5 (editors: check NEWS.help for information about editing NEWS using ReST.)
7 What's New in Python 3.1 beta 1?
8 ================================
10 *Release date: XXXX-XX-XX*
19 What's New in Python 3.1 alpha 2?
20 =================================
22 *Release date: 2009-4-4*
27 - Implement PEP 378, Format Specifier for Thousands Separator, for
30 - Issue #5666: Py_BuildValue's 'c' code should create byte strings.
32 - Issue #5499: The 'c' code for argument parsing functions now only accepts a
33 byte, and the 'C' code only accepts a unicode character.
35 - Issue #1665206: Remove the last eager import in _warnings.c and make it lazy.
37 - Fix a segfault when running test_exceptions with coverage, caused by
38 insufficient checks in accessors of Exception.__context__.
40 - Issue #5604: non-ASCII characters in module name passed to
41 imp.find_module() were converted to UTF-8 while the path is
42 converted to the default filesystem encoding, causing nonsense.
44 - Issue #5126: str.isprintable() returned False for space characters.
46 - Issue #4865: On MacOSX /Library/Python/2.7/site-packages is added to
47 the end sys.path, for compatibility with the system install of Python.
49 - Issue #4688: Add a heuristic so that tuples and dicts containing only
50 untrackable objects are not tracked by the garbage collector. This can
51 reduce the size of collections and therefore the garbage collection overhead
52 on long-running programs, depending on their particular use of datatypes.
54 - Issue #5512: Rewrite PyLong long division algorithm (x_divrem) to
55 improve its performance. Long divisions and remainder operations
56 are now between 50% and 150% faster.
58 - Issue #4258: Make it possible to use base 2**30 instead of base
59 2**15 for the internal representation of integers, for performance
60 reasons. Base 2**30 is enabled by default on 64-bit machines. Add
61 --enable-big-digits option to configure, which overrides the
62 default. Add sys.int_info structseq to provide information about
65 - Issue #4474: PyUnicode_FromWideChar now converts characters outside
66 the BMP to surrogate pairs, on systems with sizeof(wchar_t) == 4
67 and sizeof(Py_UNICODE) == 2.
69 - Issue #5237: Allow auto-numbered fields in str.format(). For
70 example: '{} {}'.format(1, 2) == '1 2'.
72 - Issue #5392: when a very low recursion limit was set, the interpreter would
73 abort with a fatal error after the recursion limit was hit twice.
75 - Issue #3845: In PyRun_SimpleFileExFlags avoid invalid memory access with
78 - Py_DECREF: Add `do { ... } while (0)' to avoid compiler warnings.
83 - Issue 2625: added missing items() call to the for loop in
84 mailbox.MH.get_message().
86 - Issue #5640: Fix _multibytecodec so that CJK codecs don't repeat
87 error substitutions from non-strict codec error callbacks in
88 incrementalencoder and StreamWriter.
90 - Issue #5656: Fix the coverage reporting when running the test suite with
93 - Issue #5647: MutableSet.__iand__() no longer mutates self during iteration.
95 - Issue #5624: Fix the _winreg module name still used in several modules.
97 - Issue #5628: Fix io.TextIOWrapper.read() with a unreadable buffer.
99 - Issue #5619: Multiprocessing children disobey the debug flag and causes
100 popups on windows buildbots. Patch applied to work around this issue.
102 - Issue #5400: Added patch for multiprocessing on netbsd compilation/support
104 - Issue #5387: Fixed mmap.move crash by integer overflow.
106 - Issue #5261: Patch multiprocessing's semaphore.c to support context
107 manager use: "with multiprocessing.Lock()" works now.
109 - Issue #5236: Change time.strptime() to only take strings. Didn't work with
110 bytes already but the failure was non-obvious.
112 - Issue #5177: Multiprocessing's SocketListener class now uses
113 socket.SO_REUSEADDR on all connections so that the user no longer needs
114 to wait 120 seconds for the socket to expire.
116 - Issue #5595: Fix UnboundedLocalError in ntpath.ismount().
118 - Issue #1174606: Calling read() without arguments of an unbounded file
119 (typically /dev/zero under Unix) could crash the interpreter.
121 - The max_buffer_size arguments of io.BufferedWriter, io.BufferedRWPair, and
122 io.BufferedRandom have been deprecated for removal in Python 3.2.
124 - Issue #5068: Fixed the tarfile._BZ2Proxy.read() method that would loop
125 forever on incomplete input. That caused tarfile.open() to hang when used
126 with mode 'r' or 'r:bz2' and a fileobj argument that contained no data or
127 partial bzip2 compressed data.
129 - Issue #2110: Add support for thousands separator and 'n' type
130 specifier to Decimal.__format__
132 - Fix Decimal.__format__ bug that swapped the meanings of the '<' and
133 '>' alignment characters.
135 - The error detection code in FileIO.close() could fail to reflect the `errno`
136 value, and report it as -1 instead.
138 - Issue #5016: FileIO.seekable() could return False if the file position
139 was negative when truncated to a C int. Patch by Victor Stinner.
144 - Issue #5391: mmap now deals exclusively with bytes.
146 - Issue #5463: In struct module, remove deprecated overflow wrapping
147 when packing an integer: struct.pack('=L', -1) now raises
148 struct.error instead of returning b'\xff\xff\xff\xff'. The
149 _PY_STRUCT_RANGE_CHECKING and _PY_STRUCT_OVERFLOW_MASKING constants
150 have been removed from the struct module.
153 What's New in Python 3.1 alpha 1
154 ================================
156 *Release date: 2009-03-07*
161 - The io module has been reimplemented in C for speed.
163 - Give dict views an informative __repr__.
165 - Issue #5247: Improve error message when unknown format codes are
166 used when using str.format() with str, int, and float arguments.
168 - Issue #5249: time.strftime returned malformed string when format string
169 contained non ascii character on windows.
171 - Issue #4626: compile(), exec(), and eval() ignore the coding cookie if the
172 source has already been decoded into str.
174 - Issue #5186: Reduce hash collisions for objects with no __hash__ method by
175 rotating the object pointer by 4 bits to the right.
177 - Issue #4575: Fix Py_IS_INFINITY macro to work correctly on x87 FPUs:
178 it now forces its argument to double before testing for infinity.
180 - Issue #5137: Make len() correctly raise a TypeError when a __len__ method
181 returns a non-number type.
183 - Issue #5182: Removed memoryview.__str__.
185 - Issue #1717: Removed builtin cmp() function, dropped tp_compare
186 slot, the C API functions PyObject_Compare and PyUnicode_Compare and
187 the type definition cmpfunc. The tp_compare slot has been renamed
188 to tp_reserved, and is reserved for future usage.
190 - Issue 1242657: the __len__() and __length_hint__() calls in several tools
191 were suppressing all exceptions. These include list() and bytearray().
193 - Issue #4707: round(x, n) now returns an integer if x is an integer.
194 Previously it returned a float.
196 - Issue #4753: By enabling a configure option named '--with-computed-gotos'
197 on compilers that support it (notably: gcc, SunPro, icc), the bytecode
198 evaluation loop is compiled with a new dispatch mechanism which gives
199 speedups of up to 20%, depending on the system, on various benchmarks.
201 - Issue #4874: Most builtin decoders now reject unicode input.
203 - Issue #4842: Don't allow trailing 'L' when constructing an integer
206 - Issue #4991: os.fdopen now raises an OSError for invalid file descriptors.
208 - Issue #4838: When a module is deallocated, free the memory backing the
209 optional module state data.
211 - Issue #4910: Rename nb_long slot to nb_reserved, and change its
214 - Issue #4935: The overflow checking code in the expandtabs() method common
215 to str, bytes and bytearray could be optimized away by the compiler, letting
216 the interpreter segfault instead of raising an error.
218 - Issue #3720: Fix a crash when an iterator modifies its class and removes its
221 - Issue #4910: Builtin int() function and PyNumber_Long/PyNumber_Int API
222 function no longer attempt to call the __long__ slot to convert an object
223 to an integer. Only the __int__ and __trunc__ slots are examined.
225 - Issue #4893: Use NT threading on CE.
227 - Issue #4915: Port sysmodule to Windows CE.
229 - Issue #4868: utf-8, utf-16 and latin1 decoding are now 2x to 4x faster. The
230 common cases are optimized thanks to a dedicated fast path and a moderate
231 amount of loop unrolling.
233 - Issue #4074: Change the criteria for doing a full garbage collection (i.e.
234 collecting the oldest generation) so that allocating lots of objects without
235 destroying them does not show quadratic performance. Based on a proposal by
237 http://mail.python.org/pipermail/python-dev/2008-June/080579.html.
239 - Issue #4604: Some objects of the I/O library could still be used after
240 having been closed (for instance, a read() call could return some
241 previously buffered data). Patch by Dmitry Vasiliev.
243 - Issue #4705: Fix the -u ("unbuffered binary stdout and stderr") command-line
244 flag to work properly. Furthermore, when specifying -u, the text stdout
245 and stderr streams have line-by-line buffering enabled (the default being
246 to buffer arbitrary chunks of data).
248 - The internal table, _PyLong_DigitValue, is now an array of unsigned chars
249 instead of ints (reducing its size from 4 to 8 times thereby reducing
250 Python's overall memory).
252 - Issue #1180193: When importing a module from a .pyc (or .pyo) file with
253 an existing .py counterpart, override the co_filename attributes of all
254 code objects if the original filename is obsolete (which can happen if the
255 file has been renamed, moved, or if it is accessed through different paths).
256 Patch by Ziga Seilnacht and Jean-Paul Calderone.
258 - Issue #4580: Fix slicing of memoryviews when the item size is greater than
259 one byte. Also fixes the meaning of len() so that it returns the number of
260 items, rather than the size in bytes.
262 - Issue #4075: Use OutputDebugStringW in Py_FatalError.
264 - Issue #4747: When the terminal does not use utf-8, executing a script with
265 non-ascii characters in its name could fail with a "SyntaxError: None" error.
267 - Issue #4797: IOError.filename was not set when _fileio.FileIO failed to open
268 file with `bytes' filename on Windows.
270 - Issue #3680: Reference cycles created through a dict, set or deque iterator
271 did not get collected.
273 - Issue #4701: PyObject_Hash now implicitly calls PyType_Ready on types
274 where the tp_hash and tp_dict slots are both NULL.
276 - Issue #4759: None is now allowed as the first argument of
277 bytearray.translate(). It was always allowed for bytes.translate().
279 - Added test case to ensure attempts to read from a file opened for writing
282 - Issue #3106: Speedup some comparisons (str/str and int/int).
284 - Issue #2183: Simplify and optimize bytecode for list, dict and set
285 comprehensions. Original patch for list comprehensions by Neal Norwitz.
287 - Issue #2467: gc.DEBUG_STATS reported invalid elapsed times. Also, always
288 print elapsed times, not only when some objects are uncollectable /
289 unreachable. Original patch by Neil Schemenauer.
291 - Issue #3439: Add a bit_length method to int.
293 - Issue #2173: When getting device encoding, check that return value of
294 nl_langinfo is not the empty string. This was causing silent build
297 - Issue #4597: Fixed several opcodes that weren't always propagating
300 - Issue #4589: Fixed exception handling when the __exit__ function of a
301 context manager returns a value that cannot be converted to a bool.
303 - Issue #4445: Replace "sizeof(PyBytesObject)" with
304 "offsetof(PyBytesObject, ob_sval) + 1" when allocating memory for
305 bytes instances. On a typical machine this saves 3 bytes of memory
306 (on average) per allocation of a bytes instance.
308 - Issue #4533: File read operation was dreadfully slow due to a slowly
309 growing read buffer. Fixed by using the same growth rate algorithm as
312 - Issue #4509: Various issues surrounding resize of bytearray objects to
313 which there are buffer exports (e.g. memoryview instances).
315 - Issue #4233: Changed semantic of ``_fileio.FileIO``'s ``close()``
316 method on file objects with closefd=False. The file descriptor is still
317 kept open but the file object behaves like a closed file. The ``FileIO``
318 object also got a new readonly attribute ``closefd``.
320 - Issue #4569: Interpreter crash when mutating a memoryview with an item size
323 - Issue #4748: Lambda generators no longer return a value.
325 - The re.sub(), re.subn() and re.split() functions now accept a flags parameter.
330 - Issue #1665206 (partially): Move imports in cgitb to the top of the module
331 instead of performing them in functions. Helps prevent import deadlocking in
334 - Issue #2522: locale.format now checks its first argument to ensure it has
335 been passed only one pattern, avoiding mysterious errors where it appeared
336 that it was failing to do localization.
