Wikipedia:Village pump (technical): Difference between revisions

From Wikipedia, the free encyclopedia
Content deleted Content added
→‎JS equivalent of FULLPAGENAME: replying to SD0001: Thanks! (Bawl!)
Line 737: Line 737:
:mw.Title.newFromText(':uSuÁrIo : example').toText() – [[User:SD0001|<span style="font-weight: bold; color: #C30">SD0001</span>]] ([[User talk:SD0001|talk]]) 15:54, 7 May 2022 (UTC)
:mw.Title.newFromText(':uSuÁrIo : example').toText() – [[User:SD0001|<span style="font-weight: bold; color: #C30">SD0001</span>]] ([[User talk:SD0001|talk]]) 15:54, 7 May 2022 (UTC)
::Thanks! [[User:NguoiDungKhongDinhDanh|<span style="font-family:Monotype Corsiva;background-image:linear-gradient(90deg,red,yellow,cyan);color:transparent;background-clip:text;-webkit-background-clip:text">'''NguoiDungKhongDinhDanh'''</span>]] 16:19, 7 May 2022 (UTC)
::Thanks! [[User:NguoiDungKhongDinhDanh|<span style="font-family:Monotype Corsiva;background-image:linear-gradient(90deg,red,yellow,cyan);color:transparent;background-clip:text;-webkit-background-clip:text">'''NguoiDungKhongDinhDanh'''</span>]] 16:19, 7 May 2022 (UTC)

== Importing Wikipedia dump on Windows ==

Hello, I would like to conduct some research on Wikipedia, but appeared that there is no documentation on how to import it into MySQL, at least for Windows case. I'm stuck with converting XML-file into SQL-file. Could somebody help me out with this? --[[User:Igor Yalovecky|Igor Yalovecky]] ([[User talk:Igor Yalovecky|talk]]) 16:30, 7 May 2022 (UTC)

Revision as of 16:30, 7 May 2022

 Policy Technical Proposals Idea lab WMF Miscellaneous 
The technical section of the village pump is used to discuss technical issues about Wikipedia. Bug reports and feature requests should be made in Phabricator (see how to report a bug). Bugs with security implications should be reported differently (see how to report security bugs).

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.


Side bar

problem TOC

Is there an opption to not have the new contents side bar as this is causing mass sandwich effect for me....so bad many articles are non readable.--Moxy- 22:06, 25 April 2022 (UTC)[reply]

You are using Vector 2022 currently. If you wish to return to previous Vector, see Special:Preferences under the Appearance tab. Izno (talk) 22:42, 25 April 2022 (UTC)[reply]
That said, I've just now been given a dose of it offwiki and it's, uh, not pretty. Izno (talk) 22:48, 25 April 2022 (UTC)[reply]
Yes Vector 2022...its the skin we tell people that use desktop view on mobile devices to use////because it normally allows you to pick full screen ...as in no "Main menu" but now we have TOC there taking up even more room then the Main meun did. Moxy- 23:23, 25 April 2022 (UTC)[reply]
I don't think anyone should be suggesting Vector 2022 for mobile use at this time. Timeless, Minerva, and Monobook with the responsive option on are all more appropriate. Izno (talk) 23:45, 25 April 2022 (UTC)[reply]
Non allow full screen. To be honest its odd we dont have full screen option on all skins...no need for "main menu" all the time....plus its the way for editing in desktop view on mobile devices.-- Moxy- 00:30, 26 April 2022 (UTC)[reply]
And neither does Vector 2022. I do know it is trivial to get full screen in Timeless. Izno (talk) 01:22, 26 April 2022 (UTC)[reply]
What?/ full page view Moxy- 03:42, 26 April 2022 (UTC)[reply]
The linked Phabricator ticket, won't help with this particular issue (currently we hide the table of contents at low resolutions if you resize your browser, but this is viewport not window size). It likely needs a modification to the meta[name="viewport"] tag to set an initial-scale that zooms out to keep it consistent with the current legacy Vector experience.
Neither Vector's are responsive, so the phone is trying to compensate.
I'll have a think about this and get back to you later in the week with a more appropriate ticket. Jdlrobson (talk) 03:51, 26 April 2022 (UTC)[reply]
The static TOC in the sidebar is an accessibility issue - specifically, it fails MOS:COLOUR (second bullet) Links should clearly be identifiable as a link to our readers. - because the subsection names are shown in black, despite being links. --Redrose64 🌹 (talk) 21:08, 26 April 2022 (UTC)[reply]
This will be fixed with T306562. – Jonesey95 (talk) 14:42, 28 April 2022 (UTC)[reply]
Hey all - thanks for bringing these issues up. A quick update from the team's side. We are aware of some of the issues with the new ToC and are working on and deploying fixes this week and next week. In particular we're looking at the list below.
- ToC on narrower screens. We have increased the threshold for which the ToC currently hides to 1000px. (See T306904 for more context). This was to make it more comfortable to read on narrower screens. This is a temporary fix, though. The conversation about the best solution will continue in T306660. By the end of this week we hope to have clear next steps on our preference of the options presented there.
- We have begun the work on reducing the margins for screens between 1000px - 1200px. This will make the table of contents smaller and create more space for text. We will track this in T307004 and will probably have the implementation ready within a few days. OVasileva (WMF) (talk) 07:45, 4 May 2022 (UTC)[reply]

Restore ToC

Hi, how can I restore the old behaviour of the ToC in Vector 2022? The new one takes up almost a third of the screen. ―  Qwerfjkltalk 15:49, 2 May 2022 (UTC)[reply]

And by old behavior you mean...? Izno (talk) 18:01, 2 May 2022 (UTC)[reply]
If you mean that the TOC takes up too much screen width, maybe add the full width gadget referenced in the next section. That way, hiding the TOC will expand the main page area. If you mean the vertical "sticky" TOC, then you need to switch back to Vector 2010 — GhostInTheMachine talk to me 11:01, 3 May 2022 (UTC)[reply]
I guess you didn't get caught by the A/B testing Izno? Vector 2010's left-hand menu took 16% of my screenwidth. Vector 2022 took 0% (yay, a great gain in working space), this ToC on the left instead of at top experiment takes a massive 28%. It's ..., it's ... bleurgh! Cabayi (talk) 12:16, 3 May 2022 (UTC)[reply]
The benefits of using Timeless. Izno (talk) 17:32, 3 May 2022 (UTC)[reply]
More annoying is that the TOC is gone on mobile (desktop mode) due to the low screen width. ―  Qwerfjkltalk 20:00, 3 May 2022 (UTC)[reply]

Vector-2022 Widening Gadget

Hi all, if you have ideas about an opt-in gadget to widen the viewport in vector-2022, feedback is welcome at MediaWiki_talk:Vector-2022.css#widener_gadget?. — xaosflux Talk 14:46, 26 April 2022 (UTC)[reply]

@Xaosflux: Would the CSS in the collapsed bit would sort of show what it would look like if it were implemented? ― Blaze WolfTalkBlaze Wolf#6545 15:26, 26 April 2022 (UTC)[reply]
Need to be able to hide TOC like the Main Menu ....for more info see Wikipedia:Village pump (technical)#‎Side bar. Causing readability issues for those who use this on mobile view for full screenMoxy- 19:37, 26 April 2022 (UTC)[reply]
When did this change? It looks absolutely horrible on my widescreen monitor, most pages barely utilize more than like 10% of the total page width, and look almost as if CSS styles failed spectacularly! I'd like to know how to revert it ASAP! KPu3uC B Poccuu (talk) 13:01, 28 April 2022 (UTC)[reply]
In Special:Preferences#mw-prefsection-rendering, you can change Skin (I use "Vector legacy (2010)"), then click Save at the bottom. Certes (talk) 14:09, 28 April 2022 (UTC)[reply]
This doesn't change anything for me. Out of all present skins, only Monobook and Timeless (I think is its name) aren't width-limited. KPu3uC B Poccuu (talk) 06:59, 3 May 2022 (UTC)[reply]
Strange. I use Vector 2010 because it displays full width. I can't see any other options or gadgets that may impact this — GhostInTheMachine talk to me 10:43, 3 May 2022 (UTC)[reply]
This worked at one point:
.mw-page-container {
  min-width: none;
  max-width: 1300px;
  padding: 0;
}
.mw-workspace-container {
  max-width: 1400px;
}
.mw-content-container {
  max-width: 1400px;
}
I don't know if it works any longer. (You change the "1300px" and "1400px" to whatever numbers work for you.) Whatamidoing (WMF) (talk) 19:44, 28 April 2022 (UTC)[reply]
@KPu3uC B Poccuu (and FYI to @Whatamidoing (WMF)) there is an experimental gadget you can load from preferences (all the way down in "testing and development") called "wide-vector-2022", you can try that to reclaim your screen from all that whitespace. — xaosflux Talk 20:00, 28 April 2022 (UTC)[reply]

Time spent on Mediawiki projects

Is there a way to see how much time you spend surfing/working on Mediawiki projects? I'm curious. - Klein Muçi (talk) 15:56, 27 April 2022 (UTC)[reply]

It would need to be client-side, logs of 'reading' pages is not available publicly for privacy reasons. — xaosflux Talk 16:02, 27 April 2022 (UTC)[reply]
@Klein Muçi This may not be exactly what you're looking for, but xtools' edit count report includes a timecard that shows what time of day and day of week your edits tend to occur on. ~ ONUnicorn(Talk|Contribs)problem solving 17:41, 27 April 2022 (UTC)[reply]
Some tasks involve offline work, such as researching in books or analysing database dumps, and there's no way to record that. Most edits I can do in seconds but a few take days to prepare and that's not recorded anywhere. Certes (talk) 17:47, 27 April 2022 (UTC)[reply]
Thank you to everyone who commented! The timecard was indeed interesting in regard to my question. I was just curious to know how many hours a day I spend at "Wikipedia" averagely. The only reason I didn't say "Wikipedia" is because, beside changing back and forth between at least 3 different languages of Wikipedias, I also switch back and forth between different projects such as WikiQuote, Meta and MediaWiki. My ideal outcome would also take into consideration the time spent at the Phabricator, the Wikimedia Code Review and SWViewer (or even other offline work as user Certes mentions) but knowing such a thing could be hard to achieved (if not impossible, considering the offline element) I was mostly hoping for a "client-side tracking script" that could track the activity in the aforementioned MediaWiki projects but it looks like such a thing isn't really a thing so...
Thank you anyway! :) - Klein Muçi (talk) 00:52, 28 April 2022 (UTC)[reply]
Klein Muçi, you might find Wikiscan interesting. It has figures for average time per day on different projects. BlackcurrantTea (talk) 14:11, 2 May 2022 (UTC)[reply]
@BlackcurrantTea, I do. Any way I can change the project, let's say, for example, from SqWiki to EnWiki or MediaWiki?
Also, considering that these stats do exist, is it so hard to have a conglomerate of all projects in one place? That would be exactly what I'm looking for. - Klein Muçi (talk) 14:15, 2 May 2022 (UTC)[reply]
At the top, there's a list of projects on the right; click on each one to see the statistics. I don't know if there's a page with totals for all projects for a single user. I see the figures as an approximation of time spent doing things that the servers can record. As Xaosflux and Certes noted, there are many things that can't be tracked. BlackcurrantTea (talk) 15:16, 2 May 2022 (UTC)[reply]

Removing messagebox and standard-talk

Hello all, I intend to remove the messagebox and standard-talk classes from Common.css Soon as a part of moving to TemplateStyles (see MediaWiki talk:Common.css/to do#description). This is because these classes already have well-known replacements (primarily {{ombox}} and {{tmbox}} and "hand-coded" tables/divs).

