Commit Graph

245 Commits

Author SHA1 Message Date
Dennis Sweeney da6c78584b
gh-90667: Add specializations of Py_DECREF when types are known (GH-30872) 2022-04-19 19:02:19 +01:00
Eric Snow 08deed1af5
bpo-46753: Add the empty tuple to the _PyRuntimeState.global_objects. (gh-31345)
https://bugs.python.org/issue46753
2022-02-28 15:15:48 -07:00
Eric Snow 81c72044a1
bpo-46541: Replace core use of _Py_IDENTIFIER() with statically initialized global objects. (gh-30928)
We're no longer using _Py_IDENTIFIER() (or _Py_static_string()) in any core CPython code.  It is still used in a number of non-builtin stdlib modules.

The replacement is: PyUnicodeObject (not pointer) fields under _PyRuntimeState, statically initialized as part of _PyRuntime.  A new _Py_GET_GLOBAL_IDENTIFIER() macro facilitates lookup of the fields (along with _Py_GET_GLOBAL_STRING() for non-identifier strings).

https://bugs.python.org/issue46541#msg411799 explains the rationale for this change.

The core of the change is in:

* (new) Include/internal/pycore_global_strings.h - the declarations for the global strings, along with the macros
* Include/internal/pycore_runtime_init.h - added the static initializers for the global strings
* Include/internal/pycore_global_objects.h - where the struct in pycore_global_strings.h is hooked into _PyRuntimeState
* Tools/scripts/generate_global_objects.py - added generation of the global string declarations and static initializers

I've also added a --check flag to generate_global_objects.py (along with make check-global-objects) to check for unused global strings.  That check is added to the PR CI config.

The remainder of this change updates the core code to use _Py_GET_GLOBAL_IDENTIFIER() instead of _Py_IDENTIFIER() and the related _Py*Id functions (likewise for _Py_GET_GLOBAL_STRING() instead of _Py_static_string()).  This includes adding a few functions where there wasn't already an alternative to _Py*Id(), replacing the _Py_Identifier * parameter with PyObject *.

The following are not changed (yet):

* stop using _Py_IDENTIFIER() in the stdlib modules
* (maybe) get rid of _Py_IDENTIFIER(), etc. entirely -- this may not be doable as at least one package on PyPI using this (private) API
* (maybe) intern the strings during runtime init

https://bugs.python.org/issue46541
2022-02-08 13:39:07 -07:00
Victor Stinner ac1f152421
bpo-46417: Use _PyType_CAST() in Objects directory (GH-30764) 2022-01-21 23:33:43 +01:00
Dennis Sweeney ad1d5908ad
bpo-46235: Do all ref-counting at once during list/tuple multiplication (GH-30346)
When multiplying lists and tuples by `n`, increment each element's refcount, by `n`, just once.

Saves `n-1` increments per element, and allows for a leaner & faster copying loop.

Code by  sweeneyde (Dennis Sweeney).
2022-01-07 21:47:58 -06:00
Eric Snow c8749b5783
bpo-46008: Make runtime-global object/type lifecycle functions and state consistent. (gh-29998)
This change is strictly renames and moving code around.  It helps in the following ways:

* ensures type-related init functions focus strictly on one of the three aspects (state, objects, types)
* passes in PyInterpreterState * to all those functions, simplifying work on moving types/objects/state to the interpreter
* consistent naming conventions help make what's going on more clear
* keeping API related to a type in the corresponding header file makes it more obvious where to look for it

https://bugs.python.org/issue46008
2021-12-09 12:59:26 -07:00
Christian Heimes 9942f42a93
bpo-45522: Allow to disable freelists on build time (GH-29056)
Freelists for object structs can now be disabled. A new ``configure``
option ``--without-freelists`` can be used to disable all freelists
except empty tuple singleton. Internal Py*_MAXFREELIST macros can now
be defined as 0 without causing compiler warnings and segfaults.

Signed-off-by: Christian Heimes <christian@python.org>
2021-10-21 06:12:20 -07:00
Mark Shannon 70945d57e7
bpo-45256: Avoid C calls for most Python to Python calls. (GH-28937)
* Avoid making C calls for most calls to Python functions.

