Code Monkey home page Code Monkey logo

manylinux's Introduction

manylinux

Email: [email protected]

Archives: https://mail.python.org/mailman/listinfo/wheel-builders

Older archives: https://groups.google.com/forum/#!forum/manylinux-discuss

The goal of the manylinux project is to provide a convenient way to distribute binary Python extensions as wheels on Linux. This effort has produced PEP 513 (manylinux1), PEP 571 (manylinux2010), PEP 599 (manylinux2014) and PEP 600 (manylinux_x_y).

PEP 513 defined manylinux1_x86_64 and manylinux1_i686 platform tags and the wheels were built on Centos5. Centos5 reached End of Life (EOL) on March 31st, 2017.

PEP 571 defined manylinux2010_x86_64 and manylinux2010_i686 platform tags and the wheels were built on Centos6. Centos6 reached End of Life (EOL) on November 30th, 2020.

PEP 599 defines the following platform tags:

  • manylinux2014_x86_64
  • manylinux2014_i686
  • manylinux2014_aarch64
  • manylinux2014_armv7l
  • manylinux2014_ppc64
  • manylinux2014_ppc64le
  • manylinux2014_s390x

Wheels are built on CentOS 7 which will reach End of Life (EOL) on June 30th, 2024.

PEP 600 has been designed to be "future-proof" and does not enforce specific symbols and a specific distro to build. It only states that a wheel tagged manylinux_x_y shall work on any distro based on glibc>=x.y. The manylinux project supports manylinux_2_24 images for x86_64, i686, aarch64, ppc64le and s390x.

Wheel packages compliant with those tags can be uploaded to PyPI (for instance with twine) and can be installed with pip:

manylinux tag Client-side pip version required CPython (sources) version embedding a compatible pip Distribution default pip compatibility
manylinux_x_y pip >= 20.3 3.10.0+ Fedora 34+, Mageia 8+, Photon OS 3.0 with updates
manylinux2014 pip >= 19.3 3.7.8+, 3.8.4+, 3.9.0+ CentOS 7 rh-python38, CentOS 8 python38, Fedora 32+, Mageia 8+, Photon OS 4.0+ (3.0+ with updates), Ubuntu 20.04+
manylinux2010 pip >= 19.0 3.7.3+, 3.8.0+ ALT Linux 9+, CentOS 7 rh-python38, CentOS 8 python38, Fedora 30+, Mageia 7+, Photon OS 4.0+ (3.0+ with updates), Ubuntu 20.04+
manylinux1 pip >= 8.1.0 3.5.2+, 3.6.0+ ALT Linux 8+, Amazon Linux 1+, CentOS 7+, Debian 9+, Fedora 25+, openSUSE 15.2+, Mageia 7+, Photon OS 1.0+, Ubuntu 16.04+

The various manylinux tags allow projects to distribute wheels that are automatically installed (and work!) on the vast majority of desktop and server Linux distributions.

This repository hosts several manylinux-related things:

Docker images

Building manylinux-compatible wheels is not trivial; as a general rule, binaries built on one Linux distro will only work on other Linux distros that are the same age or newer. Therefore, if we want to make binaries that run on most Linux distros, we have to use an old enough distro.

Rather than forcing you to install an old distro yourself, install Python, etc., we provide Docker images where we've done the work for you. The images are uploaded to quay.io and are tagged for repeatable builds.

manylinux_2_24 (Debian 9 based)

  • x86_64 image: quay.io/pypa/manylinux_2_24_x86_64
  • i686 image: quay.io/pypa/manylinux_2_24_i686
  • aarch64 image: quay.io/pypa/manylinux_2_24_aarch64
  • ppc64le image: quay.io/pypa/manylinux_2_24_ppc64le
  • s390x image: quay.io/pypa/manylinux_2_24_s390x

manylinux2014 (CentOS 7 based)

  • x86_64 image: quay.io/pypa/manylinux2014_x86_64
  • i686 image: quay.io/pypa/manylinux2014_i686
  • aarch64 image: quay.io/pypa/manylinux2014_aarch64
  • ppc64le image: quay.io/pypa/manylinux2014_ppc64le
  • s390x image: quay.io/pypa/manylinux2014_s390x

manylinux2010 (CentOS 6 based - EOL)

  • x86-64 image: quay.io/pypa/manylinux2010_x86_64
  • i686 image: quay.io/pypa/manylinux2010_i686

manylinux1 (CentOS 5 based - EOL)

Code and details regarding manylinux1 can be found in the manylinux1 branch.

Support for manylinux1 will end on January 1st, 2022.

  • x86-64 image: quay.io/pypa/manylinux1_x86_64
  • i686 image: quay.io/pypa/manylinux1_i686

All images are rebuilt using GitHub Actions / Travis-CI on every commit to this repository; see the docker/ directory for source code.

Image content

All images currently contain:

  • CPython 3.5, 3.6, 3.7, 3.8 and 3.9, installed in /opt/python/<python tag>-<abi tag>. The directories are named after the PEP 425 tags for each environment -- e.g. /opt/python/cp37-cp37m contains a CPython 3.7 build, and can be used to produce wheels named like <pkg>-<version>-cp37-cp37m-<arch>.whl.
  • Development packages for all the libraries that PEP 571/599 list. One should not assume the presence of any other development package.
  • The auditwheel tool

Note that less common or virtually unheard of flag combinations (such as --with-pydebug (d) and --without-pymalloc (absence of m)) are not provided.

Note that starting with CPython 3.8, default sys.abiflags became an empty string: the m flag for pymalloc became useless (builds with and without pymalloc are ABI compatible) and so has been removed. (e.g. /opt/python/cp38-cp38)

Building Docker images

To build the Docker images, please run the following command from the current (root) directory:

$ PLATFORM=$(uname -m) POLICY=manylinux2014 COMMIT_SHA=latest ./build.sh

Please not that the Docker build is using buildx.

Example

An example project which builds x86_64 wheels for each Python interpreter version can be found here: https://github.com/pypa/python-manylinux-demo. The repository also contains demo to build i686 and x86_64 wheels with manylinux1 tags.

This demonstrates how to use these docker images in conjunction with auditwheel to build manylinux-compatible wheels using the free travis ci continuous integration service.

(NB: for the i686 images running on a x86_64 host machine, it's necessary to run everything under the command line program linux32, which changes reported architecture in new program environment. See this example invocation)

The PEP itself

The official version of PEP 513 is stored in the PEP repository, but we also have our own copy here. This is where the PEP was originally written, so if for some reason you really want to see the full history of edits it went through, then this is the place to look.

The proposal to upgrade manylinux1 to manylinux2010 after Centos5 reached EOL was discussed in PEP 571.

The proposal to upgrade manylinux2010 to manylinux2014 was discussed in PEP 599.

The proposal for a "future-proof" manylinux_x_y definition was discussed in PEP 600.

This repo also has some analysis code that was used when putting together the original proposal in the policy-info/ directory.

If you want to read the full discussion that led to the original policy, then lots of that is here: https://groups.google.com/forum/#!forum/manylinux-discuss

The distutils-sig archives for January 2016 also contain several threads.

Code of Conduct

Everyone interacting in the manylinux project's codebases, issue trackers, chat rooms, and mailing lists is expected to follow the PSF Code of Conduct.

manylinux's People

Contributors

mayeut avatar njsmith avatar dependabot-preview[bot] avatar trishankatdatadog avatar pyup-bot avatar rmcgibbo avatar ogrisel avatar auvipy avatar manselmi avatar lkollar avatar matthew-brett avatar reaperhulk avatar jcfr avatar michael-anselmi avatar mattip avatar natefoo avatar ehashman avatar lelit avatar radarhere avatar jschueller avatar rdb avatar ijl avatar takluyver avatar segevfiner avatar joerick avatar geoffreyblake avatar armudgal avatar mephi42 avatar szilveszter avatar stefanv avatar

Watchers

James Cloos avatar  avatar

Recommend Projects

  • React photo React

    A declarative, efficient, and flexible JavaScript library for building user interfaces.

  • Vue.js photo Vue.js

    ๐Ÿ–– Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.

  • Typescript photo Typescript

    TypeScript is a superset of JavaScript that compiles to clean JavaScript output.

  • TensorFlow photo TensorFlow

    An Open Source Machine Learning Framework for Everyone

  • Django photo Django

    The Web framework for perfectionists with deadlines.

  • D3 photo D3

    Bring data to life with SVG, Canvas and HTML. ๐Ÿ“Š๐Ÿ“ˆ๐ŸŽ‰

Recommend Topics

  • javascript

    JavaScript (JS) is a lightweight interpreted programming language with first-class functions.

  • web

    Some thing interesting about web. New door for the world.

  • server

    A server is a program made to process requests and deliver data to clients.

  • Machine learning

    Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.

  • Game

    Some thing interesting about game, make everyone happy.

Recommend Org

  • Facebook photo Facebook

    We are working to build community through open source technology. NB: members must have two-factor auth.

  • Microsoft photo Microsoft

    Open source projects and samples from Microsoft.

  • Google photo Google

    Google โค๏ธ Open Source for everyone.

  • D3 photo D3

    Data-Driven Documents codes.