Wikipedia:Village pump (technical): Difference between revisions

Jump to navigation Jump to search
Tags: New topic Source
Line 459: Line 459:
{{user|Gonnym}} isn't around much these days, so I'm asking here. Under certain circumstances, {{tlx|Television network original programming category}} puts categories inside themselves unless it's configured properly; twice in the past Gonnym has fixed up such problems as requested at [[User talk:Gonnym/Archive 3#Template:Television network original programming category]], [[User talk:Gonnym/Archive 4#Television network original programming category]] and [[User talk:Gonnym/Archive 4#Zee Marathi Original Programming]], but didn't explain what actualy needed doing (there is an expln of sorts on the last one, but it's not at all clear). Now {{cl|Fox Kids}} is suffering the same problem: it's inside itself (reported at [[Wikipedia:Database reports/Self-categorized categories]]). Does anybody know what needs doing to the template in order fix the cat? --[[User:Redrose64|<span style="color:#a80000; background:#ffeeee; text-decoration:inherit">Red</span>rose64]] &#x1f339; ([[User talk:Redrose64|talk]]) 20:34, 28 September 2022 (UTC)
{{user|Gonnym}} isn't around much these days, so I'm asking here. Under certain circumstances, {{tlx|Television network original programming category}} puts categories inside themselves unless it's configured properly; twice in the past Gonnym has fixed up such problems as requested at [[User talk:Gonnym/Archive 3#Template:Television network original programming category]], [[User talk:Gonnym/Archive 4#Television network original programming category]] and [[User talk:Gonnym/Archive 4#Zee Marathi Original Programming]], but didn't explain what actualy needed doing (there is an expln of sorts on the last one, but it's not at all clear). Now {{cl|Fox Kids}} is suffering the same problem: it's inside itself (reported at [[Wikipedia:Database reports/Self-categorized categories]]). Does anybody know what needs doing to the template in order fix the cat? --[[User:Redrose64|<span style="color:#a80000; background:#ffeeee; text-decoration:inherit">Red</span>rose64]] &#x1f339; ([[User talk:Redrose64|talk]]) 20:34, 28 September 2022 (UTC)
:I have added <code>|no_cat=1</code>.[https://en.wikipedia.org/w/index.php?title=Category:Fox_Kids&diff=1112928093&oldid=1112616566] [[User:PrimeHunter|PrimeHunter]] ([[User talk:PrimeHunter|talk]]) 20:45, 28 September 2022 (UTC)
:I have added <code>|no_cat=1</code>.[https://en.wikipedia.org/w/index.php?title=Category:Fox_Kids&diff=1112928093&oldid=1112616566] [[User:PrimeHunter|PrimeHunter]] ([[User talk:PrimeHunter|talk]]) 20:45, 28 September 2022 (UTC)

== Global userpages not showing up in autocomplete search results? ==

Has anyone else noticed this? I'm not sure if this is a new bug or if I just haven't spotted it in the past. If an editor has a global user page it doesn't appear in the autocomplete results drop down - their subpages appear fine, but not the main userpage. [[Special:Contributions/192.76.8.81|192.76.8.81]] ([[User talk:192.76.8.81|talk]]) 21:03, 28 September 2022 (UTC)

Revision as of 21:03, 28 September 2022

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

If you want to report a JavaScript error, please follow this guideline. Questions about MediaWiki in general should be posted at the MediaWiki support desk. Discussions are automatically archived after remaining inactive for five days.

Search totals capped at 10,000

For over a year, I have used insource regex searching intensively. One of its most valuable features was that it gave a total number of matches.

My search for untagged, unbracketed bare URLs is insource:/\<ref( [^\>]*)?\>https?:\/\/[^ \<\>\{\}]+ *\<\/ref/i. Until this evening, it was reporting a total of ~44K hits. Now it always reports 10,000.

This is a real loss of utility. It deprives me of my main tool for monitoring progress in between the twice-monthly database downloads.

Why has this happened? BrownHairedGirl (talk) • (contribs) 22:10, 8 September 2022 (UTC)Reply[reply]

We are in the process of upgrading from Elasticsearch 6.8.23 to Elasticsearch 7.10.2. This behavior changed in ES7; now queries are early-stopped at a count of 10,000 to save CPU time.
Previously, while the number of actual documents returned would never be more than 10,000, it would still return the full count (the ~44k hits you were seeing).
It sounds like there's a setting we can flip to switch the count behavior back to how it was previously.
We'll have a phabricator task up for this in a bit so I'll circle back to edit this response with a link to that ticket in a little bit.
Thanks for reporting the behavior change, it's very helpful for us to get a feel for users' workflows. RKemper (WMF) (talk) 22:41, 8 September 2022 (UTC)Reply[reply]
And that ticket is T317374. :) Legoktm (talk) 22:42, 8 September 2022 (UTC)Reply[reply]
Thanks, @Legoktm & @RKemper (WMF).
The full count is really useful for many tasks. I use it many times a day for that search, but also as many times again for various other regex searches. It's great to be able to ask "how big is this issue" and get a prompt answer. BrownHairedGirl (talk) • (contribs) 23:09, 8 September 2022 (UTC)Reply[reply]
I know we're not supposed to say "Me too" on Phab, so here it is in the right place: "Me too!" This feature is invaluable. – Jonesey95 (talk) 23:28, 8 September 2022 (UTC)Reply[reply]
@Jonesey95. Might want to click "Award Token" on Phab and give it a thumbs up. That's an un-spammy way to support a ticket. Hope that helps. –Novem Linguae (talk) 23:59, 8 September 2022 (UTC)Reply[reply]
Yes, this is a feature I was using. I just found out that it was capped today so like @Jonesey95 and @BrownHairedGirl have said it would be helpful for the total count to show, even if it doesn't reutrn more than 10,000 results (like it did in the past). Rlink2 (talk) 13:28, 9 September 2022 (UTC)Reply[reply]
Circling back here: we've merged a patch that should restore the previous count behavior. This patch won't go live until the next backport deploy or deployment train. Our current goal is to get it out during the Monday Sep 12 US backport window, so if you all haven't heard anything by Tuesday US time, feel free to ping me/us again here or on IRC (#wikimedia-search) for an update. RKemper (WMF) (talk) 18:57, 9 September 2022 (UTC)Reply[reply]
Many thanks, @RKemper (WMF), both for your prompt action on this and for your courtesy throughout. I look fwd to the fix. BrownHairedGirl (talk) • (contribs) 19:11, 9 September 2022 (UTC)Reply[reply]
Update: We've rolled out the new setting. I tried @BrownHairedGirl's query and can confirm I see the full 44k+ hits listed. RKemper (WMF) (talk) 07:26, 13 September 2022 (UTC)Reply[reply]
Thanks for the rollout, @RKemper (WMF).
The full count is indeed being listed, @RKemper (WMF). Thanks for you work on this.
However this has confirmed another problem which I had glimpsed in the past week: that the new search is much more susceptible to timeouts.
Before the switch to Elasticsearch 7, my bare URLs search timed out only very occasionally, and only at periods of peak load, i.e. 2100–0600
However, now my bare URLs search times out on nearly every attempt. Even at 1000 UTC, I still had to try three times to get a full set, without timeout.
The behaviour within AWB is even more annoying. Since May 2022, I have at least once every day (and more often 2-3 times) used exactly that same search within AWB, to generate lists to feed to @Citation bot. It always produced 1,000 hits the maximum permitted by AWB; never any less, in hundreds of uses. But since the arrival of Elasticsearch 7, about half of those with-AWB searches return less than 1,000 hits; only morning-time UTC and early afternoon is safe. I presume that is because the search has timed out.
Note that my current bare URLs search was adopted only when the number of unbracketed, untagged bare URLs dipped below about 110,000. Before that, it timed out, so I used a crude, less complete search which excludes named refs: insource:/\<ref\>https?:\/\/[^ \<\>\{\}+\s*\<\/ref/i]
So Elasticsearch 7 is causing my current search to timeout at less than half the hit count which caused the Elasticsearch 6 to timeout.
What exactly are the upsides of Elasticsearch 7? I see only significantly degraded performance, with no benefits. Have I missed something? BrownHairedGirl (talk) • (contribs) 11:01, 13 September 2022 (UTC)Reply[reply]
Using Elasticsearch 6 blocks the use of PHP 8 (see the task chain), which constitutes a security concern as PHP 7 goes out of support soon. Elasticsearch 6 is also not supported by today's and/or tomorrow's OS vendors, which constitutes its own separate security concern. Nearly all version bumps where the feature set is not advertised widely are due to security at the end of the day. Izno (talk) 15:22, 13 September 2022 (UTC)Reply[reply]
Thanks, @Izno. I accept that security concerns will be a driving factor.
But the result of this security fix is a significant fall in performance. It would be helpful to know what processes are underway to identify and implement measures to offset the performance hit, so that users do not lose functionality. BrownHairedGirl (talk) • (contribs) 15:48, 13 September 2022 (UTC)Reply[reply]
@RKemper (WMF): any chance of an comment on whether there is a process to offset the performance hit?
Some of my other work is badly hit by this. For example, I am cleaning up the remaining bare URLs on some top-level domains. I have already done .pk, .si, .ua, .int and .ie. Now I am back to working on .fr, but since the upgrade to Elasticsearch 7 my search insource:/\<ref( [^\>]*)?\>https?:\/\/[^\/ \<\>\{\}]+?\.fr\/[^ \<\>\{\}]+ *\<\/ref/i times out more often than not. Before the upgrade, it never timed out. BrownHairedGirl (talk) • (contribs) 00:41, 16 September 2022 (UTC)Reply[reply]
Part of the upgrade requires switching traffic to a single cluster while we upgrade the other. At the moment, the search traffic is going to our secondary datacenter (codfw). This leads to higher load on the cluster and higher latency due to traffic between datacenters. I'm not confident that this is the only reason, there might well be use cases where Elasticsearch 7 has lower performances than Elasticsearch 6. But it make sense to wait until the migration is fully completed before investigating this further. You can check the progress on this Phab task: https://phabricator.wikimedia.org/T308676.
GLederrey (WMF) (talk) GLederrey (WMF) (talk) 14:58, 16 September 2022 (UTC)Reply[reply]
Thanks for that clarification, @GLederrey (WMF). That reduced capacity during switchover makes sense, and I am sure that it is at least part of the reason for the performance hit.
Thanks for the Phab link, but I don't have the headspace to learn all the complexities of managing a huge server system. I think its much better for me to leave that in the capable hands of those professionals like yourself and @RKemper (WMF) who specialise in such things.
May I assume from your reply that when the upgrade is complete someone (or some team) within WMF tech will remain "seized of the matter" (as diplomats say) and follow through on performance tests, and plan any necessary remedies? BrownHairedGirl (talk) • (contribs) 05:12, 21 September 2022 (UTC)Reply[reply]
Our metrics show that there is no significant performance degradation. That being said, our metrics are global and we don't track the performance of individual queries. If you see specific queries that are now performing worse than expected, please let us now! Ideally by opening a Phabricator ticket and tagging it with "Discovery-Search". GLederrey (WMF) (talk) 08:33, 22 September 2022 (UTC)Reply[reply]
@GLederrey (WMF): I have already given you two specific examples, above. Both extensively tested, both with very serious performance degradation. BrownHairedGirl (talk) • (contribs) 11:19, 24 September 2022 (UTC)Reply[reply]

Elasticsearch 7 features

After RKemper (WMF)'s helpful comments about the upgrade from Elasticsearch 6.8.23 to Elasticsearch 7.10.2, I did a quick google for docs on changes in functionality. However, all I found was a cluster of pages on the Elasticsearch website which extolled the software's virtues for sysops, such as more compact logfiles. Nothing about users.

Is there any documentation of any changes for users of the search?

For example, I would love the search to support Perl-style shorthand character classes such as \s, \S, \d, and \b. (I have a crush on \b). That would eliminate some tedious conversion and debugging when I take a regex from a Perl script or a C# AWB custom module to use in an insource search (or vice versa), as I often do for sanity checks.

Does the upgrade bring any such changes to the construction and/or reliability of regex searches? --BrownHairedGirl (talk) • (contribs) 07:39, 9 September 2022 (UTC)Reply[reply]

The regex support is actually provided by Lucene, which does not provide full regex support. Elasticsearch documents their regex support for v7 on this documentation page. —TheDJ (talkcontribs) 08:21, 9 September 2022 (UTC)Reply[reply]
Thanks, @TheDJ. That link is very helpful ... unlike the half-arsed version of regex which it implements. BrownHairedGirl (talk) • (contribs) 10:33, 9 September 2022 (UTC)Reply[reply]

Template:2022–23 National League 2 East

Unable to edit the league table for 2022–23 National League 2 East. When I attempt an edit, the league table for 2022–23 National League 2 North is in the template. Would appreciate some help as I am unable to work out where I have gone wrong. Jowaninpensans (talk) 20:57, 17 September 2022 (UTC)Reply[reply]

@Jowaninpensans: I have fixed the page name in the watch/edit/discuss links made by the template.[1] I clicked the "Edit" tab to edit the template when the links didn't work. PrimeHunter (talk) 21:12, 17 September 2022 (UTC)Reply[reply]
This is something that should be picked up by Wikipedia:Database reports/Invalid Navbar links, but it's not - I think because it uses manually-constructed links for edit etc., rather than a {{navbar}}. --Redrose64 🌹 (talk) 16:43, 18 September 2022 (UTC)Reply[reply]
Yes, Wikipedia:Database reports/Invalid Navbar links is only for templates which use certain other templates to make the links. Template:2022–23 National League 2 East should be converted to use {{navbar}} or something similar but it follows a convention in Category:England rugby union standings templates (and apparently also some other countries). They should all be converted but at the same time for consistency. It's also problematic that they have a watch link instead of view when most readers are IP's with no watchlist. PrimeHunter (talk) 23:53, 19 September 2022 (UTC)Reply[reply]
OOjs UI icon check-constructive.svg Done. — Qwerfjkltalk 17:42, 23 September 2022 (UTC)Reply[reply]

User talk page appearance

Probably an obvious question here but the appearance of User talk pages has radically changed even though I'm in the old Vector skin. Individual talk page comments are now enormous in size and have irrelevant notes like how many people are participating in a discussion (which is almost always 1 with a template talk page notice) and how many comments to a comment there have been (also, almost always 1). Lots of extra lines and whitespace and a larger than normal comment header.

I didn't change my Preferences and no one else has mentioned it here today so is this something that just rolled out? How can I get the normal look back? When I preview other skins, they just show me different versions of the Main Page, not a User Talk page. Oh, and I'm using a laptop, I'm not on mobile. Thanks for any information you can provide Liz Read! Talk! 21:27, 21 September 2022 (UTC)Reply[reply]

I just noticed the same thing using the desktop site on mobile, and I'm also not too fond of it. ScottishFinnishRadish (talk) 21:29, 21 September 2022 (UTC)Reply[reply]
Just FYI, you can turn this off in your preferences — it's an option titled "Show discussion activity" in the "Editing" tab — TheresNoTime (talk • they/them) 21:41, 21 September 2022 (UTC)Reply[reply]
For me it was turning off Discussion tools in the beta features tab. -- LCU ActivelyDisinterested transmissions °co-ords° 21:48, 21 September 2022 (UTC)Reply[reply]
That'll do it too :) but "Show discussion activity" will turn off the new design without turning off the rest of the features — TheresNoTime (talk • they/them) 21:50, 21 September 2022 (UTC)Reply[reply]
That's option doesn't exist for me, is it part of the new skin? I still have all the reply functions etc with the beta feature off. -- LCU ActivelyDisinterested transmissions °co-ords° 21:53, 21 September 2022 (UTC)Reply[reply]
Huh, you're quite right! My bad Face-smile.svgTheresNoTime (talk • they/them) 21:56, 21 September 2022 (UTC)Reply[reply]
@ActivelyDisinterested, if you have the Beta Feature on, then you will have an option for "Show discussion activity" at the very end of Special:Preferences#mw-prefsection-editing-discussion. If you have the Beta Featured disabled, then that option should disappear. Whatamidoing (WMF) (talk) 22:52, 21 September 2022 (UTC)Reply[reply]
I checked and you quote correct, however both seem to result in the same functionality. Maybe because I'm using the 2010 skin. I'm going to leave the beta off, so as to avoid a repeat situation with something else. -- LCU ActivelyDisinterested transmissions °co-ords° 11:27, 22 September 2022 (UTC)Reply[reply]
I wonder if the [subscribe] button is visible to you on all talk page sections? CX Zoom[he/him] (let's talk • {CX}) 14:09, 22 September 2022 (UTC)Reply[reply]
@ActivelyDisinterested: forgot ping. CX Zoom[he/him] (let's talk • {CX}) 14:10, 22 September 2022 (UTC)Reply[reply]
Yep, the subscribe button is still visable without the beta section being active. -- LCU ActivelyDisinterested transmissions °co-ords° 14:17, 22 September 2022 (UTC)Reply[reply]
To add, I'm also still getting notifications from subscribed talk page sections. -- LCU ActivelyDisinterested transmissions °co-ords° 14:17, 22 September 2022 (UTC)Reply[reply]
@ActivelyDisinterested, that's expected, because all that's "in" the Beta Feature at this moment, at this wiki, is the discussion activity stuff. The prior work has graduated "out" of the Beta Feature, and the future work isn't on the servers yet.
I'm going to make two recommendations, explain why, and then you can make your own choice.
First: It takes a week or two to get used to visual changes. The first day is always the worst/weirdest. I recommend trying to live with it for a while, and then making your decision. (I make the same recommendation all visual changes, e.g., Vector 2022. Try it out for a week or two, not just an hour or two.)
Second: If you decide to turn it off, then in this specific case I recommend that all experienced/long-time editors turn the Beta Feature back on, and then turn the discussion activity item off in the proper/permanent preferences at Special:Preferences#mw-prefsection-editing-discussion. I recommend this because if you turn on the Beta Feature and turn off the real pref, you're done forever with this set of changes. OTOH, if you turn off the Beta Feature, but leave the real pref "on" (and, at the moment, invisible), then whenever it is deployed by default/graduates from the Beta Feature, it'll get flipped back on for you later (due to prefs database limitations). So if you decide this set of changes is not for you, that's perfectly fine, and the most stable way to get rid of it is to disable the regular pref item, rather than the Beta Feature. Whatamidoing (WMF) (talk) 18:02, 22 September 2022 (UTC)Reply[reply]
I'm using the desktop version on mobile, the additional white spaces and other deadspace is far problematic at this point. Maybe I'll take another look once improvements have been made to how things are formatted. -- LCU ActivelyDisinterested transmissions °co-ords° 18:07, 22 September 2022 (UTC)Reply[reply]
Which skin are you using? You should be able to find the name at Special:Preferences#mw-prefsection-rendering-skin if you're using the desktop site. Whatamidoing (WMF) (talk) 18:11, 22 September 2022 (UTC)Reply[reply]
Seven posts above, 2010. -- LCU ActivelyDisinterested transmissions °co-ords° 18:32, 22 September 2022 (UTC)Reply[reply]
Previous feedback has been provided at mw:Talk:Talk pages project/Usability which describes the changes. Izno (talk) 21:48, 21 September 2022 (UTC)Reply[reply]
  • @PPelberg (WMF): WP:ITSTHURSDAY? Was there a phab task to force this on here? — xaosflux Talk 22:04, 21 September 2022 (UTC)Reply[reply]
    @Xaosflux: T315625 is the specific task about enabling it here (probably not the best location to raise concerns though) — TheresNoTime (talk • they/them) 22:07, 21 September 2022 (UTC)Reply[reply]
    Thank you. It does helpfully outline who it was applied to. — xaosflux Talk 22:26, 21 September 2022 (UTC)Reply[reply]
    Was there a phab task to force this on here?
    @Xaosflux: T315625 is the ticket I think you're looking for; thank you for linking it above, @TheresNoTime.
    There's also T317716 which we used to track the work around announcing this change. See: DiscussionTools Beta Feature update.
    Also, please know that it's important to us that you all:
    1. Expect the changes that we are planning to make
    2. Know who they will impact and
    3. Are clear about how you can disable them
    So, if there are improvements to how you think we can communicate the above, I'd value knowing. PPelberg (WMF) (talk) 22:30, 21 September 2022 (UTC)Reply[reply]
    @PPelberg (WMF): although these changes were mentioned in Tech/News/2022/33 (15 August 2022) with a link to the roadmap, visibly obvious changes might be worth a touch of "over communicating" with repeated mentions in Tech News and/or "reminders" here? Tech/News/2022/34 did this quite well for "The new [subscribe] button [...]", and also gave guidance on how to disable the change. Just a thought! Face-smile.svgTheresNoTime (talk • they/them) 23:10, 21 September 2022 (UTC)Reply[reply]
    It's been on this page at Wikipedia:Village pump (technical)#DiscussionTools Beta Feature update for the last 11 days, in addition to other announcements, not to mention the CentralNotice banners a few weeks back. There is no level of announcement that will ever reach every affected person, though phab:T67191 (notification that a Beta Feature you personally enabled was updated). Whatamidoing (WMF) (talk) 18:07, 22 September 2022 (UTC)Reply[reply]
    I think the confusion in T315625 is that the feature would only show for People who have the Show discussion activity setting within Special:Preferences#mw-prefsection-editing enabled - but not that something was going to enable this (what I think previously didn't exist??) option for people, yet. — xaosflux Talk 00:28, 22 September 2022 (UTC)Reply[reply]
@Liz, I'd particularly like to know more about the "enormous in size" problem. It shouldn't change the size of anything (except to the extent that adding more information means that the new information takes up some space). Whatamidoing (WMF) (talk) 22:53, 21 September 2022 (UTC)Reply[reply]
The size of the section title is now three times as high based on looking inspect element on Firefox, which is explained by the extra information, but there is also now double the padding/margin underneath the horizontal line from my testing. Also the font of the section title is now bolded which may make it feel bigger too. Personally I like the new design, but it does add more whitespace which may be the intended effect. Dreamy Jazz talk to me | my contributions 23:02, 21 September 2022 (UTC)Reply[reply]
@Dreamy Jazz, which skin are you using? Whatamidoing (WMF) (talk) 18:08, 22 September 2022 (UTC)Reply[reply]
@Whatamidoing (WMF), I'm using Vector 2022. Dreamy Jazz talk to me | my contributions 18:25, 22 September 2022 (UTC)Reply[reply]
And compared it to using incognito mode (which was therefore comparing to Vector 2010). Dreamy Jazz talk to me | my contributions 18:26, 22 September 2022 (UTC)Reply[reply]
Screenshot comparison
Here's a screenshot comparing the same talk page in three different versions:
  • Vector 2010 with the new information about the discussion activity,
  • Vector 2022 with the new information about the discussion activity, and
  • Vector 2010 again, but without the new information about the discussion activity (i.e., what it looked like last week, and what it looks like if you uncheck the pref now).
The font sizes all look the same to me. Obviously, though, if you're going to add a line saying how many people were in the conversation, etc., then that will take up some room. Is that approximately what you're seeing? Or are you seeing radically different font heights? Whatamidoing (WMF) (talk) 02:40, 23 September 2022 (UTC)Reply[reply]
I hate the big increase in vertical whitespace. It forces me to do extra scrolling, and it offers no benefit in terms of distinguishing the elements. This is a usability fail.
It may not be critical for users of huge desktop screens, but on a laptop vertical space is in short supply. BrownHairedGirl (talk) • (contribs) 23:22, 21 September 2022 (UTC)Reply[reply]
I've removed the details, but would also back a whitespace reversion - it doesn't seem to be notable adding to section header readability Nosebagbear (talk) 23:23, 21 September 2022 (UTC)Reply[reply]
(tldr) Is there a way to op out? - FlightTime (open channel) 01:10, 22 September 2022 (UTC)Reply[reply]
Special:Preferences → "Editing" → "Show discussion activity" (uncheck) — TheresNoTime (talk • they/them) 01:21, 22 September 2022 (UTC)Reply[reply]
@TheresNoTime: Thank you Face-smile.svg — Preceding unsigned comment added by FlightTime (talkcontribs) 01:29, 22 September 2022 (UTC)Reply[reply]
That's an important point worth repeating, which applies to other proposed changes too. Many (perhaps most) "desktop" readers use a laptop display about 8" high, and would rather see more information on each screenful than extra white spacing. Certes (talk) 10:55, 22 September 2022 (UTC)Reply[reply]
@BrownHairedGirl / @Dreamy Jazz we've tightened up the spacing around the headings slightly, and that'll be coming out next week. If you want a preview, it should be visible on the beta cluster now. DLynch (WMF) (talk) 15:39, 22 September 2022 (UTC)Reply[reply]
There seems to be more padding between the bottom of a section and the start of the heading on the beta cluster than there does on enwiki currently? Dreamy Jazz talk to me | my contributions 18:33, 22 September 2022 (UTC)Reply[reply]
That got adjusted too, yeah. It's more-compact inside a given section, but I think it's approximately a wash across the entire page. You can see the design decisions around it on T314449 if you're curious. DLynch (WMF) (talk) 14:47, 24 September 2022 (UTC)Reply[reply]
I don't want to turn this off just yet, but it looks terrible because the headers on talk pages now no longer respect my font choices and use a different font family from the body. Could the headers be made to use the same font as the body? Or is there a CSS hack to fix this? —Kusma (talk) 08:52, 22 September 2022 (UTC)Reply[reply]
You can override the font using the following selector:
.ext-discussiontools-visualenhancements-enabled .ext-discussiontools-init-section
In your specific case, change
body { ...
to
body, .ext-discussiontools-visualenhancements-enabled .ext-discussiontools-init-section { ... ESanders (WMF) (talk) 11:45, 24 September 2022 (UTC)Reply[reply]
Thank you! Just one thing (the most visible one) was missing: the header font is .mw-headline and also needs to be added to the CSS rule. —Kusma (talk) 20:08, 24 September 2022 (UTC)Reply[reply]

I'm curious to know what is good (enough) about this feature for it to be turned on by default; the extra talk clutter and space is annoying (now gone, but how many editors are going to need to figure out how to turn it off now?). SandyGeorgia (Talk) 05:08, 22 September 2022 (UTC)Reply[reply]

@SandyGeorgia I don't think this is on "by default", but it has been turned on for people that have opted-in to the beta feature. — xaosflux Talk 13:43, 22 September 2022 (UTC)Reply[reply]
That's correct: It is only visible to editors who have gone to Special:Preferences#mw-prefsection-betafeatures and clicked the button to enable "Discussion Tools". If you don't have the Beta Feature, you won't see it. If you don't like it, then you can turn it off (very last item on the page) permanently. The prefs have been designed so that it should avoid any "Ooops, it automatically re-enabled during the next update" problems.
Discussion Tools has multiple components:
  • Reply tool
  • New Topic tool
  • Subscribe button
  • Discussion activity and appearance (three parts)
  • Mobile site (multiple parts, basically tweaks on the previous four)
At the English Wikipedia (but not quite all of the wikis), the first three are deployed by default to all editors, including logged-out/IPs.
At the English Wikipedia (but not quite all of the wikis), the first part of the fourth item is available only to people who have enabled the Beta Features. The other parts are much smaller (mostly a line at the top of the page, a change from [reply] to Reply, and comment count in the TOC iff you are using Vector 2022) and will happen later this year. If you want to see what that looks like in various skins, click on the PatchDemo links at Wikipedia:Village pump (technical)#DiscussionTools Beta Feature update (top of this page).
Either one or two parts of the mobile category are available at about 8 wikis (including the French Wikipedia). None of that is available here, no matter what you try to enable. Whatamidoing (WMF) (talk) 17:45, 22 September 2022 (UTC)Reply[reply]
Well, fiddlesticks ... that means someone somewhere advised me to turn on Beta, and I did it, but I have no idea what I've done :( SandyGeorgia (Talk) 21:16, 22 September 2022 (UTC)Reply[reply]
I'd had it on, too - I suspect that I must have thought it provided a useful feature at some point, but have now clicked it off as I think it adds clutter and needless whitespace. Does the fact that it's a beta feature mean that there's consideration of ever moving it beyond beta and making it a permanent feature? Hog Farm Talk 21:22, 22 September 2022 (UTC)Reply[reply]
Yes of course… —TheDJ (talkcontribs) 10:36, 23 September 2022 (UTC)Reply[reply]
Fwiw - I just checked. Do not and never have had beta features enabled. Nonetheless I had to turn off the "new" features that were apparently checked by default. This is for Vector 2010. With Safari there aren't white space and font issues, but there is considerable page clutter. Victoria (tk) 21:41, 22 September 2022 (UTC)Reply[reply]
These changes, not the [reply] button.
@Victoriaearle, the new changes are in the Beta Feature. Click on the picture and see the things that the arrows are pointing at. Whatamidoing (WMF) (talk) 00:24, 23 September 2022 (UTC)Reply[reply]

New topic with wrong section link

Does anyone know why the link in this summary points to the nonexistent section "‎Depricate the signers parameter" despite the subject Depricate the <code>signers</code> parameter?, omitting the question mark? Nardog (talk) 03:49, 22 September 2022 (UTC)Reply[reply]

That was posted using Discussion Tool's ==New Topic== tool, but I wasn't able to replicate it. (I had wondered whether it just dropped the ?) Whatamidoing (WMF) (talk) 17:49, 22 September 2022 (UTC)Reply[reply]
One issue that could have caused this is T315730 (where the summary is generated incorrectly after restoring from autosave sometimes), but it's impossible to know for sure. The user might also have changed the summary themself and forgot the question mark (perhaps if they weren't sure that the <code> tags would be stripped). Matma Rex talk 20:49, 24 September 2022 (UTC)Reply[reply]

How to install CSS like we install JS?

I know how to install JS in my global.js. I wonder how to install a CSS created by someone else at my global.css. Can someone help? CX Zoom[he/him] (let's talk • {CX}) 17:10, 22 September 2022 (UTC)Reply[reply]

Assuming I understand the question correctly, if you don't want to just copy and paste it, you can use @import url("") CSS feature. So like:
@import url("https://meta.wikimedia.org/w/index.php?title=User:SMcCandlish/codefont.css&action=raw&bcache=1&maxage=86400&ctype=text/css");
from the instructions in m:User:SMcCandlish/codefont.css. (And MDN documentation). Skynxnex (talk) 18:37, 22 September 2022 (UTC)Reply[reply]
Thanks. This is what I was looking for. CX Zoom[he/him] (let's talk • {CX}) 07:04, 23 September 2022 (UTC)Reply[reply]
Keep in mind that CSS loaded like this makes loading of every other CSS line slower, so a better idea might be to import CSS as JS like this:
mw.loader.load( 'https://ru.wikipedia.org/w/index.php?title=Участник:Stjn/linter.css&action=raw&ctype=text/css', 'text/css' );
stjn 10:27, 23 September 2022 (UTC)Reply[reply]
Ah cool. I hadn't realized that loader could load CSS as well. Skynxnex (talk) 12:33, 23 September 2022 (UTC)Reply[reply]
Note loading CSS in a Javascript file will add to its processing time, and since I believe the Javascript executes after the page is visible, its appearance may then be altered after the load is complete. This might be an acceptable tradeoff depending on your specific circumstances, so you can try it out and see what works best for you. isaacl (talk) 07:03, 24 September 2022 (UTC)Reply[reply]
The bcache=1 and maxage=86400 params in the URL do nothing. Index.php calls are never cached for logged-in users. – SD0001 (talk) 06:01, 24 September 2022 (UTC)Reply[reply]
Thank you everyone. CX Zoom[he/him] (let's talk • {CX}) 07:59, 24 September 2022 (UTC)Reply[reply]

Page unexpectedly seen on my watchlist

Lately, I've seen a lot of pages on my watchlist which I cannot explain being there. Most (maybe all) of these are user talk pages for anonymous users. AFAICS, I don't have a user preference>watchlist set which might cause this. Wtmitchell (talk) (earlier Boracay Bill) 00:28, 23 September 2022 (UTC)Reply[reply]

I see that you're mainly using Huggle to post messages to IP talk pages. I'm not familiar with Huggle itself, but it probably has some setting somewhere about automatically watching a page if you edit it. Do you have that enabled? --rchard2scout (talk) 08:57, 23 September 2022 (UTC)Reply[reply]
That may be it. I've had huggle related problems in the past and huggle options are more difficult for me to figure out that WP user preferences. I've got options I think might be related to this there turned off, but Huggle may be adding all the pages related to users I've looked at there to my watchlist. I dunno. Thanks. Wtmitchell (talk) (earlier Boracay Bill) 11:32, 23 September 2022 (UTC)Reply[reply]

Bots editing Village Pump Archives

I was trying to sort by last edit date and I came across this Multiple bots seem to be updating archives. Archives are supposed to stay unchanged etc, but I think based on the history, they seem to fix one issue and leave.

Also a lot of it seems to do sigs having font's and colours, maybe the archive bots could strip those?

Or much bigger idea have the colours and fonts as the page is displayed, from the current sig, like the strikethrough for blocked users? Wakelamp d[@-@]b (talk) 12:00, 23 September 2022 (UTC)Reply[reply]

@Wakelamp this is expected behavior. History stays the same forever, "archives" are just copy-paste to make it easy for readers to reference things - these updates are for the benefit of the readers. In many cases, they are restoring the content to look how it originally was intended - as the parser changed breaking certain old markups. — xaosflux Talk 12:39, 23 September 2022 (UTC)Reply[reply]
I guess few users sort archives by last edit date so I wouldn't interfere with relevant bot edits to try to preserve an old date. You aren't supposed to make new posts to archives but fixing lint errors and formatting issues is perfectly fine whether bots or users do it. By the way, I have a script User:PrimeHunter/Search sort.js which can sort by all ten options at mw:Help:CirrusSearch#Explicit sort orders, not just the three at Special:Search. PrimeHunter (talk) 14:49, 23 September 2022 (UTC)Reply[reply]
Thank-you all. The special search looks great. Two last questions
"they seem to fix one issue and leave." Why do bots do this?
and, is there a tool for extracting topic name links from. I am categorising connected proposals, and seeing which were successful. I am rubbish at creating them, and I think some proposals, that are not marked as perennial, may be worth doing, Wiktionary has a log i think that works that way, and also shows the number of editors involved. Wakelamp d[@-@]b (talk) 06:19, 26 September 2022 (UTC)Reply[reply]
@Wakelamp, Why do bots do this? This has been discussed a lot at WP:BOTN. You might want to try the archives there. — Qwerfjkltalk 19:40, 27 September 2022 (UTC)Reply[reply]
It's better to sort by creation date instead. That way the recently created archives always show up towards the top. – SD0001 (talk) 05:58, 24 September 2022 (UTC)Reply[reply]

Help me with this stupid formatting problem

I've reformatted the "Support" section of Wikipedia:Requests for comment/Deployment of Vector (2022) to use automatic numbering (Special:Diff/1111910639)) but the complex formatting of the "Oppose" section has exceeded my ability. Could somebody who really understands how *, : and # really work in wiki-formatting please bail me out? Thanks. -- RoySmith (talk) 16:45, 23 September 2022 (UTC)Reply[reply]

 Done * Pppery * it has begun... 16:59, 23 September 2022 (UTC)Reply[reply]
@RoySmith: It's really quite easy. Consider a talk page discussion with various indent levels, and you want to reply to one of the posts: so you open up a fresh line, and copy all of the symbols (asterisk, colon or hash) from the line above, and add one more to their right. If you apply this principle to the RfC in question, it'll all fall into place. WP:COLAS has more, although it doesn't mention hashes, the principle is identical. --Redrose64 🌹 (talk) 23:40, 23 September 2022 (UTC)Reply[reply]
RoySmith, see also a demonstration at Help:Talk pages/Example. —⁠andrybak (talk) 14:17, 24 September 2022 (UTC)Reply[reply]

SelectRange doesn't scroll textarea

See test script. (run it) This utilizes the selectRange() method from OO.ui.MultilineTextInputWidget which would "Focus the input and select a specified range within the text".
In Blink-based browsers (like Chrome) the textarea doesn't scroll at all. Ideally the selection would be centered. Is that even possible?Alexis Jazz (talk or ping me) 19:26, 23 September 2022 (UTC)Reply[reply]

I believe this is textarea ridiculousness. From memory, you need to set the cursor insertion point, then focus (which will then also scroll), and then select a range. It will scroll to the cursor insertion point, but not to a selection. —TheDJ (talkcontribs) 09:05, 24 September 2022 (UTC)Reply[reply]
Thanks, this seems to have worked. Face-smile.svg For anyone else who might be interested, I used:
FTT.UITextInput.$input[0].blur();
FTT.UITextInput.$input[0].selectionEnd = RefStart; //move cursor to selection start
FTT.UITextInput.focus(); //scroll
FTT.UITextInput.$input[0].blur(); //blur so we can focus again
FTT.UITextInput.$input[0].selectionEnd = RefEnd; //move cursor to selection end
FTT.UITextInput.focus(); //scroll
FTT.UITextInput.selectRange(RefStart,RefEnd); //select
Edit: changed code to use $input which is easier to read, thx Nardog!Alexis Jazz (talk or ping me) 13:47, 24 September 2022 (UTC)Reply[reply]
A jQuery object for the textarea element already exists as property $input. Nardog (talk) 13:52, 24 September 2022 (UTC)Reply[reply]

Issue

If I try to create a section on a talk page on mobile with a link (this was the link I was trying to use), it won’t save. Any help? 47.21.202.18 (talk) 20:02, 23 September 2022 (UTC)Reply[reply]

47.21.202.18, Special:Captcha?Alexis Jazz (talk or ping me) 21:26, 23 September 2022 (UTC)Reply[reply]

Why didn't this ping work?

In Special:Diff/1111826142, I should have gotten pinged, but I never received any notification. Why not? -- RoySmith (talk) 23:16, 23 September 2022 (UTC)Reply[reply]

No signature.
Trappist the monk (talk) 23:19, 23 September 2022 (UTC)Reply[reply]
See more at WP:MENTION. No signature is the most common cause. PrimeHunter (talk) 23:27, 23 September 2022 (UTC)Reply[reply]

Transcoding system unable to convert newer audiovisual uploads?

I uploaded newer audio clips, but somehow there are error statuses, regardless of any original format. I tried purging and resetting without avail. George Ho (talk) 01:22, 24 September 2022 (UTC)Reply[reply]

It's been broken for over 2 weeks. Please see phab:T317069. —TheDJ (talkcontribs) 08:47, 24 September 2022 (UTC)Reply[reply]

CIDR range tool doesn't appear to work with deleted contribs

I can't find a previous mention of this. So, I was trying to find deleted contribs for an IP range. If I look at a single IP I can see their deleted contribs, no problem. However, if I extend that to a range, no deleted contributions appear at all, even if some exist. I have tried this on different ranges of random IPs and it seems to be standard.

Compare (Admin only, obviously):

Black Kite (talk) 06:12, 24 September 2022 (UTC)Reply[reply]

phab:T183457. Izno (talk) 08:29, 24 September 2022 (UTC)Reply[reply]
Mmm, that seems to have disappeared into a black hole, doesn't it? I'm surprised it wasn't actioned as part of phab:T163562. Black Kite (talk) 08:44, 24 September 2022 (UTC)Reply[reply]
Yeah, MusikAnimal linked that in their comment as a "welp, we forgot a thing". Izno (talk) 09:13, 24 September 2022 (UTC)Reply[reply]
In the ideal world, after the solution of something in the phab:T20493 direction, we wouldn't need two separate contribs pages. I think the reason it is like it is today is because one is looking at the revision table and one is looking at the revision archive table (it was an ancient decision that is effectively technical debt these days to have two tables). Izno (talk) 09:20, 24 September 2022 (UTC)Reply[reply]
I'm not massively technical, but does that mean that active revisions are in one table, and deleted ones in another (as opposed to them all being in one table with a field denoting whether they are deleted or not)? If so, seems ... somewhat inefficient. Black Kite (talk) 16:13, 25 September 2022 (UTC)Reply[reply]
Exactly so. I don't know about the implications on efficiency. Some possible reasons on why it was originally done that way, all of which are speculation (I assume I'd have to go hunt down the answer on mediawiki-l): 1) it was done because deletion before user admins could delete things was 'delete the revisions from the database entirely' (I don't understand why this would be a motivation), or 2) they didn't want to take the efficiency hit to find all revisions when we know the vast majority of people looking at a contribs page can see only the undeleted revisions (this doesn't seem unreasonable), or 3) they wanted to ensure deleted revisions were never visible to someone without the appropriate permissions, including in database dumps, or 4) something else. (3) doesn't seem like an issue these days given how revision deletion is implemented.
Hence, one of the later comments in the Phab discussion proposes an archived_page table (this is apparently a question of backwards compatibility) and then moving all revisions back into the revision table with a marker of deletion, similar to how revision deletion is done today. (I assume there would be some redesign of Special:MergeHistory such that moving revisions between pages would become trivial, as history moves are a rather significant use case for deleted revisions.) Izno (talk) 17:06, 25 September 2022 (UTC)Reply[reply]
  • Xaosflux It would be a really useful thing to have in the situation that I found myself (in this case, an IP range that was stealing sandboxes from regular users and creating them as their own articles, which then got deleted as they were unfinished). Black Kite (talk) 16:11, 25 September 2022 (UTC)Reply[reply]

Bullets markup on mobile site are unintuitive

What it looks like
What the code actually is

On Wikipedia's mobile site, two consecutive lines with bullets look more spaced apart than two bullets with a line gap in between them, which is very unintuitive. As in the picture, bullets 2 & 3 are more spaced apart than bullets 1 & 2, despite the fact that bullets 2 & 3 are in consecutive lines, while there is an extra line between bullets 1 & 2. See example at Talk:Nijeder Mawte Nijeder Gaan. This is a very old issue, which I finally decided to report because it is breaking my brain. How can it be fixed? CX Zoom[he/him] (let's talk • {CX}) 20:40, 24 September 2022 (UTC)Reply[reply]

I found the code defining it this way here: https://gerrit.wikimedia.org/g/mediawiki/skins/MinervaNeue/+/6401472afb31690041d1a547fb8f7bc0fcd624eb/resources/skins.minerva.base.styles/content/lists.less#29
Some more digging revealed that it was originally added in 2015 in commit https://gerrit.wikimedia.org/r/c/mediawiki/extensions/MobileFrontend/+/199765 to fix https://phabricator.wikimedia.org/T93784, which complained about poor readability of long lists, like references – it was difficult to distinguish where each item began and started, because the list items commonly take multiple lines, and at that time the list item "marker" wasn't visually indented like it is now (see screenshots in that task).
The markers were changed to be indented since then – it changed in 2019 in https://phabricator.wikimedia.org/T150377 (see screenshots there for comparison), following a discussion that pointed out the exact same readability problems, and I actually worked on it myself (I honestly completely forgot). Looks like we didn't realize at the time that the weird margin was an attempt to fix the same thing.
So to summarize, there was a reason for it once, but it no longer makes much sense. I think people would be open to changing this if you filed a task describing the problem and/or submitted a patch. Matma Rex talk 21:10, 24 September 2022 (UTC)Reply[reply]
@Matma Rex: Thank you for the detailed description. I've opened a task at phab:T318485. You may want to have a look at it. Also, I would've submitted a patch if I knew how to, but I don't know a thing about coding. CX Zoom[he/him] (let's talk • {CX}) 21:38, 24 September 2022 (UTC)Reply[reply]
CX Zoom, until a patch arrives you can add .content ul{padding-top:0.3em}.content li,.content li:last-child{margin-bottom:0.1em} to Special:MyPage/minerva.css.Alexis Jazz (talk or ping me) 21:48, 24 September 2022 (UTC)Reply[reply]

What percent of edits are reverted?

