What’s New In Python 3.8 (2024)

Assignment expressions¶

There is new syntax := that assigns values to variables as part of a largerexpression. It is affectionately known as “the walrus operator” due toits resemblance to the eyes and tusks of a walrus.

In this example, the assignment expression helps avoid callinglen() twice:

if (n := len(a)) > 10: print(f"List is too long ({n} elements, expected <= 10)")

A similar benefit arises during regular expression matching wherematch objects are needed twice, once to test whether a matchoccurred and another to extract a subgroup:

discount = 0.0if (mo := re.search(r'(\d+)% discount', advertisem*nt)): discount = float(mo.group(1)) / 100.0

The operator is also useful with while-loops that computea value to test loop termination and then need that samevalue again in the body of the loop:

# Loop over fixed length blockswhile (block := f.read(256)) != '': process(block)

Another motivating use case arises in list comprehensions wherea value computed in a filtering condition is also needed inthe expression body:

[clean_name.title() for name in names if (clean_name := normalize('NFC', name)) in allowed_names]

Try to limit use of the walrus operator to clean cases that reducecomplexity and improve readability.

See PEP 572 for a full description.

(Contributed by Emily Morehouse in bpo-35224.)

Positional-only parameters¶

There is a new function parameter syntax / to indicate that somefunction parameters must be specified positionally and cannot be used askeyword arguments. This is the same notation shown by help() for Cfunctions annotated with Larry Hastings’Argument Clinic tool.

In the following example, parameters a and b are positional-only,while c or d can be positional or keyword, and e or f arerequired to be keywords:

def f(a, b, /, c, d, *, e, f): print(a, b, c, d, e, f)

The following is a valid call:

f(10, 20, 30, d=40, e=50, f=60)

However, these are invalid calls:

f(10, b=20, c=30, d=40, e=50, f=60) # b cannot be a keyword argumentf(10, 20, 30, 40, 50, f=60) # e must be a keyword argument

One use case for this notation is that it allows pure Python functionsto fully emulate behaviors of existing C coded functions. For example,the built-in divmod() function does not accept keyword arguments:

def divmod(a, b, /): "Emulate the built in divmod() function" return (a // b, a % b)

Another use case is to preclude keyword arguments when the parametername is not helpful. For example, the builtin len() function hasthe signature len(obj, /). This precludes awkward calls such as:

len(obj='hello') # The "obj" keyword argument impairs readability

A further benefit of marking a parameter as positional-only is that itallows the parameter name to be changed in the future without risk ofbreaking client code. For example, in the statistics module, theparameter name dist may be changed in the future. This was madepossible with the following function specification:

def quantiles(dist, /, *, n=4, method='exclusive') ...

Since the parameters to the left of / are not exposed as possiblekeywords, the parameters names remain available for use in **kwargs:

>>> def f(a, b, /, **kwargs):...  print(a, b, kwargs)...>>> f(10, 20, a=1, b=2, c=3) # a and b are used in two ways10 20 {'a': 1, 'b': 2, 'c': 3}

This greatly simplifies the implementation of functions and methodsthat need to accept arbitrary keyword arguments. For example, hereis an excerpt from code in the collections module:

class Counter(dict): def __init__(self, iterable=None, /, **kwds): # Note "iterable" is a possible keyword argument

See PEP 570 for a full description.

(Contributed by Pablo Galindo in bpo-36540.)

Parallel filesystem cache for compiled bytecode files¶

The new PYTHONPYCACHEPREFIX setting (also available as-X pycache_prefix) configures the implicit bytecodecache to use a separate parallel filesystem tree, rather thanthe default __pycache__ subdirectories within each sourcedirectory.

The location of the cache is reported in sys.pycache_prefix(None indicates the default location in __pycache__subdirectories).

(Contributed by Carl Meyer in bpo-33499.)

Debug build uses the same ABI as release build¶

Python now uses the same ABI whether it’s built in release or debug mode. OnUnix, when Python is built in debug mode, it is now possible to load Cextensions built in release mode and C extensions built using the stable ABI.

Release builds and debug builds are now ABI compatible: defining thePy_DEBUG macro no longer implies the Py_TRACE_REFS macro, whichintroduces the only ABI incompatibility. The Py_TRACE_REFS macro, whichadds the sys.getobjects() function and the PYTHONDUMPREFSenvironment variable, can be set using the new ./configure--with-trace-refs build option.(Contributed by Victor Stinner in bpo-36465.)

On Unix, C extensions are no longer linked to libpython except on Androidand Cygwin.It is now possiblefor a statically linked Python to load a C extension built using a sharedlibrary Python.(Contributed by Victor Stinner in bpo-21536.)

On Unix, when Python is built in debug mode, import now also looks for Cextensions compiled in release mode and for C extensions compiled with thestable ABI.(Contributed by Victor Stinner in bpo-36722.)

To embed Python into an application, a new --embed option must be passed topython3-config --libs --embed to get -lpython3.8 (link the applicationto libpython). To support both 3.8 and older, try python3-config --libs--embed first and fallback to python3-config --libs (without --embed)if the previous command fails.

Add a pkg-config python-3.8-embed module to embed Python into anapplication: pkg-config python-3.8-embed --libs includes -lpython3.8.To support both 3.8 and older, try pkg-config python-X.Y-embed --libs firstand fallback to pkg-config python-X.Y --libs (without --embed) if theprevious command fails (replace X.Y with the Python version).

On the other hand, pkg-config python3.8 --libs no longer contains-lpython3.8. C extensions must not be linked to libpython (except onAndroid and Cygwin, whose cases are handled by the script);this change is backward incompatible on purpose.(Contributed by Victor Stinner in bpo-36721.)

f-strings support = for self-documenting expressions and debugging¶

Added an = specifier to f-strings. An f-string such asf'{expr=}' will expand to the text of the expression, an equal sign,then the representation of the evaluated expression. For example:

>>> user = 'eric_idle'>>> member_since = date(1975, 7, 31)>>> f'{user=} {member_since=}'"user='eric_idle' member_since=datetime.date(1975, 7, 31)"

The usual f-string format specifiers allow morecontrol over how the result of the expression is displayed:

>>> delta = date.today() - member_since>>> f'{user=!s} {delta.days=:,d}''user=eric_idle delta.days=16,075'

The = specifier will display the whole expression so thatcalculations can be shown:

>>> print(f'{theta=} {cos(radians(theta))=:.3f}')theta=30 cos(radians(theta))=0.866

(Contributed by Eric V. Smith and Larry Hastings in bpo-36817.)

