Skip to content

Latest commit

 

History

History
80 lines (57 loc) · 2.91 KB

README.rst

File metadata and controls

80 lines (57 loc) · 2.91 KB

pywayland

Build Status Build Coverage Documentation Status

PyWayland provides a wrapper to the libwayland library using the CFFI library to provide access to the Wayland library calls and written in pure Python.

Below is outlined some of the basics of PyWayland and how to get up and running. For more help, see the full documentation.

Current Release

PyWayland is still in a developmental state. An current version is available on the cheese shop. Current development versions can be obtained from the git repository, feedback, as well as any bug reports or fixes are highly appreciated.

Dependencies

Installing PyWayland requires the Wayland library and the headers to be installed.

PyWayland requires the cffi package to be installed. PyWayland runs and is tested against Python 3.6+, including sufficient versions of PyPy3 (see Running Tests).

See the installation guide for more information on installing required dependencies

Installing

Installation can be done through pip to pull the most recently tagged release. To see instructions on running from source, see the relevant documentation on installing from source.

Building Wayland protocols

In order to run from source, you will need to generate the interfaces to the Wayland protocol objects as defined in the wayland.xml file. By default, this file will be located in /usr/share/wayland/wayland.xml. In this case, the protocol files can be generated by the scanner module:

$ python -m pywayland.scanner

See the help for this module to use non-default locations for the input and output of the scanner.

The scanner is installed as a script pywayland-scanner when PyWayland is installed. See pywayland-scanner -h for more information.

Running Tests

PyWayland implements a (currently limited) test-suite in ./tests. The tests can be run through pytest. Be sure you build the protocol files (see Building Wayland protocols) before running the tests.