Wikipedia:Village pump (technical)/Archive 143

Problem with "Official Site" if pulled from Wikidata with multiple "Official Sites"

I'm not sure if this is a problem just with Template:Official site or with the relationship with wikidata, but it appears that Official Site is having problems if there is more than one entry in Official Site on Wikidata and none in the template. See University of North Carolina at Chapel Hill#External links. It contains {{Official website}} but if that is clicked, it tries to go to the website... "http://www.unc.edu/,%20http://www.unc.edu". Any ideas?Naraht (talk) 13:39, 5 January 2016 (UTC)

The core problem: Template:Official website is coded to use the property directly rather than a Lua call. Why is Lua necessary?
  1. There may be multiple official sites in different languages. The only way to access qualifiers (the way which Wikidata would indicate there is a language of interest) is via Lua.
  2. There may be a preferred value by ranking.
A module on Wikipedia needs to select from among the possibly multiple values, or at least display all of them sanely.
I fixed the "intermediate problem" of this case by going to Wikidata and removing the "extra" provided website, since they were the same website. --Izno (talk) 14:04, 5 January 2016 (UTC)
I've made a start at writing Module:Official website. The Wikidata code doesn't know about qualifiers or ranking yet, though. — Mr. Stradivarius ♪ talk ♪ 15:17, 5 January 2016 (UTC)
The module is now live, and understands Wikidata ranking. I didn't think that it was worth putting in language detection in the end, although if anyone knows of an actual example on Wikidata where this distinction would be useful, I would have no problem adding it. I've made one small adjustment: previously the template used class="official website", but this is actually two classes, "official", and "website", so I changed it to class="official-website". — Mr. Stradivarius ♪ talk ♪ 23:53, 6 January 2016 (UTC)
Thanks for your work on this. I assume the module behaves the same way as the template in treating locally defined values if present? Personally I think we should be more proactive in transferring data to Wikidata and removing the local values. — Martin (MSGJ · talk) 10:54, 7 January 2016 (UTC)
Yes, that's right, and I agree about the need to be proactive about moving things to Wikidata. It wouldn't be hard to switch the module to use local values only if a Wikidata value doesn't exist, or even to make it insist on using Wikidata values only. If we can come to a consensus on what to do there, then I'll be happy to make the change. — Mr. Stradivarius ♪ talk ♪ 16:22, 7 January 2016 (UTC)

WP Mobile App article description glitch

In looking at the WP App article on Newsmax Media, I see the description as "Ultraconservative American media organization". However the desktop version of the article does not use the term/descriptive "ultraconservative". The only term used is "conservative". How is it that "Ultraconservative" shows up on the mobile app? And how do we fix it? – S. Rich (talk) 21:19, 6 January 2016 (UTC)

I checked on my mobile, and I'm not seeing "ultra". I also checked the article history to see if it was just something that changed recently and it isn't. ~ ONUnicorn(Talk|Contribs)problem solving 21:25, 6 January 2016 (UTC)
Mobile uses Wikidata descriptions. "Wikidata item" under "Tools" in the desktop version of the article goes to wikidata:Q11991273. PrimeHunter (talk) 21:40, 6 January 2016 (UTC)

Looks like IP 64.132.87.246 has fixed it at Wikidata. This old dog says 'thanks all'. – S. Rich (talk) 00:34, 7 January 2016 (UTC)

Book creator toolbar

I've accidentally hidden the book creator toolbar by hitting the hide button. The book creator still works with full functionality but just without the toolbar. How do I get the toolbar back? Tom29739 (talk) 22:50, 6 January 2016 (UTC)

I managed to get it back by clearing my cookies. Tom29739 (talk) 22:50, 6 January 2016 (UTC)

Bot-fix consensus

I'm here this time as part of a bot-approval process which requires consensus that we need to fix ten thousand instances of {{Intitle}} and {{lookfrom}} on the wiki.

They are are title-oriented search links, popular in mainspace, for for use on dab pages as a bullet item in a "See also" section. They've found there way into lots of other uses, including inline usage, on similar pages where some name or topic in the title is of interest.

Currently they have the problem of messing up a printed version of their page. If you give the template a "noprint" class it leaves a stray bullet in print. The inline versions would disappear in print, and may read poorly, but there really aren't that many of them, and those 70 will be manual fixes.

Over the years they've evolved namespace parameters, labels, and alternate queries. The first part of my plan to save them is to give them a noprint bullet (in a div block), and a noprint span (for inline usage), and a print option. Since intitle is "open", I've given its sandbox version these fixes. That fixes that.

Now to remove the stray bullet from print, and to save the two templates for future use, all with no disruptions, using AWB, here is the the second part of the plan, the bot-fix. (For debating their existence in general, see the talk page at WP:NOT.) — CpiralCpiral 08:42, 6 January 2016 (UTC)

Support: Bot operator seems to have a decent knowledge of regex, and the need is present for this AWB task -- samtar whisper 08:46, 6 January 2016 (UTC)
Putting a bullet inside a div in the See also section sounds like it would have accessibility problems - see WP:LISTGAP. I think it would be a better idea to hide the extra bullet point some other way, either in MediaWiki:Print.css (if that's possible), or preferably in the MediaWiki software itself. These kinds of empty bullets were not rendered as part of the HTML output until last year, when they switched to being rendered but hidden from view. It might be that this hiding didn't work for the print view for some reason, but I would have to investigate some more to find out exactly what happened there. — Mr. Stradivarius ♪ talk ♪ 16:19, 6 January 2016 (UTC)
@Mr. Stradivarius: you're right. This should wait for the deeper solution because it certainly complicates the HTML to have div blocks in the lists. So I phabricated T123093, and the bot-fix consensus is really up to them now. My guesses are that it's a new, overlooked problem, and that CSS is impossible, and that we'll have to wait for a parser fix.
What I say there, at T123093, is that MediaWiki software correctly ignores empty bullets on screen and in print. (The second section below shows this.) It was apparently overlooked, then, that it could ever happen to become such a common idiom, that ten thousand normal bullets would ever be empty by virtue of a noprint item.
What I'll say at the bot-approval

request is that it is a fine idea to run a bot to fix the print problem. But it complicates certain patches of HTML lists for an issue that is almost certainly going to be normal priority bug-fix. So any bot fix now is temporary, and should end up, after some time, needing yet another bot-fix to cleanup the complicated HTML patches. Anytime a second, cleanup run should be done after an awaited, deeper, fix of a minor problem, this should be a classical reason for denying a bot approval request. This is because the first bot-fix makes any cleanup a low priority, plus the time interval between the two related bot-runs, say, a year, can make the need for the second run much less noticeable. — CpiralCpiral 00:37, 8 January 2016 (UTC)

Section

Perhaps divs are different than blank lines? It needs a test. Because none of the many screen readers I tried with Chrome and Firefox happened to work, if someone has access to one, please test these. The last three bullet items are each wrapped in there own "div class noprint bullets":

  • How
  • Do
  • the
  • last
  • three
  • sound?

CpiralCpiral 06:49, 7 January 2016 (UTC)

OK I got NVDA on FF to work. It treats them all the same saying "bullet" item, for these blank lines between bullets

  • how
  • do
  • the
  • last
  • three
  • sound

They all sound the same. — CpiralCpiral 08:38, 7 January 2016 (UTC)

@Cpiral: They sound very annoying to me using my primary screen reader, JAWS, like"list of 1 items, last, list end; list of 1 items, three, list end, list of 1 items, sound, list end"), and they read out similarly (but they don't sound quite as annoying) on NVDA here. Graham87 15:38, 7 January 2016 (UTC)

Section

Empty bullets seem to act correctly. It's the bullets with noprint items that are the problem. For example the third bullet with no item acts correctly, in print or on screen:


* this bullet item is normal
* <span class=noprint>this bullet item is "span class noprint", the next one is empty:</span>
*

gives


  • this bullet item is normal
  • this bullet item is "span class noprint", the next one is empty:

Please see the above in a print preview (even in a print preview of an edit preview.) — CpiralCpiral 06:49, 7 January 2016 (UTC)

This proves that the bot-fix has a deeper solution, and that it will at some point be brought into alignment. — CpiralCpiral 00:37, 8 January 2016 (UTC)

Maps

How can I create vector maps for use in Wikipedia articles? Makeandtoss (talk) 20:55, 7 January 2016 (UTC)

Do you mean maps like File:Metrolink route map 31 March 2014.svg? --Redrose64 (talk) 23:21, 7 January 2016 (UTC)
@Redrose64: Like File:Jordan location map.svg. I need to create something like this File:Jerusalem_area.svg Makeandtoss (talk) 01:44, 8 January 2016 (UTC)
The first of those was created by NordNordWest (talk · contribs) (commons:NordNordWest (talk · contribs)), who is currently active (last edited on this wiki just under a week ago, last on commons today), the second by Costello (talk · contribs) (commons:Costello (talk · contribs)) who hasn't edited in almost six years. You could ask NordNordWest directly, or if you want to throw the idea on the table for anybody to pick up, try WP:GL/MAP. --Redrose64 (talk) 10:03, 8 January 2016 (UTC)

List of admins by registration date

Out of curiosity, is there any tool or list of admin accounts by date of user registration (not date of RFA)? I'm curious how many admins still active today were editors during Wikipedia's very early days (say pre-2005). Dragons flight (talk) 19:33, 8 January 2016 (UTC)

Here is a quick list of admins by registration date. --Edgars2007 (talk/contribs) 19:44, 8 January 2016 (UTC)
Is that based on a current / recent version of the user table? Dragons flight (talk) 19:51, 8 January 2016 (UTC)
On current. But the list is static. I included time (UTC) when the query was excecuted last. --Edgars2007 (talk/contribs) 19:57, 8 January 2016 (UTC)
So of the 1327 admins, nearly 500 registered their account before 2005. A lot higher than I would have guessed. Also saw a bunch of old names that I am surprised are still active. Of course, older accounts tend to fall back into gnoming, so maybe not that surprising. Dragons flight (talk) 20:23, 8 January 2016 (UTC)
@Dragons flight: You can get a dynamically updated version of this list through the regular interface. Also keep in mind the inaccurate date bug relating to moves over redirects (fifth paragraph down in the linked section). Graham87 15:26, 9 January 2016 (UTC)

Stop using the edit filter for 30/500 protection

The following discussion is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.


MediaWiki's group and protection systems can easily be extended to handle ArbCom's 30 day/500 edit restriction, without needing to resort to the edit filter. I'd like to submit a config change for this, but the new group would need a name, and I'm not sure what it should be called. Are there any reasons that we shouldn't do this? If not, what should the group be called? Jackmcbarn (talk) 20:46, 8 January 2016 (UTC)

Well, how does your config change apply the 30/500 protection? Also, noting phab:120734 here, the general "boost up our blocking tools" task.Jo-Jo Eumerus (talk, contributions) 20:54, 8 January 2016 (UTC)
I'd add a new group, set it to autopromote after 30/500, in the same way that autoconfirmed does after 4/10, then add it as a choice to the protection dialog. It's unrelated to phab:T120734. Jackmcbarn (talk) 20:55, 8 January 2016 (UTC)
Oh, so it would be another user group? So autoconfirmed would set at 4/10, then [xxx]confirmed at 30/500? Ivanvector 🍁 (talk) 20:57, 8 January 2016 (UTC)
(edit conflict) I have suggested many, many times that we should do it if we can, and I can't think of any reason why not to do it (noting that I have no qualification to comment on technical things). Making 30/500 a system-coded protection level was specifically what I meant when I proposed it to ArbCom, not an edit filter or 3RR exclusion. As for what it should be called, what's a word for something that's in between half and full? Threequartersprotection doesn't exactly roll off the fingers. Ivanvector 🍁 (talk) 20:56, 8 January 2016 (UTC)
Well, I think that to create a new autoconfirm status in order to implement an English Wikipedia ArbCom sanction on a specific set of articles is a weird hack. I'd thought that one would need a much broader use-case for such a thing. The edit filter solution is very artificial but it does meet the AF use-case, sort of.Jo-Jo Eumerus (talk, contributions) 21:03, 8 January 2016 (UTC)
Agreed. If one is going to create a new restriction level in the interface (which all admins can access) then it should probably be on the assumption that it can and would be used more generally. That might well be a good idea though. Dragons flight (talk) 21:11, 8 January 2016 (UTC)
If at all such functionality should be implemented as part of phab:T16636, not as an ad hoc. There are certainly enough use cases for that and the blocked tasks.Jo-Jo Eumerus (talk, contributions) 21:53, 8 January 2016 (UTC)
I oppose having new user classifications created ad-hoc by arbcom; so we solve this - and tomorrow they create 100ed/60day requirements? — xaosflux Talk 21:30, 8 January 2016 (UTC)
That being said, this is the technical pump - I think the "right" way would be to indeed create another usergroup, perhaps editor, with autopromote; and use something like editor-protected. — xaosflux Talk 21:30, 8 January 2016 (UTC)
Well, in addition to making the protection level technically possible, the community sets its own restrictions on use. It would only currently be sanctioned for use on ARBPIA articles (per ArbCom). If an admin enabled it on some other page, they would be doing so against the protection policy. Like how PC/2 is technically available - I don't know if admins can actually select it now, but they could at one time. Ivanvector 🍁 (talk) 22:12, 8 January 2016 (UTC)
PC2 can still be technically applied, see list of affected pages. For the name, 'editor' seems too general and potentially confusing, maybe 'established'. Cenarium (talk) 22:23, 8 January 2016 (UTC)
PC2 is useful, and rarely is used in a WP:IAR way that usually makes sense (mostly the IAR factor is why full protection isnt quite needed). — xaosflux Talk 00:14, 9 January 2016 (UTC)
I don't like ArbCom making policy either, but a usergroup might actually give the community more say on the matter. Changes to the autopromotion criteria would from that point on require community consensus, as with all config changes. Cenarium (talk) 22:39, 8 January 2016 (UTC)
  • An advantage of using a usergroup is that we could whitelist users who don't strictly meet the criteria (by mannually adding the usergroup). I expect that ArbCom wouldn't object to such manual promotion if the requesting user has demonstrated a pattern of constructive, collaborative editing and their contributions in this area is expected to be positive. After all, this is already the case for semi-protection, with the confirmed usergroup. In addition, it might be possible to remove the usergroup if the user has proved to edit nonconstructively, even if the criteria are technically met. See also related discussion at Wikipedia:Edit_filter_noticeboard#Two_new_disallowing_filters_as_arbitration_remedy. Cenarium (talk) 22:36, 8 January 2016 (UTC)
  • I agree a new user group entirely is preferable, for reasons stated above. I thought we had already filed a phab report for this, or at least for the new form of protection? Whether WMF will do it, or the community will allow it, or whether we should be doing this 30/500 thing at all, I don't know. But I can tell you performance is not a noticeable issue with the filter implementation, and this automatic disallowing system is favourable over blindly reverting users who don't meet the 30/500 qualification -- which is what is currently being done MusikAnimal talk 23:25, 8 January 2016 (UTC)
    As far as I know there is no Phabricator task for adding another level, at this time.Jo-Jo Eumerus (talk, contributions) 23:32, 8 January 2016 (UTC)
  • The exact usergroup criteria should not be considered locked-in yet. It would be worth investigating what criteria may already be in use at other Wikis for various purposes, and seek cross-wiki consensus on some arbitrary common standard. For example a German Wikipedia process requires 2 months / 200 article edits / 50 article edits in the last year. That isn't very far off from a 30/500 threshold. Alsee (talk) 20:31, 9 January 2016 (UTC)
  • I've formally proposed a new usergroup and protection level at VPR: Wikipedia:Village_pump_(proposals)#New_usergroup_with_autopromotion_to_implement_arbitration_.2230-500.22_bans_as_a_page_protection. Cenarium (talk) 00:20, 10 January 2016 (UTC)
The discussion above is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.

Template syntax

Can somebody who understands table markup please fix the table at Gunnerus Medal? It used to be fine but when the most recent winners were added it screwed up. Thanks. BethNaught (talk) 08:12, 9 January 2016 (UTC)

  Done — JJMC89(T·C) 08:26, 9 January 2016 (UTC)
@BethNaught: The problem didn't occur when the most recent winners were added, it occurred when you removed all uses of {{STY}} without putting something else in its place - that template was used to start a new table row, so you should have left either a plain new-row marker |- instead, or even a
|- style="background: #EFEFFF;"
to preserve the previous formatting. Did you remove {{STY}} from any other pages? Those will need fixing too. --Redrose64 (talk) 11:01, 9 January 2016 (UTC)
I know the history looks different now the template is deleted, but the cell merges and alignments looked broken anyway. (Due to the TfD template, perhaps? Too late to tell, I guess.) Anyway, that was the only page the template was used on. BethNaught (talk) 11:05, 9 January 2016 (UTC)

How does Module doc work?

Hello, please can you explain in short, what makes the pages in the Module namespace display the documentation from the /doc subpage? Could the same thing be applied to the TimedText namespace? Petr Matas 00:20, 10 January 2016 (UTC)

This is done in the Scribunto extension, and the mechanism is documented here. You can edit the header for modules with a /doc subpage at MediaWiki:Scribunto-doc-page-show, and for modules without a /doc subpage at MediaWiki:Scribunto-doc-page-does-not-exist. For the same thing to be done in the TimedText namespace, code would need to be added to MediaWiki somewhere (I assume to mw:Extension:TimedMediaHandler). You would need to make a request in Phabricator to do that. Note that there have been a few problems due to the different content types on module pages and on module documentation pages, though (see phab:T61194), so people may be wary of doing the same thing as Scribunto elsewhere. — Mr. Stradivarius ♪ talk ♪ 03:17, 10 January 2016 (UTC)
Thanks! Petr Matas 03:49, 10 January 2016 (UTC)
I think the content type issue could be circumvented by placing the doc page in the talk namespace (TimedText talk:$1/doc). Petr Matas 04:25, 10 January 2016 (UTC)
That only works if you want to transclude the /doc in the TimedText talk namespace, though. It still won't display on the TimedText page (c.f TimedText talk header template proposal).Jo-Jo Eumerus (talk, contributions) 16:12, 10 January 2016 (UTC)
I have fixed the link to the proposal. Petr Matas 18:39, 10 January 2016 (UTC)

Add line

I often use an {{AddNewSection}} template for some discussions and so. The trick of this template is great. It adds a line with header and another one with inserted text. And everything could be predefined. But recently I needed to add only one line with text - numbered list item in a WikiProject page. Add section template adds one line for heading even if it is not filled in. The problem is, that if there is some extra newline between two numbered list items (add section adds one), the numbering starts from beginning. Is there any possibility (e.g. using lua module or any hack), how to achieve this?

I have this:

WikiProject contributors:
# FirstUser
# MeastheseconduseR
# Third.user.wiki

And I want to add this:

# New4thUser

But template AddNewSection or InputBox allow me to add only this:


# New4thUser

because of this layout:

<!-- heading -->
<!-- content -->

So the result looks like this and the numbering is broken:

WikiProject contributors:
# FirstUser
# MeastheseconduseR
# Third.user.wiki

# New4thUser

Does anybody know any possibility? --Dvorapa (talk) 14:21, 10 January 2016 (UTC)

Don't put blank lines in lists, not only does it restart the numbering, it's an accessibility issue. WP:LISTGAP. --Redrose64 (talk) 15:57, 10 January 2016 (UTC)
I don't want to put blank lines in list, but I don't know, how to do it. Because both AddNewSection template and InputBox add a blank line. Do you know any workaround to add just one line with list item without any blank lines before or after? --Dvorapa (talk) 16:56, 10 January 2016 (UTC)
Surely the purpose is to add a new section, not modify an existing section? New sections always get two blank lines: one above and one below the section heading. --Redrose64 (talk) 17:07, 10 January 2016 (UTC)
You understand me correctly. I don't want to add two blank lines by creating new section. I want to add one line with list item and place it at the end of a page or an existing section and I don't know, how to achieve this using any template or module. The only two existing ways to add something at the end of a page or a section I've found so far are InputBox MW extension and AddNewSection template. And they both create a new section, which clearly aren't a good choice for lists. I wanted to create an AddNewLine template if there will be anybody who will think up some workaround or even know some workaround or lua-way or so. --Dvorapa (talk) 17:38, 10 January 2016 (UTC)
All that {{AddNewSection}} does (essentially) is create a link to a URL having the query string action=edit&section=new - it is the MediaWiki software that processes that. We can't do anything about its behaviour here, you would need to file a phab: ticket, which the devs would refuse on the grounds that it's working as designed. --Redrose64 (talk) 17:58, 10 January 2016 (UTC)
That's the latest possible option I want to avoid. And you are right that it may be unsuccessful. I zhought there is any other MW extension or any other workaround I haven't found so far. --Dvorapa (talk) 18:04, 10 January 2016 (UTC)
I've found a solution using template flowlist without closing template. Could anybody write me a message into my talk page if somebody thinked up better solution? --Dvorapa (talk) 20:11, 10 January 2016 (UTC)

Why the need to double click on wikilinks?

When navigating Wikipedia on my iPad, in desktop view, I find that I have to click twice on a wikilinks to activate them. This behaviour appears to be unique to Wikipedia. It never used to be that way. The double click thing is not isolated to the same link. Clicking on link A highlights but does not activate it. With link A highlighted, clicking on link B activates link B! What causes this bizarre behaviour? It does not happen in mobile view, which I find painful on an iPad, only in desktop view. Like I said this has only been happening recently. I upgraded my iOS to the latest version but this click twice embuggerance was happening before that. Sandbh (talk) 03:24, 11 January 2016 (UTC)

Just tested in Safari on my iPad 4, seems it is happening. It doesn't happen in Puffin Browser so I'd say it might just be a Safari issue. -- numbermaniac (talk) 07:33, 11 January 2016 (UTC)

Can I use my common.css to make the top-right icons bigger?

What code could I add to my common.css file to make the top-right icons (like featured article, lock icons, etc) bigger? Especially with locks, I have trouble seeing their colour on my screen sometimes. -- numbermaniac (talk) 07:40, 11 January 2016 (UTC)

Tried this, doesn't work. --Redrose64 (talk) 10:19, 11 January 2016 (UTC)
width and height are attributes of the img element, not CSS properties, so they can't be accessed via CSS. You could do it with JavaScript, though. With JavaScript it might also be possible to substitute a higher-resolution image; just altering the width and height attributes makes the icons look a little pixelated. — Mr. Stradivarius ♪ talk ♪ 11:19, 11 January 2016 (UTC)
It is certainly possible to set image dimensions using the width: and height: CSS properties, as in this example using File:Cscr-featured.svg:
<img src="https://upload.wikimedia.org/wikipedia/en/e/e7/Cscr-featured.svg" style="width:20px;" width="40" />
<img src="https://upload.wikimedia.org/wikipedia/en/e/e7/Cscr-featured.svg" style="width:40px;" width="40" />
<img src="https://upload.wikimedia.org/wikipedia/en/e/e7/Cscr-featured.svg" style="width:80px;" width="40" />
They come out different sizes, but since the <img /> tag is not whitelisted, you need to paste that lot into a non-Wikipedia HTML doc to see what I mean. --Redrose64 (talk) 19:44, 11 January 2016 (UTC)

An HTML object off to the side with wrap-around content

I see that in many pages here, there are infoboxes which the text wraps around - that is, it avoids the text boxes, but continues directly underneath them. How does this work in the HTML? 84.228.154.145 (talk) 12:42, 11 January 2016 (UTC)

This is probably more suited to the Reference Desk than here, but the answer is the CSS declaration float: right;. Relentlessly (talk) 12:50, 11 January 2016 (UTC)

16:59, 11 January 2016 (UTC)

is there a way to auto-fill weather and climate templates?

In the course of trying to improve a city's article, I looked into adding climate information, in the standard way using one or more of the wikipedia weather templates. I've found websites describing the city's Köppen climate classification, and have found other websites with the raw numeric statistical data for temperatures etc. What has me pausing and turning to you, is that those are a lot of numbers to manually enter into a complex template, and it struck me that being that there are so many cities with wikipedia articles, maybe there was a shortcut to auto-magically fill the templates, or partially do so. —Boruch Baum (talk) 17:58, 11 January 2016 (UTC)

I'm not aware of any gadget that fills in {{weather box}} or others from a given reference but from my own experience I have to agree that it would be very helpful. On the other hand I can image that it would require some solid load of programming to create an upload mask that could proceed data from, say, CSV files or even Excel sheets. I'm working with mass data off-wiki and I can tell you that every source tends to have its own data format with different columns, headers, names and whatnot. So such a tool would most likely still need manual editing to prepare the input. De728631 (talk) 18:23, 11 January 2016 (UTC)
@De728631: Sigh, then I guess I do it manually. And I know most of our readers won't appreciate the work, because before I became an editor I didn't, and even after I became an editor, until I looked at the template, I had no idea what was involved. Oh well. Time to dive in . . . —Boruch Baum (talk) 19:10, 11 January 2016 (UTC)

Unmarking edits erroneously marked as minor

Not sure whether this issue was raised before, quick search shows it wasn't. So should we consider a way to unmark the edits erroneously marked as minor by the editor himself/herself? The proposed unmarking would remove the bolded "m" letter from page histories and user contributions and this would be restricted only to the editor who made the edit. Brandmeistertalk 22:57, 9 January 2016 (UTC)

You should make such a proposal at phabricator:. Ruslik_Zero 17:34, 10 January 2016 (UTC)
Or we can just accept that people sometimes make minor mistakes that will not impact overall quality of the project. —TheDJ (talkcontribs) 04:01, 12 January 2016 (UTC)

Microsoft no longer supporting IE 8,9, and 10

Just in case you haven't heard... Tomorrow will be the last update Microsoft will be publishing for IE 8, 9 and 10 on most systems. The only exceptions are for Vista, IE9 will still receive support for another year, and some server editions of Windows. Bgwhite (talk) 20:33, 11 January 2016 (UTC)

As far as I can see, IE8 is the only one which is becoming entirely unsupported. IE9 and 10 are more "not recommended" than completely unsupported. WMF/Wikipedia/MediaWiki have already dropped IE8 down in status, with various hacks to make stuff work in IE8 gradually being removed. Last I looked, independent stats on browsers put IE8 down at being almost completely unused and irrelevant, but 9 and 10 still had enough usage to be at least a little relevant (or irritating/annoying, depending on your perspective) to web developers. --Murph9000 (talk) 07:08, 12 January 2016 (UTC)
IE8 users who can't upgrade to IE9 (perhaps because they still use XP and can't/don't want to "upgrade" that) will probably have switched to Chrome, Firefox or Opera. --Redrose64 (talk) 08:21, 12 January 2016 (UTC)
IE9 users on Vista can't upgrade either. IE9 and 10 will NOT receive updates for Windows 7 or Windows 8. Current browser stats for Wikimedia. IE10 has less share than 8 or 9. Bgwhite (talk) 08:34, 12 January 2016 (UTC)
Yeah, now that I look at some current stats, IE9 and 10 do look like they are down in or approaching the irrelevant history category. FYI, here's the MediaWiki browser support matrix (currently listing IE support as "9+" for full support). --Murph9000 (talk) 09:17, 12 January 2016 (UTC)

False positives from recent changes at userpages

(I posted about this before but cannot find the link.)

When I search for a recreation of some username or userpage content (usually block evasion), I often get search hits for userpages that have a recent changes transclusion or such. It is a constant annoyance. I have been told to search using "intitle" or such, but this is a nuissance and extra step. Is there something that can be done like noindexing the userpages or something? Thanks. Anna Frodesiak (talk) 06:32, 7 January 2016 (UTC)

Here is an example. Search "Smartmyer". I just blocked the user and wish to check for that term elsewhere. I get lots of matches as described above. Anna Frodesiak (talk) 06:34, 7 January 2016 (UTC)

Did you try enclosing the search in quotation marks? This search returns good results for me. — Mr. Stradivarius ♪ talk ♪ 07:48, 7 January 2016 (UTC)
(non-admin comment) Under the Search field at the top of the page you can click "Everything" (which is what Mr. Stradivarius did, and by the way quotation marks are not needed for one-word terms); or you can click "Advanced" and select everything except userpages (and you can save/Remember that set-up for future searches if needed). I don't know if that helps. Softlavender (talk) 11:08, 7 January 2016 (UTC)
Thanks, Mr. Stradivarius and Softlavender. I guess I'll have to do the extra steps. It sure would be nice to just drop the term into the searchbox and that's it. Isn't there a way to make userpage recent changes transclusions non-searchable by adding something to the userpage? Anna Frodesiak (talk) 21:01, 7 January 2016 (UTC)
@Anna Frodesiak: they probably meant "insource". Insource is immune to transclusion-false-positives, so there is no need for NOINDEX. (A normal search without the insource parameter searches the content as it appears on the page, regardless of whether or not it was a mere transclusion.) Insource will also find cases where a URL or username instance in wikitext (in source) renders a label (instead of itself as the name), and would be bypassed by a normal search.
{{Search link| all: insource: smartMyer}} all: insource: smartMyer produces three results.
You can save an oft-needed search in a search link and just drop a new term into the query box on the search results page.
  1. smartMyer found zero results because it searched in article space.
  2. Search then tossed "Myer" and stemmed smart, reporting that it did so, and producing 51 thousand false positives.
  3. When everything was searched it found seven results.
The four false positives were because of transclusions.
Substrings in camelCase are indexed as "words". The beginnings of a description of this are at mw:Help:CirrusSearch. — CpiralCpiral 23:06, 9 January 2016 (UTC)
Thank you, Cpiral. I do wish there were buttons to automatically dump "insource" and "intitle" etc into the searchbox. It works well, but is a nuissance to type each time. I have updated my junkbox with your advice. Thank you kindly for the thoughtful reply. :) Anna Frodesiak (talk) 22:41, 11 January 2016 (UTC)

Follow-up

I posted here about adding some convenience buttons to the search thing-a-me-doo-dad. Feel free to yay or nay it over there. :) Anna Frodesiak (talk) 22:49, 12 January 2016 (UTC)

Watchlist is not working

My Watchlist has not green markers as it usually had. Many pages have been changed and I cannot see that. Only some bullets are enlarged (my edits). --Obsuser (talk) 02:00, 10 January 2016 (UTC)

Mine works. FF 43.0.4. Rehman 02:04, 10 January 2016 (UTC)
You might find a solution at Wikipedia:Customizing watchlists. Whatamidoing (WMF) (talk) 19:57, 12 January 2016 (UTC)
@Whatamidoing (WMF):   Fixed Yes, it’s working now beautifully. Thank you. --Obsuser (talk) 20:22, 12 January 2016 (UTC)

WTW tool

I would like a tool that can count the number of words in an article that match those listed at Wikipedia:Words to watch. Does such a tool already exist? If not, who is best to ask to create it? SpinningSpark 16:08, 9 January 2016 (UTC)

If memory serves (and it might not), I believe that User:Dank was working on something along these lines a while back. Whatamidoing (WMF) (talk) 19:56, 12 January 2016 (UTC)
Dank said on his talk page it is nothing to do with him. SpinningSpark 18:21, 13 January 2016 (UTC)

Cite option not working

Not sure if it's just me but when I click on Cite Web, News, Book etc in the toolbar nothing happens ...., I've refreshed the page and still nothing happens, Thanks, –Davey2010Talk 16:59, 11 January 2016 (UTC) *Nevermind all works now. –Davey2010Talk 03:46, 12 January 2016 (UTC)

    • @Davey2010: This very same thing has happened to me several times over the past few days – I just had this problem earlier today when editing Leslie Grossman, and had to use ProveIt instead. However, when I tried to use the toolbar again several minutes later, it worked fine!! So: 1) it's not just you that this is happening to! and 2) I have no idea what's going on here, and why the toolbar works sometimes recently but not other times. (Full disclosure: I'm using Firefox 43.0.4 on Mac 10.11.2 if that makes any difference to anybody...) --IJBall (contribstalk) 21:43, 12 January 2016 (UTC)
      • Well it's great to know it's not just me! , As I don't use it every day I unfortunately would have no idea if it's a daily thing, I'll unstrike the above, (It works for me now but if it helps I'm currently using Chrome Version 47.0.2526.106 on Windows 7). –Davey2010Talk 21:59, 12 January 2016 (UTC)

Book:Solar system not rendering

 

Hi, I'm asking on behalf of another editor, Johncogar139. When rendering Book:Solar system it fails, the full question is at WP:HD#Book= Solar System - Rendering Failed (two separate days). The error is in the image here as well. I'm asking if someone with more technical know how could answer the user on why it's failing. Thanks.  Seagull123  Φ  23:41, 12 January 2016 (UTC)

Going by Help:Books and Help:Books/Feedback, the book creator is quite broken for many cases. Broken to the extent that someone created a template, {{Bookcreatorstatus}}, to warn about it:
Status last updated 23 August 2020.
The biggest immediate issue is that it doesn't appear to give any web-visible diagnostics to help investigate what it is choking on.
--Murph9000 (talk) 01:03, 13 January 2016 (UTC)
The inability to diagnose / debug errors is covered in phab:T94308#1549095. As cscott explained in phab:T100979#1463646 he basically has no time allocated by his employer (the Wikimedia Foundation) to work on Offline Content Generator issues, which is a pity. --Malyacko (talk) 14:35, 13 January 2016 (UTC)

How filter AWB for article titles starting with "X"

Greetings, I have used AWB a few times for WP:TOTD (tip of the day) 366 articles. At WikiProject Catholicism there are thousands of articles so I would like to reduce the number for shorter running times. For example: I have setup Source" as Category; then Category of Start-Class Catholicism articles. After running the Make list, I click on Filter then choose Subject{{Dash} Category:, and Talk{{Dash} Talk: and click on Apply button.

Question: at Keep titles containing what expression should be typed to wildcard "All titles beginning with" for exmple the letter B?

I have searched here and at the AWB documentation, FAQ, etc. and do not see the answer. If this can be done, I will gladly add the answer to a new "How-to" section. Thanks. Regards,  JoeHebda (talk)  16:49, 13 January 2016 (UTC)

Use ^B and check the "Regular expression" checkbox. The caret matches the start of a text in regular expressions. SiBr4 (talk) 16:59, 13 January 2016 (UTC)
Now it filters to zero articles.  JoeHebda (talk)  17:19, 13 January 2016 (UTC)
WFM. Either your pre-filter list didn't have any "B" titles or the "Regular expressions" box wasn't checked. SiBr4 (talk) 17:48, 13 January 2016 (UTC)
Makelist showed 9753 articles which does match Wikipedia:WikiProject Catholicism/Assessment. Changed Makelist to: Source: – Category; and Category:  – Unknown-importance Catholicism articles. So now working with 2,000 + articles instead of over 9,000. At filter, the ^B still does not work even with Reg.Expr. checked ON, so am running AWB with this smaller list.
Also, wondering what "WFM." means?  JoeHebda (talk)  19:31, 13 January 2016 (UTC)
"Works for me". I forgot WikiProject categories contain talk pages rather than articles, so you need to use ^Talk:B instead (I had tested the regex using a list of "Random pages", which are articles). SiBr4 (talk) 19:43, 13 January 2016 (UTC)
That also means just filtering for Talk:B without regex works too, assuming no pages contain that text elsewhere in their title. SiBr4 (talk) 19:46, 13 January 2016 (UTC)
  Done – Now also WFM. Cheers!  JoeHebda (talk)  20:08, 13 January 2016 (UTC)

I hope that this is a good place to post this

I need help at Angel of Grief. Here is what just happened. User talk:Penarthtowncouncil posted a picture, one that should be there. However, unfortunately, the picture was posted sideways. I copied the picture, turned it the right way up and re-posted it, but now the record shows me as being the poster rather than User:Penarthtowncouncil and I am neither happy nor comfortable with this. So how do I get his/her/their name back in? thanks, Einar aka Carptrash (talk) 19:15, 14 January 2016 (UTC)

The attribution record of uploads cannot be changed, and uploads (or any other WP action) cannot be done in a way that simulate someone else acting. However, you (or someone else) can probably undo your edit, leaving the other one as the "live" one. But obviously that would leave the file itself in a state you dislike. There are probably several other solutions, but they depend on your actual concern: do you not wish to have your name associated with it, or do you simply wish that the original uploader is noted for purposes of giving credit for the work? DMacks (talk) 19:20, 14 January 2016 (UTC)
I want to make sure that 1. the image is shown in a way that makes sense, and 2. the person who took the picture, or at least posted it gets credit, (i dont care if I am mentioned or not) so I think that I shall undo my edit and then post here that someone needs to rotate the picture 90º counterclockways. Lightening it up a bit would also be a good idea. But first, undo. thanks, Carptrash (talk) 19:53, 14 January 2016 (UTC) Also, since I took three edits to get the picture and caption in, what I need, and don't know how to do is (perhaps) a rollback of my last three edits there. Anyone want to give it a try? Please? Carptrash (talk) 19:55, 14 January 2016 (UTC)
@Carptrash:, tag the new file you uploaded for deletion. Then, go to the version they uploaded (on Commons), scroll down, and there will be a link that says, "Upload a new version". Upload the picture there, and they will get proper attribution as the uploader, and you as the one who edited the image. ~ ONUnicorn(Talk|Contribs)problem solving 20:06, 14 January 2016 (UTC)
Thanks, As the old saying goes, "I've never done this sort of thing before" but I shall give it a try, thank you very much.Carptrash (talk) 20:47, 14 January 2016 (UTC) PS, I have no clue as how to tag something for deletion, or where exactly to do it. Carptrash (talk) 20:51, 14 January 2016 (UTC)
@Carptrash: You can tag it for deletion using {{db-g7}}. ~ ONUnicorn(Talk|Contribs)problem solving 21:01, 14 January 2016 (UTC)
Not any more he can't...I already deleted it:) DMacks (talk) 21:02, 14 January 2016 (UTC)
I appreciate everyone's input and suggestions here. I can now get back to trying (or something) to get blocked at Money (That's What I Want). Carptrash (talk) 21:04, 14 January 2016 (UTC)
Just for future reference, it is not necessary to re-upload an image to get it rotated on Commons. Under every image on the Commons image page there is a "request a rotation" button. This usually gets serviced promptly by a bot. SpinningSpark 23:14, 14 January 2016 (UTC)

Is there any way to make mobile link automatically jump to desktop version when I'm on desktop?

