8000 DOC: re-arrange installation docs · matplotlib/matplotlib@41641bd · GitHub
[go: up one dir, main page]

Skip to content

Commit 41641bd

Browse files
committed
DOC: re-arrange installation docs
- Note issues with subprocess32 on OSX and linux - Move note about the MVC redistributable libraries to the top section closes #8433 closes #8361
1 parent 9b1af52 commit 41641bd

File tree

1 file changed

+45
-46
lines changed

1 file changed

+45
-46
lines changed

INSTALL.rst

Lines changed: 45 additions & 46 deletions
Original file line numberDiff line numberDiff line change
@@ -19,8 +19,8 @@ Installing
1919
Installing an official release
2020
==============================
2121

22-
Matplotlib and most of its dependencies are all available as wheel packages for
23-
macOS, Windows and Linux distributions::
22+
Matplotlib and most of its dependencies are all available as wheel
23+
packages for macOS, Windows and Linux distributions::
2424

2525
python -m pip install -U pip
2626
python -m pip install -U matplotlib
@@ -33,46 +33,11 @@ macOS, Windows and Linux distributions::
3333
a larger selection of file formats, you may need to install :ref:`additional
3434
dependencies <install_requirements>`.
3535

36-
Third-party distributions of Matplotlib
37-
=======================================
38-
39-
Scientific Python distributions: Conda, Canopy...
40-
-------------------------------------------------
41-
42-
The first option is to use one of the pre-packaged Python distributions that
43-
already provide Matplotlib built-in. Both `Anaconda
44-
<https://www.continuum.io/downloads/>`_ and `Canopy
45-
<https://www.enthought.com/products/canopy/>`_ are both excellent choices that
46-
"just work" out of the box for Windows, macOS and common Linux platforms. Both
47-
of these distributions include Matplotlib and *lots* of other useful tools.
48-
49-
Linux : using your package manager
50-
----------------------------------
51-
52-
If you are on Linux, you might prefer to use your package manager. Matplotlib
53-
is packaged for almost every major Linux distribution.
54-
55-
* Debian / Ubuntu: ``sudo apt-get install python-matplotlib``
56-
* Fedora: ``sudo dnf install python-matplotlib``
57-
* Red Hat: ``sudo yum install python-matplotlib``
58-
* Arch: ``sudo pacman -S python-matplotlib``
59-
6036
.. _installing_windows:
6137

6238
Windows
6339
-------
6440

65-
We strongly recommend using `SciPy-stack compatible Python distributions
66-
<https://www.scipy.org/install.html>`_ such as WinPython, Python(x,y),
67-
Enthought Canopy, or Continuum Anaconda, which have Matplotlib and its
68-
dependencies, plus other useful packages, preinstalled.
69-
70-
For `standard Python <https://www.python.org/downloads/>`_ installations,
71-
install Matplotlib using pip_::
72-
73-
python -m pip install -U pip setuptools
74-
python -m pip install matplotlib
75-
7641
In case Python 2.7 or 3.4 are not installed for all users,
7742
the Microsoft Visual C++ 2008
7843
(`64 bit <https://www.microsoft.com/en-us/download/details.aspx?id=15336>`__
@@ -84,8 +49,27 @@ or
8449
`32 bit <https://www.microsoft.com/en-us/download/details.aspx?id=5555>`__
8550
for Python 3.4) redistributable packages need to be installed.
8651

52+
OSX
53+
---
54+
55+
If you are using python 2.7 on a Mac you may need to do::
56+
57+
xcode-select --install
58+
59+
so that *subprocess32*, a dependency, may be complied.
60+
61+
Linux
62+
-----
8763

88-
The Windows wheels (:file:`*.whl`) on the `PyPI download page
64+
On extremely old versions of linux and python 2.7 you may need to
65+
install the master version of *subprocess32* (`see comments
66+
<https://github.com/google/python-subprocess32/issues/12#issuecomment-304724113>`__)
67+
68+
69+
Test Data
70+
---------
71+
72+
The wheels (:file:`*.whl`) on the `PyPI download page
8973
<https://pypi.python.org/pypi/matplotlib/>`_ do not contain test data
9074
or example code.
9175
If you want to try the many demos that come in the Matplotlib source
@@ -101,17 +85,32 @@ To run the test suite:
10185
<https://inkscape.org/>`_;
10286
* run ``py.test path\to\tests\directory``.
10387

104-
.. note::
10588

106-
The following backends work out of the box: Agg, TkAgg, ps, pdf and svg.
107-
TkAgg is probably the best backend for interactive use from the standard
108-
Python shell or from IPython and is enabled as default.
89+
Third-party distributions of Matplotlib
90+
=======================================
10991

110-
GTK3 is not supported on Windows.
92+
Scientific Python distributions: Conda, Canopy...
93+
-------------------------------------------------
94+
95+
The first option is to use one of the pre-packaged Python distributions that
96+
already provide Matplotlib built-in. Both `Anaconda
97+
<https://www.continuum.io/downloads/>`_ and `Canopy
98+
<https://www.enthought.com/products/canopy/>`_ are both excellent choices that
99+
"just work" out of the box for Windows, macOS and common Linux platforms. Both
100+
of these distributions include Matplotlib and *lots* of other useful tools.
101+
102+
103+
Linux : using your package manager
104+
----------------------------------
105+
106+
If you are on Linux, you might prefer to use your package manager. Matplotlib
107+
is packaged for almost every major Linux distribution.
108+
109+
* Debian / Ubuntu: ``sudo apt-get install python3-matplotlib``
110+
* Fedora: ``sudo dnf install python3-matplotlib``
111+
* Red Hat: ``sudo yum install python3-matplotlib``
112+
* Arch: ``sudo pacman -S python-matplotlib``
111113

112-
For support for other backends, LaTeX rendering, animation input/output and
113-
a larger selection of file formats, you may need to install :ref:`additional
114-
dependencies <install_requirements>`.
115114

116115

117116
.. _install_from_source:

0 commit comments

Comments
 (0)
0