Wikipedia:Village pump (technical)

From Wikipedia, the free encyclopedia
Jump to navigation Jump to search
 Policy Technical Proposals Idea lab Miscellaneous 
The technical section of the village pump is used to discuss technical issues about Wikipedia. Bug reports and feature requests should be made in Phabricator (see how to report a bug). Bugs with security implications should be reported differently (see how to report security bugs).

Newcomers to the technical village pump are encouraged to read these guidelines prior to posting here. If you want to report a JavaScript error, please follow this guideline. Questions about MediaWiki in general should be posted at the MediaWiki support desk.

Frequently asked questions (FAQ) (see also: Wikipedia:FAQ/Technical)
Click "[show]" next to each point to see more details.
If something looks wrong, purge the server's cache, then bypass your browser's cache.
This tends to solve most issues, including improper display of images, user-preferences not loading, and old versions of pages being shown.
Font size changed unexpectedly?
You may have accidentally changed the font size on your browser for a particular website by pressing a shortcut key or scrollwheel without realising it. Try resetting the zoom with Ctrl+0 (typing the digit zero while holding down the control key) or adjusting the zoom with Ctrl++ or Ctrl+-. Alternatively, look for the View option on your browser's menu and reset it to 100%.
No, we will not use JavaScript to set focus on the search box.
This would interfere with usability, accessibility, keyboard navigation and standard forms. See bug 1864. There is an accesskey property on it (default to accesskey="f" in English), and for logged in users there is a gadget available in your preferences.
No, we will not add a spell-checker, or spell-checking bot.
You can use a web browser such as Firefox, which has a spell checker.
If you have problems making your fancy signature work, check Wikipedia:How to fix your signature.
If you changed to another skin and cannot change back, use this link.
Alternatively, you can press Tab until the "Save" button is highlighted, and press Enter. Using Mozilla Firefox also seems to solve the problem.
If an image thumbnail is not showing, try purging its image description page.
If the image is from Wikimedia Commons, you might have to purge there too. If it doesn't work, try again before doing anything else. Some ad blockers, proxies, or firewalls block URLs containing /ad/ or ending in common executable suffixes. This can cause some images or articles to not appear.
Numbers listed in parentheses in the "Recent changes" section, on history pages and in your watchlist are the number of added or removed bytes.
For server or network status, please see Wikimedia Metrics.
« Older discussions, 151, 152, 153, 154, 155, 156, 157, 158, 159, 160, 161, 162, 163, 164, 165, 166, 167, 168, 169, 170, 171


created timestamp[edit]


-TonyTheTiger (T / C / WP:FOUR / WP:CHICAGO / WP:WAWARD) 18:16, 16 January 2019 (UTC)

I created Template:Did you know nominations/151 North Franklin with its first published edit at 03:30, 25 December 2018. However, the scripts that I am running show "Created by TonyTheTiger (talk | contribs) on Mon, Dec 24 2018, 21:30:50 (Central Standard Time)" at the top of the page. I never noticed that this was shown in local time instead of UTC. How can I get this script to make my page show UTC times?-TonyTheTiger (T / C / WP:FOUR / WP:CHICAGO / WP:WAWARD) 03:40, 25 December 2018 (UTC)

@TonyTheTiger: first check your timezone in Special:Preferences#mw-prefsection-rendering - if you want UTC times but have set this to something else change that. If that isn't it, can you provide more information about "the scripts" you are using? — xaosflux Talk 04:09, 25 December 2018 (UTC)
User:TonyTheTiger/vector.js imports User:Eizzen/PageCreator.js. User:Eizzen/PageCreator says how to choose between local time and UTC, but UTC is supposed to be default and it doesn't work to set it with the specified method. I think User:Eizzen/PageCreator.js should change that.useUTC === true to that.options.useUTC === true. . User:Eizzen has not edited since 2017. PrimeHunter (talk) 11:45, 25 December 2018 (UTC)
PrimeHunter, Thx. How can we get this change enacted? Does Eizzen have exclusive powers?-TonyTheTiger (T / C / WP:FOUR / WP:CHICAGO / WP:WAWARD) 13:21, 25 December 2018 (UTC)
Wikipedia:Interface administrators can edit it. Some of them monitor this page. Eizzen might respond to an email, or you could copy the code to your own userspace. The latter wouldn't help other users of the script but there are only around a dozen. PrimeHunter (talk) 15:05, 25 December 2018 (UTC)
TonyTheTiger, I have now modified that page as described. —TheDJ (talkcontribs) 08:53, 28 December 2018 (UTC)
TheDJ, I am now seeing GMT instead of Central Standard Time at the top of my pages. Still not UTC.-TonyTheTiger (T / C / WP:FOUR / WP:CHICAGO / WP:WAWARD) 02:50, 3 January 2019 (UTC)
Xaosflux, would you care to comment here where I think other experts are watching.-TonyTheTiger (T / C / WP:FOUR / WP:CHICAGO / WP:WAWARD) 01:36, 9 January 2019 (UTC)
GMT and UTC are the same for all practical purposes (unless you are navigating a ship or aligning an astronomical telescope) - there's only a fraction of a second in it. --Redrose64 🌹 (talk) 11:26, 9 January 2019 (UTC)
Redrose64 year round?-TonyTheTiger (T / C / WP:FOUR / WP:CHICAGO / WP:WAWARD) 06:22, 15 January 2019 (UTC)
Yes. --Redrose64 🌹 (talk) 19:01, 15 January 2019 (UTC)

Question: extracting lists from Unused templates lists[edit]

I have just discovered this: Wikipedia:Database reports/Unused templates. Wow! I was wondering if there is any way to perform queries of the list to extract entries matching certain parameters? (eg. created before a certain year, have certain phrase or characters in the name, or have had no revisions since creation). I think that would be a useful way to filter the list to have a look more in depth at certain templates.

It goes without saying that one will need to have a closer look at the lists are generated before proposing them all for discussion / deletion (to ensure they truly aren't used in other environments such as modules, or aren't transclusionless etc.), but I think it is a useful place to start. Any ideas how I can achieve this? --Tom (LT) (talk) 01:16, 29 December 2018 (UTC)

@Tom (LT): You should look at WP:Petscan. --Izno (talk) 14:11, 29 December 2018 (UTC)
@Izno thanks --Tom (LT) (talk) 11:41, 6 January 2019 (UTC)
 Question: @Izno. PetScan is for categories. Most of the unused templates only show up on the database report so I can't actually view the categories. IF it can solve my problem (ie let me search through the database report) could you kindly show me how...? With thanks, --Tom (LT) (talk) 01:21, 13 January 2019 (UTC)
Wow, over 75,000 unused templates. It's easy to copy all the text from the first page (5000 entries) and paste that into a spreadsheet, then sort by Latest edit or First edit. I used "paste special" as text and then had to format the date columns as number with 0 decimal digits. That is enough to work on! Template:Veg-stub is on page 16 of the report. It was last edited in July 2005 but it turns out to be a redirect to Template:Vegetable-stub which is used. Getting redirects deleted can be difficult because people will argue that redirects are cheap blah blah. It would be nice to know how many of the first 5000 entries are redirects. Johnuniq (talk) 02:12, 13 January 2019 (UTC)
@Tom (LT): Petscan works on categories, templates, or other selection criteria. When you start it up, it's in the "Categories" form - but there are five other tabs along the top. These are: "Page properties"; "Templates&links"; "Other sources"; "Wikidata" and "Output". All of these (except "Output") can be used instead of "Categories" - or in combination with it. --Redrose64 🌹 (talk) 18:40, 13 January 2019 (UTC)

The totally insane URL I now get for my watchlist[edit]

I'm fine with all the updates and "improvements" made to the personal watchlist page; if I don't want to use those feature, I'm not obliged, and none of them are creating a problem. However, until these recent changes, the URL for anyone's watchlist page was Now, the URL when I go to my watchlist is

Yeah, I've noticed this for a while. And frankly I think it's really bad practice to be putting all that information into an URL; do we really need to have the colours of the buttons in the URL? The precise level of filtering that the individual user has selected? But it's really annoying me a lot now that I'm in the process of switching computers, and updating my most frequently viewed pages. How do I get back to the original, no-frills URL? Is there a preference I can select? Risker (talk) 00:25, 2 January 2019 (UTC)

@Risker: do you also want the original, no-frills watchlist - it comes with the plain URL. (Use non-JavaScript interface / Loads Watchlist without filters search or highlighting functionality) option is preferences can turn off the frills. — xaosflux Talk 00:31, 2 January 2019 (UTC)
@Xaosflux:, thank you. While I support the additional frills in principle, I wasn't actually using any of them for anything, so I'll use the "old" style. You've made my day. Risker (talk) 00:36, 2 January 2019 (UTC)
  • I wish there was a preference called "Don't change a damn thing" that just opted out of all new features. Yeah yeah, get off my lawn... (No disrespect to the developers, I'm just an old dog.) –xenotalk 01:58, 2 January 2019 (UTC)
  • The basic Watchlist link is still Special:Watchlist and afaik always has been; if you want to link from outside Wikipedia, does exactly the same thing. When you get a long list of extra parameters in the URL's query string, that's usually a sign that you've altered one or more of the Watchlist options and clicked Show. The extra parameters will include one or more redundancies (action=submit is totally pointless) since the long-form URL has parameters for most selectable options, and options that are not specified in the long-form URL will mostly default to whatever you have set at Preferences → Watchlist. The only ones without a configurable preference are the "Namespace:" menu, and the "Invert selection" and "Associated namespace" checkboxes. So if you wanted show the last six hours of changes by people other than yourself, you can start at Special:Watchlist and append the query string ?days=0.25&hidemyself=1 which gives - if you also wanted to see only edits to articles, you can append &namespace=0 to that. --Redrose64 🌹 (talk) 10:59, 2 January 2019 (UTC)
  • @Risker: Yeah, this is a tricky one. The thing is that the watchlist view now applies several filters by default, so those are all encapsulated in the query string. (At the risk of contradicting Redrose64, you don't have to have modified any of the options, because the default view is already filtered when the "enhanced" watchlist is enabled.) Even the damaging__likelybad_color=c4 and damaging__verylikelybad_color=c5 query-string components are filters (filters that happen to color their matched entries) — without those items in the query string, there's no highlighting of those items.
(That the highlight color is set in the query string is secondary — their primary purpose in appearing there is to establish the filter, so those components are indeed "needed" if that filter is to be applied. And in fact the color thing is kind of nonsense because (a) it's "magical" — c1 is blue, c2 is green, c3 is yellow, c4 is orange, c5 is red, c6 is apparently an error (not purple!), and other than the trial-and-error method I just used to discover that mapping, how would anyone possibly know?; and (b) apparently its customizability is sort of half-hearted, since if you set either filter to a non-default color, it shows two dots alongside each item — the default color followed by your configured color — indicating that the code doesn't "really" trust the user to customize the coloring.)
But the alternative to having all of those settings in the query string would be to store them in a cookie, in which case you wouldn't be able to bookmark or link differently-configured views. That's the big advantage to using query strings for customizable views, which is why it's often the preferred method for doing so. -- FeRDNYC (talk) 12:52, 9 January 2019 (UTC)

Is there a discussion about how to fix links to[edit]

(Not really a technical question:) HighBeam Research appears to have shut down a couple of weeks ago, and we appear to have something like 15,000 articles that use it as a source. Does anyone here know of a WP discussion that is happening about how to resolve the suddenly dead links in all of these citations? Thanks. – Jonesey95 (talk) 02:45, 2 January 2019 (UTC)

Hey. Normally the entire domain would be marked "dead" in IABot, then whenever it comes across it, would convert to a dead link and add an archive URL. But IABot management interface is not responding well. I will contact Cyberpower678. T212753 -- GreenC 03:13, 2 January 2019 (UTC)
If they are in a cite journal or similar, purge them all. They should never have been added in the first place. Pure preference/promotion of a commercial entity. Headbomb {t · c · p · b} 03:30, 2 January 2019 (UTC)
I agree they are last resort, but sometimes they are the only things available for on-line verification purpose when they provide the first few paragraphs of the source-text for free (example). -- GreenC 03:41, 2 January 2019 (UTC)
I could see keeping them when we have no doi/bibcode/pmid/whatever, but they're equivalent to links to Amazon for books. If you've got an identifier, purge them out. And if the links are dead, they're not much point in keeping them around. They won't even be archived, since they are commercial links that block archiving. Headbomb {t · c · p · b} 00:27, 4 January 2019 (UTC)
There has been some discussion on this topic at WT:Highbeam. —AntiCompositeNumber (talk) 14:57, 3 January 2019 (UTC)
I seem to have got it working and have set a bot job for it under job 3041.
Correction, it was not working on that job, but I have managed to set the base URL to dead, I do not have the domain changer user right on IABot [Username Needed] 15:56, 11 January 2019 (UTC)

Why does MediaWiki insert <p>...</p> inside <blockquote>...</blockquote>?[edit]

{{IPAc-en}} wraps the whole notation in class="nowrap" and uses <wbr /> to control line breaking. I recently noticed that while

Lorem {{IPAc-en|ˈ|l|ɔːr|ə|m|_|ˈ|ɪ|p|s|əm}} ipsum

results in "Lorem /ˈlɔːrəm ˈɪpsəm/ ipsum",

<blockquote>Lorem {{IPAc-en|ˈ|l|ɔːr|ə|m|_|ˈ|ɪ|p|s|əm}} ipsum</blockquote>

results in


/ˈlɔːrəm ˈɪpsəm/


Apparently MediaWiki automatically inserts <p>...</p> inside <blockquote>...</blockquote>, which generates these line breaks. Is this an intended behavior on MediaWiki's part? Nardog (talk) 12:07, 2 January 2019 (UTC)

Probably some weird exception in the mediawiki parser. The parsoid parser does it correctly. Regardless, the use of nowrap is terrible in many situations and should be avoided. I think we should consider replacing it with display:inline-block instead of nowrap. That actually does what people want it to do (put it on its own line if it won't fit on the current line), without breaking the internal line breaking where it is needed if the block doesn't fit on a line of its own. —TheDJ (talkcontribs) 14:55, 2 January 2019 (UTC)
Apparently <p>...</p> is required in XHTML so that seems to be the reason.
@TheDJ: Can you elaborate on what you mean by "replacing it with display:inline-block instead of nowrap"? White space handling is a nightmare with IPA because ˈ ˌ / - ( etc. trigger line breaks when they shouldn't, and putting the whole notation in nowrap and appending wbr to spaces is about the best solution I know. Does display:inline-block provide a better solution to it? Nardog (talk) 15:08, 2 January 2019 (UTC)
Nardog, yes I think it does. The problem with nowrap is that is ALWAYS prevents line breaks, no matter if this breaks display in other situations. This is an issue on mobile where, because people have slapped this thing on entire phrases, it will cause half the phrase to be outside of the screenspace. Using display:inline-block however makes a phrase behave in a different way that is probably more suitable. It will render on a separate line whenever it detects that the current line doesn't have 'enough' space. This thus avoids line breaks within that phrase, up until the point where the width of that phrase is as wide as an entire line, where you DO want normal line breaking to take place, to make sure the phrase is at least visible. —TheDJ (talkcontribs) 15:22, 2 January 2019 (UTC)
@TheDJ: But, again, with ˈ ˌ / - ( etc. we don't want normal line breaking to take place. For example, in "/prəˌnʌnsiˈeɪʃən, -ˌnaʊn-/", /, ˌ, ˈ, and - all trigger a break. But we want it to break only at the space. Nardog (talk) 15:44, 2 January 2019 (UTC)
@Nardog and TheDJ: Here's the interesting thing I just discovered, though: The problem isn't that MediaWiki inserts <p>...</p>, but that it inserts two pairs of <p>...</p>, one around "Lorem" and one around "ipsum". (It's really unclear to me why it's doing it that way — there are no newlines in the generated output once the Lua module is executed, just a sh*t-ton of <span>s.) However, it only does that if there isn't already a paragraph defined. IOW, this:
<blockquote><p>Lorem {{IPAc-en|ˈ|l|ɔːr|ə|m|_|ˈ|ɪ|p|s|əm}} ipsum</p></blockquote>
produces this:

Lorem /ˈlɔːrəm ˈɪpsəm/ ipsum

(Also, second observation of weird behavior: In the "Show preview" view of the Visual Editor, Nardog's second example does not render with added line breaks.) -- FeRDNYC (talk) 11:18, 9 January 2019 (UTC)

Pending changes accepted or not[edit]

Would anyone like to comment on what is happening here?

On my watch list diff, I see that an addition has been made to the list. (Say Iran-Israel-Italy) It says in bold [accepted addition]. I opt to edit the source code- and two items are missing Iran-Israel. Is this a configuation issue or a feature? --ClemRutter (talk) 10:04, 3 January 2019 (UTC)

Not sure how you did it. Open the source code of the revision and search for the items using ⌘ Command+F or Ctrl+F depending on what you're using. They're indeed there. –Ammarpad (talk) 15:24, 3 January 2019 (UTC)
Not sure how I did it either. But I went back to see- and unfortunately- my configuration will not show it still.
Here is a screen dump. It looks like a tricky one- above my pay grade. I have tried it on two computers Linux Mint 17. 3 and 18.1. (Firefox) I cannot see Iran or Israel on either. The Israel entry was edited twice in the meantime by John Cline so it must be something in my WP profile that is conflicting with the additions pending mechanism. The plot thickens? ClemRutter (talk) 21:12, 3 January 2019 (UTC)

I apologize ClemRutter, if my actions compounded the difficulty of your endeavor. I inadvertently rolled back the page's then "currently accepted version" and subsequently undid the resulting changes. Interestingly, if not for the error you'd not have mentioning me, and I'd not have this opportunity to answer your original question (I hope I am up to the task).

It appears to me that the forces of nature have conspired (for their own amusement) to choreograph some technical chicanery at our expense; tricking us by the combined effect of a truncated diff and the article's rendered output where both represent the page's "currently accepted version" (or, better yet, misrepresent it).

Upon looking closer, you'll notice that the changes, reflected in the diff, were actually added inside an HTML comment where the beginning is marked by <!--, the end by -->, and everything between is hidden except when viewing the page's source code. If that's not enough, the entire visible list is generated by the transclusion of {{List of names for secondary schools by country}}; proving that things do not always exist as they seem (at least not in this case).

I hope this has been helpful in answering your question; it was a challenge for me, and a joy. If it has not helped, follow on where confusion remains and I'll try to help further. If it has, happy editing and best regards.--John Cline (talk) 22:07, 3 January 2019 (UTC)

I am curious ClemRutter, did the comments given directly above answer your question; resolving it completely or not? In fact, I am keenly anxious to know. Thanks.--John Cline (talk) 11:49, 8 January 2019 (UTC)
Oh yes. Sorry for not saying.ClemRutter (talk) 13:59, 8 January 2019 (UTC)

──────────── @ClemRutter: The reason you got tricked is because, as John Cline says, the text being edited was commented out. Specifically, someone wrapped that entire list in an HTML comment reading:

<!-- maintained so the lists can be checked- and corrections made to the template- when done this should be deleted
IMPORTANT: if you edit the below, there will be no change to article text, you must also change the template

Since that list never was deleted, there were significant differences between the template and the source text. (Naturally, since that comment is far from sufficient to warn people off, especially given the length of the list.) I've copied the items that were added to the commented-out article source (including Iran) over to the template, rearranged a few that were out of alphabetical order, and deleted the giant commented-out list at Secondary school to avoid confusing people in the future. -- FeRDNYC (talk) 12:01, 9 January 2019 (UTC)

@ClemRutter: And all that being said, I'm now amused by the discovery that you were apparently the one who created Template:List of names for secondary schools by country in the first place (and presumably the one who wrote that comment, as well?) 😃 -- FeRDNYC (talk) 12:04, 9 January 2019 (UTC)
Yep, it was Clem. --Redrose64 🌹 (talk) 12:11, 9 January 2019 (UTC)

Unexplained errors in Javascript console[edit]

The Navigation Popups gadget created by Lupin is spewing tons and tons of errors on my console.Could somebody look into what is causing it ? — fr 12:15, 3 January 2019 (UTC)

Unexplained Javascript errors.png
@FR30799386: I have Navigation Popups enabled on my account, and I'm not getting those errors. So either the issue has been corrected since, or there's something unique to your environment that's causing them. -- FeRDNYC (talk) 11:24, 9 January 2019 (UTC)
I get these errors occasionally. The errors look very weird, and I haven't been able to reliably reproduce it. I will investigate further once I get them again. Enterprisey (talk!) 20:23, 9 January 2019 (UTC)

Combating the issue of bare URLs[edit]

Hello all,

Apologies for putting this here, I don't know really where else to put it besides Wikipedia talk:Bare URLs where it certainly wouldn't get the input it deserves.

Fetterly et al (2003, DOI: 10.1002/spe.577) noted that 40% of webpages change within a week and that one in two-hundred links disappear from the Internet in the same time period. McCown et al (2005, arXiv:cs/0511077) noted that 28% of URLs failed to respsond in a study on D-Lib Magazine. Needless to say, bare URLs are bad news for Wikipedia, because they are prone to link rot which will compromise verifiability. Despite only five articles appearing in the link rot maintenance category, the problem is much more severe. A RegEx search I have run on the latest enwiki database dump reveals that 489,767 articles, approximately 8.5% of all enwiki articles, use bare URLs. That means that every week, on average, just shy of 2,500 articles will have at least one reference turned into a dead link and we will have no contextual information to rescue the sources.

I have laid out a plan here, however I don't believe reFill has the functionality of batch processing, which rather snookers the plan. I've got the list of the 489,767 articles but I sure as hell can't use reFill on every single one of them. Does anyone have any ideas on how to go about this?

Many thanks,

SITH (talk) 23:52, 3 January 2019 (UTC)

Per WP:Link rot:
All new links added to Wikipedia are automatically saved to Wayback Machine within about 24hrs. This is done with a program called "NoMore404" which Internet Archive runs and maintains. It scans the IRC feed channels, extracts new external URLs and adds a snapshot to the Wayback. This system became active sometime after 2015, though previous efforts were also made. etc..
Any links added after 2015 there is a very-high chance of a Wayback link being available. For earlier links there is also a very-good chance due to previous archiving efforts by Wayback, and other providers. Still, some links which went dead a long time ago have no archives, and some links the providers can't or won't archive for some reason, link rot remains to some degree. -- GreenC 00:18, 4 January 2019 (UTC)
@GreenC: yep, but going through the AWB list, most of these articles are pre-2015 unfortunately. Perhaps it's a legacy issue with NoMore404, maybe we could link up a bot to reFill to clear out the backlog? SITH (talk) 00:49, 4 January 2019 (UTC)
reFill is one tool that works on bare links. There is also User:Citation bot. And IABot converts them (if the link is dead). I don't know what that error rate is for these tools, if they can safely be automated without user oversight. IABOt is fully auto, but when a link is dead the conversion isn't much to do. -- GreenC 01:03, 4 January 2019 (UTC)
One idea is extract a few thousand bare URLs at random into a test page (each surrounded by <ref></ref>). Make three copies of that test page. Run reFill on one, Citation bot on another, and IABot on the third. See what the results are - look for problems and differences. The quality of the edits will give a sense how viable it would be to run in fully automated mode. -- GreenC 01:32, 4 January 2019 (UTC)
There is also WP:Reflinks, which seems to correct a wider range than refill.Onel5969 TT me 01:38, 4 January 2019 (UTC)
One could presumably also build a bot to interface with mw:Citoid. --Izno (talk) 03:06, 4 January 2019 (UTC)
@Izno: reFill appears to be forkable on GitHub and doesn't depend on labs hosting so I could fork it and integrate it into AWB. I'd probably have to go to WP:BRFA first but that looks like the most feasible idea. — Preceding unsigned comment added by StraussInTheHouse (talkcontribs) 13:33, 4 January 2019 (UTC)
Whomever the nameless ping was from, you'll also need to check that reFill produces citations conforming to our current citation module. One or another of the gadgets does not presently. The reason I mentioned Citoid (which is similarly open source) was because it also has integrated a much-better specific site integration, I would guess. --Izno (talk) 14:44, 4 January 2019 (UTC)
reFill uses Citoid to generate citations. --AntiCompositeNumber (talk) 15:51, 4 January 2019 (UTC)
Oh, that's cool. I wasn't aware that it was using citoid. --Izno (talk) 18:30, 4 January 2019 (UTC)
Bears repeating - these tools are not run and forget. They make mistakes, quite frequently. They are designed as a starting point with further manual edits to fill in and fix. You are responsible for every edit. Focus on getting it right, the mistakes these tools makes are far more work and time to find and fix after the fact. That's why I recommended, run tests to see which tool does the best job, and see how many mistakes they are actually making. It may be we don't have any tools reliable enough for an unattended bot run. -- GreenC 15:42, 4 January 2019 (UTC)
@StraussInTheHouse: The current tool isn't very flexible in this regard, but a rewritten version under development (I'm planning to take it back from the ashes after my hiatus) provides a set of APIs that will allow for easy integration with bots. It's nowhere near completion, but the new version has much less cruft compared to the existing PHP version, and is where more development effort should go. Zhaofeng Li talk (Please {{Ping}} when replying) 15:21, 8 January 2019 (UTC)
Zhaofeng Li, that looks good, let me know when you’re fully back and I’ll help all I can! SITH (talk) 18:11, 8 January 2019 (UTC)
@StraussInTheHouse: I've got the new version back in shape in the past week, and it should be ready for public testing at this point. Zhaofeng Li talk (Please {{Ping}} when replying) 05:00, 16 January 2019 (UTC)
@GreenC: test cases on draftified versions would be the best bet. SITH (talk) 16:20, 4 January 2019 (UTC)
Yeah create a page in userspace with the URLs and run the tool on it. When done delete the page. It's userspace can do whatever you want. -- GreenC 16:25, 4 January 2019 (UTC)

Suggestion to update Reference Tooltips gadget[edit]

Reference Tooltips reworked.png

Hello! I would like to suggest to update to the new version of this gadget I've presented at mw:Topic:Ueqlcc482l9yw8gv. There are numerous bugfixes and added features like Harvard-style citations support. Tooltip style & animations are also updated to be consistent with Page Previews' style & animations. It has been tested thoroughly, including by the author of the original script, and used in Russian Wikipedia for several months with no complaints.

Jack who built the house (talk) 13:00, 4 January 2019 (UTC)

Jack who built the house, I think that this is good to go;-) Will ask at IAN. — Preceding unsigned comment added by Winged Blades of Godric (talkcontribs) 11:29, 10 January 2019 (UTC)

making plural links singular[edit]

For group article names, such as [[Christians]] it's unfortunate to have to use a pipe to get the singular version. For instance I have to say "He was a [[Christians|Christian]]". Going from singular to plural is easy, i.e., [[Christian]]s and it would be nice to be able to use such a shortcut in reverse. Something like [[Christian<s>]] or some such.Coastside (talk) 21:15, 4 January 2019 (UTC)

A similar request was declined at phab:T5527. PrimeHunter (talk) 21:28, 4 January 2019 (UTC)
@Coastside: See {{sgl}}. --Ahecht (TALK
) 22:30, 4 January 2019 (UTC)
@Ahecht: I didn't realize {{sgl}} was created as a response to this request. I thought it was an established template that I didn't know about. PrimeHunter raised concerns about readability for editors and tools. Maybe we shouldn't use this? I can change the few links I created that way. Thoughts?Coastside (talk) 16:50, 6 January 2019 (UTC)
@Coastside: The template is now substable, and when substed will look idential to a traditionally-created link (similar to how the pipe trick works). I flagged the template so that any transclusions will be automatically substed by a bot. --Ahecht (TALK
) 16:18, 7 January 2019 (UTC)
I find it less confusing to put the brackets around the name even if it is plural (e.g. I would use both [[Christian]] and [[Christians]]), and let the link redirect if necessary. Why complicate things? —Remember the dot (talk) 05:42, 8 January 2019 (UTC)
That makes sense in general. Unfortunately Christian is a disambiguation page. I think that's likely to be the case for a lot of pages on groups of people. (Whether Christians should be a page on a group of people, as opposed to say a redirect to Christianity, is a separate question.) --Trovatore (talk) 06:35, 8 January 2019 (UTC)
Good point. I should have realized that Christian is a disambiguation page and Christians is not. —Remember the dot (talk) 02:10, 10 January 2019 (UTC)

Articles added/removed from categories don't show up on watchlist anymore[edit]

So, my problem is that when I have a category page on my watchlist, following markings either don't ever show up at all or they show up with a few days of delay:

Category:X; 11:05 . . ‎User1 (talk | contribs)‎ (ArticleA removed from category)
Category:Y; 11:04 . . User2 (talk | contribs)‎ (ArticleB added to category)

My watchlist settings' "Hide categorization of pages" is not on. This feature has been very useful to me for a long time, but now it hasn't worked properly for a month or so. --Kliituu (talk) 19:06, 7 January 2019 (UTC)

I notice the same bug occur in every wiki.--3knolls (talk) 10:22, 8 January 2019 (UTC)
@3knolls and Kliituu: See T212432. –Ammarpad (talk) 10:27, 8 January 2019 (UTC)

reFill is looking for a maintainer[edit]

As you may have noticed, the reFill tool that I created has not been maintained for a while, and the unfixed bugs have led to many faulty edits that have wasted the time of people around here. There are several key problems with the current tool:

  1. An old PHP codebase with a regex-based wikitext parser, the source of many problems reported
  2. The lack of a feedback system built into the tool that will facilitate reporting bad suggested changes
  3. The absence of a maintainer that will improve the tool and help resolve issues when I'm away

For #1, there was effort to make a complete rewrite of the tool in Python, with APIs that may be used by other tools and bots. Unfortunately it was not sufficiently publicized, and few people have tested this version. I believe effort should be spent on continuing this work, bring it in parity with the current tool in terms of functionality, possibly incorporating #2.

For #3, I'm looking for a maintainer that will help fix issues reported by the community, and assist in developing the tool. As mentioned, I believe more time should be spent on developing the rewrite in #1, which will provide the community with a more stable and maintainable tool in the long term.

Last but not least, I would like to apologize for my disappearance. I left Wikipedia a year ago due to personal reasons, and I realize my lack of communication was deeply irresponsible on my part. I eventually got over my issues, but coming back has become increasingly difficult after a long hiatus. I'd like to thank User:Smuckola who pinged me on IRC and helped me make up my mind to come here and post this. Zhaofeng Li talk (Please {{Ping}} when replying) 03:21, 8 January 2019 (UTC)

@Zhaofeng Li: Z is a model Wikipedian with a long history of providing reliable engineering and support, which has overall saved far more time and effort than not. This project is a successful community pillar. Hopefully this can be somehow unified with VisualEditor, Citoid, and the WikiCite project to become everything for everyone. Whatever we use, nothing is perfect so it's on each of us to preview our work! — Smuckola(talk) 11:09, 8 January 2019 (UTC)

Page view stats per country?[edit]

The Page View Stats API includes "Get pageviews by country" (under "Pageviews data"). This works at the project level (eg. show all from Canada reading Enwiki). However there is no apparent way to get country data at a more granular level (eg. show all from Canada reading the pop music article). Is there a way to get country-specific data on a per article basis? -- GreenC 14:40, 8 January 2019 (UTC)

@GreenC: For privacy reasons this is not publicly available. It may seem silly since we're only reporting the country, but it is what it is. Say there is only one editor to the page, and only one country reported in the pageviews. Now you know what country they're located in. MusikAnimal talk 18:48, 8 January 2019 (UTC)
@MusikAnimal: OK, thank you. -- GreenC 19:01, 8 January 2019 (UTC)
I wonder if that's a real good reason though. We have a similar rationale for pages with fewer than 30 watchers, but you could just as easily (for low-visit pages) say "results unavailable due to privacy reasons". --Izno (talk) 19:36, 8 January 2019 (UTC)
Yeah, I agree there could be some logic to not expose country data if there's only so many pageviews to begin with, etc. I'm pretty sure it's a no-go, though. For privacy reasons, the current project-level country pageviews doesn't even give you exact numbers, rather they are ranges like "Spain: 5,000 - 10,000 pageviews".

There are probably also storage concerns. The normal pageviews API is quite large as it is, then you add in slots for the major countries (perhaps up to all 190+), times the number of articles, possibly for each platform (desktop/mobile) and agent (user/bot), and finally for each project. That's huge!

I didn't find a task for per-article country views on Phabricator, if you wanted to create one. Maybe the Analytics team could find a happy medium. MusikAnimal talk 20:27, 8 January 2019 (UTC)

I'll try, thanks! -- GreenC 21:26, 8 January 2019 (UTC)

Replacing coordinates[edit]

I would like to know the method in AWB by which I can replace the values of latd and longd to the current format of coord.Adithyak1997 (talk) 15:42, 8 January 2019 (UTC)

It depends on where the latd and longd appear. Can you please link to articles that you would like to convert? The Wikipedia:Coordinates in infoboxes project converted coordinates in infoboxes in about one million pages. Did we miss some? – Jonesey95 (talk) 20:20, 8 January 2019 (UTC)

Full template expansion for debugging[edit]

I was recently trying to track down some weirdness with a particular template's behavior. I eventually figured it out, but it would have been a lot easier if I could have just seen the ultimate markup that would have been produced once all templates and #invokes and such were processed, and I mean recursively here. Are there any tools that can help with this? Thanks, –Deacon Vorbis (carbon • videos) 16:09, 8 January 2019 (UTC)

@Deacon Vorbis: have you tried Special:ExpandTemplates? — xaosflux Talk 16:38, 8 January 2019 (UTC)
That seems to be exactly what I was looking for, thanks! –Deacon Vorbis (carbon • videos) 16:40, 8 January 2019 (UTC)

pipe trick's don't work in website param of citeweb template when included in references[edit]

Pipe tricks don't work in the citeweb parameter of the {{citeweb}} template when the template is included in a citation references.

This works:

{{cite web |url=|title=pipe tricks don't work in refs|website=[[Fortune (magazine)|Fortune]]|date=1 March 2016|accessdate=7 January 2019}}

"pipe tricks don't work in refs". Fortune. 1 March 2016. Retrieved 7 January 2019.

But this doesn't work (check the reference, i.e., the footnote):

<ref>{{cite web |url=|title=pipe tricks don't work in refs|website=[[Fortune (magazine)|]]|date=1 March 2016|accessdate=7 January 2019}}</ref>Check out "Fortune" in the footnote

[1]Check out "Fortune" in the footnote

Coastside (talk) 22:42, 8 January 2019 (UTC)

An ancient issue; phab:T4700 is the ticket. --Izno (talk) 22:47, 8 January 2019 (UTC)
It's nothing to do with {{cite web}}. The problem is with the <ref>...</ref> tags.. --Redrose64 🌹 (talk) 23:58, 8 January 2019 (UTC)
Yes. Compare "Fortune" with the identical wikilink in note [2]. ♦ J. Johnson (JJ) (talk) 00:30, 9 January 2019 (UTC)


  1. ^ "pipe tricks don't work in refs". [[Fortune (magazine)|]]. 1 March 2016. Retrieved 7 January 2019.
  2. ^ [[Fortune (magazine)|]] -- wikilink only, no template.

edit summary sections are off[edit]

The edit summaries of some sections for me come up as "Example{{anchor|Example}}" as opposed to simply "Example". Any explanation as to what happened? [Username Needed] 11:25, 9 January 2019 (UTC)

Markup in section headings usually breaks links to that section. --Redrose64 🌹 (talk) 12:03, 9 January 2019 (UTC)

References section in the Wikipedia's articles: adding "Show context" lable like here[edit]

Good evening, please let You look at the use of references in the IEEE website. By expanding the label "Show context", in the ending footnotes users have a preview of the paragraph(s) in which the citation is positioned.

Il would be useful if users can view the same "Show context" for the "References" section of the Wikipedia's articles.

If citation is called more times in the article by the <ref name =" " /> , the "Show context" join the multiple ranges of row line numbers where the citation it is used, limited by a fixed maxmimum leght.

Context can be visualized in:

  • an ony-text mode,
  • hypertexual mode: with the wikilinks and eventual external links. Wikilinks can also conserve the related preview, as visualized in the article.

Images and files coming from Wikimedia Commons aren't visualized in the "Show context".

Once the article is updated, the set "Show context" is dinamically synchronized to the last edit saved for the article.

Furthermore, WP editors can be enabled to associate a unique and optional parameter for the context to any given citaton template. This is waht just happens through the less used template:clarify (e.g. the clarify on it.wikipedia)- — Preceding unsigned comment added by (talk) 22:37, 9 January 2019 (UTC)

When I go to I get a "References" heading, below which are a number of blank spaces each beginning with a dot. If I scroll up the page, there is a blue box saying "Sign in to Continue Reading". Not very helpful. --Redrose64 🌹 (talk) 23:27, 9 January 2019 (UTC)
Hi. I thank for your reply. I noticed this problem now, since yesterday it was not. I am connecting with a laptop, with Windows 8.0 version 64 bit, and Google Chrome. I have closed and opened again the heading "References" and the list works rightly and it is full displayed. It may be read also by non signed users. Example:
  1. Reference n.1 (the first reference): "S. Behnke. Hierarchical Neural Networks for Image Interpretation, volume 2766 of Lecture Notes in Computer Science. Springer, 2003. 1, 2".

which is followed by three links: "Show context", "Crossref", "Google Scholar".

    • the first one (named "Show context"), if expanded, is nowm displaying the string: "We focus on deep convolutional neural networks (DNN), introduced by [11], improved by [19], refined and simplified by [1], [32], [7]." Under the text there is an unique link named "Go to the text", that can be read only for signed users.
  1. Reference n 37 (the last): "D. H. Wiesel and T. N. Hubel. Receptive fields of single neurones in the cat's striate cortex. J. Physiol., 148:574-591, 1959. 2".

The related link "Show context", if expanded below, displays the following: "The approach is inspired by Hubel and Wiesel's seminal work on the eat's primary visual cortex [37], which identified orientation selective simple cells with overlapping local receptive fields and complex cells performing down-sampling-like operations [15]."

