The `module` parameter carries semantic information about the forward ref.
Forward refs are different if they refer to different module even if they
have the same name. This affects the `__eq__`, `__repr__` and `__hash__` methods.
Co-authored-by: Andreas Hangauer <andreas.hangauer@siemens.com>
Co-authored-by: Alex Waygood <Alex.Waygood@Gmail.com>
Co-authored-by: Ken Jin <28750310+Fidget-Spinner@users.noreply.github.com>
asyncio/taskgroups.py is an adaptation of taskgroup.py from EdgeDb, with the following key changes:
- Allow creating new tasks as long as the last task hasn't finished
- Raise [Base]ExceptionGroup (directly) rather than TaskGroupError deriving from MultiError
- Instead of monkey-patching the parent task's cancel() method,
add a new public API to Task
The Task class has a new internal flag, `_cancel_requested`, which is set when `.cancel()` is called successfully. The `.cancelling()` method returns the value of this flag. Further `.cancel()` calls while this flag is set return False. To reset this flag, call `.uncancel()`.
Thus, a Task that catches and ignores `CancelledError` should call `.uncancel()` if it wants to be cancellable again; until it does so, it is deemed to be busy with uninterruptible cleanup.
This new Task API helps solve the problem where TaskGroup needs to distinguish between whether the parent task being cancelled "from the outside" vs. "from inside".
Co-authored-by: Yury Selivanov <yury@edgedb.com>
Co-authored-by: Andrew Svetlov <andrew.svetlov@gmail.com>
The module parameter carries semantic information about the forward ref.
Show to the user that forward refs with same argument but different
module are different.
Co-authored-by: Andreas Hangauer <andreas.hangauer@siemens.com>
Co-authored-by: blurb-it[bot] <43283697+blurb-it[bot]@users.noreply.github.com>
Ensure strong references are acquired whenever using `set_next()`. Added randomized test cases for `__eq__` methods that sometimes mutate sets when called.
Confirmed with @jaraco that this indeed needs a fix.
A question that came up while I was digging into the code: I think `SelectableGroups` could similarly use `__slots__ = ()`, since its purpose seems only for convenience around `dict`, not to have attributes of its own.
Automerge-Triggered-By: GH:jaraco
Numeric fields of type float, notably mtime, can't be represented
exactly in the ustar header, so the pax header is used. But it is
helpful to set them to the nearest int (i.e. second rather than
nanosecond precision mtimes) in the ustar header as well, for the
benefit of unarchivers that don't understand the pax header.
Add test for tarfile.TarInfo.create_pax_header to confirm correct
behaviour.
In [Lib/test/support/import_helper.py](https://github.com/python/cpython/blob/master/Lib/test/support/import_helper.py), the function `make_legacy_pyc` makes a call to `os.rename` which can fail when the source and target live on different devices. This happens (for example) when `PYTHONPYCACHEPREFIX` is set to a directory anywhere on disk, while a ramdisk is mounted on `/tmp` (the latter of which is the default on various Linux distros). Replacing `os.rename` with `shutil.move` fixes this.
Automerge-Triggered-By: GH:brettcannon
The calendar.LocaleTextCalendar and calendar.LocaleHTMLCalendar
classes module now use locale.getlocale(), instead of using
locale.getdefaultlocale(), if no locale is specified.
skip_if_buggy_ucrt_strfptime() of test.support now uses
locale.getpreferredencoding(False) instead of
locale.getdefaultlocale() to get the Windows code page.
encodings registers the _alias_mbcs() codec search function before
the search_function() codec search function. Previously, the
_alias_mbcs() was never used.
Fix the test_codecs.test_mbcs_alias() test: use the current ANSI code
page, not a fake ANSI code page number.
Remove the test_site.test_aliasing_mbcs() test: the alias is now
implemented in the encodings module, no longer in the site module.
It was added as part of #29222 to avoid running freeze tool tests on the
buildbots but the logic was wrong so it did not skip tests on typical posix
setup buildbots where the worker is launched from cron via an @reboot task and
thus have no USER environment variable. This uses the canonical
`getpass.getuser()` API rather than rolling its own attempt.
* Revert "bpo-45173 Remove configparser deprecations"
This reverts commit df2284bc41.
* bpo-45173: Note these configparser deprecations will be removed in 3.12
Per Pitrou:
> The original intent for the “accessor” thing was to have a variant that did all accesses under a filesystem tree in a race condition-free way using openat and friends. It turned out to be much too hairy to actually implement, so was entirely abandoned, but the accessor abstraction was left there.
https://discuss.python.org/t/make-pathlib-extensible/3428/2
Accessors are:
- Lacking any internal purpose - '_NormalAccessor' is the only implementation
- Lacking any firm conceptual difference to `Path` objects themselves (inc. subclasses)
- Non-public, i.e. underscore prefixed - '_Accessor' and '_NormalAccessor'
- Unofficially used to implement customized `Path` objects, but once once [bpo-24132]() is addressed there will be a supported route for that.
This patch preserves all existing behaviour.
Fix test_gdb.test_pycfunction() for Python built with clang -Og.
Tolerate inlined functions in the gdb traceback.
When _testcapimodule.c is built by clang -Og, _null_to_none() is
inlined in meth_varargs() and so gdb returns _null_to_none() as
the frame #1. If it's not inlined, meth_varargs() is the frame #1.
musl libc and gnu libc are not ABI compatible so we need set different
SOABI for musl and not simply assume that all linux is linux-gnu.
Replace linux-gnu with the detected os for the build from config.guess
for linux-musl*.
* Add PRECALL_FUNCTION opcode.
* Move 'call shape' varaibles into struct.
* Replace CALL_NO_KW and CALL_KW with KW_NAMES and CALL instructions.
* Specialize for builtin methods taking using the METH_FASTCALL | METH_KEYWORDS protocol.
* Allow kwnames for specialized calls to builtin types.
* Specialize calls to tuple(arg) and str(arg).
* Substitution with a list of types returns now a tuple of types.
* Substitution with Concatenate returns now a Concatenate with
concatenated lists of arguments.
* Substitution with Ellipsis is not supported.
Previously this didn't matter because there weren't any valid code paths
that could trigger a type check with a special form, but after the bug
fix for `Annotated` wrapping special forms it's now possible to annotate
something like `Annotated['ClassVar[int]', (3, 4)]`. This change would
also be needed for proposed future changes, such as allowing `ClassVar`
and `Final` to nest each other in dataclasses.
We treat Annotated type arg as class-level annotation. This exempts it from checks against Final and ClassVar in order to allow using them in any nesting order.
Automerge-Triggered-By: GH:gvanrossum
When configuring the logging stack, accept already built filters (or
just callables) in the filters array of loggers and handlers.
This facilitates passing quick callables as filters.
Automerge-Triggered-By: GH:vsajip
Add the following info to test.pythoninfo:
* windows.ver: output of the shell "ver" command
* windows.version and windows.version_caption: output of the
"wmic os get Caption,Version /value" command.
Add 'static_exceptions' list to factorize code between
_PyExc_InitTypes() and _PyBuiltins_AddExceptions().
_PyExc_InitTypes() does nothing if it's not the main interpreter.
Sort exceptions in Lib/test/exception_hierarchy.txt.
The remove_subclass() function now deletes the dictionary when
removing the last subclass (if the dictionary becomes empty) to save
memory: set PyTypeObject.tp_subclasses to NULL. remove_subclass() is
called when a type is deallocated.
_PyType_GetSubclasses() no longer holds a reference to tp_subclasses:
its loop cannot modify tp_subclasses.
The test tested that explicitly deleting the local variable bound to the exception
did not cause problems, but it did not test what it actually claimed to test, i.e.
that the variable is deleted automatically.
Ensure that directory file descriptors refer to directories different
from the current directory, and that src_dir_fd and dst_dir_fd refer
to different directories.
Add context manager open_dir_fd() in test.support.os_helper.
The _curses module now creates its ncurses_version type as a heap
type using PyStructSequence_NewType(), rather than using a static
type.
* Move _PyStructSequence_FiniType() definition to pycore_structseq.h.
* test.pythoninfo: log curses.ncurses_version.
* _testembed_Py_Initialize() now uses the PyConfig API, rather than
deprecated Py_SetProgramName().
* Reduce INIT_LOOPS from 16 to 4: test_embed now takes 8.7 seconds
rather than 14.7 seconds.
Add _PyStructSequence_FiniType() and _PyStaticType_Dealloc()
functions to finalize a structseq static type in Py_Finalize().
Currrently, these functions do nothing if Python is built in release
mode.
Clear static types:
* AsyncGenHooksType: sys.set_asyncgen_hooks()
* FlagsType: sys.flags
* FloatInfoType: sys.float_info
* Hash_InfoType: sys.hash_info
* Int_InfoType: sys.int_info
* ThreadInfoType: sys.thread_info
* UnraisableHookArgsType: sys.unraisablehook
* VersionInfoType: sys.version
* WindowsVersionType: sys.getwindowsversion()
Fix an uncaught exception during help text generation when
argparse.BooleanOptionalAction is used with default=argparse.SUPPRESS
and help is specified.
* Add RETURN_GENERATOR and JUMP_NO_INTERRUPT opcodes.
* Trim frame and generator by word each.
* Minor refactor of frame.c
* Update test.test_sys to account for smaller frames.
* Treat generator functions as normal functions when evaluating and specializing.
Use common error message for non-string attribute name in the builtin
functions getattr and hasattr.
The special check no longer needed since Python 3.0.
- skip doctest that changes depending on target system
- skip doctest that only fails on CI
- substitute in values that change depending on target system
test_capi.test_export_symbols() doesn't check if Python exports the
"PyThread_get_thread_native_id" symbol if the _thread.get_native_id()
function is not available (if the PY_HAVE_THREAD_NATIVE_ID macro is
not defined).
Skip test_builtin PTY tests on non-ASCII characters if the readline
module is loaded. The readline module changes input() behavior, but
test_builtin is not intented to test the readline module.
When the readline module is loaded, PyOS_Readline() uses the readline
implementation. In some cases, the Python readline callback
rlhandler() is called by readline with a string without non-ASCII
characters.
Undo rejected PEP-663 changes:
- restore `repr()` to its 3.10 status
- restore `str()` to its 3.10 status
New changes:
- `IntEnum` and `IntFlag` now leave `__str__` as the original `int.__str__` so that str() and format() return the same result
- zero-valued flags without a name have a slightly changed repr(), e.g. `repr(Color(0)) == '<Color: 0>'`
- update `dir()` for mixed-in types to return all the methods and attributes of the mixed-in type
- added `_numeric_repr_` to `Flag` to control display of unnamed values
- enums without doc strings have a more comprehensive doc string added
- `ReprEnum` added -- inheriting from this makes it so only `__repr__` is replaced, not `__str__` nor `__format__`; `IntEnum`, `IntFlag`, and `StrEnum` all inherit from `ReprEnum`
When Python is built without --enable-shared, the "python" program is
now linked to object files, rather than being linked to the Python
library (libpython.a), to make sure that all symbols are exported.
Previously, the linker omitted some symbols like the Py_FrozenMain()
function.
When Python is configured with --without-static-libpython, the Python
static library (libpython.a) is no longer built.
* Check --without-static-libpython earlier in configure.ac
* Add LINK_PYTHON_OBJS and LINK_PYTHON_DEPS variables to Makefile.
* test_capi now ensures that the "Py_FrozenMain" symbol is exported.
Add `string.Template.get_identifiers()` method that returns the identifiers within the template. By default, raises an error if it encounters an invalid identifier (like `substitute()`). The keyword-only argument `raise_on_invalid` can be set to `False` to ignore invalid identifiers (like `safe_substitute()`).
Automerge-Triggered-By: GH:warsaw
Move almost all private functions of Include/cpython/fileutils.h to
the internal C API Include/internal/pycore_fileutils.h.
Only keep _Py_fopen_obj() in Include/cpython/fileutils.h, since it's
used by _testcapi which must not use the internal C API.
Move EncodeLocaleEx() and DecodeLocaleEx() functions from _testcapi
to _testinternalcapi, since the C API moved to the internal C API.
In `__set_name__` there is a check for the `_value_` attribute and an attempt to add it if missing; this adds a test to cover the case for simple enums with a custom `__new__` method.
add tests that exercise the `_missing_` error path for `Flag` and `IntFlag`
Co-authored-by: Alex Waygood <Alex.Waygood@Gmail.com>
Co-authored-by: Ethan Furman <ethan@stoneleaf.us>
* bpo-20369: concurrent.futures.wait() now deduplicates futures given as arg.
* 📜🤖 Added by blurb_it.
Co-authored-by: blurb-it[bot] <43283697+blurb-it[bot]@users.noreply.github.com>
x_mul()'s squaring code can do some redundant and/or useless
work at the end of each digit pass. A more careful analysis
of worst-case carries at various digit positions allows
making that code leaner.
* bpo-46218: Change long_pow() to sliding window algorithm
The primary motivation is to eliminate long_pow's reliance on that the number of bits in a long "digit" is a multiple of 5. Now it no longer cares how many bits are in a digit.
But the sliding window approach also allows cutting the precomputed table of small powers in half, which reduces initialization overhead enough that the approach pays off for smaller exponents too. Depending on exponent bit patterns, a sliding window may also be able to save some bigint multiplies (sometimes when at least 5 consecutive exponent bits are 0, regardless of their starting bit position modulo 5).
Note: boosting the window width to 6 didn't work well overall. It give marginal speed improvements for huge exponents, but the increased overhead (the small-power table needs twice as many entries) made it a loss for smaller exponents.
Co-authored-by: Oleg Iarygin <dralife@yandex.ru>
The registry() method of functools.singledispatch() functions checks now
the first argument or the first parameter annotation and raises a TypeError if it is
not supported. Previously unsupported "types" were ignored (e.g. typing.List[int])
or caused an error at calling time (e.g. list[int]).