I am leaving this note here (atypically) because while I have done my best to remove the vast majority of uses that are "public facing" and not archived discussions/unused WikiProjects (see MediaWiki talk:Common.css/to do#messagebox; between this account and my repeat-edits account I'm in the 10k edits range mostly un-substing old AFD messages), I am essentially unable/unwilling to do anything about the 120 thousand uses on user and user talk pages. These uses are mostly WikiProject invitations and barnstars (as best I can, I've fixed the templates that generate the classes).

I do not think making changes to those pages makes sense. It is bot scale to preserve the width, centering, and background of these tables, which is not ultimately deleterious to the page content. However, I would anticipate many editors looking at their user pages and being sad.

I have previously written a transition guide for people who want to restore NavFrame-similar styles to their user pages (which had a similar issue), which I can do here as well if desired. It is fairly trivial in this case for individual editors to adjust their uses of these classes. I am not currently certain where that guide would live.

Please let me know what you think. Izno (talk) 21:33, 27 April 2022 (UTC)[reply]

It sounds to me like a reasonable idea. The old pages will still have its content preserved, which is the important thing in my opinion. as best I can, I've fixed the templates that generate the classes) is good to hear. Rlink2 (talk) 21:41, 27 April 2022 (UTC)[reply]
I've made up MediaWiki talk:Common.css/messagebox in case anyone wants to look at it and/or provide feedback. Izno (talk) 23:24, 29 April 2022 (UTC)[reply]
This is now done. Izno (talk) 21:35, 3 May 2022 (UTC)[reply]

How to get/copy raw html content?

Hi, I want to copy a specific part of a RAW Html content, but I don't know how to do it. For example, I want to get <pre>...</pre> content in this link and then use it in a Mediawiki namespace page. Thanks! ⇒ AramTalk 00:07, 29 April 2022 (UTC)[reply]

@Aram don't think that is very feasible (maybe with some Lua) - what is the end result you are trying to accomplish, perhaps there is another way to go about it. — xaosflux Talk 11:50, 29 April 2022 (UTC)[reply]
Special:ExpandTemplates may be helpful to strip template content. Otherwise look at the page source, on Firefox it is more tools > view source. --Jules (Mrjulesd) 13:05, 29 April 2022 (UTC)[reply]
@Xaosflux and @Mrjulesd I want to add a drop-down list to Special:Upload (on ckbwiki) to add non-free templates (i. e. Template:Non-free use rationale album cover) just like Licensing drop-down list. And I wrote some codes (see below), but I don't know how to get those template codes.
var albumCoverValue = "Album Cover template";

var templateTypeLabel = '<td class="mw-label"><label for="templateType">Template type:</label></td>';
var selectTag = '<select name="templateType" id="templateTypeSelector" onchange="changeOption(this)">' +
    '<option value="">Choose</option>' +
    '<option value="' + albumCoverValue + '">Album Cover</option>' +
    '<option value="poster-value">Poster</option>' +
    '</select>';

var selectTagWrapper = '<td class="mw-input">' + selectTag + '</td>';

$('.mw-htmlform-field-HTMLTextAreaField').after('<tr class="mw-htmlform-field-Non-free-use-rationale-templates">' +
    templateTypeLabel + selectTagWrapper + '</tr>');

function changeOption(selected) {
    var x = selected.options[selected.selectedIndex].value;
    document.getElementById("wpUploadDescription").value = x;
}
Try above in console. We should do something to get template codes for albumCoverValue variable. ⇒ AramTalk 18:28, 29 April 2022 (UTC)[reply]
@Aram: Try using a JSON page? NguoiDungKhongDinhDanh 19:09, 29 April 2022 (UTC)[reply]
@NguoiDungKhongDinhDanh I've never worked with JSON. Do you have any idea? ⇒ AramTalk 19:46, 29 April 2022 (UTC)[reply]
@Aram: You apparently know JS, so I think I can safely assume that you also know what JSON is. Create a JSON page (by changing content model or using a title that ends with .json) with all template codes, then use some JS code to query that page's content. NguoiDungKhongDinhDanh 19:53, 29 April 2022 (UTC)[reply]
I suggest you just add as a section within your existing form; they can be seen here: MediaWiki:Licenses; attempting to insert additional input boxes there will likely become problematic with different skins and with different releases. I sent you a note on ckbwiki where I might be able to look further. — xaosflux Talk 19:54, 29 April 2022 (UTC)[reply]
@NguoiDungKhongDinhDanh I have basic information about it, but I haven't worked on it. Of course, I have created Wikipedia JSON pages, but those were just a copy-paste from other projects (without my skills). So far, I have no idea how to get information from it, but now you've given me some ideas about how it works, and I thank you for that. I need to read some information about it so I can work with it. Thank you!
@Xaosflux I'm not technically that good. But I don't know how the MediaWiki:Licenses page can add those templates contents to the box. If we do so, it may be confused because it merges with the license section. Anyway, I'll wait for your notes. ⇒ AramTalk 20:32, 29 April 2022 (UTC)[reply]

@Aram: When you edit MediaWiki:Licenses the bulleted items work like this:

*$TemplatePageName|$Label
So for example on w:ckb:MediaWiki:Licenses there is a line:
*cc-by-3.0|Creative Commons Attribution 3.0
This uses the template at w:ckb:داڕێژە:Cc-by-3.0
So create any templates you want, apply appropriate protection to them, then edit that Licenses page to add them. Is that what you are looking for? — xaosflux Talk 20:50, 29 April 2022 (UTC)[reply]
@Xaosflux No, I'm not talking about the licensees, but about the usage of Template:Non-free use rationale. On ckbwiki, only the ckb:Template:ھۆکاری بەکارھێنانی نائازاد provided (see ckb:میدیاویکی:Onlyifuploading.js#L-20). There are more templates (see Category:Non-free use rationale templates). I mean how to add those templates usage codes to the box instead of the ckb:Template:ھۆکاری بەکارھێنانی نائازاد? Am I clear? ⇒ AramTalk 21:21, 29 April 2022 (UTC)[reply]
Think I'm missing your goal a little bit. Your Special:Upload already includes lots of non-free templated options, you can remove ones that you don't want be editing the Licenses page as well. — xaosflux Talk 21:43, 29 April 2022 (UTC)[reply]
Hello again @Xaosflux and @NguoiDungKhongDinhDanh, Sorry for reopening this section. I was about to read those documentations User:NguoiDungKhongDinhDanh had provided here and they were very useful for this case. I tried to get some information through JSON, I updated the code above according to those documentations and produced the below code; I'm not that good at coding, but anyway, I made something.
Content
// Choose a non-free template

var url = "https://test.wikipedia.org/w/api.php?" +
	new URLSearchParams({
		"action": "query",
		"format": "json",
		"prop": "revisions",
		"revids": "363256|511722",
		"formatversion": "2",
		"rvprop": "content|ids"
	});

//try {
async function getTemplates() {
	var req = await fetch(url);
	var json = await req.json();

	var query = json.query;
	var Non_free_use_rationale = query.pages[0].revisions[0].content.valueOf();
	var Non_free_use_rationale_poster = query.pages[1].revisions[0].content.valueOf();
	//var comingSoon = query.pages[0].revisions[0].content.valueOf();

	var
		optionText = "Choose", // Disabled

		optionText1 = "Non free use rationale",
		optionVal1 = Non_free_use_rationale,

		optionText2 = "Non free use rationale poster",
		optionVal2 = Non_free_use_rationale_poster;

	//optionText3 = "More coming soon...",
	//optionVal3 = comingSoon;

	var optionTags = $('#templateTypeSelector')
		.append($('<option>').text(optionText).attr('disabled', 'disabled'))
		.append($('<option>').val(optionVal1).text(optionText1))
		.append($('<option>').val(optionVal2).text(optionText2));
	//.append($('<option></option>').val(optionVal3).text(optionText3));

	var templateTypeLabel = '<td class="mw-label"><label for="templateType">Template type:</label></td>';
	var selectTag = '<select name="templateType" id="templateTypeSelector" onchange="changeOption(this)">' + optionTags + '</select>';
	var selectTagWrapper = '<td class="mw-input">' + selectTag + '</td>';

	$('.mw-htmlform-field-HTMLTextAreaField').after('<tr class="mw-htmlform-field-Non-free-use-rationale-templates">' +
		templateTypeLabel + selectTagWrapper + '</tr>');
}

function changeOption(selected) {
	var x = selected.options[selected.selectedIndex].value;
	document.getElementById("wpUploadDescription").value = x;
}
getTemplates();

//}
/*catch (e) {
	console.error(e);
}*/
Please notice that I changed the API source and it's "revids" to allow you try the the code (here: Special:Upload on testwiki) in your browser console and see a drop-down list created just above the "Licensing" drop-down list, but the problem here is an empty list, which produces [Object object], but if try the same code again just after the first time, you now can see the drop-down list we wanted. Can you tell me why it's not completely created at the first time? And if you have any other note about the code, I'm glad to hear them. Thanks! ⇒ AramTalk 13:40, 3 May 2022 (UTC)[reply]

@Aram: I'm not sure why your code doesn't work, but here's how I write it:

$(function() {
	(new mw.Api()).get({
		action: 'query',
		prop: 'revisions',
		revids: [363256, 511722],
		rvprop: ['content', 'ids'],
		format: 'json',
		formatversion: 2
	}).done(function(response) {
		var options = [];
		var text = ['Choose', 'Non free use rationale', 'Non free use rationale poster']; // etc.
		options.push({
			text: text[0]
		});
		for (let i = 0; i < response.query.pages.length; i++) {
			options.push({
				text: text[i + 1],
				value: response.query.pages[i].revisions[0].content
			});
		}
		
		$('.mw-htmlform-field-HTMLTextAreaField').after(
			$('<tr>').addClass('mw-htmlform-field-Non-free-use-rationale-templates').append(
				$('<td>').addClass('mw-label').html($('<label>').attr('for', 'templateType').text('Template type'))
			).append(
				$('<td>').addClass('mw-input').html(function() {
					var r = $('<select>').attr('name', 'templateType').attr('id', 'templateTypeSelector');
					for (let i of options) {
						console.log(i);
						if (i === options[0]) {
							r.append($('<option>').text(i.text));
						} else {
							r.append($('<option>').val(i.value).text(i.text));
						}
					}
					return r;
				})
			)
		);
		$('#templateTypeSelector').on('change', function() {
			$('#wpUploadDescription').val($(this).val());
		});
	});
});

NguoiDungKhongDinhDanh 14:29, 3 May 2022 (UTC)[reply]

thumbs up Great! @NguoiDungKhongDinhDanh: I really want to thank you! Thank you for rewriting, cleaning the code up and removing any potentially error! It's working properly now! ⇒ AramTalk 17:44, 3 May 2022 (UTC)[reply]
@Aram: Hi, you may want to make it a gadget and add a condition instead of inserting directly to MediaWiki:Common.js which will be load for every user on every page. NguoiDungKhongDinhDanh 15:41, 4 May 2022 (UTC)[reply]
NguoiDungKhongDinhDanh Done, but after creating the gadget (see ckb:میدیاویکی:Gadget-UploadTemplates.js), it doesn't worked and produced this error: JavaScript parse error (scripts need to be valid ECMAScript 5): Parse error: Unexpected token; token ; expected in file 'MediaWiki:Gadget-UploadTemplates.js' on line 33. I think this is related to phab:T75714. Do you have any idea? ⇒ AramTalk 14:36, 5 May 2022 (UTC)[reply]
Change let into var. ES6 is no supported for directly loaded JS, only inside lazy loaded JS under certain conditions. —TheDJ (talkcontribs) 14:42, 5 May 2022 (UTC)[reply]
This is highly recommended. You will also need to change line 46 to var j of options, as well as all of its usage in the loop (i to j) to avoid potential overlapping. NguoiDungKhongDinhDanh 14:48, 5 May 2022 (UTC)[reply]
@Aram: As mentioned in the task, you can move the real code to somewhere else while keeping a mw.loader.load at the gadget page. NguoiDungKhongDinhDanh 14:43, 5 May 2022 (UTC)[reply]
@TheDJ and @NguoiDungKhongDinhDanh, Thank you both! It works now. But in the line 46, I just changed the let to var (because var j of options is not working). As a result, I just loaded the gadget (ckb:میدیاویکی:Gadget-UploadTemplates.js) and changed let to var. That is all. Special thanks to you all! ⇒ AramTalk 15:40, 5 May 2022 (UTC)[reply]

Make working with templates easier: More improvements coming soon.

Hello. The last set of improvements from WMDE’s Templates project will be deployed to English Wikipedia soon:

Fundamental changes in the template dialog (example from a left-to-right wiki).

1) You’ll receive a fundamentally improved template dialog in VisualEditor and New Wikitext mode. This will make it easier to understand what is expected from a template, how to navigate the template dialog, and how to add parameters. These improvements have been active on a few wikis for some months already. The little video here summarizes the changes that were made. If you like to learn more, please visit our project page.

2) It will become easier to find and insert templates with an improved search and added warnings (in the TemplateWizard & in the template dialog of VisualEditor and New Wikitext mode).

3) In syntax highlighting (CodeMirror extension), a colorblind-friendly color scheme will be added. You’ll be able to activate it with a user setting.

The planned deployment for 1) and 2) on English Wikipedia is May 17. 3) is planned for May 10. These were the last three out of seven projects in the Templates focus area. Apart from these bigger projects, the Technical Wishes team improved a few other user-facing issues and fixed some bugs related to working with templates. You can learn more about it here.

Feedback on all the projects is much appreciated on their respective talk pages. – Thanks a lot on behalf of WMDE’s Technical Wishes team, Johanna Strodt (WMDE) (talk) 10:47, 29 April 2022 (UTC)[reply]

I'm confused. I thought this was already planned to be released previously? ― Blaze WolfTalkBlaze Wolf#6545 13:28, 29 April 2022 (UTC)[reply]
Yes some of this was previously announced, but then delayed in the week that it was supposed to go live. —TheDJ (talkcontribs) 19:17, 29 April 2022 (UTC)[reply]
Ah alright. Thanks for the clarification. ― Blaze WolfTalkBlaze Wolf#6545 19:22, 29 April 2022 (UTC)[reply]
@Blaze Wolf: Thanks for the question, and @TheDJ: +1, thanks for the clarification! We hope these changes will bring a significant improvement for working with templates. -- Have a good week, Johanna Strodt (WMDE) (talk) 07:38, 2 May 2022 (UTC)[reply]

ENGVAR in newcomer copyedit task

Hi all, I don't know where this query fits best, so I'm asking it here. Are newcomers who are guided by the Newcomers task feature, told about MOS:ENGVAR before giving them the task of copy-editing? So far, I've found quite a few Newcomer copyeditors switch British English to American one and vice-versa. If they aren't told about it already, can it be added to the guide somehow? (Noting that I know nothing about how this thing works) Thanks! CX Zoom[he/him] (let's talk • {CX}) 19:28, 29 April 2022 (UTC)[reply]

Does not seem to be mentioned anywhere. The relevant guidance, reproduced here, is:
* Pppery * it has begun... 19:52, 29 April 2022 (UTC)[reply]
So we should add "Do not change American to British spellings or vice versa" to main-rules1 ? —Kusma (talk) 07:57, 30 April 2022 (UTC)[reply]
I think we need to tell newcomers about that. Newcomers using the guide, thinking that they're making good contributions only to get reverted per ENGVAR doesn't look good. CX Zoom[he/him] (let's talk • {CX}) 13:31, 30 April 2022 (UTC)[reply]
Is there a way to replicate newcomer messages on my account. CX Zoom[he/him] (let's talk • {CX}) 13:37, 30 April 2022 (UTC)[reply]
Go to Special:Preferences#mw-prefsection-personal-homepage, turn on "Display newcomer homepage", then visit Special:Homepage * Pppery * it has begun... 13:42, 30 April 2022 (UTC)[reply]
@Trizek (WMF), I think this is for you. Whatamidoing (WMF) (talk) 17:47, 4 May 2022 (UTC)[reply]
@CX Zoom @Kusma @Pppery, thank you for starting this feedback loop.
What we usually suggest is to edit local messages to fit local rules, only if there is no other way to pass an important information. This information has to be really important, and should not break the concise texts that already exists, because of the well known TL;DR effect.
I would suggest on main-rules1 to go with "You can fix spelling and grammar errors. This might include sentences that are too long, repeated words, or incorrect punctuation. Keep the English spelling used (British, American...)."
What do you think? Trizek (WMF) (talk) 19:08, 4 May 2022 (UTC)[reply]
@Trizek (WMF): I'm not fully convinced this is better than my suggestion. The advantage of your text is that it makes it clear that American and British English are not the only possibilities; that is the main reason I have not WP:BOLDly made a change yet. "Keep the variety of English used (British, American, ...)?" "Do not change the variety of English used (British, American, ...)"? It should be short and to the point, and probably not include a link, especially not to something as terrifying as the MOS. —Kusma (talk) 19:53, 4 May 2022 (UTC)[reply]
@Kusma, it is more encouraging to have Do-s rather than Don't-s. :) This is why I suggest along the lines of "Keep the variety of English used (British, American, ...)". I agree on keeping it short and to the point, and not including the link: a single sentence is supposed to suffice to explain this part of the MOS. But the final call is yours (yours as in "the community"). Trizek (WMF) (talk) 20:16, 4 May 2022 (UTC)[reply]

User talk page warning to flag up likely THEYCANTHEARYOU problem

While waiting (!) for the mobile user notification problem (WP:THEYCANTHEARYOU) to be one day maybe (!!) fixed, would it be a) possible and b) good idea to implement in the meantime a warning which automatically appears on the top of the user talk page of a user who is likely to be not getting notifications (ie. edits only or or mostly on an affected mobile device, and does not edit their talk page)?

I realise this would do nothing to help such users receive the said notifications, but it would at least make it easier for the rest of us to identify this as a possible explanation for their lack of response to warnings etc., helping to AGF and not raise blood pressures unduly. I've seen many user talk pages where the attempts at communicating with the user get increasingly desperate, until someone points out this as the likely reason. I just thought maybe that could be done automatically by way of some clever code.

Thoughts? -- DoubleGrazing (talk) 07:25, 30 April 2022 (UTC)[reply]

I don't see how "automatically" can be done. I would prefer to just disallow editing from any remaining THEYCANTHEARYOU system, ideally combined with pointers to our many working systems instead of a broken one. (This post was written on my phone using a browser and the Monobook skin, which is both easier to use and more powerful than any of the crippled "mobile solutions"). —Kusma (talk) 08:26, 30 April 2022 (UTC)[reply]
By "automatically" I meant some sort of user-side script which I could opt in and put in my common.js or somewhere. When I then visit a user talk page, the script checks the user's edit history and displays (to me) a warning, flagging up things like 'user has never edited their user talk page' or 'user edits using iOS devices only', or whatever the relevant criteria are. Is that not doable? -- DoubleGrazing (talk) 10:04, 30 April 2022 (UTC)[reply]
@DoubleGrazing, WP:SCRIPTREQ? Both of the flags you listed above are probably doable. ― Qwerfjkltalk 10:02, 2 May 2022 (UTC)[reply]
Thanks @Qwerfjkl — didn't even know such a place existed! :) Will try there... -- DoubleGrazing (talk) 11:55, 2 May 2022 (UTC)[reply]

New vector misses hyperlinks for "table of contents"

Hi, in the new vector (2022), the "table of contents" exists in the left side bar, as the last item of it. The problem is that all of headings and subheadings listed in this table misses a hyperlink that redirect the page to the related heading/subheading. I mean, after the event of "clicking on the text" we should scroll/redirect to a heading (via a URL, that its "heading text" is preceded by a "#" sign).

For example, when we open the article Variety_(universal_algebra) in the new vector (2022), there exists a table of content in the left side bar that if we click, for example, on the "Definition" item of that list, no action happens for now, but the correct scenario is that, after clicking, we should scroll the page to the section "https://en.wikipedia.org/wiki/Variety_(universal_algebra)#Definition". Please add required hyperlinks to the table of contents items in the new vector. Thanks, Hooman Mallahzadeh (talk) 10:10, 1 May 2022 (UTC)[reply]

This is a bug in the latest version of Chrome —TheDJ (talkcontribs) 11:17, 1 May 2022 (UTC)[reply]
@TheDJ: I am using Microsoft Edge browser, and this problem exists there. Hooman Mallahzadeh (talk) 11:29, 1 May 2022 (UTC)[reply]
@TheDJ But Opera browser has no such a bug. Hooman Mallahzadeh (talk) 11:32, 1 May 2022 (UTC)[reply]
@TheDJ A browser layer, like Chrome or Edge, is under the layer of a web pages of a website, like articles of Wikipedia. So this bug is on behalf of Wikipedia not from these browsers. We should change and modify Wikipedia such that it would be compatible with all types and versions of browsers like Edge.
Note that we should write Wikipedia webpages in a version of HTML language that is the most standard and basic tags of it, so that a it should not be dependent on a specific version of HTML. The uncompatibility that we face here, exactly comes from the neglection of this principle. Hooman Mallahzadeh (talk) 11:50, 1 May 2022 (UTC)[reply]
The browser version has been out for just a couple of days, fixing bugs takes time. —TheDJ (talkcontribs) 21:04, 1 May 2022 (UTC)[reply]
Yes, Chromium, which is the rendering engine for Chrome, is also used for the same purpose in Edge. It would naturlaly occur in both. Izno (talk) 17:06, 1 May 2022 (UTC)[reply]
Although (as I note below) a device may have different Chromium versions installed with the two browsers at any given time due to different update cycles, which is why I saw different behavior between them. David Brooks (talk) 00:23, 3 May 2022 (UTC)[reply]

Vector-2022 sidebar not clickable in Edge

Subject line says it. Using Edge (version 101.0.1210.32), I can't click the TOC sidebar; also the "page-down" scrolling behavior switching between the sidebar and the main text is hard to describe or predict. The TOC does work as expected in Chrome, which is odd because the two browsers share an engine. Sorry, but I don't know of a better place to record skin problems. David Brooks (talk) 13:30, 2 May 2022 (UTC)[reply]

DavidBrooks, phab:T307360 Alexis Jazz (talk or ping me) 14:05, 2 May 2022 (UTC)[reply]
Now phab:T307271. — xaosflux Talk 14:09, 2 May 2022 (UTC)[reply]
Which is now fixed; thanks to the developers. I've been retired from the business too long, and I apologize for not also reporting my Chrome version, which was 100.0.4896.127. The Chromium (shared by Chrome and Edge) regression was introduced between those builds. David Brooks (talk) 00:20, 3 May 2022 (UTC)[reply]
@Xaosflux@DavidBrooks@Alexis Jazz Today, the same problem raised again in Edge. Hooman Mallahzadeh (talk) 09:07, 5 May 2022 (UTC)[reply]
Thanks for the note, I see phab:T307271 is reopen. — xaosflux Talk 09:45, 5 May 2022 (UTC)[reply]
@Hooman Mallahzadeh, Xaosflux, and Alexis Jazz: I'm not clear whether the WP HTML fix also works in the older browser releases (100.x) as well as 101.x, where the Chromium regression happened. Did the fix for 101 break 100? In any case, Hooman, it's probably worth while to check your Edge version. Use ..., Settings, About Microsoft Edge. David Brooks (talk) 16:21, 5 May 2022 (UTC)[reply]
@DavidBrooks This problem raised today temporarily, for about 4 or 5 hours. But now in my version of Edge browser (Version 101.0.1210.32 (Official build) (64-bit)), ToC links are clickable, too. Thanks for your reply. Hooman Mallahzadeh (talk) 16:40, 5 May 2022 (UTC)[reply]
@Hooman Mallahzadeh: Happy to help. I believe Edge does quietly update itself when it's in the mood (it can take MS a few days to roll out updates of anything Windows-related). I'm on 101.0.1210.32. David Brooks (talk) 16:47, 5 May 2022 (UTC)[reply]

Loading Watchlist very slow

For the past three days the Watchlist can take 15+ seconds just to respond to being clicked. Everything else responds normally as usual. soibangla (talk) 11:47, 1 May 2022 (UTC)[reply]

Happening for me too, even with a relatively small watchlist (sub-200 pages total). Buttons to Push Buttons (talk | contribs) 13:30, 1 May 2022 (UTC)[reply]
Same for me, sometimes 5 seconds, sometimes over 10. Kante4 (talk) 13:32, 1 May 2022 (UTC)[reply]
EVERYTHING is slowing down. Ongoing issue - see phab:T300914. I noticed it at both DYK and my watchlist a couple of days ago, and it just isn't getting better. — Maile (talk) 14:54, 1 May 2022 (UTC)[reply]
I came here to see if anyone else was experiencing this slowdown. Evidently so! It doesn't appear to be related to the size of the watchlist: I tested it with User:Smallerjim's watchlist which has only one item and it still takes about 10s, same as my main one. I haven't noticed an effect on anything other than watchlists, but if anything the delays there are increasing.  —Smalljim  15:07, 1 May 2022 (UTC)[reply]
Try turning down the enhanced filters - phab:T307214.--Snævar (talk) 15:30, 1 May 2022 (UTC)[reply]
Since MediaWiki:watchlist-messages transcludes Template:RfA watchlist notice, it's probably the same underlying issue as #Time-expired problem at Template:Centralized discussion above. I note that using a different language loads much faster, which further supports that guess. Anomie 15:34, 1 May 2022 (UTC)[reply]
I think parsing Tamzin's RfA every time might be causing it? I've tried this; does this improve things? —Kusma (talk) 15:46, 1 May 2022 (UTC)[reply]
@Kusma: Back to normal for me. Kante4 (talk) 15:48, 1 May 2022 (UTC)[reply]
Bravo. Well done, thank you. -Roxy the grumpy dog. wooF 15:52, 1 May 2022 (UTC)[reply]
Wait. How did that change affect me? -Roxy the grumpy dog. wooF 15:53, 1 May 2022 (UTC)[reply]
Much improved, thanks Kusma. Same question!  —Smalljim  15:58, 1 May 2022 (UTC)[reply]
To put what Anomie said slightly differently, the watchlist notice displaying the number of RfAs currently active was causing the slowdown, as it was changed earlier this year to parse the RfA page to count the RfAs. This change was reverted to be a hardcoded number (and so the number may lag the actual number once again). isaacl (talk) 16:02, 1 May 2022 (UTC)[reply]
I'll take my share of the blame; although I couldn't find the conversation in a quick search, I'm pretty sure I'm the one who suggested using the module to generate the count. (I didn't discuss the implementation, but I did look at it once it was implemented and didn't raise any red flags.) My apologies for the slowdown. isaacl (talk) 16:20, 1 May 2022 (UTC)[reply]
Back to normal for me, thank you! soibangla (talk) 16:04, 1 May 2022 (UTC)[reply]
Still don't understand what happened (too stupid for any technical stuff) but thanks. :D Kante4 (talk) 16:10, 1 May 2022 (UTC)[reply]
As I understand it, on everyone's watchlist there's a notice that pops up every time there's a new RfA (to encourage participation). Apparently it used to be updated manually, but when it was automated back in January it now has to read/parse(?) all of the edits in each outstanding RfA (I don't know why), and it has to do this even if the notice is no longer being shown because it's been closed. So Tamzin's current RfA, which has hundreds of edits is slowing it down. Is that anywhere near the truth, techies?  —Smalljim  16:37, 1 May 2022 (UTC)[reply]
Fairly close, since you asked. It doesn't parse the edits, but it does process the wikitext of the page enough to count the number of !votes in each section. And it doesn't actually have to do that, it just happens that the existing code that was re-used does so. Anomie 16:49, 1 May 2022 (UTC)[reply]
Wow! After reading the above, I simply dismissed the RFA notice, which never returned. And now my Watch List loads at lightening speed. — Maile (talk) 17:10, 1 May 2022 (UTC)[reply]
Hmm, not entirely convinced... I removed that RfA notice from my watchlist a couple of days ago, and although the list seems to load faster now, it's still loading noticeably slower than it used to. -- DoubleGrazing (talk) 11:54, 2 May 2022 (UTC)[reply]

Map legend in different languages at thumb and full size

For some reason at thumbnail size the legend of File:2022 Russian invasion of Ukraine.svg at bottom left appears in English, but the full size shows legend in Russian. Is it fixable? Also, the names of smaller villages and towns are illegible even at full size, but this is for Graphics Lab, it seems. Brandmeistertalk 15:53, 1 May 2022 (UTC)[reply]

@Brandmeister: You are using Russian as interface language, not? NguoiDungKhongDinhDanh 16:09, 1 May 2022 (UTC)[reply]
No, it seems, the Render this image parameter below the image is set to English and in my Preferences, Internationalisation is set to en - English. Brandmeistertalk 16:16, 1 May 2022 (UTC)[reply]
After a while playing with it myself, I guess this is the intended behaviour, and that there are no bugs to be fixed. You can view any version by appending ?lang=langcode to the URL. NguoiDungKhongDinhDanh 16:32, 1 May 2022 (UTC)[reply]
@Brandmeister: The original SVG file has text versions in multiple languages and your browser apparently chooses Russian. Mine chooses English. A thumb version is not an SVG file but a PNG rendered by MediaWiki which chooses English at the English Wikipedia, Russian at the Russian Wikipedia and French at the French Wikipedia (I didn't test the remaining languages in the SVG file). PNG is a purely graphical image format which doesn't support text in multiple languages. It doesnt support text at all but just displays an image which may contain symbols interpreted as text in some language. PrimeHunter (talk) 23:27, 3 May 2022 (UTC)[reply]
How odd, ok. Brandmeistertalk 18:30, 4 May 2022 (UTC)[reply]
@Brandmeister: It's determined by the browser and not MediaWiki when you view the original file. Your browser may have a way to set a list of preferred languages. My Firefox does. If you prefer English then you can examine browser settings or ask for help at Wikipedia:Reference desk/Computing. Remember to name your browser. The setting can also affect multilingual websites. PrimeHunter (talk) 19:32, 5 May 2022 (UTC)[reply]
I use Firefox in Russian and now understand it, it seems. But on the other hand this is counter-intuitive and at first looks like odd behavior. Brandmeistertalk 19:40, 5 May 2022 (UTC)[reply]

"Special pages"?

Hello, Village Pump Technical folks,

I was wondering if anyone here knew who or what bot was responsible for updating the Special Pages lists. Specifically, I'm interested in Special:UnusedCategories and Special:WantedCategories, lists which typically update every 3 days and are hours late updating today. The Wanted Categories helps us get rid of red link categories (per WP:REDNO) and the Unused Categories includes categories that are not included on the Database report of empty categories like maintenance categories so it is very useful. If I had my way, these lists would update daily, not every 3 days but, again, I'm not sure who to contact about this. The lists do have talk pages but if you visit them, you'll find questions posted from years ago that were never replied to so I'm not sure who is involved with maintaining these "special pages".

Any help or referral to who I could ask about this delay in posting the lists, would be much appreciated! Thanks! Liz Read! Talk! 20:02, 1 May 2022 (UTC)[reply]

MediaWiki itself updates the lists automatically and periodically. It is also reported that other things have been much slower recently, so I guess this is just another one in multiple of consequences. NguoiDungKhongDinhDanh 20:24, 1 May 2022 (UTC)[reply]
Those can't really be put on an less than 3 day update scheme, because all of the special pages collectively take 3 days to update - source: phab:T17434. Those are the responsibility of the developers.--Snævar (talk) 20:36, 1 May 2022 (UTC)[reply]
Thank you so much for this information, NguoiDungKhongDinhDanh and Snævar, this is much more than I knew before about this group of pages. If it is coming from MediaWiki, well, then there is no bot operator I can try to convince to change the timing. I'll just be grateful for the reports when they do get updated. Thanks again. Liz Read! Talk! 21:36, 1 May 2022 (UTC)[reply]
The special pages I use updated, as expected, on today, May 4th, so I guess what was broken has been fixed. Again, I appreciate the information provided here. Liz Read! Talk! 05:55, 5 May 2022 (UTC)[reply]

Dark mode by default to save energy (and the earth)

Wikipedia being a very popular website can have a high impact in reducing global warming by reducing computer resource use in server and clients. More and more devices use OLED technology, and energy can be saved by using black backgrounds. Mobile vendors have automated this, switching to dark mode when battery is low.

Wikipedia's commitment to reduce global warming is mandatory. Let's set dark mode as default. Let's change the world. — Preceding unsigned comment added by 190.19.176.171 (talk) 22:09, 1 May 2022 (UTC)[reply]

To anyone interested, if you enable the dark-mode gadget and set window.wpDarkModeAutoToggle = true in your common.js, dark mode will automatically turn on/off when your device changes its colour scheme (eg. in Mac you can set system theme=auto so that dark mode is activated at sunset). – SD0001 (talk) 07:13, 2 May 2022 (UTC)[reply]
OP: [citation needed]. – Jonesey95 (talk) 23:15, 2 May 2022 (UTC)[reply]
OLEDs don't have a backlight, so dark mode does save a small amount of energy. I believe that most recent iPhones use this technology, and that almost no desktop computers do. On a modern non-OLED display, the energy difference is on the order of 5% of the power used by the display itself, which is not necessarily the most significant source of power use by the whole computer (power use depends significantly on what you're doing, not just whether the computer is turned on). Whatamidoing (WMF) (talk) 18:27, 4 May 2022 (UTC)[reply]

Chrome browser edit window fonts

Seeking advice from anyone who uses Chrome. I use both Chrome and Firefox. In Firefox, we have a Settings tab that opens to a selection for Fonts. I've never had any issue with the Firefox fonts. In Chrome, I have the browsing window fonts the size and look I want. But no matter how I adjust Settings, Appearance, Fonts, my edit window has such teeny tiny fonts that I can't read them even with a magnifying glass. It's been this way for a while. I have one device that runs on Windows 10, and one that runs n Windows 11 - both have this issue with Chrome. Can anyone offer advice on how to adjust the edit window view in Chrome? Edge browser does the same as Chrome, but I rarely use Edge. Thanks. — Maile (talk) 22:55, 1 May 2022 (UTC)[reply]

@Maile66 this probably isn't it, but especially if you are using modern or monobook, you have kind of a mess of personal css going on. You are double importing the same personal user script in your common.css and your skin css'; but that script is then turning around and loading a gadget. If you want to use navpopups I suggest you remove all the entries to it in User:Maile66/common.css, User:Maile66/modern.css, and User:Maile66/monobook.css - and just use the gadget in preferences.
The next step to try would be to at least temporarily turn off everything in User:Maile66/common.js and see if your issue gets better. — xaosflux Talk 18:35, 4 May 2022 (UTC)[reply]
@Xaosflux: I don't use Monobook, but maybe I did in the past and it loaded. I deleted Monobook css and deleted the Modern css. So far, nothing changes on Chrome, but I haven't yet deleted that Modern css. FYI, I have a newer computer that I've just set up. On Firefox, I had these issues, until I installed NoScript, which magically corrected everything on that browser. I'm thinking I used to have NoScript on Chrome, but Chrome now blocks NoScript from being added, and deleted it from my Chrome browser some time ago. That might be why Chrome has gone weird. Wikimedia Commons itself looks strange to me on Chrome. — Maile (talk) 19:34, 4 May 2022 (UTC)[reply]
@Maile66 perhaps you have javascript disabled? See if you can see your own address on this page after clicking the button: Wikipedia:Get my IP addressxaosflux Talk 20:40, 4 May 2022 (UTC)[reply]
@Xaosflux: Yes, clicking on that link, and then clicking as instructed when it tells me to, brings up my IP address. — Maile (talk) 21:16, 4 May 2022 (UTC)[reply]

Google translate shows login message for Wikipedia

When reading a Wikipedia page through Google Translate I get this mw.notify popup: "Central login You are centrally logged in. Reload the page to apply your user settings." I don't get it when I'm logged out on Wikipedia. Would this mean Google could see my username when I read a Wikipedia page through Google Translate? Alexis Jazz (talk or ping me) 14:01, 2 May 2022 (UTC)[reply]

@Alexis Jazz when quickly looking at the debug from say "en-m-wikipedia-org.translate.goog" - I'm seeing tons of direct connections from my client to WMF sites. So I think this is a "maybe" - but not because Google is proxying your request, but possibly because your client is sending that data to Google yourself and asking them to translate it. — xaosflux Talk 14:14, 2 May 2022 (UTC)[reply]
Xaosflux, would it technically be possible for someone to set up "whatismywikimediausername.com", similar to sites like whatismyip.com? I feel like that shouldn't be possible (imagine for example a Christian school harvesting Wikimedia usernames from students to ensure they don't edit unholy subjects), but it scares me a bit that you said "maybe". Alexis Jazz (talk or ping me) 14:53, 2 May 2022 (UTC)[reply]
Cross Original Resource Sharing is blocked by default in modern browsers, however most browsers have ways you can purposefully enable this. — xaosflux Talk 15:18, 2 May 2022 (UTC)[reply]
From what what I can gather looking at the requests, all the requests that are successfully being made that include your username are saving it in cookies only accessible by mediawiki sites. The request that gets your username to add to the central login notice doesn't go through properly, but even if that were fixed, it is would still be blocked by the Cross Origin Resource Sharing (CORS) policy. That's why you see the message "You are centrally logged in. Reload the page to apply your user settings." and not "You are centrally logged in as USERNAME. Reload the page to apply your user settings." So unless the CORS policy is ignored by the browser, it wouldn't be possible to create a site like "whatismywikimediausername.com". BrandonXLF (talk) 19:38, 2 May 2022 (UTC)[reply]

Does this template do anything?

While attempting to figure out how to use a tree list, I found Template:TreeList/Branch End and Template:TreeList/Final Branch which don't seem to do anything or be connected to anything. The creator has been indeffed for years, and these are only used in an abandoned user sandbox from said indeffed user. I imagine these are useless and should probably be deleted/redirected, but as "I don't see a use for this" isn't a valid deletion rationale, I wanted to see if anyone more familiar with the template space than me could determine if these are truly junk. Hog Farm Talk 17:48, 2 May 2022 (UTC)[reply]

It would appear to be a duplicate of Template:Tree list/branching. Either redirect or send to TFD. Izno (talk) 18:02, 2 May 2022 (UTC)[reply]
I've redirected them. Hog Farm Talk 18:07, 2 May 2022 (UTC)[reply]

Tech News: 2022-18

19:32, 2 May 2022 (UTC)

Can the source editor's regex replace change case?

In some regex implementations, \U, \L, \I and \F change case (to upper, lower, intials and first, respectively). Does the regex replace in WP's source editor do that? Cheers. — Guarapiranga  11:57, 3 May 2022 (UTC)[reply]

@Guarapiranga: No, since they are not supported by JS regex. NguoiDungKhongDinhDanh 13:10, 3 May 2022 (UTC)[reply]
Wouldn't testing it be a quicker way of knowing? Izno (talk) 17:34, 3 May 2022 (UTC)[reply]
I did, of course, thank you. My question was whether it had any particular, perhaps undocumented, syntax that did that. — Guarapiranga  23:53, 5 May 2022 (UTC)[reply]
You can do some things by replacing with a subst expression using string templates. See e.g. Template:String-handling templates. PrimeHunter (talk) 18:14, 3 May 2022 (UTC)[reply]
Right! Hadn't thought of that, PrimeHunter. Thanks.
P.S.: Actually, this is what I needed. — Guarapiranga  23:55, 5 May 2022 (UTC)[reply]

Is there a Vector-2022 talk page?

I had trouble deciding where to report the Chromium regression in the Vector-2022 TOC, and only later found a discussion further up this page. I think it would be useful to have a place to focus reports, announcements, and comments rather than scatter them in this page and presumably others. Would it be possible to create its own talk page (if it doesn't exist) and, so long as it remains experimental, add static code creating a collapsible banner in the Vector-2022 skin only pointing to said page? If the page already exists, am I the only one who can't find it? David Brooks (talk) 14:21, 3 May 2022 (UTC)[reply]

@DavidBrooks probably mw:Talk:Reading/Web/Desktop Improvements is the best place for the near term, assuming you are not running in to something that only affects the deployment on the English Wikipedia. — xaosflux Talk 14:44, 3 May 2022 (UTC)[reply]
Agree, that being the talk page for mw:Skin:Vector/2022. --Redrose64 🌹 (talk) 19:49, 3 May 2022 (UTC)[reply]
I got a reply on mw:Talk:Reading/Web/Desktop Improvements#Provide a link to this page?: there already is a plan (phab:T307113) to add a pointer to that page in the Preferences/Appearance table next week. David Brooks (talk) 14:28, 4 May 2022 (UTC)[reply]

Bug in Special:Contributions - lost checkboxes

If I go to the contribs page for a given user - say Special:Contributions/Redrose64 - open up the "Search for contributions" collapsed box, and select a namespace, I am then offered two more checkboxes: "Invert selection" and "Associated namespace". Whatever I do with those two, if I click on Search, then use my browser's "back" feature and open up the "Search for contributions" collapsed box again, the chosen namespace is still selected, but the two checkboxes are no longer present. To get them back, I need to select a different namespace, then reselect the one that I wanted first time. Firefox 99.0.1, all skins. --Redrose64 🌹 (talk) 20:45, 3 May 2022 (UTC)[reply]

I filed this as T307530 and I'll submit a patch. Fun little bug. Matma Rex talk 22:32, 3 May 2022 (UTC)[reply]
Thank you --Redrose64 🌹 (talk) 23:17, 3 May 2022 (UTC)[reply]

Is iCloud Private Relay the likely cause of my IP address being blocked from editing?

On different days, trying to edit different articles, when using the Safari browser, I was always receiving the "This IP address has been blocked from editing Wikipedia" message. I do not receive the message when using Chrome.

I'm on a Macbook Air m1 (2020); macOS Monterey 12.3.1 (21E258); Safari Version 15.4 (17613.1.17.1.13) [most recent, up-to-date versions].

If I google "What is my IP" on Safari I get an IPv6 address (xxxx:xxxx:xxxx:xxx::xx:xxx); on Chrome I get an IPv4 address (xx.xxx.xx.xxx). That might be normal - I don't know much about how this works.

I might have fixed the problem, but I'm not sure. I realized that iCloud Private Relay might be the problem. So I changed the Safari privacy setting [Safari > Preferences > Privacy > Hide IP address] from “From Trackers and Websites” to “From Trackers Only”. I closed Safari and then opened it. I no longer received the "this IP address has been blocked" message. Interestingly, when I changed the privacy setting back to "From Trackers and Websites" I was still good, i.e., I'm not blocked. (I spoke too soon. The block returned after 20 minutes. I have now changed it back to "From Trackers Only" where it will stay so that I can edit Wikipedia using Safari.)

Is iCloud Private Relay the likely culprit? If so, then I know what to do if this happens again! :)

Many thanks - Mark D Worthen PsyD (talk) [he/him] 07:15, 4 May 2022 (UTC)[reply]

You will probably want to tune in to the discussion at meta:Talk:No open proxies/Unfair blocking and some set of the Phabricator tasks in this search. Izno (talk) 07:19, 4 May 2022 (UTC)[reply]
Ah, thank you Izno. Now that I know what to search for, I found this also, in the Village pump archives: Effect of Apple’s iCloud Private Relay. Much appreciated - Mark D Worthen PsyD (talk) [he/him] 07:23, 4 May 2022 (UTC)[reply]
Yes, we broadly block editing (not reading) from proxies and VPN's. You should turn that off here. — xaosflux Talk 15:05, 4 May 2022 (UTC)[reply]
You can't turn it off "here". You have to turn it off "everywhere, including in situations that most internet-savvy people consider to involve bigger-than-usual privacy risks". There is no way to use that service only for certain websites and not for others. Whatamidoing (WMF) (talk) 18:40, 4 May 2022 (UTC)[reply]
Maybe something like: Close your other tabs, switch off your Private Relay, edit Wikipedia, then turn Private Relay back on. Alternatively, just use another browser to edit here. And I'd just add that yes, it probably is the cause. Your experience varies because some Private Relay addresses are blocked, and some are not. -- zzuuzz (talk) 19:07, 4 May 2022 (UTC)[reply]
Hmm, doesn't this feature you enabled only work with Safari? Perhaps use a different browser for sites you don't want to run through Apple's proxy? — xaosflux Talk 13:03, 6 May 2022 (UTC)[reply]
That would be a solution for people like me, who normally keep multiple web browsers open anyway. I don't think it's practical for mobile editors or less tech-savvy people, though. You'd have to correctly guess what the cause is and guess that switching browsers would change your IP address (=a non-standard behavior). Whatamidoing (WMF) (talk) 20:47, 6 May 2022 (UTC)[reply]
Isn't the standard behavior not to use that VPN? You have to both be a paying customer and opt-in for it? Perhaps we can improve some of our block messages to hint people about this. Looks like we have some hints in {{CDNblock}}, but not in {{Colocationwebhost}} - both of which I'm seeing on some of these blocks. @Markworthen - are you getting more in the block message than just "This IP address has been blocked from editing Wikipedia"? — xaosflux Talk 21:14, 6 May 2022 (UTC)[reply]
@User:Xaosflux - Nothing else. Adding info about iCloud Private Relay and similar "innocent" causes would forestall frustration and perhaps reduce support requests. Given our ongoing struggle to attract and retain good editors, improving usability is a worthwhile goal. Mark D Worthen PsyD (talk) [he/him] 01:53, 7 May 2022 (UTC)[reply]