Do we have any statistics on what fraction of mainspace edits are reverted? To be somewhat more concrete, I'm looking at a user with a bit under 500 edits, of which about 10% have been reverted. I'm trying to figure out of that's a lot or not. I can see from Special:Tags that mw-reverted has been applied about 10 million times, but don't know the total number of edits that have been made since we started tagging, nor how it breaks down by namespace. -- RoySmith (talk) 16:01, 25 September 2022 (UTC)Reply[reply]

Beware that some very old edits have recently been reverted and tagged. Certes (talk) 16:18, 25 September 2022 (UTC)Reply[reply]
I got some quick numbers. Considering edits this year, by registered users, only in mainspace, 66% users had less than 5% revert rate, 73% had less than 15%, 78% had less than 25%... MarioGom (talk) 17:47, 25 September 2022 (UTC)Reply[reply]

Deleting citations to unreliable or copyvio sources

Some web sites and books, which fail WP:RS and related policies, are regularly cited in certain classes of article. For example some unreliable sources on aircraft are currently cited anything from a hundred to a thousand times or more. This is nigh-on hopeless to manually fix, but I guess not disruptive enough to warrant such banhammers as WP:DEPRECATE, WP:BLACKLIST or WP:EDITFILTER. Is there a suitable tool to (help) automate deletion of ref containers citing a given web domain? — Cheers, Steelpillow (Talk) 20:32, 25 September 2022 (UTC)Reply[reply]

WP:AWB could do that, but really you should be replacing the reference by what it was a copyvio of, or a more reliable one, rather than just deleting the reference - pretty much a manual process. Graeme Bartlett (talk) 22:14, 25 September 2022 (UTC)Reply[reply]
Thanks, but I use Linux and have never got Wine to play nicely, so AWB is out. The trouble with leaving up say a thousand cites to a source is that other editors then assume it is reliable and post more of them, faster than you can revisit them. Few of these cites actually lead to an identifiable RS in the source page or book, that is one of the main reasons these sources are unreliable in the first place. — Cheers, Steelpillow (Talk) 23:46, 25 September 2022 (UTC)Reply[reply]
Well you may still be able to use the javascript version: User:Joeytje50/JWB. Perhaps you would use the regex editor: //tools-static.wmflabs.org/meta/scripts/pathoschild.templatescript.js : You could use this to comment out the reference, or put on unreliable source template, or other copy-paste operations. Graeme Bartlett (talk) 08:47, 26 September 2022 (UTC)Reply[reply]
Thanks again for the update. It's certainly a useful UI, but does not appear to save me any of the core identify-the-whole-string-to-replace coding. I am not particularly fluent in such things, so I'll have think about it a bit. — Cheers, Steelpillow (Talk) 10:07, 26 September 2022 (UTC)Reply[reply]
@Steelpillow, you might be able to get someone to tag all the sources, since tagging can be done more mechanically than replacing bad refs with good ones. I'd suggest {{copyvio link}} or {{unreliable source?}}. That should discourage people from copying them, plus make it easier for you to find them. WhatamIdoing (talk) 16:58, 26 September 2022 (UTC)Reply[reply]

Extracting WikiProjects on a talk page using template/module

I was wondering if there is a way to extract out the names of the WikiProjects used on a given talk page, for example Talk: United States, such that when the particular template/module is used, it will create an array United States; Countries; North America; United States Government. It would be very helpful if there is a way to do just that. Thanks! CX Zoom[he/him] (let's talk • {CX}) 22:17, 25 September 2022 (UTC)Reply[reply]

Why? Izno (talk) 22:44, 25 September 2022 (UTC)Reply[reply]
Given the huge backlog of COI edit requests, I figured out that topic-wise categorisation (inbuilt from {{edit request}}) might be more helpful than a list all category. It helps in finding the requests that you might be more interested in reviewing because you primarily work in that area, potentially reducing the backlog. If template/module cannot do it, alternatively, I may have to file a bot request. CX Zoom[he/him] (let's talk • {CX}) 07:11, 26 September 2022 (UTC)Reply[reply]
Would a PetScan to intersect Category:Wikipedia conflict of interest edit requests and Category:WikiProject United States articles work for your use case? —⁠andrybak (talk) 11:05, 26 September 2022 (UTC)Reply[reply]
@Andrybak: Looks like it could work but I'm having troubles with it. Anything I search shows 0 results even though it should not. Maybe I'm putting in the wrong input. Can you please give an example input? Thanks! CX Zoom[he/him] (let's talk • {CX}) 12:27, 26 September 2022 (UTC)Reply[reply]
Here you go. Most likely you weren't setting the namespace in the Page Properties tab; by default, it'll only show pages in the main namespace, and you want Talk: instead for both of those categories. —Cryptic 13:10, 26 September 2022 (UTC)Reply[reply]
Thank you very much. CX Zoom[he/him] (let's talk • {CX}) 13:21, 26 September 2022 (UTC)Reply[reply]
@Hellknowz, is this something you could add to Wikipedia:Article alerts? WhatamIdoing (talk) 17:02, 26 September 2022 (UTC)Reply[reply]
@WhatamIdoing: In theory. The problem is that the bot can only do 1 workflow per page. The first page I clicked for an example had like 5 of them. I see many have multiple. The bot won't recognize that a page has 2+ different requests - it will list only one of them, pick one of the dates/authors and won't close or update it until all requests are done. Basically, as long as the page is in the COI request category. So it will basically show incorrect values for those cases (although it will still report them). There's no easy way for me to work around this (other than to hide the values, but then that will make the entries undated). Also, a bigger question is the scope of bot reports. Are COI requests more important than the thousands of regular ones? In theory, the bot can just have an edit request section for each project, including COI reports, which would be ideal. But that still leaves the problem of multiple requests. —  HELLKNOWZ  TALK 17:20, 26 September 2022 (UTC)Reply[reply]
For myself, I'd rather see all the pages with an open edit request, not just the COI ones. One per page is probably enough, because (a) if you're on the page anyway, you can see what other requests are there and (b) even if you don't, if you close that one, then tomorrow the bot will presumably notice and report the next one on the page.
If it could be coded in five minutes or less, I'd also rather not see extremely recent edit requests (say, less than 12–24 hours old), because that would give the regular patrollers a chance to remove empty/accidental edit requests. This should improve the signal-to-noise ratio for Article Alerts. WhatamIdoing (talk) 00:39, 28 September 2022 (UTC)Reply[reply]
"if you close that one, then tomorrow the bot will presumably notice and report the next one on the page." That's the problem - the bot won't notice that anything has changed because the page will still be in the same category. It will continue reporting the first request until every request is closed. It will continue reporting the date, the author and the section link of the old request. Even after all are closed, it will still only list the page as if it had one request during that time and was closed when the last request was closed. If a new request is added soon, the bot will think the old request was reopened.
"If it could be coded in five minutes or less, I'd also rather not see extremely recent edit requests" Unfortunately, this would take way longer to implement than that. Without going into technicals, everything assumes that "active means active" and I cannot filter entries like that without breaking things. And yeah, knowing how many empty and useless edit request I've seen that need no action expect be closed, I can see how early reports like this would be a problem. —  HELLKNOWZ  TALK 10:06, 28 September 2022 (UTC)Reply[reply]

Moving editnotices alongside the subject page

Hi everyone, yesterday I did a page move of 1948 Palestine war per a RM on its talk page. I realised that the page has an active editnotice, which also required moving. I then realised that, I and many other page movers moved pages without moving the corresponding editnotice earlier also. I wonder what can be done about it to ensure that editnotices are correctly moved. Maybe some js/CSS code that creates a check box for editnotice shown to all page movers/template editors/admins. Or maybe just improve the post-move cleanup interface message. Thoughts? CX Zoom[he/him] (let's talk • {CX}) 07:19, 26 September 2022 (UTC)Reply[reply]

I don't know how technically feasible this is but it would be very desirable. Edit notices are bound to a page as strongly as its Talk: page, which usually moves with the article automatically. Perhaps it should be a Phab request to put another tickbox on the move screen, though I don't see why anyone would untick it. (Added later: by "bound", I meant logically associated rather than technically attached.) Certes (talk) 07:42, 26 September 2022 (UTC)Reply[reply]
Such a box would be unchecked for history merges/splits done using page moves, etc. If such a thing were to be implemented, it'd probably be best to couple the talk page move with that of the edit notice, because it'd be very rare indeed to move the article and talk page but not the edit notice. Graham87 08:01, 26 September 2022 (UTC)Reply[reply]
@Certes edit notices are not "bound to a page" as strongly as its talk page (from a technical level). In the example above, 1948_Palestine_war is indeed strongly bound to Talk:1948_Palestine_war; however the software defined edit notice for that page is actually Mediawiki:editnotice-0-1948_Palestine_war. Now, we've used a bit of template magic to make the edit notices a little easier to manage here on the English Wikipedia, and we store the content of that edit notice in Template:Editnotices/Page/1948 Palestine war. So phab request isn't going to be the fix here.
@CX Zoom the "Move succeeded" results page from a move already has a check for this, and even pre-populates the move form for you with a click. Example output line is: This page has an editnotice at Template:Editnotices/Page/OLDTITLE. Please move it to Template:Editnotices/Page/NEWTITLE. (with the "move" being a link to do that). This is included with all of the other directions about cleaning up after a move. — xaosflux Talk 09:47, 26 September 2022 (UTC)Reply[reply]
Tools like rmCloser/MassMove/PageSwap, etc. don't show you the cleanup instructions at all. In fact I don't remember seeing it in the first place because I had to use PageSwap. Also, despite the instructions at the normal page move interface, I will still appreciate if moving editnotices could be done automatically like talk pages using a sitewide check box, because from a page mover's perspective they both belong to the same level. This is because the vast majority of pages needing moves don't have editnotices at all. People read instructions once or twice then start skipping them over. When you actually move a page with editnotice, you don't even realise that it has an editnotice because you skipped the instructions. A visual element like checkbox will avoid this. CX Zoom[he/him] (let's talk • {CX}) 12:14, 26 September 2022 (UTC)Reply[reply]
@CX Zoom if you are using scripts to do your moves, you are not going to see the interface pages as you mentioned - so expanding such scripts is one place where improvements could be had (follow up at script-related pages such as User talk:Andy M. Wang and User talk:Ahecht). As I noted above, we don't put editnotices in the expected place (from a software perspective) here on enwiki - so a UI box in the software isn't going to work. One thing that could possibly help, for people not using scripts, would be to have Special:MovePage display any relevant edit notices - that would be a good tip that something may be needed. — xaosflux Talk 13:28, 26 September 2022 (UTC)Reply[reply]
That makes sense, because page moves are also subject to the same sanctions that normal editing is subject to. And most, if not all, editnotices are just about the various sanctions. CX Zoom[he/him] (let's talk • {CX}) 14:22, 26 September 2022 (UTC)Reply[reply]
@CX Zoom: I opened a feature request, phab:T318572, to display EN's to the move interface. This won't solve directly for cases where you are moving without using the move interface (such as via scripts). — xaosflux Talk 14:49, 26 September 2022 (UTC)Reply[reply]
Wikipedia:Bots/Requests for approval/ProcBot 4 is approved to do exactly this. – SD0001 (talk) 13:22, 26 September 2022 (UTC)Reply[reply]
Thank you very much for this information. If a bot is already doing this thing (which I wasn't aware of), then we could just leave it like we leave double redirects nowadays. CX Zoom[he/him] (let's talk • {CX}) 14:24, 26 September 2022 (UTC)Reply[reply]

Multiple IPs for same user on same browser at same time

I tried to do some logged out testing, only to realize that I'm caught in the range block for Special:Contribs/2409:4061::/34. Same thing on test & test2 Wikipedia. So, I resorted to https://meta.wikimedia.beta.wmflabs.org where my IP address is a completely different 47.11.xx.xxx I'm using the same device, same browser. I tried moving around quite a bit and realized that on all Wikimedia wikis (Special:SiteMatrix), I have the 2409 IP. Whereas on all wmflabs wikis (SiteMatrix), I have the 47 IP. How can this be possible simultaneously? CX Zoom[he/him] (let's talk • {CX}) 20:20, 26 September 2022 (UTC)Reply[reply]

Perhaps that betalabs doesn't support IPV6, hard to tell from here since they don't normally allow anon edits. Many people ave both an IPv4 and IPv6 address, but if the remote end doesn't support IPv6 along the entire route, proxies may get invoked by your ISP as well. — xaosflux Talk 20:29, 26 September 2022 (UTC)Reply[reply]
Indeed, Wikimedia Cloud Services (including Beta Cluster, Toolforge, etc.) doesn't support IPv6 yet, so users will be forced over to IPv4. Legoktm (talk) 05:17, 27 September 2022 (UTC)Reply[reply]

Tech News: 2022-39

MediaWiki message delivery 00:27, 27 September 2022 (UTC)Reply[reply]

Replag currently at 5 hours; reports may be delayed or appear to be out of date

Resolved

Just an FYI that replication lag is currently at 5 hours. This sort of lag leads to database reports that do not run or that, when they update, contain information that is out of date. When the numbers in the linked page go back to zero, reports that are run afterwards should update correctly. – Jonesey95 (talk) 04:50, 27 September 2022 (UTC)Reply[reply]

For the record this was marked as resolved at 14:30 (UTC) on the day this was posted. Graham87 03:49, 28 September 2022 (UTC)Reply[reply]

Limitations (I'm being polite) of mobile Visual Editor?

Gosh—has anyone else been using the mobile browser-based implementation of WP's Visual Editor? Compared to the desktop version, which is generally flawless, the mobile version is so full of bugs, I continually worry about seriously damaging articles due to errors it generates that I may overlook before saving them.
You may tap the screen, and the cursor appears on a different line entirely. You select a word and backspace to delete it, and a different word is deleted instead—or parts of two different words. Sometimes, when you resume editing an article and try to scroll down to the place where you left off, the editor keeps snapping back up to the top of the content, like you've let go of a roller shade too quickly. (At least it spares us the "flap flap" noise.)
I'm always entering source-code mode to repair VE-caused errors, going back and forth to be sure what I'm seeing in the VE reflects reality. Many times I've had to open an article's existing version in a new tab to copy and restore source code where the VE inexplicably deleted it.
Why not stick with the source editor, you may ask, or the desktop version of the VE? Despite my cavalier comments, I'm a busy guy, and my WP-editing opportunities are fleeting enough that I usually have time for them only on my close-at-hand phone (running Android 10 in Firefox—guess I should mention that). And while the ability to edit source code is undeniably valuable, I find it tremendously easier to focus on content without the ubiquitous, overwhelmingly competitive plethora o' extra-textual tags.
Anyway, sorry if I seem to be kvetching. I've actually coped with the mobile VE's quirkiness for years now without saying anything. (For one thing, I didn't know where to say it.)
And one more somewhat-related thing: Why does the mobile VE allow us only to post on Talk pages, not edit or preview them? As you can imagine, this can be a very frustrating one-two punch of embarrassing inaccuracy. Dunno about you—but when addressing my fellow WP obsessives editors, I feel compelled to be at my most anal vigilant, writing-wise.
(Ironically, I see that, unlike regular article pages, this meta-page does have a Preview button. 🤷‍♂️ )
Thanks—I'll hang up now and take my answer off the air. – AndyFielding (talk) 09:33, 27 September 2022 (UTC)Reply[reply]

@AndyFielding It has limitations for sure. Some are imposed on us by the mobile environments (phone makers really want us to build apps that they have full control over rather than websites…), some are our own fault (I'm one of the developers).
Regarding the text input problems, may I hazard a guess that you're using the Gboard keyboard? There was a long-standing compatibility issue, where due to the way that the keyboard sends text, the editor would sometimes "lose track" of what is being typed, ultimately causing the text being shown on the screen to not match what the editor really contained. This can then lead to further issues like text being deleted (or typed) in the wrong place. Good news is that it seems we've finally fixed this (with a contribution from a developer at Fandom (Wikia)) – you can find the details at T312558, and test the new version now at https://en.wikipedia.beta.wmflabs.org/. There may be similar problems with other keyboards too, but this is a very popular one.
Regarding mobile talk pages, the interface there is unrelated to visual editor, only enabled on some pages (that's why you see a different behavior on this page than normal talk pages), and we're currently in the process of replacing it with a new one, based on the desktop reply tool. You can read the status updates on this work at mw:Talk pages project/Mobile. It will have a preview :)
I don't have an immediate response to the other issues, but if you can reproduce them consistently, it would be very helpful for debugging if you could record your screen while demonstrating the problem and shared the video somewhere. Matma Rex talk 21:48, 27 September 2022 (UTC)Reply[reply]
MR: Thanks for your thoughtful and informative reply. Yes, I've been using Gboard, but have alternatives at my disposal. I'll try them, as well as your beta. (Hmm, no Preview button here again, I see… Maybe it's in effect only on certain days of the week. 😄) – AndyFielding (talk) 00:51, 28 September 2022 (UTC)Reply[reply]

Re: Pushpin map of the world's FAs (discussion from March)

Hello, a little update to everyone involved in the discussion "Pushpin map of the world's FAs" from March, @Sdkb, Jéské Couriano, Pppery, The wub, Donald Albury, NguoiDungKhongDinhDanh, TheDJ, Agathoclea, Whatamidoing (WMF)):

As of yesterday, it is possible to use geopoints from Wikidata in Kartographer maps, via QID or SPARQL query. -- Best, Johanna Strodt (WMDE) (talk) 10:09, 27 September 2022 (UTC)Reply[reply]

Template substitution brings a visually different result from expected

I was creating Template:Undated/sandbox to do some experiments. I test them at User:CX Zoom/TestPage34. For all examples, the normal transclusion works fine just as anyone would expect it to. But for whatever reason, the substitution is unexpectedly different for example 2. Why did this happen? (I can think of a workaround using nested #if & #ifeq so that the template could work but maybe I hit a bug or something?) CX Zoom[he/him] (let's talk • {CX}) 10:49, 27 September 2022 (UTC)Reply[reply]

@CX Zoom: Some templates don't work with subst. {{str rightc}} tries to but fails. {{str rightc|abcdefg|5}} produces cdefg which currently says cdefg as it should. {{subst:str rightc|abcdefg|5}} currently produces blank so it matches the blank switch case | = | in Template:Undated/sandbox. PrimeHunter (talk) 03:47, 28 September 2022 (UTC)Reply[reply]
I added a safesubst to {{str rightc}} [6] and everything appears to work now. PrimeHunter (talk) 03:56, 28 September 2022 (UTC)Reply[reply]
Thank you very much. Your edit fixed the issue. CX Zoom[he/him] (let's talk • {CX}) 05:45, 28 September 2022 (UTC)Reply[reply]

Dumber search?

When I search for the word "alarmism" [7], I get hits for "alarm", which is not what I want. I know what I want and I try to tell the search algorithm what I want, but it thinks it is smarter than me and knows better. If I try to exclude "alarm", I get no hits, probably because that also excludes "alarmism".

Is there a way to stop the search algorithm from dunning-krugering like that? --Hob Gadling (talk) 15:57, 27 September 2022 (UTC)Reply[reply]

You can try something like Special:Search/Alarmism insource:/alarmism/. Add i to the end if you also want to include Alarmism, etc. Certes (talk) 16:08, 27 September 2022 (UTC)Reply[reply]
@Hob Gadling: It's better to use quotation marks like you did in the post but not the search. insource:/.../ is called a regex search. It's expensive and may not have time to search all pages. It helps if you combine it with a normal search like alarmism insource:/alarmism/ but for this purpose, you can just use "alarmism". See more at Help:Searching. PrimeHunter (talk) 16:12, 27 September 2022 (UTC)Reply[reply]
Ah! Thank you both! I did not think of Help:Searching or of quotation marks. --Hob Gadling (talk) 16:22, 27 September 2022 (UTC)Reply[reply]
This may be a goofy question, but: Why does WP return results that don't include the searched-for term (especially if the term is in quotes—which, on most search engines, indicates it's mandatory)? If that happened to me, I'd assume no pages existed with the searched-for term, but the search feature was trying to be helpful. – AndyFielding (talk) 01:03, 28 September 2022 (UTC)Reply[reply]
@AndyFielding: Help:Searching#Stem matching says: "Search results will include the roots of words included in the search string, and their various tenses (plural, past-tense, etc.). If stem matching is not wanted, use double quotes around the word or phrase you want to match verbatim." Do you have an example where a search term in quotes returns results without the term? PrimeHunter (talk) 01:15, 28 September 2022 (UTC)Reply[reply]

It's interesting to me that insource has been shown in examples, but what it does has not actually been mentioned, and it's pretty critical to know whether or not it's appropriate. IMO, the first thing you should consider is whether you want to search the page source or the rendered text. The default is the rendered text. Use insource: only if you want to search the page source. Also, it seems that Wikipedia search treats the content as a series of words, and delimiters (the characters that separate words) are not considered, e.g. unless you're using a regex search, search terms "foo/bar" and "foo+bar" will return identical results. Keep in mind also that Wikipedia search isn't google search, it's not going to guess at what you might really have wanted. Fabrickator (talk) 01:50, 28 September 2022 (UTC)Reply[reply]

AndyFieldingCertes only suggested alarmism insource:/alarmism/ as a way to get an exact match (quotes is a better way here). A regex search insource:/.../ cannot be used on the rendered result. By the way, insource:/alarmism/ doesn't find Alarmism while "alarmism" does find it. Add i like insource:/alarmism/i for a case insensitive regex search. It goes from 97 to 134 results. "alarmism" gives 129 results. The five "missing" pages can be seen with alarmism insource:/alarmism/i -"alarmism". Three only has "alarmism" in the source and two say "Alarmism's" which is treated as one word and not found by "alarmism". PrimeHunter (talk) 02:25, 28 September 2022 (UTC)Reply[reply]
Quoted text probably gives what you need here and is much more efficient than my original suggestion of insource:. There is more help at mw:Help:CirrusSearch, and an explanation of the very limited and non-standard regular expression syntax here. Certes (talk) 08:36, 28 September 2022 (UTC)Reply[reply]

Template:Cleanup

How to use "reason 2"? Eurohunter (talk) 18:36, 27 September 2022 (UTC)Reply[reply]

@Eurohunter There is no "reason 2", the 2nd unnamed parameter will be interpreted as "reason", or you can explicitly use the "reason" named parameter. — xaosflux Talk 18:44, 27 September 2022 (UTC)Reply[reply]
@Xaosflux: Second unnamed parameter doesn't works in this way (no effect). Eurohunter (talk) 18:46, 27 September 2022 (UTC)Reply[reply]
Hmm ok maybe not, looks like maybe that used to work but doesn't now - perhaps should be removed from the doc. — xaosflux Talk 18:55, 27 September 2022 (UTC)Reply[reply]
There is no mention of |2= in the template's documentation. I have removed all references to parameter |2= from the template's code, since it apparently has not worked for many years and is not present in any transclusions, as far as I can tell. I do not edit the Visual Editor's Template Data code, which is stored on the documentation page; someone else is welcome to remove references to |2= from that unprotected code. – Jonesey95 (talk) 20:16, 27 September 2022 (UTC)Reply[reply]
2 (second unnamed parameter) has been disallowed as alternative to reason, maybe unintentionally, since [8] by MSGJ in 2012. I guess any old uses of 2 were replaced years ago because it didn't work. There is no need to make it work now. PrimeHunter (talk) 20:37, 27 September 2022 (UTC)Reply[reply]
@Jonesey95 there was in Special:Diff/1112735638 (it's gone now). — xaosflux Talk 20:39, 27 September 2022 (UTC)Reply[reply]
Thanks for that cleanup. FWIW, I stand by what I wrote above. Mixing programming code and actual template documentation has always been a bad idea. I don't know why it has taken 10+ years to separate them properly. (*gets down from high horse, which shows few signs of life*)Jonesey95 (talk) 21:13, 27 September 2022 (UTC)Reply[reply]

Strange transclusions

Hello! So on the article List of The Simpsons episodes (season 21–present) all other seasons have the pages transcluded and it only shows the table. However strangely seasons 33 and 34 (which i've commented out for now while I try and figure this out) show the entire article and not just the table, despite the formatting of the transclusion link being the exact same as the previous seasons. Anyone know why this is happening? ― Blaze WolfTalkBlaze Wolf#6545 19:22, 27 September 2022 (UTC)Reply[reply]

@Blaze Wolf series 33 is lacking some <onlyinclude> tags to delineate the parts that get transcluded. Have a look at the series 32 article to see where they should go. Nthep (talk) 19:30, 27 September 2022 (UTC)Reply[reply]
Ah that would explain it. I was rather confused since I Didn't actually know how it worked. Thanks! ― Blaze WolfTalkBlaze Wolf#6545 19:32, 27 September 2022 (UTC)Reply[reply]
They took me years to get my head round and I still have to think hard about them now. Nthep (talk) 19:34, 27 September 2022 (UTC)Reply[reply]
Looks like it was just some vandalism that removed them and ended up breaking it. ― Blaze WolfTalkBlaze Wolf#6545 19:35, 27 September 2022 (UTC)Reply[reply]
The IP made much more of a mess on 33 than on 34. THey've been given a final warning for it (the edit history on 33 is a little messy since i accidentally restored the wrong version when fixing it) ― Blaze WolfTalkBlaze Wolf#6545 19:50, 27 September 2022 (UTC)Reply[reply]
Re: the last bit, you need {{User screw}} on your user page. :-) Nthep (talk) 20:02, 27 September 2022 (UTC)Reply[reply]
I have fixed countless such removals of <onlyinclude>...</onlyinclude>. An episode list often adds a DISPLAYTITLE with italics for the title of the show. Part of the episode list is often transcluded in another article about the show. Without onlyinclude the whole page including DISPLAYTITLE is transcluded. This adds the page to Category:Pages with disallowed DISPLAYTITLE modifications which I monitor for articles. I sometimes add a source comment like [9] to explain that onlyinclude is needed. PrimeHunter (talk) 20:58, 27 September 2022 (UTC)Reply[reply]

Template:Television network original programming category puts cats inside themselves

Gonnym (talk · contribs) isn't around much these days, so I'm asking here. Under certain circumstances, {{Television network original programming category}} puts categories inside themselves unless it's configured properly; twice in the past Gonnym has fixed up such problems as requested at User talk:Gonnym/Archive 3#Template:Television network original programming category, User talk:Gonnym/Archive 4#Television network original programming category and User talk:Gonnym/Archive 4#Zee Marathi Original Programming, but didn't explain what actualy needed doing (there is an expln of sorts on the last one, but it's not at all clear). Now Category:Fox Kids is suffering the same problem: it's inside itself (reported at Wikipedia:Database reports/Self-categorized categories). Does anybody know what needs doing to the template in order fix the cat? --Redrose64 🌹 (talk) 20:34, 28 September 2022 (UTC)Reply[reply]

I have added |no_cat=1.[10] PrimeHunter (talk) 20:45, 28 September 2022 (UTC)Reply[reply]

Global userpages not showing up in autocomplete search results?

Has anyone else noticed this? I'm not sure if this is a new bug or if I just haven't spotted it in the past. If an editor has a global user page it doesn't appear in the autocomplete results drop down - their subpages appear fine, but not the main userpage. 192.76.8.81 (talk) 21:03, 28 September 2022 (UTC)Reply[reply]

The article is a derivative under the Creative Commons Attribution-ShareAlike License. A link to the original article can be found here and attribution parties here. By using this site, you agree to the Terms of Use. Gpedia Ⓡ is a registered trademark of the Cyberajah Pty Ltd.