I would like to suggest an adaptation of this feature for Wikipedia's articles, that:

  • hasn't the need for users to be signed.
  • can be web archived with the list of citations (otherwise that [ in the example).

I mean that in the section "References", readers (yet signed or not) have a link named "Show context" for each number of reference, that it can be expanded, displaying the last paragraph. One possible way to identify this paragraph is like in the IEEE website "from dot to dot":

  • starting from the first dot before the tag <ref>, that originates the citation in the article.
  • ending to the first dot after tha tag </ref> , that ends the citation in the article.

For the longest periods, it can be fixed a maximum possible lenght of four (or more) lines of text, assigned to the event named "Show context", so as to have displayed the current line where is the tag <ref> in the article, adding the previous two row lines and the following one.

Maybe this feature will be hopefully introduced in Wikipedia without the need to be signed. — Preceding unsigned comment added by (talk) 14:42, 10 January 2019 (UTC)

Template transclusion including partial if statements in articles[edit]


Can someone fix {{Mergeto}} and {{Mergefrom}}? They are including a new line and "{{#if:List of TurboGrafx-16 games||", see (e.g.) List of PC Engine games and List of TurboGrafx-16 games. Thanks. ―Justin (koavf)TCM 01:55, 10 January 2019 (UTC)

See Template talk:Merge#Broken. — JJMC89(T·C) 05:07, 10 January 2019 (UTC)

Artist templates need mending[edit]

Hi, and requesting help from someone on this. Two templates, {{Evelyn De Morgan}} and {{Daniel Chester French}}, are missing their "view, discuss, edit" links at the top left. Can't figure it out, and user Another Believer suggested bringing it here. Am wondering if someone can reveal the nonexistent buttons. Thanks. Randy Kryn (talk) 14:26, 10 January 2019 (UTC)

 Done by Another Believer. Randy Kryn (talk) 14:37, 10 January 2019 (UTC)

Broken links to peer review pages.[edit]

Apparently, on occasion, an article and its subpages are moved to a new title. However, peer review pages existing in project space are not captured when subpages are moved. I became aware of this while participating in the mass-move of political campaign articles to put the dates before the name of the respective campaign or election, but it is likely that it has occurred in other contexts. There are two ways to address this, those being to move all the affected peer review pages to the new titles, or to create redirects to them. Either one would require that someone compile a list of the links that are now broken. If someone makes such a list, I'll be glad to check and move the pages. Cheers! bd2412 T 17:43, 10 January 2019 (UTC)

@BD2412: most articles moves leave a redirect behind, and articles shouldn't have subpages. Can you show some examples of this problem? — xaosflux Talk 18:20, 10 January 2019 (UTC)
To be more precise, talk pages usually move with the article, and talk pages often have subpages (typically archives). The case that was brought to my attention was from my move of William McKinley presidential campaign, 1896 to William McKinley 1896 presidential campaign, pursuant to consensus for the above-referenced mass-move of political campaign articles. It was later brought to my attention that this resulted in a broken link on the talk page of that article to Wikipedia:Peer review/William McKinley presidential campaign, 1896/archive1, which I thereafter moved to Wikipedia:Peer review/William McKinley 1896 presidential campaign/archive1. It is likely that there are other subpages of Wikipedia:Peer review that have become untethered due to page moves, since these are not in a space that is carried along with the page move. bd2412 T 18:32, 10 January 2019 (UTC)
Simplez, use an ifexist in {{Old peer review}} to categorize talk pages with nonexistant links. Galobtter (pingó mió) 18:40, 10 January 2019 (UTC)
If someone who knows how to do that would create the category, I can do the rest. Cheers! bd2412 T 19:24, 10 January 2019 (UTC)
I have added ifexist code [1] and created Category:Pages using Template:Old peer review with broken archive link. It will take time for the job queue to automatically populate the category. PrimeHunter (talk) 22:06, 10 January 2019 (UTC)
Thanks, this works. I will keep an eye on it. bd2412 T 22:27, 10 January 2019 (UTC)
@BD2412, PrimeHunter, and Galobtter: How about pages that use {{article history}} to display peer reviews but not the {{Old peer review}} template, like Talk:Charles Boycott? (Not that the link there is broken ... just using it as an example. Graham87 10:30, 11 January 2019 (UTC)
Graham87, {{article history}} does not automatically generate the link as {{old peer review}} does (you have to specify the link as e.g |action1link=Wikipedia:Peer review/Charles Boycott/archive1}}), so links don't break. Galobtter (pingó mió) 10:34, 11 January 2019 (UTC)
@Galobtter: Oh of course, oops! Graham87 11:45, 11 January 2019 (UTC)
Note that this has generated a list of 642 errors, which is a bit bigger of a task than anticipated. In other words, a technical solution may be required to fix these. bd2412 T 13:40, 11 January 2019 (UTC)
You could ask for bot help at Wikipedia:Bot requests, or manual help at Wikipedia talk:Peer review. I'm not a bot coder. The best I could do is modify {{Old peer review}} to check a list of exceptions if the normal link would be red. A list entry in a big switch could look like this: |William McKinley 1896 presidential campaign = Wikipedia:Peer review/William McKinley presidential campaign, 1896/archive1. Peer reviews can be in other places like Wikipedia:WikiProject Military history/Peer review/Murray Bourchier so we shouldn't assume a subpage of Wikipedia:Peer review. The list would still have to be compiled somehow but it might be easier than editing each talk page, or make moves or redirects. If somebody makes a script to compile a list then it wouldn't require bot approval like solutions which make 642 edits. PrimeHunter (talk) 19:04, 11 January 2019 (UTC)
Can someone make a list of the untethered peer review pages that correspond to the ~640 (I fixed a few) broken links? bd2412 T 22:22, 13 January 2019 (UTC)

Hitting "enter" in the edit summary box no longer saves the page[edit]

When I am in the edit summary box and hit "enter" rather than saving the page this action now brings up a menu of common edit summaries, even if I have already typed my edit summary. I have no use for this feature. Can I at least get the previous effect of saving the page as an option? bd2412 T 21:15, 10 January 2019 (UTC)

Testing... It saved it for me. What skin are you using? Did you change anything in your preferences or gadgets? ~ ONUnicorn(Talk|Contribs)problem solving 21:28, 10 January 2019 (UTC)
  • I am facing the same problem as User:BD2412 and it started a few hours back, so something recent has changed it. Now the only way to save a comment is to use the mouse and hit the publish button. I have not made any changes and would like the old functionality back. thanks.--DBigXray 21:31, 10 January 2019 (UTC)
Wikipedia:ITSTHURSDAY? Works fine for me, is there a chance your browser updated recently? ~ Amory (utc) 21:40, 10 January 2019 (UTC)
This is happening to me, sometime today (no browser restart or the like). "Enter" on the Edit Summary triggers the drop down of "common edit summaries" --Masem (t) 21:54, 10 January 2019 (UTC)
  • @BD2412, DBigXray, and Masem: At Preferences → Gadgets, what is your setting for "Add two new dropdown boxes below the edit summary box with some useful default summaries"? --Redrose64 🌹 (talk) 22:01, 10 January 2019 (UTC)
    It was enabled. Disabling it removes this above stated behavior. --Masem (t) 22:07, 10 January 2019 (UTC)
    Same. The issues is solved, but I note that I did not make any recent changes to preferences to cause the change. My computer did install updates within the past few days, but the behavior noted above did not manifest until earlier this afternoon. bd2412 T 22:17, 10 January 2019 (UTC)
    @Redrose64: it was selected from a long time, and still is selected (same as Masem and BD24). Unselecting the option, removed the 2 boxes of default summary. While deselecting the option solves the problem but I would still prefer to be able to use the option of hitting the enter key at the edit summary box, "while having" these 2 default edit summary boxes at the same time. Like it was before. This problem started in the last 2-3 hours so I suspect some change from the wikimedia server side triggered it. if someone could restore this back, or if there is another way to fix this, it will be appreciated. --DBigXray 22:26, 10 January 2019 (UTC)
    There are no recent changes in MediaWiki:Gadget-defaultsummaries.js --Redrose64 🌹 (talk) 23:24, 10 January 2019 (UTC)
    Something has definitely changed, though, as I had the extra boxes enabled for ages and the "unable to press Enter to submit" issue only started today. Black Kite (talk) 00:48, 11 January 2019 (UTC)
"Fails" for me as well in Chrome and FF (if the gadget is enabled). — xaosflux Talk 22:39, 10 January 2019 (UTC)
  • Additional problem - the same issue arises in the [Reason] field when trying to move a page. The fix that solved this problem for the regular edit box does not appear to be available for this one. bd2412 T 23:35, 10 January 2019 (UTC)

I've just started to notice that clicking edit doesn't save edits anymore... It used to be that if I was in the Edit Summary panel and I hit enter, the page would save. Now, doing that causes the "Common edit summaries" dropdown to open. Has anyone else seen this? --Zackmann (Talk to me/What I been doing) 23:02, 10 January 2019 (UTC)

