You've already forked python-uncompyle6
mirror of
https://github.com/rocky/python-uncompyle6.git
synced 2025-08-03 00:45:53 +08:00
1285 lines
58 KiB
Python
1285 lines
58 KiB
Python
# Copyright (c) 2015-2019 Rocky Bernstein
|
|
# Copyright (c) 2005 by Dan Pascu <dan@windowmaker.org>
|
|
# Copyright (c) 2000-2002 by hartmut Goebel <h.goebel@crazy-compilers.com>
|
|
# 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 <http://www.gnu.org/licenses/>.
|
|
"""
|
|
A spark grammar for Python 3.x.
|
|
|
|
However instead of terminal symbols being the usual ASCII text,
|
|
e.g. 5, myvariable, "for", etc. they are CPython Bytecode tokens,
|
|
e.g. "LOAD_CONST 5", "STORE NAME myvariable", "SETUP_LOOP", etc.
|
|
|
|
If we succeed in creating a parse tree, then we have a Python program
|
|
that a later phase can turn into a sequence of ASCII text.
|
|
"""
|
|
|
|
from uncompyle6.scanners.tok import Token
|
|
from uncompyle6.parser import PythonParser, PythonParserSingle, nop_func
|
|
from uncompyle6.parsers.treenode import SyntaxTree
|
|
from spark_parser import DEFAULT_DEBUG as PARSER_DEFAULT_DEBUG
|
|
from xdis import PYTHON3
|
|
|
|
class Python3Parser(PythonParser):
|
|
|
|
def __init__(self, debug_parser=PARSER_DEFAULT_DEBUG):
|
|
self.added_rules = set()
|
|
super(Python3Parser, self).__init__(SyntaxTree, 'stmts', debug=debug_parser)
|
|
self.new_rules = set()
|
|
|
|
def p_comprehension3(self, args):
|
|
"""
|
|
# Python3 scanner adds LOAD_LISTCOMP. Python3 does list comprehension like
|
|
# other comprehensions (set, dictionary).
|
|
|
|
# Our "continue" heuristic - in two successive JUMP_BACKS, the first
|
|
# one may be a continue - sometimes classifies a JUMP_BACK
|
|
# as a CONTINUE. The two are kind of the same in a comprehension.
|
|
|
|
comp_for ::= expr for_iter store comp_iter CONTINUE
|
|
comp_for ::= expr for_iter store comp_iter JUMP_BACK
|
|
|
|
list_comp ::= BUILD_LIST_0 list_iter
|
|
lc_body ::= expr LIST_APPEND
|
|
list_for ::= expr FOR_ITER store list_iter jb_or_c
|
|
|
|
# This is seen in PyPy, but possibly it appears on other Python 3?
|
|
list_if ::= expr jmp_false list_iter COME_FROM
|
|
list_if_not ::= expr jmp_true list_iter COME_FROM
|
|
|
|
jb_or_c ::= JUMP_BACK
|
|
jb_or_c ::= CONTINUE
|
|
|
|
stmt ::= set_comp_func
|
|
|
|
set_comp_func ::= BUILD_SET_0 LOAD_FAST FOR_ITER store comp_iter
|
|
JUMP_BACK RETURN_VALUE RETURN_LAST
|
|
|
|
set_comp_func ::= BUILD_SET_0 LOAD_FAST FOR_ITER store comp_iter
|
|
COME_FROM JUMP_BACK RETURN_VALUE RETURN_LAST
|
|
|
|
comp_body ::= dict_comp_body
|
|
comp_body ::= set_comp_body
|
|
dict_comp_body ::= expr expr MAP_ADD
|
|
set_comp_body ::= expr SET_ADD
|
|
|
|
# See also common Python p_list_comprehension
|
|
"""
|
|
|
|
def p_dict_comp3(self, args):
|
|
""""
|
|
expr ::= dict_comp
|
|
stmt ::= dict_comp_func
|
|
dict_comp_func ::= BUILD_MAP_0 LOAD_FAST FOR_ITER store
|
|
comp_iter JUMP_BACK RETURN_VALUE RETURN_LAST
|
|
|
|
comp_iter ::= comp_if
|
|
comp_iter ::= comp_if_not
|
|
comp_if_not ::= expr jmp_true comp_iter
|
|
comp_iter ::= comp_body
|
|
"""
|
|
|
|
def p_grammar(self, args):
|
|
'''
|
|
sstmt ::= stmt
|
|
sstmt ::= ifelsestmtr
|
|
sstmt ::= return RETURN_LAST
|
|
|
|
return_if_stmts ::= return_if_stmt come_from_opt
|
|
return_if_stmts ::= _stmts return_if_stmt
|
|
return_if_stmt ::= ret_expr RETURN_END_IF
|
|
returns ::= _stmts return_if_stmt
|
|
|
|
stmt ::= break
|
|
break ::= BREAK_LOOP
|
|
|
|
stmt ::= continue
|
|
continue ::= CONTINUE
|
|
continues ::= _stmts lastl_stmt continue
|
|
continues ::= lastl_stmt continue
|
|
continues ::= continue
|
|
|
|
|
|
kwarg ::= LOAD_CONST expr
|
|
kwargs ::= kwarg+
|
|
|
|
|
|
classdef ::= build_class store
|
|
|
|
# FIXME: we need to add these because don't detect this properly
|
|
# in custom rules. Specifically if one of the exprs is CALL_FUNCTION
|
|
# then we'll mistake that for the final CALL_FUNCTION.
|
|
# We can fix by triggering on the CALL_FUNCTION op
|
|
# Python3 introduced LOAD_BUILD_CLASS
|
|
# Other definitions are in a custom rule
|
|
build_class ::= LOAD_BUILD_CLASS mkfunc expr call CALL_FUNCTION_3
|
|
build_class ::= LOAD_BUILD_CLASS mkfunc expr call expr CALL_FUNCTION_4
|
|
|
|
stmt ::= classdefdeco
|
|
classdefdeco ::= classdefdeco1 store
|
|
|
|
expr ::= LOAD_ASSERT
|
|
assert ::= assert_expr jmp_true LOAD_ASSERT RAISE_VARARGS_1 COME_FROM
|
|
|
|
assert_expr ::= expr
|
|
assert_expr ::= assert_expr_or
|
|
assert_expr ::= assert_expr_and
|
|
assert_expr_or ::= assert_expr jmp_true expr
|
|
assert_expr_and ::= assert_expr jmp_false expr
|
|
|
|
ifstmt ::= testexpr _ifstmts_jump
|
|
|
|
testexpr ::= testfalse
|
|
testexpr ::= testtrue
|
|
testfalse ::= expr jmp_false
|
|
testtrue ::= expr jmp_true
|
|
|
|
_ifstmts_jump ::= return_if_stmts
|
|
_ifstmts_jump ::= c_stmts_opt COME_FROM
|
|
|
|
iflaststmt ::= testexpr c_stmts_opt JUMP_ABSOLUTE
|
|
|
|
iflaststmtl ::= testexpr c_stmts_opt JUMP_BACK
|
|
iflaststmtl ::= testexpr c_stmts_opt JUMP_BACK COME_FROM_LOOP
|
|
iflaststmtl ::= testexpr c_stmts_opt JUMP_BACK POP_BLOCK
|
|
|
|
# These are used to keep parse tree indices the same
|
|
jump_forward_else ::= JUMP_FORWARD ELSE
|
|
jump_absolute_else ::= JUMP_ABSOLUTE ELSE
|
|
|
|
# Note: in if/else kinds of statements, we err on the side
|
|
# of missing "else" clauses. Therefore we include grammar
|
|
# rules with and without ELSE.
|
|
|
|
ifelsestmt ::= testexpr c_stmts_opt JUMP_FORWARD
|
|
else_suite opt_come_from_except
|
|
ifelsestmt ::= testexpr c_stmts_opt jump_forward_else
|
|
else_suite _come_froms
|
|
|
|
# ifelsestmt ::= testexpr c_stmts_opt jump_forward_else
|
|
# pass _come_froms
|
|
|
|
ifelsestmtc ::= testexpr c_stmts_opt JUMP_ABSOLUTE else_suitec
|
|
ifelsestmtc ::= testexpr c_stmts_opt jump_absolute_else else_suitec
|
|
|
|
ifelsestmtr ::= testexpr return_if_stmts returns
|
|
|
|
ifelsestmtl ::= testexpr c_stmts_opt JUMP_BACK else_suitel
|
|
ifelsestmtl ::= testexpr c_stmts_opt cf_jump_back else_suitel
|
|
|
|
cf_jump_back ::= COME_FROM JUMP_BACK
|
|
|
|
# FIXME: this feels like a hack. Is it just 1 or two
|
|
# COME_FROMs? the parsed tree for this and even with just the
|
|
# one COME_FROM for Python 2.7 seems to associate the
|
|
# COME_FROM targets from the wrong places
|
|
|
|
# this is nested inside a try_except
|
|
tryfinallystmt ::= SETUP_FINALLY suite_stmts_opt
|
|
POP_BLOCK LOAD_CONST
|
|
COME_FROM_FINALLY suite_stmts_opt END_FINALLY
|
|
|
|
except_handler ::= jmp_abs COME_FROM except_stmts
|
|
END_FINALLY
|
|
except_handler ::= jmp_abs COME_FROM_EXCEPT except_stmts
|
|
END_FINALLY
|
|
|
|
# FIXME: remove this
|
|
except_handler ::= JUMP_FORWARD COME_FROM except_stmts
|
|
END_FINALLY COME_FROM
|
|
|
|
except_handler ::= JUMP_FORWARD COME_FROM except_stmts
|
|
END_FINALLY COME_FROM_EXCEPT
|
|
|
|
except_stmts ::= except_stmts except_stmt
|
|
except_stmts ::= except_stmt
|
|
|
|
except_stmt ::= except_cond1 except_suite
|
|
except_stmt ::= except_cond2 except_suite
|
|
except_stmt ::= except_cond2 except_suite_finalize
|
|
except_stmt ::= except
|
|
|
|
## FIXME: what's except_pop_except?
|
|
except_stmt ::= except_pop_except
|
|
|
|
# Python3 introduced POP_EXCEPT
|
|
except_suite ::= c_stmts_opt POP_EXCEPT jump_except
|
|
jump_except ::= JUMP_ABSOLUTE
|
|
jump_except ::= JUMP_BACK
|
|
jump_except ::= JUMP_FORWARD
|
|
jump_except ::= CONTINUE
|
|
|
|
# This is used in Python 3 in
|
|
# "except ... as e" to remove 'e' after the c_stmts_opt finishes
|
|
except_suite_finalize ::= SETUP_FINALLY c_stmts_opt except_var_finalize
|
|
END_FINALLY _jump
|
|
|
|
except_var_finalize ::= POP_BLOCK POP_EXCEPT LOAD_CONST COME_FROM_FINALLY
|
|
LOAD_CONST store del_stmt
|
|
|
|
except_suite ::= returns
|
|
|
|
except_cond1 ::= DUP_TOP expr COMPARE_OP
|
|
jmp_false POP_TOP POP_TOP POP_TOP
|
|
|
|
except_cond2 ::= DUP_TOP expr COMPARE_OP
|
|
jmp_false POP_TOP store POP_TOP
|
|
|
|
except ::= POP_TOP POP_TOP POP_TOP c_stmts_opt POP_EXCEPT _jump
|
|
except ::= POP_TOP POP_TOP POP_TOP returns
|
|
|
|
jmp_abs ::= JUMP_ABSOLUTE
|
|
jmp_abs ::= JUMP_BACK
|
|
|
|
withstmt ::= expr SETUP_WITH POP_TOP suite_stmts_opt
|
|
POP_BLOCK LOAD_CONST COME_FROM_WITH
|
|
WITH_CLEANUP END_FINALLY
|
|
|
|
withasstmt ::= expr SETUP_WITH store suite_stmts_opt
|
|
POP_BLOCK LOAD_CONST COME_FROM_WITH
|
|
WITH_CLEANUP END_FINALLY
|
|
|
|
## FIXME: Right now we have erroneous jump targets
|
|
## This below is probably not correct when the COME_FROM is put in the right place
|
|
and ::= expr jmp_false expr COME_FROM
|
|
or ::= expr jmp_true expr COME_FROM
|
|
|
|
# # something like the below is needed when the jump targets are fixed
|
|
## or ::= expr JUMP_IF_TRUE_OR_POP COME_FROM expr
|
|
## and ::= expr JUMP_IF_FALSE_OR_POP COME_FROM expr
|
|
'''
|
|
|
|
def p_misc3(self, args):
|
|
"""
|
|
except_handler ::= JUMP_FORWARD COME_FROM_EXCEPT except_stmts
|
|
END_FINALLY COME_FROM
|
|
except_handler ::= JUMP_FORWARD COME_FROM_EXCEPT except_stmts
|
|
END_FINALLY COME_FROM_EXCEPT_CLAUSE
|
|
|
|
for_block ::= l_stmts_opt COME_FROM_LOOP JUMP_BACK
|
|
for_block ::= l_stmts
|
|
iflaststmtl ::= testexpr c_stmts_opt
|
|
"""
|
|
|
|
def p_def_annotations3(self, args):
|
|
"""
|
|
# Annotated functions
|
|
stmt ::= function_def_annotate
|
|
function_def_annotate ::= mkfunc_annotate store
|
|
|
|
mkfuncdeco0 ::= mkfunc_annotate
|
|
|
|
# This has the annotation value.
|
|
# LOAD_NAME is used in an annotation type like
|
|
# int, float, str
|
|
annotate_arg ::= LOAD_NAME
|
|
# LOAD_CONST is used in an annotation string
|
|
annotate_arg ::= expr
|
|
|
|
# This stores the tuple of parameter names
|
|
# that have been annotated
|
|
annotate_tuple ::= LOAD_CONST
|
|
"""
|
|
|
|
def p_come_from3(self, args):
|
|
"""
|
|
opt_come_from_except ::= COME_FROM_EXCEPT
|
|
opt_come_from_except ::= _come_froms
|
|
opt_come_from_except ::= come_from_except_clauses
|
|
|
|
come_from_except_clauses ::= COME_FROM_EXCEPT_CLAUSE+
|
|
"""
|
|
|
|
def p_jump3(self, args):
|
|
"""
|
|
jmp_false ::= POP_JUMP_IF_FALSE
|
|
jmp_true ::= POP_JUMP_IF_TRUE
|
|
|
|
# FIXME: Common with 2.7
|
|
ret_and ::= expr JUMP_IF_FALSE_OR_POP ret_expr_or_cond COME_FROM
|
|
ret_or ::= expr JUMP_IF_TRUE_OR_POP ret_expr_or_cond COME_FROM
|
|
ret_cond ::= expr POP_JUMP_IF_FALSE expr RETURN_END_IF COME_FROM ret_expr_or_cond
|
|
|
|
or ::= expr JUMP_IF_TRUE_OR_POP expr COME_FROM
|
|
and ::= expr JUMP_IF_FALSE_OR_POP expr COME_FROM
|
|
|
|
# compare_chained1 is used exclusively in chained_compare
|
|
compare_chained1 ::= expr DUP_TOP ROT_THREE COMPARE_OP JUMP_IF_FALSE_OR_POP
|
|
compare_chained1 COME_FROM
|
|
compare_chained1 ::= expr DUP_TOP ROT_THREE COMPARE_OP JUMP_IF_FALSE_OR_POP
|
|
compare_chained2 COME_FROM
|
|
"""
|
|
|
|
def p_stmt3(self, args):
|
|
"""
|
|
stmt ::= conditional_lambda
|
|
stmt ::= conditional_not_lambda
|
|
conditional_lambda ::= expr jmp_false expr return_if_lambda
|
|
return_stmt_lambda LAMBDA_MARKER
|
|
conditional_not_lambda
|
|
::= expr jmp_true expr return_if_lambda
|
|
return_stmt_lambda LAMBDA_MARKER
|
|
|
|
return_stmt_lambda ::= ret_expr RETURN_VALUE_LAMBDA
|
|
return_if_lambda ::= RETURN_END_IF_LAMBDA
|
|
|
|
stmt ::= return_closure
|
|
return_closure ::= LOAD_CLOSURE RETURN_VALUE RETURN_LAST
|
|
|
|
stmt ::= whileTruestmt
|
|
ifelsestmt ::= testexpr c_stmts_opt JUMP_FORWARD else_suite _come_froms
|
|
"""
|
|
|
|
def p_loop_stmt3(self, args):
|
|
"""
|
|
for ::= SETUP_LOOP expr for_iter store for_block POP_BLOCK
|
|
COME_FROM_LOOP
|
|
|
|
forelsestmt ::= SETUP_LOOP expr for_iter store for_block POP_BLOCK else_suite
|
|
COME_FROM_LOOP
|
|
|
|
forelselaststmt ::= SETUP_LOOP expr for_iter store for_block POP_BLOCK else_suitec
|
|
COME_FROM_LOOP
|
|
|
|
forelselaststmtl ::= SETUP_LOOP expr for_iter store for_block POP_BLOCK else_suitel
