Wikipedia:Village pump (technical): Difference between revisions

Browse history interactively

← Previous edit Next edit →

Content deleted Content added

VisualWikitext

Inline

Revision as of 17:01, 19 December 2014

Table of contents
First discussion
End of page
New post

Shortcuts

WP:VPT
WP:VP/T
WP:TECHPUMP
WP:PUMPTECH

The technical section of the village pump is used to discuss technical issues about Wikipedia. Bugs 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. Questions about MediaWiki in general should be posted at the MediaWiki support desk.

view · edit

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

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

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

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

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

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

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

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

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

« Archives, 110, 111, 112, 113, 114, 115, 116, 117, 118, 119, 120, 121, 122, 123, 124, 125, 126, 127, 128, 129, 130, 131, 132, 133, 134, 135, 136, 137, 138, 139, 140, 141, 142, 143, 144, 145, 146, 147, 148, 149, 150, 151, 152, 153, 154, 155, 156, 157, 158, 159, 160, 161, 162, 163, 164, 165, 166, 167, 168, 169, 170, 171, 172, 173, 174, 175, 176, 177, 178, 179, 180, 181, 182, 183, 184, 185, 186, 187, 188, 189, 190, 191, 192, 193, 194, 195, 196, 197, 198, 199, 200, 201, 202, 203, 204, 205, 206, 207, 208, 209, 210, 211, 212, 213

"Under construction" tag?

Moved from Wikipedia talk:Village pump (technical) § "Under construction" tag?

– Redrose64 (talk) 23:16, 18 December 2014 (UTC)[reply]

Hi, I don't know if this is the right place, but could someone tell me if there is a "under construction" tag for articles or something similar? I'm not talking about the "expand article" or "expand section" tags. I wanted something for articles that are currently being worked on. Thank you very much, --Lecen (talk) 10:42, 4 July 2012 (UTC)[reply]

{{Under construction}}? PrimeHunter (talk) 10:54, 4 July 2012 (UTC)[reply]

I looked everywhere in the list of templates but I couldn't find it. Thank you very much, PrimeHunter. You helped me a lot now. Regards, --Lecen (talk) 11:07, 4 July 2012 (UTC)[reply]

@Lecen: - There is also {{In use}}, which asks that other editors refrain from editing the page. GoingBatty (talk) 02:20, 20 August 2013 (UTC)[reply]

Twinkle problem for at least one admin

Moved from Wikipedia talk:Village pump (technical) § Twinkle problem for at least one admin

– Redrose64 (talk) 23:16, 18 December 2014 (UTC)[reply]

I've just done it again - deleting an article with Twinkle instead of tagging it. I don't use Twinkle for deleting, only for tagging - and keep forgetting to tick the 'tag only' box. Is there any way I could (without great complication) set a default of 'ticked' for this box? I find restoring something to tag it a bit embarrassing... Peridon (talk) 20:45, 18 December 2013 (UTC)[reply]

Enable "Default to speedy tagging instead of outright deletion" at Wikipedia:Twinkle/Preferences. Jackmcbarn (talk) 21:00, 18 December 2013 (UTC)[reply]

Thanks - I didn't know you could do all that stuff.... Peridon (talk) 11:01, 19 December 2013 (UTC)[reply]

Excellent - I'd just done the same thing, have now checked that box. – Philosopher ^{Let us reason together.} 22:11, 18 December 2014 (UTC)[reply]

Duplication Detector kaput

The labs tool Duplication Detector no longer works. I assume that's because the creator seems to be banned from Wikimedia. I know about Earwig's tool. But Duplication Detector was always more workable for the DYK purposes. Is anybody going to come up with a replacement? — Maile (talk) 00:25, 10 December 2014 (UTC)[reply]

What features of Template:Dupdet made DYK easier than Copyvios report? Perhaps The Earwig can help you find the features you are missing or improve it to add whatever features you feel are missing. — {{U|Technical 13}} ^{(e • t • c)} 01:29, 10 December 2014 (UTC)[reply]

I don't know the details. That question is better asked over at DYK. But I know the Earwig tool missed too much, and Duplication Detector was better for the DYK purposes.— Maile (talk) 01:33, 10 December 2014 (UTC)[reply]

If you haven't used it since July/August of this year, then I invite you to try it out again, because I've made some substantial improvements since then. If you have, then my apologies for it not being as good. I don't have much time right now for major improvements, but if you are able to suggest some concrete things, I can work on them in the future. — Earwig ^talk 02:27, 10 December 2014 (UTC)[reply]

With regards DupDet, the problem with the tool is unrelated to the ban. The creator's other tools are still functioning. Regardless of the respective merits, we need to find somebody else to manage it. Hopefully, we'll be able to get it back online in general in a few hours, but it needs a maintainer. Um, User:The Earwig? </cheesy grin> --Moonriddengirl ^(talk) 11:51, 10 December 2014 (UTC)[reply]

Would really rather not take on a new job like that, but I guess I can do it if no one else is able... — Earwig ^talk 21:51, 10 December 2014 (UTC)[reply]

@The Earwig: perhaps I could help, with your oversight? I've been taking an interest in copyright on WP, and have some geek cred. Cheers, Basie (talk) 19:31, 11 December 2014 (UTC)[reply]

I'm still unclear about what DupDet offered that Copyvios doesn't. Is it just a cosmetic difference in tastes or is it the actual search algorithms that are different? Why can't they just be merged? — {{U|Technical 13}} ^{(e • t • c)} 22:29, 10 December 2014 (UTC)[reply]
- I don't know if they can be merged. They each have strengths, but they are different. I really like the way EarwigBot highlights the text side by side, an elegant design that makes it much easier to find content. DupDet (which at the moment has been revived by User:Jalexander-WMF) offers more options. It allows you to compare Wikipedia articles against documents (including PDFs) as well as active URLs, allows you to eliminate quotations from your search, and allows you to specify word count/character size. If they can be merged into one tool that can do it all, that would be awesome. :) --Moonriddengirl ^(talk) 12:12, 11 December 2014 (UTC)[reply]
- The Earwig tool is superior, and I use it almost exclusively. But it can't at present compare a wikipedia page with a PDF source, while the Duplication Detector can. -- Diannaa (talk) 16:23, 11 December 2014 (UTC)[reply]
  - @Diannaa: Hmm... I added PDF support a few months back, could you check that again (or give me an example where it doesn't work)? The software it uses to read PDFs isn't perfect, so I could see it having problems with certain sources. If we have examples of PDFs that it can't understand but DupeDet can, I can try to fix that. (On rereading your comment, I'm unclear if you mean PDFs on the web or PDFs that you upload? If it's the latter, that's a feature I hadn't considered, but I can look into it if people want it.) — Earwig ^talk 20:35, 11 December 2014 (UTC)[reply]
    - @The Earwig: Here's an example one that didn't work in yours, and did work in Duplication Detector: http://www.dot.gov.in/sites/default/files/AR_English_2008-09_0.pdf, used as a source in Central Monitoring System. -- Diannaa (talk) 22:06, 11 December 2014 (UTC)[reply]
      - Ah, I see. There's a 1MB-per-URL size restriction, but that PDF's 2.2 MBs. I'll work on that. — Earwig ^talk 22:20, 11 December 2014 (UTC)[reply]
        Okay, I upped the limit for PDFs to 15 MBs. I imagine some are larger than that, but I'm uncomfortable raising it much higher. — Earwig ^talk 23:23, 12 December 2014 (UTC)[reply]

Note: I updated Wikipedia:Wikimedia_Labs/Toolserver_replacements to reflect the current situation. Please update again (this tool or others), when necessary. GermanJoe (talk) 09:32, 12 December 2014 (UTC)[reply]

Slow access

I am experiencing unusually slow access to Wikipedia these days and was wondering whether this is a site-wide issue and whether I can speed it up by tweaking options. I use Firefox with MonoBook skin and very few gadgets, if any; I've switched off Java support in web browser, which brought little improvement. Thanks. Materialscientist (talk) 04:39, 13 December 2014 (UTC)[reply]

I've also been experiencing this slow access today. The page seems to load very slow, and the top(with the messages and Watchlist) seems to lag for minutes. I use Google Chrome Version 39.0.2171.95 m. I can ping en.wikipedia.org with normal results. Irritating. Dave Dial (talk) 01:11, 14 December 2014 (UTC)[reply]

@Materialscientist and DD2K: Problems like this often depend on location. Whereabouts in the world are you? (I'm in Hokkaido, Japan, and the site is working normally for me.) — Mr. Stradivarius ^{♪ talk ♪} 01:35, 14 December 2014 (UTC)[reply]

I'm in South Detroit(born and raised, lol). It's loading faster now, but it still lags a bit(the loading wheel spins for a minute or so). Something has been up, since I can load ESPN, Google and Facebook normally. Dave Dial (talk) 01:41, 14 December 2014 (UTC)[reply]

It recovered with me automagically, without me doing much, so I assume this was some issue at wikiservers. Materialscientist (talk) 22:06, 15 December 2014 (UTC)[reply]

When does TemplateData get processed?

As it stands now, to properly document a template and its parameters, documentation in two, incompatible forms is required. First there is the documentation that you should be able to read at the template's page in template space, and second, there is the JSON formatted documentation used by Visual Editor. Maintaining two versions of documentation is a pain and ultimately counter productive.

In a discussion at Help talk:Citation Style 1 an editor raised the idea that some generic standard for template parameter documentation might be implemented in such a format that it could be rendered in a human usable form and could also be scanned by an automated tool to create documentation usable by VE and other tools.

It occurred to me that a couple of templates and some module code might make that idea possible. The result of my experiment is {{template parameter doc}}, {{template parameter doc item}}, and Module:template parameter doc. It did not work. Apparently, <TemplateData>...</TemplateData> processing occurs before templates and modules are processed. Is this true? If so, could that be changed?

—Trappist the monk (talk) 15:14, 13 December 2014 (UTC)[reply]

I think your module needs to wrap the template data content with an invocation to frame:extenstionTag rather than using <TemplateData> directly. I'm not entirely sure that will work either, but it seems more likely to. Dragons flight (talk) 00:11, 14 December 2014 (UTC)[reply]

Yep, Scribunto is run after tags have been converted to strip markers, so if you just output the tag text it won't do anything. You need to preprocess the tags somehow, and frame:extensionTag is the best way. (frame:preprocess would also work, but it's slower.) — Mr. Stradivarius ^{♪ talk ♪} 00:59, 14 December 2014 (UTC)[reply]

Ding! Ding! Ding! Both work. For now, I'm using frame:preprocess because it take the entires output in one function call and is producing what I want. Thank you both.

—Trappist the monk (talk) 01:42, 14 December 2014 (UTC)[reply]

@Trappist the monk: You should also be aware that at some point in the future, TemplateData will move to its own namespace with a JSON content type. This would mean that #invoke statements would just be normal text, rather than being processed with Scribunto, so you shouldn't plan on this code working long-term. — Mr. Stradivarius ^{♪ talk ♪} 05:12, 14 December 2014 (UTC)[reply]

I don't think it actually works now so maybe it's a non-issue. I'm guessing that Visual editor and the template data editor both read the raw source file to extract template data, not the rendered page. Right? So unless the module can subst a portion of its output (the <templatedata>...</templatedata>) I think this idea may have come to its end. And, even if it could subst, that subst'd stuff would have to be deleted before any edits to the documentation could be saved. That sort of requirement is no better than the currently unacceptable requirement to maintain two separate disparate sets of documentation.

