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 |
Raymond Hettinger | 6f04adc | 2010-12-04 22:56:25 +0000 | [diff] [blame] | 14 | get rewritten. |
Raymond Hettinger | 6e6565b | 2009-06-28 20:56:11 +0000 | [diff] [blame] | 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 |
Georg Brandl | da0a211 | 2010-09-05 11:28:33 +0000 | [diff] [blame] | 39 | sufficient; the e-mail address isn't necessary. It's helpful to |
| 40 | add the issue number: |
Raymond Hettinger | 6e6565b | 2009-06-28 20:56:11 +0000 | [diff] [blame] | 41 | |
Éric Araujo | 4234ad4 | 2010-09-05 17:32:25 +0000 | [diff] [blame] | 42 | XXX Describe the transmogrify() function added to the socket |
| 43 | module. |
| 44 | |
| 45 | (Contributed by P.Y. Developer; :issue:`12345`.) |
Raymond Hettinger | 6e6565b | 2009-06-28 20:56:11 +0000 | [diff] [blame] | 46 | |
| 47 | This saves the maintainer the effort of going through the SVN log |
| 48 | when researching a change. |
| 49 | |
| 50 | This article explains the new features in Python 3.2, compared to 3.1. |
| 51 | |
Martin v. Löwis | 932e49e | 2010-12-04 13:49:32 +0000 | [diff] [blame] | 52 | PEP 384: Defining a Stable ABI |
Martin v. Löwis | 4d0d471 | 2010-12-03 20:14:31 +0000 | [diff] [blame] | 53 | ============================== |
| 54 | |
| 55 | In the past, extension modules built for one Python version were often |
| 56 | not usable with other Python versions. Particularly on Windows, every |
| 57 | feature release of Python required rebuilding all extension modules that |
| 58 | one wanted to use. This requirement was the result of the free access to |
| 59 | Python interpreter internals that extension modules could use. |
| 60 | |
| 61 | With Python 3.2, an alternative approach becomes available: extension |
Raymond Hettinger | 6f04adc | 2010-12-04 22:56:25 +0000 | [diff] [blame] | 62 | modules which restrict themselves to a limited API (by defining |
Martin v. Löwis | 4d0d471 | 2010-12-03 20:14:31 +0000 | [diff] [blame] | 63 | Py_LIMITED_API) cannot use many of the internals, but are constrained |
| 64 | to a set of API functions that are promised to be stable for several |
| 65 | releases. As a consequence, extension modules built for 3.2 in that |
| 66 | mode will also work with 3.3, 3.4, and so on. Extension modules that |
| 67 | make use of details of memory structures can still be built, but will |
| 68 | need to be recompiled for every feature release. |
| 69 | |
Raymond Hettinger | 6f04adc | 2010-12-04 22:56:25 +0000 | [diff] [blame] | 70 | .. seealso:: |
| 71 | |
Georg Brandl | 65b2eb9 | 2010-12-05 11:42:38 +0000 | [diff] [blame] | 72 | :pep:`384` - Defining a Stable ABI |
Raymond Hettinger | 6f04adc | 2010-12-04 22:56:25 +0000 | [diff] [blame] | 73 | PEP written by Martin von Loewis. |
| 74 | |
Raymond Hettinger | a5a3554 | 2010-12-05 00:39:18 +0000 | [diff] [blame] | 75 | PEP 389: Argparse Command Line Parsing Module |
| 76 | ============================================= |
| 77 | |
| 78 | A new module for command line parsing, :mod:`argparse`, was introduced to |
| 79 | overcome the limitations of :mod:`optparse` which did not provide support for |
| 80 | positional arguments (not just option), subcommands, required options and other |
Raymond Hettinger | 413abbc | 2010-12-05 07:06:47 +0000 | [diff] [blame] | 81 | common patterns of specifying and validating options. |
Raymond Hettinger | a5a3554 | 2010-12-05 00:39:18 +0000 | [diff] [blame] | 82 | |
| 83 | This module has already has wide-spread success in the community as a |
| 84 | third-party module. Being more fully featured than its predecessor, |
| 85 | :mod:`argparse`, is now the preferred module for command-line processing. The |
| 86 | older module is still being kept available because of the substantial amount of |
| 87 | legacy code that depends on it. |
| 88 | |
| 89 | .. XXX add examples that highlight the new features |
| 90 | |
| 91 | .. seealso:: |
| 92 | |
| 93 | :pep:`389` - New Command Line Parsing Module |
| 94 | PEP written by Steven Bethard. |
| 95 | |
Raymond Hettinger | 6e6565b | 2009-06-28 20:56:11 +0000 | [diff] [blame] | 96 | |
Éric Araujo | 4234ad4 | 2010-09-05 17:32:25 +0000 | [diff] [blame] | 97 | PEP 391: Dictionary Based Configuration for Logging |
| 98 | ==================================================== |
Raymond Hettinger | ef2335c | 2010-09-05 08:35:38 +0000 | [diff] [blame] | 99 | |
Raymond Hettinger | 92ba286 | 2010-09-06 01:16:46 +0000 | [diff] [blame] | 100 | The :mod:`logging` module provided two kinds of configuration, one style with |
| 101 | function calls for each option or another style driven by an external file saved |
| 102 | in a :mod:`ConfigParser` format. Those options did not provide the flexibility |
Georg Brandl | 9e75cad | 2010-09-06 06:45:47 +0000 | [diff] [blame] | 103 | to create configurations from JSON or YAML files, nor did they support |
Raymond Hettinger | 92ba286 | 2010-09-06 01:16:46 +0000 | [diff] [blame] | 104 | incremental configuration, which is needed for specifying logger options from a |
| 105 | command line. |
Raymond Hettinger | ef2335c | 2010-09-05 08:35:38 +0000 | [diff] [blame] | 106 | |
| 107 | To support a more flexible style, the module now offers |
Raymond Hettinger | 92ba286 | 2010-09-06 01:16:46 +0000 | [diff] [blame] | 108 | :func:`logging.config.dictConfig` for specifying logging configuration with |
| 109 | plain Python dictionaries. The configuration options include formatters, |
| 110 | handlers, filters, and loggers. Here's a working example of a configuration |
| 111 | dictionary:: |
Raymond Hettinger | ef2335c | 2010-09-05 08:35:38 +0000 | [diff] [blame] | 112 | |
Georg Brandl | da0a211 | 2010-09-05 11:28:33 +0000 | [diff] [blame] | 113 | {"version": 1, |
| 114 | "formatters": {"brief": {"format": "%(levelname)-8s: %(name)-15s: %(message)s"}, |
| 115 | "full": {"format": "%(asctime)s %(name)-15s %(levelname)-8s %(message)s"}, |
| 116 | }, |
| 117 | "handlers": {"console": { |
| 118 | "class": "logging.StreamHandler", |
| 119 | "formatter": "brief", |
| 120 | "level": "INFO", |
| 121 | "stream": "ext://sys.stdout"}, |
| 122 | "console_priority": { |
| 123 | "class": "logging.StreamHandler", |
| 124 | "formatter": "full", |
| 125 | "level": "ERROR", |
| 126 | "stream": "ext://sys.stderr"}, |
| 127 | }, |
| 128 | "root": {"level": "DEBUG", "handlers": ["console", "console_priority"]}} |
Raymond Hettinger | ef2335c | 2010-09-05 08:35:38 +0000 | [diff] [blame] | 129 | |
Raymond Hettinger | 92ba286 | 2010-09-06 01:16:46 +0000 | [diff] [blame] | 130 | |
Raymond Hettinger | 6f04adc | 2010-12-04 22:56:25 +0000 | [diff] [blame] | 131 | If that dictionary is stored in a file called :file:`conf.json`, it can loaded |
Raymond Hettinger | 92ba286 | 2010-09-06 01:16:46 +0000 | [diff] [blame] | 132 | and called with code like this:: |
| 133 | |
| 134 | >>> import logging.config |
| 135 | >>> logging.config.dictConfig(json.load(open('conf.json', 'rb'))) |
| 136 | >>> logging.info("Transaction completed normally") |
| 137 | >>> logging.critical("Abnormal termination") |
| 138 | |
Raymond Hettinger | ef2335c | 2010-09-05 08:35:38 +0000 | [diff] [blame] | 139 | .. seealso:: |
| 140 | |
| 141 | :pep:`391` - Dictionary Based Configuration for Logging |
| 142 | PEP written by Vinay Sajip. |
| 143 | |
Georg Brandl | 97b20da | 2010-11-16 15:15:29 +0000 | [diff] [blame] | 144 | PEP 3148: The ``concurrent.futures`` module |
| 145 | ============================================ |
| 146 | |
Raymond Hettinger | 6f04adc | 2010-12-04 22:56:25 +0000 | [diff] [blame] | 147 | Code for creating and managing concurrency is being collected in a new toplevel |
| 148 | namespace, *concurrent*. Its first member is a *futures* package which provides |
| 149 | a uniform high level interface for managing threads and processes. |
| 150 | |
| 151 | The design for :mod:`concurrent.futures` was inspired by |
| 152 | *java.util.concurrent.package*. In that model, a running call and its result |
| 153 | are represented by a :class:`~concurrent.futures.Future` object which abstracts |
| 154 | features common to threads, processes, and remote procedure calls. That object |
| 155 | supports status checks (running or done), timeouts, cancellations, adding |
| 156 | callbacks, and access to results or exceptions.XS |
| 157 | |
| 158 | The primary offering of the new module is a pair of executor classes for |
| 159 | launching and managing calls. The goal of the executors is to make it easier to |
| 160 | use existing tools for making parallel calls. They save the effort needed to |
| 161 | setup a pool of resources, launch the calls, create a results queue, add |
| 162 | time-out handling, and limit the total number of threads, processes, or remote |
Raymond Hettinger | c269ae8 | 2010-12-05 01:01:52 +0000 | [diff] [blame] | 163 | procedure calls. |
Raymond Hettinger | 6f04adc | 2010-12-04 22:56:25 +0000 | [diff] [blame] | 164 | |
| 165 | Ideally, each application should share a single executor across multiple |
