find_module() now raises a RuntimeError, instead of ImportError, on an error on
sys.path or sys.meta_path because load_package() and import_submodule() returns
None and clear the exception if a ImportError occurred.
* Rename _PyImport_FindExtension() to _PyImport_FindExtensionUnicode():
the filename becomes a Unicode object instead of byte string
* Rename _PyImport_FixupExtension() to _PyImport_FixupExtensionUnicode():
the filename becomes a Unicode object instead of byte string
filesystem encoding instead of utf-8.
imp_cache_from_source() encodes the input path to filesystem encoding and this
path is passed to make_compiled_pathname().
* _Py_fopen() and _Py_stat() come from Python/import.c
* (_Py)_wrealpath() comes from Python/sysmodule.c
* _Py_char2wchar(), _Py_wchar2char() and _Py_wfopen() come from Modules/main.c
* (_Py)_wstat(), (_Py)_wgetcwd(), _Py_wreadlink() come from Modules/getpath.c
fromlist to get __import__ to return the module desired. Now it uses the proper
approach of fetching the module from sys.modules.
Closes issue #9252. Thanks to Alexander Belopolsky for the bug report.
Call _wfopen() on Windows, or fopen() otherwise. Return the new file object on
success, or NULL if the file cannot be open or (if PyErr_Occurred()) on unicode
error.
* On non-Windows OSes: the constructor accepts bytes filenames
and use surrogateescape for unicode filenames
* On Windows: use GetFileAttributesW() instead of GetFileAttributesA()
svn+ssh://pythondev@svn.python.org/python/trunk
........
r81380 | brett.cannon | 2010-05-20 13:37:55 -0500 (Thu, 20 May 2010) | 8 lines
Turned out that if you used explicit relative import syntax
(e.g. from .os import sep) and it failed, import would still try the implicit
relative import semantics of an absolute import (from os import sep). That's
not right, so when level is negative, only do explicit relative import
semantics.
Fixes issue #7902. Thanks to Meador Inge for the patch.
........
object to Py_FileSystemDefaultEncoding with the "surrogateescape" error
handler, return a bytes object. If Py_FileSystemDefaultEncoding is not set,
fall back to UTF-8.
svn+ssh://pythondev@svn.python.org/python/trunk
........
r78826 | victor.stinner | 2010-03-10 23:30:19 +0100 (mer., 10 mars 2010) | 5 lines
Issue #3137: Don't ignore errors at startup, especially a keyboard interrupt
(SIGINT). If an error occurs while importing the site module, the error is
printed and Python exits. Initialize the GIL before importing the site
module.
........
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/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.
........
Gregory P. Smith
Antoine Pitrou
Éric Araujo
Victor Stinner
Martin v. Löwis
Nick Coghlan
Georg Brandl
Benjamin Peterson
Brian Curtin
Brett Cannon
Daniel Stutzbach
Barry Warsaw
Gregory P. Smith