What’s New In Python 3.5¶

Release:	3.5.0
Date:	September 13, 2015
Editors:	Elvis Pranskevichus <elprans@gmail.com>, Yury Selivanov <yselivanov@gmail.com>

Python 3.5 was released on September 13, 2015.

This article explains the new features in Python 3.5, compared to 3.4. For full details, see the changelog.

See also

PEP 478 - Python 3.5 Release Schedule

Summary – Release highlights¶

New syntax features:

PEP 492, coroutines with async and await syntax.
PEP 465, a new matrix multiplication operator: a @ b.
PEP 448, additional unpacking generalizations.

New library modules:

zipapp: Improving Python ZIP Application Support (PEP 441).

New built-in features:

bytes % args, bytearray % args: PEP 461 - Adding % formatting to bytes and bytearray.
b'\xf0\x9f\x90\x8d'.hex(), bytearray(b'\xf0\x9f\x90\x8d').hex(), memoryview(b'\xf0\x9f\x90\x8d').hex(): issue 9951 - A hex method has been added to bytes, bytearray, and memoryview.
memoryview (including multi-dimensional) now supports tuple indexing. (Contributed by Antoine Pitrou in issue 23632.)
Generators have new gi_yieldfrom attribute, which returns the object being iterated by yield from expressions. (Contributed by Benno Leslie and Yury Selivanov in issue 24450.)
New RecursionError exception. (Contributed by Georg Brandl in issue 19235.)
New StopAsyncIteration exception. (Contributed by Yury Selivanov in issue 24017. See also PEP 492.)

CPython implementation improvements:

When the LC_TYPE locale is the POSIX locale (C locale), sys.stdin and sys.stdout are now using the surrogateescape error handler, instead of the strict error handler. (Contributed by Victor Stinner in issue 19977.)
.pyo files are no longer used and have been replaced by a more flexible scheme that inclides the optimization level explicitly in .pyc name. (PEP 488)
Builtin and extension modules are now initialized in a multi-phase process, which is similar to how Python modules are loaded. (PEP 489).

Significantly Improved Library Modules:

collections.OrderedDict is now implemented in C, which makes it 4 to 100 times faster. (Contributed by Eric Snow in issue 16991.)
You may now pass bytes to the tempfile module’s APIs and it will return the temporary pathname as bytes instead of str. It also accepts a value of None on parameters where only str was accepted in the past to do the right thing based on the types of the other inputs. Two functions, gettempdirb() and gettempprefixb(), have been added to go along with this. This behavior matches that of the os APIs. (Contributed by Gregory P. Smith in issue 24230.)
ssl module gained support for Memory BIO, which decouples SSL protocol handling from network IO. (Contributed by Geert Jansen in issue 21965.)
traceback has new lightweight and convenient to work with classes TracebackException, StackSummary, and FrameSummary. (Contributed by Robert Collins in issue 17911.)
Most of functools.lru_cache() machinery is now implemented in C. (Contributed by Matt Joiner, Alexey Kachayev, and Serhiy Storchaka in issue 14373.)

Security improvements:

SSLv3 is now disabled throughout the standard library. It can still be enabled by instantiating a ssl.SSLContext manually. (See issue 22638 for more details; this change was backported to CPython 3.4 and 2.7.)
HTTP cookie parsing is now stricter, in order to protect against potential injection attacks. (Contributed by Antoine Pitrou in issue 22796.)

Windows improvements:

A new installer for Windows has replaced the old MSI. See Using Python on Windows for more information.
Windows builds now use Microsoft Visual C++ 14.0, and extension modules should use the same.

Please read on for a comprehensive list of user-facing changes, including many other smaller improvements, CPython optimizations, deprecations, and potential porting issues.

New Features¶

PEP 492 - Coroutines with async and await syntax¶

PEP 492 greatly improves support for asynchronous programming in Python by adding awaitable objects, coroutine functions, asynchronous iteration, and asynchronous context managers.

Coroutine functions are declared using the new async def syntax:

>>> async def coro():
...     return 'spam'

Inside a coroutine function, the new await expression can be used to suspend coroutine execution until the result is available. Any object can be awaited, as long as it implements the awaitable protocol by defining the __await__() method.

PEP 492 also adds async for statement for convenient iteration over asynchronous iterables.

An example of a simple HTTP client written using the new syntax:

import asyncio

async def http_get(domain):
    reader, writer = await asyncio.open_connection(domain, 80)

    writer.write(b'\r\n'.join([
        b'GET / HTTP/1.1',
        b'Host: %b' % domain.encode('latin-1'),
        b'Connection: close',
        b'', b''
    ]))

    async for line in reader:
        print('>>>', line)

    writer.close()

loop = asyncio.get_event_loop()
try:
    loop.run_until_complete(http_get('example.com'))
finally:
    loop.close()

Similarly to asynchronous iteration, there is a new syntax for asynchronous context managers. The following script:

import asyncio

async def coro(name, lock):
    print('coro {}: waiting for lock'.format(name))
    async with lock:
        print('coro {}: holding the lock'.format(name))
        await asyncio.sleep(1)
        print('coro {}: releasing the lock'.format(name))

loop = asyncio.get_event_loop()
lock = asyncio.Lock()
coros = asyncio.gather(coro(1, lock), coro(2, lock))
try:
    loop.run_until_complete(coros)
finally:
    loop.close()

will print:

coro 2: waiting for lock
coro 2: holding the lock
coro 1: waiting for lock
coro 2: releasing the lock
coro 1: holding the lock
coro 1: releasing the lock

Note that both async for and async with can only be used inside a coroutine function declared with async def.

Coroutine functions are intended to be run inside a compatible event loop, such as asyncio.Loop.

See also

PEP 492 – Coroutines with async and await syntax: PEP written and implemented by Yury Selivanov.

PEP 465 - A dedicated infix operator for matrix multiplication¶

PEP 465 adds the @ infix operator for matrix multiplication. Currently, no builtin Python types implement the new operator, however, it can be implemented by defining __matmul__(), __rmatmul__(), and __imatmul__() for regular, reflected, and in-place matrix multiplication. The semantics of these methods is similar to that of methods defining other infix arithmetic operators.

Matrix multiplication is a notably common operation in many fields of mathematics, science, engineering, and the addition of @ allows writing cleaner code:

S = (H @ beta - r).T @ inv(H @ V @ H.T) @ (H @ beta - r)

instead of:

S = dot((dot(H, beta) - r).T,
        dot(inv(dot(dot(H, V), H.T)), dot(H, beta) - r))

An upcoming release of NumPy 1.10 will add support for the new operator:

>>> import numpy

>>> x = numpy.ones(3)
>>> x
array([ 1., 1., 1.])

>>> m = numpy.eye(3)
>>> m
array([[ 1., 0., 0.],
       [ 0., 1., 0.],
       [ 0., 0., 1.]])

>>> x @ m
array([ 1., 1., 1.])

See also

PEP 465 – A dedicated infix operator for matrix multiplication: PEP written by Nathaniel J. Smith; implemented by Benjamin Peterson.

PEP 448 - Additional Unpacking Generalizations¶

PEP 448 extends the allowed uses of the * iterable unpacking operator and ** dictionary unpacking operator. It is now possible to use an arbitrary number of unpackings in function calls:

>>> print(*[1], *[2], 3, *[4, 5])
1 2 3 4 5

>>> def fn(a, b, c, d):
...     print(a, b, c, d)
...

>>> fn(**{'a': 1, 'c': 3}, **{'b': 2, 'd': 4})
1 2 3 4

Similarly, tuple, list, set, and dictionary displays allow multiple unpackings:

>>> *range(4), 4
(0, 1, 2, 3, 4)

>>> [*range(4), 4]
[0, 1, 2, 3, 4]

>>> {*range(4), 4, *(5, 6, 7)}
{0, 1, 2, 3, 4, 5, 6, 7}

>>> {'x': 1, **{'y': 2}}
{'x': 1, 'y': 2}

See also

PEP 448 – Additional Unpacking Generalizations: PEP written by Joshua Landau; implemented by Neil Girdhar, Thomas Wouters, and Joshua Landau.

PEP 461 - % formatting support for bytes and bytearray¶

PEP 461 adds % formatting to bytes and bytearray, aiding in handling data that is a mixture of binary and ASCII compatible text. This feature also eases porting such code from Python 2.

Examples:

>>> b'Hello %s!' % b'World'
b'Hello World!'

>>> b'x=%i y=%f' % (1, 2.5)
b'x=1 y=2.500000'

Unicode is not allowed for %s, but it is accepted by %a (equivalent of repr(obj).encode('ascii', 'backslashreplace')):

>>> b'Hello %s!' % 'World'
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
TypeError: %b requires bytes, or an object that implements __bytes__, not 'str'

>>> b'price: %a' % '10€'
b"price: '10\\u20ac'"
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
TypeError: %b requires bytes, or an object that implements __bytes__, not 'str'

See also

PEP 461 – Adding % formatting to bytes and bytearray: PEP written by Ethan Furman; implemented by Neil Schemenauer and Ethan Furman.

PEP 484 - Type Hints¶

This PEP introduces a provisional module to provide these standard definitions and tools, along with some conventions for situations where annotations are not available.

For example, here is a simple function whose argument and return type are declared in the annotations:

def greeting(name: str) -> str:
    return 'Hello ' + name

The type system supports unions, generic types, and a special type named Any which is consistent with (i.e. assignable to and from) all types.

See also

typing module documentation
PEP 484 – Type Hints

PEP written by Guido van Rossum, Jukka Lehtosalo, and Łukasz Langa; implemented by Guido van Rossum.

PEP 471 - os.scandir() function – a better and faster directory iterator¶

PEP 471 adds a new directory iteration function, os.scandir(), to the standard library. Additionally, os.walk() is now implemented using os.scandir(), which speeds it up by 3-5 times on POSIX systems and by 7-20 times on Windows systems.

See also

PEP 471 – os.scandir() function – a better and faster directory iterator: PEP written and implemented by Ben Hoyt with the help of Victor Stinner.

PEP 475: Retry system calls failing with EINTR¶

PEP 475 adds support for automatic retry of system calls failing with EINTR: this means that user code doesn’t have to deal with EINTR or InterruptedError manually, and should make it more robust against asynchronous signal reception.

Examples of functions which are now retried when interrupted by a signal instead of raising InterruptedError if the Python signal handler does not raise an exception:

open(), os.open(), io.open();
functions of the faulthandler module;
os functions: fchdir(), fchmod(), fchown(), fdatasync(), fstat(), fstatvfs(), fsync(), ftruncate(), mkfifo(), mknod(), posix_fadvise(), posix_fallocate(), pread(), pwrite(), read(), readv(), sendfile(), wait3(), wait4(), wait(), waitid(), waitpid(), write(), writev();
special cases: os.close() and os.dup2() now ignore EINTR error, the syscall is not retried (see the PEP for the rationale);
select functions: poll(), poll(), control(), poll(), select();
socket.socket() methods: accept(), connect() (except for non-blocking sockets), recv(), recvfrom(), recvmsg(), send(), sendall(), sendmsg(), sendto();
signal.sigtimedwait(), signal.sigwaitinfo();
time.sleep().

See also

PEP 475 – Retry system calls failing with EINTR: PEP and implementation written by Charles-François Natali and Victor Stinner, with the help of Antoine Pitrou (the french connection).

PEP 479: Change StopIteration handling inside generators¶

PEP 479 changes the behavior of generators: when a StopIteration exception is raised inside a generator, it is replaced with a RuntimeError. To enable the feature a __future__ import should be used:

from __future__ import generator_stop

Without a __future__ import, a PendingDeprecationWarning will be raised.

See also

PEP 479 – Change StopIteration handling inside generators: PEP written by Chris Angelico and Guido van Rossum. Implemented by Chris Angelico, Yury Selivanov and Nick Coghlan.

PEP 486: Make the Python Launcher aware of virtual environments¶

PEP 486 makes the Windows launcher (see PEP 397) aware of an active virtual environment. When the default interpreter would be used and the VIRTUAL_ENV environment variable is set, the interpreter in the virtual environment will be used.

See also

PEP 486 – Make the Python Launcher aware of virtual environments: PEP written and implemented by Paul Moore.

PEP 488: Elimination of PYO files¶

PEP 488 does away with the concept of .pyo files. This means that .pyc files represent both unoptimized and optimized bytecode. To prevent the need to constantly regenerate bytecode files, .pyc files now have an optional opt- tag in their name when the bytecode is optimized. This has the side-effect of no more bytecode file name clashes when running under either -O or -OO. Consequently, bytecode files generated from -O, and -OO may now exist simultaneously. importlib.util.cache_from_source() has an updated API to help with this change.

See also

PEP 488 – Elimination of PYO files: PEP written and implemented by Brett Cannon.

PEP 489: Multi-phase extension module initialization¶

PEP 489 updates extension module initialization to take advantage of the two step module loading mechanism introduced by PEP 451 in Python 3.4.

This change brings the import semantics of extension modules that opt-in to using the new mechanism much closer to those of Python source and bytecode modules, including the ability to use any valid identifier as a module name, rather than being restricted to ASCII.

See also

PEP 489 – Multi-phase extension module initialization: PEP written by Petr Viktorin, Stefan Behnel, and Nick Coghlan; implemented by Petr Viktorin.

PEP 485: A function for testing approximate equality¶

PEP 485 adds the math.isclose() and cmath.isclose() functions which tell whether two values are approximately equal or “close” to each other. Whether or not two values are considered close is determined according to given absolute and relative tolerances.