Dark mode

The dark mode is too extreme in my opinion, the blue text doesn't really look very good on a black background. Can I suggest to any of the developers reading this to make the dark mode a dark grey with white text and lighter grey for article links, similar to the design of Wiki Wand? Though something which places the article titles on a black background with white text I think would be a good addition too, something to embolden the article names.The black is too extreme for the article backgrounds though. Perhaps an option to have the article title and side panels with black and white and the articles white and normal that would be useful rather than the whole thing black and white. ♦ Dr. Blofeld 08:58, 4 May 2022 (UTC)[reply]

@Dr. Blofeld please follow up at Wikipedia talk:Dark mode (gadget). — xaosflux Talk 15:03, 4 May 2022 (UTC)[reply]
Thanks.Zao.♦ Dr. Blofeld 16:35, 4 May 2022 (UTC)[reply]

Petscan

Why since 1-2 months this link loads very slowly and since few days it shows only "No result for source categories" while other categories works and loads fast? Time to time there is always some problem with this tool. If it's due to Wikimedia softare update - can't someone chceck if tool works before or atleast after update implementation? Eurohunter (talk) 12:08, 4 May 2022 (UTC)[reply]

You will need to ask the developer of Petscan. Izno (talk) 17:39, 4 May 2022 (UTC)[reply]
It's a problem that has been going on for at least four months. I first noticed it in early January this year. --Redrose64 🌹 (talk) 21:33, 4 May 2022 (UTC)[reply]

Gallery pictures look blurry on Edge and Chrome (Windows)

Demonstration of a Wikipedia bug - Chrome
Demonstration of a Wikipedia bug - Edge
  • Edge version: 100.0.1185.50 (64-bit)
  • Chrome version: 100.0.4896.127 (64-bit)
  • Issue: On gallery templates, image previews look noticeably worse
Tube·of·Light 14:31, 4 May 2022 (UTC)[reply]
I'm not exactly seeing the issue here. The images look perfectly fine. ― Blaze WolfTalkBlaze Wolf#6545 14:33, 4 May 2022 (UTC)[reply]
On the top four picture thumbnails there is some weird phenomenon kinda like aliasing (the left side of the linked picture). When I click on the thumbnails, it shows clear pictures. Tube·of·Light 16:09, 4 May 2022 (UTC)[reply]
Ohhh! I see now. They're slightly pixelated. I was bit confused when you said the images were blurry. ― Blaze WolfTalkBlaze Wolf#6545 16:11, 4 May 2022 (UTC)h[reply]
I can reproduce the issue in the same article as OP. The issue here seems to be related to wrapping the gallery onto multiple lines. If all the images are on the same line you get undersized and blurry previews, if you adjust the browser window so that the gallery wraps onto multiple lines they pop back to being full quality. 163.1.15.238 (talk) 15:13, 4 May 2022 (UTC)[reply]
I checked just now and the images are blurry at zoom levels of 100% and lower, while zooming to 110% or higher fixes the issue. Tube·of·Light 16:11, 4 May 2022 (UTC)[reply]
Gallery's do some up/down scaling at certain sizes, depending on the available space in the page, to try to make them fit the entire width of the page. This scaling can definitely have this effect and there probably is a bug that scales it down by 1 pixel, even at 100%. This is pretty old code that does this, and in hindsight, probably wasn't the smartest idea to add it at all, but its been in place for 10'ish years. The whole thing is not helped by the fact that Windows isn't that good at scaling images/text. —TheDJ (talkcontribs) 09:09, 5 May 2022 (UTC)[reply]


Time restricting scripts

Hi all, I'd like to restrict several of my scripts to a certain time limit. So I'd like to know what code to use that is technically the same as the following line:

If current time is less than 1830hrs OR greater than 1930hrs, then
Run scripts

Thanks! CX Zoom[he/him] (let's talk • {CX}) 14:56, 4 May 2022 (UTC)[reply]

@CX Zoom: here is a super quick example you can build from:
date = new Date();
time = date.getHours();
if (time <= 5 || time >=12) {
  mw.notify('The time is '+time);
}
This pop ups the current hour if it is before 5 or after 12. You can replace mw.notify('The time is '+time); with your list of scripts. — xaosflux Talk 15:26, 4 May 2022 (UTC)[reply]
Note that the Date() constructor uses your browser's time zone. NguoiDungKhongDinhDanh 15:44, 4 May 2022 (UTC)[reply]

Java/Google Script help (not related to Wikipedia)

After the help I got here I was able to make this script work. Is anyone crafty enough to point me how to rewrite it in a more elegant way so I don't have to rewrite lines like I've already done? - Klein Muçi (talk) 00:23, 5 May 2022 (UTC)[reply]

VPT is not a general support forum. Either return to the reference desk or try one of the many other forums on the Internet. Izno (talk) 01:31, 5 May 2022 (UTC)[reply]
@Klein Muçi the link doesn't work for me.
While i do think generally the VPT should be restricted to wikipedia, I'm not totally opposed to this request because there are many people here who would know how to fix it but dont watch that page. Besides, the ratio of Wikimedia to non wikimedia stuff Klein posts on here is very very very high so I don't think one off topic request is a bad thing. Rlink2 (talk) 01:58, 5 May 2022 (UTC)[reply]
@Izno, is the Reference Desk located so far from the VPT as to treat like a whole different place? Just to be clear I was surprised myself when I heard about the existence of the ref-desk in the computing aspect in here (EnWiki) which basically replicated Stack Overflow, and it did take me quite some reading through its archives to make sure it was really supposed to work like that. For your personal curiosity, the whole thing is in fact 90% "related" to Wikimedia because of this and MediaWiki Gerrit notifications. It's been a while I'm struggling to deal with the large amount of emails from Wikimedia outlets. See phab:T300030 and this discussion to see what I mean. I get around 500-1000 emails per month just in regard to Wikimedia projects. You can also see this discussion with my attempts in automatic watchlist expirations and even my big support for such a feature in the 2022 Wishlist (ironically my vote is just behind yours), all attempts to cull the number of emails I get in an organic manner.
@Rlink2, maybe this one will? - Klein Muçi (talk) 02:33, 5 May 2022 (UTC)[reply]
@Klein Muçi: is the Reference Desk located so far from the VPT as to treat like a whole different place? Unequivocally. The reference desks are barely tolerated. VPT is for technical discussion focused on English Wikipedia. Izno (talk) 03:48, 5 May 2022 (UTC)[reply]
The reference desks are barely tolerated. - I wasn't aware of this reality at all. I will be more careful in regard to that aspect now. - Klein Muçi (talk) 07:54, 5 May 2022 (UTC)[reply]
I think in the future a notice saying "intrested people might be able to answer the question at the reference desk" so all discussion happens there, while still notifying the people that could answer the question. Again, 99% of people would probably be happy to answer anywhere, but just to keep 100% happy that is probably the best approach. And no one has identified any issues with him posting this here, all thats being said is that "its offtopic". Its kinda of like saying WP:IDONTLIKEIT. If someone thinks something is a problem, identify why it is a problem instead of just saying "its a problem". The time spent on saying "off topic" could be used to type out "its off topic, but i solved your question anyway here you go."

I think Klein's post was acceptable because there is no way I would have known if he hadn't posted here. Because of his post, we got the problem fixed - at the end of the day Wikipedia is not a real life or a beucracrcy, we don't need to follow rules 100% exactly. Rlink2 (talk) 23:54, 5 May 2022 (UTC)[reply]
I'm sure user Izno was only acting in good faith in protecting the integrity of the VPT domain. We have had the chance to stumble on each-other on some discussions around EnWiki on the past and I'm sure he also respects me the same as I do respect him and his contributions on this project.
I was trying to get "straight to the point" with my title thinking that it would be a trivial deed which wouldn't require too much work and words (hence also the 2 lines of text below it) but some titles can be provocative and maybe coupled with the fact that the provided link (which should have brought more context) wasn't working, the post may have appeared a bit too "alienated" with this place.
I already got my help thanks to you though so I suppose we can also end this discussion here if nothing else is to be added. Cheers! :) - Klein Muçi (talk) 02:05, 6 May 2022 (UTC)[reply]
@Klein Muçi still not working, says its "private" Rlink2 (talk) 12:42, 5 May 2022 (UTC)[reply]
@Rlink2, hmm... I'll continue this conversation on your talk page if that's okay with you. - Klein Muçi (talk) 12:44, 5 May 2022 (UTC)[reply]

Google Indexing User Sandbox Pages

Was somewhat alarmed to see Google indexing User Sandbox pages e.g. This one [6]. I typically use this as an incubation page before getting articles to minimum standards for mainspace. But, this one was somewhat surprising almost alarming. Ktin (talk) 03:37, 5 May 2022 (UTC)[reply]

Hard to say, as it is now moved and the search results already updated. But in general, Google considers almost everything a hint. Like noindex, is used by them as 'index but do not return in results'. And I'm pretty sure that if you are specific enough and its search ranking is high enough for what you are looking for it might STILL bring it to the surface. —TheDJ (talkcontribs) 08:59, 5 May 2022 (UTC)[reply]
@TheDJ: Yup. Hard to say. Still seems to be returning the sandbox result. https://www.google.com/search?q=david+walden+wiki Ktin (talk) 16:21, 5 May 2022 (UTC)[reply]
@Ktin: You could consider putting {{User sandbox}} at the top of all your sandbox pages. This will apply the __NOINDEX__ magic word. But be aware that once Google has indexed any page, it's difficult to persuade them to unindex it - whilst they may act on a page being changed from normal to NOINDEX, it's certainly not instantaneous. More at WP:NOINDEX. --Redrose64 🌹 (talk) 22:37, 6 May 2022 (UTC)[reply]

User blocker

Hi is there some template I can use to block individual editors or automated messages for prods and AFDs etc? I have a bot blocker but nothing I can code to block individual editors. Like on Twitter there ought to be a block user option from posting on your talk page.♦ Dr. Blofeld 09:22, 5 May 2022 (UTC)[reply]

