Metadata-Version: 2.4
Name: python-gmp
Version: 0.4.0
Summary: Safe bindings to the GNU GMP library
Keywords: gmp,multiple-precision,arbitrary-precision,bignum
Author-Email: Sergey B Kirpichev <skirpichev@gmail.com>
Maintainer-Email: Sergey B Kirpichev <skirpichev@gmail.com>
License-Expression: MIT
License-File: LICENSE
Classifier: Development Status :: 4 - Beta
Classifier: Intended Audience :: Developers
Classifier: Intended Audience :: Science/Research
Classifier: Natural Language :: English
Classifier: Operating System :: POSIX
Classifier: Programming Language :: C
Classifier: Programming Language :: Python :: 3
Classifier: Programming Language :: Python :: 3 :: Only
Classifier: Programming Language :: Python :: 3.9
Classifier: Programming Language :: Python :: 3.10
Classifier: Programming Language :: Python :: 3.11
Classifier: Programming Language :: Python :: 3.12
Classifier: Programming Language :: Python :: 3.13
Classifier: Programming Language :: Python :: 3.14
Classifier: Programming Language :: Python :: Free Threading :: 2 - Beta
Classifier: Programming Language :: Python :: Implementation :: CPython
Classifier: Programming Language :: Python :: Implementation :: PyPy
Classifier: Programming Language :: Python :: Implementation :: GraalPy
Classifier: Operating System :: Microsoft :: Windows
Classifier: Operating System :: POSIX
Classifier: Operating System :: Unix
Classifier: Operating System :: MacOS
Classifier: Topic :: Scientific/Engineering :: Mathematics
Classifier: Topic :: Software Development :: Libraries :: Python Modules
Project-URL: Homepage, https://github.com/diofant/python-gmp
Project-URL: Source Code, https://github.com/diofant/python-gmp
Project-URL: Bug Tracker, https://github.com/diofant/python-gmp/issues
Project-URL: Documentation, https://python-gmp.readthedocs.io/en/latest/
Requires-Python: >=3.9
Provides-Extra: tests
Requires-Dist: pytest; extra == "tests"
Requires-Dist: hypothesis; extra == "tests"
Requires-Dist: mpmath>=1.4.0a7; extra == "tests"
Provides-Extra: docs
Requires-Dist: sphinx>=8.2; extra == "docs"
Provides-Extra: develop
Requires-Dist: python-gmp[docs,tests]; extra == "develop"
Requires-Dist: pre-commit; extra == "develop"
Requires-Dist: pyperf; extra == "develop"
Description-Content-Type: text/x-rst

Python-GMP
==========

Python extension module, gmp, providing safe bindings to the GNU GMP (version
6.3.0 or later required).  This module shouldn't crash the interpreter.

The gmp can be used as a `gmpy2`_/`python-flint`_ replacement to provide
integer type (`mpz`_), compatible with Python's `int`_.  It includes few
functions (`factorial`_, `gcd`_ and `isqrt`_), compatible with the Python
stdlib's module `math`_.

This module requires Python 3.9 or later versions and has been tested with
CPython 3.9 through 3.14, with PyPy 3.11 and with GraalPy 24.2.  Free-threading
builds of the CPython are supported.

Releases are available in the Python Package Index (PyPI) at
https://pypi.org/project/python-gmp/.


Warning on --disable-alloca configure option
--------------------------------------------

Most GMP packages enable using alloca() for temporary workspace allocation.
This module can't prevent a crash in case of a stack overflow.  To avoid this,
you should compile the GMP library with '--disable-alloca' configure option to
use rather the heap for all temporary allocations.

Published on the PyPI binary wheels aren't affected by this issue.


Warning on using mp_set_memory_functions()
------------------------------------------

This extension customize memory allocation routines, used by the GMP.  Don't
import it together with other GMP bindings, like the `gmpy2`_.


Motivation
----------

The CPython (and most other Python implementations, like PyPy) is optimized to
work with small integers.  Algorithms used here for "big enough" integers
usually aren't best known in the field.  Fortunately, it's possible to use
bindings (for example, the `gmpy2`_ package) to the GNU Multiple Precision
Arithmetic Library (GMP), which aims to be faster than any other bignum library
for all operand sizes.

But such extension modules usually rely on default GMP's memory allocation
functions and can't recover from errors such as out of memory.  So, it's easy
to crash the Python interpreter during the interactive session.  Following
example with the gmpy2 will work if you set address space limit for the Python
interpreter (e.g. by ``prlimit`` command on Linux):

.. code:: pycon

   >>> import gmpy2
   >>> gmpy2.__version__
   '2.2.1'
   >>> z = gmpy2.mpz(29925959575501)
   >>> while True:  # this loop will crash interpter
   ...     z = z*z
   ...
   GNU MP: Cannot allocate memory (size=46956584)
   Aborted

The gmp module handles such errors correctly:

.. code:: pycon

   >>> import gmp
   >>> z = gmp.mpz(29925959575501)
   >>> while True:
   ...     z = z*z
   ...
   Traceback (most recent call last):
     File "<python-input-3>", line 2, in <module>
       z = z*z
           ~^~
   MemoryError
   >>> # interpreter still works, all variables in
   >>> # the current scope are available,
   >>> z.bit_length()  # including pre-failure value of z
   93882077

.. _gmpy2: https://pypi.org/project/gmpy2/
.. _python-flint: https://pypi.org/project/python-flint/
.. _mpz: https://python-gmp.readthedocs.io/en/latest/#gmp.mpz
.. _int: https://docs.python.org/3/library/functions.html#int
.. _factorial: https://python-gmp.readthedocs.io/en/latest/#gmp.factorial
.. _gcd: https://python-gmp.readthedocs.io/en/latest/#gmp.gcd
.. _isqrt: https://python-gmp.readthedocs.io/en/latest/#gmp.isqrt
.. _math: https://docs.python.org/3/library/math.html#number-theoretic-functions
