Wikipedia:Village pump (technical)/Archive 55#This week's software updates

Is it possible to use a bot to respell all the links which are currently towards Kveta Peschke, so that they link direct to Květa Peschke. Granted, there is already a redirect here, but the correct Czech spelling is Květa Peschke with a ě, and I think it would be better whether in the articles you can read the "ě". --Voletyvole (talk) 10:40, 25 January 2009 (UTC)

• Also, there are thousands of similar cases like this one with the diaretics missing from the name - maybe I can use a bot which respells many things. How do I start up this bot? --Voletyvole (talk) 10:43, 25 January 2009 (UTC)

::You need to go to Wikipedia:Bot requests. — Blue-Haired Lawyer 11:03, 25 January 2009 (UTC)

The problem with making spell-correctors depend on whatlinkshere is that too many of the lazy (unaccentuated) spellings will be completely ignored by the bot because they are not part of a link. — CharlotteWebb 13:40, 25 January 2009 (UTC)

As a frequent user I'm frustrated with having to type in the main subject everytime I would like to refer to the drop-down listings. Is there a method I could use to get back to the drop-down list witout re-entering the subject?

Thanks, Wally in Iowa173.20.5.142 (talk) 13:45, 25 January 2009 (UTC)

Are you talking about the list of previous searches you've done in the search box. Usually, it shows up after typing just a few letters after which I can use the arrow keys to select one. - Mgm|(talk) 14:34, 25 January 2009 (UTC)

Special:PrefixIndex may be of interest to you. PrimeHunter (talk) 14:52, 25 January 2009 (UTC)

----

No, I type in a search, get the drop down, select and get an article. Then I sometimes would like to go back to the drop down for another article on the subject but have to type in the original search word to get the drop down back. —Preceding unsigned comment added by Wallace Fiala (talk • contribs) 17:30, 25 January 2009 (UTC)

:Well if you turn off the automatic result thingy, it will default to your borwser's stored list, in which case you can simply double-click and select what you want. ♫ Melodia Chaconne ♫ (talk) 18:08, 25 January 2009 (UTC)

I recently added a feature [http://en.wikipedia.org/w/index.php?title=User%3AVoletyvole%2Fmonobook.js&diff=266393899&oldid=262562143 here], for correcting typos. Isn't this something which I should be better off avoiding, in order to do something else, as a non-native (room for error - natives spell far more sexily than I)? Also, I'm probably doing this wrong, and my monobook skin is pretty damn new for me and i don't think this monobook link was so perfect. I hope you understand, but I might be wasting my time in this respect to monobook. Maybe article-writing is better...this website is a crazy mind-boggler to understand!!--Voletyvole (talk) 21:57, 25 January 2009 (UTC)

How do you center class="wikitable sortable" ? Thank you Ikip (talk) 02:01, 26 January 2009 (UTC)

:Yes {{thank you}} !!! Ikip (talk) 03:00, 26 January 2009 (UTC)

When a category gets moved following a CfD, a bot leaves a message in the new category that lists the authors of the old category. For instance:

"Robot: Moved from Category:Space trading and combat simulation games. Authors: Painbearer, Marasmusine, Cydebot, Eep², SharkD"

My question is, in what order are the names listed? Are they listed in the order that edits were made? Who is the "first" editor, chronologically? SharkD (talk) 10:15, 22 January 2009 (UTC)

:Depends on how the bot is written; not all of them list the authors. In this case, I'd ask User:Cyde, as it was his bot that moved the space trading category. --Kbdank71 16:25, 22 January 2009 (UTC)

::Ah, I see. Thanks. SharkD (talk) 03:15, 23 January 2009 (UTC)

Cough, cough. It's about time we had some built-in way to move categories. Ideally the redirects would work as well as template redirects. That is, articles using the old category name would immediately appear to be using the new category name in every respect except for the wikitext of the article, which need not be immediately updated. Same thing for images would be nice… — CharlotteWebb 03:19, 25 January 2009 (UTC)

:If moving a category preserved its edit history I would be satisfied. SharkD (talk) 06:03, 27 January 2009 (UTC)

I don't know if anyone has noticed this, but if you type some text into the search field at the top of this article and simply press 'Enter' (i.e. without actually clicking the button), then you are automatically taken to the article with the same name matching the text you entered instead of to the search results. (Internet Explorer 7) SharkD (talk) 19:53, 23 January 2009 (UTC)

: Yes, because the default action is "Go", not "Search". - Jmabel | Talk 19:56, 23 January 2009 (UTC)

::Sorry for the confusion, but I meant the "Search the Village Pumps" box. SharkD (talk) 21:04, 23 January 2009 (UTC)

:::This was previously reported at the helpdesk. I don't know if anything can be done about it other than bemoan the brokenness of IE. Algebraist 00:16, 24 January 2009 (UTC)

::::Filing this in our system as bugzilla:17161 --brion (talk) 17:44, 26 January 2009 (UTC)

{{Listen|filename=NonFreeImageRemoved.svg|title=The Song At The Right|format=Ogg}}

For: Cleaning up the odd numbering and spacing at Arrondissement of Montbard

Reward: You can listen to the song at the right.

DEAD OR ALIVE. Sheriff Eddy (Howdy Partners!) 02:04, 25 January 2009 (UTC)

:I'll bite. What specifically do you need done? -- Tcncv (talk) 04:46, 25 January 2009 (UTC)

:Yeah, I don't think WP:NFCC permits "displaying" the audio sample on the village pump. — CharlotteWebb 05:13, 25 January 2009 (UTC)

::Whatever. If you check out the table at the arrondissement table, you'll see that most of it is in three columns but some is in 4 and other in two. The numbering is also messed up because a new commune was added with the merger of two small ones. ~EDDY ^{(talk/contribs/editor review)}~ 01:40, 26 January 2009 (UTC)

:::{{diff2|266448917|I rest my case}}. As for the article I don't know any universally accepted way to split it into three columns without manually numbering the communes, which is pretty lame if the list is subject to change. Also anyone copying and pasting the list would probably not want to include the numbers at the front. I suspect the numbers are not necessary or meaningful as it is an alphabetical list. Probably best to split it into three hard columns of 84 each and use asterisk-bullet points. — CharlotteWebb 16:37, 26 January 2009 (UTC)

{{Resolved|1=ukexpat (talk) 02:10, 26 January 2009 (UTC)}}

I've created a table that needs to have the column headers rotated 90 degrees. The cross-browser way of doing this that I've observed being used on Wikipedia is to use SVG images in place of the cell text. There's one problem, however: normally, clicking on an image will take visitors to the page for the image. How do I change it so that visitors will be taken to a different (i.e. arbitrary) page? Thanks. SharkD (talk) 14:08, 25 January 2009 (UTC)

:By using the ImageMap extension. Graham87 15:56, 25 January 2009 (UTC)

::Thank you. The problem now is I can't remember which article(s) this technique was used in. Does anyone recall coming across any tables that have vertically-oriented text (but in reality SVG files) in the top row? I want to look more closely at how the images themselves were sized/scaled. SharkD (talk) 16:59, 25 January 2009 (UTC)

:::Nevermind. I found the article. SharkD (talk) 17:16, 25 January 2009 (UTC)

For future reference, I added a section on the [http://en.wikipedia.org/wiki/Help:Table#Vertically-oriented_text tables help page] that describes how this is achieved. SharkD (talk) 01:47, 27 January 2009 (UTC)

Where can I find the fonts used by the SVG rasterizer? When I create images on my local system InkScape uses my local fonts, which are different than those used by Wikipedia. Thanks. SharkD (talk) 19:26, 25 January 2009 (UTC)

:See m:SVG fonts. --dapete 21:01, 25 January 2009 (UTC)

::OK, thanks. SharkD (talk) 23:31, 26 January 2009 (UTC)

I recently converted an MP4 video to the OGM format using SUPER, but Wikipedia/Wikimedia Commons only accepts OGG and OGV files. (And SUPER can't convert to OGG or OGV video.) What's the difference between .ogm and .ogv? Is there any reason Wikipedia doesn't take OGM (which is as Free as OGG or OGV)? -- King of ♥ ♦ ♣ ♠ 19:59, 25 January 2009 (UTC)

:try useing the transcoder built into VLC media player Geni 01:30, 26 January 2009 (UTC)

::It's all one and the same format, just different file extensions. ogg is the original format for audio vorbis. ogm includes some extensions for dealing with video/subtitles and metadata in this fileformat. It is the same fileformat, but you might say that it is a "newer" version of the specification. .ogm is not an official file-extension, but was used widely as a file extension to differentiate these files from pure ogg vorbis audio files. ogv and oga are the "new" file extensions that are now suggested to differentiate between pure audio (.oga .spx) and audio/video/meta (.ogv) files. But in essence, it's all one and the same fileformat that can contain "different" types of meda. See also [http://www.xiph.org/ogg/doc/rfc5334.txt] --TheDJ (talk • contribs) 11:19, 26 January 2009 (UTC)

:Note also that .ogm files frequently contain non-free codecs. Make sure you're actually using Theora as video codec and Vorbis as audio codec, or the inline player won't work. --brion (talk) 17:38, 26 January 2009 (UTC)

Is there templates like:

:{{?}}

:Code: {{?}}

...for each letter of the alphabet? thanks Ikip (talk) 02:03, 26 January 2009 (UTC)

:So far I have found:

::{{tld|A-classicon}}

::{{tld|B-classicon}}

::{{tld|C-classicon}}

::20px

: Ikip (talk) 02:07, 26 January 2009 (UTC)

::The first three are used in article classification templates, judging from the filenames. – ukexpat (talk) 02:12, 26 January 2009 (UTC)

:::I found these photos on wikicommons, I simply need to make templates:

:::

:::Thanks. Ikip (talk) 02:14, 26 January 2009 (UTC)

::::I created the templates: :Template:Gross_A Ikip (talk) 03:00, 26 January 2009 (UTC)

:::::See also commons:Category:Latin letters. PrimeHunter (talk) 12:27, 26 January 2009 (UTC)

I am creating a template: User:Ikip/n which has my {{tl|ARS-userpage}} template embedded in it. I am attempting to get all of the little pictures to line up, like User:Piotrus/Top or User:Peteforsyth's awards here:

Right now the photos are vertical, {{User:Ikip/n|Oregon|Moose|Canada}}.

How do I fix the coding on {{User:Ikip/n}}?

See also is able to allow pictures horizontally:

{{see also|20px|20px|20px}}

Welcome to play with the template. Ikip (talk) 04:58, 26 January 2009 (UTC)

:The primary problem is that this is using imagemap. imagemap has the nasty habbit of including it's results in a div. you are putting several of these imagemap div's next to eachother in a single div within {{User:Ikip/n}}. That is causing the problem. The implementation differs from User:Piotrus/Top etc in that these uses all have an individual "inline" div around each single symbol (imagemap div), whereas you are wrapping an "inline" div around all your symbols at once. --TheDJ (talk • contribs) 11:37, 26 January 2009 (UTC)

::{{Thank you}} I will look into this, although I am not quite sure how to fix in now based on what you said, but it points me in the right direction. thanks for trying to fix it yourself. :) Ikip (talk) 14:39, 26 January 2009 (UTC)

:::Actually, it isn't resolved, I removed the tag...Any more help would be appreciated. Ikip (talk) 19:04, 26 January 2009 (UTC)

::::I finished my changes (didn't have the time this afternoon). But remember... this is all VERY broken. And it will be more broken when flaggedrevs arrives, because it uses the same area on a page. When flaggedrevs arrives, we will all have to really think about what we want to do with all these icons and how to "properly" use them. It's also not working in other skins now atm of course. --TheDJ (talk • contribs) 20:40, 26 January 2009 (UTC)

I often use Google to look for reliable sources about a niche topic. And most of the time, Google returns an ocean of blogs and forums, and I have to check a dozen of result pages before finding a reliable website such as "Financial Times" or "ieee.org".

Is there a search engine that only searches within reliable websites ? "reliable" is not really biased here, some websites are clearly accepted as reliable sources, we can make a list of them, based on consensus. It is impossible to write the list of all reliable websites, but a hundred of reliable websites would be enough for general topics. The list of reliable websites could be refined by Wikipedians over time.

Does such a search engine exist ? If not, is anyone willing to create a quick mashup based on Google and filters ? Thanks for any hint/idea/cooperation :-) Nicolas1981 (talk) 11:25, 26 January 2009 (UTC)

: Google News an Google Scholar gives mostly reliable results, but certainly does not cover all reliable sites. --Apoc2400 (talk) 12:26, 26 January 2009 (UTC)

::I gave them a try, and found Google News to be useful in some cases, even though it only includes news websites and not other reference websites. Thanks ! Nicolas1981 (talk) 02:45, 27 January 2009 (UTC)

:Pubmed, JSTOR (if you are at a uni with access).Geni 17:54, 26 January 2009 (UTC)

::Pubmed is useful when working on biomedical research articles, I guess. I don't have access to JSTOR but I can imagine it gives access to some North American libraries' collections. The power of the search engine I described is that it would not look at a particular collection, but at a big array of reference websites, much more websites than editors can afford to check. This would greatly increase the odds of finding a match to your query. Thanks :-) Nicolas1981 (talk) 02:45, 27 January 2009 (UTC)

i just want to add aphoto to an article and there appears to be no way to do so

yes i went and made a new id and all that crap now i just want to [put it with the article —Preceding unsigned comment added by Koibeatu (talk • contribs) 15:07, 26 January 2009 (UTC)

:Did you upload the image as explained at WP:UPLOAD? – ukexpat (talk) 15:55, 26 January 2009 (UTC)

Can you explain to me what is wrong with the quotebox in Taking Tiger Mountain (By Strategy)? It seems to overlap with the album template. (I am using Firefox 3.0.4.) Is there any way around this? Regards, —Mattisse (Talk) 20:54, 26 January 2009 (UTC)

:I added a {{clear}} which seems to have fixed the problem, although it may be adding extra white space. —Mattisse (Talk) 21:16, 26 January 2009 (UTC)

::I tried to use the {{tl|quote box}} template code to left align the box at 50% page width, but I couldn't get it to work. – ukexpat (talk) 21:36, 26 January 2009 (UTC)

:::There was no line break between the end of {{tl|quote box}} and the next line. BTW: {{tl|quote box2}} has more features. --—— Gadget850 (Ed) ^talk - 21:48, 26 January 2009 (UTC)

::::Now the top line of the box slightly overlaps the last line of the text above (Firefox 2.0.0.20). – ukexpat (talk) 22:38, 26 January 2009 (UTC)

:::::Using seems to do it, at least in my browser. —Mattisse (Talk) 22:45, 26 January 2009 (UTC)

::::::Yup that works for me. – ukexpat (talk) 03:30, 27 January 2009 (UTC)

How technically feasible would it be to:

1. Make the "edit" tab larger and/or highlighted with different colors from the other tabs, to encourage editing?
2. Add a feature that acts like "random article", but within a specified category, or the realm of a specified portal or wikiproject?

