Wikipedia:Village pump (technical)

 Policy Technical Proposals Idea lab WMF Miscellaneous 

• Table of contents
• First discussion
• End of page
• New post

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

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

« Archives, 193, 194, 195, 196, 197, 198, 199, 200, 201, 202, 203, 204, 205, 206, 207, 208, 209, 210, 211, 212, 213

Main page on the Dutch Low Saxon Wikipedia

Hi there! Is there anyone who could assist me with setting up the mobile version of the Dutch Low Saxon main page? I've tried to set it up myself, but I can't seem to figure out how it works exactly. I also tried using an empty main page with 2 templates (consisting of only wikitext and div id) and 1 template (with a table), which is almost identical to the Frisian Wikipedia where it is displayed correctly. It would be highly appreciated. Servien (talk) 11:49, 20 October 2019 (UTC)

Servien, Frysian Wikipedia (like English) is on the old deprecated method of presenting the main page on mobile. The way to do it now, is to use CSS and make the page responsive to the dimensions of the viewport. —TheDJ (talk • contribs) 08:58, 21 October 2019 (UTC)

Thanks for the explanation, TheDJ. I'm not very familiar with CSS programming, and the 'Mobile hompage formatting' explanation is not very helpful. Is there anyone willing to assist with updating the main page, or is there a way I could still use the old methode? Nowadays one would almost need to be a programmer to design/localise the mainpage, not that practical if you only have a small community. Servien (talk) 22:10, 28 October 2019 (UTC)

I'll come and assist with this – Thjarkur (talk) 12:27, 1 November 2019 (UTC)

XFDcloser gadget?

Multiple people have suggested that my script XFDcloser be made into a gadget – see above #Preventing this in the future and Archive 176#Likelihood of User:Evad37/XFDcloser.js becoming a gadget?. Basically the script is an important part of the workflow for many users closing XFDs – it is imported by ~982 users (359 + 638 - 15). A community gadget, rather than a personal userscript, would hopefully be more stable and preemptively maintained when there are going to be breaking changes in the underlying mediawiki code.

A gadget version would still be limited to WP:extendedconfirmed users. Access could also be further restricted, if the community wants, via a whitelist like AWB (WP:AWB/CP).

Thoughts? - Evad37 [talk] 04:35, 27 October 2019 (UTC)

Pinging users from the above linked discussions: @MarioGom, Xaosflux, RL0919, Amorymeltzer, Nosebagbear, Steel1943, Headbomb, Izno, and SD0001: - Evad37 [talk] 04:35, 27 October 2019 (UTC)

• Support Certainly it feels like one of the key scripts - it shutting down generated the only backlog lasting more than a few hours AfD has had in a long while. Having some more effort put in to make sure it works pre-emptively with any change seems great. What I can't really judge on are the negatives, including the level of difficulty of getting it to that status and whether you not being able to directly alter it would be significant downsides. Nosebagbear (talk) 10:56, 27 October 2019 (UTC)
• Support XFDcloser is over 6000 lines of code, it stands to gain a lot from the ResourceLoader's caching and minimisation (see mw:ResourceLoader/Architecture), resulting in faster and bandwidth-efficient loads for users. Also, at present, the majority of users are loading the script after hopping over 2 script redirects: User:Z-man/closeAFD.js --> User:Evad37/XFDcloser.js --> User:Evad37/XFDcloser/v3.js, making things even less efficient. SD0001 (talk) 11:09, 27 October 2019 (UTC)
• Support for adding to a new section, like "Maintenance and Administration" or something, and putting a description of what access is needed to use each tool with it (e.g. "requires sysop", "requires pagemover"). — xaosflux ^Talk 15:27, 27 October 2019 (UTC)
• Support The gadget definition allows you to specify the required user rights, and the gadget won't be listed at all for those who don't have those rights (e.g. Twinkle isn't listed for unconfirmed users). So putting "requires sysop" or what have you in the description might be redundant. At any rate, I definitely support promoting XFDcloser to a gadget. It's a rather critical tool for maintenance, widely used, and this would improve both performance and maintainability. — MusikAnimal ^talk 18:03, 27 October 2019 (UTC)

Though xfdcloser can of course be used by non-sysops, so you don't want to limit to that, you'd just want to note that not all functions are available without being an admin Nosebagbear (talk) 09:47, 28 October 2019 (UTC)

The requirements to use XFDcloser should be the same as Twinkle ... or for extended confirmed. For example, non-admins (such as myself) use XFDcloser quite often. The one/only restriction is that I technically can not select the "delete" option for obvious reasons. Steel1943 (talk) 17:18, 29 October 2019 (UTC)

• Support as the initiator of Wikipedia:Village pump (technical)/Archive 176#Likelihood of User:Evad37/XFDcloser.js becoming a gadget?. I've given this some more thought, and since I no longer see the need for non-interface administrators to edit the script, I'm totally on board with this. Steel1943 (talk) 17:03, 29 October 2019 (UTC)
• Support. As essential as Twinkle. I don't recall any instances of abuse despite its wide use, so would think no further permissions-based restriction is warranted at this time. czar 16:00, 3 November 2019 (UTC)
• Support per above, but also make DiscussionCloser a gadget too. -DM

Edit screen problem at Laguna_del_Maule_(volcano)

When I was trying to edit Laguna_del_Maule_(volcano) to add information from [1], the edit screen (including the dropdowns and the input box) stopped responding to input with a thin gray bar appearing just below the button bar. Scrolling did still work but the bar on the right no longer moved. Jo-Jo Eumerus (talk, contributions) 17:10, 30 October 2019 (UTC)

Seems like that problem is now resolved. Jo-Jo Eumerus (talk, contributions) 19:33, 31 October 2019 (UTC)

Nevermind, it has reappeared. Jo-Jo Eumerus (talk, contributions) 11:35, 1 November 2019 (UTC)

Added a screenshot; it seems like part of the interface has become disconnected from the actual edit box - on only that article. Jo-Jo Eumerus (talk) 13:05, 1 November 2019 (UTC)

That looks like you either a) have a personal script causing an issue or b) your Internet browser causing an issue. What is your browser/operating system to start and/or what are your scripts? --Izno (talk) 03:16, 2 November 2019 (UTC)

Try this safe mode edit link to see if it works without scripts. — xaosflux ^Talk 03:46, 2 November 2019 (UTC)

Firefox 70.0.1 (64-Bit), aside from the Wikipedia defaults I only have syntax highlighter, Twinkle and the user scripts. I notice now that the syntax highlighting changes from a yellow-grey scheme such as in the screenshot into a pink-blue-green one within a second from opening the edit page, and that it now works. Jo-Jo Eumerus (talk) 10:04, 2 November 2019 (UTC)

Jo-Jo Eumerus, I suspect the syntax highlighter is the cause. I recommend toggling that off, and seeing if the problem disappears instantly. Whatamidoing (WMF) (talk) 15:06, 4 November 2019 (UTC)

Huh. This particular problem disappeared after toggling all preferences off and then some of them - including syntax highlighter - back on. Now the only problem is that the syntax highlighting is transposed to above the edit box, but otherwise it is normally editable. Jo-Jo Eumerus (talk) 15:09, 4 November 2019 (UTC)

Centring a file in a HTML table?

Note: When writing Wiki Markup and HTML which I don't want to display, I have changed some characters to different ones, which means code won't be exact. Check at the bottom of my post for a key.

Hello everybody, I'm encountering an irritating problem whilst editing my user page. I have these tables on my page with only one row/column which I just use to display information and I was using one to display a selection of images from Wikimedia Commons which caught my attention. When I added an image into it using [[File]] (the exact code was [[File:Russell Falls 2.jpg|thumb|[[Russell Falls]],Australia]]), it wasn't centred inside the table. I did some research and found out that you can change the alignment, which I tried, but it didn't change anything as the default was middle in the first place. Here is my full code plus a screenshot of what I'm getting:

<table align="center" style="height: 44px; border-color: #03fc6b; background-color: #59ff9e;" border="5" width="263" cellspacing="5" cellpadding="5">
<tr>
<td style="width: 231px;">
<p style="text-align: center;">&nbsp;<span style="font-family: Oswald;">PHOTOS THAT CAUGHT MY ATTENTION</span></p>
<p style="text-align: center;"><span style="font-family: Lato;">Here is a small selection of photos from Wikimedia Commons that caught my attention: [[File:Russell Falls 2.jpg|thumb|middle|[[Russell Falls]], Australia]]</span></p>
</td>
</tr>
</table>

Screenshot

If anybody can throw any light on this issue, that would be greatly appreciated. Thank you!

Key: (no longer needed) ~Xaeman (talk) 17:34, 30 October 2019 (UTC)

Try: "td style="width: 231px; text-align: center" --Jorm (talk) 17:38, 30 October 2019 (UTC)

The image parameter "middle" is for vertical alignment. You want "center" after "thumb" instead. See mw:Help:Images. – Jonesey95 (talk) 19:17, 30 October 2019 (UTC)

Thank you Jonesey95 - this worked, when I came back to my page it had already been changed, I don't know if this was someone helping me out or I just did it already without realising. Thanks anyway! ~Xaeman (talk) 16:21, 31 October 2019 (UTC)

Wikipedia alerts and notices

Every time I log into Wikipedia, my alerts and notices icons at the top of the display are highlighted in colour to indicate that I have new alerts and notices. This is almost never true. I view my alerts and notices, see that there are no new ones, only a list of old ones, and select "Mark all as read." Neither viewing my alerts and notices nor selecting "Mark all as read" is effective in suppressing them. I do not wish to waste my time checking old alerts and notices every time I log in. Can this error be fixed, please? — O'Dea (talk) 19:28, 30 October 2019 (UTC)

@O'Dea: sometimes they get stuck, try going to Special:Notifications and clearing them out. — xaosflux ^Talk 19:41, 30 October 2019 (UTC)

What action does "clearing them out" refer to? The only option presented is to mark them, one by one, as read, but I have already marked-all as read. Thanks. — O'Dea (talk) 19:45, 30 October 2019 (UTC)

@O'Dea: yes, clicking on the "Unread" tab and clicking through any that are in there. If the icon is still lit after that, log out, clear your cache, and log back in please. — xaosflux ^Talk 19:51, 30 October 2019 (UTC)

No notices appear in the "Unread" tab. As to clearing my cache, I log in from multiple devices at different locations, and clear my caches regularly, so this continuing problem is not caused by cached data or persistent log in; I log in quite often and see these attention-demanding notices. To clarify, right now in this session, no outstanding notices appear because I marked them as read at the beginning of this Wikipedia session, but they will certainly be there when I log in to a new session. — O'Dea (talk) 19:54, 30 October 2019 (UTC)

Is it always the same notices? Are the notices from the project you are currently on, or from another project (e.g. commonswiki)? — xaosflux ^Talk 20:01, 30 October 2019 (UTC)

Old notices from a variety of projects are repeated. Valid new notices appear occasionally but the old ones that I have read, and flagged as read, are highlighted anew upon logging in. — O'Dea (talk) 20:46, 30 October 2019 (UTC)

Are they "very old"? I had that problem once and had to go to each project to clear the notices, then they stayed away. — xaosflux ^Talk 22:20, 30 October 2019 (UTC)

@Xaosflux: They were a mixture of very old, old, and new. Today, I logged in after a few days' absence and saw old alerts. I flagged them as read, logged out, then logged back in and now they're gone. I'll monitor them in coming days. I am not active every day at the moment. Thanks for your interest. — O'Dea (talk) 23:06, 3 November 2019 (UTC)

@O'Dea: When you click the bell icon, the popup shows a list comprising three groups as follows (in this description, an "unread alert" is one which shows the blue dot at upper right of its entry, and a "read alert" is one which shows the empty grey circle at upper right). Group 1 comprises the unread alerts from the same wiki in reverse chronological order (newest at the top), to a maximum of 25. Group 2 comprises summaries of unread alerts on other WMF wikis, titled "More alerts from another wiki", one per wiki. Group 3 comprises read alerts from the same wiki, again in reverse chronological order. The total number of entries in groups 1 and 3 taken together cannot exceed 25, so if there are 25 unread alerts from the same wiki, no read alerts will be listed. Read alerts from other wikis are not shown. So if you have three unread alerts on this wiki and two on Commons you will get four entries with blue dots followed by 22 with the empty circle. Clicking a blue dot of a group 1 entry moves it to group 3; clicking an empty circle of a group 3 entry moves it to group 1. --Redrose64 🌹 (talk) 12:42, 4 November 2019 (UTC)

Not sure if this is related, but today I got an old alert highlighted when someone archived the talk page the alert came from. I've never noticed this behaviour in the past.   Jts1882 | talk  10:41, 7 November 2019 (UTC)

Comment highlighting script issues

I find PleaseStand's comments-highlighting script (User:PleaseStand/highlight-comments.js) quite useful to quickly find a ping and/or the part of a long discussion I was involved with. Quite a while ago its functionality changed from highlighting whole lines to highlighting just links within a line. Now it has stopped highlighting anything. Anyone have any insight as to why this may be (or, more relevant, how to fix)? — Rhododendrites ^talk \\ 03:15, 1 November 2019 (UTC)

It would have stopped working for the same reason as #mediawiki.RegExp scripts not wokring, and will have the same fix - Evad37 [talk] 06:59, 1 November 2019 (UTC)

Default keyboard

Lately something funny happened to my keyboard (only in WP). Although there is no problem in copy and paste method, I can't write certain characters (like double brackets for links ) in WP. Any suggestion to return to default keyboard? Thanks Nedim Ardoğa (talk) 08:24, 1 November 2019 (UTC)

Nedim Ardoğa, when you type those characters, do they turn into some other (strange) character? If so, then you might have accidentally enabled the IPA (software) keyboard. Whatamidoing (WMF) (talk) 15:14, 4 November 2019 (UTC)

Suggestion: substall

The subst parser function is useful. It would sometimes be useful to have a substall function too. It would subst a given template and all templates within that template. Just granpa (talk) 11:02, 1 November 2019 (UTC)

This request is phab:T4777 from 2005. Special:ExpandTemplates can do it with some copy-pasting. I agree it would be nice if wikitext could also do it. PrimeHunter (talk) 11:39, 1 November 2019 (UTC)

Thank you. Just granpa (talk) 12:07, 1 November 2019 (UTC)

Template:infobox person/Wikidata not working as intended

Resolved

Hi there!

I wrote a short biography of Michèle Ray-Gavras and I was planning to use the data from Wikidata to populate the infobox (since this is what the French version of the article does). However, It seems that only the photo was pulled from the database. Does anyone know how this template is supposed to work? -- Luk ^talk 13:36, 1 November 2019 (UTC)

From the documentation for that template: This will by default only fetch information that is sourced on Wikidata. This person's infobox-related data on Wikidata has no sources. The template is working as designed. – Jonesey95 (talk) 14:08, 1 November 2019 (UTC)

I'm responsible for Journalist appearing in the box. I was curious and added a ref. - X201 (talk) 14:19, 1 November 2019 (UTC)

You are correct! When going through the documentation, I thought using fetchwikidata=ALL would overwrite this behaviour. Thanks! -- Luk ^talk 14:31, 1 November 2019 (UTC)

Two issues (Javascript and Wiki source editor)

Greetings. I have two issues and the following are those

1. I enabled JavaScript Wikipedia:Recovering from Wikipediholism/reminder some weeks ago. Its working was smooth and didn't cause any trouble. But recently when I edited this, this problem happened. I tried to rearrange the sub-headings, published and went offline. When another editor reverted the problematic edit and informed me of that. The references were modified into this link. (skin: default vector.js, browser: Opera, browser version:63.0.3368.107, operating system: Windows , and operating system version: 10 Pro)
2. When I edit using source mode (wikitext editor) and when the syntax highlighting option is enabled, sometimes the text is entered in another place far from the blinking cursor. This is a problem only when I edit with syntax highlighting option enabled. So I turned off the option but I am curious why is it happening.Beastranger (talk) 05:19, 2 November 2019 (UTC)

Template:internetquelle compatibility with WP:CXT

Wasn't able to get a reply at Wikipedia talk:Content translation tool so reposting here:

Can anyone point me in the direction of what needs to be done to convert de:Template:Internetquelle (a dewp citation template) for use with WP:CXT? It looks like CXT should be able to convert the template directly into {{cite web}} but I don't know what needs to be linked for that to happen. Right now, enwp's {{internetquelle}} is meant to be subst'd to manually convert the German parameters to English, but isn't this something that TemplateData could do? Let me know if this question would be better handled elsewhere.

czar 15:34, 2 November 2019 (UTC)

The relevant bit of documentation seem to be mw:Content translation/Templates. The key part is setting up alias in the TemplateData for the destination template. TemplateData would look something like

<templatedata>
{
	"description": "cite web English translation from German",
	"params": {
		"author": {
			"aliases": ["autor"],
                        ....
                 },
                 ......
         }
}
</templatedata>

If you want to experiment I'd recomend setting up another template something like {{cite web trans from de}} rather than working directly with {{cite web}} as you will run into problems with protection settings. --Salix alba (talk): 18:14, 2 November 2019 (UTC)

Here's what I've found thus far. As an example, the French fr:Template:Ouvrage is linked via Wikidata to our Template:Cite book. When translating a French article with CXT, it appears to use the French template's Wikidata in which, for example, |year= is linked as an alias for |année=. My understanding is that this works because "year" is indeed an alias in that template, otherwise it would be incorrect to simply add aliases that are not supported in the template code. Then since both fr:Template:Ouvrage and Template:Cite book have |year= parameters, CXT is able to make the connection.

I attempted this myself in the template in question de:Template:Internetquelle >> our Template:Internetquelle. Since both versions use the same German parameters, by adding TemplateData to our enwp template, CXT now supports the citation's display (within the tool). Now, ideally, I would like to convert the German template directly into {{cite web}} (rather than our own version of {{internetquelle}}) but I'm not sure if we're hosting a version of the German template for any other reason or if we're prepared to add the German aliases directly into the {{cite web}} code (hence my original question). In the meantime, a bot should automatically convert/subst my local invocation of {{internetquelle}} per its documentation so this might work fine, though it appears jury rigged as a permanent solution. czar 14:48, 3 November 2019 (UTC)

Pinging User:Amire80 about the CX template-matching question. Whatamidoing (WMF) (talk) 15:19, 4 November 2019 (UTC)

Setting up auto-archiving

Resolved

It does not seem to work on Talk:Paleotempestology, despite the presence of the archive template. Jo-Jo Eumerus (talk) 12:24, 3 November 2019 (UTC)

This edit by Czar (talk · contribs) may fix it. Of course, there were several changes there, any one of which may be the actual cure. My bet is on the parameters that have no value, such as |key= - I think that the bot doesn't fall back gracefully to default values. --Redrose64 🌹 (talk) 20:55, 3 November 2019 (UTC)

Apparently this is working now. --Redrose64 🌹 (talk) 15:16, 4 November 2019 (UTC)

Access via mobile phones

Is there a working party looking specifically at problems associated with gaining access via mobile phones generally? Xyl 54 (talk) 23:46, 3 November 2019 (UTC)

Xyl 54, try mw:Reading/Web/Mobile. BlackcurrantTea (talk) 09:44, 4 November 2019 (UTC)

Xyl 54, which kind of access are you interested in? For readers, for editors, for a particular use? Whatamidoing (WMF) (talk) 15:21, 4 November 2019 (UTC)

Thank you both for replying. To answer your question; neither, really. I had a conversation with an editor from another project who'd had trouble loading an article introduction to their mobile phone, because it was too long. I hadn't considered that mobile users could have trouble with the standard layout here and wondered if anyone was looking into such matters. I searched for a project page but couldn't find one, so I thought I would ask here. Xyl 54 (talk) 22:53, 5 November 2019 (UTC)

PS: Also, I've been told there is a condensed version of WP provided by google; can you tell me what (or where) that is? Xyl 54 (talk) 22:54, 5 November 2019 (UTC)

Unexpected cite page feature

Edit an article using the (old) source editor, navigate the toolbar to the web citation screen, paste "Mystery Science Theater 3000" without the quotes into the URL field, click the search button and see what happens... Daß Wölf 00:51, 4 November 2019 (UTC)

I think it's getting it from Google or perhaps Google Books. [2] Killiondude (talk) 01:00, 4 November 2019 (UTC)

I thought it would be something cool, like adding some theatre seat and robot silhouettes to the bottom right of my screen. Instead it adds a citation to a 1999 book by Henry Krips on fetishes and psychoanalysis. Wug·a·po·des​ 08:29, 4 November 2019 (UTC)

I don't think I would've complained about that. Daß Wölf 22:56, 7 November 2019 (UTC)

Help with template gobbledygook

After butting my head on the screen for a few hours (spread over various days), I throw my hand in the air — why doesn't this work here?? (emphasis added)

{{#switch: {{Continent/Core | {{{1}}} | {{{2|simple}}} }} | {{#if: {{{table|}}} | style="text-align: center; background: Tan; {{{style|}}}" {{!}} }} Africa | Asia | North America | South America | Oceania | Antarctica | Europe = {{Continent/Core|{{{1}}}|{{{2|simple}}} }} }}

— Preceding unsigned comment added by Guarapiranga (talk • contribs) 04:26, 4 November 2019 (UTC)

Could you give more detailed information about what you want to implement? Your highlight part belongs to the switch condition. Thus, if table=1, it will require the output of {{Continent/Core | {{{1}}} | {{{2|simple}}} }} being style="text-align: center; background: Tan; {{{style|}}}" {{!}} Africa (instead of simply Africa) to give a result of {{Continent/Core|{{{1}}}|{{{2|simple}}} }}. This looks odd and possibly not what you want. --WhitePhosphorus (talk) 05:51, 4 November 2019 (UTC)

Unified view of a user's activity?

Is there a simple way to get all of, 1) contributions, 2) deleted contributions, and 3) Special:Log for a user, in a unified, chronologically ordered, list? When investigating user behavior, having to flip back and forth between those three pages, and manually correlate the timestamps, is a pain. -- RoySmith (talk) 15:01, 4 November 2019 (UTC)

@RoySmith: would a user script help? DannyS712 (talk) 06:50, 5 November 2019 (UTC)

Does such a script exist? -- RoySmith (talk) 16:26, 5 November 2019 (UTC)

Tech News: 2019-45

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

• At Special:Contributions you could see up to 5000 edits at the same time if you edited the URL. This has been lowered to 500. This is to stop requests which break the sites. [3]

Changes later this week

• MediaWiki:ipb-default-expiry can set the default length to block a user for your wiki. You will be able to use MediaWiki:ipb-default-expiry-ip to set a different default block length for IP editors. [4]
• The new version of MediaWiki will be on test wikis and MediaWiki.org from 5 November. It will be on non-Wikipedia wikis and some Wikipedias from 6 November. It will be on all wikis from 7 November (calendar).

Meetings

• You can join the technical advice meeting on IRC. During the meeting, volunteer developers can ask for advice. The meeting will be on 6 November at 15:00 (UTC). See how to join.

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

16:47, 4 November 2019 (UTC)

Google Code-In will soon take place again! Mentor tasks to help new contributors!

Hi everybody! Google Code-in (GCI) will soon take place again - a seven week long contest for 13-17 year old students to contribute to free software projects. Tasks should take an experienced contributor about two or three hours and can be of the categories Code, Documentation/Training, Outreach/Research, Quality Assurance, and User Interface/Design. Do you have any Lua, template, gadget/script or similar task that would benefit your wiki? Or maybe some of your tools need better documentation? If so, and you can imagine enjoying mentoring such a task to help a new contributor, please check out mw:Google Code-in/2019 and become a mentor. If you have any questions, feel free to ask at our talk page. Many thanks in advance! --Martin Urbanec 07:28, 5 November 2019 (UTC)

Bot to replace highjacked links

Is there a bot to disable hijacked links in refs/ELs that may now be unfit/malicious? At Wikipedia:Teahouse#Hijacked site with possible malware, Lyndaship has reported that www.clydesite.co.uk has moved to clydeships.co.uk, and the new www.clydesite.co.uk host may be malicious (certainly misleading). The layout of the new site has changed, so it's not an easy replace of the hostname (though maybe editors could help construct a translation table?). Meanwhile, disabling the links would probably be reader-friendly. Is there a bot or some other means to handle this? —[AlanM1(talk)]— 23:07, 5 November 2019 (UTC)

@GreenC: poke. --Izno (talk) 01:11, 6 November 2019 (UTC)

@AlanM1: Agree on malicious domain: http://www.clydesite.co.uk/clydebuilt/viewship.asp?id=2467 it is a sketchy domain. Normally would set the domain to Blacklisted in the IABot interface and run IABot and archive them all (since we don't know the new URLs second option is archive). But they will also require adding |url-status=usurped, to suppress displaying the malware link, which is custom coding. WaybackMedic can do it. Post/move this request to WP:URLREQ. Might also edit filter request. -- GreenC 01:44, 6 November 2019 (UTC)

@GreenC: Thanks for the quick response. I'll explore a little more to find out whether a translation table is reasonable before continuing. —[AlanM1(talk)]— 01:55, 6 November 2019 (UTC)

Query issue

I'm finding a lot of my old queries that I like to re-run periodically broken by sintax changes on Quarry. I must profess myself to be an utter noob and in need of help fixing it (I had help putting them together in the first place). The query I'm trying to fix is this one.

It is essentially a way of sorting top-wikipedian users by number of edits and other criteria (to help in narrowing the search for new editors for New Page Patrol). Another criteria that I'd like to add if possible would be X edits to Wikipedia: namespace in the last N days (like 100 in the last 6 months or something).

Any help is appreciated. — Insertcleverphrasehere ^{(or here)}_{(click me!)} 21:37, 6 November 2019 (UTC)

Your query was broken by the Actor migration. Updating the query to use the proper actor table, producing

SELECT user_id, user_name, user_editcount, user_registration
FROM `user`
JOIN actor_user on actor_user=user_id
WHERE user_editcount >= 75000
  AND user_editcount <= 80000
  AND user_registration < 20181105000000
  AND EXISTS (SELECT 1 FROM user_groups WHERE user_id=ug_user AND ug_group = 'extendedconfirmed')
  AND NOT EXISTS (SELECT 1 FROM user_groups WHERE user_id=ug_user AND ug_group = 'sysop') -- administrator
  AND NOT EXISTS (SELECT 1 FROM user_groups WHERE user_id=ug_user AND ug_group = 'patroller') -- new page patroller
  AND NOT EXISTS (SELECT 1 FROM logging_logindex
                  WHERE user_name = log_title
                    AND log_type = 'block' AND log_action = 'block'
                    AND log_timestamp >= 20181105000000)
  AND (SELECT COUNT(*) FROM logging_userindex WHERE actor_id=log_actor AND log_type = 'move' LIMIT 5) >= 5
  AND (SELECT COUNT(*) FROM revision_userindex
       WHERE rev_actor = actor_id
         AND rev_timestamp >= 20181105000000
       LIMIT 500) >= 500
  AND (SELECT COUNT(*) FROM revision_userindex
       WHERE rev_actor = actor_id
         AND rev_timestamp >= 20191005000000
       LIMIT 500) >= 50
  AND (SELECT COUNT(*) FROM revision_userindex
       JOIN page ON page_id = rev_page
       WHERE rev_actor = actor_id
         AND rev_parent_id = 0
         AND page_namespace = 0
         AND page_len >= 2000 -- length in bytes
       LIMIT 2) >= 2;

should work as a fix. * Pppery * _{it has begun...} 21:55, 6 November 2019 (UTC)

Pppery, Cheers! — Insertcleverphrasehere ^{(or here)}_{(click me!)} 22:03, 6 November 2019 (UTC)

Pppery, seems to work again. Any idea how I could add that additional criteria though? (Also sorting by # edits to a particular namespace?) — Insertcleverphrasehere ^{(or here)}_{(click me!)} 01:29, 8 November 2019 (UTC)

Developer environment for this wiki

I have raised a question here about creating developer-friendly images of each wiki, which people can use to create scripts and mediawiki extensions on their own PC without needing to request an account on Wikimedia Tools or their sister services. I hope an image can be available for install in one click. If you are experienced in how to make this happen, or know how it would be useful for this wiki, please respond either here or there. Thanks. --Gryllida/ (talk) 23:00, 6 November 2019 (UTC)

How to create a page with ALL CAPS name over existing redir

I guess this is technical, not general, so reposting here...

I'm writing a page that should be named "SNAP (programming language)". However, there is already a page called "Snap! (programming language)". It appears the system (or a user?) has created a redirect without the !, "Snap (programming language)".

I'm lost how to create a new page with the ALL CAPS name. When I try, wiki sends me to the redirect and then to Snap!. I can click back from the redir link, but that has sentence casing.

Maury Markowitz (talk) 15:29, 7 November 2019 (UTC)

Hey Maury. When you are redirected to the other page, there should be small text in the upper left hand corner that reads "redirected from REDIRECT-NAME". If you click on that link, you will be taken to the redirect itself, which you can then edit as you like. GMG^talk 15:31, 7 November 2019 (UTC)

As I noted in the post above, that sends me to the version with lower case. Try it yourself and you'll see what I mean. Maury Markowitz (talk) 15:34, 7 November 2019 (UTC)

https://en.wikipedia.org/wiki/SNAP%20(programming%20language)?redirect=no , even if SNAP (programming language) weren't a red link, should get you where you want to go. If it does not, then there's something wrong on your end, not ours. --Izno (talk) 15:39, 7 November 2019 (UTC)

That said, if your page is intended to be a replacement of the other, I might recommend editing the existing page (assuming it's the same topic) and then moving it to where you want it to go. --Izno (talk) 15:40, 7 November 2019 (UTC)

Oh yea. As Izno says. You can just edit the text of the url in the address box to say SNAP instead of Snap and it will take you there just fine. GMG^talk 15:41, 7 November 2019 (UTC)

I'm sorry, but expecting the users to edit URLs to get to the page when the en.wiki keeps redirecting them to a different one does not strike me as either a problem "on my end" nor a useful solution. The question remains, why is the wiki redirecting me to the lower-case version when I explicitly type in the upper case one? There is no upper-case version, so it does not appear this redirect should fire given the way page naming is supposed to work (ie, all caps is not the same). Maury Markowitz (talk) 15:52, 7 November 2019 (UTC)

This is how the software treats all all-caps titles entered into the search window. If you enter ABRAHAM LINCOLN into the search window, it will redirect you to the article for Abraham Lincoln, even though there has been no manually created redirect from all caps to normal case. If you type ABRAHAM LINCOLN as a wikilink, it will make a red link and take you to the non-existent page. Alternatively, if you go to Abraham Lincoln and manually change the url to say ABRAHAM LINCOLN instead, it will also take you to the non-existent page. GMG^talk 15:55, 7 November 2019 (UTC)

• ? (edit conflict) There appear to be three pages being referenced here:

  SNAP (programming language)

  This page does not exist, it could be created or another page could be moved to this name

  Snap! (programming language)

  An article

  Snap (programming_language)

  A redirect

• It is important to know what you end result will be to best direct you. Are you wanting to change the page name of the current article, or write an article about something else with a very similar name? — xaosflux ^Talk 15:43, 7 November 2019 (UTC)

I typed this, exactly, into the search box "SNAP (programming language)". This takes me through the redir. There is no obvious way to get to the redline page otherwise. What am I missing here? Are we to type in a manual URL? Maury Markowitz (talk) 15:48, 7 November 2019 (UTC)

The search box is programmed to help users find the right topic, no matter the letter case and even when it's misspelled (like without the "!"). But if you enter the phrase "SNAP (programming language)" and then click "Search" and not press Enter, it will take you to the search results where you can find the red link to edit the page. Regards SoWhy 15:56, 7 November 2019 (UTC)

Webcitation redux

resolved

There was a proposal to deprecate webcitation, since it had stopped accepting new pages to archive.

Although webcitation.org is no longer accepting new pages, it continues to serve pages archived in the past.

However Citation_bot is breaking those working links to the archives. It seems to me this is not what was agreed to in the proposal. I thought a bot was going to (1) search out all links to pages archived by webcitation.org; (2) determine if the page was also archived at archive.org; (3) if the page in question was also archived at archive.org, or, if still live, whether it could be archived at archive.org, then replace the webcitation.org url with an archive.org url.

I don't see anything in the proposal that authorized breaking working urls to archived pages.

I thought if there was no replacement at a better archive the working urls to the webcitation.org archive would be left, as is.

I went to User_talk:Citation_bot/Archive_18#WebCite_query_strings. Near as I can understand GreenC justifies this excision of working archive links because he or she has found vandals, or naive but misguided good faith contributors, have used links to webcitation so they could serve references to pages at sites we had decided to blacklist.

1. But couldn't a vandal use this technique with links to archive.org, just as easily?
2. This wasn't what was agreed to at the proposal, which only took place a month or so prior to the second discussion.

Webcitation was the first archive server I discovered, and I used it exclusively, for years, until I discovered archive.org. So, prior to this bot breaking links to it, there may have been as many as a thousand perfectly valid instances where I used it, that still pointed to it. This represented a meaningful investment of my valuable time. So, I am not happy that this bot was authorized to break those links. Of course I would have to live with that unhappiness, if a central discussion authorized this. It does not seem that this was the case.

I sure hope citationbot was replacing those links with archive.org links, when this was possible.

A complicating factor is that webcitation offered users a choice - the archive-url it returned could either be a short, but obfuscated unique link, or it could return a longer, non-opaque link, that encoded the original pages url into it, just as archive.org does. It seems to me that GreenC's justification that vandals could hide bogus improper links in archive-url fields would only be true for those short, obfuscated links.

Well, I never used the short, obfuscated links webcitation offered. I only rarely came across anyone else using them.

I'd like a status report - has this bot broken every single working link to a webcitation archive? If not I suggest the bot should have this feature disabled, immediately.

If possible, I'd like to see a bot go through all the edits where citationbot broke perfectly valid, working links to the long, non-obfuscated webcitation archives, and restore those that weren't updated with a link to archive.org. Geo Swan (talk) 19:10, 7 November 2019 (UTC)

Geo Swan, you could report this to Citation bot but AFAIK it no longer messes with Webcite links. @AManWithNoPlan:. Also you are misunderstanding the discussion at User_talk:Citation_bot/Archive_18#WebCite_query_strings I was not "justifying the excision of working archive links". -- GreenC 03:04, 8 November 2019 (UTC)

citation bot broke nothing. All it did was remove a human readable comment from the url. It was never actually used by webcitation. It turns out there are supposedly bots that add these and verify them, and they exist solely so people using Wikipedia can see what the archive is. AManWithNoPlan (talk) 03:30, 8 November 2019 (UTC)

webcitation links contain a time stamp and nothing else. Deep in their archive they simply never archived two pages at exactly the same moment. Very simple and yet very stupid. AManWithNoPlan (talk) 03:33, 8 November 2019 (UTC)

• So, I got my knickers in a knot over something cosmetic... I'd still say the removal was a lapse from the principle "if it ain't broke, don't fix it..."

• Thanks to everyone who responded. Geo Swan (talk) 03:38, 8 November 2019 (UTC)

Bad separators on multiple pages

Tracked in Phabricator
Task T233649

• Update: many of these appear to be realted, see phab:T233649 for the master ticket. — xaosflux ^Talk 01:44, 8 November 2019 (UTC)

What's with the semicolons in the watchlist?

Tracked in Phabricator
Task T237685

I just noticed today that titles of pages in my watchlist end with a semicolon. I'm a big fan of semicolons, used correctly; this one, I have to say, just looks weird. When did it start, and what was the rationale? --Trovatore (talk) 20:45, 7 November 2019 (UTC)

WP:ITSTHURSDAY - seems to be:

.mw-title::after {
    content: ';\00a0';
}

code with a bug? Will open a phab ticket. — xaosflux ^Talk 20:48, 7 November 2019 (UTC)

phab:T237685 opened. — xaosflux ^Talk 20:52, 7 November 2019 (UTC)

Thanks! --Trovatore (talk) 20:53, 7 November 2019 (UTC)

@Jdlrobson: can you take a look at that ticket? — xaosflux ^Talk 21:48, 7 November 2019 (UTC)

very curious about this one - I'm a fan of consistency where ever possible and I'm not sure why the semicolon wouldnt appear here but would appear on history. What is the problem it is solving on history but not solving on watchlist? Jdlrobson (talk) 01:39, 8 November 2019 (UTC)

Missing separator

Resolved

 – Unable to duplicate this part - possibly related to other sections. — xaosflux ^Talk 01:22, 8 November 2019 (UTC)

On a perhaps related note, a few minutes ago I lost the space between page title and revision time, e.g.

m Meteotsunami‎22:37 . . (-35‎) . . ‎LizardJr8 ( talk | contribs )

Note the missing space between "Meteotsunami" and "22:37". Daß Wölf 22:54, 7 November 2019 (UTC)

@Daß Wölf: is that on the watchlist? What interface language are you using, what skin are you using? Do you have any special watchlist settings enabled? — xaosflux ^Talk 00:49, 8 November 2019 (UTC)

@Xaosflux: yes, it's the watchlist, forgot to write that. Vector skin, JavaScript watchlist. Of all the settings that look like they might have impact, the ones I have enabled are unread changes in bold, and the subtle update marker. Daß Wölf 01:04, 8 November 2019 (UTC)

Actually, it seems to have gone away. Now for the separator I have a semicolon and a space (e.g. "Meteotsunami; 22:37". I don't remember if the semicolon was there before. Daß Wölf 01:05, 8 November 2019 (UTC)

@Daß Wölf: thanks, I hadn't been able to break it like you did - was going nuts :D The semicolon appears to be new (and unwanted), from the parent section of this above. — xaosflux ^Talk 01:10, 8 November 2019 (UTC)

Bad semicolon in history

Tracked in Phabricator
Task T237705

@Xaosflux: Again unrelated, I just went to thank you for this edit and a new semicolon has popped up in the article history: "(cur | prev ) ¤ ¤ ; 01:10, 8 November 2019‎" - bolded for clarity. Now that one I'm sure wasn't there! Daß Wölf 01:13, 8 November 2019 (UTC)

@Daß Wölf: phab:T237705 is open for the bad semicolon here. — xaosflux ^Talk 01:21, 8 November 2019 (UTC)

Yep, that's what I see too, thanks. Daß Wölf 01:35, 8 November 2019 (UTC)

Note, in come cases a double set of bad leading semicolons are also appearing, updated the phab ticket. — xaosflux ^Talk 03:54, 8 November 2019 (UTC)

Missing separators in page history

Resolved

 – This issue is no longer presenting. — xaosflux ^Talk 03:53, 8 November 2019 (UTC)

Parenthesis and pipe characters have disappeared in the page history for me (image). &safemode=1 did not solve the problem, suggesting it's not a problem with my scripts. —k6ka 🍁 (Talk · Contributions) 01:19, 8 November 2019 (UTC)

@K6ka: could you trying clearing you local cache and trying again? (I can't duplicate this one). — xaosflux ^Talk 01:27, 8 November 2019 (UTC)

@Xaosflux: The issue seems to have fixed itself (or someone fixed it behind the scenes) and I'm not getting this problem anymore. —k6ka 🍁 (Talk · Contributions) 03:48, 8 November 2019 (UTC)

Temporary CSS fix

At the moment, I get the stray semicolon before the date in Special:Contribs results. I resolved it with the following in my common.css:

.mw-changeslist-date::before {
	content:'';
}

The debugger said the content was ';\u00a0' (semicolon followed by a non-breaking space), which was then prefixed to the timestamp, after the bullet. Should I leave the \u00a0 there? It looks fine to me without it. —[AlanM1(talk)]— 03:21, 8 November 2019 (UTC)

@AlanM1: shouldn't hurt anything, keep an eye on the master phab ticket, they may fix all of this for you sooner than later. — xaosflux ^Talk 03:24, 8 November 2019 (UTC)

Semicolons on Contributions pages

Tracked in Phabricator
Task T233649

I have noticed that the entries on Contributions pages (Special:Contributions) begin with a semicolon for each entry. Please fix this. —Etewilak (talk) 08:18, 8 November 2019 (UTC)

Probably T233649? --rchard2scout (talk) 11:08, 8 November 2019 (UTC)

Make Reply-to a Gadget

Proposal: Make the popular Reply-to script a gadget, with special credit and attribution to User:Enterprisey for his coding of the script.

Rationale: The current Reply-to script is headily used, but suffers from intermittent and frequent errors depending on the page you're replying to another editor.

Hat tip: Evad37 to based on this similar proposal, and to Steel1943. --Doug Mehus (talk) 01:51, 8 November 2019 (UTC)

• @Dmehus: are you referring to reply-link? If so, not only has the author said it is stilll being tested and debugged. Bugs are still present. but you also mentioned that it has problems. We shouldn't be forking this to a public gadget if it is known to be problematic. — xaosflux ^Talk 01:53, 8 November 2019 (UTC)

Xaosflux Yes, I meant Reply-link. Thanks. But, my understanding is that being a gadget, it will be hard-coded into the MediaWiki software, no? Also, I assume Wikimedia Foundation has full-time employed software engineers who could work out any kinks? --Doug Mehus (talk) 02:16, 8 November 2019 (UTC)

@Dmehus: nope, a "gadget" is just a community-managed script. You are probably thinking of an extension (an add-on piece of server software). You can track the (slow) progress on that at phab:T207567. — xaosflux ^Talk 02:28, 8 November 2019 (UTC)

@Xaosflux: Ah, that makes sense, and thanks for pointing out that link for me! Glad to see it's "in the works"—albeit slowly. In that case, I wonder if XFDCloser/DiscussionCloser could be made into Extensions? Also, re: "gadgets," what's the point in having a community-managed script versus the way it is now (one or more editors/administrators maintain the script)? Doug Mehus (talk) 02:31, 8 November 2019 (UTC)

@Dmehus: the primary benefit is for users, they can enable or disable gadgets very easily in Special:Preferences#mw-prefsection-gadgets instead of mucking about in .js/.css files. They are generally considered "safe" as only interface-administrators may make changes to them. (Anyone can request changes of course). — xaosflux ^Talk 03:09, 8 November 2019 (UTC)

Xaosflux, Ah, yes, I knew that, too! Like Twinkle...that makes sense. Thanks. Doug Mehus (talk) 03:18, 8 November 2019 (UTC)

Read only maintenance window planned for ENWP at 14th Nov 05:00 AM UTC (one week away)

A reminder this is one week away, see https://lists.wikimedia.org/pipermail/wikitech-l/2019-October/092653.html. --JCrespo (WMF) (talk) 09:22, 8 November 2019 (UTC)

Thanks log not included in "All public logs"?

Consider these two log searches:

• All public
• Thanks

Why is the "thanks" entry not shown in the "all" search? Which part of "all" did I not understand? -- RoySmith (talk) 13:56, 8 November 2019 (UTC)