Jump to content

Wikipedia:Village pump (technical): Difference between revisions

From Wikipedia, the free encyclopedia
Content deleted Content added
Line 632: Line 632:
::Thanks for the watchlist notice! I thought it was just me for a moment. [[User:Alex Shih|Alex Shih]] ([[User talk:Alex Shih|talk]]) 05:28, 20 February 2018 (UTC)
::Thanks for the watchlist notice! I thought it was just me for a moment. [[User:Alex Shih|Alex Shih]] ([[User talk:Alex Shih|talk]]) 05:28, 20 February 2018 (UTC)
:::{{rto|Xaosflux}} I just thanked you on [//en.wikipedia.org/wiki/Special:Diff/826624253 your watchlist notice edit] and I notice the action is logged both in my thanks log and your received thanks log. Did you receive notification for this or is it just not going despite the logs showing so?–[[User:Ammarpad|Ammarpad]] ([[User talk:Ammarpad|talk]]) 06:17, 20 February 2018 (UTC)
:::{{rto|Xaosflux}} I just thanked you on [//en.wikipedia.org/wiki/Special:Diff/826624253 your watchlist notice edit] and I notice the action is logged both in my thanks log and your received thanks log. Did you receive notification for this or is it just not going despite the logs showing so?–[[User:Ammarpad|Ammarpad]] ([[User talk:Ammarpad|talk]]) 06:17, 20 February 2018 (UTC)
::::The problem appears to be that the thanks function is thanking the last unthanked edit in the page, regardless of which edit is chosen, so in many cases it is thanking the wrong edit/user. If the edit you thanked happened to be the last unthanked revision you were probably "lucky" and it will probably have worked "ok", by chance. The logs seem to be correct, but the message you receive at the time of thanking may not be. Don't rely on it at all until it's fixed would be my advice, though. --<span style="font-family:Arial;font-weight:bold;color:#004d80;"> [[User talk:Begoon|Begoon]]</span> 06:36, 20 February 2018 (UTC)
::::The problem appears to be that the thanks function is thanking the last unthanked edit in the page, regardless of which edit is chosen, so in many cases it is thanking the wrong edit/user. If the edit you thanked happened to be the last unthanked revision you were probably "lucky" and it will probably have worked "ok", by chance. The logs seem to be correct, but the feedback you receive at the time of thanking may not be. Don't rely on it at all until it's fixed would be my advice, though. --<span style="font-family:Arial;font-weight:bold;color:#004d80;"> [[User talk:Begoon|Begoon]]</span> 06:36, 20 February 2018 (UTC)

Revision as of 06:46, 20 February 2018

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

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


Archive-date parameter

Is this Wikipedia?

Wikimedia received an email from someone asking a question about a page. The URL they provided started with:

en.bywiki.com

It looks a lot like Wikipedia but is it?S Philbrick(Talk) 17:27, 7 February 2018 (UTC)[reply]

Looks like a Wikipedia mirror to me. {{3x|p}}ery (talk) 17:28, 7 February 2018 (UTC)[reply]
Domain names run from right to left in descending hierarchical order. So en.bywiki.com is a different site than en.wikipedia.org on two levels; the top level domain and the second level domain. See https://whois.icann.org/en/lookup?name=en.bywiki.com for more info. They've registered it to one of those privacy fronts, which is not a good sign. It's distinctly possible for them to have changed data between grabbing it off WP and displaying it on their site. ᛗᛁᛟᛚᚾᛁᚱPants Tell me all about it. 17:35, 7 February 2018 (UTC)[reply]
It's a live mirror which adds advertisements to Wikipedia pages. It's not controlled by the Wikimedia Foundation and it violates several rules but the advertising part is allowed. I haven't seen the email but it's possible it applies equally to the corresponding real Wikipedia page. PrimeHunter (talk) 18:15, 7 February 2018 (UTC)[reply]
Which rules does it violate? I find Wikipedia mirrors all the time while looking for sources on lesser known subjects, and most of these follow the creative commons attribution rules of Wikipedia + advertisements, which is allowed under the standard CC license used by Wikipedia. --Donald Trung (Talk) (Articles) Respect mobile users. 11:15, 12 February 2018 (UTC)[reply]
Many violate trademark law, by reusing the name and/or logos of Wikipedia. And when their attribution is proper, that is often more because they did a full copy and got our license pages and statements copied as well, rather than they actually intent to support the copyright :) —TheDJ (talkcontribs) 11:19, 12 February 2018 (UTC)[reply]

Font/margin problem

At this version of Wikipedia:WikiProject Doctor Who/Article alerts there is a box at the top containing centred text. Using Opera 36, I see all the text squeezed sideways (and stretched vertically) so that it's one word per line (even words like "a" are the sole occupants of their lines). This is (correctly) centred across the box width. Do other people experience the one-word-per-line problem?

I made these amendments with no visible effect. My alteration of the declaration font-family: trebuchet ms, sans-serif to font-family: 'Trebuchet MS', sans-serif was in accordance with CSS Fonts Module Level 3 and also Cascading Style Sheets Level 2 Revision 1. But I have narrowed it down: whichever version of Wikipedia:WikiProject Doctor Who/Article alerts that I start from, if I remove the first font family name so that the declaration is changed from either font-family: trebuchet ms, sans-serif or font-family: 'Trebuchet MS', sans-serif to font-family: sans-serif, the problem disappears. I'm pretty sure that I have Trebuchet MS installed, but the browser's "inspect element" feature shows that Arial is being used, that being the default for sans-serif. --Redrose64 🌹 (talk) 11:07, 8 February 2018 (UTC)[reply]

Redrose64, I added the underscored font-family: sans-serif above, which seemed to be missing. Is that what you intended? --Pipetricker (talk) 12:13, 8 February 2018 (UTC)[reply]
Yes, indeed, thanks. --Redrose64 🌹 (talk) 12:15, 8 February 2018 (UTC)[reply]
Have you solved this yet? --Pipetricker (talk) 20:08, 14 February 2018 (UTC)[reply]
It's behaving now, but I can't find anything that might have changed. Maybe a site CSS glitch - I don't know where to check those files. --Redrose64 🌹 (talk) 22:03, 14 February 2018 (UTC)[reply]

Section protection

Has the ability to protect individual sections ever been discussed or investigated? One uncontroversial use would be to protect AfD tags from removal, while letting the rest of the page be edited. Technically feasible? —swpbT go beyond 20:55, 8 February 2018 (UTC)[reply]

Not technically feasible; sections are arbitrary formatting things, not something that the software could separate out. Jo-Jo Eumerus (talk, contributions) 21:00, 8 February 2018 (UTC)[reply]
@Jo-Jo Eumerus: That is incorrect. See mw:Extension:ProtectSection as one technically feasible option. Nihlus 21:28, 8 February 2018 (UTC)[reply]
phab:T6375 requested it for Wikisource in 2006. It had large support at OldWikisource:Wikisource:Vote on enabling the ProtectSection extension but was declined. PrimeHunter (talk) 21:41, 8 February 2018 (UTC)[reply]
That is basically using the editfilter, but monitoring every single edit for compliance would be technically expensive. — xaosflux Talk 05:31, 9 February 2018 (UTC)[reply]
Almost everything is theoretically possible, but it would be very much against the design of wikicode and MediaWiki and is therefore extremely unlikely to ever be implemented (properly). —TheDJ (talkcontribs) 09:26, 9 February 2018 (UTC)[reply]
To editor TheDJ: For curiosity, what makes mw:Extension:ProtectSection not a proper implementation? —swpbT go beyond 16:33, 12 February 2018 (UTC)[reply]
Not TheDJ - but the big notes that it hasn't been maintained in 9 years is a red-flag for sure. — xaosflux Talk 16:47, 12 February 2018 (UTC)[reply]
@Swpb: Well almost everything in the infrastructure of MediaWiki makes assumptions about revisions of pages (this is also why section editing in VE, and watching sections are so hard to implement, as well as why Flow topics act like individual pages). A section is simply not a 'logistical unit' that can be used for anything in MediaWiki. There are three strategies to deal with this 1: Make sections pages and then collect/transclude them in a secondary page. This is what LQT did and then a page is suddenly a list of transclusions and an edit to a section would not show up in the main page when you edit the 'section' (by default). Flow does something similar and the magic required to make things like this work is.. challenging, lets put it that way, also we would jumpt from 40 million pages to some 500 million pages (just for en.wp). We can't scale to that level easily. 2: Properly retrofitting a concept of sections onto the entire infrastructure (that's a rather huge task, and would significantly change how people work with the wiki), 3: bolt on a fake concept of sections on top of every contact point where we deal with pages (which is basically what ProtectSection used to do), but that is very inefficient (you need to parse the entire page, wherever you deal with a page, while parsing a page is the most expensive operation possible within MediaWiki). —TheDJ (talkcontribs) 17:45, 12 February 2018 (UTC)[reply]

Couldn't edit article (Lyme disease) – possibly due to voting banner

I have been unable to make a null edit or even "show preview" on this page for a couple of days. Regards CV9933 (talk) 12:51, 9 February 2018 (UTC)[reply]

What happens when you try? It's only semi-protected and apart from being slow, I had no problems making a preview and a null edit with my autoconfirmred non-admin alternative account. PrimeHunter (talk) 13:57, 9 February 2018 (UTC)[reply]
If I click on edit, then preview and wait around a minute, the (Firefox) browser displays
The connection to en.wikipedia.org was interrupted while the page was loading.
  • The page you are trying to view cannot be shown because the authenticity of the received data could not be verified.
  • Please contact the web site owners to inform them of this problem.
Regards CV9933 (talk) 14:26, 9 February 2018 (UTC)[reply]
I logged in using Chrome and don't see an issue so it seems to be browser related, Cheers. CV9933 (talk) 16:39, 9 February 2018 (UTC)[reply]
I did a firefox update, as well as all the usual stuff - cleared the cache and cookies and rebooted my router with the same result.
Interestingly, I noticed a wikipedia voting banner at the top of the page and closed it. I then tried to edit the page and now all is fine. So it seems that with the banner, the time it was taking to refresh the page was too long for firefox and the error occurred. Cheers CV9933 (talk) 17:24, 9 February 2018 (UTC)[reply]
I wonder if its to do with the new code in that particular banner. This one makes an API call to see if you are an eligible voter. Seddon (WMF) (talk) 23:12, 18 February 2018 (UTC)[reply]

Template pass-through parameters

I am working on a template which generates a Template:Cite book. I would like for all unknown parameters to be passed on to {{Cite book}}, eg. {{Custom cite template|template-specific-parameter=true|author-mask=0|page=45}}, where author-mask and page aren't specified in advance. How can I do this? Daask (talk) 16:58, 9 February 2018 (UTC)[reply]

Module:Citation/CS1/Wrapper.
Trappist the monk (talk) 17:29, 9 February 2018 (UTC)[reply]
Which should really have a more generic name. {{3x|p}}ery (talk) 16:01, 10 February 2018 (UTC)[reply]
Unsupported criticism isn't very helpful. Why is a more generic name better than the existing name? What would be a better name?
Trappist the monk (talk) 16:12, 10 February 2018 (UTC)[reply]

Accidentally Changing the coin-notice Template

If I put the {{coin-notice}} template on an editor's talk page, and forget to subst it, and then add text below the template notice, it adds the added text to the template itself, which is certainly not what anyone would intend. I realize now that I should have put subst in front of the template, but if the template can be broken by a simple error, this seems to be a matter of excessive fragility that amounts to a bug. Robert McClenon (talk) 03:54, 10 February 2018 (UTC)[reply]

Right, this is because you are transcluding the header with the template alongside the "edit" button. Since it is being transcluded, the edit button is linked to the actual template and not the page itself. I'm not sure a solution is necessary when that is the intended function of transclusions and you are using the template incorrectly. Nihlus 04:04, 10 February 2018 (UTC)[reply]
Yes, a solution is necessary, because the result of an accidentally incorrect use of the template is dangerous and undesirable. Yes, a solution is necessary. Just saying that a bug is a feature doesn't make a bug a feature. Yes, a solution is necessary. Just saying that the fault is that of the humans and not of tool design doesn't excuse tools that can do very undesirable things through ordinary errors. Robert McClenon (talk) 04:32, 10 February 2018 (UTC)[reply]
It's not a bug as it is working as intended. When you transclude the header, you transclude the links that come with it (because that's what it is meant to do). You could remove the header from the transclusion, but the other {{User noticeboard notices}} have worked up until now with their headers. Also, I wouldn't call it dangerous as any misused template has the potential to cause "undesirable" issues; that's why there are instructions on how to use it on each page. Restructuring every subst-required template to be non-subst safe doesn't seem productive (outside of recursive substitution). Plus, the issues are easily fixed. Nihlus 04:47, 10 February 2018 (UTC)[reply]

@Nihlus: and @Robert McClenon: I put in a request at RPP to have this template-protected; that should solve the issue. There's been no major changes for some time, and any future ones could be easily done by a template editor. Home Lander (talk) 18:46, 10 February 2018 (UTC)[reply]

User:Home Lander - Thank you. In that case, that implies that it was working as designed but wasn't really designed to meet the requirements, because it could be accidentally edited. I worked for 45 years in information technology, and just because something is working as designed doesn't mean that there isn't a bug; it means that there isn't a code bug, but there may be a design bug. This was a design bug. Thank you. Robert McClenon (talk) 21:49, 10 February 2018 (UTC)[reply]
I'm still baffled at your insistence that this is a design bug or any bug for that matter. And I'm not sure what you having spent any time anywhere has to do with anything when you were the one using it incorrectly despite clear instructions advising you on how to use it. I have no intentions of commenting further, as it is clear you are unwilling to take any responsibility and only want to put the blame on other people (and on something so ridiculously trivial at that). Nihlus 21:55, 10 February 2018 (UTC)[reply]
I wouldn't call it a bug either but it is possible to avoid the section edit link on transclusion. I tried adding __NOEDITSECTION__ to {{Coin-notice}} but that only removed the section edit link from the template page itself and not from pages where the template is transcluded. Then I generated the section header with a parser function.[1] That worked, for example at User talk:86.1.101.6#Notice of Conflict of interest noticeboard discussion which had a section edit link to the template before but has no edit link now. It is not possible to make a section edit link to the page itself since there is no section to edit there. The template call is just a part of the preceding section, in this case the lead. PrimeHunter (talk) 23:33, 10 February 2018 (UTC)[reply]
Yep - looks good PrimeHunter, I tested it here and when I clicked "edit" by the section heading I was editing the correct page (note that "edit" isn't visible at my link because it's now an old page revision). Home Lander (talk) 02:11, 11 February 2018 (UTC)[reply]
Your test used {{subst:coin-notice}}. The issue was with {{coin-notice}} without subst. This produces no section edit link at all after my change to the template. PrimeHunter (talk) 03:07, 11 February 2018 (UTC)[reply]
When looking at the workaround discussion linked from the NOEDITSECTION documentation, I noticed that a simple way to make a section heading without an [edit] link is to use <h2>Heading</h2> markup instead of == Heading ==. Other than lacking the edit link, it seems such headings are treated like other section headings (having class="mw-headline" and an id anchor (or two) for linking). --Pipetricker (talk) 13:34, 11 February 2018 (UTC)[reply]
That method would also omit the section edit link when {{coin-notice}} is substituted so it's not suited here. PrimeHunter (talk) 16:02, 11 February 2018 (UTC)[reply]
Right, of course. I forgot about that ... --Pipetricker (talk) 16:59, 11 February 2018 (UTC)[reply]

Good spot, PrimeHunter on my link. In looking at the templates listed at Template:User noticeboard notices, things are totally inconsistent (even AN and ANI are not the same). Wonder if we'd be better off removing the section headers from all of these, it would be easier to remember that a section header needs to be included on every type. Home Lander (talk) 18:00, 11 February 2018 (UTC)[reply]

Other Noticeboard Notices

There are multiple noticeboards that have similar notices. Do others have the misfeature or whatever that they can be inserted in a way that renders the template subject to accidental editing? That is, what about ANI-notice and so on? Robert McClenon (talk) 21:17, 11 February 2018 (UTC)[reply]

I believe that the most common solution is "use WP:Twinkle".
I think I've also seen a method that changes the template to create a mess on the page when it's not been subst'd, which is another way of preventing people from transcluding and then editing it. WhatamIdoing (talk) 20:23, 12 February 2018 (UTC)[reply]
User:WhatamIdoing - How do I use Twinkle to notify of a noticeboard filing? I have Twinkle, but I am not sure what tab to use to put a notice on a talk page. Robert McClenon (talk) 21:09, 14 February 2018 (UTC)[reply]
@Robert McClenon: "tb" tab. --NeilN talk to me 21:12, 14 February 2018 (UTC)[reply]
User:NeilN, User:WhatamIdoing - Thank you. That is considerably more constructive than the comment above of another editor. I may ask to have the TB thingy expanded for other noticeboards. Robert McClenon (talk) 23:52, 14 February 2018 (UTC)[reply]

Title Blacklist and Block Capitals - Stupid Message

If I try to move an Articles for Creation submission that is in a sandbox to draft space, and the title consists mostly of capital letters, I get a message saying that the move cannot be done because the title is on the title blacklist. This statement is literally true but stupid and non-helpful. The title blacklist includes a regexp that blocks titles that consist mostly of upper case. The title blacklist also includes a lot of other things. If the reason is a straightforward user issue, such as the title having too many upper-case letters, it would be far better to do something like printing a short name for the rule in the title blacklist. Robert McClenon (talk) 03:59, 10 February 2018 (UTC)[reply]

There is no "short name for the rule". The default message includes the rule itself, but the messages have been customized locally to remove that (since 2008, for the move message). You're welcome to seek consensus to re-add it in some form. Anomie 14:03, 11 February 2018 (UTC)[reply]
If it does list the rule, then it listed it only in the form of a regexp, and I didn't think that, as a user, I was expected to be parsing a regexp. It certainly didn't say anything to the effect that there were too many capital letters. It should display something that is meant to be understood by human beings who are not acting as programmers. Just saying that the title is on the title blacklist is literally true but stupid and non-helpful. Oh well. Maybe I am again being told that our tools are not meant to be user-friendly. Robert McClenon (talk) 21:13, 11 February 2018 (UTC)[reply]
If it gave specifics, it could encourage people to devise devious methods of circumvention. --Redrose64 🌹 (talk) 11:28, 12 February 2018 (UTC)[reply]

Further Comment - It's Wrong

In looking at the message again very carefully, I see that nowhere does it say anything about what is wrong. Also, in the particular case, which is the use of too many capitals, I don't see BEANS as an issue at all. The problem is the use of block capitals, and just saying that means don't use block capitals, and, in that case, it will be okay. I also think that the general message, that the title is on the title blacklist, is wrong. The title isn't on a blacklist. That message always means that the title violates one of the rules about titles (of which there are many). A reasonable interpretation of what is meant by the title blacklist would be that the title is salted, but that isn't what it means. If the title is salted, there is a different error message. The error message should state that the title violates the rules on titles. If it says that, then that leaves to the author and the reviewer to figure out what the issue is, and that is all right. However, saying that the title is on the title blacklist may be accurate in a stupid pedantic sense only, but is fundamentally wrong as human interface. Change it! Robert McClenon (talk) 21:15, 14 February 2018 (UTC)[reply]

Redirects connected to a Wikidata item

Anyone know why the redlinked Category:Redirects connected to a Wikidata item is suddenly appearing in redirects such as Revolution (roller coaster). Thanks. --StarcheerspeaksnewslostwarsTalk to me 04:54, 10 February 2018 (UTC)[reply]

It's a tracking category and is put on there due to d:Q16877597 existing. Nihlus 05:00, 10 February 2018 (UTC)[reply]
But what's causing it to be added now, why isn't it a real category, and how can I fix the entries in which it is appearing so it goes away. Thanks. --StarcheerspeaksnewslostwarsTalk to me 05:07, 10 February 2018 (UTC)[reply]
Someone needed to create the category page. It is being added now because the change which makes the software add the category was rolled out recently (Thursday). The only way to correct the entries is to remove the redirect from the item listing it at Wikidata. I can help with this if you need it. However, there are some editors who do not think this is a bad thing, so this isn't really a category needing maintenance (just tracking). --Izno (talk) 14:01, 10 February 2018 (UTC)[reply]

Complete list of edits made by a user to pages in a certain category?

Hey, I was wondering if anyone knows of a tool that can analyze my contribs to list all the edits I made to articles included in, say, Category:Star Trek episodes? Like how editorinteract can show me a complete list of edits I've made to pages that happened to have been edited by another editor. Hijiri 88 (やや) 09:31, 10 February 2018 (UTC)[reply]

That's a BIG FL icon

The FL icon looked weird on on Grade II* listed buildings in Mendip (See this version) I corrected the icon by moving the {{Featured list}} template to the top of the page. (This violated MOS:ORDER, which requires the template to be down the bottom of the page.) The problem is Warning: Template include size is too large. Some templates will not be included. You can see the effect down the bottom; the references and external links sections are not processed properly. Is there anything more that we can do? Hawkeye7 (discuss) 20:51, 10 February 2018 (UTC)[reply]

@Hawkeye7: that page is in Category:Pages where template include size is exceeded, meaning not all templates will work on the page. This normally means: remove some templates - what looks obvious to me is - every single ROW of that table is another template, convert to a normal table. — xaosflux Talk 21:47, 10 February 2018 (UTC)[reply]
I've done some work to reduce the include size. It is under the limit for now. — JJMC89(T·C) 23:54, 10 February 2018 (UTC)[reply]
Rodw, is this one of the pages that we were discussing at Oxford in October?
Anyway, I have seen the huge FA/FL star twice before, see for instance Wikipedia:Village pump (technical)/Archive 161#Featured Star looks extremely massive. In both cases, as here, it was a WP:TLIMIT problem. --Redrose64 🌹 (talk) 00:16, 11 February 2018 (UTC)[reply]
Yes the whole series of sub lists of Grade II* listed buildings in Somerset follows the same pattern using Template:EH listed building header and Template:EH listed building row (as do sub lists of Grade I listed buildings in Somerset but these are smaller) and I would like to keep the consistent fromatting which also works for easy uploading of photos during Wiki Loves Monuments. I had a similar problem with List of civil parishes in Somerset which someone helped resolve. I thought the same issue would arise with List of ecclesiastical parishes in the Diocese of Bath and Wells but that doesn't seem to have been a problem.— Rod talk 08:38, 11 February 2018 (UTC)[reply]
Grade II* listed buildings in South Somerset doesn't have a star (yet) but is the one we discussed and you helped with. It is showing the template problem & I still have loads more building articles to do on that one.— Rod talk 11:44, 11 February 2018 (UTC)[reply]
Do you think it's worth changing MOS:ORDER to put the template at the top? Hawkeye7 (discuss) 21:27, 11 February 2018 (UTC)[reply]
How would that help? The {{featured article}} and {{featured list}} templates don't do much that would cause WP:TLIMIT to be hit, although both could be the straw that breaks the camel's back. --Redrose64 🌹 (talk) 23:03, 11 February 2018 (UTC)[reply]
It wouldn't help against breaking the template limit. It could help some pages display better when they do break the limit, but for some pages it might get worse depending on what is near the point where the limit is broken. I don't think the order of templates should be based on trying to guess the least bad result if the limit is broken. PrimeHunter (talk) 16:13, 12 February 2018 (UTC)[reply]

Odd red category

this edit to Tibet adds the article to Category:Articles with text from the Turkic languages collective, which does not ever appear to have existed. I cannot see how the edit adds the article to the category. Can anyone see what is going on? I have asked the editor who made the edit, but he is not very active nowadays. DuncanHill (talk) 02:06, 11 February 2018 (UTC)[reply]

@DuncanHill: For what it's worth, there are four pages in that non-existent category. Trying to see what they have in common here... Home Lander (talk) 02:20, 11 February 2018 (UTC)[reply]
I have created the category. It's added by {{lang|trk|...}}. PrimeHunter (talk) 03:13, 11 February 2018 (UTC)[reply]
Thank you. DuncanHill (talk) 18:48, 11 February 2018 (UTC)[reply]

Watchlist question

In the couple of months I was not editing Wikipedia there was a major change in the presentation of my Watchlist. Previously all changed pages would take up a single line, whether there were multiple edits or only a single one. Now, the articles with multiple edits show up as before but the ones with a single edit take up many lines - a minimum of 4 and some over 9 lines. It looks like the column width has been set to ~10 chars.

Is there any way to fix this or, even better, just return to the old watchlist? I see no benefit in the new version and, with the column issue, quite a detriment and loss of functionality, such as color coding editors and articles, that hurts my workflow. -- Thank you. Jbh Talk 20:49, 11 February 2018 (UTC)[reply]

This watchlist link disables your personal css and js. Does that look better? User:Jbhunley/common.js imports User:UncleDouggie/smart watchlist.js which hasn't been edited since 2011, and UncleDouggie hasn't edited since 2011. That script would be my first suspect but I haven't examined it. If you still have problems after disabling it then please name your skin at Special:Preferences#mw-prefsection-rendering, the setting of "Expand watchlist to show all changes, not just the most recent" at Special:Preferences#mw-prefsection-watchlist, and of "Group changes by page in recent changes and watchlist" at Special:Preferences#mw-prefsection-rc. PrimeHunter (talk) 22:07, 11 February 2018 (UTC)[reply]
PrimeHunter Safemode fixes it. I will clean up my common.js. Thank you for the help. Jbh Talk 22:25, 11 February 2018 (UTC)[reply]
Special:PrefixIndex/User:Jbhunley/ also shows other css and js pages. Some of them are unused depending on your skin. safemode also disables gadgets and site-wide css and js files but others haven't reported the problem so it's probably something in your own files. PrimeHunter (talk) 22:43, 11 February 2018 (UTC)[reply]
The offending script is User:Writ Keeper/Scripts/commonHistory.js [2]. I also disabled User:UncleDouggie/smart watchlist.js on general principles. Jbh Talk 22:51, 11 February 2018 (UTC)[reply]
Writ Keeper is still active and may want to know the preferences I mentioned, and your browser. PrimeHunter (talk) 23:05, 11 February 2018 (UTC)[reply]
I have the issue using Chrome on Win10 and iOS 11.x. I think I ran into it using Safari and Firefox under iOS 10.x and 11.x as well but I have not used them recently and it is not something that stuck in my memory.

I am willing to try it using different OS/browser configurations if that would help. Jbh Talk 23:18, 11 February 2018 (UTC)[reply]

Thanks. What skin are you using? Writ Keeper  23:20, 11 February 2018 (UTC)[reply]
Vector. Jbh Talk 01:12, 12 February 2018 (UTC)[reply]
Hmm, still can't reproduce (Chrome on Win10 as well). Maybe a passing bug? If it still happens after re-enabling the script, could you take a screenshot? Writ Keeper  02:56, 12 February 2018 (UTC)[reply]
I enabled the script and got the same error. One thing I noticed is that while the new filter system is running, right after a page refresh, the page looks fine. Once it has completed the error shows up in the same refresh that the filter header/menu is displayed. I took two screenshots of a single page refresh cycle. The first is the initial page load but while the filter system is still processing. The second is when it has finished and the filter menu and error show up. Where would you like me to send/upload the screenshots?

I sent you an email with my non-wiki email address since this may be easier to do via email. Also, I do not know how often I will be checking Wikipedia in the next couple of days. If you do not want to use email and I have not responded in a while leave a message on my talk page and I will get an email notice via my wiki-email. Jbh Talk 03:18, 12 February 2018 (UTC) Last edited: Added note about email. 03:32, 12 February 2018 (UTC) [reply]

@Jbhunley: For screenshots, please follow the directions at WP:WPSHOT. This makes them visible to all, so that all the followers of VPT can consider the matter; but it also keeps them within Wikimedia, since some people are unable to use third-party screenshot hosts. --Redrose64 🌹 (talk) 12:19, 12 February 2018 (UTC)[reply]
@Redrose64: Thank you for the link. I decided to send the shots directly to Writ Keeper. I am not too keen on the shots going up on Commons but I would be happy to email you copies if you want them. Jbh Talk 19:41, 12 February 2018 (UTC)[reply]
Just to add, it seems to do the same thing on modern/OSX/firefox. ~ Amory (utc) 13:35, 12 February 2018 (UTC)[reply]
of course i care :/ Writ Keeper  14:17, 12 February 2018 (UTC)[reply]
@Jbhunley, @Amorymeltzer et al: that should do it, I think. Let me know if you see any other problems. Writ Keeper  21:14, 12 February 2018 (UTC)[reply]
@Writ Keeper: That got it. Thank you. Jbh Talk 21:26, 12 February 2018 (UTC)[reply]

Raw viewership stats not updating

Found at [3]. No updates since FEB-08. Thanks, West.andrew.g (talk) 15:02, 12 February 2018 (UTC)[reply]

@West.andrew.g: there is a version from 10-Feb-2018 up there now. Is that what you were looking for? — xaosflux Talk 15:36, 12 February 2018 (UTC)[reply]
I only see a line saying:
pagecounts-2018-02-08.bz2                          10-Feb-2018 04:27           395916459
I assume that means the file was stored 10-Feb-2018 but has page counts for 2018-02-08. PrimeHunter (talk) 16:08, 12 February 2018 (UTC)[reply]
(edit conflict) @Xaosflux: I meant the last available data is from 08-FEB (coming on 10-FEB, strangely). Typical operation is a day's file is written just a couple hours after its UTC conclusion. Therefore we would expect to already have files up through 11-FEB. If you look at the consistently with which a day's file is written over the prior history, its clear something is amiss with the process right now. West.andrew.g (talk) 16:11, 12 February 2018 (UTC)[reply]
@West.andrew.g: looks to be something wrong, please open a phabricator ticket you can use phab:T132761 as a sample. — xaosflux Talk 16:26, 12 February 2018 (UTC)[reply]
https://dumps.wikimedia.org/other/pagecounts-ez/ says "Maintained by WMF Analytics". They have a mail list linked at wikitech:Analytics#Contact. There is already a thread about it with a reply and link to phab:T187073. PrimeHunter (talk) 16:28, 12 February 2018 (UTC)[reply]

21:59, 12 February 2018 (UTC)

Look for a reference within existing references

I realize this is a bit inside-baseball, and perhaps OT, but here goes...

I'm editing an article and noticed that I failed to have a cite on something that should have one. Googling turns up new references that have it, but I hate adding new cites when an existing source has the same information.

So does anyone know of a way to search a topic, but use the links in the page as the source documents? Perhaps there is some sort of "one degree of separation" technique? Maury Markowitz (talk) 22:07, 12 February 2018 (UTC)[reply]

You mean like googling:
my uncited claim site:url-for-reference1-from-current-article OR site:url-for-reference2-from-current-article ...
(Example)
and someone could make a script which picks up the references from the current article. A couple of problems with that: Google's input limit is quickly exceeded, so for articles with many references there would have to be several search querys (and if there aren't many references, there's not much point to the script). And there are many URLs for which site: doesn't work, partly due to URL parameters after a ? or # being ignored.
Also, while reusing an existing source could save a bit of work, I don't think it's a bad thing to use different sources for two citations even if one of the sources could be used for both. --Pipetricker (talk) 12:28, 15 February 2018 (UTC)[reply]

Loss of session data while using Microsoft Edge?

I don't know if this is a problem with the browser or the connection, but sometimes when I try to save an edit while using Microsoft Edge, the edit page either reloads (minus the edits I had made), or I get the "session was lost" message. What's happening here? Is this a known issue with the browser, or could it just be my connection? Narutolovehinata5 tccsdnew 01:22, 13 February 2018 (UTC)[reply]

Years ago, I had problems like that going all the way back to IE 6. One of the reasons that I gave up using Microsoft browsers. --Redrose64 🌹 (talk) 09:35, 13 February 2018 (UTC)[reply]

Two Column Edit Conflict view: New prototype ready for testing

The new paragraph-based approach

Hi, the Two Column Edit Conflict view has been a beta feature on all wikis for around 9 months now. More than 47.000 users have activated it as a beta feature. Feedback & data show that it's already an improvement to the default view of the edit conflict resolution page. But we also received mixed feedback, and we think we could do better - this is why we have developed an alternative prototype that is now ready for testing. It would be great if you find the time to try it out and comment on it! Feedback round is open until March 12, please see m:WMDE_Technical_Wishes/Edit_Conflicts/Feedback_Round_Paragraph-Based_Prototype for more information. Thanks a lot, --Birgit Müller (WMDE) (talk) 06:05, 13 February 2018 (UTC)[reply]

Browser issue: WikiProject banner content aligned to fill only left half of box

A new alignment problem seems to have arisen; the banner content aligned to fill only left half of box. I'm using Chromium on Linux. Maybe be a problem with template WPBannerMeta rather than WikiProject_Economics. Jonpatterns (talk) 14:54, 12 February 2018 (UTC)[reply]

@Jonpatterns: Which pages are affected? Does it improve if you WP:PURGE an affected page? If not, do you see this problem with any other WikiProject banners on the same page? --Redrose64 🌹 (talk) 22:02, 12 February 2018 (UTC)[reply]
Purge has no effect. I checked Firefox and the template displays correctly using that browser. Checking which pages are affected with Chromium. It seems many/all banners affected; for example economics banners: this page and Talk:Fundamental theorems of welfare economics, non-economics banners: Talk:Jeff Schroeder both the biograpghy and alternative music banner display the left alignment error. However, banners inside a bannershell display correct, such as those at Talk:Adam_Smith. Jonpatterns (talk) 22:50, 12 February 2018 (UTC)[reply]
Update, sometimes on Chromium when I go to the page it displays correctly at first, but whenever I press refresh it reverts to left alignment error. Jonpatterns (talk) 23:00, 12 February 2018 (UTC)[reply]
It's OK in Chrome Version 49.0.2623.112 m. I'm certain that this is not a problem in Template:WikiProject Economics and that although it might be in Template:WPBannerMeta, it's probably out of scope for Template talk:WPBannerMeta too. I'll send it to VPT. --Redrose64 🌹 (talk) 09:27, 13 February 2018 (UTC)[reply]
Looks like a bug with empty table cell dimensions in Chrome... We might have to kick this up to the Chromium project.. strange. —TheDJ (talkcontribs) 10:35, 13 February 2018 (UTC)[reply]
Looking at page source when {{WikiProject banner shell}} is not in use, I find that the banner is a table having two rows, the first row (with non-relevant content omitted for clarity) is:
<tr class="wpb-header">
  <td style="padding:0.3em 1em 0.3em 0.3em; width:50%;">WikiProject Economics</td>
  <th style="width:50%; padding:0.3em 0.3em 0.3em 0;">(Rated Start-class, Mid-importance)</th>
  <td class="mbox-empty-cell"><span style="display:none;">...</span></td>
</tr>
That, I think, is where the half-width is coming from. --Redrose64 🌹 (talk) 10:59, 13 February 2018 (UTC)[reply]
The version of Chromium I'm using is: Version 64.0.3282.140 (Developer Build) (64-bit). Jonpatterns (talk) 13:44, 13 February 2018 (UTC)[reply]

TOC suppression

Hi. Is some code in The Captive Slave suppressing the table of contents ( . . . or latest chapter in what am I missing)? Thanks? Alanscottwalker (talk) 15:10, 13 February 2018 (UTC)[reply]

Thanks, so much, I knew I was missing something. Alanscottwalker (talk) —Preceding undated comment added 15:38, 13 February 2018 (UTC)[reply]

Question

A few minutes ago, I received notification that there have been six failed attempts to log in to my Wikipedia account while I've been online. I'm not overly worried about this, because I already had a strong and long password and I changed it to something even longer — but I do still want to know if there's a way to track where the attempts came from, because obviously that's the kind of behaviour that we should be blocking editors for if we can track them down. Bearcat (talk) 22:57, 13 February 2018 (UTC)[reply]

@Bearcat: not at this time, it is only informative. As a sysop you may want to consider using WP:2FA though. — xaosflux Talk 00:51, 14 February 2018 (UTC)[reply]
And there are "blocks" of sorts on addresses that make lots of bad logon attempts - it is handled in the software as failed logon throttling. — xaosflux Talk 00:52, 14 February 2018 (UTC)[reply]

Is there a very easy to use registration or sign your name function/gadget for new editors?

Hi all

I'm working on an fairly large event with new users and I'm trying to find a gadget that will allow new users to register on an English Wikipedia by signing their name that doesn't involve source editor and copying and pasting tildes. Can anyone suggest anything? Thanks very much John Cummings (talk) 23:07, 13 February 2018 (UTC)[reply]

It shouldn't be too complicated to ask for four tildes. Someone might be able to make a pre-filled link for you so that if the editor is logged in it will subst their username. Killiondude (talk) 00:33, 14 February 2018 (UTC)[reply]
What do you mean by "register"? Do you mean "sign in" or "create a user account" (both? something else?). — xaosflux Talk 00:50, 14 February 2018 (UTC)[reply]
I mean sign their name. Thanks, John Cummings (talk) 09:02, 14 February 2018 (UTC)[reply]
Why the need to copy and paste tildes? - X201 (talk) 09:05, 14 February 2018 (UTC)[reply]
A typicalOne of many standard computer keyboards (the key beneath the Esc button in blue)
John Cummings-Try asking the participants to press shift and the key just beneath the Esc button four times when on source editor (Assuming they are using laptops or desktops) 150.107.215.94 (talk) 10:21, 14 February 2018 (UTC)[reply]
That doesn't produce a tilde on any of my British, American or Swedish keyboards. See Tilde#Keyboards. And I'm sure John Cummings can find the tilde on a keyboard. He asked for something that doesn't involve source editor. So why the need to copy and paste tildes? --Pipetricker (talk) 11:43, 14 February 2018 (UTC)[reply]
FWIW, the tilde on the multilingual keyboard I've been accidentally supplied is below the semicolon to the right of L. If one were to set the appropriate setting I think it would be typed with AltCar+; but as I've overridden it to US standard I type it with the usual key below Esc (Échap on mine) which is labelled #. Good luck! Ivanvector (Talk/Edits) 14:41, 14 February 2018 (UTC)[reply]
You could have them sign a Wikipedia project page, they can use the visual editor and it has a "sign" button right at the top of the interface. — xaosflux Talk 12:52, 14 February 2018 (UTC)[reply]
@Xaosflux:, do you mean a project page where you don't need to use source editor to sign your name? Can you give an example? Thanks John Cummings (talk) 14:05, 14 February 2018 (UTC)[reply]
On my keyboard (UK desktop layout, virtually no change from the IBM PC/AT's UK layout, and indeed the same layout as the one illustrated above), the key below Esc, when shifted, produces the "¬" character. Tilde is ⇧ Shift+# and is just to the left of the ↵ Enter key. --Redrose64 🌹 (talk) 14:27, 14 February 2018 (UTC)[reply]
@John Cummings: hmmm - scratch that I was thinking of the "sign" button on the source edit tool menu. OK though I have a hacky workaround for you: Use a MailingList. You don't ever need to actually use it for mailing, and it is easy to use. Head over to Wikipedia_talk:Mass_message_senders and MOVE one of those shells to where you want. You can edit it to put some text directions at the top. Then anyone can go to the page and just type in "User:John Cummings" and it will put them on the list. Have the directions include using the "User:" (or User talk: if you plan on actually sending a mailing one day) prefix in the box, and it will actually autocomplete most of the name from the list of regitered users. Think this will work for you? — xaosflux Talk 14:35, 14 February 2018 (UTC)[reply]
If you event is really large, make an event page for it, and maybe even a WP: shortcut (even if it is shortlived something like WP:John-a-thon-2018 :D; you can include a wiki link to the "sign in" page then. — xaosflux Talk 14:37, 14 February 2018 (UTC)[reply]

Thanks, I found FormWizard does exactly what I want but the set up is very complicated and you need to be an admin or something called interface editor. John Cummings (talk) 15:44, 14 February 2018 (UTC)[reply]

yes, FromWizzard will take some development time - check out those MMS's though - there is no coding to do, it is literally type your username and hit enter. — xaosflux Talk 15:55, 14 February 2018 (UTC)[reply]

OK, for the last couple of days I have been puzzled by this thread, wondering "why not teach them to use four tildes like everybody else does?". I have since come across Wikipedia talk:WikiProject Women in Red#Can I create an events page within Women in Red which makes it a lot clearer: the "fairly large event" mentioned above is Wikipedia:WikiProject Women in Red/Events/UNESCO 2018. --Redrose64 🌹 (talk) 21:08, 15 February 2018 (UTC)[reply]

Regex with template parameters

I would like to use a regex to determine if a template parameter value ends with numbers. The regex pattern itself is easy (\d+$) but I can't figure out how to test the input value. Basically, I want to do something like this (pseudo-code)

{{#if: {{regex match|{{{param|}}}|\d+$}} |<!-- true -->|<!-- false -->}} 

For the curious, this is for trying to incorporate Wikidata properties into {{Basketballstats}}. The problem is that the property value may be a number, a string which ends with a number, or just a string value, and the display is different depending on which it is. I assume this is possible with Lua but I never figured out how to use it. Thanks in advance. howcheng {chat} 02:10, 14 February 2018 (UTC)[reply]

See {{String-handling templates}} and m:Help:Calculation#Checking for a number. Probably something like (fake code):
{{#ifeq:{{str rightc|{{{parameter|}}}|1}}|{{#expr:{{str rightc |{{{parameter|}}}|1}}|number|not a number}}}} 
Try something like that. – Jonesey95 (talk) 02:58, 14 February 2018 (UTC)[reply]
Module:String. These examples return the whole source (|s=) if it matches the lua |pattern=, an empty string else. For details on regex-similar lua patterns see mw:Extension:Scribunto/Lua_reference_manual#Patterns
property value is a number:
{{#invoke:String|match|s=3 |pattern=^%d+$ |plain=false |nomatch=}} → 3
property value is a string with numeric suffix:
{{#invoke:String|match|s=property_val_3 |pattern=^[^%d]+%d+$ |plain=false |nomatch=}} → property_val_3
property value is a string:
{{#invoke:String|match|s=property_val_ |pattern=^[^%d]+$ |plain=false |nomatch=}} → property_val_
Trappist the monk (talk) 13:52, 14 February 2018 (UTC)[reply]

Technical glitch

I'm currently using the Android application and have encountered a glitch. While clicking edit at any deletion discussion at WP:DSI, the message We're sorry. The Wikipedia app has experienced an error and was terminated. Would you like to Start over or Quit? is displayed. The edit button works fine for articles and other Wikipedia pages though. To add, individual discussions somehow won't open in the mobile browser (Firefox) either. I hope someone takes notice. Thanks, MT TrainDiscuss 13:49, 14 February 2018 (UTC)[reply]

Howdy, MT Train, it looks like you've got two issues going on here. Let me see if I can separate them out. The first sounds novel. I don't see anything incredibly complicated about that page myself, so I'm not sure what is happening. I can't find anything related in our bug tracking system (Phabricator). So I created a new task to let the Android team know what's going on and to see if they can reproduce. If you can provide any more information about your device (OS and app version in particular) it would be helpful. To be clear, you're interacting with the section "Edit" links or the "Edit" link for the whole page?
Second issue, I'm sure there's a ticket for it, but could use some more information. Do you mean that on the same page (WP:DSI) you are having the issue with individual discussions, or in general anywhere on Wikipedia talk pages? Letting me know a little more would help me find and/or write a better task for the engineers. Thanks for the report! CKoerner (WMF) (talk) 21:16, 14 February 2018 (UTC)[reply]
@CKoerner (WMF): I'm still facing the issue on the Android application (Android 7.0), though now I'm back to using the desktop. I got a screenshot which I can upload to Commons if needed for reference. Regarding web browser, the version is Firefox 54.0.1 for Android, where the edit icon at either WP:DSI or WP:Articles for deletion won't work upon touch. Touch and hold and then release opens up several options including Open link in new tab, which opens as Cannot find section. To edit the deletion discussions, I've to type the entire page in the search bar (eg: WP:Articles for deletion/XYZ), which is a bit tedious if I'm to reply at several of them. Thanks, MT TrainDiscuss 05:22, 15 February 2018 (UTC)[reply]

Getting to the deletion log

Why doesn't Special:Deletion log work? Yes, I know it's at Special:Log/delete, but the other title should work to get you there. Oiyarbepsy (talk) 14:20, 14 February 2018 (UTC)[reply]

Sheesh, it appears the deletion log isn't even mentioned at Special:SpecialPages. How do you expect people to find it without already knowing? Oiyarbepsy (talk) 14:21, 14 February 2018 (UTC)[reply]
Special:SpecialPages#Recent changes and logs, third bullet, gives a direct link to Special:Log. It's not broken down into the different kinds of log, because there are over thirty of them. --Redrose64 🌹 (talk) 14:30, 14 February 2018 (UTC)[reply]
Wikipedia:Deletion log links to Special:Log/delete. We cannot create redirects on special pages here. You would have to suggest it at WP:Phabricator. There are some special page redirects, e.g. Special:PermaLink to Special:PermanentLink. PrimeHunter (talk) 14:41, 14 February 2018 (UTC)[reply]

Single sign on failing

I have noticed a number of occasions where SSO hasn't worked. Today I edited Wiktionary anonymously, despite being logged in on Wikipedia. Is this a known issue?

All the best: Rich Farmbrough, 14:50, 14 February 2018 (UTC).[reply]

I've had this at random intervals for years. It can happen for various reasons, usually connected with your login cookie - maybe it's corrupt, maybe it didn't get through, maybe one of the Wikimedia servers is out of sync and didn't recognise it. To guard against this, I've set my skin on all Wikimedia sites to MonoBook, so that if I visit a site and get served the Vector skin, it's an instant reminder to check my logged-in status. Doesn't work for wmuk: where MonoBook is not available - it's not a Wikimedia Foundation site so the WMF login cookie is not sent either. --Redrose64 🌹 (talk) 20:29, 14 February 2018 (UTC)[reply]
This doesn't help answer or solve the issue, but for anyone hoping to avoid it another trick is to change the color of your edit button. For instance, I have:
input#wpSave {
background-color: #88FF88 !important;
color: #000 !important;
}
If you ever get logged out, saving becomes really obvious, skin choice aside. ~ Amory (utc) 21:28, 14 February 2018 (UTC)[reply]
I added a missing }. The code goes in your CSS or a skin-specific css file. PrimeHunter (talk) 21:54, 14 February 2018 (UTC)[reply]

Archive box problem

Could somebody take a look at Talk:Canadian Broadcasting Corporation to figure out why the "archive box" template is listing a comprehensive directory of hundreds of pages that have nothing whatsoever to do with the archives of Talk:Canadian Broadcasting Corporation? Bearcat (talk) 18:37, 14 February 2018 (UTC)[reply]

User:ClueBot III/Master Detailed Indices/Talk:Canadian Broadcasting Corporation is the source of the list, but I don't know why it's populated that way. Chris857 (talk) 18:48, 14 February 2018 (UTC)[reply]
The bot made some mistake here. And that's not the only talk page where the problem exists, maybe because they are using the {{TALKPAGENAME}} as "archiveprefix"? Stryn (talk) 19:35, 14 February 2018 (UTC)[reply]
This edit is a quick fix, until ClueBot III overwrites it again. --Redrose64 🌹 (talk) 20:19, 14 February 2018 (UTC)[reply]
Has anybody informed Cobi (talk · contribs), or left a note at User talk:ClueBot Commons? --Redrose64 🌹 (talk) 20:32, 14 February 2018 (UTC)[reply]
@Redrose64: Would leaving the entire list but commenting out the wrong ones prevent the bot from repopulating it? Home Lander (talk) 23:01, 14 February 2018 (UTC)[reply]
archiveprefix={{TALKPAGENAME}} you can't do that. You need a real page name as a parameter, complete with the namespace. The bot passes that string verbatim to the Wikipedia API to generate a list of pages, and Wikipedia doesn't do template substitution for parameters sent to the API. I've seen this pop up a few times recently -- has someone added this to some docs or tool somewhere? Because it doesn't work, and should be corrected. It's not on the the documentation page. See User:ClueBot III#Required parameters. -- Cobi(t|c|b) 00:53, 15 February 2018 (UTC)[reply]
Most, if not all, of them seem to be added by DanielPenfield, and he was notified on January. Stryn (talk) 14:07, 15 February 2018 (UTC)[reply]

Monobook.js is no longer loading

Hello. User:Magog the Ogre/monobook.js is failing to load reliably for some reason. I've noticed this for a few days. I see no console messages. Can anyone help me debug? Magog the Ogre (tc) 23:31, 14 February 2018 (UTC)[reply]

Not that I can help, but knowing which skin you use and which web browser (name and version) is likely to be helpful. עוד מישהו Od Mishehu 09:51, 15 February 2018 (UTC)[reply]
knowing which skin you use -- presumably Monobook because otherwise User:Magog the Ogre/monobook.js wouldn't be loaded in the first place. {{3x|p}}ery (talk) 12:27, 15 February 2018 (UTC)[reply]

This is just insane

I know our system to create usernames has some limitations when it comes to creating a name that's too close to an existing name. While I don't know all the rules and per beans it's probably not a good idea to go into details, but I would've guessed that the system wouldn't allow there to be both user:Twisted Insane and user:TWISTED INSANE. I recently handled an OTRS inquiry from someone claiming to be Twisted Insane. I tracked down the account but the initial discussion did not go well. Eventually, I figured out that the username was TWISTED INSANE. Obviously, the system allowed both but should it have?

As an additional complication, the person contacting us has the name Twisted Insane, which now means we have a username corresponding to a real name but it's a different person (presumably).

User:Twisted Insane has no edits. Should anything be done or just ignore it and move on?S Philbrick(Talk) 21:37, 15 February 2018 (UTC)[reply]

@Sphilbrick: I created phab:T187516 to request that antispoof apply in this type of situation. An older request is to outright disallow these, it has been stalled as phab:T64396. — xaosflux Talk 23:59, 15 February 2018 (UTC)[reply]
Thanks. --S Philbrick(Talk) 02:39, 16 February 2018 (UTC)[reply]
To be fair, if the software had disallowed it, and the user had asked at WP:ACC for that name, I doubt any admin would have declined the request. עוד מישהו Od Mishehu 07:46, 19 February 2018 (UTC)[reply]
Correct. Per the ACC guide, ACC would create a similarly named account where the existing account is old with no edits even if AntiSpoof were triggered. — JJMC89(T·C) (ACC admin) 08:13, 19 February 2018 (UTC)[reply]

Dates displying backwards

In the article List of Presidents of Israel, the vital dates of the presidents appear backwards. For example, for David Ben-Gurion {{small|(1886–1973)}} appears as 1973-1866. I suspect that this has something to do with the Hebrew text immediately before the dates. Does anyone know how to fix this? Thanks, DuncanHill (talk) 14:01, 16 February 2018 (UTC)[reply]

Appears fine in my browser. The Rambling Man (talk) 14:05, 16 February 2018 (UTC)[reply]
I am using Edge on Win10. It doesn't make any difference which skin I select. DuncanHill (talk) 14:14, 16 February 2018 (UTC)[reply]
Macbook Pro winning. The Rambling Man (talk) 16:05, 16 February 2018 (UTC)[reply]
Fine for me as well on Chrome on Win7. Report the bug to Microsoft? Ivanvector (Talk/Edits) 16:13, 16 February 2018 (UTC)[reply]
I fixed it by changing from {{Hebrew|<Hebrew letters>}} to {{lang|He|<Hebrew letters>}}. However, the Hebrew letters look less ornate in the new version. If anyone understands both Template:Hebrew and Hebrew could they please have a look? DuncanHill (talk) 16:43, 16 February 2018 (UTC)[reply]
I don't think you should make that change for your personal benefit, as you say the characters are now displayed much less ornately and offer a much less enjoyable experience. I would work on fixing your own setup before modifying it to the detriment of all the rest of us. The Rambling Man (talk) 17:03, 16 February 2018 (UTC)[reply]
The previous version did not follow the instructions at Template:Hebrew, and the version I have changed to does. "This template only marks a string as Hebrew script, not as Hebrew language. Therefore, it is not appropriate for actual words in Hebrew. Hebrew words are marked instead like this: ‹Hebrew language string›." DuncanHill (talk) 17:09, 16 February 2018 (UTC)[reply]
Also, TRM, as you obviously have no idea what is going on here I'd appreciate if you stopped offering useless non-technical advice to someone who is asking for technical advice. DuncanHill (talk) 17:11, 16 February 2018 (UTC)[reply]
I obviously think you're obviously fixing things for yourself to the detriment of the rest of Wikipedia. Don't do it again. The Rambling Man (talk) 17:27, 16 February 2018 (UTC)[reply]
You want me never to use a template according to the instructions? Well, it's a point of view I suppose. DuncanHill (talk) 18:24, 16 February 2018 (UTC)[reply]
@DuncanHill: Can't you use {{script|Hebr|HEBREW TEXT}} for the ornate look of it? Nihlus 17:17, 16 February 2018 (UTC)[reply]
I tried that already, it still reverses the dates. DuncanHill (talk) 17:19, 16 February 2018 (UTC)[reply]
Names are words. I used the format which according to the template documentation is appropriate for words, instead of the format the documentation says is not appropriate for words. DuncanHill (talk) 17:20, 16 February 2018 (UTC)[reply]
I never stated you didn't follow the instructions; I was merely offering an alternative that you had yet to say you had tried. That being said, my computer sees it the same with all of them, which is something I wasn't paying attention to. Nihlus 17:27, 16 February 2018 (UTC)[reply]
Pretty sure that the matter of which template to use (and their differing effects) has come up in the last year or two on another talk page (possibly Template talk:Lang): it was certainly in connection with Hebrew. What we have is (a) a template that specifies what language a piece of text is written in, without specifying the font; and (b) a template that specifies which font should be used for displaying a piece of text, without specifying the language. One of them sets the dir=rtl attribute, the other doesn't. If it's not explicitly set, your browser guesses. --Redrose64 🌹 (talk) 19:53, 16 February 2018 (UTC)[reply]
Numbers like years can both be displayed left-to-right and right-to-left depending on the text next to them. It varies what browsers choose when there are no Latin letters between right-to-left text and numbers. I have inserted left-to-right marks &lrm; between the righ-to-left text and the years to help browsers choose. PrimeHunter (talk) 20:12, 16 February 2018 (UTC)[reply]
Thank you. DuncanHill (talk) 20:17, 16 February 2018 (UTC)[reply]
{{lang}} outputs a left-to-right marker when |rtl=yes:
{{lang|he|rtl=yes|דוד בן-גוריון}}<span title="Hebrew-language text"><span lang="he" dir="rtl">דוד בן-גוריון</span></span>
No need (in this case) to duplicate that.
Trappist the monk (talk) 20:25, 16 February 2018 (UTC)[reply]
Thank you. The Rambling Man (talk) 20:46, 16 February 2018 (UTC)[reply]
Aha, it was Template talk:Lang/Archive 3#Rendering Hebrew - exactly a year ago. --Redrose64 🌹 (talk) 21:41, 16 February 2018 (UTC)[reply]
My take on this is that the ornateness is really useful - for Hebrew, it can help distinguish between certain letters. Currently, the code
{{font|font=Ezra SIL|size=115%|{{lang|he|rtl=yes|דוד בן-גוריון}}}}
has the desired effect, while maintaining the lang parameter. But, it is ugly. It shouldn't be too hard to make a template that fixes the font, but doesn't pass judgement on the contents, which is probably the best way forward r/e not breaking stuff. Of course, if {{Hebrew}} can be fixed, it would be better. Bellezzasolo Discuss 12:35, 19 February 2018 (UTC)[reply]

Rangeblocklog

Looking for a tool similar to rangecontribs that could condense and display all of the block logs for a given IP range. For example, if I supply the CIDR 192.168.105.0/24, it would output the block log of any IP address that has been blocked before in that range, as well as any blocks of sub-ranges within that range. Thoughts? Ivanvector (Talk/Edits) 14:41, 16 February 2018 (UTC)[reply]

Ideally Special:Log would do this, see phab:T146628. I'm going to look into it, because it might be very similar to phab:T163562, which I worked on. In the meantime, it's probably possible to create an off-wiki tool, but it'd be quite slow. MusikAnimal talk 19:44, 16 February 2018 (UTC)[reply]
Yeah, that seems like what I have in mind. I can't contribute much to coding, but I make plenty of rangeblocks and can probably help test if/when needed. Following both tickets. Ivanvector (Talk/Edits) 07:33, 18 February 2018 (UTC)[reply]

nofollow

If nofollow were to be disabled (probably through an RFC, though it was enabled in 2007 without community consensus on the request of Jimmy Wales), would this still have negative effects related to search engine optimization? The English Wikipedia has obviously changed a lot in the intervening years, so removing it might be more feasible now, and it could possibly have positive side effects by somewhat increasing the page ranks of heavily-used sources relative to unreliable sources. Jc86035 (talk) 15:14, 16 February 2018 (UTC)[reply]

There is some discussion at phab:T44594 regarding the default status of the nofollow configuration parameter in MediaWiki. --Izno (talk) 15:43, 16 February 2018 (UTC)[reply]
There is some other chatter at phab:T54617 and elsewhere in this Phabricator search. --Izno (talk) 15:47, 16 February 2018 (UTC)[reply]
@Izno: Most of those are about removing it (or adding it) in default MediaWiki, Wikibase and/or Parsoid, which would be different to removing it from just the English Wikipedia, which is certainly better maintained than most other WMF wikis. I suppose without any evidence more recent than 2007, and very little recent discussion, it's difficult to tell whether nofollow has any harmful effects (perhaps a trial period like WP:ACTRIAL might be useful). Jc86035 (talk) 16:19, 16 February 2018 (UTC)[reply]

Is it possible with JS to automatically change the colour of the show/hide link in mw-collapsible divs which contain one mw-collapsible-content div and one other div (presumably the title)? I would have liked to change the colour of the show/hide button in {{Infobox YouTube personality/sandbox}}, which is currently dark blue on dark red. (For mw-collapsible tables and mw-collapsible divs, this is done by a few lines of JavaScript in MediaWiki:Common.js (lines 255–261 in the newer code editor); but this doesn't work if there's a mw-collapsible-content inner div and the colour of the title is handled by a separate inner div as in that template.) However for that template specifically it's probably better to use a table instead, since the div – previously a NavFrame – itself contains a table. Jc86035 (talk) 16:59, 16 February 2018 (UTC)[reply]

Do you want this changed just for you, or for everyone? — xaosflux Talk 21:52, 16 February 2018 (UTC)[reply]
@Xaosflux: For everyone; it would make sense for the collapse button to have equivalent functionality in tables and divs. (An extra class might be needed to indicate the header div.) Jc86035 (talk) 11:53, 18 February 2018 (UTC)[reply]
Header
Content
Header
Content
For some reason the collapse button in the div (on the right) isn't displaying, even though it's there in the page source. What's happened to it? Jc86035 (talk) 11:59, 18 February 2018 (UTC)[reply]

Greetings, WIkipedians. Now that I graduated WP:TWA what should I do with my contrib history? (16 Feb 2018)

Hi there, anyone who can help.

I just can't help but notice, and worry, and ponder about all of the stuff I did while in TWA, as they appear in your contribs history,and there is an army of them; AN ARMY! ; There appears to be over 50 edits made via the program (which I liked), though now it feels like a big clog. Considering that people look at your contributions, in a way, like your personal recors inspite of policy saying otherwise, thus I'm concered that I won't be able to progress here, because they seem repetious, a bit vandolous due to the quality, (considering the system tags me as doing those other wikipedians' edits [which on top of that makes it look very suspicious], among other things). Here's what it looks like:

Big list in here
  • Trial of template:

16 February 2018

After looking at those, I would start asking the user a lot of questions, which won't help things at User:Zanygenius/Please Adopt Me!, and thus I could be cabaled, blocked, or worse. Could you tell me, with this mess, what to do with it? Or frankly how? Thank you. Sincerely, User: Zanygenius(talk page) 00:06, 17 February 2018 (UTC)[reply]

As far as I know, there is nothing you can do to change history. But I wouldn't worry about it. Your first week? As long as you weren't actively behaving badly, no one cares. Just start making good contributions. --Trovatore (talk) 00:30, 17 February 2018 (UTC)[reply]
@Zanygenius: nothing, once you have 10000 edits it will be nothing :D They are obviously marked as part of TWA and are 'normal'. — xaosflux Talk 01:46, 17 February 2018 (UTC)[reply]
However, if you REALLY want them gone, you can throw a {{db-userreq}} on the ones in your userspace and an admin will delete them for you. — xaosflux Talk 01:48, 17 February 2018 (UTC)[reply]
Thanks everyone, including @Trovatoere and Xaosflux:. This should do the trick. Sincerely, User: Zanygenius(talk page) 15:33, 17 February 2018 (UTC)[reply]
@Zanygenius: Nobody cares about TWA, certainly not in a bad way. I do recent changes patrolling at times, and I skip right over the TWA stuff. I mean, in theory, there could be policy violations in it, but it's all in (your[1]) user space. It generally shows, I think, a beginner learning how Wikipedia works, and nobody is going to complain about that - the alternative is we have test edits on live pages. Bellezzasolo Discuss 12:57, 19 February 2018 (UTC)[reply]
  1. ^ it doesn't belong to you, but it's about you, but you've basically got editorial control unless your silly

Less space below the last message on an IP talk page

Is this some recent change to the software?— Vchimpanzee • talk • contributions • 17:27, 17 February 2018 (UTC)[reply]

That was the result of this edit request. The previous version used the line break tag, which is said to be generally best avoided. But if people think there should be a space, I suggest we use CSS to realize it instead of reverting the request. For instance:

Nardog (talk) 17:49, 17 February 2018 (UTC)[reply]

Welcome banner


Hi - I just completely redesigned the gif displayed above and uploaded it. So upload was successful but old version still displays on my talk page, as well as above. Yet when I click on the gif on my talk page it displays my new version. How can both versions exist simultaneously on the same damned page? Someone please kill the old version - my new version should display on all associated pages. . MarkDask 18:03, 17 February 2018 (UTC). Answered - thanks. MarkDask 18:16, 17 February 2018 (UTC)[reply]

The animated gifs are hard to tell part for others when they don't have a constant image to compare but I think the new version is displaying now. If you still see the old then try to bypass your cache. PrimeHunter (talk) 19:05, 17 February 2018 (UTC)[reply]
There is a typo in your banner, at least I don't know what else "tervetulos" could mean than Finnish language. In Finnish it is correctly written as "tervetuloa". Stryn (talk) 19:11, 17 February 2018 (UTC)[reply]
Thanks PrimeHunter, and yes Stryn I saw that. I'll upload a corrected version some time today. MarkDask 01:20, 18 February 2018 (UTC)[reply]

How to learn scripts

I would like to learn how to write scripts for use at en-wiki, probably just for my own use. I have long experience with multiple programming languages, but nothing web-based. Can anyone suggest on- and off-wiki learning resources? ―Mandruss  17:59, 17 February 2018 (UTC)[reply]

@Mandruss: Do you mean JavaScript, or Scribunto modules (Lua), or something else? Jc86035 (talk) 13:29, 18 February 2018 (UTC)[reply]
@Jc86035: I think I mean JavaScript. ―Mandruss  13:33, 18 February 2018 (UTC)[reply]

Strange edit conflicts

I have had a couple of edit conflicts with myself on the same posting recently. Has anyone else had a similar problem? Firefox on Windows 10, source editor. Save takes a loooong time, ends in edit conflict screen. Comparison of content shows that the edit was saved then it seems to have tried to save it again. Only difference being unexpanded signature in second attempt. The latest conflict on a one line talk page edit. · · · Peter (Southwood) (talk): 07:24, 18 February 2018 (UTC)[reply]

I had one today. All the best: Rich Farmbrough, 11:26, 19 February 2018 (UTC).[reply]

Watchlist layout (follow-up)

@PrimeHunter, TheDJ, and Redrose64: If you remember, on 2nd February I posted a proposal/query titled Watchlist layout and the three of you replied. Well, using Redrose's suggestion and a little perseverance I cracked it! There are actually 2 parts to this; The first is extremely straightforward and I recommend everyone to do it, which is to add a single line to your common.css to remove a completely redundant bit of whitespace:

.mw-changeslist-line-prefix { width:4px; }

The second part is slightly more drastic in that it makes much more of a difference to the layout. I suggest you try it, and if you don't like it you can remove it. Simply copy/paste the following script into your common.js:

// Remove whitespace from watchlist margin
if (wgPageName=="Special:Watchlist") {
  var wlm=document.querySelectorAll("td.mw-enhanced-rc");
  for (var i=0; i<wlm.length; i++) {
    var wlmHTML=wlm[i].innerHTML.replace(/&nbsp;/g,""), wlmChar=wlmHTML.substr(wlmHTML.length-3,1);
    if ((wlmHTML.length>5) && (wlmChar==":")) { // Move timestamp to beginning
      wlm[i].innerHTML=wlmHTML.substr(wlmHTML.length-5,5)+"&nbsp;"+wlmHTML.substr(0,wlmHTML.length-5)+"&#x202F;"; }
    else wlm[i].innerHTML=wlmHTML+"&nbsp;" }}

First it strips away all the &nbsp; then, if necessary, it takes the timestamp from the end of the string and puts it at the beginning, so that any markers (m, b, etc.) are pushed into the start of the page name, leaving everything nicely aligned and clearly spaced. Enjoy! nagualdesign 07:54, 18 February 2018 (UTC)[reply]

You shouldn't modify the dom using innerHTML, it breaks all javascript logic that is attached to it (like thanks controls for instance). Use replaceChild, appendChild, removeChild and other DOM manipulation methods. —TheDJ (talkcontribs) 09:39, 19 February 2018 (UTC)[reply]
The only changes are to the layout of table cells that have no JavaScript functionality, and having just done some preliminary tests everything appears to work fine. I appreciate your comment though and I'll read up about those methods. I'll try thanking you from my watchlist... nagualdesign 18:13, 19 February 2018 (UTC)[reply]

Automated access to deleted content

I'm interested in analysing content in pages deleted per G11 in order to try and detect other spam pages as they are created. As an admin, I access deleted revisions in a browser, and could potentially write a script to scrape it out from Special:Undelete but that would mean typing my password in plaintext, which I'm sure everyone would agree is a bad idea (and using 2FA, I presume this wouldn't be possible anyway). So is there some other way I can get hold of the data? In theory this database query with the addition of join text on ar_text_id = text_id should work, but the text table isn't publicly available. Would someone at WMF be able to run the queries on the real databases and send me the data? SmartSE (talk) 12:49, 18 February 2018 (UTC)[reply]

@MusikAnimal (WMF) and Kaldari: (not sure what your staff name is kaldari). --Izno (talk) 15:33, 18 February 2018 (UTC)[reply]
Use the action API (specifically, the prop=deletedrevisions module) to access the content, with OAuth for your script's authentication to avoid having to enter your password into it. Anomie 18:26, 18 February 2018 (UTC)[reply]
Wikimedia uses external storage for content, so it's not in the text table, even in production. I'm not actually sure how to access content without using the API, so yeah, API is the way to go :) If you are having trouble getting your script to use OAuth, Special:BotPasswords is another option (though less secure). Also, not sure if you knew about mw:ORES? It does a fair job at identifying G11 spam, and there's an API. MusikAnimal talk 03:48, 19 February 2018 (UTC)[reply]
@Anomie and MusikAnimal: Thanks. OAuth and API looks like the way to go and hopefully I can figure out how to make it work. I do know about ORES but I'm actually more interested in detecting UPE rather than the more blatant spam, but I'm curious to see whether word frequencies in G11s match up with UPE. There's an awful lot slipping through at the moment, so I presume this isn't being caught by ORES. SmartSE (talk) 13:27, 19 February 2018 (UTC)[reply]
@Anomie: I've gotten pretty close to getting it to work but adding my tokens to this gives me "mwoauth-invalid-authorization" "The authorization headers in your request are not valid: Invalid consumer". Had a look around but can't find any solutions. SmartSE (talk) 21:36, 19 February 2018 (UTC)[reply]
User:Halfak (WMF): Don't you already have a tool that does this, or is it for a different CSD criterion? Whatamidoing (WMF) (talk) 16:44, 19 February 2018 (UTC)[reply]

Pages Not Fully Loading?

I've been seeing a problem recently of Wikipedia pages not fully loading. Basically I see the title and a few paragraphs of the content, but nothing else. Not even the links along the left side. Has anybody else noticed this? It seems to be getting progressively worse. Praemonitus (talk) 15:39, 18 February 2018 (UTC)[reply]

I've seen the page content loaded but the browser still trying to load something. And I know my problem is related to Firefox version 58, but I have not reported it on Phabricator because I don't have a way to reproduce. It just randomly happens. Stryn (talk) 17:40, 18 February 2018 (UTC)[reply]
See the "Network" tab of your web browser's developer tools to debug where things get stalled. mw:Help:Locating broken scripts provides pointers. --AKlapper (WMF) (talk) 21:52, 18 February 2018 (UTC)[reply]
Thanks. I'm using Chrome 63.0.3239.132. Praemonitus (talk) 03:26, 19 February 2018 (UTC)[reply]
@Praemonitus: I'm running Chrome 64.0.3282.167 and it has sputtered a few times today when trying to load pages, particularly multiple tabs at one time, such as when I was tagging G13s earlier. Home Lander (talk) 04:16, 19 February 2018 (UTC)[reply]

Mass rollback

The mass rollback script that I am currently using seems to be out of date, as it hasn't been working for a while. Does anyone know where I can finding a currently working version? Many thanks, Alex Shih (talk) 10:21, 19 February 2018 (UTC)[reply]

User:Writ_Keeper/Scripts/massRollback.js works Galobtter (pingó mió) 10:26, 19 February 2018 (UTC)[reply]
@Alex Shih: Yes, Writ Keeper's script is great. It gives you the rollback all option in a dropdown menu next to the Twinkle dropdown - but I think you told me you don't use that! Home Lander (talk) 16:10, 19 February 2018 (UTC)[reply]
Thanks guys. I thought I was using Writ Keeper's script. It was working for me before, but just stopped working all of sudden at one point for some strange reason. I'll try again. And yes, my despise of Twinkle is probably on the same level as Ritchie333. Alex Shih (talk) 22:35, 19 February 2018 (UTC)[reply]

Redirect help

I am trying to get to the page that redirects  Religious sisters to  Nun#Distinction between a nun and a religious sister  but I find no indication on the Nun page or in the instructions at  Wikipedia:Redirect#How to edit a redirect or convert it into an article  on how to get there. It says "you must use a special technique in order to get to the redirect page itself" but I don't see where it explains that technique. How do I get to this redirect page to edit or expand it (and so to learn the process)? Thanks for any help with this! Jzsj (talk) 17:31, 19 February 2018 (UTC)[reply]

After being redirected, scroll up to the top of the page, where is should say "redirected from Religious sister". Click on that link. {{3x|p}}ery (talk) 17:34, 19 February 2018 (UTC)[reply]
Pperry already explained it, but you can also add the "redirect=no" parameter to the url, such as https://en.wikipedia.org/wiki/Religious_sister?redirect=no or https://en.wikipedia.org/w/index.php?title=Religious_sisters&redirect=no ~ Amory (utc) 17:36, 19 February 2018 (UTC)[reply]
With your help I was able to create Brother (Catholic) but I still don't find any evidence of a redirect at Brother (Christian) (or at Nun, for Religious sisters). So I don't see how to find the Brother (Catholic) page to edit it. How to get back there, please. Jzsj (talk) 18:24, 19 February 2018 (UTC)[reply]
You only se "Redirected from" when you use the redirect so click on Brother (Catholic) and scroll to the top. PrimeHunter (talk) 19:43, 19 February 2018 (UTC)[reply]

How to create scripts?

Hi. I want to create a script that fixes pages in Category:CS1 errors: dates. Unfortunately, I don't know how to code on Javascript. I want the script to be able to compare the citation date with the page's history so that it can determine when the reference was added if the error is in the access-date parameter. If the error is in the date parameter, I want the script to compare the date of when the source in question has been created and fix the error by substituting the date in which the source has been created in the correct format so that the error is fixed. Can you please help me out. Thanks. Pkbwcgs (talk) 20:28, 19 February 2018 (UTC)[reply]

Having done this manually, I'm not sure it is a task amenable to scripting. The problematic cases are partial or malformed access dates that are not close to the date of the edit which added them. — jmcgnh(talk) (contribs) 21:27, 19 February 2018 (UTC)[reply]
Right, |access-date= must specify a day, not just a month or year. I just randomly checked one. Its access-date parameter just gives the year 2008. It's a dead link, and archive.org came back with an "try again later" error message... Seems like a very tedious and time consuming task to do manually, but we have coders who know how to interface with archive.org I think. Cyberpower678 comes to mind. access-date is not a required parameter, so one way to fix it would be to simply remove the date entirely when only a partial date is specified, as that's been deemed "not good enough". Once anyone goes to the trouble of verifying that it's still accessible then they would want to update the parameter with the current date. You can't really guess what day the original editor accessed it in 2008, and they're unlikely to remember. Unless you assume they accessed it on the day that they added the reference. Then you could have a bot search the edit history to find the day that the reference was added and use that day. I could do that in PHP using the API... but I have already have a to-do list that's a mile long. wbm1058 (talk) 22:30, 19 February 2018 (UTC)[reply]

22:54, 19 February 2018 (UTC)

Transclusion count by namespace?

Template:Infobox is used (transcluded) on over 3 million pages, per THIS tool. I'm interested in breaking that down by namespace. Is there a tool that does that? For example just under 2000 talk pages transclude it, and tens of thousands of user pages transclude it. What I'm looking for is the percentage of mainspace pages {{NUMBEROFARTICLES}}: 6,864,188 that transclude it, which is a considerably smaller number than {{NUMBEROFPAGES}}: 61,195,530.

Not sure it's safe to say that 3 out of 5 article-space pages have infoboxes, though roughly that may be the case? wbm1058 (talk) 22:57, 19 February 2018 (UTC)[reply]

@Wbm1058: Special:Search is probably pretty close to the actual count. --Izno (talk) 23:08, 19 February 2018 (UTC)[reply]
Cool, Izno, thanks! How did you find out about hastemplate:infobox – is that new? It's not documented at Help:Advanced search § Parameters where the three main search parameters prefix, intitle, and incategory are documented. Now I'm wondering what other parameters there are that I might not know about.
So about 2,916,970 of 3,039,916 {{infobox}} transclusions (96%) are in mainspace, and 2,916,970 of 6,864,188 pages in mainspace that qualify as articles transclude infoboxes (52%). – wbm1058 (talk) 23:49, 19 February 2018 (UTC)[reply]
mw:Help:CirrusSearch has more documentation. PrimeHunter (talk) 00:05, 20 February 2018 (UTC)[reply]
@Wbm1058: Template:Infobox is transcluded 2,917,009 times in the main namespace, as per quarry (quarry.wmflabs.org).--Snaevar (talk) 23:42, 19 February 2018 (UTC)[reply]
That figure is pretty close to the one given by hastemplate:infobox – I suppose the difference could be explained by caching lags. – wbm1058 (talk) 23:49, 19 February 2018 (UTC)[reply]

Thanks not working

Hi, Not sure if it's just me but my thanks doesn't seem to be working?
I go to thank someone (it then says thanked as it should) but then when I thank someone else on the same page it doesn't do anything and then when I reload the page the edit I thanked no longer says "thanked" (I just have the option to rethank basically),
Thanks, –Davey2010Talk 00:15, 20 February 2018 (UTC)[reply]

Davey, see if the thanks you tried to send are in your log. I don't know why it doesn't seem to work twice for you on the same page, but I do know that "thanked" reverting to "thank" when you come back to a page is an annoying issue that has irritated me for a while, but I've just kind of put up with it. If you want me to check if I get the same issue, reply to this, and I'll see if I can thank you for both edits from the page history. -- Begoon 01:13, 20 February 2018 (UTC)[reply]
I thanked Davey2010 for his edit here, and it now says I thank Begoon. I'm going to file a bug MusikAnimal talk 01:25, 20 February 2018 (UTC)[reply]
The system tells me you thanked me twice, MA, according to my "notices"... -- Begoon 01:28, 20 February 2018 (UTC)[reply]
Yep, I tried a second time and it thanked you for your other edit. It's thanking whatever the most recent non-thanked revision is. I created a bug at phab:T187757 MusikAnimal talk 01:30, 20 February 2018 (UTC)[reply]
Well, just to maybe add to the confusion, I thanked you both, from the page history, and each time I got the expected message, and "thank" said "thanked" for both edits on the page history after the second one. My log has 2 entries, one for MA, one for Davey... -- Begoon 01:35, 20 February 2018 (UTC)[reply]
Maybe you thanked before intermediate edits were made? Because that would mean Davey and I's edits are the most recent ones that aren't yours, if that makes sense. Either that or it is magically working for you but not others. @Begoon and Davey2010: What browser/OS/skin are you using? MusikAnimal talk 01:39, 20 February 2018 (UTC)[reply]
Vector, FF 58.0.2, Win 7. And yes, you could be right, looking at the times of the edits/thanks they could well have been the last 2 edits that weren't mine (Nagual edited in the same minute, I don't have seconds to compare.) -- Begoon 01:47, 20 February 2018 (UTC)[reply]
Hi Begoon, I've never had an issue before but got your thanks (thanks! :) ),
MusikAnimal - My thank log now shows as
"01:01, 20 February 2018 Davey2010 (talk | contribs) thanked My name is not dave (talk | contribs)
00:59, 20 February 2018 Davey2010 (talk | contribs) thanked Tacyarg (talk | contribs)
00:10, 20 February 2018 Davey2010 (talk | contribs) thanked Casliber (talk | contribs)
00:09, 20 February 2018 Davey2010 (talk | contribs) thanked Dodger67 (talk | contribs)
00:09, 20 February 2018 Davey2010 (talk | contribs) thanked Callanecc (talk | contribs)"
All of which I've never thanked (I tried thanking Ritchie333, SoWhy and Lourdes on Lourdes's RFA),
I'm currently using Vector skin, Chrome, OS is Windows 7, Thanks, –Davey2010Talk 01:51, 20 February 2018 (UTC)[reply]
I get the same error and added an example with screenshots to phab:T187757. This is serious. I suggest we hide thanks links until it's fixed. We can do it by placing the below in MediaWiki:Common.css. PrimeHunter (talk) 02:11, 20 February 2018 (UTC)[reply]
.mw-thanks-thank-link {display: none;}
@PrimeHunter: I was about to suggest the same. I'm going to up the task to Unbreak Now, but there's a chance it won't get fixed until tomorrow. So yeah, let's hide it. I don't think we can hide it on mobile too, can we? MusikAnimal talk 02:22, 20 February 2018 (UTC)[reply]
I don't know a way we can hide it in mobile but I suspect the vast majority of thanks are from desktop. Pages with thanks links are harder to find in mobile and most editors are in desktop. PrimeHunter (talk) 02:31, 20 February 2018 (UTC)[reply]
I agree. I've gone ahead and hidden it with Special:Diff/826612112 MusikAnimal talk 02:32, 20 February 2018 (UTC)[reply]

I've put this in a subsection so as not to pollute the discussion of the above serious error, but I'd still be interested in an eventual answer to this, which was touched on above. While the thanks "links" on this page history for the thanks I just made today, above, still show as an unlinked "thanked" no matter how many times I refresh or purge, if I revisit the page histories for the thanks I sent yesterday the link has reverted to a blue-linked "thank", with the opportunity to do it again. Is this how it's supposed to work (ie the links revert after a period of time)? -- Begoon 02:27, 20 February 2018 (UTC)[reply]

I think you probably had the same issue. When you thank in the interface, it looks like it goes through, but if you refresh, it ends up having thanked the most recent revision. The ones you did today were probably the most recent revisions, hence why they worked. MusikAnimal talk 02:36, 20 February 2018 (UTC)[reply]
No, I think it's separate. The edits which have reverted to "thank" are correctly shown as thanked in the log. Also, at least one of them was definitely the last revision when I thanked it (and still is, as I type). Also, also, this has done this, like this, for some time (weeks, maybe months), I've just never asked or "complained" till Davey mentioned something similar. -- Begoon 02:41, 20 February 2018 (UTC)[reply]
For as long as I can remember, any time I would log out and log back in, it would reset any edits I had thanked and allow me to thank them again. They remained in the log properly. Home Lander (talk) 02:49, 20 February 2018 (UTC)[reply]
Yeah, actually I do recall seeing that as well -- where the Thanks did go through but still reverted back to a Thanks link. For me it would revert back after a few days, not immediately, if I remember correctly. Anyway I agree this is probably an unrelated issue. I don't see a bug for it in Phabricator, so let's revisit it after the above bug is fixed.

Unrelated, I love that Nihlus just Thanked me for removing Thanks =p MusikAnimal talk 02:51, 20 February 2018 (UTC)[reply]

I was testing out the Mobile to see if there were any issues. :P Nihlus 02:52, 20 February 2018 (UTC)[reply]

I'm guessing this issue arose today; all my thanks in the log from yesterday and earlier appear to be correct. Guess I'm glad I didn't try to thank anyone today. Home Lander (talk) 02:59, 20 February 2018 (UTC)[reply]

Thanks not appearing

I'm presenting problems with the thanks feature too. In the article's history the thank option disappeared from the edits, and reads as (undo | ). Only those edits I have thanked before appear as (undo | thanked). I noticed that in the Spanish Wikipedia my option is shown normally. --Jamez42 (talk) 03:24, 20 February 2018 (UTC)[reply]

@Jamez42: That's because MusikAnimal removed it for now; see above thread. Home Lander (talk) 03:29, 20 February 2018 (UTC)[reply]
I added a watchliist notice for this outage, pointing to this thread. — xaosflux Talk 04:14, 20 February 2018 (UTC)[reply]
Thanks for the watchlist notice! I thought it was just me for a moment. Alex Shih (talk) 05:28, 20 February 2018 (UTC)[reply]
@Xaosflux: I just thanked you on your watchlist notice edit and I notice the action is logged both in my thanks log and your received thanks log. Did you receive notification for this or is it just not going despite the logs showing so?–Ammarpad (talk) 06:17, 20 February 2018 (UTC)[reply]
The problem appears to be that the thanks function is thanking the last unthanked edit in the page, regardless of which edit is chosen, so in many cases it is thanking the wrong edit/user. If the edit you thanked happened to be the last unthanked revision you were probably "lucky" and it will probably have worked "ok", by chance. The logs seem to be correct, but the feedback you receive at the time of thanking may not be. Don't rely on it at all until it's fixed would be my advice, though. -- Begoon 06:36, 20 February 2018 (UTC)[reply]