Jump to content

Wikipedia:Village pump (technical): Difference between revisions

From Wikipedia, the free encyclopedia
Content deleted Content added
Line 755: Line 755:


[[User:ZuluPapa5|Zulu Papa 5 *]] ([[User talk:ZuluPapa5|talk]])
[[User:ZuluPapa5|Zulu Papa 5 *]] ([[User talk:ZuluPapa5|talk]])



LEAVE THESE TWO POSTS HERE. BECAUSE THIS SHOWS WHAT CAN HAPPEN to disrupt an editor's contribution. 18:34, 26 April 2020 (UTC)


== Why does Wikipedia keep showing edit conflicts on my own edits? ==
== Why does Wikipedia keep showing edit conflicts on my own edits? ==

Revision as of 18:34, 26 April 2020

 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).

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.


Two column edit conflict - not working

Just noting that I'm still losing work through edit conflicts, despite having activated two-column edit conflict in beta. Anybody else experiencing this since it was last mentioned here] on 6th April? Nick Moyes (talk) 14:08, 10 April 2020 (UTC)[reply]

Yup, just got a "normal" edit conflict earlier today, and I have that beta feature enabled too. Enterprisey (talk!) 02:30, 12 April 2020 (UTC)[reply]
@Nick Moyes and Enterprisey:, on what pages have you experienced this issue? --AntiCompositeNumber (talk) 04:37, 12 April 2020 (UTC)[reply]
While trying to make this edit on Talk:John Horton Conway. Enterprisey (talk!) 06:40, 12 April 2020 (UTC)[reply]
@AntiCompositeNumber: My content loss occurred whilst replying at the Teahouse. (It also occurred on a userpage on Czech Wikipedia, but there I assumed I didn't have the two column edit conflict tool enabled.) Nick Moyes (talk) 07:03, 12 April 2020 (UTC)[reply]
A patch was deployed on April 6 that disabled the two-column edit conflict page on talk pages and project-namespace pages. There was no phab task or anything associated with the patch, so I don't have any more details. Thiemo Kreuz (WMDE), can you provide some insight into what is going on? --AntiCompositeNumber (talk) 15:46, 12 April 2020 (UTC)[reply]
Hi @Nick Moyes, Enterprisey, and AntiCompositeNumber: we disabled the Two Column Edit Conflict View for talk pages, because we got the feedback that it is not helpful to solve edit conflicts on talk pages. This is why you see the old interface when having an edit conflict on a talk page.
However, we are currently developing and additional interface specifically designed to help users solve edit conflicts on talk pages. For more information, also see T230231. --For the Technical Wishes Team: Max Klemm (WMDE) (talk) 09:46, 14 April 2020 (UTC)[reply]
@Max Klemm (WMDE): Thank you for this update. There were times when there seemed to be no interface at all, just a complete loss of edits, so, in the short term, I'll try and remember to copy my replies before posting them at potentially busy fora. (Just in case!) Cheers, Nick Moyes (talk) 10:15, 14 April 2020 (UTC)[reply]
@Max Klemm (WMDE): I actually found it super useful on talk pages! Where was that feedback? --Izno (talk) 17:02, 14 April 2020 (UTC)[reply]
@Max Klemm (WMDE): I believe I had the same experience as Nick Moyes a couple of times on the Teahouse or Help Desk, with no conflict resolution facility (I think***) – just an edit window for the whole page (as opposed to just the section I was editing) with my changes missing. I was able to use the browser's (FFox) back button, accept the dialog about resending, and get back to the edit window with my edits in it. I then copied it to a text editor, restarted an edit on the section, pasted my changes back in, and saved successfully. ***It's entirely possible that the page I got had additional language on it telling me there was a conflict and what to do about it – I was too busy panicking and trying to recover to look very carefully. —[AlanM1 (talk)]— 02:17, 15 April 2020 (UTC)[reply]
Working for me... 3125ATalk!Contributions! 21:32, 22 April 2020 (UTC)[reply]

Suggestion: display certain hatnotes only if the user has been redirected from certain targets

nb: I made this suggestion at the redirection template talk page (discussion here) and on phub ( phab:T232278) but no one seems to be entirely sure where the right place for it is. Following another suggestion I'm trying it here, too.

I think we should display certain hatnotes only if the user has been redirected from certain targets. For example:

The Blue Album is an album by Weezer. It contains a song called "The World Has Turned and Left Me Here". There is also an episode of a TV show titled "The World Has Turned and Left Me Here".

If you go to the article, it currently displays this hatnote: The World Has Turned and Left Me Here" redirects here. For the Vampire Diaries episode, see The World Has Turned and Left Me Here (The Vampire Diaries).

I think this hatnote should only be displayed if the user has searched for "The World Has Turned and Left Me Here" and been redirected to Weezer (Blue Album). My logic is this:

  • If you're redirected to the Blue Album article because you were looking for an article about the TV episode, the hatnote is useful. It tells you that you're in the wrong place and gives you a link to the right article.
  • If you arrive at the Blue Album article any other way (most likely more than 99% of cases), the hatnote isn't useful - you almost certainly don't care that there is an episode of a TV show with the same title as a song on this album. It's just clutter.

It would be great if there were a way to display certain hatnotes only if the user has been redirected from certain targets. Popcornfud (talk) 15:33, 12 April 2020 (UTC)[reply]

I'd oppose this. Trying to second-guess what readers will want to see is rarely a fruitful exercise; to stick with your original example, I can easily imagine a fan looking for information on the album finding it interesting that the name of one of its songs was also used as the title of a TV show. Since no reader in the history of Wikipedia has ever found a hatnote disruptive (we're talking about one line at the top of an article), this proposal would seem to cause potential disruption to readers for no obvious benefit.
There's a high number of pages where we undoubtedly would need to display the hatnote even to those who typed in the article title verbatim (e.g. if someone searches for "The Sirens and Ulysses" there's a decent chance they're actually searching for "Ulysses and the Sirens" and vice versa). Thus, if we did go down this route, the display hatnote=yes or display hatnote=no would need to be set manually for every single redirect on Wikipedia. To put that in perspective, as of December 2019 when the report last ran there were 8,966,504 mainspace redirects; I'm guessing you'll have difficulty persuading anyone to review nine million pages for little if any obvious benefit. ‑ Iridescent 15:55, 12 April 2020 (UTC)[reply]
Iridescent, I find your first paragraph unconvincing. no reader in the history of Wikipedia has ever found a hatnote disruptive (we're talking about one line at the top of an article - well, I find it disruptive! And in the case of the Weezer article it adds two lines about two separate articles, not one. I really do think it adds a lot of clutter that helps only a vanishingly small minority of users. Adding a big lump of text to the top of the article is not a small cost from a usability/readability perspective.
I can easily imagine a fan looking for information on the album finding it interesting This is neither here nor there. If encyclopaedic information is worth including in the article then it should be included in the article proper, not delivered in a sideways fashion via a hatnote. And as a general rule, per WP:POPCULTURE, we don't typically consider references to things in other media encyclopaedically notable anyway.
However, regarding the second paragraph: I have no experience of managing the technical side of this sort of thing so it would be up to other editors to figure out if what I'm proposing could realistically be implemented at all. If the implications of that are unmanageable then the idea is indeed dead on arrival, oh well.
I'm wondering, though, if we could add an optional flag on redirects to only display a hatnote on certain conditions, so that we wouldn't have to change any existing ones - it would just be added as an exception or override where useful. Popcornfud (talk) 16:25, 12 April 2020 (UTC)[reply]
Knowing the article from whence one came is not possible in wikitext/Lua. This would be a Javascript addition that would need to be made and I would accordingly oppose it as Not Worth It. --Izno (talk) 16:28, 12 April 2020 (UTC)[reply]
I see. I assumed we had that information in a way that could be used as it says "redirected from <wherever>" when you get redirected. Obviously even if it could still be used this might easily still be Not Worth It. Popcornfud (talk) 16:42, 12 April 2020 (UTC)[reply]
Even if it were technically possible, I wouldn't be in favor of it. Just as a reader wikisurfing, hatnotes have led me to surprising and interesting connections that I wasn't aware of. As an editor who uses Special:Random a lot, it would not be helpful if the version of the article that I see is not the same version all readers see. Schazjmd (talk) 16:46, 12 April 2020 (UTC)[reply]
I'd support this if it were technically feasible without performance costs, but I don't think it is. --Yair rand (talk) 21:44, 12 April 2020 (UTC)[reply]
I think this could be done by adding a hash to the end of the redirect (e.g. #redirect [[Page title#hatnote-1]]) (which would need the hatnote template to support an ID parameter) then using css's :active pseudoclass to make it visible.  Nixinova  T  C   02:44, 13 April 2020 (UTC)[reply]
  • I like this idea a lot. I don't at all buy the "hatnotes can provide random interesting information" argument — that's not what they're supposed to be for, and they're occupying extremely valuable space at the top of the article. And it wouldn't be necessary to go through all 9 million redirects — just introduce a parameter and let it be adopted as needed. As to the technical side of things, I'm not going to be able to productively contribute to that, but if it's not possible now, perhaps we could tuck this away as a good idea to come back to once Wikipedia's coding is in an improved state. {{u|Sdkb}}talk 05:45, 13 April 2020 (UTC)[reply]
  • Having hatnotes that people will not see when looking at the page directly seems like a potential vandalism time bomb where vandals can place offensive text that will only show up when someone reaches the page from a particular redirect. 24.151.50.175 (talk) 15:54, 20 April 2020 (UTC)[reply]
    Hmm, that's a really good point. If this moves forward, something will have to be done to address that, but I'm not sure what. {{u|Sdkb}}talk 07:10, 24 April 2020 (UTC)[reply]

Editing news 2020 #1 – Discussion tools

Read this in another languageSubscription list

Screenshot showing what the Reply tool looks like
This early version of the Reply tool automatically signs and indents comments.

The Editing team has been working on the talk pages project. The goal of the talk pages project is to help contributors communicate on wiki more easily. This project is the result of the Talk pages consultation 2019.

Reply tool improved with edit tool buttons
In a future update, the team plans to test a tool for easily linking to another user's name, a rich-text editing option, and other tools.

The team is building a new tool for replying to comments now. This early version can sign and indent comments automatically. Please test the new Reply tool.

  • On 31 March 2020, the new reply tool was offered as a Beta Feature editors at four Wikipedias: Arabic, Dutch, French, and Hungarian. If your community also wants early access to the new tool, contact User:Whatamidoing (WMF).
  • The team is planning some upcoming changes. Please review the proposed design and share your thoughts on the talk page. The team will test features such as:
    • an easy way to mention another editor ("pinging"),
    • a rich-text visual editing option, and
    • other features identified through user testing or recommended by editors.

To hear more about Editing Team updates, please add your name to the "Get involved" section of the project page. You can also watch these pages: the main project page, Updates, Replying, and User testing.

PPelberg (WMF) (talk) & Whatamidoing (WMF) (talk) 15:45, 13 April 2020 (UTC)[reply]

I think this would be a nice beta feature to have. Can it be placed under "Beta" so I can try it out on enwiki? Aasim 19:02, 13 April 2020 (UTC)[reply]
You can try it on the wikis where it's already in a beta. enwiki is an unlikely candidate for betas because well, enwiki. @Whatamidoing (WMF): Does this new tool have some feature in place that prevents or fixes edit conflicts? How does it deal with them? --qedk (t c) 13:27, 14 April 2020 (UTC)[reply]
User:QEDK, the Reply tool auto-resolves most of edit conflicts (the magic is in mw:Parsoid, apparently). If it can't resolve it, you get an error message and the opportunity to copy your message (so you can either reload the page and try again, or open your normal wikitext editor and reply the old-fashioned way). I've seen edit conflicts that were auto-resolved; I've not seen complaints on any of these wikis about it failing (sample size =~500 real-world replies). If anyone wants to test this, remember that you need two accounts to trigger an edit conflict with yourself. MediaWiki overwrites your own edits rather than triggering the edit conflict tools.
Anyone who wants to try it out can enable the Beta Feature and leave a note for me on my talk page at any of those wikis, or go to https://fr.wikipedia.org/wiki/Utilisatrice:Whatamidoing_(WMF)/Brouillon?dtenable=1 and "Répondre" to your heart's content in my sandbox. Whatamidoing (WMF) (talk) 16:47, 14 April 2020 (UTC)[reply]
@Whatamidoing (WMF): That sounds good! There's a couple of issues I've come across: 1) sometimes I just get an empty box saying "OK" when clicking "Reply", not sure why. This has happened a few times, refreshing the page fixes it. 2) The "Reply" is disabled as soon as you hit "Reply" on one message, this is not ideal because sometimes you misclick and want to switch to a different editor, ideally you want a prompt asking you to switch replying to that if the editor has started typing a reply to someone else. For info, I am using Safari 13.0.4 on macOS Catalina 10.15.2. --qedk (t c) 19:09, 14 April 2020 (UTC)[reply]
  1. is weird, and is also now phab:T250234.
  2. is deliberate, because it can only handle one Reply at a time. The red "Cancel" button will bring the other Reply buttons back (first copy any text you have already typed and want to keep, of course). Whatamidoing (WMF) (talk) 22:35, 14 April 2020 (UTC)[reply]
Awesome Aasim, in the mean time, the fantastic user script WP:REPLYLINK does basically the same thing. Gaelan 💬✏️ 22:43, 14 April 2020 (UTC)[reply]
User:Gaelan and anyone else who uses it: please look at my question at User talk:Enterprisey/reply-link#How often does it work? Whatamidoing (WMF) (talk) 20:02, 16 April 2020 (UTC)[reply]
I played with it a little bit. Would I be correct to deduce that it looks for a line (or paragraph, if top-level text) that contains both (a) a link to a User: or User talk: namespace (can be cross-wiki or redlink); and (b) a time stamp in the exact format of the wiki (case-sensitive, same language and timezone, etc.)?
DT added a Reply link with fr:Modèle:U but not fr:Modèle:Non signé. It appears to want the time stamp to be in the same DOM element and not wrapped in a link or a span. Pelagic (talk) 03:58, 18 April 2020 (UTC)[reply]
Pelagic, it looks for a timestamp (exactly that wiki's default format) plus any link to the user's account (user page, user talk, or contributions). It must be one of those pages and not a subpage or redirect. (I believe that in MediaWiki, "a line" is the same as "a paragraph".)
The problem of how to detect signatures in templates such as {{u}} and {{unsigned}} is phab:T250516. If you have opinions or advice (e.g., whether it's better to fix the templates or the extension), then please share there. Whatamidoing (WMF) (talk) 15:46, 21 April 2020 (UTC)[reply]

Updating WikiProject assessment tables by bot

@Kelson: and @Audiodude:. Is there a way to make WP 1.0 bot update a given quality assessment table? I am specifically wondering about User:WP 1.0 bot/Tables/Project/WikiProject Seamounts. Jo-Jo Eumerus (talk) 16:44, 16 April 2020 (UTC)[reply]

@Jo-Jo Emerus:, you can force an update here. Seamounts doesn't appear in the drop down list, so something probably isn't configured correctly. Plantdrew (talk) 16:18, 20 April 2020 (UTC)[reply]

Mystery of the Mixing Sailors

Today, I came across a nonsensical draft whilst patrolling new pages titled Draft:Mixing’Sailors, containing the following text:

1. A rich mummy, usually green or bald.

2. To overbreak toes, i.e. to put a cat into a fish or into a dog.

3. To stab an elephant with a nostril.

Mystified by who would think up such a thing, I looked up the text on the Internet and found the same thing pasted on many other wikis and user-editable sites, as well as elsewhere on Wikipedia. Also, the page title Draft:Mixing Sailors, without the apostrophe, has actually been protected from creation using the title blacklist. Does anybody know who is spamming this stuff across the Web, and for what reason? Passengerpigeon (talk) 13:28, 18 April 2020 (UTC)[reply]

It's old and appears to always be added by French IP addresses or new accounts. User talk:81.185.159.77 is from 2009. It's called "The ideal joke" at User talk:79.85.113.185. It was globally blacklisted in 2011.[1] It's rare enough to just be one odd person. User:Croobatch posted it and also made Oil and clown colors with only content: "Bill and the maps are sticking well together because they appear to be made of swollen iron and greasy spokes." They got a checkuserblock. User:Gdodjolsoljkfvg got one the same minute by the same checkuser. They created Gsthae with tempo! with: "Moosh drives like a horn because his tie is not more than a crablike queefish. Cats dubbed with strings, a horror diet clowned about a forshipmore then skies linked a dribbloo." I don't expect to find sense in any of this. PrimeHunter (talk) 14:57, 18 April 2020 (UTC)[reply]
The lesson, as always: Wikipedia editors will always do something more outlandish than you can imagine. P.S. I know Moosh. She does drive like a horn. – Jonesey95 (talk) 15:35, 18 April 2020 (UTC)[reply]
This was dealt with using a {{schoolblock}} back in 2015. See User_talk:193.54.167.180#March_2015. – Fayenatic London 09:27, 24 April 2020 (UTC)[reply]

Save edit in Visual Editor, it's asking to save again?

I've seen this three times over the past week or two. I edit a page in the Visual Editor, save my changes, and then I'm back to VE asking me to save. For example, I just made this edit. Yet, I'm looking right now at a "Publish changes" dialog, asking me to save those exact same changes. I just clicked "Publish changes" again, to see what would happen, and it looks like it's just hung. I've got the "Review your changes" barber-pole across the top, with the strips moving to indicate operation in progress. Anybody else been seeing anything like this? -- RoySmith (talk) 21:53, 18 April 2020 (UTC)[reply]

Just had it happen again, on this edit. I hit "Publish changes", the edit got saved, and I've got another dialog asking me to save. Clicking "Review your changes" shows no changes. Clicking "Publish changes" hangs. -- RoySmith (talk) 14:55, 19 April 2020 (UTC)[reply]
Wow. Again, on this edit. -- RoySmith (talk) 15:07, 19 April 2020 (UTC)[reply]
RoySmith, can you tell me your web browser and OS? (I get something like this in Firefox on macOS, but only at Wikivoyage.) Also, the next time this happens, could you please open the page history in another tab, and see whether it has posted your changes already? (Please ping me.) Whatamidoing (WMF) (talk) 15:52, 21 April 2020 (UTC)[reply]
This is being tracked at T250620; could you update that with your observations? But, to answer your question, I'm on Chrome / MacOS, and the other editor who was able to reproduce it, the same. And, yes, every time I've checked the history in another window, the history recorded it had been saved. -- RoySmith (talk) 16:28, 21 April 2020 (UTC)[reply]

""ntsamr"-pattern spambot filter"

I noticed this hidden edit filter whilst patrolling the AbuseLog and was wondering what exactly it is. What does the "ntsamr" pattern constitute, i.e. what does this filter look out for in an edit?

Thank you, Passengerpigeon (talk) 10:54, 19 April 2020 (UTC)[reply]

Nothing to say about me really. If it's private there might not be much I can publicly say about exactly how it works, but think spambot user pages. The filter is not very reliable for various reasons, although it's not too bad. You might see some examples if you look at the logs of the deleted user pages of the blocked accounts; one such example is this another this. -- zzuuzz (talk) 11:16, 19 April 2020 (UTC)[reply]
Ah, I see the pattern now, although the two pages I viewed which that filter had flagged were nothing like that format. Passengerpigeon (talk) 11:54, 19 April 2020 (UTC)[reply]
Hello. this probably shouldn't have been flagged ? Supertoff (talk) 16:16, 21 April 2020 (UTC)[reply]

Change to CS1 (cite xxx) templates is showing warning messages to some editors using custom scripts

The following discussion is closed. Please do not modify it. Subsequent comments should be made in a new section. A summary of the conclusions reached follows.
Per WP:TALKFORK, questions and comments relating to the change should occur at Help talk:Citation Style 1#Cite book Harv warning. Requests for a script should occur at Wikipedia:User scripts/Requests#HarvErrors.js. Kees08 (Talk) 01:57, 22 April 2020 (UTC)[reply]

If you have the useful User:Ucucha/HarvErrors script installed in your common.js file, you may notice an increase in the number of brown "warning" messages that the script is generating in articles. This is due to a change in the Citation Style 1 (cite book, cite web, etc.) templates that adds |ref=harv to those templates by default in order to fix short citation links to those full citations. There was discussion of this change at Help Talk:Citation Style 1, a good (though sometimes quite active) forum to follow if you have an interest in citation formatting and templates.

No action is required on your part, or in the articles in which more messages are appearing. If the brown warning messages bother you, there is advice at the top of User:Ucucha/HarvErrors offering a few options to suppress them. (P.S. Please don't shoot the messenger.) – Jonesey95 (talk) 18:40, 19 April 2020 (UTC)[reply]

