Wikipedia:Village pump (technical)

 Policy Technical Proposals Idea lab 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).

Newcomers to the technical village pump are encouraged to read these guidelines prior to posting here. 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.

Frequently asked questions (FAQ) (see also: Wikipedia:FAQ/Technical)
Click "[show]" next to each point to see more details.
If something looks wrong, purge the server's cache, then bypass your browser's cache.
This tends to solve most issues, including improper display of images, user-preferences not loading, and old versions of pages being shown.
Font size changed unexpectedly?
You may have accidentally changed the font size on your browser for a particular website by pressing a shortcut key or scrollwheel without realising it. Try resetting the zoom with Ctrl+0 (typing the digit zero while holding down the control key) or adjusting the zoom with Ctrl++ or Ctrl+-. Alternatively, look for the View option on your browser's menu and reset it to 100%.
No, we will not use JavaScript to set focus on the search box.
This would interfere with usability, accessibility, keyboard navigation and standard forms. See bug 1864. There is an accesskey property on it (default to accesskey="f" in English), and for logged in users there is a gadget available in your preferences.
No, we will not add a spell-checker, or spell-checking bot.
You can use a web browser such as Firefox, which has a spell checker.
If you have problems making your fancy signature work, check Wikipedia:How to fix your signature.
If you changed to another skin and cannot change back, use this link.
Alternatively, you can press Tab until the "Save" button is highlighted, and press Enter. Using Mozilla Firefox also seems to solve the problem.
If an image thumbnail is not showing, try purging its image description page.
If the image is from Wikimedia Commons, you might have to purge there too. If it doesn't work, try again before doing anything else. Some ad blockers, proxies, or firewalls block URLs containing /ad/ or ending in common executable suffixes. This can cause some images or articles to not appear.
Numbers listed in parentheses in the "Recent changes" section, on history pages and in your watchlist are the number of added or removed bytes.
For server or network status, please see Wikimedia Metrics.
« Archives, 152, 153, 154, 155, 156, 157, 158, 159, 160, 161, 162, 163, 164, 165, 166, 167, 168, 169, 170, 171, 172

Contents


Adding weblinks with special characters in source text or[]

Hello everybody! How come the source text or (apparently in all the Wikis!) is still not able to convert URLs of Wiki sites into valid Wiki links, as soon as the former include special characters (for instance: https://en.wikipedia.org/wiki/%C3%9Fß)? When clicking OOjs UI icon link-ltr.svg, entering the URL, and selecting Insert link, it says: "The URL you specified looks like it was intended as a link to another wiki page. Do you want to make it an internal link?" However, upon selecting Internal link, the following error message occurs: "The requested page title contains invalid characters". Has this issue already been dealt with in the past? If yes, why hasn't it been fixed by now? (If no – since I am not at all familiar with reporting errors to Phabricator –, would somebody competent be willing to bring it up there?) Thank you in advance for any assistance. Best wishes--Hildeoc (talk) 11:36, 26 January 2019 (UTC)

When making internal wikilinks, I don't copypaste URLs. I go to the top of the page, and copy the page title, and paste that in between the double square brackets. --Redrose64 🌹 (talk) 15:47, 26 January 2019 (UTC)
@Redrose64: Thank you very much for commenting. Of course you're definitely on the safe side by proceeding that way. However, I wondered whether it couln't be possible to improve that OOjs UI icon link-ltr.svg function as to the above issue, since it's been included in the or toolbar for quite some time now.--Hildeoc (talk) 16:11, 26 January 2019 (UTC)
Hildeoc, it does for me.. Which browser are you using ? —TheDJ (talkcontribs) 12:17, 29 January 2019 (UTC)
@TheDJ: Thanks for your interest! I use Firefox Quantum. Best wishes,--Hildeoc (talk) 20:23, 29 January 2019 (UTC)

PS: @Redrose64 and TheDJ:  Note: Unfortunately, the software is apparently also still not even intelligent enough to convert URLS to Wiki sites without special characters, but only consisting of more than one term – and thus including underscores as delimiters – into regular Wiki links without underscores ... Has this ever been discussed yet?--Hildeoc (talk) 17:52, 4 February 2019 (UTC)


PPS: @Redrose64 and TheDJ:  Note: Also, it cannot identify redirects, and label them as such – as this happens, in contrast, with dabs. Same question here: Has the community already come across that issue as well?--Hildeoc (talk) 21:47, 4 February 2019 (UTC)

VisualEditor (visual and wikitext modes) can do that, and I'm sure that it wouldn't have been built to do that if ors hadn't been wanting those features. I don't expect them to be added to the 2010 mw:WikiEditor, assuming that's the mw:or you're using. Whatamidoing (WMF) (talk) 20:08, 8 February 2019 (UTC)
@Hildeoc: Tried to figure out what you did, and why you get the result you do. You say "selecting Insert link" so I assume you talking about the dialog in the 2010 WikiEditor. That dialog is not smart enough, and I don't know if anyone will put more work into it, making it work in the proposed way. The VisualEditor and WikitextEditor do recognize such links, and in VisualEditor you can simply paste the link straight into the or and it will show up as a formatted link.
Your idea is sound though, and similar functionality should work the same way whatever or you use. Jeblad (talk) 20:50, 8 February 2019 (UTC)
@Jeblad: Thank you very much for commenting! Yes, precisely, I mean the 2010 WikiEditor (which is where I get by clicking on "Edit source" here). Just for reasons of clarity: What exactly are you referring to by "Wikitext or" then?--Hildeoc (talk) 20:35, 13 February 2019 (UTC)
@Hildeoc: WikitextEditor is also referred to as mw:2017 wikitext or. It will recognize links in some dialogs, but it won't recognize them pasted into the text like VisualEditor. Jeblad (talk) 21:18, 13 February 2019 (UTC)

New gadget that adds "Read Now" buttons to book references[]

Hey guys, I have been working with Internet Archive to produce a gadget that adds Read Now links next to book references. This allows readers to quickly click through to the books without having to find a library that has it. For the missing books, users are presented with a Donate button should they wish to donate what's needed to make the book available online. I'm currently asking the community two questions. —CYBERPOWER (Chat) 18:04, 29 January 2019 (UTC)

Comment: People can try the gadget first right now. Just add importScript('User:Cyberpower678/IABooks.js'); to your common.js page as explained at Wikipedia:User_scripts#Full_manual_instructions. The working demo code is at User:Cyberpower678/IABooks.js. Thanks! Ocaasi (WMF) (talk) 04:31, 30 January 2019 (UTC)

Question 1: Given the potential usefulness to readers, should this be made into a gadget that can be enabled in preferences?

Question 2: Should the donate button have an opt-out method, ie should the gadget allow for the ability to suppress the donate buttons?

Users can test this gadget by adding importScript('User:Reinischmax/IABooks.js'); to their JS file.

