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 with_stmt(self, nodelist): 472 return self.com_with(nodelist) 473 474 def suite(self, nodelist): 475 # simple_stmt | NEWLINE INDENT NEWLINE* (stmt NEWLINE*)+ DEDENT 476 if len(nodelist) == 1: 477 return self.com_stmt(nodelist[0]) 478 479 stmts = [] 480 for node in nodelist: 481 if node[0] == symbol["stmt"]: 482 self.com_append_stmt(stmts, node) 483 return Stmt(stmts) 484 485 # -------------------------------------------------------------- 486 # 487 # EXPRESSION NODES (invoked by com_node()) 488 # 489 490 def testlist(self, nodelist): 491 # testlist: expr (',' expr)* [','] 492 # testlist_safe: test [(',' test)+ [',']] 493 # exprlist: expr (',' expr)* [','] 494 return self.com_binary(Tuple, nodelist) 495 496 testlist_safe = testlist # XXX 497 testlist1 = testlist 498 exprlist = testlist 499 500 def testlist_comp(self, nodelist): 501 # test ( (',' test)* [','] ) 502 assert nodelist[0][0] == symbol["test"] 503 return self.testlist(nodelist) 504 505 def test(self, nodelist): 506 # or_test | lambdef 507 if len(nodelist) == 1 and nodelist[0][0] == symbol["lambdef"]: 508 return self.lambdef(nodelist[0]) 509 then = self.com_node(nodelist[0]) 510 return then 511 512 def or_test(self, nodelist): 513 # and_test ('or' and_test)* | lambdef 514 if len(nodelist) == 1 and nodelist[0][0] == symbol["lambdef"]: 515 return self.lambdef(nodelist[0]) 516 return self.com_binary(Or, nodelist) 517 old_test = or_test 518 519 def and_test(self, nodelist): 520 # not_test ('and' not_test)* 521 return self.com_binary(And, nodelist) 522 523 def not_test(self, nodelist): 524 # 'not' not_test | comparison 525 result = self.com_node(nodelist[-1]) 526 if len(nodelist) == 2: 527 return Not(result, lineno=nodelist[0][2]) 528 return result 529 530 def comparison(self, nodelist): 531 # comparison: expr (comp_op expr)* 532 node = self.com_node(nodelist[0]) 533 if len(nodelist) == 1: 534 return node 535 536 results = [] 537 for i in range(2, len(nodelist), 2): 538 nl = nodelist[i-1] 539 540 # comp_op: '<' | '>' | '=' | '>=' | '<=' | '<>' | '!=' | '==' 541 # | 'in' | 'not' 'in' | 'is' | 'is' 'not' 542 n = nl[1] 543 if n[0] == token["NAME"]: 544 type = n[1] 545 if len(nl) == 3: 546 if type == 'not': 547 type = 'not in' 548 else: 549 type = 'is not' 550 else: 551 type = _cmp_types[n[0]] 552 553 lineno = nl[1][2] 554 results.append((type, self.com_node(nodelist[i]))) 555 556 # we need a special "compare" node so that we can distinguish 557 # 3 < x < 5 from (3 < x) < 5 558 # the two have very different semantics and results (note that the 559 # latter form is always true) 560 561 return Compare(node, results, lineno=lineno) 562 563 def expr(self, nodelist): 564 # xor_expr ('|' xor_expr)* 565 return self.com_binary(Bitor, nodelist) 566 567 def xor_expr(self, nodelist): 568 # xor_expr ('^' xor_expr)* 569 return self.com_binary(Bitxor, nodelist) 570 571 def and_expr(self, nodelist): 572 # xor_expr ('&' xor_expr)* 573 return self.com_binary(Bitand, nodelist) 574 575 def shift_expr(self, nodelist): 576 # shift_expr ('<<'|'>>' shift_expr)* 577 node = self.com_node(nodelist[0]) 578 for i in range(2, len(nodelist), 2): 579 right = self.com_node(nodelist[i]) 580 if nodelist[i-1][0] == token["LEFTSHIFT"]: 581 node = LeftShift([node, right], lineno=nodelist[1][2]) 582 elif nodelist[i-1][0] == token["RIGHTSHIFT"]: 583 node = RightShift([node, right], lineno=nodelist[1][2]) 584 else: 585 raise ValueError, "unexpected token: %s" % nodelist[i-1][0] 586 return node 587 588 def arith_expr(self, nodelist): 589 node = self.com_node(nodelist[0]) 590 for i in range(2, len(nodelist), 2): 591 right = self.com_node(nodelist[i]) 592 if nodelist[i-1][0] == token["PLUS"]: 593 node = Add([node, right], lineno=nodelist[1][2]) 594 elif nodelist[i-1][0] == token["MINUS"]: 595 node = Sub([node, right], lineno=nodelist[1][2]) 596 else: 597 raise ValueError, "unexpected token: %s" % nodelist[i-1][0] 598 return node 599 600 def term(self, nodelist): 601 node = self.com_node(nodelist[0]) 602 for i in range(2, len(nodelist), 2): 603 right = self.com_node(nodelist[i]) 604 t = nodelist[i-1][0] 605 if t == token["STAR"]: 606 node = Mul([node, right]) 607 elif t == token["SLASH"]: 608 node = Div([node, right]) 609 elif t == token["PERCENT"]: 610 node = Mod([node, right]) 611 elif t == token["DOUBLESLASH"]: 612 node = FloorDiv([node, right]) 613 else: 614 raise ValueError, "unexpected token: %s" % t 615 node.lineno = nodelist[1][2] 616 return node 617 618 def factor(self, nodelist): 619 elt = nodelist[0] 620 t = elt[0] 621 node = self.lookup_node(nodelist[-1])(nodelist[-1][1:]) 622 # need to handle (unary op)constant here... 623 if t == token["PLUS"]: 624 return UnaryAdd(node, lineno=elt[2]) 625 elif t == token["MINUS"]: 626 return UnarySub(node, lineno=elt[2]) 627 elif t == token["TILDE"]: 628 node = Invert(node, lineno=elt[2]) 629 return node 630 631 def power(self, nodelist): 632 # power: atom trailer* ('**' factor)* 633 node = self.com_node(nodelist[0]) 634 for i in range(1, len(nodelist)): 635 elt = nodelist[i] 636 if elt[0] == token["DOUBLESTAR"]: 637 return Power([node, self.com_node(nodelist[i+1])], 638 lineno=elt[2]) 639 640 node = self.com_apply_trailer(node, elt) 641 642 return node 643 644 def atom(self, nodelist): 645 return self._atom_dispatch[nodelist[0][0]](nodelist) 646 647 def atom_lpar(self, nodelist): 648 if nodelist[1][0] == token["RPAR"]: 649 return Tuple((), lineno=nodelist[0][2]) 650 return self.com_node(nodelist[1]) 651 652 def atom_lsqb(self, nodelist): 653 if nodelist[1][0] == token["RSQB"]: 654 return List((), lineno=nodelist[0][2]) 655 return self.com_list_constructor(nodelist[1]) 656 657 def atom_lbrace(self, nodelist): 658 if nodelist[1][0] == token["RBRACE"]: 659 return Dict((), lineno=nodelist[0][2]) 660 return self.com_dictorsetmaker(nodelist[1]) 661 662 def atom_backquote(self, nodelist): 663 return Backquote(self.com_node(nodelist[1])) 664 665 def atom_number(self, nodelist): 666 ### need to verify this matches compile.c 667 k = eval(nodelist[0][1]) 668 return Const(k, nodelist[0][1], lineno=nodelist[0][2]) 669 670 def decode_literal(self, lit): 671 if self.encoding: 672 # this is particularly fragile & a bit of a 673 # hack... changes in compile.c:parsestr and 674 # tokenizer.c must be reflected here. 675 if self.encoding not in ['utf-8', 'iso-8859-1']: 676 lit = unicode(lit, 'utf-8').encode(self.encoding) 677 return eval("# coding: %s\n%s" % (self.encoding, lit)) 678 else: 679 return eval(lit) 680 681 def atom_string(self, nodelist): 682 k = '' 683 for node in nodelist: 684 k += self.decode_literal(node[1]) 685 return Const(k, node[1], lineno=nodelist[0][2]) 686 687 def atom_name(self, nodelist): 688 return Name(nodelist[0][1], lineno=nodelist[0][2]) 689 690 # -------------------------------------------------------------- 691 # 692 # INTERNAL PARSING UTILITIES 693 # 694 695 # The use of com_node() introduces a lot of extra stack frames, 696 # enough to cause a stack overflow compiling test.test_parser with 697 # the standard interpreter recursionlimit. The com_node() is a 698 # convenience function that hides the dispatch details, but comes 699 # at a very high cost. It is more efficient to dispatch directly 700 # in the callers. In these cases, use lookup_node() and call the 701 # dispatched node directly. 702 703 def lookup_node(self, node): 704 return self._dispatch[node[0]] 705 706 def com_node(self, node): 707 # Note: compile.c has handling in com_node for del_stmt, pass_stmt, 708 # break_stmt, stmt, small_stmt, flow_stmt, simple_stmt, 709 # and compound_stmt. 710 # We'll just dispatch them. 711 return self._dispatch[node[0]](node[1:]) 712 713 def com_NEWLINE(self, *args): 714 # A ';' at the end of a line can make a NEWLINE token appear 715 # here, Render it harmless. (genc discards ('discard', 716 # ('const', xxxx)) Nodes) 717 return Discard(Const(None)) 718 719 def com_arglist(self, nodelist): 720 # varargslist: 721 # (fpdef ['=' test] ',')* ('*' NAME [',' '**' NAME] | '**' NAME) 722 # | fpdef ['=' test] (',' fpdef ['=' test])* [','] 723 # fpdef: NAME | '(' fplist ')' 724 # fplist: fpdef (',' fpdef)* [','] 725 names = [] 726 defaults = [] 727 flags = 0 728 729 i = 0 730 while i < len(nodelist): 731 node = nodelist[i] 732 if node[0] == token["STAR"] or node[0] == token["DOUBLESTAR"]: 733 if node[0] == token["STAR"]: 734 node = nodelist[i+1] 735 if node[0] == token["NAME"]: 736 names.append(node[1]) 737 flags = flags | CO_VARARGS 738 i = i + 3 739 740 if i < len(nodelist): 741 # should be DOUBLESTAR 742 t = nodelist[i][0] 743 if t == token["DOUBLESTAR"]: 744 node = nodelist[i+1] 745 else: 746 raise ValueError, "unexpected token: %s" % t 747 names.append(node[1]) 748 flags = flags | CO_VARKEYWORDS 749 750 break 751 752 # fpdef: NAME | '(' fplist ')' 753 names.append(self.com_fpdef(node)) 754 755 i = i + 1 756 if i < len(nodelist) and nodelist[i][0] == token["EQUAL"]: 757 defaults.append(self.com_node(nodelist[i + 1])) 758 i = i + 2 759 elif len(defaults): 760 # we have already seen an argument with default, but here 761 # came one without 762 raise SyntaxError, "non-default argument follows default argument" 763 764 # skip the comma 765 i = i + 1 766 767 return names, defaults, flags 768 769 def com_fpdef(self, node): 770 # fpdef: NAME | '(' fplist ')' 771 if node[1][0] == token["LPAR"]: 772 return self.com_fplist(node[2]) 773 return node[1][1] 774 775 def com_fplist(self, node): 776 # fplist: fpdef (',' fpdef)* [','] 777 if len(node) == 2: 778 return self.com_fpdef(node[1]) 779 list = [] 780 for i in range(1, len(node), 2): 781 list.append(self.com_fpdef(node[i])) 782 return tuple(list) 783 784 def com_dotted_name(self, node): 785 # String together the dotted names and return the string 786 name = "" 787 for n in node: 788 if type(n) == type(()) and n[0] == 1: 789 name = name + n[1] + '.' 790 return name[:-1] 791 792 def com_dotted_as_name(self, node): 793 assert node[0] == symbol["dotted_as_name"] 794 node = node[1:] 795 dot = self.com_dotted_name(node[0][1:]) 796 if len(node) == 1: 797 return dot, None 798 assert node[1][1] == 'as' 799 assert node[2][0] == token["NAME"] 800 return dot, node[2][1] 801 802 def com_dotted_as_names(self, node): 803 assert node[0] == symbol["dotted_as_names"] 804 node = node[1:] 805 names = [self.com_dotted_as_name(node[0])] 806 for i in range(2, len(node), 2): 807 names.append(self.com_dotted_as_name(node[i])) 808 return names 809 810 def com_import_as_name(self, node): 811 assert node[0] == symbol["import_as_name"] 812 node = node[1:] 813 assert node[0][0] == token["NAME"] 814 if len(node) == 1: 815 return node[0][1], None 816 assert node[1][1] == 'as', node 817 assert node[2][0] == token["NAME"] 818 return node[0][1], node[2][1] 819 820 def com_import_as_names(self, node): 821 assert node[0] == symbol["import_as_names"] 822 node = node[1:] 823 names = [self.com_import_as_name(node[0])] 824 for i in range(2, len(node), 2): 825 names.append(self.com_import_as_name(node[i])) 826 return names 827 828 def com_bases(self, node): 829 bases = [] 830 for i in range(1, len(node), 2): 831 bases.append(self.com_node(node[i])) 832 return bases 833 834 def com_try_except_finally(self, nodelist): 835 # ('try' ':' suite 836 # ((except_clause ':' suite)+ ['else' ':' suite] ['finally' ':' suite] 837 # | 'finally' ':' suite)) 838 839 if nodelist[3][0] == token["NAME"]: 840 # first clause is a finally clause: only try-finally 841 return TryFinally(self.com_node(nodelist[2]), 842 self.com_node(nodelist[5]), 843 lineno=nodelist[0][2]) 844 845 #tryexcept: [TryNode, [except_clauses], elseNode)] 846 clauses = [] 847 elseNode = None 848 finallyNode = None 849 for i in range(3, len(nodelist), 3): 850 node = nodelist[i] 851 if node[0] == symbol["except_clause"]: 852 # except_clause: 'except' [expr [(',' | 'as') expr]] */ 853 if len(node) > 2: 854 expr1 = self.com_node(node[2]) 855 if len(node) > 4: 856 expr2 = self.com_assign(node[4], OP_ASSIGN) 857 else: 858 expr2 = None 859 else: 860 expr1 = expr2 = None 861 clauses.append((expr1, expr2, self.com_node(nodelist[i+2]))) 862 863 if node[0] == token["NAME"]: 864 if node[1] == 'else': 865 elseNode = self.com_node(nodelist[i+2]) 866 elif node[1] == 'finally': 867 finallyNode = self.com_node(nodelist[i+2]) 868 try_except = TryExcept(self.com_node(nodelist[2]), clauses, elseNode, 869 lineno=nodelist[0][2]) 870 if finallyNode: 871 return TryFinally(try_except, finallyNode, lineno=nodelist[0][2]) 872 else: 873 return try_except 874 875 def com_with(self, nodelist): 876 # with_stmt: 'with' with_item (',' with_item)* ':' suite 877 body = self.com_node(nodelist[-1]) 878 for i in range(len(nodelist) - 3, 0, -2): 879 ret = self.com_with_item(nodelist[i], body, nodelist[0][2]) 880 if i == 1: 881 return ret 882 body = ret 883 884 def com_with_item(self, nodelist, body, lineno): 885 # with_item: test ['as' expr] 886 if len(nodelist) == 4: 887 var = self.com_assign(nodelist[3], OP_ASSIGN) 888 else: 889 var = None 890 expr = self.com_node(nodelist[1]) 891 return With(expr, var, body, lineno=lineno) 892 893 def com_augassign_op(self, node): 894 assert node[0] == symbol["augassign"] 895 return node[1] 896 897 def com_augassign(self, node): 898 """Return node suitable for lvalue of augmented assignment 899 900 Names, slices, and attributes are the only allowable nodes. 