+++++++++++ Python News +++++++++++ What's New in Python 3.4.0 Alpha 1? =================================== *Relase date: 2014-XX-XX* Core and Builtins ----------------- - Issue #17323: The "[X refs, Y blocks]" printed by debug builds has been disabled by default. It can be re-enabled with the `-X showrefcount` option. - Issue #17522: Add the PyGILState_Check() API. - Issue #16475: Support object instancing, recursion and interned strings in marshal - Use the HTTPS PyPI url for upload, overriding any plain HTTP URL in pypirc. - Issue #16795: On the ast.arguments object, unify vararg with varargannotation and kwarg and kwargannotation. Change the column offset of ast.Attribute to be at the attribute name. - Issue #17434: Properly raise a SyntaxError when a string occurs between future imports. - Issue #17117: Import and @importlib.util.set_loader now set __loader__ when it has a value of None or the attribute doesn't exist. - Issue #17327: Add PyDict_SetDefault. - Issue #17032: The "global" in the "NameError: global name 'x' is not defined" error message has been removed. Patch by Ram Rachum. - Issue #17328: Fix possible refleak in dict.setdefault. - Issue #17223: array module: Fix a crasher when converting an array containing invalid characters (outside range [U+0000; U+10ffff]) to Unicode: repr(array), str(array) and array.tounicode(). Patch written by Manuel Jacob. - Issue #17223: Fix PyUnicode_FromUnicode() for string of 1 character outside the range U+0000-U+10ffff. - Issue #17275: Corrected class name in init error messages of the C version of BufferedWriter and BufferedRandom. - Issue #7963: Fixed misleading error message that issued when object is called without arguments. - Issue #8745: Small speed up zipimport on Windows. Patch by Catalin Iacob. - Issue #5308: Raise ValueError when marshalling too large object (a sequence with size >= 2**31), instead of producing illegal marshal data. - Issue #12983: Bytes literals with invalid \x escape now raise a SyntaxError and a full traceback including line number. - Issue #16967: In function definition, evaluate positional defaults before keyword-only defaults. - Issue #17173: Remove uses of locale-dependent C functions (isalpha() etc.) in the interpreter. - Issue #17137: When an Unicode string is resized, the internal wide character string (wstr) format is now cleared. - Issue #17043: The unicode-internal decoder no longer read past the end of input buffer. - Issue #17098: All modules now have __loader__ set even if they pre-exist the bootstrapping of importlib. - Issue #16979: Fix error handling bugs in the unicode-escape-decode decoder. - Issue #16772: The base argument to the int constructor no longer accepts floats, or other non-integer objects with an __int__ method. Objects with an __index__ method are now accepted. - Issue #10156: In the interpreter's initialization phase, unicode globals are now initialized dynamically as needed. - Issue #16980: Fix processing of escaped non-ascii bytes in the unicode-escape-decode decoder. - Issue #16975: Fix error handling bug in the escape-decode bytes decoder. - Issue #14850: Now a charmap decoder treats U+FFFE as "undefined mapping" in any mapping, not only in a string. - Issue #16730: importlib.machinery.FileFinder now no longers raises an exception when trying to populate its cache and it finds out the directory is unreadable or has turned into a file. Reported and diagnosed by David Pritchard. - Issue #16906: Fix a logic error that prevented most static strings from being cleared. - Issue #11461: Fix the incremental UTF-16 decoder. Original patch by Amaury Forgeot d'Arc. - Issue #16881: Fix Py_ARRAY_LENGTH macro for GCC < 3.1. - Issue #16856: Fix a segmentation fault from calling repr() on a dict with a key whose repr raise an exception. - Issue #16367: Fix FileIO.readall() on Windows for files larger than 2 GB. - Issue #16761: Calling int() with base argument only now raises TypeError. - Issue #16759: Support the full DWORD (unsigned long) range in Reg2Py when retrieving a REG_DWORD value. This corrects functions like winreg.QueryValueEx that may have been returning truncated values. - Issue #14420: Support the full DWORD (unsigned long) range in Py2Reg when passed a REG_DWORD value. Fixes OverflowError in winreg.SetValueEx. - Issue #11939: Set the st_dev attribute of stat_result to allow Windows to take advantage of the os.path.samefile/sameopenfile/samestat implementations used by other platforms. - Issue #16772: The int() constructor's second argument (base) no longer accepts non integer values. Consistent with the behavior in Python 2. - Issue #15422: Get rid of PyCFunction_New macro. Use PyCFunction_NewEx function (PyCFunction_New func is still present for backward compatibility). - Issue #14470: Remove w9xpopen support per PEP 11. - Issue #9856: Replace deprecation warning with raising TypeError in object.__format__. Patch by Florent Xicluna. - Issue #16597: In buffered and text IO, call close() on the underlying stream if invoking flush() fails. - Issue #16722: In the bytes() constructor, try to call __bytes__ on the argument before __index__. - Issue #16421: loading multiple modules from one shared object is now handled correctly (previously, the first module loaded from that file was silently returned). Patch by Václav Šmilauer. - Issue #16602: When a weakref's target was part of a long deallocation chain, the object could remain reachable through its weakref even though its refcount had dropped to zero. - Issue #16495: Remove extraneous NULL encoding check from bytes_decode(). - Issue #16619: Create NameConstant AST class to represent None, True, and False literals. As a result, these constants are never loaded at runtime from builtins. - Issue #16455: On FreeBSD and Solaris, if the locale is C, the ASCII/surrogateescape codec is now used, instead of the locale encoding, to decode the command line arguments. This change fixes inconsistencies with os.fsencode() and os.fsdecode() because these operating systems announces an ASCII locale encoding, whereas the ISO-8859-1 encoding is used in practice. - Issue #16562: Optimize dict equality testing. Patch by Serhiy Storchaka. - Issue #16588: Silence unused-but-set warnings in Python/thread_pthread - Issue #16592: stringlib_bytes_join doesn't raise MemoryError on allocation failure. - Issue #16546: Fix: ast.YieldFrom argument is now mandatory. - Issue #16514: Fix regression causing a traceback when sys.path[0] is None (actually, any non-string or non-bytes type). - Issue #16505: Remove unused Py_TPFLAGS_INT_SUBCLASS. - Issue #16306: Fix multiple error messages when unknown command line parameters where passed to the interpreter. Patch by Hieu Nguyen. - Issue #16215: Fix potential double memory free in str.replace(). Patch by Serhiy Storchaka. - Issue #16290: A float return value from the __complex__ special method is no longer accepted in the complex() constructor. - Issue #16416: On Mac OS X, operating system data are now always encoded/decoded to/from UTF-8/surrogateescape, instead of the locale encoding (which may be ASCII if no locale environment variable is set), to avoid inconsistencies with os.fsencode() and os.fsdecode() functions which are already using UTF-8/surrogateescape. - Issue #16453: Fix equality testing of dead weakref objects. - Issue #9535: Fix pending signals that have been received but not yet handled by Python to not persist after os.fork() in the child process. - Issue #14794: Fix slice.indices to return correct results for huge values, rather than raising OverflowError. - Issue #15001: fix segfault on "del sys.modules['__main__']". Patch by Victor Stinner. - Issue #8271: the utf-8 decoder now outputs the correct number of U+FFFD characters when used with the 'replace' error handler on invalid utf-8 sequences. Patch by Serhiy Storchaka, tests by Ezio Melotti. - Issue #5765: Apply a hard recursion limit in the compiler instead of blowing the stack and segfaulting. Initial patch by Andrea Griffini. - Issue #16402: When slicing a range, fix shadowing of exceptions from __index__. - Issue #16336: fix input checking in the surrogatepass error handler. Patch by Serhiy Storchaka. - Issue #8401: assigning an int to a bytearray slice (e.g. b[3:4] = 5) now raises an error. - Issue #7317: Display full tracebacks when an error occurs asynchronously. Patch by Alon Horev with update by Alexey Kachayev. - Issue #16309: Make PYTHONPATH="" behavior the same as if PYTHONPATH not set at all. - Issue #10189: Improve the error reporting of SyntaxErrors related to global and nonlocal statements. - Issue #16086: PyTypeObject.tp_flags and PyType_Spec.flags are now unsigned (unsigned long and unsigned int) to avoid an undefined behaviour with Py_TPFLAGS_TYPE_SUBCLASS ((1 << 31). PyType_GetFlags() result type is now unsigned too (unsigned long, instead of long). - Fix segfaults on setting __qualname__ on builtin types and attempting to delete it on any type. - Issue #14625: Rewrite the UTF-32 decoder. It is now 3x to 4x faster. Patch written by Serhiy Storchaka. - Issue #16345: Fix an infinite loop when ``fromkeys`` on a dict subclass received a nonempty dict from the constructor. - Issue #16271: Fix strange bugs that resulted from __qualname__ appearing in a class's __dict__ and on type. - Issue #16197: Update winreg docstrings and documentation to match code. Patch by Zachary Ware. - Issue #16241: Document -X faulthandler command line option. Patch by Marek Šuppa. - Issue #12805: Make bytes.join and bytearray.join faster when the separator is empty. Patch by Serhiy Storchaka. - Issue #6074: Ensure cached bytecode files can always be updated by the user that created them, even when the source file is read-only. - Issue #16166: Add PY_LITTLE_ENDIAN and PY_BIG_ENDIAN macros and unified endianness detection and handling. - Issue #15958: bytes.join and bytearray.join now accept arbitrary buffer objects. - Issue #14783: Improve int() docstring and switch docstrings for str(), range(), and slice() to use multi-line signatures. - Issue #16160: Subclass support now works for types.SimpleNamespace. - Upgrade Unicode data (UCD) to version 6.2. - Issue #15379: Fix passing of non-BMP characters as integers for the charmap decoder (already working as unicode strings). Patch by Serhiy Storchaka. - Issue #15144: Fix possible integer overflow when handling pointers as integer values, by using `Py_uintptr_t` instead of `size_t`. Patch by Serhiy Storchaka. - Issue #15965: Explicitly cast `AT_FDCWD` as (int). Required on Solaris 10 (which defines `AT_FDCWD` as ``0xffd19553``), harmless on other platforms. - Issue #15839: Convert SystemErrors in `super()` to RuntimeErrors. - Issue #15448: Buffered IO now frees the buffer when closed, instead of when deallocating. - Issue #15846: Fix SystemError which happened when using `ast.parse()` in an exception handler on code with syntax errors. - Issue #15897: zipimport.c doesn't check return value of fseek(). Patch by Felipe Cruz. - Issue #15801: Make sure mappings passed to '%' formatting are actually subscriptable. - Issue #15111: __import__ should propagate ImportError when raised as a side-effect of a module triggered from using fromlist. - Issue #15022: Add pickle and comparison support to types.SimpleNamespace. Library ------- - Issue #6649: Fixed missing exit status in IDLE. Patch by Guilherme Polo. - Issue #17435: threading.Timer's __init__ method no longer uses mutable default values for the args and kwargs parameters. - Issue #17526: fix an IndexError raised while passing code without filename to inspect.findsource(). Initial patch by Tyler Doyle. - Issue #17540: Added style to formatter configuration by dict. - Issue #16692: The ssl module now supports TLS 1.1 and TLS 1.2. Initial patch by Michele Orrù. - Issue #17025: multiprocessing: Reduce Queue and SimpleQueue contention. - Issue #17536: Add to webbrowser's browser list: www-browser, x-www-browser, iceweasel, iceape. - Issue #17150: pprint now uses line continuations to wrap long string literals. - Issue #17488: Change the subprocess.Popen bufsize parameter default value from unbuffered (0) to buffering (-1) to match the behavior existing code expects and match the behavior of the subprocess module in Python 2 to avoid introducing hard to track down bugs. - Issue #17521: Corrected non-enabling of logger following two calls to fileConfig(). - Issue #17508: Corrected MemoryHandler configuration in dictConfig() where the target handler wasn't configured first. - Issue #17209: curses.window.get_wch() now handles correctly KeyboardInterrupt (CTRL+c). - Issue #5713: smtplib now handles 421 (closing connection) error codes when sending mail by closing the socket and reporting the 421 error code via the exception appropriate to the command that received the error response. - Issue #16997: unittest.TestCase now provides a subTest() context manager to procedurally generate, in an easy way, small test instances. - Issue #17485: Also delete the Request Content-Length header if the data attribute is deleted. (Follow on to issue 16464). - Issue #15927: CVS now correctly parses escaped newlines and carriage when parsing with quoting turned off. - Issue #17467: add readline and readlines support to mock_open in unittest.mock. - Issue #17192: Update the ctypes module's libffi to v3.0.13. This specifically addresses a stack misalignment issue on x86 and issues on some more recent platforms. - Issue #8862: Fixed curses cleanup when getkey is interrputed by a signal. - Issue #17443: imaplib.IMAP4_stream was using the default unbuffered IO in subprocess, but the imap code assumes buffered IO. In Python2 this worked by accident. IMAP4_stream now explicitly uses buffered IO. - Issue #17476: Fixed regression relative to Python2 in undocumented pydoc 'allmethods'; it was missing unbound methods on the class. - Issue #17474: Remove the deprecated methods of Request class. - Issue #16709: unittest discover order is no-longer filesystem specific. Patch by Jeff Ramnani. - Issue #5024: sndhdr.whichhdr now returns the frame count for WAV files rather than -1. - Issue #17460: Remove the strict argument of HTTPConnection and removing the DeprecationWarning being issued from 3.2 onwards. - Issue #16880: Do not assume _imp.load_dynamic() is defined in the imp module. - Issue #16389: Fixed a performance regression relative to Python 3.1 in the caching of compiled regular expressions. - Added missing FeedParser and BytesFeedParser to email.parser.__all__. - Issue #17431: Fix missing import of BytesFeedParser in email.parser. - Issue #12921: http.server's send_error takes an explain argument to send more information in response. Patch contributed by Karl. - Issue #17414: Add timeit, repeat, and default_timer to timeit.__all__. - Issue #1285086: Get rid of the refcounting hack and speed up urllib.parse.unquote() and urllib.parse.unquote_to_bytes(). - Issue #17099: Have importlib.find_loader() raise ValueError when __loader__ is not set, harmonizing with what happens when the attribute is set to None. - Expose the O_PATH constant in the os module if it is available. - Issue #17368: Fix an off-by-one error in the Python JSON decoder that caused a failure while decoding empty object literals when object_pairs_hook was specified. _ Issue #17385: Fix quadratic behavior in threading.Condition. The FIFO queue now uses a deque instead of a list. - Issue #15806: Add contextlib.ignored(). This creates a context manager to ignore specified exceptions, replacing the "except Exc: pass" idiom. - Issue #14645: The email generator classes now produce output using the specified linesep throughout. Previously if the prolog, epilog, or body were stored with a different linesep, that linesep was used. This fix corrects an RFC non-compliance issue with smtplib.send_message. - Issue #17278: Fix a crash in heapq.heappush() and heapq.heappop() when the list is being resized concurrently. - Issue #16962: Use getdents64 instead of the obsolete getdents syscall in the subprocess module on Linux. - Issue #16935: unittest now counts the module as skipped if it raises SkipTest, instead of counting it as an error. Patch by Zachary Ware. - Issue #17018: Make Process.join() retry if os.waitpid() fails with EINTR. - Issue #17197: profile/cProfile modules refactored so that code of run() and runctx() utility functions is not duplicated in both modules. - Issue #14720: sqlite3: Convert datetime microseconds correctly. Patch by Lowe Thiderman. - Issue #15132: Allow a list for the defaultTest argument of unittest.TestProgram. Patch by Jyrki Pulliainen. - Issue #17225: JSON decoder now counts columns in the first line starting with 1, as in other lines. - Issue #6623: Added explicit DeprecationWarning for ftplib.netrc, which has been deprecated and undocumented for a long time. - Issue #13700: Fix byte/string handling in imaplib authentication when an authobject is specified. - Issue #13153: Tkinter functions now raise TclError instead of ValueError when a string argument contains non-BMP character. - Issue #9669: Protect re against infinite loops on zero-width matching in non-greedy repeat. Patch by Matthew Barnett. - Issue #13169: The maximal repetition number in a regular expression has been increased from 65534 to 2147483647 (on 32-bit platform) or 4294967294 (on 64-bit). - Issue #17143: Fix a missing import in the trace module. Initial patch by Berker Peksag. - Issue #15220: email.feedparser's line splitting algorithm is now simpler and faster. - Issue #16743: Fix mmap overflow check on 32 bit Windows. - Issue #16996: webbrowser module now uses shutil.which() to find a web-browser on the executable search path. - Issue #16800: tempfile.gettempdir() no longer left temporary files when the disk is full. Original patch by Amir Szekely. - Issue #16564: Fixed regression relative to Python2 in the operation of email.encoders.encode_7or8bit when used with binary data. - Issue #17052: unittest discovery should use self.testLoader. - Issue #4591: Uid and gid values larger than 2**31 are supported now. - Issue #17141: random.vonmisesvariate() no more hangs for large kappas. - Issue #17149: Fix random.vonmisesvariate to always return results in [0, 2*math.pi]. - Issue #1470548: XMLGenerator now works with binary output streams. - Issue #6975: os.path.realpath() now correctly resolves multiple nested symlinks on POSIX platforms. - Issue #13773: sqlite3.connect() gets a new `uri` parameter to pass the filename as a URI, allowing to pass custom options. - Issue #17156: pygettext.py now uses an encoding of source file and correctly writes and escapes non-ascii characters. - Issue #16564: Fixed regression relative to Python2 in the operation of email.encoders.encode_noop when used with binary data. - Issue #10355: In SpooledTemporaryFile class mode, name, encoding and newlines properties now work for unrolled files. Obsoleted and never working on Python 3 xreadline method now removed. - Issue #16686: Fixed a lot of bugs in audioop module. Fixed crashes in avgpp(), maxpp() and ratecv(). Fixed an integer overflow in add(), bias(), and ratecv(). reverse(), lin2lin() and ratecv() no more lose precision for 32-bit samples. max() and rms() no more returns a negative result and various other functions now work correctly with 32-bit sample -0x80000000. - Issue #17073: Fix some integer overflows in sqlite3 module. - Issue #17114: IDLE now uses non-strict config parser. - Issue #16723: httplib.HTTPResponse no longer marked closed when the connection is automatically closed. - Issue #15359: Add CAN_BCM protocol support to the socket module. Patch by Brian Thorne. - Issue #16948: Fix quoted printable body encoding for non-latin1 character sets in the email package. - Issue #16811: Fix folding of headers with no value in the provisional email policies. - Issue #17132: Update symbol for "yield from" grammar changes. - Issue #17076: Make copying of xattrs more permissive of missing FS support. Patch by Thomas Wouters. - Issue #17089: Expat parser now correctly works with string input not only when an internal XML encoding is UTF-8 or US-ASCII. It now accepts bytes and strings larger than 2 GiB. - Issue #6083: Fix multiple segmentation faults occured when PyArg_ParseTuple parses nested mutating sequence. - Issue #5289: Fix ctypes.util.find_library on Solaris. - Issue #17106: Fix a segmentation fault in io.TextIOWrapper when an underlying stream or a decoder produces data of an unexpected type (i.e. when io.TextIOWrapper initialized with text stream or use bytes-to-bytes codec). - Issue #17015: When it has a spec, a Mock object now inspects its signature when matching calls, so that arguments can be matched positionally or by name. - Issue #15633: httplib.HTTPResponse is now mark closed when the server sends less than the advertised Content-Length. - Issue #6972: The zipfile module no longer overwrites files outside of its destination path when extracting malicious zip files. - Issue #4844: ZipFile now raises BadZipFile when opens a ZIP file with an incomplete "End of Central Directory" record. Original patch by Guilherme Polo and Alan McIntyre. - Issue #17071: Signature.bind() now works when one of the keyword arguments is named ``self``. - Issue #12004: Fix an internal error in PyZipFile when writing an invalid Python file. Patch by Ben Morgan. - Have py_compile use importlib as much as possible to avoid code duplication. - Issue #180022: Have site.addpackage() consider already known paths even when none are explicitly passed in. Bug report and fix by Kirill. - Issue #1602133: on Mac OS X a shared library build (``--enable-shared``) now fills the ``os.environ`` variable correctly. - Issue #15505: `unittest.installHandler` no longer assumes SIGINT handler is set to a callable object. - Issue #9290: In IDLE the sys.std* streams now implement io.TextIOBase interface and support all mandatory methods and properties. - Issue #13454: Fix a crash when deleting an iterator created by itertools.tee() if all other iterators were very advanced before. - Issue #12411: Fix to cgi.parse_multipart to correctly use bytes boundaries and bytes data. Patch by Jonas Wagner. - Issue #16957: shutil.which() no longer searches a bare file name in the current directory on Unix and no longer searches a relative file path with a directory part in PATH directories. Patch by Thomas Kluyver. - Issue #1159051: GzipFile now raises EOFError when reading a corrupted file with truncated header or footer. - Issue #16993: shutil.which() now preserves the case of the path and extension on Windows. - Issue #16992: On Windows in signal.set_wakeup_fd, validate the file descriptor argument. - Issue #16422: For compatibility with the Python version, the C version of decimal now uses strings instead of integers for rounding mode constants. - Issue #15861: tkinter now correctly works with lists and tuples containing strings with whitespaces, backslashes or unbalanced braces. - Issue #9720: zipfile now writes correct local headers for files larger than 4 GiB. - Issue #5066: Update IDLE docs. Patch by Todd Rovito. - Issue #16955: Fix the poll() method for multiprocessing's socket connections on Windows. - SSLContext.load_dh_params() now properly closes the input file. - Issue #16829: IDLE printing no longer fails if there are spaces or other special characters in the file path. - Issue #15031: Refactor some .pyc management code to cut down on code duplication. Thanks to Ronan Lamy for the report and taking an initial stab at the problem. - Issue #16398: Optimize deque.rotate() so that it only moves pointers and doesn't touch the underlying data with increfs and decrefs. - Issue #16900: Issue a ResourceWarning when an ssl socket is left unclosed. - Issue #13899: \A, \Z, and \B now correctly match the A, Z, and B literals when used inside character classes (e.g. '[\A]'). Patch by Matthew Barnett. - Issue #15545: Fix regression in sqlite3's iterdump method where it was failing if the connection used a row factory (such as sqlite3.Row) that produced unsortable objects. (Regression was introduced by fix for 9750). - Issue #16491: IDLE now prints chained exception tracebacks. - fcntl: add F_DUPFD_CLOEXEC constant, available on Linux 2.6.24+. - Issue #15972: Fix error messages when os functions expecting a file name or file descriptor receive the incorrect type. - Issue #8109: The ssl module now has support for server-side SNI, thanks to a :meth:`SSLContext.set_servername_callback` method. Patch by Daniel Black. - Issue #16860: In tempfile, use O_CLOEXEC when available to set the close-on-exec flag atomically. - Issue #16674: random.getrandbits() is now 20-40% faster for small integers. - Issue #16009: JSON error messages now provide more information. - Issue #16828: Fix error incorrectly raised by bz2.compress(b'') and bz2.BZ2Compressor.compress(b''). Initial patch by Martin Packman. - Issue #16833: In http.client.HTTPConnection, do not concatenate the request headers and body when the payload exceeds 16 KB, since it can consume more memory for no benefit. Patch by Benno Leslie. - Issue #16541: tk_setPalette() now works with keyword arguments. - Issue #16820: In configparser, `parser.popitem()` no longer raises ValueError. This makes `parser.clean()` work correctly. - Issue #16820: In configparser, ``parser['section'] = {}`` now preserves section order within the parser. This makes `parser.update()` preserve section order as well. - Issue #16820: In configparser, ``parser['DEFAULT'] = {}`` now correctly clears previous values stored in the default section. Same goes for ``parser.update({'DEFAULT': {}})``. - Issue #16819: IDLE method completion now correctly works for bytes literals. - Issue #9586: Redefine SEM_FAILED on MacOSX to keep compiler happy. - Issue #16787: Increase asyncore and asynchat default output buffers size, to decrease CPU usage and increase throughput. - Issue #10527: make multiprocessing use poll() instead of select() if available. - Issue #16688: Now regexes contained backreferences correctly work with non-ASCII strings. Patch by Matthew Barnett. - Issue #16486: Make aifc files act as context managers. - Issue #16485: Now file descriptors are closed if file header patching failed on closing an aifc file. - Issue #16640: Run less code under a lock in sched module. - Issue #16165: sched.scheduler.run() no longer blocks a scheduler for other threads. - Issue #16641: Default values of sched.scheduler.enter() are no longer modifiable. - Issue #16504: IDLE now catches SyntaxErrors raised by tokenizer. Patch by Roger Serwy. - Issue #16618: Make glob.glob match consistently across strings and bytes regarding leading dots. Patch by Serhiy Storchaka. - Issue #16788: Add samestat to Lib/ntpath.py - Issue #16702: test_urllib2_localnet tests now correctly ignores proxies for localhost tests. - Issue #16511: Use default IDLE width and height if config param is not valid. Patch Serhiy Storchaka. - Issue #16713: Parsing of 'tel' urls using urlparse separates params from path. - Issue #16443: Add docstrings to regular expression match objects. Patch by Anton Kasyanov. - Issue #15701: Fix HTTPError info method call to return the headers information. - Issue #16752: Add a missing import to modulefinder. Patch by Berker Peksag. - Issue #16646: ftplib.FTP.makeport() might lose socket error details. (patch by Serhiy Storchaka) - Issue #16626: Fix infinite recursion in glob.glob() on Windows when the pattern contains a wildcard in the drive or UNC path. Patch by Serhiy Storchaka. - Issue #15783: Except for the number methods, the C version of decimal now supports all None default values present in decimal.py. These values were largely undocumented. - Issue #11175: argparse.FileType now accepts encoding and errors arguments. Patch by Lucas Maystre. - Issue #16488: epoll() objects now support the `with` statement. Patch by Serhiy Storchaka. - Issue #16298: In HTTPResponse.read(), close the socket when there is no Content-Length and the incoming stream is finished. Patch by Eran Rundstein. - Add abc.ABC class to use inheritance rather than a direct invocation of ABCMeta metaclass. Patch by Bruno Dupuis. - Expose the TCP_FASTOPEN and MSG_FASTOPEN flags in socket when they're available. - Issue #15701: Add a .headers attribute to urllib.error.HTTPError. Patch contributed by Berker Peksag. - Issue #15872: Fix 3.3 regression introduced by the new fd-based shutil.rmtree that caused it to not ignore certain errors when ignore_errors was set. Patch by Alessandro Moura and Serhiy Storchaka. - Issue #16248: Disable code execution from the user's home directory by tkinter when the -E flag is passed to Python. Patch by Zachary Ware. - Issue #13390: New function :func:`sys.getallocatedblocks()` returns the number of memory blocks currently allocated. - Issue #16628: Fix a memory leak in ctypes.resize(). - Issue #13614: Fix setup.py register failure with invalid rst in description. Patch by Julien Courteau and Pierre Paul Lefebvre. - Issue #13512: Create ~/.pypirc securely (CVE-2011-4944). Initial patch by Philip Jenvey, tested by Mageia and Debian. - Issue #7719: Make distutils ignore ``.nfs*`` files instead of choking later on. Initial patch by SilentGhost and Jeff Ramnani. - Issue #13120: Allow to call pdb.set_trace() from thread. Patch by Ilya Sandler. - Issue #16585: Make CJK encoders support error handlers that return bytes per PEP 383. - Issue #10182: The re module doesn't truncate indices to 32 bits anymore. Patch by Serhiy Storchaka. - Issue #16333: use (",", ": ") as default separator when indent is specified to avoid trailing whitespace. Patch by Serhiy Storchaka. - Issue #16573: In 2to3, treat enumerate() like a consuming call, so superfluous list() calls aren't added to filter(), map(), and zip() which are directly passed enumerate(). - Issue #16549: Make json.tool work again on Python 3 and add tests. Initial patch by Berker Peksag and Serhiy Storchaka. - Issue #12848: The pure Python pickle implementation now treats object lengths as unsigned 32-bit integers, like the C implementation does. Patch by Serhiy Storchaka. - Issue #16423: urllib.request now has support for ``data:`` URLs. Patch by Mathias Panzenböck. - Issue #4473: Add a POP3.stls() to switch a clear-text POP3 session into an encrypted POP3 session, on supported servers. Patch by Lorenzo Catucci. - Issue #4473: Add a POP3.capa() method to query the capabilities advertised by the POP3 server. Patch by Lorenzo Catucci. - Issue #4473: Ensure the socket is shutdown cleanly in POP3.close(). Patch by Lorenzo Catucci. - Issue #16522: added FAIL_FAST flag to doctest. - Issue #15627: Add the importlib.abc.SourceLoader.source_to_code() method. - Issue #16408: Fix file descriptors not being closed in error conditions in the zipfile module. Patch by Serhiy Storchaka. - Issue #14631: Add a new :class:`weakref.WeakMethod` to simulate weak references to bound methods. - Issue #16469: Fix exceptions from float -> Fraction and Decimal -> Fraction conversions for special values to be consistent with those for float -> int and Decimal -> int. Patch by Alexey Kachayev. - Issue #16481: multiprocessing no longer leaks process handles on Windows. - Issue #12428: Add a pure Python implementation of functools.partial(). Patch by Brian Thorne. - Issue #16140: The subprocess module no longer double closes its child subprocess.PIPE parent file descriptors on child error prior to exec(). - Remove a bare print to stdout from the subprocess module that could have happened if the child process wrote garbage to its pre-exec error pipe. - The subprocess module now raises its own SubprocessError instead of a RuntimeError in various error situations which should not normally happen. - Issue #16327: The subprocess module no longer leaks file descriptors used for stdin/stdout/stderr pipes to the child when fork() fails. - Issue #14396: Handle the odd rare case of waitpid returning 0 when not expected in subprocess.Popen.wait(). - Issue #16411: Fix a bug where zlib.decompressobj().flush() might try to access previously-freed memory. Patch by Serhiy Storchaka. - Issue #16357: fix calling accept() on a SSLSocket created through SSLContext.wrap_socket(). Original patch by Jeff McNeil. - Issue #16409: The reporthook callback made by the legacy urllib.request.urlretrieve API now properly supplies a constant non-zero block_size as it did in Python 3.2 and 2.7. This matches the behavior of urllib.request.URLopener.retrieve. - Issue #16431: Use the type information when constructing a Decimal subtype from a Decimal argument. - Issue #15641: Clean up deprecated classes from importlib Patch by Taras Lyapun. - Issue #16350: zlib.decompressobj().decompress() now accumulates data from successive calls after EOF in unused_data, instead of only saving the argument to the last call. decompressobj().flush() now correctly sets unused_data and unconsumed_tail. A bug in the handling of MemoryError when setting the unconsumed_tail attribute has also been fixed. Patch by Serhiy Storchaka. - Issue #12759: sre_parse now raises a proper error when the name of the group is missing. Initial patch by Serhiy Storchaka. - Issue #16152: fix tokenize to ignore whitespace at the end of the code when no newline is found. Patch by Ned Batchelder. - Issue #16284: Prevent keeping unnecessary references to worker functions in concurrent.futures ThreadPoolExecutor. - Issue #1207589: Add Cut/Copy/Paste items to IDLE right click Context Menu Patch by Todd Rovito. - Issue #16230: Fix a crash in select.select() when one the lists changes size while iterated on. Patch by Serhiy Storchaka. - Issue #16228: Fix a crash in the json module where a list changes size while it is being encoded. Patch by Serhiy Storchaka. - Issue #16351: New function gc.get_stats() returns per-generation collection statistics. - Issue #14897: Enhance error messages of struct.pack and struct.pack_into. Patch by Matti Mäki. - Issue #16316: mimetypes now recognizes the .xz and .txz (.tar.xz) extensions. Patch by Serhiy Storchaka. - Issue #12890: cgitb no longer prints spurious
tags in text
mode when the logdir option is specified.
- Issue #16307: Fix multiprocessing.Pool.map_async not calling its callbacks.
Patch by Janne Karila.
- Issue #16305: Fix a segmentation fault occurring when interrupting
math.factorial.
- Issue #16116: Fix include and library paths to be correct when building C
extensions in venvs.
- Issue #16245: Fix the value of a few entities in html.entities.html5.
- Issue #16301: Fix the localhost verification in urllib/request.py for file://
urls.
- Issue #16250: Fix the invocations of URLError which had misplaced filename
attribute for exception.
- Issue #10836: Fix exception raised when file not found in urlretrieve
Initial patch by Ezio Melotti.
- Issue #14398: Fix size truncation and overflow bugs in the bz2 module.
- Issue #12692: Fix resource leak in urllib.request when talking to an HTTP
server that does not include a "Connection: close" header in its responses.
- Issue #12034: Fix bogus caching of result in check_GetFinalPathNameByHandle.
Patch by Atsuo Ishimoto.
- Improve performance of `lzma.LZMAFile`.
- Issue #16220: wsgiref now always calls close() on an iterable response.
Patch by Brent Tubbs.
- Issue #16270: urllib may hang when used for retrieving files via FTP by using
a context manager. Patch by Giampaolo Rodola'.
- Issue #16461: Wave library should be able to deal with 4GB wav files,
and sample rate of 44100 Hz.
- Issue #16176: Properly identify Windows 8 via platform.platform()
- Issue #16088: BaseHTTPRequestHandler's send_error method includes a
Content-Length header in it's response now. Patch by Antoine Pitrou.
- Issue #16114: The subprocess module no longer provides a misleading error
message stating that args[0] did not exist when either the cwd or executable
keyword arguments specified a path that did not exist.
- Issue #16169: Fix ctypes.WinError()'s confusion between errno and winerror.
- Issue #1492704: shutil.copyfile() raises a distinct SameFileError now if
source and destination are the same file. Patch by Atsuo Ishimoto.
- Issue #13896: Make shelf instances work with 'with' as context managers.
Original patch by Filip Gruszczyński.
- Issue #15417: Add support for csh and fish in venv activation scripts.
- Issue #16123: IDLE - deprecate running without a subprocess.
Patch by Roger Serwy.
- Issue #14377: ElementTree.write and some of the module-level functions have
a new parameter - *short_empty_elements*. It controls how elements with no
contents are emitted.
- Issue #16089: Allow ElementTree.TreeBuilder to work again with a non-Element
element_factory (fixes a regression in SimpleTAL).
- Issue #9650: List commonly used format codes in time.strftime and
time.strptime docsttings. Original patch by Mike Hoy.
- Issue #16034: Fix performance regressions in the new `bz2.BZ2File`
implementation. Initial patch by Serhiy Storchaka.
- `pty.spawn()` now returns the child process status returned by `os.waitpid()`.
- Issue #15756: `subprocess.poll()` now properly handles `errno.ECHILD` to
return a returncode of 0 when the child has already exited or cannot be waited
on.
- Issue #15323: Improve failure message of `Mock.assert_called_once_with()`.
- Issue #16064: ``unittest -m`` claims executable is "python", not "python3".
- Issue #12376: Pass on parameters in `TextTestResult.__init__()` super call.
- Issue #15222: Insert blank line after each message in mbox mailboxes.
- Issue #16013: Fix `csv.Reader` parsing issue with ending quote characters.
Patch by Serhiy Storchaka.
- Issue #15421: Fix an OverflowError in `Calendar.itermonthdates()` after
`datetime.MAXYEAR`. Patch by Cédric Krier.
- Issue #16112: platform.architecture does not correctly escape argument to
/usr/bin/file. Patch by David Benjamin.
- Issue #15970: `xml.etree.ElementTree` now serializes correctly the empty HTML
elements 'meta' and 'param'.
- Issue #15842: The `SocketIO.{readable,writable,seekable}` methods now raise
ValueError when the file-like object is closed. Patch by Alessandro Moura.
- Issue #15876: Fix a refleak in the `curses` module: window.encoding.
- Issue #15881: Fix `atexit` hook in `multiprocessing`. Original patch by Chris
McDonough.
- Issue #15841: The readable(), writable() and seekable() methods of
`io.BytesIO` and `io.StringIO` objects now raise ValueError when the object
has been closed. Patch by Alessandro Moura.
- Issue #16126: PyErr_Format format mismatch in _testcapimodule.c.
Patch by Serhiy Storchaka.
- Issue #15447: Use `subprocess.DEVNULL` in webbrowser, instead of opening
`os.devnull` explicitly and leaving it open.
- Issue #15509: `webbrowser.UnixBrowser` no longer passes empty arguments to
Popen when ``%action`` substitutions produce empty strings.
- Issue #12776, issue #11839: Call `argparse` type function (specified by
add_argument) only once. Before, the type function was called twice in the
case where the default was specified and the argument was given as well. This
was especially problematic for the FileType type, as a default file would
always be opened, even if a file argument was specified on the command line.
- Issue #15906: Fix a regression in argparse caused by the preceding change,
when ``action='append'``, ``type='str'`` and ``default=[]``.
- Issue #16113: Added sha3 module based on the Keccak reference implementation
3.2. The `hashlib` module has four additional hash algorithms: `sha3_224`,
`sha3_256`, `sha3_384` and `sha3_512`. As part of the patch some common
code was moved from _hashopenssl.c to hashlib.h.
- ctypes.call_commethod was removed, since its only usage was in the defunct
samples directory.
- Issue #16692: Added TLSv1.1 and TLSv1.2 support for the ssl modules.
Extension Modules
-----------------
- Issue #17192: Import libffi-3.0.12.
- Issue #12268: The io module file object write methods no longer abort early
when one of its write system calls is interrupted (EINTR).
Tests
-----
- Issue #17448: test_sax now skips if there are no xml parsers available
instead of raising an ImportError.
- Issue #11420: make test suite pass with -B/DONTWRITEBYTECODE set.
Initial patch by Thomas Wouters.
- Issue #10652: make tcl/tk tests run after __all__ test, patch by
Zachary Ware.
- Issue #11963: remove human verification from test_parser and test_subprocess.
- Issue #11732: add a new suppress_crash_popup() context manager to test.support
that disables crash popups on Windows and use it in test_faulthandler and
test_capi.
- Issue #13898: test_ssl no longer prints a spurious stack trace on Ubuntu.
- Issue #17283: Share code between `__main__.py` and `regrtest.py` in
`Lib/test`.
- Issue #17249: convert a test in test_capi to use unittest and reap threads.
- Issue #17107: Test client-side SNI support in urllib.request thanks to
the new server-side SNI support in the ssl module. Initial patch by
Daniel Black.
- Issue #17041: Fix testing when Python is configured with the
--without-doc-strings.
- Issue #16923: Fix ResourceWarnings in test_ssl.
- Issue #15539: Added regression tests for Tools/scripts/pindent.py.
- Issue #16836: Enable IPv6 support even if IPv6 is disabled on the build host.
- Issue #17479: test_io now works with unittest test discovery.
Patch by Zachary Ware.
- Issue #17066: test_robotparser now works with unittest test discovery.
Patch by Zachary Ware.
- Issue #17334: test_index now works with unittest test discovery.
Patch by Zachary Ware.
- Issue #17333: test_imaplib now works with unittest test discovery.
Patch by Zachary Ware.
- Issue #17082: test_dbm* now work with unittest test discovery.
Patch by Zachary Ware.
- Issue #17079: test_ctypes now works with unittest test discovery.
Patch by Zachary Ware.
- Issue #17304: test_hash now works with unittest test discovery.
Patch by Zachary Ware.
- Issue #17303: test_future* now work with unittest test discovery.
Patch by Zachary Ware.
- Issue #17163: test_file now works with unittest test discovery.
Patch by Zachary Ware.
- Issue #16925: test_configparser now works with unittest test discovery.
Patch by Zachary Ware.
- Issue #16918: test_codecs now works with unittest test discovery.
Patch by Zachary Ware.
- Issue #16919: test_crypt now works with unittest test discovery.
Patch by Zachary Ware.
- Issue #16910: test_bytes, test_unicode, and test_userstring now work with
unittest test discovery. Patch by Zachary Ware.
- Issue #16905: test_warnings now works with unittest test discovery.
Initial patch by Berker Peksag.
- Issue #16898: test_bufio now works with unittest test discovery.
Patch by Zachary Ware.
- Issue #16888: test_array now works with unittest test discovery.
Patch by Zachary Ware.
- Issue #16896: test_asyncore now works with unittest test discovery.
Patch by Zachary Ware.
- Issue #16897: test_bisect now works with unittest test discovery.
Initial patch by Zachary Ware.
- Issue #16852: test_genericpath, test_posixpath, test_ntpath, and test_macpath
now work with unittest test discovery. Patch by Zachary Ware.
- Issue #16748: test_heapq now works with unittest test discovery.
- Issue #10646: Tests rearranged for os.samefile/samestat to check for not
just symlinks but also hard links.
- Issue #15302: Switch regrtest from using getopt to using argparse.
- Issue #15324: Fix regrtest parsing of --fromfile, --match, and --randomize
options.
- Issue #16664: Add regression tests for glob's behaviour concerning entries
starting with a ".". Patch by Sebastian Kreft.
- Issue #13390: The ``-R`` option to regrtest now also checks for memory
allocation leaks, using :func:`sys.getallocatedblocks()`.
- Issue #16559: Add more tests for the json module, including some from the
official test suite at json.org. Patch by Serhiy Storchaka.
- Issue #16661: Fix the `os.getgrouplist()` test by not assuming that it gives
the same output as :command:`id -G`.
- Issue #16115: Add some tests for the executable argument to
subprocess.Popen(). Initial patch by Kushal Das.
- Issue #15304: Fix warning message when `os.chdir()` fails inside
`test.support.temp_cwd()`. Patch by Chris Jerdonek.
- Issue #15802: Fix test logic in `TestMaildir.test_create_tmp()`. Patch by
Serhiy Storchaka.
- Issue #15557: Added a test suite for the webbrowser module, thanks to Anton
Barkovsky.
- Issue #16698: Skip posix test_getgroups when built with OS X
deployment target prior to 10.6.
Build
-----
- Issue #17550: Fix the --enable-profiling configure switch.
- Issue #17425: Build with openssl 1.0.1d on Windows.
- Issue #16754: Fix the incorrect shared library extension on linux. Introduce
two makefile macros SHLIB_SUFFIX and EXT_SUFFIX. SO now has the value of
SHLIB_SUFFIX again (as in 2.x and 3.1). The SO macro is removed in 3.4.
- Issue #5033: Fix building of the sqlite3 extension module when the
SQLite library version has "beta" in it. Patch by Andreas Pelme.
- Issue #17228: Fix building without pymalloc.
- Issue #3718: Use AC_ARG_VAR to set MACHDEP in configure.ac.
- Issue #16235: Implement python-config as a shell script.
- Issue #16769: Remove outdated Visual Studio projects.
- Issue #17031: Fix running regen in cross builds.
- Issue #3754: fix typo in pthread AC_CACHE_VAL.
- Issue #15484: Fix _PYTHON_PROJECT_BASE for srcdir != builddir builds;
use _PYTHON_PROJECT_BASE in distutils/sysconfig.py.
- Drop support for Windows 2000.
- Issue #17029: Let h2py search the multiarch system include directory.
- Issue #16953: Fix socket module compilation on platforms with
HAVE_BROKEN_POLL. Patch by Jeffrey Armstrong.
- Issue #16320: Remove redundant Makefile dependencies for strings and bytes.
- Cross compiling needs host and build settings. configure no longer
creates a broken PYTHON_FOR_BUILD variable when --build is missing.
- Fix cross compiling issue in setup.py, ensure that lib_dirs and inc_dirs are
defined in cross compiling mode, too.
- Issue #16593: Have BSD 'make -s' do the right thing, thanks to Daniel Shahaf
- Issue #16262: fix out-of-src-tree builds, if mercurial is not installed.
- Issue #15298: ensure _sysconfigdata is generated in build directory, not
source directory.
- Issue #15833: Fix a regression in 3.3 that resulted in exceptions being
raised if importlib failed to write byte-compiled files. This affected
attempts to build Python out-of-tree from a read-only source directory.
- Issue #15923: Fix a mistake in ``asdl_c.py`` that resulted in a TypeError
after 2801bf875a24 (see #15801).
- Issue #16135: Remove OS/2 support.
- Issue #15819: Make sure we can build Python out-of-tree from a read-only
source directory. (Somewhat related to issue #9860.)
- Issue #15587: Enable Tk high-resolution text rendering on Macs with
Retina displays. Applies to Tkinter apps, such as IDLE, on OS X
framework builds linked with Cocoa Tk 8.5.
- Issue #17161: make install now also installs a python3 man page.
Documentation
-------------
- Issue 17538: Document XML vulnerabilties
- Issue #16642: sched.scheduler timefunc initial default is time.monotonic.
Patch by Ramchandra Apte
- Issue #17047: remove doubled words in docs and docstrings
reported by Serhiy Storchaka and Matthew Barnett.
- Issue #15465: Document the versioning macros in the C API docs rather than
the standard library docs. Patch by Kushal Das.
- Issue #16406: Combine the pages for uploading and registering to PyPI.
- Issue #16403: Document how distutils uses the maintainer field in
PKG-INFO. Patch by Jyrki Pulliainen.
- Issue #16695: Document how glob handles filenames starting with a
dot. Initial patch by Jyrki Pulliainen.
- Issue #8890: Stop advertising an insecure practice by replacing uses
of the /tmp directory with better alternatives in the documentation.
Patch by Geoff Wilson.
- Issue #17203: add long option names to unittest discovery docs.
- Issue #13094: add "Why do lambdas defined in a loop with different values
all return the same result?" programming FAQ.
- Issue #14901: Update portions of the Windows FAQ.
Patch by Ashish Nitin Patil.
- Issue #16267: Better document the 3.3+ approach to combining
@abstractmethod with @staticmethod, @classmethod and @property
- Issue #15209: Clarify exception chaining description in exceptions module
documentation
- Issue #15990: Improve argument/parameter documentation.
- Issue #16209: Move the documentation for the str built-in function to a new
str class entry in the "Text Sequence Type" section.
- Issue #13538: Improve str() and object.__str__() documentation.
- Issue #16489: Make it clearer that importlib.find_loader() needs parent
packages to be explicitly imported.
- Issue #16400: Update the description of which versions of a given package
PyPI displays.
- Issue #15677: Document that zlib and gzip accept a compression level of 0 to
mean 'no compression'. Patch by Brian Brazil.
- Issue #8040: added a version switcher to the documentation. Patch by
Yury Selivanov.
- Additional comments and some style changes in the concurrent.futures URL
retrieval example
- Issue #16115: Improve subprocess.Popen() documentation around args, shell,
and executable arguments.
- Issue #13498: Clarify docs of os.makedirs()'s exist_ok argument. Done with
great native-speaker help from R. David Murray.
- Issue #15533: Clarify docs and add tests for `subprocess.Popen()`'s cwd
argument.
- Issue #15979: Improve timeit documentation.
- Issue #16036: Improve documentation of built-in `int()`'s signature and
arguments.
- Issue #15935: Clarification of `argparse` docs, re: add_argument() type and
default arguments. Patch contributed by Chris Jerdonek.
- Issue #11964: Document a change in v3.2 to the behavior of the indent
parameter of json encoding operations.
- Issue #15116: Remove references to appscript as it is no longer being
supported.
- Issue #15116: Remove references to appscript as it is no longer being
supported.
Tools/Demos
-----------
- Issue #17028: Allowed Python arguments to be supplied to the Windows
launcher.
- Issue #15539: Fix a number of bugs in Tools/scripts/pindent.py. Now
pindent.py works with a "with" statement. pindent.py no longer produces
improper indentation. pindent.py now works with continued lines broken after
"class" or "def" keywords and with continuations at the start of line.
- Issue #11797: Add a 2to3 fixer that maps reload() to imp.reload().
- Issue #10966: Remove the concept of unexpected skipped tests.
- Issue #9893: Removed the Misc/Vim directory.
- Removed the Misc/TextMate directory.
- Issue #16245: Add the Tools/scripts/parse_html5_entities.py script to parse
the list of HTML5 entities and update the html.entities.html5 dictionary.
- Issue #15378: Fix Tools/unicode/comparecodecs.py. Patch by Serhiy Storchaka.
- Issue #13301: use ast.literal_eval() instead of eval() in Tools/i18n/msgfmt.py
Patch by Serhiy Storchaka.
What's New in Python 3.3.0?
===========================
*Release date: 29-Sep-2012*
Core and Builtins
-----------------
- Issue #16046: Fix loading sourceless legacy .pyo files.
- Issue #16060: Fix refcounting bug when `__trunc__()` returns an object whose
`__int__()` gives a non-integer. Patch by Serhiy Storchaka.
Extension Modules
-----------------
- Issue #16012: Fix a regression in pyexpat. The parser's `UseForeignDTD()`
method doesn't require an argument again.
What's New in Python 3.3.0 Release Candidate 3?
===============================================
*Release date: 23-Sep-2012*
Core and Builtins
-----------------
- Issue #15900: Fix reference leak in `PyUnicode_TranslateCharmap()`.
- Issue #15926: Fix crash after multiple reinitializations of the interpreter.
- Issue #15895: Fix FILE pointer leak in one error branch of
`PyRun_SimpleFileExFlags()` when filename points to a pyc/pyo file, closeit is
false an and set_main_loader() fails.
- Fixes for a few crash and memory leak regressions found by Coverity.
Library
-------
- Issue #15882: Change `_decimal` to accept any coefficient tuple when
constructing infinities. This is done for backwards compatibility with
decimal.py: Infinity coefficients are undefined in _decimal (in accordance
with the specification).
- Issue #15925: Fix a regression in `email.util` where the `parsedate()` and
`parsedate_tz()` functions did not return None anymore when the argument could
not be parsed.
Extension Modules
-----------------
- Issue #15973: Fix a segmentation fault when comparing datetime timezone
objects.
- Issue #15977: Fix memory leak in Modules/_ssl.c when the function
_set_npn_protocols() is called multiple times, thanks to Daniel Sommermann.
- Issue #15969: `faulthandler` module: rename dump_tracebacks_later() to
dump_traceback_later() and cancel_dump_tracebacks_later() to
cancel_dump_traceback_later().
- _decimal module: use only C 89 style comments.
What's New in Python 3.3.0 Release Candidate 2?
===============================================
*Release date: 09-Sep-2012*
Core and Builtins
-----------------
- Issue #13992: The trashcan mechanism is now thread-safe. This eliminates
sporadic crashes in multi-thread programs when several long deallocator chains
ran concurrently and involved subclasses of built-in container types.
- Issue #15784: Modify `OSError`.__str__() to better distinguish between errno
error numbers and Windows error numbers.
- Issue #15781: Fix two small race conditions in import's module locking.
Library
-------
- Issue #17158: Add 'symbols' to help() welcome message; clarify
'modules spam' messages.
- Issue #15847: Fix a regression in argparse, which did not accept tuples as
argument lists anymore.
- Issue #15828: Restore support for C extensions in `imp.load_module()`.
- Issue #15340: Fix importing the random module when ``/dev/urandom`` cannot be
opened. This was a regression caused by the hash randomization patch.
- Issue #10650: Deprecate the watchexp parameter of the `Decimal.quantize()`
method.
- Issue #15785: Modify `window.get_wch()` API of the curses module: return a
character for most keys, and an integer for special keys, instead of always
returning an integer. So it is now possible to distinguish special keys like
keypad keys.
- Issue #14223: Fix `window.addch()` of the curses module for special characters
like curses.ACS_HLINE: the Python function addch(int) and addch(bytes) is now
calling the C function waddch()/mvwaddch() (as it was done in Python 3.2),
instead of wadd_wch()/mvwadd_wch(). The Python function addch(str) is still
calling the C function wadd_wch()/mvwadd_wch() if the Python curses is linked
to libncursesw.
Build
-----
- Issue #15822: Really ensure 2to3 grammar pickles are properly installed
(replaces fixes for Issue #15645).
Documentation
-------------
- Issue #15814: The memoryview enhancements in 3.3.0 accidentally permitted the
hashing of multi-dimensional memorviews and memoryviews with multi-byte item
formats. The intended restrictions have now been documented - they will be
correctly enforced in 3.3.1.
What's New in Python 3.3.0 Release Candidate 1?
===============================================
*Release date: 25-Aug-2012*
Core and Builtins
-----------------
- Issue #15573: memoryview comparisons are now performed by value with full
support for any valid struct module format definition.
- Issue #15316: When an item in the fromlist for `__import__()` doesn't exist,
don't raise an error, but if an exception is raised as part of an import do
let that propagate.
- Issue #15778: Ensure that ``str(ImportError(msg))`` returns a str even when
msg isn't a str.
- Issue #2051: Source file permission bits are once again correctly copied to
the cached bytecode file. (The migration to importlib reintroduced this
problem because these was no regression test. A test has been added as part of
this patch)
- Issue #15761: Fix crash when ``PYTHONEXECUTABLE`` is set on Mac OS X.
- Issue #15726: Fix incorrect bounds checking in PyState_FindModule. Patch by
Robin Schreiber.
- Issue #15604: Update uses of `PyObject_IsTrue()` to check for and handle
errors correctly. Patch by Serhiy Storchaka.
- Issue #14846: `importlib.FileFinder` now handles the case where the directory
being searched is removed after a previous import attempt.
Library
-------
- Issue #13248: removed deprecated and undocumented difflib.isbjunk, isbpopular.
- Issue #13370: Ensure that ctypes works on Mac OS X when Python is compiled
using the clang compiler.
- Issue #13072: The array module's 'u' format code is now deprecated and will be
removed in Python 4.0.
- Issue #15544: Fix Decimal.__float__ to work with payload-carrying NaNs.
- Issue #15776: Allow pyvenv to work in existing directory with --clean.
- Issue #15249: email's BytesGenerator now correctly mangles From lines (when
requested) even if the body contains undecodable bytes.
- Issue #15777: Fix a refleak in _posixsubprocess.
- Issue ##665194: Update `email.utils.localtime` to use datetime.astimezone and
correctly handle historic changes in UTC offsets.
- Issue #15199: Fix JavaScript's default MIME type to application/javascript.
Patch by Bohuslav Kabrda.
- Issue #12643: `code.InteractiveConsole` now respects `sys.excepthook` when
displaying exceptions. Patch by Aaron Iles.
- Issue #13579: `string.Formatter` now understands the 'a' conversion specifier.
- Issue #15595: Fix ``subprocess.Popen(universal_newlines=True)`` for certain
locales (utf-16 and utf-32 family). Patch by Chris Jerdonek.
- Issue #15477: In cmath and math modules, add workaround for platforms whose
system-supplied log1p function doesn't respect signs of zeros.
- Issue #15715: `importlib.__import__()` will silence an ImportError when the
use of fromlist leads to a failed import.
- Issue #14669: Fix pickling of connections and sockets on Mac OS X by
sending/receiving an acknowledgment after file descriptor transfer.
TestPicklingConnection has been reenabled for Mac OS X.
- Issue #11062: Fix adding a message from file to Babyl mailbox.
- Issue #15646: Prevent equivalent of a fork bomb when using `multiprocessing`
on Windows without the ``if __name__ == '__main__'`` idiom.
- Issue #15678: Fix IDLE menus when started from OS X command line (3.3.0b2
regression).
C API
-----
Extension Modules
-----------------
Tools/Demos
-----------
Documentation
-------------
- Touched up the Python 2 to 3 porting guide.
- Issue #14674: Add a discussion of the `json` module's standard compliance.
Patch by Chris Rebert.
- Create a 'Concurrent Execution' section in the docs, and split up the
'Optional Operating System Services' section to use a more user-centric
classification scheme (splitting them across the new CE section, IPC and text
processing). Operating system limitatons can be reflected with the Sphinx
``:platform:`` tag, it doesn't make sense as part of the Table of Contents.
- Issue #4966: Bring the sequence docs up to date for the Py3k transition and
the many language enhancements since they were original written.
- The "path importer" misnomer has been replaced with Eric Snow's
more-awkward-but-at-least-not-wrong suggestion of "path based finder" in the
import system reference docs.
- Issue #15640: Document `importlib.abc.Finder` as deprecated.
- Issue #15630: Add an example for "continue" stmt in the tutorial. Patch by
Daniel Ellis.
Tests
-----
- Issue #15747: ZFS always returns EOPNOTSUPP when attempting to set the
UF_IMMUTABLE flag (via either chflags or lchflags); refactor affected tests in
test_posix.py to account for this.
- Issue #15285: Refactor the approach for testing connect timeouts using two
external hosts that have been configured specifically for this type of test.
- Issue #15743: Remove the deprecated method usage in `urllib` tests. Patch by
Jeff Knupp.
- Issue #15615: Add some tests for the `json` module's handling of invalid input
data. Patch by Kushal Das.
Build
-----
- Output lib files for PGO build into PGO directory.
- Pick up 32-bit launcher from PGO directory on 64-bit PGO build.
- Drop ``PC\python_nt.h`` as it's not used. Add input dependency on custom
build step.
- Issue #15511: Drop explicit dependency on pythonxy.lib from _decimal amd64
configuration.
- Add missing PGI/PGO configurations for pywlauncher.
- Issue #15645: Ensure 2to3 grammar pickles are properly installed.
What's New in Python 3.3.0 Beta 2?
==================================
*Release date: 12-Aug-2012*
Core and Builtins
-----------------
- Issue #15568: Fix the return value of ``yield from`` when StopIteration is
raised by a custom iterator.
- Issue #13119: `sys.stdout` and `sys.stderr` are now using "\r\n" newline on
Windows, as Python 2.
- Issue #15534: Fix the fast-search function for non-ASCII Unicode strings.
- Issue #15508: Fix the docstring for `__import__()` to have the proper default
value of 0 for 'level' and to not mention negative levels since they are not
supported.
- Issue #15425: Eliminated traceback noise from more situations involving
importlib.
- Issue #14578: Support modules registered in the Windows registry again.
- Issue #15466: Stop using TYPE_INT64 in marshal, to make importlib.h (and other
byte code files) equal between 32-bit and 64-bit systems.
- Issue #1692335: Move initial exception args assignment to
`BaseException.__new__()` to help pickling of naive subclasses.
- Issue #12834: Fix `PyBuffer_ToContiguous()` for non-contiguous arrays.
- Issue #15456: Fix code `__sizeof__()` after #12399 change. Patch by Serhiy
Storchaka.
- Issue #15404: Refleak in PyMethodObject repr.
- Issue #15394: An issue in `PyModule_Create()` that caused references to be
leaked on some error paths has been fixed. Patch by Julia Lawall.
- Issue #15368: An issue that caused bytecode generation to be non-deterministic
has been fixed.
- Issue #15202: Consistently use the name "follow_symlinks" for new parameters
in os and shutil functions.
- Issue #15314: ``__main__.__loader__`` is now set correctly during interpreter
startup.
- Issue #15111: When a module imported using 'from import' has an ImportError
inside itself, don't mask that fact behind a generic ImportError for the
module itself.
- Issue #15293: Add GC support to the AST base node type.
- Issue #15291: Fix a memory leak where AST nodes where not properly
deallocated.
- Issue #15110: Fix the tracebacks generated by "import xxx" to not show the
importlib stack frames.
- Issue #16369: Global PyTypeObjects not initialized with PyType_Ready(...).
- Issue #15020: The program name used to search for Python's path is now
"python3" under Unix, not "python".
- Issue #15897: zipimport.c doesn't check return value of fseek().
Patch by Felipe Cruz.
- Issue #15033: Fix the exit status bug when modules invoked using -m switch,
return the proper failure return value (1). Patch contributed by Jeff Knupp.
- Issue #15229: An `OSError` subclass whose __init__ doesn't call back
OSError.__init__ could produce incomplete instances, leading to crashes when
calling str() on them.
- Issue #15307: Virtual environments now use symlinks with framework builds on
Mac OS X, like other POSIX builds.
Library
-------
- Issue #14590: configparser now correctly strips inline comments when delimiter
occurs earlier without preceding space.
- Issue #15424: Add a `__sizeof__()` implementation for array objects. Patch by
Ludwig Hähne.
- Issue #15576: Allow extension modules to act as a package's __init__ module.
- Issue #15502: Have `importlib.invalidate_caches()` work on `sys.meta_path`
instead of `sys.path_importer_cache`.
- Issue #15163: Pydoc shouldn't list __loader__ as module data.
- Issue #15471: Do not use mutable objects as defaults for
`importlib.__import__()`.
- Issue #15559: To avoid a problematic failure mode when passed to the bytes
constructor, objects in the ipaddress module no longer implement `__index__()`
(they still implement `__int__()` as appropriate).
- Issue #15546: Fix handling of pathological input data in the peek() and
read1() methods of the BZ2File, GzipFile and LZMAFile classes.
- Issue #13052: Fix IDLE crashing when replace string in Search/Replace dialog
ended with ``\``. Patch by Roger Serwy.
- Issue #12655: Instead of requiring a custom type, `os.sched_getaffinity()` and
`os.sched_setaffinity()` now use regular sets of integers to represent the
CPUs a process is restricted to.
- Issue #15538: Fix compilation of the `socket.getnameinfo()` /
`socket.getaddrinfo()` emulation code. Patch by Philipp Hagemeister.
- Issue #15519: Properly expose WindowsRegistryFinder in importlib (and use the
correct term for it). Original patch by Eric Snow.
- Issue #15502: Bring the importlib ABCs into line with the current state of the
import protocols given PEP 420. Original patch by Eric Snow.
- Issue #15499: Launching a webbrowser in Unix used to sleep for a few seconds.
Original patch by Anton Barkovsky.
- Issue #15463: The faulthandler module truncates strings to 500 characters,
instead of 100, to be able to display long file paths.
- Issue #6056: Make `multiprocessing` use setblocking(True) on the sockets it
uses. Original patch by J Derek Wilson.
- Issue #15364: Fix sysconfig.get_config_var('srcdir') to be an absolute path.
- Issue #15041: Update "see also" list in tkinter documentation.
- Issue #15413: `os.times()` had disappeared under Windows.
- Issue #15402: An issue in the struct module that caused `sys.getsizeof()` to
return incorrect results for struct.Struct instances has been fixed. Initial
patch by Serhiy Storchaka.
- Issue #15232: When mangle_from is True, `email.Generator` now correctly
mangles lines that start with 'From ' that occur in a MIME preamble or
epilogue.
- Issue #15094: Incorrectly placed #endif in _tkinter.c. Patch by Serhiy
Storchaka.
- Issue #13922: `argparse` no longer incorrectly strips '--'s that appear after
the first one.
- Issue #12353: `argparse` now correctly handles null argument values.
- Issue #10017, issue #14998: Fix TypeError using pprint on dictionaries with
user-defined types as keys or other unorderable keys.
- Issue #15397: `inspect.getmodulename()` is now based directly on importlib via
a new `importlib.machinery.all_suffixes()` API.
- Issue #14635: `telnetlib` will use poll() rather than select() when possible to
avoid failing due to the select() file descriptor limit.
- Issue #15180: Clarify posixpath.join() error message when mixing str & bytes.
- Issue #15343: pkgutil now includes an iter_importer_modules implementation for
importlib.machinery.FileFinder (similar to the way it already handled
zipimport.zipimporter).
- Issue #15314: runpy now sets __main__.__loader__ correctly.
- Issue #15357: The import emulation in pkgutil is now deprecated. pkgutil uses
importlib internally rather than the emulation.
- Issue #15233: Python now guarantees that callables registered with the atexit
module will be called in a deterministic order.
- Issue #15238: `shutil.copystat()` now copies Linux "extended attributes".
- Issue #15230: runpy.run_path now correctly sets __package__ as described in
the documentation.
- Issue #15315: Support VS 2010 in distutils cygwincompiler.
- Issue #15294: Fix a regression in pkgutil.extend_path()'s handling of nested
namespace packages.
- Issue #15056: `imp.cache_from_source()` and `imp.source_from_cache()` raise
NotImplementedError when `sys.implementation.cache_tag` is set to None.
- Issue #15256: Grammatical mistake in exception raised by `imp.find_module()`.
- Issue #5931: `wsgiref` environ variable SERVER_SOFTWARE will specify an
implementation specific term like CPython, Jython instead of generic "Python".
- Issue #13248: Remove obsolete argument "max_buffer_size" of BufferedWriter and
BufferedRWPair, from the io module.
- Issue #13248: Remove obsolete argument "version" of `argparse.ArgumentParser`.
- Issue #14814: Implement more consistent ordering and sorting behaviour for
ipaddress objects.
- Issue #14814: `ipaddress` network objects correctly return NotImplemented when
compared to arbitrary objects instead of raising TypeError.
- Issue #14990: Correctly fail with SyntaxError on invalid encoding declaration.
- Issue #14814: `ipaddress` now provides more informative error messages when
constructing instances directly (changes permitted during beta due to
provisional API status).
- Issue #15247: `io.FileIO` now raises an error when given a file descriptor
pointing to a directory.
- Issue #15261: Stop os.stat(fd) crashing on Windows when fd not open.
- Issue #15166: Implement `imp.get_tag()` using `sys.implementation.cache_tag`.
- Issue #15210: Catch KeyError when `importlib.__init__()` can't find
_frozen_importlib in sys.modules, not ImportError.
- Issue #15030: `importlib.abc.PyPycLoader` now supports the new source size
header field in .pyc files.
- Issue #5346: Preserve permissions of mbox, MMDF and Babyl mailbox files on
flush().
- Issue #10571: Fix the "--sign" option of distutils' upload command. Patch by
Jakub Wilk.
- Issue #9559: If messages were only added, a new file is no longer created and
renamed over the old file when flush() is called on an mbox, MMDF or Babyl
mailbox.
- Issue #10924: Fixed `crypt.mksalt()` to use a RNG that is suitable for
cryptographic purpose.
- Issue #15184: Ensure consistent results of OS X configuration tailoring for
universal builds by factoring out common OS X-specific customizations from
sysconfig, distutils.sysconfig, distutils.util, and distutils.unixccompiler
into a new module _osx_support.
C API
-----
- Issue #15610: `PyImport_ImportModuleEx()` now uses a 'level' of 0 instead of -1.
- Issue #15169, issue #14599: Strip out the C implementation of
`imp.source_from_cache()` used by PyImport_ExecCodeModuleWithPathnames() and
used the Python code instead. Leads to PyImport_ExecCodeModuleObject() to not
try to infer the source path from the bytecode path as
PyImport_ExecCodeModuleWithPathnames() does.
Extension Modules
-----------------
- Issue #6493: An issue in ctypes on Windows that caused structure bitfields of
type `ctypes.c_uint32` and width 32 to incorrectly be set has been fixed.
- Issue #15194: Update libffi to the 3.0.11 release.
Tools/Demos
-----------
- Issue #15458: python-config gets a new option --configdir to print the $LIBPL
value.
- Move importlib.test.benchmark to Tools/importbench.
- Issue #12605: The gdb hooks for debugging CPython (within Tools/gdb) have been
enhanced to show information on more C frames relevant to CPython within the
"py-bt" and "py-bt-full" commands:
* C frames that are waiting on the GIL
* C frames that are garbage-collecting
* C frames that are due to the invocation of a PyCFunction
Documentation
-------------
- Issue #15444: Use proper spelling for non-ASCII contributor names. Patch by
Serhiy Storchaka.
- Issue #15295: Reorganize and rewrite the documentation on the import system.
- Issue #15230: Clearly document some of the limitations of the runpy module and
nudge readers towards importlib when appropriate.
- Issue #15053: Copy Python 3.3 import lock change notice to all relevant
functions in imp instead of just at the top of the relevant section.
- Issue #15288: Link to the term "loader" in notes in pkgutil about how things
won't work as expected in Python 3.3 and mark the requisite functions as
"changed" since they will no longer work with modules directly imported by
import itself.
- Issue #13557: Clarify effect of giving two different namespaces to `exec()` or
`execfile()`.
- Issue #15250: Document that `filecmp.dircmp()` compares files shallowly. Patch
contributed by Chris Jerdonek.
- Issue #15442: Expose the default list of directories ignored by
`filecmp.dircmp()` as a module attribute, and expand the list to more modern
values.
Tests
-----
- Issue #15467: Move helpers for `__sizeof__()` tests into test_support. Patch
by Serhiy Storchaka.
- Issue #15320: Make iterating the list of tests thread-safe when running tests
in multiprocess mode. Patch by Chris Jerdonek.
- Issue #15168: Move `importlib.test` to `test.test_importlib`.
- Issue #15091: Reactivate a test on UNIX which was failing thanks to a
forgotten `importlib.invalidate_caches()` call.
- Issue #15230: Adopted a more systematic approach in the runpy tests.
- Issue #15300: Ensure the temporary test working directories are in the same
parent folder when running tests in multiprocess mode from a Python build.
Patch by Chris Jerdonek.
- Issue #15284: Skip {send,recv}msg tests in test_socket when IPv6 is not
enabled. Patch by Brian Brazil.
- Issue #15277: Fix a resource leak in support.py when IPv6 is disabled. Patch
by Brian Brazil.
Build
-----
- Issue #11715: Fix multiarch detection without having Debian development tools
(dpkg-dev) installed.
- Issue #15037: Build OS X installers with local copy of ncurses 5.9 libraries
to avoid curses.unget_wch bug present in older versions of ncurses such as
those shipped with OS X.
- Issue #15560: Fix building _sqlite3 extension on OS X with an SDK. Also, for
OS X installers, ensure consistent sqlite3 behavior and feature availability
by building a local copy of libsqlite3 rather than depending on the wide range
of versions supplied with various OS X releases.
- Issue #8847: Disable COMDAT folding in Windows PGO builds.
- Issue #14018: Fix OS X Tcl/Tk framework checking when using OS X SDKs.
- Issue #16256: OS X installer now sets correct permissions for doc directory.
- Issue #15431: Add _freeze_importlib project to regenerate importlib.h on
Windows. Patch by Kristján Valur Jónsson.
- Issue #14197: For OS X framework builds, ensure links to the shared library
are created with the proper ABI suffix.
- Issue #14330: For cross builds, don't use host python, use host search paths
for host compiler.
- Issue #15235: Allow Berkley DB versions up to 5.3 to build the dbm module.
- Issue #15268: Search curses.h in /usr/include/ncursesw.
What's New in Python 3.3.0 Beta 1?
==================================
*Release date: 27-Jun-2012*
Core and Builtins
-----------------
- Fix a (most likely) very rare memory leak when calling main() and not being
able to decode a command-line argument.
- Issue #14815: Use Py_ssize_t instead of long for the object hash, to
preserve all 64 bits of hash on Win64.
- Issue #12268: File readline, readlines and read() or readall() methods
no longer lose data when an underlying read system call is interrupted.
IOError is no longer raised due to a read system call returning EINTR
from within these methods.
- Issue #11626: Add _SizeT functions to stable ABI.
- Issue #15146: Add PyType_FromSpecWithBases. Patch by Robin Schreiber.
- Issue #15142: Fix reference leak when deallocating instances of types
created using PyType_FromSpec().
- Issue #15042: Add PyState_AddModule and PyState_RemoveModule. Add version
guard for Py_LIMITED_API additions. Patch by Robin Schreiber.
- Issue #10053: Don't close FDs when FileIO.__init__ fails. Loosely based on
the work by Hirokazu Yamamoto.
- Issue #15096: Removed support for ur'' as the raw notation isn't
compatible with Python 2.x's raw unicode strings.
- Issue #13783: Generator objects now use the identifier APIs internally
- Issue #14874: Restore charmap decoding speed to pre-PEP 393 levels.
Patch by Serhiy Storchaka.
- Issue #15026: utf-16 encoding is now significantly faster (up to 10x).
Patch by Serhiy Storchaka.
- Issue #11022: open() and io.TextIOWrapper are now calling
locale.getpreferredencoding(False) instead of locale.getpreferredencoding()
in text mode if the encoding is not specified. Don't change temporary the
locale encoding using locale.setlocale(), use the current locale encoding
instead of the user preferred encoding.
- Issue #14673: Add Eric Snow's sys.implementation implementation.
- Issue #15038: Optimize python Locks on Windows.
Library
-------
- Issue #9803: Don't close IDLE on saving if breakpoint is open.
Patch by Roger Serwy.
- Issue #12288: Consider '0' and '0.0' as valid initialvalue
for tkinter SimpleDialog.
- Issue #15512: Add a __sizeof__ implementation for parser.
Patch by Serhiy Storchaka.
- Issue #15469: Add a __sizeof__ implementation for deque objects.
Patch by Serhiy Storchaka.
- Issue #15489: Add a __sizeof__ implementation for BytesIO objects.
Patch by Serhiy Storchaka.
- Issue #15487: Add a __sizeof__ implementation for buffered I/O objects.
Patch by Serhiy Storchaka.
- Issue #15514: Correct __sizeof__ support for cpu_set.
Patch by Serhiy Storchaka.
- Issue #15187: Bugfix: remove temporary directories test_shutil was leaving
behind.
- Issue #15177: Added dir_fd parameter to os.fwalk().
- Issue #15176: Clarified behavior, documentation, and implementation
of os.listdir().
- Issue #15061: Re-implemented hmac.compare_digest() in C to prevent further
timing analysis and to support all buffer protocol aware objects as well as
ASCII only str instances safely.
- Issue #15164: Change return value of platform.uname() from a
plain tuple to a collections.namedtuple.
- Support Mageia Linux in the platform module.
- Issue #11678: Support Arch linux in the platform module.
- Issue #15118: Change return value of os.uname() and os.times() from
plain tuples to immutable iterable objects with named attributes
(structseq objects).
- Speed up _decimal by another 10-15% by caching the thread local context
that was last accessed. In the pi benchmark (64-bit platform, prec=9),
_decimal is now only 1.5x slower than float.
- Remove the packaging module, which is not ready for prime time.
- Issue #15154: Add "dir_fd" parameter to os.rmdir, remove "rmdir"
parameter from os.remove / os.unlink.
- Issue #4489: Add a shutil.rmtree that isn't susceptible to symlink attacks.
It is used automatically on platforms supporting the necessary os.openat()
and os.unlinkat() functions. Main code by Martin von Löwis.
- Issue #15156: HTMLParser now uses the new "html.entities.html5" dictionary.
- Issue #11113: add a new "html5" dictionary containing the named character
references defined by the HTML5 standard and the equivalent Unicode
character(s) to the html.entities module.
- Issue #15114: the strict mode of HTMLParser and the HTMLParseError exception
are deprecated now that the parser is able to parse invalid markup.
- Issue #3665: \u and \U escapes are now supported in unicode regular
expressions. Patch by Serhiy Storchaka.
- Issue #15153: Added inspect.getgeneratorlocals to simplify white box
testing of generator state updates
- Issue #13062: Added inspect.getclosurevars to simplify testing stateful
closures
- Issue #11024: Fixes and additional tests for Time2Internaldate.
- Issue #14626: Large refactoring of functions / parameters in the os module.
Many functions now support "dir_fd" and "follow_symlinks" parameters;
some also support accepting an open file descriptor in place of a path
string. Added os.support_* collections as LBYL helpers. Removed many
functions only previously seen in 3.3 alpha releases (often starting with
"f" or "l", or ending with "at"). Originally suggested by Serhiy Storchaka;
implemented by Larry Hastings.
- Issue #15008: Implement PEP 362 "Signature Objects".
Patch by Yury Selivanov.
- Issue: #15138: base64.urlsafe_{en,de}code() are now 3-4x faster.
- Issue #444582: Add shutil.which, for finding programs on the system path.
Original patch by Erik Demaine, with later iterations by Jan Killian
and Brian Curtin.
- Issue #14837: SSL errors now have ``library`` and ``reason`` attributes
describing precisely what happened and in which OpenSSL submodule. The
str() of a SSLError is also enhanced accordingly.
- Issue #9527: datetime.astimezone() method will now supply a class
timezone instance corresponding to the system local timezone when
called with no arguments.
- Issue #14653: email.utils.mktime_tz() no longer relies on system
mktime() when timezone offest is supplied.
- Issue #14684: zlib.compressobj() and zlib.decompressobj() now support the use
of predefined compression dictionaries. Original patch by Sam Rushing.
- Fix GzipFile's handling of filenames given as bytes objects.
- Issue #14772: Return destination values from some shutil functions.
- Issue #15064: Implement context manager protocol for multiprocessing types
- Issue #15101: Make pool finalizer avoid joining current thread.
- Issue #14657: The frozen instance of importlib used for bootstrap is now
also the module imported as importlib._bootstrap.
- Issue #14055: Add __sizeof__ support to _elementtree.
- Issue #15054: A bug in tokenize.tokenize that caused string literals
with 'b' prefixes to be incorrectly tokenized has been fixed.
Patch by Serhiy Storchaka.
- Issue #15006: Allow equality comparison between naive and aware
time or datetime objects.
- Issue #14982: Document that pkgutil's iteration functions require the
non-standard iter_modules() method to be defined by an importer (something
the importlib importers do not define).
- Issue #15036: Mailbox no longer throws an error if a flush is done
between operations when removing or changing multiple items in mbox,
MMDF, or Babyl mailboxes.
- Issue #14059: Implement multiprocessing.Barrier.
- Issue #15061: The inappropriately named hmac.secure_compare has been
renamed to hmac.compare_digest, restricted to operating on bytes inputs
only and had its documentation updated to more accurately reflect both its
intent and its limitations
- Issue #13841: Make child processes exit using sys.exit() on Windows.
- Issue #14936: curses_panel was converted to PEP 3121 and PEP 384 API.
Patch by Robin Schreiber.
- Issue #1667546: On platforms supporting tm_zone and tm_gmtoff fields
in struct tm, time.struct_time objects returned by time.gmtime(),
time.localtime() and time.strptime() functions now have tm_zone and
tm_gmtoff attributes. Original patch by Paul Boddie.
- Rename adjusted attribute to adjustable in time.get_clock_info() result.
- Issue #3518: Remove references to non-existent BaseManager.from_address()
method.
- Issue #13857: Added textwrap.indent() function (initial patch by Ezra
Berch)
- Issue #2736: Added datetime.timestamp() method.
- Issue #13854: Make multiprocessing properly handle non-integer
non-string argument to SystemExit.
- Issue #12157: Make pool.map() empty iterables correctly. Initial
patch by mouad.
- Issue #11823: disassembly now shows argument counts on calls with keyword args.
- Issue #14711: os.stat_float_times() has been deprecated.
- LZMAFile now accepts the modes "rb"/"wb"/"ab" as synonyms of "r"/"w"/"a".
- The bz2 and lzma modules now each contain an open() function, allowing
compressed files to readily be opened in text mode as well as binary mode.
- BZ2File.__init__() and LZMAFile.__init__() now accept a file object as their
first argument, rather than requiring a separate "fileobj" argument.
- gzip.open() now accepts file objects as well as filenames.
- Issue #14992: os.makedirs(path, exist_ok=True) would raise an OSError
when the path existed and had the S_ISGID mode bit set when it was
not explicitly asked for. This is no longer an exception as mkdir
cannot control if the OS sets that bit for it or not.
- Issue #14989: Make the CGI enable option to http.server available via command
line.
- Issue #14987: Add a missing import statement to inspect.
- Issue #1079: email.header.decode_header now correctly parses all the examples
in RFC2047. There is a necessary visible behavior change: the leading and/or
trailing whitespace on ASCII parts is now preserved.
- Issue #14969: Better handling of exception chaining in contextlib.ExitStack
- Issue #14962: Update text coloring in IDLE shell window after changing
options. Patch by Roger Serwy.
- Issue #14963: Convert contextlib.ExitStack.__exit__ to use an iterative
algorithm (Patch by Alon Horev)
- Issue #14785: Add sys._debugmallocstats() to help debug low-level memory
allocation issues
- Issue #14443: Ensure that .py files are byte-compiled with the correct Python
executable within bdist_rpm even on older versions of RPM
C-API
-----
- Issue #13783: Inadvertent additions to the public C API in the PEP 380
implementation have either been removed or marked as private interfaces.
Extension Modules
-----------------
- Issue #15000: Support the "unique" x32 architecture in _posixsubprocess.c.
Documentation
-------------
- Issue #15081: Document PyState_FindModule.
Patch by Robin Schreiber.
- Issue #14814: Added first draft of ipaddress module API reference
Tests
-----
- Issue #14769: test_capi now has SkipitemTest, which cleverly checks
for "parity" between PyArg_ParseTuple() and the Python/getargs.c static
function skipitem() for all possible "format units".
- test_nntplib now tolerates being run from behind NNTP gateways that add
"X-Antivirus" headers to articles
- Issue #15043: test_gdb is now skipped entirely if gdb security settings
block loading of the gdb hooks
- Issue #14963: Add test cases for exception handling behaviour
in contextlib.ExitStack (Initial patch by Alon Horev)
Build
-----
- Issue #13590: Improve support for OS X Xcode 4:
* Try to avoid building Python or extension modules with problematic
llvm-gcc compiler.
* Since Xcode 4 removes ppc support, extension module builds now
check for ppc compiler support and automatically remove ppc and
ppc64 archs when not available.
* Since Xcode 4 no longer install SDKs in default locations,
extension module builds now revert to using installed headers
and libs if the SDK used to build the interpreter is not
available.
* Update ./configure to use better defaults for universal builds;
in particular, --enable-universalsdk=yes uses the Xcode default
SDK and --with-universal-archs now defaults to "intel" if ppc
not available.
- Issue #14225: Fix Unicode support for curses (#12567) on OS X
- Issue #14928: Fix importlib bootstrap issues by using a custom executable
(Modules/_freeze_importlib) to build Python/importlib.h.
What's New in Python 3.3.0 Alpha 4?
===================================
*Release date: 31-May-2012*
Core and Builtins
-----------------
- Issue #14835: Make plistlib output empty arrays & dicts like OS X.
Patch by Sidney San Martín.
- Issue #14744: Use the new _PyUnicodeWriter internal API to speed up
str%args and str.format(args).
- Issue #14930: Make memoryview objects weakrefable.
- Issue #14775: Fix a potential quadratic dict build-up due to the garbage
collector repeatedly trying to untrack dicts.
- Issue #14857: fix regression in references to PEP 3135 implicit __class__
closure variable (Reopens issue #12370)
- Issue #14712 (PEP 405): Virtual environments. Implemented by Vinay Sajip.
- Issue #14660 (PEP 420): Namespace packages. Implemented by Eric Smith.
- Issue #14494: Fix __future__.py and its documentation to note that
absolute imports are the default behavior in 3.0 instead of 2.7.
Patch by Sven Marnach.
- Issue #9260: A finer-grained import lock. Most of the import sequence
now uses per-module locks rather than the global import lock, eliminating
well-known issues with threads and imports.
- Issue #14624: UTF-16 decoding is now 3x to 4x faster on various inputs.
Patch by Serhiy Storchaka.
- asdl_seq and asdl_int_seq are now Py_ssize_t sized.
- Issue #14133 (PEP 415): Implement suppression of __context__ display with an
attribute on BaseException. This replaces the original mechanism of PEP 409.
- Issue #14417: Mutating a dict during lookup now restarts the lookup instead
of raising a RuntimeError (undoes issue #14205).
- Issue #14738: Speed-up UTF-8 decoding on non-ASCII data. Patch by Serhiy
Storchaka.
- Issue #14700: Fix two broken and undefined-behaviour-inducing overflow checks
in old-style string formatting.
- Issue #14705: The PyArg_Parse() family of functions now support the 'p' format
unit, which accepts a "boolean predicate" argument. It converts any Python
value into an integer--0 if it is "false", and 1 otherwise.
Library
-------
- Issue #14690: Use monotonic clock instead of system clock in the sched,
subprocess and trace modules.
- Issue #14958: Change IDLE systax highlighting to recognize all string and
byte literals supported in Python 3.3.
- Issue #10997: Prevent a duplicate entry in IDLE's "Recent Files" menu.
- Issue #14443: Tell rpmbuild to use the correct version of Python in
bdist_rpm. Initial patch by Ross Lagerwall.
- Issue #14929: Stop Idle 3.x from closing on Unicode decode errors when
grepping. Patch by Roger Serwy.
- Issue #12515: email now registers a defect if it gets to EOF while parsing
a MIME part without seeing the closing MIME boundary.
- Issue #12510: Attempting to get invalid tooltip no longer closes Idle.
Other tooltipss have been corrected or improved and the number of tests
has been tripled. Original patch by Roger Serwy.
- Issue #1672568: email now always decodes base64 payloads, adding padding and
ignoring non-base64-alphabet characters if needed, and registering defects
for any such problems.
- Issue #14925: email now registers a defect when the parser decides that there
is a missing header/body separator line. MalformedHeaderDefect, which the
existing code would never actually generate, is deprecated.
- Issue #10365: File open dialog now works instead of crashing even when
the parent window is closed before the dialog. Patch by Roger Serwy.
- Issue #8739: Updated smtpd to support RFC 5321, and added support for the
RFC 1870 SIZE extension.
- Issue #665194: Added a localtime function to email.utils to provide an
aware local datetime for use in setting Date headers.
- Issue #12586: Added new provisional policies that implement convenient
unicode support for email headers. See What's New for details.
- Issue #14731: Refactored email Policy framework to support full backward
compatibility with Python 3.2 by default yet allow for the introduction of
new features through new policies. Note that Policy.must_be_7bit is renamed
to cte_type.
- Issue #14876: Use user-selected font for highlight configuration.
- Issue #14920: Fix the help(urllib.parse) failure on locale C on terminals.
Have ascii characters in help.
- Issue #14548: Make multiprocessing finalizers check pid before
running to cope with possibility of gc running just after fork.
- Issue #14863: Update the documentation of os.fdopen() to reflect the
fact that it's only a thin wrapper around open() anymore.
- Issue #14036: Add an additional check to validate that port in urlparse does
not go in illegal range and returns None.
- Issue #14862: Add missing names to os.__all__
- Issue #14875: Use float('inf') instead of float('1e66666') in the json module.
- Issue #13585: Added contextlib.ExitStack
- PEP 3144, Issue #14814: Added the ipaddress module
- Issue #14426: Correct the Date format in Expires attribute of Set-Cookie
Header in Cookie.py.
- Issue #14588: The types module now provide new_class() and prepare_class()
functions to support PEP 3115 compliant dynamic class creation. Patch by
Daniel Urban and Nick Coghlan.
- Issue #13152: Allow to specify a custom tabsize for expanding tabs in
textwrap. Patch by John Feuerstein.
- Issue #14721: Send the correct 'Content-length: 0' header when the body is an
empty string ''. Initial Patch contributed by Arve Knudsen.
- Issue #14072: Fix parsing of 'tel' URIs in urlparse by making the check for
ports stricter.
- Issue #9374: Generic parsing of query and fragment portions of url for any
scheme. Supported both by RFC3986 and RFC2396.
- Issue #14798: Fix the functions in pyclbr to raise an ImportError
when the first part of a dotted name is not a package. Patch by
Xavier de Gaye.
- Issue #12098: multiprocessing on Windows now starts child processes
using the same sys.flags as the current process. Initial patch by
Sergey Mezentsev.
- Issue #13031: Small speed-up for tarfile when unzipping tarfiles.
Patch by Justin Peel.
- Issue #14780: urllib.request.urlopen() now has a ``cadefault`` argument
to use the default certificate store. Initial patch by James Oakley.
- Issue #14829: Fix bisect and range() indexing with large indices
(>= 2 ** 32) under 64-bit Windows.
- Issue #14732: The _csv module now uses PEP 3121 module initialization.
Patch by Robin Schreiber.
- Issue #14809: Add HTTP status codes introduced by RFC 6585 to http.server
and http.client. Patch by EungJun Yi.
- Issue #14777: tkinter may return undecoded UTF-8 bytes as a string when
accessing the Tk clipboard. Modify clipboad_get() to first request type
UTF8_STRING when no specific type is requested in an X11 windowing
environment, falling back to the current default type STRING if that fails.
Original patch by Thomas Kluyver.
- Issue #14773: Fix os.fwalk() failing on dangling symlinks.
- Issue #12541: Be lenient with quotes around Realm field of HTTP Basic
Authentation in urllib2.
- Issue #14807: move undocumented tarfile.filemode() to stat.filemode() and add
doc entry. Add tarfile.filemode alias with deprecation warning.
- Issue #13815: TarFile.extractfile() now returns io.BufferedReader objects.
- Issue #14532: Add a secure_compare() helper to the hmac module, to mitigate
timing attacks. Patch by Jon Oberheide.
- Add importlib.util.resolve_name().
- Issue #14366: Support lzma compression in zip files.
Patch by Serhiy Storchaka.
- Issue #13959: Introduce importlib.find_loader() and document
imp.find_module/load_module as deprecated.
- Issue #14082: shutil.copy2() now copies extended attributes, if possible.
Patch by Hynek Schlawack.
- Issue #13959: Make importlib.abc.FileLoader.load_module()/get_filename() and
importlib.machinery.ExtensionFileLoader.load_module() have their single
argument be optional. Allows for the replacement (and thus deprecation) of
imp.load_source()/load_package()/load_compiled().
- Issue #13959: imp.get_suffixes() has been deprecated in favour of the new
attributes on importlib.machinery: SOURCE_SUFFIXES, DEBUG_BYTECODE_SUFFIXES,
OPTIMIZED_BYTECODE_SUFFIXES, BYTECODE_SUFFIXES, and EXTENSION_SUFFIXES. This
led to an indirect deprecation of inspect.getmoduleinfo().
- Issue #14662: Prevent shutil failures on OS X when destination does not
support chflag operations. Patch by Hynek Schlawack.
- Issue #14157: Fix time.strptime failing without a year on February 29th.
Patch by Hynek Schlawack.
- Issue #14753: Make multiprocessing's handling of negative timeouts
the same as it was in Python 3.2.
- Issue #14583: Fix importlib bug when a package's __init__.py would first
import one of its modules then raise an error.
- Issue #14741: Fix missing support for Ellipsis ('...') in parser module.
- Issue #14697: Fix missing support for set displays and set comprehensions in
parser module.
- Issue #14701: Fix missing support for 'raise ... from' in parser module.
- Add support for timeouts to the acquire() methods of
multiprocessing's lock/semaphore/condition proxies.
- Issue #13989: Add support for text mode to gzip.open().
- Issue #14127: The os.stat() result object now provides three additional
fields: st_ctime_ns, st_mtime_ns, and st_atime_ns, providing those times as an
integer with nanosecond resolution. The functions os.utime(), os.lutimes(),
and os.futimes() now accept a new parameter, ns, which accepts mtime and atime
as integers with nanosecond resolution.
- Issue #14127 and #10148: shutil.copystat now preserves exact mtime and atime
on filesystems providing nanosecond resolution.
Tools/Demos
-----------
- Issue #14695: Bring Tools/parser/unparse.py support up to date with
the Python 3.3 Grammar.
Build
-----
- Issue #14472: Update .gitignore. Patch by Matej Cepl.
- Upgrade Windows library versions: bzip 1.0.6, OpenSSL 1.0.1c.
- Issue #14693: Under non-Windows platforms, hashlib's fallback modules are
always compiled, even if OpenSSL is present at build time.
- Issue #13210: Windows build now uses VS2010, ported from VS2008.
Documentation
-------------
- Issue #14588: The language reference now accurately documents the Python 3
class definition process. Patch by Nick Coghlan.
- Issue #14943: Correct a default argument value for winreg.OpenKey
and correctly list the argument names in the function's explanation.
What's New in Python 3.3.0 Alpha 3?
===================================
*Release date: 01-May-2012*
Core and Builtins
-----------------
- Issue #14699: Fix calling the classmethod descriptor directly.
- Issue #14433: Prevent msvcrt crash in interactive prompt when stdin is closed.
- Issue #14521: Make result of float('nan') and float('-nan') more consistent
across platforms.
- Issue #14646: __import__() sets __loader__ if the loader did not.
- Issue #14605: No longer have implicit entries in sys.meta_path. If
sys.meta_path is found to be empty, raise ImportWarning.
- Issue #14605: No longer have implicit entries in sys.path_hooks. If
sys.path_hooks is found to be empty, a warning will be raised. None is now
inserted into sys.path_importer_cache if no finder was discovered. This also
means imp.NullImporter is no longer implicitly used.
- Issue #13903: Implement PEP 412. Individual dictionary instances can now share
their keys with other dictionaries. Classes take advantage of this to share
their instance dictionary keys for improved memory and performance.
- Issue #11603 (again): Setting __repr__ to __str__ now raises a RuntimeError
when repr() or str() is called on such an object.
- Issue #14658: Fix binding a special method to a builtin implementation of a
special method with a different name.
- Issue #14630: Fix a memory access bug for instances of a subclass of int
with value 0.
- Issue #14339: Speed improvements to bin, oct and hex functions. Patch by
Serhiy Storchaka.
- Issue #14098: New functions PyErr_GetExcInfo and PyErr_SetExcInfo.
Patch by Stefan Behnel.
- Issue #14385: It is now possible to use a custom type for the __builtins__
namespace, instead of a dict. It can be used for sandboxing for example.
Raise also a NameError instead of ImportError if __build_class__ name if not
found in __builtins__.
- Issue #12599: Be more strict in accepting None compared to a false-like
object for importlib.util.module_for_loader and
importlib.machinery.PathFinder.
- Issue #14612: Fix jumping around with blocks by setting f_lineno.
- Issue #14592: Attempting a relative import w/o __package__ or __name__ set in
globals raises a KeyError.
- Issue #14607: Fix keyword-only arguments which started with ``__``.
- Issue #10854: The ImportError raised when an extension module on Windows
fails to import now uses the new path and name attributes from
Issue #1559549.
- Issue #13889: Check and (if necessary) set FPU control word before calling
any of the dtoa.c string <-> float conversion functions, on MSVC builds of
Python. This fixes issues when embedding Python in a Delphi app.
- __import__() now matches PEP 328 and documentation by defaulting 'index' to 0
instead of -1 and removing support for negative values.
- Issue #2377: Make importlib the implementation of __import__().
- Issue #1559549: ImportError now has 'name' and 'path' attributes that are set
using keyword arguments to its constructor. They are currently not set by
import as they are meant for use by importlib.
- Issue #14474: Save and restore exception state in thread.start_new_thread()
while writing error message if the thread leaves a unhandled exception.
- Issue #13019: Fix potential reference leaks in bytearray.extend(). Patch
by Suman Saha.
Library
-------
- Issue #14768: os.path.expanduser('~/a') doesn't works correctly when HOME is '/'.
- Issue #14371: Support bzip2 in zipfile module. Patch by Serhiy Storchaka.
- Issue #13183: Fix pdb skipping frames after hitting a breakpoint and running
step. Patch by Xavier de Gaye.
- Issue #14696: Fix parser module to understand 'nonlocal' declarations.
- Issue #10941: Fix imaplib.Internaldate2tuple to produce correct result near
the DST transition. Patch by Joe Peterson.
- Issue #9154: Fix parser module to understand function annotations.
- Issue #6085: In http.server.py SimpleHTTPServer.address_string returns the
client ip address instead client hostname. Patch by Charles-François Natali.
- Issue #14309: Deprecate time.clock(), use time.perf_counter() or
time.process_time() instead.
- Issue #14428: Implement the PEP 418. Add time.get_clock_info(),
time.perf_counter() and time.process_time() functions, and rename
time.steady() to time.monotonic().
- Issue #14646: importlib.util.module_for_loader() now sets __loader__ and
__package__ (when possible).
- Issue #14664: It is now possible to use @unittest.skip{If,Unless} on a
test class that doesn't inherit from TestCase (i.e. a mixin).
- Issue #4892: multiprocessing Connections can now be transferred over
multiprocessing Connections. Patch by Richard Oudkerk (sbt).
- Issue #14160: TarFile.extractfile() failed to resolve symbolic links when
the links were not located in an archive subdirectory.
- Issue #14638: pydoc now treats non-string __name__ values as if they
were missing, instead of raising an error.
- Issue #13684: Fix httplib tunnel issue of infinite loops for certain sites
which send EOF without trailing \r\n.
- Issue #14605: Add importlib.abc.FileLoader, importlib.machinery.(FileFinder,
SourceFileLoader, SourcelessFileLoader, ExtensionFileLoader).
- Issue #13959: imp.cache_from_source()/source_from_cache() now follow
os.path.join()/split() semantics for path manipulation instead of its prior,
custom semantics of caring the right-most path separator forward in path
joining.
- Issue #2193: Allow ":" character in Cookie NAME values.
- Issue #14629: tokenizer.detect_encoding will specify the filename in the
SyntaxError exception if found at readline.__self__.name.
- Issue #14629: Raise SyntaxError in tokenizer.detect_encoding if the
first two lines have non-UTF-8 characters without an encoding declaration.
- Issue #14308: Fix an exception when a "dummy" thread is in the threading
module's active list after a fork().
- Issue #11750: The Windows API functions scattered in the _subprocess and
_multiprocessing.win32 modules now live in a single module "_winapi".
Patch by sbt.
- Issue #14087: multiprocessing: add Condition.wait_for(). Patch by sbt.
- Issue #14538: HTMLParser can now parse correctly start tags that contain
a bare '/'.
- Issue #14452: SysLogHandler no longer inserts a UTF-8 BOM into the message.
- Issue #14386: Expose the dict_proxy internal type as types.MappingProxyType.
- Issue #13959: Make imp.reload() always use a module's __loader__ to perform
the reload.
- Issue #13959: Add imp.py and rename the built-in module to _imp, allowing for
re-implementing parts of the module in pure Python.
- Issue #13496: Fix potential overflow in bisect.bisect algorithm when applied
to a collection of size > sys.maxsize / 2.
- Have importlib take advantage of ImportError's new 'name' and 'path'
attributes.
- Issue #14399: zipfile now recognizes that the archive has been modified even
if only the comment is changed. In addition, the TypeError that results from
trying to set a non-binary value as a comment is now raised at the time
the comment is set rather than at the time the zipfile is written.
- trace.CoverageResults.is_ignored_filename() now ignores any name that starts
with "<" and ends with ">" instead of special-casing "