338 - Issue #5583: Added optional Extensions in Distutils. Initial patch by Georg
341 - Issue #1222: locale.format() bug when the thousands separator is a space
344 - Issue #5472: Fixed distutils.test_util tear down. Original patch by
347 - collections.deque() objects now have a read-only attribute called maxlen.
349 - Issue #2638: Show a window constructed with tkSimpleDialog.Dialog only after
350 it is has been populated and properly configured in order to prevent
353 - Issue #4792: Prevent a segfault in _tkinter by using the
354 guaranteed to be safe interp argument given to the PythonCmd in place of
355 the Tcl interpreter taken from a PythonCmd_ClientData.
357 - Issue #5193: Guarantee that tkinter.Text.search returns a string.
359 - Issue #5394: removed > 2.3 syntax from distutils.msvc9compiler.
360 Original patch by Akira Kitada.
362 - Issue #5334: array.fromfile() failed to insert values when EOFError was raised.
364 - Issue #5385: Fixed mmap crash after resize failure on windows.
366 - Issue #5179: Fixed subprocess handle leak on failure on windows.
368 - PEP 372: Added collections.OrderedDict().
370 - The _asdict() for method for namedtuples now returns an OrderedDict().
372 - configparser now defaults to using an ordered dictionary.
374 - Issue #5401: Fixed a performance problem in mimetypes when ``from mimetypes
375 import guess_extension`` was used.
377 - Issue #1733986: Fixed mmap crash in accessing elements of second map object
378 with same tagname but larger size than first map. (Windows)
380 - Issue #5386: mmap.write_byte didn't check map size, so it could cause buffer
383 - Issue #1533164: Installed but not listed *.pyo was breaking Distutils
386 - Issue #5378: added --quiet option to Distutils bdist_rpm command.
388 - Issue #5052: make Distutils compatible with 2.3 again.
390 - Issue #5316: Fixed buildbot failures introduced by multiple inheritance
393 - Issue #5287: Add exception handling around findCaller() call to help out
396 - Issue #5282: Fixed mmap resize on 32bit windows and unix. When offset > 0,
397 The file was resized to wrong size.
399 - Issue #5292: Fixed mmap crash on its boundary access m[len(m)].
401 - Issue #2279: distutils.sdist.add_defaults now add files
402 from the package_data and the data_files metadata.
404 - Issue #5257: refactored all tests in distutils, so they use
405 support.TempdirManager, to avoid writing in the tests directory.
407 - Issue #4524: distutils build_script command failed with --with-suffix=3.
408 Initial patch by Amaury Forgeot d'Arc.
410 - Issue #2461: added tests for distutils.util
412 - Issue #4998: The memory saving effect of __slots__ had been lost on Fractions
413 which inherited from numbers.py which did not have __slots__ defined. The
414 numbers hierarchy now has its own __slots__ declarations.
416 - Issue #4631: Fix urlopen() result when an HTTP response uses chunked
419 - Issue #5203: Fixed ctypes segfaults when passing a unicode string to a
420 function without argtypes (only occurs if HAVE_USABLE_WCHAR_T is false).
422 - Issue #3386: distutils.sysconfig.get_python_lib prefix argument was ignored
423 under NT and OS2. Patch by Philip Jenvey.
425 - Issue #5128: Make compileall properly inspect bytecode to determine if needs
426 to be recreated. This avoids a timing hole thanks to the old reliance on the
427 ctime of the files involved.
429 - Issue #5122: Synchronize tk load failure check to prevent a potential
432 - Issue #1818: collections.namedtuple() now supports a keyword argument
433 'rename' which lets invalid fieldnames be automatically converted to
434 positional names in the form, _1, _2, ...
436 - Issue #4890: Handle empty text search pattern in Tkinter.Text.search.
438 - Issue #4512 (part 2): Promote ``ZipImporter._get_filename()`` to be a
439 public documented method ``ZipImporter.get_filename()``.
441 - Issue #4195: The ``runpy`` module (and the ``-m`` switch) now support
442 the execution of packages by looking for and executing a ``__main__``
443 submodule when a package name is supplied. Initial patch by Andi
446 - Issue #1731706: Call Tcl_ConditionFinalize for Tcl_Conditions that will
447 not be used again (this requires Tcl/Tk 8.3.1), also fix a memory leak in
448 Tkapp_Call when calling from a thread different than the one that created
449 the Tcl interpreter. Patch by Robert Hancock.
451 - Issue #4285: Change sys.version_info to be a named tuple. Patch by
454 - Issue #1520877: Now distutils.sysconfig reads $AR from the
455 environment/Makefile. Patch by Douglas Greiman.
457 - Issue #1276768: The verbose option was not used in the code of
458 distutils.file_util and distutils.dir_util.
460 - Issue #5132: Fixed trouble building extensions under Solaris with
461 --enabled-shared activated. Initial patch by Dave Peterson.
463 - Issue #1581476: Always use the Tcl global namespace when calling into Tcl.
465 - The shelve module now defaults to pickle protocol 3.
467 - Fix a bug in the trace module where a bytes object from co_lnotab had its
468 items being passed through ord().
470 - Issue #2047: shutil.move() could believe that its destination path was
471 inside its source path if it began with the same letters (e.g. "src" vs.
474 - Added the ttk module. See issue #2983: Ttk support for Tkinter.
476 - Removed isSequenceType(), isMappingType, and isNumberType() from the
477 operator module; use the abstract base classes instead. Also removed
478 the repeat() function; use mul() instead.
480 - Issue 5021: doctest.testfile() did not create __name__ and
481 collections.namedtuple() relied on __name__ being defined.
483 - Backport importlib from Python 3.1. Only the import_module() function has
484 been backported to help facilitate transitions from 2.7 to 3.1.
486 - Issue #1885: distutils. When running sdist with --formats=tar,gztar
487 the tar file was overriden by the gztar one.
489 - Issue #4863: distutils.mwerkscompiler has been removed.
491 - Added a new itertools functions: combinations_with_replacement()
494 - Issue 5032: added a step argument to itertools.count() and
495 allowed non-integer arguments.
497 - Fix and properly document the multiprocessing module's logging
498 support, expose the internal levels and provide proper usage
501 - Issue #1672332: fix unpickling of subnormal floats, which was
502 producing a ValueError on some platforms.
504 - Issue #3881: Help Tcl to load even when started through the
505 unreadable local symlink to "Program Files" on Vista.
507 - Issue #4710: Extract directories properly in the zipfile module;
508 allow adding directories to a zipfile.
510 - Issue #3807: _multiprocessing build fails when configure is passed
511 --without-threads argument. When this occurs, _multiprocessing will
512 be disabled, and not compiled.
514 - Issue #5008: When a file is opened in append mode with the new IO library,
515 do an explicit seek to the end of file (so that e.g. tell() returns the
516 file size rather than 0). This is consistent with the behaviour of the
517 traditional 2.x file object.
519 - Issue #5013: Fixed a bug in FileHandler which occurred when the delay
522 - Issue #4842: Always append a trailing 'L' when pickling longs using
523 pickle protocol 0. When reading, the 'L' is optional.
525 - Add the importlib package.
527 - Issue #4301: Patch the logging module to add processName support, remove
528 _check_logger_class from multiprocessing.
530 - Issue #3325: Remove python2.x try: except: imports for old cPickle from
533 - Issue #4959: inspect.formatargspec now works for keyword only arguments
536 - Issue #3321: _multiprocessing.Connection() doesn't check handle; added checks
537 for *nix machines for negative handles and large int handles. Without this check
538 it is possible to segfault the interpreter.
540 - Issue #4449: AssertionError in mp_benchmarks.py, caused by an underlying issue
543 - Issue #1225107: inspect.isclass() returned True for instances with a custom
546 - Issue #3826 and #4791: The socket module now closes the underlying socket
547 appropriately when it is being used via socket.makefile() objects
548 rather than delaying the close by waiting for garbage collection to do it.
550 - Issue #1696199: Add collections.Counter() for rapid and convenient
553 - Issue #3860: GzipFile and BZ2File now support the context manager protocol.
555 - Issue #4867: Fixed a crash in ctypes when passing a string to a
556 function without defining argtypes.
558 - Issue #4272: Add an optional argument to the GzipFile constructor to override
559 the timestamp in the gzip stream. The default value remains the current time.
560 The information can be used by e.g. gunzip when decompressing. Patch by
563 - Restore Python 2.3 compatibility for decimal.py.
565 - Issue #3638: Remove functions from _tkinter module level that depend on
566 TkappObject to work with multiple threads.
568 - Issue #4718: Adapt the wsgiref package so that it actually works with
569 Python 3.x, in accordance with the `official amendments of the spec
570 <http://www.wsgi.org/wsgi/Amendments_1.0>`_.
572 - Issue #4796: Added Decimal.from_float() and Context.create_decimal_from_float()
573 to the decimal module.
575 - Fractions.from_float() no longer loses precision for integers too big to
578 - Issue #4812: add missing underscore prefix to some internal-use-only
579 constants in the decimal module. (Dec_0 becomes _Dec_0, etc.)
581 - Issue 4790: The nsmallest() and nlargest() functions in the heapq module
582 did unnecessary work in the common case where no key function was specified.
584 - Issue #4795: inspect.isgeneratorfunction() returns False instead of None when
585 the function is not a generator.
587 - Issue #4702: Throwing a DistutilsPlatformError instead of IOError in case
588 no MSVC compiler is found under Windows. Original patch by Philip Jenvey.
590 - Issue #4646: distutils was choking on empty options arg in the setup
591 function. Original patch by Thomas Heller.
593 - Issue #3767: Convert Tk object to string in tkColorChooser.
595 - Issue #3248: Allow placing ScrolledText in a PanedWindow.
597 - Issue #4444: Allow assertRaises() to be used as a context handler, so that
598 the code under test can be written inline if more practical.
600 - Issue #4739: Add pydoc help topics for symbols, so that e.g. help('@')
601 works as expected in the interactive environment.
603 - Issue #4756: zipfile.is_zipfile() now supports file-like objects. Patch by
606 - Issue #4574: reading an UTF16-encoded text file crashes if \r on 64-char
609 - Issue #4223: inspect.getsource() will now correctly display source code
610 for packages loaded via zipimport (or any other conformant PEP 302
611 loader). Original patch by Alexander Belopolsky.
613 - Issue #4201: pdb can now access and display source code loaded via
614 zipimport (or any other conformant PEP 302 loader). Original patch by
615 Alexander Belopolsky.
617 - Issue #4197: doctests in modules loaded via zipimport (or any other PEP
618 302 conformant loader) will now work correctly in most cases (they
619 are still subject to the constraints that exist for all code running
620 from inside a module loaded via a PEP 302 loader and attempting to
621 perform IO operations based on __file__). Original patch by
622 Alexander Belopolsky.
624 - Issues #4082 and #4512: Add runpy support to zipimport in a manner that
625 allows backporting to maintenance branches. Original patch by
626 Alexander Belopolsky.
628 - Issue #4163: textwrap module: allow word splitting on a hyphen preceded by
631 - Issue #4616: TarFile.utime(): Restore directory times on Windows.
633 - Issue #4021: tokenize.detect_encoding() now raises a SyntaxError when the
634 codec cannot be found. This is for compatibility with the builtin behavior.
636 - Issue #4084: Fix max, min, max_mag and min_mag Decimal methods to
637 give correct results in the case where one argument is a quiet NaN
638 and the other is a finite number that requires rounding.
640 - Issue #4483: _dbm module now builds on systems with gdbm & gdbm_compat
643 - Added the subprocess.check_call_output() convenience function to get output
644 from a subprocess on success or raise an exception on error.
646 - Issue #1055234: cgi.parse_header(): Fixed parsing of header parameters to
647 support unusual filenames (such as those containing semi-colons) in
648 Content-Disposition headers.
650 - Issue #4384: Added logging integration with warnings module using
651 captureWarnings(). This change includes a NullHandler which does nothing;
652 it will be of use to library developers who want to avoid the "No handlers
653 could be found for logger XXX" message which can appear if the library user
654 doesn't configure logging.
656 - Issue #3741: DISTUTILS_USE_SDK set causes msvc9compiler.py to raise an
659 - Issue #4529: fix the parser module's validation of try-except-finally
662 - Issue #4458: getopt.gnu_getopt() now recognizes a single "-" as an argument,
663 not a malformed option.
