Mirror of our build engine
You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
Martin A. Brown 6a00bd0b1e bump version to 0.7.15 11 months ago
contrib bump version to 0.7.15 11 months ago
debian bump version to 0.7.15 11 months ago
docs add entry for --version to manpage 7 years ago
etc updating the stock and sample, commented config file 8 years ago
extras adding the FO generation XSL 8 years ago
tests add a test for the new --version CLI 7 years ago
tldp bump version to 0.7.15 11 months ago
.gitignore no need to track the .pyc files 8 years ago
.travis.yml run the new long_inventory.py test, too 8 years ago
ChangeLog adjust testing for root-run tests (changelog) 7 years ago
LICENSE adding 2016 LDP copyright 8 years ago
MANIFEST.in include in MANIFEST.in and spell "docs" correctly 8 years ago
NOTES.rst some notes for self 8 years ago
README.rst adjusting to reflect changes to packaging and Python versions 8 years ago
TODO add a TODO item for improving CLI error reporting 8 years ago
requirements.txt adjust requirements 8 years ago
setup.py move VERSION to a single location 8 years ago
tox.ini add py35 to tox list 7 years ago


tldp - tools for publishing from TLDP sources

.. image:: https://api.travis-ci.org/martin-a-brown/python-tldp.svg
:target: https://github.com/tLDP/python-tldp

.. image:: http://img.shields.io/badge/license-MIT-brightgreen.svg
:target: http://opensource.org/licenses/MIT
:alt: MIT license

This package was written for the Linux Documentation Project (TLDP) to help
with management and publication automation of source documents. The primary
interface provided is a command-line tool caalled `ldptool`. The canonical
location of this software is:


The `ldptool` executable can:

- crawl through any number of source collection directories
- crawl through a single output collection
- match the sources to the outputs (based on document stem name)
- describe supported source formats (`--formats`)
- describe the meaning of document status (`--statustypes`)
- describe the collection by type and status (`--summary`)
- list out individual document type and status (`--list`)
- build the expected (non-configurable) set of outputs (`--build`)
- build and publish the outputs (`--publish`)
- produce runnable shell script to STDOUT (`--script`)

The tools in this package process source documents in the `TLDP document
repository <https://github.com/tLDP/LDP>`_ and generate the following set of
outputs from each source document.

- .pdf, PDF
- .txt, text
- -single.html, a one-page HTML document
- .html, a multipage HTML document

(We may add other output formats; an epub format is under consideration.)

Supported input formats are:

- Asciidoc
- Linuxdoc
- Docbook SGML 3.x (though deprecated, please no new submissions)
- Docbook SGML 4.x
- Docbook XML 4.x
- Docbook XML 5.x (basic support, as of 2016-03-10)

Example usages
If your attempts to run the below commands don't work or generate errors, see
also `Minimal configuration`_.

Here are some example usages against a live checkout of the LDP source
repository and a local cache of the output tree:

To see what work needs to be done, `ldptool --list`::

$ ldptool --list
orphan <unknown> Bugzilla-Guide
new DocBook XML 4.x DocBook-Demystification-HOWTO
stale DocBook XML 4.x Linux-Dictionary
broken DocBook SGML 3.x/4.x PHP-Nuke-HOWTO
stale Linuxdoc User-Group-HOWTO

To see publication status of each document:::

$ ldptool --list all | head -n 3
published Linuxdoc 3-Button-Mouse
published Linuxdoc 3D-Modelling
published Linuxdoc 4mb-Laptops

To get more information about the newer or missing files in a specific

$ ldptool --verbose --list Linux-Dictionary
stale DocBook XML 4.x Linux-Dictionary
doctype <class 'tldp.doctypes.docbook4xml.Docbook4XML'>
output dir /home/mabrown/tmp/en/Linux-Dictionary
source file /home/mabrown/vcs/LDP/LDP/guide/docbook/Linux-Dictionary/Linux-Dictionary.xml
newer source /home/mabrown/vcs/LDP/LDP/guide/docbook/Linux-Dictionary/Contributors.xml
newer source /home/mabrown/vcs/LDP/LDP/guide/docbook/Linux-Dictionary/D.xml
newer source /home/mabrown/vcs/LDP/LDP/guide/docbook/Linux-Dictionary/J.xml
newer source /home/mabrown/vcs/LDP/LDP/guide/docbook/Linux-Dictionary/O.xml
newer source /home/mabrown/vcs/LDP/LDP/guide/docbook/Linux-Dictionary/S.xml

To see what the entire source collection looks like, use `ldptool --summary`:::

$ ldptool --summary
By Status Type
source 503 3-Button-Mouse, 3D-Modelling, 4mb-Laptops, and 500 more ...
output 503 3-Button-Mouse, 3D-Modelling, 4mb-Laptops, and 500 more ...
published 503 3-Button-Mouse, 3D-Modelling, 4mb-Laptops, and 500 more ...
stale 0
orphan 0
broken 1 HOWTO-INDEX
new 0

By Document Type
Linuxdoc 226 3-Button-Mouse, 3D-Modelling, and 224 more ...
Docbook4XML 130 8021X-HOWTO, abs-guide, and 128 more ...
Docbook5XML 1 Assembly-HOWTO
DocbookSGML 146 ACP-Modem, and 145 more ...

To build and publish a single document:::

$ ldptool --publish DocBook-Demystification-HOWTO
$ ldptool --publish ~/vcs/LDP/LDP/howto/docbook/Valgrind-HOWTO.xml

To build and publish anything that is new or updated work:::

$ ldptool --publish
$ ldptool --publish work

To (re-)build and publish everything, regardless of state:::

$ ldptool --publish all

To generate a specific output (into a --builddir):::

$ ldptool --build DocBook-Demystification-HOWTO

To generate all outputs into a --builddir (should exist):::

$ ldptool --builddir ~/tmp/scratch-directory/ --build all

To build new/updated work, but pass over a trouble-maker:::

$ ldptool --build --skip HOWTO-INDEX

To loudly generate all outputs, except a trouble-maker:::

$ ldptool --build all --loglevel debug --skip HOWTO-INDEX

To print out a shell script for building a specific document:::

$ ldptool --script TransparentProxy
$ ldptool --script ~/vcs/LDP/LDP/howto/docbook/Assembly-HOWTO.xml

The `ldptool` utility is largely written to be interactive or a supervised
batch process. It uses STDERR as its logstream and sets the default loglevel
at logging.ERROR. At this log level, in `--script`, `--build` and `--publish`
mode, it should report nothing to STDERR. To increase progress verbosity,
setting the loglevel to info (`--loglevel info`) may help with understanding
what work the tool is performing. If you need to collect diagnostic
information for troubleshooting or bug reports, `ldptool` supports `--loglevel

The `ldptool` comes with support for reading its settings from the
command-line, environment or a system and/or user-specified configuration
file. If you want to generate a sample configuration file to edit and use
later, you can run:::

ldptool --dump-cfg > my-ldptool.cfg
ldptool --configfile my-ldptool.cfg --list
LDPTOOL_CONFIGFILE=/path/to/ldptool.cfg ldptool --list

Source document identification
TLDP's source repository contains many separate directories containing
documents (e.g. LDP/howto/docbook, LDP/howto/linuxdoc). Each of these
directories may contain documents; to `ldptool` each of these is a

A source document (in a `--sourcedir`) can be a file or a directory. Here are
two examples. The Assembly-HOWTO.xml is an entire document stored as a single
file. The directory BRIDGE-STP-HOWTO exists and contains its main document, a
file named BRIDGE-STP-HOWTO.sgml. In the case of a source document that is a
directory, the stem name of the primary document must match the name of the


Each document for a single run of `ldptool` can be uniquely identified by its
stem name. In the above, the stems are `Assembly-HOWTO` and
`BRIDGE-STP-HOWTO`. It is an error to have two documents with the same stem
name and the second discovered document will be ignored.

There is a directory containing the output collection. Each directory is named
by the stem name of the source document and contains the output formats for
each source document. Here are the corresponding output directories for the
above two documents:::

... and more ...

... and more ...

Minimal configuration
The most important configuration parameters that `ldptool` takes are the set
of source directories (in which to find documents) and the output directory,
in which to create the resulting outputs. It will not be able to run unless
it has at least one --sourcedir and an existing --pubdir directory.

If you have an LDP checkout in your home directory, here's an example which
would process all of the Linuxdoc HOWTO docs:::

mkdir LDP-output-tree
ldptool --sourcedir $HOME/LDP/LDP/howto/linuxdoc --pubdir LDP-output-tree

If you would like to create a sample configuration file for use later (or for
copying into the system location, `/etc/ldptool/ldptool.ini`, you can generate
your own config file as follows:::

ldptool > sample-ldptool.cfg \
--sourcedir $HOME/LDP/LDP/faq/linuxdoc/ \
--sourcedir $HOME/LDP/LDP/guide/linuxdoc/ \
--sourcedir $HOME/LDP/LDP/howto/linuxdoc/ \
--sourcedir $HOME/LDP/LDP/howto/docbook/ \
--sourcedir $HOME/LDP/LDP/guide/docbook/ \
--sourcedir $HOME/LDP/LDP/ref/docbook/ \
--sourcedir $HOME/LDP/LDP/faq/docbook/ \
--pubdir $HOME/LDP-output/ \
--loglevel info \

