blob: 9e7afee83ca3c1ea00a615eb32caae92470f0b4d [file] [log] [blame]
Skip Montanaroe5d7f7f2002-09-20 14:16:59 +00001+++++++++++
2Python News
3+++++++++++
4
Anthony Baxterb0c66302004-11-04 05:23:17 +00005(editors: check NEWS.help for information about editing NEWS using ReST.)
6
Barry Warsaw97f005d2008-12-03 16:46:14 +00007What's New in Python 3.1 alpha 0
8================================
Barry Warsawd8ccd1c2008-11-21 15:13:37 +00009
Antoine Pitrou3ad3a0d2008-12-18 17:08:32 +000010*Release date: XX-XXX-200X*
Barry Warsawd8ccd1c2008-11-21 15:13:37 +000011
12Core and Builtins
13-----------------
14
Mark Dickinsone5e298f2009-01-12 20:49:19 +000015- Issue #4910: Builtin int() function and PyNumber_Long/PyNumber_Int API
16 function no longer attempt to call the __long__ slot to convert an object
17 to an integer. Only the __int__ and __trunc__ slots are examined.
18
Martin v. Löwis7c2b66c2009-01-12 08:21:03 +000019- Issue #4893: Use NT threading on CE.
20
Martin v. Löwisec59d042009-01-12 07:59:10 +000021- Issue #4915: Port sysmodule to Windows CE.
22
Antoine Pitrouab868312009-01-10 15:40:25 +000023- Issue #4868: utf-8, utf-16 and latin1 decoding are now 2x to 4x faster. The
24 common cases are optimized thanks to a dedicated fast path and a moderate
25 amount of loop unrolling.
26
Antoine Pitrou14b78f52009-01-09 22:27:08 +000027- Issue #4074: Change the criteria for doing a full garbage collection (i.e.
28 collecting the oldest generation) so that allocating lots of objects without
29 destroying them does not show quadratic performance. Based on a proposal by
30 Martin von Löwis at
31 http://mail.python.org/pipermail/python-dev/2008-June/080579.html.
32
Antoine Pitrou8043cf82009-01-09 19:54:29 +000033- Issue #4604: Some objects of the I/O library could still be used after
34 having been closed (for instance, a read() call could return some
35 previously buffered data). Patch by Dmitry Vasiliev.
36
Antoine Pitrou05608432009-01-09 18:53:14 +000037- Issue #4705: Fix the -u ("unbuffered binary stdout and stderr") command-line
38 flag to work properly. Furthermore, when specifying -u, the text stdout
39 and stderr streams have line-by-line buffering enabled (the default being
40 to buffer arbitrary chunks of data).
41
Raymond Hettinger35631532009-01-09 03:58:09 +000042- The internal table, _PyLong_DigitValue, is now an array of unsigned chars
43 instead of ints (reducing its size from 4 to 8 times thereby reducing
44 Python's overall memory).
45
Antoine Pitroud35cbf62009-01-06 19:02:24 +000046- Issue #1180193: When importing a module from a .pyc (or .pyo) file with
47 an existing .py counterpart, override the co_filename attributes of all
48 code objects if the original filename is obsolete (which can happen if the
49 file has been renamed, moved, or if it is accessed through different paths).
50 Patch by Ziga Seilnacht and Jean-Paul Calderone.
51
Antoine Pitrouc3b39242009-01-03 16:59:18 +000052- Issue #4580: Fix slicing of memoryviews when the item size is greater than
53 one byte. Also fixes the meaning of len() so that it returns the number of
54 items, rather than the size in bytes.
55
Martin v. Löwis5c88d812009-01-02 20:47:48 +000056- Issue #4075: Use OutputDebugStringW in Py_FatalError.
57
Amaury Forgeot d'Arc374e2202009-01-01 23:05:36 +000058- Issue #4747: When the terminal does not use utf-8, executing a script with
59 non-ascii characters in its name could fail with a "SyntaxError: None" error.
