Can anyone explain this?
I just spent about two dozen times trying to submit the new page
"Chelicerata" and finally got an edit conflict. It says that someone
submitted the page while I was submitting it, but the page the other guy
submitted is blank. If it were the same as what I tried to submit, then the
other guy would be my previous attempt, and if it contained more text, then
my previous attempt succeeded and someone else then edited it.
I am a programmer here at Bomis, and I have been asked to get
pl.wikipedia.com and es.wikipedia.com running on the new (PHP/MySQL
based) software. I have also been asked to help resolve the
performance issues that Wikipedia has been having. I am new to this
list, so I'll be taking a little time to look over the archives before
I start doing anything drastic, but I wanted to introduce myself.
"Jason C. Richey" <jasonr(a)bomis.com>
I'm on vacation until next Sunday... but I will try to get to some
Wikipedia related requests, particularly for those related to the
I'll also instruct Jason to set up the pl2.wikipedia.com server,
in consultation with people on the tech list. jasonr(a)bomis.com
is his direct address.
Jason, can you help the wikipedia folks out? We're most interested in
getting the pl site on a test wikipedia with mysql, and the esperanto
Also, give them access_logs or whatever they need to help diagnose the
After much clamoring from many quarters, I decided to add to
special:whatlinkshere a list of pages that link to the current page via
redirects. Doesn't seem to have exploded so far... Feel free to improve
it / integrate it better with the rest of the list.
One potential problem is that a redirect page could, conceivably, have
additional links in the text after the #REDIRECT line, and the current
code would think those links were actual redirects and list them
accordingly. But, well, they shouldn't.
Also, I've disabled redirect-following for diffs. If I want to diff a
redirect page, goshdarnit, I want to diff the redirect page, not the
page that it (now) redirects to!
-- brion vibber (brion @ pobox.com)
I just noticed the link tables in the 'pedia are (still?) inconsistent. For
example, [] shows up on the Orphans page, but 15 (!) articles link to
Currently, we generate the skins from within the code. This really starts
getting ugly. How about:
* For each skin, there's a dummy HTML file
* It contains special markup like <a
* These could be enclosed in <showRandomPages>....</showRandomPages>
* On display, the HTML is adapted to the viewer and the page to be viewed
(no "delete this page" on special pages etc.)
This would make it easy to keep the skins up-to-date, add more skins, and
keep the code relatively clean.
The HTML for the skins could even be stored as a "wikipedia:" article.
Thinking of it, we could store the whole PHP script as protected articles
and update the running version from there... ('scuse me, going berserk;)
There are a couple of simple fixes I can make easily, like fixing the
<dl> code, recognizing ".JPG", etc. I have downloaded the latest
stuff from CVS. What would be a quick and convenient way to get
those fixes done? Posting a diff to this list? Sending a complete
file to one of the maintainers? I'm not sure I want or need CVS
checkin access; this is not something I'll be doing regularly.
After several unsuccessful attempts to find out about Willard van Orman
Quine, I get this error:
Warning: MySQL: Link to server lost, unable to reconnect in
/home/wiki-newest/work-http/databaseFunctions.php on line 11
Could not connect to MySQL server: