Prevent to detect unwanted builtin openssl.
After bump of BUILDLINK_API_DEPENDS.openssl to 1.0.1c,
buitin openssl is not acceptable for various platforms.
or rename of module name affect to other parts using this module.
Noticed by Benjamin Lorenz in tech-pkg@.
Additionally, set ALTERNATIVE for bin/smtpd.py.
This is a 2.7 series bugfix release. Most importantly, it resolves an issue that caused the interactive prompt to crash on OS X 10.9. It also includes numerous bugfixes over 2.7.5.
injecting -Wall -Wno-error does not cause the test to produce the wrong
answer. (If it does, the wrong information is installed in /usr/include,
and ultimately provokes PR 47342.)
* An ordered dictionary type
* New unittest features including test skipping, new assert methods, and test
discovery
* A much faster io module
* Automatic numbering of fields in the str.format() method
* Float repr improvements backported from 3.x
* Tile support for Tkinter
* A backport of the memoryview object from 3.x
* Set literals
* Set and dictionary comprehensions
* Dictionary views
* New syntax for nested with statements
* The sysconfig module
What's New in Python 2.7.2?
===========================
*Release date: 2011-06-11*
Library
-------
- Issue #12009: Fixed regression in netrc file comment handling.
Extension Modules
-----------------
- Issue #1221: Make pyexpat.__version__ equal to the Python version.
What's New in Python 2.7.2 release candidate 1?
===============================================
*Release date: 2011-05-29*
Core and Builtins
-----------------
- Issue #9670: Increase the default stack size for secondary threads on
Mac OS X and FreeBSD to reduce the chances of a crash instead of a
"maximum recursion depth" RuntimeError exception.
(patch by Ronald Oussoren)
- Correct lookup of __dir__ on objects. This allows old-style classes to have
__dir__. It also causes errors besides AttributeError found on lookup to be
propagated.
- Issue #1195: Fix input() if it is interrupted by CTRL+d and then CTRL+c,
clear the end-of-file indicator after CTRL+d.
- Issue #8651: PyArg_Parse*() functions raise an OverflowError if the file
doesn't have PY_SSIZE_T_CLEAN define and the size doesn't fit in an int
(length bigger than 2^31-1 bytes).
- Issue #8651: Fix "z#" format of PyArg_Parse*() function: the size was not
written if PY_SSIZE_T_CLEAN is defined.
- Issue #9756: When calling a method descriptor or a slot wrapper descriptor,
the check of the object type doesn't read the __class__ attribute anymore.
Fix a crash if a class override its __class__ attribute (e.g. a proxy of the
str type). Patch written by Andreas Stührk.
- Issue #10517: After fork(), reinitialize the TLS used by the PyGILState_*
APIs, to avoid a crash with the pthread implementation in RHEL 5. Patch
by Charles-François Natali.
- Issue #6780: fix starts/endswith error message to mention that tuples are
accepted too.
- Issue #5057: fix a bug in the peepholer that led to non-portable pyc files
between narrow and wide builds while optimizing BINARY_SUBSCR on non-BMP
chars (e.g. u"\U00012345"[0]).
- Issue #11650: PyOS_StdioReadline() retries fgets() if it was interrupted
(EINTR), for example if the program is stopped with CTRL+z on Mac OS X. Patch
written by Charles-Francois Natali.
- Issue #11144: Ensure that int(a_float) returns an int whenever possible.
Previously, there were some corner cases where a long was returned even
though the result was within the range of an int.
- Issue #11450: Don't truncate hg version info in Py_GetBuildInfo() when
there are many tags (e.g. when using mq). Patch by Nadeem Vawda.
- Issue #10451: memoryview objects could allow to mutate a readable buffer.
Initial patch by Ross Lagerwall.
- Issue #10892: Don't segfault when trying to delete __abstractmethods__ from a
class.
