* Install the libffi libtool archive. Whilst it is unused, it allows
us to automatically determine the shared library suffix.
* Switch the i386 SunOS bootstrap to .xz.
* Use PKG_CC to build lndir to help environments where the compiler is
not in $PATH.
* Apply patch from Alain O'Dea to fix issues with files residing on
read-only file systems.
* Support shared libraries on 64-bit SunOS.
* Fix typo in patches/patch-rts_Linker.c.
Fixes build on SunOS. Bump PKGREVISION.
* In module.mk, pthread is marked as direct dependency unconditionally for
packages using module.mk. It is unwanted behavior.
* Include pthread.buildlink3.mk in buildlink3.mk only for the case perl5 is
built with "threads".
Fixes potential unwanted dependeny on pthread using perl5/module.mk.
Quote from release announce:
This release contains a fix for a regression of Hash#reject in Ruby 2.1.1,
support for build with Readline-6.3 (see Bug #9578), an updated bundled
version of libyaml with psych, and some bug fixes.
For details, please refer ChangeLog.
Including many bug fixes:
* support for build with Readline-6.3
* a fix for old OpenSSL (regression in p451)
* an updated bundled version of libyaml (see Heap Overflow in YAML URI Escape Parsing (CVE-2014-2525))
For detail, please refer ChangeLog.
Remove the patch which included in upstream
Upstream changes:
Python 3.4.1
Release date: 2014-05-18
Core and Builtins
Issue #21418: Fix a crash in the builtin function super() when called without argument and without current frame (ex: embedded Python).
Issue #21425: Fix flushing of standard streams in the interactive interpreter.
Issue #21435: In rare cases, when running finalizers on objects in cyclic trash a bad pointer dereference could occur due to a subtle flaw in internal iteration logic.
Library
Issue #10744: Fix PEP 3118 format strings on ctypes objects with a nontrivial shape.
Issue #20998: Fixed re.fullmatch() of repeated single character pattern with ignore case. Original patch by Matthew Barnett.
Issue #21075: fileinput.FileInput now reads bytes from standard stream if binary mode is specified. Patch by Sam Kimbrel.
Issue #21396: Fix TextIOWrapper(..., write_through=True) to not force a flush() on the underlying binary stream. Patch by akira.
Issue #21470: Do a better job seeding the random number generator by using enough bytes to span the full state space of the Mersenne Twister.
Issue #21398: Fix an unicode error in the pydoc pager when the documentation contains characters not encodable to the stdout encoding.
Tests
Issue #17756: Fix test_code test when run from the installed location.
Issue #17752: Fix distutils tests when run from the installed location.
IDLE
Issue #18104: Add idlelib/idle_test/htest.py with a few sample tests to begin consolidating and improving human-validated tests of Idle. Change other files as needed to work with htest. Running the module as __main__ runs all tests.
Python 3.4.1rc1
Release date: 2014-05-05
Core and Builtins
Issue #21274: Define PATH_MAX for GNU/Hurd in Python/pythonrun.c.
Issue #21209: Fix sending tuples to custom generator objects with the yield from syntax.
Issue #21134: Fix segfault when str is called on an uninitialized UnicodeEncodeError, UnicodeDecodeError, or UnicodeTranslateError object.
Issue #19537: Fix PyUnicode_DATA() alignment under m68k. Patch by Andreas Schwab.
Issue #20929: Add a type cast to avoid shifting a negative number.
Issue #20731: Properly position in source code files even if they are opened in text mode. Patch by Serhiy Storchaka.
Issue #20637: Key-sharing now also works for instance dictionaries of subclasses. Patch by Peter Ingebretson.
Issue #12546: Allow x00 to be used as a fill character when using str, int, float, and complex __format__ methods.
Issue #13598: Modify string.Formatter to support auto-numbering of replacement fields. It now matches the behavior of str.format() in this regard. Patches by Phil Elson and Ramchandra Apte.
Library
Issue #21088: Bugfix for curses.window.addch() regression in 3.4.0. In porting to Argument Clinic, the first two arguments were reversed.
Issue #21469: Reduced the risk of false positives in robotparser by checking to make sure that robots.txt has been read or does not exist prior to returning True in can_fetch().
Issue #21321: itertools.islice() now releases the reference to the source iterator when the slice is exhausted. Patch by Anton Afanasyev.
Issue #9815: assertRaises now tries to clear references to local variables in the exceptions traceback.
Issue #13204: Calling sys.flags.__new__ would crash the interpreter, now it raises a TypeError.
Issue #19385: Make operations on a closed dbm.dumb database always raise the same exception.
Issue #21207: Detect when the os.urandom cached fd has been closed or replaced, and open it anew.
Issue #21291: subprocesss Popen.wait() is now thread safe so that multiple threads may be calling wait() or poll() on a Popen instance at the same time without losing the Popen.returncode value.
Issue #21127: Path objects can now be instantiated from str subclass instances (such as numpy.str_).
Issue #15002: urllib.response object to use _TemporaryFileWrapper (and _TemporaryFileCloser) facility. Provides a better way to handle file descriptor close. Patch contributed by Christian Theune.
Issue #12220: mindom now raises a custom ValueError indicating it doesnt support spaces in URIs instead of letting a split ValueError bubble up.
Issue #21239: patch.stopall() didnt work deterministically when the same name was patched more than once.
Issue #21222: Passing name keyword argument to mock.create_autospec now works.
Issue #21197: Add lib64 -> lib symlink in venvs on 64-bit non-OS X POSIX.
Issue #17498: Some SMTP servers disconnect after certain errors, violating strict RFC conformance. Instead of losing the error code when we issue the subsequent RSET, smtplib now returns the error code and defers raising the SMTPServerDisconnected error until the next command is issued.
Issue #17826: setting an iterable side_effect on a mock function created by create_autospec now works. Patch by Kushal Das.
Issue #7776: Fix Host: header and reconnection when using http.client.HTTPConnection.set_tunnel(). Patch by Nikolaus Rath.
Issue #20968: unittest.mock.MagicMock now supports division. Patch by Johannes Baiter.
Fix arbitrary memory access in JSONDecoder.raw_decode with a negative second parameter. Bug reported by Guido Vranken.
Issue #21169: getpass now handles non-ascii characters that the input stream encoding cannot encode by re-encoding using the replace error handler.
Issue #21171: Fixed undocumented filter API of the rot13 codec. Patch by Berker Peksag.
Issue #21172: isinstance check relaxed from dict to collections.Mapping.
Issue #21155: asyncio.EventLoop.create_unix_server() now raises a ValueError if path and sock are specified at the same time.
Issue #21149: Improved thread-safety in logging cleanup during interpreter shutdown. Thanks to Devin Jeanpierre for the patch.
Issue #20145: assertRaisesRegex and assertWarnsRegex now raise a TypeError if the second argument is not a string or compiled regex.
Issue #21058: Fix a leak of file descriptor in tempfile.NamedTemporaryFile(), close the file descriptor if io.open() fails
Issue #21200: Return None from pkgutil.get_loader() when __spec__ is missing.
Issue #21013: Enhance ssl.create_default_context() when used for server side sockets to provide better security by default.
Issue #20633: Replace relative import by absolute import.
Issue #20980: Stop wrapping exception when using ThreadPool.
Issue #21082: In os.makedirs, do not set the process-wide umask. Note this changes behavior of makedirs when exist_ok=True.
Issue #20990: Fix issues found by pyflakes for multiprocessing.
Issue #21015: SSL contexts will now automatically select an elliptic curve for ECDH key exchange on OpenSSL 1.0.2 and later, and otherwise default to prime256v1.
Issue #20995: Enhance default ciphers used by the ssl module to enable better security an prioritize perfect forward secrecy.
Issue #20884: Dont assume that __file__ is defined on importlib.__init__.
Issue #21499: Ignore __builtins__ in several test_importlib.test_api tests.
Issue #20879: Delay the initialization of encoding and decoding tables for base32, ascii85 and base85 codecs in the base64 module, and delay the initialization of the unquote_to_bytes() table of the urllib.parse module, to not waste memory if these modules are not used.
Issue #19157: Include the broadcast address in the usuable hosts for IPv6 in ipaddress.
Issue #11599: When an external command (e.g. compiler) fails, distutils now prints out the whole command line (instead of just the command name) if the environment variable DISTUTILS_DEBUG is set.
Issue #4931: distutils should not produce unhelpful error: None messages anymore. distutils.util.grok_environment_error is kept but doc-deprecated.
Issue #20875: Prevent possible gzip read is not defined NameError. Patch by Claudiu Popa.
Issue #11558: email.message.Message.attach now returns a more useful error message if attach is called on a message for which is_multipart is False.
Issue #20283: RE pattern methods now accept the string keyword parameters as documented. The pattern and source keyword parameters are left as deprecated aliases.
Issue #20778: Fix modulefinder to work with bytecode-only modules.
Issue #20791: copy.copy() now doesnt make a copy when the input is a bytes object. Initial patch by Peter Otten.
Issue #19748: On AIX, time.mktime() now raises an OverflowError for year outsize range [1902; 2037].
Issue #20816: Fix inspect.getcallargs() to raise correct TypeError for missing keyword-only arguments. Patch by Jeremiah Lowin.
Issue #20817: Fix inspect.getcallargs() to fail correctly if more than 3 arguments are missing. Patch by Jeremiah Lowin.
Issue #6676: Ensure a meaningful exception is raised when attempting to parse more than one XML document per pyexpat xmlparser instance. (Original patches by Hirokazu Yamamoto and Amaury Forgeot dArc, with suggested wording by David Gutteridge)
Issue #21117: Fix inspect.signature to better support functools.partial. Due to the specifics of functools.partial implementation, positional-or-keyword arguments passed as keyword arguments become keyword-only.
Issue #21209: Fix asyncio.tasks.CoroWrapper to workaround a bug in yield-from implementation in CPythons prior to 3.4.1.
asyncio: Add gi_{frame,running,code} properties to CoroWrapper (upstream issue #163).
Issue #21311: Avoid exception in _osx_support with non-standard compiler configurations. Patch by John Szakmeister.
Issue #11571: Ensure that the turtle window becomes the topmost window when launched on OS X.
Extension Modules
Issue #21276: posixmodule: Dont define USE_XATTRS on KFreeBSD and the Hurd.
Issue #21226: Set up modules properly in PyImport_ExecCodeModuleObject (and friends).
IDLE
Issue #21139: Change default paragraph width to 72, the PEP 8 recommendation.
Issue #21284: Paragraph reformat test passes after user changes reformat width.
Issue #17654: Ensure IDLE menus are customized properly on OS X for non-framework builds and for all variants of Tk.
Build
The Windows build now includes OpenSSL 1.0.1g
Issue #21285: Refactor and fix curses configure check to always search in a ncursesw directory.
Issue #15234: For BerkelyDB and Sqlite, only add the found library and include directories if they arent already being searched. This avoids an explicit runtime library dependency.
Issue #20644: OS X installer build support for documentation build changes in 3.4.1: assume externally supplied sphinx-build is available in /usr/bin.
C API
Issue #20942: PyImport_ImportFrozenModuleObject() no longer sets __file__ to match what importlib does; this affects _frozen_importlib as well as any module loaded using imp.init_frozen().
Documentation
Issue #17386: Expanded functionality of the Doc/make.bat script to make it much more comparable to Doc/Makefile.
Issue #21043: Remove the recommendation for specific CA organizations and to mention the ability to load the OS certificates.
Issue #20765: Add missing documentation for PurePath.with_name() and PurePath.with_suffix().
Issue #19407: New package installation and distribution guides based on the Python Packaging Authority tools. Existing guides have been retained as legacy links from the distutils docs, as they still contain some required reference material for tool developers that isnt recorded anywhere else.
Issue #19697: Document cases where __main__.__spec__ is None.
Tests
Issue #18604: Consolidated checks for GUI availability. All platforms now at least check whether Tk can be instantiated when the GUI resource is requested.
Issue #21275: Fix a socket test on KFreeBSD.
Issue #21223: Pass test_site/test_startup_imports when some of the extensions are built as builtins.
Issue #20635: Added tests for Tk geometry managers.
Add test case for freeze.
Issue #20743: Fix a reference leak in test_tcl.
Issue #21097: Move test_namespace_pkgs into test_importlib.
Issue #20939: Avoid various network test failures due to new redirect of http://www.python.org/ to https://www.python.org: use http://www.example.com instead.
Issue #20668: asyncio tests no longer rely on tests.txt file. (Patch by Vajrasky Kok)
Issue #21093: Prevent failures of ctypes test_macholib on OS X if a copy of libz exists in $HOME/lib or /usr/local/lib.
Tools/Demos
Add support for yield from to 2to3.
Add support for the PEP 465 matrix multiplication operator to 2to3.
Issue #16047: Fix module exception list and __file__ handling in freeze. Patch by Meador Inge.
Issue #11824: Consider ABI tags in freeze. Patch by Meador Inge.
Issue #20535: PYTHONWARNING no longer affects the run_tests.py script. Patch by Arfrever Frehtes Taifersar Arahesis.
Makefile:
See ${WRKSRC}/libgcc/config/t-slibgcc-darwin: It uses strip(1) to
create a stub library, not just to remove symbols, so we must not
let strip(1) be a no-op regardless of ${INSTALL_UNSTRIPPED} or the
build fails for missing files.
patches/patch-libgcc_config_t-slibgcc-darwin:
If we don't install libgcc_s.10.[45].dylib, our gcc links binaries
with *both* /usr/lib/libgcc_s.1.dylib and
${GCC_PREFIX}/lib/libgcc_s.1.dylib, which is certainly a bad thing.
patches/patch-libjava_Makefile.in:
Tell libtool the fact that libjvm.so is a module to be opened with
dlopen(3). This is actually needed on Mach-O systems like Darwin.
Makefile:
See ${WRKSRC}/libgcc/config/t-slibgcc-darwin: It uses strip(1) to
create a stub library, not just to remove symbols, so we must not
let strip(1) be a no-op regardless of ${INSTALL_UNSTRIPPED} or the
build fails for missing files.
patches/patch-ac:
Remove the whitespace between $(LINKER_RPATH_FLAG) and the path.
Otherwise the flag may be misinterpreted by linker (e.g. Apple's ld)
Makefile:
See ${WRKSRC}/libgcc/config/t-slibgcc-darwin: It uses strip(1) to
create a stub library, not just to remove symbols, so we must not
let strip(1) be a no-op regardless of ${INSTALL_UNSTRIPPED} or the
build fails for missing files.
patches/patch-ac:
Remove the whitespace between $(LINKER_RPATH_FLAG) and the path.
Otherwise the flag may be misinterpreted by linker (e.g. Apple's ld)
patches/patch-libgcc_config_t-slibgcc-darwin:
If we don't install libgcc_s.10.[45].dylib, our gcc links binaries
with *both* /usr/lib/libgcc_s.1.dylib and
${GCC_PREFIX}/lib/libgcc_s.1.dylib, which is certainly a bad thing.
gcc-aux builds on OpenBSD, but the default posix thread handling for Ada
tasking no longer works. All tasks get stuck in a "Thread Sleep" sleep
state. The gcc4.8.2-based egcc in OpenBSD ports supports Ada and passes
all tests. There is no difference in the configuration between that and
gcc-aux. To fix, it probably requires tasking support unique to OpenBSD
and that's out of scope for me at the moment, especially considering the
lack of a suitable development machine.
As such, I'm suspending my work on OpenBSD. This commit captures some
minor bug fixes. The addition of "--enable-checking" configuration switch
should be the default so there's no need bump gcc-aux again.
only supported by the native version, so ensure /usr/bin/nm is called. Also
explicitly pass the --arch flag to ensure the correct compiler flags are used.
On platforms where DTrace is available, we need to invoke dtrace(1)
with the original PATH. Otherwise it gets confused in the presence
of our cc wrapper because it tries to invoke the system-default gcc
with argv[0] set to "gcc", not the absolute path to gcc.
Changes from V8.4pl3 to V8.4pl4
===============================
WARNING:
The current logic of Coq is now known to be inconsistent with
Axiom prop_extensionality : forall A B:Prop, (A <-> B) -> A = B.
For more details, see:
https://gforge.inria.fr/plugins/scmgit/cgi-bin/gitweb.cgi?p=coq/coq.git;a=blob_plain;f=test-suite/failure/subterm2.v;hb=HEAD
or
https://gforge.inria.fr/plugins/scmgit/cgi-bin/gitweb.cgi?p=coq/coq.git;a=blob_plain;f=test-suite/failure/subterm3.v;hb=HEAD
Kernel
- Unsound check of elimination sort.
- Fix guard condition for nested cofixpoints.
- Univ constraints of module subtyping were not propagated.
Tactics
- A new option "Set Stable Omega" ensures that repeated identical calls
to omega will produce identical proof terms. This option is off by default
for maximal compatibility, but should be pretty safe to activate.
- The interpretation of the open_constr tactic argument was erroneously
firing type classes resolution in some corner cases. This has been
fixed. The tactic argument type open_constr_wTC is provided for retro
compatibility purposes.
- Fixing bug (fixing precedence of ltac variables over variables in
env) introduces rare and justified tactic failure.
Bug fixes
- micromega: solved an ambiguous symbol resolution.
- Coq always uses / as separator between directories on all platforms.
- remove trailing '\r' from file names returned by coqtop.
- bug correction in proving inversion principles for Function.
- ocamlbuild: minor fixes related to camlp4 and cross-compilation.
Changes from V8.4pl2 to V8.4pl3
===============================
Ide_slave XML interface
- 20120712, 20130419 : Invalidated protocol versions
- From 20130419 extra datastructure : union
(Inl "" = <union val="in_l"><string></string></union>,
Inr _ = <union val="in_r">...)
- 20130419~1 : new toplevel entry : message, not send by coptop v8.4 and not
handle by coqide v8.4. A message has a level and a content (of string).
Message levels are Debug of string, Info, Notice, Warning and Error.
- 20130425 :
* new toplevel entry : feedback, once again not send by coqtop v8.4 and not
handle by coqide v8.4. A feedback gives the id of the sentence it provides info
about and a content. Feedback contents are Processed, AddedAxiom and
GlobRef of Util.loc * string * string * string * string
* <call val="interp"> must provide an attribute id of type int. It is OK in
coqtop v8.4 to alwais send <call val="interp" id="0">
Bug fixes
- Fixing a significant efficiency leak in the code of the field tactic.
- Fix caching of local hint database in typeclasses eauto which could
miss some hypotheses.
- Fix automatic solving of obligation in program, which was not trying
to solve obligations that had no undefined dependencies left.
These changes are for NetBSD only although non-functional OpenBSD support
continues to be added. The biggest change is configuring the compiler to
use the gold linker from binutils 2.24 instead of the system linker. The
ld.bfd linker from binutils 2.24 is buggy on NetBSD and OpenBSD; it cannot
properly produce/recognize PIC files in every case, which is why gcc-aux
had been limited to the system linker. The system linker was too old for
gcc-aux and devel/gps failed to link because of it.
At the same time, the dormant dl_iterate_phdr support has been activated
for NetBSD. It seems to work; the Ada testsuite still passes perfectly.
Due to lack of testing, gold is restricted from building on NetBSD 5,
which means lang/gcc-aux will fail to build as a result. If someone
verifies gold builds on NetBSD 5, then the restriction on binutils can
be removed and lang/gcc-aux will be able to build on NetBSD 5 again.
I was never able to build gcc-aux with multilib support when it was
based on gcc 4.7. It turns out that this is because Ada support for
multilib was broken per GCC's bugzilla. It's been fixed since. After
some significant effort and many hours, I finally produced a multilib
bootstrap for Solaris. It can produce other multilib compilers (x86-64)
and it can produce 32-bit only compilers (i386). This was tested on a
stock OmniOS x86-64 installation and the Joyent dev area (i386).
Solaris has not supported static linking since Solaris 9, so it is not
possible to create a static bootstrap like the BSDs have, nor is it
possible to build the compiler statically, so that option is disabled.
That means it is unlikely that the bootstrap will run on Solaris 10 or
Solaris 11, but this theory has not been tested.
Much of the changes to the diff-* patchsets are a result of OpenBSD work.
An OpenBSD static bootstrap has been created against devel/binutils, and
as a result it fails to find system libraries such as libc and libm. That
is a side note explaining the OpenBSD additions. There are some DragonFly
and NetBSD changes in the diff-* patches though.
The bootstrap target was significantly modified, mainly to capture the
special needs of creating a Solaris bootstrap. The creation of the
bootstrap tarball was put in it's own target. The zlib and math libraries
were connected to the bootstrap option. It's not a good idea to use
system zlib on Illumos because the zlib header is not guaranteed to be
installed.
The feature to rebuild the compiler with the installed gcc-aux was fixed;
it should use it's own c++ compiler instead of the system c++ compiler.
OpenBSD 5.5 does not support "cp -a", so this was changed to "cp -RpP" for
portability reasons.