"Selected works" section in Coral Lansbury
This is what I see now in "Selected works" and "Further reading" sections that use {{cite book}}, {{cite news}}, etc. But if I remove the script that points to the errors, I lose the errors that I want to catch, namely when a long citation hasn't been used as a source. Can someone point to the change that has caused this? Someone has made ref=harv the default. Pinging Dudley and Gerda who have also complained about it. SarahSV (talk) 18:33, 21 April 2020 (UTC)[reply]
It's a little unclear to me why the solution to a script malfunctioning is to fork the script. Why isn't the original script being edited so that it stops bugging up? Jo-Jo Eumerus (talk) 18:35, 21 April 2020 (UTC)[reply]
Jo-Jo Eumerus, sorry, I can't answer that. They've said we have to ask for a new script to be written. But this change (that ref=harv is now the default) didn't have consensus. It was discussed in 2018 and again recently, this last time by just a few people. There should be wider consensus for changes that affect so many people. SarahSV (talk) 18:38, 21 April 2020 (UTC)[reply]
@Jo-Jo Eumerus: because user scripts have security risks, and you can't edit someone else's user script. Maybe interface admin can, but mortals can't. So the solution, if you're not happy with the way a script works, is to make a new one. Or have the user update the script if they are still around and willing. If people don't posses the skills to make their own fork, WP:SCRIPTREQ exists to draw on the knowledge of those who can. Headbomb {t · c · p · b} 18:51, 21 April 2020 (UTC)[reply]
SlimVirgin, have you forgotten that you have already been involved in a detailed discussion of this change at Help talk:Citation Style 1#Cite book Harv warning? Your questions have been answered there already. Please do not fork discussions.
Jo-Jo Eumerus: The script is not malfunctioning. It is showing the same no-action-necessary (but occasionally useful) warning (not error) messages that it has always shown, and it is now showing those warnings for an expanded set of templates. That change, like many changes, has resulted in disappointment from some editors who, like most of us, find change difficult. Instructions have been posted in various places telling editors how to modify the script's output, or how to use different scripts. – Jonesey95 (talk) 18:53, 21 April 2020 (UTC)[reply]
Also, the bit about consensus is obviously false. There was pretty much unanimous consent that this is an improvement to the ease of use of templates and fixes tens of thousands of articles with broken citation anchors. What we have now is 3-4 users discontent that a custom script they use doesn't work exactly the same as it did before, because the script makes assumptions that it shouldn't make, and they refuse to request a new one. Headbomb {t · c · p · b} 18:54, 21 April 2020 (UTC)[reply]
(edit conflict)Sorry, I wasn't aiming this question at you but at the people who wrote the forks. I don't think the change to the reference templates is a big deal, myself, it just needs a bot to clean out the now-unneeded |ref=harv and some repair done to User:Ucucha/HarvErrors. The problem I have is that it's more work to ask people to use new scripts rather than simply adjusting the old ones. @Headbomb:, the thing is that Interface Administrators can edit the user scripts and this isn't really a style disagreement but a bug. Jo-Jo Eumerus (talk) 18:55, 21 April 2020 (UTC)[reply]
It's not a bug, as Jonesey95 explained, the script works exactly as intended. Headbomb {t · c · p · b} 18:59, 21 April 2020 (UTC)[reply]
This is not a bug in the script. As I said in the discussion where this back-and-forth should be happening, per WP:TALKFORK: The error scripts are still useful for the task described above. If you have full citation templates in a "Works cited" section, and one of them should be but is not connected to an {{sfn}} template, the unused one will show a brown warning message. See this version of Coniston railway station (England) for an example. Only one full citation in the Sources section shows the brown warning message. In this case, it is because the sfn template that is trying to use it omits the year. – Jonesey95 (talk) 19:05, 21 April 2020 (UTC)[reply]
Trappist the monk, would you please revert this change? It needs a broader consensus. SarahSV (talk) 18:56, 21 April 2020 (UTC)[reply]
There is consensus, a minority that refuses to update their scripts does not overrule consensus to provide core functionality to CS1 templates. Headbomb {t · c · p · b} 18:59, 21 April 2020 (UTC)[reply]
Please link to the consensus. So far you've provided two links: one to a discussion in 2018 and a recent one in which very few people participated, where you requested the change. Which discussion produced the consensus? SarahSV (talk) 19:06, 21 April 2020 (UTC)[reply]
See here and links therein, but you already saw that. Headbomb {t · c · p · b} 19:09, 21 April 2020 (UTC)[reply]
Four people took part in that discussion. SarahSV (talk) 19:25, 21 April 2020 (UTC)[reply]
Answered at Help talk:Citation Style 1 § Break
Trappist the monk (talk) 19:10, 21 April 2020 (UTC)[reply]
Not answered, which is why I started discussing it here. The number of people on that other page is limited, and you were involved in making this change and therefore are defending it. SarahSV (talk) 19:25, 21 April 2020 (UTC)[reply]
That you refuse to hear the answer does not mean this was not answered. Headbomb {t · c · p · b} 19:29, 21 April 2020 (UTC)[reply]
Trappist the monk, you've suggested two scripts to help with this, but both remove the error messages we need. This change means we no longer see when citations are listed but not used, or we have to put up with big brown error messages in FR and Selected works sections that use {{cite book}}, etc. If this can't be sorted out somehow, then we need to discuss this at AN. I recently tried to add a publisher's name that has one italicized word in it, and the template produced an error message because apparently italics are not allowed in "publisher="? But why not? It seems central decisions are being made that cause editors to lose a lot of time trying to fix or work around them. I don't know what to suggest, but there needs to be some sense of gaining wider consensus before making changes that might cause a problem. SarahSV (talk) 19:37, 21 April 2020 (UTC)[reply]
Pinging Dudley Miles and Gerda Arendt so they know this is here. SarahSV (talk) 19:40, 21 April 2020 (UTC)[reply]
This change means we no longer see when citations are listed but not used, or we have to put up with big brown error messages in FR and Selected works sections that use {{cite book}}, etc. So then keep using User:Ucucha/HarvErrors.js, which apparently does display the warnings you want to see. You can't simultaneously want messages and not want them as the same time. Headbomb {t · c · p · b} 19:41, 21 April 2020 (UTC)[reply]
I want to see error messages for unused sources when I am editing an article which uses anchors. When I am following the style of an article which does not use anchors, I leave out ref=harv from cite book so that I will not get false error messages. Now that I can no longer use ref=harv there is no way User:Ucucha/HarvErrors.js or any other script can distinguish between the two cases. Dudley Miles (talk) 20:12, 21 April 2020 (UTC)[reply]
Again, wrong. CS1 and CS2 behave the same way, so however you 'ignored' errors with CS2 can be used with CS1. Typically done via the addition of |ref=none. Headbomb {t · c · p · b} 20:16, 21 April 2020 (UTC)[reply]
I do not know why you keep raising CS1 and CS2. I just asked whether they work the same way and you explained they did not. A script which relied on distinguishing sources depending on whether or not they had ref=none would not help to find unused sources in the countless articles written relying on ref=harv to indicate an anchor. It would give vast numbers of false warnings. Dudley Miles (talk) 20:28, 21 April 2020 (UTC)[reply]
I keep raising this because you keep complaining that CS1 (e.g. {{cite book}}) is problematic while simultaneously claiming that CS2 (e.g. {{citation}}) is not when they now both work exactly the same. This has been explained to you multiple times, including in the sentence right above your reply where you claimed i said that CS1 and CS2 worked differently. It's really hard to have a conversation when you read "CS1 and CS2 behave the same way" to mean "you explained they did not". Headbomb {t · c · p · b} 20:30, 21 April 2020 (UTC)[reply]
I do not think that you explained that CS1 means cite book and CS2 means citation. You just assumed that I knew and I assumed they are different protocols for which applied to both templates. Dudley Miles (talk) 20:45, 21 April 2020 (UTC)[reply]
I mean, it's just in the title of the section and at the top of Help talk:CS1#Cite book Harv warning in the big nutshell box. It's not like you ever indicated you did not know what those meant either. Headbomb {t · c · p · b} 21:00, 21 April 2020 (UTC)[reply]
I don't understand why Headbomb doesn't understand what is being said. There is some kind of miscommunication. The issue is that the script we used no longer works. It gives us error messages where there are no errors. If we disable it, we lose the error messages we want. Trappist the monk has come up with two suggested scripts but they had the same problem. The issue is that we need to be able to remove ref=harv. It was a benefit to have a set of templates that allowed that functionality. It made things easy and fast. What you saw as a flaw, we saw as an advantage. SarahSV (talk) 21:05, 21 April 2020 (UTC)[reply]
And I don't understand what you don't understand about using |ref=none to suppress anchors, rather than "removing |ref=harv" which was a solution that only worked with CS1 templates. You couldn't remove |ref=harv in {{citation}} to suppress anchors, so I really don't know why you are so distressed by CS1 and CS2 behaving exactly the same way as each other. Headbomb {t · c · p · b} 21:08, 21 April 2020 (UTC)[reply]
Your reply is an example of what I mean. This has already been explained to you. {{Citation}} isn't used as much. Most people use the templates you've just changed or caused to change. We can't go around adding ref=none to all these articles. SarahSV (talk) 21:13, 21 April 2020 (UTC)[reply]
Then learn to live with the warnings. Or, again, go make a WP:SCRIPTREQ to have a new script made that works to your liking. Headbomb {t · c · p · b} 21:21, 21 April 2020 (UTC)[reply]
You keep suggesting a new script, but as I pointed out above, it would not help. A script based on ref=none would be a good solution if it had been used from the start, but changing to it in the middle is not going to help with error checking in the many existing articles which relied ref=harv to distinguish between sources which are intended as anchors and those which are not. It may just give false error warnings. Dudley Miles (talk) 21:34, 21 April 2020 (UTC)[reply]
Giving that you're not happy with how the current script works, and that you want a different behaviour, I fail to see how a new script that worked like you wanted would "not help". You've got a path forward. If you don't want to take it, that's on you. Headbomb {t · c · p · b} 21:39, 21 April 2020 (UTC)[reply]

So that's why I am suddenly getting HarvRef errors on articles where I don't even use HarvRefs. Stuff like this explains why I manually formatted citations for years, and regret now having switched to citation templates. Please fix this. SandyGeorgia (Talk) 21:51, 21 April 2020 (UTC)[reply]

Again, as was explained multiple times, if you don't want warnings, don't use User:Ucucha/HarvErrors.js but rather User:Svick/HarvErrors.js. Or make a WP:SCRIPTREQ to have a new script that behaves how you want. Headbomb {t · c · p · b} 21:58, 21 April 2020 (UTC)[reply]
But that turns off all the warnings, including the good ones, does it not? SarahSV (talk) 21:59, 21 April 2020 (UTC)[reply]
If you went to WP:SCRIPTREQ to explain what is a "good warning" and what is a "bad warning" then you could have a script that keeps the "good warnings" and throws away the "bad ones". Headbomb {t · c · p · b} 22:30, 21 April 2020 (UTC)[reply]
I just tried it. That script turns off at least one of the good warnings. If you use a long citation but not a corresponding short one, there's no warning. It's important to be able to check that kind of thing when doing source reviews. Doing it manually can be very time-consuming. SarahSV (talk) 22:05, 21 April 2020 (UTC)[reply]
I really don't have time to deal with another citation/bot issue right now, and seriously regret having moved from manual citations to citation templates, because of the never-ending vagaries of bots and citation templates. I am unwatching this discussion. If a solution is found, Sarah has my permission to edit my userspace to implement whatever solution she finds that works. If she does that, I will do the appropriate ctrl-whatever to reload. I assume that whatever solves this problem for Sarah will solve it for me. SandyGeorgia (Talk) 22:08, 21 April 2020 (UTC)[reply]
@SandyGeorgia: if you used "manual citations", you'd have no warnings and no error flagging whatsoever. So if the warnings annoy you, then use User:Svick/HarvErrors.js and then you'd have errors, without warnings. Or use no script, and have nothing flagged. Headbomb {t · c · p · b} 22:33, 21 April 2020 (UTC)[reply]

I think we are talking past each other a little bit here. SlimVirgin, at this point there is a lot of discussion to go through, would you be able to summarize the cases in which you would like to see warnings? Kees08 (Talk) 00:17, 22 April 2020 (UTC)[reply]

It would be better to close this thread and continue the conversation at the original venue, per WP:TALKFORK, or at WP:SCRIPTREQ if a new script is desired. Multiple editors have offered to help with a new script request at WP:SCRIPTREQ. Citation Style 1 templates are discussed and decisions are made at Help Talk:Citation Style 1 and archived in the talk page archives for that page. Any discussions that are held here about CS1 templates will effectively be lost. – Jonesey95 (talk) 00:53, 22 April 2020 (UTC)[reply]
The discussion above is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.

Combining RandomInCategory with PAGENAME

I'd like to create a link to a random level 4 vital article. However, it's not as simple as using Special:Randomincategory, since Category:All Wikipedia level-4 vital articles goes to the talk page, not the article itself, and I can't get Special:Randomincategory to play nice with {{PAGENAME}}. Help? {{u|Sdkb}}talk 00:36, 20 April 2020 (UTC)[reply]

Unfortunately you can't do that with just core mediawiki, but you could do it with a user script or an external tool. --AntiCompositeNumber (talk) 01:00, 20 April 2020 (UTC)[reply]
@AntiCompositeNumber: In that case, would there be an easy way to set up a category that duplicates Category:All Wikipedia level-4 vital articles but goes to the page itself instead? This is for this conversation, so using an individualized or off-wiki method isn't really an option that I can see. {{u|Sdkb}}talk 05:47, 21 April 2020 (UTC)[reply]

File in category not shown

Hello!

Why isn't File:Rodinia Missing Link.png shown at Category:All Wikipedia files missing evidence of permission eventhough it should? The correct date category Category:Wikipedia files missing permission as of 26 March 2020 was also deleted, probably because the file couldn't be found there either. This is not the first time this happens so it must be a problem in som code somewhere.Jonteemil (talk) 05:48, 20 April 2020 (UTC)[reply]

nulledit solved the issue. I undeleted Category:Wikipedia files missing permission as of 26 March 2020 Agathoclea (talk) 06:00, 20 April 2020 (UTC)[reply]
@Agathoclea: Thanks. I tried null edit but nothing happened for me. I see the file in the catgory now when you've done it though.Jonteemil (talk) 10:36, 20 April 2020 (UTC)[reply]

JS problem

Resolved

At User:Doug Weller/common.js when I open the page to edit I get "https://en.wikipedia.Org/w/index.php? title=User:Doug_Weller/common.js&actic at line 818: Uncaught Error: Another textSelection API was already registered". Maybe connected, I often get "JavaScript Error script error when I load a Wikipedia page - both of these in boxes at the right-hand top of the page. I don't think that the bottom two scripts are installed correctly although the top one works, eg it marks the use of lulu.com at Girolamo Cavalcabo. Thanks. Doug Weller talk 09:35, 20 April 2020 (UTC)[reply]

@Doug Weller: See mw:Help:Locating broken scripts how to investigate and debug. (Especially as your common.js does not have 818 lines.) --Malyacko (talk) 13:29, 20 April 2020 (UTC)[reply]
@Doug Weller: when troubleshooting, keep in mind you have scripts loading from User:Doug Weller/common.js, meta:User:Doug Weller/global.js, and also likely User:Doug Weller/vector.js. — xaosflux Talk 13:51, 20 April 2020 (UTC)[reply]
@Xaosflux: Thanks for the reminder. I've managed to get rid of the Javascript message now by doing some cleanup. Doug Weller talk 14:09, 20 April 2020 (UTC)[reply]

Collapsible CS1 errors

I want to make CS1 errors collapsible with CSS, but I don't know how to code CSS and I need your help. --Ijoe2003 (talk) 10:32, 20 April 2020 (UTC)[reply]

What would a "collapsible CS1 error" look like? Headbomb {t · c · p · b} 11:52, 20 April 2020 (UTC)[reply]
I think they just want to hide them. – Ammarpad (talk) 12:01, 20 April 2020 (UTC)[reply]
Also, I expect that for some of these there will be bot passes in the near future, cleaning them up. Jo-Jo Eumerus (talk) 12:18, 20 April 2020 (UTC)[reply]
If hiding is the goal see Help:CS1 errors § Controlling error message display.
Trappist the monk (talk) 12:22, 20 April 2020 (UTC)[reply]
An error icon appears near the citation. When you click it, it shows an error which is occured on that cite with the close button. --Ijoe2003 (talk) 13:24, 24 April 2020 (UTC)[reply]

Removing pages from a category

In the category Category:Wikipedia non-empty soft redirected categories how can an entry be removed from this? The category talk page has 0 pageviews which led me to ask here. Adithyak1997 (talk) 14:08, 20 April 2020 (UTC)[reply]

@Adithyak1997: to remove pages from a category, you have to edit each page that is in the category and remove the category declaration. — xaosflux Talk 14:09, 20 April 2020 (UTC)[reply]
The category is added by a template. If you want to remove empty categories then try a null edit of the category page which is in the category. PrimeHunter (talk)
@Adithyak1997: Correct, null edit is the thing to do; it needs to be performed on every single one of the 60+ that presently say "(empty)". A bit tedious, it's one of the tasks that Joe's Null Bot (talk · contribs) used to do. --Redrose64 🌹 (talk) 09:17, 21 April 2020 (UTC)[reply]
It should be better now. I did some null edits. – Jonesey95 (talk) 15:08, 21 April 2020 (UTC)[reply]
I have added an explanation to the category.[2] PrimeHunter (talk) 11:26, 24 April 2020 (UTC)[reply]

Talk pages in deleted pages

Hello! I have a question that may sound a bit silly. I already checked the FAQ and I don't think I saw it there so...

When we delete a certain page, why do "have" to keep its talk page? Those get "lost" and the way I see it, only remain as a burden to the servers (albeit an insignificantly small one). Not many users, to say the least, will ever search for a talk page of a non-existent page. Are there strong reasons we keep them? Apart from the edge cases where users want to go and discuss about something related to the deletion process. Is there a way to track these kind of "friendless" talk pages? Is there a way to delete them simultaneously when we are deleting a certain page? What's the overall general logic in keeping talk pages in these situations?

I'm nearly sure something like this has been answered before and there must be reasons I'm overlooking in my logic but I couldn't find it at the FAQ page so curiosity got the best of me. - Klein Muçi (talk) 14:54, 20 April 2020 (UTC)[reply]

@Klein Muçi: we don't "have" to keep anything, and we specifically have a policy to allow for their removal (Wikipedia:Criteria_for_speedy_deletion#G8._Pages_dependent_on_a_non-existent_or_deleted_page). Some scripts and tools, such as Twinkle have options to "also" delete a talk page. From the backend, there is no requirements that these pages have an associated page, for example User talk:Xaosflux/Archive20 has no "page" in the related namespace, and if one were created deleting it wouldn't mean that that talk page should also be deleted; and if say you asked for your userpage to be deleted, there would be no need to also delete your talk page. — xaosflux Talk 15:16, 20 April 2020 (UTC)[reply]
User:Klein Muçi may have been thinking about user talk pages. Even if we delete a user page, we almost never delete the associated user talk page, even if the user asks us to. We keep it to maintain the user's history. See WP:DELTALK. -- MelanieN (talk) 15:29, 20 April 2020 (UTC)[reply]
@Xaosflux: Yes, I know of those cases and that policy because I've researched a bit this subject because of my curiosity. But I was referring mostly to the classical, unenhanced way of deleting pages. When you try to move something, you get options that allow you to decide if you also want to move its sub/talk pages. When you delete pages, there are no facilities like these. Wouldn't it be easier to have the same options when deleting pages? I believe there will be thousands, if not more, articles', categories', templates' and what not pages deleted whose talk pages still remain in the project without any specific need. This is certainly the case at SqWiki when I'm an admin. Many admins there don't even know that the talk pages still remain after deleting those kind of pages and those that do know, don't delete them anyway because they think they are needed because of something they don't know yet, because otherwise they'd have the options to delete them together with the page, like the move options.
@MelanieN: No, I know you keep them even when users get retired. Keeping that in mind, I thought maybe a reason for that exists in other namespaces too. I'm asking because, as I said, we have years (more than a decade) of a habit of not deleting talk pages in SqWiki because of the reasons specified above. And since no one really knew the reason why we were doing that (assuming they did know we weren't deleting them, because most of them thought the deletion happened simultaneously) I thought of asking here. And if no general reason really exists for that, I'm starting to think we might need to do a massive cleanup at our community. - Klein Muçi (talk) 15:44, 20 April 2020 (UTC)[reply]
@Klein Muçi: that feature has been requested for over 10 years, with noone working on it - see: phab:T21485. — xaosflux Talk 15:49, 20 April 2020 (UTC)[reply]
@Xaosflux: I see... So that cleanup I mentioned below is really needed. Any idea of what would be the best approach to this? - Klein Muçi (talk) 16:00, 20 April 2020 (UTC)[reply]
@Klein Muçi: Is a query such as this what you are looking for: quarry:query/44155? — xaosflux Talk 16:13, 20 April 2020 (UTC)[reply]
Umm, that one is broken - but something like that (I don't have time to work on it more this min). — xaosflux Talk 16:14, 20 April 2020 (UTC)[reply]
Maybe its not try that @Klein Muçi:. — xaosflux Talk 16:18, 20 April 2020 (UTC)[reply]
quarry:query/44156 has all the namespaces, though user_talk is expected I think. — xaosflux Talk 16:24, 20 April 2020 (UTC)[reply]
@Xaosflux: I mean, maybe? I'm not sure what namespaces should I search for. Giving it a quick thought, I think there is not one namespace that wouldn't be affected by that problem. And it's the first time I'm seeing this quarry tool so I can't say much about it. Besides, I'm in a dilemma if I can automatize the process or not. Maybe there are unassociated pages that are needed to be like that (like the examples you mentioned above)? (Besides x2, how deep should the quarry go? Considering that subpages also have talkpages.) As I said, I haven't researched much the actual situation in SqWiki because I thought we had a reason for keeping them, since that has been the case for over 10 years now. It's only now I'm considering the said cleanup and I was wondering if you've ever dealt with a situation like this before here, but looks like you haven't. [Text written before an edit conflict.] - Klein Muçi (talk) 16:27, 20 April 2020 (UTC)[reply]
Basically, there is no way to find these on-wiki, so you have to result to a database analysis, which is what quarry does. The primary example of an appropriate talk page without a matching page would be "user talk". If you do a database review, you could exclude pages with "/" in their titles to eliminate most subpages. — xaosflux Talk 17:29, 20 April 2020 (UTC)[reply]
@Klein Muçi: Regarding "the classical, unenhanced way of deleting pages" and "When you delete pages, there are no facilities like these." - in fact, there are. When an admin clicks the "Delete" tab, selects a reason and clicks Delete page, they then get presented with this screen, note the full-width pink box. This box is suppressed if there is no talk page, or when deleting a talk page or a User: page. --Redrose64 🌹 (talk) 10:21, 21 April 2020 (UTC)[reply]
@Redrose64: I see... Actually, we don't have that system message in our wiki. We have it set up differently. Check for yourself here. And even like this there are still some extra steps compared to what happens with move pages. But you offered me a solution for the future. If I ever am able to do that clean up I've mentioned below, I'll set that system message up like in here, hoping we won't get in a similar situation in the future again. Thank you! - Klein Muçi (talk) 11:40, 21 April 2020 (UTC)[reply]
There is also a tool to find orphan talk pages in all wikis. Stryn (talk) 16:28, 20 April 2020 (UTC)[reply]
@Stryn: Hey, that's actually pretty good! Is there a way to see the whole list? The limit of 100 results doesn't really do it for me in this case, as you've may have already thought of. Other than that, I think this is perfect for this situation, assuming the list is correct (I tried it at Sq.Wikiquote and it found out only 15 pages. The project is really young but still the low number baffled me.) - Klein Muçi (talk) 17:16, 20 April 2020 (UTC)[reply]
@Xaosflux: Do you think the tool above would be an easier solution for me? - Klein Muçi (talk) 18:01, 20 April 2020 (UTC)[reply]
Did you try downloading the data from the query I ran, #44156? You should be able to do easy text filters to remove user talk and pages with "/". — xaosflux Talk 18:21, 20 April 2020 (UTC)[reply]
@Xaosflux: No, I didn't. I've never worked with that tool before and this looked like an easier solution from the tech side at first glance. But I'll give both a try later then.
Thank you all! :) - Klein Muçi (talk) 18:37, 20 April 2020 (UTC)[reply]
On that link, it is already written and ran, just click the download link at the bottom :D — xaosflux Talk 18:43, 20 April 2020 (UTC)[reply]

@Xaosflux: Haha, yes but I meant learning a bit what to do with it without disturbing you. In what format am I supposed to download it from? And where am I supposed to feed it? Maybe AWB? Or am I off track with the logic here? - Klein Muçi (talk) 18:48, 20 April 2020 (UTC)[reply]

It really depends on what you want to do with it, you probably shoudn't just feed it in to some sort of delte-o-matic. — xaosflux Talk 19:29, 20 April 2020 (UTC)[reply]
@Xaosflux: I see. Well, I'll download it and maybe discuss it with our VP then. Thank you! :) - Klein Muçi (talk) 19:32, 20 April 2020 (UTC)[reply]
@Xaosflux: I'm seeing the downloaded file now. Was it supposed to show user talk pages? Because I'm seeing a lot of those at that file. I thought you had done your quarry in such a way as to remove those? Or have I misunderstood you? - Klein Muçi (talk) 00:27, 21 April 2020 (UTC)[reply]
The namespace is in there, if you want to do large bulk things, you're going to have to use a text editor, maybe a spreadsheet program, to filter and get the data in just the way you need it - using those is a bit beyond the scope of VPT though! Someone may have some recommended utilities, I usually just use shell scripts or text editors. — xaosflux Talk 01:14, 21 April 2020 (UTC)[reply]
@Xaosflux: Oh, so it's not a mistake. How about subpages? Are those counted too? Just want to know with what information I'm dealing with. I'll try to use Excel for sorting/filtering data.
Can you run the same quarry for SqQuote when you do find time? - Klein Muçi (talk) 01:22, 21 April 2020 (UTC)[reply]
Excel should help alot here! quarry:query/44178 is for sqq. — xaosflux Talk 01:32, 21 April 2020 (UTC)[reply]
And this query is somewhat 'simple' it does include subpages (they have a "/" in them) but it doesn't exclude them if there is a "root" companion page for example. — xaosflux Talk 01:34, 21 April 2020 (UTC)[reply]
@Xaosflux: Thank you! I hope that settles it. :) - Klein Muçi (talk) 01:35, 21 April 2020 (UTC)[reply]
Klein Muçi, you can use quarry to filter out user talk pages and subpages. See quarry:query/44190. SD0001 (talk) 16:10, 21 April 2020 (UTC)[reply]
@SD0001: thank you! I had never seen this tool before so all this is new info for me. Can you run one for SqWiki too? So I have it, just in case. I did get more responses than I hoped for in this question and I'm delighted for that. Now if only I'd get some suggestions on how to handle the situation practically regarding the high number of orphaned talk pages that we have... - Klein Muçi (talk) 16:22, 21 April 2020 (UTC)[reply]

18:44, 20 April 2020 (UTC)

I discovered this via the help desk. Apparently, for U.S. counties, if you search Google for how many coronavirus cases in a particular county, the sidebar pulls data from Wikipedia. That's par for the course. However, for some reason for counties in Wisconsin, it is not pulling the number of cases in the county as reported in our article, instead it is pulling from the column for cases per 100,000 people, then dropping the decimal. So, for example, La Crosse County, which has 25 cases, or 21.2 cases per 100,000 people, shows up on Google as having 212 cases. Milwaukee County, which has 2,191 cases or 229.6 per 100,000 people, shows up on Google as having 2,296 cases, and so on. As far as I can tell, that's the only state where Google is doing that.

For comparison, I looked at 2020 coronavirus pandemic in Michigan, which also lists how many cases per 100,000 people. The Wisconsin chart does have some columns the Michigan chart doesn't have, but overall they are similar. However, Google is correctly pulling the Michigan data from the cases column instead of the cases per 100,000 population.