665 - Added the subprocess.check_output() convenience function to get output
666 from a subprocess on success or raise an exception on error.
668 - Issue #4542: On Windows, binascii.crc32 still accepted str as binary input;
669 the corresponding tests now pass.
671 - Issue #4537: webbrowser.UnixBrowser would fail to open the browser because
672 it was calling the wrong open() function.
674 - Issue #1055234: cgi.parse_header(): Fixed parsing of header parameters to
675 support unusual filenames (such as those containing semi-colons) in
676 Content-Disposition headers.
678 - Issue #4861: ctypes.util.find_library(): Robustify. Fix library detection on
679 biarch systems. Try to rely on ldconfig only, without using objdump and gcc.
681 - Issue #5104: The socket module now raises OverflowError when 16-bit port and
682 protocol numbers are supplied outside the allowed 0-65536 range on bind()
688 - Issue #4704: remove use of cmp() in pybench, bump its version number to 2.1,
689 and make it 2.6-compatible.
691 - Ttk demos added in Demo/tkinter/ttk/
693 - Issue #4677: add two list comprehension tests to pybench.
699 - Link the shared python library with $(MODLIBS).
701 - Issue #5134: Silence compiler warnings when compiling sqlite with VC++.
703 - Issue #4494: Fix build with Py_NO_ENABLE_SHARED on Windows.
705 - Issue #4895: Use _strdup on Windows CE.
707 - Issue #4472: "configure --enable-shared" now works on OSX
709 - Issues #4728 and #4060: WORDS_BIGEDIAN is now correct in Universal builds.
711 - Issue #4389: Add icon to the uninstall entry in "add-and-remove-programs".
713 - Issue #4289: Remove Cancel button from AdvancedDlg.
715 - Issue #1656675: Register a drop handler for .py* files on Windows.
717 - Issue #4120: Exclude manifest from extension modules in VS2008.
719 - Issue #4091: Install pythonxy.dll in system32 again.
721 - Issue #4018: Disable "for me" installations on Vista.
723 - Issue #3758: Add ``patchcheck`` build target to .PHONY.
725 - Issue #4204: Fixed module build errors on FreeBSD 4.
731 - Issue #5175: PyLong_AsUnsignedLongLong now raises OverflowError
732 for negative arguments. Previously, it raised TypeError.
734 - Issue #4720: The format for PyArg_ParseTupleAndKeywords can begin with '|'.
736 - Issue #3632: from the gdb debugger, the 'pyo' macro can now be called when
737 the GIL is released, or owned by another thread.
739 - Issue #4122: On Windows, fix a compilation error when using the
740 Py_UNICODE_ISSPACE macro in an extension module.
746 - Issue #3745: Fix hashlib to always reject unicode and non buffer-api
747 supporting objects as input no matter how it was compiled (built in
748 implementations or external openssl library).
750 - Issue #4397: Fix occasional test_socket failure on OS X.
752 - Issue #4279: Fix build of parsermodule under Cygwin.
754 - Issue #4751: hashlib now releases the GIL when hashing large buffers
755 (with a hardwired threshold of 2048 bytes), allowing better parallelization
756 on multi-CPU systems. Contributed by Lukas Lueg (ebfe) and Victor Stinner.
758 - Issue #4051: Prevent conflict of UNICODE macros in cPickle.
760 - Issue #4738: Each zlib object now has a separate lock, allowing to compress
761 or decompress several streams at once on multi-CPU systems. Also, the GIL
762 is now released when computing the CRC of a large buffer. Patch by ebfe.
764 - Issue #4228: Pack negative values the same way as 2.4 in struct's L format.
766 - Issue #1040026: Fix os.times result on systems where HZ is incorrect.
768 - Issues #3167, #3682: Fix test_math failures for log, log10 on Solaris,
771 - Issue #4583: array.array would not always prohibit resizing when a buffer
772 has been exported, resulting in an interpreter crash when accessing the
776 - Issue #5228: Make functools.partial objects can now be pickled.
781 - regrtest no longer treats ImportError as equivalent to SkipTest. Imports
782 that should cause a test to be skipped are now done using import_module
783 from test support, which does the conversion.
785 - Issue #5083: New 'gui' resource for regrtest.
792 **(For information about older versions, consult the HISTORY file.)**