Jump to content

Wikipedia:Village pump (technical)/Archive 48

From Wikipedia, the free encyclopedia

Someone's asking here where the Image:WW2 Iwo Jima flag raising.jpg has gone. Could it be one of those that got lost in error a couple of weeks ago? As it's such a significant photo I think we ought to restore it somehow a.s.a.p.--82.148.54.182 (talk) 13:43, 20 September 2008 (UTC)

I suspect it was one of the images lost in that error. Unfortunately those can't be restored unless someone has a cached version or some other way of reuploading it. Stifle (talk) 14:06, 20 September 2008 (UTC)
It isn't on the original list of lost images, so that's probably not the case. Graham87 15:06, 20 September 2008 (UTC)
There's definitely something going on, I've noticed a few number of images with this same issue that aren't on that original list. In the mean time, I've rescued Image:WW2 Iwo Jima flag raising.jpg from archive.org. Anomie 15:35, 20 September 2008 (UTC)
On the very day that image deletion mistake was reported and that list of lost images was published I had to restore some of the ambox images since they were lost. And they were not on that list. So that list is not complete. Anyway, good that you fixed a new copy of the image.
--David Göthberg (talk) 20:47, 21 September 2008 (UTC)
I see that the image is 417x336 pixels, 96dpi, and is commented "No higher resolution available." There is a 600x489 pixel 150dpi copy of that image available at this US Govt URL. -- Boracay Bill (talk) 22:27, 21 September 2008 (UTC)
I actually resized it down from the 1380×1111 version at archive.org, as it was apparently decided that NFCC 3b needed to apply. Anomie 00:04, 22 September 2008 (UTC)

login screen?

Was a change made to the 'login successful screen that instead sends us back to the main page? ThuranX (talk) 22:32, 20 September 2008 (UTC)

Yes: "When a user logs in when following a link from Special:Preferences to the login page, they are now returned to Special:Preferences rather than Main Page after the login (most other pages work like this, but previously Special:Preferences didn't). (r40773, bug 12284)" (From Wikipedia:Wikipedia Signpost/2008-09-15/Technology report - Erik Baas (talk) 11:42, 21 September 2008 (UTC)

Impossible edit "creates" page with no title

The top most log entry here seems to create a new page with no title. The diff and hist buttons don't function properly. It appears to be a error of some sort but has anyone seen that behavior before? Dragons flight (talk) 02:14, 21 September 2008 (UTC)

I was testing the write api --Chris 02:16, 21 September 2008 (UTC)
According to my logs the title parameter would have been set to #redirect [[Wikipedia talk:Requests for adminship --Chris 02:19, 21 September 2008 (UTC)
I'm guessing # broke the syntax parser. But that still leaves the question of what if anything did you edit, and if you didn't edit anything why was a log entry created at all. Dragons flight (talk) 02:31, 21 September 2008 (UTC)

He created a page in the main namespace whose title is the empty string.

mysql> SELECT * FROM page WHERE page_id=19405691;
+----------+----------------+------------+-------------------+--------------+------------------+-------------+----------------+----------------+-------------+----------+
| page_id  | page_namespace | page_title | page_restrictions | page_counter | page_is_redirect | page_is_new | page_random    | page_touched   | page_latest | page_len |
+----------+----------------+------------+-------------------+--------------+------------------+-------------+----------------+----------------+-------------+----------+
| 19405691 |              0 |            |                   |            0 |                0 |           1 | 0.287517052064 | 20080921020208 |   239911065 |      190 | 
+----------+----------------+------------+-------------------+--------------+------------------+-------------+----------------+----------------+-------------+----------+
1 row in set (0.02 sec)

mysql> SELECT LENGTH(page_title) FROM page WHERE page_id=19405691;
+--------------------+
| LENGTH(page_title) |
+--------------------+
|                  0 | 
+--------------------+
1 row in set (0.00 sec)

The API didn't check for this condition correctly, so it let the edit through. In r41085 I should have helped to prevent this from happening in the future by adding a low-level check to avoid it. Ideally we'd have the lowest possible level of checking here, with constraints in the actual database, but MySQL doesn't support that. But we still need higher-level checking to return a graceful error message instead of an uncaught exception, which the API still doesn't have at the moment.

The reason this happened, I'd guess, is because the API checked for a valid Title object, and assumed that any such object would actually be valid to have in the database. It's not, necessarily. In particular, a Title can have empty text if it has a fragment, so #Foo becomes a Title with namespace = 0, text = "", and fragment = "Foo". This is fine for linking to something, but trying to add it to the database is wrong. The usual GUI editing handles this correctly, but the API apparently didn't refactor the code correctly and missed that check somewhere along the line. Someone should check if it handles interwikis wrong too ― those are another sort of Title object that can't be put in the database. —Simetrical (talk • contribs) 18:25, 21 September 2008 (UTC)

Can't delete pages

I can view articles and histories fine but when I try to delete articles I get a "Can't contact the database server: Unknown error ())". Example: I can't delete Talk:Cold (Fool's Garden song).--Commander Keane (talk) 08:19, 21 September 2008 (UTC)

Same here with article and talk namespaces. Deleting stuff in the image namespace works fine. Angus McLellan (Talk) 08:39, 21 September 2008 (UTC)
It is working fine now.--Commander Keane (talk) 08:49, 21 September 2008 (UTC)

Admin bot proposal

Since the affects bot operators I will post this here, there is a discussion regarding whether or not administrators need approval to run bots under their account. All are invited to comment. Prodego talk 14:50, 21 September 2008 (UTC)

Background colour in articles

I only wanted to suggest that the default background colour for the articles can be changed by the reader, preferably to another very light non-white colour. Me and other frequent readers spend hours reading articles, and the white background is really tough for the eyes even after changing the monitor settings. An option to change it to e.g. cyan,(or other really light colours) would be very useful for readers that spend a lot of time on the site, or readers with eyes sensitive to intense light. —Preceding unsigned comment added by 86.9.125.223 (talk) 17:31, 11 September 2008 (UTC)

If you register an account, you can change your Monobook.css file so that the background color can be anything you want. In most other scenarios (such as readers with light sensitivity), the best option isn't to tweak the background color (which could cause readability issues), but to either turn down the brightness on the monitor or to customize the colors as I pointed out. EVula // talk // // 18:27, 11 September 2008 (UTC)
With an account one can choose green text on a black background. I find this much easier on the eyes. DuncanHill (talk) 22:35, 11 September 2008 (UTC)
Your browser should also give you the option to override site styles in some fashion even if you aren't logged in. Firefox does, at least, and I expect so do all other non-IE browsers. This is not a discoverable or easily-used function, though: you have to know what CSS is to know about it. —Simetrical (talk • contribs) 16:11, 12 September 2008 (UTC)
. . . #ffffec . . .

Lorem ipsum dolor sit amet, consectetur adipisicing elit, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua.

. . . #ffffff . . .

Lorem ipsum dolor sit amet, consectetur adipisicing elit, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua.

Support off-white as a sitewide background color for mainspace. On my own MediaWiki wiki, I've set #ffffec as the background color for all pages and it does a great deal to lessen eyestrain. Indeed, I've set a similar background color in my email client and text editor.

This is not a user-specific issue, although some users may complain while others suffer silently. Did you ever wonder why slide rules are yellow? Light of different colors is refracted differently by a lens, for example that in the human eye; this is chromatic aberration. People with entirely normal eyesight will see a slightly blurred image when black text is displayed on a pure white background. This is true even when the display is homogenous, such as ink on paper. A lightly colored background reduces this effect.

Some people may feel less annoyed by chromatic abberation than others but then, there are people who are comfortable reading a book next to an operating jackhammer. That's not an argument against quiet living rooms. Nor should we expect casual readers to tinker with technical details; the project should be as presentable as possible, by default.

The current fashion of black-on-white pages is in part a reaction against the excesses of the early web, with magenta text on animated GIF backgrounds; in part it is related to the demand for white kitchens, white carpets, white sugar, and Wonder Bread -- the perceived superiority of pure whiteness. We're coming to realize that white homes are hard to keep clean and white foods hard to digest. We should also realize that white web pages are hard to read.

Xiongtalk* 12:20, 13 September 2008 (UTC)

Xiong: I took the liberty of making your whole comment use your off-white colour to make it a better example. If you don't like it just remove it.
Everyone: Since this talk page has slight blue background these examples look more yellow than they really are. They will not look as yellow if/when used for the whole article background.
--David Göthberg (talk) 14:17, 14 September 2008 (UTC)

Regular white #FFFFFF, currently used in articles.

Off-white colour #FFFBF0 that David uses in his computer.

Support - We can make this as a user friendly menu choice even for IP users, see explanation further down.

Books and newspapers are on purpose printed on off-white paper, since that is easier on the eyes. (While for instance brochures are usually printed on full white paper since that does look better, but you are not supposed to spend as much time looking at them.) I have good eyesight but sensitive eyes, so already many years ago I set my computer to show all "white" areas as a slight off-white. See my colour example to the right. I even get that colour in the Wikipedia edit window!

However I know many of my friends find the colours in my computer slightly strange. Many seem to prefer full white. (In spite that they complain they get tired eyes from the computer and can't understand how I can spend 15 hours a day 5 days a week in front of the computer.)

I am no JavaScript expert but I think we can do the following:

We probably should keep the full white colour in articles, but add a menu option in the sidebar to the left that says something like "Off-white background". That menu option can activate a JavaScript that changes the background to off-white. And it can set a JavaScript cookie so the same background is shown the next time the user visits. I think that should work even for IP users.

--David Göthberg (talk) 14:17, 14 September 2008 (UTC)

Clarification: Oppose default change, support user option if technically viable - If what Simetrical states below is true (that such IP user cookies would ruin the Squid caches) then we can not have such a menu choice. Unless of course it is added to the MediaWiki software in a smarter way. If the default should be changed then it should of course be a much lighter tint than the one I suggested above for the user option. But I don't think we should change the default colour since my experience is that most readers prefer the full white, in spite that it is bad for them. So could the devs perhaps at least think about if they could add this option in some way that wouldn't ruin the caches?
--David Göthberg (talk) 18:57, 17 September 2008 (UTC)
I can't think of any way, offhand. —Simetrical (talk • contribs) 16:34, 18 September 2008 (UTC)
See below. --Splarka (rant) 07:53, 19 September 2008 (UTC)

IIRC the developers have objected to (and intervened to revert) use of cookies for IP users in the past [dismissable sitenotice], and I believe it is technically not allowed by the privacy policy (which idiotically says "There shall be no cookies" instead of something more sensible like "We won't track you") --Random832 (contribs) 15:35, 16 September 2008 (UTC)

I doubt there's a privacy policy issue. Maybe there is, but that would be fairly idiotic, I agree. But the Squid caches use Vary: Accept-Encoding,Cookie. If you give out cookies to half of anons, Squid can't serve pages generated for one set to the other. You're fragmenting the cache. This is the reason a number of proposals related to customizability for anons have been nuked in the past. —Simetrical (talk • contribs) 16:48, 16 September 2008 (UTC)
Wikimedia Squids now use X-Vary-Options headers, which is a custom patch by Tim Starling. This allows anonymous cookies (not matching important ones) to not invalidate cache. --Splarka (rant) 07:53, 19 September 2008 (UTC)
But that would mean that the cookies wouldn't actually work, wouldn't it? Or are you thinking something like having the color change be achieved by some JavaScript that checks for the cookies, which Squid is instructed to ignore? Slightly evil, but that might work . . . —Simetrical (talk • contribs) 15:55, 19 September 2008 (UTC)
Exactly. It could even be done before page loads. A simple check of if(document.cookie.indexOf('mediawikicontentcolor')) that would then get the value and appendCSS(#content {background-color: something}). Wiktionary does this heavily, IIRC. --Splarka (rant) 05:45, 20 September 2008 (UTC)

I'd very strongly support a change of default background colour. Whether to a noticeably different buff tone or just a subtle change, unnoticeable without side-by-side comparison, it's a quick accessibility win, and should hurt no-one. Please remember that many people don;t or can't have javascript enabled. Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 20:39, 16 September 2008 (UTC)

  • Support the change (i like the yellow, but support other good colours). After looking at the yellow backgrounded (is that even a word?) text, I truly reckon that it is less of an eyestrain than white. And of couse, since Wikipedia's main slogan is something along the lines of making knowledge more accessable, the colour change could help in a small way. — ^.^ [citation needed] 12:15, 17 September 2008 (UTC)
  • Oppose default change. Turn your damned screen brightnesses down if white hurt your eyes. Newspapers don't have brightness controls, monitors do. High-contrast backgrounds aid significantly in the readability of low-resolution text, which is what most of our readers are lumbered with. Setting an off-white in addition to green-on-black as an easy option in prefs might be welcome, but changing the default isn't. Yellow would be an act of insanity. Chris Cunningham (not at work) - talk 12:33, 17 September 2008 (UTC)
    • Newspapers don't have brilliant white paper (as you can see, if you hold one up to your monitor while viewing WP); they use off-white, for the very reason this change (which is not about my eyes, not any other individual editor's, but about courtesy and consideration to all our users) has been proposed. Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 16:32, 22 September 2008 (UTC)
  • Strong oppose. This would make Wikipedia look clowny. We're not MySpace; if someone want's a different color, they can easily edit their preferences or their monobook.css. This would not only be ugly, but probably scare off readers looking for reliable information. Admiral Norton (talk) 13:28, 17 September 2008 (UTC)


Bad url with + instead of _

A few minutes ago I was on http://en.wikipedia.org/w/index.php?title=Final_Destination_books&oldid=238994860 and clicked the diff link to the previous revision. It took me to the weird http://en.wikipedia.org/wiki/Final+Destination+books?diff=238994860 (note it has '+' and not '_' in the url). The diff is displayed but there are strange tabs like redlinked article and discussion with '+' in the urls. Now I get the expected http://en.wikipedia.org/w/index.php?title=Final_Destination_books&diff=prev&oldid=238994860 when clicking the same diff link. By manually manipulating the url I can create similar effects on other pages, for example http://en.wikipedia.org/wiki/Wikipedia:Village+pump+(technical)?diff=239982153. PrimeHunter (talk) 13:18, 21 September 2008 (UTC)

That's funny. Maybe some recent change to MediaWiki broke something and was reverted? Anyway, the reason the diff itself works is that the diff code completely ignores the title: //en.wikipedia.org/w/index.php?title=NoSuchPage&diff=238994860 works just as well. It just looks up the revision ID and uses that to find the revisions to compare. You can even make diffs between revisions from two completely different pages by manually filling in the "diff=" and "oldid=" parameters. —Ilmari Karonen (talk) 16:15, 22 September 2008 (UTC)
Hah. I encountered that same problem for a few minutes yesterday, then it went away. Corvus cornixtalk 20:48, 22 September 2008 (UTC)

Bizarre diffs

I have noticed recently a number of very odd diffs, this [1] is typical, An editor is adding a comment to a discussion, but the diff shews many trivial alterations all through the page. Anyone know what is occurring? DuncanHill (talk) 02:32, 22 September 2008 (UTC)

In that case, it seems that someone's browser translated all literal non-breaking space characters in the page into regular space characters; I've heard Firefox 2 does this. Anomie 03:16, 22 September 2008 (UTC)
I've seen similar diffs without remarking on them. I see that a number of nonbreaking space ( ) chars are shown passing through without being translated into regular spaces, so I doubt that's it. -- Boracay Bill (talk) 04:06, 22 September 2008 (UTC)
A "literal nbsp" means the charater U+00A0, not any escaped version such as   or  . Every instance of that character in the section "Kurt Unblocked" was replaced with a space in that edit, and the edit summary indicates that that was the section edited. Anomie 10:48, 22 September 2008 (UTC)
Ah. The terminology threw me. Still, a google search for "literal nbsp" turned up exactly ten hits. I've been living on a small island in the Philippines for 12+ years, and had scant contact with Unicode, ISO/IEC 8859 and suchlike prior to that (or, for that matter, since). A quick look at Tutorial: Character sets & encodings in XHTML, HTML and CSS gives me the impression that modern browsers are requiredexpected to support unicode, but I do see this section which says, "First point: Always declare the encoding of your documents". Further down, there is this section, titled "HTML and XHTML served as text/html: always use a <meta> element", which says

How to do this. The meta charset declaration should appear as close as possible to the top of the head element. It looks as follows:

<meta http-equiv="Content-type" content="text/html;charset=UTF-8" />
Looking at the XHTML source for this page which I am editing, I don't see such a tag. It occurs to me that this might be a Wikipedia/Wikimedia problem instead of a Firefox problem. -- Boracay Bill (talk) 11:36, 22 September 2008 (UTC)
That's odd. What I see is:
<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">
<html xmlns="http://www.w3.org/1999/xhtml" xml:lang="en" lang="en" dir="ltr">
        <head>
                <meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
                <meta name="generator" content="MediaWiki 1.14alpha" />
Note the first <meta> tag. Anyway, I do suspect it might be a MediaWiki issue rather than a browser bug, but not for the same reason. Rather, the reason is that, when you edit a section of a page, only that section is sent to your browser and submitted back. In this case, the loss of NBSPs seems to happen in other sections on the same page, suggesting that the cause may be in the MediaWiki section-merging code. —Ilmari Karonen (talk) 12:35, 22 September 2008 (UTC)
Looking again, I see the same. My wife was calling me to dinner as I was writing my response above, and I probably rushed the response and missed the meta tag. My bad. -- Boracay Bill (talk) 23:42, 22 September 2008 (UTC)
I will just add that one of the editors to whom this has been happening has had it happen with both Firefox & Safari. DuncanHill (talk) 13:19, 22 September 2008 (UTC)
I've just notice someone asking for an explanation of this diff: [2] (just the paras in green, not the obvious spelling, etc changes). Is this the same thing? Thanks. Doug Weller (talk) 14:12, 22 September 2008 (UTC)
This [3] is slightly different, but has affected one of the editors who has had the original problem I posted. DuncanHill (talk) 15:43, 22 September 2008 (UTC)
I've seen 3 or 4. Doug Weller (talk) 15:46, 22 September 2008 (UTC)
Erik seems to be having some connection problems. May or may not be related. —Ilmari Karonen (talk) 16:04, 22 September 2008 (UTC)

The content-type META tag in the document has no effect at all if the page is served in HTTP form from a web server that includes an HTTP content-type header with a charset declaration among the headers that precede the document's content. I see that the Wikipedia server does send such a header:

Content-Type: text/html; charset=utf-8

so the tag in the document is ignored.

