Rietveld Code Review Tool
Help | Bug tracker | Discussion group | Source code | Sign in
(38259)

Delta Between Two Patch Sets: Doc/whatsnew/3.3.rst

Issue 11682: PEP 380 reference implementation for 3.3
Left Patch Set: Created 7 years, 10 months ago
Right Patch Set: Created 7 years, 8 months ago
Left:
Right:
Use n/p to move between diff chunks; N/P to move between comments. Please Sign in to add in-line comments.
Jump to:
Left: Side by side diff | Download
Right: Side by side diff | Download
LEFTRIGHT
1 **************************** 1 ****************************
2 What's New In Python 3.3 2 What's New In Python 3.3
3 **************************** 3 ****************************
4 4
5 :Author: Raymond Hettinger 5 :Author: Raymond Hettinger
6 :Release: |release| 6 :Release: |release|
7 :Date: |today| 7 :Date: |today|
8 8
9 .. Rules for maintenance: 9 .. Rules for maintenance:
10 10
(...skipping 51 matching lines...) Expand 10 before | Expand all | Expand 10 after
62 exist in parallel; over time, this compatibility should be phased out. 62 exist in parallel; over time, this compatibility should be phased out.
63 63
64 On the Python side, there should be no downside to this change. 64 On the Python side, there should be no downside to this change.
65 65
66 On the C API side, PEP 393 is fully backward compatible. The legacy API 66 On the C API side, PEP 393 is fully backward compatible. The legacy API
67 should remain available at least five years. Applications using the legacy 67 should remain available at least five years. Applications using the legacy
68 API will not fully benefit of the memory reduction, or - worse - may use 68 API will not fully benefit of the memory reduction, or - worse - may use
69 a bit more memory, because Python may have to maintain two versions of each 69 a bit more memory, because Python may have to maintain two versions of each
70 string (in the legacy format and in the new efficient storage). 70 string (in the legacy format and in the new efficient storage).
71 71
72 Functionality
73 -------------
74
72 Changes introduced by :pep:`393` are the following: 75 Changes introduced by :pep:`393` are the following:
73 76
74 * Python now always supports the full range of Unicode codepoints, including 77 * Python now always supports the full range of Unicode codepoints, including
75 non-BMP ones (i.e. from ``U+0000`` to ``U+10FFFF``). The distinction between 78 non-BMP ones (i.e. from ``U+0000`` to ``U+10FFFF``). The distinction between
76 narrow and wide builds no longer exists and Python now behaves like a wide 79 narrow and wide builds no longer exists and Python now behaves like a wide
77 build, even under Windows. 80 build, even under Windows.
78 81
79 * The storage of Unicode strings now depends on the highest codepoint in the str ing:
80
81 * pure ASCII and Latin1 strings (``U+0000-U+00FF``) use 1 byte per codepoint;
82
83 * BMP strings (``U+0000-U+FFFF``) use 2 bytes per codepoint;
84
85 * non-BMP strings (``U+10000-U+10FFFF``) use 4 bytes per codepoint.
86
87 .. The memory usage of Python 3.3 is two to three times smaller than Python 3.2,
88 and a little bit better than Python 2.7, on a `Django benchmark
89 <http://mail.python.org/pipermail/python-dev/2011-September/113714.html>`_.
90 XXX The result should be moved in the PEP and a small summary about
91 performances and a link to the PEP should be added here.
92
93 * With the death of narrow builds, the problems specific to narrow builds have 82 * With the death of narrow builds, the problems specific to narrow builds have
94 also been fixed, for example: 83 also been fixed, for example:
95 84
96 * :func:`len` now always returns 1 for non-BMP characters, 85 * :func:`len` now always returns 1 for non-BMP characters,
97 so ``len('\U0010FFFF') == 1``; 86 so ``len('\U0010FFFF') == 1``;
98 87
99 * surrogate pairs are not recombined in string literals, 88 * surrogate pairs are not recombined in string literals,
100 so ``'\uDBFF\uDFFF' != '\U0010FFFF'``; 89 so ``'\uDBFF\uDFFF' != '\U0010FFFF'``;
101 90
102 * indexing or slicing non-BMP characters returns the expected value, 91 * indexing or slicing non-BMP characters returns the expected value,
103 so ``'\U0010FFFF'[0]`` now returns ``'\U0010FFFF'`` and not ``'\uDBFF'``; 92 so ``'\U0010FFFF'[0]`` now returns ``'\U0010FFFF'`` and not ``'\uDBFF'``;
104 93
105 * several other functions in the standard library now handle correctly 94 * all other functions in the standard library now correctly handle
106 non-BMP codepoints. 95 non-BMP codepoints.
107 96
108 * The value of :data:`sys.maxunicode` is now always ``1114111`` (``0x10FFFF`` 97 * The value of :data:`sys.maxunicode` is now always ``1114111`` (``0x10FFFF``
109 in hexadecimal). The :c:func:`PyUnicode_GetMax` function still returns 98 in hexadecimal). The :c:func:`PyUnicode_GetMax` function still returns
110 either ``0xFFFF`` or ``0x10FFFF`` for backward compatibility, and it should 99 either ``0xFFFF`` or ``0x10FFFF`` for backward compatibility, and it should
111 not be used with the new Unicode API (see :issue:`13054`). 100 not be used with the new Unicode API (see :issue:`13054`).
112 101
113 * The :file:`./configure` flag ``--with-wide-unicode`` has been removed. 102 * The :file:`./configure` flag ``--with-wide-unicode`` has been removed.
114 103
115 XXX mention new and deprecated functions and macros 104 Performance and resource usage
105 ------------------------------
106
107 The storage of Unicode strings now depends on the highest codepoint in the strin g:
108
109 * pure ASCII and Latin1 strings (``U+0000-U+00FF``) use 1 byte per codepoint;
110
111 * BMP strings (``U+0000-U+FFFF``) use 2 bytes per codepoint;
112
113 * non-BMP strings (``U+10000-U+10FFFF``) use 4 bytes per codepoint.
114
115 The net effect is that for most applications, memory usage of string storage
116 should decrease significantly - especially compared to former wide unicode
117 builds - as, in many cases, strings will be pure ASCII even in international
118 contexts (because many strings store non-human language data, such as XML
119 fragments, HTTP headers, JSON-encoded data, etc.). We also hope that it
120 will, for the same reasons, increase CPU cache efficiency on non-trivial
121 applications.
122
123 .. The memory usage of Python 3.3 is two to three times smaller than Python 3.2,
124 and a little bit better than Python 2.7, on a `Django benchmark
125 <http://mail.python.org/pipermail/python-dev/2011-September/113714.html>`_.
126 XXX The result should be moved in the PEP and a link to the PEP should
127 be added here.
116 128
117 129
118 PEP 3151: Reworking the OS and IO exception hierarchy 130 PEP 3151: Reworking the OS and IO exception hierarchy
119 ===================================================== 131 =====================================================
120 132
121 :pep:`3151` - Reworking the OS and IO exception hierarchy 133 :pep:`3151` - Reworking the OS and IO exception hierarchy
122 PEP written and implemented by Antoine Pitrou. 134 PEP written and implemented by Antoine Pitrou.
123 135
124 The hierarchy of exceptions raised by operating system errors is now both 136 The hierarchy of exceptions raised by operating system errors is now both
125 simplified and finer-grained. 137 simplified and finer-grained.
(...skipping 58 matching lines...) Expand 10 before | Expand all | Expand 10 after
184 196
185 197
186 PEP 380: Syntax for Delegating to a Subgenerator 198 PEP 380: Syntax for Delegating to a Subgenerator
187 ================================================ 199 ================================================
188 200
189 PEP 380 adds the ``yield from`` expression, allowing a generator to delegate 201 PEP 380 adds the ``yield from`` expression, allowing a generator to delegate
190 part of its operations to another generator. This allows a section of code 202 part of its operations to another generator. This allows a section of code
191 containing 'yield' to be factored out and placed in another generator. 203 containing 'yield' to be factored out and placed in another generator.
192 Additionally, the subgenerator is allowed to return with a value, and the 204 Additionally, the subgenerator is allowed to return with a value, and the
193 value is made available to the delegating generator. 205 value is made available to the delegating generator.
194
195 While designed primarily for use in delegating to a subgenerator, the ``yield 206 While designed primarily for use in delegating to a subgenerator, the ``yield
196 from`` expression actually allows delegation to arbitrary subiterators. 207 from`` expression actually allows delegation to arbitrary subiterators.
197 208
198 (Implementation by Greg Ewing, integrated into 3.3 by Renaud Blanch, Ryan 209 (Implementation by Greg Ewing, integrated into 3.3 by Renaud Blanch, Ryan
199 Kelly and Nick Coghlan) 210 Kelly and Nick Coghlan, documentation by Zbigniew Jędrzejewski-Szmek and
211 Nick Coghlan)
212
213
214 PEP 3155: Qualified name for classes and functions
215 ==================================================
216
217 :pep:`3155` - Qualified name for classes and functions
218 PEP written and implemented by Antoine Pitrou.
219
220 Functions and class objects have a new ``__qualname__`` attribute representing
221 the "path" from the module top-level to their definition. For global functions
222 and classes, this is the same as ``__name__``. For other functions and classes,
223 it provides better information about where they were actually defined, and
224 how they might be accessible from the global scope.
225
226 Example with (non-bound) methods::
227 >>> class C:
228 ... def meth(self):
229 ... pass
230 >>> C.meth.__name__
231 'meth'
232 >>> C.meth.__qualname__
233 'C.meth'
234
235 Example with nested classes::
236
237 >>> class C:
238 ... class D:
239 ... def meth(self):
240 ... pass
241 ...
