Raymond Hettinger | 6e6565b | 2009-06-28 20:56:11 +0000 | [diff] [blame] | 1 | **************************** |
Raymond Hettinger | f558ddd | 2009-06-28 21:37:08 +0000 | [diff] [blame] | 2 | What's New In Python 3.2 |
Raymond Hettinger | 6e6565b | 2009-06-28 20:56:11 +0000 | [diff] [blame] | 3 | **************************** |
| 4 | |
| 5 | :Author: Raymond Hettinger |
| 6 | :Release: |release| |
| 7 | :Date: |today| |
| 8 | |
| 9 | .. $Id$ |
| 10 | Rules for maintenance: |
| 11 | |
| 12 | * Anyone can add text to this document. Do not spend very much time |
| 13 | on the wording of your changes, because your text will probably |
| 14 | get rewritten to some degree. |
| 15 | |
| 16 | * The maintainer will go through Misc/NEWS periodically and add |
| 17 | changes; it's therefore more important to add your changes to |
| 18 | Misc/NEWS than to this file. |
| 19 | |
| 20 | * This is not a complete list of every single change; completeness |
| 21 | is the purpose of Misc/NEWS. Some changes I consider too small |
| 22 | or esoteric to include. If such a change is added to the text, |
| 23 | I'll just remove it. (This is another reason you shouldn't spend |
| 24 | too much time on writing your addition.) |
| 25 | |
| 26 | * If you want to draw your new text to the attention of the |
| 27 | maintainer, add 'XXX' to the beginning of the paragraph or |
| 28 | section. |
| 29 | |
| 30 | * It's OK to just add a fragmentary note about a change. For |
| 31 | example: "XXX Describe the transmogrify() function added to the |
| 32 | socket module." The maintainer will research the change and |
| 33 | write the necessary text. |
| 34 | |
| 35 | * You can comment out your additions if you like, but it's not |
| 36 | necessary (especially when a final release is some months away). |
| 37 | |
| 38 | * Credit the author of a patch or bugfix. Just the name is |
| 39 | sufficient; the e-mail address isn't necessary. |
| 40 | |
| 41 | * It's helpful to add the bug/patch number as a comment: |
| 42 | |
| 43 | % Patch 12345 |
| 44 | XXX Describe the transmogrify() function added to the socket |
| 45 | module. |
| 46 | (Contributed by P.Y. Developer.) |
| 47 | |
| 48 | This saves the maintainer the effort of going through the SVN log |
| 49 | when researching a change. |
| 50 | |
| 51 | This article explains the new features in Python 3.2, compared to 3.1. |
| 52 | |
| 53 | |
| 54 | PEP XXX: Stub |
| 55 | ============= |
| 56 | |
| 57 | |
| 58 | Other Language Changes |
| 59 | ====================== |
| 60 | |
| 61 | Some smaller changes made to the core Python language are: |
| 62 | |
| 63 | * Stub |
| 64 | |
| 65 | |
| 66 | New, Improved, and Deprecated Modules |
| 67 | ===================================== |
| 68 | |
Raymond Hettinger | aed05eb | 2010-08-02 01:43:41 +0000 | [diff] [blame] | 69 | * The functools module now includes two new decorators for caching function |
| 70 | calls, :func:`functools.lru_cache` and :func:`functools.lfu_cache`. These can |
| 71 | save repeated queries to an external resource whenever the results are |
| 72 | expected to be the same. |
| 73 | |
Raymond Hettinger | 86f9613 | 2010-08-06 23:23:49 +0000 | [diff] [blame^] | 74 | For example, adding a caching decorator to a database query function can save |
| 75 | database accesses for popular searches:: |
Raymond Hettinger | aed05eb | 2010-08-02 01:43:41 +0000 | [diff] [blame] | 76 | |
| 77 | @functools.lfu_cache(maxsize=50) |
| 78 | def get_phone_number(name): |
| 79 | c = conn.cursor() |
| 80 | c.execute('SELECT phonenumber FROM phonelist WHERE name=?', (name,)) |
| 81 | return c.fetchone()[0] |
| 82 | |
Raymond Hettinger | 86f9613 | 2010-08-06 23:23:49 +0000 | [diff] [blame^] | 83 | The caches support two strategies for limiting their size to *maxsize*. The |
| 84 | LFU (least-frequently-used) cache works bests when popular queries remain the |
| 85 | same over time. In contrast, the LRU (least-recently-used) cache works best |
| 86 | query popularity changes over time (for example, the most popular news |
| 87 | articles change each day as newer articles are added). |
Raymond Hettinger | aed05eb | 2010-08-02 01:43:41 +0000 | [diff] [blame] | 88 | |
Raymond Hettinger | 86f9613 | 2010-08-06 23:23:49 +0000 | [diff] [blame^] | 89 | The two caching decorators can be composed (nested) to handle hybrid cases. |
Raymond Hettinger | aed05eb | 2010-08-02 01:43:41 +0000 | [diff] [blame] | 90 | For example, music searches can reflect both long-term patterns (popular |
| 91 | classics) and short-term trends (new releases):: |
| 92 | |
Raymond Hettinger | 86f9613 | 2010-08-06 23:23:49 +0000 | [diff] [blame^] | 93 | @functools.lfu_cache(maxsize=500) |
| 94 | @functools.lru_cache(maxsize=100) |
| 95 | def find_lyrics(song): |
| 96 | query = 'http://www.example.com/songlist/%s' % urllib.quote(song) |
| 97 | page = urllib.urlopen(query).read() |
| 98 | return parse_lyrics(page) |
| 99 | |
| 100 | To help with choosing an effective cache size, the wrapped function |
| 101 | is instrumented with two attributes 'hits' and 'misses':: |
| 102 | |
| 103 | >>> for song in user_requests: |
| 104 | ... find_lyrics(song) |
| 105 | >>> print find_lyrics.hits |
| 106 | 4805 |
| 107 | >>> print find_lyrics.misses |
| 108 | 980 |
Raymond Hettinger | aed05eb | 2010-08-02 01:43:41 +0000 | [diff] [blame] | 109 | |
| 110 | (Contributed by Raymond Hettinger) |
| 111 | |
Antoine Pitrou | d67075e | 2010-07-31 22:48:02 +0000 | [diff] [blame] | 112 | * The previously deprecated :func:`contextlib.nested` function has been |
| 113 | removed in favor of a plain :keyword:`with` statement which can |
| 114 | accept multiple context managers. The latter technique is faster |
| 115 | (because it is built-in), and it does a better job finalizing multiple |
| 116 | context managers when one of them raises an exception. |
| 117 | |
| 118 | (Contributed by Georg Brandl and Mattias Brändström; |
| 119 | `appspot issue 53094 <http://codereview.appspot.com/53094>`_.) |
| 120 | |
Giampaolo Rodolà | bd576b7 | 2010-05-10 14:53:29 +0000 | [diff] [blame] | 121 | * The :class:`ftplib.FTP` class now supports the context manager protocol |
| 122 | (Contributed by Tarek Ziadé and Giampaolo Rodolà; :issue:`4972`.) |
| 123 | |
Antoine Pitrou | d67075e | 2010-07-31 22:48:02 +0000 | [diff] [blame] | 124 | * The :func:`shutil.copytree` function has two new options: |
| 125 | |
| 126 | * *ignore_dangling_symlinks*: when ``symlinks=False`` (meaning that the |
| 127 | function copies the file pointed to by the symlink, not the symlink |
Georg Brandl | 7cb1319 | 2010-08-03 12:06:29 +0000 | [diff] [blame] | 128 | itself) this option will silence the error raised if the file doesn't |
Antoine Pitrou | d67075e | 2010-07-31 22:48:02 +0000 | [diff] [blame] | 129 | exist. |
| 130 | |
| 131 | * *copy_function*: a callable that will be used to copy files. |
| 132 | :func:`shutil.copy2` is used by default. |
| 133 | |
| 134 | (Contributed by Tarek Ziadé.) |
| 135 | |
| 136 | * The *sqlite3* module has some new features: |
| 137 | |
| 138 | * XXX *enable_load_extension* |
| 139 | |
| 140 | * XXX *load_extension* |
| 141 | |
| 142 | * New :class:`~sqlite3.Connection` attribute |
| 143 | :attr:`~sqlite3.Connection.in_transaction` is :const:`True` when there |
| 144 | are uncommitted changes, and :const:`False` otherwise. (Contributed |
| 145 | by R. David Murray and Shashwat Anand, :issue:`8845`.) |
| 146 | |
Antoine Pitrou | 4f2a0a8 | 2010-07-31 18:08:33 +0000 | [diff] [blame] | 147 | * The :mod:`ssl` module has a new class, :class:`~ssl.SSLContext` which |
| 148 | serves as a container for various persistent SSL data, such as protocol |
| 149 | settings, certificates, private keys, and various other options. |
| 150 | The :meth:`~ssl.SSLContext.wrap_socket` method allows to create an |
| 151 | SSL socket from such an SSL context. |
| 152 | (Added by Antoine Pitrou; :issue:`8550`.) |
| 153 | |
| 154 | The :func:`ssl.wrap_socket` constructor function now takes a |
| 155 | *ciphers* argument that's a string listing the encryption algorithms |
| 156 | to be allowed; the format of the string is described |
| 157 | `in the OpenSSL documentation |
| 158 | <http://www.openssl.org/docs/apps/ciphers.html#CIPHER_LIST_FORMAT>`__. |
| 159 | (Added by Antoine Pitrou; :issue:`8322`.) |
| 160 | |
| 161 | Various options have been added to the :mod:`ssl` module, such as |
| 162 | :data:`~ssl.OP_NO_SSLv2` which allows to force disabling of the insecure |
| 163 | and obsolete SSLv2 protocol. |
| 164 | (Added by Antoine Pitrou; :issue:`4870`.) |
| 165 | |
| 166 | Another change makes the extension load all of OpenSSL's ciphers and |
| 167 | digest algorithms so that they're all available. Some SSL |
| 168 | certificates couldn't be verified, reporting an "unknown algorithm" |
| 169 | error. (Reported by Beda Kosata, and fixed by Antoine Pitrou; |
| 170 | :issue:`8484`.) |
| 171 | |
| 172 | The version of OpenSSL being used is now available as the module |
| 173 | attributes :data:`ssl.OPENSSL_VERSION` (a string), |
| 174 | :data:`ssl.OPENSSL_VERSION_INFO` (a 5-tuple), and |
| 175 | :data:`ssl.OPENSSL_VERSION_NUMBER` (an integer). (Added by Antoine |
| 176 | Pitrou; :issue:`8321`.) |
| 177 | |
Raymond Hettinger | f558ddd | 2009-06-28 21:37:08 +0000 | [diff] [blame] | 178 | * The previously deprecated :func:`string.maketrans` function has been |
| 179 | removed in favor of the static methods, :meth:`bytes.maketrans` and |
| 180 | :meth:`bytearray.maketrans`. This change solves the confusion around which |
| 181 | types were supported by the :mod:`string` module. Now, :class:`str`, |
| 182 | :class:`bytes`, and :class:`bytearray` each have their own **maketrans** and |
| 183 | **translate** methods with intermediate translation tables of the |
| 184 | appropriate type. |
Raymond Hettinger | 6e6565b | 2009-06-28 20:56:11 +0000 | [diff] [blame] | 185 | |
Raymond Hettinger | f558ddd | 2009-06-28 21:37:08 +0000 | [diff] [blame] | 186 | (Contributed by Georg Brandl; :issue:`5675`.) |
| 187 | |
R. David Murray | d35251d | 2010-06-01 01:32:12 +0000 | [diff] [blame] | 188 | |
Antoine Pitrou | d42bc51 | 2009-11-10 23:18:31 +0000 | [diff] [blame] | 189 | Multi-threading |
| 190 | =============== |
| 191 | |
| 192 | * The mechanism for serializing execution of concurrently running Python |
| 193 | threads (generally known as the GIL or Global Interpreter Lock) has been |
| 194 | rewritten. Among the objectives were more predictable switching intervals |
| 195 | and reduced overhead due to lock contention and the number of ensuing |
| 196 | system calls. The notion of a "check interval" to allow thread switches |
| 197 | has been abandoned and replaced by an absolute duration expressed in |
| 198 | seconds. This parameter is tunable through :func:`sys.setswitchinterval()`. |
| 199 | It currently defaults to 5 milliseconds. |
| 200 | |
| 201 | Additional details about the implementation can be read from a `python-dev |
| 202 | mailing-list message |
| 203 | <http://mail.python.org/pipermail/python-dev/2009-October/093321.html>`_ |
| 204 | (however, "priority requests" as exposed in this message have not been |
| 205 | kept for inclusion). |
| 206 | |
Georg Brandl | 5e73a81 | 2010-04-22 07:02:51 +0000 | [diff] [blame] | 207 | (Contributed by Antoine Pitrou.) |
Antoine Pitrou | d42bc51 | 2009-11-10 23:18:31 +0000 | [diff] [blame] | 208 | |
Antoine Pitrou | 5bab508 | 2009-11-13 22:58:45 +0000 | [diff] [blame] | 209 | * Recursive locks (created with the :func:`threading.RLock` API) now benefit |
| 210 | from a C implementation which makes them as fast as regular locks, and |
| 211 | between 10x and 15x faster than their previous pure Python implementation. |
| 212 | |
| 213 | (Contributed by Antoine Pitrou; :issue:`3001`.) |
| 214 | |
Antoine Pitrou | e95a9ff | 2010-05-04 23:31:41 +0000 | [diff] [blame] | 215 | * Regular and recursive locks now accept an optional *timeout* argument |
| 216 | to their ``acquire`` method. (Contributed by Antoine Pitrou; :issue:`7316`) |
| 217 | Similarly, :meth:`threading.Semaphore.acquire` also gains a *timeout* |
| 218 | argument. (Contributed by Torsten Landschoff; :issue:`850728`.) |
| 219 | |
Antoine Pitrou | d42bc51 | 2009-11-10 23:18:31 +0000 | [diff] [blame] | 220 | |
Raymond Hettinger | 6e6565b | 2009-06-28 20:56:11 +0000 | [diff] [blame] | 221 | Optimizations |
| 222 | ============= |
| 223 | |
| 224 | Major performance enhancements have been added: |
| 225 | |
| 226 | * Stub |
| 227 | |
| 228 | IDLE |
| 229 | ==== |
| 230 | |
| 231 | * Stub |
| 232 | |
| 233 | |
| 234 | Build and C API Changes |
| 235 | ======================= |
| 236 | |
| 237 | Changes to Python's build process and to the C API include: |
| 238 | |
| 239 | * Stub |
| 240 | |
| 241 | |
Raymond Hettinger | f558ddd | 2009-06-28 21:37:08 +0000 | [diff] [blame] | 242 | Porting to Python 3.2 |
Raymond Hettinger | 6e6565b | 2009-06-28 20:56:11 +0000 | [diff] [blame] | 243 | ===================== |
| 244 | |
| 245 | This section lists previously described changes and other bugfixes |
| 246 | that may require changes to your code: |
| 247 | |
Victor Stinner | dcb2403 | 2010-04-22 12:08:36 +0000 | [diff] [blame] | 248 | * bytearray objects cannot be used anymore as filenames: convert them to bytes |
| 249 | |
Victor Stinner | 25e8ec4 | 2010-06-25 00:02:38 +0000 | [diff] [blame] | 250 | * PyArg_Parse*() functions: |
Victor Stinner | 3dcb5ac | 2010-06-08 22:54:19 +0000 | [diff] [blame] | 251 | |
Victor Stinner | 25e8ec4 | 2010-06-25 00:02:38 +0000 | [diff] [blame] | 252 | * "t#" format has been removed: use "s#" or "s*" instead |
| 253 | * "w" and "w#" formats has been removed: use "w*" instead |
| 254 | |