Also I'm pinging Fuzheado and OcaasiCYBERPOWER (Chat) 18:04, 29 January 2019 (UTC)

Input (New Gadget Proposal)[]

Perhaps worth noting that the Internet Archive has recently (two weeks ago) been served cease-and-desists notices from the Authors Guild (U.S.) and Society of Authors (U.K.) over this service (report). Not sure how far we should be promoting it, particularly not with quite so prominent bright green high-visibility blocks of colour. Jheald (talk) 15:54, 31 January 2019 (UTC)
When it comes to copyright, Wikimedia and the Internet Archive are really miles apart. The latter is often pushing the border to make it fairer for individuals to consume content (a position with which I strongly agree). But the Wikimedia foundation is extremely careful not to violate any copyright law. I don't think WMF council would be comfortable with links to sources that are legally challenged. - kosboot (talk) 14:22, 1 February 2019 (UTC)
Wouldn't get too carried away about Authors Guild lawsuits, they have a habit of loosing. See Authors Guild, Inc. v. Google, Inc. -- the Court found and upheld on appeal that "display of snippets from those works are non-infringing fair use". And in Authors Guild, Inc. v. HathiTrust, "the court concludes that all uses of the copyrighted works by the HDL are fair use." Also it appears the lawsuit concerns lending complete digital copies to library patrons, while this service concerns a "limited preview" akin to what Google Books and Haithi do with snippets. -- GreenC 15:57, 5 February 2019 (UTC)
On that, see https://www.techdirt.com/articles/20190128/17205941481/authors-guild-attacks-libraries-lending-digital-books.shtml Nemo 17:01, 7 February 2019 (UTC)

Proposed interim gadget : mobileUndo[]

Its been around two months since I put together the mobileUndo script (a script which allows reversion while using the mobile interface), currently around 25 people use it. I think it should be added as a gadget until the WMF implements this feature in its Advanced mobile contributions roll out. I feel that users who use the default mobile interface to Wikipedia on mobile (or for that matter anybody who uses the MinervaNeue skin) will benefit from using this script.  << FR (mobileUndo) 08:24, 2 February 2019 (UTC)

For those who want to test the script, the installation instructions are here << FR (mobileUndo) 03:39, 6 February 2019 (UTC)

Sticky headers scramble table when sorting[]

RSP Sticky Header Sort Scramble.jpg

Using FFox 63.0.1 on Win10, if I enable the "Make sure that headers of tables remain in view ..." (sticky headers) option, when I click on any of the columns at WP:RSP#Sources to sort it, the legend text is super-imposed over the headers and data rows.

If I remove the class="sortbottom" from the legend rows, the problem goes away (though the legend rows are then included in the sort). —[AlanM1(talk)]— 01:45, 8 February 2019 (UTC)

That's fun. I see the same in Fx 64 on W10. BRB, will see what happens in 65. --Izno (talk) 02:17, 8 February 2019 (UTC)
Yup, still in 65. --Izno (talk) 02:19, 8 February 2019 (UTC)

I'll note, too, that the headers are missing the vertical borders between columns as well as the rightmost border (the leftmost one is there, though). The bottom border is also missing. —[AlanM1(talk)]— 01:56, 12 February 2019 (UTC)

Yes, borders getting clobbered is a known issue. That's a Firefox problem, not a script problem. --Izno (talk) 02:22, 12 February 2019 (UTC)

File or article[]

Resolved: Already speedy deleted. — xaosflux Talk 15:42, 9 February 2019 (UTC)

Just found File:Qari_Waheed_Zafar_Qasmi.jpg something isn't right, can someone sort it out? Thanks GrahamHardy (talk) 11:23, 9 February 2019 (UTC)

The 'page' is already at MfD, nothing more to do. It's just an article in File namespace where description (of existing file) ought to go. –Ammarpad (talk) 13:34, 9 February 2019 (UTC)

An alignment issue[]

I have recently tried to replace the center tag with that of div. It landed up in an issue related to alignment. Please check it in the page Template:Wikipedia:Cleanup/Header so that the center tag can be removed.Adithyak1997 (talk) 19:03, 9 February 2019 (UTC)

I see recent s by Adithyak1997 (talk · contribs), Nardog (talk · contribs) and Izno (talk · contribs). Is there anything that still needs doing? If so, please be specific, since "an issue related to alignment" is rather too vague. --Redrose64 🌹 (talk) 22:52, 9 February 2019 (UTC)
@Redrose64:, When I posted this problem, the first section of the template was left aligned which I was not able to make it Centre aligned. But after posting it, Nardog (talk · contribs) as well as Izno (talk · contribs) solved that issue. Currently, there is no problem existing. Thanks for replying.Adithyak1997 (talk) 03:23, 10 February 2019 (UTC)

Inability to render ê in "ekhous apiston tênd anarkhian polei" in 3 March 2006 revision[]

Resolved: It's just a Unicode control character

If you click source, you can see there's supposed to be a character in between the "t" and the "nd", but it's not rendering properly. What could be the cause of that?

(Part of the reason I'm curious is, I'm using a python script to display the contents of an XML file containing that revision to the screen as it's processing it, and it always freezes up, to the point I can't even use ctrl-c to exit out but have to close the entire terminal window, when it gets to that character, so I'm wondering what's the deal with that character, anyway.) MW131tester (talk) 00:35, 10 February 2019 (UTC)

@MW131tester: I'm not seeing that in the source? What line is it on (what is a non-accented phrase near it maybe)? — xaosflux Talk 00:45, 10 February 2019 (UTC)
( conflict) The character between 't' and 'n' is Unicode character 'DEVICE CONTROL STRING' (U+0090) (UTF-8 two bytes: C2 90). See here.
Trappist the monk (talk) 00:47, 10 February 2019 (UTC)
User:Xaosflux, here's a screenshot. User:Trappist the monk, thanks for that resource. MW131tester (talk) 00:50, 10 February 2019 (UTC)
@MW131tester: ok so it seems to be garbage in - garbage out, and your script doesn't like garbage. Doesn't seem like there is anything else for us to do here? — xaosflux Talk 00:53, 10 February 2019 (UTC)
Oh, interesting, even now it allows us to put a  in there, but it won't appear in the diff or when the page is rendered. Okay, I learned something new. Thanks. MW131tester (talk) 01:05, 10 February 2019 (UTC)
The character is in the html of both the diff and the rendered page. It is your browser which does not display anything for it. Neither does mine but I guess some browsers or tools will display a placeholder symbol. My Firefox only does it in the area where I see the Unicode number 0090 in a square. PrimeHunter (talk) 01:34, 10 February 2019 (UTC)
Oh yeah, I see it when I do ctrl-u (page source). MW131tester (talk) 12:28, 10 February 2019 (UTC)

Resolving issues[]

Please check the question posted in the page Category talk:Unknown parameters. That page was viewed only by 2 persons (one is myself) after 2017 which forced me to post this message here.Adithyak1997 (talk) 08:32, 10 February 2019 (UTC)

@Adithyak1997: Most category pages have few watchers, in this case there are five. This is why you are presented with this message when ing that page. --Redrose64 🌹 (talk) 17:41, 10 February 2019 (UTC)
@Redrose64: Sorry. I didn't notice that.Adithyak1997 (talk) 18:37, 10 February 2019 (UTC)

CSS error[]

I get an error message about "persondata" (deprecated, I know) on my User:Randykitty/monobook.css. I must admit my ignorance here and am loath to make changes/delete lines without really understanding what I am doing. Perhaps someone who understands these things could have a look at this page and tell me what needs to be changed, I'd really appreciate that. Thanks! --Randykitty (talk) 09:23, 10 February 2019 (UTC)

I think you just need to remove the (top) table.persondata{...} line. —[AlanM1(talk)]— 12:01, 10 February 2019 (UTC)
Indeed, just remove the top line. --Izno (talk) 13:49, 10 February 2019 (UTC)

screen: tag and wiki-tag buttons[]

My screen has the usefull wiki markup buttons below to add tags & wiki-tags (like <sub></sub> and [[]]). Is there a topical page for these? Over there, I would like to ask:

  1. What is the logic behind the order of the options? To me it looks random chaotic. I cannot locate the right button easily.
  2. Why is <small></small> missing?
  3. Could we add option <code><nowiki></nowiki></code>?

