[3.13] gh-113433: Automatically Clean Up Subinterpreters in Py_Finalize() (gh-121067)
This change makes things a little less painful for some users. It also fixes a failing assert (gh-120765), by making sure all subinterpreters are destroyed before the main interpreter. As part of that, we make sure Py_Finalize() always runs with the main interpreter active.
When testing IDLE, don't create a Tk to avoid side effects such as
installing a PyOS_InputHook hook.
(cherry picked from commit 44eafd66882589d4f4eb569d70c49724da3e9291)
Co-authored-by: Victor Stinner <vstinner@python.org>
Victor Stinner [Wed, 26 Jun 2024 13:35:19 +0000 (15:35 +0200)]
[3.13] gh-120642: Move private PyCode APIs to the internal C API (#120643) (#121043)
gh-120642: Move private PyCode APIs to the internal C API (#120643)
* Move _Py_CODEUNIT and related functions to pycore_code.h.
* Move _Py_BackoffCounter to pycore_backoff.h.
* Move Include/cpython/optimizer.h content to pycore_optimizer.h.
* Remove Include/cpython/optimizer.h.
* Remove PyUnstable_Replace_Executor().
[3.13] gh-121018: Ensure ArgumentParser.parse_args with exit_on_error=False raises instead of exiting when given unrecognized arguments (GH-121019) (GH-121032)
[3.13] gh-120380: fix Python implementation of `pickle.Pickler` for `bytes` and `bytearray` objects in protocol version 5. (GH-120422) (GH-120832)
gh-120380: fix Python implementation of `pickle.Pickler` for `bytes` and `bytearray` objects in protocol version 5. (GH-120422)
(cherry picked from commit 7595e6743ac78ac0dd19418176f66d251668fafc)
[3.13] gh-120858: PyDict_Next should not lock the dict (GH-120859) (#120964)
PyDict_Next no longer locks the dictionary in the free-threaded build. Locking
around individual PyDict_Next calls is not sufficient because the function
returns borrowed references and because it allows concurrent modifications
during the iteraiton loop.
The internal locking also interferes with correct external synchronization
because it may suspend outer critical sections created by the caller.
(cherry picked from commit 375b723d5873f948696c7e85a97f4778d9e00ff0)
Petr Viktorin [Mon, 24 Jun 2024 18:24:19 +0000 (20:24 +0200)]
[3.13] gh-113993: Allow interned strings to be mortal, and fix related issues (GH-120520) (GH-120945)
* Add an InternalDocs file describing how interning should work and how to use it.
* Add internal functions to *explicitly* request what kind of interning is done:
- `_PyUnicode_InternMortal`
- `_PyUnicode_InternImmortal`
- `_PyUnicode_InternStatic`
* Switch uses of `PyUnicode_InternInPlace` to those.
* Disallow using `_Py_SetImmortal` on strings directly.
You should use `_PyUnicode_InternImmortal` instead:
- Strings should be interned before immortalization, otherwise you're possibly
interning a immortalizing copy.
- `_Py_SetImmortal` doesn't handle the `SSTATE_INTERNED_MORTAL` to
`SSTATE_INTERNED_IMMORTAL` update, and those flags can't be changed in
backports, as they are now part of public API and version-specific ABI.
* Add private `_only_immortal` argument for `sys.getunicodeinternedsize`, used in refleak test machinery.
* Make sure the statically allocated string singletons are unique. This means these sets are now disjoint:
- `_Py_ID`
- `_Py_STR` (including the empty string)
- one-character latin-1 singletons
Now, when you intern a singleton, that exact singleton will be interned.
* Add a `_Py_LATIN1_CHR` macro, use it instead of `_Py_ID`/`_Py_STR` for one-character latin-1 singletons everywhere (including Clinic).
* Intern `_Py_STR` singletons at startup.
* For free-threaded builds, intern `_Py_LATIN1_CHR` singletons at startup.
