Wikipedia:Village pump (technical)

 Policy Technical Proposals Idea lab WMF Miscellaneous 

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

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

Newcomers to the technical village pump are encouraged to read these guidelines prior to posting here. Questions about MediaWiki in general should be posted at the MediaWiki support desk.

« Archives, 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, 214, 215

Table cell contents spilling over.

. When viewing the following table in the mobile version of the site, some text from the first column spills over into the adjacent cell of the second column.

Team	Constructor	Chassis	Power unit	Tyre	No.	Drivers
Scuderia Ferrari	Ferrari	SF15-T	Ferrari	P	5 7	Sebastian Vettel Kimi Räikkönen
Sahara Force India F1 Team	Force India-Mercedes	VJM08	Mercedes PU106B Hybrid	P	11 27	Sergio Pérez Nico Hülkenberg
Lotus F1 Team	Lotus-Mercedes	E23 Hybrid	Mercedes PU106B Hybrid	P	8 13	Romain Grosjean Pastor Maldonado
Manor Marussia F1 Team	Marussia-Ferrari	TBA	Ferrari 059/3	P	TBA TBA	Will Stevens TBA
McLaren Honda	McLaren-Honda	MP4-30	Honda RA615H Hybrid	P	14 22	Fernando Alonso Jenson Button
Mercedes AMG Petronas F1 Team	Mercedes	F1 W06 Hybrid	Mercedes PU106B Hybrid	P	6 44	Nico Rosberg Lewis Hamilton
Infiniti Red Bull Racing	Red Bull-Renault	RB11	Renault Energy F1-2015	P	3 26	Daniel Ricciardo Daniil Kvyat
Sauber F1 Team	Sauber-Ferrari	C34	Ferrari	P	9 12	Marcus Ericsson Felipe Nasr
Scuderia Toro Rosso	Toro Rosso-Renault	STR10	Renault Energy F1-2015	P	33 55	Max Verstappen Carlos Sainz Jr.
Williams Martini Racing	Williams-Mercedes	FW37	Mercedes PU106B Hybrid	P	19 77	Felipe Massa Valtteri Bottas
Source:

Anyone got an idea what's causing this and/or how to solve this. Tvx1 22:24, 28 February 2015 (UTC)

I don't have a mobile device and it looks right for me in both desktop and https://en.m.wikipedia.org/w/index.php?title=Wikipedia:Village_pump_(technical)&mobileaction=toggle_view_mobile. The table has a coding error in {{nowrap|{{nowrap|Mercedes PU106B Hybrid}} which should only have one {{nowrap}}. Does it help to remove that:

Team	Constructor	Chassis	Power unit	Tyre	No.	Drivers
Scuderia Ferrari	Ferrari	SF15-T	Ferrari	P	5 7	Sebastian Vettel Kimi Räikkönen
Sahara Force India F1 Team	Force India-Mercedes	VJM08	Mercedes PU106B Hybrid	P	11 27	Sergio Pérez Nico Hülkenberg
Lotus F1 Team	Lotus-Mercedes	E23 Hybrid	Mercedes PU106B Hybrid	P	8 13	Romain Grosjean Pastor Maldonado
Manor Marussia F1 Team	Marussia-Ferrari	TBA	Ferrari 059/3	P	TBA TBA	Will Stevens TBA
McLaren Honda	McLaren-Honda	MP4-30	Honda RA615H Hybrid	P	14 22	Fernando Alonso Jenson Button
Mercedes AMG Petronas F1 Team	Mercedes	F1 W06 Hybrid	Mercedes PU106B Hybrid	P	6 44	Nico Rosberg Lewis Hamilton
Infiniti Red Bull Racing	Red Bull-Renault	RB11	Renault Energy F1-2015	P	3 26	Daniel Ricciardo Daniil Kvyat
Sauber F1 Team	Sauber-Ferrari	C34	Ferrari	P	9 12	Marcus Ericsson Felipe Nasr
Scuderia Toro Rosso	Toro Rosso-Renault	STR10	Renault Energy F1-2015	P	33 55	Max Verstappen Carlos Sainz Jr.
Williams Martini Racing	Williams-Mercedes	FW37	Mercedes PU106B Hybrid	P	19 77	Felipe Massa Valtteri Bottas
Source:

Does it help to remove all nowrap (may be controversial in an article), or for simplicity replace them by {{identity}} as here:

Team	Constructor	Chassis	Power unit	Tyre	No.	Drivers
Scuderia Ferrari	Ferrari	SF15-T	Ferrari	P	5 7	Sebastian Vettel Kimi Räikkönen
Sahara Force India F1 Team	Force India-Mercedes	VJM08	Mercedes PU106B Hybrid	P	11 27	Sergio Pérez Nico Hülkenberg
Lotus F1 Team	Lotus-Mercedes	E23 Hybrid	Mercedes PU106B Hybrid	P	8 13	Romain Grosjean Pastor Maldonado
Manor Marussia F1 Team	Marussia-Ferrari	TBA	Ferrari 059/3	P	TBA TBA	Will Stevens TBA
McLaren Honda	McLaren-Honda	MP4-30	Honda RA615H Hybrid	P	14 22	Fernando Alonso Jenson Button
Mercedes AMG Petronas F1 Team	Mercedes	F1 W06 Hybrid	Mercedes PU106B Hybrid	P	6 44	Nico Rosberg Lewis Hamilton
Infiniti Red Bull Racing	Red Bull-Renault	RB11	Renault Energy F1-2015	P	3 26	Daniel Ricciardo Daniil Kvyat
Sauber F1 Team	Sauber-Ferrari	C34	Ferrari	P	9 12	Marcus Ericsson Felipe Nasr
Scuderia Toro Rosso	Toro Rosso-Renault	STR10	Renault Energy F1-2015	P	33 55	Max Verstappen Carlos Sainz Jr.
Williams Martini Racing	Williams-Mercedes	FW37	Mercedes PU106B Hybrid	P	19 77	Felipe Massa Valtteri Bottas
Source:

PrimeHunter (talk) 04:56, 1 March 2015 (UTC)

Thanks for mentioning that coding error. That didn't cause the problem however. Actually it did not do any harm at all. If you click on the link to the mobile view and then reduce the width of your browser screen to the minimum you will see the text from the first column spilling over. Using the identity template doesn't solve it. Tvx1 05:38, 1 March 2015 (UTC)

I already tried the minimum width on mobile and it works for me. I get a horizontal scroll bar and no overlap. I will stop guessing. It's too hard when I don't have the problem. PrimeHunter (talk) 05:51, 1 March 2015 (UTC)

Since I'm not able to explain the issue with text, I have made a screenshot:

As you can see, content from the first column is spilling over into the second. Tvx1 21:20, 16 March 2015 (UTC)

It doesn't spill over for me. Based on the amount of spillover in your screenshot, maybe your browser doesn't reserve space for the flag icon when the column width is calculated. PrimeHunter (talk) 00:05, 20 March 2015 (UTC)

Multiple user have reported this to me though, regardless of which browser they use. Tvx1 06:54, 21 March 2015 (UTC)

I think you're spot on there. Either the flag icon isn't taken into account, either it's just counted as a 1px character. Is there a way to solve this? Tvx1 19:07, 26 March 2015 (UTC)

Does it change anything if the size of the image is increased, or removed altogether? Googol30 (talk) 04:06, 27 March 2015 (UTC)

Well, cells without flagicons don't spill over into other, no matter how wide they are. But we want to have them in that table. Tvx1 10:13, 28 March 2015 (UTC)

And does making the icons larger make the problem worse? What I'm trying to determine is how the browser is interpreting what it's given and why it isn't making the cells large enough. If the browser isn't making sufficient space for all of the elements the cell contains, should we worry about that, since it's then the fault of the browser, or find a workaround and try to fix the problem on our end, even if the mobile browser isn't working as expected? I didn't seem to catch what your mobile browser even is. If I could replicate the problem on my end, I could attempt to fix it, but the desktop version of Chrome I'm running handles element widths as expected. Let me add a table with the changes I'm thinking of for illustration:

Team	Constructor	Chassis	Power unit	Tyre	No.	Drivers
Scuderia Ferrari	Ferrari	SF15-T	Ferrari	P	5 7	Sebastian Vettel Kimi Räikkönen
Sahara Force India F1 Team	Force India-Mercedes	VJM08	Mercedes PU106B Hybrid	P	11 27	Sergio Pérez Nico Hülkenberg
Lotus F1 Team	Lotus-Mercedes	E23 Hybrid	Mercedes PU106B Hybrid	P	8 13	Romain Grosjean Pastor Maldonado
Manor Marussia F1 Team	Marussia-Ferrari	TBA	Ferrari 059/3	P	TBA TBA	Will Stevens TBA
McLaren Honda	McLaren-Honda	MP4-30	Honda RA615H Hybrid	P	14 22	Fernando Alonso Jenson Button
Mercedes AMG Petronas F1 Team	Mercedes	F1 W06 Hybrid	Mercedes PU106B Hybrid	P	6 44	Nico Rosberg Lewis Hamilton
Infiniti Red Bull Racing	Red Bull-Renault	RB11	Renault Energy F1-2015	P	3 26	Daniel Ricciardo Daniil Kvyat
Sauber F1 Team	Sauber-Ferrari	C34	Ferrari	P	9 12	Marcus Ericsson Felipe Nasr
Scuderia Toro Rosso	Toro Rosso-Renault	STR10	Renault Energy F1-2015	P	33 55	Max Verstappen Carlos Sainz Jr.
Williams Martini Racing	Williams-Mercedes	FW37	Mercedes PU106B Hybrid	P	19 77	Felipe Massa Valtteri Bottas
Source:

Of course, this is an extreme size change, but it's to troubleshoot things here. Does this make the problem worse, fix it, or make no difference whatsoever? Googol30 (talk) 23:42, 28 March 2015 (UTC)

Googol30, that makes the problem much worse. Here is a screenshot:

In case I hadn't made it clear yet, this is an issue that only occurs on the mobile version of the site. The only mobile browser I have thus far identified not to be affected by this issue is Firefox. All other mobile browser have this problem. I mostly use mobile Safari, but as said other mobile browsers are affected as well. Tvx1 00:37, 29 March 2015 (UTC)

Googol30, does my above reply hold any value for you regarding this issue? Tvx1 13:25, 30 April 2015 (UTC)

@Tvx1: yes, and I've done some research into the problem, finding that this is actually intended behavior of the nowrap attribute. Sadly, I do not have enough technical knowledge of its use within MediaWiki or on Wikipedia to properly fix this issue, so I regret to inform you that although (I think) I know what is causing the problem, I cannot properly fix it without potentially causing more problems elsewhere. Googol30 (talk) 11:04, 4 May 2015 (UTC)

And do you know anyone who could be able to fix it, or do you think it would be better to file a bug report over on Phabricator? Tvx1 15:55, 4 May 2015 (UTC)

References
References

Very slow category population from stub templates

I created this category 10 days ago, and it's still not fully populated from the stub tag. The transclusion count states there are 70 articles with the stub, but so far it's showing 56 articles in the category (it had about 40 a few days ago). Why is it taking so long to update? Thanks. Lugnuts ^{Dick Laurent is dead} 07:53, 29 April 2015 (UTC)