There is no way for you to turn off talk page messages from someone. You can turn off pings and emails, if you would like -- Guerillero Parlez Moi 12:20, 5 May 2022 (UTC)[reply]
@Dr. Blofeld: phab:T165124 is about a possible user control to block other users from your usertalk. While this could be invented, I don't expect it will ever be part of the English Wikipedia. But what are these "automated" messages you are getting? Are these really "edits" someone is maybe using scripts to help make, or is this some other process? — xaosflux Talk 12:35, 6 May 2022 (UTC)[reply]
I want something to block people with deletion prods and articles for deletion. So some sort of coding to block editors with "Proposed deletion" and "listing for deletion" notices from editing my talk page is probably possible.♦ Dr. Blofeld 12:44, 6 May 2022 (UTC)[reply]
@Dr. Blofeld right now the only way to do that would be with an edit filter, targeting your talk page - and that's not going to happen. Now, if perhaps you were only concerned with assisted edits that come from Twinkle (such as these) you could ask over at Wikipedia talk:Twinkle. Perhaps they could build a blacklist of "Users that do not want Twinkle assisted talk notices" in to that script. — xaosflux Talk 12:51, 6 May 2022 (UTC)[reply]
Is this one specific individual pestering you? Admins can set a WP:PBLOCK on a named individual that prevents them from editing certain pages. This is normally only done for a good reason, not "this user is annoying me". --Redrose64 🌹 (talk) 22:29, 6 May 2022 (UTC)[reply]

Movement Strategy Implementation Grant for the creation of an all-purpose file tool

Hello! I'll be developing an all-purpose file tool that includes the feature to detect potential copyright-violating images that also appear on the Internet. I thought it would be a good idea to request a Movement Strategy Implementation Grant from the Wikimedia Foundation. You can find more information about the scope of the tool and the grant on m:Grants:Project/MSIG/EpicPupper/Fortuna. Feel free to leave comments, questions, suggestions, or ideas on the grant talk page, and endorsements or offers to translate or localize the tool in the relevant sections. Thanks! 🐶 EpicPupper (he/him | talk) 09:24, 5 May 2022 (UTC)[reply]

New "named" group

As of gerrit:778684, there is now an implicit "named" group; currently, it applies to all users. This is causing issues in some scripts, but seems to be intended behavior on MW's end, so I haven't filed a Phab task about it. If someone does think it's unintended behavior, they're free to file.

What it does mean, though, is that a number of scripts need to be updated to exclude "named" as a meaningful group, just like they exclude "user" and often "autoconfirmed", perhaps most notably Navigation popups. I'll try to get to that and User:PleaseStand/userinfo.js presently, but I'm sure there's others. -- Tamzin[cetacean needed] (she/they) 12:12, 5 May 2022 (UTC)[reply]

Okay, looks like this is being disabled for now. Still might be a good idea to cover these bases before it does go into effect as part of IP masking. -- Tamzin[cetacean needed] (she/they) 12:45, 5 May 2022 (UTC)[reply]
Self-closing. Tim said the name "named" may change, so there's nothing that can be done now, script-update-wise. -- Tamzin[cetacean needed] (she/they) 12:51, 5 May 2022 (UTC)[reply]
Unclosed for more comments. — xaosflux Talk 13:48, 5 May 2022 (UTC)[reply]
I was about to unclose this myself, as based on further conversation with Tim, it does look like "named" will be final as the internal name, with the default message for the name still TBD (maybe "registered user"). -- Tamzin[cetacean needed] (she/they) 13:52, 5 May 2022 (UTC)[reply]
I initiated the local messages and updated the most popular gadget (navpopups) using this to put the name "Named user" on it, and not to populate navpopups with this. — xaosflux Talk 13:48, 5 May 2022 (UTC)[reply]
This is a stopgap, expect we can revert once the train comes around again and use "Registered user" or whatever it will be. — xaosflux Talk 15:32, 5 May 2022 (UTC)[reply]
This is confusing nomenclature. What exactly is a "named user" as opposed to just a "user"? I believe all "users" are registered accounts, so what's the difference from "named user"? Themaxtiger (talk) 15:56, 5 May 2022 (UTC)[reply]
@Themaxtiger In a few months/years editors that aren't logged in are going to be identified by pseudo-anonymous names rather than IP addresses (for some kind of privacy reason that hasn't really been explained), so instead of the software having groups "all" (for both IPs and logged in editors) and "users" (for registered accounts) they're going to have groups that will be something like "registered users/named users" and "unregistered users/anonymous users". This should be a mostly behind the scenes change, We'll probably have to fix a few edit filters and the like but most things where this is used (like determining whether you have the right to perform an action) should be updated by the devs. 192.76.8.77 (talk) 18:44, 5 May 2022 (UTC)[reply]
for some kind of privacy reason that hasn't really been explained → see this on meta. It gives two reasons: preventing unsavory governments from finding out who made certain edits, and something about privacy norms related to RGPD (it’s not clear from the text if Legal believes that posting the IP addresses could be a legal problem, or if they think the spirit of RGPD means there is some societal pressure to mask it).
IMO the most likely attack model is not government actors. Rather, if Alice shares a household or at least a computer with Bob (she could be a spouse, roommate, parent...), she can easily see what edits Bob made, retrieving Bob’s IP address with access to the computer takes little technical skill. (I know, IP rotate with DHCP etc. but Alice can at least find out a good portion of Bob’s edits.) TigraanClick here for my talk page ("private" contact) 13:26, 6 May 2022 (UTC) [reply]
If Bob is so worried about his abusive relationship with Alice, he can sign up for a free account. --Ahecht (TALK
PAGE
) 14:55, 6 May 2022 (UTC)[reply]

Talk page appears in Google search results

What accounts for the Talk page of this article, but not the article page, appearing in the Google index?

In this Tea house discussion, OP Llmeyers (talk · contribs) asks about the recently created Carey R. Dunne article, and why a Google search turns up only the Talk page, providing the following link: https://www.google.com/search?q=carey+r+dunne+wiki&oq=carey+r+&aqs=chrome.0.69i59l3j69i57j0i512j69i60l3.1165j0j9&sourceid=chrome&ie=UTF-8

Checking their link, I can confirm that "Talk:Carey R. Dunne - Wikipedia" appears as Google's #2 result (and that the article page does not appear anywhere on the result page). Note that their search query includes the keyword "wiki" along with the article title, and dropping that keyword changes the results so that the Talk page result appears at #24 in the results, but still doesn't surface the article page. I thought our Talk pages weren't indexed at all, but that's clearly not true. Mathglot (talk) 22:32, 5 May 2022 (UTC)[reply]

@Mathglot Seems like Google is doing their best. That article is new, and is still unreviewed, if you patrol it it should become indexed and Google will likely give it precedence. — xaosflux Talk 23:07, 5 May 2022 (UTC)[reply]
Do we know (or can we guess) what property of the talk page makes it more attractive to Google than the article? I realise that search engines are requested not to index new articles, but I assume that applies to new (and old) talk pages too. Certes (talk) 23:58, 5 May 2022 (UTC)[reply]
Yes, that's what confused me, as well. Seems to me, any noindex metatag on a mainspace article should be automatically propagated to Talk. Maybe I should add a phab ticket about this? Mathglot (talk) 00:27, 6 May 2022 (UTC) Tracked in T307753. Mathglot (talk) 00:44, 6 May 2022 (UTC)[reply]
The Google hit is https://en.wikipedia.org/wiki/Draft_talk:Carey_R._Dunne which redirects to https://en.wikipedia.org/wiki/Talk:Carey_R._Dunne. But MediaWiki doesn't make normal url redirection like HTTP 301. Instead it has JavaScript which rewrites the url in browsers with JavaScript enabled, but doesn't reload the page afterwards. The content is loaded from the original url. You see the contents of https://en.wikipedia.org/wiki/Talk:Carey_R._Dunne but you are actually on https://en.wikipedia.org/wiki/Draft_talk:Carey_R._Dunne. I guess that's how Google treats it. PrimeHunter (talk) 01:29, 6 May 2022 (UTC)[reply]
I didn't completely follow your explanation about Google vs. Mediawiki, but let's start at the top with what we do know and control, namely, with which pages we tag as noindex and which ones we don't. Here's what we have currently (a 'yes' below means that the page code contains <meta name="robots" content="noindex,nofollow"/> on the page). Note the asymmetric distribution of "noindex" metatags among the four related articles:
Given this situation, when you say "The Google hit is https://en.wikipedia.org/wiki/Draft_talk:Carey_R._Dunn", how can that possibly be? Even a targeted site search query doesn't find the Draft page. So that theory (or at least, the first part of it, which I think I understood) can't be right. Mathglot (talk) 03:46, 6 May 2022 (UTC)[reply]
Why don't we add noindex,nofollow to Talk: pages, at least when the corresponding mainspace page has noindex,nofollow? Certes (talk) 11:56, 6 May 2022 (UTC)[reply]
Your links made with {{noredirect}} link to en.wikipedia.org/w/index.php. Our robots.txt at https://en.wikipedia.org/robots.txt says Disallow: /w/ so indexing is disallowed. robots.txt tells spiders like Googlebot that they shouldn't even visit the page to see whether it has noindex. At the English Wikipedia we can only control indexing for some links to en.wikipedia.org/wiki/. If you view https://en.wikipedia.org/wiki/Draft_talk:Carey_R._Dunne with JavaScript disabled in your browser then you see the contents of the redirect target https://en.wikipedia.org/wiki/Talk:Carey_R._Dunne but you still have the former url in the adress bar and there is no noindex. Normal Internet redirects use List of HTTP status codes#3xx redirection, typically HTTP 301, and don't require JavaScript. A MediaWiki redirect is not a "real" redirect. https://en.wikipedia.org/wiki/Draft_talk:Carey_R._Dunne gives HTTP 200. It can be checked with an external tool like https://www.redirect-checker.org/, or with a feature in many browsers. PrimeHunter (talk) 16:08, 6 May 2022 (UTC)[reply]

Global vector.js

I used to put in vector.js the scripts I wanted on the desktop, but not on the mobile (is that the best way?). Now, I've decided to migrate it all to my global (meta.WM) profile, as I've started to contribute more to non-enwiki. I've tried the same 'trick' there, but to no avail. Is there a global equivalent to skin specific js files? How does one separate mobile vs desktop scripts, globally? Cheers. — Guarapiranga  11:59, 6 May 2022 (UTC)[reply]

@Guarapiranga you can't do exactly what you want. You centralwiki:User:global.(js|css) is a uniquely special page. You can read more about it here: mw:Help:Extension:GlobalCssJs. You can put conditionals or more specific class items in it (e.f. mw:Help:Extension:GlobalCssJs#Per-skin_customization). In general you can select mobile or not mobile best by applying things to the minerva skin. — xaosflux Talk 12:24, 6 May 2022 (UTC)[reply]
That clears it, @Xaosflux. Cheers. Specifically, FWIMC:

Currently the extension does not provide global CSS/JS for specific skins (mw:Help:Extension:GlobalCssJs#Per-skin_customization)

My solution then is to bracket everything with if (mw.config.get('skin') === 'vector') { ... } for now, and progressively pull out the stuff I want to see on the mobile too. Probably also a better solution than what I was doing for lang specific common.js. Thanks. — Guarapiranga  12:39, 6 May 2022 (UTC)[reply]
@Guarapiranga: Unless you change skin setting, you have three places to put user scripts, all are linked from Preferences → Appearance:
In each case, there is a corresponding .css page available. There is no provision for skin-specific pages shared by all wikis. Regarding your JavaScript code snippet, the technique for m:Special:MyPage/global.css would be different - each skin-specific rule would, in general, need to have each of its selectors prefixed with .skin-minerva - note the trailing space. So a rule like
.foo, .bar { display: none; }
would become
.skin-minerva .foo, .skin-minerva .bar { display: none; }
HTH. --Redrose64 🌹 (talk) 13:56, 7 May 2022 (UTC)[reply]

Bot issue

I've caught an issue with a bot that may need somebody here to fix. It's not a "permanently deprecate the bot entirely" level of crisis or anything, but it does need some attention.

