mpage [Wed, 6 Nov 2024 20:04:04 +0000 (12:04 -0800)]
gh-115999: Introduce helpers for (un)specializing instructions (#126414)
Introduce helpers for (un)specializing instructions
Consolidate the code to specialize/unspecialize instructions into
two helper functions and use them in `_Py_Specialize_BinaryOp`.
The resulting code is more concise and keeps all of the logic at
the point where we decide to specialize/unspecialize an instruction.
Donghee Na [Wed, 6 Nov 2024 03:35:10 +0000 (12:35 +0900)]
gh-115999: Add free-threaded specialization for CONTAINS_OP (gh-126450)
- The specialization logic determines the appropriate specialization using only the operand's type, which is safe to read non-atomically (changing it requires stopping the world). We are guaranteed that the type will not change in between when it is checked and when we specialize the bytecode because the types involved are immutable (you cannot assign to `__class__` for exact instances of `dict`, `set`, or `frozenset`). The bytecode is mutated atomically using helpers.
- The specialized instructions rely on the operand type not changing in between the `DEOPT_IF` checks and the calls to the appropriate type-specific helpers (e.g. `_PySet_Contains`). This is a correctness requirement in the default builds and there are no changes to the opcodes in the free-threaded builds that would invalidate this.
Barney Gale [Tue, 5 Nov 2024 21:19:36 +0000 (21:19 +0000)]
pathlib ABCs: defer path joining (#126409)
Defer joining of path segments in the private `PurePathBase` ABC. The new
behaviour matches how the public `PurePath` class handles path segments.
This removes a hard-to-grok difference between the ABCs and the main
classes. It also slightly reduces the size of `PurePath` objects by
eliminating a `_raw_path` slot.
T. Wouters [Tue, 5 Nov 2024 14:49:27 +0000 (15:49 +0100)]
gh-115999: Fix gdb support for libpython.so after thread-local bytecode change (#126440)
Fix the gdb pretty printer in the face of --enable-shared by delaying the attempt to load the _PyInterpreterFrame definition until after .so files are loaded.
- Explicit memory management for `None` objects (since we still try to treat immortal objects as regular objects)
- Respect possible errors of `sys.monitoring.register_callback` call
lit [Tue, 5 Nov 2024 02:58:15 +0000 (10:58 +0800)]
gh-125436: Doc: Add missing ``allow_unnamed_section`` parameter to ``ConfigParser`` documentation (#125437)
Add missing ``allow_unnamed_section`` parameter to ``ConfigParser`` doc, as well as to it's parent ``RawConfigParser``.
Split too long line on ``ConfigParser`` signature.
Add some sections about when some of ``RawConfigParser`` parameters were added.
Barney Gale [Mon, 4 Nov 2024 19:29:57 +0000 (19:29 +0000)]
GH-126363: Speed up pattern parsing in `pathlib.Path.glob()` (#126364)
The implementation of `Path.glob()` does rather a hacky thing: it calls
`self.with_segments()` to convert the given pattern to a `Path` object, and
then peeks at the private `_raw_path` attribute to see if pathlib removed a
trailing slash from the pattern.
In this patch, we make `glob()` use a new `_parse_pattern()` classmethod
that splits the pattern into parts while preserving information about any
trailing slash. This skips the cost of creating a `Path` object, and avoids
some path anchor normalization, which makes `Path.glob()` slightly faster.
But mostly it's about making the code less naughty.
mpage [Mon, 4 Nov 2024 19:13:32 +0000 (11:13 -0800)]
gh-115999: Implement thread-local bytecode and enable specialization for `BINARY_OP` (#123926)
Each thread specializes a thread-local copy of the bytecode, created on the first RESUME, in free-threaded builds. All copies of the bytecode for a code object are stored in the co_tlbc array on the code object. Threads reserve a globally unique index identifying its copy of the bytecode in all co_tlbc arrays at thread creation and release the index at thread destruction. The first entry in every co_tlbc array always points to the "main" copy of the bytecode that is stored at the end of the code object. This ensures that no bytecode is copied for programs that do not use threads.
Thread-local bytecode can be disabled at runtime by providing either -X tlbc=0 or PYTHON_TLBC=0. Disabling thread-local bytecode also disables specialization.
Concurrent modifications to the bytecode made by the specializing interpreter and instrumentation use atomics, with specialization taking care not to overwrite an instruction that was instrumented concurrently.
Zhikang Yan [Mon, 4 Nov 2024 04:08:15 +0000 (12:08 +0800)]
gh-126165: Improve docs of function `math.isclose` (#126215)
Co-authored-by: Sergey B Kirpichev <skirpichev@gmail.com> Co-authored-by: Carol Willing <carolcode@willingconsulting.com> Co-authored-by: Terry Jan Reedy <tjreedy@udel.edu>
gh-99880: document rounding mode for new-style formatting (GH-121481)
* gh-99880: document rounding mode for new-style formatting
The CPython uses _Py_dg_dtoa(), which does rounding to nearest with half
to even tie-breaking rule.
If that functions is unavailable, PyOS_double_to_string() fallbacks to
system snprintf(). Since CPython 3.12, build requirements include C11
compiler *and* support for IEEE 754 floating point numbers (Annex F).
This means that FE_TONEAREST macro is available and, per default,
printf-like functions should use same rounding mode as _Py_dg_dtoa().
mpage [Fri, 1 Nov 2024 23:10:58 +0000 (16:10 -0700)]
gh-126255: Ignore warning about JIT being deactivated when perf support is active in `test_embed.InitConfigTests.test_initconfig_api` (#126302)
Temporarily ignore warnings about JIT deactivation when perf support is active.
This will be reverted as soon as a way is found to determine at run time whether the interpreter was built with JIT. Currently, this is not possible on Windows.
Co-authored-by: Kirill Podoprigora <kirill.bast9@mail.ru> Co-authored-by: Ken Jin <28750310+Fidget-Spinner@users.noreply.github.com> Co-authored-by: Pablo Galindo <pablogsal@gmail.com>
Cody Maloney [Fri, 1 Nov 2024 21:50:49 +0000 (14:50 -0700)]
gh-120754: _io Ensure stat cache is cleared on fd change (#125166)
Performed an audit of `fileio.c` and `_pyio` and made sure anytime the
fd changes the stat result, if set, is also cleared/changed.
There's one case where it's not cleared, if code would clear it in
__init__, keep the memory allocated and just do another fstat with the
existing memory.
docs: add a more precise example in enum doc (GH-121015)
* docs: add a more precise example
Previous example used manual integer value assignment in class based declaration but in functional syntax has been used auto value assignment what could be confusing for the new users. Additionally documentation doesn't show how to declare new enum via functional syntax with usage of the manual value assignment.
mpage [Fri, 1 Nov 2024 15:53:03 +0000 (08:53 -0700)]
gh-126211: Exclude preprocessor directives from statements containing escaping calls (#126213)
The cases generator inserts code to save and restore the stack pointer around
statements that contain escaping calls. To find the beginning of such statements,
we would walk backwards from the escaping call until we encountered a token that
was treated as a statement terminator. This set of terminators should include
preprocessor directives.
Add `pathlib.Path.scandir()` as a trivial wrapper of `os.scandir()`. This
will be used to implement several `PathBase` methods more efficiently,
including methods that provide `Path.copy()`.
Alperen Keleş [Thu, 31 Oct 2024 21:44:34 +0000 (17:44 -0400)]
gh-126256: Update time.rst to use the same clock as instead of the same clock than (#126257)
Update time.rst to use `the same clock as` instead of `the same clock than`
The time documentation uses the same clock than time.monotonic instead of the same clock as time.monotonic, which is grammatically false. This PR fixes changes two instances of `the same clock than` to `the same clock as`.
Johan Förberg [Wed, 30 Oct 2024 22:08:30 +0000 (23:08 +0100)]
gh-121267: Improve performance of tarfile (#121267) (#121269)
Tarfile in the default write mode spends much of its time resolving UIDs
into usernames and GIDs into group names. By caching these mappings, a
significant speedup can be achieved.
In my simple benchmark[1], this extra caching speeds up tarfile by 8x.
Serhiy Storchaka [Wed, 30 Oct 2024 16:14:27 +0000 (18:14 +0200)]
gh-126068: Fix exceptions in the argparse module (GH-126069)
* Only error messages for ArgumentError and ArgumentTypeError are now
translated.
* ArgumentError is now only used for command line errors, not for logical
errors in the program.
* TypeError is now raised instead of ValueError for some logical errors.