I'd be interested in giving a small subset of the developers login
accounts on the machine where wikipedia runs. This will enable people
other than me to install the latest versions of the software, run update
programs, and things of that nature. This will surely be beneficial if
it works out, because I'm something of a bottleneck in the process.
Right now I'm thinking of Brion and Axel and Magnus, because they've
contributed the most code and I've talked to them the most in email
about technical stuff.
I don't want this to be a prestige thing, or for anyone's feelings to
be hurt if I don't give access to them.
If we set group permissions correctly, I think that 'root' access is not
necessary. I can still install new versions of PHP, or recompile MySQL, or
things of that nature, from time to time as needed.
--Jimbo
Conversion script tries to convert q{/foo/} to q{[[Articlename/foo|/foo]]/}
But q{/foo/} are not subpage links in UseMod wiki.
They aren't links at all.
q{/foo/} are actually used to write down pronunciacion in a few
articles. But maybe that should be used that way.
q{} are of course quotes.
I will write and ask him to subscribe here. Let's give him a big welcome.
----- Forwarded message from Guillaume Blanchard <gblanchard(a)arcsy.co.jp> -----
From: "Guillaume Blanchard" <gblanchard(a)arcsy.co.jp>
Date: Tue, 19 Mar 2002 16:21:51 +0900
To: <jwales(a)bomis.com>
Subject: French & Japanese wiki !!
Hello James,
I'm a french progrmmer living in Japan (Yokohama).
I joined the french wikipedia project last week.
It's so exciting !!
I already wrote some article about video games & egyptology.
Speaking about Wiki around me, I founded a japanese friend who wnat to help
and grow up Wiki-japan.
The probleme is the CGI actualy in the homepage is an english version.
I talk with Brion Vibber and Yussefsan about this, they say a PHP version is
under construction. I understood the way of how to help programming in
sourceforge, but how to update the server version of fr.wikipedia or
ja.wikipedia ?
Physicaly, the database is on sourceforge ?
Can I access the data base from my own homepage ?
Cheers,
Guillaume (alias Aoineko)
>From root(a)nupedia.com (Cron Daemon) Mon Mar 18 23:32:50 2002
----- End forwarded message -----
Jimbo:
When will you be able to setup a test site for Polish Wikipedia
on PHP script ? If you don't have time, we may set it up somewhere
else (just for testing, not any fork).
I have two issues with import script:
1)
It still tries to import db.orig files from UseMod wikipedia.
Patch attached. These warnings say when there is something wrong.
2)
"Most wanted" page doesn't seem to be up-to-date after import.
Many of listed pages actually exist.
Even editing them or pages that link to them doesn't help.
Affected pages are:
4th place) 53 artykulow chce Stany Zjednoczone/Podzia&lstroke; terytorialny
7th place) 40 artykulow chce Zwi&aogonek;zek chemiczny
11th place) 37 artykulow chce Cz&aogonek;steczka
And others.
What they have in common is that they contain Polish diactrics.
----
Anyway these are minor problems.
Jimbo, could you set up test site pl2.wikipedia.com with PHP
script and data imported from pl.wikipedia.com ?
Preferably after this patch is commited to CVS.
Generally it worked, but there were some problems
1)
Conversion script tried to convert .db.orig files:
Now there are articles like "Propozycje Tematow.db.o"
with contents "There was an error converting this file : file not found."
2)
Article q{"Casablanca"} (WITH quotes, i don't know how to write
it down, let's say that q{} are stronger quotes ;)), currently
a redirect to q{Casablanca (film)} was converted to q{\"Casablanca"}.
I wouldn't mind if it was just removed from database.
3)
Some articles are weird, like:
Polimery
12c12 < [?rednia masa cz?steczkowa polimer?w]? --- > Masa cz?steczkowa polimer?w
That looks like a diff which was imported as article.
I have no idea what could cause it.
I just checked in a fix for an issue that was reported on Bug Reports:
the history of pages which contain a #REDIRECT is invisible; if you
click on "Redirected from" after being redirected, you immediately get
an edit screen.
I changed that behavior: clicking on "Redirected from ..." will now
give you a regular page which shows the redirect, lets you click on
History, Edit this page, Pages that link here etc. as ususal.
I introduced a new global $redirect. If the request
&action=view&redirect=no comes in, then we operate as usual except we
don't follow redirects.
Axel
By parsing the search string, we get the benefit of full boolean
search, which is definitely cool. But our current regime has one
severe downside: any search which has a single short word or stopword
or non-existent word in it will automatically fail since we assume an
implicit "AND" between all search terms. The unadorned mysql "MATCH"
operator does not have this downside: if you search for "the chinese
wall" for example, "the" will be silently ignored and you get the
expected hit.
I am wondering if we can combine the best of those worlds. This should
decrease the number of complaints about short search terms
dramatically, maybe even to the point that we can keep the current
index size.
How about this: every subsequence of the query string which doesn't
contain any +/- boolean operators (see below) is passed to the MATCH
operator as is, which assumes an implicit "give me the best matches
you can find for these terms, the more matching terms the better".
Then we could have two additional operators: + and -. If a word is
preced by +, it *must* be presents, if a word is preceded by - it
*cannot* be present. That allows to express any complicated query we
can right now, but should result in much fewer failed searches. Does
that seem feasible?
Axel
On mar, 2002-03-12 at 12:56, Magnus Manske wrote:
> Congratulations! You succeeded where I failed miserably.
Now, if I can only figure out how to read those occasional garbled
_current_ page files...
> Suggestion: Keep all old versions as they are, and generate an "artificial"
> edit, signed with "conversion script", where the sctual "subpage
> translation" takes place. So all old versions are stored correctly, and the
> conversion itself is more transparent.
Sounds reasonable. I'll try hacking that into place later today.
> Really, great work! :)
Thanks!
-- brion vibber (brion @ pobox.com)