mirror of
https://github.com/pybind/pybind11.git
synced 2026-03-14 20:27:47 +00:00
d4f9cfbc28
* Limit busy-wait loops in per-subinterpreter GIL test Add explicit timeouts to the busy-wait coordination loops in the Per-Subinterpreter GIL test in tests/test_with_catch/test_subinterpreter.cpp. Previously those loops spun indefinitely waiting for shared atomics like `started` and `sync` to change, which is fine when CPython's free-threading and per-interpreter GIL behavior matches the test's expectations but becomes pathologically bad when that behavior regresses: the `test_with_catch` executable can then hang forever, causing our 3.14t CI jobs to time out after 90 minutes. This change keeps the structure and intent of the test but adds a std::chrono::steady_clock deadline to each of the coordination loops, using a conservative 10 second bound. Worker threads record a failure and return if they hit the timeout, while the main thread fails the test via Catch2 instead of hanging. That way, if future CPython free-threading patches change the semantics again, the test will fail quickly and produced a diagnosable error instead of wedging the CI job. * Revert "Limit busy-wait loops in per-subinterpreter GIL test" This reverts commit7847adacda. * Add progress reporter for test_with_catch Catch runner Introduce a custom Catch2 reporter for tests/test_with_catch that prints a simple one-line status for each test case as it starts and ends, and wire the cpptest CMake target to invoke test_with_catch with -r progress. This makes it much easier to see where the embedded/interpreter test binary is spending its time in CI logs, and in particular to pinpoint which test case is stuck when the free-threading builds hang. Compared to adding ad hoc timeouts around potentially infinite busy-wait loops in individual tests, a progress reporter is a more general and robust approach: it gives visibility into all tests (including future ones) without changing their behavior, and turns otherwise opaque 90-minute timeouts into locatable issues in the Catch output. * Temporarily limit CI to Python 3.14t free-threading jobs * Temporarily remove non-CI GitHub workflow files * Temporarily disable AppVeyor builds via skip_commits * Add DEBUG_LOOK in TEST_CASE("Move Subinterpreter") * Add Python version banner to Catch progress reporter Print the CPython version once at the start of the Catch-based interpreter tests using Py_GetVersion(). This makes it trivial to confirm which free-threaded build a failing run is using when inspecting CI or local logs. * Revert "Add DEBUG_LOOK in TEST_CASE("Move Subinterpreter")" This reverts commitad3e1c34ce. * Pin CI free-threaded runs to CPython 3.14.0t Update the standard-small and standard-large GitHub Actions jobs to request python-version 3.14.0t instead of 3.14t. This forces setup-python to use the last-known-good 3.14.0 free-threaded build rather than the newer 3.14.1+ builds where subinterpreter finalization regressed. * Revert "Pin CI free-threaded runs to CPython 3.14.0t" This reverts commit5281e1c20c. * Revert "Temporarily disable AppVeyor builds via skip_commits" This reverts commited11292636. * Revert "Temporarily remove non-CI GitHub workflow files" This reverts commit0fe6a42a04. * Revert "Temporarily limit CI to Python 3.14t free-threading jobs" This reverts commit60ae0e8f74. * Pin CI free-threaded runs to CPython 3.14.0t Update the standard-small and standard-large GitHub Actions jobs to request python-version 3.14.0t instead of 3.14t. This forces setup-python to use the last-known-good 3.14.0 free-threaded build rather than the newer 3.14.1+ builds where subinterpreter finalization regressed. * Switch NVHPC job to ubuntu-24.04 and disable AppVeyor * Temporarily trim workflows to focus on NVHPC job * First restore ci.yml from test-with-catch-timeouts branch, then delete all jobs except ubuntu-nvhpc7 * Change runner to ubuntu-24.04 * Use nvhpc-25-11 * Undo ALL changes relative to master (i.e. this branch is now an exact copy of master) * Change runner to ubuntu-24.04 * Use nvhpc-25-11 * Remove misleading 7 from job name (i.e. ubuntu-nvhpc7 → ubuntu-nvhpc)
.. figure:: https://github.com/pybind/pybind11/raw/master/docs/pybind11-logo.png
:alt: pybind11 logo
**pybind11 (v3) — Seamless interoperability between C++ and Python**
|Latest Documentation Status| |Stable Documentation Status| |Gitter chat| |GitHub Discussions|
|CI| |Build status| |SPEC 4 — Using and Creating Nightly Wheels|
|Repology| |PyPI package| |Conda-forge| |Python Versions|
`Setuptools example <https://github.com/pybind/python_example>`_
• `Scikit-build example <https://github.com/pybind/scikit_build_example>`_
• `CMake example <https://github.com/pybind/cmake_example>`_
.. start
**pybind11** is a lightweight header-only library that exposes C++ types
in Python and vice versa, mainly to create Python bindings of existing
C++ code. Its goals and syntax are similar to the excellent
`Boost.Python <http://www.boost.org/doc/libs/1_58_0/libs/python/doc/>`_
library by David Abrahams: to minimize boilerplate code in traditional
extension modules by inferring type information using compile-time
introspection.
