Re: Useful thing to implement for the Debian web site
[ moving discussion to debian-www@lists.debian.org ]
* James A. Treacy <treacy@debian.org> [20001124 11:18]:
> If you would like to do something useful for the web pages, here is
> a good project.
>
> Many people want to be able to view changelogs (both Debian
> and upstream) and the copyright file from the package.d.o web
> page. Currently, this is done by pointing at a copy generated
> by another program which isn't very reliable.
>
> What I'd like to have done is to extract the requested file
> directly from the .deb. You will quickly find this is not as
> easy as you think, since there are multiple locations the
> file could exist - and it may even be in a different package.
> You will probably also find a number of bugs in packages.
> Bonus points if you send mail somewhere useful so a bug can
> be filed. Culus may already have implemented a way of extracting
> these files. Either way, please make it efficient.
>
> A copy of the file is then placed in a simple db so future requests
> will simply grab the copy from the db. A timestamp will be
> needed so you can test if the db copy is more recent than
> the .deb.
>
> The script will be run on one of the development machines with
> a local copy of the archive.
Since we are starting all over anyway, let's discuss what is needed.
For example, many bugs are files because non-us packages cannot
be viewed. I suggest moving packages.d.o to a machine which is not
located in the USA (but pandora doesn't have much disk space, so
this pandora might not be an option). Any machines that could be
used?
Also, viewing the sources was requested (see #17590). This definitely
needs a machine outside the USA (showing the changelog would probably
be legal if the machine was in the USA).
--
Martin Michlmayr
tbm@cyrius.com
Reply to: