What's New in Python 2.6.4 final?
=================================
*Release date: 25-Oct-2009*
(nothing)
What's New in Python 2.6.4rc2?
==============================
*Release date: 18-Oct-2009*
Library
-------
- Issue #7115: Fixed the extension module builds that is failing when using
paths in the extension name instead of dotted names.
- Issue #7120: logging: Removed import of multiprocessing which is causing
crash in GAE.
- Issue #7149: fix exception in urllib when detecting proxy settings on OSX.
What's New in Python 2.6.4rc1?
==============================
*Release date: 07-Oct-2009*
Core and Builtins
-----------------
- Issue #7019: Raise ValueError when unmarshalling bad long data, instead
of producing internally inconsistent Python longs.
Library
-------
- Issue #7068: Fixed the partial renaming that occured in r72594.
- Issue #7064: Fixed the incompatibility with Setuptools in distutils
when running the build_ext command.
- Issue #7052: Removed nonexisting NullHandler from logging.__all__.
- Issue #7039: Fixed distutils.tests.test_sysconfig when running on
installation with no build.
Tests
-----
- Issue #7042: Fix test_signal (test_itimer_virtual) failure on OS X 10.6.
Pkgsrc changes:
- Update base pear packages:Archive_Tar-1.3.3, Console_Getopt-1.2.3
Structures_Graph-1.0.3
- Update ${FILESDIR}/go-pear.php from latest version at
http://pear.php.net/go-pear
- Update ${FILESDIR}/pear.sh from the bundled version
- Provide bootstrap PEAR.php via ${FILESDIR}/PEAR.php from
http://cvs.php.net/viewvc.cgi/pear-core/PEAR.php?view=co&pathrev=PEAR_1_4
(see FILESDIR/go-pear.php's code) and put it into ${WRKSRC} at the extract
phase
- Update PLIST
Upstream changes:
A lot! IMHO more than what can be recorded here.
Please see individual releases Changelog since version 1.5.4
at http://pear.php.net/package/PEAR.
Pict is a language in the ML tradition, formed by adding a layer of
convenient syntactic sugar and a static type system to a tiny
core. The current release includes a Pict-to-C compiler, reference
manual, language tutorial, numerous libraries, and example programs.
The core language - an asynchronous variant of Milner, Parrow, and
Walker's pi-calculus - has been used as a theoretical foundation for a
broad class of concurrent computations. The goal in Pict is to
identify high-level idioms that arise naturally when these primitives
are used to build working programs - idioms such as basic data
structures, protocols for returning results, higher-order programming,
selective communication, and concurrent objects. The type system
integrates a number of features found in recent work on theoretical
foundations for typed object-oriented languages: higher-order
polymorphism, simple recursive types, subtyping, and a powerful
partial type inference algorithm.
Pkgsrc changes:
o Enable shared libraries for NetBSD, and deal with the resulting
fallout: tests which are written in C need to point their run-path
to the build version of the shared libraries.
o The default is shared libs with .so, except for on Darwin, where
shared libs are named differently. The Darwin part is untested.
(Korean) fonts, and add (commented) dependencies on them. Ok tnn@.
Bump PKGREVISION.
XXX uming/ukai (Chinese) are not in pkgsrc, anyone can package them, or suggest
replacements?
What's New in Python 2.6.3
==========================
*Release date: 02-Oct-2009*
What's New in Python 2.6.3rc1
=============================
*Release date: 29-Sep-2009*
Core and Builtins
-----------------
- Issue #5329: Fix os.popen* regression from 2.5 with commands as a
sequence running through the shell. Patch by Jean-Paul Calderone
and Jani Hakala.
- Issue #6990: Fix threading.local subclasses leaving old state around
after a reference cycle GC which could be recycled by new locals.
- Issue #6922: Fix an infinite loop when trying to decode an invalid
UTF-32 stream with a non-raising error handler like "replace" or "ignore".
- Issue #1590864: Fix potential deadlock when mixing threads and fork().
- Issue #6844: Do not emit DeprecationWarnings when accessing a "message"
attribute on exceptions that was set explicitly.
- Issue #6846: Fix bug where bytearray.pop() returns negative integers.
- Issue #6707: dir() on an uninitialized module caused a crash.
- Issue #6540: Fixed crash for bytearray.translate() with invalid parameters.
- Issue #6573: set.union() stopped processing inputs if an instance of self
occurred in the argument chain.
