Wikipedia:Village pump (technical)#Edit window changes

Policy

Technical

Proposals

Idea lab

WMF

Miscellaneous

• Table of contents
• First discussion
• Jump to bottom
• New topic
• All village pumps

Shortcuts

• WP:VPTWP:VPT
• WP:VP/TWP:VP/T
• WP:TECHPUMPWP:TECHPUMP
• WP:PUMPTECHWP:PUMPTECH

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).

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. Discussions are automatically archived after remaining inactive for 5 days.

view · edit Frequently asked questions (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. 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 task 3864. There is an accesskey property on it (default to accesskey="f" in English). Logged-in users can enable the "Focus the cursor in the search bar on loading the Main Page" gadget in their 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. An offline spellcheck of all articles is run by Wikipedia:Typo Team/moss; human volunteers are needed to resolve potential typos. If you have problems making your fancy signature work, check Help: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. For server or network status, please see Wikimedia Status. If you cannot reach Wikipedia services, see Reporting a connectivity issue.

« Archives, 205, 206, 207, 208, 209, 210, 211, 212, 213, 214, 215, 216, 217, 218, 219, 220, 221, 222, 223, 224, 225

Copying a discussion from WP:AN:

[1]NZ-non-WANZ-members is placed in the main namespace, and is listed in e.g. the newpages feed, but is not editable (is some mass-message list?). I suppose this shouldn't be possible, but as it needs deletion an admin is needed. @Schwede66: you created this, perhaps you can explain what is happening? Fram (talk) 14:45, 3 December 2025 (UTC)

Moved to Wikipedia namespace. ~~ AirshipJungleman29 (talk) 15:09, 3 December 2025 (UTC)

Yep. Probably not the right location either (should be with some project, and when editing is listed suddenly as a "special" page). In any case, such pages shouldn't be creatable in the mainspace, serve no purpose there. Fram (talk) 15:23, 3 December 2025 (UTC)

I think there's definitely a valid task to restrict MassMessage to operating in certain namespaces. Izno (talk) 18:39, 3 December 2025 (UTC)

phab:T411661 now. Izno (talk) 18:44, 3 December 2025 (UTC)

Right then. Sorry for creating a mess. This did end up in the wrong namespace, and I wasn't awake enough to notice that. For how it happened:

• Wikipedia:Mass message senders has, under the heading "Before making your request", a line that reads: "Administrators and template editors may create lists with a special content model via Special:CreateMassMessageList."
• That opens a form at Special:CreateMassMessageList
• Nothing there indicates that you must move the list to a particular namespace after creating it.

Once you've stuffed it up, you'll probably remember for next time. But it would obviously be better for the form to be set up so that the list gets created in the correct namespace. Schwede66 23:41, 3 December 2025 (UTC)

You can consider trying it right now, but I'm guessing the page title input there is for a {{FULLPAGENAME}} and not the {{PAGENAME}}. Izno (talk) 00:03, 4 December 2025 (UTC)

Is there anything we can do (without waiting for the Phabricator ticket to be resolved) that would prevent this tool from creating pages in mainspace, or at least add a big warning so that people don't leave it in mainspace accidentally? We can't edit the Special namespace, but perhaps an edit to the relevant message names would help with the warning. Could we create an edit filter preventing anyone from putting this kind of page in mainspace? I don't understand how the tool works (perhaps my requests aren't possible), because I don't want to go testing and make a mess. Nyttend (talk) 05:44, 10 December 2025 (UTC)[reply]

@Nyttend I created Special:AbuseFilter/1394 that will warn on any attempt to create a non-wikitext page in the article namespace. --Ahecht (TALK
PAGE) 15:44, 10 December 2025 (UTC)[reply]

This should have been discussed at WP:EFN. Every edit filter has a cost and we have generally avoided using it as a guardrail for a niche mistake that occurs once in a blue moon. – SD0001 (talk) 14:12, 12 December 2025 (UTC)[reply]

@SD0001 Feel free to disable if you feel it has limited utility. --Ahecht (TALK
PAGE) 22:34, 15 December 2025 (UTC)[reply]

Over time, I have found searching the Wikipedia namespace more and more difficult as the results are increasingly discussion pages in the Wikipedia namespace. Is there an easy way to omit these? For example, of the first 20 search results for "talk header cruft", 12 are AFD, TFD, and RFD pages, 3 are peer review discussions, and 3 are featured candidate discussions. That leaves only two of twenty pages that are not obviously discussion pages. Is there a way to do a search query of the Wikipedia namespace that will omit most discussion pages?

Perhaps due to poor search terms, the most relevant pages for my query did not appear. These are:

• Wikipedia:Talk page templates
• Wikipedia:Avoid template creep

Daask (talk) 15:46, 10 December 2025 (UTC)[reply]

@Daask You can exclude some titles or some words and phrases: example. I'm not sure what you're looking for in these pages (talk? header?), maybe your search terms need some improvement too. See Help:Search. Ponor (talk) 15:59, 10 December 2025 (UTC)[reply]

I suppose I can create a custom search box with -intitle:"Articles for deletion" as you suggest. However, I thought this was likely a common problem that others have experienced and thought perhaps there was a common solution. Daask (talk) 16:22, 10 December 2025 (UTC)[reply]

You're most definitely not the only one. I have a browser shortcut to use google to search wikipedia when I want to actually find something in project or template space, because the SERP here are so frustrating. -- Avocado (talk) 15:57, 14 December 2025 (UTC)[reply]

Adding a -UTC will get rid of pages with signed comments, and it's unlikely that much baby will go out with the bathwater. Unfortunately, there's no clean way to do this, and it would probably be a better system if we had projectspace split into a place for policy/guideline/information/essays and a place for project-wide discussions. Firefangledfeathers (talk / contribs) 22:53, 10 December 2025 (UTC)[reply]

Yeah, I wish we had something like -hasmagicword:NEWSECTIONLINK -hasmagicword:NOEDITSECTION, which would work for this use case. Nardog (talk) 16:05, 14 December 2025 (UTC)[reply]

Hi all! I wanted to quickly share that the WMF Growth team is planning to release an updated version of the email new users receive to confirm their email address. The new version introduces clearer and more welcoming language that we hope will result in more new users confirming their email and continuing on their journey to becoming editors. You can read more in this Growth team update, and feel free to let us know if you have any questions or feedback! Sdkb‑WMF ^talk 00:07, 13 December 2025 (UTC)[reply]

Personally, I hate obfuscated boxes which, when clicked, do who-knows-what. Aren't prospective editors assumed to be clever enough to understand a simple sentence saying what clicking the box will do? Johnuniq (talk) 04:00, 13 December 2025 (UTC)[reply]

Lol, which Internet are you on ? ;) —TheDJ (talk • contribs) 10:00, 13 December 2025 (UTC)[reply]

I don't see what's obfuscated about a box that clearly says "Confirm your email". – SD0001 (talk) 10:49, 13 December 2025 (UTC)[reply]

What actions will be taken if the button is clicked? What are the consequences? Johnuniq (talk) 03:12, 14 December 2025 (UTC)[reply]

<sarcasm>Isn't it obvious? Your Wikipedia account will become complete.</sarcasm> – Jonesey95 (talk) 07:00, 14 December 2025 (UTC)[reply]

Confirming your email? — Qwerfjkltalk 16:03, 14 December 2025 (UTC)[reply]

If you provide an email, but don't confirm it, then you won't be able to use almost any of the email features, including email notifications or emails from other users. (I think we only use non-confirmed email addresses for password resets.)

This is because until you prove that you can receive emails (by clicking that button), we don't know that the email you provided is really yours, and we don't want to let you spam someone else with notification emails etc.

I think this email confirmation workflow is common enough across websites that it doesn't require explanation. Many don't even let you use your account at all without confirming the email (although on Wikipedia, it is only recommended, not required). Matma Rex talk 21:35, 14 December 2025 (UTC)[reply]

Uh, thanks, but I've been around long enough to be able to guess what it does. I'm thinking of other people. WTF does "Confirm your email" mean? Some small (but readable) text saying what would happen should be under the box. Johnuniq (talk) 03:40, 17 December 2025 (UTC)[reply]

You need to have a visible URL to click for people who don't click buttons without being able to see where it's taking them. SarekOfVulcan (talk) 16:05, 14 December 2025 (UTC)[reply]

On mobile, you can long-press the “Confirm your email” button, or on desktop, hover over the link, to preview the destination URL. The link points to the wiki’s local Special:ConfirmEmail page.

For users whose email accounts are configured to display messages in plain text, a plain text version is also available and displays the full URL directly. - KStoller-WMF (talk) 19:26, 14 December 2025 (UTC)[reply]

I knew about hovering, but this is the first time I heard that long-pressing works. Thanks! (Doesn't mean I don't still think you should show the URL clearly instead of obscuring it.) SarekOfVulcan (talk) 23:18, 14 December 2025 (UTC)[reply]

Also, personally, I'd prefer "Wikipedia account setup is almost complete". How do you complete an account? --SarekOfVulcan (talk) 21:58, 16 December 2025 (UTC)[reply]

Yes. Hence my sarcasm above. Feel free to comment at T396155. I get the sense that the people working on the wording of this message are maybe not native English speakers. Their reason for rejecting my proposed copy edit does not make sense to me. – Jonesey95 (talk) 02:32, 17 December 2025 (UTC)[reply]

In 2013, I created Wikipedia:Wikidata/Wikidata Sandbox, which links to d:Q4115189, Wikidata's sandbox item. A few months after that, mw:Extension:Echo went live, and for the 12 years since, I have gotten a notification every time someone adds or removes that page from the sandbox on Wikidata. I added the page to the mute list at Special:Preferences#mw-prefsection-echo, but that appears not to apply to Wikidata updates. This has been getting mildly on my nerves over time, and so I thought I'd ask if an admin could IAR-revdel my username from the first revision, but I just tested that on testwiki (creating testwiki:Page creation test and then using my alt to link it to testwikidata:Q243049 since you don't get notified if you're the linker), and I still got notified after revdelling my own username. I'm not sure whether it's that Wikibase "sees through" the revdel or just looks for the second username (also mine), but either way, I still get notified. If it's the latter, I guess this could be fixed by revdelling by username from all of the first three edits to the local Wikidata sandbox page, but that'd just pass the problem on to the second editor, which would be unfair.

Is there anything that can be done here? I can only think of two things, both kind of ugly:

• Find a random empty edit by an IP from before 21 February 2013 and merge it into the page history, making that IP the page's "creator".
• Have someone, using an account that they don't mind getting all these pings on, move the current page to Wikipedia:Wikidata/Wikidata Sandbox/old and then copy that page's contents over the resulting redirect.

There is, in fairness, also the option of telling me to suck it up and deal with the occasional pointless ping. :P -- Tamzin^{[cetacean needed]} (they|xe|🤷) 20:02, 13 December 2025 (UTC)[reply]

Are you getting echo's from here or from wikidata? If the later, did you mute the page on that project? 20:47, 13 December 2025 (UTC) — xaosflux ^Talk 20:47, 13 December 2025 (UTC)[reply]

I get the notifs here, same as I get if someone links one of my articles or DABs on Wikidata. Specifically this is MediaWiki:notification-header-page-connection. -- Tamzin^{[cetacean needed]} (they|xe|🤷) 20:56, 13 December 2025 (UTC)[reply]

@Tamzin: If you disable "Connection with Wikidata" at Special:Preferences#mw-prefsection-echo then the notificatinos will stop but also for all other pages you have created. Is that acceptable? PrimeHunter (talk) 21:08, 13 December 2025 (UTC)[reply]

