Commit Graph

86 Commits

Author SHA1 Message Date
Victor Stinner 615069eb08
bpo-44184: Fix subtype_dealloc() for freed type (GH-26274)
Fix a crash at Python exit when a deallocator function removes the
last strong reference to a heap type.

Don't read type memory after calling basedealloc() since
basedealloc() can deallocate the type and free its memory.

_PyMem_IsPtrFreed() argument is now constant.
2021-05-21 19:19:54 +02:00
Inada Naoki 3caea9adda
bpo-43651: PEP 597: Fix EncodingWarning in some tests (GH-25171)
* Fix test_float
* Fix _osx_support
* Fix test_fstring
* Fix test_gc
* Fix test_gzip
* Fix test_hashlib
* Fix unrelated whitespace issue

Co-authored-by: Ned Deily <nad@python.org>
2021-04-04 17:01:10 +09:00
Brandt Bucher c13b847a6f
bpo-41984: GC track all user classes (GH-22701) 2020-10-14 18:44:07 -07:00
Hai Shi 4660597b51
bpo-40275: Use new test.support helper submodules in tests (GH-21448) 2020-08-03 18:49:18 +02:00
Hai Shi e80697d687
bpo-40275: Adding threading_helper submodule in test.support (GH-20263) 2020-05-28 00:10:27 +02:00
Victor Stinner c6e5c1123b
bpo-39489: Remove COUNT_ALLOCS special build (GH-18259)
Remove:

* COUNT_ALLOCS macro
* sys.getcounts() function
* SHOW_ALLOC_COUNT code in listobject.c
* SHOW_TRACK_COUNT code in tupleobject.c
* PyConfig.show_alloc_count field
* -X showalloccount command line option
* @test.support.requires_type_collecting decorator
2020-02-03 15:17:15 +01:00
Pablo Galindo a2ec3f07f7
bpo-39322: Add gc.is_finalized to check if an object has been finalised by the gc (GH-17989) 2020-01-14 12:06:45 +00:00
Victor Stinner 6cac113666
bpo-38991: Remove test.support.strip_python_stderr() (GH-17490)
test.support: run_python_until_end(), assert_python_ok() and
assert_python_failure() functions no longer strip whitespaces from
stderr.
2019-12-08 08:38:16 +01:00
Neil Schemenauer 392a13bb93
bpo-38006: Add unit test for weakref clear bug (GH-16788) 2019-10-15 20:56:48 -07:00
Pablo Galindo 466326dcdf
bpo-38379: Don't block collection of unreachable objects when some objects resurrect (GH-16687)
Currently if any finalizer invoked during garbage collection resurrects any object, the gc gives up and aborts the collection. Although finalizers are assured to only run once per object, this behaviour of the gc can lead to an ever-increasing memory situation if new resurrecting objects are allocated in every new gc collection.

To avoid this, recompute what objects among the unreachable set need to be resurrected and what objects can be safely collected. In this way, resurrecting objects will not block the collection of other objects in the unreachable set.
2019-10-13 16:48:59 +01:00
Tim Peters ecbf35f933
bpo-38379: don't claim objects are collected when they aren't (#16658)
* bpo-38379:  when a finalizer resurrects an object,
nothing is actually collected in this run of gc.
Change the stats to relect that truth.
2019-10-09 12:37:30 -05:00
Victor Stinner 6876257eaa
bpo-36389: _PyObject_CheckConsistency() available in release mode (GH-16612)
bpo-36389, bpo-38376: The _PyObject_CheckConsistency() function is
now also available in release mode. For example, it can be used to
debug a crash in the visit_decref() function of the GC.

Modify the following functions to also work in release mode:

* _PyDict_CheckConsistency()
* _PyObject_CheckConsistency()
* _PyType_CheckConsistency()
* _PyUnicode_CheckConsistency()

Other changes:

* _PyMem_IsPtrFreed(ptr) now also returns 1 if ptr is NULL
  (equals to 0).
* _PyBytesWriter_CheckConsistency() now returns 1 and is only used
  with assert().
* Reorder _PyObject_Dump() to write safe fields first, and only
  attempt to render repr() at the end.
2019-10-07 18:42:01 +02:00
Min ho Kim 39d87b5471 Fix typos mostly in comments, docs and test names (GH-15209) 2019-08-30 16:21:19 -04:00
Pablo Galindo d60a79a101
bpo-36155: Check for identity on test_gc.test_get_objects (GH-12116) 2019-03-01 01:12:27 -08:00
Pablo Galindo 175421b58c bpo-36016: Add generation option to gc.getobjects() (GH-11909) 2019-02-23 12:02:06 +09:00
Victor Stinner 626bff8568
bpo-9263: Dump Python object on GC assertion failure (GH-10062)
Changes:

* Add _PyObject_AssertFailed() function.
* Add _PyObject_ASSERT() and _PyObject_ASSERT_WITH_MSG() macros.
* gc_decref(): replace assert() with _PyObject_ASSERT_WITH_MSG() to
  dump the faulty object if the assertion fails.

_PyObject_AssertFailed() calls:

* _PyMem_DumpTraceback(): try to log the traceback where the object
  memory has been allocated if tracemalloc is enabled.
* _PyObject_Dump(): log repr(obj).
* Py_FatalError(): log the current Python traceback.

_PyObject_AssertFailed() uses _PyObject_IsFreed() heuristic to check
if the object memory has been freed by a debug hook on Python memory
allocators.

Initial patch written by David Malcolm.

Co-Authored-By: David Malcolm <dmalcolm@redhat.com>
2018-10-25 17:31:10 +02:00
Zackery Spytz d8cba5d16f bpo-24596: Decref module in PyRun_SimpleFileExFlags() on SystemExit (GH-7918)
PyErr_Print() will not return when the exception is a SystemExit, so
decref the __main__ module object in that case.
2018-07-03 21:47:22 +02:00
Yury Selivanov 383b32fe10
Revert "bpo-31356: Add context manager to temporarily disable GC GH-5495
This reverts commit 72a0d218dc.

The reverted commit had a few issues so it was unanimously decided
to undo it. See the bpo issue for details.
2018-02-02 09:31:06 -05:00
Pablo Galindo 72a0d218dc bpo-31356: Add context manager to temporarily disable GC (GH-4224) 2018-01-29 12:37:09 -08:00
brainfvck c75edabbb6 bpo-31558: Add gc.freeze() (#3705)
Freeze all the objects tracked by gc - move them to a permanent generation
and ignore all the future collections. This can be used before a POSIX
fork() call to make the gc copy-on-write friendly or to speed up collection.
2017-10-16 12:49:41 -07:00
Antoine Pitrou a6a4dc816d bpo-31370: Remove support for threads-less builds (#3385)
* Remove Setup.config
* Always define WITH_THREAD for compatibility.
2017-09-07 18:56:24 +02:00
Serhiy Storchaka bae75cf3fe Issue #19527: Fixed tests with defined COUNT_ALLOCS. 2016-07-03 22:30:49 +03:00
Serhiy Storchaka a793037d80 Issue #19527: Fixed tests with defined COUNT_ALLOCS. 2016-07-03 22:27:26 +03:00
Victor Stinner 843a1fb1e7 test_gc: remove unused imports 2016-01-23 13:29:02 +01:00
Berker Peksag ce643913a9 Issue #9517: Move script_helper to the support package.
Patch by Christie Wilson.
2015-05-06 06:33:17 +03:00
Antoine Pitrou a63cc21234 Issue #23726: Don't enable GC for user subclasses of non-GC types that don't add any new fields.
Patch by Eugene Toder.
2015-04-13 20:10:06 +02:00
Serhiy Storchaka 263dcd20a3 Issue #23799: Added test.support.start_threads() for running and cleaning up
multiple threads.
2015-04-01 13:01:14 +03:00
Serhiy Storchaka 9db55004a1 Make some tests more frienly to MemoryError.
Free memory, unlock hanging threads.
2015-03-28 20:38:37 +02:00
Tim Peters 5fbc7b12f7 Issue #21435: Segfault in gc with cyclic trash
Changed the iteration logic in finalize_garbage() to tolerate objects vanishing
from the list as a side effect of executing a finalizer.
2014-05-08 17:42:19 -05:00
Serhiy Storchaka f28ba369dd Issue #20532: Tests which use _testcapi now are marked as CPython only. 2014-02-07 10:10:55 +02:00
Antoine Pitrou 796564c27b Issue #18112: PEP 442 implementation (safe object finalization). 2013-07-30 19:59:21 +02:00
Antoine Pitrou 5f454a07a0 Issue #1545463: Global variables caught in reference cycles are now garbage-collected at shutdown. 2013-05-06 21:15:57 +02:00
Antoine Pitrou d4156c1693 Issue #16351: New function gc.get_stats() returns per-generation collection statistics. 2012-10-30 22:43:19 +01:00
Antoine Pitrou 2b0218a259 Issue #13992: The trashcan mechanism is now thread-safe. This eliminates
sporadic crashes in multi-thread programs when several long deallocator
chains ran concurrently and involved subclasses of built-in container
types.

