Wikipedia:Village pump (technical)/Archive 102

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


Feedback tool

Why do some articles link to the feedback tool (eg, Clifton Gardens, New South Wales, List of rice diseases) and some ask for page rating (eg, Mazeppa (opera), Quinoline-4-carboxylate 2-oxidoreductase)? SpinningSpark 18:11, 7 August 2012 (UTC)

Articles that link to the feedback tool are either in Category:Article Feedback 5 or Category:Article Feedback 5 Additional Articles. You can add the 2nd category to any article you wish. Ryan Vesey 18:16, 7 August 2012 (UTC)
Spinningspark: didi you follow "What's this?" links to WP:Article Feedback Tool (which says "deployed on almost all pages") and WP:Article feedback (which says "only about 5 percent") and found that the explanations are not clear enough? — AlexSm 18:22, 7 August 2012 (UTC)
Since I would probably not have posted here if I had that information, then no it is not clear enough and your rhetorical question seems only to have been asked in order post the sarcasm which followed. And yes, I did follow the "What's this?" link first - it goes here and does not contain any of the information you claim. SpinningSpark 19:34, 7 August 2012 (UTC)
  • Hey guys :). So, at the moment the deciding factor is a randomised lottery; we've deployed the new version (with the feedback box) to 10 percent of articles - the rest have the old version with the page rating. We're waiting on features engineering to fix up the code a bit; they worry that if we deploy the new one to 100 percent now, bits are going to fall off the servers and explode. Okeyes (WMF) (talk) 17:53, 8 August 2012 (UTC)

Creating a map

I'm sorry, I've just joined, but would like to know how I can create map illustrations. It seems there is a standard map used to create an svg file. Where is this map and what is the recommended software? Thank you. — Preceding unsigned comment added by Kneesdata (talkcontribs) 07:52, 8 August 2012 (UTC)

Wikipedia:WikiProject Maps is a good place to start. Secretlondon (talk) 08:04, 8 August 2012 (UTC)
It obviously depends on the type of map. What sort of map do you want to make? Secretlondon (talk) 08:05, 8 August 2012 (UTC)
And Inkscape for svg maps. Wikipedia:Graphic Lab/Resources/Software, but you can base them of things like free GIS data using QGis. Secretlondon (talk) 08:07, 8 August 2012 (UTC)
Take a look at WP:USRD/MTF - it's geared towards U.S. roads, but there are some good tutorials. --Rschen7754 08:09, 8 August 2012 (UTC)

Thanks for your advice. I would like to make world and continental maps with colour codes for certain data sets, eg colouring countries with <0% growth differently from those with 0-5% and 5%+. To be honest, Wikipedia is a bit of a mystery when it comes to this sort of thing. I've occasionally edited without an account, but would like to enhance the economy sections which are often outdated. — Preceding unsigned comment added by Kneesdata (talkcontribs) 09:42, 8 August 2012 (UTC)

We have maps that already have country outlines. Filling them with colours should be pretty easy. Obviously you'll need a source for your economic data. PS You sign your comments ~~~~ Secretlondon (talk) 09:51, 8 August 2012 (UTC)
This might be what you are looking for. Secretlondon (talk) 09:55, 8 August 2012 (UTC)
There's actually a tool that's really helpful with this, but I can't find it at the moment. Perhaps someone else will know? - Jarry1250 [Deliberation needed] 11:09, 8 August 2012 (UTC)
commons:Category:Blank SVG maps and commons:Category:Blank maps contain more blank maps. (Kneesdata: Wikimedia Commons is the image repository of Wikipedia, which allows a single file to be used in all Wikipedias and Wikimedias. MKFI (talk) 13:19, 8 August 2012 (UTC)


"Preferences/Gadgets/Appearance/Add page and user options to drop-down menus on the toolbar" doesn't seem to be working properly for me today. I wanted to protect a page, but the option to do so wouldn't appear on the drop-down, only the option to unprotect the page (it was already unprotected). The behavior was inconsistent - some pages it would give me the option, and some it wouldn't.

I was trying unsuccessfully to protect List of Hollyoaks characters (2010). I had to uncheck the check box and use the option "Change protection" on the unnamed drop-down arrow to execute the action.

What's wrong?--Bbb23 (talk) 00:25, 9 August 2012 (UTC)

Whether it says "unprotect" or "protect" it takes you to the same place, and doesn't automatically put the action through. It was saying "unprotect" because that page was move protected. But that does mean the gadget could use a better name for that option. Someguy1221 (talk) 00:35, 9 August 2012 (UTC)
Thanks, Someguy, I appreciate the explanation.--Bbb23 (talk) 00:50, 9 August 2012 (UTC)

Template:WikiProject Public Art

Articles that use Template:WikiProject Public Art on the talk page all show categories which once existed, but have been deleted. Should these categories be created, with a caution, such as I have seen elsewhere on categories that are sometimes empty, that they should not be deleted even when empty? Or possibly, maybe the template is incorrectly done, and the categories should not exist at all. I hesitate to create the categories again, so I am posting here instead. --DThomsen8 (talk) 16:27, 9 August 2012 (UTC)

Have you asked the person who carried out this edit? --Redrose64 (talk) 16:52, 9 August 2012 (UTC)


Anyone have any idea why the MarkBlocked script isn't working for me anymore? I'm using Vector and Firefox 14 and the error console displays no errors when loading a page, but links to blocked accounts aren't marked for some reason. The interesting thing is that the problem seems to be limited to enwiki; the script works fine in other projects, including Commons. Has anyone else run into the same problem? Jafeluv (talk) 19:51, 9 August 2012 (UTC)

The script works fine (I personally make sure of that) but the way you're trying to call it doesn't: mw.loader needs a full URI. Follow your own instructions. — AlexSm 03:12, 10 August 2012 (UTC)
That was actually a test edit, the actual call is in my global JS page and as I said it does work on other projects. But you're right, it's not a problem in the script. After some testing it seems a replace script in my vector.js is causing it... Jafeluv (talk) 06:26, 10 August 2012 (UTC)

New Deployment of Page Curation

A new version of the Page Curation suite has been deployed. This updates Special:NewPagesFeed and enables a new curation toolbar (which is activated when reviewing pages from the feed.--Jorm (WMF) (talk) 21:18, 9 August 2012 (UTC)

Error when deleting files: "inconsistent state within the internal storage backends"

I have run into errors when trying to deleting images File:Mintransporte.png and File:ClaremontDesign.png, which were both nominated at Wikipedia:Possibly unfree files/2012 August 2. The error message looks like this, and it deletes the image description page, but not the actual uploaded image. Any help? — ξxplicit 00:40, 10 August 2012 (UTC)

We are in the process of transitioning to a new media storage backend (Swift). Some files didn't get copied over into the new system, so this error showed up when you tried to change them as the two systems must stay in sync until we git rid of the old one (NFS). The migration bug that left out some files was tracked down and those remaining files are being copied over. These errors should become rarer and then stop within a few days. Aaron Schulz 05:45, 10 August 2012 (UTC)
Oh, I also manually re-deleted those files after the backends were synced for those files. Aaron Schulz 05:46, 10 August 2012 (UTC)

Counting all articles in a category

Is there a tool or method that will count the total number of pages in a particular category including the entire "tree" of subcategories? AWB just gives the number of articles and subcategories in the "root" category. The category I'm interested in is Category: 2012 Summer Olympics. This question arose from the discussion at Talk:2012 Summer Olympics#Article imbalance. Thanks Roger (talk) 12:08, 10 August 2012 (UTC)

Category (recursive) gives the number in all subcats from the tree, not just the toplevel. Fram (talk) 12:19, 10 August 2012 (UTC)
(edit conflict) You can use the PAGESINCAT magic word. For example, {{PAGESINCAT:2012 Summer Olympics}} returns 82. (edit conflict × 2) Nevermind, I forgot that PAGESINCAT wasn't recursive. Reaper Eternal (talk) 12:20, 10 August 2012 (UTC)
That only grabs the top level category. Ryan Vesey 12:20, 10 August 2012 (UTC)
I asked this same question a couple days ago and there didn't seem to be a solution. I was dealing with maintenance categories so I just made a larger category for all the articles. That said, you might be able to use WP:Catscan and set the depth to 2 or 3 however many you need it. It will break if there are too many articles though. Ryan Vesey 12:20, 10 August 2012 (UTC)
Oh yes, for maintenance cats (monthly ones) I usually make an "all" category as well, it is sometimes much more handy than the monthly cats. But for general cats, this isn't really a good solution of course. Fram (talk) 12:32, 10 August 2012 (UTC)
One of the reasons that {{PAGESINCATEGORY:}} doesn't recurse is because category loops are permitted, hence {{PAGESINCATEGORY:}} could be infinite.
Even if each category were counted once each, it's permitted for pages to be in two different branches of the same tree (e.g. Athletics at the 2012 Summer Olympics is in both Category:2012 Summer Olympics eventsCategory:Summer Olympics events by yearCategory:Summer Olympic events; and Category:Athletics at the Summer OlympicsCategory:Summer Olympic events), so the total could well be higher than the real figure. --Redrose64 (talk) 13:59, 10 August 2012 (UTC)
Is there no way to count such an article only once? Going by the title it should be obvious that it is just one article in multiple categories. Roger (talk) 14:19, 10 August 2012 (UTC)
Could you create a list in AWB? That would give you a one-time view. Ryan Vesey 14:20, 10 August 2012 (UTC)
(ec)With the AWB method I described above, you can generate the list of articles with the option "remove duplicates" (in the "list" drop-down menu). This gives you a list of every unique article that is part of that category tree from your starting cat until the bottom cats. If this generates a list of more than +-25000 articles, this may take long (or very long) though. Fram (talk) 14:23, 10 August 2012 (UTC)
Hasn't AWB got a 25,000 cap? He'll only see greater than 25k if he's an Admin or has a bot account. - X201 (talk) 15:26, 10 August 2012 (UTC)
Will a bot account go past the cap? I could have User:Ryan Vesey Bot create that list when I get home. Ryan Vesey 16:02, 10 August 2012 (UTC)
AWB will cap it at 25000 if you just count one category, but if you specify a number or leave it open ended recursive, it will count much more than 25k. Catscan2 can also do it, but it is currently affected by the toolserver replag, so is 411 hours behind real life, which means it probably doesn't even know the Queen can skydive yet. AWB just returned 8676 unique mainspage pages, catscan was only 1841. The-Pope (talk) 16:33, 10 August 2012 (UTC)


Hi, does Geolocate not work for IPv6? GiantSnowman 16:28, 10 August 2012 (UTC)

The contributions links for IPv4 users are made by {{Anontools/ipv4}}. The Geolocate link goes to which apparently doesn't work for IPv6. The contributions links for IPv6 users are made by {{Anontools/ipv6}} which currently has no Geolocate link. I don't know whether there is a service we can use. Wikipedia:WikiProject IPv6 Readiness#Open tasks includes "Find an IPv6 Geolocator". PrimeHunter (talk) 18:43, 10 August 2012 (UTC)
There's also which also doesn't work on IPv6. GiantSnowman 18:50, 10 August 2012 (UTC)

WPBannerMeta categorization

Is it possible to modify the infobox cat parameter of {{WPBannerMeta}} so that it can support more than one category? I hope to contact various wikiprojects so articles tagged as needing infoboxes can be added to a dated category in Category:Articles covered by WikiProject Wikify and the general Category:All articles covered by WikiProject WikifyRyan Vesey 18:34, 10 August 2012 (UTC)

Actually, rather than contacting each project individually, is it possible to just add the category to {{WPBannerMeta}} so that in addition to the default category or the category specified by the project the page is categorized in the ones mentioned above? Ryan Vesey 18:36, 10 August 2012 (UTC)
As thing stand, {{WPBannerMeta}} provides two parameters related to this matter - |infobox= and |INFOBOX_CAT=. The latter is ignored unless the former is set to yes.
  • If |INFOBOX_CAT= is missing or blank, a default category is used (the project name plus articles needing infoboxes)
  • If |INFOBOX_CAT= is set to none there is no categorisation
  • If |INFOBOX_CAT= is set to any other value, that is taken as the category name.
I suspect that much more than {{WPBannerMeta}} will need to be changed - {{WPBannerMeta/core}} and {{WPBannerMeta/templatepage}} for a start. Have you brought this up at Template talk:WPBannerMeta? --Redrose64 (talk) 19:18, 10 August 2012 (UTC)
I wanted to see how to do this before I started creating consensus to do it. Consensus shouldn't be too hard, it's just internal maintenance categorization. Ryan Vesey 19:19, 10 August 2012 (UTC)
Oh, and {{WPBannerMeta/note}} --Redrose64 (talk) 19:21, 10 August 2012 (UTC)
I would see the main issue modifying WPBannerMeta to add to Category:All articles covered by WikiProject Wikify is that WPBannerMeta is used on talk pages but all the current pages in those categories are articles, so should they be mixed together? -- WOSlinker (talk) 20:17, 10 August 2012 (UTC)
There's actually a decent number of talk pages there due to {{Infobox requested}}. I don't know that I'm the biggest fan of it, but the only other option would be to put the infobox requested template on the article page. It could be slightly beneficial, if you see talk: you know the issue with that article is the lack of an infobox. Ryan Vesey 20:20, 10 August 2012 (UTC)

Bypass upload wizard when clicking redlinks

Hi there. I'm looking for a way (preference, script, etc.) to get it so that when I click a redlinked file page it takes me to the blank page in the file namspace (like it used to), rather than to the upload wizard. This would be tremendously useful to me, as getting to that page is a much easier/convienant way to view the file's deletion rational than any of the alternative methods are. Sven Manguard Wha? 17:18, 4 August 2012 (UTC)

Example: If I click on File:47th_Dai_Abdul_Qadir_Najmuddin_.JPG, I want to be taken to this page. Sven Manguard Wha? 17:19, 4 August 2012 (UTC)
I'll second that - it would be quite useful. --Philosopher Let us reason together. 21:51, 4 August 2012 (UTC)

Here's some script to put in your Special:MyPage/skin.js or Special:MyPage/common.js:

$("").filter("[href*='Wikipedia:Upload?wpDestFile=']").each(function(index, el) { el=$(el); el.attr('href', el.attr('href').replace('wiki/Wikipedia:Upload?wpDestFile=', 'w/index.php?action=edit&redlink=1&title=File:')); })

This, that, and the other (talk) 09:52, 5 August 2012 (UTC)

I think it's better to be redirected just to wiki/File:.... It looks better, it shows file usage and if implemented globally (on some project) there is no confusing for newbies edit window. — AlexSm 15:06, 5 August 2012 (UTC)
Multiple editors have complained about this situation for several years now. Wikipedia:Village pump (technical)/Archive 99#File redlink should not lead to godawful upload form is a recent one with some technical discussion. Wikipedia:Village pump (technical)/Archive 77#Redlinked files linking to upload page has a link to an underlying bugzilla item and a .js solution. DMacks (talk) 20:01, 5 August 2012 (UTC)
Yes, it would be great if this annoying behavior was fixed (especially since many editors seem to dislike it). David1217 What I've done 16:24, 11 August 2012 (UTC)

Reference bubble


Now, when I fly over a reference link, I get a bubble with the ref above my mouse, and that is great !

But what happens when I fly over a ref link which is near the top of the view ? The reference goes above the view, where I can't see it. This is not very clever. In such case, it would be nicer to hatch the ref bubble below my mouse. Look at this screen photo.

Thanks for improving that, and keep up the good work !

--Nnemo (talk) 04:16, 5 August 2012 (UTC)

This is ReferenceTooltips, shown in Preferences → Gadgets as "Reference Tooltips: hover over inline citations to see reference information without moving away from the article text". It was added to the gadgets list with this edit (see Wikipedia:Gadget/proposals#Reference Tooltips and Wikipedia:Village pump (proposals)/Archive 87#Gadget proposal), and switched on by default with this edit (see MediaWiki talk:Gadgets-definition#ReferenceTooltips on by default). I believe that the discussion page for the gadget is at User talk:Yair rand/ReferenceTooltips. --Redrose64 (talk) 10:00, 5 August 2012 (UTC)
See also this discussion.
--Nnemo (talk) 10:09, 11 August 2012 (UTC)

Why is this title triggering the blacklist

I recently tried to move Delta Opioid receptor to δ-opioid receptor, but the move page tells me that the title " δ-opioid receptor" is on the title blacklist. However I'm looking through that list and it doesn't look like there are any regex rules on that page that should be triggered by the new title... Why is it getting triggered? flaming () 13:50, 10 August 2012 (UTC)

.*[^\P{So}¦§©®°¶№™☀-⚲✁-➾`].* <casesensitive> # "Other punctuation", with some exceptions (may need more, this is a huge character class); note that single-character titles are permitted by the [[MediaWiki:Titlewhitelist|title whitelist]]
That what I get with my crude JS blacklist matching tool. — Dispenser 13:55, 10 August 2012 (UTC)
Weird, I didn't think greek letters were in the "symbol (other)" character category... Thanks flaming () 14:08, 10 August 2012 (UTC)
It's very crude. I haven't seen PHP's regex extensions in any other languages and have not fix my JS tool to recognize them. You should also check the global m:Title blacklist. What's preventing a public tool is the fear vandals will test against it. — Dispenser 14:21, 10 August 2012 (UTC)
δ is U+03B4 δ GREEK SMALL LETTER DELTA is not in Unicode general category So but Ll (Letter, lowercase). Could it be the lc problem, can you try uc: Δ-opioid receptor and add DISPLAYTITLE to the page? -DePiep (talk) 14:59, 10 August 2012 (UTC)
Ah, that page already exists (Redirect). You'll can swap pages (and add {{lowercase title}}; looked good to me in the redirect page preview). -DePiep (talk) 15:02, 10 August 2012 (UTC)
My less-crude PHP-based blacklist matching tool returns two rules matching that title:
(?!(User|Wikipedia)( talk)?:|Talk:)\P{L}*[^\p{Latin}\P{L}].*\p{Latin}.* <moveonly> # Latin + non-Latin
(?!(User|Wikipedia)( talk)?:|Talk:)\P{L}*\p{Greek}.*[^\p{Greek}\P{L}].* <moveonly> # Greek + non-Greek
In English, moving pages to titles that contain both Latin and non-Latin letters is not allowed, as is moving pages to titles that contain both Greek and non-Greek letters. This helps prevent homoglyph attacks to evade more specific blacklist rules. Anomie 20:25, 10 August 2012 (UTC)
It was an uppercase/lowercase issue. The target page already existed (so it was allowed by the blacklist!). -DePiep (talk) 23:44, 10 August 2012 (UTC)
I've made a requested move for it rather than copy-paste moving. There's already a {{lowercase}} on the page so it'll stay when it gets moved. flaming () 15:44, 10 August 2012 (UTC)
And if neither lowercase δ nor uppercase Δ are in unicode So... and I don't see either delta character (upper or lower) on the global title blacklist... So I wonder what's triggering the blacklist? flaming () 15:47, 10 August 2012 (UTC)
I don't think it is on the blacklist, but simply the page you wanted to move to already exists (so is clearly allowed). Probably because you typed lowercase and the check was positive on the (automatic) uppercase the message got distorted. You'll have to swap the two existing pages, which requires admin assistance. -DePiep (talk) 15:59, 10 August 2012 (UTC)
Yes check.svg Moved, but page not otherwise edited. --Redrose64 (talk) 16:11, 10 August 2012 (UTC)
Isn't this a bug? -DePiep (talk) 23:45, 10 August 2012 (UTC)

────────────────────────────────────────────────────────────────────────────────────────────────────I tested this: 1. Create page delta-test (content: just the {{lowercase}} template). Looks OK, title showing in lc. 2. Try: Move to δ-test. Denied, message in red: "Delta-test" cannot be moved to "Δ-test", because the title "Δ-test" is on the title blacklist. If you feel that this move is valid, please consider requesting the move first. 3. Create page δ-test (entered in lc), named Δ-test technically (uc). Accepted. Adding the lc template works fine. 4. Try (again): Move delta-test to δ-test. Same message about blacklist, and also (in red): The page could not be moved: a page of that name already exists, or the name you have chosen is not valid. [...]. Note: the blacklist is skipped for admins; I am not an admin (mw). Conclusion: The title was blacklisted for moving, but not for creating. Looks like an inconsistency to me. (My testpages to be deleted). -DePiep (talk) 15:35, 11 August 2012 (UTC)

For all those who hate the new move system

Kangaroopower has created a script that allows you to move a page without the drop down menu. You can find it at User:Kangaroopower/ajaxMove.js and add it by adding importScript('User:Kangaroopower/ajaxMove.js'); to your common.js page. He welcomes suggested improvements. Ryan Vesey 20:28, 10 August 2012 (UTC)

Just FYI another option is this dewiki gadget that modifies the normal MovePage to get rid of drop down menu. — AlexSm 21:27, 10 August 2012 (UTC)
What drop down menu? --Redrose64 (talk) 21:39, 10 August 2012 (UTC)
There is a drop down namespace menu when you attempt to move a page. This results in problems when trying to make a cross namespace move since it defaults to your original namespace. Ryan Vesey 21:41, 10 August 2012 (UTC)
Oh, that. It's hardly new (see Wikipedia:Village pump (technical)/Archive 98#screwed up REDIRECT) - I think it arrived with MediaWiki 1.19 at the end of February this year, immediately before the sudden increase in page moves from User: space to Wikipedia: space. Personally I got used to it within a day or two. --Redrose64 (talk) 22:51, 10 August 2012 (UTC)
See also the Bugzilla request to improve mediawiki.special.movePage.js to remove the namespace name when the user paste the full target name in the field. Helder 12:30, 11 August 2012 (UTC)
What happens when they want to move to article space? Do they need to use the menu? Ryan Vesey 15:15, 11 August 2012 (UTC)
Yes: they need to select "(Article)", which is at the top. Those instances in March (which I believe still occur) were because relatively-newbies didn't realise that the page they've created in user space is an article, so should go there: they see the familiar-looking "Wikipedia", and select that. --Redrose64 (talk) 17:57, 11 August 2012 (UTC)

Cite_web/smart hybrid of Cite_quick

Tangent thread from: "#Cite_quick 12x faster for short cites".

The new, experimental Template:Cite_web/smart is an attempt to totally upgrade {Cite_web} to be both faster and handle all complex {Cite} parameters known to mankind, now and in the future. The tactic is to check for rare parameters in {Cite_web/smart} and only then invoke {Citation/core}; otherwise, {Cite_web/smart} would perform like {Cite_quick} to format the parameters in the same pattern as Citation_style_1, but not really using {Citation/core} when only common parameters are used. For common parameters {Cite_web/smart} seems to run 4x faster than {Cite_web}, while for rare parameters, it runs about the same, or formatting 17 cites per second. The grand finale would be: #REDIRECT of {Cite_web} to invoke {Cite_web/smart}, for all among the 1.6 million articles which use Template:Citation/core. Note well: Although {Cite_web/smart} looks like a "fork" of {Cite_web}, it is really intended to be the next generation, perhaps ready within a few days. For people who wanted to test a template, then this is your mega-chance. It will be used in over 1 million articles (transcluded into 1,128,784 pages, June 2012), so that is why testing is so much more critical: {Cite_web/smart} is not in the little league of {Cite_quick} to be used in a few hundred slow articles. Instead, {Cite_web/smart} is the big-league replacement for {Cite_web}, which people have been wanting for years. After testing and protecting, then redirect {Cite_web} to invoke {Cite_web/smart}. Similarly, we would have a {Cite_news/smart}, etc. -Wikid77 (talk) 17:46, 11 August 2012 (UTC)

  • Already listed at TfD: The lastest round of debates at wp:TfD is for Template:Cite_web/smart, which should generate a wider discussion, and see what other people suggest:
The template already works fairly well, so I think any further tests should give mostly positive results. -Wikid77 (talk) 20:41, 11 August 2012 (UTC)

Italics (or mystical test edit)

Did I just see some test edit by Wikipedia's engineers or something? I was just looking at MOTD and saw a fellow user's signature. The end code of the italics did not work very briefly so it italicised the rest of page. i/> showed very briefly instead of ''.

Even more strangely, this does not show up in page history either. Simply south...... flapping wings into buildings for just 6 years 17:22, 11 August 2012 (UTC)

Maybe it was rejected as invalid? The I element is always used as an enclosure <i>...</i> since it has no effect if used as an empty element <i /> --Redrose64 (talk) 20:59, 11 August 2012 (UTC)

why has

how come my website has been taken out of the external links in service animal section, I am trying to relay a comprehensive and easy way to purchase ID tags and give a little information about them in the external links section. To help people that do not understand all about service animal and emotional support animal ID tags. The site is

Thanks again — Preceding unsigned comment added by Fadeous1 (talkcontribs) 19:17, 11 August 2012 (UTC)

You mean this edit? Most likely because those links fail the guideline on external links. Have you asked the editor who removed them? --Redrose64 (talk) 20:28, 11 August 2012 (UTC)

Because you are a business trying to make profit and not a direct source that is of use to the definition of service animals. Quit wasting your time.

Table cell should be pushed higher

Can someone explain why or howto: the 2nd cell is not pushed larger by the font-size. Situation is that {{braille cell/core}} sets the font-size within the cell (span the text), not in wikitable code (style between pipes).


-DePiep (talk) 10:18, 10 August 2012 (UTC)

Please? -DePiep (talk) 01:32, 11 August 2012 (UTC)
  • Look at Template:braille cell/core, for the font-size issues. A style="font-size:300%" would display 3x larger than the basic size, just as 500% would display the letter "C" as 5x larger.
  • A -- {{Braille cell/core| type=text | character=A}}
  • A -- {{Braille cell/core| type=text | character=A |size=100%}}
  • B -- {{Braille cell/core| type=text | character=B |size=300%}}
  • C -- {{Braille cell/core| type=text | character=C |size=500%}}
I guess people are unsure as to what the question is specifically. -Wikid77 (talk) 04:19, 12 August 2012 (UTC)
He wants to know why the table cell height does not auto-adjust iteself to the font height. Edokter (talk) — 09:42, 12 August 2012 (UTC)
As Edokter says. I could have been more clear then. -DePiep (talk) 12:57, 12 August 2012 (UTC)

Weird pipes display issue

For background, see the discussion here between myself and User:Giants2008. He said there were a number of pipes showing up in section headers, but neither myself nor User:TheAustinMan are able to see them. Specifically, I'd like everyone to take a look at this diff and see if there are any visible changes between edits. 24 bytes were removed, but for me, the before and after appear identical. Does anyone know what the issue is here? There are apparently similar problems in the article's references. —Torchiest talkedits 17:43, 10 August 2012 (UTC)

Oh, and I tried viewing the difference in Chrome, Firefox, and IE, and had the same results in all three: no pipes were visible. —Torchiest talkedits 17:44, 10 August 2012 (UTC)
My browser doesn't display them differently. I copy-pasted to the "Characters" field at (no Wikimedia affiliation) and clicked "View names". This says the previous version had one or two U+200C zero-width non-joiner before the en dash. PrimeHunter (talk) 18:24, 10 August 2012 (UTC)
Thanks for looking into that. I don't understand how such a bizarre character got into the section headers, and why only one editor (so far) sees it represented as a pipe. —Torchiest talkedits 18:51, 10 August 2012 (UTC)
I suspect not an actual pipe | but this symbol ISOIEC-9995-7-081--IEC-60417-6077-1--Symbol-for-ZWNJ.svg which is apparently how it should be represented on keyboards. --Redrose64 (talk) 19:04, 10 August 2012 (UTC)
The characters were added in this edit by TheAustinMan. PrimeHunter (talk) 19:15, 10 August 2012 (UTC)
Same issue [2] here: ZWNJ showing pipe-like in a styled text. Would like to know in which situation (browser) this happens. -DePiep (talk) 13:09, 11 August 2012 (UTC)
I know it can happen on IE, because that's the browser I use. It just looks like a pipe to me, as that's something I'm familiar with. I had never seen the ZWNJ character before today. Shows how techincally oriented I am. :-) Giants2008 (Talk) 16:17, 11 August 2012 (UTC)
Don't worry, ZWNJ is mostly for non-Latin scripts and typographic issues ;-). It just should not show. Can you give the IE version? (others here mentioned "All OK" on their IE). -DePiep (talk) 16:33, 11 August 2012 (UTC)
OK, at Template:Punctuation marks there are two infoboxes - one being the demo of the template itself, the other is in the documentation green box. In Firefox 14.0.1 these look the same: but in IE 7.0.5730.13 the one in the doc (only) looks like this. --Redrose64 (talk) 18:34, 11 August 2012 (UTC)
We goddim. So IE 7.0. Added a note to {{zwnj}} template /doc. Curiously, the ZWNJ was deleted from the main template July 31. Some delay occurred, but for this catch it was welcome. At last I got a chance to see it. - DePiep (talk) 18:53, 11 August 2012 (UTC)
I think my version is 7.0, but I'm not 100% sure. Does seem to fit the other evidence, though. Giants2008 (Talk) 22:31, 11 August 2012 (UTC)
Open the "Help" menu, and select "About Internet Explorer"; this should produce a box with a row beginning "Version:". --Redrose64 (talk) 22:36, 11 August 2012 (UTC)
Thanks for the advice. My version is actually 9.0, so maybe the problem is more widespread from the IE end. Giants2008 (Talk) 18:08, 12 August 2012 (UTC)

Wow, excellent detective work, everyone. Thanks for all the help figuring this out. —Torchiest talkedits 14:43, 12 August 2012 (UTC)

Finally came across this issue after being busy for the past few days. As to how I added them in the first place, I'm not very sure. TheAustinMan(Talk|Works) 20:46, 12 August 2012 (UTC)
Not your fault. The browser should not show ZWNJ. Ever. -DePiep (talk) 23:24, 12 August 2012 (UTC)
Yeah; what Microsoft are supposed to do, and what Microsoft actually do, are two different things; and probably always will be. --Redrose64 (talk) 23:38, 12 August 2012 (UTC)

WP Editor loses main text when going to a different page

So at some point, perhaps when they did the major upgrade, the behavior of the editor changed, and it automatically refreshes upon leaving to a different page and returning with the back button, contrary to the behavior of most of the net. This means one often will lose a lot of work (as I just did) if one accidentally clicks a stray link, or as had happened, had an edit conflict and instinctively hit the back button. ♫ Melodia Chaconne ♫ (talk) 04:25, 11 August 2012 (UTC)

All my 5 browsers keep the edit text when I return with the browser back button: Firefox, IE, Chrome, Opera, Safari. Which browser do you have? PrimeHunter (talk) 11:17, 11 August 2012 (UTC)
FF12, which doesn't have the problem for the majority of pages, and used to be fine (or at least earlier versions of FF) for WP. And please, no "upgrade FF" comments, as I have good reason to not do that. ♫ Melodia Chaconne ♫ (talk) 13:37, 11 August 2012 (UTC)
  • MSIE formerly reset edit-buffer on backspace key not back-key: Depending on the mode, in MS Internet Explorer, the backspace key would reset a newly edited page, to erase all changes and reset the edit-buffer to the original text. For that reason, when expecting a long edit, then first edit-preview the page, such as "Show changes" (none at first diff). Currently, the backspace key acts as browser "Back" and the edited text remains in the edit-buffer after clicking a link and returning Back. I ran the test on this page, as a "WP:" project-namespace edit, but I think even editing articles should retain the edit-buffer changes. I know it is extremely frustrating, so please keep another window nearby, or text-editor, to copy-paste changes (^A^C) every few minutes. If editing multiple articles/pages together, then save the edited text for each article in a separate local file, else a power-outage (storm or pet animal hits plug), or deletion of temporary files, could lose all prior changes to all edit-buffers. Otherwise, it is scary to make long edits, and I bet some new editors have quit because they lost some long edits. I expect much disappointment when people learn the new WYSIWYG interface can lose all tedious keystroke edits (which cannot be copy-pasted), unless there is a keystroke backup file for safety. -Wikid77 (talk) 13:21, 11 August 2012 (UTC)
But there is also work to add instant save functionality after every few keystrokes. - Jarry1250 [Deliberation needed] 16:51, 11 August 2012 (UTC)

I've been getting this from time to time as well, and it isn't browser-specific, either - just happens around these projects sometimes. Turning off the fancy js edit toolbars and other edit box gadgets seemed to resolve it at least some of the time, though, so you could try doing that. -— Isarra 06:50, 13 August 2012 (UTC)

Okay, this may seem completely random, but did the page this happened on have an editnotice? -— Isarra 07:14, 13 August 2012 (UTC)

Parser functions and newlines

Is there some way to substitute newlines in wikitext? For example,

content.{{ {{{|safesubst:}}}#ifexist:Main Page|

}}== Heading==

doesn't work, because whitespace is trimmed within template/parser function arguments. Is there a way to circumvent this? (Note: I don't want displayed line breaks, I want to substitute wikitext line breaks.) — This, that, and the other (talk) 05:12, 11 August 2012 (UTC)

There are various techniques, but they all rely on there being no whitespace at the start or end of the parser function's output. For example, you could move the full stop and the first equal inside it; if doing this, you'd also need to ensure that the "else" case is catered for:
content{{ {{{|safesubst:}}}#ifexist:Main Page|.

=|.=}}= Heading==
of course, this does mean that there is a small amount of duplication. --Redrose64 (talk) 14:02, 11 August 2012 (UTC)
  • Could use dummy tags around newline: Another technique is to place dummy HTML tags (such as "<b/>") at the start/end of the newline, to prevent removal of the newline during parsing of the if-structure. So:
Text here.{{ {{{|safesubst:}}}#ifexist:Main Page|<b/>
<b/>}}More text here.
Text here.
More text here.
However, there are some users who detest inserting dummy tags, so that could be a problem, such as issues of wp:Accessibility for screen readers. Use of "<b/>" is probably ok, but beware complaints. -Wikid77 (talk) 16:57, 11 August 2012 (UTC)
I considered (and tried) that method: but for the specific case given, it doesn't work, because ==Heading== must occur at the very start of the line - dummy tags like <b /> kill it:


so it appears as plain text. --Redrose64 (talk) 18:48, 11 August 2012 (UTC)
  • That is a good point, and using "<br>" for a newline also fails to trigger the "==Header==" so there is no direct general solution without placing the lead "=" inside the specific ifexist-structure. I would change the MediaWiki software to treat "<nl>" or such as a newline character within parameter text. -Wikid77 (talk) 19:57, 11 August 2012 (UTC)
I would like that wikicode "keep newline" solution. I can never trust my own whitespaces. (eg in template & table rows) -DePiep (talk) 20:28, 11 August 2012 (UTC)
While that would be useful for general use, the parserfunctions extension also really shouldn't be trimming out whitespace with comments at the ends as well, should it? That seems a little silly. -— Isarra 17:49, 13 August 2012 (UTC)

Closer Toolserver integration (or abusing iframes/javascript for fun and profit)

I've often wondered why there hasn't been any closer integration of toolserver tools and Wikipedia itself. The lack of integration often results in unnecessary duplication, and often clumsy workarounds (e.g. creating an account for acc, then having to manually verify that your are who you say you are). Other-times, the tool requires that you enter information that Mediawiki would already know.

While this has worked pretty well so far, I think looking towards the future, we should try and build a system that is better integrated . As an example, think something like facebook apps, but for Wikipedia. In order to give you an idea of what I'm thinking, I've created a user script that demos it. What the script does is, use javascript to create a fake special page, and then use an iframe to load the toolserver tool. For the rest of this post, I'll assume you have importScript('User:Chris G/fake toolserver apps.js'); in your Special:Mypage/vector.js.

Basic Integration

The simplest form of this, is just using the iframe and having no other data passed. Imagine something like Special:EditCount/Chris G (or for IP addresses WHOIS data). What benefits does this have? Well, it looks nice, and gives a better browsing experience (rather than hopping back and forth between Wikipedia and the toolserver).

More personalised integration

Using URL parameters we can pass information to the app, via the iframe (e.g. <iframe src="http://toolserver/~example/example.php?userid=1337>"). This can allow the tool to give a more personalised user experience. What should be noted however, is that there is no way for the app to verify the user is who they say they are. It would be trivial to pretend to be another user. Hence any apps using this would not be able to show non-public data about the user.

Full integration

Expanding on the idea above, it would be possible to pass an edit token to the app (This creates both privacy and security concerns, I'll address those in a sec). By passing the edit token, we can do two things. Firstly, we can verify the user is who they say they are. Secondly, we can perform actions on their behalf. To give a couple of examples of the kinds of things we'd be able to do with this:

  • WP:ACC - This would make it both easier for people requesting accounts, and those who create the accounts for them. Having a page like Special:RequestAccount is much more user friendly IMO. Likewise for account creators, the process can be slightly simplified. Rather than having to manual create the account via Special:Userlogin/signup, they could simply click a create button, and the app would do the rest.
  • - Currently the unblock appeal interface is pretty damn daunting for someone not familiar with Wikipedia. On top of that, it asks for a lot of redundant information that Mediawiki could fill in by itself. If it were to be integrated we could have something much simpler like Special:AppealBlock
  • Pages like WP:CHU. Rather than having to fill out a complex template that can often trip up new users, they could fill in a simple form (desired username, reason for change, etc.), and the app would then be able to edit WP:CHU on their behalf. The app could even do basic checks to make sure the requested name is valid, not already in use, etc.

I'm sure there must be heaps of others situations where this kinda thing would be cool and useful.

Food for thought


As always privacy is a big concern. The main problem is that with the personalisation, tool owners would be able to link users with IP addresses and their user agents. Depending on how much logging was used, they would also be able to link users with actions on their tools (e.g. some kind of search tool, the user might not want the tool owner to know that they spent the entire day using the tool to search through all of Wikipedia's articles on sexual positions).


Giving an app your edit token is a pretty damn dangerous thing to do. It's almost as bad as giving someone your password. (On a side note, passing an edit token via GET to the app, would leave your edit token in the web servers access logs, which is even worse. But this can be solved by sending them via POST or some javascript/HTML5 trickery).

Final Thoughts

I think the best solution to the above thoughts would be to create a dedicated app extension. Such an extension would be able to help apps verify users, without having to risk sending their edit tokens. It could also allow apps a restricted level of usage of a users' account. On using the app for the first time, users would be shown a dialogue showing them what the app can do (e.g. for acc create accounts only), and then decide whether they'll allow the app access. It could even be indicated in edit summaries/log entries (e.g. "Created account (Using WP:ACC)").

Sorry for the length. Hopefully that all makes sense. --Chris 06:45, 12 August 2012 (UTC)

Already doing this although most people use the copy/paste sign in widget on the right. However, the edittoken, which lasts only for a few hours and requires the user's unattainable HTTP-only browser cookies, cannot be used to verify the account unless you intend to change their preferences or make junk edits. Account verification is done with the watchlist token cookies by attempting to read the watchlist. These tokens are stored in cookies and can only be read by my tools. Probably more secure than TUSC, see my analysis of that. If you want I have an idea of how to get centralauth_Token (which is considered username/password equivalent) through a interface that looks like Facebook connect. — Dispenser 14:52, 12 August 2012 (UTC)
The idea of ts tools having edit tokens is pretty scary imo. In particular I imagine there are tools with XSS vulnrabilities (Most tools are not reviewed by others. Several are written by inexperianced people, etc). If they have your edit token, that is a scary thing. Bawolff (talk) 19:42, 13 August 2012 (UTC)

Finding the source of a formula image by LaTex

I asked this question in Wikipedia:Help desk, but was directed here.

I'm writing a JavaScript code that allows to edit the LaTeX code of formulas by double-clicking them, and I was wondering if it's possible to determine what's the source of a formula image just by knowing its LaTeX code. e.g. the source of is always What's the algorithm used to determine a formula's source? Is it available in JS, or at least executable via AJAX? (I was thinking about using other LaTeX rendering engines, such as codecoge's, but sometimes it doesn't give the same result as Wikipedia. It's a problem especially when the external rendering engine supports a command that Wikipedia doesn't or vice versa).

P.S. speaking of which, can you recommend any WYSIWYG online editor for LaTeX? (similar to bakoma-tex, but online, so I can integrate it in Wikipedia).

Thanks a million! (talk) 21:34, 8 August 2012 (UTC)

I wouldn't know how to answer, but if you ask at WP:VPT they might.— Vchimpanzee · talk · contributions · 20:55, 10 August 2012 (UTC)
The computing reference desk might actually be a better place, because I'm not sure whether you're asking for a Wikipedia feature.— Vchimpanzee · talk · contributions · 21:04, 10 August 2012 (UTC)

Thanks, (talk) 13:17, 12 August 2012 (UTC)

The 0cfae383362bc63d7ac429a5755fef05 is the first n characters of a cryptographic hash, so no, sorry, you won't be able to reconstruct content reliably just from the URL. However, I note you are assuming PNG conversion mode, whereas most people are now using MathJax mode, where the source is left in the HTML and consequently easy to extract. - Jarry1250 [Deliberation needed] 13:59, 12 August 2012 (UTC)
I'm not trying to extract the LaTeX code from the source, but rather the opposite – the source from the LaTeX (I can get the LaTeX from the alt attribute). My goal is to display a preview of the formula by simply changing the src attribute of the image. As for MathJax, does it support exactly the same commands as Wikimedia's rendering engine? (I want to warn the user of any mistakes, but don't have time to write a full algorithm that checks the code.) Thanks, (talk) 14:25, 12 August 2012 (UTC)
Oh, sorry. In which case, the relevant lines of PHP are "$this->hash = substr( $contents, 1, 32 );" where $contents is what the tex conversion commandline program returns. Hope that helps. To answer your question, I believe it supports everything yes, but I'm not 100% sure. - Jarry1250 [Deliberation needed] 21:37, 12 August 2012 (UTC)
You'll probably need to use the action=parse of the API to test that sort of thing. texvc isn't triggered from a 404 handler, so one would need to have previewed a page with the exact same equation on it for the src-changing to work (which honestly may be likely most of the time given how many equations have been typed in in Wikipedia's history). Bawolff (talk) 19:45, 13 August 2012 (UTC)

MediaWiki's Wiki code

Hello all,

I have no particular issue about Wikipedia, but here is the only populated place where someone could respond to me.

It is Ward Cunningham who created the wiki system, with the [[ and the ]], as well as '''. But I would like to know if any one knows where the list of wiki syntax rules can be modified in a MediaWiki? I believe that some changes could be more efficient on non-English keyboards, as well as on mobile phones.

Thank you all.

Kahlores (talk) 14:39, 12 August 2012 (UTC)

If you want to change parser rules for your local installation, you should be able to find them under w/includes/parser/Parser.php, written as (from memory) regex-based replacements - not at all easy to modify, I'm afraid. There's currently a parser rewrite underway (codenamed Parsoid) which aims to tidy up and sanitise those rules, however; you could perhaps wait for that to finish its development cycle and then piggyback. Sorry to not be of more help, - Jarry1250 [Deliberation needed] 21:31, 12 August 2012 (UTC)
Ward Cunningham did not create the wikilinking system using two square brackets; his first wiki used CamelCase links. See Wikipedia:CamelCase and Wikipedia for the full story. Graham87 02:34, 13 August 2012 (UTC)
Jarry1250, I like your modesty: I couldn't expect any better. Thanks Graham for correcting my historical mistake. Kahlores (talk) 17:49, 13 August 2012 (UTC)

Gif not animating?

Obesity state level estimates 1985-2010.gif

Why doesn't File:Obesity_state_level_estimates_1985-2010.gif work?Smallman12q (talk) 20:24, 12 August 2012 (UTC)

It works for me. Which browser do you have and what do you see? PrimeHunter (talk) 20:32, 12 August 2012 (UTC)
For me (xombrero, WebKitGTK-based) the thumb doesn't work (still image with three maps), but the full size image works with no issues. Several other animated gifs work for me, though. — Dmitrij D. Czarkoff (talk) 20:58, 12 August 2012 (UTC)
I assume you mean the lack of animation, see commons:Commons:ImageMagick limitations#GIF. — Dispenser 21:26, 12 August 2012 (UTC)
I've filed bug 39295 All the frames are the same size...I'd like to know the technical reason (what the ImageMagick error is) why its not animating.Smallman12q (talk) 22:11, 12 August 2012 (UTC)
The section heading was "Borked gif?" earlier so I didn't know the problem was animation. It doesn't animate for me either. If I read mw:Manual:$wgMaxAnimatedGifArea correctly then the problem is that 1,080 × 720 pixels × 27 frames = 20,995,200 > 12.5 million. Try reducing the size of the original file. PrimeHunter (talk) 23:21, 12 August 2012 (UTC)
Resolved: Uploaded at lower resolution 60dpi and it works. Thanks! There really should be some sort of error message to let you know the image is too big.

Smallman12q (talk) 00:23, 13 August 2012 (UTC)

Animated gifs are some of the best illustrative tools on Wikipedia. And so I agree that we need some kind of error message to help people. That is an amazing image by the way. It is interesting how the obesity trend concentrated in certain parts of the Midwest at the beginning of the timeline in 1985. There needs to be some kind of automated addition of this note to animated gif pages: "Click the escape key to stop the GIF at any point. Reload the page to restart the animation." The escape key works in Firefox and Internet Explorer, but not in Chrome. What works in Chrome?
Also, what did you use to reduce the size? That might be useful to add to a help page concerning this problem. --Timeshifter (talk) 04:29, 13 August 2012 (UTC)
I lowered the resolution of the image output in PowerPoint. I've started a page at Wikipedia:PowerPoint to animated gif.Smallman12q (talk) 21:37, 13 August 2012 (UTC)
Thanks. I added a link to that from here:
commons:Commons:Animated image resources --Timeshifter (talk) 01:11, 14 August 2012 (UTC)
However, they are troublesome to print out. TimedMediaHandler (or something) is suppose to help whenever that's release in 2018. — Dispenser 04:55, 13 August 2012 (UTC)

Very nice map, but in an animation like this there should be an extra long delay when a single run is complete (when 2010 appears in this case). Then the viewer has a hope of seeing the whole thing, and can more easily recognize that what follows is a re-run. Johnuniq (talk) 07:40, 13 August 2012 (UTC)

bugzilla:39297 for the better userfeedback when wgMaxAnimatedGifArea is tripped. —TheDJ (talkcontribs) 09:37, 13 August 2012 (UTC)

Messed up pushpin

Please assist José_C._Paz,_Buenos_Aires has a pushpin map in the middle of the text and I don't know how to fix it. Thanks. —Justin (koavf)TCM 06:29, 13 August 2012 (UTC)

The map on enwiki and eswiki are totally different. --MakecatTalk 07:57, 13 August 2012 (UTC)
I switched the infobox to location map of Argentina, so the pushpin is now on map. — Dmitrij D. Czarkoff (talk) 10:54, 13 August 2012 (UTC)

PNG instead of SVG

Please assist {{Line chart}} generates PNG when it's perfectly suited for SVG. Can someone amend this? —Justin (koavf)TCM 07:47, 13 August 2012 (UTC)

EasyTimeLine's svg support is experimental. This is bugzilla:35320 Bawolff (talk) 14:53, 13 August 2012 (UTC)
(edit conflict) It's nothing to do with {{line chart}} per se; that template inserts a <timeline>...</timeline> element into the wikicode, which is recognised by MediaWiki and processed into an image, see mw:Extension:EasyTimeline. --Redrose64 (talk) 15:06, 13 August 2012 (UTC)

Bunched-up images

I have a problem with bunched-up images – see here for example: [3]

It is caused by the use of two infoboxes (I know, I know).

I don't know if this is the right way how it's done.

I have "fixed" it for now by moving the first image down below the first infobox. Another solution would be to move the second infobox down. I'm looking for better solutions though. Tony Mach (talk) 19:23, 13 August 2012 (UTC)

When multiple box-type objects (whether infoboxes or images) appear on a page, the order that they appear in the wikicode governs the order in which their upper edges appear in the rendered page. So, if you have {infobox, infobox, image}, the image cannot appear further up the page than the second infobox, even if the image is left-aligned (the infobox being right-aligned). If you order them {infobox, image, infobox} you will find that the image sits somewhere between the upper edge of the first infobox, and the upper edge of the second.
If you want to get into the technicalities of infobox construction, here's how to effectively merge two infoboxes into one, so that the left-aligned images can go in their proper places. First, amend {{Infobox Space mission}} so that it has one extra line, containing |child={{{child|}}} - it doesn't really matter where, so long as it's inside the {{infobox}}. Then, in Apollo–Soyuz Test Project, amend the infoboxes from
{{Infobox Space mission
| mission_name      = Apollo–Soyuz Test Project 
| next_mission      = [[File:ALT mission patch.PNG|35px]] [[Approach and Landing Tests]]
{{Infobox Space mission
| mission_name      = ''Soyuz 19''
| next_mission      = [[Soyuz 20]]
|header1 = Apollo–Soyuz Test Project 
|data2 = {{Infobox Space mission
|child = yes
| mission_name      = Apollo–Soyuz Test Project 
| next_mission      = [[File:ALT mission patch.PNG|35px]] [[Approach and Landing Tests]]
|header3 = ''Soyuz 19''
|data4={{Infobox Space mission
| child = yes
| mission_name      = ''Soyuz 19''
| next_mission      = [[Soyuz 20]]
note that the two extra parameters |child=yes go inside both infoboxes - omitting these will cause double borders. See Template:Infobox space mission/testcases
BTW with the next edit, that {{refimprove}} was moved below the infoboxes: per Wikipedia:Manual of Style/Lead section#Elements of the lead it should come before the first infobox. --Redrose64 (talk) 21:16, 13 August 2012 (UTC)


Hello there, I don't know if I'm right here (the main English Wikipedia is just too big compared to my homewiki Simple English Wikipedia) but I have a simple question: Please see the formating bar above the edit text box (with that A A "Advanced", "Special characters", "Help" etc). What is that "Cite" thing? If I'm not wrong, it is used for inserting references. It can insert templates like cite-web etc. My question is, is this feature a gadget? Would it be able to enable it at the Simple English Wikipedia, which uses exactly the same templates? --weltforce (talk) 19:34, 13 August 2012 (UTC)

See Wikipedia:refToolbar 2.0#Porting to another wiki. It's installed as part of Preferences --> Editing here, but that page has instructions for installing it as a gadget. --Philosopher Let us reason together. 23:14, 13 August 2012 (UTC)
There is also some relevant discussion at Wikipedia talk:RefToolbar 2.0. Would you mind pinging me at simple:User talk:Philosopher if you decide to install it? Thanks. --Philosopher Let us reason together. 23:18, 13 August 2012 (UTC)

Table sort problem

A few hours ago, Mediawiki 1.20wmf9 was deployed to English Wikipedia. There is currently an issue with tablesorting icw certain browsers because of a bug in this release. I'd thought I preempt any reports about it. —TheDJ (talkcontribs) 19:59, 13 August 2012 (UTC)

I've just been informed the fix is now being reviewed and will hopefully arrive quickly. —TheDJ (talkcontribs) 20:01, 13 August 2012 (UTC)

Wikipedia Mobile View bug?

Accessing Wikipedia pages by Google search on my up-to-date iPhone 4S using Safari shows up in desktop view instead of automatically switching to mobile view. Can anyone confirm this for me? Scott523 (talk) 20:13, 7 August 2012 (UTC)

Go to Wikipedia:Enable mobile version and click the link. Ryan Vesey 22:00, 7 August 2012 (UTC)
-Clicking the link takes me to the desktop version of Wikipedia's homepage. Scott523 (talk) 19:53, 8 August 2012 (UTC)
-Thanks Ryan, however, clicking that Enable mobile version link (invoking the mobileRedirect.php script) doesn't solve the problem for my Mobile Safari either (iOS 5.1, iPhone); it still leaves the browser stuck in the desktop version as well.
Fortunately, the workarounds on the talk page for Enable mobile version worked (for completeness, I tried three; they all worked with varying degrees of success: enabling private browsing worked (Workaround 1), but it had the known downsides of private browsing: no history, and log in cookies aren't retained; manually deleting Wikipedia's cookies worked one time, but, any subsequent clicks on the "Desktop" link recreated the problem and I had to manually delete the cookies *again* - that proved too frustrating; I'm now using the custom javascript as a my long term workaround).
Does anyone know if has a bug tracking this annoying problem with mobileRedirect.php and iOS' Mobile Safari? sn‾uǝɹɹɐʍɯ (talk) 18:39, 14 August 2012 (UTC)
Found it - Bug 38009 tracks this issue; thank you Dreamyshade. sn‾uǝɹɹɐʍɯ (talk) 22:47, 14 August 2012 (UTC)

half-size type?

Apparently some sequence of key strokes, possibly involving the "control" key, suddenly caused Wikipedia to appear on the chrome browser with type about half as big as it has usually appeared in the past. Other web sites are not affected, nor is Wikipedia viewed with firefox affected. Another effect is that things set in bold do not appear in bold, but look like everything else. What's going on? Michael Hardy (talk) 04:17, 12 August 2012 (UTC)

Ctrl++ to increase zoom and Ctrl+- to decrease. Ctrl+0 resets the zoom to default. Users accidentally hold control while scrolling which also adjusts zoom. — Dispenser 05:03, 12 August 2012 (UTC)
This gets reported surprisingly often. My working theory is that the trend for vertical length in Wikipedia pages leads to a lot of mouse-button scrolling, and it's much easier to accidentally make drastic font-size changes if you scroll a lot with the other hand loosely resting on the keyboard! Andrew Gray (talk) 19:25, 14 August 2012 (UTC)

Set "title" attribute to the (Main) namespace wikilinks to the first sentence of lede

I believe it would be a nice option to have "title" attribute of links to articles to contain first several words (or even the whole first sentence) from the target's lead. This would help with checking for DAB links in the articles without additional page hits. AFAIK this feature isn't all that easy to implement, so I wanted to gather community's comments on topic before requesting it. — Dmitrij D. Czarkoff (talk) 13:44, 12 August 2012 (UTC)

Navigation popups does this. PrimeHunter (talk) 14:00, 12 August 2012 (UTC)
(ec) To see redirects, you can diddle your common.css, I for example have the following to show foreign-language text and redirects (of course choose colours you like):
/* see Template:Lang/doc#Applying_styles */
span[lang] { color: black; background-color: #FFE4E1; } /* misty rose */
/* highlight redirects */
.mw-redirect { background-color: #FFCF8A; }
I use the WP:POP gadget to preview the article lead without opening the page. I think it would be too clunky to try to embed parts of each target into the title attributes of each wikilink, just think how much extra information from different pages would need to be collected (or cached) before the page could be delivered. --Mirokado (talk) 14:15, 12 August 2012 (UTC)
My problem that made me go here is that I suffer from connection stability issues, so increasing the number of connections (which is the way Navigation popups work) isn't particularly helpful in my case. Achieving this on server side would probably reduce the load on servers, which would also be a positive effect. — Dmitrij D. Czarkoff (talk) 14:22, 12 August 2012 (UTC)
Personally, to me, I find it odd that MediaWiki adds an HTML class to redirect but not to disambiguation links (that I know of, and I doubt it's that much more difficult). Maybe something to poke bugzilla with? I can see the use case being plausible and not much more difficult than with what is already done with redirect links. Would that suit you Dmitrij? --Izno (talk) 23:42, 12 August 2012 (UTC)
Though I find it potentially helpful, I indeed want some lightweight (title captions only) server-side alternative to Navigation popups. BTW, I would like it to bypass redirects at all, as the goal is to make sure that I link to a right thing without requesting the target page. — Dmitrij D. Czarkoff (talk) 00:21, 13 August 2012 (UTC)
(edit conflict) The MediaWiki database does not internally distinguish disambiguation pages from other pages (unlike redirects, which are identified as such in the master page table). User scripts (such as linkclassifier) define disambiguation pages as those in Category:All disambiguation pages; but that categorisation varies from wiki to wiki and is not part of the core system. — Richardguk (talk) 00:27, 13 August 2012 (UTC)
Is there any reason that's the case for disambigs? --Izno (talk) 01:22, 13 August 2012 (UTC)
Redirection requires a page to be rendered completely differently (using the wikitext of the target page, plus a link beneath the heading); whereas disambiguation is an editorial concept and does not affect the HTML, so there is no requirement for the MediaWiki system to treat it differently.
In contrast, links to non-existent pages are treated differently (even though they could in principle be rendered in blue without the renderer having to check whether the target exists), because redlinks are treated as an essential part of the wiki functionality. So, if enough people care strongly enough about a feature, it can become part of the core system.
Richardguk (talk) 01:52, 13 August 2012 (UTC)
Except that MediaWiki does recognize disambig pages as not being content, despite being in a content namespace, via the fact that they do appear in the category "all disambigs", so MediaWiki is obviously aware of the fact that they are disambiguation pages (and I can't find the damn [Mediawiki] page which lists all such categories). Whether that awareness is re-usable is a different question, which would inhibit implementation of a class identifying disambig pages on 'proper' pages. /shrug --Izno (talk) 02:04, 13 August 2012 (UTC)
This is because the disambiguation pages contain the template {{disambiguation}} and {{Dmbox}}, which includes that hidden category. This is not Mediawiki, it's simple templates and categorization by editors. Again, it is not part of the system, it is an editorial process. —TheDJ (talkcontribs) 10:08, 13 August 2012 (UTC)
MediaWiki:Disambiguationspage suggests otherwise: "This list is used by MediaWiki to know what pages are disambiguation pages, and thus not articles. This is used by Special:Disambiguations to show any pages linking to disambig pages. This list is also used by various bots." Is it really just a page for shits and giggles, or is it used, as has been the case since eternity, as documented? --Izno (talk) 22:41, 14 August 2012 (UTC)
What about the original proposal? — Dmitrij D. Czarkoff (talk) 11:10, 13 August 2012 (UTC)
I doubt it will happen. With the amount of links we have, articles would explode in size that way, it's not efficient. Next to that, you have the problem that actually finding the lead really is a process that probably can't be 'unified' across many types of mediawiki installations. What is suited on wikipedia might not be that suitable for wikiquote for instance. JS based solutions are really the only sustainable way here. —TheDJ (talkcontribs) 16:58, 13 August 2012 (UTC)
Extension? Flexible settings? — Dmitrij D. Czarkoff (talk) 21:13, 13 August 2012 (UTC)
Things like that fragment the caching, and more options are confusing. It's also 5 to perhaps as many as 200 extra requests on the database in order to be able to retrieve the information and put it in a copy specifically for you. So that would skyrocket the render time of a page. It would become a rather slow Wikipedia for you. —TheDJ (talkcontribs) 08:16, 14 August 2012 (UTC)

Gadget for control of image display

For various reasons it would be useful to give users more direct control over image display. In the age of broadband, many have forgotten that a lot of users around the world still have crappy connections; a means to turn images off, but still leave small icons intact and easily display images on demand would be helpful. It's also a (very) partial and (hopefully) uncontroversial answer to the perennial image filter question by giving users more control over what they see (see m:Controversial_content/Brainstorming#KISS).

So, is anyone willing to volunteer to create a gadget that will

  1. put an images on/off switch at the top right of every article, defaulting to images on.
  2. in images off mode, leave the image space blank, and show images when hovering over them
  3. allow logged-in users to change their preferences, so images can be off by default. (Gadgets can do this the way Twinkle does.)

Anyone? Rd232 talk 10:58, 13 August 2012 (UTC)

Typically you'd want to do this on the php side, since images may start to download before your gadget runs (although if you're on a slow connection, probably not very much of the image would be downloaded before the js runs). [On the PHP side, you start to get into the whole splitting the parser cache issue, but if its just a binary switch, that's probably not too horrible given how much we already split the parser cache. However getting a MW extension written and enabled for something like this, given the processes involved would probably be a challange]. Bawolff (talk) 14:58, 13 August 2012 (UTC)
Thanks. I'm a bit surprised it's not possible to block image downloads in JS or even CSS; but it seems not. I wouldn't mind a JS gadget nonetheless - and setting visibility:hidden on page-load and back to normal on hover should be easy enough for someone who can do Javascript. Rd232 talk 23:09, 13 August 2012 (UTC)
Readers can already adjust their browser settings if they do not want to download any images. As for permitting icons to download, the difference between an icon and a modestly-sized image is arbitrary. JavaScript and CSS files use similar bandwidth to images on most Wikipedia pages,[citation needed] most Wikipedia pages do not contain many large images, and almost all images are displayed in reduced thumbnail size in articles. So, would a gadget really provide any significant benefit? — Richardguk (talk) 21:03, 13 August 2012 (UTC)
Most readers don't know how (and at best it's inconvenient). Icons are usually <50px, and images >=100px. JS and CSS are cached locally (aren't they?) so would only need to be loaded once per WP visit, not once per page (I think...). Not sure that JS/CSS is really that heavy either - are you thinking of the server-side rendering time? As for KB of a page: well Barack Obama is 200kb, and the 20-odd images (mostly 12-25kb) are probably double that again. Most pages have many fewer images, but it gives you some idea that an article's images may be a non-trivial burden on a sufficiently poor connection. Rd232 talk 23:09, 13 August 2012 (UTC)
Most readers also don't know how to reset their cookies, but we don't include a button for that either... Think of all the malware they could be exposed to. I'm not a fan of this kind of argumentation. Not saying that someone shouldn't make it, by all means do, but that argument is weak. —TheDJ (talkcontribs) 08:12, 14 August 2012 (UTC)
Malware? Via Wikimedia cookies? What? Rd232 talk 16:34, 14 August 2012 (UTC)


What changed with the way rollback was displayed? I accidentally made this edit for this reason. I'll certainly know in the future, but it looked like it was part of the edit summary to begin with. Ryan Vesey 19:01, 13 August 2012 (UTC)

What's most curious to me is that the rollback feature now has three distinct appearances, one for my contributions, one for my watchlist, and yet another for article histories. Someguy1221 (talk) 19:05, 13 August 2012 (UTC)
It is the result of bugzilla:18834. —TheDJ (talkcontribs) 19:35, 13 August 2012 (UTC)
Hmm, it seems like a beneficial change, do you think it would make sense for a bugzilla report to be put in for the rollback link to appear after the undo link? That would correct the problem of it appearing as part of the edit summary (or maybe I'm the only one). It is more of an issue when a summary wasn't used to begin with. Ryan Vesey 19:39, 13 August 2012 (UTC)
What undo button ? I don't have an undo button in my user contributions or watchlist.. (i'm an admin, so that might be related). Can you make a screenshot of what you see now and accurately describe how it looked before ? —TheDJ (talkcontribs) 19:52, 13 August 2012 (UTC)
It'll be a while, but I'll make one. Ryan Vesey 19:53, 13 August 2012 (UTC)
Note that the undo button only appears in page history. Ryan Vesey 19:54, 13 August 2012 (UTC)
The reason that the rollback is different (does not have the count) on RecentChanges and Watchlist is because getting the number of edits on those pages is a rather expensive operation. It is therefore disabled on WMF servers. —TheDJ (talkcontribs) 20:07, 13 August 2012 (UTC)

The old rollback links can be restored by placing this javascript in one's skin.js.

tmp = "";
var rollback_links = document.getElementsByClassName("mw-rollback-link");
for (var ind = 0; ind < rollback_links.length; ind++) {
    tmp = rollback_links[ind].innerHTML.replace(/rollback ?(more than)? [0-9]{1,2} edits?/, "rollback");
    rollback_links[ind].innerHTML = tmp;
delete tmp;

Σσς. 21:39, 13 August 2012 (UTC)

Alternatively, one can add
$('.mw-rollback-link a').text('rollback');
to personal common.js. — AlexSm 03:34, 14 August 2012 (UTC)

I find it annoying that we're suddenly misusing "rollback" as a verb. (We roll back edits; we don't "rollback" them.) Is there some way to insert a colon ("rollback: 5 edits")? —David Levy 20:38, 14 August 2012 (UTC)

What is GeoHack?

Not sure what I did, except maybe I was moving my mouse over a map on Loyal Valley, Texas. It went to GeoHack. This is the page it went to:,_Texas&params=30_34_33_N_99_00_28_W_type:city(50)_region:US-TX

What is that? Maile66 (talk) 00:13, 14 August 2012 (UTC)

See the page about the tool on the Toolserver wiki. As it says on the page there, it provides links to mapping services from coordinate links on Wikipedia. Graham87 01:31, 14 August 2012 (UTC)
That interwiki prefix doesn't work (it's going to, and only displays their version of MediaWiki:Noarticletext. Nyttend (talk) 03:21, 14 August 2012 (UTC)
Never mind — it's, which can be found at tswiki:GeoHack. The differences between this and the other links that we gave before are capitalisation and "tswiki" versus "tsswiki". Nyttend (talk) 03:28, 14 August 2012 (UTC)
Whoops, sorry. :-) Graham87 03:48, 14 August 2012 (UTC)
Anyway, to return to the original q: Maile66, the link that you posted beginning is exactly the same as the external link given in the coordinates of the article, both those in the infobox and those that appear at upper right. In this case, both of these are generated by code within {{Infobox settlement}}, but you can achieve almost exactly the same result with {{coord|30|34|33|N|99|00|28|W|type:city(50)_region:US-TX|display=inline,title}}. Coordinates links like these are a perfectly normal feature of most Wikipedia articles describing a place, structure or other fixed point. They take you to a page on toolserver from which a number of mapping services may be selected; these are configured at Template:GeoTemplate. See also WP:GEO. --Redrose64 (talk) 13:32, 14 August 2012 (UTC)
Thanks. I must have clicked on the coordinates or something. But I must say, it's a little alarming to have your browser suddenly say "GeoHack", in this day and age of worms and viruses. Maile66 (talk) 13:38, 14 August 2012 (UTC)

Watchlist gadget

Now that I've finally gotten around to closing the WP:Watchlist RfC (yes I participated, but the consensus was obvious, and no one else ever got around to it), could an admin please add a gadget to make unvisited watchlist changes bold? One of the concerns expressed in the RfC was that users would never find the feature if they had to manually change their CSS. David1217 What I've done 00:35, 14 August 2012 (UTC)

Doing...... Edokter (talk) — 16:25, 14 August 2012 (UTC)
Yes check.svg Done, but I regard it as a temporary placeholder. I still feel that supressing default software behaviour is not mandatable by community consensus. Sine the CSS for the watchlist has changed, there is a whole new range of possible styling options. Right now, the functionality is disabled for all new editors just becuae the few dislike the bold styling. The discussion should be reopened. Edokter (talk) — 16:50, 14 August 2012 (UTC)
Thank you Edokter! I feel the same as you, but I'm not going to go against an RfC with ~150 participants. And one small request: could you include a link to WP:CUSTOMWATCH in the gadget description? David1217 What I've done 18:23, 14 August 2012 (UTC)

New MW installation

Is there somewhere that lists the changes made in the latest edition of Mediawiki? I noticed that we upgraded because of the rollback change that's discussed above (the funny thing is that it was installed several days earlier at Commons), but I've not yet seen anything else that's different. Regarding the sorting-in-tables issue, the previously sortable tables in List of Indiana state historical markers in Allen County and tables with the same coding (including all of the county lists linked by List of Indiana state historical markers) are not sortable in IE8, but the same was true in the previous edition of MW. Nyttend (talk) 03:24, 14 August 2012 (UTC)

See mw:MediaWiki 1.20 and the section #Table sort problem above. — AlexSm 03:40, 14 August 2012 (UTC)
The table at List of Indiana state historical markers in Allen County and the other counties are simply not a properly formatted table, that's why they don't work. Nothing IE8 specific or 1.20wmf9 update related. I fixed a few of them, and will leave the rest up to other editors. —TheDJ (talkcontribs) 07:57, 14 August 2012 (UTC)
Confused, why did you point me to the table sort problem, since I had just referred to it? At any rate, when I produced the lists back in March, the tables were sortable. Nyttend (talk) 20:24, 14 August 2012 (UTC)
When multiple issues arise at the same time, things often get confusing for all parties. I can't say much about the behavior in march, you might be right. I do know that the table headers of those tables are not properly formatted to get the behavior that you were seeking. —TheDJ (talkcontribs) 21:09, 14 August 2012 (UTC)

Including vertical bar as part of #if ParserFunction

In Template:FoP-USonly, I am transcluding Template:Do not move to Commons, which has an optional parameter expiry. I want to have an optional parameter 2 which is passed into expiry if it exists. More specifically, I want to create an #if statement such that if {{{2|}}}, then produce |expiry={{{2}}}, else produce nothing. The problem is with that the vertical bar is also used to delimit the then/else portions, so it can't be included as part of the actual expression inside the then portion. -- King of ♠ 05:37, 14 August 2012 (UTC)

You can assign a parameter in either case but make it an unused parameter name like expirydummy with the technique in meta:Help:Template#A parameter value depending on parameters. PrimeHunter (talk) 10:40, 14 August 2012 (UTC)
Thanks! -- King of ♠ 16:35, 14 August 2012 (UTC)

Noticeboard tool

Could someone create a tool (or is there a tool) that allows you to report someone to a noticeboard and inform them? I'm thinking similar to how you can report sockpuppetry using Twinkle. Be on their user page, grab it from a drop down, choose the noticeboard (ANI/ANEW/ETC.), type the reason, and choose editors to notify. Sphilbrick has drafted a similar idea for the ANI edit window, but I'm thinking of something that doesn't even require you to go to the noticeboard first. Ryan Vesey 18:42, 14 August 2012 (UTC)

site width

why dont u limit site width to like 800px? reading an article in 2000px width isnt comfortable. if yahoo limits the width of their site, you can too... can't u? — Preceding unsigned comment added by (talk) 19:15, 14 August 2012 (UTC)

New watchlist feature

Hey all. So, we had an AFT5 deployment earlier, and one of the new features allows you to see feedback from articles you're watchlisting, via a blue link at the top of the watchlist. Hope people find this helpful; if you have any problems, give me a shout! Thanks :). Okeyes (WMF) (talk) 21:20, 14 August 2012 (UTC)

No user contributions link for this editor

When I am on User:Engineer+Attorney or User talk:Engineer+Attorney, there is no link to user contributions. Does anyone know why? Ryan Vesey 04:32, 15 August 2012 (UTC)

I figured it out, user account doesn't exist and was created by another user. Ryan Vesey 04:32, 15 August 2012 (UTC)

Cite_quick 12x faster for short cites

New Template:Cite_quick is a fast-cite alternative to {cite web} or {cite_book} or {cite news}, to produce the same format as Citation style 1 for the common parameters, but run 10-12x times faster, to allow major articles to edit-preview, or reformat, within 10 seconds, rather than 27-45 seconds or such. The NewPP preprocessor can format over 6,000 transclusions of {cite_quick} to avoid the error "Template include size is too large" when using {cite_news} or other templates based on Template:Citation/core. In prior years, other "quick" templates have been created, but they did not specify which parameters are supported as listed in Template:Cite_quick/doc.

Some articles hit template limits: Currently, the article "Arab Spring" (viewed 150,400 times per month) hits the include-size limit of 2,048,000 bytes (2000 kb), where the final use of Template:Navbox dies during expansion. Instead, by using template {cite_quick}, then all the templates will fit, and the reformat time will drop from 41 seconds to 11 seconds, where formerly, the edit-preview required 41 seconds to re-display the changes in progress. Notice how the trial revision using {cite_quick} will display much sooner than the typical 41-second reformat of prior revisions:

Look for other over-sized pages or articles with include-size errors in:

Most problems are in user pages, but many articles are on the brink of generating include-size errors.

Background and parameters: After extensive discussions of the new Template:Fcite, some users had recommended a short template which specifically lists the few parameters it would support, and clearly indicate which parameters are NOT allowed. That short template is, now, Template:Cite_quick, and it can run 10-12x times faster by omitting all the rare parameters. Only the following parameters are supported, as in major pop-culture articles:

  • title, work, journal (or: newspaper)
  • author, author2, author3, author4
  • last, first, last1, first1, last2, first2, last3, first3, last4, first4
  • publisher, agency, location, volume, issue, chapter, page, pages
  • url, archiveurl, archivedate, isbn, issn
  • date, year, accessdate

Meanwhile, the various Fcite templates, such as {{Fcite_journal}}, are being tested to support more of the rare parameters 5x faster, such as "ref=harv" and all the journal codes (arxiv, ASIN, bibcode, DOI, JSTOR, OCLC, PMC, PMID, RFC, SSRN, ZBL, etc.). Although {cite_quick} can handle most cites in common-culture articles, for medical articles or text using "ref=harv" links, then {{Fcite_journal}} or {{Fcite_book}} would be the fast alternative to handle numerous special parameters, such as editor names, JSTOR links, or PMID id numbers. In prior years, the need for faster citations had been emphasized, but some users think that the older {cite_web} or {cite_news} templates can be merely "tweaked" to suddenly run 10x faster and no longer hit the template-include-size limits. As I commented elsewhere, it is like trying to update a "covered wagon" to move like a sports car, as if a few simple changes could make the covered wagon (and all its numerous parameters) speed along at 80 miles per hour (129 km/h). For such a drastic improvement, then extensive changes would be needed. Restructuring large templates requires months of analysis and rewrites. -Wikid77 (talk) 22:43, 10 August 2012 (UTC)

  • Ohhh no... not another fork of the cite templates. We've only just finished dealing with the last lot. --Redrose64 (talk) 22:53, 10 August 2012 (UTC)
This is a documented template, {cite_quick}, and the others were undocumented {cite_book_quick}, where users wondered which parameters were supported, now explained with this template. -Wikid77 (talk) 16:31, 11 August 2012 (UTC)
  • What Redrose64 said. Please someone send this to deletion and stop forking the citation templates. Headbomb {talk / contribs / physics / books} 22:59, 10 August 2012 (UTC)
Yeah, lets keep it as it is. Why improve at all? Miami knows. -DePiep (talk) 01:42, 11 August 2012 (UTC)
Because it is not a proven improvement, maybe? It is a reduced functionality template, which is being compared, by one editor, who is the sole author and proprietor thereof, to other templates. Based on his own statistics (which could be the result of hours of rigorous testing by numerous disinterested parties but aren't) he then presents as fact that his "wiki-invention" is the finest thing since sliced bread. Everyone goes "wow" and deploys these across the wiki ... however the last time this happened, with the fcite template it broke prettty much every article it was used on. So let's not get our hopes up. pablo 01:52, 11 August 2012 (UTC)
"Restructuring large templates requires months of analysis and rewrites," which the template author apparently hasn't done. The next step in implementing this would a 100-article test, followed by review, followed by a 1000-article test, followed by an extensive review. — Arthur Rubin (talk) 03:09, 11 August 2012 (UTC)
Articles to be chosen by others. No offense indented, but selection bias is quite possible if the software author selects the test suite. — Arthur Rubin (talk) 14:23, 11 August 2012 (UTC)
Moot issue; the whole concept of reduced functionality is wrong. Br'er Rabbit (talk) 15:21, 11 August 2012 (UTC)
Why not to go directly to mw:Lua scripting? The latest status update says the deployment starts this month on Some people already started creating prototypes on for some other templates. Helder 12:27, 11 August 2012 (UTC)

I believe the editor created this to help eliminate the large number of pages that take a long time to load with the existing Cite templates. Some take a very long time to load unless you have a high speed internet connection which not everyone has. Additionally there were a lot of articles getting the "Template include size is too large" error. If someone has a way to fix either of those 2 problems with the current template (maybe changing the software to increase the template size? Then great bring them up and lets talk about it. Kumioko (talk) 14:47, 11 August 2012 (UTC)

  • What Redrose64 said. This is moar IDHT re the fcite TFD. Improve the standard templates. Br'er Rabbit (talk) 15:16, 11 August 2012 (UTC)
Yes it would be nice to modify the existing templates but we can only do some much there. The fact is they use a lot of parser functions and many templates use a lot of citations which causes overload. When we start using Lua it will help a lot but no idea when that will actually be. For what its worth there is a lot of IDHT on both sides. Both have good points and bad. Kumioko (talk) 15:28, 11 August 2012 (UTC)
There's a long history here. See:
Wikid77 is well into disruptive territory. He's not engaged the technical people involved in the citation templates. There's plenty that could be done. This isn't helpful at all. Br'er Rabbit (talk) 15:39, 11 August 2012 (UTC)

I must agree with the Rabbit. Incredibly disruptive, just like those terrible people who make new articles. Now that my sarcasm is out of the way, let me point out that Lua isn't live (or, if it is, no one told me), so it is not a valid rationale against anything. A test run would be a good idea, but extensive testing (and repairing when necessary) didn't help the last time. (The last set of templates was illogically relegated to testbed status, completely eliminating any useful possibility for the templates.) Wikid is not disrupting anything. Calling these actions disruptive is not just illogical, but incredibly insulting (and not just to Wikid, but to other editors here). Slapping these on over the standard citation templates (eliminating them) would be disruptive. Providing options is not. I wonder if some consider it disruptive when a fast food establishment gives the option of multiple meal sizes. This is no more so. --Nouniquenames (talk) 16:10, 11 August 2012 (UTC)

  • These templates are a stepping stone to better overall operation: While articles such as "Arab Spring" can reformat in 11 seconds, rather than 40-50 seconds, those cases number only in the low thousands (nation articles, major cities, top films, pop-culture, etc.). Long-term, the intention is to improve {Cite_web} or {Cite_news} to run somewhat faster, so that all news-related articles can reformat, or edit-preview, about 2x faster. That means, when clicking any prior revision, it will display 2x faster than now, but only because these recent templates show techniques for how to format citations much faster. More below, at: #Cite_web/smart hybrid of Cite_quick. -Wikid77 (talk) 16:31/17:46, 11 August 2012 (UTC)
That will be pretty soon, actually. It was just enabled on and should be installed on by next week. One of Rob's suggestions was to import a big article and try to reduce the parsing time by converting templates such as Cite to Lua. Helder 12:10, 15 August 2012 (UTC)
AFAIK the plan is to let it rest on for a while, during which time (a) en.wp can decide if it wants it and (b) the code for a cite replacement can be written. I wouldn't expect it for a couple of months, in other words - even if the deployments go to plan. - Jarry1250 [Deliberation needed] 12:47, 15 August 2012 (UTC)
I thought the plan was to let it rest on for a while so ops can make sure it doesn't overload the servers and so it can be seen whether it really speeds things up as everyone expects. As for (a), why wouldn't we want it (if we're even getting the choice)? If nothing else, killing the massive hacks in Category:String manipulation templates would be worth it. Anomie 16:33, 15 August 2012 (UTC)

Are the HTML generators out of sync on some servers?

I have noticed a strange effect at Euston railway station: in the infobox, the map caption sometimes appears to the left of the map, not below it. This seems to be a direct effect of the two bulleted lists "Lists of stations" and "External links" (each of which is formatted as WP:HLIST) not being processed properly: the space after the bullet is absent, so the list items bunch together, which forces the infobox to be over-wide. Examining the page source shows that invalid HTML has been served. Without getting into the details, the tags (but not their attributes) for the "Lists of stations" and subsequent rows are as follows:

How it appears how it should appear (obtained from London Paddington station)
    <th>Lists of stations</th>
        <li><a>National Rail</a></li>
        <th>External links</th>

at point (*) there are various correctly-nested <a>...</a> <br /> <img /> and <span>...</span> but the <br /> is misformed as <br/>

    <th>Lists of stations</th>
        <li><a>National Rail</a></li>
    <th>External links</th>

at point (*) there are various correctly-nested <a>...</a> <br /> <img /> and <span>...</span>

Served by mw4 also mw44. This one also contains a number of &#xfeff; characters. Served by srv240. This one lacks &#xfeff; characters.

Somehow, mw4 & mw44 are failing to insert closing </li></ul> at the proper places, but putting unwanted ones later on. It appears to me that the software on srv240 is more up-to-date than that on mw4 and mw44. So, I believe that the HTML generators out of sync on some servers. --Redrose64 (talk) 20:24, 14 August 2012 (UTC)

This sounds like it could be bugzilla:38273TheDJ (talkcontribs) 21:15, 14 August 2012 (UTC)
huh that is interesting, why would there be BOM's, even with tidy disabled... Bawolff (talk) 14:54, 15 August 2012 (UTC)

Need help with a template

I'm wondering if someone with template skills could update {{Infobox NASCAR race report}} such that:

  • the "Network" line is only displayed if the {{{Network}}} parameter is specified,
  • the "Announcers" line is only displayed if the {{{Announcers}}} parameter is specified,
  • the "U.S. Ratings" line is only displayed if the {{{U.S. Ratings}}} parameter is specified, and
  • the "Television" heading is only displayed if one or more of the above parameters are specified.

I had a go myself, but kept ending up with extra blank lines being displayed unless all three parameters were specified. Thanks in advance. P.S. If you feel enthused to update the template to use {{Infobox}}, go right ahead! DH85868993 (talk) 01:41, 15 August 2012 (UTC)

I don't think I can help much right now but you can try some experiments if there is a sandbox version of the template. One of the lines may look like this...
|- {{#if: {{{Network|}}} ||style="display:none;"}}
This should make the row disappear if {{{Network}}} is not set. Maybe when things cool down I may be of more help. – Allen4names (IPv6 contributions) 02:45, 15 August 2012 (UTC)
Per Wikipedia:Template sandbox and test cases, I've created a sandbox version, and some testcases. Is that what is required? --Redrose64 (talk) 10:40, 15 August 2012 (UTC)
Yes, that's great. Thanks! DH85868993 (talk) 10:55, 15 August 2012 (UTC)


I found this comment in feedback: "i totally appreciate the works of wikipedia , but it would be much better if the background is of some other color as the white color doesn't appear to be suitable for long time reading". Is it possible to create an alternative background that can be used without creating an account? Even if we made the green on black work with vector and put a button somewhere on the page to use it. Ryan Vesey 14:53, 15 August 2012 (UTC)

Most web browsers have a (Rather hidden) feature of allowing to override background/foreground colours of websites. Additionally we could use alternate css, but that is neiter a well-supported feature nor an easily discoverable feature by web browsers (For an example of how that works, go to then in firefox go to View→Page Style→<select some option>). Beyond those options, could use JS gadget to have a link at the bottom, "View page with dark background" and make sticky with cookie I suppose. Note that from what I understand most studies say black text on white background is easier for reading than green on black, or other light on dark colour schemes. Bawolff (talk) 15:00, 15 August 2012 (UTC)
Thanks, it might come down to decreasing the brightness on their computer screens, which is done on their end, not ours. Ryan Vesey 16:10, 15 August 2012 (UTC)
We also have the black skin Gadget —TheDJ (talkcontribs) 18:22, 15 August 2012 (UTC)
Yeah I quite agree Ryan, worse on my widescreen too, makes the site look most unappealing. I frequently change fonts and skins, there ought to be more choice.♦ Dr. Blofeld 18:46, 15 August 2012 (UTC)

User:Ohconfucius/script/MOSNUM dates

Does anyone know how to modify that script to remove this error I found? I decided to check 20–20–20 club using the script. No changes needed to be made, but it tried to change every instance of {{MLBY|1928}} to 1928 (the same thing occurred with other years. Can it be modified to ignore that template? Ryan Vesey 16:10, 15 August 2012 (UTC)

The function 'function ohc_delink_year_in_X()' contains two regexes for templates. Your should remove '|mlby' from them. Ruslik_Zero 16:42, 15 August 2012 (UTC)
I have no access to the script, can you make that change? Ryan Vesey 16:50, 15 August 2012 (UTC)
User:Ohconfucius/script/MOSNUM dates#Overview 2 states that the script is designed to remove year linking templates. Since Ohconfucius has retired, I suggest you make a copy of the script and modify it to suit your needs. (This is what I've done with User:GoingBatty/script/Sources.js). Good luck! GoingBatty (talk) 16:56, 15 August 2012 (UTC)
It would be best if someone adopts the script, then we can let the 'old' version load the new maintained version. That is better for the long term. —TheDJ (talkcontribs) 18:21, 15 August 2012 (UTC)

Bugs in diffs

Hi, could I please beg someone who has some influence at this project to push for the ridiculous bugs in the diff generator, that have been outstanding for probably ten years, to be fixed. This is a typical latest example:

Large identical or near-identical sections flagged as completely changed makes it very much harder to check for vandalism and bad edits. (talk) 23:00, 15 August 2012 (UTC)

Database lag

Rising steadily for the past few hours. Up to 681 seconds now. Equazcion (talk) 10:19, 2 Jul 2012 (UTC)

Any info on when this might be resolved by? Jenks24 (talk) 10:48, 2 July 2012 (UTC)
It took a minute for the page to load. This is getting out of hand.—cyberpower ChatOffline 11:13, 2 July 2012 (UTC)
I just saw it go over 800 seconds! Roger (talk) 11:32, 2 July 2012 (UTC)
Over 900 seconds now. Counter-vandalism editing gets rather hard when it takes more than 15 minutes for an edit to appear in your watchlist... Fram (talk) 11:52, 2 July 2012 (UTC)
942 seconds now. Editing in general gets hard when your watchlist doesn't want to load.—cyberpower ChatOffline 11:56, 2 July 2012 (UTC)
If you want the lag message to display minutes/seconds also, I just made a script for that -- see User:Equazcion/LagToMinutes :) Equazcion (talk) 12:02, 2 Jul 2012 (UTC)
It doesn't work right. I'm getting 280000 minutes and 21 seconds.—cyberpower ChatOffline 12:34, 2 July 2012 (UTC)
1030 1057 1133 this a record?--Gilderien Chat|List of good deeds 12:21, 2 July 2012 (UTC)
Caught is at 1000 even. W0OT! Is the number a 32-bit variable or a 64-bit variable? We wouldn't want an overflow after 9.10194444 hours... :) --Guy Macon (talk) 12:23, 2 July 2012 (UTC)
It's no record :( Highest I've seen is just over 9,000 seconds (!) Equazcion (talk) 12:27, 2 Jul 2012 (UTC)
Is there a point at which editing will (have to) be shut down to protect the integrity of the database? What are the server geeks saying about the problem? Roger (talk) 12:28, 2 July 2012 (UTC)
Well toolserver replag (edit counter) is at "4 hours, 24 minutes, 14 seconds" --Gilderien Chat|List of good deeds 12:45, 2 July 2012 (UTC)
PS. If anyone had tried my script above, it had a bug when the seconds went over 1,000, but it's been fixed now. Equazcion (talk) 13:02, 2 Jul 2012 (UTC)

Operations is now aware of the replication lag problems. They are investigating. You might want to keep an eye on server admin log with regards to resolution of the issue. —TheDJ (talkcontribs) 13:04, 2 July 2012 (UTC)

It's largely a false alarm, although it does seem to cause very slow loading of watchlists. IIRC there are five database slaves, and only one of the five is actually lagged; the other four are fine, so most of the time you will actually get an up-to-date response from the server. --R'n'B (call me Russ) 13:25, 2 July 2012 (UTC)
So far every single time I've reloaded my watchlist, the lag message appears accurate. The last change that appears in the list is from the time just before the lag amount showing. Equazcion (talk) 13:28, 2 Jul 2012 (UTC)
(/d) Yes, it's only db12 DB overview. BTW, check infra status to see how many problems operations is dealing with that you are actually NOT noticing as a user.  :D Makes you wonder how often they fix stuff and we never even thank them for it. —TheDJ (talkcontribs) 13:32, 2 July 2012 (UTC)
Hey I'm not complaining. I look at database lags kind of like forced relaxation :) Equazcion (talk) 13:37, 2 Jul 2012 (UTC)
Yes, if I look at my watchlist it only shows changes from about 40 minutes ago. But if I actually load a page it shows the current version, not a 40 minute old version. Looks like the code is written so that Special:Watchlist always reads from the slowest slave. --R'n'B (call me Russ) 13:52, 2 July 2012 (UTC)
That's the way it always is with database lags. I've never seen the lag actually cause old versions of pages to show up. It just affects the watchlist display (far as I know). Equazcion (talk) 13:54, 2 Jul 2012 (UTC)
"just affects"? Heh, for me, the watchlist is god. Having it work this badly (and it was having problems yesterday, too) is heresy.--Bbb23 (talk) 15:11, 2 July 2012 (UTC)

Lag reached 1,483 seconds. --Meno25 (talk) 13:36, 2 July 2012 (UTC)

Is it just me or has this been happening a lot more since WP switched from Godaddy to the new provider? Kumioko (talk) 13:42, 2 July 2012 (UTC)
I doubt there's any correlation, since that switch only affected who the domains were registered with. Which has absolutely nothing to do with the databases. the wub "?!" 15:46, 2 July 2012 (UTC)
Agreed was about to say the same thing.—cyberpower ChatOnline 15:53, 2 July 2012 (UTC)
  • @Equazcion: the script doesn't seem to work in Monobook, but I think you knew it might not. Bishonen | talk 14:57, 2 July 2012 (UTC).

Loading my watchlist: <!-- Served by mw3 in 30.838 secs. -->

Now 2027 seconds. This could be related? If they are changing the times on all the edits made so far today?--Gilderien Chat|List of good deeds 15:28, 2 July 2012 (UTC)
I was about to ask if the leap second could have caused this as well.Edinburgh Wanderer 15:42, 2 July 2012 (UTC)

Update: We now at 36 min. lag and loading the watchlist reports <!-- Served by mw33 in 30.903 secs. -->cyberpower ChatOnline 15:48, 2 July 2012 (UTC)

The cause is not really known yet. The server is running some scripts to get the last remaining Sha1's of revisions populated, but that should not take that much resources to cause this. Another operations guy should be waking up soon and hope is that he can properly debug the problem. The lag has been going down a bit the past half hour though (without apparent cause). Unfortunately this db server is also serving up all the watch list requests, so it's a bit annoying for the editors. —TheDJ (talkcontribs) 16:04, 2 July 2012 (UTC)
It's not just lag. Getting many "general error" screens when trying to log in or edit. North8000 (talk) 16:25, 2 July 2012 (UTC)
(edit conflict) It's not just watchlists that are affected, but contribs lists too. For example, Special:Contributions/Redrose64 shows my most recent edit as being to Cardiff Central railway station at 15:48 UTC, but I'm certain that I've made at least six edits since (to Swansea railway station, Llanelli railway station, Carmarthen railway station, Whitland railway station, Template:POV, User talk:MSGJ). --Redrose64 (talk) 16:26, 2 July 2012 (UTC)
And page histories are similarly affected/afflicted, too. Bishonen | talk 21:23, 2 July 2012 (UTC).
  • Appears lag might be sloooooowly coming down? - The Bushranger One ping only 16:31, 2 July 2012 (UTC)
    Still taking longer to load though. <!-- Served by mw42 in 31.096 secs. -->cyberpower ChatOnline 16:55, 2 July 2012 (UTC)

Update: Replag decreasing at a rate of about 15 seconds per minute. I estimate 4 hours, 27 minutes, and 44 seconds until replag is cleared at current speed.—cyberpower ChatOnline 17:08, 2 July 2012 (UTC)

Now lag is 1,255 seconds. --Meno25 (talk) 17:14, 2 July 2012 (UTC)

Replag down to 10 seconds now. Yay. Now to fix the errors.—cyberpower ChatOnline 19:07, 2 July 2012 (UTC)

Toolserver replag

OK, up-to-date watchlists and contribs, at l(e)ast(!), which is good; but still an ever-growing replag on the toolserver:

Caution: Replication lag is high, changes newer than 6 hours, 42 minutes, 25 seconds may not be shown.

What is causing this and why is it not decreasing at all for days? Jared Preston (talk) 17:48, 3 July 2012 (UTC)

Update: "Caution: Replication lag is high, changes newer than 9 hours, 18 minutes, 10 seconds may not be shown."
Why is the replag getting higher? Does anyone know when/if the server-problem will be fixed in the next few days? Jared Preston (talk) 20:15, 4 July 2012 (UTC)
No idea what its means but tool server says High replag because of inserting of many SHA1-hashes. Its getting worse rather than better.Edinburgh Wanderer 22:08, 4 July 2012 (UTC)
See, I don't understand it either. The only hashes I like are these ones. Maybe someone has dropped some on the server, now up to 10 hours, 33 minutes, 20 seconds. Jared Preston (talk) 06:22, 5 July 2012 (UTC)
You should try corned beef hash some time.  :-) But this is the "technical" discussion area, and it is a rather technical issue. The Wikipedia database that stores old page revisions had a new field added to it recently to make it easier to determine whether two past revisions are identical or not, so that we can identify reverts. Now the servers are in the process of filling in this field, which previously had been blank, and all these changes need to be copied to the Toolserver (which is a separate system from Wikipedia). The flood of changes is overloading the Toolserver and causing the high replag. Unfortunately, according to the Toolserver admins, it could take another week or two for this process to finish. --R'n'B (call me Russ) 11:55, 5 July 2012 (UTC)
Just so you guys know, this isn't happening with me. Hadger 18:58, 5 July 2012 (UTC)
Hadger, can you confirm that by checking this link? Thanks for the help, Russ – I can cope (just about) with a week to ten days! Jared Preston (talk) 19:48, 5 July 2012 (UTC)
Oh. That's what you were talking about. I thought that you were talking about watchlists and contribution pages having that appear. :) Hadger 19:30, 7 July 2012 (UTC)
It's now on the way down, having peaked at around 10:00 UTC today with just under 140000 secs (38+ hours), the current figure is about 130000 (36 hours). --Redrose64 (talk) 21:10, 7 July 2012 (UTC)
It appears to have been steadily going up all month, and now at over 160,000 (44+ hours). However, the edit counter tool lists only 16 hours. Why the discrepancy? GoingBatty (talk) 03:19, 10 July 2012 (UTC)
There are two database servers for enwiki on the Toolserver (rosemary and thyme), and they are lagged to different degrees:
[2:31pm] <Earwig> @replag
[2:31pm] <tsbot> Earwig: s1-rr-a: 17h 13m 22s [+0.26 s/s]; s1-user: 2d 34m 37s [+0.22 s/s]
The Earwig (talk) 18:34, 10 July 2012 (UTC)
Thanks for the explanation. Looks like the lags are continuing to increase and are now over 18h 41m and 2d 19h. On July 5, Russ said "according to the Toolserver admins, it could take another week or two for this process to finish." Is there any update from the admins? Thanks! GoingBatty (talk) 16:19, 13 July 2012 (UTC)
Not from the admins, but I've done my own analysis. Unless something is done about the cause, the replags are likely to continue increasing at least until the end of July, and probably for a week or two after that on s1-user. This is all due to a massive database update; eventually, it will get to the end of the database, but in the meantime all Toolserver tools are suffering. --R'n'B (call me Russ) 23:43, 13 July 2012 (UTC)
The replag has been racing up, in spurts, for the last 48 hours or so. When will this upward spiral end? Please help! Jared Preston (talk) 15:50, 19 July 2012 (UTC)
Certianly be good if we could get an official update on this.E W 18:56, 20 July 2012 (UTC)
Although the replag graphs continue to climb, the notice at the top of X!'s Edit Counter has now disappeared. This time yesterday it was showing 14 hours and some minutes, a figure which hadn't changed by more than two hours in over a week. I suppose that this means that rosemary (or thyme?) has caught up, but thyme (or rosemary?) hasn't. I've looked at the diagram and lists on tswiki:Servers but it's not at all clear which is which. --Redrose64 (talk) 15:57, 25 July 2012 (UTC)
<Dispenser> @replag
<tsbot>     Dispenser: s1-user: 1w 1d 6h 10m 35s [+0.53 s/s]; s2-user: 11s [-0.00 s/s]; s3-rr-a: 24s [+0.00 s/s]; s3-user: 24s [+0.00 s/s]
thyme is default server and is on duty as the s1-user server which hosts user databases and is slowly drifting behind. rosemary is on duty as s1-rr server, which does not have access to user created databases or temporary tables (actually, use ENGINE=MyISAM when defining temporary tables to workaround InnoDB permissions bug). — Dispenser 17:01, 25 July 2012 (UTC)
Toolserver status
Last update 10:00, 21 June 2014 (UTC)
MySQL rosemary up
MySQL daphne up
MySQL yarrow up
Replag s1 0h 0m 0s
Replag s2 0h 0m 28s
Replag s3 0h 0m 21s
Replag s4 0h 0m 4s
Replag s5 0h 0m 2s
More status Available here
Currently it is over a week, the table I included is somewhat dynamic, updated hourly, and may be easier to read then the info Dispenser posted. For the record, a week has 168 hours. So as of this post, 199 hours is 8 days and 7 hours. Vegaswikian (talk) 18:45, 25 July 2012 (UTC)
The replag managed to go down to zero yesterday evening, as Redrose noted, but now it is climbing back up again (please see X!'s Edit Counter for reference). Currently above 10 hours! Jared Preston (talk) 06:52, 26 July 2012 (UTC)
This was due to apparently some configuration changes, see jira:MNT-1258. — Dispenser 17:23, 26 July 2012 (UTC)
Do we even have a rough idea how long it is likely to take for the other to start catching up.Blethering Scot 23:24, 30 July 2012 (UTC)
My educated guess: replag on thyme will continue to increase at least through August 5, and possibly a couple of days beyond that. After that, it should start to decrease, but how long it will take to recover completely is beyond my ability to predict. --R'n'B (call me Russ) 23:29, 30 July 2012 (UTC)
The main thing is it starts catching up, from the last time it wasn't actually to bad once that process started. The last time did they not reach that point around the same time seems to be a big gap this time, might be my memory of it.Blethering Scot 23:38, 30 July 2012 (UTC)
Perhaps a bit after August 5.  ;-) --j⚛e deckertalk 06:38, 7 August 2012 (UTC)
The Wikipedia:List of Wikipedians by number of edits numbers continue to get further and further out of synch; it looks like BernsteinBot is using thyme, with increasingly inaccurate results. The replag graph shows the numbers continuing to increase. BlueMoonset (talk) 05:19, 8 August 2012 (UTC)
Yeah, it's affecting some of the tools that help ferret out unreferenced BLPs, which is why I'm noticing. Plus side, it's making our efforts to erdicate unreferenced BLPs look effective, at least for the moment.  :) I'm not looking forward to the inevitable hangover, though. --j⚛e deckertalk 01:13, 10 August 2012 (UTC)
Does anyone have an updated estimate on when it will catch up? Anylinks to other boards/sites where it is being discussed, or is it simply a case of we just have to wait? The-Pope (talk) 16:24, 10 August 2012 (UTC)
Hallelujah! it's started trending down! The-Pope (talk) 07:18, 11 August 2012 (UTC)
Hooooray! --j⚛e deckertalk 14:30, 11 August 2012 (UTC)
Hmmm, what will arrive first, Christmas or a proper working Toolserver? The Banner talk 11:12, 15 August 2012 (UTC) getting grumpy of it
Appears toolserver looking at a replag of only two hours. Blethering Scot 17:40, 16 August 2012 (UTC)
But increasing. The sharp drop at about 11:20 today may be the result of this. --Redrose64 (talk) 18:48, 16 August 2012 (UTC)
That tool is broken, the roots switched both sql-s1-user and sql-s1-rr to rosemary. thyme is only catching up at 3x real-time, so it'll be 5 days before its back to sync. DaB. is working on getting fresh dumps which may happen before that. — Dispenser 21:00, 16 August 2012 (UTC)

Edit conflicts

I don't know if this is related to the above at all, but over the last day or so I've had some crazy edit conflicts. The two most memorable:

  • a.) the edit conflict page popped up showing no conflict, but having the orange "you have a talk page message" banner too. I clicked to see the talk page message (to a new tab) but no messages, though it sent me to an IP page and it showed an IP at the top, instead of me being logged in. But when I clicked on the "edit" button in the original window gain, it showed me as logged in.
  • b.) apparently had a three-way edit conflict, and though I got the conflict screen, the amazing thing was that my edit was already saved to the page (showing my full signature).

There have been several others conflicts of various types.

Not to mention seeing the wikimedia page fairly regularly.

And yesterday the database "read only mode" notice was up for several minutes.

Hope this helps. - jc37 17:23, 2 July 2012 (UTC)

And not only did I edit conflict with meno above (lol@ irony), but it took several minutes (with the wikimedia screen and even the browser "can't connect to server" screen for several minutes) in order to post the above. I connected to other websites just fine in the interim. - jc37 17:30, 2 July 2012 (UTC)

It wasn't an edit conflict but something wierd just happened at an RfA. I wanted to comment on an oppose and the entire page got messed up and then the server gave up so I had to wait before it came back. When I hit Rollback, it blanked the page and said it reverted it edits by me to the last version by a user that didn't even contribute to an RfA. Things returned to normal after I used a previous version and restored the RfA from there but I'm curious to know what that was especially the Rollback message that was created in the edit summary.—cyberpower ChatOnline 17:42, 2 July 2012 (UTC)

Weird; Special:Contributions/Flynnster14 shows just eight edits, and none of them are to Wikipedia:Requests for adminship/Zagalejo. At least ClueBot didn't come around and warn you for blanking pages :-) Nyttend (talk) 19:21, 2 July 2012 (UTC)
That's a plus, assuming ClueBot wasn't having issues as well. I can't figure out why Rollback acted the way it did. Maybe the foundation staff could figure it out.—cyberpower ChatOnline 20:27, 2 July 2012 (UTC)
I found this.—cyberpower ChatOnline 20:32, 2 July 2012 (UTC)
Any chance this is related to the leap second? --Guy Macon (talk) 20:32, 2 July 2012 (UTC)
No, leap second fun was yesterday. Max Semenik (talk) 22:19, 2 July 2012 (UTC)

"My preferences" tabs not working

Under "my preferences" none of the tabs across the top :- Appearance, Date and time, Editing etc. are doing anything. No matter which one I select it stays on "User profile". Is this just me, or a general glitch? - Arjayay (talk) 14:43, 21 July 2012 (UTC)

Mine are working fine. (Windows 7, Firefox 14) Dismas|(talk) 15:11, 21 July 2012 (UTC)
I think I saw some mention this a couple of days back - you're using Internet Explorer? Which version? - Jarry1250 [Deliberation needed] 10:47, 22 July 2012 (UTC)
IE8 (can't upgrade with XP) Tabs still not working - Arjayay (talk) 15:13, 22 July 2012 (UTC)
Have you tried a hard refresh of the page ? See bypassing your browsercache. —TheDJ (talkcontribs) 08:06, 23 July 2012 (UTC)
Yes - Refresh, reboot, remove temporary files, wait 4 days - still not working. - Arjayay (talk) 07:50, 24 July 2012 (UTC)
9 days now and it is still not working - any other ideas? anywhere else I should be reporting this? - Arjayay (talk) 15:30, 30 July 2012 (UTC)
I still haven't seen other people with IE8 confirm this, I'd like to see confirmations before reporting. —TheDJ (talkcontribs) 15:56, 30 July 2012 (UTC)

I have the same problem with Chrome (latest version for now). Adress in adress field changes, but nothing happens. Updating page shows needed page for a second, but then page Special:Preferences#mw-prefsection-personal appears again. --AJZ (talk) 09:03, 5 August 2012 (UTC)

I have discovered that my preferences are not working. I've recently acquired a new computer: Windows 7 with IE9. I did every instruction to the letter as indicated on the prior page to this one: I've cleared my cookies, refreshed, cleared cache, removed temporary internet files. logged in and out, fretted and fumed! The preferences tabs appear, but no matter which I click, I'm just moved further down my "User Profile" page. Fylbecatulous talk 03:42, 6 August 2012 (UTC)
It seems the problem is in wiki engine. --AJZ (talk) 04:48, 6 August 2012 (UTC)
And what, precisely, does "It seems the problem is in wiki engine" mean?
And, more to the point, is it likely to be fixed in the forseeable future? Arjayay (talk) 15:24, 6 August 2012 (UTC)
Yes, with my upgrade, I'm hankering to use a gadget or two that just laughed at my old XP. thanks Fylbecatulous talk 15:29, 6 August 2012 (UTC)
It means that this problem can only be solved by fixing MediaWiki engine, not browser or anything else. --AJZ (talk) 03:25, 7 August 2012 (UTC)
I use Chrome on XP, and have not noticed the issue. Chris857 (talk) 03:46, 7 August 2012 (UTC)

Those who experience the issue could mention the skin they use, if changing the skin and language helps (like this: ...Preferences?uselang=en-gb&useskin=modern) and if resetting all preferences helps. — AlexSm 04:05, 7 August 2012 (UTC)

I think I'm using Vector - but as I can't get the Appearance tab to work, I can't check this.
I followed the link above, but just get the front (User profile) page and none of the other tabs work.
I did consider trying "restore all default settings" - but if it doesn't work, I'm left in a far, far worse situation. - Arjayay (talk) 07:53, 7 August 2012 (UTC)
If it is any help, My preferences are just as stuck on Wikimedia Commons - Arjayay (talk) 08:02, 7 August 2012 (UTC)
This is the funniest thing ~ I feel now as I did when I worked in a control room and agents would call in with some nebulous problem for me to sort and solve (with no clues to parse). I too believe I'm using Vector, but I can't find out since I can't check my preferences. (Ditto negatory response following the link above) Neither can I change the one skin I have (unless there's a secret path to circumvent having to make alterations in "my preferences"). Is there? Who's on first? Fylbecatulous talk 16:09, 7 August 2012 (UTC)
Not important anymore but you could just switch to another browser to fix your preferences or use other projects. — AlexSm 16:37, 7 August 2012 (UTC)

Possible solution: in IE go to menu Tools and uncheck "Compatibility View". There is also a button in the to address bar that could be easily pressed by accident. — AlexSm 16:37, 7 August 2012 (UTC)

Thanks for the ideas - but none worked
As stated above - it doesn't work in Commons - and it doesn't work in "Other projects" either.
Compatability view was not checked - so I tried checking it - no change - so I unchecked it again.
Any other ideas? Arjayay (talk) 15:17, 8 August 2012 (UTC)
Even if you can't get to Preferences → Appearance, it's quite a simple matter to determine which skin you're using. Try out each of these links:

and see which one produces an unchanged view of this page. --Redrose64 (talk) 17:59, 8 August 2012 (UTC)

Aha! By taking a tour of the links provided above (nice skins), I've determined that I indeed have Vector. I had already experimented with compatibility view to no avail, BTW. Fylbecatulous talk 23:19, 8 August 2012 (UTC)
Try using F12 to bring up the IE developer toolbar after switching a tab failed, and see if you can find the error console somewhere. (I don't know much about IE). —TheDJ (talkcontribs) 08:51, 9 August 2012 (UTC)

By pressing control when clicking a "My preferences" tab, the tab opens in another window, where it can be altered and saved - and these changes are implemented. Hope this works for others. (I'm sure someone will say this is obvious, and I should have tried it last month)
However, it is unrealistic to hope other editors realise / re-discover this in the future, so the root cause of the problem still needs to be identified and overcome. Has a bugzilla report been filed? (sorry I don't understand how to do bugzilla reports) - Arjayay (talk) 08:59, 9 August 2012 (UTC)

That workaround actually works, yes indeed it does. Thanks so much; it wouldn't have been obvious to me... Fylbecatulous talk 13:49, 10 August 2012 (UTC)
Ah, I think I got it. This change was recently made. It might be the culprit. I have filed bugzilla:39192 to deal with this. —TheDJ (talkcontribs) 10:59, 9 August 2012 (UTC)
Thanks very much - If I read that report right, the change was tested on Chrome, Safari, Firefox and 3 versions of Opera, but not even one version of IE - surely Wikimedia shouldn't be allowing changes to code without testing on IE8 + 9? - Arjayay (talk) 11:50, 9 August 2012 (UTC)
That is just the original message, if you read the review comments you see that IE was tested at some point. —TheDJ (talkcontribs) 13:49, 9 August 2012 (UTC)
Thank you for filing a bugzilla report :) Fylbecatulous talk 13:49, 10 August 2012 (UTC)
I note the tabs are still not working - can someone add the "workaround" to the first (User profile) page, until the coding problem has been resolved? - Arjayay (talk) 08:53, 16 August 2012 (UTC)

Watchlisting sections

Is there a way to add a only a specific section to the watchlist, and not the whole page? →Yaniv256 talk contribs 22:55, 14 August 2012 (UTC)

No. There is a hope that Wikipedia:LiquidThreads will permit such a thing, but it is not enabled on Wikipedia. Someguy1221 (talk) 23:00, 14 August 2012 (UTC)
Would you say that such a feature could be implemented on the skin side? Is it something that I as a user-programmer am likely to be able to contribute?→Yaniv256 talk contribs 06:43, 15 August 2012 (UTC)
It couldn't be done that way, since MediaWiki, as implemented on Wikipedia, does not record on a revision which section was edited. Any gadget to extract that information would have to compare the diffs itself, so it would have to be a active piece of programming and not just something in your .js or .css page. Even the section link in the edit summary is rather meaningless to the software except as a piece of hypertext. Conceivably, you could feed your watchlist RSS into a client-side program that looks for edit summaries containing links to particular section headers. Someguy1221 (talk) 07:22, 15 August 2012 (UTC)
In addition to this, section links can be altered, or even faked. So long as users are able to alter the part of the edit summary between the /* */ markers, the section link cannot be 100% reliable. --Redrose64 (talk) 13:48, 15 August 2012 (UTC)
Even if users couldn't alter that part, they could still edit the entire page resulting in no section link. Anomie 16:34, 15 August 2012 (UTC)
Thanks everone, that saved me quite a bit of time not wasted. →Yaniv256 talk contribs 20:58, 16 August 2012 (UTC)

Odd URL after edit

I just completed an edit to the History section of the Son of Beast page. When I was returned to the new version of the article after clicking save, I would expect the URL to be Instead, it was What is this pe=1 parameter? I've never seen it before. jcgoble3 (talk) 23:02, 14 August 2012 (UTC)

See Wikipedia:Village pump (technical)/Archive 101#URL change following save. PrimeHunter (talk) 23:19, 14 August 2012 (UTC)
Interesting. I will note that the change was first reported here on July 27 and Steven said it would be a two-week test beginning the Monday after that, which would be the 30th. That means this should have gone away yesterday. jcgoble3 (talk) 00:15, 15 August 2012 (UTC)
Hey all. Let me clarify: the test which delivers the post-edit feedback has already ended (it stopped on the 11th), and we now need to disable the rest of extension that tracks whether a page load is after an edit – that's the part which produces the URL change, even if you don't get any interface experiment. The removal should have happened yesterday, but we have been in a team training session that will end Wednesday. I want to apologize for the delay, and say that we will make sure it is down before the end of the week. I'll be sure to give an update here on VPT when that has happened. Steven Walling (WMF) • talk 05:18, 15 August 2012 (UTC)
Okay, this is Fixed. If and when this URL change might reappear as part of another experiment, I'll announce it here. But for now any changes or unique URLs are not due to one of our experiments. :) Thanks again for everyone's patience! Steven Walling (WMF) • talk 23:42, 16 August 2012 (UTC)

Article Feedback link on my watchlist

I'm assuming this is from the recent MW upgrade? In any case, is there a gadget to disable it? Thanks, Jenks24 (talk) 02:50, 15 August 2012 (UTC)

Does the existing "Don't show the article feedback..." button (Preferences - Appearance) not do it? And, yep; see 3 sections above :). Okeyes (WMF) (talk) 06:30, 15 August 2012 (UTC)
No, I've had that ticked since the whole Article Feedback thing started and I never see the form thing at the bottom of pages, but I still have the "Feedback from my watched pages »" link at the top of my watchlist. Jenks24 (talk) 07:04, 15 August 2012 (UTC)
It is Preferences → Appearance → Don't show the Article feedback widget on pages. It does not remove the feedback link from the watchlist. ---— Gadget850 (Ed) talk 20:51, 15 August 2012 (UTC)
It should (do both). Jared Preston (talk) 20:38, 16 August 2012 (UTC)
I don't believe it should do both. We could create a separate gadget or CSS script for the watchlist link. Removing the article feedback widget from the bottom of the page is an appearance issue, there may be editors who want to change the appearance while still having access to the feedback. Ryan Vesey 20:42, 16 August 2012 (UTC)

Support or oppose adding gadget to open search in new tab

This gadget enables the opening of search result lists and search suggestions in a new tab by using Ctrl-click (PC) or Command-click (Mac).

Please indicate your support or opposition. Of the various search gadgets to open search in a new tab I am focusing on this one. I think it gives the most options to users, and is the least intrusive. That is because it does not open search in a new tab automatically. One chooses to do so as needed. It is very handy when checking out search suggestions one-by-one. One does not have to re-enter the search term each time.

It is also useful when using the advanced search form at Special:Search. One can refine one's search and check off different boxes without losing the results of the previous searches.

The main benefit is not covering over the article one is reading. This gadget works on all search forms on Wikipedia, including archive search.

I would like this gadget to be put in My Preferences in the gadget tab: Special:Preferences#mw-prefsection-gadgets.

See JavaScript (JS) here:

See how to implement it here if you want to try it out:

  • Strong support - This gadget is very handy for me, and should be handy for most people. Checked on Arora. Can you make also a version that opens in another tab as default and uses the control for same page search? That version would probably not go on the gadget list but would be my preference. →Yaniv256 talk contribs 06:58, 16 August 2012 (UTC)
  • Support Love it (would go for enabled by default), and love it more, if you submit it as patch to MediaWiki. —TheDJ (talkcontribs) 08:40, 16 August 2012 (UTC)
  • Support enabled by default. Just can't really think of a reason not to. →Yaniv256 talk contribs 16:47, 16 August 2012 (UTC)
  • Support the initial proposal, but oppose the idea of making it default. I understand that many people would find this useful, but I'm sure that I'm one of many who would not, and we need to consider the mass number of readers who would encounter this without being able to turn it off. Nyttend (talk) 14:43, 16 August 2012 (UTC)
Not quite sure why you would not want this as a default gadget. If you don't hold the control key down at the same time as you click the submit button, then search still opens in the same page as it does now. --Timeshifter (talk) 15:54, 16 August 2012 (UTC)
I'm sorry: I misunderstood the proposal. Support making it default. I thought it was saying that a new tab would open whenever I hit "enter" or otherwise told it to start searching, just like Djsasso says; that would be quite obtrusive to me and probably to plenty of other people. Thanks to your response, I too can't understand why someone would oppose if s/he understand your proposal correctly. Nyttend (talk) 23:28, 16 August 2012 (UTC)
  • Oppose enable by default but Support it being a gadget. Hate hate hate pages that auto open tabs for you. Good webdev practice is to not open new tabs unless asked to by the user. -DJSasso (talk) 17:43, 16 August 2012 (UTC)
Even when activated by default, the gadget does not auto-open new tabs. See my reply to Nyttend just above your comment. Default only means that that the gadget is activated without having to add a checkmark to the box for it in the gadget tab. The checkmark is there automatically. You can remove the checkmark if you want to turn it off. An example of a gadget that is turned on by default is mw:Reference Tooltips in the "Browsing" section of the gadget tab page at Special:Preferences#mw-prefsection-gadgets. See also: User talk:Yair rand/ReferenceTooltips. --Timeshifter (talk) 18:03, 16 August 2012 (UTC)
  • Support but on Linux the functionality (for normal links) with Firefox is shift-click to open a new tab and ctrl-click to open a new window whereas with Chrome it is ctrl-click to open a new tab and shift-click to open a new window. Naturally these organisations would rather go bankrupt than provide the user with a uniform experience. Will it be possible to tailor this function for the browser being used? --Mirokado (talk) 17:59, 16 August 2012 (UTC)
    (update) on Linux, Firefox opens a new tab for both shift and ctrl and Chrome behaves as described above. That I think is fine unless it is easy to tweak the implementation for Firefox. --Mirokado (talk) 18:15, 16 August 2012 (UTC)
Most browsers open (or can be set to open) any normal link in a new tab via ctrl-click (PC) and command-click (Mac). So this gadget enables the same thing for search submit buttons. One does not have to do a different action for links versus search buttons. --Timeshifter (talk) 18:23, 16 August 2012 (UTC)
Firefox 14.0.1's behavior on Windows Vista is the same as what Mirokado described for Chrome: new tab on ctrl-click and and new window on shift-click. jcgoble3 (talk) 20:19, 16 August 2012 (UTC)
  • Support as default About time somebody did this. jcgoble3 (talk) 20:19, 16 August 2012 (UTC)
  • Support, but not as default If I want to open it in a new tab I use ctrl+click. I don't want it opening in a new tab every time I do a search. Ryan Vesey 20:44, 16 August 2012 (UTC)
You people listen! This is not what it means to be enabled by default. Go back and read Timeshifter's response to DJSasso, and if you still didn't get it, well, read it again. →Yaniv256 talk contribs 20:56, 16 August 2012 (UTC)
According to his response to DJSasso, you would need to go into your preferences and disable it. That is what default means, it is automatically enabled. But after re-reading this, I'm doubly confused. Is this a gadget to make Wikipedia search do something that browsers already do? I initially thought this would make the search open in a new page when you clicked rather than ctrl-clicked. Now it seems like it would just be to make it open in a new page when you ctrl-click. What change is occurring? Ryan Vesey 21:05, 16 August 2012 (UTC)
Yes! You did get it now. You just don't yet get that you got it, because it is so simple and Timeshifter kind of made it sound complex when he described how he, elegantly, did what he did and not what it does. The only thing this "gadget" does is have that ctrl+click work in the search box, where it did not work before. One could even say this is not really a "gadget" but simply the fix of a bug. Which is the reason why we want it activated by default. →Yaniv256 talk contribs 21:27, 16 August 2012 (UTC)
Timeshifter, how this little beauty works is of no interest to the user. Please don't bring it up again. →Yaniv256 talk contribs 21:34, 16 August 2012 (UTC)
(ec, I think a clarification is in order) With this activated, clicking the search button behaves as it always has done, but if you hold the appropriate (ctrl or shift) key when clicking the button, you get the alternative behaviour of opening in a new tab. Thus there should be no need for anyone to disable it (but allowing people to do so seems like good tactics). The browsers already do this for normal blue links, but do not for things like that button, so the button code must implement the alternative behaviour itself. --Mirokado (talk) 21:38, 16 August 2012 (UTC)
Hmm, my browser opens it in a new tab with ctrl-click. Yaniv is right, this seems like more of a fix then a gadget. Would this be something that could be better fixed with a bugzilla request (which would fix it on other media wiki sites outside of English wiki? Ryan Vesey 21:42, 16 August 2012 (UTC)
One minor addition to what Mirokado said. The way this thing is implemented, a ctrl+enter is just as good as a ctrl+click, at least for me, so one never needs to put down that nice cup of coffee, just to fork his nav away. →Yaniv256 talk contribs 22:07, 16 August 2012 (UTC)
  • Strong Support Changes nothing for the normal user when enabled, but makes things easier for the rest of us. Of course! --Philosopher Let us reason together. 23:00, 16 August 2012 (UTC)

Edit summaries displaying strangely

Since several months ago, I find the occasional edit on my watchlist in which the last word in the edit summary runs past the closing parentheses. Does anyone else see this? Someguy1221 (talk) 21:25, 16 August 2012 (UTC)

This sounds like Wikipedia:Village pump (technical)/Archive 100#Anomalous Edit Summary. Since other people have no access to your watchlist, please give some diff links where this has occurred. --Redrose64 (talk) 22:58, 16 August 2012 (UTC)
This is the only edit that's doing it right now. Someguy1221 (talk) 23:42, 16 August 2012 (UTC)
Yes, it looks to me like almost exactly the same problem... the difference being that there are three spaces, not 83. I take it that your browser is Chrome? --Redrose64 (talk) 23:57, 16 August 2012 (UTC)
Yup. It also disappears when I drag select it, just like in the previous thread. Weird. Thanks for pointing me to that thread. Someguy1221 (talk) 00:02, 17 August 2012 (UTC)

IPv6 edit filter logs

[4] is, clearly, a hit. But when you look at the filter log for that user, [5], you get nothing. Is this a problem with IPv6 addresses in general? That seems to me like a serious problem. Someguy1221 (talk) 00:56, 17 August 2012 (UTC)


Hello all, I've just declined an edit request at MediaWiki talk:Wrongpassword#Change back to default because it looked like it could do with more discussion. This is the page that determines what text users see if they make an incorrect login attempt. At the moment, if a user inputs the wrong password three times then they have to input a CAPTCHA as well as their username and password. Previously, it seems the CAPTCHA was required after one incorrect login attempt, but that has now changed. Now, the text that users see after one or more incorrect login attempts is "Incorrect password or confirmation code entered. Please try again." On the talk page it was suggested that this be changed back to "Incorrect password entered. Please try again." What do people think? — Mr. Stradivarius (have a chat) 05:09, 17 August 2012 (UTC)

Please give feedback on this Wikipedia skin for tablets


I designed a new Wikipedia skin for tablets and would like get some feedback. Thanks in advance. →Yaniv256 talk contribs 19:47, 12 August 2012 (UTC)

Say that I suggested that something in the line of this design should be added to the standard Wikipedia skin list, would you mind sharing any skin-based objections you would have to such a proposal? →Yaniv256 talk contribs 06:08, 14 August 2012 (UTC)

New skins always good, we should have many more available. 18:48, 15 August 2012 (UTC)

Thanks!!! →Yaniv256 talk contribs 19:49, 15 August 2012 (UTC)

WMF staff, would you be interested in accepting this skin as a proof of concept for something you will take on to add to the skin list? →Yaniv256 talk contribs 21:09, 16 August 2012 (UTC)

I changed the title of this section. Otherwise most people will not know it is about a tablet skin, and think it is just another discussion of skins. Skin discussions for desktop PCs are avoided by many people since it can be a very contentious discussion that goes nowhere. The Wikimedia Foundation would have to be involved eventually in major skin changes.
But tablet skins may interest people a lot more. I do not have a tablet, though, so I can not give much relevant feedback. Concerning your notes at User:Yaniv256/TabletChick: From Wikia experience concerning skins, I suggest you not remove any links. We need more links not less. I suggest putting them on the bottom of the page if there is no room elsewhere. And/or use lots of nested menus, hover menus, etc..
You might make a feature request at WP:Bugzilla. The people responding there will be more knowledgeable. --Timeshifter (talk) 07:08, 17 August 2012 (UTC)
Thanks! New title much better!!! →Yaniv256 talk contribs 11:16, 17 August 2012 (UTC)
People seen this may also be interested on the new MediaWiki skin, Athena. Helder 23:53, 17 August 2012 (UTC)

{{db-copyvio-notice}} and {{db-imgcopyvio-notice}}

The templates {{db-copyvio-notice}} and {{db-imgcopyvio-notice}} pass a "willprovide=no" parameter to the meta-template {{db-notice}}, in order to suppress text that invites users to request that the content be provided to them. However, the parameter is not having any effect when the template is substituted. I can't work out why this is; can anyone else see the reason? — This, that, and the other (talk) 08:08, 16 August 2012 (UTC)

Sorry, that was me - I forgot to safesubst the {{lc:}} magic word in {{db-notice}}, so it was messing up the parser function logic. Should be fixed now, but let me know if you notice anything else. — Mr. Stradivarius (have a chat) 04:22, 18 August 2012 (UTC)

Can't delete an image

I cannot delete File:LogConvergence95.gif. I get the following message: Error deleting file: The file "mwstore://local-multiwrite/local-public/6/63/LogConvergence95.gif" is in an inconsistent state within the internal storage backends. How to proceed? Magog the Ogre (talkedits) 01:26, 17 August 2012 (UTC)

Interesting. The enwiki file description page does not exist, and clicking "Create" gives an message that the image is on Commons and not enwiki. It also shows a deletion log indicating that you've deleted it twice within the past 24 hours. The full deletion log shows no undeletions. Yet the file page does not indicate that it is on Commons, and clearly suggests that it IS on enwiki by saying that it is a duplicate of the Commons version. WTF? jcgoble3 (talk) 02:12, 17 August 2012 (UTC)
I tried restoring and deleting. Nada. The problem persists. Gremlins?--Fuhghettaboutit (talk) 02:21, 17 August 2012 (UTC)
See #Error when deleting files: "inconsistent state within the internal storage backends" above. Anomie 02:33, 17 August 2012 (UTC)
...which is now at Wikipedia:Village pump (technical)/Archive 102#Error when deleting files: .22inconsistent state within the internal storage backends.22. jcgoble3 (talk) 12:58, 17 August 2012 (UTC)

MediaWiki: get URL params

moved from RefDesk/IT: I googled and wikied before asking this question, but to no avail. How do I create a template that would get URL parameters? Something like #arg: (too bad DynamicFunctions extension isn't installed for Wikipedia). (talk) 06:52, 17 August 2012 (UTC)

AFAIK, you can't (yet, at least). You could, however, make a template and then layer a user script or global script (written in JavaScript) over the top of it, or you could write your own parserfunction. - Jarry1250 [Deliberation needed] 09:54, 17 August 2012 (UTC)
Great thanks! I couldn't come up with JS idea. (talk) 11:20, 17 August 2012 (UTC)

Replacing all non-section links with TOC links

Is there a way to get all non-section links replaced with #TOC links? I am getting tired of long leads and talk page headers and would like to see them only when I want to. →Yaniv256 talk contribs 11:05, 17 August 2012 (UTC)

Could probably manipulate the html a elements using Javascript. --Izno (talk) 22:53, 17 August 2012 (UTC)

Automated filter identifying edits as potentially unconstructive

As of yesterday I am seeing the below message on about 25% of the articles I have tried to edit, mostly talk pages. I understand its purpose but it is unhelpful and unconstructive if it blocks everything we try and do. I just got it when I tried to edit the WikiProject United States talk page, when I tried to edit a user talk page and when I tried to leave a comment on Template:WikiProject United States one after the other. We should be making it easier to edit not harder under the guise of preventing vandalism. I even got it just now when I tried to save this!

A couple recommendations:

  1. Start tracking the articles it generates on to watch for false positives
  2. Don't generate this for users with more than 100 edits.
  3. Make the false positive reporting process easier. Kumioko (talk) 12:59, 17 August 2012 (UTC)
It's obviously a mistake, so there's no need for the tone. That's the edit filter. The best thing to do is probably to report the false positive on the page that should be linked in the message you quote. That will allow the editors who maintain the filters to find out which one is causing it and make any necessary adjustments. HJ Mitchell | Penny for your thoughts? 13:11, 17 August 2012 (UTC)
See Wikipedia:Administrators' noticeboard/Incidents#Is Wikipedia:Edit filter playing up?. Bring a trout. ;) AndyTheGrump (talk) 13:19, 17 August 2012 (UTC)
(ec)My tone is negative only in that I find that us doing this sort of thing against our own policies of AGF, biting newby's and a variety of others. It flies in the face of what we are trying to do here which is build an encyclopedia Anyone can edit. Unless of course we are going to be changing that to "The encyclopedia anyone "we want" to edit. Also, I am sure it is a mistake that's why I brought it here. It was a mistake to create a process that blocks an edit like this (except IMO for certain rare and exceptional circumstances), it was wrong for the error to generate in this way, it was wrong for the error to generate on a user with an edit count as high as mine (which inclines me to believe it has done it for others as well), its wrong to assume that someone who gets that, and isn't an established user, is going to figure out how to report it, need I go on? Kumioko (talk) 13:21, 17 August 2012 (UTC)
The edit filter is meant to stop only those rare occurrences. In this case, a mistake caused it to block all edits on Wikipedia. This isn't a flaw in the system and it isn't a huge conspiracy to keep people from editing. It was a simple human error. Ryan Vesey 13:31, 17 August 2012 (UTC)

No longer seeing 'redirect from' line in file namesage

Hi there. I work with files a lot, and today when I clicked on File:Jjj.jpg (part of a CCI) it took me to File:Jan Muhammad.jpg without putting on the Jan Muhammad page that it was redirected from Jjj.jpg, and without showing the former address in the usage section. This is disturbing as it makes it very hard for me to do a number of maintenance tasks that I'd otherwise ordinarily be able to do. Is anyone else having this issue, and if it's a new feature, how can I bypass it? Sven Manguard Wha? 13:40, 17 August 2012 (UTC)

That file is on Commons, not enwiki, and commons:File:Jjj.jpg shows the "redirected from" line just fine. The same occurs with File:5-WindSeeker_from_the_Eiffel_Tower.JPG (does not show that line) and commons:File:5-WindSeeker_from_the_Eiffel_Tower.JPG (does show that line). OTOH, the enwiki image redirect File:A.O..jpg shows that line with no problem. So it appears to be an issue of getting Commons to send that line over to enwiki when enwiki requests the image page or something like that. jcgoble3 (talk) 14:02, 17 August 2012 (UTC)
bugzilla:14117( EDIT: that's a different issue. Suppose I should like read these things before I hit save page button. I thought I've seen this before in bugzilla but could not find it). There's several open issues about file redirects behaving somewhat weirdly relative to normal redirects. Bawolff (talk) 14:35, 17 August 2012 (UTC)

HTTP 403 for article assessment tools at Toolserver

From time to time, I assess articles that are yet unassessed for WikiProjects that I participate in. Example: I go to the {{Wikipedia:Version 1.0 Editorial Team/XYZ articles by quality statistics}} box in the assessment department for WikiProject XYZ and click on the links to the lists at Toolserver. I just attempted to do so, but was repeatedly presented with an HTTP 403 error. What's going on with that? ~~ Lothar von Richthofen (talk) 21:29, 17 August 2012 (UTC)

The Assessment tool/bot was tuned off by its owner/creater User:CBM temporarily. The toolserver backlog was several days old so it was making the assessment updates pointless. Once the toolserver catches back up he'll turn it back on. Kumioko (talk) 21:41, 17 August 2012 (UTC)
See recent discussions at Wikipedia talk:Version 1.0 Editorial Team/Index. --Redrose64 (talk) 21:48, 17 August 2012 (UTC)

Forcing creation of a section without a heading

Is there a way to force/trick MediaWiki into creating a numbered section (e.g. a div or other block element) that does not have a corresponding H# heading? This would be very useful for making individually editable glossary entries using template-structured glossaries per MOS:GLOSSARIES. The technical problem is that a H# heading, e.g. something in ===Heading here=== markup, cannot be inside a <dl>...</dl>-based definition list structure, as used by template-structured glossaries. Nor vice-versa (you can't put <dt>...</dt> markup inside a heading). One could get around this issue by breaking entries out into templates and transcluding them, and having an edit link template edit the template page, but this would violate the principle that we don't use sub-articles or effectively "hide" article content in templates. My guess is that there is not an extant way to do this, but also that it would not be hard for developers to implement. My experience with MW's bugzilla tells me that a request to add such a feature, however, would be ignored for several years as too low-priority. So, I'm hoping there's an existing means of pulling this off already. — SMcCandlish   Talk⇒ ɖ∘¿¤þ   Contrib. 06:04, 7 August 2012 (UTC)

Anyone? Anyone? Bueller? Bueller? — SMcCandlish   Talk⇒ ɖ∘¿¤þ   Contrib. 23:35, 13 August 2012 (UTC)

It is possible, but possibly not desirable:

The following wikitext will create an empty div element to contain the manually-specified anchor ID, followed by an invisible H3 element to generate the automatic TOC entry, followed by the actual text:

<div id="{{anchorencode:example subheading}}"></div>
<h3 style="display:none;">example subheading</h3>
;example subheading
:example definition description

like this:


example subheading

example subheading
example definition description


Unfortunately, the ID of a non-displayed element is ignored when the browser looks for anchor links. But an empty element with the same ID immediately before the non-displayed element will trick the browser into jumping to the correct location (tested in Chrome; I think all browsers tolerate non-unique IDs even though they are supposed to be unique; there might be unexpected side-effects if editors have customised settings).

Note that, if any sections or sub-sections on the page have the same heading text, you would have to manually adjust the ID for the div elements corresponding to any manually-coded h elements (other than the first), because the parser automatically appends a distinguishing digit in the TOC. So this is not a completely robust solution. Clearly the above code is also likely to confuse other editors and might be difficult to maintain over time. (You could probably template the above code, turning the heading text into a parameter, to avoid the need to enter the intricate HTML and repetition of the heading in the wikitext.)

Whether such a convoluted approach is wise is another matter!

Richardguk (talk) 01:35, 14 August 2012 (UTC)

Oh, you said you wanted individually-editable glossary entries. Section editing ignores manually-created section breaks (even though the heading titles are automatically rendered in the table of contents), and there is no applicable URL for an edit link target, so the above technique does not help. — Richardguk (talk) 02:10, 14 August 2012 (UTC)
Well, thanks for trying. The trick you figured out might be good as a template option to force display of glossary entries in a a ToC (obviously this would be in short glossaries, or the ToC would become unmanageable). I guess editable divs that aren't sections will remain elusive. — SMcCandlish   Talk⇒ ɖ∘¿¤þ   Contrib. 12:34, 17 August 2012 (UTC)
I just came across {{Editpoint}}. Not sure if that might be of any use? DH85868993 (talk) 11:52, 18 August 2012 (UTC)
It just adds a link [editpoint], which is essentially another [edit] link for the same section. It doesn't create any subdivision. You can see it in use at User:Fred Gandt/sandbox/templates. --Redrose64 (talk) 14:39, 18 August 2012 (UTC)

Template called through Redirect: can we know?

So template A can be Redirected to another template B to work B's code. It there any way for B to detect & use it was called from a Redirect? (Or maybe: that it was called from template space?) -DePiep (talk) 10:13, 16 August 2012 (UTC)

no, the template only has information about the final page that it is embedded in. See bugzilla:24284. Bawolff (talk) 17:10, 16 August 2012 (UTC)
Clear then. -DePiep (talk) 00:11, 17 August 2012 (UTC)

I'm not sure I follow. The question is, can we make a template that calls another template and sets one of the parameters? We have a template {{Braille cell}} that by default has graphic output, but very commonly is specified with text output through the parameter type=text. Rather than typing that each time, or duplicating the code across two templates, could we create a template "Braille text" that calls the code at Braille cell, but supplies 'text' for the type parameter? — kwami (talk) 07:02, 17 August 2012 (UTC)

If I understand the OP correctly, they're asking about a situation where Template:Foo2 contains the text #REDIRECT [[Template:Foo1]]. In that case, I don't think there is any way for Foo1 to test whether it was invoked "as" Foo2. However, as kwami suggests, you could change Foo2 to contain something like {{Foo1|invokedasfoo2=y}}, which would tell Foo1 how it was invoked. --R'n'B (call me Russ) 11:29, 17 August 2012 (UTC)
R'n'B described the question right. I do not like the second option (code in Foo2), because we'd have to maintain both code variants (while it is only about parameter setting). Number of testing situations doubles, complexity quadruples. (An other solution: we are working to change the settings in "Foo1"). -DePiep (talk) 11:55, 17 August 2012 (UTC)
I think we have a misunderstanding here. DePiep, the complexity would hardly increase at all. All it would be do is call the bc template while setting 'type' to 'text'. Maintenance would be minimal.
R'n'B, how would we do that? — kwami (talk) 05:12, 18 August 2012 (UTC)
You would need to add {{{invokedasfoo2|}}} to the foo1 template code. The exact code would depend on the behaviour you wanted. You could get the template to add a category to the page if the template was invoked as foo2, for example. — Mr. Stradivarius (have a chat) 05:28, 18 August 2012 (UTC)
I would want foo1 to supply one of the parameters for foo2, so that we didn't need to supply it manually. That is, call foo2 and provide 'a' as the value for parameter 'x'. — kwami (talk) 05:38, 18 August 2012 (UTC)
Smart suggestion by Mr. Stradivarius. For other reasons (ask me), not applied in the current issue. Kwami and I converged on setting the default value in the base template. Consider resolved. -DePiep (talk) 16:12, 18 August 2012 (UTC)
I would still like to know how to do this. It could be very useful in other cases. — kwami (talk) 22:32, 18 August 2012 (UTC)

Firefox login bug

Has anyone else had two tabs on Firefox (14.0.1 Ubuntu 12.04) with one logged in and the other logged out? This is what I did.

  1. I logged in to Wikipedia
  2. I openned a new tab and navagated to this page via Google search

I can edit this page both logged in and logged out in the same browser window using different tabs without logging out or logging in. – Allen4names (IPv6 contributions) 04:29, 17 August 2012 (UTC)

This is an edit using the logged out tab. 2001:470:B:53:0:0:0:65 (talk) 04:31, 17 August 2012 (UTC)
Is it possible you are using the secure server in tab 1 but not in tab 2? (This could happen with either the old or the new secure server; in the latter case, it is because of the secure attribute on some cookies.) SoledadKabocha (talk) 19:48, 17 August 2012 (UTC)
HTTPS Everywhere not working. I need to figure out how to fix it. – Allen4names (IPv6 contributions) 05:47, 18 August 2012 (UTC)
There have been a couple of updates to HTTPS Everywhere recently. The latest versions are 2.2.1 (stable branch for Firefox), 3.0development.6 (unstable branch for Firefox), and 2012.8.16 (for Chrome). At least 2.2.1 contains an emergency fix for a bug from 2.2 in which some rulesets that should have been disabled by default were not. However, this shouldn't affect the Wikipedia ruleset, as it is on by default. Try forcing an update check; further troubleshooting may belong in WP:Reference desk/Computing. SoledadKabocha (talk) 06:26, 18 August 2012 (UTC)
I fixed it by downloading a ruleset and putting it into ~/$myFirefoxProfileDir/HTTPSEverywhereUserRules. Wiki(m|p)dia does not apear to be on the default list that ends with TheGlobeAndMail so I suspect there is another bug. Thanks anyway. – Allen4names (IPv6 contributions) 20:13, 18 August 2012 (UTC)

Upload and delete both in "inconsistent state"

Trying to upload a new version of File:RHS_logo.png and get the message The file "mwstore://local-multiwrite/local-public/c/c0/RHS_logo.png" is in an inconsistent state within the internal storage backends. Trying to delete an old version of File:Bing Maps Streetside.png gives Error deleting file: The file "mwstore://local-multiwrite/local-public/archive/3/38/20120808184609!Bing_Maps_Streetside.png" is in an inconsistent state within the internal storage backends. I've seen the latter before (addressed above & now in archive #102), but this is the first time I've received an upload glitch. Just thought I should report the variation on the theme... Skier Dude (talk) 01:54, 18 August 2012 (UTC)

See #Can't delete an image above. --Redrose64 (talk) 11:43, 18 August 2012 (UTC)

Preferences force edit summary

I have My preferences\Editing\"Prompt me when entering a blank edit summary" checked. It was working, although I don't forget to create an edit summary too often. However, it has stopped working. I have no idea why.--Bbb23 (talk) 10:46, 18 August 2012 (UTC)

Need some help with this unblock request

See User talk:Mmhrmhrm. His editing was blocked by ProcseeBot and I'm not sure what to do now, and he's obviously not clear at all. But he looks like a good faith editor and could use some help. Thanks. Dougweller (talk) 17:12, 18 August 2012 (UTC)

Well I made this user IP block exempt, in case there is no choice but for the use of this proxy. Graeme Bartlett (talk) 21:24, 18 August 2012 (UTC)
Check at WP:OPP first, and do some research about the IP addresses. Tell them to submit a ticket to UTRS and I'll look at it (I'll need them to give me the IP addresses) Secretlondon (talk) 22:05, 18 August 2012 (UTC)


Can somebody modify {{Dupdet}} so that it always uses links? Secure links (especially ) don't work properly! Thanks. mabdul 14:49, 18 August 2012 (UTC)

It doesn't have any https: links, nor any to Have you got the right template? --Redrose64 (talk) 15:23, 18 August 2012 (UTC)
Log in at / and go to the doc page: it creates the links automatically through the urlencoding of the article input! mabdul 15:36, 18 August 2012 (UTC)
Not on my browser (Firefox 14.0.1 on WinXP); I'm using https:// to browse, but all the links on {{Dupdet}} and {{Dupdet/doc}} are of the http:// variety. --R'n'B (call me Russ) 16:07, 18 August 2012 (UTC) is deprecated. See Wikipedia:Secure server. The problem is the use of fullurl in the url1 parameter when the first parameter to {{Dupdet}} does not start with http://. At and, {{fullurl:Name}} produces // At it produces {{Dupdet}} sticks http:// in front in either case in The first example at produces the link This has url1=http%3Ahttps%3A so duplicationdetector fails with the message "Could not load URL http: Couldn't resolve host 'http:https'." PrimeHunter (talk) 16:27, 18 August 2012 (UTC)
Mabdul, can you describe the problems you hit with and {{dupdet}}? I use only https (I'm not logged in on http, actually), and I've seen no problems with this template on my otherwise problematic xombrero... — Dmitrij D. Czarkoff (talk) 21:04, 18 August 2012 (UTC)
@PrimeHunter OK, I'm mostly online at the "deprecated" wiki(m|p)edia URLs, but still: it doesn't work: Could somebody remove it. And a) I'm still requesting a source where it says that this URL is deprecated and b) I'm still wating that the URL redirect is created that meta.wikipedia redirects to meta.wikimedia. My last information was that secure.wikim was still using directly the Florida servers and that https://en was using (in my case) the Amsterdam servers... so I will stick with the old "solution" (and it uses a different favicon for better recognition!
@Czarkoff sry my fault: this was solved through the "bug fix" of independent URI schemes in July 2011 (likely later). mabdul 00:38, 19 August 2012 (UTC)
Wikipedia:Secure server#See also links to Status of SSL access to Wikipedia? from April 2011 where User:Brion VIBBER wrote: "We've still got the old SSL proxy on which allows reaching most (hopefully all?) Wikimedia wikis over SSL on an alternate URL, but Wikimedia ops currently considers this an 'unsupported' service, which gives it a lower priority for fixes and keeps it off"
Wikipedia:Wikipedia Signpost/2011-10-10/Technology report#Improved https support comes to Wikimedia wikis says: "Also announced this week was the https switchover. Writing for the Wikimedia Foundation blog, operations engineer Ryan Lane said that the domain had been officially deprecated." PrimeHunter (talk) 01:02, 19 August 2012 (UTC)
I think I fixed this problem. Ruslik_Zero 11:19, 19 August 2012 (UTC)
@Ruslik thx.
@PrimeHunter I won't change my browser configurations as long as the old way is working stable. mabdul 11:31, 19 August 2012 (UTC)
Why not to use {{canonicalurl}}? Incnis Mrsi (talk) 12:10, 19 August 2012 (UTC)
This is also possible, of course. Ruslik_Zero 13:16, 19 August 2012 (UTC)

Image not displaying properly (a common topic)

I know that this is very common, but the logo in the infobox of OMSI isn't showing up right (it's just a black box, but it looks fine on the file page). I tried purging and bypassing the cache on the page, the file page, and the Commons file page, but it still wouldn't show up. Any idea why? David1217 What I've done 22:30, 18 August 2012 (UTC)

It works for me. I see Does that fail for you? PrimeHunter (talk) 00:07, 19 August 2012 (UTC)
(I thought this might happen.) All I see is a black box. Safari 5.1.7 on OS X 10.6.8. David1217 What I've done 01:40, 19 August 2012 (UTC)
Image is in the CMYK color model. That's never good news for a JPEG. --Redrose64 (talk) 09:54, 19 August 2012 (UTC)
Perhaps it can be converted to SVG? That might fix the problem. I don't know how to convert it, though. David1217 What I've done 15:27, 19 August 2012 (UTC)
There is no easy way of "converting" a JPEG into a SVG because they are totally different ways of looking at a picture. Many people draw SVG images using software such as InkScape; but it's possible to draw a SVG from scratch using a plain text editor, although in this case a SVG-capable browser (current versions of Chrome, Firefox, Opera or Safari) will be necessary to view the result before you upload it.
If you don't want to redraw as a SVG, you can convert the JPEG into a .PNG quite easily; or you could keep it as a JPEG, just convert it to a more widely-supported color space such as YCbCr or RGB. --Redrose64 (talk) 15:50, 19 August 2012 (UTC)
I have no clue how to do any of that, so I'll take it to the Graphic Lab. And Redrose64, do you see the image properly? David1217 What I've done 15:59, 19 August 2012 (UTC)
Yes; I see it correctly in Firefox 14.0.1, Microsoft Picture Editoe, Microsoft paint and IrfanView.
Per commons:Help:JPEG#Color model you can fix it using GIMP. --Redrose64 (talk) 16:25, 19 August 2012 (UTC)

Upload warning - inconsistent state within the internal storage backends

Here is an error message I keep getting whenever I try to upload a new version of the map CartogramPresidentialCounty1896.gif

"Upload warning

The file "mwstore://local-multiwrite/local-public/e/ee/CartogramPresidentialCounty1896.gif" is in an inconsistent state within the internal storage backends"

I encountered the problem yesterday. I uploaded the new version in the same manner as my other uploads. I've even tried renaming the file before uploading it. I've always used Firefox. Is there any way to get around this? Thanks.--Tilden76 (talk) 15:10, 19 August 2012 (UTC)

See #Upload and delete both in "inconsistent state" and #Can't delete an image above. jcgoble3 (talk) 15:34, 19 August 2012 (UTC)

Template problem

Could a kind and clever person please find out what is wrong with this template {{Fa/BeenOnMainPage|''[[In Rainbows]]''}} on this page, under the subheading "Music"? Graham Colm (talk) 16:13, 19 August 2012 (UTC)

Fixed. Template names are case-sensitive (except for the first letter). So template {{Fa/BeenOnMainPage}} does not exist, but template {{FA/BeenOnMainPage}} does. Edokter (talk) — 16:20, 19 August 2012 (UTC)
Thanks, I feel like an idiot now :-) Graham Colm (talk) 17:09, 19 August 2012 (UTC)

Autosuggest for search function on main site

Hello, I was wondering why the search box on the main page does not have a drop-down autosuggestion list. I realize that the search function has to be able to switch between languages, but surely it could be made to autosuggest in the selected language...? Just a thought. (talk) 19:36, 19 August 2012 (UTC)

Certainly possible. (I actually added it to the Wiktionary main page 9 months ago, but Wikipedia's still doesn't have it...)
This should be discussed at template, rather than here on the English Wikipedia, though. --Yair rand (talk) 20:48, 19 August 2012 (UTC)

Undisplayable image

File:Windows_8_start_screen.png will not display in articles (e.g. Metro (design language), Microsoft Windows and Windows 8), at least not when use Firefox Aurora 16 sub-ver. 2 Aug/18/2012, on Windows XP. It does however show up when you look at the images profile.

Here is a screenshot of the problem.

Icedog (talk) 20:37, 19 August 2012 (UTC) Edited 20:51 UTC

The file description page isn't showing anything useful either at the top or in the topmost row of the history table. From what I can gather on that page, image was copied from a Microsoft Windows 8 announcement, so the image is almost certainly just as buggy as anything else that those Seattle cowboys ship out. --Redrose64 (talk) 20:57, 19 August 2012 (UTC)
When I try to access the file itself, I get error 404. This makes the display issue completely logical. — Dmitrij D. Czarkoff (talk) 21:36, 19 August 2012 (UTC)
I reverted again, and now I experience no issues on Metro (design language). Solved? — Dmitrij D. Czarkoff (talk) 21:39, 19 August 2012 (UTC)
Ya. It works fine now. Icedog (talk) 02:14, 20 August 2012 (UTC)


The toolserver for viewing "Recognised content" on wikiprojectsis down and says "forbidden" since at least yesterday,Lihaas (talk) 23:55, 19 August 2012 (UTC)

Wikiprojects present their "Recognised content" in different ways so a link would have been helpful. Many of them have no relation to the toolserver. I guess you refer to a table counting articles by quality and importance like to the right at Wikipedia:WikiProject Organizations#Recognised content. For that, see Wikipedia talk:Version 1.0 Editorial Team/Index#WP 1.0 bot temporarily down 2012-8-7. PrimeHunter (talk) 01:47, 20 August 2012 (UTC)

Flexible collapsing

I recently found that I miss some more flexible collapsing then is currently possible with {{hidden}} and friends. The things I would like to be able to do are:

Label summary [show]
Details that are hidden unless the handle above is activated


summary [show] Details that are hidden unless the handle is activated

(This is needed for crowded navigation templates; I know that they can be split to reduce complexity, but this choice isn't always optimal.) Is it possible? — Dmitrij D. Czarkoff (talk) 18:00, 18 August 2012 (UTC)

You may find what you are looking for on mw:Manual:Collapsible elements / mw:RL/DM#jQuery.makeCollapsible. Helder 22:27, 18 August 2012 (UTC)
Thanks!!! Indeed that was enlightening. The only question remains:
Label summary [hide]
Details that are hidden once the handle above is activated
Is there any way to make the external handle ([hide]) react on collapsing the way it happens in {{hidden}}? (These and linked docs only describe doing so with the handle inside parent div, which can't be done in infobox.) — Dmitrij D. Czarkoff (talk) 01:23, 19 August 2012 (UTC)
I'm not sure what you want to do, so for now I will just point another topic where it is described an attempt to migrate that template to the new system: Wikipedia:Village pump (technical)/Archive 92#Deprecation of Collapsible tables and NavFrames (unfortunately it was archived and we are still sending to Wikipedia readers tree different scripts which do essentially the same thing...) Helder 02:55, 19 August 2012 (UTC)
Literally I want to be able to change the text of the "[hide]" handle to "[show]" once the hiding action is performed and vice versa. Eg. in the infobox above the handle always reads "[hide]", regardless of the state of collapsible section. — Dmitrij D. Czarkoff (talk) 03:28, 19 August 2012 (UTC)
This can be done by javascript, by moving the toggle to a new location. Ruslik_Zero 11:28, 19 August 2012 (UTC)
Label summary
Details that are hidden once the handle above is activated
Or, you can do like above. Ruslik_Zero 11:38, 19 August 2012 (UTC)
Oh, I forgot about nesting. My question is now answered in full. Thanks! — Dmitrij D. Czarkoff (talk) 17:48, 19 August 2012 (UTC)
Markup Renders as
|title = Infobox
|float = none
|bodyclass = vevent
|label1 = {{nowrap|Label1-parent}}
|data1 = Lorem ipsum
|data2 = {{infobox
 |child = yes
 |bodyclass = mw-collapsible
 |label1 = Label1-child
 |data1 = {{nowrap|Here should be the handle}}
 |data2 = <span class="mw-collapsible-content">This should be hidden once the handle above is activated</span>
|label3 = {{nowrap|Label3-parent}}
|data3 = Lorem ipsum
Label1-parent Lorem ipsum
Label1-child Here should be the handle
This should be hidden once the handle above is activated
Label3-parent Lorem ipsum
Well, it appears that child infoboxes are just converted to rows, so the |bodyclass=mw-collapsible just doesn't work. So the question of how to manage the handle remains open. — Dmitrij D. Czarkoff (talk) 21:10, 19 August 2012 (UTC)
You can do like this:
Label1-parent Lorem ipsum
This should be hidden once the handle above is activated
Label3-parent Lorem ipsum
Ruslik_Zero 11:53, 20 August 2012 (UTC)
Sure, but the whole purpose was to have collapsible section with no label and a control in another row. See template:infobox software/testcases for reference. — Dmitrij D. Czarkoff (talk) 12:20, 20 August 2012 (UTC)

Unwanted item appeared on my watchlist

Upon editing my watchlist I discovered an entry which was spam. I certainly didn't put it there and I certainly didn't watch any page with that title. I checked to see if such a page existed and it didn't nor did it say there ever was such a page. So how did this entry get on my watchlist ? Was my account hacked or is there some other way it could have happened ? Wlhquest (talk) 15:02, 19 August 2012 (UTC)

If someone moves a page you are watching, then the new page is automatically added to your watch list. So a page you were watching was moved to that spam title, then moved back to its proper name and the redirect deleted. But the spam redirect stayed on your watchlist. --B (talk) 15:04, 19 August 2012 (UTC)
Thanks. That's reassuring. It makes me think though that the deleted redirect should be automatically deleted from the watchlist. Wlhquest (talk) 15:10, 19 August 2012 (UTC)
That it stays around is actually helpful for vandalism fighting, as deleted spam or vandalism redirects are likely targets of future spam or vandalism attempts. A significant part of my watchlist consists of redlinks that I'm keeping an eye on to ensure they aren't recreated w/o a non-spam or non-vandalism reason. --Philosopher Let us reason together. 13:59, 20 August 2012 (UTC)

reference inside a reference cause no ending tag bug

ok when you make a reference group and add reference inside that reference group you get a no closing tag error

for example

[n 1][1]

the only way to fix it is to do something like this

[n 1]

but if your not using cite you get this

[n 2]

but what you are wantin it to look like is this in the notes section using cite

Shared with Dumbarton F.C. after both clubs ended the season on 29 points. A play-off game at Cathkin Park on 21 May 1891 and finished 2–2, so the clubs were declared joint champions[2]


  1. ^ a b Shared with Dumbarton F.C. after both clubs ended the season on 29 points. A play-off game at Cathkin Park on 21 May 1891 and finished 2–2, so the clubs were declared joint champions Cite error: Invalid <ref> tag; name "dumbarton" defined multiple times with different content (see the help page).
  2. ^ Shared with Dumbarton F.C. after both clubs ended the season on 29 points. A play-off game at Cathkin Park on 21 May 1891 and finished 2–2, so the clubs were declared joint champions [1]


  1. ^ "Rangers". Scottish Football League. August 2012. Archived from the original on 18 August 2012. Retrieved 18 August 2012. 
  2. ^ "Rangers". Scottish Football League. August 2012. Archived from the original on 18 August 2012. Retrieved 18 August 2012. 

if you cant reference a note which reference group n is a note about something it kinda defeating the propuse of the note to give additional information but still referenced--Andrewcrawford (talk - contrib) 18:31, 19 August 2012 (UTC)

You see that help page link at the end of the error message? That leads you to the help page for that error which explains the issue. ---— Gadget850 (Ed) talk 18:39, 19 August 2012 (UTC)
i know whay causes ut but its a bug which stop you using reference group correctlyAndrewcrawford (talk - contrib) 18:42, 19 August 2012 (UTC)
Yes; it has been reported. One of these years it will get fixed. ---— Gadget850 (Ed) talk 19:20, 19 August 2012 (UTC)
There is a workaround:
Markup Renders as
Lorem ipsum dolor sit amet, consectetur adipiscing elit, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam, quis nostrud exercitation ullamco laboris nisi ut aliquip ex ea commodo consequat. Duis aute irure dolor in reprehenderit in voluptate velit esse cillum dolore eu fugiat nulla pariatur. Excepteur sint occaecat cupidatat non proident, sunt in culpa qui officia deserunt mollit anim id est laborum.{{#tag:ref|The text is derived from sections 1.10.32–3 of Cicero's De finibus bonorum et malorum<ref>On the Boundaries of Goods and Evils</ref>}}


Lorem ipsum dolor sit amet, consectetur adipiscing elit, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam, quis nostrud exercitation ullamco laboris nisi ut aliquip ex ea commodo consequat. Duis aute irure dolor in reprehenderit in voluptate velit esse cillum dolore eu fugiat nulla pariatur. Excepteur sint occaecat cupidatat non proident, sunt in culpa qui officia deserunt mollit anim id est laborum.[2]

  1. ^ On the Boundaries of Goods and Evils
  2. ^ The text is derived from sections 1.10.32–3 of Cicero's De finibus bonorum et malorum[1]
— Dmitrij D. Czarkoff (talk) 22:00, 19 August 2012 (UTC)
The help page will lead you to WP:REFNEST, which describes that usage, as well as {{refn}}. Please let me know if there are issues with the help page.---— Gadget850 (Ed) talk 10:26, 20 August 2012 (UTC)
BTW, I see a formatting issue in the markup I placed: the order of references is calculated in order of their appearance in source, not in rendering. — Dmitrij D. Czarkoff (talk) 12:41, 20 August 2012 (UTC)


that page is acutely harder to understand than your example but thanks for understanding and the help, sorry for the late thanks only just had time to checkAndrewcrawford (talk - contrib) 20:47, 20 August 2012 (UTC)


How to identify non-printing characters in diffs?

On the surface, these two versions appear identical. The diff highlights the year 2012 but it's not clear what changed there. The version history shows that the newer version lost 3 bytes, but what are they? Is there a way to tell? Thanks, Wbm1058 (talk) 18:55, 20 August 2012 (UTC)

If you copy+paste the old text into this useful tool, you'll see that there was a U+200E character in there, a left-to-right mark. -- John of Reading (talk) 19:04, 20 August 2012 (UTC)

Vector skin colors

What is the light blue color used in the border of the left navbar? (talk) 23:10, 20 August 2012 (UTC)

It appears to be #A7D7F9. --Andrew (User:90) (talk) 04:51, 21 August 2012 (UTC)

Diffs not working

(Extracted from section above as appears to be a different problem) Diffs are not working on either the normal or secure server - is this a general problem related to the "down"? or is it (like "my preferences" above) only affecting me and a few others? - Arjayay (talk) 18:35, 6 August 2012 (UTC)

Diffs have been working fine for me. Ryan Vesey 18:44, 6 August 2012 (UTC)
They flash up momentarily and then the text disappears leaving a selection of grey bars and white boxes with blue lines around them - but no text. Arjayay (talk) 18:53, 6 August 2012 (UTC)
Diffs still just flashing up momentarily and disappearing leaving a selection of grey bars and white boxes with blue lines around them - but no text Arjayay (talk) 08:23, 7 August 2012 (UTC)
I'm getting this problem too. Not even the flashing, just the blank diffs. Enhanced diff display also non-functional. Popups is working, though. IE8, Vector skin. --Stfg (talk) 13:51, 7 August 2012 (UTC)
I'm also using IE8 and Vector - Out of interest Stfg, are your tabs under "My preferences" working? The "My preferences" problem above predated the "outage" that seems to correlate with the diffs not showing - Arjayay (talk) 15:58, 7 August 2012 (UTC)
Yes, my tabs under "My preferences" are working fine. --Stfg (talk) 17:01, 7 August 2012 (UTC)
I've got the same problem, using IE8 and monobook. I deleted emptied my monobook.js and diffs were working again. I'd prefer to keep those skripts, they were working fine so far. I hope this is helpful to find the problem. --Trigonomie (talk) 15:22, 8 August 2012 (UTC)
This is on the german wikipedia right ? I just tested your scripts (at least on Safari on Mac, since I have no windows and no IE. I could not find any glaring issues with it. No idea. but you have so much scripts, I would not be surprised if there is something broken on specifically IE, because scripts break all the time, especially if you don't maintain them (that's why gadgets are better for most users than installing skin scripts for themselves). —TheDJ (talkcontribs) 19:04, 8 August 2012 (UTC)

My diffs are still not working - and as shown above, it is not just me. Lack of diffs makes fighting vandalism particularly difficult. No-one has made any suggestions of what might be wrong, or if it should be reported elsewhere. Some help would be appreciated.
Interestingly, I had an edit conflict when trying to post the above - and the diffs on the edit conflict did work - does this information help track it down? - Arjayay (talk) 15:33, 8 August 2012 (UTC)

Some more diagnostics: templates {{diff}} and {{diff2}} have the same problem; the diff shown by an undo from the history page shows the same problem; and emptying (not deleting, just emptying) my vector.js and clearing cache (Ctrl+refresh) had no effect. Popups are displaying just what they always have. This is desperate: any more diagnostics wanted? --Stfg (talk) 16:16, 8 August 2012 (UTC)
Please try to empty both: your common.js und your vector.js --Trigonomie (talk) 16:32, 8 August 2012 (UTC)
Emptied all my .js files hard refreshed and rebooted - no change whatsoever - Arjayay (talk) 17:27, 8 August 2012 (UTC)
Hello Trigonomie. I've now emptied both vector.js and common.js. Then cleared my cache (Ctrl-refresh) and also done the action=purge job described at Wikipedia:Bypass your cache#Forcing the server to re-render. Still the same problem, I'm afraid. --Stfg (talk) 17:59, 8 August 2012 (UTC)
... and I've just tried it with several other skins: Classic, Monobook and Simple, for none of which do I have .js files. All the same. --Stfg (talk) 18:24, 8 August 2012 (UTC)
OK, any tech editor who knows what the JS console or IE developer tools are and of course has a Windows machine, would like to take a look ? I know there's few Windows using tech editors around, but please try to help your fellow OS users. —TheDJ (talkcontribs) 19:04, 8 August 2012 (UTC)

I was able to recreate those "empty boxes" in IE8 by enabling wikEdDiff in gadgets. Can it be so simple? Hard to believe the affected users didn't try checking diffs in any other projects. — AlexSm 20:32, 8 August 2012 (UTC)

Thank you, Alex. Disabling wikEdDiff has restored basic diff view to full function. I'll leave it disabled for now, but it is actually very difficult to use basic diff view when people insert or remove paragraphs breaks. To cause this, something must have happened in the last couple of days to disrupt the interaction between wikEdDiff and whatever else is in the black box. Is there a plan to resolve it? Regarding "Hard to believe ...", we non-experts don't always know where to look. Thanks again. --Stfg (talk) 21:14, 8 August 2012 (UTC)

No, it’s not that simple. DE is my homewiki. I have no problems here. In my opinion it is due to the fact that (some) scripts are processed not properly any more. WikEdDiff is no option in dewiki, WikEd is not checked. --Trigonomie (talk) 05:18, 9 August 2012 (UTC)

I have determined the culprit: [6]. --Trigonomie (talk) 08:21, 9 August 2012 (UTC)
(e/c)So, I just need to solve the "My preferences" tabs not working problem above, which has been outstanding since 21 July, and then I can go back to the basic system - hardly a solution. As User:Stfg says above, basic diff view is, er, basic, and has several limitations.
I first discovered this problem just after the "down" on 6 August - are the problems related? will they be resolved when the database lag catches up? has a bugzilla report been filed? (sorry I don't understand how to do bugzilla reports) - Arjayay (talk) 08:28, 9 August 2012 (UTC)
Sorry I had started composing my reply before User:Trigonomie's last addition - but it allowed me to post my comment without an E/C. If we have "the culprit" can it simply be reversed? - presumably the change was made to improve something else and this is just collateral damage? - Arjayay (talk) 08:35, 9 August 2012 (UTC)
I notified the maintainer of the gadget. I don't see how this problem can related to users having trouble with the My Preferences btw, that's likely a separate issue. —TheDJ (talkcontribs) 08:44, 9 August 2012 (UTC)
This is still not working, is it all down to the one individial editor? Arjayay (talk) 12:57, 16 August 2012 (UTC)
Yes. —TheDJ (talkcontribs) 14:33, 16 August 2012 (UTC)
User:Cacycle has fixed this (works for me using XP IE8 + Vector) - so it is safe to turn back on. - Arjayay (talk) 07:54, 21 August 2012 (UTC)

Sidebar script

Hey guys, it would be great if someone could create a script that puts a link to Special:NewPagesFeed in the toolbox (or anywhere in the sidebar). Just something simple that can be added to a user's personal js page, no fancy gadgets or anything. Okeyes and I were talking about it and it would allow editors who frequent the NewPagesFeed to access it more easily.16:54, 17 August 2012 (UTC) — Preceding unsigned comment added by Ryan Vesey (talkcontribs)

Haven't tested this, but it should add it to the left toolbar:

mw.util.addPortletLink ('p-tb', wgServer+wgArticlePath.replace("$1", "Special:NewPagesFeed"), 'New pages feed');

---— Gadget850 (Ed) talk 17:57, 17 August 2012 (UTC)

It works perfectly! I added it as a new script in my userspace, but it can be moved to your userspace if you wish. Ryan Vesey 18:24, 17 August 2012 (UTC)
This goes where? in the js? MathewTownsend (talk) 19:42, 17 August 2012 (UTC)
Yep, right now I have it sitting at User:Ryan Vesey/sidebar.js. You could import it using importScript('User:Ryan Vesey/sidebar.js'); in your common.js. I'll let Gadget take over the script if he wishes. Ryan Vesey 19:45, 17 August 2012 (UTC)
ok, thanks! I'll give it a try. MathewTownsend (talk) 19:46, 17 August 2012 (UTC)
I ripped the basic script from someone else. You can use it to add pretty much anything to the sidebar. ---— Gadget850 (Ed) talk 13:37, 18 August 2012 (UTC)

This is actually the advised way to do this now:

mw.util.addPortletLink ('p-tb', mw.config.get('wgArticlePath').replace("$1", "Special:NewPagesFeed"), 'New pages feed');

TheDJ (talkcontribs) 09:15, 21 August 2012 (UTC)

Some syntax not displaying on Safari and Chrome for the Mac

Dear colleagues, this may be a well-known issue, but perhaps someone could confirm it: the small and font-size-percentage syntaxes make no difference to the display on these two browsers; they work in Firefox for the Mac. Tony (talk) 07:19, 21 August 2012 (UTC)

There is not enough context to answer this question properly. Please provide a proper example. —TheDJ (talkcontribs) 08:48, 21 August 2012 (UTC)
I guess it refers to this at Help:Wiki markup#Format:
Description What you type What it looks like

Small text.

Use <small>small text</small> if needed.

A span tag can set text font-size as
being <span style="font-size:87%">87%
of prior size</span>, to match an
image caption.

Use small text if needed.

A span tag can set text font-size as being 87% of prior size, to match an image caption.

Do you see normal size for both "Small text" and "87% of prior size"? They are both smaller for me in Safari and Chrome on Windows. I don't have a Mac for testing. Which browser versions is it? PrimeHunter (talk) 11:35, 21 August 2012 (UTC)
Sorry Prime and DJ, I should have been more explicit. Yes, exactly those syntaxes you've illustrated. It's Mac OS 10.7.4 (latest), Safari 6.0, and Chrome 21.0.1180.79 (both latest, I presume), For example, the Signpost's News and notes just out, which I worked on, has 90% size for both indented block quotes. I previewed 60% just in case my eyes were not seeing it. Still no change from the normal text size on my default Safari. Tried on FF (works) and Chrome (doesn't reduce at all, like Safari). I've noticed this problem with the "small" syntax too, first about four to six months ago, at a guess. Now it's becoming an issue for me, and I'm wondering how many of our readers are not seeing these syntaxes in action. Thanks. Tony (talk) 11:48, 21 August 2012 (UTC)
Do you by any chance have the "No small fonts" gadget enabled? If not, check the "Minimum font size" setting in Chrome/Safari, which may also interfere. Edokter (talk) — 20:22, 21 August 2012 (UTC)
Thanks so much, Edoketerl. It's fixed! Tony (talk) 21:00, 21 August 2012 (UTC)

Template "Skip to bottom" changes

FYI, I twicked the way Template:Skip to bottom works so that it can be placed at any place on the page. For a working example of the new options this opens see my talk page, where I used the template to place a "back to TOC" link at the end of each section. →Yaniv256 wind roads 08:06, 21 August 2012 (UTC)

Template:TOC limit functionality

I received and interesting question via e-mail today about the functionality of {{TOC limit}} at Military history of Canada . Currently when this template is used it hides the sub headers with the option only for [hide] not show all - I was asked if there is a template that will allow our readers to expand the TOC box to show the hidden sub headers at will. I personally never though of this until it was mention - but it would be a good idea to have more functionality of this template to allow our readers to see all the sub-headers if they like. Something like [show all]/[hide] options?Moxy (talk) 20:37, 21 August 2012 (UTC)

Such a button would have to be implemented by Javascript and CSS (e.g. as a gadget). Having all of the levels visible at all times on all pages requires a simple change to a person's CSS file. --Izno (talk) 23:28, 21 August 2012 (UTC)

Display size change

All of a sudden, my Wikpedia windows are displaying in a tiny size of font. My browser (Firefox 14 in a basic display mode - no fancy skins) is still set to 16pt, and other websites that I go on are as before both in display and editing. When I change my browser size to 18, things go bigger, but look odd. Changing to 14 makes no difference from the 16. Nothing has been changed on my machine - no installation or removal, no updating, no nothing. Just happened. OS is XP Pro in Classic view, and I use Monobook. The 'Delete' reason box that I had trouble with a while back is normal size. As I say, other sites are unaffected. Only Wikipedia has shrunk. Any ideas? Is there a way of changing one's display size here? I can't see it in Preferences. Just tried IE 8 - looks OK there. Odd. Peridon (talk) 19:54, 21 August 2012 (UTC)

Try Ctrl+0 - does that fix it? --Redrose64 (talk) 20:07, 21 August 2012 (UTC)
Try a reboot. Messed up fonts are a common symptom of a computer in need of a reboot. StuRat (talk) 20:12, 21 August 2012 (UTC)
Ah. Ctrl-0 did the trick. Thanks. Any chance of an explain? Peridon (talk) 20:24, 21 August 2012 (UTC)
Might be related to this. NtheP (talk) 20:44, 21 August 2012 (UTC)
(edit conflict) You somehow zoomed in (by accidentally hitting Ctrl++ or some other way). BTW, StuRat's suggestion is more reliable, and I strongly recommend using it next time. At least that would allow to fix problem quicker and with less people involved. — Dmitrij D. Czarkoff (talk) 20:47, 21 August 2012 (UTC)
A common way of accidentally zooming is to have the Ctrl key down whilst you roll the mouse wheel. Less accidentally, Ctrl++ zooms in, and Ctrl+- zooms out. --Redrose64 (talk) 21:40, 21 August 2012 (UTC)
(ec)I'd have had to hit Ctrl dash... (Very unlikely - I'd have had to mis-hit two keys at once, and I'm a keyboard watcher - started off on typewriters where you don't get much of a second chance. I've looked at NtheP's link, and think I'll just remember Ctrl-0 rather than get into all that, but thanks, I now know what they were up to. But not why. Thanks, everyone. Peridon (talk) 21:52, 21 August 2012 (UTC)
Ctrl and mouse wheel is probably a more common way to accidentally change size. PrimeHunter (talk) 00:14, 22 August 2012 (UTC)
Some laptop touchpads now also change the text size if you make a 'pinch' gesture, probably quite easy to do accidentally if you don't know about it. the wub "?!" 15:08, 22 August 2012 (UTC)

VisualEditor/Parsoid fortnightly update - 2012-08-20 (MW 1.20wmf10)

Hey all,

I've been conscious for a while that the VisualEditor project is a major change that is coming down the tracks, and that I need to do more to make sure you all know what is happening (and make sure you have as much opportunity to tell us when we're wrong, as well as help guide the priorities for development and improvement).

Right now (and for the past few months), both the core of MediaWiki and all the extensions on Wikimedia's cluster are updated on the live site every fortnight in a new branch. Though VisualEditor is currently only deployed to, in time (when it's ready) we'll be making it available here, and on all other Wikimedia wikis. I've started to write up the technical "changelog" as a status update, giving some (hopefully) more useful, human-readable "release notes". (;-))

I thought it might be useful to copy them here — below, both for the VisualEditor tool and the new Parsoid service that is separate but vital for it to work — but I'm worried that they might be a bit spammy, so I'd love feedback not just on the content below (for which the best venue is the central feedback page) but also as to whether you would value me doing this every fortnight (or perhaps less regularly?), or if there are better, or additional fora that might be suited to getting this information for users on the English Wikipedia.


The VisualEditor was updated as part of the wider on Monday 20 August MediaWiki 1.20wmf10 branch deployment.

The most visible new item in the two weeks since 1.20wmf9 is the much-improved link inspector. This now guides users to create a link to a suggested existing article, a redlink or an external link, and replaces the previous basic functionality that did not suggest links or inform you if the target of your link existed. We have also improved the save dialogue, streamlining the interaction based on feedback from the design team.

There have also been a number of bug fixes, such as preserving spaces before and after the content in headings and other forms (so that "== Foo ==" doesn't have spacing either side of it in the editor display, but doesn't strip them either — 37935), using browsers' native deletion mechanisms which helps with support for short-cuts and internationalisation (38461), and handling of "alien nodes" (pieces of content that the editor does not know how to edit yet) so that they do not break the rest of the editor when included. However, most of the changes have been improvements to the code architecture to allow it to be re-used and extended to support new 'node types' like categories or tables when we work on these later.

A complete list of individual code commits is available in the 1.20/wmf10 changelog, and all Bugzilla bugs closed in this period on Bugzilla's list.


The Parsoid team worked on the final tasks in the JS prototype, in preparation for the C++ port. The port will allow an efficient integration with PHP and Lua, improve performance and allow the parallelization of the parser in the longer term in preparation for production use.

An important milestone we reached is the implementation and verification of the template DOM range encapsulation algorithm, which now identifies all template-affected parts of the DOM for round-tripping and protection in the mw:VisualEditor. We are currently implementing template round-tripping based on this. Other new features include oldid support so that previous versions of pages can be edited, rather than just the current one, and more complete error reporting in the web service. Wikitext escaping in the serializer is much improved, and now also handles interactions across multiple DOM nodes. An ongoing task has been improving test coverage to enable us to refactor code with more confidence and also help test the correctness of the C++ port.

Most details of the C++ port were researched. A basic build system including the selected libraries was set up, and design work on the basic data structures has started, ahead of full porting which we expect to start next iteration.

The full list of Parsoid bugs closed in the last two weeks is available in Bugzilla.

Hope this is helpful! As I said, feedback gratefully received.

Jdforrester (WMF) (talk) 22:25, 21 August 2012 (UTC)

Thanks for this update.
If you could post this to as well, that would be great. --MZMcBride (talk) 00:01, 22 August 2012 (UTC)
Of course - done. Jdforrester (WMF) (talk) 00:31, 22 August 2012 (UTC)
This is wonderful, it would be great to see such regular updates (although they can be monthly for me. —TheDJ (talkcontribs) 07:24, 22 August 2012 (UTC)

Education Program extension RfC

I've just opened up a request for comment on whether to enable the Education Program extension for managing and monitoring Wikipedia educational assignments. If it does get enabled, there are related technical (user rights) and policy (who should be able to use it, and how will user rights be assigned?) issues that will need to be sorted out. (For that reason, I'm cross-posting it here.)--Sage Ross (WMF) (talk) 12:53, 22 August 2012 (UTC)

article display class wrong with gadget on

hiya all

this article Rangers_F.C._Supporters ive rerated c class from start however i have the gadget on i cant remembewr the name so it shows the article rating but it saying stub class, i have tried refreshing and purging cache still the same, tri4d on firefox to problem started on chrome "Rangers F.C. Supporters A stub-class article from Wikipedia, the free encyclopedia"--Andrewcrawford (talk - contrib) 15:52, 22 August 2012 (UTC)

Do you mean the gadget described as "Display an assessment of an article's quality in its page header (documentation)"? --Redrose64 (talk) 16:35, 22 August 2012 (UTC)
For me it displays C-class. You should try to clear your cash one more time. Ruslik_Zero 18:56, 22 August 2012 (UTC)
working now not sure why it wasnt parsing the data right--Andrewcrawford (talk - contrib) 20:43, 22 August 2012 (UTC)

Left margin stepping on top of article

An editor slapped a template:failed verification on Willian Borges da Silva, and the page display went haywire. What's up? Wbm1058 (talk) 16:48, 22 August 2012 (UTC)

Like how? It looks OK to me. --Redrose64 (talk) 17:00, 22 August 2012 (UTC)
Never mind. Now it displays fine in my Google Chrome browser, even though no edits have been made to the page, and a few minutes ago it was consistently displaying corrrupted with several page refresh attempts. Seems a transitory problem that cured itself. Wbm1058 (talk) 17:02, 22 August 2012 (UTC)

White space

How do I create a half width line to create white space? I want to do it to Template:Botanist (edit | talk | history | links | watch | logs) to improve its appearance in articles.. -- Alan Liefting (talk - contribs) 21:10, 22 August 2012 (UTC)

Add padding. --Izno (talk) 22:50, 22 August 2012 (UTC)
Sweeeet... Cheers. -- Alan Liefting (talk - contribs) 23:07, 22 August 2012 (UTC)

Another coding glitch

I encountered another glitch in the coding and I am wondering what is going on? I've encountered this on articles and other pages around Wikipedia. See the image as an example, although it hasn't just happened to the contributions. Simply south...... flapping wings into buildings for just 6 years 17:39, 20 August 2012 (UTC)

Contributions glitch.png
One character in the HTML that was served to you has become corrupted. Specifically, the opening < of the element <a href="/wiki/Special:Contributions/Judetheobscure1" title="Special:Contributions/Judetheobscure1">contribs</a> has been changed into what appears to be a pipe character, so everything in the opening tag has been presented as plain text. Anyway, when I look at the same entry in my watchlist, it's normal. If the rogue character is a pipe, this could be a s simple as changing a single bit from 0 to 1, because < is the same as &#x3c; and | is the same as &#x7c; - a difference of 0x40, or 64 in decimal currency. --Redrose64 (talk) 19:32, 20 August 2012 (UTC)
What the, that should definitely not happen. Bits should not get randomly flipped. Bawolff (talk) 19:47, 23 August 2012 (UTC)

Revision history in mobile browsers

I was told about this problem by a friend, so forgive me if it isn't spelled out in the most complete terms. Basically, what I was told is that the previous revisions of various articles (Mystique (comics) was mentioned) are, on the Safari browser for iPad, displaying content that is actually the same as the most recent revision of the article in most cases. In other words, people looking for previous revisions of an article are actually loading pages that contain the present content of the article, aside from the "This is an old revision" note at the top of the article.

I can't think of any reason this should be, or how such a malfunction would actually take place on a technical level, but I have seen the problem in action with my own eyes. Really weird, so I thought I should tell everybody. Thanks. Evanh2008 (talk|contribs) 10:56, 23 August 2012 (UTC)

Convert Village Pump to Wikia Community Forum software

Note: Moved to Wikipedia:Village pump (proposals) by original poster (me) per discussion. --Timeshifter (talk) 20:30, 26 August 2012 (UTC)

Update on Lua/Scribunto

The long-awaited Lua templating system has been deployed to Please help test it there. It needs performance and usability testing before it can be rolled-out to Here are some useful links:

Kaldari (talk) 21:38, 22 August 2012 (UTC)

Exciting! All coders and template hackers should check this out, since it opens up vast possibilities for template programming. — This, that, and the other (talk) 11:13, 23 August 2012 (UTC)
Just a note. I think given the vast number of templates on wiki we are going to need to devise a way of identifying which ones have been converted or not into the new Lua format. I am also not that familiar with this language so I think we are going to have some growing pains for a while and it might be necessary to setup a Template transition project to help identify problems, devise a strategy for converting the templates, identifying which ones are done/need to be done etc. Kumioko (talk) 16:25, 24 August 2012 (UTC)
Lua's not intended to replace every template, only the more gratuitously convoluted and expensive ones. Probably a maintenance template with a category Category:Templates to be converted to Lua, or similar, would be sufficient. Happymelon 16:35, 24 August 2012 (UTC)
OK thanks. I can think of quite a few but do you know if there is a server process or something that the developers use to determine which ones are putting the most load on the servers/database. It seems like someone must have something somewhere and that would be a good place to start. Cite templates for example would be on the top of my list. Kumioko (talk) 18:39, 24 August 2012 (UTC)
Lua will mostly benefit templates that are currently 'deep'. Everything that is 'flat' will likely still be easier and simpler to do with templates. Things that likely qualify: many math related templates or templates that use math templates, the cite family, some mbox templates and the main infobox template, coordinates (which I actually already tackled for a a large part) etc. A WikiProject around this is indeed a good idea. I'll try to finally post my experience of converting the coordinates template during the Berlin Hackathon, I think it might provide a bit more insight. —TheDJ (talkcontribs) 19:35, 24 August 2012 (UTC)
What libraries are available in the Lua framework that MediaWiki will be using? If certain ones are enabled, it is possible to retrieve the browser session cookies and forward them to another site—not what I'd like to have a vandal do on a highly-used template. Reaper Eternal (talk) 19:43, 24 August 2012 (UTC)
Also a good question and points. I think as this starts to develop it will make having a WikiProject, Village pump page or something much more necessary to be able to discuss changes and post information. Having it on Wikia is ok but IMO that should stay more general and cross project where this one would be more specific to En Wikipedia. Kumioko (talk) 19:56, 24 August 2012 (UTC)
It's executed purely server-side and still goes through the parser -- you shouldn't be able to emit anything that you can't already in a page. Usable libraries and functions are also very restrictive. See mw:Extension:Scribunto#Lua environment and mw:Extension:Scribunto/API_specification for further details. Amalthea 20:06, 24 August 2012 (UTC)

Scroll galery


per {{Scroll gallery}}: "This template is currently not functioning, as the required JavaScript code to use it is not integrated into the English Wikipedia. " - please could you do it? It is a very useful template. Regards.--Kürbis () 15:38, 23 August 2012 (UTC)

As it says on the template page it doesn't work with Internet Explorer and while that is still the case the necessary JavaScript is unlikely to be introduced. If Microsoft fix IE might be different. NtheP (talk) 15:54, 23 August 2012 (UTC)
That and you'd have to wait for consensus to actually support it. Come back if WP:VPR#Scroll galery actually demonstrates consensus, don't forum shop. Anomie 17:39, 23 August 2012 (UTC)
That wasn't forum shopping. Per the target of the link you provided, the user was directed to post here. --Nouniquenames (talk) 18:43, 23 August 2012 (UTC)
@Anomie what are you talking about? Regards.--Kürbis () 09:31, 25 August 2012 (UTC)

Mysterious image problem

As of 19:00, 23 August 2012 (UTC)…
This exists:
This doesn't: File:FK Sutjeska Foca.gif

See the problems faced here.

How could this be possible? A server problem? We are supposed to be able to call up any file which exists via the search box, without needing the entire URL. What is the problem here? (talk) 19:00, 23 August 2012 (UTC), last modified 19:01, 23 August 2012 (UTC)

It's here: lt:File:FK Sutjeska Foca.gif, i.e. it hasn't been uploaded to en.wp or at commons so it can't be used. Amalthea 19:07, 23 August 2012 (UTC)
Can a transfer request be filed? (talk) 00:50, 24 August 2012 (UTC)
It's probably non-free so the easiest would be to just upload it here as well -- see e.g. File:F.C. Porto logo.png for how the fair-use reasoning might go. Amalthea 20:10, 24 August 2012 (UTC)
I am an IP editor and am unable to upload. May I file a request for this? (talk) 21:06, 24 August 2012 (UTC)

Hidden text in diffs

Is it just me or are all changes to hidden text now rendered invisible in diffs? For example, can anyone tell that this edit added 233 bytes of hidden text (I can't, unless I check the revision history). Evanh2008 (talk|contribs) 01:47, 25 August 2012 (UTC)

Hidden text, or commented text? Commented text, yes, there's approximately that much commented text added in that diff. What browser, version, and OS? --Izno (talk) 02:28, 25 August 2012 (UTC)
Firefox 14.0.1, Windows 7 64-bit. And I thought hidden text and commented text were the same thing. Anyway, that's what this seems to be saying. Evanh2008 (talk|contribs) 03:04, 25 August 2012 (UTC)
I also use Firefox 14.0.1, but my OS is Windoze XP. I see the following extra text: <!-- Please don't change it to Australian, while is known for including references of the country like the Great Barrier Reef, Sydney and having one to three actors that originated from Australia it is considered an American film--> --Redrose64 (talk) 09:09, 25 August 2012 (UTC)
Odd. I'm also seeing signature formatting in diffs. Where before I would just see bare code, I'm actually now seeing signatures with customised colors and formatting wholly formed. There were a few of the recent diff display changes I opted out of, but that was months ago. Could that have something to do with this? Evanh2008 (talk|contribs) 09:25, 25 August 2012 (UTC)
Sounds like you have some Javascript installed or something, as I'm also running Fx 14.0.1 on Windows 7 64b and I'm seeing the same as Redrose. Have you enabled a new gadget recently or something? --Izno (talk) 18:51, 25 August 2012 (UTC)

Status template

At User:Thekillerpenguin/vector.js, did I get the script to work right? I want to be able to update my status with this template. Tell me if there is any problem, and I'll correct it. Thekillerpenguin (talk) 03:31, 25 August 2012 (UTC)

Sorting tables with sort template

Note: Moved to Help talk:Sorting. — Preceding unsigned comment added by Timeshifter (talkcontribs) 22:27, 25 August 2012 (UTC)

WP Editor loses main text when going to a different page, II

Follow up to Wikipedia:Village pump (technical)/Archive 102#WP Editor loses main text when going to a different page.

In the last 36 hours this has started happening for me in (almost?) every page that I edit. No matter how little text there already is in the page or thread, nor how small my additional text, if I hit "Show preview" then the "back" button, my edit is lost. Until 22 August, it remembered it for one or sometimes two levels of "back"; and some months ago, it would remember it for as many as six (or more?) levels. Firefox 14.0.1, windows XP. Example pages: Wikipedia talk:WikiProject Stations; Tenby railway station; Template talk:Citation; Wikipedia:Village pump (technical) so it's not just discussion pages nor pages with edit notices.

Something is definitely instructing the browser to discard changes. --Redrose64 (talk) 14:57, 24 August 2012 (UTC)

No problems with Safari for me. —TheDJ (talkcontribs) 19:27, 24 August 2012 (UTC)
I'm getting the same problem with Firefox 3.6.6 (example page Grantham railway station) - this is a machine which I've deliberately avoided upgrading. When I last used it, two or three weeks ago, it was definitely remembering the changes to edit window content. Now, if I go for "show preview" then "back", they're being discarded. --Redrose64 (talk) 13:36, 27 August 2012 (UTC)

Edit conflict results in "Connection closed unexpectedly"

I reliably get "Connection closed unexpectedly" error if I manually merge my contribution after edit conflict. Is this a problem with my browser (Web) or a Mediawiki bug? — Dmitrij D. Czarkoff (talk) 00:35, 26 August 2012 (UTC)

It's just conceivable that it's a Mediawiki bug, but I think the first suspect would be your local software. I suggest you try using a different browser, like Firefox or Midori, to see if this is the case. If you don't get the problem in those browsers, then the problem is likely to be with your browser. If you can reproduce the problem with other browsers, let us know, and we can take a closer look. -- The Anome (talk) 12:04, 26 August 2012 (UTC)
I don't experience the problem on this particular configuration with another ISP, so this issue is almost certainly unrelated to Mediawiki. — Dmitrij D. Czarkoff (talk) 13:11, 26 August 2012 (UTC)
In all likelihood, it's some sort of configuration setting of the router in your home. It detects something that it thinks is 'bad' or might be an indication of an 'attack' on your network and terminates the connection. Issues like these are actually quite common with the gazillions of ADSL and cable routers in the world. Many of them have quite serious bugs and the consumer is left in the dark with the problems. —TheDJ (talkcontribs) 14:00, 26 August 2012 (UTC)
I have this problem only with 3G (which I'm forced to use for a while). ADSL configuration works flawlessly. — Dmitrij D. Czarkoff (talk) 16:04, 27 August 2012 (UTC)

Problem with the display of images

There is long standing problem with the display of images. Normally, when an image is displayed, any following text is wrapped round the image with some padding to prevent the text butting up too close to the image. However, when an image is left justified, bulleted items are not padded. There is an example here. Is it possible for this feature to be fixed? --Epipelagic (talk) 03:47, 26 August 2012 (UTC)

This is a problem with HTML/CSS. Bullets are not part of the 'content' of a list, they are just outside of it. Because of this it can overlap with other content. This is not a bug, this is how it's supposed to behave, it just doesn't look pretty. See also bugzilla:11782. We can add "list-style-position:inside;" but that changes the layout of lists significantly. I'm not sure it's desirable. It also doesn't fix the indentation, it just avoids that the bullets overlap the image. So it's not really bug, it's more 'undesired' behavior, but I know of no good way to define the proper behavior. —TheDJ (talkcontribs) 11:58, 26 August 2012 (UTC)
Perhaps the MOS for images needs to include something like: "Do not position a left aligned image next to text formatted as a bulleted or numbered list." Roger (talk) 12:12, 26 August 2012 (UTC)
Created a proposal for a "wrapper" class/template that you could put around a specific list to fix such an issue. It's not perfect, you'd want to solve this everywhere automatically, but that seems to be impossible. —TheDJ (talkcontribs) 12:56, 26 August 2012 (UTC)
Please see {{Flowlist}}, for dealing with such a problem. —TheDJ (talkcontribs) 12:05, 27 August 2012 (UTC)

A hidden category that's not

Category:Wikipedia pages needing cleanup from November 2009 is marked as hidden yet it shows up on at least two pages. -- Alan Liefting (talk - contribs) 00:45, 27 August 2012 (UTC)

Which pages? — Dmitrij D. Czarkoff (talk) 00:53, 27 August 2012 (UTC)
All the ones that I checked - Use of performance-enhancing drugs in sport, Matt Butcher, and some other random page. -- Alan Liefting (talk - contribs) 00:58, 27 August 2012 (UTC)
And now? — Dmitrij D. Czarkoff (talk) 01:07, 27 August 2012 (UTC)
Fixed! WOW! -- Alan Liefting (talk - contribs) 01:15, 27 August 2012 (UTC)

This category was created by user:AnomieBOT, who just placed {{Monthly clean up category}} with no parameters there, while this template only makes category hidden if |hidden= contains any value. The rest of "Wikipedia pages needing cleanup from MONTH YEAR" are equally not hidden BTW, so I invited Anomie (bot's operator) to this discussion. — Dmitrij D. Czarkoff (talk) 01:19, 27 August 2012 (UTC)

That analysis is incorrect, otherwise Category:Wikipedia pages needing cleanup from November 2010 would be showing up non-hidden on all the pages in that category. You can also look at Template:Monthly clean-up category and see that the default value for |hidden= is "y". Next time, try a purge or a null edit on the articles affected; it may be a manifestation of bug 5382 or the like. Anomie 02:24, 27 August 2012 (UTC)
Interestingly, purging pages before adding |hidden=y didn't help. Probably all maintenance categories created before 10:00, 30 May 2010 (UTC) should be purged instead. — Dmitrij D. Czarkoff (talk) 09:02, 27 August 2012 (UTC)
Why is that? Category:Wikipedia pages needing cleanup from November 2009 wasn't created until 5 July 2012, and it supposedly had a problem. Anomie 10:49, 27 August 2012 (UTC)
Oh, then it is indeed most likely to be solved with articles' purging. — Dmitrij D. Czarkoff (talk) 16:27, 27 August 2012 (UTC)


I stumbled into this page which only seems to contain garbarge. It is not accessable as a normal page: Dummkopf schicklgruber.
What is this? Should it be deleted? If so, could someone with the appropriate rights do that. Jan Arkesteijn (talk) 21:17, 27 August 2012 (UTC)

It's just a landing page. Every deleted page has it if you follow the url to the link. See this for example. Following your red link won't get you there because red links bring you to the edit page. Ryan Vesey 21:20, 27 August 2012 (UTC)
You don't say what you mean by "garbage" so I don't know whether we see the same at I see a deletion log in a red box, followed by a message starting with "Wikipedia does not have an article with this exact name." I wouldn't call that garbage. The url returns a HTTP 404 Not Found error message (this can for example be checked with, so search engines and other programs who stumble on the url can see there is no real page. PrimeHunter (talk) 23:31, 27 August 2012 (UTC)
Before, it showed a sequence of unreadable characters, but now it shows this box. Someone seems to have flipped a switch. Jan Arkesteijn (talk) 16:36, 28 August 2012 (UTC)

Watch/unwatch box

I noticed the box that pops up when you click the 'watch'/'unwatch' tab has changed. Is there a script to put (on Monobook, here, if that matters) in the CSS or whatever to go back to the old one? - The Bushranger One ping only 02:01, 28 August 2012 (UTC)

I like it. That said, it would need to be Javascript, I think. --Izno (talk) 02:05, 28 August 2012 (UTC)
Changed in what way? I see no difference. I still see the box that pushes down the page. I do like the popup box on Commons and other projects. Edokter (talk) — 09:31, 28 August 2012 (UTC)
Never mind... cleared the cache, now I see the new watchlist message. Edokter (talk) — 09:36, 28 August 2012 (UTC)
I've asked over at Help talk:Watching pages#New system? Where when what - hopefully someone there will know. -- Quiddity (talk) 23:53, 28 August 2012 (UTC)
Probably due to gerrit:17605. Anomie 00:05, 29 August 2012 (UTC)
Yes, that is it.
And the remaining improvements will be available when gerrit:19199 is merged. There is a video demonstrating the features: mw:File:Mw-notification.ogv. Helder 01:00, 29 August 2012 (UTC)
Thanks, there a script to go back to the old way? Face-smile.svg - The Bushranger One ping only 06:26, 29 August 2012 (UTC)

Google Chrome animation

Recently, I'm getting a fly-over kind of animation whenever I load a page on Wikipedia on Google Chrome (v21). Is it a new feature of Wikipedia or a bug in Chrome? Anyone else experiencing the same? EngineerFromVega 05:53, 28 August 2012 (UTC)

Can you be more specific? What do you mean with "fly-over kind of animation"? Edokter (talk) — 09:37, 28 August 2012 (UTC)
Basically, a Wikipedia page flows from right-to-left (similar to Powerpoint effect), instead of appearing like a normal webpage. This happens only when a page is not in cache, so reloading a page will not show this effect, but a 'purge' will show it. EngineerFromVega 13:07, 28 August 2012 (UTC)
Wikipedia does not have any kind of animation that involves the entire page. Edokter (talk) — 13:16, 28 August 2012 (UTC)
Does not happen in Firefox 15 Bulwersator (talk) 14:38, 28 August 2012 (UTC)
This happens only in Chrome v21. Looks like a bug in Chrome. I'll wait for the next update or try disabling my extensions one by one. EngineerFromVega 04:44, 29 August 2012 (UTC)
I'm not seeing this with Chrome 21.0.1180.83 m in Monobook or with ?useskin=vector&action=purge added to the URL. Mark Hurd (talk) 16:02, 28 August 2012 (UTC)

Gadget request - hide "Feedback from my watched pages" on watchlist

I think that tool to remove this remainder about WMF incompetence would be welcomed gadget Bulwersator (talk) 09:47, 28 August 2012 (UTC)

Overkill for a gadget. If you want to hide it, put this in your CSS:
#articlefeedbackv5-watchlist-feedback-link {
    display: none;
Edokter (talk) — 09:50, 28 August 2012 (UTC)

Article Feedback on redirects

Why is the Article Feedback Tool on redirects? See my feedback on Morris Johnson. 2001:18E8:2:1020:D59A:53C3:4F74:18D8 (talk) 14:27, 28 August 2012 (UTC)

Morris Johnson isn't a redirect. David1217 What I've done 17:00, 28 August 2012 (UTC)
Now logged in It was when I tagged it; look at the page history and you'll see that it was moved after I left the above comment. Nyttend (talk) 17:57, 28 August 2012 (UTC)
It seems plausible to me that an editor could leave feedback on whether it should have a new target, whether it should be created. Admittedly items which could be placed on the talk page, but then, AFT is kind of not great for duplicating that particular method of feedback. --Izno (talk) 23:06, 28 August 2012 (UTC)



Is it possible to include an option to also show, say 1000, 2000 and 5000 results per pages? -- Toshio Yamaguchi (tlkctb) 17:13, 28 August 2012 (UTC)

Check out . Add &limit= to the end of your url to modify it to whatever you want. Ryan Vesey 17:15, 28 August 2012 (UTC)
Wow, awesome. I didn't know this is possible. Thanks. -- Toshio Yamaguchi (tlkctb) 17:17, 28 August 2012 (UTC)

Header & tabs shrunken

When I accessed the Deaflympics article, the tabs on the top of the article (Article, Talk, etc.) seem to be shrunken. I've checked my zoom distance and it all looks fine. That is the only article that is affected. I'm using Firefox on MacBook Pro using the Vector skin. -- Luke (Talk) 18:54, 28 August 2012 (UTC)

That was odd- all but the article body was smaller. I purged it and it looks OK now. ---— Gadget850 (Ed) talk 19:20, 28 August 2012 (UTC)
  • Amazing. Was HTMLTidy broken for a while?
    Here's how the page source looked at first (note the unclosed small-tag that the Firefox then tried to correct):
<p><small><sup>1</sup> The <a href="/wiki/Republic_of_China" title="Republic of China" class="mw-redirect">Republic of China</a>, commonly known as <a href="/wiki/Taiwan" title="Taiwan">Taiwan</a>, due to complicated relations with <a href="/wiki/People%27s_Republic_of_China" title="People's Republic of China" class="mw-redirect">People's Republic of China</a>, is recognized by the name <a href="/wiki/Chinese_Taipei" title="Chinese Taipei">Chinese Taipei</a> by the <a href="/wiki/Comit%C3%A9_International_des_Sports_des_Sourds" title="Comité International des Sports des Sourds">ICSD</a> and the majority of international organizations. Nevertheless, Taiwan is a geographic location.
Then I purged it (yes, only purged, I only fixed the page source afterwards) and Tidy kicked in:
<p><small><sup>1</sup> The <a href="/wiki/Republic_of_China" title="Republic of China" class="mw-redirect">Republic of China</a>, commonly known as <a href="/wiki/Taiwan" title="Taiwan">Taiwan</a>, due to complicated relations with <a href="/wiki/People%27s_Republic_of_China" title="People's Republic of China" class="mw-redirect">People's Republic of China</a>, is recognized by the name <a href="/wiki/Chinese_Taipei" title="Chinese Taipei">Chinese Taipei</a> by the <a href="/wiki/Comit%C3%A9_International_des_Sports_des_Sourds" title="Comité International des Sports des Sourds">ICSD</a> and the majority of international organizations. Nevertheless, Taiwan is a geographic location.</small></p>
Amalthea 19:24, 28 August 2012 (UTC)
Hmm, it now looks fine. -- Luke (Talk) 19:26, 28 August 2012 (UTC)

Something wrong with Vector skin?

I just noticed that the top gradient image for the Vector skin (page-fade.png, loaded as background image for #mw-page-base) is not displaying properly in any browser; it simply shows a a thin gray band. It is not a problem with ResourceLoader, as the static link (which is valid) has the same problem. I suspect the image may be corrupted, but I can't find any fault in the image itself. Edokter (talk) — 09:28, 27 August 2012 (UTC)

Never mind... (Anyone can tell me how my display got reset to 16-bit color?) Edokter (talk) — 10:04, 27 August 2012 (UTC)
Nobody can unless you at least mention at least your OS. — Dmitrij D. Czarkoff (talk) 10:30, 27 August 2012 (UTC)
XP SP3 with ATI Radeon 4650 (and old drivers). Edokter (talk) — 11:05, 27 August 2012 (UTC)
And you wonder, why something breaks? Well, definitely it is a problem of your "OS — driver" combination, and your system logs ("Start => Administrative tools => Computer management" IIRC) could tell you something. — Dmitrij D. Czarkoff (talk) 11:16, 27 August 2012 (UTC)
That's quite enough Windows-bashing, thank you. The old drivers are there for a reason. Anyway, I traced the fault to a driver configuration error in the registry, which was introduced after I took out the card for a good clean. After reassembling, the BIOS thought it was a good idea to assign a new IRQ, mismatching the one in the registry, which sent the driver in fail-safe mode. I purged the BIOS and registry, reset all screen settings and all is well again. Edokter (talk) — 11:58, 29 August 2012 (UTC)

Small TeX formulas

When making a TeX formula, in the case of a fraction it is possible to make the formula smaller by using \tfrac instead of \frac (notice the difference between and ). Is there a general method to make a formula smaller? At User:Toshio Yamaguchi/Wieferich prime#Connection with imaginary quadratic fields I have a number of TeX formulas and want to make them smaller. Is there something I can apply to make these smaller as well, like I did with the fraction? -- Toshio Yamaguchi (tlkctb) 11:19, 29 August 2012 (UTC)

Note: Both renderings look identical using the MathJax rendering option instead of the default PNG rendering. (My preferences: Appearance: Math = N"Always render PNG", N"Leave it as TeX (for text browsers)", YesY"MathJax (experimental; best for most browsers)"). — Richardguk (talk) 11:43, 29 August 2012 (UTC)
I just enabled MathJax in my preferences and now both formulas look the same .... the same crappy. You can barely decipher anything anymore. As a note, I am using Safari. -- Toshio Yamaguchi (tlkctb) 11:48, 29 August 2012 (UTC)
I also enabled MathJax, and I would say that I have no problems with distinguishing individual symbols. That said, MathJax might be a bar for editors with substandard vision capabilities. Larger rendering appears on leftclick. I use Web. — Dmitrij D. Czarkoff (talk) 21:34, 29 August 2012 (UTC) updated 21:36, 29 August 2012 (UTC)

Huge template is huge

Can someone tell my why the navbox at the bottom of McBride & the Ride is showing up so absurdly long, yet it wraps properly on other articles it's used on (e.g. Burnin' Up the Road)? Ten Pound Hammer(What did I screw up now?) 18:30, 29 August 2012 (UTC)

Does it look better now.Blethering Scot 18:35, 29 August 2012 (UTC)
Eeyup. Ten Pound Hammer(What did I screw up now?) 18:38, 29 August 2012 (UTC)

Proposal to turn on a new extension

Hey all :). I've just started a Request for Comment on the Memento MediaWiki extension, which provides an invisible backend protocol that lets people go back in time when browsing Wikipedia's content in a standardised format shared by a lot of other services. If you've got a moment, we'd really appreciate your opinion on if it should be turned on or not :). Thanks! Okeyes (WMF) (talk) 19:24, 29 August 2012 (UTC)

What makes wikipedia too much heavy to my browser...

...when compared to her sister projects? Well, occasionally, I cannot access WP due to "connection timed out". But I can access other sister projects and log in. Also when logged in (from other projects), the logging you to other Wikimedia projects shows a broken image for WP, but loads logos for all others indicating I am not logged into WP. Even when I get enough speed to access WP, it takes very much time to load a page completely (occasionally, twinkle and popups does not get loaded). All other projects give a faster loading of pages, very quick indeed. Why is this so. Can I take any steps to reduce this occasional timing out of WP. (Well my browser is Google Chrome, but I had the same experience when using Seamonkey) ···Vanischenu「m/Talk」 01:53, 21 August 2012 (UTC)

We use a lot of citation templates on the English Wikipedia (sometimes hundreds in one article), which slows down load time considerably. I've just given up trying to edit an article where preview and diffs were taking 30 seconds plus to load. Whether that's causing your problems I don't know, but when you next experience them, take a look to see whether the page is template-heavy (go into edit mode and search for the double squiggles that signify a template -- {{ -- ). SlimVirgin (talk) 19:49, 26 August 2012 (UTC)
Can it be a cache problem? I sometimes also get problems with Wikipedia being incredibly slow to load, and mostly (when it is not server updates or anything else server-end related) bypassing the cache seems to be able to fix it. --Saddhiyama (talk) 19:52, 26 August 2012 (UTC)

Thank you so much! I am not facing this "timing out" now. This problem occurred irrespective of language version and even for article which are not created yet, possibly longer for heavier articles.···Vanischenu「m/Talk」 22:17, 30 August 2012 (UTC)

Sidebar and top navigation panels disappearing

Anybody else having this problem? When I load my watchlist, the navigation links at the top and left side do not load. They load on other pages, just not on my watchlist. Can anyone help me out?(I am so totally not technical.) I'm on Firefox 14.0.1 Thanks, Cindy(talk to me) 13:19, 26 August 2012 (UTC)

Not sure if they are the cause of this, but the following recently-added lines should not be in your User:Cindamuse/modern.js file as they are not valid JavaScript (the first two are css, and the last line is a wikilink):
table.persondata {display:table !important;} { font-weight: normal !important; }

Richardguk (talk) 20:47, 26 August 2012 (UTC)
  • I appreciate the feedback and made the changes you suggested, but it didn't resolve the issue. Would you mind taking another look to see if there's anything else? I'm hoping you have additional suggestions. Thanks, Cindy(talk to me) 04:37, 27 August 2012 (UTC)
    • Try bypassing your browser cache (Ctrl+⇧ Shift+R on Firefox). Goodvac (talk) 05:21, 27 August 2012 (UTC)
      • Thanks for the help. I went back just to make sure and bypassed the cache again, but it didn't resolve the issue. Open to any other suggestions. Thanks, Cindy(talk to me) 05:43, 27 August 2012 (UTC)
Did you completely clear the cache as described at WP:BYPASS#Mozilla Firefox and other related browsers? What is your skin at Special:Preferences#mw-prefsection-rendering? Does work? Does worl? PrimeHunter (talk) 10:00, 27 August 2012 (UTC)
To find your skin, go to Preferences → Appearance and check the Skin section (Chick through Vector). Which one is selected? ---— Gadget850 (Ed) talk 11:47, 27 August 2012 (UTC)
In that case User:Cindamuse/modern.js is irrelevant. The problem may be in User:Cindamuse/monobook.js. I didn't get the problem you describe when I copied it to User:PrimeHunter2/monobook.js, but maybe something in User:Cindamuse/monobook.js interferes with something in your preferences. Try to blank it and see if it solves the problem. If it does then you can try to experiment to see which part of the code causes the problem. PrimeHunter2 (talk) 12:09, 28 August 2012 (UTC)
  • Thanks everybody for your help. Whatever steps I took with the last suggestion worked. Not sure exactly what the specific issue was, but it works now. Thanks again! Cindy(talk to me) 14:54, 30 August 2012 (UTC)

Image Upload Issue: Free, Non Free, Does not fit extra options are not expanding

I have been trying to upload a non free image file, but, non free options (and also the other options are not opening). So I can not proceed.

  • Uploading from: Wikipedia:Upload
  • File name in my computer: Roti film poster.jpg
  • File I am trying to name: Roti film poster.jpg
  • Article: Roti (1974 film)
  • File description filled: Yes
  • Error message: No
  • Operating System: Ubuntu 12.10
  • Browser: Firefox 14.01.

Tito Dutta 21:22, 29 August 2012 (UTC)

It has been fixed but you may need to bypass your cache to get the fix. See Wikipedia talk:File Upload Wizard#Bug report. PrimeHunter (talk) 11:13, 30 August 2012 (UTC)

Table border colour

Hi, I asked this at the Help Desk, but the question expired without being fully bottomed out. I'm wondering if someone here might be able to pick it up.

1. When I copy and paste the following HTML into a .htm file and open it from IE, it displays as expected with all red borders. When I view it in Wikipedia, the borders are all black:


Why this difference?

2. Similarly, the following wiki-code displays in IE with black borders, showing that the "bordercolor" attribute is not being picked up properly. Why is this?


3. In IE, the following shows a red border around the whole table, but the internal divisions are still black (I made the border wide because otherwise the border collapse obscures the outer red border, I think). Is this the way the "style=border" attribute is supposed to work? How does it look in other browsers?

Cell A Cell B
Cell C Cell D (talk) 01:54, 30 August 2012 (UTC)

MediaWiki and/or HTML Tidy removes various obsolete or unrecognized parameters, including "bordercolor". If you view the HTML source of this page, you'll see that the "bordercolor" is not output in your tables. Anomie 02:55, 30 August 2012 (UTC)
So what is the non-obsolete / recognised way of doing this? (talk) 10:55, 30 August 2012 (UTC)
The proper and recommended way of doing this is to put the following in a stylesheet:
table {
	border: 5px solid red;
	border-collapse: collapse;

td {
	border-color: inherit;
Keφr (talk) 11:04, 30 August 2012 (UTC)
I mean what is the way of doing it in Wikipedia articles. Is that syntax possible/desirable in Wikipedia articles? (talk) 11:10, 30 August 2012 (UTC)
It is definitely not desirable to have red, five-pixel-wide table borders. Typically, tables are marked with class="wikitable", to maintain a consistent style throughout articles. If you really want them, you may install a user style matching table.wikitable or, if you register, put it in Special:MyPage/common.css. Keφr (talk) 11:25, 30 August 2012 (UTC)
Sorry, I am not suggesting having 5-pixel-wide table borders, nor am I asking about setting the border width. I am asking about setting all the borders to the same colour. The colour red was chosen merely as highly obvious/visible for the examples. Thanks, I am aware of class="wikitable" for standard-format tables. (talk) 11:34, 30 August 2012 (UTC)
Something along the lines of tr, td { border-color: inherit } should be the best way to do it. A worse way would be to set border colour for each cell individually (| style="border-color: blue;" | contents ). Still, I cannot imagine a reason why the border colour should be changed from the default at all. Keφr (talk) 11:50, 30 August 2012 (UTC)
An example arose where, for whatever reason, it was undesirable to use class="wikitable". The question then arose as to how to hand-code the syntax to make all the borders grey rather than the default black (grey being preferable to the eye). (talk) 11:56, 30 August 2012 (UTC)
Grey is the default; specifically it's #aaaaaa or     . --Redrose64 (talk) 12:08, 30 August 2012 (UTC)
The following is grey? It looks pretty black to me in IE.
C D (talk) 12:56, 30 August 2012 (UTC)
So it does to me, but you did not use class="wikitable". Which is what Redrose64 was talking about. What was the specific example? Keφr (talk) 13:09, 30 August 2012 (UTC)
I specifically said not using wikitable. (talk) 14:07, 30 August 2012 (UTC)
I don't quite get the problem.
CSS and
HTML variant
Pure CSS
variant (wordier)
Are these OK? — Dmitrij D. Czarkoff (talk) 13:52, 30 August 2012 (UTC)
The first one does not work for me in IE9 (displays as black). I do not know if it works in other browsers, which is one of the things I would be interested to learn. The second works, and is a known workaround, but is laborious and inelegant since every cell has to be individually specified. (talk) 14:05, 30 August 2012 (UTC)

──────────────────────────────────────────────────────────────────────────────────────────────────── My 2¢. — Dmitrij D. Czarkoff (talk) 14:47, 30 August 2012 (UTC)

I have disabled that link, it opened in another window, and I do not appreciate "assistance" of that form. --Redrose64 (talk) 15:55, 30 August 2012 (UTC)
  • Check the link again. The link to the image was accurate and offered in good faith. Note the space between png/ and My 2¢. The phrase "My 2¢" is not a part of the url. Cindy(talk to me) 16:19, 30 August 2012 (UTC)
I know that spaces are not found in URLs; spaces are percent-encoded as %20. I didn't copy/paste it: I clicked on it. It went to the Imageshack page, which took almost a minute to load, during which time I heard my PC fan speed up, usually a sign of high CPU demand. Checking Windows task manager, I saw that CPU usage was through the roof: and under Applications, there was an instance of Firefox which I didn't recall starting. Switching to that, I saw the aforementioned partycasino page; and upon hitting Ctrl+H to get browser history, I saw that partycasino was the very next URL visited after the Imageshack one. I closed the window containing the partycasino page: and CPU usage plummeted. That is too much of a coincidence. --Redrose64 (talk) 16:35, 30 August 2012 (UTC)
@Redrose64 et al: sorry, I should have checked the link before posting.
I put the image to another hosting, with no popups. — Dmitrij D. Czarkoff (talk) 16:40, 30 August 2012 (UTC)
Needless to say, I am now disinclined to click that link. If it contained useful/relevant information then could anyone tell me what that was? (If not, don't bother.) (talk) 17:06, 30 August 2012 (UTC) ... oh, and by the way, I don't suppose anyone knows why "bordercolor" is not supported by Wikipedia when other similar table attributes ("border", "cellpadding", "cellspacing" etc.) are? Was it just an oversight? (talk) 17:10, 30 August 2012 (UTC)
"border", "cellpadding", and "cellspacing" are valid attributes, "bordercolor" is not. See [7], [8] (WP actually uses XHTML 1.0 Transitional, but that makes no difference here).—Emil J. 18:33, 30 August 2012 (UTC)
Oh. In my HTML book it is listed as "HTML 4", and at it says "BORDERCOLOR has been in use for a while as a proprietary Internet Explorer tag. It has just now been written into the HTML 4.01 code." (talk) 23:03, 30 August 2012 (UTC)
@ you asked about the way the tables are displayed in other browsers, and I posted a partial screenshot showing them. As the ads by ImageShack caused problems for Redrose64, I uploaded image to another hosting service and I hope no issues will be found there (at list for me this link leads to an image, not a page). — Dmitrij D. Czarkoff (talk) 22:11, 30 August 2012 (UTC)
Thanks for the screenshot. For you, it seems all lines are grey in both cases, yet in my IE9 they are not, so this seems to be a bug or difference in interpretation in one or other of the browsers. For interest, which browser were you using? I don't quite understand the significance of that "About Web" panel... (talk) 22:58, 30 August 2012 (UTC)
I'm using Web (web browser). That's the significance of panel. — Dmitrij D. Czarkoff (talk) 23:37, 30 August 2012 (UTC)
Oh I see, thanks. Sorry, I have never heard of it, and it seemed an odd name. (talk) 23:43, 30 August 2012 (UTC)

Alternative to X!s tools


does anybody know of a good alternative to X!s tools? its sad to see his toolserver account expiered as it was a very good tool --Hybirdd (talk · contribs · count · email) 14:02, 30 August 2012 (UTC)

I'll have to say TParis' clones of the tools:
I'll also have to say your signature doesn't comply with WP:SIG -- no templates, please. Amalthea 14:25, 30 August 2012 (UTC)
thanks for both the link and heads up , both have now been fixed :) Hybirdd (talk) 14:27, 30 August 2012 (UTC)
Thanks. :) Amalthea 14:40, 30 August 2012 (UTC)

Please Delete this page

Can somebody please delete this page:,_2012_Olympics).svg

This image file is an identical copy of the file by the same name on Wikimedia Commons. I assume that if this file is deleted, all article links will go directly to the Wikimedia Commons version. Thanks. HowardMorland (talk) 15:49, 30 August 2012 (UTC)

I've CSD'd the copy, and so the commons version is showing through now. In the future, you can tag the image under WP:CSD#F8 to be picked up by admins to clear the redundant copy. --MASEM (t) 15:55, 30 August 2012 (UTC)


Please see Wikipedia:Help desk#incategory. Thanks, Anna Frodesiak (talk) 01:13, 31 August 2012 (UTC)


Anna Frodesiak (talk) 06:13, 31 August 2012 (UTC)

Gateway time out

I just got a "Gateway (Server time-out)" error while trying to delete a file I'd moved to Commons. The file successfully deleted, but it was about a minute before the server purged and showed the Commons file instead of the local one. Is something up? - The Bushranger One ping only 03:37, 31 August 2012 (UTC)

Can somebody tell me how to add a desired new field to the tool bar above editing page

I was wondering if somebody could show me how to program a subst:Template:Rfl into my toolbar Advanced, Special Characters, Help ,Cite and then Rfl next to it. Saves me typing out reflist and header on every article. This would greatly assist me.♦ Dr. Blofeld 16:44, 28 August 2012 (UTC)

If someone can do that, I'm looking for a way to add an empty persondata template. There are many times when it would save me a some time. Ryan Vesey 16:55, 28 August 2012 (UTC)
If you are able and willing to program it yourself, read mw:Extension:WikiEditor/Toolbar customization#Add a toolbar section. Keφr (talk) 17:05, 28 August 2012 (UTC)
(edit conflict){{rfl}} can't be substituted. Ryan Vesey 17:06, 28 August 2012 (UTC)
I looked at your edit; you tried to substitute {{RFL}}, which cannot be substituted because it does not exist. (As of this writing.) Keφr (talk) 17:14, 28 August 2012 (UTC)
Dang capitalization. Why does mediawiki care? Ryan Vesey 17:19, 28 August 2012 (UTC)

Could somebody show me the exact syntax then to programme into my monobook? It needs to be subst:rfl to show both the section and reflist template in the article rather than the rfl template. ♦ Dr. Blofeld 17:23, 28 August 2012 (UTC)

This presumes you have Preferences → Editing → Enable enhanced editing toolbar enabled.

 * Extra buttons in toolbar
 * @stats [[File:Krinkle_InsertWikiEditorButton.js]]
jQuery.getScript('//', function () {

        // Reflist
                "id": "mw-customeditbutton-reflist",
                "icon": "",
                "label": "Reflist",
                "insertBefore": "==References==\n{{reflist}}",


It uses Button ref inscription.png, which was designed for the old toolbar; I will look at creating a new icon. It does not subst the template, it just adds the markup. ---— Gadget850 (Ed) talk 15:56, 29 August 2012 (UTC)

Awesome, thanks. I was also wondering if somebody could think of a DEFAULTSORT template which will reverse the names automatically, e.g Joe Bloggs it will automatically do DEFAULTSORT:Bloggs, Joe when you press a button.♦ Dr. Blofeld —Preceding undated comment added 20:29, 29 August 2012 (UTC)

And here is one for {{persondata}}:

 * Extra buttons in toolbar
 * @stats [[File:Krinkle_InsertWikiEditorButton.js]]
jQuery.getScript('//', function () {

        // Persondata
                "id": "mw-customeditbutton-persondata",
                "icon": "",
                "label": "Persondata",
                "insertBefore": "{{persondata\n| NAME              = \n| ALTERNATIVE NAMES = \n| SHORT DESCRIPTION = \n| DATE OF BIRTH     = \n| PLACE OF BIRTH    = \n| DATE OF DEATH     = \n| PLACE OF DEATH    = \n}}",

---— Gadget850 (Ed) talk 22:45, 29 August 2012 (UTC)

Hmm... Is it possible to make these things for those, who haven't enabled enhanced editing toolbar? I really don't like it :( --Edgars2007 (talk/contribs) 23:14, 29 August 2012 (UTC)
use Greasemonkey and you can do this and much more. Frietjes (talk) 21:09, 31 August 2012 (UTC)

Dab solver and watchlist

Does anyone know how to stop pages I edit using Dab solver being added to my watchlist? Thx. NtheP (talk) 13:31, 31 August 2012 (UTC)

Probably best to ask Dispenser (talk · contribs). I just untick "watch this page" every time. Jenks24 (talk) 14:29, 31 August 2012 (UTC)
Have you setup a watchlist token in your preferences and signed in Dab solver (top right)? — Dispenser 19:16, 31 August 2012 (UTC)
I have now, cheers. NtheP (talk) 19:44, 31 August 2012 (UTC)

API limits

Is there a way to find out what the API limits are without:

  • a) sending a huge request to the server and waiting to see its response or
  • b) finding out what it is by reading the API page

The second possibility isn't feasible in case the limits change (which they do based on whether the bot has a flag or not), and the first possibility seems like kind of a waste. Magog the Ogre (t c) 17:36, 31 August 2012 (UTC)

Have you tried using the API's action=paraminfo? Note that you'll have to check meta=userinfo to determine whether you have the apihighlimits right or not to figure out which limit applies. Anomie 23:49, 31 August 2012 (UTC)
What limits are you referring too? The technical limits? The row limit? AWB, uses the API and using that (forgetting for a moment that we have rules requesting users not do it) its possible to do at least 50-100 edits a minute. Doing this would probably get you blocked pretty fast but from a strictly technical perspective its possible. There are also API limits on the number of rows you can pull down at a time without admin rights or the apihighlimits right. If you look here it tells a bit about some of the API limits as does this and browsing this lovely little doc. Kumioko (talk) 00:23, 1 September 2012 (UTC)

date mathematics

I'm not sure if this is a bug, or just a word of caution, or a quirk concerning date mathematics that I ran into. I ran into this whilst creating a simple #time command which would display the next/last month and year. I started with

{{#time:F Y|+1 month}} and {{#time:F Y|-1 month}}

which works most of the time, but not all of the time. my guess is that this is because "1 month" is not a fixed amount of time, although it appears to be treated as a fixed amount of time by the #time parser function. For example, consider the following tables:

as you can see, these are mostly what I want, except for some cases. the interesting thing is that the following appears to fix it in all cases?

my understanding is that in this second case, it is using the 1st as the reference point, but that wouldn't make sense if the previous month is a short month? for a more complete analysis see User:Frietjes/date math. any ideas of why this new version works, but the old version doesn't? Frietjes (talk) 19:27, 31 August 2012 (UTC)

Hi again. The longer expressions that "work" (that are harder to remember/easier to mistype than templates that express them as "magic words") take the internal result of the date equation (which occasionally is "strange") and express it according to the Gregorian calendar convention. Put another way, this writer can tell the wiki that his next birthday is the "31st of February" next year, and it should "translate" that into March 3: {{#time:F n Y|February 31 2013}} returns March 3 2013. If every month had the same number of days, this wouldn’t be an issue. In which case, February would have an unvarying number of days, which is the real underlying issue.
---Schweiwikist (talk) 20:05, 31 August 2012 (UTC)

this doesn't really explain what is going on. the question is why {{#time:F Y|31 March 2012 +1 month}} (May 2012) and {{#time:F Y|31 March 2012 -1 month}} (March 2012) don't work, but {{#time:F Y|March 2012 +1 month}} (April 2012) and {{#time:F Y|March 2012 -1 month}} (February 2012) do work. Neither of these is using a nested date equation. obviously the parser behind #time is treating these differently, but it's not clear what exactly is going on. Frietjes (talk) 21:05, 31 August 2012 (UTC)
With a bit more experimenting, you’ll discover that if the day of the month isn’t plugged-in, the default value returned is the first day of the month, even if it isn’t displayed in the output. Reusing your typing from above, and inserting more detailed requests (shown in bold) to the parser, without always providing day-of-month data:
"...the question is why {{#time:F d Y|31 March 2012 +1 month}} (May 01 2012) and {{#time:F d Y|31 March 2012 -1 month}} (March 02 2012) don't work, but {{#time:F d Y|March 2012 +1 month}} (April 01 2012) and {{#time:F d Y|March 2012 -1 month}} (February 01 2012) do work..."
---Schweiwikist (talk) 22:10, 31 August 2012 (UTC) (this will all be history in less than two hours, as of the time of this entry)
The answer is simple: "+1 month" and "-1 month" add or subtract one month but keep the current day number, and dates with out-of-range day numbers get normalized. So {{#time:F Y|31 March 2012 +1 month}} evaluates to {{#time:F Y|31 April 2012}}, which is normalized to {{#time:F Y|1 May 2012}} and thus outputs "May 2012". {{#time:F Y|March 2012 +1 month}}, since no day number is specified, evaluates to {{#time:F Y|1 March 2012 +1 month}} → {{#time:F Y|1 April 2012}} → "April 2012". If you really want to see what's going on, use something like {{#time:c|31 March 2012 +1 month}} → 2012-05-01T00:00:00+00:00. Anomie 00:00, 1 September 2012 (UTC)


Someone has created Wikipedia:Lay which is a redirect to Price Town. I stumbled across is when I typed "wp:lay" - I was actually looking for WP:LAY (Wikipedia:Manual of Style/Layout), but for most WP pages, the all lower case takes me to the same page as the all upper case. (Although I have found other exceptions in the past - but not ones that point to article namespace.)

I'm not sure whether I should just edit Wikipedia:Lay to point to Wikipedia:Manual of Style/Layout, or whether the redirect should be deleted - but I suspect the latter. Is there some automatic mechanism that redirects all lower-case wp:something to WP:SOMETHING if Wikipedia:something and Wikipedia:Something do not exist? (I am aware that the first character is auto-capitalised.) Mitch Ames (talk) 00:25, 1 September 2012 (UTC)

It has been deleted as CSD R3; apparently someone created the article at that odd title, and then someone else moved it to a more correct title. Anomie 00:52, 1 September 2012 (UTC)

Blocking question

I looked all over Wikipedia, which even led me into MediaWiki, for an answer to this question. When a user is blocked, they usually can edit their talk page (unless specified otherwise). Can they edit their user page? Is the answer to this question written down anywhere (other than possibly by implication)?--Bbb23 (talk) 15:36, 1 September 2012 (UTC)

No. User talkpage is a specific exception. HTH, - Jarry1250 [Deliberation needed] 15:49, 1 September 2012 (UTC)
Wikipedia:Blocking policy: " A blocked user can continue to access Wikipedia, but is unable to edit any page, except (in most cases) their own user talk page." ---— Gadget850 (Ed) talk 16:46, 1 September 2012 (UTC)
Yes, that's the comment I was referring to ("by implication"). So, do I interpret that to mean that the blocked editor cannot edit their own user page ("any page except")? And how does that work technically (whatever the bit settings are for all of this)?--Bbb23 (talk) 16:57, 1 September 2012 (UTC)
Recently I read the exact answer to this - but cannot remember exactly where. Originally, a block prevented edits to all pages, period. But, the only way that a blocked user could appeal the block was to email. So, the MediaWiki software was later amended to permit editing of one specific page - the user's own talk page. --Redrose64 (talk) 18:00, 1 September 2012 (UTC)
The logic is coded in User::isBlockedFrom():
if ( !$this->mHideName && $allowUsertalk && $title->getText() === $this->getName() &&
  $title->getNamespace() == NS_USER_TALK ) {
    $blocked = false;
    wfDebug( __METHOD__ . ": self-talk page, ignoring any blocks\n" );
Richardguk (talk) 19:35, 1 September 2012 (UTC)
(edit conflict) See mw:Manual:$wgBlockAllowsUTEdit. This and many other places clearly say user talk page. I don't know an official page which directly spells out "not user page" but you can see the actual code when it was implemented at mw:Special:Code/MediaWiki/10003, specifically $title->getNamespace() == NS_USER_TALK. PrimeHunter (talk) 19:37, 1 September 2012 (UTC)
I would prefer it spelled out in the language (I'm very literal), but I accept your and Richard's confirmation that technically the software prevents a blocked user from editing their own user page. I'll consider whether to suggest (or even boldly make) a change to the policy to make it clearer. Thanks.--Bbb23 (talk) 19:52, 1 September 2012 (UTC)

Issue with the auto-generated TOC squashing contents

At least in chrome I am experiencing an issue where the contents of the auto-TOC are squashed, and I am not alone. Bizarrely it seems to only be on articles because at least on pages like the Pump here, its fine, but at things like The Expendables 2, I see this: Image here

Prometheus (film) this: Image here

Dishonored (video game) this: Image here

Dredd this: Image here

But village pump, no issue: Image here

Darkwarriorblake (talk) 15:52, 1 September 2012 (UTC)
Related: Wikipedia:Village pump (technical)/Archive 101#Categories in Google Chrome and bugzilla:39047. It appears to be a bug in the current stable version of Chrome, which has been fixed in the upcoming version. the wub "?!" 16:42, 1 September 2012 (UTC)
Thanks for the info. Fairly sick of these chrome bugs, time to move back to IE I think. Darkwarriorblake (talk) 16:52, 1 September 2012 (UTC)
You must be desparate. Other browsers are available. --Redrose64 (talk) 17:10, 1 September 2012 (UTC)
Well I am fairly fed up of it just updating in the background and breaking stuff without any input from me, so getting desperate yes. Darkwarriorblake (talk) 19:18, 1 September 2012 (UTC)
Not sure if it is related to the category bug, but I can confirm that Webkit is calculating the table's width by one pixel short. Edokter (talk) — 17:16, 1 September 2012 (UTC)

Triple square bracket

What's the best way to wikilink something that belongs in a pair of square brackets? Since the wikilink syntax is itself double bracket, this results in triple square brackets, which the wiki parser fails to parse as a link inside square brackets; for example, [[[this doesn't work]]]. I can think of:

  • [ [[this partially works, but leaves a space between the brackets]] ]: [ example ]
  • <nowiki>[< /nowiki>[[this is long and awkward to write]]<nowiki>]< /nowiki> (remove the space before the closing nowiki element): [example]
  • [&# 8203;[[using the zero-width space is awkward to write and can result in word wrapping after the opening bracket]]&# 8203;] (remove the space in the zero-width space character reference): [​example​]

but the above workarounds all are quite clunky. Anyone know a better solution?

Lowellian (reply) 22:35, 28 August 2012 (UTC)

You only need to escape the first [ and you can use the singular <nowiki /> to break the markup; see Help:Nowiki.
Markup Renders as
[<nowiki />[[example]]]
---— Gadget850 (Ed) talk 22:39, 28 August 2012 (UTC)
Or &#91;[[example]]&#93; -> [example]. Amalthea 22:54, 28 August 2012 (UTC)
We have {{Bracket}} for this. {{Bracket|[[example]]}} produces [example]. PrimeHunter (talk) 23:17, 28 August 2012 (UTC)
Thanks, that's exactly what I was looking for! :) —Lowellian (reply) 22:12, 29 August 2012 (UTC)
Amalthea's is less typing, and doesn't send a template to do an HTML entity's job. ☺ Uncle G (talk) 23:21, 29 August 2012 (UTC)
But other editors can easily think &#91; and &#93; are errors and remove them or be confused by them. PrimeHunter (talk) 10:32, 30 August 2012 (UTC)
Even the dummy <nowiki /> trick can be screwed up by a well-meaning editor who doesn't appreciate the intent, as here. A template demonstrates that your construct is intentional, as here. --Redrose64 (talk) 11:30, 30 August 2012 (UTC)
Agree— the use of {{bracket}} makes it more clear to follow-on editors. The nowiki works, but a lot of folks just are not familiar with the singular usage. ---— Gadget850 (Ed) talk 13:04, 30 August 2012 (UTC)
This is the template happiness that I was talking about above ⇑: the ideas that people can think that HTML character entities in the text for a WWW page are "errors" and that the only things that people understand are curly braces. Uncle G (talk) 13:11, 31 August 2012 (UTC)
I'm sure a lot of editors don't know what HTML character entities look like, and don't need to for what they do. We shouldn't require such technical knowledge from casual editors. There are also editors who don't know template syntax but they are less likely to break {{Bracket|[[example]]}} than remove odd looking characters like &#;. This is also a case where the exact behaviour of the parser determines which trick works. If a parser detail is changed (whether intentionally or not), then {{Bracket|[[example]]}} will still work, or at least the template can be adapted in one place and all uses will work. And it's a cheap template. PrimeHunter (talk) 14:14, 31 August 2012 (UTC)
Yup, this is definitely template happiness. One of the symptoms is mentally excluding "{" and "}" from the set of "odd-looking characters". Thinking that &#91; is "odd" whilst not blinking an eyelid at constructions such as |[[ is another symptom. You'll be asking for {{semicolon}} to be undeleted, next, mark my words. After all, you already think that a semi-colon is an "odd-looking character". ☺ Uncle G (talk) 14:48, 31 August 2012 (UTC)
I don't know these HTML codes by heart, nor do I have an intuition for wikitext escape structures. As the OP. My keyboard does not need {{semicolon}}, but likes {{ellipsis}}. I am very template happy. -DePiep (talk) 15:39, 2 September 2012 (UTC)

Is Toolserver being cranky tonight?

Getting a lot of Bad Gateway errors while trying to use CommonsHelper. - The Bushranger One ping only 06:27, 29 August 2012 (UTC)

Some trouble with one of the web servers earlier, should be fixed now. mailing list post the wub "?!" 23:23, 29 August 2012 (UTC)
It's currently unhappy as well. Can't use UTRS. Secretlondon (talk) 13:37, 2 September 2012 (UTC)

New problems in Konqueror

My main browser since 2008 has been Konqueror; I'm using version 4.8.4 now. Despite the fact that makes me part of a tiny minority, and probably no one considers it when coding their sites, I've had very few compatibility issues over the years. The only problem I've noticed at Wikipedia is that sometimes the dismiss link for those page-top announcements doesn't work. But yesterday I suddenly noticed that the section 0 edit link, which I have enabled, has stopped working. I went to check my preferences to make sure things hadn't changed, and found that the tabs on the preferences page no longer work; all the preferences are in one flat list. Neither of these things happens with Firefox. Has a new change this week broken things? Ntsimp (talk) 19:12, 1 September 2012 (UTC)

Have you tried to clear your entire cache? PrimeHunter (talk) 19:27, 1 September 2012 (UTC)
Yes. It doesn't help with either problem. Ntsimp (talk) 20:53, 1 September 2012 (UTC)
Are you getting any Javascript errors? Or did you somehow disable Javascript entirely in your browser? Anomie 01:31, 2 September 2012 (UTC)
No and no. These are the only glitches I'm seeing. Ntsimp (talk) 04:38, 2 September 2012 (UTC)
Well, I haven't changed anything, but now the section 0 edit link is back. "My preferences" is still missing the tabs, though. Ntsimp (talk) 15:18, 2 September 2012 (UTC)

why does this template go "wide"?

I don't know where this goes, so I'm coming here, to VPT. I've loaded The Lincoln Lawyer (film) on a few browsers, logged in and not, Template:Michael Connelly is twice as wide as the page (and my browser). I don't know why. Collapsing it removes the scroll bars and returns to the proper width. Can someone help? tedder (talk) 03:03, 2 September 2012 (UTC)

A null edit fixed it for me. A purge would probably also have worked. PrimeHunter (talk) 11:49, 2 September 2012 (UTC)
This might have been caused by a wider CSS problem, now solved by the hlist fix which was coincidentally made to common.css an hour ago by User:Edokter. — Richardguk (talk) 14:00, 2 September 2012 (UTC)
Not likely, the purge seemed to have fixed it before I even applied the fix. (My fix btw only applies to hlists displayed on Special: pages.) Edokter (talk) — 14:03, 2 September 2012 (UTC)
I suspect that this is another manifestation of Wikipedia:Village pump (technical)/Archive 102#Are the HTML generators out of sync on some servers?, which I noticed just yesterday at Herne Hill railway station. I've decided that each time this happens, I'll update bugzilla:38273 with the identity of the server which generated the page. --Redrose64 (talk) 14:29, 2 September 2012 (UTC)

Spacing errors in Watchlist on Chrome

I have noticed for a while that there are spacing errors in the watchlist (and history) comments, often with the comment text moving to the right over the right parenthesis. I have traced this back to the following HTML source: <span dir="auto">...some comment...</span>. Having searched the docs for the dir attribute, I can only conclude that "auto" in not a valid value for this attribute. Removing it, or changing it to "ltr" does fix the issue. I have filed Bugzilla:39912. Edokter (talk) — 17:32, 2 September 2012 (UTC)

It seems to be valid at least in HTML5, see [9] and [10]. Anomie 17:41, 2 September 2012 (UTC)
See also Wikipedia:Village pump (technical)/Archive 100#Anomalous Edit Summary for previous discussion of this issue. IMO, even if dir="auto" is not valid XHTML, Chrome is still exhibiting a bug in halfway treating it like it has white-space:pre-wrap (an entirely unrelated property) instead of just ignoring the dir. Anomie 17:48, 2 September 2012 (UTC)
We're not serving HTML5 (yet), so technically it is invalid. The doc you linked also strongly recommends against using it and only use "auto" as a last resort, as it depends on a plethora of client-side heuristics to determine the correct directionality. Given as how Chrome is affected, that should be enough reason not to use it. Edokter (talk) — 18:05, 2 September 2012 (UTC)

Brock's Yellow-eared Bat

I know that Brock's Yellow-eared Bat was discovered by Stan Brock (philanthropist) while he worked on the Dadanawa Ranch in British Guyana, but I don't know just how to show that the Binomial name Vampyressa brocki was derived from Stanley Brock. Check talk:Stan Brock (philanthropist) for a source. --DThomsen8 (talk) 19:11, 2 September 2012 (UTC)

You should write about this fact in the text of the article. Ruslik_Zero 19:28, 2 September 2012 (UTC)
I will, but I know that there is a technique for showing how a Binomial name has been derived from the naturalist who discovered and documented the particular animal or plant. I have the sources to show how it happened in this instance. --DThomsen8 (talk) 19:41, 2 September 2012 (UTC)
Have you asked a relevant WikiProject? WT:TOL for example. --Redrose64 (talk) 20:46, 2 September 2012 (UTC)

Toolserver lag

FYI... s1 and s2 both have lags of ~12 hours that started on August 1st. If toolserver scripts are dead or dying, this is probably why. See {{toolserver}} Bgwhite (talk) 06:18, 3 September 2012 (UTC)

Search box

When I type something into the search box (Vector; Firefox), I can no longer press up or down to select the article I'm looking for. I have to place the mouse and click accordingly. What is causing this? I've also noticed that since yesterday, when I want to add a page to my watchlist, I no longer get a box of information above the page text, but see a whole new screen which reads "The page "x" has been added to your watchlist, which will list edits to this page and its associated talk page. ..." Why am I now redirected to an otherwise blank page? Someone must have been fiddling with the knobs again! Jared Preston (talk) 15:54, 2 September 2012 (UTC)

This tells me that for whatever reason, javascript isn't working for you - either your browser can't handle it (rare these days), or it's been switched off. Which browser do you use? --Redrose64 (talk) 16:41, 2 September 2012 (UTC)
Firefox 3.6.28. Old, but not that old. Shouldn't really be a problem in itself. I can still scroll up/down any page displayed in the browser, it really is only the search box which doesn't react. Jared Preston (talk) 17:26, 2 September 2012 (UTC)
FF 3.6.all versions certainly does have javascript, but it can be disabled, and sometimes acts strangely if some of the internal settings get out of synch. The fix is (usually) to explicitly disable it, and then re-enable it - this persuades FF to update its internal settings so that they're in harmony. I'm currently using FF 15.0 but I think that the method is the same with 3.6.28: open the "Tools" menu, and select "Options". This produces a box and in that you should see a "Content" icon. Click that, and below the row of icons there are some selections. Of these, the third (I think) is described as "Enable JavaScript". Clear the checkbox, then click OK; then repeat the process to turn it on again. --Redrose64 (talk) 18:37, 2 September 2012 (UTC)
Thanks, that was a good idea. I cleared the cache too, but no joy. Maybe it's something else. Jared Preston (talk) 18:49, 2 September 2012 (UTC)
Update: Two hours ago – same problem. Since about five minutes ago – resolved! And I didn't even do anything. Nice, but strange! Jared Preston (talk) 17:15, 3 September 2012 (UTC)

Prune inactive users from watchlist

I've been merrily welcoming and warning users for a couple of years now, and as a result I have 1400 users on my watchlist. Can anyone think of a neat way to remove those who have not edited in the last 12 months? -- John of Reading (talk) 16:01, 2 September 2012 (UTC)

I'm in the same situation as you (although I haven't been welcoming for nearly as long). However, if the users are inactive, then their user/user talk pages aren't getting changed, so they're not cluttering up my watchlist. So it's not really a problem for me. Just my two cents. David1217 What I've done 17:11, 3 September 2012 (UTC)

"minor edits" no longer possible

Until today, it has been possible, when saving an edit to a Wikipedia article, to click on a box labeling it a "minor edit", and it would appear so marked in the article's edit history. Suddenly that's no longer there. What happened? Was this change discussed? Michael Hardy (talk) 14:58, 3 September 2012 (UTC)

I have a minor edit tick box (using firefox). Secretlondon (talk) 15:04, 3 September 2012 (UTC)
(edit conflict) Still there for me, and in Special:RecentChanges I see a number of edits marked minor. Anomie 15:05, 3 September 2012 (UTC)
Were you accidentally logged out somehow? IP users aren't given the "minor" checkbox. -- John of Reading (talk) 15:17, 3 September 2012 (UTC)
I just did this edit, and as you can see, I was logged in. The box to mark it a minor edit was not there. Michael Hardy (talk) 15:23, 3 September 2012 (UTC)
How about the "Watch this page" checkbox? And if you look in the HTML source for the page, do you find <input name="wpMinoredit" type="checkbox" value="1" tabindex="3" accesskey="i" id="wpMinoredit" />? Anomie 15:46, 3 September 2012 (UTC)
What is your browser and skin? Have you tried to clear your entire cache? PrimeHunter (talk) 20:59, 3 September 2012 (UTC)

Fuddled up page: Foxtel Cup

Hi all,
I've looked at Foxtel Cup in the latest version Firefox, Chrome and Safari, and it just don't look right. Reverted from vector skin to monobook to look at it, but no dice. The Wikipedia globe and the editing bar are still squooshed into the article itself in whatever browser, whatever skin. The last good version is here. I've looked into the coding on the current version, but can't see any obvious unclosed {{template}}s or so forth.
What is going on here? --Shirt58 (talk) 15:12, 3 September 2012 (UTC)

A purge seems to have fixed it for me. -- John of Reading (talk) 15:22, 3 September 2012 (UTC)
And for me. Shirt "Mister Stoopid Head" 58 (talk) 15:28, 3 September 2012 (UTC)

Suppressing the "Confirm navigation" warning

I have a script I've written on Commons, but when the user clicks on the link, the "Are you sure you want to leave this page?" dialogue appears. However, I do not want it to appear because the user is not actually abandoning the changes (rather s/he is submitting them for review to my script located on the Toolserver, which returns the changes back to the user): I'd like to suppress it. This seems possible, because the "save page", "show preview", and "show changes" buttons do not generate this message. How is this done?

Just a little background (in case it's helpful): the way the Javascript works is it creates a new form with new input data, then submits that form as POST data - this is the only way that I know to submit something in JavaScript as post. Magog the Ogre (t c) 22:29, 3 September 2012 (UTC)

The code for that is in the Vector extension, modules/ext.vector.editWarning.js. It doesn't look like there is any way to turn it off directly, but it looks like you should just be able to just assign window.onbeforeunload = null; before submitting your form to override it. Anomie 01:47, 4 September 2012 (UTC)

Scribunto and citation templates

Further information: MW:Extension:Scribunto and Lua

Thanks to M. Starling, Scribunto is now on the test2 server, per the announcement in The Signpost a few days ago. I've set up a citation templates test page with a lot of templates in it there, that exercises a lot of our various citation, and cross-linkage, templates. The Beatles there is a copy of our The Beatles, and shows more results of real-world citation template use with Scribunto. Currently, the test2 wiki's version takes ~10 seconds, compared to ~17 seconds for the same wikitext on the English Wikipedia. But this is not a like-for-like comparison, as things stand right now. {{Infobox musical artist}} is missing from the test2 wiki, as are things like COinS generation.

I've implemented all of our major citation templates as very simple, usually only about three lines long, invocations of a LUA module that does all of the heavy lifting. The LUA module incorporates a fair fraction of the functionality of the English Wikipedia's {{Citation/core}}. Harvard-style shortened footnotes and cross-links work, as you can see, as do most of the parameters of citation templates. The major missing items are editor name parameters, the quote parameter, and COinS tags. I also haven't tried too hard to get the exact punctuation and item orderings, yet.

Notice that the citation template invokes the module, and that's it. LUA handles picking off the template parameters. There is no need for a system of templates that transclude templates that transclude templates and closing braces coming out of one's ears.

There are a few things that Scribunto doesn't have, yet, that I've tripped over:

  • The table. It's not possible to pick up the current page name directly from LUA yet. One workaround would be for the calling template to pass in PAGENAME.
  • The mw.text.ref function. I've had to roll my own, not very good, replacement.

Uncle G (talk) 22:57, 25 August 2012 (UTC)

  • I pasted in an article I wrote as it's got lots of list defined references and page numbers using template:r. Numbering for list defined references doesn't seem to work right - at least it is doing them as 20.1 rather than using letters. Also r isn't implemented yet. I don't know if this is helpful. Secretlondon (talk) 01:15, 26 August 2012 (UTC)
  • But starting work on this is great! Secretlondon (talk) 01:17, 26 August 2012 (UTC)
    • The reference links using numbers rather than letters isn't anything to do with Scribunto to my knowledge. That's a separate thing. You didn't have to copy over the top of the existing test page. The test2 wiki isn't exactly short of article space. I've created Dnestr radar, and it works just fine, list defined references and otherwise, now that there's a Template:R and a Template:Infobox radar. See also United States.

      You might care to look at Module:Infobox radar, which is what a simple custom infobox — {{infobox radar}} in this case — looks like using the Scribunto infobox module that I've just created.

      Uncle G (talk) 03:47, 26 August 2012 (UTC)

There is also my port of Template:Coord and it's testcases. They now make use of Module:CoordinatesTheDJ (talkcontribs) 13:21, 26 August 2012 (UTC)

  • I think the most complex templates I use are the infoboxes which have co-ordinates and locationmap. An example is Template:Infobox_military_structure. I've got uses of locationmap many if that's useful: Location map~ on A-35 anti-ballistic missile system and locationmap many on Main Centre for Missile Attack Warning. Secretlondon (talk) 15:52, 26 August 2012 (UTC)
    • We need to build up a list of test articles on the test2 wiki. We have three at the moment. The downside to doing this is that there's a lot of ancillary work needed for a like-for-like comparison, including templates that won't be Scribbled. Uncle G (talk) 17:41, 26 August 2012 (UTC)
  • "Numbering for list defined references doesn't seem to work right - at least it is doing them as 20.1 rather than using letters." That is actually the inbuilt style for Cite.php backlink labels, as is the ↑ for single backlinks. The test wikis seem to ignore the Cite.php MediaWiki interface pages. ---— Gadget850 (Ed) talk 16:15, 26 August 2012 (UTC)
    • It's largely irrelevant as far as Scribbling goes. Uncle G (talk) 17:41, 26 August 2012 (UTC)
  • For the purpose of making United States?action=purge versus United States?action=purge as much of a like-for-like comparison as it can be, we need Scribbled versions of {{convert}} and {{United States topics}}, particularly the former. I'm not about to take on the job of Scribbling every major template in the project at once. The citations templates are a large enough job for one person, and I'm already Scribbling them and infoboxes. ☺ So if you feel like tackling Module:Convert, have at it. I'm just about to Scribble Template:See also, which is used several times in United States, via Module:Headnotes and then enable more ID fields in the Scribbled citation templates. Uncle G (talk) 17:41, 26 August 2012 (UTC)

That citation template looks really awesome! I will spend some time profiling it. For {{PAGENAME}} you should be able to just use frame:preprocess('{{PAGENAME}}'). There's a cache shared across all #invoke instances which gives it a cost of about 17us per call (after a fix I just pushed out), and you could cache it on the Lua side as well. It would be good if you could add COinS metadata, since it is responsible for a significant amount of the time usage in the current citation template. -- Tim Starling (talk) 01:43, 27 August 2012 (UTC)

  • I don't want to have to call frame:preprocess() at all. The temporary workaround that I mentioned before does not. I've implemented it in Template:Allmusic (q.v.). There's an approximation to COinS metadata generated now. I say an approximation because we are lacking the mw.url.encode() function to properly encode it. But it's a big string with all of the data in, at least. Uncle G (talk) 13:38, 27 August 2012 (UTC)
  • Lua cite speed comparisons: The test2 {cite_web} using Lua Module:Citation seems to run about 72 typical citations per second, or about 4.5x times faster than enwiki {cite_web} with slow 16/second. Hence, the Lua cites are about as fast as new Template:Fcite_web (5x) but not as fast as Template:Cite_quick (runs 10x-12x faster, no COinS). However, those times depend on server load, and lately, templates copied to have been running 50%-90% "slower" on than enwiki, perhaps due to higher typical server load on So, the Lua cites on enwiki might run much faster, similar to 8x faster than enwiki {cite_web}. Also, there might be some "speed tricks" which could be introduced inside Module:Citation to be even faster. The COinS tag slows the enwiki {cite_web} due to checking many major parameters a 2nd time (near bottom of Template:Citation/core); however, I have some proposed changes to make that about 20% faster there. -Wikid77 (talk) 23:08, 28 August 2012 (UTC)
    • For the Beatles test case, my laptop seems to be about 3 times faster than test2, and it's not a fast laptop. So something funny is going on there. Maybe something was compiled with the wrong optimisation settings. -- Tim Starling (talk) 23:49, 29 August 2012 (UTC)
      • Have you got to the bottom of this? Uncle G (talk) 16:57, 4 September 2012 (UTC)

Scribunto and headnote templates

Further information: MW:Extension:Scribunto and Lua

Here's some extra Scribunto goodness to anticipate: If you look at the headnote templates test page on the test2 wiki, you'll see that the Scribbled versions of {{main}}, {{further}}, {{see also}}, and so on have no trouble with more than 15 items in the list. Uncle G (talk) 10:49, 27 August 2012 (UTC)

  • This is cool, but if there is an actual use case where a {{see also}} legitimately needs to list more than 15 links, well, Heaven help us!  :-) More seriously, if such a case exists, a "see also" section would probably be more useful than a headnote. --R'n'B (call me Russ) 13:08, 27 August 2012 (UTC)
    • More generally, there are a number of templates that will benefit from Scribbling because they contain limits seemingly derived from when the template writer became tired of copying and pasting the same lump of {{#if:}} parser function code and incrementing the numbers. TheDJ has just drawn my attention to the abomination that is {{loop15}}. Its only use on the project is to emit spaces. Believe it or not, it's used in United States, via {{spaces|2}}, to emit &nbsp;&nbsp;, which takes no more characters just to type in directly. Uncle G (talk) 14:42, 27 August 2012 (UTC)
      • What is truly scary, is the amount that that thing is actually in use right now... many thousands of transclusions. —TheDJ (talkcontribs) 14:56, 27 August 2012 (UTC)
        • In the Afghanistan article, {{nbsp}} is used to emit &nbsp;. People have become template-happy. This is why we need something like Scribunto, of course. Uncle G (talk) 17:05, 27 August 2012 (UTC)
          • What puzzles me is that all these templates require at least as much typing as writing the citations manually, with lots of extra characters generated by the template itself, so I've never understood the attraction. Writing {{nbsp}} etc doesn't seem any worse than the others. SlimVirgin (talk) 20:37, 27 August 2012 (UTC)
            • You're in the wrong section. If you were in the right section ⇑ you'd see a hint as to one part of the answer, in what M. Starling is asking for. In contrast to what happens with citation templates, writing {{nbsp}} just to generate two fewer characters of wikitext is silly. Even the headnote templates that we're discussing in this section do more than one obtains by hand. (Until TheDJ updated the site CSS, headnotes looked quite strange, to the English Wikipedia reader's eye, on the test2 wiki, because they don't do undesirable things like abuse <dl/> for indentation as hand-rolled attempts do.) Uncle G (talk) 23:05, 27 August 2012 (UTC)
            • Well, we formerly had quick {sp3}, or {sp4}, which was shorter than writing "&nbsp;&nbsp;&nbsp;&nbsp;" but some people wanted the general case as {{spaces|4}} rather than tiny "sp4" and so they deleted all those short, quick little helpers (in August 2009) which were being used to concisely shift several column items in wikitables. I do not even want to mention other short-name templates, for fear they will also delete them, to be replaced with long-winded "{general utility templature|number_of_repetitions=3}". -Wikid77 03:28, 29 August 2012 (UTC)
          • I can see a valid use for something like nbsp; as a template; there might come a time when, stylistically, we may want to change how we deal with non breaking spaces (see e.g. numbers and units), and having a template to deal with that makes some sense to me. Something like spaces on the other hand... /puzzle. --Izno (talk) 22:44, 27 August 2012 (UTC)
            • {{nbsp}} is {{spaces}}. The one is a redirect to the other. I was being generous and picking an instance where the writer typed two extra characters just to get &nbsp; back, instead of four. ☺ But the abomination here is {{loop15}}, remember. Uncle G (talk) 23:14, 27 August 2012 (UTC)
              • {{loop15}} doesn't just emit spaces: ┌{{loop15|8|─}}┘ → ┌────────┘. --Redrose64 (talk) 18:11, 28 August 2012 (UTC)
                • You mean need not rather than does not. See whether you can find a single article-space use of it to do anything other than emit spaces as a result of it in turn being emitted by {{spaces}}. For bonus points, see whether you can find article-space uses of it to emit a string of wikitext longer than "{{spaces|2}}". ☺ Uncle G (talk) 22:10, 28 August 2012 (UTC)
                • To me the issue with hatnotes is not the code repetition but their forkings for nearly similar meanings, and then variant mechanisms per fork (esp adding (disambiguation)) (doc). -DePiep (talk) 18:31, 28 August 2012 (UTC)
                  • It's the code repetition rather than the forking that extends page load time, though. Forking is just a per-wiki editorial decision, in respect of which Scribunto itself is neutral. Have a look at the bottom of Module:Headnote, by the way. I decided to do the fan-in in the module, there. One can as an alternative do it in the various individual templates. And indeed I've done exactly that with Template:Reflist and Template:Notelist, which both call the same Lua function with different arguments. The merits of the two approaches might be something that M. Starling can also profile, if we ask. Uncle G (talk) 22:10, 28 August 2012 (UTC)
                    • Staying with hatnotes, I cannot recall it's nesting or any other cost is limitative. (Of course I agree the repetitive or even nesting coding is undesired, I get Lua; and I would like to try it myself), but the issue here is that all these hatnote templates have their own prescription. It's prescription, it's setup, it's documentation. I know this is out of Lua reach, but let's not spend too much time & applause in speeding ~50 hatnote templates into ~50 somewhat faster ones. The test2 link you gave already does so module-internally. Can Lua help the usage principles? -DePiep (talk) 01:01, 29 August 2012 (UTC)
                      • As I said, Scribunto is neutral with respect to the editorial decision to have lots of little templates that are almost, but not quite, the same as one another. It doesn't force one either to do or to not do that. But it can ease the maintenance of such templates, by factoring out the common stuff that's actually the same across all of the itty-bitty-slightly-different templates.

                        Consider Module:Navigation. There's just the one place where the English Wikibooks logo filename is defined: in the table at the top. So only one thing needs to be changed in order to change that picture in {{subject bar}}, {{sister project links}}, and {{wikibooks}}. (There are probably a few more templates that I could Scribble there.) Under our current system, the logo filename is individually set in three places — {{subject bar}}, {{sister project links}}, and {{sister}} — and has to be separately updated in all three. Notice that they're currently out of synch.

                        Uncle G (talk) 17:31, 4 September 2012 (UTC)

Script error in Internet Explorer

The last couple days I get an error whenever I try and use internet explorer to view or edit Wikipedia. The error is below.