—Trappist the monk (talk) 12:40, 14 December 2014 (UTC)[reply]

There was a similar discussion at Wikipedia talk:TemplateData#Extending use; removing redundancy. There are problems with trying to share documentation between template data and normal wikitext. The biggest is the markup supported by TD, basically none, no bold, italics or links. Krinkle explained why it needs to be this way.--Salix alba (talk): 08:11, 14 December 2014 (UTC)[reply]

Yep, that's why I concocted this scheme. Editors should only have to maintain documentation in one place and in one format. This solution, though an apparent failure, was an attempt to get us at least a step closer to that target.

—Trappist the monk (talk) 12:40, 14 December 2014 (UTC)[reply]

Not come across this before...

... and I hope not again.

There was an error showing yesterday in MLDonkey. If you look at it you will see lines in infobox like "Latest stable release". So where is that in the code? Has to be a template, right? But which? {{infobox software}} is horrendous. I eventually found subpage /MLDonkey hanging off it, which someone had carefully scribbled in.

But what we have is page-specific code attached to a general-purpose template (transcluded in 1200+ pages). Undocumented. Doesn't sound right to me. Presumably there are a lot of subpages, one per invocation I imagine.

A propos, it would save a lot of time if the list of templates used in a page also included date of last change, although it probably wouldn't have helped in this case. Easier than having to trog down each likely item in turn looking for one changed in last few minutes. Unbuttered parsnip (talk) mytime= Sun 07:37, wikitime= 23:37, 13 December 2014 (UTC)[reply]

That is caused by the |frequently updated= parameter that was suppose to be fully deprecated. See Template_talk:Infobox_software/Archive_5#Edit_request_on_2_October_2013:_.22frequently_updated.22 and the two sections following for a lot of discussion about these things... Pinging some of the other editors involved... Codename Lisa — MSGJ — Patrick87 — Tothwolf — Topbanana:. — {{U|Technical 13}} ^{(e • t • c)} 23:49, 13 December 2014 (UTC)[reply]
- Here be dragons. Be extremely careful with any major changes to the way {{Infobox software}}/{{LSR}}/{{LPR}} function and display on pages. Speaking from experience, as widely used as they are, major changes can have all sorts of unforeseen consequences (which then tend to lead to lengthy and often heated discussions).
  I don't see |frequently updated= present in the current version of {{Infobox software}}? I tried to document what I could remember of |frequently updated= in that discussion but I think the parameter had already been removed by that point anyway. None of those discussions were ever about depreciating the use of {{LSR}}/{{LPR}} subtemplates themselves though. Keep in mind that those subtemplates are not only transcluded by {{Infobox software}}, but are also widely transcluded in wikitables in software comparison articles. It might also be worth noting this in Template:Infobox software/doc#Moving release data outside the article.
  One possible {{Infobox software}} improvement that comes to mind immediately, is it should be possible to display a note on a preview version of a page which transcludes {{Infobox software}} which includes links to an article's Template:Latest stable software release/ArticleName and Template:Latest preview software release/ArticleName subtemplates. In cases where they already exist, an edit link could be provided, and in cases where they don't, a note and preload links so someone could create new subtemplates could be shown.
  Years ago when I was more active here, I discovered another language Wikipedia which had a much nicer and better documented {{Infobox software}}/{{LSR}}/{{LPR}} system, but I can't remember now which it was. --Tothwolf (talk) 15:07, 14 December 2014 (UTC)[reply]
  - Here be no dragons; just people who don't read documentations and complain that such documentations don't exist although they do.
    Oh, and one minor bit of correction: |frequently updated= was never supposed to deprecated; it was removed straight away. People frequently set it to "yes" indiscriminately, or just to mean "the developer issues updates quite often". They never used it to mean "facilitate recurrently updating this number on Wikipedia". Best regards, Codename Lisa (talk) 01:39, 15 December 2014 (UTC)[reply]

To see more detail in the "list of templates used on this page", install User:Anomie/previewtemplatelastmod. -- John of Reading (talk) 08:00, 14 December 2014 (UTC)[reply]
- Won't help him. Vandals rarely leave an edit summary and even when they do, it doesn't say "hey I just vandalized this page"! Best regards, Codename Lisa (talk) 01:39, 15 December 2014 (UTC)[reply]
  - @Codename Lisa: Among other things, the script sorts the list so that the most recently-edited template is at the top. I've found this script very valuable sometimes. -- John of Reading (talk) 07:45, 15 December 2014 (UTC)[reply]
    - I don't deny its usefulness as a tool in general. And let's not forget that you are a TemplateEditor no less. But I suspect my statement remains true. Best regards, Codename Lisa (talk) 11:31, 15 December 2014 (UTC)[reply]
40 years in dp and I'd never come across such gonads as here. If it had been {{lpr}} (etc) looking for a subpage of the caller it would be fine, but to have a bespoke subpage hanging off a general purpose callee object is astonishing. Unbuttered parsnip (talk) mytime= Tue 17:12, wikitime= 09:12, 16 December 2014 (UTC)[reply]

The above appears to be referring to Special:PrefixIndex/Template:Latest stable software release where over a thousand subpages are shown, each having a single item for an infobox—example, Template:Latest stable software release/MLDonkey. Special:WhatLinksHere/Template:Latest stable software release/MLDonkey shows the subpage is used in several places. It's unusual, but probably ok. All the information could be put in a module, but that would give a single page with over a thousand items (making it hard to edit), versus the current situation of simple edits on over a thousand pages. Another drawback of a module would be that an update to any item would put over a thousand articles in the job queue. Johnuniq (talk) 09:55, 16 December 2014 (UTC)[reply]

Sure. But I wouldn't bother about his comment on Template:Latest stable software release if I were you. He seems more furious about his own lack of experience here than making a genuine comment based on actual facts. For all I know, it is a maintainable system that works; a single slash is no reason to change it. But Tothwolf, on the other hand, might actually be on to something. I hope he remembers which wiki it was. Best regards, Codename Lisa (talk) 17:56, 16 December 2014 (UTC)[reply]

I've looked through the other language wikis I have an account on via SUL and I simply can't find it now. I did notice the Italian and Russian wikipedias have both begun to use Wikidata for their implementations of {{Infobox software}}. I don't really see that it would be worth spending much more time trying to find that better {{LSR}}/{{LPR}} implementation if much of the data that is currently stored under Template:Latest stable software release/ArticleName will eventually be moving to Wikidata anyway. --Tothwolf (talk) 22:05, 16 December 2014 (UTC)[reply]

The multi-use of the MLDonkey LSR template is actually a very good example of what I was trying to explain above. The way Template:Latest stable software release/* and Template:Latest preview software release/* are widely used for software comparison articles probably should be better documented in Infobox software's documentation. For that fact, maybe this common convention should also be documented somehow in MOS:SAL, or better yet maybe even a separate MOS page for software comparisons? --Tothwolf (talk) 22:33, 16 December 2014 (UTC)[reply]

Happened again just now, with WhatsApp. This time I saw it coming. I think the mechanism is a complete pile of pooh. Unbuttered parsnip (talk) mytime= Tue 22:07, wikitime= 14:07, 16 December 2014 (UTC)[reply]

You may feel the system in place is a "pile of pooh", but the simple fact remains that it has worked extremely well for 1000s, if not 10s of 1000s of editors, both anonymous drive-by editors and veterans alike, for making quick updates to software version numbers and release dates. Not only does it make it much easier for less experienced editors to update a version number and release date, it also cuts down on a considerable amount of clutter in Recent changes.
For a system that has its roots back in 2005, I think it has actually aged quite well. You might consider it strange that data has been stored as a "subpage" in the Template: namespace, but given the limitations of Mediawiki's templates and markup language, I think as a whole it has actually worked out pretty well.
When I was more active, I used to routinely monitor Special:RecentChangesLinked/Category:Latest stable software release templates and Special:RecentChangesLinked/Category:Latest preview software release templates. Surprisingly enough, I saw little in the way of intentional vandalism, but I did see the occasional test edit, much like the one you reverted. Even more common though, were careless reverts made with Huggle to valid edits from anonymous users/ip addresses and red-linked usernames.
Btw, you don't have to make a "dummy edit" with an actual change to the page to "refresh" the transcluded data. The job queue will handle it automatically, but if you really want to force an immediate update, you can either make a null edit or append ?action=purge to the end of the url (example). --Tothwolf (talk) 23:47, 16 December 2014 (UTC)[reply]

CiteGen

Hello, folks! I've just made User:Zhaofeng Li/CiteGen which is a replacement of Cite4Wiki. It's an extension for both Firefox and Chrome which generates a {{cite web}} citation from the current page. It's different from Cite4Wiki because it uses Reflinks as the backend, whereas Cite4Wiki parses metadata in the browser. As a result, it supports more metadata fields and its functionality can be extended without the user updating the extension. The source code is available here (Patches are welcome!)

To install this on Firefox, simply click this link. For Chrome/Chromium, right click this link and select "Save link as..." to download it. Next, open Settings in Chrome and select the Extensions menu, then drag the .crx file from your file browser to Chrome.

Any thoughts on this? Thank you. Zhaofeng Li [talk... contribs...] 01:21, 14 December 2014 (UTC)[reply]

I have installed this add-on for Mozilla/5.0 (Windows NT 6.1; WOW64; rv:34.0) Gecko/20100101 Firefox/34.0 and will be testing it shortly. Thanks for developing it! — {{U|Technical 13}} ^{(e • t • c)} 00:21, 17 December 2014 (UTC)[reply]
- Zhaofeng Li, I've used it a bit today to try it out, and I have a couple things I might request. First, since I tab flip a lot (and sometimes forget to wait for it to load to copy the material), it would be great if it could cache the results locally so when I flip back it I don't have to wait for it to start over and try again. Also, it would be great if the pages it fails to work on could either offer some kind of report to help figure out why it didn't work on those pages (I've hit quite a few free pages not behind a paywall it fails on). It would be nice to have it have a checkbox that will wrap the citation in <ref>...</ref> tags (named refs are preferred of course, so they can easily be reused :D). It would also be appropriate to introduce a user option for whether it should be {{cite web or {{Cite web and it should allow the user to select the date format (MDY vs. DMY vs. YMD) if not set it should pull the browser default for the date format. Another thing that would be a great addition would be a button that an editor using the tool to work on a GA or FA could press to prevent linkrot by caching a version of the page on the way back machine or some other similar site and add the archive info to the citation template. I'm sure some of those ideas are much easier than others, and I'll throw more ideas at you as I have them. — {{U|Technical 13}} ^{(e • t • c)} 01:03, 18 December 2014 (UTC)[reply]

@Technical 13: Many thanks for your testing. Caching will certainly help, and more options will be available soon. As for pages it fails to work on, it ~~will~~ currently outputs an error message at the bottom ("HTTP Error: 403", etc). Zhaofeng Li [talk... contribs...] 07:58, 18 December 2014 (UTC)[reply]

Logged out

How come I was logged out a minute ago in the middle of watchlist checking? Glitch at WP end, or mine? Jim.henderson (talk) 01:25, 14 December 2014 (UTC)[reply]

Maybe just your login expiring? KonveyorBelt 02:28, 14 December 2014 (UTC)[reply]

Unless you see it happenning a lot, just log back in and ignore it. עוד מישהו Od Mishehu 06:58, 14 December 2014 (UTC)[reply]

Central Login failure in Firefox

Until Wednesday 10 Dec, if I logged in on e.g. English Wikipedia using my normal browser (Firefox), the Central Login system - which I think is part of WP:SUL - meant that I was automatically logged in on commons, meta, Wikidata, etc. although sometimes on first visit to a wiki it would show "Create account Log in" top right, and alongside that a bubble "Central login You are centrally logged in as Redrose64. Reload the page to apply your user settings."; pressing Ctrl+F5 would reload the page and I would be shown as logged in.