PEP 578: Python Runtime Audit Hooks¶

The PEP adds an Audit Hook and Verified Open Hook. Both are available fromPython and native code, allowing applications and frameworks written in purePython code to take advantage of extra notifications, while also allowingembedders or system administrators to deploy builds of Python where auditing isalways enabled.

See PEP 578 for full details.

PEP 587: Python Initialization Configuration¶

The PEP 587 adds a new C API to configure the Python Initializationproviding finer control on the whole configuration and better error reporting.

New structures:

• PyConfig

• PyPreConfig

• PyStatus

• PyWideStringList

New functions:

• PyConfig_Clear()

• PyConfig_InitIsolatedConfig()

• PyConfig_InitPythonConfig()

• PyConfig_Read()

• PyConfig_SetArgv()

• PyConfig_SetBytesArgv()

• PyConfig_SetBytesString()

• PyConfig_SetString()

• PyPreConfig_InitIsolatedConfig()

• PyPreConfig_InitPythonConfig()

• PyStatus_Error()

• PyStatus_Exception()

• PyStatus_Exit()

• PyStatus_IsError()

• PyStatus_IsExit()

• PyStatus_NoMemory()

• PyStatus_Ok()

• PyWideStringList_Append()

• PyWideStringList_Insert()

• Py_BytesMain()

• Py_ExitStatusException()

• Py_InitializeFromConfig()

• Py_PreInitialize()

• Py_PreInitializeFromArgs()

• Py_PreInitializeFromBytesArgs()

• Py_RunMain()

This PEP also adds _PyRuntimeState.preconfig (PyPreConfig type)and PyInterpreterState.config (PyConfig type) fields to theseinternal structures. PyInterpreterState.config becomes the newreference configuration, replacing global configuration variables andother private variables.

See Python Initialization Configuration for thedocumentation.

See PEP 587 for a full description.

(Contributed by Victor Stinner in bpo-36763.)

PEP 590: Vectorcall: a fast calling protocol for CPython¶

The Vectorcall Protocol is added to the Python/C API.It is meant to formalize existing optimizations which were already donefor various classes.Any static type implementing a callable can use thisprotocol.

This is currently provisional.The aim is to make it fully public in Python 3.9.

See PEP 590 for a full description.

(Contributed by Jeroen Demeyer, Mark Shannon and Petr Viktorin in bpo-36974.)

Pickle protocol 5 with out-of-band data buffers¶

When pickle is used to transfer large data between Python processesin order to take advantage of multi-core or multi-machine processing,it is important to optimize the transfer by reducing memory copies, andpossibly by applying custom techniques such as data-dependent compression.

The pickle protocol 5 introduces support for out-of-band bufferswhere PEP 3118-compatible data can be transmitted separately from themain pickle stream, at the discretion of the communication layer.

See PEP 574 for a full description.

(Contributed by Antoine Pitrou in bpo-36785.)

ast¶

AST nodes now have end_lineno and end_col_offset attributes,which give the precise location of the end of the node. (This onlyapplies to nodes that have lineno and col_offset attributes.)

New function ast.get_source_segment() returns the source codefor a specific AST node.

(Contributed by Ivan Levkivskyi in bpo-33416.)

The ast.parse() function has some new flags:

• type_comments=True causes it to return the text of PEP 484 andPEP 526 type comments associated with certain AST nodes;

• mode='func_type' can be used to parse PEP 484 “signature typecomments” (returned for function definition AST nodes);

• feature_version=(3, N) allows specifying an earlier Python 3version. For example, feature_version=(3, 4) will treatasync and await as non-reserved words.

(Contributed by Guido van Rossum in bpo-35766.)

asyncio¶

asyncio.run() has graduated from the provisional to stable API. Thisfunction can be used to execute a coroutine and return the result whileautomatically managing the event loop. For example:

import asyncioasync def main(): await asyncio.sleep(0) return 42asyncio.run(main())

This is roughly equivalent to:

import asyncioasync def main(): await asyncio.sleep(0) return 42loop = asyncio.new_event_loop()asyncio.set_event_loop(loop)try: loop.run_until_complete(main())finally: asyncio.set_event_loop(None) loop.close()

The actual implementation is significantly more complex. Thus,asyncio.run() should be the preferred way of running asyncio programs.

(Contributed by Yury Selivanov in bpo-32314.)

Running python -m asyncio launches a natively async REPL. This allows rapidexperimentation with code that has a top-level await. There is nolonger a need to directly call asyncio.run() which would spawn a new eventloop on every invocation:

$ python -m asyncioasyncio REPL 3.8.0Use "await" directly instead of "asyncio.run()".Type "help", "copyright", "credits" or "license" for more information.>>> import asyncio>>> await asyncio.sleep(10, result='hello')hello

(Contributed by Yury Selivanov in bpo-37028.)

The exception asyncio.CancelledError now inherits fromBaseException rather than Exception and no longer inheritsfrom concurrent.futures.CancelledError.(Contributed by Yury Selivanov in bpo-32528.)

On Windows, the default event loop is now ProactorEventLoop.(Contributed by Victor Stinner in bpo-34687.)

ProactorEventLoop now also supports UDP.(Contributed by Adam Meily and Andrew Svetlov in bpo-29883.)

ProactorEventLoop can now be interrupted byKeyboardInterrupt (“CTRL+C”).(Contributed by Vladimir Matveev in bpo-23057.)

Added asyncio.Task.get_coro() for getting the wrapped coroutinewithin an asyncio.Task.(Contributed by Alex Grönholm in bpo-36999.)

Asyncio tasks can now be named, either by passing the name keywordargument to asyncio.create_task() orthe create_task() event loop method, or bycalling the set_name() method on the task object. Thetask name is visible in the repr() output of asyncio.Task andcan also be retrieved using the get_name() method.(Contributed by Alex Grönholm in bpo-34270.)

Added support forHappy Eyeballs toasyncio.loop.create_connection(). To specify the behavior, two newparameters have been added: happy_eyeballs_delay and interleave. The HappyEyeballs algorithm improves responsiveness in applications that support IPv4and IPv6 by attempting to simultaneously connect using both.(Contributed by twisteroid ambassador in bpo-33530.)

builtins¶

The compile() built-in has been improved to accept theast.PyCF_ALLOW_TOP_LEVEL_AWAIT flag. With this new flag passed,compile() will allow top-level await, async for and async withconstructs that are usually considered invalid syntax. Asynchronous code objectmarked with the CO_COROUTINE flag may then be returned.(Contributed by Matthias Bussonnier in bpo-34616)

collections¶

The _asdict() method forcollections.namedtuple() now returns a dict instead of acollections.OrderedDict. This works because regular dicts haveguaranteed ordering since Python 3.7. If the extra features ofOrderedDict are required, the suggested remediation is to cast theresult to the desired type: OrderedDict(nt._asdict()).(Contributed by Raymond Hettinger in bpo-35864.)

cProfile¶

The cProfile.Profile class can now be used as a context manager.Profile a block of code by running:

import cProfilewith cProfile.Profile() as profiler: # code to be profiled ...

