Wikipedia:Village pump (technical)
| Policy | Technical | Proposals | Idea lab | WMF | Miscellaneous |
If you want to report a JavaScript error, please follow this how-to page. Questions about MediaWiki in general should be posted at the MediaWiki support desk. Discussions are automatically archived after remaining inactive for 5 days.
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 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. |
Is it THURSDAY or is it Firefox
[edit]Background: Vector legacy/2010, Firefox Beta (freshly updated to 154.0b1), source editor
Issue: I signed on, edited a page, and thought it weird when I tabbed out of the edit box to add an edit summary, it instead shifted to the "Insert / Wiki markup / Symbols / etc" dropdown just below the edit box. I don't remember this happening this morning when I made a few edits but I also don't remember tabbing to add an edit summary. I definitely know it wasn't happening yesterday.
Is this a THURSDAY thing or is my freshly-updated FF beta not playing nice? Thanks for the thoughts. Primefac (talk) 21:47, 23 July 2026 (UTC)
- This might be autosuggest for the source editor which was rolled out earlier this week? Kowal2701 (talk, contribs) 22:26, 23 July 2026 (UTC)
- Well, I just on a hunch turned off the Syntax highlighter and things return to "normal" tab use, so it could be (though I'm not seeing any indication of an autosuggest in my editor). Primefac (talk) 22:38, 23 July 2026 (UTC)
- I cannot reproduce it using Firefox 150.0.1, so it is likely to be Firefox.析石父 (talk) 00:55, 24 July 2026 (UTC)
- I'm on 155.0a1 and can't reproduce it. Jerod Lycett (talk) 00:12, 26 July 2026 (UTC)
- Thumb won't preview, but enjoy. If it doesn't happen in Alpha then I guess they've fixed it... Primefac (talk) 00:53, 26 July 2026 (UTC)
- I'm on alpha and can reproduce, but I'm a couple weeks out of date... I'll update and report back. LittlePuppers (talk) 04:26, 26 July 2026 (UTC)
- I updated and it's still here. I don't tab to there though, so I have no idea when/why it changed. LittlePuppers (talk) 04:32, 26 July 2026 (UTC)
- I do think it's likely a Firefox thing, my laptop only updated yesterday and right after I noticed the change in behaviour. I filed at bugzilla so we'll see how things get on. Primefac (talk) 09:43, 26 July 2026 (UTC)
- Yeah, just tried in Edge and it doesn't happen there. Do you have a link to Bugzilla, for the curious? LittlePuppers (talk) 14:59, 26 July 2026 (UTC)
- https://bugzilla.mozilla.org/show_bug.cgi?id=2057769 Primefac (talk) 15:36, 26 July 2026 (UTC)
- Just to close the loop here, it was a Firefox regression and it looks like it's in the process of being patched. Thanks all for the thoughts. Primefac (talk) 22:19, 29 July 2026 (UTC)
- @Primefac or anyone else following, I'm on Chrome, not Firefox, and this has started happening to me today so I'm inclined to think it's a Thursday thing. When my connection is slow, the page hangs noticeably before all the highlighting appears. (I haven't changed anything in my preferences recent recently, I'm using the latest version of Chrome, but this didn't happen 24 hours ago). HJ Mitchell | Penny for your thoughts? 20:37, 30 July 2026 (UTC)
- That sounds like it's probably a separate issue. LittlePuppers (talk) 20:55, 30 July 2026 (UTC)
- Agreed; that's been happening for a little bit for me, just figured it was my computer being slow to run scripts. Primefac (talk) 22:37, 30 July 2026 (UTC)
- That sounds like it's probably a separate issue. LittlePuppers (talk) 20:55, 30 July 2026 (UTC)
- @Primefac or anyone else following, I'm on Chrome, not Firefox, and this has started happening to me today so I'm inclined to think it's a Thursday thing. When my connection is slow, the page hangs noticeably before all the highlighting appears. (I haven't changed anything in my preferences recent recently, I'm using the latest version of Chrome, but this didn't happen 24 hours ago). HJ Mitchell | Penny for your thoughts? 20:37, 30 July 2026 (UTC)
- Just to close the loop here, it was a Firefox regression and it looks like it's in the process of being patched. Thanks all for the thoughts. Primefac (talk) 22:19, 29 July 2026 (UTC)
- https://bugzilla.mozilla.org/show_bug.cgi?id=2057769 Primefac (talk) 15:36, 26 July 2026 (UTC)
- Yeah, just tried in Edge and it doesn't happen there. Do you have a link to Bugzilla, for the curious? LittlePuppers (talk) 14:59, 26 July 2026 (UTC)
- I do think it's likely a Firefox thing, my laptop only updated yesterday and right after I noticed the change in behaviour. I filed at bugzilla so we'll see how things get on. Primefac (talk) 09:43, 26 July 2026 (UTC)
- I updated and it's still here. I don't tab to there though, so I have no idea when/why it changed. LittlePuppers (talk) 04:32, 26 July 2026 (UTC)
- I'm on alpha and can reproduce, but I'm a couple weeks out of date... I'll update and report back. LittlePuppers (talk) 04:26, 26 July 2026 (UTC)
- I'm on 155.0a1 and can't reproduce it. Jerod Lycett (talk) 00:12, 26 July 2026 (UTC)
"not a valid map data page"
[edit]Robben Island Marine Protected Area and 926 other articles, at this writing, are displaying messages like mapframe: Title "Robben Island Marine Protected Area.map" is not a valid map data page. They seem to come from {{maplink}}. commons:Data:Robben Island Marine Protected Area.map exists, appears to be working fine, and if I had to guess, was probably working fine at Robben Island Marine Protected Area at some point in the past. I don't see any templates used in this page that have been changed in the last few days. {{Maplink}} was last modified in 2018, and Module:Mapframe was last modified on 1 June 2026.
At the bottom of the article, a series of red error messages is displayed instead of {{Marine protected areas of South Africa map}}, which looks fine on its Template page. The above problem is present in both Parsoid and the legacy parser. – Jonesey95 (talk) 05:21, 24 July 2026 (UTC)
- I've noticed that a bunch of other articles with Template:Maplink don't have this problem until the page is purged, then the error message appears and the map disappears. Steelkamp (talk) 05:29, 24 July 2026 (UTC)
- This probably means that a WP:THURSDAY MediaWiki code update causes the issue, and pages are showing the error message when they are purged or edited. The count of articles has increased from 926 to 941 just in the last 15 minutes, so it will probably go much higher. [Update: I found T433008.]– Jonesey95 (talk) 05:37, 24 July 2026 (UTC)
- The number has gone up to 1,269 Quake1234 (talk) 14:51, 24 July 2026 (UTC)
- Yes, and Category:Pages with broken maps has increased to 1,500 pages. I hope the developers work on Fridays. – Jonesey95 (talk) 16:47, 24 July 2026 (UTC)
- They don’t —TheDJ (talk • contribs) 18:18, 24 July 2026 (UTC)
- Well, it looks like a volunteer developer from de.WP has submitted a patch. We'll see if anyone is willing to test and apply it. Off topic: Then why TF would they deploy new software on Thursdays? When I worked in IT, we NEVER changed anything at the end of the week. I liked having my weekends free and knowing that my systems were working for my customers while I was out rock-climbing or whatever. – Jonesey95 (talk) 21:14, 24 July 2026 (UTC)
- To answer the question, the reason we get new software on Thursdays is to allow the WMF to do a staggered rollout: testwikis on Tuesday, non-Wikipedias on Wednesday, Wikipedias on Thursday. No I don't know why they don't move that one-day earlier but there's a trilemma between staggered rollouts, no end-of-week deployments, and a weekly release cadence all of which are good things which are mutually incompatible. * Pppery * in solidarity 23:32, 24 July 2026 (UTC)
- IIRC, the Tuesday start was to avoid running into Monday holidays and Monday deploys of fixes for bugs found over the weekend. As for the "no deploys on Fridays" rule, looking through some recent Fridays in wikitech:Server Admin Log I do see some Friday deploys, including some that are things I really wouldn't expect given the strict language on wikitech:Deployments/Emergencies (which does include "Major loss of functionality / appearance", BTW). Anomie⚔ 00:28, 25 July 2026 (UTC)
- Feels like one of those products people would naturally discuss on Reddit once more users discover it. ~2026-42024-47 (talk) 11:06, 29 July 2026 (UTC)
- To answer the question, the reason we get new software on Thursdays is to allow the WMF to do a staggered rollout: testwikis on Tuesday, non-Wikipedias on Wednesday, Wikipedias on Thursday. No I don't know why they don't move that one-day earlier but there's a trilemma between staggered rollouts, no end-of-week deployments, and a weekly release cadence all of which are good things which are mutually incompatible. * Pppery * in solidarity 23:32, 24 July 2026 (UTC)
- Since when don't they? Anomie⚔ 23:02, 24 July 2026 (UTC)
- Well, it looks like a volunteer developer from de.WP has submitted a patch. We'll see if anyone is willing to test and apply it. Off topic: Then why TF would they deploy new software on Thursdays? When I worked in IT, we NEVER changed anything at the end of the week. I liked having my weekends free and knowing that my systems were working for my customers while I was out rock-climbing or whatever. – Jonesey95 (talk) 21:14, 24 July 2026 (UTC)
- They don’t —TheDJ (talk • contribs) 18:18, 24 July 2026 (UTC)
- Yes, and Category:Pages with broken maps has increased to 1,500 pages. I hope the developers work on Fridays. – Jonesey95 (talk) 16:47, 24 July 2026 (UTC)
- The number has gone up to 1,269 Quake1234 (talk) 14:51, 24 July 2026 (UTC)
- This probably means that a WP:THURSDAY MediaWiki code update causes the issue, and pages are showing the error message when they are purged or edited. The count of articles has increased from 926 to 941 just in the last 15 minutes, so it will probably go much higher. [Update: I found T433008.]– Jonesey95 (talk) 05:37, 24 July 2026 (UTC)
- While we are here, can we break out Category:Articles with broken maps??? Ahecht you helped me do this with Category:Articles with script errors perhaps you can assist here? Zackmann (Talk to me/What I been doing) 23:22, 28 July 2026 (UTC)
- @Zackmann08: The bug has now been fixed so Category:Pages with broken maps is down to 143 pages and it has a link to search articles so I don't see a big need for a split. The search link currently gives too many articles because search hasn't updated fully after the fix but that's a temporary issue. Category:Pages with script errors and its namespace-specific subcategories have around 1,600 pages in total so a split is more helpful there. If we want to split Category:Pages with broken maps then MediaWiki:Kartographer-broken-category could propbably make similar code to MediaWiki:Scribunto-common-error-category. I say "probably" because some MediaWiki messages don't allow wikitext and it's hard to tell which ones without trying it. PrimeHunter (talk) 12:07, 29 July 2026 (UTC)
So who busted Template:Maplink?
[edit]The |from= parameter doesn't work for any given mapdata link, and everything else is intact. ⠀⠀ .n 03:30, 27 July 2026 (UTC)
- Related to phab:T433008, perhaps? Staraction (talk · contribs) 03:44, 27 July 2026 (UTC)
- hm yeah definitely :sob: ⠀⠀ .n 03:48, 27 July 2026 (UTC)
Map error on M-47 (Michigan highway)
[edit]For some reason, the map feature on M-47 (Michigan highway) is displaying the error mapframe: Title "M-47 (Michigan highway).map" is not a valid map data page
. Anyone know how to fix this? Ping @Imzadi1979: Ten Pound Hammer (they/them) • (What did I screw up now?) 04:23, 27 July 2026 (UTC)
Display better error message?
[edit]MediaWiki:Kartographer-error-title says: Title "$1" is not a valid map data page . That's a wrong error message at the moment and displayed in thousands of articles where many editors don't know what to do. They shouldn't do anything but just wait for the bug fix. Maybe we should customize the message until a fix is deployed. I suggest: "A bug prevents map display. A fix is underway." I have tested that the message allows wikilinks. PrimeHunter (talk) 03:24, 28 July 2026 (UTC)
- this broadly makes sense to me but i would suggest not putting the ticket in. while it may provide some contrete reassurance it also might result in people coming to the ticket to urge the hurring up of a bugfix they're hard at work on.... Morwen (talk) 04:00, 28 July 2026 (UTC)
- I agree with Morwen. Please implement this in plain text ASAP. It looks like a fix may be on the way in the next MediaWiki update; it is not clear why they did not roll out a patch more quickly, given the scale of the impact and the potential for editors removing things from pages in error. – Jonesey95 (talk) 14:22, 28 July 2026 (UTC)
- I have saved it as plain text. See e.g. the effect in Robben Island Marine Protected Area where the infobox previously said mapframe: Title "Robben Island Marine Protected Area.map" is not a valid map data page. MediaWiki:Kartographer-error-title should be deleted after the bug fix is deployed but first check that the fix has propagated so the searches "not a valid map data page" or "bug prevents map display" don't give results caused by the bug. PrimeHunter (talk) 15:32, 28 July 2026 (UTC)
- A bug fix has been deployed and appears to be working. The searches "not a valid map data page" and "bug prevents map display" still give some results but I haven't found any which actually display the eror message so it's just search which hasn't updated. Category:Pages with broken maps previously had 1000+ pages. Now there are only 143 and those I examined have other errors unrelated to the bug. PrimeHunter (talk) 11:48, 29 July 2026 (UTC)
- I have saved it as plain text. See e.g. the effect in Robben Island Marine Protected Area where the infobox previously said mapframe: Title "Robben Island Marine Protected Area.map" is not a valid map data page. MediaWiki:Kartographer-error-title should be deleted after the bug fix is deployed but first check that the fix has propagated so the searches "not a valid map data page" or "bug prevents map display" don't give results caused by the bug. PrimeHunter (talk) 15:32, 28 July 2026 (UTC)
- I agree with Morwen. Please implement this in plain text ASAP. It looks like a fix may be on the way in the next MediaWiki update; it is not clear why they did not roll out a patch more quickly, given the scale of the impact and the potential for editors removing things from pages in error. – Jonesey95 (talk) 14:22, 28 July 2026 (UTC)
Maplink with thumbnail error
[edit]I hope this belongs here but right after this issue has been fixed, I've tried opening maplink templates represented by a thumbnail (instead of just a link) and it only expands said thumbnail instead of the actual interactive map (eg. the infobox of North Carolina's 1st congressional district). Interestingly, if I append #/map/0 or #/map/1 after the URL, it opens up as normal. Indeed, if a link is used in lieu of a thumbnail (eg. "Interactive map version" on the map caption in North Carolina's congressional districts), it will go to the proper URL and interactive map as normal. –twotwofourtysix(talk || edits) 10:04, 1 August 2026 (UTC)
- I've noticed this problem too! Some maps work fine (e.g. Dream Island) but others will only work in editing previews. (e.g. 2026 Minab school attack) – MrPersonHumanGuy (talk) 18:45, 1 August 2026 (UTC)
- As it turns out, the still-working maps are transclusions of {{OSM Location map}}, which doesn't have the problem that {{maplink}} currently has. – MrPersonHumanGuy (talk) 18:58, 1 August 2026 (UTC)
Re: 429 'too many requests' on Media viewer
[edit]We talked about this two months ago, and for a while it seemed fixed, but it's been back now for a few days, rendering the media viewer unusable for me. The annoying part is that the desired image shows up for a fraction of a second, and only then is replaced with a broken image symbol. -- Seelefant (talk) Seelefant (talk) 09:32, 26 July 2026 (UTC)
- Last time it happened, the recommendation by User:JTweed-WMF was
If this is still happening, please copy the details on the error page and email them to bot-traffic
User:ASarabadani (WMF) also posted
wikimedia.org so that we can investigate further.You can create a private phabricator ticket or simply send the information to asarabadani
wikimedia.org. - JTweed and ASarabadani, are these methods of reporting HTTP status 429 errors still actual? —andrybak (talk) 09:31, 31 July 2026 (UTC)
- Yes, I would say an email to bot-traffic or a Phabricator ticket are most likely to be picked up. The ID on the error page helps us to track down why the limit is being exceeded, which can be for reasons like specific blocks put in place protect infrastructure, in addition to the rate limits. We're also currently working on changes that will allow us to have different rate limits for thumbnail and original images, which will hopefully allow us to relax the limits for thumbnail requests and make reader-visible problems less likely. JTweed-WMF (talk) 12:33, 3 August 2026 (UTC)
- @JTweed-WMF and @Ladsgroup, Is there a community facing Grafana dashboard or any similar metrics dashboard to monitor spikes in 429s served? Also, similarly do we have any community facing mechanism to get a general understanding of anonymized aggregrates of how many such human-but-blocked-please-unban requests are coming in to bot-traffic, private Phabricator tickets or y'all's personal Gmail accounts? Sohom (talk) 13:55, 3 August 2026 (UTC)
- @Sohom Datta The closes I can think of is this. Status of "int" basically means serving 429s (vs hit front, hit back, miss or pass). You can also split it based on cluster (upload being images, text being html requests). That's percentage, but you can turn it into actual number via looking up the total in the main dashboard. Ladsgroupoverleg 23:04, 3 August 2026 (UTC)
- @JTweed-WMF and @Ladsgroup, Is there a community facing Grafana dashboard or any similar metrics dashboard to monitor spikes in 429s served? Also, similarly do we have any community facing mechanism to get a general understanding of anonymized aggregrates of how many such human-but-blocked-please-unban requests are coming in to bot-traffic, private Phabricator tickets or y'all's personal Gmail accounts? Sohom (talk) 13:55, 3 August 2026 (UTC)
- Yes, I would say an email to bot-traffic or a Phabricator ticket are most likely to be picked up. The ID on the error page helps us to track down why the limit is being exceeded, which can be for reasons like specific blocks put in place protect infrastructure, in addition to the rate limits. We're also currently working on changes that will allow us to have different rate limits for thumbnail and original images, which will hopefully allow us to relax the limits for thumbnail requests and make reader-visible problems less likely. JTweed-WMF (talk) 12:33, 3 August 2026 (UTC)
Tech News: 2026-31
[edit]Latest tech news from the Wikimedia technical community. Please tell other users about these changes. Not all changes will affect you. Translations are available.
Updates for editors
Content Translation now supports dark mode, fulfilling a Community Wishlist request. This brings the tool in line with the accessibility features available in the Vector 2022 and Minerva skins, helping reduce visual fatigue for users translating content. - DiscussionTools' source mode and the 2017 wikitext editor will now offer autocomplete for links (
[[), templates ({{), HTML and parser tags (<), and magic words (__), making it quicker and easier to insert links, templates, and other wiki markup while editing. - The Readers Growth team has concluded its experiment with mobile page previews and will not roll out the feature. Page Previews are a pop-up bottom sheet that appears when readers tap a blue link, showing a thumbnail, lead paragraph, and an option to open the article. The experiment showed flat retention and negative indicator metrics, suggesting that mobile web readers preferred navigating directly to linked articles rather than using page previews.
- The Reader Experience team has seen encouraging early results from the Reading Lists feature, with 93% of participating users reporting that it was useful. Reading Lists help active readers save articles for future reading and support their learning goals on Wikimedia projects. The team plans further improvements before expanding the feature to more users.
- The Explore Feed Refresh initiative was tested with new and casual Wikipedia app readers. The refreshed feed helps readers discover new and relevant content. After a 10.5% increase in engagement with the feed, Wikimedia Apps team has decided to scale the Home Feed redesign to iOS with the learnings from the Android release applied.
View all 23 community-submitted tasks that were resolved last week. For example, an issue where subject names in the Article Guidance feature were displayed with incorrect capitalization on French Wikipedia, has now been fixed. Subject names will now follow the correct capitalization rules for the language.
Updates for technical contributors
- After running several Account Creation Experiments to improve registration completion rates, a new version of the username field on Create Account has been rolled out. It includes a popover summarizing the username policy to provide clearer guidance during account creation. As part of this change, the messages
createacct-helpusernameandcreateacct-username-helpthat several communities have configured will no longer be used. If communities want to customize the guidance shown in the new popover, they can instead edit the following messages:createacct-username-policy-popover-bullet1,createacct-username-policy-popover-bullet2, andcreateacct-username-policy-popover-bullet3. - Later this week, the CodeMirror syntax highlighter will offer themes. The themes can be picked from a dropdown menu in the full CodeMirror preferences dialog. For wikitext, available themes are default, colorblind-friendly (previously the colorblind preference option on Special:Preferences#mw-prefsection-editing) and no-highlighting. For code languages (i.e., CSS/JavaScript/JSON/Vue/Lua), there are several themes available. These same themes will eventually be available for wikitext, too.
- From now on, wikis can restrict editing in the "User" namespace to only the page owner and certain user groups. Read the configuration documentation to learn more.
- Detailed code updates later this week: MediaWiki
Tech news prepared by Tech News writers and posted by bot • Contribute • Translate • Get help • Give feedback • Subscribe or unsubscribe.
MediaWiki message delivery 18:47, 27 July 2026 (UTC)
- Re "From now on, wikis can restrict editing in the "User" namespace to only the page owner and certain user groups": If anyone here sees a proposal to make this happen on the English Wikipedia, please post a link to the discussion here at VPT. There would be significant technical ramifications to making this change on this site, so technically minded editors should be involved in any such discussion. – Jonesey95 (talk) 22:07, 27 July 2026 (UTC)
- Plenty of user subpages are treated as a shared resource, to editors' mutual benefit and without objection. Moving them to Wikipedia: namespace would be a possible but fiddly piece of bureaucracy. I can see an argument for protecting only top-level user pages, but that doesn't seem to be an option. Certes (talk) 18:16, 1 August 2026 (UTC)
- A minor issue with protecting top-level user pages is that some users copy-paste wikitext of userboxes, which ends up categorizing their top-level user pages as templates: newer examples and older examples. Judging by these two searches, there are at least 2-3 new cases of these per week. If top-level pages are protected, cleaning up these miscategorizations would be more complicated.
- A similar issue is violations of WP:DRAFTNOCAT, but I don't know how prevalent it is on top-level user pages. —andrybak (talk) 18:38, 1 August 2026 (UTC)
- Plenty of user subpages are treated as a shared resource, to editors' mutual benefit and without objection. Moving them to Wikipedia: namespace would be a possible but fiddly piece of bureaucracy. I can see an argument for protecting only top-level user pages, but that doesn't seem to be an option. Certes (talk) 18:16, 1 August 2026 (UTC)
DiscussionTools' source mode and the 2017 wikitext editor will now offer autocomplete for links (
– very nice, this makes working with templates in discussion much nicer, similar to how recent introduction of template category browser and favorite templates did for editing process. —andrybak (talk) 16:36, 1 August 2026 (UTC)[[), templates ({{), HTML and parser tags (<), and magic words (__), making it quicker and easier to insert links, templates, and other wiki markup while editing.- I've noticed a bug with the new feature though, it seems to break using the up and down arrows in the text box once the popup shows up (at least in Firefox 152), until a fresh mouse click "resets" it. So far I've been too lazy to put it into Phab though. Anomie⚔ 00:32, 2 August 2026 (UTC)
Link notifications
[edit]When you create an article, and another editor links that article at another page, you receive a "link notification" in your inbox, unless you opted out of it.
Now my question is, it is possible to receive these "link notifications" for any article, your perhaps interested in? Watching it doesn't seem to do anything. Fortek67 (talk) 11:09, 29 July 2026 (UTC)
- No, not possible. Izno (talk) 15:42, 29 July 2026 (UTC)
- But it has been requested before and I've just added a tracking link to the related Phabricator task (which I'm subscribed to). Graham87 (talk) 07:25, 30 July 2026 (UTC)
- Well thanks for linking it. Although I doubt it ever gets added since this thread is 12 years old. I wonder if a user script for something like this would be difficult to develop? Fortek67 (talk) 12:47, 30 July 2026 (UTC)
- But it has been requested before and I've just added a tracking link to the related Phabricator task (which I'm subscribed to). Graham87 (talk) 07:25, 30 July 2026 (UTC)
Blocked username is sometimes struck, sometimes not
[edit]
Recently Edittttor was blocked. I'm now seeing their username in strikeout style in many places, but not all. For example, in the attached screenshot from Template:Did you know nominations/Abdul El-Sayed, their username is not struck, but the corresponding talk link is. What's going on here? RoySmith (talk) 11:16, 29 July 2026 (UTC)
- Looks like the redlink uses
<a href="https://en.wikipedia.org/wiki/User:Edittttor?action=edit&redlink=1">, which isn't properly detected by the markblocked gadget. If I switch from Parsoid to the legacy parser, the link becomes<a href="/w/index.php?title=User:Edittttor&action=edit&redlink=1">instead, which is detected properly. So either Parsoid needs to be changed to emit /w/index.php links for redlinks, or the gadget needs to be updated to remove URL parameters from /wiki/ links. --rchard2scout (talk) 11:41, 29 July 2026 (UTC) - I think this is a bug in the markblocked gadget. Since User:Edittttor is a red link, MediaWiki adds
?action=edit&redlink=1to the URL. The gadget's regex doesn't stop at?, so it ends up trying to parseEdittttor?action=edit&redlink=1as the username, which fails, and the link gets skipped. The talk page link isn't a red link, so it doesn't get the extra query string and works fine. Changingconst articleRegex = new RegExp( mw.config.get( 'wgArticlePath' ).replace( '$1', '' ) + '([^#]+)' );toconst articleRegex = new RegExp( mw.config.get( 'wgArticlePath' ).replace( '$1', '' ) + '([^#?]+)' );should fix this. – DreamRimmer ■ 11:41, 29 July 2026 (UTC)- That can't be the whole story because the redlink is struck in the block log, and in edit histories, and in Wikipedia:Peer review/Abdul El-Sayed/archive1. RoySmith (talk) 11:49, 29 July 2026 (UTC)
- @MSantos (WMF): could you take a look at this? Is this a parsoid issue? RoySmith (talk) 11:54, 29 July 2026 (UTC)
- Red links starting with
/w/index.php?are handled correctly, but links starting with/wiki/are not. Wikipedia:Peer review/Abdul El-Sayed/archive1 renders links in the former format, which is why this happens. The change above should fix the issue for/wiki/links. – DreamRimmer ■ 12:05, 29 July 2026 (UTC)- But why is parsoid producing different output for these two cases when the old parser did not? RoySmith-Mobile (talk) 17:00, 29 July 2026 (UTC)
- @RoySmith It's not. Parsoid is enabled in the Template: namespace, which is why your first link doesn't work, but not in the Wikipedia: namespace, which is why your second link does (see Wikipedia:Village pump (technical)/Parsoid#Timeline). You can check at the bottom of the page: if it used Parsoid, it will say something like "This page was last edited on 29 July 2026, at 14:31. Page was rendered with Parsoid." --Ahecht (TALK
PAGE) 18:29, 29 July 2026 (UTC)- What??? Why are we running different parsers in different namespaces? RoySmith (talk) 18:31, 29 July 2026 (UTC)
- Parsoid is being rolled out over a period of weeks or maybe months to namespaces on the English Wikipedia. It is causing a bit of confusion and difficulty in troubleshooting. In addition, some back-end processes like categorization are being handled by either the new parser or the legacy parser, causing some pages to work fine when viewed but still show up in a category that is not listed on the page, or similar weirdness. See this technical page for some information(?) about testing that the developers are doing to see if pages look different in Parsoid and the legacy parser. It appears that they are using this output to schedule the rollout to other namespaces, although it is unclear how. Diligent editors at VPT and other places have noticed and reported bugs in Parsoid, some of which have been fixed since it started being the default parser for some pages about a month ago. – Jonesey95 (talk) 19:51, 29 July 2026 (UTC)
- Ugh. I knew it was being rolled out incrementally, but I wasn't paying attention to the details and assumed that meant wiki-by-wiki, not namespace-by-namespace. I guess there will be some pain for a while until this all stabilizes. Thanks for the explanation. RoySmith (talk) 20:07, 29 July 2026 (UTC)
- We rolled out Parsoid to all namespaces couple hours back. SSastry (WMF) (talk) 22:38, 3 August 2026 (UTC)
- Cool, thanks. I've been watching the parsoid project for a long time. It's good that it's gotten to this point. I'm sure there will continue to be pain as new problems emerge, but that's the price we have to pay for progress. RoySmith (talk) 22:44, 3 August 2026 (UTC)
- We rolled out Parsoid to all namespaces couple hours back. SSastry (WMF) (talk) 22:38, 3 August 2026 (UTC)
- Ugh. I knew it was being rolled out incrementally, but I wasn't paying attention to the details and assumed that meant wiki-by-wiki, not namespace-by-namespace. I guess there will be some pain for a while until this all stabilizes. Thanks for the explanation. RoySmith (talk) 20:07, 29 July 2026 (UTC)
- Parsoid is being rolled out over a period of weeks or maybe months to namespaces on the English Wikipedia. It is causing a bit of confusion and difficulty in troubleshooting. In addition, some back-end processes like categorization are being handled by either the new parser or the legacy parser, causing some pages to work fine when viewed but still show up in a category that is not listed on the page, or similar weirdness. See this technical page for some information(?) about testing that the developers are doing to see if pages look different in Parsoid and the legacy parser. It appears that they are using this output to schedule the rollout to other namespaces, although it is unclear how. Diligent editors at VPT and other places have noticed and reported bugs in Parsoid, some of which have been fixed since it started being the default parser for some pages about a month ago. – Jonesey95 (talk) 19:51, 29 July 2026 (UTC)
- What??? Why are we running different parsers in different namespaces? RoySmith (talk) 18:31, 29 July 2026 (UTC)
- @RoySmith It's not. Parsoid is enabled in the Template: namespace, which is why your first link doesn't work, but not in the Wikipedia: namespace, which is why your second link does (see Wikipedia:Village pump (technical)/Parsoid#Timeline). You can check at the bottom of the page: if it used Parsoid, it will say something like "This page was last edited on 29 July 2026, at 14:31. Page was rendered with Parsoid." --Ahecht (TALK
- But why is parsoid producing different output for these two cases when the old parser did not? RoySmith-Mobile (talk) 17:00, 29 July 2026 (UTC)
- That can't be the whole story because the redlink is struck in the block log, and in edit histories, and in Wikipedia:Peer review/Abdul El-Sayed/archive1. RoySmith (talk) 11:49, 29 July 2026 (UTC)
Grey out Minor edit tick box for TAs
[edit]
Moved from Wikipedia:Village pump (proposals)I have been editing as a TA for months, and using the Minor Edit tick box when it seemed appropriate. I have only recently realised that this setting is ignored for TAs. This is misleading, for people who are less likely to be familiar with editing Wikipedia. Could this box be greyed out for users for whom it will be ineffective? ~2026-42114-05 (talk) 04:09, 29 July 2026 (UTC)
- I seem to be missing something, the "minor" interface selector doesn't normally appear for logged out users. Can you describe all the steps you are taking that lead to that element showing? — xaosflux Talk 13:33, 29 July 2026 (UTC)
- Perhaps this is in one of the installed mobile apps? If so, which one? — xaosflux Talk 13:34, 29 July 2026 (UTC)
- @~2026-42114-05: This post is the only edit by your current TA. If you make an edit where you see and tick minor edit then we may be able to tell how you made the edit. PrimeHunter (talk) 14:11, 29 July 2026 (UTC)
- Check the related temporary accounts and you can see they all appear to be "Android app edit". Looking at the code, https://github.com/wikimedia/apps-android-wikipedia the minorEditCheckBox does not appear to have any conditional code to disable for TAs. A bug would need filling at https://phabricator.wikimedia.org KylieTastic (talk) 14:28, 29 July 2026 (UTC)
- A previous issue was raise as T123876 but is was merged into another task T296952 and I assume got lost. KylieTastic (talk) 14:31, 29 July 2026 (UTC)
- Check the related temporary accounts and you can see they all appear to be "Android app edit". Looking at the code, https://github.com/wikimedia/apps-android-wikipedia the minorEditCheckBox does not appear to have any conditional code to disable for TAs. A bug would need filling at https://phabricator.wikimedia.org KylieTastic (talk) 14:28, 29 July 2026 (UTC)
- @~2026-42114-05: This post is the only edit by your current TA. If you make an edit where you see and tick minor edit then we may be able to tell how you made the edit. PrimeHunter (talk) 14:11, 29 July 2026 (UTC)
- Perhaps this is in one of the installed mobile apps? If so, which one? — xaosflux Talk 13:34, 29 July 2026 (UTC)
- I opened phab:T433628 on this. — xaosflux Talk 18:48, 30 July 2026 (UTC)
Birthday problem
[edit]Trying to get a decent interactive widget for the birthday problem article and had a thought: rather than just generating a bunch of random valid month-day combinations, wouldn't it be more fun to populate an experiment with real people? So fetching birthdays of random notable people. Is there any way this is doable in an article? I could prepopulate a list to randomly select from, but in addition to that being less convincing of randomness, a good-sized list is going to seriously bloat the size of the page. I know there are strict limitations on what can be fetched within an article, and something like scanning for random biographies and then parsing the dates in articles is not realistic, but I feel like I've seen wikidata templates/tools that might of use? Looking for help brainstorming wikidata possibilities, basically. — Rhododendrites talk \\ 15:00, 29 July 2026 (UTC)
- Wow, I didn't know you could do stuff like that.
- Maybe some combination of
{{Random_page_in_category}}and {{Wikidata}}? I'm not sure how you'd check that they have birthday years listed, though. (Does your calculator gadget allow loops?) I'm also not sure what the performance/server load cost of that is. But at this point, if you come back in a couple hours I might have rabbit-holed myself into finding an answer for you.... LittlePuppers (talk) 15:48, 29 July 2026 (UTC)- Never mind, Random_page_in_category just provides a link... the wikidata template should be useful though. LittlePuppers (talk) 15:58, 29 July 2026 (UTC)
- Yeah, probably best to just have a list of a few hundred popular people—that'll also make it more likely that people recognize them. LittlePuppers (talk) 16:44, 29 July 2026 (UTC)
- Never mind, Random_page_in_category just provides a link... the wikidata template should be useful though. LittlePuppers (talk) 15:58, 29 July 2026 (UTC)
- I'm not really seeing anything in the wikibase API that would let you do this. :( Morwen (talk) 16:06, 29 July 2026 (UTC)
- I think in terms of performance, it would be much better to work with a pre-calculated list (it could be stored on-wiki (with maybe a set of them that is rotated through) and be updated from time to time, perhaps using a bot). To find birthdays of notable people, I suggest using the day of the year articles (for example, July 1). isaacl (talk) 16:35, 29 July 2026 (UTC)
- To be honest it would probably be better to curate the list anyway. You don't really want war criminals or serial killers appearing there do you? Morwen (talk) 16:39, 29 July 2026 (UTC)
- Ha. It's a good point. I was thinking about all the controversy we saw years ago over the Monty Hall problem. That's why there's something like a "transparency" feature built into that widget -- I'm assuming that at least some of the people who use it might think it's rigged. So I'm carrying that mindset over to the birthday problem, and figure that a skeptical reader would be skeptical about a prepopulated list. Might be overthinking it, though. Maybe the way around that is to do an academic move of deferring to an existing list or category. I'll look around for a candidate. — Rhododendrites talk \\ 16:49, 29 July 2026 (UTC)
- The names of the people aren't what generate skepticism, though: it's the randomness of selecting the birthdays. Since Wikipedia biographies are already a selected set from the general population, I don't think trying to make a random selection out of Wikipedia biographies will help much with that concern. I don't think there's any good way around the assumption that birthdays are randomly distributed. It may be instructive to provide a way to bias the selection, to show how the result is dependent on the assumption. For example, you could simulate the likelihood of duplication during the monthly birthday celebration for a large group of acquaintances. isaacl (talk) 17:15, 29 July 2026 (UTC)
- I'm just talking about numbers. Seems easy to imagine a skeptical person not buying a random selection of 30 people from a curated list of 40, but assuaged by a random selection of 30 people from all million+ biographies on the project. Getting into biases in the widget is a fun idea, but it's already getting quite big as-is. :) — Rhododendrites talk \\ 17:26, 29 July 2026 (UTC)
- And another thought - 99% of our biographies are people who might as well just have randomly generated names for all the name recognition they have. Try doing Special:RandomInCategory/Living people 20 times in a row some time and see how long many of them you've heard of. (I just did this myself and got one, although curiously that one was someone I've met in real life so go figure.) Morwen (talk) 17:32, 29 July 2026 (UTC)
- Yep, I got 0 for 20. My methodology would probably be take the top few hundred biographies (probably doesn't have to be living people) by page views, sort out those without listed birthdays or with war crimes, and go from there. That's probably much more interesting for the general audience, even if it doesn't assuage the skeptics. LittlePuppers (talk) 17:49, 29 July 2026 (UTC)
- I realized I was thinking about the problem from the other direction – finding representative names for the randomly selected birthdays – while you are thinking of gathering a set of names and then choosing randomly from them. I agree that it would be better to find lists of names for some suitably large categories where there shouldn't be any unnatural bias.(*) (There could be an option to choose between the different categories.) I still think not querying the database dynamically for names or birth dates is better for performance.
- (*) For example, as I recall I've read that the birth dates for NHL hockey players are biased by the way players are sorted into age groups during their youth, as the oldest players within an age group generally perform better against the others in the same group. isaacl (talk) 17:44, 29 July 2026 (UTC)
- And another thought - 99% of our biographies are people who might as well just have randomly generated names for all the name recognition they have. Try doing Special:RandomInCategory/Living people 20 times in a row some time and see how long many of them you've heard of. (I just did this myself and got one, although curiously that one was someone I've met in real life so go figure.) Morwen (talk) 17:32, 29 July 2026 (UTC)
- I'm just talking about numbers. Seems easy to imagine a skeptical person not buying a random selection of 30 people from a curated list of 40, but assuaged by a random selection of 30 people from all million+ biographies on the project. Getting into biases in the widget is a fun idea, but it's already getting quite big as-is. :) — Rhododendrites talk \\ 17:26, 29 July 2026 (UTC)
- The names of the people aren't what generate skepticism, though: it's the randomness of selecting the birthdays. Since Wikipedia biographies are already a selected set from the general population, I don't think trying to make a random selection out of Wikipedia biographies will help much with that concern. I don't think there's any good way around the assumption that birthdays are randomly distributed. It may be instructive to provide a way to bias the selection, to show how the result is dependent on the assumption. For example, you could simulate the likelihood of duplication during the monthly birthday celebration for a large group of acquaintances. isaacl (talk) 17:15, 29 July 2026 (UTC)
- Ha. It's a good point. I was thinking about all the controversy we saw years ago over the Monty Hall problem. That's why there's something like a "transparency" feature built into that widget -- I'm assuming that at least some of the people who use it might think it's rigged. So I'm carrying that mindset over to the birthday problem, and figure that a skeptical reader would be skeptical about a prepopulated list. Might be overthinking it, though. Maybe the way around that is to do an academic move of deferring to an existing list or category. I'll look around for a candidate. — Rhododendrites talk \\ 16:49, 29 July 2026 (UTC)
- To be honest it would probably be better to curate the list anyway. You don't really want war criminals or serial killers appearing there do you? Morwen (talk) 16:39, 29 July 2026 (UTC)
- If you only need to select a relatively small number of names from a larger list, and you are ok with the selection only changing each time the page is parsed, you can put the list in lua. I did something similar once with a template on commons to show a random selection of quality images. Only the items selected go into the page, but its a different selection every time the page gets parsed. Bawolff (talk) 18:28, 29 July 2026 (UTC)
- I made a demo of what I mean at Module:Sandbox/bawolff/Birthday. It gives results like: Maria Tallchief was born on Jan 24. Lev Landau was born on Jan 22. John von Neumann was born on Dec 28. Bawolff (talk) 21:01, 29 July 2026 (UTC)
- Try Category:2025 deaths (11,889). There are plenty to pick from and their birth dates should be truly random. --Redrose64 🌹 (talk) 18:45, 29 July 2026 (UTC)
User:Mathbot needs a new owner
[edit]Crossposting at Wikipedia:Bots/Noticeboard#User:Mathbot needs a new owner for best coverage. User:Oleg Alexandrov has asked in a recent discussion:
Are there any poeple willing to maintain this bot? Last time I worked with Dreamrimmer who is listed as co-owner at https://toolsadmin.wikimedia.org/tools/id/mathbot. If he is around, or somebody else is willing to help adopt the bot, that would be great. It has been too long and I don't visit much or remember how things work anymore.
Note that Mathbot is a fairly important bot, as it cues up the new pages for AfD and maintains other aspects of AfD. BD2412 T 15:09, 29 July 2026 (UTC)
- Mathbot co-maintainer here. I'm around if you need anything. – DreamRimmer ■ 15:37, 29 July 2026 (UTC)
- @DreamRimmer: Mathbot's instructions currently say for questions to be asked on Oleg's talk page. Would you be okay with that instruction being changed to your talk page? BD2412 T 19:08, 29 July 2026 (UTC)
Sortable table not sorting for one column
[edit]At List of Irish counties by population, the table's "Change per year (%)" column doesn't sort some values correctly. Specifically those with one decimal place (most values have two decimal places). However when I looked at another similar table, List of Japanese prefectures by population#Prefectures of Japan ranked by population as of October 1, 2020, the "% Change" column seems to work fine for its mix of one and two decimal places. The main difference I can see between the two tables is the Irish table's use of Template:increaseNeutral on values in those columns, which the Japan table doesn't use. Thanks for any ideas, Declangi (talk) 00:04, 30 July 2026 (UTC)
- The software was guessing the data sort type was string because of the templates you identified. I've both forced it to sort by number and removed the templates. You can add separate columns or forgo them. Izno (talk) 03:04, 30 July 2026 (UTC)
- That is excellent work by you, many thanks and all looks well there now. Given that all the values in those columns are positive, not much need for embellishment. Declangi (talk) 04:08, 30 July 2026 (UTC)
- Decimals would sort correctly with string sort. This was something different. The table used {{increaseNeutral}} which adds an image with alt text. Table sorting appears to have issues if images have alt text. The third column below has the former combination in the article. Maybe it for some reason sorts the part after the decimal point as a number by itself where 8 < 75.
- Speculation without looking at the sort code: alt text is included in the text to sort (this appears consistent with results if the alt texts are different). Without a data-sort-type it treats the cell as a string but sorts digit substrings by number if the text before the number is identical. The decimal point is assumed to be a sentence-ending period so the number right after the "period" is treated by itself.
No alt text
No data-sort-type
Sorts correctlyNo alt text
data-sort-type=number
Sorts correctlyalt text
No data-sort-type
Sorts incorrectlyalt text
data-sort-type=number
Does not sort at all
0.75 0.75 0.75 0.75 0.95 0.95 0.95 0.95 0.8 0.8 0.8 0.8- PrimeHunter (talk) 11:23, 30 July 2026 (UTC)
- I found phab:T42044: "tablesorter should sort images with alt text". So it's deliberate that alt text is included in sorting. Most users don't see or hear alt text and I guess almost no editors will think of alt text when making sortable tables so this seems unfortunate. In the present case it managed to break sorting even though all the alt texts were identical. PrimeHunter (talk) 11:41, 30 July 2026 (UTC)
- Whether editors not using assistive technology consider the image probably depends on whether they consider the image as being semantic or extra/decorative. For the {{IncreaseNeutral}}/{{DecreaseNeutral}} style images, for example, it might depend on whether the image is indicating "this absolute value increased/decreased" versus "the value is an amount-changed and the image replaces the + or −". Anomie⚔ 15:16, 30 July 2026 (UTC)
- If an image is the only content in a cell then it may be helpful to include it in sorting, for example causing identical icons to be grouped although the order of the groups may often appear arbitrary to a viewer. But if a cell both has an image and text then I think sorting should ignore the image, or maybe use it as a tiebreaker if cells otherwise have identical text. PrimeHunter (talk) 11:57, 30 July 2026 (UTC)
- I found phab:T42044: "tablesorter should sort images with alt text". So it's deliberate that alt text is included in sorting. Most users don't see or hear alt text and I guess almost no editors will think of alt text when making sortable tables so this seems unfortunate. In the present case it managed to break sorting even though all the alt texts were identical. PrimeHunter (talk) 11:41, 30 July 2026 (UTC)
Display SVG not display
[edit]I tried a upload a File:Panther Creek High School (Frisco, Texas) logo.svg but nothing displayed but it works with my Brave browser. You can download logo Panther Creek High School Frisco Texas. Ranch9613 💬 03:41, 30 July 2026 (UTC)
- @Ranch9613: It looks like this SVG is failing to render, possibly due to the use of matrix() in the
gradientTransformattributes. This could be due to this librsvg bug. Although, isn't a fair use image meant to be low-resolution? i.e. a lower resolution raster version of this logo might be required anyway, and switching to that will resolve the SVG problem. SWilson (WMF) (talk) 04:19, 30 July 2026 (UTC)- We provide logos pretty wide latitude from a non-free use perspective. Izno (talk) 14:06, 30 July 2026 (UTC)
- I hope it's not time for another argument over non-free SVG "resolution" versus "level of detail". Anomie⚔ 15:19, 30 July 2026 (UTC)
- I was just going on the text given for {{Non-free use rationale 2}}, about
a low resolution image of an organization's logo in the article
. But of course the image used in the article is 250px, so it is low-res. Not re-opening any argument! :-) SWilson (WMF) (talk) 05:52, 31 July 2026 (UTC)
- I was just going on the text given for {{Non-free use rationale 2}}, about
- I hope it's not time for another argument over non-free SVG "resolution" versus "level of detail". Anomie⚔ 15:19, 30 July 2026 (UTC)
- We provide logos pretty wide latitude from a non-free use perspective. Izno (talk) 14:06, 30 July 2026 (UTC)
Script validation
[edit]Two scripts have been advertised at ITN recently:
- User:BilledMammal/DiscussionPlus.js – which starts various types of discussion
- User:Thilio/ITNHelper – which posts ITN nominations and !votes
The source code for these warns that
Code that you insert on this page could contain malicious content capable of compromising your account. ... User scripts are not centrally supported and may malfunction or become inoperable due to software changes. ... If you are unsure whether code you are adding to this page is safe, you can ask at the appropriate village pump.
This is the village pump page. So is this code safe? What testing and validation has been done? What's to stop anyone changing the scripts at any time? Note the recent security incident which indicates that such scripts are a serious risk and vulnerability.
Andrew🐉(talk) 13:52, 30 July 2026 (UTC)
- This message displays for all scripts that you might use. It is good for you to ask these questions. In some order:
- Someone needs to assess whether the scripts are safe. This is not a usual activity, and is in the domain of things other volunteers might do. Random volunteers at VPT do not usually do so, but I suppose you can ask here.
- It usually takes a deliberately hostile developer to make a script that will be malicious at the end of the day, and their efforts are usually (usually) easy to identify. From this perspective, the only testing and validation that actually needs to be done is usually on the part of the script developers in the course of reasonable script development.
- Scripts in a specific user space can be modified only by WP:INTADMINs and the user named. Scripts in MediaWiki space can be modified only by interface admins. But otherwise, yes, they can be modified at any future date. (There are ways to avoid taking updates automatically if you review a good version and decide to use that version, but you will miss good faith updates such as bug fixes and feature additions without deliberately taking time to review the scripts routinely.) Izno (talk) 14:14, 30 July 2026 (UTC)
Rater dialog is unresponsive
[edit]Is anyone else experiencing an unresponsive Rater dialog? I've reported the issue, but I'd like to see if I'm not alone here. the Stefen 𝕋ower 21:40, 30 July 2026 (UTC)
- You already raised this at User talk:Evad37/rater.js#Rater dialog is unresponsive. --Redrose64 🌹 (talk) 22:31, 30 July 2026 (UTC)
- Indeed. I posted here additionally on purpose per the second clause of my last sentence, as this page has a lot more eyes. the Stefen 𝕋ower 00:52, 31 July 2026 (UTC)
Talk page title blank character space
[edit]I am seeing a blank space between 'Talk:' and the article name on all talk pages, so Talk:Supermarine Spitfire appears as Talk: Supermarine Spitfire. I noticed it because I create many archive talk pages and went to create one just now, I always check the new title thoroughly for spelling and spacing mistakes before hitting enter,
It might be related to the recent wiki-wide appearance changes of section headers on talk pages, I have not discovered where that originated from or if it is just a skin change. I won't create any more archive pages until this is resolved, it may be ok as is but I'd rather not have to delete pages and recreate them. Nimbus (Cumulus nimbus floats by) 11:17, 31 July 2026 (UTC)
- Nimbus227, it's just a skin change. It was done a while ago to better delineate the namespace in the page title (you'll note that this page title also has a space after Wikipedia:). — Qwerfjkltalk 11:34, 31 July 2026 (UTC)
- Ok, thanks. Notices of these changes don't seem to filter through to projects or users, perhaps I missed it. Nimbus (Cumulus nimbus floats by) 12:15, 31 July 2026 (UTC)
- It’s a change introduced by discussion tools which only applies to talk pages (and pages in other namespaces used as talk pages), see Wikipedia:Village pump (technical)/Archive 231#Discussion Tools usability improvements becoming default. Johannnes89 (talk) 12:35, 31 July 2026 (UTC)
- Thanks, that's exactly what I needed. I have reverted the change through my preferences. Nimbus (Cumulus nimbus floats by) 13:44, 31 July 2026 (UTC)
- @Nimbus227: It's impossible to create a page with a space after the namespace colon in the actual page name so just create pages as usual whether or not you change preferences. The HTML says
<span class="mw-page-title-separator">:</span>, and the site CSS inserts a space with this rule: .ext-discussiontools-visualenhancements-enabled .mw-page-title-separator::after { content: ' '; }
- If you only want to remove the space then this in your CSS will do it:
.mw-page-title-separator::after { content: '' !important; }
- Some people may complain about using
!importantbut it's OK when it's in your own CSS and not imposed on others. I used it to ensure it works everywheremw-page-title-separatoris used which may change in the future. PrimeHunter (talk) 13:50, 31 July 2026 (UTC)- I have managed to do many impossible things in my time on Wikipedia!! I create the archive pages by copying the parent talk page title and pasting it into the search box and adding /Archive 1 then hitting 'go' which produces the red link for creation. I tried this earlier and the space is automatically removed on pasting it in to the search box. It's academic now as I've edited my preferences to return to the previous appearance. I appreciate functionality changes but I am a Luddite and will continue to view the talk page history for the date of its latest comments. Nimbus (Cumulus nimbus floats by) 14:07, 31 July 2026 (UTC)
- @Nimbus227 The space is actually never copied in the first place since it's a CSS feature and not part of the text. --Ahecht (TALK
PAGE) 14:46, 31 July 2026 (UTC)- I was not to know that, that's why I asked for advice here. This kind of problem could be avoided by better publicity of changes to default settings/behavior IMHO, a single post to wikiproject talk pages would be enough, most long term editors have those on their watchlists. Nimbus (Cumulus nimbus floats by) 15:11, 31 July 2026 (UTC)
- @Nimbus227 The announcement was posted here, which is the proper venue for technical changes. Wikiproject talk pages are for discussing the topic covered by that wikiproject, not larger wikipedia administration. The rollout of Discussion Tools was also discussed in Tech News last November for people that subscribe to it, although I think it should've included an item in a more recent issue for the English Wikipedia rollout. --Ahecht (TALK
PAGE) 16:41, 31 July 2026 (UTC)- I agree it doesn't belong on various pages just to spread awareness, but there are always complaints when something changes. I have long considered to start a page about how to revert interface changes (including with personal CSS and scripts), or say if it's not possible. Would there be interest in that? It would include all types of causes, also users who changed preferences wihout knowing the consequences, accidentally switched between desktop and mobile, need to update a browser or uninstall a bad extension, and so on. There would also be general tips like trying a cache bypass, or trying safemode, logout, another wiki or another browser to narrow down possible causes. PrimeHunter (talk) 16:55, 31 July 2026 (UTC)
- I don't see a particular issue with a one stop shop. Most of this information does already exist, just without a central pointer. Izno (talk) 00:55, 1 August 2026 (UTC)
- Help:User style is one place with a number of customizations, but it really could use reworking by separating different style options to simplify finding and applying specific solutions.
- (I have an idea about making a script that would allow you to check an option in one column, and the corresponding CSS rules would be populated in a box on the side that you could copy and paste into your common.css file. (In theory it could automatically update your common.css file (or common.js file if the tool were extended to handle Javscript customizations), but baby steps first.) But it's just at the conceptual stage, so I wouldn't want anyone to hold off on reworking Help:User style or creating any new help page.) isaacl (talk) 03:22, 1 August 2026 (UTC)
- @PrimeHunter Creating a page like that would be very helpful. David10244 (talk) 23:04, 2 August 2026 (UTC)
- I agree it doesn't belong on various pages just to spread awareness, but there are always complaints when something changes. I have long considered to start a page about how to revert interface changes (including with personal CSS and scripts), or say if it's not possible. Would there be interest in that? It would include all types of causes, also users who changed preferences wihout knowing the consequences, accidentally switched between desktop and mobile, need to update a browser or uninstall a bad extension, and so on. There would also be general tips like trying a cache bypass, or trying safemode, logout, another wiki or another browser to narrow down possible causes. PrimeHunter (talk) 16:55, 31 July 2026 (UTC)
- @Nimbus227 The announcement was posted here, which is the proper venue for technical changes. Wikiproject talk pages are for discussing the topic covered by that wikiproject, not larger wikipedia administration. The rollout of Discussion Tools was also discussed in Tech News last November for people that subscribe to it, although I think it should've included an item in a more recent issue for the English Wikipedia rollout. --Ahecht (TALK
- I was not to know that, that's why I asked for advice here. This kind of problem could be avoided by better publicity of changes to default settings/behavior IMHO, a single post to wikiproject talk pages would be enough, most long term editors have those on their watchlists. Nimbus (Cumulus nimbus floats by) 15:11, 31 July 2026 (UTC)
- @Nimbus227 The space is actually never copied in the first place since it's a CSS feature and not part of the text. --Ahecht (TALK
- I have managed to do many impossible things in my time on Wikipedia!! I create the archive pages by copying the parent talk page title and pasting it into the search box and adding /Archive 1 then hitting 'go' which produces the red link for creation. I tried this earlier and the space is automatically removed on pasting it in to the search box. It's academic now as I've edited my preferences to return to the previous appearance. I appreciate functionality changes but I am a Luddite and will continue to view the talk page history for the date of its latest comments. Nimbus (Cumulus nimbus floats by) 14:07, 31 July 2026 (UTC)
- @Nimbus227: It's impossible to create a page with a space after the namespace colon in the actual page name so just create pages as usual whether or not you change preferences. The HTML says
- Thanks, that's exactly what I needed. I have reverted the change through my preferences. Nimbus (Cumulus nimbus floats by) 13:44, 31 July 2026 (UTC)
- It’s a change introduced by discussion tools which only applies to talk pages (and pages in other namespaces used as talk pages), see Wikipedia:Village pump (technical)/Archive 231#Discussion Tools usability improvements becoming default. Johannnes89 (talk) 12:35, 31 July 2026 (UTC)
- Ok, thanks. Notices of these changes don't seem to filter through to projects or users, perhaps I missed it. Nimbus (Cumulus nimbus floats by) 12:15, 31 July 2026 (UTC)
I would think that the percentage of editors with wiki technical sub-pages on their watchlist is very low. A user communication option could be use of watchlist banners that can be dismissed after being read, this is used for RfA nominations and for new editions of Signpost. Nimbus (Cumulus nimbus floats by) 17:43, 31 July 2026 (UTC)
- There are too many changes that run the gamut from stylistic (this one) to potentially breaking (the rollout of Parsoid) to communicate using the watchlist et al. If you do not want to watch this page where the tech news carrying the notification was published, you can subscribe directly a page of your choice at meta:Global message delivery/Targets/Tech ambassadors. Izno (talk) 00:54, 1 August 2026 (UTC)
- @Nimbus227 and Qwerfjkl: It's not a skin change. It's a side-effect of Discussion Tools: (i) if you disable that feature, the apparent space vanishes regardless of your skin; (ii) if it's enabled, the space is seen on all skins, even unmaintained ones like Cologne Blue and Modern. --Redrose64 🌹 (talk) 10:00, 1 August 2026 (UTC)
- Redrose64, yes, I was mistaken. I've been opted into the beta feature for a while, and I vaguely remembered the VPT post about it. — Qwerfjkltalk 12:20, 1 August 2026 (UTC)
Relief maps
[edit]Is this just some screw up at my end (I did get an update this morning and Windows 11 is somewhat eccentric)? At pages like Southampton Island I'm not getting a relief map. I didn't notice this yesterday and it seems consistent throughout browsers (Chrome, Firefox, Edge). Doing this solves the problem. CambridgeBayWeather (#1 deranged), Uqaqatigijaa (talk), Huliva 16:51, 31 July 2026 (UTC)
- The same problem exists on a second computer just updated. CambridgeBayWeather (#1 deranged), Uqaqatigijaa (talk), Huliva 16:53, 31 July 2026 (UTC)
- Probably Special:Diff/1364530773/1366999373. @Hike395? Ponor (talk) 16:58, 31 July 2026 (UTC)
- I've asked Hike395 to respond here. CambridgeBayWeather (#1 deranged), Uqaqatigijaa (talk), Huliva 17:03, 31 July 2026 (UTC)
- Will investigate — hike395 (talk) 00:42, 1 August 2026 (UTC)
- I wrote new code in Module:Location map to fix a problem introduced by the
pushpin_mapcleanup by Zackmann08. My new code did not correctly handle the case whenimage1was missing. It was a simple code fix, now live.
Fixed — hike395 (talk) 01:25, 1 August 2026 (UTC)
- Thanks. CambridgeBayWeather (#1 deranged), Uqaqatigijaa (talk), Huliva 15:07, 1 August 2026 (UTC)
- I wrote new code in Module:Location map to fix a problem introduced by the
- Will investigate — hike395 (talk) 00:42, 1 August 2026 (UTC)
- I've asked Hike395 to respond here. CambridgeBayWeather (#1 deranged), Uqaqatigijaa (talk), Huliva 17:03, 31 July 2026 (UTC)
- Probably Special:Diff/1364530773/1366999373. @Hike395? Ponor (talk) 16:58, 31 July 2026 (UTC)
_logo.svg){kind=link}
_logo.svg/960px-Panther_Creek_High_School_(Frisco%2C_Texas)_logo.svg.png){kind=link}
What is breaking the infobox at MetLife Stadium?
[edit]See Special:Diff/1367094907 for the diff in where I fixed the issue. When this is not fixed, above the stadium image, it displays "[[File:|250px|MetLife Stadium is located in New York City|class=notpageimage noviewer]]" and the label "MetLife Stadium" in a white background in the center of the broken text. And this is probably because of the | pushpin_map = parameter, which is probably breaking the infobox. What else is causing this, some template? — (In solidarity 🫡) SimpleObjects-9ei 🏖️/☀️/🥵 (see talk) 00:06, 1 August 2026 (UTC)
- Something is not configured right. I couldn't say what. I've backed it up to New Jersey because that is clearly not broken. Izno (talk) 00:32, 1 August 2026 (UTC)
- Looks like it's Special:Diff/1364530773/1366999373, same as #Relief maps above. Anomie⚔ 01:14, 1 August 2026 (UTC)
- Fixed see above. — hike395 (talk) 01:25, 1 August 2026 (UTC)
Is it possible to mark my user essay as archived?
[edit]I noticed an editor randomly started editing my user essay User:guninvalid/Naming conventions of non-democratic elections despite the fact that the essay has been dormant since November and has no consensus. I'd like to mark this essay as archived. Is that possible? Just using Template:Archive doesn't feel quite right, but are there better optionsm guninvalid (talk) 17:07, 31 July 2026 (UTC)
- There are {{Historical}} and {{Failed proposal}}. Nardog (talk) 17:59, 31 July 2026 (UTC)
Television series vs web series
[edit]Hello. I'd like to inquire about consistency regarding the spelling of television series vs. web series. I've noticed some pages use the term "television series," but in the cast's filmography, the series is listed under "web series." For example, the page Knock-Off (TV series) lists the television series, but the page Jo Bo-ah lists the television series under the web series section. So, which term should I use? Television series or web series? WillsonEP09 (talk) 04:13, 1 August 2026 (UTC)
- @WillsonEP09 This doesn't seem to be a technology question ? —TheDJ (talk • contribs) 13:07, 1 August 2026 (UTC)
- @WillsonEP09: It could be discussed at Wikipedia talk:WikiProject Television but try to be more precise. I don't know what you mean by "lists" in "Knock-Off (TV series) lists the television series". If you mean "TV series" in the article name then it doesn't say "television" and it isn't a list. "Television series" vs. "web series" is not about spelling but wording. PrimeHunter (talk) 16:27, 1 August 2026 (UTC)
- To expand on TheDJ's comment, this page is for discussion on using the MediaWiki software underpinning Wikipedia, and not for questions regarding page content. Content-related questions that cover a broad category of articles but within a specific area of interest can be posed at the talk page of an appropriate active WikiProject, as mentioned by PrimeHunter. isaacl (talk) 16:36, 1 August 2026 (UTC)
- @TheDJ @PrimeHunter @Isaacl I mean why is Knock-Off (TV series) on Jo Bo-ah's filmography list the series listed under the web series section? WillsonEP09 (talk) 12:17, 2 August 2026 (UTC)
Is there a way to make topics templates on portal pages display with outlines?
[edit]For example if you scroll down to the "Topics" section on Portal:European Union#Topics then it isn't displayed the same as on Template:European Union topics. I find the latter more readable but I guess others might not prefer for it to display like that on portal pages.
I was able to go into devtools, select a tr, th, or td element, and then disable the following CSS to make the table display normally.
.mw-parser-output .PlainNavboxes .navbox, .mw-parser-output .PlainNavboxes .navbox th, .mw-parser-output .PlainNavboxes .navbox tr, .mw-parser-output .PlainNavboxes .navbox td, .mw-parser-output .PlainNavboxes .navbox-title, .mw-parser-output .PlainNavboxes .navbox-subgroup, .mw-parser-output .PlainNavboxes .navbox tr+tr>.navbox-abovebelow, .mw-parser-output .PlainNavboxes .navbox tr+tr>.navbox-group, .mw-parser-output .PlainNavboxes .navbox tr+tr>.navbox-image, .mw-parser-output .PlainNavboxes .navbox tr+tr>.navbox-list, .mw-parser-output .PlainNavboxes .navbox-even, .mw-parser-output .PlainNavboxes .navbox-odd, .mw-parser-output .PlainNavboxes .navbox span, .mw-parser-output .PlainNavboxes .navbox abbr {
background-color: transparent!important;
color: inherit!important;
border-color: transparent!important;
box-shadow: none!important;
}
But the only applies until the page is reloaded. I have the Stylus browser extension which I can use to apply custom CSS to websites which might help, like I could at least get some basic borders with it, but I wonder if there's a way to make these tables just display normally. BoardHum (talk) 06:36, 1 August 2026 (UTC)
- It's due to
{{Plain navboxes}}which uses Template:Plain navboxes/styles.css. Those!importantannotations shouldn't be necessary; and because they've been used, it means that any personal CSS must itself also use!importantto have any effect at all. --Redrose64 🌹 (talk) 09:33, 1 August 2026 (UTC)- Important is required to ensure that inline CSS is overridden, which is the clear point of the template. Izno (talk) 15:52, 1 August 2026 (UTC)
- @Redrose64 I see changing a div's class from "PlainNavboxes" to "Navboxes" makes it display normally. I guess I could set up a tampermonkey extension script that automatically makes that change each time tho I wonder if you might know a simpler way? BoardHum (talk) 00:10, 2 August 2026 (UTC)
Shortcuts with the string 'RFC' at the beginning generate corrupted anchors
[edit]Shortcut boxes made with template {{shortcut}} generate their own anchor via <span id="shortcut"> that can be conveniently linked to from a redirect to the embedded anchor when a section name is not available. I ran into a problem when trying to add a redirect to shortcut {{sh|WP:RFCSIGN}} which I recently added to WP:Rfc. The initial version of the redirect targeted Wikipedia:Requests for comment#WP:RFCSIGN, but that did not work, and I wanted to know why.
In running it down, I found something bizarre: the span id generated for the embedded anchor appears to work fine in all cases, except when the pagename of the shortcut starts with "RFC", such as "RFCSIGN". In these cases, the pagename portion of the span id is transformed, with the 'R' replaced with the Html entity R instead of the 'R'. That this is the case, can be seen in the modified redirect, which is now working using the bizarre id, although the target looks just awful and inscrutable; see Wikipedia:RFCSIGN.
I ran a bunch of tests in order to narrow down when this is happening, and this is as far as I got:
a dozen tests trying to narrow down what's going on here |
|---|
|
1. {{sh|WP:RFCXYZ}} generates: <span id="WP:RFCXYZ">, with an Html entity R instead of an 'R'
2. {{sh|WP:DFCXYZ}}: change 'R' to 'D' – this works: span id="WP:DFCXYZ"
3. {{sh|WP:RFC XYZ}}: add a space after 'RFC'; doesn't help: span id="WP:RFC_XYZ"
4. {{sh|WP:RFC ABC}}: nothing special about 'XYZ'; with 'ABC' it fails, too: span id="WP:RFC_ABC"
5. {{sh|WP:R FCXYZ}}: add a space after the 'R'; this works: span id="WP:R_FCXYZ"
6. {{sh|WP:RF CXYZ}}: or after the 'F'; this works: span id="WP:RF_CXYZ"
7. {{sh|WP:RfcXYZ}} mixed case ('Rfc') works: span id="WP:RfcXYZ",
8. {{sh|Help:RFCXYZ}} not limited to project space: this generates: <span id="Help:RFCXYZ">
9. {{sh|Help:R FCXYZ}} but the same tweaks work, eg embedded space: <span id="Help:R_FCXYZ">
10. {{sh|Help:RfcXYZ}} so does lower case 'Rfc': <span id="Help:RfcXYZ">
11. {{sh|Help:RFCxyz}} trailing lower case doesn't help: <span id="Help:RFCxyz">
12. {{sh|Help:RFCX}} neither does just one trailing char: <span id="Help:RFCX"> |
As near as I can make out, the condition that generates an entity instead of 'R' in the id is whenever the shortcut pagename begins with the upper-case string "RFC", and is longer than three characters.
The string 'RFC' does not appear in Module:Shortcut. Either there is some voodoo parsing going on, or perhaps there is some strange bit of legacy code left over for some reason I cannot guess. Any ideas? Happy to port this to Phab, but wanted to make sure there's not something I'm missing. Mathglot (talk) 19:25, 1 August 2026 (UTC)
- This is probably the cause: https://github.com/wikimedia/mediawiki/blob/5c92a51f28990523d509c6171ac658fa82096859/includes/Parser/Sanitizer.php#L914
- Solomon Ucko (talk) 20:10, 1 August 2026 (UTC)
- Yes, can confirm. I experimented with the RFC examples in the Lua sandbox first, and then went to the MediaWiki Git repository and searched with
git grep "'RFC'". The sanitization has to do with so called magic links (see functionmagicLinkCallbackin Parser.php). The PMID links are affected the same way, which I checked in my Lua sandbox: Special:Permalink/1367221459#L-20--L-42. —andrybak (talk) 20:19, 1 August 2026 (UTC)- Since WP:RFCSIGN works, as do WP:RFCST, WP:RFCBEFORE, WP:RFCCAT, WP:RFCEND, and others, what's the problem? --Redrose64 🌹 (talk) 20:27, 1 August 2026 (UTC)
- The problem is that WP:RFCSIGN has to be an ugly and unintuitive redirect to
Wikipedia:Requests for comment#WP:RFCSIGN(note the HTML character entity for the first letter R in "RFCSIGN"), while the other examples get to link to regular section headers: WP:RFCST links toWikipedia:Requests for comment#Creating an RfC, WP:RFCBEFORE toWikipedia:Requests for comment#Before starting the process, etc. —andrybak (talk) 21:12, 1 August 2026 (UTC) - The problem is that none of the older shortcuts you linked meet the criteria of this discussion, which is about a redirect targeting the built-in embedded anchor created by the shortcut itself. All four redirects you mention are section links using nearby section names as the destination, so they don't need to use the built-in anchor, whereas RFCSIGN needs to, as it is not near the section header, but way down in bullet six, not even in the same screenview (on my 15-inch laptop). Like all shortcuts do, your four shortcuts generate their own embedded anchor, and had you linked them, you would find that they all fail as well: Wikipedia:Requests for comment#WP:RFCST, Wikipedia:Requests for comment#WP:RFCBEFORE, and so on; whereas for every policy and guideline except this one, they do work. (edit conflict) Mathglot (talk) 21:19, 1 August 2026 (UTC)
- The problem is that WP:RFCSIGN has to be an ugly and unintuitive redirect to
- Andrybak, interesting; presumably at line 1734 of Parser.php. I wonder if either that routine or the Sanitizer linked by Solomon is being too aggressive in its 'sanitization'. The page at mw:Help:Magic links lists the intent for a magic link like this:
RFC 1234→ RFC 1234
- which implies, without quite stating, that the blank is required for the magic link to be created, in which case something like RFCSIGN would be left alone, but the tests show that that is not the case. Is a possible solution here to alter the code so as to require the blank for the magic link? That would leave room for links like RFCSIGN to be created without interference.
- Alternatively, I understand that code changes may be expensive, and I am willing to live with the situation as is (although I wish it were better documented; I'll probably update Template:Shortcut/doc with a note), but with one caveat. I am somewhat worried that the currently working, but admittedly very bizarre-looking redirect at Wikipedia:RFCSIGN will attract well-meaning editors to alter it. In order to prevent that, I propose to keep the existing Wikipedia:RFCSIGN § Notes section at the redirect, and add a brief explanation there about what is going on, and retain the link to this discussion.
- In the past, I have occasionally added a ==Notes== section to a redirect to explain something unusual about the redirect, but I have found that such Notes are often removed by other editors, and I usually give up trying to keep them rather than dispute it. However, removing such a note would be very undesirable in this case, as it would then open up the redirect even more to unwanted alteration. Perhaps I could create a short, mbox-based informational message template, with instructions that it could be appended to the end of redirect pages involving RFC-, PMID-, or ISBN-like shortcut destinations. If that doesn't work, we might have to request page protection if it's still a problem. However, all of this is predicated on the assumption that changing the code is hard; and if it isn't, or if it could be piggy-backed on whatever the next change in that area is, maybe that would be a cleaner solution. (edit conflict) Mathglot (talk) 21:03, 1 August 2026 (UTC)
- In addition to checking for the required space after "ISBN", "PMID", and "RFC" in the sanitization code, Sanitizer.php could also read the configuration parameter
$wgEnableMagicLinks, and skip the unnecessary sanitization altogether, depending on which kinds of magic links are disabled. —andrybak (talk) 21:17, 1 August 2026 (UTC)- Andrybak, that sounds ideal. Who/what group is most informed about the status of these magic links and whether we need sanitization here? It sounds like the proximate cause(s) has(have) now been identified; I wonder if we are at the point where this can be condensed into a Phab ticket describing the essentials, and linking this discusion as background? Or should we wait for more input first? I'm happy to open a ticket, but it's clear to me that you have a much better understanding of what is going on as well as of the relevant software, and would probably write a better report than I would, so by all means feel free to start one; just lmk if you'd rather I do so. Thanks, Mathglot (talk) 21:34, 1 August 2026 (UTC)
- I found the box banner at the top of Help:Magic links confusing; although the page is marked inactive, the box reads: Magic links were removed from the MediaWiki code in March 2021...; so, were they or weren't they? And if they were, then no reason to sanitize, iiuc, or am I confused? Related: phab:T145604. Mathglot (talk) 21:53, 1 August 2026 (UTC)
- I think that the enwiki help page is just a bit misleading. They were disabled on enwiki in 2021 (phab:T275951), but they are still very much in the source code of MediaWiki itself. The wiki mediawiki.org has them enabled, as can be seen at mw:Help:Magic links. —andrybak (talk) 22:26, 1 August 2026 (UTC)
- You give me too much credit, I just grepped the source code for clues. I can read PHP code just enough, but the only teeny-tiny amount of experience that I have with it was over a decade ago.
- Created phab:T433782 "Sanitization of magic links in attributes should be disabled if magic links are disabled". —andrybak (talk) 22:26, 1 August 2026 (UTC)
- Very, very nice; much more on point than I could have managed. Much appreciated. Mathglot (talk) 22:59, 1 August 2026 (UTC)
- I found the box banner at the top of Help:Magic links confusing; although the page is marked inactive, the box reads: Magic links were removed from the MediaWiki code in March 2021...; so, were they or weren't they? And if they were, then no reason to sanitize, iiuc, or am I confused? Related: phab:T145604. Mathglot (talk) 21:53, 1 August 2026 (UTC)
- Andrybak, that sounds ideal. Who/what group is most informed about the status of these magic links and whether we need sanitization here? It sounds like the proximate cause(s) has(have) now been identified; I wonder if we are at the point where this can be condensed into a Phab ticket describing the essentials, and linking this discusion as background? Or should we wait for more input first? I'm happy to open a ticket, but it's clear to me that you have a much better understanding of what is going on as well as of the relevant software, and would probably write a better report than I would, so by all means feel free to start one; just lmk if you'd rather I do so. Thanks, Mathglot (talk) 21:34, 1 August 2026 (UTC)
- In addition to checking for the required space after "ISBN", "PMID", and "RFC" in the sanitization code, Sanitizer.php could also read the configuration parameter
- Since WP:RFCSIGN works, as do WP:RFCST, WP:RFCBEFORE, WP:RFCCAT, WP:RFCEND, and others, what's the problem? --Redrose64 🌹 (talk) 20:27, 1 August 2026 (UTC)
- Yes, can confirm. I experimented with the RFC examples in the Lua sandbox first, and then went to the MediaWiki Git repository and searched with
- Special:ExpandTemplates shows that
{{sh|WP:RFCSIGN}}produces wikitext (not HTML but merely wikitext) with<span id="WP:&#82;FCSIGN"></span>. So the encoding is made in Module:Shortcut (by a called function) and not when the wikitext is parsed to generate HTML. There is a simple solution for known cases: Manually write<span id="WP:RFCSIGN"></span>in addition to calling {{sh}}. The "R" will not be encoded.{{anchor|WP:RFCSIGN}}also works. It uses a module but not the same as{{sh|WP:RFCSIGN}}. PrimeHunter (talk) 23:27, 1 August 2026 (UTC)- Using the workaround of an explicit {{anchor}} on the page and redirecting to that instead is one way to get rid of the ugly redirect. However, no other page other than the Rfc page requires a workaround, and it won't keep the next person from falling into the same trap. If the module is altering the uri fragment to remove an 'RFC' string based on some outdated or disabled magic link feature, shouldn't it stop doing that? Mathglot (talk) 00:18, 2 August 2026 (UTC)
- I have documented the workaround. Ideally a workaround shouldn't be needed but workarounds are often good enough when a general fix is difficult, slow or uses limited resources. I guess Module:Anchor works here because it doesn't call
mw.uri.anchorEncodelike Module:Shortcut#L-85, but that probably means there are other cases where Anchor fails and Shortcut would have worked. The coding ofanchorEncodecannot be changed at the English Wikipedia. Phabricator requests are often ignored for years, and use limited developer resources when they aren't ignored. Trying to make our own alternative version ofanchorEncodewould probably create more problems than it solves.PrimeHunter (talk) 15:06, 2 August 2026 (UTC)- We may not need a custom version of
mw.uri.anchorEncode. The real problem seems to be that the function intends to return a string that's already HTML entity-encoded for use in a tag attribute, and then `mw.html:attr()` entity-encodes it a second time. Perhaps we could simply replace theanchorDiv:tag('span'):attr('id', anchor)withanchorDiv:wikitext('<span id="' .. anchor .. '"></span>')to avoid that double-encoding. Anomie⚔ 22:29, 2 August 2026 (UTC)
- We may not need a custom version of
- I have documented the workaround. Ideally a workaround shouldn't be needed but workarounds are often good enough when a general fix is difficult, slow or uses limited resources. I guess Module:Anchor works here because it doesn't call
- Using the workaround of an explicit {{anchor}} on the page and redirecting to that instead is one way to get rid of the ugly redirect. However, no other page other than the Rfc page requires a workaround, and it won't keep the next person from falling into the same trap. If the module is altering the uri fragment to remove an 'RFC' string based on some outdated or disabled magic link feature, shouldn't it stop doing that? Mathglot (talk) 00:18, 2 August 2026 (UTC)
Broken link but syntax appears to be correct
[edit]In the Jihadi John article there's a broken link in the section about the two Japanese victims (see screenshot). But from everything I could find, the link syntax is correct. Does anyone know what the issue could possibly be? Kufern (talk) 04:33, 2 August 2026 (UTC)
- That uses {{main}} which expects titles of articles, not links. {{ill}} outputs a link. The wikitext is like this:
{{main|Example|{{ill|NotHere|ja|Title}}}}→
- I don't know what to do about it but you could presumably kludge it by not using ill—instead, put the jawiki link:
{{main|Example|:ja:NotHere}}. Johnuniq (talk) 05:31, 2 August 2026 (UTC) - Or if having it on two lines is okay, then what about these two:
- Mathglot (talk) 05:51, 2 August 2026 (UTC)
- Yeah, that's better. Johnuniq (talk) 06:23, 2 August 2026 (UTC)
Xupright
[edit]Hello guys. I would need to invent a new function called, for example, "xupright". This means that the logos in the infoboxes would have the same height. I have been using the "size=x250px" parameter in infoboxes for many years, as each ice hockey championship logo has a different width... Many users like this solution and include it in new articles. But, according to Wikipedia rules, pixels should not be used for image dimensions, so my idea is to use the parameter "xupright=1" or "upright=x1". That would be a great idea in my opinion. 😉 Thanks, Maiō T. (talk) 11:32, 2 August 2026 (UTC)
- @Maiō T.: Are you familiar with Wikipedia:Bug reports and feature requests? --Redrose64 🌹 (talk) 21:50, 2 August 2026 (UTC)
- No. 😉 Thanks Redrose, I'll open that page right away. Maiō T. (talk) 22:02, 2 August 2026 (UTC)
Whether to use LLMs to generate user scripts
[edit]I thought about generating a user script with an LLM, though LLMs are now forbidden from being used to generate new articles or rewritten ones. Is that all right? For full context: discussion link. George Ho (talk) 20:11, 2 August 2026 (UTC)
- As far as I can tell, it's all right, and I have used LLMs to sometimes assist with such things, but as with software development in general, you should at least 1) know what you're asking for; 2) try to understand what you're getting (you can have the LLM comment it to assist); 3) test the script to ensure it works as desired; 4) figure out how to fix any issues or ask the LLM to fix them (and retest); and 5) release the script as an alpha or beta at first to allow for time to wring out any issues before declaring it a "full release". Hope this helps. the Stefen 𝕋ower 21:07, 2 August 2026 (UTC)
- For the discussion, I will note I am a former professional software developer, just in case anyone is unduly alarmed by my position. The bottom line is if you have the ability and/or patience to understand what you're being given, it can and will work. the Stefen 𝕋ower 22:32, 2 August 2026 (UTC)
- Please exercise extreme caution. LLM-written code has been described as the asbestos of software: it seems to do the job but may cause major problems later when changes are needed and no human understands what it is doing or why. Certes (talk) 21:37, 2 August 2026 (UTC)
- Please don't. An awful lot of trouble was caused on 31 December 2025 by somebody running an LLM-generated script. It went to WP:ANI and it took me and several other people some hours to clean up the mess. --Redrose64 🌹 (talk) 21:48, 2 August 2026 (UTC)
- If memory serves, that was a backend process or bot rather than a user script. the Stefen 𝕋ower 22:44, 2 August 2026 (UTC)
- There's not much difference between a bot and a user running a script that affects many pages very quickly - see WP:MEATBOT. Remember, I was there at the time, and was myself one of the victims; the ANI was Wikipedia:Administrators' noticeboard/IncidentArchive1211#TattooedLeprechaun. See also history of User talk:TattooedLeprechaun from that day, which gives some idea of the context - from an innocuous-looking request by a TA, through to an indef block that was appealed and declined. --Redrose64 🌹 (talk) 07:32, 3 August 2026 (UTC)
- Found another discussion about AI agents: Wikipedia talk:Bot policy#AI agents. George Ho (talk) 19:15, 3 August 2026 (UTC)
- There's not much difference between a bot and a user running a script that affects many pages very quickly - see WP:MEATBOT. Remember, I was there at the time, and was myself one of the victims; the ANI was Wikipedia:Administrators' noticeboard/IncidentArchive1211#TattooedLeprechaun. See also history of User talk:TattooedLeprechaun from that day, which gives some idea of the context - from an innocuous-looking request by a TA, through to an indef block that was appealed and declined. --Redrose64 🌹 (talk) 07:32, 3 August 2026 (UTC)
- If memory serves, that was a backend process or bot rather than a user script. the Stefen 𝕋ower 22:44, 2 August 2026 (UTC)
- I'm not sure if that policy applies or not, but I will echo what WhatamIdoing said at the linked discussion:
Do you know enough about Javascript to know whether the LLM is doing a good job?