Thanks,» šᾦῥъ^{τ • ¢} 19:45, 25 January 2009 (UTC)

:I don't know about #2, but #1 is very easy. Any admin can just go to MediaWiki:Edit and change how the 'edit' function is displayed. -- King of ♥ ♦ ♣ ♠ 20:01, 25 January 2009 (UTC)

::Actually, MediaWiki:Edit only controls the text of the edit tab and, as far as I know, accepts only plaintext input. You'd need to use MediaWiki:Monobook.css to change its colour or size. {{Nihiltres|talk|[http://en.wikipedia.org/wiki/Special:Log?user=Nihiltres log]}} 20:23, 25 January 2009 (UTC)

Nihiltres is correct about the interface page. Getting a random article within a category would be a fun toolserver gadget if somebody wants to create it. Just include links like http://toolserver.org/~somebody/randompage.php?fromcat=Living_people (or whatever you're on about) to the sidebar. This would give BLP patrolling a new dimension at least. If this works well and proves popular it can be added to the main software perhaps. — CharlotteWebb 16:51, 26 January 2009 (UTC)

:This tool http://tools.wikimedia.de/~erwin85/randomarticle.php will give you a random article in a given category. [http://tools.wikimedia.de/~erwin85/randomarticle.php?lang=en&family=wikipedia&subcats=1&d=1&categories=Living_people This link] gives you a random BLP. --Pixelface (talk) 23:44, 26 January 2009 (UTC)

::Ah, there is nothing new under the sun. — CharlotteWebb 03:22, 27 January 2009 (UTC)

:::Wikibooks does something slightly more hacky in site JS. It gives you a "random book" by searching (client side, via the API) generator=random via query-continue until it finds a page within one of the target categories (of which there are 38), but this works because about 10% of all NS_MAIN non-redirect pages are in one of those categories. There are a few categories here that would be feasable, such as :Category:Living people. --Splarka (rant) 08:15, 27 January 2009 (UTC)

I get a 503 when visiting the secure server, but the regular en.wikipedia.org works fine. Anyone know what's up with that? (Why do I feel I'm in the wrong section...) Elm-39 - T/C 13:47, 27 January 2009 (UTC)

:Working fine for me now. Algebraist 14:23, 27 January 2009 (UTC)

::Same here. Must've been an overload or something. Elm-39 - T/C 18:17, 27 January 2009 (UTC)

::EDIT: Whoa, database server lag of 1,986 seconds. That's... about half an hour. Wonder what's going on... Elm-39 - T/C 18:51, 27 January 2009 (UTC)

The Bobbylow's userpage have several main categories (like Security, Formal sciences, etc.). I think:

• Now we must remove the categories ([http://en.wikipedia.org/w/index.php?title=User%3ABobbylow&diff=266761846&oldid=255833043 done]).
• A bot must watch categories in order to avoid userpages to be included.
• A better solution: An extension or patch that automatically avoids it. I.e.: If the category page don't have some special tag (like <includeuserpages/>) don't must include it).

:I think that was supposed to be an article draft. Elm-39 - T/C 18:20, 27 January 2009 (UTC)

::I moved it to a subpage at User:Bobbylow/Cryptography. – ukexpat (talk) 19:18, 27 January 2009 (UTC)

{{resolved}}

I looked at my new message, but the orange bar keeps showing up randomly. - Peregrine Fisher (talk) (contribs) 19:06, 27 January 2009 (UTC)

:I have noticed that too -- I suspect it's a temporary server problem, I wouldn't worry about it. – ukexpat (talk) 19:19, 27 January 2009 (UTC)

:Contributions were lagged - probably related. --NE2 19:59, 27 January 2009 (UTC)

{{resolved}}

Several times today when I've created a new page, after I save it Wikipedia says the page does not exist. This usually clears up within a few minutes. I saw the same thing a few weeks ago. davidwr/(talk)/(contribs)/(e-mail) 20:30, 27 January 2009 (UTC)

:See the posting immediately above ↑ . – ukexpat (talk) 21:05, 27 January 2009 (UTC)

::Try reloading. Some browsers cache the old page. 199.125.109.64 (talk) 02:38, 28 January 2009 (UTC)

:::There was massive lag earlier. — neuro^(talk) 02:48, 28 January 2009 (UTC)

{{y}} - resolved

In the page Great Smoky Mountains National Park I saw that the conversion of 244,442 acres gives 990.44 km2, which is incorrect. The right figure is 99.044 km2. The wiki function used is {{convert|244742|acre|km2}}. The program gives a 10x error factor this time. I don't know if this is a single error or if it is systematic. Can somebody look about this matter? Thank you, --Gabodon (talk) 21:31, 27 January 2009 (UTC)

:1 acre = 0.00404685642 square kilometers. 244,442 acres = 989.222 square kilometers. Chalk the difference up to rounding or a minor error in the formula. By the way, 640 acres = 1 square mile, and about 2.56 square kilometers = 1 square mile. Doing the math, about 250 acres = 1 square kilometer, so 990.44 square kilometers looks about right. davidwr/(talk)/(contribs)/(e-mail) 22:01, 27 January 2009 (UTC)

::That's right, I apologize but in Italy we use the comma to separate decimals so I read the figure as 244.442 acres (two hundred fourty four point ... and not two hundred fourty four thousand ... ). It's all explained, thank you. --Gabodon (talk) 22:08, 27 January 2009 (UTC)

I'd like to suggest that links to Google machine translations be provided next to inter-language wiki links in the WP sidebar. SharkD (talk) 03:23, 21 January 2009 (UTC)

:Why are machine translations desirable? {{Nihiltres|talk|[http://en.wikipedia.org/wiki/Special:Log?user=Nihiltres log]}} 03:35, 21 January 2009 (UTC)

:Yeah, using computerized translation isn't preferable when you're talking about an encyclopedia like this. It's inaccurate to say the least, and Google translations are not as open as Wikipedia in terms of suggesting better translations manually.

:That said, I'm waiting for a reason from SharkD. HУтaяtalk2mecontribs 20:44, 21 January 2009 (UTC)

::This would probably be more useful to editors - you could compare an article with, say, its French version and see if there's any information in the French version that's missing in English. Maybe this feature would be best implemented as a user script or gadget. Tra (Talk) 23:06, 21 January 2009 (UTC)

::The reason is that I'd like a simple way to search on other language wikis when I suspect that reliable information might be found there. There are a lot of Japanese video game stubs that could use expanding. SharkD (talk) 08:03, 28 January 2009 (UTC)

:::I thought the Google toolbar had this kind of functionality? — Blue-Haired Lawyer 23:40, 21 January 2009 (UTC)

:::I have a javascript that adds a translation link to any wiki page where wgContentLanguage != wgUserLanguage (so of course you have to be logged in—go-go SUL…). — CharlotteWebb 04:02, 25 January 2009 (UTC)

::::Do you have to load the target page first before this link becomes available? SharkD (talk) 08:03, 28 January 2009 (UTC)

Hey I a question why is these two edit revisions showing up red.

I makes it really hard to see what changes have been made. [http://en.wikipedia.org/w/index.php?title=East-West_Schism&diff=265320768&oldid=265309988]

LoveMonkey (talk) 05:23, 26 January 2009 (UTC)

:For some reason, MediaWiki thinks that the text on the right was replaced with the text on the left, but in reality there was nothing but a small addition... Calvin 1998 ^(t·c) 05:54, 26 January 2009 (UTC)

::Probably because the paragraph was so long. It must have totally confused MediaWiki. This, that and the other _[talk] 06:42, 26 January 2009 (UTC)

:::Confusing. What can be done? No disrespect intended, I promise. Also, why is it that all of the edits by this one editor kept doing this? Now one elses? My edits to the same article are not showing up this way, as well as other editors to the article, why just this editor? Note also now it has stopped. [http://en.wikipedia.org/w/index.php?title=Catholic%E2%80%93Orthodox_theological_differences&diff=265907949&oldid=265905419] Who can I got to, get this fixed? Since in good faith I do not like accusing people of gaming the system unless I have edit evidence. It appears from the edits that the editor has control over this (issue) since they are the only getting the result from their edits and now they stopped doing this after I have complained.LoveMonkey (talk) 14:13, 26 January 2009 (UTC)

::::Your own edits of a long paragraph have exactly the same redding effect: see all of those you have just done at http://en.wikipedia.org/w/index.php?title=Catholic%E2%80%93Orthodox_theological_differences&action=history. Soidi (talk) 16:19, 26 January 2009 (UTC)

Maybe there is some kind of byte limit in the diff generator, after which it does not bother checking for similarities and differences in each "line". — CharlotteWebb 16:58, 26 January 2009 (UTC)

:[http://svn.wikimedia.org/viewvc/mediawiki/trunk/phase3/includes/diff/DifferenceEngine.php Yes]: const MAX_LINE_LENGTH = 10000;

:Apparently this limit was added by Tim Starling, 2005-08-31 which was a long time ago. We could ask him if this is still necessary. The paragraph you were editing is about 15,000. Probably better to split it up. If you don't want a paragraph break you can add line breaks inside the ref tags, or inside an html comment like this <!--

-->. This will be ignored when viewing the page but at least it will split the paragraph into multiple "lines" of wiki-text to make the diff view less burdensome. — CharlotteWebb 17:14, 26 January 2009 (UTC)

Why use a comment? A single line break will do just fine... — Werdna • talk 09:16, 28 January 2009 (UTC)

There is a bug in the MediaWiki software that falsely marks in-use files as orphaned. The fair-use files marked as such are thus speedily deleted. Who could fix this? (Jimbo Wales would be my first guess, but I am not sure.) A whole series of properly uploaded and in-use files was deleted speedily because the files were marked orphaned. So, if anyone could point me in the direction to whom or where I should go with this, that would be great. Thank you. -BlueCaper (talk) 02:59, 28 January 2009 (UTC)

:This is not a bug, the pages just need to be purged. — neuro^(talk) 03:06, 28 January 2009 (UTC)

Jimmy isn't at all involved in technical matters. — Werdna • talk 01:21, 29 January 2009 (UTC)

At List of female tennis players we are in the middle of a big re-formatting job and have decided to use sortable wikitables. We think that the headings for each table should be centred, but that the contents should be centred only for the No., Birth, Death, and Grand Slam columns. We don't know how to do this, or if it's even possible. Any ideas? Thanks in advance! Maedin\^talk 08:20, 27 January 2009 (UTC)

:You can precede any cell content with {{xt|style{{=}}"text-align: left"|}} or {{xt|style{{=}}"text-align: center"|}} (note the single pipe | symbol). For brevity you can also use {{xt|align{{=}}left|}} or {{xt|align{{=}}center|}}, but this may not work in some circumstances. −Woodstone (talk) 08:53, 27 January 2009 (UTC)

::Using an exclamation mark (!) instead of a pipe (|) defines a cell as a header and formats it appropriately (with class="wikitable", bold, centred and the cell is darkened slightly). See Help:Table for more info. mattbr 18:53, 27 January 2009 (UTC)

:::Correct. But see how he was asking about column formatting. By default, normal content of a table will be left aligned. In mediawiki it is not possible at the moment to apply CSS styling to all cells of a single column however. If you want the contents of one row centered, you will have to write this styling for each individual cell. This is because this requires the COL / COLGROUP HTML syntax (i think..). These HTML elements are however not supported by Mediawiki, and sparsely supported by browsers (Only IE and Opera i believe). --TheDJ (talk • contribs) 21:05, 27 January 2009 (UTC)

::::Hmm, I would think you could style it properly by creating a CSS class (.wikitable-centered) that applied to the specific cells (.wikitable-centered tbody tr td). EVula // talk // ☯ // 21:14, 27 January 2009 (UTC)

:::::That doesn't help terribly much, as we would have to type class="wikitable=centered" for each cell in that column, which isn't all that superior to typing style="text-align:left". The only place this helps is for tables in which all columns are to be centered, and I think that is a distinct minority. TheDJ is correct—we really need COLGROUP syntax to make this sort of thing easier for cleaner table markup. Also note to Mattbr—the exclamation mark is ill-advised outside the table header (hence the darker shading and bold text). It is one of my pet peeves to see that markup used for standard table rows. — Andrwsc (talk · contribs) 21:33, 27 January 2009 (UTC)

::::::No, .wikitable-centered would be the class for the table itself (class="wikitable wikitable-centered"). I'm imagining that this specific class would only include the CSS to make the cells centered, so it would be interoperable with any other CSS code that is inserted into the table. EVula // talk // ☯ // 21:48, 27 January 2009 (UTC)

:::::::Ok, but that still only helps with tables in which all columns are centered, and even without that, you can always use class="wikitable" style="text-align:center" in the table header as is done today. My point is that most tables require different alignment per column settings (whether that be left, right, or centered) and we have no good solution for that. For example, a very common style of table has a text string in at least the first column (left alignment) and numeric data in several others (right aligned or center aligned, as appropriate), and there is no easy way to do that without repeating style alignment code on a per-row basis. — Andrwsc (talk · contribs) 21:56, 27 January 2009 (UTC)

::::::::Ah, this is true... though to be fair, it's a limitation of CSS, not MediaWiki. The only thing I can think of would be to have a slew of superfluous classes that affected individual rows (ie: .wikitable-center3 would affect .wikitable-center3 tbody tr td+td+td, .wikitable-center5 would affect .wikitable-center5 tbody tr td+td+td+td+td), but that's hardly an elegant solution. EVula // talk // ☯ // 22:28, 27 January 2009 (UTC)

:::::::::This isn't such a bad idea as it sounds, although I'd go with .col3-center (not to mention .col3-left and .col3-right) and so on. No reason to limit it to wikitable tables and defaults vary. — Blue-Haired Lawyer 23:18, 27 January 2009 (UTC)

:::::::::: The sketched requirement is quite generic and having a (CSS) solution would clean up a lot of articles and make editing easier. Something along the lines of .col3-center tbody tr td+td+td {style="text-align: center"}, might be workable. However "col3-center" as above would affect all columns from 3 and up. So every switch of alignment would need to be marked. Centering only column 3 in a left aligned table would require col3-center col4-left. −Woodstone (talk) 10:46, 28 January 2009 (UTC)

::Apologies to all, I mis-read the question. The exclamation mark will format the headers but not the rest of the content and shouldn't be used for non-header cells as it mis-describes the cell and doesn't help accessibility. mattbr 07:15, 28 January 2009 (UTC)

:Q: What is the main roadblock standing in the way of WikiMedia supporting colgroups (and theads and tbodies as well)? What would "break" if support for them were added right now? SharkD (talk) 07:57, 28 January 2009 (UTC)

::The various inconsistent browser implementations of <col> and <colgroup> are the major obstacle. See for example [https://bugzilla.mozilla.org/show_bug.cgi?id=915] and [https://bugzilla.mozilla.org/show_bug.cgi?id=2212]. For mediawiki's specific hesitation, see various comments in bug 986, like 986#c25. Also see [http://www.w3.org/TR/CSS21/tables.html#columns] (per the specs, alignment isn't necessarily supported in columns). --Splarka (rant) 08:28, 28 January 2009 (UTC)

:::tbody and thead, on the other hand, have been part of the overall table concept for a long time, however; I would think that those could be enabled without much issue. EVula // talk // ☯ // 17:33, 28 January 2009 (UTC)

Currently there exists the inconsistent set of templates:

• {{tl|left}} style="text-align:left"|
• {{tl|left2}} style="text-align:left"|{{{1|}}}
• {{tl|right}} <div align="right">{{{1}}}</div>
• {{tl|center}} <div align="center">{{{1}}}</div>

As an intermediate solution, while fleshing out the CSS approach, we could define a new set

• {{tl|ta-l}} style="text-align:left"|
• {{tl|ta-r}} style="text-align:right"|
• {{tl|ta-c}} style="text-align:center"|

−Woodstone (talk) 09:43, 29 January 2009 (UTC)

{{user|66.230.230.230}} is a Tor node, but the block form says it's a sensitive IP. Surely not every IP from 66.230.192.0 to 66.230.239.255 is the secure gateway (as listed on MediaWiki:Blockiptext). The only secure server IPs I know are [http://samspade.org/whois/66.230.200.219 66.230.200.219] and [http://samspade.org/whois/208.80.152.134 208.80.152.134]. Spellcast (talk) 09:06, 28 January 2009 (UTC)

:66.230.200.0/24 was our old IP range for our Tampa servers -- that covers 66.230.200.1 through 66.230.200.255. 66.230.230.230 is not and never was in our IP range. Feel free to block it or whatever. :) --brion (talk) 19:11, 28 January 2009 (UTC)

::Done, but the IPs should probably be updated at MediaWiki:Sysop.js. Spellcast (talk) 10:44, 29 January 2009 (UTC)

Has anyone done a count of the urls to find what the top ten sources for Wikipedia are? I suspect CNN, The New York Times, Time magazine and the Associated Press via secondary outlets. They all have archives online. --Richard Arthur Norton (1958- ) (talk) 12:47, 28 January 2009 (UTC)

:Hi Richard, I don't have an answer to that question but I have a project of building a list of websites that can be considered as a reference for non-niche topics, that might interest you. And if you get an anwser to your question I would be highly interested :-) See this discussion. Cheers Nicolas1981 (talk) 05:59, 30 January 2009 (UTC)

{{Resolved|We've disabled this pending it actually working correctly}}

I just created a redirect at Buffalo Ridge Railroad. Between the text and the categories, on the right side, it says "[Mark this page as patrolled]". But when I click it (http://en.wikipedia.org/w/index.php?title=Buffalo_Ridge_Railroad&action=markpatrolled&rcid=274038115), it takes me to a page with title "Cannot mark as patrolled" and text "You are not allowed to mark your own changes as patrolled. Return to Special:Newpages." --NE2 17:55, 28 January 2009 (UTC)

:Well, that would defeat the patrol system if you could patrol your own edits. — neuro^(talk) 17:58, 28 January 2009 (UTC)

::But why is this just showing up now? --NE2 18:12, 28 January 2009 (UTC)

:::The link to patrol a page should appear only when you go the page from Special:Newpages. If you didn't do that, then something funny's going on. Algebraist 18:23, 28 January 2009 (UTC)

::::Yeah, I noticed it on User:DA01's talk page before he got renamed. I had ngone there from WP:CHU to explain something to him. -Jéské Couriano ^{(v^_^v)} 18:29, 28 January 2009 (UTC)

:::::Anything potentially in the same area changed here? — neuro^(talk) 18:49, 28 January 2009 (UTC)

::::::Yes. Looks like this was the fixing of bugzilla:15936. Algebraist 18:57, 28 January 2009 (UTC)

:Was it intentional? Seems so. — neuro^(talk) 19:05, 28 January 2009 (UTC)

::The problem was that the [mark as patrolled] link would only show up if there was an rcid in the URL for the page, but this was only added by Special:NewPages. So if someone edited the page to add a cleanup tag or whatever, the patrol link wouldn't be available after they saved, unless they went back in their browser history or back to NewPages. Mr.Z-man 19:17, 28 January 2009 (UTC)

:::But you can't mark your own edits as patrolled, so why was that a problem? --NE2 20:35, 28 January 2009 (UTC)

::::You can't mark pages you create as patrolled. You can still patrol it after you edit it as long as you didn't create it. Mr.Z-man 20:37, 28 January 2009 (UTC)

:::::Oh - patrolling is on a per-page basis, not a per-edit basis? Shouldn't the error message be changed then, from "changes" to "new pages"? --NE2 20:47, 28 January 2009 (UTC)

::::::On Wikipedia, Patrolled edits only apply to new pages. Other wikis using the MediaWiki software have the option of patrolling all edits. This is the case on another wiki I am involved in. Edits made by Administrators on that wiki are marked as patrolled by default (the "autopatrol" option). -=# Amos E Wolfe ^talk #=- 22:28, 28 January 2009 (UTC)

:::::::That's the reason the MediaWiki default message refers to 'changes', but there's nothing to stop us changing MediaWiki:Markedaspatrollederror-noautopatrol to something more appropriate just for en.wikipedia. Algebraist 22:35, 28 January 2009 (UTC)

::::::::Changed message to refer to "pages", as suggested. Happy‑melon 13:56, 29 January 2009 (UTC)

{{Resolved|We've disabled this pending it actually working correctly.}}

Okay, so I get the fact that pages can now be marked as patrolled even when not clicked through from the New Pages list of articles. However, even non-patrolled non-article pages now have the "Mark as patrolled" link. Is this really necessary? I came across several Talk pages that have the link, and I don't think they need them. Gary King (talk) 22:37, 28 January 2009 (UTC)

:Same here (Wikipedia:Featured list candidates/List of furry conventions). Dabomb87 (talk) 23:42, 28 January 2009 (UTC)

We could hide these outside the mainspace with CSS like this:

div.patrollink { display: none; }

.ns-0 div.patrollink { display: block; }

This hides the link in all namespaces, and then shows it again in the mainspace. Failsafe in that browsers without CSS won't do anything, and no browser will support display:none without display:block. Happy‑melon 14:02, 29 January 2009 (UTC)

We have a special page for uncategorized articles, Special:UncategorizedPages. I was wondering if it would be possible to have a similar listing of uncategorized templates (i.e. pages in the Template namespace which do not contain a category). I think this would be very useful for maintenance and organizational purposes. --Eastlaw ^talk · _contribs 20:24, 26 January 2009 (UTC)

:Special:UncategorizedTemplates. --Splarka (rant) 08:18, 27 January 2009 (UTC)

Never mind, sorry. --Eastlaw ^talk ⁄ _contribs 10:53, 30 January 2009 (UTC)

{{Resolved|1=ukexpat (talk) 15:55, 30 January 2009 (UTC)}}

I know that stats.grok.se is able to report the number of article page views per day, but is there a similar service for the number of edits per day? Thanks! SharkD (talk) 22:02, 28 January 2009 (UTC)

:Not quite what you asked for, since it doesn't include history, but http://www.wikirage.com/ gives the most actively edited articles for various time ranges up to the present.-gadfium 01:46, 29 January 2009 (UTC)

::Thanks, anyway. SharkD (talk) 04:04, 29 January 2009 (UTC)

One or two days ago I moved the article named Aglio e Olio, about the album by the Beastie Boys, to Aglio e Olio (Album) to make room for Aglio e Olio, an article about the Italian pasta dish. I also edited the :Template:Beastie Boys to link to the new article name, a template which appears on about 100 BB's song articles. As far as I can tell, the system has purged the old template version to the new one (I've checked about ten occurrences of the template) but the backlinks at Aglio e Olio are stuck with the old Beastie Boys template version.

I realize that the backlinks are not updated immediately but via the job queue but, I'm starting to think that something isn't working. Do I need to do anything to trigger the backlink update? hydnjo talk 02:46, 29 January 2009 (UTC)

:A job queue over 1 million suggests a fairly significant backlog at the moment. You shouldn't have to do anything, the software will clear it eventually. If it is really important to you that this get done right now, you can force the parser to process a page immediately but going to each one, opening the text and hitting "Save" without changing anything. Doing that would have the effect of immediately updating all links on that page. But unless you are in some kind of rush, I wouldn't bother. It will clean itself up after a while. Dragons flight (talk) 03:39, 29 January 2009 (UTC)

::OK, and thanks for the link, I'll give it a few more days. hydnjo talk 03:49, 29 January 2009 (UTC)

:::This month some link updates have not occurred after weeks. One report concerned an edit from December 20.[http://en.wikipedia.org/w/index.php?title=Template:Country_Radio_Stations_in_Illinois&diff=259126176&oldid=240148716] 40 days later Special:WhatLinksHere/WAAG (FM) still includes WBWN which transcludes the template. PrimeHunter (talk) 04:27, 29 January 2009 (UTC)

::::Special:WhatLinksHere/WAAG (FM) has finally removed WBWN. I wonder whether somebody made a null edit after reading the above. PrimeHunter (talk) 13:03, 29 January 2009 (UTC)

:Again, I've raised this issue before. The backlog is not acceptable - it needs to be cleared down much faster than it currently is. Whenever I move a page to create a disambig article, I ensure that nothing in the article mainspace points to the disambig page. I hope other users do the same. But if it can take 40+ days for templates to refresh, no-one is going to bother checking. I kept an eye on two pages I moved prior to Christmas, and two weeks later, they still had dozens of articles pointing at them, despite the template being updated. I hope after Jimbo's last appeal for some extra cash, he's splashed out on a new server! Lugnuts (talk) 08:19, 30 January 2009 (UTC)

::We're currently investigating how to better monitor what's going on in the job queue and patch up problems with updates. --brion (talk) 17:12, 30 January 2009 (UTC)

Does there exist a template for Talk pages such that references can be contained within them without obstructing discussions? I've placed a purple box around the references in one article's Talk page, but I've already "tripped" over it and messed up the Talk page once. Thanks! SharkD (talk) 04:18, 29 January 2009 (UTC)

:I'm not sure that I understand your problem. Did you get a big red message about references? --—— Gadget850 (Ed) ^talk - 13:03, 29 January 2009 (UTC)

::I guess the problem is that if {{tl|reflist}} is placed at the bottom of a talk page source then many users will create new sections below it so it doesn't remain displayed at the bottom. Or do you want a "section reference" template which only displays references occurring in that section, so different sections can display their own references, also after archiving? PrimeHunter (talk) 15:32, 29 January 2009 (UTC)

:::The first, not the second. Ideally the template could be added to the top of the page so that editors' edits don't interfere with it, but automatically render the text at the bottom of the page as per normal. A colored background and border would also be good in order to indicate that the section should not/cannot be edited. SharkD (talk) 15:46, 29 January 2009 (UTC)

::::I was pondering something like this as well. Don't we have a footer message box template someplace? I have a real life interrupt for a few hours. --—— Gadget850 (Ed) ^talk - 16:06, 29 January 2009 (UTC)

:::::I went ahead and created a template that meets most of these requirements: {{tl|Reflist-talk}}. SharkD (talk) 22:23, 29 January 2009 (UTC)

::::::I saw that and did a quick test. Problem is that it depends on being at the bottom of the page. If someone simply clicks on +, they will create a new section below it. It is also susceptible to being archived by a bot. We need a template above the top that will always show at the bottom. --—— Gadget850 (Ed) ^talk - 01:23, 30 January 2009 (UTC)

:::::::This can't be done with a template, but it could be done with a magic word of some sort. It needs to be done at the Mediawiki level rather than the template level. — Carl (CBM · talk) 02:05, 30 January 2009 (UTC)

::::::::Could you explain specifically what these changes would be? Do you mean some sort of autogeneration of reference sections, or just a flag that will make any template fall to the bottom if included? I'm thinking JavaScript could be used to accomplish this as well. I.e, a script takes the element out of normal flow and inserts it as the last child of the parent div. SharkD (talk) 05:31, 30 January 2009 (UTC)

:::::::Yeah, I've tripped over it a few times as well, though it's easy to fix. SharkD (talk) 05:32, 30 January 2009 (UTC)

:::::::::Probably the latter, something to the effect of __MOVEREFSLISTTOBOTTOM__. It would first need to start picking up refs declared after the reflist. These don't seem to populate a second reflist either, just ignored. This should never affect articles as the "see also", "external links" sections should not contain "contentious statements" needing a ref, and their content should be supported/mitigated by the upper part of the article anyway. — CharlotteWebb 19:49, 30 January 2009 (UTC)

::::::::::The error message when the <reference/> tag is missing is always after the last section. We could put the <reference/> tag in the message so that the error actually generates the reference list. I would want more eyes on this. --—— Gadget850 (Ed) ^talk - 21:37, 30 January 2009 (UTC)

{{Resolved|1=ukexpat (talk) 15:54, 30 January 2009 (UTC)}}

I'm currently working on a website (web version of a print publication) that uses footnotes. I'd like to have it behave similarly to how they behave here, ie the entire ref line gets highlighted (background color change) when you click the anchor (and vice-versa when you click the ^ back up to the article). I don't need the footnote code automatically generated or anything; it's just a static site. Couldn't find the code after a quick perusal, so I thought I'd ask here, since it's vaguely on-topic for the pump. ;) EVula // talk // ☯ // 21:15, 29 January 2009 (UTC)

:You want the Cite.php extension. good luck. --—— Gadget850 (Ed) ^talk - 22:20, 29 January 2009 (UTC)

::I'm pretty sure that the highlighting he is referring to is somewhere in our local CSS and/or Javascript files and not part of the extension itself. Dragons flight (talk) 22:26, 29 January 2009 (UTC)

