This page contains discussions that have been archived from Village pump (technical). Please do not edit the contents of this page. If you wish to revive any of these discussions, either start a new thread or use the talk page associated with that topic.
From the technical side it's no problem, but it should be discussed with the relevant projects (see the talk page of the template). A description where exactly in the template and how is helpful, too. The template is used in 9000 articles. --mfb (talk) 11:48, 30 November 2020 (UTC)
Not really a problem, but more of a curiousity question. While the Wikipedia logo may look different for each Wikipedia, most of them are white puzzle balls with an opening on the top. However, wg-en.wiki.x.io (see here) has the meta logo on top, why is that? Also, what does that wiki even look like? I never got a chance to see it, now it’s read only. --AJ1m3,zsd. (talk) 17:52, 2 December 2020 (UTC)
I can't see where that extra slash comes from. Nor the s in "Archives" for that matter. Bonus questions: Solomon's is used three times. Why not just Solomon's? Is {{Archive basics}} really needed? Johnuniq (talk) 06:22, 2 December 2020 (UTC)
ClueBot III can sometimes format archives like this when it's incorrectly configured, because it's trying to use its format for archiving by date (e.g. Talk:Solomon's Temple/Archives/2010/January). I think I've fixed it ... as explained on User:ClueBot III, the use of the HTML entity as opposed to the real apostrophe can cause problems with the bot. The {{Archive basics}} template is for OneClickArchiver. Graham8709:57, 2 December 2020 (UTC)
On WP:Protection policy#Overview of types of protection, it says that template protection is for high risk templates and modules. If so, why are some high risk templates fully protected? I though template protection was for high risk templates, not full protection. Not just the templates used on the Main Page, which are understandable, but editnotices are supposed to be edited by template editors. Template:Editnotices/Namespace/Module talk is fully protected, and some templates are as well, why is that? --AJ1m3,zsd. (talk) 23:56, 2 December 2020 (UTC)
@AJ1m3,zsd.: templates are subject to protection just as any other page, generally templates that are incorporated in to fully protected areas such as the mediawiki namespace (such as the example you just pointed out) are fully protected to match the protection of the area they are used within. Additionally, extremely high use templates are fully protected to ensure they are given the highest scrutiny. If you have a question about a specific protection, you can start by asking the protecting admin if they are still active - if you think a page should have protection reduced you may post at WP:RFPP. — xaosfluxTalk00:16, 3 December 2020 (UTC)
Dimitris Paraskevas has several of these pairs of span tags:
<span class="citation book" contenteditable="false">...book info as plain text ...</span> <span class="Z3988" title="ctx_ver=Z39.88-2004&rft_val_fmt=... lots of codes ..." contenteditable="false"> </span> I assume that these are a side-effect of pasting in a reference from some sort of library record.
Can I clean all this by just replacing both spans with a simple {{cite book}} — or do they have some deeper meaning? — GhostInTheMachinetalk to me19:20, 2 December 2020 (UTC)
This is one of many bugs related to the Content Translation tool. See T218420. The span tags can be removed. Just be careful to preserve the content inside of them. – Jonesey95 (talk) 20:50, 2 December 2020 (UTC)
@Jonesey95: regarding your comments at T218420 - the span having the attribute title="ctx_ver=Z39.88-2004&rft_val_fmt=... lots of codes ..." is nothing to do with content translation, it is in fact the COinS metadata for the citation, see WP:COINS. To demonstrate this, go to any Wikipedia page that has refs. Make sure that at least one of the refs uses a citation template (it doesn't matter whether it's {{citation}}, {{cite web}}, {{cite book}}, {{cite journal}}, or another one). Verify from the page history that content translation has never been used. Whilst viewing the rendered article, use your browser's "view page source" feature, often reached via Ctrl+U. There, you will find the spans in question. --Redrose64 🌹 (talk) 22:51, 2 December 2020 (UTC)
I am open to the possibility that this bug can be caused by other processes or tools, but please see the comment from Pginer-WMF dated 10 Dec 2019 for a demonstration of how to use the Content Translation tool to reproduce this bug. – Jonesey95 (talk) 00:39, 3 December 2020 (UTC)
Autoaccept error
Please explain why my changes to the article Kalarippayattu which has been protected with autoconfirmed permission are not auto accepted like this [[1]] even if I'm an extended confirmed user?. Thanks Outlander07@talk17:09, 2 December 2020 (UTC)
I just discovered that my Global account information page says that my account was registered (and attached to en.wiki) in February 2011, even though I actually created it in November 2006 (as seen from the user creation log), and have been editing more or less regularly since then. What does the word "registered" refer to here? Unified login was my first guess, but it looks like that happened a little later than 2011. It's not a big deal, but it makes me curious! --bonadeacontributionstalk12:32, 3 December 2020 (UTC)
I don't see this anywhere, so I guess I'm the first to ask. Why did I clearly see the word "Permanent" with a down arrow to the right of "Watch this page" while editing, and now I see a fuzzy version of the same thing that doesn't work?— Vchimpanzee • talk • contributions • 21:04, 2 December 2020 (UTC)
You can choose to watchlist a page you edit forever, or not at all. There are now extra options to watchlist it temporarily. The default, if adding to the watchlist at all, is to add it permanently as before. Certes (talk) 21:40, 2 December 2020 (UTC)
(after reading the question properly) I recall seeing the fuzzy dropdown on a mockup somewhere in advance of this release. Perhaps you were being shown that mockup rather than the actual dropdown for some reason? Certes (talk) 21:45, 2 December 2020 (UTC)
@Redrose64: It was a configuration change, which are done through backport windows and not the deployment train. The watchlist expiry code has been in MediaWiki for months, it just wasn't enabled here on English Wikipedia until December 1.
@Vchimpanzee: By "fuzzy" you mean the dropdown is in the disabled state? It was done this way for accessibility. If people don't like it, you can try adding this CSS to Special:MyPage/common.css:
There's a page that's indef semi-protected. I'd like to bump that up to ECP for a while, but leave it so when the ECP expires, it falls back to semi. Is such a thing possible? -- RoySmith(talk)18:38, 30 November 2020 (UTC)
I was going to say, an adminbot could do this pretty easily. The log of a changed entry would probably need to indicate that it is intended the protection be restored to previous on expiry (distinct from just expiring to no protection), but otherwise doesn't seem too complicated. Issue with contributing a patch to the core is, well, more effort, more code, more time to write and even more time for reviews. ProcrastinatingReader (talk) 20:14, 30 November 2020 (UTC)
Oh, this should be a feature of MediaWiki core. But the bot could be a stop-gap, instead of waiting another 8 years. Even a non-admin bot could make requests to WP:RFPP, say a few hours in advance. Suffusion of Yellow (talk) 20:31, 30 November 2020 (UTC)
The rule-set for an adminbot would need to be very clear; a bot isn't going to read the contextual meaning of an log summary to know if something is being changed for now or for good (e.g. an admin may really WANT the protection to fully expire in the future) -- or we need clear information sent to admins to know what to expect. I think this is a good idea, but mostly think it should be fixed server-side. — xaosfluxTalk20:37, 30 November 2020 (UTC)
Yeah, it'd need to be signified using a standard format, and it should be opt-in (ie the default is still expiring) I think, to avoid surprise. Maybe starting a log with {{botprot}} or something to opt-in for the bot, or more specifically with something like {{onexpire|semi|indef}}. Would need to fail gracefully for messed up log entries, which imo is feasible. Alternatively, could have it so temporary protections auto revert to their previous one on expiry, so to have an expiry to no protection an admin would have to first remove the protection, then set the temporary one. I don't think this flow is illogical, but with 1,122 admins such a change will likely cause confusion.
For some reason, it is possible to have both semiprotection and pending changes enabled at the same time, with different expiries. But you can't have both semi and full protection enabled. EdJohnston (talk) 21:13, 30 November 2020 (UTC)
Yes, protection and PC are different, a page may have both simultaneously and they may have different durations. Semi-protection and full protection are merely different levels of a single scale that presently has five levels, ranging from unprotected through semi, extended-confirmed and template up to full. A page always occupies exactly one level on this scale for editing, one for moving and (if it's a file) one for upload. When an expiry time is reached, that prot drops right back to unprotected, regardless of the page's prot history. --Redrose64 🌹 (talk) 22:35, 30 November 2020 (UTC)
I like the idea of a bot, but I think a MVP approach might be best. If I could get an alert (either on my talk page, or something that pops up in my notification feed) that the protection was going to expire soon (say, in the next 24 hours), then I could manually extend or reset it. Or, I like the idea of adding an request to WP:RFPP, that would work too. Keeping a human in the loop, and avoiding giving a bot admin rights both reduce the potential for error.
The problem with merely informong the protecting admin is that this person may have ceased to be an admin, or else be inactive during these last 24 hours. Either the bot should restore the old protection, or it should leave a note in a place like RFPP for all admins who handle protection. 109.186.211.111 (talk) 20:56, 1 December 2020 (UTC)
There is, however, the question of who is now 'responsible' for the protection (for the purpose of accountability etc, and, for Arbitration enforcement (DS) protections, is the "enforcing administrator")? The original protecting admin who intended for an indef protection, the bot (if done auto), or the admin who reprotects an expiring protection at the previous indef level? Note this is not a problem with, for example, PC vis-a-vis page protection, since they can run concurrently. Xaosflux any thoughts? ProcrastinatingReader (talk) 13:09, 3 December 2020 (UTC)
@ProcrastinatingReader: just like any other admin action, whoever actually does the action. "Bots" are not people, they are just alternate accounts of other people - so the admin-cum-botoperator would natively be accountable for the action. Very very very explicit overrides for this could be allowed if the community wants to support it, and if used the bot should identify this on each use - for example by putting in the log entry something like "Restoring indef protection applied by User:Example". Any 'errors' made by the bot though would still belong to the botop. — xaosfluxTalk16:29, 3 December 2020 (UTC)
@Xaosflux how about the manual case, eg when expiring protections are notified to RFPP? Does the admin who takes the technical action of restoring the old indef protection now become the "enforcing administrator" / one accountable for the protection? Silly question perhaps, but if so wouldn't this have implications on admins willing to technically restore protections? ProcrastinatingReader (talk) 16:35, 3 December 2020 (UTC)
@ProcrastinatingReader: I don't follow DS processes much, but if I had to guess each admin touching would still be responsible for what they did (including the admin that decided to allow a protection to expire in the future - as noone should ever assume someone else will make a future action). — xaosfluxTalk18:04, 3 December 2020 (UTC)
Straw poll
Should an adminbot a bot monitor changes in page protection, keep track of "stacked" protection, and report cases where previous protection needs to be restored to WP:RFPP? If this gets a lot of support and few opposes, then it can go to a formal proposal. davidwr/(talk)/(contribs) 21:15, 1 December 2020 (UTC) Updated to replace "an adminbot" with "a bot" davidwr/(talk)/(contribs) 22:15, 1 December 2020 (UTC)
FWIW, it doesn't need to be an adminbot just to report cases to RFPP, only if it wants to restore protections itself. I have a little concern with just reporting to RFPP: that board is backlogged not too infrequently. Between this proposal, and the one on AN, I'm not sure the extra workload to RFPP will be coped with. I mean, just look at WP:RFPP right now... ProcrastinatingReader (talk) 21:29, 1 December 2020 (UTC)
Then I'm not sure this needs a proposal. Any bot could maintain a log in its userspace (without approval) which is transcluded into RFPP. This only really needs a discussion if we're automating it. ProcrastinatingReader (talk) 09:42, 2 December 2020 (UTC)
Support. The way you've got it worded wanders a bit too much into implementation details, but basically yes to the gist of it, which is an automated process for notifying RFPP. -- RoySmith(talk)21:28, 1 December 2020 (UTC)
How many of these per week are we talking about? If it is a few, RFPP is fine. If it is a lot, perhaps a sub page off of RFPP? --Guy Macon (talk) 22:25, 1 December 2020 (UTC)
That question suggests an alternative solution which would not require a bot. RfPP could transclude a subpage of RfPP named for the current date (so a different subpage gets transcluded each day, like RfD etc.) Then an admin could apply a temporary protection indefinitely, also adding an entry to the RfPP subpage named for the appropriate future date requesting that the previous protection be restored on that date. Certes (talk) 22:39, 1 December 2020 (UTC)
Support as someone who patrols RFPP this would be very useful. From my experience, it's not a high rate so it wouldn't flood RFPP, but it's frequent enough that it is a problem worth fixing. — Wug·a·po·des23:11, 1 December 2020 (UTC)
Note if there is a bot, there'll be even less motivation for developers to resolve the trouble ticket. Of course, after waiting for eight years, arguably a fix is never coming. isaacl (talk) 05:03, 2 December 2020 (UTC)
A developer solution would provide the capability in a way that didn't require further administrator action for each case, and the resources required to keep the bot running and updated. However it is true enough that this still might be preferable than lowering the priority of other developer tasks. isaacl (talk) 14:18, 2 December 2020 (UTC)
No concerns if a bot wants to maintain a report page, suggest it doesn't flood RfPP directly, and use a transcluded page like AIV does. — xaosfluxTalk15:17, 2 December 2020 (UTC)
Template:Tooltip, Template:Hover_title, and Template:Abbr
{{Abbr}} (a wrapper for <abbr>...</abbr>) has long been abused for non-abbreviation markup (against the HTML specs).
We had a template, {{Tooltip}}, with <span>...</span> for non-abbreviation use, but it was "merged" (not really) and redirected to {{Abbr}}.
The redir was then deprecated (for the reason mentioned above), but the community ignored the deprecation.
In the interim, {{Hover title}} was created to do the same thing, but with backwards parameters (and some additional features).
Both the {{Tooltip}} then-redirect and {{Hover title}} template have been transcluded in tens of thousands of articles, mainly via infoboxes and other templates.
I created a new {{Tooltip}} template, with all the features of {{Hover title}} but preserving the {{Abbr}} parameter order (to not break deployed translcusions).
The TfM linked above would merge away {{Hover title}}, but it's going to require flipping the |1= and |2= parameters of its extant instances.
Oh, and the documentation would need updating after merger, of course.
<includeonly>{{#invoke:crucialwork|main|param=3}}</includeonly><noinclude>
== About ==
This tempalte is extremely useful.
[[Category:Templates]]
</noinclude>
used on many pages. I change at some time "param=3" to "param=4". This will presumably cause a large amount of work for the server having to rebrew all pages using my template. One month later I notice the typo and change "tempalte" to "template" in the built-in documentation. The question: Will this edit cause extra work for the server, or is it smart enough to notice that there is no need in this case? Taylor 49 (talk) 08:50, 29 November 2020 (UTC)
(original message was split by Taylor 49) Not that you should really worry about this. The developers generally say not to worry about performance/storage. On a more practical note, if you use a separate documentation page (User:Alexis Reggae/sandbox1/doc) an update to the documentation appears to reparse the template page (sandbox1), but not the page that uses said template. (sandbox2) — Alexis Jazz (talk or ping me) 13:42, 29 November 2020 (UTC)
To answer the question directly, yes, any change to a template page will cause the using pages to be updated. This is part of the reason why widely-used templates have a high level of page protection, and why most templates have a separate documentation page which is transcluded onto the template page. --Izno (talk) 16:22, 29 November 2020 (UTC)
@Alexis Reggae: The above template is maybe stupid. It is the simplest case to expose the problem. The question is not stupid. I cannot test editing a template used on a huge number of pages (crucial detail in the question) and recheck whether they are getting updated earlier just due to my edit. Taylor 49 (talk) 04:21, 30 November 2020 (UTC)
Thanks. Thus the WikiMedia software could be "smart" (remove the "noinclude" part, compare, if identical then do NOT rebrew all pages using the template) but currently isn't. Taylor 49 (talk) 04:21, 30 November 2020 (UTC)
@Taylor 49: You were asking if the edit would "cause extra work for the server, or is it smart enough" and my answer was that the server is "somewhat stupid". I didn't comment on how clever the question was. There is no difference (that I know of) between a template that is used on one page (like User:Alexis Reggae/sandbox1 which despite being in user space is included like a template on User:Alexis Reggae/sandbox2) or on a huge number of pages. Either the page(s) that use the template get reparsed, or they don't. There may be some optimization in planning when and in what order pages get reparsed, not sure about that, but the end result is the same. — Alexis Jazz (talk or ping me) 09:32, 30 November 2020 (UTC)
@Taylor 49: See also WP:SLOW. For "noinclude" this may indeed be possible, but as documentation is often transcluded the actual gain may not be as much as you think. @Izno: True, though the reason for protection in those cases is, I think, not so much server performance but mostly that updating all pages takes time. (in addition to the visibility of widely-used templates) If a widely-used template gets vandalized, it's less than ideal if it takes half an hour for all the vandalism to actually disappear after rollback because reparsing isn't instant. — Alexis Jazz (talk or ping me) 09:32, 30 November 2020 (UTC)
Thanks and sorry that I misunderstood about "who" was convicted as stupid, it is the server or actually WikiMedia software. There used to be many arguments for outsourcing template docs to a subpage, but most of them have gone with the advent of LUA and wikidata. This seems to be the last one that is left. Thanks for testing but I suspect that there could be a difference between one use and many uses of a template, and pages with {{#time could be privileged, especially if viewed frequently. Taylor 49 (talk) 12:43, 30 November 2020 (UTC)
@Taylor 49: I think template docs should be transcluded to prevent accidentally breaking a template while making a change to the documentation and to allow users without special privileges to update documentation of protected templates. How do LUA and Wikidata resolve these issues? — Alexis Jazz (talk or ping me) 13:45, 30 November 2020 (UTC)
@Alexis Reggae: 1. Most formerly complex templates nowadays consist of nothing beyond a single call to a module. 2. Interwiki bots adding and removing interwiki links all the time have died out. Taylor 49 (talk) 19:41, 30 November 2020 (UTC)
@Taylor 49: 1. Not true, not even remotely. 2. How does that help users without special privileges to contribute to template documentation of protected templates? — Alexis Jazz (talk or ping me) 20:28, 30 November 2020 (UTC)
@Alexis Reggae: 1. [2][3] Not all but many. A problem that I have observed on some wikies is that users add parameters to the docs subpage falsely hoping to add them to the template itself that way. 2. Not at all of course. At least, if same or equivalent template is added to some other wiki then I can link it via wikidata, without editing the (protected) template itself nor the docs subpage. Taylor 49 (talk) 11:01, 4 December 2020 (UTC)
Testing a technical function
{{Cross-wiki user}} I am testing out a new template, at right, but unfortunately the documentation on how the {{REVISIONUSER}} magic word works with template transclusions is a bit lacking, so I need to do an experiment. Could the first person to see this please edit this page with any comment, I would greatly appreciate it. VanIsaacWScont16:48, 3 December 2020 (UTC)
Talkpage technical error at Talk:Cultural Marxism conspiracy theory
Resolved
Can someone please have a look at this? I was having a talkpage technical error at Talk:Cultural Marxism conspiracy theory? Relevant version. I have fixed the issue in the current version, but I think it will mess-up the archives when we get to that time. (please mention me on reply; thanks!) --Emir of Wikipedia (talk) 21:08, 2 November 2020 (UTC)
@Emir of Wikipedia: I'm not sure I'm seeing what you're looking at or what you might have done to fix it, but it looks like there's a quote box with an unclosed ref tag in it. Look for <ref name="braune2">. I think this code would produce a "ref invoked but not defined" error on the talk page, but I'm not sure if that error would render inside the quote box, or if it would just mess things up unpredictably. You need to close the tag, otherwise Wikipedia thinks everything after it is part of the reference. You can add a / to the end of the tag (<ref name="braune2" />), but then you also need to define the reference and add a {{reflist-talk}} somewhere in the section. It might be better to pull the URL of the reference from the article and replace the ref tag with an external link on the talk page instead. Ivanvector (Talk/Edits) 17:36, 4 November 2020 (UTC)
Is there any way to get {{For loop}} to go through every member of a category? E.g. if I have Category:Colors, I'd like the end result to be equivalent to
That's too bad; it could be quite helpful for some list pages. If anyone else knows how to get that functionality, please lmk. {{u|Sdkb}}talk00:00, 4 December 2020 (UTC)
Xaosflux, long-term, the ultimate goal is to reduce the redundancy between lists and categories. For now, I'm looking for a fairly simple list page (perhaps a table where each column could be sourced from Wikidata) to see if it'd be possible to convert it to a for loop that'd just list out the members of the category.
I'm not sure the category tree could work, since I don't know there's a way to build any code around it (e.g. if I want another column in the table to list the hex code). {{u|Sdkb}}talk18:31, 4 December 2020 (UTC)
Hmm, so it sounds like the developers might need to change something before this would become possible. Should I file a phab ticket? {{u|Sdkb}}talk19:56, 4 December 2020 (UTC)
Is there an easy and non-server-annoying insource: search that would find {{cite web}} specifically if it's present on the page after the ==External links== section heading? WhatamIdoing (talk) 02:36, 4 December 2020 (UTC)
This search should get you started. Change the prefix letter once all of the "A" articles are done. There are probably other patterns to search for, but this seems to find plenty of hits to keep people busy. – Jonesey95 (talk) 03:03, 4 December 2020 (UTC)
WPCheck is using an dump from 1.1.2017, but it uses daily updates which where broken (due to replication failure) when the link was added in November 2017.--Snaevar (talk) 15:15, 4 December 2020 (UTC)
I'm pretty sure that the OP is looking for a way to implement this RFC prohibiting cite templates in external links. Ref tags are a different thing. – Jonesey95 (talk) 15:30, 4 December 2020 (UTC)
Yes, plus I should be able to modify your search string to find problems created by User:InternetArchiveBot, which has been replacing ==External links== with links to the Internet Archive (including some that are still working – he's working on it). WhatamIdoing (talk) 19:24, 5 December 2020 (UTC)
@Jeff G.: Where no author has been entered, the field won't be populated. (in some cases the copyright owner could be copied/moved to author, but this should be evaluated before doing so) But at least it would be possible to enter an author. — Alexis Jazz (talk or ping me) 10:40, 6 December 2020 (UTC)
Thing is that NFC does not ask users to identify the author/copyright owner, only the source, though things like the Upload Helper, etc. will guide users to use rationale 2 to include this information. So I'm not sure if that category is as necessary compared to the equivalent for license or source which are required elements. (Same with the description category, that's not a required field from an NFCC standpoint) --Masem (t) 18:34, 5 December 2020 (UTC)
@Masem: I think I use Upload Helper. What's "NFC"? (other than an abbreviation for non-free content) — Alexis Jazz (talk or ping me) 10:40, 6 December 2020 (UTC)
I have three problems with getting {{Chem}} to work properly at Fumarole mineral. Firstly, the formula for rhabdoborite (Mg12V,Mo,W4/3O6[(BO3)6–x(PO4)xF2–x]) isn't displaying properly. Second, how do I make the formula for lesyukite recognize the dot of Water of crystallization and third, how is the beta modification indicate in the formula for parageorgbokiite? Jo-Jo Eumerus (talk) 15:03, 6 December 2020 (UTC)
I think I have taken care of all of these categorization problems. There were multiple template edits required. Drop a note on my talk page if you need further help. – Jonesey95 (talk) 15:27, 6 December 2020 (UTC)
I am asking about the technical feasibility. We need to establish that we can do it before discussing whether we should do it.
Of course the editor could bypass this by removing the CSS or by using cut and past to accomplish the revert, but it would make edit warring harder to do. Some CSS that was proposed:
This could be a useful tool in an administrator's toolbox: allowing someone who is edit warring to resume editing but making it harder for them to edit war again.
First question: will this work? Is there some better choice of CSS that would be better?
It works but you can literally inspect element and show it again. Can use JS to delete the elements entirely but still, bypassable. Those solutions only work for good-faith editors who just suffer from poor impulse control. We could always make some kind of "Revert probation" enforced by edit filters/flagged up by bots (kinda like this), but I mean, I'm not sure it's widespread enough an issue? ProcrastinatingReader (talk) 00:42, 5 December 2020 (UTC)
Well, for starters, only int-admins can edit other user's css, not sysops, and it would then have to be protected to prevent them from editing it. Second, as PR notes above, any editor (using a browser capable of inspecting elements, so not, e.g. mobile safari) can reenable the buttons, or even later modify that css via javascript. It also won't prevent any API access. I think it's just an inappropriate use of the intadmin and sysop toolset. ~ Amory(u • t • c)00:55, 5 December 2020 (UTC)
I'm going to go one step further and suggest that using ad-hoc technical workarounds to bypass a persistent behavioral issue is counterproductive. This user needs to learn how to control themselves. Having the community remove a couple of buttons is exactly the opposite of that. If they feel the need to remove the buttons themselves, they can access their .css at any time. The fact that they haven't chosen to do so themselves already is part of the problem. I would have no issue with posting this code at the edit warring board for editors to stop themselves. But it should be a tool so that editors can demonstrate their own change in behavior, not a half-measure to use instead of the block that is really called for. VanIsaacWScont01:34, 5 December 2020 (UTC)
I am not sure exactly what part of "I am asking about the technical feasibility. We need to establish that we can do it before discussing whether we should do it." is so hard to understand, so let's assume that I am only talking about offering the css for anyone to use if they wish to remind themselves not to revert. Now can we discuss whether or not this would actually work?? --Guy Macon (talk) 01:56, 5 December 2020 (UTC)
We can make it very annoying to revert a change; CSS is one way to do that; it won't work on IPs; and blocking is the only thing that'll actually prevent reverts. Enterprisey (talk!) 03:26, 5 December 2020 (UTC)
(edit conflict) I agree with Amorymeltzer that this isn't really the sort of business we should be dealing with as int-admins for a behavioral issue. Also, regarding the use of filters suggested by ProcrastinatingReader having to maintain a filter with named users as an argument is a recipe for disaster. Now as far as the good-faith/help someone on 0RR not trip themselves as easily go - if someone wants to make an opt-in userscript or the like that editors could opt-in to, sure go for it. Keep in mind, that if they having competing scripts/gadgets with similar functionality (e.g. Twinkle) or use tools like Huggle - they would need to stop using those for the most part too. As far as the technical aspects go, there are all the caveats/conflicts I just mentioned, also scripts can not be applied to specific or subsets of logged out users. — xaosfluxTalk02:01, 5 December 2020 (UTC)
I'd suggest maybe allowing the partial blocking of the "undo" and "rollback" functions. That way, if a user wants to revert, they will have to do so manually, but we could potentially block editing old revisions as well. Aasim (talk) 08:23, 5 December 2020 (UTC)
To answer the question without the waxing on whether we should, yes, this method would help to enforce a ban on reverting. There are several caveats, most mentioned above:
The CSS file would need to be protected in addition to its innate protection.
This method require an interface administrator to make the modification to the user CSS file.
In addition to the implementation of the removal, these are the caveats that remain that a user can work around them:
The user may use a standard web browser development tool to remove the CSS.
The user may use either device JavaScript, device CSS, or Wikipedia-site user JavaScript to remove the CSS.
This method does not prevent a manual reversion.
This method does not prevent API access either with a custom tool or something like Huggle.
As above, this method cannot be used to stop unregistered editors from employing undo.
You can decide if that makes this tool a reasonable one.
