Re: Blends pages, tasks pages etc.
On Sun, Jan 10, 2016 at 09:44:24PM +0100, Ole Streicher wrote:
> > On Sun, Jan 10, 2016 at 10:46:54AM +0100, Ole Streicher wrote:
> >> I then get the following output:
> >> UnicodeDecodeError: 'ascii' codec can't decode byte 0xc3 in position 1: ordinal not in range(128)
> found and fixed. My fault; I had to add client_encoding="utf-8" to the
> psycopg2.connect() argument list.
> > Pushing is fine. I'll check here with a local UDD clone which I'm using
> > due to a bad connection. Creating a mirror once a month (or week if
> > urgent changes) is way cheaper for me.
> If you have a script, could you share it?
There is an official script which is linked in the official mail
I *personally* use an adapted script in the teammetrics effort that adds
a table with normalised developer names. The UDD download is adapted to
my personal needs behind a slow connection. Please note that the latest
change with the md5sum check is not yet tested (nor the text really
done). You can find it here
> BTW, I found a few problems, and now the pages are generated. The
> remaining problem seems to be that the section "No known packages
> available but some record of interest (WNPP bug)" is empty.
Yes, that's now on my todo list after the recently announced new
> And there
> are two index file sets, one (empty) in
> /srv/blends.debian.org/www/astro/ and one (filled) in
Yes. The script tasks_udd.py outputs to tasks_udd to not interfere with
tasks. You can compare the diff with test_output.py after *also* running
the currently in production tasks.py
> When I find some time: may I clean up a bit? Things like
> print("xyz", file=fp) --> fp.write('xyz\n')
Feel free to clean up. As I said my goal is to get the remarks and the
"no known packages available" section by using the new UDD tables.