原文：Python 3.6新特性简介

Python 3.6新特性简介¶

版本:	3.6.0
日期:	December 15, 2016
编辑:	Elvis Pranskevichus
<elvis@magic.io>, Yury Selivanov
<yury@magic.io>

这篇文章介绍了与3.5相比， Python 3.6中多出的新特性。

另请参阅

PEP 494 - Python 3.6 发布时间表

摘要 - 发布亮点¶

新的语法特性：

PEP 498, 格式化字符串字面量
PEP 515, 数字字面量中的下划线
PEP 526, 变量注解中的语法
PEP 525, 异步生成器
PEP 520: 异步解析式

新的库模块

secrets: PEP 506 - 在标准库中添加了Secrets模块

CPython实现的改进：

重新实现了字典(dict)类型，以便能像PyPy的字典类型一样使用更紧凑的表达方式。与Python 3.5相比，这使字典的内存用量减少了20%到25%。
用新协定优化了类的自定义建立。
类属性定义顺序(class attribute definition order)现在被保留了
**kwargs内的元素顺序现在对应于将关键字(保留字）参数传递给函数的顺序
新增了对DTrace和SystemTap probing的支持。
新PYTHONMALLOC环境变量现在可用于调试解释器内存分配与访问错误。

标准库的重大改进：

为asyncio模块开发了新功能、显著的可用性、性能优化，以及大量的错误修复。从Python 3.6开始，asyncio模块不再是临时的了，其API也进入了稳定状态。
实现了用于支持类路径对象(path-like objects)的新文件系统路径协议。所有在路径(path)上使用的标准库函数都已更新，以便适应于新协议。
datetime模块已获得对本地时间消歧(Local Time Disambiguation)的支持。
针对typing模块的一些改进，使其不再是临时模块。
tracemalloc模块已重大改进，现用于为ResourceWarning提供更好的输出，并为内存分配错误提供更好的诊断。欲知详情，请参阅PYTHONMALLOC部分。

安全相关的改进:

新secrets模块被用于简化那些适用于管理密文的密码学安全伪随机数生成器(cryptographically strong pseudo-random numbers)的生成过程，如认证、token等。
在Linux上，现将os.urandom()改成了阻塞模式，直到系统的urandom的熵池(entropy pool)的初始化具有更高的安全性。解释请参见PEP 524。
hashlib和ssl模块现已支持OpenSSL 1.1.0。
改进了ssl模块的默认设置和特性集。
新增了hashlib模块对BLAKE2、SHA-3、SHAKE哈希算法以及scrypt()密钥导出函数的支持。

Windows上的改进:

PEP 528与PEP 529，Windows文件系统和控制台的编码已更改为UTF-8。
当用户没有指定版本（通过命令行参数或配置文件）时，py.exe启动器以交互方式使用时，不再以Python 2优先于Python 3。处理shebang行的方式保持不变 - 此处的“python”依旧指Python 2。
python.exe和pythonw.exe已标记为长路径敏感(long-path aware)，这意味着260字符路径限制可能不再适用。有关详细信息，请参阅删除MAX_PATH限制。
可以添加._pth文件以强制隔离模式(isolated mode)并完全指定所有搜索路径，以避免注册表查找和环境查找。有关详细信息，请参阅文档。
一个python36.zip文件现可用作一个地标(landmark)以臆指PYTHONHOME。有关详细信息，请参阅文档。

新特性

PEP 498: 格式化字符串

PEP 498引入了一种新的字符串：_f-strings_, 或者格式化字符串。

格式化字符串带'f'前缀，类似于str.format()接受的格式字符串。它们包含了由花括号括起来的替换字段。替换字段是表达式，它们会在运行时计算，然后使用format()协议进行格式化：

    >>> name = "Fred"
    >>> f"He said his name is {name}."
    'He said his name is Fred.'
    >>> width = 10
    >>> precision = 4
    >>> value = decimal.Decimal("12.34567")
    >>> f"result: {value:{width}.{precision}}"  # nested fields
    'result:      12.35'

又见

PEP 498 - 字符串插值。

PEP由Eric V. Smith编写和实现。

特性文档。

PEP 526: 变量注释语法

PEP 484引入了函数参数的类型注释的标准，又名类型提示。这个PEP添加了用来注释变量（包括类变量和实例变量）类型的语法：

    primes: List[int] = []

    captain: str  # Note: no initial value!

    class Starship:
        stats: Dict[str, int] = {}

正如函数注释，Python解释器不附加任何特殊意义到变量注释上，只是将它们存储在一个类或者模块的__annotations__属性中。

与静态类型语言中的变量声明相比，注释语法的目的在于提供一种简单的方式，通过抽象语法树和__annotations__属性，来为第三方工具和库指定结构化类型元数据。

又见

PEP 526 - 变量注释语法。

PEP由Ryan Gonzalez, Philip House, Ivan Levkivskyi, Lisa Roach, 和Guido van Rossum编写。由Ivan Levkivskyi实现。

使用或将要使用这个新语法的工具：mypy, pytype, PyCharm等等。

PEP 515: 数值文字中的下划线

PEP 515添加了在数值文字中使用下划线的能力，以提高可读性。例如：

    >>> 1_000_000_000_000_000
    1000000000000000
    >>> 0x_FF_FF_FF_FF
    4294967295

数字之间和任何基本符号之后允许单个下划线。不允许前置、后置或者多个连续的下划线。

字符串格式化语言现在还支持'_'选项，该选项用来通知对浮点表示类型和整型表示类型'd'，会把下划线当成千位分隔符使用。对于整型表示类型'b', 'o', 'x', 和'X', 下划线将会被插入到每4个数字之间：

    >>> '{:_}'.format(1000000)
    '1_000_000'
    >>> '{:_x}'.format(0xFFFFFFFF)
    'ffff_ffff'

又见

PEP 515 - 数值文字中的下划线

PEP由Georg Brandl和Serhiy Storchaka编写。

PEP 525: 异步生成器¶

PEP 492 引入支持原生协程和async /await的语法到Python 3.5。在Python 3.5实现里的一个值得注意的局限性就在于它不可能使用await和`yield'在同一个函数体中。而在Python 3.6中，这个限制已解除，这使得定义_异步生成器_成为可能：

    async def ticker(delay, to):
        """Yield numbers from 0 to *to* every *delay* seconds."""
        for i in range(to):
            yield i
            await asyncio.sleep(delay)

新的语法允许更快更简洁的代码。

参见 PEP 525 - 异步生成器

由Yury Selivanov撰写并实现的PEP。

PEP 530: 异步解析式¶

PEP 530 添加了对async for在list、set、dict解析式以及generator表达式中的使用支持：

    result = [i async for i in aiter() if i % 2]

此外，所有解析式都支持“await”表达式：

    result = [await fun() for fun in funcs if await condition()]

参见 PEP 530 - 异步解析式

由Yury Selivanov撰写并实现的PEP。

PEP 487: 用于建立类的更简单的自定义¶

现在可以在不使用元类的情况下自定义子类。每当创建一个新的子类时，新的__init_subclass__类方法将在基类上被调用，：

    class PluginBase:
        subclasses = []

        def __init_subclass__(cls, **kwargs):
            super().__init_subclass__(**kwargs)
            cls.subclasses.append(cls)

    class Plugin1(PluginBase):
        pass

    class Plugin2(PluginBase):
        pass

为了允许零参数 [super（）] 42 从[_init_subclass __（）](https://docs.python.org/3 .6/reference/datamodel.html#object.__init_subclass__ "object.__init_subclass__" )实现中被正确的调用并工作，自定义元类必须确保新的__classcell__命名空间输入传递到type .__ new__ （如[创建类对象](https://docs.python.org/3.6/reference/datamodel.html#class-object- creation))

参见

PEP 487 - 用于建立类的更简单的自定义

由Martin Teichmann撰写并实现的PEP。

功能文档

PEP 487: 描述符协议增强¶

PEP 487 扩展描述符协议必须包括新的可选的__set_name __（）方法。每当定义一个新类时，新方法将会调用定义中所有的描述符，并给它们提供定义类的引用，以及类命名空间中给予描述符的名字。换句话说，描述符的实例现在可以获知所有者类的属性名：

    class IntField:
        def __get__(self, instance, owner):
            return instance.__dict__[self.name]

        def __set__(self, instance, value):
            if not isinstance(value, int):
                raise ValueError(f'expecting integer in {self.name}')
            instance.__dict__[self.name] = value

        # this is the new initializer:
        def __set_name__(self, owner, name):
            self.name = name

    class Model:
        int_field = IntField()

参见

PEP 487 - 用于建立类的更简单的自定义

由Yury Selivanov撰写并实现的PEP。

[功能文档](https://docs.python.org/3.6/reference/datamodel.html#d escriptors)

PEP 519: 添加文件系统路径协议¶

文件系统路径过去被表示为str或bytes对象。这会导致那些编写操作文件系统路径代码的人，假定这些对象只能是这两种类型之一(一个代表着文件描述符的int对象将不被计入即它不是一个文件路径)。不幸的是，这种假设局限了文件系统路径表示代方法，如已经存在的pathlib，同时也包括python的一些标准库。为了解决这种情况，定义了一个由os.PathLike表示的新接口。通过实现__fspath__()方法，一个对象表示一个路径，然后，可以将文件系统路径表示为一个较低等级的str或者bytes对象。这意味着，如果一个对象实现os.PathLike或者是str或bytes，该对象被认为是path- like,它代表一个文件系统路径。你可以使用os.fspath(),os.fsdecode()或os.fsencode()显式获取str以及/或bytes来表示一个path-like对象。内建函数open()已经更新，可以接受os.PathLike对象，以及在[os][54]和os.path模块中的所有函数，以及标准库中的大多数其他函数和类。类os.DirEntry以及pathlib中相关的类也已经可以实现os.PathLike。希望对操作文件系统路径基本功能的更新能够让第三方代码在不改变任何代码，或者至少是非常少的代码（例如，在操作path-like对象之前，在代码的开头调用os.fspath()）的情况下，能够隐含地支持所有path-like objects对象。下面举一些例子说明新接口是如何让预先存在的代码简单透明地使用pathlib.Path:

    >>> import pathlib
    >>> with open(pathlib.Path("README")) as f:
    ...     contents = f.read()
    ...
    >>> import os.path
    >>> os.path.splitext(pathlib.Path("some_file.txt"))
    ('some_file', '.txt')
    >>> os.path.join("/a/b", pathlib.Path("c"))
    '/a/b/c'
    >>> import os
    >>> os.fspath(pathlib.Path("some_file.txt"))
    'some_file.txt'

（由Brett Cannon, Ethan Furman, Dusty Phillips, and Jelle Zijlstra实现）

参见

[PEP 519][55] - 添加文件系统路径协议

PEP 由Brett Cannon 和 Koos Zevenhoven撰写.

PEP 495: Local Time Disambiguation¶

In most world locations, there have been and will be times when local clocks are moved back. In those times, intervals are introduced in which local clocks show the same time twice in the same day. In these situations, the information displayed on a local clock (or stored in a Python datetime instance) is insufficient to identify a particular moment in time.

[PEP 495][56] adds the new fold attribute to instances of [datetime.datetime](https://docs.python.org/3.6/li brary/datetime.html#datetime.datetime "datetime.datetime" ) and `datetime.tim e` classes to differentiate between two moments in time for which local times are the same:

    >>> u0 = datetime(2016, 11, 6, 4, tzinfo=timezone.utc)
    >>> for i in range(4):
    ...     u = u0 + i*HOUR
    ...     t = u.astimezone(Eastern)
    ...     print(u.time(), 'UTC =', t.time(), t.tzname(), t.fold)
    ...
    04:00:00 UTC = 00:00:00 EDT 0
    05:00:00 UTC = 01:00:00 EDT 0
    06:00:00 UTC = 01:00:00 EST 1
    07:00:00 UTC = 02:00:00 EST 0

The values of the [fold](https://docs.python.org/3.6/library/datetime.html#d atetime.datetime.fold "datetime.datetime.fold" ) attribute have the value 0 for all instances except those that represent the second (chronologically) moment in time in an ambiguous case.

PEP 529: 更改windows下文件系统编码格式为UTF-8¶

使用str (Unicode) 表示文件系统路径比bytes能获得更好的效果。尽管如此，在某些情况下bytes也足以胜任并且也是正确的。

在3.6之前,使用bytes路径可能导致数据丢失。改进后, windows下现在支持使用bytes表示路径,这些bytes将以sys.getfilesy stemencoding()的方式编码，默认编码格式为'utf-8'。

不使用str方式表示路径的应用程序应当使用os.fsencode()和os.fsdecode() 以确保他们的bytes被正确编码。要回复到之前的状态, 设置 PYTHONLEGACYWINDOWSFSENCODING 或者调用 sys._enablelegacywindowsfsencoding()。

查看 PEP 529以获取更多信息并讨论可能需要变更的代码。

PEP 528: 更改windows控制台编码为UTF-8¶

windows下的默认控制台现在支持所有的Unicode字符并可以正确读取Python代码中的str对象。 sys.stdin, sys.stdout 以及 sys.stderr 现在的默认使用utf-8编码。

这一变化仅适用于使用交互控制台之时，而非重定向文件或者管道。如果要使用之前的交互控制台, 需设置 PYTHONLEGACYWINDOWSIOENCODING。

另请参阅

PEP 528 - 修改windows控制台编码为UTF-8

PEP 由Steve Dower编写和实现。

PEP 520: 保存类属性定义顺序¶

类的定义体中的属性有一个自然顺序：即源码中属性名出现的顺序。这个顺序现在保存在新的类__dict__ 的属性中.

同样，有效的缺省类和执行空间 (从type.prepare()返回)是一个保存插入顺序的映射。

另请参阅

PEP 520 - 保存类属性定义顺序

该PEP由Eric Snow编写和实现。

PEP 468: 保存关键字参数顺序¶

函数声明中的**kwargs 的顺序现在被保证是插入顺序的映射。

另请参阅

PEP 468 - 保存关键字参数顺序

该PEP由Eric Snow编写和实现。

新的字典dict类型的实现¶

字典dict类型现在使用 PyPy首创的 "紧凑" 表达方式。新字典dict() 的内存占用比Python3.5中减少20%到25%。

新的实现中保存顺序的功能被认为是不可过于依赖的(未来也许会改变，不过在将所有当前和未来的Python实现的语言规范转换为保证顺序的语法之前的几个版本中，新的dict有望被实现的; 这也能帮助保证对那些仍旧是随机迭代顺序的旧版本的向后兼容，比如Python 3.5)。

(由INADA Naoki在issue 27350提供。想法最初由Raymond Hettinger提出.)

PEP 523: Adding a frame evaluation API to CPython¶

While Python provides extensive support to customize how code executes, one place it has not done so is in the evaluation of frame objects. If you wanted some way to intercept frame evaluation in Python there really wasn't any way without directly manipulating function pointers for defined functions.

[PEP 523][65] changes this by providing an API to make frame evaluation pluggable at the C level. This will allow for tools such as debuggers and JITs to intercept frame evaluation before the execution of Python code begins. This enables the use of alternative evaluation implementations for Python code, tracking frame evaluation, etc.

This API is not part of the limited C API and is marked as private to signal that usage of this API is expected to be limited and only applicable to very select, low-level use-cases. Semantics of the API will change with Python as necessary.

PYTHONMALLOC environment variable¶

The new [PYTHONMALLOC](https://docs.python.org/3.6/using/cmdline.html

envvar-PYTHONMALLOC) environment variable allows setting the Python memory

allocators and installing debug hooks.

It is now possible to install debug hooks on Python memory allocators on Python compiled in release mode using PYTHONMALLOC=debug. Effects of debug hooks:

Newly allocated memory is filled with the byte 0xCB
Freed memory is filled with the byte 0xDB
Detect violations of the Python memory allocator API. For example, PyObject_Free() called on a memory block allocated by [PyMem_Malloc()][67].
Detect writes before the start of a buffer (buffer underflows)
Detect writes after the end of a buffer (buffer overflows)
Check that the [GIL][68] is held when allocator functions of [PYMEM_DOMAIN_OBJ][69] (ex: PyObject_Malloc()) and [PYMEM_DOMAIN_MEM][70] (ex: [PyMem_Malloc()][71]) domains are called.

Checking if the GIL is held is also a new feature of Python 3.6.

See the [PyMem_SetupDebugHooks()](https://docs.python.org/3.6/c-api/memory.h tml#c.PyMem_SetupDebugHooks "PyMem_SetupDebugHooks" ) function for debug hooks on Python memory allocators.

It is now also possible to force the usage of the malloc() allocator of the C library for all Python memory allocations using PYTHONMALLOC=malloc. This is helpful when using external memory debuggers like Valgrind on a Python compiled in release mode.

On error, the debug hooks on Python memory allocators now use the [tracemalloc](https://docs.python.org/3.6/library/tracemalloc.html#module- tracemalloc "tracemalloc: Trace memory allocations." ) module to get the traceback where a memory block was allocated.

Example of fatal error on buffer overflow using python3.6 -X tracemalloc=5 (store 5 frames in traces):

    Debug memory block at address p=0x7fbcd41666f8: API 'o'
        4 bytes originally requested
        The 7 pad bytes at p-7 are FORBIDDENBYTE, as expected.
        The 8 pad bytes at tail=0x7fbcd41666fc are not all FORBIDDENBYTE (0xfb):
            at tail+0: 0x02 *** OUCH
            at tail+1: 0xfb
            at tail+2: 0xfb
            at tail+3: 0xfb
            at tail+4: 0xfb
            at tail+5: 0xfb
            at tail+6: 0xfb
            at tail+7: 0xfb
        The block was made by call #1233329 to debug malloc/realloc.
        Data at p: 1a 2b 30 00

    Memory block allocated at (most recent call first):
      File "test/test_bytes.py", line 323
      File "unittest/case.py", line 600
      File "unittest/case.py", line 648
      File "unittest/suite.py", line 122
      File "unittest/suite.py", line 84

    Fatal Python error: bad trailing pad byte

    Current thread 0x00007fbcdbd32700 (most recent call first):
      File "test/test_bytes.py", line 323 in test_hex
      File "unittest/case.py", line 600 in run
      File "unittest/case.py", line 648 in __call__
      File "unittest/suite.py", line 122 in run
      File "unittest/suite.py", line 84 in __call__
      File "unittest/suite.py", line 122 in run
      File "unittest/suite.py", line 84 in __call__
      ...

(Contributed by Victor Stinner in issue 26516 and issue 26564.)

DTrace and SystemTap probing support¶

Python can now be built --with-dtrace which enables static markers for the following events in the interpreter:

function call/return
garbage collection started/finished
line of code executed.

This can be used to instrument running interpreters in production, without the need to recompile specific debug builds or providing application-specific profiling/debugging code.

More details in [Instrumenting CPython with DTrace and SystemTap](https://docs .python.org/3.6/howto/instrumentation.html#instrumentation).

The current implementation is tested on Linux and macOS. Additional markers may be added in the future.

(Contributed by Łukasz Langa in issue 21590, based on patches by Jesús Cea Avión, David Malcolm, and Nikhil Benesch.)