-DePiep (talk) 15:13, 10 February 2019 (UTC)

It's made by MediaWiki:Gadget-charinsert-core.js with discussion at MediaWiki talk:Edittools. I suggested <code><nowiki></nowiki></code> at MediaWiki talk:Edittools/Archive 8#code and nowiki in one step, more discussion at MediaWiki talk:Edittools/Archive 9#Individual customization? PrimeHunter (talk) 15:26, 10 February 2019 (UTC)
At Help:Edit toolbar ...the size option looks like OOjs UI icon smaller-ltr.svg .--Moxy (talk) 15:31, 10 February 2019 (UTC)
Wikidata has option #3 [1]. -DePiep (talk) 16:11, 10 February 2019 (UTC)

Notes not displaying correctly?[]

I recently noticed in two different browsers some notes displaying strangely. For example, when hovering over the first annotation ('a') on Alexander the Great, all that displays is "^[244]". Also noticed this with the infobox notes on Republic of Macedonia. S.A. Julio (talk) 15:28, 10 February 2019 (UTC)

This is because, the notes also have citations. If you hover over the "[244]" you'll then see the full citation. If it's the anotation that you want to see, then you should click the [a] to take you there. –Ammarpad (talk) 16:00, 10 February 2019 (UTC)
@Ammarpad: I don't think that is the intention though? For example, when hovering over annotation 'd' on Alexander the Great, the text "^Ἀλέξανδροςἀλέξωaléxōlit. 'ward off, avert, defend'[251][252]ἀνδρ-andr-ἀνήρanḗrlit. 'man'[253][252][254]" is displayed, when the full note states "The name Ἀλέξανδρος derives from the Greek verb ἀλέξω (aléxō, lit. 'ward off, avert, defend')[251][252] and ἀνδρ- (andr-), the stem of ἀνήρ (anḗr, lit. 'man'),[253][252] and means "protector of men".[254]". Surely the full note displayed while hovering previously? S.A. Julio (talk) 16:21, 10 February 2019 (UTC)
You're right, apparently {{cnote2}} doesn't work properly.
Check Template:Cnote2#Cnote2, last example: text[A] text[B] text[A] text[B]), scroll page to hide the "^ text text text text" lines, and hover over A - no text is displayed, only "^" character - it should display "text text text text". --MarMi wiki (talk) 16:43, 10 February 2019 (UTC)
Similar problem is with Template:Cref2 - check example at Template:Cref2#Multiple_occurrences - it works, but displays as "^ab..." or "^...". --MarMi wiki (talk) 16:59, 10 February 2019 (UTC)
There was a technical note a few weeks ago about a change in the behavior of note marks. Possibly related. – Jonesey95 (talk) 19:36, 10 February 2019 (UTC)
Tech/News/2018/50 and/or Tech/News/2018/49. --MarMi wiki (talk) 21:55, 10 February 2019 (UTC)
S.A. Julio, last week I updated the gadget to generate those previews, based on a request by Jack who built the house. Perhaps this is a side effect of that. —TheDJ (talkcontribs) 14:16, 11 February 2019 (UTC)
Thanks for the report! I fixed {{cnote2}} already, but there are more templates that try to emulate the reference mechanism but don't supply classes which the references usually go with (mw-cite-backlink, reference-text). There are two ways: 1) Add classes to these templates. 2) Update the gadget so that it would support even poorly formatted references.
Although I prepared the solution for the second option, I would suggest the first, as it would eliminate some other issues like hanging non-breaking space in the beginning of the reference text in a tooltip.
The mentioned templates are {{cnote}}, {{note}}, {{note label}}. They are all protected. I prepared versions of them with added classes in my user space: User:Jack who built the house/Template:Cnote, User:Jack who built the house/Template:Note, User:Jack who built the house/Template:Note label. @TheDJ: could you help with updating? Or, we could go with the second option and update the gadget instead. Jack who built the house (talk) 19:40, 11 February 2019 (UTC)

Sister project inline links[]

Hello, as can be seen from this version, there is apparently a glitch as to the relevant sister project inline link templates, causing that the transclusion of two or more of these in direct succession without using bullet points leads to them all being displayed in one and the same line. Does anybody have an idea how to correspondingly fix those templates (thus, regardless of their concrete transclusion)?--Hildeoc (talk) 20:38, 10 February 2019 (UTC)

I don't think that there is anything wrong with those templates. They are doing pretty much what they are advertised to do. If you want a list without bullet points wrap them in {{plain list}}:
{{plain list|
*{{wiktionary-inline}}
*{{wikiquote-inline}}
*{{Commonscat-inline|Phenomena}}}}
which gives this:
Trappist the monk (talk) 20:49, 10 February 2019 (UTC)
@Hildeoc: Yes, this is as expected (indeed, the template names contain "-inline" for a good reason) and is pretty much the same as the behaviour of plain text. If your page contains this:
The dictionary definition of phenomenon at Wiktionary
Quotations related to Phenomenon at Wikiquote
Media related to Phenomena at Wikimedia Commons
then it displays like this:

The dictionary definition of phenomenon at Wiktionary Quotations related to Phenomenon at Wikiquote Media related to Phenomena at Wikimedia Commons

It's covered by WP:LINEBREAK. --Redrose64 🌹 (talk) 00:09, 11 February 2019 (UTC)

I wonder why two out of these three tell you to use bullet points in the template doc but one doesn't. – Finnusertop (talkcontribs) 03:47, 12 February 2019 (UTC)

Expand language template problem[]

Hi I noticed that when someone add this template {{Expand language|langcode=ar}} this category appears on the top of the article and it is visible [[Category:{{{topic}}} articles needing translation from Arabic Wikipedia]]. Here is an example [2]. Can this problem be solved? Thanks--SharabSalam (talk) 21:52, 10 February 2019 (UTC)

{{Expand language}} should not be called directly but it should behave better if it's done anyway. I have changed it to accept a missing topic parameter.[3] PrimeHunter (talk) 23:01, 10 February 2019 (UTC)

Vojvodina[]

Hi there, I tried to the template in question to display the country parameter correctly by adding | country = Serbia, as before it displayed {{{country}}} and this was breaking the syntax. But I actually want the title to say "Largest cities in Vojvodina", not all of Serbia. Any help? 2607:FEA8:1DE0:7B4:5C5A:EACB:2CF0:A463 (talk) 01:00, 11 February 2019 (UTC)

It doesn't actually have to be a country. I simply said country = Vojvodina.[4] I also removed the name parameter which is only for template pages. and a province parameter which is ignored. PrimeHunter (talk) 01:55, 11 February 2019 (UTC)

All images broken?[]

Whenever I check a page on Wikipedia - or, for that matter, on any Wikimedia site, such as the Commons - the images do not display. Just a broken link. I've tried using multiple different browsers and no fewer than three physical devices (a laptop, a desktop, and a cellphone), all of which appear to have the same issue.