- Issue #8020: Avoid a crash where the small objects allocator would read
non-Python managed memory while it is being modified by another thread.
Patch by Matt Bandy.
- Issue #11004: Repaired edge case in deque.count().
- Issue #8278: On Windows and with a NTFS filesystem, os.stat() and os.utime()
can now handle dates after 2038.
- Issue #4236: Py_InitModule4 now checks the import machinery directly
rather than the Py_IsInitialized flag, avoiding a Fatal Python
error in certain circumstances when an import is done in __del__.
- issue #11828: startswith and endswith don't accept None as slice index.
Patch by Torsten Becker.
- Issue #10674: Remove unused 'dictmaker' rule from grammar.
- Issue #10596: Fix float.__mod__ to have the same behaviour as
float.__divmod__ with respect to signed zeros. -4.0 % 4.0 should be
0.0, not -0.0.
- Issue #11386: bytearray.pop() now throws IndexError when the bytearray is
empty, instead of OverflowError.
Library
-------
- Issue #12161: Cause StringIO.getvalue() to raise a ValueError when used on a
closed StringIO instance.
- Issue #12182: Fix pydoc.HTMLDoc.multicolumn() if Python uses the new (true)
division (python -Qnew). Patch written by Ralf W. Grosse-Kunstleve.
- Issue #12175: RawIOBase.readall() now returns None if read() returns None.
- Issue #12175: FileIO.readall() now raises a ValueError instead of an IOError
if the file is closed.
- Issue #1441530: In imaplib, use makefile() to wrap the SSL socket to avoid
heap fragmentation and MemoryError with some malloc implementations.
- Issue #12100: Don't reset incremental encoders of CJK codecs at each call to
their encode() method anymore, but continue to call the reset() method if the
final argument is True.
- Issue #12124: zipimport doesn't keep a reference to zlib.decompress() anymore
to be able to unload the module.
- Issue #11088: don't crash when using F5 to run a script in IDLE on MacOSX
with Tk 8.5.
- Issue #10154, #10090: change the normalization of UTF-8 to "UTF-8" instead
of "UTF8" in the locale module as the latter is not supported MacOSX and OpenBSD.
- Issue #9516: avoid errors in sysconfig when MACOSX_DEPLOYMENT_TARGET is
set in shell.
- Issue #12050: zlib.decompressobj().decompress() now clears the unconsumed_tail
attribute when called without a max_length argument.
- Issue #12062: In the `io` module, fix a flushing bug when doing a certain
type of I/O sequence on a file opened in read+write mode (namely: reading,
seeking a bit forward, writing, then seeking before the previous write but
still within buffered data, and writing again).
- Issue #8498: In socket.accept(), allow to specify 0 as a backlog value in
order to accept exactly one connection. Patch by Daniel Evers.
- Issue #12012: ssl.PROTOCOL_SSLv2 becomes optional.
- Issue #11164: Remove obsolete allnodes test from minidom test.
- Issue #11927: SMTP_SSL now uses port 465 by default as documented. Patch
by Kasun Herath.
- Issue 11999: fixed sporadic sync failure mailbox.Maildir due to its trying to
detect mtime changes by comparing to the system clock instead of to the
previous value of the mtime.
- Issue #10684: shutil.move used to delete a folder on case insensitive
filesystems when the source and destination name where the same except
for the case.
- Issue #11982: fix json.loads('""') to return u'' rather than ''.
- Issue #11277: mmap.mmap() calls fcntl(fd, F_FULLFSYNC) on Mac OS X to get
around a mmap bug with sparse files. Patch written by Steffen Daode Nurpmeso.
- Issue #10761: Fix tarfile.extractall failure when symlinked files are
present. Initial patch by Scott Leerssen.
- Issue #11763: don't use difflib in TestCase.assertMultiLineEqual if the
strings are too long.
- Issue #11236: getpass.getpass responds to ctrl-c or ctrl-z on terminal.