其他语言方面的变化¶

我们还对 Python 语言的核心做了一些小的改变：

现在在同一作用域内的 global 或 nonlocal 语句必须在受其影响的命名使用前显式出现。在先前版本中会发出 语法警告。
现在能以将 [魔术方法][72] 设置为 None 的方式来表示相应的操作不可用。例如，如果将一个类的 [__iter__()][73] 设置为None，则该类不可迭代。(由 Andrew Barnert 和 Ivan Levkivskyi 在 [issue 25958][74] 中贡献。)
在 traceback 中大量重复出现的 traceback lines 将会被略写成 "[前一行已经重复出现 {count} 次了]" (以 traceback 为例。). (由 Emanuel Barry 在 [issue 26823][75] 中贡献。)
现在模块导入在找不到模块时引发新的异常 [ModuleNotFoundError][76]（[ImportError][77] 的子类）。当前检测 ImportError 的代码（try-except中）仍然可以工作。(由 Eric Snow 在 [issue 15767][78] 中贡献。)
在类创建时，被元类调用的依赖无参数的 super() 的类方法时会正确工作。(由 Martin Teichmann 在 [issue 23722][79]中贡献。)

新增模块¶

secrets¶

新增模块 secrets 的主要目的是：提供一个显式可靠的方式来产生适合于管理诸如账户认证，令牌之类的保密信息的加密的强伪随机值。

警告

注意在 random 模块中的随机数生成器不该用于安全目的。Python 3.6 及以上版本请使用 secrets 而 Python 3.5 及更早的版本请使用 os.urandom()。

又见

[PEP 506][80] - 向标准库添加 Secrets 模块

PEP 由 Steven D'Aprano 编写及实现。

Improved Modules¶

array¶

耗尽的迭代器array.array现在将会保持被耗尽，即使被迭代的数组被拓展了。这与其他可变序列的行为保持了一致性。

由Serhiy Storchaka在issue 26492中贡献

ast¶

添加了新的ast.Constant的AST节点。它可以被外部的AST优化器出于永久聚合(constant foldind)。

由Victor Stinner在issue 26146中贡献

asyncio¶

始于Python 3.6，asyncio模块将不再是临时的了，并且它的API被认为是稳定的。

从Python 3.5开始，在asyncio模块中值得注意的变化（由于临时的状态，所有的修改都应用到了3.5.x）:

get_event_loop()函数已经被改变为，当从协同程序中调用和被回调调用时，总是返回当前正在运行的循环。（由 Yury Selivanovgong贡献于issue 28613。）
ensure_future()函数和所有使用其的函数，比如loop.run_until_complete()，现在接收所有种类的awaitable objects。（由 Yury Selivanov贡献）。
新的run_coroutine_threadsafe()函数去从其他线程中提交协同程序到事件循环中。（由Vincent Michel贡献）
新的Transport.is_closing()方法去检查是否传输是关闭的或被关闭了（由Yury Selivanov贡献）。
loop.creat_server()现在可以以列表接收主机了。（由Yann Sionneaugong贡献）
新的loop.creat_future()方法去创建一个Future对象。这允许了替代事件循环的实现，比如uvloop，去提供了一种更快的asyncio.Future的实现。（由Yury Selivanov在issue 27041中贡献）
新的loop.get_exception_handler()方法去得到一个当时的异常的句柄。（由Yury Selivanov在issue 27040中贡献）
新的StreamReader.readuntil()方法从流中读取数据读取数据直到出现分隔符字符序列。（由Mark Korenberg贡献）
StreamReader.readexactly()的表现被提升了。（由Mark Korenberg在issue 28370中贡献）
loop.getaddrinfo()方法被优化了，避免了如果地址问题已经解决，去调用系统函数getaddrinfo。（由A. Jesse Jiryu Davis贡献）
loop.stop()方法已经被更改，以便在当前迭代后立即停止循环。任何新的，作为最后一次迭代的结果的预定的回调将会被丢弃。（由Guido van Rossum在issue 25593中贡献）
Future.set_exception在通过了一个StopIteration实例的异常后，将会引发TypeError。（由Chris Angelico在issue 26221贡献）
新的loop.connect_accepted_socket()方法被服务器用来接收asyncio外部的连接，但是使用asyncio去处理。（由Jim Fulton在issue 27392中贡献）
TCP_NODELAY标志位现在被默认设置为TCP传输。（由Yury Selivanov在issue 27456中贡献）
新的loop.shutdown_asyncgens()在关闭循环之前恰当的关闭挂起的异步生成器。（由 Yury Selivanov在issue 28003中贡献）
Future和Task类现在有一个已优化的C语言的实现，使得asyncio编码速度提高了30%。（由Yury Selivanov和INADA Naoki在issue 26081和issue 28544中贡献）

binascii¶

b2a_base64()功能现在接收现在接收一个可选_newline_的关键词参数去控制是否新的一行的字符被加到返回值中。（由Victor Stinnergong贡献于issue25357。）

cmath¶

新增了方法 cmath.tau 来表示(τ) 常量. (由Lisa Roach在[issue 12345][107]中贡献, 详情请查看 **PEP 628** .)

新增的常量还有: cmath.inf 和 cmath.nan 来对应 math.inf 和 math.nan, 另外还有 cmath.infj 和 cmath.nanj 来匹配复数表达式的格式化. (由Mark Dickinson在[issue 23229][108]中贡献.)

collections¶

新的[Collection](https://docs.python.org/3.6/library/collections.abc.htm l#collections.abc.Collection "collections.abc.Collection" ) 抽象基类已经增加到可表示大小的迭代器类里面 (由Ivan Levkivskyi贡献, 文档由Neil Girdhar在issue 27598中贡献.)

新的[Reversible](https://docs.python.org/3.6/library/collections.abc.htm l#collections.abc.Reversible "collections.abc.Reversible" ) 有迭代器类的抽象基类 [__reversed__()](htt ps://docs.python.org/3.6/reference/datamodel.html#object.__reversed__ "object.__reversed__" ) 方法. (由Ivan Levkivskyi在issue 25987中贡献.)

新的 [AsyncGenerator](https://docs.python.org/3.6/library/collections.abc .html#collections.abc.AsyncGenerator "collections.abc.AsyncGenerator" ) ，是一种异步生成器类的抽象基类. (Contributed by Yury Selivanov in [issue 28720][109].)

这个[namedtuple()](https://docs.python.org/3.6/library/collections.html#coll ections.namedtuple "collections.namedtuple" ) 函数现在接受一个可选的关键字参数 module, 它指定用于__module__返回的命名元组类的属性. (由Raymond Hettinger 在[issue 17941][110]中贡献.

namedtuple()的_verbose_ 和 rename 参数是强制关键字参数. (由Raymond Hettinger 在[issue 25628][111]中贡献.)

递归方法[collections.deque](https://docs.python.org/3.6/library/collection s.html#collections.deque "collections.deque" ) 实例现在可使用pickle持久化存储. (由Serhiy Storchaka在issue 26482中贡献.)

concurrent.futures¶

The [ThreadPoolExecutor](https://docs.python.org/3.6/library/concurrent.futu res.html#concurrent.futures.ThreadPoolExecutor "concurrent.futures.ThreadPoolExecutor" ) class constructor now accepts an optional thread_name_prefix argument to make it possible to customize the names of the threads created by the pool. (Contributed by Gregory P. Smith in [issue 27664][112].)

contextlib¶

The [contextlib.AbstractContextManager](https://docs.python.org/3.6/library/ contextlib.html#contextlib.AbstractContextManager "contextlib.AbstractContextManager" ) class has been added to provide an abstract base class for context managers. It provides a sensible default implementation for enter() which returns self and leaves exit() an abstract method. A matching class has been added to the typing module as [`typing.Context Manager`](https://docs.python.org/3.6/library/typing.html#typing.ContextManage r "typing.ContextManager" ). (Contributed by Brett Cannon in issue 25609.)

datetime¶

The [datetime](https://docs.python.org/3.6/library/datetime.html#datetime.da tetime "datetime.datetime" ) and time classes have the new fold attribute used to disambiguate local time when necessary. Many functions in the datetime have been updated to support local time disambiguation. See Local Time Disambiguation section for more information. (Contributed by Alexander Belopolsky in issue 24773.)

The [datetime.strftime()](https://docs.python.org/3.6/library/datetime.html\# datetime.datetime.strftime "datetime.datetime.strftime" ) and `date.strftime( )` methods now support ISO 8601 date directives %G, %u and %V. (Contributed by Ashley Anderson in issue 12006.)

The [datetime.isoformat()](https://docs.python.org/3.6/library/datetime.html

datetime.datetime.isoformat "datetime.datetime.isoformat" ) function now

accepts an optional timespec argument that specifies the number of additional components of the time value to include. (Contributed by Alessandro Cucci and Alexander Belopolsky in issue 19475.)

The [datetime.combine()](https://docs.python.org/3.6/library/datetime.html#d atetime.datetime.combine "datetime.datetime.combine" ) now accepts an optional tzinfo argument. (Contributed by Alexander Belopolsky in issue 27661.)

decimal¶

New [Decimal.as_integer_ratio()](https://docs.python.org/3.6/library/decimal .html#decimal.Decimal.as_integer_ratio "decimal.Decimal.as_integer_ratio" ) method that returns a pair (n, d) of integers that represent the given Decimal instance as a fraction, in lowest terms and with a positive denominator:

    >>> Decimal('-3.14').as_integer_ratio()
    (-157, 50)

(Contributed by Stefan Krah amd Mark Dickinson in issue 25928.)

distutils¶

The default_format attribute has been removed from distutils.command.sdist.sdist and the formats attribute defaults to ['gztar']. Although not anticipated, any code relying on the presence of default_format may need to be adapted. See issue 27819 for more details.

email¶

The new email API, enabled via the policy keyword to various constructors, is no longer provisional. The email documentation has been reorganized and rewritten to focus on the new API, while retaining the old documentation for the legacy API. (Contributed by R. David Murray in [issue 24277][113].)

The [email.mime](https://docs.python.org/3.6/library/email.mime.html#module- email.mime "email.mime: Build MIME messages." ) classes now all accept an optional policy keyword. (Contributed by Berker Peksag in issue 27331.)

The [DecodedGenerator](https://docs.python.org/3.6/library/email.generator.h tml#email.generator.DecodedGenerator "email.generator.DecodedGenerator" ) now supports the policy keyword.

There is a new [policy](https://docs.python.org/3.6/library/email.policy.html#module- email.policy "email.policy: Controlling the parsing and generating of messages" ) attribute, [message_factory](https://docs.python.org/3.6/library /email.policy.html#email.policy.Policy.message_factory "email.policy.Policy.message_factory" ), that controls what class is used by default when the parser creates new message objects. For the [`email.policy.co mpat32`](https://docs.python.org/3.6/library/email.policy.html#email.policy.co mpat32 "email.policy.compat32" ) policy this is [Message](https://docs.pytho n.org/3.6/library/email.compat32-message.html#email.message.Message "email.message.Message" ), for the new policies it is [EmailMessage](https:/ /docs.python.org/3.6/library/email.message.html#email.message.EmailMessage "email.message.EmailMessage" ). (Contributed by R. David Murray in issue 20476.)

encodings¶

On Windows, added the 'oem' encoding to use CP_OEMCP, and the 'ansi' alias for the existing 'mbcs' encoding, which uses the CP_ACP code page. (Contributed by Steve Dower in issue 27959.)

enum¶

Two new enumeration base classes have been added to the [enum][114] module: [Flag][115] and IntFlags. Both are used to define constants that can be combined using the bitwise operators. (Contributed by Ethan Furman in issue 23591.)

Many standard library modules have been updated to use the IntFlags class for their constants.

The new enum.auto value can be used to assign values to enum members automatically:

    >>> from enum import Enum, auto
    >>> class Color(Enum):
    ...     red = auto()
    ...     blue = auto()
    ...     green = auto()
    ...
    >>> list(Color)
    [<Color.red: 1>, <Color.blue: 2>, <Color.green: 3>]

faulthandler¶

在Windows平台, [faulthandler](https://docs.python.org/3.6/library/faulthandler.html#module- faulthandler "faulthandler: Dump the Python traceback." )模块安装了一个指示Windows异常的句柄 : 详情可见于 [faulthandler.enable()](https://docs.p ython.org/3.6/library/faulthandler.html#faulthandler.enable "faulthandler.enable" ). (由 Victor Stinner 在 issue 23848中贡献.)

fileinput¶

[hook_encoded()](https://docs.python.org/3.6/library/fileinput.html#fileinpu t.hook_encoded "fileinput.hook_encoded" ) 模块现在可支持errors参数. (由 Joseph Hackman 在 issue 25788贡献.)

hashlib¶

[hashlib](https://docs.python.org/3.6/library/hashlib-blake2.html#module- hashlib "hashlib: BLAKE2 hash function for Python" ) 支持 OpenSSL 1.1.0。推荐使用的最低版本为1.0.2. (由Christian Heimes 在 [issue 26470][116]贡献.)

BLAKE2 hash 函数也被收录进这一模块. [blake2b()](https://docs.python.org/3.6/library/hashlib- blake2.html#hashlib.blake2b "hashlib.blake2b" ) 与 [blake2s()](https://docs.python.org/3.6/library/hashlib- blake2.html#hashlib.blake2s "hashlib.blake2s" ) 将长期支持BLAKE2的所有特性. (由Christian Heimes in 依据Dmitry Chestnykh 和 Samuel Neves的代码在[issue 26798][117] 贡献. 文档由Dmitry Chestnykh撰写.)

新增SHA-3哈希函数 sha3_224(), sha3_256(), sha3_384(), sha3_512(), 与SHAKE 哈希函数 shake_128() 、shake_256() . (由Christian Heimes 在 issue 16113贡献. Keccak 代码包由Guido Bertoni, Joan Daemen, Michaël Peeters, Gilles Van Assche, and Ronny Van Keer撰写.)

基于密码的密钥导出函数 scrypt() 可使用 OpenSSL 1.1.0 或更新版本. (由Christian Heimes 在 [issue 27928][118]中贡献.)

http.client¶

[HTTPConnection.request()](https://docs.python.org/3.6/library/http.client.h tml#http.client.HTTPConnection.request "http.client.HTTPConnection.request" ) and [endheaders()](https://docs.python.org/3.6/library/http.client.html#http .client.HTTPConnection.endheaders "http.client.HTTPConnection.endheaders" ) 将全部支持分块编码请求体. (由Demian Brecht and Rolf Krahl 在 [issue 12319][119]中贡献.)

idlelib and IDLE¶

对idle包做了现代化的改进与重构，使得IDLE更美观、更好用的同时令编程更易于理解、测试与改进。在IDLE的美化方面，特备针对Linux和Mac用户，我们在大多数对话框上应用了ttk插件。总之，IDLE将不再支持tcl/tk 8.4。现在要求有 tcl/tk 8.5 或 8.6。我们建议在使用时运行最新的版本.

“现代化”包括对idlelib模块的重命名和整合.重命名文件部分大写的名字是与之前版本类似的命名，例如，Tkinter和TkFont对应3.0版本中的Tkinter和tkinter.font 。因此，对idlelib在3.5环境下导入的文件通常不会工作在3.6。至少一个模块的名称需要改变（见idlelib / readme.txt），有时甚至会更多。(名称变更由Al Swiegart 与 Terry Reedy 在 [issue 24225][120]中贡献. 大多数idle补丁均会加入这一改进。)

做点补充，最终的结果是，一些idlelib类会更容易使用，将具有更好的API文档与字符串的解释。其他有用的信息会在可用时被及时添加到idlelib。

importlib¶

当无法找到被导入模块时会跳出一个新的异常提示 [ModuleNotFoundError](https://docs.pytho n.org/3.6/library/exceptions.html#ModuleNotFoundError "ModuleNotFoundError" ) ([ImportError](https://docs.python.org/3.6/library/exceptions.ht ml#ImportError "ImportError" )的一个子类) . 检测ImportError的代码（try-except）依然会工作。 (由 Eric Snow 在 [issue 15767][121]中贡献.)

[importlib.util.LazyLoader](https://docs.python.org/3.6/library/importlib.ht ml#importlib.util.LazyLoader "importlib.util.LazyLoader" )在打包好的装载器上更名为 [`create_ module()`](https://docs.python.org/3.6/library/importlib.html#importlib.abc.Lo ader.create_module "importlib.abc.Loader.create_module" ) , 去除了 [importlib.machinery.BuiltinImporter]( https://docs.python.org/3.6/library/importlib.html#importlib.machinery.Builtin Importer "importlib.machinery.BuiltinImporter" )与 [`importlib.machinery.Ext ensionFileLoader`](https://docs.python.org/3.6/library/importlib.html#importli b.machinery.ExtensionFileLoader "importlib.machinery.ExtensionFileLoader" ) 不能同 [importlib.util.LazyLoader](https://docs.python.org/3. 6/library/importlib.html#importlib.util.LazyLoader "importlib.util.LazyLoader" )一起使用的限制.

[importlib.util.cache_from_source()](https://docs.python.org/3.6/library/imp ortlib.html#importlib.util.cache_from_source "importlib.util.cache_from_source" ), [importlib.util.source_from_cache()](h ttps://docs.python.org/3.6/library/importlib.html#importlib.util.source_from_c ache "importlib.util.source_from_cache" ), 和 [`importlib.util.spec_from_file _location()`](https://docs.python.org/3.6/library/importlib.html#importlib.uti l.spec_from_file_location "importlib.util.spec_from_file_location" ) 现在可接受[path-like object](https://docs.python.org/3.6/glossary.html#term- path-like-object).

inspect

inspect.signature()函数现在支持报告由编译器为推导式和生成器表达式范围生成的隐式.0参数，就好像它们是名为implicit0的仅位置参数。 (由Jelle Zijlstra在issue 19611中贡献。)

为了减少从Python 2.7和旧版`inspect.getargspec()` API升级时的代码改动，先前记录的对`inspect.getfullargspec()`对弃用已经撤销。虽然这个函数对于单个/源Python 2/3代码库是很方便的，但是更丰富的`inspect.signature()`接口仍然是新代码的推荐方法。(由Nick Coghlan在issue 27172中贡献)

json

json.load()和json.loads()现在支持二进制输入。已编码的JSON应该用UTF-8, UTF-16, 或者UTF-32来表示。(由Serhiy Storchaka在issue 17909中贡献。)

logging

已添加新的WatchedFileHandler.reopenIfNeeded()方法来增加检查日志文件是否需要被重新打开的能力。(由Marian Horban在issue 24884中贡献。)

math

已添加tau (τ)常量到math和cmath模块中。(由Lisa Roach在issue 12345中贡献，见**PEP 628**以了解详情。)

multiprocessing

multiprocessing.Manager()返回的Proxy对象现在可以被嵌套。(由Davin Potts在issue 6766中贡献。)

os

见PEP 519的摘要，以获取关于os和os.path模块现在如何支持类路径（path-like）对象的详情。

scandir()现在支持Windows上的bytes路径。

一个新的close()方法允许显式关闭一个scandir()迭代器。scandir()迭代器现在支持上下文管理器协议。如果一个scandir()迭代器既不是被耗尽，也不是被显式关闭，那么在其析构函数中将会抛出一个ResourceWarning。(由Serhiy Storchaka在issue 25994中贡献。)

在Linux上，os.urandom()现在会阻塞，直到系统的urandom entropy池被初始化，以提高安全性。见**PEP 524**以了解缘由。

Linux的getrandom()系统调用(获取随机字节)现在作为新的os.getrandom()函数公开。(由Victor Stinner贡献，属于**PEP 524**部分)

pathlib

pathlib现在支持类路径（path-like）对象。 (由Brett Cannon在issue 27186中贡献。)

见PEP 519的摘要了解详情。

pdb

Pdb类构造器有一个新的可选_readrc_参数，用来控制是否应该读取.pdbrc文件。

pickle¶

对象，现在可以使用关键参数 __new__进行持久存储[pickle协议](https://docs.python.org/3.6/library/pickle.html

pickle-protocols)超过已有的协议版本4.

支持这种情况. (由Serhiy Storchaka在issue 24164的贡献。)

pickletools¶

现在[pickletools.dis()](https://docs.python.org/3.6/library/pickletools.html#pic kletools.dis "pickletools.dis" ) 输出的隐含备注是MEMOIZE操作码的索引。 (由Serhiy Storchaka在issue 25382的贡献。)

pydoc¶

pydoc模块已经学会遵守MANPAGER环境变量。 (由Matthias Klose 在[issue 8637][130]的贡献。)

[help()][131] 和 pydoc 现在能用指定的元组字段定义的顺序来显示列表，而不是按字母顺序。 (由Raymond Hettinger在issue 24879的贡献。)

random¶

The new choices() function returns a list of elements of specified size from the given population with optional weights. (Contributed by Raymond Hettinger in [issue 18844][132].)

re¶

Added support of modifier spans in regular expressions. Examples: '(?i:p)ython' matches 'python' and 'Python', but not 'PYTHON'; '(?i)g(?-i:v)r' matches 'GvR' and 'gvr', but not 'GVR'. (Contributed by Serhiy Storchaka in [issue 433028][133].)

