cpython/Lib/pickle.py

1381 lines
46 KiB
Python
Raw Normal View History

"""Create portable serialized representations of Python objects.
See module copyreg for a mechanism for registering custom picklers.
See module pickletools source for extensive comments.
Classes:
Pickler
Unpickler
Functions:
dump(object, file)
dumps(object) -> string
load(file) -> object
loads(string) -> object
Misc variables:
__version__
format_version
compatible_formats
"""
__version__ = "$Revision$" # Code version
from types import FunctionType, BuiltinFunctionType
from copyreg import dispatch_table
from copyreg import _extension_registry, _inverted_registry, _extension_cache
import marshal
import sys
import struct
2001-02-17 23:10:09 -04:00
import re
import io
import codecs
import _compat_pickle
2001-02-07 19:14:30 -04:00
__all__ = ["PickleError", "PicklingError", "UnpicklingError", "Pickler",
"Unpickler", "dump", "dumps", "load", "loads"]
# Shortcut for use in isinstance testing
bytes_types = (bytes, bytearray)
# These are purely informational; no code uses these.
format_version = "3.0" # File format version we write
compatible_formats = ["1.0", # Original protocol 0
"1.1", # Protocol 0 with INST added
"1.2", # Original protocol 1
"1.3", # Protocol 1 with BINFLOAT added
"2.0", # Protocol 2
"3.0", # Protocol 3
] # Old format versions we can read
# This is the highest protocol number we know how to read.
HIGHEST_PROTOCOL = 3
# The protocol we write by default. May be less than HIGHEST_PROTOCOL.
# We intentionally write a protocol that Python 2.x cannot read;
# there are too many issues with that.
DEFAULT_PROTOCOL = 3
# Why use struct.pack() for pickling but marshal.loads() for
# unpickling? struct.pack() is 40% faster than marshal.dumps(), but
# marshal.loads() is twice as fast as struct.unpack()!
mloads = marshal.loads
1995-03-09 10:08:35 -04:00
class PickleError(Exception):
"""A common base class for the other pickling exceptions."""
pass
class PicklingError(PickleError):
"""This exception is raised when an unpicklable object is passed to the
dump() method.
"""
pass
class UnpicklingError(PickleError):
"""This exception is raised when there is a problem unpickling an object,
such as a security violation.
Note that other exceptions may also be raised during unpickling, including
(but not necessarily limited to) AttributeError, EOFError, ImportError,
and IndexError.
"""
pass
# An instance of _Stop is raised by Unpickler.load_stop() in response to
# the STOP opcode, passing the object that is the result of unpickling.
class _Stop(Exception):
def __init__(self, value):
self.value = value
# Jython has PyStringMap; it's a dict subclass with string keys
try:
from org.python.core import PyStringMap
except ImportError:
PyStringMap = None
# Pickle opcodes. See pickletools.py for extensive docs. The listing
# here is in kind-of alphabetical order of 1-character pickle code.
# pickletools groups them by purpose.
MARK = b'(' # push special markobject on stack
STOP = b'.' # every pickle ends with STOP
POP = b'0' # discard topmost stack item
POP_MARK = b'1' # discard stack top through topmost markobject
DUP = b'2' # duplicate top stack item
FLOAT = b'F' # push float object; decimal string argument
INT = b'I' # push integer or bool; decimal string argument
BININT = b'J' # push four-byte signed int
BININT1 = b'K' # push 1-byte unsigned int
LONG = b'L' # push long; decimal string argument
BININT2 = b'M' # push 2-byte unsigned int
NONE = b'N' # push None
PERSID = b'P' # push persistent object; id is taken from string arg
BINPERSID = b'Q' # " " " ; " " " " stack
REDUCE = b'R' # apply callable to argtuple, both on stack
STRING = b'S' # push string; NL-terminated string argument
BINSTRING = b'T' # push string; counted binary string argument
SHORT_BINSTRING= b'U' # " " ; " " " " < 256 bytes
UNICODE = b'V' # push Unicode string; raw-unicode-escaped'd argument
BINUNICODE = b'X' # " " " ; counted UTF-8 string argument
APPEND = b'a' # append stack top to list below it
BUILD = b'b' # call __setstate__ or __dict__.update()
GLOBAL = b'c' # push self.find_class(modname, name); 2 string args
DICT = b'd' # build a dict from stack items
EMPTY_DICT = b'}' # push empty dict
APPENDS = b'e' # extend list on stack by topmost stack slice
GET = b'g' # push item from memo on stack; index is string arg
BINGET = b'h' # " " " " " " ; " " 1-byte arg
INST = b'i' # build & push class instance
LONG_BINGET = b'j' # push item from memo on stack; index is 4-byte arg
LIST = b'l' # build list from topmost stack items
EMPTY_LIST = b']' # push empty list
OBJ = b'o' # build & push class instance
PUT = b'p' # store stack top in memo; index is string arg
BINPUT = b'q' # " " " " " ; " " 1-byte arg
LONG_BINPUT = b'r' # " " " " " ; " " 4-byte arg
SETITEM = b's' # add key+value pair to dict
TUPLE = b't' # build tuple from topmost stack items
EMPTY_TUPLE = b')' # push empty tuple
SETITEMS = b'u' # modify dict by adding topmost key+value pairs
BINFLOAT = b'G' # push float; arg is 8-byte float encoding
TRUE = b'I01\n' # not an opcode; see INT docs in pickletools.py
FALSE = b'I00\n' # not an opcode; see INT docs in pickletools.py
2003-01-29 02:16:12 -04:00
# Protocol 2
PROTO = b'\x80' # identify pickle protocol
NEWOBJ = b'\x81' # build object by applying cls.__new__ to argtuple
EXT1 = b'\x82' # push object from extension registry; 1-byte index
EXT2 = b'\x83' # ditto, but 2-byte index
EXT4 = b'\x84' # ditto, but 4-byte index
TUPLE1 = b'\x85' # build 1-tuple from stack top
TUPLE2 = b'\x86' # build 2-tuple from two topmost stack items
TUPLE3 = b'\x87' # build 3-tuple from three topmost stack items
NEWTRUE = b'\x88' # push True
NEWFALSE = b'\x89' # push False
LONG1 = b'\x8a' # push long from < 256 bytes
LONG4 = b'\x8b' # push really big long
2003-01-27 17:44:25 -04:00
_tuplesize2code = [EMPTY_TUPLE, TUPLE1, TUPLE2, TUPLE3]
# Protocol 3 (Python 3.x)
BINBYTES = b'B' # push bytes; counted binary string argument
SHORT_BINBYTES = b'C' # " " ; " " " " < 256 bytes
2001-02-17 23:10:09 -04:00
__all__.extend([x for x in dir() if re.match("[A-Z][A-Z0-9_]+$",x)])
# Pickling machinery
class _Pickler:
def __init__(self, file, protocol=None, *, fix_imports=True):
"""This takes a binary file for writing a pickle data stream.
The optional protocol argument tells the pickler to use the
given protocol; supported protocols are 0, 1, 2, 3. The default
protocol is 3; a backward-incompatible protocol designed for
Python 3.0.
Specifying a negative protocol version selects the highest
protocol version supported. The higher the protocol used, the
more recent the version of Python needed to read the pickle
produced.
The file argument must have a write() method that accepts a single
bytes argument. It can thus be a file object opened for binary
writing, a io.BytesIO instance, or any other custom object that
meets this interface.
If fix_imports is True and protocol is less than 3, pickle will try to
map the new Python 3.x names to the old module names used in Python
2.x, so that the pickle data stream is readable with Python 2.x.
"""
if protocol is None:
protocol = DEFAULT_PROTOCOL
if protocol < 0:
protocol = HIGHEST_PROTOCOL
elif not 0 <= protocol <= HIGHEST_PROTOCOL:
raise ValueError("pickle protocol must be <= %d" % HIGHEST_PROTOCOL)
try:
self.write = file.write
except AttributeError:
raise TypeError("file must have a 'write' attribute")
self.memo = {}
self.proto = int(protocol)
self.bin = protocol >= 1
self.fast = 0
self.fix_imports = fix_imports and protocol < 3
def clear_memo(self):
"""Clears the pickler's "memo".
