Wikipedia talk:Linter/Archive 5

Archive 1

←

Archive 3

Archive 4

Archive 5

Milestone!!

As of 2024-05-05 20:16:47, NO error type per namespace exceeds 1 million. "Missing end tags" in "User Talkspace" just fell below 1 million. Still have a ways to go before we can say No error type exceeds 1 million total, but certainly getting there!

And for a second (smaller) milestone that is soon approaching, "Table tag to be deleted" will be pretty much gone from all of Wikipedia in the next few days. 137 at the moment with their days numbered. Zinnober9 (talk) 20:32, 5 May 2024 (UTC)

These two milestones show the power of bots (see Wikipedia:Bots/Requests for approval/Qwerfjkl (bot) 29, a task that eliminated about 60,000 Linter errors in a couple of days) and the power of diligent, plodding human editors (the "Table tag" error was at 5,479 four months ago, and every edit required a manual fix). Keep it up, team! – Jonesey95 (talk) 03:00, 6 May 2024 (UTC)

Do you already have a next target in mind for the bots? Gonnym (talk) 07:27, 6 May 2024 (UTC)

There are a significant number of bot-fixable patterns that appear on hundreds or thousands of pages, listed at User:MalnadachBot/Signature submissions. I think going after some of those, especially the higher priority problems, would make sense. It would be helpful if the bot did not have to return to a page multiple times, since that is something people complained about with the previous most-active bot. – Jonesey95 (talk) 13:03, 6 May 2024 (UTC)

Yeah, I'm already way past the ignore stage for hearing those cries. No one has shown that they can code a bot to fix all errors on a page in one go, so there is really no need to take that unachievable goal into consideration. Gonnym (talk) 16:29, 6 May 2024 (UTC)

I'm pretty sure that Legoktm's lint-fixing bot had a method of checking to ensure that it was fixing all of the errors that it showed up to fix, FWIW. I don't know how that was done in the code. Unfortunately, that bot, which is still needed, is no longer active. – Jonesey95 (talk) 16:41, 6 May 2024 (UTC)

I don't think that bot was active that much to begin with, based on the lint decrease at the time. Gonnym (talk) 16:48, 6 May 2024 (UTC)

Legobot fixed a few hundred thousand Linter errors, as far as I know; here are nearly 5,000 edits performing Lint fixes. I think the operator kept a tally somewhere. I know that I was grateful for its work. – Jonesey95 (talk) 17:30, 6 May 2024 (UTC)

Jonesey95, I was looking at this earlier, it seems it can be done with Parsoid; see mw:Parsoid/API#Wikitext -> Lint. I've got it working with some python code. — Qwerfjkltalk 17:17, 6 May 2024 (UTC)

Nice. If you can figure it out, that would be great. I have come across plenty of user signatures that comprise the only Linter error on a given User talk page. This signature, for example, is often the only error on a page; I have fixed a few hundred of them, but it is tedious. – Jonesey95 (talk) 17:30, 6 May 2024 (UTC)

Jonesey95, from the first page from that search query, User talk:Jim1138, I get:

{'type': 'missing-end-tag', 'dsr': [9646, 9723, 29, 0], 'templateInfo': None, 'params': {'name': 'span', 'inTable': True}}

As far as I can tell dsr[0] and dsr[1] are the start and end indexes, which gives:
e="color:darkblue">Abelmoschus <span style="color:green">Esculentus]] <sup>[[
(There are also around 10 other lint errors on the page that it also returns.) — Qwerfjkltalk 17:45, 6 May 2024 (UTC)

It's close but not the whole error.

The signature is:

'''―[[User:Abelmoschus Esculentus#s|<span style="color:darkblue">Abelmoschus <span style="color:green">Esculentus]] <sup>[[User talk:Abelmoschus Esculentus#s|<span style="color:orange">talk]] / [[Special:Contribs/Abelmoschus Esculentus|<span style="color:red">contribs</span>]]</sup>'''

the fix is several closing span tags there:

'''―[[User:Abelmoschus Esculentus#s|<span style="color:darkblue">Abelmoschus</span> <span style="color:green">Esculentus</span>]] <sup>[[User talk:Abelmoschus Esculentus#s|<span style="color:orange">talk</span>]] / [[Special:Contribs/Abelmoschus Esculentus|<span style="color:red">contribs</span>]]</sup>'''

Gonnym (talk) 18:28, 6 May 2024 (UTC)

I have a regex for it in User:Jonesey95/AutoEd/doi.js. – Jonesey95 (talk) 19:15, 6 May 2024 (UTC)

More interesting: I think that the above misidentification should be filed as a bug if it has not been already. I thought that this was a problem with LintHint, but if it an API problem, maybe it can be fixed. The problem was reported to the LintHint maintainer about six months ago, and appears to have something to do with non-ASCII characters causing offsets to be wrong. – Jonesey95 (talk) 19:18, 6 May 2024 (UTC)

I filed a bug at T365284. – Jonesey95 (talk) 18:24, 17 May 2024 (UTC)

Thanks for flagging the regression. The issue should be resolved now. ABreault (WMF) (talk) 23:52, 13 June 2024 (UTC)

ABreault (WMF), thanks for fixing it and coming here to tell us! I love it when bug reports can be addressed quickly. – Jonesey95 (talk) 00:19, 14 June 2024 (UTC)

Thank you. It appears to be working correctly again. —Bruce1ee^talk 00:49, 14 June 2024 (UTC)

┌──────────────────────────────┘
Jonesey95, from what I can tell it's the byte offsets. When I account for that properly I get
<span style="color:darkblue">Abelmoschus <span style="color:green">Esculentus — Qwerfjkltalk 17:00, 7 May 2024 (UTC)

@Jonesey95 A good number of those 5479 Table tags were also triggering a Fostered content error within the same issue, so early on I'd compiled a list of pages containing both errors and hit those pages first to get through more errors quicker. Was exciting how quickly both dropped. Congrats and thanks to @Gonnym for the last Table tag fix, and thanks to Primefac for assisting. Zero remain, leaving "HTML5-incompatible misnesting" as the sole surviving High Priority error type with 9,625 remaining. Zinnober9 (talk) 14:06, 7 May 2024 (UTC)

And now (2024-05-27 11:14:47) no page has more than 49 countable errors... Top 1000 pages are 49-36. Zinnober9 (talk) 11:19, 27 May 2024 (UTC)

Linter reports not updating and count issue

There's an issue with the Linter reports today (the Lint counters through a page's Page Info are all fine, and Firefly is all fine, it's just these three reports that are miscounting). I noticed that they were counting high yesterday, and today they timeout (600 sec) when manually updated. I had checked a few of the top pages and compared the counts to their page information lint counts, and they varied wildly, typically 21 or so higher than the actual. I know for a fact that we have the article space errors down to 2 errors max per article (barring any daily flareup), and all 1000 articles displaying in the report were all claiming 21+, which is thankfully not true.

While I see that there's now a lag on server 1, there wasn't any lag this morning when I found the update timeout issue when I clicked "manually update" and "broke" the Article and All reports, and doesn't obviously account for the miscounting.

• Does anyone know why the reports are timing out when told to update?

I've contacted SD0001 whose bot runs the requests, and Galobtter, who set up the report pages, but they don't know [1] [2].

• Does anyone know why they are counting high?

I'm not aware of a new error type, but realize that is a possible answer. Thanks, Zinnober9 (talk) 21:50, 14 June 2024 (UTC)

See Wikipedia:Village_pump_(technical)#Is_Toolforge_down?. It looks like toolforge is running very slowly. – Jonesey95 (talk) 17:35, 15 June 2024 (UTC)

Thanks. I updated the Pages by Lint Errors list and it spun and spun for a few minutes, but completed. Seems "mostly" back to working order again. Glad. Zinnober9 (talk) 22:54, 15 June 2024 (UTC)

Counts are back to normal now, thanks to WOSlinker's adjustment. Thank you. Zinnober9 (talk) 17:15, 16 June 2024 (UTC)

Puzzling use of <nowiki/>

From time to time I come across a puzzling use of <nowiki/>, for example: '''Bipinbabur Karansudha'<nowiki/>'' here, which of course generates a lint error. I don't understand what the purpose of this is. It breaks the bolding markup and messes up the rendered page. —Bruce1ee^talk 07:51, 19 June 2024 (UTC)

Sometimes people are trying to put an italicized word or phrase in single quotation marks, and this is the way they try to do it. See

'<nowiki/>''Naya Basat'''

in that article, which currently results in 'Naya Basat' due to some quirk of how the rendering engine works. I suppose it is possible that some previous version of the MediaWiki software rendered all of this mismatched/misnested syntax as the editors intended but that subsequent development has changed the rendering so that only some of it displays as intended. – Jonesey95 (talk) 16:26, 19 June 2024 (UTC)

Thanks, I see what they were probably trying to do. —Bruce1ee^talk 18:00, 19 June 2024 (UTC)

It also sometimes happens that later somewhere else in the text, some bolding or italics is added, which now causes the ''' to be matched up differently by the parser. I prefer using {{'}} (or {{`}} or sometimes {{'s}}) instead of nowiki. --rchard2scout (talk) 00:15, 20 June 2024 (UTC)

Yes, I've also seen that happen, and I also prefer using the templates to put a space between markup and apostrophes. —Bruce1ee^talk 00:43, 20 June 2024 (UTC)

Possible bot task?

Just posing a question here at this point to gauge its possibility. I've been seeing a repeating circumstance that probably be defined by an existing bot and run to reduce a sizable number of lint errors in a one time purge. The main issue I'm pausing at is the possible social aspect.

Could it be socially possible to have one time task where one of the bots go through the remaining linty Userspace pages, and replace any sandbox* denoted pages' content with {{Userpage blanked}} IF and only IF the user's last edit on any page was some reasonably long time ago (let's say 4+ years)?

The "IF user's last edit 4+ years" clause is important in this idea, as while {{Userpage blanked}} (via WP:stale), states that any page that hasn't been edited in twelve months or more could be blanked, I think we'd be in hell if we started blanking all forgotten pages of any user, active or not.

* I say "sandbox" to mean any page that isn't a page of the barnstar/userbox/template/stat... or any "profile adjacent" page... those should remain intact and delinted normally. As I'm not sure how intelligent a bot could be determining what a page is, or if it could determine between a profile adjacent page and a draft (eg Username/barnstars vs Username/draft title), I'm currently framing this idea for just pages with sandbox in the title (eg User/sandbox or User/sandbox/draftname) for all the linty sandbox tests and sandbox drafts. If the bot can determine a page is "User/draftname" and not "User/profileAdjacent" then great.

Currently though, I've been manually blanking and adding the {{Userpage blanked}} template to users' draft/test pages that I've found that are completely abandoned, using the typical judgement of "Blank the linty sandbox/draft page if and when the user hasn't edited anything in 4+ years." and then delinting as normal any forgotten pages of any editor with a recent (0-4yo) edit anywhere, or delinting as normal if the page is "profile adjacent".

Is this a bot logically possible task that could be run? Is this a socially possible task to run without any major backlash? Is any bot operator interested in this idea, assuming the social aspect is addressable and permissible? Zinnober9 (talk) 16:23, 3 July 2024 (UTC)

Zinnober9, the technical aspect of this is pretty easy (assuming there's some programmatic way to detect draft/sandbox/test pages). Finding consensus will probably be much harder - in general the community doesn't like editing users' pages without their permission. — Qwerfjkltalk 19:20, 3 July 2024 (UTC)

That's mainly why I am suggesting limiting the scope to draft/test pages of accounts that are long inactive and presumed gone, and suggesting an inactivity requirement far greater, and far more restrictive than that of the requirements of {{Userpage blanked}} so that we are a bit conservative and don't ruffle many feathers. If 4 years is too recent for the community, no objections to increasing it. 4 years is just the determining rule of thumb I've been using when I come across these specific pages. Hell, limiting it to a user's inactivity of 10+ years would still give a coverage range of 2001-2013, and would likely still clear plenty of errors on pages that are very unlikely to ever be looked at or returned to again by that user, or any other user.

The main thing that I think would be in our (delinters') favor is it would be a one and done task where no page would need to be revisited by this task, or by anyone else, bot or human. And if the user of the page did return, the template would clearly indicate that the action may be easily reversed by them without issue. I know people aren't too crazy about big tasks or repeating tasks, or edits on pages heavily watched pages, so that's why I thought it was possible that this might satisfy people for a nonrepeat task on forgotten/unwatched pages.

The only way I could see the task being rerun would be if we were permitted a 10+ years ago range to start with, and then later on being permitted to cover a newer like 4+ years ago (in which it would only run for the pages 4-10 years ago, no page revisits).

I think that WP:Lint gives a reliably long precedence for allowing specific edits on other's user pages, and that the {{Userpage blanked}} template, gives clear precedence for allowing the blanking of abandoned User's sandbox/draft pages. Combining these two rules and limiting this task to only the abandoned sandboxes/test pages that no one's looked at in years and almost certainly won't be returning to shouldn't cause any watchlist bombs, or cause any grief that doesn't already exist with Delinting. I do realize there will probably be some hesitation from the community suggesting another task or a new approach for some Lint clearing by this method. Zinnober9 (talk) 16:14, 4 July 2024 (UTC)

A minor milestone

I don't know if per tag statistics are shown anywhere, so in case they aren't: as of 13:47 (UTC) today, there were no more listed missing end tag errors for <table> or <div> left in article space (a couple have appeared since). There were ~4000 left about a month ago and I've been focusing on these, ~100 per day. I know missing end tags are supposed to be "low priority", but these two tags have plenty of potential for layout disruption when unclosed, so since i had some extra time...

I emphasized "listed" because of course more of these could easily be hidden behind italics/bold errors, just like plenty of <table> errors that weren't in the report were readily shown on the pages with <div> errors (I mean even before fixing the latter). Gamapamani (talk) 14:31, 4 July 2024 (UTC)

That's great. Here's some stats on other missing tags on article pages. -- WOSlinker (talk) 15:28, 4 July 2024 (UTC)

Tag	LintCount	PageCount
b	8820	8603
big	2	2
blockquote	15	14
caption	3	3
dl	5	5
i	60525	58411
p	561	546
s	1	1
small	31	31
span	111	108
sub	1	1
sup	5	4
u	1	1
ul	41	41

Nice! I knew I'd seen a little rash of those when I looked at the breakdown of the remaining errors in mainspace the other week and had been surprised they weren't triggering the "Table tag to be deleted" error.

Was also pleasantly surprised, (but it seems a majority of this was you efforts!) we got rid of 6000 missing end tags in Main this month, double the typical monthly rate of ~3000. I was just about to add a breakdown table of the current remaining errors in main, but I see WOSlinker just added that... lol. Nice. The one remaining detail that I'll add to that is that we have 65176 Articles with a single error, and 2476 Articles with two errors, so won't be too long before this report starts showing ones, or is all ones. Thank you (and everyone else!) for your great delinting efforts. Zinnober9 (talk) 15:59, 4 July 2024 (UTC)

That's awesome! Congratulations everyone for your work here! 🐸 Jdlrobson (talk) 01:19, 5 July 2024 (UTC)

surprised they weren't triggering the "Table tag to be deleted" error

Depends on the context, most deletable tag errors I've seen have semantically been missing end tag errors. The last table on a page would never get a deletable tag error when unclosed, but it can still mess with layout by subsuming the content after it (case 3). Which is why IMHO it might even make sense to lift missing end tags for tables to medium priority. And maybe divs too, since stuff like unwanted columnization of following content with unclosed {{div col}}) and scrolling with unclosed style="overflow: auto;" is pretty frequent.

1	2	3	4	5
Missing end tag	Missing end tag Fostered content	Missing end tag (Following content subsumed; the more cols Table 2 has, the "better" the result)	Fostered content Table tag that should be deleted (bogus)	Missing end tag (Following content subsumed, incl. Table 2)
{| class="wikitable" |- ! Table 1 |- | foo |- |} blah blah {| class="wikitable" |- ! Table 2 |- | bar |-	{| class="wikitable" |- ! Table 1 |- | foo |- |} blah blah {| class="wikitable" |- ! Table 2 |- | bar |- blah blah	{| class="wikitable" |- ! Table 1 |- | foo |- |} blah blah {| class="wikitable" |- ! Table 2 |- | bar | blah blah	{| class="wikitable" |- ! Table 1 |- | foo |- blah blah {| class="wikitable" |- ! Table 2 |- | bar |- |}	{| class="wikitable" |- ! Table 1 |- | foo |- | blah blah {| class="wikitable" |- ! Table 2 |- | bar |- |}

Gamapamani (talk) 05:29, 5 July 2024 (UTC)

Moving "Night-mode-unaware-background-color" out of tracking into high

Dark mode is highly likely to be released on web within the next month and has been on the official apps for some time. I think the tracking category is working exactly as it should be, and should be moved from tracking to high (or at least medium) priority.

In many articles, text will become unreadable in dark mode (or is already unreadable in apps), and we should be considering this a high priority to fix.

What is the process to changing priority here...? 🐸 Jdlrobson (talk) 18:03, 15 June 2024 (UTC)

If the population of the list of Template-space pages affected by this Linter tracking is an indication that there are display problems with those pages, I would be reluctant about going live with dark mode. If that long list is mostly showing false positives, then the Linter tracking needs to be adjusted before it will be useful.

To give a concrete example, this Template page is not "unreadable in dark mode", AFAICT, so why is it being flagged? As I said above, I don't think the rule should be promoted until the template space is largely free of tracked pages and pages affected by transcluded templates have been null-edited to remove them from the lists. Good progress has been made on that list, but having affected templates cause Linter error counts to balloon in pages that transclude them will not help anyone.

In any event, I think the first step is to "unhide" this Linter condition, set it to Low or some sort of Tracking-only priority, and list it on the Special:LintErrors page so that tools like LintHint and the links from Page Information can point editors to the specific locations in the page where code needs to be adjusted. We should also get it onto Firefly's tracking table. Simultaneously, and I can't stress this enough, ensure that WMF developer resources are available to troubleshoot false positives and other problems with the Linter detection and with the dark mode itself. If you really want to release dark mode in the next month, a significant amount of programming resources will be needed to ensure that volunteers have adequate tools to clear Template space of dark mode problems. – Jonesey95 (talk) 18:57, 15 June 2024 (UTC)

The recently added Missing end tag in heading lint error needs adding to Firefly as well. -- WOSlinker (talk) 19:14, 15 June 2024 (UTC)

Yeah, I'd like to see Missing end tag in heading errors added to Firefly too. That said, I was thinking about going after these heading errors (1545 errors on 1027 pages) after I finish off the Fostered content errors next week (well all that I can, I'm skipping some fostered pages right now for various reasons (A-J at the moment)). Zinnober9 (talk) 22:51, 15 June 2024 (UTC)

All Missing end tag in heading errors now corrected. Still would like this type to appear in Firefly's report so that we can more quickly see when they popup again, but it's now at "prune back" levels. Zinnober9 (talk) 15:22, 3 July 2024 (UTC)

> To give a concrete example, this Template page is not "unreadable in dark mode", AFAICT, so why is it being flagged?

I've said this a few times now, but I'm not sure it has registered. Vector 2022 is applying a rule

html.skin-theme-clientpref-night .mw-parser-output [style*='background'] {color: #202122;}

. This rule is temporary while we roll out dark mode and addresses about 80% of the issues in that lint, but importantly it is temporary. This is what the linter is detecting and why it is important it gets fixed.

If you disable that rule, you will see that the template becomes unreadable in dark mode.

> In any event, I think the first step is to "unhide" this Linter condition, set it to Low or some sort of Tracking-only priority, and list it on the Special:LintErrors page so that tools like LintHint and the links from Page Information can point editors to the specific locations in the page where code needs to be adjusted. We should also get it onto Firefly's tracking table.

Great. What is the process for this?

> Simultaneously, and I can't stress this enough, ensure that WMF developer resources are available to troubleshoot false positives and other problems with the Linter detection and with the dark mode itself. If you really want to release dark mode in the next month, a significant amount of programming resources will be needed to ensure that volunteers have adequate tools to clear Template space of dark mode problems

Yes. Dark mode is going to be released in July as an opt-in feature. Developer resources are available. and I assure you these lints are accurate. If tools or documentation are needed - just ask and we'll prioritize getting you those. 🐸 Jdlrobson (talk) 21:54, 25 June 2024 (UTC)

Thanks.
1. What CSS should I add to my vector-2022.css to disable the rule that hides dark mode problems?
2. A Phabricator ticket will probably be needed to unhide the rule and move it to a tracking or low-priority list and onto the Special:LintErrors page.
3. I have asked Firefly to modify their report to add the "missing end tag in heading" and "background color night mode" tracking. – Jonesey95 (talk) 22:40, 25 June 2024 (UTC)

1. You can add the following to your user CSS:

html.skin-theme-clientpref-night body .mw-parser-output [style*='background'] {color: inherit; }`

2. Thanks! Does there need to be any on wiki consensus? Or will that happen later? I can't seem to find any process for changing priorities of lints on mw:Help:Extension:Linter and there probably should be!

3. Thanks! 🐸 Jdlrobson (talk) 05:00, 26 June 2024 (UTC)

Re on-wiki consensus, the changes are to the MediaWiki code, so they affect all Wikipedias. Every change I have seen has been made in Phabricator. – Jonesey95 (talk) 15:22, 26 June 2024 (UTC)

I turned off the temporary workaround and this Template page was indeed unreadable in dark mode, with white text on a white background. I believe that I have fixed the underlying template, as it looks visually OK to me now, but without an indication on the Page information page, or a link within LintHint, it is tricky to know if the problem is fixed. I see that there are some JavaScript instructions on one of the help pages, but that stuff won't fly with regular folks like me who want to help fix problems. Some developer resources to help us gnomes with the above would be helpful this week before dark mode is rolled out next week (according to the VPT note just posted). – Jonesey95 (talk) 16:00, 26 June 2024 (UTC)

@Jdlrobson so I've set my preference to use Dark mode in the beta options. I'm looking at Template:Martial arts medalists for Iran which has completely unreadable text in the group sections. Does the new lint catch pages like this? If it doesn't, is there a way it can? Gonnym (talk) 11:07, 4 July 2024 (UTC)

Nope the lint is checking for the opposite - background without a color.

We would need to create a new lint for this, but the solution is more complex and there are more likely to be false positives so it would need more thought.

For this particular issue, you can fix it with the following change:

<div style="color: black"> 
<!-- =>  -->
<div style="color: var( --color-base );">

Jon (WMF) (talk) 18:53, 5 July 2024 (UTC)

Explanation of how dark mode is supposed to work?

