variations of the type struct and its attachments. In Py3k, all type
structs have to have all fields -- no binary backwards compatibility.
Had to change the complex object to a new-style number!
merge in bad ways, so I'll have to merge that extra-carefully (probably manually.)
Merged revisions 46495-46605 via svnmerge from
svn+ssh://pythondev@svn.python.org/python/trunk
........
r46495 | tim.peters | 2006-05-28 03:52:38 +0200 (Sun, 28 May 2006) | 2 lines
Added missing svn:eol-style property to text files.
........
r46497 | tim.peters | 2006-05-28 12:41:29 +0200 (Sun, 28 May 2006) | 3 lines
PyErr_Display(), PyErr_WriteUnraisable(): Coverity found a cut-and-paste
bug in both: `className` was referenced before being checked for NULL.
........
r46499 | fredrik.lundh | 2006-05-28 14:06:46 +0200 (Sun, 28 May 2006) | 5 lines
needforspeed: added Py_MEMCPY macro (currently tuned for Visual C only),
and use it for string copy operations. this gives a 20% speedup on some
string benchmarks.
........
r46501 | michael.hudson | 2006-05-28 17:51:40 +0200 (Sun, 28 May 2006) | 26 lines
Quality control, meet exceptions.c.
Fix a number of problems with the need for speed code:
One is doing this sort of thing:
Py_DECREF(self->field);
self->field = newval;
Py_INCREF(self->field);
without being very sure that self->field doesn't start with a
value that has a __del__, because that almost certainly can lead
to segfaults.
As self->args is constrained to be an exact tuple we may as well
exploit this fact consistently. This leads to quite a lot of
simplification (and, hey, probably better performance).
Add some error checking in places lacking it.
Fix some rather strange indentation in the Unicode code.
Delete some trailing whitespace.
More to come, I haven't fixed all the reference leaks yet...
........
r46502 | george.yoshida | 2006-05-28 18:39:09 +0200 (Sun, 28 May 2006) | 3 lines
Patch #1080727: add "encoding" parameter to doctest.DocFileSuite
Contributed by Bjorn Tillenius.
........
r46503 | martin.v.loewis | 2006-05-28 18:57:38 +0200 (Sun, 28 May 2006) | 4 lines
Rest of patch #1490384: Commit icon source, remove
claim that Erik von Blokland is the author of the
installer picture.
........
r46504 | michael.hudson | 2006-05-28 19:40:29 +0200 (Sun, 28 May 2006) | 16 lines
Quality control, meet exceptions.c, round two.
Make some functions that should have been static static.
Fix a bunch of refleaks by fixing the definition of
MiddlingExtendsException.
Remove all the __new__ implementations apart from
BaseException_new. Rewrite most code that needs it to cope with
NULL fields (such code could get excercised anyway, the
__new__-removal just makes it more likely). This involved
editing the code for WindowsError, which I can't test.
This fixes all the refleaks in at least the start of a regrtest
-R :: run.
........
r46505 | marc-andre.lemburg | 2006-05-28 19:46:58 +0200 (Sun, 28 May 2006) | 10 lines
Initial version of systimes - a module to provide platform dependent
performance measurements.
The module is currently just a proof-of-concept implementation, but
will integrated into pybench once it is stable enough.
License: pybench license.
Author: Marc-Andre Lemburg.
........
r46507 | armin.rigo | 2006-05-28 21:13:17 +0200 (Sun, 28 May 2006) | 15 lines
("Forward-port" of r46506)
Remove various dependencies on dictionary order in the standard library
tests, and one (clearly an oversight, potentially critical) in the
standard library itself - base64.py.
Remaining open issues:
* test_extcall is an output test, messy to make robust
* tarfile.py has a potential bug here, but I'm not familiar
enough with this code. Filed in as SF bug #1496501.
* urllib2.HTTPPasswordMgr() returns a random result if there is more
than one matching root path. I'm asking python-dev for
clarification...
........
r46508 | georg.brandl | 2006-05-28 22:11:45 +0200 (Sun, 28 May 2006) | 4 lines
The empty string is a valid import path.
