If you want to report a JavaScript error, please follow this guideline. Questions about MediaWiki in general should be posted at the MediaWiki support desk. Discussions are automatically archived after remaining inactive for five days.
This tends to solve most issues, including improper display of images, user-preferences not loading, and old versions of pages being shown.
No, we will not use JavaScript to set focus on the search box.
This would interfere with usability, accessibility, keyboard navigation and standard forms. See task 3864. There is an accesskey property on it (default to accesskey="f" in English). Logged-in users can enable the "Focus the cursor in the search bar on loading the Main Page" gadget in their preferences.
No, we will not add a spell-checker, or spell-checking bot.
You can use a web browser such as Firefox, which has a spell checker.
If you changed to another skin and cannot change back, use this link.
Alternatively, you can press Tab until the "Save" button is highlighted, and press Enter. Using Mozilla Firefox also seems to solve the problem.
If an image thumbnail is not showing, try purging its image description page.
If the image is from Wikimedia Commons, you might have to purge there too. If it doesn't work, try again before doing anything else. Some ad blockers, proxies, or firewalls block URLs containing /ad/ or ending in common executable suffixes. This can cause some images or articles to not appear.
This is a question about the opening of threads at the Dispute Resolution Noticeboard. In particular, exactly where is the specification that sets the Do Not Archive Until date in a new thread that is opened by what appears to be WP:Dispute resolution noticeboard/request and WP:Dispute resolution noticeboard/Header? It appears that the Do Not Archive date is being set to two weeks after the case opening date. This means that cases that have been open for more than two weeks and have no activity for 48 hours are being archived. I am planning to change the archival parameter so as to wait 72 hours rather than 48 hours, but I would also like to set the Do Not Archive Until date to start out three weeks rather than two weeks after filing. Active cases often do not get resolved in two weeks, and do not always have activity in 48 hours. So can someone please tell me where the Do Not Archive Until date comes from? That is, where is the computation done? Robert McClenon (talk) 18:20, 6 April 2024 (UTC)[reply]
Thank you, User:Redrose64. That is interesting and puzzling. Then why was Do Not Archive set to 1 April 2024 in Wikipedia:Dispute_resolution_noticeboard/Archive_244#Climate_change, which appears to have been filed on 18 March 2024? The thread was archived because there was no activity for 48 hours and it was after 1 April, so the bot was honoring the parameters. I will look at the template in more detail in a while, but I am puzzled. I thought that there might be a technical answer. I am still sure that there is a technical answer, but it is even more complicated than I would have thought. Robert McClenon (talk) 19:52, 6 April 2024 (UTC)[reply]
This is the filing edit. It rules out one theory of mine, viz. that the DNAU timeetamp was modified after filing. --Redrose64 🌹 (talk) 22:11, 6 April 2024 (UTC)[reply]
So the DNAU timestamp was indeed "wrong" for some reason. Thank you for doing the research. Robert McClenon (talk) 04:38, 7 April 2024 (UTC)[reply]
If someone used the clickable button at the top of the DRN page to make a request, the report text is defined by the javascript behind it, which has the DNAU set to 14 days. You can see the wikitext it creates at MediaWiki:DRN-wizard.js#L-189. Aidan9382(talk) 07:49, 7 April 2024 (UTC)[reply]
Thank you, User:Aidan9382. So that is indeed setting the DNAU timestamp to 14 days after filing. Whoever thought that disputes would normally be resolved in two weeks was being very optimistic, maybe because they thought that there would be dozens of volunteers, one working each dispute. How do I request that the 14 be changed to 28? Robert McClenon (talk) 15:16, 8 April 2024 (UTC)[reply]
Attempt to edit the Javascript page, see the button that says "submit edit request", click that, then make the request. An interface admin will be along Soonly. Izno (talk) 16:44, 8 April 2024 (UTC)[reply]
Change requested as discussed, and made by interface admin. Thank you for the information. Robert McClenon (talk) 04:36, 10 April 2024 (UTC)[reply]
Can we make AfD notification smart enough to notify the editor who turned an existing redirect into an article, rather than notifying the editor who merely created the redirect?
As a redirect-happy editor, I get these all the time. I make a redirect somewhere because the subject is mentioned, but not independently notable. Some other editor comes along and turns the redirect into an article. A third editor nominated the article for deletion, and who get's the notification? The editor who turned the redirect into an article? No, it's me. How hard is it to make the obvious fix to this? BD2412T 23:16, 7 April 2024 (UTC)[reply]
@BD2412 can you be more specific about this "notification" you are referring to? echo doesn't have a trigger that fires when someone discusses deleting a page. — xaosfluxTalk 23:51, 7 April 2024 (UTC)[reply]
I believe that'd be part of WP:Twinkle. Not entirely sure who maintains the scripts/how to fix this though Soni (talk) 00:13, 8 April 2024 (UTC)[reply]
@BD2412 thanks. Mediawiki doesn't do this in core, and there are many many ways someone could be notified similar to that. In your specific case you seem to be referring to this edit, and that the editor made (possibly unknowingly) a bad edit - correct? That edit claims that it was made with the help of the PageTriage extension, so you may want to report a bug about it. — xaosfluxTalk 00:13, 8 April 2024 (UTC)[reply]
@Xaosflux: (do I need to ping you, by the way? Are you watching the discussion already?) The editor didn't make a mistake; the notice is sent automatically from the editor's account when the editor uses certain tools to nominate an article for deletion. BD2412T 00:29, 8 April 2024 (UTC)[reply]
(yes please use ping I don't use "subscribe" echo notifications) As I noted above, based on the edit tag this user appeared to use an extension to help them make this edit; if this isn't the desired behavior for that extension it can be reported to the extension maintainers using the link I provided above. (Had this been a different tool, such as Twinkle, there would be a different set of volunteers to look in to it). — xaosfluxTalk 00:48, 8 April 2024 (UTC)[reply]
@Xaosflux: Not 100% sure, but it looks like it. That has been unactioned for a long damned time, though. BD2412T 01:05, 8 April 2024 (UTC)[reply]
That happens a lot. And when it is something in software, it needs to be addressed by a limited number of developers for that software (there are currently over 200 open tasks for that extension alone). This is not something that we can fix here on the English Wikipedia. — xaosfluxTalk 01:28, 8 April 2024 (UTC)[reply]
I have long had a similar annoyance, which I have mostly ignored. As an AFC reviewer, I sometimes move a sandbox draft that has been submitted for review into draft space. This creates a redirect from the sandbox to the draft, and I appear to be the originator of the draft. Either six months later, or much later, I get a notice that "my" draft has been deleted as G13, an expired draft. I think that this is the same issue, in which the creation of a redirect confuses Twinkle. Is this the same issue as User:BD2412 is reporting? Robert McClenon (talk) 15:10, 8 April 2024 (UTC)[reply]
Yes, User:Xaosflux, that is what I was describing. If a page that I moved from a sandbox to draft space is ignored for six months, I then get a G13 notice that it was deleted. Yes, I don't consider myself to have been the draft creator, but I do get the notice. No, I do not create the draft by copy-paste. I dislike copy-pastes as much as the admins who have to do history-merge to correct for them. Yes, there is a persistent minor problem with who Twinkle thinks is the author. Robert McClenon (talk) 16:15, 9 April 2024 (UTC)[reply]
Yes, I see that this was first reported five years ago. There is a persistent minor problem that has been around for so long that the bug has the status of a naturalized citizen. Robert McClenon (talk) 16:20, 9 April 2024 (UTC)[reply]
@Robert McClenon thank you, from your notes above this specific error is only coming from Twinkle, correct? Twinkle issues can generally be addressed on-wiki, it just takes someone to write patch for the script. — xaosfluxTalk 17:08, 9 April 2024 (UTC)[reply]
I don't believe that any notification is technically required, the obvious fix would be to get rid of notifications... But I don't think thats what you mean. Horse Eye's Back (talk) 15:59, 9 April 2024 (UTC)[reply]
Actually, if I am the one who turns a redirect into an article, I would want to be notified if that article is nominated for deletion. BD2412T 16:23, 9 April 2024 (UTC)[reply]
The watchlist seem to fill that role for most, personally I almost always hit the little star when making a redirect or unredirecting. Horse Eye's Back (talk) 17:00, 9 April 2024 (UTC)[reply]
Some editors don't use watchlists, and would still like to be notified that their work is being tagged for deletion. Templates on user talk pages are the "gold standard" for notices to editors, and so are required for noticeboards, where pinging is not a substitute. Some editors may not care about these notifications, but other editors either want them or should receive them, so that editors who do not care for them can ignore them. Editors who receive misdirected notifications, as are being discussed here, sometimes ridicule them. Robert McClenon (talk) 04:34, 10 April 2024 (UTC)[reply]
Thank you for mansplaining Templates to me... Including the completely irrelevant information about templates related to noticeboards when we're discussing AfD notification. Next time resist the urge. Horse Eye's Back (talk) 16:17, 12 April 2024 (UTC)[reply]
I'll preface by saying this is minor, since you can just search again without going back.
I am using Google Chrome, desktop, and today I noticed some odd behaviour with search:
I went to Special:Search (technically an archive search, but it is happening with just the base one too);
I typed a search term and successfully searched it;
I clicked back on my browser (it took me to the search page, with the search term I typed in the box - this is expected);
I then cleared the box and typed a new search term and clicked search again;
The term that was searched was the one I had searched the first time around, not the new one I typed in.
So, in short, if I search for something from Special:Search and then navigate back to before that search happened and then try to search for something else, it will always just repeat the first search. – 2804:F14:8090:C501:686B:E2D5:C69E:9592 (talk) 02:07, 8 April 2024 (UTC)[reply]
I see the same with with Edge, which is based on Chromium, so it has some commonality with Chrome. I'd guess its browser caching or some cookie wierdness. RudolfRed (talk) 03:21, 8 April 2024 (UTC)[reply]
I can't reproduce this for some reason (the input is empty for me after I press "Back"), but this is very similar to phab:T354107, just with the "Back" button. Jack who built the house (talk) 03:06, 9 April 2024 (UTC)[reply]
And it searches fine when you try?
By the way I looked at the inspect element, and it does look like the image shared([1]) in that bug report - in fact, if I go back and try searching and repeat that a few times, there will be a new element for each of my new searches (though it always searches the first one).
When I press "Back", the search input is clear for me. So the state of this page is not cached. This is probably why it works for me to search again. Jack who built the house (talk) 04:55, 10 April 2024 (UTC)[reply]
It does the same thing for me - Win11, Firefox 124.0.1 Black Kite (talk) 04:01, 9 April 2024 (UTC)[reply]
I think I understand what is causing this bug:
When clicking search it calls the setSearchSubmitTrigger function in modules/ext.advancedSearch.init.js;
$searchField is the actual searchbox element you type in, in line 60 that text is processed and saved to a variable, and in the next lines that value is put in a new, hidden, <input> element. This new element is created with a name attribute copied from the searchbox element (this hidden element is the one with name="search" in the image in that bug report);
In line 65 the value of the searchbox's name attribute ("search") is then cleared, the search then proceeds using the hidden element;
The bug then, is that when you navigate back on Chrome (and I guess Firefox), the page doesn't get (down)loaded again (browser cache?). The searchbox and this hidden element are still in the page and in this post-search state.
When you search again, it creates a new hidden element with a name attribute set to an empty string (because it's copying from the searchbox one and that one was cleared). When the search proceeds it then reuses the hidden element that has the name attribute set to "search", which is the one created in the original search.
If you stop the browser after this function is done (the bug report), these same conditions will be there when you try searching again. – 2804:F14:8090:C501:E8D8:2D74:117E:F93D (talk) 05:51, 9 April 2024 (UTC)[reply]
I have just created this draft, but the image is highly oversized. I have tried both default image size and then 250px manual. What's the problem? Kind regards 14 novembre (talk) 🇮🇹 18:45, 9 April 2024 (UTC)[reply]
Answered at WP:THQ, please don't ask same question in multiple venues. RudolfRed (talk) 19:28, 9 April 2024 (UTC)[reply]
Don't know if this is quite the correct forum to be posting this: When one views 'edit history' on any given article, selects the different versions, and then clicks on Compare select revisions, the pale yellow used to highlight deleted mark up is barely visible. Esp when only one character in a sentence or paragraph has been deleted, which is virtually impossible to discern sometimes. Is there any way to change the (very) pale yellow used to highlight deleted characters to a more visible tone or color, perhaps orange or gold? — Thanx. -- Gwillhickers (talk) 17:34, 9 April 2024 (UTC)[reply]
If you go to ?action=info for any page, you will see a table with various statistics, including two lines about how many people are watching the page, e.g.:
I am finding that the ratio shown here is not at all unusual for older articles, but the first line gets more attention from editors. They think "hundreds of editors are watching this page", when they should be thinking "almost nobody is watching this page". Is there a way we could remove/hide the irrelevant number from this info page? Or should we just change MediaWiki:Pageinfo-watchers to something like "Total number of watchlists (includes inactive editors)"? WhatamIdoing (talk) 04:53, 10 April 2024 (UTC)[reply]
I’m one of those who ‘think "hundreds of editors are watching this page", when they should be thinking "almost nobody is watching this page"’ ... I think we should have that “second line” added to these pages (or replaced the “first line”):
I bet someone here could write a .css script that would blank that out, or rename it to something like "This is the wrong line – ignore it". I checked a bunch of Special:Random pages, and most of them showed no data, due to there being too few people watching the pages. WhatamIdoing (talk) 21:11, 11 April 2024 (UTC)[reply]
For hiding that row on the page information page on Wikipedia, try #mw-pageinfo-watchers { display: none; } in one of the .css files. –Novem Linguae (talk) 22:43, 11 April 2024 (UTC)[reply]
Thanks, @Novem Linguae. That seems to have worked for me. That suggests that iff we ever decided that we wanted to do that globally, it could be done in (e.g.,) global.css. I'm going to try this out for a while. I suspect that I'll like it. WhatamIdoing (talk) 22:32, 12 April 2024 (UTC)[reply]
If it's on <30 watchlists, no number is given for either item (the second item is simply suppressed).
If it's on ≥30 watchlists, an exact number is given for both items.
If the second number is zero, it says "There may or may not be a watching user visiting recent edits" instead of a second number. WhatamIdoing (talk) 23:53, 10 April 2024 (UTC)[reply]
Is there a way we could remove/hide the irrelevant number from this info page? I would not support removing this. Presenting both numbers, and letting the user decide which they want or need, seems like an acceptable status quo here. The less than 30 thing for non admins is for security reasons. Admins can see both numbers at all times. The linked phab ticket mentions changing the second message to mention 30 days explicitly. I could get behind a change like that. –Novem Linguae (talk) 02:05, 11 April 2024 (UTC)[reply]
Under what circumstances do you think it would it be useful for to you to know that the watchlists associated with 1,991 mostly inactive (and sometimes actually dead) include the defunct Wikipedia:WikiProject Contents?
The number of active editors presently watching that page is a single-digit number. I can understand why that number would be useful to know, but not why the first has practical value. WhatamIdoing (talk) 02:56, 11 April 2024 (UTC)[reply]
As far as I can tell, pages can be watched actively, without the user being considered active there - such as through email or syndication. — xaosfluxTalk 08:45, 11 April 2024 (UTC)[reply]
Or by looking at the watchlist but not visiting the links. Nardog (talk) 09:43, 11 April 2024 (UTC)[reply]
I suspect that the scenario Nardog describes is far more common than the one Xaosflux describes. WhatamIdoing (talk) 18:32, 11 April 2024 (UTC)[reply]
That was an argument against hiding the total number of watchers. I was echoing Xaosflux. Nardog (talk) 05:58, 13 April 2024 (UTC)[reply]
I've never spent much time reading about the internal mechanism, but I believe that it works like this:
If you have some pages on your watchlist, and then get permanently locked out of your account, you are "a watcher" for all of those pages forever, even though you can't actually watch anything from that account.
If you have some pages on your watchlist, go to Special:Watchlist, close the tab without clicking on any link or visiting an article at all, and then get permanently locked out of your account, you are counted as "an active watcher" for all the pages on your entire watchlist for the next 30 days (including pages that did not have any changes made, so they weren't listed at Special:Watchlist on the day that you visited that page).
This means that there are 9 active editors with this page on their watchlist, of which an unknown number – but it is quite possibly zero – actually looked at the page in question during the last month.
So Xaosflux says, yes, there may only 9 editors who have that page on their watchlists and actually went to Special:Watchlist at any point during the last 30 days, but maybe a few more people also get e-mail messages about changes to articles on their watchlist, so the "9" might be a slight undercount.
The scenario you describe – visiting the watchlist but not checking every page – is certainly common. The "9" is probably an overcount, if the goal is to know whether anyone actually checked the specific page. WhatamIdoing (talk) 02:54, 15 April 2024 (UTC)[reply]
WhatamIdoing, I don't believe merely viewing the watchlist makes you an active watcher. — Qwerfjkltalk 07:45, 15 April 2024 (UTC)[reply]
You're correct. The count uses the same "first unseen timestamp" data used for highlighting of unseen edits on the history page. If a revision older than the configured age would be highlighted, the user is not counted as having visited recent edits. Viewing the watchlist or history page doesn't reset that, while visiting the page itself will and viewing old revisions or diffs may. Anomie⚔ 11:41, 15 April 2024 (UTC)[reply]
I am having trouble putting an anchor at a specific row in List of cover versions of Led Zeppelin songs. In particular, I want an anchor on the song "That's the Way". I tried
| rowspan="4" | {{Anchor|"That's the Way"}} "[[That's the Way (Led Zeppelin song)|That's the Way]]"
but loading the page with the anchor in the URL didn't seem to take me there. I looked at the documentation at Help:Tables and locations § Section link or map link to a row anchor which told me to try
|- id="That's the Way"
but I'm getting the same issue (also, I'm not certain how to encode quotation marks if I go the id on the row route). I'm sure I'm missing something obvious here. Kimen8 (talk) 16:31, 10 April 2024 (UTC)[reply]
This is what I figured. Something with my browser, even caching, who knows. I tried incognito and it didn't work. Good to know the current version works.
Do you know how to escape quotes using the row id anchor? Or should I just use the anchor template in that situation?
You were doing fine putting an anchor. The misunderstanding was in referring to the anchor. Uwappa (talk) 04:49, 11 April 2024 (UTC)[reply]
Table of Contents can't display some statistics symbols
(I'm posting it here since this place gets more traffic than Wikiversity). We have a statistical page in Wikiversity which includes statistical symbols in the headings. While there's no issue with those symbols shown in the body of the page, the Table of Contents is bugged whenever these symbols are used. Does anyone have the solution to this problem? OhanaUnitedTalk page 21:16, 10 April 2024 (UTC)[reply]
This is likely to be T295091, an unresolved bug from November 2021. – Jonesey95 (talk) 21:24, 10 April 2024 (UTC)[reply]
@OhanaUnited: The first problem heading is ==== Haphazard weights with estimated ratio-mean (<math>\hat{\bar{Y}}</math>) - Kish's design effect ==== which contains <math>...</math> markup. The problem that you observe is why our MOS:HEADINGS says For technical reasons, section headings should ... Not contain <math> markup. --Redrose64 🌹 (talk) 22:33, 10 April 2024 (UTC)[reply]
Why doesn't this code work?
I have a script at User:TheTechie/tut.js. The function getText is supposed to return something, but it returns undefined. Even alerting the returned wikitext returns something, but returning it or assigning it to a variable and using it in arcTostill returns undefined.
Browser: Chrome 122.0.6261.137 (either stable or LTS)
Can someone help me here? Thanks --- thetechie@wikimedia: ~/talk/$ 01:34, 11 April 2024 (UTC)[reply]
First off, are you calling getPage from somewhere? I don't see a call to it in the code currently.
Second, have you tried step debugging it in Chrome Dev tools? Press f12, ctrl shift f to search for getpage, place some breakpoints inside by clicking on the line number, then run the code. Hover over variables and hit f10 to see what the code is doing. –Novem Linguae (talk) 02:14, 11 April 2024 (UTC)[reply]
@Novem Linguae Sorry, I meant that getText stores value in a variable called wikitext, which when used in arcTo, is equal to undefined. When the wikitext got in getText is alerted, it returns something, but when the wikitext is assigned to a variable and used in arcTo it is undefined. As for the dev tools thing, I'm currently kinda busy and can't use devtools right now, I will when I get the chance. Sorry for the confusion. thetechie@wikimedia: ~/talk/$ 15:36, 11 April 2024 (UTC)[reply]
I installed your script just now and took a closer look. The getText codepath never runs because nothing calls it. Are you saying you need getText to run so you can set the wikitext variable to something other than undefined? If so you need to put getText( title ); somewhere. –Novem Linguae (talk) 16:07, 11 April 2024 (UTC)[reply]
@Novem Linguae I fixed the call issues, but it still returns undefined. (Hint: To run the offending function, click the TUT in the menu and select "Arc" and enter a page name). Thanks --- thetechie@wikimedia: ~/talk/$ 17:13, 11 April 2024 (UTC)[reply]
Unrelated tip: it's hard for other developers to read your code if you abbreviate. I'd suggest expanding TUT, arc, arch, arcTo, sm, etc. to use their full names. –Novem Linguae (talk) 22:52, 11 April 2024 (UTC)[reply]
getText doesn't have a return statement, so it won't return anything. If you want to make it return Promise<string>, it needs return statements on line 16 and line 31. See promise chaining. – SD0001 (talk) 16:20, 11 April 2024 (UTC)[reply]
@SD0001 and Novem Linguae: update: if I run the function the second time, it gets the variable and it isn't undefined. I'll be looking into this, just wanted to update you both though. thetechie@wikimedia: ~/talk/$ 17:25, 11 April 2024 (UTC)[reply]
Also I appear to have done something on accident which causes the portlet to not appear, even though the function to add a portlet is called. Any help is appreciated. thetechie@wikimedia: ~/talk/$ 21:54, 11 April 2024 (UTC)[reply]
Thanks, but although I see it I don't see a way to get it to change what I see. Doug Wellertalk 10:06, 11 April 2024 (UTC)[reply]
Click the "Time zone" field and select Europe/London. PrimeHunter (talk) 11:39, 11 April 2024 (UTC)[reply]
Which "clock" are you referring to?
Uwappa and PrimeHunter above are talking about the display of times in the watchlist, page history, and other special pages.
If you're talking about the UTCLiveClock gadget ("(S) Add a clock to the personal toolbar that displays the current time in UTC and provides a link to purge the current page (documentation)"), docs at the top of mw:MediaWiki:Gadget-UTCLiveClock.js specify what to add to your common.js or skin.js to change the timezone.
If you're talking about the timestamps shown on comments in discussion pages like this one, there's a CommentsInLocalTime gadget ("(U) Change UTC-based times and dates, such as those used in signatures, to be relative to local time (documentation)") to adjust the display, but I believe you'll have to live with UTC in the editor like the rest of us around the world do.
Is there a way to show all edits by edit summary? Via a web interface. For example, show all edits that contain the string "WP:URLREQ#herbaria4.herb.berkeley.edu" such as Special:Diff/1165643139/1218392247. -- GreenC 13:26, 11 April 2024 (UTC)[reply]
GreenC, you mean other than a SQL query? If it's just for one user you can use this tool. — Qwerfjkltalk 15:04, 11 April 2024 (UTC)[reply]
@GreenC: That tool is in the box at the bottom of your contribs page, fifth from the left as "Edit summary search". --Redrose64 🌹 (talk) 17:00, 11 April 2024 (UTC)[reply]
I should have known about this a long time ago. -- GreenC 17:41, 11 April 2024 (UTC)[reply]
Contributions link dropdown
Resolved
Is there a .css or .js way to disable the dropdown ? - FlightTime (open channel) 14:09, 11 April 2024 (UTC)[reply]
Hi there, I started having this issue today that's making it nearly impossible to edit.
When I click "review your changes" after editing a page, "template parameters changed" appears for EVERY template on the page, even if I have not changed any of them. Here is a screenshot I took (uploaded to Imgur). I tried disabling all my user scripts, and I am still having the issue.
I am using Firefox version 124.0.2, on desktop, and my browser is up to date.
I would really really appreciate any help with this. Thank you so much for any help. HeyElliott (talk) 19:13, 11 April 2024 (UTC)[reply]
Also, it's not an issue with any of my extensions, as I'm still having the issue without extensions. HeyElliott (talk) 19:21, 11 April 2024 (UTC)[reply]
Same thing happens to me. At first I thought I messed up somehow. (I'm using Chrome, so it's not the browser). Nikolaj1905 (talk) 11:07, 12 April 2024 (UTC)[reply]
Hey all, I've left the results of the investigation in the phab task, but this is a one-off disruption because of changes in contents of data-mw attributes of template content wrappers in Parsoid HTML. When cached HTML (generated by older production code) is compared with new HTML (generated by new production code that went out this week), you will see these noisy diffs reported by the visual diff code which compares HTML (not wikitext). But, you shouldn't run into this in subsequent edits. Sorry about the disruption. SSastry (WMF) (talk) 13:43, 12 April 2024 (UTC)[reply]
@SSastry (WMF): Thanks for the explanation. I thought page diffs are based on Wikitext. Can you explain why in this case it is comparing the output HTML instead of the Wikitext? RudolfRed (talk) 19:54, 12 April 2024 (UTC)[reply]
There are two kinds of diffs. One is the wikitext diffs - the kind that is most commonly used. Visual diffs need to compare the rendered output and hence it compares HTML. SSastry (WMF) (talk) 09:26, 15 April 2024 (UTC)[reply]
Thanks so much for your quick fix on this and the reply, I really really appreciate the work y'all do on this site! I've still been having this issue a few times today but I assume that's just because of the cached stuff, and it'll fix itself? (Sorry, I don't know much about caches and HTML and such.)
Thank you again, and I hope you have a great day! HeyElliott (talk) 19:57, 12 April 2024 (UTC)[reply]
Sorry to reply again, but I'm still getting the error often. Do I need to clear my cache or something? Again, sorry, I'm not super knowledgeable on this stuff. Thank you! HeyElliott (talk) 03:20, 13 April 2024 (UTC)[reply]
You don't need to do anything on your end. If the page is edited across the deployment date, that first edit will have this problem in visual diffs. You could forcibly clear the cache of the page via the purge query parameter if you wish and then look at the visual diff again - that should fix the problem in most cases. SSastry (WMF) (talk) 09:29, 15 April 2024 (UTC)[reply]
Missing Article for improvement
This week's Article for improvement seems to have gone missing. There was one last week, and there's one scheduled for next week, but there's none for this week:
citoid (VisualEditor automatic citation generator service) is not working for some refs that used to work. including NY Times. see phab:T362379. Jeremyb (talk) 18:02, 12 April 2024 (UTC)[reply]
Using the search function at the top of the page for a case with a redirect to an anchor brings the reader to the top of the redirected article, not to the anchor, which is not convenient for the reader. Is this behaviour intended? This seems to be a significant down-side of redirect to anchors for me.
The following example is for illustration (the question is not about this specific example): There is a redirect for Facial expression recognition. The search function will show Affective computing as a result. Clicking on it leads to the beginning of this article, which itself does not contain "Facial expression recognition", because the anchor leads to "Facial affect detection". Kallichore (talk) 19:08, 12 April 2024 (UTC)[reply]
I opened phab:T362442 on this. Feel free to voice support there, good chance some search dev is going to say this is a feature not a bug. — xaosfluxTalk 20:55, 12 April 2024 (UTC)[reply]
Issue with Search Feature on Wikimedia
I conducted a search for "Solar eclipse of 2024 April 8" on Wikimedia.
Expecting a concise and accurate description of the page, I was disappointed by the search results.
Upon clicking, I found metadata in Hungarian, which is unusual, as the page is new and not previously associated with Hungarian. Additionally, no image is displayed.
Metadata: Solar eclipse of 2024 April 8
<nowiki>eclipse solar del 8 de abril de 2024; 2024. április 8-i napfogyatkozás; 2024ko apirilaren 8ko eguzki eklipsea; eclipse solar del 8 d'abril de 2024; Sonnenfinsternis...
Could the search feature generate metadata that accurately describes the page's content and include an image from the page? AceSeeker (talk) 23:53, 12 April 2024 (UTC)[reply]
When you search for that string on the English Wikipedia, you get reasonable results, as far as I can see. It looks like you did this search on Wikimedia Commons, which is different from the English Wikipedia, and which editors here are not responsible for. It appears that the Commons page that is the first result uses a template called {Wikidata Infobox}. The page description that you are seeing appears to be the names of the corresponding page as listed on Wikidata at d:Q2620078. I don't know why that template on Commons pulls in that data, but you could ask about it at commons:Template talk:Wikidata Infobox. – Jonesey95 (talk) 00:07, 13 April 2024 (UTC)[reply]
Thank you for prompt response and clarifying that the issue may be related to the template {Wikidata Infobox} on Wikimedia Commons. I'll follow up on the commons:Template talk:Wikidata Infobox page to address this matter. Appreciate your assistance! AceSeeker (talk) 00:25, 13 April 2024 (UTC)[reply]
I opened a phabricator issue about this but then I realised I should have discussed it here first.
So, when using the edit button for the lead section on mobile (I'm on mobile web, I have no idea if the app is different), no section link is included in the edit summary. On the other hand, on desktop when the "Add an [edit] link for the lead section of a page" gadget is used, a section link is correctly added. In my opinion, mobile should be updated to match desktop's behavior. Nickps (talk) 14:09, 13 April 2024 (UTC)[reply]
Not sure if this is the right place, newbie editor.
So I was trying to reset the password for my old account, Lovecodeabc, so I requested password resets. However the resetting of passwords did not work, and it seems that if I send a password reset on Special:PasswordReset for my email, it works, but not for my username. Additionally, the temporary password emailed for the Lovecodeabc account does not work. I just want to have the Lovecodeabc username back. Any suggestions? Lovecodeabc(tm) (talk) 14:12, 13 April 2024 (UTC)[reply]
The only way to reset a password is by email, and only if you had previously registered and confirmed your email to an account before you forgot the password. Special:PasswordReset always requires both a username and a password, and will only send a reset if both of them match. You can also only do one reset per day in most cases. The first username you mentioned does not appear to have an email registered, unless you specifically reconfigured it to be for reset only and not for wikimail. — xaosfluxTalk 14:54, 13 April 2024 (UTC)[reply]
Special:PasswordReset only requires one of username and email address. That's why it says "Fill in one of the fields to receive a temporary password via email". User:Lovecodeabc has not specified a valid email address.[3] The message is different if you have a valid address but have chosen to disallow mails from others. @Lovecodeabc(tm): I guess you entered the email address for Lovecodeabc(tm) and the mail says so, not Lovecodeabc. PrimeHunter (talk) 16:30, 13 April 2024 (UTC)[reply]
Oops, forgot that Send password reset emails only when both email address and username are provided. is off by default! So yes, it will send you reset links for ALL the accounts you have under an email address - but again if the email address was never confirmed to the account it won't. — xaosfluxTalk 16:54, 13 April 2024 (UTC)[reply]
Well, I got a password reset email for Lovecodeabc: The emailLovecodeabc(tm) (talk) 20:26, 13 April 2024 (UTC)[reply]
@Lovecodeabc(tm): I did a test. "This user has not specified a valid email address." at https://en.wikipedia.org/wiki/Special:EmailUser?wpTarget=Lovecodeabc can both mean the account never saved an email address and never confirmed it (see Help:Email confirmation). A temporary password can be sent to an unconfirmed email address. I don't know why the password didn't work for you. It's also odd that your screenshot says "lovecodeabc" with lowercase "l" (unless it's an uppercase "L" in a weird font). The first character of usernames is automatically capitalized. If you wrote it as "lovecodeabc" when the account was created then I don't know whether the lowercase "l" is stored somewhere and can still be retrieved in some situations. Anyway, the account only has 19 edits and can just be abandoned. Special:Contributions/Lovecodeabc only shows unimportant userspace edits and Special:CentralAuth/Lovecodeabc shows no edits at other wikis. If you really want the username then you could try Wikipedia:Changing username/Usurpations. It's usually only for accounts with no edits but rare exceptions are made. First check several times in different browsers that a temporary password doesn't work. PrimeHunter (talk) 23:41, 13 April 2024 (UTC)[reply]
@Suffusion of Yellow: My account is Lovecodeabc, not Iovecodeabc. Sorry for the confusion!
@PrimeHunter: I'll look into Wikipedia:Changing username/Usurpations. I've checked Mozilla Firefox on Ubuntu 22.04 and Safari on iPadOS. Are there any issues with those two browsers? I haven't used Google Chrome to try it yet. (also, isn't Google Chrome based on Chromium/WebKit, same as Safari)?
@Lovecodeabc(tm): Is it possible you also created the account User:Iovecodeabc with your email address and wrote Iovecodeabc at Special:PasswordReset but tried to log in as Lovecodeabc with the temporary password in the mail? That would certainly explain why it didn't work. PrimeHunter (talk) 01:33, 14 April 2024 (UTC)[reply]
Oh yeah, it never occured to me to check the user page of Iovecodeabc! Lovecodeabc(tm) (talk) 01:39, 14 April 2024 (UTC)[reply]
@Lovecodeabc(tm): Everything technical appears resolved now. Lovecodeabc hasn't stored an email address (or has an unknown and unconfirmed address) so it doesn't work to write Lovecodeabc at Special:PasswordReset. Iovecodeabc has stored an email address so it works to either write Iovecodeabc or the email address, but it only sends a password for Iovecodeabc. https://en.wikipedia.org/wiki/Special:EmailUser?wpTarget=Iovecodeabc indicates the email address is not currently confirmed but that can be done as described at Help:Email confirmation. Zzyzx11 wrote there [4] that you need a confirmed email address to reset your password but an unconfirmed stored address is apparently enough for that. Confirmation is required for other email features. PrimeHunter (talk) 02:27, 14 April 2024 (UTC)[reply]
Changing "input element" to "anchor element" and then making anchor elements unselectable
Hi, after the title of articles, and after final rendering, an expression appears as
"translate links from en to fa". Bug scenario is that when we want to select the title, if we suddenly extend the selected region, for example for Wikipedia article, the texts of "Toggle the table of contents" and "en" and "fa" would be in our clipboard. So after pasting the clipboard, the total clipboard text would be:
Toggle the table of contents
Wikipedia
en
fa
To solve this problem, we should apply these styles:
.vector-page-titlebar-toc.translator-equ-wrapper{-webkit-user-select:none;/* Safari */-ms-user-select:none;/* IE 10 and IE 11 */user-select:none;/* Standard syntax */}
But the element type for "en" and "fa" is "input" and we can not make "input elements" unselectable, so please convert "input element" to "anchor element" for "en" and "fa" and then apply the above styles. Thanks, Hooman Mallahzadeh (talk) 16:46, 13 April 2024 (UTC)[reply]
I guess this is related to a userscript, not part of the standard user interface. Without that script installed I don't see quite what you're talking about. I suggest asking at fa:MediaWiki talk:Tofawiki.js or consulting Ebrahim. --Jeremyb (talk) 02:46, 14 April 2024 (UTC)[reply]
Hooman Mallahzadeh: This was proposed here User talk:Ebrahim/ArticleTranslator.js#user-select but was making breakage in some other browser so I applied another solution which supposed to not have the issue which apparently isn't imperfect in your case though I couldn't reproduce your issue. Guess the best way forward is to add your username to a blacklist for the tool so you can develop your own version and after that if your version covers all the use cases on all the different browsers we can merge the versions. Does that sound good? Just please keep your changes to minimum possible for the ease merging back the versions. Thanks −ebrahimtalk 06:27, 14 April 2024 (UTC)[reply]
Ok, I can reproduce this, it only happens on double clicks (and not in triple clicks?), and only happens in Chrome and not in Safari and Firefox. Changing those input elements to editable anchor elements isn't that easy, in fact it initially was like that before this edit and some users weren't able to edit them when needed and having that with user-select: none makes users not able to select the text inside to modify the actual language code. −ebrahimtalk 06:56, 14 April 2024 (UTC)[reply]
@Ebrahim No the problem is not resolved! To correct that, at this page and at the lines 340 and 341, you should convert input element to anchor element like this:
@Jeremyb-phone The problem for the "Toggle the table of contents" persists for all users, this code makes that unselectable.
.vector-page-titlebar-toc{-webkit-user-select:none;/* Safari */-ms-user-select:none;/* IE 10 and IE 11 */user-select:none;/* Standard syntax */}
Would you please do something to apply this style? Thanks, Hooman Mallahzadeh (talk) 08:47, 14 April 2024 (UTC)[reply]
Regarding the user script (not toggle part), unfortunately that solution will disturb the tool on other browsers and things like this can happen on double/triple clicks on Safari (at least it once it used to be). I'm testing with all three browsers (Chrome, Firefox and Safari) but suggestions that have given to me were fixing one browser issue making the other major browsers worse so maybe I can add you to the blocklist of the tool so you copy the script and use your version, then if your solution could work on other browsers on my testing I can apply it to the main script also. −ebrahimtalk 09:23, 14 April 2024 (UTC)[reply]
@Ebrahim Such problems (inconsistencies between browsers) arise from bad design of codes. Maybe you should refactor this Javascript code. Implementing codes layer by layer such that one layer be completely independent of other layers would probably solve this bug and other similar bugs. Hooman Mallahzadeh (talk) 09:49, 14 April 2024 (UTC)[reply]
Hello guys, it's been a few hours and I can't edit any page on Wikipedia not even my userspace. Prior to this, for about 30 mins I had an issue checking the page history revisions of a particular page. While checking the page history, it seemed to glitch with an error saying "Unable to stash Parsoid HTML." I couldn't compare the revisions of other editors. Normally I'd ignore such issues thinking something related to "Wiki Sever" is going on, but now it's getting me worried. Rejoy2003(talk) 14:11, 14 April 2024 (UTC)[reply]
Actually, now it's broken for me on iPhone. GoutComplex (talk) 14:36, 14 April 2024 (UTC)[reply]
I see page history perfectly, but it's just editing that won't work for me on PC. It works fine on phones. GoutComplex (talk) 14:33, 14 April 2024 (UTC)[reply]
Well I'm on phone, I can't edit nor check history properly apparently due to "glitching." Rejoy2003(talk) 14:36, 14 April 2024 (UTC)[reply]
I'm also getting ""Unable to stash Parsoid HTML" on Chrome on Windows 11. Same thing on Firefox on Windows 11. Looks like Microsoft Edge on Windows 11 works properly.
On Android 14, I get a different error, "Error, can't load the editor". ReferenceMan (talk) 14:37, 14 April 2024 (UTC)[reply]
After 1 successful edit, Microsoft Edge now gives the same error, "Unable to stash Parsoid HTML."
Same issue, Firefox on Windows 10. Mrchikkin (talk) 14:42, 14 April 2024 (UTC)[reply]
Okay I can edit now, but still there's still minor glitch or bug (along with the Parsoid HTML pop-up) while checking some history of some pages. I tried to login into my account through Opera browser, the good thing is I can edit back. Probably something's up with Google chrome or Firefox. Rejoy2003(talk) 14:45, 14 April 2024 (UTC)[reply]
Same here, the visual editing is not opeing on my phone, and opening sometimes on Desktop. Grabup (talk) 15:17, 14 April 2024 (UTC)[reply]
I'm also getting the same error message saying, "Unable to stash Parsoid HTML". But what's weird is I tried using the Legacy renderer instead of the Parsoid; But it still says: "Unable to stash Parsoid HTML". I don't understand how come a problem with the new Parsoid renderer, could be effecting the Legacy renderer also. 𝓥𝓮𝓼𝓽𝓻𝓲𝓪𝓷24𝓑𝓲𝓸 (ᴛᴀʟᴋ) 15:20, 14 April 2024 (UTC)[reply]
Whatever the problem was, it seems like its fixed now. Both visual and wikitext editors are now working in both renderers. 𝓥𝓮𝓼𝓽𝓻𝓲𝓪𝓷24𝓑𝓲𝓸 (ᴛᴀʟᴋ) 15:24, 14 April 2024 (UTC)[reply]
Still not working on mobile. Grabup (talk) 15:26, 14 April 2024 (UTC)[reply]
Same here not working for me too (Android 13 user). Rejoy2003(talk) 15:35, 14 April 2024 (UTC)[reply]
I think (though I'm not totally sure) that if you have an account/membership you can still access the pages. — Qwerfjkltalk 07:44, 15 April 2024 (UTC)[reply]
I believe that File:2024 Total Solar Eclipse (NHQ202404080102).jpg taken by NASA photographer is superior to the image now in the infobox of the article. In particular, it shows the Solar prominences much more clearly. However, I cannot figure out how to replace the image in this particular infobox. By the way, the photo was taken in Dallas, not in Indianapolis where the current image was taken. Any assistance would be appreciated. Thanks. Cullen328 (talk) 00:11, 15 April 2024 (UTC)[reply]
Zaathras and Suffusion of Yellow, thank you very much. I have been editing for 15 years and have never edited a module, and would have no idea how to do so. Cullen328 (talk) 00:38, 15 April 2024 (UTC)[reply]
@Cullen328 I'm oppose this change because the image you're proposing has noticeable noise around the corona. In addition, there is already an image of the solar eclipse in Dallas, Texas in the gallery. √2(talk) 06:03, 15 April 2024 (UTC)[reply]
sounds like the technical question is resolved and now it's an editorial issue. take it to the article's talk page? Jeremyb (talk) 06:51, 15 April 2024 (UTC)[reply]
Tables next to each other
Is there a way to put two tables next to each other instead of having one underneath the other? Flaming Hot Mess of Confusion (talk) 00:59, 15 April 2024 (UTC)[reply]
But remember that tables next to eachother on mobile are problematic most of the time. —TheDJ (talk • contribs) 12:47, 15 April 2024 (UTC)[reply]
pagelinks normalization
Hi, you might be aware that pagelinks table is being normalized (phab:T299947). I want to give enough time before dropping the columns. So here is the notice that the data has been fully populated in all Wikis except Turkish Wikipedia and Chinese Wikipedia (and these two wikis will finish in roughly a week). This means if you have tools or database reports in English Wikipedia that relies on pl_namespace or pl_title, you need to switch them to use pl_target_id ASAP. The rough estimate is that the old columns will be gone in a couple of weeks.
I was asked to make an explicit note here (one and two). Let me know if you have any question in the ticket. Thank you and my apologies for inconvenience. You can read more why we need to do this in phab:T222224. Best ASarabadani (WMF) (talk) 12:18, 15 April 2024 (UTC)[reply]
Thank you for the notification. Like the earlier templatelinks change, this will break to many on- and off-wiki tools (some of which may be unmaintained), and it's difficult to assess the likely scope of the damage. Here is a list of affected {{database report}}s, but there are plenty of other ways to use the table such as Quarry. Pinging R'n'B, who will probably be one of many with work to do. Certes (talk) 12:30, 15 April 2024 (UTC)[reply]
Moved from Wikipedia talk:Village pump
I think we should have that “second line” added to these pages (or replaced the “first line”):
You are invited to join the discussion at Module talk:Wd § Getting Template:Update tracker working with qualifiers. Sdkb talk 14:03, 10 April 2024 (UTC)
Resolved
![[1]](https://phab.wmfusercontent.org/file/data/lmqmphebmpv72akm3yyw/PHID-FILE-i2qspdrcf3kvw67yerug/image.png){kind=link}
.jpg){kind=link}