60
Hirokazu Yamamoto0f22d692009-01-01 16:03:45 +000061- Issue #4797: IOError.filename was not set when _fileio.FileIO failed to open
Hirokazu Yamamoto8ed9a802009-01-01 16:09:52 +000062 file with `bytes' filename on Windows.
Hirokazu Yamamoto0f22d692009-01-01 16:03:45 +000063
Antoine Pitrou7ddda782009-01-01 15:35:33 +000064- Issue #3680: Reference cycles created through a dict, set or deque iterator
65 did not get collected.
66
Nick Coghlanf1f2f682008-12-30 07:29:12 +000067- Issue #4701: PyObject_Hash now implicitly calls PyType_Ready on types
68 where the tp_hash and tp_dict slots are both NULL.
69
Georg Brandlccc47b62008-12-28 11:44:14 +000070- Issue #4759: None is now allowed as the first argument of
71 bytearray.translate(). It was always allowed for bytes.translate().
72
Skip Montanaro80072cb2008-12-23 03:51:14 +000073- Added test case to ensure attempts to read from a file opened for writing
74 fail.
75
Antoine Pitrou51f3ef92008-12-20 13:14:23 +000076- Issue #3106: Speedup some comparisons (str/str and int/int).
77
Antoine Pitrouf289ae62008-12-18 11:06:25 +000078- Issue #2183: Simplify and optimize bytecode for list, dict and set
79 comprehensions. Original patch for list comprehensions by Neal Norwitz.
80
Antoine Pitrou621601a2008-12-17 23:18:19 +000081- Issue #2467: gc.DEBUG_STATS reported invalid elapsed times. Also, always
82 print elapsed times, not only when some objects are uncollectable /
83 unreachable. Original patch by Neil Schemenauer.
84
Mark Dickinson54bc1ec2008-12-17 16:19:07 +000085- Issue #3439: Add a bit_length method to int.
86
Mark Dickinsonda2706b2008-12-11 18:03:03 +000087- Issue #2173: When getting device encoding, check that return value of
88 nl_langinfo is not the empty string. This was causing silent build
89 failures on OS X.
90
Jeffrey Yasskinb9083f32008-12-11 06:18:33 +000091- Issue #4597: Fixed several opcodes that weren't always propagating
92 exceptions.
93
94- Issue #4589: Fixed exception handling when the __exit__ function of a
Amaury Forgeot d'Arc10b24e82008-12-10 23:49:33 +000095 context manager returns a value that cannot be converted to a bool.
96
Mark Dickinsonfd24b322008-12-06 15:33:31 +000097- Issue #4445: Replace "sizeof(PyBytesObject)" with
98 "offsetof(PyBytesObject, ob_sval) + 1" when allocating memory for
99 bytes instances. On a typical machine this saves 3 bytes of memory
100 (on average) per allocation of a bytes instance.
101
Christian Heimesa872de52008-12-05 08:26:55 +0000102- Issue #4533: File read operation was dreadfully slow due to a slowly
103 growing read buffer. Fixed by using the same growth rate algorithm as
104 Python 2.x.
105
Antoine Pitrou5504e892008-12-06 21:27:53 +0000106- Issue #4509: Various issues surrounding resize of bytearray objects to
107 which there are buffer exports (e.g. memoryview instances).
108
Georg Brandlf9734072008-12-07 15:30:06 +0000109- Issue #4233: Changed semantic of ``_fileio.FileIO``'s ``close()``
110 method on file objects with closefd=False. The file descriptor is still
111 kept open but the file object behaves like a closed file. The ``FileIO``
112 object also got a new readonly attribute ``closefd``.
Matthias Klose7e1b8fa2008-11-26 17:22:04 +0000113
Antoine Pitroubc420402008-12-07 20:14:49 +0000114- Issue #4569: Interpreter crash when mutating a memoryview with an item size
115 larger than 1.
116
Benjamin Peterson371ccfb2008-12-27 19:03:36 +0000117- Issue #4748: Lambda generators no longer return a value.
118
Barry Warsawd8ccd1c2008-11-21 15:13:37 +0000119Library
120-------
121
Gregory P. Smithde3369f2009-01-12 04:50:11 +0000122- Issue #3826 and #4791: The socket module now closes the underlying socket
123 appropriately when it is being used via socket.makefile() objects
124 rather than delaying the close by waiting for garbage collection to do it.
125
Antoine Pitrou308705e2009-01-10 16:22:51 +0000126- Issue #3860: GzipFile and BZ2File now support the context manager protocol.
127
Thomas Heller5e4e4272009-01-08 09:34:20 +0000128- Issue #4867: Fixed a crash in ctypes when passing a string to a
129 function without defining argtypes.
130
Antoine Pitrou42db3ef2009-01-04 21:37:59 +0000131- Issue #4272: Add an optional argument to the GzipFile constructor to override
132 the timestamp in the gzip stream. The default value remains the current time.
133 The information can be used by e.g. gunzip when decompressing. Patch by
134 Jacques Frechet.
135
Mark Dickinson315a20a2009-01-04 21:34:18 +0000136- Restore Python 2.3 compatibility for decimal.py.
137
Guilherme Polo3f5f8222009-01-03 22:00:39 +0000138- Issue #3638: Remove functions from _tkinter module level that depend on
139 TkappObject to work with multiple threads.
140
Antoine Pitrou38a66ad2009-01-03 18:41:49 +0000141- Issue #4718: Adapt the wsgiref package so that it actually works with
142 Python 3.x, in accordance with the `official amendments of the spec
143 <http://www.wsgi.org/wsgi/Amendments_1.0>`_.
144
Raymond Hettinger771ed762009-01-03 19:20:32 +0000145- Issue #4796: Added Decimal.from_float() and Context.create_decimal_from_float()
146 to the decimal module.
147
Georg Brandl3a9b0622009-01-03 22:07:57 +0000148- Fractions.from_float() no longer loses precision for integers too big to
149 cast as floats.
150
Mark Dickinsonf9236412009-01-02 23:23:21 +0000151- Issue #4812: add missing underscore prefix to some internal-use-only
152 constants in the decimal module. (Dec_0 becomes _Dec_0, etc.)
153
Georg Brandl3a9b0622009-01-03 22:07:57 +0000154- Issue 4790: The nsmallest() and nlargest() functions in the heapq module
155 did unnecessary work in the common case where no key function was specified.
156
Georg Brandlb1441c72009-01-03 22:33:39 +0000157- Issue #4795: inspect.isgeneratorfunction() returns False instead of None when
158 the function is not a generator.
159
Tarek Ziadé9df8ce32008-12-30 23:09:20 +0000160- Issue #4702: Throwing a DistutilsPlatformError instead of IOError in case
161 no MSVC compiler is found under Windows. Original patch by Philip Jenvey.
162
Tarek Ziadé4450dcf2008-12-29 22:38:38 +0000163- Issue #4646: distutils was choking on empty options arg in the setup
164 function. Original patch by Thomas Heller.
165
Martin v. Löwis1f1620e2008-12-29 16:27:13 +0000166- Issue #3767: Convert Tk object to string in tkColorChooser.
167
Martin v. Löwise354d782008-12-29 16:03:04 +0000168- Issue #3248: Allow placing ScrolledText in a PanedWindow.
169
Antoine Pitrou5acd41e2008-12-28 14:29:00 +0000170- Issue #4444: Allow assertRaises() to be used as a context handler, so that
171 the code under test can be written inline if more practical.
172
Georg Brandldb7b6b92009-01-01 15:53:14 +0000173- Issue #4739: Add pydoc help topics for symbols, so that e.g. help('@')
174 works as expected in the interactive environment.
175
Antoine Pitroudb5fe662008-12-27 15:50:40 +0000176- Issue #4756: zipfile.is_zipfile() now supports file-like objects. Patch by
177 Gabriel Genellina.
178
Antoine Pitrou180a3362008-12-14 16:36:46 +0000179- Issue #4574: reading an UTF16-encoded text file crashes if \r on 64-char
180 boundary.