(fixes#1496539)
........
r46509 | georg.brandl | 2006-05-28 22:23:12 +0200 (Sun, 28 May 2006) | 3 lines
Patch #1496206: urllib2 PasswordMgr ./. default ports
........
r46510 | georg.brandl | 2006-05-28 22:57:09 +0200 (Sun, 28 May 2006) | 3 lines
Fix refleaks in UnicodeError get and set methods.
........
r46511 | michael.hudson | 2006-05-28 23:19:03 +0200 (Sun, 28 May 2006) | 3 lines
use the UnicodeError traversal and clearing functions in UnicodeError
subclasses.
........
r46512 | thomas.wouters | 2006-05-28 23:32:12 +0200 (Sun, 28 May 2006) | 4 lines
Make last patch valid C89 so Windows compilers can deal with it.
........
r46513 | georg.brandl | 2006-05-28 23:42:54 +0200 (Sun, 28 May 2006) | 3 lines
Fix ref-antileak in _struct.c which eventually lead to deallocating None.
........
r46514 | georg.brandl | 2006-05-28 23:57:35 +0200 (Sun, 28 May 2006) | 4 lines
Correct None refcount issue in Mac modules. (Are they
still used?)
........
r46515 | armin.rigo | 2006-05-29 00:07:08 +0200 (Mon, 29 May 2006) | 3 lines
A clearer error message when passing -R to regrtest.py with
release builds of Python.
........
r46516 | georg.brandl | 2006-05-29 00:14:04 +0200 (Mon, 29 May 2006) | 3 lines
Fix C function calling conventions in _sre module.
........
r46517 | georg.brandl | 2006-05-29 00:34:51 +0200 (Mon, 29 May 2006) | 3 lines
Convert audioop over to METH_VARARGS.
........
r46518 | georg.brandl | 2006-05-29 00:38:57 +0200 (Mon, 29 May 2006) | 3 lines
METH_NOARGS functions do get called with two args.
........
r46519 | georg.brandl | 2006-05-29 11:46:51 +0200 (Mon, 29 May 2006) | 4 lines
Fix refleak in socketmodule. Replace bogus Py_BuildValue calls.
Fix refleak in exceptions.
........
r46520 | nick.coghlan | 2006-05-29 14:43:05 +0200 (Mon, 29 May 2006) | 7 lines
Apply modified version of Collin Winter's patch #1478788
Renames functional extension module to _functools and adds a Python
functools module so that utility functions like update_wrapper can be
added easily.
........
r46522 | georg.brandl | 2006-05-29 15:53:16 +0200 (Mon, 29 May 2006) | 3 lines
Convert fmmodule to METH_VARARGS.
........
r46523 | georg.brandl | 2006-05-29 16:13:21 +0200 (Mon, 29 May 2006) | 3 lines
Fix#1494605.
........
r46524 | georg.brandl | 2006-05-29 16:28:05 +0200 (Mon, 29 May 2006) | 3 lines
Handle PyMem_Malloc failure in pystrtod.c. Closes#1494671.
........
r46525 | georg.brandl | 2006-05-29 16:33:55 +0200 (Mon, 29 May 2006) | 3 lines
Fix compiler warning.
........
r46526 | georg.brandl | 2006-05-29 16:39:00 +0200 (Mon, 29 May 2006) | 3 lines
Fix#1494787 (pyclbr counts whitespace as superclass name)
........
r46527 | bob.ippolito | 2006-05-29 17:47:29 +0200 (Mon, 29 May 2006) | 1 line
simplify the struct code a bit (no functional changes)
........
r46528 | armin.rigo | 2006-05-29 19:59:47 +0200 (Mon, 29 May 2006) | 2 lines
Silence a warning.
........
r46529 | georg.brandl | 2006-05-29 21:39:45 +0200 (Mon, 29 May 2006) | 3 lines
Correct some value converting strangenesses.
........
r46530 | nick.coghlan | 2006-05-29 22:27:44 +0200 (Mon, 29 May 2006) | 1 line
When adding a module like functools, it helps to let SVN know about the file.
........
r46531 | georg.brandl | 2006-05-29 22:52:54 +0200 (Mon, 29 May 2006) | 4 lines
Patches #1497027 and #972322: try HTTP digest auth first,
and watch out for handler name collisions.
........
r46532 | georg.brandl | 2006-05-29 22:57:01 +0200 (Mon, 29 May 2006) | 3 lines
Add News entry for last commit.
........
r46533 | georg.brandl | 2006-05-29 23:04:52 +0200 (Mon, 29 May 2006) | 4 lines
Make use of METH_O and METH_NOARGS where possible.
Use Py_UnpackTuple instead of PyArg_ParseTuple where possible.
........
r46534 | georg.brandl | 2006-05-29 23:58:42 +0200 (Mon, 29 May 2006) | 3 lines
Convert more modules to METH_VARARGS.
........
r46535 | georg.brandl | 2006-05-30 00:00:30 +0200 (Tue, 30 May 2006) | 3 lines
Whoops.
........
r46536 | fredrik.lundh | 2006-05-30 00:42:07 +0200 (Tue, 30 May 2006) | 4 lines
fixed "abc".count("", 100) == -96 error (hopefully, nobody's relying on
the current behaviour ;-)
........
r46537 | bob.ippolito | 2006-05-30 00:55:48 +0200 (Tue, 30 May 2006) | 1 line
struct: modulo math plus warning on all endian-explicit formats for compatibility with older struct usage (ugly)
........
r46539 | bob.ippolito | 2006-05-30 02:26:01 +0200 (Tue, 30 May 2006) | 1 line
Add a length check to aifc to ensure it doesn't write a bogus file
........
r46540 | tim.peters | 2006-05-30 04:25:25 +0200 (Tue, 30 May 2006) | 10 lines
deprecated_err(): Stop bizarre warning messages when the tests
are run in the order:
test_genexps (or any other doctest-based test)
test_struct
test_doctest
The `warnings` module needs an advertised way to save/restore
its internal filter list.
........
r46541 | tim.peters | 2006-05-30 04:26:46 +0200 (Tue, 30 May 2006) | 2 lines
Whitespace normalization.
........
r46542 | tim.peters | 2006-05-30 04:30:30 +0200 (Tue, 30 May 2006) | 2 lines
Set a binary svn:mime-type property on this UTF-8 encoded file.
........
r46543 | neal.norwitz | 2006-05-30 05:18:50 +0200 (Tue, 30 May 2006) | 1 line
Simplify further by using AddStringConstant
........
r46544 | tim.peters | 2006-05-30 06:16:25 +0200 (Tue, 30 May 2006) | 6 lines
Convert relevant dict internals to Py_ssize_t.
I don't have a box with nearly enough RAM, or an OS,
that could get close to tickling this, though (requires
a dict w/ at least 2**31 entries).
........
r46545 | neal.norwitz | 2006-05-30 06:19:21 +0200 (Tue, 30 May 2006) | 1 line
Remove stray | in comment
........
r46546 | neal.norwitz | 2006-05-30 06:25:05 +0200 (Tue, 30 May 2006) | 1 line
Use Py_SAFE_DOWNCAST for safety. Fix format strings. Remove 2 more stray | in comment
........
r46547 | neal.norwitz | 2006-05-30 06:43:23 +0200 (Tue, 30 May 2006) | 1 line
No DOWNCAST is required since sizeof(Py_ssize_t) >= sizeof(int) and Py_ReprEntr returns an int
........
r46548 | tim.peters | 2006-05-30 07:04:59 +0200 (Tue, 30 May 2006) | 3 lines
dict_print(): Explicitly narrow the return value
from a (possibly) wider variable.
........
r46549 | tim.peters | 2006-05-30 07:23:59 +0200 (Tue, 30 May 2006) | 5 lines
dict_print(): So that Neal & I don't spend the rest of
our lives taking turns rewriting code that works ;-),
get rid of casting illusions by declaring a new variable
with the obvious type.
........
r46550 | georg.brandl | 2006-05-30 09:04:55 +0200 (Tue, 30 May 2006) | 3 lines
Restore exception pickle support. #1497319.
........
r46551 | georg.brandl | 2006-05-30 09:13:29 +0200 (Tue, 30 May 2006) | 3 lines
Add a test case for exception pickling. args is never NULL.
........
r46552 | neal.norwitz | 2006-05-30 09:21:10 +0200 (Tue, 30 May 2006) | 1 line
Don't fail if the (sub)pkgname already exist.
........
r46553 | georg.brandl | 2006-05-30 09:34:45 +0200 (Tue, 30 May 2006) | 3 lines
Disallow keyword args for exceptions.
........
r46554 | neal.norwitz | 2006-05-30 09:36:54 +0200 (Tue, 30 May 2006) | 5 lines
I'm impatient. I think this will fix a few more problems with the buildbots.
I'm not sure this is the best approach, but I can't think of anything better.
If this creates problems, feel free to revert, but I think it's safe and
should make things a little better.
........
r46555 | georg.brandl | 2006-05-30 10:17:00 +0200 (Tue, 30 May 2006) | 4 lines
Do the check for no keyword arguments in __init__ so that
subclasses of Exception can be supplied keyword args
........
r46556 | georg.brandl | 2006-05-30 10:47:19 +0200 (Tue, 30 May 2006) | 3 lines
Convert test_exceptions to unittest.
........
r46557 | andrew.kuchling | 2006-05-30 14:52:01 +0200 (Tue, 30 May 2006) | 1 line
Add SoC name, and reorganize this section a bit
........
r46559 | tim.peters | 2006-05-30 17:53:34 +0200 (Tue, 30 May 2006) | 11 lines
PyLong_FromString(): Continued fraction analysis (explained in
a new comment) suggests there are almost certainly large input
integers in all non-binary input bases for which one Python digit
too few is initally allocated to hold the final result. Instead
of assert-failing when that happens, allocate more space. Alas,
I estimate it would take a few days to find a specific such case,
so this isn't backed up by a new test (not to mention that such
a case may take hours to run, since conversion time is quadratic
in the number of digits, and preliminary attempts suggested that
the smallest such inputs contain at least a million digits).
........
r46560 | fredrik.lundh | 2006-05-30 19:11:48 +0200 (Tue, 30 May 2006) | 3 lines
changed find/rfind to return -1 for matches outside the source string
........
r46561 | bob.ippolito | 2006-05-30 19:37:54 +0200 (Tue, 30 May 2006) | 1 line
Change wrapping terminology to overflow masking
........
r46562 | fredrik.lundh | 2006-05-30 19:39:58 +0200 (Tue, 30 May 2006) | 3 lines
changed count to return 0 for slices outside the source string
........
r46568 | tim.peters | 2006-05-31 01:28:02 +0200 (Wed, 31 May 2006) | 2 lines
Whitespace normalization.
........
r46569 | brett.cannon | 2006-05-31 04:19:54 +0200 (Wed, 31 May 2006) | 5 lines
Clarify wording on default values for strptime(); defaults are used when better
values cannot be inferred.
Closes bug #1496315.
........
r46572 | neal.norwitz | 2006-05-31 09:43:27 +0200 (Wed, 31 May 2006) | 1 line
Calculate smallest properly (it was off by one) and use proper ssize_t types for Win64
........
r46573 | neal.norwitz | 2006-05-31 10:01:08 +0200 (Wed, 31 May 2006) | 1 line
Revert last checkin, it is better to do make distclean
........
r46574 | neal.norwitz | 2006-05-31 11:02:44 +0200 (Wed, 31 May 2006) | 3 lines
On 64-bit platforms running test_struct after test_tarfile would fail
since the deprecation warning wouldn't be raised.
........
r46575 | thomas.heller | 2006-05-31 13:37:58 +0200 (Wed, 31 May 2006) | 3 lines
PyTuple_Pack is not available in Python 2.3, but ctypes must stay
compatible with that.
........
r46576 | andrew.kuchling | 2006-05-31 15:18:56 +0200 (Wed, 31 May 2006) | 1 line
'functional' module was renamed to 'functools'
........
r46577 | kristjan.jonsson | 2006-05-31 15:35:41 +0200 (Wed, 31 May 2006) | 1 line
Fixup the PCBuild8 project directory. exceptions.c have moved to Objects, and the functionalmodule.c has been replaced with _functoolsmodule.c. Other minor changes to .vcproj files and .sln to fix compilation
........
r46578 | andrew.kuchling | 2006-05-31 16:08:48 +0200 (Wed, 31 May 2006) | 15 lines
[Bug #1473048]
SimpleXMLRPCServer and DocXMLRPCServer don't look at
the path of the HTTP request at all; you can POST or
GET from / or /RPC2 or /blahblahblah with the same results.