- Issue #11768: The signal handler of the signal module only calls
Py_AddPendingCall() for the first signal to fix a deadlock on reentrant or
parallel calls. PyErr_SetInterrupt() writes also into the wake up file.
- Issue #11875: collections.OrderedDict's __reduce__ was temporarily
mutating the object instead of just working on a copy.
- Issue #11442: Add a charset parameter to the Content-type in SimpleHTTPServer
to avoid XSS attacks.
- Issue #11467: Fix urlparse behavior when handling urls which contains scheme
specific part only digits. Patch by Santoso Wijaya.
- collections.Counter().copy() now works correctly for subclasses.
- Issue #11474: Fix the bug with url2pathname() handling of '/C|/' on Windows.
Patch by Santoso Wijaya.
- Issue #9233: Fix json.loads('{}') to return a dict (instead of a list), when
_json is not available.
- Issue #11703: urllib2.geturl() does not return correct url when the original
url contains #fragment.
- Issue #10019: Fixed regression in json module where an indent of 0 stopped
adding newlines and acted instead like 'None'.
- Issue #5162: Treat services like frozen executables to allow child spawning
from multiprocessing.forking on Windows.
- Issue #4877: Fix a segfault in xml.parsers.expat while attempting to parse
a closed file.
- Issue #11830: Remove unnecessary introspection code in the decimal module.
It was causing a failed import in the Turkish locale where the locale
sensitive str.upper() method caused a name mismatch.
- Issue #8428: Fix a race condition in multiprocessing.Pool when terminating
worker processes: new processes would be spawned while the pool is being
shut down. Patch by Charles-François Natali.
- Issue #7311: Fix HTMLParser to accept non-ASCII attribute values.
- Issue #10963: Ensure that subprocess.communicate() never raises EPIPE.
- Issue #11662: Make urllib and urllib2 ignore redirections if the
scheme is not HTTP, HTTPS or FTP (CVE-2011-1521).
- Issue #11256: Fix inspect.getcallargs on functions that take only keyword
arguments.
- Issue #11696: Fix ID generation in msilib.
- Issue #9696: Fix exception incorrectly raised by xdrlib.Packer.pack_int when
trying to pack a negative (in-range) integer.
- Issue #11675: multiprocessing.[Raw]Array objects created from an integer size
are now zeroed on creation. This matches the behaviour specified by the
documentation.
- Issue #7639: Fix short file name generation in bdist_msi.
- Issue #11666: let help() display named tuple attributes and methods
that start with a leading underscore.
- Issue #11673: Fix multiprocessing Array and RawArray constructors to accept a
size of type 'long', rather than only accepting 'int'.
- Issue #10042: Fixed the total_ordering decorator to handle cross-type
comparisons that could lead to infinite recursion.
- Issue #10979: unittest stdout buffering now works with class and module
setup and teardown.
- Issue #11569: use absolute path to the sysctl command in multiprocessing to
ensure that it will be found regardless of the shell PATH. This ensures
that multiprocessing.cpu_count works on default installs of MacOSX.
- Issue #11500: Fixed a bug in the os x proxy bypass code for fully qualified
IP addresses in the proxy exception list.
- Issue #11131: Fix sign of zero in plus and minus operations when
the context rounding mode is ROUND_FLOOR.
- Issue #5622: Fix curses.wrapper to raise correct exception if curses
initialization fails.
- Issue #11391: Writing to a mmap object created with
``mmap.PROT_READ|mmap.PROT_EXEC`` would segfault instead of raising a
TypeError. Patch by Charles-François Natali.
- Issue #11306: mailbox in certain cases adapts to an inability to open
certain files in read-write mode. Previously it detected this by
checking for EACCES, now it also checks for EROFS.
- Issue #11265: asyncore now correctly handles EPIPE, EBADF and EAGAIN errors
on accept(), send() and recv().
- Issue #11326: Add the missing connect_ex() implementation for SSL sockets,
and make it work for non-blocking connects.
- Issue #10956: Buffered I/O classes retry reading or writing after a signal
has arrived and the handler returned successfully.
- Issue #10680: Fix mutually exclusive arguments for argument groups in
argparse.
- Issue #4681: Allow mmap() to work on file sizes and offsets larger than
4GB, even on 32-bit builds. Initial patch by Ross Lagerwall, adapted for
32-bit Windows.
- Issue #10360: In WeakSet, do not raise TypeErrors when testing for
membership of non-weakrefable objects.
- Issue #10549: Fix pydoc traceback when text-documenting certain classes.
- Issue #940286: pydoc.Helper.help() ignores input/output init parameters.
- Issue #11171: Fix detection of config/Makefile when --prefix !=
--exec-prefix, which caused Python to not start.
- Issue #11116: any error during addition of a message to a mailbox now causes
a rollback, instead of leaving the mailbox partially modified.
- Issue #8275: Fix passing of callback arguments with ctypes under Win64.
Patch by Stan Mihai.
- Issue #10940: Workaround an IDLE hang on Mac OS X 10.6 when using the
menu accelerators for Open Module, Go to Line, and New Indent Width.
The accelerators still work but no longer appear in the menu items.
- Issue #10907: Warn OS X 10.6 IDLE users to use ActiveState Tcl/Tk 8.5, rather
than the currently problematic Apple-supplied one, when running with the
64-/32-bit installer variant.
- Issue #11052: Correct IDLE menu accelerators on Mac OS X for Save
commands.
- Issue #10949: Improved robustness of rotating file handlers.
- Issue #10955: Fix a potential crash when trying to mmap() a file past its
length. Initial patch by Ross Lagerwall.
- Issue #10898: Allow compiling the posix module when the C library defines
a symbol named FSTAT.
- Issue #6075: IDLE on Mac OS X now works with both Carbon AquaTk and
Cocoa AquaTk.
- Issue #10916: mmap should not segfault when a file is mapped using 0 as
length and a non-zero offset, and an attempt to read past the end of file
is made (IndexError is raised instead). Patch by Ross Lagerwall.
- Issue #10875: Update Regular Expression HOWTO; patch by 'SilentGhost'.
- Issue #10827: Changed the rules for 2-digit years. The time.asctime
function will now format any year when ``time.accept2dyear`` is
false and will accept years >= 1000 otherwise. The year range
accepted by ``time.mktime`` and ``time.strftime`` is still system
dependent, but ``time.mktime`` will now accept full range supported
by the OS. Conversion of 2-digit years to 4-digit is deprecated.
- Issue #10869: Fixed bug where ast.increment_lineno modified the root
node twice.
- Issue #7858: Raise an error properly when os.utime() fails under Windows
on an existing file.
- Issue #3839: wsgiref should not override a Content-Length header set by
the application. Initial patch by Clovis Fabricio.
- Issue #10806, issue #9905: Fix subprocess pipes when some of the standard
file descriptors (0, 1, 2) are closed in the parent process. Initial
patch by Ross Lagerwall.
- Issue #4662: os.tempnam(), os.tmpfile() and os.tmpnam() now raise a py3k
DeprecationWarning.
- Subclasses of collections.OrderedDict now work correctly with __missing__.
- Issue #10753 - Characters ';', '=' and ',' in the PATH_INFO environment
variable won't be quoted when the URI is constructed by the wsgiref.util 's
request_uri method. According to RFC 3986, these characters can be a part of
params in PATH component of URI and need not be quoted.
- Issue #10738: Fix webbrowser.Opera.raise_opts
- Issue #9824: SimpleCookie now encodes , and ; in values to cater to how
browsers actually parse cookies.
- Issue #1379416: eliminated a source of accidental unicode promotion in
email.header.Header.encode.
