Merge branch 'feature/UpdateJs2Py' into dev

This commit is contained in:
JackDandy 2023-02-09 14:37:37 +00:00
commit 556ad8af3c
15 changed files with 266 additions and 98 deletions

View file

@ -9,6 +9,7 @@
* Update diskcache 5.1.0 (40ce0de) to 5.4.0 (1cb1425)
* Update feedparser 6.0.1 (98d189fa) to 6.0.10 (5fcb3ae)
* Update humanize 3.5.0 (b6b0ea5) to 4.0.0 (a1514eb)
* Update Js2Py 0.70 (92250a4) to 0.74 (2e017b8)
* Update profilehooks module 1.12.0 (3ee1f60) to 1.12.1 (c3fc078)
* Update Rarfile 4.0 (55fe778) to 4.1a1 (8a72967)
* Update UnRar x64 for Windows 6.11 to 6.20

View file

@ -118,20 +118,39 @@ class PyJsDate(PyJs):
def parse_date(py_string): # todo support all date string formats
date_formats = (
"%Y-%m-%d",
"%m/%d/%Y",
"%b %d %Y",
)
# Supports these hour formats and with or hour.
hour_formats = (
"T%H:%M:%S.%f",
"T%H:%M:%S",
) + ('',)
# Supports with or without Z indicator.
z_formats = ("Z",) + ('',)
supported_formats = [
d + t + z
for d in date_formats
for t in hour_formats
for z in z_formats
]
for date_format in supported_formats:
try:
try:
dt = datetime.datetime.strptime(py_string, "%Y-%m-%dT%H:%M:%S.%fZ")
except:
dt = datetime.datetime.strptime(py_string, "%Y-%m-%dT%H:%M:%SZ")
dt = datetime.datetime.strptime(py_string, date_format)
except ValueError:
continue
else:
return MakeDate(
MakeDay(Js(dt.year), Js(dt.month - 1), Js(dt.day)),
MakeTime(
Js(dt.hour), Js(dt.minute), Js(dt.second),
Js(dt.microsecond // 1000)))
except:
raise MakeError(
'TypeError',
'Could not parse date %s - unsupported date format. Currently only supported format is RFC3339 utc. Sorry!'
'Could not parse date %s - unsupported date format. Currently only supported formats are RFC3339 utc, ISO Date, Short Date, and Long Date. Sorry!'
% py_string)

View file

@ -36,8 +36,8 @@ def DaylightSavingTA(t):
return t
try:
return int(
LOCAL_ZONE.dst(datetime.datetime.utcfromtimestamp(
t // 1000)).seconds) * 1000
LOCAL_ZONE.dst(datetime.datetime(1970, 1, 1) + datetime.timedelta(
seconds=t // 1000)).seconds) * 1000
except:
warnings.warn(
'Invalid datetime date, assumed DST time, may be inaccurate...',

View file

@ -53,7 +53,7 @@ def write_file_contents(path_or_file, contents):
if hasattr(path_or_file, 'write'):
path_or_file.write(contents)
else:
with open(path_as_local(path_or_file), 'w') as f:
with codecs.open(path_as_local(path_or_file), "w", "utf-8") as f:
f.write(contents)
@ -238,6 +238,10 @@ class EvalJs(object):
self.execute_debug(code)
return self['PyJsEvalResult']
@property
def context(self):
return self._context
def __getattr__(self, var):
return getattr(self._var, var)
@ -268,14 +272,3 @@ class EvalJs(object):
else:
sys.stderr.write('EXCEPTION: ' + str(e) + '\n')
time.sleep(0.01)
#print x
if __name__ == '__main__':
#with open('C:\Users\Piotrek\Desktop\esprima.js', 'rb') as f:
# x = f.read()
e = EvalJs()
e.execute('square(x)')
#e.execute(x)
e.console()

View file

@ -6,7 +6,7 @@ def console():
@Js
def log():
print(arguments[0])
print(" ".join(repr(element) for element in arguments.to_list()))
console.put('log', log)
console.put('debug', log)

View file

@ -0,0 +1 @@
from .seval import eval_js_vm

View file

@ -602,11 +602,12 @@ class PyJsDate(PyJs):
# todo fix this problematic datetime part
def to_local_dt(self):
return datetime.datetime.utcfromtimestamp(
self.UTCToLocal(self.value) // 1000)
return datetime.datetime(1970, 1, 1) + datetime.timedelta(
seconds=self.UTCToLocal(self.value) // 1000)
def to_utc_dt(self):
return datetime.datetime.utcfromtimestamp(self.value // 1000)
return datetime.datetime(1970, 1, 1) + datetime.timedelta(
seconds=self.value // 1000)
def local_strftime(self, pattern):
if self.value is NaN:

View file

@ -38,8 +38,8 @@ def DaylightSavingTA(t):
return t
try:
return int(
LOCAL_ZONE.dst(datetime.datetime.utcfromtimestamp(
t // 1000)).seconds) * 1000
LOCAL_ZONE.dst(datetime.datetime(1970, 1, 1) + datetime.timedelta(
seconds=t // 1000)).seconds) * 1000
except:
warnings.warn(
'Invalid datetime date, assumed DST time, may be inaccurate...',

View file

@ -798,7 +798,7 @@ OP_CODES = {}
g = ''
for g in globals():
try:
if not issubclass(globals()[g], OP_CODE) or g is 'OP_CODE':
if not issubclass(globals()[g], OP_CODE) or g == 'OP_CODE':
continue
except:
continue

View file

@ -22,6 +22,11 @@ def replacement_template(rep, source, span, npar):
res += '$'
n += 2
continue
elif rep[n + 1] == '&':
# replace with matched string
res += source[span[0]:span[1]]
n += 2
continue
elif rep[n + 1] == '`':
# replace with string that is BEFORE match
res += source[:span[0]]

View file

@ -4,10 +4,13 @@ import subprocess, os, codecs, glob
from .evaljs import translate_js, DEFAULT_HEADER
from .translators.friendly_nodes import is_valid_py_name
import six
import tempfile
import hashlib
import random
DID_INIT = False
DIRNAME = os.path.dirname(os.path.abspath(__file__))
PY_NODE_MODULES_PATH = os.path.join(DIRNAME, 'py_node_modules')
DIRNAME = tempfile.mkdtemp()
PY_NODE_MODULES_PATH = os.path.join(os.path.dirname(os.path.abspath(__file__)), 'py_node_modules')
def _init():
@ -72,8 +75,10 @@ def _get_and_translate_npm_module(module_name, include_polyfill=False, update=Fa
if not os.path.exists(os.path.join(PY_NODE_MODULES_PATH,
module_filename)) or update:
_init()
in_file_name = 'tmp0in439341018923js2py.js'
out_file_name = 'tmp0out439341018923js2py.js'
module_hash = hashlib.sha1(module_name.encode("utf-8")).hexdigest()[:15]
version = random.randrange(10000000000000)
in_file_name = 'in_%s_%d.js' % (module_hash, version)
out_file_name = 'out_%s_%d.js' % (module_hash, version)
code = ADD_TO_GLOBALS_FUNC
if include_polyfill:
code += "\n;require('babel-polyfill');\n"
@ -106,7 +111,7 @@ def _get_and_translate_npm_module(module_name, include_polyfill=False, update=Fa
with codecs.open(os.path.join(DIRNAME, out_file_name), "r",
"utf-8") as f:
js_code = f.read()
os.remove(os.path.join(DIRNAME, out_file_name))
print("Bundled JS library dumped at: %s" % os.path.join(DIRNAME, out_file_name))
if len(js_code) < 50:
raise RuntimeError("Candidate JS bundle too short - likely browserify issue.")
js_code += GET_FROM_GLOBALS_FUNC

View file

@ -17,6 +17,11 @@ def replacement_template(rep, source, span, npar):
res += '$'
n += 2
continue
elif rep[n + 1] == '&':
# replace with matched string
res += source[span[0]:span[1]]
n += 2
continue
elif rep[n + 1] == '`':
# replace with string that is BEFORE match
res += source[:span[0]]

View file

@ -14,26 +14,36 @@ if six.PY3:
LINE_LEN_LIMIT = 400 # 200 # or any other value - the larger the smaller probability of errors :)
class ForController:
class LoopController:
def __init__(self):
self.inside = [False]
self.update = ''
self.update = [""]
self.label_to_update_idx = {}
def enter_for(self, update):
self.inside.append(True)
self.update = update
def enter(self, update=""):
self.update.append(update)
def leave_for(self):
self.inside.pop()
def leave(self):
self.update.pop()
def get_update(self, label=None):
if label is None:
return self.update[-1]
if label not in self.label_to_update_idx:
raise SyntaxError("Undefined label %s" % label)
if self.label_to_update_idx[label] >= len(self.update):
raise SyntaxError("%s is not a iteration statement label?" % label)
return self.update[self.label_to_update_idx[label]]
def register_label(self, label):
if label in self.label_to_update_idx:
raise SyntaxError("label %s already used")
self.label_to_update_idx[label] = len(self.update)
def deregister_label(self, label):
del self.label_to_update_idx[label]
def enter_other(self):
self.inside.append(False)
def leave_other(self):
self.inside.pop()
def is_inside(self):
return self.inside[-1]
class InlineStack:
@ -86,9 +96,10 @@ class ContextStack:
def clean_stacks():
global Context, inline_stack
global Context, inline_stack, loop_controller
Context = ContextStack()
inline_stack = InlineStack()
loop_controller = LoopController()
def to_key(literal_or_identifier):
@ -374,9 +385,14 @@ def BreakStatement(type, label):
def ContinueStatement(type, label):
if label:
return 'raise %s("Continued")\n' % (get_continue_label(label['name']))
maybe_update_expr = loop_controller.get_update(label=label['name'])
continue_stmt = 'raise %s("Continued")\n' % (get_continue_label(label['name']))
else:
return 'continue\n'
maybe_update_expr = loop_controller.get_update()
continue_stmt = "continue\n"
if maybe_update_expr:
return "# continue update\n%s\n%s" % (maybe_update_expr, continue_stmt)
return continue_stmt
def ReturnStatement(type, argument):
@ -393,24 +409,28 @@ def DebuggerStatement(type):
def DoWhileStatement(type, body, test):
inside = trans(body) + 'if not %s:\n' % trans(test) + indent('break\n')
loop_controller.enter()
body_code = trans(body)
loop_controller.leave()
inside = body_code + 'if not %s:\n' % trans(test) + indent('break\n')
result = 'while 1:\n' + indent(inside)
return result
def ForStatement(type, init, test, update, body):
update = indent(trans(update)) if update else ''
update = trans(update) if update else ''
init = trans(init) if init else ''
if not init.endswith('\n'):
init += '\n'
test = trans(test) if test else '1'
loop_controller.enter(update)
if not update:
result = '#for JS loop\n%swhile %s:\n%s%s\n' % (
init, test, indent(trans(body)), update)
else:
result = '#for JS loop\n%swhile %s:\n' % (init, test)
body = 'try:\n%sfinally:\n %s\n' % (indent(trans(body)), update)
result += indent(body)
result += indent("%s# update\n%s\n" % (trans(body), update))
loop_controller.leave()
return result
@ -429,7 +449,9 @@ def ForInStatement(type, left, right, body, each):
name = left['name']
else:
raise RuntimeError('Unusual ForIn loop')
loop_controller.enter()
res += indent('var.put(%s, PyJsTemp)\n' % repr(name) + trans(body))
loop_controller.leave()
return res
@ -445,20 +467,23 @@ def IfStatement(type, test, consequent, alternate):
def LabeledStatement(type, label, body):
# todo consider using smarter approach!
label_name = label['name']
loop_controller.register_label(label_name)
inside = trans(body)
loop_controller.deregister_label(label_name)
defs = ''
if is_iteration_statement(body) and (inside.startswith('while ') or inside.startswith(
'for ') or inside.startswith('#for')):
# we have to add contine label as well...
# 3 or 1 since #for loop type has more lines before real for.
sep = 1 if not inside.startswith('#for') else 3
cont_label = get_continue_label(label['name'])
cont_label = get_continue_label(label_name)
temp = inside.split('\n')
injected = 'try:\n' + '\n'.join(temp[sep:])
injected += 'except %s:\n pass\n' % cont_label
inside = '\n'.join(temp[:sep]) + '\n' + indent(injected)
defs += 'class %s(Exception): pass\n' % cont_label
break_label = get_break_label(label['name'])
break_label = get_break_label(label_name)
inside = 'try:\n%sexcept %s:\n pass\n' % (indent(inside), break_label)
defs += 'class %s(Exception): pass\n' % break_label
return defs + inside
@ -553,7 +578,11 @@ def VariableDeclaration(type, declarations, kind):
def WhileStatement(type, test, body):
result = 'while %s:\n' % trans(test) + indent(trans(body))
test_code = trans(test)
loop_controller.enter()
body_code = trans(body)
loop_controller.leave()
result = 'while %s:\n' % test_code + indent(body_code)
return result

View file

@ -55,16 +55,19 @@ def dbg(x):
"""does nothing, legacy dummy function"""
return ''
# Another way of doing that would be with my auto esprima translation but its much slower:
# parsed = esprima.parse(js).to_dict()
def pyjsparser_parse_fn(code):
parser = pyjsparser.PyJsParser()
return parser.parse(code)
def translate_js(js, HEADER=DEFAULT_HEADER, use_compilation_plan=False):
def translate_js(js, HEADER=DEFAULT_HEADER, use_compilation_plan=False, parse_fn=pyjsparser_parse_fn):
"""js has to be a javascript source code.
returns equivalent python code."""
if use_compilation_plan and not '//' in js and not '/*' in js:
return translate_js_with_compilation_plan(js, HEADER=HEADER)
parser = pyjsparser.PyJsParser()
parsed = parser.parse(js) # js to esprima syntax tree
# Another way of doing that would be with my auto esprima translation but its much slower and causes import problems:
# parsed = esprima.parse(js).to_dict()
parsed = parse_fn(js)
translating_nodes.clean_stacks()
return HEADER + translating_nodes.trans(
parsed) # syntax tree to python code

View file

@ -26,17 +26,19 @@ def fix_js_args(func):
return func
code = append_arguments(six.get_function_code(func), ('this', 'arguments'))
return types.FunctionType(
result = types.FunctionType(
code,
six.get_function_globals(func),
func.__name__,
closure=six.get_function_closure(func))
return result
def append_arguments(code_obj, new_locals):
co_varnames = code_obj.co_varnames # Old locals
co_names = code_obj.co_names # Old globals
co_names += tuple(e for e in new_locals if e not in co_names)
new_args = tuple(e for e in new_locals if e not in co_names)
co_names += new_args
co_argcount = code_obj.co_argcount # Argument count
co_code = code_obj.co_code # The actual bytecode as a string
@ -76,26 +78,51 @@ def append_arguments(code_obj, new_locals):
names_to_varnames = dict(
(co_names.index(name), varnames.index(name)) for name in new_locals)
is_new_bytecode = sys.version_info >= (3, 11)
# Now we modify the actual bytecode
modified = []
drop_future_cache = False
for inst in instructions(code_obj):
if is_new_bytecode and inst.opname == "CACHE":
assert inst.arg == 0
if not drop_future_cache:
modified.extend(write_instruction(inst.opcode, inst.arg))
else:
# We need to inject NOOP to not break jumps :(
modified.extend(write_instruction(dis.opmap["NOP"], 0))
continue
op, arg = inst.opcode, inst.arg
# If the instruction is a LOAD_GLOBAL, we have to check to see if
# it's one of the globals that we are replacing. Either way,
# update its arg using the appropriate dict.
drop_future_cache = False
if inst.opcode == LOAD_GLOBAL:
if inst.arg in names_to_varnames:
idx = inst.arg
if is_new_bytecode:
idx = idx // 2
if idx in names_to_varnames:
op = LOAD_FAST
arg = names_to_varnames[inst.arg]
elif inst.arg in name_translations:
arg = name_translations[inst.arg]
arg = names_to_varnames[idx]
# Cache is not present after LOAD_FAST and needs to be removed.
drop_future_cache = True
elif idx in name_translations:
tgt = name_translations[idx]
if is_new_bytecode:
tgt = 2*tgt + (inst.arg % 2)
arg = tgt
else:
raise ValueError("a name was lost in translation")
raise(ValueError("a name was lost in translation last instruction %s" % str(inst)))
# If it accesses co_varnames or co_names then update its argument.
elif inst.opcode in opcode.haslocal:
arg = varname_translations[inst.arg]
elif inst.opcode in opcode.hasname:
# for example STORE_GLOBAL
arg = name_translations[inst.arg]
elif is_new_bytecode and inst.opcode in opcode.hasfree:
# Python 3.11+ adds refs at the end (after locals), for whatever reason...
if inst.argval not in code_obj.co_varnames[:code_obj.co_argcount]: # we do not need to remap existing arguments, they are not shifted by new ones.
arg = inst.arg + len(new_locals)
modified.extend(write_instruction(op, arg))
if six.PY2:
code = ''.join(modified)
@ -113,23 +140,26 @@ def append_arguments(code_obj, new_locals):
code_obj.co_filename, code_obj.co_name,
code_obj.co_firstlineno, code_obj.co_lnotab,
code_obj.co_freevars, code_obj.co_cellvars)