See also

PEP 485 – A function for testing approximate equality: PEP written by Christopher Barker; implemented by Chris Barker and Tal Einat.

Other Language Changes¶

Some smaller changes made to the core Python language are:

Added the "namereplace" error handlers. The "backslashreplace" error handlers now works with decoding and translating. (Contributed by Serhiy Storchaka in issue 19676 and issue 22286.)
The -b option now affects comparisons of bytes with int. (Contributed by Serhiy Storchaka in issue 23681.)
New Kazakh codec kz1048. (Contributed by Serhiy Storchaka in issue 22682.)
Property docstrings are now writable. This is especially useful for collections.namedtuple() docstrings. (Contributed by Berker Peksag in issue 24064.)
New Tajik codec koi8_t. (Contributed by Serhiy Storchaka in issue 22681.)
Circular imports involving relative imports are now supported. (Contributed by Brett Cannon and Antoine Pitrou in issue 17636.)

New Modules¶

zipapp¶

The new zipapp module (specified in PEP 441) provides an API and command line tool for creating executable Python Zip Applications, which were introduced in Python 2.6 in issue 1739468, but which were not well publicized, either at the time or since.

With the new module, bundling your application is as simple as putting all the files, including a __main__.py file, into a directory myapp and running:

$ python -m zipapp myapp
$ python myapp.pyz

The module implementation has been contributed by Paul Moore in issue 23491.

See also

PEP 441 – Improving Python ZIP Application Support

Improved Modules¶

argparse¶

The ArgumentParser class now allows to disable abbreviated usage of long options by setting allow_abbrev to False. (Contributed by Jonathan Paugh, Steven Bethard, paul j3 and Daniel Eriksson in issue 14910.)

bz2¶

The BZ2Decompressor.decompress method now accepts an optional max_length argument to limit the maximum size of decompressed data. (Contributed by Nikolaus Rath in issue 15955.)

cgi¶

The FieldStorage class now supports the context management protocol. (Contributed by Berker Peksag in issue 20289.)

cmath¶

A new function isclose() provides a way to test for approximate equality. (Contributed by Chris Barker and Tal Einat in issue 24270.)

code¶

The InteractiveInterpreter.showtraceback method now prints the full chained traceback, just like the interactive interpreter. (Contributed by Claudiu Popa in issue 17442.)

collections¶

The OrderedDict class is now implemented in C, which makes it 4 to 100 times faster. (Contributed by Eric Snow in issue 16991.)

OrderedDict.items, OrderedDict.keys, OrderedDict.values views now support reversed() iteration. (Contributed by Serhiy Storchaka in issue 19505.)

The deque class now defines index(), insert(), and copy(), as well as supports + and * operators. This allows deques to be recognized as a MutableSequence and improves their substitutability for lists. (Contributed by Raymond Hettinger issue 23704.)

Docstrings produced by namedtuple() can now be updated:

Point = namedtuple('Point', ['x', 'y'])
Point.__doc__ = 'ordered pair'
Point.x.__doc__ = 'abscissa'
Point.y.__doc__ = 'ordinate'

(Contributed by Berker Peksag in issue 24064.)

The UserString class now implements __getnewargs__(), __rmod__(), casefold(), format_map(), isprintable(), and maketrans() methods to match corresponding methods of str. (Contributed by Joe Jevnik in issue 22189.)

collections.abc¶

A new Generator abstract base class. (Contributed by Stefan Behnel in issue 24018.)

New Coroutine, AsyncIterator, and AsyncIterable abstract base classes. (Contributed by Yury Selivanov in issue 24184.)

compileall¶

A new compileall option, -j N, allows to run N workers sumultaneously to perform parallel bytecode compilation. The compile_dir() function has a corresponding workers parameter. (Contributed by Claudiu Popa in issue 16104.)

The -q command line option can now be specified more than once, in which case all output, including errors, will be suppressed. The corresponding quiet parameter in compile_dir(), compile_file(), and compile_path() can now accept an integer value indicating the level of output suppression. (Contributed by Thomas Kluyver in issue 21338.)

concurrent.futures¶

The Executor.map method now accepts a chunksize argument to allow batching of tasks to improve performance when ProcessPoolExecutor() is used. (Contributed by Dan O’Reilly in issue 11271.)

contextlib¶

The new redirect_stderr() context manager (similar to redirect_stdout()) makes it easier for utility scripts to handle inflexible APIs that write their output to sys.stderr and don’t provide any options to redirect it. (Contributed by Berker Peksag in issue 22389.)

curses¶

The new update_lines_cols() function updates LINES and COLS environment variables. This is useful for detecting manual screen resize. (Contributed by Arnon Yaari in issue 4254.)

difflib¶

The charset of HTML documents generated by HtmlDiff.make_file can now be customized by using a new charset keyword-only argument. The default charset of HTML document changed from "ISO-8859-1" to "utf-8". (Contributed by Berker Peksag in issue 2052.)

The diff_bytes() function can now compare lists of byte strings. This fixes a regression from Python 2. (Contributed by Terry J. Reedy and Greg Ward in issue 17445.)

distutils¶

Both build and build_ext commands now accept a -j option to enable parallel building of extension modules. (Contributed by Antoine Pitrou in issue 5309.)

The distutils module now supports xz compression, and can be enabled by passing xztar as an argument to bdist --format. (Contributed by Serhiy Storchaka in issue 16314.)

doctest¶

The DocTestSuite() function returns an empty unittest.TestSuite if module contains no docstrings instead of raising ValueError. (Contributed by Glenn Jones in issue 15916.)

email¶

A new policy option Policy.mangle_from_ controls whether or not lines that start with "From " in email bodies are prefixed with a ">" character by generators. The default is True for compat32 and False for all other policies. (Contributed by Milan Oberkirch in issue 20098.)

A new Message.get_content_disposition method provides easy access to a canonical value for the Content-Disposition header. (Contributed by Abhilash Raj in issue 21083.)

