Wikipedia:Village pump (technical)/Archive 78

From Wikipedia, the free encyclopedia
Jump to navigation Jump to search


"You have new messages (last change)." reappears after a short delay

Something is wrong with the new messages functionality for me. Maybe 30 minutes ago I started getting these notices when there were no new messages. Sometimes there's a short delay before the message stops appearing after I've checked my messages. Once it goes away, it reappears withing a few minutes. --Ronz (talk) 19:06, 2 July 2010 (UTC)

I'm getting this too. And here I thought I was just really popular. --Bongwarrior (talk) 19:11, 2 July 2010 (UTC)
This happens when the server lags. Gary King (talk · scripts) 19:35, 2 July 2010 (UTC)

Extreme server lag

I'm getting indications that articles I've been editing don't even exist, and that people I send Talk messages to are not even registered. Server lag? Everard Proudfoot (talk) 22:43, 2 July 2010 (UTC)

When I go to Recent changes, it keeps showing the same Recent changes from about half an hour ago. I have to purge my cache to see the new changes, but even then, if I go back to the Recent changes, it shows the old version yet again. Everard Proudfoot (talk) 22:45, 2 July 2010 (UTC)
I am getting the same thing. I am in the UK, if that makes any difference. The behaviour is almost random. Sometimes recently created/deleted stuff exists, sometimes it seemingly doesn't. Even history and log pages are affected. I have to keep hitting reload several times until I see what I expect. I am trying to do some new page patrol and both Twinkle and Friendly are getting very confused and messing things up when I use them. It is confusing me too. This is the first time I have ever come across anything like this. It needs to be flagged up as a serious problem.
My theory is that there are multiple cache servers and some of them are lagging and some are not. Hence it is more or less random which cache server you get served from and hence random whether you see recent data or not. The lagging servers do not show the usual red warning about the data being old. --DanielRigal (talk) 22:56, 2 July 2010 (UTC)

It looks like this needs to be exposed publicly, because I see lots of people blanking pages. Everard Proudfoot (talk) 23:17, 2 July 2010 (UTC)

Yeah, something really weird is going on with page histories. ctr-f5 helps sometimes, sometimes not. It never happened to me before, but I thought it was a cache issue on my end, but it doesn't appear to be. --Omarcheeseboro (talk) 23:19, 2 July 2010 (UTC)
If it is affecting people in different countries then it isn't some ISP running a broken transparent proxy. I think it can only be the Wikipedia infrastructure. --DanielRigal (talk) 23:23, 2 July 2010 (UTC)

Every time I go to Recent changes, I still get the page as of 22:10. I have to clear cache every time, sometimes twice. Everard Proudfoot (talk) 23:25, 2 July 2010 (UTC)

Well, I was patrolling Recent Changes (in Canada), and Gavriil Malish is viewable half the time. Sometimes, it's visible, sometimes, it says the page does not exist. The history page does not exist, yet it show up on my watchlist. Brambleclawx 23:54, 2 July 2010 (UTC) Add Refreshing helps, but that link I provided above is switching from blue to red, and it obviously hasn't been deleted... Brambleclawx 23:55, 2 July 2010 (UTC)
Does anybody know if this is being looked into? It is causing lots of screwups and making some editing tasks close to impossible! --DanielRigal (talk) 00:23, 3 July 2010 (UTC)
I think your theory makes sense Daniel; I was looking through special:logs/newusers, and when I first got there, it showed me users who had registered 2 minutes ago, then I refreshed and it showed my users who had registered 2 hours agp. I have no idea if anyone is looking into it. Brambleclawx 00:25, 3 July 2010 (UTC)
Yes, it is being looked into. With the amount of users Wikipedia has, it takes about three seconds for someone to complain about server lag to the server admins. --Deskana (talk) 00:25, 3 July 2010 (UTC)

It appears to be an issue with one of the database servers only (db40). I've alerted our sysadmins to this and it should be sorted out soon.--Eloquence* 00:25, 3 July 2010 (UTC)

The issue should be resolved.--Eloquence* 00:44, 3 July 2010 (UTC)
Appears to be ok now. Brambleclawx 00:48, 3 July 2010 (UTC)
Yes, It seems to be. Responsiveness has also improved significantly. Maybe everybody has stopped pressing F5 every couple of seconds. ;-) Thanks. --DanielRigal (talk) 00:49, 3 July 2010 (UTC)
Thanks, Eloquence. Everard Proudfoot (talk) 01:34, 3 July 2010 (UTC)
I'm still having problems. After saving an edit the page is not updated. Using purge doesn't work. Saving the page again doesn't work. I was working on User:Droll/subpages/testcases.  –droll [chat] 03:22, 3 July 2010 (UTC)

False alarm. My bad.  –droll [chat] 03:25, 3 July 2010 (UTC)

New messages bar spazzing

I'm getting the new messages bar really frequently even after I check the message. Its been about and hour since the real new message appeared. Derild4921 23:26, 2 July 2010 (UTC)

Wait for it to fix itself. --Deskana (talk) 23:30, 2 July 2010 (UTC)


This gadget is apparently redundant since there's an "edit area font style" option in preferences. vvvt 00:32, 3 July 2010 (UTC)

If an edit conflict happens, the lower textbox shown is a textarea but it is not #wpTextbox1. The "edit area font style" affects only the main text box; the Gadget affects all textareas. The same is true for scripts such as Twinkle and closeAFD. So the two are not 100% redundant. PleaseStand (talk) 02:26, 3 July 2010 (UTC)


So it seems that there is a function named #tag somewhere. It's used in {{nowiki}} and in several other templates I've looked at. Does any one know if it's documented anywhere. I would expect it to be at MediaWiki somewhere but I can't find it. The syntax seems to be {{#tag: (tag name) | (string) }} —Preceding unsigned comment added by Droll (talkcontribs) 02:24, 3 July 2010 (UTC)

It is documented at Help:Magic words, fourth from the top. In essence, it allows a tag to be used (such as nowiki or imagemap) without the HTML opening and closing < >. Intelligentsock 02:30, 3 July 2010 (UTC)
Thanks! I guess I just looked in all the wrong places. –droll [chat] 02:34, 3 July 2010 (UTC)
(edit conflict)mw:Help:Magic words#Miscellaneous. It is most commonly used when you want the parser to expand template parameters before applying the tag's effect (especially nowiki's). For example, {{nowiki}} uses {{#tag:nowiki|{{{1}}}}} to accept a parameter and then nowiki it (producing code such as <nowiki></nowiki> instead of <nowiki>{{{1}}}</nowiki> from the parameter PleaseStand (talk) 02:39, 3 July 2010 (UTC)

Template for colors

I'm looking for a template to easily equip a template with hex no.s such as in Template:Infobox military unit. Where do I find it? E.g. {{hexcol|#######|########}}. -- (talk) 23:36, 2 July 2010 (UTC)

I am not sure what you are trying to accomplish. Can you clarify what such a template would do? Intelligentsock 23:51, 2 July 2010 (UTC)
I want to put a template to the parameter "colors=HERE" in the Template:Infobox military unit (inclusion of course). The desired template (I have seen it but I forgot it) displays one or two colors by inserting the hex code. -- (talk) 23:59, 2 July 2010 (UTC)
It's {{Color box}}, example {{color box|blue}} {{color box|yellow}} gives           and <code>{{color box|blue}} {{color box|yellow}}<code> gives          . --Scriberius (talk) 11:08, 3 July 2010 (UTC)

OGG/OGV Buttons Blank on Mac Firefox

I hope I'm not reposting this question.. When I access an OGG or OGV file with Firefox on my Mac, the buttons are blank -- see this image, taken from this article. It doesn't happen with Safari; that player is sleek and works fine. Perhaps there is something I need to adjust on my browser? I expect other users are having this problem, too. Cheers! Scartol • Tok 12:55, 3 July 2010 (UTC)

Wikipedia Waiting Time

I raised this issue in mid-June. The problem persists. It happens whether I'm logged in or not logged in. It seems to get worse the more I use WP in a given session, but that could be coincidental. When I click on something, Firefox 3.6.6 gives me a "waiting for . . ." message in the status bar. It can sometimes take 20-40 seconds to come back. Other times, no wait at all. If there is an excessive wait and I click on the link again, I usually get an instantaneous response. On some things, though, I'm reluctant to click twice because I'm not sure what effect it will have on the database, i.e., saving editing changes. My Internet download connection speed is above 10Mbps. I don't have this problem on other sites. I simply don't know what to do, but editing has become very frustrating because of it.--Bbb23 (talk) 15:38, 3 July 2010 (UTC)

Clicking save twice shouldn't cause any problems database-wise. At worst you might cause an edit conflict with yourself. Try doing a ping test to the Wikipedia servers to check for packet loss. Mr.Z-man 17:43, 3 July 2010 (UTC)
When I ping, there is no packet loss. If there's another server you want me to try, let me know.--Bbb23 (talk) 22:24, 3 July 2010 (UTC)

How can I stay signed in?

I followed advice I was given here when I found out my computer still had pages I had visited even though I appeared to have no history. Now, though, I seem to have to sign in everywhere, including Wikipedia. It's not that big a deal since my computer has the software which fills in passwords automatically, though without that I was unable to access one of my email accounts because of rules they have about passwords (though I still remember that crazy password for that address). I can remember the password if the password software messes up, since I use Wikipedia at libraries, but it's just annoying.Vchimpanzee · talk · contributions · 18:25, 3 July 2010 (UTC)

Try deleting cookies, that often 'resets' the need to constantly log in (of course, you'll still have to log in once). ♫ Melodia Chaconne ♫ (talk) 20:12, 3 July 2010 (UTC)

How to feed pages such as 'My contributions' to a self-written editor?

Hi, what I'd like to do, for example, is have a little program/macro that can interface with non-editable Wikipedia pages such as 'My contributions' (or, someone else's contributions). Ideally it would read and collect (in memory or as an appendable file) a pageful of data at a time, then virtually click on 'next 100' until the end. I can write code, but I don't know how to interface with Wikipedia, or what language is required. I'm sure this is spelled out somewhere, so if you just point me to the documentation I can figure out what I might be in for. Thank you, CliffC (talk) 22:01, 3 July 2010 (UTC)

The best option would be to use the API, located (along with some documentation) at Several output formats are available. For example, to get your 10 newest contributions in the XML format, use the following URL: Or you could use one of the tools or libraries that are already there. For example, AWB can get user's latest contributions and save them to a file. Svick (talk) 22:30, 3 July 2010 (UTC)
Thank you for that very complete answer. I never thought at all about what goes on with the bots here. I looked at that documentation and got some crazy ideas. I'll dig in when I have the time and want to be frustrated/exhilarated by getting something to work. Best, CliffC (talk) 13:22, 4 July 2010 (UTC)

Wikipedia down

Anybody care to venture as to why the insecure wikipedia is down? The secure still seems to be partially up. —Preceding unsigned comment added by (talk) 01:44, 5 July 2010 (UTC)

The rumors are that the airconditioning in the Florida center has failed, causing the machines to shutdown. It is being worked on. —TheDJ (talkcontribs) 01:51, 5 July 2010 (UTC)
Twitter has a feed of the techlog that you can follow. —TheDJ (talkcontribs) 01:54, 5 July 2010 (UTC)
For some reason, I can get on the site, but I can't log in. -- (talk) 02:15, 5 July 2010 (UTC) is up, but skinless (bits.wikimedia is down). (talk) 02:18, 5 July 2010 (UTC)
Could I suggest the very professional failure page that directs you to a Google search include a link to the Twitter feed and/or another Wikipedia Status page that I seem to recall but haven't looked for this time. Mark Hurd (If I could log in) (talk) 02:43, 5 July 2010 (UTC)
Skins are back up, but I still can't log on. —Preceding unsigned comment added by (talk) 02:46, 5 July 2010 (UTC)
Yeah, I tried to log onto my main account and my declared sock that I use on public computers, but neither one will work; I'm running Monobook, so I guess that's the problem. I'm having to use someone else's computer to type this, since mine is set to log in automatically. (talk) 02:48, 5 July 2010 (UTC)
Special:UserLogin login now working. — Richardguk (talk) 03:26, 5 July 2010 (UTC)
The insecure Wikipedia is now working. December21st2012Freak Happy Independence Day! 03:32, 5 July 2010 (UTC)

WikiNews error page still refers to Meta

(I know this is the wrong place, but it is up enough to receive edits :-) )

The error page for still refers to Meta in its title, heading and icon. Mark Hurd (talk) 02:49, 5 July 2010 (UTC)

about PWB

I would like to know which function do we use when copying a content from a wikipage or from a textfile to a wikipage or a textfile. Thanks in advance, --Jagwar - (( talk )) 10:56, 28 June 2010 (UTC)

Can you clarify what you are trying to do? Your question is worded vaguely. — Richardguk (talk) 02:15, 29 June 2010 (UTC)
I sould like to create a program that could generate several wiki pages from a single file in TXT format. Currently, I do not know how to use the script that generates constant errors when used (I don't know how to use it and what is the syntax used to make the file readable by the program) ; I am still not familiarized with the Python language. --Jagwar - (( talk )) 10:01, 2 July 2010 (UTC)
This probably isn't the right place to ask. But there is some information and an example at meta:Pywikipediabot/ and a mailing list at
You'll need to have configured your general settings in Then it looks like the simplest settings would require a file named dict.txt with the content encoded as UTF-8 and set out like this:
'''Title of first article'''
Wikitext of first article
goes here
'''Title of second article'''
Wikitext of second article
goes here

Richardguk (talk) 15:03, 2 July 2010 (UTC)

Thanks you for the indication :). But I would like to know : sometimes the generators "says" :
X:/.../ -start:{st} -end:{ed}
reading dict.txt
End of file.
Is it normal ? --Jagwar - (( talk )) 19:27, 2 July 2010 (UTC)
You probably need to post to for specific advice. — Richardguk (talk) 19:45, 2 July 2010 (UTC)
I'll see that. Thanks you and good continuation. --Jagwar - (( talk )) 12:20, 5 July 2010 (UTC)

File:Information processing system (english).svg


This might be more appropriate on Commons, but I got a positive result here the last time I had the same problem. This file was moved to Commons a few days ago, but thumbnail generation seems to be broken: is there a way to force it to be regenerated? The SVG itself looks to be okay once it's clicked through. Chris Cunningham (not at work) - talk 09:17, 30 June 2010 (UTC)

Something's broken: Error generating thumbnail Error creating thumbnail librsvg-ERROR **: _rsvg_acquire_xlink_href_resource called for external resource: " base: (null) aborting... Appending a number to the end of the URL as described at WP:Purge#For images to regenerate the thumbnail does not seem to fix the problem. And I checked the source code and there are no external links within the SVG. PleaseStand (talk) 13:19, 30 June 2010 (UTC)
I suspect the brackets in the filename are the problem. Not all characters are allowed. I've tagged the image on commons with {{rename|Information processing system English.svg}}. When an admin or filemover actions that, we'll see if that clears it up. LeadSongDog come howl! 14:39, 30 June 2010 (UTC)
Fixed by commons:User:Mormegil — it was an internal link using incorrect syntax (diff). PleaseStand (talk) 19:54, 30 June 2010 (UTC)
This file was definitely problematic. It had transformation matrices on groups, which caused the rendered size of text to be incorrect. An ungroup/group of all groups fixed this. PleaseStand (talk) 20:44, 30 June 2010 (UTC)

──────────────────────────────────────────────────────────────────────────────────────────────────── Hmmm. Was this caused by the upload process, or just a bug in the original file? Chris Cunningham (not at work) - talk 21:19, 30 June 2010 (UTC)

Which problem are you referring to? The one that caused rsvg to crash or the one that caused incorrect text sizing? PleaseStand (talk) 02:43, 3 July 2010 (UTC)
The former. As I say, the file had previously been uploaded here and thumbnail generation worked fine. Chris Cunningham (not at work) - talk 22:15, 5 July 2010 (UTC)
Most likely the SVG parser/renderer has become more strict since the initial upload. Since thumbnails are cached, thumbnails that already exist will never re-render and the problem will never show. When the file was uploaded to another location, all thumbnails had to be generated again, making the problem very clear. —TheDJ (talkcontribs) 22:54, 5 July 2010 (UTC)

Image height?

I'm hoping there is a way to discover the height of an image assuming that the name of the image file and the width are known. I realize this not likely but I thought I'd ask anyway. It would be useful in a template I've been working on. –droll [chat] 04:53, 1 July 2010 (UTC)

P.S. So all I need to know is the aspect ratio. –droll [chat] 12:24, 1 July 2010 (UTC)
I don't think you can do that. Won't one of the alternative ways to specify image size (xHeightpx or WidthxHeightpx) help you? Svick (talk) 12:31, 1 July 2010 (UTC)
Perhaps {{panorama}} might be of help? ---— Gadget850 (Ed) talk 13:04, 1 July 2010 (UTC)
And you can of course always query the api, but not from a template, which is what I suspect you are after. —TheDJ (talkcontribs) 13:51, 1 July 2010 (UTC)
  • TheDJ is right. {{Site plan}}, a.k.a {{Lageplan}} with the name transliterated, needs to know the height. I'm not sure how useful the template really is. It's like {{Superimpose}} but uses percentage coordinates instead of pixel coordinates. If the could be determined in the template it would be a good alternative for non-techies. Firebug and similar tools provide pixel detentions and the aspect ratio can be computed or the editor could just download an image, inspect the image properties and do a little math. These methods are not exactly user friendly.

    P.S. The template could use a better name. –droll [chat] 02:05, 3 July 2010 (UTC)

The MediaFunctions extension provides magic words for this (#mediaheight and #mediawidth) but is not currently installed on English Wikipedia. So the information cannot be obtained by templates on this wiki. — Richardguk (talk) 06:30, 3 July 2010 (UTC)
Thanks for the information. It's good to know it's out of my control.  –droll [chat] 07:45, 5 July 2010 (UTC)

Link doesn't work right

Maybe this is a strange way to do it, but I put links to web pages I want to go to in emails to myself so I can do it from any computer.

There was a link to [1]--at least that's what is visible in the email. [2] is what came up.

Likewise, [3] appears in the email, but clicking on it produces [4]. You'll probably send me to the Computing Reference Desk since it's not likely a Wikipedia problem, but I don't see how something like this could happen.Vchimpanzee · talk · contributions · 17:44, 1 July 2010 (UTC)

Is it an HTML format email? If so, its likely the actual HREF tag is wrong, while the text is showing what you really wanted. Do a view source. -- AnmaFinotera (talk · contribs) 18:20, 1 July 2010 (UTC)
I don't have a "View Source" button.Vchimpanzee · talk · contributions · 18:43, 1 July 2010 (UTC)
Email me if you want to establish a link outside of Wikipedia mail, then you can forward your problem to me when I respond. (Yes, that will unveil your email address) I'm suspecting "operator error" as the first port of call, but it's hard to say without the raw text. "View Source" would probably only work for web-based email clients, but a host-based client should allow you to save a message with the headers and MIME-encoding intact. For instance Outlook Express lets you save a .eml file which has all the good stuff. Franamax (talk) 20:38, 1 July 2010 (UTC)
All I've ever done is type or click (or copy and paste) on the address of the web site for the email service.Vchimpanzee · talk · contributions · 20:41, 1 July 2010 (UTC)
OK, so it sounds like you do use a web-based email client. Then you should be able to somehow or other view the HTML source of the page to see what is happening behind the scenes. Generally when I encounter a client who has a problem and I recommend a diagnostic procedure which they decide not to follow, I close the ticket as a PEBCAK, no offense meant. You could try pasting the copied link into your browser address bar to be sure you harvested the link you wanted. Franamax (talk) 22:19, 1 July 2010 (UTC)

"Somehow or other" doesn't answer the question. It doesn't seem to provide me with a way to "View Source".Vchimpanzee · talk · contributions · 22:28, 1 July 2010 (UTC)

What it sounds like to me is that instead of copying the existing source of the page when you're mailing it, you're actually sending a pointer to the current page. But without further details (what "email service" you're talking about, or the actual page source) it's difficult to be sure. To view the page source, click on View -> Page Source in your browser's menu bar. Chris Cunningham (not at work) - talk 08:46, 2 July 2010 (UTC)
Okay, we're getting closer. I don't see anything that says "View Source".Vchimpanzee · talk · contributions · 15:06, 2 July 2010 (UTC)
I figured it out. Right-click on the page, then click "View Source" ... except it doesn't seem to give me anything that's in the actual email. I recognize stuff that's displayed along with the email itself.Vchimpanzee · talk · contributions · 17:53, 2 July 2010 (UTC)
The above suggestion is, in effect, that your email is displaying the HTML equivalent of a piped link, <a href = "" ></a> (which in Wikitext is equivalent to or loosely equivalent to a misleading piped link like [[the_old_version|the_new_version]]). So just delete the old stuff and paste the URL in from plain text afresh, to ensure that you are not inadvertently pasting in hidden code. (If you are copying from HTML, such as your web browser, and don't know how to make sure it is only plain text when you paste it in, first paste it into a plain text editor such as Notepad, then copy it from Notepad and paste into your email textarea.) — Richardguk (talk) 18:17, 2 July 2010 (UTC)
I can copy it here in the box I am typing in right now. The links probably won't be links I can click on, though.Vchimpanzee · talk · contributions · 18:27, 2 July 2010 (UTC)
Oh, wait, I can click on them. And they work. It's the OTHER email address that's the problem because I have to send the links BACK, and I can't click on plain text emails FROM that address. How this email ended up not being plain text, I have no idea.Vchimpanzee · talk · contributions · 18:27, 2 July 2010 (UTC)

For the umpteenth time - this is not your personal computer help desk. This is for discussing technical issues regarding English Wikipedia. OrangeDog (τ • ε) 13:03, 5 July 2010 (UTC)

Secure problems

I notice that throws an error about containing unencrypted content in Firefox (latest update). Can this be fixed? (talk) 05:00, 4 July 2010 (UTC)

That is because the images are served as non-secure. Peachey88 (Talk Page · Contribs) 05:06, 4 July 2010 (UTC)
There is some previous discussion here. And gives a rather ugly "fix" that blocks the images. HumphreyW (talk) 06:13, 4 July 2010 (UTC)
Currently, the Wikimedia foundation does not have the financial resources to deliver media in an encrypted manner...this is why you are getting a msg about unencrypted content.Smallman12q (talk) 12:12, 4 July 2010 (UTC)
Well not only that, it is probably more of a time and manpower issue. That could be solved with throwing money at it, but apparently it isn't that high up on the priority list. Using the financial resources to fix problems like this really requires them to be of top priority. —TheDJ (talkcontribs) 12:29, 5 July 2010 (UTC)

mwEmbed + Twinkle = video not loading

When I enable the mwEmbed and Twinkle gadgets, videos (such as the one on today's featured article) don't load. Firefox's error console says:

Error: _this.moduleLoadQueue[moduleName] is undefined
Source File:,,mw.EmbedPlayer,,$j.ui,mw.PlayerControlBuilder,$j.fn.hoverIntent,$j.ui.slider,mw.PlayerSkinKskin,,$,,mw.TimedText,,mw.SwarmTransport,mw.EmbedPlayerNative&uselang=en&urid=r136
Line: 55

Chrome says:

ControlBuilder,$j.fn.hoverIntent,$j.ui.slider,mw.PlayerSkinKskin,,$,,mw.TimedText,,mw.SwarmTransport&uselang=en&urid=r136:55Uncaught TypeError: Cannot set property 'loaded' of undefined

Without Twinkle, mwEmbed works. -- Nx / talk 13:42, 5 July 2010 (UTC)

Probably clashing global variables or something. I'll ask mdale to look into it, though with Wikimania arriving, it might take a while before it is fixed. —TheDJ (talkcontribs) 14:06, 5 July 2010 (UTC)
User:AzaToth/morebits.js messes with the Array type's prototype. Here are the culprits. -- Nx / talk 15:30, 5 July 2010 (UTC)
That means mwEmbed is broken! AzaToth 15:35, 5 July 2010 (UTC)
Yeah, it's trying to use an object as a php-style associative array: for(var moduleName in moduleRequestSet) -- Nx / talk 15:38, 5 July 2010 (UTC)
Don't you mean an "Array ([])" as a "Object ({})"? AzaToth 15:40, 5 July 2010 (UTC)
Array() is an object :) But yeah, Javascript's confusing / foreach implementation is to blame here. -- Nx / talk 15:41, 5 July 2010 (UTC)
This was fixed in r69064, but it has not yet been deployed. —TheDJ (talkcontribs) 23:05, 5 July 2010 (UTC)

Image not displaying, as if deleted

I noticed today that at Crucifixion and Crucifixion in the arts, the image file File:Dionysus Crucifixion.gif displays as though the file has been deleted (a red link), as it does here, and yet the file does indeed still exist at Commons commons:File:Dionysus Crucifixion.gif. I cannot find anything wrong with how the file name is entered on these pages. Am I missing something, or is there a bug? Thanks. --Tryptofish (talk) 19:01, 5 July 2010 (UTC)

We had the same issue for the WPVG ad, see Wikipedia talk:VG#WikiProject ad redlinked. Anomie was able to fix it with some trickery, apparently using a full url with &action=purge: I would assume this is related to the recent outage. –xenotalk 19:03, 5 July 2010 (UTC)
Thanks. I see that fixes it at this page, but not at the two articles named above. Sorry for my relative lack of technical fluency, but how do I do that purge at the Commons page? --Tryptofish (talk) 19:20, 5 July 2010 (UTC)
They just needed a purge as well. The easiest way is to enable the Gadget called UTCLiveClock which is available both here and at commons, and then click the clock that appears in the top right corner which acts as a purge link. –xenotalk 19:23, 5 July 2010 (UTC)
Thanks for both the help and the gadget advice. I've just added the gadget. --Tryptofish (talk) 19:29, 5 July 2010 (UTC)

A bug (?)

Copied from misc page:

In my post at the talk page of WikiProject Chemistry i tried to bring to notice an issue which in my view affects every WikiProject and is thus a serious bug. Please see the post for details. I hope someone has a solution. --Siddhant (talk) 20:41, 5 July 2010 (UTC)

It is not a bug, it is normal behavior. A category is for performance reasons only really aware of it's own level and not of the pages in subcategories. —TheDJ (talkcontribs) 23:08, 5 July 2010 (UTC)

Vector Search Problem

I initially posted this at the Help Desk, although I kind of had it in the back of my head that I should post this here. When I search for certain items, it seems that the search bar brings up strange results. For instance, searching "Doctor Who (" turns the "(" into an "a" (an example here in this segment of a screen cap) and a search for "Doctor Who h" turns the "h" into an ":" (i'm using Doctor Who examples because that's what I was indeed looking for when I discovered this). I get the bug with other searches (like, "Star Trek o") or when I search for a single word and a space (like "Star " turns "Starting Pitcher" into "Star ing Pitcher" and "Starbucks" into "Star ucks" and "Life " turns "Life Imprisonment" into "Lifebimprisonment"). In fact, I just noticed this now, but searching for "life" then two spaces turns the searches into "Lifebo+(rest of search result phrase beginning with "Life"), three spaces turns the search to "Lifeboa+(rest of search result phrase beginning with "Life"), eg "Lifeboaimprisonment", and so on until "Life+17 spaces" full reveals "Lifeboat (shipboard)" over the search results. I get the same thing if I search words like "time", "star", "rose" and "royal" and a bunch of spaces or any spaces (even sometimes one). Bizarre! Doc StrangeMailboxLogbook 16:42, 3 July 2010 (UTC)

By the way, I have indeed tested this on another computer just to check if the problem was with my computer and/or keyboard and this bug indeed shows up on that computer when the same search queries (ie "Doctor Who (") are typed into the search bar. Others who replied to my message at the help desk noticed this bug and they mentioned that it only appears in the Vector skin. Doc StrangeMailboxLogbook 16:42, 3 July 2010 (UTC)
Yes, I've just noticed this too. Does anyone know what's going on?--Kotniski (talk) 09:36, 5 July 2010 (UTC)
Which browser ? No such problems using Safari here. —TheDJ (talkcontribs) 12:24, 5 July 2010 (UTC)
It's happening for me in IE.--Kotniski (talk) 12:38, 5 July 2010 (UTC)
Version ? Because they kinda differ a LOT. —TheDJ (talkcontribs) 22:59, 5 July 2010 (UTC)
The problem occurred for me in Firefox 3.6.6, IE6 and IE7 on multiple computers. Maybe this doesn't affect Safari. I wouldn't know. Doc StrangeMailboxLogbook 01:42, 6 July 2010 (UTC)
Filed as bugzilla:24282TheDJ (talkcontribs) 10:52, 6 July 2010 (UTC)
P.S. Everyone could have been saved a 3 day 'discussion'. If people had started by reporting their browser setup. Perhaps it is time to start new pages with a preloaded form, so that it becomes easier for people to properly report a problem. —TheDJ (talkcontribs) 10:52, 6 July 2010 (UTC)
That wasn't a personal criticism btw, it is quite a common problem in bugreporting. I have now added an editnotice in hopes that people will be reminded of these things. —TheDJ (talkcontribs) 11:11, 6 July 2010 (UTC)

New problem when editing with Chrome

I have no problems reading pages, they look normal. But when I try to edit, the edit page opens normally but within a second changes to one in which the screen is almost completely taken up with the edit field. I can see the edit toolbars (wikiedit, etc), the edit summary field, and just under that the top half of the Save Page, Show Preview, etc. buttons (I can't even remember, should the 'content that violates copyright be above these as they are in the IE browser I'm using right now?). Nothing on the left or right side of the edit field at all, and no scroll bars. It suddenly happened a few minutes ago while I was using Chrome, and all other pages look normal, it's only when editing that I have the problem. Dougweller (talk) 10:45, 6 July 2010 (UTC)

Sounds like the fullscreen mode of WikEd. In the top-right bar of WikEd click the bottom-right icon, it looks a bit like the maximize window icon of MS Windows. —TheDJ (talkcontribs) 11:00, 6 July 2010 (UTC)
Thanks, that's why, as I just discovered, clicking on preview fixes it. I clicked on that be accident, trying to click on the icon above to get down to the save page, etc buttons. Thanks again. Dougweller (talk) 11:04, 6 July 2010 (UTC)

Template sandbox and testcases templates

In my opinion the page headers provided by {{template sandbox notice}} and {{template test cases notice}} are very useful. Is there anyway that the information can be displayed on such pages system wide and without user intervention. If such a thing were to be implemented, maybe there should be a way to disable it in preferences as someone would probably object. Maybe not.  –droll [chat] 08:04, 5 July 2010 (UTC)

You mean like edit notices, but then without editing? In that case, no such a thing does not exist. These boxes are auto-added by the {{documentation}} template however. —TheDJ (talkcontribs) 23:02, 5 July 2010 (UTC)
You have lost me. Are you saying the the documentation template on the template page causes some message to appear on a different sub-page. I don't think so. So lets start over. The {{template sandbox notice}} template displayed a message on the sandbox page. At least I think it does. What I am asking is is it possible to display the message without the template. All kinds of messages appear on some of my pages. So it seems there is a mechanism that in place to the kind of work I'm talking about. It's sleepy time. talk more tomorrow.  –droll [chat] 11:17, 7 July 2010 (UTC)

anon-only options showing up on named users?

Resolved: Local issue only. –xenotalk 12:45, 7 July 2010 (UTC)

Strange, I have the option to

  • Block anonymous users only, allowing registered users to edit as normal.

when looking at the block form for a named user? –xenotalk 23:34, 6 July 2010 (UTC)

I'm not seeing it, myself. All accounts, or just one user in particular? —DoRD (talk) 00:03, 7 July 2010 (UTC)
This option is always there, but a script hides it when you enter a username. Is javascript enabled? Is it throwing any error messages? This is the script that does that: -- Nx / talk 00:15, 7 July 2010 (UTC)
Thanks, this explains it. –xenotalk 12:46, 7 July 2010 (UTC)
Works for me (Safari 5.0 under Mac OS X 10.6); the problem's certainly not universal. {{Nihiltres|talk|edits|}} 00:32, 7 July 2010 (UTC)

Making Switch parser active only when there is a certain entry in an template

Resolved: Fixed using #ifeq:. MC10 (TCGBL) 19:56, 7 July 2010 (UTC)

The m:Template:Extension changes color depending on what an editor enters as the property of {{{switch}}}

Unfortunately, the parser instructions at m:Help:Extension:ParserFunctions are not very helpful.

So if I wanted to make a section of an infobox template appear or not appear dependent upon what the editor enters in, how would I do this?

Say I have this template:

|header1 = {{#switch: {{{{{found|}}}|Idaho={{{found}}}|}} }}
| label1 =  Location
| data1 =  {{#if:{{{found|}}}|{{{found}}} }}

I know this coding is incorrect....

How would I make it that, ONLY if {{{found}}} = Idaho then header1 will be visible? But if {{{found}}} is something else, it will not be visible?

Thank you so much in advance for your help, I am so confused and don't know where else to ask on wikipedia. God bless. Adamtheclown (talk) 05:40, 7 July 2010 (UTC)

maybe try:
| label1 =  Location
| data1 =  {{{found|}}}

Bawolff (talk) 05:57, 7 July 2010 (UTC)

(ec) Bawolff on IRC was kind enough (again) to write:
after reading your post on the vp, sounds like your more after #ifeq than #switch:
{{#ifeq:{{{param}}}|some value|what to print if some value = param| what to print if not}}
Adamtheclown (talk) 06:00, 7 July 2010 (UTC)
Thanks Bawolff! you are so wonderful. Adamtheclown (talk) 06:00, 7 July 2010 (UTC)

Content Encoding Error

Any time I try to view a page that does not exist on Wikipedia (a redlink), I have gotten this message. This just recently started happening. Anyone got answers?  A p3rson  20:32, 7 July 2010 (UTC)

P.S.: Firefox 3.6.4, running on Vista Home Premium x64

Seems to be a known problem with Firefox, with somewhat vague advice for fixing your PC config given here: (talk) 21:00, 7 July 2010 (UTC)

Which license for templates?

Which license do wiki templates fall under? The wiki content license (Creative Commons) or the wiki software license (GPL2)? What about other sites that use MediaWiki? SharkD  Talk  21:29, 7 July 2010 (UTC)

Templates are but data, CC all the way. Gwen Gale (talk) 21:30, 7 July 2010 (UTC)
So, does that mean if a template were copied from Wikipedia to another MediaWiki wiki, some attribution would be required? Also, is this stated somewhere that where people can easily find it? SharkD  Talk  22:29, 7 July 2010 (UTC)
At the Terms of Use linked to at the bottom of every Wikipedia page. --Cybercobra (talk) 22:46, 7 July 2010 (UTC)
Usually when I copy past something, i link to the specific revision id from where I copied, in my edit summary. Then at the very least you have a reconstructible history. —TheDJ (talkcontribs) 00:49, 8 July 2010 (UTC)
Seems weird that templates are CC (well, for some of them it makes sense...), but whatever. SharkD  Talk  05:22, 8 July 2010 (UTC)

External links


Usually, external links are written in the way [ Google], which produces Google. But what happens when there are brackets inside the URL, and the link gets broken? Is there an alternative way to link? MBelgrano (talk) 02:21, 8 July 2010 (UTC)

Use percent-encoding. For example, replace "[" with "%5B" and "]" with "%5D". Anomie 02:53, 8 July 2010 (UTC)
(ec) You'll need to replace the offending characters as described here or use the {{urlencode:string}} magic word as described here. —DoRD (talk) 02:57, 8 July 2010 (UTC)
Thank you, the replacement worked MBelgrano (talk) 00:24, 9 July 2010 (UTC)


Resolved: Resolved by purging

I was looking something up on Hydrogen and noticed a bunch of redlinks in the periodic table in the top right. This turned out to be a failed link to Image:Transparent.gif - which is linked to by a plethora of pages. I then noticed I wasn't logged in, so I logged in. The file appeared. I logged out. The file redlinked.

Something isn't right here - I tried switching browsers to see if it was cookie-related, but the problem persists (newest Opera and IE7). The gif simply isn't there if you're not logged in.

I can imagine that some features and tools should only be presented to admins/bureaucrats/etc, so there's levels of things that are and aren't presented to different types of users. But why this applies to a utility image is beyond me - also why logged-out (ie IP) users not able to access some things that are accessible to logged-in users without any rights, except things like (semi)protection which are handled serverside. --Firien need help? 09:37, 8 July 2010 (UTC)

A while ago, that image suddenly went missing. These pages when you are not logged in are from the squid caching servers. I have purged the page of the image, and i think that should clear the pages from the squid services. —TheDJ (talkcontribs) 10:44, 8 July 2010 (UTC)
Cheers, that seems to have fixed it --Firien need help? 12:07, 8 July 2010 (UTC)

String handling

Is there any means of manipulating strings (e.g. chopping the first character)? I need something like this to manipulate a parameter. --Redaktor (talk) 14:24, 8 July 2010 (UTC)

We have an entire category of string manipulation templates. Although I don't have time to check all of them, {{str crop}} seems to be something along the lines of what you're looking for, only backwards. If there's no alternative you could try {{str_right|{{{1}}}|{{#expr:{{str_len|{{{1}}}}}-{{{2}}}}}}} (which is that template with str_right instead of str_left). Bear in mind that all of these templates are expensive and somewhat capricious. Intelligentsock 14:39, 8 July 2010 (UTC)
Many thanks. Yes, that is what I was looking for. In fact, I think {{str right|{{{1}}}|2}} will chop the first two characters, which is what I really need. --Redaktor (talk) 15:03, 8 July 2010 (UTC)

"Do not show page content below diffs" and pending changes

Pending changes seems to ignore this setting, adding &diffonly=0 to the URL. Is there any way to show only the diff to avoid the rather long rendering time when reviewing? (It seems that there should be a setting for this somewhere that furthermore should be independent from the other setting, and I still get the reviewing interface when deleting the &diffonly=0 part of the URL). I am not looking for CSS based fixes since they do not actually prevent rendering and reduce load time. PleaseStand (talk) 19:34, 8 July 2010 (UTC)

I'm a little confused by this. The "do not show page content below diffs" preference doesn't add anything to the URL. Additionally, it's certainly working (for me at least) on pending changes too. How is the &diffonly=0 being added? If you want "diffonly", it should be &diffonly=1. Calvin 1998 (t·c) 22:21, 8 July 2010 (UTC)
How to reproduce: 1) Set My preferences → Misc → Diffs → Do not show page content below diffs. 2) Go to Special:OldReviewedPages. 3) Click on any diff link (the text of each link is the word "review"). It will ignore your "do not show page content below diffs" setting (each diff URL always has &diffonly=0 at the end, which overrides the preference). There does not seem to be a preferences option to suppress the diffonly=0 or replace it with diffonly=1. PleaseStand (talk) 23:05, 8 July 2010 (UTC)
I have opened ticket 24315 for this problem. —TheDJ (talkcontribs) 23:51, 8 July 2010 (UTC)

CSS opacity

Hi there, is CSS opacity still working? See Portal:Cornwall/Random banner/1, it's not working anymore, and there haven't been any changes to that page. If it is still working does anyone know how to fix it on that page? Many thanks, --Joowwww (talk) 22:10, 8 July 2010 (UTC)

The rendered HTML contains style="/* insecure input */". It seems the CSS property filter is considered insecure by MediaWiki and so the whole style is ignored. After I removed the filter, it started working as it should in Firefox, but I think opacity doesn't work in IE. Svick (talk) 23:26, 8 July 2010 (UTC)
Correct, filter is a security problem. See also the bugreport and the commit. —TheDJ (talkcontribs) 23:43, 8 July 2010 (UTC)
I suppose if someone really wanted they could ask for filter:alpha(opacity=...) to be allowed. Changing the regex from filter\s*: to something like filter\s*:(?!\s*alpha\s*\(\s*opacity\s*=\s*\d+\s*\)\s*(?:;|$)) might work. Anomie 01:18, 9 July 2010 (UTC)

changing a page title

I believe "page view statistics" is incorrect. Shouldn't it be "page-view statistics"? Kdammers (talk) 01:40, 7 July 2010 (UTC)

No. See Page view. Gary King (talk · scripts) 06:11, 8 July 2010 (UTC)
There is nothing in the Page view explanation saying why it should be as it is rather than as I propose. General English grammar brings the two words two together with a hyphen if they are related to each other rather than the second one being related to the last word; if the second word is closer in meaning to the last word (i.e., if the first word modifies the second two words), then no hyphon is used.; "Use a hyphen to join two or more words serving as a single adjective before a noun: a one-way street chocolate-covered peanuts well-known author"

Kdammers (talk) 10:39, 9 July 2010 (UTC)

Slow Loading from Internet to Wikipedia x 2 weeks

Did some sort of over-all structural change happen about 2 weeks ago that caused the Wikipedia to now take eons to open? (Maybe that is when I started to notice it...) When I Google a subject and then from the list provided, click on the Wikipedia link, it takes forever for the page to open - even when I am on Cable or high-speed wifi or on other computers. Sometimes I just open the link in a new window and set it off to the side to eventually load at its pleasure. Wikipedia used to be one of the fastest sites to load, and now it is markedly slower than any other site I access lately. I am just wondering why. Thanks Saudade7 18:46, 8 July 2010 (UTC)

  • I should add that once I am *in* the Wiki the links work fine. It's just getting in that's really difficult. And yes, I have cleared my cache, cookies, and history and used various browsers, etc. Saudade7 18:51, 8 July 2010 (UTC)
Having an empty cache would have the effect of making the first load slow, but all subsequent pages fast. OrangeDog (τ • ε) 11:30, 9 July 2010 (UTC)

Recent Changes in Real Time

Firefox 4 beta is out, and it has Websockets support.
For those who have installed it, I wrote a tool that updates the recent changes dynamically.
Here is a demo for en.wikipedia :
And here is a script that turns the RC page of any wiki into a self-updating page:
ThomasV (talk) 07:41, 9 July 2010 (UTC)
It's cool, but it's almost impossible to click on any links, because it scrolls so fast (also Chrome doesn't like the script, it errors out with an origin mismatch) -- Nx / talk 09:33, 9 July 2010 (UTC)
For some reason Chrome requires the page to be on the same server as the websocket. I hope Firefox will not adopt the same policy in the future.
About scrolling : yes, I'm going to add something like that...
ThomasV (talk) 09:49, 9 July 2010 (UTC)
Chrome 5 implements an older version of the specification. It works on Chrome 6. -- Nx / talk 10:28, 9 July 2010 (UTC)

Deep linking to FAMSF website

I need to create links to several of the Thomas Eakins works in the collection of the De Young Museum. To see them, here, search for "Eakins." You will be presented with several paintings, and clicking on them brings you to the description pages I'd like to link to. However, the website does not use canonical URLs. Is it still possible to link to those pages? Raul654 (talk) 05:29, 10 July 2010 (UTC)

It's technically possible. You need to find the URL of the page containing the text for the picture you are interested in. Most browsers you can probably do this by right-clicking on the text and choosing something like "Properties" from the list of options your browser offers. Copy the entire URL and paste it into the wikipage but then delete everything from the first ampersand onwards. For example, if the URL is By Keyword&exhibit=&country=&century=&gallery=&building=&barCode=Search By Bar Code&accessionNumber=Search By Accession#, change this to just The record number is all it needs to display the relevant picture and text, though the resulting page does not include all the menus so you probably ought to include a separate link to the relevant section URL. Not very neat but it works.
As to whether this is permitted by the terms of the website, it seems to be such a bad website that, in my browser, clicking on "terms and conditions" at the bottom simply returns you to the main page. But others here may have some more informed advice about whether deep-linking is in principle acceptable.
Richardguk (talk) 06:42, 10 July 2010 (UTC)
Great! Thanks a bunch - that was exactly what I needed. Raul654 (talk) 15:10, 10 July 2010 (UTC)

Cite error messages not shown in IE8 – CSS "error" class

I've just learned that cite errors, such as the ones discussed here, are not shown to Internet Explorer 8 users. I see that the text is contained in the CSS class, "Error", and I presume that is the cause but does anyone know if this is an intended effect or a bug? AJCham 13:05, 10 July 2010 (UTC)

I am looking into this. ---— Gadget850 (Ed) talk 13:18, 10 July 2010 (UTC)
Is there a version of Wikipedia:Alternative text for images that shows the error messages? The error messages should never have shown in the Wikipedia namespace— they are enabled only for user, main (article), template, category, help and file namespaces by use of {{broken ref}}. We should probably enable error messages for Wikipedia namespace. ---— Gadget850 (Ed) talk 13:54, 10 July 2010 (UTC)
Sure, this is the version referred to by the user in the above talk page link. AJCham 14:00, 10 July 2010 (UTC)
I do not see any cite error messages with FF 3.6.4. ---— Gadget850 (Ed) talk 14:18, 10 July 2010 (UTC)
I was using FF 3.6.3 and have just now upgraded to 3.6.6. The errors are still visible to me, in both Monobook and Vector. My Firefox version is the Ubuntu build distributed by Canonical – don't know if that would make a difference. AJCham 14:38, 10 July 2010 (UTC)
Try purging the page here or bypass your cache with control shift R. ---— Gadget850 (Ed) talk 14:49, 10 July 2010 (UTC)
Do the people seeing the error messages on that history link have their language in their account preferences set to something besides "en"? That could even explain the original problem, if Trafford09 was logged in in IE8 and not in Firefox. For example, compare "en" versus "en-GB". Anomie 14:52, 10 July 2010 (UTC)
And I even documented that. If they don't see the link at the end to the help page, then that is the likely explanation, as it is custom. Would not explain the IE8 issue though. ---— Gadget850 (Ed) talk 14:56, 10 July 2010 (UTC)
Wonder if we could make the en-GB MediWiki pages transclude the en versions? ---— Gadget850 (Ed) talk 15:10, 10 July 2010 (UTC)
Good spot Anomie. I do indeed use en-GB. AJCham 15:12, 10 July 2010 (UTC)
I have never figured out what use it is to set the language to en-GB. The only thing that the language setting does is to set the MediaWiki messages. Here is the message you should have seen: [5] The yellow id the default, which shows with en-GB and the green is what you see with en. It isn't just the cite errors, it is a multitude of messages. ---— Gadget850 (Ed) talk 15:21, 10 July 2010 (UTC)
There are 23 messages at the moment where the MediaWiki defaults differ between en and en-gb.[6] Things like "color" vs "colour" or "-ize" vs "-ise". Anomie 16:16, 10 July 2010 (UTC)

Wikipedia robots.txt settings to crawl the live stuff?

I would like to know what are the right robots.txt settings to put in my crawler to be able to download wikipedia from online following wikipedia policy. I will apreciate any information that could help me solve this issue. Here are some screen capture i took when i put the crawler to download wikipedia.

Thanks —Preceding unsigned comment added by (talk) 21:43, 10 July 2010 (UTC)

I'm not sure what dumpd is, but this error may be because it doesn't set the User-Agent header. But, unless it's only very few pages, you shouldn't download pages from Wikipedia automatically. You should use the API or database dumps instead. Svick (talk) 22:17, 10 July 2010 (UTC)
Please do not crawl wikipedia in this mannor. Instead consider downloading a database dump (google "wikipedia database dump") or use index.php if you really need the live stuff. Example: To get the live version of the page: Apple point the crawler to Tim1357 talk 00:03, 11 July 2010 (UTC)
That's assuming he wants the wikitext. Since we're currently not updating static HTML dumps, if he wants the HTML, crawling the site is the best way to get it. Unless the crawler logs in, it will mostly hit the squid cache. Mr.Z-man 00:23, 11 July 2010 (UTC)

Stats only 5 days of July. Last stats July 6. Anybody know someone who can get this up? (talk) 01:41, 11 July 2010 (UTC)

Mathbot archiving glitch

Can anyone more clueful that me work out why Wikipedia:Articles_for_deletion/Administrator_abuse_on_Wikipedia isn't being archived by MathBot even though it's closed? I've seen malformed closes on AfDs resulting in MathBot being confused, but I can't see the problem on this one. It's causing the list at Wikipedia:Articles_for_deletion/Old to be stretched longer than it should be. It could be fixed by just commenting out the AfD on the 27 June page, but it'd be nice to fix it properly. Black Kite (t) (c) 21:47, 11 July 2010 (UTC)

Vector search dot (. period) suggestions

If you type Ph.D into the Wikipedia search box in the Vector skin, the first two suggested matches are identical, "Ph.D".

If you try the same in Monobook, the second suggestion is "Ph D" with a space instead of a dot.

In Vector, the second suggestion - although it is "Ph.D" - actually links to Ph D.  Chzz  ►  01:35, 12 July 2010 (UTC)

I think it might be because the software ignores the fullstop character when providing the 'autocomplete' search results. -- œ 10:18, 12 July 2010 (UTC)
It's strange that it would return different results based on your skin, though. Gary King (talk · scripts) 17:23, 12 July 2010 (UTC)
The Vector search uses completely different code than the other skins. This is by design and was never tested before roll-out. OrangeDog (τ • ε) 18:11, 12 July 2010 (UTC)
Hurray for untested code. Gary King (talk · scripts) 18:22, 12 July 2010 (UTC)

Images disabled, the semi-prot notice overlaps the GA note

Forgive me if this is the wrong place; feel free to point me somewhere else.

If I disable images in my browser - Firefox 3.0.19, Windows XP, monobook - and look at a semi-protected page which is a 'Good article', the text This article is semi-protected to prevent violations of Wikipedia's biographies of living persons policy. overlaps both the article title and the text This is a good article. Click here for more information. - I have uploaded a screenshot here. I'm using 1024x768.

Another user, on Firefox 3.6.6 on Ubuntu, using both Vector and Monobook, has the same effect; although at 1600 resolution it does not overlap the title part, but it does if they shrink the window.

A third user checked and saw the problem with Firefox 3.6.6 on Windows 7.

Again, apols if this is a known issue.  Chzz  ►  17:16, 10 July 2010 (UTC)

The topicon templates ({{Good article}}, protection icons, etc.) simply need alt parameter in the image wikicode, like |alt=Good article. Not sure if anybody will be willing to do the change though. — AlexSm 15:38, 13 July 2010 (UTC)
The problem is there precisely because an alt text is used. When images are disabled, the alternate text is displayed. There is not enough space there to visually display those beside one another, but if we removed it, the notices would be useless to blind users, in whose interest we use alt text in the first place. I think this can probably be counted as a known issue. If someone has a good idea of a way around the problem, I'd be interested in hearing it. I would oppose making the messages so short that they are not informative, so solutions that adjusted the icon formatting when images are disabled would be preferable. {{Nihiltres|talk|edits|}} 01:32, 24 September 2010 (UTC)


An external website says that it can only present more organized information about a user's edits if the user opts in by creating a page EditCounterOptIn.js under the user page. When attempting to start that page, the heading says

Code that you insert on this page could contain malicious content capable of compromising your account. If you are unsure whether code you are adding to this page is safe, you can ask at the appropriate village pump. The code will be executed when previewing this page under some skins, including Monobook.

and then

Warning: There is no skin "EditCounterOptIn".

Is this cause for concern? Cephal-odd (talk) 05:59, 13 July 2010 (UTC)

No, not with the edit counter. Graham87 07:35, 13 July 2010 (UTC)
1) For the opt-in, you're not placing any executable code on the page - the tool is just checking for the page's existence, and 2) A .js page is used because .js and .css pages are protected from editing by default from anyone except the owner or admins so you don't have to worry about someone else inserting malicious code. Hope that helps explain it. —DoRD (talk) 13:25, 13 July 2010 (UTC)
Thank you both for the explanation. - Cheers, Cephal-odd (talk) 20:39, 13 July 2010 (UTC)

System errors

Something is amiss with the Wikipedia system.

When I clicked on the link to a diff in the edit history of 2010 G-20 Toronto summit, the page which was displayed was the unrelated Residual-current device? A short while ago, I clicked on another link which produced an anomalous result, e.g.,

In other words, the page displayed on the monitor and the http:// at the top of the screen are mismatched?

If this this not the place to alert someone, what is the better venue?--Tenmei (talk) 18:49, 13 July 2010 (UTC)

When I had finished with the above, I clicked on the links; and a new twist had emerged. Clicking on this hyperlink produced Bearded lady? --Tenmei (talk) 18:52, 13 July 2010 (UTC)
In all cases, it looks like you're missing a few digits off the end of the diff link. is a valid diff link for 2010 G-20 Toronto summit, and is a valid link to List of G-20 summits (and goes to Busento via the redirect at Busentinus River). Anomie 20:24, 13 July 2010 (UTC)

Yep, that'd be the problem. Diff links must refer to the correct diff ID number to work, and if you have the right ID the article name listed in the URL is rather irrelevant. --erachima talk 20:57, 13 July 2010 (UTC)

Search engine choice

{{ISBN}} uses Special:BookSources to give a listing of various search engines and libraries for books. {{coord}} runs ~geohack on the toolserver which does the same for geomapping. Do we have an equivalent facility for text search? Something that lists general search engines and even better, specialized ones too, with a preloaded query? Franamax (talk) 23:13, 13 July 2010 (UTC)


Can someone who understands templates, and maybe even nesting of templates, and the best way to do things help us add a field to opt out of italic titles? Thanks. If this request should go somewhere else, please say so. - Peregrine Fisher (talk) 03:57, 14 July 2010 (UTC)

An example can be found in the coding at {{Infobox comics team and title}}. However, given the reason for having the italics template in the the 'box is consistency, if an "opt out" is added, then it would really needs three things:
  • A field name other than "opt out", "supress", or "italics off". Of the cuff, "2ndary_box" would be likely the best use.
  • An associated tracking category - Category:Comics related articles with secondary infobx
  • Clear documentation that the flag is for cases where the 'box is actually a secondary infobox, and not for turning off the comic book series only article because an particular editor dislikes the look. Also that there are composite 'boxes that should be used in place of 2+ 'boxes.
- J Greb (talk) 04:09, 14 July 2010 (UTC)

Wikipedia Topicon Problem

{{Autopatrolled}} and {{Reviewer_topicon}}, these two templates are transcluded in my userpage. When I open my userpage, sometimes the topicons seem like this: File:Mybug 02.png (the left light blue bar is the vertical scroll bar; I am using Mozilla Firefox) and sometimes like this one: File:Mybug 01.png. I have also applied it to other pages (ie my userspace and User:Alexanderps), but i have got the same result. Is there anyone having the same problem? How can I fix this problem? -- Amit6 (talk) 11:08, 11 July 2010 (UTC)

I'm guessing you're referring to this version in monobook skin. The answer is in MediaWiki:Monobook.js (current version), which was created after this discussion: MediaWiki talk:Common.js/Archive 16#topicon_part_deux. It applies the fix to vertical position of these absolutely positioned icons. The 2nd screenshot is when the fix is not active: either you have JS off in your browser or there was no CentralNotice at the moment. Which doesn't make any sense to me, since that notice is not the only thing that can screw with the icons (watch/unwatch actions make them look as ugly). Imho you need to ask sysops at MediaWiki talk:Common.js to improve the code. — AlexSm 15:28, 13 July 2010 (UTC)
The problem is that people keep changing how centralnotice works. So the 'fix' is basically useless in most cases anyway. —TheDJ (talkcontribs) 09:08, 14 July 2010 (UTC)
So, apply the fix on a permanent basis, dropping CentralNotice/wgAction checks and also removing addOnloadHook which makes icons "jump" after the page is loaded. — AlexSm 14:08, 14 July 2010 (UTC)

How I stay signed in

Referring to this question, I discovered that with Internet Explorer 8 I can stay signed in if I delete my history without "Cookies" checked.Vchimpanzee · talk · contributions · 21:45, 12 July 2010 (UTC)

Facepalm Facepalm OrangeDog (τ • ε) 12:48, 13 July 2010 (UTC)
Well, excuse me. I was advised to delete cookies, but obviously I have to do the opposite for this to work.Vchimpanzee · talk · contributions · 14:31, 14 July 2010 (UTC)
Have you considered a superior browser?xenotalk 14:32, 14 July 2010 (UTC)
There's nothing wrong with Internet Explorer 8. Changing browsers causes more problems than it solves.Vchimpanzee · talk · contributions · 15:14, 14 July 2010 (UTC)
I don't recall saying there was anything wrong with IE8 (per se), I just suggested a superior alternative for Wikipedia editing. –xenotalk 15:15, 14 July 2010 (UTC)

Watchlist bot exclusion

The user preferences panel gives an option to exclude bot edits from your watchlist, which is all well and good. However, is it possible to somehow exclude specific users from your watchlist instead? (I would ignore the archival bots and signpost distributor, specifically.) Ignoring all bots is not acceptable to me because I often have images watchlisted which I did not upload, and therefore would not receive the deletion notices for, but which I would not see bot deletion tagging for if I excluded all bots from my watchlist. Thanks. --erachima talk 20:57, 13 July 2010 (UTC)

You could periodically check for changes in file namespace only, with bots edit shown, and hide bots edit in your usual watchlist. — AlexSm 21:28, 13 July 2010 (UTC)
Javascript could be written to do this, possibly modifying User:Gary King/hide pages in watchlist.js. –xenotalk 21:32, 13 July 2010 (UTC)
Maybe. That script is old and not really well written so a new one would probably have to be written. Probably not by me, though. Gary King (talk · scripts) 19:05, 14 July 2010 (UTC)

Compare selected revision problems

I don't know if it's just me, but I can't compare revisions from a page history at the moment. I get the error messge You have either not specified a target revision(s) to perform this function, the specified revision does not exist, or you are attempting to hide the current revision.. It may just be coincidence, but I've just had the admin bit restored to my account [7] - maybe this has screwed something up. Any thoughts? (For info, I'm in classic skin running on IE6) —  Tivedshambo  (t/c) 12:51, 14 July 2010 (UTC)

Are you clicking the button, or hitting "enter"? If the latter, it is related to Wikipedia:Village pump (technical)/Archive 76#After Revision delete, default action changed for pressing "enter" on history talk page and you'll need to press the button for now. If the former, maybe a better browser might help? –xenotalk 12:54, 14 July 2010 (UTC)
I'm pressing the button. I'll try it on Firefox tonight, but at present I'm stuck in the office where there's no choice of browser :-( —  Tivedshambo  (t/c) 13:28, 14 July 2010 (UTC)
Ouch. Maybe this script might help as a workaround? –xenotalk 13:30, 14 July 2010 (UTC)
importScript('User:Superm401/Compare_link.js'); //   turn "compare selected revisions" into a copyable link
  • Bugzilla should probably be filed on this iff bugzilla:23747 going live doesn't fix it. –xenotalk 13:38, 14 July 2010 (UTC)
  • On another wiki I made the RevDel interface hidden by default because it's a bit distracting and not needed 99% of the time. Just a simple script in Sysop.js for all sysops. — AlexSm 16:42, 14 July 2010 (UTC)
I don't have any problems under Firefox and IE8, so it's probably a browser issue. —  Tivedshambo  (t/c) 18:17, 14 July 2010 (UTC)

Since auto update to Firefox 7.0.1 (been a while), whenever I try to 'Compare Selected Revisions', my system doesn't know what to do with the return index.php, and asks whether to save it or select the program to open it. I suspected was Firefox, but checking on IE 8.0.6001.18702 returning same error. I'm running on XP - Service Pack 3.

Not a show stopper, but quite frustrating. Any suggestions for settings / installations gratefully received.

Best wishes Haruth (talk) 20:50, 14 October 2011 (UTC)

Vector issues

IE update and gadget issues


My employer's IT recently updated our browsers from IE6 to IE7 (yeah, a 'lil behind the times), but in doing so, most of my gadgets no longer work, especially popups, HotCat, and the RefToolbar. Since I've never had an issue with this on my home computer on IE7 or IE8, I'm guessing it must be that some of the settings may have changed, but I'm not sure which options to begin fiddling with. Both computers use Windows XP. Any idea where to start, like security and javascript? bahamut0013wordsdeeds 17:19, 9 July 2010 (UTC)

I don't know about those three gadgets specifically, but a lot of scripts don't work in Internet Explorer because the browser acts differently from all other browsers. Specifically, it does not count empty spaces as content (other browsers read whitespace as text nodes). Gary King (talk · scripts) 18:16, 9 July 2010 (UTC)
Well, I can rule out it being the browser itself, since it does work on the same browser on my home computer, as I said. But I'm sure that the options and settings are wildly different, being a military system. I'm also pretty sure it's not my network, because they all worked fine on the Navy/Marine Corps Intranet before the update. bahamut0013wordsdeeds 18:28, 9 July 2010 (UTC)
Maybe a different JavaScript engine, or security settings that affect page display. And even if both your computers say they are running "IE7", it's possible for them to have different patches to IE7 or to Windows itself. — Richardguk (talk) 06:51, 10 July 2010 (UTC)
Is there anything I can look at to verify this, or any specific settings you think I could try tweaking? bahamut0013wordsdeeds 17:51, 13 July 2010 (UTC)
Your best bet is to use a JavaScript debugger, which will tell you if the gadgets are returning any errors when they aren't working. I've never used IE7 before; it appears that it doesn't have a built-in JavaScript debugger, so you'll have to use some of the options available. Here's a decent page that should help you on your way. Gary King (talk · scripts) 16:24, 15 July 2010 (UTC)
NMCI pushed another update, and it seems to have resolved the issue (and created others, but that's a different matter). Thanks anyway! bahamut0013wordsdeeds 19:40, 15 July 2010 (UTC)

Content Based Searching

Can we use a content based search engine for document search in Wikipedia. The above link will guide you through a mathematical search engine, which uses a Principle Component Analysis to provide most appropriate results. —Preceding unsigned comment added by Prashant.prashn (talkcontribs) 03:56, 14 July 2010 (UTC)

I suggest you make an attempt to make it scalable, and actually test it on real data (like a Wikipedia dump maybe). Theoretical papers on search are all well and good, but no-one cares if you don't have practical working code. OrangeDog (τε) 15:24, 15 July 2010 (UTC)

What's flashing by at the top of my watchlist?

For the past few days, some message has been appearing for a tenth of a second at the top of my watchlist, maybe twice a day. It's not in the page source. I'm still using monobook, so my wild guess is that they're trying to sell me on vector ... anyone know? Anyone know a url where I can see the message? - Dank (push to talk) 15:43, 15 July 2010 (UTC)

I suggest logging out, deleting all your cookies for all Wikipedia and Wikimedia sites (or just all of your cookies), and logging back in. It's likely a message that you've clicked "hide" on and forgotten about. Incidentally, I'm still using Monobook too and haven't noticed anything. --Deskana (talk) 15:47, 15 July 2010 (UTC)
Thanks. In the past, the messages I've hidden in my watchlist have been visible in the page source; this one isn't. - Dank (push to talk) 16:07, 15 July 2010 (UTC)
I notice this all the time. What I do to hide it is, when I'm loading a page, I stop the page from loading before the message disappears, then I click the "Hide" link to hide the box. It usually hides the box for a few weeks before it reappears again. Yes, it is annoying. Gary King (talk · scripts) 19:36, 15 July 2010 (UTC)
Thanks Gary, I'll try that. - Dank (push to talk) 20:58, 15 July 2010 (UTC)
That would be Central and Local notices, Due to how they are loaded, they are done then hidden as appropriate (for example if they are no longer meant to be displayed or are hidden by the user), What you are seeing are the past "OMG! We are on vector now" messages. Peachey88 (Talk Page · Contribs) 01:31, 16 July 2010 (UTC)

User Subpage

Where exactly do I go to create a user subpage?--Woogie10w (talk) 20:45, 15 July 2010 (UTC)

Somewhere like User:Woogie10w/Sandbox. OrangeDog (τε) 20:57, 15 July 2010 (UTC)
Thanks--Woogie10w (talk) 21:09, 15 July 2010 (UTC)

Changing class information in a template section

If mediawiki:common.css has coding such as this for a class:

   border: 1px solid #aaaaaa;
   background-color: #f9f9f9;
   color: black;

...and I create this template:

| image       = {{{image|}}}
| headerstyle = background-color:#FFFFFF
| data1 = {{{helpmeplease|}}

And with the "data1/helpmeplease" parameter I don't want:

border: 1px solid #aaaaaa;

What type of coding can I use to make sure there is no border?

I tried style formatting and nothing worked.

I hope I wrote this clearly, thank you so much in advance for your help! Adamtheclown (talk) 20:12, 15 July 2010 (UTC)

This is an infobox without border.
There is no Template:Bigbox, so I suppose you want general solution for all box meta-templates (like {{infobox}} or {{navbox}})? I'm afraid there is no single solution, but most of those meta-templates should provide a parameter to override the style. For example, in the case of {{infobox}}, you can use this code (the result is in on the right):
| bodystyle = border: none;
| above = Example
| data1 = This is an infobox without border.
Svick (talk) 11:18, 16 July 2010 (UTC)


I would like to make a diagram listing size ranges for Miniopterus bats from the data at List of bats of Madagascar#Family Miniopteridae. This would look similar to File:Subfossil lemur C14 ranges.svg. Is there a way to do this on-Wiki, using some intricate kind of formatting or the EasyTimeline extension, or do I need to create an image? Ucucha 16:19, 16 July 2010 (UTC)

I'm not in favor of (ab)using <timeline> for other purposes, so I would suggest creating an image. SVG is probably the best format, because it's scalable and can be relatively easily edited. Svick (talk) 16:31, 16 July 2010 (UTC)
Wouldn't it be possible to do this via some sort of table syntax, though? I would prefer that, because it would be easier to edit (I'm almost certain new species of Malagasy Miniopterus will continue to be discovered) and to adapt to other similar purposes. Ucucha 16:37, 16 July 2010 (UTC)

How to publish completed article

HOW ON EARTH do I publish my edited page?!? Or make the page "go live"? I cannot find this topic in ANY help sections, ANYWHERE. I've been looking for hours. It must just be ridiculously simple, obvious, and right in front of my face. I just want to move it from my userspace to become an actual Wikipedia article. This is very frustrating. Please help! —Preceding unsigned comment added by SportsScienes (talkcontribs) 16:53, 16 July 2010 (UTC)

International Sports Sciences Association has already been created. Ruslik_Zero 16:57, 16 July 2010 (UTC)
Also see WP:Your first article. However, since the article already exists, you can edit the one above to include all the additional material in your draft. —DoRD (talk) 17:01, 16 July 2010 (UTC)

Skin stats

Are there publicly available stats on the popularity of the various skins among Wikipedians? If so, where are they? I imagine those figures could be quite interesting in light of the recent Vector implementation. Thanks for the help. —Anonymous DissidentTalk 03:35, 16 July 2010 (UTC)

Not publicly available, I do know what it looked like before vector, March 21st 2009 for en.wp. The figures from back then are here, because I once requested them. I'll try to ask for an update of those numbers. —TheDJ (talkcontribs) 12:20, 16 July 2010 (UTC)
With the help of User:DaB., I now have some new numbers. I first have to explain something. The database is currently in a transition from one format of preferences storage, to another format. Because of this transition, these numbers are on the 269787 users who in the past months have made 'a' change to their preferences. The numbers probably also include all users who in the past had explicitly chosen the monobook skin in their preferences, these users have all been changed to vector. Users who had default skin and didn't change their preferences are probably not included in this data. As such it is very difficult to make any definitive explanations on this data, but it seems that about 34.000 users have chosen to go back to monobook. (16.000 users use modern skin). There are about 600.000 en.wp Wikipedians (people with more than 5 edits) in total. [8] The list with current known skin selections. —TheDJ (talkcontribs) 13:00, 16 July 2010 (UTC)
Is 600000 the right number? It looks like that came from column A in the first table on the page you linked, but that's accounts with at least 10 edits; column C might be more useful, which indicates that we've been trending about 40000 accounts with at least 5 edits per month recently.
It would be interesting to know skin statistics for "active" users, for some value of active (e.g. user_touched in the past few months, if the comment in tables.sql is accurate). For example, vandals and socks that managed to rack up more than 5 edits before being blocked years ago would be counted in those numbers, as will formerly-active editors who have left long before Vector. Anomie 13:43, 16 July 2010 (UTC)
I copied the stats to here: Wikipedia:Skin#Raw_skin_usage_data. Figure they'd be useful. Gary King (talk · scripts) 03:51, 17 July 2010 (UTC)
Marvelous. Please continue to update and expand the information at that page. -- Quiddity (talk) 17:22, 17 July 2010 (UTC)
Also, there is a Toolserver issue request is at in case that helps anyone coordinate on this. -- Quiddity (talk) 17:22, 17 July 2010 (UTC)


I've written a tool in 2.0 that allows automatic/assisted {{cite news}} citations with the following websites:

  • (this one is broken)

The tool and its source are available at sourceforge. If you find it useful/worthwhile, please leave a note and I'll consider further development...Enjoy.Smallman12q (talk) 23:39, 16 July 2010 (UTC)

You should add it to Wikipedia:Citation tools
Looks intriguing, although Windows-only. If I wasn't confined to linux, I'd give it a try.
I've been using the bookmarklet at User:Bazzargh/citemark for a few months. It's very simple, but does what it says on the box. -- Quiddity (talk) 17:33, 17 July 2010 (UTC)
I've developed something similar for AWB: User:Rjwilmsi/CiteCompletion. Rjwilmsi 18:38, 17 July 2010 (UTC)
It's written in .net so I will make a mono port for the 1.0 release.=DSmallman12q (talk) 21:02, 17 July 2010 (UTC)

Citation needed and paragraphs

The {{citation needed}} template is causing the paragraph following the one to which it is attached to run together. The bug is only evident when the second paragraph begins with a wikilink. For instance,

Poodles make great bacon sandwiches.[citation needed]

Animal cruelty should not be allowed on Wikipedia.

works fine. But,

Poodles make great bacon sandwiches.[citation needed]

Animal cruelty should not be allowed on Wikipedia.

results in the two lines being run together. This is not happening here on this page: it appears to only be a problem in the main namespace, which is very odd, but it is easily verified by pasting the above text in to an article and viewing it in preview. I have tried this in IE6, Firefox 3.5 and 3.6 - all with the same result.

I have no idea how to fix this. SpinningSpark 14:03, 17 July 2010 (UTC)

{{Citation needed}} uses {{fix}} which uses namespace detection, so whatever is causing the problem only shows in mainspace. I reported this two years ago; see Template talk:Fix/Archive 1#Fix at end of paragraph. I think a space after {{citation needed}} is a workaround. ---— Gadget850 (Ed) talk 14:32, 17 July 2010 (UTC)
I've no idea why, but the general problem seems to be that whitespace is ignored between a category declaration and a wikilink (and the "fix" template ends with category declarations). A workaround would probably be to add a no-break space after the category declarations in the fix template.--Kotniski (talk) 14:49, 17 July 2010 (UTC)
PS - {{fix}} suppresses the categories outside mainspace, which is why the problem only shows up in mainspace.--Kotniski (talk) 14:54, 17 July 2010 (UTC)
I've made an editprotected request at Template talk:Fix/Archive 1#Fix category breakage; someone recently moved the categories output by {{fix}} from before to after the text output. Moving them back before causes the categories to no longer be at the end of the paragraph, so it all works fine. Anomie 17:38, 17 July 2010 (UTC)
The fix has been applied by Ucucha in this edit. If you still see the problem, try purging the page. Anomie 17:52, 17 July 2010 (UTC)
Thanks, that got it. I thought it best not to apply your fix myself as I did not have the faintest idea what I was doing. SpinningSpark 19:12, 17 July 2010 (UTC)

Preference to mark all edits minor by default asked to be removed in bugzilla:24313

Per the above-linked bugzilla, and this discussion, this preference will likely be removed shortly (as users often set it and forget it, causing non-minor edits to be marked minor).