|
|
COME_FROM_LOOP
|
|
|
|
whilestmt ::= SETUP_LOOP testexpr l_stmts_opt COME_FROM JUMP_BACK POP_BLOCK
|
|
COME_FROM_LOOP
|
|
|
|
whilestmt ::= SETUP_LOOP testexpr l_stmts_opt JUMP_BACK POP_BLOCK
|
|
COME_FROM_LOOP
|
|
|
|
whilestmt ::= SETUP_LOOP testexpr returns POP_BLOCK
|
|
COME_FROM_LOOP
|
|
|
|
while1elsestmt ::= SETUP_LOOP l_stmts JUMP_BACK
|
|
else_suitel
|
|
|
|
whileelsestmt ::= SETUP_LOOP testexpr l_stmts_opt JUMP_BACK POP_BLOCK
|
|
else_suitel COME_FROM_LOOP
|
|
|
|
whileTruestmt ::= SETUP_LOOP l_stmts_opt JUMP_BACK POP_BLOCK
|
|
COME_FROM_LOOP
|
|
|
|
# FIXME: Python 3.? starts adding branch optimization? Put this starting there.
|
|
|
|
while1stmt ::= SETUP_LOOP l_stmts COME_FROM_LOOP
|
|
while1stmt ::= SETUP_LOOP l_stmts COME_FROM JUMP_BACK COME_FROM_LOOP
|
|
|
|
while1elsestmt ::= SETUP_LOOP l_stmts JUMP_BACK
|
|
else_suite COME_FROM_LOOP
|
|
|
|
# FIXME: investigate - can code really produce a NOP?
|
|
whileTruestmt ::= SETUP_LOOP l_stmts_opt JUMP_BACK NOP
|
|
COME_FROM_LOOP
|
|
whileTruestmt ::= SETUP_LOOP l_stmts_opt JUMP_BACK POP_BLOCK NOP
|
|
COME_FROM_LOOP
|
|
for ::= SETUP_LOOP expr for_iter store for_block POP_BLOCK NOP
|
|
COME_FROM_LOOP
|
|
"""
|
|
|
|
def p_generator_exp3(self, args):
|
|
'''
|
|
load_genexpr ::= LOAD_GENEXPR
|
|
load_genexpr ::= BUILD_TUPLE_1 LOAD_GENEXPR LOAD_CONST
|
|
'''
|
|
|
|
def p_expr3(self, args):
|
|
"""
|
|
expr ::= conditionalnot
|
|
conditionalnot ::= expr jmp_true expr jump_forward_else expr COME_FROM
|
|
|
|
# a JUMP_FORWARD to another JUMP_FORWARD can get turned into
|
|
# a JUMP_ABSOLUTE with no COME_FROM
|
|
conditional ::= expr jmp_false expr jump_absolute_else expr
|
|
|
|
# conditional_true are for conditions which always evaluate true
|
|
# There is dead or non-optional remnants of the condition code though,
|
|
# and we use that to match on to reconstruct the source more accurately
|
|
expr ::= conditional_true
|
|
conditional_true ::= expr JUMP_FORWARD expr COME_FROM
|
|
"""
|
|
|
|
@staticmethod
|
|
def call_fn_name(token):
|
|
"""Customize CALL_FUNCTION to add the number of positional arguments"""
|
|
if token.attr is not None:
|
|
return '%s_%i' % (token.kind, token.attr)
|
|
else:
|
|
return '%s_0' % (token.kind)
|
|
|
|
def custom_build_class_rule(self, opname, i, token, tokens, customize):
|
|
'''
|
|
# Should the first rule be somehow folded into the 2nd one?
|
|
build_class ::= LOAD_BUILD_CLASS mkfunc
|
|
LOAD_CLASSNAME {expr}^n-1 CALL_FUNCTION_n
|
|
LOAD_CONST CALL_FUNCTION_n
|
|
build_class ::= LOAD_BUILD_CLASS mkfunc
|
|
expr
|
|
call
|
|
CALL_FUNCTION_3
|
|
'''
|
|
# FIXME: I bet this can be simplified
|
|
# look for next MAKE_FUNCTION
|
|
for i in range(i+1, len(tokens)):
|
|
if tokens[i].kind.startswith('MAKE_FUNCTION'):
|
|
break
|
|
elif tokens[i].kind.startswith('MAKE_CLOSURE'):
|
|
break
|
|
pass
|
|
assert i < len(tokens), "build_class needs to find MAKE_FUNCTION or MAKE_CLOSURE"
|
|
assert tokens[i+1].kind == 'LOAD_CONST', \
|
|
"build_class expecting CONST after MAKE_FUNCTION/MAKE_CLOSURE"
|
|
call_fn_tok = None
|
|
for i in range(i, len(tokens)):
|
|
if tokens[i].kind.startswith('CALL_FUNCTION'):
|
|
call_fn_tok = tokens[i]
|
|
break
|
|
if not call_fn_tok:
|
|
raise RuntimeError("build_class custom rule for %s needs to find CALL_FUNCTION"
|
|
% opname)
|
|
|
|
|
|
# customize build_class rule
|
|
# FIXME: What's the deal with the two rules? Different Python versions?
|
|
# Different situations? Note that the above rule is based on the CALL_FUNCTION
|
|
# token found, while this one doesn't.
|
|
if self.version < 3.6:
|
|
call_function = self.call_fn_name(call_fn_tok)
|
|
args_pos, args_kw = self.get_pos_kw(call_fn_tok)
|
|
rule = ("build_class ::= LOAD_BUILD_CLASS mkfunc %s"
|
|
"%s" % (('expr ' * (args_pos - 1) + ('kwarg ' * args_kw)),
|
|
call_function))
|
|
else:
|
|
# 3.6+ handling
|
|
call_function = call_fn_tok.kind
|
|
if call_function.startswith("CALL_FUNCTION_KW"):
|
|
self.addRule("classdef ::= build_class_kw store", nop_func)
|
|
rule = ("build_class_kw ::= LOAD_BUILD_CLASS mkfunc %sLOAD_CONST %s"
|
|
% ('expr ' * (call_fn_tok.attr - 1), call_function))
|
|
else:
|
|
call_function = self.call_fn_name(call_fn_tok)
|
|
rule = ("build_class ::= LOAD_BUILD_CLASS mkfunc %s%s"
|
|
% ('expr ' * (call_fn_tok.attr - 1), call_function))
|
|
self.addRule(rule, nop_func)
|
|
return
|
|
|
|
def custom_classfunc_rule(self, opname, token, customize, next_token):
|
|
"""
|
|
call ::= expr {expr}^n CALL_FUNCTION_n
|
|
call ::= expr {expr}^n CALL_FUNCTION_VAR_n
|
|
call ::= expr {expr}^n CALL_FUNCTION_VAR_KW_n
|
|
call ::= expr {expr}^n CALL_FUNCTION_KW_n
|
|
|
|
classdefdeco2 ::= LOAD_BUILD_CLASS mkfunc {expr}^n-1 CALL_FUNCTION_n
|
|
"""
|
|
args_pos, args_kw = self.get_pos_kw(token)
|
|
|
|
# Additional exprs for * and ** args:
|
|
# 0 if neither
|
|
# 1 for CALL_FUNCTION_VAR or CALL_FUNCTION_KW
|
|
# 2 for * and ** args (CALL_FUNCTION_VAR_KW).
|
|
# Yes, this computation based on instruction name is a little bit hoaky.
|
|
nak = ( len(opname)-len('CALL_FUNCTION') ) // 3
|
|
|
|
token.kind = self.call_fn_name(token)
|
|
uniq_param = args_kw + args_pos
|
|
|
|
# Note: 3.5+ have subclassed this method; so we don't handle
|
|
# 'CALL_FUNCTION_VAR' or 'CALL_FUNCTION_EX' here.
|
|
rule = ('call ::= expr ' +
|
|
('pos_arg ' * args_pos) +
|
|
('kwarg ' * args_kw) +
|
|
'expr ' * nak + token.kind)
|
|
|
|
self.add_unique_rule(rule, token.kind, uniq_param, customize)
|
|
|
|
if 'LOAD_BUILD_CLASS' in self.seen_ops:
|
|
if (next_token == 'CALL_FUNCTION' and next_token.attr == 1
|
|
and args_pos > 1):
|
|
rule = ('classdefdeco2 ::= LOAD_BUILD_CLASS mkfunc %s%s_%d'
|
|
% (('expr ' * (args_pos-1)), opname, args_pos))
|
|
self.add_unique_rule(rule, token.kind, uniq_param, customize)
|
|
|
|
def add_make_function_rule(self, rule, opname, attr, customize):
|
|
"""Python 3.3 added a an addtional LOAD_CONST before MAKE_FUNCTION and
|
|
this has an effect on many rules.
|
|
"""
|
|
if self.version >= 3.3:
|
|
new_rule = rule % (('LOAD_CONST ') * 1)
|
|
else:
|
|
new_rule = rule % (('LOAD_CONST ') * 0)
|
|
self.add_unique_rule(new_rule, opname, attr, customize)
|
|
|
|
def customize_grammar_rules(self, tokens, customize):
|
|
"""The base grammar we start out for a Python version even with the
|
|
subclassing is, well, is pretty base. And we want it that way: lean and
|
|
mean so that parsing will go faster.
|
|
|
|
Here, we add additional grammar rules based on specific instructions
|
|
that are in the instruction/token stream. In classes that
|
|
inherit from from here and other versions, grammar rules may
|
|
also be removed.
|
|
|
|
For example if we see a pretty rare DELETE_DEREF instruction we'll
|
|
add the grammar for that.
|
|
|
|
More importantly, here we add grammar rules for instructions
|
|
that may access a variable number of stack items. CALL_FUNCTION,
|
|
BUILD_LIST and so on are like this.
|
|
|
|
Without custom rules, there can be an super-exponential number of
|
|
derivations. See the deparsing paper for an elaboration of
|
|
this.
|
|
|
|
"""
|
|
|
|
is_pypy = False
|
|
|
|
# For a rough break out on the first word. This may
|
|
# include instructions that don't need customization,
|
|
# but we'll do a finer check after the rough breakout.
|
|
customize_instruction_basenames = frozenset(
|
|
('BUILD', 'CALL', 'CONTINUE', 'DELETE', 'GET',
|
|
'JUMP', 'LOAD', 'LOOKUP', 'MAKE',
|
|
'RETURN', 'RAISE', 'SETUP',
|
|
'UNPACK'))
|
|
|
|
# Opcode names in the custom_ops_processed set have rules that get added
|
|
# unconditionally and the rules are constant. So they need to be done
|
|
# only once and if we see the opcode a second we don't have to consider
|
|
# adding more rules.
|
|
#
|
|
# Note: BUILD_TUPLE_UNPACK_WITH_CALL gets considered by
|
|
# default because it starts with BUILD. So we'll set to ignore it from
|
|
# the start.
|
|
custom_ops_processed = set(('BUILD_TUPLE_UNPACK_WITH_CALL',))
|
|
|
|
# A set of instruction operation names that exist in the token stream.
|
|
# We use this customize the grammar that we create.
|
|
# 2.6-compatible set comprehensions
|
|
self.seen_ops = frozenset([t.kind for t in tokens])
|
|
self.seen_op_basenames = frozenset([opname[:opname.rfind('_')] for opname in self.seen_ops])
|
|
|
|
# Loop over instructions adding custom grammar rules based on
|
|
# a specific instruction seen.
|
|
|
|
if 'PyPy' in customize:
|
|
is_pypy = True
|
|
self.addRule("""
|
|
stmt ::= assign3_pypy
|
|
stmt ::= assign2_pypy
|
|
assign3_pypy ::= expr expr expr store store store
|
|
assign2_pypy ::= expr expr store store
|
|
stmt ::= conditional_lambda
|
|
stmt ::= conditional_not_lambda
|
|
conditional_lambda ::= expr jmp_false expr return_if_lambda
|
|
return_lambda LAMBDA_MARKER
|
|
conditional_not_lambda
|
|
::= expr jmp_true expr return_if_lambda
|
|
return_lambda LAMBDA_MARKER
|
|
""", nop_func)
|
|
|
|
n = len(tokens)
|
|
|
|
# Determine if we have an iteration CALL_FUNCTION_1.
|
|
has_get_iter_call_function1 = False
|
|
max_branches = 0
|
|
for i, token in enumerate(tokens):
|
|
if token == 'GET_ITER' and i < n-2 and self.call_fn_name(tokens[i+1]) == 'CALL_FUNCTION_1':
|
|
has_get_iter_call_function1 = True
|
|
max_branches += 1
|
|
elif (token == 'GET_AWAITABLE' and i < n-3
|
|
and tokens[i+1] == 'LOAD_CONST' and tokens[i+2] == 'YIELD_FROM'):
|
|
max_branches += 1
|
|
if max_branches > 2:
|
|
break
|
|
|
|
for i, token in enumerate(tokens):
|
|
opname = token.kind
|
|
|
|
# Do a quick breakout before testing potentially
|
|
# each of the dozen or so instruction in if elif.
|
|
if (opname[:opname.find('_')] not in customize_instruction_basenames
|
|
or opname in custom_ops_processed):
|
|
continue
|
|
|
|
opname_base = opname[:opname.rfind('_')]
|
|
# The order of opname listed is roughly sorted below
|
|
if opname_base == 'BUILD_CONST_KEY_MAP':
|
|
# This is in 3.6+
|
|
kvlist_n = 'expr ' * (token.attr)
|
|
rule = "dict ::= %sLOAD_CONST %s" % (kvlist_n, opname)
|
|
self.addRule(rule, nop_func)
|
|
elif opname.startswith('BUILD_LIST_UNPACK'):
|
|
v = token.attr
|
|
rule = 'build_list_unpack ::= %s%s' % ('expr ' * v, opname)
|
|
self.addRule(rule, nop_func)
|
|
rule = 'expr ::= build_list_unpack'
|
|
self.addRule(rule, nop_func)
|
|
elif opname_base in ('BUILD_MAP', 'BUILD_MAP_UNPACK'):
|
|
kvlist_n = "kvlist_%s" % token.attr
|
|
if opname == 'BUILD_MAP_n':
|
|
# PyPy sometimes has no count. Sigh.
|
|
rule = ('dict_comp_func ::= BUILD_MAP_n LOAD_FAST FOR_ITER store '
|
|
'comp_iter JUMP_BACK RETURN_VALUE RETURN_LAST')
|
|
self.add_unique_rule(rule, 'dict_comp_func', 1, customize)
|
|
|
|
kvlist_n = 'kvlist_n'
|
|
rule = 'kvlist_n ::= kvlist_n kv3'
|
|
self.add_unique_rule(rule, 'kvlist_n', 0, customize)
|
|
rule = 'kvlist_n ::='
|
|
self.add_unique_rule(rule, 'kvlist_n', 1, customize)
|
|
rule = "dict ::= BUILD_MAP_n kvlist_n"
|
|
elif self.version >= 3.5:
|
|
if not opname.startswith('BUILD_MAP_WITH_CALL'):
|
|
# FIXME: Use the attr
|
|
# so this doesn't run into exponential parsing time.
|
|
if opname.startswith('BUILD_MAP_UNPACK'):
|
|
self.add_unique_rule(rule, opname, token.attr, customize)
|
|
rule = 'dict_entry ::= ' + 'expr ' * (token.attr*2)
|
|
self.add_unique_rule(rule, opname, token.attr, customize)