Because of this change, a couple extension modules compiled for 3.2.4
(those which use the trashcan mechanism, despite it being undocumented)
will not be loadable by 3.2.3 and earlier. However, extension modules
compiled for 3.2.3 and earlier will be loadable by 3.2.4.
2012-09-06 00:59:49 +02:00
Antoine Pitrou de3c73b57c pep8-ize test names 2012-04-16 21:29:58 +02:00
Antoine Pitrou 6b64fc6bff Fix refleak in test_gc 2012-04-16 21:29:02 +02:00
Kristján Valur Jónsson 69c635266e Issue #10576: Add a progress callback to gcmodule 2012-04-15 11:41:32 +00:00
Antoine Pitrou d11f76f8b9 Issue #11761: make tests for gc.get_count() less fragile 2011-04-04 19:52:56 +02:00
Antoine Pitrou e9b2a4cb59 Issue #11761: make tests for gc.get_count() less fragile 2011-04-04 19:51:33 +02:00
Antoine Pitrou b35f29a0e0 Issue #11761: make tests for gc.get_count() less fragile 2011-04-04 19:50:42 +02:00
Brett Cannon 7a54073a56 Issue #10992: make tests pass when run under coverage.
Various tests fail when run under coverage. A primary culprit is refcount tests
which fail as the counts are thrown off by the coverage code. A new decorator
-- test.support.refcount_test -- is used to decorate tests which test refcounts
and to skip them when running under coverage. Other tests simply fail because
of changes in the system (e.g., __local__ suddenly appearing).

Thanks to Kristian Vlaardingerbroek for helping to diagnose the test failures.
2011-02-22 03:04:06 +00:00
Brian Curtin 8291af2354 Fix ResourceWarning occuring on Windows. Close stdout/stderr pipes. 2010-11-01 16:40:17 +00:00
Georg Brandl 08be72d0aa Add a new warning gategory, ResourceWarning, as discussed on python-dev. It is silent by default,
except when configured --with-pydebug.

Emit this warning from the GC shutdown procedure, rather than just printing to stderr.
2010-10-24 15:11:22 +00:00
Antoine Pitrou aee47561fc Relax test condition (fix failures on FreeBSD buildbots) 2010-09-16 15:04:49 +00:00
Antoine Pitrou 2ed94eb520 Do not print additional shutdown message when gc.DEBUG_SAVEALL is set 2010-09-14 09:48:39 +00:00
Antoine Pitrou 696e03553b Issue #477863: Print a warning at shutdown if gc.garbage is not empty. 2010-08-08 22:18:46 +00:00
Benjamin Peterson 7fe73a17c5 Merged revisions 70837,70864,70878,71004,71032,71043 via svnmerge from
svn+ssh://pythondev@svn.python.org/python/trunk

........
  r70837 | gregory.p.smith | 2009-03-31 11:54:10 -0500 (Tue, 31 Mar 2009) | 9 lines

  The unittest.TestCase.assertEqual() now displays the differences in lists,
  tuples, dicts and sets on failure.

  Many new handy type and comparison specific assert* methods have been added
  that fail with error messages actually useful for debugging.  Contributed in
  by Google and completed with help from mfoord and GvR at PyCon 2009 sprints.

  Discussion lives in http://bugs.python.org/issue2578.
........
  r70864 | gregory.p.smith | 2009-03-31 14:03:28 -0500 (Tue, 31 Mar 2009) | 10 lines

  Rename the actual method definitions to the official assertFoo names.