| 166 | components so that process and thread limits can be centrally managed. This |
| 167 | solves the design challenge that arises when each component has its own |
| 168 | competing strategy for resource management. |
| 169 | |
| 170 | For an example of :class:`~concurrent.futures.ThreadPoolExecutor`, |
| 171 | see :ref:`code for threaded parallel URL reads<threadpoolexecutor-example>`. |
| 172 | |
| 173 | For an example of :class:`~concurrent.futures.ProcessPoolExecutor`, |
| 174 | see :ref:`code for computing prime numbers in parallel<processpoolexecutor-example>`. |
| 175 | |
| 176 | .. seealso:: |
| 177 | |
Raymond Hettinger | a5a3554 | 2010-12-05 00:39:18 +0000 | [diff] [blame] | 178 | :pep:`3148` - Futures -- Execute Computations Asynchronously |
Raymond Hettinger | 6f04adc | 2010-12-04 22:56:25 +0000 | [diff] [blame] | 179 | PEP written by Brain Quinlan. |
Georg Brandl | 97b20da | 2010-11-16 15:15:29 +0000 | [diff] [blame] | 180 | |
Georg Brandl | da0a211 | 2010-09-05 11:28:33 +0000 | [diff] [blame] | 181 | |
Raymond Hettinger | f95b199 | 2010-09-04 23:53:24 +0000 | [diff] [blame] | 182 | PEP 3147: PYC Repository Directories |
| 183 | ===================================== |
| 184 | |
David Malcolm | 778645a | 2010-12-07 00:32:04 +0000 | [diff] [blame] | 185 | Python's scheme for caching bytecode in *.pyc* files did not work well in |
Raymond Hettinger | f95b199 | 2010-09-04 23:53:24 +0000 | [diff] [blame] | 186 | environments with multiple python interpreters. If one interpreter encountered |
| 187 | a cached file created by another interpreter, it would recompile the source and |
| 188 | overwrite the cached file, thus losing the benefits of caching. |
| 189 | |
| 190 | The issue of "pyc fights" has become more pronounced as it has become |
Éric Araujo | 4234ad4 | 2010-09-05 17:32:25 +0000 | [diff] [blame] | 191 | commonplace for Linux distributions to ship with multiple versions of Python. |
Raymond Hettinger | f95b199 | 2010-09-04 23:53:24 +0000 | [diff] [blame] | 192 | These conflicts also arise with CPython alternatives such as Unladen Swallow. |
| 193 | |
| 194 | To solve this problem, Python's import machinery has been extended to use |
Éric Araujo | 4234ad4 | 2010-09-05 17:32:25 +0000 | [diff] [blame] | 195 | distinct filenames for each interpreter. Instead of Python 3.2 and Python 3.3 and |
| 196 | Unladen Swallow each competing for a file called "mymodule.pyc", they will now |
Raymond Hettinger | f95b199 | 2010-09-04 23:53:24 +0000 | [diff] [blame] | 197 | look for "mymodule.cpython-32.pyc", "mymodule.cpython-33.pyc", and |
Éric Araujo | 4234ad4 | 2010-09-05 17:32:25 +0000 | [diff] [blame] | 198 | "mymodule.unladen10.pyc". And to prevent all of these new files from |
Raymond Hettinger | f95b199 | 2010-09-04 23:53:24 +0000 | [diff] [blame] | 199 | cluttering source directories, the *pyc* files are now collected in a |
| 200 | "__pycache__" directory stored under the package directory. |
| 201 | |
| 202 | Aside from the filenames and target directories, the new scheme has a few |
| 203 | aspects that are visible to the programmer: |
| 204 | |
Georg Brandl | da0a211 | 2010-09-05 11:28:33 +0000 | [diff] [blame] | 205 | * Imported modules now have a :attr:`__cached__` attribute which stores the name |
| 206 | of the actual file that was imported: |
Raymond Hettinger | f95b199 | 2010-09-04 23:53:24 +0000 | [diff] [blame] | 207 | |
Raymond Hettinger | 92ba286 | 2010-09-06 01:16:46 +0000 | [diff] [blame] | 208 | >>> import collections |
| 209 | >>> collections.__cached__ |
| 210 | 'c:/py32/lib/__pycache__/collections.cpython-32.pyc' |
Raymond Hettinger | f95b199 | 2010-09-04 23:53:24 +0000 | [diff] [blame] | 211 | |
| 212 | * The tag that is unique to each interpreter is accessible from the :mod:`imp` |
Georg Brandl | da0a211 | 2010-09-05 11:28:33 +0000 | [diff] [blame] | 213 | module: |
Raymond Hettinger | f95b199 | 2010-09-04 23:53:24 +0000 | [diff] [blame] | 214 | |
Raymond Hettinger | 92ba286 | 2010-09-06 01:16:46 +0000 | [diff] [blame] | 215 | >>> import imp |
| 216 | >>> imp.get_tag() |
| 217 | 'cpython-32' |
Raymond Hettinger | f95b199 | 2010-09-04 23:53:24 +0000 | [diff] [blame] | 218 | |
| 219 | * Scripts that try to deduce source filename from the imported file now need to |
| 220 | be smarter. It is no longer sufficient to simply strip the "c" from a ".pyc" |
| 221 | filename. Instead, use the new functions in the :mod:`imp` module: |
| 222 | |
Georg Brandl | da0a211 | 2010-09-05 11:28:33 +0000 | [diff] [blame] | 223 | >>> imp.source_from_cache('c:/py32/lib/__pycache__/collections.cpython-32.pyc') |
| 224 | 'c:/py32/lib/collections.py' |
| 225 | >>> imp.cache_from_source('c:/py32/lib/collections.py') |
| 226 | 'c:/py32/lib/__pycache__/collections.cpython-32.pyc' |
Raymond Hettinger | f95b199 | 2010-09-04 23:53:24 +0000 | [diff] [blame] | 227 | |
| 228 | * The :mod:`py_compile` and :mod:`compileall` modules have been updated to |
| 229 | reflect the new naming convention and target directory. |
| 230 | |
| 231 | .. seealso:: |
| 232 | |
| 233 | :pep:`3147` - PYC Repository Directories |
| 234 | PEP written by Barry Warsaw. |
| 235 | |
Georg Brandl | da0a211 | 2010-09-05 11:28:33 +0000 | [diff] [blame] | 236 | |
Georg Brandl | 3ad4675 | 2010-12-05 07:59:29 +0000 | [diff] [blame] | 237 | PEP 3149: ABI Version Tagged .so Files |
| 238 | ====================================== |
Georg Brandl | f11c6c4 | 2010-09-03 22:20:58 +0000 | [diff] [blame] | 239 | |
Raymond Hettinger | ebea6fa | 2010-09-05 00:27:25 +0000 | [diff] [blame] | 240 | The PYC repository directory allows multiple bytecode cache files to be |
| 241 | co-located. This PEP implements a similar mechanism for shared object files by |
| 242 | giving them a common directory and distinct names for each version. |
Georg Brandl | f11c6c4 | 2010-09-03 22:20:58 +0000 | [diff] [blame] | 243 | |
Raymond Hettinger | ebea6fa | 2010-09-05 00:27:25 +0000 | [diff] [blame] | 244 | The common directory is "pyshared" and the file names are made distinct by |
| 245 | identifying the Python implementation (such as CPython, PyPy, Jython, etc.), the |
| 246 | major and minor version numbers, and optional build flags (such as "d" for |
Éric Araujo | 4234ad4 | 2010-09-05 17:32:25 +0000 | [diff] [blame] | 247 | debug, "m" for pymalloc, "u" for wide-unicode). For an arbitrary package "foo", |
Raymond Hettinger | ebea6fa | 2010-09-05 00:27:25 +0000 | [diff] [blame] | 248 | you may see these files when the distribution package is installed:: |
| 249 | |
| 250 | /usr/share/pyshared/foo.cpython-32m.so |
| 251 | /usr/share/pyshared/foo.cpython-33md.so |
| 252 | |
| 253 | In Python itself, the tags are accessible from functions in the :mod:`sysconfig` |
| 254 | module:: |
| 255 | |
| 256 | >>> import sysconfig |
| 257 | >>> sysconfig.get_config_var('SOABI') # find the version tag |
| 258 | 'cpython-32mu' |
| 259 | >>> sysconfig.get_config_var('SO') # find the full filename extension |
| 260 | 'cpython-32mu.so' |
| 261 | |
| 262 | .. seealso:: |
| 263 | |
| 264 | :pep:`3149` - ABI Version Tagged .so Files |
| 265 | PEP written by Barry Warsaw. |
Raymond Hettinger | 6e6565b | 2009-06-28 20:56:11 +0000 | [diff] [blame] | 266 | |
| 267 | |
R. David Murray | 7c0a227 | 2010-10-08 21:37:39 +0000 | [diff] [blame] | 268 | Email 5.1 |
| 269 | ========= |
| 270 | |
| 271 | The email package is extended to be able to parse and generate email messages |
| 272 | in bytes format. |
| 273 | |
| 274 | * New functions :func:`~email.message_from_bytes` and |
| 275 | :func:`~email.message_from_binary_file`, and new classes |
| 276 | :class:`~email.parser.BytesFeedParser` and :class:`~email.parser.BytesParser` |
| 277 | allow binary message data to be parsed into model objects. |
| 278 | |
| 279 | * Given bytes input to the model, :meth:`~email.message.Message.get_payload` |
| 280 | will by default decode a message body that has a |
Senthil Kumaran | 8227045 | 2010-10-15 13:29:33 +0000 | [diff] [blame] | 281 | :mailheader:`Content-Transfer-Encoding` of ``8bit`` using the charset |
| 282 | specified in the MIME headers and return the resulting string. |
R. David Murray | 7c0a227 | 2010-10-08 21:37:39 +0000 | [diff] [blame] | 283 | |
| 284 | * Given bytes input to the model, :class:`~email.generator.Generator` will |
| 285 | convert message bodies that have a :mailheader:`Content-Transfer-Encoding` of |
| 286 | 8bit to instead have a 7bit Content-Transfer-Encoding. |
| 287 | |
| 288 | * New class :class:`~email.generator.BytesGenerator` produces bytes |
| 289 | as output, preserving any unchanged non-ASCII data that was |
| 290 | present in the input used to build the model, including message bodies |
| 291 | with a :mailheader:`Content-Transfer-Encoding` of 8bit. |
| 292 | |
| 293 | (Proposed and implemented by R. David Murray, :issue:`4661`.) |
| 294 | |
| 295 | |
Raymond Hettinger | 6e6565b | 2009-06-28 20:56:11 +0000 | [diff] [blame] | 296 | Other Language Changes |
| 297 | ====================== |
| 298 | |
| 299 | Some smaller changes made to the core Python language are: |
| 300 | |
Raymond Hettinger | 43b5a85 | 2010-12-05 04:04:21 +0000 | [diff] [blame] | 301 | * :class:`bytes` and :class:`str` now have two net methods, *tranform* and *untransform*. |
| 302 | These provided analogues to *encode* and *decode* but are used for general purpose |
| 303 | string-to-string and bytes-to-bytes transformations rather than Unicode codecs. |
| 304 | |
| 305 | Along with the new methods, several non-unicode codecs been restored from Python 2.x |
| 306 | including *base64*, *bz2*, *hex*, *quopri*, *rot13*, *uu*, and *zlib*. |
| 307 | |
| 308 | >>> t = b'which witch had which witches wrist watch' |
| 309 | >>> t.transform('quopri') |
| 310 | b'which=20witch=20had=20which=20witches=20wrist=20watch' |
| 311 | |
| 312 | >>> short = t.transform('zlib_codec') |
| 313 | >>> len(t), len(short) |
| 314 | (41, 38) |
| 315 | >>> short.untransform('zlib_codec') |
| 316 | b'which witch had which witches wrist watch' |
| 317 | |
Raymond Hettinger | e5e728b | 2010-12-05 06:35:16 +0000 | [diff] [blame] | 318 | (From multiple contributors in :issue:`7475`.) |
| 319 | |
Raymond Hettinger | e5e1a98 | 2010-12-05 08:35:21 +0000 | [diff] [blame] | 320 | * String formatting for :func:`format` and :meth:`str.format` gained new |
| 321 | capabilities for the format character **#**. Previously, for integers in |
| 322 | binary, octal, or hexadecimal, it caused the output to be prefixed with '0b', |
| 323 | '0o', or '0x' respectively. Now it can also handle floats, complex, and |
| 324 | Decimal, causing the output to always have a decimal point even when no digits |
| 325 | follow it. |
Raymond Hettinger | e5e728b | 2010-12-05 06:35:16 +0000 | [diff] [blame] | 326 | |
| 327 | >>> format(20, '#o') |
| 328 | '0o24' |
| 329 | >>> format(12.34, '#5.0f') |
| 330 | ' 12.' |
| 331 | |
| 332 | (Suggested by Mark Dickinson and implemented by Eric Smith in :issue:`7094`.) |
Raymond Hettinger | 43b5a85 | 2010-12-05 04:04:21 +0000 | [diff] [blame] | 333 | |
Raymond Hettinger | c269ae8 | 2010-12-05 01:01:52 +0000 | [diff] [blame] | 334 | * The interpreter can now be started with a quiet option, ``-q``, to suppress |
| 335 | the copyright and version information in an interactive mode. |
| 336 | |
| 337 | (Contributed by Marcin Wojdyr in issue:`1772833`). |
| 338 | |
Georg Brandl | da0a211 | 2010-09-05 11:28:33 +0000 | [diff] [blame] | 339 | * The :func:`hasattr` function used to catch and suppress any Exception. Now, |
| 340 | it only catches :exc:`AttributeError`. Under the hood, :func:`hasattr` works |
| 341 | by calling :func:`getattr` and throwing away the results. This is necessary |
| 342 | because dynamic attribute creation is possible using :meth:`__getattribute__` |
Éric Araujo | 4234ad4 | 2010-09-05 17:32:25 +0000 | [diff] [blame] | 343 | or :meth:`__getattr__`. If :func:`hasattr` were to just scan instance and class |
Éric Araujo | cc6aac6 | 2010-09-07 21:35:35 +0000 | [diff] [blame] | 344 | dictionaries it would miss the dynamic methods and make it difficult to |
Georg Brandl | da0a211 | 2010-09-05 11:28:33 +0000 | [diff] [blame] | 345 | implement proxy objects. |
Raymond Hettinger | 1784ff0 | 2010-09-05 01:00:19 +0000 | [diff] [blame] | 346 | |
| 347 | (Discovered by Yury Selivanov and fixed by Benjamin Peterson; :issue:`9666`.) |
| 348 | |
Éric Araujo | 4234ad4 | 2010-09-05 17:32:25 +0000 | [diff] [blame] | 349 | * The :func:`str` of a float or complex number is now the same as its |
Raymond Hettinger | 1784ff0 | 2010-09-05 01:00:19 +0000 | [diff] [blame] | 350 | :func:`repr`. Previously, the :func:`str` form was shorter but that just |
Éric Araujo | 4234ad4 | 2010-09-05 17:32:25 +0000 | [diff] [blame] | 351 | caused confusion and is no longer needed now that the shortest possible |
Georg Brandl | da0a211 | 2010-09-05 11:28:33 +0000 | [diff] [blame] | 352 | :func:`repr` is displayed by default: |
Raymond Hettinger | bb734c6 | 2010-09-05 05:56:44 +0000 | [diff] [blame] | 353 | |
Raymond Hettinger | 92ba286 | 2010-09-06 01:16:46 +0000 | [diff] [blame] | 354 | >>> repr(math.pi) |
| 355 | '3.141592653589793' |
| 356 | >>> str(math.pi) |
| 357 | '3.141592653589793' |
Raymond Hettinger | 1784ff0 | 2010-09-05 01:00:19 +0000 | [diff] [blame] | 358 | |
Georg Brandl | da0a211 | 2010-09-05 11:28:33 +0000 | [diff] [blame] | 359 | (Proposed and implemented by Mark Dickinson; :issue:`9337`.) |
Raymond Hettinger | 6e6565b | 2009-06-28 20:56:11 +0000 | [diff] [blame] | 360 | |
Antoine Pitrou | d305200 | 2010-09-15 15:09:40 +0000 | [diff] [blame] | 361 | * :class:`memoryview` objects now have a :meth:`release()` method and support |
| 362 | the context manager protocol. This allows timely release of any resources |
| 363 | that were acquired when requesting a buffer from the original object. |
| 364 | |
Raymond Hettinger | d8fae4e | 2010-12-05 05:39:54 +0000 | [diff] [blame] | 365 | >>> with memoryview(b'abcdefgh') as v: |
| 366 | ... print(v.tolist()) |
| 367 | ... |
| 368 | [97, 98, 99, 100, 101, 102, 103, 104] |
| 369 | |
Antoine Pitrou | d305200 | 2010-09-15 15:09:40 +0000 | [diff] [blame] | 370 | (Added by Antoine Pitrou; :issue:`9757`.) |
| 371 | |
Raymond Hettinger | 92ba286 | 2010-09-06 01:16:46 +0000 | [diff] [blame] | 372 | * Mark Dickinson crafted an elegant and efficient scheme for assuring that |
| 373 | different numeric datatypes will have the same hash value whenever their |
| 374 | actual values are equal:: |
| 375 | |
| 376 | >>> assert hash(Fraction(3, 2)) == hash(1.5) == \ |
| 377 | hash(Decimal("1.5")) == hash(complex(1.5, 0)) |
| 378 | |
| 379 | (See :issue:`8188`.) |
Raymond Hettinger | 6e6565b | 2009-06-28 20:56:11 +0000 | [diff] [blame] | 380 | |
Amaury Forgeot d'Arc | ba117ef | 2010-09-10 21:39:53 +0000 | [diff] [blame] | 381 | * Previously it was illegal to delete a name from the local namespace if it |
| 382 | occurs as a free variable in a nested block:: |
| 383 | |
| 384 | >>> def outer(x): |
| 385 | ... def inner(): |
| 386 | ... return x |
| 387 | ... inner() |
| 388 | ... del x |
| 389 | |
| 390 | This is now allowed. Remember that the target of an :keyword:`except` clause |
| 391 | is cleared, so this code which used to work with Python 2.6, raised a |
| 392 | :exc:`SyntaxError` with Python 3.1 and now works again:: |
| 393 | |
| 394 | >>> def f(): |
| 395 | ... def print_error(): |
| 396 | ... print(e) |
| 397 | ... try: |
| 398 | ... something |
| 399 | ... except Exception as e: |
| 400 | ... print_error() |
| 401 | ... # implicit "del e" here |
| 402 | |
| 403 | (See :issue:`4617`.) |
| 404 | |
Antoine Pitrou | 7d15a72 | 2010-11-05 22:13:55 +0000 | [diff] [blame] | 405 | * A new warning category, :exc:`ResourceWarning`, has been added. It is |
Raymond Hettinger | c269ae8 | 2010-12-05 01:01:52 +0000 | [diff] [blame] | 406 | emitted when potential issues with resource consumption or cleanup |
Antoine Pitrou | 7d15a72 | 2010-11-05 22:13:55 +0000 | [diff] [blame] | 407 | are detected. It is silenced by default in normal release builds, but |
Raymond Hettinger | c269ae8 | 2010-12-05 01:01:52 +0000 | [diff] [blame] | 408 | can be enabled through the means provided by the :mod:`warnings` |
Antoine Pitrou | 7d15a72 | 2010-11-05 22:13:55 +0000 | [diff] [blame] | 409 | module, or on the command line. |
| 410 | |
Raymond Hettinger | d8fae4e | 2010-12-05 05:39:54 +0000 | [diff] [blame] | 411 | A :exc:`ResourceWarning` is issued at interpreter shutdown if the |
Antoine Pitrou | 7d15a72 | 2010-11-05 22:13:55 +0000 | [diff] [blame] | 412 | :data:`gc.garbage` list isn't empty. This is meant to make the programmer |
| 413 | aware that their code contains object finalization issues. |
| 414 | |
Raymond Hettinger | d8fae4e | 2010-12-05 05:39:54 +0000 | [diff] [blame] | 415 | A :exc:`ResourceWarning` is also issued when a :term:`file object` is destroyed |
Antoine Pitrou | 7d15a72 | 2010-11-05 22:13:55 +0000 | [diff] [blame] | 416 | without having been explicitly closed. While the deallocator for such |
| 417 | object ensures it closes the underlying operating system resource |
| 418 | (usually, a file descriptor), the delay in deallocating the object could |
| 419 | produce various issues, especially under Windows. Here is an example |
| 420 | of enabling the warning from the command line:: |
| 421 | |
| 422 | $ ./python -Wdefault |
| 423 | Python 3.2a3+ (py3k, Nov 5 2010, 22:58:04) |
| 424 | [GCC 4.4.3] on linux2 |
| 425 | Type "help", "copyright", "credits" or "license" for more information. |
| 426 | >>> f = open("foo", "wb") |
| 427 | >>> del f |
| 428 | __main__:1: ResourceWarning: unclosed file <_io.BufferedWriter name='foo'> |
| 429 | >>> |
| 430 | |
Raymond Hettinger | d8fae4e | 2010-12-05 05:39:54 +0000 | [diff] [blame] | 431 | (Added by Antoine Pitrou and Georg Brandl in :issue:`10093` and :issue:`477863`.) |
Antoine Pitrou | 7d15a72 | 2010-11-05 22:13:55 +0000 | [diff] [blame] | 432 | |
Raymond Hettinger | e5e1a98 | 2010-12-05 08:35:21 +0000 | [diff] [blame] | 433 | * :class:`range` objects now support *index* and *count* methods. This is |