I'm not sure if this is the right place to post this or not, but is this happening to anyone else? Gimubrc (talk) 17:07, 11 February 2019 (UTC)

Happening to me also103.215.54.84 (talk) 17:08, 11 February 2019 (UTC)
Me too. Seems to be working now. Adam9007 (talk) 17:09, 11 February 2019 (UTC)
Still out of order for me... I erased my browser's cache and cookies (everything over the past hour) and it fixed the problem. If anyone else is having issues, that should fix it.Gimubrc (talk) 17:13, 11 February 2019 (UTC)
Caused by a configuration error, see https://etherpad.wikimedia.org/p/2019-02-11-upload-cache-failure for preliminary details with a full analysis coming soon. —AntiCompositeNumber (talk) 17:43, 11 February 2019 (UTC)
But, it's not Thursday. It's only Monday :(. ~ ONUnicorn(Talk|Contribs)problem solving 17:51, 11 February 2019 (UTC)
"Config changes don't ride the train." And if that opaque statement doesn't mean more to you than it did to me, when I first heard it, then you may be interested in seeing the whole calendar at wikitech:Deployments. There's a chance for something to go wrong most days of the week. Whatamidoing (WMF) (talk) 20:16, 11 February 2019 (UTC)

There is now an incident report on this at wikitech:Incident documentation/20190211-varnish. the wub "?!" 21:47, 12 February 2019 (UTC)

Tech News: 2019-07[]

18:45, 11 February 2019 (UTC)

Navigation popups[]

On Opera, Google and Firefox (other browsers not tested) the disambiguation facility of WP:NAVPOPS seems to have disappeared, I now only get the display options, not the ability to use the tool to disambiguate links. Nthep (talk) 19:18, 11 February 2019 (UTC)

Failing "request an account" process[]

The process at these places is failing.

A notice at the first link says that the backlog is 2 months. The ideal turnaround is instant.

For those who need an orientation to the "request an account" service, it is useful for users who want to from a blocked IP address. Commonly these are shared IP addresses, such as at schools, community centers, parks, libraries, or large commercial buildings with shared wifi like some business complexes or malls. To support these people we can either coach them through removing an IP block - which usually is not possible or useful - or we can somehow get them an account.

The request an account process is supposed to be higher security than registering a Wikimedia account in the usual way. Currently it requires an email address. If we wanted more assurances then we could set up a registration process which asked for more.

I think this process needs to become automated. There are not many human volunteers who want to do this process, the talk page there is dead, this must be boring to review, and I am not able to see in the guide at Wikipedia:Request_an_account/Guide that there review aspects which a human has to do and automation cannot do.

Thoughts from others? Blue Rasberry (talk) 12:57, 12 February 2019 (UTC)

It's not easy to automate validation that requested usernames meet the username policy. It's not easy to automatically parse any comments from users who are requesting accounts. Most of the checks are automatable in a pass/fail manner, and we do have work in progress to improve that, currently blocked on a code review. We also have work in progress to do creations via OAuth, which should also increase the speed we can reach (again, blocked by code review). The truth of the matter is that probably 50-75% of requests we get will fail one of the automated checks, mostly blocks, recent contributions, multiple requests from one email or IP address - all of these need manual review. WP:ACC is meant to be the human review after the automated filters on the main on-wiki signup form have rejected; for that reason I'm fairly strongly against adding automated rejection of requests without human validation to ACC. This isn't helped by our general lack of people able to spend time working through the backlog. stwalkerster (talk) 13:46, 12 February 2019 (UTC)
It's also worth me noting that the talk page is indeed fairly dead, because most of the discussions we have are centred around specific requests or general policies, and tend to happen either on IRC or the mailing list due to the nature of the topics they contain. stwalkerster (talk) 13:46, 12 February 2019 (UTC)
Stwalkerster, time permitting of course, I will see if I can knock away some of that backlog. —CYBERPOWER (Be my Valentine) 14:35, 12 February 2019 (UTC)
I also forgot - checking the global rename log is also not automatable at the moment, pending phab:T193671. stwalkerster (talk) 20:32, 12 February 2019 (UTC)
@Stwalkerster: Okay, you have a lot of insight on this. I agree, the talk page would not matter because discussion happens in semi-private with no easily accessible record. Can you answer these questions briefly, and make your best guesses if you do not know the answer? I am making some guesses to start.
  1. How long does it take a human to review one of these requests? 2 minutes?
  2. How many requests are there a month? 100?
  3. What is the community base for addressing this queue? 10 people?
  4. How long does it take a human to learn to do evaluations in this queue? 2 hours?
  5. Under what circumstances could this queue get backlogged further?
  6. Is there anything I can do to encourage prioritizing the code review? Like for example, posting a link to this conversation somewhere or to someone?
Thanks @Cyberpower678: <3 happy valentines Blue Rasberry (talk) 00:08, 13 February 2019 (UTC)
Bluerasberry, Point 1 - 2 mins for a really easy request, maybe 30 mins max for a more difficult request. I've had some that have spanned multiple days while I discuss the matter with the blocking admin.
The present backlog is appx 3000 requests, at 2 months. So, the answer to point 2 is 1500 at this time, but going up.
Going by the IRC logs, I'd say the answer to point 3 is... "not many" :( 5-10 perhaps. We've all got our specialties and interests, mine is mostly to deal with the proxy queue.
To point 4, yeah - about 2 hours reading, and a couple more hours hands-on sounds about right.
Point 5....More requests?
To point 6, apparently the code review is with FastLizard4 as of nearly 2 years ago. It's been a long time since I started ACC, and I'm not sure I understand why that this is a point of failure spanning multiple years. SQLQuery me! 00:38, 13 February 2019 (UTC)
Bluerasberry, new ACC users can take anywhere from 2-10 minutes to assess a request. Veteran users can work as quickly as 10 seconds. My fastest time on a request is about 20 seconds. —CYBERPOWER (Be my Valentine) 01:33, 13 February 2019 (UTC)
In my humble opinion, a great non-technical way to reduce the backlog would be to stop making the recommendation for ACC so prominent. You'll see this recommendation at Help:I have been blocked, Template:Schoolblock, Template:Anonblock, Template:Rangeblock and various other places. Once upon a time we used to basically advise the user to find another IP address. ACC is a useful service but let's be honest hardly anyone is stuck on a single IP or range and unable to create an account on their own. People go to ACC because they think it's the recommended method, but in most cases it isn't. Incidentally this is a principle I'd also apply to OTRS, UTRS, functionaries, and various other places where things often get needlessly kicked. -- zzuuzz (talk) 00:31, 13 February 2019 (UTC)
( conflict) Bluerasberry: responses inline:
How long does it take a human to review one of these requests? 2 minutes? It entirely depends on who is doing it, and also on the specific request. Newbies to ACC would probably take in the region of 3-4 minutes for a simple request, more experienced users will probably take less than a minute per request. Wrinkles like AntiSpoof hits, blocks, multiple requests from the same IP, s, talk page content, etc just adds time.
How many requests are there a month? 100? You're out by an order of magnitude. It's somewhere between 1000 on a quiet month to 2500 on a busy month, and recently it's been veering towards the upper end of that range - probably as more things have started pointing to ACC in that time.
What is the community base for addressing this queue? 10 people? There's more than 10 with access, but in practice there's been about 15 people handling requests in the last month
How long does it take a human to learn to do evaluations in this queue? 2 hours? That's hard to quantify - somewhere between 15 minutes of reading the guide and a few years of experience.
Under what circumstances could this queue get backlogged further? People not spending more time working on the queue? It's currently growing...
Is there anything I can do to encourage prioritizing the code review? Like for example, posting a link to this conversation somewhere or to someone? Not really unfortunately. Those who we need the sign-off from are already on it. stwalkerster (talk) 00:33, 13 February 2019 (UTC)

SQL query requests[]

Wikidata has a wonderful forum Request a query where volunteers create SPARQL queries by request. It has never failed me within a day or less and is super helpful. Is there something like this for SQL such as for Quarry queries? -- GreenC 18:21, 12 February 2019 (UTC)

Where you would expect, and I believe it enjoys an interwiki link with the WD page. Feel free to advertise elsewhere :^). --Izno (talk) 18:27, 12 February 2019 (UTC)
OH thanks! -- GreenC 18:30, 12 February 2019 (UTC)
Izno, TIL, thanks! Watchlisted! SQLQuery me! 00:26, 13 February 2019 (UTC)

Conditional column[]

I have two tables, each 20em wide, one following the other (vertically). If there is enough room on the user's display (i.e. if it is more than about 42em), I'd like to put the second table beside the first instead of below it. I tried {{Col-break}} etc., hoping that it might be smart enough to move the second column below, but it just leaves it there and the user has to scroll right. Is there some magic to do this?

—[AlanM1(talk)]— 18:50, 12 February 2019 (UTC)

If the second table doesn't have a section heading then you can use style="display: inline-table;" per Help:Table#Positioning:
Table 1
Foo Bar
Lorem ipsum dolor sit amet,
consectetur adipiscing
Table 2
Moo Boo
Duis aute irure dolor in
reprehenderit in voluptate
PrimeHunter (talk) 19:09, 12 February 2019 (UTC)
@PrimeHunter: Very nice. Works with divs, too, including section headers. Face-smile.svg —[AlanM1(talk)]— 20:12, 12 February 2019 (UTC)

Page Deletion: User:Steeve815[]

Resolved

Dear Wikipedia's administrators,
Please help me deleting my user page. I don't think I'll be usable.
Thanks, Steeve815 (talk) 19:27, 12 February 2019 (UTC)

 Done by User:Writ Keeper --DannyS712 (talk) 19:33, 12 February 2019 (UTC)

Collapse navbox for specific article[]

Is it possible to make a navbox appear in collapsed form on a specific article, even though there is only one navbox on the page and the navbox itself is set to autocollapse (i.e. when two or more on a page)? ClippednPinned (talk) 20:49, 13 February 2019 (UTC)

@ClippednPinned: Sure, its the state parameter. What article? --DannyS712 (talk) 20:56, 13 February 2019 (UTC)
You can set it to |state=collapsed collapsed. Ruslik_Zero 20:57, 13 February 2019 (UTC)
Thanks ClippednPinned (talk) 09:48, 14 February 2019 (UTC)

F-Secure blocks Wikidata[]

Since yesterday, F-Secure blocks Wikidata. At least for me. J 1982 (talk) 13:07, 14 February 2019 (UTC)

If it is just a false positive, the powers that be at Wikidata (or others who are confident there is no malware) can take action here, if they want. 24.151.50.175 (talk) 16:01, 14 February 2019 (UTC)
Two days ago, everything was fine. Yesterday, I could at least enter Wikidata from Google without any problems and only ing via " linka" under "languages". Today, I can't enter at all. J 1982 (talk) 16:51, 14 February 2019 (UTC)
I just went to
https://www.f-secure.com/en/web/labs_global/submit-a-sample#sample-url
and submitted
https://www.wikidata.org/wiki/Wikidata:Main_Page (I also tried https://www.wikidata.org/)
The result I got was
Dear Customer, Thank you for submitting a sample to us. Your report has been sent to our Response team to be analysed. If you have chosen to be notified of the analysis result, the result will be sent to your email in a short while. F-Secure Labs
I have zero problems entering Wikidata from Google search. --Guy Macon (talk) 18:32, 14 February 2019 (UTC)
I just tried Virustotal, and all 59 antivirus engines say that Wikidata is clean.
See https://www.virustotal.com/#/url/b52bfa6ef870cd9754d8409df6ea1b49b2b45ef9d45432d8a7d90d9d7e34abce/detection
--Guy Macon (talk) 18:36, 14 February 2019 (UTC)

Weird warning[]

Hi all, I was trying to add the following text to my user page under "Searches"

*"Virendra Singh Patyal" {{Search|"Virendra Singh Patyal"|long=y}} - Feb 2019 - See {{noping|Saims243}}, a sock of {{noping|Sscomputers201}}

(It helps me search for problematic stuff)

I got the following error message: Your was not saved because it contains a new external link to a site registered on Wikipedia's blacklist.

The text below said: The following link has triggered a protection filter: britannica.com/search?query

But I don't see any part of that link on my user page, so I'm unclear on what the error was about. Help? Thanks, Cyphoidbomb (talk) 17:26, 14 February 2019 (UTC)

@Cyphoidbomb: worked for me could you try again? — xaosflux Talk 18:01, 14 February 2019 (UTC)
I copied and pasted Xaosflux's text, and did so in the same spot. AFAIK, the only thing I did differently was the whole page instead of the section. ᛗᛁᛟᛚᚾᛁᚱPants Tell me all about it. 18:52, 14 February 2019 (UTC)
@Billinghurst:Do you think you could expound upon that answer in more detail, please? I'm a software developer myself, but I've always shied away from web development and administration, so I'm not really grasping what you're saying here. You can reply at my talk page to avoid going off topic here. ᛗᛁᛟᛚᚾᛁᚱPants Tell me all about it. 23:30, 14 February 2019 (UTC)

Labs is down[]

https://tools.wmflabs.org/ is down (500 - Internal Server Error). It's been down since about 03:00 UTC. Does anybody know what's going on? Thanks, — Diannaa 🍁 (talk) 05:55, 15 February 2019 (UTC)

There are various issues, including multiple hardware failures, impacting Toolforge, ToolsDB, and other Cloud VPS instances. See phab:T216208 and mailarchive:cloud/2019-February/000548.html as well as other related tasks and emails. Expect poor performance for at least the next 24 hours. — JJMC89(T·C) 06:32, 15 February 2019 (UTC)
Update: The issues will likely not be resolved until Tuesday at the earliest. — Diannaa 🍁 (talk) 17:36, 16 February 2019 (UTC)

Something out of whack with Template:Infobox U.S. Cabinet[]

{{Infobox U.S. Cabinet}} is employed on pages dealing with US presidents. At the pages I've checked—Franklin D. Roosevelt (where I first noticed it), George H. W. Bush, and Abraham Lincoln—names of cabinet members are displayed out of chronological order when the raw text entries are in order. In January, the problem was noted at Template_talk:Infobox_U.S._Cabinet#View_problems without a response so far. Can someone here shed light on the problem? Dhtwiki (talk) 19:29, 15 February 2019 (UTC)

The error is easy to see at Template:Infobox cabinet members/doc, where each set of names at the right is ordered a/c/e/b/d. I think this is related to the 21 September 2018 s to Module:Infobox cabinet members, so this might be one for Frietjes (talk · contribs). -- John of Reading (talk) 20:42, 15 February 2019 (UTC)
Please reply at the original thread, Template_talk:Infobox_U.S._Cabinet#View_problems. Thanks. – Jonesey95 (talk) 20:59, 15 February 2019 (UTC)
Fixed – This problem now seems fixed, at least with regard to my issue. The main discussion took place on the template talk page; the code fix was made here. Dhtwiki (talk) 22:41, 16 February 2019 (UTC)

cite error[]

I get a cite error in the Matthew Whitaker article that I can't understand. As far as I can see the ref in quesion is defined. Can someone please help out? Could there be some invisible character involved?

Cite error: The named reference vladek was invoked but never defined

Thanks in advance.

