# Copyright (c) 2015-2018 Rocky Bernstein # Copyright (c) 2005 by Dan Pascu # Copyright (c) 2000-2002 by hartmut Goebel # Copyright (c) 1999 John Aycock # # This program is free software: you can redistribute it and/or modify # it under the terms of the GNU General Public License as published by # the Free Software Foundation, either version 3 of the License, or # (at your option) any later version. # # This program is distributed in the hope that it will be useful, # but WITHOUT ANY WARRANTY; without even the implied warranty of # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the # GNU General Public License for more details. # # You should have received a copy of the GNU General Public License # along with this program. If not, see . """ Common uncompyle6 parser routines. """ from __future__ import print_function import sys from xdis.code import iscode from xdis.magics import py_str2float from spark_parser import GenericASTBuilder, DEFAULT_DEBUG as PARSER_DEFAULT_DEBUG from uncompyle6.show import maybe_show_asm class ParserError(Exception): def __init__(self, token, offset): self.token = token self.offset = offset def __str__(self): return "Parse error at or near `%r' instruction at offset %s\n" % \ (self.token, self.offset) nop_func = lambda self, args: None class PythonParser(GenericASTBuilder): def __init__(self, AST, start, debug): super(PythonParser, self).__init__(AST, start, debug) # FIXME: customize per python parser version nt_list = [ 'stmts', 'except_stmts', '_stmts', 'attributes', 'exprlist', 'kvlist', 'kwargs', 'come_froms', '_come_froms', 'importlist', # Python < 3 'print_items', # PyPy: 'imports_cont', 'kvlist_n', # Python 3.6+ 'joined_str', 'come_from_loops', ] self.collect = frozenset(nt_list) # For these items we need to keep the 1st epslion reduction since # the nonterminal name is used in a semantic action. self.keep_epsilon = frozenset(('kvlist_n', 'kvlist')) # ??? Do we need a debug option to skip eliding singleton reductions? # Time will tell if it if useful in debugging # FIXME: optional_nt is a misnomer. It's really about there being a # singleton reduction that we can simplify. It also happens to be optional # in its other derivation self.optional_nt |= frozenset(('come_froms', 'suite_stmts', 'l_stmts_opt', 'c_stmts_opt')) # Reduce singleton reductions in these nonterminals: # FIXME: would love to do expr, sstmts, stmts and # so on but that would require major changes to the # semantic actions self.singleton = frozenset(('str', 'joined_str', 'store', '_stmts', 'suite_stmts_opt', 'inplace_op')) # Instructions filled in from scanner self.insts = [] def ast_first_offset(self, ast): if hasattr(ast, 'offset'): return ast.offset else: return self.ast_first_offset(ast[0]) def add_unique_rule(self, rule, opname, arg_count, customize): """Add rule to grammar, but only if it hasn't been added previously opname and stack_count are used in the customize() semantic the actions to add the semantic action rule. Stack_count is used in custom opcodes like MAKE_FUNCTION to indicate how many arguments it has. Often it is not used. """ if rule not in self.new_rules: # print("XXX ", rule) # debug self.new_rules.add(rule) self.addRule(rule, nop_func) customize[opname] = arg_count pass return def add_unique_rules(self, rules, customize): """Add rules (a list of string) to grammar. Note that the rules must not be those that set arg_count in the custom