I've left a couple of them in: zlib (third-party lib), getaddrinfo.c
(doesn't include Python.h, and probably obsolete), _sre.c (legitimate
use for the re.LOCALE flag), mpdecimal (needs to build without Python.h).
I've left a couple of them in: zlib (third-party lib), getaddrinfo.c
(doesn't include Python.h, and probably obsolete), _sre.c (legitimate
use for the re.LOCALE flag).
By setting the st_dev attribute, we can then remove some Windows-specific
code and move os.path.samefile/sameopenfile/samestat to Lib/genericpath.py
so all platforms share the same implementation.
And other fixes for Windows:
* rename, replace and link require arguments of the same type on Windows
* readlink only supports unicode filenames on Windows
* os.open() specifies the filename on OSError
Pass the original filename argument to OSError constructor, instead of trying
to encode it to or decode it from the filesystem encoding. This change avoids
an additionnal UnicodeDecodeError on Windows if the filename cannot be decoded
from the filesystem encoding (ANSI code page).
parameter from os.remove / os.unlink.
Patch written by Georg Brandl. (I'm really looking forward to George
getting commit privileges so I don't have to keep doing checkins on his
behalf.)
Many functions now support "dir_fd" and "follow_symlinks" parameters;
some also support accepting an open file descriptor in place of 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.
Removed futimens as it is now redundant.
Changed shutil.copystat to use st_atime_ns and st_mtime_ns from os.stat
and ns= parameter to utime--it once again preserves exact metadata on Linux!
time.ctime(), gmtime(), time.localtime(), datetime.date.fromtimestamp(),
datetime.datetime.fromtimestamp() and datetime.datetime.utcfromtimestamp() now
raises an OverflowError, instead of a ValueError, if the timestamp does not fit
in time_t.
datetime.datetime.fromtimestamp() and datetime.datetime.utcfromtimestamp() now
round microseconds towards zero instead of rounding to nearest with ties going
away from zero.
in order to make algorithmic complexity attacks on (e.g.) web apps much more complicated.
The environment variable PYTHONHASHSEED and the new command line flag -R control this
behavior.
in order to make algorithmic complexity attacks on (e.g.) web apps much more complicated.
The environment variable PYTHONHASHSEED and the new command line flag -R control this
behavior.
defaulting to None like the other utimes family members. It now accepts
keyword arguments because, unlike other other functions in the family,
it has a `flags` value at the end of the argument list (which
retains its 0 default).
* Replace PyUnicode_AS_UNICODE by PyUnicode_AsUnicode, PyUnicode_AS_UNICODE is
no more a real macro
* Replace Py_UNICODE by wchar_t in code specific to Windows
After 1a3e8db28d49, Windows XP could not os.stat at all due to raising
immediately when GetFinalPathNameByHandle wasn't available (pre-Vista).
The proper behavior in that situation is to just not attempt a traversal
rather than outright rejecting.
This change additionally handles a failed malloc by setting the error code
and returning false.
Patch by Hirokazu Yamamoto.
Use of DeviceIoControl to obtain the symlink path via the reparse tag was
removed. The code now uses GetFinalPathNameByHandle in the case of a
symbolic link and works properly given the added test which creates a symbolic
link and calls os.stat on it from multiple locations.
Victor Stinner also noticed an issue with os.lstat following the os.stat
code path when being passed bytes. The posix_lstat function was adjusted to
properly hook up win32_lstat instead of the previous STAT macro (win32_stat).
Rather than wrapping the C _isdir function in a Python function,
just import the C _isdir function directly. Additionally, add in the
docstring which was left out.
By changing to the Windows GetFileAttributes API in nt._isdir we can figure
out if the path is a directory without opening the file via os.stat. This has
the minor benefit of speeding up os.path.isdir by at least 2x for regular
files and 10-15x improvements were seen on symbolic links (which opened the
file multiple times during os.stat). Since os.path.isdir is used in
several places on interpreter startup, we get a minor speedup in startup time.