HandsomeFella (talk) 21:18, 15 February 2019 (UTC)

Fixed. Someone commented out the paragraph containing the full reference, so the self-closed reference didn't work. – Jonesey95 (talk) 21:29, 15 February 2019 (UTC)
Which is one of the reasons I put the full citations ("reference" is rather ambiguous) into their own section. Even if folks insist on using named-refs, having the master "references" in their own section removes a point of possible breakage. ♦ J. Johnson (JJ) (talk) 21:42, 15 February 2019 (UTC)

Bug: player overlaps music score display (css problem?)[]

View Wikipedia:Village pump (technical)#Bug: player overlaps music score display (css problem?) in the skin: 

{c' e' g' e' c' e' g' e' c' e' g' e'}

As you can see above, with default Vector skin (and most others) <score vorbis = "1"> shows the player overlapping several pixels over the score image. On en.m.wikipedia.org or with Modern skin everything shows correctly. Seems like a problem with css styles. Please help. --M5 (talk) 22:09, 15 February 2019 (UTC)

@M5: I've opened phab:T216305 for this issue to be worked on, it shouldn't require localization of css to not be broken, also fails on testwiki. — xaosflux Talk 02:13, 16 February 2019 (UTC)
I've added a set of skin links so that others may be checked easily. I find that in Modern it's fine; in Cologne Blue it's just acceptable; but in MonoBook, Timeless and Vector it has the undesirable overlap. Why do section links not work in MinervaNeue?
Anyway, part of the problem seems to be that the height of the player changes as it loads (it starts off as a dark strip, then changes to a spinny thing, then to a paler strip which is taller than the original dark strip), but whilst the containing box changes its height to compensate, the second change is not far enough. --Redrose64 🌹 (talk) 13:51, 16 February 2019 (UTC)

Why do user JS/CSS files always come back with max-age=0?[]

Every time a visit a page, everything I import with mw.loader.load() or importScript() comes back with:

Cache-Control: private, s-maxage=0, max-age=0, must-revalidate

This forces Firefox to wait for whole lot of pointless HTTP 304 responses before the scripts can run. Adding &maxage=86400&smaxage=86400 to the request doesn't work, either. It seems Varnish has been set up to strip it right back out.

Is this a bug, or a feature? I found phab:T71460, but it's been open since 2014, and I'm not following Krinkle's last comment, nor do I understand why it's gone stale. Gadgets are cached for 30 days, by comparison! I'd think the WMF folks would want to reduce server load if nothing else.

I can think of a few horrible hacks to get around this, but I'm wondering if there is actually a good reason why caching is prohibited. Suffusion of Yellow (talk) 23:01, 15 February 2019 (UTC)

@Suffusion of Yellow: depending on how you are watching, you may be getting extra 304 responses because you are watching: "Many developer tools' network panels of browsers create extraneous requests leading to 304 responses, so that access to the local cache is visible to developers." (https://developer.mozilla.org/en-US/docs/Web/HTTP/Status/304) — xaosflux Talk 23:06, 15 February 2019 (UTC)
@Xaosflux: Works fine on 127.0.0.1, where I'm accessing MediaWiki directly, with no Varnish. And adding &maxage=86400 to API requests also works on enwiki. It's just index.php where I get the 304s. My understanding is that with s-maxage=0, max-age=0, must-revalidate, the browser should be checking every time, developer tools or not. Can anyone else reproduce this? Suffusion of Yellow (talk) 23:17, 15 February 2019 (UTC)
Gadgets are loaded via the load.php endpoint, which has special caching rules applied to it. Stuff you're manually loading via importScript() hits the index.php?action=raw endpoint, which bypasses all varnish caching since you're logged-in. This isn't ideal, which is why Krinkle said "We should fix this...". Legoktm (talk) 07:33, 16 February 2019 (UTC)
@Legoktm: Now I'm even more confused. Perhaps I linked to the wrong phab ticket. I'm not asking why varnish doesn't cache this; I'm asking why my browser is being told it must-revalidate every user script after 0 seconds. If I could bypass varnish somehow, that would be good thing, because it's varnish that's doing this deliberately:
	if (req.url ~ "^/wiki/" || req.url ~ "^/w/index\.php" || req.url ~ "^/\?title=") {
			// ...but exempt CentralNotice banner special pages
			if (req.url !~ "^/(wiki/|(w/index\.php)?\?title=)Special:Banner") {
				set resp.http.Cache-Control = "private, s-maxage=0, max-age=0, must-revalidate";
			}
		}
I can get around this, by loading a script with a hack like mw.loader.load("https://en.wikipedia.org/w/index.php?title=Special:Banner&title=User:Enterprisey/script-installer.js&action=raw&ctype=text/javascript"), which gives me a sensible max-age=1209600, but I'm wondering if I should, or will this break something for me in some subtle way? Suffusion of Yellow (talk) 18:15, 16 February 2019 (UTC)
How do you expect MediaWiki to invalidate the cache then, when the page is updated? Legoktm (talk) 02:24, 17 February 2019 (UTC)
@Legoktm: I'm expecting that a script being slightly out of date is not going to be a big deal. I can always shift-reload to get the latest version, if I want. Gadgets are coming back with public, max-age=2592000, s-maxage=2592000, which is twice as long and doesn't seem to cause any problems. Suffusion of Yellow (talk) 02:36, 17 February 2019 (UTC)
The expected behavior is that user scripts update immediately. Gadgets go through ResourceLoader, which embeds versions in the URLs, so they can be cached. And ResourceLoader provides the versions in what's called the "startup" module which has a cache TTL of 5 minutes IIRC. Legoktm (talk) 04:13, 18 February 2019 (UTC)

Watchlist/mediawiki problem[]

I have just started experiencing a problem where updating my watchlist hangs while reading www.mediawiki.org. And when I stop it, I get the basic html format. This is at least partly a problem with my old Firefox browser, as it doesn't happen with my Chromium browser. But it didn't use to happen with Firefox. I haven't explicitly made any changes to my browser or system configuration, or WP personal preferences (but sometimes Strange Stuff Happens?). It started in the middle of a session (i.e., not upon logging in). I had just added some pages to my watchlist; removing a bunch of pages has not remedied the problem. Any ideas as to what is going on here? ♦ J. Johnson (JJ) (talk) 23:21, 15 February 2019 (UTC)

I'm having the same problem with Safari 12.0.3. — fourthords | =Λ= | 23:45, 15 February 2019 (UTC)
@J. Johnson and Fourthords: can you try the Safe Mode Watchlist? — xaosflux Talk 23:56, 15 February 2019 (UTC)
I still get a never-ending ellipsis undulation. — fourthords | =Λ= | 23:58, 15 February 2019 (UTC)
Likewise. A slight amendment of the problem statement: while Chromium displays a properly formatted watchlist, the little worm keeps chasing its tail, but while reading en.wikipedia.org (not mediawiki.org). Which is the same behavior I get using Safemode, except that Firefox does not display the watchlist until I stop the transfer, and then it does "basic". ♦ J. Johnson (JJ) (talk) 01:46, 16 February 2019 (UTC)
Well... The problem seems to have gone away. Don't know if it had anything do with having just rebooted the computer. (Don't usually have to do that, as I don't have run Windows.) ♦ J. Johnson (JJ) (talk) 03:48, 16 February 2019 (UTC)
This often means that the Wikimedia servers are slow to return a CSS or JavaScript file, so when your browser eventually times out from waiting, it displays what it can - which may well be unstyled as a consequence of the missing CSS/JS. --Redrose64 🌹 (talk) 13:34, 16 February 2019 (UTC)
And the problem is back. What you say makes some sense, especially this morning, as the local networking seems to be congested. But at least one additional factor is involved, as Chromium, which consistently formats the watchlist, is still waiting for something else, even after an hour. Also, I see the problem only with my watchlist; not with contributions, large articles, or anywhere else. And (aside from late last night, when I rebooted) Firefox is persistently affected, whereas if the servers were the main problem I would expect more variation in the course of the day. The problem occurrence was definitely not seen (at least, not for a long time), and now is persistently seen. The transition was around 22:00 (UTC) on 10 February. Were there any server changes around then? ♦ J. Johnson (JJ) (talk) 20:01, 16 February 2019 (UTC)
FWIW, I still have no watchlist functionality. Just a forever-pulsing elipsis. — fourthords | =Λ= | 23:34, 17 February 2019 (UTC)