I haven't modified anything and it seems to cooincide with the new popup that was implemented so I think it might have something to do with that. Its IE7. Kumioko (talk) 20:18, 31 August 2012 (UTC) ──────────────────────────────────────────────────────────────────────────────────────────────────── Now I am getting it occassionally in Firefox as well. The error is more descriptive though and here is what it says:

With the following additional text:


Kumioko (talk) 00:03, 1 September 2012 (UTC)

Without looking into console I can only guess that a js script failed to load, for instance jQuery. Ruslik_Zero 18:15, 1 September 2012 (UTC)
  • I have the exact same problem with IE and guess we're not alone.TMCk (talk) 18:52, 1 September 2012 (UTC)

Possibly related: bugzilla:39959#c1. Can anyone getting the unresponsive script error please see if they have that user preference enabled? It'll help narrow down your issue to know whether you have that particular user preference enabled. --MZMcBride (talk) 03:27, 4 September 2012 (UTC)

I can confirm that for Firefox. -- DQ (ʞlɐʇ) 03:50, 4 September 2012 (UTC)
Not confirmed for IE9, have said user preference enabled and right-clicking works fine. the wub "?!" 11:22, 4 September 2012 (UTC)
Err, right, but have you been getting the unresponsive script warnings? --MZMcBride (talk) 14:42, 4 September 2012 (UTC)
I'm using IE-8 and did get script errors and extreme long loading times for several days w/o this future enabled. But for whatever reason, today everything seems to be back to normal (I didn't make any changes to my system, not even a restart). Where there any changes made since yesterday, WP wise, that could've been responsible for this?TMCk (talk) 18:08, 4 September 2012 (UTC)

Signature customization

Hi all my username denotes a street junction. I have experimented some but would much appreciate 2 technical tips. 1st the appearance of a wooden post between "Market St." and "Diamond St." and 2nd if there is some code that tapers the <small> to 2 or 3 degrees to give the "sign" depth and 3d perception. This is my code as it stands now:

Also any better font suggestions for an old 18th/19th century street sign? Thanks!Market St.IDiamond St. 12:42, 1 September 2012 (UTC)

You might want to have ⧏Market St. Diamond St.⧐ instead. But ways better solution would be to blank the box and uncheck "treat as wiki markup" option. — Dmitrij D. Czarkoff (talk) 14:46, 1 September 2012 (UTC)
Thanks!Market St.IDiamond St. 17:58, 1 September 2012 (UTC)
BTW WP:SIGN requires that you link your user talk page, so you may prefer something like this — Market St. Diamond St. — but it is yet more wordy... — Dmitrij D. Czarkoff (talk) 21:53, 1 September 2012 (UTC)
No; the user talk page link is optional - only one link is required, but that may be any one of three. The shortcut to the specific section is WP:SIGLINK where we find Signatures must include at least one internal link to your user page, user talk page, or contributions page; this allows other editors easy access to your talk page and contributions log. The lack of such a link is widely viewed as obstructive. --Redrose64 (talk) 23:48, 1 September 2012 (UTC)
Thanks so much for all the great responses! Marketdiamond (talk) 09:21, 4 September 2012 (UTC)

Context menu very slow to open

The right-click context menu seems to be very slow to open on Wikipedia pages. Pressing the Menu key isn't any better. This behavior seems to be tied to Twinkle or some other JavaScript functionality because it only occurs when I'm logged in, and only on Wikipedia pages. I'm running Firefox 15 on Windows 7. When I right-click on a link, I may get the following error:

Any idea what's happening? — DragonLord (talk/contribs) 18:16, 3 September 2012 (UTC)

I'm using FF 15.0 under XP, but don't use Twinkle: right-click menus behave normally. --Redrose64 (talk) 18:29, 3 September 2012 (UTC)
Possibly the same as #Script_error_in_Internet_Explorer. Any information that can be added to pinpoint the issue further would be welcome. If you find a page where this occurs for your setup, consider loading the page with the ?debug=true or &debug=true and see if that gives a more exact script error. —TheDJ (talkcontribs) 18:45, 3 September 2012 (UTC)
As you can see below, I have the same problem. However this last answer is too technical for me. By the way, I'm also using FF15. Lova Falk talk 19:29, 3 September 2012 (UTC)
Which version of Windows are you using, and do you have Twinkle installed? --Redrose64 (talk) 20:09, 3 September 2012 (UTC)
  • This is the solution. It's a user preference option. -- DQ (ʞlɐʇ) 03:55, 4 September 2012 (UTC)
    • Confirmed to be due to the "right click a header to edit" javascript code. When we upgraded to jQuery 1.8.0, the IE issue reported above was the problem. Yesterday we upgraded to 1.8.1, which is probably causing the FF 15 problem. —TheDJ (talkcontribs) 11:52, 4 September 2012 (UTC)

"A script on this page may be busy, or it may have stopped responding."

Yesterday, all was fine. Today, out of the blue, whenever I want to open a link in a new tab, I have to wait for ages and then get the message: "A script on this page may be busy, or it may have stopped responding. Script:" When I stop the script, after a while I can open the link. Waiting so long drives me crazy. What can I do? Lova Falk talk 19:23, 3 September 2012 (UTC)

PS I'm sorry, I completely missed that the same question was asked less than an hour ago...Lova Falk talk 19:27, 3 September 2012 (UTC)


As an aside, are we really using so much complex javascript functionality that we have to ship the jquery library with every page load? I had no idea Wikipedia was a jquery site. Has anyone looked at using simple javascript instead, and how much bandwidth that might save? (Not to mention the time involved debugging these kinds of bugs.) —Cupco 19:58, 3 September 2012 (UTC) Looking a little further, do we use jquery for anything other than Reference Tooltips? Maybe in Twinkle? Could we use a lighter-weight library such as overlib? —Cupco 20:21, 3 September 2012 (UTC)

Almost all core JS of mediawiki has use jquery for over 2,5 years now. —TheDJ (talkcontribs) 20:42, 3 September 2012 (UTC)
(ec)I don't know much about those things but why do we need all that in the English WP while the other language wikis seem to do just fine w/o, and load rather quickly. By now, English WP sites load ridiculously slow due to scripting stuff. And no, not everybody has the most up-to-date computer that can handle that. There needs to be a change in the scripting or our slogan: WP, the encyclopedia everyone with a fast computer can edit?TMCk (talk) 20:45, 3 September 2012 (UTC)
"the other language wikis seem to do just fine w/o", the other wiki's all have exactly the same JS core functionality. —TheDJ (talkcontribs) 21:13, 3 September 2012 (UTC)
Well, there must be some major difference which is quite notable.TMCk (talk) 21:20, 3 September 2012 (UTC)
Most likely our heavy use of complicated templates and meta-templates. jQuery itself doesn't change that often and should be cached by your browser, so bandwidth isn't a big issue. And it makes writing JavaScript which works across browsers much easier. Ditching jQuery would likely result in more bugs, not less. the wub "?!" 11:17, 4 September 2012 (UTC)
  • Actually in this case it looks like it's a bug in the new jQuery version (1.8.1). Reedy has pushed a change to revert us back to the previous version, and we be testing it now. Okeyes (WMF) (talk) 14:22, 4 September 2012 (UTC)
    Okay; should now be fixed (as evidenced by me using right-click-to-section-edit to leave this message). If you're still having trouble, try a hard refresh. Okeyes (WMF) (talk) 14:27, 4 September 2012 (UTC)
    As I just mentioned at the threat further up ("Script error in Internet Explorer"), as of today everything seems to work normal again. Note: I do not use the right-click feature.TMCk (talk) 18:15, 4 September 2012 (UTC)

Could someone help me with table formatting ?

With regard to the list of target antigens in pemphigus, what is the best way to list each of the antigens such that the molecular weight is displayed right next to the corresponding antigen? Restated, how do I put line breaks between each antigen (when there is more than one per disease) and each molecular weight? Thanks in advance for your help! ---My Core Competency is Competency (talk) 21:24, 3 September 2012 (UTC)

You can use <br />. PrimeHunter (talk) 21:31, 3 September 2012 (UTC)
Or {{ubl}}. ---— Gadget850 (Ed) talk 22:35, 3 September 2012 (UTC)
Gadget's is the preferred, but either way works. --Izno (talk) 00:51, 5 September 2012 (UTC)

Scribunto comparisons

  • India — 157KiB featured article on the English Wikipedia, served by mw52 in 26.662 seconds at 2012-09-04 16:59
  • India — copied on the Test2 Wikipedia, served by mw21 in 17.526 seconds at 2012-09-04 16:59

I, WOSlinker, and Phil Boswell‎ between us have Scribbled all of the major templates included by the article on India, on the Test2 Wikipedia. This includes all of the citation templates, the conversion template, the various navigation templates such as Template:India topics, unexpected little templates such as Template:Indian population clock, and the surprisingly complex Template:Infobox country. Some minor templates have not been Scribbled, and so perform as they do here, and one template that generates an editorial category is missing. Just to bias the test in favour of the non-Scribbled, I've even added things like {{cite encyclopaedia|encyclopaedia=}} parameter spellings, to make the Scribbled templates do more work. You'll see from Citation templates test#Phil Boswell that we've also incorporated Phil Boswell's {{citation|author-format=vanc}} idea and {{citation|via=Highbeam Research|subscription=yes}} as requested at Template talk:Citation/core#Adding a parameter. And all this on a setup that runs more slowly than Tim Starling's laptop, according to Tim Starling at any rate.

Nonetheless, the results are as above. (The above, copied and pasted out of the HTML, is misleadingly precise. You'll find in practice that there's a range of ±3 seconds or thereabouts on Test2, and a slightly wider range here. I've seen Test2 as low as 15s and as high as 20s, and here sometimes longer than 30s.)

Before you say it: Yes, the articles look different. The difference in citation cross-linkage styles (Test2 using superscripted numbers not letters) and in collapsible lists are nothing to do with Scribunto. They just highlight how far different the English Wikipedia is from a vanilla installation of MediaWiki. We have "dynamic navigation bars" in MediaWiki:Common.js; different non-default settings for things like MediaWiki:Cite references link many and MediaWiki:Cite references link many format; and a whole infrastructure of custom CSS and JavaScript to get things looking as they do on the English Wikipedia. The figures above are the HTML rendering times reported by the servers, the primary target of Scribbling to make things work faster, and are unaffected by what happens with CSS and JavaScript in WWW browsers.

Uncle G (talk) 16:57, 4 September 2012 (UTC)

Some database reports aren't being updated

Can we have this sorted out somehow? Wikipedia talk:Database reports#Many of the reports aren't being updated Cheers. -- Alan Liefting (talk - contribs) 09:14, 5 September 2012 (UTC)

Creating dynamic lists

Hello, I was directed here from New Users. I have a question:

Is there a way to search lists/create dynamic lists of articles based on infobox template data?

What I envision, if it's not clear, is using the fields in an infobox the same way that a library catalog would use MARC fields - we know what the fields do, so within uses of a given infobox it should be possible to search that way (though I admit it's not as simple as the reserved fields in a MARC catalog record) to produce a list of exact values, values within a given range, and/not/or matches between fields etc.

It might even be possible to compare similar boxes this way. It would certainly help when narrowing down useful lists of articles (as a reader), because although "List of..." pages are useful, they quickly become unwieldy, as do category pages. You cannot, to my knowledge, currently eliminate articles from such lists based on, eg, whether biographical subjects are living or dead persons, their DoB, occupation, etc. other than by checking each article (if you are not already intimately familiar with the category to which it belongs).

Can I do this somehow that I've obviously failed to understand? Can I not do it because it would be a massive strain on Wikipedia's servers, or for some other technical reason? I apologise if this has been discussed in the past and I've missed it - if so, could you direct me to the archived discussion? (talk) 21:16, 5 September 2012 (UTC)

AFAIK, DBpedia is one of the better routes for doing what I understand you want to do. Sadly, tonight it seems their website is kaput. --Tagishsimon (talk) 22:59, 5 September 2012 (UTC)
That's frustrating, but thanks anyway! (talk) 00:00, 6 September 2012 (UTC)
  • Use wikisearch to find infobox parameters: Searching of infoboxes can be done, to some extent, by using wikisearch where parameter names are unique. This idea is little-known to many Wikipedians, but the wikisearch (Search:[_____]) can list matching articles based on unique parameter names in infoboxes. The tactic goes back to the early days of developing search-engine lookup, where I noticed that using unique names, or parameter names, will help search for semi-structured data (classic term "associative retrieval"). For example, a search can be made for towns whose elevation and population are set in infoboxes which use German parameter names, "Höhe" (height) and "Einwohner" (population). Run a wikisearch to find all towns at elevation 103 (metres) which have "Einwohner=" set, so hunt for:  Höhe 103 Einwohner:öhe+103+Einwohner
That wikisearch query will match about 97 articles, such as town "Gorden-Staupitz" which has parameters:
 | Bundesland Brandenburg | Landkreis Elbe-Elster | Amt Plessa | Höhe 103 | Fläche 28.38 | Einwohner 1128
Hence, the population of that town was set as "1128" inhabitants, among other towns at the same elevation, 103 metres (338 ft). When an infobox is updated, then it takes about 24 hours, overnight, before the wikisearch will match the new data. However, for such searches to work in other infoboxes, then the parameter names need to be more unique, such as a sportscar infobox having parameter "spcarname" which could then be searched. When all parameters are only common words, such as "name" or "area" or "population" then thousands of other articles match the search, so there need to be some unique parameter names designed for exclusive use in each type of infobox. However, years ago there was an "industry standard" to foolishly name data items using long common words, rather than short, mnemonic new words, for abbreviated names, as done for decades by wise computer technicians. Of course, these name-lookup tactics have been known for decades, but today's computers are still too primitive to give better results. The world even renamed the concept as "Googling" because the general public was unaware that, for decades, "multi-word searches" could be run under the control of "search brackets" which limit the scope, or width or depth, of the information to be matched. Today's browsers have a "Find" button for old simple text-strings, but no "Hunt" button to hunt for set of 3 words contained within the same paragraph, even though multi-word search has been known for decades. It is such a simple concept, but the world just "doesn't get it get it": multi-word search, it's not just for search-engines anymore. If really extended, then infoboxes would have multiple levels of detail, so that a wikisearch could also be controlled by levels of detail to search, as well as width of text paragraphs to search. Anyway, for now, just create infoboxes with more unique parameters names, to be matched during a wikisearch. -Wikid77 (talk) 04:59, 6 September 2012 (UTC)

Searching for an answer …

I have been trying to use a template with a parameter, but it ignored the value I supplied.
I realised why, and eventually found out how to make it work. Later I realised what had gone wrong – the parameter had been removed by an edit 4 years ago.
Would it be possible to do complex searches, comparable to "What Links Here"?

eg. articles that use an image but not the template version of that image based on Template:Annotated image
or articles that use a Template and supply a particular parameter

This would help me consolidate the use of these templates and make them more consistent. » Ζετα ζ (talk) 01:48, 6 September 2012 (UTC)

{{USRD}} and needs-kml

I tried adding the above flag that checks for the presence of the KML page. It displays fine, but when a KML page is added, it does not refresh the category; a null edit is required to fix it. Does anyone have any ideas as to why that is, or if this can be fixed? --Rschen7754 05:57, 6 September 2012 (UTC)

Wikipedia:Village pump (technical)#Status help

Please answer my last response in the said title. It isn't resolved yet and the problem still ensues. TruPepitoMTalk to us 06:00, 6 September 2012 (UTC)

Status help

How can this be fixed if I want it to display enable during 7 a.m. until 10 p.m. and disable during 11 p.m. until 6 a.m.? Also, what is this mainly for? Object: Status if I'm around or not.

Here's the script:

#expr:{{{{{|safesubst:}}}CURRENTHOUR}}2}} < 23|disable|enable}}

TruPepitoM (talk to us) 06:28, 1 September 2012 (UTC)

Renders as: '. Subst-protect as needed. — Dmitrij D. Czarkoff (talk) 10:26, 1 September 2012 (UTC)]]
There's a problem with the equation:
Turns out as:
. Looks like it needs two more of }}. Same object, which is
</{{DISPLAYTITLE:<font color={{#switch:{{User:TruPepitoM/Status}} |enable=green |disable =red}} face=Arnprior>{{NAMESPACE}}:</font><big><font color=green face=Neuropol>TruPepitoM</font></big>}}<center><big><big><big>USER STATUS: {{#switch:{{/Status}} |enable=[[File:Green check.svg|50px]] Active and Ready |disable=[[File:Red x.svg|50px]] Inactive}}</big></big></big>
. Please help and fix the scripts.TruPepitoM (talk to us) 14:08, 1 September 2012 (UTC)
No, the script is OK. You just shouldn't split it:
Markup Renders as

#expr:{{{{{|safesubst:}}}CURRENTHOUR}}2}} < 23|disable|enable}}


  1. ifexpr:{{safesubst:
  2. expr:042}} < 23|disable|enable}}

{{{{{|safesubst:}}}#ifexpr:{{{{{|safesubst:}}}#expr:{{{{{|safesubst:}}}CURRENTHOUR}}2}} < 23|disable|enable}}


— Dmitrij D. Czarkoff (talk) 14:35, 1 September 2012 (UTC)
No, what I mean in my response was to where should I put the two }} in
Here's what it happened earlier:
Markup Renders as



The true rendering wouldn't show up because of the lack of the two braces on the right in the script and the result. Please fix the new one above and not the original script which is
{{{{{|safesubst:}}}#ifexpr:{{{{{|safesubst:}}}#expr:{{{{{|safesubst:}}}CURRENTHOUR}}2}} < 23|disable|enable}}
TruPepitoMTalk to us 05:52, 6 September 2012 (UTC)
I think you are looking for this:
HOWEVER, be aware that CURRENTHOUR only takes into account the hour during the last time the page was PARSED and cached. It is not some sort of clock. As such, if the last time the page was parsed was at 10AM and it's now 3AM, then someone viewing the page will see 'enable'. See also the documentation. —TheDJ (talkcontribs) 16:26, 6 September 2012 (UTC)

Need to recursive subst: Cite pmid templates all the way to simple wikitext

We are preparing Birth control for translation into several languages. Presently it makes heavy use of {{Cite pmid}} templates, which are fantastic for content creation but do not exist in most foreign-language Wikipedias and seem difficult to migrate along with an article. Here is an example of what happens when trying to subst a {Cite pmid} template:

{{subst:Cite pmid|22341164|noedit}} becomes:
{{#if:22341164|{{#ifexist:Template:Cite pmid/{{#expr:22341164}}|{{cite pmid/{{#expr:22341164}}}}{{#ifeq:noedit|noedit||<span class="plainlinks noprint" style="font-size:smaller"> [{{localurl:Template:cite_pmid/{{#expr:22341164}}|action=edit&editintro=Template:Cite_pmid/editintro2}} edit]</span>}}|PMID 22341164 ({{PMID|22341164}})<br><span class="plainlinks nourlexpansion">Citation will be completed automatically in a few minutes. <small> [{{#expr:22341164}} Jump the queue] or [{{#expr:22341164}} expand by hand]</small></span>[[Category:Pages with incomplete PMID references|{{PAGENAME}}]]}}|<span class="error">Error: No PMID specified!</span>}}

But what I want is:

Chin, H. B.; Sipe, T. A.; Elder, R.; Mercer, S. L.; Chattopadhyay, S. K.; Jacob, V.; Wethington, H. R.; Kirby, D. et al. (2012). [ "The Effectiveness of Group-Based Comprehensive Risk-Reduction and Abstinence Education Interventions to Prevent or Reduce the Risk of Adolescent Pregnancy, Human Immunodeficiency Virus, and Sexually Transmitted Infections"]. ''American Journal of Preventive Medicine'' '''42''' (3): 272–294. [[Digital object identifier|doi]]:[ 10.1016/j.amepre.2011.11.006]. [[PubMed Identifier|PMID]] [ 22341164].

Is there a bot or an AWB option or a gadget or an easy procedure which can do recursive template substitution like that? I know I can copy and paste the {Cite journal...} from Template:cite pmid/22341164 Template:Cite doi/10.1016.2Fj.amepre.2011.11.006, but I hope I don't have to to that by hand for each ref. Automating that would be okay too, I think, but not as good. —Cupco 21:27, 2 September 2012 (UTC)

You can put {{Cite pmid|22341164|noedit}} in Special:ExpandTemplates and copy/paste the resulting text. Σσς. 21:47, 2 September 2012 (UTC)
Thanks for telling me about that. If it worked a little better I would try to make a curl script for it, but it makes this: <span class="citation Journal" >Chin,&#32;H. B.&#059;&#32;Sipe,&#32;T. A.&#059;&#32;Elder,&#32;R.&#059;&#32;Mercer,&#32;S. L.&#059;&#32;Chattopadhyay,&#32;S. K.&#059;&#32;Jacob,&#32;V.&#059;&#32;Wethington,&#32;H. R.&#059;&#32;Kirby,&#32;D.&#32;et al.&#32;(2012).&#32;[ "The Effectiveness of Group-Based Comprehensive Risk-Reduction and Abstinence Education Interventions to Prevent or Reduce the Risk of Adolescent Pregnancy, Human Immunodeficiency Virus, and Sexually Transmitted Infections"].&#32;''American Journal of Preventive Medicine''&#32;'''42'''&#32;(3): 272–294.&#32;[[Digital object identifier|doi]]:[ 10.1016/j.amepre.2011.11.006].&#32;[[PubMed Identifier|PMID]]&nbsp;[// 22341164]<span class="printonly">.&#32;</span>.</span><span class="Z3988" title="ctx_ver=Z39.88-2004&rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Ajournal&rft.genre=article&rft.atitle=The+Effectiveness+of+Group-Based+Comprehensive+Risk-Reduction+and+Abstinence+Education+Interventions+to+Prevent+or+Reduce+the+Risk+of+Adolescent+Pregnancy%2C+Human+Immunodeficiency+Virus%2C+and+Sexually+Transmitted+Infections&rft.jtitle=American+Journal+of+Preventive+Medicine&rft.aulast=Chin&rft.aufirst=H.+B.&"><span style="display: none;">&nbsp;</span></span>
(Scroll sideways for extra pain.) I guess I could try to expand those &#xx elements and strip out only the first span. I'll wait a little to see if someone else has done this yet. —Cupco 21:59, 2 September 2012 (UTC)
The &#xx; entities should be fine to leave as is; for example, &#32; is a space: Exam&#32;ple → Exam ple. Everything from <span class="Z3988"> onwards is the WP:COinS metadata and may safely be omitted. --Redrose64 (talk) 15:08, 3 September 2012 (UTC)

You want to work with the subtemplate {{Cite pmid/22341164}} (note the slash). You will find that this redirects to {{Cite doi/10.1016.2Fj.amepre.2011.11.006}}. Regardless, either will subst as desired. ---— Gadget850 (Ed) talk 22:28, 2 September 2012 (UTC)---— Gadget850 (Ed) talk 22:28, 2 September 2012 (UTC)

  • Edit the bar "|" to slash before using subst: Each PMID entry is a subtemplate of {Cite_pmid}, so change every "pmid|" to be "pmid/" and then the wp:Subst'ing will work:
The above cite is literally the actual result of "{subst:Cite pmid/22341164|...}" which generates a call to {Cite journal|...}. The {Cite_pmid} does seem to be overkill, where many entries are used only once, and there are over 6,000 which must be viewed separately. Thanks for noting that it also hinders interwiki copy/translation, like any non-global template which must be subst'ed before interwiki transfer. I guess I need to write a fully subst'ing template for {Cite_pmid}. -Wikid77 (talk) 06:35, 6 September 2012 (UTC)
Thank you!Cupco 02:03, 7 September 2012 (UTC)

Redundancy regarding templates excluded in print

I recently noticed that there is a lot of redundancy between Category:Exclude in print and MediaWiki:PDF Template Blacklist. Both of these appear to be used to ensure that certain templates are not included in things like Books but I'm not sure if that is all they are used for. I am also not 100% what the difference is in these 2 items but it seems that MediaWiki:PDF Template Blacklist existed and people started piling things into it and then Category:Exclude in print was created and people started piling stuff in it too. I would suggest that we remove any items from MediaWiki:PDF Template Blacklist if it already exists in Category:Exclude in print but I'm not sure what 2nd and 3rd level effects that might have. Any thoughts on why we need both of these? Kumioko (talk) 20:09, 5 September 2012 (UTC)

Just a followup
  1. there are 4854 articles on MediaWiki:PDF Template Blacklist and 22642 in Category:Exclude in print. Of those 173 are equal.
  2. There are 458 Red links on MediaWiki:PDF Template Blacklist that should probably be removed.
  3. Of the remaining templates in MediaWiki:PDF Template Blacklist it appears that the majority would be better just to add to Category:Exclude in print. So I'm not sure what purpose the MediaWiki:PDF Template Blacklist page is at that point.
  4. Many of these 4500+ templates MediaWiki:PDF Template Blacklist are navigation templates that are discouraged in PDF exports so adding Category:Exclude in print to the core Navigation template that is used to create and drive the thousands of Navigation templates would probably be an easy way to fix that. Kumioko (talk) 17:38, 6 September 2012 (UTC)

PDF renderer bug with the chem tag

I think i have found a bug in the wikipedia PDF renderer. Go to this page: Bicarbonate. In the chemical box on the right, there is the "Properties" section with a "Molecular formula" record. This record has the value "CHO3-". Now download the PDF of the page. You will see that in the PDF the same record has the incorrect value of "CHO". The same applies to Bicarbonate, where "CO32-" becomes "CO" in the PDF. This does NOT apply to non-ions like Sulfuric acid, so I think that the problem are the superscripts, while subscripts cause no problems. Particularly, it should be the "Chem|text|subscript|superscript" tag (with double curly brackets after and before). If there is no superscript, the PDF is rendered correctly. If there is the superscript, the renderer seems to ignore both the sub and the super scripts. Raffamaiden (talk) 23:53, 5 September 2012 (UTC)

This template uses {{su}}, which according to it's own documentation uses "all kinds of hacky tricks to make the output render correctly". This is why the use of hacks should be limited. The semantic meaning of sup and sub has been lost in favor of "nice output", forgoing the fact that it's not guaranteed that wikicode is output as full fledged HTML... —TheDJ (talkcontribs) 12:09, 6 September 2012 (UTC)
Adapting {{Su/Print}} should work. Getting rid of the hack of su is a better long term goal. —TheDJ (talkcontribs) 12:11, 6 September 2012 (UTC)
Yes— IE in prticular has issues with line-height. This used to cause chunks of a page to not print until we figured it out. {{Sup sub}} will probably work. ---— Gadget850 (Ed) talk 12:51, 6 September 2012 (UTC)
It looks like that the pdf writer ignores "inline-block" value of the display property. Ruslik_Zero 17:08, 6 September 2012 (UTC)

Color box template

Hi there. I would like to modify {{color box}} so that can produce a box that is half one color and half another. I plan to use this template to help indicate rank for judoka in the article infobox, and sixth to eight degree 'black belts' actually wear a red and white belt. I have a vague idea of how the template works, but I'm not sure how to go about making this change. If you could help me out, I would appreciate it. CanadianJudoka (talk) 15:33, 6 September 2012 (UTC)

The primary purpose of the template is to show coloured text on a differently-coloured background, so it takes more than one parameter; hence adding more might be a problem. It is often used as a colour dab though, and I have sometimes seen two or three butted up together - you might find this technique easier to use:         . The dividing line up the middle may be changed from black to darkgray by using {{colorbox}}:         . --Redrose64 (talk) 16:27, 6 September 2012 (UTC)
Thanks for your response. I would prefer to have just one box split down the middle, though. The template itself doesn't even necessarily have to be changed; I would be happy with just knowing what commands to use in order to produce the desired effect. CanadianJudoka (talk) 16:49, 6 September 2012 (UTC)
By copying and pasting until it worked, here is a pair of boxes with no border between and a border around:
<span style="background-color:#00FF00; color:{{{3|}}}; border:1px solid #000000; text-align:center;">{{#if:{{{2|}}}| {{{2}}} |    }}<span style="background-color:#FF0000; color:{{{3|}}}; border:0px solid #000000; text-align:center;">{{#if:{{{2|}}}| {{{2}}} |    }}</span></span>
produces         . However, it can definitely use some cleanup to use as a template. Chris857 (talk) 17:21, 6 September 2012 (UTC)
Great, thanks for that. With a little tinkering, this is what I had in mind:     
What cleanup would you suggest to turn this into a template? CanadianJudoka (talk) 17:30, 6 September 2012 (UTC)
Here you go. If you made this--> <span style="background-color:{{{1|}}}; border:1px solid #000000;">&nbsp;&nbsp;<span style="background-color:{{{2|}}};">&nbsp;&nbsp;</span></span> into a template (say we called it JudokaRank), you could then call the template like {{JudokaRank|red|white}} and it should produce     . Chris857 (talk) 17:42, 6 September 2012 (UTC)
[Changed '&'s to '&amp;' to show Chris857's means of preventing linebreaks in the middle of the box. The nowiki tag doesn't quote html elements. —Cupco 20:19, 6 September 2012 (UTC)

Thanks for your help, everyone. The new template is here: {{JudoRank}} CanadianJudoka (talk) 03:57, 7 September 2012 (UTC)

Actually, slight problem. Text added beside the box automatically starts on the next line. Is there a way to fix it so that the text will stay beside the box? CanadianJudoka (talk) 04:06, 7 September 2012 (UTC)

Sorted it out on my own. CanadianJudoka (talk) 05:09, 7 September 2012 (UTC)

Page views

Does anyone know why the page views tool has not been working on the last couple of days? Simply south...... eating shoes for just 6 years 21:13, 6 September 2012 (UTC)

Which page views tool exactly? appears to be working fine. the wub "?!" 00:43, 7 September 2012 (UTC)

Sudden problem at WP:AFD


Something weird has happened at the main page of Wikipedia:Articles for deletion. The entire log for 30 August (Thursday) has disappeared from the page. I think this just happened within the past hour. Can someone find that log and re-establish the link? Thanks. --MelanieN (talk) 00:32, 7 September 2012 (UTC)

Fixed, I just poked MathBot to run at Wikipedia:Articles for deletion/Old. the wub "?!" 00:41, 7 September 2012 (UTC)
Thank you! --MelanieN (talk) 01:06, 7 September 2012 (UTC)

Using a script without installing it

Does this method (javascript:importScript('User:Shubinator/DYKcheck.js'); dykCheck();) still work? I'm getting an error: "uncaught exception: ReferenceError: importScript is not defined". If it's not longer valid, is there an alternative? Thanks, Goodvac (talk) 22:25, 6 September 2012 (UTC)

I'm don't know what that last bit is. See User:Shubinator/DYKcheck#Using DYKcheck. ---— Gadget850 (Ed) talk 22:36, 6 September 2012 (UTC)
No, I know you can install it, but I'm wondering if there's a way to use a script without installing it. The method above worked in the past. Goodvac (talk) 22:42, 6 September 2012 (UTC)
Ah. If you use Firefox, you can then use Greasemonkey. ---— Gadget850 (Ed) talk 23:02, 6 September 2012 (UTC)
Indeed. Firefox 7 killed location bar Javascript. A simple solution is to create a bookmark with the js in as the location, and click it when you want it executed. --Splarka (rant) 08:40, 7 September 2012 (UTC)
Thanks, that worked! Goodvac (talk) 15:59, 7 September 2012 (UTC)

Forcing the edit summary window open when adding a new section

If &section=new is placed at the end of a URL, is is impossible to enter an edit summary. Is it possible to modify the url to still add a new section but also have an edit summary? This applies to the rare occasion when the new section link is being used without a new section header. Ryan Vesey 23:08, 6 September 2012 (UTC)

An actual example of the desired use, would make the question easier to answer (or workarounds to be suggested).
Is this for use on usertalkpages, or noticeboards? For yourself or for everyone? etc. More details upfront = less clarifications needed. :) -- Quiddity (talk) 23:37, 6 September 2012 (UTC)
Sorry about that. See the hack used to install scripts on Wikipedia:Teahouse/Host lounge#Tools for hosts. The button that says Install Teahouse Scripts. Ryan Vesey 01:30, 7 September 2012 (UTC)
The functionality you seek does not exist. Operating the &section=new feature without a subject is not recommended because there exists no way to provide an edit summary. Goodvac (talk) 04:07, 7 September 2012 (UTC)
But the problem is no different from simply clicking the "New section" tab at the top of this page, which also does not allow an edit summary. I find this frustrating, as I either have to use an over-long section title to explain things, or leave editors with this page on their watchlist wondering what the edit is about. - Arjayay (talk) 08:13, 7 September 2012 (UTC)

Template on article not displaying coordinates

Ysbyty Ystrad Fawr has latitude & longitude data in its infobox but there's no sign of a coordinate on the rendered page. Why? If it could be fixed, I'd be obliged. --Tagishsimon (talk) 23:15, 6 September 2012 (UTC)

If you check the documentation for {{Infobox hospital}}, you will find that latitude and longitude are "used only for adding a map, with map_type". You need to use the Coordinates parameter with {{coord}}. ---— Gadget850 (Ed) talk 23:28, 6 September 2012 (UTC)
I've added {{Coord}} to the infobox on that article. We could change the infobox so it would use the latitude & longitude parameters to generate a {{Coord}} instance if the infobox doesn't already have one, but that would cause problems with articles that have a {{Coord}} elsewhere on the page. A good example of why we should standardise on one method of coordinates (or any other type of) data entry and display, across all articles. Meh. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 10:53, 7 September 2012 (UTC)

css files

Is it possible to write a css file, and for it to be linked to a specific page when anybody views that page? (ie. not a user skin.)

Or views a page with that specific page transcluded?

Can't find anything, even in Help:Advanced_templates » Ζετα ζ (talk) 04:27, 7 September 2012 (UTC)

May want to check out Help:MediaWiki namespace. To the extent that its not covered by that, I don't think you can. Monty845 04:47, 7 September 2012 (UTC)
This is not currently possible, but is seen as a requirement to eventually fix bugzilla:35704. —TheDJ (talkcontribs) 10:01, 7 September 2012 (UTC)
Yes, inline styles don't help with Responsive Web Design. Also, css in the header can specify styles for pseudo-classes and pseudo-elements. Is it possible to use a little inline js to 'inject' a css file into a page? or does that open a whole new can of worms? I think I have a script that changes the css file of an existing link, for another. » Ζετα ζ (talk) 18:35, 7 September 2012 (UTC)
But it does help with another, totally unrelated issue, that I haven't mentioned yet :D (needs more investigation) » Ζετα ζ (talk) 18:35, 7 September 2012 (UTC)

Template code problem

Template:Infobox Ghana constituency (edit | talk | history | links | watch | logs) gives a red linked file error if there is a comment in the map1 parameter field. The articles are then thrown in Category:Articles with missing files. Can one of you code-heads sort it out. Cheers. -- Alan Liefting (talk - contribs) 09:43, 7 September 2012 (UTC)

It has now been done. -- Alan Liefting (talk - contribs) 10:19, 7 September 2012 (UTC)

Delay in updating the search index

After a couple of days of random search results (different numbers of matches when pressing refresh - usually typical of "stale slices") the search index has stopped updating; causing a backlog for us Wikignomes/ According to Help:Searching#Delay_in_updating_the_search_index this should be reported here - so this is what I am doing. Arjayay (talk) 16:09, 7 September 2012 (UTC)

Hi Arjayay, Search indexing was stopped for a while in the past 24 hours. I had to turn it off so that I could migrate data from one of our data centers to the other. Sorry for the inconvenience! It should be back up and fully up to date now. If the problem you were expereincing is still present, please let me know! Peteryoungmeister —Preceding undated comment added 16:34, 7 September 2012 (UTC)

Looking for help creating automatically updated graph

I would like to graph the total number of bytes all the dermatology project pages are currently, and how that number has changed over time. The graph would have X=months and y=bytes, similar to what was done for derm article creation. I would also be interested to have the graph automatically update monthly, similar to the article creation page. Could someone help me with this? ---My Core Competency is Competency (talk) 21:06, 7 September 2012 (UTC)

Wikipedian page fixed layout

Hello. I am Wikipedian administrator from latvian wikipedia and I am wondering, is there any discussions about Wikipedian page fixed layout implementation? I am asking this because recently I was using Wikipedia from multiple PCs with different screen resolutions and the thing is, as you might guess, the pages looks different, because width of wikipedian page is 100%. Maybe we need to start think about fixing this page width, because then we could write an article and expect the standard viewing experience in almost any screen resolution (as most of websites do these days). In that case we would be sure that page will look good in any conditions. Best regards, --Laurijs (talk) 19:46, 7 September 2012 (UTC)

  • Trending towards variable-size page layouts: Actually, many of us are actively modifying articles to custom-display for each user, such as default image-size 300px (rather than typical 220px) for sight-impaired readers who want to see larger images in an article. Wikipedia could have reformatted most large articles with larger/smaller images within 2 seconds, but unfortunately, the use of gargantuan templates will slow the 2-second reformat to become 20-45 seconds, as with most major pop-culture articles (and common medical articles). We need a faster Template:Weather_box (currently 4-5 seconds). Also, using fast-cite Template:Cite_quick can reduce article display time to about 3x-4x faster, but hand-coding of citations reduces the redisplay to about 4-5 seconds, so that is likely the best tactic for custom-viewing of major pop-culture articles. There are plans, as well, to quicken Template:Sfn to 2x (twice) as fast. However, forcing all pages to identical format is contrary to the growing diversity of hand-held devices for viewing pages. We want people to select default image-size 120px and quickly redisplay articles with tiny thumbnails, in 4 seconds, if requested. Large infoboxes only slow articles by 1 second, so other large templates are the issue. -Wikid77 (talk) 23:38, 7 September 2012 (UTC)
I'm not sure if you know this, but you can modify custom Cascading Style Sheets to remove the fluid layout (which, by the way, fluid layout is best practice so far as I'm aware). --Izno (talk) 23:54, 7 September 2012 (UTC)
Forcing fixed page layouts goes against one of the fundamental principles of the WWW. You cannot assume that the user's screen is a particular width; and so you cannot code for a specific width. In fact you can never make any assumptions about any of the screen characteristics - whether width, height, dot pitch, colour capabilities, etc. - you cannot even assume that the user has a viewable screen. All that you can assume is that the users have some means of converting a HTML document into something that they can understand, whether by sight, sound or touch. --Redrose64 (talk) 16:01, 8 September 2012 (UTC)
Something like a dynamic column layout however would be a good idea. Unfortunately CSS implementations in (mostly the older generation of used) browsers are still not really there yet. —TheDJ (talkcontribs) 21:01, 8 September 2012 (UTC)

list of my new articles?

How do I get a complete list of all articles of which I was the initial author? Michael Hardy (talk) 16:45, 8 September 2012 (UTC)

Go to 'my contributions' and in the box at the foot of the page is 'Articles created' which links to something at the toolserver. Mr Stephen (talk) 17:08, 8 September 2012 (UTC)
No wonder I couldn't find it! It's at the bottom of the page. I'd never have thought of looking there.
Thank you. Michael Hardy (talk) 19:22, 8 September 2012 (UTC)

Templates and expansion depth

Yes check.svg Done 80% - more in progress. -Wikid77 18:44, 29 Aug, 15:56, 31 Aug., 00:09, 2 September 2012 (UTC)

I was pointed here for this query but I am sure there is a forum for techie template talk.

I have been around here a while but I have only now encountered the message Page exceeded the expansion depth when editing a page. Seems they all go to Category:Pages where expansion depth is exceeded. Currently about 16.000 pages in it. Two of the templates that seem to do it are {{Automatic taxobox}} and also {{atn}} when there is no following archive page. I am no expert on templates or coding but should this be sorted out? -- Alan Liefting (talk - contribs) 22:12, 20 August 2012 (UTC)

That is excessive. Categories marked as {{Monthly clean up category}} are also in there, and they should not be. Most likely a subtemplate somewhere has recently been edited to make it more complicated - perhaps by only a tiny amount - but just sufficient to tip all these pages over the limit of 40 levels. --Redrose64 (talk) 23:02, 20 August 2012 (UTC)
One of the templates common to all those subcats is probably causing it. -- Alan Liefting (talk - contribs) 23:25, 20 August 2012 (UTC)
I thought of that - I took Category:Article sections to be split from April 2012 as one case and 5th Avenue Theatre as the other - I determined that the problem still existed if I stripped out everything except {{Monthly clean up category}} from the former, and {{Infobox Theatre}} from the latter case. Then I determined which subtemplates these two templates both used - and the only one these two have in common is {{max/2}} which hasn't changed in years. --Redrose64 (talk) 23:37, 20 August 2012 (UTC)
I think you're on the wrong track. It's probably the use of various string manipulation templates in both pages. Anomie 00:58, 21 August 2012 (UTC)
See Wikipedia:Village pump (technical)/Archive 99#Raising issue again: Page exceeded the expansion depth, Wikipedia:Village pump (technical)/Archive 99#Page exceeded the expansion depth? and Wikipedia:Village pump (technical)/Archive 100#Possible problem with the Article Feedback tool for previous discussions. My comment from the last still seems to apply for articles: "Sampling of Category:Pages where expansion depth is exceeded shows the main culprits are now {{Infobox German location}} (whenever the image_plan/Lageplan parameter is set) and species using one of several templates (even without parameters), for example {{Automatic taxobox}}, {{Speciesbox}}, {{Ichnobox}} or {{virusbox}}." All the four mentioned species templates cause it without any parameters. PrimeHunter (talk) 01:33, 21 August 2012 (UTC)
So how do we fix it? -- Alan Liefting (talk - contribs) 06:58, 21 August 2012 (UTC)
Create less complicated templates, or wait for Scribunto to be deployed (but that is still a few months away). —TheDJ (talkcontribs) 08:52, 21 August 2012 (UTC)
  • Species articles seem ok, non-fatal depth error: From working on those {Automatic taxobox} articles, I think the article contents are ok as is, because the depth-exceeded error is non-fatal there, but a long-term fix would be fine (or rewrite with Lua script). However, {Infobox_German_location} can be fixed easily (see below) to use efficient Template:Strfind_short, which avoids the 40-level limit, to even reformat 40% faster. -Wikid77 (talk) 02:57, 29 August 2012 (UTC)

────────────────────────────────────────────────────────────────────────────────────────────────────First one in the category: Category:1911 Britannica articles needing updates from August 2012. NewPP of that page: Highest expansion depth: 41/40. (Curiousity: sister categories (e.g. June 2012) with 41/40 and so categorised too is not in this list - let's "hopothese" this is just catsort that went wrong). {{Monthly clean up category}} is very suspected indeed, it is the only code on the page.

Solved the August 2012 example, individual page. -DePiep (talk) 07:34, 22 August 2012 (UTC)

Could it be this: This newPP limit is new this way (added recently to the newPP list). And the tracking category is from 8 May 2012 [11] by PrimeHunter. Has the newPP change + its category only surfaced existing pages, whose problem did not show this explicitly? -DePiep (talk) 05:33, 22 August 2012 (UTC)

Also discussed at Infobox_German_location, May 2012. -DePiep (talk) 05:39, 22 August 2012 (UTC)
List of causing/suspected templates (see also notes above):
* (please expand list here)
-DePiep (talk) 05:59, 22 August 2012 (UTC)
Continued for the cleanup talks: Category talk. Category talk:Pages where expansion depth is exceeded#Existing templates to be reviewed. -DePiep (talk) 06:18, 22 August 2012 (UTC)
Still unexplained: {{Monthly clean up category}} has ~4700 transclusions, but only ~300 show up in the excess category (as Category pages, so on top). Now see example Category:1911 Britannica articles needing updates from June 2012. Page source newPP says: expansion depth 41/40, but the category is not added. (also tried ~nulledit to trigger March 2012 -- no effect AFAICS). -DePiep (talk) 07:34, 22 August 2012 (UTC)
Depth 41/40 sometimes allowed: For some cases, a reported depth of 41/40 does not trigger "exceeded" and, remember, the top level is 1, not zero, so that 40-nested is actually level 41, counted from "1". -Wikid77 13:11, 1 September 2012 (UTC)
  • Fix Infobox_German_location to use Strfind_short: I have created a fix for the depth-exceeded limit, in Template:Infobox_German_location/sandbox2, as the update to use efficient Template:Strfind_short (7 levels), rather than slow Template:Str_find (23 levels), to avoid the 40-level limit, also center the plan-map, and even reformat 40% faster. The NewPP markup will show only 31 levels used, and 10,100 fewer preprocessor nodes, saving 2-3 seconds (40%) on reformat time. I thank everyone, above, who noted the problems and suggested the string-handling templates had exceeded the limits, in these extremely complex templates which I have spent months analyzing and updating. -Wikid77 (talk) 02:57, 29 August 2012 (UTC)
Yes check.svg Installed, thanks. -Wikid77 23:08, 29 August 2012 (UTC)
  • Fix for Template:Monthly_clean-up_category/monthname: To fix the expansion-depth limit exceeded in categories, the related subtemplate, Template:Monthly_clean-up_category/monthname needs to be updated from the /sandbox version. The fix changes the search to use efficient Template:Strfind_short (with only 9 expansion levels), rather than the prior Template:Str_find_long with 23 levels plus {Str_right} totalling to 35 levels of the 40-level limit. The /sandbox version also has more doc-text to explain the use of the template. That /monthname subtemplate is invoked from {{Template:Monthly_clean-up_category}} which is used inside hundreds of category-description pages, all exceeding the depth. -Wikid77 (talk) 17:22, 29 August 2012 (UTC)
Yes check.svg Installed, more fixes needed. -Wikid77 23:08, 29 August 2012 (UTC)
  • Beware Infobox_Theatre uses embedded Infobox_NRHP: Although infoboxes nested inside others might look good as keeping within the same box, when Template:Infobox_NRHP is set "embed=yes" inside {Infobox_Theatre}, then the expansion depth increases by 7 levels deeper. So, when using {Convert} inside the double-nested infoboxes, the depth reaches 42 deep. In such cases, {Convert} must be reduced by the round-value parameter "|1" or such to reduce nesting by 7 levels, back to 35 deep, inside the double-nested infobox. I have edited article "5th Avenue Theatre" to unnest the infoboxes, and avoid limit exceeded. -Wikid77 23:08, 29 August 2012 (UTC)
  • Essay WP:EXPANSION about reducing depth: I have written another essay, this time as general title "WP:Expansion depth limit" (WP:EXPANSION) to begin to explain more ways to keep the expansion levels, in various types of articles, from exceeding the limit. Also essay: "WP:Avoiding MediaWiki expansion depth limit". We need to de-mystify these issues, for quicker fixes, so that people do not wander for months in frustration at the bizarre "exceeded" messages. Otherwise, this could result in people overwhelmed by too much bizarreness. Long-term, we need to raise the depth limit of 40 to be 60 or 80 or such. -Wikid77 23:08, 29 August 2012 (UTC)
  • Bio/species articles use Template:Speciesbox: Although they appear to format correctly, the exceeded-limit message comes from Template:Speciesbox, which uses the huge Template:Taxobox/core with numerous parameters, and no obvious extravagance of overly nested if-else-else-else-else. By using Template:Italic_title, it uses {Str_find} which runs 23 levels deep, and perhaps that is the cause of some exceeded-limit messages. However, the whole of Template:Speciesbox always runs about 36-38 levels (of the 40-level limit), and hence some functionality is stopped in over 4,500 articles. There is no policy to "explain-your-complex-template" so we will need to write the operational notes for Template:Taxobox/core/sandbox2, to illustrate a template "call tree" to show which subtemplates are deeply nested to levels 36-38, and then use those notes to determine where to trim the huge monster, which probably needs to be reduced by just 2 or 3 levels more, as the first step. More: After drawing the call-tree, I found Template:Taxobox/taxonomy shows the taxon-name rows, and Template:Taxobox/taxonomy/3 causes the "exceeded-depth" and must be shortened below 33 levels. I have updated some template /doc pages to explain operation and submitted /sandbox updates to those templates at Template_talk:Taxobox, to fix the exceeded-depth. The trial Template:Speciesbox/sandbox2 can be used to test the fixes in any article, using all coordinated "/sandbox2" versions of templates. I think that completes this PUMPTECH topic for now. -Wikid77 (talk) 00:17/15:56, 31 Aug., revised 13:11/00:09, 2 September 2012 (UTC)

Great progress, and thanks for your persistence. Look like the taxobox is the main culprit now. Cheers. -- Alan Liefting (talk - contribs) 09:03, 7 September 2012 (UTC)

  • Multiple {Taxobox} depth problems: There are multiple depth problems after Template:Taxobox/core, so I am rewriting Template:Taxobox/taxonomy and Template:Get_regnum() and others in the 60-stack of taxon names. Many articles will then fit, but I think some others will still hit the depth limit. It is an overwhelmingly complex system, due to the lack of expansion-depth counts for each template; and perhaps the expected levels were 50 with the 41-level limit. Hence, many levels must be reduced to fit all cases. However, some can be fixed now by merely setting |taxon={{PAGENAME}}. -Wikid77 (talk) 23:35, 9 September 2012 (UTC)

Blinkity Blink!

I just ran across a username that looks like this:

~~Username~~ Talk Page

This is invoked with <span style="text-decoration: blink">.

Blink? BLINK?? BLINK??? After the pitched battle to Kill The Undead Thing in HTML, here it is, like a zombie, polluting WikiMarkup. Kill it. Kill it with Fire.

In case I was too subtle, I dislike the blink tag. see Blink element, [ ], [ ], [ ], and [ ]. --Guy Macon (talk) 14:12, 26 August 2012 (UTC)

In my opinion, absolutely annoying and disruptive. Is is possible to simply approach the user and offer some guidance and constructive rationale for changing it? Cindy(talk to me) 14:31, 26 August 2012 (UTC)
Wikipedia:Signatures#Appearance and color. Per guideline, this is inappropriate. --Izno (talk) 14:34, 26 August 2012 (UTC)
Thanks for the guidline link. I will pass it on to the users (who I am choosing not to shame here). He is a reasonable fellow so there should be no problem.
To look at the larger picture, why do we allow blinking at all? There is only one use for it... --Guy Macon (talk) 14:43, 26 August 2012 (UTC)
We allow most if not all CSS. Simply because it is bad practice doesn't mean we shouldn't. It's a freedom that authors should have to use, if even it should only be used in the rarest of occasions. --Izno (talk) 14:47, 26 August 2012 (UTC)
Well, I also came over exactly the same signature and have exactly the same feelings. I hope there is a way to cut off such annoying markup, or at least disable it in preferences. — Dmitrij D. Czarkoff (talk) 15:54, 26 August 2012 (UTC)
For the record: I just found out that "text-decoration: blink;" part isn't recognized by my browser, so I see this:

~~Username~~ Talk Page

And I still find it overly annoying. — Dmitrij D. Czarkoff (talk) 16:22, 26 August 2012 (UTC)
It's Firefox only... Mozilla probably forgot to take it out of the Netscape codebase... or didn't they? Edokter (talk) — 16:45, 26 August 2012 (UTC)
Probably left in for backwards compatability. It can be disabled in your browser by entering "about:config" into your address bar, clicking through the warning, entering "browser.blink_allowed" into the search box, and then double-clicking the line to change from true to false. Anomie 18:53, 26 August 2012 (UTC)

BTW, I think that [ ] should be promoted as signature formatting policy. — Dmitrij D. Czarkoff (talk) 18:44, 26 August 2012 (UTC)

Just did a quick test; blinks in Firefox, doesn't blink in IE, Chrome, Opera and Safari. (Latest versions, of course. I am pretty sure that if you go back far enough, Opera and IE will blink.) As for Wikimarkup allowing most if not all CSS, I can think of some clever forms of vandalism that would be invisible to most patrollers but visible to users using certain browsers:

Test 0: Test <-- Does anybody not see the word "Test" in bold here? (No CSS)
Test 1: Test <-- Does anybody not see the word "Test" in bold here? (visibility:hidden)
Test 2: Test <-- Does anybody not see the word "Test" in bold here? (display:none)
Test 3: Test <-- Does anybody not see the word "Test" in bold here? (font-size:1%)
Test 4: Test <-- Does anybody not see the word "Test" in bold here? (color:white; background-color:white)
Test 5: Test <-- Does anybody not see the word "Test" in bold here? (color:#fefefe; background-color:#ffffff)
Test 6: Test <-- Does anybody not see the word "Test" in bold here? (position:fixed; bottom:16383px; right:16383px)
In my opinion, Wikimarkup should restrict some CSS tricks. --Guy Macon (talk) 20:53, 26 August 2012 (UTC)

I see tests 0 and 3, spot tests 4 and 5 (I use modified color scheme) and don't see the rest of them. And I believe that such instances should be patrolled manually anyway — they show up in diffs. — Dmitrij D. Czarkoff (talk) 21:11, 26 August 2012 (UTC)
I think anyone who has a blink tag in their signature needs shooting, minimum. I mean some signatures are anti-social but that's the worst I've seen. Secretlondon (talk) 21:25, 26 August 2012 (UTC)
On FF13, I see Test 0, 4 and 5 show up on highlight, 1 is a blank space, 2, 3, and 6 don't even show the space for the word. --Nouniquenames (talk) 05:48, 27 August 2012 (UTC)
Perhaps we should have Wikipedia:Edit filter add a "styling added" tag to make it hard for patrollers to miss such things? I didn't realize we had so many filters until I ran across Special:AbuseFilter. --Guy Macon (talk) 21:56, 26 August 2012 (UTC)
Well, probably we should have edit filter for all or most CSS and HTML, and abuse filter for "blink" or "marque" within HTML or CSS... — Dmitrij D. Czarkoff (talk) 22:19, 26 August 2012 (UTC)
And for visibility:hidden, display:none, font-size:[>10%], color:white, color:#[e-f][*][e-f][*][e-f][*], position:fixed, etc? See Whac-A-Mole. :) --Guy Macon (talk) 22:31, 26 August 2012 (UTC)
These can have legitimate uses under some circumstances (collapsing, spans in extra large labels, templates with dark background, etc.), but when one tries to add <blink> or <marquee>, he is definitely trying to secure mass vision loss among Wikipedians. — Dmitrij D. Czarkoff (talk) 23:31, 26 August 2012 (UTC)
If you set a filter, you'll be surprised about how often we use these. Navboxes, infoboxes, user boxes article history and tons of other templates make actively use of many of these. If you want to patrol them, I don't think anyone will stop you, but I doubt it's anywhere near as problematic as you think. —TheDJ (talkcontribs) 12:09, 27 August 2012 (UTC)
Yes please stop the blinking. and while you're at it, the "background/blurring" technique that's been popping up lately - it makes it hard (for me at least) to read the signatures at all. - jc37 22:59, 26 August 2012 (UTC)
IMO signatures customization should be limited to choice of prefix ("-", "–", "—", "--", etc). — Dmitrij D. Czarkoff (talk) 23:21, 26 August 2012 (UTC)
Yep, I agree. We are all equals here so our signatures should reflect that. -- Alan Liefting (talk - contribs) 00:54, 27 August 2012 (UTC)
Very much agreed, with all 3 above. Personal aesthetics belong only on our Userpages, not forced on everyone everywhere.
(Although, a garish signature is often a handy warning-device, for anyone interacting with them... >.> ) -- Quiddity (talk) 23:13, 28 August 2012 (UTC)

See also: Wikipedia:Edit_filter/Requested#Abusable_CSS. —TheDJ (talkcontribs) 13:50, 27 August 2012 (UTC)

I'd suggest implementing User:Ais523/highlightmyname2.js as a gadget, possibly even turned-on by default. It replaces the only good reason to color one's own sig (making it visually stand out, so that you can scan down a long talkpage looking for your own comments, and any new replies). I've used it for years, with great satisfaction. My name is lime-green, whereever it appears in Wikipedia (talkpages, histories, etc) but nobody else has to see that, nor should they have to. -- Quiddity (talk) 23:13, 28 August 2012 (UTC)

I've started an RfC on signatures formatting at Wikipedia talk:Signatures#Simplifying signatures. — Dmitrij D. Czarkoff (talk) 00:13, 31 August 2012 (UTC)

Note also the draft policy at Wikipedia:Manual of Style/Accessibility/Signatures which I started some time ago, and which could be modified and adopted as a result of this RfC. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 11:46, 3 September 2012 (UTC)

What about removing the ability to customise signature markup, add

<span class="…"></span>

to the standard, and use User_style? Users could publish the css for their signature for others to copy. — Preceding unsigned comment added by Ζετα ζ (talkcontribs) 20:56, 8 September 2012 (UTC) Sorry everyone – yes it is me. » Ζετα ζ (talk) 22:21, 9 September 2012 (UTC)

I actually like slightly non-standard signatures, provided they're not in-yer-face hideous. It helps me to 'recognise' which people are involved in a discussion, and see an individual person's comments without having to use Find to locate their sig. I'm talking here about different colours, different fonts, maybe a surrounding box, etc. Not things which flash, ping, or go bump in the night ;P Pesky (talk) 06:36, 9 September 2012 (UTC)
In my opinion, the best answer is an option to turn off seeing sig customizations in my preferences, with "on" being the default. Plus, of course, a good trout slap for sigs that blink, look out of focus, etc. --Guy Macon (talk) 01:07, 10 September 2012 (UTC)

Could it be possible to watch move logs, deletion logs etc. separately?

Resolved: ···Vanischenu「m/Talk」 14:47, 9 September 2012 (UTC)

This question was first asked at the Help Desk (archived) which was not resolved there. Thanks to Vchimpanzee who guided me to ask here.

  1. Is there any way I can see move logs or deletion logs separately in Watchlist, RecentChanges, RecentChangesLinked or anything similar?
  2. Is it technically possible?
  3. And does it worth proposing to include them?

Thank you···Vanischenu「m/Talk」 22:26, 30 August 2012 (UTC)

  1. You can see the move log at Special:Log/move, and the deletion log at Special:Log/delete. These pages will show all moves and deletions, with the most recent listed first (like recent changes). I don't know any way to get anything similar for your watchlist or related changes, though.
  2. It should be technically possible for the developers to add an option to filter the watchlist or recent changes to only show logs. It might also be possible with a user script, though I don't know how.
  3. Proposing something is a good way to find out if anyone is interested. – PartTimeGnome (talk | contribs) 20:44, 4 September 2012 (UTC)

Thanks you so much. I will propose it, then.···Vanischenu「m/Talk」 14:47, 9 September 2012 (UTC)

HTML5 is coming to Wikimedia Wikis...

We're going to be enabling HTML5 [12] on all Wikimedia wikis on Monday 17th September 18:00-20:00 UTC [13]. Just a heads up for anyone that might actually care. Reedy (talk) 22:08, 5 September 2012 (UTC)

Whether or not I care (I haven't decided yet), it's great to get heads up such as this and the edit window change a few posts back. Much to be encouraged. --Tagishsimon (talk) 22:14, 5 September 2012 (UTC)
<VP stalker>Thanks! </VP stalker> :). Okeyes (WMF) (talk) 22:32, 5 September 2012 (UTC)
Have to check again, but I think this should resolve issues with the new table sorter. How about HTML Tidy? ---— Gadget850 (Ed) talk 22:47, 5 September 2012 (UTC)
Is this causing <font> to disappear. If so, you might want to make a bigger announcement as it will affect alot of people's signatures. Bgwhite (talk) 00:05, 6 September 2012 (UTC)
Case in point ---> LadyofShalott 00:11, 6 September 2012 (UTC)
Will this cause any other markup (or other things) to not work? David1217 What I've done 00:18, 6 September 2012 (UTC)
No <font> will still work, and everything else should look the same except for some tiny browser-dependent positioning changes. Actual rendering depends on browser implementation, not the standard. However it will make any pages with font tags fail to validate when they might validate today. —Cupco 00:25, 6 September 2012 (UTC)
It has been reported before that <font> and <center> will no longer work on the changeover. Bgwhite (talk) 00:38, 6 September 2012 (UTC)
That's true, HTML5 relies on the CSS template to present items. Here are the elements I know about that are being removed from HTML: basefont, big, center, font, strike (<s></s>), tt, frame, frameset and noframes. There are probably more. Some parameters we use that are also going away are: charset, coord, archive, nowrap, valign, type, width and quote a few others. We should expect some really significant fallout from this "upgrade". Forgive my pessimism but I sincerely hope that more testing has been done than some of the previous "upgrades" or we are going to have a lot of fallout and problems. Kumioko (talk) 01:47, 6 September 2012 (UTC)
Could someone else confirm these, particularly someone who has tested them on the testwiki (i.e., "how WP will behave")? The center tag is used a lot in articlespace, and strike is used extensively in all sorts of discussion pages to "mark as removed" specific comments and as markup for content-deletion changes. DMacks (talk) 05:22, 6 September 2012 (UTC)
Based upon August's dump, (if I did it right) 30,013 articles use center, 208 articles use strike and 7,379 use font. Bgwhite (talk) 06:42, 6 September 2012 (UTC)
  • Elements will not stop working. Where implemented, MediaWiki will translate to make it HTML5 valid. Otherwise, it will just output the old tag or attribute and the page will not be valid, test2wiki:User:TheDJ shows that font and big will simply remain working. In the future more internal translation will be added to convert these to proper html5 as well.
  • So after we upgrade, we will have a lot of talk pages that are not valid HTML5. This is not considered to be a problem. However, ppl should stop using the font tag in their signatures and start using span. Not because otherwise stuff will break, but because they chose to write HTML in their signature, so they should make sure they use the appropriate HTML.
  • HTML5 mode has been the default for over 3 years now. Screenscraping has been illadvised for over 7 years now. Due to the huge amount of legacy user written code on en.wp however it has been incredibly difficult to make the switch, because, well because volunteers simply don't rewrite code until it breaks. I expect some things to break again, this is not a problem. People attending this forum will know how to fix them. What is important is that the biggest user-written tools have now been either converted or can expect to break because they didn't follow trough the last time the tool broke. Waiting longer probably won't spur the last group into action.
  • Yes, things break when this kind of thing comes to en.wp. The words 'do more testing' are often flung back at developers, but people forget that there simply is nothing remotely comparable to English Wikipedia. In terms of both content, community customization and community itself, en.wp is unique and will always be. —TheDJ (talkcontribs) 06:53, 6 September 2012 (UTC)
BTW. This doesn't mean that HTML5 won't be disabled again if required, like the last times. —TheDJ (talkcontribs) 07:42, 6 September 2012 (UTC)
  • I have one correction to that statement DJ. Some of us are aware of some of the problems mentioned above however attempts to replace the HTML tags or to fix them result in various editors getting upset that we are performing minor edits that don't change the rendering of the page or something like that. The resolution is clear to me to fix these and that is to do a bot or add some code to AWB to fix these things 1, 2 or a few at a time until they are done. I have done things in the past to clean this up and had users threatening to revoke my AWB access or even block me from editing. There are some things that can be done now, before we flip the switch but we are not allowed too without going to each article manually which frankly is extremely stupid. Kumioko (talk) 10:14, 6 September 2012 (UTC)
    • There is only reason to change 'new edits' and templates. At some point the core will be fully able to take these 'older' html4 elements and attributes and translate them into HTML5 equivalents. Also, there is no rush on this front. —TheDJ (talkcontribs) 12:02, 6 September 2012 (UTC)
Here is a document describing HTML features marked as obsolete in HTML5. My experience is that "obsolete" doesn't mean "these will stop working with immediate effect", but "these may not work in HTML5-compliant browsers, and validity checkers will complain, but there is nothing to prevent the browser authors from providing these features for the convenience of legacy code". So, for example, we should not introduce new uses of <tt>...</tt> but use <kbd>...</kbd>, <var>...</var>, <code>...</code>, or <samp>...</samp> instead; but there is not an immediate rush to "fix" existing uses. --Redrose64 (talk) 12:58, 6 September 2012 (UTC)
First I want to state that I understand what you are saying, both of you, but I don't like the habit that we have on WP of kicking the can down the road for the sake of maintaining obsolete code, templates, etc. just so that we don't do minor edits. Second, if not doing these edits causes problems in other applications such as browsers then they are not minor and third, by leaving them in place, we then have to do other types of work to compensate such as modifying the CSS templates to compensate. Leaving these items in place just perpetuates the problem. Users see it working and copy it, they get in the habit of using outdated techniques. We can do most of these changes with a bot, initially the edits might be fairly high but if we stay on top of it then they won't be that bad, especially if we do them at the same time as doing other edits. Otherwise we just have trash piling up in the corner and then at some point down the road when something else happens then we have to jump through hoops to fix a problem that should have been delt with long ago but no one wanted to take the time to do it right and just whitewashed it. Kumioko (talk) 13:58, 6 September 2012 (UTC)
You are adding complexity where it is not really required I think. Remember, you can't change past revisions, so whatever is in the database, will need to work ANYWAYS, regardless of what we use as 'current' practice. So the complexity should be in the core to make sure it's wikicode is displayed as HTML5. Also do not forget that HTML5 is a living standard. It has removed the obsolete marker from elements before and might still do it again (though less likely by the month). Quite a bit of restraint (in the region of years) is not an unwise when it comes to this revisioned wikicode of ours. For the same reason. <b> and <i> were actually obsolete for a long while, it would have been a shame if they had been removed everywhere (pretty much permanently) for nothing. Perhaps some time (far far future) when we have a sane Parsoid parser and a sane DOM model, we might be able rewrite past revisions into the new DOM model. Then I would be more supportive of your approach, but for now, I and I think many developers rather take it slow with the persisted data when it comes to HTML5. —TheDJ (talkcontribs) 14:46, 6 September 2012 (UTC)
I see what your saying but I also think that we should be able to develop a way to change these things by bot as necessary and develop a standard. Right now we say we have a standard but then have a million caveats and allow anything and then call it a standard. Once we implement a change then we should fix the things to conform to that. Also using your example <b> and <i> went away for a while but there was still a way to do it using other coding. It just wasn't using <b> and <i>. In the end its you folks decision but IMO it would be better to run a bot to align these changes when they are implemented and then adjust as necessary in the future rather than to accept everything may change and do nothing. Will it fix everything of course not, but as this is a major change it will give us a solid base from which to base a standard use of coding techniques based on current standards and we can adjust as necessary going forward. Of course we need to do some testing because not everything in HTML5's new coding may be supported in all apps such as older browsers but its something to consider. Kumioko (talk) 15:11, 6 September 2012 (UTC)
I think that it would be a mistake to have a bot editing wikitext to convert it to HTML5. The fact is, wikitext is not HTML, even though it allows use of a subset of HTML tags. ''' isn't valid HTML, any more than <s> will be under HTML5. Users don't have to worry about whether the parser emits <b>, <strong>, or <span style="font-weight: bold"> when it comes across three apostrophes. We shouldn't have to worry about using <s> either. --R'n'B (call me Russ) 15:39, 6 September 2012 (UTC)
Kumioko has a point. Correct the problem now while we have time. Removal of the tags will allow the removal the hack that keeps the tags going in the Mediawiki software. It makes sure all new edits use valid HTML5. There will be only one standard and not several. There are no downsides, only benifits from the removal of the tags from the articles. In the long run, this is the way to go. Bgwhite (talk) 18:03, 6 September 2012 (UTC)
──────────────────────────────────────────────────────────────────────────────────────────────────── Potential downsides to mass-editing to "fix" existing wikitext:
  • creating unnecessary edits for hundreds of thousands of pages (including talk pages and perhaps archive subpages), which would swamp watchlists and recentchanges lists during the conversion process
  • the risk that conversion would misinterpret wikitext in some cases (for example, on pages where there are complex transclusions, mismatched brackets, badly formatted text or unexpected formatting tags), causing pages to render incorrectly without being detected
  • confusion among established users as to which tags have ceased to work and which tags they ought to use to replace them.
But the MediaWiki parser already renders some legacy markup as modern HTML and/or CSS without having to change the article wikitext (as in Sanitizer::fixDeprecatedAttributes()), and this could be extended to a wider range of deprecated tags and attributes if there is a demonstrable benefit. So valid HTML5/CSS3 output is possible without any of the above disruption.
Richardguk (talk) 22:59, 6 September 2012 (UTC)
  • As said above, talk pages would not be touched. We are left with ~308,000 pages. That is not many at all.
  • It is a straight foreword find-replace for the strike and center tags. Font tag (7,000 pages) would be different.
  • Confusion is already here. Do I use the font tag or span? So, there are multiple of ways to do one thing and more code has been created to fix tags that shouldn't be there. Why have Sanitizer::fixDeprecatedAttributes at all if there can be a good fix? How long will Sanitizer::fixDeprecatedAttributes stay around? To paraphrase a quote about developers, do not change anything until it breaks. Why not take the time to fix it properly now or end up with a quick fix in the future when it breaks. Bgwhite (talk) 00:17, 7 September 2012 (UTC)
It should also be noted that:
  • We wouldn't necessarily need to fix every problem by bot
  • Bot edits can be ignored easily by the users and should probably be marked as minor
  • We could use a standard edit message that would tip the users off to what was done so it would be further easier to ignore.
  • Adding additional hacks to the CSS style sheet will probably be needed your right, but we shouldn't rely on that or assume that we will be able to continue to fiddle with the CSS coding. Somethings are easier to mask than others.
  • Fixing these types of issues as they come up is an industry best practice employed at most major IT companies. Fairy dusting the code to hide the problems works, and will probably still be needed for the article histories. But if we stay on top of things and stop taking this there's no rush mentality then we'll never clean it up because people will still use it and copy it throughout the pedia forever.
  • The initial run will be the worst, but after that shouldn't be too bad. If we start by doing the ones with other problems first, then we can sweep up the rest to minimize the unnecessary edits. Many articles will have other problems that need to be fixed at the same time. Kumioko (talk) 01:07, 7 September 2012 (UTC)
  • Perhaps create more HTML templates or bot-edits: I was hoping the wikitable parser would translate the old keywords. However, if not, we can quickly write templates to mimic the reasonable keywords, such as {valign|top} to simulate "valign=top" or {font|color=blue}, and then perhaps run some bot-edits to call those templates, rather than the bureaucratic HTML5 (Horribly Twisted Mangled Language style="version:Five; more:later", sorry couldn't resist the joke). Meanwhile, the short tiny templates can run over 750 per second, so they will not slow pages even where not needed, but we just have to stop the rampant deletion of tiny templates. Perhaps make the short template names to be redirects of some long-winded template names, such as "{HTML_keyword/font}" and then they survive deletion because they are tedious to type and seem bureaucratically organized into tedious collections. Let's see if these tedious ones still exist:{Col-1-of-2}, {Col-2-of-2}, {Col-3-of-4}, yes tedious sets of even useless templates still survive deletion, etc. -Wikid77 (talk) 05:41, 6 September 2012 (UTC)
  • Comment: I don;t usually chip in here, but I have to say that I really think Kumioko's right, here. Whatever edits (bot) would be necessary to bring current versions of pages into line would be good; otherwise we are just piling up problems for the future. I see no reason why such edits couldn't be marked as minor, and as bot, so as not to swamp people's watchlists. I'm just kinda looking ahead to ten years down the line, when someone will be tearing their hair out and thinking "OMG, why didn't we fix this when it was so much smaller ...?" Brushing stuff under the carpet only works until there's so much fluff under there that it makes the floor surface lumpy and trips people up ;P Pesky (talk) 04:18, 9 September 2012 (UTC)
The obsolete tags will always be supported. There are literally millions of pages out there written in HTML3 which will never be upgraded, and therefore browsers will always support legacy HTML code, even when it doesn't validate against the latest committee standard. Obeying standards committees with few actual programmers on them is among the lowest priorities. —Cupco 03:54, 10 September 2012 (UTC)

Wikipedia pages not formatted properly

Hi. I've been noticing that part of the Vector stylesheet doesn't load in my browser for some reason. (talk) 02:33, 9 September 2012 (UTC)

Please give us an example of the pages you have the issue, maybe even a screenshot, and let us know what browser you use (and what version).  Hazard-SJ  ✈  05:26, 9 September 2012 (UTC)
First completely clear your cache if you haven't already. PrimeHunter (talk) 10:35, 9 September 2012 (UTC)

Another error on statistics


This website,, is not reading any views after September 3, 2012. I wonder what is going on with the site. --George Ho (talk) 09:11, 9 September 2012 (UTC)

I might be doing something wrong but when I click that link it takes me to a "Enter a wikipedia article title and press Go" page. Is there a certain article you are referencing or even this exact wiki page? Thanks Marketdiamond (talk) 16:10, 9 September 2012 (UTC)
Just a comment it is all pages you enter from 2 September onwards. Henrik has not edited since 3 June.Blethering Scot 16:35, 9 September 2012 (UTC)
It's fixed now, with a ~week gap in data. —Cupco 03:55, 10 September 2012 (UTC)

A spacing insert

It is for a wiki userbox so it is a visual thing, is there some <> that makes additional space or void characters for spacing? Thanks! Marketdiamond (talk) 16:08, 9 September 2012 (UTC)

One of the ways is non-breaking space &nbsp; Example: 1  2   3    4     5. PrimeHunter (talk) 16:29, 9 September 2012 (UTC)
Thanks! have seen that, always wondered what it really meant! Marketdiamond (talk) 16:31, 9 September 2012 (UTC)

Lua might seem slower but

The preliminary tests, on, with writing Modules in the Lua script seem to indicate only minor performance gains, certainly not 20x times faster for everything. However, there are several complicating factors:

  • Templates, copied verbatim, over to also seem 50%-90% slower, in many tests, day or night.
  • Other text, copied verbatim, over to seems to reformat 50%-90% slower, day or night.
  • Lua on enwiki might run 2x faster due to unforseen tuning improvements here.
  • Each might be faster, in some aspects, where Lua provides special features, or vice versa.
  • Mixtures might be even faster: perhaps combining markup templates and Lua script modules, carefully, might create a hybrid faster than either alone.
  • For computer technicians, writing a "standard" computer-language Lua script would be faster, for them, than learning "esoteric" arcane tricks of the error-prone template coding.
  • Competition and cooperation of text-formatting techniques, between the 2 languages, might lead to many unforseen improvements, in later months.

For faster articles, the ultimate solution is clear: almost any huge article with no templates reformats within 2 seconds. If all templates are hard-coded or wp:subst'ed within a major article, then it could be reformatted for any user-preference, such as default image-size higher than 220px, within 2 seconds. Numerous tests have shown that article "India" here, with all prior templates, reformats in 25-29 seconds due to using 801 templates, such as hundreds of Template:Citation and Template:Sfn. I have created several fast-cite versions of those templates which can reduce the reformat time of "India" to nearly 10 seconds; however, the bottom-line fact is that removing all templates allows "India" with hundreds of hard-coded footnotes, and all images, to reformat in only 1.5 seconds. Keep the huge nation infobox, and it becomes 2.5 seconds. The file servers really are lightning fast, and small templates run over 750 per second, but large templates are bizarrely slow as huge behemoths. For very popular articles, then hard-coded citations might be the best choice, while thousands of less-viewed articles can use cite-templates instead. However, until Lua is installed on enwiki, and the operational parameters are tuned for faster speeds, we still are unsure as to the live speed here. Then, long-term, we can investigate other methods of formatting data in articles, whether with improved templates or alternative modules coded in Lua script. Both systems have future potential. -Wikid77 (talk) 09:56, 6 September 2012 (UTC)

Tim yesterday committed his Lua profiler, which should help get a better grip on the exact numbers. —TheDJ (talkcontribs) 11:34, 6 September 2012 (UTC)
For the record, that was made on gerrit: 22710. Helder 18:17, 10 September 2012 (UTC)

Problem with conversion of miles to kilometres when page is rendered as PDF

Resolved: Ruslik_Zero 07:06, 10 September 2012 (UTC)

Please see Template talk:Convert#Problems with this template in Book namespace. It is not a problem with {{convert}}, nor with Book namespace, but something inside {{Infobox road/meta/length}} when rendered as a PDF. --Redrose64 (talk) 17:02, 7 September 2012 (UTC)

I've narrowed it down to the following bit of code in the {{rnd}} template
  • {{#expr:(1E5round0)E5}} which gives 10000000000.
Try converting User:WOSlinker/list to a pdf to see the error. -- WOSlinker (talk) 17:49, 7 September 2012 (UTC)
  • Feel free to hand-code conversions, when problems arise; also, adding round-value parameter "|0" in {Convert} might change the rounding result. I have a new rounding template, Template:Rndpad, in development, which avoids Template:Rnd/- for medium-sized values, such as {{rndpad|(1E5round0)E5|0}} → 10000000000, which is 1E10 with 10 zeroes. Hence, there are many alternatives. -Wikid77 (talk) 23:08, 7 September 2012 (UTC)
Thanks, I think if it were a couple of occassions on a couple of pages we could manually replace them but this is alot of transclusions on thousands of pages making it hard to replace them all. Kumioko (talk) 23:51, 7 September 2012 (UTC)
  • Try sandbox version: Someone might want to try Infobox road/meta/length/sandbox, the /sandbox version, which I have changed to round the result by {#expr: ...round...} rather than Template:Rnd, which can add trailing zeroes or x10 notation of huge numbers. The expansion depth will be reduced from 12 to 6 levels, of the 41-level limit, by omitting {rnd}. For miles/km, I think typical rounding is fine, without worry of trailing zeroes. If the sandbox makes matters worse, then just revert. In complex computerized procedures, often the quickest solution is to "try everything" as a shotgun approach, and use whichever of various options will solve the problem. Pure deduction of cause-effect connections can be very tedious and time-consuming, so try-everything might actually be faster to try multiple changes, rather than think how the internal procedures work to pinpoint the specific bug. -Wikid77 (talk) 05:02, 8 September 2012 (UTC)
  • Oh dear. I had hoped, that per WP:MULTI, people would comment at Template talk:Convert#Problems with this template in Book namespace, not here. Now we have three separate discussions going. --Redrose64 (talk) 15:45, 8 September 2012 (UTC)

WikiLove customization

Hi, following User:Quadell/common.js/wikilove.js, I'm trying to add a type of WikiLove called 'Badges'. The mockup badge I'm using is at Wikipedia:BADGE/mockup. I'm having trouble getting anything to change in Wikilove, let alone actually delivering the Badge properly. My code is at User:Ocaasi/common.js. Can you help point me in the right direction? Once I have one example I'm pretty sure I can figure out the rest. Thanks! Ocaasi t | c 18:49, 9 September 2012 (UTC)

Looks like it's a load order issue. Will investigate. Kaldari (talk) 18:27, 10 September 2012 (UTC)
Thanks! Ocaasi t | c 18:40, 10 September 2012 (UTC)

It looks like something has changed recently regarding JS load order. If you add...

mw.loader.using( 'ext.wikiLove.defaultOptions', function() {
} );

around your code, it should at least allow you to add new modules. You still won't be able to delete modules, however, as this will cause an error when the local wiki overrides are loaded. Kaldari (talk) 23:18, 10 September 2012 (UTC)

Works brilliantly! Now, do you have any idea how instead of an image+message I can deliver a template + message, as in {{Wikipedia:BADGE/mockup}} That would really be awesome. Either way, thanks! Ocaasi t | c 00:07, 11 September 2012 (UTC)

Cat handler

I am puzzled by what has happened at Raleigh Moncrief. With this edit an IP tagged the article G5, in a peculiar way using {{db-meta}} and {{cat handler}}, neither of which I have met before. When WilyD removed the visble db template with this edit, that left "cat handler" in place, and the article was still listed at CAT:CSD. It was actually deleted, and later restored, but still showed at CAT:CSD until I removed "cat handler" this morning. JohnCD (talk) 10:43, 10 September 2012 (UTC)

I suspect that the IP typed {{subst:db-g5}}, so the article received a copy of the {{db-g5}} template. The template has two big pieces, one to generate the message box, and one to deal with the categories. -- John of Reading (talk) 10:58, 10 September 2012 (UTC)
That would explain it. I'm surprised it doesn't happen more often - one tends to type "subst" automatically. It's something I will look for when I find an article in CAT:CSD that doesn't seem to have a db template. Thanks, JohnCD (talk) 12:52, 10 September 2012 (UTC)

Userbox table fix

I'm trying to make the userboxes on my userpage split into two rows so it doesn't require horizontal scrolling. Can you help me with a formatting tweak to the below code which would split the userboxes up? (nowiki's aren't there on my talk page).
<TD>{{User:West.andrew.g/STiki UserBox 1}}</TD>
<TD>{{Userbox |border-c=#000 |border-s=1 |id-c=#FFDEAD |id-s=12 |id-fc=#FFF |info-c=#A9A9A9 |info-s=8 |info-fc=#FFE4E1 |id=[[File:Aidlogo.png|55px]] |info=This user is a regular member of [irc:// #wikipedia-en-help] on [[WP:IRC|IRC]].}}</TD>
<TD>{{User wikipedia/OTRSAccess}}</TD>
<TD>{{userbox|#CCCCFF|#F8F8FF|[[File:Handshake icon.svg|45px]]|This user is a member of '''[[Wikipedia:WikiProject Cooperation |Wikiproject Cooperation]]'''}}</TD>
<TD>{{User:Henrik/ubx/User contrib|20,000|Ocaasi}}</TD>
<TD>{{User WP GLAMPMA}}</TD>

Thanks! Ocaasi t | c 18:41, 10 September 2012 (UTC)

Just insert the red line. Edokter (talk) — 19:06, 10 September 2012 (UTC)
Awesome. Thanks! Ocaasi t | c 19:52, 10 September 2012 (UTC)

When should Foreign character warning boxes be used?

Sorry if I'm in the wrong place—there are so many of these templates I couldn't find a single talk page which seemed proper for this discussion. I'm posting here since these the original purpose of these templates was to serve as technical warnings.

What is the proper use of {{Special characters}} and its derivatives, i.e. the templates in Category:Foreign character warning boxes? It seems reasonable that they shouldn't be included in articles where the foreign script in question appears only once in the first sentence, and they would be useful for articles discussing the language or its literature. But what about articles which mention several (read: less than a dozen) foreign names, but where the display of those names aren't essential to the understanding of the article itself (since the names are also represented in Latin script)? Their appearance is consistently aesthetically unappealing, whether placed below infoboxes or next to lead images. And there are even editors who apparently think that these boxes serve to list all the scripts which appear in an article. I know there are readers whose operating systems don't come pre-installed with non-Latin text support, but are these templates really still necessary? The context in which these foreign scripts appear should usually be enough for the reader to understand what they're supposed to show, whether or not it renders correctly. -- (talk) 05:36, 11 September 2012 (UTC)

Magic words question

I'm trying to make a userbox show the name of the user if the userbox is in the user OR User talk namespace, but just show the word editor if it's outside of those namespaces.

Here's the closest I came, but it didn't work (note, I have no idea what I'm doing):

{{#if: {{NAMESPACE}}=User talk | {{PAGENAME}} | editor}}

Can you help? Thanks!!!! Ocaasi t | c 01:00, 7 September 2012 (UTC)

You're pretty close. What you want is {{#ifeq:}} or {{#switch:}}. Desired implementation would either be:
{{#ifeq:{{{NAMESPACE}}}|User|{{{PAGENAME}}}|{{#ifeq:{{{NAMESPACE}}}|User talk|{{{PAGENAME}}}|editor}}}}
{{#switch:{{{NAMESPACE}}}|User|User talk={{{PAGENAME}}}|#default=editor}}
--Izno (talk) 02:04, 7 September 2012 (UTC)
Thanks! Will try it, very cool :) Ocaasi t | c 03:20, 7 September 2012 (UTC)
As you might have discovered, magic words take two brackets, not three: :
{{#switch:{{NAMESPACE}}|User|User talk={{PAGENAME}}|#default=editor}}
-DePiep (talk) 12:58, 8 September 2012 (UTC)
Oh, thank you! Yes, that's why it wasn't working! Ocaasi t | c 10:19, 11 September 2012 (UTC)

Why we have files like Category:Wikipedia image placeholders for image namespace?

Is it better than salting bad filenames? Bulwersator (talk) 15:02, 8 September 2012 (UTC)

  • And why current solution is worse than salting - files from this category pollute Category:Public domain files ineligible for copyright etc, it is rather hack than proper solution, Wikipedia:File Upload Wizard silently fails on attempt to overwrite protected file and we show files with terrible jpg artifacts as part of interface. Bulwersator (talk) 15:09, 8 September 2012 (UTC)
  • Possibly relevant to this discussion. - The Bushranger One ping only 21:12, 8 September 2012 (UTC)
  • Most likely simple inertia. The first of these placeholders was I think File:Pleasedont.gif, uploaded August 31, 2005; I don't remember when we got creation protection, but it was well after cascading protection, which was used throughout 2007 for this purpose.

    And yes, the file upload wizard deals poorly with attempts to overwrite these images, and Special:Upload is even worse—it shows neither the protection reason from the log nor the current image. Does a failed upload to a creation-protected file look any better? I don't know any to test with. (talk) 03:30, 9 September 2012 (UTC)

    • Yes, it is way better - try to upload something to File:Robotlove.gif - you will receive nice explanation instead of "Uploading in progress" sitting for hours Bulwersator (talk) 07:45, 11 September 2012 (UTC)
  • With regard to Special:Upload, "it shows neither the protection reason from the log nor the current image". Do you mean it doesn't show before you press the upload button ? What does it say AFTER pressing the upload button ? If I have the time, I'll try and fix at least Special:Upload in MediaWiki core to deal with this more elegantly. I recently already made a change to have the Commons UploadWizard deal with this problem. —TheDJ (talkcontribs) 09:03, 11 September 2012 (UTC)
    • Special:Upload - Before - nothing, after using upload button: "This image name or media file name is protected.

When uploading files to Wikipedia, please use a file name that describes the content of the image or media file you're uploading and is sufficiently distinctive that no-one else is likely to pick the same name by accident.

Examples of good file names: (...)" (tested on File:1.jpg) Bulwersator (talk) 09:21, 11 September 2012 (UTC)

      • OK, i have filed bugzilla:40161 with regards of checking for protection status of destination title and other related title checks in Special:Upload. —TheDJ (talkcontribs) 11:21, 11 September 2012 (UTC)
I agree that it is impractical to duplicate the images. It fills up Category:Public domain files ineligible for copyright with lots of almost identical copies, so it is harder to search that category. In my opinion, a better option to duplicating files would be to keep only one copy of the image and use redirects to that file instead. --Stefan2 (talk) 10:43, 11 September 2012 (UTC)
  • These could all be protected redirects instead of duplicate uploads. That prevents the individual files being categorized. That also makes prettification later on easy (e.g. the JPEG artifact issue for instance, to "fix" that a lot of uploads need to occur, whereas with a redirect only one). To ensure these are protected at all times something we can use {{R protected}} (or something like it) to auto-categorize unprotected ones. This method is fairly common on Wikimedia Commons (though many duplicates exist there as well, still). Krinkle (talk) 11:30, 11 September 2012 (UTC)
(Interesting how non-conflicting edits are sometimes merged when the edit window stays open a long time. Stefan2 already said this in the mean time :) Krinkle (talk) 14:17, 11 September 2012 (UTC))
  • Redirects are better than current solution but why not use proper tool to block creating of pages - salting? Bulwersator (talk) 14:34, 11 September 2012 (UTC)