- Issue #6070: On posix platforms import no longer copies the execute bit
from the .py file to the .pyc file if it is set.
- Issue #4547: When debugging a very large function, it was not always
possible to update the lineno attribute of the current frame.
- Issue #4618: When unicode arguments are passed to print(), the default
separator and end should be unicode also.
- Issue #6119: Fixed a incorrect Py3k warning about order comparisons of
builtin functions and methods.
- Issue #5330: C functions called with keyword arguments were not reported by
the various profiling modules (profile, cProfile). Patch by Hagen Fürstenau.
- Issue #6089: str.format can raise SystemError with certain invalid
field specifiers.
- Issue #5994: the marshal module now has docstrings.
- Issue #5981: Fix two minor inf/nan issues in float.fromhex: (1) inf
and nan strings with trailing whitespace were incorrectly rejected
and (2) the interpretation of fromhex('-nan') didn't match that of
float('-nan').
- Issue #5890: in subclasses of 'property' the __doc__ attribute was
shadowed by classtype's, even if it was None. property now
inserts the __doc__ into the subclass instance __dict__.
- Issue #5724: (See also issue #4575.) Fix Py_IS_INFINITY macro to
work correctly on x87 FPUs: it now forces its argument to double
before testing for infinity.
- Issue #4971: Fix titlecase for characters that are their own
titlecase, but not their own uppercase.
- Issue #5829: complex('1e-500') no longer raises an exception
- Issue #5787: object.__getattribute__(some_type, "__bases__") segfaulted on
some builtin types.
- Issue #5283: Setting __class__ in __del__ caused a segfault.
- Issue #5759: float() didn't call __float__ on str subclasses.
- Issue #5108: Handle %s like %S and %R in PyUnicode_FromFormatV(): Call
PyUnicode_DecodeUTF8() once, remember the result and output it in a second
step. This avoids problems with counting UTF-8 bytes that ignores the effect
of using the replace error handler in PyUnicode_DecodeUTF8().
Library
-------
- Issue #6790: Make it possible again to pass an `array.array` to
`httplib.HTTPConnection.send`. Patch by Kirk McDonald.
- Issue #6236, #6348: Fix various failures in the `io` module under AIX
and other platforms, when using a non-gcc compiler. Patch by egreen.
- Issue #6851: Fix urllib.urlopen crash on secondairy threads on OSX 10.6
- Issue #6947: Fix distutils test on windows. Patch by Hirokazu Yamamoto.
- Issue #4606: Passing 'None' if ctypes argtype is set to POINTER(...)
does now always result in NULL.
- Issue #5042: ctypes Structure sub-subclass does now initialize
correctly with base class positional arguments.
- Issue #6938: Fix a TypeError in string formatting of a multiprocessing
debug message.
- Issue #6635: Fix profiler printing usage message.
- Issue #6795: int(Decimal('nan')) now raises ValueError instead of
returning NaN or raising InvalidContext. Also, fix infinite recursion
in long(Decimal('nan')).
- Issue #6850: Fix bug in Decimal._parse_format_specifier for formats
with no type specifier.
- Issue #4937: plat-mac/bundlebuilder revers to non-existing version.plist
- Issue #6838: Use a list to accumulate the value instead of
repeatedly concatenating strings in httplib's
HTTPResponse._read_chunked providing a significant speed increase
when downloading large files servend with a Transfer-Encoding of 'chunked'.
- Issue #6794: Fix Decimal.compare_total and Decimal.compare_total_mag: NaN
payloads are now ordered by integer value rather than lexicographically.
- Issue #6117: Fix O(n**2) performance degradation when outputting lots of
small data on a buffered socket.socket.makefile() object.
- Issue #6637: defaultdict.copy() did not work when the default factory
was left unspecified. Also, the eval/repr round-trip would fail when
the default_factory was None.
- Issue #1424152: Fix for httplib, urllib2 to support SSL while working through
proxy. Original patch by Christopher Li, changes made by Senthil Kumaran.
- Issues #5155, 5313, 5331: multiprocessing.Process._bootstrap was
unconditionally calling "os.close(sys.stdin.fileno())" resulting in file
descriptor errors
- Issue #6415: Fixed warnings.warn sagfault on bad formatted string.
- Issue #6344: Fixed a crash of mmap.read() when passed a negative argument.
- Issue #5230: pydoc would report no documentation found if a module generated
a 'not found' import error when loaded; it now reports the import errors.
Thanks to Lucas Prado Melo for initial fix and collaboration on the tests.
- Issue #6274: Fixed possible file descriptors leak in subprocess.py
- Issue #6271: mmap tried to close invalid file handle (-1) when annonymous.
(On Unix)
- Issue #6258: Support AMD64 in bdist_msi.
- Issue #5262: Fixed bug in next rollover time computation in
TimedRotatingFileHandler.
- Issue #6121: pydoc now ignores leading and trailing spaces in the
argument to the 'help' function.
- Issue #6050: Don't fail extracting a directory from a zipfile if
the directory already exists.
- collections.namedtuple() was not working with the following field
names: cls, self, tuple, itemgetter, and property.
- Issue #1309352: fcntl now converts its third arguments to a C `long` rather
than an int, which makes some operations possible under 64-bit Linux (e.g.
DN_MULTISHOT with F_NOTIFY).
- Issue #1983: Fix functions taking or returning a process identifier to use
the dedicated C type ``pid_t`` instead of a C ``int``. Some platforms have
a process identifier type wider than the standard C integer type.
- Issue #4066: smtplib.SMTP_SSL._get_socket now correctly returns the socket.
Patch by Farhan Ahmad, test by Marcin Bachry.
- Issue #1655: Make imaplib IPv6-capable. Patch by Derek Morr.
- Issue #1664: Make nntplib IPv6-capable. Patch by Derek Morr.
- Issue #5971: StreamHandler.handleError now swallows IOErrors which occur when
trying to print a traceback.
- Issue 5955: aifc's close method did not close the file it wrapped,
now it does. This also means getfp method now returns the real fp.
- Issue #4875: On win32, ctypes.util.find_library does no longer
return directories.
- Issue #5692: In :class:`zipfile.Zipfile`, fix wrong path calculation when
extracting a file to the root directory.
- Issue #2245: aifc now skips chunk types it doesn't recognize, per spec.
- Issue #4305: ctypes should now build again on mipsel-linux-gnu
- Issue #5853: calling a function of the mimetypes module from several threads
at once could hit the recursion limit if the mimetypes database hadn't been
initialized before.
- Issue #5041: ctypes does now allow pickling wide character.
- Issue #5768: Fixed bug in Unicode output logic and test case for same.
- Issue #1161031: fix readwrite select flag handling: POLLPRI now
results in a handle_expt_event call, not handle_read_event, and POLLERR
and POLLNVAL now call handle_close, not handle_expt_event. Also,
dispatcher now has an 'ignore_log_types' attribute for suppressing
log messages, which is set to 'warning' by default.
- Issue #5828 (Invalid behavior of unicode.lower): Fixed bogus logic in
makeunicodedata.py and regenerated the Unicode database (This fixes
u'\u1d79'.lower() == '\x00').
- Issue #1202: zipfile module would cause a DeprecationWarning when storing
files with a CRC32 > 2**31-1.
- Issue #6163: Fixed HP-UX runtime library dir options in
distutils.unixcompiler. Initial patch by Sridhar Ratnakumar and
Michael Haubenwallner.
- Issue #4660: If a multiprocessing.JoinableQueue.put() was preempted, it was
possible to get a spurious 'task_done() called too many times' error.
- Issue #6595: The Decimal constructor now allows arbitrary Unicode
decimal digits in input, as recommended by the standard. Previously
it was restricted to accepting [0-9].
- Issue #6553: Fixed a crash in cPickle.load(), when given a file-like object
containing incomplete data.
- Issue #2622: Fixed an ImportError when importing email.messsage from a
standalone application built with py2exe or py2app.
- Issue #6455: Fixed test_build_ext under win32.
- Issue #6403: Fixed package path usage in build_ext.
- Issue #6287: Added the license field in Distutils documentation.
- Issue #6263: Fixed syntax error in distutils.cygwincompiler.
- Issue #5201: distutils.sysconfig.parse_makefile() now understands `$$`
in Makefiles. This prevents compile errors when using syntax like:
`LDFLAGS='-rpath=\$$LIB:/some/other/path'`. Patch by Floris Bruynooghe.
- Issue #6062: In distutils, fixed the package option of build_ext. Feedback
and tests on pywin32 by Tim Golden.
- Issue #1309567: Fix linecache behavior of stripping subdirectories when
looking for files given by a relative filename.
- Issue #6046: Fixed the library extension when distutils build_ext is used
inplace. Initial patch by Roumen Petrov.