Issues with infobox and sidebar[]

I have encountered issues with Template:Civil conflict sidebar and Template:Infobox civil conflict. Please see Montgomery bus boycott for examples of issues. The issue with the "Civil conflict sidebar" template is a missing white border along the right side of the box. The issue with the "Infobox civil conflict" template is located within the "Parties to the civil conflict" section of the template. The right column bullet points are the wrong color (suppose to be a blue-ish color) and have crossed into the left column. Mitchumch (talk) 07:08, 16 February 2019 (UTC)

Fixed. The combination of tables and div tags and bullet lists resulted in sensitivity to line breaks. I believe that I have fixed the "non-blue bullets" problem and a couple of other alignment problems in Template:Infobox civil conflict. I don't see any problems with the sidebar. – Jonesey95 (talk) 14:04, 16 February 2019 (UTC)
trout Self-trout (...and somehow in the process, I transcluded the sandbox version of the Infobox into Montgomery bus boycott. Sorry about that, and thanks to the ubiquitous Redrose64 for catching my mistake.) – Jonesey95 (talk) 16:37, 16 February 2019 (UTC)
Jonesey95 Thank you for fixing the "Infobox civil conflict" template. For the sidebar, there used to be a white border around the entire blue-ish colored box. Is there anything I can do to restore that attribute on the right side of the template? Mitchumch (talk) 19:55, 16 February 2019 (UTC)

Permission error?? Bug???[]

I'm an admin. When I went to restore a couple of deleted pages just now (an action I've performed a number of times during my centuries of adminship), User:Flooded with them hundreds/1.js and User:Flooded with them hundreds/common.js, I repeatedly got permission errors: "You do not have permission to view a page's deleted history, for the following reason: The action you have requested is limited to users in one of the groups: Administrators, Oversighters, Researchers, Checkusers." I was logged in to my normal, admin, account. (I know Bishzilla gets permission errors..) And when I went experimentally to view a user's deleted contributions, there was no problem; apparently I'm admin enough for that. What gives? Bug? Some special issue with those two pages? If somebody would like to help with the restoration of the pages, please give the reason "user request". But most of all I'd like to understand what went wrong, and to have it stop doing it. Bishonen | talk 10:02, 16 February 2019 (UTC).

Only interface admins can restore CSS and JS pages. The interface message is wrong. (T203083) — JJMC89(T·C) 10:06, 16 February 2019 (UTC)
Aha. Thank you. It would be great if someone corrected the message. Meanwhile, the user has requested undeletion of those pages, and there is no problem with it that I can see, so perhaps an interface admin would like to help? Bishonen | talk 10:10, 16 February 2019 (UTC).
( conflict)Pages restored - All admins are equal, but some are more equal than others. Because of phab:T202989, admins cannot see the deleted history of JS and CSS pages, and because of bugs with error messages, mediawiki assumes that you're not an admin if you can't see the deleted history (thus the error message that claims you're not an admin). Galobtter (pingó mió) 10:13, 16 February 2019 (UTC)
Thanks very much, Galobtter. Bishonen | talk 11:28, 16 February 2019 (UTC).

IABot ascribing s to wrong account[]

Why is InternetArchiveBot ascribing s like this to my old account, Noyster? When invoking the bot I am logged in using my present account, which is: Bhunacat10 (talk), 13:24, 16 February 2019 (UTC)

Bhunacat10, log out and log back in to the tool. —CYBERPOWER (Chat) 13:48, 16 February 2019 (UTC)
Thank you, but I'm simply making these s through Page history|Fix dead links and no instruction is met with about logging out & in so that one's present account name will be shown as responsible for the : Bhunacat10 (talk), 13:55, 16 February 2019 (UTC)
Also log in and out of wiki with your old account to destroy all sessions that could be linked to a cookie. — xaosflux Talk 15:30, 16 February 2019 (UTC)
Bhunacat10, when you initially use the tool, you were prompted to login first. Go to the tool, in the upper right you should see your name. Go there and click Logout on the drop down menu. —CYBERPOWER (Chat) 16:27, 16 February 2019 (UTC)
Bhunacat10, BTW, why didn't just get yourself renamed? —CYBERPOWER (Chat) 19:43, 16 February 2019 (UTC)

Twinkle[]

Using Twinkle on a mobile device (in desktop mode) has been causing a few problems lately, has anyone had this happen? For one thing, when trying to welcome an IP, the form popped up like normal, yet when I sent it, the Twinkle form went blank. When welcoming User:Guy3Fire (I know now that he is not new, my apologies) where I used the TW button on top, where it didn't want to click. Thank you for your help in advance! Sincerely, User: Zanygenius(talk page) 15:11, 16 February 2019 (UTC)

Is it possible to have a floating 'sidebar' that follows you up?[]

If you browse WP:CRAPWATCH/SETUP, it's a really really long list of stuff. If I'm halfway in the J2 section, and want to jump ahead to the W3 section, then it takes a pretty long time to scroll back up, find the TOC, and then click on W3. So I was thinking of having a small "Back to top" floating thingy that follows you as your scroll the page. Is this doable? Headbomb {t · c · p · b} 18:16, 16 February 2019 (UTC)

@Headbomb: have you tried just pressing your Home key, it should scroll you right back to the top of any page. — xaosflux Talk 18:55, 16 February 2019 (UTC)
You could also use a javascript such as Numbermaniac's User:Numbermaniac/goToTop.js. — xaosflux Talk 18:57, 16 February 2019 (UTC)
See also Wikipedia:Back to top. — xaosflux Talk 18:58, 16 February 2019 (UTC)
@Xaosflux: Facepalm Facepalm, I don't know why I didn't think of that. Thanks, that's workable. Headbomb {t · c · p · b} 19:01, 16 February 2019 (UTC)
@Headbomb: Have a look at {{Skip to top and bottom}}. --Redrose64 🌹 (talk) 08:51, 18 February 2019 (UTC)
@Redrose64: Wow, those are beautiful (if big!). Headbomb {t · c · p · b} 08:57, 18 February 2019 (UTC)
I reduced their size a bit. Still plenty big to be easily clickable, but won't take so much of the screen. And the great thing is that they can be used on every page in WP:JCW, like WP:JCW/C3. And by everyone, not just those of us with secret javascript skills.Headbomb {t · c · p · b} 09:03, 18 February 2019 (UTC)

Proxy filter blocking a URL[]

The proxy filter is blocking this URL from being added to a mainspace page:

http://infoweb.newsbank.com/iw-search/we/InfoWeb?p_action=doc&p_topdoc=1&p_docnum=1&p_sort=YMD_date:D&p_product=AWNB&p_text_direct-0=document_id=(%20127A2C205DF7A568%20)&p_docid=127A2C205DF7A568&p_theme=aggdocs&p_queryname=127A2C205DF7A568&f_openurl=yes&p

It's not a proxy URL, or is it? -- GreenC 21:31, 16 February 2019 (UTC)