181
Nick Coghlanf088e5e2008-12-14 11:50:48 +0000182- Issue #4223: inspect.getsource() will now correctly display source code
183 for packages loaded via zipimport (or any other conformant PEP 302
184 loader). Original patch by Alexander Belopolsky.
185
186- Issue #4201: pdb can now access and display source code loaded via
187 zipimport (or any other conformant PEP 302 loader). Original patch by
188 Alexander Belopolsky.
189
190- Issue #4197: doctests in modules loaded via zipimport (or any other PEP
191 302 conformant loader) will now work correctly in most cases (they
192 are still subject to the constraints that exist for all code running
193 from inside a module loaded via a PEP 302 loader and attempting to
194 perform IO operations based on __file__). Original patch by
195 Alexander Belopolsky.
196
197- Issues #4082 and #4512: Add runpy support to zipimport in a manner that
198 allows backporting to maintenance branches. Original patch by
199 Alexander Belopolsky.
200
Antoine Pitrou7c59bc62008-12-13 23:20:54 +0000201- Issue #4163: textwrap module: allow word splitting on a hyphen preceded by
202 a non-ASCII letter.
203
Lars Gustäbel43fcf3a2008-12-12 14:58:38 +0000204- Issue #4616: TarFile.utime(): Restore directory times on Windows.
205
Benjamin Peterson433f32c2008-12-12 01:25:05 +0000206- Issue #4021: tokenize.detect_encoding() now raises a SyntaxError when the
207 codec cannot be found. This is for compatibility with the builtin behavior.
208
Facundo Batista708d5812008-12-11 04:20:07 +0000209- Issue #4084: Fix max, min, max_mag and min_mag Decimal methods to
210 give correct results in the case where one argument is a quiet NaN
211 and the other is a finite number that requires rounding.
212
Skip Montanaroc1ce2862008-12-06 17:25:02 +0000213- Issue #4483: _dbm module now builds on systems with gdbm & gdbm_compat
214 libs.
215
Georg Brandlf9734072008-12-07 15:30:06 +0000216- Added the subprocess.check_call_output() convenience function to get output
217 from a subprocess on success or raise an exception on error.
218
219- Issue #1055234: cgi.parse_header(): Fixed parsing of header parameters to
220 support unusual filenames (such as those containing semi-colons) in
221 Content-Disposition headers.
222
223- Issue #4384: Added logging integration with warnings module using
224 captureWarnings(). This change includes a NullHandler which does nothing;
225 it will be of use to library developers who want to avoid the "No handlers
226 could be found for logger XXX" message which can appear if the library user
227 doesn't configure logging.
228
229- Issue #3741: DISTUTILS_USE_SDK set causes msvc9compiler.py to raise an
230 exception.
231
Georg Brandleee31162008-12-07 15:15:22 +0000232- Issue #4529: fix the parser module's validation of try-except-finally
233 statements.
234
235- Issue #4458: getopt.gnu_getopt() now recognizes a single "-" as an argument,
236 not a malformed option.
237
238- Added the subprocess.check_output() convenience function to get output
239 from a subprocess on success or raise an exception on error.
240
Amaury Forgeot d'Arcbc9d4742008-12-05 01:40:43 +0000241- Issue #4542: On Windows, binascii.crc32 still accepted str as binary input;
242 the corresponding tests now pass.
243
Amaury Forgeot d'Arcbc2ce572008-12-05 01:02:21 +0000244- Issue #4537: webbrowser.UnixBrowser would fail to open the browser because
245 it was calling the wrong open() function.
246
Fred Drake9a0a65b2008-12-04 19:24:50 +0000247- Issue #1055234: cgi.parse_header(): Fixed parsing of header parameters to
248 support unusual filenames (such as those containing semi-colons) in
249 Content-Disposition headers.
250
Matthias Klose2c7e3ee2009-01-10 17:08:25 +0000251- Issue #4861: ctypes.util.find_library(): Robustify. Fix library detection on
252 biarch systems. Try to rely on ldconfig only, without using objdump and gcc.