Match object groups can be accessed by __getitem__, which is equivalent to group(). So mo['name'] is now equivalent to mo.group('name'). (Contributed by Eric Smith in issue 24454.)

Match objects now support [index-like objects](https://docs.python.org/3.6 /reference/datamodel.html#object.__index__ "object.__index__" ) as group indices. (Contributed by Jeroen Demeyer and Xiang Zhang in issue 27177.)

readline¶

Added [set_auto_history()](https://docs.python.org/3.6/library/readline.html

readline.set_auto_history "readline.set_auto_history" ) to enable or disable

automatic addition of input to the history list. (Contributed by Tyler Crompton in [issue 26870][134].)

rlcompleter¶

Private and special attribute names now are omitted unless the prefix starts with underscores. A space or a colon is added after some completed keywords. (Contributed by Serhiy Storchaka in issue 25011 and issue 25209.)

shlex¶

The shlex has much [improved shell compatibility](https://docs.python.org/3.6/library/shlex.html#improved-shell- compatibility) through the new punctuation_chars argument to control which characters are treated as punctuation. (Contributed by Vinay Sajip in issue 1521950.)

site¶

When specifying paths to add to sys.path in a .pth file, you may now specify file paths on top of directories (e.g. zip files). (Contributed by Wolfgang Langner in issue 26587).

sqlite3¶

[sqlite3.Cursor.lastrowid](https://docs.python.org/3.6/library/sqlite3.html\# sqlite3.Cursor.lastrowid "sqlite3.Cursor.lastrowid" ) 现在支持 REPLACE 声明. (贡献者： Alex LordThorsen 于 issue 16864.)

socket¶

[ioctl()](https://docs.python.org/3.6/library/socket.html#socket.socket. ioctl "socket.socket.ioctl" ) 函数现在支持 [`SIO_LOOPBACK_FAST_PA TH`](https://docs.python.org/3.6/library/socket.html#socket.SIO\_LOOPBACK\_FAST\_ PATH "socket.SIO_LOOPBACK_FAST_PATH" ) 控制代码. (贡献者： Daniel Stokes 于 [issue 26536][135].)

[getsockopt()](https://docs.python.org/3.6/library/socket.html#socket.so cket.getsockopt "socket.socket.getsockopt" ) 常量 SO_DOMAIN, SO_PROTOCOL, SO_PEERSEC, 和 SO_PASSSEC 现已被支持. (贡献者：Christian Heimes 于 [issue 26907][136].)

[setsockopt()](https://docs.python.org/3.6/library/socket.html#socket.so cket.setsockopt "socket.socket.setsockopt" ) 现在支持 setsockopt(level, optname, None, optlen: int) 形式. (贡献者： Christian Heimes 于 [issue 27744][137].)

socket 模块现在支持地址族 AF_ALG 来和Linux Kernel crypto API连接. 添加了ALG_*, SOL_ALG 和 [sendmsg_afalg()](https://docs.python.org/3.6/library/socket.h tml#socket.socket.sendmsg_afalg "socket.socket.sendmsg_afalg" ). (贡献者： Christian Heimes 于 issue 27744 with support from Victor Stinner.)

socketserver¶

基于 [socketserver](https://docs.python.org/3.6/library/socketserver.html#module- socketserver "socketserver: A framework for network servers." ) 模块, 包括了 [http.server](https://docs.python.org/3.6/library/http.server.html#module- http.server "http.server: HTTP server and request handlers." ), xmlrpc.server 和 wsgiref.simple_server的Servers 现在支持 context manager protocol. (Contributed by Aviv Palivoda in issue 26404.)

[StreamRequestHandler](https://docs.python.org/3.6/ library/socketserver.html#socketserver.StreamRequestHandler "socketserver.StreamRequestHandler" ) 类的 wfile 属性现在实现 the `io.Buffered IOBase` 可写接口. 特别地, 调用 [write()](h ttps://docs.python.org/3.6/library/io.html#io.BufferedIOBase.write "io.BufferedIOBase.write" ) 现在可以保证完全发送数据. (贡献者： Martin Panter 于 issue 26721.)

ssl¶

[ssl][139] 已支持 OpenSSL 1.1.0. 最低推荐版本号是 1.0.2. (贡献者： Christian Heimes 于 issue 26470.)

3DES 已从默认密码套件default cipher suites中删除添加了密码套件ChaCha20 Poly1305. (贡献者： Christian Heimes 于 issue 27850 和 issue 27766.)

SSLContext 有更好的默认配置选项和密码. (贡献者： Christian Heimes 于 issue 28043.)

SSL session 可以使用新的 SSLSession 类从一个客户端连接复制到另一个客户端连接. TLS会话恢复可以加速初始握手，减少延迟并提高性能 (贡献者： Christian Heimes 于 [issue 19500][138] ，基于Alex Warhawk的草案.)

新的 [get_ciphers()](https://docs.python.org/3.6/library/ssl.html#ssl.SSL Context.get_ciphers "ssl.SSLContext.get_ciphers" ) 方法可以用来得到一个按密码优先级顺序的已启用密码列表.

所有的 constants and flags 已转换为 IntEnum 和 IntFlags. (贡献者：Christian Heimes 于 issue 28025.)

Server 和 client-side 添加 SSLContext 特定 TLS 协议. (贡献者： Christian Heimes 于 issue 28085.)

statistics¶

A new [harmonic_mean()](https://docs.python.org/3.6/library/statistics.html\# statistics.harmonic_mean "statistics.harmonic_mean" ) function has been added. (Contributed by Steven D'Aprano in issue 27181.)

struct¶

struct now supports IEEE 754 half- precision floats via the 'e' format specifier. (Contributed by Eli Stevens, Mark Dickinson in [issue 11734][139].)

subprocess¶

[subprocess.Popen](https://docs.python.org/3.6/library/subprocess.html#subpr ocess.Popen "subprocess.Popen" ) destructor now emits a [ResourceWarning](ht tps://docs.python.org/3.6/library/exceptions.html#ResourceWarning "ResourceWarning" ) warning if the child process is still running. Use the context manager protocol (with proc: ...) or explicitly call the wait() method to read the exit status of the child process. (Contributed by Victor Stinner in issue 26741.)

The [subprocess.Popen](https://docs.python.org/3.6/library/subprocess.html#s ubprocess.Popen "subprocess.Popen" ) constructor and all functions that pass arguments through to it now accept encoding and errors arguments. Specifying either of these will enable text mode for the stdin, stdout and stderr streams. (Contributed by Steve Dower in issue 6135.)

sys¶

The new [getfilesystemencodeerrors()](https://docs.python.org/3.6/library/sy s.html#sys.getfilesystemencodeerrors "sys.getfilesystemencodeerrors" ) function returns the name of the error mode used to convert between Unicode filenames and bytes filenames. (Contributed by Steve Dower in issue 27781.)

On Windows the return value of the [getwindowsversion()](https://docs.python .org/3.6/library/sys.html#sys.getwindowsversion "sys.getwindowsversion" ) function now includes the platform_version field which contains the accurate major version, minor version and build number of the current operating system, rather than the version that is being emulated for the process (Contributed by Steve Dower in [issue 27932][140].)

telnetlib¶

Telnet is now a context manager (contributed by Stéphane Wirtel in [issue 25485][141]).

time¶

The struct_time attributes tm_gmtoff and tm_zone are now available on all platforms.

timeit¶

The new [Timer.autorange()](https://docs.python.org/3.6/library/timeit.html\# timeit.Timer.autorange "timeit.Timer.autorange" ) convenience method has been added to call [Timer.timeit()](https://docs.python.org/3.6/library/timeit.ht ml#timeit.Timer.timeit "timeit.Timer.timeit" ) repeatedly so that the total run time is greater or equal to 200 milliseconds. (Contributed by Steven D'Aprano in [issue 6422][142].)

timeit now warns when there is substantial (4x) variance between best and worst times. (Contributed by Serhiy Storchaka in [issue 23552][143].)

tkinter¶

在tkinter.Variable类中添加了方法 trace_add（），trace_remove（）和trace_info（）。它们代替了之前版本中的 trace_variable（），trace（），trace_vdelete（）和 trace_vinfo（）方法，这些方法使用过时的Tcl命令，而在未来版本的Tcl中，这些Tcl命令可能不起作用。（由Serhiy Storchaka在issue 22115 提供）。

traceback

traceback模块和解释器内置的异常展示现在都省略回溯中重复行的长串，如以下例子所示：

    >>> def f(): f()
    ...
    >>> f()
    Traceback (most recent call last):
      File "<stdin>", line 1, in <module>
      File "<stdin>", line 1, in f
      File "<stdin>", line 1, in f
      File "<stdin>", line 1, in f
      [Previous line repeated 995 more times]
    RecursionError: maximum recursion depth exceeded

(由Emanuel Barry在issue 26823中贡献。)

tracemalloc

tracemalloc模块限制支持跟踪多个不同地址空间内的内存分配。

已添加新的DomainFilter过滤器类来根据地址空间（域）过滤块跟踪。

(由Victor Stinner在issue 26588中贡献。)

typing

从Python 3.6起，typing模块不再是临时的了，可以把它的API当成稳定版本使用。

由于typing模块在Python 3.5中是临时的，因此Python 3.6中引入的所有变动也已向后移植到Python 3.5.x。

typing模块很大提高了对范型别名的支持。例如，Dict[str, Tuple[S, T]]现在是一个有效的类型注释了。(由Guido van Rossum在Github #195中贡献。)

已添加typing.ContextManager类来展示contextlib.AbstractContextManager。 (由Brett Cannon在issue 25609中贡献。)

已添加typing.Collection类来展示collections.abc.Collection。 (由Ivan Levkivskyi在issue 27598中贡献。)

已添加typing.ClassVar类型构造，来标识类变量。如**PEP 526**中所述，封装在ClassVar中的一个变量注释暗示着一个给定的属性打算作为一个类变量使用，并且不应该在那个类的实例上设置它。(由Ivan Levkivskyi在Github #280中贡献。)

一个新的TYPE_CHECKING常量被静态类型检查器假设为True，但在运行时则为False。 (由Guido van Rossum在Github #230中贡献。)