Occasionally such notifications make me realize I've forgotten to check if a creation exists on other wikis, letting me make merges like d:Special:Diff/2430612829, so I'd rather not disable the thing entirely. -- Tamzin^{[cetacean needed]} (they|xe|🤷) 22:12, 13 December 2025 (UTC)[reply]

Instead of "blaming" random IP users, maybe we could import a fake 2013 edit by User:Maintenance script creating that page. Or maybe we could actually create that page by running a maintenance script (after deleting/moving the current version). Or maybe the Wikidata connections should respect that preference, it seems like a similar use case. Matma Rex talk 23:52, 13 December 2025 (UTC)[reply]

This seems like it should be an echo feature request or bug; not a need to hack around otherwise normal behavior. — xaosflux ^Talk 13:19, 14 December 2025 (UTC)[reply]

In theory I agree (in my opinion, rewriting a page's history works against one of the basic principles of a wiki and should be avoided). In practice, although it is a hack, for this specific case of a sandbox, if all of the editors of that page are amenable (without trying to sway their decision), it could be deleted and re-created, as suggested by Matma Rex. On the third hand, I don't really like putting a burden on the page editors that they might be seen as non-collaborative if they don't agree.

I am sympathetic about the desire not to see Wikidata notifications for a specific page, and that implementing a specific feature for it (particularly if editors would like to decouple it from the page link notification muting feature) may not happen anytime soon. Maybe someone is interested in writing a script to automatically acknowledge notifications meeting a configured criteria? isaacl (talk) 18:06, 14 December 2025 (UTC)[reply]

@Tamzin If you can convince either of the two importers to import this file, it will resolve your problem by inserting a fake page creation revision into that page. (No need to delete or move it.) I don't really know whether this is cool by the rules, but I haven't found a rule against it; I'm leaving it for you three to figure out.

If that turns out not to be cool, then I suggest filing a bug to make the mute preference apply to these notifications too. I can't promise that this will be prioritized though. You may be the only person facing this problem. :) Matma Rex talk 21:55, 14 December 2025 (UTC)[reply]

That almost seems like a targeted nerd-sniping attack.. and of course it worked. taavi (talk!) 22:58, 14 December 2025 (UTC)[reply]

Thank you taavi!! -- Tamzin^{[cetacean needed]} (they|xe|🤷) 16:12, 16 December 2025 (UTC)[reply]

Thinking of filing a bug but want to make sure I'm not the only one here.... Looking at Burbank, California (for example) on my iPhone using the Wikipedia app. The Mapframe won't load. Just seeing an empty square with a blue question mark. Can anyone else confirm this issue? Zackmann (^{Talk to me}/_{What I been doing}) 00:22, 14 December 2025 (UTC)[reply]

This looks like phab:T193312, open since 2018 and low priority... Sjoerd de Bruin (talk) 20:45, 14 December 2025 (UTC)[reply]

No, this is because the apps use Parsoid, and support for maps in Parsoid is not yet working. There is work happening on this problem for the last couple of weeks, but the mapimage generation is a complex process and the adaptations need to happen at multiple levels, so its a slow process. —TheDJ (talk • contribs) 17:26, 15 December 2025 (UTC)[reply]

Gotcha. HUGE relief to know it is being worked on because it is broken on a sizeable percentage of the pages on the English Wikipedia (and I'm sure a large chunk of other languages). Zackmann (^{Talk to me}/_{What I been doing}) 17:27, 15 December 2025 (UTC)[reply]

I'm setting up the Wikipedia:Meetup/Toronto/Wikipedia Day 2026 page. For the signup button, it uses the signup template and I tried to use the ordered list. But the template doesn't seem to accept this parameter despite the documentation. Anyone knows if the template for ordered list is bugged? OhanaUnited^{Talk page} 05:24, 15 December 2025 (UTC)[reply]

@OhanaUnited: Template:Meetup/signup was bugged. #tag:inputbox is needed to pass template parameters per mw:Extension:InputBox#Using InputBox in a template. Fixed by [2]. PrimeHunter (talk) 10:34, 15 December 2025 (UTC)[reply]

Thanks very much for the quick fix. OhanaUnited^{Talk page} 14:07, 15 December 2025 (UTC)[reply]

MediaWiki message delivery 19:00, 15 December 2025 (UTC)[reply]

Firstly, why does using Google Translate to translate a page from the desktop site always redirect to a translated version of the page on the mobile site?

Secondly, and I appreciate that this may not be the correct forum, why have some Wikipedias, like the German and Japanese sites, ceased to work in Google Translate in the past couple of days? Eldomtom2 (talk) 00:33, 16 December 2025 (UTC)[reply]

@Eldomtom2 Because google uses mobile browsers to consume the original website, as most people in the world use mobile. Im not familiar with the second problem. —TheDJ (talk • contribs) 01:04, 16 December 2025 (UTC)[reply]

@TheDJ, apparently not, see mw:Talk:Requests for comment/Mobile domain sunsetting#Another Google snafu with mobile domains. Izno (talk) 01:10, 16 December 2025 (UTC)[reply]

TIL.... —TheDJ (talk • contribs) 15:12, 16 December 2025 (UTC)[reply]

not sure about issue, how to reproduce it? i use timeless skin, it is a lot more functional than the mobile mode (including on smaller screens) Gryllida 02:52, 16 December 2025 (UTC)[reply]

@Eldomtom2: ?useformat=desktop (there is also a ?useformat=mobile) gives the desktop version in Google Translate for me, e.g. https://ja.wikipedia.org/wiki/Example.com?useformat=desktop. Please always give an example. Translation from Japanese and German works for me at https://translate.google.com, e.g. https://ja.wikipedia.org/wiki/Example.com and https://de.wikipedia.org/wiki/Beispieldomain. Do you have the source language set correctly? "Detect language" also works for me. PrimeHunter (talk) 09:49, 16 December 2025 (UTC)[reply]

Okay, inexplicably now they work - but in the case of the Japanese Wikipedia only if I choose "detect language".--Eldomtom2 (talk) 13:01, 16 December 2025 (UTC)[reply]

Is there any option other than WP:RX to get BBC paywalled articles? (That is, do I have to go to RX every time I need a BBC article, or is there another route?) They aren't in ProQuest, they aren't in archive.org, and the price they've put on the subscription is absurd. Is it only the US that is paywalled? SandyGeorgia (Talk) 01:35, 16 December 2025 (UTC)[reply]

i didn't know bbc was paywalled, they never asked me to subscribe, do you have a link? Gryllida 02:52, 16 December 2025 (UTC)[reply]

@SandyGeorgia: For clarity, WP:RX means asking a someone else to find the article and somehow distribute it to you. I assume you have tried https://wikipedialibrary.wmflabs.org. Yes it seems to be US only.[5]. Do you have an example of an article that is paywalled (I can't see experience the paywall in Australia). Is this 2017 BBC.com article paywalled and the archive.org stored version not available?

I think a Virtual private network (VPN) may allow you to mask your IP address as from the UK and access the articles. There could be free options, and given the small bandwidth of text articles speed is not a great concern for you. Commander Keane (talk) 02:55, 16 December 2025 (UTC)[reply]

Commander Keane thanks for the help!

I'm often at WP:RX, but I hope to not have to go there every time I need to see a BBC article! Yes, I can see that 2017 article you link; I think it's only newer articles that are affected in the US (how rude - and it's $8.99 per month to subscribe, which is steep). I use WP:TWL for many US sources, but don't see the BBC anywhere there. I don't know anything about VPNs and am too old to learn! So I'll stop citing the BBC, but I need access for a copyvio check.

The article I need is "Maduro says Trump wants Venezuela's oil. But is that the real US goal?" If anyone can email it to me, I guess I'll have to put up a cot at WP:RX. SandyGeorgia (Talk) 10:19, 16 December 2025 (UTC)[reply]

@SandyGeorgia: Check your inbox. – DreamRimmer ■ 10:39, 16 December 2025 (UTC)[reply]

@SandyGeorgia: If you Google "online VPN", you can find a number of services that'll proxy you to an IP in a given country without you having to download anything or register an account. These aren't very helpful for most of the things people use VPNs for, due to limited bandwidth and poor privacy, but should work fine for the specific use-case of accessing news articles. -- Tamzin^{[cetacean needed]} (they|xe|🤷) 10:55, 16 December 2025 (UTC)[reply]

SandyGeorgia, archive.is should work. — Qwerfjkltalk 16:26, 16 December 2025 (UTC)[reply]

@DreamRimmer, Tamzin, and Qwerfjkl: thank you all for the help. Tamzin, I'm afraid of trying to set up a VPN; old dog, new tricks, afraid I'll do something that will cause me a problem ... but I appreciate the tips and will look into it. Thx also DreamRimmer. I think what would work best going forward would be for me to have a way for someone to archive the non-paywalled version as Qwerfjkl did, as I (and anyone else who needs to) can see that ! I will ask around to see if there is an editor outside of the US who will archive.is any article I need. Thanks again, SandyGeorgia (Talk) 17:50, 16 December 2025 (UTC)[reply]

SandyGeorgia, you should be able to archive any page yourself, on archive.is. — Qwerfjkltalk 13:13, 18 December 2025 (UTC)[reply]

It would do no good for me to archive a page I can't access :) I'd be archiving the paywalled URL ... but if someone who has access can archive a BBC page, then those in the US can see it. SandyGeorgia (Talk) 15:14, 18 December 2025 (UTC)[reply]

SandyGeorgia, I think archive.is archives it on their servers, so whether you can access it or not doesn't affect it. — Qwerfjkltalk 16:17, 18 December 2025 (UTC)[reply]

Thx ... next time I need to access BBC, I'll check there! SandyGeorgia (Talk) 17:09, 18 December 2025 (UTC)[reply]

But for now, to deal with another in a long string of copyvio. SandyGeorgia (Talk) 17:55, 16 December 2025 (UTC)[reply]

You can reach out to me at any time. I will be happy to provide archived copies. – DreamRimmer ■ 18:17, 16 December 2025 (UTC)[reply]

Thanks so much, DreamRimmer; if things ever settle down for me, I may try a VPN per Tamzin, but in a pinch, I'll ping you! SandyGeorgia (Talk) 23:24, 16 December 2025 (UTC)[reply]

{{Dead link|date=August 2023 |bot=InternetArchiveBot |fix-attempted=yes }}

https://buffalonews.com/news/local/another-buffalo-woman-accuses-harvey-weinstein-of-sexual-assault/article_1619126c-98bf-5cc5-a135-8b727d4bfb33

found: https://buffalonews.com/news/local/article_1619126c-98bf-5cc5-a135-8b727d4bfb33.html

might be a pattern

Piñanana (talk) 03:07, 16 December 2025 (UTC)[reply]

You probably want WP:URLREQ. Jay8g [V•T•E] 04:27, 16 December 2025 (UTC)[reply]

if there are multiple links like this then does this request still need to go to URLREQ, or does it need to go as bug report linked below? Gryllida 05:21, 16 December 2025 (UTC)[reply]

phab:T412778? you can login with your wikipedia account and subscribe. Gryllida 05:20, 16 December 2025 (UTC)[reply]

Gryllida: I think you are assuming IABot is designed to handle URL migrations. It can not. It only sees a dead link and adds a dead link tag. There is no bug IABot is doing what it is designed to do. URL move requests are made at URLREQ, which is handled by WaybackMedic, a different bot. -- GreenC 18:51, 16 December 2025 (UTC)[reply]

Thanks Gryllida 04:51, 17 December 2025 (UTC)[reply]

If you look at portals, like Portal:Math, they have click-to-purge buttons that update the page. Recently, I added a portal to User:ProcBot/PurgeList, which calls a bot to purge a page (daily, weekly, or however the editor specifies). I saw that no other portals are listed there. I'm going down a bit of a rabbit hole now. Are any other portals (which must be purged to update their rotating content) purged by some other bot? And if not, why do we use the click-to-purge links? I can't imagine most readers would (a) understand what it means, (b) feel comfortable doing it, and (c) then click to purge the portal. Rjjⁱⁱⁱ (talk) 12:35, 16 December 2025 (UTC)[reply]

Portals often display content that updates automatically through templates and subpages, but caching can cause outdated versions to remain visible for some time. The purge link exists as a simple manual mechanism to clear the cache so the portal immediately reflects the current state of its underlying content. It is mainly intended for editors maintaining the portal rather than for readers. ProcBot previously performed scheduled purges, but that task is now inactive. – DreamRimmer ■ 13:01, 16 December 2025 (UTC)[reply]

Purging of portals is mostly to change a random selection of content and not to ensure the portal is up to date. If a transcluded page is edited then a page is automatically updated (sometimes with a delay, see Help:Job queue) so even without requested purges, there will be updates. Portals generally don't have time-sensitive content so they don't need frequent purges. I'm not aware of any bots purging portals. I don't know how many readers do it but does it matter? If a reader does want to see other articles then the link saying "Refresh with new selections below (purge)" at Portal:Mathematics doesn't sound uncomfortable to me. Most readers may not realize it affects everybody. I have also heard there is a process which automatically reparses all pages sometimes but the interval may be long and I don't know whether it happens for all namespaces. PrimeHunter (talk) 13:12, 16 December 2025 (UTC)[reply]

ProcBot did not explicitly purge portals, but was used to purge specified pages or categories, and it is now inactive. When I work on the admin newsletter, I often need to purge the main page so that content transcluded from subpages updates. Based on that experience, I assumed that portals, which also transclude templates and subpages, may likewise need purging for changes to appear. I have not worked much with portals before, so my assumption could be wrong. – DreamRimmer ■ 13:22, 16 December 2025 (UTC)[reply]

Hmm, this is more the nuts and bolts than the rationale. I think, reading between the lines, the rationale is that the purge-to-update and click-to-purge method is good enough. I am thinking about updating documentation on deleting/automating old portals, and didn't want to write any bogus advice. I will probably start that discussion somewhere else, where editors interested in portals might better find it. Rjjⁱⁱⁱ (talk) 03:45, 18 December 2025 (UTC)[reply]

I'm trying to set up talk page archiving at User talk:AmandaNP/UAA/Blacklist, but it's not working. This page is 85 sections (including an instructions header that should not be archived) and 117 KB. –LaundryPizza03 (dc̄) 15:58, 16 December 2025 (UTC)[reply]

I have moved the instructions section heading to subpage to ensure it is not archived. Everything else appears to be in order, and the bot should be able to archive in its next scheduled run in around 19 hours. – DreamRimmer ■ 17:11, 16 December 2025 (UTC)[reply]

Alas, band timelines are code that I am not good playing with.

I am trying to make a significant change to this timeline. Basically I want the whole bottom section to go - everything from "ex-Batushka/Patriarkh" down to "MK", as it is presently showing two concurrent bands unnecessarily. I managed to get rid of the actual coloured horizontal bars, but everything else remains. The vertical yellow "split" line can stay.

Can someone help? — Czello ^(music) 16:11, 16 December 2025 (UTC)[reply]

Even with system and browser settings to reduce motion, the latest redesigns can still cause motion sickness.

1. Since the Vector Redesign, Wikipedia has had wide non-scrolling sidebars. If I don't hide these before paging down, the mismatch causes motion sickness. If I do hide these, then I get a header instead, and *this* header doesn't give me trouble, but each new page gets a new set of sidebars. I tried keeping Wikipedia pages zoomed-in to *continuously* hide these, but then I get a non-scrolling pop-up instead, so the mismatch still causes motion sickness. Ymmv.

2. Now, it also has another header about a temporary account. This causes a stutter effect as *this* header hides, with 2 short jumps instead of 1 long jump, when paging down.

I don't have a phab account to post a proper bug report there. ~2025-31543-88 (talk) 00:07, 17 December 2025 (UTC)[reply]

The sidebars should stay hidden, but you might have something about your browser's cookie settings causing that not to be the case. The issue is likely to go away if you register a permanent account, which look for your saved preferences rather than for a cookie. -- Tamzin^{[cetacean needed]} (they|xe|🤷) 02:28, 17 December 2025 (UTC)[reply]

have you tried timeless skin? Need a login or an URL param. Http://en.wikipedia.org/wiki/Zebra?useskin=timeless Regards, Gryllida 04:54, 17 December 2025 (UTC)[reply]

Or indeed any skin that isn't Vector 2022, that being the only one which has sticky sidebars. --Redrose64 🌹 (talk) 23:08, 19 December 2025 (UTC)[reply]

 Dislike Vector 2022. — DVRTed (Talk) 23:40, 19 December 2025 (UTC)[reply]

I am trying to get a little fancy with a Wikidata call at Template:Infobox Russian inhabited locality, but I haven't been able to make an embedded #expr work correctly. A little help with the /sandbox would be appreciated. – Jonesey95 (talk) 07:05, 17 December 2025 (UTC)[reply]

I would like to add a source to WP:RSP, but the page has hit a technical limit. Last month WP:RSP RFC 2025 was closed and I presume work is ongoing there, but in the meantime, is there any stop-gap or temporary mitigation that could be used to make additions to that list? Thanks! Smallangryplanet (talk) 12:21, 17 December 2025 (UTC)[reply]

@Smallangryplanet Probably should ask at WT:RSP. --Ahecht (TALK
PAGE) 18:20, 17 December 2025 (UTC)[reply]

Ah, ok--will do! I wasn't sure if there was a purely technical way around this. Thanks! Smallangryplanet (talk) 19:05, 17 December 2025 (UTC)[reply]

Did something change in the login flow recently? I know 2FA was just rolled out for everybody, but I've had that for years. The problem is, LastPass has stopped filling my username and password in the login screen. The values LastPass has for those fields are correct (they work when I manually copy-paste them), it's just that LP has stopped recognizing the screen and won't auto-fill the fields. RoySmith (talk) 14:17, 17 December 2025 (UTC)[reply]

This isn't "recent", but check if you have the right domain configured in your third party app. The webui logon page is no longer at en.wikipedia.org, but at auth.wikimedia.org. — xaosflux ^Talk 14:29, 17 December 2025 (UTC)[reply]

It doesn't look like the identifiers for those fields have changed in a long time. — xaosflux ^Talk 14:34, 17 December 2025 (UTC)[reply]

Yeah, that probably explains it. I have "keep me logged in for up to a year" checked, so for me "recent" means "within the last year" :-) RoySmith (talk) 14:34, 17 December 2025 (UTC)[reply]

Can anyone else see an issue with the page User talk:Pennsy22? Near the bottom around the section "Nominations now open for the WikiProject Military history newcomer of the year and military historian of the year", I can see the footer of the page displayed on top of the talk page content — Martin (MSGJ · talk) 16:40, 17 December 2025 (UTC)[reply]

I can't see the footer; the page content is on top of it in Vector 22 (Firefox on Linux). :D Izno (talk) 16:50, 17 December 2025 (UTC)[reply]

I found the footer – it's overlaid on top of section Speedy deletion nomination of Category:Royal Norwegian Navy ship infobox templates. —⁠andrybak (talk) 17:44, 17 December 2025 (UTC)[reply]

I found an unmatched <div> - has that fixed it? -- John of Reading (talk) 17:52, 17 December 2025 (UTC)[reply]

Looks fixed to me. Thank you. —⁠andrybak (talk) 18:05, 17 December 2025 (UTC)[reply]

Thanks everyone — Martin (MSGJ · talk) 18:11, 17 December 2025 (UTC)[reply]

I'm trying to modify a Wikipedia article using a bookmarklet and don't really know what I'm doing. I noticed the following (Firefox 146.0): if I go to any website other than Wikipedia, open the developer tools/Console, type in

document.open();
document.write('Hi');
document.close();

and run it, the website will be replaced by 'Hi', as expected. But on any Wikipedia page, whether logged-in or not, I get an empty page instead: nobody says 'Hi', it is sad. I thought maybe the text color is white, but that doesn't seem to be it, and then I ran out of ideas. AxelBoldt (talk) 00:46, 18 December 2025 (UTC)[reply]

This result is apparently predictable but I don't know why. Izno (talk) 01:42, 18 December 2025 (UTC)[reply]

Aha! Thank you. AxelBoldt (talk) 02:03, 18 December 2025 (UTC)[reply]

MediaWiki overrides the document.write function to do nothing, because people used to use it to load scripts in their common.js (like in that example Izno linked), until some changes to improve page load performance caused this call to replace the whole page instead of adding to it, making these people unable to access the site (T193191).

If you really want to replace the whole page, try this way instead:

document.documentElement.innerHTML = 'Hi';

Matma Rex talk 11:36, 18 December 2025 (UTC)[reply]

 You are invited to join the discussion at Help talk:Citation Style 1 § COinS pollution shouldn't be a problem. –CopperyMarrow15 (talk ⋅ edits) 04:42, 18 December 2025 (UTC)[reply]

Hi all, does anyone know why Legobot has stopped removing/archiving old closed discussions from WP:MFD? They're starting to pile up at the bottom. Left guide (talk) 17:15, 18 December 2025 (UTC)[reply]

See also: User talk:Legoktm#Legobot –Novem Linguae (talk) 14:33, 19 December 2025 (UTC)[reply]

I tried to undo my edit to {{extract short description}}. The edit filter won't let me, claiming the presence of an external link to about.com. 😭 Help? Paradoctor (talk) 18:21, 18 December 2025 (UTC)[reply]

I skimmed through the blacklist regexes but couldn't figure out why it's happening. Awaiting someone to the rescue. — DVRTed (Talk) 18:39, 18 December 2025 (UTC)[reply]

I managed to revert the edit via rollback, albeit I have zero idea what's wrong with the spam blacklist. Tenshi! (Talk page) 18:48, 18 December 2025 (UTC)[reply]

Thanks, that solves my immediate problem. Now I only need to find out why the same code that works at User:Paradoctor/extract SD refuses cooperation in it's rightful place. Paradoctor (talk) 18:51, 18 December 2025 (UTC)[reply]

If you ask the API for details of the parse for the previous revision: Special:ApiSandbox#action=parse&oldid=861127694&prop=externallinks it returns a long list of external links that supposedly exist in it, including one link to urbanlegends.about.com. I think this is because the template documentation page uses the page 13 (number) as an example, and those links appear there, and using the template technically transcludes that page. Matma Rex talk 18:53, 18 December 2025 (UTC)[reply]

This seems related to the bug that brought me there in the first place. I'll take it from here. Thanks to everyone for the assist.

Resolved

Paradoctor (talk) 19:00, 18 December 2025 (UTC)[reply]

The spam block list log [6] confirms it was http://urbanlegends.about.com/cs/historical/a/friday_the_13th.htm which is in 13 (number). Transcluding a page with a link shouldn't trigger the spam blacklist if the link isn't produced in the page you are trying to save, but the link is in ref tags which are treated specially in some situations so I guess this is such a situation. PrimeHunter (talk) 00:05, 19 December 2025 (UTC)[reply]

I have added notes to the documentation.[7] Maybe it should also be reported in Phabricator. PrimeHunter (talk) 12:33, 19 December 2025 (UTC)[reply]

(edit conflict) It might be the transclusion at User:Berkeczsolt/sandbox which does use the example code from the doc. Tenshi! (Talk page) 19:01, 18 December 2025 (UTC)[reply]

That page wasn't involved in the edit by Paradoctor so it looks irrelevant. PrimeHunter (talk) 00:05, 19 December 2025 (UTC)[reply]

Previous discussions:

• Wikipedia:Village pump (idea lab)/Archive 60 § Bot to block Proxy/VPN IPs (ST47ProxyBot replacement)
• Wikipedia:Village pump (technical)/Archive 217 § VPNgate blocking bot

ST47ProxyBot used to block open proxies obtained through a variety of sources. However, the owner, and consequently the bot, has retired, and we're seeing a proliferation of abusive proxies and VPNs used by a variety of spammers and LTAs. Sometimes, when I look at the block log of some of these IPs, I find that they've been previously blocked by ST47ProxyBot. I think someone (an admin, per WP:ADMINBOT) should resurrect ST47ProxyBot or create something new that does the same thing. I am aware that such bots may flood block logs, but this does seem necessary to prevent abuse. Children Will Listen (🐄 talk, 🫘 contribs) 18:54, 18 December 2025 (UTC)[reply]

This would also cause a lot of collateral damage (see this discussion on Meta). It's a tough balancing act. Graham87 (talk) 03:09, 19 December 2025 (UTC)[reply]

We still do put global blocks on open proxies, but usually reactively. If someone really wants to work on something like this reporting disruptive OP's for global blocks could be possible - but someone would need to want to work on it. — xaosflux ^Talk 13:23, 19 December 2025 (UTC)[reply]

We would need a CheckUser-level bot in that case, which is obviously a big no-no. Maybe something like this can be implemented into SuggestedInvestigations? Blanket blocking obvious proxies is still the best way to go, in my opinion. Children Will Listen (🐄 talk, 🫘 contribs) 15:05, 19 December 2025 (UTC)[reply]

There's also been some thought about moving proxy blocking from being done by bots to being done by software: phab:T380917. –Novem Linguae (talk) 16:52, 19 December 2025 (UTC)[reply]

A lesser-known feature of the new IP reputation variables is that one can access them at createaccount, and this can be used to log/block accounts potentially created using proxies. There are some holes with this, but it's definitely ready for a test run. Children Will Listen (🐄 talk, 🫘 contribs) 17:02, 19 December 2025 (UTC)[reply]

Probably runs into the same issues identified by Graham87 above. Many countries around the world have rather shockingly small pools of IPv4s available. And multiple users may entirely unknowingly share them within rather short periods of each other on mobile networks. All it takes is one person to run a proxy off their device or for one device on a network to be compromised and you have an awful lot of collateral. ~2025-41540-19 (talk) 17:10, 19 December 2025 (UTC)[reply]

Yeah, that's why we should start with a logging filter first, which can then be used by CheckUsers and other people to check if these accounts are behaving abusively. The more problematic proxies (e.g. VPNGate) can then be hardblocked using another filter to prevent excess collateral. Children Will Listen (🐄 talk, 🫘 contribs) 17:13, 19 December 2025 (UTC)[reply]

Logging is much more reasonable. However, I am extremely hesitant to allow direct filter blocks. Flag for human review? No issue. But setting filters to block directly seems likely to be a cause for all kinds of collateral which is just not worth the benefit. And I say this as someone who has dealt with some really nasty LTAs over the years, including the one you are probably thinking of. ~2025-41540-19 (talk) 17:28, 19 December 2025 (UTC)[reply]

@ChildrenWillListen, the vast majority of accounts are created and never ever used, for abuse or anything else. This would add a lot of work for the CU team without likely reducing much vandalism. -- asilvering (talk) 20:37, 19 December 2025 (UTC)[reply]

We're obviously not flagging everything, only a small subset of new accounts created from obvious proxies. We have an existing edit filter that does the same thing for all anonymous users and it gets hit once to a few times a day. Children Will Listen (🐄 talk, 🫘 contribs) 20:41, 19 December 2025 (UTC)[reply]

I was aware of your meaning. -- asilvering (talk) 20:45, 19 December 2025 (UTC)[reply]

If it still logs too much, we can have a bot that will move the users who had the filter triggered to a dashboard or similar when they make their first edit. Children Will Listen (🐄 talk, 🫘 contribs) 23:23, 19 December 2025 (UTC)[reply]

That bot blocked a lot of ranges unnecessarily. A bot that only blocked, say, VPNGate or something, I could get behind, but it's my sense that ST47's bot blocked a lot of ranges that no one was using abusively. We also do already have a list of probable proxies for admins to look at and block after checking for false positives. It's backlogged to hell, because there aren't many admins who make proxy blocks and we'd rather be doing something that actually stops ongoing disruption, rather than maybe-possibly-future disruption. -- asilvering (talk) 20:44, 19 December 2025 (UTC)[reply]

Yup. As someone who used to do these blocks & handle the appeal/exemption requests, the balance between abuse level and impact to normal users just isn't worth it for most of what ST47ProxyBot blocked, or for most of what User:AntiCompositeBot/ASNBlock lists. AntiCompositeNumber (they/them) (talk) 21:50, 19 December 2025 (UTC)[reply]

Sometimes when I middle-click on a notification, I get taken to the page the new comment was made on, but the "this is the comment you are looking for" highlighting causes the comment text to be completely unreadable by making the text and background the same dark blue color. This seems to be reproducible when going to e.g. [8] with dark mode already enabled. -- Beland (talk) 23:31, 18 December 2025 (UTC)[reply]

Thanks for sharing the link where the problem can be reproduced. Looks like this is the same problem as T371695 Reply highlighting is incompatible with skin-invert class in dark mode. Matma Rex talk 01:08, 19 December 2025 (UTC)[reply]

I've subscribed to that bug; thanks for the pointer! Based on the comments there, it looks like the colorization used by {{initiated}} is what's triggering it in this case. 8( -- Beland (talk) 02:17, 19 December 2025 (UTC)[reply]

Greetings. I have a proposal for a template for use in articles that would be based on a gadget. If I'm not mistaken, this would have to be a JavaScript gadget rather than a Lua module as it requires user input. Apologies if I'm in the wrong place.

The issue I want to solve is this: when there is a series of e.g. chronological information to be conveyed visually, the best current solution is to use an animated image. For example, see the animation of a chess game on Kasparov versus the World or the chronological map on Slave states and free states. Animated images can only present information at the rate of the frames in the image; it is inconvenient or inaccessible for readers who would prefer a different rate.

This could be solved using a template that has back and forth buttons (and/or a slider) to scroll through multiple still images. There is already one template that does a more specialized version of this: Template:Owidslider is designed to convey information from Our World in Data, and is designed for hoverable SVG maps that are listed in the template data. My proposal is for a similar template that would instead allow any image files, which could be inputted as parameters.

I have never coded a MediaWiki gadget before, so I would like some help in creating this. I would also like suggestions about how to implement it. — Vigilant Cosmic Penguin 🐧 (talk | contribs) 06:55, 19 December 2025 (UTC)[reply]

@Vigilantcosmicpenguin: {{Switcher}} uses MediaWiki:Gadget-switcher.js to make a list of labeled buttons for switching between selected content, usually images. It's not very practical to go sequentially through a long list of images. PrimeHunter (talk) 13:00, 19 December 2025 (UTC)[reply]

This can be done with MediaWiki:Gadget-calculator.js; doesn't require a new gadget. For the chessboard use case in particular, @Bawolff has done some experiments which I think almost reached a production-ready state. (I also did some experiments, but I forget what state I left it in.) – SD0001 (talk) 13:53, 19 December 2025 (UTC)[reply]

Ah, thanks. Glad to see that someone is already working on this. I might fork that gadget to create a version that works with images, like in the case of the chronological map. Thanks Bawolff! — Vigilant Cosmic Penguin 🐧 (talk | contribs) 19:15, 19 December 2025 (UTC)[reply]

Hi, glad its useful. Module:Sandbox/Bawolff/Chessboard is mostly in my userspace as nobody seemed that interested in using it in articles, but happy to move to template space if there is interest in using. Bawolff (talk) 22:22, 19 December 2025 (UTC)[reply]

For anyone else unfamiliar with the history: a gadget has existed for a long time (it's deployed on Hebrew Wikipedia and I believe Russian Wikipedia) and consensus was reached in 2019 to deploy it, but the interface admins disagreed with changing the site-wide Javascript file to add a hook for on-demand loading. The gadget's author went on to pursue the approach that had been recommended: creating an extension. But getting an extension deployed is very difficult. Now that the base MediaWiki software supports on-demand loading, the existing gadget could be deployed. However, it's understandable if the gadget's author is a bit burnt out from having tried for so long to get a gadget and then an extension deployed. isaacl (talk) 23:45, 19 December 2025 (UTC)[reply]

It really is a sad case about the chessbrowswer extension. They did literally everything right, the extension is beautiful, had induvidual wmf staff buy-in, and yet it was still rejected. Part of my goals with the calculator gadget was to make something generic enough that it could be used for a variety of things without asking permission each time. Bawolff (talk) 00:16, 20 December 2025 (UTC)[reply]

If I'm understanding correctly, it's not that the calculator gadget literally can also provide slideshow functionality. However the somewhat inaptly named mw:Template gadgets feature can be used to load a gadget only for pages with a specific category (typically added by a template), which is how the calculator gadget is loaded. This should make it easier to get new gadgets approved. isaacl (talk) 23:33, 19 December 2025 (UTC)[reply]

Having looked at the sandbox module after Bawolff's response to my comment regarding history, I stand corrected: although I don't understand the mechanism, it is using the calculator gadget as a front end to create buttons that trigger changes to the display. isaacl (talk) 01:49, 20 December 2025 (UTC)[reply]

(edit conflict) @Vigilantcosmicpenguin, Module:Sandbox/Bawolff/Chessboard seems to do what you want. — Qwerfjkltalk 13:54, 19 December 2025 (UTC)[reply]

I open en.wiki about two days ago and found my display had changed from Monobook to, well, something else. It is sparce and filled with whitespace. As the menus all changed, I fumbled around until I found a setting and changed it back to Monospace, but this keeps "going away" and here I am. Any suggestions? Maury Markowitz (talk) 12:32, 19 December 2025 (UTC)[reply]

There's a link at the bottom of every page that you can use to switch between desktop and mobile version: https://en.wikipedia.org/w/index.php?mobileaction=toggle_view_desktop Ponor (talk) 12:39, 19 December 2025 (UTC)[reply]

Hmmm, the link in question is called "desktop" which is not entirely obvious, I have never clicked it, and is so far down I don't think I've ever seen it. Thanks though, I'll know what to do next time. Maury Markowitz (talk) 18:16, 19 December 2025 (UTC)[reply]

Hi all and thanks in advance. Template:Infobox officeholder is generating bogus output at this sandbox.

She is on City Council, so I wrote | office1 = Member of the [[City Council of Kansas City, Missouri]] | district1 = 5th district but the infobox renders it as Member of the City Council of Kansas City, Missouri Member of the U.S. House of Representatives from 's 5th district district. She never had anything to do with federal government.

Am I correct that this is a known issue where this output is triggered by the presence of "district=" so that my only option is to replace it with | office1 = Member of the [[City Council of Kansas City, Missouri]]<br />from the 5th district? Thanks! — Smuckola(talk) 23:48, 19 December 2025 (UTC)[reply]

@Smuckola: Yes, the district parameters can only be used for certain legislatures. PrimeHunter (talk) 00:44, 20 December 2025 (UTC)[reply]