I know we don't control Google, but can anyone figure out why they are misreading the Wiconsin chart so badly? Perhaps if we can figure out what's wrong we can fix it on our end. ~ ONUnicorn(Talk|Contribs)problem solving 19:36, 20 April 2020 (UTC)[reply]

We can't, it's the scraper they have built that is bugged. I don't know how we can leave them a note, especially that they don't even have the small "Suggest an edit" text at the bottom of COVID-related snippets. --qedk (t c) 20:58, 20 April 2020 (UTC)[reply]
@MPinchuk (WMF): Agathoclea (talk) 21:38, 20 April 2020 (UTC)[reply]
ONUnicorn, QEDK there is a mailinglist: https://lists.wikimedia.org/mailman/listinfo/covid-19-statsTheDJ (talkcontribs) 08:42, 21 April 2020 (UTC)[reply]
Thanks, TheDJ :) @ONUnicorn, as noted above, I'm pretty sure this is an issue with how Google's parser is reading the table, so I've given them a heads up about this. The Google team working on this feature is subscribed to the above mailing list, so it's a good way to directly report any issues you're seeing. I've also been monitoring various template and WikiProject talk pages for feedback/bug reports and sending out to the list on a regular basis, so pinging me works, too! MPinchuk (WMF) (talk) 13:38, 21 April 2020 (UTC)[reply]
Did not know that we had a mailing list, thanks! --qedk (t c) 14:57, 21 April 2020 (UTC)[reply]

Wisconsin Corona Virus cases incorrect on page that automatically comes up when googled

The counts by county for Wisconsin when corona virus cases is googled is so incorrect. When I go into the core they are correct but the page that comes up directly when googled is so off it is scary — Preceding unsigned comment added by 2601:40d:8001:ca60:8d21:8ed6:8014:b87e (talk) 18:39, 21 April 2020 (UTC)[reply]

You might want to post this at Talk:2020 coronavirus pandemic in Wisconsin. This page is for discussing technical issues with Wikipedia. -- LuK3 (Talk) 18:41, 21 April 2020 (UTC)[reply]
Are you talking about the "Cases overview" that comes up in the Google Knowledge Graph when you search wisconsin coronavirus cases? If so we have no control over Google, and, although they always claim their information has come from us, it quite often has either come from someone else, or has come from us but is out-of-date. If you want to complain about Knowledge Graph, click the "About this data" link underneath it and a "Send Feedback" link will pop up. ‑ Iridescent 18:46, 21 April 2020 (UTC)[reply]
See just a few sections up - this has been reported to Google. Sam Walton (talk) 18:52, 21 April 2020 (UTC)[reply]
 Done It appears to be fixed now. If you notice any other discrepancies between the Google stats card and Wikipedia, feel free to ping me or report them to this Wikimedia mailing list MPinchuk (WMF) (talk) 23:55, 21 April 2020 (UTC)[reply]

Word counter

If anyone wants to test out a new tool, I made a new word counter in ~3 hours: https://wordcount.toolforge.org Upcoming features include character count (maybe?), API (maybe??) Honestly though, I made this tool only because I saw that ArbCom currently recommends using a commercial advert-filled website and I just wanted to get some OSS into the mix. This pandemic has left me with too much free time. Facepalm Facepalm --qedk (t c) 20:54, 20 April 2020 (UTC)[reply]

Editors who would like to do word counting on-wiki, and don't mind a little JavaScript in their lives, can import User:Dr pda/prosesize. I think there is a similar script as well, but I can't put my finger on it at the moment. – Jonesey95 (talk) 23:16, 20 April 2020 (UTC)[reply]
Just adding that prosesize is pretty useful for on-wiki pages, but it counts entire pages on the site only, so there's that. --qedk (t c) 14:59, 21 April 2020 (UTC)[reply]

Loading gadgets on another Wikipedia

English Wikipedia has the gadget "(D) Reference Tooltips: hover over inline citations to see reference information without moving away from the article text (does not work if "Navigation popups" is enabled above)" and it's very useful. Our MediaWiki:Gadgets-definition sets this up using the line

ReferenceTooltips[ResourceLoader|default|type=general|dependencies=mediawiki.cookie,jquery.client]|ReferenceTooltips.js|ReferenceTooltips.css

so there are two actual gadget files and some dependencies. But this is not available at Wicipedia Cymraeg, which has no gadgets. I've got the idea that I might put appropriate code into cy:User:Redrose64/common.js and cy:User:Redrose64/common.css. The CSS file is easy - the line would be

@import "//en.wikipedia.org/w/index.php?title=MediaWiki:Gadget-ReferenceTooltips.css&action=raw&ctype=text/css";

but I'm no JavaScript expert, so would this

mw.loader.using( [ 'mediawiki.cookie', 'jquery.client' ] ).done(
  mw.loader.load( '//en.wikipedia.org/w/index.php?title=MediaWiki:Gadget-ReferenceTooltips.js&action=raw&ctype=text/javascript' );
)

work? --Redrose64 🌹 (talk) 21:31, 20 April 2020 (UTC)[reply]

Are you an interface admin of your wiki? Adithyak1997 (talk) 10:30, 21 April 2020 (UTC)[reply]
No, he isn't (and I don't know why you're asking) but anyway you can see the list of intadmins here. --qedk (t c) 15:01, 21 April 2020 (UTC)[reply]
That's for English Wikipedia (where I was an intadmin, until that was unbundled from sysop a couple of years back); but I'm asking about Wicipedia Cymraeg, for which the list of intadmins is much shorter. I rather think that if Llywelyn2000 (talk · contribs) wanted cy.wp to have gadgets, he would already have set up something at cy:MediaWiki:Gadgets-definition. Hence my desire to put code into my personal common.js --Redrose64 🌹 (talk) 20:35, 21 April 2020 (UTC)[reply]
That isn't the correct syntax. You want:
mw.loader.using( [ 'mediawiki.cookie', 'jquery.client' ] ).then(function() {
  mw.loader.load( '//en.wikipedia.org/w/index.php?title=MediaWiki:Gadget-ReferenceTooltips.js&action=raw&ctype=text/javascript' );
});
SD0001 (talk) 15:26, 21 April 2020 (UTC)[reply]
User:Redrose64 - go ahead; that would simplify the editing / adding of references. Thanks! Llywelyn2000 (talk) 05:21, 22 April 2020 (UTC)[reply]

I have gone through your query in Village Pump related to the gadget. Following is my suggestion:

1)Paste following lines in this page. This page is currently protected.

ReferenceTooltips[ResourceLoader|default|type=general|dependencies=mediawiki.cookie,jquery.client]|ReferenceTooltips.js|ReferenceTooltips.css

As you mentioned there regarding dependencies, I think it will be resolved once you add above code.Please check the last point mentioned here.
2)Paste the contents of this file in your wiki just by changing 'en' to 'cy' in the url. The content model also should not change. I think this also requires higher privileges.
3)The same needs to be done for this too.
4)Now, since gadgets have not been run on your wiki, I think something needs to be done related to ResourceLoader which I have not got info from anywhere. My suggestion is to try the above solution and if it goes wrong, I think it can easily be resolved by rolling back. Sorry if I have been too detail. Adithyak1997 (talk) 13:28, 21 April 2020 (UTC)[reply]
You have totally misunderstood me. Wicipedia Cymraeg is not "my wiki", nor am I an administrator there. I'm just an ordinary Joe who occasionally edits there, but when doing so misses the rich variety of gadgets provided here at English Wikipedia (which is my home wiki). So my question at VPT was carefully constructed in such a way that it should have been clear that what I want to do is import the en.wp gadget for my own personal use at cy.wp. You should also have noticed that I had already worked out which files were necessary. --Redrose64 🌹 (talk) 18:40, 21 April 2020 (UTC)[reply]

Change Graph:Lines default axis, labels, legend, and grid color to #54595d grey for color-inversion friendliness

{{Graph:Lines}} is maintained @ mw:Template:Graph:Lines, and I've naively made some attempts to do this myself @ mw:Template:Graph:Lines/sandbox, and made a request a few weeks ago @ mw:Template_talk:Graph:Lines, but to no avail. This would bring Graph:Lines in line with mw:Module:Graph. Could someone more knowledgeable be able to help?   ~ Tom.Reding (talkdgaf)  21:40, 20 April 2020 (UTC)[reply]

Tom.Reding, what is color inversion friendliness if i may ask ? —TheDJ (talkcontribs) 08:47, 21 April 2020 (UTC)[reply]
TheDJ, a fancy way of saying "shows up against a dark background". ‑ Iridescent 09:07, 21 April 2020 (UTC)[reply]
Tom.Reding, ah you mean to use it with something like a darkskin gadget or userscript etc ? —TheDJ (talkcontribs) 15:38, 21 April 2020 (UTC)[reply]

Editor Help

Whenever I press publish it doesn't work then I have to reload and press publish. Then it works. Video link.
Thanks,
Andrew nyr (talk, contribs) 03:16, 21 April 2020 (UTC)[reply]

User:Andrew nyr, I see you're using the visual editor. Does it sound like the problem that User:RoySmith reported in #Save edit in Visual Editor, it's asking to save again?? Whatamidoing (WMF) (talk) 18:12, 21 April 2020 (UTC)[reply]
Yeah, that looks exactly like what I'm seeing. Love the video technique :-) -- RoySmith (talk) 18:31, 21 April 2020 (UTC)[reply]

Diff: One intermediate revision by 12 users not shown

Can we get the Diff program to produce a more reasonable message than "One intermediate revision by 12 users not shown" when an intervening revdel is involved, as seen in this diff? (Possibly related discussions in archives: 4, 36, and 83.) Mathglot (talk) 05:19, 21 April 2020 (UTC)[reply]

Probably the second archive link is not related to this. But for the first archive link, I cannot reproduce the issue using what the OP described. (Open the the page history and compare the revisions that the OP described). It may have been fixed or maybe there was never a problem there. The issue in the last archive link is a different issue and it has been fixed since. Also what do you think the "x intermediate revision" message should say?– Ammarpad (talk) 10:26, 21 April 2020 (UTC)[reply]
Problem reproduced on IE 11, Opera 67, Chrome 80, and Firefox 69. I'm not sure what it should say, but it shouldn't be a logical impossibility. Maybe, "83 revisions by 12 users not shown"? If it doesn't have that information, then maybe, "Many revisions by...". Mathglot (talk) 11:55, 21 April 2020 (UTC)[reply]
uselang=qqx shows the message is MediaWiki:Diff-multi-otherusers called with (diff-multi-otherusers: 1, 12). Here at enwiki we could test whether $1 < $2 and add a possible explanation like "One intermediate revision by 12 users not shown. Some revisions may be hidden." But we don't know whether there are hidden revisions when $1 ≥ $2, and it could be confusing to only display the added text when $1 < $2. We don't have access to the real number of revisions without a MediaWiki change which should be requested at Phabricator:. PrimeHunter (talk) 13:55, 21 April 2020 (UTC)[reply]
I found that the real count is intentionally hidden to tackle unrelated issue. There were concerns of leaking "statistics of deleted revisions" too, so the real count for such cases is unlikely to be restored. Also note, what I cannot reproduce is the issue that you linked from this archive. I think I made that clear. – Ammarpad (talk) 14:46, 21 April 2020 (UTC)[reply]
Doesn't seem worth a mw software change, does it, so I guess we'll just have to live with it. Thanks, all. Mathglot (talk) 18:03, 21 April 2020 (UTC)[reply]

Text direction quirk/problem in wikitext editor

I ran into this and thought to mention it here; I ran into it in the Zhivopisny Bridge article trying to flesh out the cite which reads:

[https://www.ad-hoc.co.il/ ], 14 January 2008. Retrieved 29 January 2008.

.I went to the linked page, cut the bold text below the image there (reading פרויקטים בביצוע אד הוק בניה, עבודות גובה ושיקום מבנים, and pasting it here worked OK; maybe it's peculiar to wikilink-labeling), tried to paste it in to label the wikilink there, and got unexpected results in the edit window. I see similar unexpected results doing that here. The text I'm trying to paste in is Hebrew, with R->L direction. I'm not literate in Hebrew, I'm aware of the direction difference, but the results I get here are still unexpected. Just thought I would mention it. I'm using Win10 and Google Chrome if that matters. Wtmitchell (talk) (earlier Boracay Bill) 12:08, 21 April 2020 (UTC)[reply]

"unexpected results" is vague. Your url above is rendered because there is a newline before ]. Just remove it. Browsers may assume that numbers after right-to-left text is part of that text and should also be displayed right-to-left. You can insert a left-to-right mark &lrm; before the numbers. With &lrm;:
פרויקטים בביצוע אד הוק בניה, עבודות גובה ושיקום מבנים‎, 14 January 2008. Retrieved 29 January 2008.
Without &lrm;
פרויקטים בביצוע אד הוק בניה, עבודות גובה ושיקום מבנים, 14 January 2008. Retrieved 29 January 2008.
My Firefox renders "14" in the right place with &lrm; but to the left of the link without it. When it reaches "January" it knows the right-to-left text has ended. It's only to the left in the rendered page for me. In the source text I see it in the right place but browsers may vary. The produced html is in the original order so it all depends on the browser. PrimeHunter (talk) 13:22, 21 April 2020 (UTC)[reply]

Is there a way to find (free) files with more than one (not deleted) version in the history?