* Beef up the tests. Cover internal details (marked with `@cpython_only`).
* Add lots of assertions
Co-authored-by: Eric Snow <ericsnowcurrently@gmail.com>
[3.13] gh-119521: Rename IncompleteInputError to _IncompleteInputError and remove from public API/ABI (GH-119680, GH-120955) (GH-120944)
- gh-119521: Rename IncompleteInputError to _IncompleteInputError and remove from public API/ABI (GH-119680)
(cherry picked from commit ce1064e4c9bcfd673323ad690e60f86e1ab907bb)
[3.13] gh-120683: Fix an error in logging.LogRecord timestamp (GH-120709) (GH-120933)
The integer part of the timestamp can be rounded up, while the millisecond
calculation truncates, causing the log timestamp to be wrong by up to 999 ms
(affected roughly 1 in 8 million timestamps).
(cherry picked from commit 1500a23f33f5a6d052ff1ef6383d9839928b8ff1)
[3.13] gh-101830: Fix Tcl_Obj to string conversion (GH-120884) (GH-120905)
Accessing the Tkinter object's string representation no longer converts
the underlying Tcl object to a string on Windows.
(cherry picked from commit f4ddaa396715855ffbd94590f89ab7d55feeec07)
[3.13] gh-104855: Update Tkinter tests for Tcl/Tk 8.7 and 9.0 (GH-120824) (GH-120864)
The tests are now passed with the current version of Tcl/Tk under
development (8.7b1+ and 9.0b3+).
The following changes were also made to make the tests more flexible:
* Helper methods like checkParam() now interpret the expected error message
as a regular expression instead of a literal.
* Add support of new arguments in checkEnumParam():
- allow_empty=True skips testing with empty string;
- fullname= specifies the name for error message if it differs from the
option name;
- sort=True sorts values for error message.
* Add support of the allow_empty argument in checkReliefParam():
allow_empty=True adds an empty string to the list of accepted values.
* Attributes _clip_highlightthickness, _clip_pad and _clip_borderwidth
specify how negative values of options -highlightthickness, -padx, -pady
and -borderwidth are handled.
* Use global variables for some common error messages.
The supporting functions and structs used by the macros are also exposed for
cases where C macros are not available.
(cherry picked from commit 8f17d69b7bc906e8407095317842cc0fd52cd84a)
[3.13] gh-120773: document introspective attributes of an async generator object in the inspect module (GH-120778) (#120827)
gh-120773: document introspective attributes of an async generator object in the inspect module (GH-120778)
(cherry picked from commit 83d3d7aace32b8536f552f78dd29610344f13160)
Includes a follow-up fix to properly merge GH-119877:
* Add a link to free-threading HOWTO to the index (GH-120703)
(cherry picked from commit 45d5cab533a607716b2b41134839a59facf309cd)
Co-authored-by: Carol Willing <carolcode@willingconsulting.com> Co-authored-by: Ned Batchelder <ned@nedbatchelder.com>
[3.13] gh-120732: Fix `name` passing to `Mock`, when using kwargs to `create_autospec` (GH-120737) (#120760)
gh-120732: Fix `name` passing to `Mock`, when using kwargs to `create_autospec` (GH-120737)
(cherry picked from commit 1e4815692f6c8a37a3974d0d7d2025494d026d76)
[3.13] gh-120635: Avoid leaking processes in test_pyrepl (GH-120676) (#120741)
gh-120635: Avoid leaking processes in test_pyrepl (GH-120676)
If the child process takes longer than SHORT_TIMEOUT seconds to
complete, kill the process but then wait until it completes with no
timeout to not leak child processes.
(cherry picked from commit 0f3e36454d754026d6c510053ff1e4b22ae80cd9)
Co-authored-by: Victor Stinner <vstinner@python.org>
[3.13] gh-120726: Fix compiler warnings on is_core_module() (GH-120727) (#120729)
gh-120726: Fix compiler warnings on is_core_module() (GH-120727)
Fix compiler warnings on is_core_module() and
check_interpreter_whence(): only define them when
assertions are built.
