Plans for SkoleLinux developer gathering in Greece
I just want to inform you about my current plans to bring the CDD effort
foreward. Because I take part in the SkoleLinux developer gathering I
think it might make sense to discuss this a little bit in advance to make
sure people are informed and interested.
Currently people in Spain are working on a rewrite of the cdd-tools. So
I stalled my work on it to avoid doubling of work (see
Sergio, any new things to report, any SVN code available, any help needed???
So I thought about things which are not really mentioned there and at
Chemnitzer Linux-Tage event I remembered the point about creating CDD
documentation I mentioned at
what wee need is kind of a toll which helps building internationalized
web pages about what's inside the CDD, what could be included soon and
what's on the long term todo list. At least I was often asked about
things like that. Currently Debian-Med does this more or less manually
at the web pages for example
but this is mostly manual work and there is something which could easily
be done automagically. For instance the part "Official Debian packages"
(the "green" items - just notice the traffic light symbolic) could be
derived from the package descriptions and if we use the DDTP projects work
we could even get internationalized pages fairly easy.
For the not yet included stuff (inofficial packages = "yellow" and just
interesting projects without avialable packages ="red") this is a little bit
harder to do, but also not so important as the "green" ones. We would
perhaps need some descriptions in PO-files which can be read out by
a tool which builds some web sites and perhaps we find even a way to
make kind of a "CDD Doc-browser".
The sense of making this kind of documentation seems obvious to me: It
enables us to say: "RTFM", in case somebody is asking what is Debian-<CDD>
Do you think that this is useful or should something different be done first?
Any hints about implementation?
Regarding the implementation I would think that a daily build of this kind
of pages would be apropriate. A dynamic PHP script would not make sense to
me because we have changes to these pages at most once per day (if a
package description of an included package has changed) but probably only
after new meta package versions.
Because we decided for Python for cdd-dev-tools rewrite I guess it might
be sane to use Python scripts for parsing the packages files (original
ones an DDTP-translated ones) and a database of the not yet included
software (yellow and red ones).
I have to admit that I did not before such kind of Packages file parsing
before and do not know whether this is a sane decision compared to the
available tools inside Debian and hints would be welcome.