That is, enwiki files which like the Commons file en:File:Salt Water (2515735668).jpg have more than one file in their file history. Jo-Jo Eumerus (talk) 16:22, 21 April 2020 (UTC)[reply]

quarry:query/44209. —Cryptic 23:34, 21 April 2020 (UTC)[reply]
Thanks. Jo-Jo Eumerus (talk) 09:01, 22 April 2020 (UTC)[reply]

Detecting references with a bare URL and a ref name

Hi! I just realized that my bot does not edit bare-URL references with a reference name.

The search regex I use is:

((?<=(?<!\?)url ?= ?)|(?<=[^[]\[)|(?<=<ref>)|(?<=\* ?))(http:\/\/)?(www\.)?example\.com(?!\.)

This works to find urls in cases like:

  • *http://example.com/home (optional space between asterisk and url)
  • [http://example.com/home
  • <ref>http://example.com/home
  • url=http://example.com/home (optional space between "url" and "=" and also between "=" and the url)

However, it does not work on bare-URL references with a refname.

I am not sure what sort of thing to do here to guarantee I will capture all somewhat reasonably formatted refs. I need to be able to get all of these, at least:

  • <ref name="many-words-separated-by-hyphens">http://example.com/home
  • <ref name="words with spaces">http://example.com/home
  • <ref name=name>http://example.com/home
  • <ref name=words-with-hyphens-and-no-quotes>http://example.com/home

I think I can handle most of the refs with just a lookbehind for <ref name=.*> (demo at https://regexr.com/530uv).

However, I would like to ask for a review of this — would this cause any problems, and are there any common formats of bare URL references this would not get?


BTW here is what the regex would be with the new lookbehind:

((?<=(?<!\?)url ?= ?)|(?<=[^[]\[)|(?<=<ref name=.*>)|(?<=<ref>)|(?<=\* ?))(http:\/\/)?(www\.)?example\.com(?!\.)

(Another regex I worked on for a little was <ref name="[^"]*">http:\/\/example\.com (see https://regexr.com/530up), but I don't actually see any reason it is better than the other and it only works on refs with quotes.)

Thanks! DemonDays64 (talk) 23:16, 21 April 2020 (UTC) (please ping on reply)[reply]

Automatic redirects from non-breaking hyphens in page titles?

I was interested in doing a search-and-replace of 2019–20 coronavirus pandemic to replace all the hyphens in mentions of "COVID-19" and "SARS-CoV-2" with non-breaking hyphens[7] to prevent unwanted line breaks in the middle of those words. However, doing so broke a ton of wikilinks when I tried to preview the page, since it seems the software doesn't automatically redirect from non-breaking hyphens to breaking hyphens in page names, the same way it does for some capitalization differences. I can see this being an issue at other pages where one might want to make a similar fix, and even if I had fixed things manually at 2019–20 coronavirus pandemic, the issue would come back as the page evolves. So is there any chance we could do something to the automatic redirect software to take care of this? {{u|Sdkb}}talk 00:41, 22 April 2020 (UTC)[reply]

"COVID-19" and "SARS-CoV-2" are using hyphen-minuses instead of hyphens, so really titles with non-breaking hyphens and normal hyphens should redirect to the title with minus-hyphens. For example COVID-19 vaccine exists, but COVID‐19 vaccine and COVID‑19 vaccine do not. We could fix this by using a bot that would create the appropriate redirects and move the page to the version using hyphen-minuses when needed.BrandonXLF (talk) 21:23, 22 April 2020 (UTC)[reply]
Some other solution would be better because using something other than a hyphen will confuse editors and break searching for "COVID-19" with the normal hyphen. Johnuniq (talk) 00:03, 23 April 2020 (UTC)[reply]
@BrandonXLF and Johnuniq: yikes, hyphen-world is complicated haha. Trying to get this straight: so hyphen-minus and hyphen are separate things. And the page titles are all at hyphen-minus. Is hyphen-minus also the grammatically correct option? Is hyphen-minus what you get when you press the button on the keyboard (and thus what people will be searching for)? A bot to fix this sounds fine, if it can't be done more directly than that. Do we know what allows redirects via miscapitalization to work, and could we replicate whatever process does that? This seems more analogous to that situation than anything else. {{u|Sdkb}}talk 07:58, 24 April 2020 (UTC)[reply]
Hyphen and hyphen-minus are the same thing. I wouldn't think there is anything a bot can fix. Johnuniq (talk) 08:01, 24 April 2020 (UTC)[reply]
@Johnuniq and Sdkb: They're different. Hyphen-minus (-) is the hyphen on the keyboard and can be used as a hyphen or as a minus sign whereas a hyphen (‐) is only used as a hyphen and is not used as much and its use is discouraged here on Wikipedia. BrandonXLF (talk) 19:28, 24 April 2020 (UTC)[reply]
See hyphen-minus and hyphen. --Redrose64 🌹 (talk) 20:11, 24 April 2020 (UTC)[reply]
@BrandonXLF, Redrose64, and Johnuniq: Okay, so what we need then is a way to automatically redirect any title with a hyphen or non-breaking hyphen to a hyphen-minus. Any ideas on how to do that? I can't figure out how the software does it for capitalization — trying to go to foobar (with a lowercase f) just goes to Foobar (the page title) without even showing there was any redirect. {{u|Sdkb}}talk 01:14, 25 April 2020 (UTC)[reply]

I would like to have a random link to a mathematics page, however most Mathematics pages aren't pages in the Category:Mathematics but in one of its subcategories. So http://wikipedia.org/wiki/Special:RandomInCategory/Mathematics doesn't work. Do you have any idea on how I could generate a link that leads to an article recursively?

PS: I am very new to wikipedia (joined yesterday) and I am glad for any advice you can give me :D

TheFibonacciEffect (talk) 08:11, 22 April 2020 (UTC)[reply]

You can use the subcategory as the parameter, or am I missing something? It won't check recursively if that's what you're asking. --qedk (t c) 08:16, 22 April 2020 (UTC)[reply]
There's no way to generate such a link using just wiki-markup. But you can employ the services of User:AnomieBOT/RandomPage for this. Copy paste one of the sample usages to a page in your userspace, put |categories = {{SUBCATS|Mathematics|10}}. The bot should edit the page periodically, adding a link to a random article. SD0001 (talk) 12:05, 22 April 2020 (UTC)[reply]
phab:T48918 is exploring if a "recurse" option can be used for RandomInCategory. — xaosflux Talk 13:19, 22 April 2020 (UTC)[reply]
Thank you for your response. To clarify: I would like to have a link that I can put into my browser, similar to the Special:RandomInCategory link, so every time I open a new tab, I get to see a mathematics page and learn something new and interesting or edit an article I at least know something about.
Thanks you TheFibonacciEffect (talk) 19:03, 22 April 2020 (UTC)[reply]

ƀ

I think ƀ should redirect to Ƀ but a bot has deleted the page and now I cannot fix it. — Preceding unsigned comment added by 2a02:a457:9497:1:f938:499e:52bc:cb4c (talk) 14:48, 22 April 2020 (UTC)[reply]

I think this is related to the "Future changes" item in this Tech News. I don't know what the correct action is, but someone here (Anomie?) probably does. See also Special:Contributions/Maintenance script. – Jonesey95 (talk) 15:30, 22 April 2020 (UTC)[reply]
It was renamed by a maintenance script not bot. From the migration plans, it seems even if you recreate it now, the script will move it again. But more importantly, the title will be completely inaccessible when the Unicode upgrade take place, so there's really no point in recreating it since it'll soon not be useful anymore. – Ammarpad (talk) 17:07, 22 April 2020 (UTC)[reply]
I think the redirect shouldn't have been removed before WP starts actually using the new Unicode version. How long will the problem persist? Days? Months? — Preceding unsigned comment added by 2a02:a457:9497:1:f938:499e:52bc:cb4c (talk) 18:01, 22 April 2020 (UTC)[reply]
Special:Search/ƀ already goes directly to Ƀ, and {{uc:ƀ}} produces Ƀ. A few links at Special:WhatLinksHere/ƀ could be updated but it doesn't seem important. A page at ƀ becomes inaccessible after the link starts going to Ƀ, so the moves without leaving a redirect are deliberately made before this. PrimeHunter (talk) 18:46, 22 April 2020 (UTC)[reply]

Is there a way to add other cite template(s) to the drop-down menu in the editing box?

This is more curiosity than anything else, but I could see it being useful. I know the "Templates" drop-down (next to "Named references") has the choices "cite web, book, news, journal". If I wanted to add one - like "cite AV Media", for example, is that possible? JimKaatFan (talk) 19:28, 22 April 2020 (UTC)[reply]

JimKaatFan, you mean the dropdown of RefToolbar 2.0. It's settings are here, so if you get people to agree on adding it, it can be added. —TheDJ (talkcontribs) 08:39, 23 April 2020 (UTC)[reply]
Thanks. I wasn't looking to force changes on the whole world, but I was wondering if there's a way to modify that dropdown just for myself. Can I change those settings you linked so that the changes only apply to me when I'm logged in? Or is that a technical impossibility? JimKaatFan (talk) 15:52, 23 April 2020 (UTC)[reply]

New revision of audio clip won't load?

I uploaded the different portion of a song at File:We Are Sex Bob-Omb clip.ogg, but it loads the same revision (but just trimmed down). Actually, I reverted back and forth, but the current 14-sec one plays something different from the one I intended. --George Ho (talk) 21:56, 22 April 2020 (UTC)[reply]

@George Ho: I reset the transcode on that, is it working the way you wanted now? — xaosflux Talk 01:17, 23 April 2020 (UTC)[reply]
Still loads the same clip, which is not the one I intended. The 14-second one is the one I wanted, but somehow the 21-second version is trimmed to 14 seconds. --George Ho (talk) 01:29, 23 April 2020 (UTC)[reply]
On mobile web and mobile app, I can hear the clip that I intended to use. At first, it was strange. The audio was the same while logged-out. Then it turns out that I had to clear the cache of one of my browsers. Now I was able to hear the audio. --George Ho (talk) 07:19, 23 April 2020 (UTC)[reply]

Cache issues

So, I'm trying to get the word count java-script thing, and I have to clear my cache, but I've tried many ways but it won't let me. Any advice?? TuorEladar (talk) 22:47, 22 April 2020 (UTC)[reply]

Grab protection timestamps with SQL

I've got two queries that list indefinitely protected redirects, quarry:query/44276 and quarry:query/44280, but I'd like to add a column that lists the timestamp of the protection. How would I modify the query so that it would do that? Anarchyte (talkwork) 09:05, 23 April 2020 (UTC)[reply]

I tried this, but it timed out after 30 minutes so it's probably not the solution.
SELECT CONCAT ("[[",page_title,"]]") AS 'Redirect', log_timestamp AS 'Timestamp' FROM page
JOIN page_restrictions ON page_id = pr_page
JOIN logging ON page_id = log_page
WHERE pr_type = 'edit'
AND pr_level = 'autoconfirmed'
AND page_namespace = 0
AND page_is_redirect = 1
AND pr_expiry = 'infinity' or pr_expiry is null
AND log_type = 'protect
Anarchyte (talkwork) 09:41, 23 April 2020 (UTC)[reply]
I haven't looked at your quarry queries (should have time later today), but five quick comments based only on what's here: try logging_logindex instead of logging; pasting into the Toolforge sql optimizer would have suggested that; you need parens around AND (pr_expiry = 'infinity' OR pr_expiry IS NULL); this will entirely omit pages without a corresponding log entry (like if the page was moved, instead of being protected at its current title) - try a LEFT JOIN logging_logindex and allow log_type IS NULL as an alternative to 'protect'; WP:RAQ should be your first impulse for this sort of question, not WP:VPT. —Cryptic 15:46, 23 April 2020 (UTC)[reply]
@Cryptic: Thanks for the response. I was using these as replacements for Wikipedia:Database reports/Indefinitely semi-protected redirects but DannyS712 got the script back up and running, so this problem is moot. I didn't know WP:RAQ existed, so I'll keep it in mind for the future. Anarchyte (talkwork) 15:50, 23 April 2020 (UTC)[reply]

Hi all

I'm currently working on Wikiproject Covid-19 and collating information from UN staff members who are experts on different aspects on COVID-19 e.g agriculture, maternal health , education etc. One thing that I'm finding really difficult is that this information is best displayed in tables but VE is disabled in the Wikipedia: namespace. My understanding is this is because all Wikiprojects on English Wikipedia are in the Wikipedia: namespace which is shared with many other functions and some of these break with VE.