(cherry picked from commit a816cd67f43d9adb27ccdb6331e08c835247d1df)
The `inspect.ismethoddescriptor()` function did not check for the lack of
`__delete__()` and, consequently, erroneously returned True when applied
to *data* descriptors with only `__get__()` and `__delete__()` defined.
This exposes `PyUnstable_Object_ClearWeakRefsNoCallbacks` as an unstable
C-API function to provide a thread-safe mechanism for clearing weakrefs
without executing callbacks.
Some C-API extensions need to clear weakrefs without calling callbacks,
such as after running finalizers like we do in subtype_dealloc.
Previously they could use `_PyWeakref_ClearRef` on each weakref, but
that's not thread-safe in the free-threaded build.
[3.13] gh-120590: Fix test_pydoc in the refleak hunting mode (GH-120615) (GH-120669)
Mocking only works if sys.modules['pydoc'] and pydoc are the same,
but some pydoc functions reload the module and change sys.modules.
Ensure that sys.modules['pydoc'] is always restored after the corresponding
tests.
(cherry picked from commit 2cf47389e26cb591342d07dad98619916d5a1b15)
Fix a race in `PyMember_GetOne` and `PyMember_SetOne` for `Py_T_OBJECT_EX`.
These functions implement `__slots__` accesses for Python objects.
(cherry picked from commit 362cd2680b45a36c3467b9721ff7fc0ceb338452)
[3.13] gh-120433: Mention ``chocolatey`` for installing llvm on Windows as an alternative option (GH-120434) (#120651)
gh-120433: Mention ``chocolatey`` for installing llvm on Windows as an alternative option (GH-120434)
(cherry picked from commit 95737bbf18765a24b6585708588c9b707dc30d27)
[3.13] gh-120367: fix removal of redundant NOPs and jumps after reordering hot-cold blocks (GH-120425) (#120621)
gh-120367: fix removal of redundant NOPs and jumps after reordering hot-cold blocks (GH-120425)
(cherry picked from commit 21866c8ed296524f0ca175c0f55b43744c2b30df)
[3.13] gh-118596: Add thread-safety clarifications to the SSLContext documentation (GH-118597) (#120595)
gh-118596: Add thread-safety clarifications to the SSLContext documentation (GH-118597)
Add thread-safety clarifications to the SSLContext documentation. Per the issue:
This issue has also come up [here](https://github.com/psf/requests/pull/6667) where the matter was clarified by @tiran in [this comment](https://github.com/psf/requests/pull/6667):
> `SSLContext` is designed to be shared and used for multiple connections. It is thread safe as long as you don't reconfigure it once it is used by a connection. Adding new certs to the internal trust store is fine, but changing ciphers, verification settings, or mTLS certs can lead to surprising behavior. The problem is unrelated to threads and can even occur in a single-threaded program.
(cherry picked from commit 4f59f8638267aa64ad2daa0111d8b7fdc2499834)
[3.13] gh-112346: Always set OS byte to 255, simpler gzip.compress function. (GH-120486) (#120563)
gh-112346: Always set OS byte to 255, simpler gzip.compress function. (GH-120486)
This matches the output behavior in 3.10 and earlier; the optimization in 3.11 allowed the zlib library's "os" value to be filled in instead in the circumstance when mtime was 0. this keeps things consistent.
(cherry picked from commit 08d09cf5ba041c9c5c3860200b56bab66fd44a23)
[3.13] gh-120541: Improve the "less" prompt in pydoc (GH-120543) (GH-120562)
When help() is called with non-string argument, use __qualname__ or
__name__ if available, otherwise use "{typename} object".
(cherry picked from commit 31d1d72d7e24e0427df70f7dd14b9baff28a4f89)
projects / thirdparty / Python / cpython.git / log
