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 compiler.ast import * 29 from pyparser.pygram import syms as symbol, sym_name, tokens as token, tok_name 30 import pyparser.pyparse as parser 31 32 class WalkerError(StandardError): 33 pass 34 35 from compiler.consts import CO_VARARGS, CO_VARKEYWORDS 36 from compiler.consts import OP_ASSIGN, OP_DELETE, OP_APPLY 37 38 def parseFile(path): 39 f = open(path, "U") 40 # XXX The parser API tolerates files without a trailing newline, 41 # but not strings without a trailing newline. Always add an extra 42 # newline to the file contents, since we're going through the string 43 # version of the API. 44 src = f.read() + "\n" 45 f.close() 46 return parse(src) 47 48 def parse(buf, mode="exec"): 49 if mode == "exec" or mode == "single": 50 return Transformer().parsesuite(buf) 51 elif mode == "eval": 52 return Transformer().parseexpr(buf) 53 else: 54 raise ValueError("compile() arg 3 must be" 55 " 'exec' or 'eval' or 'single'") 56 57 def extractLineNo(ast): 58 if not isinstance(ast[1], tuple): 59 # get a terminal node 60 return ast[2] 61 for child in ast[1:]: 62 if isinstance(child, tuple): 63 lineno = extractLineNo(child) 64 if lineno is not None: 65 return lineno 66 67 def Node(*args): 68 kind = args[0] 69 if kind in nodes: 70 try: 71 return nodes[kind](*args[1:]) 72 except TypeError: 73 print nodes[kind], len(args), args 74 raise 75 else: 76 raise WalkerError, "Can't find appropriate Node type: %s" % str(args) 77 #return apply(ast.Node, args) 78 79 class Transformer: 80 """Utility object for transforming Python parse trees. 81 82 Exposes the following methods: 83 tree = transform(ast_tree) 84 tree = parsesuite(text) 85 tree = parseexpr(text) 86 tree = parsefile(fileob | filename) 87 """ 88 89 def __init__(self): 90 self._dispatch = {} 91 for value, name in sym_name.items(): 92 if hasattr(self, name): 93 self._dispatch[value] = getattr(self, name) 94 self._dispatch[token["NEWLINE"]] = self.com_NEWLINE 95 self._atom_dispatch = {token["LPAR"]: self.atom_lpar, 96 token["LSQB"]: self.atom_lsqb, 97 token["LBRACE"]: self.atom_lbrace, 98 token["BACKQUOTE"]: self.atom_backquote, 99 token["NUMBER"]: self.atom_number, 100 token["STRING"]: self.atom_string, 101 token["NAME"]: self.atom_name, 102 } 103 self.encoding = None 104 105 def transform(self, tree): 106 """Transform an AST into a modified parse tree.""" 107 if not (isinstance(tree, tuple) or isinstance(tree, list)): 108 tree = parser.st2tuple(tree, line_info=1) 109 return self.compile_node(tree) 110 111 def parsesuite(self, text): 112 """Return a modified parse tree for the given suite text.""" 113 return self.transform(parser.suite(text)) 114 115 def parseexpr(self, text): 116 """Return a modified parse tree for the given expression text.""" 117 return self.transform(parser.expr(text)) 118 119 def parsefile(self, file): 120 """Return a modified parse tree for the contents of the given file.""" 121 if type(file) == type(''): 122 file = open(file) 123 return self.parsesuite(file.read()) 124 125 # -------------------------------------------------------------- 126 # 127 # PRIVATE METHODS 128 # 129 130 def compile_node(self, node): 131 ### emit a line-number node? 132 n = node[0] 133 134 if n == symbol["encoding_decl"]: 135 self.encoding = node[2] 136 node = node[1] 137 n = node[0] 138 139 if n == symbol["single_input"]: 140 return self.single_input(node[1:]) 141 if n == symbol["file_input"]: 142 return self.file_input(node[1:]) 143 if n == symbol["eval_input"]: 144 return self.eval_input(node[1:]) 145 if n == symbol["lambdef"]: 146 return self.lambdef(node[1:]) 147 if n == symbol["funcdef"]: 148 return self.funcdef(node[1:]) 149 if n == symbol["classdef"]: 150 return self.classdef(node[1:]) 151 152 raise WalkerError, ('unexpected node type', n) 153 154 def single_input(self, node): 155 ### do we want to do anything about being "interactive" ? 156 157 # NEWLINE | simple_stmt | compound_stmt NEWLINE 158 n = node[0][0] 159 if n != token["NEWLINE"]: 160 return self.com_stmt(node[0]) 161 162 return Pass() 163 164 def file_input(self, nodelist): 165 doc = self.get_docstring(nodelist, symbol["file_input"]) 166 if doc is not None: 167 i = 1 168 else: 169 i = 0 170 stmts = [] 171 for node in nodelist[i:]: 172 if node[0] != token["ENDMARKER"] and node[0] != token["NEWLINE"]: 173 self.com_append_stmt(stmts, node) 174 return Module(doc, Stmt(stmts)) 175 176 def eval_input(self, nodelist): 177 # from the built-in function input() 178 ### is this sufficient? 179 return Expression(self.com_node(nodelist[0])) 180 181 def funcdef(self, nodelist): 182 # -5 -4 -3 -2 -1 183 # funcdef: 'def' NAME parameters ':' suite 184 # parameters: '(' [varargslist] ')' 185 186 assert len(nodelist) == 5 187 decorators = None 188 189 lineno = nodelist[-4][2] 190 name = nodelist[-4][1] 191 args = nodelist[-3][2] 192 193 if args[0] == symbol["varargslist"]: 194 names, defaults, flags = self.com_arglist(args[1:]) 195 else: 196 names = defaults = () 197 flags = 0 198 doc = self.get_docstring(nodelist[-1]) 199 200 # code for function 201 code = self.com_node(nodelist[-1]) 202 203 if doc is not None: 204 assert isinstance(code, Stmt) 205 assert isinstance(code.nodes[0], Discard) 206 del code.nodes[0] 207 return Function(decorators, name, names, defaults, flags, doc, code, 208 lineno=lineno) 209 210 def lambdef(self, nodelist): 211 # lambdef: 'lambda' [varargslist] ':' test 212 if nodelist[2][0] == symbol["varargslist"]: 213 names, defaults, flags = self.com_arglist(nodelist[2][1:]) 214 else: 215 names = defaults = () 216 flags = 0 217 218 # code for lambda 219 code = self.com_node(nodelist[-1]) 220 221 return Lambda(names, defaults, flags, code, lineno=nodelist[1][2]) 222 old_lambdef = lambdef 223 224 def classdef(self, nodelist): 225 # classdef: 'class' NAME ['(' [testlist] ')'] ':' suite 226 227 name = nodelist[1][1] 228 doc = self.get_docstring(nodelist[-1]) 229 if nodelist[2][0] == token["COLON"]: 230 bases = [] 231 elif nodelist[3][0] == token["RPAR"]: 232 bases = [] 233 else: 234 bases = self.com_bases(nodelist[3]) 235 236 # code for class 237 code = self.com_node(nodelist[-1]) 238 239 if doc is not None: 240 assert isinstance(code, Stmt) 241 assert isinstance(code.nodes[0], Discard) 242 del code.nodes[0] 243 244 return Class(name, bases, doc, code, lineno=nodelist[1][2]) 245 246 def stmt(self, nodelist): 247 return self.com_stmt(nodelist[0]) 248 249 small_stmt = stmt 250 flow_stmt = stmt 251 compound_stmt = stmt 252 253 def simple_stmt(self, nodelist): 254 # small_stmt (';' small_stmt)* [';'] NEWLINE 255 stmts = [] 256 for i in range(0, len(nodelist), 2): 257 self.com_append_stmt(stmts, nodelist[i]) 258 return Stmt(stmts) 259 260 def parameters(self, nodelist): 261 raise WalkerError 262 263 def varargslist(self, nodelist): 264 raise WalkerError 265 266 def fpdef(self, nodelist): 267 raise WalkerError 268 269 def fplist(self, nodelist): 270 raise WalkerError 271 272 def dotted_name(self, nodelist): 273 raise WalkerError 274 275 def comp_op(self, nodelist): 276 raise WalkerError 277 278 def trailer(self, nodelist): 279 raise WalkerError 280 281 def sliceop(self, nodelist): 282 raise WalkerError 283 284 def argument(self, nodelist): 285 raise WalkerError 286 287 # -------------------------------------------------------------- 288 # 289 # STATEMENT NODES (invoked by com_node()) 290 # 291 292 def expr_stmt(self, nodelist): 293 # augassign testlist | testlist ('=' testlist)* 294 en = nodelist[-1] 295 exprNode = self.lookup_node(en)(en[1:]) 296 if len(nodelist) == 1: 297 return Discard(exprNode, lineno=exprNode.lineno) 298 if nodelist[1][0] == token["EQUAL"]: 299 nodesl = [] 300 for i in range(0, len(nodelist) - 2, 2): 301 nodesl.append(self.com_assign(nodelist[i], OP_ASSIGN)) 302 return Assign(nodesl, exprNode, lineno=nodelist[1][2]) 303 else: 304 lval = self.com_augassign(nodelist[0]) 305 op = self.com_augassign_op(nodelist[1]) 306 return AugAssign(lval, op[1], exprNode, lineno=op[2]) 307 raise WalkerError, "can't get here" 308 309 def print_stmt(self, nodelist): 310 # print ([ test (',' test)* [','] ] | '>>' test [ (',' test)+ [','] ]) 311 items = [] 312 if len(nodelist) == 1: 313 start = 1 314 dest = None 315 elif nodelist[1][0] == token["RIGHTSHIFT"]: 316 assert len(nodelist) == 3 \ 317 or nodelist[3][0] == token["COMMA"] 318 dest = self.com_node(nodelist[2]) 319 start = 4 320 else: 321 dest = None 322 start = 1 323 for i in range(start, len(nodelist), 2): 324 items.append(self.com_node(nodelist[i])) 325 if nodelist[-1][0] == token["COMMA"]: 326 return Print(items, dest, lineno=nodelist[0][2]) 327 return Printnl(items, dest, lineno=nodelist[0][2]) 328 329 def del_stmt(self, nodelist): 330 return self.com_assign(nodelist[1], OP_DELETE) 331 332 def pass_stmt(self, nodelist): 333 return Pass(lineno=nodelist[0][2]) 334 335 def break_stmt(self, nodelist): 336 return Break(lineno=nodelist[0][2]) 337 338 def continue_stmt(self, nodelist): 339 return Continue(lineno=nodelist[0][2]) 340 341 def return_stmt(self, nodelist): 342 # return: [testlist] 343 if len(nodelist) < 2: 344 return Return(Const(None), lineno=nodelist[0][2]) 345 return Return(self.com_node(nodelist[1]), lineno=nodelist[0][2]) 346 347 def raise_stmt(self, nodelist): 348 # raise: [test [',' test [',' test]]] 349 if len(nodelist) > 5: 350 expr3 = self.com_node(nodelist[5]) 351 else: 352 expr3 = None 353 if len(nodelist) > 3: 354 expr2 = self.com_node(nodelist[3]) 355 else: 356 expr2 = None 357 if len(nodelist) > 1: 358 expr1 = self.com_node(nodelist[1]) 359 else: 360 expr1 = None 361 return Raise(expr1, expr2, expr3, lineno=nodelist[0][2]) 362 363 def import_stmt(self, nodelist): 364 # import_stmt: import_name | import_from 365 assert len(nodelist) == 1 366 return self.com_node(nodelist[0]) 367 368 def import_name(self, nodelist): 369 # import_name: 'import' dotted_as_names 370 return Import(self.com_dotted_as_names(nodelist[1]), 371 lineno=nodelist[0][2]) 372 373 def import_from(self, nodelist): 374 # import_from: 'from' ('.'* dotted_name | '.') 'import' ('*' | 375 # '(' import_as_names ')' | import_as_names) 376 assert nodelist[0][1] == 'from' 377 idx = 1 378 while nodelist[idx][1] == '.': 379 idx += 1 380 level = idx - 1 381 if nodelist[idx][0] == symbol["dotted_name"]: 382 fromname = self.com_dotted_name(nodelist[idx]) 383 idx += 1 384 else: 385 fromname = "" 386 assert nodelist[idx][1] == 'import' 387 if nodelist[idx + 1][0] == token["STAR"]: 388 return From(fromname, [('*', None)], level, 389 lineno=nodelist[0][2]) 390 else: 391 node = nodelist[idx + 1 + (nodelist[idx + 1][0] == token["LPAR"])] 392 return From(fromname, self.com_import_as_names(node), level, 393 lineno=nodelist[0][2]) 394 395 def global_stmt(self, nodelist): 396 # global: NAME (',' NAME)* 397 names = [] 398 for i in range(1, len(nodelist), 2): 399 names.append(nodelist[i][1]) 400 return Global(names, lineno=nodelist[0][2]) 401 402 def exec_stmt(self, nodelist): 403 # exec_stmt: 'exec' expr ['in' expr [',' expr]] 404 expr1 = self.com_node(nodelist[1]) 405 if len(nodelist) >= 4: 406 expr2 = self.com_node(nodelist[3]) 407 if len(nodelist) >= 6: 408 expr3 = self.com_node(nodelist[5]) 409 else: 410 expr3 = None 411 else: 412 expr2 = expr3 = None 413 414 return Exec(expr1, expr2, expr3, lineno=nodelist[0][2]) 415 416 def assert_stmt(self, nodelist): 417 # 'assert': test, [',' test] 418 expr1 = self.com_node(nodelist[1]) 419 if (len(nodelist) == 4): 420 expr2 = self.com_node(nodelist[3]) 421 else: 422 expr2 = None 423 return Assert(expr1, expr2, lineno=nodelist[0][2]) 424 425 def if_stmt(self, nodelist): 426 # if: test ':' suite ('elif' test ':' suite)* ['else' ':' suite] 427 tests = [] 428 for i in range(0, len(nodelist) - 3, 4): 429 testNode = self.com_node(nodelist[i + 1]) 430 suiteNode = self.com_node(nodelist[i + 3]) 431 tests.append((testNode, suiteNode)) 432 433 if len(nodelist) % 4 == 3: 434 elseNode = self.com_node(nodelist[-1]) 435 ## elseNode.lineno = nodelist[-1][1][2] 436 else: 437 elseNode = None 438 return If(tests, elseNode, lineno=nodelist[0][2]) 439 440 def while_stmt(self, nodelist): 441 # 'while' test ':' suite ['else' ':' suite] 442 443 testNode = self.com_node(nodelist[1]) 444 bodyNode = self.com_node(nodelist[3]) 445 446 if len(nodelist) > 4: 447 elseNode = self.com_node(nodelist[6]) 448 else: 449 elseNode = None 450 451 return While(testNode, bodyNode, elseNode, lineno=nodelist[0][2]) 452 453 def for_stmt(self, nodelist): 454 # 'for' exprlist 'in' exprlist ':' suite ['else' ':' suite] 455 456 assignNode = self.com_assign(nodelist[1], OP_ASSIGN) 457 listNode = self.com_node(nodelist[3]) 458 bodyNode = self.com_node(nodelist[5]) 459 460 if len(nodelist) > 8: 461 elseNode = self.com_node(nodelist[8]) 462 else: 463 elseNode = None 464 465 return For(assignNode, listNode, bodyNode, elseNode, 466 lineno=nodelist[0][2]) 467 468 def try_stmt(self, nodelist): 469 return self.com_try_except_finally(nodelist) 470 471 def suite(self, nodelist): 472 # simple_stmt | NEWLINE INDENT NEWLINE* (stmt NEWLINE*)+ DEDENT 473 if len(nodelist) == 1: 474 return self.com_stmt(nodelist[0]) 475 476 stmts = [] 477 for node in nodelist: 478 if node[0] == symbol["stmt"]: 479 self.com_append_stmt(stmts, node) 480 return Stmt(stmts) 481 482 # -------------------------------------------------------------- 483 # 484 # EXPRESSION NODES (invoked by com_node()) 485 # 486 487 def testlist(self, nodelist): 488 # testlist: expr (',' expr)* [','] 489 # testlist_safe: test [(',' test)+ [',']] 490 # exprlist: expr (',' expr)* [','] 491 return self.com_binary(Tuple, nodelist) 492 493 testlist_safe = testlist # XXX 494 testlist1 = testlist 495 exprlist = testlist 496 497 def testlist_comp(self, nodelist): 498 # test ( (',' test)* [','] ) 499 assert nodelist[0][0] == symbol["test"] 500 return self.testlist(nodelist) 501 502 def test(self, nodelist): 503 # or_test | lambdef 504 if len(nodelist) == 1 and nodelist[0][0] == symbol["lambdef"]: 505 return self.lambdef(nodelist[0]) 506 then = self.com_node(nodelist[0]) 507 return then 508 509 def or_test(self, nodelist): 510 # and_test ('or' and_test)* | lambdef 511 if len(nodelist) == 1 and nodelist[0][0] == symbol["lambdef"]: 512 return self.lambdef(nodelist[0]) 513 return self.com_binary(Or, nodelist) 514 old_test = or_test 515 516 def and_test(self, nodelist): 517 # not_test ('and' not_test)* 518 return self.com_binary(And, nodelist) 519 520 def not_test(self, nodelist): 521 # 'not' not_test | comparison 522 result = self.com_node(nodelist[-1]) 523 if len(nodelist) == 2: 524 return Not(result, lineno=nodelist[0][2]) 525 return result 526 527 def comparison(self, nodelist): 528 # comparison: expr (comp_op expr)* 529 node = self.com_node(nodelist[0]) 530 if len(nodelist) == 1: 531 return node 532 533 results = [] 534 for i in range(2, len(nodelist), 2): 535 nl = nodelist[i-1] 536 537 # comp_op: '<' | '>' | '=' | '>=' | '<=' | '<>' | '!