已添加一个新的NewType()辅助函数来为注释创建轻量单值类型：

    from typing import NewType

    UserId = NewType('UserId', int)
    some_id = UserId(524313)

静态类型检查器将会把新的类型当成原始类型的一个子类。 (由Ivan Levkivskyi在Github #189中贡献。)

unicodedata¶

The unicodedata module now uses data from [Unicode 9.0.0][148]. (Contributed by Benjamin Peterson.)

unittest.mock¶

The Mock class has the following improvements:

Two new methods, [Mock.assert_called()][149] and [Mock.assert_called_once()][150] to check if the mock object was called. (Contributed by Amit Saha in [issue 26323][151].)
The [Mock.reset_mock()][152] method now has two optional keyword only arguments: return_value and side_effect. (Contributed by Kushal Das in [issue 21271][153].)

urllib.request¶

If a HTTP request has a file or iterable body (other than a bytes object) but no Content-Length header, rather than throwing an error, AbstractHTTPHandler now falls back to use chunked transfer encoding. (Contributed by Demian Brecht and Rolf Krahl in issue 12319.)

urllib.robotparser¶

RobotFileParser now supports the Crawl-delay and Request-rate extensions. (Contributed by Nikolay Bogoychev in [issue 16099][154].)

venv¶

[venv][155] accepts a new parameter --prompt. This parameter provides an alternative prefix for the virtual environment. (Proposed by Łukasz Balcerzak and ported to 3.6 by Stéphane Wirtel in issue22829.)

warnings¶

增加一个可选参数到 warnings.warn_explicit（）函数中：引发 ResourceWarning 的已销毁对象。同时，属性也添加到 warnings.WarningMessage 中。（由Victor Stinner在issue 26568和issue 26567中提供）。

当引发 ResourceWarning 警告时，tracemalloc 模块就尝试检索分配了销毁对象的跟踪。

一个名为 example.py的例子

    import warnings

    def func():
        return open(__file__)

    f = func()
    f = None

输出命令 python3.6 -Wd -X tracemalloc=5 example.py :

    example.py:7: ResourceWarning: unclosed file <_io.TextIOWrapper name='example.py' mode='r' encoding='UTF-8'>
      f = None
    Object allocated at (most recent call first):
      File "example.py", lineno 4
          return open(__file__)
      File "example.py", lineno 6
          f = func()

“对象分配”跟踪是新的，并且只有当tracemalloc正在跟踪Python内存分配，并且 warnings 模块已经导入时才会显示。

winreg¶

新增64位整数类型 REG_QWORD. (Contributed by Clement Rouault in issue 23026.)

winsound¶

允许传递关键词参数到 Beep, [MessageBeep](https://docs.python.org/3.6/library/winsoun d.html#winsound.MessageBeep "winsound.MessageBeep" ), 和 [PlaySound](https: //docs.python.org/3.6/library/winsound.html#winsound.PlaySound"winsound.PlaySound" ) 方法。 ([issue 27982][156]).

xmlrpc.client¶

[xmlrpc.client](https://docs.python.org/3.6/library/xmlrpc.client.html# module-xmlrpc.client "xmlrpc.client: XML-RPC clientaccess." ) 模块现支持解压由apache XML-RPC实现用于数字和'None'的附加数据类型。(Contributed by Serhiy Storchaka in [issue 26885][157].)

zipfile¶

一个新的类方法 ZipInfo.from_file() 允许从文件系统文件中生成一个 ZipInfo 实例。一个新的方法 ZipInfo.is_dir() 能够用来检查ZipInfo 实例是否表示一个目录。 (Contributed by Thomas Kluyver in [issue 26039][158].)

ZipFile.open() 方法现在能够将数据写入到一个ZIP文件，也能够从中提取数据。 (Contributed by Thomas Kluyver in [issue 26039][159].)

zlib¶

compress() 以及 decompress() 函数现接受关键词参数。 (Contributed by Aviv Palivoda in [issue 26243][160] and Xiang Zhang in [issue 16764][161] respectively.)

优化

Python解释器现在使用16位编码而不是字节码，这使得你可以实现多种操作码的优化。(由Demur Rumed，Serhiy Storchaka和Victor Stinner在[issue 26647][162] 以及 [issue 28050][163]中贡献。)
[asyncio.Future][164]类现在有一种优化的C实现。(由Yury Selivanov和INADA Naoki在[issue 26081][165]中贡献。)
[asyncio.Task][166]类现在有一种优化的C实现。(由Yury Selivanov在[issue 28544][167]中贡献。)
在[typing][168]模块中通过各种改进的实施(如泛型类型的缓存)，使得其允许高达30倍的性能优化，并且减少了内存占用。
ASCII解码器的错误处理程序 surrogateescape, ignore 以及 replace 速度提高了60倍。(由Victor Stinner在[issue 24870][169]中贡献)。
ASCII和Latin1解码器的错误处理程序 surrogateescape 速度提高了3倍。(由Victor Stinner在[issue 25227][170]中贡献)。
UTF-8解码器的错误处理程序 ignore, replace, surrogateescape, surrogatepass速度提高了75倍。(由Victor Stinner在[issue 25267][171]中贡献)。
UTF-8解码器的错误处理程序ignore, replace and surrogateescape速度提高了15倍。由Victor Stinner在[issue 25301][172]中贡献)。
bytes % args现在速度可以提高到2倍。(由Victor Stinner在[issue 25349][173]中贡献)。
bytearray % args 现在速度是以前的2.5到5倍。(由Victor Stinner在[issue 25399][174]中贡献)。
对[bytes.fromhex()][175]和[bytearray.fromhex()][176]进行优化:现在比以前快2x到3.5x倍。(由Victor Stinner在[issue 25401][177]中贡献)。
对bytes.replace(b'', b'.')和bytearray.replace(b'', b'.')进行优化:速度可提高80%。 (由Josh Snider在[issue 26574][178]中贡献)。
现在[PyMem_Malloc()][179]域([PYMEM_DOMAIN_MEM][180])的分配函数使用[pymalloc memory allocator][181]，而不是C库中的malloc()函数。 pymalloc分配器针对大小小于等于512字节、生命周期较短的对象进行了优化，而当需要占用较大内存块时则使用malloc()。 (由Victor Stinner在[issue 26249][182]中贡献)。
当对大量小对象进行反序列化时，[pickle.load()][183]和[pickle.loads()][184]的速度可提高10%。(由Victor Stinner在[issue 27056][185]中贡献)。
与传递[位置参数][187]相比，将[关键字参数][186]传递给函数会带来额外的开销。现在在通过使用Argument Clinic实现的扩展功能中，这种开销将会显著的降低。(由Serhiy Storchaka在[issue 27574][188]中贡献)。
对[glob][191]模块中的[glob()][189]及[iglob()][190]进行优化;使得它们现在大概快了3-6倍。(由Serhiy Storchaka在[issue 25596][192]中贡献)。
使用[os.scandir()][194]对[pathlib][193]中的glob进行优化;使它大概快了1.5-4倍。(由Serhiy Storchaka在[issue 26032][195]中贡献)。
[xml.etree.ElementTree][196]中解析、迭代和深拷贝的性能有了显著的提高。 (由Serhiy Storchaka在[issue 25638][197]、[issue 25873][198]及[issue 25869][199]中贡献。)
由浮点数和小数创建[fractions.Fraction][200]实例的速度提高了2到3倍。(由Serhiy Storchaka在[issue 25971][201]中贡献。)

构建和C API变动

Python现在构建需要工具链中的一些C99支持。最值得注意的是，Python现在使用标准的整型类型和宏来替代诸如PY_LONG_LONG的自定义宏。欲知详情，见PEP 7和issue 17884。
带Android NDK和Android API level 21 (Android 5.0 Lollilop) 或者更高版本交叉编译CPython可以成功运行。虽然Android尚未是一个支持的平台，但是Python测试套件在Android模拟器上运行只有大约16个测试错误。见Android元问题issue 26865。
已添加--enable-optimizations配置标志。打开它将会激活昂贵的优化，例如PGO。(原始补丁由Intel的Alecsandru Patrascu在issue 26539中提出。)
当PYMEM_DOMAIN_OBJ (例如：PyObject_Malloc())和PYMEM_DOMAIN_MEM (例如：PyMem_Malloc())域的分配函数被调用的时候，现在必须使用GIL。
新的Py_FinalizeEx() API，它表示是否清除缓冲数据失败。 (由Martin Panter在issue 5319中贡献。)
PyArg_ParseTupleAndKeywords()现在支持仅位置参数。仅位置参数是由空名称定义的。 (由Serhiy Storchaka在issue 26282中贡献)。
PyTraceback_Print方法现在省略了重复行的长串，变成"[Previous line repeated {count} more times]"。 (由Emanuel Barry在issue 26823中贡献。)
新的PyErr_SetImportErrorSubclass()函数指定要引发的ImportError的子类。 (由Eric Snow在issue 15767中贡献。)
新的PyErr_ResourceWarning()函数可以被用来生成一个ResourceWarning，提供资源分配的来源。(由Victor Stinner在issue 26567中贡献。)
新的PyOS_FSPath()函数返回一个类路径(path-like)对象的文件系统表示。(由Brett Cannon在issue 27186中贡献。)
PyUnicode_FSConverter()和PyUnicode_FSDecoder()函数现在会接受类路径(path-like)对象。

其他改进

当使用--version (短格式：-V)两次的时候，Python打印sys.version以获得详细信息。

$ ./python -VV

    Python 3.6.0b4+ (3.6:223967b49e49+, Nov 21 2016, 20:55:04)
    [GCC 4.2.1 Compatible Apple LLVM 8.0.0 (clang-800.0.42.1)]

弃用的模块，包以及函数¶

新的关键词¶

在后续的Python 3.7 中，async 和 await 将作为关键字使用，因此不再建议使用它们作为变量名,类名，函数名以及模块名称。详见Python 3.5**PEP492**. 从 Python 3.6 开始, async或await的使用将会引发[DeprecationWarning](https://doc s.python.org/3.6/library/exceptions.html#DeprecationWarning "DeprecationWarning" )

已弃用的 Python 行为¶

Python 3.7 中, 在生成器中引发StopIteration异常，现在将会导致DesprecationWarning, 并且触发`RuntimeError`. 查阅[PEP 479: Change StopIteration handlinginside generators](https://docs.python.org/3.6/whatsnew/3.5.html#whatsnew- pep-479)获得相关细节.

[__aiter__()](https://docs.python.org/3.6/reference/datamodel.html#objec t.__aiter__ "object.__aiter__" )方法现在被期望返回异步迭代器，而不是之前版本中的可等待迭代器，在 Python 3.6 中调用__aiter__()将会触发[DeprecationWarning](https://doc s.python.org/3.6/library/exceptions.html#DeprecationWarning "DeprecationWarning" )。 Python 3.7 将会移除其的后向兼容性。(Contributed by Yury Selivanov in issue27243.)

反斜杠对现在不能再作为有效的转义字符串使用，其会导致 [DeprecationWarning](https://docs.python.org/3.6/library/exceptions.html#Dep recationWarning "DeprecationWarning" ). 在接下来的几个Python版本，其最终会成为一个语法错误 [SyntaxError](https://docs.python.org/3.6/library/exceptions.html#SyntaxErro r "SyntaxError" )。(Contributed by Emanuel Barry in [issue 27364][231].)

在执行相对引用, 且__spec__ 或 __package__未定义时，转而调用__name__ 和 __path__模块的行为 , 现在将会引发[ImportWarning](https://docs.python.org/3.6/library/exceptions.htm l#ImportWarning "ImportWarning" )异常. (Contributed by Rose Ames in issue 25791.)

Deprecated Python modules, functions and methods¶

弃用的 Python 模块，函数和方法

asynchat¶

为了支持asyncio， asynchat已被弃用。(Contributed by Mariatta in [issue 25002][232].)

asyncore¶

为了支持asyncio模块. [asyncore](https://docs.python.org/3.6/library/asyncore.html#module- asyncore "asyncore: A base class for developing asynchronous socket handling services." )模块已被弃用. (Contributed by Mariatta in [issue 25002][233].),

dbm¶

与其他[dbm](https://docs.python.org/3.6/library/dbm.html#module-dbm "dbm: Interfaces to various Unix "database" formats." )模块的的实现方式不同， dbm.dumb模块可以通过'rw'模式创建数据库并允许通过'r'模式来修改数据库，不过这种行为现在将被弃用，在Python 3.8 将会被移除。(Contributed by Serhiy Storchaka in issue 21708.)

distutils¶

Distribution构造器中extra_path参数现在被认为是过时的，如果在 Python 3.6 中对其进行设置的话将会引发一个警告。对这个参数的支持将会在后续 Python 版本中移除。详情请参阅issue 27919

grp¶

getgrgid()中对非整数参数的支持，现在已被移除。(Contributed by Serhiy Storchaka in [issue 26129][234].)

importlib¶

为了用来在之前的版本中支持 [`importlib.abc.Loader.exec_module()`](https://docs.python.org/3.6/library/importlib.html#im portlib.abc.Loader.exec_module "importlib.abc.Loader.exec_module" )， [importlib.machinery.SourceFileLoader.load_module()](https://docs.python .org/3.6/library/importlib.html#importlib.machinery.SourceFileLoader.load_modu le "importlib.machinery.SourceFileLoader.load_module" ) 和 [`importlib.machin ery.SourcelessFileLoader.load_module()`](https://docs.python.org/3.6/library/i mportlib.html#importlib.machinery.SourcelessFileLoader.load_module "importlib.machinery.SourcelessFileLoader.load_module" )，所以他们是 [importlib](https://docs.python.org/3.6/library/importlib.html#module- importlib "importlib: The implementation of the import machinery." ) 中`importlib.abc.Loader.load_module()` 仅存的未被弃用的实现方式，不过在 Python 3.6 中，其也已经被弃用。

[importlib.machinery.WindowsRegistryFinder](https://docs.python.org/3.6/ library/importlib.html#importlib.machinery.WindowsRegistryFinder "importlib.machinery.WindowsRegistryFinder" )类已被弃用。在 Python 3.6.0 中，它被移动到 sys.meta_path中，不过在将来这可能会发生变化。

os¶

在[os][235]中对作为路径的通用bytes-like objects，compile()以及对类函数的非正式支持，现在已经弃用。

re¶

对正则式中间的内联标志 (?letters)的支持已被弃用，并将在后续 Python 版本中移除。对正则式开头的标志的支持仍然被保留。 (Contributed by Serhiy Storchaka in [issue 22493][237].)

ssl¶

OpenSSL 0.9.8, 1.0.0 和 1.0.1 已被弃用，不再支持。以后ssl模块至少需要 OpenSSL 1.0.2 或者 OpenSSL 1.1.0

为了支持context， SSL的相关参数如：certfile, keyfile and check_hostnameftplib, [http.client](https://docs.python.org/3.6/library/http.client.html#module- http.client "http.client: HTTP and HTTPS protocol client (requires sockets)." ), imaplib, poplib, 和 smtplib 已被弃用。(Contributed by Christian Heimes in issue 28022.)

[ssl][238]的部分协议及函数，现已被弃用。后续的 OpenSSL 版本中，部分现有特征将不再可用。为了引入新的API，其他的特征也已经被弃用。 (Contributed by Christian Heimes in [issue 28022][239] 和 issue26470.)

tkinter¶

tkinter.tix模块已被弃用。原tkinter的使用者，可以使用 tkinter.ttk来代替。

venv¶

为了引入python3 -m venv命令，pyvenv 已被弃用。这可以防止混淆何种 Python 解释器将被pyvenv连接以及何种 Python 解释器将被用在虚拟环境中。(Contributed by Brett Cannon in issue 25154.)

Deprecated functions and types of the C API¶

Undocumented functions PyUnicode_AsEncodedObject(), PyUnicode_AsDecodedObject(), PyUnicode_AsEncodedUnicode() and PyUnicode_AsDecodedUnicode() are deprecated now. Use the generic codec based API instead.

Deprecated Build Options¶

The --with-system-ffi configure flag is now on by default on non-macOS UNIX platforms. It may be disabled by using --without-system-ffi, but using the flag is deprecated and will not be accepted in Python 3.7. macOS is unaffected by this change. Note that many OS distributors already use the `--with-system- ffi` flag when building their system Python.

Removed¶

API and Feature Removals¶

Unknown escapes consisting of '\' and an ASCII letter in regular expressions will now cause an error. In replacement templates for [re.sub()][240] they are still allowed, but deprecated. The [re.LOCALE][241] flag can now only be used with binary patterns.
inspect.getmoduleinfo() was removed (was deprecated since CPython 3.3). [inspect.getmodulename()][242] should be used for obtaining the module name for a given path. (Contributed by Yury Selivanov in [issue 13248][243].)
traceback.Ignore class and traceback.usage, traceback.modname, traceback.fullmodname, traceback.find_lines_from_code, traceback.find_lines, traceback.find_strings, traceback.find_executable_lines methods were removed from the [traceback][244] module. They were undocumented methods deprecated since Python 3.2 and equivalent functionality is available from private methods.
The tk_menuBar() and tk_bindForTraversal() dummy methods in [tkinter][245] widget classes were removed (corresponding Tk commands were obsolete since Tk 4.0).
The [open()][246] method of the [zipfile.ZipFile][247] class no longer supports the 'U' mode (was deprecated since Python 3.4). Use [io.TextIOWrapper][248] for reading compressed text files in [universal newlines][249] mode.
The undocumented IN, CDROM, DLFCN, TYPES, CDIO, and STROPTS modules have been removed. They had been available in the platform specific Lib/plat-*/ directories, but were chronically out of date, inconsistently available across platforms, and unmaintained. The script that created these modules is still available in the source distribution at [Tools/scripts/h2py.py][250].
The deprecated asynchat.fifo class has been removed.

Porting to Python 3.6¶

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

Changes in 'python' Command Behavior¶

The output of a special Python build with defined COUNT_ALLOCS, SHOW_ALLOC_COUNT or SHOW_TRACK_COUNT macros is now off by default. It can be re-enabled using the -X showalloccount option. It now outputs to stderr instead of stdout. (Contributed by Serhiy Storchaka in [issue 23034][251].)

