Wikipedia:Village pump (technical)/Archive 121

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


Template:Clickable button is broken

The clickable button templates, {{clickable button}} and {{clickable button 2}}, among others, have been broken for a few weeks. I assume this was due to a css or js change or the like. Compare the working buttons on Commons: commons:Template:Clickable_button. Anyone know how to fix these?--Sage Ross (WMF) (talk) 15:44, 22 November 2013 (UTC)

I assume you're referring to the lack of "button" styling? Interestingly enough, [1] makes the buttons there render correctly, but they are unstyled when not in edit mode. Chris857 (talk) 15:49, 22 November 2013 (UTC)
Yes. And I just noticed the same thing, that they show up correctly in previews.--Sage Ross (WMF) (talk) 16:07, 22 November 2013 (UTC)
They currently work for me in all tested situations. There is a previous discussion at Wikipedia:Village pump (technical)/Archive 120#Dark blue over the word "wikicode" to the point where I can hardly see it. PrimeHunter (talk) 16:41, 22 November 2013 (UTC)
jQuery.ui is no longer loaded by default. See Wikipedia:Village pump (technical)/Archive 119#jQuery UI CSS may no longer load by default and linked mailing list post. Anomie 16:50, 22 November 2013 (UTC)
Thanks, that's helpful. Is there a fix, or are we just stuck waiting until the switchover is complete? (Or do we just need to deprecate these buttons?)--Sage Ross (WMF) (talk) 20:01, 22 November 2013 (UTC)
Well we can fix them. But it adds again to the stack of javascript. So I guess the question is: Why do we need them with this styling ? —TheDJ (talkcontribs) 20:41, 22 November 2013 (UTC)
Personally, I'd think having transclusion counts of 869 and 76,859 respectively would be enough of a reason to make sure these templates are working and continue to be supported, no? Technical 13 (talk) 21:01, 22 November 2013 (UTC)
Why use JavaScript? It can be done with CSS alone. CSS is generally less of a burden for browsers then JS, and works for people who have JS disabled. CSS pseudo-classes can be used in MediaWiki:Common.css, which allow for different formatting when the user hovers over the button (:hover), clicks on the button (:active), or gives the button keyboard focus (:focus). – PartTimeGnome (talk | contribs) 23:17, 22 November 2013 (UTC)
Well sort of my point. They no longer work, because jquery-ui is no longer needed by the core on every single page (and getting rid of unneeded dependencies is a good thing). So to bring them back, you either need to load all that CSS again (either all the time, or with JS only when there are buttons on the pages), or you decouple them from jquery-ui and implement some 'light' CSS on en.wp that doesn't have so much overhead and provides similar functionality. I'm not sure where these buttons are used, which contexts, which users, so it's a bit tough to call for me. —TheDJ (talkcontribs) 18:50, 23 November 2013 (UTC)
  • I am unsure why the two templates aren't merged into one easy to use with advanced features template. I really don't care how they are supported and work, and even as a TE I don't have access to modify MediaWiki: namespace stuff, especially Common.* (which I would expect that even if MediaWiki: was added to something the could be edited by TE's in general would be fully protected anyways). So, what I'm thinking is, if you can get consensus to make {{Clickable button}} (with or without {{Clickable button 2}} merged into it) work with css in the MediaWiki:Common.css then go for it. Technical 13 (talk) 02:35, 23 November 2013 (UTC)
  • I noticed that the buttons look fine on User:TheJJJunk, but not on User:GoingBatty. I'm using FF25. GoingBatty (talk) 02:35, 25 November 2013 (UTC)
@GoingBatty: They look fine to me on both our pages. Maybe this got fixed by the time I saw this. — TheJJJunk (say hello) 15:03, 25 November 2013 (UTC)
It works on both of those pages for me as well, but not on many others (for example, {{welcome student}}). It just depends on whether something else on the page causes the required code to be loaded, I guess.--Sage Ross (WMF) (talk) 15:15, 25 November 2013 (UTC)
Hmm, the button on that page appears correctly for me as well. Even before it was switched to clicakablebutton 2. I tried adding a button on my sandbox, and that also displays correctly. I must be special. Or I have a user preference set or a script that activates it? — TheJJJunk (say hello) 15:37, 25 November 2013 (UTC)
I'm guessing what people seem to be experiencing where some buttons seem to work for some but not others is due to caching issues (at least we are going to blame it on that because it makes such a good scapegoat). Try BYPASSing and see if they still don't seem right. Technical 13 (talk) 16:28, 25 November 2013 (UTC)
I've bypassed and cleared my caches on FF25 and IE10, and User:GoingBatty still doesn't show clickable buttons for me. :-( GoingBatty (talk) 01:42, 26 November 2013 (UTC)
I've tried another computer with IE8, and I don't see the clickable buttons whether I'm logged in to WP or not. GoingBatty (talk) 14:17, 26 November 2013 (UTC)
It seems much more likely to me that the reason people see different results, is because they have different gadgets/scripts enabled, which might be loading the code that the clickable button requires. —TheDJ (talkcontribs) 15:12, 26 November 2013 (UTC)
The buttons work for me, so I logged out, and they didn't work. When logged in, I commented out all my imported scripts and blanked my .css page, and the buttons still worked. Gadgets then? — TheJJJunk (say hello) 15:22, 26 November 2013 (UTC)
@TheJJJunk: Do you mind sharing what gadgets you have enabled? Thanks! GoingBatty (talk) 16:02, 29 November 2013 (UTC)

These are all of the options I have enabled in the Gadgets section of my Preferences:

  • After rolling back an edit, automatically open the contributions of the user whose edit has been rolled back
  • Open external links in a new tab/window
  • Twinkle
  • "Ask a question" feature for the Wikimedia Foundation's "Teahouse" project
  • Reference Tooltips
  • Citation expander
  • HotCat, not checked in the gadgets, but imported through a script
  • Form for filing disputes at the dispute resolution noticeboard
  • CharInsert
  • Add an [edit] link for the lead section of a page
  • Add a "Sandbox" link to the personal toolbar area.
  • Display diffs with the old yellow/green colors and design.

If you want to go through the scripts I have included as well, I have some on my common.js and vector.js I think. — TheJJJunk (say hello) 17:51, 29 November 2013 (UTC)

Aha - enabling Twinkle fixes the clickable button, and disabling breaks it again. Anyway to have the button work without using Twinkle? I also noticed that with Twinkle disabled, the buttons don't appear when I'm reading User:GoingBatty, but they do appear in edit preview mode. Thanks! GoingBatty (talk) 00:03, 2 December 2013 (UTC)
As Anomie said way up, jquery.ui is no longer loaded by default, and the button relies on some of its styles.
Unsurprisingly activating a gadget that requires jquery.ui (like Twinkle) will make the styles available, and I assume the editing toolbar also uses jquery.ui which explains why it happens to work there.
Amalthea 10:13, 2 December 2013 (UTC)

Since we weren't getting anywhere, i've proposed to add the most efficient loader that I could come up with to fix this problem. —TheDJ (talkcontribs) 13:12, 2 December 2013 (UTC)

I have a potential solution here. Kaldari (talk) 19:20, 3 December 2013 (UTC)
Thanks to both of you! Today I see the buttons just fine on User talk:ChrisGualtieri, but still don't see them on User:GoingBatty. GoingBatty (talk) 03:13, 7 December 2013 (UTC)

Template links as parser function

Just a random idea, but would it be worthwhile to replace the {{tl}} template with a more "built-in" method of linking to templates? (i.e. a parser function or something? {{tl:ambox}} instead of {{ambox}}?) Would it even be worth it? ViperSnake151  Talk  17:17, 4 December 2013 (UTC)

I cannot phantom any reason why that would make a significant impact. —TheDJ (talkcontribs) 21:58, 4 December 2013 (UTC)
The built-in method is a normal wikilink with namespace: Template:Tl. I see no reason for anything else. PrimeHunter (talk) 23:02, 4 December 2013 (UTC)
Yeah, the {{tl}} templates are just wikilinks with some nifty css styling (templates are still linked like normal wikipages). Theopolisme (talk) 03:14, 5 December 2013 (UTC)
  • No, there's no way it's worth it. Right now, we have over a dozen tl* templates that do all sorts of linking, formatting, with or without parameters, and there are probably more that will be built in the future. It would be insane if we had to have a mediawiki update in order to add new tl* functions. This isn't even anything that would be worth it to try to do with lua - they're basically about as simple as templates get. VanIsaacWS Vexcontribs 03:31, 5 December 2013 (UTC)
  • Template:tl runs 800x/sec, #ifeq at 750x, Lua 500x: Comparing the runtime of {{tl}}, as 800x per second, to the #ifeq parser-function at 750x/second, then the operations seem about equal, where {tl} must format the template-name wikilink and that involves a check for existing page to set link color blue/red, etc. Also, {tl} has remained static, as a rapid template, not burdened with 15 layers of /core subtemplates dragging /core2 subtemplates. A tiny template, inserting one word, runs at 2,400x/second. Formerly, the Lua Scribunto interface ran tiny Lua scripts at only 300x per second, but we complained and the developers quickened Scribunto in April 2013 to 500x-600x Lua per second. Some people have imagined that complex Lua script never gets slow, but runtime tests have confirmed how checking 100 parameter names and error-check formatting those Lua parameters does, as could be expected, slow the operation of Lua, where simple checks of just a dozen parameters might run 3x faster. Hence, the current template {tl}, as 800x per second, is running perhaps 2x faster than Lua could offer. -Wikid77 19:30, 6 December 2013 (UTC)

Table sorting problems

Using IE11/Win7. I see various glitches with table sorting at List of elements. For example:

1. Load page. Click repeatedly on the sort button for the "Abundance" column (far right). Some clicks appear to do nothing, others appear to sort the table by atomic number (ascending or descending) instead.

2. Load page. Click each of the sort buttons in turn, from left to right, ("Z", "Sym", "Element". etc.). When I get to "Atomic weight" the sort ceases to function (on that or any other column).

Is this a browser bug or something in Wikipedia? (talk) 18:20, 6 December 2013 (UTC)

Hey, thanks for finding this. I think I fixed it with this edit. There are 13 columns, but there were only 12 sort keys (don't know if that's what they're called). There was also a random test edit that I reverted, which didn't seem related. So, your browser and Wikipedia are fine, just a mistake in the table. Chris857 (talk) 20:17, 6 December 2013 (UTC)
Hi Chris, all seems to be working properly now, thanks very much for sorting it out. (talk) 01:30, 7 December 2013 (UTC)

adding of categories facilitated on article creation

Is it possible for persons creating articles to be presented with a dropdown box of the categories at the time of submission. On choosing the desired category, the person can save the article. Thus work involved in looking for articles that are not categorized is eliminated?

Thank you

Anupama Srinivas (talk) 18:21, 6 December 2013 (UTC)

I oppose. Special:UncategorizedPages is made automatically so it isn't much work to find them. There are tens or hundreds of thousands of categories for articles so a dropdown box of "the categories" is not practical. And most articles should be in several categories. If we force new users to select one category then they would often make a poor choice and it's harder to find the problem articles to add better categories. It would also require a software change. The developers have better things to work on, and many new users would be frustrated when they are unable to save. And should it apply to moving drafts into mainspace? Drafts should not have categories intended for articles. PrimeHunter (talk) 20:41, 6 December 2013 (UTC)

Template trouble

The answer to this is probably really obvious but apparently my brain isn't working properly. Why is the template on Dead to Fall (and related) pointing to User:Jax 0677/Template:Dead to Fall rather than Template:Dead to Fall as it should, and how can I fix it? Thanks--Jac16888 Talk 19:16, 6 December 2013 (UTC)

Bingo, thanks--Jac16888 Talk 19:45, 6 December 2013 (UTC)

Global Wiki search?

Is there a way to run a global search across all (or multiple wikis) looking for different language versions of the same article or results from the same search term? Thanks.-Godot13 (talk) 19:44, 7 December 2013 (UTC)

@Godot13: - Have you tried a Google search - for example, search for "godot" GoingBatty (talk) 20:00, 7 December 2013 (UTC)
I haven't tried but I will. Thanks-Godot13 (talk) 02:20, 8 December 2013 (UTC)
Also include language codes in search:  Mozart es OR de OR fr OR el, because Google Search gives English-language webpages top rank, and adding the codes "fr" (French) or "de" (German deutsch) or "es" (Spanish español ) will cause those other-language webpages to list sooner among English pages. For "el" (Greek Elleniki ), the rank might be so low as to require a separate search with Greek, but a topic like "Mozart" will appear in several pages for each language. -Wikid77 21:00, 8 December 2013 (UTC)
So, this might actually be possible in the not-too-distant-future. Right now, our current search setup is such that wikis don't really know other wikis exist (other than Commons). With the new search engine we're working on, all of the wikis search data will be in the same cluster and cross-wiki queries might well be possible. I'm already looking at using them for Commons search as well as some rudimentary interwiki search.
While a "search all the wikis" would be very cool it also has the ability to be resource-intensive, so we'll have to tread carefully. It's definitely something I've thought about very recently though :) ^demon[omg plz] 01:43, 8 December 2013 (UTC)
I imagine the potential drain on the system for such searching could be huge. It would be interesting to see how efforts toward such a search function progress. Thanks.-Godot13 (talk) 02:20, 8 December 2013 (UTC)

There is a tool called Global Wikipedia Article Search at --Snaevar (talk) 03:08, 8 December 2013 (UTC)

Characters overlapping on my watchlist

This just started happening tonight. Anyone know why this overlap is happening? Dismas|(talk) 04:04, 8 December 2013 (UTC)

I don't know, but I've seen this off-and-on for months? now. (I have Chrome, XP, if it matters). Chris857 (talk) 04:41, 8 December 2013 (UTC)
It's a Chrome bug that is sometimes triggered by left-to-right markers. Edokter (talk) — 09:43, 8 December 2013 (UTC)
It doesn't seem to have anything to do with the left-to-right markers, it's that Chrome has issues with <span dir="auto"> and multiple whitespace characters. I've just checked this in Chromium 31.0.1650.57 and reduced it to a test case of just
(<span dir="auto">foo    bar</span>)
(foo bar)
and still saw the issue. See also Wikipedia:Village pump (technical)/Archive 100#Anomalous Edit Summary and bug 39912 Anomie 13:12, 8 December 2013 (UTC)
I filed a ticket in webkit for this not so long ago. —TheDJ (talkcontribs) 13:25, 8 December 2013 (UTC)
I wasn't aware of the other bug report. That one seems fixed in webkit/blink now, so quite possibly were are dealing with a related but subtly different problem with this case. —TheDJ (talkcontribs) 13:43, 8 December 2013 (UTC)
"The other bug report" being webkit bug 94832? After looking that over and the comments on the linked fix commit, it seems to me it's only related in that they both involve dir="auto". Anomie 14:18, 8 December 2013 (UTC)

I'm seeing this in Safari, not Chrome. Dismas|(talk) 02:48, 9 December 2013 (UTC)

That doesn't surprise me: both are WebKit browsers. --Redrose64 (talk) 10:06, 9 December 2013 (UTC)

Disappearing revert type edits and possible mis-identification as "can be undone"

I've encountered cases when reverting an edit which the screen says "can be undone" and where it apparently accepts the revert but the edit never shows up. From the particulars, I'm guessing that it mistakenly identified an edit as "can be undone" when in fact it can't, (which I guess is a second problem) and then it acts as if it took the edit but doesn't. I've this happen 3-4 times in the last few weeks. An example is to try to revert the 3:07 4 December edit at Punk rock. Sincerely, North8000 (talk) 14:16, 8 December 2013 (UTC)

(BTW I don't still wish to make that revert)North8000 (talk) 14:22, 8 December 2013 (UTC)
The mentioned edit [2] has already been reverted in [3]. If you undo the edit again and save then you make a null edit which isn't logged. Do you have en example where undo should make a change but doesn't? PrimeHunter (talk) 14:40, 8 December 2013 (UTC)
(edit conflict) When I bring up the "undo" diff for that edit, it shows no actual changes in the diff, making it in effect a null edit which doesn't show up in the history. The undo logic should probably detect this situation and display a more appropriate message. Anomie 14:44, 8 December 2013 (UTC)

@Primehunter, I don't remember where the other ones are. They might have been the same situation (reverting an already reverted item). Maybe the best thing would be to allow null edits and have them show up. This might help another situation which comes up. When I make a significant error in an edit summary, at first I tried null edits to enter a second correction to the edit summary. Then I noticed that they don't "take". Now someone has to make a small change in the article in order to enter an edit summary correction. North8000 (talk) 20:48, 8 December 2013 (UTC)

Block quote formatting

Take a look at the bottom half of this. Note that the attribution line spacing is very wide, I think maybe even more than the actual quote. Where do I report this for fixing? Maury Markowitz (talk) 17:11, 8 December 2013 (UTC)

You should be more specific - I do not see anything special on that page. Ruslik_Zero 18:38, 8 December 2013 (UTC)

"We don't have fighter aircraft, we have fear-enducing aircraft"

French President Charles de Gaulle, speaking of the Mirage IV in 1963.[1]

"We don't have fighter aircraft, we have fear-enducing aircraft"

French President Charles de Gaulle, speaking of the Mirage IV in 1963.[2](<small>...</small> removed here)
  1. ^ Sokolski 2004, p. 95.
  2. ^ Sokolski 2004, p. 95.
I guess it refers to the {{Quote box}} to the right. The line spacing is also large for me. It looks a little better when <small>...</small> is removed from the source parameter, but the line spacing is still larger than in the quote. PrimeHunter (talk) 19:10, 8 December 2013 (UTC)
Common problem. When you reduce the font-size in a multi-line container, line-height is not automatically adjusted with it, but in some cases remains inherited from its parent. If you reduce the font size, you have to reset the line height to 1.5em as well. BUT... I note the font-size is too small anyway. Edokter (talk) — 19:29, 8 December 2013 (UTC)
(edit conflict) The line heights are all set to "1.5em", but that depends on the font size in effect on the element where the 1.5em is calculated. The quote box has an inline style to reduce the font size to 88%, but the line-height is inherited from the page text (in my browser, it inherits 19.05px where 1.5em would now be 16.7833px). Due to the layout of the wikitext produced by the template, the quote itself gets wrapped in a <p> which has a stylesheet rule resetting the line-height (so that does get the 16.7833px), while the source text doesn't get the <p> and therefore is still inheriting the larger line-height. The <small> just makes it worse, because the line-height is still being inherited. I wonder why the line-heights are done as ems instead of plain numbers which have more sensible inheritance behavior. Anomie 19:47, 8 December 2013 (UTC)
@Anomie: That's bug 2013 tl;dr Because someone made a bad decision in 2005 and it's too much boring work and annoyance to change this now (as plain search&replace introduces minor interface changes). (There were three patches for this submitted over the years, none made it.) Matma Rex talk 20:01, 8 December 2013 (UTC)
It sounds like the problem with the most likely of those patches was basically "What if this changes how things look? Please prove to me that it won't." "I don't care enough to do that much work." Too bad. Anomie 21:01, 8 December 2013 (UTC)
@Anomie: I never looked at that one, but I did look at gerrit:55126 which was almost exactly the same and it did change layout in unpleasant ways. Compare the "before" and "after" screenshots I added there: before after. Matma Rex talk 21:09, 8 December 2013 (UTC)
What, more space around the headings? That seems easy enough for someone to fix. Anomie 21:33, 8 December 2013 (UTC)
@Anomie: Well, you're free to try :) Personally I have better things to do than to figure out the multipliers, and then manage the community here disgruntled because we moved a link by five pixels. Matma Rex talk 22:40, 8 December 2013 (UTC)
I was looking at that and just can't remember why we used ems when we updated quote box. This should be discussed at the template talk. --  Gadget850 talk 20:00, 8 December 2013 (UTC)
A site-wide fix would be easy: [p,] div, span { line-height: 1.5em; } Edokter (talk) — 20:47, 8 December 2013 (UTC)
A better fix would be along the lines of gerrit:27043. Anomie 21:01, 8 December 2013 (UTC)
That's quite extensive. Simply adding <div> and <span> to the <p> already in commonElements.css is more effective. All the other lineheights apply to specific elements anyway, with minimal risk of embedding. Edokter (talk) — 21:28, 8 December 2013 (UTC)
In retrospect, my solution also has unwanted side effects. Edokter (talk) — 21:34, 8 December 2013 (UTC)
Edokter, what does the selector [p,] div, span pick up? I know what div, span matches: any <div>...</div> element and any <span>...</span> element. I also know that e.g. div[att] matches the <div att=value>...</div> element. But what does the initial [p,] do in your example? --Redrose64 (talk) 21:43, 8 December 2013 (UTC)
Nothing really; it is already defined in commonElements.css with line-height: 1.5em;. That's why I put it in [brackets]. Edokter (talk) — 22:29, 8 December 2013 (UTC)
Still puzzled, so I want to see where this [p,] is being used, but I've drawn a blank: for a start, what is commonElements.css? MediaWiki:commonElements.css is a redlink. --Redrose64 (talk) 23:23, 8 December 2013 (UTC)
It's in MediaWiki core, specifically this folder. — This, that and the other (talk) 06:32, 9 December 2013 (UTC)

Another RFPP Twinkle issue

I requested "raise protection" on one move-protected page, but it went to "lower protection section instead. Same for other such requests. --George Ho (talk) 01:32, 9 December 2013 (UTC)

The best place to post problems with Twinkle is at WT:TW. I'll take a look at this issue; it occurs because Twinkle is getting confused by the move protection already applied to the page. — This, that and the other (talk) 06:28, 9 December 2013 (UTC)
Also please note that some (all?) of the section headings at WP:RFPP were recently altered, which broke the incoming links except when somebody came along and did this. --Redrose64 (talk) 09:32, 9 December 2013 (UTC)

Reference problem

It used to be that you could click on a "^" next to a reference and it would direct you back up to where the citation is from in the text. Those carets have disappeared, at least for me. Was there a change recently? --Coemgenus (talk) 22:00, 24 November 2013 (UTC)

You didn't give an example so I don't know whether you are looking at an article using another reference system, but I see "^" at for example Thomas F. Bayard#References. If a reference is used more than once then "^" is not a link. Instead there are letters linking to each use. PrimeHunter (talk) 22:20, 24 November 2013 (UTC)
OK, it must be a problem on my end somehow, because that article isn't working for me. I held CTRL and hit reload and they appeared for a second, then vanished. Using Chrome browser. --Coemgenus (talk) 22:23, 24 November 2013 (UTC)
All right, it came back when I disabled the "Near this page" beta feature. That's a bug, I guess. Is there someplace I should report it? --Coemgenus (talk) 22:29, 24 November 2013 (UTC)
Bugs should be reported at mw:Talk:Beta Features/Nearby Pages. -- John Broughton (♫♫) 22:42, 24 November 2013 (UTC)
According to mw:Talk:Beta Features/Nearby Pages#Disrupts footnote linking from reflist to article footnotes there is already a fix in gerrit:97143 but I guess it hasn't been deployed yet. PrimeHunter (talk) 00:28, 25 November 2013 (UTC)

Links next to citations in reflist

I'm guessing that this problem is related to "Reference problem" up above, but that section is too vague for me to be sure. As long as I can remember, the reflist has had a ^ next to each number, which when clicked takes the reader to the spot where that citation is referenced; or when <ref name=whatever> and <ref name=whatever /> are used, we get a series of letters (a, b, etc.), each of which will take the reader to the reference spot when clicked. Within this month (seemingly in the last week or two) these have disappeared without warning or explanation; any idea what's going on? It's plainly a skin problem: these letters are missing when I view pages in Monobook with Firefox 25.0.1 and with IE 8, but when I use Vector, both browsers display the links properly. Nyttend (talk) 01:45, 30 November 2013 (UTC)

Have you enabled "Near this page" at Special:Preferences#mw-prefsection-betafeatures? If so, does it help to disable it? PrimeHunter (talk) 02:23, 30 November 2013 (UTC)
I've never checked the beta features, given the mess they made of removing the orange bar. I was completely unaware that the "Near this page" even existed, so no, I've not enabled it. Nyttend (talk) 17:28, 30 November 2013 (UTC)
I forgot to specify that for the same period of time I've been having what otherwise would seem to be an unrelated problem. Whenever I go to a category with subcategories, I've traditionally been able to find the sub-subcategories by clicking the little arrow next to the name of the subcategory. However, for the last week or two, clicking the little arrow always produces a message of "Problem loading data. Please wait a moment and try again." The second sentence is a link that effectively refreshes the category page. Has anyone else witnessed such a problem? Nyttend (talk) 17:55, 30 November 2013 (UTC)
What happens if you log out and view and In Firefox 25.0.1 I see ^ in both MonoBook and Vector, both logged in and out. PrimeHunter (talk) 20:35, 30 November 2013 (UTC)
I see everything properly when logged out, and when I view them in Vector while logged in, things are just as wrong as when I'm in Monobook. I didn't feel like switching skins just for this, so I'd been logging out before in order to use Vector; I never knew about the ?useskin thing. Time to check my preferences, I guess. Thanks for diagnosing it! Nyttend (talk) 14:11, 1 December 2013 (UTC)

Reference backlinks

Reference backlinks are missing:

Markup Renders as
This is page content.<ref>LibreOffice For Starters, First Edition, Flexible Minds, Manchester, 2002, p. 18</ref>

This is page content.[1]

  1. ^ LibreOffice For Starters, First Edition, Flexible Minds, Manchester, 2002, p. 18

--  Gadget850 talk 01:17, 6 December 2013 (UTC)

I see the backlink in your example and in articles but I guess you have the issue in #Reference problem. PrimeHunter (talk) 01:50, 6 December 2013 (UTC)
I've been struck by this too. Not sure when because I haven't been editing much recently. I did have the new near-by links feature enabled but I turned it off and the problem remains. don't see it if I am logged out. A purge of my cache and restarts of the browser didn't fix the issue. Firefox. Jason Quinn (talk) 06:13, 11 December 2013 (UTC)

Toolserver seriously funked up

Replication lag is now at over three weeks and counting. Any assistance with this problem would be much appreciated. DavidLeighEllis (talk) 01:29, 27 November 2013 (UTC)

WMF+WMDE have chosen to axe the Toolserver, so most people have given up and don't care. You can use Labs: [4]. Legoktm (talk) 06:26, 27 November 2013 (UTC)
Thank you. DavidLeighEllis (talk) 15:06, 27 November 2013 (UTC)
Would it be possible to add a notice to the old version, so people know?--S Philbrick(Talk) 17:41, 27 November 2013 (UTC)
Actually, before tool server dies, we intend to turn it into a redirect page to let everyone know to change their links to point to labs.—cyberpower OnlineMerry Christmas 14:04, 1 December 2013 (UTC)
And has anyone else noticed that X!'s editcount tool on labs hasn't been accessible for a couple of days? Jared Preston (talk) 16:42, 5 December 2013 (UTC)
I haven't been able to access it recently either... DavidLeighEllis (talk) 21:51, 5 December 2013 (UTC)
The counter seems to be up and running now. DavidLeighEllis (talk) 23:19, 6 December 2013 (UTC)

Toolserver is indeed seriously funked up. For example, for November it is reporting about 550,000 views for many articles such as Filter feeder, Bait fish, Striped bass and Menhaden. Is it true that WMF+WMDE are axing it and offering no transition strategy? --Epipelagic (talk) 02:51, 11 December 2013 (UTC)

I don't think uses the toolserver, and the transition strategy is Tool Labs. Anomie 14:02, 11 December 2013 (UTC)
We've known on this page for a long time that Toolserver was having problems, see for example this thread that I kicked off. The WMF are not under any obligation to offer a transition strategy, because Toolserver wasn't theirs. It belongs to WMDE who kept asking for money to keep it going - but they didn't get enough. But we have had plenty of notice - since about January 2012 IIRC (Jarry1250 may know precisely) - that Toolserver wasn't going to last forever, so, in a manner of speaking, it's our fault for (a) not donating to WMDE and (b) not migrating the tools to Labs - or at least, somewhere other than Toolserver. --Redrose64 (talk) 14:25, 11 December 2013 (UTC)

/UserProfileIntro subpages

Example of my profile page

As you can see at this "everything" search of WP ([5]) certain new users are creating their new user pages in a subpage of their user at /UserProfileIntro. I've randomly checked a few of the pages and they were all created with the following summary: (Updating user profile introduction) (Tag: Mobile edit). November 22, 2013 is the oldest user account where this has happened, and there have been scattered instances since. I did a little Google searching and found the phrase UserProfileIntro in a script that was being developed for the mobile users. I submitted a bug report (now closed/deleted) at Bugzilla and was left a stern message that said "Closing this bug - this feature works as intended. If you have problems with it, please open a more specific bug." I'm fairly certain that new users shouldn't be directed to create their user page in a subdirectory, and that whatever is causing this is not right.

Am I nuts, or is this really a problem. If it is, can someone notify the proper people about it. I failed in that area. Thanks. --| Uncle Milty | talk | 02:30, 3 December 2013 (UTC)

Apparently the mobile people are "reinventing" user pages just for mobile. How strange. Anomie 03:20, 3 December 2013 (UTC)
Yep, that looks odd. Most of the User pages are non-existent, but there is content at the User:XXXX/UserProfileIntro page that would typically appear on the User page. Is there someone reading this page who can explain how/why this behavior counts as "works as intended"? Alternatively, is there someone here who has enough pull on Bugzilla to open this as a bug and not have it closed summarily? It doesn't seem consistent with how things have typically been done around here, but maybe there is some grand plan of which we have not been made aware. – Jonesey95 (talk) 05:46, 3 December 2013 (UTC)
Jdlrobson, do you know about this? — This, that and the other (talk) 10:54, 3 December 2013 (UTC)
This is a controlled experiment in beta to explore ways that we can in future encourage people to edit their user pages via smaller micro contributions. For the time being we want to see if people would fill these in and what kind of things they would say. We did not want to pollute user pages with this so we created it as a sub page. Let us know if you want to mark them in a way to make your lives easier or to make it clearer this is an experiment. Jdlrobson (talk) 21:48, 4 December 2013 (UTC)
I don't see anything wrong with all that. I imagine the main problems are 1) It was not clearly advertised. This is fine if it's just a small experiment but it does mean people may be confused when they come across it. 2) No real explaination was offered before now (or if it was it doesn't seem to have been clearly linked previously). The bug report just says this feature works as intended but it can't be assumed people would understand the reason why it's creating the pages as subpages. While I understand bugzilla is intended to track bug and not engage with the community, a simple explaination like you offered above, either when closing the bug report or via a link when closing the bug report would probably have stopped this thread being started. Maybe it would have been better to approach you directly but I can understand with the bug closure combined with the fact the editor just didn't see any reason why this would make sense, that they may have felt asking the community was best. Nil Einne (talk) 20:52, 9 December 2013 (UTC)
I was wondering how long it would take people to notice this :D I don't see a problem with this, we can do whatever we want with it at any time. Let them play, measure and find the best solutions. —TheDJ (talkcontribs) 11:02, 3 December 2013 (UTC)
I added an example for those are wondering where this is used. —TheDJ (talkcontribs) 12:28, 3 December 2013 (UTC)
  • Uncle Milty, it would be useful to have the closed tickets ID number. Then the ticket can simply be re-opened with some added detail instead of creating a "new" ticket that would simply get closed as a duplicate. :) Technical 13 (talk) 12:56, 3 December 2013 (UTC)
Sorry, I thought it was gone, but it was filtered out of my search. ID # 57748. --| Uncle Milty | talk | 13:47, 3 December 2013 (UTC)
I noticed bugzilla:57931 while reading this thread. Yay for bugs! Legoktm (talk) 16:38, 3 December 2013 (UTC)

Script Error in Templates

I was not sure where to post this, so I think the ANI is the best place. I noticed that Template:M1 year in topic suddenly has a script error. Judging by the templates it includes, the error seems, based on timing, to be the result of edits on Template:Navbar and Module:Navbar by Edokter (linked so she/he is aware, not because the report is about them). The Navbar pages are fully protected so I cannot edit them myself to see if they are indeed the source of the error. EvergreenFir (talk) 03:58, 4 December 2013 (UTC)

One of the greatest weaknesses of Lua modules, and something that Wikidata has to deal with all the time in its project namespace, is that if you have too many lua modules on one page, they stop loading after a while and you get that message. That might be what it is, but as I don't understand templates very well, it might be something else entirely. Sven Manguard Wha? 04:43, 4 December 2013 (UTC)
Lua's worst problem has been the miserly 10-second timeout, which then stores "Script error" for each Lua function in the rendered page. Some have imagined the Lua clock ran like traditional CPU time, but numerous tests confirmed it can double (or triple) in slow servers (a 4-second Lua slows 2.5x as 10-second and craters). -Wikid77 00:34, 7 December 2013 (UTC)
@EvergreenFir: This should be at WP:VPT. If you have scripting enabled in your browser (JavaScript, nothing to do with modules), when a "script error" occurs, you can click the red error message to see a popup box with details. You can select that text and copy it, to be pasted into a report at WP:VPT.
@Sven Manguard: Ouch, I try to follow WP:VPT and have not seen any mention of a problem with modules (apart from the total runtime of 10 seconds per page, which is plenty). Is there a discussion about that somewhere? Has Anomie commented? Johnuniq (talk) 05:46, 4 December 2013 (UTC)
Oh no, Lua 10-second timeout is way low, should be 20-25 seconds. Timing has confirmed slow servers can double/triple Lua clock and "safe" 4-second run slows to wannabe 12-seconds (aborted at 10-sec. limit). -Wikid77 00:34, 7 December 2013 (UTC)

OK, I should have clicked the first link above (Template:M1 year in topic). I had assumed that it was the usage of that template somewhere that had a problem, but it is actually on the template page. The error text is:

Lua error: Cannot pass circular reference to PHP.
Backtrace: (none)

A little digging shows the problem occurs when "{{#invoke:Year in other calendars|main}}" is previewed in a sandbox. Mr. Stradivarius may be able to throw some light on the matter, which is possibly in Module:Year in other calendars or one of the modules it calls. This section should be closed, with further discussion elsewhere. Johnuniq (talk) 06:14, 4 December 2013 (UTC)

My first thought is that it looks complicated, and that it might be a server-side change. It's not a change in Module:Year in other calendars, as that hasn't been edited recently, and it was working as late as 27 November when I made this comment. It might be a change in one of the modules it calls, which I'll investigate now. It is quite a hefty module, though (as was Template:Year in other calendars/old before it), so Sven Manguard may be on to something with the "too many modules on the page" comment. More in a second. — Mr. Stradivarius ♪ talk ♪ 07:03, 4 December 2013 (UTC)
It looks like the problems are coming from Module:Japanese calendar, but neither that or the modules it calls have been edited recently, so I think it must have been a change server-side that has broken things. I'm not sure yet about any possible fix. — Mr. Stradivarius ♪ talk ♪ 07:29, 4 December 2013 (UTC)
Also, the module calls that are causing the script errors from #invoke work fine when used from the debug console. It appears to be only module invocations using #invoke that are affected. — Mr. Stradivarius ♪ talk ♪ 07:37, 4 December 2013 (UTC)

This appears to be due to the fix for bug 48393, which was deployed yesterday. Module:Year in other calendars is returning various objects in its 'p' table. Removing those should fix the problem. If access to those objects is necessary for some other module, the p table could contain an accessor function which would return them. BJorsch (WMF) (talk) 14:15, 4 December 2013 (UTC)

Yep, that's the one, and I've fixed it now. Thanks! It was confusing because I had done the same thing in both Module:Year in other calendars and Module:Japanese calendar. I got rid of the extra objects in 'p' for the former, and used an accessor function for the latter. By the way, where can I subscribe so that I know that things like this are coming? :) — Mr. Stradivarius ♪ talk ♪ 14:47, 4 December 2013 (UTC)
That's a very good question. Unfortunately, there isn't any non-developer-oriented place for change notifications. I suppose the closest thing is [6] and [7] (if only there were a way to filter out the "Localisation updates" entries) if you want to watch changes that have been merged, and [8] will show you changes that are currently proposed, and [9] shows open bug reports and feature requests. BJorsch (WMF) (talk) 18:33, 4 December 2013 (UTC)
  • Raise Lua 10-second timeout: Meanwhile, keep pressing to raise the pitiful Lua timeout limit to 25 seconds. Many Lua timing tests have confirmed how slow servers can double/triple the Lua clock and a "safe" 4-second run slows to wannabe 12-seconds (aborted at 10-sec. limit). Should be 25-second limit, because unlike "wp:Wikimedia Foundation error" (WFE), Lua timeout stores "Script error" in the rendered page (forever) with no 70-second grace period to extend the formatting and save good results 7 minutes later. Again the Lua timeout is not a mere smaller version of the 60+10-second parser limit, and instead a Lua timeout scars the rendered page with "Script error" for every aborted Lua #invoke run after the 10-second cutoff. The Lua 10-second timeout is like narrowing a 7-foot (2-metre) road shoulder as 7x smaller along a mountain-side road at the cliff's edge (all because the Lua tires are better, except on gravel!). -Wikid77 00:34, 7 December 2013 (UTC)
    Actually, this particular script error didn't have anything to do with the 10-second time limit. It was because I was returning non-function values to PHP from the Lua modules in question, and Scribunto had changed from allowing this behaviour to prohibiting it. — Mr. Stradivarius ♪ talk ♪ 04:04, 12 December 2013 (UTC)

Wikicode or Extension for collecting comments/answers and users voting and ranking the most helpful?

I am developing a course in Wikiversity and I would like two tools to take advantage of peer learning, as in MOOCs. I wonder if there are already ways to implement these in wikis or whether we would need to develop new extensions.

  1. I want to present a question prompt with a textarea input box for the user to input an answer. Then after submitting, the user would see the prompt, the course-recommended answer, and a page of all submitted answers (which have been stored with automatically added user's signature --~~~~).
  2. I want for users to be able to click vote up or down comments/answers, so that a page will display postings in order by rank of up/down votes.

Is there a way to approach either of these functionalities with currently existing wiki tools or extension? I know we have some sort of polling tool for administration. Is it adaptable to these needs? If there isn't an existing way, is there a group I should contact that would be the logical place to propose developing these tools?

Note: In part this is a rethinking of an earlier question I posted in the Help desk here --Kenmayer (talk) 14:33, 4 December 2013 (UTC)

There might be existing extensions for upvotes/downvotes/likes/+1, but I don't know whether the devs/ops would be willing to install it on Wikiversity. Try searching on It would be possible to write a quick JS hack to do something like this (assuming the user has JavaScript enabled), but it would log the username of every voter. Not sure if you want that or not. You might also be able to write a complex template or Lua module to do something like this, but I don't think it would be very good for the students. πr2 (tc) 04:48, 6 December 2013 (UTC)
  • Careful editors could vote in a preformatted, sortable wikitable: I have noted how wp:edit-conflicts during a 20-user edit session can be prevented by separating each user response by a separator line, such as table-row token "|-" between each response stored in the page. Ideally, each of 20-30 students would have voter-id codes (V1-V30) pre-stored as "line id codes" in the empty voting table, and then each student would respond by updating the row which has their voter-id code, putting remarks in the empty row, but as a sortable table. Example:
  {| class="wikitable sortable" <--also set columns for sorting
  ! Id || Issue || Vote || Time || Remarks
  | V1 || issue #1 || vote || ~~ ~ ~~ || (remarks)
  | V1 || issue #2 || vote || ~~ ~ ~~ || (remarks)
  | V2 || issue #1 || vote || ~~ ~ ~~ || (remarks)
  | V2 || issue #2 || vote || ~~ ~ ~~ || (remarks)
  | V3 || issue #1 || vote || ~~ ~ ~~ || (remarks)
  | V3 || issue #2 || vote || ~~ ~ ~~ || (remarks)
Explain to each student to update the rows for their voter-id under the "vote" column to "yes" or "no" (or whatever choices), and connect the 5 tildes ("~~ ~ ~~") together, as the timestamp next to their vote. If the voting all occurred during the same month, then the time-of-day could be sorted by timestamp column, hours/days "0:01, 1-31 December" etc.
Also each voter could change their prior votes, if they felt misled or mistaken during an earlier response. Of course there is always the chance for glitches during editing of the wikitable format, but if various students were asked to "moderate" and check the table format, from time-to-time, then the wikitable of votes would be keep functional, and best of all: almost no edit-conflicts with 20-30 users all updating the same page within a period of minutes or hours. Also, any hostile comments could be redacted in such a system, although the "peer pressure" of all students meanwhile editing the same table would tend to "tone down" excessive remarks during voting. To reduce name-related problems, students could be asked to avoid discussing voter-ids, as semi-anonymous editing (although ids could be deduced), and privately discuss when someone else is editing under their voter-id. Due to the "magic" of the "|-" separator lines between rows, then all 20-30 students could edit the table at the same time, or announce "I'm saving edit now" to allow other students to wait 30 seconds before the next attempted Save-page operation. Even with the current primitive level of edit-conflict merging, such a voter-wikitable could be used to collect votes, allow re-votes, and sort the issues/votes into date-time order. Once sorted, then the total votes, for each related issue, would be easy to count. -Wikid77 (talk) 23:26, 6 December 2013 (UTC)

How soon do you need this? I believe that something like this will eventually be possible via WP:Flow, but I doubt that these sorts of advanced workflows would be before the end of next year (or the year after that, even). WhatamIdoing (talk) 01:15, 7 December 2013 (UTC)

So will Flow have upvotes? Will it keep track of each user's karma? I don't see why that would be needed. This is an encyclopedia, not a social network. Oh, wait, are you talking about creating a custom workflow that would be equivalent to likes/dislikes? Are there any examples of how this would work? πr2 (tc) 04:34, 7 December 2013 (UTC)
Yes, I'm thinking about some sort of custom workflow. Flow is basically a bunch of building blocks. They're supposed to add a "building block" that makes a plain vote possible (straight-up voting is far more popular at other projects, but even the English Wikipedia uses votes for a small number of things, like RFAs). So it seems to me that a custom workflow along these lines might be possible (but not any time soon, and someone would have to figure out how to set it up). WhatamIdoing (talk) 22:18, 9 December 2013 (UTC)

What links here doesn't link

According to this list, a lot of articles link to BB&T Ballpark. I had to link to similar article earlier this week and discovered the need to make that a disambiguation page, but after moving the article that was there, I found a bunch of articles where even if I clicked on edit and tried CTRL-F, I couldn't find the words "BB&T Ballpark". I know I fixed a template, but a bot has tagged the article now. I reverted the tag and reported the situation to the bot's talk page, because I'm just not finding the links.— Vchimpanzee · talk · contributions · 19:53, 6 December 2013 (UTC)

If you update a link in a template, as here, the links at Special:WhatLinksHere/BB&T Ballpark aren't updated immediately; you normally need to wait for the job queue to process them. --Redrose64 (talk) 20:00, 6 December 2013 (UTC)
I figured I had given it time.— Vchimpanzee · talk · contributions · 20:34, 6 December 2013 (UTC)
The job queue has been ropey since they reconfigured it about six months ago. It's two days since your edit - but it can take more than two weeks in some cases. --Redrose64 (talk) 21:16, 6 December 2013 (UTC)
Ive noticed the same thing a few times. Surely that isn't normally what it should be like.Blethering Scot 21:29, 6 December 2013 (UTC)
  • Pages seem to change in groups of 500 but some get stuck: (new essay: wp:Reformat) Even weeks after a change to a template has "completed" the relinking process, there might be several pages still stuck with old links, and could be cleared with null-edits. For years, it's as though a database read/write lock among several pages had caused a "relink conflict" and the pages could not be relinked when whichever wp:job_queue processed the reformatting job (wp:JQ says Wikipedia runs 3 queues at once). Also, according to wp:JQ, when more than 1,000 pages have the "invalidated cache" to trigger reformat, they are grouped as 500-pages-per-job, and checking WhatLinksHere tends to show progress as sporadic groups by hundreds of pages, either linked or delinked, in "fits and starts" where even the whole process can get delayed during the multi-day ordeal. Yet, always be prepared to null-edit the leftover stragglers when all jobs have finished relinking within the 3(?) job_queues. It would be great if a database write-lock timeout would re-queue each stuck page, to be reformatted as another job, re-added to a queue, but perhaps that requeue decision could not be handled down at the level of relinking each template used inside a page. With the Lua megatemplates, there are even more cases where changing a Lua megamodule would trigger formatting a "million" pages which formerly were reformatted as smaller groups of separate (near-redundant) fork templates, such as the 24 wp:CS1 cite-template forks now forced to all reformat 2.2 million pages for any Lua edit to the giant Module:Citation/CS1, or the new Module:Convert which would combine all the various 20,000-page, separate Convert subtemplates to always trigger reformatting all 554,000 pages for any Lua edit made to Module:Convert. I worry Lua has become used in a "pennywise and pound-foolish" architecture, running 6x-13x times faster than markup, but triggering 100x-1000x times as many page reformats as the prior smaller template forks, with /core subtemplates. Many Lua modules should use forced multi-fork, split designs (as multiple Lua modules) to trigger less system-wide reformatting of all pages. -Wikid77 15:06, 8 December 2013 (UTC)
However it happened, it's fixed now.— Vchimpanzee · talk · contributions · 19:55, 9 December 2013 (UTC)

VisualEditor template data filled by Google Code-in students?

(Edited my posts after a friendly chat with Equazcion--no info lost)

Wikimedia is participating in Google Code-in, where 13-17 years old students have completed already more than hundred technical tasks, including coding, documentation, QA, UX, and promotion. Students are completing tasks almost faster than we can create them. We need more!

If you, like me, have added VisualEditor template data to templates, you will probably agree that it becomes a boring task after the first time. Then again it is a very useful work. For GCI students, a task like this will be instructive. For, this is a chance to get work done and new contributors with a potential to stick around. Note that this work would benefit not only, but also all Wikimedia projects watching and using copies/adaptations of these templates. What do you think? It would be great if someone would step in, become a Google Code-in mentor, and coordinate this little project until January 6, when GCI ends. If you want to try before committing, just point me to 5 templates welcoming VisualEditor template data. I will create a task, and share here the news when the updated templates are ready for review. Then you can decide what comes next. Deal? PS: the last GCI task completed is a rewrite of Template:Interval in Lua.--Qgil (talk) 06:20, 8 December 2013 (UTC)

A tangential exchange was moved to User talk:Equazcion#VP Visual Editor exchange. equazcion 07:24, 8 Dec 2013 (UTC)

There are quite a few templates listed at Wikipedia:VisualEditor/TemplateData#Commonly_used_templates. Perhaps some of the infoboxes? Theopolisme (talk) 17:51, 8 December 2013 (UTC)
That list might benefit from being re-run. {{nts}} wasn't on it, and I get a transclusion count of over 16,000 for that template. Also, many of the non-English Wikipedias are much, much further behind with TemplateData work. Some templates are exactly identical, without even translating the parameter names, so it might be efficient to create TemplateData for everything all at once. Whatamidoing (WMF) (talk) 22:24, 9 December 2013 (UTC)
Thank you for the feedback and the advice. I have created five tasks and one of them has been already claimed. Let's see how it goes. Let's continue this discussion at Wikipedia_talk:VisualEditor/TemplateData. PS: Theopolisme happens to be one of these very smart and productive GCI students. Hi! --Qgil (talk) 00:34, 10 December 2013 (UTC)

advanced search needed

I am not the first one who ask for a really advanced search. e.g. here, here, and here. There should be an advanced search similar to Google or Ebay advanced search. i.e A friendlier usage of the existing search parameters. is it possible? Ykantor (talk) 07:56, 9 December 2013 (UTC)

At present, we're using Lucene, via MWSearch. I believe that CirrusSearch is being tested as a possible replacement. --Redrose64 (talk) 09:45, 9 December 2013 (UTC)
Confirming what Redrose64 wrote. If "advanced" search is wanted, it would be good to start by defining the problem that should get solved. --AKlapper (WMF) (talk) 11:19, 9 December 2013 (UTC)
Exactly, "advanced search" says nothing, unless you define what options you consider to be advanced. There are some details about the current query capabilities of Cirrus search. If we can start by making a list of things that are not possible but desired and which of those options you would want to be exposed trough an advanced search page, then that would be very helpful. —TheDJ (talkcontribs) 13:00, 9 December 2013 (UTC)
I guess it is better to have some improvement, rather than asking for the ultimate search, that requires a lot of work and thus risking getting nothing.

As said, let us start with the existing features (e.g intitle, prefix) but arrange it in a friendlier advanced search, similar to Ebay or Google one. e.g. the user may fill the intitle in a dedicated input box, even if he does not know what is the parameter inline name. Ykantor (talk) 14:17, 9 December 2013 (UTC)

As noted below, now there is Help:Advanced_search, but we have found that advanced features with non-advanced users is a rare mix, so it is better to focus on "power users" as more likely to use advanced features. In fact, in Template:Convert, which tracked each option used for 6 years, the many advanced options were used in less than 2% of pages, while 98% used basic options, despite extensive documentation plus top notes about advanced options. -Wikid77 18:21, 9 December 2013 (UTC)
  • Help:Searching and Help:Advanced_search: Currently, the page Help:Searching is written as a full introduction, and so a new page, Help:Advanced_search could be written for experienced users who want a direct, condensed list of the advanced-search features and syntax format, beyond the general wp:wikisearch. For example, a lead tilde, "~phrase" will ensure a full search, rather than showing one page named "phrase" (which has been very frustrating, because for years, WP has alternated between always show a full search or show one page as titled). I will try to write that advanced description (in my spare microseconds this month!). -Wikid77 (talk) 17:24, 9 December 2013 (UTC)
Thank you for this important tip (the ~) and for writing advanced description. Still, a friendly advanced (e.g. Ebay advanced search) should help a lot to utilize the already existing features). Ykantor (talk) 19:17, 9 December 2013 (UTC)
Definitely not something we're looking at for the first iteration of CirrusSearch (right now we're focusing on feature parity with lsearchd, but more stable & modern), but something we've talked about in the medium term. Being able to facet search results like this would be fantastic :) ^demon[omg plz] 04:07, 10 December 2013 (UTC)
Say that 1% of Wikipedia editors are using the search advanced features, but with a friendlier interface, 3% will use advanced search. Friendliness is more important than power. Ykantor (talk) 20:36, 10 December 2013 (UTC)

Tech News: 2013-49

16:10, 9 December 2013 (UTC)

Though there may be plans to deliver this newsletter by MediaWiki message delivery in the future, this one was delivered using EdwardsBot like the previous ones. [17] – PartTimeGnome (talk | contribs) 22:30, 9 December 2013 (UTC)
@PartTimeGnome: FYI, it was delivered to some pages using MassMessage. See m:User talk:Guillom#Tech-news_wasn.27t_delivered_to_Community_pages (permalink, relevant bug) for more info on this. Tech news should be delivered by MassMessage in the future (at least if the bug is fixed and a patch is merged before next week). πr2 (tc) 20:27, 10 December 2013 (UTC)

Help with Google news search

Google's news archive search recently changed to only show news results that are less than 30 days old. If any of the technical wizards here at VPT know about Google or their news search (or any other news search engine), could you please hop over to Wikipedia:Village pump (miscellaneous)/Archive 44#Google news archive search? and provide your input? We are hoping to find a technical solution to this problem. Thanks very much for your help. (talk) 03:08, 10 December 2013 (UTC)

It's been answered over there, but I might as well repeat it here. A Google representative says they are changing their news archive search and the access from Google News has gone away and will be a few months. Meanwhile, however, they explain on this page how you can access at least some old newspaper articles. (I find that you need to have Javascript enabled for the date selection to work.) -- (talk) 07:30, 20 December 2013 (UTC)

Vanishing Edits on Watchlist

Clicking on my watchlist, my most recent edit does not appear. The same happens when I click contributions. My edit (made around 22:49 EST) does not appear. Anyone know why? -- Veggies (talk) 04:33, 10 December 2013 (UTC)

Okay, now I see the "changes longer than 3,100+ seconds ago won't appear" tag. -- Veggies (talk) 04:35, 10 December 2013 (UTC)

I noticed this too and came to see if anyone had posted about it. EvergreenFir (talk) 04:36, 10 December 2013 (UTC)

Edit metadata

Of possible interest to you, people involved with this large wiki and technical things: Grants:IdeaLab/Edit_metadata. Please comment at the talk page or at the bottom (the talk page is transcluded). Thanks! Smiley.svg
Gryllida (talk) 09:10, 10 December 2013 (UTC)

Formatting problem with 2-column content

Viewing Template:POTD/2013-12-15 using Firefox 22.0 on Linux, I see essentially no space between the columns in the 2-column section, so one of the lines looks like "5. Madreporite11. Ambulacral". Actually there seems to be a slight amount of negative space, with the "e" and "1" slightly overlapping. Varying the font size in Firefox using control-plus/control-minus, this is still true at all sizes I tried; only the amount of the overlap changes a bit.

Is this because (1) the wikitext markup calling for two columns was done wrongly, (2) the HTML being generated from that wikitext is wrong, or (3) Firefox is rendering the HTML wrongly? And if it's 1 or 2, could someone please fix it?

I wonder if the cause is the fact that the sub-column (so to speak) with the item numbering gets wider as you go down the second column.

-- (talk) 20:17, 10 December 2013 (UTC)

Playing around with it (though not actually saving anything), raising the "style=width:70%" to at least 80% makes the columns better. However, getting rid of columns altogether increases overall height very minimally and eliminates all the unfortunate line breaks. Chris857 (talk) 20:34, 10 December 2013 (UTC)
Fixed size for columns is always problematic. When the POTD goes full width, the columns will be too wide. I gave the column width 12em, which is comfortable for each entry to have no linebreaks. This results in one column on the template page, but three columns on the main page (at 1280px wide screen). Edokter (talk) — 23:09, 10 December 2013 (UTC)
Thanks, Dok, that's much better. -- (talk) 04:34, 11 December 2013 (UTC)

Faulty redirect

When ever I click a link in Wikipedia that's intended to bring me to a specific section of another Wikipedia page, it instead brings me a few lines below that section and when I first go to that page the title of that section is just barely cut off from the top of the part of the page currently on the screen. Blackbombchu (talk) 21:11, 10 December 2013 (UTC)

It's the responsibility of your browser to take you to the right place. If a page has collapsing/expanding elements then it can confuse the browser if an element changes size after the browser has chosen where to place you. Do you have such an element at top of all pages? Clicking in the browser address bar and pressing enter may cause your browser to place you again, this time accurately if there are no longer elements changing size. PrimeHunter (talk) 21:26, 10 December 2013 (UTC)
That problem happened many times the most recent of which was when I clicked prose in the notification at the top of the article Protest. I made no change to the browser then I tried clicking the exact same link in the exact same article later today and this time, the link actually brought me to the title of the section instead of a few lines below it, in the same browser even. The window never changed size nor did any ad ons get added to the browser shrinking the browser screen when the link brought me a few lines below the title of the section. Blackbombchu (talk) 23:17, 10 December 2013 (UTC)
It also varies for me how accurate the browser placement is when I clicked a link on another page. As mentioned, it's your browser doing it. When you click a link like HTML anchor#Anchor, Wikipedia serves you the page HTML anchor with "#Anchor" at the end of the url: Your browser then looks at the page source for the id "Anchor", in this case <span class="mw-headline" id="Anchor">Anchor</span>, and automatically scrolls the page to that position. PrimeHunter (talk) 23:58, 10 December 2013 (UTC)

Bulk change of domain in external URLs

Greetings. Some quick searching revealed little in terms of an explicit tool to do this, but maybe this is something AWB or other semi-automated tool specializes in...

I probably have added 50-100+ external URLs to policy and discussion spaces that link back to a personal domain, where I host my academic writings and datasets relevant to Wikipedia and collaborative security. This domain has now changed, and while there is an HTTP redirect in place, administrative policies dictate that will not survive forever. The file paths are still the same on the new domain, but the domains do need to be changed. This is a touch painstaking manually. Is there a way to automate this? If so, is that solution limited to en.wp or is this something that can be done for all WMF properties (I know I have links on Wikimania wikis and Metawiki, at minimum)? Having plenty of experience coding with the Wikimedia API I know this would not be a terribly difficult thing to code up, but probably not worth the effort for my one off use. Thanks, West.andrew.g (talk) 01:30, 11 December 2013 (UTC)

@West.andrew.g: Sounds like WP:Bot requests is what you're looking for. I suggest you post all the details there, and hopefully some kind bot owner can help you out. Good luck! GoingBatty (talk) 02:35, 11 December 2013 (UTC)

Hiding the stuff above my watchlist

Watchlist screenshot.jpg

I'd love a way to hide it. It takes up half my watchlist page. Suggestions? Anna Frodesiak (talk) 07:07, 11 December 2013 (UTC)

Anna Frodesiak, I have the following in my .js file to remove the "Mark all pages visited" button in your screen shot:
Bgwhite (talk) 07:48, 11 December 2013 (UTC)
After doing some experimenting, add the following code to your .css file to remove the big "watchlist options" box:
     { display: none; }
See User:Bgwhite/vector.css for other space saving options. Bgwhite (talk) 08:04, 11 December 2013 (UTC)
Yes!! I was looking for something just like that for a while.. Thank you! :) -- œ 08:20, 11 December 2013 (UTC)
Oh, thank you soooo much. This is wonderful. Could someone please add it for me. I have no idea how to do that or where. Also, I see at User:Bgwhite/vector.css a thing to remove the logo and move the left side options up to fill that space. That would be wonderful too. Thank you very, very much! :) Anna Frodesiak (talk) 10:06, 11 December 2013 (UTC)
By the way, I use monobook. I hope that's okay. Anna Frodesiak (talk) 10:07, 11 December 2013 (UTC)
There is no easy method for hiding the line "You have 0 pages on your watchlist (excluding talk pages)." without compromising other portions of the page. The line "Pages that have been changed since you last visited them are shown with a green bullet." may be hidden using
span#mw-wlheader-showupdated { display: none; }
The button Mark all pages visited may be hidden using
form#mw-watchlist-resetbutton { display: none; }
The box "Watchlist options" may be hidden using
form#mw-watchlist-form { display: none; }
All of these would be placed in Special:MyPage/common.css: it's not necessary to use javascript for any of this. --Redrose64 (talk) 12:41, 11 December 2013 (UTC)
Yay! It worked. I made User:Anna Frodesiak/common.css and pasted those three lines and it worked. I have no idea how, but it's great. Now, I only see "You have xxxxx pages on your watchlist (excluding talk pages)." I'm fine with that. I prefer to have that line, actually.
Thank you all. I'm very grateful. :) Anna Frodesiak (talk) 13:16, 11 December 2013 (UTC)
There is also mw:Snippets/Collapsible ChangesList options. Helder 15:41, 11 December 2013 (UTC)
Thank you. :) Anna Frodesiak (talk) 21:15, 11 December 2013 (UTC)

Getting error message when trying random page in Education Program namespace

Clicking on Special:Random/Education_Program, after about a minute and a half I get this message:

A database query error has occurred. This may indicate a bug in the software.
   Function: RandomPage::selectRandomPageFromDB
   Error: 0 

Does this happen for anyone else? -- œ 11:18, 11 December 2013 (UTC)

Yes. -- John of Reading (talk) 11:29, 11 December 2013 (UTC)
This is probably related to bugzilla:58157. --Redrose64 (talk) 11:39, 11 December 2013 (UTC)
I'm not so sure. I've filed it separately as bugzilla:58324, so we'll see... — This, that and the other (talk) 11:41, 11 December 2013 (UTC)

Collapse ate the page's table of contents

I collapsed the first item (a transclusion of a GA review) on the talk page atTalk:Saengerfest and then the table of contents for the overall page went inside the "collapse". As a workaround I moved it elsewhere on the page which fixed it. Is this a "bug" or did I do something wrong? North8000 (talk) 13:25, 11 December 2013 (UTC)

Not a bug. By default the TOC is placed before the first section heading. That was a section heading in Talk:Saengerfest/GA1 so in [18] the TOC was collapsed together with the whole transcluded page. Workarounds include to remove headings from the transcluded page or place __TOC__ where you want the TOC. This overrides the default placement. PrimeHunter (talk) 13:54, 11 December 2013 (UTC)
Thanks! North8000 (talk) 14:32, 11 December 2013 (UTC)
Once a GAN is over, it shouldn't be transcluded to the talk page, but linked. the {{article history}} template is used for that, as here. --Redrose64 (talk) 14:47, 11 December 2013 (UTC)
Right now that isn't in the instructions for reviewers. What happens is that it stays transcluded for a while and then I think a bot moves it. I think that that is normally a good thing.....a GA review being visible talk page content for a while just like any other talk page content. This was a special rare case where the reviewer left mid process without passing or failing the article and a second review was started. North8000 (talk) 18:33, 11 December 2013 (UTC)

Edit buttons at talk:Engineering have disappeared

Any ideas? Please see also this discussion. Thank you. Δρ.Κ. λόγοςπράξις 06:34, 12 December 2013 (UTC)

Fixed -- John of Reading (talk) 07:42, 12 December 2013 (UTC)

Wikipedia:CSS and Javascript hacks

I don't think this is relevant to anything we have currently so should this be archived {{historical}}? TeleComNasSprVen (talkcontribs) 07:34, 12 December 2013 (UTC)

Interesting. I'm sure it could be theoretically added to at this point in time, but that content should be on Meta or MediaWikiwiki. Feel free to mark as historical. Be bold about marking things as historical if there's been no activity in years but looks like it should be kept for posterity's sake. Killiondude (talk) 07:39, 12 December 2013 (UTC)
Well it looks there's no useful information that is salvageable save the edit history because everything seems to have been already documented on Help:CSS and WP:JS etc. I was thinking of the {{historical}} tag mainly because of the history alone, but I guess it could also be transwiki'ed and then deleted (or soft-redirected). TeleComNasSprVen (talkcontribs) 08:12, 12 December 2013 (UTC)
I would be inclined to MFD it outright. There is absolutely nothing worth keeping on that page. — This, that and the other (talk) 08:52, 12 December 2013 (UTC)

Hidden comment constantly being removed

I have placed a hidden comment onto an article using the <!-- and --> tags to deter deleterious edits to the page but it seems that whenever an IP editor edits the page the comment disappears. Is this a consequence of visual editor and is there an alternative I can use so the text remains in the section of the article where it has been placed?—Ryulong (琉竜) 03:03, 3 December 2013 (UTC)

Considering that VE isn't available to IPs at this time and the edits aren't tagged with the VE tag, it seems unlikely that VE has anything to do with it. More likely you've run into an IP editor who is deleting your comment. Anomie 03:24, 3 December 2013 (UTC)
Well shit. How do I keep it in there then?—Ryulong (琉竜) 17:52, 3 December 2013 (UTC)
Semi-prot the page? --Redrose64 (talk) 17:56, 3 December 2013 (UTC)
PC1 protection maybe? Technical 13 (talk) 18:57, 3 December 2013 (UTC)
An WP:Edit notice for the page is another possibility, but have you considered whether adding this is truly necessary and appropriate? "How do I keep it in there" sounds like an WP:OWNership issue. Have you considered whether the IP might actually be improving the page by removing a message that might seem unfriendly? WhatamIdoing (talk) 19:24, 3 December 2013 (UTC)
I had put it in there because editors had been expanding the antagonists' entries to be summaries of the entire TV series from the point of view of the antagonists, which is still somewhat a problem on the article as evident here.—Ryulong (琉竜) 09:39, 7 December 2013 (UTC)
Editnotice created. --Redrose64 (talk) 15:43, 7 December 2013 (UTC)

Enhanced recent changes

I'm getting really annoyed. Enhanced recent changes isn't allowing me to expand grouped up edits (it was allowing it intermittently before, especially if I WP:BYPASSed my cache) except in really, really, rare occasions. I'm wondering if it has anything to do with removing jQuery from loading on all pages by default? Any one else experiencing this (I've tested on multiple machines with various OSes and browsers and two out of three of my accounts with the same results). Any suggestions on what may be causing it? Technical 13 (talk) 22:48, 5 December 2013 (UTC)

I personally don't have enough time to investigate this, but I'm sure you could give more details. :) Please add ?debug=true or &debug=true to the query string and check the JS error console. P.S. AFAIK, jQuery is still loaded by default, but jQuery UI is not anymore. πr2 (tc) 04:42, 6 December 2013 (UTC)
Adding ?debug=true makes enhanced recent changes work, so there is nothing appearing in the error console when I do that. The error console without ?debug=true looks like:
I hope all of this helps you figure it out PiRSquared17. Technical 13 (talk) 02:25, 7 December 2013 (UTC)
O_O. That's a lot more than I expected. Order of magnitudes more. I can't duplicated this (works for me), so could you provide your browser and skin? I'm wondering if you get a JS error when you actually click the expand links. πr2 (tc) 03:39, 7 December 2013 (UTC)
PiRSquared17, I've done some more testing to see if I can narrow it down and isolate it, and when I click the expand arrows/links and they don't expand I get nothing int the error console (I load the page and clear the log and nothing new appears). With ?debug=true set and the sections expand, I get the following JavaScript warning, but that is it: [11:54:13.836] ReferenceError: reference to undefined property queue[0] @ [19] It's a real nuisance having to add ?debug=true each time to make it work. I'll do some more digging because if no-one else is having an issue, it must be in one of the userscripts I am using or made. Technical 13 (talk) 17:03, 7 December 2013 (UTC)

View count stats

RESOLVED: fixed. -Wikid77 12:18, 12 December 2013 (UTC)

I've just noticed that the viewcounting tool appears to have stopped working. For example, Cyclone Bodil hasn't counted the views on the page for the past three days. Is there a problem? The C of E God Save the Queen! (talk) 08:29, 9 December 2013 (UTC)

It's an external tool. Problems can be reported to the operator at User talk:Henrik, but he rarely answers. PrimeHunter (talk) 13:09, 9 December 2013 (UTC)
  • Data was logged but not shown in reports: During past delays in pageview stats, the problem has been fixed within 3-5 days. There are alternative reports (also missing recent days): (monthly barchart) (list as date/number)
However, people have noted the pageview-data has been logged for the missing days, so once the reports are unstuck, then all days should appear as expected. -Wikid77 (talk) 16:33, 9 December 2013 (UTC)

Serious problem with page

I emailed a screenshot to User:Nil Einne. This library has Windows 8 and I'm not sure what version of Internet Explorer.— Vchimpanzee · talk · contributions · 18:54, 9 December 2013 (UTC)

me stares at screen with his confused face. I'm quite sure there is a context to this story, but I think you failed to add it to this post Vchimpanzee. :D —TheDJ (talkcontribs) 19:25, 9 December 2013 (UTC)
Vchimpanzee seems to have random problems with various pages every so often on the library computers they use. Some of these may be caused by misconfiguration on behalf of the library, but many seem to be just random problems not worth worrying about. They generally ask for help identifying the cause of each of these problems which is what they're doing here. I presume the problem they're having is with a wikipedia page hence why they're asking here rather than a on the RD. I'm not totally sure of the wisdom of necessity of identifying each of these errors, many of which as I've said seem to be random stuff which probably isn't going to reoccur but I guess that's up to them and those offering help. Nil Einne (talk) 19:37, 9 December 2013 (UTC)
I haven't received anything yet. While I don't particularly mind receiving this stuff, I should say I can't necessarily always offer help and by sending it to me you're ensuring no one else can use it to help. If it doesn't have identifying or other info you don't want to reveal, may I suggest one of the many external image hosting resources instead like [[20]]? Nil Einne (talk) 19:37, 9 December 2013 (UTC)
I sent it again but the one I sent is in my sent folder, although the email address was being extremely uncooperative. I've never seen anything like this on Wikipedia.— Vchimpanzee · talk · contributions · 19:50, 9 December 2013 (UTC)

You can also share an image using a service like imagebin for instance. —TheDJ (talkcontribs) 19:54, 9 December 2013 (UTC)

Thanks. Here it is.— Vchimpanzee · talk · contributions · 19:58, 9 December 2013 (UTC)
Sorry, we'll just have to hope Nil Einne gets it. I really don't know what else to do.— Vchimpanzee · talk · contributions · 20:01, 9 December 2013 (UTC)
I asked for help again and was told FTP is not allowed. I was told to send the image in an email, which I have done. Twice.— Vchimpanzee · talk · contributions · 20:28, 9 December 2013 (UTC)
I don't really know what this discussion is about, but if we're talking about a screenshot of Wikipedia, you can upload it to Wikimedia Commons. This would be preferable to using external sites. – PartTimeGnome (talk | contribs) 22:25, 9 December 2013 (UTC)
At this library we can't even see the screenshot.— Vchimpanzee · talk · contributions · 23:08, 9 December 2013 (UTC)
I'm sorry but unless Nil Einne responds I don't know what else to do. I'm not going back to the library where I was when it first happened until next Monday. I apparently can't do anything at this one. If I'm at home (I will be tomorrow) I've got no one to help me with this.— Vchimpanzee · talk · contributions · 23:11, 9 December 2013 (UTC)
You've been provided ideas already here: "if we're talking about a screenshot of Wikipedia, you can upload it to Wikimedia Commons". So this really does not depend on Nil Einne's response. --AKlapper (WMF) (talk) 15:16, 10 December 2013 (UTC)

────────────────────────────────────────────────────────────────────────────────────────────────────Providing me with an idea does not get it done. You can tell me how to do it until the cows come home but if I can't do it I can't do it.— Vchimpanzee · talk · contributions · 16:14, 10 December 2013 (UTC)

I found someone who could help me today, and the problem was that this computer did not have Word. We got on another computer that did, but we still couldn't do anything with the image. She said that computer didn't have the software. She suggested emailing Nil Einne again. I showed her the above directions and she didn't seem to know what to do. All I know is there is a Word document with the screenshot. That is what the woman at the other library created.— Vchimpanzee · talk · contributions · 17:21, 10 December 2013 (UTC)
What does Microsoft Word have to do with uploading a screenshot? If you have the png/tiff/jpg/gif/bitmap, you can just upload it to Commons, but maybe you need to crop out the logo, unless it is necessary to understand the problem. (Edit: never mind, just noticed that the link you provided to download the image is actually a docx. Unfortunately, I cannot open docx files.) πr2 (tc) 17:52, 10 December 2013 (UTC)
I really don't know why the woman at the library where this happened chose Word. Anyway, I'm at home and have gone through the process and on my computer screen is a file called "Wd0000001.wps", using Microsoft Works Word Processor. If anyone can get me past this point, I'd appreciate it.
A hint as to the nature of the problem: It actually says "This page has some issues".— Vchimpanzee · talk · contributions · 19:29, 10 December 2013 (UTC)
Just ignore MS Word, MS Works or any other word processing software: images are not text documents. You should use a program which handles images directly; fortunately, all versions of Microsoft Windows released over the last twenty years (or more) include a simple drawing program called "Paint" (other names have been used, such as "Paintbrush"). That can be used to create a .PNG file; if you're using Windows XP, it should be in Start → Programs → Accessories. Here's a condensed version of Wikipedia:Screenshots of Wikipedia:
  1. Make sure that the problem is visible on screen
  2. Press Print Screen
  3. Start Microsoft Paint
  4. Press Ctrl+V
  5. Save this file, making sure that you select "PNG (*.PNG)" at the "Save as type" dropdown
  6. Upload the file. --Redrose64 (talk) 20:02, 10 December 2013 (UTC)

────────────────────────────────────────────────────────────────────────────────────────────────────My power went out, so I've been trying to get everything back as it was. Also, I am trying to do other things. Okay, I'm ready to go.— Vchimpanzee · talk · contributions · 20:36, 10 December 2013 (UTC)

I'm stuck. I have a file called "Bennett.png". I'm on this page with the following in the summary box:{{information |Description={{en|1= }} |Source=Screenshot |Date= |Author=Wikipedia Authors, see the history of [[w: ]] |Permission={{Wikipedia-screenshot|1=en|logo=no}} }}
Vchimpanzee · talk · contributions · 20:44, 10 December 2013 (UTC)
Down at the bottom there should be a button Upload file. Click that. --Redrose64 (talk) 22:21, 10 December 2013 (UTC)
I don't know the file name. It's "Bennett.png" but surely that's not enough.— Vchimpanzee · talk · contributions · 22:39, 10 December 2013 (UTC)
The file Bennett.png does not actually exist, so you could use that name (it's not very meaningful, but we can fix that later by a file move). --Redrose64 (talk) 22:45, 10 December 2013 (UTC)
I'm talking about what the file is called now. I used paint and clicked on save. Now what do I do when typing the name of the file in the "source filename" field. I click on "Browse" and see my first name, "Public", "Computer", "Network", and a whole bunch of categories that came with the computer. Nothing about "Bennett".— Vchimpanzee · talk · contributions · 22:48, 10 December 2013 (UTC)
When you saved the file in Paint, you must have put it in some folder or other. My guess is that it's under "Desktop" → "My Documents" → "My Pictures", but since I don't have access to that computer, I really can't tell for sure.
If you are in a library, surely there's some geeky teenager about that you could ask? They tend to be very technology-savvy. --Redrose64 (talk) 23:04, 10 December 2013 (UTC)

────────────────────────────────────────────────────────────────────────────────────────────────────I'm at home. I merely saved as "Bennett.png". I didn't PUT it anywhere.— Vchimpanzee · talk · contributions · 23:07, 10 December 2013 (UTC)

I give up. I'm going to start this at a FOURTH library tomorrow and maybe someone can help me. I'm looking at "Bennett.png" right now but I have no way to send it to anyone.— Vchimpanzee · talk · contributions · 23:13, 10 December 2013 (UTC)
After clicking Browse, try clicking your first name, then click "Pictures". (If you don't see a "Pictures" option, perhaps you need to click "Documents" first, then "Pictures". I don't have access to a Windows PC to check ATM.) Hopefully your Bennett.png will be listed there – click it. – PartTimeGnome (talk | contribs) 00:37, 11 December 2013 (UTC)
Thanks, I'll try that next time. I'm here at the fourth library now and the man says he can help me.— Vchimpanzee · talk · contributions · 14:48, 11 December 2013 (UTC)
We did it! The file is File:BennettProblem.png.— Vchimpanzee · talk · contributions · 14:59, 11 December 2013 (UTC)

Looking at it, nothing seems obviously wrong, though it is odd that you are at the mobile site. Chris857 (talk) 15:07, 11 December 2013 (UTC)

I did a Bing search for more information about a newspaper article topic, and this came up. Whatever it was, it wouldn't let me look at the article's text, and it had no place to sign in, and I didn't know what I was going to do.— Vchimpanzee · talk · contributions · 15:22, 11 December 2013 (UTC)
Mobile menu
I believe that the three-bar icon at the top left corner is where you'll find the menu with the login option, as shown in this image. I've never used it, but this seems to be how it works. mw:Mobile design/Wikipedia navigation has some more information. Whatamidoing (WMF) (talk) 18:25, 11 December 2013 (UTC)
────────────────────────────────────────────────────────────────────────────────────────────────────If you want to view the page in the "normal" Wikipedia instead of the mobile version, scroll to the very bottom of the page then click "Desktop". – PartTimeGnome (talk | contribs) 00:01, 12 December 2013 (UTC)
Thank you, everyone. The advice from User:PartTimeGnome worked (except I had to double click) and I could have uploaded the photo had I stayed on the computer long enough, but I had been on a very long time and had a lot of real world stuff to do. I'll save a link when the other discussion gets archived. Someone commented it wasn't a serious problem, but if you've never seen mobile Wikipedia before and don't know what to do when you get something like that, it IS serious. Now we have advice about what to do. Why Bing sent me there is a mystery.— Vchimpanzee · talk · contributions · 20:35, 12 December 2013 (UTC)

An error about web font

I found an error about web font for Wikipedia today. For example, en:Tibetan alphabet can display Tibetan alphabets with Jomolhari, but zh:藏文字 cannot. How to deal with the issue? --Great Brightstar (talk) 14:08, 11 December 2013 (UTC)

Could you provide more details where exactly the problem can be seen on the pages, and which browser and version you use? If you are sure that you have found a bug in the software, see mw:How_to_report_a_bug --AKlapper (WMF) (talk) 13:48, 13 December 2013 (UTC)

Link to deleted contributions

Is there a way to get to that link faster than by clicking "Special:Contributions" then "deleted user contributions"? Maybe added onto the sidebar list or something? Thanks for any suggestions you can offer. Anna Frodesiak (talk) 21:37, 11 December 2013 (UTC)

This is probably not the most modern method to do it, but I use WP:POP, which allows you to hover over a username, then click "deleted contributions". —Kusma (t·c) 21:45, 11 December 2013 (UTC)
That does the trick! Many thanks. :) Anna Frodesiak (talk) 21:52, 11 December 2013 (UTC)
The below admin-only code in Special:MyPage/common.js should give you a toolbox link below "User contributions". (To curious non-admins: Removing the admin test will not give you access to deleted contributions) PrimeHunter (talk) 02:11, 12 December 2013 (UTC)
$( document ).ready( function() {
  var rights_isAdmin = (wgUserGroups.toString().indexOf('sysop') != -1);
  if (!rights_isAdmin) return; //Restrict to admins.
  if (wgNamespaceNumber != "2" && wgNamespaceNumber != "3") return;
  var username = mw.config.get('wgTitle').split( '/' )[0].replace( /\"/, "\\\"");
  mw.util.addPortletLink('p-tb', mw.util.wikiGetlink('Special:DeletedContributions/')+username,
    'Deleted contributions', 't-deleted', 'Deleted contributions of this user', null, '#t-log');
Fantastic. Even better, and exactly what I wanted. Thank you so much!!! :) Anna Frodesiak (talk) 14:19, 12 December 2013 (UTC)

Note: switched code to mw.util, since addPortletLink has been depreciated for 3 years. Theopolisme (talk) 23:58, 13 December 2013 (UTC)

Template:in title|searchterm asks about creating page "Intitle:searchterm"

Template:In title, which I see used in some "See also" sections,

Creates a link to search for page titles which contain given words. The default value of the first parameter is the current page name, and that of the second parameter is 'All pages with titles containing "searchterm".' Template:In title/doc

For example, the first example on that documentation page is

  • {{In title|Wales}}
produces All pages with a title containing Wales.

That it does; but the resulting search page begins with

You may create the page "Intitle:Wales", but consider checking the search results below to see whether the topic is already covered.

Isn't this a bug? I would have mentioned this on the Template talk page, but that hasn't been written to in over two years. (Please ping me in your reply.) --Thnidu (talk) 01:11, 12 December 2013 (UTC)

Well it's perhaps unexpected and not so user friendly. But it's not really a bug technically. That part just knows about what you entered, it has no capability to interpret the 'query' that you are making. That is all implemented at the search backend. The website interface sends it whatever you input, and the backend returns a list with results. So you can create an enhancement request in bugzilla to improve that experience. —TheDJ (talkcontribs) 01:22, 12 December 2013 (UTC)
I've found two existing bugs for this. I've added links to them at the top of this discussion. – PartTimeGnome (talk | contribs) 22:39, 12 December 2013 (UTC)
@Thnidu: That is a general feature of the search function and not specific to the template. The same thing happens if you manually enter "Intitle:Wales" in the search box to find titles containing "Wales". If whatever you entered in the search box, including search syntax like "intitle:", isn't a page name then the software says you may create it. The message comes from MediaWiki:Searchmenu-new where $1 is the contents of the search box. If MediaWiki messages can invoke modules then I suppose we could make a search message module which analyzes $1 and tries to remove search syntax, or doesn't display a message at all if the search used search syntax. I don't know whether it would be too expensive to do such tests for every search. PrimeHunter (talk) 01:27, 12 December 2013 (UTC)
Thanks, PrimeHunter. I guess it would be feasible to have {{In title}} and other such templates include some code that MediaWiki:Searchmenu-new would check for to suppress the message, but yeah, it would be a lot of work: maybe not to set up (I'm ignorant thereabouts), but to add into all the templates like that and make sure it gets added into future ones. --Thnidu (talk) 02:02, 12 December 2013 (UTC)
Templates wouldn't have to do anything. MediaWiki:Searchmenu-new would just have to check for search syntax like intitle:, incategory:, prefix:. The "You may create the page" message never makes sense for those, regardless whether the user entered them manually or got there via a template like {{in title}}. If we started manipulating $1 then we could also consider other things, for example removing quotation marks. The search "Two words" suggests to make "Two words", but Two words is far more likely to be an appropriate title. PrimeHunter (talk) 02:39, 12 December 2013 (UTC)
Although MOS:SEEALSO does not explicitly forbid search results pages (unlike WP:ELNO#EL9), I don't think that {{in title|something}} is a suitable candidate for inclusion in a "See also", because it's not a link to a related article, but might be considered equivalent to a disambiguation page. --Redrose64 (talk) 14:45, 12 December 2013 (UTC)

A tool to tag taskforce pages?

Resolved: Now at WP:BOTREQUEST. Thanks all. DePiep (talk) 14:39, 13 December 2013 (UTC)

An editor created a new WikiProject taskforce. Is there a tool that can help us to tag all relevant pages? We want to edit the project banner on talkpages: add taskforce name and assessments quality, importance. (Best would be a good manual. After this worst WP manual I'm low on patience). -DePiep (talk) 07:54, 12 December 2013 (UTC)

AWB can be used to modify text on all pages in a category. That category can be all talk pages tagged as belonging to some WikiProject, and the text can be the banner source code. This would be a semi-automated process, as AWB would show you each page in sequence, and you would tell it how to modify the template (either selecting from a list of predesignated choices, or manually entering the changes). Someguy1221 (talk) 08:13, 12 December 2013 (UTC)
Yes, I could do that. Unfort, not available for all editors. -DePiep (talk) 10:05, 12 December 2013 (UTC)
If you're looking for full automation, there are a number of bots written that do this kind of thing, but the only one that I know of that is active now is AnomieBOT (the WikiProjectWorker task). I suspect this particular task might not have been run in a while, though - Anomie, could you comment on the feasibility of this? — Mr. Stradivarius ♪ talk ♪ 08:47, 12 December 2013 (UTC)
I just completed a run for WP:WikiProject Women artists a few days ago, although before that I hadn't run it in a while and the run shook out a bug (now fixed). Anomie 14:18, 12 December 2013 (UTC)
Also, this kind of request usually goes on WP:Bot requests. Most of the people who like helping with bots watch that page, so it might get you a better response than here. — Mr. Stradivarius ♪ talk ♪ 08:53, 12 December 2013 (UTC)
Adding articles to a project/taskforce is botable? Interesting, but hard to imagine that a bot could do the logic. I'd guess a tool could make suggestions (e.g. by crossing categories) but thee tag are manual (Hotcat, AWB). -DePiep (talk) 10:05, 12 December 2013 (UTC)
It all depends on whether the logic can be made to not require human judgment. Category membership or category intersection works if there are categories that are entirely within the scope of the project. AnomieBOT can even handle things like "Contains Infobox Foo with |bar= set to 'baz'". Anomie 14:18, 12 December 2013 (UTC)
Aah, sounds very great. I'll tell the others who asked me about this. They're not all bot/AWB requesters I guess. -DePiep (talk) 15:09, 12 December 2013 (UTC)
One approach that we've taken in the past is a two-step system: First, make a list for review/editing by a human. Second, have a bot tag articles based on the human-verified list. WhatamIdoing (talk) 23:37, 12 December 2013 (UTC)

A bit of customization (monobook)

What's the CSS required to customize the "edit" tab at the top so it changes to a + sign? (I know there's some MediaWiki hack currently in place to do that for "new section".) I tried:

li#ca-edit {display: "+";}

but I'm going to guess the display property doesn't take strings. I would also like to, if possible, move the User Rights link at the bottom row of Special:Contributions into my toolbar (say, between Logs and Upload File) when I'm viewing any userpage so that it's easier for me to check at a quick glance their userrights. TeleComNasSprVen (talkcontribs) 10:42, 12 December 2013 (UTC)

Correct, the display: property does not take strings. You need to use javascript; the 'Replace the "new section" tab text with "+"' gadget is MediaWiki:Gadget-addsection-plus.js. Based on that, I offer:
$(document).ready( function () {
    var tab, tablink;
    tab = document.getElementById('ca-edit');
    if ( !tab ) {
    tablink = tab.getElementsByTagName('a')[0];
    if ( !tablink ) {
    tablink.firstChild.nodeValue = '+';
    if ( mw.config.get( 'skin' ) === 'monobook' ) { = ".4em"; = ".4em";
Put that in Special:MyPage/common.js. --Redrose64 (talk) 12:33, 12 December 2013 (UTC)
Also, would setting the CSS background as "+" image and font color as transparent also work? Link wp:Custom shows moving items on page. -Wikid77 (talk) 13:30, 12 December 2013 (UTC)
A transparent font may be invisible but the text is still there, so it still has the same width as the words "edit this page". --Redrose64 (talk) 14:38, 12 December 2013 (UTC)

A slightly simpler jQuery solution (might require something additional for monobook, it seems Looks fine for me using monobook, actually. @Redrose64: what was the padding for, if you don't mind my asking? I'm probably missing something)...

$('#ca-edit a').text('+');

Theopolisme (talk) 02:07, 13 December 2013 (UTC)

As I say, it was based on MediaWiki:Gadget-addsection-plus.js; since that script works in both MonoBook and Vector, my offering above should do so too - in fact the only change from that script is in the third non-blank line, where 'ca-addsection' was altered to 'ca-edit'.
The code
    if ( mw.config.get( 'skin' ) === 'monobook' ) { = ".4em"; = ".4em";
may be omitted; it decreases the padding from 0.8em each side, and is present because of MediaWiki talk:Gadget-addsection-plus.js#Extra spacing.
If you want to use some symbol other than a plus sign, just alter the '+' in the eleventh non-blank line: for example, if you want the symbol ✍ (U+270D: Writing Hand), change the '+' to '\u270d'. --Redrose64 (talk) 11:32, 13 December 2013 (UTC)
Thanks. TeleComNasSprVen (talkcontribs) 07:23, 14 December 2013 (UTC)

OK, I've put in some code of my own that I designed in my common.js to put in the User Rights link in my toolbox, but I don't know how to put it higher up on the list, so that it fits between logs and user contribs. Does that require css or js, and what would be the class name for the portlet link? TeleComNasSprVen (talkcontribs) 07:23, 14 December 2013 (UTC)

Move-to-commons assistant

I've been unable to make this tool work since it was ported over to WMFlabs. It takes a long time, then tells me that TUSC is verified and the file has been created, but when I look there's no file. Any idea what wrong, on what end of things, and what I can do to fix it? Moving one or two files by hand is OK, but when I've got a dozen to do, not having a tool to do it with is annoying. Beyond My Ken (talk) 03:18, 13 December 2013 (UTC)

I had the same problem but I swapped to using User:This, that and the other/For the Common Good instead - works fine for me. NtheP (talk) 09:45, 13 December 2013 (UTC)
Wow, FTCG is a great tool, very slick. --ukexpat (talk) 17:04, 13 December 2013 (UTC)
@Nthep - Thanks very much for that pointer, "For the Common Good" is indeed an excellent tool! Beyond My Ken (talk) 06:51, 14 December 2013 (UTC)

Filtered RecentChanges errors

Does anyone know why this particular Special:RecentChanges search returns database errors so frequently?

It only works about one time out of ten for me. The others I've tried work on the first try. WhatamIdoing (talk) 16:53, 13 December 2013 (UTC)

Most likely because there are so few edits tagged with this tag, meaning that MediaWiki has to scan a lot more data to get the 50 (by default) rows it wants to show you. This should probably be limited somehow. Matma Rex talk 19:45, 13 December 2013 (UTC)
I think you may be right. If I limit it to 3 days/50 items, I get a higher success rate. WhatamIdoing (talk) 19:54, 13 December 2013 (UTC)
See also #Getting error message when trying random page in Education Program namespace. --Redrose64 (talk) 21:45, 13 December 2013 (UTC)

A tiny row at the top

I asked this question in Template talk:Infobox hurricane but I got no response, so I ask it here.

A tiny row is at the top, below the title and above the category of hurricane. It corresponds to the following html code:

<th colspan="2" style="text-align:center;font-size:125%;font-weight:bold;"></th>

I cannot see why there is such row from the source code of the template. Can this row be removed?--Quest for Truth (talk) 16:59, 13 December 2013 (UTC)

Responded at that talkpage. DMacks (talk) 18:47, 13 December 2013 (UTC)

Are there server issues?

Prior to this, everything has been slow all day. For the last 2 hours, nothing pulled up connected with Wikimedia - any language version of Wikipedia, Wikisource, Commons, anything. The URL came up immediately, but it just kept telling me it was connecting and nothing filled the page. Firefox and Opera. Now, the page slowly comes up, but it looks like the style sheet is missing from everything. Kind of like that on Commons - it pulls up without a style sheet, and no images appear. I went into "page Source", and at the bottom is said " Served by mw1081" - is that the server name? — Maile (talk) 23:39, 13 December 2013 (UTC)

Hi Maile,
Do you mind saying what part of the world you are in? All's well for me, but I'm in California, and geography often matters for this problem. Whatamidoing (WMF) (talk) 02:54, 14 December 2013 (UTC)

This is giving me a database error "Function: IndexPager::buildQueryInfo (LogPager); Error: 0" when I was trying to look for entries from Cydebot from March 2012 and earlier. Is it so high volume a request that the server cannot handle it? TeleComNasSprVen (talkcontribs) 07:54, 14 December 2013 (UTC)

@TeleComNasSprVen: it may be a different problem, it sounds more like #Filtered RecentChanges errors and others that have occurred recently. --Redrose64 (talk) 08:34, 14 December 2013 (UTC)
Everything is fine this morning. I signed off earlier than usual yesterday due to this. Before I signed off, it was getting slowly better on a hit and miss basis. That is, one page would take longer than normal but eventually load correctly. The next one would not. Henrik's stats opened as normal, but I think they are not residing on Wikimedia. Oddly enough, if I was on a partially loaded article and then went to the stats, when I came back to the article it looked normal. Strange happenings. — Maile (talk) 12:14, 14 December 2013 (UTC)

Videos not loading in Gallery

Hey folks, I'm creating a 'meet Wikipedians' page for The Wikipedia Adventure. But the videos aren't playing. Any idea what's happening here: Wikipedia:TWA/People. Cheers, Ocaasi t | c 16:12, 14 December 2013 (UTC)

Fixed, needed closing </big> tag. Ocaasi t | c 19:22, 14 December 2013 (UTC)

No PNG views

I noticed for a while now that when you click on an SVG file, there is no longer a PNG viewer. I don't know if that is considered a bug, an error, or a issue there having with that. Blurred Lines 15:30, 9 December 2013 (UTC)

Please give an example and clarify what you mean by a PNG viewer. File:Example.svg is hosted at Commons but displays If you click the Commons link to commons:File:Example.svg then I see "This image rendered as PNG in other widths: 200px, 500px, 1000px, 2000px." Is that what you mean? It's made by commons:MediaWiki:Common.js and isn't in Wikipedia. PrimeHunter (talk) 16:19, 9 December 2013 (UTC)
Yes, that's what I meant. Somehow it's not showing up when I click on an SVG file. Blurred Lines 16:41, 9 December 2013 (UTC)
Do you mean it's not showing up at commons:File:Example.svg? Do you have JavaScript enabled in your browser? If you have then you should see a [show]/[hide] link at #toc on this page. If you mean another page then please give a link to a specific example. PrimeHunter (talk) 17:07, 9 December 2013 (UTC)
Yes, my JavaScript is on, and I am not talking about on Commons, I'm talking about it not showing up on Wikipedia. Like on this file, and others that have SVG format. Blurred Lines 17:11, 9 December 2013 (UTC)
I mentioned it's a Commons feature. Has it ever been at Wikipedia? PrimeHunter (talk) 17:41, 9 December 2013 (UTC)
The English Wikipedia's MediaWiki:Common.js says:
if ( mw.config.get( 'wgNamespaceNumber' ) === 6 ) {
    /* file description page scripts */
    importScript( 'MediaWiki:Common.js/file.js' );
MediaWiki:Common.js/file.js by User:TheDJ has PNG code for SVG images so maybe a problem has developed with that code. PrimeHunter (talk) 18:09, 9 December 2013 (UTC)
Ok, I have started a discussion about this on its talk page, perhaps someone with more experience on the interface can have a solution to this error. Blurred Lines 18:51, 9 December 2013 (UTC)
Fixed. I had forgotten the script even existed.. —TheDJ (talkcontribs) 19:20, 9 December 2013 (UTC)
Is this the same issue as Wikipedia:Village pump (technical)/Archive 120#Option to render a SVG images as PNG disappeared? --Redrose64 (talk) 19:53, 9 December 2013 (UTC)
Yes, that must be it. User:Astronaut should now have PNG links again. PrimeHunter (talk) 21:43, 9 December 2013 (UTC)
Indeed, I get back from vacation and the PNG rendering options are back on both Wikipedia and Commons hosted SVG files. Thank you to PrimeHunter for locating the script buried in Common.js and to The DJ for doing something to fix it. And this is despite several people assuring me that no such thing existed on Wikipedia (including Edoktor who has previously editied the script in question)! Astronaut (talk) 10:32, 16 December 2013 (UTC)

Sort problem, looking for character to sort after numbers

I'm trying to create a sortable table, currently in User:Sphilbrick/2012–13 NCAA Division I women's basketball rankings. Most entries are numbers, and I want to use numeric sorting. However, some schools in some weeks are not in the top 25. Currently, that is indicated by a "-" character. Unfortunately, that character sorts before any number, so when sorting any column, the teams outside the top 25 are listed first. I'd like to replace the character with something that sorts after numbers. I've tried en-dash and m-dash, blanks and non-breaking spaces.

I also tried specifying data-sort-type="number" for the column, but that didn't seem to help. Any suggestions?--S Philbrick(Talk) 00:28, 13 December 2013 (UTC)

Replace "-" with data-sort-value=99|-
I think that should achieve the desired effect. Documentation is at Help:Sorting#Specifying_a_sort_key_for_a_cellJonesey95 (talk) 01:02, 13 December 2013 (UTC)
Thanks, that worked.--S Philbrick(Talk) 01:31, 13 December 2013 (UTC)
Hm, that's weird. unknown elements should sort to the bottom... Can you file a bug report please ? BTW. the way this 'ought' to be done is by using data-sort-type="number" on the column header to note that the column contents should be treated as numbers. Then using an ndash or mdash (not an hyphen or minus) for the cell content. As to why that doesn't work, it will require some investigation. —TheDJ (talkcontribs) 09:44, 13 December 2013 (UTC)
I did try data-sort-type="number" and that didn't seem to work. That said, I was serially trying things, and may not have set it up properly, so I'll try to create a more controlled example, and file a bug report. My original data had a hyphen or minus,, I think, but I did try en and em dashes. I was thinking they should be last, so happy to hear that assumption wasn't unwarranted.--S Philbrick(Talk) 16:25, 13 December 2013 (UTC)
Before I attempt to write up a bug, could someone look at Help:Sorting/Test_cases.
In the first two tables, as TheDJ surmised, the sort works as one would want. Non-numeric entries are at the bottom. However, in a truncated version of the table I want to create, they sort to the top. Obviously, something is different, but I'm not seeing it. While I can replace the fields with the suggested code, the tests suggest I should not need to. I plan to create 60 or so tables, so it would be nice if I could avoid having to do a kludge for each one.--S Philbrick(Talk) 21:04, 13 December 2013 (UTC)
Hold on, I can see a difference, although I'm still confused. The first table is sorting as if it is alphabetic. Note the "25" sorts between 2 and 4, and should be at the end of the numbers. However, the third table sorts like they are numerics, "10" follows "9",not "1". I don't see what is different.--S Philbrick(Talk) 21:11, 13 December 2013 (UTC)
The 'guess' algorithm for data types only looks at the first few rows (for efficiency). Since your short tables start with dashes, the type is determined to be alphanumerical. Your larger table consists of only numbers in the first few rows, so numeric is chosen as 'detected' type. In general you should simply ALWAYS specify a data-sort-type on a column, that will bypass any guessing that the logic will attempt. —TheDJ (talkcontribs) 18:08, 14 December 2013 (UTC)
That only makes sense , it rings a (very faint) bell. Thanks.--S Philbrick(Talk) 17:42, 15 December 2013 (UTC)

Customizing starting page for all Zero users

The community now has an ability to customize starting page for Zero users. This issue seems to be dual policy and technical, so posting here as well - please see the post here. --Yurik (talk) 23:36, 15 December 2013 (UTC)

Tech News: 2013-50

08:30, 16 December 2013 (UTC)

Auto Image template- Could the parameters widthn, heightn deleted?

Auto Image template needs the parameters "|widthn=, |heightn= "

widthN – this is the width in pixels of the full resolution version of the image in Commons or Wikipedia. Examine the individual image page to obtain the full resolution.

heightN – this is the height in pixels of the full resolution version of the image in Commons or Wikipedia.

Could the template read directly those 2 parameters ? It is cumbersome to look for those numbers.

BTW the instructions are:"Parameters with numerical values, if unused, must be removed and not left blank; if left blank, they may cause undesirable side-effects. These include |widthn=, |heightn= and |total_width=. If |total_width= is omitted, the default value is 600.".

Omitting the "heightn", it would presnt the images, but the relevant image is smaller, and it does not matter what is the parameter value. Omitting the "widthn" or both parameters, it does not work. Ykantor (talk) 20:01, 12 December 2013 (UTC)

The template can not read directly those 2 parameters, but that option is being worked on - see Bugzilla: 41498.--Snaevar (talk) 21:40, 12 December 2013 (UTC)
Thank you. Ykantor (talk) 09:17, 13 December 2013 (UTC)
i think maybe it's better to remove this template. it seems that using "Gallery" with mode=packed provides the functionality, only better (for instance, this template force the editor to define the width which might cause horizontal scrollbar, and the gallery knows to adjust the width to the user screen/browser width. gallery does not force you to think about specific images dimensions, does not requires you to learn parameter names, is not limited to N images).
what's the justification for a limited and specialized template that does something the native wikicode does better? (i assume the template predated the "mode=packed" feature of gallery - my point is that once this mode was introduced, the template lost its raison d'être).
this template is used in less than 200 articles - converting them all to gallery should not be prohibitive.
peace - קיפודנחש (aka kipod) (talk) 00:39, 17 December 2013 (UTC)
The {{auto images}} template was created because of the limitations of {{multiple image}}. Those limitations exist because it is not possible to obtain the image dimensions directly: if it were, {{multiple image}} could be simplified and {{auto images}} would never have been needed. --Redrose64 (talk) 12:01, 17 December 2013 (UTC)

Interwiki font

Agains some changes to interwiki sidebar font?
My User:XXN/common.css has no effect... XXN (talk) 23:50, 13 December 2013 (UTC)

They changed the classes, so we need to use different selectors.
#p-lang li.interlanguage-link, .autonym { font-family: inherit; }
This goes in Special:MyPage/common.css, replacing the code that began #p-lang until the next closing curly bracket. --Redrose64 (talk) 00:28, 14 December 2013 (UTC)
Thank you. //XXN (talk) 01:12, 14 December 2013 (UTC)

They announce somewhere those changes? // XXN (talk) 17:43, 15 December 2013 (UTC)

Updates are announced in Tech News posts. There are too many changes to list here, but you can track any changes per release at MediaWiki, for instance mw:MediaWiki 1.23/wmf6. Edokter (talk) — 18:00, 15 December 2013 (UTC)
But this is significant update. And, because they changed class name, on some local wikis, stars for GA & FA from interwiki sidebar are not displayed more, until one of sysops will find bug. XXN (talk) 23:34, 15 December 2013 (UTC)
This was also reported at MediaWiki_talk:Common.js#LinkFA_is_broken and was fixed by this change.
You can also see the list of changes on refs/heads/wmf/1.23wmf6 (replacing the MW version by the one shown at Special:Version).
In particular, the new class was added on gerrit:94899. Helder 11:51, 18 December 2013 (UTC)

Edit filter tag: mobile edit

I see that there's a new Tag for mobile edit that pops up very frequently on my watchlist. Was there some discussion or proposal about this new filter that I missed, or maybe it was a Foundation initiative? Because there appears to be no link to the particular documentation for the filter on the special page. TeleComNasSprVen (talkcontribs) 07:23, 14 December 2013 (UTC)

Not all tags are from edit filters. The "mobile" and "VisualEditor" tags are two examples of this. Wikipedia:Tags mentions this as well, but does not contain the complete list of extensions using tags (well, the only one deployed on enwiki missing is TorBlock). I believe they come from their respective MediaWiki extensions, which were both developed by the WMF. mw:Manual:Tags is more useful documentation if you're interested in how they are there. I don't know if you understand PHP (it's pretty basic), but this is the code that adds the tag (I think). πr2 (tc) 14:24, 14 December 2013 (UTC)
The main reason for adding these is so that if something odd happens in the edit, then we will know which part of the software is responsible for the problem. I believe that about a thousand edits are made through mobile each day (for the English Wikipedia only). Whatamidoing (WMF) (talk) 22:56, 16 December 2013 (UTC)
I believe we will have more tags in the future, per:
Helder 12:04, 18 December 2013 (UTC)

Removing CSS classes on article maintenance tags

A while ago, on my urging, a unique CSS class name was added to most article maintenance tag templates, made up of the string "ambox-" plus the name of the template. You can see the effect of this when you look at the HTML source of an article like 2-Tone Entertainment: the {{orphan}} tag has a class "ambox-Orphan" and {{unreliable sources}} has "ambox-unreliable_sources". These were added as part of a past Twinkle development project.

For Twinkle's purposes, these class names are no longer required. So I am keen to get rid of them, as a matter of cleaning up after myself, and also to avoid sending unnecessary clutter whenever someone requests an article containing one of these tags.

The only question left is: is anyone else using these class names in their script, bot or tool? I strongly suspect not, as they are incomplete and inconsistent. But if you do use them, please speak up now, as I intend to go through and get rid of these unnecessary classes soon. Thanks, — This, that and the other (talk) 09:23, 16 December 2013 (UTC)

For GettingStarted (a tool to help people, especially new users, find things to work on), we're using ambox currently for a guided tour. It's possible we might want to use the specific classes in the future, though. We do already choose articles based on what kind of work they need (e.g. copy-editing). It's just that currently we don't take the template name into account when pointing to the box, but we might later want to. I think the mobile site also has special handling for these templates, so I'll give the team a heads up. Superm401 - Talk 22:33, 16 December 2013 (UTC)
In guided tours delivered to new editors, we're currently referring to the general ambox class, not the template specific ones. However, I could easily see using it in the future if we want to only refer to a specific template and not others. For instance, if we're delivering a tour of how to add references, and want to tell someone to remove a missing references template but not others. I'd recommend we keep these for now, unless they're really making a mess of our styles and/or negatively impacting performance. Steven Walling (WMF) • talk 22:30, 16 December 2013 (UTC)
In the mobile site the ambox class is used to identify cleanup templates in a page and collapse them (on some devices with small screens it really bad effects the reading experience). Please do not remove them without providing some other method (class name preferred) to identify them as without them we have no idea what they are. We also use .mbox-text, .ambox-text to identify the summary of the problem. Jon (WMF) (talk) 22:42, 16 December 2013 (UTC)
@Jon (WMF): I think he's just talking about removing the template-specific part of the class, not the ambox class entirely. Steven Walling (WMF) • talk 23:29, 16 December 2013 (UTC)
Thanks for the input. Seems like I should leave them be, then. If you do end up using this data, please be aware that these class names are very inconsistent, and they do not always match the actual template name (due to template moves over the years). Additionally, the capitalization of the first letter is not consistent. You might need to go through and fix them up before making use of them. — This, that and the other (talk) 00:10, 17 December 2013 (UTC)
Yeah some cleanup may be in order. Thanks ! Steven Walling (WMF) • talk —Preceding undated comment added 01:48, 17 December 2013 (UTC)

{{cite doi}} not expanding

Can someone tell my why the {{cite doi}} is not being expanded at Mrs Macquarie's Chair? Even if I "jump the queue", the Citation Bot seems to indicate that there is nothing to do. Mitch Ames (talk) 14:19, 16 December 2013 (UTC)

[34] sort of explains the problem, but it's not very clear:
No CrossRef record found :-( - This means that the CrossRef database that it uses for getting the info doesn't have any metadata for it. Mr.Z-man 17:05, 16 December 2013 (UTC)
A WP:GNOME expanded the citation manually. Not all dois can be expanded automatically; the vast majority can, but there are a few that just don't work. Citation Bot depends on various databases having correct information; even if you can click the link to get to the article, the information in the databases may not exist or may not be accurate. – Jonesey95 (talk) 06:07, 17 December 2013 (UTC)

To ping, or not to ping...

Somewhere in User talk:Tumandokkangcabatuan (after my one post there), I was pinged by the user. It still hasn't come up on notifications. Is there a fault on the system? Peridon (talk) 20:10, 16 December 2013 (UTC)

He's pinged me again - but no notification appears. (No, I've not turned it off in Preferences, by the way.) Peridon (talk) 20:18, 16 December 2013 (UTC)
Pings only work if the ping and the user's own signature are added in the same edit. Jackmcbarn (talk) 20:48, 16 December 2013 (UTC)
Ah, right. Thanks - I didn't know that. Peridon (talk) 21:27, 16 December 2013 (UTC)
This is a side effect of the issue that we use wiki pages for our communication. It's very difficult to distinguish in the software between a case where we edit a talk page and are just moving some sections around (archiving or correcting an incorrect archiving for instance) and the case where we "make a new post". The software currently uses "the poster has ended the new section with his own signature" to make that distinction. Until we have something like Flow or LQT, that's probably the best that can be done, without bombarding everyone with 'false' mentions. —TheDJ (talkcontribs) 21:45, 16 December 2013 (UTC)
User:SineBot seems very reliable when it signs posts (it omits to sign lots of posts but that's another matter). I don't recall ever seeing it sign a post which wasn't actually made by that user but just copied from somewhere. I don't know Sinebot's algorithm and its page says "The source code is not public at this point", but would it be possible to detect "real-looking" posts with a method similar to whatever Sinebot is doing? Or does it have too low priority when Flow is planned? (Flow or LQT has been "on the way" for as many years as I can remember). PrimeHunter (talk) 23:57, 16 December 2013 (UTC)
Re "I don't recall ever seeing it sign a post which wasn't actually made by that user but just copied from somewhere." - I do, and it happened on this very page, less than two weeks ago. Gadget850 made this post, which was then accidentally zapped by Matma Rex with this post. Matma Rex quickly noticed the problem and restored the lost post, which SineBot then signed in error. --Redrose64 (talk) 11:25, 17 December 2013 (UTC)
Good example. I'm not sure whether notifications are desirable in such cases. It might have been a falsified signature and Gadget850 would probably strongly want to know that, but it wasn't falsified (only the page history could determine it), and falsified signatures are probably very rare. PrimeHunter (talk) 13:38, 17 December 2013 (UTC)
────────────────────────────────────────────────────────────────────────────────────────────────────SineBot didn't mark the edits that added the attempted pings as unsigned, so using the same algorithm to spot user mentions would not have helped in this case. I can't say for certain without access to the source, but I believe SineBot doesn't treat an edit as a post if it doesn't add any lines to the page.
Also, users who copy posts from elsewhere (archiving pages, etc) tend to be experienced users. SineBot deliberately ignores users with more than 800 edits unless they opt-in. I suspect SineBot might sign such copied posts if done by a user with fewer edits. – PartTimeGnome (talk | contribs) 21:45, 17 December 2013 (UTC)
Ah, and Matma Rex has less than 800. --Redrose64 (talk) 00:04, 18 December 2013 (UTC)

templatedata tool(s)

there is a tool that exists on hewiki and on mw, to help create and edit "templatedata". to review it, go to mw:Special:Preferences#mw-prefsection-gadgets and enable the "TemplateData" gadget, then either edit any existing template, or create a new one (IOW, copy a template from enwiki), and notice the "templatedata" button. press it, and you'll notice that the editor knows how to read all the parameters automagically from the template. can someone please port this tool here?

thanks, peace - קיפודנחש (aka kipod) (talk) 01:11, 17 December 2013 (UTC)

I use User:NicoV/TemplateDataEditor on enwiki, which has similar functionality (and, imho, a more polished interface). Theopolisme (talk) 01:53, 17 December 2013 (UTC)
as far as i could see, the script you use does not have the main feature of the mw one, which is fully automatic generation of the tempaltedata, by reading the template parameters. please correct me if i'm wrong here. peace - קיפודנחש (aka kipod) (talk) 02:35, 17 December 2013 (UTC)
Er, yes it does. Go to any page lacking templatedata, click the TDE button in the edit window, and you should see a "Prefill the data" window. Theopolisme (talk) 02:59, 17 December 2013 (UTC)
Thanks. i did not realize this. peace - קיפודנחש (aka kipod) (talk) 04:23, 17 December 2013 (UTC)


As noted on Wikipedia talk:Special:FewestRevisions, this database report special page seems to mainly comprise of redirects from 2004 or whenever that it updates itself with, restricting the usefulness of the special page. So I have to ask, can this be fixed, or will it eventually be deprecated and ignored? TeleComNasSprVen (talkcontribs) 11:36, 17 December 2013 (UTC)

Is this bug 9736? πr2 (tc) 17:45, 17 December 2013 (UTC)
(written prior to πr2's answer) It is allready deprecated. Notice the message "Updates for this page are currently disabled. Data here will not presently be refreshed." This means that this special page will not be updated. In case you did not know, special pages are updated with an script called updateSpecialPages.php which can only be run by system administrators.--Snaevar (talk) 17:49, 17 December 2013 (UTC)
It says it was last updated "21:08, 16 December 2013". Is this incorrect? πr2 (tc) 17:54, 17 December 2013 (UTC)
A cursory review of most of these links reveal them to be nothing more than redirects. Funny thing about this special page was that I encountered this the other day and it said it was last updated, if I vaguely remember correctly, sometime in 2010, which led to my (over)prompt nomination of the corresponding Wikipedia page for deletion, and one other editor in that discussion did note this as well. Yet for some reason it updated recently causing another editor to !vote keep in the discussion and the former to change his mind. (As if it was sentient or something. Or perhaps that a simple query of the page had consequences that we had not foreseen.) TeleComNasSprVen (talkcontribs) 18:49, 17 December 2013 (UTC)
Some of the listed pages are from 2013. "A" Fort and Battery Hill Redoubt-Camp Early is from June and has not been a redirect. PrimeHunter (talk) 20:02, 17 December 2013 (UTC)
User:John Broughton on the talkpage:

Someone really should submit a bug report; as is, this is a complete waste of resources. It also would be a complete waste of resources even if redirects were eliminated: almost all of the 1000 listed articles would probably be those created within the 24 hours prior to the report being run. (There are roughly 4000 articles created each day.) To be worthwhile, this report should be limited to articles that have been around for a bit - say, those created more than 60 days prior to the report date.

[UPDATE, 12/2013] As a side note, for those who might find the estimate of 4000 new articles daily to be a bit surprising: My comment was from 2007, prior to the implementation of WP:AFC, which substantially reduced the number of new articles created each day (by disallowing new article creation by non-registered editors). -- John Broughton (♫♫) 18:19, 18 December 2013 (UTC)
But anyway, the fact that the report contains redirects creates an interesting problem. TeleComNasSprVen (talkcontribs) 20:05, 17 December 2013 (UTC)
It appears from mw:Release notes/1.10#New features since 1.9 and bugzilla:9508 that this a default MediaWiki page. Most wikis have far fewer pages than us. PrimeHunter (talk) 00:28, 18 December 2013 (UTC)

Issues with __News release section__ template

Including the {{News release section}} template brings with it some script code; see HP-UX#Characteristics. -Mardus (talk) 11:37, 17 December 2013 (UTC)

Fixed.[35] Many templates have a <noinclude> but no matching </noinclude> at the end. The software makes them work so it often goes undetected but here it caused an error in [36] when AnomieBOT II added something to the end which should have been after the missing </noinclude>. @Anomie: This could easily happen elsewhere. PrimeHunter (talk) 11:59, 17 December 2013 (UTC)
I have added the issue [37] to the documentation of Module:Unsubst. PrimeHunter (talk) 12:15, 17 December 2013 (UTC)
Ugh. Bot fixed, and I'll run a check for any other templates with this issue in a little bit. Anomie 12:55, 17 December 2013 (UTC)
It turned out that was the only template edited by the bot with that problem. Anomie 14:02, 17 December 2013 (UTC)
Thanks for checking, and fixing the bot. I have seen other templates without the ending </noinclude> and they look like errors waiting to happen when somebody does something that should normally have been OK. PrimeHunter (talk) 14:14, 17 December 2013 (UTC)

Need help changing tonnes per hectare to tons per acre

I posted a request for help on my Talk page regarding Quinoa. I received an answer, but it didn't help me. I know how to use the conversion templates to convert meters and kilometers to feet, kilograms to pounds, centimeters to inches, etc., but I have not been able to convert tonnes per hectare to either tonnes or (American short) tons per acre. Can someone help me? Thanks.CorinneSD (talk) 18:30, 17 December 2013 (UTC)

Like this? 10 tonnes per hectare (4.5 short ton/acre). Code {{convert|10|t/ha|ST/acre}}. Chris857 (talk) 18:41, 17 December 2013 (UTC)
Help:Convert has links to relevant pages, including where to ask for help. Johnuniq (talk) 19:47, 17 December 2013 (UTC)

@CorinneSD, as noted already, the typical unit-codes are "t/ha" and "ST/acre" but which unit-codes had you tried instead? The prior markup-based Convert has auto-correction for some invalid units, such as pounds "lb" to "km" will automatically reset the "km" to "kg" and show kilograms, or "lb" to "impgal" will show imperial gallons of water equivalent to a weight in pounds. Perhaps a mixed code such as "t/acre" (tonnes per acre) should auto-correct as "LT/acre" for long tons, and we can change the Convert template to perform sophisticated auto-corrections using markup, where the Lua Convert does not yet support that feature. The whole goal, behind auto-correction of units, was to provide a logical conversion, without users having to contact a noticeboard to complain (again) about several alternative conversions which are difficult to specify but could be auto-adjusted, such as "t/ha" to "acre" auto-correcting as "LT/acre". -Wikid77 (talk) 05:53, 18 December 2013 (UTC)

Draft namespace live

Subject namespaces Talk namespaces
0 (Main/Article) Talk 1
2 User User talk 3
4 Wikipedia Wikipedia talk 5
6 File File talk 7
8 MediaWiki MediaWiki talk 9
10 Template Template talk 11
12 Help Help talk 13
14 Category Category talk 15
100 Portal Portal talk 101
108 Book Book talk 109
118 Draft Draft talk 119
446 Education Program Education Program talk 447
710 TimedText TimedText talk 711
828 Module Module talk 829
2300 Gadget Gadget talk 2301
2302 Gadget definition Gadget definition talk 2303
-1 Special
-2 Media

Just in case you missed it, we now have a new Draft namespace. See Draft:Example for an example page in this namespace. Pages in this namespaces are not indexed by search engines, and IPs can create pages in this namespace. — This, that and the other (talk) 02:00, 18 December 2013 (UTC)

See more at #Draft namespace being enabled soon. PrimeHunter (talk) 02:07, 18 December 2013 (UTC)

Merging pages with infoboxes

Hello to all, I want to propose a merge between several anatomical articles with infoboxes. However, before I complete any of my proposed merges (if there is consensus in that direction), I will need to be sure I'm not destroying any wiki-data. My understanding of infoboxes is that they're trawled by bots and wikidata so that they may eventually be correlated with other-language wikipedias. Thus, if I propose a merge and remove an infobox, this won't be possible because the link to a specific anatomical structure is removed. Could an enlightened user please open the clouds for me on this issue? The only way I can see of this being done is either to have multiple small infoboxes on one page, or an infobox with multiple entries in it. Is that feasible infobox-wise? --LT910001 (talk) 07:58, 18 December 2013 (UTC)

What can we help you with? As I believe some articles may have links back to WikiData, which some helpful editor there might be willing to change, can you list the articles in question? TeleComNasSprVen (talkcontribs) 08:05, 18 December 2013 (UTC)
My question is, (1) is it possible to merge two articles with infoboxes, and if this is done, (2) is it much of a problem? --LT910001 (talk) 08:17, 18 December 2013 (UTC)
The answer is, the devil's in the details. Why don't you give a couple of examples, and then invite a couple of WikiData folks to look at it? There's at least one who posts at WT:MED about WikiData updates. WhatamIdoing (talk) 19:03, 18 December 2013 (UTC)
OK. Two examples. Example one - cervical vertebrae, Cervical vertebra 3, Cervical vertebra 4, and Cervical vertebra 5. If at some point there is consensus to merge these into the Cervical vertebrae article, these will lose the infoboxes that describe their Latin names and FMA descriptors.
Example two - the articles related to spinal vertebrae (Transverse process, Spinous process) are merged into a single article, Spinal vertebrae). How does one retain the information contained within the infoboxes (eg latin names, FMA/Gray's descriptors) if these are merged into a single article? --LT910001 (talk) 01:51, 19 December 2013 (UTC)
Well, since WikiData has not initiated their Phase 2 plan to centralize infobox information yet, I think we can still work with this. And it seems to be a relatively simple and harmless request. My suggestion about the infoboxes is to either remove the entries corresponding to the merged articles or leave them be and simply redirect the merged articles into the target article. If they are listed on a single section on the target article, or each of them have their own sections, you can even point the entries to the corresponding section using #REDIRECT [[article#section]]{{R from merge}} and it would still be alright. At least that's my take on it, but other editors might say something different. TeleComNasSprVen (talkcontribs) 02:20, 19 December 2013 (UTC)
I wonder whether the infobox could be expanded to offer names of sub-parts. It would make good sense for many articles on major anatomical features (e.g., Arm could list and link to hand, wrist, forearm, elbow, upper arm, and maybe shoulder), and could also solve this problem of having seven cervical vertebrae but not really needing separate articles for (at the very least) numbers two through six. WhatamIdoing (talk) 06:18, 19 December 2013 (UTC)
It's rare, but not unheard of to have multiple infoboxes on a single article, eg Klingon alphabets. VanIsaacWS Vexcontribs 06:38, 19 December 2013 (UTC)
I wonder if using the same type of infobox twice an WP:ACCESS problem. (Using the same image twice is.) There are also articles like Smallpox that have two different infoboxes. Many drug articles contain both {{drugbox}} and {{chembox}}. WhatamIdoing (talk) 16:19, 19 December 2013 (UTC)

The @import rule in CSS - is it blocked by MediaWiki?


CSS has the @import rule, which "allows users to import style rules from other style sheets" according to the current standard, Cascading Style Sheets Level 2 Revision 1 (CSS 2.1) Specification (W3C Recommendation 07 June 2011). It's been around since Cascading Style Sheets, level 1 (W3C Recommendation 17 Dec 1996, revised 11 Apr 2008) and is also in the proposed standard CSS Cascading and Inheritance Level 3 (W3C Candidate Recommendation, 3 October 2013).

I had the idea that I could utilise this to have a single CSS page on meta:, and import that into English Wikipedia, Commons, Welsh Wikipedia, etc. so that when something like this happens again, I would only need to update one central CSS file, instead of one per site. I tried this at m:User:Redrose64/monobook.css to import m:User:Redrose64/allsites.css; but it doesn't work. Might the MediaWiki parser be disallowing the @import rule? --Redrose64 (talk) 17:31, 18 December 2013 (UTC)

@Redrose64: You need to import the raw code, not the URL to the wiki page. For example, zu:User:PiRSquared17/common.css imports my global CSS from Meta. If you really want to have "global" CSS, see m:Forum#Global_CSS_and_JS_scripts, m:synchbot, and m:user:PiRSquared17/Global UP. πr2 (tc) 17:39, 18 December 2013 (UTC)
OK, Face-smile.svg Thank you all sorted now. --Redrose64 (talk) 18:11, 18 December 2013 (UTC)
Using import is kind of a bad idea anyway... Steven Walling (WMF) • talk 18:36, 18 December 2013 (UTC)
The alternative - the <link /> element - can only be used inside the <head>...</head> element, and so is not available for user customisation within Wikipedia. --Redrose64 (talk) 22:40, 18 December 2013 (UTC)

Thorny template issue: Dallas weatherbox

Why does {{Dallas weatherbox}} only show the mean number of precipitation days, snowy days, and monthly sunshine hours when it is transcluded? When I use Special:Expandtemplates on the Dallas weatherbox template with suppression of comments turned off, it displays nowiki tags and HTML comments where the missing values should be, but I can't figure out where these are coming from. It might have been caused by a relatively recent change somewhere (the related changes page isn't really helping me here), because the infobox displayed fine on the actual template page until I purged it. Graham87 03:15, 19 December 2013 (UTC)

It's more than just that it's all weather boxes. Look at {{Weather box}}. All the conversions are missing, the yearly summary and the colours are also gone. CambridgeBayWeather (talk) 03:21, 19 December 2013 (UTC)
I have no idea, but see User talk:Mr. Stradivarius#Module:Math. Johnuniq (talk) 03:52, 19 December 2013 (UTC)
Yes, my recent update to Module:Math probably broke something in Module:WeatherBox. I'm not quite sure what yet, though. The best solution would be to fix Module:WeatherBox, but if we really need to we can revert back to the old version of Module:Math after also reverting Module:Infobox road/length and Module:Infobox settlement. — Mr. Stradivarius ♪ talk ♪ 04:37, 19 December 2013 (UTC)
Ah, now I see it - it's the call to p._cleanNumber. I should have spotted that when I looked through the module before. — Mr. Stradivarius ♪ talk ♪ 04:39, 19 December 2013 (UTC)
I think this should have got it. Let me know if there are still any issues after purging. Sorry about that - I did look at Module:WeatherBox to see if it was compatible with my changes, but I was concentrating on the _precision_format function and neglected to properly check the _cleanNumber function. — Mr. Stradivarius ♪ talk ♪ 04:44, 19 December 2013 (UTC)
Now that a few code-related things have migrated to the Module namespace, it's harder to track down and debug things using Special:Expandtemplates, isn't it? I wonder if that's grounds for a new mediawiki feature, like "Show Lua Modules" embedded into Special:Expandtemplates. TeleComNasSprVen (talkcontribs) 04:53, 19 December 2013 (UTC)
Thanks, guys! Yes, a feature like that would be handy ... the "old-school" methods weren't really getting me anywhere. Not sure how I missed the recent math module changes, though. Graham87 05:07, 19 December 2013 (UTC)
For my part, it would have been useful to have Module:WeatherBox show up in the WhatLinksHere results for Module:Math in the module namespace. As it is, I had to use the search engine to find out which modules were using Module:Math, which brought up a lot of false positives from modules using the native Lua "math" library. Does anyone know if there's a bug open about this? — Mr. Stradivarius ♪ talk ♪ 05:50, 19 December 2013 (UTC)
Module:WeatherBox doesn't link to Module:Math, though. But if it were to have a documentation page with examples, that would have created links.
Personally, I found it easy to track down the problem, starting from Template:Dallas weatherbox and using the "Templates transcluded in this page" along with User:Anomie/previewtemplatelastmod to point me to the most-recently changed pages, and then TemplateSandbox to preview the template with an older revision of the module to confirm the issue. Anomie 12:05, 19 December 2013 (UTC)
I see - I've been thinking of the require function as conceptually equivalent to transclusion, but I can see how that wouldn't make sense if you're thinking in terms of link tables. And thanks for the link to that script - I was actually thinking that I should install it earlier after reading today's AN thread about the template vandalism incident, but I couldn't remember who wrote it or what it was called. It does make tracking these things down a lot easier. I couldn't get TemplateSandbox to work for modules, though. when I tried to preview Template:Dallas weatherbox from Module:Math or Module:WeatherBox I kept getting the message "To preview another page with this template, a page title must be specified." I get the feeling I'm missing something really obvious - can anyone see where I'm going wrong? — Mr. Stradivarius ♪ talk ♪ 14:03, 19 December 2013 (UTC)
"require" does record as a tranclusion, but only when the "require" is actually executed. Unlike with templates, modules aren't invoked automatically on their own page (which function would even be invoked?).
TemplateSandbox works fine for me with modules; that error message sounds like what you'd get if you clicked TemplateSandbox's preview button without filling in its "Page title" field. Anomie 15:37, 19 December 2013 (UTC)

──────────────────────────────────────────────────────────────────────────────────────────────────── Re: require, that makes sense, thanks. About TemplateSandbox, I found the culprit: it works when I disable User:Equazcion/UniversalTransclusionPreviews. I'm no JavaScript expert, but it looks like it only uses the default extension behaviour when called from the template namespace, when it should do that when called from the module namespace as well. Equazcion, could you confirm this, and possibly put in a fix if you have a few spare moments? — Mr. Stradivarius ♪ talk ♪ 16:16, 19 December 2013 (UTC)

Draft special page error

I don't recall exactly how I got here. I was browsing pages in & about Draft:. Then some menu option produced this: [38]

A database query error has occurred. This may indicate a bug in the software.

    Function: SpecialRecentChanges::doMainQuery
    Error: 0 

-DePiep (talk) 12:54, 19 December 2013 (UTC)

That link works for me and shows recent changes to pages in the Draft namespace. I get the url by selecting Draft in the Namespace box at Special:RecentChanges. PrimeHunter (talk) 13:00, 19 December 2013 (UTC)
If not reproduceable, then early close please. -DePiep (talk) 13:04, 19 December 2013 (UTC)
(edit conflict) The link works for me too, perhaps Steven (Steven Walling—[[User:{{{3}}}|{{{3}}}]]) can help with the error message part? Technical 13 (talk) 13:09, 19 December 2013 (UTC)
This sounds like bug 58157 Anomie 15:30, 19 December 2013 (UTC)
I don't think that it's a problem with Draft: namespace per se. Very similar problems have come up before, and if previous threads hadn't been archived, I'd merge them as a related issue. --Redrose64 (talk) 20:02, 19 December 2013 (UTC)
Yes, Anomie is correct. The bug seems to be in RecentChanges, not in Drafts namespace. Steven Walling (WMF) • talk 20:08, 19 December 2013 (UTC)

Suggest search features (lemmatisation, stemming)


I am a french user and I first apologize for my style. I would like to suggest some upgrades to the search engine.

  • All apostrophes (« ' » and « ’ ») in both search terms and result should be recognized as breaking marks (or at least somehow assimilated one for each other), so that terms containing apostrophes could be found by typing one or the other apostrophe, or none of them (e.g. L'Oréal could be recognized by searching either “L’Oréal” or even “L Oréal”). Currently it is not the case, and finding the good article is sometimes a matter of luck. (I don’t know here, but articles in the french Wikipédia share both the apostrophes, depending on keyboard configurations.)
  • That done, all non-wikicode apostrophes could be parsed as « ’ », in order to harmonize and improve Wikipedia’s typography.
  • Another request would be all circumflexes to be ignored in the searches, which would allow the user to find some secondary terms even without these accents. (Is that stemming?)

My problem is that I don’t know where I should suggest these features, and how to word it properly. (I posted this question in the suitable section but it remained unanswered, so I come here, which seems to be a more crowded place. Face-wink.svg Could anyone help me please? Thanks, --froisois (blabla) 13:20, 19 December 2013 (UTC)

Opening a request at Bugzilla is a good way to get the attention of MediaWiki developers, who would be the people to implement an improvement like this. If that's confusing, I'll be happy to help out. — Scott talk 15:08, 19 December 2013 (UTC)
Hey! I'm actually working on search full time and can answer some of your questions and file or link you to bugs. Before I go into he specific points, I have to quickly explain the state of search at WMF/on WMF wikis: We are currently migrating from an older, unmaintained, custom search system that was pretty advanced for its time (mw:Extension:Lucene-search) to one that realies on a general purpose search backend which receives much more regular maintenance (Elasticsearch). We hope that this will solve some technical problems on our side, let us get some "cheap" features, and allow us to build cool new stuff. As with all rewrites, we're scared of throwing the switch on big wikis. You can check our progress on our project page to find a wiki on which to test our progress. So, given all that, most fix and features will go into the new system and they'll appear once we switch the wiki to the system. Anyway, back to individual points:
  • Apostrophes in full text search: I had a nice short answer here but then thought about it and realized it wasn't right. So long answer it is. Languages are hard. So for English: I aggree and I'll make sure in the new search apostrophes break words (bugzilla:58701). I'm warry to do this for other languages because it might distrupt their analyzers. I believe French in particular has something for elision which needs the '. My best suggestion is to log into frwikisource, sign up for "Nouvelle recherche", search for things, and file bugs for what doesn't work. Feel free to file them like the one earlier in this paragraph.
  • Apostrophes in the page suggester: Because this is a prefix search of the entire article title including spaces I believe the right thing to do here is to cannonicalize all the apostrophes. If we want “L’Oréal” and “L Oréal” to both suggest L'Oréal then we actually need to do something different: map ' and ’ to spaces in the prefix search ((bugzilla:58702). This could create some false positives but I imagine that is worth it. Unlike during full text search, I'm pretty confident this is the right thing to do for all languages.
  • Typography: My remit is limited to search at the moment so I'm going to skip this one. Search will think of them all in the same way for the sake of finding more matches.
  • Ignoring circumflexes: I believe this is currently done for a number of languages including English. If I recall it is done for languages for which it shouldn't be. Anyway in the new system the plan its to do it for just English wikis and to turn it on for any other languages that ask for it. I don't imagine that Vietnamese wikis, for example, will want it. We actually have a lot of control over it so we can normalize just some characters if required. If you have ideas for the right way to do it for French or any other language feel free to file a bug using the same components as the bugs that I filed above.
  • What are Stemming and Lemmatisation: I like this particular source on the subject: [39]. The new search system uses kstem. I'm not sure about other languages but could find out if you are curious.
NEverett (WMF) (talk) 20:05, 19 December 2013 (UTC)
@NEverett (WMF): I couldn’t dream of a more complete answer! Thank you very much. Well, for the full text search, I really think that this feature is anxiously waited by most of French users: when I type “arbre” (“tree”), I expect the search engine to return inter alia “l’arbre” or “l'arbre” (which, by elision, are equivalent to the englishy “le arbre”). And as I type “l’arbre”, I of course want to see all articles which contain “l'arbre”. I even think that false positives can be false false positives (results like “les arbres” are of course wanted). Should I mention it on your existing bug 58701 or do I file a new bug? --froisois (blabla) 20:58, 19 December 2013 (UTC)

Using switch

Hello, I have a template called template:Days for example.

In that template is only:

style="background-color: {{color}}"

Template:color is:

{{#switch: {{{1}}}
|#default                = #739ADE
|yellow =     #CCDDAA
|green = #EFD5A4

How do I get the parser function switch to work?

If I type:


How do I get it to call:

style="background-color: #CCDDAA"

I tried an #if statement like this:

style="background-color: {{#if:{{{color|}}}|{{color}}}}"

and this also didn't work, any suggestions?

Igottheconch (talk) 23:21, 19 December 2013 (UTC)

You need to pass the {{{1}}} parameter through to Template:Color. So Template:Days would need to include something like this:
style="background-color: {{color|{{{1|}}}}}"
Does that work for you? — Mr. Stradivarius ♪ talk ♪ 23:42, 19 December 2013 (UTC)
(edit conflict) Template:Color can't read the parameter passed to Template:Days. Template:Days needs to take the parameter given to it and pass it on to Template:Color, like this:
style="background-color: {{color|{{{1}}}}}"
However, unless Template:Color is going to be used several times in Template:Days or in other templates as well, it would be better and more resource-efficient to scrap it and replace the call to it in Template:Days directly with the switch function, like this:
style="background-color: {{#switch: {{{1}}}
|#default = #739ADE
|yellow = #CCDDAA
|green = #EFD5A4
jcgoble3 (talk) 23:49, 19 December 2013 (UTC)
thank you both so much! barnstars for both. Igottheconch (talk) 02:51, 20 December 2013 (UTC)

list of all accounts, or account names containing a particular string

I'm feeling nosy about another editor's possible use of multiple accounts, and I'd like to see a list of all account names containing a particular string. For example, suppose the string is bar: I'd like to see whether someone has created an account called Foobarbaz, Babar, Barabbas, etc. I'm aware of Special:ListUsers, but I only know how to use it to find account names beginning with a particular string, or to list all account names, 500 at a time: if I searched for bar, it would show me Barabbas but not the other two examples. In other words, I'd like to do a case-insensitive search search of all account names for *bar*.

If I had a text file with a list of all account names, I could grep through that.

I glanced at and but didn't see anything public that looks like a list of account names: I saw "User account data. (private)" which presumably has people's passwords as well.

I don't want to violate the terms of service by spidering Special:ListUsers. Is there a better way? —rybec 04:28, 20 December 2013 (UTC)

On behalf of Betacommand [40] (warning: 221MB file) Legoktm (talk) 04:42, 20 December 2013 (UTC)
Interesting. I am the 8,447,206th user on that list, out of 20,375,529. Thank you for providing me with my useless fact of the day. :) jcgoble3 (talk) 05:16, 20 December 2013 (UTC)
Thank you, Legoktm and Betacommand, for the quick, guru-level answer. I was fully expecting flames. —rybec06:31, 20 December 2013 (UTC)

HTML5 question

So, I've been trying to convert Template:Page tabs to Lua at Module:Page tabs, but I've run into an HTML problem. It would be easy if I was just copying the existing HTML used in the template, but I've also been trying to make the template HTML5-compliant. It is very nearly working, but my equivalent for cellspacing="0" cellpadding="0" has left too much of a gap in the module version. Does anyone know how I could fix this? You don't have to know Lua to answer this - see my sandbox for an example with raw wikitable code and HTML. — Mr. Stradivarius ♪ talk ♪ 14:39, 20 December 2013 (UTC)

I thought we were avoiding work for HTML5-compliant markup, because we would not force all those billions of funky "<br />" to become HTML5 "<br>" and such. It is a little like forcing https-secure protocol when even Google or the U.S. Federal Government did not use https at the Library of Congress, etc. -Wikid77 16:11, 20 December 2013 (UTC)
Avoiding it as an active cleanup step yes (so don't go around changing articles just with that purpose in mind, with a bot or what else). If you are working on the new version of a template, by all means, make it as standards compliant as you can/want to. —TheDJ (talkcontribs) 17:40, 20 December 2013 (UTC)
It is not necessary to alter any <br /> tags to <br>; firstly, because both are valid HTML5; secondly, because Wikipedia pages are pushed through HTML Tidy which alters them back to <br /> again. --Redrose64 (talk) 18:07, 20 December 2013 (UTC)
@Mr. Stradivarius: fixed ! —TheDJ (talkcontribs) 20:10, 20 December 2013 (UTC)
Aha, I forgot to set one of the cell's padding to 0 - makes sense. Thank you! — Mr. Stradivarius ♪ talk ♪ 21:08, 20 December 2013 (UTC)

Searching showing 'suggestions' all of a sudden

I know the default way of displaying is to show those search suggestions, but I was not a fan and turned them off, preferring to keep by own browser-based search history. It was working fine yesterday, but all of a sudden today it's showing suggestions instead. I have no changed any prefs, and I'm using monobook skin and FF 26. ♫ Melodia Chaconne ♫ (talk) 15:51, 20 December 2013 (UTC)

See #AJAX suggestions now mandatory?, a few sections up from here. -- John of Reading (talk) 16:07, 20 December 2013 (UTC)

Displaying a part of the table

I recently removed an outdated portion of a sporting standings table, by the updated but full table. Given that there are 20 teams in a league, the table is displayed in either of two ways -

  • Display a portion of the table (Say only top 5 teams) in the relevant article, which looks better, but requires being updated at 20 locations
  • Display the entire list in every article, making it unecessarily long, but requiring update at a single location.

Considering these two alternatives, I wonder if there is any way to re-organise this current format sere is a way to transclude the full list to display only the relevant portions of a table, using some set parameters? (For example, if the parameters are 1 and 5, then only the top 5 clubs are shown)? I think a hack like that could be useful for usage in many more sporting articles.

In short, all I want is a possible way to display {{2013–14 Premier League table}} so that only a portion of the table is shown, based on how you call the template. Is it possible?