# Done modifying codestring - make the code object
if hasattr(code_obj, "replace"):
# Python 3.8+
return code_obj.replace(
code_obj = code_obj.replace(
co_argcount=co_argcount + new_locals_len,
co_nlocals=code_obj.co_nlocals + new_locals_len,
co_code=code,
co_names=names,
co_varnames=varnames)
return code_obj
else:
return types.CodeType(*args)
def instructions(code_obj):
# easy for python 3.4+
if sys.version_info >= (3, 4):
def instructions(code_obj, show_cache=True):
if sys.version_info >= (3, 11):
# Python 3.11 introduced "cache instructions", hidden by default.
for inst in dis.Bytecode(code_obj, show_caches=show_cache):
yield inst
elif sys.version_info >= (3, 4): # easy for python 3.4+
for inst in dis.Bytecode(code_obj):
yield inst
else:
@ -171,7 +201,7 @@ def write_instruction(op, arg):
chr((arg >> 8) & 255)
]
else:
raise ValueError("Invalid oparg: {0} is too large".format(oparg))
raise ValueError("Invalid oparg: {0} is too large".format(arg))
else: # python 3.6+ uses wordcode instead of bytecode and they already supply all the EXTENDEND_ARG ops :)
if arg is None:
return [chr(op), 0]
@ -191,6 +221,7 @@ def write_instruction(op, arg):
# raise ValueError("Invalid oparg: {0} is too large".format(oparg))
def check(code_obj):
old_bytecode = code_obj.co_code
insts = list(instructions(code_obj))
@ -221,24 +252,99 @@ def check(code_obj):
'Your python version made changes to the bytecode')
def signature(func):
code_obj = six.get_function_code(func)
return (code_obj.co_nlocals, code_obj.co_argcount, code_obj.co_nlocals, code_obj.co_stacksize,
code_obj.co_flags, code_obj.co_names, code_obj.co_varnames,
code_obj.co_filename,
code_obj.co_freevars, code_obj.co_cellvars)
check(six.get_function_code(check))
def compare_func(fake_func, gt_func):
print(signature(fake_func))
print(signature(gt_func))
assert signature(fake_func) == signature(gt_func)
fake_ins = list(instructions(six.get_function_code(fake_func), show_cache=False))
real_ins = list(instructions(six.get_function_code(gt_func), show_cache=False))
offset = 0
pos = 0
for e in fake_ins:
if e.opname == "NOP":
offset += 1 # ignore NOPs that are inserted in place of old cache.
else:
real = real_ins[pos]
fake = e
print("POS %d OFFSET: %d FAKE VS REAL" % (pos, offset))
print(fake)
print(real)
assert fake.opcode == real.opcode
if fake.opcode in dis.hasjabs or fake.opcode in dis.hasjrel:
pass
else:
assert fake.arg == real.arg
assert fake.argval == real.argval or fake.opname in ["LOAD_CONST"]
assert fake.is_jump_target == real.is_jump_target
pos += 1
assert pos == len(real_ins), (pos, len(real_ins))
print("DONE, looks good.")
if __name__ == '__main__':
x = 'Wrong'
dick = 3000
import faulthandler
def func(a):
print(x, y, z, a)
print(dick)
d = (x, )
for e in (e for e in x):
print(e)
return x, y, z
faulthandler.enable()
func2 = types.FunctionType(
append_arguments(six.get_function_code(func), ('x', 'y', 'z')),
six.get_function_globals(func),
func.__name__,
closure=six.get_function_closure(func))
args = (2, 2, 3, 4), 3, 4
assert func2(1, *args) == args
def func(cmpfn):
if not this.Class in ('Array', 'Arguments'):
return this.to_object() # do nothing
arr = []
for i in xrange(len(this)):
arr.append(this.get(six.text_type(i)))
if not arr:
return this
if not cmpfn.is_callable():
cmpfn = None
cmp = lambda a, b: sort_compare(a, b, cmpfn)
if six.PY3:
key = functools.cmp_to_key(cmp)
arr.sort(key=key)
else:
arr.sort(cmp=cmp)
for i in xrange(len(arr)):
this.put(six.text_type(i), arr[i])
return this
def func_gt(cmpfn, this, arguments):
if not this.Class in ('Array', 'Arguments'):
return this.to_object() # do nothing
arr = []
for i in xrange(len(this)):
arr.append(this.get(six.text_type(i)))
if not arr:
return this
if not cmpfn.is_callable():
cmpfn = None
cmp = lambda a, b: sort_compare(a, b, cmpfn)
if six.PY3:
key = functools.cmp_to_key(cmp)
arr.sort(key=key)
else:
arr.sort(cmp=cmp)
for i in xrange(len(arr)):
this.put(six.text_type(i), arr[i])
return this
func2 = fix_js_args(func)
compare_func(func2, func_gt)