- Issue #6022: a test file was created in the current working directory by
test_get_outputs in Distutils.
- Issue #5977: distutils build_ext.get_outputs was not taking into account the
inplace option. Initial patch by kxroberto.
- Issue #5984: distutils.command.build_ext.check_extensions_list checks were
broken for old-style extensions.
- Issue #5854: Updated __all__ to include some missing names and remove some
names which should not be exported.
- Issue #5810: Fixed Distutils test_build_scripts so it uses
sysconfig.get_config_vars.
- Issue #6865: Fix reference counting issue in the initialization of the pwd
module.
Extension Modules
-----------------
- Issue #6944: Fix a SystemError when socket.getnameinfo() was called
with something other than a tuple as first argument.
- Issue #6848: Fix curses module build failure on OS X 10.6.
- Fix expat to not segfault with specially crafted input.
- Issue #4873: Fix resource leaks in error cases of pwd and grp.
Build
-----
- Issue #6980: Fix ctypes build failure on armel-linux-gnueabi with
-mfloat-abi=softfp.
- Issue #6802: Fix build issues on MacOSX 10.6
- Issue 5390: Add uninstall icon independent of whether file
extensions are installed.
- Issue #6094: Build correctly with Subversion 1.7.
- Issue #5726: Make Modules/ld_so_aix return the actual exit code of the
linker, rather than always exit successfully. Patch by Floris Bruynooghe.
- Issue 5809: Specifying both --enable-framework and --enable-shared is
an error. Configure now explicity tells you about this.
Documentation
-------------
- Issue #6556: Fixed the Distutils configuration files location explanation
for Windows.
- Issue #6801 : symmetric_difference_update also accepts |.
Thanks to Carl Chenet.
Tests
-----
- Issue #5837: Certain sequences of calls to set() and unset() for
support.EnvironmentVarGuard objects restored the environment variables
incorrectly on __exit__.
This builds, and will compile and run helloworld, but hasn't been
particularly extensively tested.
No version bump since no prior x86_64 packages should have been created.
Added scheme/unsafe/ops
Added print-syntax-width
Added file-or-directory-identity
Inside: embedding applications should call scheme_seal_parameters
after initializing parameter values (currently used by Planet)
ECL 9.10.2:
===========
* Bugs fixed:
- Fixed typo in src/c/unixint.d that affected single-threaded builds
- The GMP library did not build in OS X Snow Leopard in 64-bits mode.
- The package MP is needed also in single-threaded versions (for fake
mp:with-lock, which is used in CLX).
- In CLX, there were a couple of typos in the code related to locks and ECL.
These typos only revealed in multithreaded builds of the CLX library.
- In Linux there is a problem with handlers for SIGFPE being totally ignored
by the system. The problem seems to be solved by avoiding the use of
feenableexcept() and restricting to C99 exception tests. That is bad because
we can not reliably and cheaply detect underflow exceptions.
- Under OS X, --enable-rpath works again. It was broken for about a year
due to my misunderstanding of how -install_name works and the differences
between that and -rpath.
ECL 9.10.1:
===========
* Configuration flags:
- The option --enable-slow-config is now always on in Solaris and it does not
take more time than an ordinary build using --with-system-gmp=no
- A new compiler macro ECL_NO_LEGACY can be defined to remove all deprecated
macros and functions. Have a look at the end of src/h/external.h to see
what functions are deprecated on each release -- typically they are functions
that have been renamed with one of the prefixes, ecl_, cl_, si_, ext_ or _ecl.
- Under OS X, you can use the environment variable ABI to select between
32-bits (ABI=32) or 64-bits pointers (ABI=64).
* Bignums:
- When embedded, ECL may coexist with code that uses the GMP library in
different ways, and sometimes that code may use different memory allocation
routines. In order to solve this problem ECL introduces a new option,
ECL_OPT_SET_GMP_MEMORY_FUNCTIONS, which determines whether GMP will use the
Boehm-Weiser garbage collector to allocate memory or not.
- The previous change also implies that ECL must do all bignum computations
using GMP-allocated numbers that are then automatically freed. More
precisely, this is done using big_register[0-2]_get() and
big_register_normalize() everywhere and operating destructively on those
numbers. These functions have been made aware of the fact that GMP may
use other allocation routines and always call mpz_clear() to free memory.
- The internal functions dealing with bignums have been renamed with the usual
prefix "_ecl", and some of them have been turned into macros. Do not use
those functions if possible.
- Bignums are now allocated as a compact IMMUTABLE unit that combines the lisp
object and the data for the GMP integer itself. Since it can be now
allocated as an atomic (i.e. pointerless) region of memory, the garbage
collector has to work less for marking and collecting them.
- The function ecl_alloc_simple_vector() also creates arrays as a compact unit
with pointerfree memory whenever possible.
- The functions cl_alloc_simple_{base,extended}_string() now carry the prefix
ecl_ instead of cl_, and they are simple aliases for ecl_alloc_simple_vector.
- ECL now implements weak pointers. The interface is
(ext:make-weak-pointer value) => weak-pointer
(ext:weak-pointer-value weak-pointer) => value
and it defines a new built-in type, EXT:WEAK-POINTER
- When running in multithreaded mode, the streams for *terminal-io*,
*standard-input* and *standard-output* are created using non-locking
C pipes, instead of locking C streams. This is essential to allow background
threads gain control of the console.
* Native threads:
- ECL now implements WITHOUT-INTERRUPTS and WITH-INTERRUPTS similarly to SBCL.
It creates three local macros, ALLOW-WITH-INTERRUPTS and
WITH-LOCAL-INTERRUPTS, and an additional one WITH-RESTORED-INTERRUPTS,
which executes the forms with the value of interrupts of the forms that
surround WITHOUT-INTERRUPTS.
- Symbol SI:*INTERRUPT-DISABLE* changes name to SI:*INTERRUPTS-DISABLED*.
- New function MP:PROCESS-JOIN, waits for a thread to terminate (J.C.Beaudoin)
For this to work, threads are always created in a joinable state, and they
only detach themselves when exiting and no process is waiting.
- EXT:QUIT now has a different definition
(ext:quit &optional (exit-code 0) (kill-all-threads nil))
The first difference is that QUIT will execute all unwind-protect forms
that surrounds the form that caused the exit. The second one is that we
can explicitely kill all threads and wait for them to exit before exiting.
- Two new functions MP:PROCESS-SUSPEND and MP:PROCESS-RESUME allow suspending
and resuming a thread.
- The C function ecl_import_current_thread() now automatically registers the
thread with the garbage collector. You should use ecl_release_current_thread()
before the thread exits, for consistency.
- Due to the change in ecl_import_current_thread(), ECL now needs a slightly
more modern version of the Boehm-Weiser garbage collector. Hence, even if you
try to enforce --enable-boehm=system, you may find that ECL ignores the
pre-installed library in favor of the one shipped with it.
* Bugs fixed:
- WITH-LOCK might run into a race condition when the lock acquisition process
was interrupted.
- The following functions are now interrupt safe: MP:MAKE-LOCK.
- The use of GIVEUP-LOCK in WITH-LOCK is now protected against interrupts.
- Forms in toplevel EVAL-WHEN or PROGN statements did not have a proper value
of bytecodes.definition, causing the debugger to enter an infinite chain
of errors
- In Solaris 8 ECL failed because it did not find inttypes.h
- When signbit() is not available, ECL now looks for existence of copysign.
* Signals and environment:
- The toplevel now allows handling multiple processes in the same console.
Two new commands, :w and :s, allow reading the list of processes waiting for
grabbing the console and switching to that given process.
- ECL now has a signal handling thread running in the background. When
Ctrl-C or Ctrl-Break is pressed or the process gets a SIGINT signal, then
a separate thread is spawned for handling that event. This thread typically
enters the debugger, allowing either to continue or to interrupt specific
processes.
* Packages:
- The package hash tables are not destructively extended, but rather a new
hash table is created when we need more room. ecl_sethash() now returns
the pointer to the possibly reallocated hashtable.
- ECL now follows the SBCL model, by which there is a global lock for
destructive package operations (INTERN, EXPORT, etc), but not for
query only operations (FIND-SYMBOL, PACKAGE-USE-LIST), etc.
* Windows:
- The latest version of Visual Studio introduced a number of automatically
generated files (*.implib, *.manifest, *.expdef) which are not used at
all by ECL, but which were also not deleted, causing the temporary
directory to be filled with garbage.
- FASLs created by COMPILE and LOAD in the directory for temporary files are
now properly deleted before exiting ECL.
- ECL now embeds the manifest file in the generated executables and shared
libraries.
* Other errors:
- Extended strings (i.e. strings with characters that are not base-characters)
were not identified as valid documentation strings by the interpreter.