(Contributed by Scott Sanderson in bpo-29235.)

csv¶

The csv.DictReader now returns instances of dict instead ofa collections.OrderedDict. The tool is now faster and uses lessmemory while still preserving the field order.(Contributed by Michael Selik in bpo-34003.)

curses¶

Added a new variable holding structured version information for theunderlying ncurses library: ncurses_version.(Contributed by Serhiy Storchaka in bpo-31680.)

ctypes¶

On Windows, CDLL and subclasses now accept a winmode parameterto specify flags for the underlying LoadLibraryEx call. The default flags areset to only load DLL dependencies from trusted locations, including the pathwhere the DLL is stored (if a full or partial path is used to load the initialDLL) and paths added by add_dll_directory().(Contributed by Steve Dower in bpo-36085.)

datetime¶

Added new alternate constructors datetime.date.fromisocalendar() anddatetime.datetime.fromisocalendar(), which construct date anddatetime objects respectively from ISO year, week number, and weekday;these are the inverse of each class’s isocalendar method.(Contributed by Paul Ganssle in bpo-36004.)

functools¶

functools.lru_cache() can now be used as a straight decorator ratherthan as a function returning a decorator. So both of these are now supported:

@lru_cachedef f(x): ...@lru_cache(maxsize=256)def f(x): ...

(Contributed by Raymond Hettinger in bpo-36772.)

Added a new functools.cached_property() decorator, for computed propertiescached for the life of the instance.

import functoolsimport statisticsclass Dataset: def __init__(self, sequence_of_numbers): self.data = sequence_of_numbers @functools.cached_property def variance(self): return statistics.variance(self.data)

(Contributed by Carl Meyer in bpo-21145)

Added a new functools.singledispatchmethod() decorator that convertsmethods into generic functions usingsingle dispatch:

from functools import singledispatchmethodfrom contextlib import suppressclass TaskManager: def __init__(self, tasks): self.tasks = list(tasks) @singledispatchmethod def discard(self, value): with suppress(ValueError): self.tasks.remove(value) @discard.register(list) def _(self, tasks): targets = set(tasks) self.tasks = [x for x in self.tasks if x not in targets]

(Contributed by Ethan Smith in bpo-32380)

gc¶

get_objects() can now receive an optional generation parameterindicating a generation to get objects from.(Contributed by Pablo Galindo in bpo-36016.)

gettext¶

Added pgettext() and its variants.(Contributed by Franz Glasner, Éric Araujo, and Cheryl Sabella in bpo-2504.)

gzip¶

Added the mtime parameter to gzip.compress() for reproducible output.(Contributed by Guo Ci Teo in bpo-34898.)

A BadGzipFile exception is now raised instead of OSErrorfor certain types of invalid or corrupt gzip files.(Contributed by Filip Gruszczyński, Michele Orrù, and Zackery Spytz inbpo-6584.)

IDLE and idlelib¶

Output over N lines (50 by default) is squeezed down to a button.N can be changed in the PyShell section of the General page of theSettings dialog. Fewer, but possibly extra long, lines can be squeezed byright clicking on the output. Squeezed output can be expanded in placeby double-clicking the button or into the clipboard or a separate windowby right-clicking the button. (Contributed by Tal Einat in bpo-1529353.)

Add “Run Customized” to the Run menu to run a module with customizedsettings. Any command line arguments entered are added to sys.argv.They also re-appear in the box for the next customized run. One can alsosuppress the normal Shell main module restart. (Contributed by CherylSabella, Terry Jan Reedy, and others in bpo-5680 and bpo-37627.)

Added optional line numbers for IDLE editor windows. Windowsopen without line numbers unless set otherwise in the Generaltab of the configuration dialog. Line numbers for an existingwindow are shown and hidden in the Options menu.(Contributed by Tal Einat and Saimadhav Heblikar in bpo-17535.)

OS native encoding is now used for converting between Python strings and Tclobjects. This allows IDLE to work with emoji and other non-BMP characters.These characters can be displayed or copied and pasted to or from theclipboard. Converting strings from Tcl to Python and back now never fails.(Many people worked on this for eight years but the problem was finallysolved by Serhiy Storchaka in bpo-13153.)

New in 3.8.1:

Add option to toggle cursor blink off. (Contributed by Zackery Spytzin bpo-4603.)

Escape key now closes IDLE completion windows. (Contributed by JohnnyNajera in bpo-38944.)

The changes above have been backported to 3.7 maintenance releases.

Add keywords to module name completion list. (Contributed by Terry J.Reedy in bpo-37765.)

inspect¶

The inspect.getdoc() function can now find docstrings for __slots__if that attribute is a dict where the values are docstrings.This provides documentation options similar to what we already havefor property(), classmethod(), and staticmethod():

class AudioClip: __slots__ = {'bit_rate': 'expressed in kilohertz to one decimal place', 'duration': 'in seconds, rounded up to an integer'} def __init__(self, bit_rate, duration): self.bit_rate = round(bit_rate / 1000.0, 1) self.duration = ceil(duration)

(Contributed by Raymond Hettinger in bpo-36326.)

io¶

In development mode (-X env) and in debug build, theio.IOBase finalizer now logs the exception if the close() methodfails. The exception is ignored silently by default in release build.(Contributed by Victor Stinner in bpo-18748.)

itertools¶

The itertools.accumulate() function added an option initial keywordargument to specify an initial value:

>>> from itertools import accumulate>>> list(accumulate([10, 5, 30, 15], initial=1000))[1000, 1010, 1015, 1045, 1060]

(Contributed by Lisa Roach in bpo-34659.)

json.tool¶

Add option --json-lines to parse every input line as a separate JSON object.(Contributed by Weipeng Hong in bpo-31553.)

logging¶

Added a force keyword argument to logging.basicConfig()When set to true, any existing handlers attachedto the root logger are removed and closed before carrying out theconfiguration specified by the other arguments.

This solves a long-standing problem. Once a logger or basicConfig() hadbeen called, subsequent calls to basicConfig() were silently ignored.This made it difficult to update, experiment with, or teach the variouslogging configuration options using the interactive prompt or a Jupyternotebook.

(Suggested by Raymond Hettinger, implemented by Donghee Na, andreviewed by Vinay Sajip in bpo-33897.)

math¶

Added new function math.dist() for computing Euclidean distancebetween two points. (Contributed by Raymond Hettinger in bpo-33089.)

Expanded the math.hypot() function to handle multiple dimensions.Formerly, it only supported the 2-D case.(Contributed by Raymond Hettinger in bpo-33089.)

Added new function, math.prod(), as analogous function to sum()that returns the product of a ‘start’ value (default: 1) times an iterable ofnumbers:

>>> prior = 0.8>>> likelihoods = [0.625, 0.84, 0.30]>>> math.prod(likelihoods, start=prior)0.126

(Contributed by Pablo Galindo in bpo-35606.)

Added two new combinatoric functions math.perm() and math.comb():

>>> math.perm(10, 3) # Permutations of 10 things taken 3 at a time720>>> math.comb(10, 3) # Combinations of 10 things taken 3 at a time120

(Contributed by Yash Aggarwal, Keller Fuchs, Serhiy Storchaka, and RaymondHettinger in bpo-37128, bpo-37178, and bpo-35431.)

Added a new function math.isqrt() for computing accurate integer squareroots without conversion to floating point. The new function supportsarbitrarily large integers. It is faster than floor(sqrt(n)) but slowerthan math.sqrt():

>>> r = 650320427>>> s = r ** 2>>> isqrt(s - 1) # correct650320426>>> floor(sqrt(s - 1)) # incorrect650320427

(Contributed by Mark Dickinson in bpo-36887.)

The function math.factorial() no longer accepts arguments that are notint-like. (Contributed by Pablo Galindo in bpo-33083.)

mmap¶

The mmap.mmap class now has an madvise() method toaccess the madvise() system call.(Contributed by Zackery Spytz in bpo-32941.)

multiprocessing¶

Added new multiprocessing.shared_memory module.(Contributed by Davin Potts in bpo-35813.)

On macOS, the spawn start method is now used by default.(Contributed by Victor Stinner in bpo-33725.)

os¶

Added new function add_dll_directory() on Windows for providingadditional search paths for native dependencies when importing extensionmodules or loading DLLs using ctypes.(Contributed by Steve Dower in bpo-36085.)

A new os.memfd_create() function was added to wrap thememfd_create() syscall.(Contributed by Zackery Spytz and Christian Heimes in bpo-26836.)

On Windows, much of the manual logic for handling reparse points (includingsymlinks and directory junctions) has been delegated to the operating system.Specifically, os.stat() will now traverse anything supported by theoperating system, while os.lstat() will only open reparse points thatidentify as “name surrogates” while others are opened as for os.stat().In all cases, stat_result.st_mode will only have S_IFLNK set forsymbolic links and not other kinds of reparse points. To identify other kindsof reparse point, check the new stat_result.st_reparse_tag attribute.

On Windows, os.readlink() is now able to read directory junctions. Notethat islink() will return False for directory junctions,and so code that checks islink first will continue to treat junctions asdirectories, while code that handles errors from os.readlink() may nowtreat junctions as links.

(Contributed by Steve Dower in bpo-37834.)

os.path¶

os.path functions that return a boolean result likeexists(), lexists(), isdir(),isfile(), islink(), and ismount()now return False instead of raising ValueError or its subclassesUnicodeEncodeError and UnicodeDecodeError for paths that containcharacters or bytes unrepresentable at the OS level.(Contributed by Serhiy Storchaka in bpo-33721.)

expanduser() on Windows now prefers the USERPROFILEenvironment variable and does not use HOME, which is not normally setfor regular user accounts.(Contributed by Anthony Sottile in bpo-36264.)

isdir() on Windows no longer returns True for a link to anon-existent directory.

realpath() on Windows now resolves reparse points, includingsymlinks and directory junctions.

(Contributed by Steve Dower in bpo-37834.)

pathlib¶

pathlib.Path methods that return a boolean result likeexists(), is_dir(),is_file(), is_mount(),is_symlink(), is_block_device(),is_char_device(), is_fifo(),is_socket() now return False instead of raisingValueError or its subclass UnicodeEncodeError for paths thatcontain characters unrepresentable at the OS level.(Contributed by Serhiy Storchaka in bpo-33721.)

Added pathlib.Path.link_to() which creates a hard link pointingto a path.(Contributed by Joannah Nanjekye in bpo-26978)Note that link_to was deprecated in 3.10 and removed in 3.12 infavor of a hardlink_to method added in 3.10 which matches thesemantics of the existing symlink_to method.

pickle¶

pickle extensions subclassing the C-optimized Picklercan now override the pickling logic of functions and classes by defining thespecial reducer_override() method.(Contributed by Pierre Glaser and Olivier Grisel in bpo-35900.)

plistlib¶

Added new plistlib.UID and enabled support for reading and writingNSKeyedArchiver-encoded binary plists.(Contributed by Jon Janzen in bpo-26707.)

pprint¶

The pprint module added a sort_dicts parameter to several functions.By default, those functions continue to sort dictionaries before rendering orprinting. However, if sort_dicts is set to false, the dictionaries retainthe order that keys were inserted. This can be useful for comparison to JSONinputs during debugging.

In addition, there is a convenience new function, pprint.pp() that islike pprint.pprint() but with sort_dicts defaulting to False:

>>> from pprint import pprint, pp>>> d = dict(source='input.txt', operation='filter', destination='output.txt')>>> pp(d, width=40) # Original order{'source': 'input.txt', 'operation': 'filter', 'destination': 'output.txt'}>>> pprint(d, width=40) # Keys sorted alphabetically{'destination': 'output.txt', 'operation': 'filter', 'source': 'input.txt'}

(Contributed by Rémi Lapeyre in bpo-30670.)

py_compile¶

py_compile.compile() now supports silent mode.(Contributed by Joannah Nanjekye in bpo-22640.)

shlex¶

The new shlex.join() function acts as the inverse of shlex.split().(Contributed by Bo Bayles in bpo-32102.)

shutil¶

shutil.copytree() now accepts a new dirs_exist_ok keyword argument.(Contributed by Josh Bronson in bpo-20849.)

shutil.make_archive() now defaults to the modern pax (POSIX.1-2001)format for new archives to improve portability and standards conformance,inherited from the corresponding change to the tarfile module.(Contributed by C.A.M. Gerlach in bpo-30661.)

shutil.rmtree() on Windows now removes directory junctions withoutrecursively removing their contents first.(Contributed by Steve Dower in bpo-37834.)

socket¶

Added create_server() and has_dualstack_ipv6()convenience functions to automate the necessary tasks usually involved whencreating a server socket, including accepting both IPv4 and IPv6 connectionson the same socket. (Contributed by Giampaolo Rodolà in bpo-17561.)

The socket.if_nameindex(), socket.if_nametoindex(), andsocket.if_indextoname() functions have been implemented on Windows.(Contributed by Zackery Spytz in bpo-37007.)

ssl¶

Added post_handshake_auth to enable andverify_client_post_handshake() to initiate TLS 1.3post-handshake authentication.(Contributed by Christian Heimes in bpo-34670.)

statistics¶

Added statistics.fmean() as a faster, floating-point variant ofstatistics.mean(). (Contributed by Raymond Hettinger andSteven D’Aprano in bpo-35904.)

Added statistics.geometric_mean()(Contributed by Raymond Hettinger in bpo-27181.)

Added statistics.multimode() that returns a list of the mostcommon values. (Contributed by Raymond Hettinger in bpo-35892.)

Added statistics.quantiles() that divides data or a distributionin to equiprobable intervals (e.g. quartiles, deciles, or percentiles).(Contributed by Raymond Hettinger in bpo-36546.)

Added statistics.NormalDist, a tool for creatingand manipulating normal distributions of a random variable.(Contributed by Raymond Hettinger in bpo-36018.)

>>> temperature_feb = NormalDist.from_samples([4, 12, -3, 2, 7, 14])>>> temperature_feb.mean6.0>>> temperature_feb.stdev6.356099432828281>>> temperature_feb.cdf(3) # Chance of being under 3 degrees0.3184678262814532>>> # Relative chance of being 7 degrees versus 10 degrees>>> temperature_feb.pdf(7) / temperature_feb.pdf(10)1.2039930378537762>>> el_niño = NormalDist(4, 2.5)>>> temperature_feb += el_niño # Add in a climate effect>>> temperature_febNormalDist(mu=10.0, sigma=6.830080526611674)>>> temperature_feb * (9/5) + 32 # Convert to FahrenheitNormalDist(mu=50.0, sigma=12.294144947901014)>>> temperature_feb.samples(3) # Generate random samples[7.672102882379219, 12.000027119750287, 4.647488369766392]

sys¶

Add new sys.unraisablehook() function which can be overridden to controlhow “unraisable exceptions” are handled. It is called when an exception hasoccurred but there is no way for Python to handle it. For example, when adestructor raises an exception or during garbage collection(gc.collect()).(Contributed by Victor Stinner in bpo-36829.)

tarfile¶

The tarfile module now defaults to the modern pax (POSIX.1-2001)format for new archives, instead of the previous GNU-specific one.This improves cross-platform portability with a consistent encoding (UTF-8)in a standardized and extensible format, and offers several other benefits.(Contributed by C.A.M. Gerlach in bpo-36268.)

threading¶

Add a new threading.excepthook() function which handles uncaughtthreading.Thread.run() exception. It can be overridden to control howuncaught threading.Thread.run() exceptions are handled.(Contributed by Victor Stinner in bpo-1230540.)

Add a new threading.get_native_id() function anda native_idattribute to the threading.Thread class. These return the nativeintegral Thread ID of the current thread assigned by the kernel.This feature is only available on certain platforms, seeget_native_id for more information.(Contributed by Jake Tesler in bpo-36084.)

tokenize¶

The tokenize module now implicitly emits a NEWLINE token whenprovided with input that does not have a trailing new line. This behaviornow matches what the C tokenizer does internally.(Contributed by Ammar Askar in bpo-33899.)

tkinter¶

Added methods selection_from(),selection_present(),selection_range() andselection_to()in the tkinter.Spinbox class.(Contributed by Juliette Monsel in bpo-34829.)

Added method moveto()in the tkinter.Canvas class.(Contributed by Juliette Monsel in bpo-23831.)

The tkinter.PhotoImage class now hastransparency_get() andtransparency_set() methods. (Contributed byZackery Spytz in bpo-25451.)

time¶

Added new clock CLOCK_UPTIME_RAW for macOS 10.12.(Contributed by Joannah Nanjekye in bpo-35702.)

typing¶

The typing module incorporates several new features:

• A dictionary type with per-key types. See PEP 589 andtyping.TypedDict.TypedDict uses only string keys. By default, every key is requiredto be present. Specify “total=False” to allow keys to be optional:

  class Location(TypedDict, total=False): lat_long: tuple grid_square: str xy_coordinate: tuple

• Literal types. See PEP 586 and typing.Literal.Literal types indicate that a parameter or return valueis constrained to one or more specific literal values:

  def get_status(port: int) -> Literal['connected', 'disconnected']: ...

• “Final” variables, functions, methods and classes. See PEP 591,typing.Final and typing.final().The final qualifier instructs a static type checker to restrictsubclassing, overriding, or reassignment:

  pi: Final[float] = 3.1415926536

• Protocol definitions. See PEP 544, typing.Protocol andtyping.runtime_checkable(). Simple ABCs liketyping.SupportsInt are now Protocol subclasses.

• New protocol class typing.SupportsIndex.

• New functions typing.get_origin() and typing.get_args().

unicodedata¶

The unicodedata module has been upgraded to use the Unicode 12.1.0 release.

New function is_normalized() can be used to verify a stringis in a specific normal form, often much faster than by actually normalizingthe string. (Contributed by Max Belanger, David Euresti, and Greg Price inbpo-32285 and bpo-37966).

unittest¶

Added AsyncMock to support an asynchronous version ofMock. Appropriate new assert functions for testinghave been added as well.(Contributed by Lisa Roach in bpo-26467).

Added addModuleCleanup() andaddClassCleanup() to unittest to supportcleanups for setUpModule() andsetUpClass().(Contributed by Lisa Roach in bpo-24412.)

Several mock assert functions now also print a list of actual calls uponfailure. (Contributed by Petter Strandmark in bpo-35047.)

unittest module gained support for coroutines to be used as test caseswith unittest.IsolatedAsyncioTestCase.(Contributed by Andrew Svetlov in bpo-32972.)

Example:

import unittestclass TestRequest(unittest.IsolatedAsyncioTestCase): async def asyncSetUp(self): self.connection = await AsyncConnection() async def test_get(self): response = await self.connection.get("https://example.com") self.assertEqual(response.status_code, 200) async def asyncTearDown(self): await self.connection.close()if __name__ == "__main__": unittest.main()

venv¶

venv now includes an Activate.ps1 script on all platforms foractivating virtual environments under PowerShell Core 6.1.(Contributed by Brett Cannon in bpo-32718.)

weakref¶

The proxy objects returned by weakref.proxy() now support the matrixmultiplication operators @ and @= in addition to the othernumeric operators. (Contributed by Mark Dickinson in bpo-36669.)

xml¶

As mitigation against DTD and external entity retrieval, thexml.dom.minidom and xml.sax modules no longer processexternal entities by default.(Contributed by Christian Heimes in bpo-17239.)

The .find*() methods in the xml.etree.ElementTree modulesupport wildcard searches like {*}tag which ignores the namespaceand {namespace}* which returns all tags in the given namespace.(Contributed by Stefan Behnel in bpo-28238.)

The xml.etree.ElementTree module provides a new function–xml.etree.ElementTree.canonicalize() that implements C14N 2.0.(Contributed by Stefan Behnel in bpo-13611.)

The target object of xml.etree.ElementTree.XMLParser canreceive namespace declaration events through the new callback methodsstart_ns() and end_ns(). Additionally, thexml.etree.ElementTree.TreeBuilder target can be configuredto process events about comments and processing instructions to includethem in the generated tree.(Contributed by Stefan Behnel in bpo-36676 and bpo-36673.)

xmlrpc¶

xmlrpc.client.ServerProxy now supports an optional headers keywordargument for a sequence of HTTP headers to be sent with each request. Amongother things, this makes it possible to upgrade from default basicauthentication to faster session authentication.(Contributed by Cédric Krier in bpo-35153.)

Changes in Python behavior¶

• Yield expressions (both yield and yield from clauses) are now disallowedin comprehensions and generator expressions (aside from the iterable expressionin the leftmost for clause).(Contributed by Serhiy Storchaka in bpo-10544.)

• The compiler now produces a SyntaxWarning when identity checks(is and is not) are used with certain types of literals(e.g. strings, numbers). These can often work by accident in CPython,but are not guaranteed by the language spec. The warning advises usersto use equality tests (== and !=) instead.(Contributed by Serhiy Storchaka in bpo-34850.)

• The CPython interpreter can swallow exceptions in some circ*mstances.In Python 3.8 this happens in fewer cases. In particular, exceptionsraised when getting the attribute from the type dictionary are no longerignored. (Contributed by Serhiy Storchaka in bpo-35459.)

• Removed __str__ implementations from builtin types bool,int, float, complex and few classes fromthe standard library. They now inherit __str__() from object.As result, defining the __repr__() method in the subclass of theseclasses will affect their string representation.(Contributed by Serhiy Storchaka in bpo-36793.)

• On AIX, sys.platform doesn’t contain the major version anymore.It is always 'aix', instead of 'aix3' .. 'aix7'. Sinceolder Python versions include the version number, so it is recommended toalways use sys.platform.startswith('aix').(Contributed by M. Felt in bpo-36588.)

• PyEval_AcquireLock() and PyEval_AcquireThread() nowterminate the current thread if called while the interpreter isfinalizing, making them consistent with PyEval_RestoreThread(),Py_END_ALLOW_THREADS(), and PyGILState_Ensure(). If thisbehavior is not desired, guard the call by checking _Py_IsFinalizing()or sys.is_finalizing().(Contributed by Joannah Nanjekye in bpo-36475.)

Changes in the Python API¶

• The os.getcwdb() function now uses the UTF-8 encoding on Windows,rather than the ANSI code page: see PEP 529 for the rationale. Thefunction is no longer deprecated on Windows.(Contributed by Victor Stinner in bpo-37412.)

• subprocess.Popen can now use os.posix_spawn() in some casesfor better performance. On Windows Subsystem for Linux and QEMU UserEmulation, the Popen constructor using os.posix_spawn() no longer raises anexception on errors like “missing program”. Instead the child process fails with anon-zero returncode.(Contributed by Joannah Nanjekye and Victor Stinner in bpo-35537.)

• The preexec_fn argument of * subprocess.Popen is no longercompatible with subinterpreters. The use of the parameter in asubinterpreter now raises RuntimeError.(Contributed by Eric Snow in bpo-34651, modified by Christian Heimesin bpo-37951.)

• The imap.IMAP4.logout() method no longer silently ignores arbitraryexceptions.(Contributed by Victor Stinner in bpo-36348.)

• The function platform.popen() has been removed, after having been deprecated sincePython 3.3: use os.popen() instead.(Contributed by Victor Stinner in bpo-35345.)

• The statistics.mode() function no longer raises an exceptionwhen given multimodal data. Instead, it returns the first modeencountered in the input data. (Contributed by Raymond Hettingerin bpo-35892.)

• The selection() method of thetkinter.ttk.Treeview class no longer takes arguments. Using it witharguments for changing the selection was deprecated in Python 3.6. Usespecialized methods like selection_set() forchanging the selection. (Contributed by Serhiy Storchaka in bpo-31508.)

• The writexml(), toxml() and toprettyxml() methods ofxml.dom.minidom, and the write() method of xml.etree,now preserve the attribute order specified by the user.(Contributed by Diego Rojas and Raymond Hettinger in bpo-34160.)

• A dbm.dumb database opened with flags 'r' is now read-only.dbm.dumb.open() with flags 'r' and 'w' no longer createsa database if it does not exist.(Contributed by Serhiy Storchaka in bpo-32749.)

• The doctype() method defined in a subclass ofXMLParser will no longer be called and willemit a RuntimeWarning instead of a DeprecationWarning.Define the doctype()method on a target for handling an XML doctype declaration.(Contributed by Serhiy Storchaka in bpo-29209.)

• A RuntimeError is now raised when the custom metaclass doesn’tprovide the __classcell__ entry in the namespace passed totype.__new__. A DeprecationWarning was emitted in Python3.6–3.7. (Contributed by Serhiy Storchaka in bpo-23722.)

• The cProfile.Profile class can now be used as a contextmanager. (Contributed by Scott Sanderson in bpo-29235.)

• shutil.copyfile(), shutil.copy(), shutil.copy2(),shutil.copytree() and shutil.move() use platform-specific“fast-copy” syscalls (seePlatform-dependent efficient copy operations section).

• shutil.copyfile() default buffer size on Windows was changed from16 KiB to 1 MiB.

• The PyGC_Head struct has changed completely. All code that touched thestruct member should be rewritten. (See bpo-33597.)

• The PyInterpreterState struct has been moved into the “internal”header files (specifically Include/internal/pycore_pystate.h). Anopaque PyInterpreterState is still available as part of the publicAPI (and stable ABI). The docs indicate that none of the struct’sfields are public, so we hope no one has been using them. However,if you do rely on one or more of those private fields and have noalternative then please open a BPO issue. We’ll work on helpingyou adjust (possibly including adding accessor functions to thepublic API). (See bpo-35886.)

