cpython/Doc/using/configure.rst

1014 lines
28 KiB
ReStructuredText
Raw Normal View History

****************
Configure Python
****************
Build Requirements
==================
Features required to build CPython:
* A `C11 <https://en.cppreference.com/w/c/11>`_ compiler. `Optional C11
features
<https://en.wikipedia.org/wiki/C11_(C_standard_revision)#Optional_features>`_
are not required.
* Support for `IEEE 754 <https://en.wikipedia.org/wiki/IEEE_754>`_ floating
point numbers and `floating point Not-a-Number (NaN)
<https://en.wikipedia.org/wiki/NaN#Floating_point>`_.
* Support for threads.
* OpenSSL 1.1.1 or newer for the :mod:`ssl` and :mod:`hashlib` modules.
* On Windows, Microsoft Visual Studio 2017 or later is required.
.. versionchanged:: 3.11
C11 compiler, IEEE 754 and NaN support are now required.
On Windows, Visual Studio 2017 or later is required.
.. versionchanged:: 3.10
OpenSSL 1.1.1 is now required.
.. versionchanged:: 3.7
Thread support and OpenSSL 1.0.2 are now required.
.. versionchanged:: 3.6
Selected C99 features are now required, like ``<stdint.h>`` and ``static
inline`` functions.
.. versionchanged:: 3.5
On Windows, Visual Studio 2015 or later is required.
See also :pep:`7` "Style Guide for C Code" and :pep:`11` "CPython platform
support".
.. _configure-options:
Configure Options
=================
List all ``./configure`` script options using::
./configure --help
See also the :file:`Misc/SpecialBuilds.txt` in the Python source distribution.
General Options
---------------
.. cmdoption:: --enable-loadable-sqlite-extensions
Support loadable extensions in the :mod:`_sqlite` extension module (default
is no).
See the :meth:`sqlite3.Connection.enable_load_extension` method of the
:mod:`sqlite3` module.
.. versionadded:: 3.6
.. cmdoption:: --disable-ipv6
Disable IPv6 support (enabled by default if supported), see the
:mod:`socket` module.
.. cmdoption:: --enable-big-digits=[15|30]
Define the size in bits of Python :class:`int` digits: 15 or 30 bits.
By default, the digit size is 30.
Define the ``PYLONG_BITS_IN_DIGIT`` to ``15`` or ``30``.
See :data:`sys.int_info.bits_per_digit <sys.int_info>`.
.. cmdoption:: --with-suffix=SUFFIX
Set the Python executable suffix to *SUFFIX*.
The default suffix is ``.exe`` on Windows and macOS (``python.exe``
executable), ``.js`` on Emscripten node, ``.html`` on Emscripten browser,
``.wasm`` on WASI, and an empty string on other platforms (``python``
executable).
.. versionchanged:: 3.11
The default suffix on WASM platform is one of ``.js``, ``.html``
or ``.wasm``.
.. cmdoption:: --with-tzpath=<list of absolute paths separated by pathsep>
Select the default time zone search path for :const:`zoneinfo.TZPATH`.
See the :ref:`Compile-time configuration
<zoneinfo_data_compile_time_config>` of the :mod:`zoneinfo` module.
Default: ``/usr/share/zoneinfo:/usr/lib/zoneinfo:/usr/share/lib/zoneinfo:/etc/zoneinfo``.
See :data:`os.pathsep` path separator.
.. versionadded:: 3.9
.. cmdoption:: --without-decimal-contextvar
Build the ``_decimal`` extension module using a thread-local context rather
than a coroutine-local context (default), see the :mod:`decimal` module.
See :const:`decimal.HAVE_CONTEXTVAR` and the :mod:`contextvars` module.
.. versionadded:: 3.9
.. cmdoption:: --with-dbmliborder=<list of backend names>
Override order to check db backends for the :mod:`dbm` module
A valid value is a colon (``:``) separated string with the backend names:
* ``ndbm``;
* ``gdbm``;
* ``bdb``.
.. cmdoption:: --without-c-locale-coercion
Disable C locale coercion to a UTF-8 based locale (enabled by default).
Don't define the ``PY_COERCE_C_LOCALE`` macro.
See :envvar:`PYTHONCOERCECLOCALE` and the :pep:`538`.
.. cmdoption:: --without-freelists
Disable all freelists except the empty tuple singleton.
.. versionadded:: 3.11
.. cmdoption:: --with-platlibdir=DIRNAME
Python library directory name (default is ``lib``).
Fedora and SuSE use ``lib64`` on 64-bit platforms.
See :data:`sys.platlibdir`.
.. versionadded:: 3.9
.. cmdoption:: --with-wheel-pkg-dir=PATH
Directory of wheel packages used by the :mod:`ensurepip` module
(none by default).
Some Linux distribution packaging policies recommend against bundling
dependencies. For example, Fedora installs wheel packages in the
``/usr/share/python-wheels/`` directory and don't install the
:mod:`ensurepip._bundled` package.
.. versionadded:: 3.10
.. cmdoption:: --with-pkg-config=[check|yes|no]
Whether configure should use :program:`pkg-config` to detect build
dependencies.
* ``check`` (default): :program:`pkg-config` is optional
* ``yes``: :program:`pkg-config` is mandatory
* ``no``: configure does not use :program:`pkg-config` even when present
.. versionadded:: 3.11
.. cmdoption:: --enable-pystats
Turn on internal statistics gathering.
The statistics will be dumped to a arbitrary (probably unique) file in
``/tmp/py_stats/``, or ``C:\temp\py_stats\`` on Windows. If that directory
does not exist, results will be printed on stdout.
Use ``Tools/scripts/summarize_stats.py`` to read the stats.
.. versionadded:: 3.11
WebAssembly Options
-------------------
.. cmdoption:: --with-emscripten-target=[browser|node]
Set build flavor for ``wasm32-emscripten``.
* ``browser`` (default): preload minimal stdlib, default MEMFS.
* ``node``: NODERAWFS and pthread support.
.. versionadded:: 3.11
.. cmdoption:: --enable-wasm-dynamic-linking
Turn on dynamic linking support for WASM.
Dynamic linking enables ``dlopen``. File size of the executable
increases due to limited dead code elimination and additional features.
.. versionadded:: 3.11
.. cmdoption:: --enable-wasm-pthreads
Turn on pthreads support for WASM.
.. versionadded:: 3.11
Install Options
---------------
.. cmdoption:: --prefix=PREFIX
Install architecture-independent files in PREFIX. On Unix, it
defaults to :file:`/usr/local`.
This value can be retrived at runtime using :data:`sys.prefix`.
As an example, one can use ``--prefix="$HOME/.local/"`` to install
a Python in its home directory.
.. cmdoption:: --exec-prefix=EPREFIX
Install architecture-dependent files in EPREFIX, defaults to :option:`--prefix`.
This value can be retrived at runtime using :data:`sys.exec_prefix`.
.. cmdoption:: --disable-test-modules
Don't build nor install test modules, like the :mod:`test` package or the
:mod:`_testcapi` extension module (built and installed by default).
.. versionadded:: 3.10
.. cmdoption:: --with-ensurepip=[upgrade|install|no]
Select the :mod:`ensurepip` command run on Python installation:
* ``upgrade`` (default): run ``python -m ensurepip --altinstall --upgrade``
command.
* ``install``: run ``python -m ensurepip --altinstall`` command;
* ``no``: don't run ensurepip;
.. versionadded:: 3.6
Performance options
-------------------
Configuring Python using ``--enable-optimizations --with-lto`` (PGO + LTO) is
gh-90536: Add support for the BOLT post-link binary optimizer (gh-95908) * Add support for the BOLT post-link binary optimizer Using [bolt](https://github.com/llvm/llvm-project/tree/main/bolt) provides a fairly large speedup without any code or functionality changes. It provides roughly a 1% speedup on pyperformance, and a 4% improvement on the Pyston web macrobenchmarks. It is gated behind an `--enable-bolt` configure arg because not all toolchains and environments are supported. It has been tested on a Linux x86_64 toolchain, using llvm-bolt built from the LLVM 14.0.6 sources (their binary distribution of this version did not include bolt). Compared to [a previous attempt](https://github.com/faster-cpython/ideas/issues/224), this commit uses bolt's preferred "instrumentation" approach, as well as adds some non-PIE flags which enable much better optimizations from bolt. The effects of this change are a bit more dependent on CPU microarchitecture than other changes, since it optimizes i-cache behavior which seems to be a bit more variable between architectures. The 1%/4% numbers were collected on an Intel Skylake CPU, and on an AMD Zen 3 CPU I got a slightly larger speedup (2%/4%), and on a c6i.xlarge EC2 instance I got a slightly lower speedup (1%/3%). The low speedup on pyperformance is not entirely unexpected, because BOLT improves i-cache behavior, and the benchmarks in the pyperformance suite are small and tend to fit in i-cache. This change uses the existing pgo profiling task (`python -m test --pgo`), though I was able to measure about a 1% macrobenchmark improvement by using the macrobenchmarks as the training task. I personally think that both the PGO and BOLT tasks should be updated to use macrobenchmarks, but for the sake of splitting up the work this PR uses the existing pgo task. * Simplify the build flags * Add a NEWS entry * Update Makefile.pre.in Co-authored-by: Dong-hee Na <donghee.na92@gmail.com> * Update configure.ac Co-authored-by: Dong-hee Na <donghee.na92@gmail.com> * Add myself to ACKS * Add docs * Other review comments * fix tab/space issue * Make it more clear that --enable-bolt is experimental * Add link to bolt's github page Co-authored-by: Dong-hee Na <donghee.na92@gmail.com>
2022-08-18 18:33:54 -03:00
recommended for best performance. The experimental ``--enable-bolt`` flag can
also be used to improve performance.
.. cmdoption:: --enable-optimizations
Enable Profile Guided Optimization (PGO) using :envvar:`PROFILE_TASK`
(disabled by default).
The C compiler Clang requires ``llvm-profdata`` program for PGO. On
macOS, GCC also requires it: GCC is just an alias to Clang on macOS.
Disable also semantic interposition in libpython if ``--enable-shared`` and
GCC is used: add ``-fno-semantic-interposition`` to the compiler and linker
flags.
.. versionadded:: 3.6
.. versionchanged:: 3.10
Use ``-fno-semantic-interposition`` on GCC.
.. envvar:: PROFILE_TASK
Environment variable used in the Makefile: Python command line arguments for
the PGO generation task.
Default: ``-m test --pgo --timeout=$(TESTTIMEOUT)``.
.. versionadded:: 3.8
.. cmdoption:: --with-lto=[full|thin|no|yes]
Enable Link Time Optimization (LTO) in any build (disabled by default).
The C compiler Clang requires ``llvm-ar`` for LTO (``ar`` on macOS), as well
as an LTO-aware linker (``ld.gold`` or ``lld``).
.. versionadded:: 3.6
.. versionadded:: 3.11
To use ThinLTO feature, use ``--with-lto=thin`` on Clang.
.. versionchanged:: 3.12
Use ThinLTO as the default optimization policy on Clang if the compiler accepts the flag.
gh-90536: Add support for the BOLT post-link binary optimizer (gh-95908) * Add support for the BOLT post-link binary optimizer Using [bolt](https://github.com/llvm/llvm-project/tree/main/bolt) provides a fairly large speedup without any code or functionality changes. It provides roughly a 1% speedup on pyperformance, and a 4% improvement on the Pyston web macrobenchmarks. It is gated behind an `--enable-bolt` configure arg because not all toolchains and environments are supported. It has been tested on a Linux x86_64 toolchain, using llvm-bolt built from the LLVM 14.0.6 sources (their binary distribution of this version did not include bolt). Compared to [a previous attempt](https://github.com/faster-cpython/ideas/issues/224), this commit uses bolt's preferred "instrumentation" approach, as well as adds some non-PIE flags which enable much better optimizations from bolt. The effects of this change are a bit more dependent on CPU microarchitecture than other changes, since it optimizes i-cache behavior which seems to be a bit more variable between architectures. The 1%/4% numbers were collected on an Intel Skylake CPU, and on an AMD Zen 3 CPU I got a slightly larger speedup (2%/4%), and on a c6i.xlarge EC2 instance I got a slightly lower speedup (1%/3%). The low speedup on pyperformance is not entirely unexpected, because BOLT improves i-cache behavior, and the benchmarks in the pyperformance suite are small and tend to fit in i-cache. This change uses the existing pgo profiling task (`python -m test --pgo`), though I was able to measure about a 1% macrobenchmark improvement by using the macrobenchmarks as the training task. I personally think that both the PGO and BOLT tasks should be updated to use macrobenchmarks, but for the sake of splitting up the work this PR uses the existing pgo task. * Simplify the build flags * Add a NEWS entry * Update Makefile.pre.in Co-authored-by: Dong-hee Na <donghee.na92@gmail.com> * Update configure.ac Co-authored-by: Dong-hee Na <donghee.na92@gmail.com> * Add myself to ACKS * Add docs * Other review comments * fix tab/space issue * Make it more clear that --enable-bolt is experimental * Add link to bolt's github page Co-authored-by: Dong-hee Na <donghee.na92@gmail.com>
2022-08-18 18:33:54 -03:00
.. cmdoption:: --enable-bolt
Enable usage of the `BOLT post-link binary optimizer
<https://github.com/llvm/llvm-project/tree/main/bolt>`_ (disabled by
gh-90536: Add support for the BOLT post-link binary optimizer (gh-95908) * Add support for the BOLT post-link binary optimizer Using [bolt](https://github.com/llvm/llvm-project/tree/main/bolt) provides a fairly large speedup without any code or functionality changes. It provides roughly a 1% speedup on pyperformance, and a 4% improvement on the Pyston web macrobenchmarks. It is gated behind an `--enable-bolt` configure arg because not all toolchains and environments are supported. It has been tested on a Linux x86_64 toolchain, using llvm-bolt built from the LLVM 14.0.6 sources (their binary distribution of this version did not include bolt). Compared to [a previous attempt](https://github.com/faster-cpython/ideas/issues/224), this commit uses bolt's preferred "instrumentation" approach, as well as adds some non-PIE flags which enable much better optimizations from bolt. The effects of this change are a bit more dependent on CPU microarchitecture than other changes, since it optimizes i-cache behavior which seems to be a bit more variable between architectures. The 1%/4% numbers were collected on an Intel Skylake CPU, and on an AMD Zen 3 CPU I got a slightly larger speedup (2%/4%), and on a c6i.xlarge EC2 instance I got a slightly lower speedup (1%/3%). The low speedup on pyperformance is not entirely unexpected, because BOLT improves i-cache behavior, and the benchmarks in the pyperformance suite are small and tend to fit in i-cache. This change uses the existing pgo profiling task (`python -m test --pgo`), though I was able to measure about a 1% macrobenchmark improvement by using the macrobenchmarks as the training task. I personally think that both the PGO and BOLT tasks should be updated to use macrobenchmarks, but for the sake of splitting up the work this PR uses the existing pgo task. * Simplify the build flags * Add a NEWS entry * Update Makefile.pre.in Co-authored-by: Dong-hee Na <donghee.na92@gmail.com> * Update configure.ac Co-authored-by: Dong-hee Na <donghee.na92@gmail.com> * Add myself to ACKS * Add docs * Other review comments * fix tab/space issue * Make it more clear that --enable-bolt is experimental * Add link to bolt's github page Co-authored-by: Dong-hee Na <donghee.na92@gmail.com>
2022-08-18 18:33:54 -03:00
default).
BOLT is part of the LLVM project but is not always included in their binary
distributions. This flag requires that ``llvm-bolt`` and ``merge-fdata``
are available.
BOLT is still a fairly new project so this flag should be considered
experimental for now. Because this tool operates on machine code its success
is dependent on a combination of the build environment + the other
optimization configure args + the CPU architecture, and not all combinations
are supported.
BOLT versions before LLVM 16 are known to crash BOLT under some scenarios.
Use of LLVM 16 or newer for BOLT optimization is strongly encouraged.
The :envvar:`!BOLT_INSTRUMENT_FLAGS` and :envvar:`!BOLT_APPLY_FLAGS`
:program:`configure` variables can be defined to override the default set of
arguments for :program:`llvm-bolt` to instrument and apply BOLT data to
binaries, respectively.
gh-90536: Add support for the BOLT post-link binary optimizer (gh-95908) * Add support for the BOLT post-link binary optimizer Using [bolt](https://github.com/llvm/llvm-project/tree/main/bolt) provides a fairly large speedup without any code or functionality changes. It provides roughly a 1% speedup on pyperformance, and a 4% improvement on the Pyston web macrobenchmarks. It is gated behind an `--enable-bolt` configure arg because not all toolchains and environments are supported. It has been tested on a Linux x86_64 toolchain, using llvm-bolt built from the LLVM 14.0.6 sources (their binary distribution of this version did not include bolt). Compared to [a previous attempt](https://github.com/faster-cpython/ideas/issues/224), this commit uses bolt's preferred "instrumentation" approach, as well as adds some non-PIE flags which enable much better optimizations from bolt. The effects of this change are a bit more dependent on CPU microarchitecture than other changes, since it optimizes i-cache behavior which seems to be a bit more variable between architectures. The 1%/4% numbers were collected on an Intel Skylake CPU, and on an AMD Zen 3 CPU I got a slightly larger speedup (2%/4%), and on a c6i.xlarge EC2 instance I got a slightly lower speedup (1%/3%). The low speedup on pyperformance is not entirely unexpected, because BOLT improves i-cache behavior, and the benchmarks in the pyperformance suite are small and tend to fit in i-cache. This change uses the existing pgo profiling task (`python -m test --pgo`), though I was able to measure about a 1% macrobenchmark improvement by using the macrobenchmarks as the training task. I personally think that both the PGO and BOLT tasks should be updated to use macrobenchmarks, but for the sake of splitting up the work this PR uses the existing pgo task. * Simplify the build flags * Add a NEWS entry * Update Makefile.pre.in Co-authored-by: Dong-hee Na <donghee.na92@gmail.com> * Update configure.ac Co-authored-by: Dong-hee Na <donghee.na92@gmail.com> * Add myself to ACKS * Add docs * Other review comments * fix tab/space issue * Make it more clear that --enable-bolt is experimental * Add link to bolt's github page Co-authored-by: Dong-hee Na <donghee.na92@gmail.com>
2022-08-18 18:33:54 -03:00
.. versionadded:: 3.12
.. cmdoption:: --with-computed-gotos
Enable computed gotos in evaluation loop (enabled by default on supported
compilers).
.. cmdoption:: --without-pymalloc
Disable the specialized Python memory allocator :ref:`pymalloc <pymalloc>`
(enabled by default).
See also :envvar:`PYTHONMALLOC` environment variable.
.. cmdoption:: --without-doc-strings
Disable static documentation strings to reduce the memory footprint (enabled
by default). Documentation strings defined in Python are not affected.
Don't define the ``WITH_DOC_STRINGS`` macro.
See the ``PyDoc_STRVAR()`` macro.
.. cmdoption:: --enable-profiling
Enable C-level code profiling with ``gprof`` (disabled by default).
.. cmdoption:: --with-strict-overflow
Add ``-fstrict-overflow`` to the C compiler flags (by default we add
``-fno-strict-overflow`` instead).
.. _debug-build:
Python Debug Build
------------------
A debug build is Python built with the :option:`--with-pydebug` configure
option.
Effects of a debug build:
* Display all warnings by default: the list of default warning filters is empty
in the :mod:`warnings` module.
* Add ``d`` to :data:`sys.abiflags`.
* Add :func:`sys.gettotalrefcount` function.
* Add :option:`-X showrefcount <-X>` command line option.
* Add :option:`-d` command line option and :envvar:`PYTHONDEBUG` environment
variable to debug the parser.
* Add support for the ``__lltrace__`` variable: enable low-level tracing in the
bytecode evaluation loop if the variable is defined.
* Install :ref:`debug hooks on memory allocators <default-memory-allocators>`
to detect buffer overflow and other memory errors.
* Define ``Py_DEBUG`` and ``Py_REF_DEBUG`` macros.
* Add runtime checks: code surrounded by ``#ifdef Py_DEBUG`` and ``#endif``.
Enable ``assert(...)`` and ``_PyObject_ASSERT(...)`` assertions: don't set
the ``NDEBUG`` macro (see also the :option:`--with-assertions` configure
option). Main runtime checks:
* Add sanity checks on the function arguments.
* Unicode and int objects are created with their memory filled with a pattern
to detect usage of uninitialized objects.
* Ensure that functions which can clear or replace the current exception are
not called with an exception raised.
* Check that deallocator functions don't change the current exception.
* The garbage collector (:func:`gc.collect` function) runs some basic checks
on objects consistency.
* The :c:macro:`Py_SAFE_DOWNCAST()` macro checks for integer underflow and
overflow when downcasting from wide types to narrow types.
See also the :ref:`Python Development Mode <devmode>` and the
:option:`--with-trace-refs` configure option.
.. versionchanged:: 3.8
Release builds and debug builds are now ABI compatible: defining the
``Py_DEBUG`` macro no longer implies the ``Py_TRACE_REFS`` macro (see the
:option:`--with-trace-refs` option), which introduces the only ABI
incompatibility.
Debug options
-------------
.. cmdoption:: --with-pydebug
:ref:`Build Python in debug mode <debug-build>`: define the ``Py_DEBUG``
macro (disabled by default).
.. cmdoption:: --with-trace-refs
Enable tracing references for debugging purpose (disabled by default).
Effects:
* Define the ``Py_TRACE_REFS`` macro.
* Add :func:`sys.getobjects` function.
* Add :envvar:`PYTHONDUMPREFS` environment variable.
This build is not ABI compatible with release build (default build) or debug
build (``Py_DEBUG`` and ``Py_REF_DEBUG`` macros).
.. versionadded:: 3.8
.. cmdoption:: --with-assertions
Build with C assertions enabled (default is no): ``assert(...);`` and
``_PyObject_ASSERT(...);``.
If set, the ``NDEBUG`` macro is not defined in the :envvar:`OPT` compiler
variable.
See also the :option:`--with-pydebug` option (:ref:`debug build
<debug-build>`) which also enables assertions.
.. versionadded:: 3.6
.. cmdoption:: --with-valgrind
Enable Valgrind support (default is no).
.. cmdoption:: --with-dtrace
Enable DTrace support (default is no).
See :ref:`Instrumenting CPython with DTrace and SystemTap
<instrumentation>`.
.. versionadded:: 3.6
.. cmdoption:: --with-address-sanitizer
Enable AddressSanitizer memory error detector, ``asan`` (default is no).
.. versionadded:: 3.6
.. cmdoption:: --with-memory-sanitizer
Enable MemorySanitizer allocation error detector, ``msan`` (default is no).
