and lifetime issues of dynamically allocated Py_buffer members (#9990)
as well as crashes (#8305, #7433). Many new features have been added
(See whatsnew/3.3), and the documentation has been updated extensively.
The ndarray test object from _testbuffer.c implements all aspects of
PEP-3118, so further development towards the complete implementation
of the PEP can proceed in a test-driven manner.
Thanks to Nick Coghlan, Antoine Pitrou and Pauli Virtanen for review
and many ideas.
- Issue #12834: Fix incorrect results of memoryview.tobytes() for
non-contiguous arrays.
- Issue #5231: Introduce memoryview.cast() method that allows changing
format and shape without making a copy of the underlying memory.
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.
and check the string consistency.
_PyUnicode_CheckConsistency() doesn't check the hash anymore. It should be
possible to call this function even if hash(str) was already called.
* PyUnicode_EncodeFSDefault(), PyUnicode_DecodeFSDefaultAndSize() and
PyUnicode_DecodeFSDefault() use the locale encoding instead of UTF-8 if
Py_FileSystemDefaultEncoding is NULL
* redecode_filenames() functions and _Py_code_object_list (issue #9630)
are no more needed: remove them
Redecode the filenames of:
- all modules: __file__ and __path__ attributes
- all code objects: co_filename attribute
- sys.path
- sys.meta_path
- sys.executable
- sys.path_importer_cache (keys)
Keep weak references to all code objects until initfsencoding() is called, to
be able to redecode co_filename attribute of all code objects.
(instances of int, float, complex, decimal.Decimal and
fractions.Fraction) that makes it easy to maintain the invariant that
hash(x) == hash(y) whenever x and y have equal value.
svn+ssh://pythondev@svn.python.org/python/branches/py3k
........
r81250 | victor.stinner | 2010-05-17 03:13:37 +0200 (lun., 17 mai 2010) | 2 lines
Issue #6697: Fix a crash if code of "python -c code" contains surrogates
........
r81251 | victor.stinner | 2010-05-17 03:26:01 +0200 (lun., 17 mai 2010) | 3 lines
PyObject_Dump() encodes unicode objects to utf8 with backslashreplace (instead
of strict) error handler to escape surrogates
........
r81252 | victor.stinner | 2010-05-17 10:58:51 +0200 (lun., 17 mai 2010) | 6 lines
handle_system_exit() flushs files to warranty the output order
PyObject_Print() writes into the C object stderr, whereas PySys_WriteStderr()
writes into the Python object sys.stderr. Each object has its own buffer, so
call sys.stderr.flush() and fflush(stderr).
........
r81253 | victor.stinner | 2010-05-17 11:33:42 +0200 (lun., 17 mai 2010) | 6 lines
Fix refleak in internal_print() introduced by myself in r81251
_PyUnicode_AsDefaultEncodedString() uses a magical PyUnicode attribute to
automatically destroy PyUnicode_EncodeUTF8() result when the unicode string is
destroyed.
........
_PyUnicode_AsDefaultEncodedString() uses a magical PyUnicode attribute to
automatically destroy PyUnicode_EncodeUTF8() result when the unicode string is
destroyed.
svn+ssh://pythondev@svn.python.org/python/trunk
........
r79809 | mark.dickinson | 2010-04-05 19:54:51 +0100 (Mon, 05 Apr 2010) | 1 line
Use a better NaN test in _Py_HashDouble as well.
........
svn+ssh://pythondev@svn.python.org/python/trunk
........
r79804 | mark.dickinson | 2010-04-05 19:07:51 +0100 (Mon, 05 Apr 2010) | 5 lines
Use a more robust infinity check in _Py_HashDouble.
This fixes a test_decimal failure on FreeBSD 8.0. (modf apparently
doesn't follow C99 Annex F on FreeBSD.)
........
svn+ssh://pythondev@svn.python.org/python/trunk
........
r74457 | benjamin.peterson | 2009-08-15 08:16:38 -0500 (Sat, 15 Aug 2009) | 1 line
#6707 fix a crash with dir() on an uninitialized module
........
svn+ssh://pythondev@svn.python.org/python/trunk
........
r71722 | benjamin.peterson | 2009-04-18 15:12:47 -0500 (Sat, 18 Apr 2009) | 1 line
try to initalize all builtin types with PyType_Ready to avoid problems like #5787
........
svn+ssh://pythondev@svn.python.org/python/trunk
........
r69436 | mark.dickinson | 2009-02-08 14:42:28 +0000 (Sun, 08 Feb 2009) | 10 lines
Issue #789290: make sure that hash(2**63) == hash(2.**63) on 64-bit
platforms. The previous code was fragile, depending on the twin
accidents that:
(1) in C, casting the double value 2.**63 to long returns the integer
value -2**63, and
(2) in Python, hash(-2**63) == hash(2**63).