This is a bit off-topic, but is there an explanation of the design ideas behind the dark mode? I was looking at this template page and trying to determine how to fix the current Linter errors, and I noticed that top and bottom bars above and below the navbox title were being displayed in gray instead of the intended red and blue. Also, the explicitly applied white background color was being forced to show as black, and in every navbox, there are awkward alternating gray and white thin borders on the rows. At that point, I stopped troubleshooting, because that seems broken to me, and even if I try to apply a fix, I don't know how to know if that change will result in the "right" outcome. – Jonesey95 (talk) 16:51, 26 June 2024 (UTC)

I really believe that instead of all these fixes per template that will take forever and might only look "ok", that the dark mode skin should at least at the start disable all colors and use whatever default passes AAA requirements (or at least AA). There is no way that the color contrast will pass AA requirements. Gonnym (talk) 17:49, 26 June 2024 (UTC)

I find myself in a situation that feels familiar. The WMF decides to roll out some big new change or new feature, and I try to help them with it. The documentation about the change is minimal, the set of test cases that the developers are working with is minimal, and it seems like the whole thing is going to blow up because it is not thought through very well. The catch for me is that sometimes the change works just fine, and sometimes it blows up, and my predictions in advance about which way it will go do not correlate well with the actual outcome. I just don't know. This one feels like it is not ready, but maybe nobody will depend on it or the people who use it will have a high tolerance for things not working right, so it won't matter. I don't know. I've tried to help, but ironically, it all feels like trying to make visual art in a room with no lights or windows. I think at this point I'm just going to wait until next week to see if anyone caress. – Jonesey95 (talk) 18:36, 26 June 2024 (UTC)

I say let it break, maybe that way stuff actually gets fixed correctly. All those colorful navboxes that aren't accessible in lightmode view will be horrible for everyone in darkmode. That might lead to an actual change. Gonnym (talk) 20:40, 26 June 2024 (UTC)

And maybe they'd support our delinting efforts more? *Hopeful joy* Zinnober9 (talk) 21:22, 26 June 2024 (UTC)

Are you aware of mw:Recommendations_for_night_mode_compatibility_on_Wikimedia_wikis ? 🐸 Jdlrobson (talk) 15:20, 27 June 2024 (UTC)

Yes, I've read through that a couple of times. I was unable to find an explanation of why the red and blue borders in the ice hockey navbox are turned to gray or why navbox rows have alternating white and gray borders. – Jonesey95 (talk) 17:06, 27 June 2024 (UTC)

Does this update help answer that question? https://www.mediawiki.org/wiki/Recommendations_for_night_mode_compatibility_on_Wikimedia_wikis#The_page_looks_fine_to_me,_despite_not_following_these_rules,_why? Jon (WMF) (talk) 20:34, 27 June 2024 (UTC)

Ah, disabling/updating backgrounds and borders in elements with the class .navbox ... tells me that I am probably seeing something deliberate. As has been true ever since we were asked to help, it is not clear whether any of my troubleshooting is working. I am fine to wait for the deployment to see if this blows up or is no big deal. There is plenty of other stuff to work on. – Jonesey95 (talk) 22:30, 27 June 2024 (UTC)

Special:LintErrors

Background color inline style rule exists without a corresponding text color is now showing up as a low priority item at Special:LintErrors. Only 46,386,919 errors to look at. -- WOSlinker (talk) 14:23, 4 July 2024 (UTC)

Template Action lint

Really wish that "template is up for Merge/delete/some action requiring discussion" messages didn't trigger Lint when displaying along each use of the template. The sudden, large increase in some sections and namespaces tonight are due to this and almost certainly won't need individual addressing. At least they only last a week (or until discussion closed). Zinnober9 (talk) 02:57, 7 July 2024 (UTC)

I seem to recall that this can be fixed by moving the merge/delete tag in the template, but I don't remember what it was. Pinging Jonesey95. —Bruce1ee^talk 08:27, 7 July 2024 (UTC)

I've done this edit which I think has fixed it. -- WOSlinker (talk) 10:40, 7 July 2024 (UTC)

Thanks, that should help. —Bruce1ee^talk 11:23, 7 July 2024 (UTC)

How to bypass a false result

@Anomalocaris and Bruce1ee: There is a false result, which I have reverted here. The {{limited access}} symbol needs to go next to the URL link that has limited access, in this case the link to the Internet Archive item within the {{Internet Archive}} template. This is the case with {{cite web}} when the |url-access=registration parameter is used. However, by "fixing the lint error", the {{limited access}} template is instead placed outside to the right of the {{Internet Archive}} template, which results in the symbol being place next to the Internet Archive wikilink.

This is clearly fixing a problem that does not exist. How can we bypass this? Peaceray (talk) 23:35, 6 July 2024 (UTC)

Placing {{Limited access}} in the name parameter in {{Internet Archive}} appends File:Lock-gray-alt-2.svg to the name, which generates an internal link inside an external link lint error:

[https://archive.org/details/historyofancient0000unse_y5s7 History of Ancient Egypt [[File:Lock-gray-alt-2.svg]] ] at the [[Internet Archive]]

The problem is that {{Internet Archive}} does not have a |url-access parameter, and because moving {{Limited access}} outside {{Internet Archive}} is not producing the desired result, the only solution I can see is to use {{cite book}} instead. —Bruce1ee^talk 00:33, 7 July 2024 (UTC)

Seconding {{cite book}}. Lint issues aside, it looks cleaner and doesn't have the rather clumsy arrow+indicator combo obscuring the indicator:

EDIT: The original comparison containing a lint error can be seen here. Gamapamani (talk) 04:14, 7 July 2024 (UTC)

I used {{cite book}}; sorry if I included too many details and I also omitted |via=Internet Archive, feel free to fix. —Anomalocaris (talk) 04:58, 7 July 2024 (UTC)

Update: a |url-access parameter has been added to {{Internet Archive}}. (@Peaceray: FYI). —Bruce1ee^talk 07:02, 15 July 2024 (UTC)

Thanks! Peaceray (talk) 15:07, 15 July 2024 (UTC)

User home pages

Hi, my user page was recently edited for Lint errors. I was fine with the edit, but on reading the editor's talk page I found many complaints from other editors about unwarranted interference with their user pages, both home and talk pages. I started wondering whether user pages were somehow special or sacrosanct, and started a thread at WP:HD#Etiquette for user home pages/sandboxen. Looking more closely at the main page in the 'How you can help' section, there is guidance about editing User Talk pages, but there is nothing on User Pages (home pages) at all. Should this be remedied?

WP:TALK states that "It is not necessary to bring talk pages to publishing standards, so there is no need to copy edit others' posts. Doing so can be irritating. The basic rule, with exceptions outlined below, is to not edit or delete others' posts without their permission." Furthermore, WP:UOWN states that "Bots and other users may edit pages in your user space or leave messages for you, though by convention others will not usually edit your user page itself, other than (rarely) to address significant concerns or place project-related tags."

I was wondering how the above statements sit with the general advice to Linter-Gnomes on user pages? Should they perhaps be discouraged from making edits on such pages, especially unannounced and unexplained? Cheers, >MinorProphet (talk) 11:19, 17 July 2024 (UTC)

No one owns any page on Wikipedia, not even your own user page. This isn't your web page, "home page", or social network page. Errors, however small and insignificant it might seem to you, can and should be fixed. Ideally, you should be fixing them, but as history shows, most of the users who introduce errors, don't care to fix them. Zinnober9 was correct in both fixing your errors, how they've fixed them, and even left a good edit comment describing the issue. If after all this you still have an issue, then that's one problem we can't fix. Gonnym (talk) 11:29, 17 July 2024 (UTC)

Thanks for your helpful and incisive comments, with which I tend to agree - you press 'Publish' at your peril. I am not taking sides, or complaining at all about anyone or anything. As I said I was fine with the edit, for which I thanked the editor. But others are not so happy, as you can see by the editor's talk page. I'm merely hoping to find out exactly where any consensus ("by convention") has been reached about user pages, about which many people feel quite strongly proprietorial. Maybe they shouldn't. Other guidelines about User and Talk pages include:
WP:USERTALKSTOP says: ""In general, one should avoid substantially editing another's user and user talk pages, except when it is likely edits are expected and/or will be helpful. If unsure, ask.""
WP:AVOIDABUSE says "Editing another editor's signed talk page comments is generally frowned upon, even if the edit merely corrects spelling or grammar."
WP:OWN#User pages states: "Usually others will not edit your primary user page, other than to address significant concerns (rarely); or to do routine housekeeping, such as handling project-related tags, disambiguating links to pages that have been moved, removing the page from categories meant for articles, replacing non-free content by linking to it, or removing obvious vandalism or BLP violations."" These latter seem very trivial tasks: but is everything else on user pages, e.g. in the Linter 'High Priority' list, deemed a "substantial edit"? I'm not criticizing, merely curious. Cheers, >MinorProphet (talk) 13:26, 17 July 2024 (UTC)

Fixing Linter errors fits in with "routine housekeeping". All pages are fair game for fixing errors, replacing deleted templates, adjusting wikitext to conform to MediaWiki code changes, and other maintenance that keeps Wikipedia pages rendering correctly. – Jonesey95 (talk) 13:37, 17 July 2024 (UTC)

Many thanks for a clear and concise answer which also clearly states Linter's admirable purpose. I haven't come across anything as clear is this. Does this statement (or similar) appear anywhere else, e.g. WP:Linter WP:User pages, WP:Etiquette, and the others I quoted? I think it could be more widely understood that the way a user page renders is important, and that all articles and pages on WP fall under that remit of "routine housekeeping": and that includes "your" user page, however proprietorial you may feel about it. Thanks again, >MinorProphet (talk) 16:08, 17 July 2024 (UTC)

Great discussion. You are correct that not everyone likes or is thrilled when their userspace is edited, and since I've been targeting some LINT error types that weren't as numerous and had the chance at being completely eliminated, that's meant editing mostly in userspaces to eliminate them since 72% of the remaining errors are in userspace/user talkspace. Getting some grumpy feedback just comes with this territory. Most of the time though, I've get no feedback/response at all, or I get the occasional positive thank you notification. People tend to comment more when bothered, so that's what mostly ends up on my talk page. I'm sorry if my talk page came across as a red flag due to this. I do hope you saw, however, that I responded to each and every grumpy comment about my delinting, and attempted to find neutral ground for a solution that made both parties happy.

As to discouraged from making edits on such pages, especially unannounced and unexplained, I've found it's easier to ask for forgiveness afterwards from a few people then to ask and explain the edits prior to making them for every user needing pages fixed, and I'm less likely to have my edits rejected after the fact than upfront. Also in my way of thinking, it bothers people less to go ahead and do it, take care of all known syntax errors in one edit, and be done with it all with no need for subsequent delinting or bother, than it is to ask and explain and then do the corrections, but there is no perfect way and each method has some advantages and disadvantages. One and done tends to mean fewer but bigger edits, where Errortype focused edits means smaller, clearer edits, but more subsequent edits/notifications for the remaining errortypes.

You are also correct that these issues we are fixing seem more minor, but we've also cleared up most of the more major and problematic errors in the last few years, and the remaining 3.19 million errors are mostly Misnested, Missing, Stripped, or Obsolete tags of lower, but still important, priorities. I'm still seeing pages where someone's signature is missing a closing tag and the color, font style, and/or size "leaks" onto everything that follows after it, creating some crazy and unreadable messes. I've also seen cases where people had issues with their table structuring and it mixed with other code/tables to form a distorted layout messes when read.

I do need to make clear that there is a clear and distinct line between fixing syntax errors, and correcting someone's grammar.

Fixing grammar is rarely needed, not permitted by lint or other policies, and ticks people off with no real net gain.

Correcting various syntax errors on the other hand, clears up any issues the page has of displaying content as intended and helps prevent any current issues from becoming major issues as HTML and various code improves or is updated. It also reduces the likelihood of people seeing and copying another's HTML error(s), since most people's code knowledge is "I saw this, it looked cool, so I copied it for my own use", which has caused a fair amount of replicated errors. Zinnober9 (talk) 17:04, 17 July 2024 (UTC)

@ Gonnym "most of the users who introduce errors, don't care to fix them" That, and many people don't know the errors exist, or can't figure out what's wrong or how to fix them. Thanks for replying while I was offline. Zinnober9 (talk) 17:07, 17 July 2024 (UTC)

I absolutely concur. MinorProphet (talk) 21:25, 17 July 2024 (UTC)

That's great Z9, many thanks for showing me your understanding of how it all works (or ought to). MinorProphet (talk) 21:25, 17 July 2024 (UTC)

Three errors to go; feedback requested

This month I've finished off 975 of 978 (or so) Link in Link errors, and we have three remaining due to these errors being discussed in their conversations: User talk:Beatboy16, User talk:Cyberpower678/Archive 76, and Talk:NeuroQuantology. The first two @Anomalocaris has left comments on as errors to skip, and the third is a near identical conversation of the second's error between different users (no comment currently).

While I agree that we shouldn't refactor the conversations that occurred, I'm also not of the mind to keep errors indefinitely when possible. And I realize in some cases, we just won't be able to do both. I'd like to open up the discussion about the possibility of addressing these in some agreeable way to keep the intent while addressing the error, or if these should be left alone for a better solution.

My initial thoughts are for encircling the error line(s) of each the conversations in either nowiki or pre tags, and likely adding a reply to the conversation stating the adjustment and reasoning, explaining that the reader may copy the error statement and either preview or temporarily use in their sandbox/userspace to fully understand the error, but that there isn't a need to keep the error in an active state indefinitely.

Would this be a reasonable way of handling these three errors, or is this idea getting a little too aggressive/bold on the delinting? Zinnober9 (talk) 03:20, 22 July 2024 (UTC)

Also, an additional thanks to the user(s) who in the previous few months have dropped these Link in Link errors from 3124 in March to under 1k by July. The low number was very tempting and was a refreshingly quick set to deal with after some of the more involved errors I've dealt with the past year. Grateful to you all and your efforts! Zinnober9 (talk) 03:20, 22 July 2024 (UTC)

Fixing lint errors on talk pages where the errors are part of a discussion are always tricky to handle. I think Zinnober9's suggestion of suppressing the errors with nowiki or pre and adding a note that a sandbox should be used to see the effect of the errors, probably is the best solution. —Bruce1ee^talk 06:41, 22 July 2024 (UTC)

Zinnober9's proposal is fine with me. —Anomalocaris (talk) 07:07, 22 July 2024 (UTC)

You could also put in a section link to a previous version of the page. People can click the link if they want to see the error. – Jonesey95 (talk) 12:56, 22 July 2024 (UTC)

Ah, I like the addition of view previous version. I've gone ahead and made these deactivations with these comments. If there's any issue or objections found later, we can adjust accordingly. This error is no more. Thank you all! Zinnober9 (talk) 16:38, 22 July 2024 (UTC)

That fix can probably be done in the module missing end tag and in the help fostered content to finish those off. Gonnym (talk) 17:21, 22 July 2024 (UTC)

Another milestone, and some more help needed

As of earlier today, the list of articles with lint errors, when constrained to >=2 errors per page, was empty! Special thanks to WOSlinker who I've seen help out here a lot the last few days/weeks. I've since removed that >=2 constraint, and we've got a big list of more work to do.

Next, something I'd like some help with: 2024 United States House of Representatives elections in California keeps reappearing at the top of the list, with 2-4 stripped tag errors. The page is so big that lintHint chokes on it, and I can't even edit the article with syntax highlighting enabled. Purging the page will sometimes cause the errors to disappear, but they'll eventually come back. I've made a copy of the page in my sandbox, but that one does not show any stripped tag errors in the page information. Can someone shine some light on this? Feel free to experiment in my sandbox if it helps. --rchard2scout (talk) 12:15, 23 July 2024 (UTC)

It seems as though becasuse the page is so large, sometimes there are script timeout errors, which can cause then lint errors due to failed scripts. -- WOSlinker (talk) 12:29, 23 July 2024 (UTC)

Yeah, big pages are really annoying in this regard. I'm curious if there's a logical way to fix it, either with longer timeouts (unlikely solution), or if creating some subpages of say "Districts 1-26" and "Districts 27-52" (or each District individually?), and calling them in would address it.

If the data (like the tables) were all already existing on other pages and currently copied, I'd suggest removing the copies here and calling in those tables from the other pages, but it doesn't appear that these pages are set up like that as I'm not seeing identical tables on the couple of districts I looked at. Zinnober9 (talk) 17:23, 23 July 2024 (UTC)

Probably a discussion to be had at that talk page, but splitting it (in half, in groups of tens, etc) is probably a good idea. Primefac (talk) 17:53, 23 July 2024 (UTC)

Dark mode tracking thread

I propose to use this thread to track individual questions about, and problems with, the new dark mode Linter tracking. Please report questions, problems, false positives, false negatives, and other issues here.

Possible false negatives in Template:Infobox royalty

Tracked in Phabricator
Task T369394

I'm seeing what looks like a false negative in {{infobox royalty}}. In this version of the sandbox, we have this formatting for |succession=: | headerstyle = background-color: #e4dcf6;line-height:normal;padding:0.2em 0.2em. I see a background-color defined without a corresponding color, which I think should cause a Linter error when the infobox is rendered. In this forced dark mode view of the testcases page, the value of |succession=, "Queen consort of England", shows as black text on a black background, which is an error. The testcases page reports no Linter errors.

I'm also getting a report that "from desktop, the above (name) text in black however, in mobile app it is white and does not match with the background color." On the mobile view, I'm seeing the "above" text (e.g. "King Ahab") as black text on a sort of purple background. On the dark mode mobile view of the testcases page, I'm seeing the "above" text (e.g. "King Ahab") in white on black. What could be causing the person reporting this problem to see something different? – Jonesey95 (talk) 18:00, 4 July 2024 (UTC)

Hmm odd 🤔. I am seeing this in Special:LintErrors here: https://phabricator.wikimedia.org/F56229922 so am not sure why the testcases page is not reporting an issue? Maybe worth a bug?

https://en.wikipedia.org/wiki/Special:LintErrors/night-mode-unaware-background-color?wpNamespaceRestrictions=0&titlecategorysearch=Alexander+the+Great&exactmatch=1&tag=all&template=all

The headerstyle needs to be updated to have "color:inherit;" to fix this issue.

Regarding apps, I chatted with a developer about that yesterday and it seems apps are doing a few things that need adjusting now web has dark mode. I wouldn't worry about that for now - but do feel free to get a bug on Phabricator tagged with #dark-mode and someone in apps can look into it. 🐸 Jdlrobson (talk) 02:34, 5 July 2024 (UTC)

Alexander the Great turned out to be an unlucky example. It transcluded about ten templates that all needed to be updated. The article itself, and Infobox royalty, did not need to be changed to remove the dark mode tracking lint from Alexander.

Re the header style above needing to be updated, that is the bug I am reporting. It needs to be updated, it results in invalid dark mode output, but it is not throwing a Linter error. It should. – Jonesey95 (talk) 04:37, 5 July 2024 (UTC)

DId you report this on Phabricator already? It looks like an issue with mediawiki-extensions-Linter. Jon (WMF) (talk) 18:54, 5 July 2024 (UTC)

I was being lazy or hoping someone would explain why I was wrong. T369394. – Jonesey95 (talk) 21:06, 5 July 2024 (UTC)

includeonly / noinclude Fostered Content

Tracked in Phabricator
Task T369317

Unsure if related or not, but the timing is peculiar. Today has brought on a new batch of 300some pages (most transcluded) claiming fostered content errors, due to the inclusion of includeonly and noinclude tags around startings and closings of tables. This behavior is typically not a lint error and is commonly used in the cases of transclusions. Why they are claiming errors I do not know. Anyone know what occurred? This isn't the only time I've seen this issue popup for select pages, but it's always cleared up before I get bothered enough to ask about it. Many of the pages are linked to the noinclude of the {{Ranks and Insignia of Non NATO Air Forces/OR/Blank}} template's opening bracket/parameters, and that page (really most of these pages) haven't been edited in awhile, and were not problematic when the fostered content errors were finished off to a remaining 10 or so pages two weeks ago. Zinnober9 (talk) 19:22, 4 July 2024 (UTC)

I think there was a bug report in Phabricator for this, but I am unable to find it via a Phab search. Pages like Line 4 (Shanghai Metro) appear to be affected; I remember that page and its siblings being affected by a similar problem a while ago. Since I could not find a bug, I filed a new one. T369317. – Jonesey95 (talk) 20:08, 4 July 2024 (UTC)

This phabricator bug has been fixed, so listed "fostered content" errors are real now. In articles, check for invalid recent edits (e.g. unsourced, vandalism, gibberish, terrible formatting); reverting them is often the easiest way to fix the Linter error. – Jonesey95 (talk) 00:43, 27 July 2024 (UTC)

One of the upsides of the previous error was that I've came across quite a few pages that an editor had used multiple noinclude type tags around the same section. Gonnym (talk) 11:07, 28 July 2024 (UTC)

Possible false negative in Template:DYK tools when used on a DYK page with light green background

Please see Template:Did you know nominations/St. Anne's Church, Moxi, which was showing a dark mode error. The page has a div tag that assigns a light green background. I tried added "color:inherit", but that turned the text white or light gray in dark mode. I changed that to "color:black", which displays the page properly in dark mode. The problem is that in both light and dark mode, the {{DYK tools}} box header shows black text; in the dark mode, that black text is on a black background. The page shows no Linter errors, but there is a display error. The DYK tools box header was also showing black-on-black in dark mode before I fixed the surrounding div tag. When I put {{DYK tools}} by itself in my sandbox, it shows the box header in white text on a black background. – Jonesey95 (talk) 20:32, 4 July 2024 (UTC)

I think this edit has fixed that. -- WOSlinker (talk) 21:33, 4 July 2024 (UTC)

Could really do with a full list of all those css variables and what the color values they are for both dark and light modes. Anyone know where these are? -- WOSlinker (talk) 21:35, 4 July 2024 (UTC)

What the heck is color: var(--color-base);? I don't see that anywhere on the Help page for this tracking category. Jdlrobson, do you know anyone who could improve the help page to add this useful tidbit and explain when to use it? So far, I'm limited to trial and error with color:inherit and color:black, which is not very sophisticated. – Jonesey95 (talk) 04:40, 5 July 2024 (UTC)

I've found a few of those CSS variables and have listed them at User:WOSlinker/CSSvars. -- WOSlinker (talk) 07:32, 5 July 2024 (UTC)

I updated https://www.mediawiki.org/wiki/Help:Lint_errors/night-mode-unaware-background-color#Quick_start. Hopefully that's the improvement you needed? Jon (WMF) (talk) 19:49, 5 July 2024 (UTC)