::: Can be done quite easily. The jumping is easy, it's just an anchor. The harder part is coloring the link when it is clicked on; however, if you look at MediaWiki:Common.css, the {{bcode|sup.reference:target}} part is what you want. Wrap the footnote's link in <sup class="reference"> and you're good to go. Gary King (talk) 22:30, 29 January 2009 (UTC)

::::Yeah, I've got the anchors set up (like you said... they're just anchors), but was unfamiliar with the :target CSS pseudo-class. That would explain why my searching the .js files didn't turn up anything; should have known to look for the background color in the .css files. D'oh.
I'm handling the individual footnotes as an ordered list, so I'm just applying the class to that and setting the background on the list item; it works like a charm (and I'll be combining it with a smooth-scroll jQuery script, so it'll be pretty nifty). Many thanks. :) EVula // talk // ☯ // 22:51, 29 January 2009 (UTC)

Not sure if this has been brought up before, but I have noticed in the last few months there are problems with the printing of some pages on some articles. Most pages print fine, but some pages (within the same article) will print blank. I don't think the problem is with my computer, as I have same computer and have not switched anything on my end. I think certain links are causing this problem, as in a couple of articles I noticed that if I removed a certain link (temporarily, not permanently) this solved the printing problem. --WordsExpert (talk) 21:39, 29 January 2009 (UTC)

:What links on what articles seem to be causing this? Algebraist 22:32, 29 January 2009 (UTC)

:: I have noticed it on at least a couple dozen articles, one example page 4 on Constantin Brancusi. --WordsExpert (talk) 01:07, 30 January 2009 (UTC)

:::And which browser and operating system ? --TheDJ (talk • contribs) 12:36, 30 January 2009 (UTC)

::::Using Vista and the latest version of Explorer as the browser. --WordsExpert (talk) 21:29, 30 January 2009 (UTC)

{{Resolved|1=ukexpat (talk) 15:51, 30 January 2009 (UTC)}}

I noticed something weird on the News page. When I log off, there appears to be "poo" written between the disambiguation link and the introduction. But when logged in, no such thing appears. It doesn't seem to be in the page when I try to edit it at all, and I can't tell who and when added this from the page history. This is intriguing, can anyone explain what is going on? Thanks. --Urzică (talk) 22:43, 29 January 2009 (UTC)

:Probably template vandalism. Purging the server cache seems to have fixed it. Algebraist 22:45, 29 January 2009 (UTC)

::Yes, it was vandalism to {{tl|journalism}}. Different caches are kept for logged-in and non-logged-in readers, so it's possible for vandalism to persist in the cache for one but not the other. With the present long job queue the page was taking a long time to update by itself. Algebraist 22:47, 29 January 2009 (UTC)

:::Oh, right. Thanks for explaining (and fixing) it. --Urzică (talk) 22:53, 29 January 2009 (UTC)

Why are there a number of noinclude codes showing up in the ref section of Anandabhadram? Aditya^{(talk • contribs)} 15:24, 30 January 2009 (UTC)

:I don't see them. Have you tried clearing your cache and doing a server purge? – ukexpat (talk) 15:53, 30 January 2009 (UTC)

:Anandabhadram uses {{tl|Cite web}} and was edited during the 2 minutes [http://en.wikipedia.org/w/index.php?title=Template:Cite_web&diff=next&oldid=267409828] where that template had an unmatched . The template has been fixed and the article updated (but it's possible other affected articles have not been updated yet after the template was fixed). PrimeHunter (talk) 23:13, 30 January 2009 (UTC)

I upgraded to IE8 (RC1) last night and there are problems with footnotes. (1) rather than the footnote being a superscript, it is a small subscript, and (2) down in the RefList all footnotes have the number 0. Bubba73 (talk), 18:00, 30 January 2009 (UTC)

:Sounds like an IE8 rendering/standards support issue to me. – ukexpat (talk) 18:59, 30 January 2009 (UTC)

I've got the "Prompt me when entering a blank edit summary" option enabled in my preferences. Before, when I create a new page and leave the edit summary blank, one would be automatically generated for the page's creation, like "Created page with 'new text'". I believe it still does this, but now, when I leave the edit summary blank, it warns me that I have not entered an edit summary. The problem with this is that the software used to know that it would automatically generated an edit summary for me, but now it seems as if it doesn't know this anymore? Gary King (talk) 18:04, 30 January 2009 (UTC)

If you use this inane feature it rejects the blank summary on the first attempt but adds a hidden input which prevents the same from being rejected on the second attempt. You can use javascript to add it on the first attempt under certain conditions (to skip the reload).

if(wgAction=="edit" && wgCurRevisionId == false) //if i'm editing a red link
document.getElementById("editform").innerHTML += //add this parameter to the form as the edit summary
''; //won't really be blank

: It's okay, nothing major. I don't want to change it just to work for me; I'll use an edit summary of "creating" or something. The reason I'm bringing this up, though, is because the functionality just changed recently, within the past week. I'm curious as to why it was changed. Gary King (talk) 18:50, 30 January 2009 (UTC)

I look at Special:ListGroupRights and I see, instead of the two "hiderevision" and "oversight" rights, I see this:

• Delete and undelete specific revisions of pages (deleterevision)
• Review and restore revisions hidden from Administrators (hiderevision)
• Review and restore revisions hidden from administrators (suppressrevision)
• View a previously hidden revision (oversight)

What are these rights, what is the diffence between deleterevision, hiderevision, and supressrevision? Please use {{tl|tb}} on my talk page when you have an awnser.--Ipatrol (talk) 22:56, 30 January 2009 (UTC)

:See [http://lists.wikimedia.org/pipermail/wikitech-l/2009-January/041192.html]. Happy‑melon 23:49, 30 January 2009 (UTC)

If you try adding an image on the Bad Image list, it doesn't show up, but it also vanishes the next paragraph (or several paragraphs) of text.

For example, the following code:

::Example:

::

::This text should appear in the page, but does not!

::So should this line!

::*And this list item!

::~~~~

Produces:

:Example:

:

:This text should appear in the page, but does not!

:So should this line!

:*And this list item!

:Bushytails (talk) 00:16, 31 January 2009 (UTC)

Rather than (using a non-"bad" image so something shows up):

:Example:

:Image:Strapon01.jpg

:This text should appear in the page, but does not!

:So should this line!

:*And this list item!

:Bushytails (talk) 00:16, 31 January 2009 (UTC)

Ignoring the inherent evilness of a bad images list, this apparant bug is pretty annoying... Everything until the next major change (adding a signature stops it, hence why I used it in the example) just vanishes. No trace of it is in the output html. So, is this a bug, or something else? Bushytails (talk) 00:16, 31 January 2009 (UTC)

:This is also reported in bugzilla:16039. PrimeHunter (talk) 02:24, 31 January 2009 (UTC)

MediaWiki is being upgraded on the sites now... see mw:This week's software updates for details. --brion (talk) 03:55, 28 January 2009 (UTC)

:Page doesn't exist (yet). =) —Locke Cole • t • c 04:06, 28 January 2009 (UTC)

::A little bird told me that mw:This week's updates might be a better place to look. :) {{Nihiltres|talk|[http://en.wikipedia.org/wiki/Special:Log?user=Nihiltres log]}} 04:12, 28 January 2009 (UTC)

:::Thanks for fixing the link. :) --brion (talk) 19:14, 28 January 2009 (UTC)

I don't know if this was intentional but [rollback] links now appear in the watchlist.

:Yep. --brion (talk) 19:14, 28 January 2009 (UTC)

::Would an option to turn it off be possible? It's a bit cluttery. ♫ Melodia Chaconne ♫ (talk) 20:23, 28 January 2009 (UTC)

::As well as cluttery, I find the [rollback] link in the watchlist downright scary, given my stabby way of clicking on things. I don't want to be accidentally rollbacking. Having an option to turn it off for the watchlist would be a great thing. Since they all use the same "mw-rollback-link" style, is the only option going to be a page-specific .js to make the style invisible for Special:Watchlist? Franamax (talk) 22:12, 28 January 2009 (UTC)

/* hide rollback from Recent Changes */

.page-Special_RecentChanges .mw-rollback-link {display:none;}

/* hide rollback from Watchlist */

.page-Special_Watchlist .mw-rollback-link {display:none;}

:::It only needs some CSS. But, it could be done via JS, or even undone by JS (a 'show rollbacks' link). Perhaps as a gadget. --Splarka (rant) 23:10, 28 January 2009 (UTC)

Also this revision breaks the js script "show/hide top contribs", which is odd because all that script does is hide the contributions from user contribs if that user is the most recent contributor.-- penubag  (talk) 04:23, 28 January 2009 (UTC)

:I, for one, would love to have a script or preferences option to hide these links. I have talked to a few admins who find them even scarier, as they have protected pages watchlisted. — Jake Wartenberg 16:13, 29 January 2009 (UTC)

::Nevermind. That CSS works great! — Jake Wartenberg 16:15, 29 January 2009 (UTC)

I agree with the users who find this scary. Can rollback be hidden when checking a user history as well? That is what I find the most scary. PSWG1920 (talk) 16:37, 29 January 2009 (UTC)

:Yes:

:.page-Special_Contributions .mw-rollback-link {display:none;}

:— TKD::Talk 16:57, 29 January 2009 (UTC)

In addition to being scary and cluttery, these add bunches of extra queries to RC and watchlist. I'm seeing 370 queries on enwiki RC right now, where I have rollbacker (40 for dewiki where I don't, not sure if that's gone up). Maybe this change should be reverted. —Simetrical (talk • contribs) 02:12, 30 January 2009 (UTC)

:It's absolutely brilliant and makes life a lot easier. Ty 03:59, 30 January 2009 (UTC)

::Well, that's great. I'm glad that you like it. But I think it really should be a preference, and probably not the default. — Jake Wartenberg 03:46, 31 January 2009 (UTC)

:::@Simetrical: Brion scapped it, so its performance impact is acceptable. Ergo, its performance impact is not an issue.

:::@Jake: If this inconsistency hadn't always existed, and edit logs had always consistently displayed rollback links, there wouldn't be any issue, so why should the reverse be any different? There really isn't any difference between a rollback link on a user's contributions, to a rollback link on the RC feed, to a rollback link on a watchlist: they both provide the same amount of information as to the legitimacy of the edit. This improves the consistency of the interface by standardising the appearance of the edit logs; if you want to reintroduce that inconsistency to suit your personal preferences, then that's perfectly fine, and you can do so easily with CSS. Happy‑melon 16:31, 31 January 2009 (UTC)

I was looking at Wikipedia:Deletion review#Steps to list a new deletion review and saw that Template:Newdelrev requires the user to manually enter the namespace name. I thought it might be useful to have a way to detect the template name automatically, so I made Template:Namespace for this purpose. However, now I'm unsure that Template:Newdelrev really needs it. What other applications could this be useful for? —Remember the dot ^(talk) 03:42, 30 January 2009 (UTC)

:BOWS. DOWN. IN. WORSHIP. You do realise you've just found the most important string-manipulation function we've ever wanted, and the most powerful after a regular expression test? If you can think of a way to get the last X characters of a string, then we really are set: those two ideas can be trivially extended to string slicing, then to string splitting, then to length functions, then to string comparisons. I'm not joking to say that we can build pretty much the entire StringFunctions library from those two ideas. Have a cookie {{mono|1=:D}} Happy‑melon 17:50, 30 January 2009 (UTC)

::Umm, if we are going to use this as a substring hack, it would make more sense to actually include native support for a substring function in MediaWiki. --- RockMFR 17:58, 30 January 2009 (UTC)

:::I couldn't agree more, it's hackish and the programmer within me rebels in horror. Go vote for {{phab|8455}} {{mono|1=:P}}. It's been open for two and a half years now, despite the code having being available the entire time. Maybe if we demonstrate that we need this functionality enough to hack it up from the fundamentals, the devs will be more inclined to give us the native functionality. Happy‑melon 19:36, 30 January 2009 (UTC)

StringFunctions has not been enabled in part because of its significant impact on the nature of wikitext, and the performance implications of this. Reconstructing string functions from existing parser functions is not necessarily a sensible response to this, with significant usability and performance implications. The behaviour exploited was also a bug, which broke in numerous cases. Consequently, in rev:46628, I have disabled the ability to truncate strings using the pad parser functions – this change will take effect towards the middle of next week. The advice of the technical team continues to be that if you want to write something requiring string functions, you should implement it as a parser function and not as a template. — Werdna • talk 00:49, 31 January 2009 (UTC)

:Quick note – in accordance with this principle, a parser function with the same functionality as the {{tl|namespace}} template will be added in the next few days, either by Mr.Z-man or myself. — Werdna • talk 00:57, 31 January 2009 (UTC)

::See my comments in Code Review, your patch misunderstood how pad was being (ab)used. Dragons flight (talk) 01:15, 31 January 2009 (UTC)

:::"If you want to write something requiring string functions, you should implemente it as a parser function and not as a template". Why do you think that bug has been open for thirty months then? Because we wanted to write things requiring string functions, and asked you (the royal you, don't worry I'm not blaming you) to "implement it as a parser function" so we didn't have to implement it as a template. I agree with your principle, of course I do, but I don't believe that this can somehow be interpreted as "not a sensible response" on the non-devs' part: we asked for functionality, we didn't get it, we waited an obscene length of time, still didn't get it, so we hacked it up ourselves from what we have got, the devs see it and quite rightly are scared, but they realise that, amazingly enough, we weren't joking when we asked for this functionality the better part of three years ago, and so we finally get it natively. Sounds like a win-win situation as far as I'm concerned. Happy‑melon 10:30, 31 January 2009 (UTC)

:Okay, by removing that functionality you're going to kill Wikisource's ability to automate the process of determining the first letter of an author's last name. This was a much-hoped-for feature for wikisource:Template:Author (see wikisource:Scriptorium#Remove last_initial parameter from Template:Author). Doing a substring using padright is about as intensive as doing a substring normally would be - the limitation is that you have to start the substring from position 0, rather than an arbitrary point within the string. Thus, it's really not possible to reconstruct other string functions from this. Also, what did you mean by "The behaviour exploited was also a bug, which broke in numerous cases"? Can you post some examples? —Remember the dot ^(talk) 01:03, 31 January 2009 (UTC)

::I won't explain how, but this trick plus certain other existing parser functions is sufficient to give you slicing. Once you have that, I believe you can replicate nearly the entire string library. Speaking of which, I personally believe enabling some version of StringFunctions is in the community's interest. Dragons flight (talk) 01:23, 31 January 2009 (UTC)

:::I know it is possible to combine the parser functions in an extremely unreliable and inefficient way to achieve string slicing. The problem is that the method wouldn't be reliable enough to be usable. Could you maybe e-mail me and we can discuss what we were each thinking?

:::In any case, I'm more interested in knowing what the bug was that made padright break in "numerous" cases. —Remember the dot ^(talk) 01:32, 31 January 2009 (UTC)

:rev:46630 adds parser functions for NAMESPACE, SUBJECTSPACE, and TALKSPACE that basically allow them to be used the same way as the namespace template, minus the crappy hackiness. Mr.Z-man 01:52, 31 January 2009 (UTC)

::Thank you. Is it too much to ask for a normalizing PAGENAME function also? And there's still the issue of Wikisource needing to be able to extract the first character of a string. —Remember the dot ^(talk) 02:04, 31 January 2009 (UTC)

Can I somehow filter out editors from a log, say MZMcBride from the deletion log or is all what can be done already being described at Help:Special_page#Logs?--Tikiwont (talk) 13:44, 30 January 2009 (UTC)

The delete-bot uses the same summary each time so you can hide it on your end and pay attention to the other lines.

if(wgPageName=="Special:Log")
document.body.innerHTML = document.body.innerHTML
.replace(/
/gi, "");

:Thanks, sounds useful; and this goes to monobook.js or where else?--Tikiwont (talk) 19:56, 30 January 2009 (UTC)

::Yeah. You can add other common flood-phrases Regex#Basic_concepts. — CharlotteWebb 20:12, 30 January 2009 (UTC)

:If you just want to hide entries that contain "old ip", simply add

importScript('User:MZMcBride/hideentries.js')

:to your user skin subpage (usually Special:MyPage/monobook.js). Cheers. --MZMcBride (talk) 07:44, 31 January 2009 (UTC)

::Thanks to you both. For some reasons only the second, more cryptical one, worked for me, but that is fine.--Tikiwont (talk) 08:41, 31 January 2009 (UTC)

Hey just wondering why scripts via Special:Mypage/skin.js no longer work. If this is my issue please tell me how to fix it but at this point I think its a software issue. Any ideas? Alexfusco⁵ 01:18, 31 January 2009 (UTC)

:If you mean User:Alexfusco5/monobook.js it would probably something you did editing it today (or recently). Can you check the error console log in your browser? Alternately, perhaps you changed skins or even disabled javascript accidentally? --Splarka (rant) 01:40, 31 January 2009 (UTC)

::JavaScript is on and I'm positive I didn't change skins. Furthermore, I'm pretty sure I've been having this problem for a while because I remember all scripts in monobook.js just stopping sometime in December. It's only those scripts gadgets work fine and I'm thinking it may be a software bug or a config change and I would first like to make sure I'm not the only one experiencing the problem. Alexfusco⁵ 14:54, 31 January 2009 (UTC)

:::No-one else is reporting it, so this might be a browser problem. Perhaps your browser (or some addon or extension) is blocking javascript from that exact page? Algebraist 14:59, 31 January 2009 (UTC)

::::I'm using FF3 as my browser and I don't think I have any add ons that are doing this I'm checking now just to be sure Alexfusco⁵ 15:11, 31 January 2009 (UTC)

:::::My fault when I checked my error log I wasn't scrolled down enough to see the error thanks for your help :) Alexfusco⁵ 15:21, 31 January 2009 (UTC)

I doubt anyone uses more than one skin on a regular basis but shouldn't "User:So-and-so/common.js" be recognized as something to load in all skins? — CharlotteWebb 15:04, 31 January 2009 (UTC)

:Users in fact do not change skins very often. But more importantly, a user/common.css/js edit-breaking screwup would be overly hard to work-around for some users. Imagine accidentally adding body {display:none;} in your monobook.css. You can easily work around this with ...monobook.js&action=edit&useskin=myskin but not if it was in a common.css. See bugzilla:10183 for both sides. --Splarka (rant) 22:21, 31 January 2009 (UTC)

As I write this, WT:CWNB is included in :Category:Categories for discussion notices - but it's not a category for discussion. I know that this is because of the first thread on the talk page (as I write, I tested and reverted), discussing a proposed CfD - surprise surprise.

My question here is: what reliable method is available for me to discover where exactly a category has been transcluded onto the page? Is my only option to use my wits to figure it out?

Note that I'm not talking here about mistakes with Category:blah versus :Category:blah, it's the mysterious template transclusions I'd like to pin down. Thanks! Franamax (talk) 10:39, 31 January 2009 (UTC)

:You mean figuring out whether the categorization is done inline in the article or from a template transclusion? The advanced search function may be helpful, i.e. in this case [http://en.wikipedia.org/w/index.php?title=Special%3ASearch&ns10=1&redirs=1&search=Categories+for+discussion+notices&fulltext=Advanced+search this search] which looks for the string "Categories for discussion notices" in all templates (don't put "Category:" in front, this overrides the requested namespaces and restricts the search to the category namespace). I don't know of another way other than looking at each transclusion individually. However in this case it's not a transclusion - it's right there in Wikipedia talk:Canadian Wikipedians' notice board, in the first thread generated by a variant of Template:Cfdnotice. I think these notices are meant to be cleaned up when the CfD discussion is closed. For whatever reason this one seems to have been missed. -- Rick Block (talk) 16:21, 31 January 2009 (UTC)

Hi, when I'm logged in, some infoboxes are not showing up. Examples of pages with the problem are Finnmark, Buskerud, Troms. Examples of pages without the problem are Sweden, Norrbotten and many others. I don't have any gadgets enabled and my Monobook.js is empty. What is causing this problem? --Gerrit ^{CUTED[http://www.topjaklont.org/ H]} 12:19, 31 January 2009 (UTC)

:When I log out and do Shift+Reload (Firefox 3.0.5) I get the same problem. --130.239.112.204 (talk) 12:21, 31 January 2009 (UTC)

::It seems to be fixed now! --Gerrit ^{CUTED[http://www.topjaklont.org/ H]} 12:23, 31 January 2009 (UTC)

:::Someone broke the infobox, [http://en.wikipedia.org/w/index.php?title=Template:Infobox_Fylke&diff=267606796&oldid=267195749 it was fixed]. Regards, Woody (talk) 12:25, 31 January 2009 (UTC)

{{clear}}

I'd like to make it so that all of my userpages and user talkpages on other wikimedia projects automatically soft-redirect to my en.wiki User: and User talk: pages, so that users on projects where I'm only marginally active (like meta: or commons:), will leave messages on en.wiki. Is there a bot, script, extension, or other method to do this, or do I have to do this manually every time I land on a new project?--Aervanath (talk) 15:31, 31 January 2009 (UTC)

{{Resolved}}

How do I use in templates?

{{tl|Span|A}}{{tl|Endspan}} doesn't work.

Thank you in advance. Ikip (talk) 00:36, 1 February 2009 (UTC)

:If the usual code fails (it'll work in some situations), use {{#tag:span|contents of span go here|id="A"}}. See [http://www.mediawiki.org/wiki/Help:Magic_words here] for details. Algebraist 00:46, 1 February 2009 (UTC)

::20px Thank you alg, Ikip (talk) 00:55, 1 February 2009 (UTC)

::I found it: mw:Help:Magic_words#Miscellaneous :( Ikip (talk) 01:01, 1 February 2009 (UTC)

Is it possible to display an editnotice for all blps, identified by the :Category:Living people (see Wikipedia:Village pump (proposals)#Editnotice for all blps), something similar to what has been done for disambiguation pages (see here) ? Cenarium (Talk) 23:13, 31 January 2009 (UTC)

:Is there a template or metatemplate that is consistently included on each and every BLP, and is not included on any non-BLP articles? That's the situation with the disambig pages ({{tlx|dmbox}}), so it's easy to get javascript to add the editintro automagically. This can be done equivalently for BLPs very easily if, and only if, there is such a template. Happy‑melon 23:34, 31 January 2009 (UTC)

:::Not sure if there's anything that can be done about it or how much of a difference it would make for BLPs, but the disambiguation edit notice currently only displays (at least for me) by clicking the "edit this page" tab at the top of each page -- the edit notice does not appear when I click edit from a diff. older ≠ wiser 23:40, 31 January 2009 (UTC)

::::Indeed, this is because all the script does is append a parameter to that link {{mono|1=:D}} Happy‑melon 10:21, 1 February 2009 (UTC)

:User:RockMFR/blpeditintro.js. --- RockMFR 23:41, 31 January 2009 (UTC)

::It works for me :) Thanks, Cenarium (Talk) 00:11, 1 February 2009 (UTC)

Is it possible to also add the edit intro when editing a section ? Cenarium (Talk) 18:38, 1 February 2009 (UTC)

"Display an assessment of an article's quality" doesn't work for my account. I've tried using the Gadget option (didn't work) then the script in my monobook.js (didn't work). Locke'sGhost 13:03, 1 February 2009 (UTC)

: You should post at User talk:Pyrospirit/metadata with more detailed information. Gary King (talk) 17:33, 1 February 2009 (UTC)

Am i the only one to notice this? Recently i have noticed that no matter where i am it seems to be taking ages to get onto the watchlist? Is it just me or is there a loading prob? Simply south not SS, sorry 14:03, 1 February 2009 (UTC)

How come it says that the page was created [http://en.wikipedia.org/w/index.php?title=SOJ&diff=196523070&oldid=24856203 here] when it should say the content was replaced because of the previous edit? -- Mentifisto 16:51, 1 February 2009 (UTC)

:Because the previous edit was deleted at the time, and undeleted later. See [http://en.wikipedia.org/w/index.php?title=Special:Log&page=SOJ logs]. Algebraist 17:05, 1 February 2009 (UTC)

See WP:VPR#Proposal: Enable suppressredirect rights for sysops on the English Wikipedia. Prodego ^talk 01:35, 2 February 2009 (UTC)

Does anyone know the algorithm Special:Random uses? How long has it used that algorithm?

I'm asking because it affects the reliability of studies such as User:Knulclunk/Random.

Awhile back I did my own informal tests of Random awhile back and there seemed to be topic-clustering that varied day to day. The clustering may have been some underlying thing like article-creator-name, article-creator-date, rather than subject matter, but in any case it did not look random. davidwr/(talk)/(contribs)/(e-mail) 03:36, 2 February 2009 (UTC)

:There is no built-in topic-clustering. You can see most of the relevant code in [http://svn.wikimedia.org/viewvc/mediawiki/trunk/phase3/includes/specials/SpecialRandompage.php?view=markup SpecialRandompage.php]. As far as I know, when a page is made, it is given a random number between 0 and 1. Special:Random chooses a random number x, then chooses the page with the lowest number that is greater than or equal to x. Due to this, it is possible for an individual page to appear more frequently. If you had a very small wiki, you might notice this. --- RockMFR 04:33, 2 February 2009 (UTC)

There is something in Wikipedia:Deletion review that makes my browser (IE7) sluggish. I'm guessing it's JavaScript, and think it should be fixed. SharkD (talk) 01:12, 31 January 2009 (UTC)

:One thing I've been able to identify that is causing performance problems (high CPU usage) is the following CSS rule in [http://en.wikipedia.org/skins-1.5/monobook/main.css main.css]:

.TablePager tr:hover td { background-color: #eeeeff }

:Even though this is not in use on that particular page, it still causes problems. If inclusion of this rule is not critical, it should be removed. --- RockMFR 22:48, 31 January 2009 (UTC)

::Filed at bugzilla:17294. --- RockMFR 23:16, 31 January 2009 (UTC)

:Have you tried using [http://www.mozilla.com/en-US/firefox/ Firefox]? It's much faster and has built-in spell check to boot. —Remember the dot ^(talk) 00:10, 1 February 2009 (UTC)

::I use Avant Browser. It has all the features I need, thanks. SharkD (talk) 03:14, 3 February 2009 (UTC)

I was sorting out some redirects from (the now misleading) The Untitled Office Spin-Off to Parks & Recreation and I have hit a technical snag. According to the what links here page, the former has lots of The Office related pages linking to it [http://en.wikipedia.org/wiki/Special:WhatLinksHere/The_Untitled_Office_Spin-Off] but for the life of me I can't find out how or where the links are to fix them. They don't appear to be transcluded from a template. Can anyone explain this to me? Rockpocket 06:25, 2 February 2009 (UTC)

:It looks like the link to The Untitled Office Spin-Off was just removed from :Template:Theofficeus a few hours ago. It can sometimes take a few hours for the job queue to catch up with the removal of the link, so those articles will still be listed under What links here until the job queue catches up. Don't worry about it, it should be fixed in a couple hours.--Aervanath (talk) 06:49, 2 February 2009 (UTC)

:: Ah, that makes sense. Thanks! Rockpocket 07:00, 2 February 2009 (UTC)

:::In January it could take weeks for WhatLinksHere here to update. See #Stubborn backlinks. I don't know whether this has changed. PrimeHunter (talk) 14:09, 2 February 2009 (UTC)

OK this is a strange one. We have User:Little guru and his the user's contribs. But when I check out [http://en.wikipedia.org/w/index.php?title=User:Little_guru&diff=268029449&oldid=268029448 this diff of Little guru's user page] (which I've just recently history merged), there is a username of "Little_guru" (note the underline), who seemed to have contribs from 2001 to at least March 2002. The problem is that I can't get to that users contribs - Special:Contributions/Little_guru takes me to the spaced username, and so does Special:contributions/Little%5fguru with the URL encoded. Is there any way to solve this problem, besides checking out the user's contributions at Nostalgia Wikipedia, which doesn't work either? Graham87 13:55, 2 February 2009 (UTC)

:Underscores and spaces in article names including user names are treated as the same character (space) by the mediawiki software. I'm not sure if this has always been true. I've also heard early contribution histories are not reliable due to database changes. I'm not sure when this happened, but March 2002 sounds like the right kind of timeframe. Perhaps somebody who was around then can comment. Seems like for GFDL purposes there should be an explanation of this somewhere. I looked a bit but haven't found anything. -- Rick Block (talk) 14:59, 2 February 2009 (UTC)

::I'm aware of both of those things - I think it's odd that underlines weren't converted into spaces during the software conversion. Maybe we need a sysadmin to fish out the old contribs of the username with the underline. For info on article histories before MediaWiki was installed, see Wikipedia:Usemod article histories. The diffs I found with the odd username were after the conversion to MediaWiki, which makes the situation even more puzzling. Graham87 15:08, 2 February 2009 (UTC)

:::It's possible that the user table has two entries, one with user_name = "Little_guru" and one = "Little guru". Since all text names are normalized by the input parser now, you could never get at the contribs for "Little_guru" using the name (and neither could an admin).

:::You could ask someone with DB access on the toolserver to list the user_id entries in the user table where user_name matches Littleguru. Comparing these with the rev_user in the revisions table for rev_id = 268029449 and a modern diff for Little guru would give you the answer.

:::They may have had to clone all the user entries with an "_" in the name when they did the conversion, in order to keep password authentication working. That's all I can think of anyway... Franamax (talk) 17:41, 2 February 2009 (UTC)

::::Thanks, I've [https://jira.toolserver.org/browse/DBQ-56 made a request on the toolserver]. Graham87 02:59, 3 February 2009 (UTC)

Hello. The system had an error a while ago. I am supposed to show this to the system administrators:

Request: POST http://en.wikipedia.org/w/index.php?title=Talk:Oda_Nobunaga&action=submit, from 71.207.53.254 via sq21.wikimedia.org (squid/2.6.STABLE21) to 208.80.152.27 (208.80.152.27)

Error: ERR_READ_TIMEOUT, errno [No Error] at Mon, 02 Feb 2009 15:48:02 GMT

Unfortunately I cannot tell between system administrators on Wikipedia and the regular administrators. This sometimes happens (excepting the precise dates and times of course). -BlueCaper (talk) 16:22, 2 February 2009 (UTC)

:Wikipedia is really, really slow at the moment (hence the ERR_READ_TIMEOUT). I've yet to successfully load my watchlist. — TKD::Talk 16:41, 2 February 2009 (UTC)

All the templates seem to have busted for some reason, revealing only their names and a wikilink to the appropriate page. Anyone have an idea what the frag is happening? -Jeremy ^{(v^_^v Dittobori)} 19:28, 2 February 2009 (UTC)

:By no means all templates are broken. Can you give an example of a page on which one is? Algebraist 19:31, 2 February 2009 (UTC)

::Try {{tl|user}} and {{tl|checkuser}}. Sorry; I was in the middle of undoing a Jarl sock/jobber's damage and only saw what I saw at WP:SPI. -Jeremy ^{(v^_^v Dittobori)} 19:41, 2 February 2009 (UTC)

:::That page is too big, with too many templates. The post-expand include size is 2047998 bytes, the limit being 2048000. When the software hits one of these limits, it doesn't expand any templates over the limit. See Wikipedia:Template limits for details. Algebraist 19:50, 2 February 2009 (UTC)

::::Yeowch. Thanks for the info, Algebraist. -Jeremy ^{(v^_^v Dittobori)} 20:00, 2 February 2009 (UTC)

I experience a technical problem with the Holodomor talk page. My computer (Red Hat Enterprise) restarts immediatelly after I try to open a talk page. This never happens when I do that using Windows computers, and never happens when I enter other WP pages (including talk pages) from Linux computers. I use two different Red Hat Linux computers with the same result (immediate reboot). Can anybody explain me what does it mean?

Thank you in advance,

Best regards,
--Paul Siebert (talk) 19:40, 2 February 2009 (UTC)

This issue is being actively investigated. --brion (talk)

:{{resolved|An attempt to optimize the rendering of ref tags by caching them failed to take into account variances in inline data stripping. The change has been reverted. Affected pages may need purging.}}

I recently [http://en.wikipedia.org/w/index.php?title=Visna_virus&diff=268100404&oldid=268090535 added] a reference to Visna virus. However, it keeps showing up weirdly on my computer (no matter how many times I clear the cache)...This:

[http://discovermagazine.com/1994/dec/ofmythsandmischi458 "Of Myths and Mischief"], Discover, 1 December 1994

[http://discovermagazine.com/1994/dec/ofmythsandmischi458 "Of Myths and Mischief"], Discover, 1 December 1994

[http://discovermagazine.com/1994/dec/ofmythsandmischi458 "Of Myths and Mischief"], encephalitis, 1 December 1994

:I think the ref's name ("Discover") was conflicting with another ref. I changed it to "Discover-Dec94" and it seems to work now. EVula // talk // ☯ // 20:55, 2 February 2009 (UTC)

::I saw that change...there were no other ref names remotely close ot "Discover" so I'm not sure how it worked. Then again, after I saw the correct result following your change it reverted back to "encephalitis". Maybe I need to restart my PC...and get some coffee. — Scientizzle 20:58, 2 February 2009 (UTC)

:::It's still broken for me (perhaps my edit rebroke it somehow?). At the moment, changing the ref name fixes it in preview but not in the actual article. Algebraist 20:59, 2 February 2009 (UTC)

::::Whoa, yeah. That's confusing... purging doesn't help either. *scratches head* EVula // talk // ☯ // 21:03, 2 February 2009 (UTC)

:A new caching system was just installed for ref tags. It may still have a few bugs. Dragons flight (talk) 21:01, 2 February 2009 (UTC)

::Ahhh...interesting. My switch in browsers and a complete wipe of stored internet data didn't help, so the new ref caching system seems the culprit. — Scientizzle 21:04, 2 February 2009 (UTC)

:::Yeah I noticed this on another article (William Henry Ellis), the ref system now seems to use the first wikilinks it comes across in the article as destinations for wikilinks in refs. I thought it was the citation templates at first but I made some practice edits and it seems to do that no matter what is in the ref tags. Thought I was going mad at first, hopefully it will be sorted soon - Dumelow (talk) 21:05, 2 February 2009 (UTC)

Also see this at WT:FOOTY. D.M.N. (talk) 21:31, 2 February 2009 (UTC)

:An attempt to optimize the rendering of ref tags by caching them failed to take into account variances in inline data stripping. The change has been reverted. Affected pages may need purging. --brion (talk) 21:37, 2 February 2009 (UTC)

{{resolved}}

Why does reference 9 and 11 in a [http://en.wikipedia.org/w/index.php?title=Nemesysco&oldid=268105595 previous version] of Nemesysco render strangely, displaying strings like UNIQ174fb445295d998f-nowiki-00000059-QINU? They use {{tl|Cite journal}} and {{tl|cite study}}. It is ok since I removed an empty publisher= field. Discuss at Template_talk:Cite_journal#Rendering_problem. --Apoc2400 (talk) 21:39, 2 February 2009 (UTC)

:I think this might have been the same problem as in the thread immediately above. Algebraist 21:42, 2 February 2009 (UTC)

::You are right. --Apoc2400 (talk) 21:58, 2 February 2009 (UTC)

As pointed out here, there are several redirects not working: for example, Shell House Cliffs, Zeuxis and Parrhasius, and 2007 Battle of Gaza. What's the deal? Wareh (talk) 00:15, 3 February 2009 (UTC)

:Update. Ok, this seems like voodoo to me, but they've all three been fixed by removing the space after REDIRECT. Wareh (talk) 00:18, 3 February 2009 (UTC)

::I think it's the fact that you edited the redirects that made them work again, not removing the space. A null edit would have sufficed. — Edokter • Talk • 00:40, 3 February 2009 (UTC)

:::I've cleared the first bunch at Special:shortpages up to 29 bytes. Any help would be appreciated as this is not the most fun thing to do. :-) Carlossuarez46 (talk) 01:10, 3 February 2009 (UTC)

::::So why not stop doing that and see if someone can fix the bug instead? And thanks for the big runaround when I asked what was up. Jerk. Dicklyon (talk) 05:09, 3 February 2009 (UTC)

I've noticed an error message Cite error: tags exist, but no tag was found appearing recently at the bottom of some talk pages (eg Talk:Inauguration of Barack Obama and Talk:Overpopulation). I think it might be best if it there was no warning there because quite often people copy wikitext from the article to the talk page but talk pages are layed out differently to articles and a references section doesn't seem to fit. Can we have the warning turned off for talk pages? -- Barrylb (talk) 09:57, 28 January 2009 (UTC)

:Why not just refactor the page and wikicode the <ref> tags? --—— Gadget850 (Ed) ^talk - 12:42, 28 January 2009 (UTC)

::Yes it is possible but requires constant work which is quite unproductive on talk pages just to get rid of a warning message. Could the cite extension automatically display references at the bottom of the talk page instead of a warning? -- Barrylb (talk) 13:23, 28 January 2009 (UTC)

::Hmmm... this looks like a new message— there are a number of talk pages now in :Category:Pages with incorrect ref formatting for that problem. I'm sure that cite.php could be updated in that manner, but there are a few issues. --—— Gadget850 (Ed) ^talk - 14:19, 28 January 2009 (UTC)

:::This is not a problem exclusive to enwiki; I spotted the exact same error on the Chinese Wikipedia and Meta. EVula // talk // ☯ // 22:52, 28 January 2009 (UTC)

:::: It makes the most sense to me to only have this error message appear on article pages, and nowhere else. I agree that it is unproductive to edit references on talk pages just to remove the error message, and that it also doesn't make much sense for the extension to automatically generate a reference list even if the page doesn't explicitly ask for one. Gary King (talk) 22:59, 28 January 2009 (UTC)

Ok, can we get some comment from a developer who can decide whether to implement such a change? Barrylb (talk) 00:07, 29 January 2009 (UTC)

MediaWiki:Cite error should probably be blanked... I'm in favor of hidden categories rather than large red error messages. --MZMcBride (talk) 03:48, 29 January 2009 (UTC)

:Yeah, to fully hide a "Cite error" message, this needs to be reduced to "$1". One would then want to add "Cite error: " back in to any of the other cite error messages where one wants to keep that bit of text. Provided that is done, I think it may be possible to make the no error message display only on article pages by wrapping it in a namespace detecting ParserFunction. Dragons flight (talk) 05:50, 29 January 2009 (UTC)

::MZMcBride has now created that message with {{#switch:{{NAMESPACE}}|=Cite error: $1|#default=}} as a test. If it works, namespaces should be addable/removable at will. Until domas kills us all, of course. --Splarka (rant) 08:46, 29 January 2009 (UTC)

:::The problem with that implementation is that it means that no cite errors are visible outside the mainspace, ever. It is not obvious that we want cite to fail silently in all cases as opposed to just the specific case of a missing . Dragons flight (talk) 08:55, 29 January 2009 (UTC)

::::If you just want to hide missing perhaps:

::::

{{#switch:{{NAMESPACE}}|=|#default=}}Cite error: $1

::::And in site-wide CSS: ol.references .cite-error-hidden {display:inline !important}. Though this is a dirty hack and accessability regression. --Splarka (rant) 09:16, 29 January 2009 (UTC)

So what is the current state? I am not sure I like per-namespace switching too much, as article drafts can be found in many namespaces (especially Talk and User), and it would be nice to have a meaningful error message there instead of a silent failure (or, equally bad, a hidden error message). Kusma (talk) 09:25, 29 January 2009 (UTC)

::I looked at this yesterday, but go interrupted by real life. I looked at the cite.php code; the message is controlled by MediaWiki:Cite error refs without references. I was thinking of a namespace switch as well, but that has problems. I do like the idea of an error for this, as adding refs without using {{tl|reflist}} is a problem. Perhaps we could disable it only on talk pages. It would be possible to have the message insert the missing {{tl|reflist}}, but that is probably not the best way to do this. Another way would be to have a bot that checks :Category:Pages with incorrect ref formatting and converts the ref tags on talk pages. --—— Gadget850 (Ed) ^talk - 14:32, 29 January 2009 (UTC)

:::Looking at this today, MediaWiki:Cite error now does a namespace switch; the message no longer shows on talk pages. MediaWiki:Cite error refs without references has been blanked, so articles now show a cryptic "Cite Error:" message and are placed in :Category:Wikipedia pages with broken references.

:::The error Cite error refs without references is generated if:

:::* There are {{tag|ref}} tags but no <references/> tag (usually generated with {{tl|reflist}})

:::* The reference immediately before <references/> does not have a closing </ref>

:::* There are {{tag|ref}} tags after <references/>

:::--—— Gadget850 (Ed) ^talk - 15:06, 29 January 2009 (UTC)

::::I've submitted [https://bugzilla.wikimedia.org/show_bug.cgi?id=16854 a patch to bugzilla] which would automatically add a footnotes sections at the bottom of any pages that don't have <reference/> tags, but the developers don't appear to like the idea. — Blue-Haired Lawyer 17:24, 29 January 2009 (UTC)

The message generated by MediaWiki:Cite error refs without references was removed. Thus, articles without a <references/> tag now show a cryptic Cite error: message. See "[http://en.wikipedia.org/w/index.php?title=Breaking_Free&oldid=267296207|Breaking Free]" for an example. I propose to

• Restore the former message of "Cite error: tags exist, but no tag was found".
• Create Wikipedia:Cite errors to present the current errors and how to resolve them; the talk page will give us a centralized place to discuss improvements to the cite error system.

--—— Gadget850 (Ed) ^talk - 12:44, 30 January 2009 (UTC)

Ok, I have done a comprehensive overhaul of the citation error messages. Firstly, I have made MediaWiki:Cite error transparent: all it does is pass the error message through, and also add {{tlx|broken ref}} template, which in turn adds the main tracking category Pages with incorrect ref formatting to all pages outside the Talk:, User: and MediaWiki: namespaces. This means that each error message can be handled separately. As such, I've copied the "Cite error: " part onto the front of each message, and moved the namespace switch MZMcBride added to MediaWiki:Cite error to MediaWiki:Cite error refs without references, the one which started all this off. This means that this message only is hidden outside the mainspace. All pages with this specific error are also categorised into another tracking category MZ created, Wikipedia pages with broken references. While I think this extra category is useful, it should probably be renamed to be more specific. Happy‑melon 15:36, 31 January 2009 (UTC)

:I created Help:Cite errors to explain the sometimes cryptic messages. We can use the talk page to discuss issues and improvements. --—— Gadget850 (Ed) ^talk - 14:48, 3 February 2009 (UTC)

Recently, when viewing a page's History, next to the RSS and Atom text are orange icons that represent web feeds. I personally find them distracting; if anyone else does too, then you can simply remove them by editing your monobook.css page and adding the following code to it on a new line:

a.feedlink { background: none; padding-left: 0px; }

Cheers! Gary King (talk) 19:26, 29 January 2009 (UTC)

:Thanks, works great. Just curious, where (if at all) in mediawiki space were the icons recently added? Garion96 (talk) 18:30, 31 January 2009 (UTC)

::They were [http://www.mediawiki.org/wiki/Special:Code/MediaWiki/46058 added] to the MediaWiki software in r46058. Algebraist 18:38, 31 January 2009 (UTC)

:Thanks! -- lucasbfr ^talk 16:21, 3 February 2009 (UTC)

{{Resolved}}

Is it possible for a person who uses the same template to choose whether the template is collapse or uncollapsed? :Help:Collapsing doesn't help.

For example: {{WP:ARS/Tagged}} is not collapsed by default.

Is there some way to let editors choose to make the template collapsed? Without changing the original template? Ikip (talk) 15:28, 1 February 2009 (UTC)

: Not at the time. However, if you don't know already, if there are two or more collapsible boxes on the same page (and that use the class {{bcode|autocollapse}}), then they are all automatically collapsed by default. Gary King (talk) 17:31, 1 February 2009 (UTC)

::{{thank you}} so much gary. Ikip (talk) 06:14, 4 February 2009 (UTC)

{{resolved}}

Hi there

I've got an article (Kala (album)) with lots of refs which use {{tl|cite web}} and within that use the accessdaymonth and accessyear fields. However, currently none of the accessdaymonth values are displaying. Is there an issue with the template....? -- ChrisTheDude (talk) 09:57, 3 February 2009 (UTC)

:I think |accessdate= replaced it but I'm not 100%. Bidgee (talk) 09:59, 3 February 2009 (UTC)

::I could switch it to accessdate, no problem, but how do you set the date format so that it displays in UK style (day month) rather than US style (month day).....? -- ChrisTheDude (talk) 10:53, 3 February 2009 (UTC)

:::Looks like you figured out the dateformat parameter. --—— Gadget850 (Ed) ^talk - 11:37, 3 February 2009 (UTC)

::::Yup, sussed it out in the end. :-) -- ChrisTheDude (talk) 11:50, 3 February 2009 (UTC)

Yesterday while reading the laptop article I kept seeing the phrase "So I Herd You Liek Mudkipz" appearing in the text but could not find it in the edit summary. Eventually I figured out that it was coming from :Template:convert so that {{convert|0.7|-|15.|in}} looked like this: 0.7–15. So I Herd You Liek Mudkipz (18–380 mm). Someone has fixed it and the template works normally now.

Looking at the convert template itself shows no changes, but I see this is a massively transcluding bit of code with thousands of related pages (:Category:Subtemplates_of_Template_Convert) ... so just for academic understanding, how would I see what was changed in that web of thousands of subtemplates? -- DMahalko (talk) 20:29, 3 February 2009 (UTC)

:Looking at the history of what's transcluded when you do {{convert|0.7|-|15.|in}}, the mudkipz came in through [http://en.wikipedia.org/w/index.php?title=Template:Convert/-/AoffSoff&diff=268095479&oldid=247737978 this edit]. --NE2 20:33, 3 February 2009 (UTC)

::I found it another way: At Template:convert, click "Related changes", select the Template namespace, and Template:Convert/-/AoffSoff‎ is the only recent. PrimeHunter (talk) 20:40, 3 February 2009 (UTC)

Sorry to come up with this again. The enwiki database dump from October has now, finally, either failed or been terminated [http://download.wikimedia.org/enwiki/20081008/]. (Thanks to whoever pulled the plug.) However, enwiki is now at the end of dump queue [http://download.wikimedia.org/backup-index.html] and the next dump will start only in 3 weeks or so.

Would it be possible to move enwiki to the top of the queue?

The reason is that I'm running a widely used bot which depends on database dumps, and users understandably keep asking for data updates on my talk page (since 2 months or so). I'm not interested in page content, just in the metadata tables, and these need just a few days for dumping. So starting the dump now would enable me to provide fresh data within a few days, which would improve the situation.

Thanks a lot. --B. Wolterding (talk) 18:46, 31 January 2009 (UTC)

:Brion killed it on 29 January. You'd have to poke him about getting a new one prioritised. Happy‑melon 18:52, 31 January 2009 (UTC)

Would a dump of article content be quicker than a dump that includes article history? Lightmouse (talk) 03:39, 2 February 2009 (UTC)

: Yes; in fact, the current dump procedures first dumps metadata, then content, then the history, and so the parts I'm interested in are usually finished within just a few days. --B. Wolterding (talk) 21:37, 3 February 2009 (UTC)

I am sure there are many tasks that can operate on content only. Is it possible to split the lengthy dump process into shorter ones e.g. content only? Lightmouse (talk) 12:37, 4 February 2009 (UTC)

The Article Rescue Squad has :Category:Articles that have been proposed for deletion but that may concern encyclopedic topics which simply doesn't make for good cocktail conversating; we're thinking of renaming to :Categories:Articles tagged for deletion and rescue. The articles in the cat are all pulled from our {{tl|rescue}} tag but looking at the cat page other stuff links there as well. What steps should we look to so as to not screw up everything? -- Banjeboi 09:43, 2 February 2009 (UTC)

:Should we just do this ourselves then? -- Banjeboi 13:08, 4 February 2009 (UTC)

:: What's you concern? There are currently 17 pages in the category, and {{tlx|rescue}} is [http://en.wikipedia.org/w/index.php?title=Special:WhatLinksHere/Template:Rescue&hideredirs=1&hidelinks=1 transcluded on 17 pages] as well. You can't "screw up everything" with the rename, you'll have to do a null edit on every page that's still left in the old category once you've renamed the category and changed the template accordingly, if you don't want to wait. I don't know of the internal workings, but I'd recommend first changing the template, and then renaming the category. That might spare you the null edits. --Amalthea 13:50, 4 February 2009 (UTC)

:::The template populates the current articles but I guess I'm more concerned with [http://en.wikipedia.org/w/index.php?title=Special:WhatLinksHere/Category:Articles_that_have_been_proposed_for_deletion_but_that_may_concern_encyclopedic_topics&limit=250 Pages that link to "Category:Articles that have been proposed for deletion but that may concern encyclopedic topics"]. We also have a bot that pulls from the category. o if we create the new category page, adjust the template and then redirect the old cat page to the new one would that cover things or are their other steps - or do we just go for it and solve problems that arise? -- Banjeboi 15:19, 4 February 2009 (UTC)

::::I think all or most of those links are from the ARS project banner and userbox. Change the link in those templates and see what's left. Algebraist 15:27, 4 February 2009 (UTC)

::::: Most, yes, some like User:GlassCobra link to it manually. But of course, Benjiboi, if there is a bot using the content of the category you should talk to the bot owner first, to make sure he can timely reconfigure it to use the new category. A redirect won't affect it, in all likelihood. --Amalthea 16:05, 4 February 2009 (UTC)

::::::OK, coolio, will go sort it out with them and start the changeover. Thanks all for the help. -- Banjeboi 16:06, 4 February 2009 (UTC)

I'm trying to learn how to make a box with all sorts of infor and stuff (the type that has a picture, influnces, influenced etc.) on my own user page. I know it's dumb, but at least i'm not messing with a real page right? Anyway it says I don't have a template and I'm frustrated. Please help.

--CrimsonKing2000 (talk) 04:01, 3 February 2009 (UTC)

:Find a template that's close and copy its source to your user page. Be sure to remove any categories though. davidwr/(talk)/(contribs)/(e-mail) 04:30, 3 February 2009 (UTC)

:See Help:Infobox, Wikipedia:WikiProject Biography/Infoboxes, :Category:People infobox templates. If you want to use an existing infobox type then you are restricted to the parameters it already has. PrimeHunter (talk) 12:25, 3 February 2009 (UTC)

::I suspect you are looking for {{tl|Infobox user}}. You may also be interested in Wikipedia:Userboxes. --—— Gadget850 (Ed) ^talk - 15:46, 4 February 2009 (UTC)

In [http://en.wikipedia.org/w/index.php?title=User_talk:67.142.130.25&diff=prev&oldid=268189120 this edit] I manually placed a warning and also copy/paste-moved Piano non troppo's earlier warning into the February section. If you look closely, you can see that at least one space disappears in the move ("Doing so violates Wikipedia's Wikipedia:Neutral point of view" changes to "Doing so violates Wikipedia'sWikipedia:Neutral point of view"). What could cause this type of error? Is it an OS, browser, or Wiki server issue? NJGW (talk) 04:50, 3 February 2009 (UTC)

:Probably OS; possibly the space was on a linebreak in the edit box and failed to be copied. — Edokter • Talk • 02:16, 5 February 2009 (UTC)

{{Resolved}}

Way back in 2005, User:Tim Starling [http://en.wikipedia.org/w/index.php?title=MediaWiki%3ABad_image_list&diff=17380299&oldid=12175551 added] a note to MediaWiki:Bad image list: "For performance reasons, please try to keep this page fairly short, say less than 10 KB. ". We are now at 12 kb and I'm wondering if the performance concern still exists. Anybody have an idea? - BanyanTree 11:14, 4 February 2009 (UTC)

: It still exists, but people appear to have decided it's an acceptable tradeoff in order to not see those images. (Personally I think the list is too long, but then I think any blacklist on a website with user-generated content is doomed to fail, since no matter how big you make it, any even moderately determined user can find an unlisted image) -- Gurch (talk) 12:58, 4 February 2009 (UTC)

:I asked Domas about this. He said the performance impact is negligible. --MZMcBride (talk) 13:08, 4 February 2009 (UTC)

The answer is not the obvious one! This needs a bit of deep digging and inspiration, please. It may not be possible, but surely it is! No-one will die if it can;t be done, but doing it would make life much easier.

I've been doing some work formatting List of Jat clans for legibility by migrating it from one very long list into currently 4 columns. I've noticed that a huge number of the names are not in alpha order, and suspect this is not wilful, but may be a language issue with the editors who use the page a lot and unfamiliarity with the Americish alphabet's sort order.

Is there a simple way of creating a sort "thing" for each of the sections to make life easier for contributors and readers alike, please.

Please look at the article before answering. Yeah, obvious I know, but the right answer does not seem to be a simple sortable table.

What I hope to achieve is a list that is sortable, and which automatically goes into (eg) four columns, and is easy for a novice or non skilled editor to add new members to. Exporting to excel, sorting and pasting back is not the right answer, especially singe some names are wikilinked and others are not.

Using helpme on my talk page has led me here. The guys who came running to my aid were perplexed too! See there for their thoughts. Fiddle Faddle (talk) 19:47, 4 February 2009 (UTC)

:It's unlikely that you are going to find a solution that works for all users and complies with Wikipedia:Accessibility. Even if it did, it would not be easy for unskilled editors to use. I'd recommend keeping the page as simple as possible and just clean it up when needed. --- RockMFR 21:55, 4 February 2009 (UTC)

::It was unlikely for the Wright Brothers to fly, too. But they did. Let's see, please, what is recommended. Fiddle Faddle (talk) 22:07, 4 February 2009 (UTC)

A sortable table would only be useful when there is more than one piece of information. Here you have only the name of each clan, so there is nothing else to sort by, which defeats the purpose of having buttons to automatically sort. If you just want to be able to add items in any old order when the page is saved but have them automatically sort correctly when the page loads (without having to push a button), that's a separate matter. But really it would be better to be able to click sort before saving the page. — CharlotteWebb 02:54, 5 February 2009 (UTC)

:Exactly. The problem seems to be an unwillingness to sort, or possibly an ignorance of sort order by the editors who use the page when they add a name. I know it isn't the classic use of a sortable table, but that would work if there were only a long single column, but sorting and making the page look presentable is the challenge. Fiddle Faddle (talk) 07:21, 5 February 2009 (UTC)

::Sort-on-view would only work if the javascript enabling it can correctly guess what to sort and how to sort it. This is going to be different depending on the page. Here you have one row with four cells each containing a bullet list, and your desired result is to extract the four lists, combine them into one, sort it alphabetically, and divide it back into four parts, and put these parts back in the table cells. The machine needs some way of knowing it isn't supposed to just sort the rows (doing nothing as there is only one row), or sort the columns (shifting horizontally as needed), or sort the items within each list (independently of the other lists).

::I don't think we're ever going to want separate javascript code for each page, but instructions for new types of sortable tables can be added to [http://en.wikipedia.org/skins-1.5/common/wikibits.js wikibits.js] if the various styles are reasonably well-defined and few in number. You would need to name them and specify in the wikitext which one you wanted.

::But in this case (where there is only one "right way" to sort this), your best bet would be to run a bot to clean up after edits which scupper up the alphabetical order by adding items in the wrong place.

::Ideally all browsers would support displaying one list across several columns (sorting tools would more easily know what to do), so we wouldn't need to manually chop up the list and estimate the vertical height of each sub-list. In the future I'd like to see some automatic way to fake the column-count attribute for user-agents known not to recognize it. — CharlotteWebb 15:45, 5 February 2009 (UTC)

:::Perhaps this article with its really simple lists could be made to be the causative agent for your "future way"?