  Adds unittests to make sure the old fail* names continue to work now
  and adds a comment that they are pending deprecation.

  Also adds a test to confirm that the plural Equals method variants
  continue to exist even though we're unlikely to deprecate those.

  http://bugs.python.org/issue2578
........
  r70878 | gregory.p.smith | 2009-03-31 14:59:14 -0500 (Tue, 31 Mar 2009) | 3 lines

  Issue an actual PendingDeprecationWarning for the TestCase.fail* methods.
  Document the deprecation.
........
  r71004 | benjamin.peterson | 2009-04-01 18:15:49 -0500 (Wed, 01 Apr 2009) | 1 line

  remove double underscores
........
  r71032 | michael.foord | 2009-04-01 22:20:38 -0500 (Wed, 01 Apr 2009) | 13 lines

  Better exception messages for unittest assert methods.

  - unittest.assertNotEqual() now uses the inequality operator (!=) instead
    of the equality operator.

  - Default assertTrue and assertFalse messages are now useful.

  - TestCase has a longMessage attribute. This defaults to False, but if set to True
    useful error messages are shown in addition to explicit messages passed to assert methods.

  Issue #5663
........
  r71043 | michael.foord | 2009-04-02 00:51:54 -0500 (Thu, 02 Apr 2009) | 7 lines

  Store the functions in the _type_equality_funcs as wrapped objects that are deep copyable.

  This allows for the deep copying of TestCase instances.

  Issue 5660
........
2009-04-04 16:35:46 +00:00
Antoine Pitrou 3a652b1d0a Merged revisions 70546 via svnmerge from
svn+ssh://pythondev@svn.python.org/python/trunk

........
  r70546 | antoine.pitrou | 2009-03-23 19:41:45 +0100 (lun., 23 mars 2009) | 9 lines

  Issue #4688: Add a heuristic so that tuples and dicts containing only
  untrackable objects are not tracked by the garbage collector. This can
  reduce the size of collections and therefore the garbage collection overhead
  on long-running programs, depending on their particular use of datatypes.

  (trivia: this makes the "binary_trees" benchmark from the Computer Language
  Shootout 40% faster)
........
2009-03-23 18:52:06 +00:00
Benjamin Peterson ee8712cda4 #2621 rename test.test_support to test.support 2008-05-20 21:35:26 +00:00
Christian Heimes a156e09b19 Merged revisions 60481,60485,60489-60492,60494-60496,60498-60499,60501-60503,60505-60506,60508-60509,60523-60524,60532,60543,60545,60547-60548,60552,60554,60556-60559,60561-60562,60569,60571-60572,60574,60576-60583,60585-60586,60589,60591,60594-60595,60597-60598,60600-60601,60606-60612,60615,60617,60619-60621,60623-60625,60627-60629,60631,60633,60635,60647,60650,60652,60654,60656,60658-60659,60664-60666,60668-60670,60672,60676,60678,60680-60683,60685-60686,60688,60690,60692-60694,60697-60700,60705-60706,60708,60711,60714,60720,60724-60730,60732,60736,60742,60744,60746,60748,60750-60751,60753,60756-60757,60759-60761,60763-60764,60766,60769-60770,60774-60784,60787-60845 via svnmerge from
svn+ssh://pythondev@svn.python.org/python/trunk

........
  r60790 | raymond.hettinger | 2008-02-14 10:32:45 +0100 (Thu, 14 Feb 2008) | 4 lines

  Add diagnostic message to help figure-out why SocketServer tests occasionally crash
  when trying to remove a pid that in not in the activechildren list.
........
  r60791 | raymond.hettinger | 2008-02-14 11:46:57 +0100 (Thu, 14 Feb 2008) | 1 line

  Add fixed-point examples to the decimal FAQ
........
  r60792 | raymond.hettinger | 2008-02-14 12:01:10 +0100 (Thu, 14 Feb 2008) | 1 line

  Improve rst markup
........
  r60794 | raymond.hettinger | 2008-02-14 12:57:25 +0100 (Thu, 14 Feb 2008) | 1 line

  Show how to remove exponents.
........
  r60795 | raymond.hettinger | 2008-02-14 13:05:42 +0100 (Thu, 14 Feb 2008) | 1 line

