Ethan Furman [Thu, 8 Dec 2022 06:58:08 +0000 (22:58 -0800)]
gh-100098: [Enum] insist on actual tuples, no subclasses, for auto (GH-100099)
When checking for auto() instances, only top-level usage is supported,
which means either alone or as part of a regular tuple. Other
containers, such as lists, dicts, or namedtuples, will not have auto()
transformed into a value.
Victor Stinner [Wed, 7 Dec 2022 14:22:38 +0000 (15:22 +0100)]
gh-98724: Fix Py_CLEAR() macro side effects (#99100) (#100070)
The Py_CLEAR(), Py_SETREF() and Py_XSETREF() macros now only evaluate
their arguments once. If an argument has side effects, these side
effects are no longer duplicated.
Use temporary variables to avoid duplicating side effects of macro
arguments. If available, use _Py_TYPEOF() to avoid type punning.
Otherwise, use memcpy() for the assignment to prevent a
miscompilation with strict aliasing caused by type punning.
Add _Py_TYPEOF() macro: __typeof__() on GCC and clang.
Add test_py_clear() and test_py_setref() unit tests to _testcapi.
gh-93453: No longer create an event loop in get_event_loop() (#98440)
asyncio.get_event_loop() now always return either running event loop or
the result of get_event_loop_policy().get_event_loop() call. The latter
should now raise an RuntimeError if no current event loop was set
instead of creating and setting a new event loop.
It affects also a number of asyncio functions and constructors which
call get_event_loop() implicitly: ensure_future(), shield(), gather(),
etc.
DeprecationWarning is no longer emitted if there is no running event loop but
the current event loop was set.
gh-100001: Also escape \s in http.server log messages. (#100038)
Also \ escape \s in the http.server BaseHTTPRequestHandler.log_message so
that it is technically possible to parse the line and reconstruct what the
original data was. Without this a \xHH is ambiguious as to if it is a hex
replacement we put in or the characters r"\x" came through in the original
request line.
gh-100001: Omit control characters in http.server stderr logs. (#100002)
Replace control characters in http.server.BaseHTTPRequestHandler.log_message with an escaped \xHH sequence to avoid causing problems for the terminal the output is printed to.
Petr Viktorin [Mon, 5 Dec 2022 17:02:36 +0000 (18:02 +0100)]
gh-98680: Add PyBUF_* constants to the Limited API listing (GH-100018)
``PyBUF_*`` constants are marked as part of Limited API of Python 3.11+.
These were available in 3.11.0 with `Py_LIMITED_API` defined for 3.11,
and are necessary to use the buffer API. Omitting them in `stable_abi.toml`
was a mistake.
AlexTate [Sun, 4 Dec 2022 19:37:55 +0000 (12:37 -0700)]
gh-98458: unittest: bugfix for infinite loop while handling chained exceptions that contain cycles (#98459)
* Bugfix addressing infinite loop while handling self-referencing chained exception in TestResult._clean_tracebacks()
* Bugfix extended to properly handle exception cycles in _clean_tracebacks. The "seen" set follows the approach used in the TracebackException class (thank you @iritkatriel for pointing it out)
* adds a test for a single chained exception that holds a self-loop in its __cause__ and __context__ attributes
bpo-15999: Accept arbitrary values for boolean parameters. (#15609)
builtins and extension module functions and methods that expect boolean values for parameters now accept any Python object rather than just a bool or int type. This is more consistent with how native Python code itself behaves.
Victor Stinner [Wed, 30 Nov 2022 16:22:52 +0000 (17:22 +0100)]
gh-99845: Use size_t type in __sizeof__() methods (#99846)
The implementation of __sizeof__() methods using _PyObject_SIZE() now
use an unsigned type (size_t) to compute the size, rather than a signed
type (Py_ssize_t).
Cast explicitly signed (Py_ssize_t) values to unsigned type
(Py_ssize_t).
Victor Stinner [Tue, 29 Nov 2022 11:15:21 +0000 (12:15 +0100)]
gh-99845: Clean up _PyObject_VAR_SIZE() usage (#99847)
* code_sizeof() now uses an unsigned type (size_t) to compute the result.
* Fix _PyObject_ComputedDictPointer(): cast _PyObject_VAR_SIZE() to
Py_ssize_t, rather than long: it's a different type on 64-bit Windows.
* Clarify that _PyObject_VAR_SIZE() uses an unsigned type (size_t).
Victor Stinner [Tue, 29 Nov 2022 11:12:17 +0000 (12:12 +0100)]
gh-99845: Change _PyDict_KeysSize() return type to size_t (#99848)
* Change _PyDict_KeysSize() and shared_keys_usable_size() return type
from signed (Py_ssize_t) to unsigned (size_t) type.
* new_values() argument type is now unsigned (size_t).
* init_inline_values() now uses size_t rather than int for the 'i'
iterator variable.
* type.__sizeof__() implementation now uses unsigned (size_t) type.
* PyCode_GetNumFree() casts PyCode_GetNumFree.co_nfreevars from int
to Py_ssize_t to be future proof, and because Py_ssize_t is
commonly used in the C API.
* PyCode_GetNumFree() doesn't cast its argument: the replaced macro
already required the exact type PyCodeObject*.
* Add assertions in some functions using "CAST" macros to check
the arguments type when Python is built with assertions
(debug build).
* Remove an outdated comment in unicodeobject.h.
These slots are marked "should be treated as read-only" in the
table at the start of the document. That doesn't say anything about
setting them in the static struct.
`tp_bases` docs did say that it should be ``NULL`` (TIL!). If you
ignore that, seemingly nothing bad happens. However, some slots
may not be inherited, depending on which sub-slot structs are present.
(FWIW, NumPy sets tp_bases and is affected by the quirk -- though to
be fair, its DUAL_INHERIT code probably predates tp_bases docs, and
also the result happens to be benign.)
This patch makes things explicit.
It also makes the summary table legend easier to scan.
Ronald Oussoren [Sun, 27 Nov 2022 10:56:14 +0000 (11:56 +0100)]
GH-87235: Make sure "python /dev/fd/9 9</path/to/script.py" works on macOS (#99768)
On macOS all file descriptors for a particular file in /dev/fd
share the same file offset, that is ``open("/dev/fd/9", "r")`` behaves
more like ``dup(9)`` than a regular open.
This causes problems when a user tries to run "/dev/fd/9" as a script
because zipimport changes the file offset to try to read a zipfile
directory. Therefore change zipimport to reset the file offset after
trying to read the zipfile directory.