jinx/python-uncompyle6
1
Fork 0
mirror of https://github.com/rocky/python-uncompyle6.git synced 2025-08-03 00:45:53 +08:00
Code Issues Packages Projects Releases Wiki Activity
Files
36ffd4c31f8f3201eeb68a66c0e0f8b54c561356
python-uncompyle6/uncompyle6/scanners/scanner35.py
rocky 36ffd4c31f Handle Python 3 yield from 
Start dealing with MAKE_FUNCTION flags - not done yet.
2016-05-07 11:33:18 -04:00

307 lines
14 KiB
Python
Raw Blame History

# Copyright (c) 2016 by Rocky Bernstein
"""
Python 3.5 bytecode scanner/deparser
This overlaps Python's 3.5's dis module, and in fact in some cases
we just fall back to that. But the intent is that it can be run from
Python 2 and other versions of Python. Also, we save token information
for later use in deparsing.
"""
from __future__ import print_function
import inspect
from array import array
import uncompyle6.scanners.scanner3 as scan3
import uncompyle6.scanners.dis35 as dis35
from uncompyle6.code import iscode
from uncompyle6.scanner import Token
import uncompyle6.opcodes.opcode_35
# verify uses JUMP_OPs from here
JUMP_OPs = uncompyle6.opcodes.opcode_35.JUMP_OPs
from uncompyle6.opcodes.opcode_35 import *
class Scanner35(scan3.Scanner3):
# Note: we can't use built-in disassembly routines, unless
# we do post-processing like we do here.
def disassemble(self, co, classname=None,
code_objects={}):
# Container for tokens
tokens = []
customize = {}
self.code = array('B', co.co_code)
self.build_lines_data(co)
self.build_prev_op()
# Get jump targets
# Format: {target offset: [jump offsets]}
jump_targets = self.find_jump_targets()
bytecode = dis35.Bytecode(co)
# self.lines contains (block,addrLastInstr)
if classname:
classname = '_' + classname.lstrip('_') + '__'
def unmangle(name):
if name.startswith(classname) and name[-2:] != '__':
return name[len(classname) - 2:]
return name
# free = [ unmangle(name) for name in (co.co_cellvars + co.co_freevars) ]
# names = [ unmangle(name) for name in co.co_names ]
# varnames = [ unmangle(name) for name in co.co_varnames ]
else:
# free = co.co_cellvars + co.co_freevars
# names = co.co_names
# varnames = co.co_varnames
pass
# Scan for assertions. Later we will
# turn 'LOAD_GLOBAL' to 'LOAD_ASSERT' for those
# assertions
self.load_asserts = set()
bs = list(bytecode)
n = len(bs)
for i in range(n):
inst = bs[i]
if inst.opname == 'POP_JUMP_IF_TRUE' and i+1 < n:
next_inst = bs[i+1]
if (next_inst.opname == 'LOAD_GLOBAL' and
next_inst.argval == 'AssertionError'):
self.load_asserts.add(next_inst.offset)
for inst in bytecode:
if inst.offset in jump_targets:
jump_idx = 0
for jump_offset in jump_targets[inst.offset]:
tokens.append(Token('COME_FROM', None, repr(jump_offset),
offset='%s_%s' % (inst.offset, jump_idx)))
jump_idx += 1
pass
pass
pattr = inst.argrepr
opname = inst.opname
if opname in ['LOAD_CONST']:
const = inst.argval
if iscode(const):
if const.co_name == '<lambda>':
opname = 'LOAD_LAMBDA'
elif const.co_name == '<genexpr>':
opname = 'LOAD_GENEXPR'
elif const.co_name == '<dictcomp>':
opname = 'LOAD_DICTCOMP'
elif const.co_name == '<setcomp>':
opname = 'LOAD_SETCOMP'
elif const.co_name == '<listcomp>':
opname = 'LOAD_LISTCOMP'
# verify() uses 'pattr' for comparison, since 'attr'
# now holds Code(const) and thus can not be used
# for comparison (todo: think about changing this)
# pattr = 'code_object @ 0x%x %s->%s' %\
# (id(const), const.co_filename, const.co_name)
pattr = '<code_object ' + const.co_name + '>'
else:
pattr = const
pass
elif opname in ('BUILD_LIST', 'BUILD_TUPLE', 'BUILD_SET', 'BUILD_SLICE',
'BUILD_MAP',
'UNPACK_SEQUENCE',
'MAKE_FUNCTION', 'MAKE_CLOSURE',
'DUP_TOPX', 'RAISE_VARARGS'
):
pos_args = inst.argval
if inst.opname == 'MAKE_FUNCTION':
argc = inst.argval
pos_args = (argc & 0xFF)
name_pair_args = (argc >> 8) & 0xFF
if name_pair_args > 0:
opname = 'MAKE_FUNCTION_N%d' % name_pair_args
pass
annotate_args = (argc >> 16) & 0x7FFF
if annotate_args > 0:
opname = '%s_A_%d' % [op_name, annotate_args]
pass
elif inst.opname != 'BUILD_SLICE':
customize[opname] = pos_args
opname = '%s_%d' % (opname, pos_args)
elif opname == 'JUMP_ABSOLUTE':
pattr = inst.argval
target = self.get_target(inst.offset)
if target < inst.offset:
if (inst.offset in self.stmts and
self.code[inst.offset+3] not in (END_FINALLY, POP_BLOCK)
and inst.offset not in self.not_continue):
opname = 'CONTINUE'
else:
opname = 'JUMP_BACK'
elif inst.offset in self.load_asserts:
opname = 'LOAD_ASSERT'
tokens.append(
Token(
type_ = opname,
attr = inst.argval,
pattr = pattr,
offset = inst.offset,
linestart = inst.starts_line,
)
)
pass
return tokens, {}
# FIXME: merge with scanner3 code
def detect_structure(self, offset):
"""
Detect structures and their boundaries to fix optimizied jumps
in python2.3+
"""
code = self.code
op = code[offset]
# Detect parent structure
parent = self.structs[0]
start = parent['start']
end = parent['end']
# Pick inner-most parent for our offset
for struct in self.structs:
curent_start = struct['start']
curent_end = struct['end']
if (curent_start <= offset < curent_end) and (curent_start >= start and curent_end <= end):
start = curent_start
end = curent_end
parent = struct
pass
if op in (POP_JUMP_IF_FALSE, POP_JUMP_IF_TRUE):
start = offset + self.op_size(op)
target = self.get_target(offset)
rtarget = self.restrict_to_parent(target, parent)
prev_op = self.prev_op
# Do not let jump to go out of parent struct bounds
if target != rtarget and parent['type'] == 'and/or':
self.fixed_jumps[offset] = rtarget
return
# Does this jump to right after another cond jump?
# If so, it's part of a larger conditional
if (code[prev_op[target]] in (JUMP_IF_FALSE_OR_POP, JUMP_IF_TRUE_OR_POP,
POP_JUMP_IF_FALSE, POP_JUMP_IF_TRUE)) and (target > offset):
self.fixed_jumps[offset] = prev_op[target]
self.structs.append({'type': 'and/or',
'start': start,
'end': prev_op[target]})
return
# Is it an and inside if block
if op == POP_JUMP_IF_FALSE:
# Search for other POP_JUMP_IF_FALSE targetting the same op,
# in current statement, starting from current offset, and filter
# everything inside inner 'or' jumps and midline ifs
match = self.rem_or(start, self.next_stmt[offset], POP_JUMP_IF_FALSE, target)
match = self.remove_mid_line_ifs(match)
# If we still have any offsets in set, start working on it
if match:
if (code[prev_op[rtarget]] in (JUMP_FORWARD, JUMP_ABSOLUTE) and prev_op[rtarget] not in self.stmts and
self.restrict_to_parent(self.get_target(prev_op[rtarget]), parent) == rtarget):
if (code[prev_op[prev_op[rtarget]]] == JUMP_ABSOLUTE and self.remove_mid_line_ifs([offset]) and
target == self.get_target(prev_op[prev_op[rtarget]]) and
(prev_op[prev_op[rtarget]] not in self.stmts or self.get_target(prev_op[prev_op[rtarget]]) > prev_op[prev_op[rtarget]]) and
1 == len(self.remove_mid_line_ifs(self.rem_or(start, prev_op[prev_op[rtarget]], (POP_JUMP_IF_FALSE, POP_JUMP_IF_TRUE), target)))):
pass
elif (code[prev_op[prev_op[rtarget]]] == RETURN_VALUE and self.remove_mid_line_ifs([offset]) and
1 == (len(set(self.remove_mid_line_ifs(self.rem_or(start, prev_op[prev_op[rtarget]],
(POP_JUMP_IF_FALSE, POP_JUMP_IF_TRUE), target))) |
set(self.remove_mid_line_ifs(self.rem_or(start, prev_op[prev_op[rtarget]],
(POP_JUMP_IF_FALSE, POP_JUMP_IF_TRUE, JUMP_ABSOLUTE),
prev_op[rtarget], True)))))):
pass
else:
fix = None
jump_ifs = self.all_instr(start, self.next_stmt[offset], POP_JUMP_IF_FALSE)
last_jump_good = True
for j in jump_ifs:
if target == self.get_target(j):
if self.lines[j].next == j + 3 and last_jump_good:
fix = j
break
else:
last_jump_good = False
self.fixed_jumps[offset] = fix or match[-1]
return
else:
self.fixed_jumps[offset] = match[-1]
return
# op == POP_JUMP_IF_TRUE
else:
next = self.next_stmt[offset]
if prev_op[next] == offset:
pass
elif code[next] in (JUMP_FORWARD, JUMP_ABSOLUTE) and target == self.get_target(next):
if code[prev_op[next]] == POP_JUMP_IF_FALSE:
if code[next] == JUMP_FORWARD or target != rtarget or code[prev_op[prev_op[rtarget]]] not in (JUMP_ABSOLUTE, RETURN_VALUE):
self.fixed_jumps[offset] = prev_op[next]
return
elif (code[next] == JUMP_ABSOLUTE and code[target] in (JUMP_ABSOLUTE, JUMP_FORWARD) and
self.get_target(target) == self.get_target(next)):
self.fixed_jumps[offset] = prev_op[next]
return
# Don't add a struct for a while test, it's already taken care of
if offset in self.ignore_if:
return
if (code[prev_op[rtarget]] == JUMP_ABSOLUTE and prev_op[rtarget] in self.stmts and
prev_op[rtarget] != offset and prev_op[prev_op[rtarget]] != offset and
not (code[rtarget] == JUMP_ABSOLUTE and code[rtarget+3] == POP_BLOCK and code[prev_op[prev_op[rtarget]]] != JUMP_ABSOLUTE)):
rtarget = prev_op[rtarget]
# Does the if jump just beyond a jump op, then this is probably an if statement
if code[prev_op[rtarget]] in (JUMP_ABSOLUTE, JUMP_FORWARD):
if_end = self.get_target(prev_op[rtarget])
# Is this a loop not an if?
if (if_end < prev_op[rtarget]) and (code[prev_op[if_end]] == SETUP_LOOP):
if(if_end > start):
return
end = self.restrict_to_parent(if_end, parent)
self.structs.append({'type': 'if-then',
'start': start,
'end': prev_op[rtarget]})
self.not_continue.add(prev_op[rtarget])
if rtarget < end:
self.structs.append({'type': 'if-else',
'start': rtarget,
'end': end})
elif code[prev_op[rtarget]] == RETURN_VALUE:
self.structs.append({'type': 'if-then',
'start': start,
'end': rtarget})
self.return_end_ifs.add(prev_op[rtarget])
elif op in (JUMP_IF_FALSE_OR_POP, JUMP_IF_TRUE_OR_POP):
target = self.get_target(offset)
if target > offset:
unop_target = self.last_instr(offset, target, JUMP_FORWARD, target)
if unop_target and code[unop_target+3] != ROT_TWO:
self.fixed_jumps[offset] = unop_target
else:
self.fixed_jumps[offset] = self.restrict_to_parent(target, parent)
if __name__ == "__main__":
co = inspect.currentframe().f_code
tokens, customize = Scanner35(3.5).disassemble(co)
for t in tokens:
print(t)
pass
Reference in New Issue View Git Blame Copy Permalink