* Change initialize_locals(steal=true) and _PyTuple_FromArraySteal to consume the argument references regardless of whether they succeed or fail.
2021-10-18 09:57:24 +01:00
Pablo Galindo Salgado b4903afd4d
bpo-45256: Remove the usage of the C stack in Python to Python calls (GH-28488)
Ths commit inlines calls to Python functions in the eval loop and steals all the arguments in the call from the caller for
performance.
2021-10-09 16:51:30 +01:00
Victor Stinner 79a3148099
bpo-45061: Detect refcount bug on empty tuple singleton (GH-28503)
Detect refcount bugs in C extensions when the empty tuple singleton
is destroyed by mistake.

Add the _Py_FatalRefcountErrorFunc() function.
2021-09-21 23:04:34 +02:00
Serhiy Storchaka f25f2e2e8c
Clean up initialization __class_getitem__ with Py_GenericAlias. (GH-28450)
The cast to PyCFunction is redundant. Overuse of redundant casts
can hide actual bugs.
2021-09-19 18:05:30 +03:00
Victor Stinner 818628c2da
bpo-44531: Add _PyType_AllocNoTrack() function (GH-26947)
Add an internal _PyType_AllocNoTrack() function to allocate an object
without tracking it in the GC.

Modify dict_new() to use _PyType_AllocNoTrack(): dict subclasses are
now only tracked once all PyDictObject members are initialized.
Calling _PyObject_GC_UNTRACK() is no longer needed for the dict type.

Similar change in tuple_subtype_new() for tuple subclasses.

Replace tuple_gc_track() with _PyObject_GC_TRACK().
2021-07-01 02:30:46 +02:00
Mark Shannon 069e81ab3d
bpo-43977: Use tp_flags for collection matching (GH-25723)
* Add Py_TPFLAGS_SEQUENCE and Py_TPFLAGS_MAPPING, add to all relevant standard builtin classes.

* Set relevant flags on collections.abc.Sequence and Mapping.

* Use flags in MATCH_SEQUENCE and MATCH_MAPPING opcodes.

* Inherit Py_TPFLAGS_SEQUENCE and Py_TPFLAGS_MAPPING.

* Add NEWS

* Remove interpreter-state map_abc and seq_abc fields.
2021-04-30 09:50:28 +01:00
Brandt Bucher 145bf269df
bpo-42128: Structural Pattern Matching (PEP 634) (GH-22917)
Co-authored-by: Guido van Rossum <guido@python.org>
Co-authored-by: Talin <viridia@gmail.com>
Co-authored-by: Pablo Galindo <pablogsal@gmail.com>
2021-02-26 14:51:55 -08:00
Victor Stinner bcb094b41f
bpo-43268: Pass interp rather than tstate to internal functions (GH-24580)
Pass the current interpreter (interp) rather than the current Python
thread state (tstate) to internal functions which only use the
interpreter.

Modified functions:

* _PyXXX_Fini() and _PyXXX_ClearFreeList() functions
* _PyEval_SignalAsyncExc(), make_pending_calls()
* _PySys_GetObject(), sys_set_object(), sys_set_object_id(), sys_set_object_str()
* should_audit(), set_flags_from_config(), make_flags()
* _PyAtExit_Call()
* init_stdio_encoding()
* etc.
2021-02-19 15:10:45 +01:00
Victor Stinner 0430dfac62
bpo-40521: Always create the empty tuple singleton (GH-21116)
Py_InitializeFromConfig() now always creates the empty tuple
singleton as soon as possible.

Optimize PyTuple_New(0): it no longer has to check if the empty tuple
was created or not, it is always creatd.

* Add tuple_create_empty_tuple_singleton() function.
* Add tuple_get_empty() function.
* Remove state parameter of tuple_alloc().
2020-06-24 15:21:54 +02:00
Victor Stinner 522691c46e
bpo-40521: Cleanup code of free lists (GH-21082)
Add get_xxx_state() function to factorize duplicated code.
2020-06-23 16:40:40 +02:00
Victor Stinner 04fc4f2a46
bpo-40989: PyObject_INIT() becomes an alias to PyObject_Init() (GH-20901)
The PyObject_INIT() and PyObject_INIT_VAR() macros become aliases to,
respectively, PyObject_Init() and PyObject_InitVar() functions.