A new policy option EmailPolicy.utf8 can be set to True to encode email headers using the UTF-8 charset instead of using encoded words. This allows Messages to be formatted according to RFC 6532 and used with an SMTP server that supports the RFC 6531 SMTPUTF8 extension. (Contributed by R. David Murray in issue 24211.)

enum¶

The Enum callable has a new parameter start to specify the initial number of enum values if only names are provided:

>>> Animal = enum.Enum('Animal', 'cat dog', start=10)
>>> Animal.cat
<Animal.cat: 10>
>>> Animal.dog
<Animal.dog: 11>

(Contributed by Ethan Furman in issue 21706.)

faulthandler¶

enable(), register(), dump_traceback() and dump_traceback_later() functions now accept file descriptors in addition to file-like objects. (Contributed by Wei Wu in issue 23566.)

functools¶

Most of lru_cache() machinery is now implemented in C, making it significantly faster. (Contributed by Matt Joiner, Alexey Kachayev, and Serhiy Storchaka in issue 14373.)

glob¶

iglob() and glob() functions now support recursive search in subdirectories using the "**" pattern. (Contributed by Serhiy Storchaka in issue 13968.)

heapq¶

Element comparison in merge() can now be customized by passing a key function in a new optional key keyword argument. A new optional reverse keyword argument can be used to reverse element comparison. (Contributed by Raymond Hettinger in issue 13742.)

http¶

A new HTTPStatus enum that defines a set of HTTP status codes, reason phrases and long descriptions written in English. (Contributed by Demian Brecht in issue 21793.)

idlelib and IDLE¶

Since idlelib implements the IDLE shell and editor and is not intended for import by other programs, it gets improvements with every release. See Lib/idlelib/NEWS.txt for a cumulative list of changes since 3.4.0, as well as changes made in future 3.5.x releases. This file is also available from the IDLE Help -> About Idle dialog.

imaplib¶

The IMAP4 class now supports context manager protocol. When used in a with statement, the IMAP4 LOGOUT command will be called automatically at the end of the block. (Contributed by Tarek Ziadé and Serhiy Storchaka in issue 4972.)

The imaplib module now supports RFC 5161 (ENABLE Extension) and RFC 6855 (UTF-8 Support) via the IMAP4.enable method. A new IMAP4.utf8_enabled attribute, tracks whether or not RFC 6855 support is enabled. (Contributed by Milan Oberkirch, R. David Murray, and Maciej Szulik in issue 21800.)

The imaplib module now automatically encodes non-ASCII string usernames and passwords using UTF-8, as recommended by the RFCs. (Contributed by Milan Oberkirch in issue 21800.)

imghdr¶

The what() function now recognizes the OpenEXR format (contributed by Martin Vignali and Claudiu Popa in issue 20295), and the WebP format (contributed by Fabrice Aneche and Claudiu Popa in issue 20197.)

importlib¶

The util.LazyLoader class allows for lazy loading of modules in applications where startup time is important. (Contributed by Brett Cannon in issue 17621.)

The abc.InspectLoader.source_to_code method is now a static method. This makes it easier to initialize a module object with code compiled from a string by running exec(code, module.__dict__). (Contributed by Brett Cannon in issue 21156.)

The new util.module_from_spec function is now the preferred way to create a new module. As opposed to creating a types.ModuleType instance directly, this new function will set the various import-controlled attributes based on the passed-in spec object. (Contributed by Brett Cannon in issue 20383.)

inspect¶

Both Signature and Parameter classes are now picklable and hashable. (Contributed by Yury Selivanov in issue 20726 and issue 20334.)

A new BoundArguments.apply_defaults method provides a way to set default values for missing arguments. (Contributed by Yury Selivanov in issue 24190.)

A new class method Signature.from_callable makes subclassing of Signature easier. (Contributed by Yury Selivanov and Eric Snow in issue 17373.)

The signature() function now accepts a follow_wrapped optional keyword argument, which, when set to False, disables automatic following of __wrapped__ links. (Contributed by Yury Selivanov in issue 20691.)

A set of new functions to inspect coroutine functions and coroutine objects has been added: iscoroutine(), iscoroutinefunction(), isawaitable(), getcoroutinelocals(), and getcoroutinestate(). (Contributed by Yury Selivanov in issue 24017 and issue 24400.)

stack(), trace(), getouterframes(), and getinnerframes() functions now return a list of named tuples. (Contributed by Daniel Shahaf in issue 16808.)

io¶

A new BufferedIOBase.readinto1 method, that uses at most one call to the underlying raw stream’s RawIOBase.read (or RawIOBase.readinto) method. (Contributed by Nikolaus Rath in issue 20578.)

ipaddress¶

Both IPv4Network and IPv6Network classes now accept an (address, netmask) tuple argument, so as to easily construct network objects from existing addresses. (Contributed by Peter Moody and Antoine Pitrou in issue 16531.)

A new reverse_pointer> attribute for IPv4Network and IPv6Network classes returns the name of the reverse DNS PTR record. (Contributed by Leon Weber in issue 20480.)

json¶

The json.tool command line interface now preserves the order of keys in JSON objects passed in input. The new --sort-keys option can be used to sort the keys alphabetically. (Contributed by Berker Peksag in issue 21650.)

JSON decoder now raises json.JSONDecodeError instead of ValueError. (Contributed by Serhiy Storchaka in issue 19361.)

locale¶

A new delocalize() function can be used to convert a string into a normalized number string, taking the LC_NUMERIC settings into account. (Contributed by Cédric Krier in issue 13918.)

logging¶

All logging methods (Logger log(), exception(), critical(), debug(), etc.), now accept exception instances as an exc_info argument, in addition to boolean values and exception tuples. (Contributed by Yury Selivanov in issue 20537.)

The handlers.HTTPHandler class now accepts an optional ssl.SSLContext instance to configure SSL settings used in an HTTP connection. (Contributed by Alex Gaynor in issue 22788.)

The handlers.QueueListener class now takes a respect_handler_level keyword argument which, if set to True, will pass messages to handlers taking handler levels into account. (Contributed by Vinay Sajip.)

lzma¶

The LZMADecompressor.decompress method now accepts an optional max_length argument to limit the maximum size of decompressed data. (Contributed by Martin Panter in issue 15955.)

math¶

Two new constants have been added to the math module: inf and nan. (Contributed by Mark Dickinson in issue 23185.)

A new function isclose() provides a way to test for approximate equality. (Contributed by Chris Barker and Tal Einat in issue 24270.)

A new gcd() function has been added. The fractions.gcd() function is now deprecated. (Contributed by Mark Dickinson and Serhiy Storchaka in issue 22486.)

operator¶

attrgetter(), itemgetter(), and methodcaller() objects now support pickling. (Contributed by Josh Rosenberg and Serhiy Storchaka in issue 22955.)

New matmul() and imatmul() functions to perform matrix multiplication. (Contributed by Benjamin Peterson in issue 21176.)

os¶

The new scandir() function returning an iterator of DirEntry objects has been added. If possible, scandir() extracts file attributes while scanning a directory, removing the need to perform subsequent system calls to determine file type or attributes, which may significantly improve performance. (Contributed by Ben Hoyt with the help of Victor Stinner in issue 22524.)

On Windows, a new stat_result.st_file_attributes attribute is now available. It corresponds to dwFileAttributes member of the BY_HANDLE_FILE_INFORMATION structure returned by GetFileInformationByHandle(). (Contributed by Ben Hoyt in issue 21719.)

The urandom() function now uses getrandom() syscall on Linux 3.17 or newer, and getentropy() on OpenBSD 5.6 and newer, removing the need to use /dev/urandom and avoiding failures due to potential file descriptor exhaustion. (Contributed by Victor Stinner in issue 22181.)