User:Jura1/pages with Wikidata items without any statements is a report page, automatically generated by User:ListeriaBot to keep track of a maintenance issue. It seems like a perfectly legitimate and useful list in principle — but the problem is that if one of the pages it picks up corresponds to a category, then it just lists the category like an article and causes that workpage to be filed in the category instead of just linking to the category. However, pages like that aren't supposed to be filed in articlespace categories at all, meaning that I have to remove or disable the categories — but then ListeriaBot will undo any changes I make to the list the next time it runs that job, and thus throw the page right back into categories it isn't supposed to be in again.

But ListeriaBot's primary owner hasn't edited Wikipedia since December 2021, so I don't know if leaving a talk page message for them would reach them or not.

So does anybody know if User:Magnus Manske is actually active and just not contributing here that much, or can anybody else go into ListeriaBot themselves, to ensure that if ListeriaBot has to output a category to a worklist for whatever reason, it does so as a disabled text link (e.g. leading colon or the {{cl}} template) so that the worklist doesn't get categorized as an article? Thanks. Bearcat (talk) 13:09, 6 May 2022 (UTC)[reply]

@Bearcat that user has edited globally as recently as yesterday - so yse, you should start by asking on the botop's talk page (perhaps post at User talk:ListeriaBot and leave a message on the botop talk page pointing to it). — xaosflux Talk 13:12, 6 May 2022 (UTC)[reply]
Okay, thanks, I didn't know how to check whether Magnus was still active on other projects or not. If he is, I'll just approach him directly. Thanks. Bearcat (talk) 13:15, 6 May 2022 (UTC)[reply]
FWIW to check if a user has edited recently, see their global contributions linked from the contribs page (example). Graham87 08:53, 7 May 2022 (UTC)[reply]

Module:Graph bundle interpolation curvature index (beta)

I've added the following question at Module talk:Graph, but since the edit notice there suggests posting it here (as that page is not watched by as many users), I'm copying it here too:

The default d3 curvature parameter for the bundle interpolation method, .85, can often be too "erratic" in some applications (as pointed out here, and seen there). The basis method is even more "erratic," as it's standard spline bolted at the extremities, corresponding to a bundle beta of 1, but d3 also accepts other betas. Is that an option with this module (and accompanying template)? I couldn't figure it out from perusing the code. Cheers. — Guarapiranga  11:15, 4 May 2022 (UTC)

— Preceding unsigned comment added by Guarapiranga (talkcontribs) 13:35, 6 May 2022 (UTC)[reply]

Code in software that WMF uses and is maintained by other parties is not changed by the WMF or any volunteer of WMF. This would be fixed by the Vega or D3 team. There is some work that exists to upgrade Vega to Vega 3, see phab:T223026, but no WMF team is responsible for Vega/Graph right now. Only an developer can answer the D3 question, so please lets keep this discussion to the issue at hand and not try to figure out some fixes in Vega/D3 codebase. For those that are curious, Module:Graph uses Vega, which in turn uses D3.--Snævar (talk) 16:41, 6 May 2022 (UTC)[reply]
What's to fix, Snævar? I just asked whether Module:Graph had the beta parameter already in d3 (not sure about Vega--couldn't find anything about it in its docs). If not, I aimed at inspiring someone to add it in (I looked through the code, but it wasn't evident to me where it would fit in). — Guarapiranga  04:42, 7 May 2022 (UTC)[reply]

Sandbox page in categories

User:Skyblueshaun/sandbox2 is being artificially transcluded via template into three articlespace categories; however, as WP:USERNOCAT strictly prohibits this, I have to disable or remove the categories from the page, but I can't figure out what templates the categories are coming from: none of the navigational boxes seem to be transcluding any categories that I can see, so I tried wrapping {{Infobox football club season}} in {{Suppress categories}} on the theory that the categories were coming from there, but that didn't work either. Can somebody determine where the categories are coming from so that the page can be removed from them somehow? Thanks. Bearcat (talk) 14:12, 6 May 2022 (UTC)[reply]

Fixed in Special:Diff/1086498714 * Pppery * it has begun... 14:17, 6 May 2022 (UTC)[reply]
Probably from Module:Sports table. Gonnym (talk) 14:17, 6 May 2022 (UTC)[reply]
Actually it wasn't from that module, but from a transclsuion of the entire article 2022–23 EFL League One. The approach I used to find the categories was to run {{User:Skyblueshaun/sandbox2}} through Special:ExpandTemplates and the Ctrl-F for "Category:" * Pppery * it has begun... 14:19, 6 May 2022 (UTC)[reply]
Thanks. Bearcat (talk) 14:52, 6 May 2022 (UTC)[reply]

Unfortunately, now I've got another case: User:Ɱ/sandbox and User:Ɱ/sandbox29 are both displaying in Category:Columbus Register properties. I tried using the tool mentioned above to solve this myself, but both pages are such utter profusions of templates upon templates that I just can't make heads or tails of the results in order to know where to throw a {{Suppress categories}} wrapper. Help? Bearcat (talk) 15:31, 6 May 2022 (UTC)[reply]

 Fixed in Special:Diff/1086509364 * Pppery * it has begun... 15:33, 6 May 2022 (UTC)[reply]

And now User:Mdale/wikitrust.js, which is showing up in Category:Computer graphics — but since it's obviously a user's javascript settings file rather than a sandbox, I don't know how to fix it at all since it's not a conventional category declaration and the page is protected at a usergroup level ("interface administrator") that I don't think I have, but the page still can't be left in an articlespace category. Bearcat (talk) 15:39, 6 May 2022 (UTC)[reply]

User talk:Mdale/wikitrust.js#Interface-protected edit request on 6 May 2022. For what it's worth you indeed don't have interface administrator access. Thank you for doing this cleanup, by the way. * Pppery * it has begun... 15:45, 6 May 2022 (UTC)[reply]

Is the clock right?

Hey, Village Pump Tech folks,

I have a clock in the upper-righthand corner of my browser window and lately, the time has been off by, I'm guessing, 10-20 seconds. For example, if the clock states that it is 10:49:40 UTC and I make an edit, the edit summary on the page states that it occurred at 10:50. I've only noticed this happening recently. I checked in my Preferences and I think the clock is Gadget-UTCLiveClock.js so I've left a message at the MediaWiki talk page but I thought I'd mention it here in case anyone noticed that their clocks were a bit off. Also, the time moves at an inconsistent speed, some seconds pass by quicker than others for some reason. I really don't spend my time on Wikipedia watching the clock but some tasks are time-sensitive and the time an edit happens is important. Or could this just be my laptop? Thanks for any clue you can bring to my query. Liz Read! Talk! 02:07, 7 May 2022 (UTC)[reply]

I am curious which non bot tasks on Wikipedia are so time sensitive that 10-20 seconds will make a difference? ಮಲ್ನಾಡಾಚ್ ಕೊಂಕ್ಣೊ (talk) 05:31, 7 May 2022 (UTC)[reply]
There are a few possible factors here:
  • The gadget shows the time set on your local computer, which may be off. Wikimedia's servers most likely use Network Time Protocol to set accurate time.
  • If your edit takes 10 to 20 seconds after hitting "submit" to travel the Internet and be processed and saved, that could also make the difference.
  • As for the "some seconds pass quicker", again it's running on your laptop. Your browser may be delaying or skipping some of the updates of the display, particularly if other programs are using a lot of CPU. There's also the possibility that the stopped-clock illusion plays a part.
HTH. Anomie 11:22, 7 May 2022 (UTC)[reply]
Try going to https://time.gov/. NIST is the authoritative source for time in the US (and a major collaborator in the international pool of time reference sources). They're telling me that my system clock is off by -0.019s. No wonder I'm always showing up late for meetings :-) I'd be surprised if anybody's system clock was off by more than a fraction of a second. NTP is totally ubiquitous these days, and most operating systems come with it enabled out of the box.
You can call +1 303 499 7111 to hear the time signal on your phone. In the old days, this was an exceptionally accurate way to get the time. These days, I'm not so sure. Calls today often use VOIP, which introduces all sorts of network delays. The signal was certainly accurate when it left the transmitter in Ft. Collins, Colorado, but who knows what happens by the time it comes out of your cell phone speaker. If you have a shortwave receiver, you can listen to WWV (or similar stations around the world).
Sorry, I've digressed. Can you tell I'm a time freak? But seriously, go to time.gov to see if your system clock is set right. -- RoySmith (talk) 14:47, 7 May 2022 (UTC)[reply]

Way to access infobox and templates without "View this page" (V) button

Is there any way to access infobox and other templates without "View this page" (V) button without need to go to "Edit" then copy name of template and search Template: + name of template? Eurohunter (talk) 09:53, 7 May 2022 (UTC)[reply]

I get a list of all templates used in an article below the edit box. —Kusma (talk) 10:21, 7 May 2022 (UTC)[reply]

Help with quarry

I am trying to create a query that will get a list of IP talkpages that; have not been edited in the last 5 years, the IP is not currently blocked and there have been no edits from that IP in the last 5 years. I have individual queries to get list of IP talkpages, check if an IP is blocked, check if an IP has edited after a timestamp and check whether a page has received edits after a timestamp. I am tried to combine all 4 of these to work together at quarry:query/64344 but not having any luck. ಮಲ್ನಾಡಾಚ್ ಕೊಂಕ್ಣೊ (talk) 11:20, 7 May 2022 (UTC)[reply]

@ಮಲ್ನಾಡಾಚ್ ಕೊಂಕ್ಣೊ: WT:Quarry is a good place to ask, but I've had a go in quarry:query/64387. I've joined to the current revision rather than having a subquery check for all recent ones, and used the revision_userindex alternative view for efficiency when checking for edits to other pages. It doesn't detect range blocks; that would naively need page_title between ipb_range_start and ipb_range_end, except that doesn't work: 1.2.100.0 is textually between 1.2.0.0 and 1.2.3.255 but unaffected by a block on that range. You'd need to do some intricate string manipulation to pad with zeroes. This won't be a quick query – we have a lot of IP editors – and may need splitting with and page_title like "1%" etc. (Beware that most IPv6s begin with 20.) Certes (talk) 13:47, 7 May 2022 (UTC)[reply]

JS equivalent of FULLPAGENAME

Were I on, say, the Portuguese Wikipedia, what MediaWiki Javascript function can I use to parse an arbitrary string (e.g ':uSuÁrIo : example') and get the same result as {{FULLPAGENAME:uSuÁrIo : example}} (Usuário(a):Example)?

I'm aware of mw.Title but it requires a second argument to specify namespace. I need a high-performant one so action=parse is not an option either. NguoiDungKhongDinhDanh 15:03, 7 May 2022 (UTC)[reply]

mw.Title.newFromText(':uSuÁrIo : example').toText() – SD0001 (talk) 15:54, 7 May 2022 (UTC)[reply]
Thanks! NguoiDungKhongDinhDanh 16:19, 7 May 2022 (UTC)[reply]

Importing Wikipedia dump on Windows

Hello, I would like to conduct some research on Wikipedia, but appeared that there is no documentation on how to import it into MySQL, at least for Windows case. I'm stuck with converting XML-file into SQL-file. Could somebody help me out with this? --Igor Yalovecky (talk) 16:30, 7 May 2022 (UTC)[reply]