I don't want to question why MediaWiki doesn't do this already (and why use a separate .m. link to begin with), but on user's end is there any workaround? People recently link mobile sites A LOT on social media (because apparently they're on mobile and copied directly, not their fault) and it's quite a nuisance to manually switch them every time. Thanks.--fireattack (talk) 00:25, 13 January 2016 (UTC)

+1, getting the mobile version sucks. On the other hand it needs be easy for desktop editors to check how a page renders in mobile mode. The best I can think of at the moment would be to treat it like a Redirect page: You get redirected to desktop page with a Redirect=false link at the top leading back to the mobile version. That would be annoying for editors wanting to check the mobile version, but it would properly serve the clear majority case. Alsee (talk) 01:50, 13 January 2016 (UTC)
The assumption that they're all using mobile devices might not be valid. There are a lot of readers who use the mobile website on their laptops. My guess is that they find the larger font and "airy" look easier to read. Whatamidoing (WMF) (talk) 18:46, 13 January 2016 (UTC)
Well I get your point, but just like people sometimes use desktop version on mobile as well because mobile version sucks on viewing tables, templates, etc. It happens, so yeah, not ALL people who use mobile links are using mobile devices, but MOSTLY. Anyway, here my question is if there is any way to automatically jump to desktop version as an "user setting" (i.e. I'm not asking Wikipedia to change its default behavior), since unfortunately I am not the one who prefer mobile link(s) on desktop, or laptop. --fireattack (talk) 22:33, 13 January 2016 (UTC)

fireattack, Alsee: You can add

location.host = location.host.replace( 'm.', '' );

to Special:MyPage/minerva.js. Nirmos (talk) 06:41, 14 January 2016 (UTC)

Nirmos Thanx but occasionally I specifically do need to check what's going on at mobile, so that is a non-option. Also I wasn't much concerned with the few times this annoyance happened to me. I was concerned with the global case. I expect most people will WTF when the weird/crippled mobile version unexpectedly comes up. Alsee (talk) 08:24, 14 January 2016 (UTC)
I edit consistently on a mobile device, namely an Android smartphone. I always use the desktop site, which I consider far superior to the mobile site. In my opinion, the desktop site should be renamed the "useful site" and the mobile site should be renamed the "crappy site". But that's just my opinion. When the crappy mobile site crops up, I quickly scroll to the bottom and switch to the useful desktop site on my "tiny, imposssible to use" smart phone. The one that I have written and expanded many articles on, and assisted hundreds of newbies at the Teahouse. See my essay, User:Cullen328/Smartphone editing. Cullen328 Let's discuss it 08:48, 14 January 2016 (UTC)
Thanks, Nirmos, it does work. However, it will redirect even if I'm indeed on mobile... so it's not an option. Also if you don't mind, may I ask what does this "minerva.js" come from? I searched for it but still no clue. I knew you can insert your personal script at global.js, common.js, or something like vector.js for specific skin(s) but minerva.js? Is there any document about this "magic" script name? Never mind, I found it's the skin for mobile view.--fireattack (talk) 01:57, 15 January 2016 (UTC)

#property:P373

Can anybody figure out how to fix {{#property:P373}} so it works properly on Wikipedia languages with script variants. See this discussion too. --Obsuser (talk) 22:19, 14 January 2016 (UTC)

{{#property:}} is a parser function, part of the MediaWiki software, we can't change it - phab: is where to file bug reports. P373 refers to d:P373, which is a property on Wikidata. Any problems with information on that site are best discussed on that site, somewhere like d:Wikidata:Project chat. --Redrose64 (talk) 00:07, 15 January 2016 (UTC)
OK. Thanks. --Obsuser (talk) 03:51, 15 January 2016 (UTC)

Google plus links

ok here's a stupid question from stupidland. Why is visual editor reference generator throwing lots of plus.google.com links into all the articles. I don't understand, is this a thing now? Are we helping google get their social media back on its feet? --  00:05, 15 January 2016 (UTC)

Examples? --Redrose64 (talk) 00:08, 15 January 2016 (UTC)
@Keith D: - while I was looking for examples I noticed you were removing them Keith, so I thought I'd ping you here as you might know why they were appearing. Examples of where they have been removed diff (which was added by diff), diff, diff. --  00:28, 15 January 2016 (UTC)
@Nonsenseferret:, @Redrose64: - It seems to be setting this based on the HTML on the page that is referenced in an added cite from the visual editor. If <link rel="publisher" href="http://some.site" /> or some effective variant is in the page cited, the visual editor appears to use it as the "publisher" setting for the cite tag. Not sure if intentional. SQLQuery me! 00:56, 15 January 2016 (UTC)
See Wikipedia:VisualEditor/Feedback#VE appears to put external URLs in citation publisher= parameters, Wikipedia:VisualEditor/Feedback/Archive 2015 3#cs1.7C2 .7Cpublisher= parameter and google+ links, phab:T118773. PrimeHunter (talk) 01:05, 15 January 2016 (UTC)
Thanks folks, I guess I can summarize it is basically a bug sitting on a backlog. --  01:20, 15 January 2016 (UTC)
I guess that's the place to complain about the undesirable |lang=en-GB / |lang=en-US / |lang=en-AU then. --Redrose64 (talk) 10:50, 15 January 2016 (UTC)
Apparently it's phab:T115326. --Redrose64 (talk) 10:55, 15 January 2016 (UTC)
@SQL: This edit didn't notify. You need to add the links and sign in the same edit. --Redrose64 (talk) 10:50, 15 January 2016 (UTC)

User link pings haphazardly working

For the past week Juliancolton and I have been trying to get our user pings to work over the past week or so (utilizing WP:WPTC's talk page and a userpage of his), but to no avail. No matter the link format we try, we can't get it to notify each other (our settings have it specified to work with this). However, a few other users have able to ping me by doing nothing different. We're both using Firefox (43.0.4), but not sure if that's related since other users have been able to ping us. No idea what's going on here. Any thoughts? ~ Cyclonebiskit (chat) 00:33, 15 January 2016 (UTC)

See mw:Manual:Echo#Technical details and post diffs if you think pings are missing. Always post an example of a reported problem. You didn't sign [13] so it isn't supposed to make a ping. PrimeHunter (talk) 00:59, 15 January 2016 (UTC)
I have lost track of the number of times that I've explained it again, most recent examples are Wikipedia talk:WikiProject Women#Size matters - Conundrum regarding lists and, oh, Wikipedia:Village pump (technical)/Archive 143#Google plus links right above. RTFM, guys. --Redrose64 (talk) 10:59, 15 January 2016 (UTC)
In all fairness, the manual may be hard to find when you don't know where to look or that there is one. Special:Preferences#mw-prefsection-echo displays the MediaWiki messages shown in [14] There is no general message in a position well-suited for a link to Wikipedia:Notifications#Triggering events. Would a wikilink work in the heading MediaWiki:prefs-echosubscriptions? The most common problem is attempted notifications in unsigned posts. MediaWiki:Echo-pref-tooltip-mention currently displays the default "Notify me when someone links to my user page." We could say "Notify me when someone links to my user page in a signed post." I assume the tooltip cannot have a wikilink. PrimeHunter (talk) 12:04, 15 January 2016 (UTC)
Tooltips must be plain text, because they're held in the title= attribute of a HTML tag (most often the <a> tag). Might seem strange, but HTML tags cannot contain HTML. --Redrose64 (talk) 12:42, 15 January 2016 (UTC)

#tag only for paired tags bug workaround?

Does anyone have a workaround for:

In order to include variable data within a tag such as <ref>, the #tag parser function must be used. Currently #tag only supports balanced tags, not a singular tag, thus <ref name="some-name"/> cannot be used in a template unless the some-name is hard coded.

Which is one of the reasons Template:Finedetail does not work properly.

Aoziwe (talk) 11:58, 15 January 2016 (UTC)

The Cite extension treats <ref name="foo"/> and <ref name="foo"></ref> identically since 2008, specifically so {{#tag:ref||name=foo}} can work. Template:Finedetail isn't working because the extra space in {{#tag:ref| |name=foo}} is equivalent to <ref name="foo"> </ref>, which is not the special case <ref name="foo"></ref>. Anomie 14:03, 15 January 2016 (UTC)
Thanks - if it was always that simple ! Aoziwe (talk) 15:29, 15 January 2016 (UTC)

Google Chrome crashing while editing (especially pasting)

Is anyone else having trouble with Google Chrome crashing while adding material to Wikipedia? It's been happening to me these past two to four weeks. I haven't discerned any patterns, but it seems to often be when adding (perfectly normal, not blacklisted) citations (urls). It has gotten so frustrating that I've had to often switch to Firefox, even though I find Firefox less ideal than Chrome for wiki editing for several reasons. I'd like to know if anyone else is having these experiences on Chrome. One thing I think the incidents have in common is that the articles or drafts that crash are fairly long rather than blank pages or blank sandboxes. Softlavender (talk) 09:37, 6 January 2016 (UTC)
Edited to add: It also has happened repeatedly when I tried to add a warning template to an IP vandal's talk page. Softlavender (talk) 01:25, 7 January 2016 (UTC)

See Archive 142#Article crashing Google Chrome. --Pipetricker (talk) 10:10, 6 January 2016 (UTC)
No solutions there. I've tried clearing cache, closing Chrome completely, restarting computer -- nothing works, the crashes continue. Softlavender (talk) 10:30, 6 January 2016 (UTC)
@Softlavender: Which version of Chrome are you running? -- samtar whisper 10:37, 6 January 2016 (UTC)
Samtar: It says: Google Chrome 47.0.2526.106 (Official Build) m (32-bit)

Revision 19b9e1a5713f4b9ae324bd59bbe16ca6eb91d0e0-refs/branch-heads/2526@{#532} OS Windows

(I think the only, or main, personal weirdness on my Chrome is that I have both versions of Flash [Chrome's built-in Flash, and normal Flash] blocked on Chrome, because Chrome keeps hanging when I am researching and have more than 20 or so windows open.) Softlavender (talk) 01:25, 7 January 2016 (UTC)

I've had fewer issues if I go into language settings and uncheck "Use spell check with this language". ViperSnake151  Talk  04:01, 7 January 2016 (UTC)
Can you clarify, ViperSnake151? You mean you've had fewer crashes pasting material onto Wikipedia? Also, are you un-checking spell-check for English? (I'm asking because I don't want to do that; I rely on spell-check to help with my lousy typing.) Softlavender (talk) 04:39, 7 January 2016 (UTC)
Yeah. ViperSnake151  Talk  04:41, 7 January 2016 (UTC)
Just had this happen as well. Turning off spell check works, but this is obviously a crippling solution. It also seems to work if I section edit, and avoid editing the whole article. Kuru (talk) 03:49, 8 January 2016 (UTC)
And even if crippling, would have been worth it to edit Talk:Territorial evolution of the United States/rewrite, but alas I still get the error. --Golbez (talk) 06:37, 11 January 2016 (UTC)
Same problem. Version 47.0.2526.111 on OS X. It's intermittent. I found that actually, if you click back, then forwards again, you get to an edit window where your text is preserved. Blythwood (talk) 23:36, 15 January 2016 (UTC)

Please tell me how it works? What I should do? --Дагиров Умар (talk) 20:44, 14 January 2016 (UTC)

Not my language, but I'm guessing you want to upload an image or file. Start by clicking the link in the middle of the page and follow the prompts. If you are still unsure, you might want to ask at Upload help. — Maile (talk) 22:22, 14 January 2016 (UTC)

I want to do in the English Wikipedia. And here it does not work. I want these buttons were in our wiki. --Дагиров Умар (talk) 23:12, 14 January 2016 (UTC)

The comments above are not clear, but ce:Википеди:File Upload Wizard does not appear to work. Clicking to start the upload process does nothing. Alsee (talk) 23:55, 14 January 2016 (UTC)
If it's a free-use image, upload it to Commons; then any Wikimedia project can use the image without needing upload on each one. --Redrose64 (talk) 00:02, 15 January 2016 (UTC)
@Redrose64:, I'm not entirely sure, but this individual might be trying to configure an Upload Wizard in their own language. If you go to the link provided up in the section header, and into what I think is the history, the date of creation is today. — Maile (talk) 00:15, 15 January 2016 (UTC)
— Maile Yes, that's what I'm trying to do.--Дагиров Умар (talk) 00:36, 15 January 2016 (UTC)
Namely, this is--Дагиров Умар (talk) 00:39, 15 January 2016 (UTC)
@Дагиров Умар: OK. @Redrose64: has given you the best advice. All images should be uploaded on Commons. But if you are still intent on putting an Upload Wizard on your language's Wikipedia, then the place to get advice about that is Upload help. Those are the people who know the most about it. — Maile (talk) 13:25, 15 January 2016 (UTC)
Thank you wrote here. --Дагиров Умар (talk) 14:29, 15 January 2016 (UTC)
Answered there [15]. Fut.Perf. 16:06, 15 January 2016 (UTC)

Automatic archiving at "Talk:Occupation of the Malheur National Wildlife Refuge" seems to be broken

The archive bot at Talk:Occupation of the Malheur National Wildlife Refuge is not archiving the threads. Can you please investigate this? --Jax 0677 (talk) 14:07, 15 January 2016 (UTC)

The archiving instructions needed to be tweaked after the page move. -- John of Reading (talk) 15:35, 15 January 2016 (UTC)
@Jax 0677: You posted exactly the same thread at User talk:Misza13#Talk:Occupation of the Malheur National Wildlife Refuge, a page watched by myself and also by John of Reading. Please see WP:MULTI. --Redrose64 (talk) 20:49, 15 January 2016 (UTC)

Template displaying older version

I'm finding that articles with the template Template:Free and open-source typography don't show the most up-to-date version (which has 'Overpass' listed in the top section and 'Theano Didot' in the second section). Found this happens in Chrome, Safari and TorBrowser over the last few days.

It's not a cache problem, since in testing with Tor just now I got first the most recent version, then an older version. From my tests the older version comes up more often than not. Blythwood (talk) 23:41, 15 January 2016 (UTC)

Changes to Templates in MediaWiki are not instantly reflected across all articles using the template. The wiki will eventually catch up, how long depends on the size of the job queue and how busy the server is with stuff in general (i.e. how quickly it is processing the job queue). See also WP:PURGE for a way to force an immediate update if you really need to do so. --Murph9000 (talk) 23:47, 15 January 2016 (UTC)

tagging wikiproject templates in other language using eng: counterpart as ref

is there such bot that does the following:

checks all members of a specific project in eng:, if versions in another language exists(for example zh)

if zh version exists tag it with corresponding template(zh)

I am planning to tag zhwp:pharmacology using this method Panintelize (talk) 04:15, 9 January 2016 (UTC)

I can give you the list of articles, that can be tagged. The tagging process itself would be up to you (or we can come up with another idea then). Moving further disscusion to your talk page. --Edgars2007 (talk/contribs) 09:38, 9 January 2016 (UTC)
Thanks! this is very helpful -Panintelize (talk) 15:06, 9 January 2016 (UTC)
Watch for redirects when you do this. If the project has a sizable scope, then it may have some outdated tags that on the talk pages for redirects (e.g., if an article was merged, especially into a subject that you don't really care about). I found two of those out of nearly 150 articles on an en-to-ht list. WhatamIdoing (talk) 00:36, 13 January 2016 (UTC)
Thanks for the note, WhatamIdoing. Will keep in mind that. But at least in my current way most probably I don't have to worry about it, because I don't follow redirects, I stay at them. --Edgars2007 (talk/contribs) 09:00, 17 January 2016 (UTC)

Annoying question at prefs

At Preferences, if I click on any link, I now get a question "This page is asking you to confirm that you want to leave - data you have entered may not be saved." When this feature was first added, we asked for and then got a new option at Preferences → Editing "Warn me when I leave an edit page with unsaved changes" - and it worked. Why is it now being ignored? How can I suppress that message and go straight to the page that I want? --Redrose64 (talk) 21:40, 15 January 2016 (UTC)

I'm getting this too. I have reset my preferences (adding HotCat) recently. Blythwood (talk) 23:42, 15 January 2016 (UTC)
Hey, I got it also. I went to Prefs to see what you were talking about. Didn't click on anything in Prefs. Didn't change a thing. Didn't see anything wrong, so I clicked on my Watchlink. And up pops that message.— Maile (talk) 00:10, 16 January 2016 (UTC)
I've got it too. Using Firefox as my browser. Only started happening this afternoon. Lugnuts Dick Laurent is dead 18:57, 16 January 2016 (UTC)
Hmm, it wasn't supposed to do that.. :) Should be fixed in a couple of days. Sorry about the confusion. —TheDJ (talkcontribs) 15:19, 17 January 2016 (UTC)

Inflation calculator for large values

I mostly work on historical articles where the unit of currency is in the millions or billions. I have always tried to use the inflation template to convert to today's values, like this : "$100 million ($250 million today)"

I would like to begin using fmt=eq instead of typing "today", but I cannot figure out how to do it in these examples. There is the difficult to type inflation-year example, but I am not going to use that.

Maury Markowitz (talk) 19:05, 16 January 2016 (UTC)

@Maury Markowitz: this page gets lots of attention for quick answers to such questions when necessary, but for the benefit of the template users, and those examples in their documentation, please consider asking at Template talk:Inflation. How's it going? — CpiralCpiral 21:06, 16 January 2016 (UTC)
There is an additional issue, for those who are unaware, in that there are different inflation indices for different purposes. I am not sure, however, that these specialised indices are more appropriate than a consumer price index. Certainly for calculating rebuilding costs, RIBA's building price indices are invaluable, and for economic study percent of GDP may make more sense. However for understanding the scale of expenditure RPI/CPI figures give it in terms of "about 1000 times what I spend in a year" or "about 10,000 family saloons".
All the best: Rich Farmbrough, 04:34, 17 January 2016 (UTC).

Bug in display of footnotes

Sorry to be a bore, but is the bug whereby footnotes on talk pages are displayed at the bottom of the page rather than at the bottom of the thread to which they relate ever going to be fixed? Please? Pretty please? 109.145.180.55 (talk) 03:26, 17 January 2016 (UTC)

The template {{reflist-talk}} inserts all references used at that point of insertion. --MASEM (t) 03:32, 17 January 2016 (UTC)
This has no doubt been mentioned before as a workaround, but it is unsatisfactory. The person originally adding the footnotes will in all probability not bother to do this because, to them, everything looks fine, as their thread will be the last on the page. The next person to come along and add a new thread will be the first to see the problem. They most likely don't know about that template, and anyway, it shouldn't be their problem. There is, I would suggest, essentially never a case when footnotes should correctly go at the bottom of the talk page. They always need to go at the end of the thread. The software should take care of it. It should be automatic. 109.145.180.55 (talk) 03:44, 17 January 2016 (UTC)
See T70324, currently open at Low priority. — CpiralCpiral 04:20, 17 January 2016 (UTC)
We keep getting qs like this. Most recent is Wikipedia:Village pump (technical)/Archive 142#References in section. --Redrose64 (talk) 11:16, 17 January 2016 (UTC)
Yes, it is a perennial irritation. 86.183.129.68 (talk) 21:29, 17 January 2016 (UTC)

Emails not getting through

On both of the last two occasions I've attempted to send an email through Special:EmailUser, the intended recipient has told me that they haven't received it in either their inbox or their spam folder. Does anyone have any idea why that might be? HJ Mitchell | Penny for your thoughts? 14:47, 18 January 2016 (UTC)

Wikipedia:Village pump (technical)/Archive 136#Email is not working. --Edgars2007 (talk/contribs) 14:51, 18 January 2016 (UTC)

17:56, 18 January 2016 (UTC)

Searching external links & citations

Remind me, please, what the tool or trick is for searching for a specific website in our external links (including in citations). Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 18:49, 18 January 2016 (UTC)

Special:LinkSearch? SiBr4 (talk) 19:06, 18 January 2016 (UTC)
That's it, thank you. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 19:54, 18 January 2016 (UTC)
LinkSearch can't do this: insource: "http yahoo brand edgar". — CpiralCpiral 00:27, 19 January 2016 (UTC)

Problem using the characters [ and ] in a link

I'd like to add the following external link to a page:

"https://www.digitalcommonwealth.org/search?f[subject_facet_ssim][]=Markets&f[subject_geographic_ssim][]=Boston"

But when I tried adding code containing this link to the page and I preview the result, what I see is this:

[subject_facet_ssim[]=Markets&f[subject_geographic_ssim][]=Boston]

Evidently, the presence in the link of the [ and ] characters corrupts the reference.

Can this problem be solved?

NewtonCourt (talk) 21:38, 17 January 2016 (UTC)

Percent-encode the brackets as %5B and %5D, like this: https://www.digitalcommonwealth.org/search?f%5Bsubject_facet_ssim%5D%5B%5D=Markets&f%5Bsubject_geographic_ssim%5D%5B%5D=Boston. SiBr4 (talk) 21:54, 17 January 2016 (UTC)

That did it. Thanks very much. NewtonCourt (talk) 12:27, 19 January 2016 (UTC)

Not receiving email confirmation code

After my post earlier I was informed that there are problem with Yahoo mail and the Wikipedia email system, so I created a new Gmail address and changed the address in my preferences, but I haven't received the confirmation code. I've re-tried four or five times but it's not working, and it's not going to the spam folder either (the email address is brand new and has never received any spam). Apologies if this is a known issue but I'd appreciate any help. HJ Mitchell | Penny for your thoughts? 21:21, 18 January 2016 (UTC)

Does your new address contain any special characters? Can you verify you can receive inbound mail from other sources? — xaosflux Talk 21:36, 18 January 2016 (UTC)
No special characters (it's just hjmwiki at gmail.com, the simplest thing I could think of that wasn't taken), and yes I can receive email from other sources. HJ Mitchell | Penny for your thoughts? 21:47, 18 January 2016 (UTC)
I just tested another one to Gmail successfully, did you check the "All Mail" folder? Are you checking from a browser or a client? — xaosflux Talk 23:23, 18 January 2016 (UTC)
I just used "Email this user" on your talk to test that aspect of your mail. Please report if you receive anything. Johnuniq (talk) 04:00, 19 January 2016 (UTC)
Thanks guys, the problem's solved now. HJ Mitchell | Penny for your thoughts? 17:05, 19 January 2016 (UTC)

Template bleed in to page footer

Hi. The Teahouse template placed here bleeds down on to the page footer: User talk:Murph9000#Teahouse host invitation. I notice most of the Teahouse host templates suffer from a similar issue. Cheers! {{u|Checkingfax}} {Talk} 01:44, 20 January 2016 (UTC)

The problem is almost certainly due to the position:absolute; CSS. "Do not leave space for the element."[20]. I'm looking at the way the HTML and CSS is working for it, not yet certain just how simple a fix it is. --Murph9000 (talk) 01:59, 20 January 2016 (UTC)
@Checkingfax:   Fixed It turned out to be quite simple, just a quick shuffle of the HTML elements, and eliminate the troublesome part of the CSS. Only tested in Safari right now, but the HTML and CSS is sufficiently clean looking that problems in other browsers should hopefully be unlikely. --Murph9000 (talk) 02:22, 20 January 2016 (UTC)

Handling of <br /> tags broken

It appears that Wikipedia no longer honors existing styles or font sizes when two or more line breaks are placed in text. No matter what font size is in effect, the <br /> or <br> tag produces absurdly large vertical spacing between paragraphs. Here are some examples:

Example 1

This is a paragraph with default font size.

Note the spacing between lines.

Example 2

This is a paragraph with font size "-3".

Note how the spacing between lines stays the same as in the first example, instead of becoming proportionally smaller.

Example 3 (inserted answer)   Done

This is a paragraph using actual opening and closing paragraph <p> tags with it's inline font-size: styling set to "x-small" (~ to old "-3").

Note how the spacing between lines are now proportionally smaller too, as the OP requested.


It seems to me that this used to work correctly in Wikipedia and vertical spacing followed the font size/style currently in effect, keeping them roughly in the correct proportions. It certainly doesn't behave the way HTML and XHTML normally do in a Web page. The effect doesn't depend on which Wikipedia skin is in effect, as far as I can determine. — QuicksilverT @ 16:25, 18 January 2016 (UTC)

If the font changes, the spacing should follow. I opened T124000 for vertical changes, and T123904 for horizontal. — CpiralCpiral 04:22, 19 January 2016 (UTC)
user:Hydrargyrum, once you get the line wrapping to start chiming in, it doesn't look that bad? The last column uses the Extension:poem tags MediaWiki gives us to use instead of br tags, but they seem exactly the same.

Line followed by two <br /><br />.

Note the
acceptable line spacing between this line, started with two line-breaks, and the previous line; both inside a normal font.



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



Curabitur pretium tincidunt lacus. Nulla gravida orci a odio. Nullam varius, turpis et commodo pharetra, est eros bibendum elit, nec luctus magna felis sollicitudin mauris. Integer in mauris eu nibh euismod gravida. Duis ac tellus et risus vulputate vehicula. Donec lobortis risus a elit. Etiam tempor. Ut ullamcorper, ligula eu tempor congue, eros est euismod turpis, id tincidunt sapien risus a quam. Maecenas fermentum consequat mi. Donec fermentum. Pellentesque malesuada nulla a mi. Duis sapien sem, aliquet nec, commodo eget, consequat quis, neque. Aliquam faucibus, elit ut dictum aliquet, felis nisl adipiscing sapien, sed malesuada diam lacus eget erat. Cras mollis scelerisque nunc. Nullam arcu. Aliquam consequat. Curabitur augue lorem, dapibus quis, laoreet et, pretium ac, nisi. Aenean magna nisl, mollis quis, molestie eu, feugiat in, orci. In hac habitasse platea dictumst.

Line with a <font size="-3">.

Note the unacceptable line spacing

between these line-break lines,

that start with two line-breaks,

and their previous lines.

They should have becoming proportionally-smaller vertical spacing because the font is smaller.



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



Curabitur pretium tincidunt lacus. Nulla gravida orci a odio. Nullam varius, turpis et commodo pharetra, est eros bibendum elit, nec luctus magna felis sollicitudin mauris. Integer in mauris eu nibh euismod gravida. Duis ac tellus et risus vulputate vehicula. Donec lobortis risus a elit. Etiam tempor. Ut ullamcorper, ligula eu tempor congue, eros est euismod turpis, id tincidunt sapien risus a quam. Maecenas fermentum consequat mi. Donec fermentum. Pellentesque malesuada nulla a mi. Duis sapien sem, aliquet nec, commodo eget, consequat quis, neque. Aliquam faucibus, elit ut dictum aliquet, felis nisl adipiscing sapien, sed malesuada diam lacus eget erat. Cras mollis scelerisque nunc. Nullam arcu. Aliquam consequat. Curabitur augue lorem, dapibus quis, laoreet et, pretium ac, nisi. Aenean magna nisl, mollis quis, molestie eu, feugiat in, orci. In hac habitasse platea dictumst.

Here's Epicurus in <p style="font-size:x-small;"> tags now.

Epicurus expressed

a non-aggressive attitude

characterized by his statement:

"The man who best knows how to meet external threats makes into one family all the creatures he can; and those he cannot, he at any rate does not treat as aliens;

and where he finds even this impossible,

he avoids all dealings, and, so far as is advantageous, excludes them from his life."

  Done
It really doesn't look so bad when you get away from lines and into paragraphs. — CpiralCpiral 06:39, 19 January 2016 (UTC)
If you actually set the font-size attribute's value on the paragraph element and not through the now long deprecated font element (see newly added Example 3 at top) there no longer seems to be any spacing issue(s) under the wiki-markup. If it still works outside the wiki mark-up environment; consider yourself "lucky" that "old" behaviors still work for you I suppose. -- George Orwell III (talk) 23:58, 19 January 2016 (UTC)
And I've replaced table column3 three with that solution, showing how well it works in larger forms, (and I've filed the solution at phabricator and closed that ticket.) See WP:HTML5 for all such deprecated usage. — CpiralCpiral 21:17, 20 January 2016 (UTC)

API question

The following is a copy of an email sent to OTRS:

I'm currently conducting research that requires me to know the Wikilinks (as in, hyperlinks to other Wikipedia articles) within the text of an article. For example, the page on "Pistol-Whipping" (sorry I wish I had a more tasteful example, but this one is convenient to work with for many reasons--short, only available in English, etc.), which started off only 2 sentences long (below) and currently is a bit longer with more links:

To pistol whip someone means to hit someone on the head with a gun, usually to either knock them unconscious and/or cause them pain. It is a well known phrase in american popular culture, and was once featured on an episode of "The Simpsons". I want to obtain the following word-values: pistol, gun, unconscious, pain, american, popular_culture, The_Simpsons.

Article today (current version): https://en.wikipedia.org/wiki/Pistol-whippingArticle's oldest version (day it was created): https://en.wikipedia.org/w/index.php?title=Pistol-whipping&oldid=6563362

Getting the page revisions from latest to earliest is with this JSON URL:https://en.wikipedia.org/w/api.php?action=query&format=json&prop=revisions&pageids=30221421&rvlimit=500&rvdir=newer

From the top of this page we see that the revid of the earliest version of the page is 6330157.

To know the revid (revision ID) for the latest version, we use the same URL as above but replace the final word "newer" with "older" thus getting the most recent result (yes the terminology is counter-intuitive). This ID is 693397622.

https://en.wikipedia.org/w/api.php?action=query&format=json&revids=6330157&generator=links OK all this was just to tell you that I now have the revision ID for the first version and the last version of the Wikipedia page for "Pistol-Whipping." Now this is where things get tricky.

I can get the hyperlinks for the most current version of the page with the following JSON URL: https://en.wikipedia.org/w/api.php?action=query&format=json&revids=693397622&generator=links Now, when I repeat the same URL JSON request, but with the *earliest* revision ID of the page's creation, I get the same, CURRENT links, not the ones corresponding to this old revID I'm using. See for yourself: https://en.wikipedia.org/w/api.php?action=query&format=json&revids=6330157&generator=links

I don't get the links I used at the example in the beginning of this email. Instead I get links to "Close combat", "American English", "Baton (law enforcement)", and other Wikilinks that are in the recent version of the page with revID 693397622, not 6330157 (old page) as I requested.

Can anyone help?

Thanks! --S Philbrick(Talk) 14:58, 19 January 2016 (UTC)

You want the links in an old version of a page, but generator=links doesn't seem to work for its older revids. The links reported in
https://en.wikipedia.org/w/api.php?action=query&format=json&revids=6330157&generator=links for revids=6330157
just gives you the current links:
https://en.wikipedia.org/w/api.php?action=query&format=json&revids=6563362&generator=links revids=6563362
Have you tried at mw:API talk:Web APIs_hub and at mw:API talk:Main page? — CpiralCpiral 20:00, 19 January 2016 (UTC)
Even if you try without a generator, like this, it doesn't work. I created phab:T124083 about this, so we can see what the developers say. Nirmos (talk) 20:07, 19 January 2016 (UTC)
I'm pretty sure the current API is not capable of listing links for old versions of articles. Dragons flight (talk) 10:41, 20 January 2016 (UTC)
It's been a while since I last looked at MW's internals in detail, but I seem to remember that link searches are based on a database table which only stores the current links. So, the API can do a relatively simple database query to extract stuff for the live revisions, but it would be extremely expensive to do it for any other revisions (and therefore isn't available). I could be completely wrong about that, my memory of it is vague, and quite a few versions out of date. --Murph9000 (talk) 11:18, 20 January 2016 (UTC)
If this is what you want, you basically need to import old versions into a mediawiki instance and then retrieve the information from the database. There is no other way to do it as far as I can see. Remember that you also need working versions of old templates (including the extensions that such old templates etc depend on), if you really want all links. It won't be easy (which is the reason why it's not implemented). For a subset of the articles it might be easier, but you will need to do custom work on a custom MediaWiki implementation for that as well. —TheDJ (talkcontribs) 16:30, 20 January 2016 (UTC)
As User:Δ correctly pointed out to me on IRC, if you accept that any links generated by templates etc are 'new' (instead of as they were back then), than you can approximate this by making per page api requests like: parse api. —TheDJ (talkcontribs) 20:24, 20 January 2016 (UTC)
But then there's the old page staring at us with those big blue links. It seems simple enough to extract them. If it can be done client-side, then it can be moved to the server-side. But then... it's more than API, its API plus post-processing? — CpiralCpiral 20:52, 20 January 2016 (UTC)

Weird code in articles

Does anyone know what causes stuff like this? There are a couple of articles that contain similar stuff. Rong Qiqi (talk) 03:04, 20 January 2016 (UTC)

It looks like what a WYSIWYG HTML editor would spit out. Based on the fact all those articles are based on translations from it.wiki, I suspect someone copied the page into an WYSIWYG editor, made the Italian to English translation, and then copy and pasted the code back to en.wiki, leaving the WYSIWYG elements behind. They don't seem to be anything malicious, just not useful to us. --MASEM (t) 03:21, 20 January 2016 (UTC)
The common factor with all of them is them being created with a standard "Created by translating the page "link to some other language WP article" " edit summary. I'd say it's about 99% likely that it's some automated translation tool which spews bonus garbage in its output, but can't be certain. Yes, while I make no guarantees of the impact of them, I wasn't seeing any real signs of malice, just automated garbage generation. For me, it may or may not have been something involving WYSIWYG editors (or what you see may not be what you get …). It could easily be some tool that hits the source and destination wikis directly (without user involvement at the point of transfer), and just does a bad / buggy job of it. --Murph9000 (talk) 03:31, 20 January 2016 (UTC)
That message would be MediaWiki:Cx-publish-summary, which is used by the ContentTranslation extension. — Mr. Stradivarius ♪ talk ♪ 08:13, 20 January 2016 (UTC)
Content Transcrapulator strikes again. Bgwhite (talk) 09:13, 20 January 2016 (UTC)
Would somebody please, please disable the damned thing until and unless it can be made to work properly. More garbage to clear up is really just all we need. Oh, and it can't do translation, either: "There's also a town library and a theatre, whose name's 'M. A. Galdi'." Back to the drawing board, please (or the round file would do fine too). Justlettersandnumbers (talk) 10:27, 20 January 2016 (UTC)
This is the browser extension grammarly, that corrupts the HTML in VisualEditor/CT. A ticket was already created for this and it now linked above. —TheDJ (talkcontribs) 14:54, 20 January 2016 (UTC)
Thanks everyone! Rong Qiqi (talk) 21:22, 20 January 2016 (UTC)

File redirect pages

Content on redirect pages has been shown since January 2014. However, on file redirect pages such as File:TCOP.jpg, content such as the Redr template is still not shown. GeoffreyT2000 (talk) 05:25, 20 January 2016 (UTC)

There are quite a few known issues with redirects of the file namespace, especially when it comes to redirects to shared namespaces. I suspect someone forgot to make changes to the redirect pages, to this 'exception' as well, causing the behavior to get out of sync. It's also missing the redirect arrow for instance. There are several bugreuports in phabricator around similar issues, i'll try to find a matching one. —TheDJ (talkcontribs) 10:23, 20 January 2016 (UTC)
File redirs also differ from other redirs in that they lack the "bent arrow" symbol that precedes the link. This symbol is produced by formatting the link as the only item in an unordered list, and restyling the list item marker as an image. It actually uses some inline SVG to make that image:
<?xml version="1.0" encoding="utf-8"?>
<svg xmlns="http://www.w3.org/2000/svg" width="47" height="20" viewBox="0 0 47 20">
    <g id="Layer_1">
        <path fill="none" stroke="#000" stroke-width="2" stroke-miterlimit="10" d="M14.98 2.5V11c0 1.04 1.02 1.98 2.02 1.98h6l3 .02"/>
    </g>
    <g id="Layer_3">
        <path d="M23.48 9.5l.02 7L30 13z"/>
    </g>
</svg>
so probably won't work on older versions of IE. --Redrose64 (talk) 10:37, 20 January 2016 (UTC)

Technical issue

I'm encountering an issue with Template:American Experience episodes. I've made edits to the template, but they are not appearing in other articles with the template added them.

Examples include American Experience (season 28) - the number 28 does not appear as the most current season. Nor does content about season 28 appear beneath it. See American Experience (season 27) to contrast difference between the two articles.

Second example, American Experience (season 12) - I made an edit to correct "John Brown John Brown's Holy War" to "John Brown's Holy War" on 06:34, 15 January 2016, but it still appears.

I've been seeing these issues for at least a few weeks. Originally, they appeared to be short delays in edit changes, but now appear to be an indefinite delay. What do I do? Thanks. Mitchumch (talk) 15:24, 20 January 2016 (UTC)

Update: I tried to "force" a recaching of the page by (1) opening the page in edit mode (2) make no changes and (3) press save on the suggestion of a fellow wikipedian. It worked, but I will leave this post up for documentation. Mitchumch (talk) 15:27, 20 January 2016 (UTC)
You need to look at WP:Purge. Unfortunately you need to purge each individual page on which the template is transcluded. --David Biddulph (talk) 15:30, 20 January 2016 (UTC)
The system will automatically purge the pages, but this takes time depending on the length of the job queue. You can see the changes instantaneously by doing a manual purge, as described above. Mamyles (talk) 16:09, 20 January 2016 (UTC)

System just ate a chunk of my Talk page

Hi. I just completed this edit and when the page refreshed the section and several above it were missing. See my TOC User talk:Checkingfax and you shall see this latest edited section is no longer in the TOC. OK, I will try a purge. PS: Earlier tonight X!Tools went down for a bit. Showed last post dates for editors being 2099 instead of 2016. Using: Windows 7 Pro, and Firefox browser (latest build). Cheers! {{u|Checkingfax}} {Talk} 15:57, 20 January 2016 (UTC)

UPDATE: Did a purge to no avail. Before the edit in the Diff above I had 122 items in my TOC and now have 112 items in my TOC. Cheers! {{u|Checkingfax}} {Talk} 16:01, 20 January 2016 (UTC)

UPDATE: My bad. I left an opening bracket off of a ping. Cheers! {{u|Checkingfax}} {Talk} 16:05, 20 January 2016 (UTC)

It broke the page in combination with an open {{ earlier.[21] Never leave open tags, even if they cause no problems at the time. PrimeHunter (talk) 16:10, 20 January 2016 (UTC)

Watchlist pages

So I go to a page, see that it's not on my watchlist, and think to myself, "I swear this was on there". Does this happen to anybody else? This happens here and there for me, I'd say at least once a week. I posted here because it's possible that it could be a bug. It's more than likely just me making a mistake each time. Either way, it would be interesting to see if anybody else experiences the same moments of 'confusion' as I do. —DangerousJXD (talk) 10:20, 12 January 2016 (UTC)

Not happened to me yet DangerousJXD, so it's probably just you   - either way, you might want to report it should it continue to happen? -- samtar whisper 10:31, 12 January 2016 (UTC)
It sometimes happens to me too, for pages I definitely used to watch. My guess is that I accidentally removed the checkmark at "Watch this page" before saving an edit to the page. PrimeHunter (talk) 15:06, 12 January 2016 (UTC)
I get the opposite: pages appearing in my watchlist where I have never edited either the main or talk pages. 2-3 times a month I guess. I did mention this a few years ago, my theory was that somebody else had watched a page but that it somehow got linked to my watchlist. If this is happening, it explains DangerousJXD's observation - they watched a page, but it somehow got linked to someone else's watchlist. --Redrose64 (talk) 15:50, 12 January 2016 (UTC)
Here's my guess at the three most common "accidents":
  • You accidentally bumped the WP:Keyboard shortcuts while visiting the page.
  • You pressed the spacebar when you were tabbed to that spot in the UI but thought your cursor was somewhere else.
  • You aimed for something else, but accidentally clicked the watchlist icon because a script loaded just as you were clicking.
Whatamidoing (WMF) (talk) 19:36, 12 January 2016 (UTC)
That last one has happened to me more times than I care to remember. I go to click View history and the star icon shifts under my mouse pointer at exactly the same instant as my click. I'm not aware of any instances where I failed to notice and correct, but that doesn't mean it hasn't happened. Damned frustrating for impatient people like me, having to wait an extra second for things to stop moving around. ―Mandruss  19:46, 12 January 2016 (UTC)

You might want to see Wikipedia:Village pump (technical)/Archive 143#Watchlist is not working. It can be fixed. --Obsuser (talk) 20:23, 12 January 2016 (UTC)

@Mandruss: It might be because you have Twinkle turned on. The shift occurs in order to fit the "TW" menu. GeoffreyT2000 (talk) 23:35, 12 January 2016 (UTC)
Sounds likely, but turning off Twinkle is not a solution. I'd be giving up more than I was gaining. But, as Wikipedia annoyances go, this one ain't so bad. ―Mandruss  00:13, 13 January 2016 (UTC)
Justin Lee is somehow on my watchlist. I'm not in the edit history (either main or talk); I don't see why I would ever have visited that page, let alone accidentally watched it. I don't even use Twinkle. --Redrose64 (talk) 21:12, 15 January 2016 (UTC)

Public logs

The Thanks log link is now shown first in Special:Log rather than last. GeoffreyT2000 (talk) 00:48, 15 January 2016 (UTC)

@GeoffreyT2000: Is there any particular reason for this post? Do you think they should be in a different order? — This, that and the other (talk) 08:36, 15 January 2016 (UTC)
It used to say "Show patrol log Show tag log Show review log Show thanks log" and now it says "Show thanks log Show patrol log Show tag log Show review log". GeoffreyT2000 (talk) 21:10, 16 January 2016 (UTC)
I believe that TTO is trying to figure out whether you believe that this change is a problem, or if you only wanted to say that you noticed the existence of the change. Whatamidoing (WMF) (talk) 23:22, 20 January 2016 (UTC)

Annoying font change at ExpandTemplates

At Special:ExpandTemplates, the font in the "Input text:" text box is now proportional, instead of monospace. The box is no longer full-width but is about 67% width, and the vertical gaps between the checkboxes are much larger. This means that on clicking OK, I don't get the "Result" without scrolling down. However, I do notice that the "Result" box is still monospace, and still full-width. Can we have the upper part of the page put back the way it was please? --Redrose64 (talk) 21:46, 15 January 2016 (UTC)

@Redrose64: See mw:MediaWiki UI. Jackmcbarn (talk) 23:23, 15 January 2016 (UTC)
Not a good answer, might as well have been "we want MonoBook to look like a hotch-potch of Vector, mobile site and Facebook". --Redrose64 (talk) 23:29, 15 January 2016 (UTC)
For what it's worth, I agree with you, Redrose. I also think that WMF / MediaWiki have it quite wrong in using abnormally/obnoxiously large controls and spacing on the basis of accessibility. In my opinion, they are fixing something which is outside the scope of things that web developers should do for accessibility. Making controls usable/accessible is an issue for browser developers, not web developers, in my opinion. MW should use the standard HTML controls at their normal/default size and leave it to browsers to solve any accessibility issues with them (most browsers have an easy zoom in/out feature these days, operating systems provide an accessibility magnifier, etc). MW should only concern themselves with making sure that the MW HTML works ok with the features provided by the browsers and operating system, and using HTML's tags and attributes in a way that is good for accessibility (i.e. correct tags for things and using the various accessibility attributes, but not routinely making most things abnormally large).
As for monospace in the input box, I support your call to revert to that, as dealing with code is much more accessible with monospaced fonts. --Murph9000 (talk) 00:06, 16 January 2016 (UTC)
I concur. This was a poorly conceived 'fix' of something that wasn't broken. I don't care so much about the button size or color but the input text window must be monospace and the whole thing show adapt to fit the width of my screen (like the output); there is no reason to artificially constrain the size of the input text window.
Trappist the monk (talk) 12:12, 17 January 2016 (UTC)

More general

It might be a more general change, see this page. MonoBook is supposed to be compact, with thin sharp borders to the boxes and clearly defined input boxes. I shoud not need to scroll down when I didn't before. The non-sharp smudged borders are not clear, I do not have perfect eyesight, this is now an accessibility issue. It was fine as it was before. Together with the bigger gaps, it needs to be reversed - in MonoBook at least. That goes for Special:MovePage too. --Redrose64 (talk) 10:40, 16 January 2016 (UTC)

This is the Apex theme of mw:OOjs UI, which is our library for controls and UI elements (both PHP and JS). I agree that Apex isn't terribly consistent with monobook really. I think it might be better to use the MediaWiki UI theme instead. Or someone can step up and write a whole new theme of course. I can't find right now where and why this got activated for those specific pages however. —TheDJ (talkcontribs) 15:29, 17 January 2016 (UTC)
"Where" is gerrit:255912 for ExpandTemplates and gerrit:251013 for ComparePages. The direct "why" is phab:T107037 and phab:T100161, but neither of those gives actual reasons to do so. Anomie 16:05, 17 January 2016 (UTC)
Also Special:Export. I mean, what is the point of constraining the width? --Redrose64 (talk) 19:51, 18 January 2016 (UTC)

Trouble with images

I'm on the current version of Chrome, Windows 8.1 (list of add-ons available on request).