=' | '==' 538 # | 'in' | 'not' 'in' | 'is' | 'is' 'not' 539 n = nl[1] 540 if n[0] == token["NAME"]: 541 type = n[1] 542 if len(nl) == 3: 543 if type == 'not': 544 type = 'not in' 545 else: 546 type = 'is not' 547 else: 548 type = _cmp_types[n[0]] 549 550 lineno = nl[1][2] 551 results.append((type, self.com_node(nodelist[i]))) 552 553 # we need a special "compare" node so that we can distinguish 554 # 3 < x < 5 from (3 < x) < 5 555 # the two have very different semantics and results (note that the 556 # latter form is always true) 557 558 return Compare(node, results, lineno=lineno) 559 560 def expr(self, nodelist): 561 # xor_expr ('|' xor_expr)* 562 return self.com_binary(Bitor, nodelist) 563 564 def xor_expr(self, nodelist): 565 # xor_expr ('^' xor_expr)* 566 return self.com_binary(Bitxor, nodelist) 567 568 def and_expr(self, nodelist): 569 # xor_expr ('&' xor_expr)* 570 return self.com_binary(Bitand, nodelist) 571 572 def shift_expr(self, nodelist): 573 # shift_expr ('<<'|'>>' shift_expr)* 574 node = self.com_node(nodelist[0]) 575 for i in range(2, len(nodelist), 2): 576 right = self.com_node(nodelist[i]) 577 if nodelist[i-1][0] == token["LEFTSHIFT"]: 578 node = LeftShift([node, right], lineno=nodelist[1][2]) 579 elif nodelist[i-1][0] == token["RIGHTSHIFT"]: 580 node = RightShift([node, right], lineno=nodelist[1][2]) 581 else: 582 raise ValueError, "unexpected token: %s" % nodelist[i-1][0] 583 return node 584 585 def arith_expr(self, nodelist): 586 node = self.com_node(nodelist[0]) 587 for i in range(2, len(nodelist), 2): 588 right = self.com_node(nodelist[i]) 589 if nodelist[i-1][0] == token["PLUS"]: 590 node = Add([node, right], lineno=nodelist[1][2]) 591 elif nodelist[i-1][0] == token["MINUS"]: 592 node = Sub([node, right], lineno=nodelist[1][2]) 593 else: 594 raise ValueError, "unexpected token: %s" % nodelist[i-1][0] 595 return node 596 597 def term(self, nodelist): 598 node = self.com_node(nodelist[0]) 599 for i in range(2, len(nodelist), 2): 600 right = self.com_node(nodelist[i]) 601 t = nodelist[i-1][0] 602 if t == token["STAR"]: 603 node = Mul([node, right]) 604 elif t == token["SLASH"]: 605 node = Div([node, right]) 606 elif t == token["PERCENT"]: 607 node = Mod([node, right]) 608 elif t == token["DOUBLESLASH"]: 609 node = FloorDiv([node, right]) 610 else: 611 raise ValueError, "unexpected token: %s" % t 612 node.lineno = nodelist[1][2] 613 return node 614 615 def factor(self, nodelist): 616 elt = nodelist[0] 617 t = elt[0] 618 node = self.lookup_node(nodelist[-1])(nodelist[-1][1:]) 619 # need to handle (unary op)constant here... 620 if t == token["PLUS"]: 621 return UnaryAdd(node, lineno=elt[2]) 622 elif t == token["MINUS"]: 623 return UnarySub(node, lineno=elt[2]) 624 elif t == token["TILDE"]: 625 node = Invert(node, lineno=elt[2]) 626 return node 627 628 def power(self, nodelist): 629 # power: atom trailer* ('**' factor)* 630 node = self.com_node(nodelist[0]) 631 for i in range(1, len(nodelist)): 632 elt = nodelist[i] 633 if elt[0] == token["DOUBLESTAR"]: 634 return Power([node, self.com_node(nodelist[i+1])], 635 lineno=elt[2]) 636 637 node = self.com_apply_trailer(node, elt) 638 639 return node 640 641 def atom(self, nodelist): 642 return self._atom_dispatch[nodelist[0][0]](nodelist) 643 644 def atom_lpar(self, nodelist): 645 if nodelist[1][0] == token["RPAR"]: 646 return Tuple((), lineno=nodelist[0][2]) 647 return self.com_node(nodelist[1]) 648 649 def atom_lsqb(self, nodelist): 650 if nodelist[1][0] == token["RSQB"]: 651 return List((), lineno=nodelist[0][2]) 652 return self.com_list_constructor(nodelist[1]) 653 654 def atom_lbrace(self, nodelist): 655 if nodelist[1][0] == token["RBRACE"]: 656 return Dict((), lineno=nodelist[0][2]) 657 return self.com_dictorsetmaker(nodelist[1]) 658 659 def atom_backquote(self, nodelist): 660 return Backquote(self.com_node(nodelist[1])) 661 662 def atom_number(self, nodelist): 663 ### need to verify this matches compile.c 664 k = eval(nodelist[0][1]) 665 return Const(k, nodelist[0][1], lineno=nodelist[0][2]) 666 667 def decode_literal(self, lit): 668 if self.encoding: 669 # this is particularly fragile & a bit of a 670 # hack... changes in compile.c:parsestr and 671 # tokenizer.c must be reflected here. 672 if self.encoding != 'utf-8': 673 lit = unicode(lit, 'utf-8').encode(self.encoding) 674 return eval("# coding: %s\n%s" % (self.encoding, lit)) 675 else: 676 return eval(lit) 677 678 def atom_string(self, nodelist): 679 k = '' 680 l = [] 681 for node in nodelist: 682 k += self.decode_literal(node[1]) 683 l.append(node[1]) 684 return Const(k, l, lineno=nodelist[0][2]) 685 686 def atom_name(self, nodelist): 687 return Name(nodelist[0][1], lineno=nodelist[0][2]) 688 689 # -------------------------------------------------------------- 690 # 691 # INTERNAL PARSING UTILITIES 692 # 693 694 # The use of com_node() introduces a lot of extra stack frames, 695 # enough to cause a stack overflow compiling test.test_parser with 696 # the standard interpreter recursionlimit. The com_node() is a 697 # convenience function that hides the dispatch details, but comes 698 # at a very high cost. It is more efficient to dispatch directly 699 # in the callers. In these cases, use lookup_node() and call the 700 # dispatched node directly. 701 702 def lookup_node(self, node): 703 return self._dispatch[node[0]] 704 705 def com_node(self, node): 706 # Note: compile.c has handling in com_node for del_stmt, pass_stmt, 707 # break_stmt, stmt, small_stmt, flow_stmt, simple_stmt, 708 # and compound_stmt. 709 # We'll just dispatch them. 710 return self._dispatch[node[0]](node[1:]) 711 712 def com_NEWLINE(self, *args): 713 # A ';' at the end of a line can make a NEWLINE token appear 714 # here, Render it harmless. (genc discards ('discard', 715 # ('const', xxxx)) Nodes) 716 return Discard(Const(None)) 717 718 def com_arglist(self, nodelist): 719 # varargslist: 720 # (fpdef ['=' test] ',')* ('*' NAME [',' '**' NAME] | '**' NAME) 721 # | fpdef ['=' test] (',' fpdef ['=' test])* [','] 722 # fpdef: NAME | '.' NAME | '(' fplist ')' 723 # fplist: fpdef (',' fpdef)* [','] 724 names = [] 725 defaults = [] 726 flags = 0 727 728 i = 0 729 while i < len(nodelist): 730 node = nodelist[i] 731 if node[0] == token["STAR"] or node[0] == token["DOUBLESTAR"]: 732 if node[0] == token["STAR"]: 733 node = nodelist[i+1] 734 if node[0] == token["NAME"]: 735 names.append(node[1]) 736 flags = flags | CO_VARARGS 737 i = i + 3 738 739 if i < len(nodelist): 740 # should be DOUBLESTAR 741 t = nodelist[i][0] 742 if t == token["DOUBLESTAR"]: 743 node = nodelist[i+1] 744 else: 745 raise ValueError, "unexpected token: %s" % t 746 names.append(node[1]) 747 flags = flags | CO_VARKEYWORDS 748 749 break 750 751 # fpdef: NAME | '.' NAME | '(' fplist ')' 752 names.append(self.com_fpdef(node)) 753 754 i = i + 1 755 if i < len(nodelist) and nodelist[i][0] == token["EQUAL"]: 756 defaults.append(self.com_node(nodelist[i + 1])) 757 i = i + 2 758 elif len(defaults): 759 # we have already seen an argument with default, but here 760 # came one without 761 raise SyntaxError, "non-default argument follows default argument" 762 763 # skip the comma 764 i = i + 1 765 766 return names, defaults, flags 767 768 def com_fpdef(self, node): 769 # fpdef: NAME | '.' NAME | '(' fplist ')' 770 if node[1][0] == token["LPAR"]: 771 return self.com_fplist(node[2]) 772 elif node[1][0] == token["DOT"]: 773 return node[1][1] + node[2][1] 774 return node[1][1] 775 776 def com_fplist(self, node): 777 # fplist: fpdef (',' fpdef)* [','] 778 if len(node) == 2: 779 return self.com_fpdef(node[1]) 780 list = [] 781 for i in range(1, len(node), 2): 782 list.append(self.com_fpdef(node[i])) 783 return tuple(list) 784 785 def com_dotted_name(self, node): 786 # String together the dotted names and return the string 787 name = "" 788 for n in node: 789 if type(n) == type(()) and n[0] == 1: 790 name = name + n[1] + '.' 791 return name[:-1] 792 793 def com_dotted_as_name(self, node): 794 assert node[0] == symbol["dotted_as_name"] 795 node = node[1:] 796 dot = self.com_dotted_name(node[0][1:]) 797 if len(node) == 1: 798 return dot, None 799 assert node[1][1] == 'as' 800 assert node[2][0] == token["NAME"] 801 return dot, node[2][1] 802 803 def com_dotted_as_names(self, node): 804 assert node[0] == symbol["dotted_as_names"] 805 node = node[1:] 806 names = [self.com_dotted_as_name(node[0])] 807 for i in range(2, len(node), 2): 808 names.append(self.com_dotted_as_name(node[i])) 809 return names 810 811 def com_import_as_name(self, node): 812 assert node[0] == symbol["import_as_name"] 813 node = node[1:] 814 assert node[0][0] == token["NAME"] 815 if len(node) == 1: 816 return node[0][1], None 817 assert node[1][1] == 'as', node 818 assert node[2][0] == token["NAME"] 819 return node[0][1], node[2][1] 820 821 def com_import_as_names(self, node): 822 assert node[0] == symbol["import_as_names"] 823 node = node[1:] 824 names = [self.com_import_as_name(node[0])] 825 for i in range(2, len(node), 2): 826 names.append(self.com_import_as_name(node[i])) 827 return names 828 829 def com_bases(self, node): 830 bases = [] 831 for i in range(1, len(node), 2): 832 bases.append(self.com_node(node[i])) 833 return bases 834 835 def com_try_except_finally(self, nodelist): 836 # ('try' ':' suite 837 # ((except_clause ':' suite)+ ['else' ':' suite] ['finally' ':' suite] 838 # | 'finally' ':' suite)) 839 840 if nodelist[3][0] == token["NAME"]: 841 # first clause is a finally clause: only try-finally 842 return TryFinally(self.com_node(nodelist[2]), 843 self.com_node(nodelist[5]), 844 lineno=nodelist[0][2]) 845 846 #tryexcept: [TryNode, [except_clauses], elseNode)] 847 clauses = [] 848 elseNode = None 849 finallyNode = None 850 for i in range(3, len(nodelist), 3): 851 node = nodelist[i] 852 if node[0] == symbol["except_clause"]: 853 # except_clause: 'except' [expr [(',' | 'as') expr]] */ 854 if len(node) > 2: 855 expr1 = self.com_node(node[2]) 856 if len(node) > 4: 857 expr2 = self.com_assign(node[4], OP_ASSIGN) 858 else: 859 expr2 = None 860 else: 861 expr1 = expr2 = None 862 clauses.append((expr1, expr2, self.com_node(nodelist[i+2]))) 863 864 if node[0] == token["NAME"]: 865 if node[1] == 'else': 866 elseNode = self.com_node(nodelist[i+2]) 867 elif node[1] == 'finally': 868 finallyNode = self.com_node(nodelist[i+2]) 869 try_except = TryExcept(self.com_node(nodelist[2]), clauses, elseNode, 870 lineno=nodelist[0][2]) 871 if finallyNode: 872 return TryFinally(try_except, finallyNode, lineno=nodelist[0][2]) 873 else: 874 return try_except 875 876 def com_augassign_op(self, node): 877 assert node[0] == symbol["augassign"] 878 return node[1] 879 880 def com_augassign(self, node): 881 """Return node suitable for lvalue of augmented assignment 882 883 Names, slices, and attributes are the only allowable nodes. 884 """ 885 l = self.com_node(node) 886 if l.__class__ in (Name, Slice, Subscript, Getattr): 887 return l 888 raise SyntaxError, "can't assign to %s" % l.__class__.__name__ 889 890 def com_assign(self, node, assigning): 891 # return a node suitable for use as an "lvalue" 892 # loop to avoid trivial recursion 893 while 1: 894 t = node[0] 895 if t in (symbol["exprlist"], symbol["testlist"], symbol["testlist_safe"], symbol["testlist_comp"]): 896 if len(node) > 2: 897 return self.com_assign_tuple(node, assigning) 898 node = node[1] 899 elif t in _assign_types: 900 if len(node) > 2: 901 raise SyntaxError, "can't assign to operator" 902 node = node[1] 903 elif t == symbol["power"]: 904 if node[1][0] != symbol["atom"]: 905 raise SyntaxError, "can't assign to operator" 906 if len(node) > 2: 907 primary = self.com_node(node[1]) 908 for i in range(2, len(node)-1): 909 ch = node[i] 910 if ch[0] == token["DOUBLESTAR"]: 911 raise SyntaxError, "can't assign to operator" 912 primary = self.com_apply_trailer(primary, ch) 913 return self.com_assign_trailer(primary, node[-1], 914 assigning) 915 node = node[1] 916 elif t == symbol["atom"]: 917 t = node[1][0] 918 if t == token["LPAR"]: 919 node = node[2] 920 if node[0] == token["RPAR"]: 921 raise SyntaxError, "can't assign to ()" 922 elif t == token["LSQB"]: 923 node = node[2] 924 if node[0] == token["RSQB"]: 925 raise SyntaxError, "can't assign to []" 926 return self.com_assign_list(node, assigning) 927 elif t == token["NAME"]: 928 return self.com_assign_name(node[1], assigning) 929 else: 930 raise SyntaxError, "can't assign to literal" 931 else: 932 raise SyntaxError, "bad assignment (%s)" % t 933 934 def com_assign_tuple(self, node, assigning): 935 assigns = [] 936 for i in range(1, len(node), 2): 937 assigns.append(self.com_assign(node[i], assigning)) 938 return AssTuple(assigns, lineno=extractLineNo(node)) 939 940 def com_assign_list(self, node, assigning): 941 assigns = [] 942 for i in range(1, len(node), 2): 943 if i + 1 < len(node): 944 if node[i + 1][0] == symbol["list_for"]: 945 raise SyntaxError, "can't assign to list comprehension" 946 assert node[i + 1][0] == token["COMMA"], node[i + 1] 947 assigns.append(self.com_assign(node[i], assigning)) 948 return AssList(assigns, lineno=extractLineNo(node)) 949 950 def com_assign_name(self, node, assigning): 951 return AssName(node[1], assigning, lineno=node[2]) 952 953 def com_assign_trailer(self, primary, node, assigning): 954 t = node[1][0] 955 if t == token["DOT"]: 956 return self.com_assign_attr(primary, node[2], assigning) 957 if t == token["LSQB"]: 958 return self.com_subscriptlist(primary, node[2], assigning) 959 if t == token["LPAR"]: 960 raise SyntaxError, "can't assign to function call" 961 raise SyntaxError, "unknown trailer type: %s" % t 962 963 def com_assign_attr(self, primary, node, assigning): 964 return AssAttr(primary, node[1], assigning, lineno=node[-1]) 965 966 def com_binary(self, constructor, nodelist): 967 "Compile 'NODE (OP NODE)*' into (type, [ node1, ..., nodeN ])." 