The main issue with Boost.Python—and the reason for creating such a
similar project—is Boost. Boost is an enormously large and complex suite
of utility libraries that works with almost every C++ compiler in
existence. This compatibility has its cost: arcane template tricks and
workarounds are necessary to support the oldest and buggiest of compiler
specimens. Now that C++11-compatible compilers are widely available,
this heavy machinery has become an excessively large and unnecessary
dependency.
Think of this library as a tiny self-contained version of Boost.Python
with everything stripped away that isn't relevant for binding
generation. Without comments, the core header files only require ~4K
lines of code and depend on Python (CPython 3.8+, PyPy, or GraalPy) and the C++
standard library. This compact implementation was possible thanks to some C++11
language features (specifically: tuples, lambda functions and variadic
templates). Since its creation, this library has grown beyond Boost.Python in
many ways, leading to dramatically simpler binding code in many common
situations.
Tutorial and reference documentation is provided at
`pybind11.readthedocs.io <https://pybind11.readthedocs.io/en/latest>`_.
A PDF version of the manual is available
`here <https://pybind11.readthedocs.io/_/downloads/en/latest/pdf/>`_.
And the source code is always available at
`github.com/pybind/pybind11 <https://github.com/pybind/pybind11>`_.
Core features
-------------
pybind11 can map the following core C++ features to Python:
- Functions accepting and returning custom data structures per value,
reference, or pointer
- Instance methods and static methods
- Overloaded functions
- Instance attributes and static attributes
- Arbitrary exception types
- Enumerations
- Callbacks
- Iterators and ranges
- Custom operators
- Single and multiple inheritance
- STL data structures
- Smart pointers with reference counting like ``std::shared_ptr``
- Internal references with correct reference counting
- C++ classes with virtual (and pure virtual) methods can be extended
in Python
- Integrated NumPy support (NumPy 2 requires pybind11 2.12+)
Goodies
-------
In addition to the core functionality, pybind11 provides some extra
goodies:
- CPython 3.8+, PyPy3 7.3.17+, and GraalPy 24.1+ are supported with an
implementation-agnostic interface (see older versions for older CPython
and PyPy versions).
- It is possible to bind C++11 lambda functions with captured
variables. The lambda capture data is stored inside the resulting
Python function object.
- pybind11 uses C++11 move constructors and move assignment operators
whenever possible to efficiently transfer custom data types.
- It's easy to expose the internal storage of custom data types through
Pythons' buffer protocols. This is handy e.g. for fast conversion
between C++ matrix classes like Eigen and NumPy without expensive
copy operations.
- pybind11 can automatically vectorize functions so that they are
transparently applied to all entries of one or more NumPy array
arguments.
- Python's slice-based access and assignment operations can be
supported with just a few lines of code.
- Everything is contained in just a few header files; there is no need
to link against any additional libraries.
- Binaries are generally smaller by a factor of at least 2 compared to
equivalent bindings generated by Boost.Python. A recent pybind11
conversion of PyRosetta, an enormous Boost.Python binding project,
`reported <https://graylab.jhu.edu/Sergey/2016.RosettaCon/PyRosetta-4.pdf>`_
a binary size reduction of **5.4x** and compile time reduction by
**5.8x**.
- Function signatures are precomputed at compile time (using
``constexpr``), leading to smaller binaries.
- With little extra effort, C++ types can be pickled and unpickled
similar to regular Python objects.
Supported platforms & compilers
-------------------------------
pybind11 is exercised in continuous integration across a range of operating
systems, Python versions, C++ standards, and toolchains. For an up-to-date
view of the combinations we currently test, please see the
`pybind11 GitHub Actions <https://github.com/pybind/pybind11/actions?query=branch%3Amaster>`_
logs.
The test matrix naturally evolves over time as older platforms and compilers
fall out of use and new ones are added by the community. Closely related
versions of a tested compiler or platform will often work as well in practice,
but we cannot promise to validate every possible combination. If a
configuration you rely on is missing from the matrix or regresses, issues and
pull requests to extend coverage are very welcome. At the same time, we need
to balance the size of the test matrix with the available CI resources,
such as GitHub's limits on concurrent jobs under the free tier.
About
-----
This project was created by `Wenzel
Jakob <http://rgl.epfl.ch/people/wjakob>`_. Significant features and/or
improvements to the code were contributed by
Jonas Adler,
Lori A. Burns,
Sylvain Corlay,
Eric Cousineau,
Aaron Gokaslan,
Ralf Grosse-Kunstleve,
Trent Houliston,
Axel Huebl,
@hulucc,
Yannick Jadoul,
Sergey Lyskov,
Johan Mabille,
Tomasz Miąsko,
Dean Moldovan,
Ben Pritchard,
Jason Rhinelander,
Boris Schäling,
Pim Schellart,
Henry Schreiner,
Ivan Smirnov,
Dustin Spicuzza,
Boris Staletic,
Ethan Steinberg,
Patrick Stewart,
Ivor Wanders,
and
Xiaofei Wang.
We thank Google for a generous financial contribution to the continuous
integration infrastructure used by this project.
Contributing
~~~~~~~~~~~~
See the `contributing
guide <https://github.com/pybind/pybind11/blob/master/.github/CONTRIBUTING.md>`_
for information on building and contributing to pybind11.
License
~~~~~~~
pybind11 is provided under a BSD-style license that can be found in the
`LICENSE <https://github.com/pybind/pybind11/blob/master/LICENSE>`_
file. By using, distributing, or contributing to this project, you agree
to the terms and conditions of this license.
.. |Latest Documentation Status| image:: https://readthedocs.org/projects/pybind11/badge?version=latest
:target: http://pybind11.readthedocs.org/en/latest
.. |Stable Documentation Status| image:: https://img.shields.io/badge/docs-stable-blue.svg
:target: http://pybind11.readthedocs.org/en/stable
.. |Gitter chat| image:: https://img.shields.io/gitter/room/gitterHQ/gitter.svg
:target: https://gitter.im/pybind/Lobby
.. |CI| image:: https://github.com/pybind/pybind11/workflows/CI/badge.svg
:target: https://github.com/pybind/pybind11/actions
.. |Build status| image:: https://ci.appveyor.com/api/projects/status/riaj54pn4h08xy40?svg=true
:target: https://ci.appveyor.com/project/wjakob/pybind11
.. |PyPI package| image:: https://img.shields.io/pypi/v/pybind11.svg
:target: https://pypi.org/project/pybind11/
.. |Conda-forge| image:: https://img.shields.io/conda/vn/conda-forge/pybind11.svg
:target: https://github.com/conda-forge/pybind11-feedstock
.. |Repology| image:: https://repology.org/badge/latest-versions/python:pybind11.svg
:target: https://repology.org/project/python:pybind11/versions
.. |Python Versions| image:: https://img.shields.io/pypi/pyversions/pybind11.svg
:target: https://pypi.org/project/pybind11/
.. |GitHub Discussions| image:: https://img.shields.io/static/v1?label=Discussions&message=Ask&color=blue&logo=github
:target: https://github.com/pybind/pybind11/discussions
.. |SPEC 4 — Using and Creating Nightly Wheels| image:: https://img.shields.io/badge/SPEC-4-green?labelColor=%23004811&color=%235CA038
:target: https://scientific-python.org/specs/spec-0004/