As you answered no, I would avoid it. (I do have a basic understanding of JS and would avoid it myself regardless of policy.) I would also add the caveat that if someone stumbles across this and does write a script primarily using an LLM and publishes it, please disclose such. LittlePuppers (talk) 02:20, 3 August 2026 (UTC)- And if you answered yes, you're wasting your time by involving AI. Writing good code may be hard, but fixing bad code is harder. So that highlighted question is probably rhetorical in all contexts. ―cobaltcigs 03:31, 3 August 2026 (UTC)
- If you have a working piece of software, in most cases it's much easier (practically in terms of time involved) to fix a bug than to create the whole thing from scratch with nobody else's assistance (even if you're highly skilled). And LLMs can help find the bug and provide a solution in many cases. If you have a totally rotten code set, LLMs give you the opportunity to generate something totally anew. I would invite everyone to try using LLMs to generate some code even if you don't want to use LLMs or what they produce, just to see what is possible, today, in August 2026. It's pretty mind-blowing. the Stefen 𝕋ower 04:41, 3 August 2026 (UTC)
- Your signature is invisible in Safari browser. --Redrose64 🌹 (talk) 14:58, 3 August 2026 (UTC)
- Only on mobile view. I can still see the signature on desktop view on Safari, nonetheless. George Ho (talk) 19:08, 3 August 2026 (UTC)
- Your signature is invisible in Safari browser. --Redrose64 🌹 (talk) 14:58, 3 August 2026 (UTC)
- If you have a working piece of software, in most cases it's much easier (practically in terms of time involved) to fix a bug than to create the whole thing from scratch with nobody else's assistance (even if you're highly skilled). And LLMs can help find the bug and provide a solution in many cases. If you have a totally rotten code set, LLMs give you the opportunity to generate something totally anew. I would invite everyone to try using LLMs to generate some code even if you don't want to use LLMs or what they produce, just to see what is possible, today, in August 2026. It's pretty mind-blowing. the Stefen 𝕋ower 04:41, 3 August 2026 (UTC)
- And if you answered yes, you're wasting your time by involving AI. Writing good code may be hard, but fixing bad code is harder. So that highlighted question is probably rhetorical in all contexts. ―cobaltcigs 03:31, 3 August 2026 (UTC)
Archive bots
[edit]Is there any bot that can be used to archive specific talk page discussions based on a filter rather than just date? Specifically, I'm hoping to automate archival of KiranBOT notifications regarding threads being archived, but it would be nice if I could add custom criteria for anything I want to sweep for routinely. ChompyTheGogoat (talk) 20:49, 2 August 2026 (UTC)
- @ChompyTheGogoat: See User:ClueBot III. By using the
|archivenow=parameter, it is possible to configure it to only archive threads that contain certain templates - such as{{done}}and{{resolved}}. Be aware that it's case-sensitive on all letters. --Redrose64 🌹 (talk) 21:46, 2 August 2026 (UTC)- Do you know if it's possible to PREVENT it from archiving any threads that don't qualify under that parameter? I'm currently not wanting to archive anything else automatically, although I may change my mind on that later. ChompyTheGogoat (talk) 21:57, 2 August 2026 (UTC)
- Please clarify. Are you saying that you want only threads like User talk:ChompyTheGogoat/Archive 1#Your thread has been archived to be archived, or those to be left alone and all other threads archived? --Redrose64 🌹 (talk) 22:31, 2 August 2026 (UTC)
- Only those (or other automated messages I add an
|archivenow=parameter for). The ability to define any string there is a exactly what I was looking for, but without the bot's default time-based archival of everything else. ChompyTheGogoat (talk) 22:43, 2 August 2026 (UTC)- then I would suggest to use
|archivenow=User:KiranBOT/Teahouse_archival_notification. Please note that you can have only one|archivenow=in a{{User:ClueBot III/ArchiveThis}}, and only one{{User:ClueBot III/ArchiveThis}}on a given talk page. But|archivenow=may contain multiple comma-separated strings; we do this at Wikipedia:Closure requests. --Redrose64 🌹 (talk) 22:58, 2 August 2026 (UTC)- Yes, that's what I meant by having multiple strings defined. ChompyTheGogoat (talk) 23:25, 2 August 2026 (UTC)
- The usual workaround is to set an huge number for the time-based threshold (the
|age=parameter). isaacl (talk) 23:03, 2 August 2026 (UTC)
- then I would suggest to use
- Only those (or other automated messages I add an
- Please clarify. Are you saying that you want only threads like User talk:ChompyTheGogoat/Archive 1#Your thread has been archived to be archived, or those to be left alone and all other threads archived? --Redrose64 🌹 (talk) 22:31, 2 August 2026 (UTC)
- Do you know if it's possible to PREVENT it from archiving any threads that don't qualify under that parameter? I'm currently not wanting to archive anything else automatically, although I may change my mind on that later. ChompyTheGogoat (talk) 21:57, 2 August 2026 (UTC)
Rash of script errors
[edit]Category:Articles with script errors (17) has blown up the last few days with lots of pages about elections in India, but I cannot seem to find anything that is actually broken or any errors to speak of. Can anyone point me in the right direction? Zackmann (Talk to me/What I been doing) 22:02, 2 August 2026 (UTC)
- Three of the articles I clicked on 1998 elections in India and 1998 Indian general election and 2007 Gujarat Legislative Assembly election - have "Lua error in Module:Political_party at line 105: bad argument #1 to 'find' (string expected, got nil)." in their results sections - is that not it? Morwen (talk) 22:08, 2 August 2026 (UTC)
- What doesn't make sense is that neither those pages nor Module:Political party have any recent changes that would have caused this. Zackmann (Talk to me/What I been doing) 22:11, 2 August 2026 (UTC)
- Gotcha. I am just about to develop a minimal failing test case. Morwen (talk) 22:17, 2 August 2026 (UTC)
- While I was developing a minimal failing test case - which is User:Morwen/bad election by the way - it magically started working. Go figure. Morwen (talk) 22:25, 2 August 2026 (UTC)
- @Morwen: I think I fixed that one while you were testing. Certes (talk) 22:28, 2 August 2026 (UTC)
- I've made a small edit which fixes at least one of the errors (February 2005 Bihar Legislative Assembly election). Certes (talk) 22:26, 2 August 2026 (UTC)
- Aha. Cool. We probably want to change Module:Political party as well to not be so fragile. Morwen (talk) 22:31, 2 August 2026 (UTC)
- Perhaps, though the error messages could be seen as useful ways to find missing data. A cursory search doesn't reveal what colour I should have added. They seem to use a three-colour flag. Certes (talk) 22:40, 2 August 2026 (UTC)
- The error message could very easily be made to say what entry needs adding where, in that case. Morwen (talk) 22:49, 2 August 2026 (UTC)
- I /think/ that if we changed line 81 to
local return_value = party_info[out_type] or ""then that would have produced an error saying "Value not in template. Please request that it be added." which would be slightly better: we could also change that error message to mention the specific party it's trying to look up at that point. Morwen (talk) 22:52, 2 August 2026 (UTC)- That sounds much better, especially if it can mention the type of information that is missing (in this case, colo[u]r). Certes (talk) 08:55, 3 August 2026 (UTC)
- I /think/ that if we changed line 81 to
- The error message could very easily be made to say what entry needs adding where, in that case. Morwen (talk) 22:49, 2 August 2026 (UTC)
- Perhaps, though the error messages could be seen as useful ways to find missing data. A cursory search doesn't reveal what colour I should have added. They seem to use a three-colour flag. Certes (talk) 22:40, 2 August 2026 (UTC)
- Aha. Cool. We probably want to change Module:Political party as well to not be so fragile. Morwen (talk) 22:31, 2 August 2026 (UTC)
- While I was developing a minimal failing test case - which is User:Morwen/bad election by the way - it magically started working. Go figure. Morwen (talk) 22:25, 2 August 2026 (UTC)
- Gotcha. I am just about to develop a minimal failing test case. Morwen (talk) 22:17, 2 August 2026 (UTC)
- What doesn't make sense is that neither those pages nor Module:Political party have any recent changes that would have caused this. Zackmann (Talk to me/What I been doing) 22:11, 2 August 2026 (UTC)
Oddness with User:Polygnotus/DuplicateReferences
[edit]This came up at Andrew Tate. The "Duplicate citations" script was deposited at the top of the main article (along with the listed citations which were hide/show). The instructions do state "You can add a {{Duplicate citations}} template to the page by clicking the link next to the References section header." Can any of you coders tell me if it's possible for the template to automatically be placed in articles' References sections? Do you think bolded instruction about placement at User:Polygnotus/DuplicateReferences would help? Thanks, Shearonink (talk) 18:45, 3 August 2026 (UTC)-
- I'm not sure I understand the outcome you're looking for. DuplicateReferences is a user script written by @Polygnotus. It is unlike that they are the user that used it to place that section. Are you asking for a bot that goes around and runs this code on a large number of articles? audiodude (talk) 18:58, 3 August 2026 (UTC)
- No, that is not what I am looking for. And yes, I know which editor placed the notice on the article, I did a search of the article history to find it. My issue is that having the template get placed at the very top of an article is not an optimum use of the template and if it's happened at the Tate article, it's also probably happening on other articles too. The instructions state that the template is supposed to be placed in the References section. Just asking WP-coders IF automatic placement in the References section is possible. - Shearonink (talk) 19:21, 3 August 2026 (UTC)
- It's this edit by DocWatson42 that hoisted {{Duplicated citations}} to the top of the article (which doesn't appear to have been made using the DuplicateReferences script), not the one you linked to (which correctly put it in the References section). Nardog (talk) 19:42, 3 August 2026 (UTC)
- Ah, thank you! Appreciate the education. Never mind then. Thanks, Shearonink (talk) 21:34, 3 August 2026 (UTC)
- See my comment in the template's talk page. Why is it to be placed in the "References" section when all of the other reference maintenance templates affecting the whole articles (that I am aware of—counter examples, please) are placed at the top of the article? —DocWatson42 (talk) 04:55, 4 August 2026 (UTC)
Tech News: 2026-32
[edit]Latest tech news from the Wikimedia technical community. Please tell other users about these changes. Not all changes will affect you. Translations are available.
Updates for editors
- The Reader Experience team has developed a patch demo that wraps the page toolbar onto two lines when there is not enough horizontal space for all the buttons. This aims to reduce crowding in the Vector 2022 toolbar, which can occur on some language Wikipedias at certain screen widths.
- The Reader Experience team is planning to launch Reading Lists, a Community Wishlist item, which is currently available to try in beta, as a full feature in September. Before then, volunteer translator help is needed for string translations into a number of languages. The feature supports reading and learning goals on Wikipedia.
- Next week, the table of contents on Wikimedia Commons file pages will be improved by consolidating the file page table of contents with the page table of contents. This will make it easier to understand a file page’s structure, navigate to specific sections, and share links to individual sections.
- View all 24 community-submitted tasks that were resolved last week. For example, an issue where some TIFF images failed to load after clicking their thumbnail, causing a broken image to be displayed instead of the full-size image, has now been fixed.
Updates for technical contributors
- The variable and function selector in AbuseFilter has been updated to support search and autocomplete. It will allow filter maintainers to find the desired variable or function more quickly.
- The MJPEG and VP8 formats are removed from the video player. The MP4 format (MPEG-4 Part 2) is added instead, which provides higher quality videos to older iPhone devices. It may take a few weeks to retroactively update all existing videos. The default format for modern devices stays the same (VP9/WebM).
- Detailed code updates later this week: MediaWiki
Tech news prepared by Tech News writers and posted by bot • Contribute • Translate • Get help • Give feedback • Subscribe or unsubscribe.