New get_blocking() and set_blocking() functions allow to get and set a file descriptor blocking mode (O_NONBLOCK.) (Contributed by Victor Stinner in issue 22054.)

The truncate() and ftruncate() functions are now supported on Windows. (Contributed by Steve Dower in issue 23668.)

There is a new os.path.commonpath() function returning the longest common sub-path of each passed pathname. Unlike the os.path.commonprefix() function, it always returns a valid path. (Contributed by Rafik Draoui and Serhiy Storchaka in issue 10395.)

pathlib¶

The new Path.samefile method can be used to check whether the path points to the same file as other path, which can be either an another Path object, or a string. (Contributed by Vajrasky Kok and Antoine Pitrou in issue 19775.)

The Path.mkdir method how accepts a new optional exist_ok argument to match mkdir -p and os.makrdirs() functionality. (Contributed by Berker Peksag in issue 21539.)

There is a new Path.expanduser method to expand ~ and ~user prefixes. (Contributed by Serhiy Storchaka and Claudiu Popa in issue 19776.)

A new Path.home class method can be used to get an instance of Path object representing the user’s home directory. (Contributed by Victor Salgado and Mayank Tripathi in issue 19777.)

New Path.write_text, Path.read_text, Path.write_bytes, Path.read_bytes methods to simplify read/write operations on files. (Contributed by Christopher Welborn in issue 20218.)

pickle¶

Nested objects, such as unbound methods or nested classes, can now be pickled using pickle protocols older than protocol version 4. Protocol version 4 already supports these cases. (Contributed by Serhiy Storchaka in issue 23611.)

poplib¶

A new POP3.utf8 command enables RFC 6856 (Internationalized Email) support, if a POP server supports it. (Contributed by Milan OberKirch in issue 21804.)

re¶

The number of capturing groups in regular expression is no longer limited by 100. (Contributed by Serhiy Storchaka in issue 22437.)

The sub() and subn() functions now replace unmatched groups with empty strings instead of raising an exception. (Contributed by Serhiy Storchaka in issue 1519638.)

The re.error exceptions have new attributes: msg, pattern, pos, lineno, and colno that provide better context information about the error. (Contributed by Serhiy Storchaka in issue 22578.)

readline¶

A new append_history_file() function can be used to append the specified number of trailing elements in history to the given file. (Contributed by Bruno Cauet in issue 22940.)

selectors¶

The new DevpollSelector supports efficient /dev/poll polling on Solaris. (Contributed by Giampaolo Rodola’ in issue 18931.)

shutil¶

The move() function now accepts a copy_function argument, allowing, for example, the copy() function to be used instead of the default copy2() if there is a need to ignore file metadata when moving. (Contributed by Claudiu Popa in issue 19840.)

The make_archive() function now supports the xztar format. (Contributed by Serhiy Storchaka in issue 5411.)

signal¶

On Windows, the set_wakeup_fd() function now also supports socket handles. (Contributed by Victor Stinner in issue 22018.)

Various SIG* constants in the signal module have been converted into Enums. This allows meaningful names to be printed during debugging, instead of integer “magic numbers”. (Contributed by Giampaolo Rodola’ in issue 21076.)

smtpd¶

Both SMTPServer and SMTPChannel classes now accept a decode_data keyword argument to determine if the DATA portion of the SMTP transaction is decoded using the "utf-8" codec or is instead provided to the SMTPServer.process_message method as a byte string. The default is True for backward compatibility reasons, but will change to False in Python 3.6. If decode_data is set to False, the process_message() method must be prepared to accept keyword arguments. (Contributed by Maciej Szulik in issue 19662.)

The SMTPServer class now advertises the 8BITMIME extension (RFC 6152) if decode_data has been set True. If the client specifies BODY=8BITMIME on the MAIL command, it is passed to SMTPServer.process_message via the mail_options keyword. (Contributed by Milan Oberkirch and R. David Murray in issue 21795.)

The SMTPServer class now also supports the SMTPUTF8 extension (RFC 6531: Internationalized Email). If the client specified SMTPUTF8 BODY=8BITMIME on the MAIL command, they are passed to SMTPServer.process_message via the mail_options keyword. It is the responsibility of the process_message() method to correctly handle the SMTPUTF8 data. (Contributed by Milan Oberkirch in issue 21725.)

It is now possible to provide, directly or via name resolution, IPv6 addresses in the SMTPServer constructor, and have it successfully connect. (Contributed by Milan Oberkirch in issue 14758.)

smtplib¶

A new SMTP.auth method provides a convenient way to implement custom authentication mechanisms. (Contributed by Milan Oberkirch in issue 15014.)

The SMTP.set_debuglevel method now accepts an additional debuglevel (2), which enables timestamps in debug messages. (Contributed by Gavin Chappell and Maciej Szulik in issue 16914.)

Both SMTP.sendmail and SMTP.send_message methods now support support RFC 6531 (SMTPUTF8). (Contributed by Milan Oberkirch and R. David Murray in issue 22027.)

sndhdr¶

what() and whathdr() functions now return a namedtuple(). (Contributed by Claudiu Popa in issue 18615.)

ssl¶

Memory BIO Support¶

(Contributed by Geert Jansen in issue 21965.)

The new SSLObject class has been added to provide SSL protocol support for cases when the network I/O capabilities of SSLSocket are not necessary or suboptimal. SSLObject represents an SSL protocol instance, but does not implement any network I/O methods, and instead provides a memory buffer interface. The new MemoryBIO class can be used to pass data between Python and an SSL protocol instance.

The memory BIO SSL support is primarily intended to be used in frameworks implementing asynchronous I/O for which SSLSocket‘s readiness model (“select/poll”) is inefficient.

A new SSLContext.wrap_bio method can be used to create a new SSLObject instance.

Application-Layer Protocol Negotiation Support¶

(Contributed by Benjamin Peterson in issue 20188.)

Where OpenSSL support is present, ssl module now implements Application-Layer Protocol Negotiation TLS extension as described in RFC 7301.

The new SSLContext.set_alpn_protocols can be used to specify which protocols a socket should advertise during the TLS handshake.

The new SSLSocket.selected_alpn_protocol returns the protocol that was selected during the TLS handshake. HAS_ALPN flag indicates whether APLN support is present.

Other Changes¶

There is a new SSLSocket.version method to query the actual protocol version in use. (Contributed by Antoine Pitrou in issue 20421.)

The SSLSocket class now implements a SSLSocket.sendfile method. (Contributed by Giampaolo Rodola’ in issue 17552.)

The SSLSocket.send method now raises either ssl.SSLWantReadError or ssl.SSLWantWriteError exception on a non-blocking socket if the operation would block. Previously, it would return 0. (Contributed by Nikolaus Rath in issue 20951.)

The cert_time_to_seconds() function now interprets the input time as UTC and not as local time, per RFC 5280. Additionally, the return value is always an int. (Contributed by Akira Li in issue 19940.)

New SSLObject.shared_ciphers and SSLSocket.shared_ciphers methods return the list of ciphers sent by the client during the handshake. (Contributed by Benjamin Peterson in issue 23186.)

The SSLSocket.do_handshake, SSLSocket.read, SSLSocket.shutdown, and SSLSocket.write methods of ssl.SSLSocket class no longer reset the socket timeout every time bytes are received or sent. The socket timeout is now the maximum total duration of the method. (Contributed by Victor Stinner in issue 23853.)

The match_hostname() function now supports matching of IP addresses. (Contributed by Antoine Pitrou in issue 23239.)

socket¶

Functions with timeouts now use a monotonic clock, instead of a system clock. (Contributed by Victor Stinner in issue 22043.)

A new socket.sendfile method allows to send a file over a socket by using the high-performance os.sendfile() function on UNIX resulting in uploads being from 2 to 3 times faster than when using plain socket.send. (Contributed by Giampaolo Rodola’ in issue 17552.)

The socket.sendall method no longer resets the socket timeout every time bytes are received or sent. The socket timeout is now the maximum total duration to send all data. (Contributed by Victor Stinner in issue 23853.)

The backlog argument of the socket.listen method is now optional. By default it is set to SOMAXCONN or to 128 whichever is less. (Contributed by Charles-François Natali in issue 21455.)

sqlite3¶

The Row class now fully supports sequence protocol, in particular reversed() iteration and slice indexing. (Contributed by Claudiu Popa in issue 10203; by Lucas Sinclair, Jessica McKellar, and Serhiy Storchaka in issue 13583.)

subprocess¶

The new run() function has been added. It runs the specified command and and returns a CompletedProcess object, which describes a finished process. The new API is more consistent and is the recommended approach to invoking subprocesses in Python code that does not need to maintain compatibility with earlier Python versions. (Contributed by Thomas Kluyver in issue 23342.)

sys¶

A new set_coroutine_wrapper() function allows setting a global hook that will be called whenever a coroutine object is created by an async def function. A corresponding get_coroutine_wrapper() can be used to obtain a currently set wrapper. Both functions are provisional, and are intended for debugging purposes only. (Contributed by Yury Selivanov in issue 24017.)

A new is_finalizing() function can be used to check if the Python interpreter is shutting down. (Contributed by Antoine Pitrou in issue 22696.)

sysconfig¶

The name of the user scripts directory on Windows now includes the first two components of Python version. (Contributed by Paul Moore in issue 23437.)

tarfile¶

The mode argument of the open() function now accepts "x" to request exclusive creation. (Contributed by Berker Peksag in issue 21717.)

TarFile.extractall and TarFile.extract methods now take a keyword argument numeric_only. If set to True, the extracted files and directories will be owned by the numeric uid and gid from the tarfile. If set to False (the default, and the behavior in versions prior to 3.5), they will be owned by the named user and group in the tarfile. (Contributed by Michael Vogt and Eric Smith in issue 23193.)

threading¶

Both Lock.acquire and RLock.acquire methods now use a monotonic clock for timeout management. (Contributed by Victor Stinner in issue 22043.)

time¶

The monotonic() function is now always available. (Contributed by Victor Stinner in issue 22043.)

timeit¶

A new command line option -u or --unit=U can be used to specify the time unit for the timer output. Supported options are usec, msec, or sec. (Contributed by Julian Gindi in issue 18983.)

The timeit() function has a new globals parameter for specifying the namespace in which the code will be running. (Contributed by Ben Roberts in issue 2527.)

tkinter¶

The tkinter._fix module used for setting up the Tcl/Tk environment on Windows has been replaced by a private function in the _tkinter module which makes no permanent changes to environment variables. (Contributed by Zachary Ware in issue 20035.)

traceback¶

New walk_stack() and walk_tb() functions to conveniently traverse frame and traceback objects. (Contributed by Robert Collins in issue 17911.)

New lightweight classes: TracebackException, StackSummary, and traceback.FrameSummary. (Contributed by Robert Collins in issue 17911.)

Both print_tb() and print_stack() functions now support negative values for the limit argument. (Contributed by Dmitry Kazakov in issue 22619.)

types¶

A new coroutine() function to transform generator and generator-like objects into awaitables. (Contributed by Yury Selivanov in issue 24017.)

A new CoroutineType is the type of coroutine objects created by async def functions. (Contributed by Yury Selivanov in issue 24400.)

urllib¶

A new request.HTTPPasswordMgrWithPriorAuth class allows HTTP Basic Authentication credentials to be managed so as to eliminate unnecessary 401 response handling, or to unconditionally send credentials on the first request in order to communicate with servers that return a 404 response instead of a 401 if the Authorization header is not sent. (Contributed by Matej Cepl in issue 19494 and Akshit Khurana in issue 7159.)

A new quote_via argument for the parse.urlencode function provides a way to control the encoding of query parts if needed. (Contributed by Samwyse and Arnon Yaari in issue 13866.)

The request.urlopen function accepts an ssl.SSLContext object as a context argument, which will be used for the HTTPS connection. (Contributed by Alex Gaynor in issue 22366.)

The parse.urljoin was updated to use the RFC 3986 semantics for the resolution of relative URLs, rather than RFC 1808 and RFC 2396. (Contributed by Demian Brecht and Senthil Kumaran in issue 22118.)

unicodedata¶

The unicodedata module now uses data from Unicode 8.0.0.

unittest¶

A new command line option --locals to show local variables in tracebacks. (Contributed by Robert Collins in issue 22936.)

unittest.mock¶

The Mock has the following improvements:

Class constructor has a new unsafe parameter, which causes mock objects to raise AttributeError on attribute names starting with "assert". (Contributed by Kushal Das in issue 21238.)
A new Mock.assert_not_called method to check if the mock object was called. (Contributed by Kushal Das in issue 21262.)

The MagicMock class now supports __truediv__(), __divmod__() and __matmul__() operators. (Contributed by Johannes Baiter in issue 20968, and Håkan Lövdahl in issue 23581 and issue 23568.)

wsgiref¶

The headers argument of the headers.Headers class constructor is now optional. (Contributed by Pablo Torres Navarrete and SilentGhost in issue 5800.)

xmlrpc¶

The client.ServerProxy class is now a context manager. (Contributed by Claudiu Popa in issue 20627.)

client.ServerProxy constructor now accepts an optional ssl.SSLContext instance. (Contributed by Alex Gaynor in issue 22960.)

xml.sax¶

SAX parsers now support a character stream of the xmlreader.InputSource object. (Contributed by Serhiy Storchaka in issue 2175.)

zipfile¶

ZIP output can now be written to unseekable streams. (Contributed by Serhiy Storchaka in issue 23252.)