By changing to the Windows GetFileAttributes API in nt._isdir we can figure
out if the path is a directory without opening the file via os.stat. This has
the minor benefit of speeding up os.path.isdir by at least 2x for regular
files and 10-15x improvements were seen on symbolic links (which opened the
file multiple times during os.stat). Since os.path.isdir is used in
several places on interpreter startup, we get a minor speedup in startup time.
svn+ssh://pythondev@svn.python.org/python/branches/py3k
........
r88111 | antoine.pitrou | 2011-01-19 16:21:35 +0100 (mer., 19 janv. 2011) | 4 lines
Issue #10898: Allow compiling the posix module when the C library defines
a symbol named FSTAT.
........
svn+ssh://pythondev@svn.python.org/python/branches/py3k
........
r87802 | antoine.pitrou | 2011-01-06 19:25:55 +0100 (jeu., 06 janv. 2011) | 6 lines
Issue #7858: Raise an error properly when os.utime() fails under Windows
on an existing file.
(this does not seem to be easily testable)
........
svn+ssh://pythondev@svn.python.org/python/branches/py3k
........
r87666 | amaury.forgeotdarc | 2011-01-03 01:19:11 +0100 (lun., 03 janv. 2011) | 4 lines
#8278: In the Windows implementation of stat() and utime(),
use time_t instead of int. This gives support for dates after 2038,
at least when compiled with VS2003 or later, where time_t is 64bit.
........
In order to create symlinks on Windows, SeCreateSymbolicLinkPrivilege
is an account privilege that is required to be held by the user. Not only
must the privilege be enabled for the account, the activated privileges for
the currently running application must be adjusted to enable the requested
privilege.
Rather than exposing an additional function to be called prior to the user's
first os.symlink call, we handle the AdjustTokenPrivileges Windows API call
internally and only expose os.symlink when the privilege escalation was
successful.
Due to the change of only exposing os.symlink when it's available, we can
go back to the original test skipping methods of checking via `hasattr`.
Amaury noticed that this was originally written in a way that would fail on
names that can't be encoded with the mbcs codec. Restructured the function
to work with wide names first then narrow names second, to fall in line
with the way other functions are written in posixmodule.c.
Additionally, the st_ino attribute of stat structures was not being filled
in. This was left out of the fix to #10027 and was noticed due to
test_tarfile failing when applying the patch for this issue. An earlier
version of the fix to #10027 included st_ino, but that attribute got lost
in the shuffle of a few review/fix cycles. All tests pass.
Note: This patch has no tests because as of now there is no way to create
links. #8879 adds that and the tests will go in there. I've manually observed
that existing links on my system function properly with this.
The test is semi-dumb, it just makes sure something comes back since we
don't have a solid source to validate the returned login. We can't be 100%
sure that the USERNAME env var will always match what os.getlogin() returns,
so we don't make any specific assertion there.
This uses the GetFileInformationByHandle function to return a tuple of values
to identify a file, then ntpath.sameopenfile compares file tuples, which
is exposed as os.path.sameopenfile.
svn+ssh://pythondev@svn.python.org/python/branches/py3k
........
r84489 | antoine.pitrou | 2010-09-04 19:21:57 +0200 (sam., 04 sept. 2010) | 4 lines
Issue #7736: Release the GIL around calls to opendir() and closedir()
in the posix module. Patch by Marcin Bachry.
........
svn+ssh://pythondev@svn.python.org/python/branches/py3k
........
r84063 | victor.stinner | 2010-08-15 11:33:08 +0200 (dim., 15 août 2010) | 5 lines
Issue #9605: posix.getlogin() decodes the username with file filesystem
encoding and surrogateescape error handler. Patch written by David Watson.
Reindent also posix_getlogin(), and fix a typo in the NEWS file.
........
svn+ssh://pythondev@svn.python.org/python/branches/py3k
........
r84060 | victor.stinner | 2010-08-15 11:12:51 +0200 (dim., 15 août 2010) | 4 lines
Issue #9603: posix.ttyname() and posix.ctermid() decode the terminal name
using the filesystem encoding and surrogateescape error handler. Patch
written by David Watson.
........
svn+ssh://pythondev@svn.python.org/python/branches/py3k
........
r83951 | benjamin.peterson | 2010-08-11 14:20:42 -0500 (Wed, 11 Aug 2010) | 4 lines
use pep 383 decoding for mknod and mkfifo #9570
Patch by David Watson.
........
svn+ssh://pythondev@svn.python.org/python/branches/py3k
........
r83921 | antoine.pitrou | 2010-08-10 01:39:31 +0200 (mar., 10 août 2010) | 4 lines
Issue #6915: Under Windows, os.listdir() didn't release the Global
Interpreter Lock around all system calls. Original patch by Ryan Kelly.
........
with the fact that getgroups(2) might return
more that MAX_GROUPS on OSX.
See the issue (and python-dev archives) for the
gory details. Summarized: OSX behaves rather oddly
and Apple says this is intentional.
with the fact that getgroups(2) might return
more that MAX_GROUPS on OSX.
See the issue (and python-dev archives) for the
gory details. Summarized: OSX behaves rather oddly
and Apple says this is intentional.
Added Windows support for os.symlink when run on Windows 6.0 or greater,
aka Vista. Previous Windows versions will raise NotImplementedError
when trying to symlink.
Includes numerous test updates and additions to test_os, including
a symlink_support module because of the fact that privilege escalation
is required in order to run the tests to ensure that the user is able
to create symlinks. By default, accounts do not have the required
privilege, so the escalation code will have to be exposed later (or
documented on how to do so). I'll be following up with that work next.
Note that the tests use ctypes, which was agreed on during the PyCon
language summit.
PyUnicode_DecodeFSDefault*() doesn't use surrogateescape error handler, and so
PyUnicode_FromEncodedObject(v, Py_FileSystemDefaultEncoding, "surrogateescape")
cannot be replaced by PyUnicode_DecodeFSDefault().
It's a bad idea to try to fix surrogates things in Python 3.1...
Create os.environb mapping and os.getenvb() function, os.unsetenv() encodes str
argument to the file system encoding with the surrogateescape error handler
(instead of utf8/strict) and accepts bytes, and posix.environ keys and values
are bytes.
svn+ssh://pythondev@svn.python.org/python/branches/py3k
................
r80846 | victor.stinner | 2010-05-06 02:08:46 +0200 (jeu., 06 mai 2010) | 20 lines
Recorded merge of revisions 80844-80845 via svnmerge from
svn+ssh://pythondev@svn.python.org/python/trunk
........
r80844 | victor.stinner | 2010-05-06 01:33:33 +0200 (jeu., 06 mai 2010) | 5 lines
Untabify Modules/posixmodule.c
Run Antoine Pitrou "untabify" script + manual editions (OS/2 and some
continuation lines).
........
r80845 | victor.stinner | 2010-05-06 02:03:44 +0200 (jeu., 06 mai 2010) | 4 lines
Untabify Modules/posixmodule.c (2)
Fix some more functions by hand
........
I rewrote the patch for py3k from scratch using untabify + manual editions
................
svn+ssh://pythondev@svn.python.org/python/trunk
........
r80844 | victor.stinner | 2010-05-06 01:33:33 +0200 (jeu., 06 mai 2010) | 5 lines
Untabify Modules/posixmodule.c
Run Antoine Pitrou "untabify" script + manual editions (OS/2 and some
continuation lines).
........
r80845 | victor.stinner | 2010-05-06 02:03:44 +0200 (jeu., 06 mai 2010) | 4 lines
Untabify Modules/posixmodule.c (2)
Fix some more functions by hand
........
I rewrote the patch for py3k from scratch using untabify + manual editions
svn+ssh://pythondev@svn.python.org/python/branches/py3k
........
r80421 | victor.stinner | 2010-04-23 23:41:56 +0200 (ven., 23 avril 2010) | 3 lines
Issue #8391: os.execvpe() and os.getenv() supports unicode with surrogates and
bytes strings for environment keys and values
........
r80424 | victor.stinner | 2010-04-24 00:55:39 +0200 (sam., 24 avril 2010) | 13 lines
Fix test_undecodable_env of test_subproces for non-ASCII directory
This test was introduced by r80421 (issue #8391).
The fix: copy the environment variables instead of starting Python in an empty
environement. In an empty environment, the locale is C and Python uses ASCII
for the default file system encoding. The non-ASCII directory will be encoded
using surrogates, but Python3 is unable to load a module or package with a
filename using surrogates.
See issue #8242 for more information about running Python3 with a non-ascii
directory in an empty environement.
........
POSIX systems with a C extension module. This is required in order for
the subprocess module to be made thread safe.
The pure python implementation is retained so that it can continue to be
used if for some reason the _posixsubprocess extension module is not
available.
The unittest executes tests on both code paths to guarantee compatibility.
* Moves PyLong_FromPid and PyLong_AsPid from posixmodule.c into longobject.h.
Code reviewed by jeffrey.yasskin at http://codereview.appspot.com/223077/show
svn+ssh://pythondev@svn.python.org/python/trunk
........
r78531 | gregory.p.smith | 2010-02-28 18:31:33 -0800 (Sun, 28 Feb 2010) | 2 lines
Fix for r78527. It left out updating forkpty.
........
svn+ssh://pythondev@svn.python.org/python/trunk
........
r78527 | gregory.p.smith | 2010-02-28 17:22:39 -0800 (Sun, 28 Feb 2010) | 4 lines
Issue #7242: On Solaris 9 and earlier calling os.fork() from within a
thread could raise an incorrect RuntimeError about not holding the import
lock. The import lock is now reinitialized after fork.
........
r78550 | gregory.p.smith | 2010-02-28 22:01:02 -0800 (Sun, 28 Feb 2010) | 2 lines
Fix test to be skipped on windows.
........
svn+ssh://pythondev@svn.python.org/python/branches/py3k
................
r78548 | gregory.p.smith | 2010-02-28 21:54:14 -0800 (Sun, 28 Feb 2010) | 10 lines
Merged revisions 78546 via svnmerge from
svn+ssh://pythondev@svn.python.org/python/trunk
........
r78546 | gregory.p.smith | 2010-02-28 21:43:43 -0800 (Sun, 28 Feb 2010) | 3 lines
Fixes issue #7999: os.setreuid() and os.setregid() would refuse to accept
a -1 parameter on some platforms such as OS X.
........
................
svn+ssh://pythondev@svn.python.org/python/trunk
........
r78546 | gregory.p.smith | 2010-02-28 21:43:43 -0800 (Sun, 28 Feb 2010) | 3 lines
Fixes issue #7999: os.setreuid() and os.setregid() would refuse to accept
a -1 parameter on some platforms such as OS X.
........
svn+ssh://pythondev@svn.python.org/python/trunk
........
r76636 | antoine.pitrou | 2009-12-02 21:37:54 +0100 (mer., 02 déc. 2009) | 5 lines
Issue #7333: The `posix` module gains an `initgroups()` function providing
access to the initgroups(3) C library call on Unix systems which implement
it. Patch by Jean-Paul Calderone.
........
svn+ssh://pythondev@svn.python.org/python/branches/py3k
................
r75246 | benjamin.peterson | 2009-10-04 15:32:25 -0500 (Sun, 04 Oct 2009) | 29 lines
Merged revisions 74841 via svnmerge from
svn+ssh://pythondev@svn.python.org/python/trunk
........
r74841 | thomas.wouters | 2009-09-16 14:55:54 -0500 (Wed, 16 Sep 2009) | 23 lines
Fix issue #1590864, multiple threads and fork() can cause deadlocks, by
acquiring the import lock around fork() calls. This prevents other threads
from having that lock while the fork happens, and is the recommended way of
dealing with such issues. There are two other locks we care about, the GIL
and the Thread Local Storage lock. The GIL is obviously held when calling
Python functions like os.fork(), and the TLS lock is explicitly reallocated
instead, while also deleting now-orphaned TLS data.
This only fixes calls to os.fork(), not extension modules or embedding
programs calling C's fork() directly. Solving that requires a new set of API
functions, and possibly a rewrite of the Python/thread_*.c mess. Add a
warning explaining the problem to the documentation in the mean time.
This also changes behaviour a little on AIX. Before, AIX (but only AIX) was
getting the import lock reallocated, seemingly to avoid this very same
problem. This is not the right approach, because the import lock is a
re-entrant one, and reallocating would do the wrong thing when forking while
holding the import lock.
Will backport to 2.6, minus the tiny AIX behaviour change.
........
................
svn+ssh://pythondev@svn.python.org/python/trunk
........
r74841 | thomas.wouters | 2009-09-16 14:55:54 -0500 (Wed, 16 Sep 2009) | 23 lines
Fix issue #1590864, multiple threads and fork() can cause deadlocks, by
acquiring the import lock around fork() calls. This prevents other threads
from having that lock while the fork happens, and is the recommended way of
dealing with such issues. There are two other locks we care about, the GIL
and the Thread Local Storage lock. The GIL is obviously held when calling
Python functions like os.fork(), and the TLS lock is explicitly reallocated
instead, while also deleting now-orphaned TLS data.
This only fixes calls to os.fork(), not extension modules or embedding
programs calling C's fork() directly. Solving that requires a new set of API
functions, and possibly a rewrite of the Python/thread_*.c mess. Add a
warning explaining the problem to the documentation in the mean time.
This also changes behaviour a little on AIX. Before, AIX (but only AIX) was
getting the import lock reallocated, seemingly to avoid this very same
problem. This is not the right approach, because the import lock is a
re-entrant one, and reallocating would do the wrong thing when forking while
holding the import lock.
Will backport to 2.6, minus the tiny AIX behaviour change.
........
svn+ssh://pythondev@svn.python.org/python/trunk
........
r72852 | antoine.pitrou | 2009-05-23 17:37:45 +0200 (sam., 23 mai 2009) | 5 lines
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.
........
svn+ssh://pythondev@svn.python.org/python/trunk
........
r72698 | hirokazu.yamamoto | 2009-05-17 11:52:09 +0900 | 1 line
Issue #3527: Removed Py_WIN_WIDE_FILENAMES which is not used any more.
........
r72699 | hirokazu.yamamoto | 2009-05-17 11:58:36 +0900 | 1 line
Added NEWS for r72698.
........
Charles-Francois Natali
Victor Stinner
Benjamin Peterson
Antoine Pitrou
Richard Oudkerk
Gregory P. Smith
Serhiy Storchaka
Ronald Oussoren
Brian Curtin
Andrew Svetlov
Petri Lehtinen
Jesus Cea
Christian Heimes
Trent Nelson
Georg Brandl
Larry Hastings
Stefan Krah
Victor Stinner
Brett Cannon
Victor Stinner
Charles-François Natali
Martin v. Löwis
Ross Lagerwall
Giampaolo Rodola'
Ezio Melotti
Amaury Forgeot d'Arc
Brian Curtin
Hirokazu Yamamoto
Stefan Krah
Andrew M. Kuchling
Raymond Hettinger
Senthil Kumaran
Mark Dickinson
Gregory P. Smith
Eric Smith