.. versionadded:: 3.6
.. cmdoption:: --with-undefined-behavior-sanitizer
Enable UndefinedBehaviorSanitizer undefined behaviour detector, ``ubsan``
(default is no).
.. versionadded:: 3.6
Linker options
--------------
.. cmdoption:: --enable-shared
Enable building a shared Python library: ``libpython`` (default is no).
.. cmdoption:: --without-static-libpython
Do not build ``libpythonMAJOR.MINOR.a`` and do not install ``python.o``
(built and enabled by default).
.. versionadded:: 3.10
Libraries options
-----------------
.. cmdoption:: --with-libs='lib1 ...'
Link against additional libraries (default is no).
.. cmdoption:: --with-system-expat
Build the :mod:`pyexpat` module using an installed ``expat`` library
(default is no).
.. cmdoption:: --with-system-libmpdec
Build the ``_decimal`` extension module using an installed ``mpdec``
library, see the :mod:`decimal` module (default is no).
.. versionadded:: 3.3
.. cmdoption:: --with-readline=editline
Use ``editline`` library for backend of the :mod:`readline` module.
Define the ``WITH_EDITLINE`` macro.
.. versionadded:: 3.10
.. cmdoption:: --without-readline
Don't build the :mod:`readline` module (built by default).
Don't define the ``HAVE_LIBREADLINE`` macro.
.. versionadded:: 3.10
.. cmdoption:: --with-libm=STRING
Override ``libm`` math library to *STRING* (default is system-dependent).
.. cmdoption:: --with-libc=STRING
Override ``libc`` C library to *STRING* (default is system-dependent).
.. cmdoption:: --with-openssl=DIR
Root of the OpenSSL directory.
.. versionadded:: 3.7
.. cmdoption:: --with-openssl-rpath=[no|auto|DIR]
Set runtime library directory (rpath) for OpenSSL libraries:
* ``no`` (default): don't set rpath;
* ``auto``: auto-detect rpath from :option:`--with-openssl` and
``pkg-config``;
* *DIR*: set an explicit rpath.
.. versionadded:: 3.10
Security Options
----------------
.. cmdoption:: --with-hash-algorithm=[fnv|siphash13|siphash24]
Select hash algorithm for use in ``Python/pyhash.c``:
* ``siphash13`` (default);
* ``siphash24``;
* ``fnv``.
.. versionadded:: 3.4
.. versionadded:: 3.11
``siphash13`` is added and it is the new default.
.. cmdoption:: --with-builtin-hashlib-hashes=md5,sha1,sha256,sha512,sha3,blake2
Built-in hash modules:
* ``md5``;
* ``sha1``;
* ``sha256``;
* ``sha512``;
* ``sha3`` (with shake);
* ``blake2``.
.. versionadded:: 3.9
.. cmdoption:: --with-ssl-default-suites=[python|openssl|STRING]
Override the OpenSSL default cipher suites string:
* ``python`` (default): use Python's preferred selection;
* ``openssl``: leave OpenSSL's defaults untouched;
bpo-43998: Default to TLS 1.2 and increase cipher suite security (GH-25778) The ssl module now has more secure default settings. Ciphers without forward secrecy or SHA-1 MAC are disabled by default. Security level 2 prohibits weak RSA, DH, and ECC keys with less than 112 bits of security. :class:`~ssl.SSLContext` defaults to minimum protocol version TLS 1.2. Settings are based on Hynek Schlawack's research. ``` $ openssl version OpenSSL 1.1.1k FIPS 25 Mar 2021 $ openssl ciphers -v '@SECLEVEL=2:ECDH+AESGCM:ECDH+CHACHA20:ECDH+AES:DHE+AES:!aNULL:!eNULL:!aDSS:!SHA1:!AESCCM' TLS_AES_256_GCM_SHA384 TLSv1.3 Kx=any Au=any Enc=AESGCM(256) Mac=AEAD TLS_CHACHA20_POLY1305_SHA256 TLSv1.3 Kx=any Au=any Enc=CHACHA20/POLY1305(256) Mac=AEAD TLS_AES_128_GCM_SHA256 TLSv1.3 Kx=any Au=any Enc=AESGCM(128) Mac=AEAD TLS_AES_128_CCM_SHA256 TLSv1.3 Kx=any Au=any Enc=AESCCM(128) Mac=AEAD ECDHE-ECDSA-AES256-GCM-SHA384 TLSv1.2 Kx=ECDH Au=ECDSA Enc=AESGCM(256) Mac=AEAD ECDHE-RSA-AES256-GCM-SHA384 TLSv1.2 Kx=ECDH Au=RSA Enc=AESGCM(256) Mac=AEAD ECDHE-ECDSA-AES128-GCM-SHA256 TLSv1.2 Kx=ECDH Au=ECDSA Enc=AESGCM(128) Mac=AEAD ECDHE-RSA-AES128-GCM-SHA256 TLSv1.2 Kx=ECDH Au=RSA Enc=AESGCM(128) Mac=AEAD ECDHE-ECDSA-CHACHA20-POLY1305 TLSv1.2 Kx=ECDH Au=ECDSA Enc=CHACHA20/POLY1305(256) Mac=AEAD ECDHE-RSA-CHACHA20-POLY1305 TLSv1.2 Kx=ECDH Au=RSA Enc=CHACHA20/POLY1305(256) Mac=AEAD ECDHE-ECDSA-AES256-SHA384 TLSv1.2 Kx=ECDH Au=ECDSA Enc=AES(256) Mac=SHA384 ECDHE-RSA-AES256-SHA384 TLSv1.2 Kx=ECDH Au=RSA Enc=AES(256) Mac=SHA384 ECDHE-ECDSA-AES128-SHA256 TLSv1.2 Kx=ECDH Au=ECDSA Enc=AES(128) Mac=SHA256 ECDHE-RSA-AES128-SHA256 TLSv1.2 Kx=ECDH Au=RSA Enc=AES(128) Mac=SHA256 DHE-RSA-AES256-GCM-SHA384 TLSv1.2 Kx=DH Au=RSA Enc=AESGCM(256) Mac=AEAD DHE-RSA-AES128-GCM-SHA256 TLSv1.2 Kx=DH Au=RSA Enc=AESGCM(128) Mac=AEAD DHE-RSA-AES256-SHA256 TLSv1.2 Kx=DH Au=RSA Enc=AES(256) Mac=SHA256 DHE-RSA-AES128-SHA256 TLSv1.2 Kx=DH Au=RSA Enc=AES(128) Mac=SHA256 ``` Signed-off-by: Christian Heimes <christian@python.org>
2021-05-01 15:53:10 -03:00
* *STRING*: use a custom string
See the :mod:`ssl` module.
