Re: UDD and DEHS (Was: Please provide a simple example)
On Mon, 2 Mar 2009, Stefano Zacchiroli wrote:
Looks good. Any hint how to reasonably obtain
description of binaries
long_description of binaries
homepage
license (perhaps)
Regarding (short) description I have this file generated daily using
UDD available for download:
http://master.debian.org/~zack/pts/shortdesc.txt . It is used by the
PTS, so it is relatively controlled and safe to use.
Thanks. I just solved the problem to obtain all fields I need (and
actually I try to inject even more fields into UDD to make it comparable
to the sources table). So this is bacially solved.
... but of course you can use UDD in the first place, as data comes
from it. The script generating is at
http://svn.debian.org/viewsvn/*checkout*/qa/trunk/pts/www/bin/retrieve_shortdesc.sh?content-type=text%2Fplain
AFAIK, we don't have any long description in UDD yet.
It is there. See packages.long_description. I'm able to parse the
<package>.html files in ftpnew to obtain the information I need for
fptnew_packages table which will be similar to packages.
You can get the homepage using the SOAP interface of the PTS, but is
not suitable for massive retrieval yet (see #507454), and possibly
will never be. If you want I can export that somewhere too.
Same as above.
For license you have no chance until the machine parseable copyright
format get more widespread.
http://ftp-master.debian.org/new/<source>_<version>.html#binary-<pkg>-copyright
Format-Specification: http://wiki.debian.org/Proposals/CopyrightFormat
is used.
Indeed.
This has to be done and IMHO it is sufficient for my purpose to parse
those packages who follow this format. Others will remain "unknown" /
'' or NULL (whatever you prefer for unknown values.
Kind regards
Andreas.
--
http://fam-tille.de
Reply to: