Hi all.
I'm testing the ground for writing a java aplet based mini-editor for
mediawiki (and maybe other wikis, although the lack of a standard makes that
difficult).
What I have in mind is a relatively light applet - a more specialised version
of the generic text input, that creates well formatted wiki markup, by
allowing a syntax highlighted or whysiwyg view of the article.
This wouldn't require comunication with the server, other than uploading the
article, as happens now.
The program would be GPL, and as far as possible be designed to work with the
GNU classpath.
What do you think of this idea? My main aim is to make editing easier for
non-technical users.
Jim
Hi guys...
Exist an adapt-script or workaround from mediawiki 1.2.6 to pahse3? So
in phase3 are more tables and columns as in 1.2.6...
Thanks for any help..
regards guido
JeLuF recently set up the Wikimedia servers to send their syslogs to zwinger,
and we noticed that from each of the squids, messages like the following are
regularly appearing in /var/log/messages:
Jun 15 11:00:53 coronelli squid[1192]: varyEvaluateMatch: Oops. Not a Vary
match on second attempt, 'http://en.wikipedia.org/wiki/Loop_quantum_grav
ity' 'accept-encoding="gzip,%20deflate",
cookie="enwikiSession%3dd01e31dd136539b2bdd41027d84368c8"'
Jun 15 11:00:55 coronelli squid[1192]: clientProcessHit: Vary object loop!
I've switched the messages off for now by adding "debug_options 33,0" to
squid.conf on each squid, because they were repetitive and drowning out
anything else in the log. Gabriel, do you have any idea what's wrong here?
I think you need the template namespace too, no ?
I will try to work on it, but dont know when....
Hi,
Is it possible and/or of interest to release some kind of hourly XML-dump of
the last changed pages in Wikipedia? I'm running a copy of the Wikipedia and are
trying to keep up with the changes. I believe a XML-dump could ease the load
on the server a little.
For my needs the XML-document must contain at least:
cur title
cur text (in Wiki-format, including URLs to pictures)
last change
Only three of the namespaces (from what i understand): NS_MAIN, NS_IMAGE
and NS_CATEGORY.
cur is redirect
I heard there are more people than me interested in this (Yahoo ...), what
do they want?
I've spoken to Erik Müller on the WOS in Berlin about this and he direkted
me to this list.
Hi,
currently the site is updated from cvs HEAD, not from a branch
of it. Despite this, new features are still going into HEAD, too.
Today, the rc_ip changes Tim announced to the ML went live. The
database did not contain a rc_ip column, though. This broke
saving articles for about 15 minutes.
Can we please use another branch for the main site than HEAD?
Else this will happen again. We need different branches for
production and development. It was OK to have only one branch
for the first week or two that 1.3 was running in production.
It's time to switch.
Regards,
JeLuF
>
>I keep getting
>
> Sorry! The wiki is experiencing some technical difficulties, and
>cannot contact the database server.
>
>intermixed with normal function. Maybe one of the apaches
>really doesn't
>find the database?
They find the database, but the max_connections on mysql is hit (300 actually). If we increase this number, the requests could become very very slow.
Shaihulud
Will the English Wikipedia be moving to UTF-8? I think it's appropriate,
seeing how often non-Western European characters need to be represented
in it. Also, it would be helpful for dashes, of all things. Last week I
proposed UTF-8 as a solution to that never-ending debate. Two people
agreed, and zero people said "the double-hyphen (--) will live forever!!!":
http://en.wikipedia.org/wiki/Wikipedia_talk:Manual_of_Style/Dashes
Nathan
Hi,
Is it possible and/or of interest to release some kind of hourly XML-dump of
the last changed pages in Wikipedia? I'm running a copy of the Wikipedia and are
trying to keep up with the changes. I believe a XML-dump could ease the load
on the server a little.
For my needs the XML-document must contain at least:
cur title
cur text (in Wiki-format, including URLs to pictures)
last change
Only three of the namespaces (from what i understand): NS_MAIN, NS_IMAGE
and NS_CATEGORY.
cur is redirect
I heard there are more people than me interested in this (Yahoo ...), what
do they want?
I've spoken to Erik Müller on the WOS in Berlin about this and he direkted
me to this list.
/cj
Hi,
since the access.log of maurus hit 2^32-1 bytes, squid died. Apparently
access.log was not rotated. The cron jobs are there, but the logrotate
settings where default ones. I changed them to the values browne is
using:
/var/log/squid/*.log {
# size 200M
rotate 24000
compress
delaycompress
# compresscmd /usr/local/bin/nicebzip2
# uncompresscmd /usr/bin/bunzip2
postrotate
/usr/local/squid/sbin/squid -k rotate
endscript
}
Regards,
JeLuF
brwiki didn't get added to the proper database lists, so wasn't included
in the backups used to rebuild the database after last week's crash.
It should be recoverable from the old database files, but this will
require being very very careful. Rather than risk blowing up suda, I'll
try moving the files somewhere else and working with it isolated.
(I'll also be able to recover the archive and imagelinks tables while
I'm in there.)
-- brion vibber (brion @ pobox.com)