Raymond Hettinger | dadf93c | 2010-12-05 02:56:21 +0000 | [diff] [blame] | 434 | part of an effort to make more objects fully implement the :class:`collections.Sequence` |
| 435 | :term:`abstract base class`. As a result, the language will have a more |
| 436 | uniform API. |
| 437 | |
| 438 | In addition, :class:`range` objects now support slicing and negative indices. |
| 439 | This makes *range* more interoperable with lists. |
| 440 | |
| 441 | (Contributed by Daniel Stuzback in :issue:`9213` and by Alexander Belopolsky |
| 442 | in :issue:`2690`.) |
Nick Coghlan | 37ee850 | 2010-12-03 14:26:13 +0000 | [diff] [blame] | 443 | |
Raymond Hettinger | d8fae4e | 2010-12-05 05:39:54 +0000 | [diff] [blame] | 444 | * The :func:`callable` builtin function from Py2.x was resurrected. It provides |
Raymond Hettinger | b87ba26 | 2010-12-06 04:31:40 +0000 | [diff] [blame] | 445 | a concise, readable alternative to using an :term:`abstract base class` in an |
| 446 | expression like ``isinstance(x, collections.Callable)``. |
Raymond Hettinger | d8fae4e | 2010-12-05 05:39:54 +0000 | [diff] [blame] | 447 | |
| 448 | (See :issue:`10518`.) |
Amaury Forgeot d'Arc | ba117ef | 2010-09-10 21:39:53 +0000 | [diff] [blame] | 449 | |
Raymond Hettinger | 6e6565b | 2009-06-28 20:56:11 +0000 | [diff] [blame] | 450 | New, Improved, and Deprecated Modules |
| 451 | ===================================== |
| 452 | |
Éric Araujo | 4234ad4 | 2010-09-05 17:32:25 +0000 | [diff] [blame] | 453 | * The :mod:`functools` module includes a new decorator for caching function |
Georg Brandl | da0a211 | 2010-09-05 11:28:33 +0000 | [diff] [blame] | 454 | calls. :func:`functools.lru_cache` can save repeated queries to an external |
| 455 | resource whenever the results are expected to be the same. |
Raymond Hettinger | aed05eb | 2010-08-02 01:43:41 +0000 | [diff] [blame] | 456 | |
Raymond Hettinger | 86f9613 | 2010-08-06 23:23:49 +0000 | [diff] [blame] | 457 | For example, adding a caching decorator to a database query function can save |
| 458 | database accesses for popular searches:: |
Raymond Hettinger | aed05eb | 2010-08-02 01:43:41 +0000 | [diff] [blame] | 459 | |
Georg Brandl | da0a211 | 2010-09-05 11:28:33 +0000 | [diff] [blame] | 460 | @functools.lru_cache(maxsize=300) |
| 461 | def get_phone_number(name): |
| 462 | c = conn.cursor() |
| 463 | c.execute('SELECT phonenumber FROM phonelist WHERE name=?', (name,)) |
| 464 | return c.fetchone()[0] |
Raymond Hettinger | aed05eb | 2010-08-02 01:43:41 +0000 | [diff] [blame] | 465 | |
Georg Brandl | da0a211 | 2010-09-05 11:28:33 +0000 | [diff] [blame] | 466 | >>> for name in user_requests: |
Raymond Hettinger | 7496b41 | 2010-11-30 19:15:45 +0000 | [diff] [blame] | 467 | ... get_phone_number(name) # cached lookup |
| 468 | |
| 469 | To help with choosing an effective cache size, the wrapped function is |
| 470 | instrumented for tracking cache statistics: |
| 471 | |
Raymond Hettinger | 5e20bab | 2010-11-30 07:13:04 +0000 | [diff] [blame] | 472 | >>> get_phone_number.cache_info() |
Raymond Hettinger | 7496b41 | 2010-11-30 19:15:45 +0000 | [diff] [blame] | 473 | CacheInfo(hits=4805, misses=980, maxsize=300, currsize=300) |
Raymond Hettinger | aed05eb | 2010-08-02 01:43:41 +0000 | [diff] [blame] | 474 | |
Raymond Hettinger | f309828 | 2010-08-15 03:30:45 +0000 | [diff] [blame] | 475 | If the phonelist table gets updated, the outdated contents of the cache can be |
Georg Brandl | da0a211 | 2010-09-05 11:28:33 +0000 | [diff] [blame] | 476 | cleared with: |
Raymond Hettinger | f309828 | 2010-08-15 03:30:45 +0000 | [diff] [blame] | 477 | |
Georg Brandl | da0a211 | 2010-09-05 11:28:33 +0000 | [diff] [blame] | 478 | >>> get_phone_number.cache_clear() |
Raymond Hettinger | f309828 | 2010-08-15 03:30:45 +0000 | [diff] [blame] | 479 | |
Raymond Hettinger | 6e35394 | 2010-12-04 23:42:12 +0000 | [diff] [blame] | 480 | (Contributed by Raymond Hettinger and incorporating design ideas from |
Raymond Hettinger | b87ba26 | 2010-12-06 04:31:40 +0000 | [diff] [blame] | 481 | Jim Baker, Miki Tebeka, and Nick Coghlan.) |
Raymond Hettinger | aed05eb | 2010-08-02 01:43:41 +0000 | [diff] [blame] | 482 | |
Antoine Pitrou | 7d49bc9 | 2010-09-15 15:13:17 +0000 | [diff] [blame] | 483 | * The :func:`functools.wraps` decorator now adds a :attr:`__wrapped__` attribute |
| 484 | pointing to the original callable function. This allows wrapped functions to |
| 485 | be introspected. It also copies :attr:`__annotations__` if defined. And now |
| 486 | it also gracefully skips over missing attributes such as :attr:`__doc__` which |
| 487 | might not be defined for the wrapped callable. |
| 488 | |
| 489 | (By Nick Coghlan and Terrence Cole; :issue:`9567`, :issue:`3445`, and |
| 490 | :issue:`8814`.) |
| 491 | |
Raymond Hettinger | 6e35394 | 2010-12-04 23:42:12 +0000 | [diff] [blame] | 492 | * The :mod:`itertools` module has a new function, :func:`~itertools.accumulate` |
Raymond Hettinger | a5a3554 | 2010-12-05 00:39:18 +0000 | [diff] [blame] | 493 | modeled on APL's *scan* operator and on Numpy's *accumulate* function: |
Raymond Hettinger | 6e35394 | 2010-12-04 23:42:12 +0000 | [diff] [blame] | 494 | |
| 495 | >>> list(accumulate(8, 2, 50)) |
| 496 | [8, 10, 60] |
| 497 | |
| 498 | >>> prob_dist = [0.1, 0.4, 0.2, 0.3] |
| 499 | >>> list(accumulate(prob_dist)) # cumulative probability distribution |
| 500 | [0.1, 0.5, 0.7, 1.0] |
| 501 | |
| 502 | For an example using :func:`~itertools.accumulate`, see the :ref:`examples for |
| 503 | the random module <random-examples>`. |
| 504 | |
| 505 | (Contributed by Raymond Hettinger and incorporating design suggestions |
| 506 | from Mark Dickinson.) |
| 507 | |
Antoine Pitrou | cd889af | 2010-10-06 21:13:56 +0000 | [diff] [blame] | 508 | * The :mod:`nntplib` module gets a revamped implementation with better |
| 509 | bytes / unicode semantics as well as more practical APIs. These improvements |
| 510 | break compatibility with the nntplib version in Python 3.1, which was |
| 511 | partly dysfunctional in itself. |
| 512 | |
| 513 | (Contributed by Antoine Pitrou in :issue:`9360`) |
| 514 | |
Antoine Pitrou | 7d49bc9 | 2010-09-15 15:13:17 +0000 | [diff] [blame] | 515 | * The :mod:`abc` module now supports :func:`~abc.abstractclassmethod` and |
| 516 | :func:`~abc.abstractstaticmethod`. |
| 517 | |
Raymond Hettinger | a5a3554 | 2010-12-05 00:39:18 +0000 | [diff] [blame] | 518 | These tools make it possible to define an :term:`Abstract Base Class` that |
| 519 | requires a particular :func:`classmethod` or :func:`staticmethod` to be |
| 520 | implemented. |
| 521 | |
Antoine Pitrou | 7d49bc9 | 2010-09-15 15:13:17 +0000 | [diff] [blame] | 522 | (Patch submitted by Daniel Urban; :issue:`5867`.) |
| 523 | |
Georg Brandl | da0a211 | 2010-09-05 11:28:33 +0000 | [diff] [blame] | 524 | * The previously deprecated :func:`contextlib.nested` function has been removed |
| 525 | in favor of a plain :keyword:`with` statement which can accept multiple |
| 526 | context managers. The latter technique is faster (because it is built-in), |
| 527 | and it does a better job finalizing multiple context managers when one of them |
| 528 | raises an exception. |
Antoine Pitrou | d67075e | 2010-07-31 22:48:02 +0000 | [diff] [blame] | 529 | |
| 530 | (Contributed by Georg Brandl and Mattias Brändström; |
| 531 | `appspot issue 53094 <http://codereview.appspot.com/53094>`_.) |
| 532 | |
Georg Brandl | da0a211 | 2010-09-05 11:28:33 +0000 | [diff] [blame] | 533 | * The :class:`ftplib.FTP` class now supports the context manager protocol to |
Éric Araujo | 4234ad4 | 2010-09-05 17:32:25 +0000 | [diff] [blame] | 534 | unconditionally consume :exc:`socket.error` exceptions and to close the FTP |
Raymond Hettinger | 92ba286 | 2010-09-06 01:16:46 +0000 | [diff] [blame] | 535 | connection when done:: |
Raymond Hettinger | bcbd696 | 2010-09-05 08:46:36 +0000 | [diff] [blame] | 536 | |
Raymond Hettinger | 92ba286 | 2010-09-06 01:16:46 +0000 | [diff] [blame] | 537 | >>> from ftplib import FTP |
| 538 | >>> with FTP("ftp1.at.proftpd.org") as ftp: |
| 539 | ... ftp.login() |
| 540 | ... ftp.dir() |
| 541 | ... |
| 542 | '230 Anonymous login ok, restrictions apply.' |
| 543 | dr-xr-xr-x 9 ftp ftp 154 May 6 10:43 . |
| 544 | dr-xr-xr-x 9 ftp ftp 154 May 6 10:43 .. |
| 545 | dr-xr-xr-x 5 ftp ftp 4096 May 6 10:43 CentOS |
| 546 | dr-xr-xr-x 3 ftp ftp 18 Jul 10 2008 Fedora |
Raymond Hettinger | bcbd696 | 2010-09-05 08:46:36 +0000 | [diff] [blame] | 547 | |
Raymond Hettinger | 92ba286 | 2010-09-06 01:16:46 +0000 | [diff] [blame] | 548 | Other file-like objects such as :class:`mmap.mmap` and :func:`fileinput.input` |
| 549 | also grew auto-closing context managers:: |
Giampaolo Rodolà | bd576b7 | 2010-05-10 14:53:29 +0000 | [diff] [blame] | 550 | |
Raymond Hettinger | 92ba286 | 2010-09-06 01:16:46 +0000 | [diff] [blame] | 551 | with fileinput.input(files=('log1.txt', 'log2.txt')) as f: |
| 552 | for line in f: |
| 553 | process(line) |