The mode argument of ZipFile.open method now accepts "x" to request exclusive creation. (Contributed by Serhiy Storchaka in issue 21717.)

Other module-level changes¶

Many functions in mmap, ossaudiodev, socket, ssl, and codecs modules now accept writable bytes-like objects. (Contributed by Serhiy Storchaka in issue 23001.)

Optimizations¶

The os.walk() function has been sped up by 3 to 5 times on POSIX systems, and by 7 to 20 times on Windows. This was done using the new os.scandir() function, which exposes file information from the underlying readdir or FindFirstFile/FindNextFile system calls. (Contributed by Ben Hoyt with help from Victor Stinner in issue 23605.)

Construction of bytes(int) (filled by zero bytes) is faster and uses less memory for large objects. calloc() is used instead of malloc() to allocate memory for these objects. (Contributed by Victor Stinner in issue 21233.)

Some operations on ipaddress IPv4Network and IPv6Network have been massively sped up, such as subnets(), supernet(), summarize_address_range(), collapse_addresses(). The speed up can range from 3 to 15 times. (Contributed by Antoine Pitrou, Michel Albert, and Markus in issue 21486, issue 21487, issue 20826, issue 23266.)

Pickling of ipaddress objects was optimized to produce significantly smaller output. (Contributed by Serhiy Storchaka in issue 23133.)

Many operations on io.BytesIO are now 50% to 100% faster. (Contributed by Serhiy Storchaka in issue 15381 and David Wilson in issue 22003.)

The marshal.dumps() function is now faster: 65-85% with versions 3 and 4, 20-25% with versions 0 to 2 on typical data, and up to 5 times in best cases. (Contributed by Serhiy Storchaka in issue 20416 and issue 23344.)

The UTF-32 encoder is now 3 to 7 times faster. (Contributed by Serhiy Storchaka in issue 15027.)

Regular expressions are now parsed up to 10% faster. (Contributed by Serhiy Storchaka in issue 19380.)

The json.dumps() function was optimized to run with ensure_ascii=False as fast as with ensure_ascii=True. (Contributed by Naoki Inada in issue 23206.)

The PyObject_IsInstance() and PyObject_IsSubclass() functions have been sped up in the common case that the second argument has type as its metaclass. (Contributed Georg Brandl by in issue 22540.)

Method caching was slightly improved, yielding up to 5% performance improvement in some benchmarks. (Contributed by Antoine Pitrou in issue 22847.)

Objects from random module now use two times less memory on 64-bit builds. (Contributed by Serhiy Storchaka in issue 23488.)

The property() getter calls are up to 25% faster. (Contributed by Joe Jevnik in issue 23910.)

Instantiation of fractions.Fraction is now up to 30% faster. (Contributed by Stefan Behnel in issue 22464.)

String methods find(), rfind(), split(), partition() and in string operator are now significantly faster for searching 1-character substrings. (Contributed by Serhiy Storchaka in issue 23573.)

Build and C API Changes¶

New calloc functions were added:

PyMem_RawCalloc(),

PyMem_Calloc(),

PyObject_Calloc(),

_PyObject_GC_Calloc().

(Contributed by Victor Stinner in issue 21233.)

New encoding/decoding helper functions:

Py_DecodeLocale() (replaced _Py_char2wchar()),

Py_EncodeLocale() (replaced _Py_wchar2char()).

(Contributed by Victor Stinner in issue 18395.)

A new PyCodec_NameReplaceErrors() function to replace the unicode encode error with \N{...} escapes. (Contributed by Serhiy Storchaka in issue 19676.)

A new PyErr_FormatV() function similar to PyErr_Format(), but accepts a va_list argument. (Contributed by Antoine Pitrou in issue 18711.)

A new PyExc_RecursionError exception. (Contributed by Georg Brandl in issue 19235.)

New PyModule_FromDefAndSpec(), PyModule_FromDefAndSpec2(), and PyModule_ExecDef() introduced by PEP 489 – multi-phase extension module initialization. (Contributed by Petr Viktorin in issue 24268.)

New PyNumber_MatrixMultiply() and PyNumber_InPlaceMatrixMultiply() functions to perform matrix multiplication. (Contributed by Benjamin Peterson in issue 21176. See also PEP 465 for details.)

The PyTypeObject.tp_finalize slot is now part of stable ABI.

Windows builds now require Microsoft Visual C++ 14.0, which is available as part of Visual Studio 2015.

Extension modules now include platform information tag in their filename on some platforms (the tag is optional, and CPython will import extensions without it; although if the tag is present and mismatched, the extension won’t be loaded):

On Linux, extension module filenames end with .cpython-<major><minor>m-<architecture>-<os>.pyd:
- <major> is the major number of the Python version; for Python 3.5 this is 3.
- <minor> is the minor number of the Python version; for Python 3.5 this is 5.
- <architecture> is the hardware architecture the extension module was built to run on. It’s most commonly either i386 for 32-bit Intel platforms or x86_64 for 64-bit Intel (and AMD) platforms.
- <os> is always linux-gnu, except for extensions built to talk to the 32-bit ABI on 64-bit platforms, in which case it is linux-gnu32 (and <architecture> will be x86_64).
On Windows, extension module filenames end with <debug>.cp<major><minor>-<platform>.pyd:
- <major> is the major number of the Python version; for Python 3.5 this is 3.
- <minor> is the minor number of the Python version; for Python 3.5 this is 5.
- <platform> is the platform the extension module was built for, either win32 for Win32, win_amd64 for Win64, win_ia64 for Windows Itanium 64, and win_arm for Windows on ARM.
- If built in debug mode, <debug> will be _d, otherwise it will be blank.
On OS X platforms, extension module filenames now end with -darwin.so.
On all other platforms, extension module filenames are the same as they were with Python 3.4.

Deprecated¶

New Keywords¶

async and await are not recommended to be used as variable, class, function or module names. Introduced by PEP 492 in Python 3.5, they will become proper keywords in Python 3.7.

Unsupported Operating Systems¶

Windows XP is no longer supported by Microsoft, thus, per PEP 11, CPython 3.5 is no longer officially supported on this OS.

Deprecated Python modules, functions and methods¶

The formatter module has now graduated to full deprecation and is still slated for removal in Python 3.6.

The asyncio.async() function is deprecated in favor of ensure_future().

The smtpd module has in the past always decoded the DATA portion of email messages using the utf-8 codec. This can now be controlled by the new decode_data keyword to SMTPServer. The default value is True, but this default is deprecated. Specify the decode_data keyword with an appropriate value to avoid the deprecation warning.

Directly assigning values to the key, value and coded_value of Morsel objects is deprecated. Use the set() method instead. In addition, the undocumented LegalChars parameter of set() is deprecated, and is now ignored.

Passing a format string as keyword argument format_string to the format() method of the string.Formatter class has been deprecated.