In the greater majority of cases, when I load an article, any pictures included show a broken image icon, and not the actual picture. This has been consistent for at least a week now.

Does anyone know what's causing this, and more importantly, how to fix it? Seanette (talk) 07:45, 16 January 2016 (UTC)

Pretty much could be anything. If you have any filtering stuff added to your browser, like ad blockers, etc, try testing with all stuff like that disabled. Something quick to try is to extract an image URL from the HTML source, and see what happens if you visit that URL directly (i.e. just throw it into the browser's location bar). That quick test will let you see easily if the problem is your browser failing to connect to the server for the image, or if you are connecting ok but getting some kind of error from the server.
One possibility is if you have an over-zealous ISP that is forcing some or all of your traffic through a filter. WMF use HTTPS for everything now, and certain ISP content filter configurations cause that to break due to your browser's security correctly identifying a "man in the middle" security attack/compromise via the SSL certificate mismatch. If you are from a censorship-heavy nation, your government may be forcing your ISP to attempt to block certain images on Wikimedia Commons. I do not know if that has ever happened for WMF servers, but I have witnessed it happening first hand while trying to view an entirely legal image on a general purpose image hosting site. --Murph9000 (talk) 08:00, 16 January 2016 (UTC)
I tried changing a setting in Avast Antivirus, which so far seems to be helping. This was the only site I had that issue with. I'll revisit this topic if it turns out that wasn't it. Thanks for the input! Seanette (talk) 03:47, 17 January 2016 (UTC)

Scrape Wikipedia page revision history

Does anybody know of a tool that will allow me scrape revision history of a Wikipedia page. I am actually only interested in the dates those revisions were made. —M@sssly 14:32, 20 January 2016 (UTC)

I don't know of a tool. There is the revisions API which is probably the preferred way. You could alternatively get a page dump to play with, but that's in the 20+ GB area now for all revisions. --Izno (talk) 14:41, 20 January 2016 (UTC)
You can export the history (well, the latest 1000 revisions) of individual articles as XML using Special:Export. {{Nihiltres |talk |edits}} 17:16, 20 January 2016 (UTC)
You could also get that information using SQL. If you need help, don't hesitate to ask. --Edgars2007 (talk/contribs) 08:26, 21 January 2016 (UTC)

Are there tools to track specific Jstor and Questia (etc.) citations?

I have found both Questia and Jstor access valuable tools. Is there a labs script to track how many times I've inserted either of those citations into articles? — Maile (talk) 15:06, 19 January 2016 (UTC)

You can use Special:LinkSearch to find how many times a particular URL has been used by anyone, but not how many you have added. User:Sadads or someone else from WP:The Wikipedia Library might have other ideas for you. Whatamidoing (WMF) (talk) 23:30, 20 January 2016 (UTC)
You'd also need to account for Help:CS1's/Help:CS2's Jstor linkage (I'm unsure if there are parameters for Questia off the cuff). --Izno (talk) 13:00, 21 January 2016 (UTC)
@Maile66 and Izno: We are using Special:LinkSearch as a proxy for link data, and search dois using the main Wikipedia search tool. See our current process at Wikipedia:The_Wikipedia_Library/Processes. We are working on a better solution at this phabricator Ticket, but its coming along slowly. The way the data will be stored, however, it will be really easy to figure out who added the links, so we can do more cohort based assessment for GLAM-Wiki, TWL donations, or more generally. Sadads (talk) 15:02, 21 January 2016 (UTC)
@Sadads: This won't work. --Redrose64 (talk) 15:47, 21 January 2016 (UTC)

SSL somehow blocking Wikipedia?

We've got a request for WP:IPBE in the UTRS queue from a user who claims that due to "SSL certificate rewrites" they can now only edit Wikipedia by using Tor or other proxies. I don't know much about this sort of thing, but my instincts say this is baloney. Is it? Beeblebrox (talk) 18:15, 20 January 2016 (UTC)

There are some possibilities, but it's hard to diagnose based on "SSL certificate rewrites" alone. Some ISPs have badly configured filtering which intercepts HTTPS, and good browsers with strong security preferences will correctly flag a "man in the middle attack". In some cases, that filtering is forced on them by their government. So, basically 2 cases, a clueless ISP intercepting HTTPS for no good reason, or Commons or WP has made it onto a government's web censorship list and forced the ISP to do something fundamentally incompatible with HTTPS. I've personally encountered the censorship issue in the past when viewing a HTTPS page which had entirely innocent and legal imgur images embedded. At the time in question, imgur filtering was mandated by the government to block child porn or similar, but that broke all use of imgur on HTTPS sites. If WP or Commons has made it onto a censor list, it clearly won't be due to that type of pure evil, but will be due to a clash of legitimate freedom of speech vs. evil government regime.
It is technically impossible for HTTPS to cleanly coexist with any form of content filtering performed outside the two end point systems (i.e. anywhere in the middle), by intended design.
Alternatively, it could be baloney, it's impossible to say from here, without far more detailed information to go on.
Murph9000 (talk) 18:34, 20 January 2016 (UTC)
The specific claim is that they mostly edit from their school and that it is there that they experience the problem. The odd thing is, the`account is several years old, but has only like four edits, the last one being in 2014, so there's no way to use CU to see if there is any basis to their claims. I'm sort of inclined to think that if it is their school that is eliberately blocking Wikipedia editing from their own IPs, that is their prerogative. Beeblebrox (talk) 03:27, 21 January 2016 (UTC)
Could also be a poorly configured MITM SSL Proxy server on their network. In any event, I don't see how the school would be good with their users using tor. — xaosflux Talk 03:30, 21 January 2016 (UTC)
I'm not an expert in this area, but I don't believe that it's possible to block only logged-out users from editing a page from your network. You could use a web proxy to drop all URLs that include &action=submit (which would catch previewing and saving pages, but not opening the page to look at the wikitext code), but there's no difference between logged-in and logged-out users there. Whatamidoing (WMF) (talk) 17:40, 21 January 2016 (UTC)
If I understand the assertion correctly, they are saying they can't access Wikipedia at all from these IPs, logged in or not, unless they use Tor. However, they are also claiming to have extensive knowledge of Wikipedia policies despite not even being autoconfirmed and not having made an edit at all in about eighteen months. I've gone ahead and declined the request, and told them they need to submit more specific information if they want IPBE. Thanks everyone for your input. Beeblebrox (talk) 18:04, 21 January 2016 (UTC)

What the heck is going on with the single edit tab?!?

Tech News says the Single edit tab is about to be enabled. When I go to the test server to try it, it still dumps me directly into Visual Editor by default, both when I'm logged out and when I make a new account. When I test it without javascript enabled then it completely blocks be from editing at all! And is the WMF still intending for there to be a pop up menu asking new users a question they obviously have no idea how to answer? And if so, do we really want to tell new users to blindly click to make the pop-up go away, and get chaotically dropped into a random default mode? We know from the May 2015 Visual Editor test that there was zero benefit when new users were offered two editor links, but we don't know if people randomly defaulted into VE mode will be more likely to quit before they ever figure out the extremely obscure button inside VE to activate the Wikitext editor.

Opposed to deployment until these issues get sorted out. Alsee (talk) 10:49, 12 January 2016 (UTC)

Ugh, I am strongly opposed to the visual editor, I think it adds nothing of value to a wiki, and that it actually has overall harmful elements to it. If it were up to me, I'd virtually burn the visual editor and crypto-shred the drives it used to live on! Is there anything remotely close to community consensus for further trying to ram an unnecessary editor down people's throats? --Murph9000 (talk) 12:21, 12 January 2016 (UTC)
@Alsee: If I understand correctly, then if you disabled VE for yourself, then you can ignore that message - it will be the same as it always have been. --Edgars2007 (talk/contribs) 13:44, 12 January 2016 (UTC)
Edgars2007, you seem to have misunderstood. If the test server is an accurate demonstration of what they plan to deploy, then the code is broken and deployment must be halted until this is resolved. If we assume it's just the test server that's screwed up, if we assume the WMF plan to deploy a bug-free-version of what they said it was going to be, then I strongly argue deployment should still be halted until it has a sane implementation. Alsee (talk) 14:29, 12 January 2016 (UTC)
@Alsee: Ok, yes I misunderstood. But editing at test2.wikipedia.org is as I expected. As I haven't disabled VE there, editing mode opens as VE, not wiki mode. Then I can switch to wikitext by pressing those [[]] at the top op page, which says "Switch to source editing". OK, I just know, that I have such option and I know where to look, so it could be made more clear to those, who doesn't know that. And I also can choose some preferences (look at that combobox). --Edgars2007 (talk/contribs) 14:49, 12 January 2016 (UTC)
Actually, there are a few tasks which are rendered more easily in Visual Editor (e.g. mass changes to text formatting), so I'm not a total hater of VE anymore. That said, I would strongly prefer that it would be an editor's personal preference as to whether a single "Edit" tab dumps you into "Edit source" or VE first... --IJBall (contribstalk) 21:37, 12 January 2016 (UTC)
IJBall, this is, indeed, the plan. If you have disabled VisualEditor, then nothing will change for your account at all. But if you have it enabled, then you will see this the next time you edit:
 
If you choose "Always give me the wikitext editor" (the first option), then it will always open in wikitext. Whatever you choose, you will be able to change your preferences at any time, by going to Special:Preferences#mw-prefsection-editing and choosing whatever you want from the drop-down menu:
 
The main goal is to give people what they want, regardless of whether they want one or two buttons, or wikitext or the visual editor first. Personally, I'll be choosing two buttons, because I'm used to it and I regularly use both editing systems (different tools, different strengths). But I know that many experienced editors will choose "always wikitext", and only use the visual editor for things like adding and removing columns from a table in a single click. Whatamidoing (WMF) (talk) 18:39, 13 January 2016 (UTC)

Alsee, I believe that I have explained in other discussions, more than once, that IPs at the English Wikipedia will not see the page open in the visual editor. IPs at the English Wikipedia do not have access to the visual editor by default.

Also, as I have also explained to you elsewhere, more than once, Test2 is not "an accurate demonstration of what they plan to deploy" at the English Wikipedia. Nothing's being deployed anywhere until a few bugs are resolved (e.g., using NoScript in a Javascript-capable web browser), and even when those critical bugs are fixed, the setup at test2.wikipedia is not the setup that will be used here at the English Wikipedia. (It is probably the setup that will be used at some other Wikipedias, such as Catalan, Hungarian, and Portuguese.)

Edgars, you are correct: if you disabled the visual editor in Special:Preferences, then this won't affect your account at all. After it's deployed here, if you enable the visual editor but set this preference to "Always use wikitext", then the only difference you will see is the addition of a small pencil-shaped icon in the upper right corner of the wikitext window. That icon will allow you to temporarily switch to the visual editor if you want to make a one-time edit there (e.g., to add a column to a table). Everything else will remain the same.

Also, to be absolutely clear, Alsee is wrong when he claims that Tech/News says the single edit tab is "about to be enabled". Tech/News actually says that this is a planned future change – "future", as in "any time not in the past or present", which encompasses next month and even next year. No date has been set for any deployment to any non-test wiki. The only firm decision about the deployment is that the English Wikipedia will not be the first location for the deployment. I do not even expect the team to discuss a date for the English Wikipedia until after they've seen how well this system works (or doesn't) at several other wikis. Whatamidoing (WMF) (talk) 19:53, 12 January 2016 (UTC)

Whatamidoing (WMF), yes we've discussed before. I filed T122692 Phabricator two weeks ago which has gotten no response, and your latest reply gives no response to the question I asked. It was unclear from Tech News when deployment was planed, so I took this as a potentially urgent matter. As I indicated in my second comment above I was ready to assume that the test server is a useless demo which doesn't reflect what will be deployed. So fine, we're agreed that isn't what's being done. Can you switch from explaining various things that aren't being done, to helping provide some answer on what the WMF does plan? It the plan still to give new users a dysfunction menu and give new users a random default? (Randomly default VE / default Wikitext / default two-tabs.)
Note that my posts above didn't say oppose single-edit-tab, I opposed deployment until we sort out what it is that would be deployed. Keeping two edit tabs is reasonable, having one edit tab my be reasonable if it doesn't have a bad implementation. Alsee (talk) 01:24, 13 January 2016 (UTC)
Realistically, I expect the Phabricator task you filed to be closed as invalid. Tasks are for actions that need to be done. "This is bad for users" is not an action; therefore, it is not a valid Phabricator task. I suspect that the only reason it hasn't already beed closed is due to the backlog of holidays and travel.
Test2.wikipedia is not "useless". It shows the likely configuration for the majority of Wikipedias, especially for the Wikipedias that are likely to make this transition before the English Wikipedia.
The current plan is:
  1. To fix several critical bugs before deploying the single edit tab to any production wiki.
  2. To use different configurations for different wikis.
  3. To not implement this system at the English Wikipedia first.
  4. To test (at test2.wikipedia) the configuration that is most likely to deployed to the first wiki that will see this system, rather than the configuration that is most likely to be deployed (eventually, perhaps months from now) to the largest wiki.
  5. To identify potential target dates based upon experience at these early wikis, instead of making up arbitrary dates in advance.
The "default" is:
  • to give you the wikitext editor if you have Javascript disabled, and
  • to give you the wikitext editor if you have VisualEditor disabled, and
  • otherwise, to ask you for your preference.
If you do not answer the question, then you will be given whatever editing system you used last time. This information will be collected in advance, for people who have edited before (both logged-in and logged-out). This is not a "random" default.
Note, again, that this is a general statement about the overall plan, for a hypothetical average wiki. It specifically does not apply to IP editors at wikis that do not currently permit logged-out editors to use the visual editor. If IPs are not permitted to use the visual editor at that wiki, then IPs will still not be permitted to use the visual editor at that wiki. There is nothing in here that gives any IPs access to two editing tabs. Whatamidoing (WMF) (talk) 18:39, 13 January 2016 (UTC)
I am trying to address a brand new user. The proposed menu asks them a question they don't understand and forces them to click randomly to make it go away (a bad and potentially stressful design), then sets them to an effectively random default when they click randomly.
Either new users should simply get the site default, or if the the WMF really wants a menu it needs to be a menu that won't freak out new users, one they can rationally deal with. Maybe a radio button for the various options, with the default pre-selected, and a simple "continue" button. At least that way a new user who doesn't understand the options can more comfortably just click "continue". Alsee (talk) 22:40, 13 January 2016 (UTC) Adding ping: Whatamidoing (WMF). Alsee (talk) 06:56, 17 January 2016 (UTC)
(Alsee, adding a ping to an existing comment like that won't work; see mw:Manual:Echo#Technical details. --Pipetricker (talk) 11:14, 17 January 2016 (UTC))
  1. Not all newly registered accounts are actually new editors.
  2. Whether a brand-new editor will see the dialog box at this wiki is still being discussed. They will presumably make a recommendation after the formal user testing is completed.
  3. The "potentially stressful" dialog box doesn't require a response. You can just click "OK" without making a choice. In that case, you will end up in the what you're calling the 'default' editing system, and if you switch after entering the editor, then it will remember that you switched when (if) you ever edit again.
Whatamidoing (WMF) (talk) 23:19, 20 January 2016 (UTC)

Proposing change to Sp-contributions-footer

Since Xtools is down, changing to supercount, go here. --QEDK (T 📖 C) 20:34, 21 January 2016 (UTC)

Oooh, now even SUL is down and both the tools show a 503 Error. --QEDK (T 📖 C) 20:42, 21 January 2016 (UTC)
It's a Labs outage according to this.Jo-Jo Eumerus (talk, contributions) 21:28, 21 January 2016 (UTC)

Having trouble with subheaders

The past couple of days I've noticed a problem with sub-subheaders, the ones created by typing 3 =s on either end of a word. I had written a user subpage using them instead of the larger subheaders, and when I went to edit it I couldn't without going to desktop view and then going to the window from there. I just had the same issue with a new article written by somebody else. If a sub-subheader is under a subheader, I can edit it separately on mobile, but if it's not I have to go to desktop. Has anyone else noticed this? White Arabian Filly (Neigh) 00:22, 22 January 2016 (UTC)

Terminology:
== Heading == is a heading for a section (level 2).
=== Subheading === is a heading for a subsection (level 3).
==== Sub-subheading ==== is a heading for a sub-subsection (level 4).
For reasons of standards and accessibility, you shouldn't use a subheading without a heading at the superordinate level. The guideline WP:Manual of Style/Layout#Headings and sections says "Sections should be consecutive, such that they do not skip levels from sections to sub-subsections".
Notice the different levels aren't just for styling headings in different sizes; they also convey structural information. --Pipetricker (talk) 10:21, 22 January 2016 (UTC)

Having trouble with subheaders

The past couple of days I've noticed a problem with sub-subheaders, the ones created by typing 3 =s on either end of a word. I had written a user subpage using them instead of the larger subheaders, and when I went to edit it I couldn't without going to desktop view and then going to the window from there. I just had the same issue with a new article written by somebody else. If a sub-subheader is under a subheader, I can edit it separately on mobile, but if it's not I have to go to desktop. Has anyone else noticed this? White Arabian Filly (Neigh) 00:22, 22 January 2016 (UTC)

Terminology:
== Heading == is a heading for a section (level 2).
=== Subheading === is a heading for a subsection (level 3).
==== Sub-subheading ==== is a heading for a sub-subsection (level 4).
For reasons of standards and accessibility, you shouldn't use a subheading without a heading at the superordinate level. The guideline WP:Manual of Style/Layout#Headings and sections says "Sections should be consecutive, such that they do not skip levels from sections to sub-subsections".
Notice the different levels aren't just for styling headings in different sizes; they also convey structural information. --Pipetricker (talk) 10:21, 22 January 2016 (UTC)

Hi, a short update from the WMF Community Tech team:

We did a community wishlist survey in December and got a list of things the community wanted to prioritize. We're currently investigating all the top ten wishes, but are also looking at two in more detail: migrate dead links to the Wayback Machine and pageview stats. For more details, see the status report. /Johan (WMF) (talk) 11:08, 22 January 2016 (UTC)

Dead boxes on Article wizard/Ready for submission

The submission boxes go light and inactive when I address the page Wikipedia:Article wizard/Ready for submission that I've always used. Has someone blacklisted me without telling me why?Jzsj (talk) 12:14, 22 January 2016 (UTC)

Jzsj, the submission buttons are disabled until you've typed something into the text boxes, as a result of a change in the implementation of mw:Extension:InputBox. APerson (talk!) 18:05, 22 January 2016 (UTC)

Thanks!Jzsj (talk) 19:14, 22 January 2016 (UTC)

MediaWiki's edit conflict detection broken?

I've just encountered a second case of one of my edits getting clobbered by an edit conflict. The first, I had just put down to a good faith failure to properly handle the edit conflict window (it is a bit complex for less technically skilled people). This second case, it's from a user that I believe would normally handle it properly, and they have now assured me in talk that MediaWiki did not warn them of the conflict, but just silently deleted my comment.

Edits at WP:HD where the problem occurred: [22][23][24]

Murph9000 (talk) 19:16, 20 January 2016 (UTC)

Edit conflict detection has numerous issues unfortunately - see the list at phab:T72163. --AKlapper (WMF) (talk) 13:25, 21 January 2016 (UTC)
Thanks for the info. I guess I've just been lucky in the past. I pretty much always check history and diffs after editing, and in many years of using MediaWiki have found the edit conflict detection to normally be quite reliable. It could well be that the reliability takes more of a hit on a very busy site, as my EN-WP activity is much more recent. Murph9000 (talk) 02:01, 23 January 2016 (UTC)

Logout = login

Nice, right after logout (Special:Logout) you now get the friendly message "You are globally logged in now" (and you are). Is that the great new WMF-tool not to lose any more contributors?! *eg* (Thanks to 213.169.163.106) --.js[democracy needed] 09:28, 22 January 2016 (UTC)

I used nowiki to disable your logout link. Please don't post easter eggs like that. When discussing technical issues, it is best to stick to the facts with plain language. Are you saying that Special:Logout in fact does not log out the user? Johnuniq (talk) 09:44, 22 January 2016 (UTC)

I've posted this at Wikipedia:Administrators' noticeboard/Incidents#No longer able to log out (thank you, WMF!). It happens to me as well and could cause serious problems. It is in any case unwanted behaviour, if someone explicitly clicks "log out" then the system shouldn't simply log them in again. Fram (talk) 11:27, 22 January 2016 (UTC)

I experienced this issue as well earlier today. They need to fix this. It could cause a lot of issues especially if people are editing off public computers. White Arabian Filly (Neigh) 22:25, 22 January 2016 (UTC)
At WP:ANI they declared it being fixed by 17:27, 22 January 2016 (UTC), while this official announcement leaves some questionmarks open for the future. --.js[democracy needed] 00:32, 23 January 2016 (UTC)
It was fixed, about 8 hours ago. The announcement you link is something different. Anomie 00:53, 23 January 2016 (UTC)
I don't know if this is related, but I keep getting logged out every hour or so. FoCuS contribs; talk to me! 01:29, 23 January 2016 (UTC)

User pages in Google search

I have noticed that user pages are not found in Google search anymore. I tried with a few usernames of En WP. Is it a new change? (user subpages are fine). --Tito Dutta (talk) 22:39, 22 January 2016 (UTC)

This has been in effect for years. See Template:User page "noindex=yes – prevents external search engines from indexing the page" — Maile (talk) 22:45, 22 January 2016 (UTC)
For pages that do not use {{user page}}, it's recent. See Wikipedia:Village pump (technical)/Archive 142#User pages deindexed?. Johnuniq (talk) 22:54, 22 January 2016 (UTC)
Noindex is hard-enabled for the entire User, User talk, Draft, and Draft talk namespaces. See InitialiseSettings.php (search for "wgNamespaceRobotPolicies", and look for enwiki):
# ROBOT @{
'wgNamespaceRobotPolicies' => array(
...
	'enwiki' => array(
		NS_USER => 'noindex,follow', // T104797
		NS_USER_TALK => 'noindex,follow',
		118 => 'noindex,nofollow', // draft
		119 => 'noindex,nofollow', // draft talk
	),
Murph9000 (talk) 23:02, 22 January 2016 (UTC)
Which means that if you're trying to find a conversation on a user talk page, and you're having problems finding it with Special:Search, then you used to be able to fall back to your favorite web search engine, but now you can't. If this had happened before the major upgrade to CirrusSearch, I think I would have opposed the change. As it is, almost anything you can do in a web search engine can now be done in Special:Search (and more), so it probably won't prevent people from finding what they need. WhatamIdoing (talk) 02:49, 23 January 2016 (UTC)
@WhatamIdoing: User talk pages have not been indexed for over seven years, see phab:T15890; the recent change - and what Titodutta is asking about - concerns User pages, not User talk pages, and it's phab:T104797. --Redrose64 (talk) 11:28, 23 January 2016 (UTC)

Losing edits

Red alert on Wikipedia and Mediawiki sites. If you leave the edit page, you lose your edits. This abnormal situation just started happening withing the hour. — CpiralCpiral 01:45, 22 January 2016 (UTC)

If you don't understand the complaint, to test this, press any link from a preview page. Then go back. Any change you made should still be there, per WP:TESTLINK (howto test a wikilink from preview page before saving an edit), and per User:Cpiral/Surf your cache. — CpiralCpiral 07:53, 22 January 2016 (UTC)


Turning off Preferences → Editing → live preview (which stores a local copy of the DOM) I then start getting

Document Expired. This document is no longer available. The requested document is not available in Firefox's cache. As a security precaution, Firefox does not automatically re-request sensitive documents. Click Try Again to re-request the document from the website... To display this page, Firefox must send information that will repeat any action (such as a search or order confirmation) that was performed earlier.

Sources say this is caused by Firefox honoring the instructions from Wikipedia servers of the page not to cache or store it, and that this is fixed by going to the php.ini file and changing session.cache_limiter = nocache to public instead of nocache.

On my end (FireFox 43.0.4) I've tried

  • Options → Advanced → Override automatic cache managenemt
  • safe mode
  • clearing browser cache
  • rebooting

CpiralCpiral 07:53, 22 January 2016 (UTC)

I haven't been losing edits but I have started getting "Document Expired", both at WP and at Wikivoyage. I already had Preferences → Editing → live preview turned off and have made no changes to Preferences recently. I am using Firefox and have not updated it recently – I am still using 40.0.3. So the change in behaviour is not due to browser update or any change I have made to my WP/WV preferences. @Redrose64: also reported the "Document Expired" problem above at #Problem with loading Wikimedia sites. Any solution? Or even an explanation of why this has started in the last day or so? Nurg (talk) 22:29, 22 January 2016 (UTC)
I had this problem today, and have had it in the past (can't remember when). Whatever it is, please fix it. Users shouldn't have to be jumping through hoops with special settings just to get a preview page. — Maile (talk) 22:36, 22 January 2016 (UTC)
I have just checked other browsers. The problem is not happening in Chrome, but IE 11.63.10586.0 gives "Webpage has expired", and Edge gives "This page was just here a second ago". I don't normally use these browsers so am not sure if it is new behaviour in them. Nurg (talk) 22:51, 22 January 2016 (UTC)
Chrome 48.0.2564.82 is working correctly. FireFox 43.0.4 and Explorer 11.0.9600.18163 are losing edits. — CpiralCpiral 00:17, 23 January 2016 (UTC)

T124510 thanks! Firefox is now working. Explorer is not working, but I don't know if this is a change for it or not. — CpiralCpiral 02:08, 23 January 2016 (UTC)

Yes, things are back to normal in Firefox for me too. Nurg (talk) 05:42, 23 January 2016 (UTC)
Nope, this is not a change for Explorer; it's always lost text when using the back button. Graham87 13:59, 23 January 2016 (UTC)

cookie / login change?

Anybody know if anything changed today with the login mechanism?
My bot just stopped working. —Steve Summit (talk) 03:07, 22 January 2016 (UTC)

Thank you -- that looks like it's probably it. —Steve Summit (talk) 03:50, 22 January 2016 (UTC)
WP:Huggle login and portions of WP:STiki functionality also seem to be broken, but these are not bot related on their frontends. Per [25], in this context, "our target is to have this rolling out to WMF wikis by the end of February." The timetable on the bot front isn't immediately clear in related postings, but CBNG is also having issues. West.andrew.g (talk) 05:20, 22 January 2016 (UTC)
The Feburary target is about the rollout of AuthManager. This may be related to this morning's deployment of SessionManager with -wmf.11[26]. Chenzw  Talk  07:55, 22 January 2016 (UTC)
@Scs: Your problem is that your "httpget" tool doesn't properly handle cookie expiration. See phab:T124252 for some details. You might do better to just use curl with the -d (for POST) and --cookie-jar (for cookies) options. Anomie 21:26, 22 January 2016 (UTC)
I was so not expecting you to debug (or even to be aware of the existence of!) that tool for me, but: thank you so much! It would have taken me quite a while to discover that issue. —Steve Summit (talk) 21:39, 22 January 2016 (UTC)

Double click highlght

Doubleclick:this

Did both words highlight? In edit mode, same? Did you want them to? I don't want them to. I frequently click User:XXXXX or File:XXXXX and just want the XXXXX highlighted. Is this a browser thing or Wikipedia thing? Can this be changed? Anna Frodesiak (talk) 12:01, 23 January 2016 (UTC)

It's a browser thing. In Firefox, only "Doubleclick" or "this" gets highlighted after double-clicking, not the whole phrase. If you double-click the colon itself, only the colon gets highlighted. Other browsers, like Internet Explorer/Edge, tend to be a bit overzealous when it comes to text selection. — This, that and the other (talk) 12:49, 23 January 2016 (UTC)
So it's Chrome, eh? I hereby rename it Thulium, one of the dullest metals. Thanks you for the valuable info. :) Anna Frodesiak (talk) 12:56, 23 January 2016 (UTC)

Problem with loading Wikimedia sites

User:PhilipTerryGraham

I'll keep it short and simple as I can; sometimes, when I'm editing, Wikipedia simply decides not to work. I can go for a while editing Wikipedia, then suddenly it stops loading. There's no pattern to this, it just decides randomly whenever it wants to stop loading things. Sometimes I can load things after a while, but it loads the page in a primitive HTML read-only version of the page. Sometimes it'll be 10 minutes, or half an hour, or at the worst of times a full hour before it works properly again. And then after that it stops again after a while. It has really crippled my ability to work on Wikipedia articles.

I know for a fact that it's not a problem with the site; I can easily access Wikipedia on my mobile and it works fine; no random breakdowns or anything. It also has nothing to do with my internet connection, since every other website I access works perfectly fine. The problem is limited to my access to Wikimedia sites such as Wikipedia and Wikimedia Commons. I can only deduce that it must be a problem with how I'm accessing Wikipedia on my PC, but I don't know what it is. Does anybody have the slightest clue on what's going on, or how to fix it? :s

Thanks in advance. Philip Terry Graham 15:41, 16 January 2016 (UTC)

Hi Philip Terry Graham - would it be possible to have the browser and version of the browser you're using, as well as the operating system and version? -- samtar whisper 15:50, 16 January 2016 (UTC)
Browser: Firefox 43.0.4, Operating system, Windows 10, version 1511. Philip Terry Graham 15:55, 16 January 2016 (UTC)
Thanks - during a period of Wikipedia not working, what does the browser do? Does it show the previous page while the "loading" icon spins on the tab? Has it ever shown the message that it has timed out? If it happens in a moment, can you try pinging en.wikipedia.org? -- samtar whisper 16:00, 16 January 2016 (UTC)
It simply does exactly as you described in the first scenario. There's never a "connection has timed out" or anything like that come up. The loading wheel just spins until it just stops, and I'm still on the same exact page. Philip Terry Graham 16:06, 16 January 2016 (UTC)
Okay, and when this is happening can you access other webpages? -- samtar whisper 16:08, 16 January 2016 (UTC)
Yeah, that's exactly what I do; when I can't be bothered to wait for Wikipedia to load, I just kill time on other sites like Facebook or Twitter; every other website works completely fine when Wikimedia isn't. Philip Terry Graham 16:10, 16 January 2016 (UTC)
Next time it happens can you try: Windows key + R --> "cmd" --> "ping en.wikipedia.org"? -- samtar whisper 16:23, 16 January 2016 (UTC)
@Samtar: Okay, this was the result. Philip Terry Graham 17:11, 16 January 2016 (UTC)
I'm still having problems with loading Wikipedia at times, as per my statements above. Can anyone else help out? :s Philip Terry Graham 00:19, 18 January 2016 (UTC)
Philip Terry Graham, regarding pinging and your signature:
Your above attempt to ping @Samtar probably failed (so I did it for you now), due to your signature not directly linking to a page with your current username; see mw:Manual:Echo#Technical details, and also the guideline at Wikipedia:Signatures#Internal links.
So you should change the signature in your preferences so it links to your current user page (or talk page or contributions page per the guideline).
Also, I guess (but am not certain) that pinging through a redirected page doesn't work. So trying to ping you using the link from your current signature will fail, like it probably did when User:samtar did that, which is another reason for updating your signature.
Testing to verify: I guess you get no notification for this ping @Philip Terry Graham: Am I right? --Pipetricker (talk) 11:26, 18 January 2016 (UTC)
Pinging User:PhilipTerryGraham. --Pipetricker (talk) 13:51, 18 January 2016 (UTC)
Hi hi, sorry I missed your ping! Those results suggest there is no technical reason as to why Wikipedia/media isn't loading as the hostname is being resolved to an IP (suggesting it's not a DNS issue) and the IP is reachable (not timing out). Could you try using a different browser? -- samtar whisper 12:37, 18 January 2016 (UTC)
Philip, have you tried all the usual things, like WP:BYPASSing the cache? Whatamidoing (WMF) (talk) 23:26, 20 January 2016 (UTC)
Whatamidoing, pinging User:PhilipTerryGraham by linking his former username most likely won't work; see my above comments. --Pipetricker (talk) 10:08, 21 January 2016 (UTC)

Whoops. Sorry about that. I had completely forgot to change my signature after I changed my username a while back. Anyways, I did come across Wikipedia:Bypass your cache earlier, while I was doing my own research into what I could do, but it definitely doesn't do much help when trying to open up an editing page or, most importantly, trying to save edits. Philip Terry Graham 16:50, 21 January 2016 (UTC)

Clarification: I thought it was working... I think my internet connection was simply slower earlier in the day, and I had been reloading pages too quickly and not realising it was acting normal. Caching no longer works for me. It's the old load-forever-then-stop-for-no-reason thing once more. :( Philip Terry Graham 18:06, 21 January 2016 (UTC)
Also, I can confirm that it is a Firefox-only issue; Wikipedia works fine in Microsoft Edge, at least, when it doesn't work in Firefox. However, I ain't using Edge, I want to stick to Firefox, if that's possible. Philip Terry Graham 18:11, 21 January 2016 (UTC)
Nope. It happens on Edge too, just not concurrent with Firefox. Also, Edge simply gives me a white screen, rather than just staying on the previous page, when it stops loading. I feel Murphy's law screwing around with us again! XD Philip Terry Graham 18:41, 21 January 2016 (UTC)
I am now using Google Chrome. Let's see how this goes. Philip Terry Graham 21:25, 21 January 2016 (UTC)
Nope. Google Chrome does the same thing, albeit less frequently. So, the issue is definitely not limited to a single browser. It has something to do with my PC's connection to Wikimedia sites, surely. @Samtar, Whatamidoing (WMF), and Pipetricker. --Philip Terry Graham 22:08, 21 January 2016 (UTC)
As of right now, both IE 11 and Firefox 43.0.4 have issues with Wikimedia sites. I was not having these issues before. User Edit Count brings up a page with the Wikimedia template for it, but all content is missing. User Articles Created brings up the Wikimedia page with a message at the top of the page "No webservice". Global user contributions works as it should. Edit summary search seems to work ok. — Maile (talk) 23:12, 21 January 2016 (UTC)
Fwiw... as far as Edge goes; try disabling Use page prediction to speed up browsing, improve reading and make my overall [browsing] experience better found near the very end of the view advanced settings menu of the settings menu opened by the three dots (•••) in a typical Edge toolbar. I first noticed something was off soon after the Windows Update patches (~ Jan. 13) but only recently resolved Wiki project issue by turning that (new?) option off in Edge.

I also noted that most Internet-Explorer-browser-version-to-Microsoft-operating-systems-support for other than those running Windows 10 officially changed big-time around the same time as that window's update was released. It might be the case that between what is no longer supported by Microsoft in addition to and/or in combination with what the MediaWiki code also no longer supports (i.e not Class A + both going into effect early 2016) is behind some if not all of these recently reported on-WP 'issues' (but I've been wrong before so your mileage may vary :) -- George Orwell III (talk) 09:56, 22 January 2016 (UTC)

@George Orwell III and PhilipTerryGraham: Not something I'm aware of (above) - would you like me to compile this all together for a phab bug report? -- samtar whisper 10:06, 22 January 2016 (UTC)
If it helps, Wikimedia seems to be okay on Google Chrome now, on my end. It's been two days since Chrome hiccuped when accessing Wikimedia sites, so that's a plus. Although I'd be hesitant to confirm that Chrome's a-okay; it could very well screw up for me again anytime soon, if I'm unlucky enough. Philip Terry Graham 19:28, 23 January 2016 (UTC)
In contrast, unfortunately, Microsoft Edge still suffers from the problem. The browser version is 25.10586.0.0, running on Windows 10 version 1511, for those who are wondering. Philip Terry Graham 21:54, 23 January 2016 (UTC)
Did you try disabling Use page prediction to speed up browsing, improve reading and make my overall [browsing] experience better found near the very end of the view advanced settings menu of the settings menu opened by the three dots (•••) in a typical Edge toolbar from before? And have you looked through the features given when you type about:flags in Edge's address/url field yet? -- George Orwell III (talk) 22:24, 23 January 2016 (UTC)
Yes, that was one of the very first things I did when I upgraded to Windows 10. Page prediction was a rather annoying feature in Edge, in my opinion, so I disabled it. As for About:flags, I have not heard of it until now, so I assumed all the selected options were default. I clicked "Reset all flags to default" and confirmed that no options had been changed from their default positions. Is there something here that you'd like me to tweak or play around with? Philip Terry Graham 22:52, 23 January 2016 (UTC)
I didn't know about the prediction setting until it seemed to "start working" last week after the monthly update; but that was not just when visiting Wikipedia. All the issues over the past day or two were not [directly] related to that and seem to have been nailed down/corrected here (running the same OS & browser).

Still if you're seeing similar behavior as before with Edge, I guess the root of the problem lies with something to do with caching. I know there is not much that you can "fiddle" with in Edge compared to old IEe so maybe you should check your User preferences on a lark? I found Live Preview enabled somehow the other week and was the cause of long page save waits on Wikisource for example. -- George Orwell III (talk) 23:33, 23 January 2016 (UTC).

Okay, Chrome just hiccuped for me. It behaved the same way as Firefox and Edge did; constant loading and quitting without any messages and simply not loading the next page and staying on the previous one. It started at ~12:10 GMT and lasted until 12:37 GMT. Now that it's relevant again, I'll give details: Chrome Version 48.0.2564.82 m, running on Windows 10 version 1511. Philip Terry Graham 00:49, 24 January 2016 (UTC)

User:Redrose64 and User:Jenks24

I use Firefox 43.0.4, I have two problems at the moment, which may be two symptoms of the same cause, and have appeared in the last day or so. One is slow loading of Wikimedia sites: the spinny thing which temporarily replaces the favicon in the browser tab sometimes takes a lot longer than it formerly did - whilst it's doing that, Firefox displays "waiting for en.wikipedia.org", "waiting for en.wikipedia.org", or "waiting for upload.wikimedia.org" etc. The other problem is that if I use the "back" arrow button, the previous page isn't just redisplayed - it's refetched, this doesn't just make it a lot slower than it need be, it's a pain in the ass when I'm previewing edits, since the edit window is reset to its initial state and my edit is lost. If I then try the "forward" arrow button, I get the message
Document Expired

This document is no longer available.

The requested document is not available in Firefox's cache.

    As a security precaution, Firefox does not automatically re-request sensitive documents.
    Click Try Again to re-request the document from the website.
Worse, if I carry out a non-edit action like a page move or delete, having done that an attempt to use "back" resends the request to move or delete the page. These reloading/caching problems don't occur with other sites: it's as if my browser is no longer able to cache anything from Wikimedia. Similarly, watchlist checking is now a real bind: having clicked a "diff" link and determined that the edit was OK, I click "back" - and now need to wait several seconds for the watchlist to be rebuilt and reloaded, then the Javascript to stop flicking the page up and down (collapsibles etc. at the top), before I can find the next "diff" link. It always used to cache my watchlist until I explicitly refreshed the browser tab. What's gone wrong? --Redrose64 (talk) 11:07, 22 January 2016 (UTC)
This exact same thing just happened to me. And it would not let me use my tab's arrows to "Go back one page" or "Go forward one page". Firefox 43.0.4, Windows 8.1 — Maile (talk) 21:40, 22 January 2016 (UTC)
I have the same problem as Redrose, also using Firefox 43.0.4. Jenks24 (talk) 12:51, 22 January 2016 (UTC)
IMHO; the "obvious" thing to check is to verify if this also happens on non-wikiproject based sites that are also serving content exclusively through the https//: and SSL/TLS protocols only. If the same thing happens elsewhere as it does under MW today, my guess would be that the change is/was being done by design & not accident -- and, unfortunately, you've "caught up" to the misery IE users have been experiencing for some time now. (Also see Losing Edits below - seems related to FireFox users) -- George Orwell III (talk) 13:24, 22 January 2016 (UTC)
@George Orwell III, Jenks24, Redrose64, and PhilipTerryGraham: All these issues seem related, and paired with a OTRS tech issue ticket I'm dealing with I think this could be quite widespread. I'm going to try to summarise all this up in a phab ticket -- samtar whisper 13:56, 22 January 2016 (UTC)
Issue reported at T124417 -- samtar whisper 14:08, 22 January 2016 (UTC)

Traceroute results

Could affected users please provide a traceroute to en.wikipedia.org? (Windows instructions Linux/Mac instructions) - This would be especially useful from @Redrose64 and PhilipTerryGraham: thanks! Also, could I please have the continents you are each on to better diagnose what's going on? -- samtar whisper 17:33, 22 January 2016 (UTC)

Oh gosh... Am I allowed to ask for a short as possible explain-like-i'm-five explanation of what you want me to do? It's hard to understand and visualize in my head what I'm supposed to do; I'm not exactly an expert in computers and commands, just literature and Wikipedia code. @.@ Philip Terry Graham 21:41, 22 January 2016 (UTC)
@Samtar: Europe. I get this:
Tracing route to en.wikipedia.org [91.198.174.192]
over a maximum of 30 hops:

  1    78 ms    99 ms    99 ms  192.168.1.254
  2    11 ms    10 ms    11 ms  lo0.10.Central10.pcl-bng04.plus.net [195.166.130.177]
  3    11 ms    11 ms    11 ms  irb.10.PCL-CR01.plus.net [84.93.249.81]
  4    11 ms    11 ms    11 ms  ae1.ptw-cr01.plus.net [195.166.129.0]
  5    12 ms    22 ms    24 ms  linx-1.init7.net [195.66.224.175]
  6    18 ms    18 ms    19 ms  r1ams1.core.init7.net [77.109.140.221]
  7    18 ms    24 ms    20 ms  r1ams2.core.init7.net [77.109.128.146]
  8    18 ms    18 ms    18 ms  gw-wikimedia.init7.net [77.109.134.114]
  9    19 ms    18 ms    18 ms  text-lb.esams.wikimedia.org [91.198.174.192]

Trace complete.
However, some of the problems I described are no longer affecting me - I believe they rolled back MediaWiki to the previous version, and my browser cache now seems to be behaving normally; it's also not resetting the edit window on a "back" from preview.
@PhilipTerryGraham: Start a Windows "Command Prompt", and enter the command tracert en.wikipedia.org - it takes a few seconds, and when it says Trace complete., copy everything here. --Redrose64 (talk) 02:06, 23 January 2016 (UTC)
Tracing route to en.wikipedia.org [198.35.26.96]
over a maximum of 30 hops:

  1    <1 ms    <1 ms    <1 ms  home.gateway.home.gateway [192.168.1.254]
  2    18 ms    16 ms    19 ms  bla802.ba.optusnet.com.au [198.142.128.125]
  3     *        *        *     Request timed out.
  4    16 ms    17 ms    17 ms  198.142.250.165
  5    20 ms    17 ms    17 ms  198.142.139.112
  6    17 ms    18 ms    17 ms  198.142.139.128
  7   173 ms   172 ms   216 ms  203.208.192.133
  8   175 ms   175 ms   173 ms  sjo-b21-link.telia.net [80.239.132.13]
  9   173 ms   174 ms   175 ms  wikimedia-ic-308844-sjo-b21.c.telia.net [80.239.192.66]
 10   174 ms   174 ms   173 ms  text-lb.ulsfo.wikimedia.org [198.35.26.96]

Trace complete.
Philip Terry Graham 17:31, 23 January 2016 (UTC)

Editor's new autocorrect feature—any way to turn it off?

I've been using Safari 9.0.2 for a while, so I'm assuming that English Wikipedia's regular editor, which I use, has been upgraded in the past two or three days with a new autocorrect feature.

While I'm in the middle of a word, a popup sometimes appears with a suggested full word while highlighting what I'm typing; worse, it sometimes just applies that popup even when it isn't the word I'm looking for. Assuming I notice that it's mangled my prose, I then have to stop and go back and fix something that was correct.

Is there any way to turn off this "feature" (it offered to change "feature" to "few")? If not, can there be? Some people may like it, but I find it distracting and more than a little time-consuming. Ideally, I'd like to suppress both the autocorrect and the distracting popups, but eliminating either would improve things. I was happy with the red underline indicating a probable misspelled word, since then I can use my own judgment, and I'm glad to see that it remains intact. Thanks for any help you can give. BlueMoonset (talk) 17:06, 23 January 2016 (UTC)

@BlueMoonset: can you please clarify what you mean by "regular editor"? Are you talking about the edit window? I don't get autocorrect with anything, not any gadgets in Preferences or without them. But I'm using Firefox 43.0.4, with Windows 8.1. — Maile (talk) 22:20, 23 January 2016 (UTC)
Maile, I'm talking about the regular edit window when you click on Edit to work on a Wikipedia page, whether an article, template, or talk page. As in what I'm doing right now, writing this reply to you. BlueMoonset (talk) 22:26, 23 January 2016 (UTC)
OK, well @BlueMoonset: I'm not getting that as I type. So, maybe it's specific to how Wikipedia reacts to your browser. — Maile (talk) 22:30, 23 January 2016 (UTC)
Can you try a different browser? This seems like something on your end. — xaosflux Talk 22:42, 23 January 2016 (UTC)
I'd have to install a new one, Xaosflux. All I have at the moment is Safari, and it's all I've used on Wikipedia for years. So you can confirm that this is not built-in behavior for the Wikipedia editor? BlueMoonset (talk) 22:50, 23 January 2016 (UTC)
Are you in mobile view or using a mobile application? What computer type are you using (e.g. Windows, ipad, etc)? The only autosuggest we should have on is for the search box, not the edit box. — xaosflux Talk 23:39, 23 January 2016 (UTC)
I'm not in any mobile view or application; these are regular en.wikipedia.org pages, which are all I use. The computer I've been on today is a Macbook Pro of over four years vintage, upgraded three months ago to OSX El Capitan (10.11.2). If there isn't autosuggest as part of the Wikipedia editor, then this is likely an OSX or Safari thing. Thanks for letting me know; it seems to have stopped for now after I tried a couple of things suggesting by Graeme Bartlett below. BlueMoonset (talk) 02:04, 24 January 2016 (UTC)
You may have to shut down OSX’s built-in spelling corrector. Go to system preferences, keyboard, and uncheck "Correct spelling automatically".
Or in safari there might be an edit button with spelling and grammar on the drop down, then uncheck Correct spelling automatically. Graeme Bartlett (talk) 00:08, 24 January 2016 (UTC)
Graeme Bartlett, thanks. The system preferences/keyboard was was already unchecked. I also didn't have Correct Spelling Automatically checked. I did, however, have the other two checked (Check Spelling While Typing and Check Grammar With Spelling), because I rather like have the colored underlines show up, even if half the time I end up ignoring them. However, once I unchecked those two, the correcting went away along with the popups, so I'm hoping that whatever glitch caused the issue won't object if I recheck Check Spelling While Typing, since I do make typos every now and then, and the sudden splash of color in my wake lets me know that I may have gone wrong. BlueMoonset (talk) 02:04, 24 January 2016 (UTC)

Different section opens in the edit box

This has happened quite a few times now. Whenever, I click [edit source] on a subsection (noticed it on an article first and now today at ANI), it opens some other section for editing. If I go back and click on it, the same behaviour is seen. It seems to contain the wrong edit link until I refresh the page. Using Firefox 43.0.4. Please tell me I'm not hallucinating. --QEDK (T 📖 C) 11:11, 24 January 2016 (UTC)

This can be caused by section reordering or archiving. The edit section links are tied to the order on the page at loading, so if a section is added or removed between the time you load the page and editing it, you can end up with a different section. — JJMC89(T·C) 11:18, 24 January 2016 (UTC)
I didn't pay much attention but I take lots of time so I guess it's entirely possible that it happened. --QEDK (T 📖 C) 11:32, 24 January 2016 (UTC)
@QEDK: Not knowing which article you have in mind, I cannot say what happened there; but at ANI, I'm pretty certain that your experience was caused by this edit occurring at some point between your fetching the page and clicking that edit link. --Redrose64 (talk) 12:58, 24 January 2016 (UTC)
As it happens, that was the thread which came up in place of the one I was trying to edit. --QEDK (T 📖 C) 13:27, 24 January 2016 (UTC)

Prefilled speedy-deletion reasons in deletion screen

Admin technicality: Does anybody know how and where the rules are defined on how templated deletion options are automatically pre-filled in the deletion screen? When a file page has been tagged with certain speedy-deletion templates and you then click the "delete" button, a deletion reason is automatically preselected in the "reason" dropdown accordingly. This works for {{di-orphaned fair use}}, but apparently it doesn't work for {{di-replaceable fair use}}.

How and where can this behaviour be adjusted? Fut.Perf. 16:11, 24 January 2016 (UTC)

Consider {{subst:orfud}}, this is converted to {{Di-orphaned fair use}}. In that, there is the template {{deletable image}} with the parameter |for administrators=[{{fullurl:{{FULLPAGENAME}}|action=delete&wpReason={{urlencode:[[WP:CSD#F5|F5]]: Unused non-free media file for more than 7 days}}}} delete this file] after confirming it is not in use (be sure to check redirects to the file).. That's where it comes from. --Redrose64 (talk) 16:25, 24 January 2016 (UTC)
Thanks. Actually, I just found there's also something else: [27]. This one acts not on the embedded deletion link inside the template, but on the standard deletion button at the top of the screen, via javascript in MediaWiki:Group-sysop.js. Fut.Perf. 16:39, 24 January 2016 (UTC)

Android app: "similar pages" gone

The latest version (2.3.138-r-2016-01-21) of the official Wikipedia app for Android no longer has the "similar pages" dropdown that was generated when {{other uses}}, {{for}} or similar hatnote templates were present. Surely a mistake? Qwfp (talk) 08:21, 24 January 2016 (UTC)

@Qwfp: Thanks for finding this. https://www.mediawiki.org/wiki/Wikimedia_Apps might receive more attention by the mobile application developer team. Or file a bug report directly. :) --AKlapper (WMF) (talk) 00:51, 25 January 2016 (UTC)

Edit throttling

Hi all! What are the edit throttling settings / $wgRateLimits on en.wikipedia.org? The Quixotic Potato (talk) 01:12, 25 January 2016 (UTC)

InitialiseSettings.php:
# wgRateLimits @{
'wgRateLimits' => array(
	'default' => array(
...
		'edit' => array(
			// 8 ed./min per each non-autoconfirmed, or group thereof from same IP
			'ip' => array( 8, 60 ),
			'newbie' => array( 8, 60 ),
		),
Murph9000 (talk) 01:19, 25 January 2016 (UTC)
That was quick! Thank you Murph9000! The Quixotic Potato (talk) 01:21, 25 January 2016 (UTC)

Edit count lnk broken

When I use the "edit count" link on my contributions page, it leads to a 404 error. "The URI you have requested, /xtools-ec/?user=RudolfRed&project=en.wikipedia.org, doesn't seem to actually exist.". Side note: the error message says Xtools.tools is one of the maintainers, but that account doesnt seem to exist. Clicking the link leads to "User account "Tools.xtools" is not registered." @Cyberpower678: are you still one of the maintainers?

RudolfRed (talk) 00:58, 25 January 2016 (UTC)

Works fine for me.—cyberpowerChat:Limited Access 02:03, 25 January 2016 (UTC)
https://tools.wmflabs.org/xtools/pcount/index.php is the correct link now. --Izno (talk) 03:14, 25 January 2016 (UTC)
Umm...no it's not.—cyberpowerChat:Offline 04:07, 25 January 2016 (UTC)
OK the link resolves for me now, but it says that I have 0 edits. Something's off... RudolfRed (talk) 04:25, 25 January 2016 (UTC)
xtools/pcount/index.php just mirrors xtools-ec/index.php. So, does xtools-ec//index.php. --QEDK (T 📖 C) 05:50, 25 January 2016 (UTC)

Use Special:Mypage/common.js to add a tab to the nav bar

I there a way to add a tab to the nav bar (beside the edit, history, and twinkle tabs)? The ReFill docs say to add

mw.loader.load( "https://meta.wikimedia.org/w/index.php?title=User:Zhaofeng_Li/Reflinks.js&action=raw&ctype=text/javascript" );

to Special:Mypage/common.js but this adds it to the side bar under tools. Is there a way to make it a tab (or add it to the twinkle tab which may even be better)? WikiWisePowder (talk) 18:09, 22 January 2016 (UTC)

Yeah, you can add stuff pretty much anywhere on the normal interface from user JS, that is if you write your own JS. The relevant thing is Wikipedia:User scripts/Guide#Portlets, and you want the "p-cactions" group for the links / menus / tabs just above the top of the page content. Murph9000 (talk) 20:49, 22 January 2016 (UTC)
@Murph9000: Thanks I had been looking for this but I couldn't find it. I guess I was searching for the wrong thing. WikiWisePowder (talk) 15:25, 25 January 2016 (UTC)

16:39, 25 January 2016 (UTC)

+1 template

In Template:+1, the documentation from Template:+1/doc does not get shaded green. Only the header and footer are shaded green. GeoffreyT2000 (talk) 22:19, 25 January 2016 (UTC)

  Fixed, for me, by doing a WP:PURGE. I've seen maybe a half dozen cases like that (not just template docs, but articles with huge edit links, and similar, in the last 24–48 hours. It looks to me like one of the servers was a bit dodgy for a time, while parsing source into the cached pages which are actually served to people. Murph9000 (talk) 22:21, 25 January 2016 (UTC)

Viewing tables with mobile devices

What's the story with this? Are tables hard to view? Should tables now be avoided due to the high percentage of smart phones? Anna Frodesiak (talk) 23:00, 19 January 2016 (UTC)

The mobile skin and app seem to handle this fine; have you tried it? Regardless, the answer to the last question is "no". --Izno (talk) 13:06, 20 January 2016 (UTC)
I would say that because this is an encyclopedia, you should always consider the legibility of an article, and that by it's form, a data table is always rather limited in legibility and more so on a small screen mobile device (or a 1995 era large screen for that matter). It's a presentation form suited for looking up a specific fact, if you care enough to spend the time to actually look it up. In the context of a Wikipedia article, nice graphs or properly explained prose can be much more legible and informative and are a more suited form for encyclopedias. The fact that this becomes more noticeable on a mobile device, is just a side effect of this limitation of tables in my opinion. If you think the table is not useful enough on mobile, you can therefor wonder if it's actually a proper Wikipedia article, or just 'filling', a stub or a 'start quality' article. That's how I always judge this.
Nevertheless, some things could be done to improve tables on mobile a bit further. It's just not very easy to do, and our tables are too 'free form' to to reliably do it I presume. Anyways. I encourage anyone who has seen support for large data tables on other websites to actually 'work', to present those interface and UI ideas and file them in phabricator. —TheDJ (talkcontribs) 16:18, 20 January 2016 (UTC)
I have no issues with viewing tables on a mobile phone, and even wrote one here with no problems whatsoever. White Arabian Filly (Neigh) 00:26, 22 January 2016 (UTC)

Thank you Izno,TheDJ, and White Arabian Filly for the valuable feedback. By the way, I use tables mainly when sortability would be desirable to visitors. Anna Frodesiak (talk) 00:13, 26 January 2016 (UTC)

Candidates for speedy deletion

The page Category:Candidates for speedy deletion says that it is a hidden category but it is shown on its members such as Nancy Drew (2002 film) as not being a hidden category. GeoffreyT2000 (talk) 15:26, 24 January 2016 (UTC)

A null edit to the category and to the redirect page has fixed it. I guess that this is related to today's accidental deletion of the category. -- John of Reading (talk) 15:55, 24 January 2016 (UTC)
Sorry.--S Philbrick(Talk) 03:32, 26 January 2016 (UTC)

WM Commons and Meta seem to be redirecting to the WMF site

Anyone else seeing this? Blue Rasberry (talk) 18:37, 26 January 2016 (UTC)

Yes, unable to access CommonsEdJF (talk) 19:18, 26 January 2016 (UTC)
Yes. Aboutmovies (talk) 18:40, 26 January 2016 (UTC)
Yes, and I did have initial problems logging in here too. Whenever I'm now trying to access Commons I'm getting redirected to the generic WMF page. De728631 (talk) 18:41, 26 January 2016 (UTC)
 
Well, at least the file servers are still working. De728631 (talk) 18:46, 26 January 2016 (UTC)
  • (edit conflict) Looks like some new code went wonky (needless to say). Operations is aware and working on reverting it. Everyone take deep breaths and hug a kitten while we wait for them to fix things! Kbrown (WMF) (talk) 18:42, 26 January 2016 (UTC)
Noticed it too. At first, I thought this was some sort of an unannounced change to merge Commons and MetaWiki into one. —k6ka 🍁 (Talk · Contributions) 18:43, 26 January 2016 (UTC)
Reported as phabricator:T124804. LX (talk, contribs) 18:48, 26 January 2016 (UTC)
Thanks for reporting this! There are multiple issues which are currently being investigated in #wikimedia-operations on Freenode IRC (in case you want to watch, but no further reports about it needed). --AKlapper (WMF) (talk) 18:49, 26 January 2016 (UTC)
 
Shark cat eats bad code for breakfast!

While everyone waits, here's another cute problem-solving kitty. Keilana (talk) 19:04, 26 January 2016 (UTC)

Looks like we're back up? Adam Cuerden (talk) 20:09, 26 January 2016 (UTC)

Commons image link doesn't go to Commons

What's going on with File:Codex Sinaiticus Matthew 6,4-32.JPG? It's lacking the Commons description page, and the go-to-Commons link instead takes you to https://wikimediafoundation.org/wiki/File:Codex_Sinaiticus_Matthew_6,4-32.JPG, which has no description page, and its go-to-Commons link instead takes you to https://wikimediafoundation.org/wiki/File:Codex_Sinaiticus_Matthew_6,4-32.JPG, which has no description page, and this is the Song That Never Ends, yes it goes on and on my friends...Nyttend (talk) 18:49, 26 January 2016 (UTC)

PS, now I see that the issue is wider than just this image. Nyttend (talk) 18:51, 26 January 2016 (UTC)

It works for me. Ruslik_Zero 20:26, 26 January 2016 (UTC)

N dashes: reverted edit

I edited NEMA connector page and I changed all hyphens in the text to N dashes because NEMA officially use them instead of hyphens; nevertheless my edit has been reverted on the basis that "use of en dash breaks text search within the page". Is there a soultion for this problem? A template forcing to show n dashes instead of hyphens? Thanks in advance.--Carnby (talk) 19:19, 22 January 2016 (UTC)

If we have a conflict between displaying the correct names and easy to use ones, I suggest to use the correct ones. We don't remove diacritics from names like Sauli Väinämö Niinistö just because they might be hard to type, either. But that is a policy question, not a technical one. —Kusma (t·c) 19:43, 22 January 2016 (UTC)
@Carnby: @Kusma: This was recently discussed, using the NEMA connector article as an example case, here. It's not just that it's "easy to use", by which I think is meant easy to type into the wikitext. The real problem with the en dash (and the em dash for that matter, and the subtraction symbol) is that, without awareness of en dashes and how to type them, you can't use a text search within the page to find them. And it is entirely plausible that a reader might use the broswer's text search to search the NEMA connector article for "L5-30". If it's an en dash in the wikitext they won't find anything, even though the text "L5-30" might be staring at them from their screen. This violates the principle of least astonishment. For that matter, Wikipedia's built-in search (the magnifying glass box in the URC) won't find it either. This means that hacks like having en dashes render as hyphens won't help; nobody is going to going to use Wiki search for info on "L5{{endash_rendered_as_hyphen}}30". Consensus at that Teahouse discussion was to leave them as hyphens, for what I think was good and sufficient reason. What's more important here? Adherence to a distinction of typography that (as @Chris the speller: (who is as strong a proponent of proper dash use as anybody) said at the Teahouse thread) 99.99% of the readers don't care about? or making the information in the encyclopedia findable? Jeh (talk) 20:06, 22 January 2016 (UTC)
One of our standard tests is to check what other high-quality reliable sources do. Do technical publications about NEMA connectors use hyphens or do they use dashes? We should follow the style used by relevant independent reliable sources. —Kusma (t·c) 20:39, 22 January 2016 (UTC)
This is the paper my edits were based on. I also created a lot of redirects in Commons categories (redirecting hyphens to en dashes) since someone there told me there were no problems.--Carnby (talk) 21:25, 22 January 2016 (UTC)
Well, first, I don't know how you can determine anything from that paper. The NEMA type designations appear there only in a scanned image. Neither my browser nor my PDF reader can find any characters in that chart so I have no idea how you are sure that the "-" in "L5-30" is one or the other. To my eye they look like hyphens surrounded by spaces! Which we don't do.
But anyway, I'm not disputing that NEMA might actually use en dashes. My point is that hyphens in the type designations in the article make the article more useful. And commons category names are another thing entirely. You can create redirects for those; WP:TITLE or WP:DASH, I forget which, even suggests that, and I don't think anyone will depend on browser text search to find such.
But that doesn't fix the browser in-page search problem. (For that matter, the WP search tool should equate hyphens to en dashes and even minus signs, not just for article titles but for in-page content too... but this would not fix the browser search problem either.) My "bottom line" is that the article is more useful to readers with hyphens in these type names, even though they're not exactly what's in the NEMA document (or, perhaps, other sources). Jeh (talk) 22:55, 22 January 2016 (UTC)
Emphatic agreement with the principles of least surprise, easy searchability, and usability over obscure typographical "correctness". True Unicode dashes have their place and are preferable to plain old ASCII hyphens in some circumstances, but I'd say not here. —Steve Summit (talk) 21:33, 22 January 2016 (UTC)
Half-hearted agreement. It's only a matter of time 'til browser search ambition catches up to web search ambition. Web searches use stemming on words and folding on letters to find "meaning", and to find more; I predict that all browser page-searches will one day soon allow tapping "next" again and again a few extra times as it jumps from dash to ndash and mdash. The locale of the keyboard should find all reasonable matches, so that all dashes mean the same as a keyboard dash. Then we can have the proper display Carnby is recommending ahead of their time. As for CirrusSearch L5-30 finds L5–30 and L5-30 and L5{{dash}}30 (and L5/30, L5.30, etc.). But yes, asking users to do a browser search for L5 alone and tapping next a few more times while searching for L5–30, all for the sake of "style now!" sounds like a losing debate. Meanwhile, this is an encyclopedia waiting for a browser. — CpiralCpiral 00:07, 24 January 2016 (UTC)
Good point. It's actually a little strange to me that WP's own article search doesn't do a better job of this. As it is we have a MOS recommendation to create redirects for article titles that include dashes, so people typing hyphens will still find the articles they want!
Heck, absent improvements in browsers, WP could even provide a "search in page and highlight" function that could include all "dash equivalence". And, heck, why not ask for regular expressions and "near" functions while we're at it.
(I really wish they'd work on things like that instead of the mobile editing app. Or Flow.)
Meanwhile, though, we should write WP for the browsers we have, not the browsers we wish for. :) Thank you for your response. Jeh (talk) 22:29, 26 January 2016 (UTC)

Cuneiform

(I was directed here viâ the help desk.) I’ve noticed that cuneiform appears properly on my browser only irregularly. I know I’ve installed various cuneiform fonts on several occasions. With the exception of U+122B9, which appears as a black circle, every sign on List of cuneiform signs appears properly. On the other hand, on Lugal all signs (save the one that is not text but an inline image) are seen only as a box with six numbers. The same occurs on Internet Explorer. --75.190.164.194 (talk) 07:13, 26 January 2016 (UTC)

Since most characters show up fine for you, there probably isn't a problem with your browser. Most likely none of the fonts you have installed include those specific characters. As cuneiform is a relatively recent addition to Unicode, I'm not even sure if there is any font that supports all those characters already. —Ruud 19:02, 26 January 2016 (UTC)
Everything is working as intended. List of cuneiform signs uses {{Cuneiform}} to use appropriate fonts (one of which you have installed) for the signs. Lugal does not, so your browser does not know what font to use, so you see a box with numbers. U+122B9 is supposed to look like a black circle. The font Akkadian supports all cuneiform signs in Unicode. Gorobay (talk) 23:47, 26 January 2016 (UTC)

Future IdeaLab Campaigns results

 

Last December, I invited you to help determine future ideaLab campaigns by submitting and voting on different possible topics. I'm happy to announce the results of your participation, and encourage you to review them and our next steps for implementing those campaigns this year. Thank you to everyone who volunteered time to participate and submit ideas.

With great thanks,

I JethroBT (WMF), Community Resources, Wikimedia Foundation. 23:49, 26 January 2016 (UTC)

HTML tag vs. parser function

{{angbr|<span style="font-family: serif">''т''</span>}} produces . {{angbr|{{#tag:span|''т''|style=font-family: serif}}}} produces т. Why doesn’t the former work? Gorobay (talk) 23:57, 26 January 2016 (UTC)

Because it's got an equals sign so you must explicitly number it: {{angbr|1=<span style="font-family: serif">''т''</span>}} produces т. --Redrose64 (talk) 00:07, 27 January 2016 (UTC)

Archivebot issues

There are some disconcerting issues with the archivebots (Cluebot archiving more in particular) discussed at User talk:Jimbo Wales#Conspiracy Unveiled – please contribute to the discussion there, not here. --Francis Schonken (talk) 10:38, 27 January 2016 (UTC)

SineBot not working more often than not, these days?

Until a few months ago, SineBot always worked, and signed every single unsigned post. Since about three months sgo, SineBot actually working has been more the exception than the rule, and I've constantly had to use the "unsigned" template on pages all over Wikipedia. Does anyone know why SineBot isn't working these days? Pinging Slakr, the owner of the bot, who seems to be almost completely retired these days. Is it time to give SineBot to another user, or have another user create a bot that works? Softlavender (talk) 17:50, 25 January 2016 (UTC)

I don't know what its current status is. However, SineBot never worked on every single unsigned post. It doesn't ever sign for most editors who have made 800+ edits (by design), and it doesn't catch some unsigned edits on currently busy pages (e.g., if the page changed before it can sign). Whatamidoing (WMF) (talk) 18:44, 27 January 2016 (UTC)

What's the best way to ask for help (template:help me?) on the talk page of a template? I'd like to see if we could add an optional parameter to Template:DisestcatCountryCentury and Template:EstcatCountryCentury so that would allow for a removal of the millennium category (or ignoring it). There's a large, pretty accurate CFD discussion that century categories for things that don't get into two millenniums don't need a millennium category. The only way to really do that is to substitute and delete it which is just hideous. If so, I can work it out manually on these pages and then we can delete the millennium categories under C1 and so on because that CFD will not be resolved other than manually being done. -- Ricky81682 (talk) 21:22, 27 January 2016 (UTC)

@Ricky81682:   Done, with choices. Ok, looks like you have a choice. I added the option of |milleniumcat=no to Template:DisestcatCountryCentury, and Fayenatic london (talk · contribs) made it depend in the 5th parameter in Template:EstcatCountryCentury. So, take your pick of what you prefer. I'll gladly add milleniumcat to the other one, if you want, and no problem if you prefer the other way. Murph9000 (talk) 21:56, 27 January 2016 (UTC)

File usage History anchors?

If you click on File:Canadian_Journal_of_Physics.jpg#File_usage, or File:Canadian_Journal_of_Physics.jpg#File_history why doesn't it jump to the 'File usage'/'File history' section? This should be fixed! Headbomb {talk / contribs / physics / books} 13:30, 27 January 2016 (UTC)

  • The file history and file usage headers are inserted automatically by the page, they don't work in the same way as normal headers. You'll notice there is no TOC either.Jo-Jo Eumerus (talk, contributions) 13:39, 27 January 2016 (UTC)
You can link to …#filehistory and …#filelinks, if you really want or need to. In terms of whether it should be fixed or not, I'd rate it somewhere down around the very bottom of the pile in terms of priority. Those section links have no use for normal content, they would only really be relevant on talk pages, or a link inside a template (where the anomaly shouldn't be a major concern). Murph9000 (talk) 14:13, 27 January 2016 (UTC)
(edit conflict) The displayed heading depends on user language, for example German. The html source for English says <h2 id="filehistory">File history</h2> and <h2 id="filelinks">File usage</h2>. This means you can make wikilinks File:Canadian Journal of Physics.jpg#filehistory and File:Canadian Journal of Physics.jpg#filelinks. This works in all user languages, but some anchors added automatically by MediaWiki may depend on the version and skin. PrimeHunter (talk) 14:18, 27 January 2016 (UTC)
The point is that "File usage" and "File history" anchors should be there in addition to the usual #filelinks / #filehistory per principle of least astonishment. I agree this is hardly high priority, but it should be implemented at one point. If it can be fixed by editing a local stylesheet, let's do it. If it requires some mediawiki software change, it's another story, but a ticket should be created for it. Headbomb {talk / contribs / physics / books} 14:25, 27 January 2016 (UTC)
I have no opposition to POLA. I'm pretty sure that it would require a change to MediaWiki itself, it's not a stylesheet or template thing, as far as I'm aware. Murph9000 (talk) 14:49, 27 January 2016 (UTC)
The displayed text is determined by MediaWiki:Filehist, MediaWiki:Imagelinks, and all the language variants like MediaWiki:Filehist/de, MediaWiki:Imagelinks/de. The text can be customized by admins at each wiki, and the default text can change in future MediaWiki versions (this may for example have happened when the Image namespace was renamed to File but I haven't checked). If the software starts allowing alternatives to #filelinks and #filehistory then it may create new problems with compatibility and user expectations, for example for users who set a foreign interface language and in wikis with a foreign default language. Should Commons add hundreds of id's so the default displayed text in a local wiki also works if the Commons file page is linked? Few users probably want to make these links and those who do will often be technically minded enough to look for the id in the html source. In many browsers it's easy to find by right clicking the text and choose something like "Inspect element". PrimeHunter (talk) 14:53, 27 January 2016 (UTC)
The anchor ID should be generated to match whatever the specified label text is, and an additional one to match whatever the user's local translated version of the label is, if necessary, plus #filelinks/#filehistory for backwards compatibility. So if the site language is English, but the user's preference is German, there would be #File_history and #Dateiversionen, and #File_usage and #Dateiverwendung.  — Scott talk 15:12, 27 January 2016 (UTC)
The ID cannot change on a per-user basis. It has to be static for the site. So, DE-WP could certainly have a German translation of it, but Commons, Meta, MediaWiki, Wikidata, etc could not. The server does not get the opportunity to redirect or similar based on it, so it has to be consistently available for all users regardless of language preferences. You can't give different IDs to different users unless all of the IDs are in the HTML source simultaneously for everyone. Global site preference, maybe. Per-user, nope, not practical. Murph9000 (talk) 15:32, 27 January 2016 (UTC)
Also, a HTML tag can have no more than one id= attribute, so it's one or the other but not both. --Redrose64 (talk) 23:32, 27 January 2016 (UTC)
The idea (which I don't support) is just to insert extra anchors with no content before the headings like <span id="File_history"></span><h2 id="filehistory">File history</h2>. PrimeHunter (talk) 00:00, 28 January 2016 (UTC)
@PrimeHunter: Yup, that's exactly what I was meaning when I said it could be done as a per-site localisation, an invisible tag adjacent to the tag with the standard ID. It just can't practically be done to have all the different translations at the same time. Doing it per-user is basically a recipe for confusion, as people's links would no longer be reliably shareable. We actually do exactly that for section headings on a non-WMF wiki I work with which has translations. The section headings are localised in the translated subpages, but a null span is added with the English id, so that navboxes and similar (which are common across all translations) work correctly. Murph9000 (talk) 00:09, 28 January 2016 (UTC)

Help needed with Template:Infobox drug

This template seems to be adding broken category code to articles. Near the top of DPT vaccine and BCG vaccine, you can see what I'm talking about. I couldn't see exactly where to fix this, so I'm asking for help here. Would someone please look at this protected template and see if you can find the problem? Thanks. --Auntof6 (talk) 17:49, 27 January 2016 (UTC)

Looking at the history, DePiep (talk · contribs) made some major changes to it today, between 14:28 and 15:32. It has a higher level of edit protection, so nothing most people can do about it. It needs a template editor or administrator to make any changes. Murph9000 (talk) 18:20, 27 January 2016 (UTC)
I don't see a problem on DPT vaccine or BCG vaccine. Please describe the perceived problem better, like quoting unwanted or incorrect text. Is it still there when the page is purged? PrimeHunter (talk) 19:33, 27 January 2016 (UTC)
A sub-template, {{Infobox drug/maintenance categories}} was edited at 18:06, fixing the problem. -- John of Reading (talk) 19:52, 27 January 2016 (UTC)
Indeed.   Done. Sorry it kept you bizzy. (I never met or made this error before: I added a newline as in [[Category:Some category\n|catsortlabel]] which closes the [[ brackets premature!). -DePiep (talk) 22:50, 27 January 2016 (UTC)

Thanks, all! --Auntof6 (talk) 05:54, 28 January 2016 (UTC)

Bogus notification

Special:Notifications is telling me that Mkdw reverted my edit on Korean name, but there's no such edit in the edit history. Thoughts? --SarekOfVulcan (talk) 17:14, 26 January 2016 (UTC)

I've been getting a lot of that too. I'm told that one of my edits were reverted on a page I've never edited before. It's been happening ever since the new notifications system came out. —k6ka 🍁 (Talk · Contributions) 18:35, 26 January 2016 (UTC)
Can you get a diff? Other types of Notifications give diff links, e.g. "Blythewood mentioned you at WP:RD/S" has a link to this edit, which is linked by "View changes". I disabled Notifications for reversions (it removes a temptation toward edit-warring), so maybe the answer is "obviously not". PS, reason — did the system give a reversion that was on a different page, or a reversion of an edit by another user, or something else? Might help to know more specifics about the mistake, if they're findable-outable. Nyttend (talk) 19:12, 26 January 2016 (UTC)
It gave me a link to my original edit, not a reversion. --SarekOfVulcan (talk) 19:38, 26 January 2016 (UTC)
Try to copy-paste the full entry as seen at https://en.wikipedia.org/wiki/Special:Notifications?uselang=qqx. PrimeHunter (talk) 11:42, 27 January 2016 (UTC)
You mean like:
(notification-header-reverted: Mkdw, Mkdw, Korean name, 1)
6 days ago(pipe-separator)(notification-link-text-view-changes)(pipe-separator)Mkdw
? --SarekOfVulcan (talk) 04:17, 28 January 2016 (UTC)
Yes, that's what I mean. It doesn't help explain the problem though. It has exactly the same form as a correct notification to myself about [35]:
(notification-header-reverted: JDC808, JDC808, Wikipedia:Lamest edit wars, 1)
30 days ago(pipe-separator)(notification-link-text-view-changes)(pipe-separator)JDC808
PrimeHunter (talk) 11:06, 28 January 2016 (UTC)

Many links on main site viewed on mobile need to be pressed twice

I often use the non-mobile version of Wikipedia from a phone or tablet as the mobile version is severely lacking functionality and on a tablet the full version of the site displays articles better anyway. I've noticed that many links don't work as expected any more: the first tap does nothing or only displays an underline and the second tap actually activates the link. This is normal behavior for torch-screen browsers to work around the lack of "on hover", however it will be off putting and confusing for many users, who will not expect to have to tap a second time and will find links just don't seem to work. Please fix this bug for mobile users. Safari / iOS / Vector skin. created task: T124980Pengo 21:46, 27 January 2016 (UTC)

I can sometimes see the problem described using my iOS 9.2 device, and I think I partially understand what is happening, but I'm just not certain yet what triggers it. Possibly this behavior as described in the 2012 blogpost has changed a bit in newer versions of Safari or something, possibly it's a Safari bug related to to that described :hover behavior. —TheDJ (talkcontribs) 11:37, 28 January 2016 (UTC)

Drag and drop in search box

Links, such as Wikipedia:Search (https://en.wikipedia.org/wiki/Wikipedia:Search) can no longer be dragged and dropped into the search box. GeoffreyT2000 (talk) 00:30, 28 January 2016 (UTC)

That sounds like a browser feature. What is your browser and have you changed it recently? PrimeHunter (talk) 00:34, 28 January 2016 (UTC)
I can't say it's something I've ever felt the need to do before, but I just tried it in Safari, and it works here, dragging an internal wiki link into the wiki's search box at the top right of the page. It inserted the external URL there (which I personally don't see as being particularly useful, but I don't know your use case). Murph9000 (talk) 00:37, 28 January 2016 (UTC)
(e/c) Which search box, the browser's or the one that MediaWiki provides in the skin of the page ? And please specify which browser you are using. —TheDJ (talkcontribs) 11:40, 28 January 2016 (UTC)

Explaining redirects

Hi there! I've searched for a term a few times on Wikipedia only to be redirected to some place else, often with no explanation why the redirect exists and what does it mean. I realize that I can get to the redirect page by clicking the redirected from link and sometimes the pages contain useful redirected from ... templates that do help a bit to explain the mystery meaning but sometimes this is not enough. Maybe we could have some mechanism that would after redirecting also automagically explain it a bit more (perhaps depending on the existing redirect from and other templates) i.e. it would say redirected from blabla because of blabla... So there! And thanks for all the fish! Palosirkka (talk) 15:42, 27 January 2016 (UTC)

The current software does not make it possible to show different content depending on which redirect was used. If this ability existed then I agree it would often be good to add more details. PrimeHunter (talk) 15:55, 27 January 2016 (UTC)
 
Illustration, current frugal redirect notice
I'm just suggesting augmenting the redirect notice "(redirected from Pichlemo)" with something more informative, like "redirected from Pichlemo, which is the old spelling" or whatever. Just the one-liner, should've explained better the 1st time. Palosirkka (talk) 17:49, 27 January 2016 (UTC)
It's an interesting idea, but would require fairly significant new functionality to be added to MediaWiki. As above, not currently possible with MediaWiki (the software), as there's no per-page control of that specific element of the served page. Murph9000 (talk) 17:52, 27 January 2016 (UTC)
All right, thank you for your time gentlemen! Palosirkka (talk) 18:12, 27 January 2016 (UTC)
@Palosirkka: For the Pichilemu article, you could change {{redirect|Pichilemo}} to something like {{hatnote|"Pichilemo" redirects here, which is the old spelling. For other uses, see Pichilemo (disambiguation).}} GoingBatty (talk) 18:37, 27 January 2016 (UTC)
If we wanted to do something like this then there should be consensus for a new optional explanation parameter in {{Redirect}}. It should not be done by replacing the use on specific articles with a template not intended for redirects. But when the text is also displayed to users coming from other redirects or no redirect, I prefer it to be succinct. Hatnotes are only meant to quickly guide readers to the right page and not to explain something which should be in the article text if it was important. PrimeHunter (talk) 19:26, 27 January 2016 (UTC)

Thinking outloud: Maybe the (redirected from ... ) could include the redirect category? So it could say instead: (redirected from Jose Smith Cuervo‍—‌birth name) or (redirected from Jose Cuervo‍—‌without diacritics). We have to put all the redirects in redirect categories anyway‍—‌might as well put those categories to good use. Or, am I the only one (along with Paine Ellsworth) that puts them in redirect categories? Cheers! {{u|Checkingfax}} {Talk} 11:51, 28 January 2016 (UTC)

Categorization is relatively pointless since we have to write automated checks anyway. Diacritics: 1 line of python, Section redirects: SQL query, Similar names/nicknames: powers Wikipedia:Red link recovery suggestions, Episode list redirects: Fuzzy string match with statistics? Not done yet. Just need to get wikidata people to understand the importance of redirects. — Dispenser 21:56, 28 January 2016 (UTC)

"Undefined" on edit links

On pages where VisualEditor is not disabled, I'm seeing "Edit source <undefined>" and "Edit <undefined>". (Screenshot)

I'm using Firefox 47.0a1 on Windows 10 TH2. This has only started happening recently, and I'm not getting any JavaScript errors, apart from "wg* is deprecated" warnings. nyuszika7h (talk) 22:14, 28 January 2016 (UTC)

There seems to be a small issue with caching of certain old JS code. For me this problem has already disappeared again, how is it for you now ? —TheDJ (talkcontribs) 23:34, 28 January 2016 (UTC)
This is happening to me on every page in article space (so not, for example, this page). I'm using Chrome Version 48.0.2564.97 (64-bit). Everymorning (talk) 03:08, 29 January 2016 (UTC)

Me too. First noticed as it was causing a nasty glitch on a narrow page, repeatedly scrolling left-right and redrawing that part of the page. I tried emptying my browser caches and reloading the page but the problem persists. Safari 9.0.3 (11601.4.4).--JohnBlackburnewordsdeeds 03:19, 29 January 2016 (UTC)

Looking at the HTML and the culprit looks like this:
<span class="ve-tabmessage-appendix"><undefined></span>
So something leftover from when VE was in beta, where it said '[beta]' (though the above appears next to both edit tags). I think I opted in to the beta but there is no option to opt in or out now. Tried clearing my cache again to no avail.--JohnBlackburnewordsdeeds 04:51, 29 January 2016 (UTC)

If anyone is still getting this, try adding ?uselang=qqx to the URL of any page (any language code will work). That fixed it for a few people in IRC. Something something caching is/was apparently the issue. Quiddity (WMF) (talk) 07:26, 29 January 2016 (UTC)

Ah, thanks for the pointer, though I did it another way which also worked. I changed my language via the ⚙ in the left sidebar, then changed it back. The page reloaded without the misplaced "undefined".--JohnBlackburnewordsdeeds 07:53, 29 January 2016 (UTC)

Reporting issue with wikicode, horizontal rule (hr, 4 dashes, hr template)

Greetings, For some unknown reason at Wikipedia talk:Tip of the day#(February 15) - Templates for Age and Dates to get the horizontal rule to appear, I have to add a second blank line before. The same results whether using any of these: <hr> ---- {{hr}}. Even stranger is that the next section that I added today does not have this issue at all. That section title is (February 6) - Articles about books. Regards,  JoeHebda (talk)  17:39, 27 January 2016 (UTC)

@JoeHebda: I'm unable to reproduce it on a quick test. If I edit that section and reduce the double blank lines to singles, it looks fine in page preview, with both lines visible. Have you checked if the hr shows up in the HTML source when it's not visible to you? If it is in the HTML source, but not on your screen, you may have a browser bug. Which browser (and version) are you using? Murph9000 (talk) 17:49, 27 January 2016 (UTC)
(edit conflict) In Firefox 43.0.4 I see a horizontal line below "(Note: a duplicate of Feb. 18)" at [36] Are you saying you don't see a line there? What is your browser? PrimeHunter (talk) 17:52, 27 January 2016 (UTC)
My browser is Chromodo (version of Chrome). Thanks for the View source idea. On one test when I used the four dashes, it still showed as four dashes in the view source & blank on the preview. Whenever I add the second blank line before, it shows the horizontal rule correctly. Might be a browser bug. For now since this is a tip proposal on a talk page, I did a fix with nowiki to force a partial line of dashes.  JoeHebda (talk)  19:17, 27 January 2016 (UTC)
Hi JoeHebda. The system requires section headings to have an extra break before them; maybe it is along the same lines? I have trouble with horizontal rules showing up 100% too. Cheers! {{u|Checkingfax}} {Talk} 11:43, 28 January 2016 (UTC)
@Checkingfax: No it doesn't. For a section heading to be valid, it needs to be the first item on the line, with no text after it - no preceding or succeeding blank lines are required. For a horizontal rule, it's even more lenient: the four hyphen-minus characters merely need to occur at the start of a line, you can put anything you like after them,

like this. --Redrose64 (talk) 00:55, 29 January 2016 (UTC)

Hi Redrose64. If you view-source any page on Wikipedia you will see that the system inserts an extra line break before every heading. An extra line break after a heading, although desirable, is optional.
Also, your line break above did not show up for me until I purged the page. Cheers! {{u|Checkingfax}} {Talk} 08:36, 29 January 2016 (UTC)
Using view-source (by which I think you mean "Edit source") does not modify the page content in any way unless you type something in; and even then, the system only automatically inserts a second line break before a heading if you don't edit the whole page, but instead edit the section before that heading. If you go to User:Redrose64/Sandbox14 you'll see three valid section headings; use "edit source" on that page, and you won't find any blank lines at all except after the last line ("The end.") You have my permission to make textual changes in order to see what happens. --Redrose64 (talk) 09:23, 29 January 2016 (UTC)

Text highlighting bug in wikEd

See screenshots from Walmart.

Text highlighting enabled:
File:Wal-Mart with holes in text using wikEd with highlighted text.JPG

Text highlighting disabled:
File:Wal-Mart plain text using wikEd with highlighted text disabled.JPG

Using wikEd version: 0.9.145c (September 16, 2015), Windows XP Professional SP3 v5.1 rv25.8, and Google Chrome version 47.0.2526.111 m

Cheers! {{u|Checkingfax}} {Talk} 11:31, 29 January 2016 (UTC)

Justify

Hi. The 2nd line of this post: User talk:Coffee#Havacuppa right and left justifies (like Linotype). What's up with that? Ping me back. Cheers! {{u|Checkingfax}} {Talk} 16:56, 28 January 2016 (UTC)

@Checkingfax:. It isn't right justified for me in Firefox. What is your browser? This at top of the page requests justification of lines that wrap: <div style="font-family: Berylium, sans-serif; text-align: justify;">. Maybe this in combination with the collapsing archive box can cause your browser to keep a right justification which is there while the box is expanded and causes the line to wrap. Does it wrap when you expand the archive box? PrimeHunter (talk) 22:23, 28 January 2016 (UTC)
Hi PrimeHunter. Firefox and Chrome. It must be that style= tag and the justify; parameter. It affects the whole page whether the archive box is collapsed or not. I don't care for it   Reminds me of the old days printing on a dot-matrix printer and being able to select right, left, or justified plain text. Looks cobby unless there are enough words to nearly fill the line being justified. Cheers! {{u|Checkingfax}} {Talk} 11:45, 29 January 2016 (UTC)
@Checkingfax and PrimeHunter: I use Firefox, and I see l/r justification all the way down to the bottom of the page, whether the archive box is collapsed or not, but only at soft newlines (phrases that wrap to the next line): right-justification does not occur at forced newlines. This is expected behaviour when the text-align: justify; declaration is in use, which is the case here: that <div> tag is unclosed, so the behaviour that it sets will persist to the end of the page. In the Havacuppa thread, there are two lines of text, both are quite short, so probably only wrap (and therefore justify) on narrow-screen devices.
If Coffee (talk · contribs) wants the justification to cease after the archive box, they need to add a </div> after the }} which close the {{archive box}} template. --Redrose64 (talk) 11:58, 29 January 2016 (UTC)

Help with IF expressions

Hi. Notice the "Width" label in the infobox on this page (the page currently uses this test infobox). That label is generated when any of the following parameters are used: width_min, width_avg, width_max. Depending on the parameter (min, avg, max), a custom header is generated over the value (i.e. "Average width"). I want the following to work:

  • If width_avg is filled, the custom header "Average width" should be auto generated. (seems to work at the moment)
  • There should be a break before that header, and after that header. (the after break works, before doesn't)
  • Likewise, if width_avg is not filled, the section header and the related breaks should not be gererated. (doesn't work)

Can you help fix the code please? This is what the code looks like now (it's one continuous line; broken to three here just for ease of reading):

  • {{#if:{{{width_min|}}}|'''Minimum width:'''}}<includeonly><br/></includeonly>{{{width_min|}}}<includeonly><br/></includeonly>
  • {{#if:{{{width_avg|}}}|'''Average width:'''}}<includeonly><br/></includeonly>{{{width_avg|}}}<includeonly><br/></includeonly>
  • {{#if:{{{width_max|}}}|'''Average width:'''}}<includeonly><br/></includeonly>{{{width_max|}}}<includeonly><br/></includeonly>

--Rehman 12:38, 23 January 2016 (UTC)

@Rehman: I think I fixed it. Should I also do that for depth and rate of discharge or you want to play with it yourself? --Edgars2007 (talk/contribs) 15:48, 23 January 2016 (UTC)
Thanks User:Edgars2007! I'll do the rest. Rehman 01:32, 24 January 2016 (UTC)

Hi. Need help with one more function. Looking at the infobox here, you will notice that there are two columns for the "Tributary" field. What I'm trying to do is: If any one column (left or right) is missing, I want the remaining column to ignore the column formatting, and just follow a single column in list format. Is this possible? Rehman 08:30, 24 January 2016 (UTC)

Anyone? Rehman 08:54, 25 January 2016 (UTC)
  Done, maybe. I threw together something quickly which might do the trick for you, or at least move you close to what you want. Please test thoroughly, I didn't do any real testing on it. N.B. For production use, the documentation should live in the /doc subpage, just in case you don't already know that (but it's fine as-is for sandbox). This is into personal taste, but I'm not sure about center on those lists, I think they might look cleaner with default left justification. Murph9000 (talk) 09:38, 25 January 2016 (UTC)
@Rehman: PING! Murph9000 (talk) 09:40, 25 January 2016 (UTC)
Thanks Murph! I'm on mobile at the moment, I will test it when I get home today. P.s. The doc page is only temporary. Regards, Rehman 10:46, 25 January 2016 (UTC)
@Rehman:No problem. On reflecting on it for a while, I realised my boolean logic wasn't quite right, so my second change should do the trick nicely. Cool, no problem about /doc, I just like to flag any anomalies if I spot them, in case it helps; I did suspect it was just a temp thing in sandbox. Murph9000 (talk) 11:11, 25 January 2016 (UTC)
I agree with you on the centre thing, but if its headers ("Left" and "Right") are also aligned to left, it will look messy together with the "Tributaries" label. Talking about the headers, would you be able to include the headers in the columns? It seems to not be displaying at the moment... (on a separate note, would you also know if this is possible?) Rehman 13:46, 25 January 2016 (UTC)
@Rehman: I'm not quite sure what you are asking about the headers. I'm seeing "Left" and "Right" at the top of the columns at present (in two-column mode). Or are you referring to some other headers? You should probably WP:PURGE, then reload the page, to see if that helps, if you are not seeing them, or check what you see in page preview (as that shouldn't be cached). Remember the thing about previewing a demo instance of a template on the same page as the template, it doesn't reflect changes made to the inner workings of the template until saved and purged. I think the headers would look ok left-aligned with the content, but that's just opinion, and I don't hugely care about it, I'll leave it to you to reach consensus on that with people more strongly motivated towards the content side of that infobox. Might be worth seeing if MOS has any wisdom to offer. As for the globe thing, no, I don't think there's a reasonable way to turn it off; you would need to justify a usage case and achieve consensus for a change to the templates / modules. Murph9000 (talk) 14:15, 25 January 2016 (UTC)
@Murph9000: Two column mode shows the "Left" and "Right" at the top of the columns, but doesn't show in single column mode... Is it the same for you? Thank you! Rehman 14:25, 25 January 2016 (UTC)
@Rehman: Ahhh, ok. Yes, I didn't add a header in single mode, as I thought that was what you intended. It would be easy enough to add a header if one is needed. Murph9000 (talk) 14:28, 25 January 2016 (UTC)
Oh okay. I thought something was wrong from my end. I will play around and see if adding a header is better, and do the necessary. Thank you so much for you help! Rehman 00:01, 26 January 2016 (UTC)

Hi Murph. Sorry, but I can't seem to figure that out... I am trying to make single-column uses to display the list with the header (i.e. "Left"), in the position where "Left" is displayed when both columns are used... Can you do that for me, please? Sorry for troubling... Rehman 12:57, 26 January 2016 (UTC)

@Rehman:   Done, and not a problem, tech stuff is my strong point, and I'm glad to contribute on that side of WP. I think center looks really odd for single column mode (every other thing is neatly left-aligned), but threw it in there anyway, along with some clear comments to ensure you can see which section is used for each situation, for you to make the cosmetic decisions on it (along with anyone else who has a strong interest in the content). Murph9000 (talk) 13:26, 26 January 2016 (UTC)
Thank you, and thanks for being so fast, Murph. I really appreciate it. As for the alignment, I agree with you. I will play around with it more (indents, etc), and if nothing looks better, I'll slot it back to left. Have a great day! Rehman 13:36, 26 January 2016 (UTC)

Template:Check ISO 639-1

Hi again, Murph. If you don't mind, would you be able explain what the below code (from {{Infobox river}}) does? And does it belong in a infobox?

  • {{#if:{{{native_name|}}}|<span class="nickname" {{#if:{{{native_name_lang|}}}|lang="{{Check ISO 639-1|{{{native_name_lang}}}}}"}}>{{{native_name}}}</span>}}

In other words, I want to know if the parameter native_name_lang is useful in the infobox, and/or is it something that doesnt belong in there... Thanks in advance! Rehman 13:59, 29 January 2016 (UTC)

That marks up the phrase "native name" as being from a certain language. This is useful for web browsers (they can serve a smarter font for better accessibility). It makes sense in this context because we expect native name not to be in English for the vast majority of rivers. --Izno (talk) 14:04, 29 January 2016 (UTC)
(edit conflict)@Rehman: Hi there. Check ISO 639-1 checks for a valid language code (and adds a hidden maint category for any incorrect usage, so people can find and fix errors). I don't see any immediate reason why it shouldn't be in an infobox. It's so that you can tag a foreign language name correctly in HTML terms. You can do |native_name=Rhein|native_name_lang=de for the Rhine, for example. Completely correct HTML should always have a lang="something" attribute on an element containing a section of foreign language. The HTML document itself has the default lang for the document, so lang="en" for EN-WP, and lang="de" for DE-WP, etc. If the native name is in English, you can omit the _lang parameter, as it's the same as the document's lang on this wiki. It's an important thing for things like translation, screen readers for the blind, general accessibility, etc. Murph9000 (talk) 14:18, 29 January 2016 (UTC)
Thanks a lot for the replies Izno and Murph, I get it now. I had looked everywhere, and even asked another editor, none of it answered my question, or even mentioned anything about accessibility. Thanks again. Cheers, Rehman 14:50, 29 January 2016 (UTC)
@Rehman: The problem with only allowing ISO 639-1 language codes in infoboxes is that there are a lot of valid language codes that don't conform to ISO 639-1. I was the one who wrote the module in the first place, and it was a bit of a debacle, with perfectly good language codes showing up red error messages all over the place. Some kind of language-code-checking module might be a good idea, but it would need to use a more nuanced approach. — Mr. Stradivarius ♪ talk ♪ 15:03, 29 January 2016 (UTC)
@Mr. Stradivarius: Well, having done a very quick bit of research, the current standard seems to be BCP 47, with the IANA Language Subtag Registry providing a convenient and fairly easy to parse list. In principle, it shouldn't be too hard to have a module checking against the IANA list, if the list is sufficient for our purposes, performing regular updates on it when the list is updated by IANA. Murph9000 (talk) 15:20, 29 January 2016 (UTC)
I decided to IAR, and I've removed all the template transclusions, deleted the wrapper template, and moved the module to Module:User:Mr. Stradivarius/Check ISO 639-1. Let's restore it when we have a module that conforms to BCP 47. — Mr. Stradivarius ♪ talk ♪ 00:34, 30 January 2016 (UTC)

Having trouble making edits

I copy-edit articles for Wikipedia:WikiProject Guild of Copy Editors/Requests, and today I've been editing various articles most of the day. I just started to copy-edit Bristol, and I am unable to make an edit I want to make to the lead. I am also puzzled as to why, in edit mode, all the text of the article is shaded in gray. I have had WikEd enabled for a long time. In edit mode, regular text appears as black letters on a white background, references and block quotes are on a gray background, image files and captions are in lime green, and templates such as "Citation needed" and hidden notes are in pale salmon. I just enabled Visual Editor about a week ago, but I haven't used it. (I click on "Edit source" a the top of the page, not "Edit".) In Bristol, I want to move a phrase in the lead. I highlighted the phrase, right-clicked (as I always do), clicked on "Cut", put the cursor where I wanted to place the phrase, right-clicked, clicked on "Paste", and hit enter. I've tried it four times. Each time, I get a blank screen with an error message. Can you look at the article and tell me whether the text of the article is supposed to appear shaded in gray throughout the article? I don't remember seeing an article like that before. If it is supposed to appear like that, what am I doing wrong? Corinne (talk) 02:48, 23 January 2016 (UTC)

(talk page stalker) See also User_talk:Baffle_gab1978#Bristol; I removed some unknown parameters that were only showing in the normal preview screen from the infobox. I'm unable to replicate this error because I edit without WikEd and without javascript enabled. Cheers, Baffle gab1978 (talk) 04:24, 23 January 2016 (UTC)

https://lists.wikimedia.org/pipermail/wikitech-l/2016-January/084501.html — Preceding unsigned comment added by 92.225.183.8 (talk) 08:41, 23 January 2016 (UTC)

I just tried to make the edit I wanted to make yesterday. When I tried to move a phrase earlier in the sentence through cutting and pasting, I got a blank screen with an error message. When I tried by typing the phrase where I wanted it to be, and deleting the phrase in the later location, character by character, and saved, it worked. So it's the cutting and pasting that doesn't work. Also, why is the text throughout the article shaded in gray in edit mode? Corinne (talk) 15:43, 23 January 2016 (UTC)

(talk page stalker). Hi Corinne. Gray is the color the WP:wikEd text highlighter usually uses for references but for now all body text is also being highlighted the same as a reference. I suspect it has something to do with the JavaScript used in Template:Infobox settlement and I have submitted a trouble report over there on the talk tab, and I also submitted one on the talk tab at Wikipedia talk:wikEd. Cheers! {{u|Checkingfax}} {Talk} 04:57, 24 January 2016 (UTC)
Thanks Checkingfax. As I explained a few minutes ago on my talk page, I have just now been able to resume editing after a full day of not being able to edit. I couldn't even leave a request for help on this page. All of a sudden, when I clicked on "Edit source", at first the edit window looked normal, but very quickly it didn't. It became a narrow band that stretched from the far left to the far right of the screen, and the menu at the left and the Wiki markup at the bottom showed through, in the edit window. I could type something, but when I went to save, I got the box that says, "Do you want to leave/stay on this page?" So basically I couldn't save anything I typed. I undid the edit I had just made to my talk page thinking that might help, but it didn't. I clicked on Preferences, then "Disable visual editor while in beta", and saved. That didn't help. Someone suggested I change to a different web browser, so I downloaded Firefox (I had been using Chrome), logged on to WP, typed a comment on my talk page in an edit window that looked normal, and saved. It worked. I even added another comment below that one. So this seems to have something to do with Chrome. I actually prefer the appearance of the text in edit mode when I'm using Chrome, but I will have to continue using Firefox unless someone can help me figure out what happened with the edit window. Corinne (talk) 23:29, 25 January 2016 (UTC)
Redrose64 Can you help me with this problem of the edit window when I am using Chrome? Note that this last comment is really a separate issue from the copy-and-paste issue described earlier in this section, although they may be related. Corinne (talk) 00:37, 26 January 2016 (UTC)
Hi Corinne. "Snap" which I read in one of your posts somewhere is an error message from Google Chrome that presents itself on basically a blank gray window when Chrome crashes. There is a "Feedback" link where you can describe what you were doing when Google Chrome crashed and it will automatically submit a screenshot and trouble report to Google if you allow it to. Cheers! {{u|Checkingfax}} {Talk} 05:09, 26 January 2016 (UTC)
Thanks, Checkingfax. This morning I uninstalled Google Chrome, then re-installed it. When I went to make an edit on my talk page, after I clicked "Edit", the edit window looked the same as it did before. I hesitate to send a screenshot to Google because my user name is visible. Isn't there anyone on WP who can help me with this? I prefer editing from Google Chrome because it's slightly faster than Firefox. Corinne (talk) 16:17, 26 January 2016 (UTC)
Samtar Redrose64 Can either of you, or anyone else, help me with this? I've edited with Google Chrome for the last two years with no problem, then now suddenly I can't. Corinne (talk) 16:20, 26 January 2016 (UTC)
Perhaps there is: but it's not me, I don't use Chrome. --Redrose64 (talk) 16:38, 26 January 2016 (UTC)

Hi Corinne. I switched to Chrome because IE was too slow, then switched to Firefox, then to Palemoon which is a derivative of Firefox. Palemoon offers a stripped down version that runs faster on my slow computer. Also, in Firefox/Palemoon I can copy/paste columns of data, something Chrome does not allow.

I guess you will have to try Phabricator support (see link above, top of page) and also Google tech support. I never send in the screenshots because the screenshot I would be sending them is nothing germane to the problem. Cheers! {{u|Checkingfax}} {Talk} 18:54, 26 January 2016 (UTC)

@Checkingfax: This won't work. --Redrose64 (talk) 19:55, 26 January 2016 (UTC)
Thanks for the suggestions Checkingfax. By the way, Redrose64 is right. The ping did not work. I just happened to come to this page to look. Corinne (talk) 23:12, 26 January 2016 (UTC)
Checkingfax I looked at Phabricator. It says I have to set up an account on MediaWiki, and I saw the place where it says enter a user name and a password. Even though I have an account on Wikipedia, do I still have to start a new account on WikiMedia? Can I use the same user name? I guess it will be a different password, or can it be the same password I use for WP? Corinne (talk) 23:25, 26 January 2016 (UTC)
Not sure what you mean by "WikiMedia" - Wikipedia is part of Wikimedia, as are Commons, Wikidata, Meta and the others. Your login name and password should be the same on all of these, as your Wikipedia account was created after May 2008. --Redrose64 (talk) 23:39, 26 January 2016 (UTC)
If continue you have problems, try visiting mw:MediaWiki, and make sure your account is recognised properly. Going by meta:Special:CentralAuth/Corinne, your global account does seem to be created ok on www.mediawiki.org. Once your wiki account status seems ok over there, try Phabricator again, and make sure you use the lower "MediaWiki" login/register button below the more obvious login form, to use your WMF global / SUL account. Murph9000 (talk) 00:15, 27 January 2016 (UTC)
Thank you. Corinne (talk) 02:11, 27 January 2016 (UTC) When I looked at MediaWiki, I saw my user name at the top, but it was red. It was also there at Phabricator, and also red. Is that all right, or should it be blue? If it has to be blue, how do I get it be blue? Corinne (talk) 02:13, 27 January 2016 (UTC)
@Corinne: Nah, assuming you are just talking about the usual link to your user page, red shouldn't be a problem. It just means you don't have a user page on that site. You have no obligation to create a user page anywhere, if you don't want to. If you want the red to go away across all WMF wikis, create meta:User:Corinne (but really no need, and kinda pointless creating it as blank). I just have {{#babel:en-N}} in meta:User:Murph9000, so that people on miscellaneous wikis have at least some idea that I won't understand their native language. Your local user page overrides the global one, so the page on meta only appears on wikis where you have not created a page, and you can still create a local page later (after creating a global page). Murph9000 (talk) 03:10, 27 January 2016 (UTC)
Murph9000 Thanks. A bit confusing, but I kind of understand. Enough to know I don't have to create anything. One more question. At Phabricator, do I log on with my regular WP password? Corinne (talk) 03:53, 27 January 2016 (UTC)
 

Hi Corinne. That resetting a ping trick works now and then so I tried just in case  

Murph9000 answered your question, but to reiterate: Look below all those login blanks and there is a smaller button for logging on with your current account. It says: Login or Register MediaWiki with a MediaWiki flower logo on the button to the left of that text. It should just log you right in.

  • Plain new task: click the + button to get to Phabricator. Click the Create Task link located at the top right. You will get a blank form to create a task.

Cheers! {{u|Checkingfax}} {Talk} 04:02, 27 January 2016 (UTC)

Checkingfax I filled out the New Task form at Phabricator and saved, but I don't know if I filled in the top sections right. Now it says "Assigned to Corinne" and then my user name again. Was I supposed to leave that first line blank? It said "Enter a user name", so I entered my user name, even though my user name appeared on the subsequent line. If it's wrong, is there a way I can fix it? Corinne (talk) 17:57, 27 January 2016 (UTC)

  • Corinne Which URL did you land on? I can't see a task started by you.Jo-Jo Eumerus (talk, contributions) 18:08, 27 January 2016 (UTC)
    • It's phab:T124929. I removed Corinne from the "assigned to" field. (When you fill this in, it means that you're planning to write the code that will fix the problem.) Whatamidoing (WMF) (talk) 18:37, 27 January 2016 (UTC)
Whatamidoing (WMF) Thank you for doing that. I just looked at it, and I see Krenair has asked me to take a screenshot. How do I take a screenshot, and, if I can do that, what do I do with it? Where do I put or send it? Corinne (talk) 19:28, 27 January 2016 (UTC)
Hi Corinne. I was just going to say that Whatamidoing (WMF) fixed your ticket. Be sure to bookmark the page. That is the easiest way to get back to Phab. To take a screenshot, one way is to press the PrntScrn button on the top of your keyboard which captures your whole screen to your clipboard, then paste it to Microsoft Paint (found in accessories from your Start menu), then save the image as a .JPG file, and upload it to the Commons, then put the long URL to the image in to your Phab ticket. You can also press Alt+PrntScrn which will only capture the active part of your window. Cheers! Repinging Corinne. {{u|Checkingfax}} {Talk} 21:05, 27 January 2016 (UTC)
Checkingfax, pinging a user by editing existing text doesn't work. The link to the user page and your signature must both be in the same new lines of text, not changed lines. @Corinne. --Pipetricker (talk) 23:14, 27 January 2016 (UTC)
Pipetricker: define "The link to the user page and your signature must both be in the same new lines of text, not changed lines". — Preceding unsigned comment added by Checkingfax (talkcontribs) 12:12, 28 January 2016 (UTC)
You can also do something like add a link to phab:p/Corinne/ on your user page here on EN-WP, or phab: for the front page of the site, or similar links for bugs you want to remember, as a convenient way to find stuff over there. Murph9000 (talk) 21:08, 27 January 2016 (UTC)
Thanks everyone for your comments. I took the screenshot and saved it to Microsoft Paint as a JPEG file. Do I have to upload it to Commons? Then everyone will be able to see what my talk page looks like. Isn't there another way to get it to the Phab ticket? Or can I e-mail it to the editor who is helping me? Corinne (talk) 00:48, 28 January 2016 (UTC)
O.K. I uploaded the screenshot and pasted the URL in the reply box of Phab. Corinne (talk) 01:58, 28 January 2016 (UTC)
In case you need to do this again, you can also drag and drop files directly into a Phab comment (just like most e-mail systems), or you can e-mail them to whoever asked (or me) if it contains sensitive information. I don't know if you found Wikipedia:Screenshots of Wikipedia, but it's probably the best set of directions if you need the screenshots permanently uploaded to Commons. Whatamidoing (WMF) (talk) 02:49, 28 January 2016 (UTC)
@Checkingfax and Corinne: For future ref, don't save the screenshot as a JPEG - save it as a PNG. Further information on file type, cropping (and other preparation), and licensing is given at WP:WPSHOT. --Redrose64 (talk) 10:47, 28 January 2016 (UTC)

Hi Redrose64. Nice tip on the WP:WPSHOT. Great stuff there. However, I did not just fall off the turnip truck and I do not feel that PNG is satisfactory for screenshots while JPG is. PNG is more suited for line drawings whereas screenshots need more vivacity. Should Corinne tune up the licensing on the JPG? Pinging Corinne. Cheers! {{u|Checkingfax}} {Talk} 12:12, 28 January 2016 (UTC)

AKlapper has sent me an e-mail with a few questions to which I will reply, but can I ask you something? By uploading a screenshot to Commons of the edit window open to my talk page, with my user name visible at the top, can visitors to Commons connect my e-mail address to my user name (I haven't looked at it since I uploaded it)? I hope not. Corinne (talk) 18:40, 28 January 2016 (UTC)
Only if the screenshot in question shows your email address.Jo-Jo Eumerus (talk, contributions) 19:17, 28 January 2016 (UTC)
Is anyone following the exchanges at Phab? AK suggested I try using a fresh Chrome user profile and gave me a link. It explains two ways to do it. I don't see how to use the first way. I tried the second way, but don't know what "be sure to launch with...null" means. I asked, but haven't heard back. Corinne (talk) 17:20, 29 January 2016 (UTC)
Hi Corinne. Chrome rolled out a big update yesterday. Be sure you have all your open work bookmarked then click on the 3 horizontal bars in the little square in the upper right corner of Chrome → click on: Settings → click on: About, and that will start an update check and the opportunity to install it → close and re-open all Chrome windows after the update. Cheers! {{u|Checkingfax}} {Talk} 00:04, 30 January 2016 (UTC)
Thanks, Checkingfax -- I did what you told me to do, but when I clicked "About", there was no option to install it; it just said it was installed and up-to-date. I had just installed it yesterday, so maybe that's why. I closed the window, then opened Chrome again, went to my talk page on WP, and looked at an edit window. It hadn't changed. The problem was still there. So I went to "Add and remove programs", uninstalled Chrome, then installed it, checked, and the problem is still there. In that "Settings" place, I had seen a place to report a problem, so I reported without giving my WP user name. I'd like to try what AKlapper suggested, starting with a fresh browser profile (if that is different from what I just did), but there were some things I didn't understand in his last comment. Corinne (talk) 00:36, 30 January 2016 (UTC)

Do we have a bot which could add missing WP talk templates to articles in a category tree?

A lot of articles in Category:Polish sportspeople are missing WP:BIO/WP:POLAND talk templates ({{WikiProject Poland|class=|importance=}}{{WikiProject Biographies|class=|living=}}). Is there a bot that could go through the thousands of related articles in this category and its subcategories and add those templates to the articles which are missing them from their talk? PS. If anyone replies here, please ECHO me. Thanks, --Piotr Konieczny aka Prokonsul Piotrus| reply here 20:37, 29 January 2016 (UTC)

What you need is a bot to carry out a "WikiProject Tagging" run. I know that both AnomieBOT (talk · contribs) and Yobot (talk · contribs) have done this - it's best to file a request at WP:BOTREQ; somebody should pick it up. However, they are unlikely to consider "and its subcategories" - there have been big problems with an indiscriminate recursion like that (see for example User talk:QEDK#Bot), so it's best to provide an explicit list of categories. --Redrose64 (talk) 21:00, 29 January 2016 (UTC)
Forgot to notify Piotrus. BTW see also User:Yobot#WikiProject tagging - it's good advice. --Redrose64 (talk) 21:03, 29 January 2016 (UTC)
See also Category:WikiProject tagging bots. WhatamIdoing (talk) 01:48, 30 January 2016 (UTC)

Templates for image annotation

Hello, it's the first time i'm posting here. Not sure it's the right place but it seems so. Let me know if there is a better place for my question.

I'd like to produce some visual diagrams and insert links to other wikipedia pages.

I found out there are several template for this:

Probably there are also other ones that i didn't found.

So my question is: which one is the best one? Are there best practices for annotated images?

thanks --Mikima (talk) 10:38, 29 January 2016 (UTC)

No, you found them. Those pages will lead you to any related templates. Please see those talk pages, or ask at Wikipedia:Teahouse for clarification issues about those pages. Happy editing! — CpiralCpiral 08:35, 30 January 2016 (UTC)

Cannot log in with Firefox

I get a Login Error "There was an unexpected error logging in. Please try again. If the problem persists, it may be because you have cookies disabled, and you should check that they are enabled in your browser settings" when I try to log in with Firefox (v43.0.4). Cookies appear to be allowed. I can log on successfully with Microsoft Edge and with Chrome.Nigel Ish (talk) 10:26, 30 January 2016 (UTC)

Strangly enough, I can log on with Firefox on my Android Tablet, but not on my PC.Nigel Ish (talk) 10:32, 30 January 2016 (UTC)

Complicated template syntax causing phantom redlinks, not shown, but found by tools?

Recently, the Missing Topics tool has been publicised to help projects find articles that are frequently redlinked. However, it seems to also find some phantom redlinks, which are not on the actual article, but are listed on the redlinked article's "what links here" page. One example is Hopman Cup (tennis). The Missing Topics tool and "what links here" show 22 wanted links - but none of the "what links here" pages show a redlink to Hopman Cup (tennis).

As expected, they all do have bluelinks to Hopman Cup, and they do have the complicated tennis tournament infobox, which includes a subtemplate of Template:Infobox tennis tournament year/footer, which has "#ifexist:" coding to add a " (tennis)" qualifier to a link if it doesn't exist. But as the link does exist, it doesn't add the qualifier to the published version... but it seems to do so in a secret hidden version that is found by the tools. Any idea why? Is this a bug in the template parser or just in how that template's code uses "#ifexist:"? The-Pope (talk) 16:22, 30 January 2016 (UTC)

Help:What links here#Overview says: 'The parser function #ifexist: causes a listing in "What links here" among the normal links even though no link is produced.' PrimeHunter (talk) 17:09, 30 January 2016 (UTC)

Relisting MFD discussions

Does anyone know how I can relist MFD discussions without the bot resetting it back to the original date? I know AFD has a view log or other section for the bot to catch but MFD isn't separated into separate pages like AFD is. Wikipedia:Miscellany for deletion/User:Abeiku Okai/Senya Beraku Akumase Festival keeps getting put back into the January 9th section and not into January 26th. -- Ricky81682 (talk) 09:26, 29 January 2016 (UTC)

@Legoktm: can you check on this Legobot behvaior? — xaosflux Talk 14:39, 29 January 2016 (UTC)
@Xaosflux: Looks like it's added before, twice with no response. If it's going to be not that helpful, I'm not sure what we are supposed to do. -- Ricky81682 (talk) 23:47, 30 January 2016 (UTC)
@Ricky81682: I tried tweaking the location of the timestamp, lets see if it works. — xaosflux Talk 00:11, 31 January 2016 (UTC)
It did not, I've specifically asked the botop to look at this; User:Legobot appears to be doing more overall good then the harm this is causing; if no response we can opt WP:MFD out with {{bots|deny=Legobot}} and either archive manually or request a new bot. — xaosflux Talk 01:53, 31 January 2016 (UTC)
It's a lot of good but it's been years and ignored. I don't want to have to opt the entire thing out for something this minor. Maybe I shouldn't bother relisting discussions. It's not like it isn't doable, AFD discussions work with this but I have no idea how the bot works or what it's picking up so I don't know what we can do with or add to Template:Mfd2 to get this to work. Is there somewhere else where someone with some technical skills can give us an idea? -- Ricky81682 (talk) 03:45, 31 January 2016 (UTC)

Double redirect bots

Has anyone noticed that the double redirect bots are functioning slower than they did years ago, especially in recent history? --Jax 0677 (talk) 16:33, 31 January 2016 (UTC)

Drafts not showing up in category

Hi, all,
Drafts or AFC submisions that haven't been edited for six months or more are eligible for deletion under CSD G13. I've come across many eligible drafts that are supposedly in Category:G13 eligible AfC submissions (look at Draft:Andrew Essex or Draft:Anne Ellingsen for examples) and indicate they are assigned to this category but when you look at Category:G13 eligible AfC submissions, none of these G13 submissions appear.
Do you know of any reasons why a category would appear on a draft page but those articles don't appear to be in that category when you look at its page? This is happening to dozens of old drafts which are appropriately tagged but are not seen by admins who just check the G13 eligible category to see if it contains any pages. Thanks. Liz Read! Talk! 16:02, 31 January 2016 (UTC)

Categories are cached in the categorylinks table. When a page becomes too old, Category:G13 eligible AfC submissions immediately shows up in the page footer, but it may take a few weeks until the categorylinks table is updated, and the pages won't be listed on the page Category:G13 eligible AfC submissions until that happens. If you want pages to appear in the category listing faster, then you could go to the individual pages and purge them with the forcelinkupdate option. I purged the two pages you found, so they now appear in the category listing, but there are probably other missing pages too. A bot could regularly search for pages to purge. --Stefan2 (talk) 16:31, 31 January 2016 (UTC)
I think that it was previously suggested that we send in Joe's Null Bot (talk · contribs) to work through anything transcluding {{AFC submission}} and give it a WP:NULLEDIT. Maybe this was done as a one-off: but if this was done every few weeks (but no more often than once a week), pages wouldn't be left much longer than the six-month threshold and it wouldn't tax the server kitties too much. --Redrose64 (talk) 16:58, 31 January 2016 (UTC)
Well, I don't know about forcelinkupdate but I did refresh the page/purge the cache by clicking on the clock in the upper right-hand corner (which I'm told results in a cache purge) and that didn't make any difference. The only way I found these stale drafts was going through the contributions of the HasteurBot earlier in January, to see which editors it contacted and there must be an easier way to collect these pages into the appropriate category and maybe that can be the work of a bot. It's a little time-consuming to search for stale drafts by going through them individually.
I should add that this is a recent problem, I've spent a fair amount of time deleting G13 and I know that in the fall, the G13 eligible category was FULL, it rarely had fewer than a few hundred stale drafts and I was surprised to find it empty when I first looked at it this month. Liz Read! Talk! 22:05, 31 January 2016 (UTC)
The purge link on that page only updates the page list if you are not logged in when looking at the category, I think. Logged in users should already see the most up-to-date data which is available. You need to purge the individual drafts instead of purging the category. --Stefan2 (talk) 22:14, 31 January 2016 (UTC)

Image not appearing after resizing it

In Thermal management of high-power LEDs the third image (CFD LED Free Convection Heat Sink Design.gif) is no longer displayed after I changed its size attribute to thumb. Would someone kindly have a look at this to see what the problem is? Thanks, Lambtron (talk) 17:16, 26 January 2016 (UTC)

The third image is correctly displayed for me. Have you tried bypassing your browser cache? --AKlapper (WMF) (talk) 18:35, 26 January 2016 (UTC)
The 220 px thumbnail is broken for me, giving
Error generating thumbnail
There have been too many recent failed attempts (4 or more) to render this thumbnail. Please try again later.
Manually choosing a 200px version works correctly. It does seem to be a server side issue. (Hohum @) 19:03, 26 January 2016 (UTC)
All I see is white, except for a little black X in the top left corner. I've never seen this article before, and I get the same result when I clear my cache. The browser is IE 11.63, for your reference. Nyttend (talk) 19:04, 26 January 2016 (UTC)
Dummy edit didn't help. Nyttend (talk) 19:05, 26 January 2016 (UTC)
Hey, the error message is pretty clear here. "There have been too many recent failed attempts (4 or more) to render this thumbnail. Please try again later.". Have you tried not doing any more attempts for a while and then trying again later? :) All thumbnails in that article display for me now. Matma Rex talk 21:20, 26 January 2016 (UTC)
I tried many times at various intervals and then waited several hours and tried again -- it still won't render. Browser cache clearing has no effect, in either IE or FF. How much longer must I wait, and what constitutes a "failed attempt", anyhow? Does someone else's "failed attempt" restart the server's retry timer for me, too? If so, how will I (or others experiencing this problem) ever get it to render? Lambtron (talk) 21:50, 26 January 2016 (UTC)
I even tried reuploading it as File:CFD LED Free convection heat sink design.gif (different capitalisation), but its 220px thumbnail, like the other's, fails to load properly. Nyttend (talk) 06:00, 27 January 2016 (UTC)
I tried forcing a 221px thumbnail but this also results in an error message. However, the 320px version works, so I changed the wikicode to use that one. This looks a bit ugly as this file is bigger than the other ones, but it's better than seeing nothing at all. --Stefan2 (talk) 15:06, 27 January 2016 (UTC)

Not displaying for me on any browser (IE, Chrome, or Firefox) on my PC. Softlavender (talk) 06:20, 27 January 2016 (UTC)

Played with it in a sandbox, FF 44.0. With 320px coded (original size), works fine. Anything else, same as in the article. FWIW. ―Mandruss  10:49, 27 January 2016 (UTC)

Not displaying for me on any browser (IE, FF or Chrome). Cleared browser cache and tried again 6 hours later, and then next day on another computer that has never viewed the page -- no joy. I still don't know what "too many recent failed attempts" means, so I can't even speculate what the problem is or how to work around it. BTW, page visitors will never see the "failed attempts" message; they only see what appears to be a broken image link. Lambtron (talk) 14:01, 27 January 2016 (UTC)
I see that Stefan2 changed the size attribute from thumb to 320px and now the image displays. However, this doesn't solve the problem: it won't display as thumb. Lambtron (talk) 17:52, 27 January 2016 (UTC)
I only see this as a temporary solution. It's usually inappropriate to hard-code the file size, and it looks ugly when one image is slightly larger than the other images. However, it was the only solution I could think of which would make the picture display in the article, and seeing a slightly too big image seemed better than seeing no image at all. --Stefan2 (talk) 20:50, 27 January 2016 (UTC)
  • Definitely a server side problem. It works happily as a thumb at 160px and 80px, ie, exactly an half and a quarter of the original size. I think it just that the thumb maker cannot scale this animated GIF by certain factors. But I note that the image page uses a 107px thumbnail! — RHaworth (talk · contribs) 21:01, 27 January 2016 (UTC)

@Lambtron: Try purging the pix cache by inserting two curly brackets followed by 'purge' (without the apostrophes) followed by two curly brackets, just before the call to the pix is made. Akld guy (talk) 22:26, 29 January 2016 (UTC)

@Akld guy: This won't work. Also, instead of writing that out longhand, write {{purge}} --Redrose64 (talk) 23:54, 29 January 2016 (UTC)
@Redrose64: Hi, it did work when the new version of the map at List of Auckland railway stations persistently failed to load after the map was redrawn. I didn't write the purge thing out because I thought it might do something funny to this page, but tyvm for the tip on writing it with tlx inside the brackets. Learned something and noted it for future. Akld guy (talk) 00:04, 30 January 2016 (UTC)
@Akld guy: By "this won't work" I mean that your attempt to notify Lambtron will have failed: you didn't satisfy the conditions described at WP:Echo#Triggering events. --Redrose64 (talk) 00:14, 30 January 2016 (UTC)
@Redrose64: I'm sorry, I cannot see what you're driving at. It looks like I did everything right. Can you be a little more specific? Akld guy (talk) 00:21, 30 January 2016 (UTC)
@Akld guy: In this edit you altered "Try purging the pix cache ..." to "{{ping|Lambtron}} Try purging the pix cache ..." From that I assume that your intention was to cause a notification to be sent to Lambtron. This will not have occurred, since you need to add the link in a fresh post that includes a fresh signature in the same edit. Similarly, this edit will not have notified you. --Redrose64 (talk) 10:34, 30 January 2016 (UTC)
Rose speaks the truth. For a more long-winded version, see my recent mini-essay at User talk:Boomer Vial#Pings (permalink). ―Mandruss  13:25, 30 January 2016 (UTC)

I edited the page: inserted Purge tag immediately before 3rd image and changed image size to thumb. After saving, a Purge link appears at top of page. After clicking Purge link the image is still not displayed. Then I cleared the browser cache and reloaded the page -- image still not displayed. I've removed the Purge tag but left image size as thumb. Any other suggestions? Lambtron (talk) 19:10, 30 January 2016 (UTC)

Thanks all for making it clear. Akld guy (talk) 19:20, 30 January 2016 (UTC)

Is there some other way I should be using Purge? It has no effect when used as suggested above (unless I'm misunderstanding the instructions). Lambtron (talk) 18:21, 31 January 2016 (UTC)

I can only tell you how it worked for me. I asked another editor to bring his map up to date so it would bring the List of Auckland railway stations article up to date. He did so, but the changed map repeatedly failed to appear. I inserted {{purge}} on a line by itself immediately before the call to the map. That immediately fixed the problem. Didn't need to click the 'Purge' that appeared on the article. I deleted the {{purge}}. At my request, he then made several other changes to the map, all of which needed the same purge inserted. It got to the point where I decided to simply leave the purge there so that any future updates to the map will automatically be reflected. Akld guy (talk) 19:50, 31 January 2016 (UTC)
The page is automatically purged if you edit the page. --Stefan2 (talk) 20:41, 31 January 2016 (UTC)
Yes; it sounds as if a WP:NULLEDIT would have worked. --Redrose64 (talk) 20:48, 31 January 2016 (UTC)
Apparently it was the edit itself (not the Purge tag) that fixed the problem for Akld guy, because the Purge link was never clicked. However, the problem I'm having appears to be a server-side issue that's not related to caching, because neither null edits nor Purge clicks have been effective. I'm guessing that the server is timing out while resizing the image because of the large file size and high frame count, and consequently every attempt to resize is a "failed attempt". Can anyone else corroborate this or offer ideas for how to resolve this? If not, perhaps a new version of the image should be uploaded and substituted into the article -- one that has a smaller file size and fewer frames. Lambtron (talk) 04:44, 1 February 2016 (UTC)

Viewing on iPhone

Whenever I visit this page on the iPhone, the text looks smaller than usual and the section headers are not collapsed. 172.56.30.227 (talk) 20:07, 31 January 2016 (UTC)

I've also experienced a similar problem on iPhone, on iOS 8 and 9 (I use an iPhone 5), though it is that the TOC text headers are of varying sizes. Makes it very annoying when dealing with long TOC's. Is this a registered bug perchance? Cheers, Doctor Crazy in Room 102 of The Mental Asylum 06:33, 1 February 2016 (UTC)

Qs about external links search special page

Regarding: https://en.wikipedia.org/w/index.php?title=Special:LinkSearch

  1. Is there a way to limit searches to just mainspace, or any other individual namespace?
  2. Is there a tool to return the total number of links, immediately?
  3. Is there a way to combine searches for http://www.example.com https://www.example.com http://example.com & https://example.com (or any pair of those)?
  4. Is there away to automate a batch of searches (say, return the totals for five subdomains in each of 20 languages; per namespace); or someone who could do this, please?

-- Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 15:41, 30 January 2016 (UTC)

Mw:help:linksearch: No (it only reports all namespaces). No. No, and Maybe. But forgetting you mentioned Special:LinkSearch:
For an example of how to go about applying a regexp to find external links see Help:Searching/Draft#Insource. A. (The draft is pretty much done.)
  1. Yes. Use the new insource search parameter. As in linking a namespace can go first (case insensitive, aliased, accepts whitespace delimiters).{{search link}} can take a search domain in the form of a namespace profile, (as can the Special:Search Advanced dialog).
  2. No. In general there is no tool or way to pattern the many possible text patterns indicating the general external link. But see T121379 comment by EBernhardson. A Special:Search of course returns the total number of your specific pattern, such as template: insource: "http www example com".
  3. No. Logical OR is not currently available for search parameters, (as it is for words or phrases).
  4. For searching other languages, Search can parse/analyze words in 33 languages. For batch automation, perhaps see the recent topic "reminderbot". — CpiralCpiral 19:21, 30 January 2016 (UTC)
See User:Johnuniq/Links to example which was my attempt to track down bad "example" links in articles in 2011/2012. That list is now obsolete, but I could generate it again. I started that due to a discussion somewhere, but I'm afraid I let it slip. That is generated by a script which uses mw:API:Exturlusage. I don't have time at the moment to work out an example URL, but it should be possible to do some manually. If you spell out what is wanted (I'm not sure what #4 means), I'll look at it within a day or two. Re #2: I don't think you can get a total easily; I think the API has to work its way through the list. You could probably effectively get a total with some SQL at Quarry, but I haven't done that. Johnuniq (talk) 01:53, 31 January 2016 (UTC)
Yes, counting links in SQL is possible and pretty easy. You could tell, which URLs and languages you have in mind. --Edgars2007 (talk/contribs) 17:03, 31 January 2016 (UTC)

If anyone wants to follow up on links to example.com, I have updated my list of links to example. Johnuniq (talk) 09:56, 1 February 2016 (UTC)

Reminderbot?

Do we have reminderbot-style capability anywhere? My scenario is that I'll propose something on a talk page with the idea of giving others about a week to discuss, then I'll forget about it. It would be cool to have a template or something trigger a reminder notification to me at some future time. E.g. {{remind}} could just send a ping to me pointing to whatever section had that markup after some default interval, or maybe {{remind|5|Follow up on rename discussion}} could send that reminder text in 5 days. Regards, Orange Suede Sofa (talk) 02:53, 30 January 2016 (UTC)

I did propose something similar here, and I proposed it at community wishlist. --Edgars2007 (talk/contribs) 08:04, 30 January 2016 (UTC)
It might be possible to implement something client-side via Wikipedia:WikiProject Microformats#hCalendar / hCalendar. I don't know if any browsers have built in support for it at present, but there may be extensions for browsers to provide support and a link to desktop calendar software. Murph9000 (talk) 08:19, 30 January 2016 (UTC)
It sounds like something a bot could do, just about. Although maybe it would be easier to do in user / user talk namespaces – i.e. if you had a user subpage with reminders in a format which the bot could parse, the bot could then leave you a message on your talk page (triggering a notification) at the appropriate time with whatever note you specify in the reminder. - Evad37 [talk] 00:17, 31 January 2016 (UTC)
There was a similar discussion in 2010 at Wikipedia:Bot_requests/Archive_37#Reminder_bot – pinging (still active) editors from there @Hammersoft, Anomie, and Hellknowz: - Evad37 [talk] 02:55, 31 January 2016 (UTC)
Thanks— My view is that constraining this to just user/user talk wouldn't be nearly as helpful, because the power comes from being able to do it at the same time I do something else in article talk space. "Hey editors, I think this should be renamed. What do you think? {{remind}}". If I have to maintain a separate list somewhere that I have to create & bookmark & switch to, then the value prop becomes much lower vs. any number of third party reminder tools. In any case, I'll go add support to the existing proposal. Regards, Orange Suede Sofa (talk) 20:56, 1 February 2016 (UTC)
Something similar has been proposed for WP:Flow at phab:T88781. For example, this would let admins follow up on messages that a promotional username needs to be changed, to make sure that it actually happened (or that the account was abandoned). Whatamidoing (WMF) (talk) 20:44, 1 February 2016 (UTC)

21:02, 1 February 2016 (UTC)

Lagging

Is there a way to stop underlining in the editing frame (especially if something is pasted into the code)? To clarify: if I paste anything in the edit box, a lot of text gets underlined (red wavy lines similar to this), and after that everything becomes laggy and slow till I insert any of the toolbar tool (for example, ''Italic text'' or <sup>Superscript text</sup> which clears red lines and makes everything back to normal). Does anyone know how to fix this? --Obsuser (talk) 02:03, 1 February 2016 (UTC)

@Obsuser: You could try disabling the toolbar in preferences, and see if that makes any difference. The problem might or might not be related to the toolbar. When you say that clicking something on the toolbar clears the issue, that makes me think that it may be worth that basic test (or that might solve the issue for you, if you don't need the toolbar). I never use it myself, so it's disabled, and I consequently don't know if there's a general issue of the type you describe. It may also help you get the best answers if you tell us which browser (and version) you are using, as it could easily be something specific to a particular browser. Murph9000 (talk) 02:25, 1 February 2016 (UTC)
I woudln’t actually like to disable toolbar as I have many useful gadgets there. I use Chrome updated to latest version 48.0.2564.97 m (with other everything updated too, several days ago; problem is not recent at all). I guess it is not something checking spelling (as in MS Word) as it would be ridiculous to have such tool iniside editing frame because there are urls, non-existing-only-to-Wikipedians-known words etc.
Even if problem is related to toolbar, it is not solution to disable it but to fix lagging. Here is the picture so you can see what I am talking about. NOTE: This lagging does not happen here, when we are talking about 10 or 20 or 100 rows of code, but when I edit longer articles of about 500, 1000, 10000 or more rows everything gets slow in a way I described above.
I guess it fills out some memory caps or something with so many of these underlines. Don’t know how to edit as I cannot paste anything; it sometimes crashes all Wikipedia tabs, although data is not lost as I duplicate them and then reload.
PS Note that even If I disable toolbar, I cannot disable these insertion tools below editing frame (I presume), and if I insert any symbol (for example §) from below — it simply waits for a few seconds and clears out red underlines and everything gets unlagged till next paste. Actually, not everything: text is still being underlined but only in a *paragraph I do edit (*I guess it’s in a such or similar way, maybe not paragraph, I don’t know...), not whole code for sure as lag would be present if that happens. --Obsuser (talk) 03:12, 1 February 2016 (UTC)
I disabled Use this language for spell checking in Manage languages in Chrome settings, and it should be resolved after I restart browser (cannot now because many tabs are opened). Will inform if you if this did not solve the problem, but I think it did. Thank you, my bad as it was that simple (red wavy → spelling check → just turn it off). --Obsuser (talk) 03:26, 1 February 2016 (UTC)
@Obsuser: Yeah, if you like the toolbar, fair enough. It's just good to rule it out as being the cause, since you mentioned it. I think from what you describe, that the problem is not the toolbar, and probably any small insertion is clearing the issue. It could well be a browser issue, as I'm pretty sure that those underlines are the browser's spell checker. Disabling the spell checker, to see if that changes the behaviour, is certainly a good choice for your next diagnostic. Now that we know the browser, hopefully someone who does use it for heavy editing may be able to add something. Murph9000 (talk) 03:34, 1 February 2016 (UTC)
I still didn’t restarted browser to check, but just opened new non-incognito window where it was OK. Will inform... --Obsuser (talk) 04:25, 1 February 2016 (UTC)

  Fixed. It’s resolved now. --Obsuser (talk) 23:24, 1 February 2016 (UTC)

Google Books citation tool

Per this proposal, links to Google Books should be in https://, but http://www.reftag.appspot.com generates citation having http:// link. A message was left at the creator's talk page (here), but he isn't much active. Can someone else make the necessary changes to the tools code so that it generates urls in https://. Thanks, --Skr15081997 (talk) 07:05, 2 February 2016 (UTC)

File:FiveThirtyEight Logo.svg

 – Redrose64 (talk) 17:15, 2 February 2016 (UTC)

Talk page message notification bug?

I just received a notification and email informing me that a user had left me a talk page notification, but they haven't, the diff is for the last edit to my talk page (not by them), and they haven't even edited for days. They didn't message me on another wiki, there's nothing in the filter log nor do they have any recent deleted contributions, so I'm left confused. Any ideas what this was? Sam Walton (talk) 10:34, 2 February 2016 (UTC)

@Samwalton9: If I'm reading the history right, it looks as if the message reached your talk page two minutes after you posted here. That is indeed very strange. -- John of Reading (talk) 10:59, 2 February 2016 (UTC)
@John of Reading: I got a fresh notification and email for the actual message, which all seems to have worked correctly. Just odd that I got a phantom one beforehand; it maybe has something to do with a failed edit save? Sam Walton (talk) 15:35, 2 February 2016 (UTC)
The team working on notifications is keeping an eye on these reports in T125537. If you see more occurrences of it or have more information that could help reproduce or pinpoint a cause of it, please do share. —TheDJ (talkcontribs) 17:20, 2 February 2016 (UTC)

Where are the lang links to the Username Policy

I could swear there were international links to the username policies on WP:UPOL. Now they're not there any more.—cyberpowerChat:Online 13:19, 2 February 2016 (UTC)

Just fixed it with a WP:PURGE. There's nothing in the Wikidata history to suggest that they should have disappeared for some reason. --Izno (talk) 13:21, 2 February 2016 (UTC)
THANK YOU THANK YOU.!!!—cyberpowerChat:Online 22:17, 2 February 2016 (UTC)

Page history "Thank" link

I have been seeing on page-history pages (NaNank pm) link where the "Thank" link usually is. Did something change ? or did I grab a script that might be interfering, also I thought that whachlist had a "thank" link, which I do see any more either. Mlpearc (open channel) 21:06, 2 February 2016 (UTC)

@Mlpearc: Sounds like a faulty script; it works fine for me. Note that "NaN" is very frequently a sign of broken JavaScript. Jdforrester (WMF) (talk) 21:27, 2 February 2016 (UTC)
@Jdforrester (WMF): Thank you. I'll start de-activating my .js scripts to find the culprit.   Mlpearc (open channel) 21:40, 2 February 2016 (UTC)
I have blanked all my .js and also .css pages and the error still exists. Mlpearc (open channel) 23:26, 2 February 2016 (UTC)
It doesn't seem likely but it could be a gadget. Mlpearc (open channel) 23:35, 2 February 2016 (UTC)
If it says "NaNank pm" instead of "thank" then it sounds like an attempt to interpret "thank" as a time and convert it, producing "NaN" (not a number) from "th", and adding "pm". Do you have any browser extensions or features which mess with times? What is your browser? Can you try another? PrimeHunter (talk) 01:20, 3 February 2016 (UTC)
Ah, good point. I'm using FireFox and I just tried Chrome and same error. Mlpearc (open channel) 01:27, 3 February 2016 (UTC)
Do you see it at other wikis like https://simple.wikipedia.org/w/index.php?title=Example&action=history? PrimeHunter (talk) 02:55, 3 February 2016 (UTC)

At the link you posted it shows, (undo | thank) So it seems local. Mlpearc (open channel) 03:02, 3 February 2016 (UTC)

Yes, it sounds like something in your acount. What is your skin? Do you see it in MonoBook or Modern? Have you tried your alternate accounts? PrimeHunter (talk) 03:32, 3 February 2016 (UTC)
I will definitely try all your suggestions, I appreciate all your help, I will try all you steps tomorrow and report back here. Again thanx for all your help.   Mlpearc (open channel) 03:43, 3 February 2016 (UTC)
Working fine with this account. Must be a gadget :P Mlpearc Phone (open channel) 04:08, 3 February 2016 (UTC)

HotCat & Toolbar Clock have disappeared

Both HotCat and the Toolbar Clock displayed & functioned normally thru January 12. After a 2-1/2-week hiatus, I discovered that they had disappeared when I resumed editing on Jan.30. (I actually made one edit on Jan. 26, but I didn't happen to notice anything awry.) . In addition, my pages no longer display an edit link for the lead section, and I am no longer warned when I leave an edit page with unsaved changes. I thought perhaps the settings had been lost in my preferences, but everything was properply "checked". I've even tried multiple times to get things back to normal by un-checking & saving; then re-checking & saving again -- all to no avail. I finally checked to be sure that Javascript was still properly enabled in my browswer, just in case that was involved in some way. Lastly, I just looked at the HotCat talk page, but nobody has reported losing it entirely. Sure would appreciate some help! Cgingold (talk) 19:34, 2 February 2016 (UTC)

@Cgingold: Are you using the Internet Explorer 8 browser? That browser no longer gets JavaScript support, and the changes were implemented around the timing you mentioned. Other than that I'm not sure what else it could be. Jdforrester (WMF) (talk) 19:39, 2 February 2016 (UTC)
Dammit. Not that again. (Thanks for the quick reply, JD.) Yup, that's the problem. I largely gave up editing last year when the switch to HTTPS was implemented. I can live without the clock, but HotCat is essential, since most of my edits involve adding, changing, or removing categories. This really screws everything. Why don't the bigwigs running the show at WMF give a damn about folks like me who are stuck with IE8?? AARRGGHHHHHHH!! Cgingold (talk) 19:54, 2 February 2016 (UTC)
@Cgingold: This was announced in November. Note that Microsoft stopped supporting Internet Explorer 10 last month, let alone 8. You should very seriously consider stopping using it, given the security issues. Jdforrester (WMF) (talk) 19:57, 2 February 2016 (UTC)
Crap. Don't know what I'm gonna do at this point. Just don't know.... Thx anyway for the info. Cgingold (talk) 20:16, 2 February 2016 (UTC)
From the Wikipedia point of view, the question is as follows: Is the expense (in download time and size, script development time, etc) worth it for the number of viewers/editors who actually use the browser? As time goes on, the former tends to grow while the latter tends to get smaller. עוד מישהו Od Mishehu 10:43, 3 February 2016 (UTC)

Incorrect page rendition of section titles and edit links at Cricket pitch

 

Section titles in Cricket pitch are appearing as follows:

Preparation and maintenance of the playing areaEdit with some abnormal URL behind the "Edit"

instead of looking like:

Preparation and maintenance of the playing area [edit source | edit ]

So cannot edit by section. Whole page at top edit works fine.

All other pages I have looked at look good.

Any ideas ?

Aoziwe (talk) 12:43, 31 January 2016 (UTC)

The edit links were apparently made partially with a system used for the mobile version. A purge fixed it. PrimeHunter (talk) 13:11, 31 January 2016 (UTC)
For the record, Cricket pitch#Preparation and maintenance of the playing area rendered as:
Preparation and maintenance of the playing areaEdit


The link went to https://en.wikipedia.org/wiki/Cricket_pitch#/editor/5. Compare to https://en.m.wikipedia.org/wiki/Cricket_pitch#Preparation_and_maintenance_of_the_playing_area where the edit icon goes to https://en.m.wikipedia.org/wiki/Cricket_pitch#/editor/5. The only difference in the url is .m. It was the same for all section edit links. PrimeHunter (talk) 13:24, 31 January 2016 (UTC)
I've seen similar on other pages in the last couple of weeks, it was always associated with the absence of an [edit] link for the lead section of that page; a WP:PURGE was all that was needed for all of them. --Redrose64 (talk) 15:24, 31 January 2016 (UTC)

Thanks all. Aoziwe (talk) 08:49, 1 February 2016 (UTC)

That's a known issue. Max Semenik (talk) 00:32, 3 February 2016 (UTC)
Just seen it again, when I visited Template:São Tomé and Príncipe Championships, see image top of section. The HTML for the section header containing the incorrect edit link is:
<h2><span class="mw-headline" id="Usage">Usage</span><span><a href="#/editor/T-1" title="Edit section: " data-section="T-1" class="mw-ui-icon mw-ui-icon-element mw-ui-icon-edit-enabled edit-page icon-32px">Edit</a></span></h2>
Page history at that moment had only two edits, and that duff "Edit" link was inside the documentation area (which was also visibly broken: the upper green box was finishing too early), so the only edit which could have done it was this one. NeedAGoodUsername, do you edit using a mobile device? --Redrose64 (talk) 10:51, 3 February 2016 (UTC)

Cite vs Citation templates,, harvnb template, etc.

This edit in good faith raised some questions for me:

  • the substitution of date for year breaks Harvard-style citations, in my experience.
  • I was under the impression that the Citation template was easier to use (fewer red messages).
  • author-link and now editor-link work for non-harvard-style refs, but the author-link to a harvard-style brief reference would work better in the full bibliographic entry, it seems to me. This appears to be a difficulty when mixing citation styles.
  • It appears that cite web causes us to want to change existing refs from date to year.

--Ancheta Wis   (talk | contribs) 05:16, 31 January 2016 (UTC)

Module:Citation/CS1 extracts the year portion from the value supplied to either |date= or |year=. When both of these parameters are used in a cs1|2 template, the module uses the value from |year= in the CITEREF anchor. Generally, use of both |date= and |year= should be confined to those times when the article date style is year-initial-numeric and |ref=harv and year disambiguation is required. Can you provide an example of [broken] Harvard-style citations?
{{citation}} (cs2) uses the same rendering engine and is subject to the same rules as the other (cs1) citation templates. It is distinguished from cs1 templates by the use of commas as element separators and lower case static text.
|author-link= and |editor-link= have nothing to do with reference styles. These parameters simply link the author/editor name in the rendered citation to that person's Wikipedia article.
{{cite web}} does not care if you use |date= or |year= except that it expects the values assigned to those parameters to comply with MOS:DATEFORMAT et seq.
This conversation should continue here until completed. In future, questions and issues related to the cs1|2 citation templates should be taken to Help talk:Citation Style 1.
Trappist the monk (talk) 14:04, 31 January 2016 (UTC)
This citation, Smith, A. Mark, ed. and trans. (2001), "Alhacen's Theory of Visual Perception: A Critical Edition, with English Translation and Commentary, of the First Three Books of Alhacen's De Aspectibus, the Medieval Latin Version of Ibn al-Haytham's Kitāb al-Manāẓir, 2 vols.", Transactions of the American Philosophical Society, 91 (4–5), Philadelphia: American Philosophical Society, ISBN 0-87169-914-1, OCLC 47168716{{citation}}: CS1 maint: multiple names: authors list (link) Books I-III (2001 — 91(4)) Vol 1 Commentary and Latin text via JSTOR; 91(5) Vol 2 English translation, Book I:TOCpp.339-341, Book II:TOCpp.415-6, Book III:TOCpp.559-560, Notes 681ff, Bibl. via JSTOR from Book of Optics
note 44 in Alhazen, is the base for a Harvard-style ref (Smith 2001). It works in Book of Optics. But this 2001 ref doesn't work so well in the Alhazen article. Another editor used Smith, ed. instead of last=Smith so I suppose that's why the harv Smith 2001 link doesn't work in the Alhazen article. I could harmonize the citations. --Ancheta Wis   (talk | contribs) 11:53, 1 February 2016 (UTC)
Per the above, I tried patching in last=Smith in the citations. I got "Warning: Alhazen is calling Template:Citation with more than one value for the "date" parameter. Only the last value provided will be used." ??
--Ancheta Wis   (talk | contribs) 13:16, 1 February 2016 (UTC)
One of the uses of the template is either using |date= twice, or its synonym e.g. both |date= and |year=. Removal of one will fix that problem. --Izno (talk) 13:21, 1 February 2016 (UTC)
That error in Alhazen is not a cs1|2 error and does not occur for aliases. The parameter names must be the same. In this case the offending {{citation}} template is Van Sertima which includes |date=1992 and |date=2004.
As an aside, that template also has |oclc=123168739 25416243 29837541 44748929 50901951 56571741 57559516 – seven OCLC identifiers where there should only be one. I will fix the module so that multiple oclc identifiers in |oclc= emits an error message.
As a further aside, the article suffers from inconsistency of citation style: it uses both cs1 and cs2; it uses both {{harv}} and {{harvnb}}and either of those with and without the in-source locator template {{rp}}. Much work is needed.
Trappist the monk (talk) 16:15, 1 February 2016 (UTC)
@Trappist the monk, Thank you for pointing out these problems. A. Mark Smith is Curators' Professor of History at the Univ. of Missouri, Columbia. Smith has labored long on Medieval Optics. My interest is in getting these citations right. I will patch the Van Sertima and the Smith refs for the problems I know about. --Ancheta Wis   (talk | contribs) 18:42, 1 February 2016 (UTC)
If {{harv}}, {{harvnb}} or similar templates such as {{sfn}} are in use, it should never be necessary to use {{rp}} as well, since that defeats the purpose of Shortened footnotes. --Redrose64 (talk) 21:09, 1 February 2016 (UTC)
It seems to me the benefit would be a smaller number of shortened unique footnotes, since author, year, and page number could be compressed down to a named ref to a {{harvnb}}, directly pointing to an author|year for each respective inline ref. + Each repeated author, year would be lettered by a unique named ref, indirectly pointing via the {{harvnb}} link to the full {{Citation}}. + A given page would have an {{rp}} for the unique page number.
The Smith refs could then compress his life's work down to a year and page number. I will propose this, with examples, on the Alhazen talk page, before I were to embark on the update. Consider that Book of Optics is 1000 years old, with multiple translations, languages, manuscripts, books, and volumes, written before there was a Gutenberg or Risner. Not to mention the centuries of scientists who cite Alhazen. --Ancheta Wis   (talk | contribs) 22:36, 1 February 2016 (UTC)
I am not entirely clear on what you are trying to show, but I concur with Redrose64 that {{rp}} is not needed. Note that each instance of a {harv} template (all pointing to a given citation) can have its own page number. Note also that while {harv} templates can link to a full citation anywhere in the text, including another footnote, it is much more convenient to collect all of the full citations in single section (typically named "References" or "Sources"). ~ J. Johnson (JJ) (talk) 23:04, 1 February 2016 (UTC)
Before seeing the latest, I proposed harvnb+rp on Talk:Alhazen#proposal.I am happy to go with consensus. --Ancheta Wis   (talk | contribs) 23:27, 1 February 2016 (UTC)
And looking there it appears you wanted to use a named ref to re-use a short cite (the Harv template), then use {rp} to add each cite's page number externally to the footnote (<ref>). Better to use separate footnotes, as I have explained there, with the page numbers within the footnote. ~ J. Johnson (JJ) (talk) 00:34, 2 February 2016 (UTC)
I had to put in 4 separate fixes to try to make the links to the base citation work. In these cases (notes 45, 68, 78, 116) the shorter ref to {{harv}} was the way to make the links to the base citation work. The rest was done by hand: volume, Latin para., page numbers to the Latin, to a footnote in translation, etc. --Ancheta Wis   (talk | contribs) 10:17, 2 February 2016 (UTC)
I'll help you with the details. We can talk about it there. ~ J. Johnson (JJ) (talk) 20:06, 3 February 2016 (UTC)

OTRS problems

I know this board isn't for OTRS issues, but my email inbox has been filled with over 120 "ticket-creation notifications" in about an hour and at least some of them look awfully spammy. When I went to my OTRS login screen, it looked different (it said "OTRS 5") and I didn't log in because I didn't want a hacker to get my username and password. Hope it's just the devs doing their thing again (annoying as it is), and not a security breach of some kind. Are other OTRS agents having the same problem? I'm going to delete the emails unread, but it's a timesink I don't need. All the best, Miniapolis 19:33, 3 February 2016 (UTC)

@Miniapolis: OTRS was updated to OTRS 5 today so what you're seeing at the log in screen is correct. As for 120 emails I suspect in your OTRS settings you have "Notify me when new ticket created" and as a queue has been repopulated you've been sent an email for each ticket in the queue. Nthep (talk) 20:07, 3 February 2016 (UTC)
Many thanks for the quick reply, and I'm glad things are relatively okay :-). I've been an agent for a few months, and never got ticket-creation notifications like this. Judging by the subject lines, it's another junk queue; they're all spam. I'll check my settings. All the best, Miniapolis 20:28, 3 February 2016 (UTC)

Convert template outputting oddly formatted number

  Resolved

In the article Frank Pick, in the "Frank Pick#London Passenger Transport Board – bringing it all together" section, the convert template is used to give an idea of the current value of Pick's salary. It gives an output of "£630 thousand" - now this is a bizarre way of writing an amount of money, "£630,000" would be the normal way of writing this. How do I change it to output correctly? (I feel sure I've asked this before, but can't remember when, or what the answer was). Thanks. DuncanHill (talk) 15:36, 3 February 2016 (UTC)

@DuncanHill: {{convert}} does not have the means to convert between currencies. It is for physical measures, such as length, mass, area, velocity. --Redrose64 (talk) 20:58, 3 February 2016 (UTC)
It's not {{convert}}. The issue appears to be a design decision in {{Formatprice}} which does this:
  • £{{Formatprice|630000}} → £630,000
Some opinions are at Template talk:Format price. I suggest trying to contact people who have edited that template and ask why it does that. Johnuniq (talk) 21:10, 3 February 2016 (UTC)
Sorry, my mistake - it uses the template {{Formatprice}} and {{Inflation}} DuncanHill (talk) 21:12, 3 February 2016 (UTC)
Until there is agreement to fix the template, it looks like the workaround is to use this in the article:
  • £{{Formatnum:{{Inflation|UK|10000|1933|r=-4}}}} → £760,000
Johnuniq (talk) 21:20, 3 February 2016 (UTC)
  • Thanks, I just got rid of the "formatprice" template, and it looks much better. DuncanHill (talk) 21:23, 3 February 2016 (UTC)

Geolocate not working

When I click on "geolocate" to find an IPs location I get the following message "No scripted access. See http://cdn.whatismyipaddress.com/api.html or contact support@whatismyipaddress.com" The "alternate" link to infosniper is still working. If this has already been reported elsewhere feel free to add a link to that thread here. Thanks for your time. MarnetteD|Talk 17:32, 3 February 2016 (UTC)

Worked for me. Mlpearc (open channel) 17:37, 3 February 2016 (UTC)
I forgot to mention that this has been going on for three or four days so it doesn't seem to be a temporary blip. MarnetteD|Talk 17:40, 3 February 2016 (UTC)
The link in my post works but clicking on "geolocate" does not. MarnetteD|Talk 17:40, 3 February 2016 (UTC)
I used the dropdown tab "user" on an IP's talk page. Mlpearc (open channel) 17:43, 3 February 2016 (UTC)
(edit conflict) I tried refreshing my cache and I still got the same message when clicking on your link Mlpearc. Did you try using "geolocate" on a random IP like this? Oops due to the ec you might have already answered this. MarnetteD|Talk 17:46, 3 February 2016 (UTC)
@MarnetteD: Maybe it's your browser ? are you able to try a different one ? Mlpearc (open channel) 17:49, 3 February 2016 (UTC)
Thanks for the suggestion Mlpearc. I tried Google Chrome and it worked like it is supposed to. I am not very computer literate so I have no idea what the problem is with my Firefox. I figured I'd ask in case anyone else was experiencing the same thing. Thanks again for your help. MarnetteD|Talk 20:15, 3 February 2016 (UTC)
@MarnetteD: You're very welcome. I'm also not very good when it comes to browsers. I'm using FireFox most of the time and when there's trouble Chrome usually gets me through. Cheers, Mlpearc (open channel) 22:43, 3 February 2016 (UTC)

Phabricator is down

>>> UNRECOVERABLE FATAL ERROR <<<

Call to undefined method AlmanacCreateClusterServicesCapability::getPhobjectClassConstant()

/srv/phab/phabricator/src/applications/policy/capability/PhabricatorPolicyCapability.php:18


┻━┻ ︵ ¯\_(ツ)_/¯ ︵ ┻━┻

I'd better report that in Phabr... Oh. LX (talk, contribs) 06:50, 4 February 2016 (UTC)

Yep, I reported it to tech operations earlier. Also, FYI, when something crashes on the WMF side, you can also let them know on #wikimedia-tech connect. Titoxd(?!?) 07:31, 4 February 2016 (UTC)
It was down for about about one and a half hour, but is now back. —TheDJ (talkcontribs) 08:40, 4 February 2016 (UTC)

Yes Men article displays incorrectly if Javascript disabled in browser

This page: The_Yes_Men

I've left a note on the article's talk page with my browser version and a fuller description. I'm using Trisquel 7. Clark42 (talk) 16:44, 4 February 2016 (UTC)

The template {{Anti-consumerism}} had been broken by transcluding another which is not meant for the sidebar. Removed it and purged the page and it seems OK now.--JohnBlackburnewordsdeeds 17:09, 4 February 2016 (UTC)

Is information Previewed stored anywhere (Concern about Copyright)

If the text AB AC AD AE is under copyright and that text is previewed in an article and then replaced with something else, is "AB AC AD AE" ever stored on the Wikipedia servers and would it cause a problem with infromation under copyright being stored on Wikipedia?Naraht (talk) 14:21, 4 February 2016 (UTC)

It is stored, but all such storage is of transient nature (caching). The safe harbor protections of copyright law in the USA specifically protect us from any problems with this, so it is safe. —TheDJ (talkcontribs) 14:50, 4 February 2016 (UTC)
Thank You.Naraht (talk) 17:50, 4 February 2016 (UTC)

Paragraph won't sit side-by-side on a diff

Does anyone know why the two versions of the paragraph beginning "Dopamine agonists produce significant side effects ..." won't sit next to each other in this diff? --Anthonyhcole (talk · contribs · email) 11:21, 3 February 2016 (UTC)

Can't say exactly, but the diff generator sometimes gets confused and does funny things. User:Cacycle/wikEdDiff offers an alternative view that you can use by clicking an icon on the diff. It often handles such situations better, and it in fact does in this case. ―Mandruss  11:35, 3 February 2016 (UTC)
OK. Thanks, Mandruss. Mmm. That Education project has some cool stuff. — Preceding unsigned comment added by Anthonyhcole (talkcontribs) 12:05, 3 February 2016 (UTC)
An extreme example is this bot edit. The only change to the paragraph beginning "A short distance to south the ..." was that a trailing space after the terminal full stop was removed. Similarly, in this user edit, a {{cn}} was added at the end of both paragraphs - there were no other changes to either paragraph, yet the second one fails to sync. --Redrose64 (talk) 21:48, 4 February 2016 (UTC)

Watchlist add/remove from category

  Resolved

Uggh, where did this new "feature" come from, any one see where to turn it off? I really don't want this in my watchlist, e.g.

    (diff | hist) . . Category:Candidates for speedy deletion‎; 00:01:45 . . (0)‎ . . ‎Legacypac (talk | contribs | block)‎ (Reddishgrays added to category)
    (diff | hist) . . Category:Candidates for speedy deletion‎; 00:01:19 . . (0)‎ . . ‎Legacypac (talk | contribs | block)‎ (Reddishgreys added to category)
    (diff | hist) . . Category:Candidates for speedy deletion‎; 00:01:08 . . (0)‎ . . ‎Legacypac (talk | contribs | block)‎ (Reddishgrey added to category)
    (diff | hist) . . Category:Candidates for speedy deletion‎; 00:00:55 . . (0)‎ . . ‎Legacypac (talk | contribs | block)‎ (Reddishgray added to category)
    (diff | hist) . . Category:Candidates for speedy deletion‎; 00:00:44 . . (0)‎ . . ‎Legacypac (talk | contribs | block)‎ (Reddishviolets added to category)
    (diff | hist) . . Category:Candidates for speedy deletion‎; 00:00:34 . . (0)‎ . . ‎Jjjjjjdddddd (talk | contribs | block)‎ (White van speaker removed from category)
    (diff | hist) . . Category:Candidates for speedy deletion‎; 00:00:31 . . (0)‎ . . ‎Legacypac (talk | contribs | block)‎ (Violetishreds added to category)
xaosflux Talk 00:12, 29 January 2016 (UTC)
Why are the words "diff" and "hist" not links? I want to see the edit that caused the change, and the recent edits either side of it. --Redrose64 (talk) 00:17, 29 January 2016 (UTC)
Looks like if you watch a category you get a watchlist update everytime something is added or removed from the category - while this COULD be a useful feature - it shouldn't be on without a way to screen it. — xaosflux Talk 00:19, 29 January 2016 (UTC)
Seems to have spawned from phab:T9148, guess I'll head to phab to chase it with an enhancement request :( — xaosflux Talk 00:23, 29 January 2016 (UTC)
Change request entered as phab:T125171. — xaosflux Talk 00:27, 29 January 2016 (UTC)
Thanks for adding the change request. Because I watch categories, that's all I see now! Funandtrvl (talk) 00:38, 29 January 2016 (UTC)
(edit conflict) That is not a bad feature, but it should be separately watchable. I often want to watch the changes to a procedural category, but not additions/removals (dozens of pages per day). Filtering watchlist wouldn't really help, because I want to watch other categories for their addition/removals. My watchlist is now 80% added/removed pages because I have many internal workflows watchlisted. I realize this means non-binary value for watching, which would likely be a much bigger change to MW... —  HELLKNOWZ  ▎TALK 00:43, 29 January 2016 (UTC)
Having different sorts of watch lists has been an open request for a long time! — xaosflux Talk 00:49, 29 January 2016 (UTC)
Couple of things: If you turn on the enhanced watchlist ("Expand watchlist to show all changes, not just the most recent" in preferences), all the entries relating to a single category will all collapse into a single block. And also, as far as I can tell, the checkbox that xaosflux is asking for seems to already by implemented in MediaWiki but isn't being shown on this wiki for some reason. I'll look into it, unless someone else beats me to it :) — This, that and the other (talk) 01:02, 29 January 2016 (UTC)
@This, that and the other: "expand watchlist to..." did not fix this for me, it did do what it said it was going to do, but did not have any sort of collapsing choice. — xaosflux Talk 01:11, 29 January 2016 (UTC)
Sorry, wrong option; it's on the Recent Changes tab: "Group changes by page in recent changes and watchlist". — This, that and the other (talk) 01:16, 29 January 2016 (UTC)
See phab:T125171 for an explanation of what is going on here. — This, that and the other (talk) 01:16, 29 January 2016 (UTC)
Thanks, deployment blunder - guess we can easily validate it isn't happening anymore. — xaosflux Talk 01:27, 29 January 2016 (UTC)
Looks all fixed now. — xaosflux Talk 03:49, 29 January 2016 (UTC)
Yep, the config change enabling the feature has been rolled back. This is what caused everyone to see the category membership changes (as there was no longer any way to hide them). The reason for the rollback can be tracked at https://phabricator.wikimedia.org/T125147 ·addshore· talk to me! 11:21, 29 January 2016 (UTC)
Hi all! The feature has now been deployed again! ·addshore· talk to me! 00:27, 5 February 2016 (UTC)

Script for rating articles?

We've had a query on wikimedia medicine-l from a medical editor, very active in the Arabic Wikipedia community. I'll copy it here and point the author to this thread.

I'm trying to find a script that makes it easy to insert assessment tags into (medical) articles. The one I have found was Kephir's rater, but it's more general-purpose and multi-functional in a way that makes it difficult to localize into the Arabic Wikipedia.

Any other suggested script?

Anthonyhcole (talk · contribs · email) 01:36, 4 February 2016 (UTC)

@Anthonyhcole: Did you mean the Article Feedback Tool? That tool was removed on March 3, 2014. GeoffreyT2000 (talk) 05:02, 4 February 2016 (UTC)
No, he's talking about automatic help in adding quality assessments (A Class, Good Article, Stub, etc.) to articles. These tasks are mainly performed by members of WikiProjects, who tag talk pages of articles. WhatamIdoing might know. --Anthonyhcole (talk · contribs · email) 05:12, 4 February 2016 (UTC)
seems like a very good idea ( to find a script that makes it easy to insert assessment tags[51])--Ozzie10aaaa (talk) 10:30, 4 February 2016 (UTC)
Thanks Anthonyhcole for sharing this! Back in Wikimania 2015, someone in the hackathon presented a similar script. I did not followed up with him during the conference, and now I forgot where to find it!--OsamaK (talk) 10:38, 4 February 2016 (UTC)
@OsamaK: If giving article the same class, as it's at enwiki, is fine, then I most probably can help. --Edgars2007 (talk/contribs) 10:49, 4 February 2016 (UTC)
This would be a substantial help! Providing such a script will result in the classification and subsequent editing of over 4,000 medical articles on the Arabic Wikipedia! The template used to classify articles will be identical to the one used on the English Wikipedia (i.e. {{WikiProject Medicine|class=|importance=}}). A dialog would ask users to choose the quality class (i.e. FA/GA/B/C/Start/Stub) and the importance class (Top/High/Mid/Low). Either fields can be left empty (but not both). After the choice, the template will be inserted into the talk page. --OsamaK (talk) 11:15, 4 February 2016 (UTC)
Sounds like an interesting tool that could be useful in a number of languages. User:Ladsgroup is collecting medicine related tools. Doc James (talk · contribs · email) 14:27, 4 February 2016 (UTC)
@OsamaK: I can't write a script, but I can help in another way. I can give you a list of articles, that are both here and at arwiki and are part of WikiProject Medicine here. And I can say you, what importance the article has here. Then a bot could place that info to arwiki talk pages. You're interested? --Edgars2007 (talk/contribs) 15:44, 4 February 2016 (UTC)
I have written such a bot. We are using it to track popular Arabic medical articles (i.e. we define an Arabic 'medical article' as an article that has an English version that is part of the English WikiProject Medicine). As for the importance, it cannot be imported from the English Wikipedia as it is, because the current classification used in the English Wikipedia is not really based on specific criteria. For that reason, a script is needed that would enable eager contributors to classify the articles without worrying about the wikitext.--OsamaK (talk) 18:42, 4 February 2016 (UTC)
I use Kephir's rater. User:Kaldari wrote something similar, and there are bots listed in Category:Autoassessment bots that can assess articles, usually by copying the rating that another WikiProject already gave it. You might find other tools or ideas at Wikipedia:WikiProject Council/Guide/Technical notes.
User:EpochFail and User:Nettrom also have a script for m:Research:Screening WikiProject Medicine articles for quality that can estimate the class (e.g., stub) that is IMO accurate enough to automatically tag Stub- and usually Start-class articles. (When you get into the higher levels and/or lower probability, it tends to over-rate articles that contain long lists or large numbers of references.) If you have tech skills, then I encourage you to look into integrating this tool, as it would make an excellent basis for a first effort. Your volunteers could then review the ratings at their leisure and/or as articles get expanded over time. WhatamIdoing (talk) 01:42, 5 February 2016 (UTC)
@Anthonyhcole: My script is much simpler than Kephir's and probably easier to localize: User:Kaldari/assessmentHelper.js. Kaldari (talk) 01:48, 5 February 2016 (UTC)
Re. my work with User:Nettrom to predict article class, see m:ORES. For example, this URL scores the most recent revision of Waffle (oldid=702406275): https://ores.wmflabs.org/scores/enwiki/wp10/702406275 As WhatamIdoing notes, the system is not perfect, but it should be useful. We're actively working toward increasing its fitness. Please add m:ORES to your watchlist to get updates. :) --EpochFail (talkcontribs) 02:05, 5 February 2016 (UTC)

{{reflist|2}} produces unwanted spaces when there's only one reference

In this old version of The Dark Forest, there's a {{reflist|2}} but only one reference. This produces an artificial space between any two words in the ref depending on the width of the browser. I realize that this is arguably a misuse of {{reflist|2}} but thought someone would like to know. Orange Suede Sofa (talk) 04:29, 5 February 2016 (UTC)

@Orange Suede Sofa: Yes, we know. It works as expected. The list just gets splitted into two parts, system doesn't care how much references there are. --Edgars2007 (talk/contribs) 07:36, 5 February 2016 (UTC)
@Orange Suede Sofa: See Template:Reflist#Widows and orphans continuing into "Practices". I would recommend against using multiple columns where the refs are all long and there are few of them - personally, I wait until there are at least ten, or at least six if all of them are short, as where Shortened footnotes are in use. --Redrose64 (talk) 09:44, 5 February 2016 (UTC)

Watchlist options - hide page categorisation

There's a new tick box in the watchlist options to hide page categorisations. Unticking it seems to make no difference - when the watchlist is refreshed or returned to from another page, it is once again ticked. What is the box for and why can it not successfully be unticked? DuncanHill (talk) 00:51, 5 February 2016 (UTC)

@DuncanHill: None of those checkboxes are persistent, they're all for single/short-term use. You probably want the box labelled "Hide categorization of pages" at Special:Preferences#mw-prefsection-watchlist. --Yair rand (talk) 01:19, 5 February 2016 (UTC)
(Sorry, I just realized I probably misunderstood your question. Maybe ask on phabricator.) --Yair rand (talk) 01:20, 5 February 2016 (UTC)
Well, unticking it on Prefs makes it stay unticked on the watchlist page, so that's part of it answered. What's phabricator? DuncanHill (talk) 01:37, 5 February 2016 (UTC)
@DuncanHill: See Help:Phabricator and phab:. --Redrose64 (talk) 09:50, 5 February 2016 (UTC)
  • Generally, I think the new feature of seeing page categorizations of watched categories in my watchlist is a good thing, except for some very high-traffic categories, e.g. Category:Living people, where I only want to see edits to the category page and its talk page. Is that possible? -- Michael Bednarek (talk) 02:49, 5 February 2016 (UTC)
So, does this having the box unticked mean that I will see when a page is added to a category which is on my watchlist? DuncanHill (talk) 03:20, 5 February 2016 (UTC)
Yes. That was first on automatically and drove people nuts. It's helpful for me for things like the speedy deletion categories which would I advise all admins to add to their watchlists if they turn it on. -- Ricky81682 (talk) 03:40, 5 February 2016 (UTC)
Ah, excellent, thank you. I remember it being requested several years ago. Does it also tell you when a page is removed from a category? That would be useful to spot people sneakily emptying a category before proposing it for deletion. DuncanHill (talk) 03:55, 5 February 2016 (UTC)
@DuncanHill: Yes, my watchlist has both category additions and removals. -- John of Reading (talk) 07:16, 5 February 2016 (UTC)
@John of Reading: Splendid news, thank you. DuncanHill (talk) 08:25, 5 February 2016 (UTC)
@Michael Bednarek: No, not yet. Per-category configuration is requested at phab:T109759. I won't be using this feature until that is implemented. -- John of Reading (talk) 07:16, 5 February 2016 (UTC)

IMPORTANT - archiving fails with blacklisted links

Header shortened. GermanJoe (talk) IMPORTANT! Bot has been deleting threads. This really needs attention! Vanishing archives are a really bad thing™. See details here. (Thread auto-archived w/o action.) 'Till this is fixed, admins may need to remove all to-be-blacklisted URLs from extant talk pages before blacklisting. Can someone inform the relevant admins? --Elvey(tc) 23:52, 1 February 2016 (UTC)

The example [52] in the linked page is from March 2013 by a bot which hasn't edited since October 2013. Do you have a recent example with an active bot? PrimeHunter (talk) 12:05, 2 February 2016 (UTC)
Just to be clear, this is not a "deletion", but an edit. — xaosflux Talk 12:45, 2 February 2016 (UTC)
I had fixed a similar issue in October 2015, see here for my latest message about the issue (I disabled the links manually and archived the thread afterwards). The failed archiving's diff is here. If technically possible, it would be better to save the archive first, and then analyse the action's return code (assuming a failed save via bot produces a meaningful return code). Unless the save was successful, just deleting the original content without further check seems problematic. GermanJoe (talk) 21:56, 2 February 2016 (UTC)
I have posted to the bot operator at User talk:Σ/Archive/2016/February#Archiving fails with blacklisted links. It's possible the issue was resolved after User talk:Σ/Archive/2015/November#Problem with archive bot. PrimeHunter (talk) 12:01, 5 February 2016 (UTC)

Using apostrophes in titles?

Is there any information about whether using apostrophes in titles will be supported by MediaWiki in the future? It would be pretty convenient for redirection from links like Star Wars films).--Prisencolin (talk) 02:36, 3 February 2016 (UTC)

It's not prevented by the MediaWiki software but by a setting in the English Wikipedia at MediaWiki:Titleblacklist. Admins can create the titles, and admins could remove the general block of such titles, but I don't support doing that. If you see links like Star Wars films then just fix them, e.g. with [[Star Wars films|''Star Wars'' films]] to produce Star Wars films. If a redirect was created then Star Wars films would become a working blue link but with the current software it would still render with apostrophes and not italics. PrimeHunter (talk) 02:49, 3 February 2016 (UTC)
Eh, that's the kind of unnecessary piping I would hope gets fixed in the future.--Prisencolin (talk) 04:30, 5 February 2016 (UTC)
I'm not sure I would like that "fix". If [[''Star Wars'' films]] made a link to Star Wars films then the name of the linked article would not be in the wiki source and could not be found with a Ctrl+f browser search or the search function for the edit box. It would probably also confuse many tools which would need changes to work as intended. PrimeHunter (talk) 12:13, 5 February 2016 (UTC)

Edits not showing up

I just found some refs for a draft I have in userspace, and added them. But...they didn't appear, although I know I clicked save, and I didn't get the little "success, your edit was saved!" bubble till a couple of minutes later. It's sort of annoying. Did anybody else have this, or is it a problem with me alone? White Arabian Filly (Neigh) 02:07, 5 February 2016 (UTC)

I think there was a database lag. I had a notice on my contributions page that it was like 40 seconds off but it seems to have caught up. Probably some big change to a template or something. -- Ricky81682 (talk) 03:41, 5 February 2016 (UTC)
It's still happening this morning. I think it may have something to do with the new look of mobile userpages and the mini reflist at the end of each section. White Arabian Filly (Neigh) 15:29, 5 February 2016 (UTC)

Chrome is wrongly breaking the Chart lines

Hi.

 
Chrome
 
Firefox

I'm using Template:Chart to do family trees, and I noticed that the charts are displayed differently in Chrome and Firefox/IE. Chrome tend to break the rows (lines) in an undesirable places. This example is a screenshot from Chrome (wich is bad), and this is from Firefox (the proper way of displaying). I don’t know if it is only a Template problem or only a Chrome’s one, but I wish to fix it. Or can someone explain to me if I’m doing something wrong? I’m using <br /> to break the rows. Sorry if this seems trivial, but these broken rows are making the text more difficult to read.--Daduxing (talk) 15:42, 5 February 2016 (UTC)

Would it be technically possible for developers to make it possible to disable watchlists?

I'm hoping someone might know if this would be possible to develop. If it wouldn't be, there's no point in my going further. If it might be, I'll propose it as a way of dealing with sockpuppets. Banned/indefinitely blocked users still use their watchlists and then use socks to edit. Yesterday an obvious IP sock arrived at a talk page minutes after I edited the article. Of course blocking notification would help, but it wouldn't work as well as disabling the watchlist. Doug Weller talk 08:51, 30 January 2016 (UTC)

All the sockmaster would have to do is have the raw watchlist, and then copy and paste that into the sock. It'd be a matter of seconds. You can see it yourself if you go to your watchlist and hit "Edit raw watchlist", and one good use for that is as an export/import function for a watchlist. I'm sure it would be feasible on a technical level for admins to have the ability to blow out and lock someone's watchlist, but I don't think it'd be much benefit. Seraphimblade Talk to me 09:04, 30 January 2016 (UTC)
Yeah, if the sock is smart enough, knowledgeable enough, and creative enough to figure that out. It wouldn't have occurred to me, and I'm a fucking genius. ;) It wouldn't help with some socks, but nothing's perfect. ―Mandruss  09:12, 30 January 2016 (UTC)
There might be an idea for a new feature for CheckUser in there. Comparison of watchlists. Probably shouldn't be a general admin level feature, for privacy reasons. Murph9000 (talk) 09:33, 30 January 2016 (UTC)
Comparison of watchlists: yes, that's a really good idea. -- The Anome (talk) 09:36, 30 January 2016 (UTC)
Now that, I could see being tremendously useful. I don't think we'd want to allow even checkusers to see individual pages watched (that's getting into the territory of seeing what someone has read, rather than just when and from where they've logged in), but maybe return a flag if two users have identical watchlists (and perhaps a count of how many pages they're watching; someone watching only a page or two could generate a lot of false positive duplicates, but someone watching a couple hundred is pretty unlikely to). Seraphimblade Talk to me 16:24, 30 January 2016 (UTC)
Wholly agree that viewing the contents of watchlists would be too invasive. I may support a checkuser access that accepts inputs two usernames, and outputs "There are x identical entries". — xaosflux Talk 17:00, 30 January 2016 (UTC)
Nice idea. Don't display the contents, but display the number of items in each list and the number of exact matches. For convenience I'd also add a match% figure. The math would be sqrt((IdentialEntries/ListSize1)*(IdentialEntries/ListSize2))*100. That is the geometric mean of percentage-match on each of the two lists.
I just realized that could be used to probe someone's watchlist: Use a dummy account with with test-entries in the watchlist, and use match% to test for specific entries. On the other hand Checkusers are already heavily trusted not to engage in that sort of shenanigans, and I believe Checkuser tool use is all logged. That sort of tool-abuse would be very visible in the logs. IMO that theoretical attack is probably a non-issue. Alsee (talk) 22:31, 30 January 2016 (UTC)
Re Seraphimblade's point, it should be reasonably easy for there to be a button available to admins which disables the watchlist of a blocked user, and to make it such that a disabled watchlist cannot be viewed by the user via the raw watchlist. A dedicated LTA would perhaps learn BEANS ways to workaround that by planning ahead, but the vast majority of problematic users would lose access to their watchlist and would think that was the way things work. The action would need to be logged in the same way done when removing talk-page access. Johnuniq (talk) 01:32, 31 January 2016 (UTC)
I've lodged two Phabricator tickets for these features that will be ignored for the next 10 years -- both are technically straight forward. That said, I've heard positive signs from the WMF that they want to look at technical means to tackle harassment and sockpuppetry and there's a chance we'll get these sooner rather than later. MER-C 04:53, 31 January 2016 (UTC)
  • A sockpuppet could easily circumvent this:
    1. Copy the contents of Special:EditWatchlist/raw from the blocked account, and insert the same text on that page for a new account.
    2. Clear the list for the blocked account (so that the overlap is 0 %).
    3. Don't use the new account for any editing. The new account will be a single-purpose account only used for maintaining the watchlist. Use WP:RSS#Watchlist feed with token to follow activity on the watchlist so that you don't need to log in unless you need to modify the watchlist.
However, I suppose that this could discover some sockpuppets belonging to users who are not aware of this. --Stefan2 (talk) 14:49, 31 January 2016 (UTC)
Yeah, the idea isn't that is will always confirm or exclude a sock, only that it may add evidence towards confirming in some cases. The useful result would be a positive comparison of larger watchlists, which should really be quite unlikely for genuinely unrelated accounts. Murph9000 (talk) 23:08, 31 January 2016 (UTC)
Better get a good solid baseline rate of watchlist overlap among active editors before using any such comparisons as evidence of sockpuppetry, and maybe weight the result by total number of watchers of a given page. For some reason nobody ever seems to worry about this base-rate problem when talking about page overlaps in sock cases, but it would be much worse if the tool only provided a number rather than a list of distinctive overlapping pages. Opabinia regalis (talk) 21:08, 1 February 2016 (UTC)
And to consider the size of the watchlist. A "100% matching" watchlist that contains only two or three pages (especially popular pages) is not very good evidence of anything. WhatamIdoing (talk) 19:20, 5 February 2016 (UTC)

Nowrap for references

Why don’t <ref></ref> tags get "locked" to the previous word? It is looking bad when⁠[1] "detaches" from the word and shows up in the new line at the very beginning... Is there a way to fix this problem for all Wikipedias or we have to cheat and enclose it manually with {{nowrap| and }} what makes way much bigger source codes and everything nastier and harder?

Also, this should be done for words connected by a hyphen (e.g. "two-three") which should not be let to be able to get separated as in such cases additional hyphen is needed to denote word was not syllablized but actually does contain - in itself. For example:

These two-
-three guys...

is correct. On the other hand, it would be wrong:

These two-
three guys...

I’m asking it here but it should be solved if it will be for all Wikipedias, not only one language project. I presumed this adding of hyphen is needed in English because in Serbian it is to make sure word has not been separated into syllables but actually does contain hyphen and is semi-compound word, such as semi-compound is itself.--Obsuser (talk) 04:25, 1 February 2016 (UTC)

This is a very complex issue. My best and least complex answer would be.. Sorry, this is the web, not print typesetting.
Both issues are difficult to workaround without introducing significant problems of it's own. Also the latter is not a common rule in modern English as far as I'm aware. The real place to deal with this is at the browser vendor and even the Operating System level. Whatever we do on Wikipedia will be messy and not a long term solution. —TheDJ (talkcontribs) 09:52, 1 February 2016 (UTC)
Where would it be appropriate to ask this question (first one especially which is much more important) elsewhere? Who is responsible for a maintainance of to-this-question related issues? I know it is a complex, but that doesn’t mean it it’s not resolvable. Much more complex software-related issues have been worked out successfully for solution; e.g. thousands of bugs etc. --Obsuser (talk) 23:30, 1 February 2016 (UTC)
The first one could probably be 'fixed' by automatically inserting a Word joiner in front of every single rendered reference, but this might be problematic in certain contexts (and certain languages?), but also icw older browsers and when copy pasting into other applications. This would require a software change. Feature requests like that should be filed in phabricator as mentioned in the header of this page and give 0 guarantee for resourcing towards such a thing. For the latter, I would suggest Apple and Microsoft. —TheDJ (talkcontribs) 00:32, 2 February 2016 (UTC)
Zero-width joiner (&zwj;) works nice. I might ask but nothing is going to get resolved probably. Thank you. --Obsuser (talk) 02:15, 2 February 2016 (UTC)
I'm certain that this exact request has come up before, probably here at VPT; but I'm danged if I can find it.
Anyway, if you wish to somehow connect the punctuation (period, comma, whatever) to the [ that immediately follows it in a non-breaking way without adding extra markup, there is nothing that we can do, not even by filing a request at phab: - it's down to browser behaviour, so you would need to put a change request on the developers of whichever browser that you favour (Firefox in my case). Then repeat for all the other browsers that you use (Opera, Chrome). Then go through the whole thing again for all the other browsers out there (Safari, IE, Pale Moon, etc. etc.) --Redrose64 (talk) 07:24, 2 February 2016 (UTC)
Aha, Wikipedia:Village pump (technical)/Archive 139#Orphan references. --Redrose64 (talk) 07:41, 2 February 2016 (UTC)
I actually did try to cheat with &zwj; using Google Chrome, and it did prevent [1] from being "detached". It is just problem that it increses source code size + we’d need bot that would do massive amount of changes to add this &zwj; before every single <ref + sfn, harvnb and similar ways of referencing.
I’ve already posted this to phab (Maniphest T125480) before reading the last comment here; why do you think it is about browsers? I completely disagree because I guess it is about HTML code; if you use &nbsp; it should work in Chrome, Firefox, Safari etc. Shouldn’t it?
PS "⁠" that someone added to my comment at the very beginning of the sections does not do the job. I see ugly box: ⁠!??!!?   --Obsuser (talk) 07:48, 2 February 2016 (UTC)
I’ve just read Wikipedia:Village pump (technical)/Archive 139#Orphan references. It was like "solved" but obviously it’s not. NOTE: I didn’t see this today, it’s much more older issue (at least one or even two years since I use and/or edit Wikipedia).
Do you have idea why &zwj;-&zwj; does not prevent this - from being broken (e.g. two&zwj;-&zwj;three does not do the job at all)? --Obsuser (talk) 08:03, 2 February 2016 (UTC)
It had degraded and that degradation was solved in several cases. Where it is not solved is probably caused either by browser/OS behavior, user error, or by non-standard styling. I have no problem with Safari on the Barack Obama page for instance... With regards to your example of two&zwj;-&zwj;three; well it's rather logical. You are defining "explicit allow NO break (zwj), explicit allow break(-), explicit allow NO break (zwj)". You can't remove an explicit line break option with zwj. You can only use it to prevent line breaks where there is an implicit option (which are mostly word boundaries) to add a line break. These are well defined rules used throughout computer standards. There is also something like the non-breaking hyphen btw. —TheDJ (talkcontribs) 09:12, 2 February 2016 (UTC)
OK. Thank you. --Obsuser (talk) 22:21, 2 February 2016 (UTC)
Phab crew "archived" it, probably for around 100 years or more... --Obsuser (talk) 09:34, 4 February 2016 (UTC)
It has not been "archived". However, someone did set the priority to accurately reflect the likelihood of this being addressed any time soon. The purpose of the priority isn't to communicate how annoying you find the bug. It's to communicate whether the people who are likely to actually fix it believe that this will be done sooner rather than later. In this case, the problem is purely cosmetic, and therefore it has probably been correctly triaged. Cosmetic problems are normally addressed after non-cosmetic ("Normal") problems, which in turn are usually addressed after widespread breakage ("High"). WhatamIdoing (talk) 19:36, 5 February 2016 (UTC)

Dead Links in reference citations

Are dead links updated by bots on wikipedia? IOW am I wasting my time updating them myself? --BoogaLouie (talk) 15:45, 2 February 2016 (UTC)

I think that Cyberbot II (talk · contribs) will fill in the |archiveurl= and |archivedate= parameters if a |dead link= template is placed just before the </ref>. It can take some time to reach an article though. --Redrose64 (talk) 16:31, 2 February 2016 (UTC)
Thanks for the reply. Any idea how long it is taking on average? --BoogaLouie (talk) 16:53, 2 February 2016 (UTC)
Not on average; but if you look at the bots edits to articles (not talk pages), such as this one, you'll see that it can be more than four years. I guess there's a huge backlog. --Redrose64 (talk) 10:57, 3 February 2016 (UTC)
Erm, the bot only truly went live 3 months ago, and it's still being developed.—cyberpowerChat:Online 19:41, 5 February 2016 (UTC)

Showing a specific part of an image

Hi all! I can show a specific part of an image by using some HTML code:

<div style="height:290px; width:625px;">
<span style="position: absolute; clip: rect(0px 625px 290px 0px);">
<span style="position:relative; left:-1000px; top:-2200px;">
[[File:Woody Guthrie NYWTS.jpg]]
</span>
</span>
</div>

the result looks like this: NOTE result disabled, because it was breaking the page.

Can we invent something a bit more userfriendly? Cropping and reuploading the image doesn't count.

I understand that allowing

div style="background-image: url(

is not an option.

The Quixotic Potato (talk) 19:23, 5 February 2016 (UTC)

There is already a template at {{CSS image crop}}, which uses a similar method. SiBr4 (talk) 19:39, 5 February 2016 (UTC)
Thank you! The Quixotic Potato (talk) 19:43, 5 February 2016 (UTC)

Article url not working?

At Talk:Vine Deloria, Jr. a new editor writes "the official Wikipedia URL of this page does not seem to work correctly and some kind of correction and/or redirect is needed when the link is posted on a social network it prompts people to start a new page rather than takes them to the official one - is this a bug or a problem with the punctuation characters in the Wiki URL?"

Does this make any sense to anyone? Doug Weller talk 21:46, 5 February 2016 (UTC)

Sounds like an URL encoding problem in an external website. The Quixotic Potato (talk) 22:18, 5 February 2016 (UTC)
Or possibly cut'n'pasting the URL into some text editor raw mis-handles the trailing period (trying to be smart, thinking it's a URL at the end of a sentence and therefore omitting it from the actual link). A key missing detail is the URL of the "start a new page" result when clicking on the link. DMacks (talk) 22:19, 5 February 2016 (UTC)
I have created Vine_Deloria_Jr and Vine Deloria, Jr as redirects. I hope this fixes the problem. The Quixotic Potato (talk) 22:21, 5 February 2016 (UTC)
I put a link to this discussion at Talk:Vine Deloria, Jr. --Pipetricker (talk) 00:14, 6 February 2016 (UTC)
This is almost certainly a case of copy-pasting https://en.wikipedia.org/wiki/Vine_Deloria,_Jr. to a place where some software turns the string into a link without including the ending period. Our own sofware does the same for raw url's: https://en.wikipedia.org/wiki/Vine_Deloria,_Jr. That link works via a recently created redirect. The following example currently has no redirect and fails to find Herman Spöring, Jr.: https://en.wikipedia.org/wiki/Herman_Sp%C3%B6ring,_Jr. A workaround is to either create a redirect at the title without the period, or percent-encode the period as %2E in the copied url: https://en.wikipedia.org/wiki/Herman_Sp%C3%B6ring,_Jr%2E At Template talk:No article text#Test for ending period I suggested a way to help users who haven't done either. PrimeHunter (talk) 02:15, 6 February 2016 (UTC)

A question about tables

Is there a way to make a 2-column table where the left and right columns are the same width. The template in the toolbar above the edit window varies the width of each column to achieve a uniform column length, like this: (I've added John's suggested fix, now.) --Anthonyhcole (talk · contribs · email) 07:31, 5 February 2016 (UTC)

Original version Current version
Parkinson's disease (PD), also known as idiopathic or primary parkinsonism, hypokinetic rigid syndrome, or paralysis agitans, is a degenerative disorder of the central nervous system mainly affecting the motor system. The motor symptoms of Parkinson's disease result from the death of dopamine-generating cells in the substantia nigra, a region of the midbrain. The causes of this cell death are poorly understood. Early in the course of the disease, the most obvious symptoms are movement-related; these include shaking, rigidity, slowness of movement and difficulty with walking and gait. Later, thinking and behavioral problems may arise, with dementia commonly occurring in the advanced stages of the disease, and depression is the most common psychiatric symptom. Other symptoms include sensory, sleep and emotional problems. Parkinson's disease is more common in older people, with most cases occurring after the age of 50; when it is seen in young adults, it is called young onset PD Parkinson's disease (PD, also known as idiopathic parkinsonism) is a degenerative disorder of the central nervous system. Many of the motor symptoms of PD result from the loss of pigmented dopamine-generating cells in the substantia nigra, a region of the midbrain. The causes of this cell death are poorly understood. A progressive reduction in the speed and amplitude of voluntary movement is the only physical sign present in all patients and is responsible for the common early complaints of loss of dexterity, writing difficulties, clumsiness and difficulty walking. Muscular stiffness of the limbs and trunk are also common. Trembling of one limb at rest, although not necessarily present in all cases, is the commonest symptom leading to accurate diagnosis. Dysfunction of the autonomic nervous system is also common leading to constipation, swallowing and bladder problems and disturbed temperature regulation with excessive sweating. Depression may be an early sign of the disease and frequently develops as a reaction to increasing disability and social isolation. In elderly patients there is an increased risk of cognitive impairment and dementia. Parkinson's disease is more common in older people, with most cases occurring after the age of 60; when it is seen under the age of 45 years, it is called young onset PD.

--Anthonyhcole (talk · contribs · email) 07:15, 5 February 2016 (UTC)

Yes. If you change the header line to ! style="width:50%" | Original version!! style="width:50%" | Current version you should see two columns of the same width. -- John of Reading (talk) 07:18, 5 February 2016 (UTC)
Thanks John. I replaced the header and it worked. Is there a way to get the shorter block of text to start at the top, rather than being centered in the column? (I'm, obviously, trying to replicate a diff without all the wiki-markup showing.) --Anthonyhcole (talk · contribs · email) 07:26, 5 February 2016 (UTC)
I added style="vertical-align: top;" to table. You may wish also read this, but of course you can ask your questions here. Edgars2007 (talk/contribs) 07:39, 5 February 2016 (UTC)
Yay! Thank you Edgars2007. --Anthonyhcole (talk · contribs · email) 02:23, 6 February 2016 (UTC)
Actually, I need to add a third column called "Sandbox". Would a clever person mind doing that for me? I can't understand Help:Tables. --Anthonyhcole (talk · contribs · email) 03:32, 6 February 2016 (UTC)
Something Like This
column 1 column 2 column 3

-Tagishsimon (talk) 04:19, 6 February 2016 (UTC)

That worked a treat. Thank you Tagishsimon. --Anthonyhcole (talk · contribs · email) 05:39, 6 February 2016 (UTC)

The dot next to edits on watch list

Is it my imagination, or has the little dot to the left of each edit on the watch list become smaller? Since I edit on a small laptop, I regularly have my screen at 150% resolution, and even then, the dot is so small that I can hardly tell whether it is blue (indicating it has been looked at) or green (indicating that it has not been looked at). I don't remember having this difficulty before. Would it be possible to make that dot just a little larger? Corinne (talk) 02:24, 6 February 2016 (UTC)

It hasn't changed for me; IIRC, how the dots display is browser-specific. If you're having trouble seeing it, go to Special:Preferences#mw-prefsection-gadgets and check "Display pages on your watchlist that have changed since your last visit in bold", which makes un-viewed changes far more in-your-face. ‑ Iridescent 02:48, 6 February 2016 (UTC)
IIRC it's been the case for a couple of years that in Vector skin the dots used as list item markers have varied between browsers; in MonoBook, the markers are small squares that are nicely consistent from browser to browser. The change (which was in mid 2013, I think) was basically the way that the markers were specified - instead of being the URL of a file (I think that it was a PNG) on the bits server, it became an inline graphic. In Vector, this inline file is now an SVG:
<?xml version="1.0" encoding="UTF-8"?>
<svg xmlns="http://www.w3.org/2000/svg" version="1.1" width="5" height="13">
<circle cx="2.5" cy="9.5" r="2.5" fill="#00528c"/>
</svg>
which some browsers have difficulty with, even though it's percent-encoded. MonoBook also uses an inline file, but it's a GIF in base 64 format. --Redrose64 (talk) 11:03, 6 February 2016 (UTC)

Shouldn't move log also be able to search with destination page?

For example, Google X has been moved to X (incubator). If you search with Google X, you can find the log. However, if you search with X (incubator), you got nothing. It doesn't look very intuitive for users searching for move logs.--fireattack (talk) 17:17, 5 February 2016 (UTC)

Indeed it isn't. It is often not quite easy to figure out a page's move history. phab:T66184 seems to be where this has been requested, but the task seems fairly inactive. —Kusma (t·c) 17:29, 5 February 2016 (UTC)
@Fireattack: If the page is completely free from cutpaste moves, history merges/splits and deleted revisions, you can trace the history of the page name by examining the page's edit history. Go to that link, and search for "moved page"; doing this, I find three cases: Google X Lab to Google x at 19:35, 5 May 2013; Google x to Google X at 13:03, 6 June 2013; and Google X to X (incubator) at 22:04, 13 January 2016. --Redrose64 (talk) 23:05, 5 February 2016 (UTC)
@Redrose64: Could you elaborate? I see no "search" on history page. If you mean the search function of browser, yeah it does work but the history page can only show up to 500 changes.--fireattack (talk) 00:28, 6 February 2016 (UTC)
Go to the link that I gave, which is for the whole page history; use your browser's search facility (it's Ctrl+F in most popular Windows browsers), enter the words
moved page
and start the search; it should find that text three times.
As an aside, the page history link that I gave is set to limit to 500 entries, but X (incubator) has had 468 edits (as of 23:36, 4 February 2016) since it was created, so every single edit is listed at the link that I gave at 23:05, 5 February 2016. These edits don't just include cases where the page text was altered, but also most loggable actions like page moves, protections and so on. If there had been more than 500, that wouldn't be a problem - although page histories provide clickable links to show up to 500 edits at once, it is possible to display up to 5,000 at once by setting the limit= parameter in the URL's query string, for example https://en.wikipedia.org/w/index.php?title=Wikipedia:Village_pump_(technical)&action=history&limit=5000 . --Redrose64 (talk) 00:50, 6 February 2016 (UTC)
Older page moves before MediaWiki 1.19 have a different edit summary: they simply say "moved Old tile to New title" instead of "Username moved page Old title to New title". You have to consider those separately from the ones done in MediaWiki 1.19 or later. GeoffreyT2000 (talk) 19:35, 6 February 2016 (UTC)
OK, yes, so in this case there were four moves (not three): Google X Labs to Google X Lab at 17:07, 16 November 2011. --Redrose64 (talk) 20:42, 6 February 2016 (UTC)

Double redirect bots

This thread got archived without reply. Has anyone noticed that the double redirect bots are functioning slower than they did years ago, especially in recent history? --Jax 0677 (talk) 12:06, 6 February 2016 (UTC)

Niceguyedc has.[53]Mandruss  13:17, 6 February 2016 (UTC)
I seem to remember double redirects being fixed very quickly (5-10 minutes, maybe even faster) in the past. However, without research, this is just a conjecture. The current time it takes for a double redirect to be fixed now seems to be in hours, not minutes. A look at the most recent contributions of AvicBot suggest that the length of time it takes to fix a double redirect is in a wider range. I see a couple that took over 2 hours to be fixed after the double redirect was created, but there are also others that were fixed in 20 minutes. -Niceguyedc Go Huskies! 13:38, 6 February 2016 (UTC)
Reply - Some of the redirects are taking on the order of days to be fixed. --Jax 0677 (talk) 17:26, 6 February 2016 (UTC)
Xqbot seems to have been catching up this morning. William Avery (talk) 18:20, 6 February 2016 (UTC)
@Jax 0677: Archiving bots don't care if a thread has been replied to or not. They wait for a specified duration after the last post to that thread, and then they archive it. --Redrose64 (talk) 20:38, 6 February 2016 (UTC)
Avicbot hasn't made any updates since 23rd Jan, and I wonder if it has been affected by the problem described at Wikipedia:Bot_owners'_noticeboard#Breaking_change. William Avery (talk) 21:36, 6 February 2016 (UTC)

Automatic reflists in section previews

A while back there was a thread somewhere where someone suggested there should be a preview of references when previewing edits to a section on a page. It looks like that's been turned on, but for articles which use named references which are defined in a different section, or list-defined references, what we're seeing is a preview full of reference errors. Is there a way to turn this preview off? Ivanvector 🍁 (talk) 14:20, 5 February 2016 (UTC)

Is this what is discussed at phab:T124840 and looks like a fix is in the works? I don't think there is an option to turn it off (and if they manage to make it work properly, I am not sure there should be). —Kusma (t·c) 14:27, 5 February 2016 (UTC)
Looks like that's it, yeah. I would prefer to be able to turn it off because I've been referencing fine without it for six years now and I find the format ugly. I mean, what harm can there be in giving users the option of disabling this feature we've been without for 15 years? But if this is the way it has to be, then I guess I'll survive. Ivanvector 🍁 (talk) 14:31, 5 February 2016 (UTC)
The harm would be wasting hours of my time just to please you. I've other things to do in my life. Cenarium (talk) 16:30, 5 February 2016 (UTC)
Really? Like what? Ivanvector 🍁 (talk) 19:21, 5 February 2016 (UTC)
I think that what Cenarium is trying to say, is that the little time he has available to dedicate and donate on working on references, is more effectively used when he works on things that benefit everyone, rather than spent on serving an annoyance of a single person. Working hours for one person or small group, on something that won't be useful anymore 2 weeks after finishing it, is hours of scarce resources wasted. He could also decide to just look at a blank wall for those hours and he would then probably walk away more rested and inspired due to the cathartic experience. :D —TheDJ (talkcontribs) 22:07, 5 February 2016 (UTC)
Hours was excessive in that it would take me maybe a dozen of minutes to add a preference option but I'm not sure if it would be accepted since use case would be quite small. What I could do is add a span class to the refs in section preview so that it can be easily hidden by css (wouldn't take more than a minute). But the point is that, especially when it's to satisfy just one person, I would be more motivated if asked nicely. Ivanvector, the feature implemented here has been requested on many occasions for more than ten years. Cenarium (talk) 09:50, 6 February 2016 (UTC)

I'd really appreciate the option to turn this off in Preferences as well, since I copyedit a lot of GANs and FACs and the long reflists are cumbersome to work around. I do a first, detailed pass through a section before previewing it (and fixing what I missed the first time); the sections typically have many refs, and the long lists make it a lot more difficult to bounce back and forth. An opt-out would be very helpful. Thanks and all the best, Miniapolis 14:57, 6 February 2016 (UTC)

I've written a short script that automatically hides the references and provides a button to display them. You can install it by adding importScript('User:Relentlessly/hiderefs.js'); to your common.js. Relentlessly (talk) 15:21, 6 February 2016 (UTC)
Works great. Thanks a lot! Miniapolis 17:24, 6 February 2016 (UTC)
I also came here with this concern, and I also want to say thanks for the script. It is very helpful. --Tryptofish (talk) 20:29, 6 February 2016 (UTC)
Looks like that'll work just fine, Relentlessly, many thanks. Thanks to you as well, Cenarium and everyone else who's been working on this (I don't know who you all are), I'm sure this feature will be of critical value to the vast majority of users when it is fully debugged. Your reply above about wasting hours of your time was so randomly abrasive that I assumed you were joking, but thanks for looking into this anyway, and many thanks again for the hours not wasted that you commit to the project. Ivanvector 🍁 (talk) 21:16, 6 February 2016 (UTC)
@Relentlessly: You may want to modify the JS once gerrit:268870 is deployed. Cenarium (talk) 16:47, 7 February 2016 (UTC)

I for one really like this new feature of previewing the references when previewing edits to a section on a page. I've been looking for this for years (instead I had to manually add a {{reflist}} to the bottom of the section, which sometimes I'd forget to take out before saving). So thanks to the people who worked on this. Wasted Time R (talk) 13:59, 7 February 2016 (UTC)

Time not wasted, then. ;) Cenarium (talk) 16:47, 7 February 2016 (UTC)

Image display issue

The images in the List of shipwrecks in 1799 are currently displaying centered, instead of at the right side of the page. No idea why, but the code is correct for a right display. Mjroots (talk) 17:36, 3 February 2016 (UTC)

@Mjroots: They look OK to me; am I missing something? What browser are you using? Jdforrester (WMF) (talk) 17:44, 3 February 2016 (UTC)
I'm using Firefox / Windows XP. Mjroots (talk) 17:47, 3 February 2016 (UTC)
Just done a purge but still centered. Mjroots (talk) 17:48, 3 February 2016 (UTC)
@Mjroots: Ah, do you have a quite narrow window (or high zoom level)? The images look fine to me in Chrome/Firefox/Safari/Opera normally, but at or below 768px wide they jump to the centre too. Will investigate. Jdforrester (WMF) (talk) 17:54, 3 February 2016 (UTC)
[ec] @Mjroots: This was done in this edit to Common.css by Edokter following discussion at MediaWiki talk:Common.css#Responsive_image_positioning. I rather like the change, but it's a tad surprising at first. Jdforrester (WMF) (talk) 17:57, 3 February 2016 (UTC)
This is probably related to this Common.css change. What version of Firefox? --Izno (talk) 17:56, 3 February 2016 (UTC)

Oh great, another undiscussed change forced upon us. I tried adding the code to force the image right and it doesn't work. I've got the latest version of Firefox (38.1 or something like that) and tend to run at 125% expansion on my monitor.

Can we get this change cancelled until such time as there is consensus for its introduction or there is a preference to stop it taking effect? Mjroots (talk) 18:15, 3 February 2016 (UTC)

Mjroots, you should consider correcting your claim that this was an "undiscussed change forced upon us", since it was not only actually discussed here at the English Wikipedia, but also done entirely by us, i.e., by a long-time local admin who is well-known for his expertise in this area, at the suggestion of another local editor who is highly respected for his tech skills. WhatamIdoing (talk) 19:51, 5 February 2016 (UTC)
It's also ruined the display of my User Page, which I had carefully crafted over many years to get it looking just the way I liked it.   Mjroots (talk) 18:26, 3 February 2016 (UTC)
@Jdforrester (WMF): - this change has absolutely ruined a great many articles. For example, RMS Magdalena (1948) (GA class). The forcing of the map to the centre means that the lede is forced below the map. Thus you get the infobox with acres of white space to the left, then the map, then the lede and text. Previously, the lede and text sat to the left of the infobox with the map sized to exactly fit underneath the infobox (no white space anywhere). Presumably this affects all 5 million plus articles on en-Wiki, plus all User pages with images displayed. Mjroots (talk) 18:36, 3 February 2016 (UTC)
I reverted. --Redrose64 (talk) 21:11, 3 February 2016 (UTC)
Thank you. Mjroots (talk) 21:35, 3 February 2016 (UTC)
well that experiment ended quickly. 1 objection p00f —TheDJ (talkcontribs) 22:14, 3 February 2016 (UTC)
It hasn't necessarily ended; although it has been demonstrated that further thought needs to be given to how the proposal will work, and how people can avoid unwanted side effects as described above. Mjroots (talk) 22:56, 3 February 2016 (UTC)
Not really, the unwanted side effects are actually the desired effects. —TheDJ (talkcontribs) 22:59, 3 February 2016 (UTC)
If you're using a monitor, or a browser zoomed in, to where your viewport is <768px, you are a dying breed.

As TheDJ commented, the effects were exactly as desired. --Izno (talk) 23:05, 3 February 2016 (UTC)

Many people do not have perfect eyesight (yes, me for instance) and some of those may need to take extra measures - such as zooming in. We should not be creating accessibility issues. --Redrose64 (talk) 23:56, 3 February 2016 (UTC)
Actually, a big point about this proposed change was about improving readability (also a part of accessibility) in narrow views. It's much better to reposition an image or infobox, than to have the text next to it be lines of 3-7 words long. —TheDJ (talkcontribs) 00:07, 4 February 2016 (UTC)
P.S. User:Mjroots, do you use monobook ? —TheDJ (talkcontribs) 00:13, 4 February 2016 (UTC)
@TheDJ: - No. I use the default Vector skin. Mjroots (talk) 05:56, 4 February 2016 (UTC)
"the unwanted side effects are actually the desired effects" - Are you seriously suggesting that it is desirable to force all images to display as centered at all times? It is clearly not, as the description of the problems caused by doing this at the RMS Magdalena article shows.
One way of fixing this is by adding a user preference to override the forcing of images to the centre. Mjroots (talk) 06:08, 4 February 2016 (UTC)
"at all times", no, as stated before, this only takes affect on narrow screens (or rather viewports), where without this change, paragraphs next to infoboxes would only be 3-7 words long (at their maximum), which is terrible legibility. If you are seeing this always.. then...u are using an ancient monitor (768*125% == 960 pixels. Those screens haven't been sold after the year 2000 if I think). So no, It is not always. For most people it is only if they make their browser window very small. Also, if you use the mobile website, then it always behave likes this, so it can be useful to help editors understand how content works in the mobile website. They would just make their window very small and they can see how a page will behave on a mobile device. —TheDJ (talkcontribs) 07:56, 4 February 2016 (UTC)
I can assure you my monitor is not ancient. It's a flat screen monitor, 1024 x 768 px. Thumnails are set to 220 px in my preferences. Mjroots (talk) 08:29, 4 February 2016 (UTC)
I just learned that apparently 7,5% of the world uses similar screens (although probably doesn't run their browsers at 125% in that case, which really still would be a tiny minority). Which is crazy if you consider that since 2011 there have been phones supporting that resolution. Anyway, for the sake of the health of your eyes, I advise all of you to invest 40-130 bucks to buy a new screen.
Secondly, I guess we could drop the 768 cut off point slightly down to something like 700, where possibly it might not kick in yet for you, but still give 'some' value for other people... It's annoying however in that 768 is the value that is also used as the cut off point between tablets and phones, and using multiple values will just create more confusion I fear. —TheDJ (talkcontribs) 09:08, 4 February 2016 (UTC)
I don't have $40-130 to spare, as I'm on a limited income. You say 7.5%, that's roughly 1 in 12 then, a large enough portion that it shouldn't be ignored. Either code the change so that it can be overriden, or you'll have to go the way of separate coding for smartphones, tablets and PCs. Mjroots (talk) 19:10, 4 February 2016 (UTC)
Did anyone other than Mjroots have a problem with this? WhatamIdoing (talk) 19:51, 5 February 2016 (UTC)
The edit was not live long enough for anyone but Mjroots to complain, IMO.

Regardless, his complaint that his user page is screwed up is his problem; the change in articles shouldn't and in fact is a net positive, given the lack of other known issues. --Izno (talk) 19:54, 5 February 2016 (UTC)

I wouldn't definitively say that, but at least there is not the option of finding out now either way. —TheDJ (talkcontribs) 13:07, 6 February 2016 (UTC)
I believe I adequately explained the problem caused at the RMS Magdalena (1948) article above. This was by no means the only articl affected. I've had a redesign of my user page, thanks for your concern over that issue. The edit was live for enough time for people to raise the issue, but not everyone knows where to report it. Mjroots (talk) 18:27, 7 February 2016 (UTC)

Identically looking images have different file sizes

So, I've noticed that File:York station zero post.jpg and commons:File:York station zero post.jpg look exactly identical but the latter has EXIF metadata and the former hasn't and yet the latter (according to Stefan2) appears to be smaller in byte size than the former. Does someone know why they would be? File:Flag of KDP.png and commons:File:Flag of KDP.png likewise look identical and have different filesizes, but in that case the EXIF indicates slightly different resolutions.Jo-Jo Eumerus (talk, contributions) 11:06, 8 February 2016 (UTC)

The file history sections indicate the original versions were identical but new versions have been uploaded after being processed in different ways or with different software. Image processing software often changes the file size and makes different metadata. File:York station zero post.jpg must have been rotated by a program that can increase the size of jpg files a lot. 37.78 or 37.79 dpc just sounds like a question of how to round. PrimeHunter (talk) 11:40, 8 February 2016 (UTC)
  • The files are different for some reason, so they should be tagged with {{ShadowsCommons}} until the problem has been solved. In this case, it would not be appropriate to move one of the files, though. It may be better to synchronise the files so that they are identical, or delete the local copies at FFD. --Stefan2 (talk) 12:40, 8 February 2016 (UTC)

Hidden table caption still available to screen readers?

If I do this:

{|class="wikitable""
|+style="display:none;"|Caption
!Header1
!Header2
|-
|data1
|data2
|-
|}

then MediaWiki produces this HTML:

<table>
<caption style="display:none;">Caption</caption>
<tr>
  <th>Header1</th>
  <th>Header2</th>
</tr>
<tr>
  <td>data1</td>
  <td>data2</td>
</tr>
</table>

which renders like this (table without a visible caption):

Caption
Header1 Header2
data1 data2

Since the <caption>...</caption> tags are present in the HTML source, is the caption still available to screen readers?

Trappist the monk (talk) 17:13, 6 February 2016 (UTC)

@Trappist the monk: Nope, not in the latest versions of JAWS and NVDA. Graham87 11:57, 7 February 2016 (UTC)
Well, bother. Thank you. What if I set the table caption this way: |+style="visibility:hidden;"|Caption? Does this table have a caption?
Caption
Header1 Header2
data1 data2
The caption isn't visible on the screen but the space it occupies is maintained.
Trappist the monk (talk) 12:31, 7 February 2016 (UTC)
I don't know precisely what you're using this for, but perhaps you should be using table summaries instead? Relentlessly (talk) 12:42, 7 February 2016 (UTC)
At WP:SHIPS we are considering changing the list of shipwreck lists from a more-or-less consistent text-based format (see for example List of shipwrecks in 2010) with a more structured list (see my sandbox of the same data). It has been suggested that the table captions contribute to visual clutter so I'm looking for a way to have the captions for accessibility but not show them visually.
Trappist the monk (talk) 14:13, 7 February 2016 (UTC)
Are there any accessibility issues with setting the font colour to white, so that the text is displayed like this? Mjroots (talk) 21:20, 7 February 2016 (UTC)
Yet another possibility along these same lines is to set the caption |+style="opacity:0;". To reduce the amount of white space above the table, set margin-top:-1.5em:
Caption
Header1 Header2
data1 data2
Trappist the monk (talk) 22:54, 7 February 2016 (UTC)
Screen reader software doesn't care about color, opacity or similar - it gets read out just the same as if it were completely unstyled. --Redrose64 (talk) 10:00, 8 February 2016 (UTC)
Exactly. The "visibility:hidden;" isn't read out by screen readers. Perhaps a summary would indeed be better in this case. Graham87 14:08, 8 February 2016 (UTC)

18:58, 8 February 2016 (UTC)

Default block length for Checkuserblock-account

Hi all--I propose that some clever person flip the switch to change the default block length related to Template:Checkuserblock-account from 31 hours to indefinite. I have a few reasons--first of all, we typically block such accounts indefinitely, even if we give the master account more leeway. Second, the template actually says, "It has been blocked indefinitely to prevent abuse". In addition, I'd like for this block to also have the "block other logged-in users etc." box checked, and the "prevent account creation"--are these two not the essence of CU blocks, and the creation and logging in of other accounts the essence of socking? I'm pinging Fluffernutter, since I've been told she's in charge, and Iridescent, to provide sound effects. Drmies (talk) 16:21, 8 February 2016 (UTC)

This might be more appropriate to discuss amongst Enwiki checkusers first. However, I will note that I've never had a "default block time" of 31 hours for any kind of block I have ever done, checkuserblock or otherwise. I suspect Drmies might have some sort of script that is automatically filling in that number for him. Risker (talk) 16:35, 8 February 2016 (UTC)
@Risker: as I've just confirmed on the long-suffering User:ThisIsaTest, if you disable all user-scripts and use the default Twinkle "block" option, it normally defaults to 24 hours but if you select "checkuser block" from the pull-down, it automagically changes the block duration to 31 hours. ‑ Iridescent 16:59, 8 February 2016 (UTC)
Well, then. That's an issue with Twinkle, not with default blocking, and the instructions for contacting the Twinkle developers is at Wikipedia talk:Twinkle. (It's at github, which probably means setting up an account...) I note that Drmies also has two other block default user scripts active, which may be further complicating things. Risker (talk) 17:11, 8 February 2016 (UTC)
On looking further at the "defaults" for various types of blocks as identified in Twinkle, there are several other ones that I think are problematic (including not blocking talk page access by default when leaving a template saying you're blocking talk page access...) Might be worthwhile to have some folks who regularly use Twinkle for blocking to sort it all out and make more sense of it. I've had Twinkle loaded for a long time but have never considered using it for blocking, but obviously some folks find it useful. Risker (talk) 17:19, 8 February 2016 (UTC)
Risker, I'm not the sharpest knife in the drawer; I suppose I am using Twinkle for blocking, but I'm not doing so on purpose. I do find it useful, esp. when taking care of backlogs at UAA etc., and the better the defaults, the easier my job is. So whoever can make this better and iron out some kinks, power to them. Drmies (talk) 23:28, 8 February 2016 (UTC)

URL generates error

This URL generates an error:

https://en.wikipedia.org/wiki/Władysław_T._Benda?action=raw

Normally it should download the raw wikisource such as:

https://en.wikipedia.org/wiki/Charles_Dickens?action=raw

-- GreenC 14:21, 6 February 2016 (UTC)

My guess its being caused by the presence of polish lettering (the ' ł ') while under the English Wikipedia domain and/or the presence of a dot after the middle initial "T" in the URL. It works fine in the Polish Wikipedia ...
... but they didn't include a middle initial in their title so I can't say for sure which one (or the rare combination of both) is causing the error here. -- George Orwell III (talk) 00:38, 7 February 2016 (UTC)
The period seems to be the culprit, since the redirect without middle initial does work, and other titles with period (e.g. .com, John F. Kennedy) don't. Strangely, URLs with the title in the query string, such as https://en.wikipedia.org/w/index.php?title=Władysław_T._Benda&action=raw, don't cause problems. SiBr4 (talk) 00:51, 7 February 2016 (UTC)
I created phab:T126183 for this. — This, that and the other (talk) 01:28, 8 February 2016 (UTC)
Thank you. Confirmed the period. -- GreenC 00:18, 9 February 2016 (UTC)

Navbox list1 parameter and div

I'm working on a module that effectively uses {{navbox}} to produce a navbox. Some of the more than 1,000 navboxes that will use the module use the wikitext shown on the left of the following to display a note after the list of items. Others use the wikitext on the right. Which is correct?

{{navbox                           {{navbox
|...                               |...
|list1=<div>                       |list1=<div>
*First item                        *First item
*Second item                       *Second item
<small>Note inside div</small>     </div>
</div>                             <small>Note after div</small>
}}                                 }}

The documentation at Template:Navbox/doc#Cells for listn* includes "although the text can be entered on separate lines if the entire list is enclosed within <div> </div>" which suggests the div is necessary, although I cannot see a difference when trying without the div.

A full example is in my sandbox (permalink) and there is a bit more information at Template talk:Navbox#list1 div with a note. Having the note after the div produces a slightly smaller vertical gap which seems preferable, but I am hoping for some clarity on what my module should output. Johnuniq (talk) 01:03, 9 February 2016 (UTC)

Looking for reports of users being logged in as someone else

Hi all,

we had reports of a serious authentication error that resulted in people being logged into the wrong account. We are trying to collect information about what happened, and how many users it might have happened to. If you have any knowledge of such a thing happening in the last few weeks, please tell us at one of these places:

You can find more details at https://phabricator.wikimedia.org/T126069. It looks like only a few users have been affected, but to be on the safe side, we are in the process of logging everyone out (which takes a while, so it might have happened to you this week, or it might happen in the next few days). Apologies for the inconvenience.

-- Tgr_(WMF) (talk) 23:47, 5 February 2016 (UTC)

we are in the process of logging everyone out (which takes a while... >> this was deemed unacceptably slow, so we are going to log out everyone at the same time. This is probably going to happen within a few hours, so that we can do it while most editors are asleep. I apologize for the disruption. Please report any unexpected problems (apart from having to log in again) at https://phabricator.wikimedia.org/T124440 . --Tgr (WMF) (talk) 04:23, 9 February 2016 (UTC)

Special:NewPagesFeed

Uh, so... can anyone explain why the New Pages Feed has things going all the way back to 2001?

 

There are about 94 items in here that are from more than a year ago (65 of those being more than 5 years old, 26 of those being more than 10 years old), and an additional 24 that are from more than a month ago. There are also an additional 35 redirects that are more than a month old (21 being more than a year old). Why they are in here doesn't quite add up. Zeke Essiestudy (talk) 05:59, 7 February 2016 (UTC)

@Zeke Essiestudy: When an old redirect is converted to an article, the page is added to the new pages feed so that the new article gets checked. Unfortunately, Connecticut was briefly turned into a redirect by mistake and then converted back into an article, and I guess the software isn't yet clever enough to tell this case apart. I haven't checked any of the other pages listed at the feed. -- John of Reading (talk) 08:16, 7 February 2016 (UTC)
The same also occurs when a non-main namespace page is moved to the main namespace. For example, Draft:Cerbat Mustang was moved to Cerbat Mustang by Montanabw. GeoffreyT2000 (talk) 15:31, 7 February 2016 (UTC)
Hm. Very strange. It appears as though Hulk (comics) was vandalized into a redirect yesterday, as was other media things in there, and this appears to be the culprit. But seriously, I marked it as unreviewed to see what happens and it was. It was treated like a new page just because of that vandalistic redirect but it's was made 14 years ago... I guess this is rampant on all of the 118 things that are more than a month old. That's annoying. :/ Zeke Essiestudy (talk) 15:41, 7 February 2016 (UTC)
The question is, how smart should the software be? This feature is designed to get around the creation of new articles by overwriting redirects, a not-uncommon method to get passed Special:Newpages. עוד מישהו Od Mishehu 18:59, 7 February 2016 (UTC)
The idea is that the software should scan a page that gets changed into a redirect and vice versa. If the WHOLE page's history is older than a month (I assume that's the threshold), then it doesn't go into the feed. If it's new, then it does. Zeke Essiestudy (talk) 00:57, 8 February 2016 (UTC)
@Zeke Essiestudy: There will be cases where that gives the wrong results: an article created in 2008, redirected in 2009, edited into a new article today - that should appear in the feed, I would have thought. The distinction could be the length of time the article spent as a redirect. Connecticut was a redirect for only a couple of minutes. -- John of Reading (talk) 06:57, 9 February 2016 (UTC)

The someone a there is a properly configured pywikibot?

I am not able to run no script, although 3 years ago I have enjoyed pywikibot. I tried to ask the #pywikibot , but there is not responding.

C:\Python34\pywikibot>version.py
WARNING: Configuration variable 'use_api' is defined but unknown.
Misspelled?

C:\Python34\pywikibot>

My user-config.py. Thanks.--Парис "Анима" надаль (talk) 08:37, 8 February 2016 (UTC)

@Парис "Анима" надаль: It looks like the use_api option was removed from Pywikibot. Try deleting that line from your user-config.py and see if it works. Best — Mr. Stradivarius ♪ talk ♪ 09:29, 9 February 2016 (UTC)

How to use wiki markup to make a colored border around a single image?

Adoxa (Adoxa moschatellina)

Example. Thanks.--Парис "Анима" надаль (talk) 11:35, 4 February 2016 (UTC)

Not possible using wiki markup; such borders can only be done in HTML/CSS. -- [[User:Edokter]] {{talk}} 11:56, 4 February 2016 (UTC)
By means of Template:Multiple_image#with_background_color it is possible to do the coloured borders at a few images. Can in somewise also to designate the coloured borders for one image? Thanks.--Парис "Анима" надаль (talk) 12:46, 4 February 2016 (UTC)
@Парис "Анима" надаль: Like this?
{{multiple image|caption_align=center|header_align=center
 | background color = #BBDD99
 | width = 160
 | image1 = Adoxa_moschatellina_210406.jpg
 | alt1 = Green herb with a few tiny yellow-white flowers
 | footer_background = #BBDD99
 | footer_align = center
 | footer = [[Adoxa]] (''Adoxa moschatellina'')
}}
The Quixotic Potato (talk) 10:09, 9 February 2016 (UTC)

Emojis in edit summaries

Is there any way to block something like this edit summary from re-occuring with a filter? Note if you don't have a mobile/tablet it might not be visible, but basically the edit summary is spamming flag emojis into the edit summary. I didn't notice it until I read my watchlist on my iPad this morning and could be obnoxious when desktop browsers eventually are able to read emojis full-time (I use emojis but usually never in edit summaries). Nate (chatter) 20:04, 8 February 2016 (UTC)

I don't care for it—and agree that it's confusing for those using browsers which can't handle them—but how is this any more "obnoxious" than the use of Chinese characters, for instance? ‑ Iridescent 20:18, 8 February 2016 (UTC)
Special:AbuseFilter/680 disallows emojis in the content. We can modify it also check the edit summary, but in this case that user's edit would go through anyway as they are autoconfirmed. I would reach out to them and explain this is disruptive. Pining filter author Dragons flight for input MusikAnimal talk 21:38, 8 February 2016 (UTC)
I've used emojis in edit summaries lots of times, not counting the bit we went through at RfD a while back discussing redirects with emojis in the page titles. I think a flat restriction on their use would be inadvisable as they're not inherently disruptive, but the one that you linked to is probably nothing but. Ivanvector 🍁 (talk) 21:41, 8 February 2016 (UTC)
Sure, we don't want to disallow them entirely, only for unconfirmed users as that is where the bulk of disruption comes from MusikAnimal talk 22:19, 8 February 2016 (UTC)
Yeah, I'm not saying 'ban all emojis' by any means, just have a good limit on how they're used so that we don't edits like that or Gr**p-like abuse of them. Nate (chatter) 04:17, 9 February 2016 (UTC)
I point out that it doesn't seem that anyone has informed the user of the undesirable emojis... —TheDJ (talkcontribs) 18:03, 9 February 2016 (UTC)

Sorted tables problem

Does anybody knows why there are icons to sort the columns in the first column of every row in the table?

{| class="wikitable sortable"
! Some 1 !! Some 2
|-
! style="font-weight: normal; text-align: left;" | A 
| C
|-
! style="font-weight: normal; text-align: left;" | B 
| D
|}
Some 1 Some 2
A C
B D

Such tables were displayed correctly a few days ago. However, if I change the code to

{| class="wikitable sortable"
! Some 1 !! Some 2
|-
| A 
| C
|-
| B 
| D
|}
Some 1 Some 2
A C
B D

Then I loose the nice background in the first column. Paweł Ziemian (talk) 10:24, 6 February 2016 (UTC)

There are two types of cell in tables: header cells (for information about the whole row or column) and data cells. They are not interchangeable. Header cells should not contain regular data, even if you restyle them to resemble data cells. Doing this is an accessibility issue. If you want the cells to contain regular data, but have the grey background, style it with that background:
{| class="wikitable sortable"
! Some 1 !! Some 2
|-
|style="background-color: #F2F2F2;" |A 
| C
|-
|style="background-color: #F2F2F2;" |B 
| D
|}
Some 1 Some 2
A C
B D
Start with the correct semantics, and then add presentational styles - never start with something that has inappropriate semantics even if it has the presentational styling that you want. --Redrose64 (talk) 10:34, 6 February 2016 (UTC)
Redrose64, the use of row headers is recommended by WP:DTT and there has been an unhelpful/breaking change here. I'd have thought the plainrowheaders class is probably what's intended here. Relentlessly (talk) 11:10, 6 February 2016 (UTC)
The examples above use the values "A" and "B" in the first column, and the similarity of these to "C" and "D" in the second column implies to me that "A" and "B" are regular data items. Perhaps better (i.e. more distinctive) examples could have been chosen.
If it's a true header cell, and it applies to the row, it should have the attribute scope=row - although the spec says that the omission of that attribute is equivalent to the auto state, some browsers will treat its absence as being implicitly scope=col --Redrose64 (talk) 11:21, 6 February 2016 (UTC)
Indeed, and this is precisely what I reported in the bug tracker. Relentlessly (talk) 11:32, 6 February 2016 (UTC)
I reported this on Phabricator yesterday. phab:T126007. It's a change related to VisualEditor, I think, but it's clearly an unintended consequence. Relentlessly (talk) 10:38, 6 February 2016 (UTC)
Thanks. Paweł Ziemian (talk) 13:54, 6 February 2016 (UTC)
The bug is now fixed. Relentlessly (talk) 21:15, 9 February 2016 (UTC)

Missing "rotate file" command in a lot of files in Category:Ships

How can I advice to rotate these files in Commons? --Stunteltje (talk) 19:59, 9 February 2016 (UTC)

Add {{rotate|90}} to request rotation by 90 degrees clockwise. Replace 90 with another degree count, such as 180 or 270, wherever necessary. --Stefan2 (talk) 21:03, 9 February 2016 (UTC)
Thanks. It works. --Stunteltje (talk) 22:45, 9 February 2016 (UTC)

Pending Changes madness

Why is it that if a page is on pending changes 1, and some IP makes a bad edit whilst I'm making a good one, mine needs to be independently accepted even though I have admin rights; and moreover, why can I not reject the bad edit without also rejecting my own good one? See Matt Smith (actor). All I was trying to do was add a {{pp-pc1}}. --Redrose64 (talk) 23:51, 9 February 2016 (UTC)

"You are centrally logged in..."

Why does the system keep logging me out and saying "Central login: You are centrally logged in as JohnCD. Reload the page to apply your user settings"? Five times this evening - this is getting to be a serious pain. JohnCD (talk) 21:15, 9 February 2016 (UTC)

I just got a message when I logged in saying, "No something something login attempt for this session". I've never seen that one before. Maybe the two are connected. White Arabian Filly (Neigh) 21:30, 9 February 2016 (UTC)
Probably phab:T124440. --Redrose64 (talk) 21:53, 9 February 2016 (UTC)
No, i think it might actually be T126395. An incorrectly configured server was added to the pool, and some requests would end up there and not be able to see the session data. For me it was edit saves/previews, but what JohnCD describes might have been a different effect of that same root cause. If so, then likely by the time I post this, the problem should have resolved itself. —TheDJ (talkcontribs) 23:57, 9 February 2016 (UTC)

What is the point of these web bugs?

I was trying to read the train crash article when, perhaps due to a poor connector, perhaps not, the page hung while trying to access species.wikimedia.org. This is one of the web bugs mentioned in the Privacy Policy. The actual text of the bug:

<div id="central-auth-images" style="position: absolute;"><img src="//meta.wikimedia.org/wiki/Special:CentralAutoLogin/start?type=1x1&from=enwiki" alt="" title="" width="1" height="1" style="border: none;" /><img src="//en.wiktionary.org/wiki/Special:CentralAutoLogin/start?type=1x1&from=enwiki" alt="" title="" width="1" height="1" style="border: none;" /><img src="//en.wikibooks.org/wiki/Special:CentralAutoLogin/start?type=1x1&from=enwiki" alt="" title="" width="1" height="1" style="border: none;" /><img src="//en.wikiquote.org/wiki/Special:CentralAutoLogin/start?type=1x1&from=enwiki" alt="" title="" width="1" height="1" style="border: none;" /><img src="//en.wikisource.org/wiki/Special:CentralAutoLogin/start?type=1x1&from=enwiki" alt="" title="" width="1" height="1" style="border: none;" /><img src="//commons.wikimedia.org/wiki/Special:CentralAutoLogin/start?type=1x1&from=enwiki" alt="" title="" width="1" height="1" style="border: none;" /><img src="//en.wikinews.org/wiki/Special:CentralAutoLogin/start?type=1x1&from=enwiki" alt="" title="" width="1" height="1" style="border: none;" /><img src="//en.wikiversity.org/wiki/Special:CentralAutoLogin/start?type=1x1&from=enwiki" alt="" title="" width="1" height="1" style="border: none;" /><img src="//www.mediawiki.org/wiki/Special:CentralAutoLogin/start?type=1x1&from=enwiki" alt="" title="" width="1" height="1" style="border: none;" /><img src="//www.wikidata.org/wiki/Special:CentralAutoLogin/start?type=1x1&from=enwiki" alt="" title="" width="1" height="1" style="border: none;" /><img src="//species.wikimedia.org/wiki/Special:CentralAutoLogin/start?type=1x1&from=enwiki" alt="" title="" width="1" height="1" style="border: none;" /><img src="//incubator.wikimedia.org/wiki/Special:CentralAutoLogin/start?type=1x1&from=enwiki" alt="" title="" width="1" height="1" style="border: none;" /><img src="//en.wikivoyage.org/wiki/Special:CentralAutoLogin/start?type=1x1&from=enwiki" alt="" title="" width="1" height="1" style="border: none;" /><img src="//login.wikimedia.org/wiki/Special:CentralAutoLogin/refreshCookies?type=1x1&wikiid=enwiki&proto=https" alt="" title="" width="1" height="1" style="border: none;" /></div></div><div class="printfooter">

As I understand the purpose of this is to allow the setting and/or reading of third-party cookies - actually three per site, namely CP, WMF-Last-Access, enwikibooksSession. The first is a short code, the second a date, and the third some kind of hash. Now the question I have to ask is: do we really need 14 x 3 = 42 different cookies to log into Wikimedia? If you want to use third-party cookies, why isn't one enough? And isn't it an option to just not use third-party cookies? I mean, if I go to es.wikipedia.org it loads right away, I seem to be logged in, and after that I have the three es.wikipedia.org cookies -- but I didn't have them before that. So these things seem unnecessary, they can hang and leave a page half loaded, they make it take extra long right after logging in. Why are they there? Wnt (talk) 17:37, 9 February 2016 (UTC)

Yes, each domain needs cookies in order to do a full login across all Wikimedia properties. It's the 14 that is the issue here (the 'expensive' part), not the 3. It shouldn't hang on that however, they are added and loaded fully asynchronously. Your browser might show that it is "loading" something and failing at it, but that's just fine. If it's blocking on that, then there is another problem. —TheDJ (talkcontribs) 18:00, 9 February 2016 (UTC)
You're correct, the purpose of these is to log you in to all the WMF wikis. To log you in everywhere, they need to be set on each domain, which explains most of the web bugs: .wikibooks.org, .wikinews.org, .wikipedia.org, .wikiquote.org, .wikisource.org, .wikiversity.org, .wikivoyage.org, .wiktionary.org, mediawiki.org, and wikidata.org. And we can't set just one on .wikimedia.org because externally-hosted sites such as blog.wikimedia.org are also under that domain, so we need to do Meta, Commons, Incubator, and Wikispecies each separately. Of course, whichever one corresponds to the domain you're actually on doesn't get a web bug.
The web bug hitting login.wikimedia.org (with a different path) is trying to update the cookies after the login process, which had already redirected you to loginwiki (to set an initial set of cookies) and back.
As for the cookies themselves: The "enwikibooksSession" cookie is a bit of a side effect (getting only this one is possible in some cases if your browser isn't sending the third-party cookies to loginwiki during the redirect-handshake behind those web bugs, or is blocking the redirects at some stage of the process), the really important cookies for central login are "centralauth_Session", "centralauth_User", and (if you checked "remember me") "centralauth_Token".
The "CP" and "WMF-Last-Access" aren't necessary at all for login; "CP" is apparently for indicating whether the connection is using HTTP/2 for some future optimization of ResourceLoader's bundling (why it needs to be a cookie I have no idea), while "WMF-Last-Access" is something for WMF's Analytics team to count unique visitors per day or something like that.
HTH. Anomie 00:51, 10 February 2016 (UTC)

File:FC Barcelona (crest).svg

File:FC Barcelona (crest).svg (edit | talk | history | links | watch | logs)

I'm wondering if any one can figure out why User:Lawrencedepe is still showing up in the "File usage" section for this particular file even though it was removed User:Lawrencedepe/userbox by Finnusertop per WP:NFCC#9. Just curious as to why this is happening. Thanks in advance. -- Marchjuly (talk) 01:19, 10 February 2016 (UTC)

I did a WP:Null edit on the user page and that seems to have sorted it out - Evad37 [talk] 04:12, 10 February 2016 (UTC)
Thanks for checking and fixing things Evad37. -- Marchjuly (talk) 04:58, 10 February 2016 (UTC)

Signpost book PDF fails rendering

Hi all,

Book:Wikipedia_Signpost/2016-02-03 fails rendering if the Op-ed section is included. Is there some formatting element that could be removed from that section to fix the problem? --Slashme (talk) 07:50, 10 February 2016 (UTC)

Accidentally logging out when trying to use search bar etc. on mobile

When I'm on mobile, sometimes I accidentally hit the logout button instead of the search bar or any other button on the top right. and this is something that happens on both phone and tablet. Yes I realize this is specifically what the mobile sote is supposed to solve, but that interface is still pretty limited and doesn't display things like navboxes and categories. Maybe the logout button could just be moved to the bottom of the page or something. Any thoughts?--Prisencolin (talk) 21:31, 9 February 2016 (UTC)

You mean when you use the desktop site on a touch based device ? —TheDJ (talkcontribs) 21:47, 9 February 2016 (UTC)
@Prisencolin: With CSS you can hide parts of the MediaWiki interface you dislike. Add this line:
li#pt-logout {display:none;}
to this file: Special:MyPage/common.css
and the Logout link will be hidden. Another option is to move the link, for example:
li#pt-logout {position:absolute; top:0px; left:5px;}
Or you can move the entire line a bit to the left:
div#p-personal {padding-right:200px;}
The Quixotic Potato (talk) 09:54, 10 February 2016 (UTC)
Wow I never even realized this even existed, looks like the problem is solved now. Thanks a bunch.--Prisencolin (talk) 19:07, 10 February 2016 (UTC)

input of a web link

I have a weblink http://www.surinamstars.com/Gloria Wekker.html which is showing up improperly in my document because there is a space in the title. I am sure there must be some sort of coding that must be input to tell it that it is an actual space, but I have zero clue how to do that. It is only reading the name to the end of Gloria and then giving a 404 error. [64] How do I circumnavigate this? SusunW (talk) 18:41, 10 February 2016 (UTC)

Code %20 in place of the space. For future reference, WP:Teahouse or WP:Help desk are better places for how-to questions. ―Mandruss  18:53, 10 February 2016 (UTC)
Thank you. SusunW (talk) 19:07, 10 February 2016 (UTC)

Ganglia exposed - someone suggests this is a problem

At ticket:2016021010021773 someone wrote into WP:OTRS saying that https://ganglia.wikimedia.org/latest/ is supposed to be password protected and that it must be an oversight that it is not. Any response? I am directing that person to look here for a response. I know nothing about this. Blue Rasberry (talk) 00:17, 11 February 2016 (UTC)

There is no problem. This is meant to be public. ^demon[omg plz] 00:32, 11 February 2016 (UTC)

Edit acts like it won't save

When I make an edit now (this has been going on for the past 30 minutes), I type in the summary and click save, and get a little spinner or wait symbol or whatever going around...and around...and around. It also says, "Saving edit, please wait". When I finally get fed up and back off the screen after 5 minutes, nothing happens, but a few minutes later I get an edit saved bubble. This is very annoying, especially because I'm on the verge of getting my first GA. White Arabian Filly (Neigh) 02:10, 9 February 2016 (UTC)

@White Arabian Filly: Have you tried using a different web browser than you would normally use (assuming you have more than one)? For example, if you normally use Internet Explorer, try Firefox or Chrome. The Quixotic Potato (talk) 09:54, 9 February 2016 (UTC)
I don't have more than one. My only internet access is off a smartphone. The mobile view of WP has changed in several ways recently and I think that's affecting it, because I'm not having issues with other sites. White Arabian Filly (Neigh) 16:30, 9 February 2016 (UTC)
I'm experiencing the same thing you are. I use mobile Wikipedia on Chrome browser on a phone (no app). The only way to save an edit in mobile view is to wait around 3 seconds after pressing save then press exit. If it doesn't give the "abandon your edit" message it has saved. If it does you press cancel then try again. You do have to then reload the page to actually see your changes. It's rather annoying but at least it's not broken. I'm assuming it will be fixed some day. I encounter lots of small bugs that annoy but are usually fixed a relatively short time after onset. There's been several changes lately, some are annoying but the intentional changes you get used to. These problems obviously don't occur in desktop view. —DangerousJXD (talk) 21:50, 9 February 2016 (UTC)
Yep, I've had the same issues, and normally they are fixed within a day or two...or three. One time the ability to edit on mobile was totally gone for 3 or 4 days but when fixed was quicker than before. White Arabian Filly (Neigh) 21:58, 9 February 2016 (UTC)
Reading this again, I suspect that these problems were also attributable to T126395, where there was one bad server deployed for a few hours today. If so, then now this problem should be resolved. —TheDJ (talkcontribs) 00:03, 10 February 2016 (UTC)
@White Arabian Filly, The Quixotic Potato, DangerousJXD, and TheDJ: I'm very sorry about this screw-up. It looks like something broken by my department as part of wider work, and (of course) shouldn't have happened. We have a fix on the way, and it should be deployed in a few minutes' time. Jdforrester (WMF) (talk) 22:18, 10 February 2016 (UTC)
@White Arabian Filly, The Quixotic Potato, DangerousJXD, and TheDJ: Update: this should now be fixed (tested here and on other wikis). Again, I'm very sorry this happened. If you still have issues please ping me and I'll get them fixed. Jdforrester (WMF) (talk) 22:49, 10 February 2016 (UTC)
Jdforrester (WMF), after making two edits in mobile view since this latest post, I discovered that when I press save, it closes that but I still have to reload the page to see the changes. —DangerousJXD (talk) 22:54, 10 February 2016 (UTC)
It's a mixed bag at the moment. Out of the five edits I have made in mobile view since your latest post, one of them saved properly and the rest I have had to refresh the page to see the changes. —DangerousJXD (talk) 23:14, 10 February 2016 (UTC)
@DangerousJXD: Hmm. I don't get that issue; note that there's a brief moment after saving when the old version is shown, but then it refreshes (on short pages on which I'm testing, maybe half a second). If you wait longer does it happen for you, or does it stay that way even if you wait for 10 seconds or so? Jdforrester (WMF) (talk) 23:54, 10 February 2016 (UTC)
Jdforrester (WMF), sorry for not being around to provide a faster response. Even after waiting for over a minute after it has closed the page where you fill in the edit summary and press save, it does not refresh the page or display the changes. It worked how it used to once. Edits are saving though so it's not broken. I just have to refresh the page every time I make an edit to see the changes myself and to get the "your edit is a success" message. —DangerousJXD (talk) 01:00, 11 February 2016 (UTC)
First rule of software development: shit happens. The Quixotic Potato (talk) 01:22, 11 February 2016 (UTC)

Question on search results

I noticed just now that if I type "plowback" into the search box, the drop-down list shows me the redirect plowback retained earnings but not the redirect plowback. If I hit enter, I am taken to retained earnings with a note that I was redirected from plowback. Is there a clear set of rules to predict or influence this behaviour (i.e. through rcats or something)? Ivanvector 🍁 (talk) 21:51, 10 February 2016 (UTC)

Howdy Ivanvector, I spoke to one of the search engineers and got an explanation. "Plowback retained earnings" is shown when you type "plowback" since when we have multiple redirects that go to the same page we only show one of them in the suggestions. "Plowback retained earnings" gets a higher score than "plowback" because it is a "better" match - possibly because it contains the redirect target. That's the autocompletion suggester working.
If you type 'plowback' into the search box and just press Enter/Return you perform a search for "plowback", It does not search for the redirect it showed you (plowback retained earnings) because that was not selected! Searches that come from the autocomplete bar have a special feature called 'go' (what a confusing name, right?) where if it matches the title of a page you get auto-redirected to that page. We have a task to make things a little more consistent (I'll see if I can dig it up), but that's the short explication. CKoerner (WMF) (talk) 23:52, 11 February 2016 (UTC)
Per mw:Help:CirrusSearch#Full text search, the autosuggester chose it because "Plowback retained earnings" has more links going to itself than does "Plowback"CpiralCpiral 01:14, 12 February 2016 (UTC)

Special:BookSources

Where could I find the source code that generates the ISBN input box used for Special:BookSources? I expect it's in MediaWiki somewhere but I can't find exactly where. Nikkimaria (talk) 00:41, 11 February 2016 (UTC)

As far as I know it is impossible to re-create this box on Wikipedia in wiki markup. Maybe one of our MediaWiki gurus can confirm this. Some pages on Special:Specialpages can be transcluded, such as Special:AllPages, but Special:BookSources doesn't seem to be allowed. It is probably possible to add Special:BookSources to the $wgAllowSpecialInclusion-list. Or is your goal to use this box outside of Wikipedia? The Quixotic Potato (talk) 01:41, 11 February 2016 (UTC)
No, within. Can something similar be done with wiki markup? Nikkimaria (talk) 02:08, 11 February 2016 (UTC)
Okay, this is super hacky, but try this:

Earwig talk 03:20, 11 February 2016 (UTC)
Another possibility would be:
Type the ISBN in the box below, after Special:BookSources/
- Evad37 [talk] 03:27, 11 February 2016 (UTC)
@The Earwig, Evad37, and The Quixotic Potato: Thanks. There are two reasons I was interested in this, and I don't think either would be served by these solutions
  1. IMO, the current version of Wikipedia:Book sources presents usability challenges, and I was considering ways of addressing that - one idea would be to split out the continents into different subpages. Thus, a user gets to the page by clicking on an ISBN, and when they scroll down to the Libraries section they see a header and a prefilled inputbox for each continent, which they can then click through to go wherever is most appropriate for them and still have the ISBN preloaded.
  2. I wondered about the feasibility of creating a similar inputbox-to-prefilled-linklist approach for DOIs, such that when the user does not have access to the canonical copy (eg. it's behind a paywall) they can be linked to different search engines, repositories, etc where a copy might be found. (To be clear, I'm conceiving this as a help page rather than supplanting the current DOI links in citations. See WP:Journal sources, a Wikipedia Library project).
Any ideas on how either of these things might be done? Obviously the former at least would require some more discussion before implementation, but for the moment I just want to see if it's even possible. Nikkimaria (talk) 14:02, 11 February 2016 (UTC)
Here's a hacky way of doing #2: try entering a doi at Wikipedia:Journal sources/sandbox. The Google scholar, Jurn, and articles on Wikipedia links (at least) should work. A new special page similar to Special:BookSources would definitely be better, but something like this could be used in the meantime if the other search urls are fixed (or removed) - Evad37 [talk] 01:25, 12 February 2016 (UTC)