901 """ 902 l = self.com_node(node) 903 if l.__class__ in (Name, Slice, Subscript, Getattr): 904 return l 905 raise SyntaxError, "can't assign to %s" % l.__class__.__name__ 906 907 def com_assign(self, node, assigning): 908 # return a node suitable for use as an "lvalue" 909 # loop to avoid trivial recursion 910 while 1: 911 t = node[0] 912 if t in (symbol["exprlist"], symbol["testlist"], symbol["testlist_safe"], symbol["testlist_comp"]): 913 if len(node) > 2: 914 return self.com_assign_tuple(node, assigning) 915 node = node[1] 916 elif t in _assign_types: 917 if len(node) > 2: 918 raise SyntaxError, "can't assign to operator" 919 node = node[1] 920 elif t == symbol["power"]: 921 if node[1][0] != symbol["atom"]: 922 raise SyntaxError, "can't assign to operator" 923 if len(node) > 2: 924 primary = self.com_node(node[1]) 925 for i in range(2, len(node)-1): 926 ch = node[i] 927 if ch[0] == token["DOUBLESTAR"]: 928 raise SyntaxError, "can't assign to operator" 929 primary = self.com_apply_trailer(primary, ch) 930 return self.com_assign_trailer(primary, node[-1], 931 assigning) 932 node = node[1] 933 elif t == symbol["atom"]: 934 t = node[1][0] 935 if t == token["LPAR"]: 936 node = node[2] 937 if node[0] == token["RPAR"]: 938 raise SyntaxError, "can't assign to ()" 939 elif t == token["LSQB"]: 940 node = node[2] 941 if node[0] == token["RSQB"]: 942 raise SyntaxError, "can't assign to []" 943 return self.com_assign_list(node, assigning) 944 elif t == token["NAME"]: 945 return self.com_assign_name(node[1], assigning) 946 else: 947 raise SyntaxError, "can't assign to literal" 948 else: 949 raise SyntaxError, "bad assignment (%s)" % t 950 951 def com_assign_tuple(self, node, assigning): 952 assigns = [] 953 for i in range(1, len(node), 2): 954 assigns.append(self.com_assign(node[i], assigning)) 955 return AssTuple(assigns, lineno=extractLineNo(node)) 956 957 def com_assign_list(self, node, assigning): 958 assigns = [] 959 for i in range(1, len(node), 2): 960 if i + 1 < len(node): 961 if node[i + 1][0] == symbol["list_for"]: 962 raise SyntaxError, "can't assign to list comprehension" 963 assert node[i + 1][0] == token["COMMA"], node[i + 1] 964 assigns.append(self.com_assign(node[i], assigning)) 965 return AssList(assigns, lineno=extractLineNo(node)) 966 967 def com_assign_name(self, node, assigning): 968 return AssName(node[1], assigning, lineno=node[2]) 969 970 def com_assign_trailer(self, primary, node, assigning): 971 t = node[1][0] 972 if t == token["DOT"]: 973 return self.com_assign_attr(primary, node[2], assigning) 974 if t == token["LSQB"]: 975 return self.com_subscriptlist(primary, node[2], assigning) 976 if t == token["LPAR"]: 977 raise SyntaxError, "can't assign to function call" 978 raise SyntaxError, "unknown trailer type: %s" % t 979 980 def com_assign_attr(self, primary, node, assigning): 981 return AssAttr(primary, node[1], assigning, lineno=node[-1]) 982 983 def com_binary(self, constructor, nodelist): 984 "Compile 'NODE (OP NODE)*' into (type, [ node1, ..., nodeN ])." 985 l = len(nodelist) 986 if l == 1: 987 n = nodelist[0] 988 return self.lookup_node(n)(n[1:]) 989 items = [] 990 for i in range(0, l, 2): 991 n = nodelist[i] 992 items.append(self.lookup_node(n)(n[1:])) 993 return constructor(items, lineno=extractLineNo(nodelist)) 994 995 def com_stmt(self, node): 996 result = self.lookup_node(node)(node[1:]) 997 assert result is not None 998 if isinstance(result, Stmt): 999 return result 1000 return Stmt([result]) 1001 1002 def com_append_stmt(self, stmts, node): 1003 result = self.lookup_node(node)(node[1:]) 1004 assert result is not None 1005 if isinstance(result, Stmt): 1006 stmts.extend(result.nodes) 1007 else: 1008 stmts.append(result) 1009 1010 def com_list_constructor(self, nodelist): 1011 # listmaker: test ( (',' test)* [','] ) 1012 values = [] 1013 for i in range(1, len(nodelist)): 1014 if nodelist[i][0] == token["COMMA"]: 1015 continue 1016 values.append(self.com_node(nodelist[i])) 1017 return List(values, lineno=values[0].lineno) 1018 1019 def com_dictorsetmaker(self, nodelist): 1020 # dictorsetmaker: ( (test ':' test ( (',' test ':' test)* [','])) | 1021 # (test ( (',' test)* [','])) ) 1022 assert nodelist[0] == symbol["dictorsetmaker"] 1023 nodelist = nodelist[1:] 1024 if len(nodelist) == 1 or nodelist[1][0] == token["COMMA"]: 1025 # set literal 1026 items = [] 1027 for i in range(0, len(nodelist), 2): 1028 items.append(self.com_node(nodelist[i])) 1029 return Set(items, lineno=items[0].lineno) 1030 else: 1031 # dict literal 1032 items = [] 1033 for i in range(0, len(nodelist), 4): 1034 items.append((self.com_node(nodelist[i]), 1035 self.com_node(nodelist[i+2]))) 1036 return Dict(items, lineno=items[0][0].lineno) 1037 1038 def com_apply_trailer(self, primaryNode, nodelist): 1039 t = nodelist[1][0] 1040 if t == token["LPAR"]: 1041 return self.com_call_function(primaryNode, nodelist[2]) 1042 if t == token["DOT"]: 1043 return self.com_select_member(primaryNode, nodelist[2]) 1044 if t == token["LSQB"]: 1045 return self.com_subscriptlist(primaryNode, nodelist[2], OP_APPLY) 1046 1047 raise SyntaxError, 'unknown node type: %s' % t 1048 1049 def com_select_member(self, primaryNode, nodelist): 1050 if nodelist[0] != token["NAME"]: 1051 raise SyntaxError, "member must be a name" 1052 return Getattr(primaryNode, nodelist[1], lineno=nodelist[2]) 1053 1054 def com_call_function(self, primaryNode, nodelist): 1055 if nodelist[0] == token["RPAR"]: 1056 return CallFunc(primaryNode, [], lineno=extractLineNo(nodelist)) 1057 args = [] 1058 kw = 0 1059 star_node = dstar_node = None 1060 len_nodelist = len(nodelist) 1061 i = 1 1062 while i < len_nodelist: 1063 node = nodelist[i] 1064 1065 if node[0]==token["STAR"]: 1066 if star_node is not None: 1067 raise SyntaxError, 'already have the varargs indentifier' 1068 star_node = self.com_node(nodelist[i+1]) 1069 i = i + 3 1070 continue 1071 elif node[0]==token["DOUBLESTAR"]: 1072 if dstar_node is not None: 1073 raise SyntaxError, 'already have the kwargs indentifier' 1074 dstar_node = self.com_node(nodelist[i+1]) 1075 i = i + 3 1076 continue 1077 1078 # positional or named parameters 1079 kw, result = self.com_argument(node, kw, star_node) 1080 1081 args.append(result) 1082 i = i + 2 1083 1084 return CallFunc(primaryNode, args, star_node, dstar_node, 1085 lineno=extractLineNo(nodelist)) 1086 1087 def com_argument(self, nodelist, kw, star_node): 1088 if len(nodelist) == 2: 1089 if kw: 1090 raise SyntaxError, "non-keyword arg after keyword arg" 1091 if star_node: 1092 raise SyntaxError, "only named arguments may follow *expression" 1093 return 0, self.com_node(nodelist[1]) 1094 result = self.com_node(nodelist[3]) 1095 n = nodelist[1] 1096 while len(n) == 2 and n[0] != token["NAME"]: 1097 n = n[1] 1098 if n[0] != token["NAME"]: 1099 raise SyntaxError, "keyword can't be an expression (%s)"%n[0] 1100 node = Keyword(n[1], result, lineno=n[2]) 1101 return 1, node 1102 1103 def com_subscriptlist(self, primary, nodelist, assigning): 1104 # slicing: simple_slicing | extended_slicing 1105 # simple_slicing: primary "[" short_slice "]" 1106 # extended_slicing: primary "[" slice_list "]" 1107 # slice_list: slice_item ("," slice_item)* [","] 1108 1109 # backwards compat slice for '[i:j]' 1110 if len(nodelist) == 2: 1111 sub = nodelist[1] 1112 if (sub[1][0] == token["COLON"] or \ 1113 (len(sub) > 2 and sub[2][0] == token["COLON"])) and \ 1114 sub[-1][0] != symbol["sliceop"]: 1115 return self.com_slice(primary, sub, assigning) 1116 1117 subscripts = [] 1118 for i in range(1, len(nodelist), 2): 1119 subscripts.append(self.com_subscript(nodelist[i])) 1120 return Subscript(primary, assigning, subscripts, 1121 lineno=extractLineNo(nodelist)) 1122 1123 def com_subscript(self, node): 1124 # slice_item: expression | proper_slice | ellipsis 1125 ch = node[1] 1126 t = ch[0] 1127 if t == token["DOT"] and node[2][0] == token["DOT"]: 1128 return Ellipsis() 1129 if t == token["COLON"] or len(node) > 2: 1130 return self.com_sliceobj(node) 1131 return self.com_node(ch) 1132 1133 def com_sliceobj(self, node): 1134 # proper_slice: short_slice | long_slice 1135 # short_slice: [lower_bound] ":" [upper_bound] 1136 # long_slice: short_slice ":" [stride] 1137 # lower_bound: expression 1138 # upper_bound: expression 1139 # stride: expression 1140 # 1141 # Note: a stride may be further slicing... 1142 1143 items = [] 1144 1145 if node[1][0] == token["COLON"]: 1146 items.append(Const(None)) 1147 i = 2 1148 else: 1149 items.append(self.com_node(node[1])) 1150 # i == 2 is a COLON 1151 i = 3 1152 1153 if i < len(node) and node[i][0] == symbol["test"]: 1154 items.append(self.com_node(node[i])) 1155 i = i + 1 1156 else: 1157 items.append(Const(None)) 1158 1159 # a short_slice has been built. look for long_slice now by looking 1160 # for strides... 1161 for j in range(i, len(node)): 1162 ch = node[j] 1163 if len(ch) == 2: 1164 items.append(Const(None)) 1165 else: 1166 items.append(self.com_node(ch[2])) 1167 return Sliceobj(items, lineno=extractLineNo(node)) 1168 1169 def com_slice(self, primary, node, assigning): 1170 # short_slice: [lower_bound] ":" [upper_bound] 1171 lower = upper = None 1172 if len(node) == 3: 1173 if node[1][0] == token["COLON"]: 1174 upper = self.com_node(node[2]) 1175 else: 1176 lower = self.com_node(node[1]) 1177 elif len(node) == 4: 1178 lower = self.com_node(node[1]) 1179 upper = self.com_node(node[3]) 1180 return Slice(primary, assigning, lower, upper, 1181 lineno=extractLineNo(node)) 1182 1183 def get_docstring(self, node, n=None): 1184 if n is None: 1185 n = node[0] 1186 node = node[1:] 1187 if n == symbol["suite"]: 