:Came here to raise similar concerns. We have articles like [[SS United States|SS ''United States'']] with two infoboxes tucked inside one another having extraordinarily wiiiiide boxes, to the point that articles are hard to read. I've seen broken infoboxes thinner than the current ones.
:Came here to raise similar concerns. We have articles like [[SS United States|SS ''United States'']] with two infoboxes tucked inside one another having extraordinarily wiiiiide boxes, to the point that articles are hard to read. I've seen broken infoboxes thinner than the current ones.
:Yesterday, I was set to the old 2010 display settings. Who's screwing with the new vector display? [[User:GGOTCC|GGOTCC]] ([[User talk:GGOTCC|talk]]) 22:21, 13 June 2024 (UTC)
:Yesterday, I was set to the old 2010 display settings. Who's screwing with the new vector display? [[User:GGOTCC|GGOTCC]] ([[User talk:GGOTCC|talk]]) 22:21, 13 June 2024 (UTC)
::Annnnd it's fixed! Check '''Thursday 13 June style changes''' above for more info. [[User:GGOTCC|GGOTCC]] ([[User talk:GGOTCC|talk]]) 22:23, 13 June 2024 (UTC)
If you want to report a JavaScript error, please follow this guideline. Questions about MediaWiki in general should be posted at the MediaWiki support desk. Discussions are automatically archived after remaining inactive for five days.
This tends to solve most issues, including improper display of images, user-preferences not loading, and old versions of pages being shown.
No, we will not use JavaScript to set focus on the search box.
This would interfere with usability, accessibility, keyboard navigation and standard forms. See task 3864. There is an accesskey property on it (default to accesskey="f" in English). Logged-in users can enable the "Focus the cursor in the search bar on loading the Main Page" gadget in their preferences.
No, we will not add a spell-checker, or spell-checking bot.
You can use a web browser such as Firefox, which has a spell checker.
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.
The HTML used to render all headings is being changed to improve accessibility. It will change on 22 May in some skins (Timeless, Modern, CologneBlue, Nostalgia, and Monobook). Please test gadgets on your wiki on these skins and report any related problems so that they can be resolved before this change is made in all other skins. The developers are also considering the introduction of a Gadget API for adding buttons to section titles if that would be helpful to tool creators, and would appreciate any input you have on that.
Based on a quick search, it looks like the heading change will affect almost 300 scripts, many of which have inactive maintainers. Some arbitrary highlights from the top of the list include:
A quick way to test these scripts right now, is to enable the Parsoid beta option (which already uses the new html structure) and to disable DiscussionTools, which uses a partial form of the new heading structure. —TheDJ (talk • contribs) 08:39, 22 May 2024 (UTC)[reply]
Indeed, you can already see it in Parsoid mode (but note that there are other differences – e.g. Parsoid output has <section> tags around each section, which may require a separate set of updates in some scripts).
Disabling DiscussionTools doesn't actually change anything though. The HTML structure is the same whether it's enabled or disabled, only the styles are different. Also, note that it uses a "hybrid" heading structure currently when using the default parser, as you say, but it uses the new structure when using Parsoid.
So in short, you can just use Parsoid mode to test these scripts today here on English Wikipedia, but beware that there may be extra issues. But if they work with Parsoid, they will work with the new headings too. Matma Rextalk11:25, 22 May 2024 (UTC)[reply]
The technical 13 script was blanked, so we don't have to worry about that one.
Will the fact that they're rolling this out for only some wikimedia-deployed skins at this time make the patch more complicated? If I'm reading it right, the scripts may temporarily have to support both heading styles. –Novem Linguae (talk) 09:16, 22 May 2024 (UTC)[reply]
I'm assuming ~ and feel free to correct me if i'm wrong ~ that something about this deployment is why headings no longer have numbers (for me)? Will it be possible to go back to that at some point? I find long pages almost impossible to navigate around without numbered headings, so will have to learn a new way of working if it won't be possible. Thanks, Happy days, ~ LindsayHello16:24, 27 May 2024 (UTC)[reply]
If you're speaking about the table of contents, Vector 22 does not provide numbering. Vector, Monobook, and Modern do.
If you are speaking about each actual heading, then indeed the preference is gone and indeed there is a gadget for it now. You have correctly identified that gadget as needing to be updated for this change. It looks like the necessary change to the snippet (documentation) has already been made, so someone needs to port that here. Izno (talk) 19:59, 27 May 2024 (UTC)[reply]
Thank you, Izno, helpful. I'd assumed it was a script/gadget, as so many appeared to be affected above. I shall patiently wait in hope Happy days, ~ LindsayHello11:51, 28 May 2024 (UTC)[reply]
Did you try all the old skins such as Timeless and Monobook? Vector isn't affected at all yet, and editing likely uses the API, but I can imagine the location of the header links this script places being possibly broken in old scripts. I fixed this kind of thing in 2 gadgets so far. –Novem Linguae (talk) 22:15, 5 June 2024 (UTC)[reply]
New h2 headings use serif font even when the "Vector classic typography" gadget is enabled
Vector classic typography is a gadget that forces all text to use sans-serif fonts, but even with the gadget enabled h2 headings on articles use a serif font. Incorrect behavior seen on both Firefox and Edge. TomatoFriesLAN (talk) 18:51, 6 June 2024 (UTC)[reply]
I usually spend part of the day closing AFD discussions but none of the XFDcloser options are showing up. Not even the ability to relist. I've uninstalled every installation, unchecked the XFDcloser gadget, returned everything to normal but nothing works. Do I have to reboot my computer or something? Log out and log back in? This rarely happens so I'm not sure what happened today. I've posted a message on the XFDCloser talk page but it doesn't get much activity there. LizRead!Talk!23:26, 6 June 2024 (UTC)[reply]
Izno, I see you've moved this section, and it does appear to be mentioned in the original post of this threading, but why would it only appear now? I seem to recall closing discussions earlier this week (and I suspect Liz has as well). Primefac (talk) 01:17, 7 June 2024 (UTC)[reply]
Well, I thought this thread was deleted until I found it reposted up here.
It's odd because XFDCloser was working fine this morning and then this afternoon, it just didn't load at all. But I see other editors closing discussions so I hope it isn't just me. I've had ongoing problems with XFDCloser not loading on CFD pages but it hasn't been a problem on AFD daily logs until today. Thanks for checking Novem Linguae, there are usually over 100 AFD discussions daily so if this is happening for other closers, they could pile up pretty quickly. If it matters, I use a laptop with Windows. LizRead!Talk!03:17, 7 June 2024 (UTC)[reply]
I figured out the cause. I should have a fix deployed soon.
For the record, it looks like WMF deployed mw:Heading HTML changes to old skins (monobook, timeless, modern, cologneblue) last week, vector (2010) this week, and probably minerva and vector-2022 in the coming weeks. All breakages we see today will probably be vector (2010) only.
This staggered deployment has pros and cons. It means that if someone like me does fix a bunch of gadgets today, I'll just have to go fix them all again next week when they break on vector-2022.
It would be nice if there were an API for inserting header links. phab:T337286. APIs like mw.util.addPortlet(), mw.util.addPortletLink(), etc are great for multi-skin support and for keeping HTML changes from breaking gadgets and user scripts. –Novem Linguae (talk) 05:43, 7 June 2024 (UTC)[reply]
Yeah, I don't understand all of this jargon but I am FOREVER grateful that their are editors who do. Thanks for looking into this. LizRead!Talk!06:33, 7 June 2024 (UTC)[reply]
Yeah, and I don't like the left-side menu. But thanks Novem Linguae, it looks like things are now back to normal. I can go back to my old skin! Many thanks. LizRead!Talk!07:24, 7 June 2024 (UTC)[reply]
Novem Linguae, XFDCloser disappeared again! I think you said this might happen. It came back when I changed to Vector 2022 but, ugh! I guess I'll use that skin when working in AFDLand and then change back when doing regular editing. LizRead!Talk!22:15, 8 June 2024 (UTC)[reply]
Very strange. I haven't done any work on XFDcloser since the last deploy on Thursday, and I don't see any relevant backport patches at wikitech:Server Admin Log that might have changed MediaWiki behavior this weekend. This is all quite mysterious. –Novem Linguae (talk) 08:56, 9 June 2024 (UTC)[reply]
What's the user script or gadget that puts a ¶ symbol next to headings, and when you click on it, it opens a modal with links to that section that you can copy/paste? It broke for me today and I want to fix it, but can't remember what it's called. Thanks. –Novem Linguae (talk) 03:35, 7 June 2024 (UTC)[reply]
The function showCommentLinks() (starting on line 73) adds the links. The section of code starting at line 84 finds headings in the HTML document structure previously generated by MediaWiki (which I believe is the same across skins). The section of code starting at line 93 finds headings in the currently generated HTML document structure. isaacl (talk) 15:27, 7 June 2024 (UTC)[reply]
I was hoping you'd just tell me where the links are. lol. Anyway, I put a breakpoint on line 75 and the breakpoint is not getting hit when I refresh this page. I'm missing something. –Novem Linguae (talk) 20:33, 7 June 2024 (UTC)[reply]
I'm sorry, I didn't realize you were asking about the interface. As described in the documentation, you have to select the "Toggle link2clipboard" item in the tools menu (the location of the menu depends on your skin; for Vector 2010 it's in the left sidebar). </> is prepended to the start of each comment. For headings, <h/> is also prepended. Most of the time I don't want to see the links, so I chose to require an extra step to display them. Another difference from the other script is that for the major non-Safari browsers, the link text is automatically copied to the clipboard (always without surrounding square brackets; the other script can be configured not to do that if desired). isaacl (talk) 20:51, 7 June 2024 (UTC)[reply]
Nice, that worked. Thanks a lot. Feature idea: Add a way to copy it as an external link. I do this a lot when writing GitHub or Phabricator tickets, for example. –Novem Linguae (talk) 20:59, 7 June 2024 (UTC)[reply]
As my personal frequent use case is to link to comments or sections in wikitext, I wanted a way that would provide easy access to the link without underscores ;-) (And I chose to avoid square brackets as it's easier to add them when needed than delete them, and I like to use {{section link}} when feasible.) I'll take it under consideration, though; thanks for the feedback! isaacl (talk) 21:08, 7 June 2024 (UTC)[reply]
Section header typeface
I just noticed that section headers in articles are now using a serif typeface on both Vector and Vector legacy. Sorry I couldn't find information about this elsewhere but when and why was this change made? I do not like that it uses Oldstyle figures and would like to change it in my settings or .css page to be the same sans serif font used in other headers. Thanks! Reywas92Talk17:23, 7 June 2024 (UTC)[reply]
@Reywas92 Vector headers have actually been using serif fonts by default for a long time, but you have user CSS which was overriding that. It no longer works due to some changes to heading HTML. You can either change that part of your user CSS to:
We at Wiki Project Med have built a gadget to view stacks of images such a as CT scans, which you can see here[2]. We are wanting to install it on EN WP.
Previously mentioned to User:MusikAnimalhere who want to verify community consensus first.
@Doc James about how many pages would this need to run on? We are currently experimenting with our very first implementation of Template Gadgets (see a couple sections up) right now, which I imagine would be the way we would want to implement this (and most certainly not by hooking a full page text analyzer in to common.js). — xaosfluxTalk18:49, 30 May 2024 (UTC)[reply]
@Doc James yes, where said category would come along with a template that would wrap whatever is being used. It sounds like all instances of this would use some template so that part isn't hard. What order of magnitude of pages would you expect this would get used on? — xaosfluxTalk19:23, 30 May 2024 (UTC)[reply]
For a default gadget, i'd have some concerns about the accessibility of the play button. It's not a button, and it's also not labeled.
Similar for the pager and slider in the window. This is unlabeled. It should have accessibility labels to make it possible to understand what the slider does.
The play button positioning and sizing might need a little bit more work, it seems kinda off (esp on iphone)
Might want to hide the play button on media print
Good to see that media credits are being linked.
Seems to work on mobile, but could use some additional spacing at the top controls, they are really difficult to hit because everything is so close together now.
Closing the dialog. All MW dialogs currently have close at the top (an old pattern i note due to mobile usage favoring thumb interaction at the bottom of a dialog). This does create an inconsistency, but i'm not particular concerned.
The whole ImageStackPopup-viewer is inside a label element atm. I think that's an accident?
There's been some accessibility improvements in the latest version. Button is also now hidden on print. The label thing and the close button at the bottom seem to be due to using OO.ui.alert. I'm not sure why OOUI does it that way for alert boxes. Bawolff (talk) 13:18, 31 May 2024 (UTC)[reply]
@Doc James, As one of our professors often says, "One view is no view in Radiology." From a content perspective, I am confident that these imaging stacks will enhance the quality of our radiology related articles. Looking forward to seeing this implemented soon. signed, 511KeV (talk)19:03, 30 May 2024 (UTC)[reply]
Moral support for the idea, bug-report for the implementation: the stack is scrolled by a left–right slider, but when hovering over the image the stack scrolls when I move the mouse up-and-down and not side-to-side. DMacks (talk) 19:49, 30 May 2024 (UTC)[reply]
Given the bird's-eye view with the line indicating the location of the specific scan is an up-and-down position, having the slider be side-to-side is confusing. Everything needs to be in sync. DMacks (talk) 00:09, 31 May 2024 (UTC)[reply]
I've forked ImageStackPopup over for anyone that wants to test it out in sandboxes etc, you can either manually opt-in to it in the "testing and development" gadget section, or you can load it to a page with the ?withgadget query parameter. From discussion above, this seems like it will need some extensive testing and tweaking. Nothing should currently be placed in to an article that is dependent on this right now, as readers will not be able to make use of it yet. — xaosfluxTalk23:55, 30 May 2024 (UTC)[reply]
The Vivarium template gadget being currently tested is much simpler, and we will make sure our roll out of template gadgets is done carefully. Additional discussion around if these should be able to be opted out of should also occur (i.e. not making them default+hidden). For a default here, we'll likely also use a fork, we have a bot to monitor remote changes and flag for promotion that can be used. — xaosfluxTalk23:58, 30 May 2024 (UTC)[reply]
The test version on Commons loads 250 images. Given how heavy these images are, this seems like a bad use case for a gadget and should potentially be in some sort of video instead, which won't try to download that many images all at the same time. Izno (talk) 00:24, 31 May 2024 (UTC)[reply]
That does seem like a lot if its all hitting the browser right away. Something that heavy sounds like it would be better to paginate and be done in mediaviewer perhaps. — xaosfluxTalk00:28, 31 May 2024 (UTC)[reply]
To clarify, the images get downloaded only after the user hits the play button, so only users who want to see them do the download. Perhaps that could be improved with a progress loading bar or something or the ability to cancel. The goal is to allow users to directly compare all the images all at once, so i'm not sure pagnation would work here. I agree that as a long term solution, transfering as a video with p-frames/temporal compression would probably be much more bandwidth efficient. Bawolff (talk) 05:43, 31 May 2024 (UTC)[reply]
Also, just to be clear, this gadget does not exist on commons. There is a separate gadget on commons called ImageStack, which is the inspiration for this gadget, but its a totally different gadget. Bawolff (talk) 09:46, 31 May 2024 (UTC)[reply]
Looking very nice, but I still think it needs a bit more work for mobile. I'd still say that my fingers are not 3mm x 3mm. Additionally the right positioning of the controls now gets into the scroll zone, which is possibly even worse. I can trigger the rubber banding of the scroll area, and if I zoom in, we overlap with the scrollbar of the viewport. If you switch to desktop skin on mobile, you have the same, but zoomed out 6 times so you really do need that zooming and scrollbar. —TheDJ (talk • contribs) 22:33, 1 June 2024 (UTC)[reply]
My concerns about consistency in direction of scrolling are resolved. For the record, I'm using a desktop machine. DMacks (talk) 05:32, 2 June 2024 (UTC)[reply]
See this diff, where one line was moved. The moved line ought to be preceded with curved arrows, on both the left- and right-hand sides. Instead, I see that these arrows are obscured by large black discs. If I hover my mouse over the disc, it resoves to the correct curved arrow, but returns to being a disc on moving the mouse away. This started happening in the last half hour. Firefox 126.0.1, all skins, logged in or out. I blame WP:ITSTHURSDAY. --Redrose64 🌹 (talk) 19:42, 6 June 2024 (UTC)[reply]
Can confirm. Same issue for Extended Support Release version of Firefox. Similar on Chrome and desktop site version on Mobile Firefox, except that the black disks are respectively dark blue and dark grey there. Desktop site on Mobile Chrome gives emoji-style white curved arrows in a blue box rather than plain box-less curved arrows with or without obscuring dot. AddWittyNameHere20:09, 6 June 2024 (UTC)[reply]
And I've been Ctrl+F-ing moved lines like a dummy this whole time‽ (╯°□°)╯︵ ┻━┻ And it's not really that hard to discover yourself Self-trout. —andrybak (talk) 08:44, 7 June 2024 (UTC)[reply]
I too am seeing this same exact issue on edit diff pages. Also Firefox 126.0.1 64-bit here, I think it happened on my Linux computer as well. Vector 2022 skin.
The following code does kill the black spots, but it is very dirty and I really don't want to leave it like that. Is there an edit preview event that I can hook to? setInterval(function(){$('.mw-diff-movedpara-left, .mw-diff-movedpara-right').text('');},666); — GhostInTheMachinetalk to me
The latest run of Special:WantedCategories features a cluster of redlinked categories that are being somehow autogenerated by WikiProject templates — but I can't work out where they're coming from because the category declaration does not exist in either the template or its documentation, and none of the templates have been edited recently to suddenly generate new categories that didn't exist before this week, which means they're being passed through by a new coding change somewhere other than the templates themselves, such as in a module or a template framework I'm not familiar with.
But I can't justify creating most of them either, as they mostly seem to correspond to task forces rather than full wikiprojects, and thus would never have categories at the names that have been newly autogenerated for them — and in many cases they already have categories located at different names than the ones that have been newly autogenerated for them, which are sitting on the template alongside the redlink. By and large, they seem to correspond word-for-word to the name of the template itself, meaning that most likely some edit somewhere has caused an erroneous assumption that every WikiProject template should automatically generate an eponymous category matching its own name, which is obviously not the case.
So I'm at a loss. Could somebody look into the following categories, and figure out how to resolve them?
Category:WikiProject Irish music, Category:WikiProject Private Equity, and Category:WikiProject Big 12 Conference should be created soon, they required speedy renaming.
Category:WikiProject Mathematical and Computational Biology should not be created, the banner was sent to TfD and should be merged into parent template.
Category:Article Rescue Squadron - The project lead, banner and category use "WikiProject Article Rescue Squadron", the project page is titled "Article Rescue Squadron". One style should be followed, it is either a WikiProject or not.
Category:WikiProject Article Collaboration and Improvement Drive, and Category:WikiProject Challenges should not be created. Banner sent to TfD. If kept at TfD category will need creating.
Category:WikiProject Counter-Vandalism Unit if needed for the code, should be created as a redirect. I created Category:Wikipedia:Counter-Vandalism Unit to match project, banner and sub-categories.
I'm seeing broken infoboxes on Henry Kissinger and Donald Trump (broken styling making them too big, some kind of parse error in the 'spouse' field, specific offices held replaced by a redlinked 'Ambassador to'). No obvious recent edits to either that would have broken them, and both are high profile articles that I'd expect to quickly get fixed. Both use {{infobox officeholder}}, but again I don't see obvious recent changes (last one in April). It's probably something in the infobox machinery but I don't even know where to start looking or who to notify. Polyphemus Goode (talk) 11:43, 8 June 2024 (UTC)[reply]
Wow, that was bad! Yeah, I really should have chosen disabled from the start. I took two templates that are sometimes supposed to return nothing and made them always return something. What could possibly go wrong? As for my suggestions, there is no way for WP:noinclude to break since its whole purpose is to transclude nothing but I'm not so sure about |type=disabled. I guess we'll see. Nickps (talk) 13:19, 8 June 2024 (UTC)[reply]
Yeah, but everything is obvious in hindsight; it's no big. The first step in knowing something is not knowing it, Rjjiii (talk) 20:31, 8 June 2024 (UTC)[reply]
The Tfm notice on {{If both}} keeps breaking things. Around 350 of the articles using {{Infobox YouTube personality}} are currently displaying a reference error like this on Blimey Cow: "Cite error: The named reference YouTubeStatsBlimey Cow was invoked but never defined". {{If both}} has 134047 transclusions. I suggest we just place the Tfm notice in <noinclude>...</noinclude> instead of hoping to find another method which doesn't cause errors. PrimeHunter (talk) 12:25, 10 June 2024 (UTC)[reply]
Hi, I just noticed that revdelled content can still be visible through the filter log, if you click "examine" on an edit and the revdelled content was close enough to the attempted change shown in the filter log. Can this be fixed? Air on White (talk) 02:12, 9 June 2024 (UTC)[reply]
Infoboxes and taxoboxes pushed below opening paragraph in mobile
Infoboxes and taxoboxes are now pushed below opening paragraph in mobile. Is this behaviour deliberate?
The reason I ask is I discovered this while trying to fix an issue with excess white space before taxoboxes. This occurred because of T18700 which introduced an empty paragraph with this HTML: <p class="mw-empty-elt"></p>. The workaround was to precede the taxobox with a <nowiki/> or later with <templatestyles>. This workaround no longer works after recent changes, which also introduced the shifted infoboxes and taxoboxes in mobile.
In an attempt to fix this I wrapped the taxoboxes in a <div> element and this worked in my first tests in edit preview. However it doesn't work in all cases. When the taxoobox is the first element in the article the fix works, but it does when preceded by some of the hatnote, protection and formatting templates. When it works the taxobox is no longer pushed below the first paragraph in mobile and the empty paragraph element is no longer there. It's as if the empty paragraph captures the first paragraph of the lede.
You can see this behaviour in Neoaves by wrapping the {{automatic taxobox}}<div> tags. Similarly at Lionel Messi, wrapping {{Infobox football biography}} with <div> tags moves the infobox to the normal top-right location in mobile. In this case the empty paragraph HTML code is reintroduced. If you remove all the hatnote/protection/formatting templates the empty paragraph disappears. Putting all the top templates on the same line also removes the empty paragraph.
I'm just using the mobile view on a laptop. However the issue is with desktop view. The pushing down of the infoboxes on mobile moves them away from the problematic templates at the top of the page and fixes the issue.
The mw-empty-elt element is empty but adds vertical space. Not a lot but enough for people to report on the template talk pages. Having wasted a lot of time on this over the years it's annoying to have the problem resurface. — Jts1882 | talk13:16, 10 June 2024 (UTC)[reply]
Template-transcluded redlinked categories, again
The latest run of Special:WantedCategories features two template-transcluded redlinks that I've been unable to figure out how to empty. They both result from that perennially irritating "isn't supposed to be happening but still regularly happens anyway" thing where somebody throws a template-generated category to the speedy renaming process, but the bots that handle speedy renames can't edit the templates that transclude the categories, so everything stays filed in the redlink — in both of these two cases, I was able to mostly empty out the categories, but each has one or two leftover pages that won't clear out for some other reason I can't identify because the leftover redlinks have eluded everything I've done to try to find their sources.
So could somebody look into these two categories? Thanks.
I prefer Vector Legacy as - at least on my computer - Legacy 2022 deletes the index of article contents. Vector Legacy with a menu option to switch to 2022 when needed seems a better solution. Thanks again for your help PrimeHunter. Dudley Miles (talk) 08:16, 11 June 2024 (UTC)[reply]
Template issue
Just came across a possible problem with the Template:Convert, it appears to be giving incorrect calculations. I notitced it in this article, though it doesn't appear to be limited to that page, and the errors I noted were specifically occuring when converting miles to kilometers. It seems to occur once you get into 3 and 4 digit numbers, though some rounded numbers come up correct (eg: 1,000 miles (1,600 km) ✓) but once you start adding single numbers to the end, the errors start to become larger. Eg;
1,001 miles (1,611 km) (+9.6, should be 1601.6)
1,002 miles (1,613 km) (+9.8, should be 1603.2)
1,115 miles (1,794 km) (+10, should be 1,784 km)
3,550 miles (5,710 km) (+30, should be 5,680 km)
7,077 miles (11,389 km) (+65.8, should be 11,323.2 km)
The two errors I initially noticed on that page were;
2,906 miles (4,677 km) (+27.4, should be 4,649.6 km)
2,900 miles (4,700 km) (+60, should be 4,640 km), the second entry dropped by 6 miles, but increased by 23 km...(?)
Also, I realize the template is rounding off to the next whole number, (or should be), I've only added decimals to show the fully correct number. If someone could take a look at this and either confirm there is a problem, or even better, fix the problem, or if I l've just bungled this somehow, then please let me know. Thanks - wolf15:30, 10 June 2024 (UTC)[reply]
@Thewolfchild:Amile is by definition exactly 1609.344 m. You appear to incorrectly think it's 1600m. 1,000 miles (1,600 km) only says 1,600 km because 1,000 is a round number which was probably an approximation so the exact conversion 1609.344 km or a small rounding to 1609 km would give a misleading sense of precision. PrimeHunter (talk) 15:51, 10 June 2024 (UTC)[reply]
Just so. The default rounding is to "precision comparable to that of the input value... or to two significant digits". So 1000 miles = 1609.344 km is rounded to 1600 (2SF) but 1001 miles = 1610.95... is rounded to 4SF and becomes 1611. rbrwr±16:11, 10 June 2024 (UTC)[reply]
(edit conflict)@PrimeHunter: Ah, I see. Yes, I was just going by the basic n × 1.6≈, I didn't realize the temlplate was setup (sort of) to the exact number, including three decimal places. Thanks for clarifying that bit, though it only helps solve some of problem. Why have it set to calculate to the such a high degree of precision, only to try and immediately avoid it? For example, I'm still not sure why 2900 mi comes out as 4700 km? Shouldn't it be 4667 km? Is the template assuming/or set up that, if the miles are an even hundred or thousand, the result on the km side must also round all the way to nearest hundred or thousand? And doing so to avoid a "misleading sense of precision"? Because that seems to be remarkably imprecise. Whereas 2906 mi = 4677 km, which is much more exact. (Bear with me, I haven't really edited in quite some time, so I'm trying to shake of some rust. Your assistance, as well as anyone else's here, is appreciated.) Cheers - wolf16:49, 10 June 2024 (UTC)[reply]
@Thewolfchild:Template:Convert has documentation for how to request another precision than implied by the roundness of the input. 2900 mi is exactly 4667.0976 km. The input has two zeroes so the output is by default rounded to two zeroes and becomes 4700 km. It seems reasonable to me. "2900 mi" often in practice means "around 2850 mi to 2950 mi". That's 4587 km to 4747 km. If we apply the same rule to the template result 4700 km then it becomes "around 4650 km to 4750 km" which gives a fair overlap with the expectation from "2900 mi". This type of default rounding to the same precision as the input is a common practice and not something Wikipedia has invented. It will not be changed so a suggestion at Template talk:Convert would be a waste of time. PrimeHunter (talk) 17:06, 10 June 2024 (UTC)[reply]
Request some kind of change at yet another talk page? Nah. I found what appeared to be an oddity, if not a disparity, and so reported it here. But if you're good with it, then I think we're done here. - wolf23:45, 10 June 2024 (UTC)[reply]
Tech News: 2024-24
Latest tech news from the Wikimedia technical community. Please tell other users about these changes. Not all changes will affect you. Translations are available.
Recent changes
The software used to render SVG files has been updated to a new version, fixing many longstanding bugs in SVG rendering. [9]
The HTML used to render all headings is being changed to improve accessibility. It was changed last week in some skins (Vector legacy and Minerva). Please test gadgets on your wiki on these skins and report any related problems so that they can be resolved before this change is made in Vector-2022. The developers are still considering the introduction of a Gadget API for adding buttons to section titles if that would be helpful to tool creators, and would appreciate any input you have on that.
The HTML markup used for citations by Parsoid changed last week. In places where Parsoid previously added the mw-reference-text class, Parsoid now also adds the reference-text class for better compatibility with the legacy parser. More details are available. [10]
Problems
There was a bug with the Content Translation interface that caused the tools menus to appear in the wrong location. This has now been fixed. [11]
Changes later this week
The new version of MediaWiki will be on test wikis and MediaWiki.org from 11 June. It will be on non-Wikipedia wikis and some Wikipedias from 12 June. It will be on all wikis from 13 June (calendar). [12][13]
The new version of MediaWiki includes another change to the HTML markup used for citations: Parsoid will now generate a <span class="mw-cite-backlink"> wrapper for both named and unnamed references for better compatibility with the legacy parser. Interface administrators should verify that gadgets that interact with citations are compatible with the new markup. More details are available. [14]
On multilingual wikis that use the <translate> system, there is a feature that shows potentially-outdated translations with a pink background until they are updated or confirmed. From this week, confirming translations will be logged, and there is a new user-right that can be required for confirming translations if the community requests it. [15]
Would people be open to deploying a gadget similar to wikt:MediaWiki:Gadget-UnsupportedTitles.js on the English Wikipedia? The code there is somewhat specific to Wiktionary, but the idea is that pages like https://en.wiktionary.org/wiki/:%7C get JavaScript redirected to pages describing the characters in question, and it also uses JavaScript to fix the H1. I personally care less about the second issue than the first one, and would like to enhance it further so things like Building#19 get redirected to Building No. 19 rather than a nonexistent anchor in building. That part could be done using a template gadget that only loads on pages transcluding {{technical reasons}}. Not sure if the first part is feasible that way yet. * Pppery *it has begun...04:10, 11 June 2024 (UTC)[reply]
I pretty strongly believe that page titles should display the page title as accessible by the URL, regardless of whether that's the best title. Izno (talk) 06:27, 11 June 2024 (UTC)[reply]
That's not what's proposed. Pppery clearly say they "personally care less about" ... "us[ing] JavaScript to fix the H1". – SD0001 (talk) 07:20, 11 June 2024 (UTC)[reply]
If it's what e.g. isaacl came to, then I do probably still oppose implementation - fighting with MediaWiki over just how to navigate to a page sounds like a pure lose lose situation. Izno (talk) 15:54, 11 June 2024 (UTC)[reply]
It's a "pure lose lose" situation that readers get to read about Building #19 when they search for Building #19? Don't overuse hyperbole – it spoils its impact when actually needed. – SD0001 (talk) 17:03, 11 June 2024 (UTC)[reply]
No, I'm not being hyperbolic. Please don't be a dick. I sincerely don't think there's a win to "let's fuck around with anchors". Izno (talk) 23:24, 11 June 2024 (UTC)[reply]
mediawiki.action.view.redirect.js – MediaWiki has for decades, to use your language, fucked around with anchors (to resolve sections links on redirected pages). You not thinking it's a win does not mean it's suddenly considered a bad practise. – SD0001 (talk) 15:37, 12 June 2024 (UTC)[reply]
Not sure if the first part is feasible that way yet. It won't be feasible with a template gadget, but it would have been if phab:T241524 had been implemented instead, as you could inject the parser tag into the noarticletext interface message. – SD0001 (talk) 07:25, 11 June 2024 (UTC)[reply]
Wait a minute, turns out I'm wrong. Putting the interface message in the category does have the desired effect, even though it doesn't (obviously) cause pages using the message to show up in the category. – SD0001 (talk) 20:41, 11 June 2024 (UTC)[reply]
I agree that work along these lines would be better implemented as a core feature rather than a gadget. I also don't like trying to redefine how URLs with fragment IDs work. It makes the behaviour non-standard and so the advantage of readers leveraging their experiences with the rest of the web is diminished. isaacl (talk) 15:34, 11 June 2024 (UTC)[reply]
Didn't say anything about preaching. Just saying that following common web patterns means people know what to expect. A web page-based app uses a fragment ID to access a subresource of the page, as intended. Redefining the syntax to redirect to a completely different page is a different model. Sure, it can be done, but it's something unexpected. isaacl (talk) 00:49, 12 June 2024 (UTC)[reply]
And my point is that fact is irrelevant in almost all contexts, and it's unnecessary preaching to convey that point instead of taking people where they clearly want to be taken. But whatever, it's clear that, for reasons that make no sense to me, this is being shot down and we're instead choosing to deliberately get in people's way. * Pppery *it has begun...01:03, 12 June 2024 (UTC)[reply]
The majority of readers reach Wikipedia via search engines. Making it easier for search engines to know the right index phrases for an article will help readers the most, as most of them pay little attention to the characters in the URL. (For those that do, personally I'd rather not defy their expectations by showing a page with a title that differs from what appears before the URL fragment ID.) isaacl (talk) 01:40, 12 June 2024 (UTC)[reply]
I don't really like this idea. What good is having the page under an invalid title if you can't link to it (except as an external link)? All kinds of other interfaces also won't work or will display a different title as a result (e.g. what links here, watchlist, page view counts…). Matma Rextalk20:56, 12 June 2024 (UTC)[reply]