:::I don't have any skill to create a bot I'm afraid. Fiddle Faddle (talk) 16:06, 5 February 2009 (UTC)

{{resolved}}

Bryn-crug, which was created automatically when I moved the page to Bryncrug, does not seem to be redirecting properly. There's no double redirecting involved, but I can't for the life of me see any problems with it. I've purged the page and my browser without success. Does anyone else have this problem? —  Tivedshambo  (t/c) 07:40, 5 February 2009 (UTC)

:What exactly is wrong? It seems fine to me. This, that and the other _[talk] 09:38, 5 February 2009 (UTC)

::User:Happy-melon "fixed" it by removing the space before the link, though as far as I'm aware it should have worked beforehand (especially as the page was automatically produced). Perhaps it was a problem with some database lag? —  Tivedshambo  (t/c) 09:55, 5 February 2009 (UTC)

There is a problem with the software: new code for redirect handling (to potentially allow double redirects to function) was introduced in one revision, and some bugs in it were fixed about thirty revisions later. On 28 Jan the Wikimedia wikis were updated to a version of MediaWiki between those two versions, so it included the changes but not the bugfixes. The problem appears to have been fixed, although I'm not sure exactly how (maybe the bugfixes were live-merged, as we're still running the 'broken' version), but there are still quite a few broken redirects in the database, which will remain until they are cleared out by hand. Happy‑melon 11:38, 5 February 2009 (UTC)

:The broken redirects can be fixed by a dummy edit, but not by a null edit, hence the removal of the space. Algebraist 13:08, 5 February 2009 (UTC)

:: Melon: "to potentially allow double redirects"? Yay, but the "potentially" leaves me puzzled; do you have a bug number for that (I can't find it)? --Amalthea 13:35, 5 February 2009 (UTC)

:::It's {{phab|13644}} - the code added in {{rev|45973}} makes the maximum length of redirect chains a definable config setting, allowing 0 (don't allow any hard redirects, treat all redirect pages as if they were being accessed with &redirect=no), 1 (default, current situation), or more than 1 to allow redirect chains (with some really rather cool [https://bugzilla.wikimedia.org/attachment.cgi?id=5715 new UI]). AFAIK there is no bug open at the moment to enable this anywhere on wikimedia. Happy‑melon 14:07, 5 February 2009 (UTC)

:::: Ah, thanks. I'd like it if that were upped to 2 or 3 here, that would help managing the redirects left from articles created for singles prematurely. With that feature, one could redirect
misspelled song → song → album → artist
which will just work. Currently, they will all have to point to the artist article directly, and with each article that is eventually created they all have to be retargeted. --Amalthea 15:21, 5 February 2009 (UTC)

:::Just for completeness, the code was introduced in {{rev|45973}}, fixes were implemented in {{rev|46502}} (actually 530 revisions later, not 30!) and then Wikipedia was scapped to {{rev|46424}}. Happy‑melon 16:19, 5 February 2009 (UTC)

There are quite a few articles which have had the {{tl|Dyktalk}} template subst'd onto the page, see [http://en.wikipedia.org/w/index.php?title=Special:WhatLinksHere/File:Updated_DYK_query.png&limit=5000&namespace=1 here] for a long list of over 5000. Should these occurences be fixed to use the template? -- WOSlinker (talk) 08:16, 5 February 2009 (UTC)

:If you're going through the trouble of fixing these you might as well convert them all to use {{tl|ArticleHistory}} and catch any subsequent FACs, AFDs, etc. — CharlotteWebb 15:50, 5 February 2009 (UTC)

Is this possible? What I mean by this is that, instead of getting the complete html code for a page, is it possible to get only the source of an article? This would save an awful lot of data transfer, if the only interesting is the source. For example, if I would like to create a web crawler only focusing on Wikipedia articles, everything but the source is quite uninteresting. This would also save some traffic for Wikipedia's part. --Kri (talk) 17:03, 5 February 2009 (UTC)

:The {{edit|2=Edit this page}} button serves that purpose. Dendodge Talk^Contribs 17:05, 5 February 2009 (UTC)

:To scrape the wiki-text and nothing else you would want to use [{{fullurl:Felix Pedro|action=raw}} action=raw]. To get several articles you can use the API (separate the titles by "|", e.g. [http://en.wikipedia.org/w/api.php?action=query&prop=revisions&titles=Felix_Pedro|Serge_Zwikker|World_Plus&rvprop=content]). If you want more than, say, a few thousand articles you should download a database dump. — CharlotteWebb 17:30, 5 February 2009 (UTC)

::Thank you! --Kri (talk) 17:33, 5 February 2009 (UTC)

:::If you do use the API you will want to add "&format=xml" when actually crawling, to get rid of the human-friendly formatting like colors and auto-linked urls. You will need to unescape things like &quot;, &amp;, &lt;, &gt;, etc. after you isolate the wiki-text from the tags. — CharlotteWebb 17:40, 5 February 2009 (UTC)

:::: See also mw:API for more on that. --Amalthea 17:59, 5 February 2009 (UTC)

It does work. I decided to try it when I realized people would be confused by just a redirect to the article.

It is correct under Wikipedia:Redirect#Redirects to page sections

If I could remember where I was this morning, it's not correct under there.Vchimpanzee · talk · contributions · 19:08, 5 February 2009 (UTC)

Okay, I found it.

Help:Link#Redirects with section linksVchimpanzee · talk · contributions · 19:21, 5 February 2009 (UTC)

While reading the redirect page, I discovered a problem I ran into yesterday. WSGA (defunct) has as its first edit a change to a redirect. But it's not exactly a change if there was no history before that. The second edit was a change "back" to an article. Since there's no history, I don't know who contributed to that article.

I've seen articles get moved or redirected and have no apparent history, and someone who really knew what to do was able to find what had happened.Vchimpanzee · talk · contributions · 19:12, 5 February 2009 (UTC)

: That first redirect was automatically created at WSGA (AM) when WSGA (AM) was moved to WSGA. It was expaneded into an article in November 07, and in November 08 that article was moved to WSGA (defunct).
WSGA was effectively untouched by this, it just lived for a couple of minutes at WSGA (AM). --Amalthea 19:29, 5 February 2009 (UTC)

There is no history for WSGA (AM).

I am merging WSGA (defunct) into WSEG and need to know whose contributions are to be credited.Vchimpanzee · talk · contributions · 20:36, 5 February 2009 (UTC)

Wait, I only looked at WSGA-FM for a history. I thought I was looking at WSGA but I see that while it is a disambiguation page now, it wasn't then.Vchimpanzee · talk · contributions · 20:39, 5 February 2009 (UTC)

{{tl|Colbegin}} appears to be broken, I can't get it to work now, and many of the pages which use Template:Colbegin no longer are in columns. Any ideas why? Ikip (talk) 20:08, 5 February 2009 (UTC)

:Works for me. What browser are you using? --- RockMFR 20:35, 5 February 2009 (UTC)

::Internet explorer 6.0. Can you give me an example page were it is working?

::It isn't working on Alphabet#See_also or Weaving#See_also for example. Ikip (talk) 22:01, 5 February 2009 (UTC)

:::IE doesn't support multiple columns. Algebraist 22:10, 5 February 2009 (UTC)

:::IE doesn't do a lot of things that modern web browsers can do. I highly recommend you upgrade your browser (to Safari or Firefox, preferably, but even IE7 is better). You'll enjoy the web much more that way. :) EVula // talk // ☯ // 22:17, 5 February 2009 (UTC)

::::It's definitely an IE issue. I switched between FF3 and IE7 with the IETab plugin, and it changes the behavior so that it's not in columns under IE. ♫ Melodia Chaconne ♫ (talk) 23:18, 5 February 2009 (UTC)

:::::{{tl|Colbegin}} uses CSS3 columns, which are not supported by all browsers. IE—including IE8—does not support the column selector. Neither does Opera. Thus, this template works for the ~30% who use FireFox or Safari. I added details on browser support to the template doc. --—— Gadget850 (Ed) ^talk - 23:32, 5 February 2009 (UTC)

{{resolved}}

I have tracked down a possible problem when users transclude Template:Introduction to Wikipedia to their user page, they then appear in :Category:Wikipedia basic information e.g. User:Djreload, There is a line in the template which confuses me but may be the culprit:

'{{#ifeq:{{NAMESPACE}}|{{ns:Project}}|{{PAGENAME}}}}'

Any suggestions? LeeVJ (talk) 23:22, 5 February 2009 (UTC)

:The line works correctly but the namespace test was only added a week ago in [http://en.wikipedia.org/w/index.php?title=Template:Introduction_to_Wikipedia&diff=267116067&oldid=263821883]. Affected pages can be removed from the category with a null edit. Purging is not enough. If the job queue works as intended then the pages should eventually be removed automatically but that sort of job has sometimes taken weeks recently. PrimeHunter (talk) 01:17, 6 February 2009 (UTC)

::Ah, that explains it , many thanks! LeeVJ (talk) 01:47, 6 February 2009 (UTC)

See MediaWiki talk:Common.css#Proposal: shifting italics text slightly to the left. —Preceding unsigned comment added by Army1987 (talk • contribs) 20:55, 25 January 2009 (UTC)

The magic word {{PROTECTIONLEVEL}} is now available (see {{phab|11947}}, {{rev|45587}}). So we can now update the protection templates to do this:

• when a page is not to the correct protection level, return nothing and categorize in Wikipedia pages with incorrect protection templates.

Then bots can patrol this category to remove the incorrect protection template or replace it with the appropriate one. It deprecates the expiry option in templates and the following categories: :Category:Protected pages with expiry expired, :Category:Wikipedia protected pages without expiry.

We can update directly at {{t1|pp-meta}} for the types semi, full and move and then each template individually.

When done, we may inform administrators of this in MediaWiki:Protect-text (essentially; unneeded to add expiries). Cenarium (Talk) 09:33, 29 January 2009 (UTC)

:Even more helpfully, we could in principle add the lock icons to pages automatically, if we put some code in a system message that is included on all pages. Would there be support for this? Happy‑melon 14:03, 29 January 2009 (UTC)

::Not sure, as there are different types of protection templates, and sometimes they return nothing (like {{t1|pp-move-indef}}). Except if you can enter the type of template you want in the protection log. But will it also add the categories ? Cenarium (Talk) 14:42, 29 January 2009 (UTC)

:::I disagree with the principle of having protected pages that are not marked as such with the lock icons; it's only because the protection tags are inconsistently applied that we can get away with it. Here we now have the opportunity to automatically mark pages as protected (with the lock icon, I mean) so that none of the templates need to add the icon. We still need the templates, since there can be no granularity in the messages applied universally. But it means we only need to add the templates to act as descriptors of why the page was protected, not just to indicate that it is so. So for instance, we will no longer need to add {{tlx|pp-template}} to all protected templates, as a category insertion to :Category:Protected templates based on protection level and namespace is possible. We would still need to add {{tlx|pp-semi-sock}}, for instance, because it's not possible to deduce that information from the protection state and namespace. But that template would no longer need to add the lock icon. Happy‑melon 16:55, 29 January 2009 (UTC)

::::For edit protected pages yes, but for a lot of move-protected pages, it is unnecessary to add the icon and it confuses readers and inexperienced users. (As I suggested earlier, keeping the move tab for move protected pages but allowing to disable it in the preferences would help inexperienced users to find out why the move tab sometimes disappears, the icon is inefficient as it is confounded with a semi-protection icon or missed.) And sometimes, we want the big notice, not just the icon, so it doesn't appear feasible except for templates and images. Admins should also be encouraged to use specific templates when possible. Cenarium (Talk) 17:36, 29 January 2009 (UTC)

:I was just thinking about starting this discussion. :) Here are some points:

:*I like the idea of automatic replacement of notices with tracker categories: we can get a bot to do the cleanup, or manually, or someone with AWB can sweep through every now and then.

:*I don't like the idea of automatic addition of protection icons, unless that automatic addition can be modified, removed, or replaced using a protection template. Our current system of templates allows, for the most part, the reason for protection to be reasonably clear, even on the small, icon versions: the tooltips give the expiration date if applicable and an explanation for the protection.

:*I remain neutral on the issue of indefinitely-move-protected pages: while I do think that all protection should be visible (invisible protection does confuse people), it's also true that it gets confused with semi-protection. I think we need to explore this further, but not as part of this discussion.

:*We might use this new magic word to merge a number of the generic templates: automatic selection of types semi or full could be easily implemented to simplify the template family. Whether we implement that functionality on the templates to be simplified or on {{pp-meta}} is debatable.

:*This function could add notices for pages which are both move-protected and semi-protected, which, as far as I know, currently don't mention their move-protection.

:{{Nihiltres|talk|[http://en.wikipedia.org/wiki/Special:Log?user=Nihiltres log]}} 18:04, 29 January 2009 (UTC)

YES. I hadn't thought of that, Nihiltres, but of course that's exactly what we should be doing. We still use templates, and we still have different ones for different reasons for protection, but we unify eg {{tlx|pp-vandalism}} and {{tlx|pp-semi-vandalism}} and {{tlx|pp-move-vandalism}} and let the template work out what's appropriate based on the actual protection that's on the page. We can even make the notices vanish automagically when the protection expires, and just leave a tracking category to clean out with the bots. I'm still inclined to suggest deprecating some of the pp- templates, {{tlx|pp-template}} in particular (or maybe call that by default from {{tlx|documentation}} and have it be silent when there's no protection in place), but I think this could be the best immediate use of these new functions. Happy‑melon 18:55, 29 January 2009 (UTC)

:Using a combination of namespace and protection type/level checks, the templates listed in Template:Protection templates (not counting the WP:OFFICE templates) could probably be reduced to:

:*pp-dispute

:*pp-vandalism

:*pp-sock

:*pp-indef

:*pp-protected

:Using namespace checks means the template and user-talk (which with the ability to change blocks without unblocking should be mostly deprecated now anyway) templates can be merged into the generic pp-protected, or as Happy-melon suggested, integrate the pp-template template into the documentation templates. The separate versions for move/semi/full protection shouldn't be necessary. Mr.Z-man 20:03, 29 January 2009 (UTC)

::Some more points, moving things along:

::*Merging {{pp-usertalk}} (and its semi equivalent) into the generic protected template probably isn't a good idea, since it's a specific rationale for a general template. Unless we can assert that all protections in that namespace fall under that rationale, I don't think it should be merged. Formally deprecating it, on the other hand, sounds like a good idea.

::*{{pp-template}} will have to stay on its own in some respect, since [http://en.wikipedia.org/w/index.php?title=Special%3AWhatLinksHere&target=Template%3APp-template&namespace=6 it is used for high-risk files] as well as templates, though getting {{documentation}} to call it automatically ({{#if:{{PROTECTIONLEVEL:edit}}|{{pp-template|small=yes}}}} or something) would be an improvement.

::*Do we want to use error messages for inappropriate uses of protection templates, that is, uses outside their generally-accepted scope? For example, {{pp-dispute}} should not be used on semi-protected pages, and {{pp-indef}} on full-protected ones.

::*Do we want to centralize the changes, or do you think that merging individual templates is the best plan? At present, we have protection templates along two "axes", as it were: protection reason, and protection type. {{pp-meta}} currently has a type-centred architecture. Should we be making that architecture more general? (I'm slightly against it, but there might be something interesting here.)

::Aside from that, I think we have a plan. {{Nihiltres|talk|[http://en.wikipedia.org/wiki/Special:Log?user=Nihiltres log]}} 23:20, 29 January 2009 (UTC)

:::I'm not sure we could merge protection templates, at least not move with edit: as the reasons for edit and move protection often differ (e.g. indefinitely move protected articles temporarily semi-protected against vandalism). And for example when an article is semi-protected due to vandalism, and then get fully protected due to disputes, it would give a wrong reason (and it would have to be updated anyway). I think the only case where semi and full could be merged would be pp-protected, pp-usertalk and pp-template. Some project pages are indefinitely fully protected (for example Wikipedia:Copyrights, WP:ITAAW), so we can make a pp-indef for Wikipedia namespace only, and articles are also fully protected due to sockpuppets, so pp-sock could also be created. Actually, we could use only one template, with edit= and move= parameters (e.g. edit=vandalism, move=indef), but we couldn't redirect old ones without breaking them. As for protected titles, I've updated MediaWiki:Titleprotectedwarning with {{PROTECTIONLEVEL:create}}, I suppose we could do the same with {{t1|pp-create}}. Cenarium (Talk) 04:17, 31 January 2009 (UTC)

::::The rough way to get rid of misplaced protection templates would be to use {{#ifeq:{{PROTECTIONLEVEL:edit}}|sysop|