The preference will be reset to the default; those who used previously this preference could probably use a script to restore the functionality, if desired. –xenotalk 21:30, 8 July 2010 (UTC)

This change will be occurring very shortly; in order to assist users who had been appropriately using this, is anyone available to write a simple script to restore the functionality? –xenotalk 22:38, 14 July 2010 (UTC)
So was there a point to removing it if you are also after a script to add it back? -- WOSlinker (talk) 22:59, 14 July 2010 (UTC)
if(wgAction == 'edit') {
    addOnloadHook(function minorEdit() {
       document.getElementById('wpMinoredit').checked = true;
Several points. Firstly, the existence of a preference in the interface caused some problems by implying that using the preference was always allowed and could not be restricted. Secondly, editors who misuse a script can be restricted by having an administrator remove the script and protect their skin.js. Because there's no way to stop an editor from changing their preferences, the only way to stop abuse of the preference is by blocking. Thirdly, installing a script is a bit more difficult than checking off a preference, and requires the editor to view a message reminding them that they are responsible for the use of the script, so it there should be fewer editors doing so without thinking of the consequences. Gavia immer (talk) 23:16, 14 July 2010 (UTC)
Thank you for writing the script, WOSlinker. Gavia sums it up well above. I'm sure there were a few folks out there who were happily gnoming along without incident that may be put off by this bugzilla, so that's why I've requested the script. Much obliged! –xenotalk 00:26, 15 July 2010 (UTC)
I haven't tried the above script yet, but, as is, won't it forget your current unchecked value on a refresh due to preview or showing changes? Mark Hurd (talk) 05:58, 15 July 2010 (UTC)
I don't think so. The minor-edit checkbox retains its condition while showing preview and changes. Someguy1221 (talk) 06:08, 15 July 2010 (UTC)
I think Mark was asking whether the script would re-check the checkbox after pressing "Preview". I haven't tested it, but I don't think it will as it seems wgAction is "submit" in those cases, rather than "edit". Anomie 11:46, 15 July 2010 (UTC)
If it's checked when you click the "Preview" button, it will be checked on the preview page, because the form contents are always reproduced when you preview. If you had the script auto-check on the preview page too, then if you unchecked it and previewed, it would become re-checked, which is not the desired behavior. —Aryeh Gregor (talk • contribs) 20:08, 16 July 2010 (UTC)
Anomie has it right and I have now confirmed Preview and Changes both work as expected. Mark Hurd (talk) 04:38, 19 July 2010 (UTC)

Here's something better. It remembers your last minor edit setting in a cookie. -- WOSlinker (talk) 08:20, 15 July 2010 (UTC)

if(wgAction == 'edit') {
    addOnloadHook(function minorEdit() {
       addHandler(document.getElementById('wpMinoredit'), 'change', minorEdit_change);
       if (document.cookie.indexOf("minorEdit=true")!=-1) {document.getElementById('wpMinoredit').checked = true };
    } )

function minorEdit_change() {
    var e = new Date();
    e.setTime( e.getTime() + (24*60*60*1000) ); // one day
    document.cookie = 'minorEdit='+document.getElementById('wpMinoredit').checked+';expires=' + e.toGMTString();

The above script sets future edits to the same setting as your most recent edit. It does this for one day before reverting to marking all edits to major edits. I don't see the purpose of this script and it seems extremely error prone to me, as there is no true default in this case and the "default" that is set expires without warning. I recommend against its use. Bigmantonyd (talk) 23:35, 21 March 2011 (UTC)

Could you please instruct the non-technical folks (including me) exactly what steps to take to use the script? atakdoug (talk) 06:20, 16 March 2011 (UTC)
As per Atakdoug - where in wikipedia does one apply this script - I am a gnome so most of what I do is minor. I have some understanding of Java but where to write it to? MarkDask 12:27, 17 March 2011 (UTC)
(edit conflict)Here's my experience: First, I tried this in my Common.js. When it didn't work, I then put it in my Vector.js, but it still didn't work. I gave up and let it be for a few minutes. Then I thought I'd try the first example. The weird part is that just as I began to edit Common.js (I'd already removed the code from Vector.js), I noticed that it had started to work! The minor box was checked! So I left things as they were and continued general editing. At this point, the second example above was in my Common.js, my Vector.js was empty, and all was just fine. Then a few minutes later, I noticed that it had stopped working! So, it appeared as if the code had to be in the Vector.js to work. And that is correct. I put it back into Vector.js, and it started working a few minutes later. So the easy steps are:
  • Cut and paste the second example above to your Vector.js file...
    • go to "My preferences"
    • click the "Appearance" tab
    • click and open the Vector skin's "Custom JavaScript" link
    • click "Edit"
    • paste the second example above into the Edit screen
    • save the edit
  • And you're finished! Remember that...
  1. If you use one of the other skins to edit, you'll paste the code into the "Custom JavaScript" for that skin,
  2. I could not get this to work in the "All skins" Custom JavaScript,
  3. AND MOST IMPORTANTLY wait a few minutes while you continue to edit, because the change might not be immediate.  — Paine Ellsworth ( CLIMAX )  12:47, 17 March 2011 (UTC)
Excellent - thank you Paine MarkDask 04:51, 18 March 2011 (UTC)
Well, don't thank me yet, because it looks as though this doesn't work all the time. Sometimes I open an edit page and the Minor box is checked, and other times it's not checked. I don't know if the devs are still working on this or what. – Paine Ellsworth ( CLIMAX )  07:08, 19 March 2011 (UTC)
Okay, I just converted back to the first choice, the:
if(wgAction == 'edit') {
    addOnloadHook(function minorEdit() {
       document.getElementById('wpMinoredit').checked = true;
That works (so far). – Paine Ellsworth ( CLIMAX )  07:45, 19 March 2011 (UTC)

I don't like this. The choice was made, by me, to set my edits as default. This choice was made by me. This choice is something I stand by. This choice should be respected. I do not like the notion that my edits will have to be manually set to minor while this is my default behavior. I am appalled by this move. Maki (talk) 20:33, 19 March 2011 (UTC)

I also think this is absurd. Maybe someone could convert one of these experimental scripts into a simple checkbox on my Editor Preferences that would let me toggle the minordefault value. Perhaps just unhide the one that is still there. A different repair would be to just set the value in my user record back to true and forget about either the script or checkbox. Anyway -- I've used the 1st (3rd) script in Vector skin's "Custom JavaScript" . -- LantzR (talk) 08:48, 10 June 2011 (UTC)

Switched script to [|Mark_minor_edit] (same code with comments) LantzR (talk) 10:57, 10 June 2011 (UTC)


I am currently assisting a non-technical user who is stuck with having every edit marked as minor with no way to turn it off. That hidden box needs to be turned off for all users by a 'bot. Giving the users a way to turn it back on would seem to be a Good Thing for users such as Makitk above who appear to know what they are doing, but the hidden box needs to be unchecked for users who don't do something special to keep it turned on. I am sure that there are many other users who, like mine, have been spending years inadvertently marking all edits - including major changes - as minor without really knowing what a minor edit is or why their is an m next the every edit. There is a larger issue as well: when a user interface option is hidden, careful thought needs to be given to the users who now have an unchangeable user preference. Guy Macon (talk) 15:04, 10 June 2011 (UTC)

GeSHi clarification before I go to bugzilla

I tried to use the syntax highlighting code "<syntaxhighight lang="rsplus"></syntaxhighight> (for the R programming language, but it appears that rsplus is not supported in wikipedia. Can anyone confirm that "rsplus" doesn't work or let me know if there is a bug listed for GeSHi to update the languages on wikipedia? Thanks. Protonk (talk) 01:50, 15 July 2010 (UTC)

I get the same behavior you do - the docs say rsplus is supported, but trying to use it gives an error message that doesn't list it as one of the supported options. I can't find a bug report for this, so I'd go ahead and make a new one if I were you. Gavia immer (talk) 02:23, 15 July 2010 (UTC)

  • Listed at buzilla #24383. Here's hoping I didn't miss something obvious. Protonk (talk) 03:11, 15 July 2010 (UTC)
    • I've tagged it as "1.16wmf4", which means it should get higher priority for being pushed to the main site. (X! · talk)  · @671  ·  15:06, 18 July 2010 (UTC)

Template:Location map

I create new Template:Location map2 with ideas from de, ru, fr and uk wiki. This template works also with location map templates with y, x parameters, for example: Template:Location map China1, Template:Location map Canada1, Template:Location map Russia1, Template:Location map Africa1. May be it need to include code of Template:Location map2 to Template:Location map?--Амба (talk) 21:17, 16 July 2010 (UTC)

This is a fork of {{Location map}} and is being discussed at Templates for discussion.  –droll [chat] 06:53, 18 July 2010 (UTC)

Two search boxes - any way to make them gracefully switch from side by side to one below the other?

Hey all, recently we redesigned the header for Wikipedia:Reference desk pages and added another search box for the whole of wikipedia in case people miss the one at the top. These are put side by side, see for example Wikipedia:Reference desk/Computing. Problem with this is it increases the minimum horizontal resolution for the header to above 800 pixels.

While I left a suggestion on the talk page, no one has yet came up with what would seem to me to be the obvious solution, make them gracefully switch from being side to side to one below the other as necessary. Does anyone know if this is possible and if so, how to do it?

Nil Einne (talk) 06:33, 18 July 2010 (UTC)

Put them into two divs and float them left: User:Nx/Sandbox -- Nx / talk 06:42, 18 July 2010 (UTC)
The search boxes are contained within a table in Wikipedia:Reference desk/header/howtoask. It looks straightforward to reformat the table if there's consensus to do so. —  Tivedshambo  (t/c) 06:48, 18 July 2010 (UTC)
Btw, the inputboxes are wider than 300px (at least on my computer), so the table cells were wider than 300px despite the header style suggesting that they are 300px wide. -- Nx / talk 06:52, 18 July 2010 (UTC)

Image links

Normally, if I click on a Wikipedia image, it takes me to the description page for that image. Recently, I haven't been able to click on any image at all. I'm still using the MonoBook skin, in IE8. --S-man (talk) 19:06, 18 July 2010 (UTC)

Nevermind. I reset the browser; it's all good now. --S-man (talk) 20:26, 18 July 2010 (UTC)

Cannot use parameters to create a reference inside a template


I have been experimenting with the use of a sub-template to create references to an external site which uses structured URLs, created by an existing template {{London Gazette}}.

My intention was that the parameters in my new template which creates the row in the data table would include three for the reference: {{{issue}}}, {{{date}}} and {{{page}}}. Those three parameters would be used by the row-building template to call a sub-template which would create a named reference for one of the columns, as follows: <ref name="{{{issue}}-{{{date}}}-{{{page}}}">{{London Gazette |issue = {{{issue}}} | startpage = {{{page}}} | date={{{date}}} }}</ref>

That would give me properly formatted references, with names which allow cite.php to merge refs to the same page

In my tests I have incorporated error-checking to ensure that all parameters are present, and have successfully passed all the parameters. {{London Gazette}} is called ... but all the parameters in <ref name="{{{issue}}-{{{date}}}-{{{page}}}">{{London Gazette |issue = {{{issue}}} | startpage = {{{page}}} | date={{{date}}} }}</ref> are empty.

To demonstrate this, I have set up a simple test at User:BrownHairedGirl/sandbox, in which the template {{User:BrownHairedGirl/myref]]}} is passed a one-word parameter. As you can see the parameter has a null value both in creating the reference name and in creating the text of the ref.

Is this a known bug, and is there any workaround?

It may sound like an esoteric issue, but in this case it would make it much easier to create and maintain a long series of references, particularly since each reference will be shared by an average of about 30 entries. --BrownHairedGirl (talk) • (contribs) 14:41, 18 July 2010 (UTC)

Using {{#tag:ref}} in place of <ref name="etc.">content</ref> seems to fix it. Intelligentsock 15:13, 18 July 2010 (UTC)
See WP:REFNEST for full syntax and bugs. ---— Gadget850 (Ed) talk 15:26, 18 July 2010 (UTC)
Big thanks to both of you. I didn't know about {{#tag:ref}}, but that fixes it all. --BrownHairedGirl (talk) • (contribs) 11:37, 19 July 2010 (UTC)

Prompt me when entering a blank edit summary


If the "Prompt me when entering a blank edit summary" feature is enabled in preferences and an edit summary is not entered, it prevents non-autoconfirmed users from saving edits which trigger the captcha. The edit window gets locked into a cycle of entering the captcha, being informed you didn't enter an edit summary and clicking continue to save without one, and being asked for a captcha again. Obviously this is easy to overcome by disabling the feature in preferences or entering an edit summary, but new users may be unaware of what they are doing wrong, and may assume the captcha is the problem 1230049-0012394-C (talk) 16:29, 19 July 2010 (UTC)

See bugzilla:10729 -- Nx / talk 16:33, 19 July 2010 (UTC)
Good to know it's being worked on, thanks! 1230049-0012394-C (talk) 16:35, 19 July 2010 (UTC)

Strange glitch

New pages 1.png

This appeared while viewing User talk:White Shadows. I'm running The latest Firefox on Mac OS X 10.6.3. ~NerdyScienceDude () 18:15, 19 July 2010 (UTC)

I'm running exactly the same browser on the same operating system and I'm getting what it's supposed to look like. I'm using Vector. Mr. R00t Talk 18:36, 19 July 2010 (UTC)
Yes, it's a recognised bug when transcluding special pages - bugzilla:23293. Nothing to worry about really, just, as you say, a glitch. - Jarry1250 [Humorous? Discuss.] 18:40, 19 July 2010 (UTC)

Moving to a lowercase article title

Hello. However it is that we are able to display articles such as iPhone with a leading lowercase letter, could someone please do that to Inext? - Richard Cavell (talk) 14:41, 16 July 2010 (UTC)

 Done - thanks to EmilJ. - Richard Cavell (talk) 14:45, 16 July 2010 (UTC)
I have write IPhone. I do it in that way: I clicked in editor the "Insert Link" button, typed on the keyboard "iPo" (in that cases!!!) and clicked in combobox "IPhone" item. So i can't understand: Where is the iPhone article you talking about??? And why the script change my "iPo" to "IPo"+"ne"??? Please fix that bug!
--W.M.drossel (talk) 11:17, 20 July 2010 (UTC)
The script is an intended feature to allow you to find pages quickly. You can always not use the "Insert Link" button and just type [[iPhone]] instead.
For technical reasons, all articles start with an upper-case letter, but you can add {{Lower case title}} to make it look like it starts with a lower-case letter. OrangeDog (τ • ε) 12:04, 20 July 2010 (UTC)

"Changes since I last edited" scipt

Is this script still working for other people? All I get is the "You have not edited this page! (recently)" message (even when I have). I'm still using the monobook skin.--Dodo bird (talk) 23:32, 18 July 2010 (UTC)

It has not been working for me recently in Vector either. —Ost (talk) 12:54, 19 July 2010 (UTC)
Here are the fixes necessary to make it work again. I'll see if I can make it slightly more robust against future changes. -- Nx / talk 19:44, 19 July 2010 (UTC)
Wikipedia talk:WikiProject User scripts/Scripts/Changes since I last edited#Fixed version -- Nx / talk 20:33, 19 July 2010 (UTC)
Thanks for the fix. It's working now. --Dodo bird (talk) 18:33, 20 July 2010 (UTC)

Skin checking

Is it possible to write a script (or is there one already?) that will check which skin and browser you're using and show them a different version of the page based on that? I'm currently talking about the problems at HJ Mitchell's userpage with the English flag at top. It won't look right in Vector when it look right in MonoBook and looks bad in MonoBook when it looks good in Vector. I know that there are scripts that you can add to your monobook.js or vector.js. Mr. R00t Talk 16:49, 19 July 2010 (UTC)

You mean something like like the GENDER magicword? No, not that I know of. –xenotalk 16:55, 19 July 2010 (UTC)
How about we just outlaw elements that cover parts of the userinterface ? The topicons are bad enough in my opinion. —TheDJ (talkcontribs) 17:16, 19 July 2010 (UTC)
Not a bad idea. I've moved the "toolbox" to the top location and some users have elements that then obscure my ability to click "user contributions", "logs", etc. (though this is not the case here). –xenotalk 17:18, 19 July 2010 (UTC)
Something like this was what I was thinking of. Would something like that even work? Mr. R00t Talk 18:05, 19 July 2010 (UTC)
For javascript, the skin global variable contains the skin name (right click, view source, and you'll see a bunch of globals in the js code block at the very top, including the skin), for css, the body element has a skin-specific class, e.g. "skin-vector". Of course this is not very useful, because you can only add the script or css for yourself, not anyone else visiting your userpage, so it will be broken for everyone else. -- Nx / talk 18:13, 19 July 2010 (UTC)

──────────────────────────────────────────────────────────────────────────────────────────────────── I know that. I need to be able to be able to access whatever program it uses and put it on a page. I'm hoping that it will work similar to {{REVISIONUSER}}. Mr. R00t Talk 18:19, 19 July 2010 (UTC)

File a bugzilla:. –xenotalk 18:23, 19 July 2010 (UTC)
Alright. Mr. R00t Talk 18:24, 19 July 2010 (UTC)
There's a mediawiki extension that introduces a magic word to output the skin name, but the problem with that is that it breaks the parser cache - whenever someone with a different skin visits the page, the page would have to be reparsed (alternatively, multiple copies of the page would be stored in the cache, one for each skin). It's not installed and probably will never be installed at WP for this reason. -- Nx / talk 18:24, 19 July 2010 (UTC)
bug 24444. I was thinking of placing different versions in the cache to make it work. Mr. R00t Talk 18:34, 19 July 2010 (UTC)
That would be a lot of cache duplication for a pretty much useless (IMO) feature . It would be more efficient to just do what was done for topicon, and add some css to the global skin-specific css files to position the image correctly. But I don't see how that would be useful except for userpages. -- Nx / talk 18:52, 19 July 2010 (UTC)
There is no dev that will do this I suspect. It is a stupid little feature for myspace-like userpages with a big performance hit. —TheDJ (talkcontribs) 12:17, 20 July 2010 (UTC)

Desription of wikifunctions.dll

Is there any desription to AWB library - wikifunctions.dll? Besuglov.S cont / talk 16:34, 20 July 2010 (UTC)

Exporting from OpenOffice

Are you familiar with the function of exporting from OpenOffice to MediaWiki? See Wikipedia talk:Tools#Exporting from OpenOffice.

Thank you. --Amir E. Aharoni (talk) 21:27, 20 July 2010 (UTC)

Note: I've responded at the linked talk page. AJCham 21:39, 20 July 2010 (UTC)

Telling articles from redirects

I've been around a few years and have a bloated watchlist (about 12,361 entries). I'd guess that many hundreds of them are redirects that were created automatically when pages were moved. When I sit down to delete some entries from the list I get discouraged at the thought of checking all of those to find the redirects. Does anyone know of a way to determine which articles are redirects automatically? Perhaps starting with pasting the watchlist into a page? Or using some tool? (This was inspired by trying out Huggle and finding it timing out when checking my watchlist, though I found a work-around for that).   Will Beback  talk  07:43, 18 July 2010 (UTC)

Add the line .watchlistredir { font-style:italic; } to your personal css file. This will display redirects in italics. —  Tivedshambo  (t/c) 07:58, 18 July 2010 (UTC)
Brilliant. I deleted 87 redirects from the first two letters of the alphabet alone, not counting some that I kept. BTW, do you know if there is a listing of those Wikipedia-specific CSS elements? There are some other things I'd like to change.   Will Beback  talk  08:13, 18 July 2010 (UTC)
Yes, Wikipedia:Catalogue of CSS classes. Graham87 08:40, 18 July 2010 (UTC)
Thanks!   Will Beback  talk  20:15, 18 July 2010 (UTC)
Wow, this made me actually go and finally use the CSS thing for the first time, despite seeing a lot of various other useful hints and stuff that I could have over the years. Thanks. ♫ Melodia Chaconne ♫ (talk) 14:05, 18 July 2010 (UTC)
I personally also like to add (in addition to the above):
.allpagesredirect:after { color: #808080; content: " (redirect)"}
It adds a little something more. Color can, of course, be tweaked to suit your preference. --Izno (talk) 13:22, 19 July 2010 (UTC)

Or, use Anomie's link classifier script:

importScript('User:Anomie/linkclassifier.js'); // Linkback: [[User:Anomie/linkclassifier.js]]
importStylesheet('User:Anomie/linkclassifier.css'); // Linkback: [[User:Anomie/linkclassifier.css]]

The follow the instructions at the top of that page to bypass the cache.

See User:Anomie/linkclassifier.css for a list of the other link colors. ---— Gadget850 (Ed) talk 17:51, 19 July 2010 (UTC)

I am a firm believer in Link classifer. I said it before and I'll say it again "I think this script should be stanard issue to all editors". To your question this script will turn all redirected links Green, and all Disambig's to Yellow and AFD's to pink. doesn't get any eiser than that. Mlpearc powwow 18:03, 19 July 2010 (UTC)
I don't think it should be standard issue. Changing so many elements on a page, can be rather problematic for older browsers and computers. If it was JUST css, it be a different thing, but it is using Javascript, and a script that on a page like Barack Obama can take a rather long while to execute. If you want to use it, that's your prerogative, but there are many more people who do NOT want to use it, and to who it would actually be a problematic script. So add it as a gadget perhaps, but never enabled by default. —TheDJ (talkcontribs) 12:14, 20 July 2010 (UTC)
It would probably confuse the frack out of most casual editors. ---— Gadget850 (Ed) talk 12:20, 20 July 2010 (UTC)
Linkclassifier was recently tweaked to enable run-on-demand (instead of every pageload): User:Anomie/linkclassifier#Usage. I use "p-tb" instead of "p-cactions", to place the link in the "toolbox", instead of in the tabs. None of the slow-down, all of the power :) -- Quiddity (talk) 19:41, 20 July 2010 (UTC)
I certainly don't like the idea of highlighting redirects differently from direct links, by default, in articles (as opposed to watchlists, where it might make some sense). If that were to happen, my prediction would be that we'd see a run on people "fixing" redirects by changing them to pipes. --Trovatore (talk) 21:31, 20 July 2010 (UTC)
Actually, now an idea has occurred to me. It just might make sense to color piped links differently. One of the most objectionable things about piped links is that the reader has no warning that the link may not go to an article that would be expected to have the title indicated. With a different color, that would be different. Of course it would still be bad practice to use pipes in such a way that information that ought to be given to the reader is encoded in the pipe.
So does anyone have a gadget that colors piped links differently? I might use that one. --Trovatore (talk) 21:36, 20 July 2010 (UTC)
I think you are referring to WP:EASTEREGG. ---— Gadget850 (Ed) talk 03:13, 21 July 2010 (UTC)
Also, redirected pages don't generally have talkpages, so where the talkpage is redlinked on your edit watchlist page, you can spot them straight away. Lugnuts (talk) 12:01, 21 July 2010 (UTC)

Contribs page

I'm not sure if this belongs here, and I ask to be excused if it does not. I noticed something odd that is there when I open anyone's contributions page.

(latest | earliest) View (newer 250  | older 250) (20 | 50 | 100 | 250 | 500)

There are two spaces (an extra space) between "newer 250" and the vertical bar. This applies always, no matter whose contributions I'm seeing, no matter how many edits I've chosen to view, and no matter whether I'm logged in or not. Why is it so? --Theurgist (talk) 07:05, 21 July 2010 (UTC)

I think I have found the error, see Wikipedia:MediaWiki messages#MediaWiki:Viewprevnext. Svick (talk) 09:45, 21 July 2010 (UTC)
It should be fixed now. Svick (talk) 16:29, 21 July 2010 (UTC)

Double space after a period

The MOS:FULLSTOP guideline says: "The number of spaces following the terminal punctuation of a sentence in the wiki markup makes no difference on Wikipedia because the MediaWiki software condenses any number of spaces to just one when rendering the page." Is that true regardless of what browser the reader is using? Long previous discussion here. Art LaPella (talk) 05:52, 21 July 2010 (UTC)

Yes, because this is an underlying feature of HTML in general; in most cases any amount of whitespace is treated identically to one unit of whitespace and rendered as one space. Note that doublespacing does make a difference in the formatting of the edit window text, so it's not entirely without effect. Gavia immer (talk) 06:00, 21 July 2010 (UTC)
Yes, that seems to be the standard answer. However, it might be useful to refer to this page where the original discussion started. The technical discussion began near the end of the thread. WP:MoS#Specific statement with respect to double sentence spacing or double spacing after terminal punctuation is needed. Basically, the discussion centers over whether is it possible for someone to view the spacing in a Wikipedia article with a browser that shows HTML as it was typed, as in "wysiwyg". --Airborne84 (talk) 12:33, 21 July 2010 (UTC)
In short: No, it's not possible.
While it is certainly possible for someone to create a browser that treats white space in a non-standard way, all of the major browsers (for English, anyway) treat runs of multiple whitespace characters as a single space unless specifically instructed otherwise by tags (e.g. <pre>) or CSS. And even were someone to use such a non-standard browser or edit their user css to apply the appropriate CSS directives to the entire page, MediaWiki as installed here (or, more likely, tidy) will itself collapse runs of multiple whitespace characters before serving the page to the browser. So even though there "should" be 5 spaces in [ ] according to the wikitext, if you look at the rendered page output there is only one. Anomie 15:51, 21 July 2010 (UTC)
MediaWiki would have to be changed for this to happen, but it's not impossible. —Aryeh Gregor (talk • contribs) 16:02, 21 July 2010 (UTC)
True, but the question was whether it could be done by using an unusual browser. Perhaps it was an unwarranted assumption on my part that they wouldn't care about a possibility outside their power to directly implement. Anomie 16:41, 21 July 2010 (UTC)
Yes, it would be perfectly possible in principle. The obvious way to do it would be to have MediaWiki replace a double space with a space followed by a non-breaking space. There are some HTML-based applications that do this. However, it's unlikely anyone would be interested in doing the coding for that, since on the web, the universally accepted convention is single spaces between sentences. Also, it would make a lot of pages look odd where people wrote several spaces in a row and expected it to show up as a single space in the actual article. —Aryeh Gregor (talk • contribs) 16:02, 21 July 2010 (UTC)
For that matter, the request would be possible if MediaWiki would just stop collapsing multiple spaces. Normal browsers would continue to collapse the whitespace, but the hypothetical non-standard browser mentioned would be able to choose to display them. I personally agree that changing MediaWiki in this way is not worth doing. Anomie 16:41, 21 July 2010 (UTC)
Sentence spacing was a recently featured article on the Main page. T'would be remiss not to mention :) -- Quiddity (talk) 20:18, 22 July 2010 (UTC)

image annotations

Why do images on wikimedia have option to annotate images but on wikipedia they do not? Kallimachus (talk) 01:14, 22 July 2010 (UTC)

Because the ImageAnnotator gadget is only enabled at Wikimedia Commons. Graham87 02:25, 22 July 2010 (UTC)
It seems very useful, especially when looking at complex images. We need to enable that for wikipedia. Kallimachus (talk) 05:26, 22 July 2010 (UTC)
This has been discussed here at length between December 2009 and March 2010:
Perhaps someone who is interested could read the discussions and summarise the current situation here. I think there was concern that vandals could add inappropriate annotations without easily being detected.
Richardguk (talk) 23:11, 22 July 2010 (UTC)

Transwiki Watchlists

There's way too much separation between Wiki projects. I want to have a transwiki watchlist where I can keep an eye on Wikipedia Pages in various languages as well as files I keep an eye on in the commons and entries I've contributed to in the dictionary. It's way too much of a pain to keep an eye on everything, so certain things go unwatched or unresponded to.

Am I the first one to bring this up? Don't many of you have the same complaint/problem? Chrisrus (talk) 05:40, 22 July 2010 (UTC)

There is an external tool called Watchlistr that does exactly this. And there is also an open bug requesting this. Svick (talk) 09:48, 22 July 2010 (UTC)

Edit summary moved

I've been away for a while. Today I return to make an edit and found the edit summary hidden away(on a mobile device) to the right of the page. if I wasn't really aware from before that edit summaries are desirable I would not of even realised entering an edit summary was possible now. Where did the discussion happen on moving it's position? What where people thinking? Is entering an edit summary something no longer desirable? Don't we want newbies especially to enter edit summaries as much as they can? Regards, SunCreator (talk) 10:09, 22 July 2010 (UTC)

Edit summaries are certainly very desirable, not only from newbies. I don't see any such change in Firefox (even with small window width). What kind of device and browser are you using. Could you provide screenshot/photo of the problem? Svick (talk) 10:32, 22 July 2010 (UTC)

Parser ifeq used with Magic Words


For some reason this coding does not work:

{{#ifeq:{{NAMESPACE}}|Template||'''[[New York City]]'''}}

I want to create a condition, that:

If a the namespace IS a template, '''[[New York City]]''' will NOT be shown.

If the namespace is NOT a template, '''[[New York City]]''' WILL be shown.

Any suggestions? Can I use magicwords this way? If not, is there another way to do this? Thank you for your time. Adamtheclown (talk) 14:27, 22 July 2010 (UTC)

Your code works fine, at least when I tested it on Template:Template sandbox (displayed nothing) versus Wikipedia:Sandbox displayed New York City. –xenotalk 14:32, 22 July 2010 (UTC)
Your right xeno, I just double checked myself, and came here to report this. Thank you for taking the time to do this. 14:35, 22 July 2010 (UTC)

Citing an annotation

Not sure if this is the right place...but how do you cite/reference an annotation?Smallman12q (talk) 16:03, 22 July 2010 (UTC)

We need more context to help you. An annotation where? – ukexpat (talk) 16:07, 22 July 2010 (UTC)
An annotation in an article...for example in Edmund_Evans#Process_and_techniques the last line...I'd like to do
<ref group="a">On page viii of ''The Caldecott Aesop: Twenty Fables : A Facsimile of the 1883 Edition'', a production note states that the "majority of color plates were made from the first woodblock renderings of Caldecott's work." The statement is not sufficiently clear to indicate to what extant the original blocks were used, however the cracks in the color plates suggest that the original blocks may in fact have been used.<ref>Richardson, p. 33</ref></ref>
but it gives me an error. I'd like to cite a page for the annotation.Smallman12q (talk) 16:42, 22 July 2010 (UTC)
WP:REFNEST is probably what you're after. - Jarry1250 [Humorous? Discuss.] 16:55, 22 July 2010 (UTC)
Ye nested ref is what I needed. Thanks.

{{#tag:ref|On page viii of ''The Caldecott Aesop: Twenty Fables : A Facsimile of the 1883 Edition'', a production note states that the "majority of color plates were made from the first woodblock renderings of Caldecott's work." The statement is not sufficiently clear to indicate to what extant the original blocks were used, however the cracks in the color plates suggest that the original blocks may in fact have been used.{{#tag:ref|Richardson, p. 33}}|group="a"}} This worked =).Smallman12q (talk) 17:46, 22 July 2010 (UTC)

DBPedia links

Shouldn't each article link to its DBPedia equivalent? Either at the end of the interwiki links, and/ or through a meta header: <link rel="alternate" href=""> for Birmingham, for example? Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 16:30, 22 July 2010 (UTC)

Well it's not a sister project (so far), and link attributes would require some rather big software changes. Having said that, I'd prefer linking to an open database that contains verified sets of data. So instead of collecting all the data in wikipedia in a separate DB, collecting open data in a DB, is probably more useful in the long term and a better match to our goals. Matching from wikipedia to dbpedia is probably already rather simple to do with DBpedia itself, so I don't see a good reason to do that from within Wikipedia 'manually'. —TheDJ (talkcontribs) 20:13, 22 July 2010 (UTC)

Uploading to Commons

User:guillom of the Wikimedia Multimedia Usability team has published some results on his blog of the first testing done of the usability of Wikimedia Commons upload process, and the results of the first prototype for the new upload wizard that is currently under development. See also the following movies:

Please contribute your ideas, and you can test the prototype if you want. I note that some of the results might apply to other processes as well. The "i'll just skip all this, because it is too much text" being the most notable usability issue that will likely apply to many other pages. —TheDJ (talkcontribs) 20:08, 22 July 2010 (UTC)

Thank you. As a follow-up, I'd like to add that I'm currently working on a Questions & Answers page to address the most frequently asked questions. Feedback on the prototype is of course welcome, but you will understand that we may not have the resources to answer every comment individually, especially if many are similar. The Q&A page isn't ready yet, so it'll take some time before I can publish it. So, please consider this a "soft launch": we don't want to make a lot of publicity about our prototype yet. If you happen to know about it and you want to share your opinion, that's fine. But we'll officially invite the community later to try it out, when the Q&A page lets us focus on the most useful comments. Thanks, Guillaume Paumier 20:32, 22 July 2010 (UTC)

Strange "redirect"

I was looking at a very long article, but if I went back to it the computer would have to scroll down to get back to what I was reading. However, it didn't do that. It jumped forward to an article whose link I clicked on earlier while looking at the very long article, and the back button was no longer blue, meaning it couldn't be used.

I've asked similar questions before but no one seems to have an answer for this. Sometimes with very long articles or emails the back button won't go back to them, or the forward button won't go back to them.Vchimpanzee · talk · contributions · 20:11, 16 July 2010 (UTC)

And it just happened again. I used the back button several times and it wouldn't go forward again because THIS page is so long. I should also add I have IE8 and Vista.Vchimpanzee · talk · contributions · 20:13, 16 July 2010 (UTC)

Have you ever tried Firefox? I used to think Internet Explorer was the bees knees until I actually gave Firefox a chance. Never looked back. This is surely a browser issue, not Wikipedia's. –xenotalk 20:15, 16 July 2010 (UTC)
If this is what I think it is, then it isn't specific to Wikipedia. What might be happening is that when you hit "Back", the previous page is a redirect that sends you to another URL. This then wipes out your "Forward" pages since you are creating a new series of pages, starting with the redirected page, which is essentially the same as you clicking on a link, at least for the purpose of this example. Gary King (talk · scripts) 20:19, 16 July 2010 (UTC)
I don't like Firefox. I've had to use it at libraries. Also, making a major change at home will just make things worse for me.
I know there wasn't an actual redirect, but the behavior of the computer was like when there is one.Vchimpanzee · talk · contributions · 20:30, 16 July 2010 (UTC)
Some wikilinks redirect to a section in an article, which is the same as a redirect and would wipe your Forward history, too. Gary King (talk · scripts) 20:35, 16 July 2010 (UTC)
This is, again, a browser issue but it's like when I click on "Back" and get something I've already backed away from. I wonder if there's a glitch in the function of IE8? Only this time, instead of scrolling down to the specific point in the Wikipedia article (which is sort of like a redirect), it sent me back to an article I had clicked on "back" from.Vchimpanzee · talk · contributions · 20:55, 16 July 2010 (UTC)

I'm seeing the forward button turn gray instead of black when the computer scrolls down in a long article or other type of page. It doesn't usually happen in a short article.Vchimpanzee · talk · contributions · 20:58, 18 July 2010 (UTC)

There is probably nothing we can do here - this seems like an IE issue, not a Wikipedia issue. –xenotalk 16:44, 19 July 2010 (UTC)
I posted this here but didn't get a lot of help. Someone, somewhere, must know the answer to this.Vchimpanzee · talk · contributions · 18:44, 19 July 2010 (UTC)
At the risk of being repetitive and annoying, the answer is that you are using a sub-standard browser. I'm sure that you can customize Firefox to function how you like. If it's just tabbed browsing that you don't want, install the Tab Killer add-on. –xenotalk 18:50, 19 July 2010 (UTC)
I don't like change. Why does no one understand that? IE8 is perfectly good and this is just an annoying glitch that makes me curious. A wholesale change would be much worse.Vchimpanzee · talk · contributions · 20:05, 19 July 2010 (UTC)
Ok. You should probably try to find some kind of dedicated IE support forum; most of our technically-minded users do not use IE and probably won't be inclined to help find a solution to this browser problem. –xenotalk 20:20, 19 July 2010 (UTC)
Or perhaps give a specific example of a page producing the error; I for one am not going to try to find a page that is "very long" with a redirect to a section near the end of the page. I can't even understand the statement of the problem, so I'm glad there are other editors that seem to be able to interpret the issue. —Ost (talk) 21:05, 19 July 2010 (UTC)

I'm sure I mentioned Deepwater Horizon oil spill. I planned to reproduce the sequence of events on the computing reference desk (people were responding more here), but I never got around to it. I can't today because there's not enough time.Vchimpanzee · talk · contributions · 13:14, 20 July 2010 (UTC)

And have you tried Chrome, a rising browser that is getting popular these days ? Or Opera ? There are 5 leading browsers at the moment, and IE is, as a matter of fact, the worst of them. It's only popular and used because Microsoft is using his monopoly in a very unfair fashion.
Basically, you can ask a question to the Microsoft Community. I absolutely won't guarantee that your issue will be fixed (IE isn't popular for solving issues at all, quite the contrary), but you're likely to get an answer. Dodoïste (talk) 08:40, 21 July 2010 (UTC)
My problems will be worse if I add these strange browsers. That I guarantee.Vchimpanzee · talk · contributions · 14:17, 21 July 2010 (UTC)
Also, Firefox has behaved very poorly at the library where I have to use it if ads cause problems. When the computer freezes up there is an ad on the screen with moving information but not what I came to the site to see, or everything is ther but I can't do anything. Or there is some information at the bottom of the screen about some web site relating to an ad. If I try to click or go anywhere else, I get the message "stopped". As if that communicated any information.Vchimpanzee · talk · contributions · 15:16, 21 July 2010 (UTC)
I know I promised I'd stop pestering you about Firefox, but just to response to your issue there, Adblock is an excellent Firefox add-on that can rid you of those meddlesome ads. –xenotalk 15:18, 21 July 2010 (UTC)
I use Firefox at a library. So it's not up to me. As for that being a solution at home, it's yet another way the solution would be worse than the problem.Vchimpanzee · talk · contributions · 16:55, 21 July 2010 (UTC)
More here.Vchimpanzee · talk · contributions · 20:29, 23 July 2010 (UTC)

robots.txt, __ NOINDEX__ and {{userspace draft}}

The __NOINDEX__ and {{noindex}} directives aren't that well-known or well documented.

I was creating subpages under User:Geo Swan for years before I became aware of them.

__NOINDEX__ directive and robots.txt seem to work differently. A robots.txt, with the appropriate contents, tells well-behaved web-crawling robots, like those that search engines use to find contents, will honor the robots from indexing the files in that directory, and all subdirectories -- while the __NOINDEX__ directive only applies to the article on which it has been placed.

I wonder, since, to a web-browser's robots, the subpages of User pages are just files and directories, couldn't a properly drafted robots.txt eliminate the need to put the __NOINDEX__ directive on all our subpages?

If a file has already been hidden from web-browsers, then a {{userspace draft}} would be redundant, and unnecessary, wouldn't it?

Thanks! Geo Swan (talk) 23:21, 21 July 2010 (UTC)

An admin would need to put every user who used noindex on the robots.txt, so it would be rather inconvenient (users couldn't do it themselves) and robots.txt would become rather huge {{userspace draft}} is used on more than 10,000 userpages; even if everyone who used the template had 10 subpages, that would still be 1,000 more things to list on robots.txt. Mr.Z-man 01:05, 22 July 2010 (UTC)
I thought that robots that honored robots.txt directives, look for a robots.txt in every directory they traverse. So, if I create a subpage User:Geo Swan/Guantanamo/robots.txt shouldn't it be able to block compliant robots from everything in that directory?
I would still appreciate feedback on whether the {{Userspace draft}} tag is redundant when web browsers have already been blocked, via robots or NOINDEX. Geo Swan (talk) 02:29, 23 July 2010 (UTC)
From what I've seen, robots.txt only applies in the root directory. Additionally, robots.txt files need to follow a very specific format, just listing entries on a wiki page won't work because of all the other HTML. {{userspace draft}} is the only one of the 3 that provides a visual note in addition to noindexing; if you don't care about the box, then yes, its redundant. Mr.Z-man 03:01, 23 July 2010 (UTC)
Only one robots.txt works per domain. gets only one for the whole site. —Aryeh Gregor (talk • contribs) 21:09, 23 July 2010 (UTC)

Wikitext weirdness

I noticed a minor fault in how (presumably) MediaWiki renders a page, and will describe it in case anyone wants to follow up. The events took place at WT:Requests for comment/Jagged 85#Cleaning up the problem:

  1. At 06:06, 20 June 2010, a new section was added, and the wikitext included a <p> tag (with no closing tag) (diff). Looking at the html source for the page (permalink) shows that something (HTML Tidy?) added a closing </p>, and everything was good.
  2. Several comments were later added with varying indents (colons at left margin). All good.
  3. At 09:08, 28 June 2010, I appended a new comment: blank line and no colon (diff). All good.
  4. At 08:25, 22 July 2010, John Vandenberg appended a new comment: blank line and no colon (diff). Bug! His comment is rendered with no line break before mine (as if there were no blank line between the two comments) (permalink). I tried adding ?action=purge to the page URL, and I examined the wikitext for any strange characters, but can't see any. Bug still present.
  5. I mentioned this at User talk:John Vandenberg#Wikitext weirdness.
  6. At 11:00, 22 July 2010, Jack Merridew fixed the problem by replacing the <p> with a blank line (diff).

Conclusion: the <p> caused MediaWiki to misrender two paragraphs below. Johnuniq (talk) 00:08, 23 July 2010 (UTC)

Not unexpected: the unclosed <p> element contains the indented text, which is parsed according to its own wikitext rules, then the paragraph tag is implicitly closed by HTML Tidy at the next section heading. HTML tag soup does not mix well with wikitext porridge. In fact wikitext porridge does not always mix well with wikitext porridge. It's a quirky syntax at the best of times. — Richardguk (talk) 00:36, 23 July 2010 (UTC)
MediaWiki did rework the initial post by ImperfectlyInformed; it properly wrapped the text as two paragraphs.view source As the thread continued with indented replies, it kept working fine; the indented stuff was not contained by the paragraphs (aside: the whole mechanism of using definition lists for indented replies is semantic mush; fix *that*). It worked when Johnuniq stared a new paragraph, too. But somehow the embedded <p> element addled things such that when JohnV added another implied paragraph (i.e. blank line), MedaiWiki got confused; it failed to parse Johnuniq's && JohnV's posts such that it generated two paragraph; it goofed, and wrapped them in just one; JohnV's blank line being reduced to a mere space. MediaWiki got it wrong, here, and the improperly formed embedded <p> was what triggered the mis-generation of served-code.
Embedded markup can be useful, but most people can't produce well-formed markup; that's why we have wiki-syntax: it's very simple. There's something amiss in sanitizer.php, or nearby, that needs fixing-up. Cheers, Jack Merridew 22:46, 23 July 2010 (UTC)

Need CSS/javascript help with Template:ref_label

Template:ref label wraps its content in a span, rather than a div. This causes problems when the background color is set, because if the line spacing is wide enough then the underlying background color seeps through the cracks. There is a thread about this on the template talk page.

The CSS that is set in Common.css [10] refers directly to the span tag, not just the .citation class name. I have not looked up the javascript that takes care of changing the background color.

I don't know enough about our CSS/javascript setup to try to fix this problem. If someone with the knowledge has time to fix it, it would be appreciated. — Carl (CBM · talk) 00:56, 23 July 2010 (UTC)

I believe this was fixed for {{ref}}; see Template talk:Ref#Text color. ---— Gadget850 (Ed) talk 02:05, 23 July 2010 (UTC)
I don't think that's the same problem. Here's an exaggerated example of what's wrong with the style in question. Make sure it wraps onto multiple lines and look at the gaps between them. — Carl (CBM · talk) 02:13, 23 July 2010 (UTC)
xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx xxx
Looking at the other discussion, it looks like you are talking about the spacing between the lines. ---— Gadget850 (Ed) talk 03:12, 23 July 2010 (UTC)
It is not just that template. Click on (b 2010):
b, a (2010). Lorem ipsum dolor sit amet, consectetur adipiscing elit, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam, quis nostrud exercitation ullamco laboris nisi ut aliquip ex ea commodo consequat. Duis aute irure dolor in reprehenderit in voluptate velit esse cillum dolore eu fugiat nulla pariatur. Excepteur sint occaecat cupidatat non proident, sunt in culpa qui officia deserunt mollit anim id est laborum.
The above uses {{harvrefcol}} which is classed with citation, so I suspect it is something to do with that class. Compare to this link.[1]


  1. ^ Lorem ipsum dolor sit amet, consectetur adipiscing elit, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam, quis nostrud exercitation ullamco laboris nisi ut aliquip ex ea commodo consequat. Duis aute irure dolor in reprehenderit in voluptate velit esse cillum dolore eu fugiat nulla pariatur. Excepteur sint occaecat cupidatat non proident, sunt in culpa qui officia deserunt mollit anim id est laborum.
The above is classed with reference. The reference list is an ordered list inside a div, whereas the template uses a span. I think you are going to have to convert it to use a div. If you get it working, the same fix should be applied to other reference templates. ---— Gadget850 (Ed) talk 03:38, 23 July 2010 (UTC)
Yes, that's what I said originally: it needs to be updated to use a div. That means Common.css has to be edited to match, and also possibly the javascript that does the highlighting needs to be edited. — Carl (CBM · talk) 03:57, 23 July 2010 (UTC)
You can do a sandbox version and add the CSS to your personal CSS for testing, then request that it be added to the site CSS. I don't believe there is any JS involved in the highlighting. ---— Gadget850 (Ed) talk 04:15, 23 July 2010 (UTC)
I don't think that's a good idea. This is a typical example of what should be an inline element, changing it to a block element for the sake of a highlight color, does not make sense. Is not having a continious background over multiple lines really such a big problem ? —TheDJ (talkcontribs) 07:26, 23 July 2010 (UTC)
Cite places the list inside a div, so I don't understand the issue. I have a test at User:Gadget850/t8. ---— Gadget850 (Ed) talk 11:50, 23 July 2010 (UTC)

Citation Checking in Show Preview

When I make changes to an article, I click on Show preview to see my changes. If I put in a Wiki link, for example, I can click on the link in the preview and make sure it works. However, if I make a citation change and click on the superscript citation, I can't see how the citation resolves. Thus, I have to Save page, check the resolved citation, and if I screwed up, go back in and fix it. Is there a way to do what I want without saving the page (short of using a sandbox)?--Bbb23 (talk) 01:14, 23 July 2010 (UTC)

Add this to Special:MyPage/skin.js and purge the page per the instructions at the top of the page:

importScript('User:Anomie/ajaxpreview.js'); // Linkback: [[User:Anomie/ajaxpreview.js]]

At the bottom of the edit window, you will find a button marked Ajax Preview w/Refs. ---— Gadget850 (Ed) talk 02:02, 23 July 2010 (UTC)

Or simply type a temporary references tag at the end of your wikitext:
<references />
then preview – but don't do this unless you can remember to remove the temporary line before saving! — Richardguk (talk) 05:46, 23 July 2010 (UTC)
You can also simply avoid section edit on those occasions. Editing the whole page can be slower, but preview then shows the whole page including the way the cites have resolved. (Anomie's script sounds useful though; must have a look at that myself.) PL290 (talk) 12:11, 23 July 2010 (UTC)

Thanks very much for the suggestions.--Bbb23 (talk) 22:55, 23 July 2010 (UTC)

Main page problem

A recurring problem that makes WP look amateurish is the main page spacing. On the right is a story about Kosovo but there is a picture of a man that looks (pardon my ignorance) like he might be from south India. If the pictures can be lined up, this would be better. Suomi Finland 2009 (talk) 22:11, 23 July 2010 (UTC)

The image belongs to the entry with "(pictured)":
See Wikipedia:FAQ/Main Page#Why are the images on "In the news" and "On this day" not aligned next to each relevant entry? PrimeHunter (talk) 23:04, 23 July 2010 (UTC)

Possible issue with alt-left-arrow in IE8

For a number of months now, a nontechnical Wikipedia editor named Vchimpanzee has occasionally reported problems to the Computing Reference Desk involving a disappearing forward-history in IE8. (Here is his most recent thread, a lengthy attempt to be meticulous in his bug report.) I just reproduced the problem on my machine under IE8 but am having trouble reproducing the problem reliably. My one repro case was that I edited WP:SANDBOX, clicked Link A, then clicked Link B, then pressed Alt-left-arrow on the keyboard, which took me to Link A as expected, but the browser's "Forward" button turned from blue to gray, indicating the forward-history had been lost. Then, the Back command brought me, surprisingly, to Link B.

Is it possible that some feature of Wikipedia's use of AJAX is interfering with the alt-left-arrow or alt-right-arrow keyboard shortcuts of IE8? Without a good repro case that is my first guess at the source of the issue. Comet Tuttle (talk) 21:15, 23 July 2010 (UTC)

<insert obligatory get a better browser remark here> If you suspect it is ajax, disable javascript and try reproducing it again. But this looks more like a browser bug. -- Nx / talk 22:05, 23 July 2010 (UTC)
Something else that might 'invalidate' the forward history is a change in the windows hash (using the TOC to navigate to a section for instance). —TheDJ (talkcontribs) 15:58, 24 July 2010 (UTC)

Dialogs in vector

This is it!

I am having a unique problem with dialogs in Apple Safari 5, Google Chrome, and Opera 10 on Mac OS X Snow Leopard (does not occur in FF4). The best way to explain it is with the image above. Any ideas to fix this? mono 01:28, 24 July 2010 (UTC)

Try clearing your cache. -- Nx / talk 08:56, 24 July 2010 (UTC)
Hmm, error in the editor. "TypeError: Result of expression '$('<div />').attr('id','context',context).appendTo($('body')).each(module.init).dialog' [undefined] is not a function." Reported as bugzilla:24528. Let's hope it gets fixed quickly. —TheDJ (talkcontribs) 15:45, 24 July 2010 (UTC)

Secure server

I've noticed a tendency towards people using (the secure server) when giving diffs. I suppose I can understand that for highly-sensitive abuse issues, but is there is any reason for it on general noticeboards like WP:WQA and WP:RSN? I understand the possibility of man-in-the-middle attacks and so forth, but that seems highly unlikely, so I'm curious whether there is a technical reason for using secure diffs. Johnuniq (talk) 01:39, 24 July 2010 (UTC)

Only that it is often easier to copy-and-paste diff links verbatim rather than using templates such as {{diff}} (which use the secure gateway if you are using it, unencrypted access otherwise) or changing the URLs to point to the standard server. The trend you describe could indicate an increase in the number of people using the secure gateway (or an increase in laziness, depending on your perspective). PleaseStand (talk) 03:08, 24 July 2010 (UTC)
I'm one of the lazy group ;) I notice when logged in via secure that standard HTTP links are automatically reformatted to point at secure. Does this mean that there's a missing feature to point secure links back at HTTP for users on HTTP? Rjwilmsi 09:11, 24 July 2010 (UTC)
There is a script MediaWiki:Common.js/secure.js that rewrites links for users of the secure server. There is no reverse script, because it is a intensive task to rewrite all urls and something that should be avoided in normal browsing. Ideally this is solved in the mediawiki core of course, but this not an easy task. —TheDJ (talkcontribs) 15:48, 24 July 2010 (UTC)
People use secure links, because that is what they copy paste from their addressbar, or from the history of a page. It is understandable that most people have no desire to "rewrite" a url to another form and prefer to copy paste. It has nothing to do with security, other than the security environment that people choose to contribute from (schools, libraries etc are good spots to use the secure server). —TheDJ (talkcontribs) 15:51, 24 July 2010 (UTC)
  • Users of the secure server can use the following script to re-write secure diffs to be non-secure diffs. –xenotalk 15:53, 24 July 2010 (UTC)
importScript('user:js/urldecoder.js') // decodes external links to be internal
Description pages is here: user:js/urldecoder. — AlexSm 16:35, 24 July 2010 (UTC)

Glitch with script

I found a glitch with the script today...when using search (with the old monobook style)...if you do a transwiki search such as "commons:main" you will be taken to the unsecure version of the commons main page.Smallman12q (talk) 01:29, 25 July 2010 (UTC)

How to filter page revision history to see all my edits to a page

I was wondering is there a smart way to filter a page's revision history to see all my edits to the page? Thanks Rjwilmsi 09:12, 24 July 2010 (UTC)

You can do it with API, e.g. prop=revisions&rvuser=Rjwilmsi, there is also a script User:Ale jrb/Scripts#User History. — AlexSm 15:32, 24 July 2010 (UTC)

How to filter specific images in a page, but view others

I want a code that disables some images in a page, but enable others when written on my_user_page/monobook.css. The code img {display: none;} makes all images disappear. I want to write a guide to disable only some images in the page for religious people. Dbachmann advised to ask any of your mates on how to install the Adblock Extension, but I am planning to write a simple guide without using a program outside Wikipedia by just changing the code img {display: none;} that is used in the guide to disable viewing pictures in Muhammad page. This is my user page, For this experiment, I have added 4 pictures, 2 of them are OK and 2 of them are bad for religious people. I want to disable the first and the fourth images on this page but enable the second and the third images at the same time with a simple code on User:Kavas/monobook.css. Is it possible? I have to say that this is not a censorment, because today users can change monobook.css by following a guide to disable viewing all images in a page, but I want to change the guide to filter only specific images. The aim is to write a guide to disable specific images in Muhammad page by changing someone's monobook.css. Kavas (talk) 15:31, 25 July 2010 (UTC)

You could do it in reasonably modern browsers with a rule something like this: a[href="/wiki/File:Aziz_efendi-muhammad_alayhi_s-salam.jpg"] img, a[href="/wiki/File:Siyer-i_Nebi_151b.jpg"] img {
    display: none;
You could also do a whilelist with two rules, one to hide all images and a second to unhide the acceptable images, something like this: img {
    display: none;
} a[href="/wiki/File:Aziz_efendi-muhammad_alayhi_s-salam.jpg"] img, a[href="/wiki/File:Siyer-i_Nebi_151b.jpg"] img {
    display: block;
I'm told this should work in at least IE 7+, Firefox 3.0+, Safari 4+, and Chrome 4+. HTH. Anomie 16:12, 25 July 2010 (UTC)

Template:Sørlandsbanen needs optimization, 613 ifexist

Hi. Template:Sørlandsbanen includes 613 Template:BS, which itself contains one ifexist. The result in the parser count is: “Expensive parser function count: 613/500”. Could someone optimize it? By the way, State highways in New Jersey is in a similar case. Thanks. Dodoïste (talk) 23:19, 25 July 2010 (UTC)

Is there a reason not to make {{BS}} require the value "none" when none is intended instead of using #ifexist at all? In other words, is there reason to use a value that does not exist yet in the expectation that it will be created in the future (and should be automagically used then)? Anomie 03:17, 26 July 2010 (UTC)

Latest change at Mediawiki:Common.css

That latest change disabled my Firefox (3.6.8) browser (using Windows 7) from choosing the font I have specified for IPA in my own user:Mahmudmasri/vector.css. I need help. The latest change made it only possible for my browser to choose my specified font if there were texts written as {{IPA|/tekst/}} only after I wrote that: span[class|=IPA] rather than .IPA, but I still have a problem: all the IPA wikitables such as Help:IPA conventions for English#Vowels don't choose my specified font for IPA. TheDJ told me, “span.IPA or adding !important before the ; should work as well. Because the CSS is now added by Javascript, it is higher in importance than things in your skin css.” I tried what he told me by many ways, but failed. I need help. --Mahmudmasri (talk) 03:11, 26 July 2010 (UTC)

I can't help with your specific problem, but giving sitewide CSS priority over user CSS sounds like complete breakage to me. I know that the ability to set user-specific styles from the software is always going to be fragile, but it shouldn't be required to set !important for user styles. Firstly, the software could take care of it if it is a universal requirement, and secondly, actually setting all user styles to !important could break inline CSS styles. I understand that there is an ongoing effort to condense the loading of CSS and javascript files as much as possible, but this seems like something that ought to be done right. Gavia immer (talk) 03:22, 26 July 2010 (UTC)
Well, thanks Gavia immer for the reply. I wasn't actually requesting that Mediawiki:Common.css latest change be reverted for me. I was just requesting help making/editing my own page(s) that might help overcome the issue, because I tried too much & nothing seems to work with IPA-Wikitables. --Mahmudmasri (talk) 08:42, 26 July 2010 (UTC)
I was about to give the same answer as Gavia immer. Dodoïste (talk) 10:33, 26 July 2010 (UTC) Ps. I edited this message, it previously contained "+1 Gavia immer", I just made it clearer.
What does it mean? --Mahmudmasri (talk) 13:25, 26 July 2010 (UTC)
If anyone has a method of using platform specific CSS, then I welcome hearing it. I don't know how to target "Windows" from pure CSS. Having said that, usage of !important is no requirement. Just use specificity of CSS by prepending a more important element to the front. Like body .IPA for instance. Problem solved. —TheDJ (talkcontribs) 15:51, 26 July 2010 (UTC)
Fixed by not making it span specific and giving it a higher specificity than a normal .IPA rule. —TheDJ (talkcontribs) 15:49, 26 July 2010 (UTC)
Wow! That was nice. Thanks so much TheDJ :) --Mahmudmasri (talk) 05:16, 27 July 2010 (UTC)

Toolserver gone down?

What's going on with the Toolserver? I just clicked a recently-added {{coord}} template and received a HTTP 404 Not Found error; thinking that it was perhaps bad coords, I went to, only to get an identical result there. I'm running IE 8.0.6001.18928, although I doubt that causes any problems. I've tried refreshing, but the same result comes up; however, other websites aren't giving me any problems at all. Nyttend (talk) 15:50, 26 July 2010 (UTC)

SSI error: recursion exceeded ? –xenotalk 15:52, 26 July 2010 (UTC)
  • Should be Fixed now. –xenotalk 15:57, 26 July 2010 (UTC)
Both and the coord template (here's the link I was clicking) work fine now. Thanks for the help. Nyttend (talk) 16:03, 26 July 2010 (UTC)

Strange main page

I'd love to hear an explanation of this. This "secret" page was just exploited by a vandal. — Martin (MSGJ · talk) 21:05, 26 July 2010 (UTC)

Oh, I think it's like the μ/µ problem we had with a template once. Two different symbols which look identical. — Martin (MSGJ · talk) 21:11, 26 July 2010 (UTC)
Yes. The "a" in "Main" is cyrillic. Algebraist 22:17, 26 July 2010 (UTC)

Tables & TOC's

Query: can one force a TOC entry from inside a table? if so, how? Д-рСДжП,ДС 21:56, 26 July 2010 (UTC)

Yes. The header markup == is only recognized at the beginning of a line, thus you must have a newline between the row markup and the header markup:
---— Gadget850 (Ed) talk 22:15, 26 July 2010 (UTC)
Not sure if this is what you are looking for but: You can use the template {{anchor}} to create an invisible link to a particular page-section, that won't show up in the actual TOC. Also related templates in its "See also" section.
If you gave us an example page and context, we'd be able to help more clearly. HTH. -- Quiddity (talk) 22:25, 26 July 2010 (UTC)
Thanks a bunch - both pieces of information were very helpful. Д-рСДжП,ДС 21:10, 27 July 2010 (UTC)

Dealing with Shared IPs. For anonymous users extend the user identifier with a browser cookie

Proposal: For anonymous users extend the user identifier (IP address) with a browser cookie. E.g. "".

Reason: Can help discriminating between good faith contributors and vandals operating from the shared IP address. Currently it is difficult / almost impossible to determine which particular user is editing from the shared IP address. This task is difficult even for humans, but automatic vandalism detection tools are particularly affected. Banning these shared IP contributors or forcing these users to register is ineffective and contradictory to the 'anyone can edit' principle. On the other hand adding a browser cookie to the anonymous user identifier (IP address) is fairly simple and potentially may resolve shared IP problem.

Comments? --Dc987 (talk) 03:00, 27 July 2010 (UTC)

Determined vandals already know how to IP-hop, what makes you think they won't figure out how to delete or change their cookie in under 5 minutes? Anomie 03:19, 27 July 2010 (UTC)
This proposal primarily addresses 'discrimination between multiple users from a shared IP problem', not 'IP-hoping determined vandals problem'. --Dc987 (talk) 04:28, 27 July 2010 (UTC)
But if a vandal can appear to be a separate user for each edit, and possibly even set their cookie to the same as that of a legitimate "shared" user, how does it improve the situation at all? Anomie 11:01, 27 July 2010 (UTC)
Not all IP hopping is deliberate. Mostly is just a byproduct of commercial ISP's use of Dynamic DNS with limited lease times. Conversely most IP sharing is from schools, transit WiFi, and internet cafés where legit and vandal IP edits comingle. Most of the vandals are not terribly sophisticated, they are the modern equivalent of the magic marker graffitist writing "For a good time call ..." Better managing these frees up time and resources for the few who are more adept. —Preceding unsigned comment added by LeadSongDog (talkcontribs) 13:42, 27 July 2010 (UTC)
I know not all IP-hopping is deliberate. But I don't involve myself in vandal fighting, and I've still seen often enough some vandal brag about how he can't be easily blocked because he can IP-hop at will. "Cookie-hopping" will almost certainly be quite a bit easier.
The idea here is that these cookies will somehow allow you to identify which edits are done by a vandal versus by an innocent user sharing the IP. But since it will almost certainly be trivial for the vandal to get another cookie (and therefore temporarily appear to be another innocent user) and may be quite easy for the vandal to clone an innocent's cookie, I can't see how this will accomplish enough to be worth the trouble. Anomie 16:12, 27 July 2010 (UTC)

As a checkuser, I would prefer not to see this implemented. It would be best for me to not say why. --Deskana (talk) 16:13, 27 July 2010 (UTC)

'Nuff said. Point taken. LeadSongDog come howl! 16:30, 27 July 2010 (UTC)
I see your point(s). Yes, this, and cloning an innocent user's cookie are problematic. I guess the only solution here is not showing cookies to regular users. Storing them in the database and xml dump, but otherwise keeping them out of view. XML dump format can easily accept an extra field. But adding an extra field to the database/mw_revision table creates extra overhead. An alternative is using mw_revision.rev_user_text to store the IP address and cookie (e.g. and strip the cookie whenever mw_revision.rev_user_text is used instead of the user identifier. Any developers here who can tell how much of an overhead it is? --Dc987 (talk) 21:30, 27 July 2010 (UTC)
If its not actually displayed anywhere, then what's the point? Mr.Z-man 22:08, 27 July 2010 (UTC)
Plus you can't keep cookies secret. They're stored as plain text on the user's machine. Even if using them had some advantage, they are trivial to disable. OrangeDog (τ • ε) 22:48, 27 July 2010 (UTC)

────────────────────────────────────────────────────────────────────────────────────────────────────I have previously proposed this at Wikipedia:Village pump (proposals)/Archive 60#Effectively warning school IPs in a slightly different form. Essentially, the cookie would be temporary (1 hour expiry) and only used for the purpose of user warnings. The user talk pages themselves (including name) would be the same, and a user would still be warned if he clears his cookies (to the extent that the current system works). Comments? PleaseStand (talk) 00:49, 28 July 2010 (UTC)

Cookies can be kept secret if they are only visible to administrators, checkusers, bot operators and researchers (with the access to database/page dumps). Yes, cookies are trivial to disable, but disabling the cookie would provide some identification by itself. And as long as the cookie is hidden from regular users it wouldn't be possible for a vandal to replicate other users cookies.
Again, the primary goal of this proposal is to identify users sharing the same IP. Currently there is no way to do that at all. --Dc987 (talk) 01:27, 28 July 2010 (UTC)
Hiding cookies only prevents people from hijacking the identity of a legitimate user, it doesn't do anything about people simply deleting their cookies to get a new one. Mr.Z-man 02:09, 28 July 2010 (UTC)
And it wouldn't help at all for users with aggressive privacy settings. If you treat editors as "third-class citizens" for not accepting cookies as a matter of course, our reputation for being "the encyclopedia that anyone can edit" goes right in the toilet. Gavia immer (talk) 02:37, 28 July 2010 (UTC)
Yes. The solution is far from ideal. Users with new cookies and aggressive privacy settings would get a 'clean' (specific to that IP) record. On the other hand if this information would help to identify some large percent of the shared IP users (with regular browser settings, not bothering to reset cookies) it may be worthwhile. An alternative to that could be enforcing shared IPs (engaged in vandalism) into registering user accounts. But that's not in the spirit of 'anyone can edit' rule. Any other alternatives? --Dc987 (talk) 05:49, 28 July 2010 (UTC)

Limits of tables

I'm looking for some advice on the technical viability of a list which is most clearly structured as a large (648-row) 7-column sortable table.

It's a list of the Members of Parliament elected at the United Kingdom general election, 1832. Other similar lists in this series have been implemented as a table split into sub-sections (by initial letter of the constituency name), for example in List of MPs elected in the United Kingdom general election, 1857. This works well enough if the constituency name is known in advance, but for other uses (such as looking for MPs returned for Welsh boroughs), it's not much good, because the sectional structure prevents it from being sorted on other attributes.

So when I was preparing the data for 1832, I kept other data fields as I started building the list, and have now produced a draft with about 45% of the MPs listed: see User:BrownHairedGirl/1832 MPs with gazette. I'm quite pleased with how it works from a reader's POV: the all-one-table sortable columns work well, and I do like having each constituency explicitly referenced to the the authoritative source, the London Gazette.

However, I do wonder whether it's just too big. :(

The markup of 648 references to 32 footnotes and 648 sort-keyed entries is obviously fairly bulky, but the gzipped page is so far only only 44kB, which doesn't seem excessive in terms of download bandwidth, although the rendering of that much markup is probably fairy resource-intensive. The HTML served up the server is currently about 540kB uncompressed, and when the list is complete it will probably be about 750kB.

However, my concerns arise from editing the table. My PC has lots of RAM, but it still seems unhappy about editing the section with the list: after the edit box appears, it takes about a minute of processing before it's ready to allow me to start editing the box. Fine after that, but the delay is significant; my 4GB 1.3Gz laptop is far from cutting edge, but it's no fossil, and I can imagine older PCs struggling very badly. The contents of the editbox are currently about 210Kb, and once complete will be about 300Kb; that's big.

When I save changes, there is a lot of template processing to be done, and it takes a minute or two for the rendered page to be displayed. That's not a problem for me, and it's server issue which won't become more acute with older client PCs.

So, is this big table a step too far? And if so, how should I resolve the problem?

I'm very reluctant to go back to the static segmented table as used in the List of MPs in 1857.

The first option I tried was stripping out the gazette references, at User:BrownHairedGirl/1832 MPs no gazette. That reduces the editbox size by about 50Kb, which probably is not enough to make a great difference.

So I have been wondering about splitting the list into sub-pages, of the form User:BrownHairedGirl/1832 MPs with gazette/A, User:BrownHairedGirl/1832 MPs with gazette/B etc, and then transcluding the subpages into the main page. That will certainly resolve the problems with editing ... but am I right to think that apart from the problems caused by stuffing 200kB into the editbox, the other points are all manageable?

(Sorry if I have missed some relevant guideline. I have checked Help:Table, WP:TABLE and WP:LIST, but can't find anything which addresses this issue). --BrownHairedGirl (talk) • (contribs) 15:01, 23 July 2010 (UTC)

Are you using Internet Explorer? IE is notoriously slow at rendering larger pages.
You could optimise the page by removing repeated templates (individually using subst or for whole blocks of text via Special:ExpandTemplates) and possibly manually simplifying the resulting wikitext.
But it is likely that readers who are not logged in will benefit from caching and lack of gadgets, so they probably already see the page download and render faster than you do.
Richardguk (talk) 22:07, 23 July 2010 (UTC)
Eeek!!! Me use exploder???? Does the Pope pray to Mecca?? {mock horror and indignation ...}
No, I'm using Firefox, though currently I'm using it under Vista (rather than my usual Linux setup), while I'm doing some WP:AWBing. Vista is so horrible that I am happy to blame any probs on Vista.
I'm reluctant to subst/expand any templates, because that will complicate maintenance, but I can see that it may be necessary. I'll see how it looks after I have completed the list and split it up into sub-sections. You're probably right that cacheing will eliminate template-parseing delays for readers, so I'll check that from a logged-out browser when I'm done.
Thanks for your help. --BrownHairedGirl (talk) • (contribs) 15:32, 24 July 2010 (UTC)

Update on table limits

On the off-chance that anyone is interested, this is just a quick note to say that I have completed the list and moved it to mainpsace at List of MPs elected in the United Kingdom general election, 1832. I think that by splitting up the list into sub-sections and slightly reducing the number of template calls, it all works acceptably, apart from taking a long time to save changes to the main list. I have left a few notes on the technical issues at the list's talk page. --BrownHairedGirl (talk) • (contribs) 13:12, 28 July 2010 (UTC)

The two hyphens standing before the signature

Something that annoys me is the two hyphens standing before the signature when I press the signature button to sign my edit:--~~~~ Is there any way to remove this, because it annoys me to remove them every time I sign my edits. MichaelJackson231 (talk) 08:24, 26 July 2010 (UTC)

You could just type 4 tildes manually. --Cybercobra (talk) 08:31, 26 July 2010 (UTC)
You could use User:MarkS/extraeditbuttons.js to rewrite the toolbar. –xenotalk 15:53, 26 July 2010 (UTC)
3) You could use the ~~~~ link that is below the edit window, instead of the button above. :) -- Quiddity (talk) 17:22, 26 July 2010 (UTC)
(Tangentially, for when this was last discussed (afaik), see MediaWiki talk:Edittools/Archive 5#Signature button (2006), Wikipedia:Village pump (proposals)/Archive 5#Should the default signature have two hyphens in front of it? (2007), Wikipedia:Village pump (proposals)/Archive 13#Disparity between signature (~~~~) toolbar buttons and character map (2008), and MediaWiki talk:Common.js/Archive 15#Mdash in signature button (2008). HTH. -- Quiddity (talk) 17:42, 26 July 2010 (UTC)
If it annoys you to remove them, then why not simply not remove them and get on with helping to write this encyclopedia? That's what we are supposed to be doing here. Phil Bridger (talk) 01:53, 27 July 2010 (UTC)
I second Cybercobra's motion; I always type the tildes, and I had to search for a while to find the signature button. Nyttend (talk) 17:13, 28 July 2010 (UTC)

Turn off talk page creation for IP users when the page does not exist

Let's say Example page does not exist. An IP user cannot create Example page. They can create Talk:Example page. New page patrollers have a blind side to talk page creation on the assumption the article page exists, and once created, nobody is watching it. Or worse, the talk page becomes a fake article written by some banned user, and a new page patroller helps them out by moving it from talk to article space.

So, is it a Wikimedia feature to not allow talk page creation when the article doesn't exist? Should it be? SchmuckyTheCat (talk) 20:33, 27 July 2010 (UTC)

Good point. We have Category:Talk pages with no main page, but it is empty and I don't know if it is used. Fences&Windows 20:51, 27 July 2010 (UTC)
I think there was a bot that maintained that category that has not been run in years. I'm wondering if there isn't already a MediaWiki feature that can just be turned on. SchmuckyTheCat (talk)
There is a periodically updated report of such talk pages at Wikipedia:Database reports/Orphaned talk pages. Svick (talk) 15:17, 28 July 2010 (UTC)
I don't believe the decision to disallow article creation by IP users was ever supposed to be some sort of inviolable principle that can't have workarounds for the determined. Back when IPs could create articles, something like 70% of new articles were created by IPs, and a significant fraction of them had to go to AfD, bogging down the process. The restriction stanched the flow somewhat. --Trovatore (talk) 21:05, 27 July 2010 (UTC)
Of course the determined will do it, but it shouldn't be easy or successful via trickery. We've created a front door process for those with good faith at Article for Creation. SchmuckyTheCat (talk) 02:24, 28 July 2010 (UTC)
Perhaps this thread should go to WP:VP/PR, but wherever it is, it's a good idea. Why don't we ask the developers to prevent anyone, even admins, from creating talk pages for articles that don't exist? I can't think of why we ever would need to create a talk page for a nonexistent article; the only good reasons for the existence of a talk page for a nonexistent article all apply only in situations where the article has been deleted. If someone came up with a situation in which a talk page was somehow needed for a never-created article, they could create text at the article, create the talk page, and nominate the article for speedy deletion. Nyttend (talk) 17:10, 28 July 2010 (UTC)
I can think of three reasons offhand: talk page archives, user talk pages, and testing (in a user sandbox or template test case subpage) for templates that whine when placed on a non-talk page. The first applies even if you only meant namespace 1 rather than all talk namespaces. Anomie 17:42, 28 July 2010 (UTC)

un-usable diffs

Looks like changing a few words in the middle of the first sentence of a paragraph causes the software to think that it's a different paragraph.

See [11]. The first paragraph should be side-to-side with the new first paragraph, since only 3 or 4 words have. The following five paragraphs are pushed down even if they didn't change at all.

I'm seriously considering printing the diff so I can highlight it with a red pen. --Enric Naval (talk) 13:03, 28 July 2010 (UTC)

  • My educated guess is that what much more probably caused it to think that there were different paragraphs was that xe deleted the blank line between the section heading and the first paragraph. By the way, if you use <references></references> or {{reflist|refs=}} you can place all citations in the references section instead of in the middle of the prose. That makes it easier to distinguish, as in cases such as this, between changes to citation template use and changes to prose. (There are a couple of prose changes in there that the changes to the templates are obscuring.) It also makes the prose easier to read in the raw wikitext. (It can be made to closely resemble Harvard style in the raw wikitext.) Uncle G (talk) 13:38, 28 July 2010 (UTC)
    • (thanks for the "refs=" tip, I just implemented it for two huge references) Ah, this looks like "removal of blank line" + "alteration of the first sentence of the following paragraph" = "total mismatch of paragraphs"? --Enric Naval (talk) 23:47, 28 July 2010 (UTC)
      • Actually, I was talking about something like this, where I've done one section as an example. Uncle G (talk) 03:06, 29 July 2010 (UTC)

Template oddity

I made a couple of minor edits[12] to this template and now it won't display correctly... any idea what's going on? I'm obviously missing something very simple!--Pondle (talk) 16:50, 28 July 2010 (UTC)

One missing bracket [13] - fixed. —DoRD (talk) 16:59, 28 July 2010 (UTC)
I found the same thing as DoRD. For future reference, sometimes I find it useful to use WikEd with templates (the color coding and such help). Killiondude (talk) 17:01, 28 July 2010 (UTC)

Organising watchlist

Hello, I've googled and searched Wikipedia for info on this but it doesn't seem possible. Basically i just want to organise my fast growing watchlist into categories for easy reference. I've tried editing the raw watchlist to change the wording, but that seems to change links and it a bit of a faff anyway. Is there really no way to add pages to categories, in the same way as adding a bookmark to your browser and deciding which section to put it in? Any help much appreciated and apologies in advance if i've been thick and missed something. —Preceding unsigned comment added by Richyratton (talkcontribs) 10:18, 28 July 2010 (UTC)

One way around this would be:
  1. Create a page in your userspace for each group of articles (one page per group), simply containing a list of links to the articles in the relevant group.
  2. Then, on another userspace page, create links to Special:RecentChangesLinked to each of the above list pages.
Hope that helps.
Richardguk (talk) 11:04, 28 July 2010 (UTC)

Many thanks Richard - still seems a bit complicated, but then I don't really understand how wikipedia links - for instance, i was looking for a 'reply' button to get back to you on this, but it seems i have to edit the whole thing? Is there some techy reason in the coding why browser style bookmarking can't be implemented? kind regards. —Preceding unsigned comment added by Richyratton (talkcontribs) 13:06, 28 July 2010 (UTC)

Perhaps you should take this to the village pump for proposals. Simply put, you're asking for something that doesn't exist, but which would definitely be quite useful. I'm surprised that what you want doesn't already exist, because it would definitely be useful, and I suspect that the developers could do it. Thanks for making a very good suggestion! By the way, Richardguk's idea could work, but please be aware that nobody else can know whether a page is on your normal watchlist, but if you create a system of pages like Richard proposes, anyone will be able to see them. That's fine if you don't care about others seeing your watchlist, but if you ever decide that you don't want people looking at it, you might have trouble. Finally, no there's no "reply" button; to reply to something, click the "edit" button at the top of the relevant section header; in this case, click the "edit" button next to the big "Organising watchlist" letters. Nyttend (talk) 15:29, 28 July 2010 (UTC)
I agree, the ability to create folders or categories in which to group watchlist entries would be very handy. People have got used to not having it, so there may be some "what's the point" inertia, but I agree, make the suggestion at the pump and see what happens. PL290 (talk) 18:08, 28 July 2010 (UTC)

Thanks everyone, have suggested it at village pump for proposals. --Richyratton (talk) 10:45, 29 July 2010 (UTC)

769 dead links to the US State Department

(Reported at the administrator noticeboard, they suggested bringing it here.)

Sometime before January 21, 2010, the US State Department changed its website address from to without providing any sort of redirection for existing links. That change broke all citations to They are still broken.

The situation was reported at EAR here. An EAR editor confirmed the problem and reported it to WikiProject United States Government here. Trying to follow a citation today, I discovered it still exists.

A Linksummary shows ~769 links to, all of them broken. Clicking on any returns server not found. Linksearch en (https) - meta - de - fr - simple - wikt:en - wikt:frMER-C X-wiki • Reports: Links on en - COIBot - COIBot-Local • Discussions: tracked - advanced • COIBot-Link, Local, & XWiki Reports - Wikipedia: en - fr - de • Google: searchmeta • Yahoo: backlinks • Domain:

I don't have any bright ideas, but I think this should be addressed. I'll volunteer to fix a couple of dozen (it's not a one-for-one fix, things have moved around) if a project is set up. Interestingly, we have more of these now than we did in January. --CliffC (talk) 21:22, 28 July 2010 (UTC)

While I appreciate you raising this issue, there isn't anything administrators particularly can do about this. You might have more luck raising the issue at the technical village pump or filing a bot request to fix the deadlinks in batches. Regards, Skomorokh 22:58, 28 July 2010 (UTC)
If things have moved around in a predictable way, it could be a job for Wikipedia:AutoWikiBrowser/Tasks. Rd232 talk 08:18, 29 July 2010 (UTC)
No, the new addresses are unpredictable – in the sample I checked, plugging '' into the url gets you to the State Department (motto: Engaging the world"), where a catch-all message says "The Page You Are Looking for Has Moved. This Web site has been redesigned and restructured, and the page you are looking for may no longer be available." If a bot did anything, I suppose it could substitute the '' and flag the citation 'failed verification' to get the attention of regular editors of the affected articles. I'll ask the folks at the village pump to take a look. --CliffC (talk)
Are you sure that there is no pattern, not even a complex one, for most of the addresses? Changes like this usually produce a small number of very common links with unpredictable changes (reflecting departmental or server directory restructuring) plus a large number of obscure links which were moved on a series of patterns (typically, from sub-pages of one department/directory to sub-pages of a new department/directory). Once the pattern is identified, a bot could easily fix both types. — Richardguk (talk) 12:41, 29 July 2010 (UTC)
Upon a second look, sometimes there is a pattern. Omar al-Tikriti has an EL titled "Treasury Department Blocks Assets of Saddam Hussein's Nephews", here is the old link vs. the assumed new link, based on an exact match to the EL title, no apparent match:
However, at Dick Dale, citation ""Guitarist Dick Dale Brought Arabic Folk Song to Surf Music", there is a match. Old vs. new:
We might want to flag those where a new link cannot be developed as 'fails verification' and let the article experts resolve them. Simpler links have multiple possibilities;
Finally, although I can't recall which subjects I used the site's search feature to look for in the past, some articles seem to have disappeared entirely. --CliffC (talk) 14:42, 29 July 2010 (UTC)
{{Failed verification}} isn't for the purpose you advocate; as its page mentions, that's what {{dead link}} is for. Personally I like the idea of temporary task force or letting these get fixed organically. I haven't look at many links, but everyone one I have checked has had an archive at the Internet Archive. Tools like WP:CHECKLINKS can at least be used to find the archive if the new location cannot be found. —Ost (talk) 18:26, 29 July 2010 (UTC)
(For completeness, copied the following from the administrator noticeboard) --CliffC (talk) 23:52, 29 July 2010 (UTC)
As a general approach, find the broken url, identify when it was cited (this may require digging back through version history), search for the old url. Find the archived version just before the date cited. Open that archived url. Search for that first sentence on the new domain, e.g. this, and compare what it finds to the archived version. Per WP:SAYWHEREYOUGOTIT, cite the archived version and tag the cite with {{update}} and a comment to help editors link to the new one once they can verify that it still supports the statements for which it is cited. LeadSongDog come howl! 13:28, 29 July 2010 (UTC)
The idea of Googling for the first sentence is a good one. I picked a random article, Dhabihu'llah Mahrami, used WP:CHECKLINKS to find the archived State Department link, opened it, Googled for the first sentence, found a current copy at State, updated link. Lather, rinse, repeat, 768 to go. --CliffC (talk) 23:52, 29 July 2010 (UTC)

──────────────────────────────────────────────────────────────────────────────────────────────────── Here are two links found at random that are quite similar:

--CliffC (talk) 00:34, 30 July 2010 (UTC)

I can fairly easily craft some code in AWB that will replace the old link to a new one based on the Dick dale example above. Then we can see whats left and fix those if that seems like it might be ok with everone here. Of course they are not all going to work but hopefully well get more than not. --Kumioko (talk) 18:22, 22 October 2010 (UTC)

duplicate editnotices in help talk namespace

Hi, if one edits say Help talk:Moving a page, you get two editnotices saying the same thing ( Template:Editnotices/Namespace/Help talk ), I started to look into it but then noticed a sign saying 'caution technical: your head might explode!' so have turned up here before that happened! Lee∴V (talkcontribs) 13:57, 29 July 2010 (UTC)

I see Template:Editnotices/Namespace/Help talk only once there and MediaWiki:Talkpagetext before that. Svick (talk) 14:14, 29 July 2010 (UTC)
Maybe I've mucked up my preferences! Just tried whilst logged out and only get one, followed by page length followed by 'You are not currently logged in. Editing this way will ...' message, ( logged in it is msg, page length, msg ), think I'm using WikEd. Lee∴V (talkcontribs) 14:35, 29 July 2010 (UTC)
The MediaWiki:Talkpagetext message is before these in both cases. Lee∴V (talkcontribs) 14:37, 29 July 2010 (UTC)
Apparently, this is a feature of wikiEd. See User talk:Cacycle/wikEd#Would you consider making the duplicate edit notices optional? with an explanation how to disable this. Svick (talk) 14:50, 29 July 2010 (UTC)
Thanks Svick, strange feature! I'll have to reread the explanation again with a clearer head! Lee∴V (talkcontribs) 15:01, 29 July 2010 (UTC)

Table formatting

Hi, I'm hoping this is obvious to someone ... at Template:Lexus, I'd like to suppress the white border between a couple of cells, specifically the first row of timelines against "Entry-level luxury" where the two halves of 1989 are distinguished. The only reason they're distinguished is to help control the width of each column; that whole table uses %ages for widths and so any unmanaged column width can be out of proportion. I've tried various keywords and CSS approaches, with no success. TIA --AndrewHowse (talk) 20:30, 29 July 2010 (UTC)

Example.png and Template:Image label begin

I occasionally clean up user tests by looking at where Example.jpg, Example.png etc. are used. The "what links here" for Example.png (this page) comes up with Shanghai (among a few others), which gets it from {{Shanghai Labelled Map}} which in turn gets it from {{Image label begin}}. As most of the users of this template don't show up, I'm guessing that there's something wrong about the use of this template specific to its use in Shanghai. Perhaps a a parameter wasn't supplied properly, but I confess I can't fathom what. Can someone help figure out what's wrong, so we can clear the link list? Thanks. -- Finlay McWalterTalk 22:45, 29 July 2010 (UTC)

Major API breakage in the latest MediaWiki update

An hour or so ago, we were apparently updated to r70061. This seems to have broken the use of "max" as a value for the various limit parameters in the API; attempting to use them will now give a fatal error. This has already been reported as bug 24564, hopefully they fix it soon. This may cause user scripts and other tools that use the API to stop working. Anomie 03:49, 28 July 2010 (UTC)

Tim has already rolled out the fix for this issue onto the WMF cluster. Peachey88 (T · C) 06:26, 28 July 2010 (UTC)
Not completely fixed, the backlinks, deletedrevs, and revisions modules are still b0rken. Anomie 14:45, 28 July 2010 (UTC)
Now it seems to be fixed. Anomie 11:10, 30 July 2010 (UTC)

Persistent page on watchlist


For at least three or four years, the phantom page on my watchlist "Wikipedia:BJAODN/C" has been on my watchlist. For some reason it will not disappear from my watchlist despite any removal actions. The page is long gone except for the watchlist...even creating then deleting it long ago did not work. Today I use Camino 2 (= Firefox 3.0), but this issue is much older and goes back to 2007 or so.

Can anybody provide any sort of technical assistance? Raymie Humbert (local radar | current conditions) 04:41, 30 July 2010 (UTC)

When I had a problem with "sticky pages" on my watchlist, the following worked:
  1. Go to Special:Watchlist/raw, the plain text version of your watchlist.
  2. Copy the entire plain text copy of your watchlist somewhere you can keep it safe for the next step - for instance, save it as a text file on your home computer, or copy it to a page in your userspace - be aware that userspace pages are somewhat public, so if you want to keep your watchlist private you should be careful with this option.
  3. Delete the entire contents of the plain text watchlist and save it as completely empty. After saving it, verify that you have no pages on your watchlist at all.
  4. On your saved copy of the watchlist, find the "sticky" entry and remove it from the text.
  5. Now, copy the saved version back to Special:Watchlist/raw. Do not press "Save" until you have verified that the "sticky" entry is not present.
  6. After saving, verify that your watchlist works and that the "sticky" entry is gone.
Hope this helps. Gavia immer (talk) 05:00, 30 July 2010 (UTC)
Thank you, Gavia! It worked like a charm. Raymie Humbert (local radar | current conditions) 20:29, 30 July 2010 (UTC)

Wikiklink underlining randomly missing


Today, now and again -- but only occasionally -- I get whole pages that have the wikilinks not underlined (this is a setting you can choose in your prefs, mine is set to always). Anyone else have this happening? ♫ Melodia Chaconne ♫ (talk) 00:12, 31 July 2010 (UTC)

It has something to do with the notice that appears periodically at the top of my screen about jobs. --Golbez (talk) 00:37, 31 July 2010 (UTC)
I have passed along this information via IRC. The author of the centralnotice will take care of it. —TheDJ (talkcontribs) 00:57, 31 July 2010 (UTC)
Thanks for reporting the issue. I've fixed the problem that was causing it and the fix should be propagated to all the other wikis within the next hour. --Az1568 (talk) 01:46, 31 July 2010 (UTC)

Odd text reversal

When I go to edit the Lujerului metro station article the text is backwards (right-to-left), yet when viewing the article normally, Mediawiki renders it the correct way (left-to-right).

It appears that this edit simply reversed the text in the infobox template, and which as a result affected the entire article for some reason.

Can someone tell me why MediaWiki does this? I'm pretty sure it's not a bug but I just don't know what the purpose of this 'feature' is. My guess is that it's a side-effect of the transwiki import process from right-to-left language Wikipedias. -- œ 07:14, 31 July 2010 (UTC)

The edit added some odd Unicode characters; one of them must have been a Right-to-left mark. I've fixed it. Graham87 08:22, 31 July 2010 (UTC)
Hmmm, apparently not; all the Unicode characters used in that diff were numbers 8534–8538 decimal, or 2156–215A hex, and are found at the start of the table at the number forms article. Since the IP made several similar edits to the one at the Lujerului metro station, they were probably exploiting a bug, maybe one in Internet Explorer, that causes that particular sequence of characters to be interpreted like a right-to-left mark. There are some advantages to using screen readers; even though JAWS couldn't tell me what those characters were, it could tell me their code points! Graham87 08:46, 31 July 2010 (UTC)
Interesting that it may be a bug.. I am using Firefox btw. -- œ 08:55, 31 July 2010 (UTC)
I don't see a bug. I copy-pasted the start of the edit into the Characters field at and clicked "View names". It started with "U+202B: RIGHT-TO-LEFT EMBEDDING", and similar things followed. PrimeHunter (talk) 09:10, 31 July 2010 (UTC)
Oopss, I misremembered the numbers, hence my above post. I was in a bit of a rush. Graham87 09:35, 31 July 2010 (UTC)
Where is this right-to-left character? Why can't I see it? I'm unable to copy-paste any extraneous characters, when I go to select the start of that edit I'm only able to select the backwards "{{Bucharest metro stations" and nothing else before or after. Could this be related to the font I'm using or something in my preferences that's not allowing me to view or select certain unicode characters? -- œ 17:45, 31 July 2010 (UTC)
The character doesn't actually display on the screen in modern browsers, it just affects the bi-directional text layout algorithm to hint that characters should be laid out right-to-left (as in Arabic or Hebrew); see Unicode control characters#Bidirectional text control for more details. You probably are able to select it (it should be to the right of the "first" character in the text), but as it is non-printing you get no visual feedback. In my Firefox I can even (carefully) use the Delete button to delete these extraneous characters, but I personally like to use It's All Text! to open it in Vim which displays these characters visually. Anomie 18:40, 31 July 2010 (UTC)
Yes that's what I'm saying.. I'm unable to select it.. there's nothing there to the right of the first character. I know what it's for and what it does and that i'm aware that it wouldn't be visible on screen but I thought I'd be able to select it and I can't, it's just not there, that's why i'm wondering if it's because of certain settings in my preferences. I'm using Firefox with the regular plain editor that vector uses. -- œ 02:39, 1 August 2010 (UTC)

<math> tag does not inherit style sheet when generating a LaTeX png

I'm not sure if this is a bug, or a deliberate feature, but when the <math> tag decides to generate a LaTex PNG it defaults to producing a black-on-white image, even when the tag is embedded inside a table that has a background colour, for instance

Plain text Complexs maths starts and ends

Should the LaTex PNG generator be told the the style of the region it is in, and default to the style? I know it does support foregrounds, but does not appear to support the colorbox or pagecolor functions to set the background. CS Miller (talk) 12:15, 31 July 2010 (UTC)

<div style="display:inline-block;background-color:#2d2f92"><math>\pagecolor{Blue}\color{White}\mbox{Pagecolor works, but it also makes the background transparent.}</math></div> In order to pass that style information to the PNG rendered, MediaWiki would have to parse the whole page (including all user css) to determine just which styles apply. Anomie 12:41, 31 July 2010 (UTC)
The PNG rendering is cached from before transparent backgrounds were switched on. Any new equation will have binary transparency. — Dispenser 12:53, 31 July 2010 (UTC)
Yeah, we just have to live with the fact that until MathML and mathfonts become standard in all browsers and Operating systems, rendering of Math will mostly just suck. —TheDJ (talkcontribs) 12:57, 31 July 2010 (UTC)
This means you can force transparency by making irrelevant changes to the code:
Plain text Complexs maths starts and ends
Unfortunately, the result looks absolutely awful. Algebraist 13:03, 31 July 2010 (UTC)
This doesn't look to shabby
{| border="1" style="background: darkgreen; color: yellow" 
| Plain text
| <math>Simple maths</math>
| Complexs maths starts <math>\definecolor{Darkgreen}{rgb}{0.145,0.255,0.090}\pagecolor{Darkgreen}\color{Yellow}\sqrt[3]{x^3+y^3 \over 2}</math> and ends
HumphreyW (talk) 13:16, 31 July 2010 (UTC)
The reason I initially asked was because on the Cavendish experiment page, in the table at the end, the units are sometimes rendered as PNG. The table they are in has a light-gray background which almost, but not quite, blends into the white of the PNG. It's a bit distracting once you notice it. As I said, it would be nice if the PNG had the correct background. However, if the user doesn't have white as the default background for normal pages, then it could go wrong, but no worse that normal. CS Miller (talk) 14:00, 31 July 2010 (UTC)
For me the entire table at the bottom of Cavendish experiment is 100% text with no PNGs. I have my settings to use HTML text where possible. So you could probably rewrite the table to use just superscripts and normal text. There isn't really any need to use the math tags in that table. HumphreyW (talk) 23:27, 31 July 2010 (UTC)

Work at Wikipedia

What is this? It's not a link. It is, however, kind of annoying.Vchimpanzee · talk · contributions · 16:09, 31 July 2010 (UTC)

It's a link for me. Idea, were you trying to open it in a new tab/window using Internet Explorer ? Perhaps you experienced the same bug that the pageheaders have... —TheDJ (talkcontribs) 17:18, 31 July 2010 (UTC)

Bad sort on sortable wikitables

See, e.g., Divorce demography. Sorting the table by the column "Crude divorce rate" in descending order puts the United States (3.6), correctly after Moldova (4.14), but incorrectly ahead of Puerto Rico (3.67) and Ukraine (3.66). --Wfaxon (talk) 17:01, 31 July 2010 (UTC)

Due to all the text that is in the columns, the text sorting is triggered, instead of the number sorting. See also Help:Sorting. You can use {{Nts}} on each number in that column to correct for this problem. —TheDJ (talkcontribs) 17:22, 31 July 2010 (UTC)

comments not being displayed

Something is blocking the display of comments here: Noloop (talk) 17:31, 31 July 2010 (UTC)

Some one did not close a ref tag.Slatersteven (talk) 17:53, 31 July 2010 (UTC)

Henrik's traffic counter

Henrik's traffic counter is down. I'm desparate to know the number of page views for a DYK... Please fix it... Kayau Voting IS evil 01:02, 1 August 2010 (UTC)

Image notes

I added image notes to two photos stored on Commons, File:State Govt Complex.jpg and File:Michigan Ave from Boji Tower.jpg. Is there anyway to get them to show up on the enwp file description page? I added both to Capitol Loop, but because of the number of state buildings in the first photo (7) I wouldn't want to try to list them in the caption. The easiest thing to do would be point the reader to click on the photo to see them, but the image notes only seem to work on the Commons file page. Imzadi 1979  22:45, 30 July 2010 (UTC)

That would require the ImageAnnotator gadget to be locally installed on en.wikipedia. Making this addition was suggested in this discussion, but that wasn't conclusive. It looks like this page was prepared to host a proper consensus-building discussion in that regard, but seems to have fallen idle without reaching a proper proposal. It was again proposed here, again without going anywhere. I'd hazard that this isn't a very controversial feature, but someone who cares about this feature is probably going to have to take some more assertive action to get it approved here. -- Finlay McWalterTalk 23:37, 30 July 2010 (UTC)
I propose we just add it as a gadget first and see where it takes us. I just checked if we have the latest versions of all the required files, and I have set up a config that only shows the annotations on File description pages, with no option to edit. What do you guys think ? —TheDJ (talkcontribs) 01:39, 31 July 2010 (UTC)
Well, I'd appreciate any method that let's me annotate which building is which in File:State Govt Complex.jpg, since there are 7 state office buildings, including MDOT's building in downtown Lansing. Trying to list them in the photo's caption in the article would be unwieldy. I could try to describe them in on the image page, but that's cumbersome as well. Imzadi 1979  01:57, 31 July 2010 (UTC)
As mentioned the last few times this came up, I think that a gadget is the right way to do this, and that enabling annotations for casual viewers is the wrong way to do it. Given the amount of interest that is now popping up, I'd love to see a gadget implemented. Gavia immer (talk) 01:43, 31 July 2010 (UTC)
I have installed the gadget. If you enable it, you can view the annotations of File:State Govt Complex.jpg. The sitewide config file for annotations is MediaWiki:ImageAnnotatorConfig.js. I personally wouldn't mind loading this code in view only mode for all users, for the file namespace, but now people can test that out, and we can have a more informed discussion about it. —TheDJ (talkcontribs) 14:28, 31 July 2010 (UTC)
Thanks, it seems to be working well. So that it gets some usage, I intend to publicise this in the tech section of the Signpost. To that end, I think it's fair to say this is a "trial" for now, yes? I've also asked at the Commons Village Pump for their collective experiences on it (which are, as far as I know, positive), which I'll summarise for said Signpost piece. -- Finlay McWalterTalk 18:17, 31 July 2010 (UTC)
Tech news in the signpost seems a tad busy this week, perhaps next week is better edition for an invitation. —TheDJ (talkcontribs) 13:45, 2 August 2010 (UTC)

Direct link to a tab in Special:preferences

Is it possible to directly link a user to a tab in hir preferences page? Something like [[Special:Preferences/Watchlist]]. If necessary, a small bit of Javascript could query the URL the page was loaded from, detect the tab in it, and the click on it. CS Miller (talk) 16:18, 1 August 2010 (UTC)

This is bugzilla:16013. I believe the usability team is now in the process of working on these scripts, so perhaps they can take care of it. —TheDJ (talkcontribs) 13:30, 2 August 2010 (UTC)
Thanks. CS Miller (talk) 19:30, 2 August 2010 (UTC)

Link from Special:preferences to Special:Watchlist

Following on from the above question, I think it would be useful if the [[Special:preferences]] Watchlist tab had a direct link to [[Special:Watchlist]] or even [[Special:Watchlist/edit]]. A typical user's response to the preferences/Watchlist tab might be 'yes, but how do I actually edit my watchlist?' CS Miller (talk) 16:18, 1 August 2010 (UTC)

Opened a ticket as bugzilla:24632TheDJ (talkcontribs) 13:32, 2 August 2010 (UTC)
Thanks. CS Miller (talk) 19:30, 2 August 2010 (UTC)

Remove"Privacy Policy" "About Wikipedia" "Disclaimers" from printable version in Vector


When printing from the Vector skin, all page are appended with

  • Privacy Policy
  • About Wikipedia
  • Disclaimers

This is unnecessary and was removed in monobook. Could someone please remove this for the Vector skin...I don't want to waste pages every time I print...18:53, 1 August 2010 (UTC) —Preceding unsigned comment added by Smallman12q (talkcontribs)

Temporarily fixed in MediaWiki:Print.css, and fixed upstream in r70333. —TheDJ (talkcontribs) 13:20, 2 August 2010 (UTC)

Repeating citation templates

With <ref name="title"> one can repeatedly cite the same source by using <ref name="title" />. Is there something comparable for the {{citation}} template? ENeville (talk) 19:35, 1 August 2010 (UTC)

<ref> and {{citation}} are usually used together, so if you want to repeat some citation, you just repeat the whole <ref>. Why do you want that? Svick (talk) 20:52, 1 August 2010 (UTC)
Uh-oh, I fear I'm too ignorant to understand the clarification. I just wanted to use a citation template (for the benefit of standardized data formatting) to create an initial footnote, then use something (some kind of <ref> tag?) to refer to that same footnote again after that. Or is there a version of {{citation}} for followup citations? Succinctly, I guess, I'm trying to exploit the data formatting of {{citation}} without having the clunkiness of repeating its giant template throughout the text. ENeville (talk) 21:28, 1 August 2010 (UTC)
I guess I don't understand. If you want to repeatedly cite some source using {{citation}}, you can use something like
Some text.<ref name="the book">{{citation | first = First | last = Last | title = The Book}}</ref> More text.<ref name="the book" />
which renders as:

Some text.[1] More text.[1]

  1. ^ a b Last, First, The Book 
That way, you don't repeat the whole citation the second time you use it. Svick (talk) 21:44, 1 August 2010 (UTC)
Either I misunderstand your intention or this is likely to be a duh moment for you. {{citation}} and friends usually go inside a ref. Use the same method to duplicate a ref whether you have a template or plain text inside the ref. Write <ref name="title">{{citation|...}}</ref> one of the times and <ref name="title" /> the others. If you want to change a page number or other detail in the different uses of the citation then this doesn't work. PrimeHunter (talk) 21:42, 1 August 2010 (UTC)
Oohhh. I thought {{citation}} created the whole superscript numeral dealio, and I was getting too many error messages in my tryouts to troubleshoot what was up. Thanks much. ENeville (talk) 22:36, 1 August 2010 (UTC)
The Cite extension uses <ref>...</ref> tags to create the numeric in-text citation, with the actual citation included inside the <ref>...</ref> tags but showing in the reference list. The citation can be formatted with a template. ---— Gadget850 (Ed) talk 22:40, 1 August 2010 (UTC)
OK, I think I have a much clearer picture now of the technical break-down. :-) ENeville (talk) 00:54, 2 August 2010 (UTC)

COinS metadata for articles

Our citation templates, and infoboxes for books and journals, emit COinS metadata. Can we find a way to also emit the same thing for each article, about the article itself? Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 20:18, 1 August 2010 (UTC)

Once again, do more than two people think it's a good idea? OrangeDog (τε) 21:23, 1 August 2010 (UTC)
What do you mean about the article itself?Smallman12q (talk) 23:05, 1 August 2010 (UTC)

────────────────────────────────────────────────────────────────────────────────────────────────────OrangeDog: Yes; though I don't recall discussing this with you previously. Smallman12q: The metadata that would allow someone to cite the article itself. For example, for the article Pink Floyd (viewed today and last edited yesterday), would emit:

  • item_type = web page
  • website_title = Wikipedia
  • language = en
  • title = Pink Floyd
  • accessdate = 2010-08-02
  • date = 2010-08-01
  • url =
  • rights = CC-BY-SA 3.0 License and GFDL

Though we might include a permalink for the specific version, and a time as well as date. This would allow people who cite the page to collect the citation information into a tool like Zotero and then output it into their document or web page in one of a number of citation styles. Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 17:51, 2 August 2010 (UTC)

A template on every page would do the job, but it would be more efficient and reliable to add it to the MediaWiki software. You need to file a Bugzilla enhancement. ---— Gadget850 (Ed) talk 18:04, 2 August 2010 (UTC)
(edit conflict) It seems like we'd be asking for trouble by encouraging people to cite the current revision instead of a specific version; therefore, is it technically feasible to output the present oldid in the metadata every time the page is rendered? (And of course, the corresponding oldid when an old version is rendered.)
What's the meaning of date in this context, and how is it different from accessdate? (Date of article creation? If so, your example is incorrect.) Now I get it: date of last revision vs. date of rendering. TheFeds 18:23, 2 August 2010 (UTC)
Could you give us a comparison of the workflow someone would use now to cite an article, versus using a tool like Zotero to interpret the metadata? Do you see this as being a substantial improvement over the status quo?
And what are your thoughts about embedding the metadata on the Special:Cite page instead (that is generated by the "Cite this page" link in article space)? TheFeds 18:19, 2 August 2010 (UTC)
I believe OrangeDog was referring to at least this discussion. —Ost (talk) 18:30, 2 August 2010 (UTC)
Such functionality should be implemented into the Special:Cite page, and as a matter of fact, already has been since June 21, 2007, something actually stated in the COinS websiteTheDJ (talkcontribs) 19:32, 2 August 2010 (UTC)
That's a sub-optimal place to put it. Furthermore, the fact that the span, which contains only a space, is styled as display:none means that it is not found by tools such as the OpenURL Referrer extension for Firefox, thereby defeating the purpose of including it. Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 23:08, 2 August 2010 (UTC)
Fixes can be made MediaWiki:Cite_text, if suggested. —TheDJ (talkcontribs) 03:06, 3 August 2010 (UTC)
The span with the style display:none is located within the COinS span. Why would that have any effect at all? What's supposed to be inside the span? Looks like nothing is supposed to go there, so shouldn't that extension just act as if that's the case instead of breaking? Reach Out to the Truth 03:51, 3 August 2010 (UTC)
It's not the extension that's breaking. The extension inserts a user-defined sting into the span and makes it a link to a user-defined library service. Wikipedia is preventing that link from displaying. Why are we making an "empty" (white-space) span not display? Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 10:23, 3 August 2010 (UTC)
If so, that would be very silly, since that discussion is about microformats, not COinS, and even includes comments from me, pointing out that they are not the same thing. Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 23:08, 2 August 2010 (UTC)
Yes people, please learn to say "semantic markup for metadata". Not that anyone but Andy cares :D —TheDJ (talkcontribs) 03:06, 3 August 2010 (UTC)
You couldn't be more wrong: Wikipedia to Add Meaning to Its Pages. Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 10:23, 3 August 2010 (UTC)
Cares about the name, not the feature of course. —TheDJ (talkcontribs) 13:18, 3 August 2010 (UTC)
Oh well; in that case... you still couldn't be more wrong. Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 15:08, 3 August 2010 (UTC)

Problem deleting page

Having closed an MFD for User:Cats & dogs forever/Sandbox/Secret page as delete I am trying to delete it, but (for several hours now) whenever I try I get a message saying "Our servers are currently experiencing a technical problem" with an error message like:

Request: GET, from via (squid/2.7.STABLE6) to ()

Error: ERR_ACCESS_DENIED, errno [No Error] at Mon, 02 Aug 2010 20:45:12 GMT

Windows XP SP3, Firefox 3.5.11

JohnCD (talk) 20:54, 2 August 2010 (UTC)

I've deleted it without issue. I don't know why you had the problem you did. -- Finlay McWalterTalk 21:12, 2 August 2010 (UTC)
Thanks. "errno [No error]" is one for the line book. JohnCD (talk) 21:32, 2 August 2010 (UTC)
That error usually comes from bad deletion links in templates that escape ampersands incorrectly into &amp; or %26amp%3B which are blocked on Wikimedia servers (eg;). If you encounter those, just go to the page manually and click delete. If /that/ doesn't work, try copying the link and converting %26amp%3B directly to %26. --Splarka (rant) 08:09, 3 August 2010 (UTC)
Ah, I see, no doubt it was the ampersand in this title that was causing the problem. Thanks. JohnCD (talk) 09:46, 3 August 2010 (UTC)

Suggestion for an alternative to rangeblocks

Currently, large rangeblocks are sometimes used against persistent, banned vandals that rapidly change IP addresses. Some of these long term abusers have caused hundreds of thousands of IP addresses to be blocked for extended periods of time, potentially preventing many new users from ever contributing. Is there any possibility of developing a new system for IP and unconfirmed users analogous to pending changes? In this system, new IP's and unconfirmed users from an IP range used by a dynamic IP vandal would be classified as "provisional users," whose edits would require approval by a reviewer before they were visible to the general community. Users who had their edits marked as vandalism by a the reviewers could be blocked by an automated process to prevent increased workload on administrators. This way, innocent new contributors would be less impacted by the actions of a few vandals. VQuakr (talk) 03:33, 3 August 2010 (UTC)

Disruptive IP users already can be stopped by Abuse Filter, for registered users your proposal would reveal their location. — AlexSm 14:37, 3 August 2010 (UTC)
I like the idea of making editors or ranges subject to pending changes rather than only articles. "For registered users your proposal would reveal their location." I don't think there's a list of the locations of all current range blocks, is there? I don't see how falling into this system would reveal where someone is editing from. Fences&Windows 03:42, 4 August 2010 (UTC)

Relative and absolute in templates

How do I keep the items which are absolute in my template, in the template? In other words, when I use absolute in a template, then use that template on another page, how do I keep the absolute elements in the template.

Relative is much harder to use, as each new item added effects those items around it.

Thanks in advance! I hope this question is clear...

Adamtheclown (talk) 06:14, 3 August 2010 (UTC)

Your question is not very clear, but e.g. if you do no not want the result to depend on the page, avoid using {{PAGENAME}} etc.--Patrick (talk) 23:12, 3 August 2010 (UTC)

Referenced desk was archived--or WAS it?

I looked through my contributions and clicked on this topic, and replied to it.

But in my contributions, this is the address where the reply went, even though the above topic is still on the reference desk. What's going on?Vchimpanzee · talk · contributions · 20:34, 3 August 2010 (UTC)

Wikipedia:Reference desk/Computing transcludes some of its recent archives, so everything happened as it should. (Even though I have no idea why do they do it like this.) Svick (talk) 21:24, 3 August 2010 (UTC)

769 bad links to the US State Department

I see that 769 dead links to the US State Department got archived; if anyone is ever interested enough to fire up a project to fix these, please drop a note on my talk page as I don't normally have the pump on my watchlist. I left a note at WT:WikiProject United States Government as well. --CliffC (talk) 21:31, 3 August 2010 (UTC)

Trouble with staying logged in

I logged in yesterday and clicked on the link to my watchlist. But instead of taking me to my watchlist a screen came up that said "you must be logged in to view your watchlist," even though I had logged in. So I clicked the link to log in, and suddenly, without typing in anything, I was logged in again. This happened to me twice yesterday. What's going on? The Raptor Let's talk/My mistakes; I mean, er, contributions 23:28, 20 July 2010 (UTC)

This has happened to me, too, especially last July (which is exactly 1 year ago!), when I was most active. To anyone reading this, please see the related HD thread (WP:HD#Problem with staying logged in). Kayau Voting IS evil 04:54, 21 July 2010 (UTC)
They told me to bring it here, so I did. Anybody else have anything to say about this? The Raptor Let's talk/My mistakes; I mean, er, contributions 13:40, 21 July 2010 (UTC)
It just did it again. The Raptor Let's talk/My mistakes; I mean, er, contributions 03:52, 23 July 2010 (UTC)
Yeah, I've been having problems staying logged in as well. I keep checking the "stay logged in for 30 days", but I have to log in everytime I start a new session, even though it the same computer. Its getting very frustrating, since it's been going on for about 4 or 5 days now. Wizard191 (talk) 13:23, 26 July 2010 (UTC)
Note: it's now getting more serious again, in my area. It should be a universal problem given the variety of places where it's happened. Kayau Voting IS evil 03:53, 28 July 2010 (UTC)
Yeah, it's still happening to me too. Wizard191 (talk) 12:23, 28 July 2010 (UTC)
Ich auch. Skomorokh 20:44, 29 July 2010 (UTC)
It's getting really silly now. Happens at Wikibooks too. Kayau Voting IS evil 01:56, 1 August 2010 (UTC)
Is there anywhere else or anyone else we can report this problem to? The Raptor You rang?/My mistakes; I mean, er, contributions 16:20, 1 August 2010 (UTC)
None, except bugzilla, that I know of. Kayau Voting IS evil 11:44, 2 August 2010 (UTC)
It's working properly for me today. Wizard191 (talk) 22:32, 2 August 2010 (UTC)

──────────────────────────────────────────────────────────────────────────────────────────────────── It's been working fine for me, too. Wonder if it's been fixed? The Raptor You rang?/My mistakes; I mean, er, contributions 14:36, 3 August 2010 (UTC)

Since I first experienced that, it's been happening rather inconsistently... Kayau Voting IS evil 04:55, 4 August 2010 (UTC)
Just happened again. Are you sure there's nothing else we can do about it? The Raptor You rang?/My mistakes; I mean, er, contributions 22:02, 4 August 2010 (UTC)

edit conflicts...

  1. Joe Bloggs opens the edit window for section A
  2. I open the edit window for section B
  3. Joe Bloggs saves
  4. I save and get an edit conflict message
  5. I desperately try to find the section I was editing, copy it from the lower box to the upper box, and save
  6. Joe Bloggs's edit is gone!

after the first time this appeared I didn't save when I got an edit conflict like this one. I copied the thing I wrote from the diff and re-added it to the section instead. Kayau Voting IS evil 08:58, 3 August 2010 (UTC)

You're not supposed to do 5, you're supposed to merge the two edits, leaving both in what you save, if appropriate. OrangeDog (τε) 11:44, 3 August 2010 (UTC)
And the problem is, it doesn't appear in the diff so I cannot do that! Kayau Voting IS evil 04:54, 4 August 2010 (UTC)
Joe's edits are in the upper edit box. Do not desperately paste over them. OrangeDog (τε) 17:59, 4 August 2010 (UTC)

"Off by one" error in footnote linking


I've started seeing this weird bug occasionally: two entries in the References section have the same number. In the screenshot at right, the 14 refs are numbered 1–13 with two different #8 listed. After clicking various intra-page links for a while, the displayed eventually heals back to a proper 1–14 consecutive list. The numbering in the article itself and the crosslinks to/from the footnotes are correct (i.e., even when the bug is visible, the [number] inline marks are properly numbered and go to the "correct" (matched, and what would be that number if the reflist were consecutive) item. For example, in the screenshot, there is a #14 in the Godwin's law section, that links to ref #13. DMacks (talk) 18:11, 3 August 2010 (UTC)

Those numbers are generated by your browser, so it's probably its bug. Svick (talk) 21:20, 3 August 2010 (UTC)
The numbers are generated by the Cite extension. Does the problem resolve if you refresh? ---— Gadget850 (Ed) talk 23:25, 3 August 2010 (UTC)
But <references /> is rendered as a <ol>, so it it is up to the browser to actually number them. OTOH, the numbers from <ref> are generated by the extension directly. Svick (talk) 01:04, 4 August 2010 (UTC)
That's right— I was thinking of the id number. ---— Gadget850 (Ed) talk 02:50, 4 August 2010 (UTC)
Alright, it's definitely a browser issue, triggered by something in WP skins/css/etc. User:DMacks/multicolumn-test is my test-case (by editing down source-dump of the original demo case), just a simple numbered list with no ref tags or other templating has same effect in a local .html file. Completely reproducible on some browsers, doesn't happen at all on others. If I tried removing any one more of the scripting or other features and the problem goes away in my test-case. Definitely weird! DMacks (talk) 20:01, 4 August 2010 (UTC)

Table: sortable and autocollapse don't play nice?

If this is a FAQ please excuse me (and please provide a reference) but none of my search efforts have turned up an answer.

From the looks of it, the table classes sortable and autocollapse don't seem to work together. I can make a table sortable, I can make a table collapsible, I can make a table both sortable and collapsable, but no matter what I try I can't make a table sortable and autocollapsible. Is this a bug, a deliberate feature or am I doing something wrong? In case it matters (I deem it unlikely to be a browser/OS issue, but you never know): I'm using Opera 10.10 for Linux through Linux emulation onFreeBSD. Skysmurf (talk) 02:59, 4 August 2010 (UTC)

The help page is Help:Collapsing, in this diff you're confusing collapsed and autocollapse, the latter only works if there are 2 or more collapsible tables on the page. — AlexSm 03:49, 4 August 2010 (UTC)
Okay, the penny has dropped. Thanks for helping! Skysmurf (talk) 12:53, 4 August 2010 (UTC)

Constant glitches while editing

Almost every time I try to edit a page, glitches in the text show up after I'm finished. These glitches randomly change bits of text throughout the article, such as "Metastasis" becoming "Metast9 GM" or "Metas<s��s" and "security cameras" becoming "securitawaymeras". (Here's an example: [14].) It usually happens in areas of the article no where near the place I was actually editing. It never happens in the preview of the edit, only after I save the page. Using the sandbox first really doesn't help because even if I can correct the errors there, more will just appear when I go to the article itself. I have no idea why this is happening, and I even tried using a different web browser. (I was using Google Chrome before, and am now using Firefox, and nothing has changed.) My operating system is Windows Vista. Tesseract-7 (talk) 06:25, 4 August 2010 (UTC)

These are control characters. An often seen cause for them is trying to use the accesskeys, or other keyboard sequences. If you make an error with those, or if you use them in a context where they don't work, it is easy to accidentally enter a character, that although not visible in editing, will show up after saving. —TheDJ (talkcontribs) 11:53, 4 August 2010 (UTC)

Text in editor

Is there a way to enlarge the text which appears in the editor while editing? Maybe adding something in my own CCS page? What about choosing the font that appears in the editor while editing? --Mahmudmasri (talk) 16:53, 4 August 2010 (UTC)

In firefox you can enlarge the text by hitting Ctrl + +. I'm sure every browser has something comparable to that. Killiondude (talk) 17:09, 4 August 2010 (UTC)
Thanks, but are there other ways? --Mahmudmasri (talk) 17:27, 4 August 2010 (UTC)
textarea { font-size: 150% }
(or whatever value suits you) to your personal CSS.—Emil J. 18:26, 4 August 2010 (UTC)

How to convert references to a non-templated state?

Example: International recognition of Kosovo is at the edge of max number of templates/article. Most of them are from references. A naive solution is to convert them. How? Materialscientist (talk) 10:05, 4 August 2010 (UTC)

Looking at the HTML source:
NewPP limit report
Preprocessor node count: 258155/1000000
Post-expand include size: 1887812/2048000 bytes
Template argument size: 672021/2048000 bytes
Expensive parser function count: 6/500
So, yes, a few more templates will take you over the post-expand limit. There are 473 cite templates, 198 flag icons which in turn call {{flag}} and {{country flag}} which then calls a country data template, 151 uses of {{smallsup}} and 70 uses of {{dts}}. ---— Gadget850 (Ed) talk 11:15, 4 August 2010 (UTC)

Right. This problem is intrinsic to many large WP articles. Many of them are also relatively important, but are really difficult to edit (some give wikimedia crash windows). That is why posting here. Country flags and references eat up most templates. Stripping flags would make some articles ugly (changing to non-templated flags?). Non-templated refs don't harm anything, but there are hundreds of them, that requires some automated solution. Materialscientist (talk) 12:34, 4 August 2010 (UTC)

I'm a bit busy at the moment, but later I could look at converting the citations to text with an UltraEdit macro. Might be simpler to convert the flags to images. —Preceding unsigned comment added by Gadget850 (talkcontribs) 12:52, 5 August 2010 (UTC)
Yes, but is there a way to do that without cluttering much the article code? Materialscientist (talk) 13:02, 4 August 2010 (UTC)
The easiest (and possibly best) solution would be to split the page into subarticles. Its currently the 30th longest article by raw byte count. The HTML alone is almost 1 MB; when you include all the images, its nearly 1.5 MB. Mr.Z-man 22:19, 4 August 2010 (UTC)
Now that I have had the time to really look at it, I agree with Mr.Z-man on splitting the article. Other options would be to remove the flags entirely, expand some of the templates (use Special:ExpandTemplates to test) or leave it as is and hope no one runs over the limit. ---— Gadget850 (Ed) talk 22:58, 4 August 2010 (UTC)
Well the flags don't seem to add that much encyclopedic content (and i just plain don't like them used in that sense), I'm thinking just try the article without them and see where it goes from there. Peachey88 (T · C) 04:10, 5 August 2010 (UTC)
From the techie point, flag templates do their job, namely allow adding an image and a linked country with a simple 3-letter code, reducing the clutter. Reference templates do it other way around - add code clutter and a template w/o improving the output (I know their advantages, that is not the point here). Thus it is logical to leave flags and strip ref templates, and it is doable (even with simple Word macro substitutions, but this is tiresome). I am just not versed enough in wiki-compatible programming to do that. Splitting is obvious. BTW, even auto replacement or deletion of flags is not that trivial because it requires knowledge of the flag abbreviations, and the linked above Special:ExpandTemplates adds much clutter of its own. Materialscientist (talk) 04:21, 5 August 2010 (UTC)
In the case of International recognition of Kosovo (I am one of its regular contributors), I think splitting the article will be difficult without seriously compromising its quality. There are bits that could be moved elsewhere, but these won't make much of a difference. And I feel that in this article the flags are vital in terms of its look and feel. I would be unhappy with the removal of the cite templates, but I understand that as a last resort this may be a necessary evil. However, I think that there is definitely scope for tidying up the text and removing unnecessary references. This will have to be a manual job, and it won't be easy, but it has needed doing for some time. This technical problem could provide the impetus for a major cleanup. (Of course, one other solution is to persuade more governments to recognise Kosovo's independence - we could then move them from the Non-Recognisers table (several references each) to the Recognisers table (just one or two refs each)). Bazonka (talk) 17:13, 5 August 2010 (UTC)
Why can't each of the tables be split to a separate article and have a short prose text with the names of the countries in the main article instead? Most people checking the article most likely don't want all those details anyway and those who do can check the sub-article without problems. Regards SoWhy 17:43, 5 August 2010 (UTC)
I agree that some form of WP:FORKing is the ideal solution, probably of International recognition of Kosovo#States which do not formally recognise Kosovo as independent to its own List of states which do not formally recognise Kosovo as sovereign. I also don't agree that the flagicons are worth the real-estate on the page just to reduce clutter in the source, but I notice that they are used often in other articles. I question the repeated UN and EU icons for international membership column as I would expect the flags to take the place of the text of that the text would be sufficient, but it is hardly a concern if the bloated table with per country overviews of position were moved. Along with a {{main}} link, this section could be altered to give an overview of common or peculiar positions. —Ost (talk) 18:17, 5 August 2010 (UTC)
I think that Talk:International recognition of Kosovo would be a more appropriate place for this non-technical discussion. I am going to copy some of these comments to there. Bazonka (talk) 18:49, 5 August 2010 (UTC)

Citation toolbox widget

Created importScript("User:Smith609/toolbox.js") some days ago on a modern.js It did not create a widget on my toolbox. I also tried it with single quotation marks importScript('User:Smith609/toolbox.js') and it still did nothing. Left a talk page message for Smith609 several days ago, but have had no response. I have Firefox 3.6.8 and my Java plug-ins are enabled. It is the same issue if I use my IE 8.0.6001.18702 Can anybody offer some insight? Maile66 (talk) 19:46, 4 August 2010 (UTC)

It adds the tools BELOW the toolbox, just so you know. I also fixed a couple of escaping issues that the script had. Other than that, it seems to work for me with both the modern and monobook skins (not the default vector). Also note that JAVA is not the same as JavaScript. Your modern.js file is also in order. The only thing I can think of is that you have a Gadget enabled in your preferences that is not compatible with the Modern skin, such things can cause other scripts to fail. —TheDJ (talkcontribs) 20:41, 4 August 2010 (UTC)
Thanks for the try, anyway. My JavaScript was already enabled, also. The only thing below my Toolbox are the three standard Print/Report options: Create a book, PDF, Print. I unchecked all my gadgets and purged the browser cache. Even shut down the browser and re-opened. It's not the gadgets. A puzzlement, indeed.
--Maile66 (talk) 21:14, 4 August 2010 (UTC)
Are you sure you are using the modern skin ? The following link should always take you to the appropriate script file for your skin: Special:MyPage/skin.jsTheDJ (talkcontribs) 21:18, 4 August 2010 (UTC)
Yep, I'm in Modern, and the link takes me to the skin. I'll tell you what I've been attempting, if it makes any difference. I only want to do one thing: add ISBNs automatically And while I can do it by saving this link somewhere, I wanted a widget. That doesn't seem to be happening.
--Maile66 (talk) 21:43, 4 August 2010 (UTC)
Your firefox has an error console option in one of the menus. It should report Javascript errors, if any are occurring. Perhaps that information can be of help in pinpointing the problem, otherwise I really don't know what else to suggest. Someone else have ideas ? —TheDJ (talkcontribs) 23:18, 4 August 2010 (UTC)
One thing, when I'm in the Edit screen, below the Toolbox, there is a section for "scripts", but the "Custom regex" is grayed out. There are some error messages that may be of use. I am going to email you one of those messages, as I am not a programmer and do not know what they indicate. Maile66 (talk) 23:32, 4 August 2010 (UTC)

Got an error message that might mean something, when I click on the grayed out Custom regex: "Custom is not defined Javascript:custom()" --Maile66 (talk) 12:35, 5 August 2010 (UTC)

More of the Error Console's error code message. I do not use Google Chrome, if this is what it's referring to. Mean anything to anybody?: Error: uncaught exception: [Exception... "Component returned failure code: 0x80004005 (NS_ERROR_FAILURE) [nsIWebProgress.removeProgressListener]" nsresult: "0x80004005 (NS_ERROR_FAILURE)" location: "JS frame :: chrome://global/content/bindings/browser.xml :: removeProgressListener :: line 394" data: no] Error: uncaught exception: [Exception... "Component returned failure code: 0x80070057 (NS_ERROR_ILLEGAL_VALUE) [nsIWebNavigation.loadURI]" nsresult: "0x80070057 (NS_ERROR_ILLEGAL_VALUE)" location: "JS frame :: chrome://global/content/viewSource.js :: viewSource :: line 221" data: no] --Maile66 (talk) 20:26, 5 August 2010 (UTC)

Page move oddity

Green_Bay-Titirangi_United was recently moved from Green Bay/Titirangi. Now while the talk page it is at and displays "Talk:Green Bay-Titirangi United" as the heading, gives a heading of "Green Bay/Titirangi". (no sign of redirect message and address bar still shows Green Bay-Titirangi United). diff at shows this to be the last diff, as does oldID: . History shows another edit which has an even later diff which when you go t onext edit seems to go back to the previous "move" edit. I've cleared my cache, restarted firefox, screwed up my lip and held my head at a 37.5degree tilt and it still doesn't want to behave itself. Is it just me or is someone else seeing this?--ClubOranjeT 09:18, 5 August 2010 (UTC)

Ok, I managed to "fix" it by restoring to . The history still seems to have a bit of an issue pretty much as described and the "restore" does not register in the history or my own "contributions", but at least the article looks right now.--ClubOranjeT 09:34, 5 August 2010 (UTC)
Very peculiar. I don't get the header display problems (I made the move, but I'm using Safari) - but the later diff shown is one that - although it says i made it - I didn't make (it isn't listed in my contributions, either)! I changed the infobox and text from a slash to a hyphen, then moved the page - why it recorded that I then reverted the text changes, I don't know. I've no idea what goes on here, though I suspect it might have something to do with the problematic former page name (the "/" may have caused some glitch, perhaps?) Grutness...wha? 09:40, 5 August 2010 (UTC)

Pending changes not automatically accepted?

Has anyone seen this [15]? I had to accept my own revisions (see page history)? Thanks! Plastikspork ―Œ(talk) 15:13, 5 August 2010 (UTC)

I had to accept my own edit two days ago, I forget on what article. I think an edit conflict might have been involved. Fences&Windows 00:01, 6 August 2010 (UTC)
The previous edit was never accepted, and you didn't revert in a way that would be automatically approved. The software doesn't know for sure you actually reviewed the previous revision, so you have to approve it manually. Reach Out to the Truth 00:21, 6 August 2010 (UTC)

Article section heading - incoming link notes

Sometimes, a section heading will have a note, explaining that "article foo links to here" or "redirect foo links to here", so that editors don't accidentally break incoming links when they change the wording of the heading.

e.g. something like: == History <!-- the redirect [History of foo] links to this section --> ==

Is this practice described or recommended anywhere? I can't find examples, but I'm sure I've seen it in action, somewhere... Thanks. -- Quiddity (talk) 20:19, 5 August 2010 (UTC)

Wikipedia:Redirect#Targeted and untargeted redirects. — AlexSm 20:46, 5 August 2010 (UTC)
Perfect. Thank you. -- Quiddity (talk) 21:24, 5 August 2010 (UTC)

How do we best achieve the reverse of this, for {main} links? - We desire a system that is able to list which articles are using {{Main}} to link to the target-article.

e.g. Alphabet#History includes the template {{Main|History of the alphabet}}, but we wish to know how many other articles also include it.

Another editor has proposed a slightly cumbersome method at VPR, and hence I'm wondering if there is a more elegant solution. (Note: manually searching for the exact text doesn't seem to work at all [16], however an automated method is what we're requesting...). Thanks. -- Quiddity (talk) 21:24, 5 August 2010 (UTC)