I can see it in Firefox. But in Safari, the article ends with that section heading. And if I search for "Hiatt" on the page, it's highlighted in the invisible section, but I still can't see any text. I'm using Safari 17.4 and running Mac OS Sonoma 14.4. I also posted this to the article's Talk page. Peterh6658 (talk) 02:17, 23 March 2024 (UTC)[reply]

 Fixed. – Jonesey95 (talk) 04:41, 23 March 2024 (UTC)[reply]

Thanks! Why was it visible in Firefox but not Safari? Peterh6658 (talk) 07:14, 28 March 2024 (UTC)[reply]

@Peterh6658: Going to the relevant section in this old version, using Firefox 124, I see that the page is in two columns at that point. The left-hand column has the "Soundtracks" and "As Session Musician" subsections, plus the "Films" and "Written works" sections and the beginning (4 refs) of the "References" section; on the right, we have the remainder of the references, plus the "External links" and navboxes. This two-column layout ends after the navboxes but before the categories box, so something in the MediaWiki software is injecting sufficient </div> tags at that point to finish the page neatly. Viewing the HTML source (Ctrl+U) shows that this is indeed the case: there are <div class="div-col"> tags at each of the four positions that the Wikitext has a {{div col}}, there are </div> tags at each of the two positions that the Wikitext has a {{div col end}}, plus four </div> tags before the category box where I would expect there to be two - one to terminate the last navbox and one to end the page content.

This imbalance was caused by some earlier edits that removed a {{div col end}} from the bottom of the "Soundtracks" subsection and subsequently replaced it with a {{div col}}. This meant that with the {{div col}} being still at the top of the "Soundtracks" section, and another just below the "As Session Musician" subheading, there were then three nested {{div col}}, only one of which was subsequently closed. It could be that Safari doesn't like nesting in this manner; I can't test it, because the most recent version of Safari for Windows (Safari 5.1.7) was way back in May 2012. --Redrose64 🌹 (talk) 16:25, 28 March 2024 (UTC)[reply]

Thanks, Redrose64 🌹, for teaching us "Drawling, Stretching, and Fainting in Coils." (reply inspired by the song of the latter's same name by the band Bruford on their first proper album, said song being in turn inspired by you-know-who) Peterh6658 (talk) 17:45, 28 March 2024 (UTC)[reply]

The Mock Turtle's Story is where I remember it from. --Redrose64 🌹 (talk) 19:04, 28 March 2024 (UTC)[reply]

Just encountered this Twitter thread where someone looked at a page history on mobile, and thought that the bytes ± indicator was actually part of an upvote/downvote system. This isn't the first time I've seen someone make this mistake. On desktop, the bytes ± number appears right next to the total bytes and lots of other dense information, which I think makes it clearer that it's not a voting system, but on mobile it's more ambiguous. Perhaps it would make sense to add the word "bytes" on the mobile site? Just a letter "B" could work if "bytes" doesn't fit – not everybody would know what it means, but at least they'd be less likely to think it's a voting score. –IagoQnsi (talk) 21:54, 24 March 2024 (UTC)[reply]

The "hover text" on that explains it is bytes, but it is otherwise absent there. Adding a units label to that would best be done upstream, you may make an enhancement request with the details. — xaosflux ^Talk 10:17, 25 March 2024 (UTC)[reply]

The hover text in mobile page histories is MediaWiki:Tooltip-last which says "Difference with preceding revision". In user contributions it is MediaWiki:Rc-change-size-new which says "$1 bytes after change of this size" (I added "of this size" in 2017). Desktop uses MediaWiki:Rc-change-size-new in both page histories and contributions. PrimeHunter (talk) 16:08, 25 March 2024 (UTC)[reply]

Looks like you get the "size-new" if you have "advanced mode" enabled in mobile, and the other if you are not in advanced mode. Both are not very useful for the use case of non-hovered viewing. — xaosflux ^Talk 17:12, 25 March 2024 (UTC)[reply]

In this case we seem to have quite a few choices, but since Cifaldi has taken an interest it may be worth a Twitter user telling him that he's welcome to donate photos he owns the copyright to, which can influence which photo(s) of him we use in the infobox of Frank Cifaldi. — Bilorv (talk) 22:45, 29 March 2024 (UTC)[reply]

Wikipedia:Syndication#Watchlist feed with token describes how to use an RSS feed with a watchlist tokem. However, when I try it, the filtering parameters don't seem to work. The documentation for these parameters is at mw:API:Watchlist feed. Here are some tests:

• Watchlist feed with hours=72 (hours does seem to work)
• Watchlist feed with hours=72 and wltype=categorize - this shows an edit for me, which isn't a category change
• Watchlist feed with hours=72 and wlshow=bot - this shows an user edit for me, which isn't a bot edit

Am I missing something here? — Qwerfjkltalk 21:22, 27 March 2024 (UTC)[reply]

@Qwerfjkl try building your query using the guided form here, see if you get the results you expect, and if the string is in the same format. If the string format is different, the documentation may be outdated. If the results are wrong, you can open a bug report on it. — xaosflux ^Talk 12:52, 29 March 2024 (UTC)[reply]

Xaosflux, just tried it, seem to be a bug in the api. I'll try to file a ticket. — Qwerfjkltalk 13:04, 29 March 2024 (UTC)[reply]

phab:T361367 — Qwerfjkltalk 13:18, 29 March 2024 (UTC)[reply]

I’m not sure if this is a Wikipedia issue or a me issue but I thought I might as well mention it.

The issue is how the tense alveolar lateral approximate symbol (a.k.a ⟨l⟩ with a tense symbol) having a weird space after it which is odd since it isn’t there on Wiktionary. E.g. /l͈/ /l͈o/ /el͈o/ display similar to /l͈ / /l͈ o/ /el͈ o/. 2001:BB6:B84C:CF00:39A0:EB40:5711:F8C2 (talk) 12:41, 28 March 2024 (UTC)[reply]

You must be referring to the IPA on the Áed Oirdnide article. After a small bit of messing around, I found that this seems to be an issue with MinervaNeue, as there is no space when I view it in desktop mode (Vector 2022). — Mugtheboss (talk) 12:13, 29 March 2024 (UTC)[reply]

More likely, it’s just font specific issue, and it just depends on the character fallback chain that the browser calculates depending on the fonts installed on the device. —TheDJ (talk • contribs) 19:31, 29 March 2024 (UTC)[reply]

Hi, what is the process for directly generating data into an infobox from Wikidata? Riad Salih (talk) 18:15, 29 March 2024 (UTC)[reply]

Riad Salih, it is done using Module:WikidataIB. An example of a template which uses this module is Template:Infobox bridge. —⁠andrybak (talk) 12:04, 30 March 2024 (UTC)[reply]

I thought that we discouraged this, on several grounds - such as verifiability and the potential for undetected vandalism. --Redrose64 🌹 (talk) 13:49, 30 March 2024 (UTC)[reply]

The relevant RfC did not reach a general consensus, but to quote the close there is a consensus on one point: if Wikipedia wants to use data from Wikidata, there needs to be clear assurances on the reliability of this data. I don't think anyone has yet come up with a convincing way to provide that assurance. Mike Christie (talk - contribs - library) 15:14, 30 March 2024 (UTC)[reply]

There's other 1,000 articles where the 'reference' from Wikidata is just an error message, see Category:Module:Wd reference errors. -- LCU ActivelyDisinterested «@» °∆t° 21:36, 30 March 2024 (UTC)[reply]

Correction: the reference from Wikidata is being rendered as an error message. Actual data does exist on Wikidata that the template can't render properly. * Pppery * _{it has begun...} 23:02, 30 March 2024 (UTC)[reply]

That's a difference without a distintion, the effect in Wikipedia is the same. The reference in Wikipedia is an error message. -- LCU ActivelyDisinterested «@» °∆t° 23:52, 30 March 2024 (UTC)[reply]

Resolved

At 2018–19 Big Ten Conference men's basketball season, the rankings section has borders while neither the coaches section nor the preseason national polls section shows borders, but they all seem to be styled by the {{CollegePrimaryStyle}}. What makes one section present cell padding/border and the others not?-TonyTheTiger (T / C / WP:FOUR / WP:CHICAGO / WP:WAWARD) 18:27, 29 March 2024 (UTC)[reply]

@TonyTheTiger: Many of the cells are missing quotation marks in style="...". {{CollegePrimaryStyle|Illinois Fighting Illini}} produces background-color:#13294B;color:#FF5F05;box-shadow: inset 2px 2px 0 #FF5F05, inset -2px -2px 0 #FF5F05;. Without quotation marks, only some of it is included in the style. The documentation at {{CollegePrimaryStyle}} should probably be updated to say that if it's assigned directly to style= in a cell then use quotation marks. The documentation seems to currently assume it's used as a parameter in a template which adds quotation marks. PrimeHunter (talk) 19:14, 29 March 2024 (UTC)[reply]

@TonyTheTiger: To expand on that: style= is a HTML attribute, and attribute values must be quoted unless they consist entirely of the 64 characters A-Z, a-z, 0-9, hyphen-minus and full stop. When you use something like style=background-color:#13294B;color:#FF5F05;box-shadow: inset 2px 2px 0 #FF5F05, inset -2px -2px 0 #FF5F05; this contains several characters which are not among the 64: there are three colons, four hashes, three semicolons, several spaces and a comma. How this is treated will depend upon the browser. Some may reject the whole attribute outright; some will ignore everything after the first 'invalid' character (the first colon), some may ignore the lack of quotes and process the whole value as if it were quoted. --Redrose64 🌹 (talk) 20:39, 29 March 2024 (UTC)[reply]

@PrimeHunter and Redrose64:, Neither style=“{{CollegePrimaryStyle|Illinois Fighting Illini}}; width=75” nor style=“{{CollegePrimaryStyle|Illinois Fighting Illini}}"; width=75 seems to be the proper correction for style={{CollegePrimaryStyle|Illinois Fighting Illini}}; width=75 in the Preseason national polls section.-TonyTheTiger (T / C / WP:FOUR / WP:CHICAGO / WP:WAWARD) 21:23, 29 March 2024 (UTC)[reply]

Well, no. The first one is invalid because you're trying to put width=75 inside a style= attribute, and it's not valid CSS. The second one is invalid because you've put a semicolon after the closing quote - after the closing quote of an attribute's value, only two characters are valid: (i) a space; (ii) a greater-than character >. --Redrose64 🌹 (talk) 21:57, 29 March 2024 (UTC)[reply]

style=“{{CollegePrimaryStyle|Illinois Fighting Illini}}” width=75 does not work either.-TonyTheTiger (T / C / WP:FOUR / WP:CHICAGO / WP:WAWARD) 22:04, 29 March 2024 (UTC)[reply]

I don't see any recent (later than 20:39 today) related edits at 2018–19 Big Ten Conference men's basketball season - are you doing this elsewhere? --Redrose64 🌹 (talk) 22:20, 29 March 2024 (UTC)[reply]

I am not going to save stuff that makes the table break when I hit show preview. How would you make those other two tables show the secondary color for the borders?-TonyTheTiger (T / C / WP:FOUR / WP:CHICAGO / WP:WAWARD) 22:44, 29 March 2024 (UTC)[reply]

User:TonyTheTiger/sandbox/2018–19 Big Ten Conference men's basketball season test is the failed test.-TonyTheTiger (T / C / WP:FOUR / WP:CHICAGO / WP:WAWARD) 04:27, 30 March 2024 (UTC)[reply]

@PrimeHunter and Redrose64:, thoughts on my test?-TonyTheTiger (T / C / WP:FOUR / WP:CHICAGO / WP:WAWARD) 04:46, 30 March 2024 (UTC)[reply]

@TonyTheTiger: You didn't use straight quotation marks.[1] PrimeHunter (talk) 12:02, 30 March 2024 (UTC)[reply]

The HTML spec permits the use of either U+0022 QUOTATION MARK (") or U+0027 APOSTROPHE (') as value delimiters; by implication, the “ and ” characters are not delimiters but are treated as part of the value itself. --Redrose64 🌹 (talk) 13:02, 30 March 2024 (UTC)[reply]

• In a related matter, can you help me interpret the differences between {{CollegePrimaryStyle}} and {{NCAA color cell}}. It seems the former tries to use the true primary and secondary school colors, but the latter seems to use one of the two that constrasts well against white text. Is this correct?-TonyTheTiger (T / C / WP:FOUR / WP:CHICAGO / WP:WAWARD) 16:00, 30 March 2024 (UTC)[reply]

  In both cases, the work is done through Module:College color. --Redrose64 🌹 (talk) 16:06, 30 March 2024 (UTC)[reply]

  What do those 5 colmns of contrast mean and how do these two draw from the three columns of colors?-TonyTheTiger (T / C / WP:FOUR / WP:CHICAGO / WP:WAWARD) 17:17, 30 March 2024 (UTC)[reply]

  There are three columns of colour dabs, headed 1, 2 and 3. The first three of the five contrast columns show how each of those relate to one of the others - 1/2 is for colours 1 and 2 in conjunction (either text of colour 1 on a background of colour 2 or text of colour 2 on a background of colour 1); similarly 2/3 is for colours 2 and 3 in conjunction; and 3/1 is for colours 3 and 1 in conjunction. The last two, 1/w and 1/b, are for colour 1 in conjunction with white or black respectively. --Redrose64 🌹 (talk) 18:31, 30 March 2024 (UTC)[reply]

  It seems that {{CollegePrimaryStyle}} uses 1 & 3 and {{NCAA color cell}} uses 1 & 2. Is that correct? Or are there some conditions?-TonyTheTiger (T / C / WP:FOUR / WP:CHICAGO / WP:WAWARD) 18:58, 30 March 2024 (UTC)[reply]

  I think I am up to speed. Thx.-TonyTheTiger (T / C / WP:FOUR / WP:CHICAGO / WP:WAWARD) 04:50, 31 March 2024 (UTC)[reply]

I created a redirect page, and a human editor reviewed it. Unfortunately the title was a typo. I moved the redirect to the correctly-spelt title, and User:DannyS712 bot III okayed that.

In my notifications, it then appeared that the human had okayed the new title, and the bot had okayed the old typo one. I don't think this is intended behaviour. For redirects, it might give the impression that an editor has approved a highly inappropriate redirect.

I now think I should have RfDd the redirect before moving: Wikipedia:Redirects_for_discussion/Log/2024_March_31#Medical-grade_silicon

HLHJ (talk) 17:04, 31 March 2024 (UTC)[reply]

I started typing a reply to a discussion at ANI and a new edit happened so I refreshed to see it (that edit closed the discussion).
The problem now, is that every time I go to ANI it tries to do the "Changes recovered" thing where it scrolls to that topic and says Loading..., and normally would have loaded my reply, but since it's closed it gives up after a bit.
This happens every time I go there now, I have tried starting another reply on another section and cancelling that, but no luck. – 2804:F14:8093:5F01:BD93:DDC2:7C48:C2EC (talk) 01:03, 1 April 2024 (UTC)[reply]

You have to temporarily unclose the thread (I usually just comment out the top and bottom templates), this will allow you to close the reply window, and you then revert the changes so the thread is closed again. I ran into this exact problem yesterday and had to temporarily re-open and re-close a thread.[2][3] -- LCU ActivelyDisinterested «@» °∆t° 13:50, 2 April 2024 (UTC)[reply]

I just noticed today that I see the following script warnings when using "Show preview" for any edit that includes use of {{marriage}}.

Script warning: templatename used with unknown parameter(s): paramname.

Script warning: templatename used with deprecated parameter(s): paramname.

Script warning: templatename used with duplicate parameter(s): paramandvalue.

There have not been any recent changes to {{marriage}}, so I guess that there has been a change in an underlying module or template.

• OS: Microsoft Windows 10
• Browser: Microsoft Edge Version 123.0.2420.65 (Official build) (64-bit)
• Skin: Vector legacy (2010)

I also see it when logged out using:

• Google Chrone Version 123.0.6312.86 (Official Build) (64-bit)
• Opera One(version: 109.0.5097.35)

Any ideas on how to track this down?  — Archer1234 (t·c) 02:54, 1 April 2024 (UTC)[reply]

Fixed. * Pppery * _{it has begun...} 03:11, 1 April 2024 (UTC)[reply]

With this edit to Module:If preview, for the record. Graham87 (talk) 10:16, 1 April 2024 (UTC)[reply]

PEIS is becoming an issue for one of our larger articles, Donald Trump. An editor is currently stating because of PEIS we cannot pre-emptively archive sources as prescribed at WP:DEADREF and WP:ARCHIVEEARLY. The current limit is 2MB, and has apparently been set to that for decades. The initial reasoning was to prevent denial-of-service attacks through large complex pages overwhelming MediaWiki. But with the significant change in processing power since that initial limit (and the growing complexity of pages on Wikipedia), it is time to revisit that limit. I'd actually suggest at least doubling it to 4MB, but a higher value may be reasonable because (again) computational power has increased a lot since that initial limit was created. The linked task has been active since 2018 (six years ago). I'm not entirely sure what is needed to change this setting, but if you were waiting for an actual article to hit the limit, we're here. —Locke Cole • t • c 17:35, 1 April 2024 (UTC)[reply]

FWIW this was also discussed last December; the Phabricator discussion you link above reminds me very much of the one about increasing the Lua memory limit, and I hope he won't mind me saying this, but my suggestion is: ping Tim Starling and/or leave him a talk page message—and if you can identify any other dev who seems to have expertise in this area, ping them too—so he can at least give a second/informed opinion on whether increasing this limit is reasonable. Maybe the answer is no, but maybe the answer is yes, since much has changed since the limit was set 18(?) years ago. -sche (talk) 20:33, 1 April 2024 (UTC)[reply]

Looks like he has email turned on, so sent him an email directly with a summary of the situation. —Locke Cole • t • c 22:22, 1 April 2024 (UTC)[reply]

At the time (18 years ago), long articles about US presidents like George W. Bush took about 30 seconds to parse, and by imposing limits, we were trying to keep user-visible latency down to around that figure. Now today Donald Trump takes 9 seconds to parse, so we can congratulate ourselves for directing some small part of the enormous hardware performance gains to an improvement in user-visible latency. But, you might say, human patience remains the same, so we should increase the limits by a factor of 3 or so, allocating the remaining performance dividend to editors so that they can increase the size and complexity of articles by that amount. However, a cloud on the horizon approaches — Parsoid read views are imminently coming, and Parsoid does Donald Trump in 21 seconds. They have staked their claim on half of the gap between current and historical performance. Wirth's law in our context is the process by which performance improvements are greedily consumed by various stakeholders, all hungry for clock cycles, up to the limit of human tolerance for latency, and you might well complain that editors here are getting the crumbs. Parsoid performance is quite sensitive to the number of tags on the page, and if you were optimising for it, removing unnecessary archive links would be a reasonable step to take. So the limit is still doing its job. -- Tim Starling (talk) 01:20, 2 April 2024 (UTC)[reply]

... removing unnecessary archive links would be a reasonable step to take. Archive links aren't what I'd call "unnecessary", especially as they've taken a hammer (disallowed ALL archive links for URLs that are still live) to the situation instead of a scalpel (disallowed archive links for specific sources known to be reliable/persistent).

Regardless, the overarching point is, we have language at WP:CITE that says we should be proactive in archiving sources to prevent links from going stale/dead. All things being equal, around 99.9% of the project follows this guidance and archives are added either automatically or semi-automatically using scripts. What we're running in to here is a technical restriction, however, in WP:PEIS. On an article as critical as one for someone facing numerous legal challenges, being the current party-nominee for a Presidential election, and otherwise being very much in the public eye, it seems like pre-emptively archiving sources would be more desirable, not less so.

So the limit is still doing its job. Forgive me, but the initial reason for this was seemingly related to fears of denial-of-service attacks abusing markup and causing the site to stall or fail completely. Not to force editors to limit what can be done with articles. —Locke Cole • t • c 03:05, 2 April 2024 (UTC)[reply]

That article needs to do better at WP:SUMMARY given its current size of 425 kb wikitext. Cut the article down by a third to a half of its current size and you won't be running into template expansion issues anymore. Izno (talk) 05:03, 2 April 2024 (UTC)[reply]

I agree. An article that is close to hitting the limit is a great candidate for WP:SPINOUTs. Even if we have the technical ability to raise the limit, it may be beneficial to keep it the same, in order to discourage the creation of articles that are too big. –Novem Linguae (talk) 05:16, 2 April 2024 (UTC)[reply]

@Izno and Novem Linguae: You'll get no argument from me. My primary concern is going into a US election season with archives banned for a page that is BEGGING to be as verifiable as possible. Sometimes sources are paywalled. Sometimes sources go down temporarily. Sometimes sources move because of a site restructure or an article getting renamed. The list of reasons to have archives available on a page like this is far longer than the list of reasons not to. My goal with this was to find the least disruptive (to the regular editors of Donald Trump to make the page editable and get back to including archive links. If someone thinks they can split it off even more than it already is, be my guest. —Locke Cole • t • c 05:27, 2 April 2024 (UTC)[reply]

@Locke Cole: Given the Donald Trump article is already split into myriad existing subarticles, so long as the sources used in Donald Trump are the ones used in the relevant main articles (if they're not they can be copied to the main articles), they can be archived at the relevant main articles and so preserved in case they are later needed. CMD (talk) 05:05, 3 April 2024 (UTC)[reply]

You're welcome to bring that up at WT:CITE where archiving is prescribed as a way to avoid dead links. I do think it's a bad idea to carve out exceptions like this though, as these archive links are also useful to our readers. Having them absent entirely on some pages creates an inconsistent experience, and again, WP:V is policy and WP:CITE works to standardize how we make our articles verifiable. —Locke Cole • t • c 05:17, 3 April 2024 (UTC)[reply]

Completely agree. If you've got to the point of fiddling around with reference templates to try and make you article work, your likely well past the point it's should better summarised or split. -- LCU ActivelyDisinterested «@» °∆t° 13:53, 2 April 2024 (UTC)[reply]

Worth bearing in mind that the limit here is only partially technical - it also serves to limit articles to a reasonable size. Computers may be able to handle doubling or even quadrupling the limit, but would the articles produced actually end up being readable? If you attempt to print out the current Donald trump article it runs to 26 sides of A4, with 12 pages being the article text, that seems awfully long for what is supposed to be a high level overview of a topic. Would a 50 or 100 page long article actually represent an improvement? 86.23.109.101 (talk) 10:36, 2 April 2024 (UTC)[reply]

In that case there should be a limit on article size, not post-expand include size (which in 90% of cases is driven by nested templates-within-templates or modules-within-templates such as {{navbox}} and {{navboxes}}, {{flag}}, {{coord}}, and {{cite web}} (et al.), none of which appreciably increase article size). --Ahecht (TALK
PAGE) 18:35, 2 April 2024 (UTC)[reply]

You could easily reduce the WP:PEIS of that article from ~1.8MB to 1.1MB by using {{#invoke:cite web|‍}}, {{#invoke:cite news|‍}}, {{#invoke:cite book|‍}}, etc. instead of {{cite web}}, {{cite news}}, {{cite book}}, etc. --Ahecht (TALK
PAGE) 15:22, 2 April 2024 (UTC)[reply]

That's a good tip. Implemented here, it brought the PEIS down to Post‐expand include size: 1261011/2097152 bytes as expected. —Locke Cole • t • c 15:40, 2 April 2024 (UTC)[reply]

IMO that's a bad suggestion. We normally avoid using #invoke or other parser functions directly in articles as it makes the wikitext even harder to understand. Anomie⚔ 16:01, 2 April 2024 (UTC)[reply]

It's fiddling at the edges of a problem, rather than dealing with the scale of the article. -- LCU ActivelyDisinterested «@» °∆t° 16:09, 2 April 2024 (UTC)[reply]

(edit conflict) I agree. But it's better than the references not working. The fact that PEIS almost always results in tech hacks that don't even clearly improve parsing time rather than the article getting smaller. It's an interesting question (which Tim Starling (WMF) could possibly answer) whether {{#invoke:cite web|...}} is actually meaningfully faster to parse that {{cite web|...}}. Because it's largely things like that the the PEIS limit is forcing people to do, not split articles.

The only exception I can think of to this, where an article was split which probably wouldn't have been otherwise, is List of Ang Probinsyano seasons FKA List of Ang Probinsyano episodes. * Pppery * _{it has begun...} 16:10, 2 April 2024 (UTC)[reply]

Just pointing out that at least 73 pages were directly split off from or are related to Donald Trump. I don't understand what these two edits are supposed to do on the page or how the second one follows from the first one: Reduce WP:PEIS and WP:PEIS improvements from WP:VPT courtesy of User:Ahecht. The first "improvement" increased the size by 270 bytes, the second one by 5,842. Space4Time3Continuum2x🖖 16:52, 2 April 2024 (UTC)[reply]

"Post-expand include size" is not the size of the article. It roughly means the total amount of wikitext generated by templates/modules, with wikitext being sent through a template/module more than once being counted multiple times. Thus, by reducing the number of layers of templates by one, those edits did improve the PEIS. * Pppery * _{it has begun...} 16:56, 2 April 2024 (UTC)[reply]

Let's be very clear. This discussion is about increasing the PEIS limit and related topics, not about the Donald Trump article. No matter what happens here, any change to the article's current consensus item 25 would need a new consensus at Talk:Donald Trump. The same goes for any talk about splitting or trimming, and there's currently an open discussion about that. Thanks guys. ―Mandruss ☎ 04:39, 3 April 2024 (UTC)[reply]

The buttons of connected gadgets, for example, the wikifier after updating the page are mixed in the middle and back to the right. Where can this be fixed so that the buttons do not move and leave in one place, right or left, please help? In Russian, all the gadget buttons on the left. Here is the screen where I noted these buttons, when they went to the middle. Thank you! Takhirgeran Umar (talk) 22:01, 1 April 2024 (UTC)[reply]

Your post is confused. ce.wikipedia.org links to an English Wikipedia article but you apparently refer to source edit pages at ce:. ce:MediaWiki:Gadgets-definition has the line:

• AlignTemplateParameters[ResourceLoader|default|hidden|type=general|actions=edit|dependencies=ext.gadget.registerTool,jquery.client]|AlignTemplateParameters.js

It loads ce:MediaWiki:Gadget-AlignTemplateParameters.js by default but hides it in preferences. The gadget adds some buttons to the edit toolbar. Compare [4] to safemode which doesn't have the extra buttons. The position of the buttons vary between page loads for me and you apparently want the position to be fixed. I don't know how to control this. PrimeHunter (talk) 22:59, 1 April 2024 (UTC)[reply]

@PrimeHunter. On this screenshot everything is ok, I want them not to go left or to the center. Sorry I don’t know English, I write with Google translator, I hope now more understandable. Takhirgeran Umar (talk) 23:24, 1 April 2024 (UTC)[reply]

Here they went to the center. It is necessary that they stay all the time on the right. Takhirgeran Umar (talk) 23:30, 1 April 2024 (UTC)[reply]

I see this problem through the Mozilla Firefox browser. Takhirgeran Umar (talk) 07:07, 2 April 2024 (UTC)[reply]

MediaWiki message delivery 03:33, 2 April 2024 (UTC)[reply]

I just realized, that it is possible to indent tables.

::::{| class="wikitable collapsible collapsed"
! table header
|-
| table cell
|}

This seems like an accidental feature, that could easily be broken with the next CSS change.
But I think it is quite useful on talk pages, and should be kept.
Especially discussions about illustrations become unwieldy, when too many thumbs pile up on the right.

So my suggestion is just, to recognize this as a feature, that should be kept. Watchduck (quack) 14:13, 2 April 2024 (UTC)[reply]

@Watchduck: Please see Help:Table#Indenting tables.

Indenting with colons (ab)uses HTML definition lists, and in HTML, tables may be used inside all three kinds of list - definition, ordered and unordered. It's highly unlikely that the feature will be removed, certainly not with the next CSS change. --Redrose64 🌹 (talk) 14:47, 2 April 2024 (UTC)[reply]

In addition, indenting tables is liable to cause this lint. Please don't. Izno (talk) 20:30, 2 April 2024 (UTC)[reply]

@Redrose64 and Izno: It would probably be easy to replicate the indentation with CSS classes. I think it would make sense to combine wikitable collapsible collapsed into something like talktable. The indentation depth could be added with a dash, so the first line would look like {| class="talktable-4". But counting colons is not something people like to do. Would it not be possible to make indented tables official, so that ::::{| would be translated into a table tag with class indented-table-4? — Preceding unsigned comment added by Watchduck (talk • contribs) 23:05, 2 April 2024 (UTC)[reply]

Feel free to file a feature request. --Redrose64 🌹 (talk) 23:21, 2 April 2024 (UTC)[reply]

No, it's not really reliably possible to do that in the current parser. It may be more possible in the future.

As for the text, there has been discussion about supporting multiline content in lists better already that would involve some new wikitext and would help us support more use cases than just tables. You can look on Phabricator for that if you want. Izno (talk) 00:53, 3 April 2024 (UTC)[reply]

The bot is removing the "vital article" parameter from every talk page, in alphabetical order. Can someone put it to sleep until its operator can sort it out? Pinging @Kanashimi and MSGJ:, who might know what is going on. ~~ AirshipJungleman29 (talk) 15:19, 2 April 2024 (UTC)[reply]

@AirshipJungleman29: Does User talk:Cewbot/Stop this stop it when you fill in the form? ——Serial Number 54129 15:26, 2 April 2024 (UTC)[reply]

No. ~~ AirshipJungleman29 (talk) 15:36, 2 April 2024 (UTC)[reply]

Bloody hell they last edited 17 and 6 hours ago! ——Serial Number 54129 15:28, 2 April 2024 (UTC)[reply]

Rogue bot, clearly. I have no knowledge of technical stuff, so I didn't want to break anything further. ~~ AirshipJungleman29 (talk) 15:31, 2 April 2024 (UTC)[reply]

Me neither! I hope you don't mind; I copied it over to ANI for more eyes. etc., as it certainly counts a (hopefully temporary!) 'chronic behavioural problem'  :) ——Serial Number 54129 15:36, 2 April 2024 (UTC)[reply]

I've blocked it for now. HJ Mitchell | Penny for your thoughts? 15:42, 2 April 2024 (UTC)[reply]

Thank you for putting it out of its misery. LegalSmeagolian (talk) 15:47, 2 April 2024 (UTC)[reply]

Thanks @HJ Mitchell:, very much for stepping up there. Now... can anyone mass-rollback over 1000 edits?!  :) ——Serial Number 54129 15:52, 2 April 2024 (UTC)[reply]

@Serial Number 54129 I can from a technical perspective with a few clicks. But might it not be better to wait for the botop to fix the bot and then let it fix its own errors? Is it an emergency now that the immediate problem is resolved. HJ Mitchell | Penny for your thoughts? 15:56, 2 April 2024 (UTC)[reply]

ooo... i didn't notice this follow up. I have rollback all the changes up till when the issue started after seeing someone rollbacking the changes one by one in my watchlist, and I thought I would save them the trouble and time spent going through some 1,600 edits. – robertsky (talk) 21:19, 2 April 2024 (UTC)[reply]

@Robertsky: About that, now that I'm looking closer, it seems your rollbacks and @Remsense's undos reverted multiple edits at once in various talk pages - only the most recent edit in each talk page appears to have been part of this malfunction (usually 10 bytes changes?).

I'm hoping the bot will fix those when it's fixed. Does anyone know if it will? – 2804:F14:80EC:AB01:D0C2:97E3:6645:A903 (talk) 21:46, 2 April 2024 (UTC)[reply]

Robertsky is duly thanked for his service :) ——Serial Number 54129 16:48, 2 April 2024 (UTC)[reply]

Odd that at WP:Database reports/Vital articles update report it says "36135 talk page(s) to edit (The amount of talk pages to edit exceeds the value of talk_page_limit_for_editing on the configuration page. Do not edit the talk pages at all.).", and yet the bot was happily trying to edit anyways. – 2804:F1...45:A903 (talk) 20:59, 2 April 2024 (UTC)[reply]

In case this helps any, here are the 250 edits near and before the beginning of the malfunction: [8] – 2804:F1...45:A903 (talk) 21:18, 2 April 2024 (UTC)[reply]

In general, this is the wrong forum for a bot malfunction. You should post in some order to the operator's talk page, WP:BOTN, and/or one of WP:AN/WP:ANI for the appropriate attention. ANI is most likely to get an immediate block, and you have to notify the operator of the post there regardless. Izno (talk) 21:45, 2 April 2024 (UTC)[reply]

@HJ Mitchell I'm sorry, I didn't notice that the configuration was corrupted. I've restored the profile, please unblock the robot, thank you. Kanashimi (talk) 22:24, 2 April 2024 (UTC)[reply]

Also, it should be enough to report it on the User talk:Cewbot/Stop page, without blocking it, because this bot is also responsible for other operations. Kanashimi (talk) 22:26, 2 April 2024 (UTC)[reply]

I've taken some precautions so that in the future, when this happens, the robot will stop editing. Kanashimi (talk) 22:31, 2 April 2024 (UTC)[reply]

Progress report: I am re-running the robot to recover the error. Kanashimi (talk) 01:27, 3 April 2024 (UTC)[reply]

 Done Kanashimi (talk) 03:25, 3 April 2024 (UTC)[reply]

Is it possible to format a regular table so that there is a large entry or two on a lower row, the way that tv show episode tables are formatted? —blindlynx 22:39, 2 April 2024 (UTC)[reply]

@Blindlynx Yes, you can use colspan to specify how many columns a single cell will span:

{| class="wikitable"
| R1 C1
| R1 C2
| R1 C3
|-
| colspan = "3" | R2
|}

--Ahecht (TALK
PAGE) 23:51, 2 April 2024 (UTC)[reply]

I assume from your edits that this is about List of genocides#List of genocides. In a sortable table you can add class="expand-child" from Help:Sortable tables#Keeping some rows together. You may also want something like style="border-bottom:solid 2px" to show what belongs together:

PrimeHunter (talk) 02:49, 3 April 2024 (UTC)[reply]