Wikipedia:Village pump (technical): Difference between revisions

Browse history interactively

← Previous edit Next edit →

Content deleted Content added

VisualWikitext

Inline

Revision as of 23:25, 13 December 2019

Table of contents
First discussion
End of page
New post

Shortcuts

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

Newcomers to the technical village pump are encouraged to read these guidelines prior to posting here. If you want to report a JavaScript error, please follow this guideline. Questions about MediaWiki in general should be posted at the MediaWiki support desk.

view · edit

Frequently asked questions (see also: Wikipedia:FAQ/Technical)

Click "[show]" next to each point to see more details.

If something looks wrong, purge the server's cache, then bypass your browser's cache.
This tends to solve most issues, including improper display of images, user-preferences not loading, and old versions of pages being shown.

No, we will not use JavaScript to set focus on the search box.
This would interfere with usability, accessibility, keyboard navigation and standard forms. See task 3864. There is an `accesskey` property on it (default to `accesskey="f"` in English). Logged-in users can enable the "Focus the cursor in the search bar on loading the Main Page" gadget in their preferences.

No, we will not add a spell-checker, or spell-checking bot.
You can use a web browser such as Firefox, which has a spell checker.

If you have problems making your fancy signature work, check Help:How to fix your signature.

If you changed to another skin and cannot change back, use this link.
Alternatively, you can press Tab until the "Save" button is highlighted, and press Enter. Using Mozilla Firefox also seems to solve the problem.

If an image thumbnail is not showing, try purging its image description page.
If the image is from Wikimedia Commons, you might have to purge there too. If it doesn't work, try again before doing anything else. Some ad blockers, proxies, or firewalls block URLs containing /ad/ or ending in common executable suffixes. This can cause some images or articles to not appear.

For server or network status, please see Wikimedia Status. If you cannot reach Wikipedia services, see Reporting a connectivity issue.

« Archives, 194, 195, 196, 197, 198, 199, 200, 201, 202, 203, 204, 205, 206, 207, 208, 209, 210, 211, 212, 213, 214

Prosesize as gadget

Hello, I've rewritten User:Dr pda/prosesize.js to be more modern and maintainable at User:Galobtter/scripts/prosesize.js. Most of the changes are internal, with the main user-facing changes are to remove the arbitrary disabling of the script on certain skins (e.g timeless) when the script works just fine on all desktop skins and a fixing of a bug in the calculation of reference text size. As prosesize is a very widely used user script (>2000 uses [1]), it seems an ideal candidate to become a gadget. If this becomes a gadget, I also propose replacing the content of User:Dr pda/prosesize.js to be mw.loader.load( ['ext.gadget.prosesize'] ), i.e to load the gadget directly so the old code no longer needs to be maintained. Galobtter (pingó mió) 07:31, 14 November 2019 (UTC)[reply]

If this becomes a gadget, I'd instead suggest changing User:Dr pda/prosesize.js to: if ( mw.user.options.get('gadget-prosesize') === null ) new mw.Api().saveOption('gadget-prosesize', '1'); instead, so that the gadget gets enabled directly rather than get loaded via the user js page; and Special:GadgetUsage would also show an accurate number of usages. SD0001 (talk) 08:07, 14 November 2019 (UTC)[reply]

That seems reasonable. Galobtter (pingó mió) 18:00, 14 November 2019 (UTC)[reply]

There is a typo at <small><i>(See <a href="//en.wikipedia.org/wiki/Wikipedia:Prosesize">here</a> for details.)<i></small>; should be a closing </i> there. Also, color me salty about mw.notify( 'Prosesize does not work with the Visual Editor.' );, but can't do anything about that on my part. --Izno (talk) 23:23, 14 November 2019 (UTC)[reply]

Thanks, fixed. Word count should at least be workable with VE (html size cannot be due to the differing html used by VE and the regular editor); it is my intention to get prosesize working with VE eventually but the last time I tried to deal with anything VE it went badly so I'll need to find the time to do it properly. Galobtter (pingó mió) 05:52, 15 November 2019 (UTC)[reply]

Izno and Galobtter, there is a page at mw:VisualEditor/Gadgets that allegedly contains some useful information and a couple of worked examples, whenever you feel up to tackling it. Ping me if you get stuck, and I'll see if I can find someone who understands it for you. Whatamidoing (WMF) (talk) 00:19, 26 November 2019 (UTC)[reply]

Thanks, I will see if I'll find the time to do so (definitely would be nice). Galobtter (pingó mió) 07:22, 8 December 2019 (UTC)[reply]

Highlighting in the new script seems busted. Also, there's a (somehow blue) link to Wikipedia:Prosesize, which is obviously wrong/missing something. Headbomb {t · c · p · b} 23:35, 14 November 2019 (UTC)[reply]

Headbomb, oops, the code for loading the css was broken. I haven't created the page yet, but Wikipedia:Prosesize would be the page describing the gadget. Galobtter (pingó mió) 05:47, 15 November 2019 (UTC)[reply]

@Galobtter: working now, thanks. Headbomb {t · c · p · b} 05:50, 15 November 2019 (UTC)[reply]

Quick bump for more comments before this gets archived. Galobtter (pingó mió) 05:22, 19 November 2019 (UTC)[reply]

Support gadgetizing (gadgetifying?). User seems relatively inactive and community maintenance would be a good thing for the 2k users. Wug·a·po·des 03:41, 22 November 2019 (UTC)[reply]

Support. Popular userscripts that meet the WP:Gadget criteria are better off as gadgets (including in terms of maintainability and performance) - Evad37 [talk] 10:37, 24 November 2019 (UTC)[reply]
I was reading Evad's tech report and upon viewing Wikipedia:User scripts/Most imported scripts had the thought that the prose size scripts would make for good gadget candidates. Lo, the next paragraph was a link to this discussion! So, yes, support, but I have a few questions. "File size" is vague—if it's the size of the HTML document, call it "HTML document size" or "total size"? On USS Chesapeake (1799)'s text-only prose size, prosesize.js spits out "24 kB" and User:Shubinator/DYKcheck.js spits out "24986 characters"—should yours round up to 25 kB if it's going to leave off digits? (Would it be worth combining those DYK features into this one? I gave up on prosesize.js because DYKcheck.js was more human-readable: characters instead of kB.) And is there a reason why this function isn't baked into the sidebar's "Page information" link? (not watching, please {{ping}}) czar 01:16, 2 December 2019 (UTC)[reply]
Yeah I have I renamed file size to "HTML document size" and rounded the bytes rather than truncated it - thanks for the suggestions! Galobtter (pingó mió) 06:16, 4 December 2019 (UTC)[reply]
@Czar, UnnamedUser, Headbomb, Wugapodes, Evad37, SD0001, and Whatamidoing (WMF): Pinging those commenting above that I've added this as a gadget in the section testing and development per the support above. I'll eventually move the gadget to the section Browsing. Galobtter (pingó mió) 07:22, 8 December 2019 (UTC)[reply]

Too often I get this "Edit conflict: Talk - Someone else has changed this page since you started editing it, resulting in an edit conflict."

Whenever I help out editing an article or talk page, I nearly always get this notification "Edit conflict: Talk - Someone else has changed this page since you started editing it, resulting in an edit conflict." I've learned to live with it and work around it:

copy/pasting my text and then hit the blue "Publish changes" button a second time; sometimes it then publishes my text;
if it doesn't, I have my text in the copy/paste memory of my computer and can again paste my text and then hit the blue "Publish changes" button a 3rd time and then mostly my changes are finally in the article or talk page.

Anybody know's what's going on? Is it because I have too many observations and reversed edits by moderators? I really study the reasons in case my additions are not accepted, at times I don't agree and I find the - how are these wikipedians called - I think "arbiters" - extremely blunt - especially when I've put 10 minutes to an hour of editing and documenting an article or clearly stated I wanted to start an article and need to park it somewhere to continue it later because my kids of wife walks in and need my attention. Anyway - thy. --SvenAERTS (talk) 08:27, 14 November 2019 (UTC)[reply]

It often happens to me when I've spent time making a single substantial edit while other editors are making many small edits. So, where possible, the solution is to reduce the size of your edits: work by making many small edits rather than a few longer ones, then you're less likely to be interrupted in this way. Where this isn't possible, you just have to get round it as you say you do. Peter coxhead (talk) 10:16, 14 November 2019 (UTC)[reply]

What's happening is described at Help:Edit conflict. If you're planning on making a long edit, make a small edit first to add the {{inuse}} tag to the top of the page. Don't forget to remove it afterwards, though. Yunshui ^雲 _水 10:30, 14 November 2019 (UTC)[reply]

@SvenAERTS: This edit suggests that you clicked Publish changes twice, perhaps inadvertently - I have a mouse that sometimes double-clicks when I intended a single click, because of contact bounce. --Redrose64 🌹 (talk) 21:21, 15 November 2019 (UTC)[reply]

@Redrose64: @Peter coxhead: @Yunshui: thx all. By the way, what is the name of the coding language with these double accolades and piping - is that java or php or python? ThySvenAERTS (talk) 09:44, 19 November 2019 (UTC)[reply]

Technically it's a markup language rather than a programming/coding language - it's called Wikitext. Yunshui ^雲 _水 09:47, 19 November 2019 (UTC)[reply]

Constant edit conflicts

So it seems, every time I make an edit, even if the last edit was months ago, that when I click "Submit", I get the edit conflict screen. I've disabled any of the functions in Beta and Preferences that might have caused it, and I still get it. Anyone know why? And for that matter, it seems the wiki software will immediately revert the edit in my name so I'll in in the history "PrussianOwl +380", and in the next line "PrussianOwl -380". I asked this at the Help Desk and they said to try asking here. Thanks, PrussianOwl (talk) 02:39, 15 November 2019 (UTC)[reply]

@SvenAERTS: especially when I've put 10 minutes to an hour of editing and documenting an article or clearly stated I wanted to start an article and need to park it somewhere to continue it later because my kids of wife walks in and need my attention.

In that case you should first flesh out the whole article in a subpage of your user page, either as a draft or in a sandbox, before submitting it for publishing in the main space. Guarapiranga (talk) 11:02, 22 November 2019 (UTC)[reply]

PrussianOwl, have you tried editing in mw:safemode? Whatamidoing (WMF) (talk) 00:22, 26 November 2019 (UTC)[reply]

Thanks, I'll have to try that! UPDATE: It works! It must be a script I'm using then right? UPDATE 2: So I tried removing all the scripts I installed one by one since I started seeing the problem, and I still saw it. I also made sure to delete my cache between each attempt. I already checked my Beta and preferences and disabled anything involving edit conflicts or saving edits, so I'm out of ideas. Also, if I click away from the edit conflict screen, my edits show up on the page as live. PrussianOwl (talk) 02:28, 26 November 2019 (UTC)[reply]

Do you have anything (especially anything that isn't popular) enabled in the gadgets for Special:Preferences? Or global scripts (on Meta) that you might have missed? Whatamidoing (WMF) (talk) 20:19, 27 November 2019 (UTC)[reply]

Nothing global, but quite a few gadgets on. Nothing that involves saving or edit conflicts though. PrussianOwl (talk) 03:10, 29 November 2019 (UTC)[reply]

@PrussianOwl:. I have seen totally unrelated gadgets break the strangest things. If you haven't solved the problem yet, you might try turning off all the gadgets in a particular section of gadgets and see if that helps. Then if that doesn't work try turning off another section of gadgets. Eventually you will find the section of gadgets where the problem lies. Then turn off the gadgets in that section one by one. Before doing any of this take a screenshot of the whole gadget preferences page. So you can remember how to get back to what you like. Firefox has a built-in screenshot tool that lets you get the whole page including the part of the page not showing on the screen. -- Timeshifter (talk) 23:57, 5 December 2019 (UTC)[reply]

@Timeshifter: Good idea, I'll have to try that! PrussianOwl (talk) 01:52, 6 December 2019 (UTC)[reply]

@PrussianOwl: Or you could just save the gadget preferences page before you start changing it: Firefox > File menu > save page as > web page, complete. I tried it, and it works fine. Some pages don't save well, or the saved page does not open correctly. -- Timeshifter (talk) 15:00, 6 December 2019 (UTC)[reply]

Slowdown on WP

Hi. Has anyone else noticed a slowdown over the last 12/24hrs on EN WP? Pages seem to take longer to load, sometimes taking a while to access talkpages too. I've also noticed that the tools box on the left-hand side of the page also takes an age to load too. I've disabled some scripts in my custom css/javascript pages, but that doesn't seem to have helped. Using Firefox in the UK if that helps. Thanks! Lugnuts ^{Fire Walk with Me} 11:15, 23 November 2019 (UTC)[reply]

I've noticed it, too. —[AlanM1(talk)]— 11:25, 23 November 2019 (UTC)[reply]

Thanks Alan. Other slowdowns I've noticed - taking ages to save a talkpage, clicking on the "what links here" and trying to view the diffs via the edit history. Please can someone switch something off, and then switch it back on again! Lugnuts ^{Fire Walk with Me} 12:17, 23 November 2019 (UTC)[reply]

Same here, though does not appear to affect Wikipedia namespace. Keith D (talk) 12:19, 23 November 2019 (UTC)[reply]

Well whatever was happening, it seems to be OK now (famous last words...) Lugnuts ^{Fire Walk with Me} 15:14, 23 November 2019 (UTC)[reply]

Been happening on and off for me, but a little longer than 24h. More like a day and a half, I think. Assumed it was my internet connection being a PITA again, but if others experienced the same, it probably wasn't on my side for once. Agree that it currently seems to be okay, but I thought that at least thrice yesterday as well, so... AddWittyNameHere 15:24, 23 November 2019 (UTC)[reply]

It was driving me nuts earlier today. Doug Weller talk 16:03, 23 November 2019 (UTC)[reply]

I'm still finding WP sluggish, even right now. It was very slow earlier today, though. --IJBall (contribs • talk) 04:26, 25 November 2019 (UTC)[reply]

@IJBall: Yeah, been experiencing some slowness on my side today again as well. Always fun when trying to do things that are already mind-numbingly boring repetitive on the best of days... AddWittyNameHere 05:44, 25 November 2019 (UTC)[reply]

Perhaps some clues can be found in "I was referred here: Wikimedia Foundation error" (below) Pi314m (talk) 22:09, 1 December 2019 (UTC)[reply]

Hmmm, same sort of thing is happening again (18:30 UK time) for about the last hour or so. Lugnuts ^{Fire Walk with Me} 18:37, 7 December 2019 (UTC)[reply]

possible add'l clue: see Could slowing down be due to TLS 1.2? Pi314m (talk) 06:30, 13 December 2019 (UTC)[reply]

AfD stats down?

Is AfD stats at tools.wmflabs down? I havent contributed to much other than AfD's recently. So one can easily see AfD votes in my contrib history, but they are not being reflect at wmf tools. —usernamekiran(talk) 17:19, 29 November 2019 (UTC)[reply]

pinging Enterprisey. —usernamekiran(talk) 17:22, 29 November 2019 (UTC)[reply]

its working again. But I think Enterprisey should see this. He seems to be offline since 2 days now. —usernamekiran(talk) 20:55, 30 November 2019 (UTC)[reply]

That's interesting - there's nothing in any of the logs from the 29th. usernamekiran, are there any specific contributions that aren't showing up on your stats page at the moment? Enterprisey (talk!) 21:42, 2 December 2019 (UTC)[reply]

@Enterprisey: hi, apologies for the delayed response. I am not sure what logs you are referring to. When I first posted here, my afd contributions (votes), were not being showed up for 2-3 days. Later they all appeared, and I again posted it here. —usernamekiran(talk) 02:30, 8 December 2019 (UTC)[reply]

Getting signed out when moving between projects

I'm quite certain this has been beaten to death before. But has anyone else been sporadically getting signed out when moving between enwiki and Commons? Frequently when I find self-written vanity articles or drafts that contain Commons images, I'll open the images and cross over to Commons to nominate them for deletion. Suddenly, this is sporadically causing me to be "logged out" when I get to Commons (or enwiki, if I'm moving the other way). If I click "Log in", it seems to "kick" immediately and show that I'm logged in to my account (I don't have to enter a username or password at all, it just goes right away). Last evening, this problem resulted in my IP being accidentally revealed on multiple pages (used a script to nominate an image) and I had to discreetly get ahold of an oversighter on Commons to hide it. I use Windows 10 64-bit with Google Chrome, and haven't noticed any problems up until recently. Anyone know what gives? Home Lander (talk) 17:56, 29 November 2019 (UTC)[reply]

I have had this happen at least half the time when moving between projects over the past two weeks - probably 5-6 times/day. About half the time, I will get the message saying "refresh, you're logged on centrally", but the other half of the time, I need to actually go through and log in with my password. Windows 10, current Firefox. Risker (talk) 19:44, 29 November 2019 (UTC)[reply]
Risker, I don't recall that I've ever even seen such a message. Usually once I sign into one project, I'm signed into anything I cross over to. Home Lander (talk) 21:09, 29 November 2019 (UTC)[reply]
Usually when this happens, I just have to wait a few seconds for it to fix itself. Occasionally, the not-too-drastic WP:BYPASS is necessary. Tip: set a global user preference that makes it immediately obvious whether you are logged-in or not - I have set MonoBook skin, which means that if I see a page that is displayed in Vector skin, I'm not logged in so know not to make edits. --Redrose64 🌹 (talk) 23:40, 29 November 2019 (UTC)[reply]

@Home Lander: The message that Risker mentions reads
Central login
You are centrally logged in as Redrose64. Reload the page to apply your user settings.
it's in a small box upper right, just below the Search box, and about the same width. --Redrose64 🌹 (talk) 00:13, 8 December 2019 (UTC)[reply]

Equation \operatorname macros?

For writing $\operatorname {E} [X]$ is there a standard way to use \DeclareMathOperator or \newcommand to avoid typing \operatorname{E} all the time? Wqwt (talk) 22:10, 2 December 2019 (UTC)[reply]

@Wqwt: Even if you could, it would only be effective within the same <math>...</math> tags. Each block is parsed separately, there's no way for state to be preserved between them. Easiest thing might be to copy it into your local clipboard, so you can "type" it repeatedly with just Ctrl+v. Or you can look into clipboard managers or macro tools for your local browser/OS... -- FeRDNYC (talk) 07:20, 3 December 2019 (UTC)[reply]

@FeRDNYC: I see. It seems MathJax on its own (for example used in math.stackexchange.com) supports it. Wqwt (talk) 20:01, 3 December 2019 (UTC)[reply]

@Wqwt: I secretly wish for some mechanism to add a standard header that gets added to the beginning of any <math>...</math> blocks, possibly separate ones: one global, and then the ability to add per-page ones for unusual things. However, I have no idea how expensive such a feature would be, or even if it would make more sense for anything like this to be stricter and more behind the scenes. –Deacon Vorbis (carbon • videos) 20:31, 3 December 2019 (UTC)[reply]

@Deacon Vorbis and Wqwt: As I see it, it's less an issue of computational expense than of other editors' ability to read and edit the formula source. The important difference between a site like math.stackexchange.com and Wikipedia is that the questions and answers at the stack exchange sites are, for the most part, single-editor content. There are some exceptions where posts have been migrated to community-ownership, and of course anyone can suggest edits to any posts there, but generally the post still has a named author who wrote the lion's share of it, and on a source-code level nobody is really required to understand it but that author.

Wikipedia, though, is a different animal entirely, and being a collaboratively-edited resource, features that allow editors to redefine the source language of the shared content would just be making it more difficult for other users to edit that content. Especially since the stackexchange MathJax implementation even allows you to do things like \def\sin{\operatorname{cos}}, and if you do then <math>\sin x</math> absolutely will show up as

\cos x

every time you use it. And even if it's not done maliciously, if I see some <math>...</math> in an article that includes a \gangnam operator because another editor defined it somewhere else on the page, how am I supposed to make sense of that?

It's fine if someone wants to redefine the language for convenience in their personal content, but as a contribution to a shared resource like any Wikipedia article, it misses the mark on the third pillar by a pretty wide margin — enough that you could even make a reasonable argument that it violates the fourth pillar as well.

(Stack Exchange, in fact, have had problems due to the fact that their \newcommand, \def, etc. implementations used to be page-scoped, which allowed posters to cause trouble for other posters either accidentally or maliciously. They've since locked things down so that the commands are scoped to the individual post, which helped reduce the degree to which each user's content could impact everyone else, but that's obviously not an option for Wikipedia articles because the entire page is a single collaborative document.) -- FeRDNYC (talk) 06:25, 7 December 2019 (UTC)[reply]

@FeRDNYC: I see where you're coming from. I still think source code readability is important enough that perhaps we can have a compromise like @Deacon Vorbis: said: a community-consensus list of standard macros for very common math operators, such as

\operatorname {E} [X],\operatorname {Pr} (X)

. Absolute value and norm are particularly annoying to do "spaced correctly" (instead of using |) and are much easier with macros: \DeclarePairedDelimiter\abs{\lvert}{\rvert} and \DeclarePairedDelimiter\norm{\lVert}{\rVert}. See https://tex.stackexchange.com/questions/43008/absolute-value-symbols/416214 But I don't have the technical or bureaucratic know-how to push this proposal forward. Wqwt (talk) 06:38, 8 December 2019 (UTC)[reply]

@Wqwt: Yeah, I could get behind that. My real concern is that it's not an ad hoc thing where each editor effectively has their own dialect of the language, because that's just anarchy. But a set of standard macros, used sparingly — or, actually I mean a relatively small set of them, because they're used frequently — make sense to me.

It's too bad that templates can't work inside <math>...</math>, since that could be a solution to writing macros in a "standard" wiki-oriented way. But of course MathJax/TeX makes far too frequent use of curly braces itself for that to ever be possible. And going the other direction is typically limited to solutions that operate outside the math domain entirely, like {{oiint}}, since the template can take <math>...</math> formulas as arguments it can't easily manipulate their contents (without becoming a complete MathJax parser). -- FeRDNYC (talk) 18:54, 8 December 2019 (UTC)[reply]

Can this happen automatically or only manually?

At the top of this talk page edit, the revision time shows as "Revision as of 00:54, 31 October 2019”. But the time stamp in each edit segment of that edit shows as "00:30, 31 October 2019 (UTC)”. It looks like it was done manually by copying the preview of a draft done at 00:30 and pasting into an edit published at 00:54. I'm considering correcting the 00:30 time stamps to avoid confusion, but want to make sure there was no way this could have arisen automatically. Thoughts? Humanengr (talk) 18:26, 4 December 2019 (UTC)[reply]

@Humanengr: none that I'm aware of, but I don't think you should change anyone else's text timestamps. The "actual" timestamp for anything we would care about is what is in the revision timestamp. This possibly came from an edit conflict, going back, copy-pasting like you suggested, etc. — xaosflux ^Talk 18:29, 4 December 2019 (UTC)[reply]

It looks like they just didn't know how to sign so they manually mimicked signatures. User talk:Samp4ngeles#Signing comments supports this. Just ignore it. PrimeHunter (talk) 21:56, 4 December 2019 (UTC)[reply]

I'll note that this can sometimes happen for normal edits due to edit stashing, although the timestamp is more likely to be off by only a minute or two. Essentially, the edited page's wikitext is parsed (and the signature expanded) immediately after you stop typing, before you click "Publish". Matma Rex talk 18:24, 10 December 2019 (UTC)[reply]

BLP noticeboard not search engine indexed, but many archives are

I noticed that, while WP:BLPN is not indexed on search engines (and to my mind, certainly shouldn't be), many of its archives are: [2]. Those often include discussion of potentially sensitive BLP matters. After a discussion with Xaosflux, it seems that, while the robots.txt exclusion file is set up properly, being transcluded or linked elsewhere still might lead search engines to crawl these, even when they're intended not to be. On Xaosflux' advice, I'd like to get a few opinions about the best way to handle this so that these will be reliably excluded from search engines, and I'm not fully conversant with exactly how Mediawiki handles that. Any expertise in that would be much appreciated. If there is a solution to ensure newly created archives are also not indexed, that would be even better. Seraphimblade ^{Talk to me} 19:14, 4 December 2019 (UTC)[reply]

One difference is that for the archives our only directive not in use is in robots.txt, on the main page we are asserting a noindex on the page itself. — xaosflux ^Talk 19:24, 4 December 2019 (UTC)[reply]
The <meta name="robots" content="noindex,follow"/> directive is in the page source, due to the NOINDEX magic word in Wikipedia:Biographies of living persons/Noticeboard/Header. — xaosflux ^Talk 19:28, 4 December 2019 (UTC)[reply]
On archive pages of other boards such as WP:ANI the NOINDEX is in an archive box that is also present on the archive pages - so that is one way to do it. — xaosflux ^Talk 19:31, 4 December 2019 (UTC)[reply]

The archives currently don't have NOINDEX - no Category:Noindexed pages - and the entry on robots.txt is formatted in such a way that it won't cover subpages if my understanding of the syntax is correct. I dunno if the correct solution would be to edit MediaWiki:Robots.txt or a Phab level solution is needed (or a bot directed addition of NOINDEX to each archive). Jo-Jo Eumerus (talk) 19:30, 4 December 2019 (UTC)[reply]

We could easily have a bot add something to those pages for now, but getting the future archives to auto-have it like how ANI's do may be better. — xaosflux ^Talk 19:32, 4 December 2019 (UTC)[reply]

For what it's worth, ANI archives are NOINDEXed through Template:Administrators' noticeboard navbox all, which the archive bots add to each archive as they are directed by the {{{archiveheader}}} parameter in {{User:MiszaBot/config}}. So one way to do NOINDEX for all BLPN archives created in the future would be to append {{NOINDEX}} to the same parameter at WP:BLPN. Of course, one would have to update all the past archives as well. Jo-Jo Eumerus (talk) 19:53, 4 December 2019 (UTC)[reply]

That's an easy enough bot run if this will be the "fix". — xaosflux ^Talk 20:19, 4 December 2019 (UTC)[reply]

@Jo-Jo Eumerus: (also see note above) - robots.txt matching should be midstring greedy and match those pages - but if you see a specific syntax error with it please let me know! — xaosflux ^Talk 19:31, 4 December 2019 (UTC)[reply]

Many of the archives are indexed because of url variations with percent-encoding. For example, https://en.wikipedia.org/wiki/Wikipedia%3ABiographies_of_living_persons%2FNoticeboard%2FArchive165 is indexed by Google while the canonical url https://en.wikipedia.org/wiki/Wikipedia:Biographies_of_living_persons/Noticeboard/Archive165 is blocked by https://en.wikipedia.org/robots.txt. %3A instead of colon is covered by an extra entry in robots.txt but %2F instead of slash is not. I see five potential ways to fix or reduce such problems:

Add __NOINDEX__ to the wikitext, typically via a template adding {{NOINDEX}}. This will place noindex in the html of any url variation (except some external mirrors).
Add more url variations to robots.txt (or cut url's off before characters with url variations).
Find out where search engines see the url variations. If the origin is under our control then replace them with the canonical url's.
Get MediaWiki to automatically noindex non-canonical url's.
Get MediaWiki to automatically redirect non-canonical url's to the canonical url which is covered by robots.txt.

Option 4 and 5 would be Phabricator requests and cannot be done locally. PrimeHunter (talk) 01:04, 5 December 2019 (UTC)[reply]

@PrimeHunter: are you confident about the %2F problem - it is hard to think it has never come up before? — xaosflux ^Talk 01:22, 5 December 2019 (UTC)[reply]

It came up at MediaWiki talk:Robots.txt#Google thinks it's cute, we need to blacklist Wikipedia%3AArticles_for_deletion%2F. I don't see another explanation. Consider Archive165 site:en.wikipedia.org. If I click the little green triangle and then "Cached" at the entry for "Wikipedia:Biographies of living persons/Noticeboard/Archive165" then I get "This is Google's cache of https://en.wikipedia.org/wiki/Wikipedia%3ABiographies_of_living_persons%2FNoticeboard%2FArchive165". If I do the same for "Wikipedia:Teahouse/Questions/Archive 164" (which is allowed by robots.txt) then I get "This is Google's cache of https://en.wikipedia.org/wiki/Wikipedia:Teahouse/Questions/Archive_164". https://de.wikipedia.org/robots.txt has some entries from de:MediaWiki:Robots.txt with all four combinations of %3A versus colon and %2F versus slash. PrimeHunter (talk) 02:00, 5 December 2019 (UTC)[reply]

@PrimeHunter: thanks for the update.

I've added the %2F version to the robots file, this will take a bit to propagate. I think the NOINDEX options are better for this use case though, if someone wants to get it fixed for future archives we can run a bot after the old ones very easily. — xaosflux ^Talk 03:35, 5 December 2019 (UTC)[reply]
If a bot's going to do that, then maybe it should NOINDEX all the Talk: archives of BLPs as well. There are probably tens of thousands of BLP talk pages with archives, and most are more poorly watched than BLP/N. Suffusion of Yellow (talk) 03:49, 5 December 2019 (UTC)[reply]
@Suffusion of Yellow: a bot run to put a future-standard template on ~200 pages is one thing, dealing with a huge number of talk pages is another - I'd be more supportive of a phab change to make "Talk namespace subpages" be NOINDEX by default. — xaosflux ^Talk 04:00, 5 December 2019 (UTC)[reply]
I think this should fix it so far as future pages [3], though please check to make sure I didn't do anything dumb. If we can do the old pages by bot, hopefully that should take care of it? Seraphimblade ^{Talk to me} 05:04, 5 December 2019 (UTC)[reply]

Whitelines before and after template

Why are there whitelines before and after the {{Schneersohn family tree}} template, see e.g. Schneersohn. Debresser (talk) 23:48, 4 December 2019 (UTC)[reply]

@Debresser: Well, it's happening because when there is no |align= specified, the template sets a style of margin: 1em auto;. Why, I do not yet know, but I'll look further at the history. —[AlanM1(talk)]— 00:20, 5 December 2019 (UTC)[reply]

@Debresser: At this edit, the behavior was changed so that, if |align= is not specified, the style margin: 1em auto; is set. Previously, it looks like the style was a margin of 0, which is typical for a navbox. The current version produces a 1em empty space above and below. I'll also note that it has a white background by default, instead of the usual purple-ish used for navboxes. I'm inclined to remove the margin style or put it back to being an unconditional 0, as it was before the edit. Also, I think the background color should default to that of other navboxes, as it's kind of ugly when the margin is removed IMO. I'm concerned that I don't have the time at the moment to research how it's used everywhere. Would someone more familiar with navboxes handle this and/or comment? —[AlanM1(talk)]— 00:45, 5 December 2019 (UTC)[reply]

Thank you for your comments. I removed the culprit.[4] Can somebody please make the edit to add the blue color, in line with other navboxes? Debresser (talk) 19:21, 5 December 2019 (UTC)[reply]

Table totals and total percentages

Is there a way of making 2-step calculations in tables without running into node-count errors?

I've for instance fully automated this article, converting what previously were pasted numbers into the formulas that calculate them and pulling the input info from a single main source, but, for the life of me, can't figure out how to calculate percentages from totals without running into node-count errors. Best I could do was use a pasted running total onto the page and hide it in a span tag. I looked into doing that with subst, but couldn't figure out how. I stared at this code for half an hour, and couldn't see how it could help either. Guarapiranga (talk) 20:18, 5 December 2019 (UTC) This may be unrelated to the 2-step calc, though, as the same page for North America blows over the preprocessor node count limit even without attempting to do the 2-step calc. The culprit is obviously {{country population}} (mea culpa), which transcludes numbers from List of countries by population (United Nations). I realise each transclusion in the template adds to the load, even if it's unused in the output, and {{country population}} makes 8 of them, but I don't see how else to do it (and how that this on a list of ~200 countries adds up to +1M preprocessor nodes. Guarapiranga (talk) 01:12, 6 December 2019 (UTC)[reply]

Put the data into a module and use another module to do the calculations? Seems a better way to me since all data and the calculations that depend on the data are in just two modules. Another module can be written that would take the raw data from wherever it comes from and properly format it for your application. I did something similar for the {{lang}} and {{lang-??}} templates and the modules that support them. See Module:Lang.

—Trappist the monk (talk) 01:29, 6 December 2019 (UTC)[reply]

Thanks, Trappist the monk. Had a look at Module:Lang. I see how it works, but can't see how to make it work for data that is not static like language codes. Those population numbers don't change all that often—once a year, if they don't need corrections—with the daily update being done via exponential interpolation, but if they're in a list article they can be updated and corrected by common editors, with no need of delving into Lua code. Also, not sure if this is related, but I have a heap of trouble when I try to subst Module:ISO 3166 so I can get ISO codes for labelling sections (I have to subst twice, and the page rendering times out a couple of times on a full list of countries). Guarapiranga (talk) 01:48, 6 December 2019 (UTC)[reply]

You can use the list article data from a Lua module. Load the content of the page with the title object and then just use gsub to grab the sections for the required contries. This method just loads the list article contents once and keeps the transclusion size and depth numbers down. Jts1882 | talk 10:04, 6 December 2019 (UTC)[reply]

@Guarapiranga: I should have added that I have a rudimentary lua function that retrieves labeled sections (see function p.section() at Module:Sandbox/Jts1882/Test). Invoking it via {{section/sandbox}} can retrieve the population figures from List of countries by population (United Nations), e.g. {{section/sandbox|List of countries by population (United Nations)|BRA_0}} --> Section not found. Jts1882 | talk 13:20, 7 December 2019 (UTC)[reply]

@Jts1882: May I ask what feature Module:Sandbox/Jts1882/Test provides that isn't supported by the built-in #section parser function. * Pppery * _{it has begun...} 23:42, 7 December 2019 (UTC)[reply]

@Pppery:For some reason there is an issue with transcluding labelled sections from templates using other templates. So if I want to reuse section "Magnolids" in {{Phylogeny/APG IV}}, #section doesn't work, but the same labelled section can be retrieved from my sandbox or if it was the sole content of a template. The problem is something to do with the template nesting and the order of transclusion.

Description Code Output Comment

Using #section to get section from my sandbox

{{#section:User:Jts1882/sandbox|Magnolids}}

	Canellales Cronquist 1957

	Piperales von Berchtold & Presl 1820

	Magnoliales de Jussieu ex von Berchtold & Presl 1820

	Laurales de Jussieu ex von Berchtold & Presl 1820

success

Using #section to get section from template

{{#section:Template:Phylogeny/APG IV|Magnolids}}

fails

Using lua module function

{{#invoke:Sandbox/Jts1882/Test|section|Template:Phylogeny/APG IV|Magnolids}}

	Canellales Cronquist 1957

	Piperales von Berchtold & Presl 1820

	Magnoliales de Jussieu ex von Berchtold & Presl 1820

	Laurales de Jussieu ex von Berchtold & Presl 1820

success

There are also benefits on post-expand include. If I understand correctly, multiple calls with #section add the whole page size for each call, which I think is the issue in this question (as the country population list is quite large). The module version only adds the size of the content returned. Jts1882 | talk 08:28, 8 December 2019 (UTC)[reply]

That indeed seems to be the case, Jts1882, as I quickly found out trying the same method with the List of countries by past and future population. Guarapiranga (talk) 04:17, 10 December 2019 (UTC)[reply]

The language data aren't all that static. Since November 2017, Module:Language/data/iana languages has been updated from the IANA language-subtag-registry file eleven times. That particular module is 7900+ lines so updating it manually would be a headache. So I wrote a tool, Module:Language/data/iana/make that reads the subtag-registry file and makes the ~Language/data/iana languages and the other registry-derived modules so that no one has to manually update them.

For your South American population example, the data are purportedly taken from a United Nations document – so says the citation. If the data are in a list article they can be updated and corrected by common editors then some sort of provision must be made to source those changes. It seems to me better to use the data from the 'official' source in your table. If there are other sources that disagree with the 'official' source, that should be noted and cited in the article text. Further, editors do, quite often, break wikitables so whatever it is that reads the wikitable must be very robust because a simple omission of a single pipe in a wikitable could potentially break multiple articles dependent on that wikitable. Of course the wikitable could be a protected template but that would preclude correction by common editors. What a muddle.

I cannot speak to your troubles with Module:ISO 3166 except to say that a module solution could mw.loadData ('Module:ISO 3166/data/National') and use that directly – no substing required.

—Trappist the monk (talk) 12:54, 6 December 2019 (UTC)[reply]

Significantly slow page opening when using FireFox and being logged in and the global language is set to British English

I asked about this at the Help Desk, but someone told me to ask here instead. It does not seem to happen on Chrome, but on FireFox. I have tried disabling all extensions, and that did not solve it. Page opening delay (at the "Waiting" stage) varies, but mostly about 3~6 seconds, sometimes reaches 10 seconds. — Preceding unsigned comment added by Sin Jeong-hun (talk • contribs) 00:17, 6 December 2019 (UTC)[reply]

@Sin Jeong-hun: if you change your language to "English" is it better? — xaosflux ^Talk 13:13, 6 December 2019 (UTC)[reply]

@Xaosflux: Yes, setting it to "English" makes it better, so I changed it from British English to English as a workaround. Sin Jeong-hun (talk) 00:56, 9 December 2019 (UTC)[reply]

@Sin Jeong-hun: Can you give a link to a page where this happens? I tried it with English Baccalaureate and see no significant difference in performance between en and en-gb. I do show 3 extra requests, but those are related to gadgets that I have enabled (MoreMenu, Cat-a-lot, and HotCat apparently have to load localized messages for other than lang=en), and they are only a few tens of ms out of an average of about 2000 ms total. —[AlanM1(talk)]— 09:57, 7 December 2019 (UTC)[reply]

Use the Performance tab in the Developer Console and you will get a breakdown of what is taking what time. --qedk (t 桜 c) 10:17, 7 December 2019 (UTC)[reply]

@AlanM1:As far as I had tested, it happened in almost all pages. For example, when I posted this question, I opened the "Dog" page, and then clicked "Wolf" and then "North America", etc. It happened in ALL pages. I used the Developer's Tool in FireFox and the delay I was talking about was the "waiting" for the main HTML page, not its resources, etc. So, it does not happen, and you re using FireFox? Mind you, it did not happen for me when I used Chrome. I just test it again. See the screen captures.

I had changed the global language to British English, and opened the "greywolf" page by clicking the link on the "dog" page. Notice the 6390 ms delay.

Then, I changed the global language to English, and clicked the "North America" page on the "wolf" page. Notice that the delay is now 600 ms.

[[User:Sin Jeong-hun|Sin Jeong-hun}} (talk) 01:12, 9 December 2019 (UTC)[reply]

{{Re|Sin Jeong-hun|| Once you open it, if you refresh the page is it faster the second time? I'm not sure but I doubt we maintain a cache of every language layout for every page. — xaosflux ^Talk 14:35, 10 December 2019 (UTC)[reply]

Script for grouping contributions by page

Does anyone have a script to group Special:contributions by page? Guarapiranga (talk) 21:04, 6 December 2019 (UTC)[reply]

Found User:PerfektesChaos/js/listPageOptions but it didn't work for the Special:Contributions page (as the 'advanced options' in Special:Preferences#mw-prefsection-rc don't either). Guarapiranga (talk) 20:29, 7 December 2019 (UTC)[reply]

Characters not displaying

Hi, the superscript characters in {{Surviving ocean going ships}} aren't displaying for me. What might the reason be? Don't get too excited, I'm running Firefox 47.0.1 on XP SP3 and it may well be just too old to cope It's not really not important, just technical curiosity. Cheers, MinorProphet (talk) 19:39, 7 December 2019 (UTC)[reply]

Between the OS and browser, I'd say it's the age of your system. --Izno (talk) 20:36, 7 December 2019 (UTC)[reply]

Can you see them here?

Falie^⚓ (1919)
Hanne Marie^⛵ (1919)

— xaosflux ^Talk 20:38, 7 December 2019 (UTC)[reply]

Xaosflux Hi, thanks for replying, no I can't.

Izno Hi, can you describe the actual characters? What font/character set do they come from?

On that first one above it is a unicode character, U+2693. Is it appearing as a box with numbers in it or just not at all? — xaosflux ^Talk 23:57, 7 December 2019 (UTC)[reply]

The box with numbers. MinorProphet (talk) 02:43, 8 December 2019 (UTC)[reply]

The other is Unicode U+26F5. The first is an anchor and the second is a sailboat. --Izno (talk) 01:08, 8 December 2019 (UTC)[reply]

@MinorProphet: The two characters ⚓ 'ANCHOR' (U+2693) and ⛵ 'SAILBOAT' (U+26F5) are both in the Unicode Miscellaneous Symbols block. If you're getting little boxes containing four hex digits (numbers 0-9 and letters A-F) instead, that means that no suitable font is installed on your machine; it's known as "tofu". IIRC with Firefox 51.0 they introduced built-in emoji support, not requiring special font installation. As for Firefox under XP SP3, you should be able to take Firefox up to at least version 52.9, if not further. --Redrose64 🌹 (talk) 10:35, 8 December 2019 (UTC)[reply]

Redrose64 Hi RR, I've tried FF 52.9, but it seemed very slow compared to 47 and would often grind to a halt and crash horribly. Haha! I just found SymbolA.ttf and Unifont.ttf, installed them both, and now I can see anchor⚓ & sailboat⛵. Fantastic work everyone, thank you for your helpful comments - one satisfied customer.

Done MinorProphet (talk) 13:30, 8 December 2019 (UTC)[reply]

Nice work. For future reference if people are searching the archives, Noto fonts has information about an emoji font that might also work. – Jonesey95 (talk) 15:48, 8 December 2019 (UTC)[reply]

Ahh, {{de-indent}}, my favourite template, CR/LF for the heptaphobes. Just for the record—for those with XP SP3 etc.,—I found SymbolA.ttf here, and Unifont.ttf here. MinorProphet (talk) 22:07, 8 December 2019 (UTC)[reply]

PS: I can now see Chinese / Japanese / East Asian characters as well... Like, 1337 totally recommend. MinorProphet (talk) 00:30, 9 December 2019 (UTC)[reply]

Book Creation Tool Not Allowing Creation of PDFs

I am currently trying to create a book with a variety of articles. When I go to the "Manage Your Book" page, the option to "Download" is greyed out.

Upon clicking the "Learn More" link, I noticed that discussion on the tool had largely concluded (since there has been no update since July of this year) and a new version of the PDF renderer has been launched. Is there a reason for the lack of an e-book downloading functionality? Or has the tool simply not been re-instated. — Preceding unsigned comment added by 70.176.68.225 (talk) 20:22, 7 December 2019 (UTC)[reply]

Working for me, example - can you be more specific, exactly what page are you on, what are you clicking, what is happening? — xaosflux ^Talk 20:40, 7 December 2019 (UTC)[reply]

Initial table sort

Tracked in Phabricator
Task T33332

I see this has been asked several times in the past, but I was wondering whether there's been any progress toward implementing configurable initial table sorts. In the past, one frequent answer to this question has been: just save it sorted already. That obviously doesn't resolve the issue for tables dynamically updated through formulas (e.g. population clocks) or transclusion from other pages. At the end, what this rigidity ends up promoting are frivolous edits to keep rank tables in the right order. Ideally, one should simply be able to specify in the data-sort-type tag, in addition to the data type, also whether the first sort should be ascending or descending, and whether the table should be pre-sorted by the ascribed column as soon as loaded onto the screen, e.g.:

! data-sort-type="alpha; first: ascending; initial: 3;" | Countries
! data-sort-type="number; first: descending; initial: 1;" | Population
! data-sort-type="date; first: ascending; initial: 2;" | Date

Guarapiranga (talk) 03:50, 8 December 2019 (UTC)[reply]

@Guarapiranga: you could try re-opening phab:T33332 (add some more details to it). — xaosflux ^Talk 04:07, 8 December 2019 (UTC)[reply]

Done, Xaosflux. Guarapiranga (talk) 13:14, 8 December 2019 (UTC)[reply]

WP:Accessibility is a requirement; the tables displayed on a page with such 'default sorting' will not match the content provided to non-sighted readers. I think this is a flat reject. --Izno (talk) 15:56, 8 December 2019 (UTC)[reply]

That would only be an issue for client-side sorting though - server side sorting prior to rendering the page would allow for consistency. — xaosflux ^Talk 17:14, 8 December 2019 (UTC)[reply]

Tracked in Phabricator
Task T240114

Whatever happens with this, it would also be nice to give the Visual Editor the capability to rearrange a table in alphabetical order. I am talking about rearranging it in the wikitext. Then one saves the page. Then the initial sort is now alphabetical. That means all future edits by individual editors do not mess up the sort order.

Converting a table to initial alphabetical order now is difficult for most editors. So there are many tables completely out of whack order-wise. Because someone spent the time to order it long ago in rank order. But over time that changes with each edit. So the table is no longer in any order at all initially. -- Timeshifter (talk) 17:17, 8 December 2019 (UTC)[reply]

@Timeshifter: I've opened a FR for this for you at phab:T240114. — xaosflux ^Talk 17:31, 8 December 2019 (UTC)[reply]

Thanks Xaosflux. I had to think awhile to figure out that FR stands for feature request. :) -- Timeshifter (talk) 20:26, 8 December 2019 (UTC)[reply]

(talk page stalker) Xaosflux, please forgive the intrusion: which "alphabetical sort order"? As defined by what? This - for me - approaches the central problem of Tables (may they live for ever). I have battled with sort order in Tables™ during all my time on WP. Again, this is mere technical curiosity - could you perhaps point me towards why sorting tables seems particularly difficult for us mere mortal editors? >MinorProphet (talk) 23:21, 8 December 2019 (UTC)[reply]

@MinorProphet: See Help:Sorting and the sections on initial alphabetical sorting. See also: Commons:Commons:Convert tables and charts to wiki code or image files#Sort by rank or alphabetically. Use LibreOffice Calc. -- Timeshifter (talk) 00:07, 9 December 2019 (UTC)[reply]

@MinorProphet: for the VE active-editing solution, I don't think this has anything explicitly to do with "alphabetic" sorting, the suggestion would be to allow any column to be sorted by its sort type, then save to the source however it is currently sorted for the person making the edit - just as if they manually edited all the rows in the source. — xaosflux ^Talk 00:12, 9 December 2019 (UTC)[reply]

@Xaosflux: Wow, is that a quantum change understandable by yer average WP:AWARE editor, or did you just blow my mind?

@Timeshifter:, thanks for the links. MinorProphet (talk) 01:13, 9 December 2019 (UTC)[reply]

@MinorProphet: just that columns could have different data types to "sort" by, such as alphabetic, numeric, date, etc. While integers (e.g. 3,4,5,1 -> 1,3,4,5) normally "sort" alphabetically OK, floating point numbers usually don't (e.g. 3.1, 31, 300, 40, 20 should be 3.1, 20, 31, 40, 300 not their alphabetic sort of 20, 300, 31, 3.1, 40). — xaosflux ^Talk 01:21, 9 December 2019 (UTC)[reply]

Sooooo.... As I understand it (as a mathematical deficent), there is no internationally-agreed, RFC-acknowledged, "Universal Default Sort Protocol" (UDSP). Maybe the world awaits. NB I once attempted to code a generic bubble-sort in Borland Turbo Basic. I still bear the scars. So as not to bore you with RFC 2324], I leave you with RFC 2549. >MinorProphet (talk) 03:27, 9 December 2019 (UTC)[reply]

Not sure what the accessibility issue is exactly, but the VE sorting button, while desirable, does nothing to ameliorate the problem with dynamically updated ranking lists that are either perpetually out of order, or constantly inviting frivolous edits. Guarapiranga (talk) 04:14, 9 December 2019 (UTC)[reply]

@Guarapiranga: I wouldn't get your hopes up. I think it would be taxing on the servers. Combining dynamic formulas and transclusions and sorting is a lot of work for the servers. I notice similar problems when I go to sites that offer these big menus of data to choose from: OECD, WHO, CDC, etc.. They are often a total pain to use. And they are often slow. It depends on the time of day, the luck of the draw, and how many people are accessing the server at the moment. Alphabetically organized data is much less taxing on everything and everybody. -- Timeshifter (talk) 04:46, 9 December 2019 (UTC)[reply]

Are you saying country rankings, for instance, should be sorted alphabetically rather than ranked, Timeshifter? Guarapiranga (talk) 05:02, 9 December 2019 (UTC)[reply]

@Guarapiranga: Yes. Thanks for that list link. It goes in my to-do folder of bookmarks for categories of lists. Until the recent editing and improvement of the {{Rank}} template for easily adding a fixed row-number column there was no easy way to rank alphabetical lists. Now we just add the row-number column and tell people to click the column head of any sortable column they want ranked. I have advocated for an easy way to add a fixed row-number column for years. See: Help:Sorting#Auto-ranking or adding a row numbering column (1,2,3) next to a table and phab:T42618: "jquery.tablesorter: Add support for a 'fixed' column of row numbers." -- Timeshifter (talk) 07:39, 9 December 2019 (UTC)[reply]

It has been pointed out to me though that row alignment is off when viewed in mobile browsers. This is when using the previously mentioned {{Rank} template. See discussion here:

https://en.wikipedia.org/wiki/Talk:List_of_countries_by_population_(United_Nations)#Alignment_problem

Initial alphabetical sorting is very useful when combined with the ranking provided by a fixed row-number column that stays aligned. This is another reason for an integrated, fixed row number column as proposed at phab:T42618. -- Timeshifter (talk) 07:11, 11 December 2019 (UTC)[reply]

New users creating drafts in template namespace

I've been noticing lately that an increasing number of new users are creating what are supposed to be drafts in template namespace instead. Special:AbuseFilter/994 is intended to detect these creations. However, a stranger trend has also emerged - titles like this, with "Template:Please leave this line alone" followed by their intended title. How exactly are new users coming up with this? Is there a link somewhere that could be modified to stop this? Home Lander (talk) 19:34, 8 December 2019 (UTC)[reply]

Home Lander, that comes from the first line of Wikipedia:Sandbox. Editors change the template (which isn't obviously a template), then click the red link. —Kusma (t·c) 19:53, 8 December 2019 (UTC)[reply]

It's a frustrating problem that I agree seems to be growing over time. I also see titles starting with "Infobox", e.g. Template:Infobox Tennis Player and Template:Infobox musical artist ==. I asked about this at WT:AFC a while back. My suggestion was to put some language in MediaWiki:Newarticletext, shown just for new users creating templates, that briefly explains what templates are and that articles should go at WP:WIZARD or WP:DRAFTS. I don't know how successful it will be but seems worth a try. — MusikAnimal ^talk 04:06, 9 December 2019 (UTC)[reply]

I don't think that will have much of an effect. These are probably people who don't read the instructions, just play around and notice that they are not allowed to create [[]] redlinks, but are allowed to create {{}} redlinks. Simplest way out is to allow them to create articles ;) —Kusma (t·c) 08:31, 9 December 2019 (UTC)[reply]

Wouldn't the actual simplest way out be to make WP:ACPERM apply to the template namespace? Why should users who aren't allowed to create articles be able to create templates? --Ahecht (TALK
PAGE) 16:37, 9 December 2019 (UTC)[reply]

Yes, that would also work. I'm still opposed to WP:ACPERM, so I don't support extending it to other namespaces to fix problems caused by WP:ACPERM. —Kusma (t·c) 10:12, 11 December 2019 (UTC)[reply]

Include one template inside another

Resolved

I'd like to be able to include {{Infobox nuclear reactor}} inside {{Infobox rocket engine}}. Any ideas on how this could be done? Hawkeye7 (discuss) 22:51, 8 December 2019 (UTC)[reply]

@Hawkeye7: See Wikipedia:WikiProject Infoboxes/embed, there's multiple approaches. --qedk (t 桜 c) 23:14, 8 December 2019 (UTC)[reply]

Thanks for that! Hawkeye7 (discuss) 01:06, 9 December 2019 (UTC)[reply]

Glad to be of help.

--qedk (t 桜 c) 21:41, 9 December 2019 (UTC)[reply]

Well I'm close. I've made a parent template capable of holding a module and the child capable of being one. (User:Hawkeye7/Sandbox7) The only thing I'd still like is for the child to emit a header, but only if it is a child. Hawkeye7 (discuss) 03:02, 10 December 2019 (UTC)[reply]

@Hawkeye7: I've made the required changes (added an optional title too since I wasn't sure which header, you can remove that). Feel free to ping me if you need any more help! --qedk (t 桜 c) 09:41, 10 December 2019 (UTC)[reply]

That's great! Thanks for your help! Hawkeye7 (discuss) 09:43, 10 December 2019 (UTC)[reply]

Wikitable header centering: desktop vs mobile

Continuing from Special:MobileDiff/929887874 (per Redrose64's recommendation)

Done, Snaevar: T240106. Guarapiranga (talk) 00:33, 9 December 2019 (UTC)[reply]

So where's our visual differ?

Now that we have a much-maligned visual editor, where is the much-maligned visual diff? It would really be nice to improve the diff tool in various ways. One thing is that I can't compare A/B images that were swapped, except if I switching revision previews back and forth. And many times formatting, text flow, wrapping, all becomes an issue for newly added diffs, especially people hunting vandalism such as myself. What else would you like to see in a next-generation diff engine? Where is this in Phabricator? Elizium23 (talk) 05:22, 9 December 2019 (UTC)[reply]

mw:VisualEditor/Diffs, go to Special:Preferences#mw-prefsection-betafeatures to enable it. Galobtter (pingó mió) 05:29, 9 December 2019 (UTC)[reply]

Weird collapse template formatting

In Wikipedia:Deletion review/Log/2019 December 8, there's two {{collapse}} sections. In one, the "show" button is on the right margin. In the other, it's to the left of the descriptive text. What's going on there? -- RoySmith (talk) 15:26, 9 December 2019 (UTC)[reply]

It looks like the difference in positioning between using classes mw-collapsible and collapsible, although both are using the former here. If you remove the text between the two collapsible boxes the "show" is on the right as expected in preview. I'm guessing that some CSS is interfering, possibly related to the some mismatched HTML tags. The toggle for the collapsible elements is very sensitive. Jts1882 | talk 16:12, 9 December 2019 (UTC)[reply]

The collapse template, which creates a table within a set of <div>...</div> tags, is being used within an HTML list, which looks like it is causing this strangeness. – Jonesey95 (talk) 17:08, 9 December 2019 (UTC)[reply]

Yes, never indent either the start or end tags for a collapsible box. --Redrose64 🌹 (talk) 19:26, 9 December 2019 (UTC)[reply]

Tech News: 2019-50

Latest tech news from the Wikimedia technical community. Please tell other users about these changes. Not all changes will affect you. Translations are available.

Recent changes

Wikimedia projects use Translatewiki to translate the wiki interface. You can now use WatchTranslations to watch projects there. You would get an email if there are missing translations to your language. [5]
There is a new dataset you can use. It shows the number of editors per country per month for a number of countries. You can read the documentation and download the dataset.

Changes later this week

The new version of MediaWiki will be on test wikis and MediaWiki.org from 10 December. It will be on non-Wikipedia wikis and some Wikipedias from 11 December. It will be on all wikis from 12 December (calendar).

Meetings

You can join the technical advice meeting on IRC. During the meeting, volunteer developers can ask for advice. The meeting will be on 11 December at 16:00 (UTC). See how to join.

Future changes

You can test a new reference tool. It makes it possible to reference different parts of a source without repeating all information. You can test it on the beta cluster. You can see an example article. [6]

Tech news prepared by Tech News writers and posted by bot • Contribute • Translate • Get help • Give feedback • Subscribe or unsubscribe.

16:35, 9 December 2019 (UTC)

Nonsense from mobile editors

Moved from Talk:Main Page

– Suffusion of Yellow (talk) 21:47, 9 December 2019 (UTC)[reply]

Pretend you have never edited Wikipedia. Do not pretend you are a complete idiot.

Go to the Main Page, and switch to mobile view
Scroll to the bottom. Note the big button labeled "Talk". Click it.
This brings you to a very simple page. At the top, there is a big blue button labeled "Add discussion". Click it.
Now you will see another simple page, with one field labeled "Subject", and another labeled "What is on your mind?" (!). That's it. No edit notice. No chance to see Talk:Main Page/HelpBox. No mention of what this page is about. Literally just an invitation to share your thoughts.

And yet we're surprised by all the off-topic stuff that gets posted here.

Anything we can do about this? Before anyone says "edit filter", I think think the first choice should always be to alert users before they make a mistake, so as not to waste their time. Suffusion of Yellow (talk) 18:04, 9 December 2019 (UTC)[reply]

And no preview option. Thought that was hidden behind "publish", like in VE or the 2017 editor, but no. Suffusion of Yellow (talk) 18:06, 9 December 2019 (UTC)[reply]

I did try to find which MediaWiki namespace page sets that default text "What is on your mind?", but couldn't find anything. @Whatamidoing (WMF) and Whatamidoing:, help, please? Jo-Jo Eumerus (talk) 18:18, 9 December 2019 (UTC)[reply]

@WhatamIdoing: Jo-Jo Eumerus (talk) 18:18, 9 December 2019 (UTC)[reply]

Thanks for the ping, Jo-Jo Eumerus. That's MediaWiki:Mobile-frontend-talk-add-overlay-content-placeholder. It looks like it has said that since 2013. Whatamidoing (WMF) (talk) 18:26, 9 December 2019 (UTC)[reply]

Just one more reason we should do away with the mobile version. Yes, people are using their phones more and more. BUT if the phone can't display the entirety of Wikipedia without us having to chop it down, then that phone really SHOULDN'T be used for viewing (much less editing) this site. Mobile phone users complaining about their access to this site (or lack there of) are like radio listeners complaining that they can't see the TV image on their devices. --Khajidha (talk) 19:18, 9 December 2019 (UTC)[reply]

This would be rather drastic, but we could hide the talk button by putting something like this in MediaWiki:mobile.css:
```
.page-Main_Page #page-secondary-actions .talk {
	display: none;
}
```
It's not there for IP editors, anyway. Here's what we'd be missing out on. Suffusion of Yellow (talk) 19:59, 9 December 2019 (UTC)[reply]

This seems like a broader problem with the mobile interface for all talk pages, not just this one. I too question why we allow any edits from the mobile version (it's great for viewing). However I feel like this is something for WP:VPT to resolve on a site-wide basis, not for T:MP to come up with a kludge for just one page. Oddly, I don't see the 'talk' button on my phone, but I do if I switch my desktop browser to mobile view *shrug*. Modest Genius ^talk 20:28, 9 December 2019 (UTC)[reply]
@Modest Genius: Agree it's a sitewide problem, but the talk link is at least harder to find on other pages. Only the MP has the special big button at the bottom, at least for me. Are you logged in on your phone? Suffusion of Yellow (talk) 20:40, 9 December 2019 (UTC)[reply]
Ah that might explain the difference - I was logged in on desktop, not on my phone. When I do see the 'talk' button it appears on every page, not just the MP. Modest Genius ^talk 21:20, 9 December 2019 (UTC)[reply]
Oh, I see now. "Advanced" mode (didn't know I had that set) hides it from other pages. In the non-advanced mobile view, it's on every page. Suffusion of Yellow (talk) 21:39, 9 December 2019 (UTC)[reply]
No, mobile view is a very useful editor and I use it all the time, and the new mobile editor is has been extremely useful. Just because you personally do not use it does not mean it is useless. And your analogy is stupid. Nixinova T C 06:04, 10 December 2019 (UTC)[reply]
Mobile user with 10000+ edits here. No thanks, I would not like to be forced to permanently change to the desktop version. Glades12 (talk) 10:22, 10 December 2019 (UTC)[reply]
@Glades12: I agree. The question is, how can we clue in new editors as to what talk pages are all about, if we can't show them edit notices, or talk page headers? Suffusion of Yellow (talk) 17:29, 10 December 2019 (UTC)[reply]
Is there a technical reason that we can't display editnotices on mobile? --Ahecht (TALK
PAGE) 19:22, 10 December 2019 (UTC)[reply]
@Ahecht: phab:T201595 has been open for over a year, with little progress. Suffusion of Yellow (talk) 19:28, 10 December 2019 (UTC)[reply]
Maybe pending changes protection for Talk:Main Page? – Muboshgu (talk) 16:48, 10 December 2019 (UTC)[reply]
@Muboshgu: Reverting the off-topic edits isn't all that difficult. It's the WP:BITE issue of being reverted, and sometimes warned, for doing exactly what you were told to do: Fill in a box labeled "What is on your mind?". Suffusion of Yellow (talk) 17:29, 10 December 2019 (UTC)[reply]
"What is on your mind?" is lovely and warm and fuzzy and inclusive and welcoming and "reaching out", and nothing whatsoever to do with building an encyclopaedia, so expect serious repercussions from WMF if we try to do anything about it. DuncanHill (talk) 20:01, 10 December 2019 (UTC)[reply]
The English Wikipedia community should not refuse to talk action because of speculation that the WMF might dislike it. Go ahead, modify (blank?) MediaWiki:Mobile-frontend-talk-add-overlay-content-placeholder to be more encyclopedic. * Pppery * _{it has begun...} 20:15, 10 December 2019 (UTC)[reply]
Any ideas? That message is, as the name implies, placeholder text, so it can't include a link to Wikipedia:Talk page guidelines, etc. ~~QQX doesn't reveal anywhere else we could inject text.~~ Suffusion of Yellow (talk) 20:24, 10 December 2019 (UTC)[reply]

Hmmm, wonder if something like "Comments should be relevant to {{BASEPAGENAME}}" would work there... Suffusion of Yellow (talk) 20:27, 10 December 2019 (UTC)[reply]

And there's MediaWiki:mobile-frontend-talk-add-overlay-submit, which might allow links. Suffusion of Yellow (talk) 20:32, 10 December 2019 (UTC)[reply]
MediaWiki:Mobile-frontend-editor-licensing-with-terms looks like the best message to (ab)use to stick wikitext in there to me. (Presumably, though I have no real means of testing, one could add conditionals that depend on the page name to only include content on certain pages -- would a {{editnotice load}} call work?) * Pppery * _{it has begun...} 20:44, 10 December 2019 (UTC)[reply]
Could try this on testwiki, but there's a shouty editnotice on every MediaWiki: page telling me not to do that. Suffusion of Yellow (talk) 20:49, 10 December 2019 (UTC)[reply]

See reply on my talk :) — xaosflux ^Talk 00:32, 11 December 2019 (UTC)[reply]
In general messages related to licensing are almost always the wrong thing to try to stick anything in. — xaosflux ^Talk 00:33, 11 December 2019 (UTC)[reply]

Why is this link red?

On User talk:146.95.196.199, the link to Wikipedia:Articles for deletion/Macau cuisine, a existing page, is red. --146.95.196.199 (talk) 00:32, 10 December 2019 (UTC)[reply]

Works for me looks blue when I look at it, could you hover over it, and copy the URL you are seeing to here for further research? — xaosflux ^Talk 00:35, 10 December 2019 (UTC)[reply]

Thank you! It's blue now. The entry was created half an hour ago but my talk page leads to "title=Wikipedia:Articles_for_deletion/Macau_cuisine&redlink=1" until recent no matter how many times I refreshed it. --146.95.196.199 (talk) 00:44, 10 December 2019 (UTC)[reply]

If this happens again try to WP:PURGE the page. If that doesn't work, try a WP:NULLEDIT. Suffusion of Yellow (talk) 00:47, 10 December 2019 (UTC)[reply]

Thank you! --146.95.196.199 (talk) 01:16, 10 December 2019 (UTC)[reply]

Why doesn't `bgcolor` work on this table?

Vaccination schedule#Worldwide (but works on others on the same page) Guarapiranga (talk) 01:54, 10 December 2019 (UTC)[reply]

Something to do with class mw-datatable? Remove that from the table declaration and all that is supposed to be blue is blue. I don't know what that class does or where it is defined ...

I hope you are asking because you intend to make that table accessible to those who have trouble distinguishing one color from another ...

—Trappist the monk (talk) 02:07, 10 December 2019 (UTC)[reply]

Aha, bingo! mw-datatable highlights rows as one hovers the cursor over them.

What do you recommend to make that table accessible to those who have trouble distinguishing one color from another, Trappist the monk? Guarapiranga (talk) 04:00, 10 December 2019 (UTC)[reply]

MOS:COLOR.

—Trappist the monk (talk) 09:09, 10 December 2019 (UTC)[reply]

Guarapiranga, because bgcolor is a deprecated HTML attribute that gets overriden by inline CSS styles and default styling of tables, like for instance mw-datatable applies. It doesn't work, because it's so old that anything else takes precedence over it. —TheDJ (talk • contribs) 08:44, 10 December 2019 (UTC)[reply]

It's not just deprecated but was marked as obsolete from HTML 5.0 onwards. Instead of e.g. bgcolor=LightSkyBlue you could use style="background-color:LightSkyBlue;", or even better, set up Template Styles. --Redrose64 🌹 (talk) 12:15, 10 December 2019 (UTC)[reply]

Security warning issues

Browsing via Win7/IE11

(Help Desk directed me here)

On Sunday I was machine-advised:

Wikipedia https://en.wikipedia.org/sec-warning

Your Browser's Connection Security is Outdated.

Wikipedia is making the site more secure. You are using an old web browser that will not be able to connect to Wikipedia in the future. Please update your device or contact your IT administrator.

I'm using Internet Explorer 11, the highest-numbered release.

I'm using Windows 7 Professional, for which Support from Microsoft ends Jan. 13, 2020,

... and with it IE 11 on Win7.

Aside from Windows 8 (NINE was the German "Nein" = No), Windows 10 will not have a Windows 11 successor - 10 will be followed by 12.

(reminds of "Why is nine afraid of seven - because "seven ate nine")

Even Microsoft Security Essentials will continue to support Win7.. for a while, so why is Wikipedia scared of non-malevolent users of Win7/IE11, and how safe is it for other wiki people to use, if they're on a highway with so many non-inoculatees?

Isn't it time, if Wiki has some clout, to publicize that those running Win7/IE11 are still welcome, and let Microsoft/Dell (from whom I've been buying) know that Apple will be happy to eat their lunch, if they close this part of the dining room. Pi314m (talk) 08:41, 10 December 2019 (UTC)[reply]

::*update* @ https://wikitech.wikimedia.org/wiki/HTTPS/Browser_Recommendations

(recommended by Wiki's Help Desk)

::: Edit History has 22 updates in 2017, none in 2018, one in 2019, and Win7/IE11 is still recommended. Pi314m (talk) 08:56, 10 December 2019 (UTC)[reply]

@Pi314m: Could it be that you have not installed all security updates for IE11 or use another software in connection with your browsing (which the message mentions in the bottom (Also it could be interference from corporate or personal "Web Security" software which actually downgrades connection security.))? I just tried accessing Wikipedia from the IE installed on my Windows 7 (version 11.0.9600.19540) and I did not get any such warning. Regards So Why 09:03, 10 December 2019 (UTC)[reply]
Pi314m, This is the TLSv1.0 and TLSv1.1 deprecation effort. Both protocols that are deprecated by all vendors coming january. So that means that your Windows 7 + IE11 combination doesn't support TLS.1.2 While Windows 7 does come with TLSv1.2 (starting from Service Pack 1) it is not enabled by default I think. I'm trying to get some clarification on that. The advise will probably be: Use MS Edges/Firefox/Chrome instead of IE11 or modify your Windows registry to enable TLS 1.2 —TheDJ (talk • contribs) 12:08, 10 December 2019 (UTC)[reply]
FWIW, I definitely didn't make any changes to my configuration but "Enable TLS 1.2" as mentioned by xaosflux below is activated in my settings. So Pi314m has likely either a non-updated version of Windows 7 or somehow deactivated that setting. Regards So Why 13:21, 10 December 2019 (UTC)[reply]
Just for fun I fired up a W7/IE11 and it worked just fine, in MSIE11 there is a checkbox to "Enable TLS 1.2" in Internet Options - > Advanced - > (Security section). — xaosflux ^Talk 12:31, 10 December 2019 (UTC)[reply]
Also I have no idea what you are editing with, but you have made a mess with all those <p> tags - hopefully you never introduce those to an article. — xaosflux ^Talk 12:32, 10 December 2019 (UTC)[reply]

Tracked in Phabricator
Task T238038

Since this isn't likely to be the last question we get that this here, should we have the page at https://en.wikipedia.org/sec-warning detect the user's OS and browser and then display the appropriate advice from https://wikitech.wikimedia.org/wiki/HTTPS/Browser_Recommendations instead of making them scroll past a huge list of languages (why are we even displaying all those languages on en.wikipedia.org?), click on a tiny link, and then scroll through a huge page? --Ahecht (TALK
PAGE) 14:45, 10 December 2019 (UTC)[reply]
@Ahecht: the best place to suggest improvements to sec-warning is probably going to be at: phab:T238038. — xaosflux ^Talk 14:53, 10 December 2019 (UTC)[reply]

Ahecht, "why are we even displaying all those languages on en.wikipedia.org" because technically there is only one page, shown on different domains. —TheDJ (talk • contribs) 16:27, 10 December 2019 (UTC)[reply]

Browsing via Win7/IE11 (MORE)

I've done lots of edits in the past hours since reporting the SEC-WARNING; moments ago was the SECOND time. I did not change any settings. Pi314m (talk) 23:29, 10 December 2019 (UTC)[reply]

New security warning: syntax error and API challenge

The new security warning page "Your Browser's Connection Security is Outdated" has two problems. First, when wrapped using the query API it doesn't parse as valid XML. The .NET method XmlDocument.Load, applied to the incoming stream, complains:

The 'img' start tag on line 21 position 38 does not match the end tag of 'a'. Line 22, position 3.

I see the offending HTML has an <img> tag posing as an attribute of an <a> tag. That does need to be fixed.

Second: as I'm using code based on HttpClient.GetStreamAsync() to read a page, I don't know how to fix the "please get more secure" message. Maybe I've not set it to use SSL properly, and I'll try to research what needs fixing, but has anyone else come across this as an API regression and can suggest a fix? David Brooks (talk) 18:49, 11 December 2019 (UTC)[reply]

@DavidBrooks: the sec-warning page isn't managed here on the English Wikipedia, please open a bug report on phabricator for any corrections needed there. Check your TLS version settings, you should have version 1.2 enabled. — xaosflux ^Talk 19:05, 11 December 2019 (UTC)[reply]

In case anyone else has the same problem: I'm still using fx 4.5. Apparently upgrading to 4.8 would help (and maybe have to set the SslProtocols property of the HttpClientHandler), but I have a number of legacy apps. If staying with 4.5, a fix is to add this before constructing the HttpClientHandler:

System.Net.ServicePointManager.SecurityProtocol = SecurityProtocolType.Tls12;

Unfortunately this won't automatically upgrade to another protocol should it become required; the SystemDefault setting is not available in 4.5. David Brooks (talk) 21:04, 11 December 2019 (UTC)[reply]

I'll add: I had completely lost sight of the fact that the API is returning HTML (valid HTML at that) when it is expecting well-formed XML in the usual API syntax. It just happened to contain an unclosed <IMG> element that is OK in HTML5. The phabricator bug opines that it should at least return a normal XML return with an <warnings> entry. David Brooks (talk) 05:26, 12 December 2019 (UTC)[reply]

Access to Wikipedia from my home computer was cut off 22 days before the announced deadline

I'll have to buy a new computer or device in order to meet the stupid new encryption protocol requirements, but I assumed in good faith that I would have until Jan 1st. to do so, since that what it says on the page https://en.wikipedia.org/sec-warning . Instead, access was cut off on December 9th, 22 days earlier than the announced deadline, meaning I can't access Wikipedia or any Wikimedia.org sites other than the https://en.wikipedia.org/sec-warning page itself from my home computer. I'm currently on a library computer, but I really can't spend much time at the library, and I doubt that I'll practically be able to buy a new device much before Jan 1st anyway, due to various holiday limitations and obligations, so this moving of the encryption-protocol guillotine to 22 days before its previously-announced date is highly inconvenient and annoying to me... AnonMoos (talk) 00:45, 12 December 2019 (UTC)[reply]

AnonMoos, You probably don't need to buy a new device but just need to update your browser: see wikitech:HTTPS/Browser_Recommendations for advice on that. Galobtter (pingó mió) 00:56, 12 December 2019 (UTC)[reply]

Nope, I most definitely need to buy a new device -- Mozilla 3.6.28 is about the best my existing computer can do. AnonMoos (talk) 01:08, 12 December 2019 (UTC)[reply]

Sounds like you are a prisoner of Windows! Have you ever considered installing Linux on the old computer, the common way to breathe new life into old hardware. Elizium23 (talk) 04:23, 12 December 2019 (UTC)[reply]

Or given Firefox early version history, possibly Mac OS X Tiger (EOL 2012) or PowerPC Macs (similar date I imagine for hardware support). (How do computers get that old and not die on their own?)

Linux is probably the next best step.

As for the date, it was likely some percentage of computers and not all; the last time this was done (for Windows XP + IE7), I think the final date for all users was published. Users excluded were gradually added in waves prior to that date (I think it was 10/25/50/100% of readers by week or something). --Izno (talk) 04:57, 12 December 2019 (UTC)[reply]

I would strongly tend to doubt whether Linux is any kind of magic silver bullet for my situation. I have a Mandrake 9.1 DVD, which I know works on my computer, but do the double-digit versions of FireFox necessary to get around the encryption protocol nonsense run on Mandrake 9.1? Would a significantly more recent version of the operating system work on my hardware? AnonMoos (talk) 23:25, 13 December 2019 (UTC)[reply]

Also, I get the same result whether I'm logged-in or not. AnonMoos (talk) 23:25, 13 December 2019 (UTC)[reply]

How I read it was that the sec-warning page will forcibly be shown until January 1, after which you won't be able to connect to Wikipedia at all (including the sec-warning page). — MusikAnimal ^talk 04:59, 12 December 2019 (UTC)[reply]

Ah yes, that's how it is. --Izno (talk) 05:06, 12 December 2019 (UTC)[reply]

MusikAnimal, easy to misinterpret though... —TheDJ (talk • contribs) 09:17, 12 December 2019 (UTC)[reply]

sec-warning redux

Not 67.81.173.178 but I can spill the details out. I have a first generation iPad from Christmas 2011 and still works okay.

The sad truth is after I looked at an intermediate revision of a Nickelodeon themed article named Sam and Cat on Monday I went back to the contribs page of MPFitz1968 instead it redirected back to a security warning aka Browser Connection Security Issues.

It was never supposed to happen at all as what the article about said message on Wikitech because it interferes with older browsers and models. That must be fixed.

Ceasing transmissions,

47.16.146.238 (talk) 20:25, 12 December 2019 (UTC)[reply]

Could slowing down be due to TLS 1.2?

Now that people are disabling TLS 1.0 and 1.1, leaving 1.2, is it possible that the CPU requirement for TLS 1.2 is sufficiently higher to be the cause of the reported slowdowns? Here's a recent example:

If you report this error to the Wikimedia System Administrators, please include the details below.
- Request from 64.24.9.9 via cp1089 frontend, Varnish XID 366290093
- Error: 503, Backend fetch failed at Fri, 13 Dec 2019 05:23:26 GMT

Pi314m (talk) 06:15, 13 December 2019 (UTC)[reply]

That error is a 503 error, meaning there was an issue with the Wikimedia servers and not with the user's computer. Galobtter (pingó mió) 09:11, 13 December 2019 (UTC)[reply]

Pi314m, if you have TLS1.2 enabled, it is preferred over older stuff. Everyone capable of TLSv1.2 was already on it, those not capable are still connecting in the exact same way (unless they themselves changed something), they are just not reaching Wikimedia webpages (and after january 1 they aren't reaching wikimedia at all any longer). Technically: for a small percentage of users the web routing was changed, but the connection method has not changed at all until january 1. —TheDJ (talk • contribs) 09:38, 13 December 2019 (UTC)[reply]

Any technical reasons why we can't thank robots?

Tracked in Phabricator
Task T205224

Such blatant mechanophobia! Guarapiranga (talk) 12:54, 10 December 2019 (UTC)[reply]

Until bots have their own thought process and intelligence + emotions, it will be helpful if you thank the bot operator who will be encouraged by your gesture. --DBig Xrayᗙ 13:09, 10 December 2019 (UTC)[reply]

That is my intention, DBigXray, but my advances in telling the bots to send my thanks to their operators have been coldly thwarted by the interface (API error, or sumthin'). Guarapiranga (talk) 13:26, 10 December 2019 (UTC)[reply]

Don't anthropomorphize robots. They hate it when you do that. -- RoySmith (talk) 13:34, 10 December 2019 (UTC)[reply]

The robots sincerely thank you and hope more editors will flock to give them praise. Martinevans123 (talk) 13:36, 10 December 2019 (UTC)[reply]

@Guarapiranga: Since this is VPT I'll give a tech answer: it is because we haven't enabled it in the extension configuration ($wgThanksSendToBots). There is a task request open you can participate in if you would like to follow up: phab:T205224. — xaosflux ^Talk 14:04, 10 December 2019 (UTC)[reply]

Chimed in there, Xaosflux, thanks (yes, I was being serious, despite the humour). Guarapiranga (talk) 14:27, 10 December 2019 (UTC)[reply]

We also can't thank anonymous editors. Glades12 (talk) 16:20, 10 December 2019 (UTC)[reply]

@Glades12: We can quite safely assume a logged-in editor is always the same person (except when someone uses a shared accout, against our policies), but we can't assume that in case of IP-editors. Sometimes multiple anonymous users use the same proxy, so the IP address does not identify an editor, and then any thanks may be misdirected. --CiaPan (talk) 16:25, 10 December 2019 (UTC)[reply]

Glades12, because you don't know who they are the next time they use the website. —TheDJ (talk • contribs) 16:26, 10 December 2019 (UTC)[reply]

I already understood that; should probably have clarified. Glades12 (talk) 16:29, 10 December 2019 (UTC)[reply]

Nothing gets sent if I type in my registered email address to reset my password

Hi, this is User:My name is not dave here. Despite still having Wikipedia related emails in my email folder for my registered email address, I still get nothing when I try to send a password reset email. I recently had a few problems with Google storing and retrieving my passwords, so I reset the whole thing. Now I'm locked out of my account because the damn password reset thing doesn't send an email to my registered email address. Any solutions to this, or do I need to grit my teeth in a sense? 143.167.200.66 (talk) 14:22, 10 December 2019 (UTC)[reply]

Hi My name, a few things you can try. First check if your email provider has any sort of spam/junk filtering that could be obscuring the email from you. You could also try having the email sent in a slightly different way by requesting the password reset from another project, such as meta:Special:PasswordReset. — xaosflux ^Talk 14:27, 10 December 2019 (UTC)[reply]

fatal error

When previewing Cut (Hunters and Collectors album) § References I get this oh-so-helpful error message:

[Xe-HRgpAADoAAG87AKgAAADJ] 2019-12-10 16:26:46: Fatal exception of type "TypeError"

If I remove {{reflist}} then I can preview the section. Previewing the whole page does not cause the error.

and it isn't even Thursday ...

—Trappist the monk (talk) 16:30, 10 December 2019 (UTC)[reply]

Probably related to T240248 Reedy (talk) 16:40, 10 December 2019 (UTC)[reply]

Simpler example with the same error:

{{Reflist|refs=<ref name="Name">Reference</ref>}}

. PrimeHunter (talk) 16:48, 10 December 2019 (UTC)[reply]

Should be fixed now... Reedy (talk) 17:27, 10 December 2019 (UTC)[reply]

Infobox image display appears to differ between Desktop and Mobile view

I don't have this problem fully diagnosed, and I am probably off on the wrong track, but can someone please take a look at the discussion at Template talk:Infobox election#Mobile display issue? The problem is that there are three images of candidates displayed in {{Infobox election}}, and things look fine in the Desktop view, but in the Mobile view, one image is much larger than the other two. – Jonesey95 (talk) 17:21, 10 December 2019 (UTC)[reply]

Query on Template:div col in meta:

Hello,

I am in the process of putting together an overview of instructions for how to add new water fountains and photos of fountains to the open data databases that will power our application. I followed the div col template instructions from wikipedia for our page, but there seems to be some differences between the template usage for the same tags on meta.wikimedia.org

When I the following tags to give some headers for the two columns, neither the colwidth nor the text align seem to be taken into account.

{{div col|colwidth=30em|style="text-align:center"}} Instructions Illustration {{div col end}}

I am new to all this. Thank you for your help. I am choosing the col template over the wikitable format because the floating aspect of the second column is more appropriate for mobile phone visualization.

here is the link to our page : meta:Wikimedia_CH/Project/European_Water_Project

BR, Stuart S2rapoport (talk) 17:29, 10 December 2019 (UTC)[reply]

I see two columns, one headed "*Instructions" and one headed "*Illustrations", so it looks like it is working. {{Div col}} on meta does not accept |style=, so that part is being ignored. That said, the columns do not look pretty, and your specified width is not working, because the images are large. You may be better off using a wikitable, as in the subsequent section. – Jonesey95 (talk) 17:45, 10 December 2019 (UTC)[reply]

Wikitables are not flexible and look terrible on the mobile phone. How could I reproduce the div col template with html tags to be able to have full flexibility? What are all the parameters for Div col on meta? Except for the headers the current behaviour on th first table is best for mobiles ans is what I want for the instructions.

Is there a template that allows one to have text plus a toggle text which reveals an image?

Thanks for your help! Stuart — Preceding unsigned comment added by S2rapoport (talk • contribs) 22:44, 10 December 2019 (UTC)[reply]

Meta's template is, not surprisingly, meta:Template:Div col. At the time that it was created in 2009, it was a straight copy of ours; but things have moved on since then. It accepts the following parameters:

|colwidth=
|cols= - defaults to 2 (the first positional parameter is an alias for this)
|gap= - defaults to 1em
|small=

The |cols= parameter is no longer provided in our template, but the other three have basically the same function as with ours. --Redrose64 🌹 (talk) 00:20, 11 December 2019 (UTC)[reply]

@Redrose64 @Jonesey95 thank you again for your advice.

Would it be possible to point me to a wikipedia or meta page where someone has a style section on their page and html. What I have now with the meta:Template:Div col is good enough for our needs, so it will do for now.

Is it possible to use wikipedia templates on meta pages and get the more modern template definitions?

Thx, Stuart S2rapoport (talk) 09:23, 11 December 2019 (UTC)[reply]

It is possible to copy the code of a template from en.WP to other Wikimedia instances, but doing so can break existing transclusions of the target template, and if the template has dependencies that are not present or work differently on the target site, the template may not work correctly. You might want to ask on the VPT equivalent at meta for an experienced template editor to enhance Template:div col on meta to work more like the one here. – Jonesey95 (talk) 15:25, 11 December 2019 (UTC)[reply]

User:MusikAnimal is an admin on meta, he might know who to ask. EdJohnston (talk) 18:30, 11 December 2019 (UTC)[reply]

meta:Babel, I suppose? meta:Template:Div col has some 500+ transclusions, mind you. I would be a little cautious about breaking changes. Maybe you could use a bot or script to fix existing transclusions, or just create a new template altogether. — MusikAnimal ^talk 19:11, 11 December 2019 (UTC)[reply]

I very much appreciate your advice. The Div col template will sufficient for the next couple of months. We need to focus on some other more pressing technological challenges right now. How can I come back to this message in early February and reach out to you for specific advice when I have time to properly experiment ? I want to make this open data fountain and photo addition instruction page as clear and user friendly as possible - as many of our application users may be non-technical.

If you have any suggestions of how I can make the instructions clearer and friendlier - especially when using a mobile phone, I very welcome constructive input. meta:Wikimedia_CH/Project/European_Water_Project

S2rapoport (talk) 16:41, 12 December 2019 (UTC) Stuart[reply]

Did you mean ...? Showing results for ...

When I make a typo when doing a search, the suggested search term or the results I am shown is often for something entirely different from what I expect. When I try to retype, I now notice I often get what I WAS expecting immediately below.

It seems whatever comes up should be closer to what we intended.— Vchimpanzee • talk • contributions • 19:09, 10 December 2019 (UTC)[reply]

Gadget and userscript unit testing

FYI

– Pointer to relevant discussion elsewhere.

I've created a script to make on-wiki unit testing a bit easier to implement. Please see WP:IANB § Gadget and userscript unit testing, and comment there (not here). - Evad37 [talk] 01:09, 11 December 2019 (UTC)[reply]

Edit count aggregations?

Do we have edit count aggregations available in the replica database? Or anyplace else I can query from within toolforge? Right now I'm doing:

           SELECT count(*)                                                                                                                                                                                                                                                
           FROM revision_userindex                                                                                                                                                                                                                                        
           JOIN actor ON rev_actor = actor_id                                                                                                                                                                                                                             
           WHERE actor_name = %(username)s

which works, but is grossly inefficient. — Preceding unsigned comment added by RoySmith (talk • contribs) 04:20, 11 December 2019 (UTC)[reply]

@RoySmith: If you're okay with some approximation, you should use the user_editcount column on the user table. Otherwise try using actor_revision instead of just actor, which runs a bunch of subqueries behind the scenes. — MusikAnimal ^talk 04:38, 11 December 2019 (UTC)[reply]

Thanks for that. I knew about the alternate views, but the many flavors of actor is a new twist. This schema never ceases to surprise me. As for user_editcount, it's not clear exactly how up to date that is, but I'll go with it for now. For my purposes, the smaller end of the spectrum is more interesting. Maybe I'll do something like use user_editcount as long as it's >= 1000 and do a revision_userindex scan otherwise. I'll also need deleted edit counts, which I get from archive_index. I suspect those will all be small enough that a full count won't be too expensive. -- RoySmith (talk) 17:22, 11 December 2019 (UTC)[reply]

Does Wp-mirror work?

Does mw:Wp-mirror work?

Its latest release was in 2014. Its documentation hasn't been updated since 2014. Its mailing lists appear to have fizzed out in 2015.

It's mentioned in instructional documentation in several places, such as at Wikipedia:Database download#Mirror building, which is a disservice if it doesn't actually work anymore.

I look forward to your replies.

Sincerely, — The Transhumanist 12:33, 11 December 2019 (UTC)[reply]

MoreMenu gadget is getting an upgrade

I wanted to give a heads up so there's less surprise. Over the past several months, I've rewritten meta:MoreMenu (aka "Dropdown menus") from scratch and have slowly been migrating wikis to using the new version. English Wikipedia is next! There are some 16,000 users, though I suspect only a fraction are still active. The new version is fully localizable, extendable, and it now works in the Timeless skin. If you use the Vector skin, you'll notice minimal changes. If you use Monobook or Modern, you'll see a slightly bigger font size (which matches the default font size for the skin), but everything will largely function the same. You can try the new version now by installing it globally (the preferred method), just make sure you first uninstall the gadget here. If there are any concerns, please let me know. I plan to make this upgrade in the coming days. Best, — MusikAnimal ^talk 21:27, 11 December 2019 (UTC)[reply]

Template (or module) returning number of search results

Is there a template (or module) that returns the number of search results (with or without a link to the results page)? Cheers. Guarapiranga (talk) 05:49, 12 December 2019 (UTC)[reply]

Nope. And it's impossible to make such a template or module as they can't access the API. SD0001 (talk) 12:11, 12 December 2019 (UTC)[reply]

Visited watchlist pages return to bold

Tracked in Phabricator
Task T218511

The visited pages on my watchlist keep returning to bold even though there are no further edits since my visit. What gives? WWGB (talk) 13:57, 12 December 2019 (UTC)[reply]

This is a bug that has been coming and going for at least nine months now (see link to the right). It has gotten worse in the past few days. – Jonesey95 (talk) 15:35, 12 December 2019 (UTC)[reply]

Noticed it too, WWGB. Guarapiranga (talk) 23:34, 12 December 2019 (UTC)[reply]

Me, too. Really annoying. – Tea2min (talk) 07:51, 13 December 2019 (UTC)[reply]

Range block not working?

Whoops!

– Working at intended. SQL ^{Query me!} 21:32, 12 December 2019 (UTC)[reply]

Not sure where better to ask this. Figure someone here will know what's going on. See the log here. This /17 range was blocked in September for six months (compare), but this IP is editing as of today. Not sure how to brain. Is anyone aware of this happening on any other project? GMG ^talk 20:28, 12 December 2019 (UTC)[reply]

GreenMeansGo, Odd! It should cover that, see [7], 86.179.128.0/17 should encompass 86.179.128.0 - 86.179.255.255. What's even odder is that .116 and .118 both show the rangeblock. I'd raise a ticket on phab. I tried it on the testwiki, and it seems to work: [8] SQL ^{Query me!} 20:38, 12 December 2019 (UTC)[reply]

@SQL: Well I started T240615. Not sure if I did it correctly though. Actually, looking at q:Special:Contributions/86.179.128.0/17, at least two IPs have edited through the range block since September. GMG ^talk 20:48, 12 December 2019 (UTC)[reply]

Oooooooooh. I'm an idiot. They're editing their own talk pages. Derp. GMG ^talk 20:52, 12 December 2019 (UTC)[reply]

Making first row of table a heading

I'm embarrassed to ask this because I've done this dozens of time, but I forgot how to make the first row of a table a heading row using VE.

Desired location: Brittney Ezell (the stats table)

Example of correct one: Shalonda_Enis#Alabama_statistics

I tried looking in help:table, and VE User guide but I did not see it.

It's something simple, but I don't recall it.S Philbrick (Talk) 21:37, 12 December 2019 (UTC)[reply]

@Sphilbrick: the control has moved around a few times, in VE highlight the entire row of the table, then all the way at the top of the screen (just under the "article", "talk", etc tabs) there is a pull down to change from "content cell" to "header cell". Not a great UX layout for that control IMHO. — xaosflux ^Talk 00:27, 13 December 2019 (UTC)[reply]

Guarapiranga (talk) 00:34, 13 December 2019 (UTC)[reply]

Perhaps an option could be added here:

Guarapiranga (talk) 00:39, 13 December 2019 (UTC)[reply]

@Guarapiranga: yes, I think the row/column control should have it - I think at once point it had a little "gear" icon there where you could make "header" or "sort" options, but it is now at the top. — xaosflux ^Talk 00:50, 13 December 2019 (UTC)[reply]

Xaosflux, Thanks, maybe you have explained why I couldn't do it. I know I've done it before, but I did a lot in 2015, and perhaps it was changed. S Philbrick (Talk) 01:07, 13 December 2019 (UTC)[reply]

Mobile interface bug

Tracked in Phabricator
Task T240651

I realize WP:ITSTHURSDAY, so probing if this is a problem for others since I use the advanced mode (whatever that means) on mobile. The buttons which would edit or unwatch a page now have the first letter of the link text overlaid so you'll have an "E..." over the pencil or "U..." over the star. Wug·a·po·des 02:45, 13 December 2019 (UTC)[reply]

Userbox in table not moving over to next column / heading

Not sure if this or the help desk is the better place to ask this since it's not really a "bug", but over here on my sandbox page you can see that the UPS and Monopoly userboxes are in the "About me" section, when I would expect they would move over to "Random" because I have used the || formatting in the page source, which should move those two userboxes over to the next column in the table. You can also see the || formatting next to the rightmost userbox, which must mean I must have messed something up. I highly doubt this will help, but for reference, I'm on Chrome on a Windows 10 v1903 machine. Anyone know what I could be doing wrong? --PlanetJuice (talk) 03:32, 13 December 2019 (UTC)[reply]

Questions that are clearly "how to" questions are for WP:Help desk or WP:Teahouse as you suspected. That's something that's widely misunderstood (it's currently already misunderstood on this page), and I'm feeling too lazy to move this discussion. Just good info to know for future reference. ―Mandruss ☎ 03:39, 13 December 2019 (UTC)[reply]

@PlanetJuice: the problem appears to be with Template:User en-us-N, note if you use Template:User en-N there instead it fixes your issue. I don't have time to troubleshoot that template right now, but feel free to give it a look! — xaosflux ^Talk 04:26, 13 December 2019 (UTC)[reply]

This edit should fix it. But I don't see why we need all those variants, can't we all agree that a tomato should be called a tomato instead? --Redrose64 🌹 (talk) 08:46, 13 December 2019 (UTC)[reply]

Toe-mah-toe or toe-may-toe? Quick, there's only one right answer. --Izno (talk) 18:46, 13 December 2019 (UTC)[reply]

Toe-may-toe for me :P Well, thanks for all the help - I'll make sure to look at the userbox source for any future issues!

@@ Line 476: / Line 476: @@
 :::::I would strongly tend to doubt whether Linux is any kind of magic silver bullet for my situation.  I have a Mandrake 9.1 DVD, which I know works on my computer, but do the double-digit versions of FireFox necessary to get around the encryption protocol nonsense run on Mandrake 9.1?  Would a significantly more recent version of the operating system work on my hardware? [[User:AnonMoos|AnonMoos]] ([[User talk:AnonMoos|talk]]) 23:25, 13 December 2019 (UTC)
+:::::Also, I get the same result whether I'm logged-in or not. [[User:AnonMoos|AnonMoos]] ([[User talk:AnonMoos|talk]]) 23:25, 13 December 2019 (UTC)
 :How I read it was that the sec-warning page will forcibly be shown until January 1, after which you won't be able to connect to Wikipedia at all (including the sec-warning page). <span style="font-family:sans-serif">&mdash; <span style="font-weight:bold">[[User:MusikAnimal|<span style="color:black; font-style:italic">MusikAnimal</span>]] <sup>[[User talk:MusikAnimal|<span style="color:green">talk</span>]]</sup></span></span> 04:59, 12 December 2019 (UTC)