Rename _PyObject_INIT() and _PyObject_INIT_VAR() static inline
functions to, respectively, _PyObject_Init() and _PyObject_InitVar(),
and move them to pycore_object.h. Remove their return value:
their return type becomes void.

The _datetime module is now built with the Py_BUILD_CORE_MODULE macro
defined.

Remove an outdated comment on _Py_tracemalloc_config.
2020-06-16 01:28:07 +02:00
Victor Stinner bcb198385d
bpo-40887: Don't use finalized free lists (GH-20700)
In debug mode, ensure that free lists are no longer used after being
finalized. Set numfree to -1 in finalization functions
(eg. _PyList_Fini()), and then check that numfree is not equal to -1
before using a free list (e.g list_dealloc()).
2020-06-08 02:14:47 +02:00
Victor Stinner 69ac6e58fd
bpo-40521: Make tuple free list per-interpreter (GH-20247)
Each interpreter now has its own tuple free lists:

* Move tuple numfree and free_list arrays into PyInterpreterState.
* Define PyTuple_MAXSAVESIZE and PyTuple_MAXFREELIST macros in
  pycore_interp.h.
* Add _Py_tuple_state structure. Pass it explicitly to tuple_alloc().
* Add tstate parameter to _PyTuple_ClearFreeList()
* Each interpreter now has its own empty tuple singleton.
2020-06-04 23:38:36 +02:00
Victor Stinner fe2978b3b9
bpo-39573: Convert Py_REFCNT and Py_SIZE to functions (GH-20429)
Convert Py_REFCNT() and Py_SIZE() macros to static inline functions.
They cannot be used as l-value anymore: use Py_SET_REFCNT() and
Py_SET_SIZE() to set an object reference count and size.

Replace &Py_SIZE(self) with &((PyVarObject*)self)->ob_size
in arraymodule.c.

This change is backward incompatible on purpose, to prepare the C API
for an opaque PyObject structure.
2020-05-27 14:55:10 +02:00
Dong-hee Na 7d847e29d7
bpo-39573: Fix buildbot failure for tupleobject.c (GH-20391) 2020-05-26 02:25:28 +09:00
Sergey Fedoseev e682b26a6b
bpo-34397: Remove redundant overflow checks in list and tuple implementation. (GH-8757) 2020-05-25 07:54:40 -07:00
Victor Stinner b4b53868d7
bpo-40521: Disable free lists in subinterpreters (GH-19937)
When Python is built with experimental isolated interpreters, disable
tuple, dict and free free lists.

Temporary workaround until these caches are made per-interpreter.

Add frame_alloc() and frame_get_builtins() subfunctions to simplify
_PyFrame_New_NoTrack().
2020-05-05 19:55:29 +02:00
Victor Stinner ae00a5a885
bpo-40428: Remove PyTuple_ClearFreeList() function (GH-19769)
Remove the following function from the C API:

* PyAsyncGen_ClearFreeLists()
* PyContext_ClearFreeList()
* PyDict_ClearFreeList()
* PyFloat_ClearFreeList()
* PyFrame_ClearFreeList()
* PyList_ClearFreeList()
* PySet_ClearFreeList()
* PyTuple_ClearFreeList()

Make these functions private, move them to the internal C API and
change their return type to void.

Call explicitly PyGC_Collect() to free all free lists.

Note: PySet_ClearFreeList() did nothing.
2020-04-29 02:29:20 +02:00
Victor Stinner e5014be049
bpo-40268: Remove a few pycore_pystate.h includes (GH-19510) 2020-04-14 17:52:15 +02:00
Victor Stinner a15e260b70
bpo-40170: Add _PyIndex_Check() internal function (GH-19426)
Add _PyIndex_Check() function to the internal C API: fast inlined
verson of PyIndex_Check().

Add Include/internal/pycore_abstract.h header file.

Replace PyIndex_Check() with _PyIndex_Check() in C files of Objects
and Python subdirectories.
2020-04-08 02:01:56 +02:00
Guido van Rossum 48b069a003
bpo-39481: Implementation for PEP 585 (#18239)
This implements things like `list[int]`,
which returns an object of type `types.GenericAlias`.
This object mostly acts as a proxy for `list`,
but has attributes `__origin__` and `__args__`
that allow recovering the parts (with values `list` and `(int,)`.