I agree with Matma Rex. This kind of reminds me of phab:T315893 and phab:T338151, which I am also disinclined to support. We should keep all the various systems (what links here, watchlist, page view counts, wikilinks, how the title displays when loading the page) in sync with each other, and try to avoid adding more complexity than what we already have (the redirect system, CirrusSearch accepting case insensitivity, unnecessary space character in mobile talk page titles, etc.). By continuing to pile on complexity to the title system, I see a lot of potential for technical debt here. Debugging becomes difficult as the system gets so complex that few can wrap their head around the entire thing. –Novem Linguae (talk) 21:48, 12 June 2024 (UTC)[reply]
Appearance dropbox menu
I have just noticed a new icon and associated dropdown menu "Appearance" just to the right of my user page link.(pair of spectacles?) It gives radiobox options for small, standard and large text. I like the idea, but it does nor work as I would expect. The selections change text size for existing article text and preview windows but edit window text size is unchanged by the selection. I assume this is a new feature, which I welcome, but the text size in the edit window is the one I really need to make bigger. Cheers, · · · Peter Southwood(talk): 15:00, 11 June 2024 (UTC)[reply]
Thanks for the feedback @Pbsouthwood! Could I ask which editor you're using? Currently, we have the Visual Editor keeping the same size as the text, and the wikitext editor keeping the smaller size. If you're using the wikitext editor, I'd be curious if you could tell me a bit more about why the wikitext editor would work better at this size for you? We're currently collecting feedback on the feature that we hope to use for future changes and configurations. OVasileva (WMF) (talk) 15:25, 11 June 2024 (UTC)[reply]
Describing my experience: right now I'm using the page zoom feature to make the all text sizes a bit bigger, both when reading and editing. If I reset this to 100% and use the accessibility for reading feature, then I can read with a larger font but would be editing (and previewing) with a smaller one. I'd have to increase the page zoom level anyway, and turn it back down when reading pages. So I'd rather just set page zoom once, as it will apply for all pages. isaacl (talk) 15:42, 11 June 2024 (UTC)[reply]
I always use the wikitext editor. Partly habit, partly to maintain my skills and partly because I get frustrated when visual editor doesn't do what I want. No doubt it is great for some things, but so far I have not found out what. My eyes are deteriorating and get tired quickly, so editing on a larger font helps a lot. I generally zoom in the whole page to see what I have just typed and make corrections. It is easier on desktop where I have a mouse at hand. On laptop I have no mouse because of no suitable surface for it most of the time so zoom in and out with touchpad. Before the tablet failed I would zoom with two fingers there as well. It would be nice to have the edit screen follow the others. Might also be nice to have a fourth font size option, even larger. Some eyes are worse than others, and more accessibility is better. Cheers · · · Peter Southwood(talk): 16:06, 11 June 2024 (UTC)[reply]
To be clear, for proofreading the wikitext I need bigger text to spot the errors. I can read the smaller text but spotting the errors needs better resolution for reduced eyestrain. Cheers · · · Peter Southwood(talk): 16:51, 11 June 2024 (UTC)[reply]
Just wanted to note that Olga has answered there. In short, this is about scannability and information density on pages that don't include long-from text. I'm not sure but perhaps we should move the discussion about font size on special pages on MediaWikiwiki? People from other wikis could chime in. SGrabarczuk (WMF) (talk) 22:53, 12 June 2024 (UTC)[reply]
Perhaps it should also be about legibility in general, and accessibility for visually impaired users, unless there is a technical issue making it too difficult, or you have plausible reason to believe it would be unwanted by enough users to make it a problem. Cheers, · · · Peter Southwood(talk): 07:15, 13 June 2024 (UTC)[reply]
I understood the version in permalink/1228796863, or you have plausible reason to believe it would not be wanted by enough users to make it a problem., but or you have plausible reason to believe it would be unwanted by enough users to make it a problem. says something very different and I'm having trouble making sense of it in context. -- Shmuel (Seymour J.) Metz Username:Chatul (talk) 11:23, 13 June 2024 (UTC)[reply]
Sorry to confuse. What I wanted to convey is that it does not really matter if a fair number of people fail to find it desirable, but it could be problematic if there are enough people who actively do not want it, which is the revised meaning. Wikipedia does have a history of people objecting strongly to things they feel have been pushed onto them without due process. Cheers, · · · Peter Southwood(talk): · · · Peter Southwood(talk): 13:32, 13 June 2024 (UTC)[reply]
Update - Today at Quarry, the list of "Recent Queries" show them all as Failed for the last 24 hours or so. Once more, asking for help fixing this issue. Thanks. JoeNMLC (talk) 13:26, 12 June 2024 (UTC)[reply]
Thanks @Xaosflux for the quick response. Good to know that issue is being addressed. Going forward I saved above info. in my (offline) notepad Query file. Thanks again. Regards, JoeNMLC (talk) 14:48, 12 June 2024 (UTC)[reply]
Hi all, I'm noticing an issue with {{Cite Cambridge History of China}}. If I do {{Cite Cambridge History of China|volume=1}} it links to the wrong volume (volume 6 instead of 1):