Georg Brandl | da0a211 | 2010-09-05 11:28:33 +0000 | [diff] [blame] | 554 | |
Raymond Hettinger | 92ba286 | 2010-09-06 01:16:46 +0000 | [diff] [blame] | 555 | (Contributed by Tarek Ziadé and Giampaolo Rodolà in :issue:`4972`, and |
| 556 | by Georg Brandl in :issue:`8046` and :issue:`1286`.) |
Antoine Pitrou | 696e035 | 2010-08-08 22:18:46 +0000 | [diff] [blame] | 557 | |
Georg Brandl | 3ad4675 | 2010-12-05 07:59:29 +0000 | [diff] [blame] | 558 | .. mention os.popen and subprocess.Popen auto-closing of fds |
| 559 | |
Antoine Pitrou | cd889af | 2010-10-06 21:13:56 +0000 | [diff] [blame] | 560 | * :class:`gzip.GzipFile` now implements the :class:`io.BufferedIOBase` ABC |
| 561 | (except for ``truncate()``), has a :meth:`~gzip.GzipFile.peek` method, |
| 562 | and supports unseekable as well as zero-padded file objects. |
| 563 | |
| 564 | (Contributed by Antoine Pitrou, Nir Aides and Brian Curtin in :issue:`9962`, |
| 565 | :issue:`1675951`, :issue:`7471` and :issue:`2846`.) |
| 566 | |
| 567 | The :mod:`gzip` module also gains the :func:`~gzip.compress` and |
| 568 | :func:`~gzip.decompress` functions for easier in-memory compression and |
| 569 | decompression. |
| 570 | |
| 571 | (Contributed by Anand B. Pillai in :issue:`3488`.) |
| 572 | |
Andrew M. Kuchling | 4ea04a3 | 2010-08-18 22:30:34 +0000 | [diff] [blame] | 573 | * The :mod:`os` module now has the :const:`ST_RDONLY` and :const:`ST_NOSUID` |
| 574 | constants, for use with the :func:`~os.statvfs` function. |
Raymond Hettinger | bcbd696 | 2010-09-05 08:46:36 +0000 | [diff] [blame] | 575 | |
Andrew M. Kuchling | 4ea04a3 | 2010-08-18 22:30:34 +0000 | [diff] [blame] | 576 | (Patch by Adam Jackson; :issue:`7647`.) |
| 577 | |
Amaury Forgeot d'Arc | 4b6fdf3 | 2010-09-07 21:31:17 +0000 | [diff] [blame] | 578 | * :func:`os.getppid` is now supported on Windows. Note that it will continue to |
| 579 | return the same pid even after the parent process has exited. |
| 580 | |
| 581 | (Patch by Jon Anglin; :issue:`6394`.) |
| 582 | |
Antoine Pitrou | d67075e | 2010-07-31 22:48:02 +0000 | [diff] [blame] | 583 | * The :func:`shutil.copytree` function has two new options: |
| 584 | |
Raymond Hettinger | db9044e | 2010-09-06 01:29:23 +0000 | [diff] [blame] | 585 | * *ignore_dangling_symlinks*: when ``symlinks=False`` so that the function |
| 586 | copies the file pointed to by the symlink, not the symlink itself. This |
Georg Brandl | da0a211 | 2010-09-05 11:28:33 +0000 | [diff] [blame] | 587 | option will silence the error raised if the file doesn't exist. |
Antoine Pitrou | d67075e | 2010-07-31 22:48:02 +0000 | [diff] [blame] | 588 | |
Raymond Hettinger | db9044e | 2010-09-06 01:29:23 +0000 | [diff] [blame] | 589 | * *copy_function*: is a callable that will be used to copy files. |
Antoine Pitrou | d67075e | 2010-07-31 22:48:02 +0000 | [diff] [blame] | 590 | :func:`shutil.copy2` is used by default. |
| 591 | |
| 592 | (Contributed by Tarek Ziadé.) |
| 593 | |
Georg Brandl | da0a211 | 2010-09-05 11:28:33 +0000 | [diff] [blame] | 594 | * Socket objects now have a :meth:`~socket.socket.detach()` method which puts |
| 595 | the socket into closed state without actually closing the underlying file |
| 596 | descriptor. The latter can then be reused for other purposes. |
Antoine Pitrou | e43f9d0 | 2010-08-08 23:24:50 +0000 | [diff] [blame] | 597 | |
| 598 | (Added by Antoine Pitrou; :issue:`8524`.) |
| 599 | |
Raymond Hettinger | d881f31 | 2010-09-05 08:54:32 +0000 | [diff] [blame] | 600 | * The :mod:`sqlite3` module has two new capabilities. |
Antoine Pitrou | d67075e | 2010-07-31 22:48:02 +0000 | [diff] [blame] | 601 | |
Georg Brandl | da0a211 | 2010-09-05 11:28:33 +0000 | [diff] [blame] | 602 | The :attr:`Connection.in_transit` attribute is true if there is an active |
| 603 | transaction for uncommitted changes. |
Antoine Pitrou | d67075e | 2010-07-31 22:48:02 +0000 | [diff] [blame] | 604 | |
Raymond Hettinger | d881f31 | 2010-09-05 08:54:32 +0000 | [diff] [blame] | 605 | The :meth:`Connection.enable_load_extension` and |
| 606 | :meth:`Connection.load_extension` methods allows you to load SQLite extensions |
| 607 | from ".so" files. One well-known extension is the fulltext-search extension |
| 608 | distributed with SQLite. |
Antoine Pitrou | d67075e | 2010-07-31 22:48:02 +0000 | [diff] [blame] | 609 | |
Georg Brandl | da0a211 | 2010-09-05 11:28:33 +0000 | [diff] [blame] | 610 | (Contributed by R. David Murray and Shashwat Anand; :issue:`8845`.) |
Antoine Pitrou | d67075e | 2010-07-31 22:48:02 +0000 | [diff] [blame] | 611 | |
Georg Brandl | da0a211 | 2010-09-05 11:28:33 +0000 | [diff] [blame] | 612 | * The :mod:`ssl` module has a new class, :class:`~ssl.SSLContext` which serves |
| 613 | as a container for various persistent SSL data, such as protocol settings, |
| 614 | certificates, private keys, and various other options. The |
| 615 | :meth:`~ssl.SSLContext.wrap_socket` method allows to create an SSL socket from |
| 616 | such an SSL context. (Added by Antoine Pitrou; :issue:`8550`.) |
Antoine Pitrou | 4f2a0a8 | 2010-07-31 18:08:33 +0000 | [diff] [blame] | 617 | |
Antoine Pitrou | 0ee4c9f | 2010-10-08 16:46:17 +0000 | [diff] [blame] | 618 | A new function, :func:`ssl.match_hostname`, helps implement server identity |
| 619 | verification for higher-level protocols by implementing the rules of |
| 620 | HTTPS (from :rfc:`2818`), which are also suitable for other protocols. |
| 621 | (Added by Antoine Pitrou, :issue:`1589`). |
| 622 | |
Georg Brandl | da0a211 | 2010-09-05 11:28:33 +0000 | [diff] [blame] | 623 | The :func:`ssl.wrap_socket` constructor function now takes a *ciphers* |
| 624 | argument that's a string listing the encryption algorithms to be allowed; the |
| 625 | format of the string is described `in the OpenSSL documentation |
| 626 | <http://www.openssl.org/docs/apps/ciphers.html#CIPHER_LIST_FORMAT>`__. (Added |
| 627 | by Antoine Pitrou; :issue:`8322`.) |
Antoine Pitrou | 4f2a0a8 | 2010-07-31 18:08:33 +0000 | [diff] [blame] | 628 | |
Antoine Pitrou | 7d15a72 | 2010-11-05 22:13:55 +0000 | [diff] [blame] | 629 | When linked against a recent enough version of OpenSSL, the :mod:`ssl` |
| 630 | module now supports the Server Name Indication extension to the TLS |
| 631 | protocol, allowing for several "virtual hosts" using different certificates |
| 632 | on a single IP/port. This extension is only supported in client mode, |
| 633 | and is activated by passing the *server_hostname* argument to |
| 634 | :meth:`SSLContext.wrap_socket`. |
| 635 | (Added by Antoine Pitrou, :issue:`5639`.) |
| 636 | |
Antoine Pitrou | 4f2a0a8 | 2010-07-31 18:08:33 +0000 | [diff] [blame] | 637 | Various options have been added to the :mod:`ssl` module, such as |
Georg Brandl | da0a211 | 2010-09-05 11:28:33 +0000 | [diff] [blame] | 638 | :data:`~ssl.OP_NO_SSLv2` which allows to force disabling of the insecure and |
| 639 | obsolete SSLv2 protocol. (Added by Antoine Pitrou; :issue:`4870`.) |
Antoine Pitrou | 4f2a0a8 | 2010-07-31 18:08:33 +0000 | [diff] [blame] | 640 | |
Georg Brandl | da0a211 | 2010-09-05 11:28:33 +0000 | [diff] [blame] | 641 | Another change makes the extension load all of OpenSSL's ciphers and digest |
| 642 | algorithms so that they're all available. Some SSL certificates couldn't be |
| 643 | verified, reporting an "unknown algorithm" error. (Reported by Beda Kosata, |
| 644 | and fixed by Antoine Pitrou; :issue:`8484`.) |
Antoine Pitrou | 4f2a0a8 | 2010-07-31 18:08:33 +0000 | [diff] [blame] | 645 | |
Georg Brandl | da0a211 | 2010-09-05 11:28:33 +0000 | [diff] [blame] | 646 | The version of OpenSSL being used is now available as the module attributes |
| 647 | :data:`ssl.OPENSSL_VERSION` (a string), :data:`ssl.OPENSSL_VERSION_INFO` (a |
| 648 | 5-tuple), and :data:`ssl.OPENSSL_VERSION_NUMBER` (an integer). (Added by |
| 649 | Antoine Pitrou; :issue:`8321`.) |
Antoine Pitrou | 4f2a0a8 | 2010-07-31 18:08:33 +0000 | [diff] [blame] | 650 | |
Antoine Pitrou | afb078d | 2010-11-05 22:18:28 +0000 | [diff] [blame] | 651 | * :class:`http.client.HTTPSConnection`, :class:`urllib.request.HTTPSHandler` |
| 652 | and :func:`urllib.request.urlopen` now take optional arguments to allow for |
| 653 | server certificate checking against a set of Certificate Authorities, |
| 654 | as recommended in public uses of HTTPS. |
| 655 | (Added by Antoine Pitrou, :issue:`9003`.) |
| 656 | |
Raymond Hettinger | dc2f9b5 | 2010-12-05 07:02:45 +0000 | [diff] [blame] | 657 | * The command call, ``python -m unittest`` can now accept file paths instead |
| 658 | of module names for running specific tests (:issue:`10620`). |
Antoine Pitrou | d305200 | 2010-09-15 15:09:40 +0000 | [diff] [blame] | 659 | |
Raymond Hettinger | dc2f9b5 | 2010-12-05 07:02:45 +0000 | [diff] [blame] | 660 | * The :mod:`unittest` module has two new methods, |
| 661 | :meth:`~unittest.TestCase.assertWarns` and |
| 662 | :meth:`~unittest.TestCase.assertWarnsRegex` to check that a given warning type |
Raymond Hettinger | 413abbc | 2010-12-05 07:06:47 +0000 | [diff] [blame] | 663 | is triggered by the code under test: |
Antoine Pitrou | d305200 | 2010-09-15 15:09:40 +0000 | [diff] [blame] | 664 | |
Raymond Hettinger | dc2f9b5 | 2010-12-05 07:02:45 +0000 | [diff] [blame] | 665 | >>> with self.assertWarns(DeprecationWarning): |
| 666 | ... legacy_function('XYZ') |
Ezio Melotti | 2baf1a6 | 2010-11-22 12:56:58 +0000 | [diff] [blame] | 667 | |
Raymond Hettinger | dc2f9b5 | 2010-12-05 07:02:45 +0000 | [diff] [blame] | 668 | In addition, the naming in the module has ungone a number of clean-ups. |
| 669 | For example, :meth:`assertRegex` is the new name for :meth:`assertRegexpMatches` |
| 670 | which was misnamed because the test uses :func:`re.search`, not :func:`re.match`. |
| 671 | |
| 672 | To improve consistency, some of long-standing method aliases are being |
| 673 | deprecated in favor of the preferred names: |
| 674 | |
| 675 | - replace :meth:`assert_` with :meth:`.assertTrue` |
| 676 | - replace :meth:`assertEquals` with :meth:`.assertEqual` |
| 677 | - replace :meth:`assertNotEquals` with :meth:`.assertNotEqual` |
| 678 | - replace :meth:`assertAlmostEquals` with :meth:`.assertAlmostEqual` |
| 679 | - replace :meth:`assertNotAlmostEquals` with :meth:`.assertNotAlmostEqual` |
| 680 | |
| 681 | Likewise, the ``TestCase.fail*`` methods deprecated in Python 3.1 are expected |
| 682 | to be removed in Python 3.3. See also the :ref:`deprecated-aliases` section in |
| 683 | the :mod:`unittest` documentation. |
Ezio Melotti | 2baf1a6 | 2010-11-22 12:56:58 +0000 | [diff] [blame] | 684 | |
| 685 | (Contributed by Ezio Melotti; :issue:`9424`.) |
Antoine Pitrou | d305200 | 2010-09-15 15:09:40 +0000 | [diff] [blame] | 686 | |
Georg Brandl | da0a211 | 2010-09-05 11:28:33 +0000 | [diff] [blame] | 687 | * The previously deprecated :func:`string.maketrans` function has been removed |
| 688 | in favor of the static methods, :meth:`bytes.maketrans` and |
Raymond Hettinger | f558ddd | 2009-06-28 21:37:08 +0000 | [diff] [blame] | 689 | :meth:`bytearray.maketrans`. This change solves the confusion around which |
Georg Brandl | da0a211 | 2010-09-05 11:28:33 +0000 | [diff] [blame] | 690 | types were supported by the :mod:`string` module. Now, :class:`str`, |
Raymond Hettinger | f558ddd | 2009-06-28 21:37:08 +0000 | [diff] [blame] | 691 | :class:`bytes`, and :class:`bytearray` each have their own **maketrans** and |
Georg Brandl | da0a211 | 2010-09-05 11:28:33 +0000 | [diff] [blame] | 692 | **translate** methods with intermediate translation tables of the appropriate |
| 693 | type. |
Raymond Hettinger | 6e6565b | 2009-06-28 20:56:11 +0000 | [diff] [blame] | 694 | |
Raymond Hettinger | f558ddd | 2009-06-28 21:37:08 +0000 | [diff] [blame] | 695 | (Contributed by Georg Brandl; :issue:`5675`.) |
| 696 | |
Giampaolo Rodolà | 42382fe | 2010-08-17 16:09:53 +0000 | [diff] [blame] | 697 | * :class:`~poplib.POP3_SSL` class now accepts a *context* parameter, which is a |
| 698 | :class:`ssl.SSLContext` object allowing bundling SSL configuration options, |
| 699 | certificates and private keys into a single (potentially long-lived) |
| 700 | structure. |
| 701 | |
| 702 | (Contributed by Giampaolo Rodolà; :issue:`8807`.) |
| 703 | |
Giampaolo Rodolà | b383dbb | 2010-09-08 22:44:12 +0000 | [diff] [blame] | 704 | * :func:`socket.create_connection` now supports the context manager protocol |
| 705 | to unconditionally consume :exc:`socket.error` exceptions and to close the |
| 706 | socket when done. |
| 707 | |
| 708 | (Contributed by Giampaolo Rodolà; :issue:`9794`.) |
| 709 | |
Giampaolo Rodolà | 977c707 | 2010-10-04 21:08:36 +0000 | [diff] [blame] | 710 | * :class:`asyncore.dispatcher` now provides a |
| 711 | :meth:`~asyncore.dispatcher.handle_accepted()` method |
| 712 | returning a `(sock, addr)` pair which is called when a connection has actually |
| 713 | been established with a new remote endpoint. This is supposed to be used as a |
| 714 | replacement for old :meth:`~asyncore.dispatcher.handle_accept()` and avoids |
| 715 | the user to call :meth:`~asyncore.dispatcher.accept()` directly. |
| 716 | |
| 717 | (Contributed by Giampaolo Rodolà; :issue:`6706`.) |
Georg Brandl | da0a211 | 2010-09-05 11:28:33 +0000 | [diff] [blame] | 718 | |
Nick Coghlan | 543af75 | 2010-10-24 11:23:25 +0000 | [diff] [blame] | 719 | * The :mod:`tempfile` module has a new context manager, |
| 720 | :class:`~tempfile.TemporaryDirectory` which provides easy deterministic |
| 721 | cleanup of temporary directories. |
| 722 | |
| 723 | (Contributed by Neil Schemenauer and Nick Coghlan; :issue:`5178`.) |
| 724 | |
R. David Murray | 7dff9e0 | 2010-11-08 17:15:13 +0000 | [diff] [blame] | 725 | * The :mod:`smtplib` :class:`~smtplib.SMTP` class now accepts a byte string |
| 726 | for the *msg* argument to the :meth:`~smtplib.SMTP.sendmail` method, |
| 727 | and a new method, :meth:`~smtplib.SMTP.send_message` accepts a |
| 728 | :class:`~email.message.Message` object and can optionally obtain the |
| 729 | *from_addr* and *to_addrs* addresses directly from the object. |
| 730 | |
| 731 | (Contributed by R. David Murray, :issue:`10321`.) |
| 732 | |
Nick Coghlan | e0f0465 | 2010-11-21 03:44:04 +0000 | [diff] [blame] | 733 | * The :mod:`inspect` module has a new function :func:`getgenatorstate` |
| 734 | to easily identify the current state of a generator as one of |
| 735 | ``GEN_CREATED``, ``GEN_RUNNING``, ``GEN_SUSPENDED`` or ``GEN_CLOSED``. |
| 736 | |
| 737 | (Contributed by Rodolpho Eckhardt and Nick Coghlan, :issue:`10220`.) |
| 738 | |
Raymond Hettinger | e5e1a98 | 2010-12-05 08:35:21 +0000 | [diff] [blame] | 739 | .. XXX: Create a new section for all changes relating to context managers. |
| 740 | .. XXX: Various ConfigParser changes |
Nick Coghlan | e0f0465 | 2010-11-21 03:44:04 +0000 | [diff] [blame] | 741 | .. XXX: Mention inspect.getattr_static (Michael Foord) |
Nick Coghlan | 9fc443c | 2010-11-30 15:48:08 +0000 | [diff] [blame] | 742 | .. XXX: Mention urllib.parse changes |
| 743 | Issue 9873 (Nick Coghlan): |
| 744 | - ASCII byte sequence support in URL parsing |
| 745 | - named tuple for urldefrag return value |
| 746 | Issue 5468 (Dan Mahn) for urlencode: |
| 747 | - bytes input support |
| 748 | - non-UTF8 percent encoding of non-ASCII characters |
| 749 | Issue 2987 for IPv6 (RFC2732) support in urlparse |
Nick Coghlan | e0f0465 | 2010-11-21 03:44:04 +0000 | [diff] [blame] | 750 | |
Nick Coghlan | 7bb30b7 | 2010-12-03 09:29:11 +0000 | [diff] [blame] | 751 | * The :mod:`pydoc` module now provides a much improved Web server interface, |
| 752 | as well as a new command-line option to automatically open a browser |
| 753 | window to display that server. |
| 754 | |
| 755 | (Contributed by Ron Adam; :issue:`2001`.) |
| 756 | |
Georg Brandl | 3ad4675 | 2010-12-05 07:59:29 +0000 | [diff] [blame] | 757 | .. XXX add something about pdb additions: |
| 758 | |
| 759 | * new commands interact, (un)display, longlist, source, until lineno |
| 760 | * -c option that executes commands as if given in .pdbrc |
| 761 | * SIGINT handler to break a continued program |
| 762 | |
| 763 | .. XXX add optimize flags for py_compile/compileall (issue10553) |
| 764 | |
Raymond Hettinger | 3f9734c | 2010-12-07 01:47:52 +0000 | [diff] [blame^] | 765 | * The new :mod:`sysconfig` module makes it straight-forward to discover |
| 766 | installation paths and configuration variables which vary across platforms and |
| 767 | installs. |
| 768 | |
| 769 | The module offers access simple access functions for platform and version |
| 770 | information: |
| 771 | |
| 772 | * :func:`~sysconfig.get_platform` returning values like *linux-i586* or |
| 773 | *macosx-10.6-ppc*. |
| 774 | * :func:`~sysconfig.get_python_version` returns a Python version string in |
| 775 | the form, "3.2". |
| 776 | |
| 777 | It also provides access to the paths and variables corresponding to one of |
| 778 | seven named schemes used by :mod:`distutils`. Those include *posix_prefix*, |
| 779 | *posix_home*, *posix_user*, *nt*, *nt_user*, *os2*, *os2_home*: |
| 780 | |
| 781 | * :func:`~sysconfig.get_paths` makes a dictionary containing installation paths |
| 782 | for the current installation scheme. |
| 783 | * :func:`~sysconfig.get_config_vars` returns a dictionary of platform specific |
| 784 | variables. |
| 785 | |
| 786 | There is also a convenient command-line interface:: |
| 787 | |
| 788 | C:\Python32>python -m sysconfig |
| 789 | Platform: "win32" |
| 790 | Python version: "3.2" |
| 791 | Current installation scheme: "nt" |
| 792 | |
| 793 | Paths: |
| 794 | data = "C:\Python32" |
| 795 | include = "C:\Python32\Include" |
| 796 | platinclude = "C:\Python32\Include" |
| 797 | platlib = "C:\Python32\Lib\site-packages" |
| 798 | platstdlib = "C:\Python32\Lib" |
| 799 | purelib = "C:\Python32\Lib\site-packages" |
| 800 | scripts = "C:\Python32\Scripts" |
| 801 | stdlib = "C:\Python32\Lib" |
| 802 | |
| 803 | Variables: |
| 804 | BINDIR = "C:\Python32" |
| 805 | BINLIBDEST = "C:\Python32\Lib" |
| 806 | EXE = ".exe" |
| 807 | INCLUDEPY = "C:\Python32\Include" |
| 808 | LIBDEST = "C:\Python32\Lib" |
| 809 | SO = ".pyd" |
| 810 | VERSION = "32" |
| 811 | abiflags = "" |
| 812 | base = "C:\Python32" |
| 813 | exec_prefix = "C:\Python32" |
| 814 | platbase = "C:\Python32" |
| 815 | prefix = "C:\Python32" |
| 816 | projectbase = "C:\Python32" |
| 817 | py_version = "3.2b1" |
| 818 | py_version_nodot = "32" |
| 819 | py_version_short = "3.2" |
| 820 | srcdir = "C:\Python32" |
| 821 | userbase = "C:\Documents and Settings\Raymond\Application Data\Python" |
| 822 | |
| 823 | |
Antoine Pitrou | d42bc51 | 2009-11-10 23:18:31 +0000 | [diff] [blame] | 824 | Multi-threading |
| 825 | =============== |
| 826 | |
Georg Brandl | da0a211 | 2010-09-05 11:28:33 +0000 | [diff] [blame] | 827 | * The mechanism for serializing execution of concurrently running Python threads |
| 828 | (generally known as the GIL or Global Interpreter Lock) has been rewritten. |
| 829 | Among the objectives were more predictable switching intervals and reduced |
| 830 | overhead due to lock contention and the number of ensuing system calls. The |
| 831 | notion of a "check interval" to allow thread switches has been abandoned and |
| 832 | replaced by an absolute duration expressed in seconds. This parameter is |
| 833 | tunable through :func:`sys.setswitchinterval()`. It currently defaults to 5 |
| 834 | milliseconds. |
Antoine Pitrou | d42bc51 | 2009-11-10 23:18:31 +0000 | [diff] [blame] | 835 | |
| 836 | Additional details about the implementation can be read from a `python-dev |
| 837 | mailing-list message |
| 838 | <http://mail.python.org/pipermail/python-dev/2009-October/093321.html>`_ |
Georg Brandl | da0a211 | 2010-09-05 11:28:33 +0000 | [diff] [blame] | 839 | (however, "priority requests" as exposed in this message have not been kept |
| 840 | for inclusion). |
Antoine Pitrou | d42bc51 | 2009-11-10 23:18:31 +0000 | [diff] [blame] | 841 | |
Georg Brandl | 5e73a81 | 2010-04-22 07:02:51 +0000 | [diff] [blame] | 842 | (Contributed by Antoine Pitrou.) |
Antoine Pitrou | d42bc51 | 2009-11-10 23:18:31 +0000 | [diff] [blame] | 843 | |
Antoine Pitrou | 5bab508 | 2009-11-13 22:58:45 +0000 | [diff] [blame] | 844 | * Recursive locks (created with the :func:`threading.RLock` API) now benefit |
Georg Brandl | da0a211 | 2010-09-05 11:28:33 +0000 | [diff] [blame] | 845 | from a C implementation which makes them as fast as regular locks, and between |
| 846 | 10x and 15x faster than their previous pure Python implementation. |
Antoine Pitrou | 5bab508 | 2009-11-13 22:58:45 +0000 | [diff] [blame] | 847 | |
| 848 | (Contributed by Antoine Pitrou; :issue:`3001`.) |
| 849 | |
Georg Brandl | da0a211 | 2010-09-05 11:28:33 +0000 | [diff] [blame] | 850 | * Regular and recursive locks now accept an optional *timeout* argument to their |
Raymond Hettinger | 09e4ebb | 2010-09-06 19:55:51 +0000 | [diff] [blame] | 851 | :meth:`acquire` method. (Contributed by Antoine Pitrou; :issue:`7316`.) |
Georg Brandl | da0a211 | 2010-09-05 11:28:33 +0000 | [diff] [blame] | 852 | |
Antoine Pitrou | e95a9ff | 2010-05-04 23:31:41 +0000 | [diff] [blame] | 853 | Similarly, :meth:`threading.Semaphore.acquire` also gains a *timeout* |
Georg Brandl | da0a211 | 2010-09-05 11:28:33 +0000 | [diff] [blame] | 854 | argument. (Contributed by Torsten Landschoff; :issue:`850728`.) |
Antoine Pitrou | e95a9ff | 2010-05-04 23:31:41 +0000 | [diff] [blame] | 855 | |
Antoine Pitrou | d42bc51 | 2009-11-10 23:18:31 +0000 | [diff] [blame] | 856 | |
Raymond Hettinger | 92ba286 | 2010-09-06 01:16:46 +0000 | [diff] [blame] | 857 | Optimizations |
| 858 | ============= |
Raymond Hettinger | 6e6565b | 2009-06-28 20:56:11 +0000 | [diff] [blame] | 859 | |
Raymond Hettinger | 92ba286 | 2010-09-06 01:16:46 +0000 | [diff] [blame] | 860 | A number of small performance enhancements have been added: |
Raymond Hettinger | 6e6565b | 2009-06-28 20:56:11 +0000 | [diff] [blame] | 861 | |
Antoine Pitrou | d305200 | 2010-09-15 15:09:40 +0000 | [diff] [blame] | 862 | * Python's peephole optimizer now recognizes patterns such ``x in {1, 2, 3}`` as |
Raymond Hettinger | 92ba286 | 2010-09-06 01:16:46 +0000 | [diff] [blame] | 863 | being a test for membership in a set of constants. The optimizer recasts the |
| 864 | :class:`set` as a :class:`frozenset` and stores the pre-built constant. |
| 865 | |
| 866 | Now that the speed penalty is gone, it is practical to start writing |
| 867 | membership tests using set-notation. This style is both semantically clear |
| 868 | and operationally fast:: |
| 869 | |
| 870 | extension = name.rpartition('.')[2] |
| 871 | if extension in {'xml', 'html', 'xhtml', 'css'}: |
| 872 | handle(name) |
| 873 | |
| 874 | (Patch and additional tests by Dave Malcolm; :issue:`6690`). |
| 875 | |
Antoine Pitrou | d305200 | 2010-09-15 15:09:40 +0000 | [diff] [blame] | 876 | * Serializing and unserializing data using the :mod:`pickle` module is now |
Raymond Hettinger | dadf93c | 2010-12-05 02:56:21 +0000 | [diff] [blame] | 877 | several times faster. |
| 878 | |
| 879 | (Contributed by Alexandre Vassalotti, Antoine Pitrou |
Antoine Pitrou | ff150f2 | 2010-10-22 21:41:05 +0000 | [diff] [blame] | 880 | and the Unladen Swallow team in :issue:`9410` and :issue:`3873`.) |
Antoine Pitrou | d305200 | 2010-09-15 15:09:40 +0000 | [diff] [blame] | 881 | |
Raymond Hettinger | c269ae8 | 2010-12-05 01:01:52 +0000 | [diff] [blame] | 882 | * The `Timsort algorithm <http://en.wikipedia.org/wiki/Timsort>`_ used in |
| 883 | :meth:`list.sort` and :func:`sorted` now runs faster and used less memory |
| 884 | when called with a :term:`key function`. Previously, every element of |
| 885 | a list was wrapped with a temporary object that remembered the key value |
| 886 | associated with each element. Now, an array of keys and values are |
| 887 | sorted in parallel. This save the memory consumed by the sort wrappers, |
| 888 | and it saves time lost from during comparisons which where delegated |
| 889 | by the sort wrappers. |
| 890 | |
| 891 | (Patch by Daniel Stuzback in :issue:`9915`.) |
| 892 | |
Raymond Hettinger | dadf93c | 2010-12-05 02:56:21 +0000 | [diff] [blame] | 893 | * JSON decoding performance is improved and memory consumption is reduced |
Raymond Hettinger | 413abbc | 2010-12-05 07:06:47 +0000 | [diff] [blame] | 894 | whenever the same string is repeated for multiple keys. Also, JSON encoding |
Raymond Hettinger | dadf93c | 2010-12-05 02:56:21 +0000 | [diff] [blame] | 895 | now uses the C speedups when the ``sort_keys`` argument is true. |
| 896 | |
| 897 | (Contributed by Antoine Pitrou in :issue:`7451` and by Raymond Hettinger and |
| 898 | Antoine Pitrou in :issue:`10314`.) |
| 899 | |
| 900 | * The fast-search algorithm in stringlib is now used by the :meth:`split`, |
| 901 | :meth:`rsplit`, :meth:`splitlines` and :meth:`replace` methods on |
| 902 | :class:`bytes`, :class:`bytearray` and :class:`str` objects. Likewise, the |
| 903 | algorithm is also used by :meth:`rfind`, :meth:`rindex`, :meth:`rsplit` and |
| 904 | :meth:`rpartition`. |
| 905 | |
| 906 | (Patch by Florent Xicluna in :issue:`7622` and :issue:`7462`.) |
| 907 | |
Raymond Hettinger | d8fae4e | 2010-12-05 05:39:54 +0000 | [diff] [blame] | 908 | There were several other minor optimizations. Set differencing now runs faster |
| 909 | when one operand is much larger than the other (Patch by Andress Bennetts in |
| 910 | :issue:`8685`). The :meth:`array.repeat` method has a faster implementation |
| 911 | (:issue:`1569291` by Alexander Belopolsky). The :class:`BaseHTTPRequestHandler` |
| 912 | has more efficient buffering (:issue:`3709` by Andrew Schaaf). The |
| 913 | multi-argument form of :func:`operator.attrgetter` now function runs slightly |
| 914 | faster (:issue:`10160` by Christos Georgiou). And :class:`ConfigParser` loads |
| 915 | multi-line arguments a bit faster (:issue:`7113` by Łukasz Langa). |
| 916 | |
Antoine Pitrou | d305200 | 2010-09-15 15:09:40 +0000 | [diff] [blame] | 917 | |
Victor Stinner | 47ce965 | 2010-10-29 00:57:35 +0000 | [diff] [blame] | 918 | Unicode |
| 919 | ======= |
Victor Stinner | 94908bb | 2010-08-18 21:23:25 +0000 | [diff] [blame] | 920 | |
Alexander Belopolsky | 507e3f8 | 2010-12-02 00:05:57 +0000 | [diff] [blame] | 921 | Python has been updated to Unicode 6.0.0. The new features of the |
| 922 | Unicode Standard that will affect Python users include: |
| 923 | |
| 924 | * adds 2,088 characters, including over 1,000 additional symbols—chief |
Raymond Hettinger | c74d518 | 2010-12-02 01:38:25 +0000 | [diff] [blame] | 925 | among them the additional emoji symbols, which are especially |
| 926 | important for mobile phones; |
Alexander Belopolsky | 507e3f8 | 2010-12-02 00:05:57 +0000 | [diff] [blame] | 927 | |
| 928 | * corrects character properties for existing characters including |
Alexander Belopolsky | 507e3f8 | 2010-12-02 00:05:57 +0000 | [diff] [blame] | 929 | |
Raymond Hettinger | c74d518 | 2010-12-02 01:38:25 +0000 | [diff] [blame] | 930 | - a general category change to two Kannada characters (U+0CF1, |