242 >>> C.D.__name__
243 'D'
244 >>> C.D.__qualname__
245 'C.D'
246 >>> C.D.meth.__name__
247 'meth'
248 >>> C.D.meth.__qualname__
249 'C.D.meth'
250
251 Example with nested functions::
252
253 >>> def outer():
254 ... def inner():
255 ... pass
256 ... return inner
257 ...
258 >>> outer().__name__
259 'inner'
260 >>> outer().__qualname__
261 'outer.<locals>.inner'
262
263 The string representation of those objects is also changed to include the
264 new, more precise information::
265
266 >>> str(C.D)
267 "<class '__main__.C.D'>"
268 >>> str(C.D.meth)
269 '<function C.D.meth at 0x7f46b9fe31e0>'
200 270
201 271
202 Other Language Changes 272 Other Language Changes
203 ====================== 273 ======================
204 274
205 Some smaller changes made to the core Python language are: 275 Some smaller changes made to the core Python language are:
206 276
207 * Stub 277 * Added support for Unicode name aliases and named sequences.
208 278 Both :func:`unicodedata.lookup()` and ``'\N{...}'`` now resolve name aliases,
209 Added support for Unicode name aliases and named sequences. 279 and :func:`unicodedata.lookup()` resolves named sequences too.
210 Both :func:`unicodedata.lookup()` and ``'\N{...}'`` now resolve name aliases, 280
211 and :func:`unicodedata.lookup()` resolves named sequences too. 281 (Contributed by Ezio Melotti in :issue:`12753`)
212 282
213 (Contributed by Ezio Melotti in :issue:`12753`) 283 * Equality comparisons on :func:`range` objects now return a result reflecting
214 284 the equality of the underlying sequences generated by those range objects.
215 285
216 Equality comparisons on :func:`range` objects now return a result reflecting 286 (:issue:`13021`)
217 the equality of the underlying sequences generated by those range objects. 287
218 288 * The ``count()``, ``find()``, ``rfind()``, ``index()`` and ``rindex()``
219 (:issue:`13021`) 289 methods of :class:`bytes` and :class:`bytearray` objects now accept an
220 290 integer between 0 and 255 as their first argument.
221 291
222 New, Improved, and Deprecated Modules 292 (:issue:`12170`)
223 ===================================== 293
224 294 * Memoryview objects are now hashable when the underlying object is hashable.
225 * Stub 295
296 (Contributed by Antoine Pitrou in :issue:`13411`)
297
298
299 New and Improved Modules
300 ========================
226 301
227 array 302 array
228 ----- 303 -----
229 304
230 The :mod:`array` module supports the :c:type:`long long` type using ``q`` and 305 The :mod:`array` module supports the :c:type:`long long` type using ``q`` and
231 ``Q`` type codes. 306 ``Q`` type codes.
232 307
233 (Contributed by Oren Tirosh and Hirokazu Yamamoto in :issue:`1172711`) 308 (Contributed by Oren Tirosh and Hirokazu Yamamoto in :issue:`1172711`)
234 309
235 310
(...skipping 23 matching lines...) Expand all
259 >>> import codecs 334 >>> import codecs
260 >>> encoder = codecs.getincrementalencoder('hz')('strict') 335 >>> encoder = codecs.getincrementalencoder('hz')('strict')
261 >>> b''.join(encoder.encode(x) for x in '\u52ff\u65bd\u65bc\u4eba\u3002 Bye. ') 336 >>> b''.join(encoder.encode(x) for x in '\u52ff\u65bd\u65bc\u4eba\u3002 Bye. ')
262 b'~{NpJ)l6HK!#~} Bye.' 337 b'~{NpJ)l6HK!#~} Bye.'
263 338
264 This example gives ``b'~{Np~}~{J)~}~{l6~}~{HK~}~{!#~} Bye.'`` with older Python 339 This example gives ``b'~{Np~}~{J)~}~{l6~}~{HK~}~{!#~} Bye.'`` with older Python
265 versions. 340 versions.
266 341
267 (:issue:`12100`) 342 (:issue:`12100`)
268 343
344 The ``unicode_internal`` codec has been deprecated.
345
269 crypt 346 crypt
270 ----- 347 -----
271 348
272 Addition of salt and modular crypt format and the :func:`~crypt.mksalt` 349 Addition of salt and modular crypt format and the :func:`~crypt.mksalt`
273 function to the :mod:`crypt` module. 350 function to the :mod:`crypt` module.
274 351
275 (:issue:`10924`) 352 (:issue:`10924`)
276 353
277 curses 354 curses
278 ------ 355 ------
279 356
357 * If the :mod:`curses` module is linked to the ncursesw library, use Unicode
358 functions when Unicode strings or characters are passed (e.g.
359 :c:func:`waddwstr`), and bytes functions otherwise (e.g. :c:func:`waddstr`).
360 * Use the locale encoding instead of ``utf-8`` to encode Unicode strings.
361 * :class:`curses.window` has a new :attr:`curses.window.encoding` attribute.
280 * The :class:`curses.window` class has a new :meth:`~curses.window.get_wch` 362 * The :class:`curses.window` class has a new :meth:`~curses.window.get_wch`
281 method to get a wide character 363 method to get a wide character
282 * The :mod:`curses` module has a new :meth:`~curses.unget_wch` function to 364 * The :mod:`curses` module has a new :meth:`~curses.unget_wch` function to
283 push a wide character so the next :meth:`~curses.window.get_wch` will return 365 push a wide character so the next :meth:`~curses.window.get_wch` will return
284 it 366 it
285 367
286 (Contributed by Iñigo Serna in :issue:`6755`) 368 (Contributed by Iñigo Serna in :issue:`6755`)
287 369
370 abc
371 ---
372
373 Improved support for abstract base classes containing descriptors composed with
374 abstract methods. The recommended approach to declaring abstract descriptors is
375 now to provide :attr:`__isabstractmethod__` as a dynamically updated
376 property. The built-in descriptors have been updated accordingly.
377
378 * :class:`abc.abstractproperty` has been deprecated, use :class:`property`
379 with :func:`abc.abstractmethod` instead.
380 * :class:`abc.abstractclassmethod` has been deprecated, use
381 :class:`classmethod` with :func:`abc.abstractmethod` instead.
382 * :class:`abc.abstractstaticmethod` has been deprecated, use
383 :class:`staticmethod` with :func:`abc.abstractmethod` instead.
384
385 (Contributed by Darren Dale in :issue:`11610`)
386
288 faulthandler 387 faulthandler
289 ------------ 388 ------------
290 389
291 New module: :mod:`faulthandler`. 390 New module: :mod:`faulthandler`.
292 391
293 * :envvar:`PYTHONFAULTHANDLER` 392 * :envvar:`PYTHONFAULTHANDLER`
294 * :option:`-X` ``faulthandler`` 393 * :option:`-X` ``faulthandler``
295 394
296 time 395 time
297 ---- 396 ----
298 397
299 * The :mod:`time` module has new :func:`~time.clock_getres` and 398 * The :mod:`time` module has new :func:`~time.clock_getres` and
300 :func:`~time.clock_gettime` functions and ``CLOCK_xxx`` constants. 399 :func:`~time.clock_gettime` functions and ``CLOCK_xxx`` constants.
301 :func:`~time.clock_gettime` can be used with :data:`time.CLOCK_MONOTONIC` to 400 :func:`~time.clock_gettime` can be used with :data:`time.CLOCK_MONOTONIC` to
302 get a monotonic clock. 401 get a monotonic clock.
303 402
304 (Contributed by Victor Stinner in :issue:`10278`) 403 (Contributed by Victor Stinner in :issue:`10278`)
305 404
306 405
307 ftplib 406 ftplib
308 ------ 407 ------
309 408
310 The :class:`~ftplib.FTP_TLS` class now provides a new 409 The :class:`~ftplib.FTP_TLS` class now provides a new
311 :func:`~ftplib.FTP_TLS.ccc` function to revert control channel back to 410 :func:`~ftplib.FTP_TLS.ccc` function to revert control channel back to
312 plaintext. This can be useful to take advantage of firewalls that know how to 411 plaintext. This can be useful to take advantage of firewalls that know how to
313 handle NAT with non-secure FTP without opening fixed ports. 412 handle NAT with non-secure FTP without opening fixed ports.
314 413
315 (Contributed by Giampaolo Rodolà in :issue:`12139`) 414 (Contributed by Giampaolo Rodolà in :issue:`12139`)
415
416
417 imaplib
418 -------
419
420 The :class:`~imaplib.IMAP4_SSL` constructor now accepts an SSLContext
421 parameter to control parameters of the secure channel.
422
423 (Contributed by Sijin Joseph in :issue:`8808`)
424
425
426 io
427 --
428
429 The :func:`~io.open` function has a new ``'x'`` mode that can be used to create
430 a new file, and raise a :exc:`FileExistsError` if the file already exists.
431
432 (Contributed by David Townshend in :issue:`12760`)
433
434
435 lzma
436 ----
437
438 The newly-added :mod:`lzma` module provides data compression and decompression
439 using the LZMA algorithm, including support for the ``.xz`` and ``.lzma``
440 file formats.
441
442 (Contributed by Nadeem Vawda and Per Øyvind Karlsen in :issue:`6715`)
316 443
317 444
318 math 445 math
319 ---- 446 ----
320 447
321 The :mod:`math` module has a new function: 448 The :mod:`math` module has a new function:
322 449
323 * :func:`~math.log2`: return the base-2 logarithm of *x* 450 * :func:`~math.log2`: return the base-2 logarithm of *x*
324 (Written by Mark Dickinson in :issue:`11888`). 451 (Written by Mark Dickinson in :issue:`11888`).
325 452
(...skipping 178 matching lines...) Expand 10 before | Expand all | Expand 10 after
504 631
505 (Contributed by David Watson in :issue:`6560`, based on an earlier patch by 632 (Contributed by David Watson in :issue:`6560`, based on an earlier patch by
506 Heiko Wundram) 633 Heiko Wundram)
507 634
508 * The :class:`~socket.socket` class now supports the PF_CAN protocol family 635 * The :class:`~socket.socket` class now supports the PF_CAN protocol family
509 (http://en.wikipedia.org/wiki/Socketcan), on Linux 636 (http://en.wikipedia.org/wiki/Socketcan), on Linux
510 (http://lwn.net/Articles/253425). 637 (http://lwn.net/Articles/253425).
511 638
512 (Contributed by Matthias Fuchs, updated by Tiago Gonçalves in :issue:`10141`) 639 (Contributed by Matthias Fuchs, updated by Tiago Gonçalves in :issue:`10141`)
513 640
641 * The :class:`~socket.socket` class now supports the PF_RDS protocol family
642 (http://en.wikipedia.org/wiki/Reliable_Datagram_Sockets and
643 http://oss.oracle.com/projects/rds/).
514 644
515 ssl 645 ssl
516 --- 646 ---
517 647
518 The :mod:`ssl` module has new functions: 648 * The :mod:`ssl` module has two new random generation functions:
519 649
520 * :func:`~ssl.RAND_bytes`: generate cryptographically strong 650 * :func:`~ssl.RAND_bytes`: generate cryptographically strong
521 pseudo-random bytes. 651 pseudo-random bytes.
522 * :func:`~ssl.RAND_pseudo_bytes`: generate pseudo-random bytes. 652 * :func:`~ssl.RAND_pseudo_bytes`: generate pseudo-random bytes.
653
654 (Contributed by Victor Stinner in :issue:`12049`)
655
656 * The :mod:`ssl` module now exposes a finer-grained exception hierarchy
657 in order to make it easier to inspect the various kinds of errors.
658
659 (Contributed by Antoine Pitrou in :issue:`11183`)
660
661 * :meth:`~ssl.SSLContext.load_cert_chain` now accepts a *password* argument
662 to be used if the private key is encrypted.
663
664 (Contributed by Adam Simpkins in :issue:`12803`)
665
666 * Diffie-Hellman key exchange, both regular and Elliptic Curve-based, is
667 now supported through the :meth:`~ssl.SSLContext.load_dh_params` and
668 :meth:`~ssl.SSLContext.set_ecdh_curve` methods.
669
670 (Contributed by Antoine Pitrou in :issue:`13626` and :issue:`13627`)
671
672 * SSL sockets have a new :meth:`~ssl.SSLSocket.get_channel_binding` method
673 allowing the implementation of certain authentication mechanisms such as
674 SCRAM-SHA-1-PLUS.
675
676 (Contributed by Jacek Konieczny in :issue:`12551`)
677
678 * You can query the SSL compression algorithm used by an SSL socket, thanks
679 to its new :meth:`~ssl.SSLSocket.compression` method.
680
681 (Contributed by Antoine Pitrou in :issue:`13634`)
523 682
524 683
525 shutil 684 shutil
526 ------ 685 ------
527 686
528 * The :mod:`shutil` module has these new fuctions: 687 * The :mod:`shutil` module has these new fuctions:
529 688
530 * :func:`~shutil.disk_usage`: provides total, used and free disk space 689 * :func:`~shutil.disk_usage`: provides total, used and free disk space
531 statistics. (Contributed by Giampaolo Rodolà in :issue:`12442`) 690 statistics. (Contributed by Giampaolo Rodolà in :issue:`12442`)
532 * :func:`~shutil.chown`: allows one to change user and/or group of the given 691 * :func:`~shutil.chown`: allows one to change user and/or group of the given
533 path also specifying the user/group names and not only their numeric 692 path also specifying the user/group names and not only their numeric
534 ids. (Contributed by Sandro Tosi in :issue:`12191`) 693 ids. (Contributed by Sandro Tosi in :issue:`12191`)
535 694
695 smtplib
696 -------
697
698 The :class:`~smtplib.SMTP_SSL` constructor and the :meth:`~smtplib.SMTP.starttls `
699 method now accept an SSLContext parameter to control parameters of the secure
700 channel.
701
702 (Contributed by Kasun Herath in :issue:`8809`)
703
536 urllib 704 urllib
537 ------ 705 ------
538 706
539 The :class:`~urllib.request.Request` class, now accepts a *method* argument 707 The :class:`~urllib.request.Request` class, now accepts a *method* argument
540 used by :meth:`~urllib.request.Request.get_method` to determine what HTTP method 708 used by :meth:`~urllib.request.Request.get_method` to determine what HTTP method
541 should be used. For example, this will send a ``'HEAD'`` request:: 709 should be used. For example, this will send a ``'HEAD'`` request::
542 710
543 >>> urlopen(Request('http://www.python.org', method='HEAD')) 711 >>> urlopen(Request('http://www.python.org', method='HEAD'))
544 712
545 (:issue:`1673007`) 713 (:issue:`1673007`)
546 714
715 sched
716 -----
717
718 * :meth:`~sched.scheduler.run` now accepts a *blocking* parameter which when
719 set to False makes the method execute the scheduled events due to expire
720 soonest (if any) and then return immediately.
721 This is useful in case you want to use the :class:`~sched.scheduler` in
722 non-blocking applications. (Contributed by Giampaolo Rodolà in :issue:`13449` )
723
724 * :class:`~sched.scheduler` class can now be safely used in multi-threaded
725 environments. (Contributed by Josiah Carlson and Giampaolo Rodolà in
726 :issue:`8684`)
727
728 * *timefunc* and *delayfunct* parameters of :class:`~sched.scheduler` class
729 constructor are now optional and defaults to :func:`time.time` and
730 :func:`time.sleep` respectively. (Contributed by Chris Clark in
731 :issue:`13245`)
732
733 * :meth:`~sched.scheduler.enter` and :meth:`~sched.scheduler.enterabs`
734 *argument* parameter is now optional. (Contributed by Chris Clark in
735 :issue:`13245`)
736
737 * :meth:`~sched.scheduler.enter` and :meth:`~sched.scheduler.enterabs`
738 now accept a *kwargs* parameter. (Contributed by Chris Clark in
739 :issue:`13245`)
740
547 Optimizations 741 Optimizations
548 ============= 742 =============
549 743
550 Major performance enhancements have been added: 744 Major performance enhancements have been added:
551 745
552 * Stub 746 * Thanks to the :pep:`393`, some operations on Unicode strings has been optimize d:
747
748 * the memory footprint is divided by 2 to 4 depending on the text
749 * encode an ASCII string to UTF-8 doesn't need to encode characters anymore,
750 the UTF-8 representation is shared with the ASCII representation
751 * the UTF-8 encoder has been optimized
752 * repeating a single ASCII letter and getting a substring of a ASCII strings
753 is 4 times faster
553 754
554 755
555 Build and C API Changes 756 Build and C API Changes
556 ======================= 757 =======================
557 758
558 Changes to Python's build process and to the C API include: 759 Changes to Python's build process and to the C API include:
559 760
560 * Stub 761 * The :pep:`393` added new Unicode types, macros and functions:
561 762
763 * High-level API:
764
765 * :c:func:`PyUnicode_CopyCharacters`
766 * :c:func:`PyUnicode_FindChar`
767 * :c:func:`PyUnicode_GetLength`, :c:macro:`PyUnicode_GET_LENGTH`
768 * :c:func:`PyUnicode_New`
769 * :c:func:`PyUnicode_Substring`
770 * :c:func:`PyUnicode_ReadChar`, :c:func:`PyUnicode_WriteChar`
771
772 * Low-level API:
773
774 * :c:type:`Py_UCS1`, :c:type:`Py_UCS2`, :c:type:`Py_UCS4` types
775 * :c:type:`PyASCIIObject` and :c:type:`PyCompactUnicodeObject` structures
776 * :c:macro:`PyUnicode_READY`
777 * :c:func:`PyUnicode_FromKindAndData`
778 * :c:func:`PyUnicode_AsUCS4`, :c:func:`PyUnicode_AsUCS4Copy`
779 * :c:macro:`PyUnicode_DATA`, :c:macro:`PyUnicode_1BYTE_DATA`,
780 :c:macro:`PyUnicode_2BYTE_DATA`, :c:macro:`PyUnicode_4BYTE_DATA`
781 * :c:macro:`PyUnicode_KIND` with :c:type:`PyUnicode_Kind` enum:
782 :c:data:`PyUnicode_WCHAR_KIND`, :c:data:`PyUnicode_1BYTE_KIND`,
783 :c:data:`PyUnicode_2BYTE_KIND`, :c:data:`PyUnicode_4BYTE_KIND`
784 * :c:macro:`PyUnicode_READ`, :c:macro:`PyUnicode_READ_CHAR`, :c:macro:`PyUni code_WRITE`
785 * :c:macro:`PyUnicode_MAX_CHAR_VALUE`
786
787
788
789 Deprecated
790 ==========
562 791
563 Unsupported Operating Systems 792 Unsupported Operating Systems
564 ============================= 793 -----------------------------
565 794
566 OS/2 and VMS are no longer supported due to the lack of a maintainer. 795 OS/2 and VMS are no longer supported due to the lack of a maintainer.
567 796
568 Windows 2000 and Windows platforms which set ``COMSPEC`` to ``command.com`` 797 Windows 2000 and Windows platforms which set ``COMSPEC`` to ``command.com``
569 are no longer supported due to maintenance burden. 798 are no longer supported due to maintenance burden.
570 799
571 800
801 Deprecated Python modules, functions and methods
802 ------------------------------------------------
803
804 * The :mod:`packaging` module replaces the :mod:`distutils` module
805 * The ``unicode_internal`` codec has been deprecated because of the
806 :pep:`393`, use UTF-8, UTF-16 (``utf-16-le`` or ``utf-16-le``), or UTF-32
807 (``utf-32-le`` or ``utf-32-le``)
808 * :meth:`ftplib.FTP.nlst` and :meth:`ftplib.FTP.dir`: use
809 :meth:`ftplib.FTP.mlsd`
810 * :func:`platform.popen`: use the :mod:`subprocess` module. Check especially
811 the :ref:`subprocess-replacements` section.
812 * :issue:`13374`: The Windows bytes API has been deprecated in the :mod:`os`
813 module. Use Unicode filenames, instead of bytes filenames, to not depend on
814 the ANSI code page anymore and to support any filename.
815
816
817 Deprecated functions and types of the C API
818 -------------------------------------------
819
820 The :c:type:`Py_UNICODE` has been deprecated by the :pep:`393` and will be
821 removed in Python 4. All functions using this type are deprecated:
822
823 Unicode functions and methods using :c:type:`Py_UNICODE` and
824 :c:type:`Py_UNICODE*` types:
825
826 * :c:macro:`PyUnicode_FromUnicode`: use :c:func:`PyUnicode_FromWideChar` or
827 :c:func:`PyUnicode_FromKindAndData`
828 * :c:macro:`PyUnicode_AS_UNICODE`, :c:func:`PyUnicode_AsUnicode`,
829 :c:func:`PyUnicode_AsUnicodeAndSize`: use :c:func:`PyUnicode_AsWideCharString `
830 * :c:macro:`PyUnicode_AS_DATA`: use :c:macro:`PyUnicode_DATA` with
831 :c:macro:`PyUnicode_READ` and :c:macro:`PyUnicode_WRITE`
832 * :c:macro:`PyUnicode_GET_SIZE`, :c:func:`PyUnicode_GetSize`: use
833 :c:macro:`PyUnicode_GET_LENGTH` or :c:func:`PyUnicode_GetLength`
834 * :c:macro:`PyUnicode_GET_DATA_SIZE`: use
835 ``PyUnicode_GET_LENGTH(str) * PyUnicode_KIND(str)`` (only work on ready
836 strings)
837 * :c:func:`PyUnicode_AsUnicodeCopy`: use :c:func:`PyUnicode_AsUCS4Copy` or
838 :c:func:`PyUnicode_AsWideCharString`
839 * :c:func:`PyUnicode_GetMax`
840
841
842 Functions and macros manipulating Py_UNICODE* strings:
843
844 * :c:macro:`Py_UNICODE_strlen`: use :c:func:`PyUnicode_GetLength` or
845 :c:macro:`PyUnicode_GET_LENGTH`
846 * :c:macro:`Py_UNICODE_strcat`: use :c:func:`PyUnicode_CopyCharacters` or
847 :c:func:`PyUnicode_FromFormat`
848 * :c:macro:`Py_UNICODE_strcpy`, :c:macro:`Py_UNICODE_strncpy`,
849 :c:macro:`Py_UNICODE_COPY`: use :c:func:`PyUnicode_CopyCharacters` or
850 :c:func:`PyUnicode_Substring`
851 * :c:macro:`Py_UNICODE_strcmp`: use :c:func:`PyUnicode_Compare`
852 * :c:macro:`Py_UNICODE_strncmp`: use :c:func:`PyUnicode_Tailmatch`
853 * :c:macro:`Py_UNICODE_strchr`, :c:macro:`Py_UNICODE_strrchr`: use
854 :c:func:`PyUnicode_FindChar`
855 * :c:macro:`Py_UNICODE_FILL`: use :c:func:`PyUnicode_Fill`
856 * :c:macro:`Py_UNICODE_MATCH`
857
858 Encoders:
859
860 * :c:func:`PyUnicode_Encode`: use :c:func:`PyUnicode_AsEncodedObject`
861 * :c:func:`PyUnicode_EncodeUTF7`
862 * :c:func:`PyUnicode_EncodeUTF8`: use :c:func:`PyUnicode_AsUTF8` or
863 :c:func:`PyUnicode_AsUTF8String`
864 * :c:func:`PyUnicode_EncodeUTF32`
865 * :c:func:`PyUnicode_EncodeUTF16`
866 * :c:func:`PyUnicode_EncodeUnicodeEscape:` use
867 :c:func:`PyUnicode_AsUnicodeEscapeString`
868 * :c:func:`PyUnicode_EncodeRawUnicodeEscape:` use
869 :c:func:`PyUnicode_AsRawUnicodeEscapeString`
870 * :c:func:`PyUnicode_EncodeLatin1`: use :c:func:`PyUnicode_AsLatin1String`
871 * :c:func:`PyUnicode_EncodeASCII`: use :c:func:`PyUnicode_AsASCIIString`
872 * :c:func:`PyUnicode_EncodeCharmap`
873 * :c:func:`PyUnicode_TranslateCharmap`
874 * :c:func:`PyUnicode_EncodeMBCS`: use :c:func:`PyUnicode_AsMBCSString` or