The job queue was altered for Parsoid (IIRC) in mid 2013, since when it hasn't worked as well as it did previously. Judging by mw:API:Purge, visiting https://en.wikipedia.org/w/api.php?action=purge&forcerecursivelinkupdate&titles=Template:Mexico-cycling-bio-stub may get the counts straight again. --Redrose64 (talk) 08:59, 29 April 2015 (UTC)

Thanks. I've created a couple of similar categories in the last two days - will check their progress in June! Lugnuts ^{Dick Laurent is dead} 09:45, 29 April 2015 (UTC)

In the past couple of years, I have seen category population from templates take up to 100 days. It can be quite frustrating. I submitted a bug on Bugzilla somewhere, and there is a post in the VPT archives, I believe, but it's been a while. If you search for my user name, you might find it. – Jonesey95 (talk) 13:08, 1 May 2015 (UTC)

Its the bonus from one template fits all. Anybody know whats the setting for the job queue is? (fifo, random or timestamp) Christian75 (talk) 19:54, 2 May 2015 (UTC)

Still stuck at 58. Lugnuts ^{Dick Laurent is dead} 18:02, 2 May 2015 (UTC)

Right, I've made this edit. It is not a WP:NULLEDIT, since a diff exists (and it shows in the page history, my contribs, etc.) but it has no effect on the action of the template. However, it should get the transcluding pages into the job queue. --Redrose64 (talk) 19:01, 2 May 2015 (UTC)

62! Lugnuts ^{Dick Laurent is dead} 07:34, 5 May 2015 (UTC)

62! Still. Lugnuts ^{Dick Laurent is dead} 13:15, 8 May 2015 (UTC)

I've just looked through all 70 transclusions of the template, and they all say that they are in the category. It occurs to me that this may not be a job-queue problem, but a category-counting bug (tracked as phab:23230). I'm just about to bot-null-edit all of the transclusions to see what happens. Watch this space... — Mr. Stradivarius ^{♪ talk ♪} 14:57, 8 May 2015 (UTC)

@Lugnuts and Mr. Stradivarius: As Technical 13 noted elsewhere on this page, the job queue has been down for some days, so when a template is changed, the changes are only propagating through to articles as and when those articles are edited; and only at that point are the cats populated, one article at a time. I've been deliberately refraining from WP:NULLEDITing the transcluding pages, in order to see when the job queue picks up again. --Redrose64 (talk) 15:33, 8 May 2015 (UTC)

• According to Legoktm, the job queue has been back up. There is something else going on because I've watched it go from 19.7 million yesterday morning to about 21.8 million right now and still climbing at a steady rate. I'm wondering at what threshold it will be time for a few more resources to be given to the queue to knock it back down to acceptable levels (or what an acceptable level even is). — {{U|Technical 13}} (e • t • c) 15:50, 8 May 2015 (UTC)

Well the null edits worked, so it was definitely the job queue that was the problem. Sorry for ruining the job-queue test. (Although maybe something in userspace could be used to test that instead?) — Mr. Stradivarius ^{♪ talk ♪} 15:56, 8 May 2015 (UTC)

Edit Tags

I just noticed that there is a new feature on Commons that allows users to edit tags on edits (Example)? I seem to remember reading about this somewhere, but what is the purpose and is it coming to en.wiki soon? If so, how will it be used? Sorry if this has already been asked... Thanks! EoRdE6^{(Come Talk to Me!)} 19:20, 29 April 2015 (UTC)

I suspect that it will - English Wikipedia tends to get new MediaWiki versions a day or two after Commons. phab:T20670 might be relevant. --Redrose64 (talk) 19:53, 29 April 2015 (UTC)

I found where you read it - Wikipedia:Village pump (technical)/Archive 136#Tech News: 2015-17, under "Future changes". --Redrose64 (talk) 19:58, 29 April 2015 (UTC)

Within the past few minutes, I noticed something the same thing on the English Wikipedia. Dustin (talk) 21:41, 29 April 2015 (UTC)

Though I can see some benefit to removing tags if they're inappropriate, I'm not sure I see the benefit to editing or manually adding tags. I like knowing that tagged edits have been tagged as a result of a strict rule set (an edit filter) which can be checked and improved if it's not tagging the correct edits. Under what circumstances would someone want to add or edit a tag? If an edit is problematic surely we can just deal with it there and then. Sam Walton (talk) 23:01, 29 April 2015 (UTC)

It also seems that any user (unconfirmed accounts included) can edit tags - shouldn't we have some higher bar restricting this feature? Especially seeing as there doesn't appear to be any obvious way to track tag changes. Sam Walton (talk) 23:03, 29 April 2015 (UTC)

Nevermind, I just read the original email (I should have RTFM), and it seems my concerns are largely unnecessary. Sam Walton (talk) 23:10, 29 April 2015 (UTC)

I still don't understand the purpose here, I can click edit, look at the tags (ie. Possible libel or vandalism) and do absolutely nothing. Can anyone elaborate on what is the point of this? The email didn't really help. EoRdE6^{(Come Talk to Me!)} 00:17, 30 April 2015 (UTC)

Unrelated note, should the HHVM tag be marked as inactive now in Special:Tags? EoRdE6^{(Come Talk to Me!)} 00:17, 30 April 2015 (UTC)

Should WP:Tags be updated? --I am k6ka ^{Talk to me!} _{See what I have done} 01:52, 30 April 2015 (UTC)

So another little gadget that's useless / of little use to the community .... Why am I not surprised ?..... –Davey2010^Talk 02:17, 30 April 2015 (UTC)

This feature was implemented by a volunteer developer. [1]. --(ʞɿɐʇ) ɐuɐʞsǝp 02:28, 30 April 2015 (UTC)

If this feature is mainly to let bots and scripts add tags, and let experienced users remove false-positive tags, then why not have the checkboxes and edit tags button hidden by default, and let the experienced users unhide them with CSS (maybe as a gadget, and maybe unhidden by default for all admins)? - Evad37 [talk] 02:44, 30 April 2015 (UTC)

Yeah I feel like as it stands it serves only to really confuse new users (and experienced ones :p). Also, what is the point of removing false positive tags? Not only pointless you could spend days doing that and achieve nothing. EoRdE6^{(Come Talk to Me!)} 02:48, 30 April 2015 (UTC)

The following CSS seems to hide it (though I would appreciate it if a more experienced coder would check)

/* == Hide tag editing button & checkboxes == */
input[name^='ids['],
.mw-history-editchangetags-button { display: none; }

- Evad37 [talk] 05:09, 30 April 2015 (UTC)

I was the volunteer developer who implemented this, and I have to say I was never particularly comfortable with cluttering up people's history pages with a feature they are rarely likely to use. (Administrators won't really notice the difference, since they already see the column of checkboxes as part of the revision deletion feature.) Do you think the untagging feature should be restricted to sysops by default? Or should something different happen to it? — This, that and the other (talk) 10:09, 30 April 2015 (UTC)

Yes, restricting to sysops and removing the clutter for other editors seems like a good idea. In most cases, tags shouldn't be added or removed manually, as that defeats their purpose (software tracking and identifying potentially harmful edits). - Evad37 [talk] 12:51, 30 April 2015 (UTC)

Currently, tags applied by the software (e.g. AbuseFilter tags and tags like "visualeditor") cannot be manually removed. That feature is yet to be implemented; the only tags that can be removed are ones that are "applied by users and bots", of which we appear to have none here yet. So until the feature becomes more useful, I would not oppose a restriction of the relevant permission to sysops. — This, that and the other (talk) 13:05, 30 April 2015 (UTC)

Bots or scripts adding tags to their edits (presumably through the API) I can understand, but what would be the point of users manually adding tags (by pushing the edit tags button), and why should any of these "applied by users and bots" tags be easily removable by any user? - Evad37 [talk] 14:19, 30 April 2015 (UTC)

Just give it to sysops. There's really no reason for users to be removing tags, and it just adds clutter. Kharkiv07^Talk 16:01, 30 April 2015 (UTC)

Kharkiv07, just in case you missed it, not only is there no reason (yet) but also there are no tags to be added or removed, I have tried. I would say yes to hiding it until someone can come forward with a good use and management system. EoRdE6^{(Come Talk to Me!)} 18:07, 30 April 2015 (UTC)

That's what I meant, strange as I said it, regardless thanks for the clarification. Kharkiv07^Talk 18:22, 30 April 2015 (UTC)

I think this should be removed asap. To add that much clutter to an already-cluttered page needs a lot more justification than this has—or will ever have, even with more functionality. ―Mandruss ☎ 15:11, 1 May 2015 (UTC)

I agree. It reminds me of the first Andy Griffith Show episode, where they erected a stop sign at a place where they intended to build a road someday. ←Baseball Bugs ^{What's up, Doc?} carrots→ 18:05, 1 May 2015 (UTC)

This appears to have been removed by someone responding to our concerns (thank you whoever you are). At least it's gone from my page histories, and I haven't done anything to make that happen. I don't know whether it would be necessary to log out/in to pick up the change. ―Mandruss ☎ 18:57, 1 May 2015 (UTC)

So does anybody knows what is going on...? Because I would really like to see them, they refuse to show for me. -- [[User:Edokter]] {{talk}} 19:19, 1 May 2015 (UTC)

Seems they have been removed, possibly centrally. Gone from Commons too. Wish they would occasionally tell us whats going on or keep an updates board... EoRdE6^{(Come Talk to Me!)} 19:26, 1 May 2015 (UTC)

If you are an admin, the only difference on history pages was an extra button, adjacent to the existing Delete/undelete selected revisions button. I don't recall the wording. If you are not an admin, you got that extra button (level with the Compare selected revisions button, but positioned far right), plus a whole column of checkboxes just to the left of the time/date stamp. So the change was much more obvious for non-admins. --Redrose64 (talk) 20:30, 1 May 2015 (UTC)

@Edokter: phab:T97773 - the new features now hide themselves until an admin has set up some tags that can be added and removed. -- John of Reading (talk) 20:57, 1 May 2015 (UTC)

When was the WP:VPR discussion on this? ―Mandruss ☎ 21:13, 1 May 2015 (UTC)

This change was announced in m:Tech/News/2015/17. If you want to know what's happening to the software, you should subscribe and read it.

Volunteer devs aren't actually required to discuss their ideas for MediaWiki software in advance with this community (or any of the communities at the other other 800+ WMF wikis). Having 800+ discussions is just not practical. MediaWiki development is its own community. That community's discussions usually happen at mediawiki.org or on Phab:. You will find relevant links in the Tech/News announcement. WhatamIdoing (talk) 19:00, 5 May 2015 (UTC)

Most of us here are volunteers, but the rest of us are required to get something resembling a community consensus before making controversial changes that affect everyone. MediaWiki development does not in any sense represent the larger community. You're speaking of how things are; while I don't much expect to change that, there's nothing wrong with objecting, calmly and constructively, when things are not as they should be. ―Mandruss ☎ 19:07, 5 May 2015 (UTC)

I am also no longer seeing them. Thank you! ←Baseball Bugs ^{What's up, Doc?} carrots→ 21:43, 1 May 2015 (UTC)

• See original feature request. This was requested for bots and scripts, and admins when cleanup is required. Cenarium (talk) 15:14, 8 May 2015 (UTC)

See also discussion for global default at meta. The consensus on enwiki in the original discussion was to grant it to bots and sysops, which I'm also proposing as global default. Cenarium (talk) 15:18, 8 May 2015 (UTC)

• I personally see it as productive for TEs and AFs to also get the permissions needed to create and edit tags (managetags), and I've commented as such (or am about to) on every discussion I've seen that is relevant to this. It doesn't make sense that this wonderful new feature isn't accessible to EF people (who have been using tags long since before this was available to scripts and bots) and TEs who are more likely to create and maintain scripts and should be trusted enough to manage tags related to their scripts. I've also suggested something similar for if/when the new Gadgets 2.0 usergroup gets defined on MW: (IIRC). — {{U|Technical 13}} (e • t • c) 15:27, 8 May 2015 (UTC)

Weird things

I am using chrome browser on a phone. I have noticed some weird things that all started yesterday. I am not sure whether these are bugs or intensional changes and they all aren't present on the desktop version. I'll just list them:

• Several links (including the links to the side menu, the watchlist, the notifications side menu, edit buttons, the "This page has issues" link, the names of articles as presented on the page itself, and possibly others) are all of a sudden slightly bigger, slightly, smaller, or are spaced out differently.
• I reported a real deal bug here. I was going to report it here but I did it there.
• The mobile watchlist all of a sudden is outdated and wrong. For example It'll say a page changed 2 hours ago but if you go to the page you'll see it was actually 10 minutes ago.
• The writing on user contributions and recent changes pages on both desktop and mobile views, is all squashed together with no spaces in between.

There could be others I'm not thinking of right now; if I think of any others I'll add it to the list. —DangerousJXD (talk) 08:19, 1 May 2015 (UTC)

Thanks for the reports @DangerousJXD. I've created Phabricator tasks to cover all bugs but the one you reported over at Wikipedia talk:Notifications. The tasks are as follows: T97791, T97792, and T97794. You'll note that I've only been able to reproduce the icon sizing issue in the main menu. Could you provide an example of a page that has an incorrect last changed date? —Phuedx (WMF) (talk) 20:23, 1 May 2015 (UTC)

I would just like to add that the spinning page loading icon is another icon that is still smaller than normal.

I'm not sure how to give you an example of the watchlist thing, Phuedx (WMF). I will clarify the page history of pages isn't the issue, just the watchlist. If I go into the watchlist right now, the top article (this, not that it matters) says it was changed 2 days ago. On the page itself, it says 5 days ago. That is the best example I can give. When will these things be fixed? —DangerousJXD (talk) 21:55, 1 May 2015 (UTC)

I'm still having all these issues. The notifications system is broken; the rest are just little annoying things that I really don't care about. —DangerousJXD (talk) 11:55, 5 May 2015 (UTC)

Talk:Hillary Rodham Clinton/April 2015 move request subdivision

Talk:Hillary Rodham Clinton/April 2015 move request has gotten to be around 400k, leading to a complaint about the length. In order to ease that situation, I split off as subpages Talk:Hillary Rodham Clinton/April 2015 move request/Discussion and Talk:Hillary Rodham Clinton/April 2015 move request/Discussion/Closing and structural issues. This solution was reverted, so I am wondering, is there any way to prevent the page from causing loading issues and the like without splitting it up? bd2412 T 20:40, 2 May 2015 (UTC)

No. —TheDJ (talk • contribs) 20:59, 2 May 2015 (UTC)

I don't understand what the problem is or why everything has to be "centralized". Why not just follow another link and have two adjacent browser tabs? Dustin (talk) 21:02, 2 May 2015 (UTC)

That is what I tried to do; it was reverted by another editor who feels that it is important that everything be "centralized". bd2412 T 21:03, 2 May 2015 (UTC)

Many people have submitted their rationale for opposition or support of the proposal into the "discussion" section, as it was not clearly demarcated at the start. As a result, removing the discussion section would amount to balkanising the request, and marginalising the well-thought rationales that are found that in that section. Huw is just one of the editors who placed his whole opinion in the "discussion" section. RGloucester — ☎ 21:05, 2 May 2015 (UTC)

It's 2015, if someone can't load a 300k web page then that's not an important concern; time to stop catering to people running IE6 on windows 95 or Nokia flip-phones. Tarc (talk) 23:34, 2 May 2015 (UTC)

IBTD and actually support closing accounts of users thinking that mobile devices or old hardware are irrelevant. –Be..anyone (talk) 09:25, 3 May 2015 (UTC)

There's a balance, and people are asked to be considerate towards others. —TheDJ (talk • contribs) 13:20, 3 May 2015 (UTC)

"Considerate" is one thing, but breaking pages because of length in this era is just absurd. Should we continue support for lynx and NCSA Mosaic too? 300k is not inordinately large. Tarc (talk) 13:34, 3 May 2015 (UTC)

• Except that it isn't just a 300K web page, It's over 400K of wikitext and ends up being a 2.81 MB (2,949,120 bytes) web page all inclusive. — {{U|Technical 13}} (e • t • c) 17:27, 3 May 2015 (UTC)

• Where on earth are you getting almost 2.9M from? Even if you count all of the actual html and such behind the page...and that is not an accurate measure of page load at all...that comes out to 684k. Tarc (talk) 14:58, 5 May 2015 (UTC)

• First of all, I don't believe any editor has stated that they are having trouble actually loading the page. I believe the complaint/s was/is that the page is just too damn long. That can't be helped, for the most part. Since many editors have weighed in with their reasoning and policy arguments. If there are any loading problems, perhaps editors shouldn't insert misleading and irrelevant images on the page. Especially pictures that take up the whole page and push the boundaries of the page beyond the regular single page view. What we should not do is remove the discussion. Thanks. Dave Dial (talk) 14:01, 3 May 2015 (UTC)

  The proposal to split the page is neutral, as it would not seek to characterize the relevance of particular comments. bd2412 T 14:07, 3 May 2015 (UTC)

• I know you were trying to be neutral when you made the change, and your subsequent proposals. I just do not believe that splitting the page up is a good idea. Dave Dial (talk) 14:53, 3 May 2015 (UTC)

I like my picture thank you very much. I only used that size as unlike normal pictures it was actually the topic of that discussion and it was unreadable in thumbnail view. EoRdE6^{(Come Talk to Me!)} 16:00, 3 May 2015 (UTC)

• Because there is so much other stuff on every page such as css style sheets and javascript and the mediawiki UI itself, it's normal that people start having page size issues between 150-200K worth of wiki text. Pages really should never exceed 300K of wiki text if we want people to be able to actually load them and contribute. Currently, that page is about 2.9MB inclusive for me with 400K of wikitext. That is over the boundaries for some mobile devices. — {{U|Technical 13}} (e • t • c) 17:27, 3 May 2015 (UTC)

WP slow when saving changes

Anyone else having this issue? Seems to take an age to post any edits. Started this morning and continues to now. Takes approx. 10 seconds to save a change. Lugnuts ^{Dick Laurent is dead} 12:31, 3 May 2015 (UTC)

@Lugnuts: Same issue here. --NeilN ^{talk to me} 14:22, 3 May 2015 (UTC)

+1 ―Mandruss ☎ 14:42, 3 May 2015 (UTC)

I had this problem last night too... EoRdE6^{(Come Talk to Me!)} 15:57, 3 May 2015 (UTC)

Thanks both. Seems OK now. Lugnuts ^{Dick Laurent is dead} 16:50, 3 May 2015 (UTC)

Possible issue with "Automatically accepted" being improperly issued after reversions

I think I've found an issue with non-"Pending changes reviewer" editors getting their revisions marked "Automatically accepted" if they revert a previous reversion from a Pending changes reviewer. I noticed this happening yesterday at the article Charley Webb (though the issue seemed to go away when another Pending changes reviewer (Davey2010) reverted the reversion that a new editor (Bjcressy) had made to my reversion of their changes (did you get all that?!)). But now I've just noticed it again Sajal Ali, where another relatively new editor, Saqibbsse (talk · contribs), just reverted my reversion of their changes, and had their revision "Automatically accepted"). So, is this some kind of "bug" that needs dealing with?... Thanks in advance. --IJBall (talk) 16:23, 3 May 2015 (UTC)

It's more widespread than I suggested above – at Rebel Wilson, at least two editors who do not have "Pending changes reviewer" status have had their recent edits marked "Automatically accepted", including the most recent edit there by User:JosephSpiral. It seems there might be something kind of significantly wrong with the "Automatically accepted" tagging system currently... --IJBall (talk) 17:30, 3 May 2015 (UTC)

@IJBall: I suggest you read up on how pending changes works again. There are no bugs here, any editor who is auto confirmed (has account, ten days, x edits) will have their edits Automatically accepted. Only unconfirmed editors (IP's and less than 10day accounts) must have their edits reviewed, and that is the only purpose of the pending changes right, to review those edits. Having EVERY edit needing review is known as Pending Changes 2 (PC2) and is not implemented anywhere on en.wiki. EoRdE6^{(Come Talk to Me!)} 17:44, 3 May 2015 (UTC)

@EoRdE6:, thanks for the response, and I will admit that the PC1 vs. PC2 thing is something I was dimly aware off, but had forgotten the details of. However, that still doesn't explain yesterday's situation at Charley Webb where the Bjcressy (talk · contribs) edits were "automatically accepted" when the account had been registered that same day (i.e. certainly not for 10 days). But what you're saying probably does explain what happened at Rebel Wilson. --IJBall (talk) 17:52, 3 May 2015 (UTC)

Also, for context, previous discussion of the Bjcressy situation took place at the Help desk... --IJBall (talk) 17:55, 3 May 2015 (UTC)

Note also that User:Saqibbsse is also not on the "auto-confirmed" list yet either. --IJBall (talk) 17:57, 3 May 2015 (UTC)

@IJBall: Bjcressy registered on 2 May 2014 and made that edit on 2 May 2015, so yes, he was autoconfirmed. Jackmcbarn (talk) 18:01, 3 May 2015 (UTC)

@EoRdE6: PC2 is in use, on these three pages. They are test pages, it is true: but it means that the PC2 level is implemented somewhere on en.wiki. --Redrose64 (talk) 18:03, 3 May 2015 (UTC)

(edit conflict) @IJBall: (i) Who are the users at Rebel Wilson who should not have had their edits "automatically accepted"? (ii) There is no problem with edits by Saqibbsse showing as "automatically accepted": looking at the history of Sajal Ali, only this edit was shown as "automatically accepted". This was their twelfth edit, they had been registered for almost three weeks, and by that time had been autoconfirmed for 9 hours 50 mins. --Redrose64 (talk) 18:16, 3 May 2015 (UTC)