dictionary. """ for rule in rules: if len(rule) == 0: continue opname = rule.split('::=')[0].strip() self.add_unique_rule(rule, opname, 0, customize) return def add_unique_doc_rules(self, rules_str, customize): """Add rules (a docstring-like list of rules) to grammar. Note that the rules must not be those that set arg_count in the custom dictionary. """ rules = [r.strip() for r in rules_str.split("\n")] self.add_unique_rules(rules, customize) return def cleanup(self): """ Remove recursive references to allow garbage collector to collect this object. """ for dict in (self.rule2func, self.rules, self.rule2name): for i in list(dict.keys()): dict[i] = None for i in dir(self): setattr(self, i, None) def debug_reduce(self, rule, tokens, parent, last_token_pos): """Customized format and print for our kind of tokens which gets called in debugging grammar reduce rules """ def fix(c): s = str(c) last_token_pos = s.find('_') if last_token_pos == -1: return s else: return s[:last_token_pos] prefix = '' if parent and tokens: p_token = tokens[parent] if hasattr(p_token, 'linestart') and p_token.linestart: prefix = 'L.%3d: ' % p_token.linestart else: prefix = ' ' if hasattr(p_token, 'offset'): prefix += "%3s" % fix(p_token.offset) if len(rule[1]) > 1: prefix += '-%-3s ' % fix(tokens[last_token_pos-1].offset) else: prefix += ' ' else: prefix = ' ' print("%s%s ::= %s (%d)" % (prefix, rule[0], ' '.join(rule[1]), last_token_pos)) def error(self, instructions, index): # Find the last line boundary start, finish = -1, -1 for start in range(index, -1, -1): if instructions[start].linestart: break pass for finish in range(index+1, len(instructions)): if instructions[finish].linestart: break pass if start > 0: err_token = instructions[index] print("Instruction context:") for i in range(start, finish): if i != index: indent = ' ' else: indent = '-> ' print("%s%s" % (indent, instructions[i])) raise ParserError(err_token, err_token.offset) else: raise ParserError(None, -1) def get_pos_kw(self, token): """Return then the number of positional parameters and represented by the attr field of token""" # Low byte indicates number of positional paramters, # high byte number of keyword parameters args_pos = token.attr & 0xff args_kw = (token.attr >> 8) & 0xff return args_pos, args_kw def nonterminal(self, nt, args): n = len(args) # # Use this to find lots of singleton rule # if n == 1 and nt not in self.singleton: # print("XXX", nt) if nt in self.collect and n > 1: # # Collect iterated thingies together. That is rather than # stmts -> stmts stmt -> stmts stmt -> ... # stmms -> stmt stmt ... # if not hasattr(args[0], 'append'): # Was in self.optional_nt as a single item, but we find we have # more than one now... rv = GenericASTBuilder.nonterminal(self, nt, [args[0]]) else: rv = args[0] pass # In a list-like entity where the first item goes to epsilon, # drop that and save the 2nd item as the first one if len(rv) == 0 and nt not in self.keep_epsilon: rv = args[1] else: rv.append(args[1]) elif n == 1 and args[0] in self.singleton: rv = GenericASTBuilder.nonterminal(self, nt, args[0]) del args[0] # save memory elif n == 1 and nt in self.optional_nt: rv = args[0] else: rv = GenericASTBuilder.nonterminal(self, nt, args) return rv def __ambiguity(self, children): # only for debugging! to be removed hG/2000-10-15 print(children) return GenericASTBuilder.ambiguity(self, children) def resolve(self, list): if len(list) == 2 and 'function_def' in list and 'assign' in list: return 'function_def' if 'grammar' in list and 'expr' in list: return 'expr' # print >> sys.stderr, 'resolve', str(list) return GenericASTBuilder.resolve(self, list) ############################################### # Common Python 2 and Python 3 grammar rules # ############################################### def p_start(self, args): ''' # The start or goal symbol stmts ::= sstmt+ ''' def p_call_stmt(self, args): ''' # eval-mode compilation. Single-mode interactive compilation # adds another rule. call_stmt ::= expr POP_TOP ''' def p_stmt(self, args): """ pass ::= _stmts ::= stmt+ # statements with continue c_stmts ::= _stmts c_stmts ::= _stmts lastc_stmt c_stmts ::= lastc_stmt c_stmts ::= continues lastc_stmt ::= iflaststmt lastc_stmt ::= forelselaststmt lastc_stmt ::= ifelsestmtc c_stmts_opt ::= c_stmts c_stmts_opt ::= pass # statements inside a loop l_stmts ::= _stmts l_stmts ::= returns l_stmts ::= continues l_stmts ::= _stmts lastl_stmt l_stmts ::= lastl_stmt lastl_stmt ::= iflaststmtl lastl_stmt ::= ifelsestmtl lastl_stmt ::= forelselaststmtl lastl_stmt ::= tryelsestmtl l_stmts_opt ::= l_stmts l_stmts_opt ::= pass suite_stmts ::= _stmts suite_stmts ::= returns suite_stmts ::= continues suite_stmts_opt ::= suite_stmts # passtmt is needed for semantic actions to add "pass" suite_stmts_opt ::= pass else_suite ::= suite_stmts else_suitel ::= l_stmts else_suitec ::= c_stmts else_suitec ::= returns stmt ::= assert stmt ::= classdef stmt ::= call_stmt stmt ::= ifstmt stmt ::= ifelsestmt stmt ::= whilestmt stmt ::= while1stmt stmt ::= whileelsestmt stmt ::= while1elsestmt stmt ::= for stmt ::= forelsestmt stmt ::= try_except stmt ::= tryelsestmt stmt ::= tryfinallystmt stmt ::= withstmt stmt ::= withasstmt stmt ::= del_stmt del_stmt ::= DELETE_FAST del_stmt ::= DELETE_NAME del_stmt ::= DELETE_GLOBAL stmt ::= return return ::= ret_expr RETURN_VALUE # "returns" nonterminal is a sequence of statements that ends in a RETURN statement. # In later Python versions with jump optimization, this can cause JUMPs # that would normally appear to be omitted. returns ::= return returns ::= _stmts return """ pass def p_function_def(self, args): ''' stmt ::= function_def function_def ::= mkfunc store stmt ::= function_def_deco function_def_deco ::= mkfuncdeco store mkfuncdeco ::= expr mkfuncdeco CALL_FUNCTION_1 mkfuncdeco ::= expr mkfuncdeco0 CALL_FUNCTION_1 mkfuncdeco0 ::= mkfunc load_closure ::= load_closure LOAD_CLOSURE load_closure ::= LOAD_CLOSURE ''' def p_generator_exp(self, args): ''' expr ::= generator_exp stmt ::= genexpr_func genexpr_func ::= LOAD_FAST FOR_ITER store comp_iter JUMP_BACK ''' def p_jump(self, args): """ _jump ::= JUMP_ABSOLUTE _jump ::= JUMP_FORWARD _jump ::= JUMP_BACK # Zero or more COME_FROMs - loops can have this _come_froms ::= COME_FROM* # One or more COME_FROMs - joins of tryelse's have this come_froms ::= COME_FROM+ # Zero or one COME_FROM # And/or expressions have this come_from_opt ::= COME_FROM? """ def p_augmented_assign(self, args): ''' stmt ::= aug_assign1 stmt ::= aug_assign2 # This is odd in that other aug_assign1's have only 3 slots # The store isn't used as that's supposed to be also # indicated in the first expr aug_assign1 ::= expr expr inplace_op store aug_assign1 ::= expr expr inplace_op ROT_THREE STORE_SUBSCR aug_assign2 ::= expr DUP_TOP LOAD_ATTR expr inplace_op ROT_TWO STORE_ATTR