Zackmann08, I have, and it's annoying me to no end. These sort of "updates" are completely unnecessary. -- /Alex/21 00:37, 11 January 2019 (UTC)
FWIW, I am not having this issue on testwiki or test2wiki. — xaosflux Talk 01:02, 11 January 2019 (UTC)
And it just stopped happening here.... — xaosflux Talk 01:03, 11 January 2019 (UTC)
@Alex 21 and DBigXray: are you still having the issue? — xaosflux Talk 01:04, 11 January 2019 (UTC)
Never mind, with gadget enabled it is still failing here. — xaosflux Talk 01:05, 11 January 2019 (UTC)
@Alex 21 and DBigXray: FWIW once I disabled the 2 fields, problem solved. :-) --Zackmann (Talk to me/What I been doing) 04:19, 11 January 2019 (UTC)
@Zackmann08:, have you tried doing a page move? For me, disabling the fields does not seem to fix the issue for the edit summary for that action. bd2412 T 13:41, 11 January 2019 (UTC)
  • If any intadmin (xaosflux?) wants to put a quick patch in to restore this functionality while we try to figure out what went wrong, they can copy over the content of User:Enterprisey/Gadget-defaultsummaries.js and then G6 it. Not a very good fix (it's just another keydown handler to check for enter presses, and then trigger the save button), of course, but this issue seems to be disrupting some workflows. Enterprisey (talk!) 06:43, 11 January 2019 (UTC)
  • Gonna go ahead and make a wild guess that the OOUI update to v0.30.0 that was rolled out 20:00–22:00 UTC as part of the normal update to 1.33.0-wmf.12 (all from wikitech:Deployments#Thursday,_January_10 and mw:MediaWiki_1.33/wmf.12. The underlying issue might not exactly be trivial to find. Edit request filed. Enterprisey (talk!) —Preceding undated comment added 07:28, 11 January 2019 (UTC)
    @Enterprisey: I promoted your code to the gadget, this does seem to be bigger problem as there are some other workflows mentioned above. Probably needs a dev ticket opened. — xaosflux Talk 12:05, 11 January 2019 (UTC)
  • phab:T213544 opened. — xaosflux Talk 14:48, 11 January 2019 (UTC)

This issue is resolved now. Matma Rex talk 00:44, 15 January 2019 (UTC)

Thank you, I've reverted the workaround on MediaWiki:Gadget-defaultsummaries.js. — xaosflux Talk 02:54, 15 January 2019 (UTC)
  • Possibly related issue Inserting wiki markup via the bottom tabs after editing an edit summary will lead to the markup being in the summary, even when you have since edited the main article body. [Username Needed] 15:31, 11 January 2019 (UTC)

Is there a way to turn off the saving when hitting the ↵ Enter in the edit summary? I occasionally accidentally hit ↵ Enter instead of ', saving an incomplete summary or incomplete edit. I normally use the Alt+⇧ Shift+S (or click Publish) to save my edits, and would rather have it ignore ↵ Enter. I tried it with the "Add two new dropdown..." option both set and not set; it made no difference. —[AlanM1(talk)]— 11:06, 16 January 2019 (UTC)

User:Anomie/nosubmitsummary.js seems to still work for me. Anomie 12:53, 16 January 2019 (UTC)
@AlanM1: ^-- that, also the issue describe above was a bug, and was fixed. — xaosflux Talk 12:55, 16 January 2019 (UTC)

Responsive data tables for mobile Wikipedia[edit]

I think that the styling for tables on mobile can use a revamp. The horizontal scroll makes it hard to read information on smaller devices. Since Wikipedia already uses a separate subdomain for mobile, why not make tables responsive? Perhaps a column collapse like this: — Preceding unsigned comment added by LionelAsh (talkcontribs) 03:50, 11 January 2019 (UTC)

Bug when trying to move Roosevelt Senior High School (Washington, D.C.) to Theodore Roosevelt High School (Washington, D.C.)[edit]

I'm trying to move Roosevelt Senior High School (Washington, D.C.) to Theodore Roosevelt High School (Washington, D.C.) (as that is the current name used on the school website) when I get:

  • "[XDg0zgpAICwAALjkrO0AAABX] 2019-01-11 06:16:46: Fatal exception of type "MWException""

I would like to report this to Phabricator, but I also want to notify you guys here so I can have assistance moving the page WhisperToMe (talk) 06:18, 11 January 2019 (UTC)

@WhisperToMe: I've moved it for you by deleting the target page first instead of using the delete when moving over an existing page functionality. I had the same issue a few times earlier but forgot to note the exception for a bug report. — JJMC89(T·C) 06:31, 11 January 2019 (UTC)



I note that the blocking form has changed in this week's update, though I'm not entirely sure why or how it's an improvement. What would be useful though, is the form actually being at the top of the page, so you can fill it in without having to scroll down? A trivial thing, I know, but it all looks a bit amateur, and has done for a while. Black Kite (talk) 15:36, 11 January 2019 (UTC)

@Black Kite: you should be getting a bunch of text in that large empty space at the top (this isn't new) - are you blocking scripts or perhaps set your interface language to some English variant? — xaosflux Talk 15:39, 11 January 2019 (UTC)
The text you should see is MediaWiki:Blockiptext. — xaosflux Talk 15:40, 11 January 2019 (UTC)
Also note, it is so low on the page, because we have filled it up with all those directions (you could hide them with css as well). Look for example at the page in another language: de language block screen. — xaosflux Talk 15:43, 11 January 2019 (UTC)
The text isn't showing because of the #blockiptext { display: none;} line in User:Black Kite/monobook.css. {{Sensitive IP addresses}} isn't hidden though, so it pushes the rest of the form down. It probably started happening around the end of March 2018 when Special:Block switched to OOUI. --AntiCompositeNumber (talk) 16:01, 11 January 2019 (UTC)
Only the first three lines (and the comment lines also) of User:Black Kite/monobook.css are valid CSS - all the rest (i.e. this lot) is JavaScript and should probably be at User:Black Kite/monobook.js - I would fix it myself, but I can no longer do so (it's unlikely that I ever will be able to fix it, since I cannot be an interface admin without WP:2FA, and my hardware won't handle it). --Redrose64 🌹 (talk) 16:58, 11 January 2019 (UTC)
I suspect a lot of my CSS/JS is there to fix issues that have since been fixed by gadgets or scripts etc. - is any of it necessary any more? If not I'll just delete it. Black Kite (talk) 17:49, 11 January 2019 (UTC)
You can't put JavaScript into a CSS page. At best, it will be ignored; at worst, some strange things may happen. Either way, it won't be executed as JavaScript, so will not perform its intended purpose. If you merely revert this edit either you'll see no difference, or you'll see an improvement. --Redrose64 🌹 (talk) 18:05, 11 January 2019 (UTC)
Thanks. Black Kite (talk) 15:41, 12 January 2019 (UTC)
Related to the new block screen: I've taken the liberty of modifying the text label for the "disable talk page access" option. Before I changed it, it was labeled Editing their own talk page, which is not terribly helpful or descriptive. I used Twinkle's Prevent this user from editing their own talk page while blocked instead. The change is at Mediawiki:ipb-disableusertalk; I couldn't find any previous interface messages related to this option (other than MediaWiki:Ipballowusertalk, which obviously isn't relevant). Writ Keeper  18:15, 15 January 2019 (UTC)

Dupe ref name check could be made smarter[edit]

Now it trips just because parameters are in different order even if their content matches... E.g.

[1] [1]


  1. ^ a b "aa". Cite error: Invalid <ref> tag; name "bob" defined multiple times with different content (see the help page).

Somebody please make it smarter. Now articles get ugly errors when the only error is in the check! --Palosirkka (talk) 17:05, 11 January 2019 (UTC)

It doesn't expand templates. It compares the raw text between the <ref> and </ref> tags. You get the same thing with plain text and an extra space:[1][1]


  1. ^ a b cc Cite error: Invalid <ref> tag; name "kate" defined multiple times with different content (see the help page).
You would need to file a ticket at phab: to get this changed. --Redrose64 🌹 (talk) 17:23, 11 January 2019 (UTC)

Caching and slow to populate category.[edit]

About a month ago I adding some deprecated parameter tracking to {{Infobox song}} with this edit. The purpose is to populate Category:Pages using infobox song with deprecated parameters with pages using the old parameters. What I'm confused about is every few hours new pages are showing up in the category despite having no new edits. For example, Air Man ga Taosenai was just placed in the category this morning. Why is it taking SO long for the template to update these pages? Usually I've found that this happens within 24 hours or so. Am I missing something? --Zackmann (Talk to me/What I been doing) 19:03, 11 January 2019 (UTC)

The WP:Job queue has been unreliable in this respect since mid-2013, I've given up waiting for them to fix it. The thing to do is to ask Joe Decker (talk · contribs) nicely, if they wouldn't mind sending in Joe's Null Bot (talk · contribs) to process every page that transcludes {{Infobox song}}. Here's the list, it has over 63,000 entries. --Redrose64 🌹 (talk) 20:34, 11 January 2019 (UTC)
Taking a few weeks, or even a few months, is not unusual, especially for high-use templates (more than a few thousand transclusions). – Jonesey95 (talk) 21:07, 11 January 2019 (UTC)
The rendering of pages using a template is usually updated quickly in my experience. It's the link tables used elsewhere like on category pages which can be very slow. The search "Pages using infobox song with deprecated parameters" finds updated article pages, even for a hidden category like this. I currently get 826 hits. They display the category (assuming "Show hidden categories" is enabled at Special:Preferences#mw-prefsection-rendering) but Category:Pages using infobox song with deprecated parameters currently displays as empty. PrimeHunter (talk) 21:32, 11 January 2019 (UTC)
@Redrose64: Joe's Null Bot is currently down. I have also added the phab task that tracks this (below the other phab task tracked already) --DannyS712 (talk) 23:40, 11 January 2019 (UTC)

Gallery slideshow[edit]

Hello, considering the aforementoned issue, can somebody here help me along with this inquiry of mine from HD (which, unfortunately, remained unanswered)?--Hildeoc (talk) 22:55, 11 January 2019 (UTC)

OneClickArchiver creating unwanted pages[edit]

Is there a way to disable OCA on pages such as WP:UAA, where there is no reason to "archive" the posts? It results in the creation of a new page that serves no purpose. See the log for Wikipedia:Administrator intervention against vandalism/Archive 1 for what I mean. I also saw a similar situation with WP:RFP, which the accidental clicking of "Archive" resulted in the creation of a competently unnecessary page several times. funplussmart (talk) 15:39, 12 January 2019 (UTC)

Cannot log on through TOR[edit]

I have Wikipedia:IP block exemption. Today I attempted to log on through the Tor browser and got this error:

There seems to be a problem with your login session; this action has been canceled as a precaution against session hijacking. Please resubmit the form.