Security scanners that look for /cgi-bin/phf will therefore report
lots of vulnerabilities.
Fix: add a .rpc_paths attribute to the SimpleXMLRPCServer class,
and report a 404 error if the path isn't on the allowed list.
Possibly-controversial aspect of this change: the default makes only
'/' and '/RPC2' legal. Maybe this will break people's applications
(though I doubt it). We could just set the default to an empty tuple,
which would exactly match the current behaviour.
........
r46579 | andrew.kuchling | 2006-05-31 16:12:47 +0200 (Wed, 31 May 2006) | 1 line
Mention SimpleXMLRPCServer change
........
r46580 | tim.peters | 2006-05-31 16:28:07 +0200 (Wed, 31 May 2006) | 2 lines
Trimmed trailing whitespace.
........
r46581 | tim.peters | 2006-05-31 17:33:22 +0200 (Wed, 31 May 2006) | 4 lines
_range_error(): Speed and simplify (there's no real need for
loops here). Assert that size_t is actually big enough, and
that f->size is at least one. Wrap a long line.
........
r46582 | tim.peters | 2006-05-31 17:34:37 +0200 (Wed, 31 May 2006) | 2 lines
Repaired error in new comment.
........
r46584 | neal.norwitz | 2006-06-01 07:32:49 +0200 (Thu, 01 Jun 2006) | 4 lines
Remove ; at end of macro. There was a compiler recently that warned
about extra semi-colons. It may have been the HP C compiler.
This file will trigger a bunch of those warnings now.
........
r46585 | georg.brandl | 2006-06-01 08:39:19 +0200 (Thu, 01 Jun 2006) | 3 lines
Correctly unpickle 2.4 exceptions via __setstate__ (patch #1498571)
........
r46586 | georg.brandl | 2006-06-01 10:27:32 +0200 (Thu, 01 Jun 2006) | 3 lines
Correctly allocate complex types with tp_alloc. (bug #1498638)
........
r46587 | georg.brandl | 2006-06-01 14:30:46 +0200 (Thu, 01 Jun 2006) | 2 lines
Correctly dispatch Faults in loads (patch #1498627)
........
r46588 | georg.brandl | 2006-06-01 15:00:49 +0200 (Thu, 01 Jun 2006) | 3 lines
Some code style tweaks, and remove apply.
........
r46589 | armin.rigo | 2006-06-01 15:19:12 +0200 (Thu, 01 Jun 2006) | 5 lines
[ 1497053 ] Let dicts propagate the exceptions in user __eq__().
[ 1456209 ] dictresize() vulnerability ( <- backport candidate ).
........
r46590 | tim.peters | 2006-06-01 15:41:46 +0200 (Thu, 01 Jun 2006) | 2 lines
Whitespace normalization.
........
r46591 | tim.peters | 2006-06-01 15:49:23 +0200 (Thu, 01 Jun 2006) | 2 lines
Record bugs 1275608 and 1456209 as being fixed.
........
r46592 | tim.peters | 2006-06-01 15:56:26 +0200 (Thu, 01 Jun 2006) | 5 lines
Re-enable a new empty-string test added during the NFS sprint,
but disabled then because str and unicode strings gave different
results. The implementations were repaired later during the
sprint, but the new test remained disabled.
........
r46594 | tim.peters | 2006-06-01 17:50:44 +0200 (Thu, 01 Jun 2006) | 7 lines
Armin committed his patch while I was reviewing it (I'm sure
he didn't know this), so merged in some changes I made during
review. Nothing material apart from changing a new `mask` local
from int to Py_ssize_t. Mostly this is repairing comments that
were made incorrect, and adding new comments. Also a few
minor code rewrites for clarity or helpful succinctness.