There's already a test for this in test_hash.
........
from object.h; don't inherit tp_compare slot on subclasses; and raise TypeError
when initializing a type that has a nonzero tp_compare slot. Fix up
comparison-related comments in object.c and code.h.
svn+ssh://pythondev@svn.python.org/python/trunk
........
r68560 | amaury.forgeotdarc | 2009-01-13 00:36:55 +0100 (mar., 13 janv. 2009) | 6 lines
#3720: Interpreter crashes when an evil iterator removes its own next function.
Now the slot is filled with a function that always raises.
Will not backport: extensions compiled with 2.6.x would not run on 2.6.0.
........
svn+ssh://pythondev@svn.python.org/python/trunk
........
r67654 | georg.brandl | 2008-12-07 16:42:09 -0600 (Sun, 07 Dec 2008) | 2 lines
#4457: rewrite __import__() documentation.
........
r67676 | benjamin.peterson | 2008-12-08 20:03:03 -0600 (Mon, 08 Dec 2008) | 1 line
specify how things are copied
........
r67677 | benjamin.peterson | 2008-12-08 20:05:11 -0600 (Mon, 08 Dec 2008) | 1 line
revert unrelated change to installer script
........
r67681 | jeremy.hylton | 2008-12-09 15:03:10 -0600 (Tue, 09 Dec 2008) | 2 lines
Add simple unittests for Request
........
r67692 | amaury.forgeotdarc | 2008-12-10 18:03:42 -0600 (Wed, 10 Dec 2008) | 2 lines
#1030250: correctly pass the dry_run option to the mkpath() function.
........
r67725 | benjamin.peterson | 2008-12-12 22:02:20 -0600 (Fri, 12 Dec 2008) | 1 line
fix incorrect example
........
r67761 | benjamin.peterson | 2008-12-14 11:26:04 -0600 (Sun, 14 Dec 2008) | 1 line
fix missing bracket
........
r67784 | georg.brandl | 2008-12-15 02:33:58 -0600 (Mon, 15 Dec 2008) | 2 lines
#4446: document "platforms" argument for setup().
........
r67785 | georg.brandl | 2008-12-15 02:36:11 -0600 (Mon, 15 Dec 2008) | 2 lines
#4611: fix typo.
........
r67787 | georg.brandl | 2008-12-15 02:58:59 -0600 (Mon, 15 Dec 2008) | 2 lines
#4578: fix has_key() usage in compiler package.
........
r67788 | georg.brandl | 2008-12-15 03:07:39 -0600 (Mon, 15 Dec 2008) | 2 lines
#4568: remove limitation in varargs callback example.
........
r67802 | amaury.forgeotdarc | 2008-12-15 16:29:14 -0600 (Mon, 15 Dec 2008) | 4 lines
#3632: the "pyo" macro from gdbinit can now run when the GIL is released.
Patch by haypo.
........
r67848 | benjamin.peterson | 2008-12-18 20:28:56 -0600 (Thu, 18 Dec 2008) | 1 line
fix typo
........
r67849 | benjamin.peterson | 2008-12-18 20:31:35 -0600 (Thu, 18 Dec 2008) | 1 line
_call_method -> _callmethod and _get_value to _getvalue
........
r67850 | raymond.hettinger | 2008-12-19 03:06:07 -0600 (Fri, 19 Dec 2008) | 9 lines
Fix-up and clean-up docs for int.bit_length().
* Replace dramatic footnote with in-line comment about possible round-off errors in logarithms of large numbers.
* Add comments to the pure python code equivalent.
* replace floor() with int() in the mathematical equivalent so the type is correct (should be an int, not a float).
* add abs() to the mathematical equivalent so that it matches the previous line that it is supposed to be equivalent to.
* make one combined example with a negative input.
........
r67862 | benjamin.peterson | 2008-12-19 20:48:02 -0600 (Fri, 19 Dec 2008) | 1 line
copy sentence from docstring
........
r67863 | benjamin.peterson | 2008-12-19 20:51:26 -0600 (Fri, 19 Dec 2008) | 1 line
add headings
........
r67864 | benjamin.peterson | 2008-12-19 20:57:19 -0600 (Fri, 19 Dec 2008) | 1 line
beef up docstring
........
r67880 | benjamin.peterson | 2008-12-20 16:49:24 -0600 (Sat, 20 Dec 2008) | 1 line
remove redundant sentence
........
r67882 | benjamin.peterson | 2008-12-20 16:59:49 -0600 (Sat, 20 Dec 2008) | 1 line
add some recent releases to the list
........
PyUnicode_AsStringAndSize -> _PyUnicode_AsStringAndSize to mark
them for interpreter internal use only.
We'll have to rework these APIs or create new ones for the
purpose of accessing the UTF-8 representation of Unicode objects
for 3.1.
The repr() of a string now contains printable Unicode characters unescaped.
The new ascii() builtin can be used to get a repr() with only ASCII characters in it.
PEP and patch were written by Atsuo Ishimoto.
svn+ssh://pythondev@svn.python.org/python/trunk
........
r60246 | guido.van.rossum | 2008-01-24 18:58:05 +0100 (Thu, 24 Jan 2008) | 2 lines
Fix test67.py from issue #1303614.
........
r60248 | raymond.hettinger | 2008-01-24 19:05:54 +0100 (Thu, 24 Jan 2008) | 1 line
Clean-up and speed-up code by accessing numerator/denominator directly. There's no reason to enforce readonliness
........
r60249 | raymond.hettinger | 2008-01-24 19:12:23 +0100 (Thu, 24 Jan 2008) | 1 line
Revert 60189 and restore performance.
........
r60250 | guido.van.rossum | 2008-01-24 19:21:02 +0100 (Thu, 24 Jan 2008) | 5 lines
News about recently fixed crashers:
- A few crashers fixed: weakref_in_del.py (issue #1377858);
loosing_dict_ref.py (issue #1303614, test67.py);
borrowed_ref_[34].py (not in tracker).
........
r60252 | thomas.heller | 2008-01-24 19:36:27 +0100 (Thu, 24 Jan 2008) | 7 lines
Use a PyDictObject again for the array type cache; retrieving items
from the WeakValueDictionary was slower by nearly a factor of 3.
To avoid leaks, weakref proxies for the array types are put into the
cache dict, with weakref callbacks that removes the entries when the
type goes away.
........
r60253 | thomas.heller | 2008-01-24 19:54:12 +0100 (Thu, 24 Jan 2008) | 2 lines
Replace Py_BuildValue with PyTuple_Pack because it is faster.
Also add a missing DECREF.
........
r60254 | raymond.hettinger | 2008-01-24 20:05:29 +0100 (Thu, 24 Jan 2008) | 1 line
Add support for trunc().
........
r60255 | thomas.heller | 2008-01-24 20:15:02 +0100 (Thu, 24 Jan 2008) | 5 lines
Invert the checks in get_[u]long and get_[u]longlong. The intent was
to not accept float types; the result was that integer-like objects
were not accepted.
Ported from release25-maint.
........
r60256 | raymond.hettinger | 2008-01-24 20:30:19 +0100 (Thu, 24 Jan 2008) | 1 line
Add support for int(r) just like the other numeric classes.
........
r60263 | raymond.hettinger | 2008-01-24 22:23:58 +0100 (Thu, 24 Jan 2008) | 1 line
Expand tests to include nested graph structures.
........
r60264 | raymond.hettinger | 2008-01-24 22:47:56 +0100 (Thu, 24 Jan 2008) | 1 line
Shorter pprint's for empty sets and frozensets. Fix indentation of frozensets. Add tests including two complex data structures.
........
r60265 | amaury.forgeotdarc | 2008-01-24 23:51:18 +0100 (Thu, 24 Jan 2008) | 14 lines
#1920: when considering a block starting by "while 0", the compiler optimized the
whole construct away, even when an 'else' clause is present::
while 0:
print("no")
else:
print("yes")
did not generate any code at all.
Now the compiler emits the 'else' block, like it already does for 'if' statements.
Will backport.
........
r60266 | amaury.forgeotdarc | 2008-01-24 23:59:25 +0100 (Thu, 24 Jan 2008) | 2 lines
News entry for r60265 (Issue 1920).
........
r60269 | raymond.hettinger | 2008-01-25 00:50:26 +0100 (Fri, 25 Jan 2008) | 1 line
More code cleanup. Remove unnecessary indirection to useless class methods.
........
r60270 | raymond.hettinger | 2008-01-25 01:21:54 +0100 (Fri, 25 Jan 2008) | 1 line
Add support for copy, deepcopy, and pickle.
........
r60271 | raymond.hettinger | 2008-01-25 01:33:45 +0100 (Fri, 25 Jan 2008) | 1 line
Mark todos and review comments.
........
r60272 | raymond.hettinger | 2008-01-25 02:13:12 +0100 (Fri, 25 Jan 2008) | 1 line
Add one other review comment.
........
r60273 | raymond.hettinger | 2008-01-25 02:23:38 +0100 (Fri, 25 Jan 2008) | 1 line
Fix-up signature for approximation.
........
r60274 | raymond.hettinger | 2008-01-25 02:46:33 +0100 (Fri, 25 Jan 2008) | 1 line
More design notes
........
r60276 | neal.norwitz | 2008-01-25 07:37:23 +0100 (Fri, 25 Jan 2008) | 6 lines
Make the test more robust by trying to reconnect up to 3 times
in case there were transient failures. This will hopefully silence
the buildbots for this test. As we find other tests that have a problem,
we can fix with a similar strategy assuming it is successful. It worked
on my box in a loop for 10+ runs where it would have an exception otherwise.
........
r60277 | neal.norwitz | 2008-01-25 09:04:16 +0100 (Fri, 25 Jan 2008) | 4 lines
Add prototypes to get the mathmodule.c to compile on OSF1 5.1 (Tru64)
and eliminate a compiler warning in floatobject.c. There might be
a better way to go about this, but it should be good enough for now.
........
svn+ssh://pythondev@svn.python.org/python/trunk
........
r59923 | raymond.hettinger | 2008-01-11 19:04:55 +0100 (Fri, 11 Jan 2008) | 1 line
Speed-up and simplify code urlparse's result objects.
........
r59924 | andrew.kuchling | 2008-01-11 20:33:24 +0100 (Fri, 11 Jan 2008) | 1 line
Bug #1790: update link; remove outdated paragraph
........
r59925 | thomas.heller | 2008-01-11 20:34:06 +0100 (Fri, 11 Jan 2008) | 5 lines
Raise an error instead of crashing with a segfault when a NULL
function pointer is called.
Will backport to release25-maint.
........
r59927 | thomas.heller | 2008-01-11 21:29:19 +0100 (Fri, 11 Jan 2008) | 4 lines
Fix a potential 'SystemError: NULL result without error'.
NULL may be a valid return value from PyLong_AsVoidPtr.
Will backport to release25-maint.
........
r59928 | raymond.hettinger | 2008-01-12 00:25:18 +0100 (Sat, 12 Jan 2008) | 1 line
Update the opcode docs for STORE_MAP and BUILD_MAP
........
r59929 | mark.dickinson | 2008-01-12 02:56:00 +0100 (Sat, 12 Jan 2008) | 4 lines
Issue 1780: Allow leading and trailing whitespace in Decimal constructor,
when constructing from a string. Disallow trailing newlines in
Context.create_decimal.
........
r59930 | georg.brandl | 2008-01-12 11:53:29 +0100 (Sat, 12 Jan 2008) | 3 lines
Move OSError docs to exceptions doc, remove obsolete descriptions
from os docs, rework posix docs.
........
r59931 | georg.brandl | 2008-01-12 14:47:57 +0100 (Sat, 12 Jan 2008) | 3 lines
Patch #1700288: Method cache optimization, by Armin Rigo, ported to
2.6 by Kevin Jacobs.
........
r59932 | georg.brandl | 2008-01-12 17:11:09 +0100 (Sat, 12 Jan 2008) | 2 lines
Fix editing glitch.
........
Multiple reinitializations of Python 3.0 failed on a system without a hardcoded default fs encoding. The patch makes sure that the default fs encoding is freed and reset to NULL on e.g. Linux.
I've also taken the liberty to increase the debugging in Objects/object.c:_Py_ForgetReference(). The method is used to validate the reference chain.
Reinitialization still fails in the 3rd round of my test suite:
* ob
object : <refcnt 0 at 0x821c840>
type : str
refcount: 0
address : 0x821c840
* op->_ob_prev->_ob_next
object : <refcnt 0 at 0x821c840>
type : str
refcount: 0
address : 0x821c840
* op->_ob_next->_ob_prev
object : bytearray(b'')
type : bytearray
refcount: 1
address : 0x826b838
Fatal Python error: UNREF invalid object
without calling PyType_Ready().
Question 1: Should the interpreter register all types with PyType_Ready()?
Many types seem to avoid it.
Question 2: To reproduce the problem, run the following code:
def f():
while True:
for a in iter(range(0,1,10**20)):
pass
f()
And watch the memory used by the process.
How do we test this in a unittest?
Barry Warsaw
Benjamin Peterson
Antoine Pitrou
Victor Stinner
Stefan Krah
Georg Brandl
Gregory P. Smith
Victor Stinner
Martin v. Löwis
Brian Curtin
Raymond Hettinger
Éric Araujo
Mark Dickinson
Skip Montanaro
Amaury Forgeot d'Arc
Nick Coghlan
Christian Heimes
Marc-André Lemburg