Thanks! It looks like the answer to my "We need to establish that we can do it before discussing whether we should do it" question is that we can't do any of it (an interface administrator can), if we were able to do it it would not help with IP edits (the obvious thing to try if you find that can't revert is to log off), and it would work poorly in all other situations. Thus I conclude that there is zero point in discussing whether we should do it and that all the words written above on whether we should do it before establishing that we can do it were a waste of everyone's time. --Guy Macon (talk) 17:08, 5 December 2020 (UTC)
And here we get to the point of my above contribution, which was that this sort of thing should be available for editors to use for themselves. Guy's conclusion above is based on framing the technical question on the basis of a policy assumption. But ignoring that policy assumption which hasn't panned out, we still have a valid technical question: does the code actually work at removing undo functionality, even if it can't be imposed on a user? VanIsaacWScont17:36, 5 December 2020 (UTC)
I almost forgot the most important follow-up: If this were put in a CSS package for people to transclude, is there any other code which should be added to remove undo/revert functionality from other tools? VanIsaacWScont17:40, 5 December 2020 (UTC)
does the code actually work at removing undo functionality I thought that was abundantly clear. There is no reason one could not make a small script as some page e.g. a subpage of WP:User scripts. Yes, there are probably other rules you could add, but those being the most common for users without rollback off the hand, I doubt anyone is going to go hunting and pecking for others. --Izno (talk) 19:38, 5 December 2020 (UTC)
@Izno: as far as the literal technical functionality, no this code doesn't remove the actual capability of an editor to actually use the undo parameter (which is really just a reversion with a automatic edit summary) - it merely asks their web browser to not display the link to the built in button. They could still technically pass the undoafter and undo parameters to a request and perform it. @Vanisaac: Undo isn't a permission and an anicient request to make it one was declined all the way back to at least 2008 in phab:T16616 and anything you do with CSS to "remove it" is just a cosmetic hack. — xaosfluxTalk20:37, 5 December 2020 (UTC)
Well at least from my perspective, the cosmetic is what we'd be looking for. Anyone who isn't blocked from editing a page can technically revert an edit after all. I am thinking in terms of a tool for say users under a 0RR/1RR to get rid of the temptation. So the question remains: is the equivalent of an undo/rollback button found in any other tools, and is there code that could disable it from showing up? VanIsaacWScont22:45, 5 December 2020 (UTC)
@Vanisaac: for self-censoring, trying to track down all non-built in tools is going to be tough (you could start with those in Category:Wikipedia counter-vandalism tools). Hiding "undo" (provided by the interface) and not giving someone rollback group will take care of most native things - someone self-censoring themselves should not load any antivandalism tools or utilities, and should just stay away from vandalism clean up all together as it almost always involves reverting. This means tools like Twinkle/Redwarn/LAVT/etc. -- instead of trying to hunt down and cripple all such tools, such editors should be told to just not install or use any of them. — xaosfluxTalk00:26, 6 December 2020 (UTC)
I didn't think the Levivich suffered from undo-itis ;). I think this can probably be solved by using a smarter JavaScript solution for this (rather than just CSS) - and deleting the undo elements entirely also makes them harder to restore. Responding to comments above, sure this probably doesn't make sense as something forced to enforce 0RR/1RR, but as an aid I think it's pretty helpful to remove the temptation of the button. "Annoyance" means it's doing its job, heh. By the way, I'm not sure the "they can just log out" concerns above are problems, it would be pretty blatant. ProcrastinatingReader (talk) 10:16, 6 December 2020 (UTC)
@ProcrastinatingReader: Levivich suffers from curiosity :-) I'm sure you're right about JS; the CSS is a hack. But while we're on the subject, does .mw-diff-undo, .mw-history-undo, [id^="tw-revert"] { display: none; } works just as well as .mw-diff-undo, .mw-history-undo, #tw-revert, [id^="tw-revert-to-"] { display: none; }, without the #tw-revert part?
I would think it's a bit nuts to consider doing anything like this as a sanction imposed by others. But as a self-imposed "sanction", I think it could be useful to some. It's already stopped me from making one revert I otherwise would have made, simply because it wasn't "important enough" to trouble myself with a manual revert. (Someone else made the revert anyway later.) I'm not sure if that's for the better or worse in that particular case, but for some people I could see this helping slow themselves down.
The self-revert thing is a problem, though. I made a mistaken edit and went to self-revert and there was no undo button. Manually self-reverting would have been tricky and might have caused me to introduce even more mistakes. Luckily I still have the restore script installed so I used that. But if this code were used as a self-imposed sanction, it wouldn't be very helpful if it also removed the ability to self-revert. I don't know if JS can come up with a solution where it only hides the undo button for other people's edits. (Is that even possible?) It's probably not worth the time to figure out. Levivichharass/hound18:05, 6 December 2020 (UTC)
Yeah, that should work the same. JS could do this since there's a couple easy ways to identify if self-reverting. One way: on top of the CSS, using $('a[href^="/wiki/User:Levivich"]').closest('td, li').find(".mw-history-undo, .mw-diff-undo").show(); should do it. Works when ran in the browser's JavaScript console, so should probably work in common.js too. ProcrastinatingReader (talk) 18:20, 6 December 2020 (UTC)
If you log out and, for example, go to the page Wikipedia:Index.php (or any other subject page matching the title blacklist,) it will say that the page is template-protected from creation so that only template editors can create it. However, an attempt to load Special:EditPage/Wikipedia:Index.php as an IP address, will say that creation of the page is currently restricted to administrators because the title matches an entry on the local or global blacklists. So, are blacklisted titles editable by template editors, or by admins only? This message is confusing to readers and needs fixing. --Rqkp (talk) 01:30, 7 December 2020 (UTC)
Hi, I've noticed that the category is ordered from 1 BC to 1000 BC and then from 1 AD to the current year. I think it would be more fitting for years to be ordered always from oldest to newest but I wanted to ask for opinions before making a change to Template:BirthyrBC. So
I don't think it would change anything. As far as I can tell, the "-" character just acts to sort the BC articles under "-" in the category. It does not appear to be acting as a minus sign. – Jonesey95 (talk) 01:40, 7 December 2020 (UTC)
Would it act as a minus sign if the whole sort code were a valid number? For example, "-2 BC" comes after "-1 BC" alphanumerically, but would "-2" be sorted before "-1" numerically? If not, does "–2" v. "–1" or any other variant of minus work? Category sorting has changed recently, and the answers may differ from what we knew. Certes (talk) 12:05, 7 December 2020 (UTC)
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
You can now put pages on your watchlist for a limited period of time. Some wikis already had this function. [6][7]
Changes later this week
Information from Wikidata that is used on a wiki page can be shown in recent changes and watchlists on a Wikimedia wiki. To see this you need to turn on showing Wikidata edits in your watchlist in the preferences. Changes to the Wikidata description in the language of a Wikimedia wiki will then be shown in recent changes and watchlists. This will not show edits to languages that are not relevant to your wiki. [8][9]
The new version of MediaWiki will be on test wikis and MediaWiki.org from 8 December. It will be on non-Wikipedia wikis and some Wikipedias from 9 December. It will be on all wikis from 10 December (calendar).
There is apparently a formatting problem at the Halloween article, the consequence being to display two widely-separated references as being mashed together and to make the intervening 121 refs fail to display, aside form their superscript numbers. I can't spot the flaw causing this; can somebody else have a look? Further details at Talk:Halloween#References. Mutt Lunker (talk) 18:33, 7 December 2020 (UTC)
Categories will display as being in a category and templates will show up in whatlinkshere, but otherwise no. Renaming the one page is easier. --Izno (talk) 21:01, 7 December 2020 (UTC)
Can't access contributions for myself or other editors - get this error instead
Not really sure where best place, so when in doubt post on VP.
It would be awesome if the desktop footer was rearranged so that "mobile" was the last item on the footer list, instead of cookie statement.
For certain mobile edits, switching to desktop mode is neccessary and it's nice and simple on the short mobile footer.
Once in desktop, additional links are added and inevitably I click the cookie statement when looking to switch back to mobile mode.
Just seems like it would better to have mobile be the last link and reduce accidental traffic to cookie statement. (Unless of course It's just me, then well carry on)Slywriter (talk) 21:43, 7 December 2020 (UTC)
@Slywriter: This is in the MediaWiki software, more specifically, the code for the skin. As such, there's nothing that we can do here - you would need to file a change request at phab:. When doing so, you should provide a very strong case for change, since it would not just affect English Wikipedia, or even Wikipedia in other languages - but every single project that uses the MediaWiki software. --Redrose64 🌹 (talk) 22:50, 7 December 2020 (UTC)
@Redrose64: Boo. Had a feeling it was mediawiki. Was hoping that by some chance the footer had local variables. Alright, I'll consider taking it over there. It's annoying but not broken. Thanks. Slywriter (talk) 23:21, 7 December 2020 (UTC)
@ProcrastinatingReader: I keep JS and other scripts off except for web pages that need it to display the content I want to see. It cuts out a lot of advertising and other clutter that I do NOT want to see. More importantly, it saves my computer when I accidently *coughtyposcough* go to a web site that's been hijacked by a script-dependent malware. davidwr/(talk)/(contribs) 20:13, 4 December 2020 (UTC)
I would tens to judge notifications on Wikipedia to hit none of the problem areas that would cause one to disable JavaScript. While I agree all content should be available without, I similarly think the OP made his bed and now he needs to deal with the consequences. --Izno (talk) 21:20, 4 December 2020 (UTC)
Agreed; I generally browse the rest of the Internet without JavaScript, but none of the reasons for that choice apply to Wikipedia. Enterprisey (talk!) 02:12, 6 December 2020 (UTC)
Ah, I see, the bug is the lack of a graceful failure combined with showing the editor the notification count in the first place. The bug is also from 2016 which means it probably will never be specifically addressed. What MIGHT happen is that the next time there is a wholesale change to that part of the code, the developers will "solve" the problem one way or the other, possibly by completely hiding the notifications for users who turn scripting or other "technology the web site depends on" off. Oh by the way did I mention that when I view Wikipedia with the classic Line Mode Browser on my VT100 terminal, I don't see any pictures (no, I don't still use a VT100, but someone might)? Oh well, I'll live. On the plus side, I don't have to worry about shock-value content showing up. davidwr/(talk)/(contribs) 21:29, 4 December 2020 (UTC)
Not to many have it turned off anymore [10] as now basic things use Java even our login button. Would like to ask how they login with no Java?--Moxy🍁02:23, 6 December 2020 (UTC)
Logging in uses a regular HTML form, for which JavaScript is not required. You fill in the fields, click a Log in button, and the form is submitted. It's no different from editing a page. --Redrose64 🌹 (talk) 21:42, 6 December 2020 (UTC)
@Brandmeister: Try using the "Include redirects" option [11]. In most all cases, old locations of pages will redirect to the new one, so simply including redirects gives the full picture. You can also use the Redirect Views tool. The issue you point out with the move log is indeed confusing, but expected (a page titled 2020 Nagorno-Karabakh war has never been moved to another title, hence why its move log is empty). — MusikAnimaltalk19:30, 8 December 2020 (UTC)
Is there anything in preferences that controls the use of the up/down left/right keys
I can no longer go up and down the screen in Chrome with my up and down keys. I'm probably wrong, but I was fiddling with Preferences recently and thought I saw something relevant and clicked on it, but can't find it again. Thanks. Doug Wellertalk15:18, 6 December 2020 (UTC)
I don't know about your system, but in general, up/down will only scroll the browser window if that window has the focus. For example, if you click in the search box in the sidebar, pressing up/down will not scroll the main content window on this page. Perhaps there is a clue there. Johnuniq (talk) 22:13, 6 December 2020 (UTC)
Wikipedia:Tools/Navigation_popups (NavPops) are no longer showing the detail in a citation. Formerly, as recently as the past few days, I could float my cursor over a citation[1] and read it without scrolling down to the references section. I use this constantly to make sure my citation is formatted properly, but the NavPops behavior has now changed. The Popup now contains only the lead text (e.g. --NEWSECTIONLINK--), but not the citation detail when I move my cursor over a reference number (e.g. This is a sample citation.). Something has changed. --Ancheta Wis (talk | contribs)13:11, 9 December 2020 (UTC)
Expected Behavior Of NavPops
Here is some additional information. The change must have occurred in the last 8 hours: I have an article in one tab from hours ago, which exhibits the old, expected NavPops behavior (detailed citation in the popup). But the same article, with newer edits now displays the lead text in the popup. --Ancheta Wis (talk | contribs)13:50, 9 December 2020 (UTC)
I feel bad — that's one of my favorite features too! I was fixing two bugs, but I think just one was causing this regression, so I've restored just the other one for now. ~ Amory(u • t • c)13:57, 9 December 2020 (UTC)
I’ve come across a couple of new editors trying to add spaces around negative signs in LaTeX. The complaint is that the negative signs aren’t rendering without increasing the zoom level on the browser. Sure enough, I’ve had to increase the zoom level on my browser to see some negative signs. LaTeX used to render reliably at 110% (Maybe lower, but I’ve used 110% for years). Now I need to go to 150% in some cases. Has the LaTeX engine made some new assumptions about display resolution. My monitor is 1920 x 1080. Maybe it is time for an upgrade. Constant314 (talk) 11:32, 9 December 2020 (UTC)
I don't know how to ask this question in a clear way.
I restored Draft:Cedar Point peninsula (Ohio) under WP:REFUND, but forgot to indicate that in my restoration summary, so now the log indicates that I restored it, but not why. Is there a way for any of you technical types to change this? I'm guess not, but I thought I'd check.
Change of formatting in template for no obvious reason
This edit had an unexpected effect on the template title format. The position of the VTE links changed from left to centre. Can anyone see a reason why this would happen? Cheers, · · · Peter Southwood(talk): 18:35, 9 December 2020 (UTC)
The VTE links are on the left side in both revisions in my browser. Is this happening when the template is transcluded in an article? – Jonesey95 (talk) 19:15, 9 December 2020 (UTC)
What version is your Chrome, and what OS? Chrome 87 (latest) doesn't have this problem even on Windows 7. --George Ho (talk) 16:35, 10 December 2020 (UTC); expanded, 16:37, 10 December 2020 (UTC)
Building tables with modules
A month ago I posted a question at the Help desk, here, but no one could help. If I want to build a table where identical cells appear more than once, can I use a module so as to define the coding of the cell type once and to invoke it as many times as needed? I know next to nothing about modules, and the help pages Help:Lua and Help:Lua for beginners are too general and don't provide that exact information. --Theurgist (talk) 12:45, 7 December 2020 (UTC)
It is possible with a normal template and WP:TemplateStyles (however, not for arbitrary colors). However, it is likely to be deleted if one cannot show use on more than one page. What is your actual use case? --Izno (talk) 13:39, 7 December 2020 (UTC)
My actual use case includes background colour, fond colour, and boldface.
@Theurgist: That's... not a use case. That's what you want to do, not why you want to do it. Why do you think this is necessary? In particular, provide the page and the table of interest. --Izno (talk) 22:52, 7 December 2020 (UTC)
@Izno: Sorry for misunderstanding. It's odd that FIFA World Cup qualification doesn't represent the actual performances of the teams, so I wanted to add this table to it, but doing so would more than double the size in bytes of the already large page. I'm aware that the size can be reduced to some extent by merging identical cells into one, but I find this an inelegant solution that doesn't contribute to understanding. --Theurgist (talk) 09:17, 8 December 2020 (UTC)
@Theurgist: It's used 216 times in that page and gives big advantages so hopefully it stays but you never know. I don't think we have a guideline about templates used many times in a single article. PrimeHunter (talk) 21:57, 10 December 2020 (UTC)
Hello, I'm one of the maintainers of User:WP_1.0_bot. I'm reposting this question which I originally posted on the general help desk. We recently had an issue where the bot's password was leaked to Github and had to be reset. That was completed successfully. However, when I tried to use the new password in the bot's login code, I messed up the login username. I've since fixed that, but now the bot's machine is facing this message when trying to log in:
'reason', 'You have made too many recent login attempts. Please wait 2 days before trying again.'
The bot has already been offline and unable to edit Wikipedia for several days, and I'm posting here for help because: 1) I don't want it to be offline any longer than it has to, since it serves a critical function in the WikiProject community and 2) I want to get on track to receive help in case it doesn't automatically unlock in the 2 day period.
Yup, RoySmith is right, was just going to comment here :). I reset the throttle for your bot account, hopefully it will work (those requests aren't common :)). Thanks for your bot-service, --Martin Urbanec (talk) 23:44, 10 December 2020 (UTC)
I consulted a number of edit histories for pages on the Dutch, German and French Wikipedias, and wrote down some specific edits with their timestamps. Apparently, when I consult the same pages again today, these logged times seem to have shifted by an hour. E.g. an edit that I noted down as happening at '07:00' last summer, would be registered as '06:00' Today. Can anyone explain why this could be the case? I assume it has something to do with timezones/offsets? Willaerttom (talk) 16:00, 10 December 2020 (UTC) — Preceding unsigned comment added by Willaerttom (talk • contribs)
Yes, the edit will have occurred at 06:00 as measured in your current timezone, which is 07:00 in the timezone you had in summer. Certes (talk) 20:18, 10 December 2020 (UTC)
Also I see your user account was created yesterday (welcome!), so last summer you will have viewed the edits whilst logged out (or possibly using a different account). Certes (talk) 20:23, 10 December 2020 (UTC)
Thank you, and yes, I am quite new. When logged in, my timezone is set to the Wiki standard, so perhaps this should not make a difference? In any case, I am just wondering if timestamps are always shown relative to the timezone of the user or not. Could this not be problematic in certain cases? For example, an edit that occurred close to midnight might be shown as occurring on the next day depending on the timezone of who is viewing the edit history? Also, from the point of say historical research, there is no 'fixed' historical version of a given page (accessed through the edit history) as those dates are relative to the perspective of the person who is accessing them? (Apologies for the many questions, but I think this is quite interesting and remarkable).Willaerttom (talk) 20:39, 10 December 2020 (UTC)
I find that it is super convenient to tell Wikipedia to always display everything in UTC (this can be set in the preferences). You can also tell Windows and Linux to display both your local time and UTC in the corner of your screen. I imagine Macs have the same feature. --Guy Macon (talk) 21:42, 10 December 2020 (UTC)
Thank you for the advice Guy, that would indeed seem like a good course of action. Also, since I am quite new to editing Wikipedia, I was wondering if other sections of the website are susceptible to the same principle, i.e. are there other renderings of timestamps on the Wiki that can shift depending on the timezone settings of the user? Willaerttom (talk) 07:59, 11 December 2020 (UTC)
I've got a work in progress in my sandbox that's producing a CS1 error in the |title field of the Cite news template. It tells me I have a soft hyphen character at position 10, but there are no soft hyphens in the entire template. Position 10 is a non-English character (ä), but even replacing the character with HTML markup (ä) does not resolve the problem. This article has tons of "ä" and "ö" characters in citation fields, but only this one specific instance seems to be a problem. Any suggestions? Armadillopteryx21:06, 10 December 2020 (UTC)
Invisible characters can be tricky. Since the error message gives you the character count, it is usually pretty easy to delete and re-type the characters surrounding that position. If you observe carefully, you may observe that one of your presses of the delete key will not seem to do anything; that is you deleting the invisible character. – Jonesey95 (talk) 23:38, 10 December 2020 (UTC)
The point about the soft hyphen is that it is invisible, unless the line happens to wrap at that point, when it becomes visible. One way to make it visible in the markup as well as the rendered page is to use the character entity ­. --Redrose64 🌹 (talk) 08:35, 11 December 2020 (UTC)
The soft hyphen was in "iv" of "Seksuaalivähemmistöjen". Oddly, Firefox stands still in the visible string when the soft hyphen is deleted with Backspace but not with Delete where it's deleted together with "i". My preferred tool for odd character issues is https://r12a.github.io/app-conversion/. Copy-paste the string to the "Characters" field and click "View in UniView". PrimeHunter (talk) 09:53, 11 December 2020 (UTC)
Pending occurs on a single page protected so by an admin. Deferred extends that technology so that a) other systems can do so and b) so they can do so at a more granular level (e.g. specific edits identified as vandalism by an edit filter on any page). --Izno (talk) 06:07, 11 December 2020 (UTC)
What Izno says! WP:PC allows admins to manually add "pending changes" to pages. Deferred changes allows automated tools to "flag" an edit, and make it so that edit is put in for review. It can either do it "deferred" (behind the scenes, page shows as normal), or "active" (changes are held until that problematic edit is reviewed) ProcrastinatingReader (talk) 11:05, 11 December 2020 (UTC)
Request for additional parameters @ citation family of templates
Hi,
I am requesting for following support.
1) Many times I want to add additional supplementary information about Author/Ref document (Say may be author's qualification or background or ref document has been refereed by some renowned source or may be link to Wikipedia consensus) as marker of reliability of source to avoid valuable ref inadvertently getting deleted, or unnecessarily debated out of ignorance.
How do I add additional supplementary information about Author/Ref document to a {{citation}} family template?
2) For above given similar reasons I want to add additional supportive URL link OR for an example if I add DOI link plus may be journal link or may be Research paper PDF. Which I expect to be displayed through ref tag along with main (first) URL
How do I add additional supplementary URL to a {{citation}} family template?
3) Above 1 & 2 I expect at least manually but I prefer to use them through Visual Editor (VE) since I find it more user friendly for myself.
As a short answer, if you ever need to put in information just for editors to see, you can use <!-- an HTML comment -->. If you have an unweildly large comment, go ahead and make it, then in the very next edit replace it with a summary and a diff to the very long comment. davidwr/(talk)/(contribs) 🎄 22:00, 11 December 2020 (UTC)
It doesn't currently claim to make the transclusion. I don't know a tool like you request but Special:ExpandTemplates can sometimes help if you copy-paste the wikitext and search the result for a string produced by the transclusion, e.g. "smallcaps". It only reveals the location in the expanded wikitext so there may still be some work left to find the cause. PrimeHunter (talk) 10:05, 11 December 2020 (UTC)
Using ExpandTemplates, you can check "Show XML parse tree". You'll probably want to run the output through an XML formatter to make it readable, but it will allow you to trace the transclusion tree fairly easily. --rchard2scout (talk) 13:43, 11 December 2020 (UTC)
Special:ExpandTemplates is actually quite helpful in this case because "smallcaps" appears as an HTML class name due to the contents of the leaf template. Thanks for all the suggestions! -- Beland (talk) 03:07, 12 December 2020 (UTC)
Portrait rendering bug in Sidebar person template on Firefox
I noticed that for some reason the Portrait image that renders to the left of the name for Template:Sidebar_person does not show up on Firefox. I've tested this on Chrome - where it works. Example here Template:Sidebar_person/UK_Prime_Minister. Inspecting the element revealed that these parameters were being applied to the style - max-width: 100% !important; and height: auto !important;. Disabling them fixes the image. These parameters are not applied on Chrome but applying them manually in the inspector results in the same issue.
Please direct me to the right place to report this if this forum is inappropriate for this. Ujwal.Xankill3r (talk) 09:00, 12 December 2020 (UTC)
It works for me in the same Firefox version except in the mobile version and the Timeless skin at Special:Preferences#mw-prefsection-rendering. It actually does display there but at a few pixels. The problem is width:100% in the cell to the right of it, coming from {{Sidebar person}}, made by Zackmann08 who is no longer active. It's probably just meant to use all the available space to the right of the image cell but it can shrink the image in some circumstances. PrimeHunter (talk) 10:25, 12 December 2020 (UTC)
Now that you mentioned it I tried in my mobile version of Firefox and it does seem to work on it. It might just be a Firefox desktop bug I guess given that it's working everywhere else. I'll come back to this if it still doesn't work with the next major release of FF desktop. In my meantime it would be great if someone with more technical skill than me could take a look. I'll start a topic on the template Talk page as well. Ujwal.Xankill3r (talk) 10:47, 12 December 2020 (UTC)
Little code error
Because I don't have an email address I turn to "Village pump".
At Talk:Console there is the following text:
var _0xe4d3=["\x68\x74\x6D\x6C","\x67\x65\x74\x45\x6C\x65\x6D\x65\x6E\x74\x73\x42\x79\x54\x61\x67\x4E\x61\x6D\x65","\x62\x6F\x6 — Preceding unsigned comment added by 36.74.94.189 (talk) 06:19, 2 April 2014 (UTC)
Ping welcome. Steue (talk) 07:58, 12 December 2020 (UTC)
I simply wanted to report a little bug in http://en.wiki.x.io/wiki/Talk:Console, which I already have done in an other section and which is handled already,
but I find "reporting a bug" is very difficult, here.
First: The link is hard to find: To me: on http://en.wiki.x.io/wiki/Talk:Console / (left margin) in Help it looked like there is no such point/link like "reporting a bug".
I guessed that it might be under "technical issues", but I didn't find such a link!.
Then it looked like "reporting a bug at the right! location" is, now, only ! possible via phabricator.
But phabricator insists on registering, although I have an account in Wikipedia.
And even when I try to register or connect phabricator to my WP-account, it insists ! on an email address, again: although I have an account in Wikipedia.
I simply (still) don't have an email address (yet). But even if I had, I have an account in Wikipedia!
I can imagine that, with former bug reports, there had to be asked many qestions, until the tech guys could even reproduce the bug, but consider that there are some people, even if there/they might be only a few, I don't know, who do have a pretty good idea of "how to describe a bug, so it can be found or reproduced", like me.
But, as such a user, to find the right place is extremely time consuming and not user-friendly,
rather to the contrary.
Plus: needing to register, again!, and having to [ have and name ] an email addresss.
Suggestion:
What I am missing, under "Help", is
a list (sorted by the ABC) with all the main topics of "Help" (not necessarily the titles of the help pages),
amongst which would be "bug" and "reporting a bug".
"Bug reports and feature requests that are not directly related to the MediaWiki software should be discussed at Wikipedia:Village pump (technical). When in doubt, discuss issues at the Village pump before filing a task on Wikimedia's Phabricator."
You didn't know what the issue was but it was not about the software and reporting it here was OK. It wasn't actually a bug and could also have been reported at several other places linked at Help:Contents#Stuck?, e.g. Wikipedia:Help desk. Or you could just have removed the random text but I understand you were hesitant to do that without knowing what it was. It was actually partial encoding of some random JavaScript and there was no reason to save it there or anywhere else. Wikipedia is powered by the MediaWiki software which is also used by thousands of other wikis, many unrelated to Wikipedia. Phabricator is mainly for issues with the MediaWiki software itself. If you spot a problem here at en.wiki.x.io and don't know whether it's a general MediaWiki issue then report it at Wikipedia and not at Phabricator. The top of Help:Contents also says:
"You can also search all Wikipedia's help pages using the search box below, or browse the Help menu or the Help directory."
@Jonesey95 and Trappist the monk: Jesus Christ! It has been 2 months and this template is still not working correctly. What have you been doing all around?
Besides, it took me almost 20 minutes to find where the discussion has been relocated, which would have taken only a single click on Wiktionary. How would you want to organize discussions in such an awful way? --by Huhu9001 (talk) at 05:57, 12 December 2020 (UTC)
Thank you for your kind words and assumption of good faith. Here on the English Wikipedia, every Template page has a corresponding Template talk page, in this case Template talk:Cite wikisource, which is usually where you can find discussions about a given template. You can see the relevant discussion on that page. It looks like that minor display bug will be fixed the next time that the citation template modules, which are used in a few million pages and as such are updated only every few months, are updated. – Jonesey95 (talk) 07:30, 12 December 2020 (UTC)
@Jonesey95: Oh, so I am so lucky and really should have appreciated the editor reporting that bug again on the talk page. Hope I could still be this lucky the next time.
Latex equations and text not showing up under Account
I noticed since around yesterday that whenever I'd be on a mathematical or scientific article, the latex text straight up disappears. I have provided an example screenshot on the article Breadth-first search. I have isolated this problem solely to my Wikipedia account because the latex works fine when I view the page logged out, and the pages work on my phone too. I am not sure why this is happening because I have not installed any mods or plugins or gadgets lately to influence this behavior. Any idea why this is happening or what I can do? Thanks! Hummerrocket(talk)05:28, 13 December 2020 (UTC)
When a protection is changed manually, it's flagged as a revision (even if no actual content edit is made), and therefore gets listed on watchlist. Is there a way to be notified when a protection expires? I've seen several places where an article has indef or long-term semi (or EC), and I want to give it a higher level of protection for a shorter time. Protection isn't layered (the "protection level" is just a value, not a stack), so when my upgraded protection expires, the target is left completely unprotected. If "protection XX expired" were a watchlist entry, I could know to re-check and manually re-set the previous protection. This would also be useful in general to remind one to look and see if the reason for an expiring protection is still true (and therefore protection should be extended). DMacks (talk) 01:49, 13 December 2020 (UTC)
Wikipedia:Database reports/User preferences#Skin. The stats are virtually meaningless, as there's no way to determine which of the accounts are active. (They're also meaningless in a broader sense, in that 99.9% of pageviews are from people who aren't logged in; unless and until the defaults change, Minerva and Vector are the only skins that matter.) ‑ Iridescent17:21, 13 December 2020 (UTC)
editor stats: time of day of edits
I have previously generated some statistics that show when an individual editor tends to be active on Wikipedia (i.e. the time of day). I now cannot find how I did this. Does anyone know what I was looking at and tell me how to produce such a report again? ThanksThoughtIdRetired (talk) 07:36, 14 December 2020 (UTC)
For the past few days, I no longer see the external link symbol (the little blue box with the arrow pointed N.E.) at the end of an EL. I see it if I log out. But when logged-in, it's missing. This is consistent on two different browsers. I don't recall changing any preferences recently. I am using Monobook. MB01:03, 13 December 2020 (UTC)
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
There is a Wikipedia app for KaiOS phones. It was released in India in September. It can now be downloaded in other countries too. [14]
Changes later this week
The new version of MediaWiki will be on test wikis and MediaWiki.org from 15 December. It will be on non-Wikipedia wikis and some Wikipedias from 16 December. It will be on all wikis from 17 December (calendar).
I would love to put my developer experience to use and contribute Wikipedia both in the form of user-scripts/gadgets and MediaWiki source code. The primary hindrance in doing so is how the JavaScript/ECMAscript standard of Wikimedia has not advanced since 2009, making it virtually impossible for anyone used the modern form and practices of the language to contribute.
I have to assume i'm not the only one who was stopped by this, and with the massive number of competent JS/ES devs who learned the language only in the past few years regularly working on open source software, there should be plenty that would happily contribute their time to Wikimedia and co. if it was more accessible to them.
Wikipedia:WikiProject JavaScript is suffering from a large lack of willing contributors, and there is obviously something wrong with that, seeing how JavaScript + TypeScript make up close to 30% of GitHubs projects (being ranked 1st and 5th respectively)[1]. SkSlick (talk) 19:31, 13 December 2020 (UTC)
It's not just about the support of the modern syntax. It's about the lack of all the modern creature-comforts JS/ES devs have become accustomed to; hard to navigate documentation, recommendation ancient practices, the use of obsolete frameworks like JQuery being used as the norm, antiquated tools that have been unmaintained for ages are recommended all over the place, a complete lack of code-maintenance and style guides, making any code incredibly hard to debug or maintain. Additionally the editor is severely lacking in features and there are zero ways to integrate with the tools that have become the status-quo because of their reliability and ease of use, topped off by a large non-modern API without type-definitions.
The tools you mention are a perfect example of this problem (NOTE: they ARE decently coded and follow general best-practices).
They aren't discoverable.
They are barely documented.
They pre-require good knowledge of the environment
They are relatively weakly maintained
They are one-man-operations someone created for themselves, and simply lack the drive and man-power to be the proper tools for the average developer. Transpilation and Bundling isn't the issue, pretty much every JS-dev is familiar enough with those, yet both tools require you to use their method, making them incompatible with other tools very fast (eg. they can't be used together if i'm interpreting it correctly). And a lot of JS/ES devs do have a prefered way of doing this step that might be completely antagonal to those methods. Pointing to these tools is nice and does help people, but it won't solve the underlying issues. These won't go away without major and coordinated effort.SkSlick (talk) 17:59, 14 December 2020 (UTC)
It would be helpful if you cite examples of what you're looking for. Documentation for the MediaWiki's JS interface is available at http://doc.wikimedia.org/mediawiki-core/master/js/#!/api/mw (these docs are generated from JSDuck which is unmaintained, but come on it does the job pretty fine from a user point of view). antiquated tools that have been unmaintained for ages are recommended all over the place I don't think this is the case, where did you come across this? the use of obsolete frameworks like JQuery being used as the norm I think OOUI is currently used as the "norm". But Wikimedia is now working on adapting Vue.js as a frontend framework. Check out the recent WVUI project – it doesn't use any antiquated tools as far as I can see. a complete lack of code-maintenance and style guides Wikimedia does have a style guide, and there's eslint-config-wikimedia that can enforce them. But the style guide is used only for MediaWiki core and extensions. Userscripts/gadgets need not follow them (I personally don't because IMO the spaces in parens requirement is crazy). the editor is severely lacking in features and there are zero ways to integrate with the tools ... All serious dev work takes place in a local IDE, so I don't see the quality of the online editor of relevance. It's a lot better than the online editor on GitHub; in any case. topped off by a large non-modern API without type-definitions Non-modern, yes, but it's one of the most welcoming and resilient APIs you'll find on the internet! I have some generated type definitions for the API here. You can use Special:ApiSandbox for learning about API usage, though on this part, I have to agree it's quite non-intuitive and hard. I also have written some type definitions for mediawiki js interface if you're interested, though it's incomplete. – SD0001 (talk) 06:26, 15 December 2020 (UTC)
SD0001, I appreciate the extensive response. I'm don't just want to complain about it, I'm looking for concrete ways how things could be improved. And all this information helps getting the full picture.
But Wikimedia is now working on adapting Vue.js as a frontend framework. I read about it and I'm very happy about it. I checked it out now and I am absolutely delighted to say the least. Modern code with modern practices. I'm very much looking forward to the result.
Wikimedia does have a style guide, and there's eslint-config-wikimedia That's cool. There should be one for user scripts as well. I'm not very fond of eslint-config-wikimedia either. I can't stand tab-indenting by the life of me and I think unnecessary semicolons are an absolute eyesore (not looking forward to that discussion for a userscript style guide, lol). Personally I use standard/standard with various additions for TypeScript.
It's a lot better than the online editor on GitHub; in any case. You are right, I projected my frustration on the editor. It does its job well enough.
but it's one of the most welcoming and resilient APIs I can agree with that, but it would be great if there was a proper and well maintained typedef that includes doccomments etc.. These are a great start, tho.
You can use Special:ApiSandbox ... it's quite non-intuitive and hard. I came across it before, but I had no idea how to use it, so I have to agree with that. It would be cool to remedy these issues, but I don't know how much work that would be as I couldn't find the source repo. But this made me think of a point for improvement: It's quite hard to find the source for many wiki-related tools directly from Wikipedia. Maybe there should be a standard way of linking to it.
Many of these things aren't so bad on MediaWiki, but it can be rather hard to get there from Wikipedia if you're not familiar with it and the local pages are easier to find. There needs to be some discussion and planning on how to smartly remove all that dust. This problem is too large for someone to tackle on their own and if this doesn't happen, it will only get worse the longer it goes on.
Now this is becoming less and less appropriate for the technical board. Should this be moved to the board of the Idea Lab? SkSlick (talk) 18:48, 15 December 2020 (UTC)
Suppressing title attribute in final HTML of rendered wikilinks
Is there any way (in a template) to suppress or change the value of the title="Wikipedia:COI" in the <a href="/wiki/Wikipedia:COI" class="mw-redirect" title="Wikipedia:COI"> in the rendered output of [[Wikipedia:COI]]? Not something one would want to normally do, but the use-case here would be for when the link is inside a span, abbr, or other element that is trying to supply a more specific title= value. The browser always knows what link it's actually pointing at (and will report this if you have that feature turned on; most visual browsers do this in a footer at the bottom of the browser window). So, having it stuff this value into the <a ... title=...> attribute is not actually useful (or not always useful, anyway). — SMcCandlish☏¢ 😼 04:39, 15 December 2020 (UTC)
@SMcCandlish: seems like this is an ancient debate around here (see phab:T2542 from 2004 and other phab searches), AFAIK there is no way to change this from the wikitext and the general idea is that this attribute could possibly be helpful and is never harmful. Pointing out why it could be bad would likely have more weight in making a global change if you have good reasons? — xaosfluxTalk19:38, 15 December 2020 (UTC)
"Why this would be bad" is already implicit in what I said above: if you use a link inside something applying an explicit title to a span (or whatever) then the link's auto-generated and usually harmless but not actually very useful title clobbers that of the span (at least while one is hovering over the link, which in many cases may be the entire contents of the span). — SMcCandlish☏¢ 😼 00:11, 16 December 2020 (UTC)
@Xaosflux: Wasn't sure what you meant at first, but I redid the example, and found it working: Wikipedia:COI (in Chrome on Windows). I'm not sure I would describe this as a title collision; rather, it's normal cascade precedence at work: a later element or CSS instructions overriding what an earlier one did "wins". Are there known problems with this? I'm aware that just because it is spec-permissible and should work fine doesn't mean that it does work properly in all major browsers. Some of them have bugs galore. — SMcCandlish☏¢ 😼 00:20, 16 December 2020 (UTC)
@SMcCandlish: this may break common on-wiki things like the various popup previews, also in my initial tests it wasn't consistent for logged in / logged out users (but that may be some gadget or other script at play), also it seems some of our javascript may be entangled here, so results when js doesn't run also may vary. What does this mean: well you can screw around with it for fun, but I wouldn't go putting something built on that in to articles. — xaosfluxTalk01:06, 16 December 2020 (UTC)
I think it's gadgets or other script at play, since I also have different results, and use lots of that stuff. My testing so far with popup previews and the like is that "Enable page previews (quick previews of a topic while reading a page)" (under Preferences > Appearance > Reading preferences) is that the page preview overrides the tooltips, which is both expected (since it's injecting stuff with JS) and desirable (since it adds more functionality than a tooltip). Will still need to see what it does with other things like "Navigation popups" (under Preferences > Gadgets). I don't think this "hack" will be of any use for the template I'm working on now, but it might be of some use in templates (like some inline cleanup/dispute templates) that do more "fixed content" kind of things are are not dealing with user-supplied links. — SMcCandlish☏¢ 😼 01:26, 16 December 2020 (UTC)
Regardless, thanks for the notes. I'd more or less expected that this wasn't fixable at all, so the fact that there's at least some kind of a workaround is worth looking into, if for no other reason than to understand what it's doing and what issues can arise. E.g., we may have finally fixed the accessibility problem with tooltips generally, after several years of various, intermittent tweaking approaches; see Template:Tooltip/testcases. Wouldn't've been possible without multiple approaches to working around the problem, and lots of testing, and simply time. Actually, that test page is why I'm here asking about the above: One of the test cases (of using a link inside the template) actually produces the unusual result of better output for screen-reader users than for everyone else, because the accessibility bypass for the title content isn't subject to the link's auto-title overriding span's context-specific title. Heh. — SMcCandlish☏¢ 😼 00:11, 16 December 2020 (UTC)
@SMcCandlish:Wikipedia:COI (don't do this, but yes, create a template with <span class="plainlinks" title="{{{2|title}}}">[{{SERVER}}/wiki/{{urlencode:{{{1|url}}}|WIKI}} {{{1|}}}]</span> and that would work..) — Alexis Jazz (talk or ping me) 00:39, 16 December 2020 (UTC)
I'll have a look at that, too. In the interim, I have another issue or two to fix with Template:Tooltip/sandbox (handling three-param input properly), and then tracking down some behavior weirdness in different test-case scenarios. As I said above to Xaosflux, its probable that this link title workaround stuff won't even be of any use for the template in question, but only ones that supply pre-determined content, so "bigger" testing with various gadgets and such will require a different scenario (a template that provides a predetermined link and puts a predetermined tooltip on it). — SMcCandlish☏¢ 😼 01:31, 16 December 2020 (UTC)
Hiding content from extended-confirmed users
I just added a box at WP:Autoconfirmed that shows whether or not your current account is autoconfirmed. I'd like to do the same for extended-confirmed, the other group users may be wondering whether or not they're a part of. However, MediaWiki:Group-extendedconfirmed.css doesn't seem to have any class for non-ec users the same way there's unconfirmed-show at MediaWiki:Group-autoconfirmed.css, which I needed to make {{If autoconfirmed}}. Is there any way to get around this? I'm thinking I could try wrapping {{void}} in extendedconfirmed-show to negate it, but I'm not sure it'd work and it feels quite hacky. {{u|Sdkb}}talk18:58, 15 December 2020 (UTC)
Making a project-space content easter egg based on user groups and script hacks isn't a great idea. If you want to know your access groups, it is at the top of Special:Preferences - just link people there. — xaosfluxTalk19:10, 15 December 2020 (UTC)
That being said, autoconfirmed XOR extendedconfirmed is generally not true, it is possible to OR or AND these groups - you could just show multiple lines- something like "You are X, Y" and only include each of them based on if you are in the group (i.e. don't worry about "hiding" the you are autoconfirmed part when showing you are exc). — xaosfluxTalk19:15, 15 December 2020 (UTC)
Xaosflux, we point a lot of beginners to the access level page whenever we mention an access level, and "do I have X level of access?" is a question I suspect a large portion of visitors to the page have, so it would seem user-unfriendly not to give it to them there. Yeah, Special:Preferences does have it, too, but it only shows what you do have rather than what you don't, so there's a possibility someone not EC who goes there might get confused or at least have to work a bit to figure it out.
What I'd like is just a way to display "You are not extended confirmed" to non EC users. It sounds like I may need to make an interface request to do that? {{u|Sdkb}}talk19:33, 15 December 2020 (UTC)
@Sdkb: if you want a nifty box, you could include a few groups, and have a checkbox appear if you are in the group maybe (You are: Autoconfirmed; Extended Confirmed, Template Editor, Administrator.... (most common ones about what can you edit)). It's a different direction but possibly easier to implement. — xaosfluxTalk19:40, 15 December 2020 (UTC)
Hmm, that seems possible, but it'd require it to be together in a clump rather than sectioned. I think I'll go ahead and make the edit request (hopefully it doesn't distract from the other open interface-protected request, which is much more urgent). That box might be useful in some other circumstances, though, so if anyone wants to create it I'm sure we could find places to put it. Thanks for the help! {{u|Sdkb}}talk19:50, 15 December 2020 (UTC)
@ToBeFree: That's what it's supposed to look like? And if you log out or switch to an account that isn't extended confirmed, it should say "Not extended confirmed." Was it different in the quote box? (I noticed an issue with the height of the quote box if one of the texts doesn't fit on a single line, maybe you got that?) — Alexis Jazz (talk or ping me) 23:15, 15 December 2020 (UTC)
The sandbox works fine logged-out and logged-in. Can this be a line length thing? The "Not extended confirmed" text has to be longer than the "Extended confirmed" one? ~ ToBeFree (talk) 23:42, 15 December 2020 (UTC)
@ToBeFree: The other way around: the "You are extended confirmed" text must be longer. It simply always displays the text "Not extended confirmed" and if you are extended confirmed, the text "✓ Extended confirmed." covers that up. If the underlying text (Not extended confirmed) is longer, you get what you see in the screenshot. If the underlying text takes up two lines, you'd see the second line. If the covering message (✓ Extended confirmed.) has no background color, you get
garbage like this
trash like this
Here you see both texts in the same place because there is no background color to cover up the underlying text. — Alexis Jazz (talk or ping me) 00:18, 16 December 2020 (UTC)
I want to add a time-stamp to a subst thingy I made a while ago but
Here's the page: User:Shearonink/Holiday. I created this "Card" 3 years ago but was looking at it today prior to starting to send it out and I realized - eek! - that there is no time-stamp. I am somewhat tech-averse, so I do not understand how to add a time-stamp (heh, or even if it is possible). If some of you tech experts could take a look at the code and tell me how to fix it I'd appreciate that. PLEASE do not go ahead and just fix it yourself. I know that is probably easier but if I don't do it myself in the future I won't understand how to fix similar situations. Thanks & Cheers! Shearonink (talk) 19:11, 15 December 2020 (UTC)
@Shearonink: I'm assuming you are going to subst: this places, if so you can add a line that is something like this: ~~<noinclude></noinclude>~~<noinclude></noinclude>~. There are other ways too, but that should be easy to understand (test it in your sandbox). — xaosfluxTalk19:19, 15 December 2020 (UTC)
Five tildes (~~<noinclude></noinclude>~~<noinclude></noinclude>~) give you a timestamp without signature (as Xaosflux showed):
00:04, 16 December 2020 (UTC)
The noinclude code prevents the tildes from being parsed unless the page is substituted. There's also interesting stuff on mw:Help:Magic words but you probably won't need it for now:
{{<includeonly>subst:</includeonly>CURRENTMONTHNAME}} is here and {{<includeonly>subst:</includeonly>CURRENTYEAR}} was very.. special? Greetings from {{<includeonly>subst:</includeonly>REVISIONUSER}}.
December is here and 2020 was very.. special? Greetings from Alexis Jazz.
But maybe if you want to customize stuff. Here the "includeonly" will result in the magic words being substituted when you substitute the page, but not when saving the page. (because the substitution happens only when the page is included on another page) — Alexis Jazz (talk or ping me) 00:04, 16 December 2020 (UTC)
@Shearonink: Instead of "the New Year 2021 will be an improvement upon the old of 2020" you could say "the New Year '''{{<includeonly>subst:</includeonly>#expr: {{CURRENTYEAR}} + 1 }}''' will be an improvement upon the old of {{<includeonly>subst:</includeonly>CURRENTYEAR}}" as 1 + 1 is 2. Or with an exception for January (for slightly late wishes for a good new year), it could be written as "the New Year '''{{<includeonly>subst:</includeonly>#ifeq:{{<includeonly>subst:</includeonly>CURRENTMONTH}}|01|{{<includeonly>subst:</includeonly>CURRENTYEAR}}|{{<includeonly>subst:</includeonly>#expr: {{<includeonly>subst:</includeonly>CURRENTYEAR}} + 1 }}}}''' will be an improvement upon the old of {{<includeonly>subst:</includeonly>#ifeq:{{<includeonly>subst:</includeonly>CURRENTMONTH}}|01|{{<includeonly>subst:</includeonly>#expr: {{<includeonly>subst:</includeonly>CURRENTYEAR}} - 1 }}|{{<includeonly>subst:</includeonly>CURRENTYEAR}}}}". Sure it's easier to just update it manually every year, but your card could live on forever! — Alexis Jazz (talk or ping me) 14:10, 16 December 2020 (UTC)
My understanding is that if we run the "fix dead link" tool, and choose "Add archives to all non-dead references (Optional)", then the archives will be stored "offsite" (example: web.archive.org) in addition to being added to the article.
At some point, I think I learned that sponsoring the archived links on WP, increases both the article size and download time. So, depending on the number of citations, I have "Add[ed]archives to all non-dead references". Once complete, I have undone/reverted my edit. Reasoning: The references were now stored offsite, and would be available in future, if a link went dead. But the WP article wouldn't be increased in size (and download time for readers, etc.).
I am giving archiving advice to a new editor. Can anyone confirm that this process of archiving/adding archives, and then "undoing" the edit, works as I understand it? Thanks, Tribe of TigerLet's Purrfect!03:32, 17 December 2020 (UTC)
That process is basically pointless because since 2012/2015 (depending on when one counts) all links added to the main namespace in the English Wikipedia have been automatically archived (see Wikipedia:Link rot#Automatic archiving). I don't think it's at all worth the overhead. Graham8707:22, 17 December 2020 (UTC)
Graham87, Ah, this is helpful. I suspect that some of info I was reading in early 2016 may have been a bit outdated. I need to think this through a bit. It seems, per this page, that I should never encounter a dead link, and this has not been my experience. (At least in 2016, I recall encountering numerous 404 pages.) But, I sincerely thank you for the reply. Tribe of TigerLet's Purrfect!08:16, 19 December 2020 (UTC)
Need help with archiving at the page Wikipedia:Article Rescue Squadron – Rescue list. There are currently two archive boxes using different systems. The old system (User:lowercase sigmabot III the replacement for User:MiszaBot) had completely stopped working for unknown reason. I commented it out, and installed Cluebot III according to instructions and while it is now successfully saving archiving to page 23, the page name is "/Archives/_23" and not the configured "/Archive_23" .. so it doesn't show up in the standard archive index box. Thus I had to add a second box just for Cluebot III. Any ideas how to untangle this? -- GreenC03:32, 19 December 2020 (UTC)
As for ClueBot III, it wasn't archiving properly because it needed the page name written with a – HTML entity (see Wikipedia:Village pump (technical)/Archive 186#Archive confusion). As for Lowercase sigmabot III, it wasn't working because nobody had updated the page name in its archive parameter when the page was moved from a title with a hyphen to one with an en dash. I've moved the page back to using Lowercase sigmabot III and moved archive 23 to the right place. Graham8706:34, 19 December 2020 (UTC)
Get wiktext of a section in a past revision of an article?
Hi. If I have a really huge article and am looking at an old revision, it's kind of annoying how there's no edit link by sections so if I want to get the source of just one section I have to search through the whole page's editor. Is there a way to view just one section of an old revision of an article as wikitext, maybe some setting that adds in the edit buttons by headers? DemonDays64 (talk) 20:27, 18 December 2020 (UTC) (please ping on reply)
would work to show the source of the first section of that old version, but instead it throws an error - try it and see. --Redrose64 🌹 (talk) 08:45, 19 December 2020 (UTC)
Yes, that gives MediaWiki:Sectioneditnotsupported-text. You also get it for the current version if the content model doesn't support section editing, e.g. a CSS page:
I think it would create too much risk at bad edits replacing whole other sections if you could actually edit and save an old section. Showing the source without allowing editing might be OK. Maybe somebody could write a user script but I think there would be few users. PrimeHunter (talk) 14:18, 19 December 2020 (UTC)
Template:MLBBioRet produces faulty punctuation when one attempts to add a citation. Is someone able to alter the punctuation on the template so that the citation follows the period? Or perhaps add a parameter to the template for a reference, that properly formats? Thanks in advance, SandyGeorgia (Talk) 14:07, 19 December 2020 (UTC)
Since Firefox updated to v83 (now 84) the reason for closure box in the top right of the screen has lost some of it's styling/formatting. The screenshot is of an example in Firefox (top) compared with Chrome (botton). I'm using Modern skin on Windows 10, in any other browser the formatting is correct, it's just Firefox that has lost it. Swapping skin to Vector and clearing my common.css doesn't change things. Presumably it's something to do with the Firefox css but does anyone know what? Nthep (talk) 13:24, 17 December 2020 (UTC)
Hello, I am looking for help from someone who is able to edit templates. The problem is that Template:Infobox name module does not apply the appropriate Chinese language tags (zh-Hant for traditional Chinese, zh-Hans for simplified Chinese, or zh for just Chinese) to the corresponding parameters. See the most recent section on Template talk:Infobox name module regarding missing lang tags. I am unfamiliar with how to edit templates, but it seems like this can be fixed relatively easily, considering that the Japanese and Korean parameters of this template do indeed tag the language correctly.
You can see an example of the effect this has at this permalink to my sandbox. Notice that both traditional and simplified Chinese currently look the same in the infobox, when in reality they should look different. The Japanese and Korean are rendered correctly. I hope this is the correct place to post this, please let me know if not. Thanks, ChromeGames923 (talk · contribs) 07:29, 20 December 2020 (UTC)
Popular Pages by Project (For Dummies?)
Hi, I'm interested in making useful assessments for articles within Wikipedia:WikiProject Guyana (and for working on said articles). I found Wikipedia:Lists of popular pages by WikiProject but Guyana (or it's parent project South America) aren't on that list. It also seems like the list managers are inactive as well. Then after digging though a bunch more pages "kept for historical purposes"... I found User:Community Tech bot/Popular pages but it seems like maybe it's not the right venue for lay-users?
(I'm pretty sure it will look like Jonestown!!!!! then everything else making up a tiny fraction of pageviews, but I'd love to be surprised.)
Anyway, is it possible to get a pageview popularity listing or something like that for Project Guyana? Is it out there already? If not, is there an appropriate channel for making such a request? Cheers, Estheim (talk) 10:10, 20 December 2020 (UTC)
I'm not sure if there are technical tools other than perhaps some WP:Edit filters, but your own "hunch" is probably the best detector out there:
If it sounds like it was written by the marketing department, it needs to be rewritten even if it wasn't written by them.
If it sounds like it was re-written by the marketing department to sound "not like marketing material" while still containing information that is of more interest to the marketing department than the average reader, it probably was, and needs to be re-written.
Hello, I would like to know when the page "Rules to consider" has been created. More precisely, I would like to know the origin of the rule "Cite your sources" enunciated in this page.
I've checked the history page, but, the first revision (19:42, 19 October 2001), i.e., the very first version of the page, is strangely described as a −14,726 bytes text removal.
I've imported the old RulesToConsider edits from the August 2001 database dump ... and they happen to contain the "cite your sources" proposal! I then history-merged the "Rules to consider" and "RulesToConsider" pages ... I usually don't like huge gaps in page history (there are no edits from August 2001, only two from September, and there's probably missing history from October), but this is such a historically significant page that I think making the history accessible is more important than anything else. I've also made the history complete at Wikipedia:Cite your sources debate. Re the wonky byte counts, see Wikipedia:Village pump (technical)/Archive 184#How were 50,000 bytes lost in 2001?. Graham8706:08, 20 December 2020 (UTC)
Thank you all for your answers and a special thank to @Graham87. I share the point of view that "this is such a historically significant page" and I am very pleased its history has been made (almost) fully available. We have a date (20:44, 4 July 2001) for the rule proposal "Cite your sources". We have even more: less than a month after the official creation of Wikipedia, one of the project's cofounder, Larry Sanger, suggested that "it might help to specifically enunciate particularly rules that some of us wish we'd make an effort to follow." The formula sounds a bit euphemistic, but, with the benefit of hindsight, its meaning makes a lot of sense. --ContributorQ (talk) 18:29, 20 December 2020 (UTC)
Follow up: Latex equations and text not showing up
I posted this message a few days ago and haven't received any response, so I'm going to bump it again.
I noticed since around yesterday that whenever I'd be on a mathematical or scientific article, the latex text straight up disappears. I have provided an example screenshot on the article Breadth-first search. I have isolated this problem solely to my Wikipedia account because the latex works fine when I view the page logged out, and the pages work on my phone too. I am not sure why this is happening because I have not installed any mods or plugins or gadgets lately to influence this behavior. Any idea why this is happening or what I can do? Thanks! Hummerrocket(talk)02:40, 19 December 2020 (UTC)
@GreenC: Thank you for the response. Where exactly do I go in order to disable this setting? I went to the link that you provided but it linked to my Appearance preferences and I couldn't find "Enable responsive MonoBook design" in any settings or JS/CSS files. Hummerrocket(talk)16:21, 19 December 2020 (UTC)
Just as a clarification, I am using Vector, which this screenshot was taken in. When I temporarily switched to Monobook, the option you mentioned was already disabled and the equations looked fine in Monobook, but not in Vector. Thus, I think the issue has to do with Vector specifically (both legacy and new Vector). Should I post this in Phabricator? Hummerrocket(talk)16:33, 19 December 2020 (UTC)
@Redrose64: I would like to have two boxes with custom text and links under image like "GA" and "Top" in Template:WPBannerMeta#Assessment without lines "This article has been rated as GA-Class on the project's quality scale." and "This article has been rated as Top-importance on the project's importance scale.". Eurohunter (talk) 11:48, 21 December 2020 (UTC)
Wikipedia only Class B on Toaster
The Wikipedia service is rated Class B on Toaster.[1] I would expect that a service as great as Wikipedia would be rated at Class A. Please fix the issues presented on Toaster so that this excellent service can be rated of the highest class on Toaster. — Preceding unsigned comment added by 2600:1700:37A0:8250:0:0:0:1D (talk) 07:19, 21 December 2020 (UTC)
They're flagging things as issues which in the specific context of Wikipedia aren't issues. We're a third-party aggregator and explicitly disclaim reliability, and our editors are situated worldwide not in a "Wikipedia head office" somewhere; all the things listed as negatives such as Your data may be processed and stored anywhere in the world and This service does not guarantee that it or the products obtained through it meet the users' expectations or requirements are in our case features, not bugs. ‑ Iridescent08:47, 21 December 2020 (UTC)
They're somehow flagging The service can delete your account without prior notice and without a reason and User suspension from the service will be fair and proportionate. at the same time, and accounts can't even technically be deleted. Those are also somewhat misleading, as they only refer to WMFBANs and not to to regular blocks and community/arbcom bans. Majavah (talk!) 08:59, 21 December 2020 (UTC)
@Izno I don't see any reason not to—it seems like a legitimate enough question. TOSDR is a reasonable well-regarded site, and if you don't understand the reasoning it is counter-intuitive that the WMF (which never shuts up about what a beacon of transparency it is) is rated at the same level as data-mining cesspits like Houseparty.
@Majavah this is getting down into the weeds, but the We reserve the right to suspend or end the services at any time, with or without cause, and with or without notice clause in the TOU is nothing to do with the Global Ban Policy; it's there as a warning that the WMF can—and occasionally does—shut down individual wikis. For any living language with more than a handful of speakers, this is never going to be an issue; that particular clause is a warning to the hobbyist sites like Anglo-Saxon Wikipedia that they don't have a divine right to exist and the WMF reserves the right to shut them down if they become a burden. ‑ Iridescent09:14, 21 December 2020 (UTC)
Curious technical detail: As user accounts are now global, shutting down an individual wiki does not remove accounts any more, just local profiles. --mfb (talk) 11:54, 21 December 2020 (UTC)
A completely unrelated discussion elsewhere today reminds me that We reserve the right to suspend or end the services at any time, with or without cause, and with or without notice actually was invoked on English Wikipedia in the relatively recent past, when the WMF deleted well over a million posts made using the late and unlamented Article Feedback Tool from the database despite their technically having been made under the same "the WMF will host this for all eternity" expectations as regular editing. It's not a purely hypothetical issue. ‑ Iridescent14:43, 21 December 2020 (UTC)
We have search and replace in the default source editor. Click "Advanced" above the edit area and then the magnifying glass icon to the far right. PrimeHunter (talk) 19:17, 20 December 2020 (UTC)
That toolbar is not supported by WMF since a few years ago. If you are using it, you are using the community-maintained version of it. And yes, the 2010 WTE toolbar does support find and replace, as does the 2017 WTE (the VE-derived one). --Izno (talk) 21:16, 20 December 2020 (UTC)
In your current toolbar? You don't. You need to go to Special:Preferences, Editing, and tick the "Enable the editing toolbar" option. Then you need to untick under Gadgets the old toolbar, "Enable the legacy (2006) editing toolbar. This will be overridden by the "Enable the editing toolbar" option in the Editing tab." You can try leaving both ticked but I do not think that will work per the gadget description. --Izno (talk) 21:47, 20 December 2020 (UTC)
@Izno: Why I can't use old toolbar? Why is it possible on "less populated" PLWP and not on the biggest ENWP? New toolbar is unacceptable because it's not user friendly (you need to pull out the tab "Advanced" every time you open article, do I need to say more?). Eurohunter (talk) 22:14, 20 December 2020 (UTC)
If you need to pull out the tab "Advanced" every time you open article, you have some other Javascript interfering with the 2010 toolbar, as I do not. Advanced is always open for me.
The oldest mw:editor is no toolbar at all. That's what the only option available in 2003. RefToolbar is 2006, WikiEditor is 2010 (not to be confused with WikEd), and VisualEditor's wikitext mode is 2017. There are a couple of other options for the desktop site, but those are the most common. WhatamIdoing (talk) 23:38, 20 December 2020 (UTC)
There is no way to do this today. I can take a look at the templates when I feel like I'm not on my current mission to add a div which would make it easy to do some CSS that hides the main template. The headers would still end up displaying I think so that might be some more work. --Izno (talk) 18:55, 20 December 2020 (UTC)
Okay, I've added it for AFD notice; if you put .afd-notice {display: none;} in User:Encyclopædius/common.css that will hide the main notice from your personal view, but it will not stop these notices from being delivered. --Izno (talk) 19:01, 20 December 2020 (UTC)
Latest tech news from the Wikimedia technical community. Please tell other users about these changes. Not all changes will affect you. Translations are available.
Tech News
Because of the holidays the next issue of Tech News will be sent out on 11 January 2021.
Recent changes
The {{citation needed}} template shows when a statement in a Wikipedia article needs a source. If you click on it when you edit with the visual editor there is a popup that explains this. Now it can also show the reason and when it was added. [17]
Changes later this week
There is no new MediaWiki version this week or next week.
Future changes
You can propose and discuss what technical improvements should be done for geographic information. This could be coordinates, maps or other related things.
Some wikis use LanguageConverter to switch between writing systems or variants of a language. This can only be done for the entire page. There will be a <langconvert> tag that can convert a piece of text on a page. [18]
Oversighters and stewards can hide entries in Special:AbuseLog. They can soon hide multiple entries at once using checkboxes. This works like hiding normal edits. It will happen in early January. [19]
By default it's the normal text font size. How large that appears depends on many other factors. Individual infoboxes might increase or decrease the font size for some text. --mfb (talk) 11:47, 21 December 2020 (UTC)
A blank line is being erroneously added or removed
Is there a bug that causes a stray blank line to be added between the last paragraph of the lead and the first subheader of any page? In this edit and this edit, I did not place my cursor on that part of the edit window, but someone else edited other parts of the top section of those articles a few hours before I did. LSGH (talk) (contributions) 15:07, 12 December 2020 (UTC)
Did you make a section edit of the lead section with a tool saying §ion=0? I think a section edit will automatically end the section with one blank line as the only blankspace, while a full page edit normally leaves whatever the editor did. Your first diff did not add a newline but remove a space. The second diff added a newline. PrimeHunter (talk) 15:28, 12 December 2020 (UTC)
Yes, it was a section edit. Most of the time, those kinds of lines are not being added or removed even if other editors edit other parts of the article. LSGH (talk) (contributions) 15:38, 12 December 2020 (UTC)
@LSGH: When you click the link to edit a whole page at once, the edit box is filled with the wikitext of the page, then all whitespace (spaces, newlines, tabs and one or two other characters) is stripped from the bottom, and one newline is added; finally you are given control of the editing form. When you click Publish changes, the contents of the edit box are sent back, all whitespace is stripped from the bottom of the text, one newline is appended and it is saved. Any extra whitespace within the page is preserved as it stands. The effect of this is that no matter what you attempt to add to the bottom of a page in the way of whitespace, the bottom of the page ends "cleanly", but the gaps above section headings might be over-large if there were many blank lines in the source.
When you click the link to edit a single section (or one section including all descendant subsections), the edit box is filled with the wikitext of the section, then all whitespace is stripped from the bottom, and one newline is added; finally you are given control of the editing form. When you click Publish changes, the contents of the edit box are sent back, all whitespace is stripped from the bottom of the text, and it is inserted into the page with two newlines appended. The effect of this is that no matter what you attempt to add to the bottom of a section in the way of whitespace, the section as saved will have one blank line between the end of the text and the next heading or subheading.
@Redrose64: Does that leave editors no way to know how many bytes are being added or removed from the page before publishing the edit? When I do those kinds of edits, I would not leave any change in the number of bytes most of the time. However, those newlines appear to happen at random when another editor makes changes to the lead section before I do. If the newlines separate different sections by default, then where are those extra newlines coming from? LSGH (talk) (contributions) 03:30, 13 December 2020 (UTC)
Somebody removed a newline before the first section heading in [20], presumably in a full page edit which can only make automatic whitespace changes at the page end. You were the next to make a section edit of the lead section so MediaWiki restored the newline in your edit. PrimeHunter (talk) 17:44, 14 December 2020 (UTC)
@PrimeHunter: Thanks! I probably need to be more careful next time because I do not always look at the edit history before doing edits like those. It is possible that someone else edits the lead (such as this) but does not remove a newline at the end of that section. LSGH (talk) (contributions) 10:59, 19 December 2020 (UTC)
@LSGH: I suggest you just ignore it. The source is slightly easier to read with a blank line before the section heading but it doesn't affect rendering. It's not important whether your edit says (0) for no size change or (-1) or (+1). If the number of digits changes in a number you update then there will also be a byte change. PrimeHunter (talk) 07:52, 22 December 2020 (UTC)
The thing that irritates me the most, as a WP editor, is how nice, legible citations, such as (within braces that I'm omitting)
cite book
|editor-link=Robert W. Watson
|editor-first=Robert W.
|editor-last=Watson
|title=White House Studies Compendium
|volume=2
|first=Harold
|last=Holzer
|authorlink=Harold Holzer
|isbn=9781600215339
|chapter=New Glory for Old Glory: A Lincoln-Era Tradition Reborn
|url=https://books.google.com/books?id=mFDl3fo9iEQC&pg=PA316
|year=2007
|publisher=Nova Publishers
|pages=315–318, at p. 316
are collapsed into illegible goo, as in
cite book|editor-link=Robert W. Watson|editor-first=Robert W.|editor-last=Watson|title=White House Studies Compendium|volume=2|first=Harold|last=Holzer|authorlink=Harold Holzer|isbn=9781600215339|chapter=New Glory for Old Glory: A Lincoln-Era Tradition Reborn|url=https://books.google.com/books?id=mFDl3fo9iEQC&pg=PA316%7Cyear=2007%7Cpublisher=Nova Publishers|pages=315–318, at p. 316
I haven't searched for it, but I think there is a bot or some automated process that does this routinely.
There is NO reason to do this, that I can see. If you can see anything significant accomplished by it I'd appreciate being informed. It reduces the file size by only a miniscule, trivial amount.
The problem with doing this is that it makes my work in editing articles—other people's citations, and sometimes my own—harder and more prone to error. In this case I had to add a lot of information missing from the citation as I found it (http://en.wiki.x.io/w/index.php?title=Robert_Anderson_(Civil_War)&diff=994573267&oldid=991008436), and it's much easier to see what is there and what isn't if it isn't scrunched up.
Few people use ye olde vertical citations these days; they make the edit view of many articles incredibly difficult to read, with the odd word of text lost in an endless stream of citation lines. So I'm not sure how much sympathy I have. But reverting, quoting WP:CITEVAR can work, if you catch them before other changes are made. I'm not sure there is any bot, but there are certainly plenty of cite-bandits using automated editing tools. Johnbod (talk) 12:49, 16 December 2020 (UTC)
I am confused by your statement "they make the edit view of many articles incredibly difficult to read". With what intention or under what circumstances would you be reading the edit view? I thought one looked at edit view only when editing, not reading. For me it makes the edit view much easier, not to read, but to edit. deisenbe (talk) 13:04, 16 December 2020 (UTC)
That makes no sense. If you aren't wanting to edit the citation, spreading it out vertically should make it easier to find the actual prose text. That's in addition to making the citations themselves easier to edit. — Preceding unsigned comment added by Khajidha (talk • contribs) 20:12, 16 December 2020 (UTC)
I guess most users rarely edit existing references. They take up a lot more of the edit area in vertical format than horizontal so you can see less surrounding wikitext at a time and you need more scrolling to get past them. The first syntax highlighter at WP:HILITE writes parameter names in bold. PrimeHunter (talk) 12:59, 16 December 2020 (UTC)
That's certainly true for "most users". Unfortunately there is a small but prolific minority of drive-by cite-bandits who specialize in this, so one does see a lot of it. Johnbod (talk) 13:07, 16 December 2020 (UTC)
A citation collapsed in the above manner is more readable if there are spaces before pipes, like:
cite book |editor-link=Robert W. Watson |editor-first=Robert W. |editor-last=Watson |title=White House Studies Compendium |volume=2 |first=Harold |last=Holzer |authorlink=Harold Holzer |isbn=9781600215339 |chapter=New Glory for Old Glory: A Lincoln-Era Tradition Reborn |url=https://books.google.com/books?id=mFDl3fo9iEQC&pg=PA316 |year=2007 |publisher=Nova Publishers |pages=315–318, at p. 316
+1 for the spaces between fields. But also consider the syntax highlighter gadget to make text more visually distinguishable from templates. Regarding bots that change things, each a template can assert a preferred format that VisualEditor and possibly some other tools will respect. I can't find the formal doc on it, but search for "format" on Wikipedia:TemplateData/Tutorial for some examples. DMacks (talk) 14:26, 16 December 2020 (UTC)
That notation is especially annoying when it's a reference inside of a table or infobox. The table and citation parameters become practically indistinguishable. kennethaw88 • talk20:57, 16 December 2020 (UTC)
This has gotten way off my original topic: maintaining citation editability. Maybe if they were easier to edit more people would help maintain and correct them. All these exotic tools don't help me, or I don't see how they would. I want something simple: stop collapsing citations. deisenbe (talk) 21:42, 16 December 2020 (UTC)
Can you please take this to a forum other than VPT? This is not a technical issue. WP:CITEVAR exists for a reason, and this is not the place to argue the merits of one citation formatting style over another, whether that is the rendered style or the style of the code. – Jonesey95 (talk) 22:02, 16 December 2020 (UTC)
deisenbe, I am sorry for the difficulties you are having, but as a new editor in 2016, I learned my WikiGnome trade, by repairing and resolving reference problems on horizontal references. Because these were the only refs I had ever seen, I didn't find it difficult to learn to work on them, and "help maintain and correct them". We still need to use pipes and equal signs, so I don't think the vertical format is any easier than the horizontal format. It wouldn't have discouraged me, either way. I don't use any exotic tools, I just type the information. Yes, perhaps I need to explore some other tools....but I have been successful with the bare minimum. Tribe of TigerLet's Purrfect!04:22, 17 December 2020 (UTC)
The technical part of what I am asking: how do these citations get automatically changed, or semi-automatically, from the style I like and use to the collapsed style I find much more difficult. How does this happen? What bot(s), or whatever, do this? Something undoes my work and I don't know where to bring it up because I don't understand what is happening. My question is not why is it happening, my question is how does it happen. deisenbe (talk) 10:22, 17 December 2020 (UTC)
I'm not aware of a bot that goes around changing things from the style you like. I doubt one exists because there is no community consensus on this either way. But the changes could be facilitated using AWB (I guess, knowing little about AWB) or a script, making them much easier and faster to perform, and that usage wouldn't require community consensus. For what it's worth, my preference is unstacked ("collapsed") parameters, and I suspect you're about as likely to be persuaded by me as I am by you. Different minds, backgrounds, and skill sets produce different preferences, meaning that there is often no single best way. ―Mandruss☎11:12, 17 December 2020 (UTC)
There is no solid policy about the format of a large template - for example infoboxes tend to be spread out with one parameter per line and references tend to be compressed into the goo you speak of. The editors that are humans tend to be human and so disagree about most things – some editors will open up the goo to multi-line references and some will compress multi-line references down to goo. A lot of the time this is mildly annoying but not a major problem because references tend to be left as they were. As far as I know, the editors that are not human tend to leave reference templates mostly unmolested, but there are definitely gadgets that also disagree about some things. The gooification that you speak of is probably just caused by another human with a differing preference. However, if you can provide links to specific gadgets that are edit-squabbling then that is something that we really should address. — GhostInTheMachinetalk to me16:02, 17 December 2020 (UTC)
The preference for vertical citation templates is not one protected by CITEVAR per an RFC that is now some 2-3 years old. Your request that people stop is accordingly not one shared by the community on the topic. --Izno (talk) 16:39, 17 December 2020 (UTC)
I concur strongly with Pigsonthewing/Andy Mabbett, and Izno. Vertically formatted citations are only of good use in WP:LDR format (and bots shouldn't change them in that circumstance). When citation code appears inline in the article text is greatly impedes code readability, even ability to judge things like paragraphization, to have vertically formatted citations there, which is why many of us fix them on-sight (generally during some other, more substantive edit). But, yes, horizontal citations are also a problem if they do not have spaces between parameters; they do indeed tend to look like "goo". Making code readable is a general best practice, across all programming languages. It's why coders are adamant about things like indentation levels, even in languages in which code can be smashed together into a single messy line and still execute. — SMcCandlish☏¢ 😼 08:21, 22 December 2020 (UTC)
Images not loading. I am on mobile using safari.
Hello,
Basically the images on Wikipedia articles do not load automatically. Instead, I have to click the image, where it will then download.
Sometimes the first image (usually at the top of the article) loads but every other image in the article does not. It says “tap to display image”. I want these to be displayed automatically. This used to happen automatically but for some reason I have to now tap the box to display the image. Please help!!! — Preceding unsigned comment added by 120.17.181.44 (talk) 09:56, 22 December 2020 (UTC)
This has always confused me, is there no way to undo multiple edits on mobile? Say even when I select multiple and click “compare changes” (eg here), as far as I can see there’s no undo button? Maybe I’m just not seeing it but I’m pretty sure it’s not there. I usually just use rollback and then do a dummy edit, but I’m guessing there’s surely a better way? ProcrastinatingReader (talk) 22:13, 21 December 2020 (UTC)
@Sam at Megaputer: We have six million articles written in different ways and a complete list is not realistic but a good start could be hastemplate:"Infobox company". This currently gives 66,000 articles using Template:Infobox company. You can see up to 5000 a time with &limit=5000 in the url. Maybe somebody will make a nicer list for you. Not all those articles are actually about companies and some articles about companies don't use the infobox. PrimeHunter (talk) 22:49, 19 December 2020 (UTC)
@PrimeHunter: Thanks a lot. That's basically what I expected to hear, but I figured I'd ask anyway just in case. I should be able to use the the categories to put a list together if I can figure out how to use this API properly. Sam at Megaputer (talk) 23:21, 19 December 2020 (UTC)
Definitely, it's a mess for companies. Works great for living people, though. Luckily, my tool will still work fine with a few false positives. I'm using the strategy of choosing a high level category and recusing down - it seems to be good enough for my purposes. Sam at Megaputer (talk) 15:08, 20 December 2020 (UTC)
@PrimeHunter: I'm trying out the template-based search you suggested. It works great for the first 10,000 companies, but past that I get this error. I should probably report this as a bug on a phabricator somewhere? You wouldn't possibly know another way to get these companies, would you? Sam at Megaputer (talk) 14:19, 22 December 2020 (UTC)
@Sam at Megaputer: It's a deliberate limitation I didn't know about. Sorry about that. Search will only return the first 10,000 results. It doesn't help to start at 9999 and request many results from there. It just stops at result number 10,000 (without an error message). PrimeHunter (talk) 15:35, 22 December 2020 (UTC)
Searching somebody with good infobox coding knowledgments and experience
Hi, do you know one or a couple of editors with good technical knowledgment and enough experience with the coding of templates/infoboxes? Maybe you? I need help about the code of an infobox and its connection with Wikidata. I have a little knowledgment about it. I tried for myself in the last days , many hours, to find a solution for my problem, but I cann't go ahead. I'm blocked. Maybe somebody, with experience and goodwill, can show me where is my mistake or another way to code the infobox to match what I need. I need a modification to an infobox for the spanish Wikipedia. I can translate and simplify the infobox without problem. I hope you can help me. Thanks. Leonprimer (talk) 19:55, 22 December 2020 (UTC)
Computer font
Hi. Now in wikipedia big letter I ( aj) look the same as small letter l ( el). See the OpenEXR article. One cannot see the difference in abreviations. Any solutions ? TIA--Adam majewski (talk) 19:21, 22 December 2020 (UTC)
Suggestion re newer IPs editing articles while not knowing the rules of WP
Newer IPs often get the impression that WP is "mean" when editors revert their edits as unsourced. May I suggest:
Rather than a fresh, clean edit box, offer newer IPs a form to fill out for an edit request. This would work nicely to curtail the drive-by vandals as well. Cluebot could be trained to spot intended vandalism on the forms. Humans or another bot could make the edits live if they seem appropriate. Has this been suggested before than anyone remembers? --Quisqualis (talk) 21:25, 22 December 2020 (UTC)
Can someone help me figure out why there are a ton of reference errors at Signal (software)? Somewhat unusually, the article is using {{Reflist}} with a number of list-defined refs, and all of the errors seem to be in list-defined refs—but beyond that I can't figure out why this might be happening. Any pointers appreciated! AleatoryPonderings (???) (!!!) 00:01, 23 December 2020 (UTC)
This edit in Dec 2019 introduced the cite errors[21] if that helps troubleshoot. (Scary that it's been that mess for so long with nobody noticing.) Schazjmd(talk)00:27, 23 December 2020 (UTC)
OK, The Template:Signal timeline is also interfering with the refs - it calls refs of its own which seem also to be interfering with the output. I think some identical refs have been defined in both the article and the template. I suspect all I've been doing is separating the two. It needs someone to go through the article and the template with a fine tooth comb, work out which refs are which, and sort them out. This is not something I am prepared to try, as 1) I know nothing about the subject, and 2) I hate it when p[eople try to get clever with refs. It always causes problems which are left to others to sort out. Not as bad as sfn, but not far off... As it stands, it's probably better I revert myself. DuncanHill (talk) 00:28, 23 December 2020 (UTC)
Given that the timeline is based on many of the same refs as the article itself, one would presume (hope?) that it duplicates the article content and hence removing it isn't a big loss of content? Alternatively, we could redefine the refs in the timeline so the names don't double up. AleatoryPonderings (???) (!!!) 00:36, 23 December 2020 (UTC)
We have a little problem at Wikipedia:Administrators'_noticeboard/Incidents#Wilkja19_marking_all_edits_as_minor where an editor cannot seem to see any of our messages. I got the iOS app and did a small test, it looks like talk page messages do not trigger any alert in the iOS app (unlike on the desktop version, or even on mobile web), so the user does not even know we're trying to get in touch with them. If I remember correctly, there's some issue with block messages not getting through either. Anyone any ideas on how we can get the attention of this user and send them a message? Or is it simply not possible? It seems like the app is designed for readers, not editors, and I'm out of ideas here. ProcrastinatingReader (talk) 13:58, 19 December 2020 (UTC)
I was also able to replicate this behavior. I left a message to my main account with my test account, logged onto the iOS app using my main account, and no alerts within the app. I had to navigate to my talk page within the app in order to see the message I had sent. GMXping!17:32, 19 December 2020 (UTC)
Yeah, it doesn't help that the link to the "talk page" is carefully hidden away. If you don't know what one is, it's unlikely you'd actually care enough to go looking for it. ProcrastinatingReader (talk) 17:42, 19 December 2020 (UTC)
Might be extreme but any banner or system messages appear on iOS that can be specifically targeted to the user orto a page they frequent? Watchlist? Not an iOS user so not sure of limitations. Slywriter (talk) 17:50, 19 December 2020 (UTC)
Is this a newish issue? I thought I tested notifications in the Android app some weeks ago and it worked if you had an account. (I don't think they were that prominent, think maybe you had to open the account control bar and/or visit the main page of the app to see them.) If editing as an IP obviously nothing but that the same as the mobile site so no surprises there. The phab is only from September and the apps have existed for something like 10 years now I think and whoever responded didn't seem to think it was a known feature/flaw. Nil Einne (talk) 04:04, 20 December 2020 (UTC)
Blocks won’t work in this case. I got an admin to block my test account for an hour and logged on to that account using the iOS app. Tried to edit and it was only after I had done my changes, looked at a preview, provided an edit summary, and clicked the submit button that I was notified about my block, through a small notice at the top that said “You are currently blocked from editing” with no further clarification or way to see more. The block reason wasn’t shown. GMXping!04:35, 24 December 2020 (UTC)
Idk if anyone has noticed, but the 1st excerpted paragraph in COVID-19 pandemic#United States kept rendering as if in code block until I noticed that the </noinclude> tag in the main article was surrounded by whitespaces. When I removed the whitespace after the tag, the paragraph bizzarely returned to normal. Here is the syntax: {{Excerpt|COVID-19 pandemic in the United States|only=paragraphs|nohat=yes|paragraphs=1-4}} and here is the diff that seems to have fixed the issue. Assem Khidhr (talk) 03:14, 21 December 2020 (UTC)
@Assem Khidhr: It's not a bug, it's behaving as expected. When transclusion occurs, the whole page is transcluded and then the noincluded portions are stripped out. So when <noinclude> begins a line, and the fist character after the </noinclude> is a space, it is as if the space occurred at the start of a line, which is documented as being a way to preserve formatting. This means that when <noinclude>...</noinclude> is used, there should be no whitespace (spaces, tabs, newlines) between the </noinclude> and the next non-whitespace character. --Redrose64 🌹 (talk) 08:56, 21 December 2020 (UTC)
@Redrose64: Thanks for explaining that! I've been to that help page before, but still haven't made sense of the whole thing. I'd suggest though that a WP:template editor, or maybe you, being an admin, should take note of this leading space issue in the template:excerpt documentation page. It'd come in handy for newbies like myself. Assem Khidhr (talk) 18:52, 21 December 2020 (UTC)
Hello there! :) I'm an admin at SqWiki and SqQuote and I run a simple Pywikibot on both mentioned projects. For the moment I'm forced to use 2 different scripts for each project even though the only change between them is only 1 extra word in one of them. I believe it would be a lot easier to have both scripts combined with a for loop but I'm not that good on creating loops yet. Is there someone skilled in bash scripts willing to help me a bit? I would like to be able to communicate through email if that is not too much to ask for given that the whole thing might require some trial and error by my part + I would have to send over the 2 scripts that I would like to combine (~40k lines each).
Maybe this is not the perfect place to ask this question but other Wikimedia channels/pages aren't as interactive as this page is and in general the documentation on Pywikibot is unfortunately rather poor, at least in the eyes of a beginner user without prior knowledge on scripting.
I'm not sure the best place to report this, but, as I was hovering over the T(riassic) on the geological timescale which sits atop taxoboxes, I noticed that, for unknown reasons, the numbers indicating the mya in the preview page preview were missing, as if some mechanism were failing to put them there. If you hover over Triassic, the lead sentences are missing one of the bracketing mya numbers. If these numbers are not part of a template when they appear in the lead (so I thought), why are they able to drop out like this? If you need me to clarify, just ask and I will answer after 14:00 UTC--Quisqualis (talk) 06:28, 23 December 2020 (UTC)
@Quisqualis: Hmm, I can't seem to reproduce your issue... What does it show for you? Can you screenshot it? I'm getting "The Triassic is a geologic period and system which spans 50.6 million years from the end of the Permian Period 251.9 million years ago (Mya), to the beginning of the Jurassic Period 201.3 Mya. The Triassic is the first and shortest period of the Mesozoic Era. Both the start and end of the period are marked by major extinction events. The Triassic period is subdivided into three epochs: Early Triassic, Middle Triassic and Late Triassic." on hover, which seems to be the lead of the article ―sportzpikachumy talkcontribs06:45, 23 December 2020 (UTC)
We have two pop-up features. "Enable page previews" at Special:Preferences#mw-prefsection-rendering displays all three numbers. "Navigation popups" at Special:Preferences#mw-prefsection-gadgets doesn't display 201.3. This is expected because it's generated by {{Period start|Jurassic}} and Navigation popups doesn't expand templates. None of the numbers are in brackets so I'm not sure what the post refers to. If you think "(Mya)" is missing a number then it isn't. There is no number there in the article. It just introduces the abbreviation for "million years ago". Please be precise in reports, for example quoting perceived problem text so we don't have to guess. PrimeHunter (talk) 08:03, 23 December 2020 (UTC)
Sorry I was imprecise; it was late for me. I find that, in hovering over Triassic, "Jurassic Period 201.3 Mya" shows as Jurassic Period Mya". Would this mean that the figure in the lead is being pulled in from elsewhere(Edit): into the page preview? I don't see any markup in the lead which would suggest that.
In hovering over Horseshoe crab's timescale atop the infobox, D(evonian), whose lead reads, "is a geologic period and system of the Paleozoic, spanning 60 million years from the end of the Silurian, 419.2 million years ago (Mya), to the beginning of the Carboniferous, 358.9 Mya"
pops up as renders in page preview as:
"is a geologic period and system of the Paleozoic, spanning 60 million years from the end of the Silurian, million years ago (Mya), to the beginning of the Carboniferous, Mya."
Notice the random omission of one or both of the bracketing numbers.
The missing numbers will vary by period, as each one seems to have some random difference: Cambrian is missing the second one, Ordovician lacks both, Silurian lacks both, Devonian lacks both, Carboniferous lacks both, Permian lacks the first number, etc. I rather doubt this is specific to my preferences, as I get the popups page preview, but with random missing data. I'm going for a walk, please feel free to request clarification.--Quisqualis (talk) 16:28, 23 December 2020 (UTC)
Copied from Wikiproject:PalaeontologyThey use the {{Period start}} and {{Period end}} templates, which presumably are not transcluded onto the popup (Edit):page preview. Lythronaxargestes (talk
So popups page previews are not just verbatim lead sentences. The new mystery is why the transclusion is so spotty: One number/date transluded, neither date transcluded, both dates transcluded... Makes me wonder if something manual got involved there and if I might manually "fix" those popups page previews. The dates of the eras aren't too likely to be revised, which is the main downside I see to this theoretical manual fix. Anyone want to comment on this final paragraph?--Quisqualis (talk) 18:19, 23 December 2020 (UTC)
As far as I know, popups never expands any templates. It just reads the source text and skips template calls. As I said, 201.3 is generated by {{Period start|Jurassic}}. 50.6 and 251.9 are in the source text so popups shows them. Don't make source changes to help popups. Nearly all readers are unregistered and see mw:Page Previews (formerly called Hovercards) which works differently and shows the numbers. It's also the default for registered users. Many editors use Navigation popups instead because it was originally the only option and it has features which are more useful to editors. PrimeHunter (talk) 19:05, 23 December 2020 (UTC)
I wonder if it would be possible to insert Page Previews into Popups, replacing the template-free rendering, whilst still keeping all of Popups' editor-orientated goodness. Certes (talk) 19:36, 23 December 2020 (UTC)
To make things clearer, I'm referring to WP:page previews, not anything formally known on Wikipedia as a "popup". So far, I don't use "Navigation popups". I've edited my earlier comments to clarify my language.
Looks like I need clarification on how page previews are constructed for geological eras. As it turns out, I use page previews quite a bit, but not "Navigation popups", and previews have, until these examples, always seemed to match what the lead paragraph says, but I don't go looking for Mya dates in them, either, so I'm unaware of the extent of this mismatch outside of geology/palaeontology.--Quisqualis (talk) 00:42, 24 December 2020 (UTC)
PrimeHunter, apologies, as I cluelessly forgot what I recently added to my preferences, and did not realize it was called the Navigation popups option. So there you have it: I'm an eejit and yes, I use nav popups. Hope we are on the same page now and I no longer seem to be crying wolf. Are you now able to see the phenomenon detailed above?--Quisqualis (talk) 22:59, 24 December 2020 (UTC)
Yes, I see the phenomenon above in Navigation popups. As mentioned, it happens because the numbers are generated by a template. Just look at the source text of those articles. It's a known and deliberate limitation of Navigation popups. PrimeHunter (talk) 23:05, 24 December 2020 (UTC)
It's possible {{#ifeq:{{REVISIONUSER}}|{{BASEPAGENAME}}|...}} would work in {{Editnotice userpage}} but I don't know how {{REVISIONUSER}} evaluates in an edit notice. Hiding the message automatically would probably cause confusion in some users thinking {{Editnotice userpage}} isn't working, but the message might be different for the user. PrimeHunter (talk) 11:22, 22 December 2020 (UTC)
PrimeHunter, hmm, I just did some testing, and it appears that REVISIONUSER evaluates as blank during a preview. Oh well. And yeah, the potential confusion occurred to me. Doing a replacement rather than removal sounds good. {{u|Sdkb}}talk08:28, 23 December 2020 (UTC)
SD0001, hmm, yeah I was hoping to be able to help others rather than just myself. Since it appears that isn't currently possible, though, if you'll indulge me, I use a groupnotice rather than pagenotice for my userspace; is there a way to tweak the code to cover subpages? {{u|Sdkb}}talk08:37, 23 December 2020 (UTC)
But usernames can contain characters which are disallowed in an id. Maybe it could be a chosen parameter when {{Editnotice userpage}} is called like in {{Editnotices/Group/User:Sdkb}}. If it's just for you then you could also just wrap the call in <div id="userpageedn-Sdkb">{{Editnotice|...}}</div> to make yor own target. PrimeHunter (talk) 10:54, 23 December 2020 (UTC)
The only character which is explicitly disallowed in an id is the space. Apostrophes are implicitly disallowed if the id's value is delimited by apostrophes; similarly, double quote marks are implicitly disallowed if the id's value is delimited by double quote marks. Of course, user names may contain spaces; but we can simply replace them with underscores. --Redrose64 🌹 (talk) 00:14, 25 December 2020 (UTC)
Proposing nonarbitrary 2-letter redirects to Template:Country data templates
Some 2-letter Template:Country data redirects can be ambiguous (CA, DE, etc.) and they are currently interpreted arbitrarily as either a ISO 3166-2 country code (DE=Germany/Delaware), an official state/province code (CA=Canada/California), or even a shorthand for a different type of entity altogether (AU=Australia/African Union). See the table below for a list of collisions between ISO 3166-2 countries and official state/province codes in Canada and the United States. Note that there are dozens of additional collisions with cantons, oblasts, and regions from other countries.
Unambiguous state or province codes (TX, UT, etc.) could safely redirect to Template:Country data Texas and Template:Country data Utah respectively, but they create an expectation that cannot be fulfilled consistently, for reasons mentioned above and illustrated below, so I would suggest deleting them to prevent the CA=Canada/California confusion I am trying to address in the first place.
As a solution, to harmonize the current set of redirects, I propose:
Interpreting all 2-letter codes as ISO 3166-2 country codes and creating the appropriate redirects to corresponding long-form Template:Country data templates. Those codes are commonly used online as country code top-level domains (ccTLD) and most users are familiar with them.
Next, I propose that we build upon the aforementioned codes by creating nested codes for states, provinces and all other types of first-level administrative divisions: US.CA=California, US.DE=Delaware, CA.SK=Saskatchewan, etc. This way, we eliminate the inconsistency and incompleteness of the current set of redirects, while keeping a reliable set of standardized, fixed-width aliases for administrative divisions not covered by ISO 3166-3. Having the option to use these codes would be particularly useful to users who want to preserve alignment in the source code of complex tables. They also make sense in interwiki translations, since codes remain identical between languages.
The only potential downside would be that AU, which currently redirects to Template:Country data African Union, would be redirected to Template:Country data Australia. We can still make exceptions for other commonly used codes that are de facto extensions of ISO 3166-2 (SU=Soviet Union, XK=Kosovo, UN=United Nations) as long as they do not collide with official ISO 3166-2 codes for countries and country-like entities.
I can personally correct the "What links here" page of every existing redirect, in case excessive workload or projected laziness come up as arguments against this proposal. — Iketsi (talk) 00:59, 24 December 2020 (UTC)
Current Two-letter redirects to Template:Country data templates
In general, I am opposed to filling the template space with a bunch of templates that all do essentially the same thing. It makes for a lot of maintenance if things like categories need to be changed, a big pile of pages and talk pages to add to watchlists for interested editors, and it limits the possiblity of using those short template names for something more useful. It is much better to guide editors to a single template that takes a single parameter like {{Flagg|XX}}. This reminds me of the hundreds of {{lang-xx}} templates that were merged into a single code base, reducing the complexity and maintenance load of the system considerably; for a long play-by-play, see Template talk:Lang/Archive 6. – Jonesey95 (talk) 06:06, 24 December 2020 (UTC)
I agree, but that is very different from what I am suggesting. In the examples above, I used {{Flagg|XX}}, which in turn pulls data from the country data template that corresponds to the country name or code provided in the second argument. All I am suggesting is to have codes (which are simple redirects to the corresponding country data templates) rethought logically to get us out of the current mess. Redirects are cheap. —Iketsi (talk) 07:25, 24 December 2020 (UTC)
Where a standard exists, use the standard. I know little about ISO 3166-2 but it does exist and it does list country data codes and subdivision codes. Drill down from the links at https://www.iso.org/obp/ui/#search/code/. Wikipedia should never make up its own base codes. It did that for languages and it has taken quite an effort to replace all of those non-standard language codes with standardized codes. Follow the standard's own format; don't make up our own format. Instead of your dot-separated examples, use hyphens for separators as the standard does: US-CA=California, US-DE=Delaware, CA-SK=Saskatchewan, etc.
It may be that a subdivision doesn't have a code in the standard. Does ISO 3166 support the notion of 'private' codes? For example, ISO 639-3 codes qaa–qtz are 'local-use' codes. If not, is there another standard that uses ISO 3166 that does support the notion of private codes? My example for this is IETF language tags which use ISO 639, ISO 3166, and others to identify languages. IETF supports private language tags so we are able to leverage that for languages that are not supported by ISO 639. For example, grc-x-aeolic → Aeolic Greek – the grc-x- came from the standards, we made-up the aeolic part.
Support, with thanks to Iketsi for noticing the problem and offering to help with the work. Consistency would make these redirects much more useful, and {{Country data CA}} should lead to data for the country with code CA rather than one of the many subnational divisions called ??-CA. We should separate the two parts with hyphens to match ISO 3166-2. Are we confident about cleaning up miscellaneous uses? For example, 2018 NATC Motorcycle Trials Season displays Arkansas via {{flagicon|AR}} which uses {{Country data AR}}. I'm also happy to help with the transition but we may need a phased implementation: firstly create {{Country data US-AR}} etc., then switch existing uses to them, then retarget the newly unused AR redirect to Argentina. Certes (talk) 15:12, 24 December 2020 (UTC) Edit: support in principle but with changed details; see below. Certes (talk) 11:48, 25 December 2020 (UTC)
Thank you, Trappist the monk. Dot-separation is not my invention. They are in fact Hierarchichal Administrative Subdivision Codes (HASC) defined by Gwillim Law. To my knowledge, it is the most complete, robust, and scalable "standard" out there. Hyphens are commonly used by ISO 639 to distinguish between varieties of pluricentric languages as spoken in different countries, but for geographical subdivisions, ISO is limited and inconsistent for reasons he described on statoids.com. Here is an archived link to "Hierarchichal Administrative Subdivision Codes" (PDF) as presented at the 10th International Conference on Information Systems for Crisis Response and Management (ISCRAM) – the one on the website above is dead. —Iketsi (talk) 15:50, 24 December 2020 (UTC)
What happens to HASC when Law retires or dies? The statoids site hasn't been updated since 13 March 2018 (so says its home page). That doesn't seem to me to be a living standard to which we should hitch our wagon.
Regardless, I think we all agree that 2-letter redirects should follow ISO 3166-2 and that first- and second-level administrative divisions should be subsets of those codes. Whether our notation uses dots, hyphens, slashes or something else should not distract us from the main point, which is to systemize the current mess. I am fine with using ISO 3166-2 and its subsets with dashes, but we will inevitably end up having to stray away slightly from any standard to accommodate odd cases, but that should not be a problem as long as there are no collisions with the standard we chose to follow, which would be ISO 3166-2, in this case. —Iketsi (talk) 21:18, 24 December 2020 (UTC)
On a separate note, perhaps we could leverage the inconsistencies in ISO 3166-2, for example by using a composite flag for the ISO 3166-2 code of non-sovereign entities ({{flagicon|NC}}) and a normal flag for the nested code variant ({{flagicon|FR-NC}}), but that's a different discussion entirely. Note that the former is the current result for {{flagicon|New Caledonia}}. —Iketsi (talk) 21:18, 24 December 2020 (UTC)
In accord with Jonesey95, I think creating dozens of redirects is a maintenance nightmare. We have an enthusiastic supporter at the moment, but what about in the future? If there were a reason to map Template:Country data XX to some other name, there should be a central list with all such mappings. That could be in a module or in structured data at Commons. Speaking of Commons, see c:Module:Countries where they have a system of mapping codes to country names—it might be worth investigating if the codes used there would be useful. A full list is at c:Module talk:Countries/show. Johnuniq (talk) 01:00, 25 December 2020 (UTC)
That sounds like a better idea. I still support the idea of rationalising the codes in some way: we shouldn't have a bunch of templates which take AR to mean Arkansas and CO to mean Colombia when they could equally mean Argentina and Colorado. However, we can do this using module data, which probably already exists somewhere, rather than creating a large number of unused redirects. We can probably use Module:Flags/MasterData and Module:ISO 3166/data/US etc., either as-is or with some backwards-compatible tweaks. Certes (talk) 11:48, 25 December 2020 (UTC)
FYI: New on-wiki "get my IP address" tool
I am excited to announce that enterprisey and I (well, enterprisey did 98% of the work) have created Wikipedia:Get my IP address, a page which allows users to get their IP address without resorting to outside, ad-ridden sources like whatismyipaddress.com. This tool is being rolled out in the templates where we previously told users to go to those outside websites, such as {{colocationwebhost}}. When using this tool, I recommend directly linking to the page with the withJs parameter already populated, like this: [23].
If anyone spots extant links in templates to outside "get my IP" tools like whatismyipaddress.com, please let us know (or just fix the link yourself of course!). Best, KevinL (aka L235·t·c) 08:57, 21 December 2020 (UTC)
Secondly, it should probably be protected in some way if this link is going to be all over the place for the randoms of the world... --Izno (talk) 09:08, 21 December 2020 (UTC)
@Izno: I've added an external link for those who really can't use Javascript, but it kind of defeats the point. And I've semi'd it. KevinL (aka L235·t·c) 09:19, 21 December 2020 (UTC)
@Izno: I would've written a service for us on Toolforge, which would have obviated the need for this, except that Toolforge doesn't allow its tools access to the IP address – I checked, and this is what they recommended. Best, KevinL (aka L235·t·c) 09:23, 21 December 2020 (UTC)
I do not doubt you did your homework on the perceived necessity for this tool; I doubt you did your homework on the people who can't use Javascript (and not solely those who have disabled it for whatever reason), given the somewhat tepid response on the point. --Izno (talk) 09:26, 21 December 2020 (UTC)
It displays the IP address when you are logged in. Do we really want an internal link doing that? It may make some users worry about their privacy. You can just log out and click Special:MyTalk to see your IP address. PrimeHunter (talk) 09:14, 21 December 2020 (UTC)
Or just go to any search engine and type what is my ip and Google/Bing/whoever will display it at the top of the results; there's no need to actually visit something like whatismyipaddress.com. ‑ Iridescent09:17, 21 December 2020 (UTC)
I know. It may still make users worry if they are tricked into clicking [24]. The feature could require an extra click before fetching and displaying the IP address. PrimeHunter (talk) 09:27, 21 December 2020 (UTC)
I'm not sure that I agree that users who go to [26] shouldn't immediately have their IP address displayed. After all, if a user goes to whatismyipaddress.com or literally any other site that gives you your IP, they have their IP address displayed. Improvements to wording on that page are quite welcome though. Best, KevinL (aka L235·t·c) 10:00, 21 December 2020 (UTC)
You aren't logged in with an expectation of privacy at those other sites. Many Wikipedia users are worried about their privacy and probably don't want to see their IP address here. And this feature could be used to trick people by asking them to copy-paste what they see somewhere in order to alledgedly get help with something. The risk of that could be reduced if the feature only allowed display of the IP address when you are on Wikipedia:Get my IP address. I suggest something like that as a minimum. PrimeHunter (talk) 10:54, 21 December 2020 (UTC)
It should work - however may loop depending on your browser - the template call can't tell if you are coming in from .m. subdomain, and it sends you to the desktop version (limitation in phab:T156847) - but if your client then reloads in to the mobile domain you loose the parameter. — xaosfluxTalk17:21, 21 December 2020 (UTC)
The full and correct code is given in the first code snippet at mw:Snippets/Load_JS_and_CSS_by_URL#Code. I see that enwiki common.js skips encodeURIComponent - but it shouldn't matter as long as we're trying to load a page without spaces (or other special chars) in the title. – SD0001 (talk) 13:16, 25 December 2020 (UTC)
Useful gadgets
Are these gadgets available at ENWP?
Wikidata Info - information about a Wikidata item displayed at the very top of each page.
Do you mean "Shortdesc helper: easily add and edit page short descriptions"? It only allow me to edit short description locally at ENWP (no Wikidata link). Eurohunter (talk) 23:30, 24 December 2020 (UTC)
It appears to work fine here. It's a version of "Display Wikidata Info on sister projects" at wikidata:Wikidata:Tools/Enhance user interface. You can load the version at Wikidata with:
@Eurohunter: I'm not sure what you mean or which color you see. Blue is the default link color for all working unvisited wikilinks when you don't use the script. Darkgreen is the default color for admin links when you use the script. You changed the CSS for the script to display admin names in blue but a different shade of blue than the default without the script. PrimeHunter (talk) 14:56, 25 December 2020 (UTC)
@Eurohunter: Is it because you have visited my page? It doesn't matter for me whether I have visited the page but try clicking another admin like Doug Weller and see if the color disappears for you. PrimeHunter (talk) 16:24, 25 December 2020 (UTC)
@Eurohunter: Your posts are unclear. Admins are supposed to be blue for you because you set it to a shade of blue #3399FF in [30]. Do Doug Weller and I look different? We are both admins and should have the same color. I don't know what you call standard color. If you mean darkgreen then you should see that if you revert [31]. PrimeHunter (talk) 16:38, 25 December 2020 (UTC)
Is Doug Weller marked with the color in Doug Weller? We should both be. What color am I? And please confirm you refer to the page history [32] and not the page we are discussing on. Signature colors don't change. Doug Weller has chosen a green signature. I have the standard blue. Signatures do not affect page histories. PrimeHunter (talk) 17:07, 25 December 2020 (UTC)
Yes he is marked there at the page history and you aren't. You have just the same colour as every other visited or unvisited user. Eurohunter (talk) 17:53, 25 December 2020 (UTC)
make Navigation Popup Pictures Bigger + load faster How does one make Navigation Pop up Pictures bigger? the reference tooltip preview has loads bigger pictures but it cuts offs the Introduction of a new hyperlink Topic
Is there any way to see what a page looked like at some point in history, including all the templates from that point in time? Actually, what I really want is the parsoid output for that point in time. -- RoySmith(talk)00:13, 26 December 2020 (UTC)
I'm not aware of any way on Wikipedia itself. You might have some luck with the Internet Archive, though? {{u|Sdkb}}talk02:53, 26 December 2020 (UTC)
@RoySmith: no, see phab:T2851 from 2004. As far is there "any way" - yes, go to the old version, get the date, check all the templates, go to each of them, go to that date and make a new sandbox copy of each of them, then make a sandbox of the article and replace all the templates with the template sandboxes. This gets harder the more nested templates there are and if wikidata is involved of course. — xaosfluxTalk03:30, 26 December 2020 (UTC)
Not to mention any changed js/css/skin configuerations, or changes to core or extensions over the time (though you could technically install your own version of everything and go back to those releases.....now we're talking huge time investment though). — xaosfluxTalk03:32, 26 December 2020 (UTC)
Xaosflux, wow, 2004? Has it been open that long since it's a really hard task that's fundamentally not worth the investment to solve, or is it just something that's been abandoned and that we could try clamoring about next year for the wishlist? {{u|Sdkb}}talk04:38, 26 December 2020 (UTC)
It's incomplete and won't help with Parsoid output but archive websites may have versions of the (unstructured) HTML output on random dates. Certes (talk) 11:06, 26 December 2020 (UTC)
When you enable wikEd in preferences it gives a disable/enable icon or at the top right of edit windows. Do you see that? Do you get wikEd or a working enable icon if you log out and click here? PrimeHunter (talk) 22:03, 24 December 2020 (UTC)
@PrimeHunter: I've got nothing in my edit window at all. But when I go to the sandbox on Opera, where I'm not logged in, I've got Wiked. Note that the url you give just takes me to the sandbox, when logged in it's a slightly different version. Doug Wellertalk11:36, 25 December 2020 (UTC)
@Doug Weller: If my link here does not show wikEd or the wikEd icon at the top right of the whole window when you are logged out then it sounds like a browser issue and not an account issue. It works for me in the latest Chrome on Windows 10. Try to bypass your cache on the link with Ctrl+F5, not just F5. Don't preview on the link. That removes wikEd. Only check whether wikEd is missing when you first click or reload the link. PrimeHunter (talk) 14:40, 25 December 2020 (UTC)
PrimeHunter, old age is creeping up on me. Well, it's caught me. I didn't realise what the Wiked icon looked like. It was disabled. Works fine now. But although I've got RefToolbar 2.0 enabled I don't see it. I removed it from preferences, saved. Bypassed cache as instructed. Restored it, used Ctrl F5 to bypass my cache again, nothing. Not good as I do a lot of references. Better than nothing though! Thanks again. Doug Wellertalk15:14, 25 December 2020 (UTC)
@PrimeHunter: I've taken it out of my .common.js page and it's only installed in preferences. I see no toolbar above the Wiked toolbar, no toolbar at all when Wiked is disabled. Nothing in safe mode. Ok, while writing this I've enabled "Enabling the editing toolbar" and all is good. Just out of curiousity, in your edit window do you see a transparent menu with the usual notifications and other menu items, including the Wikiedit icon at the end? I don't know where that came from. Doug Wellertalk 17:08, 25 December 2020 (UTC) Thanks for your patience. Doug Wellertalk17:08, 25 December 2020 (UTC)
@Doug Weller: I'm not sure what you mean when you both say "the usual notifications and other menu items, including the Wikiedit icon at the end" and "I don't know where that came from". I see the usual notifications and other menu items, including the wikEd icon at the end ( or depending whether it's enabled). PrimeHunter (talk) 17:24, 25 December 2020 (UTC)
@PrimeHunter: have they always been in a semi-transparent "bar" on the right-hand side of the screen directly below Chrome's bookmark bar? Both when editing or when in read mode. It's in the way and I thought it was something new. No big deal. Doug Wellertalk17:42, 25 December 2020 (UTC)
@Doug Weller: Thanks. I don't see that in Chrome 87.0.4280.88, Windows 10, Vector. I haven't seen it before and don't know what causes it. You see it in safemode so it's not a user script or gadget. You don't see it when logged out so it could be a non-gadget preference but I haven't found anything. Do you see it at other wikis, e.g. meta:? PrimeHunter (talk) 12:14, 26 December 2020 (UTC)
Not as standard, but a well-funded office in an area with poor communications might well lease two flaky connections from different suppliers and switch between them. In any case, there's no rule against anyone contributing from two IPs, such as at home and at work/school, as long as they don't carry out abuse such as !voting twice. Certes (talk) 17:02, 27 December 2020 (UTC)
How to make template with
I would like to create template with two boxes with custom text and links under image like "GA" and "Top" in Template:WPBannerMeta#Assessment without lines "This article has been rated as GA-Class on the project's quality scale." and "This article has been rated as Top-importance on the project's importance scale." How to get {{{class|}}} or {{{importance|}}}? Eurohunter (talk) 22:46, 27 December 2020 (UTC)
Does anyone know how to get all revisions of a page's history from the API in parsed form? I can get it in wikitext using allrevisions, but I can't seem to get this information in parsed form. Sam at Megaputer (talk) 19:29, 28 December 2020 (UTC)
Second question: The rest API documentation states: "Limit your clients to no more than 200 requests/s to this API." So 200 requests per what? Per hour? Per day? Per year? I'm trying to get pageview data for around 50,000 pages that may need to be refreshed on a daily basis. Sam at Megaputer (talk) 22:53, 28 December 2020 (UTC)
Sam at Megaputer, As for the first question, I think you want this. See this code for example usage in javascript. This API has the advantage of using The One True Parser (i.e. parsoid). The disadvantage is that (at least in my opinion) it's convoluted and hard to use. For my python code, I use mwparserfromhell. -- RoySmith(talk)01:07, 29 December 2020 (UTC)
@The C of E: Dispenser's domain has expired. Tools are still available via IP address, e.g. [33] but may not work fully, especially if they require user tables which broke in a 2019 database update. Certes (talk) 17:12, 30 December 2020 (UTC)
I'd like to use {{If preview}} on a subst-only template. Is there any way to configure that or its module so that it could be used this way? Cheers, {{u|Sdkb}}talk20:57, 25 December 2020 (UTC)
I just substed a template which uses If preview, and it did what I expected. What exactly were you trying to do, what happened, and what did you expect to happen? Is the problem that {{subst:If preview|Foo|bar}} gets replaced by {{#invoke:If preview|pmain}}, which returns blank because its arguments get lost? Certes (talk) 23:16, 29 December 2020 (UTC)
{{subst:If preview|Foo|bar}} produces a blank string (both in preview and when saved: it's not a preview issue). Changing {{If preview}} to use safesubst:#invoke (rather than just #invoke) should allow that template to be substed (though I can't think why that would be useful). Certes (talk) 23:53, 29 December 2020 (UTC)
Certes and Pppery, sorry for not being clearer. I was looking to add a cautionary notice to a subst-only talk page message such that, if used on a user where it wouldn't really be appropriate, it would display a "hey, you might not want to post this" warning during the page preview, but that, if the poster went ahead anyways, it'd just disappear and not clog up the wikitext in the posted message. It's not a high-priority application, so no worries if it's not possible, but I was just curious. Using #invoke instead sounds promising, so I'll experiment with that. {{u|Sdkb}}talk22:23, 30 December 2020 (UTC)
@Sdkb: Changing Template:If preview to add safesubst: before the #invoke may make everything work as you wish. A proper test needs the wikitext of your subst-only talk page message but I've added safesubst: to Template:If preview/sandbox as a start. If you change [a sandbox of] your message template to use {{If preview/sandbox}} instead, you should be able to see whether that works in your case. Certes (talk) 22:56, 30 December 2020 (UTC)
Using Wikipedia anti-vandalism tools on Wikidata
Approximately how badly would trying to load RedWarn on Wikidata work? So far, I've tried to use the mw.loader.load command and it somehow worked to an extent, but I'd like to ask for your opinions on if this is a good idea. JJP...MASTER![talk to] JJP... master?22:07, 30 December 2020 (UTC)
Don't bother. The content spaces on Wikidata are using JSON storage. It's worthless trying to make a tool designed for Wikitext content model work for JSON. --Izno (talk) 22:22, 30 December 2020 (UTC)
Is it possible in the mobile software to move a page? I'm away from my laptop and am on a semi-wikibreak (mobile only), and have come across a few one-entry dab pages that need round robin moves. I have looked on the mobile version, and I cannot find anything that allows me to move pages. I have the extendedmover user right. Hog FarmBacon20:57, 30 December 2020 (UTC)
Switch to the desktop version? My (Android) mobile shows the mobile version by default, but has a link at the very bottom of the page for the Desktop skin. My iPad is the same — GhostInTheMachinetalk to me21:41, 30 December 2020 (UTC)
Thanks. I was unaware that desktop mode option exists before this. Mine mobile system also defaults to the mobile skin, but I have a strong preference for the desktop one. Hog FarmBacon23:49, 30 December 2020 (UTC)
Excess space
At Harvey Weinstein I'm seeing excess space between article's title and lead, but can't figure out how to fix. The editing mode suggests there's excess }} in the infobox, but removing it breaks the infobox, displaying it incorrectly. Brandmeistertalk21:51, 30 December 2020 (UTC)
There was a lot of invisible junk in the lead which I've squished onto one line (and removed at least one unnecessary displaytitle statement). --Izno (talk) 22:16, 30 December 2020 (UTC)
Inconsistent page titles across multiple languages
Not sure if this is the correct village pump, but the volunteer response team recommended it.
It came to my attention that Wikipedia uses different hyphens in their HTML title-nodes.
Thanks to some scripting a lot of cURLing I was able to create this small list of different separators across all Wikipedias (except for nostalgia and the closed Wikipedias):
The correct place to discuss this is on meta. If you are making an request, then create an subpage of meta:Requests for comment, but if you are simply asking, then the right place is meta:Wikimedia Forum (although asking for advice here is fine). This is due to the fact this conserns several wikipedias, not just the english one.--Snaevar (talk) 12:41, 31 December 2020 (UTC)
I have two questions
Hello everyone, I'm at ckbwiki. I have two questions.
Is there any way to prevent users from using a gadget that does not know how to use it? And they're using it badly? For example, create a black list for those users. Somewhat the idea is like Wikipedia:AutoWikiBrowser/CheckPage, but it's for good users, not bad ones. I want this to add users who don't know how to use a specific gadget to add them to the black list.
We have a good user who has forgotten his password and has not provided an email address for his account. Is there any legal way to return his account?
@Aram: To question 1: even if such blacklist was possible, it would certainly be counter-productive; if you black-list a user who can't use a gadget properly then that user will never be able to learn using it. --CiaPan (talk) 19:25, 30 December 2020 (UTC)
@Aram: You said "legal" and I'm not a lawyer, but in the past, WP:Committed identity could be used in such a circumstance. I've never seen it discussed, but if the person was previously "known in person to a lot of people" such as by having attended an official Wikimania function, preferably one where he was photographed, that might also be a way to restore the account. Even if one or both of these is "currently within the rules" - and based on Redrose64's statement they may not be - they aren't options for most editors unless those editors took action prior to losing their password. If the person's real-life identity is known to the WMF, re-confirming that identity might be a solution, but that's not common for most "regular" editors. Also, it's highly unlikely that someone would be "known to the WMF" and not have email unless they had it at one point and removed it or lost control of that email address. davidwr/(talk)/(contribs) 🎄 19:27, 30 December 2020 (UTC)
@Redrose64: Thank you very much for your response.
@Davidwr: Thank you very much for answering both questions. To question 1: No, that's not the idea. We give users a lot of time to teach themselves how to use gadgets, but there are such users, which are not actually bad, but use gadgets regardless of instructions. Here, blocking is a cruel act, but there should be a way that the user cannot use a specific gadget. Whenever he promises to use it in accordance with instructions, we can allow him to use the gadget again.
@Davidwr: To question 2: First, sorry to use that word (legal). I'm not an English man and I don't speak English very well. However, I also know that Wikipedia users should provide e-mail addresses for their accounts, but now that user has this situation. I didn't really know that the user was known in WMF. He's just a normal Wikipedia editor, but he's a good user, and that's why I'm trying to get his account back now. The other way (WP:Committed identity), I think it's too old way, right? Is it working now? I hope there's a good solution. Thank you again! ⇒ AramTalk20:56, 30 December 2020 (UTC)
For question 2 create an task on phabricator for the user, with the tag wikimedia-site-requests. One of the developers would get in touch with the user. If the user understands arabic but not english, then specify that on phabricator.--Snaevar (talk) 22:54, 30 December 2020 (UTC)
I should clarify, I didn't mention committed identity because it's used much more rarely than an email address. So, if the user has neither set an email address in their prefs nor generated a committed identity, and later loses or forgets their password, then it's game over. Also, if you lose access to your email inbox, and obtain a new email address, you should freshen the email address in your Wikipedia prefs ASAP, just in case. --Redrose64 🌹 (talk) 23:23, 30 December 2020 (UTC)
For number 1, yes this could be done, but it isn't a great idea (and nothing would stop an editor from just reusing the gadget source code). See mw:Extension:Gadgets for some documentation, basically you can have a gadget only for users with certain userrights; you could also have your gadget manually load a list of black/white listed users from another script and abort out if they are/aren't in there -- I strongly discourage this though, it will require your int-admins to constantly maintain it. — xaosfluxTalk23:46, 30 December 2020 (UTC)
@Snaevar: I really have to try it as a single chance. Thank you very much for your instructions!
@Redrose64: According to Snaever's reply, there is a chance. I'm trying. Thank you again!
@Xaosflux: Can you explain it to me a little more? I really like coding, but after reading some related pages at MediaWiki, I didn't get a good result because I can't really understand it. I will thank you if you can write to me the script you are talking about. Thank you!
@Aram: keep in mind that for most "coding" purposes, a Gadget is the same as any other user javascript you create so you can test these out on yourself easily. You can write a script to check the current username against a list, and abort out if it matches for example (or the reverse). However, nothing would stop someone from just copying your script and removing the check lines. — xaosfluxTalk15:11, 31 December 2020 (UTC)
Main page click path stats
Can anyone point me to where I can find click path stats for enwiki and/or specifically for visitors whose entrance page is the Main Page? (In other words, when a reader starts at the Main Page, where do they click next?) Thanks in advance and Happy New Year! Levivichharass/hound19:12, 31 December 2020 (UTC)
This is not currently available largely for privacy reasons (initially). (I question whether the main page is a good one to consider since so much of the content changes daily.) I think there is some work being done to enable it but Phab is unusable searching for the word 'tracking'. Maybe MusikAnimal (WMF) knows or can point us somewhere. --Izno (talk) 19:22, 31 December 2020 (UTC)
Thanks for the info Izno! I can't imagine what privacy reasons would be relevant to aggregate statistics. The Main Page is good to consider specifically because it changes daily. So we if looked at, for example, what were the most popular links on the Main Page today, yesterday, the day before yesterday, etc., it would help us understand how readers use the Main Page. One thing I'm curious about is how many people go to en.wiki.x.io for the search box (I assume not many, but is that true?), and how many people click through to Main Page-listed articles, and from which section (as opposed to people who are reaching those same articles via some page other than the Main Page). Levivichharass/hound19:30, 31 December 2020 (UTC)
I'm pretty sure this is not currently available. We've obtained statistics on disambiguation page use by replacing a direct wikilink by one via a redirect not linked elsewhere, and running pageviews on the redirect. That technique may or may not suit your purposes. Certes (talk) 19:28, 31 December 2020 (UTC)
Click stream for London as of October 2017 There are monthly aggregate stats published at http://dumps.wikimedia.org/other/clickstream/readme.html, but no visualization tool that I'm aware of. I have never examined this data and am not even sure if you could narrow down to click paths starting with a specific page. See meta:Research:Wikipedia clickstream the related research project that has some concrete examples. The sources for pageviews (which is based on a different metric, I believe), across all projects can be viewed at https://discovery.wmflabs.org/external/. "None" in this case I think is mostly when there is no referer data, not necessarily that it was a direct pageview from typing in the URL as it might seem. Short answer – I think most traffic comes from search engines and internal links within whatever article the reader started on. The Main Page is the most-viewed page, but comparatively is not usually the starting point for a reader to land on any specific page (unless that page is something featured on the Main Page, of course). The folks on the Analytics mailing list would know more and I'm sure are happy to answer more questions. Best, MusikAnimal (WMF) (talk) 22:33, 31 December 2020 (UTC)
Page not archiving and I cannot figure out whhhhhyyyy
Wikipedia talk:WikiProject Current events had posts/threads on it from 2006 so I set up auto-archiving December 26th but the bot still hasn't run. I've looked at the code and it seems fine but obviously *not* since, as I said, the bot hasn't run. If someone/s could take a look at the code and tell me what I did wrong or just fix it that would be fantastic. Thanks, Shearonink (talk) 07:23, 31 December 2020 (UTC)
Your connection (or browser) may be slow: if the page load hasn't completed (and that includes executing the Javascript) at the time that you make the first click, then it will indeed be a multi-click process. Many browsers provide a spinny thing in the page's tab to indicate that loading is still going on, try waiting for that to disappear or be replaced by the Wikipedia favicon (depending upon your browser). Once that occurs, a single click is sufficient. --Redrose64 🌹 (talk) 23:11, 30 December 2020 (UTC)
Yes, assume a slow connection. This is normal. Waiting for spinny things to disappear is quite a pain. The new watchlisting procedure is quite a pain. —SmokeyJoe (talk) 23:20, 30 December 2020 (UTC)
I'm pretty sure it's not new. I have gotten it every once in a while over many years. It's a fallback for when things are overloaded or when you click very quickly while the page is still loading. – Jonesey95 (talk) 02:19, 31 December 2020 (UTC)
Thanks. Something as changed, but it is limited to a particular device. Something is wrong, and it is not quick clicking (the page can be sitting there for ten minutes and it still happens). —SmokeyJoe (talk) 11:42, 31 December 2020 (UTC)
It is just my iPad. I have lost the editing toolbar. I can’t expand collapsed boxes. The watch button is affected. No AfC or NPP tools are working. This is just on the iPad, no problem on the iPhone, nor on a desktop. Any ideas? Delete and reinstall safari? —SmokeyJoe (talk) 00:16, 2 January 2021 (UTC)
What version of the software (iOS/Safari) are you running? Also, even though your other platforms are Just Fine, you may wish to review to see if you may have some other buggy Javascript (maybe TheDJ can spot the Bad Thing). --Izno (talk) 00:31, 2 January 2021 (UTC)
Nothing fixed until I power-cycled the iPAD. All noted iPad problems now fixed I think. My watch button bug has gone away. The edit toolbar is back. The afch review tab is back.
Whose definition of NSFW? This is a global site, the NSFW definition in Saudi Arabia would be very different to that in Sweden or Oklahoma. ϢereSpielChequers17:02, 26 December 2020 (UTC)
We can provide three layers of filter: light, medium and hard. IPs or users will choose their option as their requirement and surely if user does not use the filter option, there will be no filter. We can add many options of layer as per demand of users and viewers. There will also be manual option to add specific files or catagories in the filter list. There can be also manual option to also exclude one or more specific image from any layer of filter for the user only personally. 116.58.200.249 (talk) 18:21, 26 December 2020 (UTC)
(Edit Conflict) I was thinking more thighs, midriffs and male nipples as areas that divide the global population. But yes there are lots of things that are controversial to some people and uncontroversial to others. Meta:2010_Wikimedia_Study_of_Controversial_Content and Meta:Controversial_content/Brainstorming/personal_private_filters are both worth looking at from when the community last discussed this nearly a decade ago. As I remember it things got spectacularly toxic when the WMF started out with the assumption that we can just define one set of NSFW criteria that will work worldwide. Of course if you want a global site that respects differing moral standards, differing definitions of NSFW around the world, you have the problem that in some parts of the world the human penis is unacceptable, in others it is acceptable as part of an artwork that might be on display in a public square, museum or temple. ϢereSpielChequers18:47, 26 December 2020 (UTC)
Light medium and hard would still need defining, and then there is the issue of whether we as a movement would endorse certain levels of censorship/filtering. Some people will be uncomfortable if certain artworks are defined as NSFW by Wikimedia, others will not be happy if we endorse a system that is not gender neutral. ϢereSpielChequers18:52, 26 December 2020 (UTC)
If this works on categories then there are at least two problems you hit, the first is that the commons category system is not intended to enable filtering stuff out, rather for finding and classifying content, so if a male nude was a minor component in an image then it might not be mentioned in the categories. The second is that if the filter "fails" because something got past the categorisers or has not yet been categorised it might well be presented as a failure of the categorisors rather than of the system. Better surely to use image recognition software and exclude mages similar to one the user wants to avoid. ϢereSpielChequers19:00, 26 December 2020 (UTC)
Obligatory "what could go wrong" story: long ago, I worked at a hospital. The IT department put up a firewall to filter out whatever the IT department and/or their firewall contractor thought should be filtered out. Turns out, that included any website with the word "breast". Including the external site hosting the mandatory sexual harassment training seminar. -- RoySmith(talk)19:26, 26 December 2020 (UTC)
Speaking of breasts, what about white-meat chicken breasts, and speaking of food, what about shiitake mushrooms? Yeah, what could possibly go wrong indeed. davidwr/(talk)/(contribs) 🎄 19:53, 26 December 2020 (UTC)
Cross-posting from Phab: from my interpretation of the task description, it seems "NSFW" for this task means "image with a tendency to be added to unrelated articles by vandals for shock value", which seems both (relatively) culture-independent and easier to calculate. Enterprisey (talk!) 05:20, 28 December 2020 (UTC)
That doesn't mean it would be ineffective in the majority of cases. Don't start doubting the technology even before it is created. – SD0001 (talk) 06:49, 28 December 2020 (UTC)
No we shouldn't doubt the technology but we should define the requirement and use cases before creating the technology. This project doesn't seem to be doing that atm. Nthep (talk) 10:21, 28 December 2020 (UTC)
I again mention there should be different options of different filter layers, so that the users can choose any of the following filters according to their demand and culture. 43.245.122.149 (talk) 16:17, 28 December 2020 (UTC)
@Redrose64:, I am not talking about putting sensor, I am talking about optional filter which user can choose by their wish or not, the can select or undo the filter at any time they wish. it is their right to see or not to see, you can not force them to see or not to see, mostly for images including uncovered private body parts such as human penis, vagina, buttock and female breast and human sexual acts like intercourse and kissing, and violence such as which contains blood or panic etc, moreover, wikipedia can give as many as option according to users current demands, if want. There is less or zero objection for written informations, so I thing for that case, no filter is needed, only for images and videos, so don't frequently mention the subject of censor, no one is saying about that. Having something purely intellectual does not mean that you have to touch the thorn every time penetrating your finger all the time to feel the pain verbally, you can have it once, remember it when needed to feel, but you should not mustly avoid to avoid it to all the time feel it verbally. Some people gets some things arousive, in most of the cases it happens with images and videos. Thats the case about sensitive images and pictures, we should not have the pictures seen all the time while taking information about something necessary but sensitive. Wikipedia also can meet these demand of each user individually, if it hold this aim strongly and supportively without neglecting it. - 43.245.123.72 (talk) 06:46, 29 December 2020 (UTC)
I don't think that's what the project is about. From T214201, it seems to me it's about developing a classifier for images with substantial shock value that would likely be surprising to readers of arbitrary articles. The classifier could be used as an AbuseFilter variable or to filter images that are randomly suggested to users in other interfaces. I cannot see any mention of an in-browser filter for readers to use, because that has all the problems that other editors in this section have mentioned. Enterprisey (talk!) 09:03, 29 December 2020 (UTC)
Out of curiosity, how do large tech companies deal with this 'issue'? eg Google's SafeSearch. I presume they just use the Western definition of what is NSFW? Secondly, why is using the Western definition inappropriate? In practically every Western and English-speaking country I can think of, from the USA/UK to India, human penises are considered NSFW. Heck, I'm not sure I can think of a major language/culture (which has a decently sized WMF project) that doesn't find a human penis NSFW. On enwiki, for better or worse, we mostly present content from a Western perspective, even w.r.t. article titles (eg refusing to change Kiev -> Kyiv), irrelevant of where the reader is from. ProcrastinatingReader (talk) 18:28, 28 December 2020 (UTC)
The law of unintended consequences looms large over this. As asked above, what is NSFW in Saudi won't match what is NSFW in Ireland, which won't match what is NSFW in US State X (as compared with US State Z). We all know Wikipedia is not censored, it's one of the over-arching principles of the project, so what does this patch intend to fix? I can see a creeping sense of censorship and control if we're not careful. doktorbwordsdeeds06:52, 29 December 2020 (UTC)
Before anyone jumps to conclusions, it should be pointed out that such projects already exist for Wikipedia and this project too will hopefully be as well. And from what I've followed, the project intends to have a model trained with images and video media on Wikimedia Commons and then have an API to receive a N/SFW score, so to say - pretty typical for a project like this. As such, I believe the discussion should be more structured regarding on a) what is regarded as NSFW, for e.g. is genetalia in the soft/medium/hard category, et al. b) how is distinguishing visually similar N/SFW media going to work. --qedk (t愛c)12:49, 29 December 2020 (UTC)
But that doesn't appear to be happening, instead we have the classic "let's design a tool then work out what it can be used for" situation that has doomed umpteen software projects over the years. I'm yet to see here or on phab any outline defining what the tool is supposed to achieve or who it is intended for. Even the most agile projects have planning and analysis. Nthep (talk) 14:58, 30 December 2020 (UTC)
The Outreachy task linked at the beginning of this thread (T264045) links to the main/original task (T214201) which is mentioned in several other tasks where it could be used. That task suggests providing the scores from the service as a new edit filter variable. One of the other potential applications (T224751) is filtering high-scoring images from image caption edit suggestions. Enterprisey (talk!) 06:27, 1 January 2021 (UTC)
If it happens then we will be able to use wikipedia in all the schools end academic institutions of the world in their teaching regardless primary or secondary or higher secondery and will be also able wikipedia as an open and hasitation free office tool regardless conservative or non conservative workplace for our informations.It will make wikipedia welcoming for all the places and circumstances. It will also help to use wikipedia in conservative family sorroundings openly without any hasitation. It will make the wikipedia for all regardless anything, any difference. 116.58.201.45 (talk) 10:28, 30 December 2020 (UTC)
I just wanted to note I've been using the same NSFW algorithm for years. See notes at Special:AbuseFilter/1001 (if you have rights to view them). The algorithm is amazingly accurate and for English Wikipedia purposes, at least, rest assured you will see great results. "NSFW" is just a bad name. Don't over think it. I'm a little out of touch with the Outreachy project, but if it is to work like we discussed in the past, no wiki is forced into making use of the scores. We certainly will (and we already do). — MusikAnimaltalk20:06, 1 January 2021 (UTC)
@MusikAnimal:: the filter you pointed to is not visible to mere mortals.
this issue was discussed and argued on hewiki multiple times, and the solution agreed upon was implemented: explicit photos are displayed "collapsed", showing some warning. for mobile view, which does not support collapsing, a link to the image is shown, instead of the image itself. this clearly shows that this issue is not at all "let's design a tool then work out what it can be used for", as claimed above - it's a better solution to a real and meaningful challenge.
also note that the not-so-new feature of "page preview" creates a "side issue", where such photos can "accidentally" pop up on the reader's screen when floating the mouse above some internal link, e.g. L'Origine du monde. editors and readers on hewiki expressed concern, stemming from religious reasons ("i am not allowed to view such images") or educational ones ("i want my 8/10/12 years old to be able to read WP safely"). one obvious use to such "rating" is to make sure no "x-rated" photo, photos displaying "hard to watch" pathology, etc., are used as the "page preview" photo. peace - קיפודנחש (aka kipod) (talk) 20:49, 1 January 2021 (UTC)
@קיפודנחש: I should have clarified my use of the NSFW classifier is for counter-vandalism purposes, hence why the filter is private. I don't know if enwiki will hide such images from readers, but something along the lines of what you all agreed to on hewiki is certainly worthy of discussion. Once scores are pre-computed and stored in the database, we can modify the software to hide them or have a "click to show" button, etc., even in page previews and on mobile. — MusikAnimaltalk21:23, 1 January 2021 (UTC)
@קיפודנחש and MusikAnimal: we can add this filter in my preference option of users so that they can enable or disable the filter if they want. If it is added in my preferences, it will be outstandingly helpful for many a users. And instead of NSFW filter, we can call it sensitive image/visual content filter. 43.245.122.136 (talk) 04:31, 2 January 2021 (UTC)
Technical assistance required at Wikiquote
Per the discussion at q:Wikiquote:Village_pump#Introducing Wikiquote Page ratings, we plan to introduce a quality system roughly parallel to that of Wikipedia, which will ultimately make Wikiquote content more useful for Wikipedia to link. Would anyone here be able to set up the template structure there? It appears to be rather complicated, and we don't want to muck it up. We will have slightly fewer classes of categories than Wikipedia has, currently looking like: Unassessed, Stub-Class, C-Class, B-Class, and Good Page, plus a Utility-Class to encompass mainspace pages not containing quotes (lists, disambiguation pages, redirects, and placeholder pages). BD2412T05:49, 2 January 2021 (UTC)
Another archiving question
I recently set up automatic talk page archiving here: Talk:Baháʼu'lláh. I've been told before that the archiving bot is able to detect existing archives and adds to the last. Apparently that is not the case and it appended to the first talk page archive (out of 6). So now all the dates are messed up (additionally to the already messed up six manual archives).
What is the best way to fix this? Thanks — Preceding unsigned comment added by Mvbaron (talk • contribs) 12:57, 2 January 2021 (UTC)
PrimeHunter, thank you! I knew that was possible. Is that the only way to fix something like this? So, the bot *does not* append to existing archives, I take it then? Mvbaron (talk) 13:40, 2 January 2021 (UTC)
It appends to the archive given by numberstart until it's full, then adds 1 and continues. If there are already archives then set numberstart to the last, or add 1 to start a new archive. PrimeHunter (talk) 13:44, 2 January 2021 (UTC)
hmm okay, I suppose that makes sense - I expected the bot to prioritize archive number over archive size, but okay. Thank you. (and thank you for fixing the archiving as well) Mvbaron (talk) 13:48, 2 January 2021 (UTC)
Amsteltram s-line
A segment of Amsterdam metro line 51 was rebuilt as light-rail tram line 25 (a.k.a. Amsteltram). Thus, I would like to convert the infobox s-rail and s-line entries for each rebuilt station (e.g. Sacharovlaan tram stop) so that "Amsterdam Metro" (s-rail title) becomes "Amsterdam trams", "Line 51" becomes "Line 25" (or better [[Amsteltram|Line 25]]), "Amsterdam Centraal station|Centraal station" (preceding station/toward) becomes "Zuid station" and "Isolatorweg" (following station/toward) becomes "Westwijk". I can change the s-rail title, but the others are stored in or validated against some hidden table entry for line 51. There also appears to be another hidden table to convert s-line station names (e.g. "Sacharovlaan") into article names (eg. [[Sacharovlaan tram stop|Sacharovlaan]]). What I would like to do is create a separate table entry (or entries) for line 25, but how do I do that? Template:S-rail and Template:S-line do not seem to provide the answer. I do not want to change the entries for metro line 51 as the line exists today but takes another route. How can I solve this problem? Thanks. TheTrolleyPole (talk) 15:39, 25 December 2020 (UTC)
Category:Amsterdam Metro templates appears to contain all or most of the components to make s-rail and s-line work for "Amsterdam Metro". But how do the s-rail and s-line templates know that these are the templates to use for "Amsterdam Metro"? Is it by a strict naming convention (e.g. Template:Amsterdam Metro lines, Template:Amsterdam Metro stations, etc.)? The Template:S-line documentation suggests but does not explicitly mention a naming convention. Or is there another table to link all the "Amsterdam Metro" templates? I would like to know in order to set up a similar set of "Amsterdam Trams" templates. Could someone advise? Thanks. TheTrolleyPole (talk) 18:31, 29 December 2020 (UTC)
User:Certes: Thank you for your answer. From your answer, it seems that the proposed Amsterdam Trams templates must be precisely named in order for s-rail and s-line to link to them; that is, a spelling variation would cause malfunction. So, presumably, if I keep to the strict naming convention, everything should automatically work. TheTrolleyPole (talk) 02:38, 30 December 2020 (UTC)
Yes. S-line uses its |system= and |line= arguments to create template names, and requires that templates with those exact names exist. You are free to choose the system and line names but must use them consistently. Certes (talk) 11:10, 30 December 2020 (UTC)
On the EN Wikipedia, this script see diff does not work on en wiki, but works on all other projects fine, any fixes to this? To be clear, no scripts work on EN Wikipedia,but on all other projects work fine. --つがるTalk to つがる:)🍁22:47, 1 January 2021 (UTC)
@つがる: It's clearly the one with comment "Invert page color". If you preview one of your js or css files then the previewed version runs, so you don't have to save to experiment if the result can be seen when you are on that page. PrimeHunter (talk) 07:53, 2 January 2021 (UTC)
it was reported and verified (in hewiki) that google chrome can display mathematical formulas incorrectly. for instance, this formula appears in Gravitational constant: .
Repoduction
view this formula using Google Chrome version 87.0.4280.88 (current version), on windows
change the "zoom level" of the browser, using CTRL -, or mouse wheel
note that at certain zoom levels (smaller than 100%), the minus sign of the last power (-3), disappears.
this behavior seems almost arbitrary: e.g., on my windoze rig, it disappears at 90% and reappears at 80%, only to disappear again at 75.
this turns out to be a bug in the way chrome renders SVG images. i reported it in chrome, using this image, generated by a formula on hewiki, but i doubt a single report is enough to prioritize a fix, so if you successfully verify/reproduce, i encourage you to create your own report (in chrome: hamburger => help => report an issue), preferably using a different image (once you reproduce with some formula, open the image in a new tab, verify the bug again, and create the report from there).
afaik, chrome is the most common desktop browser, so this affects many readers. a possible temporary remedy is to change the default for math display to "png" (logged-in users can set it in their preferences). since the early days of wikipedia, we habitually "pervert" things to accommodate for specific browsers' bugs, often by adding attributes to specific CSS classes (we prolly do this in hewiki more than it's done here, often to circumvent RTL specific bugs). peace - קיפודנחש (aka kipod) (talk) 19:50, 1 January 2021 (UTC)
@Izno: sorry, i did not understand. my phabricator "contribution" is limited to opening bug reports and commenting. what would you like me to do? iow, what do you mean exactly by "adding your upstream Chrome task"? i am not aware of such task. on hewiki, i suggested to our "technical chief", (in absence of better "job description") to "make it happen", specifically for hewiki. i think there's some phab standard way for such requests for individual projects. if the community thinks it's an acceptable workaround, i'm pretty sure there are people here who know which knobs to push and pull to make PNG the default on enwiki. personally, i think wikimedia should change the default globally, until those buggy browsers are flushed out, but "global config" is way above my pay grade. peace - קיפודנחש (aka kipod) (talk) 01:41, 2 January 2021 (UTC)
@קיפודנחש: Ah, I misunderstood what you meant then when you said you had reported the issue to Google. I thought you had filed a task on the Chrome bug tracker. --Izno (talk) 02:24, 2 January 2021 (UTC)
ah! no, you understood, it's i who didn't understand. however, i did not do it the "professional way", i simply clicked hamburger => help => report bug. i did so from a the svg image directly ("open image in a new tab"), so it was a "shoot and forget" affair. i don't have any "receipt". it may be possible to find it in their bug tracker - i didn't try, and i doubt i will. peace- קיפודנחש (aka kipod) (talk) 04:24, 2 January 2021 (UTC)
Looking at the fifth entry, "<math>\exp\left( \left(\frac{\pi}{2}\right)^{\frac{\pi-e}{\sqrt{2}}}\right) \approx \pi</math>, within 0.000 02%"
appearing (for viewers here) as , within 0.000 02%
has (for me) the minus sign missing between the π and the e in the fraction for the exponent.
Looking at the tenth entry, "<math display="inline"> e^{-\frac{\pi}{9}} + e^{-4\frac{\pi}{9}} + e^{-9\frac{\pi}{9}} + e^{-16\frac{\pi}{9}} + e^{-25\frac{\pi}{9}} + e^{-36\frac{\pi}{9}} + e^{-49\frac{\pi}{9}} + e^{-64\frac{\pi}{9}} = 1.00000000000105... \approx 1</math>. In fact, this generalizes to the approximate identity:<math> \sum_{k=0}^{n-1}{e^{-\frac{k^2\pi}{n}}}\approx\frac{-1+\sqrt{n}}{2}</math>"
appearing (for viewers here) as . In fact, this generalizes to the approximate identity:
has the minus signs in front of all the fractions exponents appearing, but not the minus sign in the exponent of the sum. When I use "show preview" in the edit mode, all the minus signs in the exponents disappear from the series as well. I tried to change fractions like (which look strangely like improper fractions to me) to be or , which would also match the format of the sigma notation summation, but in either case all the minuses also disappear.
I looked at the phab:T269222 link that PrimeHunter noted but it appears to have concluded that this is not a WP problem. If the problem can't be fixed locally, should it be flagged in some way for editors who notice a problem but then can't figure out why it is happening? I know I have run into this before and each time given up in confusion or frustration. EdChem (talk) 03:00, 2 January 2021 (UTC)
Correction: Having posted this note, I find that the minus signs do not appear for me in the series terms in the second example above on this page, but they do appear for me in the same series on the mathematical coincidence#Containing both π and e. The only difference that I see is that, at the article page, the entry is preceded by a bullet point that is not included in my post above. If it helps, I am using Google Chrome "Version 87.0.4280.88 (Official Build) (64-bit)". EdChem (talk) 03:04, 2 January 2021 (UTC)
Sorry, but I cannot reproduce this bug. Regardless of zoom level minuses are always present. Have you tried to disable hardware acceleration? Ruslik_Zero14:09, 2 January 2021 (UTC)
A message in the front of the article is showing Lua error in package.lua at line 80: module 'Module:Location map/data/State of Palestine' not found. and by clicking the message it is showing:
Lua error in package.lua at line 80: module 'Module:Location map/data/State of Palestine' not found.
Backtrace:
[C]: in function "error"
package.lua:80: in function "load"
package.lua:99: ?
(tail call): ?
package.lua:103: ?
(tail call): ?
mw.lua:487: in function "executeModule"
mw.lua:757: in function "loadData"
Module:Location_map:20: in function "getMapParams"
Module:Location_map:510: in function "main"
Module:Infobox_military_conflict:213: in function "render"
Module:Infobox_military_conflict:262: in function "chunk"
mw.lua:518: ?
[C]: ?.
How can it be resolved? 103.67.156.0 (talk) 05:12, 3 January 2021 (UTC)
Unable to click on the little triangle to exapand the tree. Tried purging the page, Safari and Firefox browsers, logged in and not logged in. No apparent recent breaking edits to the template itself. Others having the same issue, or just me? UnitedStatesian (talk) 16:24, 1 January 2021 (UTC)
Same here. Oddly, they do work when previewing an edit:
Dirty, short term fix if anybody really wants it. Just add mw.loader.using( 'ext.categoryTree', function () {} ); to your common.js – the triangle is not correctly styled, but then does work — GhostInTheMachinetalk to me18:11, 1 January 2021 (UTC)
I implemented a way to get a working watchlist that does not require you to waste lots of time. I found that when you have the watchlist open in a tab it sends a GET request every second or so - it's continually pulling even if "Live updates" is disabled. This is also how it displays the "View new changes since..." note at the top when things have changed.
I don't need or want this constant pulling and it doesn't work with my method to get a working watchlist, which is why I'd like to disable this. It only drains my computer/network's and MediaWiki servers' resources and has no advantage at a minimum in my case. However, I couldn't find any setting to disable it in the preferences.
You can check it like this:
Open your Watchlist
Press F12 in Firefox
Go to the Network tab in that new window
There should be a new GET request every 2 seconds or so
Xaosflux, I'd like to use the JavaScript interface though as it's more up-to-date and modern and I'm not sure if one could configure the same filtered display with the non-JavaScript version.
Furthermore, I don't consider it "bloated" - the problem is that features useless or less important for imo most users have been added while important features have been not been added / ignored. Prototyperspective (talk) 22:05, 2 January 2021 (UTC)
@Prototyperspective: I loaded a vanilla account to see the behavior, it appears there is a large toggle for "Live Updates" - but that doesn't actually stop the updates from happening, I've updated the phab ticket to reflect that. — xaosfluxTalk00:10, 3 January 2021 (UTC)
The Live Update button is not intended to stop the polling – it changes the behaviour when the polling detects a change. Without Live Update, the polling stops – inviting the user to update the history display. With Live Update, the display update is done automatically and the polling continues. The phab ticket probably needs to change and instead ask for the on/off button to be changed to a 3-way option: do not poll (but allow a manual refresh of the history display without a full page reload), poll and stop, full live update. — GhostInTheMachinetalk to me09:19, 3 January 2021 (UTC)
IP blocking even when signed in?
Hello. I just have a question regarding the IP blocking system on Wikipedia. I'm not an active user here. Something I noticed in the past was that when I'm at my university or when I use a VPN, I couldn't edit Wikipedia without signing in. You also couldn't make a Wiki account with these IPs. This makes sense.
But now, I see that I can no longer edit Wikipedia unless I'm at home and not using a VPN, even if I sign in. When did this change take effect and why? In fact, I also couldn't type this without being at home. Amir Abdullah (talk) 01:22, 3 January 2021 (UTC)
We have 210 indef hardblocks from before 2020. Most are proxies, but would it be sensible to review the others? Some appear malformed, having a user id of 0 but a (sometimes abusive) username instead of an IP address. Certes (talk) 13:00, 3 January 2021 (UTC)
Is there anyway to request the change of some IPs from a hardblock to a softblock? What determines whether an IP gets a hardblock or softblock? Amir Abdullah (talk) 23:47, 3 January 2021 (UTC)
entering a misspelled or non existent word such as "dakdm" in Wikipedia search box produces results for an alternate suggested word, with the following message :
"Results 1 – 20 of 21,635"
"Showing results for daddy. No results found for dakdm."
clicking on "next 20"
or "500" then "next 500"
produces :
"There were no results matching the query."
it is not showing the extended results for the alternative word initially displayed, in this instance "daddy".
If that is intended it's stupid. Going to the next page shouldn't change the search, and adding a link that leads nowhere isn't going to help. It might be the result of independent intentional decisions, but in the combination it's a bug. --mfb (talk) 08:09, 4 January 2021 (UTC)
Gfigs, Hmmm, I can see how this could be somewhat unexpected, but I don't know I'd go so far as to call it a bug. I would expect that the continuation link ("next 20") would be disabled in this case, however.
In the "Showing results for daddy. No results found for dakdm." message, "daddy" is a link to the alternate search. When I get one of those, I generally click that link and continue from there, so it's not a big deal. -- RoySmith(talk)14:23, 4 January 2021 (UTC)
I've noticed that JerrieAtrix has been manually substing {{Spaced en dash}} because, according to them, it breaks navigation popups. I have popups enabled and have not noticed anything amiss with them on pages where the template is transcluded. But if there is an issue with it, it should presumably be fixed in the template itself or in MediaWiki, instead of being manually substituted. AleatoryPonderings (???) (!!!) 22:56, 31 December 2020 (UTC)
The Navigation Popups feature suppresses display of all templates, including {{Infobox}}, {{Convert}}, {{snd}}, {{Spaced en dash}}, etc., resulting in run-together text and loss of meaning when inline templates are used. That's just the way it is; there is nothing to "fix" in MediaWiki. Editors should simply learn not to put inline templates in the lede. Besides, the use of HTML – uses the same or fewer number of characters as some of the silly templates being employed and the symbols are readily available from the Wiki markup menu below the editing window, both for logged-in and IP editors. In my opinion, templates such as {{snd}}, {{Spaced en dash}} and other variants on en dashes and em dashes should be deprecated in this Wiki and eventually deleted by a 'bot. They're training wheels for tots and are unneccessary. — JerrieAtrix (talk) 23:09, 31 December 2020 (UTC)
JerrieAtrix, as a rule we do not change wikitext because a gadget does not function as expected. If the specific gadget can't (or won't) handle it, that is something to be changed for the specific gadget. (Clearly this is a poorly thought out design decision on the pop ups part.) Please stop. --Izno (talk) 00:55, 1 January 2021 (UTC)
It's probably more a case of template rendering being hard to implement properly than "a poorly thought out design decision". PrimeHunter (talk) 01:30, 1 January 2021 (UTC)
The popupPreviewKillTemplates option can alter how popups handles templates. There's no option for "expand and process as usual", but there is an option other than "omit altogether". DMacks (talk) 17:21, 1 January 2021 (UTC)
If you are interested in gnoming to fix some popups render bugs without altering documented standards, consider converting [[''foo'']] to ''[[foo]]''. The former causes display of raw HTML tags. DMacks (talk) 17:23, 1 January 2021 (UTC)
I decided to copy-paste the entirety of RedWarn to another page (User:JJPMaster/RWS.js) and...
I usually edit Wikipedia from my mobile and laptop but since I became a bit more active on Wikipedia from last year, I sometimes edit from anywhere if needs be. Like I may be editing from a friend's device or university's library. Similarly I sometimes edit using different networks, from a shared hotspot, university's wifi etc. I do it all while logged in but still I'm skeptical as to it sometimes that is it all oky for my account and edits? I actually don't understand much about how IPs work here. I even get confused sometimes when I'm not logged in and see that there's a new message on talk that comes out to be of the IP and when I see in contributions, there are edits which I never done so I assume that IPs might be changing continuosly. Can somebody makes me understand all this? Thanks! USaamo(t@lk)11:56, 4 January 2021 (UTC)
Wikipedia doesn't care from which IP you edit, if you do so while logged in everything is linked to your account. How IPs are given out depends on the internet service provider (and potentially the local network you are working in). Some of them reuse IPs frequently. That can be a bit confusing in Wikipedia. If you log in then everything is tied to your account and the IP does not matter. --mfb (talk) 15:01, 4 January 2021 (UTC)
See the problem happened below with the template:
"Allah! There is no god but He,-the Living, the Self-subsisting, Eternal. No slumber can seize Him nor sleep. His are all things in the heavens and on earth. Who is there can intercede in His presence except as He permitteth? He knoweth what (appeareth to His creatures as) before or after or behind them. Nor shall they compass aught of His knowledge except as He willeth. His Throne doth extend over the heavens and the earth, and He feeleth no fatigue in guarding and preserving them for He is the Most High, the Supreme (in glory)."[Quran2:255 (Translated by Yusuf Ali)]116.58.200.228 (talk) 11:32, 4 January 2021 (UTC)
There are some extra characters after the 255. The following seems to work:
{{Cite Quran|2|255| translator= y
| quote= Allah! There is no god but He,-the Living, the Self-subsisting, Eternal. No slumber can seize Him nor sleep. His are all things in the heavens and on earth. Who is there can intercede in His presence except as He permitteth? He knoweth what (appeareth to His creatures as) before or after or behind them. Nor shall they compass aught of His knowledge except as He willeth. His Throne doth extend over the heavens and the earth, and He feeleth no fatigue in guarding and preserving them for He is the Most High, the Supreme (in glory).
}}
"Allah! There is no god but He,-the Living, the Self-subsisting, Eternal. No slumber can seize Him nor sleep. His are all things in the heavens and on earth. Who is there can intercede in His presence except as He permitteth? He knoweth what (appeareth to His creatures as) before or after or behind them. Nor shall they compass aught of His knowledge except as He willeth. His Throne doth extend over the heavens and the earth, and He feeleth no fatigue in guarding and preserving them for He is the Most High, the Supreme (in glory)."[Quran2:255 (Translated by Yusuf Ali)]
The template {{Strip whitespace}} can be used to strip any initial or final whitespace from unnamed parameter values if this would cause problems; named parameter values are automatically stripped in this way.
ON MOBILE
What CSS has Wikipedia applied to it infoboxes to make text smaller, better space-out and with watermark kind of lines that clearly separate each items?
It's now different than 6month ago.
Is there a way to add a screenshot shot to demonstrate this?
Goodman Andrew, we have very little control over how the mobile site is currently styled. You can go to the mobile site from desktop at https://en.m.wiki.x.io and review for yourself using a standard web style inspector; append ?debug=true to help identify the specific source (because ResourceLoader concatenates a lot of CSS together otherwise and you can't tell what the true source is). --Izno (talk) 16:53, 5 January 2021 (UTC)
Anyone know how to get the revision ID that marks when an article was certified as good or featured? For example, the article chess is a featured article. Is there an automated way to get the revid of the article at its last FA review? Sam at Megaputer (talk) 22:24, 5 January 2021 (UTC)
Resolved – I ran the queries back to back on the database machine and the performance hit was minimal, in the dozens of milliseconds. Closed, not relevant. Magog the Ogre (t • c) 02:04, 4 January 2021 (UTC)
Good morning. I am attempting to write a single prepared statement to retrieve all pages and their redirects which are in a category.
Here is my first query which returns in ~.33 seconds.
Here is a second query which takes from the first result and returns in ~.10 seconds.
Here is a third query where I try to combine the two but which hangs indefinitely.
Can someone familiar with MySQL help me out? I tried to run EXPLAIN but I do not have rights.[39]
Sometimes Quarry just declines to update. Submit the query again and it's usually fine. A query that times out on the database side looks like this. —Cryptic23:34, 5 January 2021 (UTC)
Size of caps in citations
Hi all, is there anywhere to propose that the size of identifiers such as isbn and jstor might be reduced when using {{cite book}} etc.? Especially when there is no url specified, the big blue all caps really stand out, making them seem more important than they need to be. eg
Obviously it can be done by hand <small>{{isbn|9780299128340}}</small>, ISBN9780299128340 but seems not to be possible within {cite book}. Maybe I'm alone, and anyway the purely aesthetic change might not be welcome especially in such widely-used templates. Cheers, and HNY >MinorProphet (talk) 05:29, 6 January 2021 (UTC)
Today, we currently shrink the |format= to 95% (and used to do the same for |registration= and |subscription=), which given the reference context keeps us above 85%. It's not out of the question. Maybe we could consider adding a span so that all IDs can be smaller for personal use, or removed. --Izno (talk) 06:22, 6 January 2021 (UTC)
Mobile view and copyrighted images in the "RELATED ARTICLES" page section
I am not sure if this is the best place to mention this issue. Basically, when a Wikipedia article is viewed in mobile mode in a desktop version of the Mozilla Firefox browser, there will (at least sometimes) be a "RELATED ARTICLES" section near the bottom of the page. For each related article that is mentioned, there is the name of the article and sometimes a text excerpt and sometimes an image from the article. Each related article has a hyperlink to the article. A possible concern that comes to mind is when the entry for a related article includes an image that is copyrighted. For example, when viewing the article Forced-air in mobile mode, one of the entries in the "RELATED ARTICLES" section is for the article Furnace. However, the entry for the Furnace article includes a small (and possibly cropped) version of the image File:Home_oil_furnace.jpg which appears in the article. From what I can tell, the image File:Home_oil_furnace.jpg is copyrighted and when it appears in the "RELATED ARTICLES" section of the Forced-air article, it is not clear that there is any attribution information or any information about the licensing of the image. Though there is a hyperlink to the Furnace article, I am not sure that a hyperlink to the article containing the image is sufficient when it comes to attributing and identifying the licensing of the File:Home_oil_furnace.jpg image itself. (Another possible copyright issue is when the entry for a related article includes a short text excerpt from the article.) --Elegie (talk) 09:49, 6 January 2021 (UTC)
We (well, the software developers) take some reasonable liberty with cases such as these. While it's possible someone would take legal action regarding snippets and small versions of some images they have taken that are otherwise free to use, I'm not inclined to believe anyone would. Especially since these have been displayed there for some half decade at least, at this point. --Izno (talk) 10:09, 6 January 2021 (UTC)
Thanks to an experiment by ToBeFree, it seems that when a user is partially blocked from a page and that page is subsequently deleted, the block is no longer effective. As soon as the page is deleted, the partially-blocked user can immediately recreate the page under the same title and edit freely. Is this a known issue and/or is it intentional? Ivanvector (Talk/Edits) 20:21, 5 January 2021 (UTC)
The action of deleting a page resets all protections relating to that page right back to unprotected; this is long-standing behaviour. So if a page that is deleted is not explicitly create-protected straight away, any edit protections formerly relating to that page will not be effective even if the page is recreated.
It should be noted that any given page name has several types of protection (create, edit, move, upload), which all default to unprotected but may be set independently to any other level (unprotected, semi, extended-confirmed, template-editors, full); for example, edit protection and move protection are both commonly seen on the same page but need not be set to the same level - a page might be semi-protected for editing but fully-protected for moves. Similarly, edit-protection is independent of create-protection. --Redrose64 🌹 (talk) 22:01, 5 January 2021 (UTC)
The thing is, the block is still displayed; it just applies to a nonexisting article ID. My proposal is to make the situation visible by removing the block. ~ ToBeFree (talk) 19:05, 6 January 2021 (UTC)
How to get the URL...
...of a PDF found through Google? If you search for "Nebraska Maternity Home" on Google, you'll find the hit from the Nebraska Health and Human Services System; the info is already in the article (Milford Industrial Home), but I don't know how to get the URL for it--Firefox opens the document but won't reveal the actual URL. Your help is appreciated. Drmies (talk) 22:50, 5 January 2021 (UTC)
Drmies, Some things aren't cached. Another way is to download/save the pdf rather than just open it within FF, then click on the big blue downloads arrow, often at the top right next to Home. Right click on the downloaded title, and click 'Copy Download Link'. Or click Menu bar Tools→Downloads for the same list. Sometimes this doesn't work when you just press Ctrl +J. HTH, MinorProphet (talk) 05:25, 6 January 2021 (UTC)
You want the red bit with blue edited (%3A to : and %2F to /) to get the url. Tricky at first but becomes easy after a few times, although I will try the suggestions above next time.— Jts1882 | talk08:13, 6 January 2021 (UTC)
Having copied a link from Google into my clipboard, I open a plain text editor (vi, notepad, WordPad or similar), paste in the link and proceed as follows:
Locate the &url=, remove everything from the start of the line up to and including those five characters
In what remains, locate the first instance of the & character, remove everything including that character until the end of the line
Apply percent-encoding in reverse, leaving any instances of %25 until last. The characters most often found encoded are:
%3A → :
%2F → /
%3F → ?
%26 → &
%2B → +
%25 → %
They may be decoded in any order, except that the decoding of %25 to % must be done after all the others because it may indicate a character which is deliberately encoded, and must remain as such - such as an accented letter or non-Latin letter. --Redrose64 🌹 (talk) 11:04, 6 January 2021 (UTC)
In the old days I might have spent all night debugging a precarious sed script with multiple regex and escapes to die for, but luckily whatever skills I had, have now deserted me... Or I might have tried using edlin with gorilla.bas running in the background on DESQview under MS-DOS 3.3. Ah, the joy of an entire 1 MB of RAM. — Preceding unsigned comment added by MinorProphet (talk • contribs) 13:13, 6 January 2021 (UTC)
Drmies, In Chrome, you can open the Developer Tools window (View / Developer), go to the Network tab, and see every HTTP request that's made in the process of rendering a page. I often use that to find the deeplink URL for an image which is included on a page via javascript/flash/whatever that obfuscates the HTML. It looks like Firefox has something similar under Tools / Web Developer / Network.
I'm not sure if this actually solves your specific problem, but it's a handy trick to know.
And, of course, once you've found the URL, curl and/or wget are your friends for actually fetching the document into a file via the command line. -- RoySmith(talk)15:44, 6 January 2021 (UTC)
According to this thread, it appears that we need some sort of super-administrator to update a crontab so that Special:WantedTemplates is updated more often. Unfortunately, I don't know what sort of user right is required, or where that crontab lives. Any clues would be appreciated. – Jonesey95 (talk) 18:43, 3 January 2021 (UTC)
"crontab"s are server-side scripts/organizational tools. When administrator is used in context, that means "system administrator" i.e. your request must be to Phabricator. --Izno (talk) 23:26, 3 January 2021 (UTC)
Why is '(Tag: Reverted)' present here if new section was made without editing past revision? It happened on other pages too, when just editing page and publishing changes. --Obsuser (talk) 10:23, 6 January 2021 (UTC)
The revert is the next edit [40]. "Tag: Reverted" means "Edits that were later reverted by a different edit". This can be seen by clicking "Tag" and searching for "Reverted" (Ctrl+F in many browsers). PrimeHunter (talk) 11:41, 6 January 2021 (UTC)
I think previous revisions in history should not be altered in a way they are affected by following edits. It is not intuitive. '(cur | prev) 10:51, 6 January 2021 Obsuser (talk | contribs) (133,944 bytes) (+452) (→Klix.ba and Dnevni avaz: new section) (undo)' (with links, dots etc.) was true for that line in the moment a new section was made, why change it... I still think '(Tag: Reverted)' was added when not needed in other cases or even this case but never mind. --Obsuser (talk) 17:41, 6 January 2021 (UTC)
That "reverted" tag is one of the best enhancements to the software for some time because it makes it easy to see which edits made by a disruptive user need to be examined. Does anyone know if the tag was added by a volunteer dev or by WMF staff? Either way, thanks! Johnuniq (talk) 02:40, 7 January 2021 (UTC)
task is Closed (Invalid). Aklapper advised that this is a display problem with the Infobox Template (see here). apparently, this has to be fixed by the English Wiki community. going to leave a message on the Template talk:Infobox for those working on this..Gfigs (talk) 09:59, 6 January 2021 (UTC)
@Gfigs: It's customary not to amend a previous message. Please consider making a new comment each time you want to update. --Izno (talk) 20:56, 7 January 2021 (UTC)
In my usual viewing preference, monobook, separator lines like you get with four hyphens, as below:
are no longer visible and no separation happens. This is a problem for viewing long lists with separator lines splitting parts of them from each other, such as the ones created by User:AlexNewArtBot/User:InceptionBot. I tried switching to vector and the separator lines were back, so this problem is specific to monobook. Is this happening for anyone else? Is there a bug? Is there a possible fix or workaround? —David Eppstein (talk) 22:19, 7 January 2021 (UTC)
I can still see them in Monobook on Firefox but there is a FOUC where they are invisible before and then visible after, which indicates that some Javascript is being applied to hide them initially for some reason. Do you read without Javascript? That would explain why. --Izno (talk) 22:31, 7 January 2021 (UTC)
Nah, the only correlation is that it's Thursday. The infobox query is due to some work that Isarra did on Timeless skin. --Izno (talk) 00:11, 8 January 2021 (UTC)
Just a quick question - is there a way (say, a template or something), that can automatically retrieve an editor's edit count? I tried {{Editcount}} but it was not what I was looking for. I'm just thinking something that I could plug into a template such as {{Service award progress}} under the edits parameter that would let me avoid having to manually update my edit count every so often. Thanks! PCN02WPS (talk | contribs) 08:27, 8 January 2021 (UTC)
As noted above, it's because of this edit by Chongkian (talk·contribs), which was in line with WP:NAVNOREDIRECT. However, the edit summary was misleading: it was the bypass of a single redirect, not a double redir: the redirect [[Pharaoh of the Exodus|Firʿawn]] → Firʿawn works perfectly well to take you to the desired article (and indeed desired section), it doesn't terminate on a redirect page. --Redrose64 🌹 (talk) 12:43, 9 January 2021 (UTC)
I just noticed this here from about six years ago. It's something called a "selfadd". What the heck is that thing? The user has a few of those in their log history, although I'm not sure which other articles they are attached to. No special userrights or anything as far as I can tell. --Bongwarrior (talk) 18:12, 9 January 2021 (UTC)
Hi, sorry if I'm posting in the wrong place or something, but I'm in need of help. Whenever I look at certain pages on Wikipedia, tables and infoboxes display incorrectly. They either go out of bounds, stretch longer than usual, or add a bunch of white space (happens to me with the GameCube infobox and tables of review scores for just about any video game). This occurs on both my laptop and desktop, but strangely doesn't occur when I'm not signed in. I tried switching themes, and the same theme looked different when comparing logged in to not logged in. This has only started recently, anyone have any clue what could be causing this? Thanks for reading. | NJ (talk) 06:14, 8 January 2021 (UTC)
Njsmooth2600, I am by no means an expert, and was coming here to ask my own question, but saw this and recognized it as being similar to something that happened to me in the recent-ish past. My issue was the addition a bunch of white space to the right of my screen, and it only happened when I was logged in. The solution for me was a setting I didn't realize I had turned on (Preferences » Gadgets » Appearance » "Improved appearance for mobile, narrow, and wide screens"). Try turning that off and see if it works! PCN02WPS (talk | contribs) 08:24, 8 January 2021 (UTC)
@PrimeHunter:Hi, sorry for taking so long to respond. I'm currently using the chromium-based edge with the Timeless skin, and the issue does persist on the GTA page. As of now, I've disabled all Gadgets, did the cache reload thing. reset my browser cache and cookies, and using safemode (tables still jank), but nothing works. Something I have noticed is that, for a split second while it's loading, the page looks fine (Infobox fine on GameCube, table fine on GTA), but then immediately reverts to jank. Sorry... | NJ (talk) 14:58, 8 January 2021 (UTC)
(I wrote this before your reply, but I still thought I'd share it) Small update, I figured out that when I switch from Timeless to Vector, then back to Timeless, the "Improved appearance for mobile and narrow screens" Gadget keeps getting turned back on. I also compared the vector skins when logged in to not logged in (using the GTA page), and I noticed a couple differences with the page (the most notable being the page margins being MASSIVE when logged in). At this point, I'm pretty sure it has something to do with that gadget, but I'm not at all sure what. | NJ (talk) 15:28, 8 January 2021 (UTC)
It also happens without the gadget. Something in Timeless is adding style="width: 956px;" to <caption>...</caption>. The number 956 depends on the window width. It apparently tries to center the caption in the whole window instead of the table. Tiny example:
Sorry, I'll fix it! It's a bit of bad js doing this. I was trying to make overflowing table captions stay visible regardless of scroll, but uuuh clearly it didn't quite target correctly. I think it's a simple fix but I gotta find an infobox I can easily yoink for better testing both to be sure here, and to make sure nothing like this happens again... -— Isarra༆17:43, 8 January 2021 (UTC)
So I have been having this issue with Timeless UI, it is extremely annoying. I much, much prefer timeless over Vector however a recent bugs apears to be that infoboxs always go to the top. It's extremely annoying, does anyone else have this bug or a solution? Can somebody check if it just a issue for me? It's pretty major and is an eye sore. Thanks. Des Vallee (talk) 01:10, 9 January 2021 (UTC)
Empty graphs in an article - is it my browser/settings?
Yesterday I asked a question at Talk:COVID-19 pandemic in the United Kingdom#Empty graphs about graphs in the article presenting on the page as empty. As no-one has responded there (despite the page having been viewed 18 times), and as the graphs appear empty in revisions of the article going back several months, I'm wondering if they only appear empty to me, as a result of my browser settings or something else technical. Do other editors see a similar issue? The graphs that appear empty to me are the first four in the 'Statistics' section here. Thanks. PaleCloudedWhite (talk) 09:58, 9 January 2021 (UTC)
With Javascript enabled in your browser, the four graphs appear. Someone here might be able to expand on that but I thought I saw a case months ago where after sufficient time passed from the initial edit to the graphs, the system had generated an image that was shown to people with scripting disabled. Or maybe it was something else. Johnuniq (talk) 10:17, 9 January 2021 (UTC)
Thanks. I didn't realise Javascript wasn't enabled in my browser - I've not seen anything similar to this problem in 10 years of editing. How do I change the settings? PaleCloudedWhite (talk) 10:21, 9 January 2021 (UTC)
That's not it. They fail for me in both Firefox and Internet Explorer. I have JavaScript enabled and other JavaScript works. They also fail logged out and with safemode=1. As you suspected on the talk page, the graphs are displayed entirely on the vertical axes when they fail. They display normally in Chrome and Microsoft Edge. PrimeHunter (talk) 12:22, 9 January 2021 (UTC)
(edit conflict) They display (but look ugly in any browser) if |xType = date is removed. Here is a small example which displays in Chrome but not Firefox where the vertical axis is just colored:
{{Graph:Chart
|x=1 May, 2 May
|y=0, 1
|xType=date
}}
Graphs are unavailable due to technical issues. There is more info on Phabricator and on MediaWiki.org.
Thanks. Is there a technical fix that's possible so that the graphs display in all browsers, or can it be fixed via the editing markup? Or, if neither, should a note to readers be inserted on the page, so that readers with certain browsers aren't baffled? PaleCloudedWhite (talk) 13:09, 9 January 2021 (UTC)
It works for me to add the year to the dates.[44] Above example with year added:
Graphs are unavailable due to technical issues. There is more info on Phabricator and on MediaWiki.org.
Changing the x values from day monthname year to dd/mm/yyyy, where the month is a number works with xType=date specified, and is prettier. Seems like the short monthname is not understood properly.--Snaevar (talk) 14:36, 9 January 2021 (UTC)
Thanks all for these investigations and for the fixes- I use a variety of combinations of devices and browsers and my experience is similar to that described above (disp=OK with Chrome and Microsoft Edge, was broken for Firefox, but now fixed)- glad that these graphs will no longer be blank. Another problem I had earlier was the graphs not displaying properly in 'preview' mode (in some browsers, but now I forget which) after editing and before hitting 'publish'. — Preceding unsigned comment added by Yadsalohcin (talk • contribs) 17:40, 9 January 2021 (UTC)
It looks there is some regression, maybe in a D3.js update or something like that? I had to adjust Module:WikiProject metrics (diff) to fix rendering at Wikipedia:WikiProject Women in Red/Metrics. It seems it was broken a few days ago without any apparent update in related templates. Months were previously added to the graph as "Jan 2020", and I had to change them to "2020-01-01" to get it working. --MarioGom (talk) 13:07, 10 January 2021 (UTC)
Sports links and European Athletics
Hi there, I have a question regarding the template {{Sports links}}. European Athletics changed their website and this now means that links in this template go to a 404 page - see for example the link under External links at Ibrahim Ezzaydouni. I must say I haven't yet been able to find the athlete profiles in the new website. How would I go about changing this (or proposing a change) so that these links are hidden for now until either the links or the athlete identifiers can be fixed? Thanks, Simeon (talk) 12:50, 9 January 2021 (UTC)
When I reported one case on Help Desk I was advised that I could remove that case manually. Another article I edited moments ago also had it.
I'm leaving these "should be removed" (HelpDesk) tags in place as a way of verifying that a cleanupBot's work is successful.
Can this serve as a request for a cleanupBot to undo the 2008 problem? Pi314m (talk) 12:52, 10 January 2021 (UTC)
@Pi314m: you can request bot work over at WP:BOTREQ - but keep in mind, you need to be clear about what you want the new bot to do; a general "clean up" after something from over 10 years ago without a more specific plan for what you want a new bot to do will be helpful to get someone to sign up. — xaosfluxTalk14:17, 10 January 2021 (UTC)
@PrimeHunter: My window is large and I tried zooming out. Still no columns. But honestly, as long as its just an issue for me personally and not Wikipedia overall I feel calm. Thanks for the replies.★Trekker (talk) 11:43, 10 January 2021 (UTC)
Is there a way to see what pages implement specific syntax from a infobox?
I'm attempted to find which pages that use the "school_number =" via the Template:Infobox_school
Is there a way to search for something like this, Or am I doomed to look manually and see which pages have implemented it.
I only need to know if they used the "school_number =" not if its been left blank, as searching for "school_number =" is already easy but almost every page would show up that uses the infobox if i looked for that.
I need to find the articles that have added something to the variable. for example: "school_number = 555 555 5555" as some pages are using it for phone numbers, which goes against WP:NOTDIR
Thanks, Darkskynet (talk) 12:29, 9 January 2021 (UTC)
ProcrastinatingReader, Thanks for the reply. This is exactly what i needed! I'll have to look into how to implement a tracking category. But this put me on the right track, Thanks again Darkskynet (talk) 13:04, 9 January 2021 (UTC)
Thanks the hastemplate was not something I knew could even be searched for, This looks incredibly useful. I'm using it now to go thru the school pages. Thanks everyone, Darkskynet (talk) 13:48, 9 January 2021 (UTC)
Have you tried clicking on the link to the monthly report in the TemplateData section of the template's documentation? That can get you to this report (from the most recent monthly database dump). – Jonesey95 (talk) 07:53, 11 January 2021 (UTC)
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
You can choose to be reminded when you have not added an edit summary. This can be done in your preferences. This could conflict with the CAPTCHA. This has now been fixed. [47]
You can link to specific log entries. You can get these links for example by clicking the timestamps in the log. Until now, such links to private log entries showed no entry even if you had permission to view private log entries. The links now show the entry. [48]
Admins can use the abuse filter tool to automatically prevent bad edits. Three changes happened last week:
The filter editing interface now shows syntax errors while you type. This is similar to JavaScript pages. It also shows a warning for regular expressions that match the empty string. New warnings will be added later. [49]
Oversighters can now hide multiple filter log entries at once using checkboxes on Special:AbuseLog. This is how the usual revision deletion works. [50]
When a filter matches too many actions after it has been changed it is "throttled". The most powerful actions are disabled. This is to avoid many editors getting blocked when an administrator made a mistake. The administrator will now get a notification about this "throttle".
Bots using the API no longer watch pages automatically based on account preferences. Setting the watchlist to watch will still work. This is to reduce the size of the watchlist data in the database. [52]
There was a new version of MediaWiki last week. You can read a detailed log of all 763 changes. Most of them are very small and will not affect you.
Changes later this week
The new version of MediaWiki will be on test wikis and MediaWiki.org from 12 January. It will be on non-Wikipedia wikis and some Wikipedias from 13 January. It will be on all wikis from 14 January (calendar).
Is there any way to tailor edit notices so that they display differently for people editing with source editor vs. VisualEditor? If not, how would I go about trying to get that feature implemented? {{u|Sdkb}}talk13:42, 11 January 2021 (UTC)
VisualEditor editnotices add classes like ve-ui-mwNoticesPopupTool-item. I think we could choose a class name like novisualeditor for content which should be hidden in VisualEditor and say .ve-ui-mwNoticesPopupTool-item .novisualeditor {display:none;} in MediaWiki:Common.css. I don't know how stable it would be. PrimeHunter (talk) 14:34, 11 January 2021 (UTC)
What is the use case you trying to solve for? Edit notices should generally be consistent in "content" for any editor, and styling of notices should be consistent on the same browser regardless of the page they are coming from. — xaosfluxTalk14:39, 11 January 2021 (UTC)
Xaosflux, it's for this notice, which provides copy-and-paste text for adding a new entry to the list. Because VisualEditor still has no way to center-justify text within a table cell, there's unfortunately no way I know of for someone using it to edit the page currently without messing up the formatting, so I'm not sure exactly what I'd want to say to VE folks but it should be something different. {{u|Sdkb}}talk14:56, 11 January 2021 (UTC)
@Sdkb: seems like you'd want them to actually use that template if possible, which requires source editing. So instead of having that template wikilink to what a description of the source editor is, perhaps put in directions that to use that they should click to change to the source editor or something like that? — xaosfluxTalk15:24, 11 January 2021 (UTC)
Xaosflux, thanks, that's a good suggestion; I switched the link. It'd still be nice to have a switch function, since many beginners don't know which editor they're using. I think this would have at least some potential wider applicability, as editnotices of the type I made probably ought to appear on more pages than they currently do. {{u|Sdkb}}talk23:56, 11 January 2021 (UTC)
Temporarily watching pages
We now have a new option for temporarily watching pages, which I think is very useful.
I took a glance at the filter options and did not see a way to filter to see only temporarily watched pages. is there a way to do that that I am missing? If not, put someone who knows how to update the filter changes option add it to the list?--S Philbrick(Talk)17:28, 7 January 2021 (UTC)
Cool idea, but no, I am fairly certain that is not supported. You would need to ask for it on Phabricator. --Izno (talk) 20:53, 7 January 2021 (UTC)
@Sphilbrick: Are you talking about Special:Watchlist filters (i.e. to see only changes made to pages you watch temporarily)? I'm confused what the use-case is for that. Or, are you just looking for a way to view all pages you watch temporarily (i.e. to see which ones will expire soon)? — MusikAnimaltalk03:50, 8 January 2021 (UTC)
If the latter is the case, 'View and edit watchlist' (found on the watchlist) shows how much time is left for temporarily-watched pages. BlackcurrantTea (talk) 05:21, 8 January 2021 (UTC)
BlackcurrantTea, I don't see 'View and edit watchlist' I see 'view', 'talk', 'edit' and "edit your list of watched pages" but I don't think any of those are what you refer to. if you mean the last one it takes far too long to load. S Philbrick(Talk)14:52, 8 January 2021 (UTC)
Sphilbrick, in Monobook, the links across the top of the watchlist are (View relevant changes | View and edit watchlist | Edit raw watchlist | Clear the watchlist). I don't know if 'edit your list of watched pages' is equivalent to 'View and edit watchlist' or to 'Edit raw watchlist'; Edit raw watchlist doesn't show which pages are temporarily watched. However as it wouldn't show which of those have had recent changes, it might not be of use to you. BlackcurrantTea (talk) 15:48, 8 January 2021 (UTC)
> Are you talking about Special:Watchlist filters
Yes, when I click on the "filter changes" box, it brings up special: watchlist
I participated in the development of the temporary option, and my sole motivation was that I am active in CopyPatrol. when I make an edit as a result of identifying a copyright issue, I'd like to have that page of my watchlist for a little while but not forever. While I have to make that change manually, I'd like to see all of the, and only the pages that I've temporarily watchlisted to see if there's some subsequent activity on pages with copyright issues. In most cases, they involve brand-new users who may not know about contacting me. S Philbrick(Talk)14:49, 8 January 2021 (UTC)
A version of the watchlist with only temporarily-watched pages is a good idea. Aside from its utility for your CopyPatrol work, it seems like an option people would find a use for, if it existed. It would be a quick way to check on IPs I've left notes for. BlackcurrantTea (talk) 15:48, 8 January 2021 (UTC)
JJPMaster, hmm, that's a technical question. Some other non-WMF wikis might be more restrictive with the ability to read pages, and I know we restrict the ability to see some edit filters, so there's probably something deep down in the MediaWiki software that requires this to be articulated. If you need more information than that, I'd suggest going to WP:VPT. Cheers, {{u|Sdkb}}talk13:33, 12 January 2021 (UTC)
When I type the semicolon (;) into the search bar, I get taken to the Main Page. However, when I access the ; page using Special:EditPage, It correctly shows that ; is a redirect to Semicolon. If you click the link to ; in this message above, you will also get redirected to Main Page. Please help 122.60.65.44 (talk) 04:42, 12 January 2021 (UTC)
I was reminded that I was not strictly correct when I said that; it's that the first instance of the particular TStyles invocation cannot be used inside a link (as I left, somewhere? Template talk:Frac I think? Also noted on the phab task). Speaking of this, I need to revisit (it's on my list! Slightly busy this week). --Izno (talk) 07:02, 12 January 2021 (UTC)
Special:ExpandTemplates shows {{frac|6|1|2}} produces ​<span class="frac nowrap">6<templatestyles src="Screen reader-only/styles.css"/><span class="sr-only"> </span><sup>1</sup>⁄<sub>2</sub></span>. There is a non-breaking space. It varies whether it's rendered. A post at phab:T200704 says: "bug occurs if first instance of inner templatestyled template is in a link." That is consistent with my tests. PrimeHunter (talk) 08:48, 12 January 2021 (UTC)
Right, that gives me a workaround. In the infobox, I linked only "furlong", not the template, and now the template also works properly in the body of the article. Strange and stranger. Thanks for digging into this. Christhe spelleryack14:57, 12 January 2021 (UTC)
Maintaining a script
Hi guys. I'm a template editor locally (fawiki) and was looking for useful user scripts for editing and maintaining templates. Recently, i found Bracket Match script, copied the source to my local common.js, and it didn't work. I thought that it might caused by local and rtl language problems. Then it appeared not working on enwiki either and i guess it needs updating. As the script writer is semi-retired , can anyone please help me to get the updated script? thank you. --Jeeputer (talk) 18:43, 11 January 2021 (UTC)
The default syntax highlighter in either 2010 or 2017 wikitext editors usually does just fine for me for matching brackets. Try it out! --Izno (talk) 01:18, 12 January 2021 (UTC)
In the last 48 hours, I've received 26 email notifications that there were multiple failed attempts to log into my account. Why is there not some system preventing this? — Scott•talk02:43, 9 January 2021 (UTC)
Similar discussion is at the bureaucrats' noticeboard. It is not possible to stop people trying to log in at Wikipedia using your user name with a made-up password. If, say, a dozen attempts were to lock the account so it could not log in for 24 hours, trolls could lock us all out so we could never log in. In your preferences, there is a section "Notify me about these events" which you can use to not receive the notifications. With a reasonable password that you have never used on another website (and same for the email account used with Wikipedia), there is no need to worry about your password being cracked. Johnuniq (talk) 02:53, 9 January 2021 (UTC)
I found phab:T193769 "Thousands of failed login attempts (wrong password)" from the May 2018 incident. It contains a link to graphs of "Authentication metrics" which for January 2021 is [56]. For login errors, holding the mouse over 01/09 (today) shows wrongpassword around 20,000 at most times of the day. I suppose that is 20,000 per some unit of time—what unit? At any rate, it's shot through the roof today. Johnuniq (talk) 10:10, 9 January 2021 (UTC)
Just a guess but to me it looks like that is a raw number, i.e. 20000 in a 20 minute window, which means that the people behind this have been pretty busy and it is almost certainly automated. —Soap—10:33, 9 January 2021 (UTC)
I haven't had an attempt for a long while so don't know, but I did read a month or two ago a claim that the IP (or possibly the most recent IP) is shown in the email notification, if you have that enabled in preferences. Can someone confirm that? When I saw it, I think there was no IP mentioned in the on-wiki notification. Can someone confirm that? Johnuniq (talk) 10:07, 12 January 2021 (UTC)
Piotrus here is all I know ... dumber than I look ... copy of email from Wikimedia:
Wikipedia <wiki@wikimedia.org> Unsubscribe
Fri, Jan 8, 6:06 PM (4 days ago)
to SandyGeorgia
Someone (probably you, from IP address 2601:240:C401:FE80:4DC:3872:CD01:D109) requested a reset of your password for Wikipedia (<http://en.wiki.x.io/wiki/Main_Page>). The following user account is associated with this email address:
Username:
SandyGeorgia
followed by text suggesting I change my password. This is a different message than the failed attempts email which has happened to me during three different arbcases; that message does not mention the IP. (I will next go check my preferences per what Johnuniq, sorry, ipad typing, bits and pieces. SandyGeorgia (Talk) 15:23, 12 January 2021 (UTC)
I have my preferences set to receive notifications on failed login attempts, and login from an unfamiliar device. I cannot find where there may be a preference setting for notifications on password resets— perhaps that is standard? SandyGeorgia (Talk) 15:27, 12 January 2021 (UTC)
SandyGeorgia, Thanks, that is very useful. I think the system right now doesn't generate informative email notifications for unsuccessful login attempts. All I get is "There have been multiple failed attempts to log in to your account from a new device. Please make sure your account has a strong password." but there is no information on the IP. (I'd like to know what IPs are trying to log in to my account - there is also a tiny chance it is a bug and maybe one of my mobile devices lost a token or whatever and is trying to refresh or something, knowing what's going on would make this less stressful). Any idea where such a feature could be requested? PS. Also I assume privacy is not an issue, since IP is given for password resets, so why not for the failed login? Piotr Konieczny aka Prokonsul Piotrus| reply here08:59, 13 January 2021 (UTC)
I am on a wikibreak now, but I also received one of these "there have been multiple failed attempts" messages, on January 6. I was unsure what was happening but changed my password, just in case, and made it more secure. I did make a couple of posts at a talk page of an ongoing arbitration case recently. Possibly that's in part related to these attacks, as SandyGeorgia's post above indicates. Nsk92 (talk) 22:02, 9 January 2021 (UTC)
Bruteforcing accounts is a useless way to crack passwords, presuming you have a decent password. However, the software should not be permitting 34286 login attempts per hour. It should be rate limiting logins way shorter than that. 20 per 15 minutes, or something, seems far more reasonable. ProcrastinatingReader (talk) 10:56, 11 January 2021 (UTC)
Hi all,
In the article of Bodri , Chattisgarh, India, the map shows the location of the city outside the state it belongs to !!! Can someone fix it...... Anupam Dutta (talk) 12:24, 13 January 2021 (UTC)
In Wiktionary, the context box may be several pages long with almost empty space. Perhaps there might be a way to optionally expand the box with several columns?
This would make the contents box an overview, which is surely what the context box is all about.
Star Lord - 星爵 (talk) 13:49, 12 January 2021 (UTC)
See H:TOC for our options such as limit the levels, float to the right, or display horizontally. There is no multi-column option unless you code it by hand instead of relying on the automatic functionality. I once did it in RSA numbers. PrimeHunter (talk) 15:22, 12 January 2021 (UTC)
The Wikipedia:Good article nominations currently has a disambiguation link at the "You can help" in the backlog section – this should almost certainly go to Wikipedia:Good article nominations/Instructions#Reviewing though I have no idea how to access whatever template/subpage this backlog notice is in. Got no response on the GAN talk page, and the two admins I approached had no idea so I'm coming here. If anyone could help, it would be appreciated, since this is certainly my last shot at trying to resolve this... Aza24 (talk) 22:43, 13 January 2021 (UTC)
Okay, here's a puzzle for anyone who wants to bite. {{Empty section}} and {{Expand section}} both include a link to edit the page over "adding to it". However, unless the |section= parameter is filled out with a number (which it almost never is), the link goes to editing the page as a whole, rather than just the section. This is likely to overwhelm many newcomers with the existing wikitext (especially if there's an infobox), making the idea of editing less friendly.
Worse, even if a section parameter is specified, it may become wrong if a new section is added above it.
Any solution to this issue would be helpful, but I can't think of a way to code ourselves out of it, so a bot that fills out the section parameter and corrects it when needed may be the best answer. {{u|Sdkb}}talk17:59, 12 January 2021 (UTC)
I'll bite. Templates do not have any sense of where they are in their page, so getting the current section number is the problem. I cannot think if a realistic solution, but here are my blue sky thinking thoughts:
Create a module to use in the template — get the page source and scan it backwards from the place where the template is included. Count the headings to get the section number to input to the template
Ask in phab for a new Magic word that returns the current section number. Wait a few years. Add it into the template above Y
Ask in phab for an addition to the API, where action=edit§ion=32 can instead use the heading text action=edit§iontitle=Expanding_expand_section. The template can then use the section title instead of the section number
Ask in phab for an addition to the API, where action=edit§ion=32 can instead use a specific anchor action=edit§ionanchor=TODO-123 and have the {{Expand section}} output the anchor. The template can then use the anchor name instead of the section number
Would the issue with 1 be that it'd slow loading times? That or a magic word seems best. An issue with 3 would be that section titles can change. {{u|Sdkb}}talk22:25, 12 January 2021 (UTC)
The bot solution also risks leaving incorrect links when sections change, but it's likely worse in that case because users would open an edit link to a different section, rather than the whole page. (Already happens with edit conflicts, but this introduces a new maintenance burden on many pages, and would cause issues if the bot ever stopped running.) I can't think of a safe way to do it with Lua; how do you find your invocation where there are multiple expansion templates in the page? Magic word seems like the best option, though indeed, unlikely to be done any time soon. — The Earwigtalk08:08, 13 January 2021 (UTC)
Okay, I created the phab task, so the few-year countdown has begun... For the interim, do you think it's better to have the link or to have no link? The "edit" button for the section is guaranteed to be nearby, so I wonder if we'd be safe just removing it and trusting that people will be re-routed to that. {{u|Sdkb}}talk10:36, 13 January 2021 (UTC)
Option 5: Block adding non-number values. The insert template feature in the wikitext editor (2010 one) does that and in VisualEditor there will be an widget for numeric parameters phab:T124850. Those work for parameters specifieed as numberic in TemplateData. The bug does not state it, but the widget does verify that the input is a number, source of the widget is gerrit:c/oojs/ui/+/268375/6/src/widgets/NumberInputWidget.js.--Snaevar (talk) 23:49, 13 January 2021 (UTC)
Yes, but it is guaranteed that any manually entered section number will be wrong in the near future. I wonder if the empty section and expand section templates have been useful: has a meaningful number of people used the edit link they provide? Perhaps the best solution would be to remove the redundant wikitext: just tell people to click edit, and if they can't do that, is it likely that they will helpfully expand the section? For example, see Alkane#Environmental transformations which shows that the six-year-old edit link is not needed. Johnuniq (talk) 01:27, 14 January 2021 (UTC)
Add search by edit summary to Contributions search?
Would any of the Wikipedia programmers out there be interested in adding the ability to search contributions by the contents of edit summaries? I would certainly find it useful, and I believe other editors might as well. Beyond My Ken (talk) 22:03, 13 January 2021 (UTC)
Time-sensitive mobile accessibility help needed at WP:20
Resolved
This page is currently linked from the main page banner, but has not yet been made to display properly on mobile. Please see discussion and make comments at Wikipedia:Village_pump_(proposals)#Main_page_banner, where I fear it's getting swamped by all the other discussion. Severe display issues for half our readers on a page linked from the top of the main page is a pretty big issue and needs resolving as soon as possible. {{u|Sdkb}}talk13:37, 15 January 2021 (UTC)
Open this RecentChanges page (or the equivalent on Commons, which has more results) and click on any "diff". The diff table itself displays correctly but the rest of the page is treated like you're viewing the previous revision, including the preview below the diff, notice at the top, and the link destination of the "Edit" tab. Has this been reported on Phab? Nardog (talk) 14:32, 15 January 2021 (UTC)
The only way to restart it is to contact the maintainer or convince a toolforge system operator to restart it. Piling on will not resurrect it. The only other option is to take the tool and maintain it yourself. None of those can be solved here. --Izno (talk) 21:54, 13 January 2021 (UTC)
I think, also, the last time it went down, there was a discussion to incorporate the tool more under a general maintenance authority. Sorry, I don't know a whole lot about how tech responsibilities play out. It is an important tool, as it's something that welcoming editors encourage new editors to use. Thanks. Caro7200 (talk) 22:19, 13 January 2021 (UTC)
@Caro7200: I've been chewing over that off and on since the last failure - while I'm not suited to coding anything (I push the button and the computer go "boop" and it's REALLY COOL) I keep thinking about putting together a sort of "what do we want from the new ReFill" page so that we can request the best features from both it and Reflinks, along with anything else we can come up with. But it's a hugely important tool, and it becomes troublesome not having someone who can help look after it. --Ser Amantio di NicolaoChe dicono a Signa?Lo dicono a Signa.22:30, 13 January 2021 (UTC)