Then, you can run the same configuration again with:::

ldptool --configfile sample-ldptool.cfg

The `ldptool` program tries to locate all of the tools it needs to process
documents. Each source format requires a certain set of tools, for example, to
process DocBook 4.x XML, `ldptool` needs the executables xmllint, xstlproc,
html2text, fop and dblatex. It also requires the XSL files for generating FO,
chunked HTML and single-page HTML. All of the items are configurable on the
command-line or in the configuration file, but here's a sample config file

xslchunk = /usr/share/xml/docbook/stylesheet/ldp/html/tldp-sections.xsl
xslsingle = /usr/share/xml/docbook/stylesheet/ldp/html/tldp-one-page.xsl
fop = /usr/bin/fop
dblatex = /usr/bin/dblatex
xsltproc = /usr/bin/xsltproc
html2text = /usr/bin/html2text
xslprint = /usr/share/xml/docbook/stylesheet/ldp/fo/tldp-print.xsl
xmllint = /usr/bin/xmllint

The above stanza was generated by running `ldptool --dump-cfg` on an Ubuntu
14.04 system which had all of the software dependencies installed. If your
distribution does not supply ldp-docbook-xsl, for example, you would need to
fetch those files, put them someplace in the filesystem and adjust your
configuration file or command-line invocations accordingly.

Software dependencies
There are a large number of packages listed here in the dependency set. This
is because the supporting software for processing Linuxdoc and the various
DocBook formats is split across many upstream packages and repositories.

The generated python packages (see below) do not include the explicit
dependencies to allow the package manager (e.g. apt, zypper, dnf) to install
the dependencies. This would be a nice improvement.

Here are the dependencies needed for this tool to run:

Ubuntu / Debian
- linuxdoc-tools{,-text,-latex}
- docbook{,-dsssl,-xsl,-utils}
- htmldoc{,-common}
- xsltproc
- fop
- sgml2x
- opensp
- openjade
- ldp-docbook-xsl
- ldp-docbook-dsssl
- html2text
- docbook5-xml
- docbook-xsl-ns
- jing
- asciidoc
- libxml2-utils

- htmldoc
- openjade
- sgmltool
- html2text
- docbook{,5}-xsl-stylesheets
- docbook-dsssl-stylesheets
- docbook-utils-minimal
- docbook-utils
- jing
- asciidoc
- libxml2-tools
- libxslt-tools

There are a few additional data files that are needed, specifically, the TLDP
XSL and DSSSL files that are used by the respective DocBook SGML (openjade) and
DocBook XML (xsltproc) processing engines to generate the various outputs.

On Debian-based systems, there are packages available from the distributor
called ldp-docbook-{xsl,dsssl}. There aren't any such packages for RPM (yet).

Supported Python versions
This package was developed against Python-2.7.8 and Python-3.4.1 (on
OpenSUSE). It has been used on Python-2.7.6 (Ubuntu-14.04) and Python-3.4.2 and Python-2.7.9 (on Debian 8).

Continuous Integration testing information and coverage can be reviewed at
`this project's Travis CI page <https://travis-ci.org/martin-a-brown/python-tldp/>`_.

This is a pure-Python package, and you should be able to use your favorite
Python tool to install it on your system. The python-tldp package (`ldptool`)
requires a large number of other packages, most of which are outside of the
Python ecosystem. There's room for improvement here, but here are a few

Build an RPM::

python setup.py sdist && rpmbuild -ta ./dist/python-tldp-${VERSION}.tar.gz

There's a generated file, `contrib/tldp.spec`, which makes a few changes to the
setuptools stock-generated specfile. It adds the dependencies, marks the
configuration file as %config(noreplace), adds a manpage and names the binary
package `python-tldp`.

Build a DEB::

Check to see if the package is available from upstream. It may be included in
the Debian repositories already::

apt-cache search tldp

The quick and dirty way is as follows::

python setup.py --command-packages=stdeb.command bdist_deb

But, there is also a `debian` directory. If you are working straight from the
git checkout, you should be able to generate an installable (unsigned) Debian
package with::

bash contrib/debian-release.sh -us -uc

Install using pip:

Unknown. Because the tool relies so heavily on system-installed non-Python
tools, I have not bothered to try installing the package using pip. It should
work equivalently as well as running the program straight from a checkout.
If you learn anything here or have suggestions, for me, please feel free to
send them along.


* `Canonical python-tldp repository <https://github.com/tLDP/python-tldp>`_
* `Source tree on GitHub <https://github.com/tLDP/LDP>`_
* `Output documentation tree (sample) <http://www.tldp.org/>`_