| 931 | U+0CF2), which has the effect of making them newly eligible for |
| 932 | inclusion in identifiers; |
| 933 | |
| 934 | - a general category change to one New Tai Lue numeric character |
| 935 | (U+19DA), which would have the effect of disqualifying it from |
| 936 | inclusion in identifiers unless grandfathering measures are in place |
| 937 | for the defining identifier syntax. |
Alexander Belopolsky | 507e3f8 | 2010-12-02 00:05:57 +0000 | [diff] [blame] | 938 | |
Éric Araujo | 4234ad4 | 2010-09-05 17:32:25 +0000 | [diff] [blame] | 939 | The :mod:`os` module has two new functions: :func:`~os.fsencode` and |
Victor Stinner | 47ce965 | 2010-10-29 00:57:35 +0000 | [diff] [blame] | 940 | :func:`~os.fsdecode`. Add :data:`os.environb`: bytes version of |
| 941 | :data:`os.environ`, :func:`os.getenvb` function and |
| 942 | :data:`os.supports_bytes_environ` constant. |
Victor Stinner | e8d5145 | 2010-08-19 01:05:19 +0000 | [diff] [blame] | 943 | |
Georg Brandl | 326c57d | 2010-11-26 12:10:06 +0000 | [diff] [blame] | 944 | ``'mbcs'`` encoding doesn't ignore the error handler argument any more. By |
Victor Stinner | 47ce965 | 2010-10-29 00:57:35 +0000 | [diff] [blame] | 945 | default (strict mode), it raises an UnicodeDecodeError on undecodable byte |
| 946 | sequence and UnicodeEncodeError on unencodable character. To get the ``'mbcs'`` |
| 947 | encoding of Python 3.1, use ``'ignore'`` error handler to decode and |
| 948 | ``'replace'`` error handler to encode. ``'mbcs'`` supports ``'strict'`` and |
| 949 | ``'ignore'`` error handlers for decoding, and ``'strict'`` and ``'replace'`` |
| 950 | for encoding. |
| 951 | |
| 952 | On Mac OS X, Python uses ``'utf-8'`` to decode the command line arguments, |
| 953 | instead of the locale encoding (which is ISO-8859-1 if the ``LANG`` environment |
| 954 | variable is not set). |
| 955 | |
| 956 | By default, tarfile uses ``'utf-8'`` encoding on Windows (instead of |
| 957 | ``'mbcs'``), and the ``'surrogateescape'`` error handler on all operating |
| 958 | systems. |
Antoine Pitrou | d305200 | 2010-09-15 15:09:40 +0000 | [diff] [blame] | 959 | |
Victor Stinner | 94908bb | 2010-08-18 21:23:25 +0000 | [diff] [blame] | 960 | |
Raymond Hettinger | 1fa7682 | 2010-12-06 23:31:36 +0000 | [diff] [blame] | 961 | Documentation |
| 962 | ============= |
| 963 | |
| 964 | The documentation continues to be improved. |
| 965 | |
| 966 | A table of quick links has been added to the top of lengthy sections such as |
| 967 | :ref:`built-in-funcs`. In the case of :mod:`itertools`, the links are |
| 968 | accompanied by tables of cheatsheet-style summaries to provide an overview and |
| 969 | memory jog without having to read all of the docs. |
| 970 | |
| 971 | In some cases, the pure python source code can be helpful adjunct to the docs, |
| 972 | so now some modules feature quick links to the latest version of the source |
| 973 | code. For example, the :mod:`functools` module documentation has a quick link |
| 974 | at the top labeled :source:`functools Python source code <Lib/functools.py>`. |
| 975 | |
| 976 | The docs now contain more examples and recipes. In particular, :mod:`re` module |
| 977 | has an extensive section, :ref:`re-examples`. Likewise, the :mod:`itertools` |
| 978 | module continues to be updated with new :ref:`itertools-recipes`. |
| 979 | |
| 980 | |
| 981 | IDLE |
| 982 | ==== |
Raymond Hettinger | 6e6565b | 2009-06-28 20:56:11 +0000 | [diff] [blame] | 983 | |
Raymond Hettinger | dadf93c | 2010-12-05 02:56:21 +0000 | [diff] [blame] | 984 | * The format menu now has an option to clean-up source files by strip trailing |
| 985 | whitespace (:issue:`5150`). |
Raymond Hettinger | 6e6565b | 2009-06-28 20:56:11 +0000 | [diff] [blame] | 986 | |
| 987 | |
| 988 | Build and C API Changes |
| 989 | ======================= |
| 990 | |
| 991 | Changes to Python's build process and to the C API include: |
| 992 | |
Georg Brandl | da0a211 | 2010-09-05 11:28:33 +0000 | [diff] [blame] | 993 | * The C functions that access the Unicode Database now accept and return |
| 994 | characters from the full Unicode range, even on narrow unicode builds |
Raymond Hettinger | 1784ff0 | 2010-09-05 01:00:19 +0000 | [diff] [blame] | 995 | (Py_UNICODE_TOLOWER, Py_UNICODE_ISDECIMAL, and others). A visible difference |
Georg Brandl | da0a211 | 2010-09-05 11:28:33 +0000 | [diff] [blame] | 996 | in Python is that :func:`unicodedata.numeric` now returns the correct value |
| 997 | for large code points, and :func:`repr` may consider more characters as |
| 998 | printable. |
Raymond Hettinger | 6e6565b | 2009-06-28 20:56:11 +0000 | [diff] [blame] | 999 | |
Raymond Hettinger | 1784ff0 | 2010-09-05 01:00:19 +0000 | [diff] [blame] | 1000 | (Reported by Bupjoe Lee and fixed by Amaury Forgeot D'Arc; :issue:`5127`.) |
| 1001 | |
Georg Brandl | da0a211 | 2010-09-05 11:28:33 +0000 | [diff] [blame] | 1002 | * Computed gotos are now enabled by default on supported compilers (which are |
Raymond Hettinger | db9044e | 2010-09-06 01:29:23 +0000 | [diff] [blame] | 1003 | detected by the configure script). They can still be disabled selectively by |
Georg Brandl | da0a211 | 2010-09-05 11:28:33 +0000 | [diff] [blame] | 1004 | specifying ``--without-computed-gotos``. |
Raymond Hettinger | 1784ff0 | 2010-09-05 01:00:19 +0000 | [diff] [blame] | 1005 | |
Georg Brandl | da0a211 | 2010-09-05 11:28:33 +0000 | [diff] [blame] | 1006 | (Contributed by Antoine Pitrou; :issue:`9203`.) |
| 1007 | |
Amaury Forgeot d'Arc | feb7307 | 2010-09-12 22:42:57 +0000 | [diff] [blame] | 1008 | * The option ``--with-wctype-functions`` was removed. The built-in unicode |
| 1009 | database is now used for all functions. |
| 1010 | |
| 1011 | (Contributed by Amaury Forgeot D'Arc; :issue:`9210`.) |
| 1012 | |
Skip Montanaro | 961aaf5 | 2010-10-17 22:22:24 +0000 | [diff] [blame] | 1013 | * Hash values are now values of a new type, Py_hash_t, which is defined to |
| 1014 | be the same size as a pointer. Previously they were of type long, which |
Raymond Hettinger | 1fa7682 | 2010-12-06 23:31:36 +0000 | [diff] [blame] | 1015 | on some 64-bit operating systems is still only 32 bits long. As a result |
| 1016 | of this fix, :class:`set` and :class:`dict` can now hold more than ``2**32`` |
| 1017 | entries on builds with 64-bit pointers (previously, they could grow to |
| 1018 | that size but their performance degraded catastrophically). |
Skip Montanaro | 961aaf5 | 2010-10-17 22:22:24 +0000 | [diff] [blame] | 1019 | |
| 1020 | (Contributed by Benjamin Peterson; :issue:`9778`.) |
| 1021 | |
Raymond Hettinger | 6e6565b | 2009-06-28 20:56:11 +0000 | [diff] [blame] | 1022 | |
Raymond Hettinger | f558ddd | 2009-06-28 21:37:08 +0000 | [diff] [blame] | 1023 | Porting to Python 3.2 |
Raymond Hettinger | 6e6565b | 2009-06-28 20:56:11 +0000 | [diff] [blame] | 1024 | ===================== |
| 1025 | |
Georg Brandl | da0a211 | 2010-09-05 11:28:33 +0000 | [diff] [blame] | 1026 | This section lists previously described changes and other bugfixes that may |
| 1027 | require changes to your code: |
Raymond Hettinger | 6e6565b | 2009-06-28 20:56:11 +0000 | [diff] [blame] | 1028 | |
Antoine Pitrou | cd889af | 2010-10-06 21:13:56 +0000 | [diff] [blame] | 1029 | * The :mod:`nntplib` module was reworked extensively, meaning that its APIs |
| 1030 | are often incompatible with the 3.1 APIs. |
| 1031 | |
Raymond Hettinger | 1fa7682 | 2010-12-06 23:31:36 +0000 | [diff] [blame] | 1032 | * :class:`bytearray` objects can no longer be used as filenames; instead, |
| 1033 | they should be converted to :class:`bytes`. |
Victor Stinner | dcb2403 | 2010-04-22 12:08:36 +0000 | [diff] [blame] | 1034 | |
Victor Stinner | 25e8ec4 | 2010-06-25 00:02:38 +0000 | [diff] [blame] | 1035 | * PyArg_Parse*() functions: |
Victor Stinner | 3dcb5ac | 2010-06-08 22:54:19 +0000 | [diff] [blame] | 1036 | |
Victor Stinner | 25e8ec4 | 2010-06-25 00:02:38 +0000 | [diff] [blame] | 1037 | * "t#" format has been removed: use "s#" or "s*" instead |
| 1038 | * "w" and "w#" formats has been removed: use "w*" instead |
| 1039 | |
Georg Brandl | 60203b4 | 2010-10-06 10:11:56 +0000 | [diff] [blame] | 1040 | * The :c:type:`PyCObject` type, deprecated in 3.1, has been removed. To wrap |
| 1041 | opaque C pointers in Python objects, the :c:type:`PyCapsule` API should be used |
Éric Araujo | 4234ad4 | 2010-09-05 17:32:25 +0000 | [diff] [blame] | 1042 | instead; the new type has a well-defined interface for passing typing safety |
Georg Brandl | da0a211 | 2010-09-05 11:28:33 +0000 | [diff] [blame] | 1043 | information and a less complicated signature for calling a destructor. |
Victor Stinner | 0cbec57 | 2010-09-12 20:32:57 +0000 | [diff] [blame] | 1044 | |
Raymond Hettinger | 6f04adc | 2010-12-04 22:56:25 +0000 | [diff] [blame] | 1045 | * The :func:`sys.setfilesystemencoding` function was removed because |
Raymond Hettinger | 1fa7682 | 2010-12-06 23:31:36 +0000 | [diff] [blame] | 1046 | it had a flawed design. |