inplace_op ::= INPLACE_ADD inplace_op ::= INPLACE_SUBTRACT inplace_op ::= INPLACE_MULTIPLY inplace_op ::= INPLACE_TRUE_DIVIDE inplace_op ::= INPLACE_FLOOR_DIVIDE inplace_op ::= INPLACE_MODULO inplace_op ::= INPLACE_POWER inplace_op ::= INPLACE_LSHIFT inplace_op ::= INPLACE_RSHIFT inplace_op ::= INPLACE_AND inplace_op ::= INPLACE_XOR inplace_op ::= INPLACE_OR ''' def p_assign(self, args): ''' stmt ::= assign assign ::= expr DUP_TOP designList assign ::= expr store stmt ::= assign2 stmt ::= assign3 assign2 ::= expr expr ROT_TWO store store assign3 ::= expr expr expr ROT_THREE ROT_TWO store store store ''' def p_forstmt(self, args): """ for_iter ::= GET_ITER FOR_ITER for_block ::= l_stmts_opt _come_froms JUMP_BACK for ::= SETUP_LOOP expr for_iter store for_block POP_BLOCK _come_froms forelsestmt ::= SETUP_LOOP expr for_iter store for_block POP_BLOCK else_suite _come_froms forelselaststmt ::= SETUP_LOOP expr for_iter store for_block POP_BLOCK else_suitec _come_froms forelselaststmtl ::= SETUP_LOOP expr for_iter store for_block POP_BLOCK else_suitel _come_froms """ def p_import20(self, args): """ stmt ::= import stmt ::= import_from stmt ::= import_from_star stmt ::= importmultiple importlist ::= importlist alias importlist ::= alias alias ::= IMPORT_NAME store alias ::= IMPORT_FROM store alias ::= IMPORT_NAME attributes store import ::= LOAD_CONST LOAD_CONST alias import_from_star ::= LOAD_CONST LOAD_CONST IMPORT_NAME IMPORT_STAR import_from ::= LOAD_CONST LOAD_CONST IMPORT_NAME importlist POP_TOP importmultiple ::= LOAD_CONST LOAD_CONST alias imports_cont imports_cont ::= import_cont+ import_cont ::= LOAD_CONST LOAD_CONST alias attributes ::= LOAD_ATTR+ """ def p_list_comprehension(self, args): """ expr ::= list_comp list_iter ::= list_for list_iter ::= list_if list_iter ::= list_if_not list_iter ::= lc_body list_if ::= expr jmp_false list_iter list_if_not ::= expr jmp_true list_iter """ def p_set_comp(self, args): """ comp_iter ::= comp_for comp_iter ::= comp_body comp_body ::= gen_comp_body gen_comp_body ::= expr YIELD_VALUE POP_TOP comp_iter ::= comp_if comp_if ::= expr jmp_false comp_iter """ def p_expr(self, args): ''' expr ::= _mklambda expr ::= LOAD_FAST expr ::= LOAD_NAME expr ::= LOAD_CONST expr ::= LOAD_GLOBAL expr ::= LOAD_DEREF expr ::= binary_expr expr ::= list expr ::= compare expr ::= dict expr ::= and expr ::= or expr ::= unary_expr expr ::= call expr ::= unary_not expr ::= subscript expr ::= subscript2 expr ::= yield binary_expr ::= expr expr binary_op binary_op ::= BINARY_ADD binary_op ::= BINARY_MULTIPLY binary_op ::= BINARY_AND binary_op ::= BINARY_OR binary_op ::= BINARY_XOR binary_op ::= BINARY_SUBTRACT binary_op ::= BINARY_TRUE_DIVIDE binary_op ::= BINARY_FLOOR_DIVIDE binary_op ::= BINARY_MODULO binary_op ::= BINARY_LSHIFT binary_op ::= BINARY_RSHIFT binary_op ::= BINARY_POWER unary_expr ::= expr unary_op unary_op ::= UNARY_POSITIVE unary_op ::= UNARY_NEGATIVE unary_op ::= UNARY_INVERT unary_not ::= expr UNARY_NOT subscript ::= expr expr BINARY_SUBSCR attribute ::= expr LOAD_ATTR get_iter ::= expr GET_ITER yield ::= expr YIELD_VALUE _mklambda ::= mklambda expr ::= conditional ret_expr ::= expr ret_expr ::= ret_and ret_expr ::= ret_or ret_expr_or_cond ::= ret_expr ret_expr_or_cond ::= ret_cond stmt ::= return_lambda return_lambda ::= ret_expr RETURN_VALUE_LAMBDA LAMBDA_MARKER return_lambda ::= ret_expr RETURN_VALUE_LAMBDA compare ::= compare_chained