968 l = len(nodelist) 969 if l == 1: 970 n = nodelist[0] 971 return self.lookup_node(n)(n[1:]) 972 items = [] 973 for i in range(0, l, 2): 974 n = nodelist[i] 975 items.append(self.lookup_node(n)(n[1:])) 976 return constructor(items, lineno=extractLineNo(nodelist)) 977 978 def com_stmt(self, node): 979 result = self.lookup_node(node)(node[1:]) 980 assert result is not None 981 if isinstance(result, Stmt): 982 return result 983 return Stmt([result]) 984 985 def com_append_stmt(self, stmts, node): 986 result = self.lookup_node(node)(node[1:]) 987 assert result is not None 988 if isinstance(result, Stmt): 989 stmts.extend(result.nodes) 990 else: 991 stmts.append(result) 992 993 def com_list_constructor(self, nodelist): 994 # listmaker: test ( (',' test)* [','] ) 995 values = [] 996 for i in range(1, len(nodelist)): 997 if nodelist[i][0] == token["COMMA"]: 998 continue 999 values.append(self.com_node(nodelist[i])) 1000 return List(values, lineno=values[0].lineno) 1001 1002 def com_dictorsetmaker(self, nodelist): 1003 # dictorsetmaker: ( (test ':' test ( (',' test ':' test)* [','])) | 1004 # (test ( (',' test)* [','])) ) 1005 assert nodelist[0] == symbol["dictorsetmaker"] 1006 nodelist = nodelist[1:] 1007 if len(nodelist) == 1 or nodelist[1][0] == token["COMMA"]: 1008 # set literal 1009 items = [] 1010 for i in range(0, len(nodelist), 2): 1011 items.append(self.com_node(nodelist[i])) 1012 return Set(items, lineno=items[0].lineno) 1013 else: 1014 # dict literal 1015 items = [] 1016 for i in range(0, len(nodelist), 4): 1017 items.append((self.com_node(nodelist[i]), 1018 self.com_node(nodelist[i+2]))) 1019 return Dict(items, lineno=items[0][0].lineno) 1020 1021 def com_apply_trailer(self, primaryNode, nodelist): 1022 t = nodelist[1][0] 1023 if t == token["LPAR"]: 1024 return self.com_call_function(primaryNode, nodelist[2]) 1025 if t == token["DOT"]: 1026 return self.com_select_member(primaryNode, nodelist[2]) 1027 if t == token["LSQB"]: 1028 return self.com_subscriptlist(primaryNode, nodelist[2], OP_APPLY) 1029 1030 raise SyntaxError, 'unknown node type: %s' % t 1031 1032 def com_select_member(self, primaryNode, nodelist): 1033 if nodelist[0] != token["NAME"]: 1034 raise SyntaxError, "member must be a name" 1035 return Getattr(primaryNode, nodelist[1], lineno=nodelist[2]) 1036 1037 def com_call_function(self, primaryNode, nodelist): 1038 if nodelist[0] == token["RPAR"]: 1039 return CallFunc(primaryNode, [], lineno=extractLineNo(nodelist)) 1040 args = [] 1041 kw = 0 1042 star_node = dstar_node = None 1043 len_nodelist = len(nodelist) 1044 i = 1 1045 while i < len_nodelist: 1046 node = nodelist[i] 1047 1048 if node[0]==token["STAR"]: 1049 if star_node is not None: 1050 raise SyntaxError, 'already have the varargs indentifier' 1051 star_node = self.com_node(nodelist[i+1]) 1052 i = i + 3 1053 continue 1054 elif node[0]==token["DOUBLESTAR"]: 1055 if dstar_node is not None: 1056 raise SyntaxError, 'already have the kwargs indentifier' 1057 dstar_node = self.com_node(nodelist[i+1]) 1058 i = i + 3 1059 continue 1060 1061 # positional or named parameters 1062 kw, result = self.com_argument(node, kw, star_node) 1063 1064 args.append(result) 1065 i = i + 2 1066 1067 return CallFunc(primaryNode, args, star_node, dstar_node, 1068 lineno=extractLineNo(nodelist)) 1069 1070 def com_argument(self, nodelist, kw, star_node): 1071 if len(nodelist) == 2: 1072 if kw: 1073 raise SyntaxError, "non-keyword arg after keyword arg" 1074 if star_node: 1075 raise SyntaxError, "only named arguments may follow *expression" 1076 return 0, self.com_node(nodelist[1]) 1077 result = self.com_node(nodelist[3]) 1078 n = nodelist[1] 1079 while len(n) == 2 and n[0] != token["NAME"]: 1080 n = n[1] 1081 if n[0] != token["NAME"]: 1082 raise SyntaxError, "keyword can't be an expression (%s)"%n[0] 1083 node = Keyword(n[1], result, lineno=n[2]) 1084 return 1, node 1085 1086 def com_subscriptlist(self, primary, nodelist, assigning): 1087 # slicing: simple_slicing | extended_slicing 1088 # simple_slicing: primary "[" short_slice "]" 1089 # extended_slicing: primary "[" slice_list "]" 1090 # slice_list: slice_item ("," slice_item)* [","] 1091 1092 # backwards compat slice for '[i:j]' 1093 if len(nodelist) == 2: 1094 sub = nodelist[1] 1095 if (sub[1][0] == token["COLON"] or \ 1096 (len(sub) > 2 and sub[2][0] == token["COLON"])) and \ 1097 sub[-1][0] != symbol["sliceop"]: 1098 return self.com_slice(primary, sub, assigning) 1099 1100 subscripts = [] 1101 for i in range(1, len(nodelist), 2): 1102 subscripts.append(self.com_subscript(nodelist[i])) 1103 return Subscript(primary, assigning, subscripts, 1104 lineno=extractLineNo(nodelist)) 1105 1106 def com_subscript(self, node): 1107 # slice_item: expression | proper_slice | ellipsis 1108 ch = node[1] 1109 t = ch[0] 1110 if t == token["DOT"] and node[2][0] == token["DOT"]: 1111 return Ellipsis() 1112 if t == token["COLON"] or len(node) > 2: 1113 return self.com_sliceobj(node) 1114 return self.com_node(ch) 1115 1116 def com_sliceobj(self, node): 1117 # proper_slice: short_slice | long_slice 1118 # short_slice: [lower_bound] ":" [upper_bound] 1119 # long_slice: short_slice ":" [stride] 1120 # lower_bound: expression 1121 # upper_bound: expression 1122 # stride: expression 1123 # 1124 # Note: a stride may be further slicing... 