It is not a realistic solution to ask experts and new users to learn how to edit tables in source editor and having tables easily editable would be extremely helpful in many Wikiprojects. I would like a template that I can put at the top of these pages in Wikiprojects which would say something like:

'Click here to edit this page with Visual Editor' and the link would simply add ?veaction=edit to the end of the URL

Would creating a template to do this be possible?

Thanks

John Cummings (talk) 10:39, 23 April 2020 (UTC)[reply]

We already have {{VEFriendly}}. I think the main reason VisualEditor is not linked by default in Wikipedia space is that it's often used for discussions, like this page. PrimeHunter (talk) 10:50, 23 April 2020 (UTC)[reply]
Oh this is perfect, thanks so much :) John Cummings (talk) 11:07, 23 April 2020 (UTC)[reply]
I've made a couple of suggestions for improvements to the template on the talk page (the grammar is weird) https://en.wikipedia.org/wiki/Template_talk:VEFriendly
Thanks
John Cummings (talk) 11:18, 23 April 2020 (UTC)[reply]

Why doesn't the contribs page for IPs show the Xtools Edit count page (example)? Also, the "Edit summary search" and "Articles created" links are missing that registered users have.

Xtools stats are useful for analyzing problematic IPs, so the link would be useful to have there. Is there a technical barrier? --Pudeo (talk) 13:44, 23 April 2020 (UTC)[reply]

No technical barrier, all that would need to happen is an admin making an edit to MediaWiki:Sp-contributions-footer-anon. * Pppery * it has begun... 13:59, 23 April 2020 (UTC)[reply]
Thanks. I dropped a suggestion on that template's talkpage. --Pudeo (talk) 06:44, 24 April 2020 (UTC)[reply]

Solve issue with date format in Template:Graph:Chart

Hi. I would like to ask for some technical help regarding an issue/bug with the Template:Graph:Chart. In a nutshell, when using the date format for the x-axis in combination with the line graph type, everything (almost, see below) goes fine; however in combination with the rect graph type (i.e. bar charts) the dates are not formatted properly and are shown in Unix date format. Furthermore, the date format – for any graph type, line, rect, etc – appears to be incompatible with the choice of xMin and xMax values, again for the x-axis. I would suggest that it's better to direct helpers to the existing discussion in the template talk page, where there are also graphical examples of the issues, and per WP:MULTI. Thanks! --Ritchie92 (talk) 08:56, 24 April 2020 (UTC)[reply]

Safari (iPadOS) screws up when rendering Wikipedia (desktop)

I had noticed this issue before, but for some reasons, I didn’t bother to ask about this. Anyways, the issue is this:

  1. Letters are misaligned, so sometimes words will overlap
  2. If there’s a section title which should fit in one line in the navigation box, but has two words, Safari renders it in 2 lines
  3. The issue gets resolved when reloading the page

Device: iPad (2017)
OS: latest iPadOS (no matter what I update it to, the bug stays)
When it began: I’ve forgotten (it’s been so long)
Has anyone else had to deal with this?
RedBulbBlueBlood9911 (talk) 10:35, 24 April 2020 (UTC)[reply]
Post-posting comment: I noticed that this issue is only on Wikipedia, and that it usually happens between an ordinary word and a linked word. By the way, I don’t think different browsers can fix this (Apple demands that all browsers use WebKit, and I think I noticed this issue on Chrome on an iPhone 7) - RedBulbBlueBlood9911 (talk) 10:40, 24 April 2020 (UTC)[reply]

Request, copy Template:Portal_navigation from Meta to English Wikipedia, 'Lua error in package.lua at line 80: module 'Module:Is rtl' not found.'

Hi all

I'm working on usability for a Wikiproject and want to use https://meta.wikimedia.org/wiki/Template:Portal_navigation however I don't understand how to copy it into English Wikipedia, I tried to copy it to Wikidata and it didn't work and someone had to fix it. If anyone knows how to copy it across I'd really appreciate it.

Thanks very much

John Cummings (talk) 10:42, 24 April 2020 (UTC)[reply]

Update: I copied it across to Wikipedia but I'm getting a big red error message 'Script error: No such module "Portal navigation".' any help would be really appreciated. John Cummings (talk) 13:14, 24 April 2020 (UTC)[reply]
The only thing Template:Portal navigation does is to invoke Module:Portal navigation so you need to copy meta:Module:Portal navigation at a minimum. I don't know whether other things are needed. PrimeHunter (talk) 13:19, 24 April 2020 (UTC)[reply]
Thanks @PrimeHunter:, I've done this but now I get the big red error message 'Lua error in package.lua at line 80: module 'Module:Is rtl' not found.' John Cummings (talk) 13:30, 24 April 2020 (UTC)[reply]
OH MY GOD I MADE IT WORK :) John Cummings (talk) 13:41, 24 April 2020 (UTC)[reply]

Whenever I receive notifications for either https://en.wikipedia.org/wiki/James_Gordon_(British_Army_officer,_died_1783) or https://en.wikipedia.org/wiki/Cyril_Lloyd_(British_Army_officer) the second part of the link is not live, but black. It's just annoying, rather than serious, but can anything be done? I have to search for the pages instead of simply linking to them. Thanks. Anne (talk) 12:42, 24 April 2020 (UTC)[reply]

@Arbil44: "notifications" usually refers to the two icons to the right of your username at top of Wikipedia pages. Are you referring to notification mails? If so, do you have "HTML" or "Plain text" at "Email format:" at Special:Preferences#mw-prefsection-echo? "Plain text" sends a url as a plain text string, e.g. https://en.wikipedia.org/wiki/James_Gordon_(British_Army_officer,_died_1783). Your own mail software or webmail service may turn it into a clickable link and may omit an ending part in parentheses if it guesses it isn't part of the url. I don't know how much "HTML" encodes but maybe it will send a full clickable link so your mail software doesn't have to guess where the url ends. PrimeHunter (talk) 13:12, 24 April 2020 (UTC)[reply]
Thank you PrimeHunter. Your kind response is a bit too technical for me to fully understand, but I have several pages on my watch list and the links go straight to the relevant pages with no problem, but not the two I mentioned. Here is the latest notification which has caused me to ask the question now. I have emboldened the bit which isn't live in my email, but I suspect will go 'blue' here! Dear Arbil44,
The Wikipedia page James Gordon (British Army officer, died 1783) has been changed on 20 April 2020 by Necrothesp, see https://en.wikipedia.org/wiki/James_Gordon_(British_Army_officer,_died_1783)for the current revision.
To view this change, see https://en.wikipedia.org/w/index.php?title=James_Gordon_(British_Army_officer,_died_1783)&diff=next&oldid=944367213
Anne (talk) 13:32, 24 April 2020 (UTC)[reply]
@Arbil44: You were referring to notification mails as I suspected. Click Special:Preferences#mw-prefsection-echo and see what is says below "Email format:". If it says "Plain text" then change it to "HTML" and click "Save" at the bottom. PrimeHunter (talk) 13:40, 24 April 2020 (UTC)[reply]
Thank you again. Did as you suggested and it is already showing "HTML", so what now?! Keep it simple for me to understand please! Anne (talk) 13:47, 24 April 2020 (UTC)[reply]
@Arbil44: Watchlist mails are enabled at Special:Preferences#mw-prefsection-personal and not Special:Preferences#mw-prefsection-echo. They are not a part of the Echo notification system so the Echo preference has no effect on them. They always come as plain text (phab:T15303 from 2008 has a HTML request). It appears there is currently no solution for you, apart from changing to mail software which interprets the url string differently and includes a part in parentheses when it makes a link. But it's impractical to change mail software just for this. My own mail service (Danish webmail so not practical for others) includes the parentheses. There are also mail systems which don't make a clickable link for any url so you are better off than some. PrimeHunter (talk) 14:32, 24 April 2020 (UTC)[reply]
Thank you PrimeHunter. I appreciate knowing the situation. It was annoying me and I supposed it would be fixable, but now I know it isn't, then I will stop being irritated (or at least try)! Much appreciate your assistance, thank you. Anne (talk) 14:51, 24 April 2020 (UTC)[reply]
@Arbil44: It's not fixable for you. It could be improved in MediaWiki for some users but hasn't been done. Here are some technical details you are free to ignore since they don't help you. Wikipedia sends a mail like this:
The Wikipedia page Wikipedia:Village pump (proposals) has been changed
on 24 April 2020 by Shnizzedy, see
https://en.wikipedia.org/wiki/Wikipedia:Village_pump_(proposals) for the
current revision. 

To view this change, see
https://en.wikipedia.org/w/index.php?title=Wikipedia:Village_pump_(proposals)&diff=next&oldid=952863338

For all changes since your last visit, see
https://en.wikipedia.org/w/index.php?title=Wikipedia:Village_pump_(proposals)&diff=0&oldid=952863338
Note that none of it is clickable. It's all plain text. When mail software displays the mail, it may automatically change some of the text to clickable links according to an algorithm chosen by the software. The algorithm in your current mail software apparently says that a link ends if parentheses are reached, so https://en.wikipedia.org/wiki/Wikipedia:Village_pump_(proposals) becomes https://en.wikipedia.org/wiki/Wikipedia:Village_pump_(proposals) instead of https://en.wikipedia.org/wiki/Wikipedia:Village_pump_(proposals). Many mail services do this. The usual complaint we get is not about watchlists but that a link doesn't work if a user mails the url. In that situation there is a workaround which works in most mail software: Replace ( by %28 and ) by %29: https://en.wikipedia.org/wiki/Wikipedia:Village_pump_%28proposals%29. Our software doesn't do this when it sends plain text mails about watchlist changes, and we have no HTML option for this. In HTML you specify exactly where a url starts and ends so there is no ambiguity. Another frequent problem is url's ending in a period like Martin Luther King Jr. which has url https://en.wikipedia.org/wiki/Martin_Luther_King_Jr. Nearly all software will think that the period is ending a sentence and is not part of the url. This includes MediaWiki which produces: https://en.wikipedia.org/wiki/Martin_Luther_King_Jr. It works out OK in this case because we made a redirect at Martin Luther King Jr without the period. PrimeHunter (talk) 15:06, 24 April 2020 (UTC)[reply]

Well, I certainly did you the courtesy of reading, rather than ignoring, your response, but I followed you only up to a point! Anyway, if I have to change my software/provider/whatever in order to get clickable links, well that would be a bit unnecessary. Anyway, you have been most helpful in explaining that I must just put up with the situation. One small irony is that I put this page on my watchlist, while this discussion was taking place, and the emailed link to this thread does not take me to this thread! Don't worry about explaining that since this matter is now 'resolved'! Anne (talk) 15:15, 24 April 2020 (UTC)[reply]

HELP! with medical cases chart

Help. Someone please explain to me how Template:Medical cases chart renders the colors for recoveries etc. I am trying to add a "Recoveries" light blue color tab to the top of Template:2019–20 coronavirus pandemic data/United States/West Virginia medical cases chart but I don't see how to do that. I have been poring over the medical cases chart Template and the .doc Template subpage but I confess - not a coder, don't understand the template explanations. Please don't fix it for me, help me understand so I can do it myself and know how to do it next time. Shearonink (talk) 17:38, 24 April 2020 (UTC)[reply]

@Shearonink: Template:2019–20 coronavirus pandemic data/United States/West Virginia medical cases chart calls {{Medical cases chart}} with |recoveries=n to omit the legend for recoveries. Just remove it or change to |recoveries=y. PrimeHunter (talk) 18:26, 24 April 2020 (UTC)[reply]
Oh THANK YOU. Sheesh, I don't understand this present chart as well as the previous style, where everything was clearly labeled, row by row...... I wish the documentation or tamplet page itself explained that as clearly as you just did. Shearonink (talk) 18:48, 24 April 2020 (UTC)[reply]

Hi. I wanted to make a personal fork of the English Wikpedia's Featured and Good Articles (over 40,000 total). Uaing this ExtremeTech.com article as a guide, I downloaded a "Dump" file of the English Wikipedia (over 6,000,000 articles).