After I logged out on Wednesday, my next login was yesterday morning (13 Dec) - and it only logged me in on en.wp: if I go to commons, or any other Wikimedia site, I'm shown as logged out ("Create account Log in"). This would be OK, except that the "You are centrally logged in" bubble does not appear, not even if I try a hard refresh of the page. This is particularly annoying on Commons, because without being logged in there, it opens images in Media Viewer. I've logged out, cleared cookies, tried again, started off on different sites (there was a period some months ago when Central Login only worked properly if I logged in on Meta first, even though my "home" wiki is en.wp), but whatever I try, no luck. The only way to be logged in on all sites is to log in to each site separately, which defeats the purpose of Central Login and indeed of SUL.

I suspect a browser problem. If I switch browsers from Firefox to Opera, it works as it should (I get the "You are centrally logged in" bubble when visiting another wiki for the first time after logging in). Firefox was recently upgraded from 33.something to 34.0.5, so I suspect that something has changed in Firefox - is there some configuration relating to cookies which might affect it? --Redrose64 (talk) 10:20, 14 December 2014 (UTC)[reply]

After trying several more times after posting the above, Commons and Meta are now recognising my en.wp login after some six hours. Since I didn't upgrade Firefox, restart it, or even alter any Firefox settings during that time, it can't be a browser problem. Most odd. --Redrose64 (talk) 16:29, 14 December 2014 (UTC)[reply]

Regression in backlink symbol

Occasionally I come across a page where in the references list, the backlinks are malformed. When this happens, there are two effects: for all refs, an arrow is shown instead of a caret; and for those refs used two or more times, numbers (2.0 2.1 etc.) appear instead of letters (a b etc.). If I WP:PURGE the page it fixes itself. It's not a caching issue at my end, and is replicable in various circumstances - on different browsers, whether logged in or out, using Vector or MonoBook, and when viewing the article proper (e.g. Epping tube station#References) or a diff (e.g. last revision); it's as if there was a software regression. This has happened more frequently in the last few weeks. --Redrose64 (talk) 22:44, 14 December 2014 (UTC)[reply]

I see it to (but then I purged). My guess is the Cite extention somehow fails to replace the default messages with the locally defined one (MediaWiki:Cite references link one and MediaWiki:Cite references link many). -- [[User:Edokter]] {{talk}} 23:06, 14 December 2014 (UTC)[reply]

Edokter is correct. What you are seeing is the default styling. We've had a very similar issue where cite error messages would show in a non-English language because of the last editor's language preferences, until the page was purged. I don't recall seeing this on the Cite bug list. -- Gadget850^talk 23:20, 14 December 2014 (UTC)[reply]

Ping notifications working?

I and another editor have tried repeatedly to ping some users with Template:ping (Template:Reply to), at Talk:Houston_Riot_(1917)#Requested_move_14_December_2014, but we don't see any notifications. I tried a test at my talk page, without the template, too: User_talk:Dicklyon#Ping_test; still neither of us saw a notification. Each time, I used a four-tilde signature. Is there some magic I'm missing, or are notifications broken? Dicklyon (talk) 01:58, 15 December 2014 (UTC)[reply]

You have to ping them and sign in the same edit or it doesn't work. See User:Floquenbeam/Pinging. --Floquenbeam (talk) 02:03, 15 December 2014 (UTC)[reply]

I did that every time. It's hard to see from the history, since the four-tilde signatures are expanded, but each one is new, not a copy of an old signature. Dicklyon (talk) 02:15, 15 December 2014 (UTC)[reply]

There are many factors that can cause this. See Template_talk:Reply_to#Ping or do a search on Phabricator for echo tickets. Signatures must include a non-colon prefixed link to the pinger's userpage and can not overwrite a previous ping (best to delete the signature (leave the old timestamp) and add a new four tilde signature at the end). Happy pinging! — {{U|Technical 13}} ^{(e • t • c)} 02:23, 15 December 2014 (UTC)[reply]

I can't see how that would be the issue. My signature is pretty vanilla. Dicklyon (talk) 02:29, 15 December 2014 (UTC)[reply]

- Yeah, I've found it best (if slightly annoying) to undo the edit, then re-add the corrected ping and the corrected signature together, rather than overtype it. Also, Dicklyon what did you mean by "we didn't see any notifications"? You wouldn't have seen any notifications when you ping other people, it isn't a talk page thing it's a notification button thing. You can only ever see it when someone pings you. --Floquenbeam (talk) 02:27, 15 December 2014 (UTC)[reply]

I mean notifying myself didn't work, and RGloucester and my attempts to notify each other didn't work. Dicklyon (talk) 02:31, 15 December 2014 (UTC)[reply]

You can't notify yourself. And, as I said, your ping test on your user page didn't include a new ping and a new signature at the same time. I don't see where RGloucester tried to ping you. --Floquenbeam (talk) 02:40, 15 December 2014 (UTC)[reply]

From where I am standing, it seems working because I did receive your ping. I am referring to the edit in which you called me "Asshole" and the associated talk page edit. It is possible that the editor is ignoring you. Also editors can disable notifications, in whole or for certain person. Codename Lisa (talk) 02:32, 15 December 2014 (UTC)[reply]

I called you "asshole" in an edit summary. What's not working is a ping caused by mentioning a user name. Actually, my edit summary of "asshole" was more a reaction to your edit summary than to you, per WP:NPA, since I know nothing about you and expect you're actually a nice guy in real life. Dicklyon (talk) 02:33, 15 December 2014 (UTC)[reply]

Okay, based on that, I looked at your actual edits. Talk:Houston_Riot_(1917)#Requested_move_14_December_2014 failed to ping anyone because you exceeded the hard coded limit of 20 pings per post. You can see in this edit specifically, you did not add a new signature to the end, you attempted to overwrite an existing signature, which will not work. this test failed because there was no signature at all, and attempt to fix it failed because you only added a signature. You have to link to the user page of the person you want to ping and have an entirely new signature with a clean link to your userpage for it to work. Hope this clears it up. — {{U|Technical 13}} ^{(e • t • c)} 02:43, 15 December 2014 (UTC)[reply]

Thanks, those make a bit of sense. What about this one? Would I need to do intervening edits to remove the ping and signature, so that adding the new one is "clean"? I can try that. Why the 20 limit? Dicklyon (talk) 03:26, 15 December 2014 (UTC)[reply]

An intervening edit like that should do it. A whole new post after the original should also work. If there are more than 20 pings then there is probably a significant risk that it's either disruptive or an accident like copying a lot of signed posts with user page links and adding your own signed post in the same edit. PrimeHunter (talk) 03:44, 15 December 2014 (UTC)[reply]

See mw:Help:Echo#Technical details and if you still think a notification is missing then post a diff to the edit. Don't just say which page it was on. PrimeHunter (talk) 03:27, 15 December 2014 (UTC)[reply]

I have added the total limit of 20 to the documentation for {{reply to}}.[1] PrimeHunter (talk) 03:37, 15 December 2014 (UTC)[reply]

Using a tilde signature in a template

When I join or initiate a discussion on a Talk page, I usually end my first entry in the thread with something like

If you would like to discuss this with me, please {{ping}} me. --~~~~

where of course the "~~~~" turns into my timestamped signature. This way, I can be available for discussion without cluttering my watchlist.

Since I do this fairly often, I'm trying to make a template in my userspace that would expand to that ↑↑. But I haven't been able to make the final result include the expanded date/time. Either I still get plain old "~~~~" in the result, or I get the timestamp in the template itself, which will never be correct when the template is invoked. Advice, please? And, of course, please {{ping}} me. --Thnidu (talk) 02:54, 15 December 2014 (UTC)[reply]

@Thnidu: You can use ~~<includeonly></includeonly>~~. Cenarium (talk) 03:06, 15 December 2014 (UTC)[reply]

Thnidu, try ~~<noinclude />~~. — {{U|Technical 13}} ^{(e • t • c)} 03:07, 15 December 2014 (UTC)[reply]

@Cenarium and Technical 13: Thanks, that works with subst:.

--Thnidu (talk) 04:04, 15 December 2014 (UTC)[reply]

Blacklisted links ... or not?

Dear editors: Can someone who understands the blacklisted links process please take a look at the history of this page ITM LAW School? Thanks. —Anne Delong (talk) 03:02, 15 December 2014 (UTC)[reply]

The history links at Special:Contributions/Cyberbot II shows the bot is edit warring with itself on lots of pages (maybe a case for Wikipedia:Lamest edit wars#(Ro)bot wars). It has been reported to the operator at User talk:cyberpower678#Cyberbot II is churning, churning. PrimeHunter (talk) 03:15, 15 December 2014 (UTC)[reply]

Expired security certificate

I'm using IE9 to view the site (not my choice -- mandated by the company) and as of this morning, I'm getting errors indicating that the site's security certificates are expired. Anyone else seeing this? I can navigate around it, but it's quite annoying. WikiDan61^ChatMe!_ReadMe!! 14:35, 15 December 2014 (UTC)[reply]

@WikiDan61: See Wikipedia:Village pump (technical)/Archive 131#SSL 3.0 discontinued. --Redrose64 (talk) 16:41, 15 December 2014 (UTC)[reply]

@WikiDan61: You won't be able to browse at all if it's related to SSLv3. It's more likely that your company intercepts your HTTPS traffic with a rouge certificate which is expired. In this case, contact your IT department. Zhaofeng Li [talk... contribs...] 04:46, 16 December 2014 (UTC)[reply]

@Zhaofeng Li: I'm pretty sure you mean a rogue certificate, not a rouge certificate! A lot of native English users make this error, too. --Thnidu (talk) 04:53, 19 December 2014 (UTC)[reply]

PHP API for #ifexist function?

The "getpage" function I'm using (see here), gets the content of a page, and returns FALSE on error. The problem is that it returns FALSE even when the error is simply that the page does not exist, making it impossible for the application to know when the error is really that the server is down, or the bits got lost in the ether, or who-knows-what-problem beyond the application's control. Indeed the page merely not existing is not a problem, as the application in that circumstance just wants to create a new page rather than append content to an existing page. What I really want is a return code that confirms that the page actually doesn't exist and needs to be created. This is what I believe is needed to address this reported problem. Help from PHP programmers knowledgeable about the MediaWiki API would be appreciated. – Wbm1058 (talk) 14:49, 15 December 2014 (UTC)[reply]

https://en.wikipedia.org/wiki/Special:ApiSandbox#action=query&prop=info&format=xml&titles=FooBarBaz is an example of how to see if a page exists through the api. If the "missing" attribute exists, the page does not. — {{U|Technical 13}} ^{(e • t • c)} 16:11, 15 December 2014 (UTC)[reply]

I'm having trouble making this work in actual php code. My attempt to check the 'missing' attribute with this code change didn't work for me. Jackmcbarn can you help? Wbm1058 (talk) 19:44, 15 December 2014 (UTC)[reply]

 if(isset($x['missing'])) return "<missing>";

would have to be some variation of

 if(isset($x['query']['pages']['-1']['missing']) || $x['query']['pages']['-1']['missing'] == "") return "<missing>";

to use that to determine if missing was there. — {{U|Technical 13}} ^{(e • t • c)} 19:58, 15 December 2014 (UTC)[reply]

That specific code returned "<missing>" for every item passed to it, including everything that wasn't. Can you explain what you're doing there? It seems kind of convoluted to me. Wbm1058 (talk) 20:35, 15 December 2014 (UTC)[reply]

I've no way to test my ideas, so play with it, you might need to replace the OR ( || ) with an AND ( && ) or you may only need the second half of the condition (== ""). I'm not sure that isset() will catch that it exists if its value is null or empty, so you need to be careful with that. :) — {{U|Technical 13}} ^{(e • t • c)} 20:43, 15 December 2014 (UTC)[reply]

If multiple parameters are supplied then isset() will return TRUE only if all of the parameters are set. So it is not sufficient to check only whether 'missing' is set? Wbm1058 (talk) 22:46, 15 December 2014 (UTC)[reply]

isset() determines if a variable is set and is not NULL. Returns TRUE if var exists and has value other than NULL, FALSE otherwise. So if indeed "missing": "" is that NULL or not, and if it is, what's the point? I don't get it. Shouldn't "missing": "TRUE" be returned by the function for "FooBarBaz"? Wbm1058 (talk) 23:05, 15 December 2014 (UTC)[reply]

(edit conflict) If I understand your question, think of it like a tree. ['missing'] is a leaf that is attached to a branch labeled ['-1'], which is in turn attached to the trunk of the tree, ['pages'], and finally to the root, $x['query']. If you are a worm, and you want to get to the leaf, you need to start at the root, climb up the trunk, and across the branch to get to the leaf. :) isset() is suppose to return if that value is defined, but I don't remember how it does with things that are defined as null or empty, so checking http://php.net/manual/en/function.isset.php I find that Determine if a variable is set and is not null. — {{U|Technical 13}} ^{(e • t • c)} 23:11, 15 December 2014 (UTC)[reply]

Whatever. You found the link that I got the information from which I shared in my last two edits. I appreciate that you're trying to help, but I think it's better at this point to wait for someone who knows this stuff to respond. @Cyberpower678: I hate to pull you away from your expensive studies, but if you can quickly give an answer, I'd appreciate it. Wbm1058 (talk) 23:25, 15 December 2014 (UTC)[reply]

@Wbm1058: See how Reflinks deals with this in an unclean way. Zhaofeng Li [talk... contribs...] 23:33, 15 December 2014 (UTC)[reply]

@Zhaofeng Li: Thanks, that seems helpful. I notice it uses format=json rather than the format=php that the "library function" I have uses. I'm not familiar with JSON (JavaScript Object Notation). Is there an advantage in using that? Wbm1058 (talk) 02:21, 16 December 2014 (UTC)[reply]

@Wbm1058: I think they are the same. However, unserialize() is always available,^[2] while json_decode() is available by default from 5.2.^[3] But anyway, considering most distros have at least PHP 5.3, compatibility isn't an issue. Zhaofeng Li [talk... contribs...] 03:55, 16 December 2014 (UTC)[reply]

Never use unserialize() on foreign data, it's a security risk. Legoktm (talk) 04:44, 16 December 2014 (UTC)[reply]

Wow, didn't know that before. That's quite true since unserialize() can be used to execute foreign code when initialising the classes. @Wbm1058: You really should look into JSON, then. Zhaofeng Li [talk... contribs...] 06:14, 16 December 2014 (UTC)[reply]

@Technical 13: ~~By the way, to correct your mistake,~~ isset() can only be used with variables. Passing anything else to it (for instance an expression ~~in your example~~) will fail (It would output "Can't use function return value in write context" which was unhelpful, but now it gives an error message with a correct example).^[4]^[5] As an an unrelated note, empty() was a similar function which only accepts variables, leading to widespread confusions and steeper learning curve. This was changed in PHP 5.5.^[6] Sadly, the learning curve is still there, since PHP 5.3/5.4 are still popular. [Insert rants about PHP design here] Zhaofeng Li [talk... contribs...] 12:52, 16 December 2014 (UTC)[reply]

Zhaofeng Li, look again, my example doesn't put the expression in the isset(), it puts the isset() in an expression. It says IF isset($x['query']['pages']['-1']['missing']) OR if $x['query']['pages']['-1']['missing'] == "" THEN return "<missing>";. — {{U|Technical 13}} ^{(e • t • c)} 12:57, 16 December 2014 (UTC)[reply]

Ah, I was having some eye problems. Sorry about that. :( Take it as a rant against PHP's design, then. Zhaofeng Li [talk... contribs...] 13:17, 16 December 2014 (UTC)[reply]

No worries, honest mistake. :) — {{U|Technical 13}} ^{(e • t • c)} 13:40, 16 December 2014 (UTC)[reply]

There is another item on my problems list: see User talk:RMCD bot § Unicode, diacritics problems. Will using json_decode() rather than unserialize() address this issue, or is there something else going on? Wbm1058 (talk) 17:35, 16 December 2014 (UTC)[reply]

@Legoktm: The "Bot classes for interacting with mediawiki" is the function library that the I bots I took over use. I believe that Chris G was the person coordinating it. As I recall I pulled the source off of his account on the former toolserver, but now when I look there I just see a "We've moved!" message. I also note that Chris G has been rather inactive for the past few months, since they turned their bots over to you. Looking at botclasses.php I see a single occurrence of the unserialize() function which you say is insecure for foreign data. However this is in the low-level query() function which "sends a query to the api". This query function seems to be used by virtually every other function in this library, and thus it appears that the entire library is insecure. It's a bit upsetting to learn that I've been running (potentially) insecure scripts on my personal computer for the last two years. Is there a new central location where this library is maintained, where I can look for a newer version, which is hopefully secure? Perhaps this is something that the highly paid programming staff of the Wikimedia Foundation can provide to their volunteer programmers and bot operators? Wbm1058 (talk) 14:36, 16 December 2014 (UTC)[reply]

@Legoktm: I see from Wikipedia:Bots/Requests for approval/Legobot 33 there is a link to the source code for the harej-bots on GitHub. Observe that your copy of botclasses.php there uses unserialize() in the query() function. So a fix would benefit both of us. Wbm1058 (talk) 20:48, 16 December 2014 (UTC)[reply]

I see: unserialize – Warning Do not pass untrusted user input to unserialize(). Unserialization can result in code being loaded and executed due to object instantiation and autoloading, and a malicious user may be able to exploit this. Use a safe, standard data interchange format such as JSON (via json_decode() and json_encode()) if you need to pass serialized data to the user.

I suppose that the encyclopedia that anyone can edit might have untrusted user input in it? I wonder if it's possible to track down everyone who is still using the botclasses library. How serious an issue is this? I might take a stab at the necessary code changes soon if nobody else does. Wbm1058 (talk) 16:31, 17 December 2014 (UTC)[reply]

@Wbm1058: Just replace format=php with json, and replace the unserialize line with return json_decode($ret,true);. By the way, the real risk is not from the editors, since the API will properly escape whatever they write. The problem is you don't know whether the API you are talking to is the real one or not (Using HTTPS will help a lot), and a fake one could return a maliciously-crafted string as the result. Zhaofeng Li [talk... contribs...] 04:51, 18 December 2014 (UTC)[reply]

@Zhaofeng Li: Thanks, I made the JSON changes and it worked fine. But when I changed "http:" to "https:" it responded with "Login error: ". Does the API not support secure http or is there something else I need to change to get it to work? Wbm1058 (talk) 16:59, 19 December 2014 (UTC)[reply]

Tech News: 2014-51

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 software changes

You may see a new tool tested on the mobile site of the English Wikipedia. It asks simple questions to make the article better. In the future, your answers will go to Wikidata. [7]
You can watch a video to learn how to use Phabricator to manage projects. [8]
You can test a new version of the Content Translation tool.

Software changes this week

The new version of MediaWiki (1.25wmf12) has been on test wikis and MediaWiki.org since December 10. It will be on non-Wikipedia wikis from December 16. It will be on all Wikipedias from December 17 (calendar). [9]
You can now find and replace text in VisualEditor. You can open the tool via the menu, Ctrl+F or Cmd+F. [10]
VisualEditor now doesn't change the space at the end of a text that you make bold. [11]
The "Apply changes" button in the reference window of VisualEditor is now disabled until you make a change. [12]

Future changes

The star image of the watchlist may change soon. [13] [14]
You won't be able to use Phabricator on December 18, 2014. It will be down between 00:00 and 08:00 (UTC). [15] [16]

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

16:44, 15 December 2014 (UTC)

Number of unwatched pages

Is there a way to find the number of unwatched pages? They are listed at Special:UnwatchedPages but I suspect that, even 500 at a time, it would take a long time to get to the end of the list by using "next 500". JohnCD (talk) 17:46, 15 December 2014 (UTC)[reply]

John, you can speed up the process by changing limit=500 to limit=5000 in the URL address bar. You might also be interested in my User:Technical 13/Scripts/Gadget-listStyles userscript which will convert the bullets into a numbered list on most pages (I don't have permission to view that special page, so I haven't tested it on that specific page yet, if it doesn't work, let me know). — {{U|Technical 13}} ^{(e • t • c)} 19:33, 15 December 2014 (UTC)[reply]

Thanks. The problem is that I suspect the number of unwatched pages is very large, in the 2 - 3 million range, though that's what I want to check; so even stepping by 5,000 would take some time. If I'm right, do you think your script would cope? JohnCD (talk) 22:47, 15 December 2014 (UTC)[reply]
- Considering what my script does, sure. You could even just load it through the console on the last page (the one with no more "next 5,000") and it will tell you how many are on that page. Then you just have to go backwards and count the pages to get the larger number. That will still take quite a while for 2+ million pages. Let me see if I can find a better way. :) — {{U|Technical 13}} ^{(e • t • c)} 23:19, 15 December 2014 (UTC)[reply]

I left a note at Wikipedia talk:Special:UnwatchedPages in case anyone is watching that page and has ideas. Hmm... I can't load AWB on this computer, but I wonder if there is any way for it to count the entries? (Such as by having it make a list of all pages on that special page?) – Philosopher ^{Let us reason together.} 23:33, 15 December 2014 (UTC)[reply]
- I'm not sure why only administrators can view that specific special page, but due to that I am unable to use that method to find out for you. — {{U|Technical 13}} ^{(e • t • c)} 00:01, 16 December 2014 (UTC)[reply]
- I believe AWB caps its list at 25,000 items, which is better than 5,000 – but Magioladitis should be able to give the definitive answer. Wbm1058 (talk) 00:20, 16 December 2014 (UTC)[reply]
  - I've tried looking at the pageview statistics for random pages, and have never found a page that is viewed less than three or four times per day. I guess enough people browse random pages that every page gets a few hits. So there are no unwatched ones. DOwenWilliams (talk) 00:54, 16 December 2014 (UTC)[reply]
    - There may be no unviewed pages, but being viewed does not put a page on anyone's watch-list. JohnCD (talk) 10:15, 16 December 2014 (UTC)[reply]
  - It's admin only because it's vandal bait. I believe there once was a case where an admin gave a list to a banned vandal who then proceeded to trash a bunch of pages that no one was watching. Oiyarbepsy (talk) 01:01, 16 December 2014 (UTC)[reply]
    - I think that is too narrow of a scope. I'd like to see it expanded to be available to more established users. It should be available to CUs, OSs, TEs, and probably Rollbackers as well. Is there anyone else that might be interested in developing a proposal to expand the usergroups with access to that page to include those? — {{U|Technical 13}} ^{(e • t • c)} 01:11, 16 December 2014 (UTC)[reply]
      - I would oppose that, because CUs and OSs are required-to-be admins anyway, and TEs/Rollbackers go through a lot less rigorous of a process to get those userrights. The issue that Oiyarbepsy mentions did not end well, and we do not want to see it repeated. Quiddity (talk) 18:47, 18 December 2014 (UTC)[reply]