........
r46599 | neal.norwitz | 2006-06-02 06:45:53 +0200 (Fri, 02 Jun 2006) | 1 line
Convert docstrings to comments so regrtest -v prints method names
........
r46600 | neal.norwitz | 2006-06-02 06:50:49 +0200 (Fri, 02 Jun 2006) | 2 lines
Fix memory leak found by valgrind.
........
r46601 | neal.norwitz | 2006-06-02 06:54:52 +0200 (Fri, 02 Jun 2006) | 1 line
More memory leaks from valgrind
........
r46602 | neal.norwitz | 2006-06-02 08:23:00 +0200 (Fri, 02 Jun 2006) | 11 lines
Patch #1357836:
Prevent an invalid memory read from test_coding in case the done flag is set.
In that case, the loop isn't entered. I wonder if rather than setting
the done flag in the cases before the loop, if they should just exit early.
This code looks like it should be refactored.
Backport candidate (also the early break above if decoding_fgets fails)
........
r46603 | martin.blais | 2006-06-02 15:03:43 +0200 (Fri, 02 Jun 2006) | 1 line
Fixed struct test to not use unittest.
........
r46605 | tim.peters | 2006-06-03 01:22:51 +0200 (Sat, 03 Jun 2006) | 10 lines
pprint functions used to sort a dict (by key) if and only if
the output required more than one line. "Small" dicts got
displayed in seemingly random order (the hash-induced order
produced by dict.__repr__). None of this was documented.
Now pprint functions always sort dicts by key, and the docs
promise it.
This was proposed and agreed to during the PyCon 2006 core
sprint -- I just didn't have time for it before now.
........
number of tests, all because of the codecs/_multibytecodecs issue described
here (it's not a Py3K issue, just something Py3K discovers):
http://mail.python.org/pipermail/python-dev/2006-April/064051.html
Hye-Shik Chang promised to look for a fix, so no need to fix it here. The
tests that are expected to break are:
test_codecencodings_cn
test_codecencodings_hk
test_codecencodings_jp
test_codecencodings_kr
test_codecencodings_tw
test_codecs
test_multibytecodec
This merge fixes an actual test failure (test_weakref) in this branch,
though, so I believe merging is the right thing to do anyway.
PyObject_Unicode(). This problem was originally reported from Coverity
and addresses mail on python-dev "checkin r43015".
This inlines the conversion of the string to unicode and cleans
up/simplifies some code at the end of the PyObject_Unicode().
We really need a complete C API test module for all public APIs
and passing good and bad parameter values.
Will backport.
In C++, it's an error to pass a string literal to a char* function
without a const_cast(). Rather than require every C++ extension
module to put a cast around string literals, fix the API to state the
const-ness.
I focused on parts of the API where people usually pass literals:
PyArg_ParseTuple() and friends, Py_BuildValue(), PyMethodDef, the type
slots, etc. Predictably, there were a large set of functions that
needed to be fixed as a result of these changes. The most pervasive
change was to make the keyword args list passed to
PyArg_ParseTupleAndKewords() to be a const char *kwlist[].
One cast was required as a result of the changes: A type object
mallocs the memory for its tp_doc slot and later frees it.
PyTypeObject says that tp_doc is const char *; but if the type was
created by type_new(), we know it is safe to cast to char *.
conversion using the proper magic slot (e.g., __int__()). Also move conversion
code out of PyNumber_*() functions in the C API into the nb_* function.
Applied patch #1109424. Thanks Walter Doewald.
- weakref.ref and weakref.ReferenceType will become aliases for each
other
- weakref.ref will be a modern, new-style class with proper __new__
and __init__ methods
- weakref.WeakValueDictionary will have a lighter memory footprint,
using a new weakref.ref subclass to associate the key with the
value, allowing us to have only a single object of overhead for each
dictionary entry (currently, there are 3 objects of overhead per
entry: a weakref to the value, a weakref to the dictionary, and a
function object used as a weakref callback; the weakref to the
dictionary could be avoided without this change)
- a new macro, PyWeakref_CheckRefExact(), will be added
- PyWeakref_CheckRef() will check for subclasses of weakref.ref
This closes SF patch #983019.
The builtin eval() function now accepts any mapping for the locals argument.
Time sensitive steps guarded by PyDict_CheckExact() to keep from slowing
down the normal case. My timings so no measurable impact.