My questions are:

1. Is there an easy way to download only the Featured and Good articles (along with the necessary Template pages) rather than a dump of the whole encyclopedia? Or:


2. Is there an easy and efficient way to auto-delete articles; so that after setting it up, I can just delete the articles I don't want to keep? Such as a script or bot I can use, or point me in the direction of how to code one?

Thanks.--IBBishops (talk) 18:53, 24 April 2020 (UTC)[reply]

What you asking for is very similar to what Version_1.0_Editorial_Team has been doing. Ruslik_Zero 20:20, 24 April 2020 (UTC)[reply]

What you want to do is not terribly difficult, but it does require some programming skills. There is an API (Application Programming Interface which lets you do anything you can do via the web front end. There are client libraries which let you access this via a wide variety of programming languages. The basic gist would be to (have your program) look at Category:Good articles and Category:Featured articles and iterate over the entries, downloading the wikitext for each page and saving it locally. Figuring out which templates you need for each pages is only a little bit more complicated. You'll need to parse the wikitext and (possibly recursively) iterate over the transcluded templates, and download those via the api. Again, there's parsers available for a wide variety of programming languages. For somebody experienced in programming and familiar with the API and parser libraries, you could probably knock this off in an afternoon. If you're an experienced programmer but have never used these libraries, probably a couple of days to get used to how they work. -- RoySmith (talk) 21:26, 24 April 2020 (UTC)[reply]

Deleting talk page history

I noticed this editor somehow deleted their talk page history. I'm not sure how, but it's unfortunate because there were several warnings there. Thank you. Magnolia677 (talk) 21:51, 24 April 2020 (UTC)[reply]

Magnolia677 That is a consequence of archiving the history and text by moving it rather than copy-pasting. It is a valid, though rarer, form of archiving. You can see it in the first revision of that page. --Izno (talk) 22:07, 24 April 2020 (UTC)[reply]
@Izno: How unfortunate. Thanks! Magnolia677 (talk) 22:13, 24 April 2020 (UTC)[reply]
The history still exists, but is now found by looking at the history of the archive page. --Redrose64 🌹 (talk) 23:01, 24 April 2020 (UTC)[reply]

Question

I am not sure if this is the correct place for me to post this.

I am working on a tool that allows the review and answering of protected page edit requests. You can view the tool here. My question is how do we create an edit form so that reviewers do not have to switch between different tabs? Has someone already created one? And I am talking about the edit form that is kind of like the traditional Wikipedia edit form.Aasim 23:21, 24 April 2020 (UTC)[reply]

Awesome Aasim, since I don't know if you're doing this for fun or because you think we lack things to help us process protected edits, consider taking a look at User:Jackmcbarn/editProtectedHelper which can be used to directly answer requests and User:AnomieBOT/TPERTable (and related) which can be used to review requests. --Izno (talk) 23:36, 24 April 2020 (UTC)[reply]
(EPH could use an adopter if you are familiar with Javascript. --Izno (talk) 23:37, 24 April 2020 (UTC))[reply]
@Izno: I have taken a look at the tables (and even have the editprotectedhelper script) installed. I am working on this both so I can learn more about MW gadgets/APIs and so other editors can have an additional tool to use on Wikipedia. Yes, we have all these categories, but it takes a ton of clicking to go to the article, edit the article, go back to the talk page, go back to the list, etc. I hope to make this a little faster with this script. My question was: How do we make an "editform"-ish because I have tried making an editform-ish but it did not work. I think I figured it out. Thanks for the suggestion, though! I will definitely consider looking at the EPH script, but I think for now I will just work on my second tool :) Have you tried checking here? Aasim 08:33, 25 April 2020 (UTC)[reply]

Javascript: remembering dropdown menu selection?

Disclaimer: I'm asking on behalf of another wiki, so I understand if help may be limited in that regard. I am also a newbie with Javascript, so I only understand the barest of basics. I also apologise for the external links as I was unable to get the code to function properly on here.

On another wiki that I edit, StrategyWiki, there are two templates: Template:Control selector and Template:Control. Control selector takes in a string where each item is appended with a comma and renders it as a dropdown menu that functions as the switch, while Control works as the output and displays different content for the item selected in Control selector. You can observe how this works over on this test page: select an item from the dropdown menu on the right and observe how the icon to its left changes.

Control selector is dependent on Javascript in the wiki's Mediawiki:common.js file. Below is the code excerpted from the file:

Code excerpt from SW's MediaWiki:common.js
===Control Selector template===
*/
// Created by [[User:Prod]] with help from [[User:DrBob]]
// jQueryfied by [[User:Skizzerz]]
function selectControlSet( ) {
  $('span.controlOpt').each(function() {
    if ($(this).hasClass('control' + $('#control_selector_select').val()))
      $(this).show();
    else
      $(this).hide();
  });
}

function createControlSelector( ) {
  var controlDiv = $('#control_selector_inner');

  if (controlDiv.length == 0) return false;

  var ControlSelector = $('<select class="ControlSet" id="control_selector_select" />');
  ControlSelector.change(selectControlSet);

  var sysTexts = controlDiv.text().split(',');
  for (var i=0; i < Math.min(10, sysTexts.length); i++) {
    ControlSelector.append($('<option>', {value: i}).text($.trim(sysTexts[i])));
  }

  controlDiv.empty().append(ControlSelector);
  controlDiv.parent().css('display', 'block');
}
$( createControlSelector );

My question: is there a way for the dropdown menu item in Control selector to be remembered on future page visits? Using the test page above as a reference, is it possible to select the second (SNES) or third (GBA) item in the drop down menu, refresh the page, and have it stay on that dropdown menu item instead of reverting back to the first (PSX)? Doing some more research on the topic has led me to believe that user cookies may be involved somehow (this Stack Overflow thread mentions using jquery).

Your assistance is most appreciated. —Tenryuu 🐲 ( 💬 • 📝 ) 23:47, 24 April 2020 (UTC)[reply]

Have you tried using sessionStorage or localStorage? This basically puts the data on a user's computer. This is independent of whether a user is logged in or not. You can read more about it here: https://www.w3schools.com/html/html5_webstorage.asp. Hope this helps! Aasim 05:24, 25 April 2020 (UTC)[reply]
Awesome Aasim, that seems like it could be what we would use; thanks! Unfortunately I'm not much of a tech wiz, so I'm not sure how we'd incorporate GetItem and SetItem into the code. Would we add that somewhere around var sysTexts? —Tenryuu 🐲 ( 💬 • 📝 ) 18:01, 25 April 2020 (UTC)[reply]
A lot of the examples I've seen involved typing text and not so much selecting a dropdown menu item, so I'm not sure how much of that was applicable. :P —Tenryuu 🐲 ( 💬 • 📝 ) 18:03, 25 April 2020 (UTC)[reply]

Max page size

What is the largest possible size of a page? I once hit 2.1MB but the software can't let me go much farther. Is there a limit, and if so what is it? Thx Eumat114 formerly The Lord of Math (Message) 08:26, 25 April 2020 (UTC)[reply]

I don't think there is a maximum size for pages, but there are limitations on use of templates and processing time. One is the output of templates cannot exceed to 2MB (which is why pages with 500 plus citations using cite templates get into trouble). However, my understanding is that if it is simple text on the page it can be much bigger. —  Jts1882 | talk  08:37, 25 April 2020 (UTC)[reply]
The source size limit is determined by mw:Manual:$wgMaxArticleSize. We use the MediaWiki default 2 MB = 2×1024×1024 bytes = 2,097,152 bytes. The limit on post-expand include size is also 2 MB, meaning a page can reach up to 4 MB in total after template expansions. PrimeHunter (talk) 10:15, 25 April 2020 (UTC)[reply]

WikiProject assessment category names

I initiated Wikipedia:Categories_for_discussion/Log/2020_April_14#WikiProject_Vancouver thinking that it would be OK to have "WikiProject" in the parent category names but with no need to rename the detailed sub-cats, e.g. Category:WikiProject Vancouver articles by quality parenting Category:B-Class Vancouver articles.

The category pages look OK like that, but Template:WikiProject Vancouver has red links now.

If I insert a line

|ASSESSMENT_CAT = WikiProject Vancouver articles

it fixes those two red links, but requires all the sub-cats to be renamed.

If necessary I could rename the sub-cats under G6, but is there an easier fix, please?

Other projects somehow allow inconsistencies, e.g. WikiProject Women includes "WikiProject" in the quality categories but not the importance categories. I have no idea how that works through template:WikiProject Women.– Fayenatic London 19:53, 25 April 2020 (UTC)[reply]

Ah, for a workaround I will simply redirect the red links that offend the WikiProject template: Category:Vancouver articles by importance, Category:Vancouver articles by quality. – Fayenatic London 21:34, 25 April 2020 (UTC)[reply]

Do rcats (redirect categories) influence the search engine?

I have long known, or believed to know, that the redirect categories will influence how redirects are treated both in the search results and in the the drop-down list of suggestions in the search box. So that, all other things being equal, a redirect tagged with {{R from misspelling}} will for example be ranked lower than one tagged with {{R from alternative name}}. However, I've just realised that I don't know how I've learned that, and I don't find it in the documentation at mw:Extension:CirrusSearch. Is this really a thing or have I been deceiving myself? – Uanfala (talk) 21:54, 25 April 2020 (UTC)[reply]

The only weightage other than relevance is by article-class: MediaWiki:Cirrussearch-boost-templates. --qedk (t c) 22:05, 25 April 2020 (UTC)[reply]
Is "relevance" just a function of edit distance between the search term and the potential target? signed, Rosguill talk 22:18, 25 April 2020 (UTC)[reply]
It's a major part definitely, as any indexing and searching tool will be (I believe you are talking about Levenshtein distance?) - while not exactly (since there are also more factors) the same thing, at a certain level, yes. There's an entire blog on this if you want to take a look: https://www.elastic.co/blog/found-elasticsearch-from-the-bottom-up --qedk (t c) 22:25, 25 April 2020 (UTC)[reply]
No, "relevance" also depends on the popularity of the page, apparently determined by the number of incoming links. E.g, a search for Donald brings up Donald Trump at the top (after Donald itself), while all other lesser-known Donalds are further down the list. SD0001 (talk) 14:46, 26 April 2020 (UTC)[reply]
Pinging Thryduulf, who I remember had mentioned something about rcats and search suggestions. – Uanfala (talk) 14:54, 26 April 2020 (UTC)[reply]
I believe it used to be the case that redirects tagged as unprintworthy (directly or indirectly) were excluded from the search suggestions (possibly for anything other than an exact match) but this is no longer the case. Redirects that are unprintworthy will generally have fewer incoming links and so naturally de-prioritised, but phab:T24251 is asking for a way to exclude some pages from the search results/search suggestions (it's been around since 2012 and triaged as low priority though so don't hold your breath). It has been a long-standing principle at RfD though that we don't delete redirects solely to avoid them appearing in the search suggestions box if they are otherwise useful, as the suggestions are only shown to a subset of users (mainly those with javascript enabled who use the internal search engine, the official Android app links to the target directly) and it is not knowable what proportion of users this is. Thryduulf (talk) 15:45, 26 April 2020 (UTC)[reply]

Unsortable table in a template

Please view question here and reply there, as this is where the helpee is looking for help. (I've also advised them of this forum's existence in case they need it in the future). Thank you and best regards, --Gryllida (talk) 05:28, 26 April 2020 (UTC)[reply]

Why does Wikipedia keep showing edit conflicts on my own edits

Why does Wikipedia keep showing edit conflicts on my own edits?

I'll make an edit to a section or a page > send Publish Changes, receive an Edit Conflict page > and have to cut & paste from that to a New Edit page > sometimes it works.

No others contributing to the page at the time.

Fustrating!

File:Https://en.wikipedia.org/wiki/Frustration
Zulu Papa 5 * (talk)


LEAVE THESE TWO POSTS HERE. BECAUSE THIS SHOWS WHAT CAN HAPPEN to disrupt an editor's contribution. 18:34, 26 April 2020 (UTC)

Why does Wikipedia keep showing edit conflicts on my own edits?

Why does Wikipedia keep showing edit conflicts on my own edits? I'll make an edit to a section or a page > send Publish Changes, receive an Edit Conflict page > and have to cut & paste from that to a New Edit page > sometimes it works. No others contributing to the page at the time. Fustrating!

File:Https://en.wikipedia.org/wiki/Frustration

Zulu Papa 5 * (talk)