So, I've enabled the warning on a blank edit summary and, while it has helped me a lot, it's also really annoying in one specific case. When I want to make a redirect, the H:AES is more than enough since it perfectly describes what the edit is. If people want to know the motivation behind the redirect's creation, they can look at the WP:RCATS I've provided. So, in that case, I'm forced to click Publish twice so the edit goes through, even though I know that it is going to have an edit summary and one that I'm perfectly happy with. I wish there were a way to disable the warning when an automatic summary is generated in the case of redirect creation(obviously excepting the default undo summary which should still generate a warning). To be clear, I'm not asking to change how the setting works for everyone. I'm asking for a way to toggle it between the current behavior and the one I described above. Nickps (talk) 11:33, 13 June 2024 (UTC) After reading m:Help:Edit_summary#Automatic_summaries, I changed my request to only apply to redirect creation Nickps (talk) 11:38, 13 June 2024 (UTC)[reply]
"Prompt me when entering a blank edit summary (or the default undo summary)" is a software feature, not a community gadget, to request changes to its behavior you may file a feature request at phabricator. — xaosfluxTalk13:55, 13 June 2024 (UTC)[reply]
I'm not sure that AES is always the only thing needed when creating a redirect, yes it is useful but it only tells "what" was done, not "why" it was done, which is part of the usefulness of edit summaries. — xaosfluxTalk14:15, 13 June 2024 (UTC)[reply]
For example, we often tag or categorize our redirects with more information, like "redirect from misspelling", etc -- and not everyone would know about that, but if they say so in their edit summary, others can figure out their intent. — xaosfluxTalk14:17, 13 June 2024 (UTC)[reply]
I don't disagree with that, but, since most redirects are created for obvious reasons, I still think it should be left to the editor's discretion whether to rely on AES or not (that's why I'm asking for a second toggle instead of a change in the default behavior). I don't leave a custom summary on most redirects I make, and I don't think there's anything wrong with that. In the very rare case I think the redirect is needs an explanation, I will write a custom summary. In any case, the warning mostly gets in the way when I make redirects. Nickps (talk) 15:12, 13 June 2024 (UTC)[reply]
I am not seeing infobox borders anymore (atWalter Cronkite and Google for example). Is that a "me" issue, or is something broken? (I am on Debian/Firefox, so it might very well be a "me"/specific issue.)
The infoboxes look different for me as well, starting recently. But I'm not sure it's a technical problem, it could be a recently released change? Simeon (talk) 19:14, 13 June 2024 (UTC)[reply]
Images are the same cause. I'll file a separate task for them since it's not obvious to me what the best resolution is for them. Izno (talk) 20:07, 13 June 2024 (UTC)[reply]
For me, at least, the font size in infoboxes has changed to 90% of the default size instead of 88%, which it has been forever. In Vector 2010, the font size in infoboxes is still 88%. I am looking at John Dalton, for example. I have the (formerly default) "small" font size selected as my prose body font preference in the new radio-button switcher on the right-side toolbar. – Jonesey95 (talk) 19:50, 13 June 2024 (UTC)[reply]
Line spacing inside infoboxes? Yes, that would be this change. Line spacing outside? Probably worth a different section Izno (talk) 20:52, 13 June 2024 (UTC)[reply]
Did looking around and the infoboxes now look like they do on the Minerva skin, and the hatnotes on the top of the article also look like Minerva now. Not sure if this is intentional or not. --JackFromWisconsin (talk | contribs) 19:40, 13 June 2024 (UTC)[reply]
I'm not surprised this happened, I will poke the relevant task. And yes, the relevant task also caused the hatnote differences below. Izno (talk) 19:45, 13 June 2024 (UTC)[reply]
This seems to have broken a lot of infoboxes. The career history of every association football player is a misaligned mess; see the screenshot I've attached for an example. It seems like this change needs to be reverted until it's more polished. –IagoQnsi (talk) 20:12, 13 June 2024 (UTC)[reply]
Consider that your specific example is also how it displays on mobile and you should consider how best to remedy that regardless. This just made the issue visible for desktop as well. There probably needs to be some work done on the template to support small resolutions. Izno (talk) 20:25, 13 June 2024 (UTC)[reply]
A four-column table should be four columns. Not four items placed at seemingly-random horizontal alignments. I checked, and the HTML has the data as a table with several rows and four cells per row. Now, HTML tables go right back to HTML 3.2 (27 years ago), and it's always been the case that tables having multiple rows and multiple columns are presented in such a way that each cell is the same width as the other cells in the same column. How can this have been screwed up so badly? It looks as if all of the cells in a row have been merged into one, with proportionate spacing between the items. --Redrose64 🌹 (talk) 22:19, 13 June 2024 (UTC)[reply]
Look at the documentation for {{hatnote}} under Vector 2022. A WP:THURSDAY just happened; is there some change in MediaWiki that would've caused this? The CSS indicates that the change is intended to apply to any responsive skin. Aaron Liu (talk) 19:21, 13 June 2024 (UTC)[reply]
(putting this here because it seems related, feel free to move if needed) I noticed on mobile web that section headings that include both linked and unlinked text place each type of text into a separate column. I first encountered it at Wikipedia talk:WikiProject Film but I'm seeing it elsewhere. Any idea what might have caused it? RunningTiger123 (talk) 21:20, 13 June 2024 (UTC)[reply]
I don't know what caused it, but as per MOS:NOSECTIONLINKS there shouldn't be links in section headings in any case - Arjayay (talk)
Unrelated cause to the above, I think this is the heading change previously notified about and how DiscussionTools is or was dealing with stuff in headings? Matma RexIzno (talk) 21:32, 13 June 2024 (UTC)[reply]
Vector 2022 is responsive. I can reproduce that at screen width below 1120 pixels, the search box disappears, but a flat button with magnifying glass icon (or in dark mode) appears. Clicking on it shows the search box. To focus on the search box, you can also use accesskeyF.
Came here to raise similar concerns. We have articles like SSUnited States with two infoboxes tucked inside one another having extraordinarily wiiiiide boxes, to the point that articles are hard to read. I've seen broken infoboxes thinner than the current ones.