TheOriginalSoni (talk) 23:42, 1 December 2013 (UTC)

Thanks to TheOriginalSoni for inviting me to this discussion. I agree that the best solution would be a centrally updated template with the ability to only show certain sections of the table when the focus of the article is a specific team; after all, there is no point showing the teams at the bottom when the focus is on the top. I would love to know if this is possible, although I'm sure it would take such intricate coding that it may not be worth making such a template when it would need updating every few days (and therefore more chance of breaking the template). – PeeJay 02:39, 2 December 2013 (UTC)
I made a test at {{2013–14 Premier League table/sandbox}} with code choosing which rows to display based on the already passed team parameter, so the season articles don't have to be updated when the team positions change. I had problems controlling pipes and newlines so the source isn't pretty. Here is how it looks with the parameter team=ALL:
Pos Team Pld W D L GF GA GD Pts Qualification or relegation
1 Manchester City (C) 38 27 5 6 102 37 +65 86 2014–15 UEFA Champions League group stage
2 Liverpool 38 26 6 6 101 50 +51 84
3 Chelsea 38 25 7 6 71 27 +44 82
4 Arsenal 38 24 7 7 68 41 +27 79 2014–15 UEFA Champions League play-off round
5 Everton 38 21 9 8 61 39 +22 72 2014–15 UEFA Europa League group stage[a]
6 Tottenham Hotspur 38 21 6 11 55 51 +4 69 2014–15 UEFA Europa League play-off round[a]
7 Manchester United 38 19 7 12 64 43 +21 64
8 Southampton 38 15 11 12 54 46 +8 56
9 Stoke City 38 13 11 14 45 52 −7 50
10 Newcastle United 38 15 4 19 43 59 −16 49
11 Crystal Palace 38 13 6 19 33 48 −15 45
12 Swansea City 38 11 9 18 54 54 0 42
13 West Ham United 38 11 7 20 40 51 −11 40
14 Sunderland 38 10 8 20 41 60 −19 38
15 Aston Villa 38 10 8 20 39 61 −22 38
16 Hull City 38 10 7 21 38 53 −15 37 2014–15 UEFA Europa League third qualifying round[b]
17 West Bromwich Albion 38 7 15 16 43 59 −16 36
18 Norwich City (R) 38 8 9 21 28 62 −34 33 Relegation to the 2014–15 Football League Championship
19 Fulham (R) 38 9 5 24 40 85 −45 32
20 Cardiff City (R) 38 7 9 22 32 74 −42 30
Updated to match(es) played on February 2015. Source: Barclays Premier League Official League Table
Rules for classification: 1) points; 2) goal difference; 3) number of goals scored
(C) Champion; (R) Relegated.
  1. ^ a b Since the winners of 2013–14 FA Cup and 2013–14 Football League Cup (Arsenal and Manchester City) qualified for the 2014–15 UEFA Champions League, the 5th and 6th placed teams qualified for the group stage and the play-off round of 2014–15 UEFA Europa League.
  2. ^ Hull City qualified for the 2014–15 UEFA Europa League third qualifying round as runners-up of the 2013–14 FA Cup since winners Arsenal qualified for 2014–15 UEFA Champions League.
With team=ARS:
Pos Team Pld W D L GF GA GD Pts Qualification or relegation
2 Liverpool 38 26 6 6 101 50 +51 84 2014–15 UEFA Champions League group stage
3 Chelsea 38 25 7 6 71 27 +44 82
4 Arsenal 38 24 7 7 68 41 +27 79 2014–15 UEFA Champions League play-off round
5 Everton 38 21 9 8 61 39 +22 72 2014–15 UEFA Europa League group stage[a]
6 Tottenham Hotspur 38 21 6 11 55 51 +4 69 2014–15 UEFA Europa League play-off round[a]
Updated to match(es) played on February 2015. Source: Barclays Premier League Official League Table
Rules for classification: 1) points; 2) goal difference; 3) number of goals scored
  1. ^ a b Since the winners of 2013–14 FA Cup and 2013–14 Football League Cup (Arsenal and Manchester City) qualified for the 2014–15 UEFA Champions League, the 5th and 6th placed teams qualified for the group stage and the play-off round of 2014–15 UEFA Europa League.
With team=MNU:
Pos Team Pld W D L GF GA GD Pts Qualification or relegation
5 Everton 38 21 9 8 61 39 +22 72 2014–15 UEFA Europa League group stage[a]
6 Tottenham Hotspur 38 21 6 11 55 51 +4 69 2014–15 UEFA Europa League play-off round[a]
7 Manchester United 38 19 7 12 64 43 +21 64
8 Southampton 38 15 11 12 54 46 +8 56
9 Stoke City 38 13 11 14 45 52 −7 50
Updated to match(es) played on February 2015. Source: Barclays Premier League Official League Table
Rules for classification: 1) points; 2) goal difference; 3) number of goals scored
  1. ^ a b Since the winners of 2013–14 FA Cup and 2013–14 Football League Cup (Arsenal and Manchester City) qualified for the 2014–15 UEFA Champions League, the 5th and 6th placed teams qualified for the group stage and the play-off round of 2014–15 UEFA Europa League.
With team=CPA:
Pos Team Pld W D L GF GA GD Pts
9 Stoke City 38 13 11 14 45 52 −7 50
10 Newcastle United 38 15 4 19 43 59 −16 49
11 Crystal Palace 38 13 6 19 33 48 −15 45
12 Swansea City 38 11 9 18 54 54 0 42
13 West Ham United 38 11 7 20 40 51 −11 40
Updated to match(es) played on February 2015. Source: Barclays Premier League Official League Table
Rules for classification: 1) points; 2) goal difference; 3) number of goals scored
You control which rows to display for each team by adding its 3-letter code to a switch statement for each of those rows. When team positions change, those codes have to be manually swapped around in the template. You can choose to display more than 3 rows by adding the team code to more rows. PrimeHunter (talk) 04:41, 2 December 2013 (UTC)
  • PrimeHunter, The formatting of this truncated table is exactly how I expected it to be.
I also created a subpage at {{2013–14 Premier League table/sandbox/1}} with the text "ARS" which I intended to replace at the code's switch expressions. You can see I've replaced one instance of ARS in the code to see that it works.
This way the places do not have to be replace everywhere on the table, making it less likely to break the code, albeit more clumsy. Also, whenever teams change positions, the respective subpages would have to be altered to have the correct rankings.
While this code works, I was kindof hoping for a less hacky of a way. Is there any way to do that?
TheOriginalSoni (talk) 06:09, 2 December 2013 (UTC)
As one (of very few) editing this template so far, I must say I really like this idea of having this table. A reason for having it is to have something that is continuosly updated, which do not happen on all club articles. To make it only show parts would be fantastic, then it would be better for each of the articles. If/when this is done some visual things maybe can be done (yellow color of team unneccesary?) to make it look even better.
My only concern is that it can not be to complicated to update, since that makes a bigger risk for errors. Today a lot of IP edit league table and if we should use it even on the main PL article with team=ALL then there could possibly be a lot of correcting to do. The more articles it is used in, the more editors will try and edit it. But in total I am very positive and I have tried editing in the sandbox above (just show preview, never saved it) and to me it is not that hard updating it, but I could imagine other editors "forgetting" to update the parameters of which rows to be shown. QED237 (talk) 12:03, 2 December 2013 (UTC)
User:PrimeHunter, I love you. This is exactly what I was expecting. Now to sort out some minor user-friendliness issues and it'll be perfect! – PeeJay 12:08, 2 December 2013 (UTC)
Thanks. 20 subtemplates like {{2013–14 Premier League table/sandbox/1}} for team positions would be impractical. I have changed {{2013–14 Premier League table/sandbox}} to get all 20 teams from {{2013–14 Premier League table/p}} by supplying a parameter for the position. PrimeHunter (talk) 13:04, 2 December 2013 (UTC)
  • Looks good to go. Great work, PrimeHunter.
Could we use another less obstrusive colour, like #F5D44E to highlight? TheOriginalSoni (talk) 13:27, 2 December 2013 (UTC)

Might I suggest writing this in Lua? By moving it to Lua you could just input the wins, draws, losses, goals for, and goals against for each team, and then the template could automatically calculate the matches played, the goal difference, and the points. It could also automatically update each team's position in the table, and you could also pass a parameter to limit the number of entries. All the complicated bits could be put in the module, and the template invocation would look something like this:

{{#invoke:league table|premier
|limit = 5
|team1 = [[Arsenal F.C.|Arsenal]]
|team1_wins = 10
|team1_draws = 1
|team1_losses = 2
|team1_for = 27
|team1_against = 10
|team2 = [[Chelsea F.C.|Chelsea]]
|team2_wins = 8
|team2_draws = 3
|team2_losses = 2
|team2_for = 24
|team2_against = 11

If there's any interest in this, I can try and code something up tomorrow. — Mr. Stradivarius ♪ talk ♪ 14:08, 2 December 2013 (UTC)

Mr Stradivarius, that sounds great. I'd like to see how that would look. – PeeJay 14:40, 2 December 2013 (UTC)
I don't edit league tables but it sounds great. You may need more parameters for a general module. Some leagues can give punishments where points are deducted from a team (they can even have a negative number of points). Automatic computation of positions would give complications with some tie rules like La Liga#Competition format. PrimeHunter (talk) 14:53, 2 December 2013 (UTC)

Okay so now there has been no edits for a while, what is the next step? Is there something more to be done? Should I we go to WT:FOOTY to find consensus for instering it on the team articles? QED237 (talk) 00:02, 6 December 2013 (UTC)

Mr. Stradivarius started on Module:League table but hasn't edited it for a week. Should we go ahead with the system in {{2013–14 Premier League table/sandbox}}? PrimeHunter (talk) 03:06, 11 December 2013 (UTC)
Don't delay the deployment on my account - I am planning to make the module quite general, so that it could apply to all leagues and all sports, and getting it right may take some time. That, and I also have a few other Lua projects on the go at the moment. The Lua solution can always be deployed later, when it is ready. — Mr. Stradivarius ♪ talk ♪ 03:58, 11 December 2013 (UTC)

──────────────────────────────────────────────────────────────────────────────────────────────────── I have pinged WikiProject Football about this discussion. If nobody objects in a couple days, I think we should be good to go. TheOriginalSoni (talk) 21:34, 11 December 2013 (UTC)

This is a good idea, I've thought about how to do it myself a couple of times in the recent past. The only suggestion I'd make is that I think having five teams in the cropped league table allows a greater context for the position in the table. this is especially true for teams near the top and the bottom of the table. => Spudgfsh (Text Me!) 21:46, 11 December 2013 (UTC)
As I have said I really support this and I would also say that always showing 5 teams would be the best. If a team is first, second or third show teams 1-5, if a team is fourth show 2-6, fifth 3-7, and so on. QED237 (talk) 11:52, 12 December 2013 (UTC)
Should we have the table both on the main PL article or only on the season articles? Would be nice to have same table everywhere. QED237 (talk) 11:54, 12 December 2013 (UTC)
That would be my preference. it was on 2013-14 in English football last time I looked. => Spudgfsh (Text Me!) 12:01, 12 December 2013 (UTC)
I agree, that would be the best, but I doubt it will be pain free on 2013–14 Premier League and when I inserted this template "full table" on season articles I met some resistance. Some of the arguments was that this template is "ugly" and that the table currently on 2013–14 Premier League looks better. QED237 (talk) 12:06, 12 December 2013 (UTC)

──────────────────────────────────────────────────────────────────────────────────────────────────── To me we should also discuss the yellow color, good or change it? And also should we not show when the table was last updated even on the truncated small tables? QED237 (talk) 12:23, 12 December 2013 (UTC)

I've been been berated for my choice of colours in the past (see here Talk:Doctor_Who_missing_episodes#"1.4 List of lost episodes"). The choice of colour should satisfy WP:Colour. => Spudgfsh (Text Me!) 21:27, 12 December 2013 (UTC)
Personally I would display the whole table so you no not have to go off somewhere else to find the rest of the information, so hope that there is a display all even when a team name is specified. Keith D (talk) 14:02, 12 December 2013 (UTC)
I understand your opinion, but the reason for starting this is that in general editors feel that it is unneccesary to display the whole table. The whole table already exists, but due to the fact that we should not show the whole table it was brought here to see if there was a simple way to show parts of the table. QED237 (talk) 14:32, 12 December 2013 (UTC)
Literally just wanted to stop by and say this is brilliant. I have been waiting for something like this for years but never had the know-how or time to go about doing it myself! Support, support, support. Qed, thanks for bringing this to my attention, and please let me know again if it comes in! Andre666 (talk) 22:09, 12 December 2013 (UTC)
If you should thank someone it is User:TheOriginalSoni who started this discussion. I have never heard of Village pump before, I just pasted the info along to the talkpages of the articles. QED237 (talk) 22:22, 12 December 2013 (UTC)
I have updated {{2013–14 Premier League table}} to use the code in {{2013–14 Premier League table/sandbox}}. It's set to always display 5 teams on team articles and use the highlight color #F5D44E suggested above. I have not added the template to team articles not currently using it. The football editors can do that. PrimeHunter (talk) 01:56, 14 December 2013 (UTC)
PrimeHunter, Sounds perfect. I have made some sample changes to the current template at User:TheOriginalSoni/sandbox2 so the legend at the end and the "Last updated" gets displayed only under certain conditions. Please see User:TheOriginalSoni/sandbox4 to view how each transclusion would call the template. TheOriginalSoni (talk) 04:47, 14 December 2013 (UTC)
Are there any plans to do this for the other similar templates in Category:England football league tables? => Spudgfsh (Text Me!) 10:05, 14 December 2013 (UTC)
I must say I really like what User:TheOriginalSoni has done at User:TheOriginalSoni/sandbox2. If noone opposes I will insert that later. It is great to show the explanations for the lines when neeeded. QED237 (talk) 11:47, 14 December 2013 (UTC)
@Spudgfsh: If you want we now has the bases for doing that to. We have this template code it is just to make a new template and change the teams and make a new table. When this has settled for PL i can try and make that if you want to. QED237 (talk) 11:50, 14 December 2013 (UTC)
I've some free time in the next couple of weeks, I'll do some of the grunt work if you like. => Spudgfsh (Text Me!) 15:15, 14 December 2013 (UTC)
If you want to do it that is great, because I dont really have the time myself. I dont think it will take much time (just a lot of copy and paste) but I could still use the time elsewhere. QED237 (talk) 15:42, 14 December 2013 (UTC)

May I suggest a color change? I think the current cool might be to dark. I have tried looking at snook color contrast to find good colors with good contrast to make the text easy to see according to WP:COLOUR. My suggestion is to use the same light green background color as in the tables in 2013–14 UEFA Champions League which is #CCFFCC. Perhaps the first place line can then change color to gold/yellow instead. QED237 (talk) 12:09, 14 December 2013 (UTC)

Couple of questions from me: should the table, as well as highlighting the chosen team, make the numbers bold for the selected team? I think that's what a lot of us have done before, just wondering if you think it's necessary. Also, should there be a guide as there is in other templates below the table explaining what letters mean? For example: Pld = Matches played; W = Matches won; D = Matches drawn; L = Matches lost; GF = Goals for; GA = Goals against; GD = Goal difference; Pts = Points or similar? Andre666 (talk) 14:20, 14 December 2013 (UTC)

I would say bolding the number would be unneccessary when team is highlighted and it would complicate the coding very much. However the explanation in the bottom might be very good. QED237 (talk) 14:50, 14 December 2013 (UTC)
Sounds good. Also, is there an extra line break in the code or anything? In use it leaves quite a large space below... Andre666 (talk) 16:35, 14 December 2013 (UTC)
Yes I have been asking myself about that large blank space. I think it disappears when inserting info below the table, which I consider to do soon since no editor has opposed yet. QED237 (talk) 22:17, 14 December 2013 (UTC)


Having just updated Template:2012–13 Premier League table to be like this I can say that there needs to be some inline documentation on the current table. It is not obvious that you have to edit Template:2012–13 Premier League table/p to ensure that the correct lines are displayed. It's less of an issue for old tables as there ordering will not change. => Spudgfsh (Text Me!) 16:39, 14 December 2013 (UTC)

Is that why one that I'm using is now showing the full table? Andre666 (talk) 17:01, 14 December 2013 (UTC)
No, someone has edited an old version of the 13/14 template and removed all of the new content. => Spudgfsh (Text Me!) 17:02, 14 December 2013 (UTC)
I added that to the documentation [41] when I made the changes to {{2013–14 Premier League table}}. It should be in the documentation for all templates using this system. I'm not a football editor and others are free to change things without consulting me. PrimeHunter (talk) 17:28, 14 December 2013 (UTC)
That is true and, of course, it should be there but very few editors actually read the documentation before editing the table. This is especially true of the IP editors. only when you use the table will it be obvious that something is has been missed and it's not easy to find if you don't know it's there. hence the inline comment suggestion => Spudgfsh (Text Me!) 17:36, 14 December 2013 (UTC)
Done.[42] I don't plan to monitor the usage but feel free to change the notice if it isn't working. PrimeHunter (talk) 19:39, 14 December 2013 (UTC)

Technical question

What does {{{!}} mean and why is it two of them ahead of "last updated" at the bottom? QED237 (talk) 00:06, 15 December 2013 (UTC)

It says {{!}} and not {{{!}}. It just transcludes Template:! which makes a pipe. I don't know why the original template said || at the start of a line but I simply converted it to {{!}}{{!}}. PrimeHunter (talk) 01:17, 15 December 2013 (UTC)
Okay thanks for the answer. I have seen that "last updated" is not aligned to the left, maybe that has something to do with it? QED237 (talk) 01:30, 15 December 2013 (UTC)

Old table at 2013–14 Premier League

Great job has been done and I just wanted to ask how hard it would be if we instead wanted to do this to the table currently at 2013–14 Premier League (the table below). This is the table on main article so some club season article has used it (parts of it, header, bottom and the team rows wanted) when they did not want to use the entire template since we could not show selected rows before. One of the arguments against the template above has been that is is "ugly" and therefore I ask how hard it will be to apply what we have done above to the table below (and make template of it)? And also if we should use the template above or the one below? Maybe it will be easier to update the one below? So sorry if I bring this up so late after all work people have done, just wanted to ask since I am not very good at this. QED237 (talk) 00:33, 15 December 2013 (UTC)

Qualification or relegation
1 Arsenal 16 11 2 3 33 17 +16 35 2014–15 UEFA Champions League group stage
2 Chelsea 16 10 3 3 32 18 +14 33
3 Manchester City 16 10 2 4 47 18 +29 32
4 Everton 16 8 7 1 27 15 +12 31 2014–15 UEFA Champions League Play-off round
5 Liverpool 15 9 3 3 34 18 +16 30 2014–15 UEFA Europa League play-off round
6 Newcastle United 16 8 3 5 21 22 −1 27
7 Tottenham Hotspur 15 8 3 4 15 16 −1 27
8 Southampton 16 6 6 4 20 15 +5 24
9 Manchester United 15 6 4 5 22 19 +3 22
10 Swansea City 15 5 4 6 21 20 +1 19
11 Aston Villa 15 5 4 6 16 18 −2 19
12 Hull City 16 5 4 7 13 19 −6 19
13 Stoke City 16 4 6 6 15 20 −5 18
14 Cardiff City 16 4 5 7 12 22 −10 17
15 Norwich City 15 5 2 8 14 28 −14 17
16 West Bromwich Albion 16 3 6 7 17 22 −5 15
17 West Ham United 16 3 5 8 13 19 −6 14
18 Crystal Palace 16 4 1 11 11 24 −13 13 Relegation to 2014–15 Football League Championship
19 Fulham 16 4 1 11 15 30 −15 13
20 Sunderland 16 2 3 11 12 30 −18 9

Updated to games played on 14 December 2013.
Source: Barclays Premier League Official League Table
Rules for classification: 1) points; 2) goal difference; 3) number of goals scored
(C) = Champion; (R) = Relegated; (P) = Promoted; (E) = Eliminated; (O) = Play-off winner; (A) = Advances to a further round.
Only applicable when the season is not finished:
(Q) = Qualified to the phase of tournament indicated; (TQ) = Qualified to tournament, but not yet to the particular phase indicated; (RQ) = Qualified to the relegation tournament indicated; (DQ) = Disqualified from tournament.

QED237 (talk) 00:33, 15 December 2013 (UTC)

And example from 2013–14 Arsenal F.C. season‎:
Qualification or relegation
1 Arsenal 16 11 2 3 33 17 +16 35 2014–15 UEFA Champions League group stage
2 Chelsea 16 10 3 3 32 18 +14 33
3 Manchester City 16 10 2 4 47 18 +29 32
4 Everton 16 8 7 1 27 15 +12 31 2014–15 UEFA Champions League Play-off round
5 Liverpool 15 9 3 3 34 18 +16 30 2014–15 UEFA Europa League play-off round

Updated to games played on 14 December 2013.
Source: Barclays Premier League
Rules for classification: 1) points; 2) goal difference; 3) number of goals scored
(C) = Champion; (R) = Relegated; (P) = Promoted; (E) = Eliminated; (O) = Play-off winner; (A) = Advances to a further round.
Only applicable when the season is not finished:
(Q) = Qualified to the phase of tournament indicated; (TQ) = Qualified to tournament, but not yet to the particular phase indicated; (RQ) = Qualified to the relegation tournament indicated; (DQ) = Disqualified from tournament.

QED237 (talk) 00:35, 15 December 2013 (UTC)

  • I had the same idea just sometime back. It shouldn't be particularly difficult, I think though probably a bit hectic. Overall it will end up saving time for update, so we might as well give it a shot. TheOriginalSoni (talk) 03:09, 15 December 2013 (UTC)
I'll take a look at it and edit the sandbox for the current season template. It shouldn't be too hard to convert the current template to this style. It won't be until later today or tomorrow until I try.=> Spudgfsh (Text Me!) 09:29, 15 December 2013 (UTC)
There is no rush, I just wanted to put what was on my mind out there and hear if it could be done. It sounds really good. QED237 (talk) 12:19, 15 December 2013 (UTC)
I have had a first stab on the sandbox but there is a slight issue with regards to extra spacing (as you can see above). The only difference I've made is to make each qualification etc column entry a single row rather than a couple of rows (it wouldn't take too much to fix but I've run out of brainpower). what do you think? => Spudgfsh (Text Me!) 20:48, 15 December 2013 (UTC)
I have tried looking at it but had no luck at all so far. QED237 (talk) 19:28, 17 December 2013 (UTC)
All my experience tells me it should be a simple, but everything I've tried causes more spaces or breaks it completely => Spudgfsh (Text Me!) 19:38, 17 December 2013 (UTC)
Where is this "extra spacing" to which you allude? --Redrose64 (talk) 20:29, 17 December 2013 (UTC)
You can see it in this discussion above. It is okay in full table. For Arsenal it is big spacing for Everton, and then above table and on Swansea. For bottom team there is larger space above table. QED237 (talk) 20:36, 17 December 2013 (UTC)

────────────────────────────────────────────────────────────────────────────────────────────────────I was using my sandbox to see all options at once. it only looks ok for the full table. => Spudgfsh (Text Me!) 20:40, 17 December 2013 (UTC)

I got the impression that it was one of the two in this section that showed extra space. Apparently not; but I do see big spaces in my sandbox. I haven't looked at the code, but I'm pretty sure that it's going to concern conditional statements (such as {{#if:}} etc.) with newlines both before and after. --Redrose64 (talk) 21:32, 17 December 2013 (UTC)
Right,  Done, see here. The trick is to prevent all newlines except when you really need them - that is, before each {{Fb cl team}} and {{Fb cl2 qr}} - and for those, put a soft space at the end of the preceding line. This soft space needs to be a character entity &#32; otherwise the MediaWiki parser strips it again, and so strips the newline as well. --Redrose64 (talk) 01:04, 18 December 2013 (UTC)
Awesome, now it is really great! Thank you! Now the only question remaining is what table to ues (most likely this last one?) And if we should highlight the team in question or not since there already are background colors in the table? And if not should we instead see if we can bold the team in question? QED237 (talk) 01:12, 18 December 2013 (UTC)
And also should the vte-buttons be somewhere to make the template easy to access? Perhaps below table? QED237 (talk) 01:14, 18 December 2013 (UTC)
I knew there had to be an elegant solution. I'd put the vte buttons in the header of the table if it looked ok but otherwise they would be better at the bottom. => Spudgfsh (Text Me!) 10:16, 18 December 2013 (UTC)
The problem is that it is a template header and it would remove "pos" so I think bottom is the best. I will see if I can remember how it is done. The risk with the vte button is the bigger risk of unexperienced users and IP editors interested in destroying the table can find it. Or perhaps they want to make a good edit but accidentally "destroy" something. QED237 (talk) 12:14, 18 December 2013 (UTC)
Now placed on best spot I could find (bottom below text). QED237 (talk) 12:47, 18 December 2013 (UTC)

Highlighting or bolding

The last question (I think) before all this amazing accomplishment is finished is highlightning of teams. Should we hihlight the teams at all? Should we still use the brown/orange color to highlight teams even when the table now has different background colors for qualification/relegation? Or is it possible to make the team names bolded instead? QED237 (talk) 12:57, 18 December 2013 (UTC)

The purpose of the template when the highlighting is applied is to limit the table to the teams around the one selected. In the for mid table teams it's clearly the middle row of the table that is important but for teams at the top or bottom it can be less clear. It all boils down to whether the context the table is being used in is enough to highlight the important row or not.
When it comes to using bold to highlight a row, it's not a good idea as it is dependent on how the users are set up as to whether they see the highlighting. I've been gently nudged away from using it in the past and now think they were right. => Spudgfsh (Text Me!) 15:03, 18 December 2013 (UTC)
I am not a nativfe english speaker so do you mean not to highlight or bold? I think when teams are on team article they know what team they re looking for anyway. To me I would say to not mess with the colors (overwrite) that already is there for qual/rel and I have seen editors removing the bolding to. QED237 (talk) 15:56, 18 December 2013 (UTC)
The issue tends to be with using bold text. => Spudgfsh (Text Me!) 16:08, 18 December 2013 (UTC)
I took the liberty to "gather" the links to champions league and relegation to make it look like the original table. That is to make champions league span over the first three rows and so on. I finally succeded and the result can be seen at User:Spudgfsh/sandbox. Someone with good coding experience may want to look at it, what I did was that if rows 1-5 is shown I use rows=3 for the template and if rows 2-6 is shown (team has pos 4) i use rows=2 and so on with some if statements and switches. This made the highlighting look bad so I suggest it being removed? If no one opposes I will remove it later and then I think it can be used in the articles. Or would you rather keep the highlighting and all the relegation and qualification on separate rows? QED237 (talk) 22:59, 18 December 2013 (UTC)
  • I just came here to say that the current sandbox version looks perfect, looks wise. We could reduce the subtext at the end, but otherwise it looks great. Also, for teams ranked 2,3 19 or 20, the qualification/relegation section should be yellow, not the original red/green. TheOriginalSoni (talk) 13:12, 19 December 2013 (UTC)
That is unfotunately a problem I can not solve. The color of the qualification/relegation section will always be the same as the first team (team ranked 1 and 17). That is due to the way Template:Fb cl2 qr and Template:Fb cl team work so I dont think there is much that can be done there. One way is to remove highlighning everywhere or else return to separate rows for qualification/relegation section.QED237 (talk) 13:37, 19 December 2013 (UTC)
Pos Qualification or relegation
2 2014–15 UEFA Champions League group stage
Sorted, with these edits, see User:Spudgfsh/sandbox.
What is somewhat less easy (but is still possible at the expense of rather greater complication, see demo at right) is to make the "2014–15 UEFA Champions League group stage" and "Relegation to 2014–15 Football League Championship" cells so that the yellow is the height of one row for the full width. This example assumes that the third row is highlighted, and omits all columns except the first and last --Redrose64 (talk) 21:13, 19 December 2013 (UTC)
In my opinion I would prefer what you say is possible with highlighting one row rather then the whole "qualification column" unless it is very complicated and will affect the updating of the table. I think it should be the way you describe above or having it as it was a few days ago with rows=1 everywhere (the qualification text was written three times). This highlighting seems as the best option especially since I had no luck bolding the entire row, which should be the aleternative to this yellow highlighting. QED237 (talk) 21:34, 19 December 2013 (UTC)
This idea of mine to have a two-colour background for the multi-row cells in the "Qualification or relegation" column will only work if there is an odd number of rows to be spanned by that cell (i.e. a table of five rows 1-5, 3-7, 14-18 or 16-20). Where the number of rows spanned by that cell is even (i.e. five rows 2-6 or 15-19), the "2014–15 UEFA Champions League group stage" and "Relegation to 2014–15 Football League Championship" links cannot be vertically centred - they would need to be displaced up or down. This is because in my demo at right, the apparent three-row cell is in fact three single-row cells with the intermediate borders suppressed; so an apparent two-row cell would need to be constructed as two single-row cells. So I think we'd better forget about that idea. --Redrose64 (talk) 23:09, 19 December 2013 (UTC)
Okay I totally understand. Then there is not much left to do/decide before putting it into use. Either we keep the template as it is, or we use single rows for qualification and relegation to get the whole row highlighted. The last alternative is to not highlight at all (but it seems that editor want some sort of highlighting or bolding of the team row). This since there is no way to bold an entire row (of the specified team). QED237 (talk) 23:16, 19 December 2013 (UTC)
Have a look at the Spurs row now... --Redrose64 (talk) 23:26, 19 December 2013 (UTC)
Wow, I would never have thought of that. And I guess we can use that in if-statements as well. So now we need to find consensus wheter to use 1. Bolding, 2. Highlighting as it is now, 3. Highlightning with single qual rows, 4. Nothing to show team. QED237 (talk) 23:34, 19 December 2013 (UTC)
There is a small problem in that if rows 1 or 18 are bolded, the "2014–15 UEFA Champions League group stage" and "Relegation to 2014–15 Football League Championship" links - which are level with the next row down - will also be bolded. To get those links bolded when rows 2, 3, 19 or 20 are bolded means that similar extra code needs to be added to the |bc= parameter of the {{Fb cl2 qr}} and {{Fb cl3 qr}}. --Redrose64 (talk) 23:49, 19 December 2013 (UTC)
Okay I see, Not to sure exactly how you mean (my english not the best) but it seems solve-able with the same type of switch as for the background color for qualification and relegation. QED237 (talk) 23:57, 19 December 2013 (UTC)
I'd just like to say, if a team is specified when calling the template, we should highlight that team when the template is displayed. This should be done by bolding the entire row that team occupies in the table, including their position, games played/won/drawn/lost, goals for/against and points. Just my opinion anyway. – PeeJay 18:22, 19 December 2013 (UTC)
@PeeJay2K3:I know everyone do not agrre if we should bold or not (on some articles this has been edited back and forth) but I can agree with you. The problem is I dont know at the moment how to solve this, but when I have some time over, I can give it a try. QED237 (talk) 20:13, 19 December 2013 (UTC)
Just wanted to drop by to say I'm very happy with the current template, and yes I think we should highlight the team selected just so it's as clear as possible. Question though: is the 'Qualification or relegation' column going to be there even when no teams in the five shown are qualifying for anything, or will it be hidden? The latter would be preferable IMO. Andre666 (talk) 18:47, 19 December 2013 (UTC)
@Andre666: Solved with a very long switch (perhaps not the best solution, but I think it works). QED237 (talk) 20:11, 19 December 2013 (UTC)
While the code looks a bit complicated, as long as the editors only move the {{Fb cl team}} lines it will remain maintainable. Well done all. => Spudgfsh (Text Me!) 21:52, 19 December 2013 (UTC)

I am now taking this one last time (due to low amount of "voting")to WT:FOOTY to decide if we should use

  1. Background color
  2. Bolding
  3. Background color and bolding
  4. Separate the right qualification column
  5. Nothing

to show the team in question. Then when editor have had their say (together with this discussion), I will instert the template and start using it. QED237 (talk) 11:14, 20 December 2013 (UTC)

Conclusion (part of table)

The discussion on WT:FOOTY resulted in the use of bolding, since WP:ACCESS forbids using color to identify list item ("Especially, do not use colored text or background unless its status is also indicated using another method") and also many would like bolding anyway since we should always show the relegation and qualification color and not hide it with our yellow.

If no one has something to add I will insert the sandbox we worked on to the template later tonight or tomorrow before the matches and see how it works tomorrow when matches are played. Thank you everyone for all the work that has been done. Great job!. QED237 (talk) 21:24, 20 December 2013 (UTC)

  • Qed237 For the teams in the top 5, there are too many colours in the background for bolding to be an effective way to highlight the team. I prefer highlighting, but since that may not be an option, we might want to change the background colour for the teams to something like c1ebeb or ccffff (for team 5), ccffcc (for team 4) and b9ffb9 (for team 1-3). I tried using the suggested colours at User:TheOriginalSoni/sandbox4 but it probably doesn't look that good, so we might want to explore some other colours. Maybe we bold and italicise the chosen team? TheOriginalSoni (talk) 03:21, 21 December 2013 (UTC)
About the change of colors I doubt it will "work". The current color have been figured out pretty well so that they are not to dark but it still must be a difference in the green between group stage and qualification. It would be to hard to find better and change them. In you suggestion they are very close to eachother. The italics part I am not sure of, it was more of "we should have bolding" and that was what was used on the articles before this template. The thing is that we should not need to mark the team at all, it should be mentioned in text above the table and if any we should bold that row. QED237 (talk) 10:52, 21 December 2013 (UTC)

So what about the Championship, League One and League Two tables? Are these going to be updated with the new format as well? Will they even be updated at all as they haven't been done this week??? — Preceding unsigned comment added by (talk) 21:14, 22 December 2013 (UTC)

I've updated a version of the championship table in its sandbox {{{:2013–14 Football League Championship table/sandbox}}}. the results can be seen at the bottom of my sandbox (User:Spudgfsh/sandbox). there's still something not quite right with it (i'll work it out in the morning if someone else hasn't by then) => Spudgfsh (Text Me!) 23:19, 22 December 2013 (UTC)
I've got a couple of extra spaces in the code somewhere but I don't have the brainpower for that kind of debugging now. feel free to fix or copy to the real template => Spudgfsh (Text Me!) 23:44, 22 December 2013 (UTC)
Championship has been done and is now in place in the season article => Spudgfsh (Text Me!) 10:05, 23 December 2013 (UTC)
League one done and in the article. It currently displays the note about Coventry all the time but there is no easy way of turning it off when they are not displayed. any thoughts? => Spudgfsh (Text Me!) 15:03, 23 December 2013 (UTC)
YesY top four English leagues done for this season. => Spudgfsh (Text Me!) 16:30, 23 December 2013 (UTC)
Great. About the coventry note a switch could be used but the problem is that when teams change place in the table, the switch must also be change. QED237 (talk) 16:56, 23 December 2013 (UTC)
That was the problem I had. It can be added at the end of the season without issue but it might just have to remain as it is at the moment.=> Spudgfsh (Text Me!) 17:00, 23 December 2013 (UTC)

Feedback: Don't suppose there's any way of restricting the six lines of boilerplate beneath the table section to relevant stuff such as abbreviations actually used in the table section, as-of date and reference? This stuff is obviously needed with the full league table in the relevant league season article, but beneath a five-team subsection in a club season page, several lines of irrelevant stuff in a font so small it's MoS-non-compliant really is intrusive.

And I'd have thought the V/T/E links shouldn't be included in the transcluded part? cheers, Struway2 (talk) 12:45, 23 December 2013 (UTC)

You definately have some good poinys there. The info has the same size as the table. At the moment I would say the top 2 or 3 lines is useful. However this is controlled by Template:Fb cl footer that we should not change so it is nothing we can do, unless we choose to remove all of the info. I think the parameter "season_over=yes" removes the last two rows so that may be used in a if statement I guess.

About V/T/E my thoughts has always been that a template should be accessable were it is used, but perhaps it is unneccesary from club articles, I dont know. QED237 (talk) 14:35, 23 December 2013 (UTC)

I thought the v/t/e was for navigation templates? These aren't navigation templates. cheers, Struway2 (talk) 17:13, 23 December 2013 (UTC)
Okay I did not know it was only for navigation templates? Where is that info? Now Christmas but could perhaps look at it later. If you have the info, please give it to me. QED237 (talk) 00:31, 24 December 2013 (UTC)

Watchlist key

How do I get rid of it? I've been here nearly 8 years, I know what a minor edit is... GiantSnowman 19:55, 12 December 2013 (UTC)

Add ".mw-changeslist-legend {display: none;}" to your common.css.--v/r - TP 20:00, 12 December 2013 (UTC)
Is there no intention of adding an option to Special:Preferences#mw-prefsection-watchlist? GiantSnowman 20:03, 12 December 2013 (UTC)
I have no idea, I didn't even notice it until you said something.--v/r - TP 20:05, 12 December 2013 (UTC)
Doubtful. I'd say a gadget would be easy enough to make, but chances are that by the time someone wants to hide it they've probably been around long enough to be able to find WP:VPT and to edit their common.css once they get the answer. ;) Anomie 20:24, 12 December 2013 (UTC)
Could a "Dismiss" button be added here? John Reaves 16:04, 13 December 2013 (UTC)
I too would have liked a "Dismiss" button. That list is quite annoying. Exactly where should I add ".mw-changeslist-legend {display: none;}"? I didn't quite understand the bit about common.css. Manxruler (talk) 16:42, 13 December 2013 (UTC)
@Manxruler:: User:Manxruler/common.css--v/r - TP 16:44, 13 December 2013 (UTC)
Thanks. That fixed it. :) Manxruler (talk) 16:51, 13 December 2013 (UTC)

The new cluttered looking legend on Special:Watchlist

I'm not sure who added the new cluttered looking legend box in the #mw-header-options section of Special:Watchlist, but I think not only is it incomplete and may result in lot of new questions on the WP:HD and WP:TEAHOUSE specifically asking what (-625) numbers mean. I also think it looks extremely cluttered and ugly. I've added:

    display: none;

to my common.css to hide that section for me (and you can too if you agree with me). Since I'm not sure who else to ask, Andre, would you happen to know who's idea it was to add that and point them to this post so we can discuss some better options for it? Thanks. Technical 13 (talk) 15:23, 13 December 2013 (UTC)

Bugzilla: 52005 -- WOSlinker (talk) 15:36, 13 December 2013 (UTC)
I like it. I think there should be a link to some expanded info on a Help: page, which would extend its usefulness beyond merely duplicating the abbr tooltip text. Otherwise I think it's a good idea and the design seems tasteful to me, not cluttered. I think it may just take some a bit of getting used to. A dynamic minimize feature with a persistent state might help though (but it should be continually "expand"-able in case new tags show up that people are confused by). equazcion 15:37, 13 Dec 2013 (UTC)
I would say that the need for a legend is a proof of our failure to adapt the watch list in general to something that can be used by novices and en.wp superusers alike. —TheDJ (talkcontribs) 15:43, 13 December 2013 (UTC)
I agree, but until we do manage to, this is definitely a useful addition. — Scott talk 20:10, 13 December 2013 (UTC)
The English Wikipedia links to Help:Watching pages in MediaWiki:Watchlist-details which is displayed at top of the watchlist. Looking at there is no interface message in the legend box which is well suited for a link to a help page, but it could be stuffed into one of the explanations like MediaWiki:Recentchanges-label-plusminus. PrimeHunter (talk) 16:26, 13 December 2013 (UTC)
WOSlinker was faster than me providing a link to the corresponding ticket in Bugzilla. :) Thanks! --AKlapper (WMF) (talk) 16:22, 13 December 2013 (UTC)
I think it wouldn't be unreasonable to add a new message to that box (empty by default) that could be filled with additional information. Matma Rex talk 19:42, 13 December 2013 (UTC)
I like the new watchlist legend. I expect to develop a solid case of banner blindness around it in a few days, but it reminds me of how long it took me to figure out what "(top)" meant (now "current"; I wasn't the only person who was confused), and I can see that it might be useful for less experienced editors. WhatamIdoing (talk) 19:57, 13 December 2013 (UTC)

I just submitted for review a patch that would make the box collapsible (collapse state would be persistent, like with the tables of contents): . Matma Rex talk 22:11, 13 December 2013 (UTC)

An old-timer, I don't at all mind the legend on my real computer with wide screen. My Dell XPX-10 tablet is much smaller, and yes it makes clutter there. Jim.henderson (talk) 22:17, 13 December 2013 (UTC)

As a laptop user, I found quite unobtrusive; in fact I didn't really even notice it until this thread pointed it out to me. This is definitely going to help the newer folks to the watchlist, but someone did say the wording could be shortened. I'm also wondering, since someone pointed it out in the Bugzilla ticket discussion, would this feature make the tooltips for each of the symbols obsolete, and would they be deprecated in the future? TeleComNasSprVen (talkcontribs) 07:40, 15 December 2013 (UTC)

Speaking of which, why isn't there a single MediaWiki message containing the whole text for the legend somewhere, and just a scattering of various MediaWiki messages for each line? That way, we can change the MediaWiki message however we want specific to this project, and any of the more CSS/JS savvy users out there can base their code off the MediaWiki message if they so choose (say set to
{display: none;}
for example). TeleComNasSprVen (talkcontribs) 07:46, 15 December 2013 (UTC)
Technical 13 has already shown how to hide the whole legend box. The box is a textbook example of an association list: a <dl>...</dl> enclosing several <dt>...</dt><dd>...</dd> pairs, used properly (unlike the colon markup for discussion page threads, which is an improper use of the association list - and not just because we don't provide <dt>...</dt> elements). Most of the <dd>...</dd> elements have neither class nor id; however, if an individual element has no classes or ids of its own, but is enclosed or preceded by elements that do have classes or ids, it's possible to construct a selector that pinpoints the one that has neither class nor id. Thus, if you wanted a yellow background for the text "This edit was performed by a bot" (but not the b itself), that text is given by the HTML <dd>This edit was performed by a bot</dd>, which has neither class nor id; but the immediately preceding element is <dt class="botedit">b</dt> which does have a class, so you would use
dt.botedit + dd { background: yellow; }
Extending that, here are the selectors (plus an empty declaration block) for both items on each individual row:
dt.newpage, dt.newpage + dd {  } /* N This edit created a new page (also see list of new pages) */
dt.minoredit, dt.minoredit + dd {  } /* m This is a minor edit */
dt.botedit, dt.botedit + dd {  } /* b This edit was performed by a bot */
dt.wikibase-edit, dt.wikibase-edit + dd {  } /* D Wikidata edit */, + dd {  } /* (±123) The page size changed by this number of bytes */
Notice how each row has two simple selectors, separated by a comma: the first works on the boldface letter; the second on the text definition. The declaration blocks are empty, but you can insert any CSS that you like, such as good old display: none;. If you can determine a CSS selector for an element, you can write javascript that targets that same element. --Redrose64 (talk) 09:10, 15 December 2013 (UTC)
Ah thanks Redrose. I had completely forgotten Technical 13's suggestion applied to the whole box. I normally just use "View source" on my web browser and look for the classes and ids to modify in CSS, but it was always a mystery to me how to find the ones that did not have classes and ids. Thanks again, this is new information to me. TeleComNasSprVen (talkcontribs) 10:03, 15 December 2013 (UTC)

It needs to be made clear that it is in fact a legend, and what it's a legend for, as many watchlists will not show any NmbD flags. When I (a seasoned editor) first saw it, it took me a while to figure out what it actually was. Rwxrwxrwx (talk) 23:43, 15 December 2013 (UTC)

How about improving the legend by adding a link to a page describing what actually goes under the "Wikidata edit" flag? I'd say that edit flag is the most confusing. — Dsimic (talk) 22:05, 19 December 2013 (UTC)
That text comes from MediaWiki:Wikibase-rc-wikibase-edit-title. If we can decide which page it should link to, I'm pretty sure that we could link it, because MediaWiki:Recentchanges-legend-newpage has a link, so it's not one of those MediaWiki: pages where only plain text is permitted. --Redrose64 (talk) 00:04, 20 December 2013 (UTC)
Thank you for a great insight! Hmm, maybe a link to WP:Wikidata would be good enough? — Dsimic (talk) 00:21, 20 December 2013 (UTC)
I have added [43] an explanation to Help:Watching pages. I think that makes my below heading suggestion "Legend:" sufficient. PrimeHunter (talk) 00:38, 20 December 2013 (UTC)
Adding link to Help:Watching pages somewhere within the legend box sounds really useful to me! Though, I'd prefer something like "Legend: (help)", as it should be much more intuitive, and it's also somewhat inline with the layout of default user signatures... Thoughts? — Dsimic (talk) 01:25, 20 December 2013 (UTC)
I have changed MediaWiki:Recentchanges-legend-heading from "Legend:" to "Legend: (help)". The message is displayed at both Special:RecentChanges and Special:Watchlist so I made it a section link to Help:Watching pages#How to read a watchlist which explains the legend in more detail and is helpful for both special pages (the watchlist is just a filter of recent changes). PrimeHunter (talk) 17:49, 21 December 2013 (UTC)
Awesome, thank you very much! Looks great, and it's all quite intuitive this way. — Dsimic (talk) 01:09, 22 December 2013 (UTC)


The legend has just been given a little "Legend:" heading and an option to collapse it (that should stick between page reloads). Thanks for the input (and for not being too mean this once ;) ). Matma Rex talk 19:31, 19 December 2013 (UTC)

Thanks. Since the English Wikipedia has a relevant help page, I suggest we change the new heading in Mediawiki:Recentchanges-legend-heading from "Legend:" to "Legend:". The message is also displayed when the box is collapsed so it should be short. PrimeHunter (talk) 19:50, 19 December 2013 (UTC)

Draft namespace being enabled soon

This is advance notice that, pending final code review, we're planning to enable the new Draft namespace on Tuesday, December 17th (in the afternoon PST).

The basic functionality of the new namespace is outlined in our technical requirements. There are also future ideas that our team plans to build and/or test. If you want to try it out now, you can test the new namespace on the Wikimedia Labs test wiki for English Wikipedia (RecentChanges in the Draft/Draft talk namespace there).

I view this as not the end of the work on Drafts namespace, but rather just the beginning. I'm hoping that in the new year, many of the open questions about drafts –how categories should work, whether to allow creation of drafts if a page already exists, and more– can be answered through simple discussion and some A/B testing. Please see the List of open questions I've posted for discussion.

I think our next step is to figure out how to help new editors get their drafts reviewed and published in a timely manner. There is still a huge backlog at Articles for Creation –tens of thousands of old stale drafts, thousands more at the front of the queue– and the new namespace as it will be launched won't solve this problem. We'll need to build better tools, experiment, and try some new approaches if we're going to simply shifting the backlog.

Learn more

Steven Walling (WMF) • talk 23:04, 12 December 2013 (UTC)

It seems like there's a word missing near the end of the last sentence... do you mean something like, "if we're going to AVOID simply shifting the backlog"? Neutron (talk) 00:01, 14 December 2013 (UTC)
  • Good news on the ETA of the Draft namespace but I'm not so sure that the AfC project would actually welcome any further Foundation involvement in the development of the new 'Draft' mainspace for AfC purposes, especially as I am still confused about the repeated statements by one staff member that AfC is absolutely not within the WMF remit. Let us not forget too easily that such backlogs (and those at NPP) are due partly (or even mainly) to the fact that there is still no proper landing page for new users which IMO is a much more serious and pressing priority. The rejection of WP:ACTRIAL (emphasis on trial) demonstrated that something in this direction clearly needs to be done - and not after another two years of tinkering with stats first to prove the necessity. Page Curation was offered as a consolation but although it is a brilliant piece of engineering it does not address the core issues.
My suggestion that the New Pages Feed/Curation tool could be cloned and adapted to the feeds to AfC have been neatly side-stepped. Technically it would be possible, but because it would need assistance from the Foundation to implement it, there would be the usual re-debate in the office over whether the community really knows what it wants. Kudpung กุดผึ้ง (talk) 20:54, 14 December 2013 (UTC)
Perhaps a bot to move the pages from AFC to the Draft namespace might be in order.--v/r - TP 00:14, 13 December 2013 (UTC)
I think that's partially up the AFC project members to decide. Personally, I'd recommend against moving them en masse right now. If we move articles immediately, it might make it harder to get through the backlog, since we haven't figured out how AFC Helper Script or a fork of it would work on Drafts namespace. Instead, I think we should begin by testing recommending the draft namespace to new authors, and then doing reviews in a much more lightweight way, without the complex system of scripts, bots, and templates AFC employs right now. AFC was designed for a specific use case, and I think that we probably don't want the replicate some of the worst practices right away, like putting review comments and templates in to the drafts directly. Steven Walling (WMF) • talk 00:39, 13 December 2013 (UTC)
The helper script should be rewritten, or at least given a substantial cleansing. It's become a convoluted mess as numerous developers have passed it along, and I think it would benefit quite a bit from some direction -- style guides, using built-in MediaWiki and jQuery functionality rather than weird hacks, etc. With that said, I won't have time to seriously look into this until (after) the holidays. Theopolisme (talk) 02:14, 13 December 2013 (UTC)
If we want to add a simple interface to more quickly publish articles that are ready, then that's something we could take care of on my team. Adding additional buttons to the edit interface on only drafts (e.g. adding a "publish" button) is something that is much easier now that there will be separate namespace. Steven Walling (WMF) • talk 02:25, 13 December 2013 (UTC)
Indeed it will, and that's something I'm definitely excited about for the future. It's more about the extensive feature requirements outlined at Wikipedia:AFCH#Reviewing_an_article_submission that I'm "concerned" about -- and it will require more discussion by AFC folks before decisions can be made. Regardless, thank you so much for helping the Draft namespace come to fruition! Theopolisme (talk) 03:21, 13 December 2013 (UTC)

Steven (WMF) Please don't use hyperbole when facts will do. Category:G13_eligible_AfC_submissions has ~25k submissions that are eligible for G13, but we're slowly working our way through in terms of delaying the deletion or deleting that there is no deadline on (we've had these for a long time and we've already cleaned up ~3 years (50k submissions) in the 3 months that my bot has been on the task). The Pending submissions backlog is ~2k right now and could use some loving to get it down to a more reasonable number. Using hyperbole only causes FUD and ill will against the AfC process/project. Hasteur (talk) 18:34, 13 December 2013 (UTC)

Steven (WMF) What's the namespace ID going to be for this new namespace (per WP:NS)? What would the mapping from the current structure to the new namespace look like (i.e. Wikipedia talk:Articles for Creation/FoObAz => Drafts:FoObAz)? Hasteur (talk) 18:38, 13 December 2013 (UTC)

Per bugzilla:57315#c45, it looks like the ID for Draft: will be 118, and Draft talk: will be 119. equazcion 20:33, 13 Dec 2013 (UTC)
@Steven and equazcion (Steven (WMF)Equazcion): Just want to be sure so I can start thinking about HasteurBot tasks (G13 checking, Replacing "Drafts missing a template" checks, etc.). Hasteur (talk) 19:57, 14 December 2013 (UTC)

Great! Another namespace. More bureaucracy. Debresser (talk) 17:11, 14 December 2013 (UTC)

@Debresser: Not sure if you participated in the original RFC, but just in case you didn't see it or the WP:DRAFTS guide... using the draft namespace will be entirely optional for logged in editors. If you want to go on creating articles just like you have before, you can do that. Steven Walling (WMF) • talk 21:20, 14 December 2013 (UTC)

PST dates in WMF edits

I often despair of WMF edits - If they don't follow the rules, why do they expect others to do so? Let us just take the top line "Tuesday, December 17th (in the afternoon PST)". Please read WP:DATEFORMAT and please use UTC - thanks. Arjayay (talk) 20:19, 13 December 2013 (UTC)

@Arjayay: Umm, what? This strikes me as an extremely petty thing to complain about, and that guideline clearly only applies to articles anyway (if we're trying to out-lawyer each other…). You're seemingly trying to create problems where there aren't any. Matma Rex talk 22:10, 13 December 2013 (UTC)
equazcion is correct about the namespace IDs. And Arjayay: the guideline you cite is part of the Manual of Style, which (as a Wikipedian editing since 2006) I know quite well. It is not, however, a rule about announcements or conversation in general, so I don't really understand why you'd be hostile and snipe at me about it. Steven Walling (WMF) • talk 21:55, 13 December 2013 (UTC)
I must agree that applying that guideline to talk page posts is slightly daft. However, using non-UTC dates in public announcements is slightly inconsiderate. A lot of editors outside the US don't know what the US time zones are, whereas UTC is more universally understood (and is the default time zone for this wiki).
So no one else needs to look it up, PST's afternoon is after 8pm UTC. – PartTimeGnome (talk | contribs) 00:01, 14 December 2013 (UTC)
I hear you about it being a custom that's more considerate. Noted for the future. Steven Walling (WMF) • talk 00:17, 14 December 2013 (UTC)
There are valid concerns about WMF practices. This is not one of them and it belittles the concerns that actually do matter. The MOS applies to content pages. It would be a bit more considerate if they used UTC, but saying this puts you in despair is a bit... dramatic. Killiondude (talk) 00:25, 14 December 2013 (UTC)

Page creation by IP editors

I understand from the bugzilla thread that the Draft namespace is now enabled on the en Beta Wikipedia. I went there (following the link from bugzilla) and attempted to create a new page in the namespace while not logged in. I then registered (apparently Beta does not use the single-sign-on feature) and tried again to create a page. Perhaps Beta has some different means of page creation that i missed? But if the Draft ns does not permit page creation by IP editors, then it lacks one of the most basic requirements specified in the document you link to above. DES (talk) 22:09, 13 December 2013 (UTC)

It should be configured the same as English Wikipedia. In any case, as you can see from examples like this, the Draft namespace will definitely permit creation of drafts by anonymous editors. Steven Walling (WMF) • talk 22:59, 13 December 2013 (UTC)
  • I don't think it's absolutely necessary to move the existing submissions to the new Draft mainspace. I can imagine them staying as they are until the backlogs are cleared, while all new submissions go to the Draft namespace. Kudpung กุดผึ้ง (talk) 21:00, 14 December 2013 (UTC)
At least we should not add the current backlog of G13-applivcable submissions. That would still leave us 6 months under the old system, and the process of working them through will last at least another 6 months after that until all the revisions are dealt with. I'm not really happy with that. But the virtue of not entering them (as Kudpung suggests) is that the additional time of organizing the necessary moves and programming might well take a considerable time itself, and delay figuring out what to do with the new submissions. It will also let us correct the errors and necessary reconsiderations in the new system while there are still only a small number of items. .Perhaps in say 6 or 7 or 8 months when the active submissions under the old system are down to a manageable number, we could then reconsider moving them. DGG ( talk ) 22:58, 14 December 2013 (UTC)
I commit to both you DGG and Kudpung to begin modifying my bot code to support the new namespace and prefix (G13 nudging/nominating, "Submissions missing a submission template" tasks) Hasteur (talk) 01:58, 15 December 2013 (UTC)
Thanks Hasteur. I think there may be other considerations as well. For instance, we probably don't actually want review templates or comments on drafts directly, now that we have Talk pages. Also: per conversation on WT:CSD, you can see that it is unclear whether G13 will apply to all drafts, or just those editors choose to submit to AFC. Steven Walling (WMF) • talk 04:19, 15 December 2013 (UTC)
I don't think there's anything close to a consensus to expand G13 to non-AfC drafts, and I would argue that until the Draft namespace has been going for the half year it would take for G13 to even be applicable, we probably shouldn't be trying to guess what the namespace is going to look like, or how best to perform housekeeping maintenance there. VanIsaacWS Vexcontribs 05:38, 15 December 2013 (UTC)
Stevan, there is an advantage to comments on the draft pages, where they will be seen by reviews making later comments, as well as anyone who stops by to improve the article. There is also an advantage in having them on the user talk pages, where they will be sure to see them if they ever log in again. Since WP is not Paper, we can put ithem in both places--which is what I have been doing manually if I think the ed. is still around. But this is the sort of detail that should not be hard-coded into the system, but that we here can modify as needed. There are dozens more such. The WMF shouldn't have to get involved in them. DGG ( talk ) 06:18, 15 December 2013 (UTC)
@Vanisaac: The CSD G-criteria "apply to all namespaces" although some criteria have namespace exemptions: G2, for example, "does not apply to pages in the user namespace". G13 is the only one of the G-criteria that excludes by implication, not by explicit statement. It's something of an anomaly. --Redrose64 (talk) 11:17, 15 December 2013 (UTC)
Hmm, that's interesting. My memory of the original closure was that any AfC, no matter where it was located, was valid. There seems to be a bit of exclusive verbiage right now. But like I said, until we've had a draft namespace for 6 months, it's irrelevant, although I would support opening up the current language to ensure that AfCs qualify no matter where they are hosted. VanIsaacWS Vexcontribs 11:29, 15 December 2013 (UTC)
Vanisaac Redrose64 I forget where I saw it, but there was a discussion on the CSD talk page that endorsed the viewpoint that pages outside the current AfC prefix were also eligible for G13 as long as they had been marked with the AfC template. Up to now I've been very conservative in applying the rules(only in the AfC prefix), but I think it's time to extend the rules to match the community's consensus. Hasteur (talk) 18:38, 15 December 2013 (UTC)
Amended the above post, because {{U|Vanisaac|Redrose64}} is equivalent to [[User:Vanisaac|Redrose64]] which is not what you intended, I'm sure. Anyway, I think that you mean WT:CSD#Wikipedia:Drafts. Any change to the CSD criteria should not be discussed here, but at WT:CSD (with a courtesy note to WP:VPP). --Redrose64 (talk) 20:31, 15 December 2013 (UTC)


It looks like we're all set to enable this in a few hours. I'll post again when it's complete. Steven Walling (WMF) • talk 21:51, 17 December 2013 (UTC)

Yep, as you can see with pages like Draft:Example and its talk page, we're  Done for now. Steven Walling (WMF) • talk 00:33, 18 December 2013 (UTC)

No, If I click on I don't get told about Draft:John Wainwright Evans at all! Josh Parris 09:29, 18 December 2013 (UTC)
The English Wikipedia can test for the draft page and choose a message in MediaWiki:Newarticletext. Make suggestions at MediaWiki talk:Newarticletext. PrimeHunter (talk) 10:17, 18 December 2013 (UTC)
Josh Parris I attempted to make the case that we needed to solve the lack of draft information on Search and redlinks before launching, but people yelled and screamed bloody murder at me. I'm happy to help fix that going forward. You can see some of our design concepts for this in the PDF linked from here. Steven Walling (WMF) • talk 18:40, 18 December 2013 (UTC)

I have a wide variety of draft articles in my userspace, in various stages of completion (see, e.g., User:BD2412/End of history, User:BD2412/World domination, User:BD2412/List of churches in New York City, User:BD2412/Charles F. Lettow, User:BD2412/J. Bratton Davis United States Bankruptcy Courthouse), as well as sections of my sandbox that I would like to turn into drafts. Some of these are super-stubby; some of them are copied from public domain sources and need to be adapted for encyclopedic use. At what point can I move these to draft space, and will it be appropriate to move all of them there? bd2412 T 14:12, 20 December 2013 (UTC)

That's not a technical question but a matter of practices at the English Wikipedia. See Wikipedia:Drafts for what we have now, but this is new and may change. You can move them now, later or never. It's optional. The main issue for established users may be whether they want others to be more aware of a draft and maybe edit it, or they prefer to keep it more private. PrimeHunter (talk) 22:11, 20 December 2013 (UTC)
As you can see from looking at list of drafts that there's plenty being created there now. I think you have userspace drafts you want to make draft namespace pages, there's nothing stopping you from moving your drafts there. Steven Walling (WMF) • talk 23:07, 20 December 2013 (UTC)
Thanks, I wasn't sure if we were waiting for some protocol to be enacted. I will go ahead with these. Cheers! bd2412 T 00:59, 22 December 2013 (UTC)

Other uses?

Will this be possible:
To be used for Template Sandbox and Test Case pages? and thus in some way linked to live templates?
And for scratchspace pages for collaborative rewrites of live articles without working on the actual live article, while a rewrite is in progress? (which currently occur on talk subpages? and necessary for some processes such as CopyVio), and thus in some way linked to live pages?
-- (talk) 01:26, 19 December 2013 (UTC)
Honestly I don't think so. The purpose of the namespace is for drafts of articles. Pages are, roughly speaking, meant to be temporary and eventually either moved to mainspace or deleted. Template sandboxes and test pages should probably stay subpages, since they're meant to be around permanently. Steven Walling (WMF) • talk 23:12, 20 December 2013 (UTC)
It might be interesting to ask bot operators to test new bot tasks in the draft namespace before approving them to run in mainspace. Thoughts? GoingBatty (talk) 14:28, 21 December 2013 (UTC)
While this may be possible in some cases, it doesn't seem to me to be viable as a general rule. Many bot tasks are cleaning up issues in existing articles (which may not be present in the Draft namespace) or are performing tasks outside of the article namespace. Anomie 15:37, 21 December 2013 (UTC)

Search (monobook)

I'm not sure if it's related to the AJAX issue up above, but here goes: When I search for something in the search bar, and AJAX (I think) auto-suggests the relevant search results for me; I hit enter on the result and it immediately brings up the page. I hit backspace, and my browser is suddenly redirected either back to the page I searched for, or to the a blank search menu. Is this a problem just on my end? Has this been reported on Bugzilla yet and if so can I take a look at the ticket for it? TeleComNasSprVen (talkcontribs) 02:06, 21 December 2013 (UTC)

See here: is that the same as what you observe? --Redrose64 (talk) 14:18, 21 December 2013 (UTC)
Based on Redrose64's earlier description, this seems to be a bug in Firefox. I'm guessing you use Firefox too? I'm not sure there's a way we could work around it. I don't know if anyone's logged this on MediaWiki's or Firefox's Bugzilla ( – PartTimeGnome (talk | contribs) 17:50, 21 December 2013 (UTC)
Yes, thanks Redrose64 for the link. It's an exact description of what I was experiencing. I'm not too familiar with Mozilla's open-source community to report a bug to them as well as I would to WMF, I would think. TeleComNasSprVen (talkcontribs) 10:52, 22 December 2013 (UTC)

undone vandalism still being displayed -- template caching problem?

At Spanish Armada, the first line is being displayed as

{do your homework quicker{About|the Spanish fleet that sailed against England in 1588|the Armada Española, the modern navy of Spain|Spanish Navy}}

The revision history shows that this is from a vandalizing anonymous edit, which was immediately undone; nevertheless the vandalized version is still being displayed (both in the live article and in the permalink version). Joriki (talk) 00:01, 22 December 2013 (UTC)

I added ?action=purge to the URL and pressed Enter. That fixed the problem. Johnuniq (talk) 00:04, 22 December 2013 (UTC)
  • Beware browser cache copies and delays: It was not a template-related delay. For several months/years, some people have noticed how a "Save page" operation would occasionally redisplay the prior revision, even though a view-history confirmed the update had been saved. The timing of cache-update operations, versus saving a new revision, involves some complex problems of interleaving the prior cache-copy with a browser's cache files. Hence, using option "?action=purge" will help to update the cached page and ensure a new cache-copy is also written into the browser cache files. The whole situation can be so confusing, and we probably need an essay to identify all the related issues. Plus, imagine the horror if a new user saves a long, complex edit-session and sees all the changes omitted if the prior revision is redisplayed after edit. Also note how a 60-second timeout error can lose all changes, upon wp:Wikimedia Foundation error, unless reformatted within a 75-second grace period (or such), which will eventually post the saved revision for view-history but perhaps 6-8 minutes after Save-page.... -Wikid77 06:51, 22 December 2013 (UTC)
This crops up a lot, particularly with ClueBot NG's vandalism reverts. The most recent report I know of was last month, and there were a bunch more the month before that. I'd suggest adding this to the FAQ, but the first FAQ entry already covers it. – PartTimeGnome (talk | contribs) 22:34, 22 December 2013 (UTC)

Wikipedia Adventure message substitutions

Hey folks! The javascript API calls in The Wikipedia Adventure substitute messages like WP:TWA/Welcome. I'm trying to get the static {current date} and {basepagename} (username) to appear in the messages on userpages. I've been using 'noinclude' tags, but these aren't working. It was suggested to change the code a bit, curious what you think:

MediaWiki:Guidedtour-tour-twa1.js change in sendMessage function with var text = "{" + "{" + "subst" + ":" + msgPage + "}" + "}"

Thoughts? Cheers and thanks! Ocaasi t | c 18:17, 22 December 2013 (UTC)

Draft Namespace: Automatic warning

I'd like to propose that there be an automatic header on every page in the draft namespace making it clear that it's a draft and not a real article. That or, alternative and a bit more drastic idea, the Wikipedia background be changed from white to an orange color for the draft namespace.--v/r - TP 18:28, 22 December 2013 (UTC)

This would be better at Wikipedia talk:Drafts where many other things about drafts are discussed. PrimeHunter (talk) 19:25, 22 December 2013 (UTC)

Afd and robots.txt

This google search is, for me, popping Wikipedia:Articles_for_deletion/Michael_Steele_credit_report_incident in the results. It was my understanding from our robots.txt and google policy that shouldn't happen. NE Ent 18:56, 22 December 2013 (UTC)

Google indexed the url Our robots.txt has an entry for the %3A encoding by disallowing /wiki/Wikipedia%3AArticles_for_deletion/, but the %2F encoding is not caught. Our normal url at Wikipedia:Articles_for_deletion/Michael_Steele_credit_report_incident is without any of the two encodings. We could reduce such encoding loopholes by adding {{NOINDEX}} to certain templates, possibly only when a {{PAGENAME}} check shows it's one of the pages we don't want indexed. PrimeHunter (talk) 19:16, 22 December 2013 (UTC)
(edit conflict) Our robots.txt excludes pages beginning with "/wiki/Wikipedia:Articles_for_deletion/" an "/wiki/Wikipedia%3AArticles_for_deletion/". But it doesn't exclude pages beginning with "/wiki/Wikipedia%3AArticles_for_deletion%2F", as the link included in that search does. Anomie 19:17, 22 December 2013 (UTC)


I've been having a problem where changes in a transcluded page aren't showing up for a while. (the GA review page at Talk:Drexler–Smalley debate on molecular nanotechnology. When I changed browsers and even computers I'm still getting an older version on the talk page. But going to the transcluded page directly shows the updated one.

Is it Wikipedia or the person's browser that "builds" the page with the transcluded content? And either way, how could they be showing 2 different versions of the transcluded page? Is one working off of an older copy rather than the actual page? Thanks. North8000 (talk) 19:20, 22 December 2013 (UTC)

It is due to caching on the server side. (The page with the transcluded content is built by the Wikimedia servers.) If you want to force the page to update you can make a null edit. Ruslik_Zero 19:26, 22 December 2013 (UTC)
It's Wikipedia but there can be a long delay. See Wikipedia:Purge. PrimeHunter (talk) 19:28, 22 December 2013 (UTC)
Thanks! North8000 (talk) 20:00, 22 December 2013 (UTC)

Purging pages when updating template

Hi. There is a template, Template:2013–14 Premier League table‎, with subtemplate Template:2013–14 Premier League table/p‎. This template is currently being used on 21 pages for the simple reason that this table is updated directly and before the tables on some of the articles were outdated. The problem now is that even if the template is updated, the pages affected by this change is not updated until someone makes an edit on that page. This means that the table is just as outdated as it was before (until someone makes an edit to that page).

I tried looking around for solution and on Help:Template I found the information "When a template is changed (when the template or one of its subtemplates is edited), the change will be reflected on all pages on which the template is transcluded. However the change may not become visible on all pages immediately – a previously cached version of a page, based on the previous version of the template, may continue to be displayed for some time. Use the purge function to force a page to be displayed using the latest versions of templates. (This includes the template page itself, if it contains usage examples.)". Therefore I went to WP:PURGE that informs about Template:Purge that can be added on pages to purge them and update all content on that page.

Now I would like to know if anyone know a way to solve this and update all the pages where Template:2013–14 Premier League table‎ is being used? Is it possible to somehow purge all pages when the template is updated (when pressing save)? Perhaps some link or something can be pressed on the template paqe to purge all of the pages?

I would appreciate all answers. Thank you! QED237 (talk) 20:28, 22 December 2013 (UTC)

Pages using the template should automatically be updated by the job queue at some time, but that time varies a lot (in my experience from seconds to days, and sometimes months when there are problems). I don't think there is any way to automatically jump the job queue for edits to a specific template. The articles shouldn't have purge links, but the template could have noincluded manual purge links for the articles. For example "Liverpool (purge)". PrimeHunter (talk) 20:53, 22 December 2013 (UTC)
Okay, I understand. I guess it takes a lot of time at the moment because when I looked at the pages they were not updated so I made "null edits". From what I could see and understand the job queue was very long (4.6 million) now. I will consider purge-links in the template if it will continue to be this slow updating. It is not possible to make a "purge all"-link to purge all pages? QED237 (talk) 21:39, 22 December 2013 (UTC)
The API can purge multiple pages at once. For example, the following link will purge three pages if clicked: [44]. To purge more pages, add their titles to the end of that URL, with %7C between each one. (The link will only work for logged-in users.) – PartTimeGnome (talk | contribs) 22:53, 22 December 2013 (UTC)
Thanks, I will try that next time. QED237 (talk) 23:21, 22 December 2013 (UTC)
@PrimeHunter:@PartTimeGnome:There was still a few pages that had not been updated since the matches about 7 hours ago so I used the link and inserted all 21 pages and it worked! Now I am thinking about inserting it in the template (in the noinclude info) with something like "If the pages with this template dont update (might take a while), try and purge with this link (will only work for logged-in users).". Thanks to both of you for the information and help! QED237 (talk) 00:20, 23 December 2013 (UTC)
For the template itself I would just add something like {{purge|refresh}}. We don't want to bother readers too much with the technical ins and outs of how the pages get updated. We can leave your API purge link on the template documentation instead, so that users who update the template can use after they finish making their edits. Also, I believe it is possible to make an API link that will update all pages transcluded from Template:2013–14 Premier League table‎ without having to list each page individually. I'm not good enough with the API to know how to make it, though - could someone who knows the API better confirm whether this is possible or not? — Mr. Stradivarius ♪ talk ♪ 13:29, 23 December 2013 (UTC)
@Mr. Stradivarius: not sure what you are looking for but this link purges all the pages but there is probably a better solution. QED237 (talk) 15:22, 23 December 2013 (UTC)

Scrolling back to top in File Upload Wizard?

After the File Upload Wizard form is filled out and "upload locally" button is clicked, the form collapses leaving basically an empty screen in front of the person uploading a file. It would be really great if the page could scroll to the top after the form is submitted, so people can see what's happening with the upload process. This way, looking at an empty screen is quite confusing. Thoughts? — Dsimic (talk) 23:56, 22 December 2013 (UTC)

With which browser and version does this happen, on which exact page (URL)? --AKlapper (WMF) (talk) 17:07, 23 December 2013 (UTC)
I've experienced the described behavior with a few versions of Firefox (24–26), and the URL is, reached through Hope that helps. — Dsimic (talk) 23:41, 23 December 2013 (UTC)

Tech News: 2013-51

08:25, 23 December 2013 (UTC)

Geohack forbidden?

I'm trying to use Geohack and it keeps coming up with Error 403 Forbidden. Simply south...... eating lexicological sandwiches for just 7 years 15:05, 23 December 2013 (UTC)

It works for me, for example at top of Paris. Please give the url which doesn't work and say where you got it. PrimeHunter (talk) 16:36, 23 December 2013 (UTC)
Tool labs was down earlier due to an XFS issue, which might have caused such errors. Anomie 18:04, 23 December 2013 (UTC)

GeoGroupTemplate maps acting up is the result of going to a page in my userspace and clicking the GeoGroupTemplate, just like it should be. Edit the URL to change "Athens" to the name of another page in the same group, e.g. "Perry", and you get the map from User:Nyttend/Ohio NRHP/Perry. Until a few days ago, you could click the box on the right side of the map (below "Traffic \/") and select both Athens and Perry, so the markers for both would appear at the same time. Suddenly it's not doing that; you can only toggle on/off the one that's current selected, so going to Perry gets rid of the Athens locations. Have we changed something? This was working normally until a week or so ago, long after Google made the big change to its API; I haven't seen any changes by Google that would cause problems here. Nyttend (talk) 21:26, 23 December 2013 (UTC)

PS, I was afraid that Google had surreptitiously gotten rid of this feature totally, but that's not the case. If you change the name in the bar between "Google" and the magnifying glass (where you type the name of the place you want to go, if looking at a blank map), it continues to work as I expected it. Nyttend (talk) 21:28, 23 December 2013 (UTC)

Open all pages in edit mode

Hello. How can I open all pages of a list of pages to edit mode? I read Wikipedia:Linky tutorial but I couldn't find understand. Is there another way? Xaris333 (talk) 02:16, 24 December 2013 (UTC)

Hi. Could you please clarify? What kind of list? You could run JS like
if (mw.config.get('wgPageName')=="Special:BlankPage/list") {
  var list = prompt("Paste list of page names").split('\n');
  for (var i = 0; i < list.length; i++) {
    open(""+encodeURIComponent(list[i].replace(/ /g,'_')));
and go to Special:BlankPage/list and paste in a list of page names separated by newlines. I really don't know what you mean, so I can't be more specific. Why don't you understand the Linky tutorial? See Wikipedia:Linky_tutorial#Optimize_your_speed_using_Linky-lists. πr2 (tc) 18:34, 24 December 2013 (UTC)

IBM Watson as fact checker- REST API access

Would it be possible to write a software tool that analyses an article and makes a list of the stated facts and feed this facts to the REST API from IBM Watson ? so Watson does cross reference checking for us ? and if it returns a negative the tool can post a line on the talkpage of the article ? [65] Mion (talk) 02:20, 24 December 2013 (UTC)

As a design note, if the speed of analysis is high enough it might be better to go from an article level to edit level so the facts in the contribution to the article can be checked in realtime, as a test case for the plugin we could let Watson check all edits and reverts for facts of the Reviewer group.Mion (talk) 03:21, 24 December 2013 (UTC)

Odd font problems

For the past several weeks I've had the problem that when I open a page, often the title, headers, image captions, content in tables, and even italicised text will be in a different serif font. It goes away to the normal font after I click a couple times but it's getting very annoying, especially the inconsistency on page. Anyone know what's causing this and how to get it to stop? Thanks, Reywas92Talk 21:20, 24 December 2013 (UTC)

This is a bug in Google Chrome: the wub "?!" 21:55, 24 December 2013 (UTC)

Amy Hānaialiʻi Gilliom

The title of the page Amy Hanaialiʻi Gilliom is incorrect, as it is missing a macron. I attempted to fix this using {{DISPLAYTITLE}} but that didn't work. I tried to perform a page move but it threw up an error message (something about non-Latin script). Out of curiosity I tried to create a new page called Amy Hānaialiʻi Gilliom and it worked! I basically copy/pasted the contents from the original article, including the talk page. Could somebody please complete the job by merging the two pages (ie. copy the history over to the new page and delete the old, or whatever)? If there's some reason not to have the page renamed then please just delete the new page instead. Thanks in advance.

Also, as soon as I created the page I got a message on my talk page about copyright violation. I know nothing about the subject of the article, so won't be editing the page. Perhaps someone would be kind enough to add/remove the necessary templates so that the baby doesn't get thrown out with the bathwater? Regards, nagualdesign (talk) 07:32, 18 December 2013 (UTC)

The copyvio notice is because a bot found the "new" article on a mirror site at its "old" name. I've marked the "new" article for speedy deletion per WP:CSD#A10. If you wish the article to be renamed, please don't cut-and-paste, but take it to WP:RM. --Redrose64 (talk) 13:00, 18 December 2013 (UTC)
I speedied the "new" article but don't want to do the move as sources seem mixed and use both versions. I've also had to revert text to a quite old version as there was an obvious copyvio from Mountain Apple. Dpmuk (talk) 15:16, 18 December 2013 (UTC)
Well I have moved it, but anyone is free to move back over the redirect. Self identification has that "ā". Graeme Bartlett (talk) 20:12, 18 December 2013 (UTC)
Thank you very much, Graeme. It seems that you're the only person who understood the request and was willing and able to sort the problem out, rather than just pasting templates here and there. Sorry for any trouble caused. All the best for the holidays.
Regarding the copyvio, I don't think MTV are claiming any sort of copyright. It clearly says at the bottom of the page, "Source: Wikipedia" "Text from this biography licensed under creative commons license". nagualdesign (talk) 22:49, 18 December 2013 (UTC)
The copyvio isn't from MTV, but rather from Mountain Apple and as the editor who added the information says in the edit summary that it's from Mountain Apple it's a pretty clear cut case of copyvio. I reverted all of that edit as, although I could find a source for only half of it, it seems likely that it's all a copyvio given that it was added by the same editor.
As for the naming issue there are one or two places on her own website that uses "a" rather than "ā", for example under Amy Live on the main page and other sites also showed a bit of a split so it wasn't even completely clear to me what she self-identifies as (even if it did seem likely that it was "ā"). Given that I'd have liked to seen more of a case for the move, but, at the same time, think it's very likely that the article would have ended up at this title so have no objection to the move. Dpmuk (talk) 00:47, 25 December 2013 (UTC)
Good enough for me. nagualdesign (talk) 01:15, 25 December 2013 (UTC)

Image render problem?

File:QuincyMineNo2Shafthouse.jpg, on Wikipedia, looks all blue and pink like the colors are inverted. However, downloading it and opening in an image viewer, the colors are correct. Currently on Firefox on linux. Anyone else see this, or know what might be happening? Chris857 (talk) 14:39, 18 December 2013 (UTC)

Looks ok on IE on Windows 7... what page is this on? --Mdann52talk to me! 14:48, 18 December 2013 (UTC)
It looks bad on the file description page, direct link, and every article it appears in. Chris857 (talk) 14:55, 18 December 2013 (UTC)
Looks fine under Windows XP SP3, using: Chrome 32; Firefox 25.0.1; IE 8; Opera 12.16; Safari 5.1.7. --Redrose64 (talk) 17:16, 18 December 2013 (UTC)
When I get home from work, I'll see how it looks there. Chris857 (talk) 17:18, 18 December 2013 (UTC)
Well, looks fine on WinXP SP3, Chrome, here at home. Chris857 (talk) 03:20, 19 December 2013 (UTC)
Hmm, the image has an icc profile of sRGB, so it shouldn't be a problem with colour management. As a test, you might want to see how your browser handles colour managed files [66] just to see if something funky is going on in your web browser. Bawolff (talk) 05:18, 19 December 2013 (UTC)
In Windows XP I've produced all three possible outcomes, demonstrating that it's certainly browser-dependent. Opera and Safari "supports these ICC version 4 and version 2 profiles"; Chrome and Firefox 25 "supports ICC version 2 profiles only"; but IE 8 "does not support these ICC profiles". --Redrose64 (talk) 12:23, 19 December 2013 (UTC)
Firefox 3.6.6 also "supports ICC version 2 profiles only". --Redrose64 (talk) 11:59, 26 December 2013 (UTC)

Just reformat every article every day

The delay for reformatting pages in the wp:Job_queue(s) has stretched into weeks, and I think it would be much faster, now, to just reformat all 4.4 million articles every day. Perhaps leave talk-pages or template /doc pages in the job queues. A check of job-queue size during the past hour showed a backlog of 8.2 million jobs, being processed at over 30 per second, or 108,000 per hour (30×60 seconds × 60 minutes per hour).

Okay, so here's the deal: the Lua version of Template:Convert was deployed 10 days ago, at 02:15, 11 December 2013, and it has not yet finished reformatting all the 554,000 pages which use {convert}. How do we know? We can edit many unreformatted pages and still see all subtemplates linked from the markup-based Convert of 24 June 2013, which will delink upon edit-preview. However, the queue speed runtime should be only 554,000/108,000 per hour = 5.13 hours to process 554,000 jobs in a job queue. Well, 11 days later is 264 hours (11×24), and that would be 51x times slower than 5.13 hours (as 264/5.13) to process 554,000 pages. Or 14-day reformat would be 65x times slower. Are there any plans to just reformat and re-cache every article every day? -Wikid77 (talk) 08:38, 20 December 2013 (UTC)

Sorry, what's the edit-preview again? I heard it's supposed to help debug Lua, but given that I returned and missed all the kerfuffle about when Lua was introduced... TeleComNasSprVen (talkcontribs) 10:03, 20 December 2013 (UTC)
You can see which templates are transcluded in a page by edit the page, and look at the bottom of the page. If the templates has changed, like {{convert}} and the page hasnt been updated by the jobqueue, then you will see the old templates at the bottom. A Null edit will update the page, and if you press edit again, the correct templates are in the list. I think "edit-preview" is when you press "show preview" when you edit an article – Christian75 (talk) 10:55, 20 December 2013 (UTC)
Where do you have the numbers from? Module:Math (1.3 millions transclution) has been changed recently too. The old template:Convert had a lot of subpages (which was changed without the job queue had to update all pages which used template:convert), and the new module is one large one (every page has to be updated if there is a spelling error or another minor edit). The same story goes for a lot of other modules. Christian75 (talk) 10:55, 20 December 2013 (UTC)
Job queue is about update the linklist, not about updating the page content shown. What you could be seeing is an old linklist (update job still in the queue), while the actual page content is updated already (because it was thrown out of the cache). Wikid77, what did the page show before the edit started (or in preview)? Old format (old templte code used) or current format (new template code used)? -DePiep (talk) 11:10, 20 December 2013 (UTC)
The page's cache-copy has also shown old results, while Edit has shown old links, as in page "USS_Portland_(LSD-37)" having old/new fathom conversion:
  • {convert/old |2621|fathom} -   {{convert/old|2621|fathom}}
  • {convert[lua]|2621|fathom} -   {{convert/q|2621|fathom}}
The page cache-copy still had old result "15,730 ft" but edit-preview shows new Lua "15,726 ft". It would be great if the page cache updated within 24 hours with relinks later, but both were delayed over 10 days, as if the 2 jobs (for refreshLinks2 and htmlCacheUpdate) were always queued near each other. -Wikid77 18:23, 20 December 2013 (UTC)
The job queue invalidates the pages (htmlcacheupdate), and update the links (refreshlinks/refreslinks2). It has to parse the page (expand all templates) before it can update the link table. But I didnt say it changed the actual content. I (tried) to say that you can see which templates are used on the page at the bottom of the edit window. A null edit update the links too, and the links used on the page (at the bottom of the edit window) will change (if the templates has been changed, and it hasnt been refreshed yet). Christian75 (talk) 11:34, 20 December 2013 (UTC)
I actually was replying to Wikid77, who seems to think that the page content itself is oudated (because and when the linklist below, in editmode, is). I tried to note that the linklist may be outdated, but the page content is not.
Re Christian75, 11:34: "job queue invalidates the pages" - are you sure? As I read it, the cache is invalidated at the edit save moment (which makes sense). At the same moment, a job is added to the JQ (symchronous). Later, asynchronous, coming from the JQ the job updates the linklist. -DePiep (talk) 15:04, 20 December 2013 (UTC)
Yes, if you save, lets say, {{convert}}, the server has to invalidate 550000+ pages (all pages which transclude the template). The last line of the section you qouted says " But if there are more than about 1000 to do, it takes a long time. By default, jobs are added when more than 500 pages need to be invalidated, one job per 500 operations." (the page you save is invalidated) Christian75 (talk) 15:16, 20 December 2013 (UTC)
{ec} Ch75, you're right, it says so. But ... but ... that means then while page X is in a JQ job for a few days, readers are served up the old version, each view request? Hard to believe, and did not experience that over here. -DePiep (talk) 16:57, 20 December 2013 (UTC)
I have been looking at the job queue over the last few days while wondering whether switching to Module:Convert broke anything. Does anyone know why WP:JQ gives a link with the note "graphs are available, but are broken as of 2013"? Will that be fixed? I have used the API for "jobs", and admired the graphs at ganglia (for the "global queue in all Wikimedia projects"). The queue seemed to rise to around 7 million before Module:Convert was switched on, and steadily reduced over a few days, then quickly cleared to under 0.5 million recently, but something hammered it at 2013-12-19 19:00, and it rose to over 10 million in about an hour. Any guesses for what was responsible? I'm just curious, because the system seems to be working, although as Wikid77 says, it's taking quite a long time for the pages with {{convert}} to settle down. The module and the template have not changed since going live at 2013-12-11 02:15, and I cannot see any impact from switching to Module:Convert in the ganglia graphs. — Preceding unsigned comment added by Johnuniq (talkcontribs) 11:19, 20 December 2013 (UTC)
The job queue is a little strange. Because an edit to any page (with at least 500 transclusion or what the mediawiki configuration is at wikipedia) will add two jobs to the job queue (one refreshLinks2 and one htmlCacheUpdate). When the job queue runs a refreshLinks2 it will expand to lot of refreshLinks2 with 500 links in each. When the job queue runs the new refreshLinks2 it will expand to 500 refreshLinks. And finally when the queue runs refreshLinks it will update the links on the page. (htmlCacheUpdate expands to htmlCacheUpdate with 500 pages in each) Christian75 (talk) 11:34, 20 December 2013 (UTC)

Is this question/issue related to the delays in articles being added to categories when templates in those articles are modified in a way that changes the template's category? For example, changes were made to Module:Citation/CS1 on November 3, 2013, and articles are still being added to CS1 error categories (without any edits taking place) as a result of these changes (47 days later, by my calculation). I don't know enough of the back-end mechanics to understand why it takes 47 days for an article to receive what is essentially a null edit. To give a specific example: {{PBB/2069}}, which has not been edited since May 12, just appeared in Category:Pages containing cite templates with deprecated parameters today, based on the changes made to the CS1 module on November 3.

Anything like this that takes 47+ days in 2013 is not working right, as far as I can tell. Here's a link to the changes that were made: Help_talk:Citation_Style_1#Update_to_the_live_CS1_module_week_of_2013-11-03. Thanks for any insight you can provide. – Jonesey95 (talk) 00:42, 21 December 2013 (UTC)

That is exactly the same problem, and so I say demote the 3(?) wp:JQ queues and just trigger reformatting all 4.4 million articles once every day (or 3 days), rather than re-re-re-reformatting thousands of navbox pages over and over and over. In fact, if {convert} triggers 554,000 pages to reformat in 14 days, then Module:Citation/CS1, in 2.2 million pages, should reformat in 56 days (2,200/554×14), at the current queue speeds. -Wikid77 (talk) 09:51, 21 December 2013 (UTC)
It will take long time to update all articles. A better idea is to start write the modules so all pages doesnt have to be updated every time a minor detail is changed. Like the module:portal which was rewritten so all 5 million pages didnt have to be updated every time a project changed its image... (but still if you change portals starting with b, it has to update 1.5 million transclution). Christian75 (talk) 18:52, 22 December 2013 (UTC)


Reformat major pages every day

Well, if "[i]t will take long time to update [reformat] all articles" then perhaps just reformat the major pages daily (or 3-day periods). I am concerned, when a template is changed, how some major articles still keep the outdated cache page for weeks (or months). Just like false claims of "every page is often vandalized" (not true, many go years with no vandalism), some people think major articles are edited daily, or weekly, but instead many major topics remain unedited for months, awaiting further work from a busy volunteer distracted by fiddling with stubs which are read a few times per week, when a page read 50x per day also needs rework. Perhaps all major articles should link a dummy template which could be revised, every 3 days, to trigger wp:job_queue relisting, to ensure those pages are reformatted, sooner, with results of recent template updates. -Wikid77 (talk) 07:45, 27 December 2013 (UTC)

Help extract gps coordinates

I would like to add coordinates to the article Federal University of Ceará, here is the google link:[67], I tried {{coord|42|17|47|S|56|51|36|W|source:ptwiki_region:PT_scale:1000_type:landmark|display=title}} but it didn`t work. --Hanyangprofessor2 (talk) 08:38, 20 December 2013 (UTC)

Well, that seems to work as "{{coord|42|17|47|S|56|51|36|W|...}}" but what problem did you have? Did you want to show a map of Brazil? -Wikid77 09:56, 20 December 2013 (UTC)
@Hanyangprofessor2: I've successfully added coordinates to the article, using exactly the same code that you give above, unchanged. It seems to work fine. Looking at the revision history for that page, your name doesn't appear against any edit for the last eight years - perhaps you didn't save your change, or maybe you clicked the Save page button, but there was some error? --Redrose64 (talk) 20:14, 20 December 2013 (UTC)
@User:Wikid77, User:Redrose64: The problem is that the GPS coordinates we thought we extracted (I was doing it with some students of mine) are off; they locate the university few hundreds km from where it should be - in the sea :) Google Maps shows the correct location. So - why are our GPS coordinates off? What are the correct ones, and how do you get them? --Piotr Konieczny aka Prokonsul Piotrus| reply here 03:03, 21 December 2013 (UTC)
Site is located in Brazil
Site in Brazil
{{Location map
| Brazil
| width   = 240 | float= right
| border  = | caption = Site in Brazil
| alt     = | relief  = 
| AlternativeMap = 
| label=Site | label_size  = 
| position=  | background=
| mark=      | marksize    = 
| link1= <!--03°40′26″S 40°14′20″ -->
| lat_deg=03 | lat_min=40 | lat_sec=26 | lat_dir=S
| lon_deg=40 | lon_min=14 | lon_sec=20 | lon_dir=W

The above map, displayed by Template:Location_map for correct placement on most browsers, shows the site at 03°40′26″S 40°14′20 (far north of the original coordinates: 42|17|47|S|56|51|36|W), and so I tend to use coordinates of a nearby town, as perhaps a more accurate map location. -Wikid77 (talk) 09:23, 21 December 2013 (UTC)

@Wikid77 I am not sure I understand your reply: were you able to find the correct GPS coordinates based on the Google Map link I provided? The article still has incorrect GPS coords. --Piotr Konieczny aka Prokonsul Piotrus| reply here 11:08, 26 December 2013 (UTC)
the google link in the the OP posted leads here: 3°44′30″S 38°32′20″W / 3.7415729075568422°S 38.539000041782856°W / -3.7415729075568422; -38.539000041782856
HTH - קיפודנחש (aka kipod) (talk) 20:29, 26 December 2013 (UTC)

I am having a problem editing

When I VIEW the Wikipedia articles, they display fine. However, when I EDIT a page, the references and associated material are not displaying, and when I make any change they are lost when I SAVE. I have run lots of antivirus checks. I was having this problem in Mozilla Firefox (which I have uninstalled and will reinstall) but it is also happening in Internet Explorer. Any help would be greatly appreciated, as I am currently out of the editing business, as I don't want to be destroying articles. Thanks. 7&6=thirteen () 15:14, 20 December 2013 (UTC)

I see the mangled revision you saved, as dif558, which dropped all reftags from the page, to compare the effects after a "Save page" has been run. Is it still happening when you try editing? -Wikid77 16:11, 20 December 2013 (UTC)
Yes. Sigh. I did not deliberately do any of that. I changed three words, and it it dropped all the references that had links in them. I had a similar problem about a month ago, and it cleared up on its own. I think that when I ran an antimalware scan it successfully removed something. Unfortunately, it isn't doing it this time. 7&6=thirteen () 16:35, 20 December 2013 (UTC)
A few days ago I filed WP:EF/R#Removal of HTML and extension tags for this problem. --Redrose64 (talk) 18:01, 20 December 2013 (UTC)
found this Process.Start() and ShellExecute() fails with URLs on Windows 8. Does that relate? 7&6=thirteen () 18:40, 20 December 2013 (UTC)
It's not just reftags with URLs, but also "&nbsp" and perhaps the format tags "<small>" or "<br>". -Wikid77 04:21, 21 December 2013 (UTC)
No, that is not related. Browsers do not use those API calls when you submit an edit. The page you link to relates to clicking a link in some other application (not a web browser) and it failing to open a web browser. – PartTimeGnome (talk | contribs) 17:44, 21 December 2013 (UTC)
  • Other users drop all HTML tags or &nbsp entities: When it seems the reftags "<ref>...</ref>" have been omitted, then try to insert other tags, such as "<small>testing</small>" and see if the "Show preview" button will display the small-font "testing". Other people have also had trouble with "&nbsp;" being replaced with space, rather than &#160; as typical for most users, and that includes &ndash or other HTML entities. It seems the edit-buffer has omitted all HTML-type markup tags or &entities, as some option of text-only feature, but a computer virus might be responsible, although unlikely to occur for multiple users together. More later. -Wikid77 04:21, 21 December 2013 (UTC)
While we're at it, there's also this sort of edit, which removes all occurrences of "" from values of |url= in citation templates. I've seen a few of these. They create lots of URL errors, since they leave behind the rest of the URL, after the ".com" or whatever is at the end of the domain name. When I corresponded with the editor who created the first one I saw, he had no idea what happened and was quite disturbed by it. – Jonesey95 (talk) 07:12, 21 December 2013 (UTC)
I used many forms of antivirus and malware removal. It did not solve the problem.
I upgraded to Windows 8.1 from Windows 8 which did not solve the problem.
I then switched from Firefox and Internet Explorer to Google Chrome, and the problem is resolved. I am back. 7&6=thirteen () 15:36, 21 December 2013 (UTC)

And the problem returned with Chrome. I was able to edit one article, then bam. I am completely out of ideas. Other than to use another computer. 7&6=thirteen () 16:18, 21 December 2013 (UTC)

  • Well, since direct deduction has failed to find an answer, perhaps try a "shotgun search" and try everything, looking at Internet forums about "dropping OR omitting" tags. I found page "Tag omission" which indicates how there is an option to auto-remove some kinds of tags. I would also reset Special:Preferences to the default options, or logout to use IP-address editing, to see if the username is causing all tags and &entities to be omitted. I know it can be frustrating unable to Edit, so try to turn this experimenting with a myriad of options into a "learning experience" to review other aspects, or write new articles without tags, or start page "Crapnology". -Wikid77 (talk) 17:02, 21 December 2013 (UTC)
I returned all the wikipedia editing preferences to the default position. No change. When I click on edit (this time in Chrome) it shows the article without any of the references. When I edit, it then drops them in the saved document. I did not touch any of the references (indeed, I couldn't as it doesn't show in the text to be edited). 7&6=thirteen (talk) 18:22, 21 December 2013 (UTC)
Found this Browser Issues at MediaWiki and Wikipedia:Browser notes. 7&6=thirteen () 22:13, 21 December 2013 (UTC) 19:14, 21 December 2013 (UTC)
Curious. When you switched to another browser the problem appeared to be gone on the first edit, but then returned? That sounds like some kind of retained state, such as might involve a cookie. Try deleting your cookies. Also, try editing without logging in (i.e., as an "IP") and see if you get the same behavior. Does this behavior also affect your Talk and Sandbox pages? And: can you try editing using Visual Editor? ~ J. Johnson (JJ) (talk) 20:49, 21 December 2013 (UTC)
Cleaned out the cache ("from the beginning of time") and the cookies in Google Chrome. For now, I am seeing articles in the EDIT mode with the references and links. I hope this works. Thank you. If it doesn't, or thing change, I'll let you know. User:J. Johnson. 7&6=thirteen (talk) 21:26, 21 December 2013 (UTC)
I logged in. Tried to edit Wayfarer (dinghy) to add seaworthiness. Didn't do it because it reverted back again to the same display, no references. I have not tried editing as an IP, but will. 7&6=thirteen (talk) 21:32, 21 December 2013 (UTC)

I edited as an IP, and had the same problems. Returned to Chrome and used Ctrl-Shift-D. At the very end of the tools, there is a "revert browser to factory settings". I did that. I think it worked! Have made two edits so far without the dreaded 'burning of the books' so to speak. 7&6=thirteen () 22:13, 21 December 2013 (UTC)

It worked! As of today, no issues. 7&6=thirteen () 19:15, 24 December 2013 (UTC)

7&6=thirteen, is everything still working for you? Can you edit pages with non-breaking spaces now? Whatamidoing (WMF) (talk) 19:53, 26 December 2013 (UTC)

Yes. Everything is copacetic. 7&6=thirteen () 19:56, 26 December 2013 (UTC)
I'm glad it's working for you. If it ever develops again, or if anyone else encounters this, then I'd really like to know if VisualEditor breaks in a different-but-related way before you reset Chrome. Using Chrome and Windows 8 appears to be a common thread in a rare (once every few thousand edits) error that I'd like to get sorted out. Whatamidoing (WMF) (talk) 20:10, 26 December 2013 (UTC)
Doing a reset of the browser settings in Firefox also cured its problem. Had to be done separately for the browser (resetting Chrome did not affect it). I have not messed with Internet Explorer yet. 7&6=thirteen () 00:29, 27 December 2013 (UTC)

Watchlist option "all talkspaces only"

In watchlist and similar lists, I could use the option All talkspaces only and All subjectspaces only. -DePiep (talk) 11:52, 23 December 2013 (UTC)

Specific enhancement requests can be filed in Bugzilla. --AKlapper (WMF) (talk) 17:08, 23 December 2013 (UTC)
@DePiep: I agree, that would be a great feature. I was talking to someone a few days ago, about the different headspaces we have to be in, for editing articles vs project pages vs discussion areas.
Note the existing bugzilla:11499 (2007, "Add talk/no talk namespace filters to Special:Contributions") which might cover this, with a minor rewording of the title?
It would be great if this took into account the discussion areas that are in non-standard namespaces, eg the Village pumps and similar boards (I think they're all confined to just WP namespace?). Possibly Category:Non-talk pages that are automatically signed could help with this?
Also, this is particularly relevant for WP:Flow, as it aims to make everything better/clearer for all aspects related to discussion & collaboration. If you have great ideas about how to rethink any aspects of discussion participation, the dev team would love to hear them.
(On a purely personal level, I've been contemplating how it might be interesting to (optionally) split the watchlist window into 2 columns, with discussions on the right, and everything else on the left, so that an edit to [Talk:Foo] would be easier to see if it was near the same time as an edit to [Foo] (and other page-ordering variations). If I ever find free-time, I'll try to make a mockup of how this might look. :) Quiddity (WMF) (talk) 21:46, 23 December 2013 (UTC)
  • I agree that such an option would be helpful. Just so it does not become overly complicated, it could be made an opt-in option. TheOriginalSoni (talk) 15:44, 25 December 2013 (UTC)

I need a list of pages transcluding two templates

Can anyone make a list for me of pages transcluding {{multicol}} and {{col-end}} but not (the correct, <div>-tag-closing) {{multicol-end}}? I'm hoping that the list is not very long. Whatamidoing (WMF) (talk) 00:22, 24 December 2013 (UTC)

Here you go. Anomie 01:04, 24 December 2013 (UTC)
FYI, one easy tool to give you this information would be AutoWikiBrowser - you need approval to use it for editing, but not for generating lists. עוד מישהו Od Mishehu 18:11, 25 December 2013 (UTC)
More than 300 pages. Ugh, but thank you for the list.
(AWB unfortunately does not appear to work on Mac OS.) Whatamidoing (WMF) (talk) 01:34, 26 December 2013 (UTC)
@Whatamidoing (WMF): In the future, you might want to try CatScan. Happy holidays! GoingBatty (talk) 19:08, 26 December 2013 (UTC)
Thanks for the suggestion. By the way, if anyone ever creates a class for how to use CatScan, please sign me up. The last time I tried to use it, it took me half an hour to find the intersection of two categories. Whatamidoing (WMF) (talk) 19:44, 26 December 2013 (UTC)

Help move this key from under chart to side...

Disability-adjusted life year for epilepsy per 100,000 inhabitants in 2004.

Hello, we've got this epidemiology chart here: Epilepsy#Epidemiology. There's 3 parts, the chart, the caption underneath, and underneath that caption there's the color key. Because epidemiology probably isn't going to be expanded much further I'd like to move the color key from underneath the caption to be to the right of the graph itself. This would prevent the color key from poking down into the History section below. Yes I know about the {{-}} template but then that makes ugly whitespace. Anybody know how to make that happen?? Some sort of column or table thing? Much appreciated, thanks... Zad68 16:45, 24 December 2013 (UTC)

  • Tall boxes adjust to new text, moved higher: I have moved that image+legend higher on the page, to retain the tall, portrait-style shape which will better adjust to any added text in the future. In general, a landscape-width box tends to clutter the typesetting, with less oportunity to keep related phrases near the box, and so moving a tall box higher on the page tends to allow more flexibility when adding, or removing, the nearby text. There is no need to force every illustration to remain within the confines of section headers. -Wikid77 (talk) 21:25, 24 December 2013 (UTC)
Epilepsy world map - DALY - WHO2004.svg
Disability-adjusted life year for epilepsy per 100,000 inhabitants in 2004.
Wikid, what you've done is an WP:ACCESS violation because it hides the image from people who skip to the ==Epidemiology== section and can't visually see what's hanging down from the previous section.
Zad68, I think that putting it in a table would work, as shown here. I've given the default image size, plus 2 pixels left and right for the table's column widths. You might choose to increase the image's size and/or decrease the legend's size (e.g., by using {{caption}} on each of the two columns), but if you do, then you should probably test it with several different font sizes. WhatamIdoing (talk) 01:49, 26 December 2013 (UTC)

Thank you both Wikid77 and WhatamIdoing, very helpful! Zad68 20:17, 26 December 2013 (UTC)

Maybe this way: File:Rail gauge world.png? -DePiep (talk) 21:31, 26 December 2013 (UTC)

Turn off nutty auto-"correct"?

How do I turn off the semi-idiot auto-correct spell check here, that's driving me nuts? For instance, when I type "Cochiti" (a pueblo in northern New Mexico), and my over-active touch-pad highlights it, I get "Couchette"! Very disconcerting, and annoying. Help? TIA, Pete Tillman (talk) 23:07, 24 December 2013 (UTC)

That's your computer, not Wikipedia, so it would be a question for the reference desk rather than the village pump. That said, you will need to specify your device model and operating system version. Someguy1221 (talk) 23:48, 24 December 2013 (UTC)
Nope, only happens in the editing window here. New(ish) problem, old computer ;-[
2008-ish MacBook 2.1 running OS 10.6.8 (Snow Leopard), fwtw. --Pete Tillman (talk) 01:06, 25 December 2013 (UTC)
Actually, your cmt spurred me to turn off the spell-check in Firefox. Still get Couchette for Cochiti--- some sort of auto-replace, triggered by the [space] after the word -- could it be typeit4me? [turns off]. Aha! -- upgraded to a new version recently. Now, turn off auto-typo correction. Cochiti . Done!. Imagine green checkmark here .... Thanks, Pete Tillman (talk) 01:31, 25 December 2013 (UTC)
--S Philbrick(Talk) 22:03, 26 December 2013 (UTC)

Checking templates

This template {{NCAA Women's Division I Basketball Tournament}}

is not working properly. See collapsed version:


I think it is one of the last two edits, however, I do not know how to see how a prior version of a template rendered to see where the problem first occurs.

The seem to work in preview mode, but something is wrong.--S Philbrick(Talk) 14:10, 26 December 2013 (UTC)

The <noinclude> tag was misplaced inside the template. Edokter (talk) — 14:17, 26 December 2013 (UTC)
Thanks, much better.--S Philbrick(Talk) 16:09, 26 December 2013 (UTC)

Syntax in a template

I have this syntax in a template.

| label7 = Height
| data7 = {{{height|}}}

I want, if the parameter height is complete with a number, to show, automatically [[Metre|m]]. For example, by put 1,73 will show 1,73 m. Please don't suggest the use of any other templmte. Thx. Xaris333 (talk) 20:22, 26 December 2013 (UTC)

So we shouldn't suggest using {{convert}}, which will probably do what you want to do and also make your numbers more accessible to English-speaking readers in backwards countries like the U.S. where feet and inches are still used for height? OK, I won't suggest it. – Jonesey95 (talk) 21:50, 26 December 2013 (UTC)
I only want to put metres so I don't need {{convert}}. Thx anyway. Xaris333 (talk) 23:57, 26 December 2013 (UTC)
You would do that like this:
| data7 = {{#if: {{{height|}}} | {{{height}}} [[Metre|m]] }}
However, it is probably better to use {{convert}} in situations like this, because not everyone likes their height to be displayed in metres. — Mr. Stradivarius ♪ talk ♪ 11:22, 27 December 2013 (UTC)

Protection messages

I usually see protection messages with settings saying "[edit=autconfirmed] [move=sysop]" or something similar, but lately I've been seeing more custom-looking like messages with "[edit=Allow autoconfirmed users only] [move=Allow administrators only]". Is this part of a user preference, or some other MediaWiki change that I'm not aware of? TeleComNasSprVen (talkcontribs) 21:26, 26 December 2013 (UTC)

These are usually set in MediaWiki's LocalSettings.php a directory within a MediaWiki installation and is controlled with $wgRestrictionLevels. Mlpearc (open channel) 23:09, 26 December 2013 (UTC)
This was changed almost a year ago (Gerrit change 9359), so that non-English wikis can see the text in their own languages. This originally applied both to the edit summary on the dummy edit and to Special:Log, but the latter was rolled back (Gerrit change 42249) to avoid breaking clients that parse the protection levels out of the IRC feed. There is no user preference to control this. See Wikipedia:Village pump (technical)/Archive 107#Logs for some contemporary discussion. Anomie 01:59, 27 December 2013 (UTC)

Help for Wikipedia:AutoWikiBrowser/Custom Modules

I need help for apply Wikipedia:AutoWikiBrowser/Custom Modules#Add template if not already on talk page. Xaris333 (talk) 06:54, 27 December 2013 (UTC)

@Xaris333: It would be better to ask your question at Wikipedia talk:AutoWikiBrowser/Custom Modules, with specifics as to what help you're looking for. Thanks! GoingBatty (talk) 07:05, 27 December 2013 (UTC)

<noref><ref> ... </ref> </noref> tagging?

Once a <ref> tag is in wikitext, it's hard to handle it specifically. This is a topic where one wants the reference out of legal parameter input. AFAIK, Lua has no stable solution either. Let me suggest to create a new tag that throws any ref out. Consistently, the opposite my be available too.

<noref> ABC<ref>author: A, book: B, ...</ref></noref> → ABC
<refonly> ABC<ref>author: A, book: B, ...</ref></refonly>(remaining code shown): <ref>author: A, book: B, ...</ref>

Inspired by WP:noinclude. -DePiep (talk) 07:19, 27 December 2013 (UTC)

<ref> is easy to handle. As a PCRE regular expression, /<ref(?:\s[^>]*)?(?:\/>|(?<!\/)>.*?<\/ref>)/s. For maximum compatibility, remove/hide comments, <nowiki>, and the like first. Lua is slightly more difficult since its patterns don't have alternation or quantified groups, but it should still be simple enough to write a function to do it. {{#tag:ref}}, on the other hand, is a fair bit harder since you have to actually parse the templates out of the wikitext, but it's still very doable. But then once you get into silliness like {{sfn}}, you're about out of luck without a lot of work. Anomie 13:11, 27 December 2013 (UTC)

Bad link in Fall and Rise of Reginald Perrin


This link
Many exterior scenes were filmed at sites in Cheltenham including Eldorado Road and the Beehive public house. Reginald Perrin TV Locations at Gloucestershire On Screen
does not end up where it is supposed to. -- ALGRIF talk 10:14, 27 December 2013 (UTC)

It looks like the site moved to Google Sites. I've updated the link to the new location. — Mr. Stradivarius ♪ talk ♪ 10:43, 27 December 2013 (UTC)

AJAX suggestions now mandatory?

I cannot find a way to disable AJAX search suggestions anymore. I have had them disabled for a long time but I noticed today that the suggestions were back. The checkbox is not in Special:Preferences Search. Could you please tell me where to look or revert the preferences to how they used to be? Thanks. Connor Behan (talk) 05:42, 20 December 2013 (UTC)

It's been removed from the preferences bugzilla:52812. Basically it's one of those options that was cluttering the preferences and almost no one used it. —TheDJ (talkcontribs) 07:39, 20 December 2013 (UTC)
As someone who used it and hates the suggestions, can you advice of the monobook.js/commons.css hack to get rid of them? Among other reasons I dislike it, it screws up the accessibility of the search verses go buttons which it's hard for me to imagine functioning without (but then again, I think the foisted change to Vector as the default was the worst thing this website ever did).--Fuhghettaboutit (talk)
If another JS fan made an unsubstantiated claim that almost no one used wiki markup would the visual editor then be mandatory? This goes against Wikipedia's principles of consensus. Connor Behan (talk) 17:33, 20 December 2013 (UTC)
As stated so many times before, the MW software is not the subject of Wikipedia consensus. At most it acts as an influence on some of the decisions made. You don't get to disable your search suggestions on google either. Your options are to disable JS altogether, or to spend your own time to get rid of it. As to accessibility ... if you can't find the advanced link in your search suggestions and CAN find the advanced button, well then think of all the poor people who can't make heads or tails of the preferences view. What you actually mean I suppose, is that you want to disable the "I feel lucky" behavior of the search dialog. I suggest to file a bug report for that specific action and I give you high chances that you might actually receive the option. —TheDJ (talkcontribs) 17:55, 20 December 2013 (UTC)
I also hate the suggestions. This isn't google. I can make heads or tails of the preferences view, and I don't like others deciding these things for me. The suggestions box hides "go" and "search" when it drops down.--Smarkflea (talk) 18:02, 20 December 2013 (UTC)
it makes tons of sense to remove options that only tiny number of users use (similar thing happened a while ago, when several skins that were only used by tiny number of users were removed).
it is very easy for a registered user to hide the suggestions: add to Special:MyPage/common.css the following line:
.suggestions-results { display: none; }
if my conclusion that you actually use monobook skin is correct, it may be better to add the line to Special:MyPage/monobook.css instead. this way, if/when you'll switch to vector, the suggestions will be visible (your rationale to remove them is that they hide the "go" and "search" buttons - this rationale does not exist in vector). peace - קיפודנחש (aka kipod) (talk) 19:32, 20 December 2013 (UTC)
That code doesn't work, though I notice it's formatted differently from the other code I have. ♫ Melodia Chaconne ♫ (talk) 21:01, 20 December 2013 (UTC)
@Melodia: I take it you refer to these edits: the formatting is fine - in CSS, newlines and spaces are pretty much interchangeable (there are a few exceptions, none of which apply here); and CSS is also very tolerant of variations in spacing. In fact, in that particular CSS rule, all of the spaces are optional (as is the semicolon). --Redrose64 (talk) 00:27, 21 December 2013 (UTC)
Ah, you are a monobook user, that explains. Perhaps we should disable search suggestions on monobook altogether. They don't make much sense in that layout indeed. actually, if I check, scratch that. You have both go (enter) and search (containing in the dropdown) fully available, though I agree that UI wise it could use a bit of work.—TheDJ (talkcontribs) 19:45, 20 December 2013 (UTC)
I like AJAX (using Monobook). If I need to use the "Search" button I hit the "escape" key or click outside of the box, making the suggestions disappear. It's only a minor inconvenience compared to how useful the search suggestions are. Killiondude (talk) 21:23, 20 December 2013 (UTC)
I'm with Killiondude on this one, the search within Monobook is useful. NtheP (talk) 21:50, 20 December 2013 (UTC)
I find the "escape" key or click outside of the box to be a major inconvenience. Is there a code for monobook? And please be deatailed, I don't know what I'm doing... Smarkflea (talk) 22:16, 20 December 2013 (UTC)
As one of the few people still using monobook, I agree that search suggestions are useful. πr2 (tc) 00:13, 29 December 2013 (UTC)
It does not make sense to remove an option if it is the only way to obey web standards (HTML5 autocomplete attribute). Pressing escape will make the suggestions go away but it will not let the browser return to its defined behaviour of displaying a list of things you recently searched for. See Wolfram Alpha for a site that implements this horribly. Connor Behan (talk) 23:07, 20 December 2013 (UTC)
Search option obscuring search and go buttons.png
@קיפודנחש: Thanks for the hack קיפודנחש. Unfortunately, while in a strict sense it works, it doesn't solve the problem. What happens now is that I get no suggestions, but the dropdown box still appears and covers the search and go buttons. It's easier to show what I mean than explain further so I've posted a screenshot above. Any ideas?

On a more general note, I do not find this a trivial inconvenience. It means every single time I want to use the search field I need to type in, click off somewhere else on the screen then go back, or a few hundred extra clicks every week. I'd be curious how many people actually had this turned off, and keep in mind that because of the way things work, it's highly likely that of that number of persons who had it off, a large percentage are core users.--Fuhghettaboutit (talk) 23:19, 20 December 2013 (UTC)

@Fuhghettaboutit:: you are correct - my testing was sloppy. however, i think that the following should work (again, either Special:MyPage/common.css or Special:MyPage/monobook.css)
.suggestions { max-height: 0; overflow: hidden; }
peace - קיפודנחש (aka kipod) (talk) 23:48, 20 December 2013 (UTC)
This worked for me in monobook, not in the other one. Thank you for being more responsive than the site workers. Smarkflea (talk) 00:40, 21 December 2013 (UTC)
@Fuhghettaboutit: A fulltext search "suggestion" is provided on the default Vector skin. Matma Rex talk 23:26, 20 December 2013 (UTC) Um actually, it's also provided in the screenshot above. So I have no idea waht you're getting at now :( – click "containing" for fulltext search, press "Enter" for normal search. Matma Rex talk 23:29, 20 December 2013 (UTC)
Perfect, it's working. Thanks much קיפודנחש. @Matma Rex: Matma, I'm not really sure what you are not understanding. The issue was that the disabling of the Ajax suggestions, formerly done through a command in one's preferences was no longer available and so users like me who do not want any such suggestions, had it suddenly enabled. The screenshot above shows that the first hack to re-institute turning it off worked (as you can see there is no list of suggestions), but that it was still showing below the search field the "containing <search term>", so it was still obscuring the Search and Go buttons that are behind that in Monobook. The Vector skin, which I despise, does not have the incredibly useful Go button.--Fuhghettaboutit (talk) 00:10, 21 December 2013 (UTC)
@Fuhghettaboutit: What I am not understanding is why would you want to disable these :) And the Vector skin has only one search button which acts exactly like the "Go" button in Monobook.
(edit conflict) @Fuhghettaboutit:: it is absolutely within your god-given rights to despise the vector skin, but neither vector nor monobook actually need the "go" button. once you are done typing, just press <Enter>, which has exactly the same effect as removing your hand from the keyboard and moving it to the mouse, and clicking "Go". (unless i'm missing something?) peace - קיפודנחש (aka kipod) (talk) 00:26, 21 December 2013 (UTC)
In Monobook skin, there is a difference between clicking Go and pressing ↵ Enter if you use Firefox 25.0 or later. The difference only becomes apparent when you return to the page using either ← Backspace or Alt+Left arrow.
  • If you enter a search term and click Go, using ← Backspace or Alt+Left arrow will send you back, the "search" window is cleared and you stay where you wanted to get back to
  • If you enter a search term and press ↵ Enter, using ← Backspace or Alt+Left arrow will send you back; but as soon as you're back, Firefox sends you forward again as if you had pressed ↵ Enter again
Clicking the "back" button has the same effect for both: the the "search" window is cleared and you stay where you wanted to get back to. This means that in order to stay back, you need to either click Go at the start, or click the "back" button at the end. Since these actions can only be performed using the mouse, not with the keyboard, it's an accessibility violation. --Redrose64 (talk) 00:47, 21 December 2013 (UTC)
(Indent wherever) While the search suggestions are gone, it still doesn't show the browser-based history. I imagine that's something that can't be fixed by CSS but I still don't get why the /option/ was disabled. This seems to be a theme on the Internet lately, kill options that work in the name of assuming 'everyone' wants some newfangled annoying feature. ♫ Melodia Chaconne ♫ (talk) 01:31, 21 December 2013 (UTC)
(Replying to the posts at me a few above) There are many reason I hate the Vector skin which are beyond the scope here but why having the separated out buttons is superior is simple. First, let me tell you that having those Ajax suggestions manically populating and changing as I'm typing is simply annoying to me – visual nails on a chalkboard. The issue is not the Go button but not having both Go and Search. I reversed them in my mind because as you know, what we have is a button labeled search with the function of the go button. Anyway, the problem is that we are defaulted to Go always unless the entry is not indexed. We can only** do an actual search for an indexed entry by scrolling to the bottom of the Ajax suggestions and clicking "containing..." and note that lots of users have no idea one can do an actual search instead of go (e.g. one; two) because having that arcane way to force a search is inobvious; they don't notice it at the bottom of the list, or don't realize what it's for if they do. Meanwhile, anyone who does not like the Ajax suggestions (and I am not alone), and turns it off but continues to use Vector, must turn to the alternative methods to access search. Separating out the two functions as we had it until 2010 was cleaner, less confusing for users and allowed for more efficient navigation.--Fuhghettaboutit (talk) 04:18, 21 December 2013 (UTC)
   **Yes, I am aware one can add a tilde to force a search, or invoke with the search field empty to reach an actual search button, but that's only going to be known to a tiny subset of people.
@Fuhghettaboutit:: you are correct that in order to get the "search" button functionality, one needs to do more than just hitting ↵ Enter. however, the functionality is still available via keyboard only, without resorting to mouse: all you need to do is use the arrow keys to get tho the little "containing" box at the bottom of the suggestions, and _then_ hit ↵ Enter. this may seem annoying ("why do i need to press so many times just to get the functionality i now have with a single mouse click?"), but in reality, it's a tiny change to habits: instead of hitting repeatedly, you can hit exactly once.
to recap:
  • for the Go functionality: type the term you want to search, and hit ↵ Enter
  • for the Search functionality: type the term you want to search, and hit +↵ Enter
peace - קיפודנחש (aka kipod) (talk) 16:58, 22 December 2013 (UTC)
Fuhghettaboutit, it appears that only 112 out of approximately 126,141 had this enabled. Two-thirds of active editors never change any of their prefs. Of the one-third that did, only 112 used this. That means that fewer than one out of a thousand active users had this enabled, and fewer than one out of 300 who actually changed any pref at all used it. WhatamIdoing (talk) 00:07, 22 December 2013 (UTC)
I may be missing something, but where does it say that the 41,626 active editors in the data had all changed at least one preference? Regardless, this should never have been an issue because Javascript features that change as we type, slow down our browsers and give celebrities free advertising should be opt-in, not opt-out. The few of us who kept our accounts free of this are being punished for not complaining earlier. Connor Behan (talk) 00:49, 22 December 2013 (UTC)
If 112 people use something, I think the people in charge should allow them to continue. I don't see how an extra button clogs anything up. Smarkflea (talk) 01:00, 22 December 2013 (UTC)
Connor, this information is in this message: "Unchanged prefs aren't included". This message is the first link in the paragraph in which Erik said to read the caveats before jumping to conclusions. (Also, if you look at the complete dataset, you'll see that zero users have default prefs.) WhatamIdoing (talk) 05:02, 23 December 2013 (UTC)
Thanks for the info WhatamIdoing! You're right that is a surprisingly small number of users, though if we figure that it's mostly out of our 5,000 most active users it's 2.24%. Since I have the hack to turn it off, I'm happy from a personal standpoint, but I question the methodology of how it was turned off, which strikes me as emblematic of the way these things are done often affecting many more people – someone decides and it's presented as a fait accompli to the community. (The comparison to changes made to Google near the beginning of this thread could not be more inapt.)--Fuhghettaboutit (talk) 13:41, 22 December 2013 (UTC)
User:MZMcBride started an RFC at mw:Requests for comment/Core_user_preferences over a year ago, and there was also discussion on the mailing lists. From my quick scan of the history, the first time a staff member posted on that page was more than two weeks after the discussion began. That doesn't suggest to me that this was something driven by WMF devs or presented to anyone as a fait accompli. In the future, I hope that meta:Tech/News will make these proposals be a surprise to fewer people. WhatamIdoing (talk) 05:02, 23 December 2013 (UTC)
Ahh, not just me then who hates this auto-complete/suggest twatery then? Another case pf power-tripping people changing things because they can. Lugnuts Dick Laurent is dead 15:23, 22 December 2013 (UTC)
Lugnuts: I think your comment here is a much better example of twattery than removing a rarely used user preference. If you have an actual (read: technical) issue, please share. Otherwise, kindly shut up. :-) --MZMcBride (talk) 13:57, 24 December 2013 (UTC)
And a Merry Christmas to you. Whoever the fuck you are. Lugnuts Dick Laurent is dead 14:00, 24 December 2013 (UTC)
I suppose it's forgivable that you wouldn't know who I am given how new I am around here. A very merry Christmas to you too! --MZMcBride (talk) 14:14, 24 December 2013 (UTC)

Wikidata links

Articles with a Wikidata equivalent have a "Data item" link in the "Tools" menu. For other articles, it would be useful if a "Make data item" (or similar) link could be provided. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 22:08, 24 December 2013 (UTC)

If there is no Wikidata page, then under the "Languages" heading there should still be a "Add links" item. This does require Javascript though. --Redrose64 (talk) 13:03, 25 December 2013 (UTC)
There is; but this is for adding Interwiki links, not creating a Wikidata entry for articles which exist only in English. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 13:13, 28 December 2013 (UTC)
Hmm. In the past I have edited the article, pasted in [[d:Special:ItemByTitle/enwiki/{{FULLPAGENAME}}]], previewed without saving and followed that link. For example, d:Special:ItemByTitle/enwiki/Wikipedia:Village pump (technical)/Archive 121. That should take you to the Q page, and allow you to create it if there isn't one --Redrose64 (talk) 14:40, 28 December 2013 (UTC)
Yes it would, but my point is that a pre-formatted link under "tools" would be much more convenient, and more likely to result in busy editors (e.g. me) creating Wikidata items. [In fact, better still would be a link to a tool which would pre-populate the Wikidata item with data from an infobox, templates like {{authority control}}, and categories, but that's another matter.] Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 15:49, 28 December 2013 (UTC)

Looking for scripts/gadgets

Can someone point me in the right direction, I'm trying to find two functions I would like to install in my project, one is added to signatures and displays how long ago the signature was placed ie. (1 month, 20 days ago) (3 years, 13 days ago) the second is a lightly colored edit window when editing a protected page, I've noticed it when editing template protected templates. I either can not find the scripts or gadgets or are over looking them. Thank you. Mlpearc (open channel) 23:01, 26 December 2013 (UTC)

Are you talking about User:Gary King/comments in local time.js? As far as edit window styling goes,
.mw-textarea-protected {
  background-color: #FFDBDB;
added to your site's Common.css file (or to a more specific page -- for example, Group-templateeditor.css in the template editor scenario) will do the trick. Theopolisme (talk) 03:12, 27 December 2013 (UTC)
I thanked you for this edit, but wanted to post something tangible :P Thank you. Mlpearc (open channel) 22:11, 27 December 2013 (UTC)

JavaScript and redirects issues?

About an hour ago, many things requiring JavaScript stopped working here on Wikipedia. For example, messages counter on top of the pages no longer displays a "big tooltip" when clicked, no custom JavaScript code works, there are no "advanced diff" buttons on pages displaying differences between revisions etc. It seems not to be caused by my browser, as JavaScript works just fine on other web sites. Also, it's not up to a bad uplink on my side, as in Firebug I see no failed attempts to load files containing JavaScript code. Tested in Firefox 26.

At the same time, redirects with anchors seem to be no longer working, for example Second-stage boot loader redirect page. That means, they do perform a redirect, but the actual anchor isn't passed to the destination page. Tested in Firefox 26 and Internet Explorer 11.

Is it just me experiencing those? Anyone else having similar issues? — Dsimic (talk) 00:27, 27 December 2013 (UTC)

As another example, syntax highlighting while editing articles doesn't work, and that's a JavaScript plugin here on Wikipedia. Also tested the redirect pages in Chrome 31, and no anchors are passed to the destination page after redirection. — Dsimic (talk) 00:40, 27 December 2013 (UTC)

I tried your anchor link and it worked for me in Firefox 26, and clicking the gray square with the 0 also opens the list of previous notifications I had received. So I guess this counts as a "Cannot reproduce"... --AKlapper (WMF) (talk) 01:20, 27 December 2013 (UTC)
Thank you for checking, I'll try to debug this further and report back. — Dsimic (talk) 01:26, 27 December 2013 (UTC)
That was totally caused on my side. I'd say it was something up to the antivirus software, which doubles as a kind of HTTP content scanner. Though, those redirect pages passing no anchors were really weird. — Dsimic (talk) 02:07, 27 December 2013 (UTC)
@Dsimic: Redirects to anchors require JavaScript to function properly, too. Matma Rex talk 03:06, 27 December 2013 (UTC)
@Matma Rex: Thank you for an insight! Just out of curiosity, why is JavaScript required there? Those are native HTML anchors. — Dsimic (talk) 03:22, 27 December 2013 (UTC)
@Dsimic: All I can say is, just because somebody years ago made them work that way :) Note that, when you click on, say, housecat, the page URL is still despite the page being a redirect to cat (that is, a MediaWiki redirect is not a HTTP redirect). On one hand, this allows links like housecat#Anatomy to work like you'd expect, but it also requires redirects to anchors to be implemented as a piece of JavaScript that jumps to an appropriate section after the page has loaded (since second-stage boot loader is and not – the anchor part of the URL only changes to "#SECOND-STAGE" after page load). Matma Rex talk 09:41, 27 December 2013 (UTC)
Thank you for a detailed explanation, Matma Rex! But still, as we know JavaScript technically shouldn't be needed there... For example, second-stage boot loader link is actually unfolding to as the URL, but a plain HTTP redirect could be issued once the redirect page is loaded, and any additional anchors could be processed on the server side, possibly overriding the anchor specified in the redirect link itself. Though, that would be introducing one more HTTP request to be made in the process, and that's probably the reason why a JavaScript-based solution was choosen back at the time. An alternative approach — involving no JavaScript code and no additional HTTP requests — would be to initially unfold the links to their final destination URLs, but that probably wasn't used because it migh be confusing to people reading the articles and looking to the URLs behind links. — Dsimic (talk) 19:50, 27 December 2013 (UTC)

Mediawiki message

In a previous discussion, it was noted that this does not properly render:

{{red|{{MediaWiki:Cite error refs without references|nocat=true}}}}
Renders as MediaWiki:Cite error refs without references

That should transclude the contents of MediaWiki:Cite error refs without references in red, and should suppress the error category. I see it as:

There are <ref> tags on this page, but the references will not show without a {{reflist}} template (see the help page).

Is anyone seeing this differently? If so, what language is set in your preferences? --  Gadget850 talk 15:50, 27 December 2013 (UTC)

You have a rule in your common.css to unhide span.brokenref, which is used by {{broken ref}} (in combination with a rule in MediaWiki:Common.css) to hide reference errors in most namespaces, including this one. Anomie 17:03, 27 December 2013 (UTC)
<headthunk> Of course— and I helped implement that. Need to add a switch to {{broken ref}} so this can be used in documentation. Thanks. --  Gadget850 talk 17:27, 27 December 2013 (UTC)
OK, 'nocat' now suppresses the brokenref class so this can be properly used in documentation. --  Gadget850 talk 17:41, 27 December 2013 (UTC)

Rfc for spacing mixed numbers

Again, people have noticed unusual spaces inside mixed numbers (space before fraction part), and I have tagged the discussion as RfC:

This topic of typesetting fractions, possibly hiding a space within a span-tag, crosses into technical means to format fractions with templates ({{frac}}, {{sfrac}}, {{convert}}, etc.), yet support wp:Accessibility for screen readers, as a mix of policy/style and technical issues. As an RfC, we can gain a broader, 30-day consensus to settle the issues in all the various facets, now starting early to gather opinions at year-end in 2013. -Wikid77 (talk) 18:15, 27 December 2013 (UTC)

Wikipedia error. Paragraph stealing sentence from following paragraph

Sometimes even with a blank space between paragraphs, it moves a sentence to the wrong place. I am unable to duplicate it outside of one article though. Is there a place on Wikipedia to point out errors like this? Notice at [68] if you go to the part saying When discussing growing up in Pre-Civil-Rights-Era Louisiana, that sentence will be part of the previous paragraph. If you hit the "edit this page" tab, and look at that spot though, its got a blank line between that and the previous paragraph it gets stuck to. It should be in a separate paragraph. By moving the quote to a different line after it, I fixed the problem, after first trying something else. So, having a sentence with a quote in the same sentence, sometimes makes that sentence appear in the previous paragraph. Odd. Dream Focus 21:00, 27 December 2013 (UTC)

I think it's to do with the {{quote}} template. If you insert a single newline before the phrase "He added, paraphrasing a Biblical passage", and preview, it formats correctly with "When discussing growing" at the start of a new paragraph. --Redrose64 (talk) 21:15, 27 December 2013 (UTC)
you may be right, but i think this is a bug, nonetheless. i think it's worth reporting to bugzilla (though i'm not reporting it myself ATM). peace - קיפודנחש (aka kipod) (talk) 22:07, 27 December 2013 (UTC)

This is about the html tag <blockquote>...</blockquote> and not about MediaWiki, although it's possible {{quote}} could be modified to avoid the issue. Simplified example where I see Word1 and Word2 on the same line:


Word1 Word2


The example produces this code (view source to see the code):


Word1 Word2


The same still happens if we simplify further:


Word1 Word2


If the above source lines are saved in a html file on my own computer and viewed by IE, Firefox or Chrome without any MediaWiki then the same still happens. I don't know whether it's an intentional part of the blockquote specification in html but it isn't something to report to Bugzilla. PrimeHunter (talk) 22:51, 27 December 2013 (UTC)

i think you do not fully understand the way wiki parser works. what you describe (with saving html on home computer) has nothing to do with "blockquote" specifically, it has to do with the way browsers treat white spaces. saving this stuff to a file on your local computer and viewing it with a browser is not a valuable test. browsers, in general ignore multiple whitespaces (spaces, tabs, and newlines) and collapse any sequence of whitespaces to a single space. html specification wants you to use either a "br" or a "p" tag to create a new line, and just ignores all those linefeed characters.
the wiki parser, on the other hand, is supposed to convert any 2 or more consecutive newlines to "p" elements. it clearly did not do what it's supposed to do in this case, which is the bug. peace - קיפודנחש (aka kipod) (talk) 23:17, 27 December 2013 (UTC)
I wasn't thinking. Well, at least the simple example may help narrow down the issue. PrimeHunter (talk) 23:51, 27 December 2013 (UTC)
(edit conflict)I'll have to pick through the spec, but I am quite sure that <blockquote> must start and end with a newline. And there is the possibility that HTMLTidy is at work here as well. I cleaned up the article a bit, but the particular quote had already been removed. --  Gadget850 talk 23:20, 27 December 2013 (UTC)
I've not checked, but it would really surprise me if an HTML spec said that. HTML is generally not sensitive to new lines and treats them the same as whitespace. Many HTML documents can be minified to just one line. (There is the odd exception, such as inside <pre> tags and inline JavaScript.) Wikitext, on the other hand, is sensitive to new lines; see Bgwhite's comment below. – PartTimeGnome (talk | contribs) 00:29, 29 December 2013 (UTC)
This is a bug that was "introduced" in a fix for bug 6200 Start with looking at comment #79 and onwards. The bug is in all quote templates and blockquote. The reason the bug happens is, "Yeah, it's a side effect of the old parser handling tags on a line-by-line basis. But I think it's too established a behavior to change now, and Parsoid is backwards-compatible with this behavior." The work-around is to start and end with a newline.
bug 5718 is related. There is a similar problem in ref tags. Adding blank lines in ref results in nothing happening. Bgwhite (talk) 08:21, 28 December 2013 (UTC)

Dab solver tool: link in tools menu

The Dab solver tool to which DPL bot includes a link every time it leaves one of its helpful messages on a talk page, is very useful. It would be useful also to have a link to it in the "tools" menu on the left-hand side of articles. How can we get that added? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 11:02, 28 December 2013 (UTC)

I have already added this link for myself with the following lines in my vector.js page:
// add a "Dab solver" link to the toolbar
jQuery(document).ready(function (){
   if( mw.config.get('wgIsArticle') ){    // only add toolbox link on actual pages
       var dabsolverUrl = '//' + mw.config.get('wgPageName');
       mw.util.addPortletLink('p-tb', dabsolverUrl, 'Dab solver', 't-dabsolve', "Invoke Dab Solver on this page");
Unfortunately, the usefulness of this tool is rapidly declining as the Toolserver database is not being updated and the tool author has declined to port it to Tool Labs. --R'n'B (call me Russ) 15:27, 28 December 2013 (UTC)
Thank you. I take it that the code is not available for someone else to so so? Should we perhaps be mandating that all code on Tool Labs must be publicly available and open source? Or is that already done? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 15:55, 28 December 2013 (UTC)
Well, "we" don't and can't mandate anything at all for projects that aren't the English Wikipedia. That's a decision that would have to be taken by Toolserver, not us. I believe that in the past, they have declined to do so, but I'm certain that someone else here will know far more about it than I do. WhatamIdoing (talk) 18:53, 28 December 2013 (UTC)
For Tool Labs, open-source licenses are required: see wikitech:Nova Resource:Tools/Help#Licensing. Nothing can be done about Toolserver tools besides contacting the author and asking them to release the source. Anomie 19:14, 28 December 2013 (UTC)
Thank you for clarifying. (By "we", of course, I meant the Wikimedia community.) Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 19:58, 28 December 2013 (UTC)

WikiProject template selector

The AfC helper script by Timotheus Canens and Mabdul, includes a really useful tool for selecting WikiProject templates, which it then adds to the talk pages of articles which it publishes. It featurtes a rating tool ("stub", "start", etc), and options to add "needs image" and "needs infobox" properties. It would be useful to have this tool in stand-alone mode when creating or improving new articles during normal editing; perhaps as an option in Twinkle (I'll drop a pointer to this comment on the Twinkle and AfC script talk pages). Would that be possible, please? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 11:10, 28 December 2013 (UTC)

I wrote such a thing once already, but I am somewhat embarrassed of it. I really need to rewrite it, but I lack the time to commit myself to it fully. Keφr 14:39, 28 December 2013 (UTC)

I actually wrote the WikiProject selector feature in question -- and sure, I'll be happy to convert it to a stand-alone tool or work with other devs to integrate it in other tools. Theopolisme (talk) 18:35, 28 December 2013 (UTC)

Thank you - please do. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 19:55, 28 December 2013 (UTC)
I'm working on incorporating parts of this tool into Twinkle. Obviously, appropriate attribution will be provided where needed. I'll discuss this further on other pages. — This, that and the other (talk) 04:19, 29 December 2013 (UTC)
Thanks - please let me know if you need testers. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 11:07, 29 December 2013 (UTC)

New column graphs on WP:Wikipedia with dislocated axis labels

Could someone with knowledge of making Wikipedia graphs look at the "Internal Quality Control" subsection on WP:Wikipedia? The labels on the axis appears to be dislocated when running under Windows 8. BillMoyers (talk) 15:09, 28 December 2013 (UTC)

WP:"Wikipedia" is a redlink; do you mean WP:Wikipedia? That doesn't have an "Internal Quality Control" subsection. --Redrose64 (talk) 15:41, 28 December 2013 (UTC)
It must be about Wikipedia#Internal quality control and assessment of importance. PrimeHunter (talk) 15:51, 28 December 2013 (UTC)
Yes, PrimeHunter, that's it. Graph labels are really dislocated. BillMoyers (talk) 16:02, 28 December 2013 (UTC)
I guess that's why I was being quizzed at User talk:Redrose64#Possible experience on a specific app. --Redrose64 (talk) 18:52, 28 December 2013 (UTC)
Hello Redrose64 and PrimeHunter, Redrose probably remembers the "skyscraper graphs" conversation from a week ago. This shift to 4-color graphs works ok, yet the tick marks and axis-labels still do not line up under my Windows8 set-up. Do they all line up on your screen? Zero on the bottom, up to 2 million on the top-most tick mark? Can anyone see if the graph works on their screen, or is it only a Windows8 issue? BillMoyers (talk) 22:02, 28 December 2013 (UTC)
────────────────────────────────────────────────────────────────────────────────────────────────────(edit conflict) I found having two pie charts side-by-side to the right of the bar graph meant there wasn't enough room to correctly show the bar graph at 1024×768. Bizarrely, this caused the graph to be split, with the tick marks and labels being shown first, and the rest underneath. I've placed the pie charts on top of each other instead. The bar graph now shows correctly when full screen at 1024×768, but still goes wrong at lower screen resolutions and window sizes.
A proper fix for this would probably be best done in Module:Chart. As well as fixing the bizarre splitting of the graph, it might be an idea to wrap the graph with style="overflow: auto", so scrollbars appear if the graph doesn't fit. קיפודנחש (kipod) is probably the best one to ask about this, as they created this module and have done the most work on it. – PartTimeGnome (talk | contribs) 01:08, 29 December 2013 (UTC)
  • I see the labels centered where the tick marks should be on Win7. Technical 13 (talk) 00:49, 29 December 2013 (UTC)
not sure what the correct fix should be. unfortunately, i don't know how much time i'll have for finding and implementing a good fix for it, in Module:Chart. i haven't touched it in six months or so. for now, i added a "clear:both" element (aka {{-}}) before the bar chart, which seems to solve the local issue in the article Wikipedia. if anyone is willing to fix Module:Chart, they are very welcome. peace - קיפודנחש (aka kipod) (talk) 07:47, 29 December 2013 (UTC)

Prevent link from becoming bold


Hi, is there any way to prevent a link from being displayed in bold when it appears on the same page as the link's target? (The context is {{Nuclear-powered icebreakers of Russia}}, which uses bold to highlight ships still in service. When this appears on the article of a retired or scrapped ship, it seems to incorrectly show that ship as in service. Obviously if there is no way to prevent this then another type of formatting can be chosen to highlight active ships.) (talk) 18:35, 28 December 2013 (UTC)

Selflinks are supposed to be bold in navboxes so the reader knows they are on that article. At least that's what many readers have become used to and changing it would be confusing, although it could be done with {{#ifeq:{{PAGENAME}}|Arktika (icebreaker)|Arktika|[[Arktika (icebreaker)|Arktika]]}}, and so on for other links. Bold "selflinks" aren't actually links so they are black and not blue. I think this gives sufficient hint that it's different from the other bolds and doesn't necessarily mean in service, but the explanation could be changed to say ships in service have bold links (a ship in service would be bold without a link on its own article). Another indication like in {{Toronto Huskies coach navbox}} could also be used. PrimeHunter (talk) 19:37, 28 December 2013 (UTC)
(edit conflict) (i) Link to a redirect: [[WP:VPT]]WP:VPT; (ii) make sure the link has a fragment identifier - if you want to link to the page and not a section, use #top as in [[Wikipedia:Village pump (technical)#top]]Wikipedia:Village pump (technical)#top.
In both cases you can pipe the link if you don't want the workaround to show: [[WP:VPT|Wikipedia:Village pump (technical)]]Wikipedia:Village pump (technical) and [[Wikipedia:Village pump (technical)#top|Wikipedia:Village pump (technical)]]Wikipedia:Village pump (technical). --Redrose64 (talk) 19:43, 28 December 2013 (UTC)
The answer is not to "use bold to highlight ships still in service". the use of bold (or italics, smallcaps, or colour, alone to convey information is an accessibility barrier, in breach of our own guidelines and international standards. Instead, use a symbol, like a dagger or hash. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 19:53, 28 December 2013 (UTC)
  • Thanks for the suggestions. I have changed the template so that it no longer uses bold to highlight active ships. (talk) —Preceding undated comment added 21:02, 28 December 2013 (UTC)

Arbcom ballot paper

Can you get me a arbcom ballot paper! Thanks, (talk) 19:25, 28 December 2013 (UTC)

Voting in 2013 was 26 November to 9 December so you are either too late or far too early. See Wikipedia:Arbitration Committee Elections December 2013#Timeline. PrimeHunter (talk) 19:42, 28 December 2013 (UTC)
Also, since you don't appear to have an account, you weren't eligible to vote in the 2013 election. See point 2 ("Voting period") in the link PrimeHunter supplied for more information on the suffrage requirements.
FWIW, the ballot paper was at Special:SecurePoll/vote/360. That link no longer works since the election is over. – PartTimeGnome (talk | contribs) 00:15, 29 December 2013 (UTC)

Embedding templates in Infoboxes

Having been asked a couple of times lately, how to embed templates within infoboxes, I've started:

Please help to develop/ deploy them. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 22:33, 28 December 2013 (UTC)

Strange behaviour of top tabs

Can someone have a look at this query raised at the Teahouse and see if they have any ideas on what the problem might be, Wikipedia:Teahouse/Questions#why are they waggling at me?. Thanks. NtheP (talk) 00:09, 29 December 2013 (UTC)

  • I'm pretty sure this is resolved. I'm guessing that Edokter's gadget was being used unknowingly and consolidating the tabs into a drop-down. Technical 13 (talk) 00:22, 29 December 2013 (UTC)

MediaWiki, edits

It's possible from the same account to make at the same time multiple edits on wikipedia? I.E. at 21:10:15, in the same 15th second to make multiple edits? MediaWiki software does not give an session conflict error? XXN (talk) 01:42, 29 December 2013 (UTC)

  • XXN Sure, it's possible. Could you be more specific as to the user account or pages? Technical 13 (talk) 01:51, 29 December 2013 (UTC)
You might get an edit conflict if you try to make two concurrent changes to the same page, but there's no problem editing different pages at the same time.
However, unregistered users and recently created accounts have a limit on how many times they can click "Save page" in a 60-second period. They can have as many articles open for editing as they like, but can't save more than this number in the same minute. – PartTimeGnome (talk | contribs) 02:19, 29 December 2013 (UTC)
Retrieved from "