| 1 | """Parse tree transformation module.
|
| 2 |
|
| 3 | Transforms Python source code into an abstract syntax tree (AST)
|
| 4 | defined in the ast module.
|
| 5 |
|
| 6 | The simplest ways to invoke this module are via parse and parseFile.
|
| 7 | parse(buf) -> AST
|
| 8 | parseFile(path) -> AST
|
| 9 | """
|
| 10 |
|
| 11 | # Original version written by Greg Stein (gstein@lyra.org)
|
| 12 | # and Bill Tutt (rassilon@lima.mudlib.org)
|
| 13 | # February 1997.
|
| 14 | #
|
| 15 | # Modifications and improvements for Python 2.0 by Jeremy Hylton and
|
| 16 | # Mark Hammond
|
| 17 | #
|
| 18 | # Some fixes to try to have correct line number on almost all nodes
|
| 19 | # (except Module, Discard and Stmt) added by Sylvain Thenault
|
| 20 | #
|
| 21 | # Portions of this file are:
|
| 22 | # Copyright (C) 1997-1998 Greg Stein. All Rights Reserved.
|
| 23 | #
|
| 24 | # This module is provided under a BSD-ish license. See
|
| 25 | # http://www.opensource.org/licenses/bsd-license.html
|
| 26 | # and replace OWNER, ORGANIZATION, and YEAR as appropriate.
|
| 27 |
|
| 28 | from .ast import *
|
| 29 | from .consts import CO_VARARGS, CO_VARKEYWORDS
|
| 30 | from .consts import OP_ASSIGN, OP_DELETE, OP_APPLY
|
| 31 |
|
| 32 | from ..pytree import type_repr
|
| 33 | from pgen2 import token
|
| 34 |
|
| 35 |
|
| 36 | symbol = None
|
| 37 |
|
| 38 | def Init(sym):
|
| 39 | """Replacement for the stdlib symbol module.
|
| 40 |
|
| 41 | Args:
|
| 42 | sym: module
|
| 43 |
|
| 44 | The stdlib module is generated from pgen.c's output data. pgen2 derives it
|
| 45 | from the grammar directly.
|
| 46 |
|
| 47 | """
|
| 48 | global symbol
|
| 49 | symbol = sym
|
| 50 | _InitGlobals()
|
| 51 |
|
| 52 |
|
| 53 | # Various constants
|
| 54 | _doc_nodes = []
|
| 55 | _legal_node_types = []
|
| 56 | _assign_types = []
|
| 57 | # NOTE: This is somewhat duplicated in pytree.py as type_repr.
|
| 58 | _names = {}
|
| 59 |
|
| 60 | def _InitGlobals():
|
| 61 | _doc_nodes.extend([
|
| 62 | symbol.expr_stmt,
|
| 63 | symbol.testlist,
|
| 64 | symbol.testlist_safe,
|
| 65 | symbol.test,
|
| 66 | symbol.or_test,
|
| 67 | symbol.and_test,
|
| 68 | symbol.not_test,
|
| 69 | symbol.comparison,
|
| 70 | symbol.expr,
|
| 71 | symbol.xor_expr,
|
| 72 | symbol.and_expr,
|
| 73 | symbol.shift_expr,
|
| 74 | symbol.arith_expr,
|
| 75 | symbol.term,
|
| 76 | symbol.factor,
|
| 77 | symbol.power,
|
| 78 | ])
|
| 79 |
|
| 80 | _legal_node_types.extend([
|
| 81 | symbol.funcdef,
|
| 82 | symbol.classdef,
|
| 83 | symbol.stmt,
|
| 84 | symbol.small_stmt,
|
| 85 | symbol.flow_stmt,
|
| 86 | symbol.simple_stmt,
|
| 87 | symbol.compound_stmt,
|
| 88 | symbol.expr_stmt,
|
| 89 | symbol.print_stmt,
|
| 90 | symbol.del_stmt,
|
| 91 | symbol.pass_stmt,
|
| 92 | symbol.break_stmt,
|
| 93 | symbol.continue_stmt,
|
| 94 | symbol.return_stmt,
|
| 95 | symbol.raise_stmt,
|
| 96 | symbol.import_stmt,
|
| 97 | symbol.global_stmt,
|
| 98 | symbol.exec_stmt,
|
| 99 | symbol.assert_stmt,
|
| 100 | symbol.if_stmt,
|
| 101 | symbol.while_stmt,
|
| 102 | symbol.for_stmt,
|
| 103 | symbol.try_stmt,
|
| 104 | symbol.with_stmt,
|
| 105 | symbol.suite,
|
| 106 | symbol.testlist,
|
| 107 | symbol.testlist_safe,
|
| 108 | symbol.test,
|
| 109 | symbol.and_test,
|
| 110 | symbol.not_test,
|
| 111 | symbol.comparison,
|
| 112 | symbol.exprlist,
|
| 113 | symbol.expr,
|
| 114 | symbol.xor_expr,
|
| 115 | symbol.and_expr,
|
| 116 | symbol.shift_expr,
|
| 117 | symbol.arith_expr,
|
| 118 | symbol.term,
|
| 119 | symbol.factor,
|
| 120 | symbol.power,
|
| 121 | symbol.atom,
|
| 122 | ])
|
| 123 |
|
| 124 | if hasattr(symbol, 'yield_stmt'):
|
| 125 | _legal_node_types.append(symbol.yield_stmt)
|
| 126 | if hasattr(symbol, 'yield_expr'):
|
| 127 | _legal_node_types.append(symbol.yield_expr)
|
| 128 |
|
| 129 | _assign_types.extend([
|
| 130 | symbol.test,
|
| 131 | symbol.or_test,
|
| 132 | symbol.and_test,
|
| 133 | symbol.not_test,
|
| 134 | symbol.comparison,
|
| 135 | symbol.expr,
|
| 136 | symbol.xor_expr,
|
| 137 | symbol.and_expr,
|
| 138 | symbol.shift_expr,
|
| 139 | symbol.arith_expr,
|
| 140 | symbol.term,
|
| 141 | symbol.factor,
|
| 142 | ])
|
| 143 |
|
| 144 | # Do this first because NT_OFFSET (non-terminal offset) conflicts with
|
| 145 | # file_input.
|
| 146 | for k, v in token.tok_name.items():
|
| 147 | _names[k] = v
|
| 148 | for k, v in symbol.number2symbol.items():
|
| 149 | _names[k] = v
|
| 150 |
|
| 151 |
|
| 152 | # comp_op: '<' | '>' | '=' | '>=' | '<=' | '<>' | '!=' | '=='
|
| 153 | # | 'in' | 'not' 'in' | 'is' | 'is' 'not'
|
| 154 | _cmp_types = {
|
| 155 | token.LESS : '<',
|
| 156 | token.GREATER : '>',
|
| 157 | token.EQEQUAL : '==',
|
| 158 | token.EQUAL : '==',
|
| 159 | token.LESSEQUAL : '<=',
|
| 160 | token.GREATEREQUAL : '>=',
|
| 161 | token.NOTEQUAL : '!=',
|
| 162 | }
|
| 163 |
|
| 164 |
|
| 165 | class WalkerError(StandardError):
|
| 166 | pass
|
| 167 |
|
| 168 |
|
| 169 | def asList(nodes):
|
| 170 | l = []
|
| 171 | for item in nodes:
|
| 172 | if hasattr(item, "asList"):
|
| 173 | l.append(item.asList())
|
| 174 | else:
|
| 175 | if type(item) is type( (None, None) ):
|
| 176 | l.append(tuple(asList(item)))
|
| 177 | elif type(item) is type( [] ):
|
| 178 | l.append(asList(item))
|
| 179 | else:
|
| 180 | l.append(item)
|
| 181 | return l
|
| 182 |
|
| 183 | def extractLineNo(ast):
|
| 184 | if not isinstance(ast[1], tuple):
|
| 185 | # get a terminal node
|
| 186 | return ast[2]
|
| 187 | for child in ast[1:]:
|
| 188 | if isinstance(child, tuple):
|
| 189 | lineno = extractLineNo(child)
|
| 190 | if lineno is not None:
|
| 191 | return lineno
|
| 192 |
|
| 193 | def Node(*args):
|
| 194 | kind = args[0]
|
| 195 | if kind in nodes:
|
| 196 | try:
|
| 197 | return nodes[kind](*args[1:])
|
| 198 | except TypeError:
|
| 199 | print(nodes[kind], len(args), args)
|
| 200 | raise
|
| 201 | else:
|
| 202 | raise WalkerError, "Can't find appropriate Node type: %s" % str(args)
|
| 203 | #return apply(ast.Node, args)
|
| 204 |
|
| 205 | class Transformer(object):
|
| 206 | """Transform a parse tree into an AST."""
|
| 207 |
|
| 208 | def __init__(self):
|
| 209 | self._dispatch = {}
|
| 210 | for value, name in symbol.number2symbol.items():
|
| 211 | if hasattr(self, name):
|
| 212 | self._dispatch[value] = getattr(self, name)
|
| 213 | self._dispatch[token.NEWLINE] = self.com_NEWLINE
|
| 214 | self._atom_dispatch = {token.LPAR: self.atom_lpar,
|
| 215 | token.LSQB: self.atom_lsqb,
|
| 216 | token.LBRACE: self.atom_lbrace,
|
| 217 | token.BACKQUOTE: self.atom_backquote,
|
| 218 | token.NUMBER: self.atom_number,
|
| 219 | token.STRING: self.atom_string,
|
| 220 | token.NAME: self.atom_name,
|
| 221 | }
|
| 222 | self.encoding = None
|
| 223 |
|
| 224 | def transform(self, node):
|
| 225 | ### emit a line-number node?
|
| 226 | n = node[0]
|
| 227 |
|
| 228 | if n == symbol.encoding_decl:
|
| 229 | self.encoding = node[2]
|
| 230 | node = node[1]
|
| 231 | n = node[0]
|
| 232 |
|
| 233 | if n == symbol.single_input:
|
| 234 | return self.single_input(node[1:])
|
| 235 | if n == symbol.file_input:
|
| 236 | return self.file_input(node[1:])
|
| 237 | if n == symbol.eval_input:
|
| 238 | return self.eval_input(node[1:])
|
| 239 | if n == symbol.lambdef:
|
| 240 | return self.lambdef(node[1:])
|
| 241 | if n == symbol.funcdef:
|
| 242 | return self.funcdef(node[1:])
|
| 243 | if n == symbol.classdef:
|
| 244 | return self.classdef(node[1:])
|
| 245 |
|
| 246 | raise WalkerError('unexpected node type %r' % type_repr(n))
|
| 247 |
|
| 248 | # --------------------------------------------------------------
|
| 249 | #
|
| 250 | # PRIVATE METHODS
|
| 251 | #
|
| 252 |
|
| 253 | def single_input(self, node):
|
| 254 | ### do we want to do anything about being "interactive" ?
|
| 255 |
|
| 256 | # NEWLINE | simple_stmt | compound_stmt NEWLINE
|
| 257 | n = node[0][0]
|
| 258 | if n != token.NEWLINE:
|
| 259 | return self.com_stmt(node[0])
|
| 260 |
|
| 261 | return Pass()
|
| 262 |
|
| 263 | def file_input(self, nodelist):
|
| 264 | doc = self.get_docstring(nodelist, symbol.file_input)
|
| 265 | if doc is not None:
|
| 266 | i = 1
|
| 267 | else:
|
| 268 | i = 0
|
| 269 | stmts = []
|
| 270 | for node in nodelist[i:]:
|
| 271 | if node[0] != token.ENDMARKER and node[0] != token.NEWLINE:
|
| 272 | self.com_append_stmt(stmts, node)
|
| 273 | return Module(doc, Stmt(stmts))
|
| 274 |
|
| 275 | def eval_input(self, nodelist):
|
| 276 | # from the built-in function input()
|
| 277 | ### is this sufficient?
|
| 278 | return Expression(self.com_node(nodelist[0]))
|
| 279 |
|
| 280 | def decorator_name(self, nodelist):
|
| 281 | listlen = len(nodelist)
|
| 282 | assert listlen >= 1 and listlen % 2 == 1
|
| 283 |
|
| 284 | item = self.atom_name(nodelist)
|
| 285 | i = 1
|
| 286 | while i < listlen:
|
| 287 | assert nodelist[i][0] == token.DOT
|
| 288 | assert nodelist[i + 1][0] == token.NAME
|
| 289 | item = Getattr(item, nodelist[i + 1][1])
|
| 290 | i += 2
|
| 291 |
|
| 292 | return item
|
| 293 |
|
| 294 | def decorator(self, nodelist):
|
| 295 | # '@' dotted_name [ '(' [arglist] ')' ]
|
| 296 | assert len(nodelist) in (3, 5, 6)
|
| 297 | assert nodelist[0][0] == token.AT
|
| 298 | assert nodelist[-1][0] == token.NEWLINE
|
| 299 |
|
| 300 | assert nodelist[1][0] == symbol.dotted_name
|
| 301 | funcname = self.decorator_name(nodelist[1][1:])
|
| 302 |
|
| 303 | if len(nodelist) > 3:
|
| 304 | assert nodelist[2][0] == token.LPAR
|
| 305 | expr = self.com_call_function(funcname, nodelist[3])
|
| 306 | else:
|
| 307 | expr = funcname
|
| 308 |
|
| 309 | return expr
|
| 310 |
|
| 311 | def decorators(self, nodelist):
|
| 312 | # decorators: decorator ([NEWLINE] decorator)* NEWLINE
|
| 313 | items = []
|
| 314 | for dec_nodelist in nodelist:
|
| 315 | assert dec_nodelist[0] == symbol.decorator
|
| 316 | items.append(self.decorator(dec_nodelist[1:]))
|
| 317 | return Decorators(items)
|
| 318 |
|
| 319 | def decorated(self, nodelist):
|
| 320 | assert nodelist[0][0] == symbol.decorators
|
| 321 | if nodelist[1][0] == symbol.funcdef:
|
| 322 | n = [nodelist[0]] + list(nodelist[1][1:])
|
| 323 | return self.funcdef(n)
|
| 324 | elif nodelist[1][0] == symbol.classdef:
|
| 325 | decorators = self.decorators(nodelist[0][1:])
|
| 326 | cls = self.classdef(nodelist[1][1:])
|
| 327 | cls.decorators = decorators
|
| 328 | return cls
|
| 329 | raise WalkerError()
|
| 330 |
|
| 331 | def funcdef(self, nodelist):
|
| 332 | # -6 -5 -4 -3 -2 -1
|
| 333 | # funcdef: [decorators] 'def' NAME parameters ':' suite
|
| 334 | # parameters: '(' [varargslist] ')'
|
| 335 |
|
| 336 | if len(nodelist) == 6:
|
| 337 | assert nodelist[0][0] == symbol.decorators
|
| 338 | decorators = self.decorators(nodelist[0][1:])
|
| 339 | else:
|
| 340 | assert len(nodelist) == 5
|
| 341 | decorators = None
|
| 342 |
|
| 343 | lineno = nodelist[-4][2]
|
| 344 | name = nodelist[-4][1]
|
| 345 | args = nodelist[-3][2]
|
| 346 |
|
| 347 | if args[0] == symbol.varargslist:
|
| 348 | names, defaults, flags = self.com_arglist(args[1:])
|
| 349 | else:
|
| 350 | names = defaults = ()
|
| 351 | flags = 0
|
| 352 | doc = self.get_docstring(nodelist[-1])
|
| 353 |
|
| 354 | # code for function
|
| 355 | code = self.com_node(nodelist[-1])
|
| 356 |
|
| 357 | if doc is not None:
|
| 358 | assert isinstance(code, Stmt)
|
| 359 | assert isinstance(code.nodes[0], Discard)
|
| 360 | del code.nodes[0]
|
| 361 | return Function(decorators, name, names, defaults, flags, doc, code,
|
| 362 | lineno=lineno)
|
| 363 |
|
| 364 | def lambdef(self, nodelist):
|
| 365 | # lambdef: 'lambda' [varargslist] ':' test
|
| 366 | if nodelist[2][0] == symbol.varargslist:
|
| 367 | names, defaults, flags = self.com_arglist(nodelist[2][1:])
|
| 368 | else:
|
| 369 | names = defaults = ()
|
| 370 | flags = 0
|
| 371 |
|
| 372 | # code for lambda
|
| 373 | code = self.com_node(nodelist[-1])
|
| 374 |
|
| 375 | return Lambda(names, defaults, flags, code, lineno=nodelist[1][2])
|
| 376 | old_lambdef = lambdef
|
| 377 |
|
| 378 | def classdef(self, nodelist):
|
| 379 | # classdef: 'class' NAME ['(' [testlist] ')'] ':' suite
|
| 380 |
|
| 381 | name = nodelist[1][1]
|
| 382 | doc = self.get_docstring(nodelist[-1])
|
| 383 | if nodelist[2][0] == token.COLON:
|
| 384 | bases = []
|
| 385 | elif nodelist[3][0] == token.RPAR:
|
| 386 | bases = []
|
| 387 | else:
|
| 388 | bases = self.com_bases(nodelist[3])
|
| 389 |
|
| 390 | # code for class
|
| 391 | code = self.com_node(nodelist[-1])
|
| 392 |
|
| 393 | if doc is not None:
|
| 394 | assert isinstance(code, Stmt)
|
| 395 | assert isinstance(code.nodes[0], Discard)
|
| 396 | del code.nodes[0]
|
| 397 |
|
| 398 | return Class(name, bases, doc, code, lineno=nodelist[1][2])
|
| 399 |
|
| 400 | def stmt(self, nodelist):
|
| 401 | return self.com_stmt(nodelist[0])
|
| 402 |
|
| 403 | small_stmt = stmt
|
| 404 | flow_stmt = stmt
|
| 405 | compound_stmt = stmt
|
| 406 |
|
| 407 | def simple_stmt(self, nodelist):
|
| 408 | # small_stmt (';' small_stmt)* [';'] NEWLINE
|
| 409 | stmts = []
|
| 410 | for i in range(0, len(nodelist), 2):
|
| 411 | self.com_append_stmt(stmts, nodelist[i])
|
| 412 | return Stmt(stmts)
|
| 413 |
|
| 414 | def parameters(self, nodelist):
|
| 415 | raise WalkerError
|
| 416 |
|
| 417 | def varargslist(self, nodelist):
|
| 418 | raise WalkerError
|
| 419 |
|
| 420 | def fpdef(self, nodelist):
|
| 421 | raise WalkerError
|
| 422 |
|
| 423 | def fplist(self, nodelist):
|
| 424 | raise WalkerError
|
| 425 |
|
| 426 | def dotted_name(self, nodelist):
|
| 427 | raise WalkerError
|
| 428 |
|
| 429 | def comp_op(self, nodelist):
|
| 430 | raise WalkerError
|
| 431 |
|
| 432 | def trailer(self, nodelist):
|
| 433 | raise WalkerError
|
| 434 |
|
| 435 | def sliceop(self, nodelist):
|
| 436 | raise WalkerError
|
| 437 |
|
| 438 | def argument(self, nodelist):
|
| 439 | raise WalkerError
|
| 440 |
|
| 441 | # --------------------------------------------------------------
|
| 442 | #
|
| 443 | # STATEMENT NODES (invoked by com_node())
|
| 444 | #
|
| 445 |
|
| 446 | def expr_stmt(self, nodelist):
|
| 447 | # augassign testlist | testlist ('=' testlist)*
|
| 448 | en = nodelist[-1]
|
| 449 | exprNode = self.lookup_node(en)(en[1:])
|
| 450 | if len(nodelist) == 1:
|
| 451 | return Discard(exprNode, lineno=exprNode.lineno)
|
| 452 | if nodelist[1][0] == token.EQUAL:
|
| 453 | nodesl = []
|
| 454 | for i in range(0, len(nodelist) - 2, 2):
|
| 455 | nodesl.append(self.com_assign(nodelist[i], OP_ASSIGN))
|
| 456 | return Assign(nodesl, exprNode, lineno=nodelist[1][2])
|
| 457 | else:
|
| 458 | lval = self.com_augassign(nodelist[0])
|
| 459 | op = self.com_augassign_op(nodelist[1])
|
| 460 | return AugAssign(lval, op[1], exprNode, lineno=op[2])
|
| 461 | raise WalkerError, "can't get here"
|
| 462 |
|
| 463 | def print_stmt(self, nodelist):
|
| 464 | # print ([ test (',' test)* [','] ] | '>>' test [ (',' test)+ [','] ])
|
| 465 | items = []
|
| 466 | if len(nodelist) == 1:
|
| 467 | start = 1
|
| 468 | dest = None
|
| 469 | elif nodelist[1][0] == token.RIGHTSHIFT:
|
| 470 | assert len(nodelist) == 3 \
|
| 471 | or nodelist[3][0] == token.COMMA
|
| 472 | dest = self.com_node(nodelist[2])
|
| 473 | start = 4
|
| 474 | else:
|
| 475 | dest = None
|
| 476 | start = 1
|
| 477 | for i in range(start, len(nodelist), 2):
|
| 478 | items.append(self.com_node(nodelist[i]))
|
| 479 | if nodelist[-1][0] == token.COMMA:
|
| 480 | return Print(items, dest, lineno=nodelist[0][2])
|
| 481 | return Printnl(items, dest, lineno=nodelist[0][2])
|
| 482 |
|
| 483 | def del_stmt(self, nodelist):
|
| 484 | return self.com_assign(nodelist[1], OP_DELETE)
|
| 485 |
|
| 486 | def pass_stmt(self, nodelist):
|
| 487 | return Pass(lineno=nodelist[0][2])
|
| 488 |
|
| 489 | def break_stmt(self, nodelist):
|
| 490 | return Break(lineno=nodelist[0][2])
|
| 491 |
|
| 492 | def continue_stmt(self, nodelist):
|
| 493 | return Continue(lineno=nodelist[0][2])
|
| 494 |
|
| 495 | def return_stmt(self, nodelist):
|
| 496 | # return: [testlist]
|
| 497 | if len(nodelist) < 2:
|
| 498 | return Return(Const(None), lineno=nodelist[0][2])
|
| 499 | return Return(self.com_node(nodelist[1]), lineno=nodelist[0][2])
|
| 500 |
|
| 501 | def yield_stmt(self, nodelist):
|
| 502 | expr = self.com_node(nodelist[0])
|
| 503 | return Discard(expr, lineno=expr.lineno)
|
| 504 |
|
| 505 | def yield_expr(self, nodelist):
|
| 506 | if len(nodelist) > 1:
|
| 507 | value = self.com_node(nodelist[1])
|
| 508 | else:
|
| 509 | value = Const(None)
|
| 510 | return Yield(value, lineno=nodelist[0][2])
|
| 511 |
|
| 512 | def raise_stmt(self, nodelist):
|
| 513 | # raise: [test [',' test [',' test]]]
|
| 514 | if len(nodelist) > 5:
|
| 515 | expr3 = self.com_node(nodelist[5])
|
| 516 | else:
|
| 517 | expr3 = None
|
| 518 | if len(nodelist) > 3:
|
| 519 | expr2 = self.com_node(nodelist[3])
|
| 520 | else:
|
| 521 | expr2 = None
|
| 522 | if len(nodelist) > 1:
|
| 523 | expr1 = self.com_node(nodelist[1])
|
| 524 | else:
|
| 525 | expr1 = None
|
| 526 | return Raise(expr1, expr2, expr3, lineno=nodelist[0][2])
|
| 527 |
|
| 528 | def import_stmt(self, nodelist):
|
| 529 | # import_stmt: import_name | import_from
|
| 530 | assert len(nodelist) == 1
|
| 531 | return self.com_node(nodelist[0])
|
| 532 |
|
| 533 | def import_name(self, nodelist):
|
| 534 | # import_name: 'import' dotted_as_names
|
| 535 | return Import(self.com_dotted_as_names(nodelist[1]),
|
| 536 | lineno=nodelist[0][2])
|
| 537 |
|
| 538 | def import_from(self, nodelist):
|
| 539 | # import_from: 'from' ('.'* dotted_name | '.') 'import' ('*' |
|
| 540 | # '(' import_as_names ')' | import_as_names)
|
| 541 | assert nodelist[0][1] == 'from'
|
| 542 | idx = 1
|
| 543 | while nodelist[idx][1] == '.':
|
| 544 | idx += 1
|
| 545 | level = idx - 1
|
| 546 | if nodelist[idx][0] == symbol.dotted_name:
|
| 547 | fromname = self.com_dotted_name(nodelist[idx])
|
| 548 | idx += 1
|
| 549 | else:
|
| 550 | fromname = ""
|
| 551 | assert nodelist[idx][1] == 'import'
|
| 552 | if nodelist[idx + 1][0] == token.STAR:
|
| 553 | return From(fromname, [('*', None)], level,
|
| 554 | lineno=nodelist[0][2])
|
| 555 | else:
|
| 556 | node = nodelist[idx + 1 + (nodelist[idx + 1][0] == token.LPAR)]
|
| 557 | return From(fromname, self.com_import_as_names(node), level,
|
| 558 | lineno=nodelist[0][2])
|
| 559 |
|
| 560 | def global_stmt(self, nodelist):
|
| 561 | # global: NAME (',' NAME)*
|
| 562 | names = []
|
| 563 | for i in range(1, len(nodelist), 2):
|
| 564 | names.append(nodelist[i][1])
|
| 565 | return Global(names, lineno=nodelist[0][2])
|
| 566 |
|
| 567 | def exec_stmt(self, nodelist):
|
| 568 | # exec_stmt: 'exec' expr ['in' expr [',' expr]]
|
| 569 | expr1 = self.com_node(nodelist[1])
|
| 570 | if len(nodelist) >= 4:
|
| 571 | expr2 = self.com_node(nodelist[3])
|
| 572 | if len(nodelist) >= 6:
|
| 573 | expr3 = self.com_node(nodelist[5])
|
| 574 | else:
|
| 575 | expr3 = None
|
| 576 | else:
|
| 577 | expr2 = expr3 = None
|
| 578 |
|
| 579 | return Exec(expr1, expr2, expr3, lineno=nodelist[0][2])
|
| 580 |
|
| 581 | def assert_stmt(self, nodelist):
|
| 582 | # 'assert': test, [',' test]
|
| 583 | expr1 = self.com_node(nodelist[1])
|
| 584 | if (len(nodelist) == 4):
|
| 585 | expr2 = self.com_node(nodelist[3])
|
| 586 | else:
|
| 587 | expr2 = None
|
| 588 | return Assert(expr1, expr2, lineno=nodelist[0][2])
|
| 589 |
|
| 590 | def if_stmt(self, nodelist):
|
| 591 | # if: test ':' suite ('elif' test ':' suite)* ['else' ':' suite]
|
| 592 | tests = []
|
| 593 | for i in range(0, len(nodelist) - 3, 4):
|
| 594 | testNode = self.com_node(nodelist[i + 1])
|
| 595 | suiteNode = self.com_node(nodelist[i + 3])
|
| 596 | tests.append((testNode, suiteNode))
|
| 597 |
|
| 598 | if len(nodelist) % 4 == 3:
|
| 599 | elseNode = self.com_node(nodelist[-1])
|
| 600 | ## elseNode.lineno = nodelist[-1][1][2]
|
| 601 | else:
|
| 602 | elseNode = None
|
| 603 | return If(tests, elseNode, lineno=nodelist[0][2])
|
| 604 |
|
| 605 | def while_stmt(self, nodelist):
|
| 606 | # 'while' test ':' suite ['else' ':' suite]
|
| 607 |
|
| 608 | testNode = self.com_node(nodelist[1])
|
| 609 | bodyNode = self.com_node(nodelist[3])
|
| 610 |
|
| 611 | if len(nodelist) > 4:
|
| 612 | elseNode = self.com_node(nodelist[6])
|
| 613 | else:
|
| 614 | elseNode = None
|
| 615 |
|
| 616 | return While(testNode, bodyNode, elseNode, lineno=nodelist[0][2])
|
| 617 |
|
| 618 | def for_stmt(self, nodelist):
|
| 619 | # 'for' exprlist 'in' exprlist ':' suite ['else' ':' suite]
|
| 620 |
|
| 621 | assignNode = self.com_assign(nodelist[1], OP_ASSIGN)
|
| 622 | listNode = self.com_node(nodelist[3])
|
| 623 | bodyNode = self.com_node(nodelist[5])
|
| 624 |
|
| 625 | if len(nodelist) > 8:
|
| 626 | elseNode = self.com_node(nodelist[8])
|
| 627 | else:
|
| 628 | elseNode = None
|
| 629 |
|
| 630 | return For(assignNode, listNode, bodyNode, elseNode,
|
| 631 | lineno=nodelist[0][2])
|
| 632 |
|
| 633 | def try_stmt(self, nodelist):
|
| 634 | return self.com_try_except_finally(nodelist)
|
| 635 |
|
| 636 | def with_stmt(self, nodelist):
|
| 637 | return self.com_with(nodelist)
|
| 638 |
|
| 639 | def with_var(self, nodelist):
|
| 640 | return self.com_with_var(nodelist)
|
| 641 |
|
| 642 | def suite(self, nodelist):
|
| 643 | # simple_stmt | NEWLINE INDENT NEWLINE* (stmt NEWLINE*)+ DEDENT
|
| 644 | if len(nodelist) == 1:
|
| 645 | return self.com_stmt(nodelist[0])
|
| 646 |
|
| 647 | stmts = []
|
| 648 | for node in nodelist:
|
| 649 | if node[0] == symbol.stmt:
|
| 650 | self.com_append_stmt(stmts, node)
|
| 651 | return Stmt(stmts)
|
| 652 |
|
| 653 | # --------------------------------------------------------------
|
| 654 | #
|
| 655 | # EXPRESSION NODES (invoked by com_node())
|
| 656 | #
|
| 657 |
|
| 658 | def testlist(self, nodelist):
|
| 659 | # testlist: expr (',' expr)* [',']
|
| 660 | # testlist_safe: test [(',' test)+ [',']]
|
| 661 | # exprlist: expr (',' expr)* [',']
|
| 662 | return self.com_binary(Tuple, nodelist)
|
| 663 |
|
| 664 | testlist_safe = testlist # XXX
|
| 665 | testlist1 = testlist
|
| 666 | exprlist = testlist
|
| 667 |
|
| 668 | def testlist_comp(self, nodelist):
|
| 669 | # test ( comp_for | (',' test)* [','] )
|
| 670 | assert nodelist[0][0] == symbol.test
|
| 671 | if len(nodelist) == 2 and nodelist[1][0] == symbol.comp_for:
|
| 672 | test = self.com_node(nodelist[0])
|
| 673 | return self.com_generator_expression(test, nodelist[1])
|
| 674 | return self.testlist(nodelist)
|
| 675 |
|
| 676 | def test(self, nodelist):
|
| 677 | # or_test ['if' or_test 'else' test] | lambdef
|
| 678 | if len(nodelist) == 1 and nodelist[0][0] == symbol.lambdef:
|
| 679 | return self.lambdef(nodelist[0])
|
| 680 | then = self.com_node(nodelist[0])
|
| 681 | if len(nodelist) > 1:
|
| 682 | assert len(nodelist) == 5
|
| 683 | assert nodelist[1][1] == 'if'
|
| 684 | assert nodelist[3][1] == 'else'
|
| 685 | test = self.com_node(nodelist[2])
|
| 686 | else_ = self.com_node(nodelist[4])
|
| 687 | return IfExp(test, then, else_, lineno=nodelist[1][2])
|
| 688 | return then
|
| 689 |
|
| 690 | def or_test(self, nodelist):
|
| 691 | # and_test ('or' and_test)* | lambdef
|
| 692 | if len(nodelist) == 1 and nodelist[0][0] == symbol.lambdef:
|
| 693 | return self.lambdef(nodelist[0])
|
| 694 | return self.com_binary(Or, nodelist)
|
| 695 | old_test = or_test
|
| 696 |
|
| 697 | def and_test(self, nodelist):
|
| 698 | # not_test ('and' not_test)*
|
| 699 | return self.com_binary(And, nodelist)
|
| 700 |
|
| 701 | def not_test(self, nodelist):
|
| 702 | # 'not' not_test | comparison
|
| 703 | result = self.com_node(nodelist[-1])
|
| 704 | if len(nodelist) == 2:
|
| 705 | return Not(result, lineno=nodelist[0][2])
|
| 706 | return result
|
| 707 |
|
| 708 | def comparison(self, nodelist):
|
| 709 | # comparison: expr (comp_op expr)*
|
| 710 | node = self.com_node(nodelist[0])
|
| 711 | if len(nodelist) == 1:
|
| 712 | return node
|
| 713 |
|
| 714 | results = []
|
| 715 | for i in range(2, len(nodelist), 2):
|
| 716 | nl = nodelist[i-1]
|
| 717 |
|
| 718 | # comp_op: '<' | '>' | '=' | '>=' | '<=' | '<>' | '!=' | '=='
|
| 719 | # | 'in' | 'not' 'in' | 'is' | 'is' 'not'
|
| 720 | n = nl[1]
|
| 721 | if n[0] == token.NAME:
|
| 722 | type = n[1]
|
| 723 | if len(nl) == 3:
|
| 724 | if type == 'not':
|
| 725 | type = 'not in'
|
| 726 | else:
|
| 727 | type = 'is not'
|
| 728 | else:
|
| 729 | type = _cmp_types[n[0]]
|
| 730 |
|
| 731 | lineno = nl[1][2]
|
| 732 | results.append((type, self.com_node(nodelist[i])))
|
| 733 |
|
| 734 | # we need a special "compare" node so that we can distinguish
|
| 735 | # 3 < x < 5 from (3 < x) < 5
|
| 736 | # the two have very different semantics and results (note that the
|
| 737 | # latter form is always true)
|
| 738 |
|
| 739 | return Compare(node, results, lineno=lineno)
|
| 740 |
|
| 741 | def expr(self, nodelist):
|
| 742 | # xor_expr ('|' xor_expr)*
|
| 743 | return self.com_binary(Bitor, nodelist)
|
| 744 |
|
| 745 | def xor_expr(self, nodelist):
|
| 746 | # xor_expr ('^' xor_expr)*
|
| 747 | return self.com_binary(Bitxor, nodelist)
|
| 748 |
|
| 749 | def and_expr(self, nodelist):
|
| 750 | # xor_expr ('&' xor_expr)*
|
| 751 | return self.com_binary(Bitand, nodelist)
|
| 752 |
|
| 753 | def shift_expr(self, nodelist):
|
| 754 | # shift_expr ('<<'|'>>' shift_expr)*
|
| 755 | node = self.com_node(nodelist[0])
|
| 756 | for i in range(2, len(nodelist), 2):
|
| 757 | right = self.com_node(nodelist[i])
|
| 758 | if nodelist[i-1][0] == token.LEFTSHIFT:
|
| 759 | node = LeftShift([node, right], lineno=nodelist[1][2])
|
| 760 | elif nodelist[i-1][0] == token.RIGHTSHIFT:
|
| 761 | node = RightShift([node, right], lineno=nodelist[1][2])
|
| 762 | else:
|
| 763 | raise ValueError, "unexpected token: %s" % nodelist[i-1][0]
|
| 764 | return node
|
| 765 |
|
| 766 | def arith_expr(self, nodelist):
|
| 767 | node = self.com_node(nodelist[0])
|
| 768 | for i in range(2, len(nodelist), 2):
|
| 769 | right = self.com_node(nodelist[i])
|
| 770 | if nodelist[i-1][0] == token.PLUS:
|
| 771 | node = Add([node, right], lineno=nodelist[1][2])
|
| 772 | elif nodelist[i-1][0] == token.MINUS:
|
| 773 | node = Sub([node, right], lineno=nodelist[1][2])
|
| 774 | else:
|
| 775 | raise ValueError, "unexpected token: %s" % nodelist[i-1][0]
|
| 776 | return node
|
| 777 |
|
| 778 | def term(self, nodelist):
|
| 779 | node = self.com_node(nodelist[0])
|
| 780 | for i in range(2, len(nodelist), 2):
|
| 781 | right = self.com_node(nodelist[i])
|
| 782 | t = nodelist[i-1][0]
|
| 783 | if t == token.STAR:
|
| 784 | node = Mul([node, right])
|
| 785 | elif t == token.SLASH:
|
| 786 | node = Div([node, right])
|
| 787 | elif t == token.PERCENT:
|
| 788 | node = Mod([node, right])
|
| 789 | elif t == token.DOUBLESLASH:
|
| 790 | node = FloorDiv([node, right])
|
| 791 | else:
|
| 792 | raise ValueError, "unexpected token: %s" % t
|
| 793 | node.lineno = nodelist[1][2]
|
| 794 | return node
|
| 795 |
|
| 796 | def factor(self, nodelist):
|
| 797 | elt = nodelist[0]
|
| 798 | t = elt[0]
|
| 799 | node = self.lookup_node(nodelist[-1])(nodelist[-1][1:])
|
| 800 | # need to handle (unary op)constant here...
|
| 801 | if t == token.PLUS:
|
| 802 | return UnaryAdd(node, lineno=elt[2])
|
| 803 | elif t == token.MINUS:
|
| 804 | return UnarySub(node, lineno=elt[2])
|
| 805 | elif t == token.TILDE:
|
| 806 | node = Invert(node, lineno=elt[2])
|
| 807 | return node
|
| 808 |
|
| 809 | def power(self, nodelist):
|
| 810 | # power: atom trailer* ['**' factor]
|
| 811 | node = self.com_node(nodelist[0])
|
| 812 | for i in range(1, len(nodelist)):
|
| 813 | elt = nodelist[i]
|
| 814 | if elt[0] == token.DOUBLESTAR:
|
| 815 | return Power([node, self.com_node(nodelist[i+1])],
|
| 816 | lineno=elt[2])
|
| 817 |
|
| 818 | node = self.com_apply_trailer(node, elt)
|
| 819 |
|
| 820 | return node
|
| 821 |
|
| 822 | def atom(self, nodelist):
|
| 823 | return self._atom_dispatch[nodelist[0][0]](nodelist)
|
| 824 |
|
| 825 | def atom_lpar(self, nodelist):
|
| 826 | if nodelist[1][0] == token.RPAR:
|
| 827 | return Tuple((), lineno=nodelist[0][2])
|
| 828 | return self.com_node(nodelist[1])
|
| 829 |
|
| 830 | def atom_lsqb(self, nodelist):
|
| 831 | if nodelist[1][0] == token.RSQB:
|
| 832 | return List((), lineno=nodelist[0][2])
|
| 833 | return self.com_list_constructor(nodelist[1])
|
| 834 |
|
| 835 | def atom_lbrace(self, nodelist):
|
| 836 | if nodelist[1][0] == token.RBRACE:
|
| 837 | return Dict((), lineno=nodelist[0][2])
|
| 838 | return self.com_dictorsetmaker(nodelist[1])
|
| 839 |
|
| 840 | def atom_backquote(self, nodelist):
|
| 841 | return Backquote(self.com_node(nodelist[1]))
|
| 842 |
|
| 843 | def atom_number(self, nodelist):
|
| 844 | ### need to verify this matches compile.c
|
| 845 | k = eval(nodelist[0][1])
|
| 846 | return Const(k, lineno=nodelist[0][2])
|
| 847 |
|
| 848 | def decode_literal(self, lit):
|
| 849 | if self.encoding:
|
| 850 | # this is particularly fragile & a bit of a
|
| 851 | # hack... changes in compile.c:parsestr and
|
| 852 | # tokenizer.c must be reflected here.
|
| 853 | if self.encoding not in ['utf-8', 'iso-8859-1']:
|
| 854 | lit = unicode(lit, 'utf-8').encode(self.encoding)
|
| 855 | return eval("# coding: %s\n%s" % (self.encoding, lit))
|
| 856 | else:
|
| 857 | return eval(lit)
|
| 858 |
|
| 859 | def atom_string(self, nodelist):
|
| 860 | k = ''
|
| 861 | for node in nodelist:
|
| 862 | k += self.decode_literal(node[1])
|
| 863 | return Const(k, lineno=nodelist[0][2])
|
| 864 |
|
| 865 | def atom_name(self, nodelist):
|
| 866 | return Name(nodelist[0][1], lineno=nodelist[0][2])
|
| 867 |
|
| 868 | # --------------------------------------------------------------
|
| 869 | #
|
| 870 | # INTERNAL PARSING UTILITIES
|
| 871 | #
|
| 872 |
|
| 873 | # The use of com_node() introduces a lot of extra stack frames,
|
| 874 | # enough to cause a stack overflow compiling test.test_parser with
|
| 875 | # the standard interpreter recursionlimit. The com_node() is a
|
| 876 | # convenience function that hides the dispatch details, but comes
|
| 877 | # at a very high cost. It is more efficient to dispatch directly
|
| 878 | # in the callers. In these cases, use lookup_node() and call the
|
| 879 | # dispatched node directly.
|
| 880 |
|
| 881 | def lookup_node(self, node):
|
| 882 | return self._dispatch[node[0]]
|
| 883 |
|
| 884 | def com_node(self, node):
|
| 885 | # Note: compile.c has handling in com_node for del_stmt, pass_stmt,
|
| 886 | # break_stmt, stmt, small_stmt, flow_stmt, simple_stmt,
|
| 887 | # and compound_stmt.
|
| 888 | # We'll just dispatch them.
|
| 889 | return self._dispatch[node[0]](node[1:])
|
| 890 |
|
| 891 | def com_NEWLINE(self, *args):
|
| 892 | # A ';' at the end of a line can make a NEWLINE token appear
|
| 893 | # here, Render it harmless. (genc discards ('discard',
|
| 894 | # ('const', xxxx)) Nodes)
|
| 895 | return Discard(Const(None))
|
| 896 |
|
| 897 | def com_arglist(self, nodelist):
|
| 898 | # varargslist:
|
| 899 | # (fpdef ['=' test] ',')* ('*' NAME [',' '**' NAME] | '**' NAME)
|
| 900 | # | fpdef ['=' test] (',' fpdef ['=' test])* [',']
|
| 901 | # fpdef: NAME | '(' fplist ')'
|
| 902 | # fplist: fpdef (',' fpdef)* [',']
|
| 903 | names = []
|
| 904 | defaults = []
|
| 905 | flags = 0
|
| 906 |
|
| 907 | i = 0
|
| 908 | while i < len(nodelist):
|
| 909 | node = nodelist[i]
|
| 910 | if node[0] == token.STAR or node[0] == token.DOUBLESTAR:
|
| 911 | if node[0] == token.STAR:
|
| 912 | node = nodelist[i+1]
|
| 913 | if node[0] == token.NAME:
|
| 914 | names.append(node[1])
|
| 915 | flags = flags | CO_VARARGS
|
| 916 | i = i + 3
|
| 917 |
|
| 918 | if i < len(nodelist):
|
| 919 | # should be DOUBLESTAR
|
| 920 | t = nodelist[i][0]
|
| 921 | if t == token.DOUBLESTAR:
|
| 922 | node = nodelist[i+1]
|
| 923 | else:
|
| 924 | raise ValueError, "unexpected token: %s" % t
|
| 925 | names.append(node[1])
|
| 926 | flags = flags | CO_VARKEYWORDS
|
| 927 |
|
| 928 | break
|
| 929 |
|
| 930 | # fpdef: NAME | '(' fplist ')'
|
| 931 | names.append(self.com_fpdef(node))
|
| 932 |
|
| 933 | i = i + 1
|
| 934 | if i < len(nodelist) and nodelist[i][0] == token.EQUAL:
|
| 935 | defaults.append(self.com_node(nodelist[i + 1]))
|
| 936 | i = i + 2
|
| 937 | elif len(defaults):
|
| 938 | # we have already seen an argument with default, but here
|
| 939 | # came one without
|
| 940 | raise SyntaxError, "non-default argument follows default argument"
|
| 941 |
|
| 942 | # skip the comma
|
| 943 | i = i + 1
|
| 944 |
|
| 945 | return names, defaults, flags
|
| 946 |
|
| 947 | def com_fpdef(self, node):
|
| 948 | # fpdef: NAME | '(' fplist ')'
|
| 949 | if node[1][0] == token.LPAR:
|
| 950 | return self.com_fplist(node[2])
|
| 951 | return node[1][1]
|
| 952 |
|
| 953 | def com_fplist(self, node):
|
| 954 | # fplist: fpdef (',' fpdef)* [',']
|
| 955 | if len(node) == 2:
|
| 956 | return self.com_fpdef(node[1])
|
| 957 | list = []
|
| 958 | for i in range(1, len(node), 2):
|
| 959 | list.append(self.com_fpdef(node[i]))
|
| 960 | return tuple(list)
|
| 961 |
|
| 962 | def com_dotted_name(self, node):
|
| 963 | # String together the dotted names and return the string
|
| 964 | name = ""
|
| 965 | for n in node:
|
| 966 | if type(n) == type(()) and n[0] == 1:
|
| 967 | name = name + n[1] + '.'
|
| 968 | return name[:-1]
|
| 969 |
|
| 970 | def com_dotted_as_name(self, node):
|
| 971 | assert node[0] == symbol.dotted_as_name
|
| 972 | node = node[1:]
|
| 973 | dot = self.com_dotted_name(node[0][1:])
|
| 974 | if len(node) == 1:
|
| 975 | return dot, None
|
| 976 | assert node[1][1] == 'as'
|
| 977 | assert node[2][0] == token.NAME
|
| 978 | return dot, node[2][1]
|
| 979 |
|
| 980 | def com_dotted_as_names(self, node):
|
| 981 | assert node[0] == symbol.dotted_as_names
|
| 982 | node = node[1:]
|
| 983 | names = [self.com_dotted_as_name(node[0])]
|
| 984 | for i in range(2, len(node), 2):
|
| 985 | names.append(self.com_dotted_as_name(node[i]))
|
| 986 | return names
|
| 987 |
|
| 988 | def com_import_as_name(self, node):
|
| 989 | assert node[0] == symbol.import_as_name
|
| 990 | node = node[1:]
|
| 991 | assert node[0][0] == token.NAME
|
| 992 | if len(node) == 1:
|
| 993 | return node[0][1], None
|
| 994 | assert node[1][1] == 'as', node
|
| 995 | assert node[2][0] == token.NAME
|
| 996 | return node[0][1], node[2][1]
|
| 997 |
|
| 998 | def com_import_as_names(self, node):
|
| 999 | assert node[0] == symbol.import_as_names
|
| 1000 | node = node[1:]
|
| 1001 | names = [self.com_import_as_name(node[0])]
|
| 1002 | for i in range(2, len(node), 2):
|
| 1003 | names.append(self.com_import_as_name(node[i]))
|
| 1004 | return names
|
| 1005 |
|
| 1006 | def com_bases(self, node):
|
| 1007 | bases = []
|
| 1008 | for i in range(1, len(node), 2):
|
| 1009 | bases.append(self.com_node(node[i]))
|
| 1010 | return bases
|
| 1011 |
|
| 1012 | def com_try_except_finally(self, nodelist):
|
| 1013 | # ('try' ':' suite
|
| 1014 | # ((except_clause ':' suite)+ ['else' ':' suite] ['finally' ':' suite]
|
| 1015 | # | 'finally' ':' suite))
|
| 1016 |
|
| 1017 | if nodelist[3][0] == token.NAME:
|
| 1018 | # first clause is a finally clause: only try-finally
|
| 1019 | return TryFinally(self.com_node(nodelist[2]),
|
| 1020 | self.com_node(nodelist[5]),
|
| 1021 | lineno=nodelist[0][2])
|
| 1022 |
|
| 1023 | #tryexcept: [TryNode, [except_clauses], elseNode)]
|
| 1024 | clauses = []
|
| 1025 | elseNode = None
|
| 1026 | finallyNode = None
|
| 1027 | for i in range(3, len(nodelist), 3):
|
| 1028 | node = nodelist[i]
|
| 1029 | if node[0] == symbol.except_clause:
|
| 1030 | # except_clause: 'except' [expr [(',' | 'as') expr]] */
|
| 1031 | if len(node) > 2:
|
| 1032 | expr1 = self.com_node(node[2])
|
| 1033 | if len(node) > 4:
|
| 1034 | expr2 = self.com_assign(node[4], OP_ASSIGN)
|
| 1035 | else:
|
| 1036 | expr2 = None
|
| 1037 | else:
|
| 1038 | expr1 = expr2 = None
|
| 1039 | clauses.append((expr1, expr2, self.com_node(nodelist[i+2])))
|
| 1040 |
|
| 1041 | if node[0] == token.NAME:
|
| 1042 | if node[1] == 'else':
|
| 1043 | elseNode = self.com_node(nodelist[i+2])
|
| 1044 | elif node[1] == 'finally':
|
| 1045 | finallyNode = self.com_node(nodelist[i+2])
|
| 1046 | try_except = TryExcept(self.com_node(nodelist[2]), clauses, elseNode,
|
| 1047 | lineno=nodelist[0][2])
|
| 1048 | if finallyNode:
|
| 1049 | return TryFinally(try_except, finallyNode, lineno=nodelist[0][2])
|
| 1050 | else:
|
| 1051 | return try_except
|
| 1052 |
|
| 1053 | def com_with(self, nodelist):
|
| 1054 | # with_stmt: 'with' with_item (',' with_item)* ':' suite
|
| 1055 | body = self.com_node(nodelist[-1])
|
| 1056 | for i in range(len(nodelist) - 3, 0, -2):
|
| 1057 | ret = self.com_with_item(nodelist[i], body, nodelist[0][2])
|
| 1058 | if i == 1:
|
| 1059 | return ret
|
| 1060 | body = ret
|
| 1061 |
|
| 1062 | def com_with_item(self, nodelist, body, lineno):
|
| 1063 | # with_item: test ['as' expr]
|
| 1064 | if len(nodelist) == 4:
|
| 1065 | var = self.com_assign(nodelist[3], OP_ASSIGN)
|
| 1066 | else:
|
| 1067 | var = None
|
| 1068 | expr = self.com_node(nodelist[1])
|
| 1069 | return With(expr, var, body, lineno=lineno)
|
| 1070 |
|
| 1071 | def com_augassign_op(self, node):
|
| 1072 | assert node[0] == symbol.augassign
|
| 1073 | return node[1]
|
| 1074 |
|
| 1075 | def com_augassign(self, node):
|
| 1076 | """Return node suitable for lvalue of augmented assignment
|
| 1077 |
|
| 1078 | Names, slices, and attributes are the only allowable nodes.
|
| 1079 | """
|
| 1080 | l = self.com_node(node)
|
| 1081 | if l.__class__ in (Name, Slice, Subscript, Getattr):
|
| 1082 | return l
|
| 1083 | raise SyntaxError, "can't assign to %s" % l.__class__.__name__
|
| 1084 |
|
| 1085 | def com_assign(self, node, assigning):
|
| 1086 | # return a node suitable for use as an "lvalue"
|
| 1087 | # loop to avoid trivial recursion
|
| 1088 | while 1:
|
| 1089 | t = node[0]
|
| 1090 | if t in (symbol.exprlist, symbol.testlist, symbol.testlist_safe, symbol.testlist_comp):
|
| 1091 | if len(node) > 2:
|
| 1092 | return self.com_assign_tuple(node, assigning)
|
| 1093 | node = node[1]
|
| 1094 | elif t in _assign_types:
|
| 1095 | if len(node) > 2:
|
| 1096 | raise SyntaxError, "can't assign to operator"
|
| 1097 | node = node[1]
|
| 1098 | elif t == symbol.power:
|
| 1099 | if node[1][0] != symbol.atom:
|
| 1100 | raise SyntaxError, "can't assign to operator"
|
| 1101 | if len(node) > 2:
|
| 1102 | primary = self.com_node(node[1])
|
| 1103 | for i in range(2, len(node)-1):
|
| 1104 | ch = node[i]
|
| 1105 | if ch[0] == token.DOUBLESTAR:
|
| 1106 | raise SyntaxError, "can't assign to operator"
|
| 1107 | primary = self.com_apply_trailer(primary, ch)
|
| 1108 | return self.com_assign_trailer(primary, node[-1],
|
| 1109 | assigning)
|
| 1110 | node = node[1]
|
| 1111 | elif t == symbol.atom:
|
| 1112 | t = node[1][0]
|
| 1113 | if t == token.LPAR:
|
| 1114 | node = node[2]
|
| 1115 | if node[0] == token.RPAR:
|
| 1116 | raise SyntaxError, "can't assign to ()"
|
| 1117 | elif t == token.LSQB:
|
| 1118 | node = node[2]
|
| 1119 | if node[0] == token.RSQB:
|
| 1120 | raise SyntaxError, "can't assign to []"
|
| 1121 | return self.com_assign_list(node, assigning)
|
| 1122 | elif t == token.NAME:
|
| 1123 | return self.com_assign_name(node[1], assigning)
|
| 1124 | else:
|
| 1125 | raise SyntaxError, "can't assign to literal"
|
| 1126 | else:
|
| 1127 | raise SyntaxError, "bad assignment (%s)" % t
|
| 1128 |
|
| 1129 | def com_assign_tuple(self, node, assigning):
|
| 1130 | assigns = []
|
| 1131 | for i in range(1, len(node), 2):
|
| 1132 | assigns.append(self.com_assign(node[i], assigning))
|
| 1133 | return AssTuple(assigns, lineno=extractLineNo(node))
|
| 1134 |
|
| 1135 | def com_assign_list(self, node, assigning):
|
| 1136 | assigns = []
|
| 1137 | for i in range(1, len(node), 2):
|
| 1138 | if i + 1 < len(node):
|
| 1139 | if node[i + 1][0] == symbol.list_for:
|
| 1140 | raise SyntaxError, "can't assign to list comprehension"
|
| 1141 | assert node[i + 1][0] == token.COMMA, node[i + 1]
|
| 1142 | assigns.append(self.com_assign(node[i], assigning))
|
| 1143 | return AssList(assigns, lineno=extractLineNo(node))
|
| 1144 |
|
| 1145 | def com_assign_name(self, node, assigning):
|
| 1146 | return AssName(node[1], assigning, lineno=node[2])
|
| 1147 |
|
| 1148 | def com_assign_trailer(self, primary, node, assigning):
|
| 1149 | t = node[1][0]
|
| 1150 | if t == token.DOT:
|
| 1151 | return self.com_assign_attr(primary, node[2], assigning)
|
| 1152 | if t == token.LSQB:
|
| 1153 | return self.com_subscriptlist(primary, node[2], assigning)
|
| 1154 | if t == token.LPAR:
|
| 1155 | raise SyntaxError, "can't assign to function call"
|
| 1156 | raise SyntaxError, "unknown trailer type: %s" % t
|
| 1157 |
|
| 1158 | def com_assign_attr(self, primary, node, assigning):
|
| 1159 | return AssAttr(primary, node[1], assigning, lineno=node[-1])
|
| 1160 |
|
| 1161 | def com_binary(self, constructor, nodelist):
|
| 1162 | "Compile 'NODE (OP NODE)*' into (type, [ node1, ..., nodeN ])."
|
| 1163 | l = len(nodelist)
|
| 1164 | if l == 1:
|
| 1165 | n = nodelist[0]
|
| 1166 | return self.lookup_node(n)(n[1:])
|
| 1167 | items = []
|
| 1168 | for i in range(0, l, 2):
|
| 1169 | n = nodelist[i]
|
| 1170 | items.append(self.lookup_node(n)(n[1:]))
|
| 1171 | return constructor(items, lineno=extractLineNo(nodelist))
|
| 1172 |
|
| 1173 | def com_stmt(self, node):
|
| 1174 | result = self.lookup_node(node)(node[1:])
|
| 1175 | assert result is not None
|
| 1176 | if isinstance(result, Stmt):
|
| 1177 | return result
|
| 1178 | return Stmt([result])
|
| 1179 |
|
| 1180 | def com_append_stmt(self, stmts, node):
|
| 1181 | result = self.lookup_node(node)(node[1:])
|
| 1182 | assert result is not None
|
| 1183 | if isinstance(result, Stmt):
|
| 1184 | stmts.extend(result.nodes)
|
| 1185 | else:
|
| 1186 | stmts.append(result)
|
| 1187 |
|
| 1188 | def com_list_constructor(self, nodelist):
|
| 1189 | # listmaker: test ( list_for | (',' test)* [','] )
|
| 1190 | values = []
|
| 1191 | for i in range(1, len(nodelist)):
|
| 1192 | if nodelist[i][0] == symbol.list_for:
|
| 1193 | assert len(nodelist[i:]) == 1
|
| 1194 | return self.com_list_comprehension(values[0],
|
| 1195 | nodelist[i])
|
| 1196 | elif nodelist[i][0] == token.COMMA:
|
| 1197 | continue
|
| 1198 | values.append(self.com_node(nodelist[i]))
|
| 1199 | return List(values, lineno=values[0].lineno)
|
| 1200 |
|
| 1201 | def com_list_comprehension(self, expr, node):
|
| 1202 | return self.com_comprehension(expr, None, node, 'list')
|
| 1203 |
|
| 1204 | def com_comprehension(self, expr1, expr2, node, type):
|
| 1205 | # list_iter: list_for | list_if
|
| 1206 | # list_for: 'for' exprlist 'in' testlist [list_iter]
|
| 1207 | # list_if: 'if' test [list_iter]
|
| 1208 |
|
| 1209 | # XXX should raise SyntaxError for assignment
|
| 1210 | # XXX(avassalotti) Set and dict comprehensions should have generator
|
| 1211 | # semantics. In other words, they shouldn't leak
|
| 1212 | # variables outside of the comprehension's scope.
|
| 1213 |
|
| 1214 | lineno = node[1][2]
|
| 1215 | fors = []
|
| 1216 | while node:
|
| 1217 | t = node[1][1]
|
| 1218 | if t == 'for':
|
| 1219 | assignNode = self.com_assign(node[2], OP_ASSIGN)
|
| 1220 | compNode = self.com_node(node[4])
|
| 1221 | newfor = ListCompFor(assignNode, compNode, [])
|
| 1222 | newfor.lineno = node[1][2]
|
| 1223 | fors.append(newfor)
|
| 1224 | if len(node) == 5:
|
| 1225 | node = None
|
| 1226 | elif type == 'list':
|
| 1227 | node = self.com_list_iter(node[5])
|
| 1228 | else:
|
| 1229 | node = self.com_comp_iter(node[5])
|
| 1230 | elif t == 'if':
|
| 1231 | test = self.com_node(node[2])
|
| 1232 | newif = ListCompIf(test, lineno=node[1][2])
|
| 1233 | newfor.ifs.append(newif)
|
| 1234 | if len(node) == 3:
|
| 1235 | node = None
|
| 1236 | elif type == 'list':
|
| 1237 | node = self.com_list_iter(node[3])
|
| 1238 | else:
|
| 1239 | node = self.com_comp_iter(node[3])
|
| 1240 | else:
|
| 1241 | raise SyntaxError, \
|
| 1242 | ("unexpected comprehension element: %s %d"
|
| 1243 | % (node, lineno))
|
| 1244 | if type == 'list':
|
| 1245 | return ListComp(expr1, fors, lineno=lineno)
|
| 1246 | elif type == 'set':
|
| 1247 | return SetComp(expr1, fors, lineno=lineno)
|
| 1248 | elif type == 'dict':
|
| 1249 | return DictComp(expr1, expr2, fors, lineno=lineno)
|
| 1250 | else:
|
| 1251 | raise ValueError("unexpected comprehension type: " + repr(type))
|
| 1252 |
|
| 1253 | def com_list_iter(self, node):
|
| 1254 | assert node[0] == symbol.list_iter
|
| 1255 | return node[1]
|
| 1256 |
|
| 1257 | def com_comp_iter(self, node):
|
| 1258 | assert node[0] == symbol.comp_iter
|
| 1259 | return node[1]
|
| 1260 |
|
| 1261 | def com_generator_expression(self, expr, node):
|
| 1262 | # comp_iter: comp_for | comp_if
|
| 1263 | # comp_for: 'for' exprlist 'in' test [comp_iter]
|
| 1264 | # comp_if: 'if' test [comp_iter]
|
| 1265 |
|
| 1266 | lineno = node[1][2]
|
| 1267 | fors = []
|
| 1268 | while node:
|
| 1269 | t = node[1][1]
|
| 1270 | if t == 'for':
|
| 1271 | assignNode = self.com_assign(node[2], OP_ASSIGN)
|
| 1272 | genNode = self.com_node(node[4])
|
| 1273 | newfor = GenExprFor(assignNode, genNode, [],
|
| 1274 | lineno=node[1][2])
|
| 1275 | fors.append(newfor)
|
| 1276 | if (len(node)) == 5:
|
| 1277 | node = None
|
| 1278 | else:
|
| 1279 | node = self.com_comp_iter(node[5])
|
| 1280 | elif t == 'if':
|
| 1281 | test = self.com_node(node[2])
|
| 1282 | newif = GenExprIf(test, lineno=node[1][2])
|
| 1283 | newfor.ifs.append(newif)
|
| 1284 | if len(node) == 3:
|
| 1285 | node = None
|
| 1286 | else:
|
| 1287 | node = self.com_comp_iter(node[3])
|
| 1288 | else:
|
| 1289 | raise SyntaxError, \
|
| 1290 | ("unexpected generator expression element: %s %d"
|
| 1291 | % (node, lineno))
|
| 1292 | fors[0].is_outmost = True
|
| 1293 | return GenExpr(GenExprInner(expr, fors), lineno=lineno)
|
| 1294 |
|
| 1295 | def com_dictorsetmaker(self, nodelist):
|
| 1296 | # dictorsetmaker: ( (test ':' test (comp_for | (',' test ':' test)* [','])) |
|
| 1297 | # (test (comp_for | (',' test)* [','])) )
|
| 1298 | assert nodelist[0] == symbol.dictorsetmaker
|
| 1299 | nodelist = nodelist[1:]
|
| 1300 | if len(nodelist) == 1 or nodelist[1][0] == token.COMMA:
|
| 1301 | # set literal
|
| 1302 | items = []
|
| 1303 | for i in range(0, len(nodelist), 2):
|
| 1304 | items.append(self.com_node(nodelist[i]))
|
| 1305 | return Set(items, lineno=items[0].lineno)
|
| 1306 | elif nodelist[1][0] == symbol.comp_for:
|
| 1307 | # set comprehension
|
| 1308 | expr = self.com_node(nodelist[0])
|
| 1309 | return self.com_comprehension(expr, None, nodelist[1], 'set')
|
| 1310 | elif len(nodelist) > 3 and nodelist[3][0] == symbol.comp_for:
|
| 1311 | # dict comprehension
|
| 1312 | assert nodelist[1][0] == token.COLON
|
| 1313 | key = self.com_node(nodelist[0])
|
| 1314 | value = self.com_node(nodelist[2])
|
| 1315 | return self.com_comprehension(key, value, nodelist[3], 'dict')
|
| 1316 | else:
|
| 1317 | # dict literal
|
| 1318 | items = []
|
| 1319 | for i in range(0, len(nodelist), 4):
|
| 1320 | items.append((self.com_node(nodelist[i]),
|
| 1321 | self.com_node(nodelist[i+2])))
|
| 1322 | return Dict(items, lineno=items[0][0].lineno)
|
| 1323 |
|
| 1324 | def com_apply_trailer(self, primaryNode, nodelist):
|
| 1325 | t = nodelist[1][0]
|
| 1326 | if t == token.LPAR:
|
| 1327 | return self.com_call_function(primaryNode, nodelist[2])
|
| 1328 | if t == token.DOT:
|
| 1329 | return self.com_select_member(primaryNode, nodelist[2])
|
| 1330 | if t == token.LSQB:
|
| 1331 | return self.com_subscriptlist(primaryNode, nodelist[2], OP_APPLY)
|
| 1332 |
|
| 1333 | raise SyntaxError, 'unknown node type: %s' % t
|
| 1334 |
|
| 1335 | def com_select_member(self, primaryNode, nodelist):
|
| 1336 | if nodelist[0] != token.NAME:
|
| 1337 | raise SyntaxError, "member must be a name"
|
| 1338 | return Getattr(primaryNode, nodelist[1], lineno=nodelist[2])
|
| 1339 |
|
| 1340 | def com_call_function(self, primaryNode, nodelist):
|
| 1341 | if nodelist[0] == token.RPAR:
|
| 1342 | return CallFunc(primaryNode, [], lineno=extractLineNo(nodelist))
|
| 1343 | args = []
|
| 1344 | kw = 0
|
| 1345 | star_node = dstar_node = None
|
| 1346 | len_nodelist = len(nodelist)
|
| 1347 | i = 1
|
| 1348 | while i < len_nodelist:
|
| 1349 | node = nodelist[i]
|
| 1350 |
|
| 1351 | if node[0]==token.STAR:
|
| 1352 | if star_node is not None:
|
| 1353 | raise SyntaxError, 'already have the varargs identifier'
|
| 1354 | star_node = self.com_node(nodelist[i+1])
|
| 1355 | i = i + 3
|
| 1356 | continue
|
| 1357 | elif node[0]==token.DOUBLESTAR:
|
| 1358 | if dstar_node is not None:
|
| 1359 | raise SyntaxError, 'already have the kwargs identifier'
|
| 1360 | dstar_node = self.com_node(nodelist[i+1])
|
| 1361 | i = i + 3
|
| 1362 | continue
|
| 1363 |
|
| 1364 | # positional or named parameters
|
| 1365 | kw, result = self.com_argument(node, kw, star_node)
|
| 1366 |
|
| 1367 | if len_nodelist != 2 and isinstance(result, GenExpr) \
|
| 1368 | and len(node) == 3 and node[2][0] == symbol.comp_for:
|
| 1369 | # allow f(x for x in y), but reject f(x for x in y, 1)
|
| 1370 | # should use f((x for x in y), 1) instead of f(x for x in y, 1)
|
| 1371 | raise SyntaxError, 'generator expression needs parenthesis'
|
| 1372 |
|
| 1373 | args.append(result)
|
| 1374 | i = i + 2
|
| 1375 |
|
| 1376 | return CallFunc(primaryNode, args, star_node, dstar_node,
|
| 1377 | lineno=extractLineNo(nodelist))
|
| 1378 |
|
| 1379 | def com_argument(self, nodelist, kw, star_node):
|
| 1380 | if len(nodelist) == 3 and nodelist[2][0] == symbol.comp_for:
|
| 1381 | test = self.com_node(nodelist[1])
|
| 1382 | return 0, self.com_generator_expression(test, nodelist[2])
|
| 1383 | if len(nodelist) == 2:
|
| 1384 | if kw:
|
| 1385 | raise SyntaxError, "non-keyword arg after keyword arg"
|
| 1386 | if star_node:
|
| 1387 | raise SyntaxError, "only named arguments may follow *expression"
|
| 1388 | return 0, self.com_node(nodelist[1])
|
| 1389 | result = self.com_node(nodelist[3])
|
| 1390 | n = nodelist[1]
|
| 1391 | while len(n) == 2 and n[0] != token.NAME:
|
| 1392 | n = n[1]
|
| 1393 | if n[0] != token.NAME:
|
| 1394 | raise SyntaxError, "keyword can't be an expression (%s)"%n[0]
|
| 1395 | node = Keyword(n[1], result, lineno=n[2])
|
| 1396 | return 1, node
|
| 1397 |
|
| 1398 | def com_subscriptlist(self, primary, nodelist, assigning):
|
| 1399 | # slicing: simple_slicing | extended_slicing
|
| 1400 | # simple_slicing: primary "[" short_slice "]"
|
| 1401 | # extended_slicing: primary "[" slice_list "]"
|
| 1402 | # slice_list: slice_item ("," slice_item)* [","]
|
| 1403 |
|
| 1404 | # backwards compat slice for '[i:j]'
|
| 1405 | if len(nodelist) == 2:
|
| 1406 | sub = nodelist[1]
|
| 1407 | if (sub[1][0] == token.COLON or \
|
| 1408 | (len(sub) > 2 and sub[2][0] == token.COLON)) and \
|
| 1409 | sub[-1][0] != symbol.sliceop:
|
| 1410 | return self.com_slice(primary, sub, assigning)
|
| 1411 |
|
| 1412 | subscripts = []
|
| 1413 | for i in range(1, len(nodelist), 2):
|
| 1414 | subscripts.append(self.com_subscript(nodelist[i]))
|
| 1415 | return Subscript(primary, assigning, subscripts,
|
| 1416 | lineno=extractLineNo(nodelist))
|
| 1417 |
|
| 1418 | def com_subscript(self, node):
|
| 1419 | # slice_item: expression | proper_slice | ellipsis
|
| 1420 | ch = node[1]
|
| 1421 | t = ch[0]
|
| 1422 | if t == token.DOT and node[2][0] == token.DOT:
|
| 1423 | return Ellipsis()
|
| 1424 | if t == token.COLON or len(node) > 2:
|
| 1425 | return self.com_sliceobj(node)
|
| 1426 | return self.com_node(ch)
|
| 1427 |
|
| 1428 | def com_sliceobj(self, node):
|
| 1429 | # proper_slice: short_slice | long_slice
|
| 1430 | # short_slice: [lower_bound] ":" [upper_bound]
|
| 1431 | # long_slice: short_slice ":" [stride]
|
| 1432 | # lower_bound: expression
|
| 1433 | # upper_bound: expression
|
| 1434 | # stride: expression
|
| 1435 | #
|
| 1436 | # Note: a stride may be further slicing...
|
| 1437 |
|
| 1438 | items = []
|
| 1439 |
|
| 1440 | if node[1][0] == token.COLON:
|
| 1441 | items.append(Const(None))
|
| 1442 | i = 2
|
| 1443 | else:
|
| 1444 | items.append(self.com_node(node[1]))
|
| 1445 | # i == 2 is a COLON
|
| 1446 | i = 3
|
| 1447 |
|
| 1448 | if i < len(node) and node[i][0] == symbol.test:
|
| 1449 | items.append(self.com_node(node[i]))
|
| 1450 | i = i + 1
|
| 1451 | else:
|
| 1452 | items.append(Const(None))
|
| 1453 |
|
| 1454 | # a short_slice has been built. look for long_slice now by looking
|
| 1455 | # for strides...
|
| 1456 | for j in range(i, len(node)):
|
| 1457 | ch = node[j]
|
| 1458 | if len(ch) == 2:
|
| 1459 | items.append(Const(None))
|
| 1460 | else:
|
| 1461 | items.append(self.com_node(ch[2]))
|
| 1462 | return Sliceobj(items, lineno=extractLineNo(node))
|
| 1463 |
|
| 1464 | def com_slice(self, primary, node, assigning):
|
| 1465 | # short_slice: [lower_bound] ":" [upper_bound]
|
| 1466 | lower = upper = None
|
| 1467 | if len(node) == 3:
|
| 1468 | if node[1][0] == token.COLON:
|
| 1469 | upper = self.com_node(node[2])
|
| 1470 | else:
|
| 1471 | lower = self.com_node(node[1])
|
| 1472 | elif len(node) == 4:
|
| 1473 | lower = self.com_node(node[1])
|
| 1474 | upper = self.com_node(node[3])
|
| 1475 | return Slice(primary, assigning, lower, upper,
|
| 1476 | lineno=extractLineNo(node))
|
| 1477 |
|
| 1478 | def get_docstring(self, node, n=None):
|
| 1479 | if n is None:
|
| 1480 | n = node[0]
|
| 1481 | node = node[1:]
|
| 1482 | if n == symbol.suite:
|
| 1483 | if len(node) == 1:
|
| 1484 | return self.get_docstring(node[0])
|
| 1485 | for sub in node:
|
| 1486 | if sub[0] == symbol.stmt:
|
| 1487 | return self.get_docstring(sub)
|
| 1488 | return None
|
| 1489 | if n == symbol.file_input:
|
| 1490 | for sub in node:
|
| 1491 | if sub[0] == symbol.stmt:
|
| 1492 | return self.get_docstring(sub)
|
| 1493 | return None
|
| 1494 | if n == symbol.atom:
|
| 1495 | if node[0][0] == token.STRING:
|
| 1496 | s = ''
|
| 1497 | for t in node:
|
| 1498 | s = s + eval(t[1])
|
| 1499 | return s
|
| 1500 | return None
|
| 1501 | if n == symbol.stmt or n == symbol.simple_stmt \
|
| 1502 | or n == symbol.small_stmt:
|
| 1503 | return self.get_docstring(node[0])
|
| 1504 | if n in _doc_nodes and len(node) == 1:
|
| 1505 | return self.get_docstring(node[0])
|
| 1506 | return None
|
| 1507 |
|
| 1508 |
|
| 1509 | def debug_tree(tree):
|
| 1510 | l = []
|
| 1511 | for elt in tree:
|
| 1512 | if isinstance(elt, int):
|
| 1513 | l.append(_names.get(elt, elt))
|
| 1514 | elif isinstance(elt, str):
|
| 1515 | l.append(elt)
|
| 1516 | else:
|
| 1517 | l.append(debug_tree(elt))
|
| 1518 | return l
|