The memo is the data structure that remembers which objects the
pickler has already seen, so that shared or recursive objects are
pickled by reference and not by value. This method is useful when
re-using picklers.
"""
self.memo.clear()
def dump(self, obj):
"""Write a pickled representation of obj to the open file."""
# Check whether Pickler was initialized correctly. This is
# only needed to mimic the behavior of _pickle.Pickler.dump().
if not hasattr(self, "write"):
raise PicklingError("Pickler.__init__() was not called by "
"%s.__init__()" % (self.__class__.__name__,))
if self.proto >= 2:
self.write(PROTO + bytes([self.proto]))
self.save(obj)
self.write(STOP)
def memoize(self, obj):
"""Store an object in the memo."""
# The Pickler memo is a dictionary mapping object ids to 2-tuples
# that contain the Unpickler memo key and the object being memoized.
# The memo key is written to the pickle and will become
# the key in the Unpickler's memo. The object is stored in the
# Pickler memo so that transient objects are kept alive during
# pickling.
# The use of the Unpickler memo length as the memo key is just a
# convention. The only requirement is that the memo values be unique.
# But there appears no advantage to any other scheme, and this
2003-01-27 20:24:43 -04:00
# scheme allows the Unpickler memo to be implemented as a plain (but
# growable) array, indexed by memo key.
if self.fast:
return
assert id(obj) not in self.memo
memo_len = len(self.memo)
self.write(self.put(memo_len))
self.memo[id(obj)] = memo_len, obj
2003-01-27 17:25:41 -04:00
# Return a PUT (BINPUT, LONG_BINPUT) opcode string, with argument i.
def put(self, i, pack=struct.pack):
if self.bin:
if i < 256:
return BINPUT + bytes([i])
else:
return LONG_BINPUT + pack("<i", i)
return PUT + repr(i).encode("ascii") + b'\n'
2003-01-27 17:25:41 -04:00
# Return a GET (BINGET, LONG_BINGET) opcode string, with argument i.
def get(self, i, pack=struct.pack):
if self.bin:
if i < 256:
return BINGET + bytes([i])
else:
return LONG_BINGET + pack("<i", i)
return GET + repr(i).encode("ascii") + b'\n'
2001-01-14 20:50:52 -04:00
def save(self, obj, save_persistent_id=True):
# Check for persistent id (defined by a subclass)
pid = self.persistent_id(obj)
if pid is not None and save_persistent_id:
self.save_pers(pid)
return
# Check the memo
x = self.memo.get(id(obj))
if x:
self.write(self.get(x[0]))
return
# Check the type dispatch table
t = type(obj)
f = self.dispatch.get(t)
if f:
f(self, obj) # Call unbound method with explicit self
return
# Check for a class with a custom metaclass; treat as regular class
try:
issc = issubclass(t, type)
except TypeError: # t is not a class (old Boost; see SF #502085)
issc = 0
if issc:
self.save_global(obj)
return
# Check copyreg.dispatch_table
reduce = dispatch_table.get(t)
if reduce:
rv = reduce(obj)
else:
# Check for a __reduce_ex__ method, fall back to __reduce__
reduce = getattr(obj, "__reduce_ex__", None)
if reduce:
rv = reduce(self.proto)
else:
reduce = getattr(obj, "__reduce__", None)
if reduce:
rv = reduce()
else:
raise PicklingError("Can't pickle %r object: %r" %
(t.__name__, obj))
# Check for string returned by reduce(), meaning "save as global"
if isinstance(rv, str):
self.save_global(obj, rv)
return
# Assert that reduce() returned a tuple
if not isinstance(rv, tuple):
raise PicklingError("%s must return string or tuple" % reduce)
2001-01-14 20:50:52 -04:00
# Assert that it returned an appropriately sized tuple
l = len(rv)
if not (2 <= l <= 5):
raise PicklingError("Tuple returned by %s must have "
"two to five elements" % reduce)
# Save the reduce() output and finally memoize the object
self.save_reduce(obj=obj, *rv)
def persistent_id(self, obj):
# This exists so a subclass can override it
return None
def save_pers(self, pid):
# Save a persistent id reference
if self.bin:
self.save(pid, save_persistent_id=False)
self.write(BINPERSID)
else:
self.write(PERSID + str(pid).encode("ascii") + b'\n')
def save_reduce(self, func, args, state=None,
listitems=None, dictitems=None, obj=None):
2003-06-29 13:59:59 -03:00
# This API is called by some subclasses
# Assert that args is a tuple
if not isinstance(args, tuple):
raise PicklingError("args from save_reduce() should be a tuple")
# Assert that func is callable
Merged revisions 55407-55513 via svnmerge from svn+ssh://pythondev@svn.python.org/python/branches/p3yk ................ r55413 | fred.drake | 2007-05-17 12:30:10 -0700 (Thu, 17 May 2007) | 1 line fix argument name in documentation; match the implementation ................ r55430 | jack.diederich | 2007-05-18 06:39:59 -0700 (Fri, 18 May 2007) | 1 line Implements class decorators, PEP 3129. ................ r55432 | guido.van.rossum | 2007-05-18 08:09:41 -0700 (Fri, 18 May 2007) | 2 lines obsubmit. ................ r55434 | guido.van.rossum | 2007-05-18 09:39:10 -0700 (Fri, 18 May 2007) | 3 lines Fix bug in test_inspect. (I presume this is how it should be fixed; Jack Diedrich, please verify.) ................ r55460 | brett.cannon | 2007-05-20 00:31:57 -0700 (Sun, 20 May 2007) | 4 lines Remove the imageop module. With imgfile already removed in Python 3.0 and rgbimg gone in Python 2.6 the unit tests themselves were made worthless. Plus third-party libraries perform the same function much better. ................ r55469 | neal.norwitz | 2007-05-20 11:28:20 -0700 (Sun, 20 May 2007) | 118 lines Merged revisions 55324-55467 via svnmerge from svn+ssh://pythondev@svn.python.org/python/trunk ........ r55348 | georg.brandl | 2007-05-15 13:19:34 -0700 (Tue, 15 May 2007) | 4 lines HTML-escape the plain traceback in cgitb's HTML output, to prevent the traceback inadvertently or maliciously closing the comment and injecting HTML into the error page. ........ r55372 | neal.norwitz | 2007-05-15 21:33:50 -0700 (Tue, 15 May 2007) | 6 lines Port rev 55353 from Guido: Add what looks like a necessary call to PyErr_NoMemory() when PyMem_MALLOC() fails. Will backport. ........ r55377 | neal.norwitz | 2007-05-15 22:06:33 -0700 (Tue, 15 May 2007) | 1 line Mention removal of some directories for obsolete platforms ........ r55380 | brett.cannon | 2007-05-15 22:50:03 -0700 (Tue, 15 May 2007) | 2 lines Change the maintainer of the BeOS port. ........ r55383 | georg.brandl | 2007-05-16 06:44:18 -0700 (Wed, 16 May 2007) | 2 lines Bug #1719995: don't use deprecated method in sets example. ........ r55386 | neal.norwitz | 2007-05-16 13:05:11 -0700 (Wed, 16 May 2007) | 5 lines Fix bug in marshal where bad data would cause a segfault due to lack of an infinite recursion check. Contributed by Damien Miller at Google. ........ r55389 | brett.cannon | 2007-05-16 15:42:29 -0700 (Wed, 16 May 2007) | 6 lines Remove the gopherlib module. It has been raising a DeprecationWarning since Python 2.5. Also remove gopher support from urllib/urllib2. As both imported gopherlib the usage of the support would have raised a DeprecationWarning. ........ r55394 | raymond.hettinger | 2007-05-16 18:08:04 -0700 (Wed, 16 May 2007) | 1 line calendar.py gets no benefit from xrange() instead of range() ........ r55395 | brett.cannon | 2007-05-16 19:02:56 -0700 (Wed, 16 May 2007) | 3 lines Complete deprecation of BaseException.message. Some subclasses were directly accessing the message attribute instead of using the descriptor. ........ r55396 | neal.norwitz | 2007-05-16 23:11:36 -0700 (Wed, 16 May 2007) | 4 lines Reduce the max stack depth to see if this fixes the segfaults on Windows and some other boxes. If this is successful, this rev should be backported. I'm not sure how close to the limit we should push this. ........ r55397 | neal.norwitz | 2007-05-16 23:23:50 -0700 (Wed, 16 May 2007) | 4 lines Set the depth to something very small to try to determine if the crashes on Windows are really due to the stack size or possibly some other problem. ........ r55398 | neal.norwitz | 2007-05-17 00:04:46 -0700 (Thu, 17 May 2007) | 4 lines Last try for tweaking the max stack depth. 5000 was the original value, 4000 didn't work either. 1000 does work on Windows. If 2000 works, that will hopefully be a reasonable balance. ........ r55412 | fred.drake | 2007-05-17 12:29:58 -0700 (Thu, 17 May 2007) | 1 line fix argument name in documentation; match the implementation ........ r55427 | neal.norwitz | 2007-05-17 22:47:16 -0700 (Thu, 17 May 2007) | 1 line Verify neither dumps or loads overflow the stack and segfault. ........ r55446 | collin.winter | 2007-05-18 16:11:24 -0700 (Fri, 18 May 2007) | 1 line Backport PEP 3110's new 'except' syntax to 2.6. ........ r55448 | raymond.hettinger | 2007-05-18 18:11:16 -0700 (Fri, 18 May 2007) | 1 line Improvements to NamedTuple's implementation, tests, and documentation ........ r55449 | raymond.hettinger | 2007-05-18 18:50:11 -0700 (Fri, 18 May 2007) | 1 line Fix beginner mistake -- don't mix spaces and tabs. ........ r55450 | neal.norwitz | 2007-05-18 20:48:47 -0700 (Fri, 18 May 2007) | 1 line Clear data so random memory does not get freed. Will backport. ........ r55452 | neal.norwitz | 2007-05-18 21:34:55 -0700 (Fri, 18 May 2007) | 3 lines Whoops, need to pay attention to those test failures. Move the clear to *before* the first use, not after. ........ r55453 | neal.norwitz | 2007-05-18 21:35:52 -0700 (Fri, 18 May 2007) | 1 line Give some clue as to what happened if the test fails. ........ r55455 | georg.brandl | 2007-05-19 11:09:26 -0700 (Sat, 19 May 2007) | 2 lines Fix docstring for add_package in site.py. ........ r55458 | brett.cannon | 2007-05-20 00:09:50 -0700 (Sun, 20 May 2007) | 2 lines Remove the rgbimg module. It has been deprecated since Python 2.5. ........ r55465 | nick.coghlan | 2007-05-20 04:12:49 -0700 (Sun, 20 May 2007) | 1 line Fix typo in example (should be backported, but my maintenance branch is woefully out of date) ........ ................ r55472 | brett.cannon | 2007-05-20 12:06:18 -0700 (Sun, 20 May 2007) | 2 lines Remove imageop from the Windows build process. ................ r55486 | neal.norwitz | 2007-05-20 23:59:52 -0700 (Sun, 20 May 2007) | 1 line Remove callable() builtin ................ r55506 | neal.norwitz | 2007-05-22 00:43:29 -0700 (Tue, 22 May 2007) | 78 lines Merged revisions 55468-55505 via svnmerge from svn+ssh://pythondev@svn.python.org/python/trunk ........ r55468 | neal.norwitz | 2007-05-20 11:06:27 -0700 (Sun, 20 May 2007) | 1 line rotor is long gone. ........ r55470 | neal.norwitz | 2007-05-20 11:43:00 -0700 (Sun, 20 May 2007) | 1 line Update directories/files at the top-level. ........ r55471 | brett.cannon | 2007-05-20 12:05:06 -0700 (Sun, 20 May 2007) | 2 lines Try to remove rgbimg from Windows builds. ........ r55474 | brett.cannon | 2007-05-20 16:17:38 -0700 (Sun, 20 May 2007) | 4 lines Remove the macfs module. This led to the deprecation of macostools.touched(); it completely relied on macfs and is a no-op on OS X according to code comments. ........ r55476 | brett.cannon | 2007-05-20 16:56:18 -0700 (Sun, 20 May 2007) | 3 lines Move imgfile import to the global namespace to trigger an import error ASAP to prevent creation of a test file. ........ r55477 | brett.cannon | 2007-05-20 16:57:38 -0700 (Sun, 20 May 2007) | 3 lines Cause posixfile to raise a DeprecationWarning. Documented as deprecated since Ptyhon 1.5. ........ r55479 | andrew.kuchling | 2007-05-20 17:03:15 -0700 (Sun, 20 May 2007) | 1 line Note removed modules ........ r55481 | martin.v.loewis | 2007-05-20 21:35:47 -0700 (Sun, 20 May 2007) | 2 lines Add Alexandre Vassalotti. ........ r55482 | george.yoshida | 2007-05-20 21:41:21 -0700 (Sun, 20 May 2007) | 4 lines fix against r55474 [Remove the macfs module] Remove "libmacfs.tex" from Makefile.deps and mac/mac.tex. ........ r55487 | raymond.hettinger | 2007-05-21 01:13:35 -0700 (Mon, 21 May 2007) | 1 line Replace assertion with straight error-checking. ........ r55489 | raymond.hettinger | 2007-05-21 09:40:10 -0700 (Mon, 21 May 2007) | 1 line Allow all alphanumeric and underscores in type and field names. ........ r55490 | facundo.batista | 2007-05-21 10:32:32 -0700 (Mon, 21 May 2007) | 5 lines Added timeout support to HTTPSConnection, through the socket.create_connection function. Also added a small test for this, and updated NEWS file. ........ r55495 | georg.brandl | 2007-05-21 13:34:16 -0700 (Mon, 21 May 2007) | 2 lines Patch #1686487: you can now pass any mapping after '**' in function calls. ........ r55502 | neal.norwitz | 2007-05-21 23:03:36 -0700 (Mon, 21 May 2007) | 1 line Document new params to HTTPSConnection ........ r55504 | neal.norwitz | 2007-05-22 00:16:10 -0700 (Tue, 22 May 2007) | 1 line Stop using METH_OLDARGS ........ r55505 | neal.norwitz | 2007-05-22 00:16:44 -0700 (Tue, 22 May 2007) | 1 line Stop using METH_OLDARGS implicitly ........ ................
2007-05-22 15:11:13 -03:00
if not hasattr(func, '__call__'):
raise PicklingError("func from save_reduce() should be callable")
save = self.save
write = self.write
# Protocol 2 special case: if func's name is __newobj__, use NEWOBJ
if self.proto >= 2 and getattr(func, "__name__", "") == "__newobj__":
# A __reduce__ implementation can direct protocol 2 to
# use the more efficient NEWOBJ opcode, while still
# allowing protocol 0 and 1 to work normally. For this to
# work, the function returned by __reduce__ should be
# called __newobj__, and its first argument should be a
# new-style class. The implementation for __newobj__
# should be as follows, although pickle has no way to
# verify this:
#
# def __newobj__(cls, *args):
# return cls.__new__(cls, *args)
#
# Protocols 0 and 1 will pickle a reference to __newobj__,
# while protocol 2 (and above) will pickle a reference to
# cls, the remaining args tuple, and the NEWOBJ code,
# which calls cls.__new__(cls, *args) at unpickling time
# (see load_newobj below). If __reduce__ returns a
# three-tuple, the state from the third tuple item will be
# pickled regardless of the protocol, calling __setstate__
# at unpickling time (see load_build below).
#
# Note that no standard __newobj__ implementation exists;
# you have to provide your own. This is to enforce
# compatibility with Python 2.2 (pickles written using
# protocol 0 or 1 in Python 2.3 should be unpicklable by
# Python 2.2).
cls = args[0]
if not hasattr(cls, "__new__"):
raise PicklingError(
"args[0] from __newobj__ args has no __new__")
if obj is not None and cls is not obj.__class__:
raise PicklingError(
"args[0] from __newobj__ args has the wrong class")
args = args[1:]
save(cls)
save(args)
write(NEWOBJ)
else:
save(func)
save(args)
write(REDUCE)
2001-01-14 20:50:52 -04:00
if obj is not None:
self.memoize(obj)
# More new special cases (that work with older protocols as
# well): when __reduce__ returns a tuple with 4 or 5 items,
# the 4th and 5th item should be iterators that provide list
# items and dict items (as (key, value) tuples), or None.
if listitems is not None:
self._batch_appends(listitems)
if dictitems is not None:
self._batch_setitems(dictitems)
if state is not None:
save(state)
write(BUILD)
# Methods below this point are dispatched through the dispatch table
dispatch = {}
def save_none(self, obj):
self.write(NONE)
dispatch[type(None)] = save_none
def save_bool(self, obj):
if self.proto >= 2:
self.write(obj and NEWTRUE or NEWFALSE)
else:
self.write(obj and TRUE or FALSE)
dispatch[bool] = save_bool
def save_long(self, obj, pack=struct.pack):
if self.bin:
# If the int is small enough to fit in a signed 4-byte 2's-comp
# format, we can store it more efficiently than the general
# case.
# First one- and two-byte unsigned ints:
if obj >= 0:
if obj <= 0xff:
self.write(BININT1 + bytes([obj]))
return
if obj <= 0xffff:
self.write(BININT2 + bytes([obj&0xff, obj>>8]))
return
# Next check for 4-byte signed ints:
high_bits = obj >> 31 # note that Python shift sign-extends
if high_bits == 0 or high_bits == -1:
# All high bits are copies of bit 2**31, so the value
# fits in a 4-byte signed int.
self.write(BININT + pack("<i", obj))
return
if self.proto >= 2:
encoded = encode_long(obj)
n = len(encoded)
if n < 256:
self.write(LONG1 + bytes([n]) + encoded)
else:
self.write(LONG4 + pack("<i", n) + encoded)
cPickle.c: Full support for the new LONG1 and LONG4. Added comments. Assorted code cleanups; e.g., sizeof(char) is 1 by definition, so there's no need to do things like multiply by sizeof(char) in hairy malloc arguments. Fixed an undetected-overflow bug in readline_file(). longobject.c: Fixed a really stupid bug in the new _PyLong_NumBits. pickle.py: Fixed stupid bug in save_long(): When proto is 2, it wrote LONG1 or LONG4, but forgot to return then -- it went on to append the proto 1 LONG opcode too. Fixed equally stupid cancelling bugs in load_long1() and load_long4(): they *returned* the unpickled long instead of pushing it on the stack. The return values were ignored. Tests passed before only because save_long() pickled the long twice. Fixed bugs in encode_long(). Noted that decode_long() is quadratic-time despite our hopes, because long(string, 16) is still quadratic-time in len(string). It's hex() that's linear-time. I don't know a way to make decode_long() linear-time in Python, short of maybe transforming the 256's-complement bytes into marshal's funky internal format, and letting marshal decode that. It would be more valuable to make long(string, 16) linear time. pickletester.py: Added a global "protocols" vector so tests can try all the protocols in a sane way. Changed test_ints() and test_unicode() to do so. Added a new test_long(), but the tail end of it is disabled because it "takes forever" under pickle.py (but runs very quickly under cPickle: cPickle proto 2 for longs is linear-time).
2003-02-01 22:57:53 -04:00
return
self.write(LONG + repr(obj).encode("ascii") + b'L\n')
dispatch[int] = save_long
def save_float(self, obj, pack=struct.pack):
if self.bin:
self.write(BINFLOAT + pack('>d', obj))
else:
self.write(FLOAT + repr(obj).encode("ascii") + b'\n')
dispatch[float] = save_float
def save_bytes(self, obj, pack=struct.pack):
if self.proto < 3:
self.save_reduce(bytes, (list(obj),), obj=obj)
return
n = len(obj)
if n < 256:
self.write(SHORT_BINBYTES + bytes([n]) + bytes(obj))
else:
self.write(BINBYTES + pack("<i", n) + bytes(obj))
self.memoize(obj)
dispatch[bytes] = save_bytes
def save_str(self, obj, pack=struct.pack):
if self.bin:
encoded = obj.encode('utf-8', 'surrogatepass')
n = len(encoded)
self.write(BINUNICODE + pack("<i", n) + encoded)
else:
obj = obj.replace("\\", "\\u005c")
obj = obj.replace("\n", "\\u000a")
2007-05-04 17:30:19 -03:00
self.write(UNICODE + bytes(obj.encode('raw-unicode-escape')) +
b'\n')
self.memoize(obj)
dispatch[str] = save_str
2001-02-09 16:06:00 -04:00
def save_tuple(self, obj):
write = self.write
proto = self.proto
n = len(obj)
if n == 0:
if proto:
write(EMPTY_TUPLE)
else:
write(MARK + TUPLE)
return
save = self.save
memo = self.memo
if n <= 3 and proto >= 2:
for element in obj:
save(element)
# Subtle. Same as in the big comment below.
if id(obj) in memo:
get = self.get(memo[id(obj)][0])
write(POP * n + get)
else:
write(_tuplesize2code[n])
self.memoize(obj)
return
# proto 0 or proto 1 and tuple isn't empty, or proto > 1 and tuple
# has more than 3 elements.
write(MARK)
for element in obj:
save(element)
if id(obj) in memo:
# Subtle. d was not in memo when we entered save_tuple(), so
# the process of saving the tuple's elements must have saved
# the tuple itself: the tuple is recursive. The proper action
# now is to throw away everything we put on the stack, and
# simply GET the tuple (it's already constructed). This check
# could have been done in the "for element" loop instead, but
# recursive tuples are a rare thing.
get = self.get(memo[id(obj)][0])
if proto:
write(POP_MARK + get)
else: # proto 0 -- POP_MARK not available
write(POP * (n+1) + get)
return
# No recursion.
self.write(TUPLE)
self.memoize(obj)
dispatch[tuple] = save_tuple
def save_list(self, obj):
write = self.write
if self.bin:
write(EMPTY_LIST)
else: # proto 0 -- can't use EMPTY_LIST
write(MARK + LIST)
self.memoize(obj)
self._batch_appends(obj)
dispatch[list] = save_list
_BATCHSIZE = 1000
def _batch_appends(self, items):
# Helper to batch up APPENDS sequences
save = self.save
write = self.write
if not self.bin:
for x in items:
save(x)
write(APPEND)
return
items = iter(items)
Merged revisions 55007-55179 via svnmerge from svn+ssh://pythondev@svn.python.org/python/branches/p3yk ........ r55077 | guido.van.rossum | 2007-05-02 11:54:37 -0700 (Wed, 02 May 2007) | 2 lines Use the new print syntax, at least. ........ r55142 | fred.drake | 2007-05-04 21:27:30 -0700 (Fri, 04 May 2007) | 1 line remove old cruftiness ........ r55143 | fred.drake | 2007-05-04 21:52:16 -0700 (Fri, 04 May 2007) | 1 line make this work with the new Python ........ r55162 | neal.norwitz | 2007-05-06 22:29:18 -0700 (Sun, 06 May 2007) | 1 line Get asdl code gen working with Python 2.3. Should continue to work with 3.0 ........ r55164 | neal.norwitz | 2007-05-07 00:00:38 -0700 (Mon, 07 May 2007) | 1 line Verify checkins to p3yk (sic) branch go to 3000 list. ........ r55166 | neal.norwitz | 2007-05-07 00:12:35 -0700 (Mon, 07 May 2007) | 1 line Fix this test so it runs again by importing warnings_test properly. ........ r55167 | neal.norwitz | 2007-05-07 01:03:22 -0700 (Mon, 07 May 2007) | 8 lines So long xrange. range() now supports values that are outside -sys.maxint to sys.maxint. floats raise a TypeError. This has been sitting for a long time. It probably has some problems and needs cleanup. Objects/rangeobject.c now uses 4-space indents since it is almost completely new. ........ r55171 | guido.van.rossum | 2007-05-07 10:21:26 -0700 (Mon, 07 May 2007) | 4 lines Fix two tests that were previously depending on significant spaces at the end of a line (and before that on Python 2.x print behavior that has no exact equivalent in 3.0). ........
2007-05-07 19:24:25 -03:00
r = range(self._BATCHSIZE)
while items is not None:
tmp = []
for i in r:
try:
x = next(items)
tmp.append(x)
except StopIteration:
items = None
break
n = len(tmp)
if n > 1:
write(MARK)
for x in tmp:
save(x)
write(APPENDS)
elif n:
save(tmp[0])
write(APPEND)
# else tmp is empty, and we're done
def save_dict(self, obj):
write = self.write
if self.bin:
write(EMPTY_DICT)
else: # proto 0 -- can't use EMPTY_DICT
write(MARK + DICT)
self.memoize(obj)
self._batch_setitems(obj.items())
dispatch[dict] = save_dict
if PyStringMap is not None:
dispatch[PyStringMap] = save_dict
def _batch_setitems(self, items):
# Helper to batch up SETITEMS sequences; proto >= 1 only
save = self.save
write = self.write
if not self.bin:
for k, v in items:
save(k)
save(v)
write(SETITEM)
return
items = iter(items)
Merged revisions 55007-55179 via svnmerge from svn+ssh://pythondev@svn.python.org/python/branches/p3yk ........ r55077 | guido.van.rossum | 2007-05-02 11:54:37 -0700 (Wed, 02 May 2007) | 2 lines Use the new print syntax, at least. ........ r55142 | fred.drake | 2007-05-04 21:27:30 -0700 (Fri, 04 May 2007) | 1 line remove old cruftiness ........ r55143 | fred.drake | 2007-05-04 21:52:16 -0700 (Fri, 04 May 2007) | 1 line make this work with the new Python ........ r55162 | neal.norwitz | 2007-05-06 22:29:18 -0700 (Sun, 06 May 2007) | 1 line Get asdl code gen working with Python 2.3. Should continue to work with 3.0 ........ r55164 | neal.norwitz | 2007-05-07 00:00:38 -0700 (Mon, 07 May 2007) | 1 line Verify checkins to p3yk (sic) branch go to 3000 list. ........ r55166 | neal.norwitz | 2007-05-07 00:12:35 -0700 (Mon, 07 May 2007) | 1 line Fix this test so it runs again by importing warnings_test properly. ........ r55167 | neal.norwitz | 2007-05-07 01:03:22 -0700 (Mon, 07 May 2007) | 8 lines So long xrange. range() now supports values that are outside -sys.maxint to sys.maxint. floats raise a TypeError. This has been sitting for a long time. It probably has some problems and needs cleanup. Objects/rangeobject.c now uses 4-space indents since it is almost completely new. ........ r55171 | guido.van.rossum | 2007-05-07 10:21:26 -0700 (Mon, 07 May 2007) | 4 lines Fix two tests that were previously depending on significant spaces at the end of a line (and before that on Python 2.x print behavior that has no exact equivalent in 3.0). ........
2007-05-07 19:24:25 -03:00
r = range(self._BATCHSIZE)
while items is not None:
tmp = []
for i in r:
try:
tmp.append(next(items))
except StopIteration:
items = None
break
n = len(tmp)
if n > 1:
write(MARK)
for k, v in tmp:
save(k)
save(v)
write(SETITEMS)
elif n:
k, v = tmp[0]
save(k)
save(v)
write(SETITEM)
# else tmp is empty, and we're done
def save_global(self, obj, name=None, pack=struct.pack):
write = self.write
memo = self.memo
if name is None:
name = obj.__name__
module = getattr(obj, "__module__", None)
if module is None:
module = whichmodule(obj, name)
try:
__import__(module, level=0)
mod = sys.modules[module]
klass = getattr(mod, name)
except (ImportError, KeyError, AttributeError):
raise PicklingError(
"Can't pickle %r: it's not found as %s.%s" %
(obj, module, name))
else:
if klass is not obj:
raise PicklingError(
"Can't pickle %r: it's not the same object as %s.%s" %
(obj, module, name))
if self.proto >= 2:
code = _extension_registry.get((module, name))
if code:
assert code > 0
if code <= 0xff:
write(EXT1 + bytes([code]))
elif code <= 0xffff:
write(EXT2 + bytes([code&0xff, code>>8]))
else:
write(EXT4 + pack("<i", code))
return
# Non-ASCII identifiers are supported only with protocols >= 3.
if self.proto >= 3:
write(GLOBAL + bytes(module, "utf-8") + b'\n' +
bytes(name, "utf-8") + b'\n')
else:
if self.fix_imports:
if (module, name) in _compat_pickle.REVERSE_NAME_MAPPING:
module, name = _compat_pickle.REVERSE_NAME_MAPPING[(module, name)]
if module in _compat_pickle.REVERSE_IMPORT_MAPPING:
module = _compat_pickle.REVERSE_IMPORT_MAPPING[module]
try:
write(GLOBAL + bytes(module, "ascii") + b'\n' +
bytes(name, "ascii") + b'\n')
except UnicodeEncodeError:
raise PicklingError(
"can't pickle global identifier '%s.%s' using "
"pickle protocol %i" % (module, name, self.proto))
self.memoize(obj)
dispatch[FunctionType] = save_global
dispatch[BuiltinFunctionType] = save_global
dispatch[type] = save_global
# Pickling helpers
def _keep_alive(x, memo):
"""Keeps a reference to the object x in the memo.
Because we remember objects by their id, we have
to assure that possibly temporary objects are kept
alive by referencing them.
We store a reference at the id of the memo, which should
normally not be used unless someone tries to deepcopy
the memo itself...
"""
try:
memo[id(memo)].append(x)
except KeyError:
# aha, this is the first one :-)
memo[id(memo)]=[x]
# A cache for whichmodule(), mapping a function object to the name of
# the module in which the function was found.
classmap = {} # called classmap for backwards compatibility
def whichmodule(func, funcname):
"""Figure out the module in which a function occurs.
Search sys.modules for the module.
Cache in classmap.
Return a module name.
If the function cannot be found, return "__main__".
"""
# Python functions should always get an __module__ from their globals.
mod = getattr(func, "__module__", None)
if mod is not None:
return mod
if func in classmap:
return classmap[func]
for name, module in list(sys.modules.items()):
if module is None:
continue # skip dummy package entries
2003-02-06 12:23:01 -04:00
if name != '__main__' and getattr(module, funcname, None) is func:
break
else:
name = '__main__'
classmap[func] = name
return name
# Unpickling machinery
class _Unpickler:
def __init__(self, file, *, fix_imports=True,
encoding="ASCII", errors="strict"):
"""This takes a binary file for reading a pickle data stream.
The protocol version of the pickle is detected automatically, so no
proto argument is needed.
The file-like object must have two methods, a read() method
that takes an integer argument, and a readline() method that
requires no arguments. Both methods should return bytes.
Thus file-like object can be a binary file object opened for
reading, a BytesIO object, or any other custom object that
meets this interface.
Optional keyword arguments are *fix_imports*, *encoding* and *errors*,
which are used to control compatiblity support for pickle stream
generated by Python 2.x. If *fix_imports* is True, pickle will try to
map the old Python 2.x names to the new names used in Python 3.x. The
*encoding* and *errors* tell pickle how to decode 8-bit string
instances pickled by Python 2.x; these default to 'ASCII' and
'strict', respectively.
"""
self.readline = file.readline
self.read = file.read
self.memo = {}
self.encoding = encoding
self.errors = errors
self.proto = 0
self.fix_imports = fix_imports
def load(self):
"""Read a pickled object representation from the open file.
Return the reconstituted object hierarchy specified in the file.
"""
# Check whether Unpickler was initialized correctly. This is
# only needed to mimic the behavior of _pickle.Unpickler.dump().
if not hasattr(self, "read"):
raise UnpicklingError("Unpickler.__init__() was not called by "
"%s.__init__()" % (self.__class__.__name__,))
2001-11-09 12:15:04 -04:00
self.mark = object() # any new unique object
self.stack = []
self.append = self.stack.append
read = self.read
dispatch = self.dispatch
try:
while 1:
key = read(1)
if not key:
raise EOFError
assert isinstance(key, bytes_types)
dispatch[key[0]](self)
except _Stop as stopinst:
return stopinst.value
2003-01-27 21:41:51 -04:00
# Return largest index k such that self.stack[k] is self.mark.
# If the stack doesn't contain a mark, eventually raises IndexError.
# This could be sped by maintaining another stack, of indices at which
# the mark appears. For that matter, the latter stack would suffice,
# and we wouldn't need to push mark objects on self.stack at all.
# Doing so is probably a good thing, though, since if the pickle is
# corrupt (or hostile) we may get a clue from finding self.mark embedded
# in unpickled objects.
def marker(self):
stack = self.stack
mark = self.mark
k = len(stack)-1
while stack[k] is not mark: k = k-1
return k
def persistent_load(self, pid):
2009-01-10 13:05:44 -04:00
raise UnpicklingError("unsupported persistent id encountered")
dispatch = {}
def load_proto(self):
proto = ord(self.read(1))
if not 0 <= proto <= HIGHEST_PROTOCOL:
2007-08-27 20:18:54 -03:00
raise ValueError("unsupported pickle protocol: %d" % proto)
self.proto = proto
dispatch[PROTO[0]] = load_proto
def load_persid(self):
pid = self.readline()[:-1].decode("ascii")
self.append(self.persistent_load(pid))
dispatch[PERSID[0]] = load_persid
def load_binpersid(self):
pid = self.stack.pop()
self.append(self.persistent_load(pid))
dispatch[BINPERSID[0]] = load_binpersid
def load_none(self):
self.append(None)
dispatch[NONE[0]] = load_none
def load_false(self):
self.append(False)
dispatch[NEWFALSE[0]] = load_false
def load_true(self):
self.append(True)
dispatch[NEWTRUE[0]] = load_true
def load_int(self):
data = self.readline()
if data == FALSE[1:]:
val = False
elif data == TRUE[1:]:
val = True
else:
try:
val = int(data, 0)
except ValueError:
val = int(data, 0)
self.append(val)
dispatch[INT[0]] = load_int
def load_binint(self):
self.append(mloads(b'i' + self.read(4)))
dispatch[BININT[0]] = load_binint
def load_binint1(self):
self.append(ord(self.read(1)))
dispatch[BININT1[0]] = load_binint1
def load_binint2(self):
self.append(mloads(b'i' + self.read(2) + b'\000\000'))
dispatch[BININT2[0]] = load_binint2
2001-01-14 20:50:52 -04:00
def load_long(self):
val = self.readline()[:-1].decode("ascii")
if val and val[-1] == 'L':
val = val[:-1]
self.append(int(val, 0))
dispatch[LONG[0]] = load_long
def load_long1(self):
n = ord(self.read(1))
data = self.read(n)
self.append(decode_long(data))
dispatch[LONG1[0]] = load_long1
def load_long4(self):
n = mloads(b'i' + self.read(4))
data = self.read(n)
self.append(decode_long(data))
dispatch[LONG4[0]] = load_long4
def load_float(self):
self.append(float(self.readline()[:-1]))
dispatch[FLOAT[0]] = load_float
def load_binfloat(self, unpack=struct.unpack):
self.append(unpack('>d', self.read(8))[0])
dispatch[BINFLOAT[0]] = load_binfloat
def load_string(self):
orig = self.readline()
rep = orig[:-1]
2007-08-27 20:18:54 -03:00
for q in (b'"', b"'"): # double or single quote
if rep.startswith(q):
if not rep.endswith(q):
2007-08-27 20:18:54 -03:00
raise ValueError("insecure string pickle")
rep = rep[len(q):-len(q)]
break
else:
raise ValueError("insecure string pickle: %r" % orig)
self.append(codecs.escape_decode(rep)[0]
.decode(self.encoding, self.errors))
dispatch[STRING[0]] = load_string
def load_binstring(self):
len = mloads(b'i' + self.read(4))
data = self.read(len)
value = str(data, self.encoding, self.errors)
self.append(value)
dispatch[BINSTRING[0]] = load_binstring
def load_binbytes(self):
len = mloads(b'i' + self.read(4))
self.append(self.read(len))
dispatch[BINBYTES[0]] = load_binbytes
def load_unicode(self):
self.append(str(self.readline()[:-1], 'raw-unicode-escape'))
dispatch[UNICODE[0]] = load_unicode
def load_binunicode(self):
len = mloads(b'i' + self.read(4))
self.append(str(self.read(len), 'utf-8', 'surrogatepass'))
dispatch[BINUNICODE[0]] = load_binunicode
def load_short_binstring(self):
len = ord(self.read(1))
data = bytes(self.read(len))
value = str(data, self.encoding, self.errors)
self.append(value)
dispatch[SHORT_BINSTRING[0]] = load_short_binstring
def load_short_binbytes(self):
len = ord(self.read(1))
self.append(bytes(self.read(len)))
dispatch[SHORT_BINBYTES[0]] = load_short_binbytes
def load_tuple(self):
k = self.marker()
self.stack[k:] = [tuple(self.stack[k+1:])]
dispatch[TUPLE[0]] = load_tuple
def load_empty_tuple(self):
self.append(())
dispatch[EMPTY_TUPLE[0]] = load_empty_tuple
def load_tuple1(self):
self.stack[-1] = (self.stack[-1],)
dispatch[TUPLE1[0]] = load_tuple1
def load_tuple2(self):
self.stack[-2:] = [(self.stack[-2], self.stack[-1])]
dispatch[TUPLE2[0]] = load_tuple2
def load_tuple3(self):
self.stack[-3:] = [(self.stack[-3], self.stack[-2], self.stack[-1])]
dispatch[TUPLE3[0]] = load_tuple3
def load_empty_list(self):
self.append([])
dispatch[EMPTY_LIST[0]] = load_empty_list
def load_empty_dictionary(self):
self.append({})
dispatch[EMPTY_DICT[0]] = load_empty_dictionary
def load_list(self):
k = self.marker()
self.stack[k:] = [self.stack[k+1:]]
dispatch[LIST[0]] = load_list
def load_dict(self):
k = self.marker()
d = {}
items = self.stack[k+1:]
for i in range(0, len(items), 2):
key = items[i]
value = items[i+1]
d[key] = value
self.stack[k:] = [d]
dispatch[DICT[0]] = load_dict
# INST and OBJ differ only in how they get a class object. It's not
# only sensible to do the rest in a common routine, the two routines
# previously diverged and grew different bugs.
# klass is the class to instantiate, and k points to the topmost mark
# object, following which are the arguments for klass.__init__.
def _instantiate(self, klass, k):
args = tuple(self.stack[k+1:])
del self.stack[k:]
instantiated = False
if (not args and
isinstance(klass, type) and
not hasattr(klass, "__getinitargs__")):
Merged revisions 55270-55324 via svnmerge from svn+ssh://pythondev@svn.python.org/python/branches/p3yk ........ r55271 | fred.drake | 2007-05-11 10:14:47 -0700 (Fri, 11 May 2007) | 3 lines remove jpeg, panel libraries for SGI; there is more IRIX stuff left over, I guess that should be removed too, but will leave for someone who is sure ........ r55280 | fred.drake | 2007-05-11 19:11:37 -0700 (Fri, 11 May 2007) | 1 line remove mention of file that has been removed ........ r55301 | brett.cannon | 2007-05-13 17:38:05 -0700 (Sun, 13 May 2007) | 4 lines Remove rexec and Bastion from the stdlib. This also eliminates the need for f_restricted on frames. This in turn negates the need for PyEval_GetRestricted() and PyFrame_IsRestricted(). ........ r55303 | brett.cannon | 2007-05-13 19:22:22 -0700 (Sun, 13 May 2007) | 2 lines Remove the md5 and sha modules. ........ r55305 | george.yoshida | 2007-05-13 19:45:55 -0700 (Sun, 13 May 2007) | 2 lines fix markup ........ r55306 | neal.norwitz | 2007-05-13 19:47:57 -0700 (Sun, 13 May 2007) | 1 line Get the doc building again after some removals. ........ r55307 | neal.norwitz | 2007-05-13 19:50:45 -0700 (Sun, 13 May 2007) | 1 line Get test_pyclbr passing again after getstatus was removed from commands. This "test case" was weird since it was just importing a seemingly random module. Remove the import ........ r55322 | brett.cannon | 2007-05-14 14:09:20 -0700 (Mon, 14 May 2007) | 3 lines Remove the compiler package. Will eventually need a mechanism to byte compile an AST. ........
2007-05-14 19:03:55 -03:00
value = _EmptyClass()
value.__class__ = klass
instantiated = True
if not instantiated:
try:
value = klass(*args)
except TypeError as err:
2007-08-27 20:18:54 -03:00
raise TypeError("in constructor for %s: %s" %
(klass.__name__, str(err)), sys.exc_info()[2])
self.append(value)
def load_inst(self):
module = self.readline()[:-1].decode("ascii")
name = self.readline()[:-1].decode("ascii")
klass = self.find_class(module, name)
self._instantiate(klass, self.marker())
dispatch[INST[0]] = load_inst
def load_obj(self):
# Stack is ... markobject classobject arg1 arg2 ...
k = self.marker()
klass = self.stack.pop(k+1)
self._instantiate(klass, k)
dispatch[OBJ[0]] = load_obj
def load_newobj(self):
args = self.stack.pop()
cls = self.stack[-1]
obj = cls.__new__(cls, *args)
self.stack[-1] = obj
dispatch[NEWOBJ[0]] = load_newobj
def load_global(self):
module = self.readline()[:-1].decode("utf-8")
name = self.readline()[:-1].decode("utf-8")
klass = self.find_class(module, name)
self.append(klass)
dispatch[GLOBAL[0]] = load_global
def load_ext1(self):
code = ord(self.read(1))
self.get_extension(code)
dispatch[EXT1[0]] = load_ext1
def load_ext2(self):
code = mloads(b'i' + self.read(2) + b'\000\000')
self.get_extension(code)
dispatch[EXT2[0]] = load_ext2
def load_ext4(self):
code = mloads(b'i' + self.read(4))
self.get_extension(code)
dispatch[EXT4[0]] = load_ext4
def get_extension(self, code):
nil = []
obj = _extension_cache.get(code, nil)
if obj is not nil:
self.append(obj)
return
key = _inverted_registry.get(code)
if not key:
raise ValueError("unregistered extension code %d" % code)
obj = self.find_class(*key)
_extension_cache[code] = obj
self.append(obj)
def find_class(self, module, name):
# Subclasses may override this.
if self.proto < 3 and self.fix_imports:
if (module, name) in _compat_pickle.NAME_MAPPING:
module, name = _compat_pickle.NAME_MAPPING[(module, name)]
if module in _compat_pickle.IMPORT_MAPPING:
module = _compat_pickle.IMPORT_MAPPING[module]
__import__(module, level=0)
mod = sys.modules[module]
klass = getattr(mod, name)
return klass
def load_reduce(self):
stack = self.stack
args = stack.pop()
func = stack[-1]
try:
value = func(*args)
except:
print(sys.exc_info())
print(func, args)
raise
stack[-1] = value
dispatch[REDUCE[0]] = load_reduce
def load_pop(self):
del self.stack[-1]
dispatch[POP[0]] = load_pop
def load_pop_mark(self):
k = self.marker()
del self.stack[k:]
dispatch[POP_MARK[0]] = load_pop_mark
def load_dup(self):
self.append(self.stack[-1])
dispatch[DUP[0]] = load_dup
def load_get(self):
i = int(self.readline()[:-1])
self.append(self.memo[i])
dispatch[GET[0]] = load_get
def load_binget(self):
i = self.read(1)[0]
self.append(self.memo[i])
dispatch[BINGET[0]] = load_binget
def load_long_binget(self):
i = mloads(b'i' + self.read(4))
self.append(self.memo[i])
dispatch[LONG_BINGET[0]] = load_long_binget
def load_put(self):
i = int(self.readline()[:-1])
self.memo[i] = self.stack[-1]
dispatch[PUT[0]] = load_put
def load_binput(self):
i = self.read(1)[0]
self.memo[i] = self.stack[-1]
dispatch[BINPUT[0]] = load_binput
def load_long_binput(self):
i = mloads(b'i' + self.read(4))
self.memo[i] = self.stack[-1]
dispatch[LONG_BINPUT[0]] = load_long_binput
def load_append(self):
stack = self.stack
value = stack.pop()
list = stack[-1]
list.append(value)
dispatch[APPEND[0]] = load_append
def load_appends(self):
stack = self.stack
mark = self.marker()
list = stack[mark - 1]
list.extend(stack[mark + 1:])
del stack[mark:]
dispatch[APPENDS[0]] = load_appends
2001-01-14 20:50:52 -04:00
def load_setitem(self):
stack = self.stack
value = stack.pop()
key = stack.pop()
dict = stack[-1]
dict[key] = value
dispatch[SETITEM[0]] = load_setitem
def load_setitems(self):
stack = self.stack
mark = self.marker()
dict = stack[mark - 1]
for i in range(mark + 1, len(stack), 2):
dict[stack[i]] = stack[i + 1]
del stack[mark:]
dispatch[SETITEMS[0]] = load_setitems
def load_build(self):
stack = self.stack
state = stack.pop()
inst = stack[-1]
setstate = getattr(inst, "__setstate__", None)
if setstate:
setstate(state)
return
slotstate = None
if isinstance(state, tuple) and len(state) == 2:
state, slotstate = state
if state:
inst_dict = inst.__dict__
intern = sys.intern
for k, v in state.items():
if type(k) is str:
inst_dict[intern(k)] = v
else:
inst_dict[k] = v
if slotstate:
for k, v in slotstate.items():
setattr(inst, k, v)
dispatch[BUILD[0]] = load_build
def load_mark(self):
self.append(self.mark)
dispatch[MARK[0]] = load_mark
def load_stop(self):
value = self.stack.pop()
raise _Stop(value)
dispatch[STOP[0]] = load_stop
# Helper class for load_inst/load_obj
class _EmptyClass:
pass
# Encode/decode longs in linear time.
import binascii as _binascii
def encode_long(x):
r"""Encode a long to a two's complement little-endian binary string.
Note that 0 is a special case, returning an empty string, to save a
byte in the LONG1 pickling context.
>>> encode_long(0)
b''
>>> encode_long(255)
b'\xff\x00'
>>> encode_long(32767)
b'\xff\x7f'
>>> encode_long(-256)
b'\x00\xff'
>>> encode_long(-32768)
b'\x00\x80'
>>> encode_long(-128)
b'\x80'
>>> encode_long(127)
b'\x7f'
>>>
"""
if x == 0:
return b''
if x > 0:
ashex = hex(x)
assert ashex.startswith("0x")
njunkchars = 2 + ashex.endswith('L')
nibbles = len(ashex) - njunkchars
if nibbles & 1:
# need an even # of nibbles for unhexlify
ashex = "0x0" + ashex[2:]
elif int(ashex[2], 16) >= 8:
# "looks negative", so need a byte of sign bits
ashex = "0x00" + ashex[2:]
else:
# Build the 256's-complement: (1L << nbytes) + x. The trick is
# to find the number of bytes in linear time (although that should
# really be a constant-time task).
ashex = hex(-x)
assert ashex.startswith("0x")
njunkchars = 2 + ashex.endswith('L')
nibbles = len(ashex) - njunkchars
if nibbles & 1:
cPickle.c: Full support for the new LONG1 and LONG4. Added comments. Assorted code cleanups; e.g., sizeof(char) is 1 by definition, so there's no need to do things like multiply by sizeof(char) in hairy malloc arguments. Fixed an undetected-overflow bug in readline_file(). longobject.c: Fixed a really stupid bug in the new _PyLong_NumBits. pickle.py: Fixed stupid bug in save_long(): When proto is 2, it wrote LONG1 or LONG4, but forgot to return then -- it went on to append the proto 1 LONG opcode too. Fixed equally stupid cancelling bugs in load_long1() and load_long4(): they *returned* the unpickled long instead of pushing it on the stack. The return values were ignored. Tests passed before only because save_long() pickled the long twice. Fixed bugs in encode_long(). Noted that decode_long() is quadratic-time despite our hopes, because long(string, 16) is still quadratic-time in len(string). It's hex() that's linear-time. I don't know a way to make decode_long() linear-time in Python, short of maybe transforming the 256's-complement bytes into marshal's funky internal format, and letting marshal decode that. It would be more valuable to make long(string, 16) linear time. pickletester.py: Added a global "protocols" vector so tests can try all the protocols in a sane way. Changed test_ints() and test_unicode() to do so. Added a new test_long(), but the tail end of it is disabled because it "takes forever" under pickle.py (but runs very quickly under cPickle: cPickle proto 2 for longs is linear-time).
2003-02-01 22:57:53 -04:00
# Extend to a full byte.
nibbles += 1
nbits = nibbles * 4
x += 1 << nbits
assert x > 0
ashex = hex(x)
cPickle.c: Full support for the new LONG1 and LONG4. Added comments. Assorted code cleanups; e.g., sizeof(char) is 1 by definition, so there's no need to do things like multiply by sizeof(char) in hairy malloc arguments. Fixed an undetected-overflow bug in readline_file(). longobject.c: Fixed a really stupid bug in the new _PyLong_NumBits. pickle.py: Fixed stupid bug in save_long(): When proto is 2, it wrote LONG1 or LONG4, but forgot to return then -- it went on to append the proto 1 LONG opcode too. Fixed equally stupid cancelling bugs in load_long1() and load_long4(): they *returned* the unpickled long instead of pushing it on the stack. The return values were ignored. Tests passed before only because save_long() pickled the long twice. Fixed bugs in encode_long(). Noted that decode_long() is quadratic-time despite our hopes, because long(string, 16) is still quadratic-time in len(string). It's hex() that's linear-time. I don't know a way to make decode_long() linear-time in Python, short of maybe transforming the 256's-complement bytes into marshal's funky internal format, and letting marshal decode that. It would be more valuable to make long(string, 16) linear time. pickletester.py: Added a global "protocols" vector so tests can try all the protocols in a sane way. Changed test_ints() and test_unicode() to do so. Added a new test_long(), but the tail end of it is disabled because it "takes forever" under pickle.py (but runs very quickly under cPickle: cPickle proto 2 for longs is linear-time).
2003-02-01 22:57:53 -04:00
njunkchars = 2 + ashex.endswith('L')
newnibbles = len(ashex) - njunkchars
if newnibbles < nibbles:
ashex = "0x" + "0" * (nibbles - newnibbles) + ashex[2:]
if int(ashex[2], 16) < 8:
# "looks positive", so need a byte of sign bits
cPickle.c: Full support for the new LONG1 and LONG4. Added comments. Assorted code cleanups; e.g., sizeof(char) is 1 by definition, so there's no need to do things like multiply by sizeof(char) in hairy malloc arguments. Fixed an undetected-overflow bug in readline_file(). longobject.c: Fixed a really stupid bug in the new _PyLong_NumBits. pickle.py: Fixed stupid bug in save_long(): When proto is 2, it wrote LONG1 or LONG4, but forgot to return then -- it went on to append the proto 1 LONG opcode too. Fixed equally stupid cancelling bugs in load_long1() and load_long4(): they *returned* the unpickled long instead of pushing it on the stack. The return values were ignored. Tests passed before only because save_long() pickled the long twice. Fixed bugs in encode_long(). Noted that decode_long() is quadratic-time despite our hopes, because long(string, 16) is still quadratic-time in len(string). It's hex() that's linear-time. I don't know a way to make decode_long() linear-time in Python, short of maybe transforming the 256's-complement bytes into marshal's funky internal format, and letting marshal decode that. It would be more valuable to make long(string, 16) linear time. pickletester.py: Added a global "protocols" vector so tests can try all the protocols in a sane way. Changed test_ints() and test_unicode() to do so. Added a new test_long(), but the tail end of it is disabled because it "takes forever" under pickle.py (but runs very quickly under cPickle: cPickle proto 2 for longs is linear-time).
2003-02-01 22:57:53 -04:00
ashex = "0xff" + ashex[2:]
if ashex.endswith('L'):
ashex = ashex[2:-1]
else:
ashex = ashex[2:]
cPickle.c: Full support for the new LONG1 and LONG4. Added comments. Assorted code cleanups; e.g., sizeof(char) is 1 by definition, so there's no need to do things like multiply by sizeof(char) in hairy malloc arguments. Fixed an undetected-overflow bug in readline_file(). longobject.c: Fixed a really stupid bug in the new _PyLong_NumBits. pickle.py: Fixed stupid bug in save_long(): When proto is 2, it wrote LONG1 or LONG4, but forgot to return then -- it went on to append the proto 1 LONG opcode too. Fixed equally stupid cancelling bugs in load_long1() and load_long4(): they *returned* the unpickled long instead of pushing it on the stack. The return values were ignored. Tests passed before only because save_long() pickled the long twice. Fixed bugs in encode_long(). Noted that decode_long() is quadratic-time despite our hopes, because long(string, 16) is still quadratic-time in len(string). It's hex() that's linear-time. I don't know a way to make decode_long() linear-time in Python, short of maybe transforming the 256's-complement bytes into marshal's funky internal format, and letting marshal decode that. It would be more valuable to make long(string, 16) linear time. pickletester.py: Added a global "protocols" vector so tests can try all the protocols in a sane way. Changed test_ints() and test_unicode() to do so. Added a new test_long(), but the tail end of it is disabled because it "takes forever" under pickle.py (but runs very quickly under cPickle: cPickle proto 2 for longs is linear-time).
2003-02-01 22:57:53 -04:00
assert len(ashex) & 1 == 0, (x, ashex)
binary = _binascii.unhexlify(ashex)
return bytes(binary[::-1])
def decode_long(data):
r"""Decode a long from a two's complement little-endian binary string.
>>> decode_long(b'')
0
>>> decode_long(b"\xff\x00")
255
>>> decode_long(b"\xff\x7f")
32767
>>> decode_long(b"\x00\xff")
-256
>>> decode_long(b"\x00\x80")
-32768
>>> decode_long(b"\x80")
-128
>>> decode_long(b"\x7f")
127
"""
nbytes = len(data)
if nbytes == 0:
return 0
ashex = _binascii.hexlify(data[::-1])
n = int(ashex, 16) # quadratic time before Python 2.3; linear now
if data[-1] >= 0x80:
n -= 1 << (nbytes * 8)
return n
# Use the faster _pickle if possible
try:
from _pickle import *
except ImportError:
Pickler, Unpickler = _Pickler, _Unpickler
# Shorthands
def dump(obj, file, protocol=None, *, fix_imports=True):
Pickler(file, protocol, fix_imports=fix_imports).dump(obj)
def dumps(obj, protocol=None, *, fix_imports=True):
f = io.BytesIO()
Pickler(f, protocol, fix_imports=fix_imports).dump(obj)
res = f.getvalue()
assert isinstance(res, bytes_types)
return res
def load(file, *, fix_imports=True, encoding="ASCII", errors="strict"):
return Unpickler(file, fix_imports=fix_imports,
encoding=encoding, errors=errors).load()
def loads(s, *, fix_imports=True, encoding="ASCII", errors="strict"):
if isinstance(s, str):
raise TypeError("Can't load pickle from unicode string")
file = io.BytesIO(s)
return Unpickler(file, fix_imports=fix_imports,
encoding=encoding, errors=errors).load()
# Doctest
def _test():
import doctest
return doctest.testmod()
if __name__ == "__main__":
_test()