253
Antoine Pitrou2a013ea2008-12-27 20:39:28 +0000254Tools/Demos
255-----------
256
257- Issue #4677: add two list comprehension tests to pybench.
258
Georg Brandlb1441c72009-01-03 22:33:39 +0000259
260Build
261-----
262
Martin v. Löwise97637d2009-01-11 09:45:11 +0000263- Issue #4895: Use _strdup on Windows CE.
264
Georg Brandlb1441c72009-01-03 22:33:39 +0000265- Issue #4472: "configure --enable-shared" now works on OSX
266
267- Issues #4728 and #4060: WORDS_BIGEDIAN is now correct in Universal builds.
268
269- Issue #4389: Add icon to the uninstall entry in "add-and-remove-programs".
270
271- Issue #4289: Remove Cancel button from AdvancedDlg.
272
273- Issue #1656675: Register a drop handler for .py* files on Windows.
274
275- Issue #4120: Exclude manifest from extension modules in VS2008.
276
277- Issue #4091: Install pythonxy.dll in system32 again.
278
279- Issue #4018: Disable "for me" installations on Vista.
280
281- Issue #3758: Add ``patchcheck`` build target to .PHONY.
282
283- Issue #4204: Fixed module build errors on FreeBSD 4.
284
285
286C-API
287-----
288
289- Issue #4720: The format for PyArg_ParseTupleAndKeywords can begin with '|'.
290
291- Issue #3632: from the gdb debugger, the 'pyo' macro can now be called when
292 the GIL is released, or owned by another thread.
293
294- Issue #4122: On Windows, fix a compilation error when using the
295 Py_UNICODE_ISSPACE macro in an extension module.
296
297
Mark Dickinsone675f082008-12-11 21:56:00 +0000298Extension Modules
299-----------------
300
Martin v. Löwisa9b54c72009-01-11 09:30:50 +0000301- Issue #4279: Fix build of parsermodule under Cygwin.
302
Antoine Pitroubcd5cbe2009-01-08 21:17:16 +0000303- Issue #4751: hashlib now releases the GIL when hashing large buffers
304 (with a hardwired threshold of 2048 bytes), allowing better parallelization
305 on multi-CPU systems. Contributed by Lukas Lueg (ebfe) and Victor Stinner.
306
Martin v. Löwis5c88d812009-01-02 20:47:48 +0000307- Issue #4051: Prevent conflict of UNICODE macros in cPickle.
308
Antoine Pitrou31f30b12009-01-02 17:34:35 +0000309- Issue #4738: Each zlib object now has a separate lock, allowing to compress
310 or decompress several streams at once on multi-CPU systems. Also, the GIL
311 is now released when computing the CRC of a large buffer. Patch by ebfe.
312
Georg Brandlb1441c72009-01-03 22:33:39 +0000313- Issue #4228: Pack negative values the same way as 2.4 in struct's L format.
314
Martin v. Löwis05bfe1f2008-12-29 18:21:47 +0000315- Issue #1040026: Fix os.times result on systems where HZ is incorrect.
316
Mark Dickinsone675f082008-12-11 21:56:00 +0000317- Issues #3167, #3682: Fix test_math failures for log, log10 on Solaris,
318 OpenBSD.
Amaury Forgeot d'Arcefae8c42008-11-21 23:08:09 +0000319
Antoine Pitrou3ad3a0d2008-12-18 17:08:32 +0000320- Issue #4583: array.array would not always prohibit resizing when a buffer
321 has been exported, resulting in an interpreter crash when accessing the
322 buffer.
323
Martin v. Löwis841edad2008-11-30 10:36:49 +0000324
Brett Cannon7317c1e2008-11-25 19:19:17 +0000325Docs
326----
327
Georg Brandl52b92cc2008-05-25 09:43:03 +0000328
Skip Montanaroe5d7f7f2002-09-20 14:16:59 +0000329**(For information about older versions, consult the HISTORY file.)**