• The mmap.flush() method now returns None onsuccess and raises an exception on error under all platforms. Previously,its behavior was platform-dependent: a nonzero value was returned on success;zero was returned on error under Windows. A zero value was returned onsuccess; an exception was raised on error under Unix.(Contributed by Berker Peksag in bpo-2122.)

• xml.dom.minidom and xml.sax modules no longer processexternal entities by default.(Contributed by Christian Heimes in bpo-17239.)

• Deleting a key from a read-only dbm database (dbm.dumb,dbm.gnu or dbm.ndbm) raises error (dbm.dumb.error,dbm.gnu.error or dbm.ndbm.error) instead of KeyError.(Contributed by Xiang Zhang in bpo-33106.)

• Simplified AST for literals. All constants will be represented asast.Constant instances. Instantiating old classes Num,Str, Bytes, NameConstant and Ellipsis will returnan instance of Constant.(Contributed by Serhiy Storchaka in bpo-32892.)

• expanduser() on Windows now prefers the USERPROFILEenvironment variable and does not use HOME, which is not normallyset for regular user accounts.(Contributed by Anthony Sottile in bpo-36264.)

• The exception asyncio.CancelledError now inherits fromBaseException rather than Exception and no longer inheritsfrom concurrent.futures.CancelledError.(Contributed by Yury Selivanov in bpo-32528.)

• The function asyncio.wait_for() now correctly waits for cancellationwhen using an instance of asyncio.Task. Previously, upon reachingtimeout, it was cancelled and immediately returned.(Contributed by Elvis Pranskevichus in bpo-32751.)

• The function asyncio.BaseTransport.get_extra_info() now returns a safeto use socket object when ‘socket’ is passed to the name parameter.(Contributed by Yury Selivanov in bpo-37027.)

• asyncio.BufferedProtocol has graduated to the stable API.

• DLL dependencies for extension modules and DLLs loaded with ctypes onWindows are now resolved more securely. Only the system paths, the directorycontaining the DLL or PYD file, and directories added withadd_dll_directory() are searched for load-time dependencies.Specifically, PATH and the current working directory are no longerused, and modifications to these will no longer have any effect on normal DLLresolution. If your application relies on these mechanisms, you should checkfor add_dll_directory() and if it exists, use it to add your DLLsdirectory while loading your library. Note that Windows 7 users will need toensure that Windows Update KB2533623 has been installed (this is also verifiedby the installer).(Contributed by Steve Dower in bpo-36085.)

• The header files and functions related to pgen have been removed after itsreplacement by a pure Python implementation. (Contributed by Pablo Galindoin bpo-36623.)

• types.CodeType has a new parameter in the second position of theconstructor (posonlyargcount) to support positional-only arguments definedin PEP 570. The first argument (argcount) now represents the totalnumber of positional arguments (including positional-only arguments). The newreplace() method of types.CodeType can be used to make the codefuture-proof.

• The parameter digestmod for hmac.new() no longer uses the MD5 digestby default.

Changes in the C API¶

• The PyCompilerFlags structure got a new cf_feature_versionfield. It should be initialized to PY_MINOR_VERSION. The field is ignoredby default, and is used if and only if PyCF_ONLY_AST flag is set incf_flags.(Contributed by Guido van Rossum in bpo-35766.)

• The PyEval_ReInitThreads() function has been removed from the C API.It should not be called explicitly: use PyOS_AfterFork_Child()instead.(Contributed by Victor Stinner in bpo-36728.)

• On Unix, C extensions are no longer linked to libpython except on Androidand Cygwin. When Python is embedded, libpython must not be loaded withRTLD_LOCAL, but RTLD_GLOBAL instead. Previously, usingRTLD_LOCAL, it was already not possible to load C extensions whichwere not linked to libpython, like C extensions of the standardlibrary built by the *shared* section of Modules/Setup.(Contributed by Victor Stinner in bpo-21536.)

• Use of # variants of formats in parsing or building value (e.g.PyArg_ParseTuple(), Py_BuildValue(), PyObject_CallFunction(),etc.) without PY_SSIZE_T_CLEAN defined raises DeprecationWarning now.It will be removed in 3.10 or 4.0. Read Parsing arguments and building values for detail.(Contributed by Inada Naoki in bpo-36381.)

• Instances of heap-allocated types (such as those created withPyType_FromSpec()) hold a reference to their type object.Increasing the reference count of these type objects has been moved fromPyType_GenericAlloc() to the more low-level functions,PyObject_Init() and PyObject_INIT().This makes types created through PyType_FromSpec() behave likeother classes in managed code.

  Statically allocated types are not affected.

  For the vast majority of cases, there should be no side effect.However, types that manually increase the reference count after allocatingan instance (perhaps to work around the bug) may now become immortal.To avoid this, these classes need to call Py_DECREF on the type objectduring instance deallocation.

  To correctly port these types into 3.8, please apply the followingchanges:

  • Remove Py_INCREF on the type object after allocating aninstance - if any.This may happen after calling PyObject_New,PyObject_NewVar, PyObject_GC_New(),PyObject_GC_NewVar(), or any other custom allocator that usesPyObject_Init() or PyObject_INIT().

    Example:

    static foo_struct *foo_new(PyObject *type) { foo_struct *foo = PyObject_GC_New(foo_struct, (PyTypeObject *) type); if (foo == NULL) return NULL;#if PY_VERSION_HEX < 0x03080000 // Workaround for Python issue 35810; no longer necessary in Python 3.8 PY_INCREF(type)#endif return foo;}

  • Ensure that all custom tp_dealloc functions of heap-allocated typesdecrease the type’s reference count.

    Example:

    static voidfoo_dealloc(foo_struct *instance) { PyObject *type = Py_TYPE(instance); PyObject_GC_Del(instance);#if PY_VERSION_HEX >= 0x03080000 // This was not needed before Python 3.8 (Python issue 35810) Py_DECREF(type);#endif}

  (Contributed by Eddie Elizondo in bpo-35810.)

• The Py_DEPRECATED() macro has been implemented for MSVC.The macro now must be placed before the symbol name.

  Example:

  Py_DEPRECATED(3.8) PyAPI_FUNC(int) Py_OldFunction(void);

  (Contributed by Zackery Spytz in bpo-33407.)

• The interpreter does not pretend to support binary compatibility ofextension types across feature releases, anymore. A PyTypeObjectexported by a third-party extension module is supposed to have all theslots expected in the current Python version, includingtp_finalize (Py_TPFLAGS_HAVE_FINALIZEis not checked anymore before reading tp_finalize).

  (Contributed by Antoine Pitrou in bpo-32388.)