|
|
|
|
# FIXME: start here. The LHS should be unmap_dict, not dict.
|
|
# FIXME: really we need a combination of dict_entry-like things.
|
|
# It just so happens the most common case is not to mix
|
|
# dictionary comphensions with dictionary, elements
|
|
if 'LOAD_DICTCOMP' in self.seen_ops:
|
|
rule = 'dict ::= %s%s' % ('dict_comp ' * token.attr, opname)
|
|
self.addRule(rule, nop_func)
|
|
rule = """
|
|
expr ::= unmap_dict
|
|
unmap_dict ::= %s%s
|
|
""" % ('expr ' * token.attr, opname)
|
|
else:
|
|
rule = "%s ::= %s %s" % (kvlist_n, 'expr ' * (token.attr*2), opname)
|
|
self.add_unique_rule(rule, opname, token.attr, customize)
|
|
rule = "dict ::= %s" % kvlist_n
|
|
else:
|
|
rule = kvlist_n + ' ::= ' + 'expr expr STORE_MAP ' * token.attr
|
|
self.add_unique_rule(rule, opname, token.attr, customize)
|
|
rule = "dict ::= %s %s" % (opname, kvlist_n)
|
|
self.add_unique_rule(rule, opname, token.attr, customize)
|
|
elif opname.startswith('BUILD_MAP_UNPACK_WITH_CALL'):
|
|
v = token.attr
|
|
rule = 'build_map_unpack_with_call ::= %s%s' % ('expr ' * v, opname)
|
|
self.addRule(rule, nop_func)
|
|
elif opname.startswith('BUILD_TUPLE_UNPACK_WITH_CALL'):
|
|
v = token.attr
|
|
rule = ('starred ::= %s %s' % ('expr ' * v, opname))
|
|
self.addRule(rule, nop_func)
|
|
|
|
elif opname_base in ('BUILD_LIST', 'BUILD_SET', 'BUILD_TUPLE',
|
|
'BUILD_TUPLE_UNPACK'):
|
|
v = token.attr
|
|
|
|
is_LOAD_CLOSURE = False
|
|
if opname_base == 'BUILD_TUPLE':
|
|
# If is part of a "load_closure", then it is not part of a
|
|
# "list".
|
|
is_LOAD_CLOSURE = True
|
|
for j in range(v):
|
|
if tokens[i-j-1].kind != 'LOAD_CLOSURE':
|
|
is_LOAD_CLOSURE = False
|
|
break
|
|
if is_LOAD_CLOSURE:
|
|
rule = ('load_closure ::= %s%s' % (('LOAD_CLOSURE ' * v), opname))
|
|
self.add_unique_rule(rule, opname, token.attr, customize)
|
|
if not is_LOAD_CLOSURE or v == 0:
|
|
# We do this complicated test to speed up parsing of
|
|
# pathelogically long literals, especially those over 1024.
|
|
build_count = token.attr
|
|
thousands = (build_count//1024)
|
|
thirty32s = ((build_count//32) % 32)
|
|
if thirty32s > 0:
|
|
rule = "expr32 ::=%s" % (' expr' * 32)
|
|
self.add_unique_rule(rule, opname_base, build_count, customize)
|
|
pass
|
|
if thousands > 0:
|
|
self.add_unique_rule("expr1024 ::=%s" % (' expr32' * 32),
|
|
opname_base, build_count, customize)
|
|
pass
|
|
collection = opname_base[opname_base.find('_')+1:].lower()
|
|
rule = (('%s ::= ' % collection) + 'expr1024 '*thousands +
|
|
'expr32 '*thirty32s + 'expr '*(build_count % 32) + opname)
|
|
self.add_unique_rules([
|
|
"expr ::= %s" % collection,
|
|
rule], customize)
|
|
continue
|
|
continue
|
|
elif opname_base == 'BUILD_SLICE':
|
|
if token.attr == 2:
|
|
self.add_unique_rules([
|
|
'expr ::= build_slice2',
|
|
'build_slice2 ::= expr expr BUILD_SLICE_2'
|
|
], customize)
|
|
else:
|
|
assert token.attr == 3, "BUILD_SLICE value must be 2 or 3; is %s" % v
|
|
self.add_unique_rules([
|
|
'expr ::= build_slice3',
|
|
'build_slice3 ::= expr expr expr BUILD_SLICE_3',
|
|
], customize)
|
|
elif (opname in frozenset(('CALL_FUNCTION',
|
|
'CALL_FUNCTION_EX',
|
|
'CALL_FUNCTION_EX_KW',
|
|
'CALL_FUNCTION_VAR',
|
|
'CALL_FUNCTION_VAR_KW'))
|
|
or opname.startswith('CALL_FUNCTION_KW')):
|
|
|
|
if opname == 'CALL_FUNCTION' and token.attr == 1:
|
|
rule = """
|
|
dict_comp ::= LOAD_DICTCOMP LOAD_CONST MAKE_FUNCTION_0 expr
|
|
GET_ITER CALL_FUNCTION_1
|
|
classdefdeco1 ::= expr classdefdeco2 CALL_FUNCTION_1
|
|
"""
|
|
if self.version < 3.5:
|
|
rule += """
|
|
classdefdeco1 ::= expr classdefdeco1 CALL_FUNCTION_1
|
|
"""
|
|
self.addRule(rule, nop_func)
|
|
|
|
self.custom_classfunc_rule(opname, token, customize, tokens[i+1])
|
|
# Note: don't add to custom_ops_processed.
|
|
|
|
elif opname_base == 'CALL_METHOD':
|
|
# PyPy only - DRY with parse2
|
|
|
|
args_pos, args_kw = self.get_pos_kw(token)
|
|
|
|
# number of apply equiv arguments:
|
|
nak = ( len(opname_base)-len('CALL_METHOD') ) // 3
|
|
rule = ('call ::= expr ' +
|
|
('pos_arg ' * args_pos) +
|
|
('kwarg ' * args_kw) +
|
|
'expr ' * nak + opname)
|
|
self.add_unique_rule(rule, opname, token.attr, customize)
|
|
elif opname == 'CONTINUE':
|
|
self.addRule('continue ::= CONTINUE', nop_func)
|
|
custom_ops_processed.add(opname)
|
|
elif opname == 'CONTINUE_LOOP':
|
|
self.addRule('continue ::= CONTINUE_LOOP', nop_func)
|
|
custom_ops_processed.add(opname)
|
|
elif opname == 'DELETE_ATTR':
|
|
self.addRule('del_stmt ::= expr DELETE_ATTR', nop_func)
|
|
custom_ops_processed.add(opname)
|
|
elif opname == 'DELETE_DEREF':
|
|
self.addRule("""
|
|
stmt ::= del_deref_stmt
|
|
del_deref_stmt ::= DELETE_DEREF
|
|
""", nop_func)
|
|
custom_ops_processed.add(opname)
|
|
elif opname == 'DELETE_SUBSCR':
|
|
self.addRule("""
|
|
del_stmt ::= delete_subscr
|
|
delete_subscr ::= expr expr DELETE_SUBSCR
|
|
""", nop_func)
|
|
custom_ops_processed.add(opname)
|
|
elif opname == 'GET_ITER':
|
|
self.addRule("""
|
|
expr ::= get_iter
|
|
attribute ::= expr GET_ITER
|
|
""", nop_func)
|
|
custom_ops_processed.add(opname)
|
|
elif opname == 'JUMP_IF_NOT_DEBUG':
|
|
v = token.attr
|
|
self.addRule("""
|
|
stmt ::= assert_pypy
|
|
stmt ::= assert2_pypy", nop_func)
|
|
assert_pypy ::= JUMP_IF_NOT_DEBUG assert_expr jmp_true
|
|
LOAD_ASSERT RAISE_VARARGS_1 COME_FROM
|
|
assert2_pypy ::= JUMP_IF_NOT_DEBUG assert_expr jmp_true
|
|
LOAD_ASSERT expr CALL_FUNCTION_1
|
|
RAISE_VARARGS_1 COME_FROM
|
|
assert2_pypy ::= JUMP_IF_NOT_DEBUG assert_expr jmp_true
|
|
LOAD_ASSERT expr CALL_FUNCTION_1
|
|
RAISE_VARARGS_1 COME_FROM,
|
|
""", nop_func)
|
|
custom_ops_processed.add(opname)
|
|
elif opname == 'LOAD_BUILD_CLASS':
|
|
self.custom_build_class_rule(opname, i, token, tokens, customize)
