#165 [[User:0]]

[Toby Bartels]

Anybody notice this new user lately?
Seems to be popping up all over the place.
An interesting user name, although unfortunately
we do also a regular article [[:0]] (a redirect).

Actually, [[User:0]] is something that shows up
in histories (also Watchlist, Recentchanges) a lot,
in the place of what should be an ordinary user,
such as me.

Low priority on this, I suspect, unless it's
a symptom of an actual lack of info in the database.

[Lee Daniel Crocker]

Logged In: YES
user_id=3076

I think it was just the one time I accidentally deleted some
info from the database, but I believe it's restored. If you can
point me to somewhere (an exact URL) it shows up still, or a
way to create such an entry, I'll be able to do something.

[Toby Bartels]

Logged In: YES
user_id=573955

I don't know how ot create it, but here's an exact URL:

http://www.wikipedia.com/w/wiki.phtml?
title=User:Toby_Bartels/todo&action=history

I didn't give one before, because I didn't think that it was static, but
upon further reflection, I guess that it is.

[Lee Daniel Crocker]

Logged In: YES
user_id=3076

I only see on such entry on the list, and it's from 7/19, the
day before the installation of the new software.

[Toby Bartels]

Logged In: YES
user_id=573955

Yes, that's the one. Now change the URL to

http://old.wikipedia.com/wiki/wiki.phtml?
title=User:Toby_Bartels/todo&action=history

It shows up correctly there (and is the last edit, hmm ...),
so the problem is in the new software, whatever it is.

I never kept close track, but I don't think that I ever saw [[User:0]]
more than once on a single page, and always a couple days old.
Quite possibly, this only infected the most recent edit on given page
before the switch to Phase III. But not every page, only some.
Might this be related to the bug that dropped some of the history?
It too treated the most recent edit on a given page differently,
combining all changes since April 19 into a single edit at that date.

[Lee Daniel Crocker]

Logged In: YES
user_id=3076

Yes, a few entries got lost in the timeframe of 7/19-7/20.
That appears to be about 10 articles in the whole database,
and there are another dozen or so that were previously bad
(old mistakes of Magnus, I imagine). Since it's just a one-
time loss and the software doesn't create them, and there's
really little I can do to fix them, and it isn't vital information
anyway, I'm just closing this.