• The functions PyNode_AddChild() and PyParser_AddToken() now accepttwo additional int arguments end_lineno and end_col_offset.

• The libpython38.a file to allow MinGW tools to link directly againstpython38.dll is no longer included in the regular Windows distribution.If you require this file, it may be generated with the gendef anddlltool tools, which are part of the MinGW binutils package:

  gendef - python38.dll > tmp.defdlltool --dllname python38.dll --def tmp.def --output-lib libpython38.a

  The location of an installed pythonXY.dll will depend on theinstallation options and the version and language of Windows. SeeUsing Python on Windows for more information. The resulting library should beplaced in the same directory as pythonXY.lib, which is generally thelibs directory under your Python installation.

  (Contributed by Steve Dower in bpo-37351.)

CPython bytecode changes¶

• The interpreter loop has been simplified by moving the logic of unrollingthe stack of blocks into the compiler. The compiler emits now explicitinstructions for adjusting the stack of values and calling thecleaning-up code for break, continue andreturn.

  Removed opcodes BREAK_LOOP, CONTINUE_LOOP,SETUP_LOOP and SETUP_EXCEPT. Added new opcodesROT_FOUR, BEGIN_FINALLY, CALL_FINALLY andPOP_FINALLY. Changed the behavior of END_FINALLYand WITH_CLEANUP_START.

  (Contributed by Mark Shannon, Antoine Pitrou and Serhiy Storchaka inbpo-17611.)

• Added new opcode END_ASYNC_FOR for handling exceptions raisedwhen awaiting a next item in an async for loop.(Contributed by Serhiy Storchaka in bpo-33041.)

• The MAP_ADD now expects the value as the first element in thestack and the key as the second element. This change was made so the keyis always evaluated before the value in dictionary comprehensions, asproposed by PEP 572. (Contributed by Jörn Heissler in bpo-35224.)

Demos and Tools¶

Added a benchmark script for timing various ways to access variables:Tools/scripts/var_access_benchmark.py.(Contributed by Raymond Hettinger in bpo-35884.)

Here’s a summary of performance improvements since Python 3.3:

Python version 3.3 3.4 3.5 3.6 3.7 3.8-------------- --- --- --- --- --- ---Variable and attribute read access: read_local 4.0 7.1 7.1 5.4 5.1 3.9 read_nonlocal 5.3 7.1 8.1 5.8 5.4 4.4 read_global 13.3 15.5 19.0 14.3 13.6 7.6 read_builtin 20.0 21.1 21.6 18.5 19.0 7.5 read_classvar_from_class 20.5 25.6 26.5 20.7 19.5 18.4 read_classvar_from_instance 18.5 22.8 23.5 18.8 17.1 16.4 read_instancevar 26.8 32.4 33.1 28.0 26.3 25.4 read_instancevar_slots 23.7 27.8 31.3 20.8 20.8 20.2 read_namedtuple 68.5 73.8 57.5 45.0 46.8 18.4 read_boundmethod 29.8 37.6 37.9 29.6 26.9 27.7Variable and attribute write access: write_local 4.6 8.7 9.3 5.5 5.3 4.3 write_nonlocal 7.3 10.5 11.1 5.6 5.5 4.7 write_global 15.9 19.7 21.2 18.0 18.0 15.8 write_classvar 81.9 92.9 96.0 104.6 102.1 39.2 write_instancevar 36.4 44.6 45.8 40.0 38.9 35.5 write_instancevar_slots 28.7 35.6 36.1 27.3 26.6 25.7Data structure read access: read_list 19.2 24.2 24.5 20.8 20.8 19.0 read_deque 19.9 24.7 25.5 20.2 20.6 19.8 read_dict 19.7 24.3 25.7 22.3 23.0 21.0 read_strdict 17.9 22.6 24.3 19.5 21.2 18.9Data structure write access: write_list 21.2 27.1 28.5 22.5 21.6 20.0 write_deque 23.8 28.7 30.1 22.7 21.8 23.5 write_dict 25.9 31.4 33.3 29.3 29.2 24.7 write_strdict 22.9 28.4 29.9 27.5 25.2 23.1Stack (or queue) operations: list_append_pop 144.2 93.4 112.7 75.4 74.2 50.8 deque_append_pop 30.4 43.5 57.0 49.4 49.2 42.5 deque_append_popleft 30.8 43.7 57.3 49.7 49.7 42.8Timing loop: loop_overhead 0.3 0.5 0.6 0.4 0.3 0.3

The benchmarks were measured on anIntel® Core™ i7-4960HQ processorrunning the macOS 64-bit builds found atpython.org.The benchmark script displays timings in nanoseconds.

macOS 11.0 (Big Sur) and Apple Silicon Mac support¶

As of 3.8.10, Python now supports building and running on macOS 11(Big Sur) and on Apple Silicon Macs (based on the ARM64 architecture).A new universal build variant, universal2, is now available to nativelysupport both ARM64 and Intel 64 in one set of executables.Note that support for “weaklinking”, building binaries targeted for newerversions of macOS that will also run correctly on older versions bytesting at runtime for missing features, is not included in this backportfrom Python 3.9; to support a range of macOS versions, continue to targetfor and build on the oldest version in the range.

(Originally contributed by Ronald Oussoren and Lawrence D’Anna in gh-85272,with fixes by FX Coudert and Eli Rykoff, and backported to 3.8 by Maxime Bélangerand Ned Deily)

urllib.parse¶

The presence of newline or tab characters in parts of a URL allows for someforms of attacks. Following the WHATWG specification that updates RFC 3986,ASCII newline \n, \r and tab \t characters are stripped from theURL by the parser in urllib.parse preventing such attacks. The removalcharacters are controlled by a new module level variableurllib.parse._UNSAFE_URL_BYTES_TO_REMOVE. (See bpo-43882)

Changes in the Python API¶

Starting with Python 3.8.12 the ipaddress module no longer acceptsany leading zeros in IPv4 address strings. Leading zeros are ambiguous andinterpreted as octal notation by some libraries. For example the legacyfunction socket.inet_aton() treats leading zeros as octal notation.glibc implementation of modern inet_pton() does not acceptany leading zeros.

(Originally contributed by Christian Heimes in bpo-36384, and backportedto 3.8 by Achraf Merzouki.)

tarfile¶

• The extraction methods in tarfile, and shutil.unpack_archive(),have a new a filter argument that allows limiting tar features than may besurprising or dangerous, such as creating files outside the destinationdirectory.See Extraction filters for details.In Python 3.12, use without the filter argument will show aDeprecationWarning.In Python 3.14, the default will switch to 'data'.(Contributed by Petr Viktorin in PEP 706.)