.. versionadded:: 3.7
bpo-43998: Default to TLS 1.2 and increase cipher suite security (GH-25778) The ssl module now has more secure default settings. Ciphers without forward secrecy or SHA-1 MAC are disabled by default. Security level 2 prohibits weak RSA, DH, and ECC keys with less than 112 bits of security. :class:`~ssl.SSLContext` defaults to minimum protocol version TLS 1.2. Settings are based on Hynek Schlawack's research. ``` $ openssl version OpenSSL 1.1.1k FIPS 25 Mar 2021 $ openssl ciphers -v '@SECLEVEL=2:ECDH+AESGCM:ECDH+CHACHA20:ECDH+AES:DHE+AES:!aNULL:!eNULL:!aDSS:!SHA1:!AESCCM' TLS_AES_256_GCM_SHA384 TLSv1.3 Kx=any Au=any Enc=AESGCM(256) Mac=AEAD TLS_CHACHA20_POLY1305_SHA256 TLSv1.3 Kx=any Au=any Enc=CHACHA20/POLY1305(256) Mac=AEAD TLS_AES_128_GCM_SHA256 TLSv1.3 Kx=any Au=any Enc=AESGCM(128) Mac=AEAD TLS_AES_128_CCM_SHA256 TLSv1.3 Kx=any Au=any Enc=AESCCM(128) Mac=AEAD ECDHE-ECDSA-AES256-GCM-SHA384 TLSv1.2 Kx=ECDH Au=ECDSA Enc=AESGCM(256) Mac=AEAD ECDHE-RSA-AES256-GCM-SHA384 TLSv1.2 Kx=ECDH Au=RSA Enc=AESGCM(256) Mac=AEAD ECDHE-ECDSA-AES128-GCM-SHA256 TLSv1.2 Kx=ECDH Au=ECDSA Enc=AESGCM(128) Mac=AEAD ECDHE-RSA-AES128-GCM-SHA256 TLSv1.2 Kx=ECDH Au=RSA Enc=AESGCM(128) Mac=AEAD ECDHE-ECDSA-CHACHA20-POLY1305 TLSv1.2 Kx=ECDH Au=ECDSA Enc=CHACHA20/POLY1305(256) Mac=AEAD ECDHE-RSA-CHACHA20-POLY1305 TLSv1.2 Kx=ECDH Au=RSA Enc=CHACHA20/POLY1305(256) Mac=AEAD ECDHE-ECDSA-AES256-SHA384 TLSv1.2 Kx=ECDH Au=ECDSA Enc=AES(256) Mac=SHA384 ECDHE-RSA-AES256-SHA384 TLSv1.2 Kx=ECDH Au=RSA Enc=AES(256) Mac=SHA384 ECDHE-ECDSA-AES128-SHA256 TLSv1.2 Kx=ECDH Au=ECDSA Enc=AES(128) Mac=SHA256 ECDHE-RSA-AES128-SHA256 TLSv1.2 Kx=ECDH Au=RSA Enc=AES(128) Mac=SHA256 DHE-RSA-AES256-GCM-SHA384 TLSv1.2 Kx=DH Au=RSA Enc=AESGCM(256) Mac=AEAD DHE-RSA-AES128-GCM-SHA256 TLSv1.2 Kx=DH Au=RSA Enc=AESGCM(128) Mac=AEAD DHE-RSA-AES256-SHA256 TLSv1.2 Kx=DH Au=RSA Enc=AES(256) Mac=SHA256 DHE-RSA-AES128-SHA256 TLSv1.2 Kx=DH Au=RSA Enc=AES(128) Mac=SHA256 ``` Signed-off-by: Christian Heimes <christian@python.org>
2021-05-01 15:53:10 -03:00
.. versionchanged:: 3.10
The settings ``python`` and *STRING* also set TLS 1.2 as minimum
protocol version.
macOS Options
-------------
See ``Mac/README.rst``.
.. cmdoption:: --enable-universalsdk
.. cmdoption:: --enable-universalsdk=SDKDIR
Create a universal binary build. *SDKDIR* specifies which macOS SDK should
be used to perform the build (default is no).
.. cmdoption:: --enable-framework
.. cmdoption:: --enable-framework=INSTALLDIR
Create a Python.framework rather than a traditional Unix install. Optional
*INSTALLDIR* specifies the installation path (default is no).
.. cmdoption:: --with-universal-archs=ARCH
Specify the kind of universal binary that should be created. This option is
only valid when :option:`--enable-universalsdk` is set.
Options:
* ``universal2``;
* ``32-bit``;
* ``64-bit``;
* ``3-way``;
* ``intel``;
* ``intel-32``;
* ``intel-64``;
* ``all``.
.. cmdoption:: --with-framework-name=FRAMEWORK
Specify the name for the python framework on macOS only valid when
:option:`--enable-framework` is set (default: ``Python``).
Cross Compiling Options
-----------------------
Cross compiling, also known as cross building, can be used to build Python
for another CPU architecture or platform. Cross compiling requires a Python
interpreter for the build platform. The version of the build Python must match
the version of the cross compiled host Python.
.. cmdoption:: --build=BUILD
configure for building on BUILD, usually guessed by :program:`config.guess`.
.. cmdoption:: --host=HOST
cross-compile to build programs to run on HOST (target platform)
.. cmdoption:: --with-build-python=path/to/python
path to build ``python`` binary for cross compiling
.. versionadded:: 3.11
.. cmdoption:: CONFIG_SITE=file
An environment variable that points to a file with configure overrides.
Example *config.site* file::
# config.site-aarch64
ac_cv_buggy_getaddrinfo=no
ac_cv_file__dev_ptmx=yes
ac_cv_file__dev_ptc=no
Cross compiling example::
CONFIG_SITE=config.site-aarch64 ../configure \
--build=x86_64-pc-linux-gnu \
--host=aarch64-unknown-linux-gnu \
--with-build-python=../x86_64/python
Python Build System
===================
Main files of the build system
------------------------------
* :file:`configure.ac` => :file:`configure`;
* :file:`Makefile.pre.in` => :file:`Makefile` (created by :file:`configure`);
* :file:`pyconfig.h` (created by :file:`configure`);
* :file:`Modules/Setup`: C extensions built by the Makefile using
:file:`Module/makesetup` shell script;
* :file:`setup.py`: C extensions built using the ``setuptools`` package.
Main build steps
----------------
* C files (``.c``) are built as object files (``.o``).
* A static ``libpython`` library (``.a``) is created from objects files.
* ``python.o`` and the static ``libpython`` library are linked into the
final ``python`` program.
* C extensions are built by the Makefile (see :file:`Modules/Setup`)
and ``python setup.py build``.
Main Makefile targets
---------------------
* ``make``: Build Python with the standard library.
* ``make platform:``: build the ``python`` program, but don't build the
standard library extension modules.
* ``make profile-opt``: build Python using Profile Guided Optimization (PGO).
You can use the configure :option:`--enable-optimizations` option to make
this the default target of the ``make`` command (``make all`` or just
``make``).
* ``make buildbottest``: Build Python and run the Python test suite, the same
way than buildbots test Python. Set ``TESTTIMEOUT`` variable (in seconds)
to change the test timeout (1200 by default: 20 minutes).
* ``make install``: Build and install Python.
* ``make regen-all``: Regenerate (almost) all generated files;
``make regen-stdlib-module-names`` and ``autoconf`` must be run separately
for the remaining generated files.
* ``make clean``: Remove built files.
* ``make distclean``: Same than ``make clean``, but remove also files created
by the configure script.
C extensions
------------
Some C extensions are built as built-in modules, like the ``sys`` module.
They are built with the ``Py_BUILD_CORE_BUILTIN`` macro defined.
Built-in modules have no ``__file__`` attribute::
>>> import sys
>>> sys
<module 'sys' (built-in)>
>>> sys.__file__
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
AttributeError: module 'sys' has no attribute '__file__'
Other C extensions are built as dynamic libraries, like the ``_asyncio`` module.
They are built with the ``Py_BUILD_CORE_MODULE`` macro defined.
Example on Linux x86-64::
>>> import _asyncio
>>> _asyncio
<module '_asyncio' from '/usr/lib64/python3.9/lib-dynload/_asyncio.cpython-39-x86_64-linux-gnu.so'>
>>> _asyncio.__file__
'/usr/lib64/python3.9/lib-dynload/_asyncio.cpython-39-x86_64-linux-gnu.so'
:file:`Modules/Setup` is used to generate Makefile targets to build C extensions.
At the beginning of the files, C extensions are built as built-in modules.
Extensions defined after the ``*shared*`` marker are built as dynamic libraries.
The :file:`setup.py` script only builds C extensions as shared libraries using
the :mod:`distutils` module.
The :c:macro:`PyAPI_FUNC()`, :c:macro:`PyAPI_API()` and
:c:macro:`PyMODINIT_FUNC()` macros of :file:`Include/pyport.h` are defined
differently depending if the ``Py_BUILD_CORE_MODULE`` macro is defined:
* Use ``Py_EXPORTED_SYMBOL`` if the ``Py_BUILD_CORE_MODULE`` is defined
* Use ``Py_IMPORTED_SYMBOL`` otherwise.
If the ``Py_BUILD_CORE_BUILTIN`` macro is used by mistake on a C extension
built as a shared library, its ``PyInit_xxx()`` function is not exported,
causing an :exc:`ImportError` on import.
Compiler and linker flags
=========================
Options set by the ``./configure`` script and environment variables and used by
``Makefile``.
Preprocessor flags
------------------
.. envvar:: CONFIGURE_CPPFLAGS
Value of :envvar:`CPPFLAGS` variable passed to the ``./configure`` script.
.. versionadded:: 3.6
.. envvar:: CPPFLAGS
(Objective) C/C++ preprocessor flags, e.g. ``-I<include dir>`` if you have
headers in a nonstandard directory ``<include dir>``.
Both :envvar:`CPPFLAGS` and :envvar:`LDFLAGS` need to contain the shell's
value for setup.py to be able to build extension modules using the
directories specified in the environment variables.
.. envvar:: BASECPPFLAGS
.. versionadded:: 3.4
.. envvar:: PY_CPPFLAGS
Extra preprocessor flags added for building the interpreter object files.
Default: ``$(BASECPPFLAGS) -I. -I$(srcdir)/Include $(CONFIGURE_CPPFLAGS) $(CPPFLAGS)``.
.. versionadded:: 3.2
Compiler flags
--------------
.. envvar:: CC
C compiler command.
Example: ``gcc -pthread``.
.. envvar:: CXX
C++ compiler command.
Example: ``g++ -pthread``.
.. envvar:: CFLAGS
C compiler flags.
.. envvar:: CFLAGS_NODIST
:envvar:`CFLAGS_NODIST` is used for building the interpreter and stdlib C
extensions. Use it when a compiler flag should *not* be part of the
distutils :envvar:`CFLAGS` once Python is installed (:issue:`21121`).
In particular, :envvar:`CFLAGS` should not contain:
* the compiler flag ``-I`` (for setting the search path for include files).
The ``-I`` flags are processed from left to right, and any flags in
:envvar:`CFLAGS` would take precedence over user- and package-supplied ``-I``
flags.
* hardening flags such as ``-Werror`` because distributions cannot control
whether packages installed by users conform to such heightened
standards.
.. versionadded:: 3.5
.. envvar:: COMPILEALL_OPTS
Options passed to the :mod:`compileall` command line when building PYC files
in ``make install``. Default: ``-j0``.
.. versionadded:: 3.12
.. envvar:: EXTRA_CFLAGS
Extra C compiler flags.
.. envvar:: CONFIGURE_CFLAGS
Value of :envvar:`CFLAGS` variable passed to the ``./configure``
script.
.. versionadded:: 3.2
.. envvar:: CONFIGURE_CFLAGS_NODIST
Value of :envvar:`CFLAGS_NODIST` variable passed to the ``./configure``
script.
.. versionadded:: 3.5
.. envvar:: BASECFLAGS
Base compiler flags.
.. envvar:: OPT
Optimization flags.
.. envvar:: CFLAGS_ALIASING
Strict or non-strict aliasing flags used to compile ``Python/dtoa.c``.
.. versionadded:: 3.7
.. envvar:: CCSHARED
Compiler flags used to build a shared library.
For example, ``-fPIC`` is used on Linux and on BSD.
.. envvar:: CFLAGSFORSHARED
Extra C flags added for building the interpreter object files.
Default: ``$(CCSHARED)`` when :option:`--enable-shared` is used, or an empty
string otherwise.
.. envvar:: PY_CFLAGS
Default: ``$(BASECFLAGS) $(OPT) $(CONFIGURE_CFLAGS) $(CFLAGS) $(EXTRA_CFLAGS)``.
.. envvar:: PY_CFLAGS_NODIST
Default: ``$(CONFIGURE_CFLAGS_NODIST) $(CFLAGS_NODIST) -I$(srcdir)/Include/internal``.
.. versionadded:: 3.5
.. envvar:: PY_STDMODULE_CFLAGS
C flags used for building the interpreter object files.
Default: ``$(PY_CFLAGS) $(PY_CFLAGS_NODIST) $(PY_CPPFLAGS) $(CFLAGSFORSHARED)``.
.. versionadded:: 3.7
.. envvar:: PY_CORE_CFLAGS
Default: ``$(PY_STDMODULE_CFLAGS) -DPy_BUILD_CORE``.
.. versionadded:: 3.2
.. envvar:: PY_BUILTIN_MODULE_CFLAGS
Compiler flags to build a standard library extension module as a built-in
module, like the :mod:`posix` module.
Default: ``$(PY_STDMODULE_CFLAGS) -DPy_BUILD_CORE_BUILTIN``.
.. versionadded:: 3.8
.. envvar:: PURIFY
Purify command. Purify is a memory debugger program.
Default: empty string (not used).
Linker flags
------------
.. envvar:: LINKCC
Linker command used to build programs like ``python`` and ``_testembed``.
Default: ``$(PURIFY) $(CC)``.
.. envvar:: CONFIGURE_LDFLAGS
Value of :envvar:`LDFLAGS` variable passed to the ``./configure`` script.
Avoid assigning :envvar:`CFLAGS`, :envvar:`LDFLAGS`, etc. so users can use
them on the command line to append to these values without stomping the
pre-set values.
.. versionadded:: 3.2
.. envvar:: LDFLAGS_NODIST
:envvar:`LDFLAGS_NODIST` is used in the same manner as
:envvar:`CFLAGS_NODIST`. Use it when a linker flag should *not* be part of
the distutils :envvar:`LDFLAGS` once Python is installed (:issue:`35257`).
In particular, :envvar:`LDFLAGS` should not contain:
* the compiler flag ``-L`` (for setting the search path for libraries).
The ``-L`` flags are processed from left to right, and any flags in
:envvar:`LDFLAGS` would take precedence over user- and package-supplied ``-L``
flags.
.. envvar:: CONFIGURE_LDFLAGS_NODIST
Value of :envvar:`LDFLAGS_NODIST` variable passed to the ``./configure``
script.
.. versionadded:: 3.8
.. envvar:: LDFLAGS
Linker flags, e.g. ``-L<lib dir>`` if you have libraries in a nonstandard
directory ``<lib dir>``.
Both :envvar:`CPPFLAGS` and :envvar:`LDFLAGS` need to contain the shell's
value for setup.py to be able to build extension modules using the
directories specified in the environment variables.
.. envvar:: LIBS
Linker flags to pass libraries to the linker when linking the Python
executable.
Example: ``-lrt``.
.. envvar:: LDSHARED
Command to build a shared library.
Default: ``@LDSHARED@ $(PY_LDFLAGS)``.
.. envvar:: BLDSHARED
Command to build ``libpython`` shared library.
Default: ``@BLDSHARED@ $(PY_CORE_LDFLAGS)``.
.. envvar:: PY_LDFLAGS
Default: ``$(CONFIGURE_LDFLAGS) $(LDFLAGS)``.
.. envvar:: PY_LDFLAGS_NODIST
Default: ``$(CONFIGURE_LDFLAGS_NODIST) $(LDFLAGS_NODIST)``.
.. versionadded:: 3.8
.. envvar:: PY_CORE_LDFLAGS
Linker flags used for building the interpreter object files.
.. versionadded:: 3.8