Besides that, in response to someone else: what it means to say that browsers "support Unicode" is that they are supposed to be able to deal with all Unicode characters. But the encoding used to generate the data sent to the browser, or the encoding of the file that the browser is serving, unless it's UTF-8 or some other encoding designed to support Unicode, does not support all of Unicode. For example, US-ASCII, ISO-8859-1, and Windows-1252 are all very limited character sets with less than 256 characters in each. When any of those is used, any characters to be sent that are outside of the encoding being used can only be sent in escaped &#nnnn; or &#xnnnn; format or, for a very few of them, using some predefined mnemonic such as &nbsp; or &oacute; or &Oacute;. —Largo Plazo (talk) 18:31, 22 September 2008 (UTC)

I'm no expert, but the "... for a very few of them ..." part above conflicts with my understanding. W3C XHTML 1.0 Recommendation 26 January 2000, revised 1 August 2002 section C.12. Using Ampersands in Attribute Values (and Elsewhere) says:

In both SGML and XML, the ampersand character ("&") declares the beginning of an entity reference (e.g., &reg; for the registered trademark symbol "®"). Unfortunately, many HTML user agents have silently ignored incorrect usage of the ampersand character in HTML documents - treating ampersands that do not look like entity references as literal ampersands. XML-based user agents will not tolerate this incorrect usage, and any document that uses an ampersand incorrectly will not be "valid", and consequently will not conform to this specification. In order to ensure that documents are compatible with historical HTML user agents and XML-based user agents, ampersands used in a document that are to be treated as literal characters must be expressed themselves as an entity reference (e.g. "&amp;"). For example, when the href attribute of the a element refers to a CGI script that takes parameters, it must be expressed as http://my.site.dom/cgi-bin/myscript.pl?class=guest&amp;name=user rather than as http://my.site.dom/cgi-bin/myscript.pl?class=guest&name=user.

Perhaps browser conformance to this is spotty (??). (also see http://www.w3.org/TR/xhtml1/dtds.html#h-A2))-- Boracay Bill (talk) 23:42, 22 September 2008 (UTC)

Is it possible that this is (was) related to the editor gadget wikEd which became functional under Safari and Chrome yesterday? I had problems to track this down, but I have added a possible fix to wikEd (assuming that browsers sometimes converted &nbsp; into &#160; or \u00a0). Cacycle (talk) 03:39, 23 September 2008 (UTC)

Layout question

Hi, someone suggested I come here to ask my question. On my talk, I'm having an issue with a nav thing I built. I htink it may be an unclosed tag or something, but I can't figure out where. It's causing formatting issues with my TOC & archive boxes; if you look on my talk page you'll see two infoboxes I placed to illustrate the problem more clearly. The first box is FUBAR, the second is displaying correctly. Is anyone here able to shed light on what I did wrong? Thanks in advance. Prince of Canada t | c 17:05, 22 September 2008 (UTC)

Well, it's definitely something unclosed. I threw in a </table> tag and the uboxen both show up. I'll leave it to you to find the bad tag in your NavTemplate though. :) Franamax (talk) 18:29, 22 September 2008 (UTC)
Well, it's so nice to have the servers back, I fixed your NavTemplate after all. You were opening three tables and only closing two. Franamax (talk) 18:33, 22 September 2008 (UTC)
Many thanks. Still trying to figure out why the ClueBot links to my archive are still displaying.. weird. Prince of Canada t | c 18:35, 22 September 2008 (UTC)

Error

Does anyone know what the error was or still is on this website for the past half hour? Simply south (talk) 18:01, 22 September 2008 (UTC)

Template modifications not appearing

Over the last few days I have encountered a problem when modifying templates - the modifications appear on the template's pages when I save my edits, but not on the templates on the pages in which they are used. For example, I modified some parameters in the DCWorld template, but after I saved my edits, and went on a page on the which the template was used, (1992 Davis Cup), my modifications didn't appear. I have encountered this problem several times on several templates in the past days, and I don't know if it is because I do something wrong in my edits, if the problem comes from my computer, or from wikipedia. --Oxford St. (talk) 15:04, 23 September 2008 (UTC)

This happens because updating transcluded templates isn't done immediately, but is instead put on the job queue to do gradually. This reduces load on the servers, but means you have to wait a while. If necessary, you can purge the cache of the pages where the template is used; this will force the servers to update immediately. Algebraist 15:11, 23 September 2008 (UTC)
OK, thanks. --Oxford St. (talk) 15:27, 23 September 2008 (UTC)

The following two pages appear to have identically arranged AfDM templates:

Why, in the first, does the "this article's entry" link lead to the discussion in ordinary display mode, while in the second the corresponding link leads to the discussion in edit mode? —Largo Plazo (talk) 15:57, 23 September 2008 (UTC)

Because the servers hadn't caught up with TW at the time you viewed the page in question. It's fine now. In future, purging should fix this, if you don't feel like waiting. Algebraist 16:01, 23 September 2008 (UTC)
Ah, thanks. —Largo Plazo (talk) 16:32, 23 September 2008 (UTC)

Parameter and default changes to Birth/Death date templates

Following discussion at Template talk:Birth date and age, please note there is consensus to add a new parameter and modify the default operation of the following related templates:

Specifically, the new mf=yes parameter would signify which dates are to be expressed in the American-based month-first date format (e.g. September 24, 2008). The default operation of the template would be the day-first format (e.g. 24 September 2008, currently specified by df=yes parameter), which is more common internationally and therefore more appropriate as a default.

However, to minimise format disruption to articles, it is intended that there be a grace period to pre-tag selected articles with the mf=yes parameter. This generally applies to American subjects, or other specific cases where month-first dates are used.

A request was made at WP:BOTREQ to determine if a bot could do most of the pretagging. Some manual inspection and parameter pre-tagging will likely be needed; editors with AWB or other such tools may wish to assist this process. After the pre-tagging is sufficiently complete, the templates will be changed to use the new default and the mf=yes parameter. Dl2000 (talk) 03:29, 24 September 2008 (UTC)

Are you aware of this problem? Just making sure

Twice today I got this error message a lot. I hope it translates okay.

Long error message

(20 pages of generic error message removed, http://svn.wikimedia.org/viewvc/mediawiki/trunk/debs/squid/debian/errors/error.html )

Request: GET http://en.wikipedia.org/wiki/Special:Search?search=oliver+platt, from 208.80.152.41 via sq37.wikimedia.org (squid/2.6.STABLE21) to () Error: ERR_CANNOT_FORWARD, errno (11) Resource temporarily unavailable at Mon, 22 Sep 2008 17:43:45 GMT

Vchimpanzee · talk · contributions · 18:05, 22 September 2008 (UTC)

Added collapse box for very long errmsg. Franamax (talk) 18:13, 22 September 2008 (UTC)
Its normal for it to do this although normally it doesn't last for more than a few minutes. However, this seems to have occurred longer, in the past half hour. Simply south (talk) 18:08, 22 September 2008 (UTC)
See here. Looks like they got it fixed. I figured that the black hole had finally crossed the Atlantic. Got the tin-foil shaped just right for my head too, now I don't need it... Franamax (talk) 18:15, 22 September 2008 (UTC)
The fileserver that holds our NFS home directories and misc files and logs experienced a kernel crash, then turned up some disk errors on reboot. (Possibly two failed drives, which may hose the array.) Ideally this wouldn't disturb production web serving, but various debugging logs were being saved onto this server, and this caused the web servers to hang waiting for NFS to come back up.
We've disabled the internal debug logging for now, and the site's back up and running while we poke at recovering or replacing the fileserver. --brion (talk) 18:37, 22 September 2008 (UTC)

Just so you'll know, it was happening to me about three hours ago too.Vchimpanzee · talk · contributions · 18:45, 22 September 2008 (UTC)

You only need to copy/paste the 2 lines of details at the bottom for such error pages. It is somewhat annoying that the javascript hides the bulk of the text just above these lines, causing them ALL to be copied in several browsers. Anyway, you can see the static text of this error page any time you want by putting &amp; or %26amp%3B in the URI: http://en.wikipedia.org/%26amp%3B .. or in svn --Splarka (rant) 07:16, 23 September 2008 (UTC)
I am constantly experiencing errors when I try to edit Wikipedia pages this week. Talk comments and small edits that should rightly require 5 minutes maximum are dragging out for half an hour due to waiting, error messages, reloads, and "edit conflicts" when the thing goes through but I'm not told that it goes through. This is very aggravating, and it makes me not want to work on Wikipedia. (Even posting this message has been an exercise in aggravation – I've gotten eighteen error messages here!) I'm working on a US public school network, so I don't expect the kind of speed I would get at home – but this is beyond the pale. (Because of my position on the network and limited access to control panels, etc, there's not much I can do to try to isolate the problem on my end, but insofar as simple page loads take a long time – and other websites do not – I don't think the problem is on my end.) Is some kind of work being done on the servers? Can we get an estimated date when these headaches will cease? Thanks in advance for any guidance anyone can give. Scartol • Tok 17:21, 24 September 2008 (UTC)

Deleted image?

Image:COATARMS.gif is... less than functional. The history is there, but the thumb is broken, as is the regular link. Is this an unfortunate victim of that image deletion issue a few weeks back?

The history thumbnail still works, though; if push comes to shove, I could just re-upload that (or find another copy of the image). Still, being the lazy person that I am, I'd rather not have to do anything. :) EVula // talk // // 18:43, 22 September 2008 (UTC)

Yes, looks like the same problem. I found what I believe is the original at [4] and reuploaded it. —Ilmari Karonen (talk) 22:04, 22 September 2008 (UTC)
Awesome, thanks. :) EVula // talk // // 03:55, 23 September 2008 (UTC)
That image has too general a name (and the all-capitals title is not very endearing, either). Waltham, The Duke of 13:00, 23 September 2008 (UTC)
True, that's a common problem with non-free logos: people just download them off websites and upload them using whatever name they had on the original site (like Logo001.gif). We've salted a few of the most obvious titles (such as Image:Logo.gif, Image:Logo.png and Image:Logo.jpg) to prevent that, but there's really no way to guess all the possible variants like Logo2.gif, Logo100x53.jpg, Logo(TransparentBkgrd).jpg or Logo1114677515.jpg in advance (examples courtesy of Special:PrefixIndex/Image:Logo, in case you're curious).
Of course, with free images we're more likely to get filenames straight off the digital camera instead: see e.g. Special:PrefixIndex/Image:DSC. The real solution will be to get MediaWiki to support renaming images. (I've heard most of the necessary code is already in place.) In the mean time, I've come to feel that it's not really worth giving a fuck. It'll probably be fixed in a few years at most; in the mean time, the badly named images don't really cause much harm unless they're so obvious that people keep reuploading unrelated images over them (and even then, trying to do something about it can be a pain). I'd recommend just making sure your image names are descriptive and letting the rest of the mess be for now. —Ilmari Karonen (talk) 14:13, 23 September 2008 (UTC)
I'm not exactly the image-contributor type, but sure; I am rather fussy about titles, anyway (including such small things as underscores, which I find utterly unnecessary). Thanks for the overview. Waltham, The Duke of 14:35, 23 September 2008 (UTC)
For what it's worth, I agree with you. I didn't upload the image; hmwith did, and while she's too busy for wiki-stuff, I'm watching her talk page, hence the thread. EVula // talk // // 18:39, 24 September 2008 (UTC)
does MediaWiki:Titleblacklist apply here? That would seem to be a reasonable solution. -Steve Sanbeg (talk) 19:21, 24 September 2008 (UTC)
I don't think so; this is just an inconvenience, not something we need to put a foot down over. EVula // talk // // 04:05, 25 September 2008 (UTC)

PHP fatal error in ExternalStoreDB.php

Resolved

It took me more than five tries each to make both of these edits: [5] [6]. On some of the unsuccessful attempts, there was the Wikimedia error page with this error message:

PHP fatal error in /usr/local/apache/common-local/php-1.5/includes/ExternalStoreDB.php line 127:
Call to a member function nextSequenceValue() on a non-object

Around that time, the edit rate dropped to zero. Is anyone else getting these? If so, please file a bug because I need the sleep. MER-C 13:14, 24 September 2008 (UTC)

I'm having this issue too. -- Logical Premise Ergo? 14:47, 24 September 2008 (UTC)
The server administrators are looking into this. — Carl (CBM · talk) 15:34, 24 September 2008 (UTC)
Some breakage in recent code updates, has been fixed/rolled back. --brion (talk) 19:51, 24 September 2008 (UTC)

Database error

I am unable to edit the User_talk:08sbailey (http://en.wikipedia.org/wiki/User_talk:08sbailey) page as it continually gives me the text

The database did not find the text of a page that it should have found, named "User talk:08sbailey".

This is usually caused by following an outdated diff or history link to a page that has been deleted.

If this is not the case, you may have found a bug in the software. Please report this, making note of the URL.

WikiDan61ChatMe!ReadMe!! 17:47, 24 September 2008 (UTC)

Yes, there are known problems with the site today. I will pass this error along to the site admins just to be sure they are aware of this particular message. The cause is being investigated, and once it is fixed an explanation will be provided. — Carl (CBM · talk) 17:58, 24 September 2008 (UTC)
One of the devs deleted the broken page; you should be able to recreate it if you want. — Carl (CBM · talk) 20:56, 24 September 2008 (UTC)

Interlanguage Extension

There's a MediaWiki extension called "Interlanguage" that is designed to make the maintenance of interwiki links much easier. For more info see:

I tried it in its test wiki. It worked very well and seemed very useful.

What is the right way to ask to enable it in Wikipedia? I filed a request in MediaZilla:

Further comments are welcome. If you think that it's useful, vote for it or something (although I don't know how important those votes actually are.)

Is there anything else I should do?

Thanks in advance. --Amir E. Aharoni (talk) 06:48, 15 September 2008 (UTC)

I took a look. It seems that technically it is fairly sound, but only fairly. There still are some issues:
  1. The central interlanguage.wikimedia.org wiki has to be deployed and tested.
  2. There still is the grand discussion about how the articles on the central interlanguage wiki should be named. Should they be in English (but some things don't even have an English name), or in the language of the first article on the subject (well, that would make many titles just ??????? for many of us and I can't even cut and paste the names of articles in some of those languages), or should they perhaps simply use numbers (perhaps not as user friendly but at least works technically)?
  3. There are some other issues how to describe on the interlanguage page what it is about. Should for instance the interwiki links on that page be put in small templates with link + short description? Or should we instead (or perhaps as an additional option) use such a template on the Wikipedias? That is, interlanguage link + description, then a bot can copy that to the interlanguage wiki page.
The interwiki system seems to be very new, just some month. So I think it needs to be discussed and thought about a bit more. But it is good that you announced it here Amir, so more people get to know about it and can take a look.
Now please don't discuss these issues here, those were just examples. Go to the pages that Amir linked to above and their talk pages.
--David Göthberg (talk) 11:48, 15 September 2008 (UTC)
Yes, the best place for discussion about the implementation in Wikipedia is meta:Talk:A newer look at the interlanguage link. --Amir E. Aharoni (talk) 16:58, 15 September 2008 (UTC)
As for the cutting and pasting—most likely you are copying the text properly—you're simply not using a Unicode font and/or the correct encoding. UTF-16 is the native encoding of Windows and Java, so it should work in Notepad, MS Word, Open Office, etc. I've run into the problem, though, where some third-party text editors encode text in ANSI by default, and have to be reconfigured to use UTF-16 (or UTF-8 if UTF-16 is not available).SharkD (talk) 17:03, 16 September 2008 (UTC)
You may find this site to be helpful. SharkD (talk) 17:12, 16 September 2008 (UTC)
There is no format called "ANSI". ANSI is the American National Standards Institute, which may or may not endorse any character standards, but if it does I haven't heard of them. As I recall, Notepad calls ISO 8859-1/Latin-1 "ANSI" for some reason, but the names I just gave are its proper and specific ones. I seem to recall that it also calls UTF-16 "Unicode". —Simetrical (talk • contribs) 16:43, 18 September 2008 (UTC)

I suggested something similar to this a year ago in a de-railed mailing list thread, see [7]. — CharlotteWebb 17:09, 16 September 2008 (UTC)

Simetrical: You must be young. We old computer geeks know exactly what SharkD meant with "ANSI". He of course meant the old 7 or 8 bit ASCII code. Which sort of was issued by ANSI. So the two names have always been mixed up all since the very early days of computing.
SharkD: And yes, I use an operating system that doesn't have proper support for UTF-16. My notepad only uses 8 bit ISO 8859-15 (the European version of ISO 8859-1). That is, I use a Swedish version of WinME. (And don't laugh. I got it with this computer when I bought it in spring 2001, and for some odd reason it haven't needed a reinstall yet. Must be a world record for a WinME.) But actually, I can usually copy and paste 16 bit Asian characters as long as I only cut and paste directly between edit windows in Firefox tabs and don't store them in some editor in-between. But I can't copy the characters from a rendered page, even though it is within the same browser. Besides, it is hard to know that one copies the right string when all one sees is "????????". And that problem will remain even with newer OSs for some years to come. (Until the day that all computers come with all the worlds characters. That'll be the day! :)))
So my conclusion is that the interlanguage extension needs to use characters in their page names that all people can handle well in their computers. And that pretty much limits it to a-z and 0-9. Which kind of will suck hard for us that write with other characters. (Well, I as a Swede just use 5 more letters: åäöéü.) So perhaps the "simplest" and mostly neutral solution is to have the interlanguage pages have names made of numbers? Anyway, we shouldn't be discussing this here. It has already been discussed over at the pages we linked to above. But what I mean is that the interlanguage extension perhaps is not ready to be deployed before we have the answers to these issues. Or of course, it might be done the wiki way: Deploy an incomplete "solution" and then learn from that and make a better solution in the next round.
--David Göthberg (talk) 15:46, 21 September 2008 (UTC)
This article at least suggests it should be possible to have Unicode-supporting applications in Win9x/ME. You might want to try one of these text editors and/or one of the freeware fonts found here. The last screenshot, here, in particular shows Unicode fonts working properly in IE5 on Win95 with the help of a utility called MagicWin 98. It's been so long since I have used Win9x I can't remember what I did when I encountered web pages with Japanese characters. If at any time you see a series of boxes instead of question marks, it means your application properly supports Unicode, but your font is missing those particular characters. SharkD (talk) 17:12, 21 September 2008 (UTC)
The overwhelming majority of computers today support Unicode. Pretending we can still use ASCII is simply not an option for the majority of the world's people, who don't use scripts that are compatible with ASCII. Even if everyone is capable of typing ASCII, a considerable number of people are not as comfortable with it as with their native script. Anyone whose computer does not support at least the BMP of Unicode is not going to be able to correctly view all MediaWiki pages, it's as simple as that. We are not going to avoid Unicode in page names, anywhere, for the sake of the tiny minority who can't handle it.

But I think that an extension whose purpose is to improve interlanguage capabilities, but which forces preference to be given to a specific language, is not a great idea. There's probably some better way to design it.

Also, does this have issues with transitivity? For instance, if English has a single article for X and Y and a separate one for Z (not at all a far-fetched scenario), while German has a single article for Y and Z and a separate one for X, can it handle "X and Y" <=> "X", "X and Y" <=> "Y and Z", and "Z" <=> "Y and Z" without assuming that "Z" <=> "X" as well?

Keep in mind I haven't actually looked at the extension.  :) —Simetrical (talk • contribs) 18:13, 21 September 2008 (UTC)

SharkD: Well, it might sound strange but I don't want to have support for all those new things. See, I learnt during my web mastering days in the 1990s that I as a web master should have worse, not better, browsers than my users. Thus I am more likely to make web pages that work for all my visitors. (Since newer browsers usually are backwards compatible.) Besides, I don't read Japanese anyway.
And web masters should of course have several browsers installed so they can test properly. And we Wikipedia editors are indeed web masters. We are the web masters for one of the biggest web sites in the world. And the need to test with old browsers is probably especially true for us who work with things like the templates, CSS and javascript here. (It is so nice to have a good reason to blame when I am too lazy to upgrade! :)
I do my day to day job here with Firefox 2, since I prefer that one. (Well, Firefox 3 isn't available for WinME...) But every now and then I fire up my sucky old IE 5.5 and my Opera 9.02 (not upgraded since 2006) and test the things we work with here. And every now and then I discover and solve some problem. (Sometimes involving our site wide CSS code.) Of course, we need testing with the latest browsers too, but there is no lack of Wikipedia editors who use the latest versions of the different browsers. I am just about to write up a report to the infobox coders here that their CSS is slightly broken in older browsers and how to fix that...
--David Göthberg (talk) 18:30, 21 September 2008 (UTC)
Simetrical: Well, people have suggested a simple solution to the page name problem on the interlanguage server: To use numbers as names. Numbers are truly international and used everywhere. And instead on the page then list the page name in all languages that link to the interlanguage page. I mean, seeing "????????" is much less clear than seeing "2138687". Hey, I can even store that number in an old text editor or send it through email or even type it in by hand on any keyboard in the world. I mean I can check that I copy and paste the right link when I paste "2138687", while I can't check that with "????????". No computer today has all the characters for all the languages. So if we choose any other language than numbers then we have problems. I doubt that a Chinese user will have the Somali or Sami font installed, and so on. And note, if I wonder what a number means it is just one click away to see the name of it in a whole bunch of languages on that interlanguage page. While if I see "????????" and "????????" I won't even notice if they are the same or different names so I won't even know if I need to check them.
For example: With the current interlanguage suggestion if a Chinese user is the first to make an interlanguage link for his "???" article, then the interlanguage article will be named "???". That would mean that on EVERY other language we would have to type {{interlanguage|???|Sandwich|A food item made of two or more slices of leavened bread with one or more layers of a filling.}}. It would be easier to handle for the other 250 projects or so it the link instead was only in numbers like this: {{interlanguage|21138687|Sandwich|A food item ...}}. Note, the "Sandwich|A food item..." part that we fill in here would turn up on the "???" / 21138687 interlanguage page as a link back to the English Wikipedia. On other languages they would fill in their corresponding data so they would only see the number and text in their own language. Here's what you perhaps would see on the interlanguage page:
2138687
  • ka:სენდვიჩი – სენდვიჩი სენდვიჩი სენდვიჩი
  • en:Sandwich – A food item made of two or more slices of leavened bread with one or more layers of a filling.
  • sv:Smörgås – Är en maträtt som består av en brödskiva med ett eller flera pålägg.
  • Other languages...
Would you rather prefer that the page title was ???, seriously?
--David Göthberg (talk) 19:14, 21 September 2008 (UTC)

Problems after recent Firefox v 2.0.0.17 update

I am an old stick-in-the-mud and I am still using FF 2.0. Yesterday it was updated to v2.0.0.17 and since then I am constantly getting script errors when attempting to read any WP page. I am posting this with Google Chrome 0.2.149.30 with no problems at all. Any other FF 2.0 users having the same problems? I know I can update to FF3.0 but until I check whether my favourite extensions are compatible, I am not going to. Thanks. – ukexpat (talk) 14:23, 24 September 2008 (UTC)

Please quote the exact error message. If you see only CSS errors/warnings, it's not a big deal, just ignore them. —AlexSm 15:41, 24 September 2008 (UTC)
Nope it's not a CSS warning, it's the "Unresponsive script" message box with the three buttons - Stop, Debug, Continue. I have to hit stop otherwise it keeps popping up. – ukexpat (talk) 16:16, 24 September 2008 (UTC)
Have you tried blanking your monobook.js to see if anything there's causing it? Algebraist 16:19, 24 September 2008 (UTC)
Just tried that - still the same problem. – ukexpat (talk) 00:25, 25 September 2008 (UTC)
Does it happen on Special:Blankpage ? Does it happen hen you are logged out? Does it happen on another language or sister project, or project? --Splarka (rant) 07:34, 25 September 2008 (UTC)
It was happening on all WP and sister project pages, but I have figured it out. I have WikEd installed in my monobook.js and also as a Greasemonkey script so I can use it on any wiki that uses the Wikimedia software. Disabling the GM script fixes the problem. I guess the latest Firefox update caused or highlighted a script conflict or similar issue. Thanks for all your suggestions, normal service will now be resumed. – ukexpat (talk) 13:25, 25 September 2008 (UTC)

Wikipedia has a problem, all right

Resolved

I am constantly experiencing errors when I try to edit Wikipedia pages this week. Talk comments and small edits that should rightly require 5 minutes maximum are dragging out for half an hour due to waiting, error messages, reloads, and "edit conflicts" when the thing goes through but I'm not told that it goes through. This is very aggravating, and it makes me not want to work on Wikipedia. (Even posting this message has been an exercise in aggravation – I've gotten more than twenty "Wikipedia has a problem" error messages here!)

I'm working on a US public school district's network, so I don't expect the kind of speed I would get at home – but this is beyond the pale. (Because of my position on the network and limited access to control panels, etc, there's not much I can do to try to isolate the problem on my end, but insofar as simple page loads take a long time – and other websites do not – I don't think the problem is on my end.) Is some kind of work being done on the servers? Can we get an estimated date when these headaches will cease? Thanks in advance for any guidance anyone can give. Scartol • Tok 17:24, 24 September 2008 (UTC)

Curse, foam, rage, whine and scream. This is beyond absurd, and it's been going on for several days. At least some explanation of why this is so severe and ongoing might help lower the frustration level. SandyGeorgia (Talk) 17:29, 24 September 2008 (UTC)
Until just a few minutes ago I didn't have these issues. Now I can see what's got everyone hot. -Jéské (v^_^v Kacheek!) 17:31, 24 September 2008 (UTC)
I'm getting this problem on Wikisource and Meta as well. It is very frustrating. EVula // talk // // 17:32, 24 September 2008 (UTC)


Re everyone: there have been several technical issues this week, and it's frustrating for everyone. However, I think the site has a pretty reasonable record for uptime overall; the last few couple are an exception rather than the norm. At the moment, it appears just to be unfortunate coincidence to have several problems in a row. — Carl (CBM · talk) 17:36, 24 September 2008 (UTC)


It's happening all the time now, and has been on and off for the last few days. Errors like this:

Sorry! This site is experiencing technical difficulties.

Try waiting a few minutes and reloading.

(Can't contact the database server: Unknown error (10.0.2.103))

Sort it out and quick. Lugnuts (talk) 17:38, 24 September 2008 (UTC)

Don't worry, it will be solved before the deadline. GRBerry 17:43, 24 September 2008 (UTC)
Carl, thanks for your feedback. I'm sure it is a lousy coincidence with several technical problems all lined up making each one exponentially worse. In a way, the general reliability of the Wikipedia/-media servers makes this all so much more frustrating. It's like a sudden loss of Internet access – we don't realize how nice the smooth road is until we reach an unpaved patch.
However, it's also very frustrating for us clueless noobs to be so far in the dark with regard to these issues. I'd love to see some sort of blog about the server status, so we can check and see if there's an acknowledged problem, and – if so – what's being done to address it. In my experience, tech folks are wizards when it comes to fixing insanely complicated issues but not so good at keeping lesser beings informed about their work. (This is a good spot for me to say that I can't imagine the gargantuan task of keeping the Wiki* servers active. Many thanks as always for those toiling in the salt mines.)
When I first saw your edit summary, Sandy, I was worried that maybe the "status site" I suggested above might already exist, and I was being chewed out for not checking it first. =) Glad to see I'm not the only one gnashing his teeth into a gruesome powder. Scartol • Tok 17:48, 24 September 2008 (UTC)

The philosophy that "it's all right because there is no deadline" assumes that editors' time and effort are of no value and that they will go right on editing when their work gets eaten by the malfunctioning machine. Bland reassurances about "coincidences" or "pretty good uptime" just do not cut it. I would like to know what malfunction, glitch, operator error or resource shortage is responsible for this, and what is being done to fix it. Supposedly there are discussions on IRC, per the error message. Someone, somewhere, surely has a clue as to why the system is broken. Lately it has been pretty glitchy in general. Oh well, there is real world stuff to do. Edison (talk) 17:53, 24 September 2008 (UTC)

I've noticed too that it lags a little bit, but other than that crash a couple of days ago, I'm humming along just fine. hbdragon88 (talk) 17:55, 24 September 2008 (UTC)


Again, re everybody: the site admins tend to focus on figuring out and fixing the problem first, and reporting second. The problems from Sep 22 were reported in this email. As you can see there, once the immediate problem is fixed the admins do consider how to avoid the same problem in the future. They have already implemented the UDP system that is mentioned in the email.

The exact reason for today's problems isn't known yet; the devs have been investigating for some time. Based on the server admin log, it appears to have something to do with processes that are blocking other processes from updating the database when an edit is recorded. Of course, once the exact problem is found, it will be fixed, and we can expect another email explaining what happened. — Carl (CBM · talk) 17:56, 24 September 2008 (UTC)

Just tell us "The rasufractor canullated due to a moth caught between the contacts of a relay. The contacts were burnished and a screen was placed on the window to discourage bugs in the future." Or any other technospeak which might have ben lifted from Startrek, along with some assurance that it will not be a continuing problem or one that only gets worse with time. Thanks. Edison (talk) 18:16, 24 September 2008 (UTC)
Cap'n she canna take much more....  – ukexpat (talk) 18:30, 24 September 2008 (UTC)
I've also experienced this problem over on Wikiversity, Wikibooks and other's it seems fine now but a few days ago I tried to access my account but couldn't for several hours due to the error message. Dark Mage 18:45, 24 September 2008 (UTC)

We've tracked down today's problems to a combination of a couple of things:

  1. There've been ongoing database locking issues with the site statistics updates -- these would all block on each other, making page saves very slow at times
  2. ... which held open database connections, causing the text storage servers to start locking out new connections ...
  3. ... which exacerbated problems with the failover behavior of recent changes to the storage and load balancing code.

The code changes have been rolled back, fixing the slow site load behavior. (doing this correctly unfortunately was a bit painful, as we had to restore the broken code for a while in order to pick out what was going on enough to fully revert it again.)

Domas believes the main culprit on the database locking is actually an issue with our mail server -- some actions (such as creation of new accounts) would involve both mail and updates to the site statistics table. With overload to the mail server, and a very simple local mail client called from MediaWiki, the outgoing mail would sometimes hang, while the transaction was still open, causing the locks, causing other updates to stall.

As a temporary measure I've disabled the site stats updates, fixing the failures on page save. (They'll need to be re-updated after we've totally resolved it.)

We're looking at the way the mail servers are set up to see if we can ensure that internal connections don't stall the way they were; we should also be able to rearrange the transactions so that things are committed before the mail goes out! --brion (talk) 19:45, 24 September 2008 (UTC)

Yes, I've noticed that things are a lot faster at this point. Infinite thanks to everyone working to make the needed repairs. You rock. Scartol • Tok 19:54, 24 September 2008 (UTC)
Hmm, out of curiosity these users mentioned here and here are being created every few minutes over on Wikiversity - apparently taking part in some project scheme, would they have caused a strain in the servers - some of us are wondering what and why they are being created. Dark Mage 19:59, 24 September 2008 (UTC)
I'll bet it's some kind of ungodly probing into the dark mysteries of the universe, like in The Mist. This may only be the beginning of our troubles. =) Scartol • Tok 00:45, 25 September 2008 (UTC)
Or a weird Star Trek-type pure energy alien life form - faced with that you are bound to lose a few redshirts along the way! – ukexpat (talk) 13:28, 25 September 2008 (UTC)

I have a rather strange problem with Special:WhatLinksHere/2008 Christmas special (Doctor Who)... There are more then 50 links pointing to this redirect, even though all links have been changed since this page was moved to The Next Doctor two days ago. Most articles do not actually link to the redirect, nor do any templates on these pages. Is the database slow to catch up or am I overlooking something? EdokterTalk 20:12, 24 September 2008 (UTC)

The job queue is getting longer, and so the servers are getting behindhand. You can null-edit pages to remove them from whatlinkshere, but it'll slow the servers further and there's not much point. Algebraist 21:00, 24 September 2008 (UTC)
I tried editing and purging the articles, but they somehow stay on the WhatLinksHere page. Pages that do have an actual link, do get removed from the list after I remove the link from that page. EdokterTalk 22:51, 24 September 2008 (UTC)
Purging isn't enough. As Algebraist said, you can null-edit pages which is something different. Can you give an example of an article you have edited without it being removed from the WhatLinksHere page? PrimeHunter (talk) 23:07, 24 September 2008 (UTC)
OK, that does seem to work, though it's not really a solution. Is this something that will clear up by itself? EdokterTalk 10:19, 25 September 2008 (UTC)
Yes. Algebraist 10:38, 25 September 2008 (UTC)

The automated redirect fixer hasn't fixed any redirects for over 48 hours now. Is this related to the other ongoing technical problems, or perhaps yet another problem? --Russ (talk) 09:57, 25 September 2008 (UTC)

In Wikipedia:Suspected sock puppets/Absad, it appears that the subpages used are user talk subpages. These subpages are not detectible from the subpages link below the User contributions page. For example, from Special:Contributions/Absad, selecting Absad User subpages shows no subpages. However, Absad User talk subpages shows subpages. In the box containing "0tat: Subpages · User rights · Edit and action count · Interiot · Edit summary usage · Images uploaded · Articles created · SUL accounts · Global contribs", is it possible to divide Subpages into "User subpages" and "User talk subpages"? (I don't know where this box comes from, else I would post there.) Thanks. -- Suntag 16:50, 25 September 2008 (UTC)

Found it. See MediaWiki:Sp-contributions-footer talk page for follow up. -- Suntag 16:53, 25 September 2008 (UTC)

Edit message

As far as I remember it's possible to create a MediaWiki page which causes a message to come up before the edit box when editing a page. See for example WP:ACC and most of the MediaWiki talk namespace. What's the syntax for that? Stifle (talk) 18:13, 25 September 2008 (UTC)

See WP:Editnotice for explanation. —AlexSm 18:16, 25 September 2008 (UTC)

Classic skin and main page

Recently when I start a wikipedia browsing session at the main page, I have not been logged in. However if I jump to another page I am logged in. When I go back to the main page I then am still logged in. What happens with the first visit of a session that the log-in cookie gets ignored? Is it fixable? (PC XP MSIE7 Classic skin) -- SGBailey (talk) 22:42, 25 September 2008 (UTC)

I am wondering if there is a modifyPorletLink or "foo"PortletLink or something, and if there is any list of these "special" functions. Thanks...
Ajl772 (talk) 22:46, 25 September 2008 (UTC)

You shouldn't need pre-written functions for simple portletlink modifications. What do you want to do? Algebraist 23:18, 25 September 2008 (UTC)
well, I think I figured it out, so never mind, but if you have any suggestions, feel free to mention them.
Ajl772 (talk) 23:32, 25 September 2008 (UTC)
Since you asked, you could make it a bit shorter by using
document.getElementById("pt-mytalk").firstChild.innerHTML = "Talk";
and so on. Doesn't actually achieve anything, mind. Algebraist 23:39, 25 September 2008 (UTC)
For you, perhaps, but if you imported the script into your "monobook" or whatever you would see what i did achieve or I could explain:
Ajl772 | My Talk | My Preferences | My Watchlist | My Contributions | online | busy | around | offline | Log out is just too long, and it gets shortened down to Ajl772 | talk | prefs | watch | contribs | on | bzy | arnd | off | Log out (the on,bzy... is from Xenocidic's "statusChanger2")
Ajl772 (talk) 23:52, 25 September 2008 (UTC)
Sorry, I meant my change doesn't achieve anything, it just makes the code look tidier. Algebraist 23:54, 25 September 2008 (UTC)
That's okay. Usually though, shorter code == less load/processing time.
Ajl772 (talk) 23:56, 25 September 2008 (UTC)

Accessibility of NavFrame

I raised an issue at Wikipedia talk:NavFrame:

Using Firefox 3 with Javascript disabled by the very widely used "NoScript" extension, I find that the "collapsible" sections on {{Infobox Ship Begin/doc}} are already collapsed with no option to expand them.

and AlexSm has replied to point out that he raised the issue on MediaWiki talk:Common.js a year ago but nothing has been done. I view this as a serious failing; can someone assist in finding a remedy, please? Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 18:51, 22 September 2008 (UTC)

I'm no expert, but isn't the entire point of NoScript to block JavaScript on any website not allowed? Calvin 1998 (t·c) 05:21, 23 September 2008 (UTC)
That is the entire point, but "collapsed" sections should be "uncollapsed" by default. Some people do not have javascript at all (and many more have it disabled for security reasons) but they should still be able to see the entire infobox/navbox/whatever content. — CharlotteWebb 13:53, 23 September 2008 (UTC)

Note that while the JS code has been recently fixed by Ilmari Karonen, I suspect that style="display:none" is yet to be fixed in some templates, for example in Template:Hidden. —AlexSm 20:46, 24 September 2008 (UTC)

Yes, someone still needs to go around and fix all those hardcoded style attributes. —Ilmari Karonen (talk) 20:28, 26 September 2008 (UTC)

Problems loading up Wikipedia

Since mid-to-late August, I have been having problems loading up Wikipedia articles in both the latest editions of Firefox and Internet Explorer (Windows XP). It usually gets as far as showing the article's title in my browser's title bar and sometimes, if I'm lucky, perhaps the first few lines of the article. Sometimes the page does eventually load up, sometimes it doesn't. I am on a broadband connection and other webpages load up fine. There have been times, especially this week, when the Wikipedia (Wiktionary & Wikimedia) articles would load as normal.

Things I have tried:

  • Called my local ISP and they are able to access Wikipedia with no problems. Claim (as always) that problem is on my end.
  • Used a laptop (older and wonkier than this desktop) that's connected via the same WiFi network and successfully connected to Wikimedia projects.
  • Temporarily disabled antivirus and spyware software
  • On this desktop, opened a connection from my broadband connection to AOL and successfully connected to Wikipedia using the AOL client. They essentially used the same browser (MSIE 7.0)!
  • Emptied temporary internet files and cookies
  • Tracert to Wikipedia and everything is under 100ms, more or less.
  • Did a Google Image search for "site:wikimedia.org" and some images load, others either (usually larger ones) do not load or do so slowly.

I'm running out of ideas here. I am currently connected to Wikipedia via the Secure connection (secure.wikimedia.org) and everything loads normally and quickly, but I want to connect normally. No major changes have been made to my computer, I should add. Could anyone help? Thanks in advance --Chris S. (talk) 00:47, 24 September 2008 (UTC)

Try reading this: http://www.netheaven.com/pmtu.html -- Tim Starling (talk) 03:51, 24 September 2008 (UTC)
Hello and thanks for your reply. I read and tried the remedy suggested. However, there are no registry entries for PMTUD that I could edit. Does anyone else have any other ideas? This problem still persists. --Chris S. (talk) 04:29, 27 September 2008 (UTC)

Quick JS question

We have a precoded function to test whether an object has a certain class ('hasClass'). How would I, with or without this function, test to see if an object is contained within an object of a particular class? Happymelon 13:58, 26 September 2008 (UTC)

The only way I know is to go up the DOM tree (with parentNode) and checking each node on the way. —AlexSm 15:01, 26 September 2008 (UTC)

Reflist

When doing a section edit, a preview of any references does not get displayed. Is there any way of getting a checkbox in the vicinity of the "Minor edit" checkbox to force a reflist display at the end of such a preview? -- SGBailey (talk) 14:30, 26 September 2008 (UTC)

You can manually add display of references as described at Wikipedia:Footnotes#Previewing a single section edit. If a software change is made then I would prefer to display the references by default or always. See also Wikipedia:Village pump (proposals)/Archive 6#Footnotes in preview (I haven't tried this). PrimeHunter (talk) 14:47, 26 September 2008 (UTC)

¿¡Avril Lavigne's birthday!?

Instead of the usual "Help us by donating today!" message, I've found this text in the article about Inkatha Freedom Party: "The Wikipedia Community would like to wish Avril Lavigne a happy birthday!" How the hell could this happen??? --NaBUru38 (talk) 22:54, 26 September 2008 (UTC)

It was vandalism to the transcluded template {{South Africa topics}}. It has been reverted and the vandal account has been blocked. This is unfortunately common. If the problem goes away when you purge the page then it's a signal that the template vandalism has been fixed. PrimeHunter (talk) 23:12, 26 September 2008 (UTC)

Current squad / navbox

Please can someone help me to work out a way of combining the current squad and the navbox on Association football pages, I have asked a couple of experienced editors and at Wikiproject football but apart from a few positive comments and an unsuccessful attempt to help I got nowhere.

Using Arsenal de Sarandí as an example the proposal is that since the navbox and current squad display almost exactly the same information, could somebody modify the navbox so that it displays something similar to the current squad when coded like {{Arsenal de Sarandí squad|team article=yes}} along with the [v] [d] [e] functions to allow the transcluded squad to be easily modified from the team article and a "last updated" header. There are several reasons this would be beneficial I will outline some of them:

  1. Keeping similar information in the same place and avoiding repetition makes sense for obvious reasons.
  2. Casual editors often edit one or the other, rarely both, meaning that the 2 ways of displaying the squads often contain contradictory information, combining the functions would make this kind of contradiction impossible.
  3. Editors like myself who spend/waste our time updating player and team articles during the transfer window currently have to make at least 5 edits per player transfer (1. Update player article 2. Remove player from former team article 3. Remove player from former team navbox 4. Place player into new team article 5. Place player into new team navbox) combining the current squad and the navbox would reduce this to three edits, reducing transfer window workload by 40%, (which is a lot of edits considering in just one league (Primera División Argentina) the clubs amassed nearly 300 transfers in the most recent bi-annual transfer frenzy).
  4. Introducing a "last updated" parameter would make it easier to search out and update outdated squad information.

If anyone could help me by amending Template:Football squad or making an alternative I would be immensely grateful. If nobody can help perhaps someone could point me towards someone who can? EP 22:17, 25 September 2008 (UTC)

Wikipedia:Requested templates is a place to ask experts to help with template modifications as well as with new templates. -- John Broughton (♫♫) 00:06, 28 September 2008 (UTC)

How Google handles page moves on Wikipedia

This is a question for Google experts about how Google treats page moves on Wikipedia. When moving a page to a new name, is it necessary to delete the redirect at the old name to persuade Google to index the page under the new name? (Presumably after some time has passed, we could re-create old redirect, and then Google might continue to index the real page, and ignore the redirect as a duplicate.) That is, as long as the original page name exists as a redirect, will Google continue to see the redirect page as the "real" page, and ignore the new page as a duplicate? (That is the behavior I think I am seeing from Google.)

Back around 21 February 2008, at my suggestion, Sbowers3 moved several FAQ pages to put them into a single hierarchy of subpages under the Wikipedia:FAQ top-level page. This was to allow {{Google custom}} to search all the FAQ pages. Our assumption at the time was that Google would index the FAQ pages under what we as humans recognize as their "real" names (WP:FAQ/Business, WP:FAQ/Editing, etc.). I guess nobody must have checked to see if Google really would do that, because recently Commander Keane called my attention to a problem with the search on the FAQ pages. I checked, and lo, the search that I assumed would work, doesn't really. The following example searches illustrate the problem. The terms "business" and "FAQ" appear on at least two FAQ pages, so we would expect to find them:

Type this To get this What it produces, or searches for
{{Google custom|en.wikipedia.org/wiki/Wikipedia:FAQ|business FAQ|Search the FAQ for: business FAQ}} Search the FAQ for: business FAQ Search WP:FAQ and (hopefully) its subpages for: business FAQ
{{Google custom|en.wikipedia.org/wiki/Wikipedia:|business FAQ|Search the Wikipedia: namespace for: business FAQ}} Search the Wikipedia: namespace for: business FAQ Search the Wikipedia: namespace for: business FAQ
{{Google custom|en.wikipedia.org|business FAQ|Search the English Wikipedia for: business FAQ}} Search the English Wikipedia for: business FAQ Search the English Wikipedia for: business FAQ

The above three searches are progressively more general. The first search fails to find the Wikipedia:FAQ/Business subpage; it only finds the top-level Wikipedia:FAQ page. The last two searches find the Wikipedia:FAQ/Business subpage as the first result. However, while they show Wikipedia:FAQ/Business as the page title, they actually link to the old page name which is now a redirect: Wikipedia:Business' FAQ. This effect is subtle, and I probably never noticed it before because I usually try searching on whole namespaces, or on subpage trees that have been stable against page moves. When the search is on a large-enough subset of Wikipedia to find the page one wants, one may not even notice that the found page is actually a redirect, or the implications of this may not sink in.

More technical details, if necessary:

--Teratornis (talk) 23:48, 25 September 2008 (UTC)

Deleting redirects to force Google to do this is a bad idea. One alternative could be for the software to automatically noindex redirects. Though, it seems to me that this is Google's problem to fix. Wikipedia is a large enough site that they should be able to tweak their system to index our articles by their current title rather than a redirect title. --- RockMFR 00:10, 26 September 2008 (UTC)
Google is bigger. SharkD (talk) 00:11, 26 September 2008 (UTC)
I think RockMFR meant that Wikipedia is a large enough site, not that WP is bigger...
Ajl772 (talk) 00:14, 26 September 2008 (UTC)
"Large enough"? With 14.6 million pages? You can't be serious. Waltham, The Duke of 01:57, 26 September 2008 (UTC)
Wikipedia "redirects" are a special MediaWiki feature and not URL redirection. If the only difference between the content of the url's http://en.wikipedia.org/wiki/Wikipedia:Editing_FAQ and http://en.wikipedia.org/wiki/Wikipedia:FAQ/Editing is a line saying "(Redirected from Wikipedia:Editing FAQ)" then I don't think we should expect Google to change their software to recognize this. http://www.google.com/support/webmasters/bin/answer.py?answer=66359 says:
Google tries hard to index and show pages with distinct information. This filtering means, for instance, that if your site has a "regular" and "printer" version of each article, and neither of these is blocked in robots.txt or with a noindex meta tag, we'll choose one of them to list.
I don't blame Google for not always making the choice we would prefer. Having our software put a noindex tag on a page with "Redirected from ..." sounds good to me (assuming search engines discover and index the target page). PrimeHunter (talk) 14:32, 26 September 2008 (UTC)
Google's choice is heavily influenced by which version has the most internal and external links to it. If people frequently link to Wikipedia:Editing FAQ but rarely to Wikipedia:FAQ/Editing, then Google is likely to give priority to the former, even though the "real" page might be at the later. This is also a reason not to use NOINDEX on redirects in general since the weight that Google gives to links may apply to a specific target, and if you hide that target because you don't like its name then that content may become harder to find. Whether using NOINDEX may be useful in specific cases though is an open question that would probably have to be determined by experimentation. Dragons flight (talk) 14:44, 26 September 2008 (UTC)
Is it possible without developer involvement to noindex http://en.wikipedia.org/wiki/Wikipedia:Editing_FAQ and other redirects to Wikipedia:FAQ/Editing without noindexing http://en.wikipedia.org/wiki/Wikipedia:FAQ/Editing? For example, can code on Wikipedia:FAQ/Editing detect whether it was reached by a redirect and only place __NOINDEX__ in that case? PrimeHunter (talk) 15:40, 26 September 2008 (UTC)
I just had an idea - how about replacing redirects such as Wikipedia:Editing FAQ and Wikipedia:Business' FAQ with soft redirects? Then the redirects should become distinct from the content pages to Google, and presumably Google will index them separately. Functionality to Wikipedia users would be only slightly worse (they would have to click once more to follow a redirect). If this solution works, it has the advantage of not requiring administrator or developer assistance. In the meantime, I will look for templates that may be currently adding many links to the redirects, and update them to link to the canonical FAQ pages. An example was {{Db-spam-notice}} (diff). --Teratornis (talk) 18:45, 26 September 2008 (UTC)

Google will do some advanced searching of websites by using "inurl" and "intitle". For example; the search link below will find all the pages with "Wikipedia:FAQ" in the URL. Add search terms such as "business" to find more info.

Here is how to search the titles of those "Wikipedia:FAQ" pages:

The above link pulls up the business FAQ page. One can add additional search terms to these types of searches too. Add the additional search terms before the "inurl" and "intitle" stuff to avoid problems.

Why can't Wikipedia's search engine do any of this? Please see Special:PrefixIndex. Special:PrefixIndex/Wikipedia:FAQ pulls up all the subpages of Wikipedia:FAQ. It will not search those pages though. It would be nice if Special:Search (the main Wikipedia search engine) did that. It would also be nice if it was a right-clickable link on the sidebar of every Wikipedia page, and if it did more of what people actually want in the way of search. Such as searching Wikipedia pages that start with specific names such as "Wikipedia:FAQ". But of course the "Just Say No" admins blocked the sidebar link idea. Why? Who knows. See MediaWiki_talk:Sidebar#Advanced search is hard to find. It is not for us mere editors to question the wisdom of the Great Wizards behind the curtain of Oz. ;) And there is little money to develop the many frequently requested enhancements (such as specialized searches) due to the other "Just Say No" ideologues who oppose optional advertising for those who choose to allow it. See Wikipedia:Advertising. --Timeshifter (talk) 18:28, 26 September 2008 (UTC)

I had thought about using Google's standard search with inurl and intitle options. If all else fails we can do that. That has the ergonomic disadvantage of linking to a search form pre-filled with these search options, which the user must then type text alongside of. That's not a fatal flaw, but it may be more confusing to non-technical users than presenting them with the clean Google custom search form, which keeps the search-limiting parameters outside the input box for the user's search keywords. Thus I would prefer to use the Google custom search if possible. Speaking of which, do any of our Google experts know where Google hides its technical documentation on its custom search? I only learned how to use the options I know about (edit: those options are: domains and sitesearch) by seeing someone else's use of them. --Teratornis (talk) 18:47, 26 September 2008 (UTC)
See google.com and the advanced search link there. It is http://www.google.com/advanced_search and there is an "advanced search tips" link on that page for further info. There are more links and one can go very deep into search methods. I forgot about "site". But it does not seem to work in finding the business FAQ.
See
http://www.google.com/search?q=site:wikipedia.org/wiki/Wikipedia:FAQ
http://www.google.com/search?q=site:wikipedia.org/wiki/Wikipedia:FAQ+intitle:business
"site" seems to work best with whole domains. See:
http://www.google.com/search?q=site:wikipedia.org+Wikipedia:FAQ+business
It may be easier to create a subpage box, and tell people to drill down as necessary to find info. Sometimes searching doesn't always work. See:
Wikipedia:Graphic Lab/Resources
Wikipedia:Graphic Lab/Resource list box
Wikipedia:Graphics Lab/Resources

Stale/Archival

--Timeshifter (talk) 19:29, 26 September 2008 (UTC)

Google's advanced search options seem to be distinct from its custom search, which has a different URL:
It seems Google has added more documentation than I saw the last time I looked (which might have been last year):
In particular, there seems to be an option to search specific Web pages now, so that might be a way to cobble up a custom search that would search on the Wikipedia FAQ pages (or any other collection of pages). If I see a practical way to do it, I might try to extend the {{Google custom}} template to take an optional list of Wikipedia page titles. The Google Co-op article mentions the Custom Search Engine (CSE), which may be what {{Google custom}} uses. Since I started the {{Google custom}} template, I should probably read the friendly manuals to get some idea of what else the template could do. --Teratornis (talk) 02:59, 27 September 2008 (UTC)
It looks like the {{Google custom}} searches listed higher up in this discussion are the "site:" searches I pointed out also. For some reason they don't work as well as some of the "inurl" searches. Some variations of the "inurl" searches don't work that well either. I don't know why. It is all so inconsistent. I really like some of the featured custom searches linked from http://www.google.com/coop/cse/ and http://www.google.com/coop/cse/examples/GooglePicks
If one could create a custom Google searchform that only searched within subpages of a particular Wikipedia page that would be great. Maybe there might be a way to incorporate some of that functionality into {{Google custom}}. --Timeshifter (talk) 04:40, 27 September 2008 (UTC)
Most of the examples in Template:Google custom#How to search subpage trees within Wikipedia do work. For example, the {{Google help desk}} template has worked well for me dozens of times. Subpage trees are searchable with {{Google custom}} except for the following cases I know about so far:
  • Google actually indexes redirects that are outside the subpage tree.
  • The subpage tree is in the (article) Talk: namespace. (Google does not index Wikipedia's article talk pages, but it does index other namespace talk pages.)
  • The search involves pages that Wikipedia's robots.txt file excludes. (This is annoying, for example, when one wants to find an articles for deletion discussion. robots.txt excludes those.)
--Teratornis (talk) 06:16, 27 September 2008 (UTC)
I was thinking that Google's custom searching could be incorporated into our Template:Google custom somehow. Google's custom searching is more specific than its other search methods. Its other search methods are very inconsistent. It is not just the redirects and robots.txt exclusions that are a problem. Try the "site:" or "inurl:" searches with different parts of a URL, and see what I mean. I have a website and a mirror of it. Google is very inconsistent in what it indexes and pulls up in "site:" or "inurl:" searches of that website.
Since Google does not index certain pages due to various reasons, then maybe the Wikipedia search engine could be developed further to do searches of pages found with Special:PrefixIndex. --Timeshifter (talk) 04:55, 28 September 2008 (UTC)

Always available to help

I am relatively new to Wikipedia/MediaWiki and at times I am asked to help out and edit certain pages. I research and attempt to do the best I can however, realize that I am still learning and most importantly, that I am a 1 man operation working out of my home. I have no one to go to to get support so I have to work with what I have. Now , im thinking that I will be criticized for making this general note on a technical site but please understand how very busy I am (not to say that you are not) and sometimes the time it takes to say just one thing takes an hour to find where to put the message. Thank you --Ptw007 (talk) 13:33, 26 September 2008 (UTC)

What are you referring to? Special:Contributions/Ptw007 shows your only edits are adding a '§' character and two signatures in meaningless places with meaningless edit summaries. PrimeHunter (talk) 13:53, 26 September 2008 (UTC)
Wikipedia has huge amounts of documentation on how to edit. See the editor's index to Wikipedia. -- John Broughton (♫♫) 00:02, 28 September 2008 (UTC)

A strange rendering problem with an SVG

Image:Nuvola apps package graphics.svg, the SVG version of Image:Nuvola apps package graphics.png, seems to render strangely when used in templates (e.g. here). The paint palette appears dark brown for some reason, when the SVG appears fine on the PNG's page. Any idea what is causing this discrepancy? It Is Me Here (talk) 16:25, 26 September 2008 (UTC)

I also notice that the red and yellow paints are not showing up unless the image is scaled up. The same happens using The Gimp or the ImageMagick convert command to convert the image to a png.
If I edit the SVG file to remove the filter effects from the image, both The Gimp and ImageMagick display the image correctly. Apparently at a certain point the SVG renderer(s) used by these programs stops displaying a blurred object at all; this is apparently a bug in those renderers. Anomie 16:56, 26 September 2008 (UTC)
I see - is there any way you could fix Image:Nuvola apps package graphics.svg to make it display correctly? It Is Me Here (talk) 17:31, 26 September 2008 (UTC)
Done, although I can't promise anyone on Commons won't undo it. Anomie 23:19, 26 September 2008 (UTC)
Thank you very much! It Is Me Here (talk) 07:35, 27 September 2008 (UTC)

google earth wikipedia links. Out of Date?

I was browsing the Google Earth, and clicked on a Wikipedia link which took me here> [8]

Which told me the image file does not exist. I did find a suitable image file thus> [9]

But my point is: the link as displayed in Google-Earth is WRONG. Maybe the article has been edited, I do not know. (But if so this could be one of hundreds of out of date Wikipedia links on Google-Earth.)

Is there not a policy in place, whereby revisions to Wikipedia articles should, where applicable, include revisions to Google-Earth links?

Who is it that put all these Wikipedia links into Google-Earth in the first place? And should not THEY be responsible for the revision and updating of such links? -Or better still have the process automated?

It really does defeat the object of having these two wonderful landmarks and paragons of the internet (Google-Earth and Wikipedia) working in tandem, (of itself a wonderful idea) if it is not maintained.

However I am not criticising; if I were capable of fixing the problem myself I would do so. But I felt the need to draw it to (someone's) attention.

--Macthefork (talk) 18:02, 26 September 2008 (UTC)

I believe this is done entirely by Google. Mr.Z-man 19:44, 26 September 2008 (UTC)
Yes, see the Google Earth Geographic Web Layer FAQ, which includes information about how Wikipedia and Google Earth interact. Graham87 08:13, 27 September 2008 (UTC)

Files vs. images

Would it be all right if we changed MediaWiki:Nstab-image to the MediaWiki default of "File", like the Commons has? It's well-known that the "Image" namespace contains sound and video too. —Remember the dot (talk) 03:19, 27 September 2008 (UTC)

No objection from me. :) --brion (talk) 18:00, 27 September 2008 (UTC)
Alright, thanks. I did it - hopefully this won't be too controversial. —Remember the dot (talk) 02:48, 28 September 2008 (UTC)

My talk page

Could someone make the section [edit] links reappear on my talk page? I think the problem is hidden somewhere at User talk:Editorofthewiki/Header. Your friend Eddy O. D. Wiki[citation needed] 13:51, 27 September 2008 (UTC)

Done. The problem was in the {{User:Cuyler91093/Box-header}} template you were using, it requires the parameter EDIT=1 to not remove the section edit links. Anomie 16:40, 27 September 2008 (UTC)
It appears User talk:Editorofthewiki transcludes User talk:Editorofthewiki/Header which transcludes User:Cuyler91093/Box-header which redirects to User:The Obento Musubi/Box-header which conditionally adds __NOEDITSECTION__. I don't think there is a way to override __NOEDITSECTION__ so you must avoid transcluding it. I haven't examined how to best do that or whether you transclude it in other ways. PrimeHunter (talk) 16:44, 27 September 2008 (UTC)

Image upload messed up

I revised Image:Arthropod cuticle 01.png and uploaded - the thumbnails are OK, but the main pic has the new height but the old content, so it's stretched. -- Philcha (talk) 14:11, 27 September 2008 (UTC)

Did you clearing your cache? If that fails, you could also try purging the image page. Anomie 16:26, 27 September 2008 (UTC)

help with footnotes please

I tried to add citations to my additions to the auteur theory article, but the footnotes don't appear at the bottom of the page, and I can't figure out what I'm doing wrong. Minaker (talk) 20:37, 27 September 2008 (UTC)

You needed a reference section. I added it to the article. Garion96 (talk) 20:54, 27 September 2008 (UTC)

Thanks, Garion. Minaker (talk) 21:34, 27 September 2008 (UTC)

Javascript Error

I am getting the following error:

Line: 877
Char: 3
Error: Object expected
Code: 0
URL: http://en.wikipedia.org/wiki/ *whatever*

I believe it is in connection with User:Ajl772/statusChanger2.js but I don't see anything in there that would cause a problem... Is there something that I didn't see/think-of that is screwing up?
Ajl772 21:40, 27 September 2008 (UTC)

And then at other times (randomly, I can't seem to make it happen every time) it says: Line: 43
Char: 3
Error: Object expected
Code: 0
URL: http://en.wikipedia.org/wiki/ *whatever*

Any help would be much appreciated.
Ajl772 21:51, 27 September 2008 (UTC)

I'm probably mistaken, but it might be a problem with the modifyForm variable; try placing var modifyForm = 0; on a new line below var statis = "";. —Animum (talk) 22:43, 27 September 2008 (UTC)
(edit conflict)
Nevermind, I've fixed it. Thanks for the help anyway...
Ajl772 22:47, 27 September 2008 (UTC)

You know, we did block that status bot for a reason... — Werdna • talk 04:10, 28 September 2008 (UTC)

2,563,508 articles in English: for how long?

For how long will the Main Page continue to say "2,563,508 articles in English"?
-- Wavelength (talk) 03:08, 28 September 2008 (UTC)

Until the devs turn the statistics updating back on? Anomie 03:53, 28 September 2008 (UTC)

Systems administrators. — Werdna • talk 04:09, 28 September 2008 (UTC)

Problem uploading media

I made a spoken version of an article and saved it as a .ogg. It's 14 MB and meets all other requirements I can find. But every time I try to upload it, I get an error page. Neither Safari nor Firefox has been able to upload it. Any ideas? Huadpe (talk) 04:43, 28 September 2008 (UTC)

An old deletion summary which disparages the subject of the deleted page

Recently, when deleting a page, I discovered that a page by that name had been deleted back in 2005, with a summary which includes disparaging content. How do I get this summary changed/removed? עוד מישהו Od Mishehu 04:56, 28 September 2008 (UTC)

Only developers can change or remove log items, and they only do so in exceptional circumstances. Graham87 06:18, 28 September 2008 (UTC)
Systems administrators, even. Graham87 06:23, 28 September 2008 (UTC)

Can we grant move to not yet autoconfimred users for their own articles?

Quite a few duplicate articles and subsequent confusion result from fresh users posting their article to then discover that they should have named or capitalized them differently. While many are not aware of the move option, those who are cannot use it until they're autoconfirmed. In either case they're tempted to create a copy. Would it be possible and helpful to allow page moves in any case for the articles a user has created to at least evade some of the duplicating, redirecting and tagging that may follow?--Tikiwont (talk) 14:07, 24 September 2008 (UTC)

I'm pretty certain that the pagemove right can't get cut up that finely. EVula // talk // // 18:38, 24 September 2008 (UTC)
The "reupload" right ("allows overwriting existing images and files") is cut up this finely, as there is also a "reupload-own" right ("allows overwriting existing images and files uploaded by oneself"). This division allows newbs to fix their own mistakes (sideways image, for example) without being able to jack with other people's images until a higher access level. — CharlotteWebb 18:48, 24 September 2008 (UTC)
Support This makes complete sense, I mean, from a vandalism perspective, minimal disruption could come from a new user moving the 3r09gew09j they created to HAGGER??????? and it helps new users to rename their pages. If there isn't already a "move-own" privilege then one should be implemented. Bugzilla anyone? — ^.^ [citation needed] 10:17, 25 September 2008 (UTC)

Wait for the abuse filter. — Werdna • talk 04:45, 27 September 2008 (UTC)

Which means what? That the abuse filter would make this proposal obsolete or feasible? --Tikiwont (talk) 12:31, 28 September 2008 (UTC)
I assume that Werdna was referring to the HAGGER-type moves, knowing what the abuse filter is. עוד מישהו Od Mishehu 13:00, 28 September 2008 (UTC)

XLinkBot reverted an edit to Bendigo, Victoria, but replaced the page with the "The database did not find the text of a page that it should have found, named "Bendigo, Victoria" (revision#: 203)." etc. message. Diff link: http://en.wikipedia.org/w/index.php?title=Bendigo,_Victoria&curid=62578&diff=240921785&oldid=240921691 I have reverted the page back to the last good version. --Richmeistertalk 16:13, 25 September 2008 (UTC)

I've put a notice on XLinkBot's talk page. x42bn6 Talk Mess 20:23, 27 September 2008 (UTC)
See User talk:XLinkBot#Very weird revert on Girls (Sugababes song) for Dirk Beetstra's response, apparently was a bug, and as I just noticed it's still not quite working yet: [10]. I put that diff on the bot's talkpage, but it happens very rarely now, so I don't think it needs an emergency shutdown. --AmaltheaTalk 17:48, 28 September 2008 (UTC)
The first one was in a buggy version of XLinkBot that ran for almost 5 hours (the data I saw it retrieving seemed correct, but it later went wrong on real reverts), that has been resolved.
The last one that Amalthea now reported is a quirk of the api, it seems; the bot does not retrieve the history of a page (resulting in perlwikipedia returning '1' ...). Here it resulted in a bad revert, since then it also crashed twice for the same reason (it killed itself before this bug was reported ..). I am catching this now in the bot, but it is strange that it happens these last couple of hours, the bot did the same before for months without problem (the crashes are in parts of the code which were already in User:ShadowBot/User:AntiSpamBot). Thanks for spotting this, hope this explains. --Dirk Beetstra T C 22:20, 28 September 2008 (UTC)

Software changes to assist blind users?

Graham87 and I were talking about ways to make geometrical images more accessible to the blind readers of Wikipedia. Here's a small extension of the MediaWiki software that came up in our discussions, and I would very much appreciate people's ideas on it. Thanks! :)

The proposed extension has two parts. The first part is to enable an option in User preferences (perhaps under the "Misc" tab?) that turns off the downloading/rendering of images. This option might have advantages for other users as well, since image downloading/rendering can be slow, e.g., 4 MB GIF animations; hence, browsing Wikipedia might be much faster if you just had to render text. The second part is to enable something like the alt attribute as an optional field within the [[Image]] link to provide descriptive text to render in place of the image. Different instances of the sameimage could have different ALT texts, to allow authors to emphasize different aspects of the image. If the ALT text is omitted by the author, the description from the Image page itself could be fetched from WP or from the Commons.

The caption of an image usually doesn't serve to describe it, at least in geometry articles. Rather, the caption usually describes some argument or some conclusion that can be derived from the image.

Image rendering printers exist for the blind, but they're extremely expensive. Until they come down in price, this seems to be a good method for making Wikipedia articles more accessible to the blind, wouldn't you all agree? Willow (talk) 18:26, 26 September 2008 (UTC)

Why not just turn off images in the browser? Mr.Z-man 19:43, 26 September 2008 (UTC)

Ummm — because I didn't know that I could do that? (sheepish)

What an elegant solution to the first part! :) The only drawback with that solution is that my nice LaTeX equations are unformatted. Can we turn off the images without turning off the equation rendering? But I can also re-write my equations in non-LaTeX ways, if need be.

As for the second half — is there an equally simple way to send an ALT message to the image along with its caption? It'd be awesome if anyone had a trick for that! :) Willow (talk) 20:12, 26 September 2008 (UTC)

The image title and the alt attribute are both generated from the image caption, and I don't know of a way to change that. This effectively results in the image caption being spoken twice with JAWS, once with the wikilinks indicated and once without them. See bug 368. Graham87 08:09, 27 September 2008 (UTC)
This should be fixed in r41364, assuming that doesn't get reverted for some reason. —Simetrical (talk • contribs) 21:11, 28 September 2008 (UTC)

I came here to announce that I'd made a template {{Alt Image}} as a workaround for the problem until the MediaWiki software got fixed, but I see with pleasure that I'm a Johanna-come-lately. ;) Thank you, Simetrical! You're awesome. It's been too long since our paths have crossed, but I'm glad that they did so in such a happy way. :) Could you write up some documentation on how your method works? I'll do the same for mine. Once your system is working, all we need to do is convince everyone else to start using it for accessibility, at the very least for Featured and Good Articles. :) Willow (talk) 21:45, 28 September 2008 (UTC)

user identity

I have stumbled onto quite a few userpages which show Template:User committed identity. Would a user prefers SHA-1 over SHA-512? Is SHA-512 better than SHA-1? Thanx in advance for answering my question. BTW how to produce a SHA-512 or SHA-1 commitment? Ktsquare (talk) 01:33, 27 September 2008 (UTC)

Did you read the documentation on the {{User committed identity}} page? It seems to answer your questions. I'm not trying to be rude, I just don't see what question the documentation does not answer. --Teratornis (talk) 03:14, 27 September 2008 (UTC)
What operating system do you use? If it is not unix-like, then you may not already have the programs the template documentation mentions. --Teratornis (talk) 03:16, 27 September 2008 (UTC)
For example, if you use Microsoft Windows, you could install Cygwin to get programs like these:
$ which md5sum
/usr/bin/md5sum

$ which sha512sum
/usr/bin/sha512sum
That's directly out of the Cygwin window on the computer I'm using at the moment. --Teratornis (talk) 03:21, 27 September 2008 (UTC)
Just use http://www.johnmaguire.us/tools/hashcalc/ and be done with it. --MZMcBride (talk) 03:22, 27 September 2008 (UTC)
But if you're on a LAN, your input to the page (i.e. your passphrase) can be sniffed over the wire by a clever adversary, who'd then have both your passphrase and the hash that the server returns. Pegasus «C¦ 12:04, 28 September 2008 (UTC)
If you have someone that willing to compromise your Wikipedia account, you're probably screwed anyway. But there's also a small downloadable tool, [11] so people don't have to go through the trouble of installing Cygwin and learning how to use it just to do a SHA-512 hash. Mr.Z-man 16:55, 28 September 2008 (UTC)
The network neighbourhood is often less secure than the real-world neighbourhood, so i'd prefer offline tools for things as sensitive as this. :/ Pegasus «C¦ 03:52, 29 September 2008 (UTC)

For Mac OS X, the easiest method is to open Terminal (a default utility) and write md5 -s "your string here" where you replace the text within the quotes with your string. {{Nihiltres|talk|log}} 23:23, 28 September 2008 (UTC)

DISPLAYTITLE addition?

per this request: Wikipedia:Requested_templates#Article_titles_in_italics. it seems to me it would be possible to fix the Magic Word {{DISPLAYTITLE}} so that it can handle span elements or limited wikitext markup, and thus generate italics and such in page titles. the original concern was over articles concerning scientific names where all or part of the name is conventionally always rendered in italics. the DISPLAYTITLE variable currently allows for minor modifications of the title (so long as the produced title is equivalent to the actual page title) but it doesn't recognize formatting changes as equivalents. We'd have to change DISPLAYTITLE so that it recognizes formatting differences as equivalents (I'd have to see the PHP code to know exactly how to do that - if someone could point me to the code I could look at it - but it shouldn't be too difficult). it seems like a worthwhile change to me, though we'd have to restrict it to prevent weird title abuses. --Ludwigs2 23:43, 27 September 2008 (UTC)

What you're describing is exactly what Template:wrongtitle already does, although I haven't gotten around to writing a PHP implementation of this for use in MediaWiki. Whether we want italic titles is another matter. —Remember the dot (talk) 02:44, 28 September 2008 (UTC)
Strangely, template:wrongtitle's documentation says nothing about JavaScript which is required for many cases. As for MediaWiki implementation, bugzilla:14226 "Allow sup and sub in {DISPLAYTITLE:}" comes to mind. —AlexSm 03:55, 28 September 2008 (UTC)
ah, interesting. I had no idea that template existed. I'll pass on that information to the concerned parties. do you think I should advise them against using it for the time being, because of the 'italics in titles' concern?
Alex, in what cases is javascript required? I'm happy to take the time to update the documentation. --Ludwigs2 20:09, 28 September 2008 (UTC)

Has Rollback changed?

I just did a rolllback, and instead of just telling me that there was a rollback performed, it showed me the diff of what I was rolling back from. That's new. Corvus cornixtalk 06:06, 28 September 2008 (UTC)

Sme Thing Happened with me. Uhh, i take it that they did change it. Here II MusLiM HyBRiD II 13:13, 28 September 2008 (UTC)
Noticed it yesterday (My time). Love the idea. :) Bidgee (talk) 13:18, 28 September 2008 (UTC)
I thought you left! And, could someone tell us if its a perm. fix or bug? II MusLiM HyBRiD II 13:23, 28 September 2008 (UTC)
I hate it, because it doesn't tell us that the rollback actually worked. You have to go look at the article's history to see if you actually rolled back anything. Corvus cornixtalk 19:08, 28 September 2008 (UTC)
I would to, if it were happening to me, but it isn't. Doug Weller (talk) 19:09, 28 September 2008 (UTC)
I wonder if admin rollback is working differently from non-admin rollback. I'm going to do a test in my space to see if it's still happening. Corvus cornixtalk 19:39, 28 September 2008 (UTC)
Now it says "Edits by xxx have been changed", so I guess I can live with that. Corvus cornixtalk 19:40, 28 September 2008 (UTC)

And now moves are not working

I tried to move James settelmyer to a capitalized version of the page. I clicked on "Move" three times, nothing ever happened. Corvus cornixtalk 06:24, 28 September 2008 (UTC)

I don't know - it seems to have been moved. עוד מישהו Od Mishehu 06:52, 28 September 2008 (UTC)

I was able to move Matt lee with no problem. Corvus cornixtalk 07:04, 28 September 2008 (UTC)

Watchlist not updating

Resolved

My watchlist doesn't show any updates since 05:48 UTC (40 minutes ago). Could this be related to the two reports above? -- Zsero (talk) 06:28, 28 September 2008 (UTC)

Same. SandyGeorgia (Talk) 07:01, 28 September 2008 (UTC)
I am experiencing the same problems, as of 5:48. Should we send something in to Bugzilla? —Anonymous DissidentTalk 07:25, 28 September 2008 (UTC)
Two hours, no change. I don't know what or where Bugzilla is; I'm reporting here :-)). I do find it strange that one of the top ten websites on the 'net can't/doesn't update its users when it has technical issues, though. The last time I whined, I was told WP:TIND: great motivation for the volunteers. Just an indication that someone knows of the issue and is working on it is always nice ... SandyGeorgia (Talk) 07:55, 28 September 2008 (UTC)
Hello, is this thing on? My watchlist is dead too, as of 5:43 UTC. Ominously, MediaWiki was updated at 5:38. Franamax (talk) 07:56, 28 September 2008 (UTC)
Ditto - still an ongoing problem here. Ghmyrtle (talk) 07:58, 28 September 2008 (UTC)
I just checked at #wikimedia-tech and it appears that no sysadmins are awake currently, so we must wait for someone to wake up. I don't suppose there is some sort of alarm that could be triggered. That would be nice. __meco (talk) 08:01, 28 September 2008 (UTC)
I was asked to post this link: http://svn.wikimedia.org/viewvc/mediawiki?view=rev&revision=41277. __meco (talk) 08:09, 28 September 2008 (UTC)
http://svn.wikimedia.org/viewvc/mediawiki?view=rev&revision=41337 is the one that broke it according to https://wikitech.leuksman.com/view/Server_admin_log . Matthewedwards (talk contribs  email) 08:24, 28 September 2008 (UTC)
That's incorrect - the entry meant that the site now runs on the version of MediaWiki checked out of SVN at r41337. It was previously at r41097, so any revision between the two could have caused borkage. MER-C 11:03, 28 September 2008 (UTC)
Hey, meco, thanks for letting us know and thanks for the link (which is gibberish to me ... I think it says "Time to pack it in for the day"). SandyGeorgia (Talk) 08:11, 28 September 2008 (UTC)
I was just thinking the same thing. Bed seems to be calling me. Matthewedwards (talk contribs  email) 08:16, 28 September 2008 (UTC)

If it were a critical issue, we could give a sysadmin a call. Since it's just watchlists, I don't think it's sensible to do so. Werdna 08:12, 28 September 2008 (UTC)

Special:RecentChanges still works though, and watchlisted items still appear emboldened. Also, a notification has been placed by the admins to the top of the Watchlist, "Watchlists are currently broken! This will be fixed shortly." but yeah, we have to wait for a dev or whatever to wake up. Matthewedwards (talk contribs  email) 08:16, 28 September 2008 (UTC)
Recentchanges is a very poor replacement indeed. It only shows the last 500 edits which amounts to the last five minutes of editing (7 minutes if limited to main namespace). __meco (talk) 08:41, 28 September 2008 (UTC)
Well, I'm basically stymied. I have 11,000 items in my watchlist and it is my most important working tool here. __meco (talk) 08:36, 28 September 2008 (UTC)
and I thought 2,500 was big! :o ATM I'm removing a few that I know longer need but it's a really pain no to have the watchlist for articles that I'm interested in and having to check the articles the hard way. Bidgee (talk) 10:17, 28 September 2008 (UTC)
So... Who's in the mood for some tic-tac-toe? Waltham, The Duke of 09:04, 28 September 2008 (UTC)
I'm surprised that the Watchlist is so low priority... considering that the whole of WP could be vandalised while its offline and we wouldn't know! ~~ [Jam][talk] 09:14, 28 September 2008 (UTC)
WP:BEANS. :) PeterSymonds (talk) 09:23, 28 September 2008 (UTC)
Fair enough.
So bored... ooo... beans... ~~ [Jam][talk] 09:25, 28 September 2008 (UTC)
And there would be a very interesting statistic - Of the reversions of some random seletion of vandalistic edits, how many of the reverters reached the article via watchlist, how many via RC patroll, and how many via other? Someguy1221 (talk) 09:30, 28 September 2008 (UTC)
Just spotted the message on my watchlist saying they're currently broken! Lugnuts (talk) 09:50, 28 September 2008 (UTC)
So, when it says occurred at 05:48am, September 28, 2008, is it getting the time from my prefs? Whoever placed the notice should qualify that. Yngvarr (t) (c) 10:23, 28 September 2008 (UTC)

I find that the last part of the message (This problem will be fixed shortly) is a little misleading since it's not worked for at least 3 hours and should be This problem will be fixed ASAP/As soon as possible. Bidgee (talk) 10:34, 28 September 2008 (UTC)

(ec) Heh, agreed with Bidgee; "will be fixed shortly" seems to be the sort of statement that would get people to hit F5 every few seconds, when really the solution is to go do something else... such as bake. Cornbread anyone? —/Mendaliv//Δ's/ 10:43, 28 September 2008 (UTC)

It's fixed. Tim lives in Australia (just near me), but nobody thought it important enough to summon him. It's 10pm on a Sunday, here :) — Werdna • talk 11:54, 28 September 2008 (UTC)

diffeerence in source and page

[12] on this page the wikiproject pakistan's rating is B-class in the source but on the page it appears as C-class.User talk:Yousaf465

The {{WP Pakistan}} banner is coded to display C instead of B class unless certain other parameters are also specified. Anomie 19:01, 28 September 2008 (UTC)

Is there any way to make

http://atlas.nrcan.gc.ca/site/english/maps/topo/map?mapsize=428%20380&lat=44.8919444464&long=-63.8291666677&mapxy=2407416.76111+141348.507625&scale=5000000&feature_na=Mount+Uniacke&location1=12&unique_key=c8c62e2daffc11d892e2080020a0f4c9&searchstring=Mount%20Uniacke&entity=UNP&layers=fapfeature+fapfeature%20north_arrow%20other_features%20million_grid%20t50k_grid%20grid_50k_3%20roads%20hydrography%20boundary%20builtup%20vegetation%20populated_places%20railway%20power_network%20manmade_features%20designated_areas%20water_features%20water_saturated_soils%20relief%20contours%20toponymy%20contour%20nodata_ntdb_50k&urlappend=%26unique_key%3Dc8c62e2daffc11d892e2080020a0f4c9%26map.layer[textzoom03]%3DFEATURE+POINTS+2407434.2005+141353.603457+END+TEXT+%22Mount%2BUniacke%22+END%26map.layer[textzoom46]%3DFEATURE+POINTS+2407434.2005+141353.603457+END+TEXT+%22Mount%2BUniacke%22+END%26map.layer[arrowzoom03]%3DFEATURE+POINTS+2407434.2005+141353.603457+END+END

work correctly?

Right now it produces

[textzoom03%3DFEATURE+POINTS+2407434.2005+141353.603457+END+TEXT+%22Mount%2BUniacke%22+END%26map.layer[textzoom46]%3DFEATURE+POINTS+2407434.2005+141353.603457+END+TEXT+%22Mount%2BUniacke%22+END%26map.layer[arrowzoom03]%3DFEATURE+POINTS+2407434.2005+141353.603457+END+END

I know the problem is caused by the ] after textzoom03 but I can't figure out how to fix it. Clicking on the link textzoom03 will take you to the correct page but causes a popup error and the map will not load. However, the coordinates are shown which is the point of the reference. CambridgeBayWeather Have a gorilla 23:27, 27 September 2008 (UTC)

Perhaps {{urlencode:}} – a pseudo-template that can be used to "wrap" an unusual URL (for example, one that has brackets within in). (Description is from the index section WP:EIS#URL). -- John Broughton (♫♫) 23:36, 27 September 2008 (UTC)
Replace the brackets with %5B and %5D, like this: [13] Anomie 23:42, 27 September 2008 (UTC)
Thanks that works perfectly. Now if there just was a way to not have a 9 line reference in the middle of a stub article. CambridgeBayWeather Have a gorilla 23:49, 27 September 2008 (UTC)
You could omit any unneeded parameters (&foo=bar) of the url, probably most of them. Your coordinates are still off by about 1.2 miles according to Google Maps, which says 44°54′14.30″N 63°50′32.27″W / 44.9039722°N 63.8422972°W / 44.9039722; -63.8422972. — CharlotteWebb 20:47, 29 September 2008 (UTC)

editing templates -- help please

I'm trying to edit the template for The Band by adding the rockumentary Festival Express. For some reason, it's just not working. When I save changes, and view the template on its own, the changes are visible, but when you look at the template at the bottom of the article, the changes are gone. But when I try to edit it again, the changes are still there in the "edit" page for the template. What's up with that? Minaker (talk) 10:37, 29 September 2008 (UTC)

It takes a variable amount of time before template edits are propagated to pages which transclude the template. Just wait, or purge the page if you want to to see the result right away. PrimeHunter (talk) 13:24, 29 September 2008 (UTC)

Speedy deletion notices

When I've been adding speedy deletion notices today, I've noticed that some are not appearing in the "Candidates for speedy deletion" category. For instance, William blair hickman, tagged with {{db-bio}} only appears in "Importance or significance not asserted pages for speedy deletion" but should appear in both according to the template. Something wrong there... Stephenb (Talk) 10:59, 29 September 2008 (UTC)

CAT:CSD has been marked as a Hidden category, meaning the category link is not displayed on articles. This is because it's not a grouping useful to readers; it's a 'backend' category that should only be accessible from the 'back door'. For some reason, the other CSD categories haven't been marked as such... that's probably an oversight. Happymelon 11:20, 29 September 2008 (UTC)
Ah, thanks for the explanation. Either a recent change or the first time I've noticed! Stephenb (Talk) 11:27, 29 September 2008 (UTC)
You can select "Show hidden categories" under Misc at Special:Preferences. PrimeHunter (talk) 13:28, 29 September 2008 (UTC)

CorenSearchBot confused?

On User talk:Mfishnitinp, CorenSearchBot seems confused, referring to a (deleted?) page that doesn't exist, and resulting in this error message. --Geronimo20 (talk) 00:47, 30 September 2008 (UTC)

The url in your link is http://en.wikipedia.org/wiki/Quota+Management+System?diff=241674123. It was not made by CorenSearchBot but by User:La Pianista in [14]. The url is similar to the one I reported at [15]. The deleted page is http://en.wikipedia.org/wiki/Quota_Management_System with the normal '_' between words and not '+'. As an administrator I can see that it was created and edited many times by Mfishnitinp. PrimeHunter (talk) 01:10, 30 September 2008 (UTC)
By the way, the article was deleted and recreated between the two posts by CorenSearchBot and I don't see any problem in those two posts. PrimeHunter (talk) 01:14, 30 September 2008 (UTC)
So why doesn't this activity show up in Mfishnitinp's edit history? --Geronimo20 (talk) 02:03, 30 September 2008 (UTC)
Only administrators can see edits to deleted pages. PrimeHunter (talk) 02:06, 30 September 2008 (UTC)
Thank you :) --Geronimo20 (talk) 02:10, 30 September 2008 (UTC)

When to use hidden/collapsible sections

Split from the thread above, #Accessibility of NavFrame.

See also:


We really need some solid recommendations about when/not to use the "hidden" code, outside of footer-navboxes. -- Quiddity (talk) 20:25, 24 September 2008 (UTC)

I would like to point out that the issue raised in this section was about "hidden box" behaviour (once it's used somewhere), not about "when to use it"; it might be better to start a separate section for that. —AlexSm 20:46, 24 September 2008 (UTC)
Good point. Thread duly split/refactored.
  • Are there any more links to relevant discussions about hidden-collapsible sections?
  • The various hiding-templates often get used to hide content that some editors simply cannot agree on whether to display or not (see the "influences" sections in some Writer-infoboxes (e.g. William Gibson), the Ponte Vecchio experiment, the vertical navboxes linked above, etc). Is this a usage we want to encourage or discourage?
  • What code should be used? Wikipedia:NavFrame says it is deprecated, but it is widely used by all of the hiding templates ({{Hidden}}, {{Show}}, {{Hidden begin}}, {{HiddenMultiLine}}, {{Hidden section top}}, {{Hidden infoboxes}}) none of which mention deprecation.
  • Any suggestions as to what we should be using for guidelines? Or where we should be discussing it? -- Quiddity (talk) 04:42, 1 October 2008 (UTC)

Math rendering bug

The antialiasing on superscript z's (e.g. [16]) makes z look more like a than a z. Probably a bug with some related software (TeX?) but I don't know where to post this. Doodle77 (talk) 13:10, 30 September 2008 (UTC)

I filed the bug as bugzilla:15777. But there isn't much active development on the mediawiki TeX system, so it may be a long time before the problem is solved. — Carl (CBM · talk) 13:35, 30 September 2008 (UTC)

RE: Watchlist problem help

Hello, I am wondering if anyone can help my technical problem. Whcih is quite important:

Everytime I try to enter "My Watchlist", I get this message:

Redirect Loop

Redirection limit for this URL exceeded. Unable to load the requested page. This may be caused by cookies that are blocked.

The browser has stopped trying to retrieve the requested item. The site is

redirecting the request in a way that will never complete.

  • Have you disabled or blocked cookies required by this site?


  • NOTE: If accepting the site's cookies does not resolve the problem, it is probably a server configuration issue and not your computer.

I really need some help on this one, thanks. Police,Mad,Jack (talk · contribs) 16:15, 30 September 2008 (UTC)

This is under discussion here already. Try bypassing your cache to see the Watchlist again. Regards SoWhy 16:17, 30 September 2008 (UTC)
Sorry, but that doesn't solve the problem for me at all. Instead, things run as slow as treacle as well as getting the iedntical error message that was posted above. I've cleared the cache, bypassed the cache, set the cache to zero and tried all the solutions suggested in the link you provided: none work. Firefox 3.0.3 being used here, but the problem also persists if I use Internet Explorer instead.  DDStretch  (talk) 17:12, 30 September 2008 (UTC)

I cant even log in now, I'm User:Police,Mad,Jack, I really need some help. 217.42.11.95 (talk) 17:15, 30 September 2008 (UTC)

Just to confirm, the IP user is me. But I can log in now, but no luck with the watchlist. Police,Mad,Jack (talk · contribs) 17:19, 30 September 2008 (UTC)

Bypassing the cache doesn't work for me. BigDuncTalk 17:28, 30 September 2008 (UTC)
It seems now that a non-intentional blunder was committed by an admin, as you can see in my above post. —Cesar Tort 17:58, 30 September 2008 (UTC)

Two sections? O.O

Uhh, yeah, I posted something in a talk page of a article. Just see my images below. Another bug? This never happened to me before, Never!

CLICK ME

CLICK ME

Thanks.. II MusLiM HyBRiD II 15:52, 30 September 2008 (UTC)

Diff is here. One explanation would be that you hit the Save page button twice. Gary King (talk) 15:56, 30 September 2008 (UTC)

Ever since the statistics page was updated to the new layout I don't think it has been logging the stats properly. It only seems to log about two edits a minute when that number should be much higher. Is the page broken? Can the statistics be considered accurate anymore? --Andrew Kelly (talk) 21:07, 24 September 2008 (UTC)

Well, the stats updates are disabled at the moment due to DB problems (see up a couple sections). --69.17.48.227 (talk) 21:24, 24 September 2008 (UTC)
Ah, I see. Thanks for that, guess I missed the obvious! Hopefully they will be updated soon. --Andrew Kelly (talk) 21:35, 24 September 2008 (UTC)
I hope so too. I've been logging the English article counts daily (several times per day) for 6 months, and this is the first time I am losing significant data. (by the way, I just added a graph of the English article counts on my home page in case anybody is interested in seeing the data.) HowiAuckland (talk) 04:16, 26 September 2008 (UTC)
How close are the problems with the statistics page going? I ask this because there was a bigger bump in the article count today than there has been in the past few days, but I still can't believe that less than 1,000 articles were added over this time.SPNic (talk) 14:07, 28 September 2008 (UTC)
Indeed. The number of articles are increasing in number again, but not at a typical rate. The Special:Statistics page was being updated, even since the page was changed, just not at the normal rate, ... a very slow rate. That continues, but at an ever quicker, but still slow, pace. HowiAuckland (talk) 04:25, 29 September 2008 (UTC)
The "Job queue length" also shows far to low values. Some days ago and also today I made changes to most of the mbox templates such as {{imbox}}. That is updates that affects more than 1 million pages. But both times the "Job queue length" hardly reacted. Today I reloaded the the Special:Statistics page about every 15 seconds or so during those saves and for some minutes after it to keep a close eye on it. But the "Job queue length" stayed at around 90 thousand all the time.
--David Göthberg (talk) 01:43, 2 October 2008 (UTC)
Well, the wikimedia server admin log shows brion shutting off the whatever that updates the statistics, and doesn't say anything about it being re-enabled. Calvin 1998 (t·c) 01:54, 2 October 2008 (UTC)

How to list subpages?

Template:Annotated image is rapidly acquiring a library of subpages containing annotated images built with it, e.g. Template:Annotated image/Arthropod head problem. Is there any automated way to build / display an index of these sub-pages, e.g. in Template:Annotated image/doc? -- Philcha (talk) 13:25, 30 September 2008 (UTC)

How about {{Special:Prefixindex/Template:Annotated image/}}, which renders as:
  • Annotated image/Arthropod head problem/doc
  • Annotated image/Extinction
  • Annotated image/Extinction/doc
  • Annotated image/Himachal Pradesh
  • Annotated image/Porifera cell types
  • Annotated image/Porifera cell types/doc
  • Annotated image/Skeletal anatomy of T rex right leg
  • Annotated image/Skeletal anatomy of T rex right leg/doc
  • Annotated image/Snail radula working
  • Annotated image/Snail radula working/doc
  • Annotated image/Trochophore larva
  • Annotated image/Trochophore larva/doc
  • Annotated image/doc
  • Annotated image/doc/Samples
  • Annotated image/sandbox
  • Annotated image/testcases
  • Hemmingsen 13:58, 30 September 2008 (UTC)
    That looks really useful, thanks!
    Is there any way to exclude the /doc pages? -- Philcha (talk) 19:12, 30 September 2008 (UTC)
    No way that I can think of. If you used another system for naming the subpages with for example "Template:Annotated image/images/Extinction" for the image and "Template:Annotated image/doc/Extinction" for the corresponding documentation, it would just be a matter of adding "images/" to the prefix, but moving everything around is probably more trouble than it's worth. Maybe someone else has a better idea? Hemmingsen 15:23, 1 October 2008 (UTC)
    Could use a grep tool: subpages without /, but it's not automated, i.e. you cannot transclude it. —AlexSm 16:30, 1 October 2008 (UTC)

    Can't move-protect

    So after Neanderthal was page-move vandalised, I try protecting non-admin moves indefinitely (as the article probably never needs to be moved). For editing, I click "Allow all users" (set to indefinite) and for moving, I click "Block all non-admin users" (set to indefinite), yet the form returns with "Expiry time is invalid." Spellcast (talk) 09:11, 1 October 2008 (UTC)

    Worked fine for me. Maybe you or your browser added a space to the expiry field or something? Happymelon 09:17, 1 October 2008 (UTC)
    Nope, I selected "indefinite" in the drop down menu yet it didn't work. Spellcast (talk) 09:19, 1 October 2008 (UTC)
    I tried it on a different page that won't need moving and it worked :S I guess this can be considered resolved Spellcast (talk) 04:27, 2 October 2008 (UTC)

    Wikipedia, the image-free encyclopedia

    I see dead images. Or, more accurately, I see no images. They don't have the broken image placeholders, they just don't appear. When I click on an image's link, it doesn't appear on the image description page either, and when I click on the link there, I get a 403 error. Aylad ['ɑɪlæd] 12:35, 1 October 2008 (UTC)

    Sometimes it's a server issue. Try refreshing by holding Ctrl-F5 to purge your browser's cache. If that doesn't work, could you give us some examples of broken images or images missing from their pages? Thanks, UltraExactZZ Claims ~ Evidence 14:02, 1 October 2008 (UTC)
    It was all images on all article pages... didn't check any other namespace. The Ctrl+F5 didn't help when I tried it, but perhaps a half hour after my post, the images came back. It may have been the malfunctioning content filter I just got an email about here at work. Thanks anyway. Aylad ['ɑɪlæd] 14:23, 1 October 2008 (UTC)

    Allow this user to edit own talk page while blocked

    Resolved

    This seems to be a new addition to blocking options today. While it is checked by default, my own testing indicates that any pre-existing blocked IP addresses are currently unable to edit their talk pages. This seems such a fundamental shift in policy I wonder if was actually intended? -- zzuuzz (talk) 13:51, 28 September 2008 (UTC)

    That's probably a bug. —Simetrical (talk • contribs) 20:20, 28 September 2008 (UTC)
    That was one of my features, and needs a query run on enwiki to fix it. I'll poke a sysadmin. Matt/TheFearow (Talk) 07:39, 29 September 2008 (UTC)
    Why didn't you just make the default state "able to edit talk pages" instead of having to run a database query? —Simetrical (talk • contribs) 17:23, 29 September 2008 (UTC)
    Because the default for most wikis is for users to not be able to edit their own talk pages while blocked. I did realise a way around that, and should be committing the fixed one soon, but this was already created. It's a simple enough query. Matt/TheFearow (Talk) 22:43, 29 September 2008 (UTC)
    Would someone also be able to check again, to see if this was fixed or not? Matt/TheFearow (Talk) 22:51, 29 September 2008 (UTC)
    What sort of fix is planned? Are you going to take away the check box about the blocked user editing their own talk page? EdJohnston (talk) 02:52, 30 September 2008 (UTC)
    It's a simple database query, "UPDATE ipblocks SET ipb_allow_usertalk = 1;". But I don't know whether it was done yet or not (if not, it's simple enough). Matt/TheFearow (Talk) 07:28, 30 September 2008 (UTC)
    No it's not been fixed, for IP addresses at least - I don't have any blocked user accounts I can check with. I haven't filed a bug either, so we're currently relying on your poking to get it fixed. Two related points of interest - it works the opposite way to the prevent e-mail option, which is somewhat counter-intuitive; and it has affected blocks with tools like Huggle and some blocking scripts which presumably do not pro-actively set this value. For Wikipedia use at least, it seems to be made the wrong way round. Thanks for poking. -- zzuuzz (talk) 08:27, 30 September 2008 (UTC)
    (unindent) These blocking scripts need updating, it's a simple one line fix to them. It was set this way because that's the way the option is, and it makes more sense from a coding standpoint. I've added a tag to the bug so it can be updated as soon as a shell notices it. Matt/TheFearow (Talk) 12:02, 30 September 2008 (UTC)
    Should be fixed now, Tim just ran the query. Can someone confirm this? (note: it will not show up on IPBlockList till r41535 is synced, which will happen sometime in future) Matt/TheFearow (Talk) 09:50, 2 October 2008 (UTC)
    Yes, thanks. -- zzuuzz (talk) 09:58, 2 October 2008 (UTC)

    Are people making changes NOT shown in history?

    I've been working on Community Reinvestment Act today and weird changes keep showing up, without my or anyone else evidently making them. It's happened at least twice. Is this a system hangup or can people somehow manage to do this? Thanks. Carol Moore 20:33, 28 September 2008 (UTC)Carolmooredc

    Are you saying that there are changes that don't appear in the diffs, or that there are changes that do appear in the diffs, but that were not actually made by the editors that the diffs show? — Carl (CBM · talk) 20:46, 28 September 2008 (UTC)
    It could be that people make changes to things that are transcluded into articles. Green caterpillar (talk) 20:51, 28 September 2008 (UTC)
    the only non- or semi-protected template on that page that's been changed recently is {{USStat}} (changed on the 22nd). the other three have changes on the 1st and the 4th. it would help to know, though, what the 'weird' changes are. --Ludwigs2 21:34, 28 September 2008 (UTC)
    Ok, I found a couple edits that don't show up in any diffs, around the time a lot of anon IPs were editng (and vandalizing). Maybe the diffs just failed to show the changes at all for some reason. I looked a couple times so don't think I missed them if they were there. Carol Moore 13:09, 29 September 2008 (UTC)Carolmooredc
    In rare cases a diff will show wrong results but we cannot say whether that happened here when you haven't given an example of text that appeared or disappeared in the article. Incorrect diffs can usually be fixed by purging: append &action=purge to the end of the diff url. PrimeHunter (talk) 13:21, 29 September 2008 (UTC)
    It just happened in another article about an hour ago. I went over it a couple times to make sure. However, when I went to back to try to find it just now, the correct diff had re-appeared. So maybe it's just a lag in the system which I didn't give time to catchup yesterday (there were so many various edits on that article hard to find at that point anyway, not to mention now). So i'll hope it's just that. Carol Moore 18:53, 29 September 2008 (UTC)Carolmooredc
    I am beginning to suspect that an editor with power to delete Diffs and edit summaries may be doing so in Community Reinvestment Act. I keep finding edits that are not in the DIFFs and I have to go back to my original a few hours back to catch some of them! Last night I made an edit that I clearly remember looking at in the history - now it's gone. This is a hot topic right now, there's a lot of POV deleting of things that make this Act look bad, things people don't want reverted, and there are people worried about losing a lot of money if this Act loses credibility. Plus it's a hot McCain/Obama issue. So if some admin knows how to look at the history for deletions, I'd appreciate it. At least we'd know if it's systematic tech problems with wikipedia or pov vandalism. THANKS!! Carol Moore 13:43, 2 October 2008 (UTC)Carolmooredc
    There are no deleted contributions to Community Reinvestment Act, nor are there any oversighted revisions in the log (which goes back to 2006). --jpgordon∇∆∇∆ 14:20, 2 October 2008 (UTC)

    <--- Thanks. This is getting me nuts. Back to the original problem, I still find that even though I systematically go through the changes on the change pages, I later find changes that I did not find on them. (And I don't think it's me ignoring a scroll.) So does wikipedia sometimes goof? Thanks! Carol Moore 15:19, 2 October 2008 (UTC)Carolmooredc

    Why was Wikipedia down; just now?

    Resolved

    I heard it was something about an update gone wrong? Can anyone clear it up? 82.73.59.139 (talk) 14:41, 30 September 2008 (UTC)

    According to the Server admin log, RobH updated a config file and crashed the site with it :D Happymelon 14:59, 30 September 2008 (UTC)
    Liiies. PeterSymonds (talk) 18:34, 30 September 2008 (UTC)
    Did I break the internet or not? Jennavecia (Talk) 18:42, 30 September 2008 (UTC)
    Redirect config problem happened on the 30th, took a while to clear out of caches. --brion (talk) 20:26, 2 October 2008 (UTC)

    Can't load a page

    Resolved

    Completely weird. I can load everything else, but Wikipedia talk:Featured article criteria won't load for me ... help? I can't even get a URL to add a purge. SandyGeorgia (Talk) 15:04, 30 September 2008 (UTC)

    When the page doesn't load, the URL should be in the address bar, where you can add "&action=purge" "?action=purge" to purge the page. That works for me. I purged the criteria page now. Gary King (talk) 15:05, 30 September 2008 (UTC)
    You left off the question mark, Gary ... thanks, it worked this time. SandyGeorgia (Talk) 15:08, 30 September 2008 (UTC)


    Redirect config problem happened on the 30th, took a while to clear out of caches. --brion (talk) 20:26, 2 October 2008 (UTC)

    Apparent server problem

    Resolved

    After the recent server problem I can access e.g. Nudibranch or Osphradium OK, but Chromodoris quadricolor gives me a message "Redirection limit for this URL exceeded. Unable to load the requested page." Looks like one of the servers needs some TLC.

    Just noticed that the special symbols below the edit box are not coming up as links. -- Philcha (talk) 15:09, 30 September 2008 (UTC)

    Some pages need to be purged by adding "?action=purge" to the end of the URL and hitting Enter. Gary King (talk) 15:13, 30 September 2008 (UTC)
    I had the same error message for my watchlist. I've tried to sort my cookies out and decided to log off and back in, and now the Login page has the same error message. 91.106.106.87 (talk) 15:14, 30 September 2008 (UTC)
    Redirect config problem happened on the 30th, took a while to clear out of caches. --brion (talk) 20:26, 2 October 2008 (UTC)

    Watchlists down?

    Resolved

    I can't get my watchlist to load (infinite spin on IE, redirect loop on Firefox and Chrome). Anyone else? Bongomatic (talk) 15:13, 30 September 2008 (UTC)

    Yup, same here. Rest of the site is extremely slow to load as well. Ghmyrtle (talk) 15:15, 30 September 2008 (UTC)
    Same here (related to the direct above topic). D.M.N. (talk) 15:16, 30 September 2008 (UTC)
    (ec) See the section above. Much wackiness at present. I get my watchlist, but all 'collapsible' sections are expanded and showing no collapse option. You might try changing your watchlist settings. I'm guessing it might be the 'show all revisions' option which is having trouble. --CBD 15:17, 30 September 2008 (UTC)
    (ec) Yeah. WP:BYPASS solved the problem for me. — Emil J. 15:19, 30 September 2008 (UTC)
    Went here to report it as well. They work when you bypass the cache but only once. If you click "My Watchlist" again, the error returns. SoWhy 15:22, 30 September 2008 (UTC)
    It's been working for me for a while even without bypassing the cache at all, so I guess YMMV. Gary King (talk) 15:24, 30 September 2008 (UTC)
    I'm seeing trouble as well, and I'm experiencing that ja.wikipedia.org's watchlist is redirected to en.wikipedia.org. --Fukumoto (talk) 15:26, 30 September 2008 (UTC)
    fr too. Someone has broken the Web. GillesC (talk) 15:48, 30 September 2008 (UTC)
    Also de and es. It.wikipedia's watchlist still works. --Al Pereira(talk) 15:53, 30 September 2008 (UTC)
    Watchlist works on Hebrew wikipedia. RolandR (talk) 16:03, 30 September 2008 (UTC)
    de.wikipedia works after bypassing cache--Dany3000 (talk) 16:43, 30 September 2008 (UTC)

    (outdent) Sorry, but bypassing the cache doesn't solve the problem for me at all. Instead, things run as slow as treacle as well as getting the identical error message as before: "Redirection limit for this URL exceeded. Unable to load the requested page. This may be caused by cookies that are blocked." I've cleared the cache, bypassed the cache, set the cache to zero, unblocked all relevant cookies, deleted all cookies, thus trying all the solutions suggested and more: none work. I'm principally using Firefox 3.0.3, but the problem of things timing out or just not loading all occurring with Internet Explorer as well. I think it's a server problem, and I note someone did something to ot just a little while ago which crashed the site.  DDStretch  (talk) 17:21, 30 September 2008 (UTC)

    It even affected the Spanish Wikipedia, as you can see in this thread. —Cesar Tort 17:40, 30 September 2008 (UTC)
    Redirect config problem happened on the 30th, took a while to clear out of caches. --brion (talk) 20:25, 2 October 2008 (UTC)

    Unable to log in

    Resolved

    I can't access the Wikipedia login page. Firefox gives me a message "Firefox has detected that the server is redirecting the request for this address in a way that will never complete." and it says "This problem can sometimes be caused by disabling or refusing to accept cookies". But I have not disabled my cookies or anything. Is this something wrong with the page? 202.124.190.212 (talk) 15:18, 30 September 2008 (UTC)

    I purged it. Try now. Gary King (talk) 15:21, 30 September 2008 (UTC)
    Nope. Still the same. Not working in IE7 either. 202.124.189.180 (talk) 15:27, 30 September 2008 (UTC)
    Changing to Level 3 header - appears to be the related to above. D.M.N. (talk) 15:31, 30 September 2008 (UTC)
    Try bypassing your cache. It solves the problem above. SoWhy 15:44, 30 September 2008 (UTC)
    Done, but I can still not get it to load. :| D.M.N. (talk) 17:05, 30 September 2008 (UTC)
    I've had the same symptoms for the last couple of hours. Slightly borked wikipedia :( --217.28.34.132 (talk) 17:31, 30 September 2008 (UTC)
    Redirect config problem happened on the 30th; took a while to clear it out of proxy caches. --brion (talk) 20:25, 2 October 2008 (UTC)

    whatever happened to Cortado?

    Didn't we have some sort of Java-based OGG player on Wikipedia at one point? What happened to it? ViperSnake151 19:45, 1 October 2008 (UTC)

    It's still there. --brion (talk) 20:22, 2 October 2008 (UTC)

    AOL show/hide problem?

    With AOL 9.0 under Windows Vista, the "show/hide" option isn't showing up on WikiProjectBannerShell, so that all banners are displayed, nor is it there on navboxes, which are all open. Things seem to be fine under IE 7.0, Firefox 2.0 and Safari 3.1.2. The problem just developed recently. Any thoughts? Ed Fitzgerald t / c 21:34, 1 October 2008 (UTC)

    It sounds like you either have javascript disabled or you're getting a javascript error which is stopping the script from running. Does this still happen if you are not logged in? Anomie 23:07, 1 October 2008 (UTC)
    No, if I log out, "show/hide" comes back. Hmmmm. Ed Fitzgerald t / c 00:10, 2 October 2008 (UTC)
    Other features - such as enhanced recent changes - which say they require Javascript seem to be working when I'm logged in. Ed Fitzgerald t / c 01:10, 2 October 2008 (UTC)

    OK, I'm confused. I appear to have javascript enabled, as features that require javascript are working, but if I log out or use a different broswer while logged-in, there's no problem. What had changed recently that would affect show/hide only in AOL? Ed Fitzgerald t / c 03:25, 2 October 2008 (UTC)

    Try it with &useskin, like modern (on whatever page it doesn't work on). If that works, maybe blame your /monobook.js --Splarka (rant) 07:08, 2 October 2008 (UTC)
    You seem to have hit the nail on the head. Everything's fine with the modern skin in place, but still out of whack when I go back to monobook. Any idea what I should look for that's wrong in the file? As far as I can tell, it hasn't been touched since May 31st.Ed Fitzgerald t / c 07:15, 2 October 2008 (UTC)
    I restored on old version of the monobook.js and "show/hide" seems to work fine now. I should be able to find what's in the file that's causing the problem. Thanks for the tip. Ed Fitzgerald t / c 07:24, 2 October 2008 (UTC)

    Two days ago, I edited Template:Susquehanna_River_System to fix a disambiguation. I changed a link in the template to point to Huntington Creek (Pennsylvania) rather than Huntington Creek.

    But looking at the pages that link to Huntington Creek, all the pages using the river template are still listed as pointing to the disambiguation page.

    Isn't this supposed to be handled in the job queue(http://www.mediawiki.org/wiki/Manual:Job_queue)? I thought that these types of changes go through pretty quickly, especially for a template that "only" shows up in a couple hundred pages.

    Did something break? DeFaultRyan (talk) 18:52, 2 October 2008 (UTC)

    The job queue is well over 9,000 right now (actually more like 80,000). The couple hundred "jobs" triggered by your edit will be handled chronologically (see Queue (data structure)). Check back in a few hours or do a null edit to each affected page if you are impatient. — CharlotteWebb 19:29, 2 October 2008 (UTC)
    Thanks for the response, but that's why I mentioned that this template changed took place two days ago. I did a bit of reading up on the job queue (and I'm a software developer with background in data structures), and supposedly, it processes a job each time a request is made to the wiki. That, along with watching the job queue fluctuate leads me to believe that this change should have gone through within an hour or two, max. It has now been over 48 hours since the edit. This is why I wonder if there is a bug that prevented these changes from getting onto the job queue. In addition, I was under the impression that a job of this size (less than 200 updates) is too small to get onto the job queue, but I could be wrong on that last point. DeFaultRyan (talk) 20:01, 2 October 2008 (UTC)
    On another note, is there any kind of template substitution, inclusion, or some other voodoo that could cause this behavior? I really need to read up on the details of that stuff... DeFaultRyan (talk) 20:07, 2 October 2008 (UTC)
    It's definitely a job queue issue - I null edited Susquehanna River and it dropped off the list. --NE2 03:41, 3 October 2008 (UTC)

    Google Chrome issue?

    Can someone explain what this is? Is this an error in the VoA javascript (or Google Chrome's rendering thereof) or this an error in another Javascript rendering in Google Chrome? The Evil Spartan (talk) 02:28, 3 October 2008 (UTC)

    Perhaps it would be prudent to describe the issue you observe. I observe no issue, except that the logo happens to be loaded from a non-secure server, so there's a mixed content warning. — Werdna • talk 02:58, 3 October 2008 (UTC)
    The issue (non-secure diff link) seems to be the insertion of <script type="text/javascript"...; I guess that's why it's better to call scripts with importScriptURI() instead of document.write(). —AlexSm 03:41, 3 October 2008 (UTC)

    Printing problem

    When I printed the printable version of the Dinosaur article, all the images (except the lead image) turned out to be out of place and superimposed on text, while the Print Preview was fine. What do you think is the problem? If it helps, I'm using Firefox 3.0.3, a Canon S600, and I didn't print the references (printed only 22 out of 29 pages). Thank you. Eklipse (talk) 12:01, 1 October 2008 (UTC)

    Looks ok for me on Firefox 3.0.3, Mac OS X. --brion (talk) 20:24, 2 October 2008 (UTC)
    No I don't mean on screen, but on printed paper. Eklipse (talk) 15:23, 3 October 2008 (UTC)

    Template expert wanted

    WikiProject Geographical coordinates could do with some consulting services from a wikipedia template expert to advise & assist in the development of template driven categorisation of pages. The discussion leading to this request is here. thanks. --Tagishsimon (talk) 12:30, 2 October 2008 (UTC)

    Experts reportedly hang out at Wikipedia:Requested templates (the page is for modifying existing templates as well as new ones). -- John Broughton (♫♫) 22:54, 3 October 2008 (UTC)

    Very weird protection logs on WWE SmackDown vs. Raw 2009

    There appears to be some very weird logs to do with the WWE SmackDown vs. Raw 2009 article. The log (as shown on my watchlist) is also very weird, which involves a page move back and forth titles:

    1. (Move log); 10:27 . . IMatthew (Talk | contribs) moved Talk:WWE SmackDown vs. RAW 2009 to Talk:WWE SmackDown vs. Raw 2009 over redirect (revert)
    2. (Move log); 10:27 . . IMatthew (Talk | contribs) moved WWE SmackDown vs. RAW 2009 to WWE SmackDown vs. Raw 2009 over redirect (revert)
    3. (Protection log); 10:27 . . IMatthew (Talk | contribs) protected WWE SmackDown vs. Raw 2009 (moved WWE SmackDown vs. RAW 2009 to WWE SmackDown vs. Raw 2009)
    4. (Move log); 10:24 . . Josephjames21 (Talk | contribs) moved Talk:WWE SmackDown vs. Raw 2009 to Talk:WWE SmackDown vs. RAW 2009 (that is the title of the game, it's RAW not Raw on the cover.)
    5. (Protection log); 10:24 . . Josephjames21 (Talk | contribs) protected WWE SmackDown vs. RAW 2009 (moved WWE SmackDown vs. Raw 2009 to WWE SmackDown vs. RAW 2009)
    6. (Move log); 10:24 . . Josephjames21 (Talk | contribs) moved WWE SmackDown vs. Raw 2009 to WWE SmackDown vs. RAW 2009 (that is the title of the game, it's RAW not Raw on the cover.)

    The move was reverted as it is against the MOS and per previous titles. However the issue I'm having here is within the protection log. IMatthew (talk · contribs) and Josephjames21 (talk · contribs) are not admins thus cannot protect articles. Am I missing something here, or is there a new feature where if an article is moved, the protection moves with it? D.M.N. (talk) 15:11, 3 October 2008 (UTC)

    Yes it is a new feature, now when a protected page is moved the redirect at the old title is protected with the same level as the page and a log entry is created for the new title, as the logs don't move with the page. Though personally, I find it rather annoying and not especially helpful. Mr.Z-man 15:53, 3 October 2008 (UTC)
    New protection log entries will look better pretty soon, appropriate new messages have been added (but not live on Wikipedia yet), see bugzilla:15754. —AlexSm 16:07, 3 October 2008 (UTC)
    I only hope this automatic protection of the redirect does not add a dummy edit to the history of the redirect (as a manual protection does). This would sneakily make the page-move more difficult to revert and should be considered a bug if it does work this way. — CharlotteWebb 17:44, 3 October 2008 (UTC)

    moving multiple subpages

    List of asteroids is divided up into about 1600 subpages of the form "List of asteroids/x–y". Is there a way to move them all at once, the way we can with subdocs of talk pages?

    (They really should be titled "List of minor planets", as many of the objects on the list, such as Pluto, are not considered asteroids.)

    Thanks, kwami (talk) 20:33, 3 October 2008 (UTC)

    It would be a trivial task for an automated script: try asking at WP:BOTREQ. Happymelon 20:51, 3 October 2008 (UTC)
    Wouldn't it be easier to remove from these lists any items which are not "asteroids"? — CharlotteWebb 21:24, 3 October 2008 (UTC)
    Almost impossible. First, there's no agreed definition of "asteroid", even within the IAU. Secondly, it's a numbered list of minor planets, so removing some would leave a partial list; we'd need to supply tens of thousands of redirects for people looking up the missing numbers. Easier just to use the technically correct term. kwami (talk) 22:03, 3 October 2008 (UTC)

    Thanks for the link. I'll see if we get some discussion or debate before putting the request through. kwami (talk) 22:41, 3 October 2008 (UTC)

    How good is Wikipedia's quality control?

    We've started an article called Wikipedia:Quality control, but there's definitely something missing, though I don't have a clue what it is.

    The Transhumanist 21:48, 3 October 2008 (UTC)

    I assume you mean FlaggedRevisions. How is this relevant to VPT? Happymelon 21:51, 3 October 2008 (UTC)
    Well, there are a lot of things in the "Quality" topic of the editor's index that aren't covered by that page (I think calling it an "article" is a mistake; the term "article" is normally limited to a page in mainspace). More generally, describing the various initiatives to improve quality seems a bit odd to me, since so many things, from WikiProjects to noticeboards to article talk pages to mailing lists to protecting pages, is about improving quality (otherwise, why bother?) So by the time you exhaustively list everything that is done to affect quality, you'll have listed most of what happens here at Wikipedia. -- John Broughton (♫♫) 22:53, 3 October 2008 (UTC)

    A Stupid Question

    Hello. I've made a bit of a boo-boo (a mistake). I just created the page whose current title is A l tibawi. I wanted the title of the article to appear as A. L. Tibawi. If anyone could correct this for me and/or explain how to do it I would be very grateful. Thankyou. Langdell (talk) 01:40, 4 October 2008 (UTC)

    I moved to the subject's complete name. Pegasus «C¦ 01:54, 4 October 2008 (UTC)

    Rollback option

    Please see discussion for more information. - jc37 02:48, 4 October 2008 (UTC)

    "Remove titles"

    On the "View and edit watchlist" page, there is a button at the bottom, clicking on which removes the selected pages from the watchlist. So far so good. When a list is long, however (say over 100 items), it takes much scrolling to get to the button, and sometimes one only wishes to clear the list of mainspace or userspace pages or something else closer to the top.

    I therefore ask the honourable colleagues to comment on the suggested addition of a second button at the top. It's not exactly crucial, but I find that it would aid usability. (It's also standard practice in various sites; my e-mail inbox, for one thing, has buttons both at the top and at the bottom and it only lists 25 items per page.)

    The only problem I can see is how exactly to fit the button. But again, we have many resourceful people here; I'm sure someone will "figure it out". Waltham, The Duke of 01:29, 26 September 2008 (UTC)

    Maybe I should have posted this at the Proposals section? Waltham, The Duke of 09:07, 28 September 2008 (UTC)
    How about adding this code (untested!) to your monobook.js:
    if (wgCanonicalNamespace == "Special" && wgCanonicalSpecialPageName == "Watchlist") {
        addOnloadHook(function () {
            var form = document.getElementsByTagName("form")[0];
            if (!form || !/[?&]action=edit(&|$)/.test(form.getAttribute("action")))
                return;  // not the edit page
            var heading = document.getElementsByTagName("h2")[0];
            if (!heading) return;
            var buttonP = document.createElement("p");
            buttonP.innerHTML = '<input value="Remove titles" type="submit">';
            heading.parentNode.insertBefore(buttonP, heading);
        });
    }
    
    Ilmari Karonen (talk) 21:22, 2 October 2008 (UTC)
    In fact, I just saved a slightly fancier version as User:Ilmari Karonen/watchlisteditdoublesubmit.js, so just as simple importScript("User:Ilmari Karonen/watchlisteditdoublesubmit.js"); will do. —Ilmari Karonen (talk) 21:39, 2 October 2008 (UTC)
    And it works! Thank you very much; it's a small but meaningful improvement. Waltham, The Duke of 18:10, 4 October 2008 (UTC)

    Google Chrome partial page loads

    I am constantly getting partial page loading, and this is only happening in Google Chrome, it was happening in Vista (recently downupgraded) to XP and Ubuntu, where I run Chrome on Wine. The same issue occurs, I keep on getting cut off code snippets all over the page, sometimes making buttons like 'save page' not load or have missing/additional text/code in them, and often the top bar disappears altogether. I cannot think but to put this down to an incompatibility between Chrome and Wikipedia after all these reinstalls of different OSes, and I don't see how hardware could be particularly relevant here. Is there a Wikipedia side fix for this, or is it something that Google are going to have to be informed about for fixing in their next version? — neuro 09:46, 4 October 2008 (UTC)

    I have no such trouble editing with Chrome running under Vista. Some user scripts work a little funky with it, but Wikipedia pages generally come through just fine. —Ashanda (talk) 09:54, 4 October 2008 (UTC)
    Hm, not sure what is going on then. I have just had this, got a screenshot. I am going to post more soon, because there are more that aren't necessarily like this one: [17] — neuro 09:57, 4 October 2008 (UTC)

    date a page was created

    I am looking for a template or variable for a template that will display the date a page was created. Thanks GtstrickyTalk or C 15:43, 2 October 2008 (UTC)

    There are variables for the date/time of the "top" edit, but not for the "bottom" edit. Talk to the devs. — CharlotteWebb 16:37, 2 October 2008 (UTC)

    Is there a use case for it? — Werdna • talk 03:30, 6 October 2008 (UTC)

    diffs with page blanking, CSS irritant (Safari only, maybe)

    I've noticed this when viewing diff pages in Safari, where the current version is a page-blank, like this one: [18]. Safari for some unknown reason renders the right hand table cell (class="diff-ntitle") with a minimal width of 28 pixels, and gives the left hand table cell (class="diff-otitle") all the remaining space, creating a long unreadable column of text on the right hand side that's a pain to work with. Firefox renders the page correctly on my Mac, so this may be Safari specific. This should be fairly simple to fix, CSS-wise, but I can't find where the table class (diff) is defined in order to know where to put in a request. can someone point me in the right direction, or (blessed be) fix the problem? --Ludwigs2 06:05, 3 October 2008 (UTC)

    The file is diff.css, plus a small fix for very old Mozilla diff.js. Note that bugzilla:13875 "Empty diffs appear as long columns of text when viewed in Safari" looks very similar to your problem. —AlexSm 13:17, 3 October 2008 (UTC)
    How odd. Apparently Safari has a bug with table-layout:fixed tables where some columns are never used without colspan. It's not a Wikipedia or MediaWiki issue, this HTML is sufficient to illustrate the problem in Safari 3.1.2 (Mac and Windows (under Wine anyway)):
    <html>
    <head>
    <title>Test case</title>
    </head>
    <body>
    <table border="1" style="table-layout:fixed;width:100%">
     <col /><col /><col /><col />
     <tr>
      <td colspan="2">a b c d e f</td>
      <td colspan="2">a b c d e f</td>
     </tr>
    </table>
    </body>
    </html>
    
    If you save the above to a file and load it in the browser, it should have 2 columns each 50% of the width; in my tests I see one 50% column, one 25% column, and empty space for the remaining 25%. Someone who actually uses Safari should probably file a bug with Apple. Does it also occur with Google Chrome (which also uses WebKit for HTML rendering)? Anomie 16:30, 3 October 2008 (UTC)
    ok, I've created a temporary (user-by-user) solution - just add the code here to your monobook.css page. I'll post it over at bugzilla as well. it might cause minor problems with long URLs expanding the table, but that's the lesser of two evils. what I think is happening is that Safari is being over literal: it interprets the second table cell (with colspan=2) as though it were the third column, and applies the 2% width to it over-zealously (probably because of the 'fixed' element). that certainly looks like a Safari bug, and I'll go ahead and report it to apple. in the meantime, can anyone think of a more generalized solution to the problem? there are a couple that occur to me, but they aren't very elegant. --Ludwigs2 03:57, 4 October 2008 (UTC)
    Well, if the devs want to adjust the code to always output a dummy row with no colspans (i.e. <tr><td>&nbsp;</td><td>&nbsp;</td><td>&nbsp;</td><td>&nbsp;</td></td>) at the end of the diff table, that should work around it. It might work to add the same using javascript. Anomie 15:54, 4 October 2008 (UTC)

    (undent) ah, actually, that's a much better idea. this code:

    addOnloadHook( function() {
    	var url = document.URL;
    	if ( url.search(/&diff=/) == -1 ) return;
    	var tbody = document.getElementsByTagName("TBODY")[0];
    	var row = document.createElement("TR")
    	var td = new Array
    	for (i=0; i<4; i++) {
    		td[i] = document.createElement("TD")
    		td[i].appendChild(document.createTextNode(""))
    		row.appendChild(td[i]);
    	}
    	tbody.appendChild(row);
    } );

    works perfectly. I assume there's a shortcut way of browser sniffing on wikipedia? if someone could give me the code to check for Safari, this could be added as a patch for the problem. it could probably be run as is - I doubt it would goof things up in other browsers - but better safe than sorry... --Ludwigs2 20:48, 4 October 2008 (UTC)

    It might be even easier to just change the diff code not to use colspans in the first place. It doesn't particularly need them: <td>&nbsp;</td><td>&nbsp;</td> would do just as fine as <td colspan="2">&nbsp;</td>. —Ilmari Karonen (talk) 17:15, 5 October 2008 (UTC)
    yeah, you're probably right, though they may be preliminary moves in some future development. me, I'm happy I have a personal solution. that being sid, though, is there anything else that needs to be done to bring this to the attention of the developers, or will they pick it up here or over at bugzilla? --Ludwigs2 19:01, 5 October 2008 (UTC)
    The devs appears to be aware (there's a comment from Brion on the bug), someone just needs to code up the fix and commit it. I might take a look at it myself, though my experience with the diff code is limited. —Ilmari Karonen (talk) 19:46, 5 October 2008 (UTC)