A lot of this is old hat. There were Most-watched pages (configuration) and other reports on the Toolserver. Nobody has unfucked Labs yet (cf. phabricator:T59617). For individual pages, you can use the info action (example). This data is also available via the MediaWiki API. For unwatched pages specifically, you'd want access via Labs or equivalent, I imagine (individual lookups are rough). An SQL query for this is trivial. Admins have access to Special:UnwatchedPages (as noted) and aren't subject to a threshold on the info action. --MZMcBride (talk) 04:17, 16 December 2014 (UTC)[reply]

@Technical 13: Concerning visibility of Special:UnwatchedPages, please see comments at Wikipedia:Village pump (technical)/Archive 92#js to add watchlist pages; Wikipedia:Village pump (technical)/Archive 103#Unusual search advice; and the comment by Azylber at 00:06, 28 January 2013 at Wikipedia:Village pump (technical)/Archive 107#Have we lost the count of the number of watchers?. --Redrose64 (talk) 11:53, 16 December 2014 (UTC)[reply]

Formatting, signatures not working on a talk page

See the last section of [17]. Thanks. Dougweller (talk) 22:03, 15 December 2014 (UTC)[reply]

@Dougweller: That is due to an unclosed <nowiki> added with this edit. SiBr₄ (talk) 22:10, 15 December 2014 (UTC)[reply]

Thanks. Dougweller (talk) 22:15, 15 December 2014 (UTC)[reply]

I cleaned it up. Wbm1058 (talk) 22:42, 15 December 2014 (UTC)[reply]

Socks (cat)

If the article Socks (cat) is in the category "Category:Individual cats", then why does it not show up anywhere when you check the category page? Ten Pound Hammer • ^{(What did I screw up now?)} 23:40, 15 December 2014 (UTC)[reply]

Mediawiki hiccup. It's fixed. DS (talk) 23:57, 15 December 2014 (UTC)[reply]

Occasionally such glitches happen - see, for example, Wikipedia talk:Special:UncategorizedPages#Additive identity. There is always the olption of null-editing the page. עוד מישהו Od Mishehu 06:47, 16 December 2014 (UTC)[reply]

Can't log in

Tracked in Phabricator
Task T75462

On Friday, attempting to log in from work (not too far away from where I live). resulted in a 503 error message all day. I remained logged in at home until some time today. Now I can't get logged back in ... but only to Wikipedia; all other Foundation projects I have accounts on (Commons, Wikivoyage, Wikiquote etc.) work just fine. I have tried multiple browsers and other Internet access points in the area; no luck. So I am posting via my IP address in the hope someone will be able to make sense of the error message from my recent attempt to log in:

Request: POST http://en.wikipedia.org/w/index.php?title=Special:UserLogin&action=submitlogin&type=login& returnto=Wikipedia:Village+pump+(technical), from 10.64.32.106 via cp1067 cp1067 ([10.64.0.104]:3128), Varnish XID 2153850834
Forwarded for: 173.85.86.72, 10.64.32.106, 10.64.32.106
Error: 503, Service Unavailable at Tue, 16 Dec 2014 03:10:52 GMT

If this requires any personal follow up, email me from my userpage or (if that doesn't work) use my Commons talk page (I have email notification set up there). Daniel Case (talk) 03:23, 16 December 2014 (UTC) — Preceding unsigned comment added by 173.85.86.72 (talk) [reply]

I have no trouble, so I don't know... If, as yyou say, you can log in to other Wikimedia projects, can you then come into Wikipedia using your SUL? עוד מישהו Od Mishehu 07:02, 16 December 2014 (UTC)[reply]

This sounds like phab:T75462 and the last comment asks for coordinating with someone whose account is affected so help is welcome in that ticket. --AKlapper (WMF) (talk) 10:21, 16 December 2014 (UTC)[reply]

Reading through that ticket, it is exactly the problem I'm having. Points of perhaps useful information:

I do not have an SUL, just like some of the other people having the problem.
I am getting the usual error message if I mistype my password, just like the other people having the problem.
It is only occurring on enwiki, which, if I did have an SUL, would be my home wiki. Nowhere else.
To reiterate, it is not specific to Firefox, even though that's my primary browser. It has occurred without fail in IE, Chrome and (when I tried logging in from my iPad) Safari.
At this point I very much believe it is specific to my account, although no one else that I know of logs on at any of the other points I've tried from. I will try creating an alternate account (which I probably should have done a while ago anyway) and see if that works. Daniel Case (talk) 20:27, 16 December 2014 (UTC) — Preceding unsigned comment added by 173.85.86.72 (talk) [reply]

OK, I was just able to log on. If someone implemented one of the older bug patches mentioned most recently in the ticket, it worked. Daniel Case (talk) 23:42, 16 December 2014 (UTC)[reply]

Combining duplicate references

Moved from Wikipedia talk:Village pump (technical) § Combining duplicate references

– Redrose64 (talk) 23:18, 18 December 2014 (UTC)[reply]

Dear editors: At one time the "Expand citations" used to combine duplicate references. Also Reflinks used to do this. These don't seem to have this function now. Is there another gadget or process for this? —Anne Delong (talk) 12:22, 16 December 2014 (UTC)[reply]

Anne Delong, are you using Zhaofeng Li's reflinks? Perhaps he could add that for you or explain how to make it happen. :) — {{U|Technical 13}} ^{(e • t • c)} 13:38, 16 December 2014 (UTC)[reply]

Thanks, Technical 13. I was hoping for an existing process, but you are right that Zhaofeng Li is the most likely person to help with this, and I have left a message.—Anne Delong (talk) 18:09, 16 December 2014 (UTC)[reply]

This is actually pretty easy with visual editor. Delete the duplicate, select cite from the menu, and choose re-use. Visual editor will name the ref for you and make it all work swimmingly. Oiyarbepsy (talk) 18:20, 16 December 2014 (UTC)[reply]

For the Reflinks functionality, please continue the discussion at User talk:Zhaofeng Li#Combine duplicate references? Zhaofeng Li [talk... contribs...] 23:40, 18 December 2014 (UTC)[reply]

Campaignbox

I'm having trouble getting Template:Campaignbox Syrian Civil War to look like Template:Campaignbox Afghan War. The |border= parameter doesn't work somehow. Fitzcarmalan (talk) 12:52, 16 December 2014 (UTC)[reply]

Having the same problem in Template:Campaignbox Iraqi insurgency (2011–present) and Template:Campaignbox French Revolutionary Wars. Fitzcarmalan (talk) 12:59, 16 December 2014 (UTC)[reply]

Each of the used templates must implement the border parameter:

|border    = {{{border|}}}<!-- allows this template to be nested -->

PrimeHunter (talk) 13:05, 16 December 2014 (UTC)[reply]

PrimeHunter, can you show me how it is done in the Syrian box, for example? I will do the same with the other ones. Fitzcarmalan (talk) 13:12, 16 December 2014 (UTC)[reply]

The first used template: [18]. PrimeHunter (talk) 13:16, 16 December 2014 (UTC)[reply]

Thanks! Fitzcarmalan (talk) 13:29, 16 December 2014 (UTC)[reply]

Phabricator Profile Pic Placeholder

Does anyone else realize that this placeholder is a silhouette of the Pokemon Psyduck? Who's idea was that? While I am a huge fan of Pokemon, isn't this a copyvio?—cyberpower _Offline^{Merry Christmas} 17:43, 16 December 2014 (UTC)[reply]

You mean https://phabricator.wikimedia.org/rsrc/image/avatar.png is´Psyduck? That looks right. A phabricator search for Psyduck shows mention in phabricator:T65 and phabricator:T256, but no discussion of copyright issues. PrimeHunter (talk) 18:05, 16 December 2014 (UTC)[reply]

I'm pretty sure it doesn't meet the threshold for originality and is PD as such. — {{U|Technical 13}} ^{(e • t • c)} 18:27, 16 December 2014 (UTC)[reply]
- Why wouldn't it meet the threshold for originality? If we can tell that it's Psyduck, which on its own certainly meets the threshold, doesn't that imply that there's enough unique material in the silhouette to qualify for copyright? {{Nihiltres|talk|edits}} 19:04, 16 December 2014 (UTC)[reply]
We have the 'serious' mode enabled in Phabricator, but by default, phab has a lot of non-seriousness and cultural references in much of it's UI texts. I presume this is just one more, but now in outside of the magic that is switched by modes. —TheDJ (talk • contribs) 19:03, 16 December 2014 (UTC)[reply]

Infobox DOB problem

Apologies in advance if this is the wrong place but there are that many forums on here now that I don't know which is which any more.

Anyway, I've recently added full dates of birth to infoboxes in a bunch of articles (specifically Jackie Thompson, Donald Hodgen, Mo Courtney, Winkie Dodds and James Millar (loyalist)) and whilst I have included both day and month only the month is showing up. Have I screwed up somewhere (probable as the technical aspects of editing often go wrong in my hands), is it a glitch, is it my browser (Firefox) or is something else going on? Any help appreciated. Keresaspa (talk) 19:53, 16 December 2014 (UTC)[reply]

@Keresaspa: - it appears that you are using the birth year and age template, rather than the birth date and age template. Connormah (talk) 19:55, 16 December 2014 (UTC)[reply]

Ah, I didn't notice that. I knew it would be me screwing up a technical thingy :D Thanks muchly, Connormah. Keresaspa (talk) 19:57, 16 December 2014 (UTC)[reply]

@Keresaspa: For those people where mdy dates are inappropriate, you should also use the |df=yes parameter to set the display of day-first dates. --Redrose64 (talk) 22:39, 16 December 2014 (UTC)[reply]

Search function much faster

On the Help Desk and in The Teahouse, people have been told for years it could take several days for the Wikipedia search function to find a certain article.

Yesterday, while looking for further information on a topic for which I had just edited a related article, I was amazed to find the edits I had just made minutes earlier, on the screen of the Wikipedia search function!— Vchimpanzee • talk • contributions • 21:33, 16 December 2014 (UTC)[reply]

November 19 mw:Help:CirrusSearch became default on the English Wikipedia. Before that it had been an option called "New search" at Special:Preferences#mw-prefsection-betafeatures for a long time, but it's gone there now it's the default. PrimeHunter (talk) 22:00, 16 December 2014 (UTC)[reply]

Thank you.— Vchimpanzee • talk • contributions • 22:48, 16 December 2014 (UTC)[reply]

Possible to form URL to download article source directly?

Given a specific article title, say Boekenweek, is it possible download its text source directly? I was thinking along the lines of by adding to the article's URL. Maybe there's better way. Jason Quinn (talk) 22:52, 16 December 2014 (UTC)[reply]

When you say "text source" do you mean wikitext, parsed text (what renders on the page when reading), or raw HTML? — {{U|Technical 13}} ^{(e • t • c)} 22:55, 16 December 2014 (UTC)[reply]
- Wikitext. Basically the contents of the edit window. Jason Quinn (talk) 23:12, 16 December 2014 (UTC)[reply]
  - You can play around in the ApiSandbox. A place to start might be https://en.wikipedia.org/wiki/Special:ApiSandbox#action=query&prop=revisions&format=json&rvprop=content&rawcontinue=&titles=Boekenweek - Good luck! — {{U|Technical 13}} ^{(e • t • c)} 23:29, 16 December 2014 (UTC)[reply]

@Jason Quinn: use the basic URL http://en.wikipedia.org/w/index.php?action=raw&ctype=text/plain&title= and append the page name, with spaces replaced by underscores - i.e. the URL http://en.wikipedia.org/w/index.php?action=raw&ctype=text/plain&title=Wikipedia:Village_pump_(technical) will return the source for this page. --Redrose64 (talk) 00:28, 17 December 2014 (UTC)[reply]

That works. Thank you. Jason Quinn (talk) 11:12, 17 December 2014 (UTC)[reply]

And if you're going to do that, in an automatic fashion, please amend your user agent appropriately. Ironholds (talk) 06:04, 18 December 2014 (UTC)[reply]

Lost tools

I seem to have lost many of the tools that made editing here easier, including:

The special characters, four-tilde group etc from below the edit window
User:Ohconfucius/script/MOSNUM dates.js
User:PleaseStand/segregate-refs.js
User:Ohconfucius/script/EngvarB.js
User:Scottywong/diffconverter.js
MediaWiki:Gadget-HotCat.js
Something (can't remember what) that gave me page sizes (total, readable text etc.)

Has there been a change in the software, or could it be because I tried, without success, to install Cat-a-lot by following the instructions here? Can anyone advise on how to get those things back? Justlettersandnumbers (talk) 10:32, 17 December 2014 (UTC)[reply]

If it is because of Cat-a-lot, all you would need to do is undo the addition and purge your cache. עוד מישהו Od Mishehu 10:40, 17 December 2014 (UTC)[reply]

Thank you, Od Mishehu, that has indeed resolved the problem, so it looks as if there is something adrift with the Cat-a-lot instructions or implementation. That is a very useful tool (I've used it on Commons), and it would be excellent if it could be made to work here too. Thanks for your advice, Justlettersandnumbers (talk) 13:46, 17 December 2014 (UTC)[reply]

Hmm, yeah the instructions on that page don't work indeed and will require some updating. JS pages don't allow you to 'render' wikicode anymore... This is part of the stricter separation of the different types of content that the developers are trying to create. Slowly we are moving in a direction where something is either wikicode or javascript, but not both (including the complexities and peculiarities of both). —TheDJ (talk • contribs) 15:29, 17 December 2014 (UTC)[reply]

@Justlettersandnumbers: You attempted to install it by substing a template. Templates are Wiki markup, and Javascript pages expect Javascript, not Wiki markup; so I don't think that subst: would work either. The proper way is like this, as described at User:קיפודנחש#Cat-A-Lot. --Redrose64 (talk) 15:58, 17 December 2014 (UTC)[reply]

Thank you, Redrose64, that has worked and I now have the tools I used before and Cat-a-lot too. I'm most grateful. I did of course see that the subst-ing had failed, but had no idea how to fix it. I attempted to install it in that way because that is what the installation instructions told me to do. It seems that with a few exceptions all other installation instructions on that page also involve subst-ing an {{iusc}} template. Thanks again, Justlettersandnumbers (talk) 16:34, 17 December 2014 (UTC)[reply]

Red, that's entirely wrong and not the problem. {{subst:InstallUserScript}} ({{subst:Iusc}}) is a script specifically designed for substituting user scripts on to the script page. The issue is with the deprecated importScriptURI and importStylesheetURI (see mw:ResourceLoader/Legacy JavaScript#importScript for details) that are used on User:קיפודנחש/cat-a-lot.js. קיפודנחש, or a willing administrator, can you please update the script to use mw.loader.load() instead so that this will work for users again? Thanks. — {{U|Technical 13}} ^{(e • t • c)} 16:43, 17 December 2014 (UTC)[reply]

@TheDJ, Redrose64, and Technical 13: Actually, none of those answers are right. You can still subst templates on JavaScript pages currently. The only reason it didn't work in Special:Diff/637096033 was the unclosed <pre> tag before it. Also, there's no problem at all with importScriptURI or importStylesheetURI yet. Jackmcbarn (talk) 17:04, 17 December 2014 (UTC)[reply]

User:Jackmcbarn is right, of course, and the problem was the "pre" tag. however, due to the pings, i looked at this again, and found that my little three-liner that basically sets a single global variable and then loads one script and one stylesheet from commons was broken: apparently, the comons script developed a new dependency ("jquery.ui.resizable"). so something good came out of it - i made it a four-liner, and now it works again... peace - קיפודנחש (aka kipod) (talk) 04:28, 18 December 2014 (UTC)[reply]

Diffusing template error categories

I've started a discussion about possibly diffusing the template error subcategories of Category:Wikipedia template parameter issues if anyone is interested. The discussion is here. — Mr. Stradivarius ^{♪ talk ♪} 13:00, 17 December 2014 (UTC)[reply]

hover not working

Hi, I'm using a different computer, but still logged in and using Firefox. The hover functions (eg diffs, history and the drop down with (un)watch etc all seem to have gone. I can't see anything in Preferences that I've unticked, what's wrong? Jimfbleak - talk to me? 19:31, 17 December 2014 (UTC)[reply]

Could it be that JavaScript is disabled on that different computer? -- [[User:Edokter]] {{talk}} 19:59, 17 December 2014 (UTC)[reply]

Don't know, how do I find out? Jimfbleak - talk to me? 07:21, 18 December 2014 (UTC)[reply]

@Jimfbleak: Do you have Hovercards turned on in Special:Preferences#mw-prefsection-betafeatures? That has code to make only one or the other of Hovercards and Navpopups display at once - it's slightly buggy in a few browsers hence you might not have encountered the problem before. To fix: You can either click the "cog icon" in a Hovercard to access its preferences menu (screenshots) (which lets us select between the 2 options, or turn it off entirely), or opt-out of the BetaFeature. Either should work fine.

If not, then try either a hard-refresh (ctrl-F5 or WP:REFRESH), or check if your browser has JS turned on/off at http://www.javatester.org/javascript.html . --Quiddity (WMF) (talk) 19:03, 18 December 2014 (UTC)[reply]

Thanks, I'm going to be away for a day or so, but I'll try these when I get back Jimfbleak - talk to me? 19:16, 18 December 2014 (UTC)[reply]

Narrow screen tabs

I have two Internet Explorer 10 browser windows running at part-width, so this one doesn't have room for all the tabs on a normal unprotected article like Matthew 2:23. I get "Article," "Talk," a space, "Read," "Edit," and "More" with a dropdown. If I click "More," I just get "View history." No tabs saved. No real difference. Could you change this, so "More" appears only if two or more tabs are hidden? 65.210.65.16 (talk) 19:58, 17 December 2014 (UTC)[reply]

This only happens when there isn't enough space horizontally to fit in the "View history" tab. The skin is smart enough not to do this when there is enough space, and not to do this if the width of the "More" tab is greater than the width of "View history" (which happens in some languages, but not English). On my computer, the difference is 14 pixels, it varies depending on your browser, OS and available fonts. Matma Rex talk 20:20, 17 December 2014 (UTC)[reply]

history in mobile beta?

Is there a way to access an article's history page in the betatest version of the mobile website? I've never been able to find one. Once I did get to an article's history page when I wasn't trying to, but I have no idea how it happened. Please {{ping}} me for discussion. --Thnidu (talk) 07:48, 18 December 2014 (UTC)[reply]

@Thnidu: If you click the text "last edited..." at the top of the page, it should take you to the edit history. It took me a couple of minutes of clicking different things to get that. I think it needs to be clearer that that is the history link, especially for copyright attribution purposes. Sarahj2107 (talk) 09:05, 18 December 2014 (UTC)[reply]

@Sarahj2107: By golly, you're right! Thank you. --Thnidu (talk) 09:38, 18 December 2014 (UTC)[reply]

Hidden heading

On the article 2 + 2 = 5, there is a heading 'Self-evident truth'. However on my screen, and that of another user, the heading is not visible. That it exists, can be known because it is in the table of contents, and when you view the raw code (with the Edit button), you'll see that it is coded correctly. Apparently the heading is hidden behind the 'Quote box' which is (or should be) above it. When I added a series of hard line-breaks and pressed 'Show preview', the heading became visible, but this is an awkward solution, since it is difficult to guess the number of line-breaks needed (when you insert too many of them, you get whitespace after the quote box) and this may well depend on the user's screen format. Bever (talk) 08:53, 18 December 2014 (UTC)[reply]

I have added a {{clear}} template just above the Self-evident truth header. This should solve the problem without the need to use line-breaks. Sarahj2107 (talk) 08:57, 18 December 2014 (UTC)[reply]

It did, thank you. Bever (talk) 09:12, 18 December 2014 (UTC)[reply]

Am I missing something?

I'm trying to use {{replace}} or alternatively {{str rep}} but it doesn't work. Is there something I need to do? NB this is for a standard page.
I have tried {{str rep | 9°39'N, 123°52'E | N | 3=latNS= }} and it just gives 9°39'N, 123°52'E (unless it works here, that is 9°39'N, 123°52'E). --Unbuttered parsnip (talk) mytime= Thu 20:55, wikitime= 12:55, 18 December 2014 (UTC)[reply]

For some reason, the templates do not strip whitespace between pipes and parameter values:

{{str rep|Foo|o|a}} → Fao
{{str rep|Foo| o |a}} → Foo
{{replace|Foo|o|a}} → Faa
{{replace|Foo| o |a}} → Foo

In the second and fourth cases, the template searches for the string " o " instead of "o", which it cannot find. I'm assuming it's a bug/feature of Lua to take whitespace at the start and end of parameters into account. SiBr₄ (talk) 13:12, 18 December 2014 (UTC)[reply]

(3 ec later) See Help:Template#Parameters, looks like intentional behavior (or at least documented behavior). GermanJoe (talk) 13:30, 18 December 2014 (UTC)[reply]

Never realized that. Is there a reason for the difference in behavior between named and unnamed parameters? SiBr₄ (talk) 14:22, 18 December 2014 (UTC)[reply]

The basic behavior is mentioned in the general meta help for templates as well. Maybe some old-timer in technical questions has more background knowledge - I'm just following the documentation :). GermanJoe (talk) 14:55, 18 December 2014 (UTC)[reply]

It isn't clear to me what it is that you are trying to accomplish with this:

{{str rep | 9°39'N, 123°52'E | N | 3=latNS= }}

Find: N

in string: 9°39'N, 123°52'E

and replace it with: latNS=

Taking out the leading and trailing spaces thus:

{{str rep |9°39'N, 123°52'E|N|3=latNS=}}

gives this result:

9°39'latNS=, 123°52'E

But, why?

—Trappist the monk (talk) 13:49, 18 December 2014 (UTC)[reply]

This is still at the development stge, and was just a test. I want to take coordinates output from wikidata (which is as shown) and turn it into a param string for use inside an infobox. I always develop by prototyping, so no point in getting too far ahead only to have to rework. And I need to make sure I had the concept right, as some documentation is laconic. "Trappist" is such a coincidence! --Unbuttered parsnip (talk) mytime= Fri 02:49, wikitime= 18:49, 18 December 2014 (UTC)[reply]

So far, so good, but now I have hit a problem (of course). The output of #property, putatively 9°39'N, 123°52'E isn't actually as straight forward as that. Presumably the arcminutes sign is an apostrophe, but that's not allowed directly. As far as I can tell, it's not ' either. In fact I don't know what it is - using {{string|pos}} to show the chars, it displays

1.9
2.°
3.3
4.9
5.&
6.3
7.9
1.N
2.,
3.1
4.2
5.3
6.°
7.5
8.2
9.&
10.3
11.9
1.E

Note that sequencing restarted after (first) 7 and 11, but the character itself is not displayed/displayable.
Note too that these coords have no arcseconds. I guess the chars (first) 5-7 and 9-11 are showing the html entity code for apostrophe, &39; except that the ; isn't showing so it could be anything. I also can't find any template to break out a char into hex.

Any thoughts? --Unbuttered parsnip (talk) mytime= Fri 12:31, wikitime= 04:31, 19 December 2014 (UTC)[reply]

Actually... It should be ' (') so you are missing the # too. The MediaWiki parser does this, and there is a ticket someplace about it but I'm too tired to dig it up. I can assure you that it is in fact using ' based on your results for that particular instance. That doesn't mean that it will use if for every instance, and you need to be prepared for that. — {{U|Technical 13}} ^{(e • t • c)} 04:48, 19 December 2014 (UTC)[reply]

Please post code and page link another time, or save the code and post a diff. The hardest part was figuring out what you were already doing. I guess you were on Cogon, Tagbilaran trying code like {{replace|{{#property:P625}}|'|m}} if you for example want to replace the apostrophe by m. It works if you do {{replace|{{#property:P625}}|'|m}}. If you write {{subst:#property:P625}} on the page and click "Show changes" then you can see the apostrophe becomes '. PrimeHunter (talk) 05:20, 19 December 2014 (UTC)[reply]
- Well I was using Cogon as a testbed, but I never saved it there. I just used it because it's small. I save to local file.
- I now have the parsing doing what I want. If you're interested the string latd= {{replace|{{replace|{{replace|{{str rep|{{str rep|{{str rep|{{str rep|{{str rep|{{str rep|{{str rep|{{str rep|{{str rep |{{#invoke:Wikidata|getRawValue|P625|FETCH_WIKIDATA}}|,| {{pipe}}longd{{=}}}}|°|{{pipe}} latm{{=}} }}|°|{{pipe}} longm{{=}} }}|'|{{pipe}} lats{{=}} }}|'|{{pipe}} longs{{=}} }}|N|{{pipe}}latNX{{=}} N}}|S|{{pipe}}latNS{{=}} S}}|E|{{pipe}}longEY{{=}} E}}|W|{{pipe}}longEW{{=}} W}}|"| }}|NX|NS}}|EY|EW}} produces what I was after - here latd= 9| latm= 39| lats= |latNS= N |longd= 123| longm= 52| longs= |longEW= E. Which is fine. It's in the format that {{infobox settlement}} wants. But it doesn't work directly, I guess because it gets treated as a single parameter with a lot of pipes in. So how can I get this code into the infobox? Should I just build a template which emits this stuff? My guess is that wouldn't work either, because it gets parsed too late.
- (BTW my thinking is to produce a more robust template, one that could emit in other infobox formats, using |model=)
- (BTW2 - subst: didn't seem to work)
- Unbuttered parsnip (talk) mytime= Fri 15:50, wikitime= 07:50, 19 December 2014 (UTC)[reply]
  - Use {{subst:!}} instead of {{Pipe}} and {{subst:=}} instead of {{=}}. You should actually be substituting all the parser funtions and template calls if you want it to work right in the end. — {{U|Technical 13}} ^{(e • t • c)} 08:06, 19 December 2014 (UTC)[reply]

The symbol for arcminutes isn't an apostrophe, but a prime, U+2032 ′ or ′ - it's closer to the vertical than a curly (typographic) apostrophe, U+2019 ’ or ’. Similarly that for arcseconds is a double prime, U+2033 ″ or ″ - it's closer to the vertical than a curly (typographic) quote, U+201D ” or ”. The prime and double prime are the symbols emitted by the {{coord}} template. Note that some fonts may show all of these symbols at the same angle - even vertical. --Redrose64 (talk) 09:07, 19 December 2014 (UTC)[reply]

Problems with page curation

Page curation isn't working properly for me right now. The mark as patrolled button doesn't work and the tagging button doesn't work. However, the next article and info buttons do work. This problem started around 12 hours ago, as far as I know. Mac OS 10.7.5 with Safari 6.1.6. Oiyarbepsy (talk) 15:29, 18 December 2014 (UTC)[reply]

Block during page edit

When we need to record something in a block log, it's common to perform a 1-second block; for example, I unblocked a user recently without clear explanation, so I reblocked him for 1 second with a fuller explanation. What happens if the block occurs while you're editing a page? You click "edit", and while you're writing away, someone blocks you for 1 second, and of course this is long past when you hit "save". Will the software accept it, since you're not blocked at the moment you hit "save", or will it have some odd editconflict-type screen? Nyttend (talk) 15:29, 18 December 2014 (UTC)[reply]

I just tested it on testwiki and there was no edit conflict or any indication on the user's end that there was a one second block[19]. My question is, how are you blocking for one second when that is not a listed option on MediaWiki:Ipboptions. Yes, I'm aware it can be typed in to "other duration", but that seems odd because if that is a used parameter, I would expect it to have been added to make things easier and ensure duration. I'm also wondering why there is redundancy in the options on that page; indefinite:indefinite is double listed. — {{U|Technical 13}} ^{(e • t • c)} 15:57, 18 December 2014 (UTC)[reply]

I only see it in the list once - at the top of the drop-down box, but there appears to have been an agreement back in 2006 (MediaWiki talk:Ipboptions#Indef) to have it at both the top and the bottom. Perhaps the software has been changed since then to remove redundant entries? – Philosopher ^{Let us reason together.} 20:47, 18 December 2014 (UTC)[reply]

As you say, just typing it into the "other". I'm glad that 1 second isn't an option: there's no reason to place such a short block if you're not adding a log entry, and if it were an option, someone might pick it by accident when intending to place a much longer block for someone who really is misbehaving. Thank you for the technical confirmation! Nyttend (talk) 21:56, 18 December 2014 (UTC)[reply]

Interestingly (well, to me), you can block someone for zero seconds: [20]. --Floquensock (talk) 22:18, 18 December 2014 (UTC)[reply]

Weird. I just blocked ThisIsaTest for 0 years after getting an "invalid expiry" warning when I tried to place 0-decade and 0-century blocks. However, the block log says "0 seconds", not "0 years". Nyttend (talk) 22:38, 18 December 2014 (UTC)[reply]

Try blocking someone for "a potato". (Note: please don't actually do this.) Matma Rex talk 23:01, 18 December 2014 (UTC)[reply]

Why not? Will something go wrong? It looks like it just won't have the desired effect, and ThisIsaTest won't care: the sole purpose of the account is for blocking tests and blocking practice. Nyttend (talk) 01:19, 19 December 2014 (UTC)[reply]

You can put in all kinds of nonsense in that field, and get unusual results. — xaosflux ^Talk 01:59, 19 December 2014 (UTC)[reply]

See also Gnu standard format. — xaosflux ^Talk 02:07, 19 December 2014 (UTC)[reply]

I see that "a potato" is apparently -3600 seconds. — Mr. Stradivarius ^{♪ talk ♪} 03:05, 19 December 2014 (UTC)[reply]

More script errors: "The time allocated for running scripts has expired"

I see in the archives some technical discussion regarding this error (see this for a somewhat recent example. I don't know the particulars of why this happens, or how to resolve it, but it is currently happening at the Chris Sarandon article (twice at the top of the page, and eight times in references and external links). Assistance please? --Hammersoft (talk) 16:10, 18 December 2014 (UTC)[reply]

OK, seems to have cleared up now...no idea how or why. --Hammersoft (talk) 19:09, 18 December 2014 (UTC)[reply]

Worry about it if it persists after a null edit. I introduced a lot of script errors the last time I updated the modules that render Citation Style 1 templates. In that particular case, there were four pages sequentially updated one after the other. For a brief period of time the various combinations of new and old code were incompatible and so caused errors in several hundred articles. So, wait a bit and then try a null edit.

—Trappist the monk (talk) 19:28, 18 December 2014 (UTC)[reply]

List of Wikipedians by article count

Could a tech-savvy person please offer some advice at Wikipedia talk:List of Wikipedians by article count#Before reviving? That list is moribund, and was never implemented with the opt-out feature of lists such as Wikipedia:List of Wikipedians by number of edits. A number of people would like to see the list being regularly updated, and another number of people want to be sure that opting out is taken care of. I'd be happy to help with this, but am clueless about how to run or build such a thing. Can someone advise? Sminthopsis84 (talk) 17:23, 18 December 2014 (UTC)[reply]

Usercheck template out of date?

Moved from Wikipedia talk:Arbitration Committee Elections December 2014 § Usercheck template out of date?

I was just looking at the {{Usercheck}} template used for these elections, and I noticed that it seems to be out of date. At some point, ArbCom deprecated the Wikipedia:Requests for arbitration/ scheme in favor of a new Wikipedia:Arbitration/Requests/ scheme. This skews the results returned by the template into thinking that editors who joined wikipedia after the date of that change have no experience at all and have never been involved in any kind of arbitration. This misinformation of the people could, as a result of this information, skewed the results of this election. I'm not calling foul, and I actually did better than I expected. I just wonder what the results might have been if the information was correct. In an effort to prevent this very likely issue for future elections, I would like to propose a modification to the template. I would like to propose that the link be changed:

From: Mr vili / Example

arb / arb

Using this code

{{#ifexist: Wikipedia:Requests for arbitration/{{ucfirst:{{{User|{{{1}}}}}}}}|[[Special:Prefixindex/Wikipedia:Requests for arbitration/{{ucfirst:{{{User|{{{1}}}}}}}}|arb]]|[[{{SITENAME}}:Requests for arbitration#{{ucfirst:{{{User|{{{1}}}}}}}}|<span title="There is no {{SITENAME}}:Requests for arbitration/{{ucfirst:{{{User|{{{1}}}}}}}}" style="color:gray;">arb</span>]]}}

To:

1

arb / arb

Using this code

{{#ifexist: Wikipedia:Arbitration/Requests/Case/{{ucfirst:{{{User|{{{1|{{REVISIONUSER:Wikipedia:Administrators' noticeboard/Incidents‎‎}}}}}}}}}}|[[Special:Prefixindex/Wikipedia:Arbitration/Requests/Case/{{ucfirst:{{{User|{{{1|{{REVISIONUSER:Wikipedia:Administrators' noticeboard/Incidents‎‎}}}}}}}}}}|arb]]|<span class="plainlinks" title="There are no specific cases naming {{ucfirst:{{{User|{{{1|{{REVISIONUSER:Wikipedia:Administrators' noticeboard/Incidents‎‎}}}}}}}}}}; Click to search for experience">[{{fullurl:Special:Search|profile=default&fulltext=Search&search={{urlencode:"{{ucfirst:{{{User|{{{1|{{REVISIONUSER:Wikipedia:Administrators' noticeboard/Incidents‎‎}}}}}}}}}}"|WIKI}}+prefix:Wikipedia/Arbitration/Requests}} <span style="color: #808080;">arb</span>]</span>}}

