On 06/11/2013 10:31 AM, Joachim Breitner wrote:
Hi,
Am Dienstag, den 11.06.2013, 08:26 +0200 schrieb Dirk Hünniger:
So the version I sent in yesterday works now without any changes but
does not produce the list of figures and the list of contributors, since
toolsever entirely gave up. So currently I can not even work on the
problem. So if you think you can add the current revision to debian,
just add it. Otherwise it might take a longer time until I will be able
to add a new version, because I currently have very little Idea how to
fix that problem.
It seems to work for me, e.g.
https://toolserver.org/~daniel/WikiSense/Contributors.php?wikilang=en&wikifam=.wikipedia.org&page=Debian&since&until&grouped=on&hideanons=on&max=100000&format=html
but only sometimes...:
Database Error: User 'daniel_www' has exceeded the
'max_user_connections' resource (current value: 15) (sql-toolserver) on
sql-toolserver/toolserver
Have you talked to the toolserver admin whether the way you use it is ok
for them, if they can increase the limit or maybe work on an interface
that works better?
Scraping HTML doesn’t seem to be too stable anyways. Can’t you use the
CSV or other formats?
Greetings,
Joachim
I don't feel like discussing with them. I attempted several times. In
particular the devoloper of Contributors php knows that I am using
Contributors.php is this way. Its very unlikely that they will make it
more stable. Toolserver is the place where all unstable code of
Wikipedia and friends runs.I can make some heuristics like try with 10
seconds timeout ten seconds ten times, which produces good results. I
also can get the list of contributors directly from the mediawikiserver
processing the version history with my own html parser, which is the
same parser as I use for articles. I also tried processing with an xml
parser but this failed since mediawikis results are not always well
formatted. Today I feel quite tired an I don't really feel like
implementing the way using the version history right now, although this
is the most right way of doing it I can think out. I might still I might
do sooner or later, but I don't know when. Also Pierre Neidhards will
very likely switch the repo form SVN to GIT in the next few days. So it
will take time until I can resolve that.
Yours Dirk
--
To UNSUBSCRIBE, email to [email protected]
with a subject of "unsubscribe". Trouble? Contact [email protected]
Archive: http://lists.debian.org/[email protected]