At Rebel Wilson, none – EoRdE6 pointed out that long-standing editors are autoconfirmed, and both of the "accepted" ones at Rebel Wilson were long-standing. I had thought that Bjcressy and Saqibbsse weren't autoconfirmed, but with what Jackmcbarn said, and rechecking, I see that both are auto-confirmed as well. Thus, there's nothing to report – which is why I wrote the 'mea culpa' below. Again, sorry for the confusion – now that I've been reminded about the details of PC1, it shouldn't happen again. --IJBall (talk) 18:20, 3 May 2015 (UTC)

Ack! Well, that's embarrassing!!... And I guess User:Saqibbsse has been registered for more than 10 days... Oh, well, sorry for wasting everyone's time! Lesson learned!! --IJBall (talk) 18:08, 3 May 2015 (UTC)

Email notifications not consistent

I've recently enabled notifications to send email alerts. It was working fine but in the past week, it failed to alert me for at least two notifications. One was a thanked edit (on 30 April) and the other was a talk page message (27 April). I've checked the known bugs at WP:ECHO and bugzilla but couldn't find anything related to this. In the email, searched through the categories including spam for those two email alerts. -Ugog Nizdast (talk) 17:14, 3 May 2015 (UTC)

Watchlist not loading

When I click on watchlist, it hangs. If I try to back out, I get logged out. Anyone else having this problem? Victoria (tk) 18:11, 3 May 2015 (UTC)

Roughly speaking, how many pages do you have on your watchlist? To the nearest thousand. --Redrose64 (talk) 18:31, 3 May 2015 (UTC)

Only 2000. After posting here, I got bumped off again. I can log in through the main page, I can see my contribs, it hangs if I try to go the watchlist, and I get logged out if I try to refresh or re-load. I've cleared caches, done a restart but it keeps happening. Was fine earlier in the day. Safari 6.1.6. Victoria (tk) 18:34, 3 May 2015 (UTC)

It might be a corrupt cookie. Try doing a deliberate logout, then log in. If it still logs you out without you asking, you could clear your cookies. There were two ways of doing this in Safari 5.1.7 (they may differ for 6.1.6), both start at Cogwheel → Preferences → Privacy. Then you can either: (i) go for Remove All Website Data... Remove Now or (ii) Details... and in the little search window, enter "wikipedia.org"; highlight the row that is shown and click Remove. --Redrose64 (talk) 19:11, 3 May 2015 (UTC)

Thanks, Redrose64 I've deleted all cookies, removed WP from "topsites" and from toolbar, emptied cache again, restarted again, and it still hangs. It might be something else, or something other than the watchlist. To log in I have to navigate to the main page from the log in page, and log in from there, and I can get here by following my contribs. As soon as I post, I get logged out. When I am logged in, I can't load the watchlist. Victoria (tk) 19:51, 3 May 2015 (UTC)

It sounds a bit like Wikipedia:Village pump (technical)/Archive 130#Problems logging in and Wikipedia:Village pump (technical)/Archive 132#Severe problems with log-in. I don't know what fixed it there. --Redrose64 (talk) 20:44, 3 May 2015 (UTC)

I get the change password message every day and I've been ignoring it. That, I think, is a separate issue.

So, in Monobook and Modern I cannot load the watchlist. In Vector and Cologne Blue I can. I have css and js pages for one of those skins but not both - so I leave you to you all to figure out. For now I've switched from Monobook to Vector and will see whether I get bumped out after making this post. Victoria (tk) 21:22, 3 May 2015 (UTC)

• I'm back and wondering if anyone here can help or tell me where to seek help. I've managed to log in via the mainpage and I changed my skin to Vector. Then I went to the raw watchlist, dumped it all into a text file, saved and was able to access the watchlist via all the skins. I then copied the watchlist back and got a message that there were too many entries - I don't know the exact count but it's between 1800 and 2000 pages - and now it's hanging with all the skins. I'm guessing there's a problem with the count. Suggestions? Help? Anything? Victoria (tk) 20:54, 4 May 2015 (UTC)
• Adding: through the watchlist tab on preferences I was able to clear the entire watchlist and it loads fine in all skins. Then I added back about 100 pages and got a message telling me there were too many pages. So basically I can't have a watchlist? Victoria (tk) 21:13, 4 May 2015 (UTC)

• Hi Victoria, the usual troubleshooting advice (and I have no idea whether this is the right thing to do here) is to check whether it's happening with other browsers. Next, set up an account (e.g. Victoriaearle1) with default preferences, then copy-paste your watchlist to it, or restore your main account's preferences to the default if you prefer, and see what happens. The latter will tell you whether it's something in your preferences that's causing a problem. Scripts that we add can cause clashes too. Is your monobook.js the same as your vector.js? Sarah (SV) ^(talk) 21:34, 4 May 2015 (UTC)

(edit conflict) 100 is very few. I've got over 20,000 on my watchlist and the main problem that I have is that when I go to Edit raw watchlist, alphabetical ordering breaks down about half-way through the list. Try visiting that page, copy its contents off somewhere, and then go for Clear the watchlist to force it to start from a blank sheet. --Redrose64 (talk) 21:40, 4 May 2015 (UTC)

Thanks, SlimVirgin I actually tried to see if I could log in as Truthkeeper88 and I could! I went through the same thing there, copied the watchlist out, it was fine, copied it back and it hung. Thanks Redrose64 for suggesting Clear the watchlist - I've always wondered what that was for. Will give it a try and report back. Right now I have a single page (this one) on watch and it's working. But would probably like to have a few more. Victoria (tk) 21:47, 4 May 2015 (UTC)

Okay, so I hit clear watchlist, and tried to load my watchlist and it hung. Had to force quit, log in again through the main page, go to preferences, at the raw watchlist deleted the one page that's there, this one, got the "there are too many pages" message. So something's wrong. Victoria (tk) 21:57, 4 May 2015 (UTC)

Forgot to ping above so adding. Btw - Redrose64 would it be worth trying to change the watchlist token? Victoria (tk) 22:34, 4 May 2015 (UTC)

I don't know. I never use one. --Redrose64 (talk) 23:03, 4 May 2015 (UTC)

Redrose64, I noticed one in preferences at the "Watchlist" tab with a link to change it, if necessary. I never created it, just assumed everyone had one. Haven't a clue how it got there. Would someone else know about this? Victoria (tk) 23:24, 4 May 2015 (UTC)

The watchlist token (in preferences) is not relevant. I have never used it, but if you visit your watchlist, in the "tools" box in the left sidebar, there is an "Atom" link. Copying the URL of the link into a text document shows that it contains the watchlist token—anyone who knows that token can view a web feed, presumably listing changes to articles that are on the watchlist. What is the exact text (copy/paste) of the "too many pages" error? It is sometimes useful to view the html source of a page to see if there are any clues in comments (for example, there is a NewPP report showing certain resource usage—conceivably there could be information with some technical detail about your error). Sometimes programmers re-use error messages for unlikely problems, so the actual error might be quite unrelated to the number of items on the list. If you have access to another computer where you have not logged in to Wikipedia, and if you are confident that computer is secure without malware, you should log out of your main computer, then go to the other computer and log in there. Does the problem recur? Did you try another browser on your main computer (preferably one where you have not logged on to Wikipedia before)? What is needed is a clear demonstration of whether the problem is related to corruption on your system, or whether your account is corrupted at Wikipedia (very unlikely). Johnuniq (talk) 02:15, 5 May 2015 (UTC)

Hi Johnuniq thanks for the help. Clicking the Atom brings up the dialogue box for Apple email, which I've never used.

Image is attached, showing the error message when I try to copy in the watchlist, and may or may not help for these many errors I found when looking at the source. When I tried save the watchlist the 16 or so errors jumped to 75 and then the machine hung again. I don't currently have access to a machine I've never used to access WP, but work will on that tomorrow. Maybe this img will give you a clue, if not, if there's anything there that's identifiable, pls delete it.

I only use one browser, but will download another tomorrow. Thanks for the suggestions. Victoria (tk) 03:01, 5 May 2015 (UTC)

Hi Johnuniq, more info. We do have a machine I've never used to log into WP, so I tried that. But because I've deleted my watchlist, there was none. So I logged back into my machine, copied in the watchlist from the text file, got the error message again, logged out and logged back into the other machine - and could see my watchlist. So, it's either a Safari issue, (the other machine is a PC w/ Google Chrome), or my machine. I'll download Google Chrome tomorrow to see if that works. If not, I guess I have corrupt data on my machine, which isn't good news. Thanks for the suggestions. Victoria (tk) 03:34, 5 May 2015 (UTC)

Re the Atom link: I was just trying to show what the purpose of the watchlist token is (it is to get a web feed of your watchlist which is not relevant for this discussion). There is no need to click the Atom thingy.

The screen shot shows that you are seeing MediaWiki:Watchlistedit-too-many. Searching all the MediaWiki source for "Watchlistedit-too-many" shows it is only used once, namely in SpecialEditWatchlist.php (that link is for anyone interested; nothing for a user there). That applies when editing a watchlist. I do not understand the source which outputs the error if the number of titles currently being processed is 100 or more, and someone more familiar with the source would need to see how the count could exceed that limit (it obviously works with watchlists that contain thousands of titles, so something should be splitting it up).

One thing I would suggest is that you experiment by clearing the watchlist, then adding very boring titles which use only simple English text with no accented characters or funny dashes. During these experiments, you may have pasted titles into the edit watchlist window with broken characters (to do it correctly, your system and your editor would need to be working in UTF-8 encoding). For example, use titles like 1 and 2 and Car. If you keep your computer switched on, you might try shutting it down with a full reboot. I don't expect either of these thoughts to help, but they are simple and rule out some things. Johnuniq (talk) 04:53, 5 May 2015 (UTC)

Hi Johnuniq, ok I tried that, adding a few titles at a time, very boring titles and it was working, so I got bolder and added titles with accents and so on, and all was well. I managed to slowly 1000 pages in batches with no problems, but when I tried to add the next 25, I noticed the alphabetization was lost and 8 were at the bottom of the edit window in the raw watch, out of order. I looked at the watchlist and those 8 had just been updated, 4 with with Kasparbot and the others general edits, which rearranged the sequence in the raw watchlist. So, I know I can get to 1000 but beyond that things get a little iffy. Don't know whether this is helpful. Thanks for the advice, btw. Victoria (tk) 16:16, 5 May 2015 (UTC)

Update - the alphabetization issue seemed to go away, I managed to get to 1300 pages and it came back. I got a new error message in source code and so deleted some scripts, [2], [3]. That brought back the edit notice, which I hadn't realized wasn't displaying, so that's good. But at 1300 pages I stopped adding them because whenever a page on my watch is edited it gets moved to the bottom of the list in the raw list window. So I downloaded Google chrome, and have the same problem. At this point I'm giving up. Victoria (tk) 00:11, 6 May 2015 (UTC)

Just to add during these experiments I found that I can still log in as Truthkeeper88 and under that name today have a complete and fully alphabetized watchlist. But yesterday it wouldn't load. Also under that acct all there are no scripts. Victoria (tk) 00:20, 6 May 2015 (UTC)

Your .css pages shouldn't be causing trouble like this, they didn't contain any scripts. common.css contained six CSS rules: the first two hide certain items on a watchlist; the third de-bolds some text in a watchlist; the last three affect how the bullets in a watchlist are displayed. monobook.css contained two CSS rules: both hide certain items on a watchlist.

All of those eight rules appear syntactically valid (some are identical to CSS rules that I use myself), and none can have caused the example error messages that you gave in that screenshot; they look very like they are from a JavaScript processor. I see that you have three JavaScript pages: of those, two (User:Victoriaearle/monobook.js and User:Victoriaearle/vector.js) are empty, which is never a problem; however, User:Victoriaearle/common.js is not, so try (i) removing the first script only; (ii) removing the second script only; (iii) blanking the page entirely. In this way you may identify one which causes errors. --Redrose64 (talk) 09:13, 6 May 2015 (UTC)

Thanks Redrose64, Johnuniq and SlimVirgin for helping with this; it's now resolved. I thought I'd blanked User:Victoriaearle/common.js but apparently missed it, and it was almost certainly the script for the orange bar, [4], that caused the problem. An additional problem, too, is that I'd turned on the gadget that prevented those green bullets from displaying, and looking at the source I found it was causing an error: Screenshot here.

After removing the gadget and the orange bar script, the banners reappeared and the watchlist is now loading. It's still not alphabetizing correctly but I don't often look at the raw watchlist so don't know if that's normal behavior or new.

Something to keep in mind, I suppose, is that given the many combinations of scripts + browsers + various other configs, what works for one person might not work for another. For whatever reason, my upper limit for a watchlist seems to be 2000 (I'd just gone beyond that when this happened and have now pared back). Anyway, thanks again for all the help and suggestions. Victoria (tk) 15:24, 7 May 2015 (UTC)

Background to first paragraph in last post above is at Wikipedia:Village pump (technical)/Archive 112#Reset code. The raw watchlist always used to sort in namespace order, and alphabetical order within each namespace. I first noticed that it was getting out of order in about March this year, but I put it down to excessive size. --Redrose64 (talk) 15:40, 7 May 2015 (UTC)

Thanks for linking to that thread. I'm not sure that had anything to do with the hang but it does throw up a warning as seen in the screenshot (the warning only displays when the gadget is selected). As for the raw watchlist sorting, I think it does have to do with size. It was consistently sorting alphabetically as I added pages in small batches; at 1000 pages it started to sort by most recent page edited. Whether that causes a hang, I don't know. I suspect the culprit was the script for the orange bar - I'd noticed it was sort of hiccuping before displaying. Anyway, again, thanks for all the suggestions. Victoria (tk) 15:59, 7 May 2015 (UTC)

CATEGORY:PAGES WITH LOGIN REQUIRED REFERENCES OR SOURCES

The all-caps in the heading is intentional. This category is populated when {{registration required}} is inserted into the |format= parameter of a CS1 template. The citation templates capitalize the format so if, for example, someone enters |format=pdf, the citation template will display "(PDF)".

I recently fixed all of the articles in this category, but new ones were added since then. So, I'd like to ask: Is this a proper use of |format=? If it is, how should this category be handled (category redirect, fixing the citation templates)? If not, is there an automated tool that is doing this? Anon126 (notify me of responses! / talk / contribs) 19:32, 3 May 2015 (UTC)

I don't know about an automated tool, but it's certainly a misuse of |format=, and also of {{registration required}}. The correct way is to either set the parameter |registration=yes inside the CS1 template, or to put {{registration required}} outside the CS1 template. --Redrose64 (talk) 19:38, 3 May 2015 (UTC)

We definitely don't want Category:PAGES WITH LOGIN REQUIRED REFERENCES OR SOURCES to be a visible red category on articles. I have created the category so it can be hidden and display an explanation. It would be possible for citation templates to test the format parameter for certain inappropriate values like the output of {{registration required}}, but I'm not sure it would be worth using server resources on that. PrimeHunter (talk) 23:42, 3 May 2015 (UTC)

Whatever you do, registration=yes and subscription=yes should be handled in a consistent way, e.g., if you want a hidden tracking category for "registration" it could include the worse case "subscription", or track the latter in a different way, depending on your goals. –Be..anyone (talk) 03:05, 4 May 2015 (UTC)

Might I suggest sending this to Help talk:CS1, since I see a feature request there in that suggestion? @Trappist the monk: --Izno (talk) 14:34, 4 May 2015 (UTC)

Tech News: 2015-19

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

Recent changes

• You can now chat with other users in Phabricator. [5]
• There was a rare problem with VisualEditor. The text of another wiki could be added to your edit. The issue is now fixed. [6] [7]

Changes this week

• The new version of MediaWiki has been on test wikis and MediaWiki.org since April 29. It will be on non-Wikipedia wikis from May 5. It will be on all Wikipedias from May 6 (calendar).
• It is now easier to add a link in VisualEditor. You can see an image and a description about the page. [8]

Meetings

• You can join a meeting with the Language team. The meeting will be on May 5 at 14:30 (UTC). [9]
• You can join the next meeting with the Editing team. During the meeting, you can tell developers which bugs are the most important. The meeting will be on May 6 at 18:00 (UTC). See how to join.
• You can join an event in France on May 22−25. You need to sign up before May 8. [10]

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

15:22, 4 May 2015 (UTC)

Question about toolserver

I've noticed that Wikiblame hosted at wikipedia.ramselehof.de always seems rock solid while the tools hosted at tools.wmflabs.org are about as stable as a bowl of jello. Are they run by the same group? --NeilN ^{talk to me} 20:43, 4 May 2015 (UTC)

Nope. Toolserver is run by the DE folks whereas the Labs server is hosted at by the WMF folks. Giraffasaurus (talk) 20:51, 4 May 2015 (UTC)

Any chance the WMF group can take lessons from the DE folks? Yes, I know this is pointed. --NeilN ^{talk to me} 21:00, 4 May 2015 (UTC)

Unfortunately, it's even worse, NeilN. WMDE is not running this, WMDE has decided to shut down the toolserver (with pressure "support" by WMF) and i don't think WMDE is in any position to give lessons. The reliable wikipedia.ramselehof.de is a website by a german Wikipedian, de:user:Flominator. And the reliable http://vs.aka-online.de (global Wikipedia search, WPPageHistStat /article edit history overview, RCHistStat) is also a website by a german Wikipedian, de:user:Aka. And the fantastically helpful scripts-compendium by de:user:Schnark is also... well, you get my point. And some toolserver-tools are gone that i really miss (mydiff/wikisense, and geolocation/coordinate and stats of gesichteteVersionen and... mw:Tool_Labs/Collection_of_issues_after_Toolserver_shutdown) :-( I better stop. --Atlasowa (talk) 08:32, 5 May 2015 (UTC)

Question about a script

@Jackmcbarn, Anomie, Writ Keeper, and Technical 13: Greetings folks, I found a useful script at User:Equazcion/CatListMainTalkLinks.js that adds a link when viewing categories to view the associated talk page or main page when the appropriate page is viewed in the category (Article shows up as Article (talk) and Article talk shows up as Talk:Article (main)). This is a very useful script to folks like me who add WikiProject banners (or to see if a talk page is missing its associated article page). The code seems to indicate that it should light the corresponding page in red if that page doesn't exist. The script does not appear to be working however and it does not show in red if its missing. I was going to leave User:Equazcion a note about it but it looks like they stopped editing and User:Technical13 says they are busy and suggests I talk to you abou it. Since its probably in appropriate to fiddle with somone else's code under their user page I copied it here. I'm not trying to claim ownership, I'm just trying to facilitate fixing the problem. I noticed there were some missing namespaces and added those, but I can't figure out why the applicable pages aren't showing in red like they are supposed too. Its probably something simple in the code but I don't see what it is.

Is there any chance you all might be willing to take a look at this page and see why its not working right? Sorry to post this again here but I thought it might be better since the others might be busy. Thanks in advance for the help. Please let me know if you have any questions. Giraffasaurus (talk) 20:49, 4 May 2015 (UTC)

@Giraffasaurus: It's probably because the API call is specifying http, whereas Wikipedia gets served by default as https; thus, the API call gets blocked by the browser since it's insecure. Replacing:

url:"http://en.wikipedia.org/w/api.php?action=query&titles=" + encodeURIComponent(cloneTitle) + "&format=xml",

with:

url:mw.util.wikiScript('api') + "?action=query&titles=" + encodeURIComponent(cloneTitle) + "&format=xml",

should fix it, I think. Writ Keeper ⚇♔ 23:15, 4 May 2015 (UTC)

Outstanding! That did fix the problem thank you. It works like a champ now. You can mark this problem as resolved. Giraffasaurus (talk) 23:29, 4 May 2015 (UTC)

I've copied over the changes to the main script at User:Equazcion/CatListMainTalkLinks.js. Graham87 10:38, 6 May 2015 (UTC)

Clickable button 2 and images

Hello,

I am using a Clickable button and have spent 2 hours attemping to insert an image into it, with noincludes, includes, transclusions, you name it, to no avail. This is the closest I've come so far:

Talk

Is it possible to remove the "leftovers" surrounding the button, i.e. the brackets and text? Buffaboy ^talk 23:23, 4 May 2015 (UTC)

@Buffaboy: Try |link=, like [[File:Example.png|frameless|link=Whatever]] Zhaofeng Li [talk ♦ contribs] 01:05, 5 May 2015 (UTC)

@Zhaofeng Li: I've tried that already. Buffaboy ^talk 03:50, 5 May 2015 (UTC)

Buffaboy, what do you actually want? I looked at User:Buffaboy/Navbar/Talk, but I can't figure out what your goal is. An image that takes you to a page? Or—tell me what URL you want to end up with, and (separately) show me what you want it to look like. Whatamidoing (WMF) (talk) 19:19, 5 May 2015 (UTC)

@Whatamidoing (WMF): I'm trying to use a {{clickable button 2}} to link to my talkpage but I would like to embed an image within it. I think limitations with Lus and the template itself prevent this however.Buffaboy ^talk 19:24, 5 May 2015 (UTC)

Buffaboy, why do you want to do that?

If you tell me what the end result is supposed to look like, and what's supposed to happen when someone clicks on it, then I might be able to give you a solution. (It won't use that particular template, though, because that template doesn't do what you want.) Whatamidoing (WMF) (talk) 19:37, 5 May 2015 (UTC)

Use {{Clickable button}}. {{Clickable button 2}} automatically inserts a pair of square brackets at the start and end, with the assumption that you're providing a wikilink. This behaviour cannot be overridden. Alakzi (talk) 19:26, 5 May 2015 (UTC)

Job queue

What's up with it? I started checking it when I got frustrated with links not updating. It was like 3 million (which I thought was very high), but then it steadily grew over the last week to current 15.7 million. That's almost half of Wikipedia's pages! What's going on? That does not sound like your run-of-the-mill slowdown. Renata (talk) 03:09, 5 May 2015 (UTC)

Is someone needed to make 15 million edits? I'll do it. Cheers! bd2412 T 03:26, 5 May 2015 (UTC)

19.9 million now... I do agree that seems odd its growing so fast, but not knowing what it normally rests at I couldn't tell you if that is normal or not... EoRdE6^{(Come Talk to Me!)} 14:28, 7 May 2015 (UTC)

• I'm poking in -operations on IRC to try and get the job queue turned back on. Apparently, it was accidentally disabled last week. — {{U|Technical 13}} (e • t • c) 15:33, 7 May 2015 (UTC)
  • User:Technical 13, it's still steadily going up. 21 million now. Renata (talk) 16:08, 8 May 2015 (UTC)
    • Continued above in Very slow category population from stub templates. — {{U|Technical 13}} (e • t • c) 16:50, 8 May 2015 (UTC)
      • Filed as phab:T98621. --Glaisher (talk) 17:26, 8 May 2015 (UTC)

Broken anchor links

Templates {{anchor}} and {{vanchor}} (and probably others) can be used as mentioned at WP:ANCHOR. Several discussions have argued about where {anchor} should be placed:

• Before the heading (where subsequent editing will almost certainly damage or move the anchor).
• After the heading (then, clicking a link with the anchor sometimes does not show the heading at the target, although someone attempted to fix that at some stage by inserting some html for a vertical offset, I think).
• In the heading (which works perfectly for readers, but is ugly for editors, and which leads to broken links in diffs and history pages).

An IP (contribs) has changed a lot of anchors, and wants to do more. The IP has changed Template:Anchor/doc (three edits) and Template:Visible anchor/doc to say that {anchor} should not be used in a heading.

My question for VPT concerns whether there is any fix for the broken link issue in the last point above. Examples:

• This diff shows "(→{{visible anchor|Public address}}es: no reference)" at the top of the right-hand side, and the arrow is a link. However, the link is broken because the anchor portion is #.7B.7Bvisible_anchor.7CPublic_address.7D.7Des (should be #Public_addresses). The same edit is shown on the history as the last of these three, and the link in the arrow is similarly broken.
• This diff shows the same link problems using {anchor}.

I like using span tags for anchors, but they are a inexplicable mumbo-jumbo for subsequent editors. Is there a solution to the broken links problem? Could the problems, at least in principle, be fixed with something like a new magic word to add anchors? Johnuniq (talk) 08:18, 5 May 2015 (UTC)

Mostly, the anchors I’ve changed were simply wrapping the section heading, i.e. == {{visibleanchor|Lorem ipsum}} ==, which is all kinds of wrong. But I have no objection to my other anchor edits being reverted if there’s no better alternative. I thought it was common for alternate anchors to be included beneath the heading, though. —174.141.182.82 (talk) 09:02, 5 May 2015 (UTC)

That's not something we need to consider here (and I agree that anchors-in-the-heading are a pain for editors). I gave an outline only to list the issues—what is needed is a technical solution that avoids the problems mentioned. Johnuniq (talk) 09:22, 5 May 2015 (UTC)

I agree this is a problem that could have a technical fix. I'd like to note that if I link to the anchor Wikipedia:Manual of Style#Section management, the result does not show the section heading and is likely to be confusing to the reader, but if I link to the shortcut MOS:SECTIONS, which is on the same line in the edit window, it goes to the right place. However in another example, this anchor Wikipedia:Manual of Style#US again doesn't show the heading, but neither does this shortcut MOS:U.S.. The wiki text looks the same as the previous example, but the performance is different. So sometimes it seems like {{Shortcut}} works better than {{Anchor}} but not always.  SchreiberBike | ⌨  17:28, 5 May 2015 (UTC)

MOS:US redirected to {{anchor|US}} rather than the section title. I’ve just fixed it to match the heading. I don’t think {{shortcut}} includes an anchor itself (in the case of US, it’d be a duplicate anchor anyway), but relies on redirects being properly set. Correction: {{shortcut}} does include anchors, but using the full shortcut name, e.g. WP:MOS#MOS:U.S. takes you to the shortcut box at MOS:US. Anyway, I think the behavior of {{anchor}} varies by browser; on mine (Mac/Safari 8.0.5), such links land slightly above the section headings, farther up than direct section links. —174.141.182.82 (talk) 17:58, 5 May 2015 (UTC)

You have probably seen Template talk:Anchor#Positioning. I gave up on that a year ago, but you are welcome to tackle it. -- Gadget850 ^talk 21:08, 5 May 2015 (UTC)

Accessing image when obscured by popup caption

Tracked in Phabricator
Task T93393

A reader noticed a problem in a specific article, but I believe the problem is likely to be more widespread. In some cases, it may be difficult or impossible to open up a larger version of an image. To see an example, go to the Galley in Roman Baths Roman_Baths_(Bath)#GalleryIf you hover a mouse over any image, a popup will display a caption. In some cases, you can still click on the image, to access the larger version of the image itself.

With my monitor and resolution, I see the following:

Roman baths screen shot 1

My mouse is over the second image in the second row. There is enough of an image displayed to click on it. However, the situation may be different for different monitors or screen resolution settings. I simulated a different setting by opening in a window that was slightly less than full screen. In this case, the popup obscures the image so that it is impossible to click on the image.

Roman baths screen shot 2

I do not see an easy way to get to the image. (There is a difficult way, go to the source and find the file name, but this would not constitute a reasonable work-around for general readers.) While this reader identified one particular situation, it seems likely that this would be a common problem, especially for diligent editors who provide descriptive captions. Any thoughts on a solution? --S Philbrick(Talk) 13:36, 5 May 2015 (UTC)

@Sphilbrick: Click on the previous image, then use the back button to return to the Roman Baths article. In some browsers, the border of that image is now styled differently - in Firefox, it's a fine-pitch dotted line. Use the Tab ↹ key twice to get to the image in question, which should now have that border styling, and press ↵ Enter. --Redrose64 (talk) 14:49, 5 May 2015 (UTC)

Frankly, I didn’t follow your instructions. (I did see that if I clicked on one of the other images, I can use the right and left scroll arrows to cycle though the images.) However, my point is, unless the right answer is obvious to most readers, and the two of us just failed to know the obvious, we have a situation where casual readers will not find it easy to access an image. I think we need a technical solution, e.g. some way to render the caption so that it is offset or presented in some way that the reader can click on the image. I expect a Wikipedia editor to know how to get to the image, but do you agree that our main audience is very likely to be stymied?--S Philbrick(Talk) 15:10, 5 May 2015 (UTC)

This really needs to be addressed at the Wikimedia developer's level. Far too often we devise overly-complex workarounds for problems in the way the Wikimedia software behaves. They are being paid to fix this sort of problem.

That being said, here is my favorite overly-complex workaround for this sort of problem: :) In FireFox there is an add-on called "Nuke Anything" ( https://addons.mozilla.org/en-US/firefox/addon/nuke-anything-enhanced/ ) It allows you to delete anything you can click on or highlight with the mouse or shift + arrows. It also lets you highlight a selection and delete everything else -- perfect for printing a page without the banner ads at the top or the long list of stupid comments at the bottom. --Guy Macon (talk) 16:03, 5 May 2015 (UTC)

It looks like this was reported about six weeks ago as phab:T93393. Whatamidoing (WMF) (talk) 19:28, 5 May 2015 (UTC)

Thanks, that is very helpful. I asked the person reporting it to monitor this page. I see some specific workarounds that may help that reader in the short-term, and it now is clear that the general problem is known to the developers. I realize it may take some time to address, but I trust it will eventually be fixed. — Preceding unsigned comment added by Sphilbrick (talk • contribs) 20:18, 6 May 2015 (UTC)

Hiring technical people

The Community Tech team, whose remit is doing stuff that experienced editors actually need done, has started hiring. If you're interested or know someone who ought to be interested, then please look at the job posts for the Community Tech Developer and Community Tech Engineering Manager positions. It's usual in the WMF for community members to be given preference in the hiring process (just like multilingual candidates are always preferred), but I believe that is even more true for this team. If you apply, be sure to include your username. Whatamidoing (WMF) (talk) 19:35, 5 May 2015 (UTC)

Bug reporting. Excess whitespace, trying to render Mongolian script

Tracked in Phabricator
Task T98565

There's a browser bug that means that when I try to load any page with Mongolian text what I get is a large block of whitespace, for me specifically its a problem with Chrome running on Android 4.4.2. I've tried the pages with Firefox, and things seem to render just fine with Firefox.--KTo288 (talk) 21:48, 5 May 2015 (UTC)

Ridiculous

Mobile web Chrome. I made this edit; now I have been experiencing a lot of issues. (1) Beta was enabled. (2) Picture loading was disabled. (3) I had this weird box full of technical writing at the top of pages (it is gone now; after I clicked on it, everything got worse). (4) I get a prompt nonstop that my edit was saved even though I'm not making edits (severe). (5) Pages are taking longer to load as if there is something new being loaded (this is the worst). (6) The page isn't loaded properly for a second then it is fine. (7) There is a big space on edit summary pages. (8) The page in which I thank editors is a mess. (9) Of course there is the notifications system bug above. I have tried logging out (while I was logged out testing things for a few minutes, all these things are still happening). I have tried making a null edit. I have tried clearing some "website settings" in Chrome. I tried setting my preferences to default. I tried turning off my phone. Of course everything is all fine on desktop view. It is not my phone because other websites are fine. This needs to be resolved. —DangerousJXD (talk) 23:25, 5 May 2015 (UTC)

It would be great if you could take a screenshot of what you see. Also, have you tried using another mobile browser? Is this on Android or iOS? Additionally, though you are experiencing a trial in your life, it's best not to demand results. Editing via mobile Wikipedia is new and full of bugs. Patience is required. Killiondude (talk) 03:45, 6 May 2015 (UTC)

Killiondude, thanks for responding. I have the high tech knowledge of a stereotypical 70 year old so uploading a screen shot is a no no. I just tried using the browser on my phone called "browser" but it's the same. It is Android. I am sorry if it sounded like I was "demanding results". —DangerousJXD (talk) 03:58, 6 May 2015 (UTC)

We do have a guide at WP:WPSHOT, but I don't know how applicable it is to mobiles. --Redrose64 (talk) 14:42, 6 May 2015 (UTC)

It would be too hard to time it right to take a picture of the 4th thing I listed for example. As I said, uploading pictures is not for me. Notifications appear to be fine now. 1 and 2 are obviously fine now (because I turned them off). 3 as I said is gone (that is what I think made the first 2 items happen. 4 and 5 are the most important thing here. 6, 7 and 8 are still a thing but they are minor compared to 4 and 5. —DangerousJXD (talk) 22:10, 6 May 2015 (UTC)

I can report that the 4th and 6th issues I reported in my initial post appear to be have been fixed. —DangerousJXD (talk) 04:22, 7 May 2015 (UTC)

I have discovered that the loading issue is (most of the time) non-existent for the loading involved when you save an edit; otherwise, the loading issue is still happening. It is a pain to navigate. —DangerousJXD (talk) 08:56, 8 May 2015 (UTC)

Google book citation tool problem?