1125 1126 items = [] 1127 1128 if node[1][0] == token["COLON"]: 1129 items.append(Const(None)) 1130 i = 2 1131 else: 1132 items.append(self.com_node(node[1])) 1133 # i == 2 is a COLON 1134 i = 3 1135 1136 if i < len(node) and node[i][0] == symbol["test"]: 1137 items.append(self.com_node(node[i])) 1138 i = i + 1 1139 else: 1140 items.append(Const(None)) 1141 1142 # a short_slice has been built. look for long_slice now by looking 1143 # for strides... 1144 for j in range(i, len(node)): 1145 ch = node[j] 1146 if len(ch) == 2: 1147 items.append(Const(None)) 1148 else: 1149 items.append(self.com_node(ch[2])) 1150 return Sliceobj(items, lineno=extractLineNo(node)) 1151 1152 def com_slice(self, primary, node, assigning): 1153 # short_slice: [lower_bound] ":" [upper_bound] 1154 lower = upper = None 1155 if len(node) == 3: 1156 if node[1][0] == token["COLON"]: 1157 upper = self.com_node(node[2]) 1158 else: 1159 lower = self.com_node(node[1]) 1160 elif len(node) == 4: 1161 lower = self.com_node(node[1]) 1162 upper = self.com_node(node[3]) 1163 return Slice(primary, assigning, lower, upper, 1164 lineno=extractLineNo(node)) 1165 1166 def get_docstring(self, node, n=None): 1167 if n is None: 1168 n = node[0] 1169 node = node[1:] 1170 if n == symbol["suite"]: 1171 if len(node) == 1: 1172 return self.get_docstring(node[0]) 1173 for sub in node: 1174 if sub[0] == symbol["stmt"]: 1175 return self.get_docstring(sub) 1176 return None 1177 if n == symbol["file_input"]: 1178 for sub in node: 1179 if sub[0] == symbol["stmt"]: 1180 return self.get_docstring(sub) 1181 return None 1182 if n == symbol["atom"]: 1183 if node[0][0] == token["STRING"]: 1184 s = '' 1185 for t in node: 1186 s = s + eval(t[1]) 1187 return s 1188 return None 1189 if n == symbol["stmt"] or n == symbol["simple_stmt"] \ 1190 or n == symbol["small_stmt"]: 1191 return self.get_docstring(node[0]) 1192 if n in _doc_nodes and len(node) == 1: 1193 return self.get_docstring(node[0]) 1194 return None 1195 1196 1197 _doc_nodes = [ 1198 symbol["expr_stmt"], 1199 symbol["testlist"], 1200 symbol["testlist_safe"], 1201 symbol["test"], 1202 symbol["or_test"], 1203 symbol["and_test"], 1204 symbol["not_test"], 1205 symbol["comparison"], 1206 symbol["expr"], 1207 symbol["xor_expr"], 1208 symbol["and_expr"], 1209 symbol["shift_expr"], 1210 symbol["arith_expr"], 1211 symbol["term"], 1212 symbol["factor"], 1213 symbol["power"], 1214 ] 1215 1216 # comp_op: '<' | '>' | '=' | '>=' | '<=' | '<>' | '!=' | '==' 1217 # | 'in' | 'not' 'in' | 'is' | 'is' 'not' 1218 _cmp_types = { 1219 token["LESS"] : '<', 1220 token["GREATER"] : '>', 1221 token["EQEQUAL"] : '==', 1222 token["EQUAL"] : '==', 1223 token["LESSEQUAL"] : '<=', 1224 token["GREATEREQUAL"] : '>=', 1225 token["NOTEQUAL"] : '!=', 1226 } 1227 1228 _legal_node_types = [ 1229 symbol["funcdef"], 1230 symbol["classdef"], 1231 symbol["stmt"], 1232 symbol["small_stmt"], 1233 symbol["flow_stmt"], 1234 symbol["simple_stmt"], 1235 symbol["compound_stmt"], 1236 symbol["expr_stmt"], 1237 symbol["print_stmt"], 1238 symbol["del_stmt"], 1239 symbol["pass_stmt"], 1240 symbol["break_stmt"], 1241 symbol["continue_stmt"], 1242 symbol["return_stmt"], 1243 symbol["raise_stmt"], 1244 symbol["import_stmt"], 1245 symbol["global_stmt"], 1246 symbol["exec_stmt"], 1247 symbol["assert_stmt"], 1248 symbol["if_stmt"], 1249 symbol["while_stmt"], 1250 symbol["for_stmt"], 1251 symbol["try_stmt"], 1252 symbol["suite"], 1253 symbol["testlist"], 1254 symbol["testlist_safe"], 1255 symbol["test"], 1256 symbol["and_test"], 1257 symbol["not_test"], 1258 symbol["comparison"], 1259 symbol["exprlist"], 1260 symbol["expr"], 1261 symbol["xor_expr"], 1262 symbol["and_expr"], 1263 symbol["shift_expr"], 1264 symbol["arith_expr"], 1265 symbol["term"], 1266 symbol["factor"], 1267 symbol["power"], 1268 symbol["atom"], 1269 ] 1270 1271 _assign_types = [ 1272 symbol["test"], 1273 symbol["or_test"], 1274 symbol["and_test"], 1275 symbol["not_test"], 1276 symbol["comparison"], 1277 symbol["expr"], 1278 symbol["xor_expr"], 1279 symbol["and_expr"], 1280 symbol["shift_expr"], 1281 symbol["arith_expr"], 1282 symbol["term"], 1283 symbol["factor"], 1284 ] 1285 1286 _names = {} 1287 for k, v in sym_name.items(): 1288 _names[k] = v 1289 for k, v in tok_name.items(): 1290 _names[k] = v 1291 1292 def debug_tree(tree): 1293 l = [] 1294 for elt in tree: 1295 if isinstance(elt, int): 1296 l.append(_names.get(elt, elt)) 1297 elif isinstance(elt, str): 1298 l.append(elt) 1299 else: 1300 l.append(debug_tree(elt)) 1301 return l