@GreenC: "infoweb.newsbank.com" is the hit on that filter. It was added by @Beetstra: a couple of weeks ago. — xaosflux Talk 21:35, 16 February 2019 (UTC)
Also FWIW, that link doesn't go to any actual content for those of us on the normal public internet... — xaosflux Talk 21:40, 16 February 2019 (UTC)
Ok great, I'll delete them. Editors will often try to hide blacklisted URLs behind web shortened URLs (eg. webcitation.org) but my bot keeps finding them when it expands the URL to long form and then is blocked trying to upload. -- GreenC 21:41, 16 February 2019 (UTC)
@GreenC: FWIW, I've exempted bots from that filter since they are still operating correctly - even though the links are of little use. — xaosflux Talk 21:46, 16 February 2019 (UTC)

@Beetstra and Xaosflux: That filter has got to go, or at least be made more specific. NewsBank URL's work perfectly fine if you know how to export them. For example: https://infoweb.newsbank.com/resources/openurl?ctx_ver=z39.88-2004&rft_dat=document_id%3Anews%252F0F6DAD8C9AE71B1B&rft_id=info%3Asid%2Finfoweb.newsbank.com&rft_val_format=info%3Aofi%2Ffmt%3Akev%3Amtx%3Actx&svc_dat=AWNB&req_dat=1028A39C75C2B899. (For any NewsBank users reading this, you go to Export in the toolbar, then Export Article Links, and copy the OpenURL link.) Toohool (talk) 01:05, 17 February 2019 (UTC)

@Toohool: the urls that I ran into were all useless proxies and therefore I assumed this was the same problem as it was in the same set. I see that we need to be more specific. Is it the /iw-search/ path that is the only one we should not use? --Dirk Beetstra T C 05:03, 17 February 2019 (UTC)
@Beetstra: Looking through the LinkSearch results, even blocking the /iw-search path is too strict. Some of those links work fine. It seems to be those that have a "f_openurl" parameter that work. There are also a lot of links with paths that start with /resources/doc/nb that only seem to work if you're already logged in to NewsBank, but that's a valid use case (those links should just be flagged as subscription-only). Toohool (talk) 05:48, 17 February 2019 (UTC)
@Toohool: I have removed the newsbank part for now. I saw an or who performed 10-15 s, and many of those were clear proxies. Also these links did not work for me (see the first link above - no content, no information as to what it is supposed to link to), so I added these in the same go. Maybe I was mistaken on this one, or we need to be way more specific (we could consider to update the warning: MediaWiki:Abusefilter-warning-proxy-link to reflect your instructions). --Dirk Beetstra T C 06:06, 17 February 2019 (UTC)

Speedy painless tool to review complex citation templates[]

When reviewing existing reference citations, sometimes in the default window we see that the prior or formatted the citation template in a column, with a line break after each parameter. Other ors prefer to collapse these by removing the line-breaks. Its all fun and good faith. In this example, the prior version shows columnar formatting, and the revision shows the same code with the line breaks removed. Every once in awhile, I'd like to do a wiki-gnomish review of the citation parameters, but its painfully slow when the linebreaks are absent.

Do we have a tool that could read a "collapsed" citation and display the parameters in columnar format in a pop up window? Would you use such a tool if it existed? If it doesn't exist, would this be something to suggest for Twinkle or somewhere else?

Thanks for thoughts NewsAndEventsGuy (talk) 11:44, 17 February 2019 (UTC)

I don't know about a tool, but here's how I do it: open two tabs for the article. In one tab, scroll down to the references. In the other tab, the article. Look through the references on the rendered page, and when you find an error, switch over to the "" tab to make changes. To check overall consistency, I like using {{ref info}} as well. – Jonesey95 (talk) 17:25, 17 February 2019 (UTC)

Bot request[]

I have currently made a request in the page Wikipedia:Bots/Requests for approval/Adithyakbot. I would like to know whether I can the whole topic present in that page. The doubt arises due to the reason that I have 'submitted' a request.Adithyak1997 (talk) 14:44, 17 February 2019 (UTC)

You can that page. – Jonesey95 (talk) 17:26, 17 February 2019 (UTC)
@Adithyak1997: and until you transclude it to Wikipedia:Bots/Requests for approval it isn't really "submitted" yet anyway. — xaosflux Talk 17:50, 17 February 2019 (UTC)

Tool that tracks WikiProject banner removals?[]

One of the functionalities previously provided by the Wikipedia:Version 1.0 Editorial Team/Index assessment logs was tracking the removal of articles from a WikiProject (when a banner is removed or when an article is deleted). User:WP 1.0 bot hasn't been working for some four months now, and I'd like to know if there are alternative tools that could provide said information. It's relatively trivial to track project banner additions or other changes to talk pages via Special:RecentChangesLinked, but this of course doesn't work for removals. --Paul_012 (talk) 17:48, 17 February 2019 (UTC)

Wikidata inverse property in Lua[]

I've been playing around with a module to extract Wikidata less painfully. For example {{#invoke:Sandbox/Certes|page_property|Persian cat|instance of}} produces "cat breed". (Some inputs produce a list of multiple values; I picked a simple example.) I'm using entity:formatPropertyValues(property) from mw:Extension:Wikibase Client/Lua. Is there any way to detect this relationship from the other end, so that I can write {{...|inverse_page_property|Cat breed|instance of}} to display "Abyssinian, ..., Persian cat, ..."? I can do this in WQDS with SPARQL, but it would be useful to have access to it in a template, either directly or via a Lua module. Certes (talk) 19:53, 17 February 2019 (UTC)

I don't think there is. {{3x|p}}ery (talk) 19:55, 17 February 2019 (UTC)

Our notification process has demonstrated inexplicable behavior[]

I am perplexed by the selective nature of our notifications process regarding this recent posting. Therein, I "mentioned" two ors (without an attributable error) yet: only one notification was delivered, and the other, inexplicably, was not.[6] Please enlighten me if it can be explained, or assist in reporting if, instead, it's a "bug". Thank you.--John Cline (talk) 23:55, 17 February 2019 (UTC)

Could it be because the text you removed also mentioned one of the ors? The one with a net increase in mentions got pinged; the other didn't. Certes (talk) 00:16, 18 February 2019 (UTC)
Thank you Certes. That is a reasonable assumption and most likely will bear out the cause. It might even be trivial to those "in the know"? But I don't think it comports with mainstay expectations regardless of cognitive skill.--John Cline (talk) 02:16, 18 February 2019 (UTC)

Pinged out of the blue[]

Why did Special:Diff/883844394 made by SwagGangster ping me?—CYBERPOWER (Chat) 00:54, 18 February 2019 (UTC)

Perhaps because it accidentally transcluded an entire talk page on which you had been mentioned. Certes (talk) 01:07, 18 February 2019 (UTC)
You didn't say which type of ping it was. The transcluded page Talk:Viggo Mortensen had two links to User:Cyberpower678/FaQs#InternetArchiveBot but a subpage link should not produce a mention notification. Was it a notification that a page you created had been linked? PrimeHunter (talk) 02:27, 18 February 2019 (UTC)
PrimeHunter, it was a mention to my email. Most likely my bot was pinged. I get emails when someone mentions IABot. —CYBERPOWER (Around) 03:16, 18 February 2019 (UTC)
Cyberpower678 That was my fault, I pinged 22 users because I accidentally transcluded the article on the Talk Page. 😅 SwagGangster 03:22, 18 February 2019 (UTC)
SwagGangster, lol —CYBERPOWER (Around) 03:26, 18 February 2019 (UTC)
Back to fighting the endless round of vandalism I go. SwagGangster 03:37, 18 February 2019 (UTC)