  Fix markup.
........
  r60797 | christian.heimes | 2008-02-14 13:47:33 +0100 (Thu, 14 Feb 2008) | 1 line

  Implemented Martin's suggestion to clear the free lists during the garbage collection of the highest generation.
........
  r60798 | raymond.hettinger | 2008-02-14 13:49:37 +0100 (Thu, 14 Feb 2008) | 1 line

  Simplify moneyfmt() recipe.
........
  r60810 | raymond.hettinger | 2008-02-14 20:02:39 +0100 (Thu, 14 Feb 2008) | 1 line

  Fix markup
........
  r60811 | raymond.hettinger | 2008-02-14 20:30:30 +0100 (Thu, 14 Feb 2008) | 1 line

  No need to register subclass of ABCs.
........
  r60814 | thomas.heller | 2008-02-14 22:00:28 +0100 (Thu, 14 Feb 2008) | 1 line

  Try to correct a markup error that does hide the following paragraph.
........
  r60822 | christian.heimes | 2008-02-14 23:40:11 +0100 (Thu, 14 Feb 2008) | 1 line

  Use a static and interned string for __subclasscheck__ and __instancecheck__ as suggested by Thomas Heller in #2115
........
  r60827 | christian.heimes | 2008-02-15 07:57:08 +0100 (Fri, 15 Feb 2008) | 1 line

  Fixed repr() and str() of complex numbers. Complex suffered from the same problem as floats but I forgot to test and fix them.
........
  r60830 | christian.heimes | 2008-02-15 09:20:11 +0100 (Fri, 15 Feb 2008) | 2 lines

  Bug #2111: mmap segfaults when trying to write a block opened with PROT_READ
  Thanks to Thomas Herve for the fix.
........
  r60835 | eric.smith | 2008-02-15 13:14:32 +0100 (Fri, 15 Feb 2008) | 1 line

  In PyNumber_ToBase, changed from an assert to returning an error when PyObject_Index() returns something other than an int or long.  It should never be possible to trigger this, as PyObject_Index checks to make sure it returns an int or long.
........
  r60837 | skip.montanaro | 2008-02-15 20:03:59 +0100 (Fri, 15 Feb 2008) | 8 lines

  Two new functions:

    * place_summary_first copies the regrtest summary to the front of the file
      making it easier to scan quickly for problems.

    * count_failures gets the actual count of the number of failing tests, not
      just a 1 (some failures) or 0 (no failures).
........
  r60840 | raymond.hettinger | 2008-02-15 22:21:25 +0100 (Fri, 15 Feb 2008) | 1 line

  Update example to match the current syntax.
........
  r60841 | amaury.forgeotdarc | 2008-02-15 22:22:45 +0100 (Fri, 15 Feb 2008) | 8 lines

  Issue #2115: __slot__ attributes setting was 10x slower.
  Also correct a possible crash using ABCs.

  This change is exactly the same as an optimisation
  done 5 years ago, but on slot *access*:
  http://svn.python.org/view?view=rev&rev=28297
........
  r60842 | amaury.forgeotdarc | 2008-02-15 22:27:44 +0100 (Fri, 15 Feb 2008) | 2 lines

  Temporarily let these tests pass
........
  r60843 | kurt.kaiser | 2008-02-15 22:56:36 +0100 (Fri, 15 Feb 2008) | 2 lines

  ScriptBinding event handlers weren't returning 'break'. Patch 2050, Tal Einat.
........
  r60844 | kurt.kaiser | 2008-02-15 23:25:09 +0100 (Fri, 15 Feb 2008) | 4 lines

  Configured selection highlighting colors were ignored; updating highlighting
  in the config dialog would cause non-Python files to be colored as if they
  were Python source; improve use of ColorDelagator.  Patch 1334. Tal Einat.
........
  r60845 | amaury.forgeotdarc | 2008-02-15 23:44:20 +0100 (Fri, 15 Feb 2008) | 9 lines

  Re-enable tests, they were failing since gc.collect() clears the various freelists.
  They still remain fragile.

  For example, a call to assertEqual currently does not make any allocation
  (which surprised me at first).
  But this can change when gc.collect also deletes the numerous "zombie frames"
  attached to each function.
........
2008-02-16 07:38:31 +00:00