Resubmitting gives the same error. Clearing browsing/download/form/search history, site preferences, offline website data, active logins, cookies and cache and then restarting the computer didn't help. Trying to log on to several other WMF projects didn't help.

Previously I could log on through TOR, and I can still log on normally with Firefox through my normal ISP.

Has something changed?

I searched for "precaution against session hijacking"[2], read every thread, and tried everything anyone suggested. Nothing worked.

Note: I have IP block exemption because I often am at a remote site in China where industrial espionage is a real problem (I do consulting work in the toy industry) and my US employer requires that when in China I access the Internet through Tails and Tor. There is a lot of waiting around at the remote site so I have plenty of time to edit Wikipedia -- if I can log on. --Guy Macon (talk) 18:25, 12 January 2019 (UTC)

@Guy Macon: this seems to be related to phab:T151770 - can you try connecting with Tor, but using a different browser? — xaosflux Talk 20:48, 12 January 2019 (UTC)
As far as I know you need to use the Tor browser (Firefox based) to access Tor, so simply using a different browser won't work. I will give the suggested change to Firefox configuration in the phab a try and get back to you.. --Guy Macon (talk) 06:06, 13 January 2019 (UTC)
Nope. Made the change, cleared everything and restarted, still getting the error message. :( --Guy Macon (talk) 06:55, 13 January 2019 (UTC)
Since we are at the desperate stage, are you logged on somewhere? For example, if you had logged on as normal without Tor because you were in a safe location, did you log off there before trying the Tor logon? Perhaps you did not have to in the past, but if convenient it would be worth testing. Johnuniq (talk) 08:32, 13 January 2019 (UTC)
Good suggestion. I am indeed logged on normally. I occasionally fire up my Tor browser and do some small bit of noncontroversial typo fixing just so I know it will work the next time I get a call to go to China and fix a production problem. It doesn't seem like that should matter, but it is easy enough to test. I will get back to you after trying that. --Guy Macon (talk) 16:09, 13 January 2019 (UTC)
This could be due to attempted http inspection (interception) where the session has some sort of invalid certificate inserted that is rejected by the browser. Graeme Bartlett (talk) 05:44, 13 January 2019 (UTC)
I am pretty sure that it is a Wikipedia error message, not a browser error message. It is in the HTML that Wikipedia sends:
 <div class="error"><p>There seems to be a problem with your login session; this action has been canceled as a precaution against session hijacking. Please resubmit the form.</p></div>
I also changed my Tor Exit not to one on another continent. Didn't help. Tor is completely encrypted from the Tor Browser on my PC to the exit node, so if something is inserted it would have to be between the exit node and Wikipedia. The fact that is still happens when I change exit nodes would imply that someone is intercepting everything that goes to and from Wikipedia. The simpler explanation is that Wikipedia is cancelling my login and sending me the above error message. --Guy Macon (talk) 06:06, 13 January 2019 (UTC)
The above message is from MediaWiki:Sessionfailure. –Ammarpad (talk) 06:42, 13 January 2019 (UTC)
Thanks! I thought that it was one of our error messages. That Wikilink was to a page that was deleted ten years ago. Did you mean [$wgReadOnly ]? --Guy Macon (talk) 16:16, 13 January 2019 (UTC)
Guy Macon, the page Ammarpad linked is still a MediaWiki message despite being deleted - mediawiki messages have defaults and a page is only necessary when changing that default. Galobtter (pingó mió) 16:40, 13 January 2019 (UTC)
(edit conflict) MediaWiki:Sessionfailure sounds right. It displays the exact message you quoted. Messages in the MediaWiki namespace have a default value which is used if a wiki has not created or has deleted the corresponding page. The link is blue, indicating the message name is known to MediaWiki even though there is no wiki page. PrimeHunter (talk) 16:42, 13 January 2019 (UTC)
@Guy Macon: I get exactly the same uninformative message when I try to log on with cookies disabled. Could your browser or Tor client could be blocking cookies from for some reason? Suffusion of Yellow (talk) 22:23, 13 January 2019 (UTC)
...aaand it logged in just fine after I told it to accept cookies from and Thanks! I can't help but wonder, though, what changed. I test this once a month or so so and last month I was able to log in even though I wasn't accepting cookies.
OK, my problem is solved, but what can we do to help others? How about a help page that tells the user about the cookies, suggests clearing the cache, etc. rather than an uninformative "Please resubmit the form" error message? Who do we talk to to get a better error message? --Guy Macon (talk) 22:46, 13 January 2019 (UTC)
A local version of MediaWiki:Sessionfailure can be created and edited by administrators. It could link Help:Logging in. PrimeHunter (talk) 22:59, 13 January 2019 (UTC)
I asked for help at AN.[3] --Guy Macon (talk) 23:11, 13 January 2019 (UTC)
@Guy Macon: I'm assuming your logon failure isn't limited to the English Wikipedia? If so I suggest you open a phab ticket. — xaosflux Talk 12:26, 14 January 2019 (UTC)
When I found that I suddenly could not log on to the English Wikipedia, I tried the french, Wiktionary, Meta, and half a dozen others -- no luck. When I enabled cookies on Wikipedia, I was able to log on, and the unified login logged me on to all of the other projects, even ones that still have cookies blocked. So yes, it appears to be project wide. This implies that as a bare minimum we should fix the help page for all languages and all WMF wikis. There is also a strong argument for fixing the underlying problem instead of simply making the error message suggest turning on cookies; first, anyone using Tor is especially likely to turn off cookies, block trackers, etc. Second, it isn't only the WMF that uses our software. A bunch of independent wikis use it as well. Sites like Brickepedia use our software but do not have our error pages unless they come "baked in" with the software. Third, it isn't our business to force the users to accept cookies, and indeed other than logging on, Wikipedia works just fine with cookies disabled. --Guy Macon (talk) 16:56, 14 January 2019 (UTC)
@Guy Macon: thanks for the update, just making sure we captured everything:
  1. You are able to logon and edit via TOR now
  2. The primary technical issue was that you were blocking cookies, and allowing the authentication cookies resolved this for you
  3. Updating our interface message (and further the mediawiki default) for MediaWiki:Sessionfailure to include information about cookies may be useful.
Thanks for your patience as well! — xaosflux Talk 17:04, 14 January 2019 (UTC)
That appears to cover it. --Guy Macon (talk) 17:25, 14 January 2019 (UTC)
@Guy Macon: thanks for the update, I've updated our local message to reference cookie handling, and have requested a core update in phab:T213763 - coincidentally 10 years to the date after phab:T19029 was opened for Session failure warning message ('sessionfailure') gives bad advice. — xaosflux Talk 22:00, 14 January 2019 (UTC)

Category tree trace tool[edit]

Is there any tool which can help me understand why a specific article is in a specific category tree? For example, what is the root from Category:Mammals to Phronima? עוד מישהו Od Mishehu 12:20, 13 January 2019 (UTC)

I think that Pintoch (talk · contribs) was working on something like this. --Redrose64 🌹 (talk) 18:42, 13 January 2019 (UTC)
@Od Mishehu and Redrose64: There is vCat (but it's not my tool). − Pintoch (talk) 08:25, 14 January 2019 (UTC)


Hi. I have BrandonXLF's invert script here but for some reason it went away from my menu bar. --Thegooduser Life Begins With a Smile :) 🍁 21:40, 13 January 2019 (UTC)

Have you tried asking at User talk:BrandonXLF? — xaosflux Talk 21:46, 13 January 2019 (UTC)
Outside that, start looking for conflicts - remove all your imported scripts except for that one and see if it is still working. — xaosflux Talk 21:47, 13 January 2019 (UTC)
Xaosflux If I remove all my scripts how can I see which ones I have Installed? Can I rollback my edits after? --Thegooduser Life Begins With a Smile :) 🍁 21:48, 13 January 2019 (UTC)
Asked it on BrandonXLF's talk page. Thegooduser Life Begins With a Smile :) 🍁 21:56, 13 January 2019 (UTC)
@Thegooduser: certainly, your script page (User:Thegooduser/vector.js) has a history that you can revert to just like any other page. — xaosflux Talk 22:14, 13 January 2019 (UTC)
@Thegooduser: The latest version moved the link from your "p-personal" toolbar to the "p-tb" toolbar - check the tools side menu (the one with "page information," "permanent link," etc.) --DannyS712 (talk) 22:04, 13 January 2019 (UTC)

Help with modules[edit]

I noticed that a lot of templates on Wikipedia use modules/Lua/Scribunto. I know modules can do almost the same things as templates. However, I was wondering if I can set two arguments to do the same thing. For example, with {{Infobox film}}, I can use both "film-name" and "film name" arguments. They do the same exact thing (in the template's code is "| data1 = {{{film_name|{{{film name|}}}}}}"). Now, say the template was converted to Lua — how would it be written? —CaiusSPQR (talk) 23:22, 13 January 2019 (UTC)

That assignment might be written:
local data1 = frame.args['film_name'] or frame.args['film name']
Perhaps read WP:LUA.
Trappist the monk (talk) 00:17, 14 January 2019 (UTC)
Thank you. I'm a bit confused though. Should I use "data1" or both "frame.args" in the rest of the code? —CaiusSPQR (talk) 00:23, 14 January 2019 (UTC)
Were it me, I would use data if I needed to use that value more than once so that every time I needed the value assigned to |film_name= or |film name= I didn't have to do the 'or'ing every time.
Trappist the monk (talk) 00:32, 14 January 2019 (UTC)
Thank you so much for your help! 😁 —CaiusSPQR (talk) 00:34, 14 January 2019 (UTC)
Some lua tips:
Depending on what you want to do, you can also use a template wrapper for parameter conversion: {{#invoke|...|data1={{{film_name|{{{film name|}}}}}}|...}}.
frame.args['film_name'] can be shortened to frame.args.film_name (table).
Lua treats empty parameters ("" or spaces) differently than template: Module:Arguments#Trimming and removing blanks. MarMi wiki (talk) 02:09, 14 January 2019 (UTC)

FileExporter beta feature[edit]

Johanna Strodt (WMDE) 09:41, 14 January 2019 (UTC)

Tech News: 2019-03[edit]

@Trizek (WMF): I think this edition got truncated. Graham87 04:13, 15 January 2019 (UTC)
@Graham87: yes it was (meta:). You can read the entire issue here: meta:Tech/News/2019/03. — xaosflux Talk 04:55, 15 January 2019 (UTC)
@Graham87:, yeah, I'm very sorry about that. I've fixed it above. Trizek (WMF) (talk) 13:45, 15 January 2019 (UTC)

Black box appearing[edit]


Today I have logged onto Wikipedia and there seems to be a solid black box on the articles box and what activity I am doing. For example: This page right now will have a black box around "Article" and a black box around "Edit source". Another example would be if I was on a random page and there would be a black box around "Article" and a black box around "Read".

Thank you AmericanAir88(talk) 22:59, 14 January 2019 (UTC)

What is your browser? What is your skin at Special:Preferences#mw-prefsection-rendering. Does it happen if you log out? PrimeHunter (talk) 23:10, 14 January 2019 (UTC)
@PrimeHunter: It happens when I log out. I use safari. I use the default skin. AmericanAir88(talk) 23:45, 14 January 2019 (UTC)
Is it on a computer or iOS device? Safari dropped the Windows version in 2012 so I cannot test it on a computer. I don't see it on an iPad. It sometimes helps with interface rendering to clear the entire cache. PrimeHunter (talk) 00:01, 15 January 2019 (UTC)
@PrimeHunter: Clearing the cache did not work. This is on a MacBook Air. AmericanAir88(talk) 00:39, 15 January 2019 (UTC)
MacBook Air uses macOS and not iOS so I cannot help properly. PrimeHunter (talk) 00:47, 15 January 2019 (UTC)
@PrimeHunter: Is there someone I can speak to who has expertise in MacOs? AmericanAir88(talk) 02:50, 15 January 2019 (UTC)
Using Safari on MacBook Air and everything appears normal to me, logged-in and out. –Ammarpad (talk) 07:52, 15 January 2019 (UTC)
I don't see any changes, either, AmericanAir88. Have you tried restarting? After that, as today's not Thursday, I'd look at updates (Safari, OS X, or a Safari extension) and new extensions, and double-check that I hadn't changed any preferences. BlackcurrantTea (talk) 16:58, 15 January 2019 (UTC)

List of Wikipedians by article count[edit]

I know I've raised this before, but does anyone have the time and skills to be able to fix this list so it updates automatically each day, as per the number of edits list? Many thanks. Lugnuts Fire Walk with Me 09:07, 15 January 2019 (UTC)

@Lugnuts: that really needs to be done by a bot, it looks like it used to be done by MZMcBride's bot - so you could start by asking MZMcBride. Alternatively, someone else could take it up, and you could ask at WP:BOTREQ. — xaosflux Talk 13:50, 15 January 2019 (UTC)
Thank you. IIRC, I've asked MZMcBride before about this, and I think there was something technical stopping it from being updated. I'll do a bit of digging, and see what I find. Lugnuts Fire Walk with Me 13:58, 15 January 2019 (UTC)

Code correction[edit]

Hi - I don't do code so would like this line looked at

div style="padding: 14px; background: #FFFAF0;<!--#ddcef2;--> border-style: groove; border-width: 5px; border-color: green; font-size: 95%; text-align:left; font-family: Euphema, arial;"

Its the top line on my talk page and its in pink; is the code deprecated perhaps? Thanks your time Face-smile.svg MarkDask 15:22, 15 January 2019 (UTC)

Added nowiki tags to your div above. I suspect that your browser doesn't ignore the commented color value (violet) (tested on Firefox based/IE11/Chrome based and the background color is FFFAF0), or you just need to refresh (F5/Ctrl+F5) the page. MarMi wiki (talk) 16:49, 15 January 2019 (UTC)
And comments in css (in style too?) are done by /* and */ - CSS Comments. MarMi wiki (talk) 16:57, 15 January 2019 (UTC)
Yes, according to CSS Style Attributes (linked from Semantics, structure, and APIs of HTML documents) the value of the style attribute must match the syntax of the contents of a CSS declaration block (excluding the delimiting braces). So comments inside a style= attribute can only be of the /* ... */ form. --Redrose64 🌹 (talk) 21:35, 15 January 2019 (UTC)
(edit conflict) @Markdask: I suspect this is because you have a <div> element that is missing a </div> closing tag. This happens on talk pages when people want unusual affects sometimes. Are you trying to keep that box border around future sections that other people add, or just around your welcome box? — xaosflux Talk 16:50, 15 January 2019 (UTC)

AFD transclusion[edit]

Bit of an odd situation I wanted to see if somebody on the technical side could look into. Earlier today, User:Cyberbot I detected Wikipedia:Articles for deletion/John Dingfelder as "not correctly transcluded to the log", and added it to today's AFD list accordingly — but the page is, and always has been, correctly transcluded to the daylog for the day it was created, January 9 (and there's no indication in the edit history of that page that it got temporarily vandalized, the only explanation I can think of for why a discussion on it might appear untranscluded six days later.) And it's not a case of the discussion having been relisted for further consideration without getting readded to the new daylog, either, as January 9 isn't seven days ago yet and the discussion's already on a consensus track and won't even need to be relisted at all.

So could somebody take a few minutes to look into why a bot might make this mistake? Thanks. Bearcat (talk) 16:25, 15 January 2019 (UTC)

Find unusually short articles[edit]

How can I query (or search, if no such list exists) the English wikipedia for articles obviously having a problem because they are meaninglessly short, i.e. zero prose, or less than 10 words, or even less than 100 words. Searching wikipedia is likely to be an enormous task, and such a script might need admin or even higher access to the wiki, Nonetheless, can it be done?

If there is a list, or when there is one, what can/should be done with it? These articles should be moved to some kind of editors-only (requires login to access) sandbox until they are suitable for publication, but I know of no such global structure. Articles can be so severely impaired for reasons other than length that they shouldn't be in the encyclopedia. Maybe deleting them isn't the best policy, because we'd rather eat dirt than delete a page because, well, it just doesn't measure up, and not for any admin or technical reason. Sbalfour (talk) 22:14, 15 January 2019 (UTC)

See Special:ShortPages. It excludes disambiguation pages but includes set index articles like surnames with two entries. PrimeHunter (talk) 22:31, 15 January 2019 (UTC)

File names which have been changed[edit]

I'm going to ask this here since I figure it's a technical question and because there are probably more people watching this page than there are watching WT:File names. What happens to a file name once it has been changed? Is a redirect to the new name created? I came across File:Kedi (2006 film).jpg while checking on some non-free images; this file used to be named File:Kedi.jpg, which if you click on the link you'll see is now a completely different file uploaded to Commons. It's probably not a big deal except that there is a link to the original file name in Wikipedia:Non-free content review/Archive 55#File:Kedi.jpg which now lead to a completely different file and the log shows that "File:Kedi.jpg" was deleted per WP:F2. I know that {{Shadows Commons}} is used in cases where two files (one on Commons and one on Wikipedia) have the same name; but, I'm not sure if that applies in a case like this. -- Marchjuly (talk) 05:51, 16 January 2019 (UTC)

Date format in signatures as viewed by unregistered users[edit]

Is there a reason why the date format in signatures is dd mmm yyyy when viewed by unregistered users? I presume it's just the way the css is configured but is there any background for that configuration rather than mmm dd, yyyy? Nthep (talk) 13:28, 16 January 2019 (UTC)

@Nthep: signature dates are not stored as any sort of typed data value, they are simply text and readers see them just as they are in the wikitext. In fact everyone sees them this way unless they specifically implement a hack to see them some other way. — xaosflux Talk 14:06, 16 January 2019 (UTC)
I get that but was there any reason why it's the way round it is, other than that's the way the developer set it? Nthep (talk) 14:29, 16 January 2019 (UTC)
I think that day-month-year is more common around the world than month-day-year, which IIRC is pretty much just a North American thing (or maybe more generally an English thing?). It also orders the time units by length. Writ Keeper  14:34, 16 January 2019 (UTC)
It's just a North American thing, I live in the UK and we use DDMMYY. [Username Needed] 15:10, 16 January 2019 (UTC)
See also Date format by country#Usage map. Some wikis use another date format. PrimeHunter (talk) 18:42, 16 January 2019 (UTC)
Is there an example somewhere of such a "hack" I can use to make the standard sig times display the same way as my (localized ISO-3166) history/contrib times (e.g. to convert "01:23, 16 January 2019 (UTC)" to "2019-01-15T17:23")? —[AlanM1(talk)]— 15:12, 16 January 2019 (UTC)
@AlanM1: there is a Gadget for that, see Wikipedia:Comments in Local Time for documentation and limitations. — xaosflux Talk 15:20, 16 January 2019 (UTC)
I think you'll also want to check out Special:Preferences#mw-prefsection-rendering, specifically the "Date format" and "Time format" sections. ~ Amory (utc) 15:23, 16 January 2019 (UTC)