There is also an approximate notion of type variables;
e.g. `list[T]` has a `__parameters__` attribute equal to `(T,)`.
Type variables are objects of type `typing.TypeVar`.
2020-04-07 09:50:06 -07:00
Dong-hee Na 87ec86c425
bpo-37207: Add _PyArg_NoKwnames() helper function (GH-18980) 2020-03-16 15:06:20 +01:00
Dong-hee Na c98f87fc33
bpo-37207: Use _PyArg_CheckPositional() for tuple vectorcall (GH-18986) 2020-03-16 15:04:14 +01:00
Hai Shi c81609e44e
Fix a possible refleak in tupleobject.c (GH-19018) 2020-03-15 19:37:49 +00:00
Dong-hee Na 9ee88cde1a
bpo-37207: Use PEP 590 vectorcall to speed up tuple() (GH-18936)
Master:

./python.exe -m pyperf timeit "tuple((1, 2, 3, 4, 5))"
Mean +- std dev: 361 ns +- 15 ns

PEP-590:

./python.exe -m pyperf timeit "tuple((1, 2, 3, 4, 5))"
Mean +- std dev: 203 ns +- 13 ns
2020-03-13 14:57:00 +01:00
Andy Lester 557287075c
bpo-39573: Use Py_IS_TYPE() macro to check for types (GH-18809)
Co-authored-by: Victor Stinner <vstinner@python.org>
2020-03-06 23:53:17 +01:00
Andy Lester dffe4c0709
bpo-39573: Finish converting to new Py_IS_TYPE() macro (GH-18601) 2020-03-04 14:15:20 +01:00
Victor Stinner a93c51e3a8
bpo-39573: Use Py_REFCNT() macro (GH-18388)
Replace direct acccess to PyObject.ob_refcnt with usage of the
Py_REFCNT() macro.
2020-02-07 00:38:59 +01:00
Victor Stinner 49932fec62
bpo-39542: Simplify _Py_NewReference() (GH-18332)
* Remove _Py_INC_REFTOTAL and _Py_DEC_REFTOTAL macros: modify
  directly _Py_RefTotal.
* _Py_ForgetReference() is no longer defined if the Py_TRACE_REFS
  macro is not defined.
* Remove _Py_NewReference() implementation from object.c:
  unify the two implementations in object.h inline function.
* Fix Py_TRACE_REFS build: _Py_INC_TPALLOCS() macro has been removed.
2020-02-03 17:55:04 +01:00
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
Sergey Fedoseev 755d4ef824 Correct overflow check in PyTuple_New() (GH-14838) 2019-09-09 13:40:58 -07:00
Zackery Spytz 60bd1f88f2 bpo-36030: Fix a possible segfault in PyTuple_New() (GH-15670) 2019-09-04 15:58:04 +02:00
Victor Stinner bed4817d52
Make PyXXX_Fini() functions private (GH-15531)
For example, rename PyTuple_Fini() to _PyTuple_Fini().

These functions are only declared in the internal C API.
2019-08-27 00:12:32 +02:00
Sergey Fedoseev 4fa10dde40 bpo-36030: Improve performance of some tuple operations (GH-12052) 2019-08-14 16:10:33 +02:00
Serhiy Storchaka 18b711c5a7
bpo-37648: Fixed minor inconsistency in some __contains__. (GH-14904)
The collection's item is now always at the left and
the needle is on the right of ==.
2019-08-04 14:12:48 +03:00
Jeroen Demeyer 530f506ac9 bpo-36974: tp_print -> tp_vectorcall_offset and tp_reserved -> tp_as_async (GH-13464)
Automatically replace
tp_print -> tp_vectorcall_offset
tp_compare -> tp_as_async
tp_reserved -> tp_as_async
2019-05-30 19:13:39 -07:00
Victor Stinner 331a6a56e9
bpo-36763: Implement the PEP 587 (GH-13592)
* Add a whole new documentation page:
  "Python Initialization Configuration"
* PyWideStringList_Append() return type is now PyStatus,
  instead of int
* PyInterpreterState_New() now calls PyConfig_Clear() if
  PyConfig_InitPythonConfig() fails.
* Rename files:

  * Python/coreconfig.c => Python/initconfig.c
  * Include/cpython/coreconfig.h => Include/cpython/initconfig.h
  * Include/internal/: pycore_coreconfig.h => pycore_initconfig.h

* Rename structures

  * _PyCoreConfig => PyConfig
  * _PyPreConfig => PyPreConfig
  * _PyInitError => PyStatus
  * _PyWstrList => PyWideStringList

* Rename PyConfig fields:

  * use_module_search_paths => module_search_paths_set
  * module_search_path_env => pythonpath_env

* Rename PyStatus field: _func => func
* PyInterpreterState: rename core_config field to config
* Rename macros and functions:

  * _PyCoreConfig_SetArgv() => PyConfig_SetBytesArgv()
  * _PyCoreConfig_SetWideArgv() => PyConfig_SetArgv()
  * _PyCoreConfig_DecodeLocale() => PyConfig_SetBytesString()
  * _PyInitError_Failed() => PyStatus_Exception()
  * _Py_INIT_ERROR_TYPE_xxx enums => _PyStatus_TYPE_xxx
  * _Py_UnixMain() => Py_BytesMain()
  * _Py_ExitInitError() => Py_ExitStatusException()
  * _Py_PreInitializeFromArgs() => Py_PreInitializeFromBytesArgs()
  * _Py_PreInitializeFromWideArgs() => Py_PreInitializeFromArgs()
  * _Py_PreInitialize() => Py_PreInitialize()
  * _Py_RunMain() => Py_RunMain()
  * _Py_InitializeFromConfig() => Py_InitializeFromConfig()
  * _Py_INIT_XXX() => _PyStatus_XXX()
  * _Py_INIT_FAILED() => _PyStatus_EXCEPTION()

* Rename 'err' PyStatus variables to 'status'
* Convert RUN_CODE() macro to config_run_code() static inline function
* Remove functions:

  * _Py_InitializeFromArgs()
  * _Py_InitializeFromWideArgs()
  * _PyInterpreterState_GetCoreConfig()
2019-05-27 16:39:22 +02:00
Zackery Spytz 14514d9084 bpo-36946: Fix possible signed integer overflow when handling slices. (GH-13375)
The final addition (cur += step) may overflow, so use size_t for "cur".
"cur" is always positive (even for negative steps), so it is safe to use
size_t here.

Co-Authored-By: Martin Panter <vadmium+py@gmail.com>
2019-05-17 10:13:03 +03:00
Jeroen Demeyer 351c67416b bpo-35983: skip trashcan for subclasses (GH-11841)
Add new trashcan macros to deal with a double deallocation that could occur when the `tp_dealloc` of a subclass calls the `tp_dealloc` of a base class and that base class uses the trashcan mechanism.

Patch by Jeroen Demeyer.
2019-05-10 19:21:10 +02:00
Sergey Fedoseev 234531b446 bpo-36030: Add _PyTuple_FromArray() function (GH-11954) 2019-02-25 17:59:12 +01:00
Serhiy Storchaka bb86bf4c4e
bpo-35444: Unify and optimize the helper for getting a builtin object. (GH-11047)
This speeds up pickling of some iterators.

This fixes also error handling in pickling methods when fail to
look up builtin "getattr".
2018-12-11 08:28:18 +02:00
Victor Stinner bcda8f1d42
bpo-35081: Add Include/internal/pycore_object.h (GH-10640)
Move _PyObject_GC_TRACK() and _PyObject_GC_UNTRACK() from
Include/objimpl.h to Include/internal/pycore_object.h.
2018-11-21 22:27:47 +01:00
Victor Stinner 621cebe81b
bpo-35081: Rename internal headers (GH-10275)
Rename Include/internal/ headers:

* pycore_hash.h -> pycore_pyhash.h
* pycore_lifecycle.h -> pycore_pylifecycle.h
* pycore_mem.h -> pycore_pymem.h
* pycore_state.h -> pycore_pystate.h

Add missing headers to Makefile.pre.in and PCbuild:

* pycore_condvar.h.
* pycore_hamt.h
* pycore_pyhash.h
2018-11-12 16:53:38 +01:00