[ 784825 ] fix obscure crash in descriptor handling
Should be applied to release23-maint and in all likelyhood
release22-maint, too.
Certainly doesn't apply to release21-maint.
I'm finding some pretty baffling output, like reprs consisting entirely
of three left parens. At least this will let us know what type the object
is (it's not str -- there's no quote character in the repr).
New tool combinerefs.py, to combine the two output blocks produced via
PYTHONDUMPREFS.
new line.
New pvt API function _Py_PrintReferenceAddresses(): Prints only the
addresses and refcnts of the live objects. This is always safe to call,
because it has no dependence on Python's C API.
Py_Finalize(): If envar PYTHONDUMPREFS is set, call (the new)
_Py_PrintReferenceAddresses() right before dumping final pymalloc stats.
We can't print the reprs of the objects here because too much of the
interpreter has been shut down. You need to correlate the addresses
displayed here with the object reprs printed by the earlier
PYTHONDUMPREFS call to _Py_PrintReferences().
even farther down, to just before the call to
_PyObject_DebugMallocStats(). This required the following changes:
- pystate.c, PyThreadState_GetDict(): changed not to raise an
exception or issue a fatal error when no current thread state is
available, but simply return NULL without raising an exception
(ever).
- object.c, Py_ReprEnter(): when PyThreadState_GetDict() returns NULL,
don't raise an exception but return 0. This means that when
printing a container that's recursive, printing will go on and on
and on. But that shouldn't happen in the case we care about (see
first bullet).
- Updated Misc/NEWS and Doc/api/init.tex to reflect changes to
PyThreadState_GetDict() definition.
Arranged that all the objects exposed by __builtin__ appear in the list
of all objects. I basically peed away two days tracking down a mystery
leak in sys.gettotalrefcount() in a ZODB app (== tons of code), because
the object leaking the references didn't appear in the sys.getobjects(0)
list. The object happened to be False. Now False is in the list, along
with other popular & previously missing leak candidates (like None).
Alas, we still don't have a choke point covering *all* Python objects,
so the list of all objects may still be incomplete.
_Py_AddToAllObjects() that simply inserts an object at the front of
the doubly-linked list of all objects. Changed PyType_Ready() (the
closest thing we've got to a choke point for type objects) to call
that.
a doubly-linked list, exposed by sys.getobjects(). Unfortunately, it's not
really all live objects, and it seems my fate to bump into programs where
sys.gettotalrefcount() keeps going up but where the reference leaks aren't
accounted for by anything in the list of all objects.
This patch helps a little: if COUNT_ALLOCS is also defined, from now on
type objects will also appear in this list, provided at least one object
of a type has been allocated.
Don't access tp_descr_{get,set} of a descriptor without checking the
flag bits of the descriptor's type. While we know that the main type
(the type of the object whose attribute is being accessed) has all the
right flag bits (or else PyObject_Generic{Get,Set}Attr wouldn't be
called), we don't know that for its class attributes!
Will backport to 2.2.
was broken because new-in-2.3 code added a tp_as_mapping slot to tuples.
Repaired that.
Added basic docs to check_recursion().
The code that intended to exempt tuples and strings was also broken here,
and in 2.2: these should use PyXYZ_CheckExact(), not PyXYZ_Check() -- we
can't know whether subclass instances are immutable. This part (and this
part alone) is a bugfix candidate.
Py_Init crash". refchain cannot be cleared because objects can live across
Py_Finalize() and Py_Initialize() if they are kept alive by circular
references.
macros. The 'op' argument is then the result from PyObject_MALLOC,
and that can of course be NULL. In that case, PyObject_Init[Var]
would raise a SystemError with "NULL object passed to
PyObject_Init[Var]". But there's nothing the caller of the macro can
do about this. So PyObject_Init[Var] should call just PyErr_NoMemory.
Will backport.
Guido van Rossum
Thomas Wouters
Neal Norwitz
Georg Brandl
Martin v. Löwis
Jeremy Hylton
Neil Schemenauer
Brett Cannon
Armin Rigo
Tim Peters
Michael W. Hudson
Raymond Hettinger
Fred Drake
Thomas Heller
Jack Jansen