|author||Adam Jackson <firstname.lastname@example.org>||Wed Oct 21 10:30:03 2020|
|committer||GitHub <email@example.com>||Wed Oct 21 10:30:03 2020|
api: Fix genheaders.py to emit valid comments (#114) When we switched to using SPDX identifiers instead of explicit Apache 2.0 text, the resulting generated headers ended up with a nested /* inside the license comment.
The EGL-Registry repository contains the EGL API and Extension Registry, including specifications, reference pages and reference cards, and the enumerant registry. It is also used as a backing store for the web view of the registry at https://www.khronos.org/registry/egl/ ; commits to the master branch of this repository will be reflected there.
In the past, the EGL registry was maintained in a public Subversion repository. The history in that repository has not been imported to github, but it is still available at https://cvs.khronos.org/svn/repos/registry/trunk/public/egl/ .
Interesting files in this repository include:
EGL enumerants are documented in api/egl.xml . New ranges can be allocated by proposing a pull request to master modifying this file, following the existing examples. Allocate ranges starting at the lowest free values available (search for “Reservable for future use”). Ranges are not officially allocated until your pull request is accepted into master. At that point you can use values from your assigned range for API extensions.
Extension specification documents can be added by proposing a pull request to master, adding the specification .txt file and related changes under extensions/<vendor>/filename.txt. Your pull request must also:
<extension>examples, search for “Next free extension number”, and use the lowest available extension number).
# Validate XML changes make validate # Verify headers build and are legal C make clobber make make tests
Sometimes extension text files contain inappropriate UTF-8 characters. They should be restricted to the ASCII subset of UTF-8 at present. They can be removed using the iconv Linux command-line tool via
iconv -c -f utf-8 -t ascii filename.txt
(see internal Bugzilla issue 16141 for more).
We may transition to an asciidoc-based extension specification format at some point.
This section is not complete (see https://github.com/KhronosGroup/EGL-Registry/issues/92).
To validate the XML and build the headers you will need at least GNU make, ‘jing’ for the ‘make validate’ step (https://relaxng.org/jclark/jing.html), and Python 3.5 and the lxml.etree Python library (https://pypi.org/project/lxml/) for the ‘make’ step. The ‘make tests’ step requires whatever the C and C++ compilers configured for GNU make are, usually gcc and g++.
All of these components are available prepackaged for major Linux distributions and for the Windows 10 Debian WSL.