875 :c:func:`PyUnicode_EncodeCodePage` (with ``CP_ACP`` code_page)
876 * :c:func:`PyUnicode_EncodeDecimal`,
877 :c:func:`PyUnicode_TransformDecimalToASCII`
878
879
572 Porting to Python 3.3 880 Porting to Python 3.3
573 ===================== 881 =====================
574 882
575 This section lists previously described changes and other bugfixes 883 This section lists previously described changes and other bugfixes
576 that may require changes to your code. 884 that may require changes to your code.
577 885
578 Porting Python code 886 Porting Python code
579 ------------------- 887 -------------------
580 888
581 * Issue #12326: On Linux, sys.platform doesn't contain the major version 889 * :issue:`12326`: On Linux, sys.platform doesn't contain the major version
582 anymore. It is now always 'linux', instead of 'linux2' or 'linux3' depending 890 anymore. It is now always 'linux', instead of 'linux2' or 'linux3' depending
583 on the Linux version used to build Python. Replace sys.platform == 'linux2' 891 on the Linux version used to build Python. Replace sys.platform == 'linux2'
584 with sys.platform.startswith('linux'), or directly sys.platform == 'linux' if 892 with sys.platform.startswith('linux'), or directly sys.platform == 'linux' if
585 you don't need to support older Python versions. 893 you don't need to support older Python versions.
586 894
587 Porting C code 895 Porting C code
588 -------------- 896 --------------
589 897
590 * Due to :ref:`PEP 393 <pep-393>`, the :c:type:`Py_UNICODE` type and all 898 * Due to :ref:`PEP 393 <pep-393>`, the :c:type:`Py_UNICODE` type and all
591 functions using this type are deprecated (but will stay available for 899 functions using this type are deprecated (but will stay available for
592 at least five years). If you were using low-level Unicode APIs to 900 at least five years). If you were using low-level Unicode APIs to
593 construct and access unicode objects and you want to benefit of the 901 construct and access unicode objects and you want to benefit of the
594 memory footprint reduction provided by the PEP 393, you have to convert 902 memory footprint reduction provided by the PEP 393, you have to convert
595 your code to the new :doc:`Unicode API <../c-api/unicode>`. 903 your code to the new :doc:`Unicode API <../c-api/unicode>`.
596 904
597 However, if you only have been using high-level functions such as 905 However, if you only have been using high-level functions such as
598 :c:func:`PyUnicode_Concat()`, :c:func:`PyUnicode_Join` or 906 :c:func:`PyUnicode_Concat()`, :c:func:`PyUnicode_Join` or
599 :c:func:`PyUnicode_FromFormat()`, your code will automatically take 907 :c:func:`PyUnicode_FromFormat()`, your code will automatically take
600 advantage of the new unicode representations. 908 advantage of the new unicode representations.
601 909
602 Other issues 910 Other issues
603 ------------ 911 ------------
604 912
605 .. Issue #11591: When :program:`python` was started with :option:`-S`, 913 .. Issue #11591: When :program:`python` was started with :option:`-S`,
606 ``import site`` will not add site-specific paths to the module search 914 ``import site`` will not add site-specific paths to the module search
607 paths. In previous versions, it did. See changeset for doc changes in 915 paths. In previous versions, it did. See changeset for doc changes in
608 various files. Contributed by Carl Meyer with editions by Éric Araujo. 916 various files. Contributed by Carl Meyer with editions by Éric Araujo.
609 917
610 .. Issue #10998: -Q command-line flags are related artifacts have been 918 .. Issue #10998: the -Q command-line flag and related artifacts have been
611 removed. Code checking sys.flags.division_warning will need updating. 919 removed. Code checking sys.flags.division_warning will need updating.
612 Contributed by Éric Araujo. 920 Contributed by Éric Araujo.
613
LEFTRIGHT

RSS Feeds Recent Issues | This issue
This is Rietveld 894c83f36cb7+