- Issue #5258/#10642: if site.py encounters a .pth file that generates an error,
it now prints the filename, line number, and traceback to stderr and skips
the rest of that individual file, instead of stopping processing entirely.
- Issue #10750: The ``raw`` attribute of buffered IO objects is now read-only.
- Issue #10242: unittest.TestCase.assertItemsEqual makes too many assumptions
about input.
- Issue #10611: SystemExit should not cause a unittest test run to exit.
- Issue #6791: Limit header line length (to 65535 bytes) in http.client,
to avoid denial of services from the other party.
- Issue #10404: Use ctl-button-1 on OSX for the context menu in Idle.
- Issue #9907: Fix tab handling on OSX when using editline by calling
rl_initialize first, then setting our custom defaults, then reading .editrc.
- Issue #4188: Avoid creating dummy thread objects when logging operations
from the threading module (with the internal verbose flag activated).
- Issue #9721: Fix the behavior of urljoin when the relative url starts with a
';' character. Patch by Wes Chow.
- Issue #10714: Limit length of incoming request in http.server to 65536 bytes
for security reasons. Initial patch by Ross Lagerwall.
- Issue #9558: Fix distutils.command.build_ext with VS 8.0.
- Issue #10695: passing the port as a string value to telnetlib no longer
causes debug mode to fail.
- Issue #10107: Warn about unsaved files in IDLE on OSX.
- Issue #10406: Enable Rstrip IDLE extension on OSX (just like on other
platforms).
- Issue #10478: Reentrant calls inside buffered IO objects (for example by
way of a signal handler) now raise a RuntimeError instead of freezing the
current process.
- Issue #10497: Fix incorrect use of gettext in argparse.
- Issue #10464: netrc now correctly handles lines with embedded '#' characters.
- Issue #1731717: Fixed the problem where subprocess.wait() could cause an
OSError exception when The OS had been told to ignore SIGCLD in our process
or otherwise not wait for exiting child processes.
- Issue #9509: argparse now properly handles IOErrors raised by
argparse.FileType.
- Issue #9348: Raise an early error if argparse nargs and metavar don't match.
- Issue #8982: Improve the documentation for the argparse Namespace object.
- Issue #9343: Document that argparse parent parsers must be configured before
their children.
- Issue #9026: Fix order of argparse sub-commands in help messages.
- Issue #9347: Fix formatting for tuples in argparse type= error messages.
Extension Modules
-----------------
- Stop using the old interface for providing methods and attributes in the _sre
module. Among other things, this gives these classes ``__class__``
attributes. (See #12099)
- Issue #10169: Fix argument parsing in socket.sendto() to avoid error masking.
- Issue #12051: Fix segfault in json.dumps() while encoding highly-nested
objects using the C accelerations.
- Issue #12017: Fix segfault in json.loads() while decoding highly-nested
objects using the C accelerations.
- Issue #1838: Prevent segfault in ctypes, when _as_parameter_ on a class is set
to an instance of the class.
- Issue #678250: Make mmap flush a noop on ACCESS_READ and ACCESS_COPY.
Build
-----
- Issue #11217: For 64-bit/32-bit Mac OS X universal framework builds,
ensure "make install" creates symlinks in --prefix bin for the "-32"
files in the framework bin directory like the installer does.
- Issue #11411: Fix 'make DESTDIR=' with a relative destination.
- Issue #10709: Add updated AIX notes in Misc/README.AIX.
- Issue #11184: Fix large-file support on AIX.
- Issue #941346: Fix broken shared library build on AIX.
- Issue #11268: Prevent Mac OS X Installer failure if Documentation
package had previously been installed.
- Issue #11079: The /Applications/Python x.x folder created by the Mac
OS X installers now includes a link to the installed documentation.
- Issue #11054: Allow Mac OS X installer builds to again work on 10.5 with
the system-provided Python.
- Issue #10843: Update third-party library versions used in OS X
32-bit installer builds: bzip2 1.0.6, readline 6.1.2, SQLite 3.7.4
(with FTS3/FTS4 and RTREE enabled), and ncursesw 5.5 (wide-char
support enabled).
- Don't run pgen twice when using make -j.
- Issue #7716: Under Solaris, don't assume existence of /usr/xpg4/bin/grep in
the configure script but use $GREP instead. Patch by Fabian Groffen.
- Issue #10475: Don't hardcode compilers for LDSHARED/LDCXXSHARED on NetBSD
and DragonFly BSD. Patch by Nicolas Joly.
- Issue #10655: Fix the build on PowerPC on Linux with GCC when building with
timestamp profiling (--with-tsc): the preprocessor test for the PowerPC
support now looks for "__powerpc__" as well as "__ppc__": the latter seems to
only be present on OS X; the former is the correct one for Linux with GCC.
- Issue #1099: Fix the build on MacOSX when building a framework with pydebug
using GCC 4.0.
IDLE
----
- Issue #11718: IDLE's open module dialog couldn't find the __init__.py
file in a package.
Tests
-----
- Issue #12205: Fix test_subprocess failure due to uninstalled test data.
- Issue #5723: Improve json tests to be executed with and without accelerations.
- Issue #11910: Fix test_heapq to skip the C tests when _heapq is missing.
- Fix test_startfile to wait for child process to terminate before finishing.
- Issue #11719: Fix message about unexpected test_msilib skip on non-Windows
platforms. Patch by Nadeem Vawda.
- Issue #7108: Fix test_commands to not fail when special attributes ('@'
or '.') appear in 'ls -l' output.
- Issue #11490: test_subprocess:test_leaking_fds_on_error no longer gives a
false positive if the last directory in the path is inaccessible.
- Issue #10822: Fix test_posix:test_getgroups failure under Solaris. Patch
by Ross Lagerwall.
- Issue #6293: Have regrtest.py echo back sys.flags. This is done by default
in whole runs and enabled selectively using ``--header`` when running an
explicit list of tests. Original patch by Collin Winter.
- Issue #775964: test_grp now skips YP/NIS entries instead of failing when
encountering them.
- Issue #7110: regrtest now sends test failure reports and single-failure
tracebacks to stderr rather than stdout.
Python 2.7 is intended to be the last major release in the 2.x series.
The Python maintainers are planning to focus their future efforts on
the Python 3.x series.
This means that 2.7 will remain in place for a long time, running
production systems that have not been ported to Python 3.x.
Two consequences of the long-term significance of 2.7 are:
* It's very likely the 2.7 release will have a longer period of
maintenance compared to earlier 2.x versions. Python 2.7 will
continue to be maintained while the transition to 3.x continues, and
the developers are planning to support Python 2.7 with bug-fix
releases beyond the typical two years.
* A policy decision was made to silence warnings only of interest to
developers. :exc:`DeprecationWarning` and its
descendants are now ignored unless otherwise requested, preventing
users from seeing warnings triggered by an application. This change
was also made in the branch that will become Python 3.2. (Discussed
on stdlib-sig and carried out in :issue:`7319`.)
In previous releases, :exc:`DeprecationWarning` messages were
enabled by default, providing Python developers with a clear
indication of where their code may break in a future major version
of Python.
However, there are increasingly many users of Python-based
applications who are not directly involved in the development of
those applications. :exc:`DeprecationWarning` messages are
irrelevant to such users, making them worry about an application
that's actually working correctly and burdening application developers
with responding to these concerns.
You can re-enable display of :exc:`DeprecationWarning` messages by
running Python with the :option:`-Wdefault <-W>` (short form:
:option:`-Wd <-W>`) switch, or by setting the :envvar:`PYTHONWARNINGS`
environment variable to ``"default"`` (or ``"d"``) before running
Python. Python code can also re-enable them
by calling ``warnings.simplefilter('default')``.