compare ::= compare_single compare_single ::= expr expr COMPARE_OP # A compare_chained is two comparisions like x <= y <= z compare_chained ::= expr compare_chained1 ROT_TWO POP_TOP _come_froms compare_chained2 ::= expr COMPARE_OP JUMP_FORWARD # Non-null kvlist items are broken out in the indiviual grammars kvlist ::= # Positional arguments in make_function pos_arg ::= expr expr32 ::= expr expr expr expr expr expr expr expr expr expr expr expr expr expr expr expr expr expr expr expr expr expr expr expr expr expr expr expr expr expr expr expr expr1024 ::= expr32 expr32 expr32 expr32 expr32 expr32 expr32 expr32 expr32 expr32 expr32 expr32 expr32 expr32 expr32 expr32 expr32 expr32 expr32 expr32 expr32 expr32 expr32 expr32 expr32 expr32 expr32 expr32 expr32 expr32 expr32 expr32 ''' def p_store(self, args): ''' # Note. The below is right-recursive: designList ::= store store designList ::= store DUP_TOP designList ## Can we replace with left-recursive, and redo with: ## ## designList ::= designLists store store ## designLists ::= designLists store DUP_TOP ## designLists ::= ## Will need to redo semantic actiion store ::= STORE_FAST store ::= STORE_NAME store ::= STORE_GLOBAL store ::= STORE_DEREF store ::= expr STORE_ATTR store ::= store_subscr store_subscr ::= expr expr STORE_SUBSCR store ::= unpack ''' def parse(p, tokens, customize): p.customize_grammar_rules(tokens, customize) ast = p.parse(tokens) # p.cleanup() return ast def get_python_parser( version, debug_parser=PARSER_DEFAULT_DEBUG, compile_mode='exec', is_pypy = False): """Returns parser object for Python version 2 or 3, 3.2, 3.5on, etc., depending on the parameters passed. *compile_mode* is either 'exec', 'eval', or 'single'. See https://docs.python.org/3.6/library/functions.html#compile for an explanation of the different modes. """ # If version is a string, turn that into the corresponding float. if isinstance(version, str): version = py_str2float(version) # FIXME: there has to be a better way... # We could do this as a table lookup, but that would force us # in import all of the parsers all of the time. Perhaps there is # a lazy way of doing the import? if version < 3.0: if version == 1.5: import uncompyle6.parsers.parse15 as parse15 if compile_mode == 'exec': p = parse15.Python15Parser(debug_parser) else: p = parse15.Python15ParserSingle(debug_parser) elif version == 2.1: import uncompyle6.parsers.parse21 as parse21 if compile_mode == 'exec': p = parse21.Python21Parser(debug_parser) else: p = parse21.Python21ParserSingle(debug_parser) elif version == 2.2: import uncompyle6.parsers.parse22 as parse22 if compile_mode == 'exec': p = parse22.Python22Parser(debug_parser) else: p = parse22.Python22ParserSingle(debug_parser) elif version == 2.3: import uncompyle6.parsers.parse23 as parse23 if compile_mode == 'exec': p = parse23.Python23Parser(debug_parser) else: p = parse23.Python23ParserSingle(debug_parser) elif version == 2.4: import uncompyle6.parsers.parse24 as parse24 if compile_mode == 'exec': p = parse24.Python24Parser(debug_parser) else: p = parse24.Python24ParserSingle(debug_parser) elif version == 2.5: import uncompyle6.parsers.parse25 as parse25 if compile_mode == 'exec': p = parse25.Python25Parser(debug_parser) else: p = parse25.Python25ParserSingle(debug_parser) elif version == 2.6: import uncompyle6.parsers.parse26 as parse26 if compile_mode == 'exec': p = parse26.Python26Parser(debug_parser) else: p = parse26.Python26ParserSingle(debug_parser) elif version == 2.7: import uncompyle6.parsers.parse27 as parse27 if compile_mode == 'exec': p = parse27.Python27Parser(debug_parser) else: p = parse27.Python27ParserSingle(debug_parser) else: import uncompyle6.parsers.parse2 as parse2 if compile_mode == 'exec': p = parse2.Python2Parser(debug_parser) else: p = parse2.Python2ParserSingle(debug_parser) pass pass pass else: import uncompyle6.parsers.parse3 as parse3 if version == 3.0: import uncompyle6.parsers.parse30 as parse30 if compile_mode == 'exec': p = parse30.Python30Parser(debug_parser) else: p = parse30.Python30ParserSingle(debug_parser) elif version == 3.1: import uncompyle6.parsers.parse31 as parse31 if compile_mode == 'exec': p = parse31.Python31Parser(debug_parser) else: p = parse31.Python31ParserSingle(debug_parser) elif version == 3.2: import uncompyle6.parsers.parse32 as parse32 if compile_mode == 'exec': p = parse32.Python32Parser(debug_parser) else: p = parse32.Python32ParserSingle(debug_parser) elif version == 3.3: import uncompyle6.parsers.parse33 as parse33 if compile_mode == 'exec': p = parse33.Python33Parser(debug_parser) else: p = parse33.Python33ParserSingle(debug_parser) elif version == 3.4: import uncompyle6.parsers.parse34 as parse34 if compile_mode == 'exec': p = parse34.Python34Parser(debug_parser) else: p = parse34.Python34ParserSingle(debug_parser) elif version == 3.5: import uncompyle6.parsers.parse35 as parse35 if compile_mode == 'exec': p = parse35.Python35Parser(debug_parser) else: p = parse35.Python35ParserSingle(debug_parser) elif version == 3.6: import uncompyle6.parsers.parse36 as parse36 if compile_mode == 'exec': p = parse36.Python36Parser(debug_parser) else: p = parse36.Python36ParserSingle(debug_parser) else: if compile_mode == 'exec': p = parse3.Python3Parser(debug_parser) else: p = parse3.Python3ParserSingle(debug_parser) p.version = version # p.dump_grammar() # debug return p class PythonParserSingle(PythonParser): def p_call_stmt_single(self, args): ''' # single-mode compilation. Eval-mode interactive compilation # drops the last rule. call_stmt ::= expr PRINT_EXPR ''' def python_parser(version, co, out=sys.stdout, showasm=False, parser_debug=PARSER_DEFAULT_DEBUG, is_pypy=False): """ Parse a code object to an abstract syntax tree representation. :param version: The python version this code is from as a float, for example 2.6, 2.7, 3.2, 3.3, 3.4, 3.5 etc. :param co: The code object to parse. :param out: File like object to write the output to. :param showasm: Flag which determines whether the disassembled and ingested code is written to sys.stdout or not. :param parser_debug: dict containing debug flags for the spark parser. :return: Abstract syntax tree representation of the code object. """ assert iscode(co) from uncompyle6.scanner import get_scanner scanner = get_scanner(version, is_pypy) tokens, customize = scanner.ingest(co) maybe_show_asm(showasm, tokens) # For heavy grammar debugging # parser_debug = {'rules': True, 'transition': True, 'reduce' : True, # 'showstack': 'full'} p = get_python_parser(version, parser_debug) return parse(p, tokens, customize) if __name__ == '__main__': def parse_test(co): from uncompyle6 import PYTHON_VERSION, IS_PYPY ast = python_parser('2.7.13', co, showasm=True, is_pypy=True) ast = python_parser(PYTHON_VERSION, co, showasm=True, is_pypy=IS_PYPY) print(ast) return parse_test(parse_test.__code__)