Here you can find set of scripts, which take documentation from various OpenLMI repositories and generate an overall document with all the docs.
The docs are stored in this git repository, so readthedocs.org can pull them easily.
During the processing, various cmake and make targets are called. Therefore
it's necessary to have all installed packages, that OpenLMI depends on.
Following list was gathered using yum-builddeps *.spec on Fedora 21:
- cmake
- plantuml
- graphviz
- tar
- dia
- cim-schema
- dbus-devel
- glib2-devel
- json-glib-devel
- konkretcmpi-devel
- konkretcmpi-python
- libselinux-devel
- libsemanage-devel
- libsss_simpleifp-devel
- systemd-devel
- libuser-devel
- libxml2-devel
- lm_sensors-devel
- pciutils-devel
- python-setuptools
- python-sphinx
- python-sphinx_rtd_theme
- python-devel
- sblim-cmpi-devel
- systemd-devel
- openlmi-providers (for class2rst and mof files in /usr/share/mof)
- python-IPy
- openwsman-python
Run setup.sh. It clones necessary repositories into _ext/.
Run build.sh. It refreshes the repositories in _ext/ directory
(git pull), reorganizes whatever is necesary (using _build/) directory
and produces (huge) tree of .rst files in doc/.
All changes in doc/, incl. new files, must be checked into the repository,
so readthedocs.org can find it there.
-
doc/python/is populated with python libraries we want to document (lmishell, openlmi-scripts, ...) and their dependencies (pywbem; it is not available on readthedocs.org builders). -
Links to figures in original .rst files are translated to appropriate parent directories. In other words, following is perfectly OK and it will be updated to point to the right file::
.. figure:: pic/my.svg :target: _static/my.svg