Dear editors: I have been regularly using the Google book citation tool at http://reftag.appspot.com . Recently I tried to click on some of the links in the citations that I had previously created, only to find that they don't work (for example the first two references in Wikipedia talk:Articles for creation/Matthew Murray (writer). Is this tool no longer functioning properly, or has Google changed its URL structure, or is there some other explanation? Has this problem already been noted, and, if so, can someone point me to the discussion? —Anne Delong (talk) 03:11, 6 May 2015 (UTC)

Oh my. It's not the tool but Google books itself giving a "500 error The server encountered an error and could not complete your request". I just tried and couldn't access some random books. I assume Google will fix this quickly. -Ugog Nizdast (talk) 04:26, 6 May 2015 (UTC)

For me (in the UK) http://books.google.com/books?id=ZsjpAwAAQBAJ&pg=PA124 fails and http://books.google.co.uk/books?id=ZsjpAwAAQBAJ&pg=PA124 is working. Thincat (talk) 09:12, 6 May 2015 (UTC)

Well, that's a relief in one way, since it's good to know it's not a glitch in the tool, but let's hope that Google puts things back the way they were. I am remembering when the Canadian Encyclopedia decided to reorganize the structure of their database and changed all the URLs, without redirects. This was rather embarrassing, since I discovered it by demonstrating in front of a group of genealogists at a live streamed meeting, how the citation tool could be useful... —Anne Delong (talk) 11:57, 6 May 2015 (UTC)

This is being discussed at the Google Books API forum here because the correct forum, linked to from here doesn't exist. This is perhaps the best thread. Google will be looking into the matter when they have finished their present game of DotA. Thincat (talk) 08:05, 7 May 2015 (UTC)

My Samsung 10.1" 32gb tablet is only displaying Geo = {IPv6: true} when I go to wikipedia on it

Greetings! My wife was looking something up on Wikipedia and when she followed a link from Google to the page, she encountered this: GEO = {"city" : "Fresno" . . . lots of stuff including our actual latitude and longitude and my tablet's IP! Needless to say, she was perplexed. We're, neither of us, super technical people, so she handed the tablet off to me. She told me she was asked about "being a Guinea Pig" for something in Wikipedia, but she assured me she responded "No". She also said she MAY have tapped something within the Wikipedia page. I have no idea what it could have been to change the display to radically. In any case, I did what everyone does. I restarted the tablet. LOL Now, instead of a long trail of pretty personal information, I get Geo = {IPv6: true} Is there some way to change the display in Wikipedia back to the nice, normal, incredibly informative view or are we trapped with me being forever told that IPv6 is true? LOL Thank you very much. — Preceding unsigned comment added by Sak1776 (talk • contribs) 10:14, 6 May 2015 (UTC)

A similar problem has been reported a couple of times on the Help Desk. Those specify that it has something to do with the Chrome browser. Dismas|^(talk) 11:37, 6 May 2015 (UTC)

Hey! That is very weird! Certainly a bug. I've opened a bug on phabricator: phabricator.wikimedia.org/T98311 to talk about the issue. Phabricator is where bug reports go and where programmers deal with them and fix them. Feel free to go there and post whatever you need (you can use your Wikipedia account). It would be very useful if you could post there (or here) more information. We would need the address/url where you saw that text (or if it happens everywhere), and information about what is your browser and version, and what is your operating system and version. If you don't know where to look those things up, the easiest way is going to whatismybrowser.com and whatismyandroidversion.com from the tablet's internet browser. With that information it should be easy to replicate and we'll get to fix the problem. Thanks! JHernandez (WMF) (talk) 11:46, 6 May 2015 (UTC)

It seems like the bug was already identified and has been fixed and deployed, so you shouldn't experience the problem any more. See https://phabricator.wikimedia.org/T98309 for context. JHernandez (WMF) (talk) 08:10, 7 May 2015 (UTC)

Short footnotes not showing up in citation counts

I've taken to using {{sfn}} and Harvard style references in articles I'm heavily involved in, and ordinary <ref></ref> now for ordinary footnote-worthy side commentary. I notice that when I check articles I created on Special:NewPagesFeed the articles with just sfn's and no ref's are showing up as having "No citations". See for example John Elder (pastor) versus 7½ Cents. This doesn't actually fool patrollers. For all I know it's red meat that draws their attention. I assume it would not be hard to count {{sfn}} and its variants. Choor monster (talk) 11:25, 6 May 2015 (UTC)

Not to mention pages that use parenthetical referencing. If it were a pure count of <ref>...</ref>, then even if {{sfn}} were taken into account, Actuary would fail: but it's a featured article. --Redrose64 (talk) 14:39, 6 May 2015 (UTC)

The point being that the feed is to help reviewers, not to do the job for them. The presence of a flag doesn't mean the article is bad; the absence doesn't make it good. If your article is good (as John Elder (pastor) very obviously is) then you have nothing to worry about. Relentlessly (talk) 16:08, 6 May 2015 (UTC)

I wasn't worried. But thanks for the comments.

Thinking of NPP, are redirects turned into articles patrolled? Choor monster (talk) 11:24, 7 May 2015 (UTC)

No, because they're not new pages. The patrolling will have been done when the redir was first created, if its creator had the autopatroller right; otherwise, at some point within the following thirty days. --Redrose64 (talk) 16:34, 7 May 2015 (UTC)

That makes sense, but it does seem like a loophole in the patrolling process. Choor monster (talk) 14:22, 8 May 2015 (UTC)

"Collections", mobile front beta

I use the beta version of the Mobile FrontEnd on my smartphone, Samsung Verizon Android Galaxy S-3, SCH-I535 OS v4.4.2. In the menu accessed by the three-horizontal-lines icon at the top left there is an item "Collections". When I click on it, it shows a list of just one item, "Watchlist (private)".

The implication is that I can develop and add more collections and make some of them public, but there's no indication of how I would do any of that. Searching for Collections or WP: Collections or Wikipedia help: Collections gets me nowhere. What is this crypto-feature about? To discuss this, please {{Ping}} me. --Thnidu (talk) 17:33, 6 May 2015 (UTC)

@Thnidu: See mw:Extension:Gather. PrimeHunter (talk) 20:57, 6 May 2015 (UTC)

Thanks, I guess. And now the "Collections" entry is gone altogether.* In case anyone else wonders what happened to it, I've put up a brief summary at Wikipedia:Collections, with a wikilink to mw:Gather and a xref to this section. --Thnidu (talk) 02:41, 7 May 2015 (UTC)

* The"Collections" item is back on the menu, apparently still as I described it. --Thnidu (talk) 04:59, 7 May 2015 (UTC)

"Tagging information" on the mobile site

I was reading a Wikipedia article on my mobile device when a box popped up asking me to "Improve Wikipedia by tagging information on this page". I followed through with it, curious what it was. It asked me to "Select tags that correctly describe article name". After, it told me that "Wikipedia is experimenting with new ways for people to contribute...Entries are submitted to Wikidata and are freely licensed under...". Curious, I checked my Wikidata contributions (since I was logged in) and the page history of the Wikidata page about the article. Neither had any record of what I had just done... So what is this feature and what purpose does it serve? Thanks! EoRdE6^{(Come Talk to Me!)} 18:41, 6 May 2015 (UTC)

See mw:Extension:WikiGrok. It includes: "Right now (during prototype testing) claims are recorded via EventLogging (and are not yet being pushed to Wikidata)". PrimeHunter (talk) 20:48, 6 May 2015 (UTC)

Wikidata date errors

Seeing these on a few pages, suspect there are a lot more that will appear in the category as it updates. Lemery, Batangas, Orangeville, Ontario and Lake of Bays all have identical script errors due to trying to access date information to do with population data. But except 1 the pages haven't changed, the data hasn't changed, the script hasn't changed in the last couple of days in any way that I can see that would cause it. Has something changed in WP's Lua implementation that might be the cause?--JohnBlackburne^words_deeds 23:17, 6 May 2015 (UTC)

Try pulling the data directly in the article using #property; it returns nothing. Alakzi (talk) 03:14, 7 May 2015 (UTC)

This is an problem in the module, Module:Wikidata. The module fetches an timestamp from wikidata, but looks for the day in said timestamp in the wrong place. In line 240 in said module the module thinks that the day is "0:", which is actually a part of the time in the timestamp. This string then gives an error at line 247. Now, this error does not show up anymore because the module is no longer used on those pages.--Snaevar (talk) 15:54, 7 May 2015 (UTC)

Manila, Parañaque (fixed those, using local values for now) and possibly a few more articles have the same broken function - not sure how to search for that specific module call, only a few of the 3000+ Wikidata-transclusions (all with the mentioned date functionality) seem to be affected. GermanJoe (talk) 16:31, 7 May 2015 (UTC)

They appear in Category:Pages with script errors, seven now. It's only ones accessing dates, and seemingly only for census dates. I notice Lake of Bays has been fixed by removing the 'broken' code but it's not the code that's broken, that code had been in since March not generating errors. It also seems unlikely to be the data for the same reason, it mostly hasn't changed recently. The script was last modified by RexxS, and that change might be related as it’s to do with the date, but reverting (checked with a preview) doesn't seem to help. Besides it was seven days ago which is a long time for a script change to percolate to the error category.--JohnBlackburne^words_deeds 17:02, 7 May 2015 (UTC)

It would be a long time, in the normal course of things; but as Technical 13 noted elsewhere on this page, the job queue has been down for some days, so when a template or module is changed, the changes are only propagating through to articles as and when those articles are edited; and only at that point are the error messages shown and cats populated, one article at a time. --Redrose64 (talk) 20:16, 7 May 2015 (UTC)

(copied from Wikipedia:Templates for discussion/Log/2015 May 7:) Sorry, I've been out all day, campaigning in the UK General Election. It seems that the way Wikidata stores dates has changed in the last week. As you can see at Module talk:Wikidata #Module_talk:Wikidata, I dumped the claims for European green toad and showed that P574.datavalue.value.time was stored as "+00000001768-00-00T00:00:00Z" on 29 April, but if you try {{#invoke:Wikidata|Dump|claims}} today, you get the time as "+1768-00-00T00:00:00Z". Because we didn't have the mw formatting functions when I initially wrote the module, I relied on pulling year, month, day, etc. out of the zero-padded string using substrings at fixed positions. Of course now that the year is in the second to fifth characters, instead of in the ninth to twelfth characters, we get errors. Many thanks to Alakzi for spotting the problem and applying a fix. Using the current mw date format functions should harden the code against changes to the way data is stored and provide a less error-prone module in future. --RexxS (talk) 20:56, 7 May 2015 (UTC)

Looking for volunteers to help us label edits as damaging/not

A screenshot of the "wiki labels" interface for evaluating a diff

Hey folks!

If you have not been following our work, we have been working on developing the infrastructure that will help introduce smarter Artificial Intelligence based tools to Wikipedia that utilize the most cutting edge Machine Learning algorithms.

Hence, we are looking for volunteers to help us build new and better quality control tools for Wikipedia. We need help manually assessing the quality of edits to train machine learning models that will help with quality control work. We've built a fancy tool to make the work quick, easy and maybe even fun. :) If you're interested, check out our project page WP:Labels/Edit quality.

-- A Certain White Cat ^chi? 08:20, 7 May 2015 (UTC)

Someone's fucked around with the UKIP page

https://en.wikipedia.org/wiki/UK_Independence_Party

This is a big issue given it is election day. I can't find the offending diff, anyone else, (quickly this article is getting lots of views today). EoRdE6^{(Come Talk to Me!)} 15:18, 7 May 2015 (UTC)

No clue how it was done, I tried reverting the last edit but that did nothing (feel free to undo that now that it's fixed)

ElectrifiedSpork (talk) 15:19, 7 May 2015 (UTC)

Someone want to clue the rest of us in what you were seeing, even if not how? DMacks (talk) 15:21, 7 May 2015 (UTC)

The article starts to load normally, then is overlain or replaced in some way with a "Vote Labour" message and a big picture of Ed Milliband. DuncanHill (talk) 15:22, 7 May 2015 (UTC)

Ah yes, several admins reverted and protected a few things to stop that. DMacks (talk) 15:24, 7 May 2015 (UTC)

http://i.imgur.com/3km8RFv.png is what it looked like. Care to explain what happened? Must've been someone with admin privileges. ElectrifiedSpork (talk) 15:25, 7 May 2015 (UTC)

It wasn't an admin. And the accusation is insulting to the admins who have been cleaning up the vandalism. - X201 (talk) 15:46, 7 May 2015 (UTC)

Someone edited a page that was transcluded, not that page itself. DMacks (talk) 15:28, 7 May 2015 (UTC)

Template:UK_MP_links, among others, were the culprits. Are all protected now, so no WP:BEANS there. Offending users have been blocked. Mamyles (talk) 15:30, 7 May 2015 (UTC)

UK election page, too

The page 56th_United_Kingdom_general_election also shows that partisan political ad. Attempt to view the revision diff redirect to the same ad, so can't revert. The problem vanished as I logged in. This does not seem to be a birthday attack, reloading does not help. This is a big problem, needs a dev fast. HLHJ (talk) 15:30, 7 May 2015 (UTC)

Purging the page fixed it, just the template vandalism. Template editor protection is implemented and IP's are getting proxy blocks. SHould be getting better. Keep and eye on them all though. EoRdE6^{(Come Talk to Me!)} 15:35, 7 May 2015 (UTC)

Related Changes in the left sidebar will help tack down this sort of vandalism. -- Gadget850 ^talk 16:38, 7 May 2015 (UTC)

Another way is to install Anomie's script, see User:Anomie/previewtemplatelastmod, and then on the page with the problem, go for the "Edit this page" tab (not a section edit link). Go to the bottom, and under "Pages transcluded onto the current version of this page (help):" (which you may need to expand), the most-recently edited templates and modules will be listed first. The culprit will probably be the first one listed. --Redrose64 (talk) 20:06, 7 May 2015 (UTC)

Displaytitle with quotation marks

Currently Template:DISPLAYTITLE doesn't seem to have the quotation mark option and article titles that typically require quotation marks per manual style are displayed without them (e.g. "Marry the Night"). The quotation marks are not listed among prohibited characters at Wikipedia:Naming_conventions_(technical_restrictions)#Forbidden_characters and I don't see any prior discussion of that at template's talk. In something like {{DISPLAYTITLE:"Marry the Night"}} the quotation marks are ignored and the title is displayed without them. Are there any actual technical restrictions to that or there were prior objections? Brandmeister^talk 15:44, 7 May 2015 (UTC)

Unless the quotation mark is a part of the proper title itself, e.g. "Weird Al" Yankovic, I don't think you'd use them. Formatting text within the article body isn't the same as formatting an aviatrices title orper. Tarc (talk) 15:51, 7 May 2015 (UTC)

Quotation marks which are not actually in the page's title are invalid per mediawikiwiki:Help:Magic words#Displaytitle. --Izno (talk) 18:01, 7 May 2015 (UTC)

If we wanted the page title to be "Marry the Night", we could move the page to that title. However, the Manual of Style does not recommend quotation marks in titles. —Remember the dot ^(talk) 18:29, 7 May 2015 (UTC)

There was a RFC, with a fairly large amount of input, about this just last January. You can see the consensus that was reached here Wikipedia talk:Article titles/Archive 50#RFC: Quotation marks in displayed article titles. MarnetteD|Talk 18:49, 7 May 2015 (UTC)

need help with userpage design strategies

So, basically, I'm creating an awesome userpage to showcase all of the work I'm doing, and it will be an incredible upgrade over my current one. However, I'm not too experienced with CSS tables or CSS in general, and I'm really not that experienced with tables in Wiki markup.

Here is where the work is being done, and I have had an issue trying to restore the gold interior container as seen in this diff. Plus I can't seem to get the padding right anymore. The reason I have made it more complicated now is because I've attempted to pare down the "tabs" seen in the diff, just so I could have one wide column with about 3 or 4 boxes and a narrow column with a couple boxes, including one with a scrollable list of infoboxes.

I've put a lot of time into it (probably too much) and it is taking away from article editing time. Anyone have any suggestions to fix the page so I can continue? Buffaboy ^talk 15:58, 7 May 2015 (UTC)

Template:R from ASCII

Four category pages display Template:Redirects from titles without diacritics/Explanation which states when Template:R from ASCII and Template:R to ASCII should be used, but they are deprecated, so the advice seems to need updating. – Fayenatic London 17:21, 7 May 2015 (UTC)

Paine Ellsworth, have you seen this? I can't remember where the central venue is for R from ... templates. --Redrose64 (talk) 18:55, 7 May 2015 (UTC)

Hi, ya'll – yes, I am aware of this situation and want to look into it. It seems to me to need more study, as I've found several items on the What links here pages of the rcats that don't appear to fall into the instructions found on the /Explanation subpage. Another seemingly mangled mess that I want to untangle and provide more clarification. I haven't yet taken the opportunity to give it the time it deserves. – Paine  19:09, 7 May 2015 (UTC)

How to adapt the languages list?

• I posted this question in the Miscellaneous section; they replied I should post it here in the Technical section. Anyhow they suggested it could be done by creating the CSS below. I tried but it doesn't work yet. Ceinturion (talk) 21:31, 7 May 2015 (UTC)

The list of language editions at the left of each Wikipedia page is a set of exotic languages that rarely contains the three languages that are useful for me (English, Dutch, German). Is there a way to adapt the list to my needs? Ceinturion (talk) 07:12, 7 May 2015 (UTC)

@Ceinturion: This is really a WP:VPT question. However, it's possible to selectively hide interlanguage links that you're not interested in using CSS, if you know the language codes; to hide all except a select few is also possible, but slightly more complicated - what we do is to hide them all and then selectively unhide the ones that you want to see. For example, to hide all except the links to the English, Dutch and German Wikipedias, which are language codes "en", "nl" and "de" respectively, you would use

/* Hide all interlanguage links - Vector, MonoBook and Modern skins */
li.interlanguage-link {
  display: none;
}
/* Unhide interlanguage links for English, Dutch and German - Vector, MonoBook and Modern */
li.interlanguage-link.interwiki-en,
li.interlanguage-link.interwiki-nl,
li.interlanguage-link.interwiki-de {
  display: list-item;
}

which would be added to Special:MyPage/common.css. --Redrose64 (talk) 12:12, 7 May 2015 (UTC)

Or to meta:Special:MyPage/global.css to work at all wikis. PrimeHunter (talk) 12:26, 7 May 2015 (UTC)

I tried both but it doesn't work yet. Any suggestions? Ceinturion (talk) 21:53, 7 May 2015 (UTC)

What is your browser and skin? For me in Firefox on Windows Vista it works in all skins except Cologne Blue. PrimeHunter (talk) 23:54, 7 May 2015 (UTC)

I am using IE 11 on Windows 7, and skin Monobook. Neither did the changed common.css work on another browser (Chrome 42) and other skins (Vector, Modern, Cologne Blue) either. I clicked on (Ctrl) Reload. Ceinturion (talk) 08:53, 8 May 2015 (UTC)

All skins except Cologne Blue also work for me with IE 9 and Chrome 42 on Windows Vista. IE 9 is the latest version which runs on Vista. PrimeHunter (talk) 11:17, 8 May 2015 (UTC)

My code works for Vector, Monobook and Modern skins, in Firefox, Chrome, Opera, Safari and IE8. It doesn't work for Cologne Blue, because the "list" of interlanguage links is actually a single paragraph. In Cologne Blue, the links may be hidden and selectively unhidden, by using CSS based on the above - by altering all the li. to span. and the list-item to inline - that is, use

/* Hide all interlanguage links - Cologne Blue skin */
span.interlanguage-link {
  display: none;
}
/* Unhide interlanguage links for English, Dutch and German - Cologne Blue skin */
span.interlanguage-link.interwiki-en,
span.interlanguage-link.interwiki-nl,
span.interlanguage-link.interwiki-de {
  display: inline;
}

This is specific to one skin, so it's best placed in Special:MyPage/cologneblue.css; however, the result looks messy, because the pipes and non-breaking spaces between those links can't be hidden (unless we resort to JavaScript). It works in Firefox, Chrome, Opera, Safari and IE8. --Redrose64 (talk) 12:18, 8 May 2015 (UTC)

Never finishes loading

My problem came up a week or so ago and was deemed to be phab:T66721. Pages just don't finish loading, or if they do, it takes over a minute of spinning blue "loading" circles, with no apparent change. Sometimes (less than half the time), a forced reload worked.

The proposed workaround was to double the local storage. I did that, but it still (almost) never finishes loading pages! While I'm typing this, it says "Transferring data from upload.wikimedia.org". Reading a diff on a talk page gives me "Transferring data from commons.wikimedia.org".

This only happens to me in Firefox 37/Mac OS 10.10. In Safari 8 (different account, probably fewer scripts) I don't have this problem. Any ideas what I should try next? WhatamIdoing (talk) 04:33, 8 May 2015 (UTC)

I'm having the same problem with Firefox 37.0.2/Windows 7 and have noticed this oddity occurring myself. What user scripts/gadgets are you using? I have a sneaking suspicion it's one of them but I want to see if there's overlap first. --Izno (talk) 04:53, 8 May 2015 (UTC)

This happens to me too! I switched from Firefox to Chrome for this very same reason, it still didn't stop. I'm currently on Chrome 41.0.2272.118/Linux. I tried commenting out all my scripts too, didn't work. I concluded that it was my network problem but now seeing this post makes it more confusing. It seems to be random, some days it's horrible while on others it works fine. -Ugog Nizdast (talk) 10:33, 8 May 2015 (UTC)

I had this a while back and it was due to having the reflinks script in my Custom JavaScript page. Try removing that text and see what happens. Lugnuts ^{Dick Laurent is dead} 13:15, 8 May 2015 (UTC)

Table Alignment

Is there any way to set a table's alignment to ‘none’ so that it can be placed above floated multimedia objects that are above it in the source code (see Labour Party (UK)#Electoral_performance for an example of the problem I'm trying to solve)? Esszet (talk) 14:48, 8 May 2015 (UTC)

try using {{stack}} with the other objects. -- Gadget850 ^talk 14:54, 8 May 2015 (UTC)