The platform.dist() and platform.linux_distribution() functions are now deprecated and will be removed in Python 3.7. Linux distributions use too many different ways of describing themselves, so the functionality is left to a package. (Contributed by Vajrasky Kok and Berker Peksag in issue 1322.)

The previously undocumented from_function and from_builtin methods of inspect.Signature are deprecated. Use new inspect.Signature.from_callable() instead. (Contributed by Yury Selivanov in issue 24248.)

The inspect.getargspec() function is deprecated and scheduled to be removed in Python 3.6. (See issue 20438 for details.)

The inspect getfullargspec(), getargvalues(), getcallargs(), getargvalues(), formatargspec(), and formatargvalues() functions are deprecated in favor of inspect.signature() API. (Contributed by Yury Selivanov in issue 20438.)

Use of re.LOCALE flag with str patterns or re.ASCII is now deprecated. (Contributed by Serhiy Storchaka in issue 22407.)

Removed¶

API and Feature Removals¶

The following obsolete and previously deprecated APIs and features have been removed:

The __version__ attribute has been dropped from the email package. The email code hasn’t been shipped separately from the stdlib for a long time, and the __version__ string was not updated in the last few releases.
The internal Netrc class in the ftplib module was deprecated in 3.4, and has now been removed. (Contributed by Matt Chaput in issue 6623.)
The concept of .pyo files has been removed.
The JoinableQueue class in the provisional asyncio module was deprecated in 3.4.4 and is now removed. (Contributed by A. Jesse Jiryu Davis in issue 23464.)

Porting to Python 3.5¶

This section lists previously described changes and other bugfixes that may require changes to your code.

Changes in the Python API¶

PEP 475: System calls are now retried when interrupted by a signal instead of raising InterruptedError if the Python signal handler does not raise an exception.
Before Python 3.5, a datetime.time object was considered to be false if it represented midnight in UTC. This behavior was considered obscure and error-prone and has been removed in Python 3.5. See issue 13936 for full details.
The ssl.SSLSocket.send() method now raises either ssl.SSLWantReadError or ssl.SSLWantWriteError on a non-blocking socket if the operation would block. Previously, it would return 0. (Contributed by Nikolaus Rath in issue 20951.)
The __name__ attribute of generator is now set from the function name, instead of being set from the code name. Use gen.gi_code.co_name to retrieve the code name. Generators also have a new __qualname__ attribute, the qualified name, which is now used for the representation of a generator (repr(gen)). (Contributed by Victor Stinner in issue 21205.)
The deprecated “strict” mode and argument of HTMLParser, HTMLParser.error(), and the HTMLParserError exception have been removed. (Contributed by Ezio Melotti in issue 15114.) The convert_charrefs argument of HTMLParser is now True by default. (Contributed by Berker Peksag in issue 21047.)
Although it is not formally part of the API, it is worth noting for porting purposes (ie: fixing tests) that error messages that were previously of the form “‘sometype’ does not support the buffer protocol” are now of the form “a bytes-like object is required, not ‘sometype’”. (Contributed by Ezio Melotti in issue 16518.)
If the current directory is set to a directory that no longer exists then FileNotFoundError will no longer be raised and instead find_spec() will return None without caching None in sys.path_importer_cache which is different than the typical case (issue 22834).
HTTP status code and messages from http.client and http.server were refactored into a common HTTPStatus enum. The values in http.client and http.server remain available for backwards compatibility. (Contributed by Demian Brecht in issue 21793.)
When an import loader defines exec_module() it is now expected to also define create_module() (raises a DeprecationWarning now, will be an error in Python 3.6). If the loader inherits from importlib.abc.Loader then there is nothing to do, else simply define create_module() to return None. (Contributed by Brett Cannon in issue 23014.)
The re.split() function always ignored empty pattern matches, so the "x*" pattern worked the same as "x+", and the "\b" pattern never worked. Now re.split() raises a warning if the pattern could match an empty string. For compatibility use patterns that never match an empty string (e.g. "x+" instead of "x*"). Patterns that could only match an empty string (such as "\b") now raise an error.
The Morsel dict-like interface has been made self consistent: morsel comparison now takes the key and value into account, copy() now results in a Morsel instance rather than a dict, and update() will now raise an exception if any of the keys in the update dictionary are invalid. In addition, the undocumented LegalChars parameter of set() is deprecated and is now ignored. (Contributed by Demian Brecht in issue 2211.)
PEP 488 has removed .pyo files from Python and introduced the optional opt- tag in .pyc file names. The importlib.util.cache_from_source() has gained an optimization parameter to help control the opt- tag. Because of this, the debug_override parameter of the function is now deprecated. .pyo files are also no longer supported as a file argument to the Python interpreter and thus serve no purpose when distributed on their own (i.e. sourcless code distribution). Due to the fact that the magic number for bytecode has changed in Python 3.5, all old .pyo files from previous versions of Python are invalid regardless of this PEP.
The socket module now exports the CAN_RAW_FD_FRAMES constant on linux 3.6 and greater.
The cert_time_to_seconds() function now interprets the input time as UTC and not as local time, per RFC 5280. Additionally, the return value is always an int. (Contributed by Akira Li in issue 19940.)
The pygettext.py Tool now uses the standard +NNNN format for timezones in the POT-Creation-Date header.
The smtplib module now uses sys.stderr instead of previous module level stderr variable for debug output. If your (test) program depends on patching the module level variable to capture the debug output, you will need to update it to capture sys.stderr instead.
The str.startswith() and str.endswith() methods no longer return True when finding the empty string and the indexes are completely out of range. (Contributed by Serhiy Storchaka in issue 24284.)
The inspect.getdoc() function now returns documentation strings inherited from base classes. Documentation strings no longer need to be duplicated if the inherited documentation is appropriate. To suppress an inherited string, an empty string must be specified (or the documentation may be filled in). This change affects the output of the pydoc module and the help() function. (Contributed by Serhiy Storchaka in issue 15582.)

Changes in the C API¶

The undocumented format member of the (non-public) PyMemoryViewObject structure has been removed. All extensions relying on the relevant parts in memoryobject.h must be rebuilt.
The PyMemAllocator structure was renamed to PyMemAllocatorEx and a new calloc field was added.
Removed non-documented macro PyObject_REPR which leaked references. Use format character %R in PyUnicode_FromFormat()-like functions to format the repr() of the object. (Contributed by Serhiy Storchaka in issue 22453.)
Because the lack of the __module__ attribute breaks pickling and introspection, a deprecation warning now is raised for builtin type without the __module__ attribute. Would be an AttributeError in future. (Contributed by Serhiy Storchaka in issue 20204.)
As part of PEP 492 implementation, tp_reserved slot of PyTypeObject was replaced with a tp_as_async slot. Refer to Coroutine Objects for new types, structures and functions.