1188 if len(node) == 1: 1189 return self.get_docstring(node[0]) 1190 for sub in node: 1191 if sub[0] == symbol["stmt"]: 1192 return self.get_docstring(sub) 1193 return None 1194 if n == symbol["file_input"]: 1195 for sub in node: 1196 if sub[0] == symbol["stmt"]: 1197 return self.get_docstring(sub) 1198 return None 1199 if n == symbol["atom"]: 1200 if node[0][0] == token["STRING"]: 1201 s = '' 1202 for t in node: 1203 s = s + eval(t[1]) 1204 return s 1205 return None 1206 if n == symbol["stmt"] or n == symbol["simple_stmt"] \ 1207 or n == symbol["small_stmt"]: 1208 return self.get_docstring(node[0]) 1209 if n in _doc_nodes and len(node) == 1: 1210 return self.get_docstring(node[0]) 1211 return None 1212 1213 1214 _doc_nodes = [ 1215 symbol["expr_stmt"], 1216 symbol["testlist"], 1217 symbol["testlist_safe"], 1218 symbol["test"], 1219 symbol["or_test"], 1220 symbol["and_test"], 1221 symbol["not_test"], 1222 symbol["comparison"], 1223 symbol["expr"], 1224 symbol["xor_expr"], 1225 symbol["and_expr"], 1226 symbol["shift_expr"], 1227 symbol["arith_expr"], 1228 symbol["term"], 1229 symbol["factor"], 1230 symbol["power"], 1231 ] 1232 1233 # comp_op: '<' | '>' | '=' | '>=' | '<=' | '<>' | '!=' | '==' 1234 # | 'in' | 'not' 'in' | 'is' | 'is' 'not' 1235 _cmp_types = { 1236 token["LESS"] : '<', 1237 token["GREATER"] : '>', 1238 token["EQEQUAL"] : '==', 1239 token["EQUAL"] : '==', 1240 token["LESSEQUAL"] : '<=', 1241 token["GREATEREQUAL"] : '>=', 1242 token["NOTEQUAL"] : '!=', 1243 } 1244 1245 _legal_node_types = [ 1246 symbol["funcdef"], 1247 symbol["classdef"], 1248 symbol["stmt"], 1249 symbol["small_stmt"], 1250 symbol["flow_stmt"], 1251 symbol["simple_stmt"], 1252 symbol["compound_stmt"], 1253 symbol["expr_stmt"], 1254 symbol["print_stmt"], 1255 symbol["del_stmt"], 1256 symbol["pass_stmt"], 1257 symbol["break_stmt"], 1258 symbol["continue_stmt"], 1259 symbol["return_stmt"], 1260 symbol["raise_stmt"], 1261 symbol["import_stmt"], 1262 symbol["global_stmt"], 1263 symbol["exec_stmt"], 1264 symbol["assert_stmt"], 1265 symbol["if_stmt"], 1266 symbol["while_stmt"], 1267 symbol["for_stmt"], 1268 symbol["try_stmt"], 1269 symbol["with_stmt"], 1270 symbol["suite"], 1271 symbol["testlist"], 1272 symbol["testlist_safe"], 1273 symbol["test"], 1274 symbol["and_test"], 1275 symbol["not_test"], 1276 symbol["comparison"], 1277 symbol["exprlist"], 1278 symbol["expr"], 1279 symbol["xor_expr"], 1280 symbol["and_expr"], 1281 symbol["shift_expr"], 1282 symbol["arith_expr"], 1283 symbol["term"], 1284 symbol["factor"], 1285 symbol["power"], 1286 symbol["atom"], 1287 ] 1288 1289 _assign_types = [ 1290 symbol["test"], 1291 symbol["or_test"], 1292 symbol["and_test"], 1293 symbol["not_test"], 1294 symbol["comparison"], 1295 symbol["expr"], 1296 symbol["xor_expr"], 1297 symbol["and_expr"], 1298 symbol["shift_expr"], 1299 symbol["arith_expr"], 1300 symbol["term"], 1301 symbol["factor"], 1302 ] 1303 1304 _names = {} 1305 for k, v in sym_name.items(): 1306 _names[k] = v 1307 for k, v in tok_name.items(): 1308 _names[k] = v 1309 1310 def debug_tree(tree): 1311 l = [] 1312 for elt in tree: 1313 if isinstance(elt, int): 1314 l.append(_names.get(elt, elt)) 1315 elif isinstance(elt, str): 1316 l.append(elt) 1317 else: 1318 l.append(debug_tree(elt)) 1319 return l