|
|
# Note: don't add to custom_ops_processed.
|
|
elif opname == 'LOAD_CLASSDEREF':
|
|
# Python 3.4+
|
|
self.addRule("expr ::= LOAD_CLASSDEREF", nop_func)
|
|
custom_ops_processed.add(opname)
|
|
elif opname == 'LOAD_CLASSNAME':
|
|
self.addRule("expr ::= LOAD_CLASSNAME", nop_func)
|
|
custom_ops_processed.add(opname)
|
|
elif opname == 'LOAD_DICTCOMP':
|
|
if has_get_iter_call_function1:
|
|
rule_pat = ("dict_comp ::= LOAD_DICTCOMP %sMAKE_FUNCTION_0 expr "
|
|
"GET_ITER CALL_FUNCTION_1")
|
|
self.add_make_function_rule(rule_pat, opname, token.attr, customize)
|
|
pass
|
|
custom_ops_processed.add(opname)
|
|
elif opname == 'LOAD_ATTR':
|
|
self.addRule("""
|
|
expr ::= attribute
|
|
attribute ::= expr LOAD_ATTR
|
|
""", nop_func)
|
|
custom_ops_processed.add(opname)
|
|
elif opname == 'LOAD_LISTCOMP':
|
|
self.add_unique_rule("expr ::= listcomp", opname, token.attr, customize)
|
|
custom_ops_processed.add(opname)
|
|
elif opname == 'LOAD_SETCOMP':
|
|
# Should this be generalized and put under MAKE_FUNCTION?
|
|
if has_get_iter_call_function1:
|
|
self.addRule("expr ::= set_comp", nop_func)
|
|
rule_pat = ("set_comp ::= LOAD_SETCOMP %sMAKE_FUNCTION_0 expr "
|
|
"GET_ITER CALL_FUNCTION_1")
|
|
self.add_make_function_rule(rule_pat, opname, token.attr, customize)
|
|
pass
|
|
custom_ops_processed.add(opname)
|
|
elif opname == 'LOOKUP_METHOD':
|
|
# A PyPy speciality - DRY with parse3
|
|
self.addRule("""
|
|
expr ::= attribute
|
|
attribute ::= expr LOOKUP_METHOD
|
|
""",
|
|
nop_func)
|
|
custom_ops_processed.add(opname)
|
|
elif opname.startswith('MAKE_CLOSURE'):
|
|
# DRY with MAKE_FUNCTION
|
|
# Note: this probably doesn't handle kwargs proprerly
|
|
|
|
if opname == 'MAKE_CLOSURE_0' and 'LOAD_DICTCOMP' in self.seen_ops:
|
|
# Is there something general going on here?
|
|
# Note that 3.6+ doesn't do this, but we'll remove
|
|
# this rule in parse36.py
|
|
rule = """
|
|
dict_comp ::= load_closure LOAD_DICTCOMP LOAD_CONST
|
|
MAKE_CLOSURE_0 expr
|
|
GET_ITER CALL_FUNCTION_1
|
|
"""
|
|
self.addRule(rule, nop_func)
|
|
|
|
args_pos, args_kw, annotate_args = token.attr
|
|
|
|
# FIXME: Fold test into add_make_function_rule
|
|
if self.version < 3.3:
|
|
j = 1
|
|
else:
|
|
j = 2
|
|
if is_pypy or (i >= j and tokens[i-j] == 'LOAD_LAMBDA'):
|
|
rule_pat = ('mklambda ::= %sload_closure LOAD_LAMBDA %%s%s' %
|
|
('pos_arg ' * args_pos, opname))
|
|
self.add_make_function_rule(rule_pat, opname, token.attr, customize)
|
|
|
|
if has_get_iter_call_function1:
|
|
rule_pat = ("generator_exp ::= %sload_closure load_genexpr %%s%s expr "
|
|
"GET_ITER CALL_FUNCTION_1" % ('pos_arg ' * args_pos, opname))
|
|
self.add_make_function_rule(rule_pat, opname, token.attr, customize)
|
|
|
|
if has_get_iter_call_function1:
|
|
if (is_pypy or (i >= j and tokens[i-j] == 'LOAD_LISTCOMP')):
|
|
# In the tokens we saw:
|
|
# LOAD_LISTCOMP LOAD_CONST MAKE_FUNCTION (>= 3.3) or
|
|
# LOAD_LISTCOMP MAKE_FUNCTION (< 3.3) or
|
|
# and have GET_ITER CALL_FUNCTION_1
|
|
# Todo: For Pypy we need to modify this slightly
|
|
rule_pat = ('listcomp ::= %sload_closure LOAD_LISTCOMP %%s%s expr '
|
|
'GET_ITER CALL_FUNCTION_1' % ('pos_arg ' * args_pos, opname))
|
|
self.add_make_function_rule(rule_pat, opname, token.attr, customize)
|
|
if (is_pypy or (i >= j and tokens[i-j] == 'LOAD_SETCOMP')):
|
|
rule_pat = ('set_comp ::= %sload_closure LOAD_SETCOMP %%s%s expr '
|
|
'GET_ITER CALL_FUNCTION_1' % ('pos_arg ' * args_pos, opname))
|
|
self.add_make_function_rule(rule_pat, opname, token.attr, customize)
|
|
if (is_pypy or (i >= j and tokens[i-j] == 'LOAD_DICTCOMP')):
|
|
self.add_unique_rule('dict_comp ::= %sload_closure LOAD_DICTCOMP %s '
|
|
'expr GET_ITER CALL_FUNCTION_1' %
|
|
('pos_arg ' * args_pos, opname),
|
|
opname, token.attr, customize)
|
|
|
|
if args_kw > 0:
|
|
kwargs_str = 'kwargs '
|
|
else:
|
|
kwargs_str = ''
|
|
|
|
# Note order of kwargs and pos args changed between 3.3-3.4
|
|
if self.version <= 3.2:
|
|
rule = ('mkfunc ::= %s%sload_closure LOAD_CONST %s'
|
|
% (kwargs_str, 'expr ' * args_pos, opname))
|
|
elif self.version == 3.3:
|
|
rule = ('mkfunc ::= %s%sload_closure LOAD_CONST LOAD_CONST %s'
|
|
% (kwargs_str, 'expr ' * args_pos, opname))
|
|
elif self.version >= 3.4:
|
|
rule = ('mkfunc ::= %s%s load_closure LOAD_CONST LOAD_CONST %s'
|
|
% ('expr ' * args_pos, kwargs_str, opname))
|
|
|
|
self.add_unique_rule(rule, opname, token.attr, customize)
|
|
|
|
if args_kw == 0:
|
|
rule = ('mkfunc ::= %sload_closure load_genexpr %s'
|
|
% ('pos_arg ' * args_pos, opname))
|
|
self.add_unique_rule(rule, opname, token.attr, customize)
|
|
|
|
if self.version < 3.4:
|
|
rule = ('mkfunc ::= %sload_closure LOAD_CONST %s'
|
|
% ('expr ' * args_pos, opname))
|
|
self.add_unique_rule(rule, opname, token.attr, customize)
|
|
|
|
pass
|
|
elif opname_base.startswith('MAKE_FUNCTION'):
|
|
# DRY with MAKE_CLOSURE
|
|
if self.version >= 3.6:
|
|
# The semantics of MAKE_FUNCTION in 3.6 are totally different from
|
|
# before.
|
|
args_pos, args_kw, annotate_args, closure = token.attr
|
|
stack_count = args_pos + args_kw + annotate_args
|
|
if closure:
|
|
if args_pos:
|
|
rule = ('mklambda ::= %s%s%s%s' %
|
|
('expr ' * stack_count,
|
|
'load_closure ' * closure,
|
|
'BUILD_TUPLE_1 LOAD_LAMBDA LOAD_CONST ',
|
|
opname))
|
|
else:
|
|
rule = ('mklambda ::= %s%s%s' %
|
|
('load_closure ' * closure,
|
|
'LOAD_LAMBDA LOAD_CONST ',
|
|
opname))
|
|
self.add_unique_rule(rule, opname, token.attr, customize)
|
|
|
|
rule = ('mkfunc ::= %s%s%s%s' %
|
|
('expr ' * stack_count,
|
|
'load_closure ' * closure,
|
|
'LOAD_CONST ' * 2,
|
|
opname))
|
|
self.add_unique_rule(rule, opname, token.attr, customize)
|
|
|
|
if has_get_iter_call_function1:
|
|
rule_pat = ("generator_exp ::= %sload_genexpr %%s%s expr "
|
|
"GET_ITER CALL_FUNCTION_1" % ('pos_arg ' * args_pos, opname))
|
|
self.add_make_function_rule(rule_pat, opname, token.attr, customize)
|
|
rule_pat = ("generator_exp ::= %sload_closure load_genexpr %%s%s expr "
|
|
"GET_ITER CALL_FUNCTION_1" % ('pos_arg ' * args_pos, opname))
|
|
self.add_make_function_rule(rule_pat, opname, token.attr, customize)
|
|
if is_pypy or (i >= 2 and tokens[i-2] == 'LOAD_LISTCOMP'):
|
|
if self.version >= 3.6:
|
|
# 3.6+ sometimes bundles all of the
|
|
# 'exprs' in the rule above into a
|
|
# tuple.
|
|
rule_pat = ("listcomp ::= load_closure LOAD_LISTCOMP %%s%s "
|
|
"expr GET_ITER CALL_FUNCTION_1" % (opname,))
|
|
self.add_make_function_rule(rule_pat, opname, token.attr, customize)
|
|
rule_pat = ("listcomp ::= %sLOAD_LISTCOMP %%s%s expr "
|
|
"GET_ITER CALL_FUNCTION_1" % ('expr ' * args_pos, opname))
|
|
self.add_make_function_rule(rule_pat, opname, token.attr, customize)
|
|
|
|
if is_pypy or (i >= 2 and tokens[i-2] == 'LOAD_LAMBDA'):
|
|
rule_pat = ('mklambda ::= %s%sLOAD_LAMBDA %%s%s' %
|
|
(('pos_arg ' * args_pos),
|
|
('kwarg ' * args_kw),
|
|
opname))
|
|
self.add_make_function_rule(rule_pat, opname, token.attr, customize)
|
|
continue
|
|
|
|
if self.version < 3.6:
|
|
args_pos, args_kw, annotate_args = token.attr
|
|
else:
|
|
args_pos, args_kw, annotate_args, closure = token.attr
|
|
|
|
if self.version < 3.3:
|
|
j = 1
|
|
else:
|
|
j = 2
|
|
|
|
if has_get_iter_call_function1:
|
|
rule_pat = ("generator_exp ::= %sload_genexpr %%s%s expr "
|
|
"GET_ITER CALL_FUNCTION_1" % ('pos_arg ' * args_pos, opname))
|
|
self.add_make_function_rule(rule_pat, opname, token.attr, customize)
|
|
|
|
if is_pypy or (i >= j and tokens[i-j] == 'LOAD_LISTCOMP'):
|
|
# In the tokens we saw:
|
|
# LOAD_LISTCOMP LOAD_CONST MAKE_FUNCTION (>= 3.3) or
|
|
# LOAD_LISTCOMP MAKE_FUNCTION (< 3.3) or
|
|
# and have GET_ITER CALL_FUNCTION_1
|
|
# Todo: For Pypy we need to modify this slightly
|
|
rule_pat = ("listcomp ::= %sLOAD_LISTCOMP %%s%s expr "
|
|
"GET_ITER CALL_FUNCTION_1" % ('expr ' * args_pos, opname))
|
|
self.add_make_function_rule(rule_pat, opname, token.attr, customize)
|
|
|
|
# FIXME: Fold test into add_make_function_rule
|
|
if is_pypy or (i >= j and tokens[i-j] == 'LOAD_LAMBDA'):
|
|
rule_pat = ('mklambda ::= %s%sLOAD_LAMBDA %%s%s' %
|
|
(('pos_arg ' * args_pos),
|
|
('kwarg ' * args_kw),
|
|
opname))
|
|
self.add_make_function_rule(rule_pat, opname, token.attr, customize)
|
|
|
|
if args_kw == 0:
|
|
kwargs = 'no_kwargs'
|
|
self.add_unique_rule("no_kwargs ::=", opname, token.attr, customize)
|
|
else:
|
|
kwargs = 'kwargs'
|
|
|
|
if self.version < 3.3:
|
|
# positional args after keyword args
|
|
rule = ('mkfunc ::= %s %s%s%s' %
|
|
(kwargs, 'pos_arg ' * args_pos, 'LOAD_CONST ',
|
|
opname))
|
|
self.add_unique_rule(rule, opname, token.attr, customize)
|
|
rule = ('mkfunc ::= %s%s%s' %
|
|
('pos_arg ' * args_pos, 'LOAD_CONST ',
|
|
opname))
|
|
elif self.version == 3.3:
|
|
# positional args after keyword args
|
|
rule = ('mkfunc ::= %s %s%s%s' %
|
|
(kwargs, 'pos_arg ' * args_pos, 'LOAD_CONST '*2,
|
|
opname))
|
|
elif self.version > 3.5:
|
|
# positional args before keyword args
|
|
rule = ('mkfunc ::= %s%s %s%s' %
|
|
('pos_arg ' * args_pos, kwargs, 'LOAD_CONST '*2,
|
|
opname))
|
|
elif self.version > 3.3:
|
|
# positional args before keyword args
|
|
rule = ('mkfunc ::= %s%s %s%s' %
|
|
('pos_arg ' * args_pos, kwargs, 'LOAD_CONST '*2,
|
|
opname))
|
|
else:
|
|
rule = ('mkfunc ::= %s%sexpr %s' %
|
|
(kwargs, 'pos_arg ' * args_pos, opname))
|
|
self.add_unique_rule(rule, opname, token.attr, customize)
|
|
|
|
if opname.startswith('MAKE_FUNCTION_A'):
|
|
if self.version >= 3.6:
|
|
rule = ('mkfunc_annotate ::= %s%sannotate_tuple LOAD_CONST LOAD_CONST %s' %
|
|
(('pos_arg ' * (args_pos)),
|
|
('call ' * (annotate_args-1)), opname))
|
|
self.add_unique_rule(rule, opname, token.attr, customize)
|
|
rule = ('mkfunc_annotate ::= %s%sannotate_tuple LOAD_CONST LOAD_CONST %s' %
|
|
(('pos_arg ' * (args_pos)),
|
|
('annotate_arg ' * (annotate_args-1)), opname))
|
|
if self.version >= 3.3:
|
|
# Normally we remove EXTENDED_ARG from the opcodes, but in the case of
|
|
# annotated functions can use the EXTENDED_ARG tuple to signal we have an annotated function.
|
|
# Yes this is a little hacky
|
|
rule = ('mkfunc_annotate ::= %s%sannotate_tuple LOAD_CONST LOAD_CONST EXTENDED_ARG %s' %
|
|
(('pos_arg ' * (args_pos)),
|
|
('call ' * (annotate_args-1)), opname))
|
|
self.add_unique_rule(rule, opname, token.attr, customize)
|
|
rule = ('mkfunc_annotate ::= %s%sannotate_tuple LOAD_CONST LOAD_CONST EXTENDED_ARG %s' %
|
|
(('pos_arg ' * (args_pos)),
|
|
('annotate_arg ' * (annotate_args-1)), opname))
|
|
else:
|
|
# See above comment about use of EXTENDED_ARG
|
|
rule = ('mkfunc_annotate ::= %s%sannotate_tuple LOAD_CONST EXTENDED_ARG %s' %
|
|
(('pos_arg ' * (args_pos)),
|
|
('annotate_arg ' * (annotate_args-1)), opname))
|
|
self.add_unique_rule(rule, opname, token.attr, customize)
|
|
rule = ('mkfunc_annotate ::= %s%sannotate_tuple LOAD_CONST EXTENDED_ARG %s' %
|
|
(('pos_arg ' * (args_pos)),
|
|
('call ' * (annotate_args-1)), opname))
|
|
self.addRule(rule, nop_func)
|
|
elif opname == 'RETURN_VALUE_LAMBDA':
|
|
self.addRule("""
|
|
return_lambda ::= ret_expr RETURN_VALUE_LAMBDA
|
|
""", nop_func)
|
|
custom_ops_processed.add(opname)
|
|
elif opname == 'RAISE_VARARGS_0':
|
|
self.addRule("""
|
|
stmt ::= raise_stmt0
|
|
raise_stmt0 ::= RAISE_VARARGS_0
|
|
""", nop_func)
|
|
custom_ops_processed.add(opname)
|
|
elif opname == 'RAISE_VARARGS_1':
|
|
self.addRule("""
|
|
stmt ::= raise_stmt1
|
|
raise_stmt1 ::= expr RAISE_VARARGS_1
|
|
""", nop_func)
|
|
custom_ops_processed.add(opname)
|
|
elif opname == 'RAISE_VARARGS_2':
|
|
self.addRule("""
|
|
stmt ::= raise_stmt2
|
|
raise_stmt2 ::= expr expr RAISE_VARARGS_2
|
|
""", nop_func)
|
|
custom_ops_processed.add(opname)
|
|
elif opname == 'SETUP_EXCEPT':
|
|
self.addRule("""
|
|
try_except ::= SETUP_EXCEPT suite_stmts_opt POP_BLOCK
|
|
except_handler opt_come_from_except
|
|
|
|
tryelsestmt ::= SETUP_EXCEPT suite_stmts_opt POP_BLOCK
|
|
except_handler else_suite come_from_except_clauses
|
|
|
|
tryelsestmt ::= SETUP_EXCEPT suite_stmts_opt POP_BLOCK
|
|
except_handler else_suite come_froms
|
|
|
|
tryelsestmtl ::= SETUP_EXCEPT suite_stmts_opt POP_BLOCK
|
|
except_handler else_suitel come_from_except_clauses
|
|
|
|
stmt ::= tryelsestmtl3
|
|
tryelsestmtl3 ::= SETUP_EXCEPT suite_stmts_opt POP_BLOCK
|
|
except_handler COME_FROM else_suitel
|
|
opt_come_from_except
|
|
""", nop_func)
|
|
custom_ops_processed.add(opname)
|
|
elif opname_base in ('UNPACK_EX',):
|
|
before_count, after_count = token.attr
|
|
rule = 'unpack ::= ' + opname + ' store' * (before_count + after_count + 1)
|
|
self.addRule(rule, nop_func)
|
|
elif opname_base in ('UNPACK_TUPLE', 'UNPACK_SEQUENCE'):
|
|
rule = 'unpack ::= ' + opname + ' store' * token.attr
|
|
self.addRule(rule, nop_func)
|
|
elif opname_base == 'UNPACK_LIST':
|
|
rule = 'unpack_list ::= ' + opname + ' store' * token.attr
|
|
self.addRule(rule, nop_func)
|
|
custom_ops_processed.add(opname)
|
|
pass
|
|
pass
|
|
|
|
self.check_reduce['aug_assign1'] = 'AST'
|
|
self.check_reduce['aug_assign2'] = 'AST'
|
|
self.check_reduce['while1stmt'] = 'noAST'
|
|
self.check_reduce['while1elsestmt'] = 'noAST'
|
|
self.check_reduce['ifelsestmt'] = 'AST'
|
|
self.check_reduce['annotate_tuple'] = 'noAST'
|
|
self.check_reduce['kwarg'] = 'noAST'
|
|
self.check_reduce['try_except'] = 'AST'
|
|
|
|
# FIXME: remove parser errors caused by the below
|
|
# self.check_reduce['while1elsestmt'] = 'noAST'
|
|
return
|
|
|
|
def reduce_is_invalid(self, rule, ast, tokens, first, last):
|
|
lhs = rule[0]
|
|
if lhs in ('aug_assign1', 'aug_assign2') and ast[0][0] == 'and':
|
|
return True
|
|
elif lhs == 'annotate_tuple':
|
|
return not isinstance(tokens[first].attr, tuple)
|
|
elif lhs == 'kwarg':
|
|
arg = tokens[first].attr
|
|
if PYTHON3:
|
|
return not isinstance(arg, str)
|
|
else:
|
|
return not (isinstance(arg, str) or isinstance(arg, unicode))
|
|
elif lhs == 'while1elsestmt':
|
|
|
|
n = len(tokens)
|
|
if last == n:
|
|
# Adjust for fuzziness in parsing
|
|
last -= 1
|
|
|
|
if tokens[last] == 'COME_FROM_LOOP':
|
|
last -= 1
|
|
elif tokens[last-1] == 'COME_FROM_LOOP':
|
|
last -= 2
|
|
if tokens[last] in ('JUMP_BACK', 'CONTINUE'):
|
|
# These indicate inside a loop, but token[last]
|
|
# should not be in a loop.
|
|
# FIXME: Not quite right: refine by using target
|
|
return True
|
|
|
|
# if SETUP_LOOP target spans the else part, then this is
|
|
# not while1else. Also do for whileTrue?
|
|
last += 1
|
|
while last < n and isinstance(tokens[last].offset, str):
|
|
last += 1
|
|
if last == n:
|
|
return False
|
|
# 3.8+ Doesn't have SETUP_LOOP
|
|
return self.version < 3.8 and tokens[first].attr > tokens[last].offset
|
|
|
|
elif rule == ('try_except',
|
|
('SETUP_EXCEPT', 'suite_stmts_opt', 'POP_BLOCK',
|
|
'except_handler', 'opt_come_from_except')):
|
|
come_from_except = ast[-1]
|
|
if come_from_except[0] == 'COME_FROM':
|
|
# There should be at last two COME_FROMs, one from an
|
|
# exception handler and one from the try. Otherwise
|
|
# we have a try/else.
|
|
return True
|
|
pass
|
|
elif lhs == 'while1stmt':
|
|
|
|
# If there is a fall through to the COME_FROM_LOOP. then this is
|
|
# not a while 1. So the instruction before should either be a
|
|
# JUMP_BACK or the instruction before should not be the target of a
|
|
# jump. (Well that last clause i not quite right; that target could be
|
|
# from dead code. Ugh. We need a more uniform control flow analysis.)
|
|
if last == len(tokens) or tokens[last-1] == 'COME_FROM_LOOP':
|
|
cfl = last-1
|
|
else:
|
|
cfl = last
|
|
assert tokens[cfl] == 'COME_FROM_LOOP'
|
|
|
|
if tokens[cfl-1] != 'JUMP_BACK':
|
|
cfl_offset = tokens[cfl-1].offset
|
|
insn = next(i for i in self.insts if cfl_offset == i.offset)
|
|
if insn and insn.is_jump_target:
|
|
return True
|
|
|
|
# Check that the SETUP_LOOP jumps to the offset after the
|
|
# COME_FROM_LOOP
|
|
if (0 <= last < len(tokens)
|
|
and tokens[last] in ('COME_FROM_LOOP', 'JUMP_BACK')):
|
|
# jump_back should be right before COME_FROM_LOOP?
|
|
last += 1
|
|
while last < len(tokens) and isinstance(tokens[last].offset, str):
|
|
last += 1
|
|
if last < len(tokens):
|
|
offset = tokens[last].offset
|
|
assert tokens[first] == 'SETUP_LOOP'
|
|
if offset != tokens[first].attr:
|
|
return True
|
|
return False
|
|
elif rule == ('ifelsestmt',
|
|
('testexpr', 'c_stmts_opt', 'jump_forward_else', 'else_suite', '_come_froms')):
|
|
# Make sure the highest/smallest "come from" offset comes inside the "if".
|
|
come_froms = ast[-1]
|
|
if not isinstance(come_froms, Token):
|
|
return tokens[first].offset > come_froms[-1].attr
|
|
return False
|
|
|
|
return False
|
|
|
|
class Python30Parser(Python3Parser):
|
|
|
|
def p_30(self, args):
|
|
"""
|
|
jmp_true ::= JUMP_IF_TRUE_OR_POP POP_TOP
|
|
_ifstmts_jump ::= c_stmts_opt JUMP_FORWARD POP_TOP COME_FROM
|
|
"""
|
|
|
|
class Python3ParserSingle(Python3Parser, PythonParserSingle):
|
|
pass
|
|
|
|
def info(args):
|
|
# Check grammar
|
|
p = Python3Parser()
|
|
if len(args) > 0:
|
|
arg = args[0]
|
|
if arg == '3.5':
|
|
from uncompyle6.parser.parse35 import Python35Parser
|
|
p = Python35Parser()
|
|
elif arg == '3.3':
|
|
from uncompyle6.parser.parse33 import Python33Parser
|
|
p = Python33Parser()
|
|
elif arg == '3.2':
|
|
from uncompyle6.parser.parse32 import Python32Parser
|
|
p = Python32Parser()
|
|
elif arg == '3.0':
|
|
p = Python30Parser()
|
|
p.check_grammar()
|
|
if len(sys.argv) > 1 and sys.argv[1] == 'dump':
|
|
print('-' * 50)
|
|
p.dump_grammar()
|
|
|
|
if __name__ == '__main__':
|
|
import sys
|
|
info(sys.argv)
|