Changes in the Python API¶

[open()][252] will no longer allow combining the 'U' mode flag with '+'. (Contributed by Jeff Balogh and John O'Connor in [issue 2091][253].)
[sqlite3][254] no longer implicitly commits an open transaction before DDL statements.
On Linux, [os.urandom()][255] now blocks until the system urandom entropy pool is initialized to increase the security.
When [importlib.abc.Loader.exec_module()][256] is defined, [importlib.abc.Loader.create_module()][257] must also be defined.
[PyErr_SetImportError()][258] now sets [TypeError][259] when its msg argument is not set. Previously only NULL was returned.
The format of the co_lnotab attribute of code objects changed to support a negative line number delta. By default, Python does not emit bytecode with a negative line number delta. Functions using frame.f_lineno, PyFrame_GetLineNumber() or PyCode_Addr2Line() are not affected. Functions directly decoding co_lnotab should be updated to use a signed 8-bit integer type for the line number delta, but this is only required to support applications using a negative line number delta. See Objects/lnotab_notes.txt for the co_lnotab format and how to decode it, and see the [PEP 511][260] for the rationale.
The functions in the [compileall][261] module now return booleans instead of 1 or 0 to represent success or failure, respectively. Thanks to booleans being a subclass of integers, this should only be an issue if you were doing identity checks for 1 or 0. See [issue 25768][262].
Reading the port attribute of [urllib.parse.urlsplit()][263] and [urlparse()][264] results now raises [ValueError][265] for out-of-range values, rather than returning [None][266]. See [issue 20059][267].
The [imp][268] module now raises a [DeprecationWarning][269] instead of [PendingDeprecationWarning][270].
The following modules have had missing APIs added to their __all__ attributes to match the documented APIs: [calendar][271], [cgi][272], [csv][273], [ElementTree][274], [enum][275], [fileinput][276], [ftplib][277], [logging][278], [mailbox][279], [mimetypes][280], [optparse][281], [plistlib][282], [smtpd][283], [subprocess][284], [tarfile][285], [threading][286] and [wave][287]. This means they will export new symbols when import * is used. (Contributed by Joel Taddei and Jacek Kołodziej in [issue 23883][288].)
When performing a relative import, if __package__ does not compare equal to __spec__.parent then [ImportWarning][289] is raised. (Contributed by Brett Cannon in [issue 25791][290].)
When a relative import is performed and no parent package is known, then [ImportError][291] will be raised. Previously, [SystemError][292] could be raised. (Contributed by Brett Cannon in [issue 18018][293].)
Servers based on the [socketserver][294] module, including those defined in [http.server][295], [xmlrpc.server][296] and [wsgiref.simple_server][297], now only catch exceptions derived from [Exception][298]. Therefore if a request handler raises an exception like [SystemExit][299] or [KeyboardInterrupt][300], [handle_error()][301] is no longer called, and the exception will stop a single-threaded server. (Contributed by Martin Panter in [issue 23430][302].)
[spwd.getspnam()][303] now raises a [PermissionError][304] instead of [KeyError][305] if the user doesn't have privileges.
The [socket.socket.close()][306] method now raises an exception if an error (e.g. EBADF) was reported by the underlying system call. (Contributed by Martin Panter in [issue 26685][307].)
The decode_data argument for the [smtpd.SMTPChannel][308] and [smtpd.SMTPServer][309] constructors is now False by default. This means that the argument passed to [process_message()][310] is now a bytes object by default, and process_message() will be passed keyword arguments. Code that has already been updated in accordance with the deprecation warning generated by 3.5 will not be affected.
All optional arguments of the [dump()][311], [dumps()][312], [load()][313] and [loads()][314] functions and [JSONEncoder][315] and [JSONDecoder][316] class constructors in the [json][317] module are now [keyword-only][318]. (Contributed by Serhiy Storchaka in [issue 18726][319].)
Subclasses of [type][320] which don't override type.__new__ may no longer use the one-argument form to get the type of an object.
As part of [PEP 487][321], the handling of keyword arguments passed to [type][322] (other than the metaclass hint, metaclass) is now consistently delegated to [object.__init_subclass__()][323]. This means that type.__new__() and type.__init__() both now accept arbitrary keyword arguments, but [object.__init_subclass__()][324] (which is called from type.__new__()) will reject them by default. Custom metaclasses accepting additional keyword arguments will need to adjust their calls to type.__new__() (whether direct or via [super][325]) accordingly.
In distutils.command.sdist.sdist, the default_format attribute has been removed and is no longer honored. Instead, the gzipped tarfile format is the default on all platforms and no platform-specific selection is made. In environments where distributions are built on Windows and zip distributions are required, configure the project with a setup.cfg file containing the following:


    formats=zip

This behavior has also been backported to earlier Python versions by Setuptools 26.0.0.

In the [urllib.request][326] module and the [http.client.HTTPConnection.request()][327] method, if no Content-Length header field has been specified and the request body is a file object, it is now sent with HTTP 1.1 chunked encoding. If a file object has to be sent to a HTTP 1.0 server, the Content-Length value now has to be specified by the caller. (Contributed by Demian Brecht and Rolf Krahl with tweaks from Martin Panter in [issue 12319][328].)
The [DictReader][329] now returns rows of type [OrderedDict][330]. (Contributed by Steve Holden in [issue 27842][331].)
The [crypt.METHOD_CRYPT][332] will no longer be added to crypt.methods if unsupported by the platform. (Contributed by Victor Stinner in [issue 25287][333].)
The verbose and rename arguments for [namedtuple()][334] are now keyword-only. (Contributed by Raymond Hettinger in [issue 25628][335].)
On Linux, [ctypes.util.find_library()][336] now looks in LD_LIBRARY_PATH for shared libraries. (Contributed by Vinay Sajip in [issue 9998][337].)
The [imaplib.IMAP4][338] class now handles flags containing the ']' character in messages sent from the server to improve real-world compatibility. (Contributed by Lita Cho in [issue 21815][339].)
The mmap.write() function now returns the number of bytes written like other write methods. (Contributed by Jakub Stasiak in [issue 26335][340].)
The [pkgutil.iter_modules()][341] and [pkgutil.walk_packages()][342] functions now return [ModuleInfo][343] named tuples. (Contributed by Ramchandra Apte in [issue 17211][344].)
[re.sub()][345] now raises an error for invalid numerical group references in replacement templates even if the pattern is not found in the string. The error message for invalid group references now includes the group index and the position of the reference. (Contributed by SilentGhost, Serhiy Storchaka in [issue 25953][346].)
[zipfile.ZipFile][347] will now raise [NotImplementedError][348] for unrecognized compression values. Previously a plain [RuntimeError][349] was raised. Additionally, calling [ZipFile][350] methods on a closed ZipFile or calling the [write()][351] method on a ZipFile created with mode 'r' will raise a [ValueError][352]. Previously, a [RuntimeError][353] was raised in those scenarios.
when custom metaclasses are combined with zero-argument [super()][354] or direct references from methods to the implicit __class__ closure variable, the implicit __classcell__ namespace entry must now be passed up to type.__new__ for initialisation. Failing to do so will result in a [DeprecationWarning][355] in 3.6 and a [RuntimeWarning][356] in the future.

在C API上的改动¶

旧的[PyMem_Malloc()][357]分配算符族将使用新的[pymalloc分配算符][358]，替换掉了系统原有的’malloc()’。应用程序在没有捕获GIL时调用[PyMem_Malloc()][359]会导致崩溃。可通过设置[pythonmalloc][360]环境变量去’debug’来实现应用程内存的分配。详情请看[问题26429][361]。
[Py_Exit()][362]（和新的主注释器），如果刷新缓存失败，将覆盖当前状态。详情请看[问题5319][363]

CPython 字节码的更改¶

在Python 3.6中将会出现几个关于[字节码][364]要更新。 3.6.

新的Python注释器将使用16位词码去替代老的字符码。（感谢Demur Rumed在Serhiy Storchaka和Victor Stinner[问题26647][365]和[问题28050][366]中做出贡献后的审查）
将启用新的[FORMAT_VALUE][367]和[BUILD_STRING][368] 操作码作为一部分格式化字符串。（感谢Eric Smith在[问题25483][369]和Serhiy Storchaka在[问题27078][370]的贡献）
将启用新的[BUILD_CONST_KEY_MAP][371]操作码并伴随常量关键字优化创建字典的过程。（感谢Serhiy Storchaka 提出的[问题27140][372].）
为了更好的表现和更快捷的启动, 已经对调用操作码的功能进行了深度重做。[MAKE_FUNCTION][373], [CALL_FUNCTION][374],[ CALL_FUNCTION_KW][375] 以及 BUILD_MAP_UNPACK_WITH_CALL 这些操作码已经做了修改，并添加了新的CALL_FUNCTION_EX 和 BUILD_TUPLE_UNPACK_WITH_CALL。而CALL_FUNCTION_VAR, CALL_FUNCTION_VAR_KW 和MAKE_CLOSURE则在新的Python中被移除。（感谢Demur Rumed在[问题27905][376]中的贡献以及Serhiy Storchaka在[问题27213][377]、[问题28257][378]中的贡献）
添加了[SETUP_ANNOTATIONS][379] 和 [STORE_ANNOTATION][380]操作码，用于支持新的[变量注释语法][381]。（感谢Ivan Levkivskyi在[问题27985][382]中的贡献）

(C) [Copyright][383] 2001-2016, Python Software Foundation. The Python Software Foundation is a non-profit corporation. Please donate. Last updated on Dec 15, 2016. Found a bug? Created using [Sphinx][384] 1.3.3.

[54]: https://docs.python.org/3.6/library/os.html#module-os "os:

Mathematical functions (sin() etc.)." [125]: https://docs.python.org/3.6/library/cmath.html#module-cmath "cmath:

Mathematical functions for complex numbers." [126]: https://bugs.python.org/issue12345 [127]: https://docs.python.org/3.6/library/os.html#module-os "os:

Miscellaneous operating system interfaces." [236]: https://bugs.python.org/issue25791 [237]: https://bugs.python.org/issue22493 [238]: https://docs.python.org/3.6/library/ssl.html#module-ssl "ssl: TLS/SSL

Files

What’s New In Python 3.6.md

Latest commit

History

What’s New In Python 3.6.md

File metadata and controls

Python 3.6新特性简介¶

摘要 - 发布亮点¶

新特性

PEP 498: 格式化字符串

PEP 526: 变量注释语法

PEP 515: 数值文字中的下划线

PEP 525: 异步生成器¶

PEP 530: 异步解析式¶

PEP 487: 用于建立类的更简单的自定义¶

PEP 487: 描述符协议增强¶

PEP 519: 添加文件系统路径协议¶

PEP 495: Local Time Disambiguation¶

PEP 529: 更改windows下文件系统编码格式为UTF-8¶

PEP 528: 更改windows控制台编码为UTF-8¶

PEP 520: 保存类属性定义顺序¶

PEP 468: 保存关键字参数顺序¶

新的 字典dict类型的实现¶

PEP 523: Adding a frame evaluation API to CPython¶

PYTHONMALLOC environment variable¶

envvar-PYTHONMALLOC) environment variable allows setting the Python memory

DTrace and SystemTap probing support¶

其他语言方面的变化¶

新增模块¶

secrets¶

Improved Modules¶

array¶

ast¶

asyncio¶

binascii¶

cmath¶

collections¶

concurrent.futures¶

contextlib¶

datetime¶

datetime.datetime.isoformat "datetime.datetime.isoformat" ) function now

decimal¶

distutils¶

email¶

encodings¶

enum¶

faulthandler¶

fileinput¶

hashlib¶

http.client¶

idlelib and IDLE¶

importlib¶

inspect

json

logging

math

multiprocessing

os

pathlib

pdb

pickle¶

pickle-protocols)超过已有的协议版本4.

pickletools¶

pydoc¶

random¶

re¶

readline¶

readline.set_auto_history "readline.set_auto_history" ) to enable or disable

rlcompleter¶

shlex¶

site¶

sqlite3¶

socket¶

socketserver¶

ssl¶

statistics¶

struct¶

subprocess¶

sys¶

telnetlib¶

新的字典dict类型的实现¶