2

arb / arb

Using this code

{{#ifexist: Wikipedia:Arbitration/Requests/Case/{{ucfirst:{{{User|{{{1|{{REVISIONUSER:Wikipedia:Administrators' noticeboard/Incidents‎‎}}}}}}}}}}|[[Special:Prefixindex/Wikipedia:Arbitration/Requests/Case/{{ucfirst:{{{User|{{{1|{{REVISIONUSER:Wikipedia:Administrators' noticeboard/Incidents‎‎}}}}}}}}}}|arb]]|[[Wikipedia:Arbitration/Requests/Case#{{ucfirst:{{{User|{{{1|{{REVISIONUSER:Wikipedia:Administrators' noticeboard/Incidents‎‎}}}}}}}}}}|<span style="color: #808080;">arb</span>]]}}

3

arb (search) / arb (search)

Using this code

{{#ifexist: Wikipedia:Arbitration/Requests/Case/{{ucfirst:{{{User|{{{1|{{REVISIONUSER:Wikipedia:Administrators' noticeboard/Incidents‎‎}}}}}}}}}}|[[Special:Prefixindex/Wikipedia:Arbitration/Requests/Case/{{ucfirst:{{{User|{{{1|{{REVISIONUSER:Wikipedia:Administrators' noticeboard/Incidents‎‎}}}}}}}}}}|arb]]|[[Wikipedia:Arbitration/Requests/Case#{{ucfirst:{{{User|{{{1|{{REVISIONUSER:Wikipedia:Administrators' noticeboard/Incidents‎‎}}}}}}}}}}|<span style="color: #808080;">arb</span>]]}} <span class="plainlinks" title="There are no specific cases naming {{ucfirst:{{{User|{{{1|{{REVISIONUSER:Wikipedia:Administrators' noticeboard/Incidents‎‎}}}}}}}}}}; Search all cases.">([{{fullurl:Special:Search|profile=default&fulltext=Search&search={{urlencode:"{{ucfirst:{{{User|{{{1|{{REVISIONUSER:Wikipedia:Administrators' noticeboard/Incidents‎‎}}}}}}}}}}"|WIKI}}+prefix:Wikipedia/Arbitration/Requests}} search])</span>

For these examples, I'm throwing in the username of the last person to edit WP:AN/I (Mr vili) in order to get a wide variety of results. Thank you for your time and consideration on this. — {{U|Technical 13}} ^{(e • t • c)} 16:45, 18 December 2014 (UTC)[reply]

~~Partial Oppose~~- I oppose changing the link into a search link if no pages are found as it makes the template less useful in other places as a quick spot check. Note in particular mobile editors often can't see the pop-up notes on links easy, and so making the link blue always hides information. Usercheck is used in other places besides arb elections and in most cases experience in dealing with arbitration cases is not what is being searched. PaleAqua (talk) 18:43, 18 December 2014 (UTC)[reply]
- PaleAqua, I've modified the proposed replacement to address your concerns. :) — {{U|Technical 13}} ^{(e • t • c)} 19:02, 18 December 2014 (UTC)[reply]
  - Thanks, Click to search for experience still seems mostly targeted towards elections and also seems to be a bit NPOV. And a little bike sheding but how about split the search link out. i.e. something like "arb (search)" where arb is only linked if a page is found and search is always available. PaleAqua (talk) 19:18, 18 December 2014 (UTC)[reply]
    - I'd be happy to make other modifications once I get back to a computer. What would you think about using my exact example replacement in a new {{ACEcandidate}} wrapper template? Where else is this used exactly? I'd love to make it suit those needs better as well. — {{U|Technical 13}} ^{(e • t • c)} 19:38, 18 December 2014 (UTC)[reply]
    - I've added a few more options. I hope there is one everyone likes. :) — {{U|Technical 13}} ^{(e • t • c)} 22:14, 18 December 2014 (UTC)[reply]
      - 3 of course is my choice, going to drop out of the discussion for now though so you can hopefully get other input as I have no desire to go to further down the bike trail I started us towards. PaleAqua (talk) 01:59, 19 December 2014 (UTC)[reply]
        I missed you wanted search always available. I've fixed it in example 3. :) — {{U|Technical 13}} ^{(e • t • c)} 02:15, 19 December 2014 (UTC)[reply]

Links to log entries in Atom feed

For usergroup changes, the watchlist feed links to "diff=0" of the user's userpage (e.g. https://en.wikipedia.org/w/index.php?title=User:GoingBatty&diff=0) instead of Special:Log. Is it a known bug? Zhaofeng Li [talk... contribs...] 03:50, 19 December 2014 (UTC)[reply]

Orphaned non-free use files that are not orphaned

I regularly patrol Category:Orphaned non-free use Wikipedia files to see if the files are right for deletion and fix what I can to keep proper files on Wikipedia. Recently I have been noticing when an editor blanks either a section or the entire article and is quickly reverted within a minute, in the cases listed below by User:ClueBot NG, the files show that they are not being used in any article on Wikipedia. Eventually this ends up in where ever User:Stefan2 finds the list and he does his great work and tags them an being orphaned. Once I find them, I know a null edit is all it takes for me to make the article reappear in the file's usage section and then I remove the orphaned fair use template from the file.

Examples from yesterday's list that I have not done a null edit on so the blank usage sections can be seen.

Adobe InDesign
1. File:Adobe InDesign CS6.png
The Hitchhiker's Guide to the Galaxy
But weirdly enough, the file in the infobox, File:H2G2 UK front cover.jpg, has the article listed in its usage section.

This seems like more extra work for both Stefan2, myself and any other editor who might be fighting vandalism or just regular editing. Is there anything that can be done in these cases or is this something we will just have to live with. Aspects (talk) 07:09, 19 December 2014 (UTC)[reply]

I think we'll have to wait to hear from Stefan to see which, if any, tool/script he uses to tag them as orphaned... If he is using a script, the script could do a check to see if the images are actually orphaned before applying the tag to save everyone effort (except the script code writer who would have to set it up XD). — {{U|Technical 13}} ^{(e • t • c)} 07:39, 19 December 2014 (UTC)[reply]

@Technical 13: The problem here is that the "what links here" information isn't being updated after ClueBot's revert. How is a script supposed to check whether an image is "actually orphaned" if it can't trust "what links here"? -- John of Reading (talk) 08:40, 19 December 2014 (UTC)[reply]

John, it's a script, it can purge to update the tables and then check WLH. — {{U|Technical 13}} ^{(e • t • c)} 09:21, 19 December 2014 (UTC)[reply]

@Technical 13:A purge or null edit of the file page does not fix the problem; you have to guess which article has the image and do a null edit there. A script won't be able to guess correctly. -- John of Reading (talk) 09:55, 19 December 2014 (UTC)[reply]

Since we're talking about non-free files, WP:NFCC#10c says that there must be "a separate, specific non-free use rationale for each use of the item" together with "the name of each article ... in which fair use is claimed for the item". Therefore, there should be a valid WP:FUR on the file description page naming the article where the image is intended to be used; it will probably be linked (but might not be), so follow that link, and WP:NULLEDIT the resulting article. Do this for every FUR on the file description page (there might be more than one, example), and if after processing all the FURs in this way, there are still no articles listed under "File usage", the file may be safely put up for WP:CSD#F5. If there are some left, check that each one of the linked articles has a FUR on the file page - if you find an article which uses the image but there is no FUR explicitly naming that article, the image may be removed from the article under WP:NFCC#10c; if after doing this, there are no articles listed under "File usage", F5 also applies. I don't think the task can be fully automated: many FURs are in template form, with an |Article= parameter which may be read, but a valid FUR may be constructed using no templates whatsoever. --Redrose64 (talk) 14:19, 19 December 2014 (UTC)[reply]

If MediaWiki doesn't update Special:WhatLinksHere when User:ClueBot NG edits pages, then there is a bug in MediaWiki which should be fixed. If Special:WhatLinksHere can't be trusted, then this can cause problems in various parts of Wikipedia, not only with F5 tagging.

If MediaWiki doesn't update Special:WhatLinksHere when User:ClueBot NG edits pages, then maybe User:ClueBot NG could POST action=purge with the forcelinkupdate parameter after each edit to reduce the damage.

I could maybe write a script which searches for articles based on WP:NFCC#10c requirements and purges the articles. I'll try to see if there is something I can do here and how difficult it would be to write that script. When I tag the files, I open them in my browser and tag them using Twinkle. --Stefan2 (talk) 15:11, 19 December 2014 (UTC)[reply]

Template:Imdb-film-year

I was wondering if somebody could create a template. Sort of like Template:Imdb-japan-year but to cover all countries/languages. They're to go at the end of all the film year list like American films of 1955, South Korean films of 1982, Spanish films of 1990 etc. I think it would need some sort of coding like "if Japanese =jp" etc. Bascially there's a common format linking to a list on imdb, see the url for Japanese films of 1964 vs American films of 1964 It is controlled solely by the short country code jp for Japan, us for USA etc. What I want is to be able for the template to read what the year is in something like List of Japanese films of 1964 and for it to automatically or have to do very little in the current List of Japanese films of 1975. I want to be able to have one template I can use as an external link in all the country year lists linking to the appropriate page. So whoever creates it would need to find the index of countries on imdb and code it so it reads all shortcodes for each country by year. Do you follow? Ideally I want something which will read a year list like List of Australian films of 1987 and all I will have to do is add a Template:Imdb-film-year in the external link and it will read the Australian and 1987 in the title and correctly link to the correct page on imdb.♦ Dr. Blofeld 13:35, 19 December 2014 (UTC)[reply]

Something like: {{Imdb-country-year|jp|1964}}? Or do you want it to be fully automatic where it might read the article title, for example Spanish films of 1990 and from that extract the country and year and populate the appropriate places in the url?

—Trappist the monk (talk) 14:15, 19 December 2014 (UTC)[reply]

@@ Line 250: / Line 250: @@
 I suppose that the encyclopedia that anyone can edit might have untrusted user input in it? I wonder if it's possible to track down everyone who is still using the botclasses library. How serious an issue is this? I might take a stab at the necessary code changes soon if nobody else does. [[User:Wbm1058|Wbm1058]] ([[User talk:Wbm1058|talk]]) 16:31, 17 December 2014 (UTC)
 :{{ping|Wbm1058}} Just replace <code>format=php</code> with json, and replace the unserialize line with <code>return json_decode($ret,true);</code>. By the way, the real risk is not from the editors, since the API will properly escape whatever they write. The problem is you don't know whether the API you are talking to is the real one or not (Using HTTPS will help a lot), and a fake one could return a maliciously-crafted string as the result. <span style="text-shadow: 0.1em 0.1em 0.2em black">[[User:Zhaofeng Li|Zhaofeng Li]]</span> <small><nowiki>[</nowiki>[[User talk:Zhaofeng Li|talk...]] [[Special:Contributions/Zhaofeng Li|contribs...]]<nowiki>]</nowiki></small> 04:51, 18 December 2014 (UTC)
+::{{ping|Zhaofeng Li}} Thanks, I made the JSON changes and it worked fine. But when I changed "http:" to "https:" it responded with "Login error: ". Does the API not support secure http or is there something else I need to change to get it to work? [[User:Wbm1058|Wbm1058]] ([[User talk:Wbm1058|talk]]) 16:59, 19 December 2014 (UTC)
 == [[m:Special:MyLanguage/Tech/News/2014/51|Tech News: 2014-51]] ==