MAYBE(MANY(ExtAttr, ',')), ']'] | |
359 | |
360 def ExtAttr(): | |
361 return [Id, MAYBE(OR(['=', ExtAttrValue], ExtAttrArgList))] | |
362 | |
363 def ExtAttrValue(): | |
364 return OR(ExtAttrFunctionValue, re.compile(r'[\w&0-9:\-\|]+')) | |
365 | |
366 def ExtAttrFunctionValue(): | |
367 return [Id, ExtAttrArgList] | |
368 | |
369 def ExtAttrArgList(): | |
370 return ['(', MAYBE(MANY(Argument, ',')), ')'] | |
371 | |
372 # Annotations - used in the FremontCut IDL grammar: | |
373 def _Annotations(): | |
374 return MANY(Annotation) | |
375 | |
376 def Annotation(): | |
377 return ['@', Id, MAYBE(_AnnotationBody)] | |
378 | |
379 def _AnnotationBody(): | |
380 return ['(', MAYBE(MANY(AnnotationArg, ',')), ')'] | |
381 | |
382 def AnnotationArg(): | |
383 return [Id, MAYBE(['=', AnnotationArgValue])] | |
384 | |
385 def AnnotationArgValue(): | |
386 return re.compile(r'[\w&0-9:/\-\.]+') | |
387 | |
388 ###################### END GRAMMAR ##################### | |
389 | |
390 # Return the grammar's root rule: | |
391 return MANY(_Definition) | |
392 | |
393 def _whitespace_grammar(self): | |
394 return OR(re.compile(r'\s+'), | |
395 re.compile(r'//.*'), | |
396 re.compile(r'#.*'), | |
397 re.compile(r'/\*.*?\*/', re.S)) | |
398 | |
399 def _pre_process(self, content, defines, includePaths): | |
400 """Pre-processes the content using gcc. | |
401 | |
402 WebKit IDLs require pre-processing by gcc. This is done by invoking | |
403 gcc in a sub-process and capturing the results. | |
404 | |
405 Returns: | |
406 The result of running gcc on the content. | |
407 | |
408 Args: | |
409 content -- text to process. | |
410 defines -- an array of pre-processor defines. | |
411 includePaths -- an array of path strings. | |
412 """ | |
413 # FIXME: Handle gcc not found, or any other processing errors | |
414 gcc = 'gcc' | |
415 cmd = [gcc, '-E', '-P', '-C', '-x', 'c++']; | |
416 for define in defines: | |
417 cmd.append('-D%s' % define) | |
418 cmd.append('-') | |
419 pipe = subprocess.Popen(cmd, stdin=subprocess.PIPE, | |
420 stdout=subprocess.PIPE, stderr=subprocess.PIPE) | |
421 (content, stderr) = pipe.communicate(content) | |
422 return content | |
423 | |
424 def parse(self, content, defines=[], includePaths=[]): | |
425 """Parse the give content string. | |
426 | |
427 The WebKit IDL syntax also allows gcc pre-processing instructions. | |
428 Lists of defined variables and include paths can be provided. | |
429 | |
430 Returns: | |
431 An abstract syntax tree (AST). | |
432 | |
433 Args: | |
434 content -- text to parse. | |
435 defines -- an array of pre-processor defines. | |
436 includePaths -- an array of path strings used by the | |
437 gcc pre-processor. | |
438 """ | |
439 if self._syntax == WEBKIT_SYNTAX: | |
440 content = self._pre_process(content, defines, includePaths) | |
441 | |
442 return self._pegparser.parse(content) | |
OLD | NEW |